The role of the endoderm in the development and evolution of the pharyngeal arches

Abstract

The oro-pharyngeal apparatus has its origin in a series of bulges found on the lateral surface of the embryonic head, the pharyngeal arches. Significantly, the development of these structures is extremely complex, involving interactions between a number of disparate embryonic cell types: ectoderm, endoderm, mesoderm and neural crest, each of which generates particular components of the arches, and whose development must be co-ordinated to generate the functional adult oro-pharyngeal apparatus. In the past most studies have emphasized the role played by the neural crest, which generates the skeletal elements of the arches, in directing pharyngeal arch development. However, it is now apparent that the pharyngeal endoderm plays an important role in directing arch development. Here we discuss the role of the pharyngeal endoderm in organizing the development of the pharyngeal arches, and the mechanisms that act to pattern the endoderm itself and those which direct its morphogenesis. Finally, we discuss the importance of modification to the pharyngeal endoderm during vertebrate evolution. In particular, we focus on the emergence of the parathyroid gland, which we have recently shown to be the result of the internalization of the gills.

Introduction

A conserved feature of all vertebrate embryos is the presence of a series of bulges on the lateral surface of the head, the pharyngeal arches; it is within these structures that the nerves, muscles and skeletal components of the pharyngeal apparatus are laid down. The pharyngeal arches are separated by endodermal outpocketings, the pharyngeal pouches, which are structures of considerable importance (Fig. 1). These endodermal outpocketings contact the ectoderm at defined points along the rostrocaudal axis, and expand along the dorsoventral axis to generate a narrow slit-like morphology. The two most anterior pouches, 1 and 2, form first, at about stage 12 in the chick, followed by the third pouch and lastly the fourth pouch (Veitch et al. 1999). The first pouch lies between the first and second arches, the second pouch between the second and third arches, the third pouch between the third and fourth arches, and the last, fourth pouch, between arches 4 and 6. A fifth arch never establishes itself in amniotes. The pharyngeal pouches thus come to separate the neural crest and mesodermal cells of the arches and to define the anterior and posterior limits of each arch.

The pharyngeal pouches display overt regionalization and are highly ordered (Veitch et al. 1999) (Fig. 2). The pouches are polarized structures. For example, whereas the rostral half of each pouch expresses Bmp-7, the caudal half expresses FGF-8 and the dorsal aspect of each pouch is marked via its expression of Pax-1. Finally, each pouch has an individual sense of identity. Shh expression is a prominent early feature of the caudal endoderm of the second arch, and individual pouches mark the anterior limits of expression of Hox genes within the pharyngeal endoderm; Hox-a2 has a rostral boundary at the second pouch, Hox-a3 at the third pouch and Hox-a4 at the most caudal pouch (Manley & Capecchi, 1995; Sakiyama et al. 2001). Later in development the pouches also form particular derivatives. In humans, the thyroid forms from the ventral aspect of the second pouch, the thymus from the third pouch, and the parathyroid from the third and fourth pouches.

A number of studies have suggested that the patterning of the arches was dependent upon the neural crest (Noden, 1983; Kontges & Lumsden, 1996); it has now become apparent, however, that the formation of the pouches and the establishment of regionalization within the pharyngeal endoderm does not require the neural crest (Veitch et al. 1999). Importantly, it was found that pharyngeal arches could form in the absence of the neural crest, and that these crestless arches were still normally patterned and had a sense of individual identity; Bmp-7 was still expressed at the rostral half of each pouch, FGF-8 was expressed at the caudal half of each pouch, Pax-1 was expressed dorsally within the pouches and shh expression was still associated with the second pouch. This view was confirmed by a subsequent study in mice analysing mutations in the homeobox genes Hoxa1 and Hoxb1, which are involved in patterning the hindbrain (Gavalas et al. 2001). In these particular double mutants rhombomere 4 has lost the ability to generate neural crest cells, and thus the second arch crest population is not generated. However, this does not affect the formation of the second arch nor its epithelial regionalisation.

A number of studies have also shown that the formation of the pouches is central to the development of the pharyngeal arches. In the zebrafish mutant, van gogh (vgo), the neural crest-derived cartilages of the arches are disorganized, often fusing with each other, or, in the more posterior arches, failing to form altogether (Piotrowski & Nusslein-Volhard, 2000). This is not, however, due to defects in the neural crest, which form and are segregated into three streams, but lies in alterations to the pharyngeal endoderm. In the vgo mutant the endoderm fails to segment and the pharyngeal pouches do not form. More recently, it has been shown that the Vgo gene encodes the zebrafish orthologue of Tbx-1, a gene which has been strongly implicated as a key factor in DiGeorge syndrome in humans (Jerome & Papaioannou, 2001; Lindsay et al. 2001; Merscher et al. 2001). Studies in mice have also shown that mutation in the Tbx-1 gene results in defects in pharyngeal pouch formation in this species. This in turn affects the development of the aortic arches, the major blood vessels of the pharyngeal arches, and derivatives of the pharyngeal pouches, the parathyroid and thymus. These defects phenocopy many aspects evident in DiGeorge syndrome patients.

These studies strongly suggested that the pharyngeal endoderm is a key player in organizing the arches, and direct evidence supporting this view has recently been demonstrated in chick embryos. When specific domains of the pharyngeal endoderm are ablated at early stages there is a corresponding failure in the development of the neural crest-derived skeletal elements (Couly et al. 2002). Thus, ablation of the most rostral pharyngeal endoderm causes the reduction or absence of the nasal cartilage, while ablation of a next caudal-most piece affects the development of the meckel's cartilage. Moreover, if rostral neural crest cells are exposed to an ectopic strip of pharyngeal endoderm, then they will generate a supernumerary jaw. Interestingly, the orientation of the ectopic strip of endoderm also determines the orientation of the skeletal elements of the additional jaw.

The neural crest cells, however, are not completely passive in this process. Rather, the response of different populations of the neural crest to endodermal cues is dependent upon the transcription factors that they express, most notably their Hox gene repertoire. During normal development, the first arch crest is Hox negative, the second arch crest expresses Hox-a2, and the third arch crest express members of the third paralogous Hox group: Hox-a3, Hox-b3, Hox-d3 (Hunt et al. 1991). In mice lacking Hox-a2 the second arch exhibits a transformation, and first arch jaw elements form within it (Gendron-Maguire et al. 1993; Rijli et al. 1993). Correspondingly, if Hox-a2 expression is forced in the neural crest cells of the first arch, then jaw development is inhibited and second arch elements form (Grammatopoulos et al. 2000; Pasqualetti et al. 2000). The fact that Hox genes allow neural crest cells to respond differentially to endoderm cues is also supported by the observation that if Hox-positive crest was exposed to an ectopic strip of rostral pharyngeal endoderm then, unlike the rostral crest cells which do not express Hox genes and form a jaw in this situation, these cells do not form a jaw (Creuzet et al. 2002).

The endoderm is also responsible for the induction of particular arch components. Past studies in amphibia have supported a role for the endoderm in directing crest cells towards a chondrogenic fate (Epperlein, 1974). More recent studies in zebrafish have further supported this view and have given indications into the molecular basis of this inductive interaction. The zebrafish mutants bonnie and clyde (bon) and casanova (cas), which do not produce endoderm, fail to form pharyngeal cartilages (David et al. 2002), and it has further been shown that the neural crest cells are induced to form cartilage via the action of FGF-3 and FGF-8 emanating from the endoderm (David et al. 2002; Walshe & Mason, 2003). The endoderm also acts to induce the epibranchial placodes (Begbie et al. 1999). These are focal thickenings of the embryonic ectoderm that form immediately dorsal and caudal of the clefts between the pharyngeal arches, and they produce the neuroblasts that migrate and condense to form the distal cranial sensory ganglia: the geniculate, petrosal and nodose. These structures are induced to form in the ectoderm via the action of Bmp-7 (Begbie et al. 1999). This molecule is expressed in the endoderm where it contacts the ectoderm, it will elicit the production of epibranchial neurons when added to cranial ectoderm in culture, and if its function is blocked pharyngeal endoderm is unable to induce the formation of the placodes.

Patterning the pharyngeal endoderm

Although it has been established that the endoderm plays an important role in patterning the pharyngeal apparatus, there have been few direct studies on the patterning of the pharyngeal endoderm itself. Recent studies, however, have suggested that retinoids play a role in the patterning of the pharyngeal endoderm. The first evidence for this came from studies of mice in which the retinoic acid synthetic enzyme, Retinaldehyde-specific dehydrogenase type2 (Raldh2), had been inactivated (Niederreither et al. 1999). In these animals, only the first pharyngeal arch is evident, and the caudal arches are absent. Another study in mice, in which embryos were treated with a pan-RAR antagonist, also provided support for a role for retinoids in the patterning of the pharyngeal endoderm (Wendling et al. 2000). It was found that inhibiting retinoid action at particular times in development specifically perturbed the development of the third and fourth arches, but that the first two arches formed. More recently, an analysis of the zebrafish mutant, neckless, has shown that the role of retinoic acid in directing the development of the pharyngeal endoderm is also evident in this species (Begemann et al. 2001). The neckless phenotype has been shown to be underlain by a point mutation in Raldh2, and in these it has been found that the development of the caudal arches is affected, but that the first two form.

To detail accurately the deficits that occur in the patterning of the endoderm in the absence of retinoids we analysed the development of this tissue in vitamin A-deficient (VAD) quail embryos (Quinlan et al. 2002). These embryos are born to mothers that were fed a diet lacking vitamin A, and as such the embryos have no retinoids (Maden et al. 1996). We found that the development of the pharyngeal endoderm is affected from the very earliest stages in these animals in that this territory extends further caudally than it normally would. We also observed that as development proceeds the first pouch invariably forms, the second struggles to form, and that there is never any indication of the formation of the third or fourth pharyngeal pouches. We did find, however, that dorsoventral patterning of the pharyngeal endoderm seems to be unaffected. The fact that while pouches 3 and 4 are absolutely dependent upon retinoic acid for their development, and that the second pouch is dependent to a lesser extent while the first pouch does not require retinoids, is explained by an analysis of the expression pattern of Raldh2 (Fig. 3). This gene is expressed in the lateral mesoderm flanking the endoderm and that from stage 12 onwards has a fixed rostral limit underlying the otic vesicle, i.e. at the level of the second pouch. Thus, while pouches 2–4 form in the vicinity of Raldh2-expressing lateral mesoderm, the first pouch, which forms anteriorly, does not. Clearly, the development of the first pouch is under a distinct developmental programme, which is as yet undefined. Interestingly, a similar situation is observed in Tbx-1 mutants, where again the caudal pouches fail to form but the first pouch develops normally (Jerome & Papaioannou, 2001; Lindsay et al. 2001; Merscher et al. 2001). Given that the second pouch is delineated in the absence of retinoid signalling but that its development is subsequently abnormal it is possible that the developmental programme of the first pouch extends to the level of the second pouch, and it is at this axial position that these two distinct programmes abut or overlap.

Morphogenesis of the pharyngeal pouches

The morphogenesis of the pharyngeal pouches must involve a remodelling of the endoderm, such that the initial outpocketing that defines the pouch elaborates along the dorsoventral axis, adopting a narrow slit-like morphology. Interestingly, studies into a variety of other morphogenetic episodes in both vertebrates and invertebrates have described the involvement of actin cables that function at the level of the tissue to bring about specific movements (Jacinto et al. 2001) and recently we have also demonstrated that actin cables play a role in directing pouch morphogenesis (Quinlan et al. 2004). There is a pronounced and localized accumulation of f-actin within each of the pouches throughout the prolonged period of their dorsoventral expansion. These actin fibres are organized into a two-dimensional ‘web’ of supracellular actin cables that run just below the apical plasma membrane of the pharyngeal endodermal cells (Fig. 4). Importantly, this web of actin is not found throughout the pharyngeal endoderm, but instead shows a marked localized accumulation, being most abundant in regions where pouches are forming and generally at much lower abundance in the interpouch endoderm. Moreover, these actin cables are connected via N-cadherin-based adherens junctions, and, again, N-cadherin expression is localized to the pouch endoderm and is not associated with interpouch regions.

If the actin cables are disrupted by treatment with cytochalasin-D, which inhibits the new assembly of actin filaments, defects in the morphogenesis of the pouches are observed (Quinlan et al. 2004). Application of cytochalasin-D at early stages of pouch morphogenesis does not result in a failure in pouch formation, but affects the elongation of the pouches, which is not directed dorsoventrally. Consequently, the pouches fail to generate the typical narrow, slit-like morphology but rather display a diamond-shaped morphology (Fig. 4). Injection of cytochalasin-D into the pharyngeal cavity of embryos at later stages also affected the morphology of the pouches, with the pouch endoderm assuming a contorted morphology, which is again due to growth not being directed appropriately. The fact that disrupting the actin cables does not affect growth as such but rather the direction of growth suggests that the actin cables are functioning as a constraining force upon the endodermal sheet directing the pouches primarily to elongate along the dorsoventral axis and thus generate their typical narrow, slit-like shape. Besides directing the dorsoventral expansion of the pouches, the actin cables could also act to give rigidity to the pouch endoderm.

Alterations to the pharyngeal endoderm during vertebrate evolution

A number of important alterations have occurred to the pharyngeal apparatus during vertebrate evolution, and some of these are likely to have their basis in modifications to the pharyngeal endoderm. One of the most obvious changes that has altered is the number of pharyngeal arches that develop. Thus, whereas lamprey have nine arches, teleosts have seven and amniotes have five. Although this suggests that the general trend is towards a reduction in arch number, it should be noted that considerably more arches have been described in ostracoderm fossils, and that within chondrichthyans, although most species have seven arches (five gills) there are also species with eight arches, the six-gilled shark, and species with nine arches, seven-gilled shark. In keeping with the current view of the prime role of endoderm in organizing arch development, it is likely that all of these alterations in arch number have their origins in alterations in the number of pharyngeal pouches.

The pouches generate a number of specialized epithelial structures: the thyroid, parathyroid, thymus and ultimobranchial body. In all vertebrates the thyroid arises from the ventral aspect of the second pouch, but such a static situation is not seen for the other pouch derivatives (Kardong, 1998). Thus, whereas the most posterior pouch generates the ultimobranchial bodies in most vertebrate classes, mammals do not form this structure. Rather, the cells fulfilling the function of the ultimobranchial body, the parafollicular cells, are incorporated into the body of the thyroid. Variations are also seen in the formation of the thymus and the parathyroid glands. In fish, thymic primordia are generated by all the pouches except the first. However, in avians the thymus arises from pouches 3 and 4, whereas in humans it is only generated by the third pouch. The parathryoid gland is not formed in fish, but is only found in tetrapods. In humans and chick it emerges from pouches 3 and 4, but in mice it is exclusively generated by the third pouch.

The emergence of the parathyroid gland is an evolutionary modification that is believed to have been of great importance to the emergence of the tetrapods. This structure plays a pivotal role in regulating extracellular calcium homeostasis, which is important to many physiological processes, such as muscle contraction, blood coagulation and synaptic activity. The parathyroid detects changes in the levels of calcium in blood via the calcium sensing receptor, CasR, which in turn modulates the secretion of parathyroid hormone (PTH), which acts to release calcium from internal stores such as bone, as well as acting to modulate renal ion transport (Brown, 2001). The evolution of the tetrapods, and the shift from an aquatic to a terrestrial environment, was believed to have required new controls for regulating calcium homeostasis, and thus the evolution of the parathyroid glands, and of PTH, was a key event in facilitating this transition. This freed the tetrapods from relying upon calcium uptake from the water, as utilized by fish, by giving them the ability to regulate their serum calcium levels internally.

Studies in mice have demonstrated that the transcription factor Gcm-2 is a key regulator of parathyroid development. The expression of this gene is restricted to the parathyroids, and if this gene is mutated the parathyroids fail to form (Kim et al. 1998; Gunther et al. 2000; Gordon et al. 2001). Recently, we have exploited the rigid association between Gcm-2 and the parathyroid and conducted a phylogenetic analysis to gain insight into how the parathyroid evolved (Okabe & Graham, 2004). Interestingly, we found that it is not only tetrapods that possess the Gcm-2 gene but that this gene is also present in dipnoi and teleost fish, as well as in chondrichthyans (Fig. 5). We also found that Gcm-2 in zebrafish and dogfish, as in amniotes, is expressed in the pharyngeal pouches, and the structures that derive from the pouches, the internal gill buds in fish and the parathyroids in tetrapods (Fig. 5). We also tested the function of Gcm-2 in zebrafish, and showed that this gene is required for the elaboration of the internal gill buds from the pharyngeal pouches. Finally, through searching the zebrafish and fugu genome sequences we have identified PTH-encoding genes, and we show that these genes are expressed in the gills, as is the calcium-sensing receptor. Thus, both the tetrapod parathyroid and the gills of fish contribute to the regulation of extracellular calcium levels. It is therefore reasonable to suggest that the parathyroid gland evolved as a result of the transformation of the gills into the parathyroid glands of tetrapods and the transition from an aquatic to a terrestrial environment. This interpretation would also explain the positioning of the parathyroid gland within the pharynx in the tetrapod body. Were the parathyroid gland to have emerged de novo with the evolution of the tetrapods it could, as an endocrine organ, have been placed anywhere in the body and still exert its effect.

Concluding remarks

The long held view that the neural crest played the key role in patterning the pharyngeal arches must now be reconsidered. Neural crest cells obviously have a major role to play in the development of the arches, as it is this cell type that generates the skeletal and connectives tissue derivatives of the arches, but the pharyngeal endoderm plays a more prominent role in directing early arch patterning. This has a number of important ramifications. To understand how the basic organization of the pharyngeal apparatus is directed, we need to study the pharyngeal endoderm itself. Furthermore, when considering the aetiological basis of birth defects with pharyngeal abnormalities, alterations to the development of the pharyngeal endoderm should be considered. Finally, alterations to the development of the pharyngeal endoderm are likely to have played important roles in the evolution of the pharyngeal apparatus.

References