Department of Craniofacial Development and Stem Cell Biology, Dental Institute, King's College London, Guy's Hospital, London Bridge, London, United Kingdom
Division of Oral Anatomy, Department of Oral Biological Science, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan
Correspondence to: Atsushi Ohazama, Division of Oral Anatomy, Department of Oral Biological Science, Niigata University, Graduate School of Medical and Dental Sciences, 2–5274, Gakkocho-dori, Chuo-ku, Niigata 951–8514, Japan. E-mail: email@example.com
R-spondin family proteins are secreted glycoproteins containing cysteine rich domains and a thrombospondin type I domain, and consist of the four members (R-spondin1-4). R-spondin proteins have been shown to interact with the Frizzled8 and Lrp6 receptors to activate Wnt signalling (Nam et al., 2006). Recently, Lgr4, Lgr5, and Lgr6, have been identified as receptors for R-spondins (Carmon et al., 2011; de Lau et al., 2011). R-spondin/Lgr signaling plays important roles in many biological processes including regulating limb, lung, nail and placenta development in addition to stimulating adult stem cell proliferation, and activate Wnt signaling (de Lau et al., 2011).
Teeth develop from sequential and reciprocal interactions between epithelium and mesenchyme. The first morphological sign of tooth development is a thickening jaw epithelium. Tooth development proceeds in three stages: bud, cap, and bell, because the thickened epithelium progressively takes the form of the bud, cap and bell configuration. Primary enamel knots, signaling centers appear as thickened inner enamel epithelium at the early cap stage, but disappear by the late cap stage. Subsequently, epithelial cells differentiate into enamel-producing ameloblasts, while dentin-producing odontoblasts are differentiated from mesenchymal cells (dental papilla). It is known that multiple signaling pathways, including TNF, BMP, FGF, WNT, and SHH all play roles in regulating tooth development (Tucker and Sharpe, 2004, Ohazama and Sharpe, 2004, 2008). Tooth position, number, and shape are consistent in mammals, and are believed to be controlled by an intricate network between these pathways whose activities are regulated by the balance between ligands, activators, inhibitors, and receptors. The role of R-spondin/Lgr signaling in tooth development, however, remains unclear.
Murine dentition is an excellent experimental model to study the mechanisms of tooth development, because mice are the commonly studied mammals for investigating the genetics and molecular mechanisms of organogenesis. Unlike human, mice have only one incisor and three molars in each jaw quadrant that are divided by a tooth-less region, the diastema. It is believed that mice lost teeth in the diastema during evolution approximately 100Mya (Meng et al., 1994; Ji et al., 2002). Rudimentary tooth primordia is known to present in the diastema as remnants of evolutionary lost teeth. The bud-like structures (MS) that form anterior to the first molar tooth germ are visible at embryonic day (E) 12.5 and Shh expression is observed in only MS epithelium at the stage. At E13.5, another epithelial buds (R2) become prominent posterior to MS. Shh is expressed in R2, while MS loses Shh expression and regresses at this stage. R2 is involved in the formation of the mesial part of the first molar (Prochazka et al., 2010). The expression of negative feedback regulators of Fgf signaling, Sprouty2 and Sprouty4 are observed in the embryonic diastema, and mice with mutations in these molecules show supernumerary teeth in the diastema (Klein et al., 2006). The Shh inhibitor, Gas1, is expressed in the diastema and Gas1 mutant mice also exhibit extra teeth in the diastema (Ohazama et al., 2009). The mutation of a secreted Bmp antagonist and an inhibitor of Wnt signaling, Wise, also results in diastema tooth formation (Ohazama et al., 2008, Ahn et al., 2010, Kassai et al., 2005). Change (either overexpression or reduction) of the TNF signaling pathway proteins, Eda and Edar, has been shown to lead to supernumerary tooth formation in the diastema (Mustonen et al., 2003; Tucker et al., 2004, Peterková et al., 2005). The data from these mice suggest that mice have retained the genetic potential for the development of evolutionary lost teeth.
In addition to the presence of the diastema, mice have unique incisor teeth as they grow continuously throughout life with enamel being present only on the labial side. Continuous growth is supported by dental stem cells localized at the proximal end of incisor. The labial cervical loop at the proximal end of the incisor has been shown to contain an epithelial stem cell niche that houses the ameloblast progenitors (Harada et al., 1999). Mesenchymal stem cells are also present at the proximal end of incisors (Lapthanasupkul et al., 2012). Although Lgr5 has previously been shown to be expressed in the labial cervical loop, the role of R-spondin/Lgr signaling in the rodent incisor cervical loop, however, remains unclear (Suomalainen and Thesleff, 2010).
We show here dynamic spatio-temporal expression of R-spondins and Lgrs during murine odontogenesis. Although craniofacial malformations have been shown in R-spondin2 mutant mice, the role of R-spondin2 in tooth development is not understood (Nam et al., 2007a; Aoki et al., 2008; Bell et al., 2008; Yamada et al., 2009; Jin et al., 2011). To address this question, we examined tooth development in R-spondin2 mutant mice. Supernumerary teeth were found in the diastema, indicating that R-spondin2 plays a role in regulating tooth development in the diastema.
Expression of R-spondins and Lgrs in Molar Tooth Development
We carried out comparative in situ hybridization analysis of molar tooth development at E10.5, E11.5, E13.5, E14.5, and E18.5 in mice. This period encompasses molar tooth development from initiation to the onset of cytodifferentiation.
Initiation of tooth development (E10.5)
Initiation begins before the tooth anlagen is morphologically visible. The first signals are derived from tooth presumptive epithelium at E9.5–10.5 (Ferguson et al., 2000). R-spondin2 was expressed in the presumptive tooth mesenchyme at E10.5, but no expression of R-spondin1, R-spondin3, or R-spondin4 were detectable in the presumptive tooth region (Fig. 1A,F,K,P). Lgr4 was strongly expressed in the presumptive tooth epithelium, while weak expression was also observed in underneath mesenchyme (Fig. 2A). Faint expression of Lgr6 could be detected in the presumptive tooth epithelium and no expression of Lgr5 was observed in the presumptive tooth region at E10.5 (Fig. 2F,K).
Thickening tooth epithelium (E11.5)
Thickening of the oral epithelium takes place from E11.5 and R-spondin1 was weakly expressed in both tooth epithelium and mesenchyme at this time (Fig. 1B). Strong R-spondin2 expression was observed in the aboral side of first branchial arch mesenchyme (Fig. 1G). No expression of R-spondin3 or R-spondin4 was observed in the presumptive tooth region at E11.5 (Fig. 1L–Q). Lgr4 was strongly expressed in tooth epithelium, whereas weak expression of Lgr6 was observed in tooth epithelium (Fig. 2B,L). Lgr5 expression was found in the presumptive tooth region (Fig. 2G).
Bud stage (E13.5)
By E13.5, the tooth epithelium invaginates into underlying mesenchyme to form the epithelial bud. R-spondin1 is strongly expressed in mesenchyme, while weak expression was also observed in basal epithelium of tooth bud epithelium (Fig. 1C). Expression of R-spondin2 and R-spondin3 were observed in mesenchyme lingual to bud tooth epithelium (Fig. 1H,M). R-spondin4 was weakly expressed in basal epithelium of tooth bud epithelium and mesenchyme (Fig. 1R). Lgr4 was strongly expressed in bud tooth epithelium (Fig. 2C). Expression of Lgr5 was found in mesenchyme buccal to bud tooth epithelium, while weak expression was observed in collar of tooth epithelium (Fig. 2H). Weak expression of Lgr6 could be detected in tooth bud epithelium (Fig. 2M).
Cap stage (E14.5)
By E14.5, the bud basal epithelium develops into the inner and the outer enamel epithelium, and the mesenchyme develops into the dental papilla and the dental follicle. R-spondin1 was weakly expressed in mesenchyme and cervical loop epithelium (Fig. 1D). Expression of R-spondin2 and R-spondin4 were observed in the dental papilla, while R-spondin4 also showed weak expression in cervical loop epithelium (Fig. 1I,S). No expression of R-spondin3 could be detected in tooth germs (Fig. 1N). Lgr4 was expressed in collar of tooth epithelium and outer enamel epithelium (Fig. 2D). Weak expression of Lgr5 was observed in tooth epithelium, while Lgr6 showed expression in both tooth epithelium and mesenchyme (Fig. 2I,N).
The terminal differentiation of dentin-forming odontoblasts from dental papilla cells and the enamel-forming ameloblasts from the internal epithelium occurs between E18 to P0. Weak expression of R-spondin1 was observed in ameloblasts and developing pulp (Fig. 1E). R-spondin2 was expressed in ameloblasts at the presumptive cusp region, and dental papillae cells facing apical tooth epithelium (Fig. 1J; Nam et al., 2007b). Strong R-spondin4 expression was observed in developing pulp, whereas odontoblasts and ameloblasts showed only weak expression (Fig. 1T; Nam et al., 2007b). No expression of R-spondin3 could be detected in tooth germ at this stage (Fig. 1O). Lgr4 was weakly expressed in odontoblasts beneath developing cusps (Fig. 2E). Strong expression of Lgr6 was observed in ameloblasts and odontoblasts, while no expression of Lgr5 could be detected in tooth germs (Fig. 2J,O).
Expression of R-spondins and Lgrs in Incisor Tooth Development
The R-spondins and Lgrs we examined showed similar expression patterns in incisor development compared with those in molars (data not shown). Lgr5, an epithelial stem cell marker in the intestine and hair follicle, has previously been shown to be expressed in incisor stem cell niche, labial cervical loop, although main body of incisors exhibited no obvious Lgr5 expression (Fig. 3E; data not shown, Barker et al., 2007; Suomalainen and Thesleff, 2010). Because expression of R-spondins and other Lgrs (Lgr4 and Lgr6) in the cervical loop remain unclear, their expression were examined in the labial cervical loop on sagittal sections at E18.5. Weak expression of R-spondin1 was observed in the lingual aspect of the cervical loop (Fig. 3A). R-spondin2 expression was observed in entire incisor developing dental pulp extending to cervical loop area (Fig. 3B). No obvious R-spondin3 expression could be detected at the proximal end of incisors (data not shown). R-spondin4 showed expression in the lingual aspect of the cervical loop and the lingual side of developing dental pulp, while weak expression was also observed in mesenchyme facing the cervical loop (Fig. 3C). Lgr4 was clearly expressed in the labial aspect of the cervical loop (Fig. 3D). Lgr6 expression was observed in mesenchyme facing ameloblasts, which extend to the cervical loop area (Fig. 3F).
The expression of R-spondins and Lgrs during murine tooth development is summarized in Figure 4, Figure 5, Supplementary Tables S1–S3. (Supplementary Tables are available online.) R-spondins and Lgrs thus show dynamic spatio-temporal expression in murine odontogenesis.
Tooth Development in R-spondin2 Mutant Mice
Although craniofacial malformations have been described in R-spondin2 mutant mice, tooth abnormalities were not reported (Nam et al., 2007a; Aoki et al., 2008; Bell et al., 2008; Yamada et al., 2009; Jin et al., 2011). To address this question, we examined teeth from R-spondin2 mutant mice. Frontal sections of mutant heads showed no significant differences in molar or incisor tooth germs (Fig. 6B,D,F). Sagittal sections and 3D reconstruction of tooth epithelium in R-spondin2 mutants, however, showed the presence of supernumerary tooth germs in the diastema of the mandible (8/12), but not in the maxilla (Fig. 6H,I,K, data not shown). Significant changes were not detected in the labial cervical loop of lower incisors in R-spondin2 mutants (Fig. 6M,O).
Expression of R-spondin2 and Lgrs in the Diastema
The formation of extra tooth germs in the diastema prompted us examine R-spondin2 expression in the diastema. The region of the diastema and tooth germs were first identified by expression of tooth marker gene, Pitx2, which was expressed in tooth germs, but was not in the diastema at E12.5 and E13.5 (Fig. 7B,D). At E12.5, R-spondin2 expression was observed at the lingual side of the diastema and region surrounding the tooth region (Fig. 7A,E). At E13.5, the buccal domain of R-spondin2 expression extended into the diastema, while the lingual domain of R-spondin2 expression became weak (Fig. 7C,F). Although R-spondin2 mutants have teeth in the diastema anterior to the first molar, radioactive in situ hybridization on sagittal sections showed no expression of R-spondin2 corresponding to the location of extra tooth formation in R-spondin2 mutants (Figs. 6K,H, and 7G,H). R-spondin2 expression in the diastema is thus slightly distant from (buccaly and lingually) the first molar tooth germs (Fig. 7L). Receptors of R-spondin2, Lgr5, and Lgr6, however, showed expression in the diastema anterior to the first molar tooth germs, while Lgr4 and Lgr6 expression were also found in anterior tooth epithelium (Fig. 7J–L). Unlike R-spondin2, R-spondin1, R-spondin3, and R-spondin4 were expressed in the diastema anterior to the first molar tooth germs (Fig. 8).
We show here dynamic spatio-temporal expression of R-spondins and Lgrs during murine odontogenesis. Although R-spondin2 expression was observed in tooth development, R-spondin2 mutants showed no significant anomalies in molars or incisors. The expression pattern of R-spondin4 is similar to that of R-spondin2 in tooth development. It is thus possible that R-spondin4 compensates for R-spondin2 function in tooth development. On the other hand, R-spondin2 mutants did show extra tooth formation in the diastema, although all R-spondins are expressed in the diastema. These suggest that R-spondin2 inhibits odontogenesis in the diastema and R-spondin2 function in the diastema is not compensated by other R-spondins. No expression of R-spondin2 was observed in the diastema corresponding to the location of extra tooth formation in R-spondin2 mutants, but other R-spondins were expressed in the region. It is conceivable that the role of other R-spondins in the diastema is different from that of R-spondin2.
Down-regulation of Wnt signaling by overexpression of Wnt inhibitor, Dkk1, results in failure of tooth development, while overexpression of Wnt signaling leads to multiple extra teeth formation (Andl et al., 2002, Järvinen et al., 2006). Furthermore, mutations of an inhibitor of Wnt signaling, Wise, show extra tooth formation in the diastema, suggesting that Wnt signaling is required for the initiation of tooth development in the diastema (Kassai et al., 2005; Ohazama et al., 2008; Ahn et al., 2010). R-spondin2 mutants did, however, show extra tooth formation in the diastema, even though significant reduction of Wnt signaling activity has been shown in R-spondin2 mutant mandibles at E10.5 (Jin et al., 2011). Both up-and down-regulation of Eda has been shown to lead to extra tooth formation in the diastema, suggesting that the fine-tuning of signaling pathways is important for inhibiting tooth formation in the diastema (Peterkova et al., 2006; Mustonen et al., 2003).
R-spondins activate Wnt signaling, but they are unable to initiate Wnt signaling and only synergistically activate Wnt signaling with Wnt ligands (Jin et al., 2011). In addition to ligands, R-spondin function is also sensitive to the presence of the Wnt inhibitor, Dkk1 (de Lau et al., 2011). Wnt ligands and Dkks also show dynamic spatio-temporal expression in the diastema (Porntaveetus et al., 2012). The level of Wnt signaling activity in the diastema is thus easily altered by the balance between Wnt ligands, inhibitors, and modulators like R-spondins.
In addition to Wnt signaling, R-spondin2 is also known to be related to other signals in cochlea development (Mulvaney et al., 2013). Changes of the TNF receptor member, Edar results in extra tooth formation in the diastema, and R-Spondin2 has been shown to be involved with another TNF receptor family member, Troy (Fafilek et al., 2013). Changes in Fgf and Shh signaling pathways also lead to the diastema tooth formation, and the expression of Fgfs and Shh are altered in limb development of R-spondin2 mutants (Klein et al., 2006; Aoki et al., 2008; Yamada et al., 2009; Ohazama et al., 2009). It is possible that R-spondin2 regulates tooth development in the diastema through Fgf and Shh.
We also show here dynamic spatio-temporal expression of R-spondins and Lgrs at the proximal end of incisors where contain stem cell niches. R-spondin2 and R-spondin4 are mainly expressed in mesenchyme of the region, whereas their receptors, Lgr4 and Lgr5 are expressed in the epithelial cervical loop. No significant morphological changes of cervical loops was found in R-spondin2 mutants, suggesting that R-spondin2 is functionally dispensable in the incisor epithelial stem cell niche.
Production and Analysis of Transgenic Mice
R-spondin2 mutant mice were generated from the mouse line produced by Yamada et al. (2009). The day on which vaginal plugs were found was considered as embryonic day (E) 0.5. To accurately assess the age of embryos, somite pairs were counted and the stage confirmed using morphological criteria such as relative size of maxillary and mandibular primordia, extent of nasal placode invagination, and the size of limb buds. Mouse heads were fixed in 4% paraformaldehyde (PFA), embedded, and serially sectioned at 8 μm. Sections were split over 4–10 slides and prepared for histology and radioactive in situ hybridization. Decalcification using 0.5M EDTA was performed after fixation of newborn mice.
In Situ Hybridization
Radioactive with 35S-UTP-radiolabeled riboprobes and whole mount with digoxigenin (DIG) or fluorescein-labeled riboprobes in situ hybridization was performed as described previously by Ohazama et al., 2008.
A set of images obtained from serial sections of the embryo under study was aligned with the alignment module contained in DeltaViewer software (version 2.1.1, http://delta.math.sci.osaka-u.ac.jp/DeltaViewer/), which is specialized for automatic alignment and three-dimensional reconstruction from serial sections (Yamada et al., 2007).
We thank Jeong Kyo Yoon for R-Spondin plasmids and, Alasdair Edgar and Alex Huhn for technical supports. Y.O.-K. is supported by Nihon University. M.K. and K.K. are supported by JSPS International Program for Young Researcher Overseas Visits.