Cell Tissue Res (2004) 315:85–95 DOI 10.1007/s00441-003-0818-x

REGULAR ARTICLE

Ann Huysseune · Jean-Yves Sire

The role of epithelial remodelling in tooth eruption in larval zebrafish

Received: 28 February 2003 / Accepted: 15 September 2003 / Published online: 29 October 2003
Springer-Verlag 2003

Abstract Based on light and transmission electronmicroscopic observations on erupting first-generation teeth in the zebrafish, Danio rerio, we propose a biphasic mechanism for tooth eruption: (1) formation of an epithelial crypt prior to eruption of the tooth, possibly as a result of constraints in the epithelium resulting from the growth of adjacent tooth germs, and (2) detachment of cellular interdigitations both within the pharyngeal epithelium, at the pharyngeal epithelium/enamel organ boundary, and between the outer and inner dental epithelium, resulting in the exposure of the tooth tip in the crypt, immediately after tooth ankylosis. Later, further detachment of interdigitations between the inner and the outer enamel epithelium unfolds the epithelium even more and leads to a more pronounced exposure of the tooth tip. The presence of small patches of non-collagenous matrix on the outer surface of the tooth close to where it merges with the attachment bone is interpreted as a device to prevent complete detachment of the enamel organ. The biphasic nature of the mechanism for tooth eruption is supported by observations on in vitro cultured heads. First-generation teeth develop normally and crypts are formed, as under in vivo conditions, but the teeth fail to erupt. Taken together, our observations suggest that epithelial remodelling plays a crucial role in eruption of the teeth in this model organism. This work was supported by grants from the FWO (no. 31513695) and the ‘Bijzonder Onderzoeksfonds’ of Ghent University (no. 01102995) A. Huysseune ()) Biology Department, Ghent University, Ledeganckstraat 35, 9000 Gent, Belgium e-mail: [email protected] Tel.: +32-9-2645229 Fax: +32-9-2645344 J.-Y. Sire Equipe “Evolution & Dveloppement du Squelette dermique”, CNRS UMR 8570, Universit Paris 6, France

Keywords Tooth eruption · Epithelial morphogenesis · Zebrafish, Danio rerio (Teleostei)

Introduction Tooth eruption is defined as the movement of the tooth from its site of development within the jaw to its functional position in the mouth (Massler and Schour 1941). Despite numerous histological, experimental or clinical studies (reviewed in, a.o., Marks et al. 1988; Marks and Schroeder 1996; Wise 1998), tooth eruption in mammals is a process that is still not fully understood. Eruption of mammalian teeth requires the resorption of bone overlying the crown to create an eruption pathway. Therefore, many tooth eruption studies have concentrated on the cellular events associated with alveolar bone resorption, and the role played by the dental follicle and the enamel epithelium in this process (e.g., Marks 1987; Bidwell et al. 1995). More specifically, factors that regulate the process of alveolar bone resorption indirectly or directly, like epidermal growth factor (EGF) (Thesleff 1987; Wise et al. 1992; Shroff et al. 1996), colony stimulating factor-1 (CSF-1) (Grier et al. 1998; Van Wesenbeeck et al. 2002), parathyroid hormone-related protein (PTHrP) (Philbrick et al. 1998), and osteoprotegerin (Wise et al. 2002), have been the prime target of such studies, and the molecular signals are thought to act in a cascade regulating the initiation of eruption (Wise and Lin 1995). The periodontal ligament, a derivative of the dental follicle, is probably not responsible for generating tractional forces during eruption (Shore and Berkovitz 1979; Pycroft et al. 2002), but this does not preclude a role of its cells in tissue remodelling during the eruptive process (Tsubota et al. 2002; Beertsen et al. 2002). The amount of data collected on factors involved in alveolar bone resorption during eruption stands in sharp contrast to the poor knowledge about a later, albeit less prominent, stage in the eruption process, mucosal penetration (see, e.g., Shibata et al. 1995; Kaneko et al. 1997).

86

Unlike in mammals, teeth in teleost fish can develop in an extraosseous (extramedullary) or in an intraosseous (intramedullary) location, both in the oral and the pharyngeal region (reviewed by Trapani 2001). In the former, the teeth develop on the surface of the bone to which they will eventually attach. In intraosseous development, the tooth develops within the medullary cavity of the bone, and like in mammals, an eruption pathway has to be created to allow passage of the tooth. Teleosts display tooth renewal throughout life (both in the extraand intraosseous situation) and the process of tooth eruption is therefore a repetitive event. On various occasions, a temporal coincidence between deposition of attachment bone, or ankylosis, on the one hand, and eruption on the other hand, has been noticed (e.g., cichlids: Huysseune 1983; Huysseune and Sire 1998; callichthyids: Huysseune and Sire 1997). In zebrafish, a widely used model for vertebrate development, teeth develop extraosseously and erupt almost immediately after ankylosis (Huysseune et al. 1998; Van der heyden et al. 2000). Both for the first-generation and replacement teeth, we have suggested the possible role of remodelling of the epithelium in tooth eruption in this species (Huysseune et al. 1998; Van der heyden et al. 2000). Because teeth in the zebrafish develop in an extraosseous position, eruption is defined here as piercing of the tip through the pharyngeal epithelium, irrespective of whether the tip of the tooth is still hidden by folds of the epithelium, or distinctly protrudes into the pharyngeal cavity. In the present paper, we have tried to elucidate the mechanism leading to the eruption of first-generation teeth in the zebrafish. To this end, in vivo stages of tooth development and eruption have been studied using light (LM) and transmission electron microscopy (TEM). In addition, we have examined larval heads cultured in vitro in order to establish whether the culture conditions used interfere with eruption of the developing teeth. On the basis of our detailed morphological observations, we propose a hypothesis on the mechanism of tooth eruption in this model organism.

Materials and methods Zebrafish larvae and juveniles, raised at 28C, were fixed at 48, 56, 72 and 80 h postfertilization (hPF) and at 4–6, 10 and 30 days postfertilization (dPF), in a mixture of 1.5% glutaraldehyde and 1.5% paraformaldehyde in 0.1 M cacodylate buffer and processed for embedding in Epon according to procedures described elsewhere (Huysseune and Sire 1992). Stages up to 10 dPF were selected because they cover the period prior to, and after, eruption of the first three teeth to develop in larval zebrafish (for details of patterning, see Van der heyden and Huysseune 2000). Older animals (30 dPF) served to check for the morphological characters of the epithelium around erupted replacement teeth. Serial 1- or 2-m sections, stained with toluidine blue, were observed using light microscopy and served to make 3D reconstructions of the epithelium and erupting teeth using the Jandell PC-3D software. In order to determine the position of the epithelial surface with respect to an erupting tooth, handmade reconstructions were prepared of tooth 4V1 (the very first tooth to form in the dentition,

cf. Van der heyden and Huysseune 2000) and of the epithelial surface in two successive stages (one prior to eruption, the other immediately after eruption). These reconstructions were subsequently superimposed, taking the cartilaginous support (the ceratobranchials of the fifth branchial arch) as a reference point. In some specimens, serial sectioning was interrupted for ultrathin sectioning and subsequent TEM observations. Ultrathin sections were contrasted with uranyl acetate and lead citrate and observed with a Philips TEM operating at 80 kV. Heads of 48, 56, 72 and 80 hPF zebrafish were cultured in a serum-free chemically defined medium (Koumans and Sire 1996) at 25.5C for 3 and 7 days, respectively (six replicas for each combination). Explants were fixed, serially sectioned and observed by LM and TEM, as described above.

Results The zebrafish has pharyngeal teeth only, which develop on the ceratobranchials of the fifth branchial arch (ceratobranchials V). By 10 dPF, three of the five firstgeneration teeth in the ventral tooth row have developed and erupted (first 4V1, followed by 3V1 and 5V1) (for details on patterning, and illustrations of the zebrafish dentition, see Huysseune et al. 1998; Van der heyden and Huysseune 2000). The results pertain especially to teeth 4V1 and 3V1. In vivo development Three-dimensional reconstructions of the dentigerous area show the progressive folding of the surface of the ventral pharyngeal epithelium in relation to tooth development, from 56 to 80 hPF (Fig. 1). Superimposed 3D reconstructions of tooth 4V1 prior to and after eruption show that the proximal part of the tooth elongates and that there is a possible, albeit slight, rotational movement of the tooth. The overall position of the epithelial surface remains unchanged throughout the eruption process, but invaginations of the epithelium surround the tooth tips after eruption (Fig. 2). In 48 hPF larvae, there is no pharyngeal lumen yet, despite the visible anlage of the first tooth germ, 4V1 (Fig. 3). Thus, the prospective dorsal and ventral pharyngeal epithelium are not yet separated. In 56 hPF larvae, the pharyngeal lumen has formed in this area and the two epithelia are now separated. The ventral epithelium has a fairly smooth surface, apart from a small depression lying exactly opposite the boundary between germs 4V1 and 3V1 (Figs. 1A, 4). This depression represents the anlage of the primary crypt on each side. Tooth germ 4V1 is now in a stage of cytodifferentiation, with matrix being deposited. Larvae of 72 hPF show a gutter-like depression on each side of a longitudinal, anteroposteriorly oriented slight elevation of the pharyngeal epithelium (Fig. 1B). Tooth 4V1 is now well differentiated but not yet attached (Fig. 5). It is flanked by two less-developed germs (3V1 and 5V1). TEM micrographs of the pharyngeal epithelium in the dentigerous area reveal two to three cell layers

87

Fig. 1A–C Drawings based on 3D reconstructions of the ventral pharyngeal epithelium and of the first developing teeth in larval zebrafish. Different shades of grey reflect surfaces that lie more superficially (more dorsally, turned towards the pharyngeal cavity: white or light grey), or deeper (more ventrally, turned away from the pharyngeal cavity: darker grey). A 56 hPF: the pharyngeal epithelium is fairly smooth; a shallow depression is present at this stage but is not very pronounced yet; tooth 4V1 has some matrix deposited (not pictured); B 72 hPF: anlagen of the primary crypts are forming medial to the two 4V1 teeth, which are still unerupted; C 80 hPF: the primary crypts are now well pronounced and both 4V1 teeth have erupted (their tip pierces the lateral-facing wall of the crypt). Orientations as indicated: Ca caudal, Cr cranial, D dorsal, L lateral, M medial, V ventral; eruption of the tooth tip follows a medially and dorsally directed orientation

Fig. 2A, B Handmade reconstructions based on drawings of serial sections. Surface of the ventral pharyngeal epithelium (PE) and position of unerupted (A 72 hPF) and erupted (B 80 hPF) 4V1 teeth; epithelial surface and outline of the ceratobranchial V cartilage (CBV) in A is projected in drawing B (dashed lines) using the cartilage as an ‘external’ reference point; note the similar overall position of the epithelial surface prior to and after eruption of the teeth

resting on a smooth basal lamina (Fig. 6). The basal layer of this epithelium, facing the non-dental mesenchyme, consists of cuboidal cells, whereas cells of the superficial layer are more flattened. The epithelium also houses differentiating mucous cells. The distal membrane of the superficial cells is electron-dense. It is smooth except at the junction of adjacent cells where a protrusion flanks each side of the junctional complex (Fig. 7). Scanning electron micrographs reveal that such protrusions form ridges (data not shown) and are therefore appropriately called microridges (cf. Elliot 2000). In 80 hPF larvae, the gutter-like depressions have deepened and form what is further called the primary crypts. The primary crypts run on both sides of a crestlike elevation along the midline of the pharyngeal cavity (Figs. 1C, 8). The basal lamina follows the curvature of the epithelial surface. Tooth 4V1 is now attached. Its tip pierces the lateral wall of the primary crypt but does not protrude above the overall surface of the lining of the pharyngeal cavity floor. Tooth germs 3V1 and 5V1 are still in a phase of differentiation and are not yet attached. Their tip is still surrounded by the enamel organ and lies well below the epithelial surface. At 4 dPF, the primary crypts are well developed; they face the boundary between the enamel organs of tooth 4V1 and 3V1 (Fig. 9). At the ultrastructural level, the epithelium has thickened, and abundant mucous cells have clearly differentiated (Fig. 10). The crypt is lined by several cells which, like in the pharyngeal epithelium elsewhere, display hardly any superficial ornamentation. Microridges are present adjacent to the junctional complexes only. The epithelial cells located between the floor of the crypt and the enamel organ of tooth 4V1 display no special characteristics. Yet, at a higher magnification, it is clear that adjacent cells are in close relationship with each other via long and thin cytoplasmic processes (Fig. 11). However, desmosomal contacts are rarely observed between these cytoplasmic processes. The morphological characteristics of the cells surrounding the now erupted portion of tooth 4V1 are well visible in 5 dPF specimens. Close to where the tooth protrudes into the primary crypt, the cells that line the crypt show elaborate cytoplasmic processes on their surface (Fig. 12). In some cases, these processes interdigitate with those of cells at the opposite side of the crypt. In 4 and 5 dPF larvae, a second slit-like invagination, called the secondary crypt, appears medial to each primary crypt, and is distinctly present at 6 dPF (Fig. 13). Together, the primary and secondary crypts delimit an epithelial crest. Prior to its eruption, the tip of tooth 3V1 lies embedded within this crest. At 6 dPF, tooth 3V1 has just erupted or is about to do so (Fig. 14). Its tip emerges from the epithelial crest between the primary and secondary crypts and protrudes into the secondary crypt. Tooth 5V1 is also about to erupt (Fig. 15). Three more tooth germs are present at this stage, at most in a stage of cytodifferentiation. These have not yet attached to the ceratobranchial cartilage. At the

88

89 1

ultrastructural level, the exposed part of tooth 4V is surrounded by epithelial cells, the surface of which is ornamented with numerous, distinct microridges (Fig. 16). These cells adjoin the cells of the pharyngeal epithelium proper. Deeper towards the bottom of the crypt, where the tooth emerges from the epithelium, the microridge-bearing epithelial cells are folded back and cover the tooth matrix over a short distance (Fig. 17). This covering layer extends into a thin cytoplasmic sleeve, on average 0.15–0.2 m thick, surrounding the proximal portion of the exposed part of the tooth. Some microrFigs. 3–27 Light (Figs. 3–5, 8, 9, 13–15, 22, 26–27) and transmission electron (Figs. 6, 7, 10–12, 16–21, 23–25) micrographs of the development of the pharyngeal region in the dentigerous area of zebrafish larvae and juveniles from 48 hPF up to 30 dPF. Left body side is to the right in the pictures. If one body side is figured, it is the left. Fig. 3 48 hPF (just hatched). Arrowheads indicate the position of the basal lamina (dotted line) constituting the dorsal (two arrowheads at the top) and ventral (two arrowheads at the bottom) boundary of the pharyngeal epithelium, respectively. The developing tooth 4V1 is indicated on each body side. Scale bar 25 m Fig. 4 56 hPF. Large arrowheads indicate a small depression in the epithelial surface on both sides of the midline. The two buds of tooth 3V1 (just initiated) are indicated by two small arrowheads; also indicated are teeth 4V1 (with some matrix already deposited, small white arrows). Scale bar 25 m Fig. 5 72 hPF. The floor of the pharyngeal cavity presents two gutter-like depressions (arrowheads) opposite the boundary between germs 4V1 and 3V1. Teeth 4V1 have not yet erupted. The germs of 5V1, which are in a similar stage of development as 3V1, are positioned at a level more posterior than that shown. The boxed area is enlarged in Fig. 6. Scale bar 25 m Fig. 6 Higher magnification of the boxed area in Fig. 5, showing the gutter-like depression. White arrows (extreme left and extreme right of figure) point to the basal lamina at the limit between the pharyngeal epithelium and the non-dentigerous mesenchyme. Arrowhead points to the boundary between the enamel organs of tooth 3V1 and 4V1. Mucous cells have started to differentiate (asterisk). The pharyngeal epithelium on the right side of the picture is three-layered. Scale bar 5 m Fig. 7 Epithelial surface lining the gutter-like depression in another specimen of 72 hPF. Microridges (arrowheads) ornament the surface of the superficial epithelial cells at their junction. Scale bar 2 m Fig. 8 80 hPF. Section immediately anterior to where teeth 4V1 erupt. Their tips emerge from the lateral wall of the primary crypts (arrowheads). Compare with Fig. 1C. Scale bar 25 m Fig. 9 4 dPF. Section anterior to the tip of tooth 4V1. Primary crypts are indicated by arrowheads. The section is slightly oblique, one body side (showing both 3V1 and 4V1) being further advanced. The boxed area is enlarged in Fig. 10. Scale bar 25 m Fig. 10 Higher magnification of the boxed area in Fig. 9. The enamel organ of tooth 4V1 is marked with a black asterisk. Arrowheads point to where the pharyngeal epithelium proper is connected to the enamel organ of 4V1. The boxed area is enlarged in Fig. 11. Scale bar 2 m Fig. 11 Higher magnification of the boxed area in Fig. 10. The plasma membranes of adjacent epithelial cells are interdigitating (arrowheads). Scale bar 0.5 m Fig. 12 5 dPF. Section adjacent to the tip of tooth 4V1. Superficial cells are interconnected by long cell processes (arrowhead). Adjacent to tooth 4V1 lies tooth 3V1. Scale bar 5 m

idges of the epithelial cells opposite the tooth make contact with this sleeve (Fig. 17). Further distally, the microridges make contact with the naked tooth surface and still more distally, towards the tooth tip, turn away from the tooth surface (Fig. 16). Both the microridgebearing cells and the cells that extend into a sleeve are considered to be part of the former enamel epithelium (outer and inner dental epithelial cells, respectively). Indeed, in another specimen, folds of the cell membrane of the outer dental epithelium cells are suggestive of prospective microridge formation; as a whole, the enamel organ of 4V1 is separated from the cells of the adjacent, pharyngeal epithelium by widened intercellular spaces (Fig. 18). Tooth 3V1 presents a picture that is different from that of 4V1. This difference is probably explained by the different orientation of the two teeth. In contrast to tooth 4V1, which is transversely oriented, tooth 3V1 is oriented anteroposteriorly. In a series of transverse sections, the upper part of tooth 3V1 runs inside the epithelial crest between primary and secondary crypts (Fig. 19), before eventually piercing the secondary crypt more posteriorly in the series. Below where tooth 3V1 emerges from the epithelium, it is surrounded by its enamel organ, probably covered itself by a superficial thin layer of flattened, ornamented, epithelial cells (Fig. 19). The cells lining both the primary and secondary crypts bear long cellular processes as well as prominent microridges. Both the inner and outer dental epithelium cells contain large bundles of thin filaments (Fig. 20). In the deep part of the secondary crypt the superficial cells encircling the enamel organ of 3V1 show long cell processes, which contact either similar cell processes, or microridges, at the opposite side of the crypt (Fig. 21). At 10 dPF, tooth 4V1 emerges much further from the mucosa than before (Fig. 22). The topmost cell of the outer dental epithelium shows microridges distally, as well as along the membrane facing the cytoplasmic sleeve extending from the topmost inner dental epithelial cell1 (Fig. 23). Deeper, the outer dental epithelial cells are still interdigitated with the inner dental epithelial cells (Fig. 24). Close to the tooth base, the surface of the tooth shows small, more or less regularly spaced, patches composed of electron-dense, granular, non-collagenous matrix (Fig. 25). Depending on the orientation of the tooth, and on the level of sectioning, these patches can be seen either to be surrounded by cytoplasm of the inner dental epithelial cells, or to be connected to the peripheral, granular tooth matrix, which itself represents the remains of the basal lamina (Fig. 25). The material being decalcified, it is unknown whether these patches are mineralized. 1 Note that in the present context we prefer to use the term ‘inner dental epithelial cell’ rather than ‘ameloblast’ in later stages because zebrafish teeth are only partly covered by a hypermineralized—enameloid—layer; functional ameloblasts are therefore limited to the distal part of the tooth. The boundary between functional ameloblasts and inner dental epithelial cells is not visible after the decalcification procedure used.

90

91

During larval development the number of mucous cells increases dramatically. The first mucous cells to be identified as such are observed at 72 hPF and they are extremely numerous at 30 dPF, making up almost the only cell type present in the superficial layer of the pharyngeal epithelium (Fig. 26). At this stage, the entire epithelial covering of the dentigerous area is highly folded. The tooth tips erupt in deep crypts, and the epithelium bordering these crypts (called crypt epithelium) is characterized by numerous mucous cells. Yet, a thin epithelial sleeve deriving from the former inner dental epithelium is always present around the proximal portion of the erupted part of the tooth, as in larval stages. Depending on the orientation of the tooth, and of the section, teeth can be found entirely encircled by such crypt epithelium (Fig. 26). It is also clear that replacement teeth develop in the bottom of a crypt, at the interface of the former enamel epithelium and the crypt epithelium proper.

Development under in vitro conditions In heads of larval zebrafish, cultured in vitro for a variable number of days, the tooth germs present at the time of explanting (depending on the stage: 4V1 alone, 4V1 and 3V1, or 4V1 together with 3V1 and 5V1) undergo normal morphogenesis and cytodifferentiation (Fig. 27). During the incubation period, tooth 4V1 also deposits attachment bone and becomes ankylosed to the perichondral bone surrounding the ceratobranchial cartilage, as in normal development. Opposite the interface between teeth 4V1 and 3V1, a crypt is distinctly present, corresponding to the primary crypt in normal development. However, the tip of 4V1 lies completely embedded in the epithelium adjacent to the crypt, and has failed to erupt, although the incubation time is largely sufficient to allow eruption to be completed in normal conditions.

Discussion Fig. 13 6 dPF. Note the distinct presence of a primary crypt (black arrowhead), and medial to it a secondary crypt (white arrowhead). Tooth 4V1 erupts into the primary crypt; tooth 3V1 will eventually erupt into the secondary crypt. Scale bar 25 m Fig. 14 6 dPF. Level of eruption of tooth 3V1 (white asterisks); on the left side, its tip pierces the lateral wall of the secondary crypt (black arrowhead). A white arrowhead indicates the position of the primary crypt. Scale bar 50 m Fig. 15 6 dPF. Level posterior to that shown in Fig. 14. Eruption of tooth 5V1 (asterisk). Scale bar 50 m Fig. 16 6 dPF. Erupted part of tooth 4V1. Arrowheads point to microridges. A pharyngeal epithelial cell is indicated by an asterisk. Scale bar 2 m Fig. 17 Same tooth (4V1) as in Fig. 16; more towards the bottom of the crypt. A thin cytoplasmic sleeve (arrowheads) extending from the topmost inner dental epithelial cell (IDE) covers the tooth where it emerges from the epithelium. Microridges of the opposite cell surface (asterisks) make contact with this sleeve. Scale bar 1 m Fig. 18 Same erupted tooth (4V1) in another 6 dPF specimen. Highly interdigitated inner dental epithelium (IDE) and outer dental epithelium (ODE) cells are partly detached from each other (arrowhead). As a whole, these cell layers are partly detached from the adjacent pharyngeal epithelium (asterisk). Scale bar 2 m Fig. 19 6 dPF. Upper part of tooth 3V1 surrounded by cells of the epithelial crest. This section level is anterior to the eruption site of tooth 3V1. Note the deep secondary crypt along the medial side of tooth 3V1 (black arrowhead). The primary crypt is partially visible on its lateral side (white arrowhead). The boxed area is enlarged in Fig. 20. Scale bar 5 m Fig. 20 Higher magnification of the boxed area in Fig. 19, showing the inner dental epithelial cells (IDE) at the surface of tooth 3V1. Note the distinct presence of bundles of thin filaments (arrowhead). Scale bar 0.5 m Fig. 21 Same specimen. Deep part of the secondary crypt (arrowhead). Black asterisk indicates the enamel organ of tooth 3V1. In this area, the enamel organ is still covered by a superficial epithelial layer that is part of the original pharyngeal epithelium (white asterisk). Scale bar 1 m Fig. 22 10 dPF. Slightly oblique section, with erupted teeth 3V1 and 4V1 on one side, and erupted tooth 5V1 at the other side. All other germs are of replacement teeth. Scale bar 50 m

In earlier studies on the zebrafish dentition, we were repeatedly struck by the sudden exposure of a tooth into the pharyngeal cavity shortly after its attachment to the underlying cartilage, hence without apparent concomitant growth or movement of the tooth (Huysseune et al. 1998; Van der heyden et al. 2000). This was confirmed in the present study by superimposing reconstructions of the same tooth in successive developmental stages (Fig. 2). These reconstructions indeed show that the mutual positions of the epithelial surface and the tooth remain essentially unchanged throughout the eruption process. A slight rotation, if confirmed, could bring the tooth tip closer to the crypt, but is not sufficient to expose the tip into the pharyngeal cavity. Earlier we already proposed that some sort of remodelling of the epithelium might be involved in the eruption process (cf. Huysseune et al. 1998; Van der heyden et al. 2000), reminiscent of the formation of epithelial troughs around the developing tooth plates in the dipnoan Neoceratodus forsteri described by Kemp (1995). The tooth pattern in the zebrafish having been completely resolved (Van der heyden and Huysseune 2000), it was now possible to monitor individual teeth throughout their development and eruption process. The observations on the events accompanying eruption of first-generation teeth can be summarized as follows: once a lumen has been formed in the pharyngeal area, the ventral epithelium first forms shallow depressions, which become progressively more gutter-like. We have called these gutter-like depressions crypts. The first crypt forms medial to where tooth 4V1 will erupt, the second medial to tooth 3V1. Teeth erupt when their tips emerge through the lateral-facing wall of these crypts. Initially, the tooth tips lie within these crypts and do not protrude above the overall epithelial surface. Later, the epithelial surface becomes more elaborately folded, retracts, and the tooth tip now protrudes into the pharyngeal cavity.

92

Fig. 23 Tooth 4V1 in a 10 dPF zebrafish. Note the limit between the outer dental epithelium (white asterisk) and the pharyngeal epithelium (mucous cell, black asterisk), and the presence of the cytoplasmic sleeve (arrowheads), extending from the inner dental epithelium (IDE), along the tooth surface. The cell marked by white asterisk is the topmost outer dental epithelium cell and bears many microridges. The boxed areas are enlarged in Figs. 24 and 25, respectively. Scale bar 5 m Fig. 24 Higher magnification of the uppermost boxed area in Fig. 23. Cells of the outer (ODE) and inner dental epithelium (IDE) are interdigitated proximally (white arrowhead). Microridges ornamenting the outer dental epithelium cells seem to derive from such interdigitations (black arrowheads). Scale bar 1 m

Fig. 25 Higher magnification of the lowermost boxed area in Fig. 23. The outer surface of the tooth (remains of the basal lamina, BL) bears small patches of a non-collagenous matrix (black arrowhead), of the same granular nature as the basal lamina, and similar in structure to deposits encircled by the cytoplasm of the inner dental epithelium cells (white arrowheads) (D dentine, OD odontoblast). Scale bar 1 m Fig. 26 30 dPF. Note the elaborate folding of the epithelium, with crypts (black arrowheads) encircling erupted teeth. The lining of these crypts is characterized by the presence of numerous mucous cells (small white asterisks). Every erupted tooth is covered partly by a thin cytoplasmic sleeve (white arrowheads). Note the presence of a replacement tooth germ (large white asterisk) budding off from an area in the bottom of the crypt at the limit between the crypt

93

Based on these observations, we have deduced a hypothesis on the mechanism of tooth eruption in zebrafish, which is pictured in a schematic way in Fig. 28. Tooth eruption in larval zebrafish is proposed to be a biphasic process, with (1) the formation of a crypt, and (2) exposure of the tooth tip in this crypt by retraction of the pharyngeal epithelium. The first process, the formation of the crypt, is thought to result from mechanical constraints in the epithelium, possibly provoked by the growth of the developing tooth germs. Indeed, the epithelium is always folded (i.e., a crypt is formed) opposite the medial margin of a developing tooth germ. Because of its orientation and shape, a tooth will necessarily erupt in the crypt along its lateral border. The close developmental relationship between the presence, and localization, of crypts and the development of tooth germs is further suggested by the one-to-one relationship between crypts and teeth which are about to erupt. Crypts are also present in specimens cultured in vitro, where the first-generation teeth develop normally. All these observations seem to support the hypothesis that crypts form in association with the development of tooth germs. There is no direct observation that could explain why the crypts are formed at the edge of tooth germ anlagen, rather than directly opposite to them. Yet, the (thinner) pharyngeal epithelium in these areas may well be less resistant to mechanical forces than the (thicker) epithelium opposite tooth germ anlagen, resulting in folding taking place preferentially in these sites. If this speculation is correct, growth of the germs would automatically lead to the correct positioning of the crypts. This dependence could be tested by explanting heads in vitro at a stage early enough to prevent formation of first-generation teeth, and to examine whether or not crypts are formed. The second process, exposure of the tooth within the crypt, starts by detachment of the long interdigitations that keep the epithelial cells of the pharyngeal epithelium proper together. First, the epithelial cell layer covering the enamel organ detaches from the outer dental epithelium cells, and then the outer detaches from the inner dental epithelium, both probably as a result of disruption of intercellular contacts. The cytoplasmic processes resulting from the disconnection of the interdigitations appear to turn progressively into microridges. Indications for intercellular detachment in vivo are seen in (1) the fact that along the membrane of a single cell interdigitations can grade into microridge-like protrusions, and (2) the abundance of extremely tall microridges (seen as further

epithelium and the remains of the enamel organ of the functional tooth. Scale bar 25 m Fig. 27 Pharyngeal region of a 48 hPF specimen, cultured in vitro for 3 days. Note the distinct presence of primary crypts (white arrowheads). The tip of tooth 4V1 (asterisk) is completely surrounded by epithelium and has not erupted. In normal conditions, loss of epithelial integrity in the area indicated by a black arrowhead should have allowed the tip to become exposed. Scale bar 25 m

Fig. 28 Interpretative scheme of events associated with the eruption of a first-generation tooth in larval zebrafish. The drawings represent four (1–4) successive stages in the formation and eruption of tooth 4V1 and 3V1; large arrows show the succession of these stages. The boxed area in 4 is enlarged in 5. 1
Formation of primary crypts (arrowheads) in the ventral pharyngeal epithelium opposite the boundary between the developing tooth germs 4V1 and 3V1 (labelled); tooth 4V1 is in a stage of morphogenesis. 2
Deepening of primary crypts (arrowheads) and onset of formation of secondary crypts (small arrows); matrix has been deposited for tooth 4V1; tooth 3V1 is in a stage of morphogenesis. 3
Eruption of teeth 4V1 in the primary crypts (arrowheads) and deepening of the secondary crypts (arrows); matrix has been deposited for tooth 3V1. 4
Further exposure of tooth 4V1. Lighter shade of grey epithelium, darker shade of grey tooth matrix, CBV ceratobranchial V cartilage (cartilage light grey, surrounded by perichondral bone, darker grey, in 3 and 4), IDE inner dental epithelium, ODE outer dental epithelium, PE ventral pharyngeal epithelium, TM tooth matrix

support that they derive from long cytoplasmic interdigitations). The patches of non-collagenous matrix observed on the surface of the proximal part of erupted teeth, and which most likely represent a secondary deposit, possibly play a role in preventing complete detachment of the enamel organ from the tooth. The idea that a second process is involved and that eruption is probably biphasic, with two independent phases, is supported by several observations: (1) crypts are formed prior to eruption of the teeth, (2) the crypt stretches over a far greater length than just in that particular area where the tooth tip will become exposed,

94

suggesting that the formation of crypts is a process that is independent of the eruption process, and (3) observations on in vitro cultured head explants. In the latter conditions, the well-formed teeth do not erupt, despite the presence of crypts. We speculate that the culture conditions used do not permit the cells to achieve disruption of their interdigitations and so the cells cannot detach from one another, a process required to permit the tooth tip to penetrate into the crypt. Although the two processes, formation of the crypt and eruption of the tooth through detachment of cellular processes, are assumed to succeed each other in time, it is likely that further deepening of the crypt is also aided by a similar detachment of interdigitations, as has been observed for tooth 3V1. Different from the situation in mammals (Marks and Schroeder 1996), the enamel organ in zebrafish remains broadly connected to the pharyngeal epithelium throughout the development of the tooth. This hampers identification of the different epithelial cell layers during the eruption process. The cell layer at the contact with the proximal part of the erupted tooth can be safely interpreted as the former inner dental epithelium, and the epithelial sleeve that extends from it is therefore part of the former inner dental epithelial cells. This sleeve completely surrounds the proximal portion of erupted teeth. Its function might be to prevent food particles from entering the space between the tooth and the surrounding soft tissues. The former outer dental epithelium can only be identified by its topographical relationship to the inner dental epithelium, e.g., where interdigitations with the inner dental epithelial cells have persisted. If the cells of the outer dental epithelium indeed turn into microridgebearing cells, as we have deduced from our observations, they cover the more distal portion of the erupted tooth. Unless markers become available to identify the different epithelial cell layers covering the tooth prior to eruption, their fate nevertheless remains uncertain. So far, we have dealt with first-generation teeth only. Replacement teeth also start to form early in postembryonic development [the first replacement tooth for 4V1 (i.e., 4V2) is initiated between 3 and 4 dPF], and they continue to form on the various tooth positions throughout life (see Van der heyden et al. 2000; Van der heyden and Huysseune 2000 for details). Formation of replacement teeth is initiated at the contact zone between the enamel organ of the predecessor and the pharyngeal epithelium proper (Van der heyden et al. 2000) in the bottom of the crypt of the predecessor. Whether the replacement tooth erupts in a pre-existing crypt, or the epithelium is remodelled for each tooth anew, is unknown. To answer this question, a detailed monitoring of replacement teeth is required. Very few researchers investigating fish dentitions have paid attention to eruption. Berkovitz (1977), working on trout dentitions, noted that in some of the early teeth the tip barely penetrated the oral epithelium. This suggests that the tooth tips lie in folds of the epithelium, which, in turn, suggests that crypts might also form in this species.

Tooth eruption has been most thoroughly studied in mammalian teeth. Yet, to our knowledge, such a kind of epithelial remodelling during eruption of mammalian teeth has not been reported. Different from zebrafish, the enamel organ in mammals is isolated from the overlying oral epithelium by bone and soft connective tissues. During eruption, mammalian teeth have to move from their site of development within the jaw to their functional position in the mouth. Marks and Schroeder (1996) distinguished five different phases in this process, only one of which is mucosal penetration. The occlusal surface of the tooth crown, covered with the reduced enamel epithelium (i.e., the ameloblasts and other remaining cells of the dental organ), first moves orally to the gum surface. During mucosal penetration, the reduced enamel epithelium fuses with the oral epithelium, and the degeneration of this epithelial plug allows passage of the tooth (Shibata et al. 1995). Both Shibata et al. (1995) and Kaneko et al. (1997) found apoptotic cells possibly involved in disruption of the epithelium in mice and rats, respectively, and thus in the formation of an opening through which the tooth erupts. So far, we have not been able, with TUNEL staining, to detect apoptoses in the epithelial tissues surrounding the erupting teeth in larval zebrafish (unpublished observations). Yet, we suspect that some cells of the dental organ die at the moment of eruption. Further detailed studies are needed to confirm this. Different from mammals, the importance of epithelial remodelling for tooth eruption in the zebrafish might well be linked to the small size of the teeth with respect to the thickness of the pharyngeal epithelium (at least in young ontogenetic stages), and/or their location on the jaw surface (extraosseous tooth formation, cf. Trapani 2001), requiring hardly any movement. The eruption process proposed here for the zebrafish teeth might therefore either be proper to extramedullary teeth, or represent a final phase in the eruption of intramedullary teeth. In teleosts which have an intraosseous tooth development, tooth eruption is more reminiscent of the eruption of mammalian teeth, since teeth have to move from their intramedullary location towards their functional position on top of the jaw. In cichlids, e.g., pharyngeal replacement teeth develop in an intramedullary position, and here, too, epithelial crypts surround the exposed part of the tooth (personal observations). Therefore, the presence of crypts in teleosts appears to be independent of the place of formation of teeth, i.e., extra- or intraosseous. In conclusion, we propose that the eruption of the firstgeneration teeth in the zebrafish results from two, independent events: (1) the formation of crypts, which become positioned opposite the edges of developing tooth germs, and (2) the disruption of contacts within the pharyngeal epithelium and within the enamel organ layers. Several issues remain unresolved. First, we need to confirm that the formation of primary, secondary, etc., crypts is dependent upon tooth germ development. Second, further evidence for detachment of intercellular contacts preceding tooth eruption has to be collected. Finally, the involvement of these mechanisms in the

95

eruption of replacement teeth during juvenile and adult life stages of the zebrafish has to be clarified. Acknowledgements The authors acknowledge expert technical assistance from Miss N. Van Damme and Miss F. Allizard. TEM work was carried out at the Centre de Microscopie Electronique, Universit Paris 6–CNRS. Many thanks to Dr. C. Van der heyden for critically reading the manuscript.

References Beertsen W, Holmbeck K, Niehof A, Bianco P, Chrysovergis K, Birkedal-Hansen H, Everts V (2002) On the role of MT1-MMP, a matrix metalloproteinase essential to collagen remodeling, in murine molar eruption and root growth. Eur J Oral Sci 110:445–451 Berkovitz BKB (1977) The order of tooth development and eruption in the rainbow trout (Salmo gairdneri). J Exp Zool 201:221–226 Bidwell JP, Fey EG, Marks SC (1995) Nuclear matrix-intermediate filament proteins of the dental follicle/enamel epithelium and their changes during tooth eruption in dogs. Arch Oral Biol 40:1047–1051 Elliott DG (2000) Integumentary system. In: Ostrander GK (ed) The laboratory fish. Academic Press, San Diego, pp 271–306 Grier RL, Zhao LH, Adams CE, Wise GE (1998) Secretion of CSF1 and its inhibition in rat dental follicle cells: implications for tooth eruption. Eur J Oral Sci 106:808–815 Huysseune A (1983) Observations on tooth development and implantation in the upper pharyngeal jaws in Astatotilapia elegans (Teleostei, Cichlidae). J Morphol 175:217–234 Huysseune A, Sire JY (1992) Development of cartilage and bone tissues of the anterior part of the mandible in cichlid fish: a light and TEM study. Anat Rec 233:357–375 Huysseune A, Sire JY (1997) Structure and development of teeth in three armoured catfish, Corydoras aeneus, C. arcuatus and Hoplosternum littorale (Siluriformes, Callichthyidae). Acta Zool 78:69–84 Huysseune A, Sire JY (1998) Structure and development of first generation teeth in the cichlid Hemichromis bimaculatus (Teleostei, Cichlidae). Tissue Cell 29:679–697 Huysseune A, Van der heyden C, Sire JY (1998) Early development of the zebrafish (Danio rerio) pharyngeal dentition (Teleostei, Cyprinidae). Anat Embryol 198:289–305 Kaneko H, Ogiuchi H, Shimono M (1997) Cell death during tooth eruption in the rat: surrounding tissues of the crown. Anat Embryol 195:427–434 Kemp A (1995) Marginal tooth-bearing bones in the lower jaw of the recent Australian lungfish, Neoceratodus forsteri (Osteichthyes, Dipnoi). J Morphol 225:345–355 Koumans JTM, Sire JY (1996) An in vitro, serum-free organ culture technique for the study of development and growth of the dermal skeleton in fish. In Vitro Cell Dev Biol 32:612–626 Marks SC (1987) Tooth eruption: the regulation of a localized, bilaterally symmetrical metabolic event in alveolar bone. Scan Microsc 1:1125–1133

Marks SC, Gorski JP, Cahill DR, Wise GE (1988) Tooth eruption— a synthesis of experimental observations. In: Davidovitch Z (ed) The biological mechanisms of tooth eruption and root resorption. EBSCO Media, Birmingham, Alabama, pp 161–169 Marks SC, Schroeder HE (1996) Tooth eruption: theories and facts. Anat Rec 245:374–393 Massler M, Schour I (1941) Studies in tooth development: theories of eruption. Am J Orthodont Oral Surg 27:552–576 Philbrick WM, Dreyer BE, Nakchbandi IA, Karaplis AC (1998) Parathyroid hormone-related protein is required for tooth eruption. Proc Natl Acad Sci USA 95:11846–11851 Pycroft JM, Hann A, Moxham BJ (2002) Apoptosis in the connective tissues of the periodontal ligament and gingivae of rat incisor and molar teeth at various stages of development. Conn Tissue Res 43:265–279 Shibata S, Suzuki S, Tengan T, Yamashita Y (1995) A histochemical study of apoptosis in the reduced ameloblasts of erupting mouse molars. Arch Oral Biol 40:677–680 Shore RC, Berkovitz BKB (1979) An ultrastructural study of periodontal ligament fibroblasts in relation to their possible role in tooth eruption and intracellular collagen degradation in the rat. Arch Oral Biol 24:155–164 Shroff B, Kashner JE, Keyser JD, Hebert C, Norris K (1996) Epidermal growth factor and epidermal growth factor-receptor expression in the mouse dental follicle during tooth eruption. Arch Oral Biol 41:613–617 Thesleff I (1987) Does epidermal growth factor control tooth eruption? J Dent Child 54:321–329 Trapani J (2001) Position of developing replacement teeth in teleosts. Copeia 2001:5–51 Tsubota M, Sasano Y, Takahashi I, Kagayama M, Shimauchi H (2002) Expression of MMP-8 and MMP-13 mRNAs in rat periodontium during tooth eruption. J Dent Res 81:673–678 Van der heyden C, Huysseune A (2000) Dynamics of tooth formation and replacement in the zebrafish (Danio rerio) (Teleostei, Cyprinidae). Dev Dyn 219:486–496 Van der heyden C, Huysseune A, Sire JY (2000) Development and fine structure of pharyngeal replacement teeth in juvenile zebrafish (Danio rerio) (Teleostei, Cyprinidae). Cell Tissue Res 302:205–219 Van Wesenbeeck L, Odgren PR, MacKay CA, D’Angelo M, Safadi FF, Popoff SN, Van Hul W, Marks SC (2002) The osteopetrotic mutation toothless (tl) is a loss-of-function frameshift mutation in the rat Csf-1 gene: Evidence of a crucial role for CSF-1 in osteoclastogenesis and endochondral ossification. Proc Natl Acad Sci USA 99:14303–14308 Wise GE (1998) The biology of tooth eruption. J Dent Res 77:1576–1579 Wise GE, Lin F (1995) The molecular biology of initiation of tooth eruption. J Dent Res 74:303–306 Wise GE, Lin F, Fan W (1992) Localization of epidermal growth factor and its receptor in mandibular molars of the rat prior to and during prefunctional tooth eruption. Dev Dyn 195:121–126 Wise GE, Yao S, Zhang Q, Ren Y (2002) Inhibition of osteoclastogenesis by the secretion of osteoprotegerin in vitro by rat dental follicle cells and its implications for tooth eruption. Arch Oral Biol 47:247–254