131
Evolution of the Vertebrate Neural Crest
Kalcheim 1999; Hall 2008a, 2018). Importantly, many of the
same transcription factors and signaling pathways that play
roles in controlling the neural crest stem cell state are redeployed to direct transit to specifc lineage states, including SoxD/E families, Foxd3, Wnt, and BMP. However, in
this new context, these transcription factors and pathways
must control a new set of regulatory targets that will promote and characterize a differentiated state. As the list of
neural crest-derived structures is vast, here we focus on two
derivatives where work in Xenopus has made fundamental
contributions that have informed work in other models: the
craniofacial skeleton and melanocytes. The full set of neural
crest derivatives have been comprehensively reviewed elsewhere (Le Douarin and Kalcheim 1999; Hall 2008a, 2018;
Schock and LaBonne 2020).
The formation of the neural crest-derived head skeleton
requires an early phase of patterning followed by differentiation via chondrogenesis and, eventually, osteogenesis
(Gross and Hanken 2005). The three-dimensional structure
of the head skeleton is delineated early in development by
Wnt and BMP signaling (Jacox et al. 2016). These pathways
create a Cartesian coordinate system of transcription factors
and other signaling molecules that interact along the dorsalventral (DV) and anterior-posterior (AP) axes. The function
of this system is to specify cell type and structural identity
of specifc skeletal elements. For the DV axis, this involves
nested and combinatorial expression of dlx, alx, msx, prrx,
emx, and hand transcription factors. The precise combination of these nested expression patterns is what confers skeletal identity. For example, hand1, hand2, dlx1, dlx2, dlx5 , and
dlx6 are all expressed in the ventral pharynx and lower jaw
precursors, whereas a dlx1/dlx2-positive and hand1/hand2negative domain is expressed dorsally and generates the
upper jaw skeleton (Square et al. 2015). Similarly, expression of barx, nkx, gdf5, and zax genes, along with signaling by Satb2 and Endothelins, prefgures the jaw joint in the
intermediate pharynx (Newman et al. 1997; Reisoli et al.
2010; Square et al. 2015; Lukas and Olsson 2018; Lukas et
al. 2020). Finally, recent work has shown that intercellular
signaling by the Endothelin pathway—one of just a few signaling pathways that is truly unique to vertebrates—is crucial for establishing the DV patterning program, a feature
conserved between Xenopus and lamprey, a jawless vertebrate (Bonano et al. 2006; Kawasaki-Nishihara et al. 2011;
Square et al. 2016, 2020).
A similar patterning role for hox genes occurs along
the AP axis. An anterior hox-negative expression domain
generates oropharyngeal elements such as Meckel’s cartilage, with increasingly nested hox expression domains more
posteriorly prefguring the gill-bearing ceratobranchial
cartilages. The importance of this AP hox expression code
in Xenopus was demonstrated by knockdown and overexpression experiments. When hoxa2 was over-expressed in
post-migratory neural crest, homeotic transformations were
observed wherein the anterior jaw skeleton of the f rst arch
was transformed into a hyoid skeleton, a second-arch derivative (Pasqualetti et al. 2000). Complete knockdown of the
entire f rst hox paralogous group (hoxa1, hoxb1, and hoxd1)
caused more severe defects, including an inability of cranial
neural crest to colonize the pharyngeal arches (McNulty et
al. 2005).
After patterning, differentiation of cranial neural crest
cells requires expression of several transcription factors
known to be important for chondrogenesis and osteogenesis
across vertebrates (Square et al. 2016). These include members of the SoxE (Sox8, Sox9), SoxD (Sox5, Sox6), and Runx
(Runx2) transcription factor families and FGF, BMP, and
retinoic acid signaling pathways (Golub et al. 2000; Eimon
and Harland 2001; Spokony et al. 2002; Lee et al. 2004;
Kerney et al. 2007; Tahir et al. 2014). Interactions between
neural crest cells and surrounding tissues (mesoderm and
endoderm) are also required for proper differentiation, and
this likely refects the requirement for continued input from
signaling pathways (Seufert and Hall 1990). Overall, these
tissue interactions and transcription factor activities promote
expression of proteins (e.g. Collagens, Aggrecans, Lectins)
that mark the transition from progenitor regions to differentiated skeletal elements (Seufert et al. 1994; Evanson and
Milos 1996; Kerney et al. 2010).
Melanocytes are responsible for creating the colorful patterns of pigmentation observed in the skin, feathers, and scales
of phylogenetically diverse vertebrates, including amphibians (Dupin and Le Douarin 2003). In Xenopus, as in other
vertebrates, the transcription factor Sox10 acts as a key regulator of melanocyte development by activating expression of
mitf and is also required, along with Endothelin signaling, for
proper melanocyte migration (Aoki et al. 2003; KawasakiNishihara et al. 2011). Indeed, neural crest-derived melanocyte precursors in the head are readily identifed in Xenopus
embryos days before they actually differentiate by expression of the endothelin receptor, ednrb2 (Square et al. 2016).
The progression from melanocyte precursor to fully differentiated melanocyte in both Xenopus and Rana proceeds
along the dorsal-ventral axis controlled by an antagonism
between melanocyte-stimulating and melanocyte-inhibiting
factors and their corresponding receptors, and defects in
these features may underlie the albinism of natural Xenopus
mutants (Fukuzawa and Ide 1987; Fukuzawa and Bagnara
1989). Signaling interactions between neural crest-derived
melanocyte precursors and the surrounding tissues through
which they migrate strongly infuence both the migratory
routes of these cells and their subsequent patterns of melanophore pigmentation (MacMillan 1976; Milos and Wilson
1986; Frunchak and Milos 1990).
8.4. THE ORIGINS OF NEURAL
CREST POTENTIAL
One of the most remarkable features of the neural crest is its
ability to produce such a large and diverse array of cell types
and structures. This impressive developmental potential is
most obvious when considering that neural crest cells produce derivatives associated with both ectodermal (neurons)
and mesodermal (cartilage, bone) germ layers and also make
Evolution of the Vertebrate Neural Crest
Kalcheim 1999; Hall 2008a, 2018). Importantly, many of the
same transcription factors and signaling pathways that play
roles in controlling the neural crest stem cell state are redeployed to direct transit to specifc lineage states, including SoxD/E families, Foxd3, Wnt, and BMP. However, in
this new context, these transcription factors and pathways
must control a new set of regulatory targets that will promote and characterize a differentiated state. As the list of
neural crest-derived structures is vast, here we focus on two
derivatives where work in Xenopus has made fundamental
contributions that have informed work in other models: the
craniofacial skeleton and melanocytes. The full set of neural
crest derivatives have been comprehensively reviewed elsewhere (Le Douarin and Kalcheim 1999; Hall 2008a, 2018;
Schock and LaBonne 2020).
The formation of the neural crest-derived head skeleton
requires an early phase of patterning followed by differentiation via chondrogenesis and, eventually, osteogenesis
(Gross and Hanken 2005). The three-dimensional structure
of the head skeleton is delineated early in development by
Wnt and BMP signaling (Jacox et al. 2016). These pathways
create a Cartesian coordinate system of transcription factors
and other signaling molecules that interact along the dorsalventral (DV) and anterior-posterior (AP) axes. The function
of this system is to specify cell type and structural identity
of specifc skeletal elements. For the DV axis, this involves
nested and combinatorial expression of dlx, alx, msx, prrx,
emx, and hand transcription factors. The precise combination of these nested expression patterns is what confers skeletal identity. For example, hand1, hand2, dlx1, dlx2, dlx5 , and
dlx6 are all expressed in the ventral pharynx and lower jaw
precursors, whereas a dlx1/dlx2-positive and hand1/hand2negative domain is expressed dorsally and generates the
upper jaw skeleton (Square et al. 2015). Similarly, expression of barx, nkx, gdf5, and zax genes, along with signaling by Satb2 and Endothelins, prefgures the jaw joint in the
intermediate pharynx (Newman et al. 1997; Reisoli et al.
2010; Square et al. 2015; Lukas and Olsson 2018; Lukas et
al. 2020). Finally, recent work has shown that intercellular
signaling by the Endothelin pathway—one of just a few signaling pathways that is truly unique to vertebrates—is crucial for establishing the DV patterning program, a feature
conserved between Xenopus and lamprey, a jawless vertebrate (Bonano et al. 2006; Kawasaki-Nishihara et al. 2011;
Square et al. 2016, 2020).
A similar patterning role for hox genes occurs along
the AP axis. An anterior hox-negative expression domain
generates oropharyngeal elements such as Meckel’s cartilage, with increasingly nested hox expression domains more
posteriorly prefguring the gill-bearing ceratobranchial
cartilages. The importance of this AP hox expression code
in Xenopus was demonstrated by knockdown and overexpression experiments. When hoxa2 was over-expressed in
post-migratory neural crest, homeotic transformations were
observed wherein the anterior jaw skeleton of the f rst arch
was transformed into a hyoid skeleton, a second-arch derivative (Pasqualetti et al. 2000). Complete knockdown of the
entire f rst hox paralogous group (hoxa1, hoxb1, and hoxd1)
caused more severe defects, including an inability of cranial
neural crest to colonize the pharyngeal arches (McNulty et
al. 2005).
After patterning, differentiation of cranial neural crest
cells requires expression of several transcription factors
known to be important for chondrogenesis and osteogenesis
across vertebrates (Square et al. 2016). These include members of the SoxE (Sox8, Sox9), SoxD (Sox5, Sox6), and Runx
(Runx2) transcription factor families and FGF, BMP, and
retinoic acid signaling pathways (Golub et al. 2000; Eimon
and Harland 2001; Spokony et al. 2002; Lee et al. 2004;
Kerney et al. 2007; Tahir et al. 2014). Interactions between
neural crest cells and surrounding tissues (mesoderm and
endoderm) are also required for proper differentiation, and
this likely refects the requirement for continued input from
signaling pathways (Seufert and Hall 1990). Overall, these
tissue interactions and transcription factor activities promote
expression of proteins (e.g. Collagens, Aggrecans, Lectins)
that mark the transition from progenitor regions to differentiated skeletal elements (Seufert et al. 1994; Evanson and
Milos 1996; Kerney et al. 2010).
Melanocytes are responsible for creating the colorful patterns of pigmentation observed in the skin, feathers, and scales
of phylogenetically diverse vertebrates, including amphibians (Dupin and Le Douarin 2003). In Xenopus, as in other
vertebrates, the transcription factor Sox10 acts as a key regulator of melanocyte development by activating expression of
mitf and is also required, along with Endothelin signaling, for
proper melanocyte migration (Aoki et al. 2003; KawasakiNishihara et al. 2011). Indeed, neural crest-derived melanocyte precursors in the head are readily identifed in Xenopus
embryos days before they actually differentiate by expression of the endothelin receptor, ednrb2 (Square et al. 2016).
The progression from melanocyte precursor to fully differentiated melanocyte in both Xenopus and Rana proceeds
along the dorsal-ventral axis controlled by an antagonism
between melanocyte-stimulating and melanocyte-inhibiting
factors and their corresponding receptors, and defects in
these features may underlie the albinism of natural Xenopus
mutants (Fukuzawa and Ide 1987; Fukuzawa and Bagnara
1989). Signaling interactions between neural crest-derived
melanocyte precursors and the surrounding tissues through
which they migrate strongly infuence both the migratory
routes of these cells and their subsequent patterns of melanophore pigmentation (MacMillan 1976; Milos and Wilson
1986; Frunchak and Milos 1990).
8.4. THE ORIGINS OF NEURAL
CREST POTENTIAL
One of the most remarkable features of the neural crest is its
ability to produce such a large and diverse array of cell types
and structures. This impressive developmental potential is
most obvious when considering that neural crest cells produce derivatives associated with both ectodermal (neurons)
and mesodermal (cartilage, bone) germ layers and also make