Endodermal/ectodermal interfaces during pharyngeal segmentation in vertebrates

Abstract

A key event in the formation of the pharyngeal arches is the outpocketing of the endodermal pharyngeal pouches and the establishment of contact with the overlying ectoderm. However, relatively little is known about how the endoderm and ectoderm relate to each other at these points of contact and the extent to which this differs between the pouches. We have therefore detailed the interactions between the pharyngeal pouches and ectoderm in the chick embryo. Unlike the other pouches, the first pouch does not sustain direct contact with the ectoderm but separates after initial contact. Contrastingly, a perforation is formed between the second pouch and cleft that creates an external opening into the pharynx. Finally, the third and fourth pouch endoderm can be seen to bulge outwards through the ectoderm, although external openings to the pharyngeal lumen are not established. To understand whether these behaviours represent derived or ancestral features, we characterised the pharyngeal ectodermal–endodermal interfaces in the shark embryo. We found that the pouches of the posterior gill-bearing arches in this species also displayed the outward bulging of the endoderm into the ectoderm, although openings were established. We further used genetic tools to detail unambiguously the relationship between the endoderm and ectoderm in zebrafish and mouse embryos and again found that the posterior pouches break through the ectoderm. Thus different pharyngeal pouches establish different topological relationships with the overlying ectoderm and the posterior pouches initiate the developmental programme for the formation of gills, be they amniotes or anamniotes.

Introduction

A prominent feature of all vertebrate embryos is the presence of a series of bulges on the lateral surface of the head, the pharyngeal arches (Graham & Richardson, 2012). The development of these, however, is complex, as they comprise of a number of disparate embryonic populations: ectoderm, endoderm, neural crest and mesoderm (Graham & Smith, 2001). While it was historically believed that it was the neural crest cells that directed the development of the arches, it has now become apparent that the segmentation of the endoderm is central to this process (Graham & Richardson, 2012). The first indication of pharyngeal arch development is the formation of outpocketings within the pharyngeal endoderm, the pharyngeal pouches (Veitch et al., 1999; Crump et al., 2004). These entities form at distinct positions along the anteroposterior axis and they will contact the overlying ectoderm, which invaginates to meet them and form the pharyngeal clefts. The points of contact between the pouches and clefts define the posterior and anterior limits of each arch (Fig. 1). Thus the neural crest and the mesoderm migrate into pre-existing pharyngeal segments formed around the segmentation of the endoderm. Mutants that fail to form the pharyngeal pouches also subsequently fail to form pharyngeal arches (Piotrowski & Nusslein-Volhard, 2000).

Fig. 1.

Pharyngeal arch anatomy. (A) Schematic representation showing a lateral view of a chick embryo. The black dotted line marks the coronal plane through which the section in (B) has been taken. (B) This section represents a ventral view of the embryo (anterior at the top). The pharyngeal clefts and pharyngeal pouches have been labelled with orange and pink arrows, respectively, and the white holes in each arch represent the aortic arch arteries. The black dotted box around the 3rd pharyngeal pouch demarcates the region which is magnified in (C). (C) The image has been rotated so that the lateral view is at the top. 1–6, pharyngeal arches; ov, otic vesicle; pc, pharyngeal cleft; pp, pharyngeal pouch; red, ectoderm; green, endoderm; blue, mesoderm

Besides acting to segment the pharynx, the endodermal pouches and ectodermal clefts also give rise to distinct and important derivatives (Larsen, 1997). In amniotes, the first pharyngeal pouch/cleft contributes to the internal auditory canal and its covering, the tympanic membrane. The second pharyngeal pouch gives rise to the palatine tonsil in humans, although in chick the second pouch does not give rise to any lymphoid tissue or indeed any distinct tissue at all, but instead contributes to the mesenchyme associated with third arch structures (Hamilton & Hinsch, 1957). The third pouch will develop into the thymus and inferior parathyroid glands, and the fourth pouch gives rise to the superior parathyroid glands. Finally, an outpocketing of the endoderm in the posterior pharynx gives rise to the ultimobranchial bodies. These epithelial glands remain separate in the chick, but in humans they migrate toward and become associated with the thyroid gland before differentiating into the parafollicular or C-cells that produce calcitonin (Fagman et al., 2006; Fagman & Nilsson, 2010).

A pervasive influence on the organisation of the pharyngeal region is its evolutionary history and thus it is also important to understand how the development of the pharynx has been modified during evolution. The pharyngeal arches are a defining feature of the phylotypic stage of vertebrates and lend that stage its name, pharyngula. However, as development progresses beyond the phylotypic stage, differences amongst the different vertebrate groups become apparent (Graham & Richardson, 2012). Within most chondrichthyans, the underlying pattern of pharyngeal segmentation is preserved in the form of the array of gill slits. In osteichthyans, gills still form from the pharyngeal pouches but these are covered by the posterior expansion of the second arch and the formation of the operculum. This is a large flap which acts both to protect the gills and to help draw water into the pharynx. Contrastingly, in tetrapods, pharyngeal segmentation undergoes extensive remodelling and these modifications promote adaptation to a terrestrial lifestyle. The gills are lost as primary respiratory function is shifted to the lungs. This transition also required novel ways of regulating calcium homeostasis and thus the parathyroid gland evolved. In amniotes, the second arch expands to cover the posterior arches during embryogenesis but now the caudal edge of the second arch fuses to the subjacent tissue, which results in the posterior arches becoming enclosed in a cavity, the cervical sinus of His, which is later eradicated (Richardson et al., 2012).

The contacting of the ectoderm by the pharyngeal pouches is central to pharyngeal segmentation and the relationship between these two tissues in the developing pharynx has long been a point for discussion (Kastschenko, 1887; Mall, 1887; Mangold et al., 1981; Waterman, 1985; Goette, 1990). However, as no cell lineage analysis was performed in these studies, it remains unclear how the ectoderm and endoderm actually relate to each other and how that relationship develops. Furthermore, while it has been noted that perforations are apparent externally (Kastschenko, 1887), the route through which these are generated has not been well defined. One study in chick, using electron and light microscopy, did examine the points of contact between the endoderm and ectoderm but primarily focused on the second pouch and corresponding cleft (Waterman, 1985). It was noted that the endoderm and ectoderm contact each other via small focal junctions where the basement membrane becomes discontinuous, and that these junctions subsequently enlarge before the endoderm and ectoderm interdigitate and a small perforation appears. It was, however, unclear whether these observations of the second pouch can be extended to the other pouches and this is an important issue, as they will subsequently generate different derivatives.

Therefore, one aim of this study was to document carefully the interface between the endoderm and the ectoderm at their points of contact and to scrutinise how these mature, and for this we used the chick as a model amniote. In doing this, we also compared and contrasted the relationship between the endoderm and ectoderm at each of the pouches/clefts to ascertain which aspects were shared by all the pouches and which were specific to individual pouches. An important point that emerged from this was that, beyond the initial association between the pharyngeal endoderm and the overlying ectoderm, the pouches displayed distinct morphogenetic programmes with the relative topology of the endoderm and ectoderm differing between the pouches as development progressed. A notable feature of the posterior pouches was that the endoderm pushed into the ectoderm towards the external environment.

A second aim of this study was to place these developmental events in a phylogenetic context. To do this we additionally analysed the relationships between the ectoderm and endoderm in shark embryos. We found that the developmental programme associated with the posterior pouches in chick: the insertion of the endoderm into the ectoderm is a general feature of all the gill-bearing arches in the shark. To further assess whether this behaviour is shared across the vertebrates, we used genetic lineage-tracing strategies in zebrafish and mouse. We found in zebrafish that the development of the posterior gill-bearing arches was also associated with the endoderm breaking through the ectoderm and that similarly in mice, the posterior pharyngeal pouches break through the ectoderm. These results demonstrate that in amniotes the posterior pharyngeal pouches still execute a developmental programme that is involved in gill formation in non-tetrapod gnathostomes.

Material and methods

Embryo collection

All embryo collection was carried out as prescribed by the UK Animals (Scientific Procedures) Act 1986. Fertile hen's eggs were incubated at 38 °C to the required stages (HH st) (Hamburger & Hamilton, 1992) and the embryos were fixed in 4% paraformaldehyde (PFA). Scyliorhinus canicula embryos were taken from the egg cases, anaesthetised (MS222), staged (Ballard et al., 1993) and fixed in 4% PFA. Transgenic sox17:GFP zebrafish embryos were obtained by natural spawning and grown at 28.5 °C in egg water (0.3 g L⁻¹ Instant Ocean Salt, 1 mg L⁻¹ Methylene Blue). Embryos were staged as described (Kimmel et al., 1995) and fixed in 4% PFA. Transgenic Sox17-2A-iCre;R26R mice were supplied by Dr Albert Basson and Dr Abigail Tucker, King's College London. The day the vaginal plug was found was demarcated as E0.5. Embryos were staged according to the day collected and immediately used for x-gal staining.

Immunohistochemistry

Previously fixed embryos were washed three times 30 min in phosphate-buffered saline (PBS)/1% TritonX-100 (PBSTx) before being washed in a blocking solution of 10% goat serum in PBSTx twice for 1 h at room temperature. The relevant primary antibody was diluted in the blocking solution with 0.02% sodium azide and the embryos incubated at 4 °C for 1–2 weeks. Embryos were then rinsed in blocking solution and washed three times for 1 h in blocking solution before adding the secondary antibody diluted in blocking solution with 0.02% sodium azide. This was incubated at 4 °C for 1–2 weeks. The primary antibodies used were rabbit anti-laminin 1 : 100 (Sigma), mouse anti-β catenin 1 : 200 (Sigma) and mouse anti-GFP 1 : 500 (green fluorescent protein; Roche). Secondary antibody was Alexa 488-conjugated goat anti-mouse IgG, and Alexa 568 goat anti-rabbit IgG, both used at 1 : 1000 (Molecular Probes). For sectioning, embryos were washed into PBS, embedded in gelatin, fixed, and vibratomed to 50-μm slices.

Lysotracker staining

The vital lysosomal dye, LysoTracker red (molecular probes L-7528 RED), was diluted to 1 : 100 in PBS, and pre-warmed at 37 °C. This was then added to an equal volume of fresh PBS on each embryo and incubated at 37 °C in the dark for 30 min. Embryos were then rinsed 4× in PBS and fixed in 4% PFA overnight at 4 °C. Embryos were rinsed in PBS, then dehydrated into 100% methanol, and viewed under fluorescence.

CCFSE staining

CCFSE [5-(and-6)-carboxy-2′,7′-dichlorofluorescein diacetate, succinimidyl ester; Invitrogen] was dissolved in anhydrous DMSO to 50 mm (stored at −20 ^°C). A working concentration of 250 μm in PBS was used for all chicken embryo treatments; approximately 20 μL was applied in ovo.

Results

Morphogenesis of the pharyngeal pouches in the chick, Gallus gallus

Pouch 1

Figure 2 shows a series of developmental stages from when the first and second pharyngeal pouches arise at stage 13 until all four pouches have formed by stage 19. At stage 13 (n = 6) the first pouch is evident as a small out-pocketing with close association to the overlying ectoderm (Fig. 2A-C). It is flattened in shape and shallow, and displays slightly disorganised apicobasal polarity with rather diffuse apical β-catenin and basal laminin labelling. By stage 17 (n = 6) the first pouch has established a coherent basement membrane between the ectoderm and endoderm, separating the two tissue layers (Fig. 2D-F,J). However, although the pouch itself remains flattened with a slight evaginated morphology, the ectodermal cleft presents as a deep invagination meeting that pouch. By stage 19 (n = 6) the ectoderm and endoderm have separated from one another (Fig. 2G-I). However, the basement membrane is disorganised at the pharyngeal cleft, suggesting a movement of the ectoderm away from the endoderm.

Fig. 2.

Morphogenesis at the ectoderm/endoderm interface of the first pharyngeal pouch. Confocal sections of chick embryos following immunofluorescence. White arrows point to the first pharyngeal pouch in each image. (A-C) During initial pouch formation, increased levels of β-catenin and laminin at the first pharyngeal pouch interface suggest a rearrangement of these proteins as the epithelia grow and remodel. (D-F) As the pouch develops, the epithelia remain in apposition to one another, separated by a distinct basement membrane. The white box in (F) is magnified in (J). (G-I) At a later stage of development the epithelia appear to have moved away from one another, with intense laminin labelling indicating a rearrangement of the basement membrane. (J) Magnified view of the white box in (F) clearly shows the basement membrane separating cells of the epithelia as visualised with nuclear DAPI (4',6' diamino-2-phenylindole·2HCl) staining.

Pouch 2

Although the first and second pouches simultaneously bud off from the pharyngeal endoderm at stage 13, they come to display quite different features. At stage 14 the epithelia are in apposition to each other but are separated by an intact basement membrane (Fig. 3A-C,J). In contrast to the first pouch, breakdown of the membrane is visible at the second pouch at a relatively earlier time point, by stage 15/16 (n = 3 stage 16) (Fig. 3D-F,K). As development progresses, a single layer connecting the second and third pharyngeal arches can be seen (Fig. 3G-I,L). It is also apparent that there has been a substantial thinning of this interface, particularly at the anterior-most portion directly adjacent to the second pharyngeal arch, where the beginning of a small perforation has appeared (Fig. 3G-I,L). This perforation expands by stage 19 until the entire interface region has broken through (data not shown).

Fig. 3.

Morphogenesis at the ectoderm/endoderm interface of the second pharyngeal pouch. Confocal sections of chick embryos following immunofluorescence. White arrows point to the second pharyngeal pouch in each image. (A-C) At stage 14 the basement membranes of each epithelial layer have made contact and are intact. The region within the white box in (C) is magnified in (J) showing the intact basement membrane. (D-F) By stage 15/16 the basement membrane separating the endoderm and ectoderm is in the process of breaking down, as evidenced by its spotty appearance. The boxed region in (F) is magnified in (K). (G-H) At stage 18 the basement membrane has completely degraded and is no longer visible. This is accompanied by a thinning of the epithelia as they have fused together. The boxed region in (I) is magnified in (L).

Pouch 3

The third pharyngeal pouch appears posterior to, and at some time after (Stage 14), the second pharyngeal pouch (Fig. 4A). It is first evident as an out-pocketing of the pharyngeal endoderm with intact apicobasal polarity, indicated by clear expression of β-catenin at the apical surface and laminin expression along the basement membrane (Fig. 4B,C,J). At this stage the endoderm is evaginating in a postero-lateral direction, while the ectoderm overlying the third pouch is flat and has its own complete basement membrane extending parallel to the pouch (Fig. 4C,J). Subsequently, the ectoderm and endoderm contact each other and as the pouch continues its development, it grows toward the external surface of the embryo and causes the ectoderm to bulge outwards (Fig. 4D-F,K). The basement membrane has begun to break down at stage 17, with only patches of laminin seen across the interface (Fig. 4F,K). By stage 19 the basement membrane has disappeared at the interface, indicating that this region is composed of a single layer. This layer has also bulged so much that it extends further laterally than the adjacent pharyngeal arches do and is exposed to the external environment (Fig. 4G-I).

Fig. 4.

Morphogenesis at the ectoderm/endoderm interface of the third pharyngeal pouch. Confocal sections of chick embryos following immunofluorescence. White arrows point to the third pharyngeal pouch in each image. (A-C) The third pouch is first evident at stage 14 with distinct and separate basement membranes along the basal surface of both the ectoderm and endoderm tissue layers. The boxed region is (C) is magnified in (J) to show clearly the intact basement membrane. (D-F) At stage 17 the basement membrane begins to break down as evidenced by the spotted labelling of the laminin protein. The endoderm can also be seen pushing against the ectoderm, causing it to bulge outward. The boxed region in (F) is magnified in (K) to show more clearly the degrading basement membrane. (G-I) By stage 19 the basement membrane has broken down and the epithelial cells directly interact, resulting in a thinning of the ectoderm/endoderm interface, which now presents as a single layer with no basement membrane and showing a morphology of bulging out toward the external surface.

Pouch 4

The fourth pharyngeal pouch buds off from the anterior portion of the developing fourth arch endoderm at around stage 15 (Fig. 5A-C). The epithelium has apicobasal polarity and a clear basement membrane can be visualised (Fig. 5B,C). This pouch continues to deepen in a posterior direction, but by stage 17, changes orientation to move toward the external surface (Fig. 5D-F). The basement membrane is still intact. However, by stage 19 the endoderm has moved in an outward direction to meet the overlying ectoderm and a single basement membrane is visible at the interface between the two epithelia (Fig. 5G-I). As with the third pouch, the endoderm can be seen to be have made contact with the ectoderm.

Fig. 5.

Morphogenesis at the ectoderm/endoderm interface of the fourth pharyngeal pouch. Confocal sections of chick embryos following immunofluorescence. White arrows point to the fourth pharyngeal pouch in each image. (A-C) The fourth pouch buds off the anterior endoderm of the developing fourth arch around stage 15 and expand posteriorly. (D-F) By stage 17 the pouch has deepened and it continues expanding, following a change in direction, out towards the overlying ectoderm. The basement membranes of the ectoderm and endoderm are distinct and intact. (G-I) At stage 19 the endoderm has made contact with the ectoderm and their basement membranes appear to have fused.

Cell lineage tracing reveals endodermal/ectodermal interfaces at the pouches

The immunofluorescence analysis showed that as the endoderm and ectoderm meet at the pouches/clefts, their basement membranes become closely associated and it is difficult to distinguish these epithelia from each other. We therefore used cell tracker labelling to ascertain how the ectoderm and endoderm relate to each other as they come into contact and how this interface matures. To do this, CCFSE, a lipid-soluble dye, was applied to the external surface of the embryo to label the ectoderm at stages prior to the pouches establishing contact. CCFSE passively diffuses into cells it comes into contact with, but it is not fluorescent until its acetate groups are cleaved by intracellular esterases (Griffith & Hay, 1992). Once chemically altered by the cell it cannot diffuse any further, thus making it useful as an in vivo cell tracker.

Figure 6 shows the results of such labelling. In all cases, robust labelling of the ectoderm was achieved and this can be seen both as labelling of the exterior of the embryo and through the labelling of neuroblasts migrating internally from the epibranchial placodes (Fig. 6J,L, pink asterisks). The first pharyngeal pouch has a relatively flat morphology, with the overlying labelled ectodermal cleft invaginating to meet it (Fig. 6D). As development progresses, these two epithelial sheets move away from each other and this cell lineage tracing confirms that no swapping or sharing of cells has occurred and that both sheets remain intact at all times (Fig. 6D-F). Lineage tracing of the ectoderm at the level of the second pouch also reveals no intercalation of the ectoderm and endoderm with each other. At stage 17/18 (n = 4) the interface has begun to thin considerably but the epithelial sheets remain distinct and there is no evidence of interdigitation (Fig. 6G,P). By stage 19 (n = 6), this thin layer perforates at its most anterior point adjacent to the expanding second pharyngeal arch (Fig. 6H,I,Q). Labelling at the third pharyngeal pouch highlights a different behaviour, in that the endoderm pushes against the overlying ectoderm as it expands toward the external surface, causing the ectoderm to bulge outwards (Fig. 6J-L,R). Thus, by stage 19, the endoderm appears to be displacing the ectoderm by growing into it, yet still never mixing with the ectodermal cells, until it pushes through the ectoderm so it is bulging out of the embryo (Fig. 6K,L,R). The fourth pharyngeal pouch interface shows a similar behaviour to that of the third. The pouch is still expanding toward the ectoderm at stage 17/18, but by stage 19 the epithelia have almost made contact (Fig. 6M,N). At stage 20 (n = 8) it is apparent that the endoderm of the pouch has pushed through the ectoderm in a similar manner to that seen in the third pouch, therefore making contact with the external environment (Fig. 6O,R).

Fig. 6.

Ectodermal cell lineage tracing using CCFSE. Confocal sections of chick embryos following CCFSE treatment. White dotted lines represent the basal surface of the epithelia or the endoderm/ectoderm interface. (A-C) Overview of the morphology of all pouches present in the embryo at the indicated stage of development. White arrows point to the pouches and each is labelled individually. (D-F) Magnified images of the first pouch. Cell tracing reveals that the first pharyngeal pouch and cleft do not exchange or share any cells. (G–I) Magnified images of the second pouch. (G) Stage 17/18 ectodermal labelling reveals that the epithelial cells remain distinct once contact has been made. The boxed region in (G) is magnified in (P) and more clearly shows that ectodermal cells remain on the external surface. (H) At stage 19, the endoderm at the anterior and posterior portions of the pouch are in contact with the ectoderm and continue to grow outwards, while the interface itself perforates. (I) By stage 20 the endoderm has grown out into the ectoderm so it is now in contact with the external environment. The boxed region in (I) is magnified in (Q). (J-L) Magnified images of the third pouch. (J) At stage 17/18 the third pouch/cleft interface has made contact but each epithelial sheet retains its own territory. (K) By stage 19 the endoderm continues expanding out into the overlying ectoderm and this ectoderm begins to thin as the endoderm starts displacing it. (L) At stage 20 the endoderm has pushed through the ectoderm and is now in contact with the external environment. The boxed region is magnified in (R). (M-O) Magnified images of the fourth pouch. (M) The fourth pouch is expanding toward the ectoderm at stage 17/18. (N) By stage 19 the pouch and cleft have made contact but remain as distinct epithelia and do not intercalate. (O) At stage 20 the pouch has pushed through the overlying ectoderm and also makes contact with the external environment, as seen with the third pouch.

Foci of cell death precede endodermal displacement of the ectoderm

The cell-tracing studies showed that following breakdown of the basement membrane at pouches 2, 3 and 4, the ectoderm and endoderm do not intercalate at their interface but the ectoderm is displaced by the expansion of the endoderm. As cell death is evident during many epithelial remodelling events, including neural tube closure and eye formation, we tested whether it is associated with the morphogenesis of the pharyngeal pouches; to do this we used Lysotracker Red. Staining at various developmental stages revealed highly localised foci of cell death, both spatially and temporally, associated with the displacement of the ectoderm. At stage 15 (n = 3) the epithelia at the second pouch/cleft interface are in apposition with each other, and there is intense staining in the ectoderm (Fig. 7A,D,G). This cell death lasts only for a short time, and by stage 17 it is no longer evident at the second pouch/cleft interface (Fig. 7B,E). The short period correlates with the time following the breakdown of the basement membrane, and the direct contact between the epithelia. At stage 19 (n = 3) cell death is seen again specifically in the ectoderm but now at the third pouch/cleft interface (Fig. 7C,F,H). Thus, these bursts of apoptosis may allow the ectoderm to make way for the endoderm to push through at these specific regions.

Fig. 7.

Lysotracker Red staining reveals cell death in the ectoderm at the pouch interface. Confocal sections of chick embryos following Lysotracker Red staining. (A) A stage 15 embryo reveals Lysotracker Red staining specifically in the second pharyngeal pouch (pp2). The boxed region is enlarged in (D). (D) Cell death is evident mostly in the ectoderm at the interface, which is marked by a white dotted line. This region is further magnified in (G) to show dying cells among non-dying cells whose nuclei are stained with DAPI. (B) The arrows point to the first (pp1), second (pp2) and third (pp3) pharyngeal pouches that are enlarged in (E). (E) No apoptosis is evident and cell death appears to have halted at this time. (C) An overview of pouch morphology at stage 19 reveals specific Lysotracker Red staining over the third pharyngeal pouches. The boxed region is enlarged in (F). (F) The dotted line demarcates the interface, showing that the majority of apoptotic cells are also located in the ectoderm. This region is further magnified in (H), where dying cells can be seen located nearest the external edge of the pouch interface and at the anterior-most region closest to the second arch.

Pharyngeal pouch morphogenesis in the shark, Scyliorhinus canicula

The morphogenesis of the pharynx is profoundly shaped by its evolutionary history and it is therefore important to understand the extent to which the modes of development of the pharyngeal pouches uncovered in the chick represent a derived or ancestral state. To address this issue, we have analysed the relationships between the pharyngeal pouches and clefts, and how these alter during development in a chondrichthyan species, S. canicula. We have used anti-laminin immunofluorescence to investigate the morphology of the pouch/cleft interfaces over three different stages (n = 3 embryos at each stage). At stage 19, where only three (of a total of six) pouches have developed, it can be seen that the first pharyngeal pouch is already closely associated with the ectoderm, whereas the second pouch has evaginated outward and has made contact with the overlying ectoderm (Fig. 8A). A distinct basement membrane is visible on the lateral sides of the second pouch, although it appears to have degraded over the middle portion (Fig. 8B). The third pharyngeal pouch endoderm has also just made contact with the overlying ectoderm at a small focal point (Fig. 8C, white arrowhead), while lateral of this point the endoderm and ectoderm remain separate. Five pharyngeal pouches are present by stage 21 (Fig. 8D). The first pharyngeal pouch and cleft are still separated by a basement membrane, but this is somewhat disorganised, and the first pouch has not broken through (Fig. 8D). However, the second pouch has broken through and established contact with the external environment (Fig. 8D). At this stage the third pharyngeal pouch/cleft interface has also thinned greatly and no basement membrane is evident, while the fourth pouch has a basement membrane which, as evidenced by its spotty appearance, has started to break down (Fig. 8E,K). It is also apparent that the fourth cleft shows a distinct thinning and the endodermal pouch is pushing through it towards the external surface (Fig. 8E,K). The most posterior fifth pouch/cleft interface present at this stage reveals an intact basement membrane separating the endoderm and ectoderm, with both epithelial layers retaining the same thickness (Fig. 8F). By stage 22, the second, third and fourth pharyngeal pouches have all broken through (Fig. 8G,H). However, the fifth pouch has now budded off the posterior pharynx and has a distinct basement membrane separating the epithelia, although it does appear to be starting to degrade at its anterior aspect (Fig. 8G,I,L). Thus a general morphogenetic program for the interaction of the endoderm and ectoderm appears to apply to all the gill-forming posterior pouches in the shark. The epithelia make contact and their basement membranes fuse and break down, allowing direct contact between these epithelia. The ectodermal layer then thins as the endodermal pouch pushes through and out into the external environment when it perforates, allowing a connection between the pharyngeal lumen and the external environment.

Fig. 8.

Location of the pharyngeal pouch/cleft interface in shark embryos. Confocal sections of shark embryos following anti-laminin immunofluorescence (anterior to the left). (A) A view of all pharyngeal pouches present at stage 19, with white arrows pointing toward each pouch and labelled appropriately. (B) A magnified view of the second pharyngeal pouch, showing a continuous basement membrane at the lateral edges of the pouch with less prominent labelling across the middle of the interface, indicating a discontinuation (arrowheads). (J) Magnification of (B) reveals clear laminin labelling at the edges of the interface but none across the middle, showing this part of the basement membrane has broken down. (C) A magnified view of the third and most posterior pharyngeal pouch at this stage reveals an intact basement membrane separating the endodermal pouch from the overlying ectoderm. The point of contact between the ectoderm and endoderm is labelled with a white arrowhead. (D) An overview of all five pouches seen at stage 21. The first pouch/cleft interface is intact with strong laminin labelling, whereas the second pouch has broken through and the third pouch is about to. (E) The interface at the fourth pharyngeal pouch is beginning to thin, particularly in the ectodermal layer, and a discontinuous basement membrane is present (white arrowheads). This image is magnified in (K). (F) An intact and distinct basement membrane can be seen separating the two epithelia (white arrowheads), with both epithelial layers each retaining a two- to three-cell-deep thickness. (G) This image gives a general overview of all pouches present at stage 22. (H) The third pouch has broken through, allowing communication between the pharyngeal lumen and the external environment. (I) The basement membrane of the fifth and posterior pouch is beginning to break down (arrowhead) and will eventually perforate as more posterior pouches form with continued development. The image in (I) is magnified in (L).

Lineage tracing demonstrates that the displacement of the ectoderm by the endoderm is a conserved feature of the development of the posterior pharyngeal pouches

Our analysis of pharyngeal pouch morphogenesis suggested that in both chick and dogfish the endoderm of the posterior pouches inserts into the ectoderm and thus is exposed to the external environment. To determine whether this is found more generally within other vertebrate classes and to determine reliably whether the endoderm does push out into the external environment, we have used genetic tools that specifically label the endoderm in zebrafish and mouse to detail the relationship between the ectoderm and endoderm in these species.

Lineage tracing of the pharyngeal endoderm in zebrafish, Dano rerio

To define the position of the endoderm relative to the ectoderm in zebrafish, we used a transgenic zebrafish line, Tg (sox17:GFP), in which the Sox17 regulatory region directs expression of a GFP transgene, allowing visualisation of the endoderm (Chung & Stainier, 2008). At 48 hpf (n = 4) the second arch has begun growing posteriorly, covering the third and fourth arches (Fig. 9A, pink asterisks). Strong sox17 expression can be seen in the endoderm and it is clear that this layer reaches the external environment (Fig. 9A,C). This is more pronounced at later stages, and at 72 hpf (n = 6) strong Sox17 expression is still evident along the external surface of the pharyngeal arches (Fig. 9B,D, white asterisks). In addition, the inner surface of the operculum strongly expresses Sox17, indicating that this part of the flap is of endodermal origin (Fig. 9B,D). A magnified view of the 72 hpf pharyngeal pouches reveals that the anterior half of the pouch contributes to the posterior portion of the anterior adjacent pharyngeal arch, and thus that the endoderm becomes part of the external surface of the arches (Fig. 9D).

Fig. 9.

The pharyngeal pouch/cleft interface of transgenic Sox17 zebrafish and mice. (A-D) Confocal images of sox17:gfp zebrafish coronal sections (anterior to the top). (A) The endoderm of all the pharyngeal pouches has pushed out toward the external surface of the embryo. The posterior pharynx is blocked from view because of the still enlarged yolk (orange asterisk). The operculum (pink asterisk) covers the first two posterior arches and its inner lining expresses Sox17. The boxed region is magnified in (C) showing that the pouch (pp5; white arrow) has grown outward past the adjacent pharyngeal arches. (B) Confocal section of a 72 hpf zebrafish reveals the pharyngeal arches (white asterisks) with endoderm contributing to their external lining. The boxed region is magnified in (D) showing the anterior half of the pouch extends over the posterior portion of the anterior adjacent arch. (E-J) Sox17iCre;R26R mice (anterior to the left). (E) A coronal section through the pharyngeal arches at E9.5 reveals the morphology of the pharyngeal pouches. The first pouch (pp1) is separated from the overlying ectoderm by mesenchyme. The second pouch (pp2) interface has thinned so only a narrow band of endoderm connects the second and third arches. The third pouch (pp3) is in contact with the overlying ectoderm and appears to bulge through it, with the ectoderm much thinner than the endoderm. The black box is magnified in (G) and the red box in (I). (F) A coronal section through the arches at E10.5 reveals the morphology of the pouches at this later stage. The first pouch (pp1) is still separate from the cleft. The second pouch (pp2) has now broken through, with endoderm from the anterior border of the pouch beginning to extend posteriorly over the anterior surface of the third arch as the second arch expands caudally. The region inside the black box is magnified in (H). The third pouch (pp3) endoderm has broken through and is now in contact with the external environment. The region in the red box is magnified in (J).

Lineage tracing of the pharyngeal endoderm in mouse, Mus musculus

Sox17-2A-iCre mice (Engert et al., 2009) were crossed with R26R (Rosa 26 reporter) mice to track the development of the pharyngeal pouches and to define their relationship with the ectoderm. A coronal section through the arches reveals the morphology of each pouch and the nature of its interaction with the overlying ectoderm. At E9.5 (n = 3), the ectoderm and endoderm are separate from one another at the first pouch/cleft interface, as was described in the chick in the previous section (Fig. 9E, pp1). The second pouch at this stage has evaginated toward the external surface, leaving only a thin layer of endoderm connecting the second and third pharyngeal arches (Fig. 9G), while the third pouch is pushing into the ectoderm (Fig. 9I). By E10.5 (n = 3), it can be seen that the first pharyngeal pouch remains separate from the overlying ectoderm (Fig. 9F, pp1), while the second pharyngeal pouch has broken through, allowing the endoderm to make contact with the external environment (Fig. 9H). It is also now apparent that the third pharyngeal pouch has become more triangular in shape, and has pushed through the ectoderm to make contact with the external environment (Fig. 9J).

Discussion

Pharyngeal segmentation is established and defined by contacts between the pharyngeal pouches and the overlying ectoderm. Yet, we have little understanding of how these tissues relate to each other at these interfaces, how they alter during early development, and the extent to which this differs between each pouch. We have therefore conducted a detailed morphological analysis of the endodermal/ectodermal interfaces during pharyngeal development in the chick embryo. As expected we find that there is a general rostral to caudal maturation, with pouches 1 and 2 contacting the ectoderm at relatively early stages, followed by pouch 3 and then pouch 4. We also noted similarities between the different pouches and their interactions with the corresponding overlying ectoderm. We found that for most of the pouches, the establishment of the interface between the endoderm and ectoderm is followed by the breakdown of the basement membrane between these epithelia. However, we found no evidence of intercalation or interdigitation, the ectoderm and endoderm remaining separate at all times. Beyond these commonalities, our studies also revealed significant differences in the behaviour of each of the pouches. This is an important point, as previous studies have tended to assume that lessons learned from one pouch can be extended to the others and that data from all the pouches can be pooled, as they will all develop via similar routes (Waterman, 1985; Miller et al., 1993). We found that in contrast to the others, the interaction between the first pouch and cleft was relatively transient, with these structures separating soon after their initial contact and the space being filled by mesenchyme. The second pouch was also distinct in that the endoderm and ectoderm come together to form a layer that thins significantly, followed by the formation of a perforation which generates an external opening that persists as the second pharyngeal arch subsequently expands posteriorly. The behaviour of the most posterior pouches, 3 and 4, was markedly different from that of both pouches 1 and 2, but was relatively similar to each other. The development of pouches 3 and 4 is distinguished by the growth of the endoderm such that it bulges outwards, inserts into, and displaces, the ectoderm, as revealed by cell tracker labelling of the ectoderm. However, it should be noted that perforations of the endoderm do not form at these pouches and indeed the external endodermal protrusions that are established will come to be covered and internalised by the expanding second arch. We further found that the outwards expansion of the endoderm at pouches 2, 3 and 4 was preceded by foci of apoptosis appearing in the overlying ectoderm, suggesting that the induction of cell death might facilitate the perforation formed at the second pouch and the insertion of the endoderm of the third and fourth pouches through the ectoderm and out into the surface environment. These ectodermal foci of cell death may help explain the documented formation of openings, transient or otherwise, at the 1st, 2nd and 3rd pharyngeal clefts (Mangold et al., 1981; Kastschenko, 1887) even if, apart from the second pouch, the underlying pouch endoderm does not perforate.

Our observations further highlight the significant differences that exist between the development of the anterior and posterior pharyngeal arches. The two most anterior pharyngeal pouches form at the same time, whereas the more posterior pouches form sequentially. Furthermore, it is apparent that the development of the posterior pouches is also under the control of distinct signalling pathways. It has been shown that retinoid and wnt signalling is required for the formation of the more posterior but not the anterior pouches (Quinlan et al., 2002; Choe et al., 2013). One key gene for the formation of the posterior but not anterior pharyngeal pouches is Tbx1 (Piotrowski et al., 2003; Xu et al., 2005). In mutants that lack Tbx1 function, the posterior pharyngeal pouches fail to form and the neural crest fails to become segregated. A detailed study in mice has shown that this gene is required to drive proliferation within the endoderm (Xu et al., 2005) and this would suggest that the control of proliferation might be central to the pushing out of the endoderm into the ectoderm. This viewpoint would also be generally consistent with the study of Miller et al. (1993), which suggested that differential proliferation plays a role in pharyngeal arch morphogenesis.

Pharyngeal segmentation is a key vertebrate characteristic but it is one that has also undergone substantial evolutionary modification, in particular with the emergence of the tetrapods (Graham & Richardson, 2012). We therefore sought to compare the lessons learned from the chick with an analysis of the development of the endodermal/ectodermal interfaces in the shark embryo. The shark is particularly useful, as it represents a model that can give us insights into how these events proceed in chondrichthyans, the sister group to the osteichthyans, and thus help inform us of the basal gnathostome condition. As with chick, we find that there is a general rostral to caudal establishment of the pharyngeal endodermal/ectodermal interfaces and again a similar initial sequence of events occurs at these points, with the basement membrane breaking down after the establishment of contact between the epithelia. Yet, in the shark the pouch endoderm continues to impinge upon the ectoderm until it perforates and openings come to be associated with all of the pouches (Fig. 7).

The comparisons between pharyngeal morphogenesis in chick and shark suggest that the growth of the endoderm and its displacement of the ectoderm might be an ancestral feature of gnathostomes. If this were the case, rather than being an example of convergence, one would expect it to be observable in representatives of other vertebrate clades. We therefore used genetic techniques specifically to follow the fate of pharyngeal pouches in a teleost, the zebrafish, and a mammal, the mouse. Significantly, we found that in both these species the pouch endoderm could be seen to evaginate and to break through the ectoderm but, beyond this point, differences were apparent. In the zebrafish, as in the shark, perforations formed and a clear opening was established, whereas in the mouse, perforations did not form. Thus, collectively our data demonstrate that in all gnathostomes the posterior pharyngeal pouches execute a similar developmental programme which results in the endoderm becoming exposed to the external environment as a result of this tissue displacing the overlying ectoderm. In animals that form gills, exemplified here by dogfish and zebrafish, this process continues such that openings are made, which is necessary to allow oxygenated water taken in through the mouth to move out through the pharynx, passing over the gills. However, in amniotes, which do not respire via gills, this developmental process does not progress as far and permanent openings are not established.

This work supports our previous analyses that demonstrated that the gills were not lost with the evolution of the tetrapods but were transformed into the parathyroid glands (Okabe & Graham, 2004). In anamniotes, gills derive from the posterior pharyngeal pouches, whereas in amniotes the parathyroid is generated from these structures, and this has long been viewed as representing a major distinction between the amniotes and other vertebrates. However, we have previously shown that both the gills and the parathyroid glands express the gene gcm2 and require this gene for their formation. The results from the present study further reinforces this view in that it is now apparent that the morphogenesis of the posterior pouches in gill-bearing and non-gill-bearing animals is the same; in both, the endoderm pushes out and through the ectoderm. Thus the fact that the anterior and posterior pharyngeal arches are underpinned by different developmental programmes indicates that the ancestral functional division between an anterior region of the pharynx associated with feeding and a posterior gill-forming region still exists in amniotes.

Author contributions

This study was devised by A.G. and all experimental work was carried out by V.S. The manuscript was written by A.G. and V.S.