Why Does the Face Predict the Brain? Neural Crest Induction, Craniofacial Morphogenesis, and Neural Circuit Development

Abstract

Mesenchephalic and rhombencephalic neural crest cells generate the craniofacial skeleton, special sensory organs, and subsets of cranial sensory receptor neurons. They do so while preserving the anterior-posterior (A-P) identity of their neural tube origins. This organizational principle is paralleled by central nervous system circuits that receive and process information from facial structures whose A-P identity is in register with that in the brain. Prior to morphogenesis of the face and its circuits, however, neural crest cells act as "inductive ambassadors" from distinct regions of the neural tube to induce differentiation of target craniofacial domains and establish an initial interface between the brain and face. At every site of bilateral, non-axial secondary induction, neural crest constitutes all or some of the mesenchymal compartment for non-axial mesenchymal/epithelial (M/E) interactions. Thus, for epithelial domains in the craniofacial primordia, aortic arches, limbs, the spinal cord, and the forebrain (Fb), neural crest-derived mesenchymal cells establish local sources of inductive signaling molecules that drive morphogenesis and cellular differentiation. This common mechanism for building brains, faces, limbs, and hearts, A-P axis specified, neural crest-mediated M/E induction, coordinates differentiation of distal structures, peripheral neurons that provide their sensory or autonomic innervation in some cases, and central neural circuits that regulate their behavioral functions. The essential role of this neural crest-mediated mechanism identifies it as a prime target for pathogenesis in a broad range of neurodevelopmental disorders. Thus, the face and the brain "predict" one another, and this mutual developmental relationship provides a key target for disruption by developmental pathology.

Keywords: 22q11 deletion syndrome; inductive signaling; neural crest; olfactory; placodes; sensory pathways.

Figure 1

Neural crest mediated mesenchymal/epithelial (M/E) induction prefigures nasal/forebrain (Fb), craniofacial, heart, and limb morphogenesis. (A) A summary of the sites of non-axial M/E induction and their morphogenetic endpoints. The arrows point to the embryonic regions illustrated in panel B. (B) A summary of the relationship between anterior-posterior (A-P) regionally specified neural crest and the sites of M/E induction that establish the nose and Fb, the face, the heart, and the limbs. At each site, a primarily neural crest-derived population of A-P specified mesenchymal cells is opposed to the adjacent surface ectoderm, which is also axially specified. (C) Subsets of neural crest-derived mesenchymal cells, labeled with a knock-in reporter transgene (βgeo6; LaMantia et al., 2000; Bhasin et al., 2003) at each of the sites of M/E induction produce the morphogenetic signaling molecule retinoic acid (RA). These cells drive locally patterned expression of several target genes in placodal domains (purple shading) immediately adjacent to the RA-producing mesenchymal cells.

Figure 2

The distribution of RA producing neural crest at sites of M/E induction and its relationship to epithelia and mesenchymal sources of additional cardinal inductive signals. (A) Subsets of frontonasal (panels 1, 2), branchial arch (panels 5, 6), aortic arch (panels 7, 8), and forelimb bud (Flb; panels 9–11) are labeled by the βgeo6 reporter. These cells are coincident with Raldh2-expressing cells in the frontonasal mesenchyme (FnM; panels 3, 4) as well as other sites of non-axial M/E interaction. The dotted lines in panels 1, 7, and 9 indicate the approximate plane of the sections shown in panels 2, 8, 10, and 11, respectively. (B) In situ hybridization identifies local expression of cardinal inductive/morphogenetic signals Fgf8, Shh, and Bmp4 in epithelial as well as mesenchymal domains in the frontonasal process (FnP), mandibular arch (Ba1b), and Flb. Fgf8 and Shh are limited to epithelial domains in the medial nasal process (mnp) while Bmp4 is enhanced in the lateral nasal process epithelium. Fgf8 is found in a limited dorsal-lateral epithelial domain in Ba1b, Shh in a medial domain, and Bmp4 in a dorsal medial location. In the Flb, these three cardinal signals define the apical ectodermal ridge (aer) and zone of polarizing activity (zpa), two embryologically defined signaling regions that drive limb morphogenesis and patterning (REFS). (C) Schematic summary of the localization and signaling interactions (arrows) of local mesenchymal and epithelial sources of RA, Fgf8, Shh, and Bmp4 in the frontonasal mass/Fb (top) and Flb (bottom) during the initial establishment of these sites of non-axial neural crest-mediated M/E induction (Embryonic day E9.5 in the mouse) and as signaling and morphogenesis moves forward (E11.5). The direction of the arrows was determined using in vitro mesenchymal/epithelial co-cultures or isolated explants of the epithelium or mesenchyme alone (LaMantia et al., 2000; Bhasin et al., 2003).

Figure 3

The sequence of neural crest-mediated M/E induction and its consequences for local patterning, neuronal differentiation, and initial establishment of the axon growth and targeting from the olfactory placode. The top panels show the stepwise initial development of the olfactory placode (op; blue shading), olfactory epithelium (OE; blue shading), and olfactory receptor neurons (ORNs) and their axons that constitute the nascent olfactory nerve (ON; red). A sub-population of mesencephalic/diencephalic neural crest cells in the FnM (green) produce RA and establish domains of RA-mediated gene expression (blue shading) in the Fb as well as the olfactory periphery. This Fb domain will generate the olfactory bulb (OB), the target of the axons from the OE via the ON. The middle panels summarize the inductive, patterning, sensory neuron differentiation and axon outgrowth, peripheral and brain morphogenetic events diagramed in the top panels. The bottom panels show the disruption of neural crest-mediated M/E interaction in the Pax6 ^Sey/Sey mutant and its consequences for each subsequent step of initial olfactory pathway formation (top panel adapted from LaMantia et al., 1993; middle and bottom panels adapted from LaMantia et al., 1993, 2000; Anchan et al., 1997; Balmer and LaMantia, 2004; Tucker et al., 2010).

Figure 4

The relationship between nascent cranial sensory neurons, ectodermal placode, and neural crest-derived neural progenitors and neuroblasts in the cranial sensory ganglia at midgestation (E10.5) in the mouse. (A) The neuronal microtubule protein βIII-tubulin is expressed in newly generated ORNs (OE, top panel) as well as cranial sensory neurons in the trigeminal (gCN V), facial (gCN VII), spiral (gCN VIII), glossopharyngeal (gCN IX), and vagal (gCN X) cranial nerve ganglia and their axons that extend toward central (OE, gCNV, VII, VIII, IX, and X) as well as peripheral (gCNV, VII, VIII, IX, and X) targets. (B) Six1 (red), a marker of placode-associated cells and Wnt1:Cre recombination-mediated expression of a conditional GFP reporter allele (green), shows the relationship between placode-associated cells and neural crest-derived cells in the OE, FnM and cranial ganglia. Cells in the OE are labeled exclusively by Six1. Cells in the FnM are uniformly labeled by the Wnt1:Cre reporter, but a subset of them in the lateral nasal process also expresses Six1. Each of the cranial nerve ganglia, except for gCN VIII, is composed of primarily Six1-expressing placode-derived cells. The mesenchyme between the cranial nerve ganglia and the hindbrain at this stage of development has cells that express Six1 as well as the Wnt1:Cre reporter, as is the case for the cranial epithelium in the periphery. (C) Relationship between Six1-expressing, Wnt1:Cre reporter-expressing, and HuC/D-expressing cells in the OE and cranial nerve ganglia. In the OE, HuC/D-expressing newly generated neurons (blue) are scattered through the epithelium and have downregulated Six1 (arrows). In addition, there is a population of HuC/D expressing neurons (arrowheads) in the FnM that have also downregulated Six1 and are not labeled by the Wnt1:Cre reporter. These cells are most likely the GnRH-expressing neurons that migrate from the OE to enter the ventral Fb along newly extending ORN axons at this stage of development. In gCN V, gCN VII, and gCN IX/X, HuC/D-expressing neurons are coincident with cells labeled by Six1, the Wnt1:Cre reporter, or both (adapted from Karpinski et al., 2016).

Figure 5

A large subset of mouse orthologues of the genes on human Chr. 22 deleted in DiGeorge/22q11 Deletion syndrome (22q11DS) are expressed in the developing or adult brain as well as sites of neural crest-mediated M/E induction at midgestation. (A) The location on mouse chromosome 16 of 28/32 orthologues of the genes in the minimal critical deleted region of human chromosome 22 whose heterozygous deletion causes 22q11DS. (B) PCR, in situ hybridization, immunoblotting, and immunlocalization identify expression of 22 of the 28 murine 22q11 orthologues in the developing and adult mouse brain. (C) Multiple 22q11 orthologues are expressed uniformly at sites of neural crest-mediated M/E induction as well as in the nascent central nervous system in the midgestation mouse embryo (E10.5). The purple-blue label shows in situ hybridization labeling of mRNA for several 22q11 genes at these sites, and the brown label shows the localization of proteins encoded by three of the deleted genes (panel A, B, adapted from Meechan et al., 2015; panel C adapted from Motahari et al., 2019).

Figure 6

The mutant gene in a rare monogenic disorder characterized clinically by craniofacial, limb, and Fb anomalies is initially expressed focally and maximally at sites of neural crest-mediated M/E induction. (A) Craniofacial, brain, hand (forelimb), and foot (hindlimb) anomalies in a 19-year-old male. This individual also had developmental delay, poor academic performance, and poor language skills from an early age. (B) Mapping and confirming the causal mutant gene for this Mendelian, monogenic disorder. The mutant gene SPG20, is a microtubule-interacting trafficking molecule involved in multiple signaling and metabolic cellular processes. The mutation in this Omani pedigree is a novel SPG20 two base pair deletion mutation that results in undetectable expression of SPG20 in fibroblasts from affected individuals in the pedigree, as well as undectectable Spartin protein expression. (C) Localization of Spg20, the murine orthologue of SPG20 by in situ hybridization in an E10.5 mouse embryo, shows focal, selective expression at sites of neural crest-mediated M/E induction, including FnM and Fb, the maxillary process (mx), and as well as the nascent mandibular process (ba1), the hyoid process (ba2), the heart (h), and Flb. (D) qPCR in microdissected frontonasal mass/Fb, branchial arches, h and Flb confirms enhanced expression of Spg20 at these M/E inductive sites. These expression levels, especially for the fnm/fb, ba1/2, and flb, are substantially elevated above the expression level detected in the whole E10.5 embryo (wh; adapted from Manzini et al., 2010).

Figure 7

Early disruption of hindbrain patterning alters anterior cranial nerve differentiation, prefiguring anomalous oropharyngeal sensory/motor function that likely contributes to suckling, feeding, and swallowing (S/F/S) difficulties in early post-natal LgDel mouse pups who carry a heterozygous deletion of the 28 murine orthologues of the genes deleted in 22q11DS. (A) The five cranial nerves that contribute to sensory/motor control of S/F/S have begun to differentiate by E10.5 in the mouse. In this preparation, they have been immunolabeled in the whole by the early marker for neuron and axons, βIII-tubulin, and visualized in a high-resolution confocal image in which embryo volume/depth is color coded, with warm colors representing structures close to the viewer and cooler colors representing those deeper in the embryo. The inset shows the multiple small axon fascicles that characterize the maxillary branch (Mx) of the trigeminal nerve (V) and the single fascicle of axons that forms as the mandibular branch in typically developing WT embryos. (B) The A-P array of S/F/S contributing cranial nerves is prefigured in E9.5 embryos by a gradient of RA-signaling that distinguishes posterior (r5,6) from anterior (r2,3) rhombomeres in the developing hindbrain. This posterior RA-dependent patterning, as well as opposing anterior signaling via Fgfs and Wnts, specifies the precursors of the cranial sensory neurons and hindbrain motor neurons that then differentiate as the cranial nerves within 24 h. (C) In LgDel E9.5 embryos, the gradient of RA signaling is enhanced in and shifted beyond posterior rhombomeres; it now elicits RA-regulated gene expression in anterior rhombomeres. Within a day, anterior cranial nerves, V (trigeminal) and VII (facial) are dysmorphic. The multiple axon fascicles normally seen in the Mx of V are diminished, the mandibular branch is similarly hypotrophic, and the facial nerve (VII) lacks its nascent anterior branch (arrow). In addition, the posterior cranial nerves IX (glossopharyngeal) and X (vagus) have either small axonal anastomoses (arrowhead) or in extreme cases are fused. (D) When RA signaling levels are diminished genetically by heterozygous deletion of the RA synthetic gene Raldh2 in LgDel embryos (“Rescue”), the pattern of RA-dependent gene expression in the anterior rhombomeres returns to that seen in the WT. In parallel, initial differentiation of the nascent trigeminal and facial nerve is restored to the WT state. The ophthalamic (Op), Mx, and mandibular branches of the trigeminal nerve (V) extend toward their targets as in the WT with similar degrees of fasciculation. The facial nerve branches appropriately (arrow). The fusion of the posterior cranial nerves IX and X persists, most likely because this reflects the disrupted differentiation of cardiovascular targets due to Tbx1 heterozygous deletion, independent of hindbrain RA-dependent A-P patterning. (E–H) Schematics of the relationship between RA-dependent hindbrain patterning and the growth and trajectory of individual trigeminal motor and sensory axons in the WT embryo. Individual trigeminal motor axons, as well as primarily placodal derived trigeminal sensory axons, respond differently as they interact with neural crest derived mesenchymal substrates in the periphery whose A-P identity has been presumably altered by enhanced RA signaling in the anterior rhombomeres.

Figure 8

Coordination of A-P identity in the nascent central nervous system and peripheral sites of neural crest-mediated M/E induction prefigures coordinated differentiation of peripheral sensory organs, the heart and viscera, and the limbs as well as neural circuits that control each structure in the peripheral and central nervous system. (A) A summary of the A-P locations in the neural tube that generate neural crest and the brain regions they reflect. (B) The potential relationship between peripheral structures, sensory organs, and sensory ganglia induced or patterned by non-axial neural crest-mediated M/E interactions and the central neural circuits that control the function of these structures. It is unclear whether the coordination of peripheral induction via the neural crest in the A-P axis and corresponding regions of the neural tube has a direct influence on the regional differentiation of anterior Fb regions that process relevant information, with the exception of the OB. The locations of the relevant regions of the cerebral cortex that receive thalamic (diencephalic) inputs from relay nucleic for vision (eye), audition (ear), and somatosensation (sensory cranial ganglia) are indicated for completeness.