Cell Tissue Res (2005) 321: 375–389 DOI 10.1007/s00441-004-1036-x

REGULAR A RTICLE

C. Van der heyden . F. Allizard . J.-Y. Sire . A. Huysseune

Tooth development in vitro in two teleost fish, the cichlid Hemichromis bimaculatus and the cyprinid Danio rerio Received: 29 June 2004 / Accepted: 2 November 2004 / Published online: 21 June 2005 # Springer-Verlag 2005

Abstract A technique for organotypic in vitro culture with serum-free medium was tested for its appropriateness to mimic normal odontogenesis in the cichlid fish Hemichromis bimaculatus and the zebrafish Danio rerio. Serial semithin sections were observed by light microscopy to collect data on tooth patterning and transmission electron microscopy was used to compare cellular and extracellular features of tooth germs developing in vitro with the situation in vivo. Head explants of H. bimaculatus from 120 h post-fertilization (hPF) to 8.5 days post-fertilization (dPF) and of zebrafish from 45 hPF to 79 hPF and adults kept in culture for 3, 4 or 7 days revealed that tooth germs developed in vitro from explants in which the buccal or pharyngeal epithelium was apparently undifferentiated and, when present at the time of explantation, they continued their development up to a stage of attachment. In addition, the medium allowed the morphogenesis and cytodifferentiation of the tooth germs similar to that observed in vivo and the establishment of a dental pattern (place and order of

tooth appearance and of attachment) that mimicked that in vivo. Organotypic culture in serum-free conditions thus provides us with the means of studying epithelial-mesenchymal interactions during tooth development in teleost fish and of analysing the genetic control of either mandibular or pharyngeal tooth development and replacement in these polyphyodont species. Importantly, it allows heads from embryonically lethal (zebrafish) mutants or from early lethal knockdown experiments to develop beyond the point at which the embryos normally die. Such organotypic culture in serum-free conditions could therefore become a powerful tool in developmental studies and open new perspectives for craniofacial research. Keywords Teeth . Dentition . Serum-free in vitro culture . Zebrafish . Danio rerio . Cichlid . Hemichromis bimaculatus (Teleostei)

Introduction The in vitro infrastructure at the Ghent laboratory was financed through a grant of the “Bijzonder Onderzoeksfonds” of Ghent University (BOF: 01102995) and a “Krediet aan navorsers” (no. 31513695) of the Fonds voor Wetenschappelijk onderzoek (FWO-Vlaanderen). This study also benefitted from an exchange program between the “Centre National de Recherche Scientifique” (CNRS) and the “Ministerie van de Vlaamse Gemeenschap”. Research performed by C. Van der heyden was partly financed through a specialization grant of the Flemish Institute for the Advancement of Scientific-Technological Research in Industry (IWT). C. Van der heyden . A. Huysseune (*) Biology Department, Ghent University, K.L. Ledeganckstraat 35, 9000 Gent, Belgium e-mail: [email protected] Tel.: +32-9-2645229 Fax: +32-9-2645344 F. Allizard . J.-Y. Sire Equipe “Evolution and Development of the Skeleton”, FRE2696, Case 7077, Université Paris 6, 7 Quai St.-Bernard, 75251 Paris Cedex 05, France

Vertebrate teeth are formed by reciprocal interactions between an odontogenic epithelium and the underlying mesenchyme (Stock et al. 1997; Thesleff 2000, 2003). A large number of genes have been shown to be expressed during mammalian tooth morphogenesis and differentiation (see, for example, the on-line data base: http://www.bite-it.helsinki. fi; Nieminen 1998). In mouse, teeth are autonomous units, i.e. presumptive germs cultured in oculo before the placode stage develop and form teeth in isolation from the jaw (Lumsden 1979). Whether this is true of other vertebrate lineages remains to be seen. Mammals are the most frequently used animal models to study aspects of tooth development but they replace their teeth only once (diphyodont) or are even monophyodont (rat, mouse). In contrast, most non-mammalian taxa replace their teeth throughout life (polyphyodont condition). To date, the mechanism by which a replacement tooth is initiated, be it in a diphyodont or polyphyodont species, is not known. The polyphyodont condition such as that found in teleost fish raises several questions. (1) Is the genetic control

376

of tooth development in teleosts similar to that known in mammals? (2) Is this control similar during mandibular and pharyngeal tooth development? (3) Does the frequent replacement that characterizes polyphyodonty represent renewed initiation, involving the same molecular cascades responsible for the initiation of the first-generation teeth, or is replacement the outcome of a separate (or truncated) developmental cascade? In other words, is the epithelium and/or mesenchyme committed to an odontogenic fate for all future generations of teeth once the first-generation teeth have been formed? Some of these questions might be adequately addressed by analysis of odontogenesis under standardized conditions in vitro. Yamada et al. (1980) were the first to describe a chemically defined serum-free medium permitting the morphogenesis and cytodifferentiation of murine molar tooth germs in vitro. Molar tooth germs explanted at the cap stage formed dentin and enamel within 10 days of culture. The authors thus provided evidence that neither serum nor embryonic extract is required for tooth organogenesis to be completed. To our knowledge, only two papers deal with in vitro organ culture systems developed especially for the study of skeletal tissues in fish. Miyake and Hall (1994) have established an in vitro culture method based on a medium containing Leibovitz L15 (L-15) and horse or fetal bovine serum, whereas Koumans and Sire (1996) have developed a serum-free medium. The latter, which has the advantage of being chemically defined, has been demonstrated to be useful for developmental studies of skeletal elements in the cichlid fish, Hemichromis bimaculatus, and is used routinely in our research. The present study aims at testing whether tooth development (including the steps of initiation, morphogenesis, cytodifferentiation and attachment) in teleost fish can be mimicked in vitro by using a serum-free medium. It serves as a baseline study to test, in a later phase, the influence of certain morphogens, e.g. growth factors, on tooth development and replacement. We have used a model organism, the zebrafish (Danio rerio), for which a growing body of data on tooth development is being collected (Huysseune et al. 1998; Van der heyden and Huysseune 2000; Van der heyden et al. 2000, 2001; Wautier et al. 2001). The major disadvantage of the zebrafish as a model organism in tooth studies is that the teeth are restricted to the pharyngeal jaws (in association with the fifth branchial arches). Therefore, we have chosen to perform parallel in vitro culture experiments on another teleost possessing both oral and pharyngeal teeth, viz. the cichlid H. bimaculatus. Like for the zebrafish, the morphological establishment of the tooth pattern in H. bimaculatus is well known (Huysseune 1990; Huysseune and Sire 1992, 1997). Knowledge of the tooth pattern is essential for the setting up and assessment of in vitro cultures aimed at addressing some of the questions raised above. In this study, we have cultured head explants of early postembryonic stages of H. bimaculatus and of D. rerio and explants of the pharyngeal jaws of adult zebrafish and compared these results with the situation in vivo by using

light- and transmission-electron-microscopical (TEM) observations. We show that, starting from explants in which the epithelium is morphologically undifferentiated, tooth germs form in vitro, that odontogenesis proceeds as in vivo, and that teeth attach to the jaw bone by (unmineralized) bone of attachment formed in vitro.

Materials and methods Materials H. bimaculatus larvae from 120, 136, 144 and 150 h postfertilization (hPF) and 8.5 days post-fertilization (dPF; approximately 200 hPF) were obtained from laboratoryreared animals. The eggs were kept at 25°C and hatched at 72 hPF. Larvae were fed Artemia nauplii and commercial powder starting from the free-swimming phase, i.e. at 6 dPF. D. rerio larvae from 45 hPF to 79 hPF were obtained from laboratory-reared animals and raised at 27±0.5°C (hatching occurred at 48 hPF). In addition, adults were used for the explantation of isolated pharyngeal jaws. All specimens were killed in accordance to the Belgian law on the protection of laboratory animals (KB d.d., 14 November 1993) by an overdose of the anaesthetic MS222. Because of the small size of the larval fish (the head length being less than 1 mm), the culture of isolated jaws was impossible. The larvae were decapitated (Fig. 1) and the heads (in the case of Danio) or the anterior part of the head (in the case of Hemichromis) were used as explants; in adults, the pharyngeal jaws were dissected free from the head.

Fig. 1 Larvae of H. bimaculatus (A 6 dPF) and of D. rerio (B 48 hPF) showing the level at which the head was sectioned (dashed line). Arrows indicate the approximate position of the oral (Hemichromis) and pharyngeal (Danio) jaws, respectively. Bars 1 mm (A), 250 μm (B)

377

Control specimens of the same broods were fixed at the time of explantation and at the time of fixation of the explants. The number of specimens used for each type of experiment and the number of controls are given in Tables 1, 2. Methods In vitro cultures were performed according to the procedure described by Koumans and Sire (1996). Head explants and isolated jaws were first washed three times for 3 min in L15 and then twice for 3 min in L-15 containing 300 μg/ml gentamycin (Sigma) and 500 IU/ml nystatin (Sigma; i.e. ten times the normal concentration) in order to prevent fungal and bacterial contaminations. The samples were subsequently rinsed again in prepared L-15 medium (according to Koumans and Sire 1996) and incubated by complete submersion in prepared L-15 in six-well tissue culture plates (Corning, 4 ml medium/well). Plates were incubated at 25–27°C in an incubator with a normal water-saturated atmosphere. Medium was changed every other day. Explants were kept in culture for 3 or 7 days. Explants at the end of the culture period and control specimens were fixed in a mixture of paraformaldehyde and glutaraldehyde and embedded in Epon as described by Huysseune and Sire (1992). The material was not decalcified. All explants and control specimens were serially cross-sectioned (1 μm or 2 μm thick) and the sections were stained with toluidine blue. Some series were interrupted for ultrathin sectioning. Thin (80 nm) sections were contrasted with uranyl acetate and lead citrate and viewed in a Philips 201 EM operating at 80 kV. Table 1 Culture experiments performed on H. bimaculatus and results of the explants in terms of number of tooth germs or teeth (t0 age at explantation, n number of days in culture)

a

Number of head explants placed in culture b Number of control specimens fixed at the time of explantation c Number of tooth germs or teeth observed in the controls at the time of explantation (brackets number of times this number of tooth germs/teeth was found versus total number of hemimandibles examined) d Number of tooth germs or teeth observed in the explants (brackets number of times this number of tooth germs/teeth was found versus total number of hemimandibles examined)

Age at explantation (t0)

3

136 hPF

3

6

7

6

3

3

7

3

3

6

150 hPF

The descriptions concerning the tooth pattern are based on light-microscopic and TEM observations of control series and on previously published data of the oral tooth pattern of H. bimaculatus (Huysseune 1990) and of the pharyngeal tooth pattern of D. rerio (Van der heyden and Huysseune 2000). Between hatching (72 hPF) and 8.5 dPF, the lower jaw dentition of H. bimaculatus develops eight of the nine tooth positions found on each hemimandible (cf. Huysseune 1990). They are numbered 1–8 from the symphysis backwards. The fully developed zebrafish pharyngeal dentition consists of three rostro-caudal tooth rows located on each of the paired fifth ceratobranchials (ceratobranchialia V, the ventral parts of the last branchial arch). During the experimental period (i.e. from 45 hPF to 79 hPF), the larvae develop teeth on the ventral (V) row only. This row normally consists of five tooth positions, which are referred to as positions 1V to 5V (Van der heyden and Huysseune 2000), but only first-generation teeth in positions 3V, 4V and 5V and their respective successors develop prior to the age of 10 dPF. In vitro mandibular tooth development in H. bimaculatus Appearance of new tooth germs The number of mandibular teeth and tooth germs, irrespective of their stage of development, present in controls (i.e. at the time of explantation) and in cultured explants is

No. of days No. of No. of in culture explantsa controls (n) at t0b

120 hPF 3 (hatching at 72 hPF) 7

144 hPF

Results

2

No. of tooth germs/teeth in controls (t0)c

No. of successful explants

No. of tooth germs/ teeth in explants (t0+n)d

1 (4/4)

2

3 5 3 4 6 4 5 6 3 4 6 6 7 5 6 7 8

3

3

2

3 (4/4)

5

3

2

5 (4/4)

3 2

6

5 6 7 8

(3/12) (2/12) (4/12) (3/12)

6

(2/4) (2/4) (2/6) (2/6) (2/6) (1/10) (7/10) (2/10) (1/6) (1/6) (4/6) (5/6) (1/6) (3/4) (1/4) (5/12) (7/12)

378 Table 2 Culture experiments performed on D. rerio and results of the explants in terms of number of tooth germs or teeth present (abbreviations and explanations of column headings as in Table 1; hemimandibles to be replaced by separate pharyngeal jaws)

Age at explantation (t0)

No. of tooth No. of days in No. of No. of No. of culture (n) explants controls at germs/teeth in successful controls (t0) t0 explants

No. of tooth germs/teeth in explants (t0+n)

45 hPF

3

6

3

1 (6/6)

6

48 hPF (hatching) 55 hPF

4

2

2

1 (4/4)

2

2 (2/12) 3 (9/12) 4? (1/12) 3 (4/4)

3

6

7

2 (14/14)

4

7

6

69 hPF

3

6

3

3 (6/6)

6

79 hPF

7 3 7

6 6 6

4

4 (8/8)

3 4 1

6

3 (2/8) 4? (6/8) 1 (1/12) 2 (1/12) 3 (4/12) 4? (6/12) 3 (10/12) 4? (2/12) 3 (6/6) 4 (8/8) 3 (1/2) 4 (1/2)

summarized in Table 1. At least one tooth germ was present on each hemimandible at the time of explantation in all experiments and several teeth formed during the culture period (up to five new germs in explants made at 120 hPF and cultured for 7 days). The pattern of the mandibular dentition obtained in vitro (Fig. 2B, C) resembled the situation in vivo (Fig. 2A), i.e. the first three germs developed in position 4, followed by

positions 2 and 6; the following two germs developed in positions 5 and 7, followed by the germ in position 3 and finally the germs in positions 1 and 8. Similarly, the order of tooth attachment in vitro (Fig. 2B, C) followed the pattern of tooth initiation and coincided with the attachment pattern observed in vivo (compare with Huysseune 1990). First, the tooth in position 4 attached, followed by an almost simultaneous attachment of teeth in positions 2 and

Fig. 2 H. bimaculatus. Representation of the oral dentition in several ontogenetic stages at the time of explantation (At0 control in vivo), after 3 (Bt0+3d) and 7 days (Ct0+7d) of in vitro culture. The explants showing the highest number of teeth are represented. Tooth positions are numbered from rostral to caudal; tooth germs and teeth are represented by dots (white dots initiation and morphogenesis

stages, grey dots cytodifferentiation stage, black dots attached tooth, hPF hours post-fertilization, asterisk after Huysseune 1990). Note that, because of the higher water temperature, the development of the dentition in the current experiments is slightly advanced compared with the data published by Huysseune (1990)

379

6, and so on. There was, however, a delay of approximately 50 h in tooth development in vitro compared with the situation in vivo (compare, for example, the control at 144 hPF with the explant of 120 hPF cultured for 3 days, i.e. aged 192 h; Fig. 2).

cartilage (Fig. 4). After 7 days, a thick layer of dentin had formed (Fig. 5). The germ in position 6, which had formed a visible amount of matrix at the moment of explantation (144 hPF; Table 1, Figs. 2, 6), deposited attachment bone after 3 days in culture (Fig. 7) and was fully attached after 7 days (Fig. 8).

Morphogenesis, differentiation and attachment Ultrastructure of teeth grown in vitro Tooth germs present at the time of explantation and new germs initiated after explantation continued their development in vitro, in a manner resembling the situation in vivo (cf. Huysseune and Sire 1997). For example, in a 144-hPF specimen, the tooth germ in position 3 was not present at the time of explantation (Table 1, Figs. 2A, 3). After 3 days in culture, a germ starting matrix deposition (cytodifferentiation stage) had formed between position 4 and Meckel’s

Fig. 3-8 H. bimaculatus. Light micrographs of the tooth in position 3 (Figs. 3, 4, 5) and in position 6 (Figs. 6, 7, 8), at the time of explantation (t0=144 hPF, Figs. 3, 6), after 3 days in culture (t0 + 3d, Figs. 4, 7), and after 7 days in culture (t0 + 7d, Figs. 5, 8) Fig. 3 At the moment of explantation, a tooth in position 3 is still absent (arrowhead), whereas the tooth at the adjacent position 4 (asterisk) has started matrix deposition (MC Meckel’s cartilage). Bar 10 μm Fig. 4 After 3 days in culture, the tooth in position 3 (arrowhead) has differentiated between the tooth in position 4 (asterisk) and Meckel’s cartilage (MC). Bar 10 μm

Mandibles of 8.5-dPF-old H. bimaculatus explants, cultured in vitro for 7 days, were examined at the ultrastructural level for details on the further deposition, in vitro, of predentin and attachment bone. The tooth in position 3 was chosen as an example for the description below (Figs. 9, 10, 11, 12).

Fig. 5 After 7 days in culture, the tooth in position 3 has produced matrix in vitro (arrowhead), whereas its neighbour in position 4 (asterisk) has further developed to the attachment stage (MC Meckel’s cartilage). Bar 10 μm Fig. 6 Tooth in position 6 (asterisk) at the time of explantation (t0=144 hPF), showing the start of matrix deposition (arrowhead). MC Meckel’s cartilage. Bar 10 μm Fig. 7 Tooth in position 6 (asterisk) after 3 days in culture. Matrix production has continued (arrowhead newly forming attachment bone, MC Meckel’s cartilage). Bar 10 μm Fig. 8 Same tooth (asterisk) after 7 days in vitro. The attachment bone (arrowhead) is now prominent (MC Meckel’s cartilage). Bar 10 μm

380

The predentin matrix deposited in vitro contained thin collagen fibrils randomly dispersed throughout the matrix (Fig. 10). Several of them anchored in the lamina densa of the basement membrane lining the ameloblast-matrix in-

terface. Numerous electron-dense structures, which were 5 nm in diameter on average and surrounded by a membrane and which probably were matrix vesicles, were dispersed in the matrix (Fig. 11). Short odontoblastic prolongations

381 3 Fig. 9-12 Light (Fig. 9) and transmission electron micrographs (Figs. 10, 11, 12) of teeth or tooth germs on the mandible of a H. bimaculatus head explanted at 8.5 dPF and cultured in vitro for 7 days. Fig. 9 Developing tooth in position 3 on the two hemimandibles. The section is slightly oblique, so that the tooth is observed both at the level of its shaft (left box) and at its attachment site (right box). An arrowhead indicates the unmineralized attachment bone formed in vitro (MC Meckel’s cartilage). Bar 20 μm Fig. 10 Detail of the tooth in position 3 sectioned in the mid-region of the shaft (left box in Fig. 9), showing the predentin (PD) deposited in vitro. This matrix is composed of collagen fibrils (black arrowheads), some of which anchor into the lamina densa of the basement membrane (white arrowhead). AM Ameloblast, OD odontoblast. Bar 1 μm Fig. 11 Same section as in Fig. 10. Detail of the odontoblasts (OD) and ameloblasts (AM), surrounding the predentin (PD) formed in vitro (black arrowheads matrix vesicles, white arrowhead lamina densa of the basement membrane, G Golgi region, M mitochondrion). Bar 0.5 μm Fig. 12 Section through the attachment region of the tooth in position 3 (right box in Fig. 9). Detail of the (unmineralized) attachment bone matrix (AB) formed in vitro and surrounded by osteoblasts (OB). The arrowheads indicate matrix vesicles (PB perichondral bone surrounding Meckel’s cartilage). Bar1 μm

penetrated into the predentin (Figs. 10, 11). Both odontoblasts and ameloblasts showed a healthy aspect (Fig. 11). The plump odontoblasts possessed the features of actively secreting cells, as indicated by a well-developed rough endoplasmic reticulum (RER), prominent Golgi regions and secretory vesicles (Figs. 10, 11). The ameloblasts similarly contained prominent Golgi regions and several mito-

Fig. 13 D. rerio. Representation of the observed pattern of dentition at several ontogenetic stages, at the time of explantation (At0= control in vivo), and after 3 days (Bt0+3d) and 7 days (Ct0+7d) of in vitro culture. The dentitions having the largest number of teeth are represented. Small individual differences and left-right differences sometimes occur (line internal margin of the pharyngeal jaw, even

chondria and were associated with each other by finger-like cell processes (Figs. 10, 11). The matrix of the attachment bone, deposited in vitro, resembled the predentin matrix, from which it could be distinguished only by a larger number of matrix vesicles and the absence of a basal lamina between the matrix and the lining cells (Fig. 12). Large plump secretory osteoblasts, in the prolongation of the pulpal odontoblasts, lined the attachment bone, whereas at the opposite side, the attachment bone matrix was separated from the perichondral bone surrounding Meckel’s cartilage by an osteoblastic layer. In vitro pharyngeal tooth development in D. rerio Appearance of new tooth germs In control larvae at 45 hPF (i.e. a few hours before hatching) and 48 hPF (i.e. around hatching), only one tooth germ was present on each side, in position 4V (Fig. 13A). After 3 days in culture, this germ had developed further and one, or more frequently two, new germs had appeared, in positions 3V and 5V, respectively (Fig. 13B). In control larvae of 55 hPF, two (and in some cases three) tooth germs were present: a well-differentiated germ in position 4V and less developed germs in positions 3V and 5V. After 3 days in culture, three teeth were present: an attached tooth in position 4V and two well-differentiated germs in positions 3V and 5V, respectively (Fig. 13B).

when the fifth ceratobranchial cartilages have not yet differentiated, as at 48-hPF and younger animals). Tooth germs and teeth are represented by dots (white, grey, black different developmental stages, cf. Fig. 2). Dashed lines positions of teeth 3V, 4V and 5V, Ca caudal, R rostral

382

383 3 Fig. 14-15 D. rerio. Light micrographs of one side of the pharyngeal dentition at the time of explantation (t0=45 hPF, Fig. 14) and after 3 days in culture (t0+3d, Fig. 15) Fig. 14 Pharyngeal epithelium at the time of explantation. Only the tooth in position 4V is present in a stage of morphogenesis. The pharyngeal cavity is not yet open (white arrowheads boundary between pharyngeal epithelium and surrounding mesenchyme, i.e. position of the basement membrane). Bar 10 μm Fig. 15 After 3 days in vitro, the tooth in position 4V is in a cytodifferentiation phase, showing the onset of matrix production. Teeth in positions 3V and 5V are now clearly present, both in a stage of morphogenesis. Note the polarized cells of the enamel organ (future ameloblasts) of the tooth in position 3V. The pharyngeal cavity has opened in vitro. Bar 10 μm Fig. 16-22 D. rerio. Light micrographs of one side of the pharyngeal dentition at the time of explantation (t0=55 hPF, Fig. 16), after 3 days in culture (t0+3d, Figs. 17, 19, 21) and after 7 days in culture (t0+7d, Figs. 18, 20, 22) Fig. 16 At the time of explantation, the tooth germ in position 4V is in a cytodifferentiation stage, whereas slight outbulgings of the epithelium indicate the onset of formation of tooth germs in position 3V (white arrowhead) and 5V (black arrowhead). Bar 10 μm Fig. 17 Tooth germ in position 3V has reached the cytodifferentiation stage after 3 days in culture. As in vivo, the germ develops medial to the tooth in position 4V (CBV ceratobranchial V cartilage). Bar 10 μm Fig. 18 After 7 days in culture, the tooth germ in position 3V has formed a considerable amount of matrix. Bar 10 μm Fig. 19 The tooth germ in position 5V follows the same developmental trajectory as tooth germ 3V (4V tooth germ 4V). After 3 days in culture, it has reached the cytodifferentiation stage. Note the presence of an epithelial crypt which has appeared in vitro (white arrowhead). Bar 10 μm Fig. 20 After 7 days in culture, tooth germ in position 5V has deposited a distinct amount of matrix. Bar 10 μm Fig. 21 The tooth in position 4V, which had started matrix deposition at the moment of explantation (cf. Fig. 16) has attached after 3 days in culture onto the perichondral bone surrounding the ceratobranchial V cartilage (CBV). The white arrowhead indicates the crypt that has formed in vitro. Bar 10 μm Fig. 22 The tooth in position 4V is still attached after 7 days in vitro; there are no signs of shedding (CBV ceratobranchial V cartilage). Bar 10 μm Fig. 23-25 D. rerio. Light micrographs of one side of the pharyngeal dentition at the time of explantation (t0=79 hPF, Fig. 23), after 3 days in culture (t0+3d, Fig. 24) and after 7 days in culture (t0+7d, Fig. 25) Fig. 23 A fourth tooth germ (white arrowhead), the first replacement tooth in position 4V, is present as an epithelial bulging at 79 hPF. Tooth germs present at positions 3V, 4V and 5V are also indicated. Bar 20 μm Fig. 24 This fourth tooth germ has undergone morphogenesis after 3 days in culture (white arrowhead). As in vivo, this tooth germ is located ventral to the tooth in position 4V and is flanked by tooth germs in positions 3V and 5V. Bar 20 μm Fig. 25 The replacement tooth germ (asterisk) has undergone cytodifferentiation after 7 days in culture and a noticeable amount of matrix is present. The tooth germ lateral to it is the germ in position 5V. Bar 20 μm

After 7 days in culture, no new germs had been added but the three teeth present had developed further (Fig. 13C). In 69-hPF controls, three tooth germs were clearly distinguishable at positions 4V, 3V and 5V. In explants taken at 69 hPF and allowed to develop in vitro for 3 and 7 days, no new germs were added and the three germs continued to develop (Fig. 13B, C). In 79-hPF larvae, four tooth germs or teeth were clearly distinguishable, the oldest of which (position 4V) was

attached and the youngest of which (the successor in position 4V) was a mere bulge on the epithelium (Fig. 13A). In explants allowed to develop in vitro for 3 and 7 days, no new germs were added, although the youngest germ in position 4V had developed further (Fig. 13B, C). In general, the order of tooth formation in vitro followed that in vivo, i.e. first 4V, followed by the almost simultaneous appearance of germs 3V and 5V and then the successor in position 4V. Tooth development in vitro nevertheless showed a delay compared with odontogenesis in vivo comparable to that observed for H. bimaculatus, i.e. approximately 50– 60 h. Morphogenesis, differentiation and attachment At 45 hPF, only the tooth germ in position 4V was present and was in a phase of morphogenesis: the epithelium had clearly invaginated and formed a bell-shaped enamel organ surrounding a small number of mesenchymal cells (Fig. 14). The epithelium adjacent to this tooth germ (i.e. the point of development of the future germs in positions 3V and 5V) showed no evidence of differentiation. After 3 days in culture, the pharyngeal cavity was open and three tooth germs were now clearly present: the tooth germ in position 4V, with a slight but distinct amount of matrix, and germs in positions 3V and 5V, both in a stage of morphogenesis (Fig. 15). Thus, germs that were absent at the time of explantation (or, at least, were morphologically indistinguishable) developed in vitro but, perhaps because of the experimental set-up (none of these explants were cultured for 7 days), reached the stage of morphogenesis only. When the explants were made slightly later (t0=55 hPF), two or three tooth germs were present: a tooth germ in position 4V, with a small but distinct amount of matrix, and germs in positions 3V and 5V (Fig. 16). The slight bulging of the epithelium at these two locations showed that these germs were still in a phase of initiation and had hardly entered the phase of morphogenesis (Fig. 16). The further development of the germs in positions 3V and 5V exhibited remarkable parallels. After 3 days in culture, the germ in position 3V showed the onset of matrix deposition (Fig. 17) and a noticeable amount of matrix was produced after 7 days in culture (Fig. 18). Similarly, the tooth germ in position 5V was in the cytodifferentiation stage after 3 days in culture (Fig. 19). By 7 days in culture, this tooth had a distinct amount of matrix (Fig. 20). Thus, germs that were in a stage of initiation at the moment of explantation underwent morphogenesis and cytodifferentiation over a period of 7 days in vitro. On the other hand, germs depositing matrix at the time of explantation (tooth in position 4V, Fig. 16) ended up attached in vitro, even after as little as 3 days (Fig. 21). This situation did not change after 7 days of in vitro culture (Fig. 22). Despite the finding of a crypt (Figs. 19, 21), the tooth in position 4V did not erupt. Similar to the condition in vivo, tooth 4V fully ankylosed to the perichondral bone surrounding the ceratobranchial V cartilage, without the involvement of bony apolamellae. In contrast, teeth in

384

positions 3V and 5V, when explanted prior to the onset of matrix formation, did not become attached, not even after 7 days in culture. In vivo, germs in positions 3V and 5V became attached around 144 hPF (Van der heyden and Huysseune 2000). Even considering a delay of approximately 50 h attributable to in vitro conditions, germs 3V and 5V, explanted at 55 hPF, should in theory have had sufficient time to attach after 7 days in culture. Finally, a fourth tooth germ, located ventral to the functional tooth in position 4V, was in a stage of morphogenesis when the heads were explanted at 79 hPF (Fig. 23). This germ, the first replacement tooth in position 4V,

showed the onset of matrix deposition after 3 days in culture (Fig. 24) and was well-differentiated after 7 days in culture (Fig. 25). In the current experimental set-up, i.e. with a maximum of 7 days in culture, no germs underwent the entire process of morphological initiation, morphogenesis, differentiation and attachment in vitro. Ultrastructure of teeth formed in vitro In control specimens at 69 hPF, germs in positions 3V and 4V could be seen in a single cross section (Fig. 26). The

385 3 Fig. 26-34 TEM micrographs of zebrafish tooth germs in positions 3V and 4V in a control at 69 hPF (Figs. 26, 29, 30), in a specimen explanted at 69 hPF and kept in culture for 3 days (Figs. 27, 31, 32), and in a control specimen at the moment of explant fixation, i.e., 6 dPF (Figs. 28, 33, 34) Fig. 26 Tooth germs in positions 3V and 4V at the moment of explanting. Tooth germ 4V has produced a small amount of matrix. PC pharyngeal cavity. Bar 10 µm Fig. 27 General overview of the tooth germs in positions 3V and 4V after 3 days in vitro. The mineralized tip of the tooth in position 4V (white arrowhead) was present at the time of explanting and is lined at its inner surface by predentin formed in vitro (black arrowhead). DP dental papilla; EO enamel organ. Bar 10 µm Fig. 28 General overview of the teeth in positions 3V and 4V at 6 dPF in vivo. In both teeth the dentin matrix is mineralized. DP dental papilla; EO enamel organ. Bar 5 µm Fig. 29 Detail of the tooth germ in position 3V shown in Fig. 26. The dental papilla cells (DPC) and the inner dental epithelial cells (IDE) are rather undifferentiated and are separated by a basal lamina (arrowhead). Bar 1 µm Fig. 30 Detail of the tip of the tooth in position 4V shown in Fig. 26. The odontoblasts (OD) have differentiated and formed a noticeable amount of dentin-like matrix, which is mineralized. Distinct patches of mineral crystals (arrowheads) extend up to the ameloblast (AM) surface. The basement membrane is no longer discernable. Bar 1 µm Fig. 31 Detail of the matrix of the tooth germ in position 3V (shown in Fig. 27). A thin layer of predentin (PD) has been formed in vitro. Odontoblastic processes (arrowheads) extend into the predentin matrix. The lamina densa of the basement membrane is still present (asterisk), lining the ameloblast (AM) surfaces. OD odontoblast. Bar 0.5 µm Fig. 32 Detail of the predentin (PD) and dentin (D) of the tooth in position 4V (shown in Fig. 27). The predentin consists of densely packed collagen fibrils, parallel to the tooth surface. Collagen fibrils are no longer discernable in the mineralized dentin, which was formed in vivo prior to explanting. AM ameloblast, OD odontoblast. Bar 0.5 µm Fig. 33 The tooth in position 3V (shown in Fig. 28) has formed a large amount of predentin (PD) and dentin (D). A basement membrane lining the ameloblast (AM) surfaces is no longer visible. Bar 0.5 µm Fig. 34 Detail of the tooth in position 4V (shown in Fig. 28). Odontoblasts and ameloblasts (AM) are separated from each other by a considerable layer of predentin (PD) and mineralized dentin (D). Bar 0.5 µm

tooth germ in position 3V was at the stage of morphogenesis (Figs. 26, 29), with the dental papilla being surrounded by the dental epithelium. Both the dental papilla cells (preodontoblasts) and inner dental epithelium cells (preameloblasts) were still undifferentiated. The germ in position 4V had formed a thin layer of dentin-like matrix (Figs. 26, 30). The dental papilla cells and the inner dental epithelium cells had clearly differentiated into secretory odontoblasts and ameloblasts, respectively. Both types of cells were characterized by the presence of many mitochondria and an RER-rich cytoplasm. The matrix had started to mineralize. Patches of mineral crystals appeared to cluster along the dentinal collagen fibrils. The basal lamina located between the ameloblasts and the dentin surface could no longer be distinguished. In heads explanted at 69 hPF and kept in culture for 3 days (Figs. 27, 31, 32), the tooth in position 4V became attached and the tooth in position 3V was at a stage of

cytodifferentiation (Fig. 27). The newly deposited matrix of the tooth in position 3V was unmineralized and consisted of collagen fibrils that were oriented parallel to the tooth surface (Fig. 31). The odontoblasts contained a large amount of RER and showed conspicuous cytoplasmic prolongations that extended well into the predentin-like matrix. A basal lamina was still present between the surface of the matrix and the inner dental epithelium. Whereas the tip of the tooth in position 4V was wellmineralized, mineralization did not extend downwards along the shaft and the bone of attachment (Fig. 27). The unmineralized matrix formed in vitro, whereas the tooth tip had been mineralized at explantation (Figs. 27, 32). The odontoblasts were in a secretory phase as indicated by their cytoplasmic content rich in RER cisternae (Fig. 32). The predentin matrix formed in vitro consisted of densely packed collagen fibrils following the long axis of the tooth (Fig. 32). In control specimens fixed at 6 dPF (i.e. approximately similar in age to the explants of 69 hPF, cultured for 3 days), the tooth in position 3V showed a thick layer of matrix that was partly mineralized (Figs. 28, 33). A layer of densely packed unmineralized collagen fibrils constituted the predentin matrix bordering the odontoblast surface. Near the mineralization front, mineral crystals aggregated around the collagen fibrils. The odontoblasts that deposited the dentin matrix were highly differentiated with their cytoplasm filled with RER cisternae and containing numerous mitochondria. Opposite the mineralized dentin and facing the inner dental epithelium cells, the basal lamina could no longer be distinguished (Fig. 33). The external (i.e. dentin) layer of the tooth in position 4V was well mineralized, whereas the internal side of the tooth facing the pulp cavity was bordered by an unmineralized predentin layer of about 1 μm thick (Figs. 28, 34 for detail). The tooth in position 4V had been attached for a number of days and its dental papilla showed a typically loose aspect (Fig. 28). The dental papilla cells still had a small amount of RER. Mineral crystals aligned along the densely packed collagen fibrils. A basal lamina was absent. The inner dental epithelium cells were flattened and closely adjoined each other. The isolated adult pharyngeal jaws cultured in vitro showed a healthy aspect and well-preserved structures, demonstrating that the in vitro culture technique used was appropriate for the incubation of isolated adult organs. We found evidence that attachment of teeth continued in vitro if it had started prior to explantation (data not shown).

Discussion In this study, we have investigated two species that belong to two phylogenetically distant teleost families (Cichlidae and Cyprinidae) and that possess either oral and pharyngeal teeth (cichlids) or pharyngeal teeth only (cyprinids). We

386

demonstrate that tooth development in vivo can be mimicked in vitro in both species when either head explants from larvae or jaws isolated from adults are incubated for a number of days in a chemically defined medium. Tooth initiation in vitro Our experiments reveal that tooth germs can be initiated de novo under the conditions used in our in vitro cultures (as with Hemichromis explants of 120 hPF or Danio explants of 45 hPF). Reports in the literature on the possibility that tooth germs in vertebrates can be generated de novo are somewhat contradictory. Conflicting results may be partly explained by the culture conditions used. For example, in oculo grafts of toothless mandibular arches of mice in homologous hosts yield teeth, even though a dental lamina is absent at the onset of the experiments (Lumsden 1979; Lumsden and Buchanan 1986). Mouse teeth can be initiated in vitro in a serum-containing medium supplemented with growth factors, even when the dental lamina is not present at the onset of the experiments (Chai et al. 1998; Amano et al. 1999). In contrast, Ferguson et al. (1983) have not been able to obtain any dental structures on cultured mandibulae of Alligator mississippiensis (neither in serum-free nor in serum-containing media) when explanted prior to mandibular differentiation and prior to the presence of tooth rudiments. To our knowledge, this is the first report concerning the in vitro initiation of teeth in teleost fish. In mice, excess of retinol and retinal initiates supernumerary tooth germs in the diastemal dental lamina in vitro (Kronmiller et al. 1994, 1995). Other factors, such as epidermal growth factor, give results comparable to retinol and retinal excess (Kronmiller 1995). The addition of alltrans retinoic acid to the culture medium apparently does not alter the tooth pattern obtained in 48 hPF, 72 hPF and 10 dPF zebrafish head explants (C. Van der heyden et al., unpublished results). Nerve growth factor appears to be an indispensable medium supplement for murine tooth initiation and morphogenesis (Amano et al. 1999). Surprisingly, innervation has been claimed to play no role in murine tooth initiation (Lumsden and Buchanan 1986). In teleosts, initiation of replacement teeth is arrested in denervated lower jaws in adults of the cichlid Tilapia mariae (Tuisku and Hildebrand 1994); innervation therefore seems to be required at least for the development of replacement teeth. Koumans and Sire (1996) observed the development of cichlid tooth germs in a serum-free medium but were not sure whether these germs had been initiated at the moment of explantation. Our results confirm the formation de novo of tooth germs in vitro. In contrast to our results, no new tooth germs were initiated in head explants of larval medaka and zebrafish in serum-containing medium (Miyake and Hall 1994). A possible explanation of these contradictory results is either an adverse effect of the serum or the absence of essential substances, such as dexamethasone, a synthetic glucocorticoid known to have a positive effect on epithelial differentiation (Simo et al. 1992), in the medium.

Because partial (Hemichromis) or entire heads (Danio) were explanted in our experimental set-up, we cannot neglect the possible actions of endogenous substances, such as certain growth factors, or the possible effect of the presence of nerves. However, these influences should be reduced given the absence of blood flow in the culture and the transection of cranial nerves caused by decapitation. We cannot exclude, nevertheless, that the variety of tissues present in the explants could be a source of paracrine factors that could possibly stimulate the development of new germs. To avoid unwanted influences from residual endogenous substances, explantation of isolated jaws is far more desirable than explantation of entire heads. However, the dissection of the jaws requires the tissue to be sectioned at multiple sites and, given the small size of the material and the difficulty of the removal of the extremely small jaws of the larval fish, the tissues might easily be damaged during dissection, most probably leading to unwanted variation in the results. So far, we do not know precisely at which time the oral and pharyngeal epithelium becomes committed to its odontogenic fate, nor at which time the ectomesenchyme becomes competent to respond. If there is a phase in which the epithelium is committed but does not show morphological signs of tooth formation, the initiation of new germs might not be surprising and may be attributable to this time lag between “molecular initiation” and “morphological initiation”. In a recent study, Laurenti et al. (2004) have shown that, in the zebrafish, the even-skipped gene eve1 is expressed in the placode area of the tooth germ in position 4V, but not in 3V and 5V. However, the initial expression domain might be large enough to cover the initiation sites of both 3V and 5V. If the expression domain associated with 4V renders competence to an area that also covers 3V and 5V, this may account for the initiation of 3V and 5V in explants in which these two germs are not yet morphologically present. Experiments are underway in the zebrafish with explants earlier than 45 hPF to examine whether all three germs can be initiated prior to their morphological appearance, and if so, from what moment in development onwards. Westergaard (1983, 1988) and Westergaard and Ferguson (1986, 1987) have hypothesized that tooth initiation is linked to jaw development (the bone support), thereby implying that tooth germs will not differentiate unless they receive some signals, in an interactive dialogue, that the jaw is large/old enough for a new germ to develop. In our in vitro cultures, teeth are initiated together with the development of the cartilaginous ceratobranchials and their perichondral bone, suggesting some relationship between tooth initiation and jaw growth. Schilling et al. (1996) have however reported the presence of teeth in zebrafish belonging to the flathead group of mutants lacking the cartilaginous fifth ceratobranchials, thus providing a strong argument against the view of Westergaard and Ferguson (1986, 1987). Similarly, some experiments in amphibian embryos have led to larvae developing teeth in the absence of a bone support (Signoret 1960).

387

From morphogenesis to attachment Most in vitro studies of murine teeth concern explants containing tooth rudiments, thus focusing mostly on stages of morphogenesis and differentiation (e.g. Koch 1967; Sakakura 1986; Mark et al. 1992). By using calf serum and by applying the so-called floatation method, Sakakura (1986) obtained normal three-dimensional development of cap-stage mouse molar tooth germs. In our cultures, the shape (and size and orientation) of the teeth differentiated in vitro appears to be similar to that of teeth having grown normally in vivo. This should nevertheless be confirmed by three-dimensional reconstructions. Mark et al. (1992) have demonstrated that exogenous retinoids are indispensable for mouse molar morphogenesis. In our experiments, despite no retinoids having been added to the medium, the teeth seem to undergo normal morphogenesis. We suppose that either the explants contain enough endogenous retinoids or that teeth in fish do not need this molecule for their further development. Rather unexpectedly, teeth 3V and 5V, in contrast to tooth 4V, fail to attach in the zebrafish explants, despite their having been given enough time to attach in vitro. One possible explanation is that the delay, observed to occur in the cultures, builds up over time and that these teeth would eventually attach. The observation that, in one experiment in which 3V and 5V were explanted at a stage of differentiation, they became attached to the ceratobranchial by an unmineralized connection after 7 days in culture is in favour of this hypothesis. Alternatively, the lack of ankylosis of certain teeth (albeit through unmineralized attachment bone) can possibly also be explained by their attachment site in vivo. Indeed, the tooth in position 4V is the only tooth of the ventral row ankylosing directly, and completely, onto the perichondral bone surrounding the fifth ceratobranchial cartilage. Given the general delay of ossification of the cranial bones in our culture conditions, the failure of 3V and 5V to attach might be the result of the bony apolamella to which these teeth normally attach not being sufficiently developed and the attachment site not being easily changed. This suggests an interaction between the cells located at the tooth base and those located at the surface of the bone support. In H. bimaculatus, attachment generally occurs in vitro. In vivo, the teeth attach to the perichondral bone via an annular collar of attachment bone (Huysseune and Sire 1997). This difference in attachment site in vivo between both species may well account for the differences observed in vitro. The finding that, in our cultures of Danio, both a pharyngeal lumen and crypts form but that tooth 4V, despite becoming attached, fails to erupt is significant. It supports our earlier conclusion (Huysseune and Sire 2004) that crypt formation in the epithelium and exposure of the tooth within the crypt are independent processes. Horne-Badovinac et al. (2001) have reported that lumen formation in the digestive tract of zebrafish is linked to a clustering of adherens junctions. If such a process causes lumen and crypt formation, it would indeed differ from the disruption of cellu-

lar interdigitations assumed to provoke exposure of the tooth tip in the crypt (Huysseune and Sire 2004). Ultrastructure of teeth developed in vitro Our TEM results for both fish species reveal strong similarities in the structural characteristics of the cells and matrices of the teeth cultured in vitro compared with those in vivo, except for the lack of mineralization. This “normality” of the teeth cultured in vitro is amply shown by the presence of an enamel organ with differentiated ameloblasts and an outer dental epithelium, the presence of a dental papilla with differentiated odontoblasts and by the nature of the matrix deposited. In short, the phenotype that we observe in vitro is similar to that observed in vivo, as can also be appreciated by comparing the data with published histological and ultrastructural details of developing first-generation teeth in Hemichromis (Huysseune and Sire 1997) and Danio (Huysseune et al. 1998). The use of markers to identify the structures developed in vitro as genuine teeth would therefore, in our view, be redundant. Moreover, the molecular analysis of teleost odontogenesis has only recently been initiated and, thus, no markers are as yet available for the identification of dental matrix proteins in teleost fish. In addition, the growing evidence that markers supposedly specific for dental matrices are expressed in non-dental structures in tetrapods (e.g. Leiser et al. 2002; Gluhak-Heinrich et al. 2003) should lead us to use them with great caution, especially in view of the virtual lack of knowledge regarding the molecular composition of fish tooth matrices. A general feature observed in our in vitro system is the slower development of the teeth (a delay close to 50 h in both species). Nevertheless, the pattern observed in vivo is maintained in vitro. Such a developmental delay has also been noted in the experiments carried out by Koumans and Sire (1996) and may be attributable to the epithelia having to cover the sectioned tissues before differentiation can proceed (J.Y. Sire, personal observations). Miyake and Hall (1994) have demonstrated faster development of the parasphenoid in vitro vs. in vivo, although other cranial bones do not show such an enhanced growth. Re-evaluation of our in vitro material with respect to the cranial bones may elucidate this issue. The presence of a large number of matrix vesicles in the absence of mineralization under in vitro conditions provides an alternative way to administer treatments inhibiting mineralization in vivo when studying these structures (e.g. Takano et al. 2000). A lack of mineral deposition in the cultured teeth of both H. bimaculatus and D. rerio is easily explained by the absence of supplemental organic phosphate, such as ß-glycerophosphate, in the culture medium (Nijweide and Burger 1990). In rodent teeth, matrix mineralization apparently requires serum and/or growth-promoting exogenous factors, such as embryonic extracts (Yamada et al. 1980). In the culture conditions used by Sakakura (1986), mineralization of both dentin and enamel of murine

388

molars occurs. Other calcifiable matrices cultured in vitro also mineralize in serum-containing medium (Thesleff 1976; Sakakura 1986 and references therein). Fish skeletal elements cultured in the serum-containing medium developed by Miyake and Hall (1994) also lack mineralization, which the authors attribute (in a general way) to the absence of factors promoting mineralization. In conclusion, we have shown that the serum-free medium developed by Koumans and Sire (1996) allows tooth germs in H. bimaculatus and D. rerio to develop in vitro from an apparently undifferentiated buccal/pharyngeal epithelium and, when present at the time of explantation, to continue their development up to a stage of attachment. In addition, the medium allows the morphogenesis and cytodifferentiation of the tooth germs as is observed in vivo and the establishment of a dental pattern (place and order of tooth appearance and of attachment) that mimics that seen in vivo. The tooth phenotype obtained in vitro is indistinguishable from that observed in vivo. Molecular data will nevertheless be needed to confirm that the regulatory cascades involved in the epitheliomesenchymal interactions governing tooth development are conserved in vitro. To date, only a few teleost homologues of mammalian genes have been examined for their expression pattern during wild-type zebrafish tooth formation, revealing both similarities and significant differences (Laurenti et al. 2004; Jackman et al. 2004); this should be compared with the expression patterns of more than 200 genes in rodent (mainly mouse) teeth (cf. the database http://www.bite-it.helsinki.fi/; Nieminen 1998). Once more data become available regarding the genes involved in the development of zebrafish teeth in vivo, these can clearly be compared with data concerning the situation in vitro. Because our culture conditions allow new tooth germs to be initiated in vitro and because morphogenesis and cytodifferentiation continue in much the same way as under in vivo conditions, this method will enable tests to be carried out regarding the effects of different exogenous substances on the stimulation or inhibition of the different stages of odontogenesis under standardized conditions. Importantly, the culture conditions described here should allow heads from embryonically lethal (zebrafish) mutants to be cultured beyond the point at which the embryos normally die. Similarly, studies involving knockdown of genes with antisense morpholino oligonucleotides causing early lethality might benefit from the use of head explants to explore the effects on (later) craniofacial development. Thus, organotypic culture in serum-free conditions may become a powerful tool in developmental studies and open new perspectives for craniofacial research. Acknowledgements The authors thank Ing. G. De Wever and N. Van Damme for their valuable help in making the serial sections. Dr. T. Moens is acknowledged for numerous comments on earlier drafts of this paper. TEM work and preparation of the photographic prints were carried out at the “Service de Microscopie Electronique de l’IFR de Biologie integrative—CNRS—Université Paris 6”.

References Amano O, Bringas P, Takahashi I, Takahashi K, Yamane A, Chai Y, Nuckolls GH, Shum L, Slavkin HC (1999) Nerve growth factor (NGF) supports tooth morphogenesis in mouse first branchial arch explants. Dev Dyn 216:299–310 Chai Y, Bringas P, Shuler C, Devaney E, Grosschedl R, Slavkin HC (1998) A mouse mandibular culture model permits the study of neural crest cell migration and tooth development. Int J Dev Biol 42:87–94 Ferguson MWJ, Honig LS, Bringas P, Slavkin HC (1983) Alligator mandibular development during long term organ culture. In Vitro 19:385–393 Gluhak-Heinrich J, Ye L, Bonewald LF, Feng JQ, MacDougall M, Harris SE, Pavlin D (2003) Mechanical loading stimulates dentin matrix protein 1 (DMP1) expression in osteocytes in vivo. J Bone Miner Res 18:807–817 Horne-Badovinac S, Lin D, Waldron S, Schwarz M, Mbamalu G, Pawson T, Jan YN, Stainier DYR, Abdelilah-Seyfried S (2001) Positional cloning of heart and soul reveals multiple roles for PKCλ in zebrafish organogenesis. Curr Biol 11:1492–1502 Huysseune A (1990) Development of the anterior part of the mandible and the mandibular dentition in two species of Cichlidae (Teleostei). Cybium 14:327–344 Huysseune A, Sire JY (1992) Bone and cartilage resorption in relation to tooth development in the anterior part of the mandible in cichlid fish: a light and TEM study. Anat Rec 234:1–14 Huysseune A, Sire JY (1997) Development of first generation teeth in Hemichromis bimaculatus (Teleostei: Cichlidae). Tissue Cell 29:679–697 Huysseune A, Sire JY (2004) The role of epithelial remodelling in tooth eruption in larval zebrafish.Cell Tissue Res 315:85-95 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 Jackman WR, Draper BW, Stock DW (2004) Fgf signaling is required for zebrafish tooth development. Dev Biol 274:139–157 Koch WE (1967) In vitro differentiation of tooth rudiments of embryonic mice. 1. Transfilter interaction of embryonic incisor tissues. J Zool 165:155–170 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 Anim 32:612– 626 Kronmiller JE (1995) Spatial distribution of epidermal growth factor transcripts and effects of exogenous epidermal growth factor on the pattern of the mouse dental lamina. Arch Oral Biol 40:137– 143 Kronmiller JE, Beeman CS, Kwiecien K, Rollins T (1994) Effects of the intermediate retinoid metabolite retinal on the pattern of the dental lamina in vitro. Arch Oral Biol 39:839–845 Kronmiller JE, Nguyen T, Berndt W (1995) Instruction by retinoic acid of incisor morphology in the mouse embryonic mandible. Arch Oral Biol 40:589–595 Laurenti P, Thaëron C, Allizard F, Huysseune A, Sire JY (2004) Cellular expression of eve1 suggests its requirement for the differentiation of the ameloblasts and for the initiation and morphogenesis of the first tooth in the zebrafish (Danio rerio). Dev Dyn 230:727–733 Leiser Y, Dafni L, Mao Z, Taylor A, Rosenfeld E, Shay B, Fermon E, Haze A, Deutsch D (2002) The expression of the acidic tuftelin protein in developing and mineralizing enamel and in non-mineralizing soft tissues. J Dent Res 81 (Special Issue B): B307 Lumsden AGS (1979) Pattern formation in the molar dentition of the mouse. J Biol Buccale 7:77–103 Lumsden AGS, Buchanan JAG (1986) An experimental study of timing and topography of early tooth development in the mouse embryo with an analysis of the role of innervation. Arch Oral Biol 31:301–311

389 Mark MP, Bloch-Zupan A, Ruch JV (1992) Effects of retinoids on tooth morphogenesis and cytodifferentiations, in vitro. Int J Dev Biol 36:517–526 Miyake T, Hall BK (1994) Development of in vitro organ culture techniques for differentiation and growth of cartilage and bones from teleost fish and comparisons with in vivo skeletal development. J Exp Zool 268:22–43 Nieminen P, Pekkanen M, Åberg T, Thesleff I (1998) A graphical WWW-database on gene expression in tooth. Eur J Oral Sci 106:7–11 Nijweide PJ, Burger EH (1990) Mechanisms of bone formation in vitro. In: Hall BK (ed) Bone, vol 1. The osteoblast and osteocyte. Telford, Caldwell, pp 303–326 Sakakura Y (1986) A new culture method assuring the threedimensional development of the mouse embryonic molar tooth in vitro. Calcif Tissue Int 39:271–278 Schilling TF, Piotrowski T, Grandel H, Brand M, Heisenberg CP, Jiang YJ, Beuchle D, Hammerschmidt M, Kane DA, Mullins MC, Eeden FJM van, Kelsh RN, Furutani-Seiki M, Granato M, Haffter P, Ödenthal J, Warga RM, Trowe T, Nüsslein-Volhard C (1996) Jaw and branchial arch mutants in zebrafish. I. Branchial arches. Development 123:329–344 Signoret J (1960) Céphalogenèse chez le triton Pleurodeles waltlii Michah après traitement de la gastrula par le chlorure de lithium. Mém Soc Zool Fr 32:1–117 Simo P, Simon-Assmann P, Arnold C, Kedinger M (1992) Mesenchyme-mediated effect of dexamethasone on laminin in cocultures of embryonic gut epithelial cells and mesenchyme-derived cells. J Cell Sci 101:161–171 Stock DW, Weiss KM, Zhao Z (1997) Patterning of the mammalian dentition in development and evolution. BioEssays 19:481–490 Takano Y, Sakai H, Baba O, Terashima T (2000) Differential involvement of matrix vesicles during the initial and appositional mineralization processes in bone, dentin, and cementum. Bone 26:333–339 Thesleff I (1976) Differentiation of odontogenic tissues in organ culture. Scand J Dent Res 84:353–356

Thesleff I (2000) Genetic basis of tooth development and dental defects. Acta Odontol Scand 58:191–194 Thesleff I (2003) Epithelial-mesenchymal signalling regulating tooth morphogenesis. J Cell Sci 116:1647–1648 Tuisku F, Hildebrand C (1994) Evidence of neural influence on tooth germ generation in a polyphyodont species. Dev Biol 65:1–9 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 der heyden C, Wautier K, Huysseune A (2001) Tooth succession in the zebrafish (Danio rerio). Arch Oral Biol 46:1051– 1058 Wautier K, Van der heyden C, Huysseune A (2001) A quantitative analysis of the pharyngeal tooth shape in the zebrafish (Danio rerio) (Teleostei, Cyprinidae). Arch Oral Biol 46:67–75 Westergaard B (1983) A new detailed model for mammalian dentitional evolution. Z Zool Syst Evolutionsforsch 21:67–78 Westergaard B (1988) The pattern of embryonic tooth initiation in reptiles. Mém Mus Nat Hist Nat Série C Paris 53:55–63 Westergaard B, Ferguson MJW (1986) Development of the dentition in Alligator mississippiensis. Early embryonic development in the lower jaw. J Zool Lond 210:575–597 Westergaard B, Ferguson MJW (1987) Development of the dentition in Alligator mississippiensis. Later development in the lower jaws of embryos, hatchlings and young juveniles. J Zool Lond 212:191–222 Yamada M, Bringas P, Grodin M, MacDougall M, Cummings E, Grimmett J, Weliky B, Slavkin HC (1980) Chemically-defined organ culture of embryonic mouse tooth organs: morphogenesis, dentinogenesis and amelogenesis. J Biol Buccale 8:127– 139