The morphogenesis of most organs is governed by interactions between epithelial and mesenchymal tissues. The vertebrate teeth develop as a consequence of inductive interactions between the stomodeal epithelium and underlying neural crest-derived ectomesenchyme. The presumptive dental epithelium governs the initiation of tooth development, and subsequently the dental mesenchyme regulates the morphogenesis of the epithelial enamel organ (Kollar and Baird, 1970; Mina and Kollar, 1987; Lumsden, 1988). During the bell stage, tooth shape is established by epithelial folding morphogenesis, and the dentin and enamel forming odontoblasts and ameloblasts, respectively, differentiate, also as a consequence of tissue interactions (for review, see Thesleff and Sharpe, 1997). Conserved signaling molecules belonging to the FGF (fibroblast growth factor) Hh (hedgehog), BMP (bone morphogenetic protein), and Wnt families mediate inductive signaling during embryogenesis of different animals including Drosophila and vertebrates (Hogan, 1999). These signaling molecules are also involved in the regulation of tooth development, and during recent years the molecular details of these signaling networks have started to be elucidated (Vainio et al., 1993; Satokata and Maas 1994; Chen et al., 1996; Kratochwill et al., 1996; Neubüser et al., 1997; Bei and Maas, 1998; Ferguson et al., 1998; Hardcastle et al., 1998; Dassule and McMahon, 1998; Tucker et al., 1998; St. Amand et al., 2000). There is increasing evidence that different signaling pathways interact, and it is apparent that specific combinations of signals and transcription factors determine the outcome of the inductive interactions in tooth development (Bei and Maas, 1998). It is of interest that signals in all four families (FGF, BMP, Shh, Wnt) show striking coexpression in the dental epithelium at specific stages of development, further suggesting interactions between the pathways (for review, see Jernvall and Thesleff, 2000).
FGFs have important regulatory functions in the formation of numerous organs such as limb, midbrain, lung, hair, pituitary gland, and feather (Hebert et al., 1994; Peters et al., 1994; Cohn et al., 1995; Crossley et al., 1996a,b; Widelitz et al., 1996; Bellusci et al., 1997; Ericson et al., 1998; Celli et al., 1998; Min et al., 1998; Takuma et al., 1998; Martinez et al., 1999; Sekine et al., 1999; De Moerlooze et al., 2000). In mammals, the FGF family consists of 19 structurally related growth factors at present (for review, see Wilkie et al., 1995; Yamasaki et al., 1996; Smallwood et al., 1996; McWhirter et al., 1997; Miyake et al., 1998; Hoshikawa et al., 1998; Ohbayashi et al., 1998; Nishimura et al., 1999), and their biological effects are mediated through four high-affinity tyrosine kinase receptors (FGFR1–4) (for review, see Wilkie et al., 1995), three (FGFR1–3) of which produce IIIb and IIIc RNA isoforms through alternative splicing (for review, see Johnson and Williams, 1993) with different ligand-binding properties (Ornitz et al., 1996).
FGFs were first implicated in tooth development when intense and restricted Fgf-4 expression was observed in the epithelium of cap stage tooth (Niswander and Martin, 1992). This cell population was identified as the enamel knot, and it was suggested to act as a signaling center regulating tooth shape (Jernvall et al., 1994). Subsequently, Fgf-9, Shh, and several members of the Bmp and Wnt families have been shown to be expressed in the enamel knot, supporting the role of the enamel knot as a signaling center (Bitgood and McMahon, 1995; Vaahtokari et al., 1996; Dassule and McMahon, 1998; Kettunen and Thesleff, 1998; Sarkar and Sharpe, 1999). Fgf-4 was shown to be expressed also in the secondary enamel knots appearing during bell stage at the sites of tooth cusps, and FGF-4 was suggested to regulate the growth of the cusps (Jernvall et al., 1994). FGF-8, on the other hand, has been implicated as an early epithelial signal during tooth initiation (Jernvall and Thesleff, 2000; Peters and Balling, 1999; Tucker and Sharpe, 1999). It regulates the mesenchymal expression of a number of key transcription factors that are associated with the acquisition of odontogenic potential in the mesenchyme (Neubüser et al., 1997; Bei and Maas, 1998; Ferguson et al., 1998).
To further evaluate the roles of FGFs in tooth development, we analyzed the roles of FGF-3 (int-2), FGF-7 (KGF) and FGF-10 in developing mouse teeth. Fgf-3 transcripts have been localized in developing organs including teeth (Wilkinson et al., 1988, 1989; Thesleff et al., 1990), and it is needed for middle air and tail development (Mansour et al., 1993). Fgf-7 is expressed more widely than Fgf-3 during development (Mason et al., 1994; Finch et al., 1995) and is needed for normal hair development (Guo et al., 1996). Fgf-10 was found by sequence homology to Fgf-3 and Fgf-7 (Yamasaki et al., 1996), and it is involved in pancreas and prostate development (Miralles et al., 1999; Thomson and Cunha, 1999) and necessary for limb, lung, and external genitalia development (Bellusci et al., 1997; Ohuchi et al., 1997; Xu et al., 1998; Min et al., 1998; Sekine et al., 1999; Haraguchi et al., 2000). Our results suggest that FGF-3 and FGF-10 have redundant functions as mesenchymal signals regulating epithelial morphogenesis of the tooth and that their expressions are differentially regulated. In addition, FGF-3 may participate in signaling functions of the primary enamel knot. The dynamic expression patterns of different Fgfs in dental epithelium and mesenchyme and their interactions suggest existence of regulatory signaling cascades between epithelial and mesenchymal FGFs during tooth development.
MATERIALS AND METHODS
Preparation of Tissues for In Situ Hybridization
Mice (CBA × NMRI) were mated overnight, and the appearance of the vaginal plug was taken as day 0 of embryogenesis (EO). The developmental stage of the teeth was also confirmed by morphological criteria. Tissues were fixed in 4% paraformaldehyde in PBS (pH 7.2) for 24–48 hr at +4°C, dehydrated, embedded in paraffin, and serially sectioned at 7 μm. Sections were placed on triethoxysilane- and acetone-treated slides, dried overnight at +37°C, and stored at +4°C until used.
RNA In Situ Hybridization
The RNA probes were synthesized from a murine 500-bp Fgf-3 genomic DNA fragment coding for a 3′-untranslated sequence inserted into the pGEM-4 vector (Promega, Madison, WI), from a rat 750-bp Fgf-10 cDNA fragment inserted into the pGEM-T vector (Promega), from a mouse 584-bp Fgf-10 cDNA fragment inserted into the Bluescript KS II+ vector (Stratagene, La Jolla, CA), from a mouse 202-bp Fgf-7 cDNA fragment inserted into the Bluescript KS II+ vector (Stratagene), from a rat 2.6-kb Shh cDNA fragment inserted into the Bluescript SK+ vector (Stratagene), and from a murine 600-bp Lef-1 cDNA fragment (3′-truncation of the GLI clone at NdeI site) inserted into the Bluescript (Stratagene). Radioactive in situ hybridization procedures were conducted as described by Wilkinson and Green (1990), with some modifications (Luukko et al., 1996). Nonradioactive whole-mount in situ hybridization was performed principally as described by Henrique et al. (1995). Brightfield and darkfield illuminations were digitized by using an Olympus BX-50 microscope (Olympus, Tokyo, Japan), Cohu 4912-5000 CDD video camera (Cohu, San Diego, CA) and Macintosh computer (Apple Computer, Cupertino, CA). Figures were processed by using NIH Image 1.61 (public domain program, U.S. National Institutes of Health) and Adobe Photoshop 5.0 (Adobe Systems, San Jose, CA).
Tissue Recombination and Bead Experiments
The regions of the presumptive molar teeth were carefully dissected from the mandibles of E11–14 embryos, incubated for 5 min in 0.75% pancreatin (GibcoBRL), 2.25% trypsin (Difco) in room temperature, and dental mesenchymes and epithelia were separated from each other with 25-gauge sterile needles (Terumo Corporation, Leuven, Belgium) under a stereomicroscope after a short regeneration time. Isolated dental epithelia and mesenchymes were cultured on pieces of polycarbonate Nuclepore filters (pore size 0.1 μm; Costar, Cambridge, MA) supported by stainless steel grids in Trowell-type cultures in Dulbecco's minimum essential medium (DMEM) (Sigma, St. Louis, MO) supplemented with 10% fetal bovine serum (GibcoBRL, Paisley, Scotland), penicillin 10 IU/ml and streptomycin 10 mg/ml (GibcoBRL). A drop (1 μl or 5 μl) of Matrigel (Collaborative Biochemical Products, Bedford, MA) with reduced amount of growth factors was added on the dental epithelia when they were cultured alone >24 hr. In some cases, the dental mesenchymes were recombined and cultured in contact with the dental epithelia. For delivery of FGF recombinant proteins and BSA heparin acrylic beads (Sigma) were used. For delivery of BMP, TGF-β1, Shh, and BSA, Affi-Gel blue agarose beads (100–200 mesh, 75–150-μm diameter; Bio Rad Laboratories, Hercules, CA) were used. BSA was used at a concentration of 1 mg/ml. FGF-4 and -8 (b form) (R&D Systems, Abingdon, UK) were used at concentrations of 25, 50, and 100 μg/ml, FGF-10 at a concentration of 72 and 120 μg/ml, and BMP-2 (a kind gift from Genetics Institute, Cambridge, MA) and Shh (a kind gift from Ontogeny, Cambridge, MA) at concentrations of 75 μg/ml and 600 μg/ml, respectively. TGFβ1 (Boehringen-Mannheim, Mannheim) was used at a concentration of 1 μg/ml. Beads were washed in phosphate-buffered saline (PBS) and incubated with the proteins for 30 min at 37°C. After culture in a humidified atmosphere of 5% CO2 in air at +37°C, the explants were washed in ice-cold 100% methanol and fixed in 4% paraformaldehyde (PFA). Cultures were then dehydrated, embedded in paraffin, and serially sectioned for section in situ hybridization or 5-bromo-2′-deoxyuridine (BrdU) immunostaining.
Wnt-6 Cell Line
NIH3T3 cell line expressing Wnt-6 gene was prepared essentially as described in Pear et al., 1993. The coding region of Wnt-6 cDNA (Gavin et al., 1990) was cloned into the retroviral expression vector pLNCX with CMV promoter (Miller and Rosman, 1989) and then transfected into Bosc23 packaging cells (Fugene reagent; Boehringer-Mannheim, Mannheim, Germany) (Pear et al., 1993). The viral supernatant of these cells was harvested and used to infect NIH3T3 cells, which were then placed under neomycin (1 mg/ml) selection for 3 weeks. NIH3T3 cell line expressing no gene was used as control cell line. A clump of Wnt-6 or control NIH3T3 cells were put on top of dental mesenchymes and cultured and fixed as described above. Induction of Lef-1 expression by Wnt-6 NIH3T3 cells in the dental mesenchyme (18 of 18 positive) was used as a positive control to indicate that the cells worked. Control cells did not induce Lef-1 (18 of 18 negative). In addition, Fgf-10 sense probe was used as a control (18 of 18 negative). Earlier Wnt10b NIH3T3 cells have been reported to induce Lef-1 expression in the dental mesenchyme (Dassule and McMahon, 1998). Induction of Lef-1 and Fgf-10 was analyzed by using whole-mount in situ hybridization (see above).
Cell Proliferation Assay
The effect of FGF-10 on dental cell proliferation was analyzed by labeling the explants with BrdU and 5-fluoro-2′-deoxyuridine (FdU) (Amersham, Buckinghamshire, UK) after 24- or 72-hr culture. The labeling reagent was diluted according to the manufacturer's instructions (1 μl labeling reagent to 1,000 μl culture media), and the tissues were incubated for 1 or 4 hr in the culture medium, submerged in cold 100% methanol for 10 sec, and fixed with 4% PFA. Some of the explants were analyzed as whole mounts and some as sections. After PBS washes, the incorporated BrdU was detected by indirect immunoperoxidase method with monoclonal antibodies against BrdU (Amersham, Burlingame, CA) and biotinylated secondary antibody (Jackson ImmunoReseach laboratories, West Grove, PA). For color reaction the Vectastain ABC Elite Kit (Vector Laboratories, Burlingame, CA) and 3-amino-9-ethyl carbazol (AEC) or 3,3′diaminobenzidine (DAB) (Vector Laboratories, DAKO) as substrate were used.
Localization of Fgf3 mRNAs During Molar Tooth Development
The expression of Fgfs-3, -7, and -10 mRNAs were analyzed from the initiation of the first molar formation (E10) to the completion of crown morphogenesis (P2) from serial sections using in situ hybridization. Expression of Fgf-3 has been earlier reported in the mouse tooth (Wilkinson et al., 1989; Thesleff et al., 1990; Bei and Maas, 1998; Harada et al., 1999), but its expression patterns have not been comprehensively studied during tooth formation. No specific Fgf-3 hybridization signal was detected in the dental tissues during the initiation and early bud stage of tooth development (E10–E12). At the late bud stage (E13), a weak expression appeared at the tip of the epithelial bud and in the adjacent condensed dental mesenchyme (Fig. 2A). During the cap stage (E14–E15), Fgf-3 was expressed in the dental papilla mesenchyme and in the enamel knot (Fig. 2B). At the subsequent bell stage (E16–E18), transcripts were restricted to the cuspal regions of the dental papilla mesenchyme including preodontoblasts and differentiating odontoblasts (Fig. 2C). Transcripts were not seen in the secondary enamel knots. Postnatally, the expression of Fgf-3 was down-regulated in the developmentally advanced cuspal regions of the dental papilla, and expression shifted gradually to the less differentiated cervical regions of the dental papilla mesenchyme (Fig. 2D). No transcripts were seen in terminally differentiated odontoblasts secreting predentin matrix (Fig. 2D).
Localization of Fgf-10 mRNAs During Tooth Development
Fgf-10 transcripts were observed in the region of the presumptive dental epithelium and in the adjacent dental and jaw mesenchyme before any morphological sign of tooth formation at E10 (Fig. 1A). Subsequently, Fgf-10 expression disappeared from the dental epithelium, and only a weak expression was observed in the mesenchyme next to the dental epithelium (Fig. 1B,C). At the bud stage (E13), Fgf-10 transcripts were largely restricted to the mesenchymal cells under to the oral epithelium(Fig. 2E). At the cap stage (E14–E15), an intense Fgf-10 expression was observed in the mesenchymal cells of the dental papilla and follicle (Fig. 2F). Later at the bell stage (E16–E18), Fgf-10 expression persisted in the dental papilla cells, but expression was no longer detected in the dental follicle (Fig. 2G). Some transcripts were also observed in the mesenchymal cells lining the dental lamina (data not shown). Postnatally, Fgf-10 mRNAs were gradually down-regulated in the dental papilla mesenchyme. However, expression remained in the mesenchymal cells next to the dental epithelium, whereas (Fig. 2H) terminally differentiated odontoblasts secreting predentin matrix showed no hybridization signal (Fig. 2H).
Localization of Fgf-7 During Tooth Development
Expression of Fgf-7 transcripts was observed in the developing bone surrounding the developing tooth germ. In addition, Fgf-7 expression was seen in the developing muscles of the first branchial arch. Hence, the expression of Fgf-7 appeared not to be developmentally regulated during molar tooth morphogenesis.
Localization of Fgf-3,Fgf-7, and Fgf-10 mRNAs in Incisor Tooth Germs
The expression patterns of Fgf-3 and Fgf-10 in the developing incisor tooth germs correlated to that in molars (Fig. 3 and data not shown). Unlike mouse molars, the incisors grow continuously, and expression of both Fgf-3 and Fgf-10 was detected in the dental papilla mesenchyme adjacent to the epithelial crevical loop (Fig. 3A,B). Secretory odontoblasts did not express Fgf-3 and Fgf-10. However, Fgf-10 transcripts were also present in the mesenchymal cells next to the tip of the cervical loop (Fig. 3B). Transcripts of Fgf-7 were seen in mesenchyme next to the outer dental epithelium and cervical loop (Fig. 3C). In addition, developing bone showed Fgf-7 expression (Fig. 3C).
FGF-10 Stimulates Cell Proliferation in Dental Epithelium But Not in Dental Mesenchyme
We have earlier shown that FGF-4 expressed in the epithelial primary and secondary enamel knots can trigger cell proliferation in the isolated dental mesenchyme and epithelium of the cap stage tooth (Jernvall et al., 1994; Kettunen et al., 1998). In addition, other epithelially expressed FGFs (FGF-8 and –9) can stimulate cell proliferation in isolated dental mesenchymes during E11–E14 (Kettunen et al., 1998). To investigate the effects of FGF-10 on cell proliferation, beads soaked in recombinant human FGF-10 were placed on isolated E12–E14 dental epithelia and mesenchymes. Dental epithelia were cultured for 72 hr, the last 4 hr with BrdU and FdU, and analyzed morphologically and then immunohistochemically for BrdU incorporation. Dental mesenchymes were cultured for 24 hr and the last hour with BrdU and FdU. It is of interest that FGF-10 beads enhanced cell proliferation in isolated dental epithelia (Fig. 4B, 30 explants). The dental epithelia cultured with FGF-10 beads (Fig. 4B) were remarkably bigger in size than those cultured with BSA beads (Fig. 4A, 25 explants). Immunohistochemical analysis showed that cells next to the FGF-10 beads incorporated BrdU in E14 dental epithelia (Fig. 4D, 6 explants, 6 of 6 positive). No BrdU incorporation was seen in cells next to the BSA beads in E14 dental epithelia (Fig. 4C, 5 explants, 5 of 5 negative). Our earlier studies have indicated that beads releasing FGFs-4, -8, and -9 induce a translucent zone in the surrounding mesenchyme after a 24-hr culture and that BrdU incorporation is increased in this area (Kettunen and Thesleff, 1998; Kettunen et al., 1998). However, neither translucent zone nor increased BrdU incorporation was seen around FGF-10 releasing beads in the dental mesenchymes (data not shown). Hence, the dental mesenchyme appeared not to have mitogenic response to FGF-10. These results show that FGF-10 promotes epithelial growth by enhancing cell proliferation and perhaps also survival.
FGFs-4 and -8 Stimulate the Expression of Fgf-3 But Not Fgf-10 in the Dental Mesenchyme
The role of epithelial genes as regulators of Fgf-3 and Fgf-10 expression was analyzed in tissue recombination cultures of E13 and E14 dental epithelia and mesenchymes. After 24 hr of culture, Fgf-3 was induced in dental mesenchyme next to the E13 dental epithelium (2 of 2 positive result) (Fig. 5D). However, Fgf-10 was induced in only one of three mesenchymes in E13 recombinant cultures (not shown). At E14, Fgf-10 was induced in 2 of 2 explants (Fig 5E). Fgf-4, -8, and -9, which are expressed in the dental epithelium, stimulate mesenchymal cell proliferation and the expression of several transcription factors shown to be needed for proper tooth development (Kettunen and Thesleff, 1998; Kettunen et al., 1998; for review, see Peters and Balling, 1999). To study their ability to regulate the expression of Fgf-3 and -10 mRNAs, recombinant FGF-4 and -8 proteins were placed on isolated E11–E13 dental mesenchymes. After a 24-hr culture, the expression of Fgf-3 and Fgf-10 was analyzed by radioactive in situ hybridization in serial tissue sections. It is of interest that FGF-4, and -8 induced Fgf-3 expression in the dental mesenchyme (Fig. 5A,B), whereas Fgf-10 expression was not induced (Fig. 5G). BSA-soaked beads had no effect on Fgf-3 or Fgf-10 expression (Fig. 5G,H). We also tested the effects of other potential epithelial signaling molecules on the expression of Fgf-3 and Fgf-10 in dental mesenchyme. Bmp-2 and Shh are expressed intensely in the enamel knot (Vaahtokari et al., 1996) and Tgf-β1 in the epithelial cervical loops (Vaahtokari et al., 1991). When applied with beads on dental mesenchyme, all three failed to stimulate the expression of Fgf-3 and Fgf-10 (Table 1). Several members of the Wnt-signaling family are expressed in the dental epithelium, and hence, they are putative regulators of mesenchymal expression of Fgf-10 (Dassule and McMahon, 1998; Sarkar and Sharpe, 1999). In addition, Wnt3 has been shown to maintain Fgf-10 expression in chicken limb (Kengaku et al., 1998). However, Wnt-6-producing NIH3T3 cells did not induce Fgf-10 expression when applied on the dental mesenchyme. Furthermore, the effects of FGF-10 on the expression of Shh in dental epithelium were analyzed. Locally applied FGF-10 had no effects on Shh expression. (Table 1).
Table 1. Induction of Fgf-3 and Fgf-10 Expression in the Dental Mesenchyme by the Dental Epithelium and Signaling Molecules
Dynamic Expression Patterns of Fgf-3 and Fgf-10 Suggest Functions During Several Stages of Tooth Morphogenesis
Our detailed in situ hybridization analysis using serial sections revealed spatiotemporally changing expression patterns for Fgf-3 and Fgf-10 mRNAs during tooth formation. In contrast, Fgf-7 was expressed in the developing bone surrounding the developing molar tooth germ. In the incisor, Fgf-7 expression was seen in the mesenchyme adjacent to the epithelial cervical loop. Fgf-3 and Fgf-10 showed predominantly mesenchymal expression patterns except for transient epithelial expressions. The localization of both Fgfs in dental papilla mesenchyme was associated with the active phase of tooth morphogenesis from cap to early bell stage, which involves rapid growth, and epithelial folding morphogenesis. During postnatal development, Fgf-3 and Fgf-10 mRNAs were detected in the coronal region of the dental papilla, whereas secreting odontoblasts showed no transcripts. These patterns suggest roles for Fgf-3 and Fgf-10 in the regulation of epithelial morphogenesis and cell differentiation.
FGFs-3 and -10 May Act as Autocrine Signals in the Dental Epithelium and Have Redundant Functions With FGFs-4, -8, and -9
Fgf-3 and Fgf-10 were transiently expressed in dental epithelium: Fgf-10 in the presumptive dental epithelium at E10 and Fgf-3 in the primary enamel knot at E14. In these locations they are coexpressed with the epithelially restricted Fgfs, namely Fgf-8 and -9 during initiation, and Fgf-4 and -9 in the enamel knot (Kettunen and Thesleff, 1998). Earlier results have indicated that the epithelial FGFs, together with the signals from other families such as Wnt, BMP, and Hh act primarily as paracrine signals on the dental mesenchyme (for review, see Peters and Balling, 1999; Tucker and Sharpe, 1999; Jernvall and Thesleff, 2000). During tooth initiation, the epithelial FGFs upregulate the expression of essential genes for tooth formation in the dental mesenchyme such as Msx-1, Pax-9, and Activin βA (Satokata and Maas, 1994; van Genderen et al., 1994; Peters et al., 1998; Matzuk et al., 1996; Chen et al., 1996; Neubüser et al., 1997; Ferguson et al., 1998; Bei and Maas, 1998; Kettunen and Thesleff, 1998). The epithelial FGFs, however, may exert effects also on dental epithelium, and we previously showed that FGF-4 stimulates the proliferation of isolated dental epithelium and mesenchyme of the cap stage tooth (Jernvall et al., 1994). Furthermore, receptors for FGF-3 and FGF-10, namely FGFR1b and FGFR 2b, are present in the dental epithelium and, indeed, we showed that FGF-10 stimulates the proliferation of the dental epithelium in culture. Hence, we suggest that FGF-10 and FGF-3 expressed in the epithelium during initiation and in the primary enamel knot, respectively, may act as autocrine regulators of dental epithelium, and that their functions may be redundant with the epithelially restricted FGFs-4, -8, and -9.
FGF-10 and FGF-3 May Be Redundant Mesenchymal Signals Regulating Epithelial Morphogenesis
Fgf-3 and Fgf-10 were also intensely expressed in the dental papilla mesenchyme during the cap stage, when the tooth grows rapidly and the epithelium undergoes folding morphogenesis. Previously, primary enamel knot markers FGF-4 and -9 have been suggested to regulate epithelial morphogenesis. The enamel knot cells do not express FGF receptors; they remain nonproliferative and undergo apoptosis in the distal part of the enamel knot (Jernvall et al., 1998). On the basis of our present results we suggest that FGF-10 and possibly FGF-3 (we did not have access to FGF-3 protein) have redundant functions in the regulation of tooth morphogenesis at the cap stage by stimulating epithelial cell proliferation in the cervical loops. This is supported by the findings that the expression of the tyrosine kinase receptors for FGF-10 and FGF-3, FGFR1b and FGFR2b (Mathieu et al., 1995; Igarashi et al., 1998; Lu et al., 1999) are restricted to the dental epithelium which is site of intense cell division during tooth morphogenesis (Kettunen et al., 1998). The redundant roles for FGF-3 and -10 in tooth formation is further supported by the observations that tooth morphogenesis is arrested in transgenic mice expressing a soluble dominant-negative form of FGFR2b and in knockouts of FGFR2b (Cell et al., 1998; De Moerlooze et al., 2000), whereas no major tooth defects have been reported in FGF-3- and FGF-10-deficient mice (Mansour et al., 1993; Min et al., 1998; Sekine et al., 1999). The schematic Figure 6 depicts the proposed associations between epithelial and mesenchymal FGFs and their receptors in the cap stage tooth germ.
Fgf-3 and Fgf-10 Are Both Associated With Epithelial-Mesenchymal Interactions, But Differentially Regulated by Epithelial FGFs
During epithelial budding (E11–E12), the tooth-forming capacity shifts to the dental mesenchyme (Mina and Kollar, 1987; Lumsden, 1988), and this is associated with the upregulation of a number of mesenchymal genes associated with signaling networks (Peters and Balling, 1999; Jernvall and Thesleff, 2000). Among these are Msx-1, Pax-9, and Lef-1 transcription factors that are regulated by epithelial FGF, BMP, and Wnt signals (Dassule and MacMahon, 1998; Bei and Maas, 1998). Subsequent morphogenesis of the tooth from the bud to cap stage involves the induction of the enamel knot probably by mesenchymal signals, and thereafter the dental mesenchyme continues to instruct epithelial morphogenesis (Kollar and Baird, 1970). BMP-4 is a putative mesenchymal signal (Vainio et al., 1993) for the induction of the enamel knot, and BMP-4 beads can induce the enamel knot markers p21 and Msx-2 in dental epithelium (Jernvall et al., 1998). Furthermore, the analysis of tooth development in Msx-1-deficient mice, in which tooth development arrests at the bud stage, has indicated that Msx-1 is required for the expression of Bmp-4 and Fgf-3 in the mesenchyme, and furthermore that tooth development in the mutants can be partially rescued by BMP-4 (Chen et al., 1996). The appearance of Fgf-3 transcripts in the dental mesenchyme at the late bud stage suggests that FGF-3 is a mesenchymal signal acting on epithelium and that its expression is regulated by epithelial signals. Indeed, our tissue recombination experiments showed that dental epithelium can at least maintain and possibly induce mesenchymal Fgf-3 expression in culture.
Although the expression of both Fgf-3 and Fgf-10 was shown to depend on dental epithelium (Fig. 5D,E), our bead implantation studies showed that they are differentially regulated in the early dental mesenchyme. FGF-4 and -8 induced Fgf-3 expression in E11–E13.5 mesenchymal cells around the beads (Fig. 5A,B), whereas Fgf-10 expression was not stimulated (Fig. 5G). Other signals expressed in the early dental epithelium, including BMP-2, TGFβ1, and Shh, had no stimulatory effects on either Fgf-3 or Fgf-10 expression in the dental mesenchyme. Fgf-10 expression was neither induced by Wnt6 producing NIH3T3 cells. Recently, during the course of our studies, Bei and Maas (1998) also showed that FGFs (FGF-1, -2, and -8) but not BMP-4 regulate Fgf-3 expression in dental mesenchyme. In addition, they showed that mesenchymal Fgf-3 expression depends on Msx-1 like that of Bmp-4, suggesting that the FGF and BMP signaling pathways are integrated in the dental mesenchyme at the level of Msx-1. The ability of FGF-4 and -8 to induce Fgf-3 in dental mesenchyme indicates that there is a regulatory cascade between epithelial and mesenchymal FGFs in the tooth as has been shown in limb development (Xu et al., 1998). Fgf-10 was not regulated by FGFs, Shh, BMP-2, TGFβ1, or Wnt6, and hence, the putative signaling factor or combination of signals regulating Fgf-10 expression in tooth remain(s) open.
FGF-3 and FGF-10 May Regulate the Differentiation of Tooth-Specific Cells
The differentiation of odontoblasts and ameloblasts, i.e., cells producing the tooth-specific extracellular matrices dentin and enamel, respectively, takes place at the interface between the epithelium and mesenchyme, and their differentiation is regulated by interactions between the two cell lineages (for review, see Thesleff and Hurmerinta, 1981; Ruch, 1987). During the advanced bell stage, i.e., the stage of terminal cell differentiation, Fgf-3 and Fgf-10 expression continued in the dental mesenchymal cells, in preodontoblasts, and in postmitotic odontoblasts beneath the inner dental epithelium; therefore, they showed close association with cell differentiation at the epithelial-mesenchymal interface. Expression was not detected in the odontoblasts secreting predentin matrix. This down-regulation of Fgf-3 and Fgf-10 expression in the odontoblasts correlates with their maturation. In addition, the down-regulation of Fgf-3 and Fgf-10 was associated with the last cell divisions in the neighboring inner dental epithelial cells differentiating into ameloblasts. Hence, we suggest that FGF-3 and FGF-10 may be important regulators of epithelial cell proliferation during tooth morphogenesis. In addition, down-regulation of Fgf-3 and Fgf-10 expression may be needed for the maturation of odontoblasts and/or cessation of epithelial cell proliferation and subsequent terminal differentiation of ameloblasts. We recently showed that FGF-10, which is expressed in the mesenchyme of continuously growing incisors (see Fig. 3B), may regulate ameloblast differentiation already at the level of stem cells (Harada et al., 1999). It stimulates the proliferation of ameloblast progenitors in the cervical loops of incisors and also upregulate expression of Lunatic Fringe, a modulator of Notch signaling. Hence, FGF-10 may actually couple cell fate decisions and proliferation in the dental epithelial stem cells (Harada et al., 1999).
We thank David Wilkinson for Fgf3, Sabine Werner for Fgf-7, Saverio Bellusci for Fgf-10, Thomas Edlund for Shh, and Rudolf Grosschedl for Lef-1 probe. The BMP protein was a kind gift from Genetics Institute, Inc. (Cambridge, MA). Shh was a kind gift from Ontogeny, Inc. (Cambridge, MA). The expert technical assistance of Kaija Kettunen, Merja Mäkinen, and Riikka Santalahti is gratefully acknowledged.