sign in sign up
<<
Home , Cardiac neural crest, Cranial neural crest, Eye development, Mesoderm, Neural crest, Rhombomere

Seminars in & Developmental 16 (2005) 683–693

Review Specification of cell formation and migration in mouse Paul A. Trainor ∗

Stowers Institute for Medical Research, 1000 E. 50th Street, Kansas City, MO 64110, USA

Available online 25 July 2005

Abstract

Of all the model organisms used to study development, rodents such as mice most accurately reflect human craniofacial development. Collective advances in mouse and mouse continue to shape our understanding of neural crest cell development and by extrapolation the etiology of human congenital and facial birth defects. The aim of this review is to highlight the considerable progress being made in our understanding of cell patterning in mouse embryos. © 2005 Elsevier Ltd. All rights reserved.

Keywords: Mouse; Neural crest; Craniofacial; Induction; Migration

Contents

1. Introduction ...... 683 2. Mouse neural crest cells: formation, migration, differentiation ...... 684 3. Specification of mouse neural crest cell formation ...... 685 4. Specification of mouse neural crest ...... 687 5. Specification of multipotency versus restricted potency of mouse neural crest cells ...... 690 6. Conclusions ...... 690 Acknowledgements ...... 690 References ...... 691

1. Introduction form very early during craniofacial development and they generate the majority of the , , connective and The craniofacial complex is anatomically the most sophis- peripheral in the head (Fig. 1). Not only are ticated part of the body and to function properly it requires neural crest cells crucial to head development but they are the orchestrated integration of the viscerocranium and also synonymous with craniofacial evolution. neurocranium, the central and peripheral nervous systems, Craniofacial abnormalities are largely attributed to defects facial muscles, , vasculature and . in the formation, migration and differentiation of neural crest Not surprisingly this integration often goes awry such that cells and the origins of particular congenital syndromes can craniofacial abnormalities are the most common congenital be traced back specifically to problems in one or more of these malformation constituting at least a third of all birth defects. phases of neural crest cell development. For example, First Neural crest cells, are a migratory population that Arch syndrome broadly describes craniofacial abnormalities characterized by malformation of the , , lower jaw ∗ Tel.: +1 816 926 4414; fax: +1 816 926 2051. and palate. Treacher Collins syndrome and Pierre Robin syn- E-mail address: [email protected]. drome are two of the more extreme examples falling into this

1084-9521/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.semcdb.2005.06.007 684 P.A. Trainor / Seminars in Cell & 16 (2005) 683–693

Fig. 1. Migration and differentiation of neural crest cells. Sox10 in situ hybridization (purple stain) labels the distinct segregated streams of neural crest cells as they leave the of mouse embryos at 8.5 dpc (A) and as they condense to contribute to the cranial ganglia at 9.5 dpc (B). Neurofilament immunohistochemistry highlights the differentiation of neural crest cells in the cranial ganglia into sensory that project axons into the branchial arches at 10.5 dpc (C). Neural crest cells also give rise to the majority of the cartilage (blue) and bone (red) that constitute the viscerocranium and neurocranium (D, sagittal view; E, superior view). category. The clinical abnormalities associated with Treacher earliest known markers of neural crest cell formation and the Collins and Pierre Robin syndromes are thought to arise due onset of their expression has been used to study the spatial to defects in the migration of neural crest cells. In contrast, and temporal competence of the to initiate neural craniosynostoses, which are characterized by the premature crest induction in response to different signals. fusion of the bony plates in the skull are related to problems Murine neural crest cell formation and migration com- with neural crest cell differentiation [1–3]. Consequently it is mences at approximately the 4–5 stage in the region important to understand the distinct mechanisms, which reg- of the caudal midbrain and rostral [7] and proceeds ulate the formation, migration and differentiation of neural simultaneously as a wave rostrally towards the forebrain and crest cells. caudally towards the tail (Fig. 1). In avian embryos neu- ral crest cell migration commences after neural tube closure however this is not the case in mammalian embryos such as 2. Mouse neural crest cells: formation, migration, mice where neural crest cell formation and migration com- differentiation mences well before fusion of the bilateral halves of the neural plate. Typically there is a narrow temporal window during Murine neural crest cells are generated transiently along which neural crest cells are induced to delaminate and emi- almost the entire vertebrate axis at the interface between the grate from the dorsal neural tube and although this period surface and the neural plate of the , in a varies between species, in mice it typically lasts mice 7–9 h at region that is referred to as the neural plate border. During this each axial level [8]. The neural crest can be subdivided rostro- induction process, neuroepithelial cells undergo an epithelial caudally into at least four distinct major axial populations; to mesenchymal transformation at which point they delam- cranial, cardiac, vagal and trunk, each of which migrates inate and begin to emigrate from the neural tube, a process along unique pathways, contributing to specific cell and tis- that requires significant cytoarchitectural and cell adhesive sue types that are characteristic of their axial level of origin. changes. The induction of the neural crest cells is typically The cranial neural crest, which is the focus of this assayed by the expression of members of the Snail (Snail review can be divided into forebrain, midbrain and hindbrain and Slug) zinc-finger transcription factors gene family [4,5], domains of migrating neural crest cells. Cranial neural crest which play key roles in the epithelial to mesenchymal trans- cells do not appear to migrate randomly, rather they follow formation process by repressing the precise, species and region specific pathways moving subec- E- [6]. Snail/Slug transcription factors are among the todermally over the surface of the cranial [8–10]. P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 685

Cranial neural crest cells typically migrate in discrete seg- These results are eagerly anticipated as they will substantially regated streams, the pattern of which is highly conserved aid comparative analysis of neural crest cell patterning in in vertebrate species as disparate as amphibians, teleosts, diverse organisms and further our understanding of neural avians, marsupials and mammals (reviewed in [11]). Briefly, crest cell and craniofacial evolution. forebrain and rostral midbrain neural crest cells colonise the Collectively these analyses performed in mice together frontonasal and periocular regions. Caudal midbrain derived with those from avians, fish and frogs demonstrate that neural crest cells populate the maxillary component of the cardiac, vagal and cells produce neurons, first branchial arch [9,10]. The hindbrain is divided into glial cells, secretory cells and pigment cells—contributing seven distinct segments known as [12] and to the peripheral , , neural crest cells emigrate from each of the rhombomeres, and . Cranial neural crest cells however but predominantly from rhombomeres 2, 4 and 6 in discrete exhibit an even more surprising diversity of derivatives, segregated streams that populate the first, second and third giving rise to pigment cells, nerve ganglia, , branchial arches respectively [9,10]. and connective tissue, as well as most of the bone and Exquisite fate mapping analyses particularly in avians cartilage of the head (Fig. 1). using the quail-chick system, have revealed that neu- ral crest cells derived from each axial region of the cranial neural plate and in particular from each individual rhom- 3. Specification of mouse neural crest cell formation bomere generate specific and unique components of the cran- iofacial complex including the viscero- and neurocraniums Neural crest cell formation requires contact mediated sig- as well as the peripheral nervous system [13–16]. Despite the nals between the neural plate and paraxial tissues such as generation of lineage fate maps in mice [9,10], the limitations the and/or the mesoderm [21]. Three key of mouse embryo culture prevented a long-term differentia- signaling pathways (BMP, FGF, Wnt) intersect at the neu- tion analysis of the extensive fates of murine neural crest cells ral plate border, each of which is critical for neural crest cell although obvious parallels were drawn with avians. induction. Previously, BMP4/7 signaling in the ectoderm and Recently however, genetic recombination experiments neuroepithelium in the form of a gradient was finally permitted the long-term tracing of murine neural crest considered to be the key factor in inducing neural crest cell cells. Initially P0-cre mice were used in combination with formation [5,21–23]. However, more recently, the emphasis ␤- conditional reporters, which confirmed much has shifted in favour of Wnt signaling from the surface and of what was expected in terms of the fates of mammalian neu- in chick embryos Wnt6 is key, since it is widely expressed ral crest cells [17]. However, P0-cre mediated recombination in domains of epithelial to mesenchymal transitions during only occurred in subsets of neural crest cells and exhibited embryogenesis [24]. In contrast in zebrafish, Wnt8 signaling some ectopic expression leaving many questions in mam- from the surface ectoderm is the main protagonist of neural malian neural crest differentiation unanswered. crest cell induction [25].InXenopus, although Wnt signal- Shortly thereafter, a more specific neural crest cell pro- ing is also critical for neural crest cell induction, the situation moter from the Wnt1 gene was used in combination with is more complex since FGF signaling from the underlying ROSA26-lacZ conditional reporter mice to indelibly mark mesoderm also independently induces neural crest cell for- the progeny of the cranial and cells dur- mation [26–28]. The specifics of neural crest cell induction in ing embryonic, fetal and post-natal development [18–20]. these species are discussed in detail elsewhere in this volume, Murine cranial neural crest cells contribute to the forma- however, the contact mediated induction is considered to be a tion of the condensed dental , , conserved feature of vertebrate neural crest cell formation as , dentine matrix, pulp, cementum, periodontal it has been rigorously tested in chick and frog embryos due to ligaments, in Meckel’s cartilage, mandible, the their amenability to tissue manipulation, recombination and articulating disk of the temporomandibular joint and the explant culture. branchial arch nerve ganglia [20]. In the neurocranium, the In the case of mammalian embryos, there is currently a meninges and frontal are derived from neural crest paucity of data linking these different signaling pathways cells as is the suture mesenchyme, which separates the bony in neural crest cell induction mechanisms. Part of the prob- plates [18]. In the cardiac region, neural crest cells contribute lem is that many of the Wnt, Bmp and Fgf genes described as to the , conotruncal cushions, and playing roles in neural crest cell induction in non-mammalian adult derivatives of the third, fourth and sixth pharyngeal embryos also play key roles during and/or neural arch arteries [19]. Not surprisingly, the overall patterns of plate induction. Therefore, many knockout models of these migration and long-term differentiation are highly conserved genes exhibit embryonic lethality prior to the onset of neu- between mammalian and avian embryos. ral crest cell formation or no at this specific stage Further refinements to lineage and fate determination in due to functional redundancy. For example, genetic ablation mammalian embryos in terms of using axial level specific of Wnt1 or Wnt3a in mouse embryos failed to demonstrate promoters such as those that delineate a single a conserved role for Wnt signaling in murine neural crest subpopulation of neural crest cells are no doubt in progress. cell induction. Wnt1 was shown to be important for mid- 686 P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 brain patterning while Wnt3a was required for formation of A role for BMP signaling and BMP2 in particular in neural the [29,30]. However, in the absence of crest cell formation in mice is further supported by analyses of both Wnt1 and Wnt3a there is a marked deficiency in neu- Bmp2 null mutant mice which can survive to the 15–17 somite ral crest derivatives. Double mutant mice exhibit skeletal stage [45]. These embryos exhibit no evidence of migrating defects and a significant reduction in cranial and spinal sen- neural crest cells as assayed by the expression of the neural sory neurons as well as . [31]. Further evidence crest cell marker Crabp1 [44]. We cannot conclusively rule of a role for Wnt signaling in neural crest cell determination out that neural crest cells formed but failed to migrate because came from targeted inactivation of downstream components markers of neural crest cell formation such as Snail were of the . Inactivation of ␤-catenin, in the not used in this study. However, an absence of neural crest dorsal neural tube of mouse embryos [32,33] or null muta- cell formation seems to be the most likely scenario, which tions in APC [34] result in severe defects in cranial neural suggests that roles for the surface ectoderm and also BMP crest derivatives including the cranial and dorsal root gan- signaling in neural crest cell formation are conserved even in and the craniofacial skeletal elements. To date there is mouse embryos. no literature available describing a null in Wnt6. Interestingly, recent work in avians has reported an inte- This implies that Wnt signaling in mouse embryos appears gration between BMP and Wnt signaling with the cell cycle to be more important for the lineage specification of neural during avian trunk neural crest cell induction [46]. The over- crest cell differentiation rather than neural crest cell induc- expression of in the neural tube inhibits G1/S tran- tion. In support of this idea, Wnt/␤-catenin signal activation sition, which in turn abrogates BMP-induced neural crest in emigrating neural crest stem cells (NCSCs), regulates cell cell delamination. Similarly, interfering with ␤-catenin and fate decisions by promoting the formation of sensory neural LEF/TCF also inhibits G1/S transition and neural crest cell cells in vivo at the expense of other neural crest derivatives formation. Exogenous BMP can stimulate Wnt1 transcrip- [35]. tion, which implies that BMP signaling regulates G1/S tran- Bmp4 and Bmp7 have been implicated in neural crest cell sition and consequently, neural crest cell formation through induction in avians [22,36], however in Bmp4 or the canonical Wnt signaling pathway [46]. It remains to be Bmp7 do not produce obvious deficiencies in murine cranial seen however if this mechanism also holds true in the mouse neural crest development [37–39]. Functional redundancy during neural crest cell induction. may explain the absence of an effect, however even mouse Apart from the Bmp2 mutants, currently there is at least embryos mutant for both Bmp7 and Bmp5 also exhibit nor- one other clear example of murine neural crest cell induction mal patterns of neural crest cell formation and migration [40]. failure and that occurs in a restricted domain in the Hoxa1/b1 Furthermore, Bmpr1a (Alk3) and Bmpr1b (Alk2), which are double null mutant mice [47]. Hoxa1/b1 double mutants lack expressed in the neural tube of mouse embryos at the time of all second branchial arch skeletal elements and exhibit strik- neural crest cell induction, do not affect the formation of neu- ing external and middle malformations. Lineage tracing ral crest cells when conditionally inactivated [41,42]. They do in cultured double mutant embryos together with analyses of however affect the later differentiation of neural crest cells. neural crest cell markers such as Sox10 revealed a complete In contrast to chick embryos, Bmp4 expression is not absence of second branchial arch neural crest cells. What detected in the dorsal neural tube or surface ectoderm of is significant about this particular study was the demonstra- mouse embryos during the formation of neural crest cells tion that wild type cells transplanted into the double mutant [43]. Rather Bmp2 is expressed in a small region of the surface embryos could migrate into the second branchial arch. This ectoderm that abuts the neural plate suggesting that it may suggested that there was no inherent defect in the endogenous play a role in murine neural crest cell patterning [43,44].In migration pathway but rather that the neural crest cells at this order to disrupt BMP signaling in premigratory and migra- axial level failed to form [47]. It is unlikely that Hox genes tory neural crest cells, without the problems of functional play a universal role in neural crest cell induction because redundancy and early embryonic lethality potentially, Xeno- they are not expressed throughout the anterior regions of the pus noggin was expressed under the control of the Hoxa2 head where neural crest cells clearly form. Hox genes are in the hindbrain of transgenic mouse embryos [44]. clearly well recognized for their roles in anterior–posterior Hoxa2 is expressed in the neural tube and in the second patterning, however new roles are slowly being uncovered and third branchial arch mesenchyme. Consequently, noggin for these genes in dorso-ventral patterning as well [48].Inthe overexpression in the Hoxa2 domain led to a spatially spe- Hoxa1/b1 double mutants it therefore seems likely that the cific depletion of second and third branchial arch neural crest absence of neural crest cell formation was the secondary con- derivatives which included components of the cranial gan- sequence of dorso-ventral patterning transformation defects glia, the stapes and styloid process, greater horn and body of which respecifies the dorsal neural plate border territory to a the hyoid bone, and cricoid [44]. Underly- non-neural crest identity. ing the malformation or absence of these skeletal elements Overall, there are surprisingly and disappointingly few was the observation that the migration of cranial neural crest mouse mutants that exhibit a clear absence of migrating neu- cells was nearly completely abolished in the caudal branchial ral crest cells. Snail/Slug are widely used as indicators of arch region. neural crest cell formation since they mark the epithelial to P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 687 mesenchymal transition that neural crest cells undergo dur- the up-regulation of cadherin-7 and cadherin-11 [6,54–57]. ing their formation and emigration from the neural tube. By A breakdown of the cytoskeletal components in the basement directly repressing E-cadherin, Snail/Slug promotes the for- membrane surrounding the neural tube where the crest cells mation of neural crest cells and the onset of their migration emigrate is also required [58]. from the neural tube [6]. Unfortunately, null mutations in Slug Hindbrain derived cranial neural crest cells migrate in demonstrate that Slug is not required for neural crest cell gen- distinct segregated streams adjacent to the even numbered eration [49]. This may possibly be explained by functional rhombomeres and into the branchial arches in mouse embryos redundancy with Snail, however to date there is no litera- [10]. Consequently, neural crest free zones exist adjacent to ture describing a null allele of Snail or a phenotype with a rhombomeres 3 and 5 (Fig. 2). In contrast to the even num- defect in neural crest cell formation or migration. Although bered rhombomeres, we know from lineage tracing in mouse it is widely believed that the same contact mediated induc- embryos that generally fewer neural crest cells emigrate from tive mechanisms underlie murine neural crest cell formation, the odd numbered rhombomeres and also that rather than interactions between the neural ectoderm and surface ecto- migrating laterally these neural crest cells migrate anteriorly derm or the neural ectoderm and paraxial mesoderm have yet or posteriorly to join the adjacent even rhombomere neu- to be definitively tested because of the technical difficulty ral crest streams [43]. The segregation of these neural crest of manipulating mouse embryos due to their small size and cell populations is critical to prevent fusions of the cranial complex in vitro culture requirements. ganglia and skeletal elements and also to prevent mixing of neural crest cells with different genetic constitutions, which as described below are critical for axial neural crest and cran- 4. Specification of mouse neural crest cell migration iofacial patterning. The presence of neural crest free zones and the segre- One of the key steps in neural crest cell development is gation of neural crest cells into discrete streams is not an their migration from the neural tube through the extracel- intrinsic property of the vertebrate hindbrain or the neural lular matrix to their final destinations throughout the body crest cells. In ErbB4 null mutant mice neural crest cells [50]. The formation of neural crest cells is concomitant with from rhombomere 4 acquire the ability to migrate through the epithelial to mesenchymal transition of neuroepithelial the dorsal mesenchyme adjacent to rhombomere 3, which is cells and their emigration from the neural tube. Therefore, normally free of neural crest cells to join the first branchial it is extremely difficult to distinguish between a failure in arch stream of neural crest cells (Fig. 2) [59]. Consequently, neural crest cell induction versus a complete failure of neu- this leads to fusions of the trigeminal and facio-acoustic ral crest cell migration largely because the latter depends on ganglia. This aberrant rostral migration is not autonomous the former. Currently, there does not appear to be a situa- to the neural crest cells, since wild type neural crest cells tion where neural crest cells are induced to form but fail to transplanted into Erbb4 mutants also exhibit this aberrant migrate and consequently most of our analyses of the specifi- migration. Conversely, when ErbB4 mutant neural crest cells cation of neural crest cell migration in mice relate to aberrant are homotopically transplanted into wild type embryos they streams of neural crest cells in vivo and in vitro emulate the normal endogenous neural crest migratory pat- assays. tern from rhombomere 4 into the second branchial arch [60]. As described above Snail/Slug are key regulators of both The aberrant migration of neural crest cells in ErbB4 mutants neural crest cell induction and migration. The inhibition is due to as yet unidentified changes in the paraxial mes- of Slug blocks neural crest migration in [51,52]. enchyme environment. Since ErbB4 is normally expressed Similarly, antisense oligonucleotides directed against Slug in rhombomeres 3 and 5 this phenotype reflects defects in in avian embryos also blocks the ability of neuroepithe- signalling between the hindbrain and the adjacent environ- lial cells to emigrate from the neural tube [4]. Conversely, ment branchial arches in mutant embryos. over-expression of Slug in the chick neural tube leads to an Fusions of the cranial ganglia have also been described increased production of migrating neural crest cells [53]. The in the neuregulin and Krox20 null mutant mice [61–63], caveat of these analyses is that Slug/Snail are expressed in however, currently no neural crest cell lineage tracing or neural crest cells during their formation and also during their gene expression data is available to confirm the suspicion migration. What is still critically missing from analyses of that neural crest cells also migrate aberrantly in these mice. neural crest cell determination in mammalian embryos is an Interestingly, neuregulin is a for ErbB4,soitisnot assessment of the function of Slug/Snail during the migration surprising that it too may also play a role in regulating the of neural crest cells completely independently of their for- segregated pathways of cranial neural crest cell migration. mation. To achieve this will require conditional or inducible Coincidentally, Krox20 is expressed in rhombomeres 3 and inactivation of Slug/Snail in migrating neural crest cells in a 5, which overlaps with ErbB4. To date no direct link between strict spatiotemporally controlled manner. Krox20 and ErbB4 has been established, but given that Neural crest cell delamination and migration from the neu- Krox20 is one of the earliest determinants of rhombomere ral tube is aided by the down-regulation of cell adhesion 3 identity, it seems plausible (although not yet proven) that molecules such as NCAM, N-cadherin, and cadherin-6b and ErbB4 expression is disrupted in the Krox20 mutants which 688 P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693

Fig. 2. Normal and abnormal patterns of neural crest cell migration. Lineage tracing of neural crest cells in mouse embryos with DiI (fluorescent red dots) demonstrates that rhombomere 2 derived neural crest cells contribute to the first branchial arch (A and B) whereas rhombomere four derived neural crest cells contribute to the second branchial arch (C and D). Normally cranial neural crest cells migrate as distinct segregated streams into the adjacent branchial arches (E). Mixing between first and second branchial arch streams is observed dorsally adjacent to the neural tube in ErbB4 mutants (F). Extensive mixing of neural crest streams is observed in Twist mutants (G). Mixing is also observed in Tbx1 mutants but only ventrally in the vicinity of the first pharyngeal pouch (H). would account for the same ganglia fusions as observed in are disrupted resulting in poor colonisation by neural crest ErbB4 null mutant mice. cells. Consequently, a stream of Hoxa2 expressing neural Twist mutant embryos also exhibit aberrant neural crest crest cells originating from rhombomere four migrate aber- cell migration such that second branchial arch neural crest rantly into the first branchial arch, which results in inner cells stray from their subectodermal route towards the sec- ear abnormalities [67]. Tbx1 is not expressed in neural crest ond branchial arch and invade the normally neural crest free cells but rather in the paraxial mesoderm and core regions of mesenchyme adjacent to rhombomere 3 (Fig. 2) [64,65]. the branchial arches. This suggests that Tbx1 is acting non- Consequently, Twist mutant mice also display fusions of the cell autonomously in the branchial region during craniofacial trigeminal and facio-acoustic ganglia, similar to the ErbB4 . This data also provocatively suggests that mutants [65]. Interestingly, the expression pattern of ErbB4 craniofacial malformations such as those seen in DiGeorge is normal in the Twist mutant hindbrain, which implies that syndrome in which Tbx1 mutations are implicated may occur Twist could act downstream of ErbB4 or be part of an addi- as a secondary consequence of defects in neural crest cell tional important pathway that regulates the segregated migra- migration. The primary defect possibly lies in the mesoderm tion pathways of cranial neural crest cells. It remains to be or , which are the main regions of Tbx1 expression. determined whether the mesodermal or neural expression These results demonstrate that ErbB4, Twist and Tbx1 may domain of Twist regulates the migration patterns of cranial all be acting non-cell autonomously in directing the migration neural crest cells, but conditional Twist mutants will eventu- pathways of cranial neural crest cells and to date surprisingly ally answer this question. few molecules that intrinsically influence the path finding Previously it has been proposed that the cranial mesoderm of murine cranial neural crest cells have been identified. influences the migration pathways of cranial neural crest cells Evidence obtained primarily from analyses in frog embryos in mouse embryos [43] and consequently this implies that suggests that bi-directional Eph/ephrin cell signalling plays not all craniofacial malformations arise due to defects intrin- an important intrinsic role in keeping the neural crest cell sic to the neural crest. Rather, primary defects in the cranial streams segregated [68]. This is also supported by a recent mesoderm or other tissues that the neural crest cells contact study in mouse embryos, where it was recently shown that could lead to secondary abnormalities in both the migration ephrinB1 acts cell autonomously in neural crest cells to regu- and patterning of cranial neural crest cells [43,66]. This idea late craniofacial development [69]. EphrinB1 deficient mice appears to be supported by neural crest cell migration anoma- exhibit high degree of cleft palate and tympanic ring defects lies in Tbx1 mutant mice (Fig. 2). In Tbx1 mutant embryos, and the neural crest cells from the post-otic region of the patterning of the second and more caudal branchial arches hindbrain display aberrant migration by invading territories P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 689 normally devoid of neural crest cells which ultimately led to the second arch [75]. The hindbrain derived neural crest nerve fasciculation and branching defects. Each of the studies cells migrate toward the second branchial arch, but failing detailed above describe subpopulations of neural crest cells to enter, they instead accumulate in a region proximal to the migrating aberrantly but they do not reveal the underlying second branchial arch. While this defect in neural crest migra- mechanism that controls neural crest cell migration along tion does not reflect a neural crest cell specific function of there normal pathways. This instead appears to be a mor- Fgfr1, the second branchial arch exhibited by these mutants phogen regulated event. appear to result from abnormal patterning during early devel- BMPs are not only critical for neural crest induction, but opment of the branchial arch region. The initial generation they are also involved in neural crest migration. In mouse of neural crest and the segmentation and patterning of the embryos in which Alk2 (Bmpr1b) is conditionally deleted hindbrain appear normal in these mutants. Taken together, from neural crest cells, the initial formation and migration this suggests Fgfr1 patterns the branchial arch region cre- of neural crest cells remains unaffected. However, the later ating permissive environments that enable neural crest cell migration of mutant neural crest cells to the outflow tract of migration. the is dramatically impaired [41]. The effect of BMP Cranial neural crest cells are also susceptible to retinoic signaling on changes in cell adhesion and the cytoskeletal acid-mediated teratogenesis [76] and in vitro and in vivo components within the neural tube remains to be clearly studies have demonstrated that interferes with established, however one target of BMP signaling is the GTP- neural crest migration [77–81]. Targeted inactivation of the binding protein, rhoB [70]. Ectopic expression of noggin in mouse retinaldehyde dehydrogenase 2 (Raldh2), the the avian neural tube, leads to a reduction in the expression of responsible for early embryonic retinoic acid synthesis, leads rhoB as well as cadherin-6B which prevents neural crest emi- to prenatal death from a lack of aorticopulmonary septation gration from the dorsal neuroepithelium [71]. Collectively, [82]. Raldh2 null embryos exhibit impaired development of these studies support a role for BMP in neural crest migra- their posterior (third to sixth) branchial arches. Post-otic neu- tion from the neural tube through the regulation of rhoB and ral crest cells fail to establish segmental migratory pathways . and are misrouted caudally. The disruption of cardiac neural Although a role for FGF signaling in murine neural crest crest cell migration leads to the absence of outflow tract septa- cell formation has not been determined, in vitro migra- tion. Similar defects in vagal neural crest leads to agenesis of tion assays have revealed that exogenous FGF2 (basic FGF) the enteric ganglia, a condition reminiscent of Hirschprung’s and FGF8 exhibit a chemo-attractive activity that influ- disease in . Raldh2 expression is restricted to the pos- ences the migration of mesencephalic mouse neural crest terior most pharyngeal mesoderm, highlighting the potential cells [72]. Anti-FGF2 and anti-FGF8 neutralising antibod- importance of the mesoderm during the migration phase of ies inhibit the chemotaxic response. In vivo, FGF2 activity is neural crest cell development. In support of these genetic detected predominantly in target regions such as the mandibu- studies, treating headfold-stage mouse embryos with a pan- lar mesenchyme, which is colonised by mesencephalic neu- RAR antagonist in vitro and in vivo results in a complete ral crest cells. This characteristic distribution supports the absence of the third and fourth branchial arches [83]. Conse- notion that FGF2 acts as a chemo-attractant in the mouse quently, neural crest cells normally destined for the third and embryo that directs mesencephalic neural crest cell migra- fourth arches migrate ectopically. tion. The domains of FGF8 activity are largely restricted to the Further roles for retinoic acid in the migration of neural branchial arch suggesting it may not be involved crest cells have been observed in a stage-dependent manner in in guiding neural crest cell migration, however FGF2 activ- rat embryos [80]. Early-stage retinoic acid treatment induced ity can be promoted by FGF8. This has led to the hypothesis an ectopic caudal migration of the anterior hindbrain (rhom- that FGF8 activity in the mandibular arch epithelium is a pre- bomeres (r) 1 and 2) crest cells into the second branchial arch requisite for the differential localisation of FGF2 and that in and acousticofacial . In contrast, late-stage treatment turn, the distribution of FGF2 is essential for chemotaxis of did not disturb the segmental migration pattern of hindbrain- mesencephalic neural crest cell migration [72]. In vivo this derived crest cells, although it did induce branchial arch is not the case since the conditional inactivation of Fgf8 in fusions. Anterior hindbrain-derived neural crest cells pop- the branchial arch epithelium does not disrupt the migration ulated the anterior half of the fused arch while neural crest of cranial neural crest cells into the branchial arches. Instead cells derived from the pre-otic hindbrain (r3 and r4) occupied FGF8 activity in the branchial arch surface ectoderm appears its posterior half [80]. Similar treatments of mouse embryos to play an important functional role in neural crest cell sur- with retinoic acid in culture result in the ectopic migration of vival [73]. second arch neural crest cells into the first arch [84], which Further evidence for a role of FGF signaling in neural interestingly mimics the Tbx1 mutant neural crest cell migra- crest cell migration has come from analyses of the Fgfr1 null tion phenotype. mutant mice. Fgfr1 is expressed in multiple cellular com- One of the potential targets of retinoic acid is the c-Jun ponents of the branchial arches [74] and recent studies have N-terminal kinase (JNK) pathway [85]. Cardiac neural crest shown that perturbation of Fgfr1 function during branchial migration is inhibited when retinoic acid blocks JNK phos- arch development results in failure of the neural crest to enter phorylation, which is critical for neural crest outgrowth. Fur- 690 P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 thermore, the use of dominant-negative constructs to perturb This issue of multipotency versus lineage restriction has upstream and downstream components of the JNK pathway also been addressed with respect to melanocytes. A sub- also reduce cardiac neural crest outgrowth, implying that JNK population of cells in the dorsomedial domain of the neu- is not only a target for the action of retinoic acid, but is also ral tube, which expresses the tyrosine kinase, Kit, critical for the migration of cardiac neural crest cells in vitro. migrates exclusively into the developing dermis [96]. Con- Neural crest cells express retinoic acid receptors (RARs) and sequently, these cells activate definitive lineage cellular retinoic acid binding proteins (CRABPs) however markers indicating they are melanocyte progenitor cells. This the primary source of retinoic acid comes from the paraxial particular sub-population is generated predominantly at the mesoderm. midbrain–hindbrain junction and at the cervical and lower Collectively, these studies demonstrate that appropriate trunk levels. Other cells within the dorsal neural tube that are migration may not be an intrinsic property of neural crest cells Kit negative, express p75 and migrate ventrally giving rise but rather is directed non-cell autonomously in an extremely to neurons and glia. Interestingly, the p75 positive cells are intricate fashion that requires complex interactions between located ventrolateral to the Kit positive cells, suggesting that signals from the neuroepithelium, together with the paraxial there is in vivo neural crest cell lineage segregation within mesoderm, ectoderm and endoderm tissues [43,59]. the mouse neural tube [96]. This poses an important ques- tion that currently remains unresolved and that is what are the signaling pathways that establish this lineage segrega- 5. Specification of multipotency versus restricted tion during the period of neural crest cell formation? Is the potency of mouse neural crest cells interface between FGF, WNT and BMP signaling at the neu- ral plate border only stimulating the induction of neural crest One of the many intriguing features of neural crest cell cells or do each of these signaling pathways have the capacity is their pluripotency. A single neural crest cell will differ- to specify sublineages within the neural crest. entiate into any of several distinct cell types depending on its location within the embryo. However, it is still not certain whether most individual neural crest cells that leave the neural 6. Conclusions tube are pluripotent or whether the majority of the popula- Mammalian neural crest cells are a mulitpotent migratory tion is already restricted to certain fates. Some of the strongest stem cell population that generates an impressively broad evidence for pluripotency comes from single neuroepithelial array of cell and tissue types during craniofacial development. cell labeling in chick embryos, which revealed that progeny Currently, BMP2 signaling specifically from the surface ecto- of a single neural crest cell could become sensory neurons, derm immediately adjacent to the neural plate appears to be melanocytes, adrenomedullary cells and also glia [86,87]. the key player in murine neural crest cell induction. This Similar results were obtained from in vitro clonal assays in is somewhat different to chick, frog and zebrafish embryos mice [88], where it was demonstrated that these cells have where the emphasis lies with Wnt signaling. Collectively, the capacity to self renew and form neurons, glia and smooth BMP, Wnt and FGF signaling all play key roles in regulating muscle [89]. Consequently they have been termed neural either early or late steps in neural crest cell migration. Surpris- crest stem cells [90]. In contrast analyses in zebrafish provide ingly, these signaling pathways also mediate major aspects equally strong evidence for restricted lineage determination of neural crest cell differentiation. Hence, repeated use of the since similar labeling experiments revealed that most neural same signaling pathways elicits completely different effects crest cells differentiate into only one cell type [91,92]. on neural crest cell formation, migration and differentiation. Evidence in mouse embryos suggests that some popu- The studies described in this review highlight the complex- lations of neural crest cells are committed at the time of ity in regulating neural crest cell patterning and the balance emigration from the neural tube or very soon after since they between signals acquired in the neuroepithelium during for- express transcription factors that constrain the cell types they mation versus signals received from the extrinsic tissue envi- can produce [93,94]. Using the Cre recombinase system to ronments during migration. This intricate regulation affects permanently mark subsets of neural crest cells that transiently not only the migratory properties of neural crest cells but also expressed the Ngn2 or Wnt1 revealed that dramatically influences their differentiation and has been cru- Ngn2 positive progenitors were four-fold more likely than cial to vertebrate neural crest and craniofacial evolution. Wnt1 positive cells to contribute to sensory and sympathetic ganglia [95]. Within the dorsal route ganglia however, both Ngn2 and Wnt1 positive populations were equally likely to Acknowledgements generate neurons or glia. Hence this suggests that very early in neural crest migration, Ngn2 may mark a sub-population P.A.T. is supported by research funds from the Stowers that is already fate biased. Furthermore, this data also implies Institute, a Basil O’Connor Research Scholar Award (#5- that in the trunk neural crest cells become restricted to sensory FY03-16) from the March of Dimes and grant RO1 DE or autonomic lineages before committing to either neuronal 016082-01 from the National Institute of Dental and Cranio- or glial fates. facial Research. P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 691

References [23] Mayor R, Guerrero N, Martinez C. Role of FGF and noggin in neural crest induction. Dev Biol 1997;189(1):1–12. [1] Poswillo D. The pathogenesis of the Treacher Collins syndrome [24] Garcia-Castro MI, Marcelle C, Bronner-Fraser M. Ectodermal Wnt (mandibulofacial dysostosis). Br J Oral Surg 1975;13(1):1–26. function as a neural crest inducer. Science 2002;297(5582):848–51. [2] Poswillo D. The aetiology and pathogenesis of craniofacial defor- [25] Lewis JL, Bonner J, Modrell M, Ragland JW, Moon RT, Dorsky RI, mity. Development 1988;103(Suppl):207–12. et al. Reiterated Wnt signaling during zebrafish neural crest devel- [3] Moore KL. The developing human. Clin Orient Embryol 1988. opment. Development 2004;131(6):1299–308. [4] Nieto MA, Sargent MG, Wilkinson DG, Cooke J. Control of cell [26] Monsoro-Burq AH, Wang E, Harland R. Msx1 and Pax3 cooperate behavior during vertebrate development by slug, a zinc finger gene. to mediate FGF8 and WNT signals during Xenopus neural crest Science 1994;264:835–9. induction. Dev Cell 2005;8:167–78. [5] Mayor R, Morgan R, Sargent MG. Induction of the prospective neu- [27] Monsoro-Burq AH, Fletcher RB, Harland RM. Neural crest induction ral crest of Xenopus. Development 1995;121(3):767–77. by paraxial mesoderm in Xenopus embryos requires FGF signals. [6] Cano A, Perez-Moreno MA, Rodrigo I, Locascio A, Blanco MJ, del Development 2003;130(14):3111–24. Barrio MG, et al. The transcription factor snail controls epithelial- [28] LaBonne C, Bronner-Fraser M. Neural crest induction in Xenopus: mesenchymal transitions by repressing E-cadherin expression. Nat evidence for a two-signal model. Development 1998;125:2403–14. Cell Biol 2000;2(2):76–83. [29] Thomas K, Capecchi M. Targeted disruption of the murine int-1 [7] Nichols DH. Formation and distribution of neural crest mesenchyme proto-oncogene resulting in severe abnormalities in midbrain and to the first region of the mouse embryo. Am J Anat cerebellar development. Nature 1990;346:847–50. 1986;176(2):221–31. [30] Liu P, Wakamiya M, Shea MJ, Albrecht U, Behringer RR, Bradley [8] Serbedzija G, Fraser S, Bronner-Fraser M. Vital dye analysis of cra- A. Requirement for Wnt3 in vertebrate axis formation. Nat Genet nial neural crest cell migration in the mouse embryo. Development 1999;22:361–5. 1992;116:297–307. [31] Ikeya M, Lee SM, Johnson JE, McMahon AP, Takada S. Wnt sig- [9] Osumi-Yamashita N, Ninomiya Y, Doi H, Eto K. The contribution nalling required for expansion of neural crest and CNS progenitors. of both forebrain and midbrain crest cells to the mesenchyme in Nature 1997;389(6654):966–70. the frontonasal mass of mouse embryos. Dev Biol 1994;164(2): [32] Brault V, Moore R, Kutsch S, Ishibashi M, Rowitch DH, McMahon 409–19. AP, et al. Inactivation of the beta-catenin gene by Wnt1-Cre-mediated [10] Trainor PA, Tam PPL. Cranial paraxial mesoderm and neural results in dramatic brain malformation and failure of cran- crest of the mouse embryo-codistribution in the craniofacial mes- iofacial development. Development 2001;128(8):1253–64. enchyme but distinct segregation in the branchial arches. Develop- [33] Hasegawa S, Sato T, Akazawa H, Okada H, Maeno A, Ito M, ment 1995;121(8):2569–82. et al. in neural crest cells by functional loss of APC [11] Kulesa P, Ellies DL, Trainor PA. Comparative analysis of neural crest tumor suppressor gene. Proc Natl Acad Sci USA 2002;99:297– cell death, migration, and function during vertebrate embryogenesis. 302. Dev Dyn 2004;229:14–29. [34] Hari L, Brault V, Kleber M, Lee HY, Ille F, Leimeroth R, et al. [12] Vaage S. The segmentation of the primitive neural tube in chick Lineage-specific requirements of beta-catenin in neural crest devel- embryos (Gallus domesticus). Advances in , embryology and opment. J Cell Biol 2002;159:867–80. cell biology 1969;41:1–88. [35] Lee HY, Kleber M, Hari L, Brault V, Suter U, Taketo MM, et al. [13] Noden D. The role of the neural crest in patterning of avian cranial Instructive role of Wnt/beta-catenin in sensory fate specification in skeletal, connective, and muscle tissues. Dev Biol 1983;96:144–65. neural crest stem cells. Science 2004;303(5660):1020–3. [14] Couly GF, Coltey PM, Le Douarin NM. The triple origin of skull [36] Liem Jr KF, Tremml G, Roelink H, Jessell TM. Dorsal differenti- in higher —a study in quail-chick chimeras. Development ation of neural plate cells induced by BMP-mediated signals from 1993;117:409–29. epidermal ectoderm. Cell 1995;82(6):969–79. [15] Couly G, LeDouarin N. The fate map of cephalic neural primordium [37] Winnier G, Blessing M, Labosky PA, Hogan BL. Bone morpho- at the presomitic to the 3-somtie stage in the avian embryo. Devel- genetic protein-4 is required for mesoderm formation and patterning opment 1988;103(Suppl.):101–13. in the mouse. Genes Dev 1995;9:2105–16. [16] Kontges G, Lumsden A. Rhombencephalic neural crest segmen- [38] Dudley AT, Lyons KM, Robertson EJ. A requirement for bone mor- tation is preserved throughout craniofacial . Development phogenetic protein-7 during development of the mammalian kidney 1996;122(10):3229–42. and . Genes Dev 1995;9:2795–807. [17] Yamauchi Y, Abe K, Mantani A, Hitoshi Y, Suzuki M, Osuzu F, [39] Luo G, Hofmann C, Bronckers AL, Sohocki M, Bradley A, et al. A novel transgenic technique that allows specific marking Karsenty G. BMP-7 is an inducer of nephrogenesis, and is also of the neural crest cell lineage in mice. Dev Biol 1999;212:191– required for and skeletal patterning. Genes Dev 203. 1995;9:2808–20. [18] Jiang X, Iseki S, Maxson RE, Sucov HM, Morriss-Kay GM. Tissue [40] Solloway MJ, Robertson EJ. Early embryonic lethality in Bmp5; origins and interactions in the mammalian skull vault. Dev Biol Bmp7 double mutant mice suggests functional redundancy within 2002;241:106–16. the 60A subgroup. Development 1999;126:1753–68. [19] Jiang X, Rowitch DH, Soriano P, McMahon AP, Sucov HM. [41] Dudas M, Sridurongrit S, Nagy A, Okazaki K, Kaartinen V. Cran- Fate of the mammalian cardiac neural crest. Development iofacial defects in mice lacking BMP type I receptor Alk2 in neural 2000;127(8):1607–16. crest cells. Mech Dev 2004;121:173–82. [20] Chai Y, Jiang X, Ito Y, Bringas Jr P, Han J, Rowitch DH, et al. Fate [42] Stottmann RW, Choi M, Mishina Y, Meyers EN, Klingensmith of the mammalian cranial neural crest during and mandibular J. BMP receptor IA is required in mammalian neural crest cells morphogenesis. Development 2000;127(8):1671–9. for development of the cardiac outflow tract and ventricular [21] Selleck MA, Bronner-Fraser M. Origins of the avian neural myocardium. Development 2004;131:2205–18. crest: the role of neural plate-epidermal interactions. Development [43] Trainor PA, Sobieszczuk D, Wilkinson D, Krumlauf R. Signalling 1995;121(2):525–38. between the hindbrain and paraxial tissues dictates neural crest [22] Selleck MA, Garcia-Castro MI, Artinger KB, Bronner-Fraser M. migration pathways. Development 2002;129(2):433–42. Effects of Shh and Noggin on neural crest formation demonstrate [44] Kanzler B, Foreman RK, Labosky PA, Mallo M. BMP signaling that BMP is required in the neural tube but not ectoderm. Develop- is essential for development of skeletogenic and neurogenic cranial ment 1998;125(24):4919–30. neural crest. Development 2000;127(5):1095–104. 692 P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693

[45] Zhang H, Bradley A. Mice deficient for BMP2 are nonviable and [66] Trainor P, Krumlauf R. Patterning the cranial neural crest: Hind- have defects in /chorion and cardiac development. Develop- brain segmentation and plasticity. Nat Rev Neurosci ment 1996;122:2977–86. 2000;1:116–24. [46] Burstyn-Cohen T, Kalcheim C. Association between the cell cycle [67] Moraes F, Novoa A, Jerome-Majewska LA, Papaioannou VE, Mallo and neural crest delamination through specific regulation of G1/S M. Tbx1 is required for proper neural crest migration and to stabilize transition. Dev Cell 2002;3:383–95. spatial patterns during middle and development. Mech Dev [47] Gavalas A, Trainor P, Ariza-McNaughton L, Krumlauf R. Syn- 2005;122:199–212. ergy between Hoxa1 and Hoxb1: the relationship between arch [68] Smith A, Robinson V, Patel K, Wilkinson DG, The EphA4. EphB1 patterning and the generation of cranial neural crest. Development receptor tyrosine kinases and ephrin-B2 ligand regulate targeted 2001;128(15):3017–27. migration of branchial neural crest cells. Curr Biol 1997;7:561– [48] Davenne M, Maconochie M, Neun R, Brunet J-F, Chambon P, Krum- 70. lauf R, et al. Hoxa2 and Hoxb2 control dorsoventral patterns of neu- [69] Davy A, Aubin J, Soriano P. Ephrin-B1 forward and reverse signaling ronal development in the rostral hindbrain. 1999;22:677–91. are required during mouse development. Genes Dev 2004;18:572–83. [49] Jiang R, Lan Y, Norton CR, Sundberg JP, Gridley T. The Slug gene [70] Liu JP, Jessell TM. A role for rhoB in the delamination of is not essential for mesoderm or neural crest development in mice. neural crest cells from the dorsal neural tube. Development Dev Biol 1998;198(2):277–85. 1998;125(24):5055–67. [50] Hay ED. An overview of epithelio-mesenchymal transformation. [71] Sela-Donenfeld D, Kalcheim C. Regulation of the onset of neural Acta Anat (Basel) 1995;154(1):8–20. crest migration by co-ordinated activity of BMP4 and noggin inthe [51] LaBonne C, Bronner-Fraser M. Snail-related transcriptional repres- dorsal neural tube. Development 1999;126:4749–62. sors are required in Xenopus for both the induction of the neu- [72] Kubota Y, Ito K. Chemotactic migration of mesencephalic neural ral crest and its subsequent migration. Dev Biol 2000;221(1):195– crest cells in the mouse. Dev Dyn 2000;217:170–9. 205. [73] Trumpp A, Depew MJ, Rubenstein JL, Bishop JM, Martin GR. Cre- [52] Carl TF, Dufton C, Hanken J, Klymkowsky MW. Inhibition of neural mediated gene inactivation demonstrates that FGF8 is required for crest migration in Xenopus using antisense slug RNA. Dev Biol cell survival and patterning of the first branchial arch. Genes Dev 1999;213(1):101–15. 1999;13(23):3136–48. [53] del Barrio MG, Nieto MA. Overexpression of Snail family mem- [74] Wilke TA, Gubbels S, Schwartz J, Richman JM. Expression of bers highlights their ability to promote chick neural crest formation. fibroblast receptors (FGFR1, FGFR2, FGFR3) in the Development 2002;129(7):1583–93. developing head and face. Dev Dyn 1997;210:41–52. [54] Akitaya T, Bronner-Fraser M. Expression of cell adhesion molecules [75] Trokovic N, Trokovic R, Mai P, Partanen J. Fgfr1 regulates patterning during initiation and cessation of neural crest cell migration. Dev of the pharyngeal region. Genes Dev 2003;17(1):141–53. Dyn 1992;194(1):12–20. [76] Lammer E, Chen D, Hoar R, Agnish A, Benke P, Braun J, et al. [55] Nakagawa S, Takeichi M. Neural crest cell–cell adhesion controlled Retinoic acid embryopathy. New Engl J Med 1985;313:837–41. by sequential and subpopulation-specific expression of novel cad- [77] Thorogood P, Smith L, Nicol A, McGinty R, Garrod D. Effects of herins. Development 1995;121(5):1321–32. vitamin A on the behaviour of migratory neural crest cells in vitro. [56] Inoue T, Chisaka O, Matsunami H, Takeichi M. Cadherin-6 expres- J Cell Sci 1982;57:331–50. sion transiently delineates specific rhombomeres, other neural tube [78] Wiley MJ, Cauwenbergs P, Taylor IM. Effects of retinoic acid on the subdivisions, and neural crest subpopulations in mouse embryos. Dev development of the facial in hamsters: early changes involv- Biol 1997;183(2):183–94. ing cranial neural crest cells. Acta Anat (Basel) 1983;116(2):180– [57] Nakagawa S, Takeichi M. Neural crest emigration from the neu- 92. ral tube depends on regulated cadherin expression. Development [79] Pratt RM, Goulding EH, Abbott BD. Retinoic acid inhibits migra- 1998;125(15):2963–71. tion of cranial neural crest cells in the cultured mouse embryo. J [58] Erickson CA, Perris R. The role of cell–cell and cell–matrix Craniofac Genet Dev Biol 1987;7:205–17. interactions in the morphogenesis of the neural crest. Dev Biol [80] Lee YM, Osumi-Yamashita N, Ninomiya Y, Moon CK, Eriksson 1993;159(1):60–74. U, Eto K. Retinoic acid stage-dependently alters the migration [59] Golding J, Trainor P, Krumlauf R, Gassman M. Defects in pathfind- pattern and identity of hindbrain neural crest cells. Development ing by cranial neural crest cells in mice lacking the Neuregulin 1995;121:825–37. receptor ErbB4. Nat Cell Biol 2000;2:103–9. [81] Mulder GB, Manley N, Maggio-Price L. Retinoic acid-induced [60] Golding JP, Dixon M, Gassmann M. Cues from neuroepithe- thymic abnormalities in the mouse are associated with altered lium and surface ectoderm maintain neural crest-free regions pharyngeal morphology, thymocyte maturation defects, and within cranial mesenchyme of the developing chick. Development altered expression of Hoxa3 and Pax1. 1998;58: 2002;129(5):1095–105. 263–75. [61] Meyer D, Birchmeier C. Multiple essential functions of neuregulin [82] Niederreither K, Vermot J, Messaddeq N, Schuhbaur B, Chambon in development. Nature 1995;378:386–90. P, Dolle P. Embryonic retinoic acid synthesis is essential for heart [62] Schneider-Maunoury S, Topilko P, Seitanidou T, Levi G, Cohen- morphogenesis in the mouse. Development 2001;128:1019–31. Tannoudji M, Pournin S, et al. Disruption of Krox-20 results in [83] Wendling O, Ghyselinck NB, Chambon P, Mark M. Roles of retinoic alteration of rhombomeres 3 and 5 in the developing hindbrain. Cell acid receptors in early embryonic morphogenesis and hindbrain pat- 1993;75:1199–214. terning. Development 2001;128(11):2031–8. [63] Swiatek PJ, Gridley T. Perinatal lethality and defects in hindbrain [84] Trainor P, Krumlauf R. Plasticity in mouse neural crest cells development in mice homozygous for a targeted mutation of the zinc reveals a new patterning role for cranial mesoderm. Nat Cell Biol finger gene Krox-20. Genes Dev 1993;7:2071–84. 2000;2:96–102. [64] Soo K, O’Rourke MP, Khoo PL, Steiner KA, Wong N, Behringer [85] Li J, Molkentin JD, Colbert MC. Retinoic acid inhibits cardiac neural RR, et al. Twist function is required for the morphogenesis of the crest migration by blocking c-Jun N-terminal kinase activation. Dev cephalic neural tube and the differentiation of the cranial neural crest Biol 2001;232:351–61. cells in the mouse embryo. Dev Biol 2002;247:251–70. [86] Bronner-Fraser M, Fraser S. Cell lineage analysis reveals multipo- [65] Ota MS, Loebel DA, O’Rourke MP, Wong N, Tsoi B, Tam PP. tency of some avian neural crest cells. Nature 1988;335:161–4. Twist is required for patterning the cranial and maintaining [87] Bronner-Fraser M, Fraser S. Developmental potential of avian trunk the viability of mesodermal cells. Dev Dyn 2004;230:216–28. neural crest cells in situ. Neuron 1989;3(6):755–66. P.A. Trainor / Seminars in Cell & Developmental Biology 16 (2005) 683–693 693

[88] Stemple DL, Anderson DJ. Isolation of a stem cell for neurons and [93] Lo L, Sommer L, Anderson DJ. MASH1 maintains competence for glia from the mammalian neural crest. Cell 1992;71(6):973–85. BMP2-induced neuronal differentiation in post-migratory neural crest [89] Shah NM, Groves AK, Anderson DJ. Alternative neural crest cell cells. Curr Biol 1997;7:440–50. fates are instructively promoted by TGFbeta superfamily members. [94] Ma Q, Chen Z, del Barco Barrantes I, de la Pompa JL, Anderson DJ. Cell 1996;85(3):331–43. neurogenin1 is essential for the determination of neuronal precursors [90] Morrison SJ, White PM, Zock C, Anderson DJ. Prospective iden- for proximal cranial sensory ganglia. Neuron 1998;20:469–82. tification, isolation by flow cytometry, and in vivo self-renewal of [95] Zirlinger M, Lo L, McMahon J, McMahon AP, Anderson DJ. Tran- multipotent mammalian neural crest stem cells. Cell 1999;96:737–49. sient expression of the bHLH factor neurogenin-2 marks a subpop- [91] Raible DW, Eisen JS. Restriction of neural crest cell fate in the trunk ulation of neural crest cells biased for a sensory but not a neuronal of the embryonic zebrafish. Development 1994;120:495–503. fate. Proc Natl Acad Sci USA 2002;99:8084–9. [92] Schilling TF, Kimmel CB. Segment and cell type lineage restric- [96] Wilson YM, Richards KL, Ford-Perriss ML, Panthier JJ, Murphy tions during pharyngeal arch development in the zebrafish embryo. M. Neural crest cell lineage segregation in the mouse neural tube. Development 1994;120(3):483–94. Development 2004;131:6153–62.