An essential role for Fgfs in endodermal pouch formation influences later craniofacial skeletal patterning

Fibroblast growth factor (Fgf) proteins are important regulators of pharyngeal arch development. Analyses of Fgf8 function in chick and mouse and Fgf3 function in zebrafish have demonstrated a role for Fgfs in the differentiation and survival of postmigratory neural crest cells (NCC) that give rise to the pharyngeal skeleton. Here we describe, in zebrafish, an earlier, essential function for Fgf8 and Fgf3 in regulating the segmentation of the pharyngeal endoderm into pouches. Using time-lapse microscopy, we show that pharyngeal pouches form by the directed lateral migration of discrete clusters of endodermal cells. In animals doubly reduced for Fgf8 and Fgf3, the migration of pharyngeal endodermal cells is disorganized and pouches fail to form. Transplantation and pharmacological experiments show that Fgf8 and Fgf3 are required in the neural keel and cranial mesoderm during early somite stages to promote first pouch formation. In addition, we show that animals doubly reduced for Fgf8 and Fgf3 have severe reductions in hyoid cartilages and the more posterior branchial cartilages. By examining early pouch and later cartilage phenotypes in individual animals hypomorphic for Fgf function, we find that alterations in pouch structure correlate with later cartilage defects. We present a model in which Fgf signaling in the mesoderm and segmented hindbrain organizes the segmentation of the pharyngeal endoderm into pouches. Moreover, we argue that the Fgf-dependent morphogenesis of the pharyngeal endoderm into pouches is critical for the later patterning of pharyngeal cartilages.     

Requirements for endoderm and BMP signaling in sensory neurogenesis in zebrafish

Cranial sensory neurons largely derive from neurogenic placodes (epibranchial and dorsolateral), which are ectodermal thickenings that form the sensory ganglia associated with cranial nerves, but the molecular mechanisms of placodal development are unclear. Here, we show that the pharyngeal endoderm induces epibranchial neurogenesis in zebrafish, and that BMP signaling plays a crucial role in this process. Using a her5:egfp transgenic line to follow endodermal movements in living embryos, we show that contact between pharyngeal pouches and the surface ectoderm coincides with the onset of neurogenesis in epibranchial placodes. By genetic ablation and reintroduction of endoderm by cell transplantation, we show that these contacts promote neurogenesis. Using a genetic interference approach we further identify bmp2b and bmp5 as crucial components of the endodermal signals that induce epibranchial neurogenesis. Dorsolateral placodes (trigeminal, auditory, vestibular, lateral line) develop independently of the endoderm and BMP signaling, suggesting that these two sets of placodes are under separate genetic control. Our results show that the endoderm regulates the differentiation of cranial sensory ganglia, which coordinates the cranial nerves with the segments that they innervate.     

An integrin-dependent role of pouch endoderm in hyoid cartilage development

Pharyngeal endoderm is essential for and can reprogram development of the head skeleton. Here we investigate the roles of specific endodermal structures in regulating craniofacial development. We have isolated an integrinα5 mutant in zebrafish that has region-specific losses of facial cartilages derived from hyoid neural crest cells. In addition, the cranial muscles that normally attach to the affected cartilage region and their associated nerve are secondarily reduced in integrinα5⁻ animals. Earlier in development, integrinα5 mutants also have specific defects in the formation of the first pouch, an outpocketing of the pharyngeal endoderm. By fate mapping, we show that the cartilage regions that are lost in integrinα5 mutants develop from neural crest cells directly adjacent to the first pouch in wild-type animals. Furthermore, we demonstrate that Integrinα5 functions in the endoderm to control pouch formation and cartilage development. Time-lapse recordings suggest that the first pouch promotes region-specific cartilage development by regulating the local compaction and survival of skeletogenic neural crest cells. Thus, our results reveal a hierarchy of tissue interactions, at the top of which is the first endodermal pouch, which locally coordinates the development of multiple tissues in a specific region of the vertebrate face. Lastly, we discuss the implications of a mosaic assembly of the facial skeleton for the evolution of ray-finned fish.     

Development and evolution of chordate cartilage

Deuterostomes are a monophyletic group of animals containing vertebrates, lancelets, tunicates, hemichordates, echinoderms, and xenoturbellids. Four out of these six extant groups-vertebrates, lancelets, tunicates, and hemichordates-have pharyngeal gill slits. All groups of deuterostome animals that have pharyngeal gill slits also have a pharyngeal skeleton supporting the pharyngeal openings, except tunicates. We previously found that pharyngeal cartilage in hemichordates and cephalochordates contains a fibrillar collagen protein similar to vertebrate type II collagen, but unlike vertebrate cartilage, the invertebrate deuterostome cartilages are acellular. We found SoxE and fibrillar collagen expression in the pharyngeal endodermal cells adjacent to where the cartilages form. These same endodermal epithelial cells also express Pax1/9, a marker of pharyngeal endoderm in vertebrates, lancelets, tunicates, and hemichordates. In situ experiments with a cephalochordate fibrillar collagen also showed expression in pharyngeal endoderm, as well as the ectoderm and the mesodermal coelomic pouches lining the gill bars. These results indicate that the pharyngeal endodermal cells are responsible for secretion of the cartilage in hemichordates, whereas in lancelets, all the pharyngeal cells surrounding the gill bars, ectodermal, endodermal, and mesodermal may be responsible for cartilage formation. We propose that endoderm secretion was primarily the ancestral mode of making pharyngeal cartilages in deuterostomes. Later the evolutionary origin of neural crest allowed co-option of the gene network for the secretion of pharyngeal cartilage matrix in the new migratory neural crest cell populations found in vertebrates.     

Alcama mediates Edn1 signaling during zebrafish cartilage morphogenesis

The zebrafish pharyngeal cartilage is derived from the pharyngeal apparatus, a vertebrate-specific structure derived from all three germ layers. Developmental aberrations of the pharyngeal apparatus lead to birth defects such as Treacher-Collins and DiGeorge syndromes. While interactions between endoderm and neural crest (NC) are known to be important for cartilage formation, the full complement of molecular players involved and their roles remain to be elucidated. Activated leukocyte cell adhesion molecule a (alcama), a member of the immunoglobulin (Ig) superfamily, is among the prominent markers of pharyngeal pouch endoderm, but to date no role has been assigned to this adhesion molecule in the development of the pharyngeal apparatus. Here we show that alcama plays a crucial, non-autonomous role in pharyngeal endoderm during zebrafish cartilage morphogenesis. alcama knockdown leads to defects in NC differentiation, without affecting NC specification or migration. These defects are reminiscent of the phenotypes observed when Endothelin 1 (Edn1) signaling, a key regulator of cartilage development is disrupted. Using gene expression analysis and rescue experiments we show that Alcama functions downstream of Edn1 signaling to regulate NC differentiation and cartilage morphogenesis. In addition, we also identify a role for neural adhesion molecule 1.1 (nadl1.1), a known interacting partner of Alcama expressed in neural crest, in NC differentiation. Our data shows that nadl1.1 is required for alcama rescue of NC differentiation in edn1^−/− mutants and that Alcama interacts with Nadl1.1 during chondrogenesis. Collectively our results support a model by which Alcama on the endoderm interacts with Nadl1.1 on NC to mediate Edn1 signaling and NC differentiation during chondrogenesis.     

Wnt Signaling Interacts with Bmp and Edn1 to Regulate Dorsal-Ventral Patterning and Growth of the Craniofacial Skeleton

Craniofacial development requires signals from epithelia to pattern skeletogenic neural crest (NC) cells, such as the subdivision of each pharyngeal arch into distinct dorsal (D) and ventral (V) elements. Wnt signaling has been implicated in many aspects of NC and craniofacial development, but its roles in D-V arch patterning remain unclear. To address this we blocked Wnt signaling in zebrafish embryos in a temporally-controlled manner, using transgenics to overexpress a dominant negative Tcf3, (dntcf3), (Tg(hsp70I:tcf3-GFP), or the canonical Wnt inhibitor dickkopf1 (dkk1), (Tg(hsp70i:dkk1-GFP) after NC migration. In dntcf3 transgenics, NC cells in the ventral arches of heat-shocked embryos show reduced proliferation, expression of ventral patterning genes (hand2, dlx3b, dlx5a, msxe), and ventral cartilage differentiation (e.g. lower jaws). These D-V patterning defects resemble the phenotypes of zebrafish embryos lacking Bmp or Edn1 signaling, and overexpression of dntcf3 dramatically reduces expression of a subset of Bmp receptors in the arches. Addition of ectopic BMP (or EDN1) protein partially rescues ventral development and expression of dlx3b, dlx5a, and msxe in Wnt signaling-deficient embryos, but surprisingly does not rescue hand2 expression. Thus Wnt signaling provides ventralizing patterning cues to arch NC cells, in part through regulation of Bmp and Edn1 signaling, but independently regulates hand2. Similarly, heat-shocked dkk1+ embryos exhibit ventral arch reductions, but also have mandibular clefts at the ventral midline not seen in dntcf3+ embryos. Dkk1 is expressed in pharyngeal endoderm, and cell transplantation experiments reveal that dntcf3 must be overexpressed in pharyngeal endoderm to disrupt D-V arch patterning, suggesting that distinct endodermal roles for Wnts and Wnt antagonists pattern the developing skeleton.     

Differentiation of cartilage from cranial neural crest in the axolotl (Ambystoma mexicanum)

Explants of cranial neural crest from neurula-stage Ambystoma mexicanum embryos form cartilage nodules in 10-14 days, when cultured with pharyngeal endoderm. The time course of formation of the nodules, and their appearance, correspond closely to that observed for visceral cartilage in vivo. Endoderm from any area of the sheet surrounding the pharyngeal cavity can induce cartilage formation, but endoderm from regions posterior to the pharyngeal cavity cannot. No other tissues are required for induction in vitro. Cranial neural crest cultured without inductive endoderm did not yield cartilage when taken prior to the stage at which migration begins, indicating that the crest was not determined for cartilage formation at this time. However, a small proportion of the cultures from neural crest taken during the early migratory phase did form cartilage, suggesting that interactions leading to their determination had begun.     

Zebrafish wnt9a is expressed in pharyngeal ectoderm and is required for palate and lower jaw development

Wnt activity is critical in craniofacial morphogenesis. Dysregulation of Wnt/β-catenin signaling results in significant alterations in the facial form, and has been implicated in cleft palate phenotypes in mouse and man. In zebrafish, we show that wnt9a is expressed in the pharyngeal arch, oropharyngeal epithelium that circumscribes the ethmoid plate, and ectodermal cells superficial to the lower jaw structures. Alcian blue staining of morpholino-mediated knockdown of wnt9a results in loss of the ethmoid plate, absence of lateral and posterior parachordals, and significant abrogation of the lower jaw structures. Analysis of cranial neural crest cells in the sox10:eGFP transgenic demonstrates that the wnt9a is required early during pharyngeal development, and confirms that the absence of Alcian blue staining is due to absence of neural crest derived chondrocytes. Molecular analysis of genes regulating cranial neural crest migration and chondrogenic differentiation suggest that wnt9a is dispensable for early cranial neural crest migration, but is required for chondrogenic development of major craniofacial structures. Taken together, these data corroborate the central role for Wnt signaling in vertebrate craniofacial development, and reveal that wnt9a provides the signal from the pharyngeal epithelium to support craniofacial chondrogenic morphogenesis in zebrafish.     

Requirement for endoderm and FGF3 in ventral head skeleton formation

The vertebrate head skeleton is derived in part from neural crest cells, which physically interact with head ectoderm, mesoderm and endoderm to shape the pharyngeal arches. The cellular and molecular nature of these interactions is poorly understood, and we explore here the function of endoderm in this process. By genetic ablation and reintroduction of endoderm in zebrafish, we show that it is required for the development of chondrogenic neural crest cells, including their identity, survival and differentiation into arch cartilages. Using a genetic interference approach, we further identify Fgf3 as a critical component of endodermal function that allows the development of posterior arch cartilages. Together, our results reveal for the first time that the endoderm provides differential cues along the anteroposterior axis to control ventral head skeleton development and demonstrate that this function is mediated in part by Fgf3.     

sucker encodes a zebrafish Endothelin-1 required for ventral pharyngeal arch development

Mutation of sucker (suc) disrupts development of the lower jaw and other ventral cartilages in pharyngeal segments of the zebrafish head. Our sequencing, cosegregation and rescue results indicate that suc encodes an Endothelin-1 (Et-1). Like mouse and chick Et-1, suc/et-1 is expressed in a central core of arch paraxial mesoderm and in arch epithelia, both surface ectoderm and pharyngeal endoderm, but not in skeletogenic neural crest. Long before chondrogenesis, suc/et-1 mutant embryos have severe defects in ventral arch neural crest expression of dHAND, dlx2, msxE, gsc, dlx3 and EphA3 in the anterior arches. Dorsal expression patterns are unaffected. Later in development, suc/et-1 mutant embryos display defects in mesodermal and endodermal tissues of the pharynx. Ventral premyogenic condensations fail to express myoD, which correlates with a ventral muscle defect. Further, expression of shh in endoderm of the first pharyngeal pouch fails to extend as far laterally as in wild types. We use mosaic analyses to show that suc/et-1 functions nonautonomously in neural crest cells, and is thus required in the environment of postmigratory neural crest cells to specify ventral arch fates. Our mosaic analyses further show that suc/et-1 nonautonomously functions in mesendoderm for ventral arch muscle formation. Collectively our results support a model for dorsoventral patterning of the gnathostome pharyngeal arches in which Et-1 in the environment of the postmigratory cranial neural crest specifies the lower jaw and other ventral arch fates.     

tgfβ3 regulation of chondrogenesis and osteogenesis in zebrafish is mediated through formation and survival of a subpopulation of the cranial neural crest

Zebrafish tgfβ3 is strongly expressed in a subpopulation of the migrating neural crest cells, developing pharyngeal arches and neurocranial cartilages. To study the regulatory role of tgfβ3 in head skeletal formation, we knocked down tgfβ3 in zebrafish and found impaired craniofacial chondrogenesis, evident by malformations in selected neurocranial and pharyngeal arch cartilages. Over-expressing tgfβ3 in embryos resulted in smaller craniofacial cartilages without any gross malformations. These defects suggest that tgfβ3 is required for normal chondrogenesis. To address the cellular mechanisms that lead to the observed malformations, we analyzed cranial neural crest development in morphant and tgfβ3 over-expressing fish. We observed reduced pre-migratory and migratory cranial neural crest, the precursors of the neurocranial cartilage and pharyngeal arches, in tgfβ3 knockdown embryos. In contrast, only the migratory neural crest was reduced in embryos over-expressing tgfβ3. This raised the possibility that the reduced number of cranial neural crest cells is a result of increased apoptosis. Consistent with this, markedly elevated TUNEL staining in the midbrain and hindbrain, and developing pharyngeal arch region was observed in morphants, while tgfβ3 over-expressing embryos showed marginally increased apoptosis in the developing pharyngeal arch region. We propose that both Tgfβ3 suppression and over-expression result in reduced chondrocyte and osteocyte formation, but to different degrees and through different mechanisms. In Tgfβ3 suppressed embryos, this is due to impaired formation and survival of a subpopulation of cranial neural crest cells through markedly increased apoptosis in regions containing the cranial neural crest cells, while in Tgfβ3 over-expressing embryos, the milder phenotype is also due to a slightly elevated apoptosis in these regions. Therefore, proper cranial neural crest formation and survival, and ultimately craniofacial chondrogenesis and osteogenesis, are dependent on tight regulation of Tgfβ3 protein levels in zebrafish.