The tooth is an advantageous model system for analyzing molecular mechanisms regulating organ formation and cell differentiation (for review, see Peters and Balling,1999; Peterkova et al.,2000; Miletich and Sharpe,2003; Thesleff,2003). The formation of the tooth is controlled by reciprocal signaling between ectoderm-derived epithelium and neural crest-derived mesenchyme. There is increasing evidence for several signaling molecules belonging to fibroblast growth factor, bone morphogenetic protein, Hedgehog, and Wnt families to mediate these interactions (Thesleff and Mikkola,2002; Miletich and Sharpe,2003). Many of the key signals are specifically expressed in the putative epithelial signaling centers, the primary and secondary enamel knots (PEK and SEK, respectively; for a review see Thesleff and Jernvall,1997), which together with recently characterized tertiary enamel knot (TEK) and the ameloblast-free ridge (AFR) are suggested to regulate molar tooth morphogenesis and final crown shape (Luukko et al.,2003).
The Wnts form a family of secreted lipid-modified signaling proteins (>19 members in man), and they serve important roles in morphogenesis, determination of cell polarity, and regulation of cell proliferation and differentiation during embryogenesis (Cadigan and Nusse,1997; Peifer and Polakis,2000; Willert et al.,2003). In vertebrates, Wnt signaling is transduced through at least three different distinct intracellular signaling pathways, including the canonical Wnt/β-catenin pathway, the Wnt/Ca2+ pathway and the Wnt “planar polarity” pathway (Cadigan and Nusse,1997; Kuhl et al.,2000; Nelson and Nusse,2004). The Wnt signaling pathway is highly regulated by extracellular factors, and the antagonists of the pathway can functionally be divided into two classes. The first class, which includes the secreted Frizzled-related protein (sFRP) family, Wnt inhibitory factor-1 (WIF-1), and Cerberus, primarily binds to the Wnt proteins (for a review, see Kawano and Kypta,2003). The second class consists of certain members of the Dickkopf (Dkk) family. These are secreted proteins that modulate Wnt signaling by binding to the LRP5/6 component of the Wnt receptor complex (Glinka et al.,1998; Krupnik et al.,1999; Wu et al.,2000; Brott and Sokol,2002; Mao et al.,2002). To date, the Dkk family comprises four members (Dkk-1 to Dkk-4) and a unique Dkk-3–related protein named Soggy (Glinka et al.,1998; Krupnik et al.,1999).
Several Wnts show overlapping expression in the developing mouse tooth, mostly in the epithelial tissues (Dassule and McMahon,1998; Sarkar and Sharpe,1999; Nadiri et al.,2004). In addition, the Wnt receptor MFz6 and antagonist/agonist MFrzb1 and MFrp2 are expressed in the developing tooth (Sarkar and Sharpe,1999). Wnt signaling is essential for odontogenesis in mouse and man (van Genderen et al.,1994; Andl et al.,2002; Lammi et al.,2004). Dkk1 overexpression in the early oral epithelium results in cessation of tooth formation at the bud stage (Andl et al.,2002), suggesting roles for Wnts in tooth formation in vivo. Dkk3 expression has been shown previously in the dental mesenchyme of embryonic day (E) 12 mouse embryo (Monaghan et al.,1999), but the expression of Dkks during tooth formation has not been investigated systematically. Here, we compare cellular mRNA expression patterns of all known mouse Dkk genes (Dkk1, -2, and -3) before and during initiation of mouse mandibular incisor and molar development, as well as during molar tooth morphogenesis and crown calcification.
Comparison of Dkk-1, -2, and -3 mRNA Expression During the Initiation of Dentition Development
No specific expression of Dkk1, -2, and 3 was seen in the mandibular arches of E9.5 embryos (Fig. 1A–C). At E10, before any morphological sign of tooth development, Dkk1 expression was seen in the oral ectoderm and mesenchyme in the distal area of the mandibular process where the incisor teeth later start to develop (Fig. 1D). Dkk2 expression was restricted to the mesenchymal area where the paired mandibular arch had fused (Fig. 1E). Only a few Dkk3 transcripts were found in the mandibular process mesenchyme, whereas in the maxillary process, Dkk3 expression was observed in the presumptive area of the molar tooth epithelium (Fig. 1F). At the histological onset of tooth formation, at E11.5, Dkk1 and Dkk3 expression continued in the presumptive mandibular incisor mesenchyme and maxillary molar epithelium, respectively (Fig. 1J–O). In addition, Dkk3 appeared in the mandibular incisor epithelium (Fig. 1L). No expression was seen for Dkk1 and Dkk3 in the mandibular molar tooth germ (Fig. 1M,O). Dkk2 was observed in the mesenchyme around the vestibular lamina epithelium (Fig. 1N). The expression domains of Dkk1, -2, and -3 in the E10–E11.5 mandibular processes was also confirmed by three-dimensional reconstructions as visualized in Figure 1G–I,P–R.
Dynamic Expression of Dkk1, -2, and -3 During Morphogenesis and Crown Calcification of the Mandibular First Molar
Dkk1, -2, and -3 showed spatiotemporally regulated expression domains in the mandibular first molar tooth germ during crown morphogenesis and calcification. At the bud stage (E13), Dkk1 transcripts were seen for the first time in the molar tooth germ, in the cells located at the middle, oral side of the invaginated epithelial bud (Fig. 2A). In contrast, Dkk3 was detected in the condensed dental mesenchyme (Fig. 2C), and Dkk2 was prominently expressed in the mesenchyme beneath the oral epithelium next to the tooth germ (Fig. 2B).
At the cap stage (E14), Dkk1 and -2 transcripts were observed in the cervical region of the mesenchymal dental papilla and in the dental follicle, respectively (Fig. 2D,E), whereas Dkk3 mRNAs were specifically expressed in the epithelial PEK, mostly at the buccal side (Fig. 2F). Similar expression domains of the Dkks continued in the late cap stage (E15) and early bell stage (E16) tooth germs. The Dkk1 expression continued in the dental papilla except for the areas next to the inner dental epithelium (Fig. 2G,J). Dkk2 expression in the dental follicle was highest at the lingual side of the tooth germ (Fig. 2H,K), whereas Dkk3 was now observed in the SEKs at the sites of the developing cusps (Fig. 2I,L). Some Dkk3 transcripts were also observed in the dental follicle.
Before birth, at the bell stage (E18), Dkk1 expression in the dental papilla shifted to the cuspal region (Fig. 2M). In particular, the preodontoblasts, which appear under the inner dental epithelium at the tips of the future cusps, showed prominent Dkk1 expression. Similar Dkk2 expression, although weaker was observed in the dental papilla, but preodontoblasts were devoid of transcripts (Fig. 2N). Weak Dkk3 expression continued in the dental follicle mesenchyme around the enamel organ, and some transcripts were also observed in the outer dental epithelium (Fig. 2O).
At postnatal day (PN) 2, intense Dkk1 expression was evident in the preodontoblasts, whereas odontoblasts and dental lamina showed weaker signal (Fig. 2P). Ameloblasts showed transient expression of Dkk3 before the onset of enamel matrix secretion (Fig. 2R,S). Some Dkk3 transcripts were also observed in the dental papilla mesenchyme as well as in the mesenchyme next to the dental lamina (Fig. 2R). Dkk2 expression was restricted to the dental lamina (Fig. 2Q).
Similar expression domains of the Dkks were seen in the E18 incisor when enamel and dentin formation had occurred (Fig. 3A–C). In addition, Dkk1 showed prominent expression in developing mandibular bone during E13–PN2 (Fig. 2A,D,G,J,M,P).
We found that mRNAs for all three mouse Dkks are present in the developing tooth with distinct spatiotemporally regulated expression domains. Dkk1 and -3 expression shifted between epithelium and mesenchyme, whereas Dkk2 expression was restricted to the mesenchymal tissues. Dkk1 is a specific inhibitor of the canonical Wnt signaling pathway, whereas Dkk2 has the ability to act both as an agonist and an antagonist in this pathway (Wu et al.,2000; Brott and Sokol,2002; Mao and Niehrs,2003). The biological role of Dkk3 is unclear but recent studies suggest that it may also act as an inhibitor in Wnt signaling (Caricasole et al.,2003; Hoang et al.,2004). Thus, all three Dkks are putative regulators of Wnt signaling during tooth formation.
Possible Roles for Dkk1 in Dental Patterning
Mammalian dentition is patterned; the incisor teeth develop in the distal region of the mandibular process, whereas the molars are located in the proximal part. The restricted expression of Dkk1 in the incisor region suggests that Dkk1 signaling may function in dental patterning. These expression patterns correlate with the expression of genes such as Bmp4, Islet1, and Msx1, which all mark the developing incisor region and appear to regulate dental patterning (Tucker et al.,1998; Mitsiadis et al.,2003). In support for this function, Dkk1 controls distal limb patterning, and its expression can be activated by Bmp4 (Mukhopadhyay et al.,2001; Grotewold and Ruther,2002).
Possible Roles of Dkks in Tooth Morphogenesis
During molar morphogenesis Dkk3 was specifically expressed in the PEK and SEK signaling centers (Thesleff and Jernvall,1997). The PEK is essential for initiation of molar-specific folding morphogenesis (Kratochwil et al.,2002), whereas SEKs control the number and patterning of the cusps (Jernvall et al.,1994). Experimental and genetic evidence indicates that Wnt-signaling regulators, such as Lef1, and Axin2 expressed in the enamel knot serve essential roles in tooth morphogenesis (van Genderen et al.,1994; Sarkar and Sharpe,2000; Kratochwil et al.,2002; Lammi et al.,2004). Thus, modulation of Wnt pathway in the epithelial signaling centers may serve important regulatory roles in tooth morphogenesis. The expression of Dkks in the dental mesenchyme suggests that combinatorial Dkk expression domains may also contribute in defining the spatiotemporal effects of Wnt signaling in the tissue.
Possible Roles for Dkks in Crown Calcification
The differentiation of amelo- and odontoblasts and subsequent dentin and enamel formation is regulated by epithelial–mesenchymal interactions between the dental papilla and inner dental epithelium cells and involves the activity of signaling molecules (for review, see Lesot et al.,2001). We found that in addition to being expressed in the developing mandibular bone, Dkk1 was prominently up-regulated in the preodontoblasts and the expression continued in the secretory odontoblasts. In contrast, Dkk3 was transiently expressed in the ameloblasts before enamel matrix secretion. There is evidence that Wnt signaling and its modulation by Dkks is involved in bone formation (Kato et al.,2002; Rawadi et al.,2003; Akiyama et al.,2004; MacDonald et al.,2004). In addition, overexpression of Wnt3 in dental epithelium results in the loss of ameloblasts and a reduction of enamel in postnatal incisors, suggesting a role for the Wnt pathway in enamel formation (Millar et al.,2003). Thus, Dkks may be involved in differentiation of the amelo- and odontoblast cell lineages and subsequent dental hard tissue formation. In summary, the dynamic expression of Dkk1, -2, and -3 during tooth development indicates that the modulation of Wnt-signaling by Dkks may serve important functions during dental patterning and tooth formation.
MATERIALS AND METHODS
Preparation of Tissues
NMRI mice were mated overnight, and the appearance of a vaginal plug was taken as day 0 of embryonic development (E0). The animals were killed by cervical dislocation or decapitation. Heads (E9.5–E13) and lower jaws ([E14–E18] and 2-day postnatal [PN2]) were fixed in 4% paraformaldehyde (PFA). After fixation, the E14.5–PN2 tissues were decalcified in 12.5% ethylenediaminetetraacetic acid and 2.5% PFA in phosphate-buffered saline. The tissues were embedded in paraffin, and cut into 7-μm frontal sections. This study was approved by the Animal Welfare Committee of the Department for Biomedicine, University of Bergen.
Dkk Probes for In Situ Hybridization
Three Dkk probes were generated by reverse transcriptase-polymerase chain reaction (RT-PCR) from total RNA isolated from E10 to E13 mouse heads and subcloned into the pGEM-T Easy Vector (Promega Corporation, Madison, WI): a 692-bp Dkk1 fragment spanning the region between 152 and 843 in accession no. AF030433; a 528-bp Dkk2 fragment spanning the region between 969 and 1496 in accession no. NM_020265; and a 590-bp Dkk3 fragment spanning the region between 464 and 1053 in accession no. NM_015814. The following primer oligonucleotides were used: Dkk1 (forward 5′-ATT CCA ACG CGA TCA AGA AC-3′ and reverse 5′-GCA GGT GTG GAG CCT AGA AG-3′), Dkk2 (forward 5′-ACC CTT GCA GCA GTG ATA AG-3′ and reverse 5′-TGG CTT TGG AAG AGT AGG TG-3′), and Dkk3 (forward 5′-ACC AGA GTG GAC AGG TGG TC -3′ and reverse 5′-CAC TTC CCC TAT GAA GCC AA-3′).
In Situ Hybridization
In situ hybridization on sections was performed as described in Luukko et al., 1996. 35S-UTP–labeled antisense and sense cRNA probes were generated by in vitro transcription. No specific signal was detected in sections hybridized with the control sense probes (not shown). Bright- and darkfield images of the in situ sections were photographed by using a Zeiss Axioskop 2 microscope (Carl Zeiss Jena GmbH, Jena, Germany) and a SPOT Insight digital camera (Diagnostic Instruments, Sterling Heights, MI). The figures were prepared with Adobe Photoshop 6.0 program (Adobe Systems, San Jose, CA).
Three-Dimensional Reconstruction of Gene Expression
Three-dimensional (3D) reconstruction of the mandibular processes was generated from 8-μm serial frontal bright- and darkfield images (approximately 70 sections from each field). The processing of the images was done using custom-made scripts and programs made with Java Advanced Imaging and Java 3D (Sun Microsystems, CA; http://java.sun.com). The 3D reconstructions were rendered with perspective camera view in Visualization Toolkit (Kitware, Clifton Park, New York; http://www.kitware.com). Approximately 70% transparent 3D surface was generated using Marching Cubes function in Visualization Toolkit from the outlines of the jaw epithelium. Hybridization signal with intensity over 190 was considered to represent positive gene expression. The sections were median-filtered to reduce the background hybridization signal in the 3D image.
We wish to thank Ms. Kjellfid Haukanes, Ms. Helen Olsen and Ms. Anne Nyhaug for their skillful technical assistance. This study was supported by the Norwegian Cancer Society, the L. Meltzer's Foundation, the University of Bergen, and the Research Council of Norway.