Biomineralization, Life-Time of Odontogenic Cells and Differential Expression of the Two Homeobox Genes MSX-1 and DLX-2 in Transgenic Mice



Msx and Dlx homeobox genes encode for transcription factors that control early morphogenesis. More specifically, Msx-1, Msx-2, and Dlx-2 homeobox genes contribute to the initial patterning of the dentition. The present study is devoted to the potential role of those homeobox genes during the late formation of mineralized tissues, using the rodent incisor as an experimental system. The continuously erupting mandibular incisor allows (1) the coinvestigation of the whole sequences of amelogenesis and dentinogenesis, aligned along the main dental axis in a single sample in situ and (2) the differential characterization of transcripts generated by epithelial and ectomesenchymal odontogenic cells. Northern blot experiments on microdissected cells showed the continuing expression of Msx-2 and Dlx-2 in the later stages of dental biomineralization, differentially in epithelial and ectomesenchymal compartments. Transgenic mice produced with LacZ reporter constructs for Dlx-2 and Msx-1 were used to detect different components of the gene expression patterns with the sensitive β-galactosidase histoenzymology. The results show a prominent epithelial involvement of Dlx-2, with stage-specific variations in the cells involved in enamel formation. Quantitative analyses identified specific modulations of Dlx-2 expression in ameloblasts depending on the anatomical sites of the incisor, showing more specifically an inverse linear relationship between the Dlx-2 promoter activity level and enamel thickness. This investigation extends the role of homeoproteins to postmitotic stages, which would control secretory cell activity, in a site-specific manner as shown here for Dlx-2.


Skeletogenic cells derive from three distinct lineages, the lateral plate mesoderm, the sclerotome, and the neural crest, which form most of the cranial skeleton.(1) Early patterning and development of teeth and craniofacial bones determine the correct morphology of these mineralized tissues and are a model system for understanding the pattern formation during head development.(2) Our present hypothesis is that differentiated skeletogenic cells that are phenotypically identical, for instance, osteoblasts expressing osteocalcin, bone sialoprotein, and osteonectin, may be different in their regulatory mechanisms depending on their anatomical site, in relation with their specific patterning and differentiation pathways.(3) This assumption is based on (1) an observed variation in the expression levels of osteocalcin in developing bone and teeth,(4) (2) the differential gene expression patterns(5,6) and site-specific functions of Msx-1, Msx-2, and Dlx-2 in transgenic null mutant mice(7,8) during early development, and (3) the direct link established between Msx-2/Dlx-5 transcription factors and osteocalcin transcription activity in osteoblasts.(5,9–13) Here, this hypothesis is explored in the tooth because it constitutes an exemplary system in the developmental biology of mineralized tissues.(14–17)

The formation of mineralized dental tissues involves the coordinated activities of several secretory units,(18) derived from buccal epithelium for enamel and cranial neural crest cells (ectomesenchyme) for dentine and cementum.(19) Dental morphogenesis and the terminal differentiation of odontogenic cells are achieved through a series of interactions between the dental epithelium and ectomesenchymal cells.(20,21) These interactions involve cross-talk between diffusible signaling molecules such as bone morphogenetic protein (BMP2-BMP4)/Sonic hedgehog (SHH) fibroblast growth factor (FGF-8) and transcription factors such as members of the Msx/Dlx/Pax families(14,22–25) Mammalian Hox genes have been shown to have a role in the patterning of the axial skeleton(1) and the developmental expression of non-Hox transcription factors in facial ectomesenchyme cells controls the patterning of facial bones and teeth.(2,15,17) More specifically, expression patterns of homeobox genes of Msx/Dlx families, orthologous to fruit fly muscle segment (Msh) and distal-less (Dll) genes, respectively, have been investigated during early stages of mouse odontogenesis.(26–30) From those data, Msx and Dlx genes have been proposed to contribute to odontogenic patterning.(15,29) This concept has been supported by phenotypes of mice with null mutations for Msx-1 and Dlx-1/Dlx-2.(7,8,31) Such mice exhibit major and site-specific defects in teeth and other craniofacial skeletal units, that is, arrest of molar development at the bud stage for Msx-1 null mutant and absence of maxillary molars in Dlx-1/Dlx-2 null mutant. Functional analyses based on recombination experiments in organotypic cultures have furthermore established that Msx-1, Msx-2, and Dlx-2 contribute to the epithelial-mesenchymal interactions that initiate tooth development.(24,28,32)

Therefore, Msx and Dlx homeobox gene families have attracted much attention because of their role in skeletal morphogenesis.(7,8,15,25,28,31,33,34) In contrast, few studies have focused on the involvement of those transcription factors in the process of biomineralization and more specifically at postnatal stages of development in vivo. In adult tissues of other biological systems, the expression of several homeobox genes has been reported: Dlx-3, Msx-1, and Msx-2 in the skin(35–37) and Msx-1 and Msx-2 in the mammary glands,(38) suggesting that these developmental genes may contribute to other functions besides early patterning and development. Furthermore, several human gene mutations have been reported (DLX-3, MSX-2, and MSX-1), which were related to major abnormalities in the formation of craniofacial mineralized tissues.(39–41) Their phenotypes suggest that Msx/Dlx transcription factors (enamel defects in the trichodento-osseous syndroma for instance in Ref. 39) may be key factors not only in skeletal patterning but also in later control of the activity of skeletogenic cells during postnatal development.

The specific aim of this study is to analyze the expression of three of those homeobox genes, Msx-1, Msx-2, and Dlx-2, during dental biomineralization at postnatal stages by Northern blot analysis because they were the first transcription factors to be clearly studied in early development.(7,8,15,26–31,42) Furthermore, in situ investigation of promoter-driven LacZ expression also was done in transgenic mice (Msx-1 and Dlx-2).(31,43) The mandibular continuously erupting incisor of rodent was used here because it allows the different stages of amelogenesis and dentinogenesis to be studied on one sample along the main anatomical axis.(44–46)


Msx-1/nLacZ mice

Heterozygous mice were produced by crossing wild-type mice C57BL/6 (C. River, Saint-Aubin les Elboeuf, France) with heterozygous Msx-1/nLacZ transgenic mice generated as previously described.(31) Briefly, the transgenic animals bear an nLacZ reporter gene inserted into the second exon of the mouse Msx-1 gene in the region coding for the third helix of the homeodomain. The heterozygous mutants were back-crossed for 10 generations onto a C57BL/6J background. Typing of the heterozygous was realized by polymerase chain reaction (PCR) using two sets of primers as previously described.(31)

Dlx-2/LacZ mice

Transgenic mice were produced using a Dlx-2/LacZ reporter construct containing 3.7 kilobases (kb) upstream of the translation start site and thus utilizing the endogenous Dlx-2 promoter in B6CBA mice (Haston, U.K.).(43) Crossing a single line, confirmed by Southern analysis with wild-type mice produced heterozygotes and wild types used in this study.


Wild-type mice mandibles (C57BL/6 mice) were fixed by immersion with 4% paraformaldehyde (Sigma, la Verpillière, France) in phosphate-buffered saline (PBS), pH 7.4 (the fixative solution), overnight at 4°C and rinsed in PBS. Half-mandibles were microradiographed on High Resolution Film SO-343 (Kodak Professional, Paris, France) by microfocal X-ray generator (Tubix, Paris, France) at a focal distance of 56 cm for 15 minutes (power setting at 8 mA and 15 kV). Standardized enlargements were used to locate incisor areas concerned by the successive presecretion, secretion, and maturation stages of enamel formation.

Whole-mount staining of embryos and microdissected incisors from Msx-1/nLacZ and Dlx-2/LacZ mice

Embryos and half-mandibles of postnatal mice were microdissected in PBS and then fixed in the fixative solution for 2 h at 4°C. The samples were then rinsed for 2 h in PBS at 4°C and stained for β-galactosidase activity as previously described.(31) The samples were fixed again for 24 h at 4°C with the fixative solution and rinsed at 4°C in PBS for 2 h. Incisors were extracted from half-mandibles in PBS under a dissecting stereomicroscope (Stereo Star, American optical). Both embryos and incisors were photographed using Kodak 160T film. Wild-type embryos and microdissected incisors raised on the same conditions were treated similarly, as controls for potential endogenous β-galactosidase activity.

Whole-mount in situ hybridization with Dlx-2 riboprobes

Antisense (AS) and sense (S) Dlx-2 riboprobes were synthesized from a 1.7-kb complimentary DNA (cDNA) fragment, cloned in a Bluescript derivative plasmid called E61.(47)Dlx-2 plasmid was linearized with EcoRI restriction endonuclease (Gibco BRL, Life Technologies, Cergy-Pontoise, France) and AS and S probes were obtained using, respectively, T3 and T7 RNA polymerase (Gibco BRL, Life Technologies, Cergy-Pontoise, France) and incorporating digoxygenin-UTP (Boehringer-Mannheim, Meylan, France) according to manufacturers instructions. Wild-type mice were killed and half-mandibles were fixed overnight in the fixative solution. Incisors were microdissected in PBS under the stereomicroscope and dehydrated in a graded series of ethanol. Whole-mount in situ hybridization was carried out according to Wilkinson(48) but excluding proteinase K treatment. Stained incisors were stored at 4°C in PBS with 0.4% paraformaldehyde and photographed using Kodak 160T film.

In situ hybridization with Dlx-2 riboprobes

Whole heads of E18.5 mice embryos and microdissected half-mandibles of postnatal mice were fixed overnight at 4°C by immersion in the fixative solution and rinsed for 2 h at 4°C in PBS. Undecalcified samples were cut into 10-μm serial sections with a cryostat (Bright instrument company LTD, Huntingdon, U.K.) and sections were deposited on 50 mg/ml poly-l-lysine (Sigma, la Verpillière, France) coated slides. In situ hybridization was carried out with Dlx-2 riboprobes (AS and S) prepared as described for whole-mount in situ hybridization but using (35S)-UTP labeled probes. In situ hybridization was performed as previously described.(4,49) Briefly, cryostat sections were pretreated with proteinase K (Sigma, la Verpillière, France), hybridized with 20 μl of labeled probes containing 60,000 cpm/μl radioactivity, in a humid chamber overnight at 50°C and washed under high-stringency conditions. The slides were dipped into NTB2 autoradiographic emulsion (Kodak, Paris, France) and exposed for 5 weeks at 4°C. After developing the film, sections were stained with hematoxylin (Merck, Nogent-sur Marne, France), dehydrated, and mounted under coverslip. Sections were observed and photographed with a Leitz Orthoplan photomicroscope (Leitz/Leica, Rueil-Malmaison, France) using bright and dark field illumination.

Dlx-2/LacZ half-mandible serial sections and quantitative analyses of Dlx-2/LacZ expression

After staining of β-galactosidase activity and fixation overnight at 4°C, the half-mandibles were rinsed 4 h in PBS at 4°C and decalcified for 4 weeks at 4°C in PBS with 4.13% disodium ethylenediamine-tetraacetic acid (Sigma, la Verpillière, France) and 0.2% paraformaldehyde, pH 7.4. After dehydration, half-mandibles were wax-embedded and serial sections of 8 μm were cut (Microtome Leica RM 2145, Leica, France). After rehydration, the slides were stained according to the Van-Gieson technique. Finally, the slides were dehydrated and mounted with DePeX (BDH laboratory, Poole, U.K.). Sections were observed and photographed with the Leitz Orthoplan photomicroscope using Kodak 160T film. Wild-type mice raised on the same conditions were treated similarly as controls for β-galactosidase activity staining.

Frontal sections from Dlx-2/LacZ mandibles (n = 5) were used to measure β-galactosidase activity resulting from the transgene expression in incisor enamel secretion stage ameloblast using a Quantimet Q570 (Leitz/Leica, Ruel-Malmaison, France). Measurements were made along six lines perpendicular to the enamo-dentinal junction for each section on (1) ameloblasts on each anatomical site and (2) their corresponding enamel matrix. Image-acquiring and quantification procedures were those previously described.(50) Briefly, images were acquired with a microscope (Aristoplan Leitz; Leitz/Leica, Ruel-Malmaisan, France) equiped with a fixed camera CCD (Sony, Paris, France) that enables a 31 mm × 31 mm square to be shot corresponding to the field measured on the slide. Quantification was realized in inverse grey levels (white levels) and linear mode. The quantimet integrated the white levels along a linear segment, which uncovers the entire length of the secretory stage ameloblast. The enamel thickness also was measured in each site. Relationships between the enamel thickness and grey level (complementary to the white level) corresponding to the β-galactosidase staining reflecting Dlx-2 promoter activity was established using linear regression for the calculation of the ratio.

Northern blotting

The animals used for Northern blotting experiments, because of the size of samples (mandibular incisor epithelium and mesenchyme), were 56-day-old Sprague-Dawley rats (C. River, Saint-Aubin les Elboeuf, France; n = 20). The rat mandibles were removed and incisors were microdissected as previously described(51) in order to separate dental epithelium and mesenchyme. Total RNAs were isolated from dental mesenchyme and epithelium using guanidium thiocyanate–phenol chloroform procedure (Euromedex, Souf-fleweyersheim, France). Graded concentration of total RNAs (1, 5, 10, 15, and 20 μg) were fractionated electrophoretically on a 1% agarose-formaldehyde gel and transferred on to a nylon membrane (Amersham, les Ulis, France). Membranes were successively prehybridized and then hybridized with mice Dlx-2, Msx-2, and glucose glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probes 32P-labeled by random priming (Rediprime II, Amersham Pharmacia Biotech, France). All membranes were stripped between different hybridizations by washing in 50% formamide and 10 mM phosphate buffer, pH 6.5, for 30 minutes at 65°C. The hybridized blots were autoradiographed using Kodak films (Amersham).

Figure FIG. 1..

Microradiography of a 13-day-old mouse half-mandible. The first and second molars (m1, m2) are clearly visible at this age, which corresponds to their eruption stage. The crown of the third molar (m3) is formed. Concerning the incisor (i), the enamel (e) synthesis process can be described on the labial part with the successive presecretion (ps), secretion (s), and maturation (m) stages, ([d] dentine).


Description of the experimental models

Dental biomineralization in the mandibular incisor: The processes of cell differentiation, enamel, and dentine deposition may be followed along the main axis of the mouse mandibular incisor. The present study focuses on 13-day-old mice as an experimental system because the maturation stage of enamel formation is already reached at this stage. Microradiography of 13-day-old mice half-mandibles enabled us to define areas of incisor corresponding to the stages of enamel formation (Fig. 1). The presecretion area corresponding to the apical part of the incisor does not contain enamel matrix. In the secretion area—the central part of the incisor—the enamel matrix thickness gradually increases whereas in the maturation area the enamel matrix thickness is constant and undergoes mineralization process.

Msx-1/nLacZ and Dlx-2/LacZ whole-mount stainings: Msx-1/nLacZ and Dlx-2/LacZ whole-mount staining in E10.5 embryos (Figs. 2A–2D) showed the developmental patterns during embryonal stages of development that are visualized with those constructs.(31,43) The β-galactosidase activity was located in both mandibular and maxillary odontogenic areas.

Figure FIG. 2..

Whole-mount β-galactosidase histoenzymology in E10.5 (A and B) Msx-1/nLacZ heterozygous embryo and (C and D) Dlx-2/LacZ heterozygous embryo. Msx-1 and Dlx-2 transgenes are expressed in maxilla (mx) and mandible (md) areas.

Distribution of Dlx-2 and Msx-1 expression in the mandibular incisor

The mandibular mouse incisor is characterized by the spatial restriction of enamel deposition on the labial face. On the lingual part, instead of enamel, dentine is overlayed by cementum, which enables attachment to alveolar bone through periodontal ligament fibers. Examination of micro-dissected incisors of 56-day-old Dlx-2/LacZ expressing mice (Fig. 3A) revealed expression in the dental epithelium on the labial surface. Whole-mount in situ hybridization at the same age confirmed that this was endogenous Dlx-2 expression (Figs. 3B and 3C), which is not homogenous and shows variations during enamel formation as shown in these vestibular views of incisors (Figs. 3B and 3C). In contrast, microdissected Msx-1/nLacZ ± mice incisors (Fig. 3D) revealed that there is no Msx-1 expression in the incisor at postnatal stages of development. The Msx-1 expression in the mesenchyme ended at E18.5, that is, just before birth (data not shown). In situ hybridization with Dlx-2 riboprobes provided similar results and confirmed the Dlx-2 pattern of expression in ameloblasts. In E18.5 embryos (Fig. 4), Dlx-2 appeared to be highly expressed in dental cells from epithelial origin and weakly in the dental mesenchyme of the incisor because there is an extinction of Dlx-2 expression in the mesenchyme around neonatal stage.

Figure FIG. 3..

(A) Microdissected incisors of 56-day-old Dlx-2/LacZ heterozygous mice (upper) and wild-type mice (down) after staining for β-galactosidase activity. The Dlx-2 trans-gene expression is only detected in the enamel organ (eo) and shows variations during the enamel formation. (B) Whole-mount in situ hybridization on 56-day-old mice microdissected incisor with Dlx-2 AS riboprobe labeled with digoxigenine versus (C) Dlx-2 s riboprobe in the medial part of the incisor corresponding to the transition between presecretion and secretion and in a vestibular view. Dlx-2 transcripts are expressed in the enamel organ (eo) but not in the mesenchyme (m). (D) Microdissected incisors of 14-day-old Msx-1/nLacZ heterozygous mice after staining for β-galactosidase activity. No transgene expression is detected.

Developmental stages of enamel formation

Epithelial cells are characterized by their morphology; during the presecretion stage, the apical loop contained several cell rows that appeared morphologically quite similar (Fig. 5A). Progressively, the supra-ameloblastic cells and inner dental epithelium showed different features. Supra-ameloblastic cells appeared to be composed of the external dental epithelium, stellate reticulum, and stratum intermedium (Fig. 5A). The inner dental epithelium showed an increasing cell polarity as assessed by the shape and position of nuclei and the observed length of the cell body (Figs. 5A and 6A). The secretion stage was characterized by the evidence of a progressive enamel matrix deposition by polarized ameloblasts with distinct steps: inner aprismatic, inner prismatic, external prismatic, and external aprismatic, related to the organization of the apical process of the ameloblast (Figs. 5B, 6B, and 6C). Maturation stage involved a series of morphological events: shortening of ameloblasts during the transition and then after periodic morphomodulations of the apical membrane structure and distal intercellular junctions in ruffle and smooth-ended ameloblasts (Figs. 6D–6F). Furthermore, supra-ameloblastic cells appeared to be restricted to a single layer, the papillary layer (Fig. 6F). The developmental pattern of Dlx-2 promoter activity was established on longitudinal sections of the central incisor areas corresponding to the optimal thickness of enamel.

Dlx-2 transgene expression in odontogenic epithelial cells

In the apical part of enamel presecretion zone of the incisor, the so-called dental loop, Dlx-2/LacZ transgene expression, was detected in undifferentiated dental cells of epithelial origin but not significantly in undifferentiated ectomesenchymal cells (Fig. 5A) as previously shown for this construct in early steps of development.(43) During the major part of the presecretion stage, the transgene expression was maintained in epithelial cells, including preameloblasts and supra-ameloblastic cells (Fig. 6A). During late enamel presecretion, the transgene expression appeared to decrease in polarizing ameloblasts (Fig. 5B). In the enamel secretion zone, the β-galactosidase activity was only detected at high levels in supra-ameloblastic cells. The secretory ameloblasts did not express the transgene activity (Figs. 6B and 6C). However, in other planes of section where the enamel thickness was lower, secretion stage ameloblasts accumulated β-galactosidase staining albeit at a lower level than supra-ameloblastic cells.

Figure FIG. 4..

In situ hybridization on longitudinal sections of E18.5 wild-type embryos with Dlx-2 AS riboprobe [(A) bright and (B) dark field] and sense riboprobe [(C) bright and (D) dark field]. Dlx-2 transcripts are detected in epithelial cells (ec) but less significantly in the mesenchymal cells (mc).

Dlx-2 developmental pattern during the sequence of enamel deposition and mineralization

During the secretion stage, the intensity of the Dlx-2/LacZ expression was related to the thickness of extracellular enamel matrix. At the very beginning, where ameloblasts did not appear to secrete a morphologically distinct enamel matrix (Fig. 6A), the intensity of the staining was identical in the supra-ameloblastic and ameloblastic cells. When the enamel thickness increased, the levels of Dlx-2/LacZ expression significantly decreased (Fig. 5B). When the enamel deposition ended, jointly with changing morphology of the ameloblastic and supra-ameloblastic cells, the staining pattern was different. At the end of the enamel secretion stage and during the transition between enamel secretion and maturation zones, Dlx-2/LacZ expression was de novo gradually detected in ameloblasts (Figs. 6D and 6E). In the enamel maturation zone, the transgene was highly expressed in both ameloblasts and supra-ameloblastic cells (Fig. 6F). In contrast, ectomesenchymal cells facing enamel did not show significant histoenzymatic reactivity for β-galactosidase.

Dlx-2 expression in secretory ameloblasts is correlated with enamel thickness

Frontal sections of Dlx-2/LacZ mice half-mandibles confirmed the high Dlx-2 expression in ameloblasts during presecretion and maturation stages and the quasi-absence of expression during secretion stage in the areas corresponding to the optimal thickness of enamel. However, the transgene expression appeared to vary in secretory ameloblasts according to the anatomical sites, being maximum where enamel thickness was minimal (Fig. 7A). The revealed β-galactosidase activity was measured systematically on not wax sections that were counterstained using a quantimet and related to the varying enamel thickness from the central to the lateral parts of the incisor. The selected areas were chosen in the medial part of the secretion stage in order to avoid the presence of presecretion stage ameloblasts. Calculation of the ratio between the relative optical density inside ameloblasts and the corresponding enamel thickness (Fig. 7B) showed a linear relationship (r = 0.97 ± 0.03).

Dlx-2 and Msx-2 transcript expression in the dental epithelium and mesenchyme

Northern blotting experiments (Fig. 8) revealed that (1) there was cross-hybridization between mice probes and rat messenger RNA (mRNA) for Dlx-2 and Msx-2, (2) the size of rat Dlx-2 transcript (2.6 kb) and rat Msx-2 transcripts (1.4 kb and 2.4 kb) were similar to those previously reported for mice Dlx-2(42) and Msx-2,(13) and (3) the Dlx-2 transcript was only detected in the dental epithelium, which confirmed the data obtained by in situ hybridization, whereas Msx-2 transcripts were detected in both dental mesenchyme and epithelium.


Mineralized tissue formation results from the coordinated activity of highly differentiated cells. The deposition of extracellular matrix containing different sets of proteins is followed by a tissue-specific biomineralization.(52) Convergent regulations involving systemic hormones, local growth factors, cytokines, and also transcription factors contribute to the final control of gene activity in cells of mineralized tissues.(12,53,54) Among those regulatory key molecules, Msx/Dlx family members were shown to act on osteocalcin and alkaline phosphatase promoter activity.(5,9–13) Furthermore, Msx-1, Msx-2, Dlx-2, and Dlx-5 proteins have been investigated in in vitro experiments, which show that their transcriptional properties display reciprocal inhibition.(55,56) To cooperate, transcription factors have to be coexpressed in the same tissues.

Figure FIG. 5..

Sections of Dlx-2/LacZ 13-day-old mouse half-mandible along the main axis (longitudinal sections). (A) In the apical part of the incisor, the dental loop, which corresponds to the presecretion area, only undifferentiated cells from epithelial origin (ue), preameloblasts (pab), presecretory ameloblasts (psab), and supra-ameloblastic cells (sac) express the transgene. Magnification ×420. (B) During the transition between presecretion and secretion stages, the trans-gene expression decreases inside ameloblasts (ab) after the course of the enamel matrix (em) deposition, whereas the transgene expression is constant in supra-ameloblastic cells (sac). Magnification ×840.

Our present aim was to investigate the three first homeobox genes shown to constitute key factors in the patterning of teeth and craniofacial bones.(7,15,30) Furthermore, the developmental pattern of Msx-1 and Dlx-2 was analyzed in transgenic mice. When compared with previous data describing early development(5,31,57) our data show that Msx-1 involvement is restricted to the initial tooth germ formation, whereas Msx-2 and Dlx-2 were expressed continuously in teeth from the earliest stages of dental lamina until dental biomineralization achievement at postnatal stages.(5,31,57) The investigation of their developmental pattern (this study; first postnatal days for Msx-2)(5) showed cell- and stage-specific involvement of those homeobox genes in mineralized tissue formation. Msx-2 was distributed in both epithelial and mesenchymal compartments, suggesting that this transcription factor may play a role in both amelogenesis and dentinogenesis processes. In those tissues, Msx-2 has been proposed to be able to act reversely on osteocalcin and the tissue nonspecific alkaline phosphatase, based on data obtained in bone cells in vitro.(5) Interestingly, alkaline phosphatase gene shows distinct tissue-specific developmental expression patterns in teeth,(58) which show similarity with Dlx-2 in the epithelial compartment.

The molecular parameters of dental mineralized tissue formation are conveniently investigated in the continuously erupting rodent incisor. These include fine analysis of mineral formation,(45) calcium/phosphate handling,(59–61) detailed cellular events,(62,63) and regulation of gene expression during the lifetime of odontogenic cells.(4,49,51,64–70) These studies show that distinct developmental patterns are related to the succession of matrix deposition and secondary hypermineralization of enamel with one maximum for the expression of matrix proteins—amelogenin, ameloblastin/amelin/sheatlin, and enamelin—when enamel deposition occurs, that is, the secretion stage and two maximums for proteins involved in calcium and phosphate handling (i.e., alkaline phosphatase, calcium adenosine triphosphatase [ATPase], calbindin-D9k, calbindin-D28k, and parvalbumin).

Figure FIG. 6..

Reverse fluctuations of Dlx-2/LacZ staining during secretion and maturation stages in a 13-day-old mouse incisor. During the secretion stage, Dlx-2 transgene expression is progressively undetectable (B and C) in ameloblasts (ab) in contrast to (A) presecretion stage and is evident in supra-ameloblastic cells (sac). During the transition between secretion and maturation stages, the trans-gene expression is de novo gradually (from D to E) increasing in ameloblasts, until a level of expression similar (F) to the adjoining supra-ameloblastics cells (sac) in maturation stage ameloblasts. No difference in the staining was observed in ruffle-ended and smooth-ended ameloblasts. Magnification ×840.

One of the most striking results was the modulation of Dlx-2 gene expression in the epithelial cells throughout the stages of amelogenesis. In ameloblasts, the first optimal expression of Dlx-2 was associated with the enamel presecretion stage and the second one was associated with the enamel maturation stage. These data suggest that Dlx-2 gene expression may switch on and off during two successive activation/inactivation steps in the lifetime of the same postmitotic cell (ameloblast). In contrast, supra-ameloblastic cells show continuous, even Dlx-2 expression levels throughout amelogenesis, suggesting that these activation/inactivation signals controlling Dlx-2 expression in the ameloblasts are not acting on those cells. This differential behavior of the ameloblastic and supra-ameloblastic cells is a constant feature of odontogenesis and might be related to reciprocal interactions with adjoining ectomesenchymal cells via diffusible signaling molecules such as BMP-2/BMP-4 and their cognate receptors.(71) Cross-talk between BMP-2/BMP-4 and Msx-1, Msx-2, and Dlx-2 have been shown clearly in early dental development by dissociation/reassociation experiments, addition of recombinant BMP-2/BMP-4 via agarose beads onto epithelial and mesenchymal tissues from control and null mutant mice.(24,28,32,33,72) Such explanations cannot be used in late stages of tooth biomineralization because ameloblasts strongly adhere to enamel surface and odontoblasts are embedded in dentine through their large distal process.(62,63) However, the present study supports the notion that such cross-talk may be acting on postmitotic odontogenic cells because dentine proteins have been shown to be present in enamel matrix and vice versa.(73,74)

Figure FIG. 7..

(A) Frontal sections of a Dlx-2/LacZ heterozygous 13-day-old mouse half-mandible at secretion stage. The transgene is detected in supra-ameloblastic cells (sac) at high levels and also variably in some ameloblasts (ab) depending on their anatomical sites. Arrows correspond to the extremities of the six segments used for the measurements ([e] enamel). Magnification ×840. (B) Graph of the linear inverse relationship between the Dlx-2 transgene expression inside ameloblasts and enamel thickness. Linear regression is shown for a representative frontal section realized in the middle secretion (present frontal section).

Interestingly, signaling molecules that are switched on and off during epithelial-mesenchymal signaling interactions in early development also show fluctuations during those critical steps of enamel biomineralization, for example, epidermal growth factor (EGF) receptors;(66) adenomatous polyposis coli protein (APC), a protein involved in the Wint signaling pathway;(69) and as shown here, Dlx-2. The minimal expression of Dlx-2 observed here corresponds to the major step of enamel matrix deposition and thus to the optimal expression levels of mRNAs encoding matrix proteins such as amelogenin and enamelin.(68,75)Dlx-2 may be proposed as a morphogenetic antagonist to the increase of enamel thickness. This is based on the reverse developmental patterns of Dlx-2 expression and the activity of enamel matrix proteins deposition(63) and the inverse linear relationship between Dlx-2 gene activity and enamel thickness. Here, Dlx-2 was shown at the same developmental stage, that is, in secretion stage ameloblasts, to be expressed variably depending on the anatomical site on frontal sections of the incisor. This data may be related to another report on amelogenin mRNA levels in different anatomical sites of the molars.(76) Quantitative in situ hybridization showed that amelogenin mRNA levels were related to the enamel thickness, which vary depending on the anatomical site, suggesting that the regulation of amelogenin gene activity in ameloblasts is the key factor in the establishment of the mineralized morphogenesis. In the perspective of investigations on the role of Msx and Dlx transcription factors, dental mineralized tissues may be studied more easily than bone tissue because no remodeling modifies the general shape of these skeletal units.

Figure FIG. 8..

Northern blotting with graded amounts (1, 5, 10, 15, and 20 μg) of total RNA extracted from 56-day-old rat microdissected mandibular incisors. Msx-2 mRNAs (1.4 kb and 2.4 kb) are expressed in both mesenchyme and epithelium whereas Dlx-2 mRNA (2.6 kb) is only detected in the epithelium. The GAPDH is used as a control for RNA graded series deposition.


This study was supported by grants of the EA2380 laboratory, University Paris VII. Travel grants to A.B. also were obtained from EEC (COST B8 action).