Laminin alpha2 is essential for odontoblast differentiation regulating dentin sialoprotein expression.

Laminin alpha2 is subunit of laminin-2 (alpha2beta1gamma1), which is a major component of the muscle basement membrane. Although the laminin alpha2 chain is expressed in the early stage of dental mesenchyme development and localized in the tooth germ basement membrane, its expression pattern in the late stage of tooth germ development and molecular roles are not clearly understood. We analyzed the role of laminin alpha2 in tooth development by using targeted mice with a disrupted lama2 gene. Laminin alpha2 is expressed in dental mesenchymal cells, especially in odontoblasts and during the maturation stage of ameloblasts, but not in the pre-secretory or secretory stages of ameloblasts. Lama2 mutant mice have thin dentin and a widely opened dentinal tube, as compared with wild-type and heterozygote mice, which is similar to the phenotype of dentinogenesis imperfecta. During dentin formation, the expression of dentin sialoprotein, a marker of odontoblast differentiation, was found to be decreased in odontoblasts from mutant mice. Furthermore, in primary cultures of dental mesenchymal cells, dentin matrix protein, and dentin sialophosphoprotein, mRNA expression was increased in laminin-2 coated dishes but not in those coated with other matrices, fibronectin, or type I collagen. Our results suggest that laminin alpha2 is essential for odontoblast differentiation and regulates the expression of dentin matrix proteins.

Laminin ␣2 is subunit of laminin-2 (␣2␤1␥1), which is a major component of the muscle basement membrane. Although the laminin ␣2 chain is expressed in the early stage of dental mesenchyme development and localized in the tooth germ basement membrane, its expression pattern in the late stage of tooth germ development and molecular roles are not clearly understood. We analyzed the role of laminin ␣2 in tooth development by using targeted mice with a disrupted lama2 gene. Laminin ␣2 is expressed in dental mesenchymal cells, especially in odontoblasts and during the maturation stage of ameloblasts, but not in the pre-secretory or secretory stages of ameloblasts. Lama2 mutant mice have thin dentin and a widely opened dentinal tube, as compared with wildtype and heterozygote mice, which is similar to the phenotype of dentinogenesis imperfecta. During dentin formation, the expression of dentin sialoprotein, a marker of odontoblast differentiation, was found to be decreased in odontoblasts from mutant mice. Furthermore, in primary cultures of dental mesenchymal cells, dentin matrix protein, and dentin sialophosphoprotein, mRNA expression was increased in laminin-2 coated dishes but not in those coated with other matrices, fibronectin, or type I collagen. Our results suggest that laminin ␣2 is essential for odontoblast differentiation and regulates the expression of dentin matrix proteins.
Tooth development is regulated by sequential and reciprocal interactions between neural crest-derived mesenchymal cells and the oral environment (1-3); however, the precise molecular mechanisms mediating interactions between epithelium and mesenchymal cells are not clear, although basement membrane (BM) 1 components have been shown to play important roles in these regulatory events. In addition, the extracellular matrix layer, whose main components are laminin, collagen IV, nidogen, and sulfated proteoglycan, and the BM layer are both considered to be involved with cell proliferation and differentiation (4,5).
The laminin family is composed of BM proteins that have been implicated in diverse functions of epithelial and mesenchymal cells. Each member is a heterotrimer composed of ␣, ␤, and ␥ chains, and five ␣, three ␤, and three ␥ chains have been identified and are known to form at least 15 heterotrimer structures. Most laminin family members have been found to have combinations of ␤1␥1 or ␤2␥1 chains with one of the five ␣ chains, although laminin-5 has a unique chain composition, ␣3␤3␥2.
Laminin-2 (with ␣2, ␤1, and ␥1 chains), also known as merosin, is a major component of BM proteins in skeletal muscle and the peripheral nervous system (6), and the absence of the laminin ␣2 chain causes merosin-deficient congenital muscular dystrophy (MD-CMD) (7), which characteristically involves skeletal muscle along with the peripheral and central nervous systems (8). MD-CMD causes degradation, regeneration, interstitial fibrosis, and adipose tissue infiltration in skeletal muscle. Dystrophic (dy/dy) mice also display a severe reduction in laminin ␣2 chain expression and are accepted as an animal model of MD-CMD (6, 9 -11). We generated a null mutant lacking the laminin ␣2 chain using a gene-targeting technique to examine the molecular pathophysiology of MD-CMD (12). The mice showed symptoms similar to those seen in MD-CMD and a shorter life span than dy/dy mice (13).
During tooth development, the mRNA of three laminin ␣ chains, ␣1, ␣2, and ␣4, is expressed in tooth mesenchymal cells, whereas two other types, laminin ␣3 and ␣5 chain mRNA, are found in epithelial cells (14,15). Furthermore, laminin ␣5 mRNA is widely expressed in tooth epithelium, with the corresponding protein distributed along the tooth basement membrane during the embryonic stage and diminished at the start of enamel matrix production (14). Laminin ␣3 expression is slight in the embryonic stage and then dramatically increases during terminal differentiation of ameloblasts, following degradation of the tooth BM (14,16). On the other hand, laminin ␣2 mRNA is detected in mesenchymal cells in the tooth germ. In the earlier stages, mesenchymal expression is seen around the epithelial bud, while in later stages (E15-18) laminin ␣2 mRNA expression becomes stronger in dental sac cells than in dental papilla cells. By using immunohistostaining, laminin ␣2 can be detected in the tooth germ BM before E15, although later (E15-18) staining is lost from the dental BM, the area between the inner dental epithelium and dental papilla mesenchyme layers, whereas it remains strong in the BM area between the outer dental epithelium and dental sac mesenchyme (14). However, the expression and molecular mechanisms of laminin ␣2 in postnatal tooth development have not been clearly shown.
In the present study, we examined tooth formation and dentin sialoprotein (DSP) expression in laminin ␣2 knockout mice, a mutant strain with thin enamel and a widely opened dentinal tube, which are caused by a reduction of dentin formation and odontoblast differentiation. Laminin-2 enhances the expression of dentin sialophosphoprotein (DSPP) and dentin matrix protein (DMP) in primary cultured dental mesenchymal cells. Our present findings suggest that interactions between laminin ␣2 and odontoblasts regulate their differentiation and are important for dentinogenesis.

EXPERIMENTAL PROCEDURES
Scanning Electron Microscope (SEM) Analysis-Incisors were taken from wild-type and laminin ␣2 null mice and coated with gold and photographed using scanning electron microscopy at 20 kV (S-3500, Hitachi Ltd., Tokyo, Japan). To observe the enamel crystals and dentinal tubes, the specimens were embedded in epoxy resin, cut with an ISOMET low speed saw (Buehler, Lake Bluff, IL), and then treated with 40% phosphoric acid for 10 s and 10% sodium hypochlorite for 30 s, prior to coating with gold.
Preparation of Tissue Sections-Laminin ␣2 null mice were generated by gene targeting and housed in a pathogen-free animal facility. Standard Nagasaki University guidelines were followed to monitor their health status as well as the housing and breeding practices. To prepare the heads of 3-week-old mice, each animal was anesthetized and then fixed by perfusion with 4% paraformaldehyde/PBS. The maxilla was dissected out, post-fixed overnight at 4°C in 4% paraformaldehyde/PBS, and decalcified with 250 mM EDTA/PBS for 2 weeks, then dehydrated in xylene through a graded ethanol series, and embedded in paraplastic paraffin (Oxford Laboratories). Sections were cut at 10 m on a microtome (RM2155, LEICA, Inc.). For detailed morphological analyses of molars and incisors, sections were stained with Harris hematoxylin and eosin Y (Sigma). For staining of cultured cells, cells were fixed with 4% paraformaldehyde, 0.5% Triton X-100, PBS for 5 min and then 4% paraformaldehyde/PBS for 10 min.
Dental Epithelial and Mesenchymal Cell Cultures-For dental mesenchymal cell cultures, P3 mouse molars were dissected and treated with 0.1% collagenase, 0.05% trypsin, 0.5 mM EDTA for 10 min, after which the dental mesenchyme was separated from the dental epithelium. Separated dental mesenchyme samples were treated with 0.1% collagenase, 0.05% trypsin, 0.5 mM EDTA for 15 min and then collected using a pipette and placed into the wells (21). Dental mesenchymal cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) with 10% fetal calf serum, whereas a dental epithelial cell line (HAT-7) was cultured in Dulbecco's modified Eagle's medium/F-12 with 10% fetal calf serum (22).
Cell Proliferation and Cell Binding-Dental epithelial cells (HAT-7) and dental mesenchymal cells at 1.0 ϫ 10 5 were cultured in a 60-mm diameter dish coated with or without laminin-2 (merosin, Invitrogen), type I collagen (Cellmatrix, NITTA GERATIN, Japan), and fibronectin (human fibronectin, Invitrogen) for 5 days. At 1, 3, and 5 days after plating, cells were treated with 0.05% trypsin, 0.5 mM EDTA, and their numbers were counted under a microscope. For cell binding, dental epithelial and mesenchymal cells were detached with 0.05% EDTA, washed with Dulbecco's modified Eagle's medium containing 0.1% bovine serum albumin, and resuspended to a concentration of 3.5 ϫ 10 5 /ml. Assays were performed in 96-well round-bottomed microtiter plates (Immulon-2HB, Dynex Technologies, Inc., Chantilly, VA). Wells were coated overnight at 4°C with laminin 2, type I collagen, or fibronectin, then diluted with PBS, and blocked with 3% bovine serum albumin for 1 h at 37°C. After washing, the cells were added to a plate and incubated for 60 min at 37°C. Attached cells were stained for 10 min with 0.2% crystal violet (Sigma) in 20% methanol. After washing with H 2 O, the cells were dissolved in 10% SDS, and absorbance at 600 nm was measured.

Laminin ␣2 Expression in Differentiated Odontoblasts and
Maturation of Ameloblasts-Laminin ␣2 is known to be expressed in dental mesenchymal cells and localizes in the BM area between the dental epithelium and mesenchyme layers in the early stage of tooth germ development (14); however, its expression in later stages, especially in the postnatal period, has not been clearly identified. We performed immunostaining of incisor samples from 3-week-old mice to identify the localization of laminin ␣2, which was found expressed in the preameloblast and pre-odontoblast interfaces ( Fig. 1A, a), as well as in odontoblasts, the outer side of the dental epithelium including the papillary cell layer, capillaries, and the muscle BM ( Fig. 1A, a and b). However, laminin ␣2 expression was not observed in the cervical loop region or the secretory stage of ameloblasts ( Fig. 1B, a). Later, in the early maturation ( Fig.  1B, b) and late maturation (Fig. 1B, c) stages, laminin ␣2 appeared in the side of the enamel. On the other hand, laminin ␣2 expression was detected in odontoblasts at all stages ( Fig.  1C). To confirm the expression in odontoblasts, immunostaining of primary cultured dental mesenchymal cells was performed, and laminin ␣2-positive cells were found expressed in DSP (Fig. 1D), which is a marker of odontoblasts, indicating that the odontoblasts produced laminin ␣2.
Decreased Dentin Formation and Amelogenesis Imperfecta in Laminin ␣2 Null Mice-For the targeted disruption of LAMA2, a PGK-neo gene was inserted into the laminin ␣2 chain (12). These mice had been back-crossed into the BALB/c strain at least 10 times, before the tooth phenotype was analyzed. To examine laminin ␣2 expression in the masseter muscle, we performed immunostaining of 3-week-old wild-type, heterozygous, and null mutant mice, using rat anti-laminin ␣2 chain monoclonal antibody 4H8 -2, which recognizes the 300-kDa portion of the protein (23). The results confirmed the absence of the laminin ␣2 chain in mutant strain muscles. Laminin ␣2positive staining was observed in the muscle BM ( Fig. 2A, a  and b) in the wild-type and heterozygote mice but not in the mutants ( Fig. 2A, c). In the same mice used to analyze laminin ␣2 expression in the masseter muscle, the color of the incisor surface in the mutants was found to be white when compared with the wild-type and heterozygote mice, indicating amelogenesis imperfecta (AI) (Fig. 2B). For detailed analyses of the incisor enamel and dentin, scanning microscopic examinations were performed. The overall tooth length and shape of the upper incisors were not different between the heterozygote and mutant mice (Fig. 2C); however, the cross-sectional surface of the incisor was decreased, as well as the enamel and dentin thickness by ϳ13 and 25%, respectively, in the mutant speci-mens (Fig. 2, D and E). Furthermore, the surface of the superficial enamel was rough as compared with that of the heterozygote incisors, resulting in the white color (Fig. 3d). In contrast, the size and structure of the enamel were not different between the two types of mice. On the other hand, the dentinal tubes were opened wide in the mutant mice following treatment with phosphoric acid and sodium hypochlorite, indicating immature dentin in the surrounding area (Fig. 3e). In the SEM analysis, enamel surface and dentin structures were not different between the wild-type and heterozygote mice (data not shown).
Decreased Expression of DSP in Laminin ␣2 Null Mice-To analyze the differentiation of ameloblasts and odontoblasts in mutant mice, we performed immunohistochemistry examinations using antibodies to AMEL, AMBN, and DSP. The AMEL and AMBN expression patterns during the secretory stage of ameloblasts were not altered in mutant incisors as compared with those from the wild-type and heterozygote mice (Fig. 4A). Furthermore, by using hematoxylin and eosin staining of dentin and odontoblasts, the width of the predentin and shape of odontoblasts in laminin ␣2 null mice were shown not to be different from those of the wild-type and heterozygote mice. However, DSP expression in odontoblasts was dramatically reduced in the mutant mice teeth (Fig. 4B). In addition, DMP, DSPP, and osteopontin mRNA expressions in dissected incisors were also decreased in two of the mutant mice (Fig. 5), whereas DMP, DSPP, and osteopontin mRNA were highly expressed in those from the wild type. DMP mRNA expression in two mutant strains was dramatically decreased by ϳ60%, whereas DSP and osteopontin mRNA were decreased by ϳ70% (Fig. 5). These results suggest that the differentiation of odontoblasts in the mutant mice was inhibited by the absence of the laminin ␣2 chain.
Laminin-2 Inhibits the Expression of Enamel Matrix and Enhances Dentin Matrix Proteins-BM components in ameloblasts, including laminin and collagen IV, are known to disappear in the secretory stage and reappear in the maturation stage (14,24,25). In the present study, laminin ␣2 showed an expression pattern similar to other BM proteins (Fig. 1). To analyze the effect of laminin-2 on ameloblasts, a rat dental epithelial cell line (HAT-7) was cultured in several extracellular matrix-coated dishes for 3 days, after which marker gene expression was analyzed by RT-PCR. In HAT-7 cells, AMEL expression was inhibited by laminin 2, fibronectin, and type I collagen (Fig. 6A). Furthermore, enamelin and MMP-20 (enamelysin) expression was also decreased in those matrix cultures (Fig. 6A), whereas AMBN expression in HAT-7 was not detected under any of the tested conditions (data not shown). These results suggest that laminin-2 inhibits ameloblast differentiation in the secretory stage. Laminin ␣2 is known to be expressed in dental mesenchymal cells in the early stage and then transiently disappears in mesenchymal cells facing the inner dental epithelium. In the present study, this expression reappeared and continued during dentin formation. To identify the role of laminin ␣2 with odontoblast differentiation, dental mesenchymal cells from P3 molar tooth germ samples were cultured in laminin-2, fibronectin, and type I collagen-coated dishes, and ϳ15% of the cells were found to be DSP-positive by immunostaining (data not shown). Furthermore, the expression of DSPP and DMP was increased 3-4-fold in the laminin-2-coated dish, whereas there were no changes in those coated with fibronectin or type I collagen (Fig. 6B).
No Effect by Laminin-2 on Proliferation and Cell Binding of Dental Mesenchymal Cells-In general, the extracellular matrix is important for proliferation and differentiation of dental epithelium and mesenchymal cells. However, long term cultures of dental epithelial cells was shown to lead to an enhance-FIG. 1. Laminin ␣2 expression in incisor ameloblasts and odontoblasts. A, 3-week-old mice incisor cells were immunostained with anti-laminin ␣2 antibody as described under "Experimental Procedures." a, laminin ␣2 expression was found in the pre-ameloblast and odontoblast interface (arrow), and it continued in the secretory stage (b). B, higher magnification of incisor ameloblasts in each stage. Laminin ␣2 expression was not observed in secretory stage ameloblasts (a). Later, in the early maturation (b) and late maturation (c) stages, laminin ␣2 appeared in the side of the enamel. It was also detected in the outer side of the dental epithelium as well as the papillary cell layer in all stages. C, higher magnification of incisor odontoblasts in each stage corresponding to the ameloblast developmental stages as shown in B. Laminin ␣2 expression was detected in odontoblasts at all stages (a-c). D, primary cultured dental epithelial cells were stained with anti-DSP (a) and laminin ␣2 (b) antibodies. si, stratum intermedium; pa, papillary cell layer; am, ameloblast; od, odontoblast; em, enamel matrix; de, dentin. ment of cell proliferation and decrease of the enamel matrix protein amelogenin in collagen-coated dishes (26). To determine whether the decrease of amelogenin and increase of DMP and DSPP mRNA in laminin-2-coated dishes was dependent on cell proliferation, we analyzed the effect of matrices on cell proliferation (Fig. 7). Cells were cultured with each matrix, and their numbers were counted after 5 days. HAT-7 proliferation in the type I collagen dish was increased (Fig. 7A), but not in the laminin-2 or fibronectin dishes, with similar results ob-served in the primary cultured dental mesenchymal cells (Fig.  7B). Type I collagen coating is known to have a greater effect on the expression of enamel matrix proteins than laminin-2 and fibronectin. Our results suggest that the decrease of enamel matrix expression in the type I collagen-coated dish may have been caused by an increase of cell proliferation. However, the decrease of amelogenin seen in dental epithelial cells and the increase of DMP and DSPP mRNA seen in dental mesenchymal cells cultured with laminin-2 may not have been dependent on cell proliferation.
Laminin is associated with several extracellular matrix and cell receptors that are important for cell binding. We hypothesized that laminin-2 directly binds to dental epithelial and mesenchymal cells and regulates the gene expression of dentin and enamel matrices. To analyze the cell binding activity of laminin-2, cells were incubated in laminin-2-coated microtiter plates as described under the "Experimental Procedures" (Fig.  8). We found that dental epithelial cell binding to laminin-2 was significantly weaker than that to fibronectin and type I collagen (Fig. 8A), whereas dental mesenchymal cells showed little or no binding to the laminin-2-coated plate (Fig. 8B). DISCUSSION This is the first known analysis of laminin ␣2 expression in differentiated ameloblasts and odontoblasts during odontogenesis and tooth development in mutant mice. Laminin ␣2 chain mRNA and protein are expressed in several organs besides striated muscles, including the central nervous system (6), thyroid gland, thymus, kidney, testis, skin, and digestive tract. In addition, laminin ␣2 is also expressed in the tooth germ during amelogenesis and dentinogenesis in a stage-specific manner (14). In the present study, laminin ␣2 chain mRNA was found intensely expressed in dental sac cells. Furthermore, results from immunostaining for laminin ␣2 expression were in conformity with the detected mRNA patterns, indicating a role for the laminin ␣2 chain in the production of BM proteins by dental mesenchymal cells. Salmivirta et al. (14) reported that no laminin ␣ chains (␣1, ␣2, and ␣4) expressed by mesenchymal cells were found in secretory odontoblasts. However, they analyzed molars from embryonic stage and postnatal day 1 mice, during which odontoblasts are partially differentiated. For this reason, we used incisors from 3-week-old mice to investigate the expression of laminin ␣2 chain in fully differentiated ameloblasts and odontoblasts. The BM of the cervical  , c, and d) and mutant (b, d, and e) mice was performed as described under "Experimental Procedures." Higher magnified images of superficial enamel (c and d) and dentinal tubes (d and e) in the boxed region of a and b are also shown. The superficial enamel of the mutant incisors was rough (d) as compared with that from the heterozygotes (c). The dentinal tubes were opened wide in the mutant specimens (e) following treatment with phosphoric acid and sodium hypochlorite.

FIG. 2. Tooth abnormalities in 3-week-old lama2؊/؊ mice.
A, immunostaining with laminin ␣2 in masseter muscle tissues from wild-type (a), heterozygote (b), and mutant (c) mice. B, the color of the mutant incisor surfaces was white (c) as compared with the wild-type (a) and heterozygote (b) mice. C, overall tooth shape was not different between the heterozygotes (a) and mutants (b) in the SEM analysis. D, dentin thickness in the incisor cross-sectional surface area was decreased in the mutant (b) as compared with the heterozygote (a) mice in the SEM analysis. E, dentin thickness in the wild-type and heterozygote mice was not different, but dentin width in the mutants was decreased ϳ25%. Statistical analysis was performed using ANOVA (**, p Ͻ 0.05).
loop in the incisor did not express the laminin ␣2 chain, indicating that there is no expression in inner dental epithelium or dental mesenchymal cells in molars. In contrast, expression was observed in the interface between pre-ameloblasts and odontoblasts, whereas laminin ␣2 transiently disappeared during odontoblast differentiation.
Because dy 3K /dy 3K mice have no laminin ␣2 chain, they are excellently suited for analysis of the biological functions of the laminin ␣2 chain in various organs (12). In the present study, we analyzed laminin ␣2 null mutant mice in order to identify the role of laminin ␣2 in the tooth development. These mice showed irregular enamel surface structures and a decrease of dentin formation, whereas ameloblasts expressed the laminin ␣2 chain in the maturation stage but not in the secretory stage (Fig. 1). Furthermore, enamel abnormalities were only observed in superficial enamel, because of the restriction of laminin ␣2 chain expression, resulting in white colored incisors. However, the structure and size of the enamel crystals were not different between heterozygote and mutant mice, because there was no laminin ␣2 expression in secretory stage ameloblasts.
Amelogenin and ameloblastin, tooth-specific extracellular FIG. 5. Decreased expression of DMP, DSPP, and osteopontin mRNA in lama2؊/؊ mice. Incisors were dissected from the maxillas of wild type (wt) and two mutant (K1 and K2) strains of 3-week-old mice, and then mRNA was isolated and amplified using quantitative RT-PCR, real time PCR, methods with specific primer sets as described under "Experimental Procedures." The expressions of DMP, DSPP, and osteopontin mRNA were decreased in both mutant strains. G3PDH mRNA was used as the control. G3PDH expression was not different between each sample (data not shown). mRNA expression in the mutant samples was compared with that in the wild type. Statistical analysis was performed using ANOVA (*, p Ͻ 0.01).
FIG. 6. Laminin-2 inhibits the expression of enamel matrix and enhances dentin matrix proteins. A dental epithelial cell line (HAT-7) (A) and primary dental mesenchymal cells (B) were cultured in dishes coated with or without laminin-2, fibronectin, or type I collagen for 2 days. mRNA was isolated and amplified using a quantitative RT-PCR method with specific primer sets as described under "Experimental Procedures." AMEL and enamelin (ENAM) expression were decreased in dental epithelial cells in laminin-2, fibronectin, and type I collagen-coated dishes. DMP and DSPP expression were increased in dental mesenchymal cells cultured in laminin-2-coated dishes. G3PDH mRNA was used as the control. G3PDH expression was not different between each sample (data not shown). These experiments were repeated at least three times with similar results. mRNA expressions in each matrix in HAT-7 and in primary cultured dental mesenchymal cells were compared with non-coated dishes. Statistical analysis was performed using ANOVA (*, p Ͻ 0.01).

FIG. 4. Decreased expression of DSP in lama2؊/؊ mice.
A, immunostaining with anti-ameloblastin (a, c, and e) and amelogenin (b, d, and f) antibodies in secretory stage ameloblasts from wildtype (a and b), heterozygote (c and d), and mutant (e and f) mice. Ameloblastin and amelogenin expressions were not different. B, hematoxylin and eosin staining (a, c, and e) and immunostaining with anti-DSP antibody (b, d, and f) in differentiated odontoblasts from wild-type (a and b), heterozygote (c and d), and mutant (e and f) mice. The width of the predentin and shape of the odontoblasts in the mutants (e) were not different from the wildtype (a) or heterozygote mice (c). DSP expression was dramatically decreased in mutant odontoblasts (f). am, ameloblast; od, odontoblast; de, dentin; pde, predentin; dp, dental pulp. matrix proteins, are specifically expressed in secretory ameloblasts (17,(27)(28)(29)(30). In patients, numerous mutations have been found in amelogenin coding sequences, with the most common genetic disorder, AI, affecting enamel (31)(32)(33). In another study, targeted disruption of the amelogenin gene locus in mice caused a hypoplastic enamel phenotype similar to AI, confirming the important role of amelogenin in enamel formation (34). Ameloblastin is also thought to have a relationship with the autosomal dominant type of AI (35). In the present laminin ␣2 null mutants, the expression of these enamel matrix proteins was not changed, indicating that ameloblast differentiation in the secretory stage was undisturbed, because there was no expression of laminin ␣2 found. Furthermore, other BM components, including collagen IV and, notably, BM proteins, were not expressed in secretory stage ameloblasts and reappeared in the maturation stage, indicating that ameloblasts may adhere to enamel surfaces via an extracellular matrix. In fact, laminin-2 was shown to have cell binding activity identical to a dental epithelial cell line, HAT-7, and did not affect cell proliferation in either HAT-7 cells or primary cultured dental mesenchymal cells.
In odontoblasts, laminin ␣2 expression continued during dentinogenesis. A decrease in dentin width and the occurrence of clearly opened dentinal tubes were phenotypic changes more severe than those seen in the enamel, as biomineralization of the dentin extracellular matrix requires complex interactions among several collagenous and non-collagenous molecules. Surprisingly, the expressions of Dspp mRNA and DSP protein were decreased in laminin ␣2 mutant odontoblasts, as Dspp was primarily found expressed by odontoblasts and pre-ameloblasts. Dspp mRNA is translated into a single protein, Dspp, and cleaved into two peptides, dentin DSP and dentin phosphoprotein (DPP), which become localized within the dentin matrix (36 -40). Furthermore, the expression pattern of Dspp in odontoblasts is similar to that of laminin ␣2. Recently, mutations in this gene were identified in human dentinogenesis imperfecta II (Online Mendelian Inheritance in Man accession number 125490) and dentin dysplasia II (Online Mendelian Inheritance in Man accession number 125420) syndromes (41,42). Dspp-null mice were also generated and found to develop tooth defects similar to human dentinogenesis imperfecta III with enlarged pulp chambers, an increased predentin zone width, hypomineralization, and pulp exposure (43). Interestingly, the levels of biglycan and decorin, small leucine-rich proteoglycans, were increased in the widened predentin zone and in void spaces among the calcospherites in the dentin of those null mice. Decorin also functions as an inhibitor of mineralization during primary ossification of mouse embryo bones (44), whereas biglycan facilitates the initiation of apatite formation and inhibits the growth of apatite (45). These results suggest that Dspp is essential for dentin mineralization, including the potential regulation of proteoglycan levels, and is involved in the results of the present study, as the expression of DSP and DMP was decreased in the incisors of laminin ␣2 null mice, indicating an inhibition of odontoblast differentiation. Similar results were observed in primary cultured dental mesenchymal cells.
It was recently reported that a lack of the laminin ␣2 chain results in apoptosis of myogenic cells in vitro, as this chain appears to promote myotube stability by preventing cell death (46). Similar results were observed in the muscle tissue cells of the present laminin ␣2 mutant mice, as there was a markedly high number of apoptotic nuclei that were terminal dUTP nick-end labeling-positive, as compared with wild-type littermates (12). However, in the tooth germ specimens, apoptosis was not observed in ameloblasts or odontoblasts (data not shown), and no changes in the numbers or morphology of these cells were detected. These results suggest that a reduction of dentin formation does not depend on apoptosis of dental epithelium and mesenchymal cells.
Laminin ␣2 did not enhance proliferation and caused no cell binding activity in the primary cultured dental mesenchymal cells; however, it did regulate the expression of dentin matrix. In contrast, type I collagen and fibronectin showed cell binding activity and enhanced the proliferation of dental mesenchymal cells. Matrix expression in primary cultured tooth germ cells was affected by cell proliferation and cell binding (26). In fact, type I collagen enhanced proliferation of both dental epithelial and mesenchymal cells and showed cell binding activity, as well as in a previous report (26). These results suggest that laminin-2 has a different function with dental mesenchymal cells, especially in contrast to odontoblasts. In addition, laminin-2 coating enhanced the expression of Dsp and Dmp, and this effect was specific to laminin-2 and corresponded to the expression of these genes in laminin ␣2 null mutant mice.
We also considered what kind of molecules had an interaction with laminin ␣2 and regulated the expression of dentin matrix proteins during dentin formation. Laminin interacts FIG. 7. Proliferation of HAT-7 and primary dental mesenchymal cells in laminin-2, fibronectin, and type I collagen-coated dishes. A dental epithelial cell line (HAT-7) (A) and primary dental mesenchymal cells (B) were cultured in dishes coated with or without laminin-2 (LN2), fibronectin (FN), or type I collagen (ColI) for 5 days. Cells were prepared and counted as described under "Experimental Procedures." Laminin-2 coating did not affect the proliferation of HAT-7 and primary dental mesenchymal cells. Type I collagen coating enhanced cell proliferation by both types of cells. These experiments were repeated at least three times with similar results. Cell proliferation in each matrix in HAT-7 and in primary cultured dental mesenchymal cells was compared with non-coated dishes. Statistical analysis was performed using ANOVA (*, p Ͻ 0.01).
FIG. 8. Binding of HAT-7 and primary dental mesenchymal cells to laminin-2, fibronectin, and type I collagen. A dental epithelial cell line (HAT-7) (A) and primary dental mesenchymal cells (B) were cultured and plated on microtiter plates coated with laminin-2, fibronectin, or type I collagen. A, cell binding activity was analyzed as described under "Experimental Procedures." HAT-7 cells bound to laminin-2, although the binding activity was lower than to fibronectin and type I collagen. B, dental mesenchymal cells showed little or no binding to laminin-2-coated plates.
with several types of extracellular matrix and cell receptors. For example, laminin 5 (␣3␤3␥2) differentially regulates the anchorage and motility of epithelial cells through integrin ␣ 6 ␤ 4 and ␣ 3 ␤ 1 , respectively (47). Furthermore, targeted disruption of the LAMA3 gene, which encodes the ␣3 subunit of laminin-5, caused an abnormality similar to human junctional epidermolysis bullosa and disturbed ameloblast differentiation (16). These results suggest that laminin is important for tooth formation. The laminin-type G (LG) domain modules consist of ϳ190 residues at the C-terminal of the laminin ␣1 to ␣5 chains (4); however, the function of the G domain of the laminin ␣2 chain, which is shared by laminin-2 and -4, has not been extensively studied. It has also been demonstrated that the absence of ␣2 chains in two mutant mouse strains caused severe muscular dystrophy, presumably because of a strong muscular matrix interaction (11,12). Other indications for the potential functions of ␣2LG modules have come from previous studies of laminin-2 and -4, which demonstrated cell adhesion through ␤ 1 integrin and heparin binding (48), and a distinct interaction with ␣-dystroglycan (49,50). Moreover, the ␤ 1 integrin is expressed in dental mesenchymal cells and differentiated odontoblasts and may interact with laminin ␣2 during dentin formation. In addition, the heparan sulfate proteoglycan perlecan also binds to the LG module of laminin ␣2 (51), is expressed in dental mesenchymal cells, and localized in the BM of the tooth germ (data not shown). Our preliminary experiment showed that cell binding of dental epithelium to laminin-2 was inhibited by the addition of EDTA and RGD peptide. Furthermore, the enhancement of DMP and DSPP mRNA expression in dental mesenchymal cells by laminin-2 was inhibited by the addition of heparin. These results suggest that laminin-2 interacts with integrins in dental epithelial cells and heparan sulfate proteoglycan in dental mesenchymal cells, and regulates the cell binding involved with dental epithelium and dentin matrix expression. These molecules may be candidates for the partner of laminin ␣2 in the tooth germ and are now under investigation in our laboratory using inhibitory antibodies to integrins and recombinant proteins for perlecan.
The present results indicate that absence of the laminin ␣2 chain and altered regulation of dentin matrix proteins may be causative factors that contribute to mineralization defects in tooth developments disorders.