Cell Tissue Res (2005) 319: 71–89 DOI 10.1007/s00441-004-0950-2

REGULAR A RTICLE

Sidney Delgado . Tiphaine Davit-Béal . Françoise Allizard . Jean-Yves Sire

Tooth development in a scincid lizard, Chalcides viridanus (Squamata), with particular attention to enamel formation Received: 5 August 2003 / Accepted: 15 June 2004 / Published online: 3 November 2004 # Springer-Verlag 2004

Abstract Comparative analysis of tooth development in the main vertebrate lineages is needed to determine the various evolutionary routes leading to current dentition in living vertebrates. We have used light, scanning and transmission electron microscopy to study tooth morphology and the main stages of tooth development in the scincid lizard, Chalcides viridanus, viz., from late embryos to 6-year-old specimens of a laboratory-bred colony, and from early initiation stages to complete differentiation and attachment, including resorption and enamel formation. In C. viridanus, all teeth of a jaw have a similar morphology but tooth shape, size and orientation change during ontogeny, with a constant number of tooth positions. Tooth morphology changes from a simple smooth cone in the late embryo to the typical adult aspect of two cusps and several ridges via successive tooth replacement at every position. First-generation teeth are initiated by interaction between the oral epithelium and subjacent mesenchyme. The dental lamina of these teeth directly branches from the basal layer of the oral epithelium. On replacement-tooth initiation, the dental lamina spreads from the enamel organ of the previous tooth. The epithelial cell population, at the dental lamina extremity and near the bone support surface, proliferates and differentiates into the enamel organ, the inner (IDE) and outer dental epithelium being separated by stellate reticulum. IDE differentiates into ameloblasts, which produce enamel matrix components. In the region facing differentiating IDE, mesenchymal cells differentiate into dental papilla and give rise to odontoblasts, which first deposit a layer of predentin matrix. The first elements of the enamel matrix are then synthesised by ameloblasts. Matrix mineralisation starts in the upper region of the S. Delgado . T. Davit-Béal . F. Allizard . J.-Y. Sire (*) Equipe “Evolution & Développement du Squelette”, Université Paris 6, CNRS FRE 2696, Case 7077, 7 Quai St.-Bernard, 75251 Paris cedex 05, France e-mail: [email protected] Tel.: +33-1-44273572 Fax: +33-1-44275653

tooth (dentin then enamel). Enamel maturation begins once the enamel matrix layer is complete. Concomitantly, dental matrices are deposited towards the base of the dentin cone. Maturation of the enamel matrix progresses from top to base; dentin mineralisation proceeds centripetally from the dentin–enamel junction towards the pulp cavity. Tooth attachment is pleurodont and tooth replacement occurs from the lingual side from which the dentin cone of the functional teeth is resorbed. Resorption starts from a deeper region in adults than in juveniles. Our results lead us to conclude that tooth morphogenesis and differentiation in this lizard are similar to those described for mammalian teeth. However, Tomes’ processes and enamel prisms are absent. Keywords Tooth . Development . Resorption . Transmission-electron microscopy . Scanning-electron microscopy . Chalcides viridanus (Squamata)

Introduction Our extensive investigations into the development of the dermal skeleton in various vertebrate lineages have enabled us to perform comparative developmental analyses of its various elements (for reviews, see Huysseune and Sire 1998; Sire et al. 2002; Sire and Huysseune 2003). These comparative studies clearly demonstrate how crucial such developmental comparisons can be for understanding evolutionary relationships between the various elements of the dermal skeleton (e.g. scales, dermal bones, fin rays, denticles, teeth). In particular, these comparisons have been important for unraveling the history of dental tissues (dentin and enamel) and of a number of dentalrelated tissues (Sire and Huysseune 2003). The review of Sire and Huysseune (2003) includes comparisons of the developmental processes leading to the formation of the various elements of the dermal skeleton in actinopterygian fish. This lineage possesses the richest diversity not only in terms of the number of species, but also in the number of different types of dermal skeletal elements.

72

A recent comparative analysis of tooth development (morphogenesis and differentiation at the tissue and cellular level) in vertebrates has highlighted the lack of such comprehensive information for various reptilian lineages (Sire et al. 2002). Indeed, previous studies have generally been limited to specific aspects of tooth development, as in investigations of tooth development in squamates (lizards and snakes). Since the 1960s, most studies of reptilian (sauropsid and crocodilian) teeth have been devoted to a comparison of dentition patterns. Authors have especially focused their attention on the question as to how replacement waves are generated (Edmund 1960, 1969; Cooper 1966; Cooper et al. 1970; Rieppel 1978; Rocek 1980). Following Edmund’s (1960, 1969) work, knowledge of the pattern of tooth replacement has generated a wide debate in terms of evolutionary implications (Osborn 1970, 1984; DeMar 1972, 1974; Osborn and Crompton 1973) and of the embryological formation of such patterns (e.g. Osborn 1971, 1972; Westergaard 1986; Westergaard and Ferguson 1986). In contrast, histological studies of tooth development in reptiles are either old (e.g. Harrison 1901; Woerdeman 1919; Poole 1957) or have been used as a tool to understand the pattern of developing teeth (e.g. Osborn 1971). However, if we wish to compare tooth development in all vertebrate lineages from an evolutionary perspective, we require a description covering the main stages of tooth development for squamates, from tooth initiation to attachment, and the resorption processes. The aim of the present work has been to fill this gap by studying in detail the various steps of tooth development in a scincid lizard, Chalcides viridanus. This work is based on recent results concerning the dentition and tooth replacement pattern in this species (Delgado et al. 2003); tooth replacement in C. viridanus follows a regular reproducible pattern that allows the accurate prediction of the developmental stage of a tooth at a given position. We have used scanning electron microscopy (SEM) to study changes in tooth shape, size and orientation during the ontogeny of C. viridanus; the number of tooth positions in this lizard does not increase with age. Light (LM) and transmission electron microscopy (TEM) has been employed to study the various steps of tooth development, from early initiation to complete differentiation and attachment. Our results lead us to conclude that tooth morphogenesis and differentiation in this lizard are similar to those described for mammalian teeth. However, Tomes’ processes and enamel prisms are absent.

Materials and methods Biological material Since 1996, the Canarian skink C. viridanus (Gravenhorst) has been bred in our laboratory, where it reproduces under controlled conditions, i.e. a “winter” period of 2.5 months from the end of November to mid February (but with a room temperature never falling below 10°C) and a

“temperate” period (e.g. room temperature) for the rest of the year. Males and females were distributed in several tanks (80 cm long/40 cm wide, with a sandy bottom) and fed (except during the “winter” period) twice a week with insect larvae (mostly maggots) and pieces of fruit and tomatoes. Canarian skinks are viviparous. Although mating has not been observed, 2–3 baby skinks per female have been obtained every year since the summer of 1996, from the second half of June to the end of July. Soon after birth, the baby skinks were placed into small tanks (40/20 cm) and fed with small maggots. Skinks reach sexual maturity at the age of 3 years (J. Castanet, personal communication). Individuals below this age are considered as juveniles. Several specimens, from stage 39 embryos (approximately 2 weeks before birth) to 70-month-old-adults, were used. Stage 39 embryos were estimated according to the table of development provided by Dufaure and Hubert (1961) for Lacerta vivipara. All lizards were measured (from snout to vent, SVL), killed with an overdose of anaesthetic and decapitated, following which the jaws were dissected. All animals used in this study were reared in our laboratory facilities (agreement no. A-75-05-11) and killed according to the guidelines of the French Ethics Committee. Methods SEM procedure The right lower jaw quadrants of five specimens at 1, 4.5, 17, 20 and 70 months of age (35, 39, 60, 70 and 75 mm SVL, respectively) were immersed in 10% sodium hypochlorite for 20–40 min (depending on jaw size). Soft tissues were carefully removed by using forceps and thin brushes. The samples were then dehydrated in a graded series of ethanol, air-dried, glued onto a copper support, coated with a thin layer of gold/palladium, and observed in a JEOL SEM 35 operating at 25 kV. LM and TEM procedures The anterior region of the lower jaw of five specimens (embryo stage 39, 30 mm; 1-day-old, 41 mm; 1-monthold, 43 mm; 5-month-old, 42 mm; 60-month-old adult female, 72 mm SVL) was sectioned into small pieces and fixed for 2 h, at room temperature, in a mixture containing 1.5% glutaraldehyde and 1.5% paraformaldehyde in 0.1 M cacodylate buffer (pH 7.4). Some samples were decalcified during either 7 (small specimens) or 14 days at 4°C in the same fixative to which 5% EDTA was added. After being rinsed in the same buffer overnight at 4°C, the samples were postfixed in 1% osmium tetroxide in cacodylate buffer and rapidly rinsed once more. They were next dehydrated in a graded series of ethanol and embedded in Epon for transverse or longitudinal section-

73

ing. Several samples were processed into uninterrupted series of transverse or longitudinal semi-thin sections (1 μm thick), which were stained with toluidine blue. In a few samples, the series were interrupted at appropriate locations for ultra-thin sectioning. Thin sections were contrasted with uranyl acetate and lead citrate and viewed in a Philips 201 transmission electron microscope operating at 60 kV. Selection of appropriate stages of tooth development C. viridanus replaces its teeth continuously during its entire life span. Thus, numerous teeth were available at various developmental stages (from initiation to attachment) in our growth series, even in adults, except for the first-generation teeth, which could only be studied in the embryo. To select the appropriate stages for TEM, we referred to our previous study in which we showed that tooth replacement in this lizard follows a regular pattern (Delgado et al. 2003). This pattern allows the accurate prediction of the developmental stage of a tooth at a given position in the jaw provided that the developmental stage of at least one tooth is known in the jaw quadrant. To this end, starting from the anterior region, the jaw was transversely sectioned at 1 μm steps until the first replacement tooth was completely sectioned. Once the developmental stage of this tooth was defined, we knew from our previous work that the following replacement tooth was 20% less developed. In this way, we were able to calculate the distance along the jaw to the next replacement tooth at the required stage (the distances were available from SEM data).

Results Tooth shape, wear and resorption during ontogeny SEM observations on the jaws of a growth series of C. viridanus allowed us to collect data (1) on the shape, size, and orientation of the teeth (in juveniles vs in adults and in anterior vs posterior positions), (2) on the relationship of the teeth with regard to the surface of the bony support, i.e. the dentary bone, (3) on aspects of tooth wear, and (4) on the resorption features of the functional teeth in relation to the presence of developing replacement teeth. These data, in addition to the recently described pattern of tooth replacement (Delgado et al. 2003) greatly helped to localise the tooth germs versus the position of the functional teeth and to interpret the location and orientation of the sections. In C. viridanus, the teeth are located in a dental groove, a depression of the jaw bone bordered by a labial and a lingual wall, the former being higher than the latter (Fig. 1a–d). The teeth are pleurodont in their mode of attachment to the bone, i.e. they are ankylosed to the inner side of the labial wall of the jawbone. During ontogeny, the shape of the dental groove and that of the teeth

changes. As a consequence, the relationships of the teeth to the supporting bone also change. In adults, the dental groove is deeper and narrower and the labial wall is higher than in juveniles. More than half of the tooth height protrudes above the labial wall margin in juveniles versus only a third or a quarter in adults (compare Fig. 1a, d). On each jaw quadrant, there is a gradient in tooth size along the row. The teeth located at the rear are larger than the anterior ones. In 1-month-old specimens, the teeth are on average 450 μm high with a crown diameter of 150 μm, whereas tooth height reaches 1 mm with a crown diameter of 400 μm in 70-month-old adults. Tooth shape is similar in each dental arcade (homodonty) but it varies during ontogeny as the tooth row becomes more regular. In juveniles, teeth are conical and separated on average by 200-μm-wide spaces. The crown is either straight or slightly curved backwards and the broad dentin base is surrounded by bone tissue (Fig. 1a, b). In adults, teeth are cylindrical and separated on average by 100-μm-wide spaces. The width is similar from the top to the base and the curvature of the crown is more pronounced than in juveniles (Fig. 1c, d). The crowns are curved inwards, perpendicular to the contours of the arcade. During ontogeny, the teeth acquire their characteristic shape by means of tooth replacement (Fig. 1e–g). The crown surface (enamel) is smooth in 1-month-old lizards and a single cusp is distinguishable on the lingual side, underlined by a ridge (Fig. 1e). In 17-month-old juveniles, this ridge is reinforced and the lingual crown surface is ornamented with two or three vertical crests (Fig. 1f). The main ridge progressively differentiates into a marked secondary cusp in 20-month-old lizards, whereas the vertical crests become prominent and their number increases at the lingual surface (Fig. 1g). In 1-month-old and 4.5-month-old juveniles, the base of the dentin cone is 300 μm in diameter, twice the crown diameter (Fig. 1a, b, h, i). At this stage, the tooth base is entirely surrounded by bone extending into the space separating successive teeth. In adults, the dentin base and the crown have a similar diameter and bone is less visible around the tooth bases, which are closer to one another than in juveniles (Fig. 1j). A single foramen is observed at the base of the tooth at the lingual side (Fig. 1a–d). These foramina lead into the pulp cavity and permit the passage of blood vessels and, probably, of the nerve supply (see also Fig. 3a). The location of the foramina often coincides with the first resorbed region of the teeth (Fig. 1a–d). Several foramina are also present at the labial side of the dentary bone (see Fig. 3b). Detailed observations of the tooth crown reveal that this region is often subjected to wear (Fig. 1k–m), a feature that is barely visible at low magnification (Fig. 1a–d). Even in 1-month-old juveniles and in all other specimens examined, numerous worn teeth are observed in each arcade, long before their replacement. However, we have not been able to observe the crown surface of teeth in an advanced state of resorption (probably with more pronounced wear) because they are shed during preparation of the jaws for SEM. First, the enamel layer is worn

74

Fig. 1 Scanning electron micrographs of right lower jaw quadrants in selected specimens from a growth series of the scincid lizard, C. viridanus (a, e, h, k at 1 month, b, i at 4.5 months, f at 17 months, c, g at 20 months, d, j, l, m at 70 months). Anterior is left and all micrographs are presented in lingual view. The developing replacement teeth and the functional teeth in an advanced phase of resorption were lost during bleaching of the jaws, leaving some tooth positions open (tooth positions are numbered 1–16 in a, d). a– d Low magnification. Note the change in shape, size and orientation of the teeth during ontogeny. More tooth positions are lacking in young (a, b) than in old (c, d) specimens, indicating that more teeth are in advanced stages of resorption (and hence of replacement) in young lizards (arrow in a foramen located at the lingual side of the

tooth base; note that the resorption process acting in this region enlarges most foramina). e–g High magnification of the crown of newly functional anterior teeth (positions 4 or 5). Note the progressive differentiation during ontogeny of the cusps and of the ornamentations of the enamel covering the lingual side. h–j Base of recently attached teeth. Note the progressive modification of the dentin cone, from conical to cylindrical, during ontogeny. k–m Various aspects of tooth wear located at the tooth tip and occurring as early as 1 month of life (k). Advanced wear is observed in old specimens in which a large part of the enamel is worn, exposing the dentin surface, which is sometimes also eroded (l, m). Bars 500 μm (a–c), 1 mm (d), 25 μm (e), 50 μm (f, g, m), 100 μm (h, i), 200 μm (j), 10 μm (k), 100 μm (l)

75

(or partially broken) at the tip of the cusp at the lingual surface (Fig. 1k) and then the eroded area extends lingually to the whole surface of the crown tip (Fig. 1l, m). In such teeth, the enamel is usually worn enough to expose the dentin surface, which is also partially eroded. Such considerable wear indicates that the enamel layer is three times as thick in adults than in juveniles (65 μm vs 20 μm, respectively, at the tooth tip) and is twice as thick on the labial compared with the lingual side (65 μm vs 35 μm, respectively, at the tooth tip; 65 μm vs 20 μm at the middle of the crown; Fig. 1k–m). In all examined teeth, wear is always limited to the top of the crown and to the lingual side only. The labial side and the walls of the enamel cover are not affected. Wear does not extend to the dentin cone, even in the oldest specimen examined (approximately 6-year-old); only the upper part of the dentin is eroded in the case of extensive wear. Probably,

tooth replacement prevents such wear penetrating more deeply. In C. viridanus, each tooth position has been found to contain at least one developing replacement tooth, irrespective of ontogenetic stage (Delgado et al. 2003). Depending on the developmental stage of the germ and on the age of the lizard, the location of the tooth germ can vary with regard to the functional tooth that it will replace. This is reflected in the manner in which the base of the functional tooth is resorbed (Fig. 2). All tooth germs start their morphogenesis and first differentiation stages at the lingual side. Here, they are protected from injuries by the functional teeth and by the dental groove. In juveniles, the tooth germ is located laterally, i.e. it develops adjacent to the tooth base and slightly posteriorly to the functional tooth. In consequence, resorption starts at the dentin base and is restricted to the dentin surface at the lingual side

Fig. 2 Aspects of tooth resorption in relation to the presence of growing replacement teeth in C. viridanus (a at 1 month, b at 4.5 months, c at 17 months, d at 20 months, e at 70 months). The replacement teeth have been lost during bleaching of the jaw, except in e. The location of the resorption of the dentin cone changes during ontogeny. Dentin is resorbed along the lingual side in young

lizards (a–c) and from below in old specimens (d, e), where the tooth germ enters the pulp cavity. In e, the crown of the replacement tooth is completely formed before the dentin cone develops and before the apex of the predecessor is worn. Evidence of osteoclastic activity is clearly visible on the surface of the tooth base (arrows). Bars 50 μm (a, c), 100 μm (b–d), 200 μm (e)

76

(Fig. 2a, c). The dentin wall is resorbed facing the growing germ and the pulp cavity is opened. As the germ grows, the entire lingual surface of the dentin cone is resorbed and the developing tooth penetrates progressively into the pulp cavity. Most of the labial and upper region of the functional tooth remains unresorbed during this process (Fig. 2b). In subadult and adult lizards the tooth germ is also located laterally but more deeply in the dental groove. In consequence, resorption takes place below the entire

tooth base and the germ penetrates into the pulp cavity during its early development (Fig. 2d, e). As the tooth germ grows, the dentin walls are resorbed from below (Fig. 2e) and the upper part of the functional tooth is shed instead of being entirely resorbed. As observed in Fig. 2e, the crown of the developing tooth is completed long before the dentin cone has started to be formed. No obvious differences in tooth morphology have been observed among sexes, but the small number of mature

77 3 Fig. 3 Right lower jaw of C. viridanus, labial right. Transverse sections (1 μm thick), toluidine blue staining (am ameloblasts, be buccal epithelium, db dentary bone, de dentin, dl dental lamina, do dental organ, en enamel, ide inner dental epithelium, mc medullar cavity, me mesenchyme, n nerve, Ode outer dental epithelium, pc pulp cavity, sr stellate reticulum, tg tooth germ). a, b Firstgeneration tooth, 1 day post-hatching. c–f Various types of tooth initiation in the embryo. g Tooth initiation in an adult. a Section level through the lingual foramen of the functional tooth (arrow capillary blood vessel passing through the foramen and reaching the region of the developing replacement tooth). Note the typical pleurodont attachment of the tooth and the different structure of the dentary bone and the dentin. A canal connects the pulp cavity to the medullar cavity of the bone support. A nerve has penetrated the canal. b Same tooth sectioned at a more anterior level. A large foramen pierces the dentary bone at the labial side of the jaw. It contains a blood vessel and a nerve (arrow) that penetrate the soft tissues of the lip region. Note the primary type of attachment of the tooth (compare g). c Primary type of initiation. The bilayered dental lamina penetrates deep into the mesenchyme close to the forming dentary bone. Inset: Detail of the extremity of the dental lamina interacting with a mesenchymal cell population. d Secondary type of initiation. The dental lamina extends at the lingual side from the upper region of the dental epithelium surrounding a well-formed tooth. e Detail of the extremity of the dental lamina in d, showing the interaction with the mesenchymal cell population. f Detail of the connection between the dental lamina and the tooth. g In the adult, three teeth constitute a tooth family: a functional tooth (1) that has recently attached to the dentary bone support, a replacement tooth (2) that has entered an advanced stage of enamel maturation, and a tooth germ (3) that is developing. Left The extremity of the dental lamina runs deeper in the mesenchyme. These three teeth are linked to one another by a single dental lamina coming from the buccal epithelium (arrows limits or cement lines between the matrix of the attachment bone and that of the supporting dentary bone). Bars 100 μm (a, b, g), 50 μm (c, d), 20 μm (inset, c, f), 10 μm (e)

specimens available does not permit us to be certain of this. Tooth development in C. viridanus Serially sectioned jaws of the growth series of C. viridanus provide various stages of tooth development. For convenience, we have selected developmental stages from various specimens to illustrate the different steps of tooth development, with a special focus on enamel formation. Most data were obtained at the light-microscopical level (Figs. 3, 4, 5, 7), and only details dealing with enamel formation are presented at the transmission electron-microscopical level (Fig. 6). A previous study of the dentition and pattern of tooth replacement in C. viridanus has revealed that the number of tooth positions is complete when the lizard is born and that no new position is acquired during further growth (Delgado et al. 2003). At birth, the young scincid is equipped with functional teeth occupying each of the 16 tooth positions. These first-generation teeth are firmly attached to the bone support in a typical pleurodont manner. Each position shows a well-developed replacement tooth located at the lingual side of the functional tooth (Fig. 3a, b). In the latter, enamel and dentin are well mineralised and the pulp cavity contains numerous mesenchymal cells, capillary blood vessels, nerve endings

and odontoblasts located along the dentin surface where they deposit secondary dentin (Fig. 3a, b). At the lingual side and slightly posteriorly, the base of the dentin shaft of all non-resorbed functional teeth is pierced by a foramen that permits a capillary blood vessel to reach the region of the developing replacement tooth (Fig. 3a, see also Fig. 1a–d). The pulp cavity communicates also with the medullar cavity of the dentary bone, which contains nerve and blood vessels. At the labial side, the dentary bone is sometimes pierced by a foramen that permits nerves and capillary blood vessels to penetrate the lip region from the medullar cavity (Fig. 3b). Tooth initiation The initiation of a first-generation tooth (called primary initiation) at each position occurs only during embryonic life. This initiation starts as a local thickening of the basal layer of the buccal epithelium, called the placode stage (not shown). A bilayered lamellar extension of the basal layer of the buccal epithelium (called dental lamina) subsequently penetrates deeply into the mesenchyme and reaches the surface of a developing bone, which will become the bone support of the future tooth (Fig. 3c). In contrast to primary initiation, secondary initiation is defined as the initiation of every replacement tooth occurring at a given position. This second type of initiation is therefore the main process of tooth initiation, if one takes into account that each tooth position is replaced several times during the lifetime of a lizard. The bilayered dental lamina emerging from the upper region of the dental epithelium (lingual side) of the previous tooth of the same family extends deep into the mesenchyme where a new replacement tooth will be initiated (Fig. 3d). Both types of initiation processes are morphologically similar. The extension of the dental lamina is 10 μm thick and composed of two undifferentiated cell layers derived from the basal layer of the buccal epithelium (Figs. 3c–e). The existence of epithelial–mesenchymal interactions preceding tooth development is suggested by (1) a thickening of the extremity of the dental lamina resulting from an increase in cell volume and (2) an accumulation of mesenchymal cells around this extremity (inset in Fig. 3c, e). At this stage the dental lamina has extended to about 250 μm from the surface of the buccal epithelium and is located approximately 75 μm away from the surface of the developing bone (Fig. 3c). The distal region of the dental lamina and, particularly, its labial (inner) side is involved in the first steps of tooth morphogenesis and differentiation, whereas the lingual (outer) layer apparently is not (Fig. 3d, f). The lingual layer seems to be involved in conserving the continuity of the dental lamina, whereas the cells from which the three layers composing the dental organ will differentiate (outer dental epithelium, stellate reticulum and inner dental epithelium) seem to be derived from the labial layer (Fig. 3d, f). The secondary type of tooth initiation begins after the previous tooth has started its differentiation.

78

Fig. 4 Embryo (stage 39) of C. viridanus. Right lower jaw, labial right. Transverse sections (1 μm thick), toluidine blue staining. Various stages of tooth development (am ameloblasts, be buccal epithelium, db dentary bone, de dentin, dl dental lamina, dp dental papilla, em enamel matrix, ide inner dental epithelium, me mesenchyme, od odontoblasts, ode outer dental epithelium, pd predentin, sr stellate reticulum, tg tooth germ). a The cells located at the extremity of the dental lamina form a cup surrounding the dental papilla cells. b Detail of a showing the region of interaction between the dental papilla and the dental lamina cells. c A bell has formed. The cells of the dental organ and of the dental papilla have started to

differentiate. The first elements of the tooth matrix (predentin) have been deposited (arrowheads). d This tooth is well differentiated. Dentin and enamel matrices have been deposited and have started to mineralise (arrow new short extension of the dental lamina). e Deposition of the tooth tissues continues and the mineralisation extends, while the dental lamina penetrates deeper into the mesenchyme (arrowhead surface of the dentary bone). f Detail of the tooth base showing well-differentiated ameloblasts depositing enamel matrix and odontoblasts depositing predentin matrix. Bars 50 μm (a, e), 20 μm (b–d, f)

In embryos in which the teeth are not yet functional, i.e. they are not attached to the bone support and they do not pierce the surface of the buccal epithelium, there are only two teeth at a single position: a well-differentiated tooth with maturing enamel and a considerable layer of dentin and a recently initiated tooth germ. In contrast, in 1– to 5month-old juveniles and adults, in which the teeth are functional, three teeth can be found on a single position, together constituting a tooth family: the functional tooth (attached to the bone support and piercing the epithelium), a well-differentiated replacement tooth (that stimulates the resorption of the base of the functional tooth) and a tooth germ in which the first tooth tissues are being deposited (Fig. 3g). These three teeth are linked to one another by the dental lamina, which extends deeper into the mesenchyme to initiate a fourth tooth.

mesenchymal cells around the extremity of the dental lamina (Fig. 4a). The dental lamina has formed an extension in a labial direction and is now shaped as a bell surrounding a small group of mesenchymal cells (approximately 30 μm in diameter), which constitutes the dental papilla. Dental papilla cells have condensed and changed from elongated to wide, whereas most of the mesenchymal cells surrounding the papilla remain undifferentiated and organised in concentric layers around the bell-shaped tooth germ (Fig. 4b). In a subsequent stage, the epithelial cells of the bell differentiate into the so-called dental organ (100 μm high, 80 μm wide at its base). The lower edges of the bell extend deeper into the mesenchyme and three layers can now be distinguished: the inner dental epithelium, the stellate reticulum and the outer dental epithelium (Fig. 4c). The inner dental epithelial cells ( or preameloblasts) have differentiated in the region facing the dental papilla, whereas the outer dental epithelial cells remain undifferentiated. Between these two layers, the stellate reticulum is composed of clear and rounded cells separated by large intercellular spaces. The labial side of the bell is thicker than the lingual

Tooth morphogenesis and cytodifferentiation Embryo (stage 39) Epithelial–mesenchymal interactions result in the accumulation of a dense population of

79

side, the ameloblasts being in a further state of differentiation than those at the opposite side (Fig. 4c). The dental papilla is now 40–50 μm in diameter and the cell number has considerably increased. The cells facing the differentiating ameloblasts have differentiated into polarised odontoblasts and the first unmineralised collagenous matrix of the predentin has been deposited. In the next stage of tooth development, the epithelial bell is well organised (now 160 μm high) and enamel matrix has been deposited by 20 μm high, 3 μm wide, polarised ameloblasts, the nuclei of which are located distally. The tooth cusp forms at the labial side of the enamel organ (Fig. 4d). Predentin is deposited by the odontoblasts located in the upper region of the dental papilla. The first deposited dental matrices (enamel and dentin) have mineralised and the mineralisation has proceeded centrifugally (Fig. 4d). In the basal region of the dental organ and from the lingual surface of the outer dental epithelium, the dental lamina branches off again (Fig. 4d). This bilayered dental lamina eventually elongates into the mesenchyme, whereas the tooth continues its growth (Fig. 4e). The enamel matures and maturation reaches the upper region of the tooth, whereas enamel matrix is still deposited at the tooth base by well-differentiated ameloblasts (Fig. 4f). The odontoblasts are now organised into a pseudo-epithelium and actively produce the predentin matrix in the upper region of the tooth (Fig. 4e). Preceding enamel matrix deposition, the predentin deposition extends towards the tooth base, which is now close to the surface of the bone support. However, the tooth is not yet attached (Fig. 4e, f). The cell populations occupying the pulp cavity are not clearly distinct, and neither nerves nor blood vessels are visible at this time. All well-formed teeth examined in this embryo (approximately 15 days before birth) had reached this developmental stage. No teeth were found to be attached to the bone support, whereas all functional teeth were attached in the 1-day-old specimen studied (Fig. 3a, b). Juveniles and adults Except for the attachment to the bony support, the above descriptions are also valid for tooth development in adult, subadult and juvenile specimens. The only characters that vary are related to lizard growth and concern the size and shape of the forming teeth. Below, we particularly focus on the different steps of enamel formation. The first part presents a general overview of amelogenesis at the LM level (Fig. 5) and the second part is a detailed description of selected stages at the TEM level (Figs. 6, 7). The ameloblasts start to deposit enamel matrix after the first elements of the predentin matrix are deposited by the odontoblasts (see Fig. 6). Immature enamel matrix is highly metachromatic on sections stained with toluidine blue (Fig. 5a). The enamel matrix is first deposited at the tooth tip and the enamel layer extends over a short distance on the tooth sides. The secretory ameloblasts are tall, narrow and closely juxtaposed, perpendicular to the tooth surface and their nuclei are located in the distal cytoplasm. Subsequently, the enamel matrix thickens and

is deposited over the whole tooth surface, provided that predentin has previously been deposited (Fig. 5b). As the tooth grows, the aspect of the covering ameloblasts at the tooth tip changes and numerous intercellular spaces appear between these cells compared with the neighbouring ameloblasts, which remain closely juxtaposed. These changes are concomitant with the initiation of the maturation process in the enamel matrix (Fig. 5c). At the tooth tip, the enamel matrix maturation starts where the dentin matrix has been previously mineralised. Maturation next progresses towards the base of the tooth (Fig. 5e). Enamel maturation starts at the top, extends to the base and results in an enamel layer in which no organic matrix remains visible after decalcification (Fig. 5f, g). At the tip of the tooth only, some sparse dark granules persist in the enamel matrix and finally disappear when enamel maturation is completed. In the meantime, enamel matrix continues to be deposited at the tooth base where it matures late, viz., when the tooth is nearly attached (Fig. 5f, g). At the ultrastructural level, prior to enamel matrix deposition, the epithelial and mesenchymal cell populations, i.e. ameloblasts and odontoblasts, face one another. They are polarised cells separated by an uninterrupted basal lamina (Fig. 6a). Along this epithelial–mesenchymal interface, the ameloblast surface is regular and smooth, whereas numerous cytoplasmic prolongations emanating from the odontoblasts reach the basal lamina. At this stage, the ameloblasts are juxtaposed and elongated and their cytoplasm features do not indicate active protein synthesis. Their large nucleus is located in the mid-region of the cell and the cytoplasm contains a small number of mitochondria, cisternae of the rough endoplasmic reticulum (RER) and vesicles (Fig. 6a). In contrast, the cytoplasm of the facing odontoblasts contains numerous RER cisternae and mitochondria, indicative of active protein synthesis. By this time, a small amount of collagenous matrix has been deposited in the extracellular space separating the odontoblasts from the ameloblasts. Slightly later, the space between both cell populations has enlarged and a thick layer of collagenous matrix has been deposited by the odontoblasts (Fig. 6b). This unmineralised predentin matrix contains numerous cytoplasmic prolongations from the odontoblasts. The latter show intense activity as indicated by the large numbers of mitochondria and RER cisternae. Opposite the layer of predentin, the ameloblasts have not changed, except for an increase of the amount of cytoplasmic organelles (Fig. 6b). A basal lamina is still visible lining the ameloblast surface. The unmineralised predentin layer adjoining the epithelial–mesenchymal interface, i.e. the earliest deposited collagenous matrix, is different from the predentin elsewhere. Indeed, numerous collagen fibrils are oriented perpendicular to the ameloblast surface and reach the basal lamina. This feature is clearly seen at the next stage, in which the predentin matrix has been impregnated with electron-dense non-collagenous proteins and in which the basal lamina has disappeared (Fig. 6c). The cytoplasm of the ameloblasts facing the predentin matrix now contains a

80

large number of RER cisternae, mitochondria and vesicles with electron-dense material, all features indicative of highly active protein synthesis (Fig. 6c). In addition, the

cell membrane of the ameloblasts is no longer smooth and regular. Extracellular spaces have formed along the predentin surface. These spaces are filled with an

81 3 Fig. 5 Juvenile (5 months old) and adult (70 months old) C. viridanus. Transverse sections (1 μm thick), toluidine blue staining. Right lower jaw, labial right. EDTA-decalcified samples. Various stages of differentiation of replacement teeth, with particular attention to enamel formation (am ameloblasts, db dentary bone, de dentin, dl dental lamina, dp dental papilla, em enamel matrix, en enamel, ide inner dental epithelium, od odontoblasts, ode outer dental epithelium, pd predentin, sr stellate reticulum). a Early deposition of the tooth tissues. The enamel matrix is highly metachromatic. The limit between the dentin and the predentin is visible as a dense line. b Extension of the surface covered by enamel matrix (arrow). The arrowhead indicates the delineation between dentin and predentin. c The enamel matrix is deposited at the dentin surface along the tooth sides. The cusps are visible at the apex of the tooth where the maturation process has started (arrow). d A thick layer of enamel matrix has been deposited. The cusps are readily seen and enamel maturation is only visible in the mid region of the upper part of the enamel (arrow). Note the absence of ameloblast Tomes’ processes penetrating the enamel matrix. e Enamel maturation first spreads in the upper region of the tooth. Here, the enamel matrix is progressively removed from the region close to the dentin outwards. f Section through the upper region of a tooth. Enamel maturation is well advanced in this region of the tooth, proceeding from apex to base where the matrix is still being deposited. In the maturing region, the remaining enamel matrix is only visible as small dark granules. g Section through the medial region of a tooth. Enamel maturation reaches the tooth base where enamel matrix is, however, still being deposited. Note the numerous resorption lacunae on the surface of the dentary bone facing the developing tooth (arrowheads). Bars 20 μm

electron-lucent matrix, barely recognisable as collagenous material (Fig. 6d). By the next stage, intercellular spaces have appeared between adjacent ameloblasts, whereas their nuclei are now localised in the distal part of the cell. The cytoplasm bordering the predentin surface is rich in organelles and small vesicles. The ameloblast cell membrane is irregular and a number of short cytoplasmic extensions are visible (Fig. 6e). They penetrate into the space between the ameloblasts and the predentin surface, where they delimit small zones, in which patches of an electron-dense matrix have appeared. These patches represent the first elements of the enamel matrix (Fig. 6e). They are deposited within the collagenous matrix of the upper layer of the predentin. This region will constitute the so-called dentin–enamel junction. The enamel matrix is easily recognisable from the predentin matrix by its electron-dense aspect and the absence of collagen fibrils (Fig. 6e). The enamel matrix deposits then enlarge within the extracellular space facing the ameloblast surface. They form large dense fibrillar patches, with the fibrils arranged parallel to one another and oriented perpendicularly to the ameloblast surface. These fibrils are embedded within a thin granular background substance (Fig. 6f). In a later stage, the patches enlarge, fuse to each other and constitute a homogeneous layer of enamel matrix (Fig. 6g). This electron-dense layer thickens because of the activity of the bordering ameloblasts. As the enamel layer thickens, their surface becomes more and more regular. Dark lines, parallel to the cell surface, appear within the enamel matrix and are suggestive of a periodic growth of this layer (Fig. 6g). Meanwhile, the highly active odontoblasts have deposited a thick layer of predentin,

the upper part of which has been impregnated by noncollagenous electron-dense matrix, indicating the onset of the mineralisation process. At the level of the cusps, the enamel organ is organised differently from that in the other regions. In particular, the orientation of the ameloblasts differs on both sides of the cusp surface where both layers are perpendicular (Fig. 7a). The ameloblasts that deposit the enamel matrix are rich in RER cisternae, mitochondria and Golgi apparatus, from which numerous small vesicles arise (Fig. 7b). Some of these vesicles are seen fusing with the smooth cell membrane facing the enamel matrix. The deposition of the enamel matrix is completed in thickness before the start of its maturation at the dentin–enamel junction. The latter is irregular and composed of small patches of enamel matrix randomly dispersed within the collagenous matrix (Fig. 7c). As enamel maturation continues, it spreads throughout the whole thickness of the enamel layer, but only where enamel was first deposited. Enamel maturation progresses from this region in the upper part of the tooth towards the more recently deposited matrix, along the tooth base, so that there is a gradient in enamel maturation from the tooth tip to base. On our decalcified samples the process of enamel matrix maturation can be followed indirectly as a function of the degree of degradation (by proteolysis) of the enamel matrix. The more the enamel matrix is removed, the more maturation has advanced. When maturation starts, the cytoplasm of the ameloblasts facing the mineralising region is no different from that at the previous stage, except for the presence of small vacuoles (Fig. 7a, d). Clear spaces appear among the fibrillar material of the enamel matrix, throughout the thickness of the layer, near the surface (Fig. 7d) and near the dentin– enamel junction (Fig. 7e). As enamel maturation progresses, the enamel organic matrix disappears progressively, but with a little delay at the level of the growth lines (Fig. 7f). The cytoplasmic content of the ameloblasts bordering the maturing enamel has not changed, except for the presence of large vacuoles along the enamel surface and for the irregular aspect of the cell membrane of the ameloblasts (Fig. 7g). At a latter stage of maturation, the number of organelles is reduced and the intercellular spaces in which numerous interdigitations are visible have expanded (Fig. 7h). The enamel matrix has largely disappeared and the remaining matrix has lost its fibrillar aspect. The organic material now appears globular and a thin basal lamina has formed at the enamel surface, underlining the ameloblast cell membrane (Fig. 7i). At the end of enamel maturation, the organic matrix has completely disappeared. The bordering ameloblasts are interdigitated and their cytoplasm is poor in organelles and electron-dense because of their wealth of filaments (Fig. 7j). After decalcification, the dentin– enamel junction conserves the traces of the interaction between the enamel and dentin matrices in the form of a thin layer of small electron-dense granules (Fig. 7k). The presence of odontoblast prolongations within the enamel layer indicates that, in some regions, enamel proteins have

82

83 3 Fig. 6 One-month-old C. viridanus, left lower jaw, EDTAdecalcified. Aspects of enamel matrix deposition at the electronmicroscopical level (am ameloblasts, em enamel matrix, od odontoblasts, pd predentin). a Interaction between odontoblasts (polarised cells left) and ameloblasts (polarised cells right). Note the numerous cytoplasmic extensions of the odontoblasts. They reach the basal lamina lining the layer of ameloblasts (arrowheads). b Deposition of predentin by the odontoblasts. The basal lamina is still present (arrowheads). c The basal lamina has disappeared and the ameloblasts are now in close contact with the predentin surface. Note the collagen fibrils oriented perpendicular to the cell surface (arrows). d An extracellular space containing a barely distinguishable matrix has been created between the ameloblast and the predentin surface (arrowheads remnants of the basal lamina). e First stages of enamel matrix deposition. Small patches of enamel matrix are deposited on the predentin surface (arrowheads) and subsequently thicken (arrow). f Detail of a large patch of recently deposited enamel matrix. g Low magnification showing (right to left) a layer of immature enamel matrix, a predentin layer impregnated with non-collagenous background substances and the recently deposited predentin matrix. Note the absence of ameloblast Tomes’ processes penetrating the enamel layer. Bars 2 μm (a, b), 1 μm (c, e), 500 nm (d, f), 5 μm (g)

impregnated the upper dentin layer over a larger thickness; this part of the dentin–enamel junction is completely removed by proteolysis (Fig. 7j, k). From all the stages of tooth development and of enamel formation examined in C. viridanus, it is clear that (1) ameloblast Tomes’ processes, similar to those described during mammalian enamel formation, are lacking, and (2) enamel prisms, typical of mammalian enamel, do not form. Resorption In C. viridanus, continuous tooth replacement throughout life implies that functional teeth are resorbed and their remains shed prior to attachment and eruption of the new tooth. Indeed, within a tooth family, the replacement tooth will occupy exactly the same position as the functional tooth. Since the tooth germ develops near the base of the functional tooth, its growth provokes the progressive resorption of the latter. The resorption starts at the lingual side (i.e. facing the developing tooth) close to the tooth base (Fig. 8a). At this stage, the replacement-tooth germ is well formed. The enamel layer is maturing and the surface of the dental organ adjoins the base of the dentin cone of the functional tooth. Clastic cells are first observed at the dentin surface of the latter in the region facing the dental organ of the replacement tooth. We call odontoclasts “multinucleated resorbing cells” as they erode the dentin cone, whatever their location (on the outer tooth surface or inside the pulp cavity); osteoclast cells have the same morphology but resorb the surface of the bone support. First, the odontoclasts are seen along the outer surface of the lingual side of the dentin cone of the functional tooth. They remove the dentin matrix from outside until the dentin cone is pierced (Fig. 8b). Odontoclasts next penetrate the pulp cavity and resorb the dentin wall from inside, whereas other odontoclasts continue to remove the dentin matrix at the outside (Fig. 8c). At this time, the

replacement tooth has completed its maturation process and is now located for a large part within the pulp cavity of the functional tooth. The dentin walls of the latter are now extremely reduced because of clastic resorption. Shortly before eruption of the replacement tooth, the labial wall of the dentin is also attacked by odontoclasts (Fig. 8d). We have never observed odontoclasts resorbing the enamel layer. Hence, this is probably the only part of the functional tooth that is shed. During growth of the replacement tooth, osteoclasts are frequently observed acting along the dentary bone surface, in a region that corresponds to the region at which the new tooth will attach.

Discussion To our knowledge, the present work is the first to provide details concerning tooth development in a lizard, particularly with respect to amelogenesis, both at the light- and electron-microscopical level. Breeding squamates in captivity is not easy and the ability to obtain a growth series of C. viridanus from a laboratory-bred population has obviously been a great advantage for this study, especially because we have been able to investigate specimens of known age. Tooth morphology In the lizard C. viridanus, tooth shape, size and orientation change during ontogeny. Starting from a generalised shape (i.e. small straight conical smooth teeth with large spaces between them) in recently hatched lizards, the teeth in premature juveniles and in adults acquire a morphology specific to the species (i.e. large, posteriorly curved, rectangular, cuspidate and close-set). All of these important modifications would not be possible in the absence of continuous tooth replacement. In young lizards, firstgeneration teeth (i.e. the first teeth to be functional in ontogeny) differ from adult teeth in their external morphology, although they are similar in structure so that they appear as miniatures of adult teeth. This contrasts with the first-generation teeth in actinopterygian and urodele amphibian larvae, in which first-generation teeth differ structurally from their adult counterparts (i.e. in having a thin enameloid cover, dentin without embedded odontoblastic processes and a pulp cavity devoid of blood vessels and of nerve endings). These features are related to the extremely small size of the teeth in the larvae compared with the adults (Sire et al. 2002). Our study clearly shows that the morphological changes observed during ontogeny in C. viridanus are acquired progressively and not suddenly from one generation to the other. Such progressive modifications are, for instance, obvious in the formation of the ornamentations (cusps and ridges) and in the change of shape from conical to cylindrical. Unfortunately, our growth series has not allowed the monitoring of successive tooth generations

84

85 3 Fig. 7 Same specimen of C. viridanus as in Fig. 6. Some aspects of enamel matrix maturation at the electron-microscopical level (am ameloblasts, de dentin, em enamel matrix, en enamel, pd predentin). a Section through the cusp (arrow). Enamel maturation has started in the upper region of the tooth (asterisk). Note the different orientation of the ameloblasts on both sides of the cusp. b Cytoplasmic content of an ameloblast at the onset of enamel matrix maturation. c Aspect of the dentin–enamel junction at the onset of maturation. Note the presence of small patches of enamel matrix among the collagen fibrils of the predentin. d Aspect of the cytoplasm of an ameloblast in a region in which the maturation of the enamel matrix is progressing. e Detail of the enamel matrix close to the dentin–enamel junction at which maturation has started. f Enamel matrix in a region subjected to maturation. Note the presence of horizontal dark lines within the maturing enamel matrix. g Cytoplasm of an ameloblast facing the maturing enamel matrix. Note the large vacuole with a thin granular content. h Features of an ameloblast in a region in which the enamel maturation is nearly complete. Note the large number of interdigitations between the ameloblasts. i Remains of enamel matrix by the end of the maturation process. Note the presence of a thin basal lamina forming at the interface between the enamel and the ameloblasts (arrow). j Recently matured enamel (large empty space) at the level of the cusp (arrow, compare a). k Aspect of the dentin–enamel junction in a region in which enamel maturation is complete. Note the presence of odontoblast extensions at the dentin–enamel junction (long arrow) and within the enamel layer (short arrow). Bars 5 μm (a, j), 500 nm (b, c, f, g, i), 1 μm (d, e, h, k)

within a tooth family; this was especially true of the young stages in which teeth are frequently replaced. Indeed, the functional period of a tooth is short in young reptiles (1– 1.5 months; Cooper 1966) but increases with age (from several months to more than 1 year; Edmund 1960, 1962; Cooper 1966). Moreover, in C. viridanus, the number of tooth positions does not change during ontogeny, the larger size of the adult teeth being accommodated by the change in their shape from conical to cylindrical and by the reduction in the space between adjacent positions (Delgado et al. 2003). This may, however, not be the general condition in lizards and should be checked in another species: for example, is a tooth developing at a new position in a 5-month-old specimen, morphologically similar to functional teeth at 5 months or to firstgeneration teeth? Bleaching the jaw to remove the soft tissues generally stimulates the shedding of most tooth germs. Fortunately, one young replacement tooth remained in place in one of our specimens (a 6-year-old adult). We observed that the tooth cap was finished (i.e. well mineralised), whereas the dentin cone and the tooth base were absent, probably because they were not as yet mineralised and were removed by bleaching. In addition, the upper part of the tooth (enamel covering) was identical in shape, size and ornamentation to that of the functional tooth to be replaced. This indicates that, in this 6-year-old adult, (1) the tooth no longer changes its morphology when replaced at this age, and (2) that the tooth cap is completed long before the replacement tooth is attached. We suggest that the main morphological changes that involve the enamel cap (cusps and ornamentation) should be monitored during ontogeny by comparing the aspect of the tooth surface in the functional tooth with that of its successors in numerous tooth positions. This should provide information on the

time at which the tooth cap is completed and could lead, by means of histological and gene expression studies conducted at appropriate stages of the morphogenesis of the tooth tip, to the identification of the factors that determine the formation of tooth ornamentation, particularly the cusps. In C. viridanus, as in all reptiles examined so far (Sire et al. 2002), only one functional tooth is present at a given position and the replacement tooth does not attach until the functional tooth is shed. This is not the case in the larvae and juveniles of the zebrafish, Danio rerio, in which replacement teeth attach close to the functional teeth and do not stimulate their shedding, in contrast to the adult condition (Van der Heyden and Huysseune 2000). Except for a slight variation in size, the morphology of the teeth is the same for all teeth in a lizard jaw quadrant (i.e. the teeth are homodont). Homodonty is the rule in most lizards (Edmund 1960). Numerous teeth were found to be worn in our specimens. Under our breeding conditions, such wear cannot be provoked by the processing of the provided food itself (soft food: tomatoes and maggots) but could be explained by the grains of sand that are often glued to the food (the bottom of the tanks is sandy). However, tooth wear is not extensive, probably because the teeth do not occlude. This is the general condition in squamates, except in some agamids in which the teeth occlude and are subjected to strong erosion (Cooper et al. 1970). The mode of tooth attachment in C. viridanus is pleurodont (i.e. the tooth base is ankylosed to the inner side of the labial wall). This is the general condition observed in reptiles (e.g. Edmund 1960), with however the exception of a few acrodont species, such as agamids, and the thecodont crocodilians. During ontogeny, the mode of tooth attachment does not change in C. viridanus but the relationships with the dental groove and the replacement teeth appear to vary a little; the replacement tooth germs form more deeply in adults than in juveniles stimulating tooth resorption from the base rather than from the side. Osteoclasts and odontoclasts were observed to resorb the bone and dentin surfaces facing the growing tooth germ. These cells were termed according to the tissue that they were resorbing but not with regard to their origin, which seems to be the same. Osteoclasts are involved in preparing the bone surface for the attachment of a new tooth. Odontoclasts are involved in the resorption of the dentin walls of the functional tooth, thus creating space for the growth of the replacement tooth germ. The process of tooth replacement does not appear to depend on the state of wear or on the durability of the functional tooth. The initiation of a new tooth in a given family seems to be controlled by an internal clock, the timing of which extends during the life span of the lizard. This “automatic” tooth replacement is in contrast with the observations reported by Huysseune (2000) in Astatoreochromis niloticus. When fed hard food (molluscs), the teeth of this cichlid fish show increasing wear and a relationship has been established between the degree of wear and the speed of tooth replacement.

86

Fig. 8 Adult (70 months old) C. viridanus, EDTA decalcified sample. Transverse sections (1 μm thick), toluidine blue staining. Resorption of a functional tooth in relation to various developmental stages of the replacement tooth (be buccal epithelium, db dentary bone, de dentin, dl dental lamina, do dental organ, en enamel, pc pulp cavity, tg tooth germ). Left lower jaw, labial right. a The enamel matrix of this replacement tooth is partially mineralised but the region is not visible at this section level. The basal region of the dentin cone is resorbed on its lingual side, which is in contact with the dental organ of the growing tooth (arrow osteoclast resorbing the surface of the dentary bone). b Mineralisation of the enamel matrix

of the replacement tooth is nearly complete. A large part of the dentin cone of the functional tooth has been resorbed and the pulp cavity has been opened (arrow odontoclast). c The replacement tooth, in which the enamel is mature, has stimulated the resorption of most of the dentin cone at the lingual side of the functional tooth and of a large part of the dentin surface from inside the pulp cavity (arrowheads resorption lacunae). The pulp cavity is largely opened. d The replacement tooth is close to attaching to the bone support. The functional tooth is entirely resorbed except for part of the labial side of the dentin cone, which is still attached to the dentary bone surface (arrow odontoclast). Bars 20 μm (a), 50 μm b–d

Tooth development

opment has been investigated only recently in our laboratory. Sire et al. (2002) has studied the first-generation teeth in C. viridanus and C. sexlineatus, the dentition and replacement pattern of the teeth having been examined by Delgado et al. (2003). In C. viridanus, the first teeth develop early in embryos but most of them are degenerative and never become functional (Sire et al. 2002). These degenerative (rudimentary) teeth are replaced by functional teeth before birth or soon thereafter, as described for other reptiles (e.g. Röse 1894; Woerdeman 1919; Edmund 1969; Osborn 1971; Westergaard 1986; Westergaard and Ferguson 1986). The replacement tooth pattern in C.

The histological development of reptilian dentition has been previously reported in several species: in Lacerta vivipara by Osborn (1971), in Sphenodon punctatus, Cnemaspis kandiana (a gekkonid), Lacerta agilis and Anguis fragilis (slow worm) by Westergaard (1986) and in Alligator mississipiensis by Westergaard and Ferguson (1986). However, these studies provide only few comparative data and only old studies can be used as a comparison with the present study (e.g. Röse 1894; Woerdeman 1919; Poole 1957). In scincids, tooth devel-

87

viridanus is predictable (Delgado et al. 2003), a feature used in the present study to search for representative stages of tooth development in the jaws of a growth series of C. viridanus and, in particular, to study various stages of amelogenesis. Initiation In crocodiles and in various lepidosaurian species, the first (degenerative) teeth are initiated directly from the oral epithelium before the dental lamina is formed (Westergaard 1986; Westergaard and Ferguson 1986). In Alligator mississipiensis, Westergaard and Ferguson (1986) have shown that the dentition starts by the initiation of one tooth long before other teeth appear. The early initiation stimulus may originate in the epithelium (called placode stage by Westergaard and Ferguson 1986) followed by the local induction of the mesenchyme, which further condenses into a dental papilla. Unfortunately, the only embryonic stage available in C. viridanus was too old (2 weeks before birth) to observe the placode stage of the first-generation teeth. Degenerative teeth were not seen, probably because they had disappeared (i.e. been resorbed) at this embryonic stage. The first functional teeth were initiated at the extremity of a narrow straight dental lamina branching directly off the buccal epithelium. The first replacement teeth were formed in connection with the dental lamina. This is in accordance with previous findings claiming that all functional teeth in reptiles develop from a well-formed dental lamina (Westergaard 1986). As soon as a tooth starts differentiating in C. viridanus, a new dental lamina extends into the mesenchyme from the upper region of its dental organ and a new tooth is initiated. This has been observed in the embryo and in juveniles and adults. It signifies that when a new replacement tooth starts to be initiated the former tooth is not as yet functional, i.e. neither erupted nor attached to the bone support. This raises the question of the identity of the factors that control the initiation of tooth replacement and may also explain the replacement waves that are welldocumented in the literature (Edmund 1960, 1969; Osborn 1970, 1984). One hypothesis to explain the initiation of a new replacement tooth could be that the new extension of the dental lamina from the outer dental epithelium is related to the state of differentiation of the previous tooth in the family. This hypothesis has to be tested by comparing the differentiation stages of numerous, serially sectioned, developing teeth at different positions and in a growth series. In humans with their diphyodont dentition, in contrast to the polyphyodont condition in squamates, the replacement of deciduous teeth occurs in a similar way. The permanent tooth is initiated and differentiates during the odontogenesis of the deciduous tooth (Ruch 2001). The dental lamina of the replacement tooth extends, at the lingual side, from the surface of the well-differentiated enamel organ. In contrast, the initiation of the molars, which have no deciduous precursors, occurs directly from

a dental lamina emerging from the buccal epithelium (Ruch 2001), as shown for the first-generation teeth in reptiles. Except for the embryo and 1-day-old C. viridanus, in which only one or two teeth were present at a given position, three teeth from a single family were always present in all specimens studied, the third tooth being at the bell stage with the predentin and enamel matrix having been deposited but not yet mineralised. All teeth of a single family were connected to the dental lamina at the level of its outer dental epithelium. In addition, the dental lamina extended from the outer dental epithelium of the third and youngest tooth suggesting that the initiation of a fourth tooth had started. This finding confirms the hypothesis proposed by Delgado et al. (2003) regarding the probable existence of a third, not yet mineralised, tooth germ in addition to the two teeth visible in radiographs. Cytodifferentiation and matrix formation: similarities to and differences from mammalian teeth In C. viridanus, first-generation and replacement teeth invariably develop as described in other squamate species: epithelial cells of the dental lamina and a mesenchymal cell population come into contact and differentiate into a dental organ and a dental papilla, respectively, at a short distance from the bone support surface (Osborn 1971; Ogawa 1977). We have repeatedly observed that the labial side of the dental lamina is involved in tooth differentiation, whereas its lingual side is involved in the initiation of the replacement tooth. In C. viridanus, the dental organ differentiates from the proliferating zone located at the extremity of the dental lamina as described in mammals (Ruch 2001). This organ is composed of three layers: the inner (ameloblast layer) and outer dental epithelia, which are separated by a thin layer that is interpreted here as the stellate reticulum (Ogawa 1977). This organisation differs from that observed in mammalian teeth, in which a fourth layer, called the stratum intermedium, has been described (Ruch 2001). The following differences between C. viridanus and mammalian tooth structure and development are noteworthy. (1) The pleurodont mode of attachment in the lizard, characterised by the tooth base being ankylosed to the bone support, differs from the mammalian thecodont mode, in which a periodontal ligament links the tooth to the bone support. Various modes of tooth attachment, including the presence of developed pedicels in fish and lissamphibians, have been described in non-mammalian lineages (Shellis 1982; Huysseune and Sire 1998; Trapani 2001). However, a thecodont attachment has been reported only in the crocodiles, which, unlike mammals, replace their teeth continuously (Poole 1967; Edmund 1969; Berkovitz and Sloan 1979). (2) Tomes’ processes are absent in the lizard. Such ameloblast cytoplasmic extensions, on the contrary, are a strong characteristic of mammalian enamel formation (Nanci and Goldberg 2001). This is, however, the only clear difference between lizard

88

and mammal ameloblasts. Indeed, in C. viridanus, the ameloblasts, although not producing Tomes’ processes, show the same cellular changes as reported during mammalian amelogenesis, with a secretory and a maturing phase (Zaki and Macrae 1978). (3) Enamel prisms are also absent in the lizard. In non-mammalian species, prismless enamel has been related to the lack of Tomes’ processes and, probably, to the lack of stepped ameloblast alignment (Grine et al. 1979). This particular structure of enamel without prisms has been previously reported for most squamate teeth (see Sander 2001 for a detailed review of the crystal organisation in reptile enamel). In C. viridanus, the predentin matrix starts to be deposited by the odontoblasts (and probably start to mineralise) before the immature enamel matrix is deposited by the ameloblasts. In this lizard, the enamel matrix therefore appears to be deposited on the dentine matrix with a slight delay compared with that in mammals (Goldberg et al. 2001). Contrary to mammalian enamel, which starts to mineralise quickly after the first deposition of the organic matrix, the deposition of the immature enamel matrix in C. viridanus is completed prior to maturation. The end of the active role of the ameloblasts in enamel matrix deposition and the initiation of maturation are thus probably related and, hence, ameloblasts may play a role in the initiation of this process. Nevertheless, we cannot reject the possibility that some mineral crystals could be present in the immature enamel matrix; appropriate techniques (e.g. Von Kossa staining) should therefore be used to detect the presence of calcium in these teeth but this was not within the scope of our study. Such a delay in mineralisation has also been described in tooth enameloid of teleost fish (Prostak and Skobe 1986) but not in shark teeth (Risnes 1989). Otherwise, the maturation process of C. viridanus enamel proceeds as reported for mammals and for other taxa, such as lissamphibians and crocodiles (Yamashita and Ichijo 1983): it starts above the mineralised dentin surface, spreads towards the surface and then extends to both sides. The first-differentiated odontoblasts of C. viridanus are polarised, with prominent cytoplasmic extensions that reach the ameloblast surface, whereas they are nonpolarised and devoid of extensions in mammals. Therefore, a typical mantle dentin is lacking in the lizard tooth. As a consequence of odontoblast polarisation, most of the first-deposited collagen fibrils are oriented perpendicular to the ameloblast surface. This differs from mammalian mantle dentin in which the fibrils are not similarly oriented. In the lizard, this collagen fibril arrangement at the dentin surface allows the enamel proteins to penetrate deeply into the outer layer of the predentin matrix. Such a feature is interpreted, in the lizard, as an improvement in the strength of the dentin–enamel junction. The interpenetration of both matrices at the dentin–enamel junction has also been described in a basal actinopterygian fish (Sire et al. 1987). In C. viridanus, a number of cytoplasmic extensions of the odontoblasts do not retract when dentin mineralises, so that dentinal canaliculi are visible at the level of the dentin–enamel junction. This is not the case in

mammals, in which the mantle dentin is devoid of canaliculi (Linde and Goldberg 1993). In conclusion, except for the few differences indicated above, the processes involved during tooth development in the lizard C. viridanus are similar to those reported for mammalian tooth morphogenesis and cytodifferentiation. Apart from the absence of Tomes’ processes and prisms, the cellular changes during lizard amelogenesis are basically similar to those reported in mammals. Therefore, this squamate species, which has the advantage of being polyphyodont and easy to bred in the laboratory, is a favourable model for further investigations of amelogenesis. Acknowledgements We are grateful to Prof. Ann Huysseune (Ghent University) for numerous comments and suggestions on the manuscript. We also thank two referees for their constructive remarks. Prof. Jacques Castanet (Université Pierre et Marie Curie) is acknowledged for the gift of embryos, juveniles and adults of C. viridanus. SEM and TEM was carried out at the “Service de Microscopie électronique-Université Paris 6 et CNRS”.

References Berkovitz BKB, Sloan P (1979) Attachment tissue of the teeth in Caiman sclerops (Crocodilia). J Zool Lond 187:179–194 Cooper JS (1966) Tooth replacement in the slow worm (Anguis fragilis). J Zool Lond 150:235–248 Cooper JS, Poole DFG, Lawson R (1970) The dentition of agamid lizards with special reference to tooth replacement. J Zool Lond 162:85–98 Delgado S, Davit-Beal T, Sire J-Y (2003) The dentition and tooth replacement pattern in Chalcides (Squamata; Scincidae). J Morphol 256:146–159 DeMar RE (1972) Evolutionary implications of Zahnreihen. Evolution 26:435–450 DeMar RE (1974) On the reality of Zahnreihen and the nature of reality in morphological studies. Evolution 28:328–330 Dufaure JP, Hubert J (1961) Table de développement du lézard vivipare: Lacerta (Zootoca) vivipara Jacquin. Arch Anat Microsc 50:309–327 Edmund AG (1960) Tooth replacement phenomena in the lower vertebrates. Contrib R Ont Mus Life Sci Div 52:1–190 Edmund AG (1962) Sequence and rate of tooth replacement in the Crocodilia. Contrib R Ont Mus Life Sci Div 56:1–42 Edmund AG (1969) Dentition. In: Gans C, Bellairs Ad’A, Parsons TS (eds) Biology of reptilia, vol I. Academic Press, London, pp 117–200 Goldberg M, Farges J-C, Magloire H (2001) Structure des dents (dentines). In: Piette E, Goldberg M (eds) La dent normale et pathologique. DeBoeck Université, Bruxelles, pp 55–72 Grine FE, Vrba ES, Cruickshank ARI (1979) Enamel prisms and diphyodonty: linked apomorphies of Mammalia. S Afr J Sci 75:114–120 Harrison HS (1901) The development and succession of teeth in Hatteria punctata. Q J Microsc Sci 44:161–219 Huysseune A (2000) Developmental plasticity in the dentition of a heterodont polyphyodont fish species. In: Teaford MF, Smith MM, Ferguson MWS (eds) Development, function and evolution of teeth. Cambridge University Press, Cambridge, pp 231– 241 Huysseune A, Sire J-Y (1998) Evolution of patterns and processes in teeth and tooth-related tissues in non-mammalian vertebrates. Eur J Oral Sci 106 (Suppl 1):437–481 Linde A, Goldberg M (1993) Dentinogenesis. Crit Rev Oral Biol Med 4:679–728

89 Nanci A, Goldberg M (2001) Structure des dents (émail). In: Piette E, Goldberg M (eds) La dent normale et pathologique. DeBoeck Université, Bruxelles, pp 39–54 Ogawa T (1977) A histological study of the gekko tooth. Shigaku Odontol 64:1377–1388 Osborn JW (1970) New approach to Zahnreihen. Nature 225:343– 346 Osborn JW (1971) The ontogeny of tooth succession in Lacerta vivipara Jacquin (1787). Proc R Soc Lond [Biol] 179:261–289 Osborn JW (1972) On the biological improbability of Zahnreihen as embryological units. Evolution 26:601–607 Osborn JW (1984) From reptile to mammals: evolutionary considerations of the dentition with emphasis on tooth attachment. Symp Zool Soc Lond 52:549–574 Osborn JW Crompton AW (1973) The evolution of mammalian from reptilian dentitions. Breviora 399:1–18 Poole DFG (1957) The formation and properties of the organic matrix of reptilian tooth enamel. Q J Microsc Sci 98:349–367 Poole DFG (1967) Phylogeny of tooth tissues: enameloid and enamel in recent vertebrates, with a note on the history of cementum. In: Miles AEW (ed) Structural and chemical organization of teeth. Academic Press, New York, pp 11–149 Prostak K, Skobe Z (1986) Ultrastructure of the dental epithelium during enameloid mineralization in a teleost fish, Cichlasoma cyanoguttatum. Arch Oral Biol 31:73–85 Rieppel O (1978) Tooth replacement in Anguinomorph lizards. Zoomorphologie 91:77–90 Risnes S (1989) Shark tooth morphogenesis. An SEM and EDX analysis of enameloid and dentin development in various shark species. J Biol Buccale 18:237–248 Rocek Z (1980) Intraspecific and ontogenetic variation of the dentition in the green lizard Lacerta viridis (Reptilia, Squamata). Vest Cs Spolec Zool 44:272–278 Röse C (1894) Ueber die Zahnentwicklung der Crocodile. Morphol Arbeit 3:195–228 Ruch JV (2001) Développement dentaire normal. In: Piette E, Goldberg M (eds) La dent normale et pathologique. DeBoeck Université, Bruxelles, pp 5–17

Sander PM (2001) Primless enamel in amniotes: terminology, function and evolution. In: Teaford M, Ferguson MWJ, Smith MM (eds) Development, function and evolution of teeth. Cambridge University Press, New York, pp 92–106 Shellis RP (1982) Comparative anatomy of tooth attachment. In: Berkovitz BKB, Moxham BJ, Newman HN (eds) The periodontal ligament in health and disease. Pergamon, Oxford, pp 3–24 Sire J-Y, Huysseune A (2003) Formation of skeletal and dental tissues in fish: a comparative and evolutionary approach. Biol Rev 78:219–249 Sire J-Y, Géraudie G, Meunier FJ, Zylberberg L (1987) On the origin of the ganoine: histological and ultrastructural data on the experimental regeneration of the scales of Calamoichthys calabaricus (Osteichthyes, Brachyopterygii, Polypteridae). Am J Anat 180:391–402 Sire J-Y, Davit-Beal T, Delgado S, Van Der Heyden C, Huysseune A (2002) The first generation teeth in non-mammalian lineages: evidence for a conserved ancestral character? Microsc Res Techn 59:408–434 Trapani J (2001) Position of developing teeth in teleosts. Copeia 1:35–51 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 Westergaard B (1986) The pattern of embryonic tooth initiation in reptiles. Mém Mus Natn Hist Nat Paris (Série C) 53:55–63 Westergaard B, Ferguson MWJ (1986) Development of the dentition in Alligator mississipiensis. Early embryonic development in the lower jaw. J Zool Lond 210:575–597 Woerdeman MW (1919) Beiträge zur Entwicklungsgeschichte von Zähnen und Gebiss der Reptilien. Beitrag I. Die Anlage und Entwicklung des embryonalen Gebisses als Ganzes und seine Beziehung zur Zahnleiste. Arch Mikrosk Anat 92:104–192 Yamashita Y, Ichijo T (1983) Comparative studies on the structure of the ameloblasts. In: Suga S (ed) Mechanisms of tooth enamel formation. Quintessence, Tokyo, pp 91–107 Zaki AE, Macrae EK (1978) Fine structure of the secretory and nonsecretory ameloblasts in the frog. I. Fine structure of the secretory ameloblasts. Am J Anat 148:161–194