Advertisement
Regular article Epithelial and mesenchymal cell biology| Volume 181, ISSUE 5, P1659-1671, November 2012

Type VII Collagen Deficiency Causes Defective Tooth Enamel Formation due to Poor Differentiation of Ameloblasts

Published:September 03, 2012DOI:https://doi.org/10.1016/j.ajpath.2012.07.018
      Recessive dystrophic epidermolysis bullosa (RDEB) is caused by mutations in the gene encoding type VII collagen (COL7), a major component of anchoring fibrils in the epidermal basement membrane zone. Patients with RDEB present a low oral hygiene index and prevalent tooth abnormalities with caries. We examined the tooth enamel structure of an RDEB patient by scanning electron microscopy. It showed irregular enamel prisms, indicating structural enamel defects. To elucidate the pathomechanisms of enamel defects due to COL7 deficiency, we investigated tooth formation in Col7a1−/− and COL7-rescued humanized mice that we have established. The enamel from Col7a1−/− mice had normal surface structure. The enamel calcification and chemical composition of Col7a1−/− mice were similar to those of the wild type. However, transverse sections of teeth from the Col7a1−/− mice showed irregular enamel prisms, which were also observed in the RDEB patient. Furthermore, the Col7a1−/− mice teeth had poorly differentiated ameloblasts, lacking normal enamel protein–secreting Tomes' processes, and showed reduced mRNA expression of amelogenin and other enamel-related molecules. These enamel abnormalities were corrected in the COL7-rescued humanized mice expressing a human COL7A1 transgene. These findings suggest that COL7 regulates ameloblast differentiation and is essential for the formation of Tomes' processes. Collectively, COL7 deficiency is thought to disrupt epithelial–mesenchymal interactions, leading to defective ameloblast differentiation and enamel malformation in RDEB patients.
      Mesenchymal–epithelial interactions are thought to play essential roles in the development of epithelial organs, including the epidermis, hair follicles, and teeth. Various soluble factors, cell surface markers, and signal molecules have been reported to be involved in mesenchymal–epithelial interactions.
      • Maas R.
      • Bei M.
      The genetic control of early tooth development.
      • Liu F.
      • Chu E.Y.
      • Watt B.
      • Zhang Y.
      • Gallant N.M.
      • Andl T.
      • Yang S.H.
      • Lu M.M.
      • Piccolo S.
      • Schmidt-Ullrich R.
      • Taketo M.M.
      • Morrisey E.E.
      • Atit R.
      • Dlugosz A.A.
      • Millar S.E.
      Wnt/beta-catenin signaling directs multiple stages of tooth morphogenesis.
      Several mouse models with epithelial mesenchymal junction (EMJ) component deficiencies have been developed.
      • Natsuga K.
      • Shinkuma S.
      • Nishie W.
      • Shimizu H.
      Animal models of epidermolysis bullosa.
      Of these, type XVII collagen (COL17)-deficient mice and laminin332-deficient mice show tooth malformation
      • Asaka T.
      • Akiyama M.
      • Domon T.
      • Nishie W.
      • Natsuga K.
      • Fujita Y.
      • Abe R.
      • Kitagawa Y.
      • Shimizu H.
      Type XVII collagen is a key player in tooth enamel formation.
      • Ryan M.C.
      • Lee K.
      • Miyashita Y.
      • Carter W.G.
      Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells.
      and junctional epidermolysis bullosa (EB) caused by deficiencies of these molecules shows the abnormal tooth formation of amelogenesis imperfecta.
      • Fine J.D.
      • Eady R.A.
      • Bauer E.A.
      • Bauer J.W.
      • Bruckner-Tuderman L.
      • Heagerty A.
      • Hintner H.
      • Hovnanian A.
      • Jonkman M.F.
      • Leigh I.
      • McGrath J.A.
      • Mellerio J.E.
      • Murrell D.F.
      • Shimizu H.
      • Uitto J.
      • Vahlquist A.
      • Woodley D.
      • Zambruno G.
      The classification of inherited epidermolysis bullosa (EB): report of the Third International Consensus Meeting on Diagnosis and Classification of EB.
      COL17 and laminin332 play a crucial role in hemidesmosome stability and epithelial mesenchymal attachment. The teeth of COL17-deficient mice exhibit reduced yellow pigmentation, diminished iron deposition, delayed calcification, and irregular enamel prisms. Furthermore, poorly differentiated ameloblasts are revealed in COL17-deficient mice. The teeth of laminin332-deficient mice also have remarkable abnormalities, including disturbance of ameloblast differentiation and reduced enamel deposition. These findings indicate that COL17 and laminin332 deficiency disrupts epithelial–mesenchymal interactions, leading to defective ameloblast differentiation and enamel malformation.
      • Asaka T.
      • Akiyama M.
      • Domon T.
      • Nishie W.
      • Natsuga K.
      • Fujita Y.
      • Abe R.
      • Kitagawa Y.
      • Shimizu H.
      Type XVII collagen is a key player in tooth enamel formation.
      • Ryan M.C.
      • Lee K.
      • Miyashita Y.
      • Carter W.G.
      Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells.
      Thus, it is important to study the pathomechanisms of enamel malformation in mice with defects in other basement membrane zone components.
      Anchoring fibrils are subcellular adhesion structures in the basement membrane zone just beneath the lamina densa between the mesenchymal and epithelial cells, which bind the epithelial cells to the underlying mesenchymal tissue.
      • Tidman M.J.
      • Eady R.A.
      Evaluation of anchoring fibrils and other components of the dermal-epidermal junction in dystrophic epidermolysis bullosa by a quantitative ultrastructural technique.
      Type VII collagen (COL7) is expressed in stratified and complex epithelia, such as the epidermis of skin and the oral mucosal epithelium, and is identified as the major protein component of anchoring fibrils. COL7 plays a crucial role in anchoring fibril formation and mediates dermal–epidermal adherence.
      • Sakai L.
      • Keene D.
      • Morris N.
      • Burgeson R.
      Type VII collagen is a major structural component of anchoring fibrils.
      Anchoring fibrils consist of a central collagenous triple-helical domain flanked by two noncollagenous, globular domains, the NC1 and NC2 domains of COL7. It was reported that the NC1 domain of COL7 binds laminin332 with high affinity.
      • Rousselle P.
      • Keene D.R.
      • Ruggiero F.
      • Champliaud M.-F.
      • van der Rest M.
      • Burgeson R.E.
      Laminin 5 binds the NC-1 domain of type VII collagen.
      The NC1 domain binds predominantly to the β3 chain of laminin332, but also to the γ2 chain.
      • Chen M.
      • Marinkovich M.P.
      • Jones J.C.
      • O'Toole E.A.
      • Li Y.Y.
      • Woodley D.T.
      NC1 domain of type VII collagen binds to the beta3 chain of laminin 5 via a unique subdomain within the fibronectin-like repeats.
      We, therefore, hypothesized that COL7 in anchoring fibrils also plays an important role in mesenchymal–epithelial interactions in tooth formation.
      Recessive dystrophic epidermolysis bullosa (RDEB), which is caused by COL7 deficiency, shows mucocutaneous blistering in response to minor trauma, followed by milium and scar formation, joint contractures, and strictures of the esophagus.
      • Christiano A.
      • Greenspan D.
      • Hoffman G.
      • Hoffman G.
      • Zhang X.
      • Tamai Y.
      • Lin A.
      • Dietz H.
      • Hovnanian A.
      • Uitto J.
      A missense mutation in type VII collagen in two affected siblings with recessive dystrophic epidermolysis bullosa.
      RDEB patients also present with a distinct pattern of oral involvement consisting of microstomia, ankyloglossia, vestibule obliteration, and dental caries.
      • Wright J.T.
      Oral manifestations in the epidermolysis bullosa spectrum.
      • Harris J.C.
      • Bryan R.A.
      • Lucas V.S.
      • Roberts G.J.
      Dental disease and caries related microflora in children with dystrophic epidermolysis bullosa.
      It has been speculated in the literature that the tendency of RDEB patients to have higher caries index is due to structural abnormalities of the teeth. However, it is unknown to what extent the apparently high susceptibility to enamel caries in these cases is due to disease-related altered enamel structure or to low oral hygiene owing to blisters and erosions in the oral mucosa. The debate largely subsided after it was demonstrated that the chemical composition of enamel in RDEB patients is normal.
      • Kirkham J.
      • Robinson C.
      • Strafford S.M.
      • Bonass W.A.
      • Brookes S.J.
      • Wright J.T.
      The chemical composition of tooth enamel in recessive dystrophic epidermolysis bullosa: significance with respect to dental caries.
      Furthermore, it was demonstrated that RDEB patients have enamel surface defects that show similar frequency and distribution to those in healthy controls.
      • Wright J.T.
      • Fine J.D.
      • Johnson L.B.
      Developmental defects of enamel in humans with hereditary epidermolysis bullosa.
      Thus, it has been suggested that tooth malposition and the cross-bite relationship between maxilla and mandible could play a major role in promoting enamel caries.
      • Shah H.
      • McDonald F.
      • Lucas V.
      • Ashley P.
      • Roberts G.
      A cephalometric analysis of patients with recessive dystrophic epidermolysis bullosa.
      Poor oral hygiene conditions can result from the inability to brush the teeth, as tooth brushing is painful and evokes blister formation. This eventually leads to the early onset and severe manifestations of caries. However, it has not been fully elucidated whether the frequent dental caries is due to disease-related structural disruption of the enamel itself or to poor oral hygiene owing to RDEB oral mucosal lesions, because enamel formation is easily disrupted and enamel defects may reflect more than just genetic abnormalities. Enamel defects can also be attributed to environmental factors that cause chronological hypoplasia of the enamel during the enamel formation period.
      • Murrell D.F.
      • Pasmooij A.M.
      • Pas H.H.
      • Marr P.
      • Klingberg S.
      • Pfendner E.
      • Uitto J.
      • Sadowski S.
      • Collins F.
      • Widmer R.
      • Jonkman M.F.
      Retrospective diagnosis of fatal BP180-deficient non-Herlitz junctional epidermolysis bullosa suggested by immunofluorescence (IF) antigen-mapping of parental carriers bearing enamel defects.
      In this context, as a preliminary study, we observed the tooth enamel structure of an RDEB patient and found irregular enamel prisms indicating structural defects of the enamel.
      We previously established Col7a1 knockout (Col7a1−/−) mice,
      • Heinonen S.
      • Männikkö M.
      • Klement J.F.
      • Whitaker-Menezes D.
      • Murphy G.F.
      • Uitto J.
      Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
      a potentially useful model for investigations into the pathomechanisms of enamel defects that arise from defects in anchoring fibrils caused by COL7 deficiency. In the present study, to clarify the roles of COL7 in tooth formation, we studied the detailed process of tooth formation in Col7a1 knockout (Col7a1−/−) mice.
      • Heinonen S.
      • Männikkö M.
      • Klement J.F.
      • Whitaker-Menezes D.
      • Murphy G.F.
      • Uitto J.
      Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
      Tooth development is mediated by reciprocal interdependent epithelial–mesenchymal interactions resulting in the differentiation of epithelial cells into ameloblasts, and mesenchymal cells into odontoblast cells.
      • Thesleff I.
      • Partanen A.M.
      • Vainio S.
      Epithelial-mesenchymal interactions in tooth morphogenesis: the roles of extracellular matrix, growth factors, and cell surface receptors.
      Enamel formation, known as amelogenesis, is divided into three main stages: i) presecretory; ii) secretory; and iii) maturation. In the presecretory stage, epithelial cells differentiate into ameloblasts. In the secretory stage, ameloblasts synthesize and secrete tissue-specific proteins into the developing enamel extracellular matrix. In the maturation stage, removal of organic components and water from the extracellular matrix is followed by the formation of calcium hydroxyapatite crystals, resulting in conversion of the enamel extracellular matrix to fully mineralized enamel.
      We show that COL7 has important roles in tooth formation, especially in enamelization and ameloblast differentiation, suggesting the importance of anchoring fibrils in the mesenchymal–epithelial interaction during tooth formation. In this context, we speculate that the high frequency of enamel caries in RDEB patients might be attributable to enamel malformation caused by the disturbed differentiation of ameloblasts.

      Materials and Methods

      Analysis of an RDEB Patient's Dentition

      We obtained teeth from a 9-year-old female patient with RDEB (Figure 1, A and B). She harbored compound heterozygous mutations in COL7A1, c.6574 + 1G>C and c.8109 + 2T>A, described elsewhere.
      • Sawamura D.
      • Goto M.
      • Yasukawa K.
      • Sato-Matsumura K.
      • Nakamura H.
      • Ito K.
      • Nakamura H.
      • Tomita Y.
      • Shimizu H.
      Genetic studies of 20 Japanese families of dystrophic epidermolysis bullosa.
      Immunofluorescence staining of skin for COL7 showed that the expression level of COL7 was reduced compared with a normal control (Figure 1, C and D), whereas type IV collagen (COL4) and laminin332, which are the major protein components of the lamina densa and anchoring filaments, respectively, in the EMJ, showed equal brightness at the EMJ in the patient and the normal control (Figure 1, E–H). Dental caries was seen on the enamel surface, although she had dental treatment as soon as tooth eruption occurred. She had presented with microstomia; therefore, the permanent first premolars were extracted to prevent tooth malposition. The extracted premolars were prepared for structural observations as follows. The teeth were carefully cleaned and were observed macroscopically. After overnight air-drying, the teeth were sputter-coated with carbon CC-40F (Meiwa-shouji, Osaka, Japan) and observed with a Hitachi S-4000 scanning electron microscope (SEM) (Hitachi Electronics, Tokyo, Japan) operated at 10 kV. For the observation of enamel structure including enamel rods, the teeth were embedded in polyester resin (Rigolic; Ouken Co., Tokyo, Japan). Labiolingual ground sections of both 150-μm and 1000-μm thicknesses were prepared with a rotary diamond saw (Speadrap ML521; Maruto, Tokyo, Japan) and emery papers. The ground sections were etched for 50 seconds in 1 N hydrochloric acid, and the section of 150-μm thickness was processed for H&E staining and observation by light microscopy. The section of 1000-μm thickness was similarly observed with SEM.
      Figure thumbnail gr1
      Figure 1Analysis of the RDEB patient's dentition. A and B: The RDEB patient with compound heterozygous mutations in COL7A1 exhibits mucocutaneous blistering in response to minor trauma. CH: Immunofluorescence staining of the patient's skin for COL7 shows reduced expression of COL7 (C) compared with the normal control (D), whereas type IV collagen (COL4) (E and F) and laminin332 (G and H), which are major protein components of the lamina densa and the anchoring filaments, respectively, in the EMJ, show a brightness at the EMJ that is the same in the patient (E and G) and the normal control (F and H). Separation between epidermis and dermis (asterisk) is seen in the patient's skin (C and G). I and J: Dental caries with whitish color is seen on the enamel surface of the permanent premolar (I: labial surface, J: occlusal surface). K and L: By SEM, the dental caries, indicated by the dotted rectangle in I, shows a rough and pitted appearance on the enamel surface of the patient. A high-power view of the enamel surface is seen in L. M and N: The overall structure of the enamel layer is similar in the labiolingual section of the patient's premolar and the healthy control at the light microscopic level. O and P: Ultrastructurally, in the section of the patient's premolar, the pattern of enamel rods is disrupted in the enamel layer (O) in contrast with the healthy control (P). Scale bars: 10 μm (K); 1 μm (L); 1 mm (M and N); 10 μm (O and P). de, dentin; en, enamel.

      Generation of Col7a1−/− Mice and COL7-Rescued Humanized Mice

      The procedure for generating Col7a1+/− mice has been described.
      • Heinonen S.
      • Männikkö M.
      • Klement J.F.
      • Whitaker-Menezes D.
      • Murphy G.F.
      • Uitto J.
      Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
      The genotype of the Col7a1+/− mice was verified by PCR of the Col7a1 (GenBank accession number U32177) and the neo genes with a template of genomic DNA from tail samples. Col7a1+/− mice were clinically normal and indistinguishable from their wild-type littermates (Col7a1+/+). Heterozygous mice were intercrossed to produce Col7a1-null (Col7a1−/−) offspring. The procedures for screening Col7a1−/− mice by PCR, Northern and Western blotting, histology, electron microscopy, and immunofluorescence are described elsewhere.
      • Heinonen S.
      • Männikkö M.
      • Klement J.F.
      • Whitaker-Menezes D.
      • Murphy G.F.
      • Uitto J.
      Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
      The phenotypic features of the Col7a1 knockout (Col7a1−/−) mice closely resemble those seen in RDEB (OMIM: 226600), caused by null mutations in the COL7A1 gene (GenBank accession number L23982), as previously described.
      • Heinonen S.
      • Männikkö M.
      • Klement J.F.
      • Whitaker-Menezes D.
      • Murphy G.F.
      • Uitto J.
      Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
      The Col7a1−/− mice are readily identified at birth by large, fluid-filled blisters that develop primarily on the ventral side of the animals and by large hemorrhagic blisters on the paws. Skin blisters and erosions readily form from minor trauma. The Col7a1−/− mice skin showed the subepidermal blistering associated with a lack of COL7, and the mice entirely lacked ultrastructurally recognizable anchoring fibrils.
      Procedures for generating COL7-rescued humanized mice have been described elsewhere.
      • Ito K.
      • Sawamura D.
      • Goto M.
      • Nakamura H.
      • Nishie W.
      • Sakai K.
      • Natsuga K.
      • Shinkuma S.
      • Shibaki A.
      • Uitto J.
      • Denton C.P.
      • Nakajima O.
      • Akiyama M.
      • Shimizu H.
      Keratinocyte-/fibroblast-targeted rescue of Col7a1-disrupted mice and generation of an exact dystrophic epidermolysis bullosa model using a human COL7A1 mutation.
      Briefly, we crossed transgenic mice (C57BL/6 background) expressing human COL7A1 cDNA (Col7a1+/+, COL7A1+) driven by the squamous epithelium-specific K14 promoter with heterozygous Col7a1+/− mice. Mice that carried both the heterozygous null mutation in Col7a1 and the transgene of human COL7A1 (Col7a1+/−, COL7A1+) were bred to produce rescued Col7a1−/−, COL7A1+ COL7-humanized mice. The rescued mice lacked the abnormal manifestations seen in the Col7a1−/− mice.
      • Ito K.
      • Sawamura D.
      • Goto M.
      • Nakamura H.
      • Nishie W.
      • Sakai K.
      • Natsuga K.
      • Shinkuma S.
      • Shibaki A.
      • Uitto J.
      • Denton C.P.
      • Nakajima O.
      • Akiyama M.
      • Shimizu H.
      Keratinocyte-/fibroblast-targeted rescue of Col7a1-disrupted mice and generation of an exact dystrophic epidermolysis bullosa model using a human COL7A1 mutation.
      In the COL7-rescued mice (Col7a1−/−, human COL7A1+ mice), intact anchoring fibrils are already observed at the presecretory stage of ameloblast differentiation. Thus, the rescue of deficiency in COL7 by the human gene occurs very early in dental development: the presecretory stage or earlier.

      Structural Analysis of Mouse Dentition

      Tissue samples of mice were incubated in hot (approximately 90°C) distilled water for several minutes and soaked in 10% Tasinase (Kyowa-hakkou, Tokyo, Japan) at 37°C for 6 hours. Incisors and molars were taken from maxillomandibular tissue by removal of soft tissue. The teeth were carefully cleaned and were observed macroscopically. After overnight air drying, the teeth were sputter-coated with carbon CC-40F (Meiwa-shouji) and observed with a Hitachi S-4000 SEM operated at 10 kV. For the observation of enamel rod inclination, incisors were embedded in polyester resin (Rigolic; Ouken Co.). Transverse labiolingual ground sections of 1-mm thickness were prepared with a rotary diamond saw and emery papers. The ground sections were etched for 30 seconds in 0.1 N hydrochloric acid and were observed similarly.

      Chemical and Mineralization Analysis

      Qualitative and distributive elemental analyses were performed for polyester resin–embedded transverse ground sections of incisors and molars that had been prepared with a rotary diamond saw and emery paper. These analyses were done using a Hitachi S-2380 SEM, operated at 15 kV, and energy-dispersive X-ray spectrometry.
      To demonstrate the patterns of mineralization, micro-computed tomography (CT) scanning was performed using an R_mCT2 (Rigaku, Tokyo, Japan). The maxillae and mandibles with incisors were obtained and dehydrated by passage through a series of ascending concentrations of ethanol solution before being embedded in polyester resin. The samples were placed longitudinally into a sample holder for CT scanning. The scanning protocol was set at an X-ray energy setting of 90 kV and 160 μA performed for one full 360° rotation. From the scans, a three-dimensional reconstruction of the sample was made using software supplied with the scanner, resulting in datasets with an isotropic voxel size of 20 μm.

      Preparation of Tissue Sections and Immunohistochemistry

      Under anesthesia with ether inhalation, intracardiac perfusions of the mice were performed with a fixative solution containing 4% paraformaldehyde in PBS, at pH 7.4. Postfixation was ensured by immersion of the dissected maxilla and mandible in the fixative solution overnight at 4°C.
      The maxillae and mandibles with incisors and molars were processed for histological analysis by decalcification at 4°C for up to 2 weeks in PBS solution (pH 7.4) containing 5% EDTA. After extensive washing in PBS, the samples were dehydrated in increasing concentrations of ethanol and Lemosol (Wako, Osaka, Japan) before being embedded in paraffin. Serial longitudinal sections of the incisors of the paraffin-embedded specimens (5 μm) were processed for H&E staining and for Berlin blue staining to detect iron deposits.
      For immunohistochemistry, neonatal mice were sacrificed and the tissue samples were embedded in optimal cutting temperature compound (Sakura Finetechnical Co., Tokyo, Japan) for frozen sectioning. Frozen tissue sections were cut sagittally to a thickness of 6 μm until incisors and molars were exposed. After air drying for several minutes, sections were washed in PBS and incubated with a primary antibody—either anti-mouse COL7 polyclonal antibody (Calbiochem, Darmstadt, Germany; final dilution of 1:200), anti-mouse COL4 polyclonal antibody (Abcam, Cambridge, UK; final dilution of 1:400), or anti-mouse laminin332 polyclonal antibody (Abcam; final dilution of 1:50)—at 37°C for 30 minutes. The sections were then incubated with a secondary antibody—fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (H+L; Jackson ImmunoResearch Laboratories, Newmarket, UK; final dilution, 1:50, 1:200, 1:50)—at 37°C for 30 minutes and incubated with 10 μg/mL of propidium iodide at 37°C for 10 minutes for nuclear counterstaining. Sections were observed under an Olympus Fluoview confocal laser-scanning microscope (Olympus, Tokyo, Japan).

      Ultrastructural Analysis during Tooth Formation

      As above, mice incisors and molars were obtained from the maxillomandibular tissue fixed with modified Karnovsky's fixative [final concentration of 2% paraformaldehyde and 2.5% glutaraldehyde in 0.05 mol/L cacodylate buffer solution (pH 7.4)], and decalcified in 5% EDTA (pH 7.4) at 4°C for 2 weeks. After decalcification, samples were postfixed in 1% osmium tetroxide at 4°C for 2 hours and stained en bloc with 1% uranyl acetate at 4°C for 20 minutes. The samples were dehydrated through a graded series of ethanol and embedded in Epon 812 (TAAB Laboratories, Aldermaston, UK). Ultrathin sections were cut in the sagittal direction to include both the separated enamel organ and the dental papilla. Sections were stained with uranyl acetate and lead citrate, and observed under a Hitachi H-7100 transmission electron microscope.

      Cell Cultures and Immunolabeling

      For dental epithelial cell cultures, maxillary and mandibular incisors and molars from mice were dissected. Tooth samples were treated with 0.25% trypsin for 10 minutes and pipetted up and down intensely. Dental epithelial cells, dental mesenchymal cells and various other cells were isolated from tooth buds. To separate dental epithelial cells from the other cells, the cells were cultured in epidermal keratinocyte medium containing a small amount of bovine pituitary extract (CnT-57; CELLnTEC Advanced Cell Systems, Bern, Switzerland) for 7 days. After obtaining a sufficient number of dental progenitor epithelial cells, we changed the culture medium to epidermal keratinocyte medium containing 0.07 mmol/L calcium (CnT-02; CELLnTEC Advanced Cell Systems) to induce differentiation, and cultured the cells for 10 days.
      For fluorescence staining, the cells were fixed with 70% ethanol for 10 minutes and washed with PBS. The cells were incubated with a primary antibody anti-mouse amelogenin polyclonal antibody (Hokudo, Sapporo, Japan), final dilution of 1:100, or with anti-mouse ameloblastin polyclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), final dilution of 1:50, at 37°C for 30 minutes. The cells were then incubated with the secondary antibody (FITC-conjugated goat anti-rabbit IgG; H+L; Jackson ImmunoResearch Laboratory, West Grove, PA), with a final dilution of 1:50, at 37°C for 30 minutes and incubated with 10 μg/mL of propidium iodide at 37°C for 10 minutes to visualize the nucleus. The cells were observed under an Olympus Fluoview confocal laser-scanning microscope (Olympus, Tokyo, Japan).

      RT-PCR Analysis

      To study Col7a1 mRNA expression in dental epithelial cells and ameloblasts, total RNA from tooth buds or cultured dental epithelial cells was extracted using the RNeasy Mini Kit (Qiagen, Tokyo, Japan) according to the manufacturer's instructions. RNA concentration was measured spectrophotometrically and samples were stored at −80°C until being used for reverse transcription-PCR (RT-PCR). We reverse transcribed RNA using SuperScript II (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. The following primers specific for mouse Col7a1 sequence (NM: 007738) were used for RT-PCR: 5′-CGAGGAAGAGATGGTGAAGC-3′ (RT-F), and 5′-TTGCCTGAAGCACCATGTAG-3′ (RT-B). As a control, we used the primers for mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH; NM: 001001303): 5′-TTAGCCCCC CTGGCCAAGG-3′ (mGAPDH-F) and 5′-CTTACTCCTTGGAGGCCATG-3′ (mGAPDH-B), which amplified a 541-bp fragment.

      Real-Time RT-PCR Analysis

      To quantitatively analyze the mRNA expression levels of tooth formation–associated proteins [ie, amelogenin, ameloblastin, enamelin, tuftelin, enamelysin, and dentin sialophosphoprotein (DSPP)], in tooth buds from the Col7a1+/+, Col7a1−/−, and COL7-rescued mice, cDNA samples were analyzed using the ABI Prism 7000 Sequence Detection System (Applied Biosystems, Foster City, CA). Primers and probes specific for amelogenin, ameloblastin, enamelin, tuftelin, enamelysin, DSPP, and the control housekeeping genes coding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and β-actin were obtained from the TaqMan Gene Expression Assay (Applied Biosystems; Probe ID: Mm00711644_g1, Mm00477485_m1, Mm00516922_m1, Mm00449139_m1, Mm00600244_m1, Mm00515666_m1, Mm99999915_gl, and Mm00607939_sl).
      Differences between the mean cycle threshold (Ct) values of mRNA expression of tooth formation–associated proteins and those of GAPDH or ß-actin were calculated as ΔCtCol7a1−/− mice = Cttooth protein − CtGAPDH (or other housekeeping genes) and those of ΔCt for the Col7a1+/+ teeth as ΔCtcalibrator = Cttooth protein − CtGAPDH (or other housekeeping genes). We were able to obtain similar results from the GAPDH and β-actin standard; thus, we describe the results of the GAPDH standard in the present study. Final results for Col7a1−/− tooth samples/Col7a1+/+ tooth samples (%) were determined by 2−(ΔCt Col7a1−/− − ΔCtcalibrator).
      Using similar methods, we quantitatively analyzed the tooth formation–associated protein mRNA expression levels in the primary dental epithelial cells cultured from the Col7a1+/+, Col7a1−/−, and COL7-rescued mice.

      Western Blot Analysis

      To study COL7 and laminin332 expression in the ameloblasts cultured from the Col7a1+/+ and Col7a1−/− mice, we used proteins extracted from cultured ameloblasts prepared with radioimmunoprecipitation assay buffer comprising 50 mmol/L Tris-HCl (pH 7.5), 150 mmol/L NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, and Roche Protease Cocktail Tablet (Roche, Basel, Switzerland). The proteins were fractionated by SDS-PAGE and transferred to polyvinylidene difluoride membranes for immunoblotting. Membrane blocking and incubation with antibodies were performed in Tris-buffered saline with 2% nonfat dry milk. The following primary antibodies were used: anti-mouse COL7 polyclonal antibody (Calbiochem; final dilution of 1:1000), anti-mouse laminin332 polyclonal antibody (Abcam; final dilution of 1:1000), and anti-mouse β-actin monoclonal antibody (Sigma, St. Louis, MO; final dilution of 1:1000). The following secondary antibodies were used: horseradish peroxidase–conjugated goat anti-rabbit IgG or horseradish peroxidase–conjugated goat anti-mouse IgG (Invitrogen; final dilution of 1:5000). Signals were revealed with chemiluminescence reagents and photographed by LAS-1000 mini digital camera (Fujifilm, Tokyo, Japan).

      Results

      Structural Defects of the Enamel in the RDEB Patient

      We examined teeth from a 9-year-old female patient with RDEB resulting from COL7 deficiency (Figure 1, I–P). Under macroscopic observation, dental caries with a whitish appearance was widely seen on the enamel surface of permanent premolars (Figure 1, I and J). The enamel surface of the patient was rough and pitted under SEM (Figure 1, K and L). In the labiolingual section of the patient's teeth, the overall structure of the enamel layer was similar to healthy control at the light microscopic level (Figure 1, M and N). However, ultrastructurally, the pattern of enamel rods was disrupted in the enamel layer in contrast with the normal human control (Figure 1, O and P).

      COL7 Expression Pattern in the EMJ of Teeth in the Col7a1−/− Mice

      To clarify COL7 expression during tooth formation, we immunostained tissue sections of neonatal incisors and molars. In the incisors, we were able to observe the presecretory and secretory stages of tooth formation (Figure 2D). COL7 was expressed in the EMJ between ameloblasts and odontoblasts at the presecretory stage in wild-type (Col7a1+/+) mice and heterozygous (Col7a1+/−) mice (data not shown). Due to elongation of Tomes' processes, the basement membrane became discontinuous, and COL7 expression was gradually reduced and became absent in places at the secretory stage. In molars, we were able to observe the presecretory stage of tooth formation. In the molars of Col7a1+/+ and Col7a1+/− mice, COL7 was expressed in the EMJ between ameloblasts and odontoblasts at the presecretory stage (data not shown).
      Figure thumbnail gr2
      Figure 2COL7 expression in the tooth of Col7a1+/+ mice, and COL7 absence in the tooth of Col7a1−/− mice. A: Mice incisors are continuously elongating; shown are the teeth of 4-month-old mice. B and C: Schematic of mouse upper incisor. A sagittal section is shown in C. In the root of these incisors, teeth develop through three main stages: i) presecretory (*); ii) secretory (**); and iii) maturation (***). In the presecretory stage, epithelial cells differentiate into ameloblasts (blue), and mesenchymal cells differentiate into odontoblasts (green). In the secretory stage, ameloblasts and odontoblasts secrete enamel matrix (sky blue) and dentin (gray). In the maturation stage, mineralization of the enamel matrix to enamel (yellow) occurs. Red line: basement membrane zone (BMZ). D: Immunofluorescence staining for COL7 in neonatal mice incisors reveals that COL7 is expressed in the EMJ between ameloblasts and odontoblasts at the presecretory stage (white arrowheads) of a Col7a1+/+ mouse. At the secretory stage, COL7 expression (white arrowheads) is gradually reduced and becomes absent (white arrow). In the Col7a1−/− mice, no COL7 staining is observed in any part of the EMJ. COL4 is expressed in the EMJ (white arrowheads) in both Col7a1+/+ and Col7a1−/− mice incisors. At the secretory stage of incisors, the basement membrane becomes discontinuous, and COL4 expression is gradually reduced and absent in places (white arrow). Scale bars = 40 μm. Shown are neonatal mice incisors. E: RT-PCR assay reveals that Col7a1 mRNA (506-bp band) is expressed in Col7a1+/+ mouse teeth (left lane) and cultured ameloblasts from Col7a1+/+ mice (second right lane). Col7a1 mRNA is not expressed in either Col7a1−/− mouse teeth (second left lane) or in cultured ameloblasts from Col7a1−/− mice (right hand lane). Shown are 3-day-old mice incisors and molars. F: Ultrastructural features of the basement membrane zone at the presecretory stage. Anchoring fibrils, which anchor lamina densa to the underlying mesenchymal tissue, and anchoring filaments crossing the lamina lucida are seen in the Col7a1+/+ mouse (left, white arrowheads), but there are no apparent anchoring fibrils in the Col7a1−/− mice, even though the lamina densa is normally formed (right). am, ameloblasts; IP, inner attachment plaques; LD, lamina densa; LL, lamina lucida; od, odontoblast. Scale bar = 250 nm. Shown are neonatal mice molars.
      In the Col7a1−/− mice, COL7 expression was not observed in the EMJ underlying the ameloblasts at any stage of tooth development.
      We also studied COL4 expression during tooth formation (Figure 2D) because COL4 is a major protein component of the lamina densa in the EMJ.
      • Natsuga K.
      • Shinkuma S.
      • Nishie W.
      • Shimizu H.
      Animal models of epidermolysis bullosa.
      In the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice, COL4 was expressed in the EMJ between ameloblasts and odontoblasts at the presecretory stage. At the secretory stage, the basement membrane became discontinuous, and COL4 expression was gradually reduced and absent in places.
      In addition, we examined expression of laminin332, a component of anchoring filaments, during tooth formation, because this protein interacts directly with COL7 and is already known to affect ameloblast differentiation. In the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice, laminin332 was seen at the EMJ between ameloblasts and odontoblasts at the presecretory stage (data not shown). At the secretory stage, laminin332 staining was weak and negative in places (data not shown).
      We subsequently analyzed mouse Col7a1 mRNA expression in tooth buds or cultured dental epithelial cells by RT-PCR (Figure 2E). The 506-bp fragment of mouse Col7a1 mRNA was detected in the Col7a1+/+ mouse teeth in vivo and in cells cultured from the Col7a1+/+ mouse teeth in vitro, but mouse Col7a1 mRNA was not detected in either teeth or cultured cells from the Col7a1−/− mice (Figure 2E).
      In addition, we analyzed mouse COL7 and laminin332 protein expression in cultured dental epithelial cells by Western blotting. COL7 was not detected in cell lysates from the Col7a1+/+ or Col7a1−/− mice (data not shown), although weak laminin332 protein expression was detected in the cultured cells from the Col7a1+/+ and Col7a1−/− mice (data not shown).
      The basement membrane on the basal surface of the ameloblasts separates the ameloblasts from mesenchymal tissue/odontoblasts. Beneath the lamina densa, most of the COL7 molecules form semicircular loop structures called anchoring fibrils whose terminals originate and terminate in the lamina densa, similar to those at the dermoepidermal junction of the skin.
      • Shimizu H.
      • Ishiko A.
      • Masunaga T.
      • Kurihara Y.
      • Sato M.
      • Bruckner-Tuderman L.
      • Nishikawa T.
      Most anchoring fibrils in human skin originate and terminate in the lamina densa.
      Ultrastructurally, in the underlying mesenchymal tissue, anchoring fibrils enable the lamina densa to link or encircle mesenchymal collagen fibers or other components to anchor the basal lamina to underlying structures. Thin anchoring filaments cross the lamina lucida zone, subjacent to hemidesmosomes, and extend into the lamina densa in the Col7a1+/+ mice (Figure 2F, left) and the Col7a1+/− mice (data not shown). In the Col7a1−/− mice, there were no apparent anchoring fibrils, even though the lamina densa was normally formed (Figure 2F, right).

      Dental Phenotype of the Col7a1−/− Mice

      The Col7a1−/− mice demonstrated a striking skin phenotype noted at birth or shortly thereafter: A large blister developed primarily on the ventral side of the mice, often extending to cover the extremities (Figure 3A). Some of the affected mice died within the first week of life from complications of the disease, whereas the others survived beyond the first week of life and demonstrated normal hair growth that appeared to coincide with lessening of the blistering tendency in the skin (Figure 3A).
      Figure thumbnail gr3
      Figure 3Dental phenotype of Col7a1−/− mice. A: Clinical presentation of Col7a1−/− mice (lower) compared with Col7a1+/+ littermate (upper). Typical blister and scar seen on paws (inset). Malformed ears (inset). B: At 2 weeks of age, Col7a1+/+ and Col7a1−/− mice have whitish incisors (left). There is no apparent difference in enamel shape and surface structure between these mice. At 4 months of age, incisors of the Col7a1+/+ and the Col7a1−/− mice show similar yellowish color (right). There is also no apparent difference in shape and surface structure between incisors of the Col7a1+/+ and Col7a1−/− mice. Scale bars: 500 μm. C: By SEM, the enamel surface of incisal teeth and molars in the Col7a1+/+ and Col7a1−/− mice appear smooth and unpitted. In the molars, tooth wear is equal between Col7a1+/+ and Col7a1−/− mice. Scale bars = 200 μm. Shown are 4-month-old mice incisors and molars.
      At 2 weeks of age, shortly after eruption of incisors, the wild-type (Col7a1+/+), heterozygous (Col7a1+/−), and Col7a1−/− mice had whitish incisors (Figure 3B). There was no apparent difference in enamel shape and surface structure among these three types of mice under macroscopic observation. The COL7-rescued mice (mouse Col7a1−/−, human COL7A1+) had similar incisors, as did these three types of mice (data not shown). At 4 months of age, incisors of the Col7a1+/+ and Col7a1−/− mice showed similar yellowish color. There was also no apparent difference in shape or surface structure between incisors of the Col7a1+/+ and Col7a1−/− mice (Figure 3B). As for the oral mucosa and soft tissue manifestations, the Col7a1−/− mice showed extreme fragility of the oral and perioral mucosa. In the maxillary mucosa, the continual process of blister formation and healing resulted in loss of normal anatomical oral features, such as the palatal rugae, leaving a smooth and ulcerated roof of the mouth (data not shown).
      Under SEM, the enamel surfaces of incisal teeth and molars in the Col7a1+/+(Figure 3C), Col7a1+/− (data not shown), and Col7a1−/− (Figure 3C) mice all appeared smooth and unpitted. In the molars, tooth wear was equal between the Col7a1+/+ and Col7a1−/− mice. In transverse sections of the Col7a1−/− mice incisors (Figure 4, C and F), the enamel rod inclination was irregular, and the rods had lost the normal network arrangement that was seen in the Col7a1+/+ mice (Figure 4, B and E) and Col7a1+/− mice (data not shown). In the COL7-rescued mice (Col7a1−/−, human COL7A1+ mice), enamel rod formation was restored to normal (Figure 4, D and G), confirming that the enamel changes were caused by COL7 deficiency.
      Figure thumbnail gr4
      Figure 4SEM of the transverse sections of incisors. A: Model of an upper incisor. A transverse section of the incisor (green line) is shown at the bottom. The enamel layer, indicated by the rectangle, is enlarged in B and C. High-power views of the enamel rod inclinations in BD are shown in EG, respectively. BG: In the Col7a1−/− mouse, irregular inclinations of enamel rods without a normal network arrangement are observed (C and F), in contrast to the regular network of enamel rods observed in the Col7a1+/+ incisor (B and E). The normal, regular network of enamel rods has been restored in the COL7-humanized mouse (D and G). Black dotted line in EG: enamel rod. de, dentin; en, enamel. Scale bars: 100 μm (A); 20 μm (BD); 5 μm (EG). Shown are 15-day-old mice incisors.

      Chemical and Mineralization Analysis of the Teeth

      Backscatter electron images of the transverse labiolingual sections of the incisors in the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice revealed calcium and phosphorus to be homogeneously distributed in all samples (data not shown). Iron was also homogeneously distributed on the enamel surface. There were no significant differences in the atomic concentrations of calcium and phosphorus between enamel of 4-month-old Col7a1+/+ and Col7a1−/− mice. The calcium concentration was 25.4 ± 1.2% for Col7a1+/+ and 24.9 ± 1.5% for Col7a1−/−, and the phosphorus concentration was 19.6 ± 0.60% for Col7a1+/+ and 18.9 ± 1.6% for Col7a1−/−. Energy-dispersive X-ray spectrometry spectra at a number of spots showed that there was no significant difference between the Col7a1+/+ and Col7a1−/− mice in terms of the concentrations of calcium, phosphorus, iron, and other ions on the surface of the enamel (data not shown).
      To compare the mineralization patterns of teeth between the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice, micro-CT scanning was performed on the maxillomandibular incisors. Calcifications of skull and incisor were clearly visible. In the incisors of the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice, the radio-opacity of the enamel increased gradually toward the incisal edge, from the enamel secretory stage to the maturation stage. Mineralization reached its maximum at the incisal edge (data not shown). To objectively evaluate the mineralization level in the enamel layers, we measured the mean CT number of the incisal edge in Hounsfield units (HU). There was no significant difference in the CT number of the incisal edge between the 20-day-old Col7a1+/+ and Col7a1−/− mice. That number was 1070.26 ± 33.40 HU for Col7a1+/+ and 1078.61 ± 33.51 HU for Col7a1−/−. These results show that mineralization of the enamel in the Col7a1−/− incisors was similar to that in the Col7a1+/+ incisors.

      Defective Amelogenesis in Col7a1−/− Mice

      We observed the tooth development at each of the three stages of enamel formation: presecretory, secretory, and maturation (Figure 5, A–C). There is no apparent blistering in the EMJ between ameloblasts and odontoblasts in the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice. Ameloblast size and enamel matrix thickness in the Col7a1−/− mice were similar to those in the Col7a1+/+ and Col7a1+/− mice (Figure 5, D, F, H, and J). The Tomes' processes of the Col7a1+/+ and Col7a1+/− mice were triangular and orderly. By contrast, the processes of the Col7a1−/− mice were deformed and difficult to clearly visualize in H&E-stained sections (Figure 5, D and F).
      Figure thumbnail gr5
      Figure 5Malformed Tomes' processes in Col7a1−/− mice and defective amelogenesis in Col7a1−/− mice. A: Three stages of ameloblast differentiation are observed in sagittal sections of mice incisors: i) presecretory stage (asterisk); ii) secretory stage (double asterisk); and iii) maturation stage (triple asterisk). B and C: The secretory and maturation stage of enamel formation is enlarged in DK. Rectangles with red and blue dotted lines are enlarged in DG and in HK, respectively. DG: At the secretory stage, Tomes' processes have formed and enamel matrix is produced by ameloblasts. D and F: In the secretory stage, the processes of ameloblasts are malformed and blurred (arrows) in the Col7a1−/− mouse (F), compared with well-organized lattice-like structures of the Tomes' processes (arrows) in the Col7a1+/+ mice (D). The thickness of the enamel matrix seems similar between the Col7a1+/+ (D) and Col7a1−/− (F) mice. E and G: At the secretory stage, in the Col7a1−/− mice, the Tomes' processes (white arrows) are hypoplastic and irregular in size and width, showing a less prominent appearance compared with normal Tomes' processes in the Col7a1+/+ mouse. HK: At the maturation stage, the cell structure and organelles of Col7a1−/− mature ameloblasts appear normal; however, the enamel rods are rough and irregularly distributed (asterisk) compared with those of the Col7a1+/+ mice. am, ameloblast; de, dentin; em, enamel matrix; en, enamel; od, odontoblast; tp, Tomes' processes. Scale bars: 400 μm (B and C); 40 μm (D, F, H, and J); 4 μm (E and G); 200 nm (I and K). Shown are 5-day-old mice incisors.
      Furthermore, enamel formation of the teeth of the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice was studied ultrastructurally. In the presecretory to the early secretory stage, the overall structure of ameloblasts was similar between the Col7a1+/+ and Col7a1−/− mice. In the presecretory stage, anchoring fibrils were identified in the Col7a1+/+ mice (Figure 2F, left) and Col7a1+/− mice, but not in the Col7a1−/− mice (Figure 2F, right). However, in the early secretory stage, only thin filamentous material was observed beneath the basal lamina at the EMJ in the Col7a1+/+, Col7a1+/−, and Col7a1−/− mice (data not shown). In the secretory stage, secretory ameloblasts appeared as tall columnar cells with intact Tomes' processes producing enamel matrix in the Col7a1+/+ (Figure 5E) and Col7a1+/− (data not shown) mice.
      In the Col7a1−/− mice, the Tomes' processes were disorganized and irregular in size and width, showing a less prominent appearance compared with those of the Col7a1+/+ mice (Figure 5G). There was no obvious abnormality in the other structural components of the ameloblasts. During the secretory stage, the electron density of the enamel matrix of the Col7a1−/− mice was similar to that of Col7a1+/+ and Col7a1+/− mice.
      In the maturation stage, mature ameloblasts were columnar cells that had rough endoplasmic reticulum, lysosomes, mitochondria, small vacuoles and Golgi apparatuses in the apical and mid portions. The cell structure and organelles of Col7a1−/− mature ameloblasts appeared normal, but the enamel rods were rougher and more irregularly distributed (Figure 5, J and K) than those of the Col7a1+/+ mice (Figure 5, H and I) and Col7a1+/− mice (data not shown).
      In the COL7-rescued mice (Col7a1−/−, human COL7A1+ mice), the overall structure of ameloblasts was similar to those in the Col7a1+/+ mice, and the Tomes' processes were restored to normal (data not shown).
      We performed Berlin blue staining in the maturation stage of 11-day-old Col7a1+/+ and Col7a1−/− mice teeth. In both mice, positive staining of Berlin blue was detected in the ameloblast cytosol, indicating the accumulation of iron, during the maturation stage (data not shown). There was no obvious difference in the Berlin blue staining patterns between the Col7a1+/+ and Col7a1−/− mice, suggesting that the aberrant differentiation of ameloblasts in the Col7a1−/− mice did not affect iron metabolism.

      Assay of Ameloblast Differentiation

      We examined the expression of enamel proteins in the teeth in vivo and in cultured dental epithelial cells in vitro using real-time RT-PCR analysis.
      • Fukumoto S.
      • Kiba T.
      • Hall B.
      • Iehara N.
      • Nakamura T.
      • Longenecker G.
      • Krebsbach P.H.
      • Nanci A.
      • Kulkarni A.B.
      • Yamada Y.
      Ameloblastin is a cell adhesion molecule required for maintaining the differentiation state of ameloblasts.
      • Fukumoto S.
      • Yamada A.
      • Nonaka K.
      • Yamada Y.
      Essential roles of ameloblastin in maintaining ameloblast differentiation and enamel formation.
      • Masuya H.
      • Shimizu K.
      • Sezutsu H.
      • Sakuraba Y.
      • Nagano J.
      • Shimizu A.
      • Fujimoto N.
      • Kawai A.
      • Miura I.
      • Kaneda H.
      • Kobayashi K.
      • Ishijima J.
      • Maeda T.
      • Gondo Y.
      • Noda T.
      • Wakana S.
      • Shiroishi T.
      Enamelin (Enam) is essential for amelogenesis: eNU-induced mouse mutants as models for different clinical subtypes of human amelogenesis imperfecta (AI).
      mRNA expression of the major enamel proteins produced by ameloblasts in vivo, including amelogenin, ameloblastin, enamelin, enamelysin, and DSPP, was significantly decreased in the Col7a1−/− teeth, except for the expression of tuftelin (Figure 6A), which was only slightly reduced. Real-time RT-PCR analysis demonstrated that in the teeth of COL7-rescued mice, in vivo mRNA expression of the major enamel proteins was restored (data not shown). In dental epithelial cells cultured from the Col7a1+/+ mice, mRNA expression of amelogenin, ameloblastin, and tuftelin was confirmed, although mRNA expression of enamelin, enamelysin, and DSPP was absent. In the Col7a1−/− mice, mRNA expression of amelogenin and tuftelin in cultured cells were significantly lower than in the Col7a1+/+ mice. Ameloblastin expression in the cultured Col7a1−/− cells was higher than in the cells cultured from the Col7a1+/+ mice (data not shown). Real-time RT-PCR analysis revealed that, in the COL7-rescued mice, mRNA expression of amelogenin and tuftelin in cultured cells was restored (data not shown). Immunocytological examinations revealed granular expression of amelogenin in the ameloblasts cultured from the Col7a1+/+ (Figure 6B) and COL7-rescued mice (data not shown), although no expression of amelogenin was observed in the cells cultured from the Col7a1−/− mice (Figure 6C). Expression of ameloblastin was similar in cells cultured from the Col7a1+/+, Col7a1−/−, and COL7-rescued mice (data not shown).
      Figure thumbnail gr6
      Figure 6Expression of enamel proteins in Col7a1−/− ameloblasts. A: mRNA expression of all of the enamel proteins examined (amelogenin, ameloblastin, enamelin, tuftelin, enamelysin, and DSPP) was down-regulated in ameloblasts of the teeth of the Col7a1−/−mice in vivo. (3-, 4-, and 5-day-old Col7a1+/+ mice incisors and molars, n = 3; 3-, 4-, and 5-day-old Col7a1−/− mice incisors and molars, n = 3). B and C: Protein expression (FITC, green) of amelogenin is decreased in ameloblasts cultured from the Col7a1−/− mice (C), relative to that in ameloblasts cultured from the Col7a1+/+ mice (B). Insets: high-power views. Scale bars = 10 μm. Shown are 8-day-old mice incisors.

      Discussion

      RDEB is a hereditary blistering skin disease with tissue separation occurring beneath the lamina densa of the epidermal basement membrane zone. RDEB is characterized by mucocutaneous blistering in response to minor trauma, followed by ulceration and scarring mainly in the hands and feet, joint contractures, strictures of the esophagus and dental abnormalities.
      • Christiano A.
      • Greenspan D.
      • Hoffman G.
      • Hoffman G.
      • Zhang X.
      • Tamai Y.
      • Lin A.
      • Dietz H.
      • Hovnanian A.
      • Uitto J.
      A missense mutation in type VII collagen in two affected siblings with recessive dystrophic epidermolysis bullosa.
      Molecular genetic studies have revealed that RDEB is caused by mutations in the gene that encodes COL7, a major component of anchoring fibrils.
      • Christiano A.
      • Greenspan D.
      • Hoffman G.
      • Hoffman G.
      • Zhang X.
      • Tamai Y.
      • Lin A.
      • Dietz H.
      • Hovnanian A.
      • Uitto J.
      A missense mutation in type VII collagen in two affected siblings with recessive dystrophic epidermolysis bullosa.
      • Uitto J.
      • Pulkkinen L.
      • Christiano A.M.
      Molecular basis of the dystrophic and junctional forms of epidermolysis bullosa: mutations in the type VII collagen and kalinin (laminin 5) genes.
      • Christiano A.M.
      • Anhalt G.
      • Gibbons S.
      • Bauer E.A.
      • Uitto J.
      Premature termination codons in the type VII collagen gene (COL7A1) underlie severe, mutilating recessive dystrophic epidermolysis bullosa.
      Most RDEB patients exhibit oral mucous membrane blistering, and prevalent tooth abnormalities with caries.
      • Wright J.T.
      Oral manifestations in the epidermolysis bullosa spectrum.
      • Harris J.C.
      • Bryan R.A.
      • Lucas V.S.
      • Roberts G.J.
      Dental disease and caries related microflora in children with dystrophic epidermolysis bullosa.
      The present study revealed that the secretory ameloblasts of the Col7a1−/− mice lacked normal enamel protein-secreting Tomes' processes and exhibited disturbed enamel matrix secretion, which resulted in imperfect amelogenesis demonstrated by malformed enamel rods and irregular enamel matrix (Figure 7).
      Figure thumbnail gr7
      Figure 7Diagrams of normal enamel formation in Col7a1+/+ mice and defective enamel formation in Col7a1−/− mice. At the presecretory stage, hypoplasia of anchoring fibrils is the only apparent abnormality in the Col7a1−/− teeth (right), whereas anchoring fibrils are seen in the Col7a1+/+ teeth (left). From the secretory to the maturation stages, normal enamel matrix is formed by Tomes' processes, resulting in intact enamel formation in the Col7a1+/+ teeth (left). In the Col7a1−/− teeth (right), disrupted Tomes' processes produce disturbed enamel matrix, leading to irregular enamel formation.
      Mice have only one set of dentition, whereas humans have two: primary and secondary. As for the total number and the type of teeth, mice have 16 teeth, and these are classified as incisors and molars, whereas humans usually have 20 primary teeth and 32 permanent teeth, and these are classified as incisors, canines, premolars and molars. Due to these differences, the tooth abnormalities demonstrated in Col7a1−/− mice are not perfectly analogous to the tooth abnormality in human RDEB patients, because RDEB affects all types of teeth and both the primary and secondary dentition in humans. However, the physiological processes of enamel formation are similar in all types of teeth of both human and mouse.
      • Miletich I.
      • Sharpe P.T.
      Normal and abnormal dental development.
      • Fleischmannova J.
      • Matalova E.
      • Tucker A.S.
      • Sharpe P.T.
      Mouse models of tooth abnormalities.
      Thus, we believe that the present Col7a1−/− mice are a practical and useful model for studying RDEB dental abnormalities.
      Some of the Col7a1−/− mice showed growth retardation as early as 3 days of age and died within the first week of life from complications of the disease. A sequence of events associated with the development of the teeth begins with neural crest cell migration at embryonic days 8.5 to 10 (E8.5 to E10); formation of the dental lamina is at E12.
      • Slavkin H.C.
      Regulatory issues during early craniofacial development: a summary.
      Ameloblast differentiation and crown formation are observed in newborn postnatal mice. Thus, we believe that the present Col7a1−/− mouse is an adequate model for studying the pathomechanisms of enamel defects caused by COL7 deficiency.
      Micro-CT observation demonstrated that enamel calcification patterns were similar between the Col7a1−/− and Col7a1+/+ mice. In addition, similar deposition patterns of chemical components were revealed in the enamel of the Col7a1−/− and Col7a1+/+ mice by SEM–energy-dispersive X-ray spectrometry. With Berlin blue staining, iron accumulation patterns were also identical in the ameloblasts of Col7a1−/− and Col7a1+/+ incisors. Deposition of iron and other ions is known to occur with maturation of the enamel matrix and mineralization at maturation stage, and the present results suggest that disturbed differentiation of ameloblasts in Col7a1−/− mice does not affect the mineralization in tooth enamel formation.
      Subsequently, we studied the developmental processes of the teeth in the Col7a1−/− mice. The teeth develop through the presecretory, secretory, and maturation stages.
      • Smith C.E.
      Cellular and chemical events during enamel maturation.
      At the presecretory stage, hypoplasia of anchoring fibrils was the only apparent abnormality in the Col7a1−/− mice teeth. The Tomes' processes are known to be involved in the secretion of enamel matrix,
      • Smith C.E.
      Cellular and chemical events during enamel maturation.
      and, in ameloblasts in the secretory stage, disturbed Tomes' process formation was observed in the Col7a1−/− mice, although enamel matrix was seen around the malformed Tomes' processes.
      Ameloblasts at the maturation stage showed no apparent abnormalities, although the crystal structure of the enamel matrix was disturbed in the Col7a1−/− mice. SEM revealed that enamel rods were malformed and irregular in the enamel of the Col7a1−/− mice. These morphological abnormalities were not observed in the rescued COL7-humanized mice, thus suggesting that the abnormalities were direct effects of the COL7 deficiency. These results clearly indicate that tooth malformation in Col7a1−/− mice, and probably in COL7A1-deficient RDEB patients, is caused by aberrant differentiation of ameloblasts. These abnormal ameloblasts lacked normal Tomes' processes and secreted reduced amounts of enamel matrix, resulting in irregular enamel rod inclination (Figure 7).
      The enamel rod inclination of teeth from the RDEB patient was irregular in the enamel layer, compared with that of a normal human control, although she does not show a complete absence of COL7. These abnormalities are most likely a consequence of a lack of COL7 causing aberrant ameloblast differentiation, similar to those in Col7a1−/− mice. Ameloblasts in Col7a1−/− mice express reduced amounts of amelogenin, ameloblastin, and other enamel proteins. These enamel proteins are secreted from Tomes' processes at the secretory stage and form the enamel matrix. Thus, the results of the present in vivo enamel protein expression study further support the idea that ameloblast differentiation, especially the formation of enamel protein-secreting Tomes' process, is disturbed in Col7a1−/− mice. Amelogenin is the most abundant of the proteins secreted by the ameloblasts, accounting for approximately 90% of total enamel protein. Mice null for amelogenin have been reported to produce hypoplastic enamel matrix without well-defined enamel rods.
      • Bartlett J.D.
      • Skobe Z.
      • Lee D.H.
      • Wright J.T.
      • Li Y.
      • Kulkarni A.B.
      • Gibson C.W.
      A developmental comparison of matrix metalloproteinase-20 and amelogenin null mouse enamel.
      Ameloblasts cultured without interaction with mesenchymal tissue cannot differentiate sufficiently to form columnar epithelium.
      • Morotomi T.
      • Kawano S.
      • Toyono T.
      • Kitamura C.
      • Terashita M.
      • Uchida T.
      • Toyoshima K.
      • Harada H.
      In vitro differentiation of dental epithelial progenitor cells through epithelial-mesenchymal interactions.
      In the present study, such insufficiently differentiated ameloblasts expressed mRNA for amelogenin, ameloblastin, and tuftelin, but not other enamel proteins, including enamelin and enamelysin. Cultured ameloblasts from Col7a1−/− mice expressed ameloblastin mRNA almost as much as ameloblasts from Col7a1+/+ mice do, although Col7a1−/− cells expressed reduced amounts of amelogenin and tuftelin mRNA. We confirmed by RT-PCR analysis that Col7a1 mRNA is expressed from cultured ameloblasts (Figure 2E), although we were unable to detect COL7 protein expression from cultured ameloblasts by Western blotting. We consider that the expressed amount of COL7 protein was so tiny that we could not detect a COL7 band by Western blotting in the present study. COL7 expression might be associated with ameloblast differentiation, and COL7 deficiency in Col7a1−/− cells might lead to the decreased expression of the major ameloblast differentiation-associated proteins, although there is no direct evidence of regulatory effects by COL7 in the differentiation of cultured ameloblasts.
      In the Col7a1−/− mice, ameloblast differentiation was retarded, resulting in malformation of Tomes' processes. The present results in Col7a1−/− mice clearly demonstrate that COL7, a component of the anchoring fibrils involved in basement membrane adhesion, also regulates differentiation of odontogenic epithelial cells including ameloblasts and plays essential role in enamelization.
      COL17 is known to be important components of hemidesmosomes. COL17 deficiency results in junctional EB, a hereditary blistering skin disease with tissue separation occurring within the lamina lucida of the epidermal basement membrane zone. Remarkable abnormalities, including disturbance of ameloblast differentiation and reduced enamel deposition, have also been reported in the incisors of COL17-disrupted mice.
      • Asaka T.
      • Akiyama M.
      • Domon T.
      • Nishie W.
      • Natsuga K.
      • Fujita Y.
      • Abe R.
      • Kitagawa Y.
      • Shimizu H.
      Type XVII collagen is a key player in tooth enamel formation.
      These facts further support the idea that interactions between ameloblasts and mesenchymal tissue via the basement membrane are crucial for ameloblast differentiation and function. Ultrastructural changes of Tomes' processes have been observed in COL17-disrupted Col17a1−/− mice,
      • Asaka T.
      • Akiyama M.
      • Domon T.
      • Nishie W.
      • Natsuga K.
      • Fujita Y.
      • Abe R.
      • Kitagawa Y.
      • Shimizu H.
      Type XVII collagen is a key player in tooth enamel formation.
      similar to those observed in Col7a1−/− mice in the present study. During the maturation stage, teeth calcification was delayed in the Col17a1−/− mice, and reduced iron deposition was revealed in the enamel of Col17a1−/− incisors,
      • Asaka T.
      • Akiyama M.
      • Domon T.
      • Nishie W.
      • Natsuga K.
      • Fujita Y.
      • Abe R.
      • Kitagawa Y.
      • Shimizu H.
      Type XVII collagen is a key player in tooth enamel formation.
      but not of Col7a1−/− mice in our study. These findings suggest that a lack of COL7 and a lack of COL17 have similar detrimental effects on ameloblast differentiation and enamel formation, although COL17 deficiency appears to have more severe disruptive effects on enamel epithelium than COL7 deficiency has.
      Concerning laminin332, another causative molecule of junctional EB, patients with laminin332 deficiency also show abnormal tooth formation (amelogenesis imperfecta).
      • Fine J.D.
      • Eady R.A.
      • Bauer E.A.
      • Bauer J.W.
      • Bruckner-Tuderman L.
      • Heagerty A.
      • Hintner H.
      • Hovnanian A.
      • Jonkman M.F.
      • Leigh I.
      • McGrath J.A.
      • Mellerio J.E.
      • Murrell D.F.
      • Shimizu H.
      • Uitto J.
      • Vahlquist A.
      • Woodley D.
      • Zambruno G.
      The classification of inherited epidermolysis bullosa (EB): report of the Third International Consensus Meeting on Diagnosis and Classification of EB.
      Laminin332-deficient mice exhibit disturbed ameloblast differentiation and reduced enamel deposition.
      • Ryan M.C.
      • Lee K.
      • Miyashita Y.
      • Carter W.G.
      Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells.
      During the maturation stage of teeth, tissue organization was completely disrupted in the enamel epithelium of the laminin332-deficient mice. Thus, laminin332 is thought to play an important role in regulating ameloblast differentiation during tooth development.
      • Ryan M.C.
      • Lee K.
      • Miyashita Y.
      • Carter W.G.
      Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells.
      In addition, it is known that COL7 interacts with laminin332 at the EMJ.
      • Rousselle P.
      • Keene D.R.
      • Ruggiero F.
      • Champliaud M.-F.
      • van der Rest M.
      • Burgeson R.E.
      Laminin 5 binds the NC-1 domain of type VII collagen.
      Specifically, the NC1 domain of COL7 binds with high affinity to the β3 chain of laminin332.
      • Chen M.
      • Marinkovich M.P.
      • Jones J.C.
      • O'Toole E.A.
      • Li Y.Y.
      • Woodley D.T.
      NC1 domain of type VII collagen binds to the beta3 chain of laminin 5 via a unique subdomain within the fibronectin-like repeats.
      In this context, we can speculate that COL7 deficiency results in loss of intact binding between COL7 and laminin332, which might lead to disturbed regulation of ameloblast differentiation by laminin332. This disturbed regulation may cause abnormal ameloblast differentiation and enamel defects in the COL7-deficient RDEB model mice and RDEB patients.
      It is known that heterozygous carriers of the LAMB3 defect and some with COL17A1 defects display dental enamel defects, whereas those with COL7A1 defects do not show these.
      • Murrell D.F.
      • Pasmooij A.M.
      • Pas H.H.
      • Marr P.
      • Klingberg S.
      • Pfendner E.
      • Uitto J.
      • Sadowski S.
      • Collins F.
      • Widmer R.
      • Jonkman M.F.
      Retrospective diagnosis of fatal BP180-deficient non-Herlitz junctional epidermolysis bullosa suggested by immunofluorescence (IF) antigen-mapping of parental carriers bearing enamel defects.
      • McGrath J.A.
      • Gatalica B.
      • Li K.
      • Dunnill M.G.
      • McMillan J.R.
      • Christiano A.M.
      • Eady R.A.
      • Uitto J.
      Compound heterozygosity for a dominant glycine substitution and a recessive internal duplication mutation in the type XVII collagen gene results in junctional epidermolysis bullosa and abnormal dentition.
      Similarly, in the present study, heterozygous mice carrying Col7a1 defects (Col7a1+/− mice) showed no apparent tooth abnormalities. These facts suggest that haploinsufficiency of COL7 does not cause dental enamel defects. Therefore, a certain deficiency level of COL7 (more than 50% reduction of COL7) may be required to cause enamel abnormality.
      Our results show that disruption of the Col7a1 gene leads to insufficient interaction between enamel epithelium and the underlying mesenchyme via the EMJ, resulting in defective ameloblast differentiation. Consequently, the Col7a1−/− mice exhibit ameloblasts with malformed Tomes' processes and diminished secretion of enamel matrix at the secretary stage. We consider that these mechanisms contribute to the immature and irregular enamel formation seen in Col7a1−/− mice. Given these findings from the model mice in the present study, we speculate that the enamel structure in RDEB patients may be impaired and that this is why there is the greater risk of caries. In conclusion, epithelial–mesenchymal interactions via the EMJ are important for tooth morphogenesis, and anchoring fibrils consisting of COL7 are thought to be involved, via interaction between COL7 and laminin332, in the regulation of proliferation and differentiation of tooth-forming cells including ameloblasts.

      Acknowledgments

      We thank Mari Miura for her helpful discussion and oral treatment of the patient and Yoshiyuki Honma, Natsumi Ushijima, Yuko Hayakawa, and Kaori Sakai for their technical assistance.

      References

        • Maas R.
        • Bei M.
        The genetic control of early tooth development.
        Crit Rev Oral Biol Med. 1997; 8: 4-39
        • Liu F.
        • Chu E.Y.
        • Watt B.
        • Zhang Y.
        • Gallant N.M.
        • Andl T.
        • Yang S.H.
        • Lu M.M.
        • Piccolo S.
        • Schmidt-Ullrich R.
        • Taketo M.M.
        • Morrisey E.E.
        • Atit R.
        • Dlugosz A.A.
        • Millar S.E.
        Wnt/beta-catenin signaling directs multiple stages of tooth morphogenesis.
        Dev Biol. 2008; 313: 210-224
        • Natsuga K.
        • Shinkuma S.
        • Nishie W.
        • Shimizu H.
        Animal models of epidermolysis bullosa.
        Dermatol Clin. 2010; 28: 137-142
        • Asaka T.
        • Akiyama M.
        • Domon T.
        • Nishie W.
        • Natsuga K.
        • Fujita Y.
        • Abe R.
        • Kitagawa Y.
        • Shimizu H.
        Type XVII collagen is a key player in tooth enamel formation.
        Am J Pathol. 2009; 174: 91-100
        • Ryan M.C.
        • Lee K.
        • Miyashita Y.
        • Carter W.G.
        Targeted disruption of the LAMA3 gene in mice reveals abnormalities in survival and late stage differentiation of epithelial cells.
        J Cell Biol. 1999; 145: 1309-1323
        • Fine J.D.
        • Eady R.A.
        • Bauer E.A.
        • Bauer J.W.
        • Bruckner-Tuderman L.
        • Heagerty A.
        • Hintner H.
        • Hovnanian A.
        • Jonkman M.F.
        • Leigh I.
        • McGrath J.A.
        • Mellerio J.E.
        • Murrell D.F.
        • Shimizu H.
        • Uitto J.
        • Vahlquist A.
        • Woodley D.
        • Zambruno G.
        The classification of inherited epidermolysis bullosa (EB): report of the Third International Consensus Meeting on Diagnosis and Classification of EB.
        J Am Acad Dermatol. 2008; 58: 931-950
        • Tidman M.J.
        • Eady R.A.
        Evaluation of anchoring fibrils and other components of the dermal-epidermal junction in dystrophic epidermolysis bullosa by a quantitative ultrastructural technique.
        J Invest Dermatol. 1985; 84: 374-377
        • Sakai L.
        • Keene D.
        • Morris N.
        • Burgeson R.
        Type VII collagen is a major structural component of anchoring fibrils.
        J Cell Biol. 1986; 103: 1577-1586
        • Rousselle P.
        • Keene D.R.
        • Ruggiero F.
        • Champliaud M.-F.
        • van der Rest M.
        • Burgeson R.E.
        Laminin 5 binds the NC-1 domain of type VII collagen.
        J Cell Biol. 1997; 138: 719-728
        • Chen M.
        • Marinkovich M.P.
        • Jones J.C.
        • O'Toole E.A.
        • Li Y.Y.
        • Woodley D.T.
        NC1 domain of type VII collagen binds to the beta3 chain of laminin 5 via a unique subdomain within the fibronectin-like repeats.
        J Invest Dermatol. 1999; 112: 177-183
        • Christiano A.
        • Greenspan D.
        • Hoffman G.
        • Hoffman G.
        • Zhang X.
        • Tamai Y.
        • Lin A.
        • Dietz H.
        • Hovnanian A.
        • Uitto J.
        A missense mutation in type VII collagen in two affected siblings with recessive dystrophic epidermolysis bullosa.
        Nat Genet. 1993; 4: 62-66
        • Wright J.T.
        Oral manifestations in the epidermolysis bullosa spectrum.
        Dermatol Clin. 2010; 28: 159-164
        • Harris J.C.
        • Bryan R.A.
        • Lucas V.S.
        • Roberts G.J.
        Dental disease and caries related microflora in children with dystrophic epidermolysis bullosa.
        Pediatr Dent. 2001; 23: 438-443
        • Kirkham J.
        • Robinson C.
        • Strafford S.M.
        • Bonass W.A.
        • Brookes S.J.
        • Wright J.T.
        The chemical composition of tooth enamel in recessive dystrophic epidermolysis bullosa: significance with respect to dental caries.
        J Dent Res. 1996; 75: 1672-1678
        • Wright J.T.
        • Fine J.D.
        • Johnson L.B.
        Developmental defects of enamel in humans with hereditary epidermolysis bullosa.
        Archs Oral Biol. 1993; 38: 945-955
        • Shah H.
        • McDonald F.
        • Lucas V.
        • Ashley P.
        • Roberts G.
        A cephalometric analysis of patients with recessive dystrophic epidermolysis bullosa.
        Angle Orthod. 2002; 72: 55-60
        • Murrell D.F.
        • Pasmooij A.M.
        • Pas H.H.
        • Marr P.
        • Klingberg S.
        • Pfendner E.
        • Uitto J.
        • Sadowski S.
        • Collins F.
        • Widmer R.
        • Jonkman M.F.
        Retrospective diagnosis of fatal BP180-deficient non-Herlitz junctional epidermolysis bullosa suggested by immunofluorescence (IF) antigen-mapping of parental carriers bearing enamel defects.
        J Invest Dermatol. 2007; 127: 1772-1775
        • Heinonen S.
        • Männikkö M.
        • Klement J.F.
        • Whitaker-Menezes D.
        • Murphy G.F.
        • Uitto J.
        Targeted inactivation of the type VII collagen gene (Col7a1) in mice results in severe blistering phenotype: a model for recessive dystrophic epidermolysis bullosa.
        J Cell Sci. 1999; 112: 3641-3648
        • Thesleff I.
        • Partanen A.M.
        • Vainio S.
        Epithelial-mesenchymal interactions in tooth morphogenesis: the roles of extracellular matrix, growth factors, and cell surface receptors.
        J Craniofac Genet Dev Biol. 1991; 11: 229-237
        • Sawamura D.
        • Goto M.
        • Yasukawa K.
        • Sato-Matsumura K.
        • Nakamura H.
        • Ito K.
        • Nakamura H.
        • Tomita Y.
        • Shimizu H.
        Genetic studies of 20 Japanese families of dystrophic epidermolysis bullosa.
        J Hum Genet. 2005; 50: 543-546
        • Ito K.
        • Sawamura D.
        • Goto M.
        • Nakamura H.
        • Nishie W.
        • Sakai K.
        • Natsuga K.
        • Shinkuma S.
        • Shibaki A.
        • Uitto J.
        • Denton C.P.
        • Nakajima O.
        • Akiyama M.
        • Shimizu H.
        Keratinocyte-/fibroblast-targeted rescue of Col7a1-disrupted mice and generation of an exact dystrophic epidermolysis bullosa model using a human COL7A1 mutation.
        Am J Pathol. 2009; 175: 2508-2517
        • Shimizu H.
        • Ishiko A.
        • Masunaga T.
        • Kurihara Y.
        • Sato M.
        • Bruckner-Tuderman L.
        • Nishikawa T.
        Most anchoring fibrils in human skin originate and terminate in the lamina densa.
        Lab Invest. 1997; 76: 753-763
        • Fukumoto S.
        • Kiba T.
        • Hall B.
        • Iehara N.
        • Nakamura T.
        • Longenecker G.
        • Krebsbach P.H.
        • Nanci A.
        • Kulkarni A.B.
        • Yamada Y.
        Ameloblastin is a cell adhesion molecule required for maintaining the differentiation state of ameloblasts.
        J Cell Biol. 2004; 167: 973-983
        • Fukumoto S.
        • Yamada A.
        • Nonaka K.
        • Yamada Y.
        Essential roles of ameloblastin in maintaining ameloblast differentiation and enamel formation.
        Cells Tissues Organs. 2005; 181: 189-195
        • Masuya H.
        • Shimizu K.
        • Sezutsu H.
        • Sakuraba Y.
        • Nagano J.
        • Shimizu A.
        • Fujimoto N.
        • Kawai A.
        • Miura I.
        • Kaneda H.
        • Kobayashi K.
        • Ishijima J.
        • Maeda T.
        • Gondo Y.
        • Noda T.
        • Wakana S.
        • Shiroishi T.
        Enamelin (Enam) is essential for amelogenesis: eNU-induced mouse mutants as models for different clinical subtypes of human amelogenesis imperfecta (AI).
        Hum Mol Genet. 2005; 14: 575-583
        • Uitto J.
        • Pulkkinen L.
        • Christiano A.M.
        Molecular basis of the dystrophic and junctional forms of epidermolysis bullosa: mutations in the type VII collagen and kalinin (laminin 5) genes.
        J Invest Dermatol. 1994; 103: 39S-46S
        • Christiano A.M.
        • Anhalt G.
        • Gibbons S.
        • Bauer E.A.
        • Uitto J.
        Premature termination codons in the type VII collagen gene (COL7A1) underlie severe, mutilating recessive dystrophic epidermolysis bullosa.
        Genomics. 1994; 21: 160-168
        • Miletich I.
        • Sharpe P.T.
        Normal and abnormal dental development.
        Hum Mol Genet. 2003; 12: R69-R73
        • Fleischmannova J.
        • Matalova E.
        • Tucker A.S.
        • Sharpe P.T.
        Mouse models of tooth abnormalities.
        Eur J Oral Sci. 2008; 116: 1-10
        • Slavkin H.C.
        Regulatory issues during early craniofacial development: a summary.
        Cleft Palate J. 1990; 27: 101-109
        • Smith C.E.
        Cellular and chemical events during enamel maturation.
        Crit Rev Oral Biol Med. 1998; 9: 128-161
        • Bartlett J.D.
        • Skobe Z.
        • Lee D.H.
        • Wright J.T.
        • Li Y.
        • Kulkarni A.B.
        • Gibson C.W.
        A developmental comparison of matrix metalloproteinase-20 and amelogenin null mouse enamel.
        Eur J Oral Sci. 2006; 114: 18-23
        • Morotomi T.
        • Kawano S.
        • Toyono T.
        • Kitamura C.
        • Terashita M.
        • Uchida T.
        • Toyoshima K.
        • Harada H.
        In vitro differentiation of dental epithelial progenitor cells through epithelial-mesenchymal interactions.
        Arch Oral Biol. 2005; 50: 695-705
        • McGrath J.A.
        • Gatalica B.
        • Li K.
        • Dunnill M.G.
        • McMillan J.R.
        • Christiano A.M.
        • Eady R.A.
        • Uitto J.
        Compound heterozygosity for a dominant glycine substitution and a recessive internal duplication mutation in the type XVII collagen gene results in junctional epidermolysis bullosa and abnormal dentition.
        Am J Pathol. 1996; 148: 1787-1796