70
Xenopus
performed with egg extracts, the critical role of Axin in
regulating pathway activity was deciphered. The ODE-based
model was later extended by including Dkk1 and Axin2
feedback showing that the pathway may show oscillatory
behavior under some conditions (Wawra et al., 2007).
A more complex mathematical model describes a Wnt signaling network with negative cross-regulation of canonical
and non-canonical Wnt signaling during dorso-ventral axis
formation in Xenopus. Wnt11b was earlier shown to activate CamkII and to be required for ventral development
in Xenopus (Kuhl et al., 2000). However, Wnt11b was also
shown to be required for dorsal development in Xenopus
(Tao et al., 2005). This contradiction was resolved by the
observation that Wnts can activate different signaling
branches in a concentration-dependent manner; high concentrations of Wnt3a favored canonical, whereas low concentrations favored non-canonical, Wnt signaling (Nalesso
et al., 2011). Together with the cross-inhibition of different
Wnt signaling branches, this implicated a switch-like behavior of the Wnt signaling network (Kestler and Kuhl, 2011).
Indeed, the function of Wnt11b during dorso-ventral axis
formation could be recapitulated in a mathematical model
that also refects the fndings of gain- and loss-of-function
studies in Xenopus (Strang et al., 2017). Taken together, this
work showed that mathematical models in combination with
experiments in Xenopus embryos can be used to gain new
insights into the mechanisms underlying Wnt-mediated signal transduction.
Mathematical models also proved helpful for analyzing
gene regulatory network architecture involved in embryonic pattern formation. For example, during early heart
development, Wnt6 signaling is involved in patterning the
lateral plate, cardiogenic mesoderm into heart muscle (myocardium, toward the inside), and non-muscular heart tissue (pericardium, toward the outside) (Lavery et al., 2008).
Mathematical modeling alerted us that we were missing an
additional important component, which led to study of the
role of Sfrp1 in this context. We demonstrated that Sfrp1
interacts with Wnt6 in an intricate gene regulatory network
(Gibb et al., 2013): Wnt6 signaling from the ectoderm adjacent to the cardiogenic mesoderm promotes nearby pericardium differentiation and conf nes sfrp1 expression and
myocardium differentiation toward the inside of the cardiogenic mesoderm. While sfrp1 is initially only expressed
in a few cells beyond the reach of strong Wnt6-mediated
repression, after being secreted, Sfpr1 in turn inhibits Wnt6
signaling and thus reduces the reach of this Wnt6-mediated
repression (and pericardium differentiation) and thereby
increases the myocardium domain.
Mathematical modeling opened our eyes. While previously we had been constrained by our Wnt6-centered perspective, modeling provided us with an alternative outlook:
wnt6 expression just provides positional information about
where and how far away the edge of the cardiogenic mesoderm is located. sfrp1 expression is really in charge by using
this positional information to carve out an appropriately
sized and positioned heart muscle.
6.3.3. CONTEXT-SPECIFIC WNT SIGNALING
Wnt signaling is one of remarkably few molecular cell-to-cell
signaling pathways that are used repeatedly during embryonic development, both in different embryonic tissues and at
different stages of development. The early Xenopus embryo
provided us with an experimentally accessible model system
to investigate mechanisms determining context-specif c Wnt
signaling function.
In one example, before zygotic gene activation (ZGA),
maternal Wnt signaling promotes subsequent dorsal embryonic cell fate (illustrated in the famous axis duplication essay
discussed previously), yet after ZGA, zygotic Wnt8a functions to promote essentially the opposite, subsequent ventral
(and lateral) mesodermal cell fate (reviewed by Zylkiewicz
et al., 2014). Dorsal-promoting maternal and ventral-promoting zygotic Wnt8a signaling are both mediated by canonical
Wnt/β-Catenin pathway mechanisms (Hamilton et al., 2001);
thus, the relevant context-specifc Wnt signaling mechanisms
are to be found downstream of β-Catenin in the regulation of presumably two different classes of direct Wnt target
genes: direct maternal Wnt/β-Catenin target genes (normally
expressed early in prospective dorsal cells) and direct zygotic
Wnt8a/β-Catenin target genes (normally expressed later in
prospective ventrolateral mesoderm). Since nuclear β-Catenin
is considered the hallmark of active canonical Wnt signaling
(see previously), genome-wide transcriptome analysis (RNAseq) was combined with mapping of physical β-Catenin protein association to gene loci on chromosomes (ChIP-seq) to
identify direct Wnt/β-Catenin target genes comprehensively in
these two contexts, before and after ZGA (Afouda et al., 2020;
Nakamura et al., 2016). Wnt signaling-regulated physical
β-Catenin protein association to gene loci on chromosomes is,
surprisingly, not suffcient for transcriptional regulation. Wnt
signaling initiates β-Catenin association to many gene loci,
with additional context-specifc mechanisms combining with
Wnt signaling to determine which of these β-Catenin associated genes are expressed for the correct context-specif c Wnt
target gene response (Nakamura and Hoppler, 2017).
Two different classes were originally expected in early
Xenopus development: dorsal/maternal and ventral/zygotic
direct Wnt target genes. Our analysis suggests not two but a
useful defnition of about fve different classes of direct Wnt
target genes. This includes two classes of maternal dorsal
Wnt target genes (both co-regulated by Nodal signaling but
with genes in the slightly later expressed class additionally
regulated by products of the slightly earlier expressed class
in a feed-forward regulatory loop, Afouda et al., 2020); a few
universal Wnt target genes directly regulated by both dorsal
maternal and ventral zygotic Wnt signaling (including axin2
and sp5); and two classes of specifc zygotic Wnt8a/β-Catenin
target genes (Nakamura et al., 2016), one class co-regulated
by BMP signaling (see also Hoppler and Moon, 1998) and one
class co-regulated by FGF signaling (Haremaki et al., 2003).
This reveals quite a remarkable complexity of Wnt target genes for just the early stages of Xenopus embryonic
development.
Xenopus
performed with egg extracts, the critical role of Axin in
regulating pathway activity was deciphered. The ODE-based
model was later extended by including Dkk1 and Axin2
feedback showing that the pathway may show oscillatory
behavior under some conditions (Wawra et al., 2007).
A more complex mathematical model describes a Wnt signaling network with negative cross-regulation of canonical
and non-canonical Wnt signaling during dorso-ventral axis
formation in Xenopus. Wnt11b was earlier shown to activate CamkII and to be required for ventral development
in Xenopus (Kuhl et al., 2000). However, Wnt11b was also
shown to be required for dorsal development in Xenopus
(Tao et al., 2005). This contradiction was resolved by the
observation that Wnts can activate different signaling
branches in a concentration-dependent manner; high concentrations of Wnt3a favored canonical, whereas low concentrations favored non-canonical, Wnt signaling (Nalesso
et al., 2011). Together with the cross-inhibition of different
Wnt signaling branches, this implicated a switch-like behavior of the Wnt signaling network (Kestler and Kuhl, 2011).
Indeed, the function of Wnt11b during dorso-ventral axis
formation could be recapitulated in a mathematical model
that also refects the fndings of gain- and loss-of-function
studies in Xenopus (Strang et al., 2017). Taken together, this
work showed that mathematical models in combination with
experiments in Xenopus embryos can be used to gain new
insights into the mechanisms underlying Wnt-mediated signal transduction.
Mathematical models also proved helpful for analyzing
gene regulatory network architecture involved in embryonic pattern formation. For example, during early heart
development, Wnt6 signaling is involved in patterning the
lateral plate, cardiogenic mesoderm into heart muscle (myocardium, toward the inside), and non-muscular heart tissue (pericardium, toward the outside) (Lavery et al., 2008).
Mathematical modeling alerted us that we were missing an
additional important component, which led to study of the
role of Sfrp1 in this context. We demonstrated that Sfrp1
interacts with Wnt6 in an intricate gene regulatory network
(Gibb et al., 2013): Wnt6 signaling from the ectoderm adjacent to the cardiogenic mesoderm promotes nearby pericardium differentiation and conf nes sfrp1 expression and
myocardium differentiation toward the inside of the cardiogenic mesoderm. While sfrp1 is initially only expressed
in a few cells beyond the reach of strong Wnt6-mediated
repression, after being secreted, Sfpr1 in turn inhibits Wnt6
signaling and thus reduces the reach of this Wnt6-mediated
repression (and pericardium differentiation) and thereby
increases the myocardium domain.
Mathematical modeling opened our eyes. While previously we had been constrained by our Wnt6-centered perspective, modeling provided us with an alternative outlook:
wnt6 expression just provides positional information about
where and how far away the edge of the cardiogenic mesoderm is located. sfrp1 expression is really in charge by using
this positional information to carve out an appropriately
sized and positioned heart muscle.
6.3.3. CONTEXT-SPECIFIC WNT SIGNALING
Wnt signaling is one of remarkably few molecular cell-to-cell
signaling pathways that are used repeatedly during embryonic development, both in different embryonic tissues and at
different stages of development. The early Xenopus embryo
provided us with an experimentally accessible model system
to investigate mechanisms determining context-specif c Wnt
signaling function.
In one example, before zygotic gene activation (ZGA),
maternal Wnt signaling promotes subsequent dorsal embryonic cell fate (illustrated in the famous axis duplication essay
discussed previously), yet after ZGA, zygotic Wnt8a functions to promote essentially the opposite, subsequent ventral
(and lateral) mesodermal cell fate (reviewed by Zylkiewicz
et al., 2014). Dorsal-promoting maternal and ventral-promoting zygotic Wnt8a signaling are both mediated by canonical
Wnt/β-Catenin pathway mechanisms (Hamilton et al., 2001);
thus, the relevant context-specifc Wnt signaling mechanisms
are to be found downstream of β-Catenin in the regulation of presumably two different classes of direct Wnt target
genes: direct maternal Wnt/β-Catenin target genes (normally
expressed early in prospective dorsal cells) and direct zygotic
Wnt8a/β-Catenin target genes (normally expressed later in
prospective ventrolateral mesoderm). Since nuclear β-Catenin
is considered the hallmark of active canonical Wnt signaling
(see previously), genome-wide transcriptome analysis (RNAseq) was combined with mapping of physical β-Catenin protein association to gene loci on chromosomes (ChIP-seq) to
identify direct Wnt/β-Catenin target genes comprehensively in
these two contexts, before and after ZGA (Afouda et al., 2020;
Nakamura et al., 2016). Wnt signaling-regulated physical
β-Catenin protein association to gene loci on chromosomes is,
surprisingly, not suffcient for transcriptional regulation. Wnt
signaling initiates β-Catenin association to many gene loci,
with additional context-specifc mechanisms combining with
Wnt signaling to determine which of these β-Catenin associated genes are expressed for the correct context-specif c Wnt
target gene response (Nakamura and Hoppler, 2017).
Two different classes were originally expected in early
Xenopus development: dorsal/maternal and ventral/zygotic
direct Wnt target genes. Our analysis suggests not two but a
useful defnition of about fve different classes of direct Wnt
target genes. This includes two classes of maternal dorsal
Wnt target genes (both co-regulated by Nodal signaling but
with genes in the slightly later expressed class additionally
regulated by products of the slightly earlier expressed class
in a feed-forward regulatory loop, Afouda et al., 2020); a few
universal Wnt target genes directly regulated by both dorsal
maternal and ventral zygotic Wnt signaling (including axin2
and sp5); and two classes of specifc zygotic Wnt8a/β-Catenin
target genes (Nakamura et al., 2016), one class co-regulated
by BMP signaling (see also Hoppler and Moon, 1998) and one
class co-regulated by FGF signaling (Haremaki et al., 2003).
This reveals quite a remarkable complexity of Wnt target genes for just the early stages of Xenopus embryonic
development.