How to Cite this Article: Cordero DR, Brugmann S, Chu Y, Bajpai R, Jame M, Helms JA. 2011. Cranial neural crest cells on the move: Their roles in craniofacial development. Am J Med Genet Part A 155:270–279.
Malformations involving the craniofacial regions are observed in three-fourths of human birth defects (see Centers for Disease Control, Birth Defects Research and Prevention at http://www.cdc.gov/ncbddd/bd/centers.htm). In part, this may be attributable to the intricate means by which the craniofacial region is assembled during embryonic development. Tissues of the craniofacial complex are primarily derived from neural crest cells (NCCs), a population of transiently migratory cells that originate from the dorsal aspect of the neural tube during embryogenesis (Fig. 1), then migrate to populate the frontonasal process (FNP) and the first, second, third and fourth pharyngeal arches [Lièvre and Douarin, 1975; Lumsden et al., 1991; Hunt et al., 1991a,b]. NCCs contribute to neural, skeletal, dermal, and mesenchymal structures. Because of this pluripotency they have sometimes been referred to as the fourth germ layer [Hall, 2000].
A tightly controlled spatial and temporal signaling network is required for the induction, migration, proliferation, and differentiation of NCCs that give rise to craniofacial structures. During migration, and even after arrival at their final destination, interactions between NCCs and the adjacent surface ectoderm, neuroectoderm and endoderm are necessary for normal development of the craniofacial region [Cordero et al., 2004; Creuzet et al., 2005, 2006; Sandell and Trainor, 2006]. Aberrations in any of the multi-step processes involved in regulating NCC behavior can result in craniofacial malformations. Clinicians who evaluate patients with such malformations will find that a working knowledge of NCC biology is beneficial for understanding the etiologies of these diseases.
DERIVATIVES OF THE NEURAL CREST
NCCs contribute to multiple cell types and tissues throughout the body (enteric nervous system, glia, neurons, melanocytes, as well as connective tissues, chondrocytes, and myofibroblasts lining the blood vessels). In the face, however, they are the prime contributor. Specifically, the facial skeleton and the vast majority of facial connective tissues are derived exclusively from cranial NCCs. As a consequence, genetic diseases that affect aspects of NCC generation, migration, proliferation, or differentiation are more likely to manifest themselves in the craniofacial region. Therefore, we turn our attention first to the initiation of NCC development.
THE EMERGING NEURAL CREST
There are currently two theories regarding the initiation of NCCs specification. One theory that has been put forth is that NCCs are specified during gastrulation, and is based upon detailed studies in the avian model system [Basch et al., 2006]. In this study it was shown that a restricted region of epiblast, expressing Pax7, contributes to neural folds and migrating NCCs [Basch et al., 2006]. Inhibition of Pax7 protein production prevented the expression of several NCC markers like Slug (Snail2), Sox9, Sox10, and HNK-1 [Basch and Bronner-Fraser, 2006] independent of mesoderm or neural plate. A second group also showed that specification of NCC occurs during gastrulation but is dependent upon signals from dorso-lateral mesoderm, requiring both Wnt activation and BMP inhibition and a second step at neurulation, which requires both Wnt and BMP activation in adjacent tissues [Steventon et al., 2009]. This is in contrast to the classic theory that suggests that NCCs are specified around or at the time of neural tube closure.
In the second or “classical” model of NCC induction, planar interactions between the neural ectoderm and non-neural ectoderm, as well as signaling from underlying paraxial mesoderm, are required for NCC formation [Selleck and Stern, 1991; Noden and Trainor, 2005]. A number of signaling pathways have been shown to be necessary for NCC generation and/or survival including two bone morphogenetic protein (BMP) antagonists, chordin and noggin [Anderson et al., 2006], fibroblast growth factors (FGFs) [Mayor et al., 1997], and Wnt signals [Garcia-Castro et al., 2002]. With the key molecular pathways identified, the present challenge is to understand how they are integrated in such a way as to induce the formation of this unique set of cells. Part of this integration relies on tightly regulated spatiotemporal expression of pathway activators. For example, the position of presumptive neural crest domain correlates broadly with the BMP4 expression domain [Ezin et al., 2009] and with the anterior-most extent of Wnt pathway activity [Li et al., 2009]. However, the integration of Wnt and BMP signaling pathways to produce NCCs is poorly understood. In some experimental models, BMP signals induce cells at the neural plate border to become NCCs, but only when Wnt signaling is inhibited [Patthey et al., 2009]. In other experimental models, Wnt signaling is sufficient for the generation of NCCs. For example, Wnt signaling induces the transcription factor Gbx2, which in turn is necessary for neural crest induction [Byrd and Meyers, 2005; Li et al., 2009].
Although both theories propose a different timing of specification, both agree that cranial NCCs are generated from the border between neural and non-neural ectoderm [Ruffins and Bronner-Fraser, 2000; Knecht and Bronner-Fraser, 2002; Crane and Trainor, 2006]. In addition the data from various model systems must be carefully compared [Steventon et al., 2005] and their extrapolation to human development must be critically assessed.
NEURAL CREST MIGRATION
Independent of the timing of NCC specification, contact-mediated signaling between tissues in the dorsal neural tube results in neural ectoderm cells at the neuroectoderm and non-neuroectoderm border undergoing epithelial-mesenchymal transition (EMT), which is required for NCC migration. Contact-mediated signaling between tissues in the dorsal neural tube stimulates the cells at the neural/non-neural border to undergo a transition from an epithelial to a mesenchymal phenotype [Noden and Trainor, 2005]. This transition into a highly invasive phenotype is a hallmark of NCCs and one they share with metastatic cells [Radisky and LaBarge, 2008].
Epithelial-to-mesenchymal transition (EMT) is a multi-step process: For NCC migration to occur, NCCs must lose their apico-basal polarity, and simultaneously disassemble intercellular adhesion complexes that are required for epithelial formation [Thiery and Sleeman, 2006; Acloque et al., 2009]. The disassembly of intercellular adhesion complexes involves members of the Cadherin family. The expression of Cadherins are associated with the formation of epithelial and/or stable cell phenotypes and must be down-regulated by transcriptional repressors to complete the transition from a sedentary epithelium to motile mesenchyme [Obrink, 1986]. Furthermore, tight junctions, structures involved in maintaining cell–cell contacts and paracellular permeability and cell polarity [Lal-Nag and Morin, 2009] must also be eliminated in order to allow for cell migration. In terms of neural crest development and migration, Cadherins play an essential role. The neural plate can be delineated from the non-neural ectoderm by the expression of N-cadherin expression and the loss of E-cadherin expression. Further specification of the neural plate is acquired by induction of cadherin-6B expression in the dorsal neural folds, defining the domain that gives rise to premigratory neural crest cells. As neural crest cells undergo EMT in preparation of migration, cadherin-6B and N-cadherin are down-regulated, and cadherin-7 expression is induced in migratory neural crest cells [Taneyhill, 2008]. This process is initiated by expression of genes such as zinc-finger-like transcription factors like Snail1 and Slug (Snail 2) [Nieto et al., 1994; Nieto, 2008]. For example, over-expression of Snail leads to repression of E-cadherin gene transcription [Cano et al., 2000; Peinado et al., 2004a]. When this dynamic concert of cadherin gene expression fails to occur in the proper manner, NCC migration is disrupted and the result is a perturbation in craniofacial development.
SIP1 and SNAIL [SNAIL1 and SNAIL2 (SLUG)] are known transcriptional repressors that control Cadherin activation [Comijn et al., 2001]. Micro-deletions or mutations in SIP1 transcriptional repressor have been identified in ∼200 patients with Mowat–Wilson syndrome (OMIM 235730) [Dastot-Le Moal et al., 2007; Saunders et al., 2009] an autosomal dominant disorder characterized by distinct facial [Saunders et al., 2009] dysmorphisms, microcephaly, mental retardation, and epilepsy [Mowat et al., 1998, 2003; Horn et al., 2004; Zweier et al., 2005; Garavelli et al., 2009]. SIP1 normally represses E-cadherin expression [Comijn et al., 2001] and in mice carrying null mutations in Sip1, E-Cadherin levels stay abnormally high [Van de Putte et al., 2007]. Consequently, cells at the neural plate boundary cannot undergo an EMT and instead remain tethered to the epithelium. The end result is that NCC migration is delayed and affected mice exhibit abnormal morphology of the neural plate [Van de Putte et al., 2007].
The transcription factor Snail also represses E-cadherin gene transcription, through histone deacetylase-mediated chromatin remodeling [Peinado et al., 2004a,b]. Over-expression of Snail1 (which should down regulate E-cadherin function) is sufficient to induce EMT [Ikenouchi et al., 2003]. Snail1-null mice are early embryonic lethals. Conditional loss of Snail1 in murine neural crest cells has shown that Snail1 is not required for neural crest cell delamination and migration [Murray and Gridley, 2006a,b]. Redundant functions of Snail1 and Slug in cranial neural crest cells most likely account for a lack of phenotypic presentation in Snail1 conditional knock-out mice [Murray et al., 2006]. To date, no mutations in SNAIL have been associated with congenital malformations although the role of this gene in promoting epithelial-to-mesenchymal transitions in metastatic cancers is an area of active research [Herfs et al., 2010; Larriba et al., 2010; Wu and Zhou, 2010].
Slug, a marker of neural crest cells in Xenopus, zebrafish, and chick embryos, is also involved in NCC migration. Slug functions via its up-regulation of rhoB [del Barrio and Nieto, 2002], a member of the Rho family of GTPases, which regulates actin organization and membrane trafficking [Prendergast, 2001]. RhoB expression pattern correlates with regions of NCC delamination, suggesting that NCCs that have already gone through the EMT and are ready to migrate express RhoB [del Barrio and Nieto, 2004]. Furthermore, inactivation of RhoB impairs neural crest cell delamination [Liu and Jessell, 1998].
Homozygous deletions in Slug/SMAI2 that lead to the absence of the Slug protein have been identified in two unrelated patients with Waardenburg syndrome, type 2D (WS2D) (OMIM 608890) [Sanchez-Martin et al., 2002]. WS2D is a disorder primarily affecting the neural crest or neural crest derivatives, and is characterized by abnormal pigmentation, sensorineural hearing loss, and subtle differences in facial features [Read and Newton, 1997].
It is not surprising, then, that transcriptional misregulation of sets of genes involved in epithelial to mesenchymal transition of NCC like Snail and Slug may result in craniofacial defects. Such a mechanism is hypothesized to contribute to pathogenesis of CHARGE syndrome (OMIM 214800), a sporadic autosomal dominant disorder with cardinal features consisting of ocular coloboma, choanal atresia, heart malformations, growth restriction, genital and ear abnormalities (external and abnormal semicircular canals) [Verloes, 2005; Sanlaville and Verloes, 2007]. The face may also be square-shaped with a narrow bifrontal diameter, broad nasal bridge, malar flattening, cleft lip and palate bridge and a small mouth [Hall, 1979; Hittner et al., 1979; Oley et al., 1988; Felix et al., 2006; Jongmans et al., 2006; Sanlaville and Verloes, 2007]. Human mutations in the chromodomain helicase DNA-binding domain-7 member (CHD7) result in CHARGE syndrome [Vissers et al., 2004; Johnson et al., 2006; Jongmans et al., 2006; Sanlaville et al., 2006; Wincent et al., 2009] and account for 2/3 of the patients with this disorder [Sanlaville and Verloes, 2007]. Studies of CHD7 suggests that through enhancer elements it regulates transcription of genes involved in proliferation, differentiation and migration of cells that give rise to CHARGE-affected tissues [Zentner et al., 2010]. During human embryonic development CHD7 is expressed in the central nervous system and neural crest mesenchyme of the pharyngeal arches [Sanlaville et al., 2006]. It is felt that the underlying mechanism of this disorder involves abnormalities in NCCs [Siebert et al., 1985]. Knockdown studies in Xenopus laevis embryos using morpholino oligonucleotides targeting CHD7 resulted in abnormal migration of NCCs into the pharyngeal arches and a phenotype consistent with CHARGE syndrome [Bajpai et al., 2010]. Expression analysis in these embryos did not reveal down-regulation of genes (Pax3, Msx1, Zic1) involved in induction and formation of neural plate border but did show a down-regulation of genes (Slug/Snail, Twist, Sox9) that are involved in migration of NCC from the border. These data suggest that CHD7 is involved in the regulation of gene expression programs necessary for the formation of multipotent, migratory neural crest population but not in inductive events for specification of the neural plate border [Bajpai et al., 2010]. Human mutations in CHD7 may result in the same changes in NCC gene expression programs leading to the CHARGE phenotype. Mutations in SEMA3E which may be involved in NCC guidance have also been identified in CHARGE syndrome patients [Lalani et al., 2004].
GUIDANCE CUES FOR NEURAL CREST MIGRATION
NCCs follow well-delineated paths from the dorsal neural tube into the craniofacial region, and achieve this feat largely by communicating with the surrounding neural, facial, and pharyngeal epithelia, and cephalic mesoderm [Noden and Trainor, 2005]. NCCs utilize both repulsive and attractive factors as migratory guidance cues but the identities of most of these guidance cues have not been elucidated. There are, however, some notable exceptions: FGF-2 and FGF-8 both function in a chemotactic manner for NCCs [Kubota and Ito, 2000]. Semaphorins and Ephrin molecules also function as NCC guidance factors.
Perhaps the best studied of these guidance cues are the Ephrins, a family of ligands for the Eph receptors. Ephs constitute a subfamily of receptor tyrosine kinases with multiple functions [Nakamoto, 2000] including the modulation of the adhesion proteins such as members of the integrin family [Murai and Pasquale, 2003; Peinado et al., 2004b; Davy and Soriano, 2007]. Semaphorin proteins are involved in cellular processes such as axon guidance and cell migration. Mutations in semaphorin proteins are hypothesized to cause aberrant NCC migration and hence play an important role in pathogenesis of the CHARGE syndrome [Lalani et al., 2004].
NEURAL CREST PROLIFERATION AND DIFFERENTIATION
Once NCCs arrive at their final destination, how do they know what to do? Two main theories have been proposed. The first theory suggests that NCCs are intrinsically programmed [Noden, 1983] with a facial patterning “blueprint,” and that they carry this molecular patterning information with them upon departure from the neural tube. The second theory suggests that NCCs acquire facial patterning information from the environment in which they eventually find themselves [del Barrio and Nieto, 2002]. This debate has been experimentally addressed in a number of ways, most recently by using quail and duck chimaeras. Quail NCCs transplanted into duck embryos produce ducks with a quail-like face (the “quck”), whereas duck NCCs put into a quail embryo produce quails with a duck-like bill (the “duail”) [Schneider and Helms, 2003]. These data argue that NCCs contain some sort of molecular blueprint for the structures they will eventually form. Other data argue the opposite point: that the fate of NCCs is only determined once they arrive in the facial prominences, and that fate can be altered by simply adjusting the molecular signals that the NCCs see Hu et al. . So are NCCs pre-patterned or plastic? The prevailing opinion is that NCCs retain their multipotency even into late embryonic stages of development, which may explain why teratogens can exert their untoward effects even late in human gestation. On the other hand, even malformed faces retain definitive, species-specific characteristics and these appear to be immutable to a large degree.
Once in the prominences, NCCs proliferate and the facial structures begin to take shape. Many factors induce and regulate this proliferation. Abnormal regulation of proliferation contributes to Treacher Collins syndrome (TCS) (OMIM 154500), which manifests with severe craniofacial hypoplasia and dysplasia [Sakai and Trainor, 2009]. This autosomal dominant disorder is caused by mutations in TCOF1, which encodes for the Treacle protein. Treacle is a nuclear protein involved in ribosomal DNA gene transcription. Elegant experiments in mice have shown that haploinsufficiency (TCOF1 +/−) produces a significant decrease in the number of NCCs [Dixon et al., 2006]. This decrease in NCC is secondary to extensive neuroepithelial apoptosis and an accompanying decrease in NCC migration and proliferation [Dixon et al., 2006]. Apoptosis in TCS is caused by the activation of the p53 tumor suppressor pathway. It is hypothesized that mutation in the Treacle protein leads to deficient ribosome biogenesis and causes nuclear stress activation, which stabilizes p53. This stabilization results in activation of pro-apoptic genes and leads to the high degree of apoptosis observed in TCS. Inhibition of p53 provides a possible therapeutic approach in preventing the craniofacial birth defects seen in TCS or other neurocristopathies [Jones et al., 2008].
Other important factors involved in regulation of proliferation of NCC include the secreted proteins Sonic Hedgehog [Jeong et al., 2004] and Wnts [Brugmann et al., 2010]. Disruptions to these two pathways have profound effects on craniofacial morphogenesis. For example, Hedgehog signal transduction requires functional primary cilia; consequently, mutations in components of the primary cilia apparatus result in abnormal Hedgehog signaling, aberrant proliferation of neural crest cells and craniofacial malformations [Brugmann et al., 2010]. Based on these and other data, modulation of Hedgehog signaling may constitute a therapeutic approach to rescuing some types of facial dysmorphologies [Han et al., 2009].
NEURAL CREST DERIVATIVES IN THE FACE
NCCs make an integral contribution to the elaborate program of craniofacial development; consequently, we will devote the following paragraphs to outlining how the face develops, and the interactions between NCCs and adjacent cell populations that are required for normal craniofacial morphogenesis. The embryonic vertebrate face is made up of seven outgrowths or prominences: the singular frontonasal (FNP), and the paired lateral nasal, maxillary and mandibular prominences. We will consider each prominence separately because NCCs that populate each arch have distinct molecular signatures and interact with different epithelia during the development.
THE FACIAL PROMINENCES
The FNP gives rise to the forehead, middle of the nose, upper lip, philtrum, and primary palate [Helms et al., 2005]. Proper development of the FNP requires interactions between NCC and two epithelia, the forebrain and facial ectoderm [Hu and Helms, 1999]. Normally, the lateral region of the FNP fuses with the lateral nasal prominences (LNP) and medial region of the maxillary prominences (see below) to create the alae and columellae of the nose. The maxillary and mandibular prominences are both derived from the first pharyngeal arch. The maxillary prominences (MXP) [Helms et al., 2005] produce the upper jaw while the mandibular prominences (MNP) produce the lower jaw. The maxilla and mandible are both first arch derivatives and both require interactions between surface ectoderm, NCC, mesoderm and pharyngeal endoderm for their proper development [Couly et al., 2002; Ruhin et al., 2003].
INTERACTIONS BETWEEN NEURAL CREST CELLS AND CEPHALIC MESODERM
Although NCCs occupy the pharyngeal arches, they also have at their center a mass of mesodermal cells. These mesoderm-derived cells are surrounded by NCCs and externally covered by ectoderm and internally by endoderm [Graham, 2003] (Fig. 2). This elaborate arrangement has great significance for craniofacial development because disruptions in the interactions between mesoderm, NCC, and the epithelia have profound effects on craniofacial development. A functional mouth, for example, depends upon the coordinated development of the facial skeleton, which is derived from NCCs, and its associated musculature, which is derived from mesoderm. New data demonstrate that the NCCs act as the conductor for this coordinated morphogenetic program: signals emanating from NCCs instruct and inform mesodermal cells to differentiate into myoblast precursors, and then to organize themselves around the developing skeletal elements [Rinon et al., 2007; Grenier et al., 2009].
Much of what we now know about these tissues interactions originates from studies conducted in zebrafish that outline some of the functions of Endothelin (Edn) genes in craniofacial patterning. Edn1 encodes a secreted peptide Endothelin 1 (Edn 1) which signals from facial epithelia and mesoderm to the intervening NCCs that will later form the pharyngeal arch 1 (PA1) derived skeleton [Miller et al., 2000; Kimmel et al., 2001]. These zebrafish studies laid the foundation for understanding how Edn signaling regulated mammalian craniofacial development: for example, mice with a homozygous mutation in Edn1 have severely malformed bones in the jaw and throat [Kurihara et al., 1994] and defective musculature as well [Kimmel et al., 2003]. Data are now emerging from genetic studies in humans, which indicate that mutations in the Endothelin pathway contribute to Waardenburg–Shah syndrome (WS4A) (OMIM #277580; see Edery et al. ) and Hirschsprung disease (also known as aganglionic megacolon, OMIM #142623; see Brooks et al. ). Both Waardenburg and Hirschsprung disease are considered neurocristopathies; that is, diseases whose etiologies are attributable to defects in NCC behavior. In addition to other NCC-related defects, patients with WS4A may exhibit heterochromia, ocular ptosis, and hypertelorism.
The craniofacial features of the velo-cardio-facial/DiGeorge syndrome (DGS) (OMIM #188400) appear to result as a consequence of perturbation of crosstalk between the pharyngeal arch components. In addition to the spectrum of craniofacial dysmorphisms that includes cleft palate, neonatal hypocalcemia secondary to parathyroid hypoplasia, T cell deficiency resulting from hypoplasia or aplasia of the thymus gland, and a number of cardiac malformations that include the outflow tract [Ryan et al., 1997] are observed. The majority of cases involve a 1.5–3.0 Mb hemizygous interstitial deletion of chromosome 22q11.2 although terminal deletions and translocations that include 22q11.2 occur. This microdeletion includes between 30 and 50 genes. TBX1 a Tbox family of binding domain transcription factors is within this 22q11.2 critical region [Scambler, 2010]. In the developing pharyngeal arch Tbx1 is expressed in the endoderm and mesoderm as well as the epithelium of the palatal shelves and frontonasal process [Zoupa et al., 2006]. Decreased Tbx1 expression in the endoderm during pharyngeal arch and pouch development and epithelia of the palatal shelves [Zoupa et al., 2006] is believed to result in secondary NCC abnormalities (extrinsic to NCCs) [Walker and Trainor, 2006]. Supporting this idea, Tbx1 knockout mice show pharyngeal arch and pouch defects, which underlie the mouse phenotype and recapitulate the human phenotype [Walker and Trainor, 2006]. The cononical Wnt-beta-catenin signaling pathway has recently been shown to regulate Tbx1 expression. In mice beta-catenin inactivation results in malformations within the spectrum of DGS phenotypes and may be involved modifying the severity of DGS [Huh and Ornitz, 2010].
NEURAL CREST CELLS AND OTHER CRANIOFACIAL MALFORMATIONS
Identifying a molecular basis for one of the most common craniofacial malformations, clefting, has also been pursued heavily. Secondary palate is a product of MXPs while the primary palate is derived from the FNP. The palatal shelves themselves consist of NCC mesenchyme covered by surface and oral epithelia. The palatal shelves must perform an elaborate outgrowth in order for fusion to occur: first, the MXPs must expand and then must undergo a rotation from their initial vertical position to a horizontal position dorsal to the tongue, and finally fuse. If the embryonic tongue is too large, or if it fails to descend because it is ankylosed, this can impede the elevation of the palatal shelves and consequently cause secondary palatal clefting.
Secondary Palatal Clefting Caused by TBX22 Mutations
In the last decade, investigators have determined that a X-linked condition where cleft palate is accompanied by ankyloglossia (CPX) (OMIM #303400) may be caused by mutations in a member of T-box transcription factor, TBX22 [Braybrook et al., 2001]. Unlike other types of clefting that have been attributed to environmental factors, CPX is recognized as an X-linked semi-dominant condition. Members of T-box family play an important role in vertebrate development, especially in mesoderm specification [Papaioannou and Silver, 1998]. Precisely how TBX22 regulates mesoderm-NCC interactions has yet to be determined.
Secondary Palatal Clefting Associated With Defects in Wnt Signaling
There are other causes for secondary palatal clefting, attributable to defects in MXP outgrowth. For example, in mice with mutation in Wnt 9b, in particular, the basis for the clefting phenotype appears to be insufficient growth of the maxillary prominences [Lan et al., 2006]. Consequently, these prominences, from which the palatal shelves derive, fail to approximate and the result is palatal clefting [Juriloff et al., 2006]. In mammals, Wnt signaling is critical for the proliferation of NCCs within the MXPs [Brugmann et al., 2006, 2007] and disruptions in Wnt signaling specifically and profoundly influence the ability of NCC-derived facial prominences to fully develop [Brugmann et al., 2007]. As a consequence of down-regulated Wnt signaling, the FNP and MXPs are truncated and the result is palatal (and facial) clefting.
Palatal Clefting and FOXE1 Defects
FOXE1 (Forkhead box protein E1) is a member of a transcription factor family that is involved in embryonic pattern formation. Through positional cloning, candidate gene sequencing, and developmental gene expression analyses, a strong correlation has been discovered in between mutations in FOXE1 and the occurrence of cleft lip and palate in humans [Moreno et al., 2009]. Foxe1 is expressed in the secondary palate epithelium of mice [Dathan et al., 2002] and human embryos [Trueba et al., 2005]. Furthermore, mice with a null mutation in Titf2/Foxe1 have cleft palate and thyroid anomalies [De Felice et al., 1998]. The Foxe1 is expressed at the point of fusion between the MXPs and the FNP, further supporting its role in palatogenesis [Moreno et al., 2009].
Palatal Clefting and IRF6
Another gene associated with facial clefting is the IRF6 (interferon regulatory factor) gene, which is part of a larger family of transcription factors that bind to specific DNA sequences and regulate gene expression. In mice, disruption in Irf6 function results in clefting phenotypes [Ingraham et al., 2006; Rahimov et al., 2008]. In humans, mutations in IRF6 cause Van der Woude syndrome and popliteal pterygium syndrome, two clefting disorders [Little et al., 2009]. There is also an increase risk of isolated cleft lip and palate in humans with variations in IRF6 [Blanton et al., 2005]. The occurrence of clefting in Irf6 mutant mice is hypothesized to be caused by a defect in elevation of palatine shelves, secondary to inappropriate adhesions between the palatal shelves and oral epithelium [Ingraham et al., 2006].
Palatal Clefting and EGF Receptor Mutations
Once the palatal shelves have approximated, they must fuse. For this to occur, the medial edge epithelium must be removed in order for the NCC-derived mesenchyme to become confluent. This typically happens when the epithelial cells covering the palatal mesenchyme undergo programmed cell death and/or cell migration. Transforming growth factor-alpha (TGFα), an epidermal growth factor receptor (EGFR) ligand that is expressed in facial epithelia, plays a role in this fusion process. For example, Egfr−/− mice have midline defects that produce an elongated primary palate, micrognathia, and a high incidence of cleft palate [Miettinen et al., 1999]. In vitro experiments suggest that a delay in epithelial degeneration occurs in the absence of EGFR signaling. The molecular basis for this delay in epithelial degeneration and clefting seen in Egfr−/− embryos is associated with the loss of function of matrix metalloproteinases (MMPs), which are endopeptidases that cleave the extracellular matrix. These experiments illustrate the delicate balance between matrix remodeling and epithelial seam removal that is required for proper craniofacial fusions.
In conclusion, disruptions in the rate, the timing, or the extent of outgrowth of the facial prominences can all result in facial clefting and often these disruptions are attributable to defects in NCC behavior. Clefting malformations occur in approximately 1 out of 700 births, making it one of the most prevalent craniofacial birth defect; and the role of NCCs in this process cannot be overemphasized.
PLURIPOTENCY AND MULTIPOTENCY OF THE NEURAL CREST
NCCs, like hematopoietic stem cells, exhibit a hierarchical progression from being initially pluripotent to becoming progressively more restricted in their developmental potential. By means of in vitro serial subcloning, Le Douarin and coworkers identified both multipotent and oligopotent NCC progenitors that differed in their developmental repertoire, including their ability to self-renew [Trentin et al., 2004]. Whether NCCs actually retain their pluripotency into adulthood, however, is not entirely clear.
One approach to addressing this question is to follow the fate of NCCs in the adult animal. NCCs give rise to the embryonic facial skeleton, and they are also responsible for maintaining facial skeletal elements into adulthood [Leucht et al., 2008]. When NCC-derived skeletal stem cells are transplanted, they retain the ability to differentiate into chondrocytes and osteoblasts [Leucht et al., 2008]. These studies indicate that adult NCC-derived skeletal stem cells retain a bi-potential fate but they fall short of demonstrating that adult NCCs retain their self-renewing capacity.
THE FUTURE OF CRANIOFACIAL DEVELOPMENT
Much of our present knowledge of normal craniofacial development has come from genetic mutations in humans, genetically engineered mice and the results of embryonic exposure to teratogens in both humans and animals. Future advancements in our understanding of craniofacial development will come from utilizing new technologies and exploiting the knowledge of gene regulation provided by other disciplines, in particular the field of epigenetics. Understanding the importance of epigenetic regulation during neural crest cell development has already made an impact on understanding the etiology of craniofacial syndromes [Bajpai et al., 2010]. Furthermore, techniques such as next-generation sequencing can serve to identify patient-specific mutations in sporadic as well as congenital cases of craniofacial disorders, whereas recent advances in neural crest cell culture provide model systems of both indefinitely self-renewing primary human neural crest cultures [Thomas et al., 2008] and lineage-specific differentiation of pluripotent human embryonic stem cells into neural crest cells [Bajpai et al., 2010]. The exploitation of these techniques along with continued careful studies in craniofacial biology will hopefully provide an avenue for the accurate diagnosis and possible treatment of craniofacial disorders.
The authors would like to note the valuable insights and suggestions from Alan Shanske, Danny Huylebroeck, Paul Trainor, and James Fraser. The manuscript described was supported by Award Number K12HD001255 from the Eunice Kennedy Shriver National Institute of Child Health & Human Development and the Eleanor and Miles Shore Scholars in Medicine, Harvard Medical School to D.C. The content is solely the responsibility of the authors and does not necessarily represent the official views of the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health. This work was also sponsored by the National Institute of Health (NIH) grants K99DE019853-01 to S.A.B.