Abstract
N-cadherin is a well-studied classic cadherin involved in multiple developmental processes and is also known to have a signaling function. Using the zebrafish (Danio rerio) as a model, we tested the hypothesis that tooth morphogenesis is accompanied by dynamic changes in N-cadherin distribution and that absence of N-cadherin disturbs tooth development. N-cadherin, encoded by the gene cdh2, is absent during the initiation and morphogenesis stages of both primary (first-generation) and replacement teeth, as demonstrated by immunohistochemistry. However, N-cadherin is up-regulated at the onset of differentiation of cells of the inner dental epithelium and the dental papilla, i.e., the ameloblasts and odontoblasts, respectively. In the inner dental epithelium, N-cadherin is co-expressed with E-cadherin, excluding the occurrence of cadherin switching such as observed during human tooth development. While early lethality of N-cadherin knockout mice prevents any functional study of N-cadherin in mouse odontogenesis, zebrafish parachute (pac) mutants, deficient for N-cadherin, survive beyond the age when primary teeth normally start to form. In these mutants, the first tooth forms, but its development stops at the early cytodifferentiation stage. N-cadherin deficiency also completely inhibits the development of the other first-generation teeth, possibly due to the absence of N-cadherin signaling once the first tooth has differentiated.
Introduction
The zebrafish, Danio rerio, has become one of the most important vertebrate models for studying early development and the molecular basis of diseases. Unlike the mouse, which has an evolutionarily derived dentition with continuously growing incisors and non-replacing molars, the zebrafish displays polyphyodonty, i.e., teeth are continuously replaced throughout life (Huysseune et al., 1998). The teeth are restricted to the last pair of ceratobranchials (pharyngeal jaws). The first-generation (or primary) teeth form during embryonic and early larval life and start to be replaced long before the full set of 11 tooth loci on each body side is established. This feature offers the major advantage, compared with the mouse model, for the study of tooth development and replacement even in mutants that display early lethality. Otherwise, many characteristics of zebrafish tooth development are shared with mammals. Thus, as in mammals, teeth pass through initiation, morphogenesis, and cytodifferentiation stages (Huysseune et al., 1998; Laurenti et al., 2004) before attaching to the ceratobranchial bone and erupting. Likewise, many of the molecular cues regulating tooth development in zebrafish are shared with those controlling mammalian odontogenesis (reviewed in Stock, 2007).
While much emphasis in odontogenesis research is placed on the role of signaling molecules, their receptors, downstream signaling pathways, and transcription factors, much less attention has been paid to the role of cell-cell adhesion in the morphogenetic changes that characterize early stages of tooth formation. During zebrafish primary tooth development, a local thickening (placode) of the pharyngeal epithelium invaginates into the underlying mesenchyme to form an epithelial bud. In contrast, replacement teeth form from a successional lamina. Unlike the placode of first-generation teeth, the successional lamina develops as an outgrowth of the crypt epithelium surrounding the functional tooth (Huysseune, 2006). The formation of placode and successional lamina, and the development of the tooth itself, require substantial morphogenesis of both epithelial and mesenchymal cell layers. We hypothesized that these morphogenetic events are accompanied, and possibly preceded, by changes in cell-cell adhesion, analogous to what has been proposed for hair follicle morphogenesis (Jamora et al., 2005).
Cell-cell adhesion is provided by adherens junctions and desmosomes operating through a wide range of junctional proteins. The cadherin superfamily is comprised of a diversity of molecules with structural characteristics enabling them to have different functions at different times during development (Halbleib and Nelson, 2006; Hulpiau and van Roy, 2009). One large group consists of the classic cadherins, including, among others, E-cadherin, N-cadherin, R-cadherin, and cadherin-11.
In the zebrafish, N-cadherin, encoded by the gene cdh2, is expressed ubiquitously from late blastula stages onward. It becomes restricted to the central nervous system at around 24 hpf (Lele et al., 2002). In the mouse, N-cadherin is initially expressed at the gastrulation stage when epiblast cells down-regulate E-cadherin and undergo an epithelial-mesenchymal transition. This process includes the up-regulation of N-cadherin in the nascent mesoderm. In later stages and in adults, N-cadherin is expressed in neural tissue, cells of mesenchymal origin, and diverse endoderm-derived epithelial tissues in which E- and N-cadherin co-exist (Radice et al., 1997; Straub et al., 2011). E-cadherin and N-cadherin share many structural and functional features. Both establish calcium-dependent homophilic cell-cell adhesion with their extracellular domains and are connected to catenins at their intracellular domains. However, in contrast to E-cadherin, N-cadherin can interact with receptors of fibroblast growth factors (FGFR) at the cell surface and thus modulate their signaling functions (Williams et al., 1994; Stepniak et al., 2009).
Here, we used immunohistochemistry combined with thin-section histology to examine the detailed cellular distribution of N-cadherin during tooth development and replacement in the zebrafish. Furthermore, we explored the role of N-cadherin during odontogenesis by analyzing parachute (pac) mutants, deficient in N-cadherin (Jiang et al., 1996; Lele et al., 2002). We also examined whether E-cadherin, normally expressed throughout odontogenesis in the epithelial-derived part of the tooth (Verstraeten et al., 2010), is affected by the absence of N-cadherin in this mutant.
Materials & Methods
Zebrafish Collection
Zebrafish were mated and eggs were raised at 28.5°C in a 10-hour dark/14-hour light cycle. The embryos were sacrificed every 4 hrs starting 40 hrs post-fertilization (hpf). Parachute/ N-cadherin mutant zebrafish with mutated allele tm101b were obtained from the Nüsslein-Volhard Lab (Tübingen, Germany). This allele carries a point mutation that causes a premature stop codon at the beginning of the EC domain 4 and likely produces a null phenotype (Lele et al., 2002). In total, 27 embryos, displaying the abnormal tail morphogenesis typical of the parachute mutation, were collected from the ages of 60 to 112 hpf (Table). The mutation becomes lethal 4 to 5 days post-fertilization (dpf). Embryos were sacrificed by an overdose of the anesthetic MS222 and fixed overnight at 4°C in 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS). Adult zebrafish were similarly sacrificed by an overdose of MS222. Their pharyngeal jaws were microdissected under a Leica MZ Apo dissecting microscope and fixed overnight at 4°C in 4% PFA in PBS.
Analysis of the Tooth Phenotype in the pactm101b Mutant and Comparison with Wild-type (WT)
Teeth were scored based on serial 1- to 4-µm-thick sections. The mutation is early-lethal; several embryos were dying at the time of fixation, suggesting variable expressivity of the mutation, and were of too poor quality to be scored (numbers indicated in column “poor tissue quality”). All 27 embryos (of a batch of 115 embryos) showed the parachute phenotype with abnormal tail morphogenesis, indicating complete penetrance of the mutation. We tested the difference in presence/absence of teeth between pac and WT for the 60- to 64-hpf interval using the GENMOD procedure (SAS Institute, Inc.) for a binomial distribution and with the logit link function (a) χ2 = 16.61, p < 0.0001, as well as the difference in the presence/absence of teeth 3V1 and 5V1 between pac and WT for the 68- to 112-hpf interval (b) χ2 = 33.10, p < 0.0001.
Abbreviations: EM, early morphogenesis; LM, late morphogenesis; ED, early cytodifferentiation; LD, late cytodifferentiation; ATT, attached tooth (stages as defined in Borday-Birraux et al., 2006).
Tissue Processing
After fixation, embryos were de-pigmented, dehydrated in an increasing methanol series, and stored in 100% methanol at -20°C. The dissected pharyngeal jaws were decalcified in Morse’s solution (10% sodium citrate, 22.5% formic acid) for several days at 4°C, embedded in paraffin according to standard procedures, and sectioned at 5 µm.
Whole-mount immunohistochemistry (IHC) of embryonic and larval stages with mouse anti-N-cadherin antibody (1/300, BD Transduction Laboratories, Lexington, KT, USA) was performed as described previously (Verstraeten et al., 2012). To detect N-cadherin in paraffin sections, we used an antigen retrieval step (95°C in citrate buffer for 20 min). Sections were incubated overnight at 4°C with primary antibody (1/500 mouse anti-N-cadherin, BD Transduction Laboratories, Lexington, KT, USA). After removal of the primary antibody, the secondary antibody (1/300 anti-mouse-biotin labeled antibody, Dako Cytomations, Glostrup, Denmark) was applied for 1 hr at room temperature (RT). Finally, the sections were incubated with streptABComplex labeled with peroxidase (Dako Cytomations, Glostrup, Denmark) for 45 min at RT. 3,3′-diaminobenzidine (DAB) was used for visualization. pac mutants were subjected to whole-mount immunodetection or in situ hybridization for E-cadherin as described previously (Verstraeten et al., 2012). All whole mounts, including all 27 mutant embryos, were embedded in epon and serially sectioned at 4 µm.
Results
N-cadherin is Up-regulated at the Onset of Cytodifferentiation of Zebrafish Teeth
N-cadherin was expressed in advanced stages of tooth development only. Thus, the placode from which first-generation teeth arise showed no N-cadherin expression (Figs. 1A, 1A′). Similarly, the successional lamina giving rise to replacement teeth did not express N-cadherin (Figs. 2A, 2A′). The mesenchyme was also negative. Likewise, during the morphogenesis stage, N-cadherin protein was not detectable in the enamel organ or mesenchyme, whether of first-generation (Figs. 1B, 1B′) or replacement teeth (Figs. 2B, 2B′).
N-cadherin protein distribution during development of first-generation teeth.
N-cadherin protein distribution during tooth replacement in an adult zebrafish.
N-cadherin was up-regulated during the early cytodifferentiation stage in first-generation and in replacement teeth. N-cadherin was expressed both in the inner dental epithelium (IDE) and in some cells of the dental papilla (Figs. 1B, 1B′, 1C, 1C′, 2C, 2C′). During the late cytodifferentiation stage, expression of N-cadherin in the enamel organ was maintained, while a much stronger expression was observed in the differentiated odontoblasts adjacent to the matrix. Cells elsewhere in the dental papilla showed no N-cadherin expression (Figs. 1D, 1D′, 2D, 2D′).
After attachment of the tooth to the ceratobranchial bone and concomitant eruption, N-cadherin continued to be expressed in the odontoblasts. The reduced enamel organ, in contrast, lost its N-cadherin expression (Figs. 1D, 1D′). The highly polarized, actively secreting odontoblasts of young functional teeth maintained strong N-cadherin immunoreactivity (Fig. 2A). In more mature functional teeth, odontoblasts lost their polarized appearance along with most of their N-cadherin immunoreactivity (Fig. 2D).
Tooth Development is Arrested in N-cadherin-deficient Parachute Zebrafish
Embryonic development in the pactm101b (parachute) mutants appeared to be delayed compared with that in the wild type (WT), as inferred from the small brain volume, and the retardation in yolk resorption, pharyngeal lumen formation, and cranial cartilage differentiation (compare Figs. 3B, 3D, 3F with 3A, 3C, 3E). Tooth 4V1 (the first primary tooth to form) was first detected in the pactm101b mutants at 68 hpf (Table), i.e., close to 20 hrs later than during normal development (Borday-Birraux et al., 2006). This tooth did not progress beyond the early cytodifferentiation stage (Figs. 3B, 3D), not even at 112 hpf, when the mutation became lethal (Fig. 3F, Table). A replacement tooth normally starts to develop in position 4V at about 80 hpf, but no such tooth was observed in any of the specimens studied. We were also unable to detect teeth 3V1 and 5V1, even in specimens of 104 and 112 hpf (Figs. 3B, 3B′, 3D, 3D′, 3F, 3F′, Table). These 2 teeth, which flank tooth 4V1 in the WT, normally start to develop about 8 hrs after the appearance of 4V1, when tooth 4V1 reaches the early cytodifferentiation stage (Figs. 3A, 3C, 3E). In the mutants, E-cadherin was expressed in the enamel organ of tooth 4V1 in a manner similar to that in the wild type (compare Figs. 3A, 3C with 3B, 3D) (see Verstraeten et al., 2010).
Pharyngeal region of the pactm101b mutant (
Discussion
Zebrafish teeth are initiated very precociously (around 48 hpf), and even their replacement starts early (at around 80 hpf), i.e., before many mutations become lethal. This animal model thus offers an unparalleled advantage over the late-developing mouse dentition for functional studies, in particular to reveal how budding morphogenesis is governed by cell adhesion dynamics. We have exploited this opportunity and show that (1) N-cadherin protein, encoded by the gene cdh2, is up-regulated at the start of cytodifferentiation, and (2) its absence leads to an arrest of formation of the first primary tooth at the early cytodifferentiation stage.
N-cadherin is absent during the initiation and morphogenesis stages of first-generation or replacement teeth, and up-regulated at the start of cytodifferentiation in the cells of the IDE and the dental papilla, i.e., ameloblasts and odontoblasts, respectively. N-cadherin expression persists in the odontoblasts until they lose their polarized aspect (Appendix Fig.). This resembles the pattern of N-cadherin immunoreactivity during human tooth development (Heymann et al., 2002). Briefly, in humans, there is an absence of staining in the dental bud, weak N-cadherin staining in the cells that constitute the inner and outer dental epithelium (IDE, ODE), and increased staining in the more differentiated cells (pre-ameloblasts and ameloblasts). Strong N-cadherin immunoreactivity was also detected in differentiating and functional human odontoblasts (Heymann et al., 2002). In the mouse, N-cadherin expression is initiated after pre- odontoblasts and pre-ameloblasts have fully differentiated (Bartlett et al., 2010). Thus, in all species examined, N-cadherin expression is associated with differentiation of odontogenic cells. During cytodifferentiation, ameloblasts and odontoblasts undergo extensive polarization and cytological changes related to their secretory function. This is consistent with reports of N-cadherin as a marker of mesenchymal differentiation and its role in enhancing cellular differentiation in tissues such as corneal endothelium and cartilage (Larue et al., 1996; Delise and Tuan, 2002; Derycke and Bracke, 2004; Koh et al., 2008).
Verstraeten et al. (2010) reported expression of E-cadherin in the enamel organ but not in the mesenchyme throughout zebrafish tooth development (Appendix Fig.). Thus, cells of the IDE and ODE express both E-cadherin and N-cadherin. During human tooth development, E- and N-cadherin have mutually exclusive patterns of expression (Heymann et al., 2002). Xiao and Tsutsui (2012) concluded that when post-natal human oral epithelial cells are co-cultured with dental pulp stem cells, E-cadherin mediates epithelial invagination, while N-cadherin initiates mesenchymal condensation. The continued expression of E-cadherin in the zebrafish enamel organ until well after attachment (Verstraeten et al., 2010) excludes the occurrence of a switch from E- to N-cadherin. Such cadherin switching is associated with the initiation of cell motility and/or the definition of a different tissue layer (Takeichi, 2011; Sierant and Bartlett, 2012). In Xenopus ectodermal tissues, differential expression of cadherin molecules results in different types of actin assembly and therefore different types of tissue movements (Nandadasa et al., 2009). However, here we show co-expression of E- and N-cadherin in the same tissue layer, and possibly even in the same cells. Adherens junctions containing heterodimers of E- and N-cadherin might define a poorly studied special epithelial differentiation pathway during normal development (Straub et al., 2011). Interestingly, such heterodimers were typically observed in endoderm-derived tissues (Straub et al., 2011). Although still a matter of contention (Huysseune et al., 2009), the enamel organ of zebrafish teeth is considered to be derived from the supposedly endodermal basal layer of the pharyngeal epithelium.
Investigation of N-cadherin-deficient zebrafish revealed that tooth 4V1 never displays more than a thickened basal lamina, and therefore never progresses beyond the early cytodifferentiation stage. The general developmental delay in these mutants, estimated by us to be about 24 hrs, is insufficient to explain this phenotype at 112 hpf. In addition, E-cadherin expression in these tooth germs was found to be normal. Therefore, we conclude that tooth development is arrested at the onset of cytodifferentiation due to N-cadherin deficiency, suggesting a vital requirement for N-cadherin in odontoblast and ameloblast differentiation. During zebrafish tooth development, enameloid, a mixed ameloblast and odontoblast secretory product, is deposited first, followed by dentin (Huysseune et al., 1998; Stock, 2007). The role of N-cadherin in differentiation, as inferred from our expression data, thus fits with the absence of distinctive enameloid and ensuing dentin matrices in the parachute mutant. In human teeth, N-cadherin might be required for the ameloblast transformation and polarization necessary for enamel matrix secretion (Heymann et al., 2002). It has also been suggested to allow ameloblasts to move in rows during the secretory stage of enamel development (Sierant and Bartlett, 2012). In the human pulp cavity, N-cadherin could mediate direct binding between odontoblasts and the forming nerve plexus (Heymann et al., 2002). Regrettably, homozygous N-cadherin knockout mice die by day 10 of gestation (Radice et al., 1997), too early for any effects on the dentition to be observed.
The first successor tooth normally starts its development to replace 4V1 when the latter has attached and erupted. Its absence in the parachute mutants is therefore likely linked to the arrest of 4V1 differentiation. Interestingly, teeth 3V1 and 5V1, which normally start their development approximately 8 hrs after 4V1, were also absent, despite sufficient time to allow for their development and despite 4V1 having reached early cytodifferentiation. An intriguing possibility is that tooth 4V1 might act as a trigger for initiation of the other first-generation teeth (Huysseune and Witten, 2006). Given that blocking FGF receptors in the zebrafish by the pharmacological compound SU5402 prevents tooth formation (Jackman et al., 2004), and given that N-cadherin directly interacts with FGF receptors (Williams et al., 1994; Sierant and Bartlett, 2012), we speculate that N-cadherin deficiency in 4V1 leads to failure in FGFR-mediated signaling and failure of the formation of adjacent teeth. Whatever the mechanism that prevents the adjacent teeth from forming, analysis of our data from the pac mutants supports the idea that the development of tooth 4V1 could be regulated differently from that of teeth 3V1 and 5V1 (Laurenti et al., 2004).
In conclusion, analysis of our data on zebrafish tooth development demonstrates the up-regulation of N-cadherin expression during the cytodifferentiation stage as well as co-expression of E- and N-cadherin in the inner dental epithelium. N-cadherin deficiency does not prevent the first tooth from starting to develop, but stops its development at the early cytodifferentiation stage and completely inhibits the development of the other first-generation teeth, possibly due to the absence of N-cadherin signaling.
Footnotes
Acknowledgements
We thank the Nüsslein-Volhard Lab (Tübingen, Germany) for providing the parachute mutant zebrafish, Tommy D’heuvaert and Mieke Soenens for technical support, and Sam Vandenplas and Steven Van Belleghem for statistical advice.
This work was supported by a GOA research grant (BOF08/GOA/019) to JvH, FVR, and AH. BV acknowledges a grant from the Agency for Innovation by Science and Technology (IWT).
The authors declare no potential conflicts of interest with respect to the authorship and/or publication of this article.
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
