Matrix Vol. 12/1992, pp. 448-455 © 1992 by Gustav Fischer Verlag, Stuttgart
Demonstration of Type III Collagen in the Dentin of Mice KENGO NAGATA 1 , YU HSIN HUANG 2 , YASUYOSHI OHSAKI 1 , TOSHIO KUKITA 1 , MINORU NAKATA2 and KOJIRO KURISU 1,3 Second Department of Anatomy] and Department of Pediatric Dentistry 2, Faculty of Dentistry, Kyushu University, Fukuoka,]apan.
Abstract It has been reported that, although type III collagen is present in human dentin where there is dentinogenesis imperfecta and in reparative dentin, it is absent in normal dentin. In a preliminary study, however, we observed evidence showing that small amounts of fibers showing positive labeling for type III collagen are present in the molars of normal mice. In the present study; in order to localize type III in normal dentin, immunofluorescent and immunoelectron microscopic examinations of the molars of normal mice were carried out using affinity-purified antibodies to mouse type III and type I collagen. The fibers positive for type III collagen were much more frequently observed in the root than in the crown. These fibers ran in peritubular dentin or near that in parallel to them. The incidence of the existence of dentinal tubules associated with type III collagen-positive fibrils either in or near peritubular dentin was low. These fibrils positive for type III collagen showed a clear cross-banding. In dentinal tubules, unusual collagen aggregations, segment long-spacing-like and fibrous long-spacing-like structures which were intensively stained for type I collagen but weakly so for type III collagen were seldom observed. Type III collagen-positive fibers often extended towards the pulp beyond the odontoblast layer, suggesting that these fibers were produced, at least partly, by the pulp cells. Key words: collagen antibodies, dentin, type I and III collagen.
Introduction Over fourteen genetically distinct types of collagen have been identified in recent years. Of these, the molecules of interstitial collagens (I, II, III, V, XI) which form fibers in the extracellular matrix of connective tissue, have been extensively characterized (Linsenmayer, 1981; Kiihn, 1987; Holzmann et aI., 1989). Although type III collagen has been observed in the dentin of patients with osteogenesis imperfecta (Sauk et aI., 1980; Gagr, 1985; Lukinmaa et aI., 1985; Lukinmaa, 1988), in hereditary opalescent dentin (Sauk et aI., 1980) and in human reparative dentin (Magloire et aI., 1988), most investigators have 3 Present address: Kojiro Kurisu, First Department of Oral Anatomy, Faculty of Dentistry, Osaka University, 1-8 Yamadaoka, Suita, Osaka 565, Japan.
agreed that type III collagen is absent from normal dentin, as well as from bone matrix (Cournil et aI., 1979; Tung et aI., 1985; Takita et aI., 1987; Andujar et aI., 1988; Andujar et aI., 1991). Recently, Becker et aI. (1986), using normal human teeth, immunohistochemically demonstrated that, although procollagen type III was virtually absent from intertubular dentin, dentinal tubules contained procollagen type III positive fibers, which appeared to be thicker than the tubular type I procollagen fibers, and that these fibers sometimes extended into the pulpal space through the strongly reactive predentin matrix which was positive for type III collagen. This report suggested that type III collagen is present in normal dentin. In addition, our preliminary study using the immunofluorescent method for the localization of type III collagen in normal mouse dentin demonstrated the presence of a few fibers positive for type III collagen (unpub-
Type III Collagen in Dentin Iished data). However, it is not easy to determine, by light microscopy, whether fibers positive for type III collagen exist in intertubular dentin or in dentinal tubules. Therefore, immunoelectron microscopic examinations should be carried out for the localization of type III collagen in normal dentin. In the present study, we obtained the immunofluorescent and immunoelectron microscopic localization of type III collagen in the normal dentin of mouse molars. Distribution of type I collagen in relation to type III collagen was also examined. Type III collagen was often observed in the normal dentin matrix of the root or crown region, but not in the dentinal tubules. Types III collagen-positive fibers ran parallel to the dentinal tubules either in peritubular dentin or within the vicinity of the dentinal tubules. They sometimes extended through the predentin into the pulpal space. These observations support those of Becker et al. (1986) except for the fact that type III collagen fibers were not found to be present in the dentinal tubules, but in the dentin matrix.
Materials and Methods Tissue preparation
The materials used were the maxillary molars of CFl mice, 13 days to 4 months old. Preparation of materials was as described by Huang et al. (1991). Mice were perfused from the left ventricle with a fixative (0.1 M cacodylate buffer containing 0.1 % glutaraldehyde and 4% paraformaldehyde). Maxillae removed from the mice were immersed in the fixative for 12-24h at 4°C. They were then decalcified with 10% EDTA for 2-3 weeks at 4°C. Samples for light microscopic examinations were embedded in paraffin. For post-embedding immunoelectron microscopy, the samples embedded in gelatin were cut with a Microslicer (Dohsaka, Kyoto, japan) into thin slices with a thickness of 100 !-tm. The slices were them embedded in Lowicryl K4M and polymerized with UV light at - 20°C for 5 h and at 4 °C for 2 h. Preparation ofantigens and antibodies
Types I and III collagen were extracted from adult CFl mouse skin with acetic acid and pepsin, respectively, and then purified by differential salt precipitation and DEAEcellulose chromatography, as described by Chung and Miller (1974). Their purity was checked by SDS-PAGE. Antitype I and anti-type III collagen antibodies (titer 7) were raised in rabbits. Anti-type III collagen antibody (titer 9) used for the double-staining method were raised in guinea pigs. These antibodies (lgG) were then purified by affinity chromatography using Sepharose-bound antigens. Both antibody preparations were adsorbed by the other antigens in order to make them monospecific for type I or type III
449
collagen. Affinity-purified non-labeled and FITC-labeled goat anti-rabbit IgG were purchased from Organon Teknika Corp. (West Chester, PAl. Affinity-purified nonlabeled and rhodamine-labeled goat anti-guinea pig IgG were purchased from Chemicon International Inc. (Temecula, CAl. The labeling of non-labeled antibodies with colloidal gold was performed as described by Slot et al. (1984; 1985). Immunofluorescent microscopy
The double-staining technique was performed as described by Engel et al. (1980). Briefly, paraffin sections were incubated at room temperature consecutively with (1) 2% bovine serum albumin (BSA) for 1 h, (2) 100 !-tg/ml rabbit anti-type I plus 100 !-tg/ml guinea pig anti-type III collagen antibodies for 1 hand (3) FITC-labeled goat antirabbit IgG plus rhodamine-labeled goat anti-guinea pig IgG diluted 1:500 in PBA (0.1 M PBS containing 1% BSA and 0.1 M lysine-HCl) for 30 min. Between each step, samples were thoroughly rinsed with PBS. In the control sections, antibody solutions adsorbed with corresponding antigens, PBS, normal rabbit or normal guinea pig IgG were used in place of primary antibodies to verified the specificity of primary as well as secondary antibodies. They were then sealed with a PBS-buffered glycerol to p-phenylenediamine had been added in order to reduce the fading of the fluorescence, as recommended by johnson and Arujo (1981). Immunoelectron microscopy
The post-embedding technique was carried out as follows: The ultra-thin sections mounted on nickel grids coated with 1% collodion film were incubated sequentially in a moist chamber at room temperature with PBA for 30 min, rabbit anti-type I or anti-type III collagen antibodies for 1 h, and colloidal gold (0 11 nm)-labeled antirabbit IgG for 20 min. Between each step, the sections were thoroughly rinsed with PBS. In the simultaneous doublestaining, a mixture of rabbit anti-type I and guinea pig antitype III collagen antibodies was used for the primary antibodies, and that of colloidal gold (011 nm)-labeled goat anti-rabbit IgG and colloidal gold ( 6 nm)-labeled goat anti-guinea pig IgG was used for the secondary antibodies. Control sections were treated with PBS, normal serum or antibodies adsorbed with corresponding antigens in place of primary antibodies. The sections were finally stained with uranyl acetate and examined with a JEOL 200CX electron microscope.
Results Immunofluorescent examination
In the longitudinal section of roots (Fig. 1), there were thick fibrous structures positive for type III collagen which
450
K. Nagata et al.
Figs. 1- 3 are immunofluorescent micrographs of the mandibular first molar in 4-month-old mice that were stained for either type I (a) or type III (b) collagen. Fig. 1: Section longitudinal to the dentinal tubules of the root which was double-stained for type I (a) and type III collagen (b). (a) Type I collagen is homogeneously distributed in dentin (D), the staining of which is less intense than that of predentin (Pd). Cytoplasm of odontoblasts (arrows) is also positive. (b) Thick fibrous structures showing a positive reaction for type III collagen (arrows) can be observed parallel to the dentinal tubules. Some of them (arrow heads) extend through the predentin into the pulp space (P). PL: periodontal ligaments. x 520. Bar = 25 !-tm. Fig. 2: Section transverse to the dentinal tubules of the root which was stained for either type I (a) or type III collagen (b). (b) Positive reactions can be observed as spots in dentin. PL: periodontal ligaments, D: dentin. x 520. Bar = 25 !-tm. Fig. 3: Longitudinal section to the dentinal tubules of crown dentin which was stained for either type I (a) or type III collagen (b). (b) Positive fibrous structures were rarely observed (arrows) in dentin. D: dentin. x 520. Bar = 25 !-tm. Fig. 4: Control section of root incubated with normal serum in place of primary antibodies. Positive staining was not observed. x 520. Bar = 25 !-tm.
were parallel to the dentinal tubules (Fig. 1 b). Some of them extended through the predentin into the pulpal space. Cytoplasm of odontoblasts did not show any positive reaction for type III collagen. On the other hand, type I collagen was homogeneously distributed in the dentin matrix, the inten-
sity of which was weaker than that of predentin (Fig. 1 a). Cytoplasm of odontoblasts was also positive for type I collagen. In the transverse section of roots, type III collagenpositive spots were observed in dentin (Fig. 2). Fibrous structures, positive for type III collagen, were observed in
Type III Collagen in Dentin
451
Figs. 5 -13 except for 9 and 10. Immunoelectron micrographs for type III or type I collagen in the root dentin of 4-month-old mice. Fig. 5: Section longitudinal to dentinal tubule stained for type III collagen. Fibril bundles running parallel with dentinal tubules show positive staining (arrow). x 21,000. Bar = 0.5I!m. The inset is shown at high magnification. X 57,000. Fig. 6: Section longitudinal to dentinal tubule stained for type III collagen. Close relation between the positive fibrils and the peritubular dentin reveals that the positive fibrils run either within the peritubular dentin or the vicinity of the dentinal tubules. X 21,000. Bar = 0.5I!m. Fig. 7: Section transeverse to dentinal tubule stained for type III collagen. Positive labeling can be observed in the peritubular dentin. DT: dentinal tubule. X 21,000. Bar = 0.5I!m. Fig. 8: Section transverse to dentinal tubules stained for type III collagen. Labeling is absent in the intertubular dentin, including the peritubular dentin. X 21,000. Bar = 0.5 ~m.
the same pattern as those in crown dentin (Fig. 3), but the frequency of their appearance was much less than that in root dentin. Control sections incubated with normal sera in place of primary antibodies showed negative staining (Fig. 4). Control sections incubated with PBS or the antibody solution adsorbed with corresponding antigens in place of primary antibodies also showed negative staining for both type III and type I collagen, as shown previously (Huang et al., 1991). The staining status of these control sections revealed that the antibodies used were monospecific for their corresponding antigens.
Immunoelectron examination
In the section longitudinal to the dentinal tubules, positive staining for type III collagen was observed on fiber bundles running parallel with the dentinal tubules (Fig. 5). These fibers were observed either within peritubular dentin, or within the vicinity of the dentinal tubules (Fig. 6). In the section transverse to the dentinal tubules, labeling for type III collagen was observed either in peritubular dentin or within the vicinity of some dentinal tubules (Fig. 7). However, in a large number of the dentinal tubules, no labeling was observed either at peritubular dentin or intertubular dentin (Fig. 8). In about 10% of the tubules at the
452
K. Nagata et al.
Fig. 9 and 10: Conventional electron micrograph of root dentin in rhe mandibular molar of 4-month-old mice. The dentin was embedded in Epon as described in Materials and Methods. FLS-like structure can be observed within the dentinal tubule (Fig. 9. x 57,000. Bar = 0.2 Ilm). SLS-like structures can be observed within the dentinal tubule (Fig. 10. x 21,000. Bar = 0.5 Ilm). Fig. 11: Obliquely sectioned FLS-like structures double-stained for type I and type III collagen. Note that these structures are intensively stained for type I collagen (0 11 nm) but weakly so for type III collagen (0 6 nm). FLS-like structure shows the same staining pattern as that of a SLS-like structure (arrow). x 31,000. Bar = 0.5 Ilm. Fig. 12: Section containing SLS-like structures double-stained for type I and type III collagen. Note these structures are intensively stained for type I collagen (0 11 nm) as in a SLS-like structure shown in Fig. 11. x 42,000. Bar = 0.2 Ilm. Fig. 13: Control section incubated with normal serum in place of primary antibodies. Labeling is absent in all areas. X 21,000. Bar = 0.5 Ilm.
Type III Collagen in Dentin middle portion of the root, positive labeling was observed in peritubular dentin. Two types of collagen aggregates, segment long-spacing (SLS)-like and fibrous long-spacing (FLS)-like structures were observed in some of the dentinal tubules of the root dentin at a low frequency (Figs. 9, 10, 11 and 12). FLS-like structure had about 160 nm periodicity (Figs. 9 and 11), and SLS-like structure had a symmetric banding pattern about 320 nm in length, but with a varying width (Figs. 10 and 12). These structures were intensively stained for type I collagen but weakly so for type III collagen (Figs. 11 and 12). Control sections incubated with normal sera showed no labeling, suggesting the high immunospecificity of the staining method used (Fig. 13).
Discussion There has been a consensus that type III collagen is not present in dentin, except for that of patients suffering from type IV osteogenesis imperfecta with dentinogenesis imperfecta (Sauk et al., 1980; Gage, 1985; Lukinmaa et al., 1988; Lukinmaa, 1988) and in reparative dentin (Magloire et al., 1988). Recently, however, Becker et al. (1986) reported the presence of type III collagen in the dentinal tubules of normal human molars by the use of immunohistochemical methods at the light microscopic level. In the present study, in order to obtain a critical localization of type III collagen in normal dentin, we carried out immunohistochemical examinations at light as well as electron microscopic levels, using affinity purified specific antibodies to type III and type I collagen, from which those cross-reacting to other antigens were eliminated. Our results demonstrated that type III collagen fibers certainly exist in, or close to, the peritubular dentin of normal dentin, but that they were absent within the dentinal tubules, as suggested by Becker et al. (1986). We assume that other researchers failed to identify type III collagen as being present within dentinal tubules, let alone in or close to peritubular dentin, since it is not easy to obtain a critical localization of immunohistochemically positive fibers in thick sections at the light microscopic level. Becker et al. (1986) was also able to observe type III collagen-positive fibers extending from the dentin to the pulp through the odontoblastic layer, which was consistent with our findings. In the dentinal tubules, we observed two types of unusual collagen aggregates, SLS-like and FLS-like structures, both of which showed positive labeling for type I collagen, but weakly so for type III collagen. These structures have been observed in predentin as well in the dentin of normal rat incisors (Warshawsky, 1972; Weinstock, 1977). The presence of uniform structures has been reported in predentin and in the incisors of rats treated with either vinblastine (Miake and Takuma, 1985) or sodium fluoride (Araki, 1989). SLS in the longitudinal section had centrosymmetrical striations associated with filamentous materials at both
453
ends. From this characteristic structure, it has been supposed that SLS consists of laterally packed antipolar procollagen molecule, in which the N-terminal end of one, overlays the C-terminal end of another (Chapman and Armitage, 1972; Weinstock, 1977; Miake and Takuma, 1985; Araki, 1989). FLS-like structures are supposed to be FLS IV which was thought to have been formed by antipolar dimers aggregating side by side, with sliding by one half of them (Armitage and Chapman, 1971; Chapman and Armitage, 1972). Some reporters indicated the possibility that these unusual collagen aggregates were artifacts resulting from fixation or from decalcification. Araki (1989), however, neglected the possibility of them being artifacts, because these unusual collagen structures were observed in predentin and in the dentin of the incisors of rats treated with NaF, but they were not present in those of the controls. It is thought that these collagen structures arose as a result of the inhibition of the normal processing of the N- and Cterminals of procollagen molecules. In the present study, these unusual collagen structures were observed only in dentinal tubules, but not elsewhere in the matrix of predentin or dentin. In addition, they showed positive labeling for type I collagen, but weakly so for type III collagen. These findings indicate that the type III collagen-positive fibers in dentinal tubules reported by Becker et al. (1986) are different from the unusual collagen structures described above. The origin of the type III collagen fibers in predentin and dentin was not clarified in this study. However, the observation that these fibers often extend towards the pulp beyond the odontoblast layer, suggests that these fibers are produced by pulp cells, but not by odontoblasts. Becker et al. (1986) revealed a faint staining of the cytoplasm of the odontoblasts for type III collagen. However, we failed to detect any positive staining of the cytoplasm of the odontoblasts for type III collagen, although we found positive staining for type I collagen, indicating odontoblasts do not synthesize type III, whereas they do type I collagen. Andujar et al. (1991) revealed that type III collagen mRNAs were not detected in odontoblasts during mouse molar tooth development by in situ hybridization. It is possible, however, that the amount of type III collagen synthesized by odontoblasts is too small to be detected by the immunohistochemical method employed in this study. The role of type III collagen in normal dentin is not clear from this study. The cross-bandings of collagen fibrils in dentin matrix showing positive labeling for type I collagen were more obscure compared with those of fibrils positive for type III collagen, indicating that the latter fibrils are not calcified at all, or else are only slightly calcified. Although type III collagen was believed to be absent from mineralized tissue, Keene et al. (1991) revealed that type III collagen were present in discrete fiber bundles throughout the human bone cortex. Type III collagen is predominant in distensible tissue such as the walls of blood vessels, the digestive tract and the lung, this type of collagen may
454
K. Nagata et al.
participate in the elasticity of tissue (Miller, 1976). The presence of uncalcified fibrils in dentin may have some effect on the physical property of dentin. However, type III collagen fibrils are often observed in the dentin of an unphysiological state, for example, in the dentin of odontogenesis imperfecta (Sauk et aI., 1980; Gage, 1985; Lukinmaa et aI., 1985; Lukinmaa, 1988) or in reparative dentin (Magloire et aI., 1988), suggesting the possibility that as a result of the unphysiological condition of those cells relating to dentinogenesis, especially odontoblasts, the inclusion of type III collagen may be induced in peritubular dentin or dentin matrix. There is some controversy concerning the presence of cross-banding in type III collagen fibrils. Several studies have demonstrated that type III collagen can be observed in fibrils without cross-banding in human skin (Fleischmajer et ai, 1981) and in dog gingiva (Cho et aI., 1987). Amenta et al. (1986) showed that type I collagen existed on thick and cross-banded fibrils, while type III collagen existed exclusively on thin, beaded noncross-banded fibrils. Keene et al. (1987), however, using a monoclonal antibody specific for human type III collagen, revealed that type III collagen was present on cross-banded fibrils in skin, tendons, and amnion, regardless of the fibril diameter. Huang et al. (1991) also demonstrated the presence of type III collagen in crossbanded fibrils in the mouse periodontal ligament. In the present study, type III collagen was found in the crossbanded fibrils, supporting the presence of type III collagen fibrils with cross-banding.
Acknowledgements A part of this study was supported by Grant-in-Aid for Scientific Research from the Ministry of Education of Japan (Project 61571860).
References Amenta, P. S., Gay, S., Vaheri, A. and Martinez-Hernandez, A.: The extracellular matrix is an integrated unit: Ultrastructural localization of collagen type I, III, IV, VI, fibronectin, and laminin in human term placenta. Collagen ReI. Res. 6: 125 -152, 1986. Andujar, M. B., Hartmann, D.J., Emonard, H. and Magloire, H.: Distribution and synthesis of type I and type III collagens in developing mouse molar tooth root. Histochemistry 88: 131-140, 1988. Andujar, M. B., Couble, P., Couble, M. L. and Magloire, H.: Differential expression of type I and type III collagen genes during tooth development. Development 111: 691-698, 1991. Araki, S.: Ultrastructral changes in rat-incisor odontoblasts and dentin caused by administration of sodium fluoride. Shika Gakuho 89: 49-91,1989. Armitage, P. M. and Chapman,]. A.: New fibrous long spacing form of collagen. Nature 229: 151-152,1971. Becker,J., Schuppan, D., Benzeian, H., Bals, T., Hahn, E. G., Cantaluppi, C. and Reichart, P.: Immunohistochemical distribution
of collagens type IV, V and VI and of pro-collagens types I and III in human alveolar bone and dentine. J. Histochem. Cytochem. 34: 1417-1429,1986. Chapman,].A. and Armitage, P.M.: An analysis of fibrous longspacing forms of collagen. Conn. Tissue Res. 1: 31-37, 1972. Cho, M. L., Lee, Y. L. and Garant, P. R.: Immunocytochemical localization of extracellular matrix components in beagle periodontium: I. Collagen type I and type III in healthy gingival connective tissue. J. Period. Res. 22: 313-319, 1987. Chung,E. and Miller, E.].: Collagen polymorphism: Characterization of molecules with the chain composition [a1(III)h in human tissues. Science 183: 1200-1201,1974. Cournil, I., Lebrond, C. P., Pomponio,]., Hand, A. R., Sederlof, L. and Martin, G. R.: Immunohistochemical localization of procollagens. I. Light microscopic distribution of procollagen I, III and IV antigenicity in the rat incisor tooth by the indirect peroxidase-anti-peroxidase method. J. Histochem. Cytochem. 27: 1059-1069,1979. Engel, D., Schroeder, H. E., Gay, R. and Clagett,].: Fine structure of cultured human gigival fibroblasts and demonstration of simultaneous synthesis of types I and III collagen. Archs. Oral BioI. 25: 283-296, 1980. Fleischmajer, R., Tuderman, L., Raisher, L., Wiestner, M., Perlish,J. S. and Graves, P. N.: Ultrastructural identification of extension aminopropeptides of type I and type III collagens in human skin. Proc. Natl. Acad. Sci. USA 78: 7360-7364, 1981. Gage, I. P.: Dentinogenesis imperfecta. A new perspective. Austr. Dent. j. 30: 285 - 290,1985. Holzmann, B., McIntyre, B. W. and Weissman, I. L.: Identification of a murine Peyer's patch-specific lymphocyte homing receptor as an integrin molecule with an a chain homologous to VLA-4a. Cell 56: 37-46, 1989. Huang, Y. H., Ohsaki, Y. and Kurisu, K.: Distribution of type I and type III collagen in the developing periodontal ligament of mice. Matrix 11: 100-108,1991. Johnson,G.D. and Araujo,G.M.C.N.: A simple method of reducing the fading of immunofluorescence during microscopy. J. Immunol. Methods 42: 349-350, 1981. Keene, D. R., Sakai, L. Y., Bachinger, H. P. and Burgeson, R. E.: Type III collagen can be present on banded collagen fibrils regardless of fibril diameter. J. Cell Bioi. 105: 2393-2402, 1987. Keene, D. R., Sasaki, L. Y. and Burgeson, R. E.: Human bone contains type III collagen, type IV collagen, and Fibrillin: Type III collagen is present on specific fibers that may mediate attachment of tendons, ligaments and periosteum to calcified bone cortex. J. Histochem. Cytochem. 39: 56-69, 1991. Kuhn, K.: The classical collagens. In: Types I, II and III. In: Structure and Function of Collagen Types. ed. by Mayne, R. and Burgeson, R.E., Academic Press, New York, 1987, pp. 1-42. Linsenmayer, T. F.: Collagen. In: Cell Biology of Extracellular Matrix, ed. by Hay,E.D., Plenum Press, New York, 1981, pp.5-37. Lukinmaa, P. L., Ranta, H., Ranta, K., Peltonen, L. and Hietanen,].: Demineralization of dentin with EDTA in organic solvent: Immunofluorescence of collagen in ostegenesis imperfecta and normal teeth. Collagen Rei. Res. 5: 505 - 512,1985. Lukinmaa, P. L.: Immunofluorescent localization of type III collagen and the N-terminal propeptide of type III procollagen in dentin matrix in osteogenesis imperfecta. J. Craniofac. Genet. Develop. Bioi. 8: 235-243, 1988. Magloire, H., Joffre, A. and Hartman, D.].: Localization and synthesis of type III collagen and fibronectin in human reparative dentine. Histochemistry 88: 141-149,1988.
Type III Collagen in Dentin Miake, Y. and Takuma, S.: Unusual collagen aggregates induced in rat incisor dentin after vinblastine administration. Calcif. Tissue Int. 37: 501-510, 1985. Miller, E.].: Biochemical characteristics and biological significance of the genetically-distinct collagens. Mol. Cell Biochem. 13: 165-192,1976. Sauk,].]., Gay, R., Miller, E.J. and Gay, S.: Immunohistochemical localization of type III collagen in the dentin of patients with osteogenesis imperfecta and hereditary opalescent dentin. J. Oral Path. 9: 210- 220,1980. Slot,J. W. and Geuze, H.J.: Gold markers for single and double immunolabelling of ultrathin cryosection. In: Immunolabelling for Electron Microscopy, ed. by Polak,]. M. and Varndell, I. M., Elsevier Science Publishing, Amsterdam, 1984, pp. 129-142. Slot,J. W. and Geuze, H.].: A new method of preparing gold pobes for multiple labeling cytochemistry. Eur. J. Cell Bioi. 38: 87-93,1985. Takita, K., Ohsaki, Y., Nakata, M. and Kurisu, K.: Immuno-
455
fluorescence localization of type I and type III collagen and fibronectin in mouse dental tissues in late development and during molar eruption. Archs. Oral BioI. 32: 273-279, 1987. Tung, P. S., Domenicucci, C., Wasi, S. and Sodek,].: Specific immunohistochemical localization of osteonectin and collagen types I and III in fetal and adult porcine dental tissues. J. Histochem. Cytochem. 33: 531-540, 1985. Warshawsky, H.: The presence of atypical collagen fibrils in EDTA decalcified predentin and dentin of rat incisors. Archs. Oral Bioi. 17: 1745-1754,1972. Weinstock, M.: Centrosymmetrical cross-banded structures in the matrix of rat incisor predentin and dentin. J. Ultrastruct. Res. 61: 218-229, 1977. Dr. Kengo Nagata, Second Department of Anatomy, Faculty of Dentistry, Kyushu University, Maidashi, Higashi-ku, Fukuoka 812,Japan.
Demonstration of Type III Collagen in the Dentin of Mice KENGO NAGATA 1 , YU HSIN HUANG 2 , YASUYOSHI OHSAKI 1 , TOSHIO KUKITA 1 , MINORU NAKATA2 and KOJIRO KURISU 1,3 Second Department of Anatomy] and Department of Pediatric Dentistry 2, Faculty of Dentistry, Kyushu University, Fukuoka,]apan.
Abstract It has been reported that, although type III collagen is present in human dentin where there is dentinogenesis imperfecta and in reparative dentin, it is absent in normal dentin. In a preliminary study, however, we observed evidence showing that small amounts of fibers showing positive labeling for type III collagen are present in the molars of normal mice. In the present study; in order to localize type III in normal dentin, immunofluorescent and immunoelectron microscopic examinations of the molars of normal mice were carried out using affinity-purified antibodies to mouse type III and type I collagen. The fibers positive for type III collagen were much more frequently observed in the root than in the crown. These fibers ran in peritubular dentin or near that in parallel to them. The incidence of the existence of dentinal tubules associated with type III collagen-positive fibrils either in or near peritubular dentin was low. These fibrils positive for type III collagen showed a clear cross-banding. In dentinal tubules, unusual collagen aggregations, segment long-spacing-like and fibrous long-spacing-like structures which were intensively stained for type I collagen but weakly so for type III collagen were seldom observed. Type III collagen-positive fibers often extended towards the pulp beyond the odontoblast layer, suggesting that these fibers were produced, at least partly, by the pulp cells. Key words: collagen antibodies, dentin, type I and III collagen.
Introduction Over fourteen genetically distinct types of collagen have been identified in recent years. Of these, the molecules of interstitial collagens (I, II, III, V, XI) which form fibers in the extracellular matrix of connective tissue, have been extensively characterized (Linsenmayer, 1981; Kiihn, 1987; Holzmann et aI., 1989). Although type III collagen has been observed in the dentin of patients with osteogenesis imperfecta (Sauk et aI., 1980; Gagr, 1985; Lukinmaa et aI., 1985; Lukinmaa, 1988), in hereditary opalescent dentin (Sauk et aI., 1980) and in human reparative dentin (Magloire et aI., 1988), most investigators have 3 Present address: Kojiro Kurisu, First Department of Oral Anatomy, Faculty of Dentistry, Osaka University, 1-8 Yamadaoka, Suita, Osaka 565, Japan.
agreed that type III collagen is absent from normal dentin, as well as from bone matrix (Cournil et aI., 1979; Tung et aI., 1985; Takita et aI., 1987; Andujar et aI., 1988; Andujar et aI., 1991). Recently, Becker et aI. (1986), using normal human teeth, immunohistochemically demonstrated that, although procollagen type III was virtually absent from intertubular dentin, dentinal tubules contained procollagen type III positive fibers, which appeared to be thicker than the tubular type I procollagen fibers, and that these fibers sometimes extended into the pulpal space through the strongly reactive predentin matrix which was positive for type III collagen. This report suggested that type III collagen is present in normal dentin. In addition, our preliminary study using the immunofluorescent method for the localization of type III collagen in normal mouse dentin demonstrated the presence of a few fibers positive for type III collagen (unpub-
Type III Collagen in Dentin Iished data). However, it is not easy to determine, by light microscopy, whether fibers positive for type III collagen exist in intertubular dentin or in dentinal tubules. Therefore, immunoelectron microscopic examinations should be carried out for the localization of type III collagen in normal dentin. In the present study, we obtained the immunofluorescent and immunoelectron microscopic localization of type III collagen in the normal dentin of mouse molars. Distribution of type I collagen in relation to type III collagen was also examined. Type III collagen was often observed in the normal dentin matrix of the root or crown region, but not in the dentinal tubules. Types III collagen-positive fibers ran parallel to the dentinal tubules either in peritubular dentin or within the vicinity of the dentinal tubules. They sometimes extended through the predentin into the pulpal space. These observations support those of Becker et al. (1986) except for the fact that type III collagen fibers were not found to be present in the dentinal tubules, but in the dentin matrix.
Materials and Methods Tissue preparation
The materials used were the maxillary molars of CFl mice, 13 days to 4 months old. Preparation of materials was as described by Huang et al. (1991). Mice were perfused from the left ventricle with a fixative (0.1 M cacodylate buffer containing 0.1 % glutaraldehyde and 4% paraformaldehyde). Maxillae removed from the mice were immersed in the fixative for 12-24h at 4°C. They were then decalcified with 10% EDTA for 2-3 weeks at 4°C. Samples for light microscopic examinations were embedded in paraffin. For post-embedding immunoelectron microscopy, the samples embedded in gelatin were cut with a Microslicer (Dohsaka, Kyoto, japan) into thin slices with a thickness of 100 !-tm. The slices were them embedded in Lowicryl K4M and polymerized with UV light at - 20°C for 5 h and at 4 °C for 2 h. Preparation ofantigens and antibodies
Types I and III collagen were extracted from adult CFl mouse skin with acetic acid and pepsin, respectively, and then purified by differential salt precipitation and DEAEcellulose chromatography, as described by Chung and Miller (1974). Their purity was checked by SDS-PAGE. Antitype I and anti-type III collagen antibodies (titer 7) were raised in rabbits. Anti-type III collagen antibody (titer 9) used for the double-staining method were raised in guinea pigs. These antibodies (lgG) were then purified by affinity chromatography using Sepharose-bound antigens. Both antibody preparations were adsorbed by the other antigens in order to make them monospecific for type I or type III
449
collagen. Affinity-purified non-labeled and FITC-labeled goat anti-rabbit IgG were purchased from Organon Teknika Corp. (West Chester, PAl. Affinity-purified nonlabeled and rhodamine-labeled goat anti-guinea pig IgG were purchased from Chemicon International Inc. (Temecula, CAl. The labeling of non-labeled antibodies with colloidal gold was performed as described by Slot et al. (1984; 1985). Immunofluorescent microscopy
The double-staining technique was performed as described by Engel et al. (1980). Briefly, paraffin sections were incubated at room temperature consecutively with (1) 2% bovine serum albumin (BSA) for 1 h, (2) 100 !-tg/ml rabbit anti-type I plus 100 !-tg/ml guinea pig anti-type III collagen antibodies for 1 hand (3) FITC-labeled goat antirabbit IgG plus rhodamine-labeled goat anti-guinea pig IgG diluted 1:500 in PBA (0.1 M PBS containing 1% BSA and 0.1 M lysine-HCl) for 30 min. Between each step, samples were thoroughly rinsed with PBS. In the control sections, antibody solutions adsorbed with corresponding antigens, PBS, normal rabbit or normal guinea pig IgG were used in place of primary antibodies to verified the specificity of primary as well as secondary antibodies. They were then sealed with a PBS-buffered glycerol to p-phenylenediamine had been added in order to reduce the fading of the fluorescence, as recommended by johnson and Arujo (1981). Immunoelectron microscopy
The post-embedding technique was carried out as follows: The ultra-thin sections mounted on nickel grids coated with 1% collodion film were incubated sequentially in a moist chamber at room temperature with PBA for 30 min, rabbit anti-type I or anti-type III collagen antibodies for 1 h, and colloidal gold (0 11 nm)-labeled antirabbit IgG for 20 min. Between each step, the sections were thoroughly rinsed with PBS. In the simultaneous doublestaining, a mixture of rabbit anti-type I and guinea pig antitype III collagen antibodies was used for the primary antibodies, and that of colloidal gold (011 nm)-labeled goat anti-rabbit IgG and colloidal gold ( 6 nm)-labeled goat anti-guinea pig IgG was used for the secondary antibodies. Control sections were treated with PBS, normal serum or antibodies adsorbed with corresponding antigens in place of primary antibodies. The sections were finally stained with uranyl acetate and examined with a JEOL 200CX electron microscope.
Results Immunofluorescent examination
In the longitudinal section of roots (Fig. 1), there were thick fibrous structures positive for type III collagen which
450
K. Nagata et al.
Figs. 1- 3 are immunofluorescent micrographs of the mandibular first molar in 4-month-old mice that were stained for either type I (a) or type III (b) collagen. Fig. 1: Section longitudinal to the dentinal tubules of the root which was double-stained for type I (a) and type III collagen (b). (a) Type I collagen is homogeneously distributed in dentin (D), the staining of which is less intense than that of predentin (Pd). Cytoplasm of odontoblasts (arrows) is also positive. (b) Thick fibrous structures showing a positive reaction for type III collagen (arrows) can be observed parallel to the dentinal tubules. Some of them (arrow heads) extend through the predentin into the pulp space (P). PL: periodontal ligaments. x 520. Bar = 25 !-tm. Fig. 2: Section transverse to the dentinal tubules of the root which was stained for either type I (a) or type III collagen (b). (b) Positive reactions can be observed as spots in dentin. PL: periodontal ligaments, D: dentin. x 520. Bar = 25 !-tm. Fig. 3: Longitudinal section to the dentinal tubules of crown dentin which was stained for either type I (a) or type III collagen (b). (b) Positive fibrous structures were rarely observed (arrows) in dentin. D: dentin. x 520. Bar = 25 !-tm. Fig. 4: Control section of root incubated with normal serum in place of primary antibodies. Positive staining was not observed. x 520. Bar = 25 !-tm.
were parallel to the dentinal tubules (Fig. 1 b). Some of them extended through the predentin into the pulpal space. Cytoplasm of odontoblasts did not show any positive reaction for type III collagen. On the other hand, type I collagen was homogeneously distributed in the dentin matrix, the inten-
sity of which was weaker than that of predentin (Fig. 1 a). Cytoplasm of odontoblasts was also positive for type I collagen. In the transverse section of roots, type III collagenpositive spots were observed in dentin (Fig. 2). Fibrous structures, positive for type III collagen, were observed in
Type III Collagen in Dentin
451
Figs. 5 -13 except for 9 and 10. Immunoelectron micrographs for type III or type I collagen in the root dentin of 4-month-old mice. Fig. 5: Section longitudinal to dentinal tubule stained for type III collagen. Fibril bundles running parallel with dentinal tubules show positive staining (arrow). x 21,000. Bar = 0.5I!m. The inset is shown at high magnification. X 57,000. Fig. 6: Section longitudinal to dentinal tubule stained for type III collagen. Close relation between the positive fibrils and the peritubular dentin reveals that the positive fibrils run either within the peritubular dentin or the vicinity of the dentinal tubules. X 21,000. Bar = 0.5I!m. Fig. 7: Section transeverse to dentinal tubule stained for type III collagen. Positive labeling can be observed in the peritubular dentin. DT: dentinal tubule. X 21,000. Bar = 0.5I!m. Fig. 8: Section transverse to dentinal tubules stained for type III collagen. Labeling is absent in the intertubular dentin, including the peritubular dentin. X 21,000. Bar = 0.5 ~m.
the same pattern as those in crown dentin (Fig. 3), but the frequency of their appearance was much less than that in root dentin. Control sections incubated with normal sera in place of primary antibodies showed negative staining (Fig. 4). Control sections incubated with PBS or the antibody solution adsorbed with corresponding antigens in place of primary antibodies also showed negative staining for both type III and type I collagen, as shown previously (Huang et al., 1991). The staining status of these control sections revealed that the antibodies used were monospecific for their corresponding antigens.
Immunoelectron examination
In the section longitudinal to the dentinal tubules, positive staining for type III collagen was observed on fiber bundles running parallel with the dentinal tubules (Fig. 5). These fibers were observed either within peritubular dentin, or within the vicinity of the dentinal tubules (Fig. 6). In the section transverse to the dentinal tubules, labeling for type III collagen was observed either in peritubular dentin or within the vicinity of some dentinal tubules (Fig. 7). However, in a large number of the dentinal tubules, no labeling was observed either at peritubular dentin or intertubular dentin (Fig. 8). In about 10% of the tubules at the
452
K. Nagata et al.
Fig. 9 and 10: Conventional electron micrograph of root dentin in rhe mandibular molar of 4-month-old mice. The dentin was embedded in Epon as described in Materials and Methods. FLS-like structure can be observed within the dentinal tubule (Fig. 9. x 57,000. Bar = 0.2 Ilm). SLS-like structures can be observed within the dentinal tubule (Fig. 10. x 21,000. Bar = 0.5 Ilm). Fig. 11: Obliquely sectioned FLS-like structures double-stained for type I and type III collagen. Note that these structures are intensively stained for type I collagen (0 11 nm) but weakly so for type III collagen (0 6 nm). FLS-like structure shows the same staining pattern as that of a SLS-like structure (arrow). x 31,000. Bar = 0.5 Ilm. Fig. 12: Section containing SLS-like structures double-stained for type I and type III collagen. Note these structures are intensively stained for type I collagen (0 11 nm) as in a SLS-like structure shown in Fig. 11. x 42,000. Bar = 0.2 Ilm. Fig. 13: Control section incubated with normal serum in place of primary antibodies. Labeling is absent in all areas. X 21,000. Bar = 0.5 Ilm.
Type III Collagen in Dentin middle portion of the root, positive labeling was observed in peritubular dentin. Two types of collagen aggregates, segment long-spacing (SLS)-like and fibrous long-spacing (FLS)-like structures were observed in some of the dentinal tubules of the root dentin at a low frequency (Figs. 9, 10, 11 and 12). FLS-like structure had about 160 nm periodicity (Figs. 9 and 11), and SLS-like structure had a symmetric banding pattern about 320 nm in length, but with a varying width (Figs. 10 and 12). These structures were intensively stained for type I collagen but weakly so for type III collagen (Figs. 11 and 12). Control sections incubated with normal sera showed no labeling, suggesting the high immunospecificity of the staining method used (Fig. 13).
Discussion There has been a consensus that type III collagen is not present in dentin, except for that of patients suffering from type IV osteogenesis imperfecta with dentinogenesis imperfecta (Sauk et al., 1980; Gage, 1985; Lukinmaa et al., 1988; Lukinmaa, 1988) and in reparative dentin (Magloire et al., 1988). Recently, however, Becker et al. (1986) reported the presence of type III collagen in the dentinal tubules of normal human molars by the use of immunohistochemical methods at the light microscopic level. In the present study, in order to obtain a critical localization of type III collagen in normal dentin, we carried out immunohistochemical examinations at light as well as electron microscopic levels, using affinity purified specific antibodies to type III and type I collagen, from which those cross-reacting to other antigens were eliminated. Our results demonstrated that type III collagen fibers certainly exist in, or close to, the peritubular dentin of normal dentin, but that they were absent within the dentinal tubules, as suggested by Becker et al. (1986). We assume that other researchers failed to identify type III collagen as being present within dentinal tubules, let alone in or close to peritubular dentin, since it is not easy to obtain a critical localization of immunohistochemically positive fibers in thick sections at the light microscopic level. Becker et al. (1986) was also able to observe type III collagen-positive fibers extending from the dentin to the pulp through the odontoblastic layer, which was consistent with our findings. In the dentinal tubules, we observed two types of unusual collagen aggregates, SLS-like and FLS-like structures, both of which showed positive labeling for type I collagen, but weakly so for type III collagen. These structures have been observed in predentin as well in the dentin of normal rat incisors (Warshawsky, 1972; Weinstock, 1977). The presence of uniform structures has been reported in predentin and in the incisors of rats treated with either vinblastine (Miake and Takuma, 1985) or sodium fluoride (Araki, 1989). SLS in the longitudinal section had centrosymmetrical striations associated with filamentous materials at both
453
ends. From this characteristic structure, it has been supposed that SLS consists of laterally packed antipolar procollagen molecule, in which the N-terminal end of one, overlays the C-terminal end of another (Chapman and Armitage, 1972; Weinstock, 1977; Miake and Takuma, 1985; Araki, 1989). FLS-like structures are supposed to be FLS IV which was thought to have been formed by antipolar dimers aggregating side by side, with sliding by one half of them (Armitage and Chapman, 1971; Chapman and Armitage, 1972). Some reporters indicated the possibility that these unusual collagen aggregates were artifacts resulting from fixation or from decalcification. Araki (1989), however, neglected the possibility of them being artifacts, because these unusual collagen structures were observed in predentin and in the dentin of the incisors of rats treated with NaF, but they were not present in those of the controls. It is thought that these collagen structures arose as a result of the inhibition of the normal processing of the N- and Cterminals of procollagen molecules. In the present study, these unusual collagen structures were observed only in dentinal tubules, but not elsewhere in the matrix of predentin or dentin. In addition, they showed positive labeling for type I collagen, but weakly so for type III collagen. These findings indicate that the type III collagen-positive fibers in dentinal tubules reported by Becker et al. (1986) are different from the unusual collagen structures described above. The origin of the type III collagen fibers in predentin and dentin was not clarified in this study. However, the observation that these fibers often extend towards the pulp beyond the odontoblast layer, suggests that these fibers are produced by pulp cells, but not by odontoblasts. Becker et al. (1986) revealed a faint staining of the cytoplasm of the odontoblasts for type III collagen. However, we failed to detect any positive staining of the cytoplasm of the odontoblasts for type III collagen, although we found positive staining for type I collagen, indicating odontoblasts do not synthesize type III, whereas they do type I collagen. Andujar et al. (1991) revealed that type III collagen mRNAs were not detected in odontoblasts during mouse molar tooth development by in situ hybridization. It is possible, however, that the amount of type III collagen synthesized by odontoblasts is too small to be detected by the immunohistochemical method employed in this study. The role of type III collagen in normal dentin is not clear from this study. The cross-bandings of collagen fibrils in dentin matrix showing positive labeling for type I collagen were more obscure compared with those of fibrils positive for type III collagen, indicating that the latter fibrils are not calcified at all, or else are only slightly calcified. Although type III collagen was believed to be absent from mineralized tissue, Keene et al. (1991) revealed that type III collagen were present in discrete fiber bundles throughout the human bone cortex. Type III collagen is predominant in distensible tissue such as the walls of blood vessels, the digestive tract and the lung, this type of collagen may
454
K. Nagata et al.
participate in the elasticity of tissue (Miller, 1976). The presence of uncalcified fibrils in dentin may have some effect on the physical property of dentin. However, type III collagen fibrils are often observed in the dentin of an unphysiological state, for example, in the dentin of odontogenesis imperfecta (Sauk et aI., 1980; Gage, 1985; Lukinmaa et aI., 1985; Lukinmaa, 1988) or in reparative dentin (Magloire et aI., 1988), suggesting the possibility that as a result of the unphysiological condition of those cells relating to dentinogenesis, especially odontoblasts, the inclusion of type III collagen may be induced in peritubular dentin or dentin matrix. There is some controversy concerning the presence of cross-banding in type III collagen fibrils. Several studies have demonstrated that type III collagen can be observed in fibrils without cross-banding in human skin (Fleischmajer et ai, 1981) and in dog gingiva (Cho et aI., 1987). Amenta et al. (1986) showed that type I collagen existed on thick and cross-banded fibrils, while type III collagen existed exclusively on thin, beaded noncross-banded fibrils. Keene et al. (1987), however, using a monoclonal antibody specific for human type III collagen, revealed that type III collagen was present on cross-banded fibrils in skin, tendons, and amnion, regardless of the fibril diameter. Huang et al. (1991) also demonstrated the presence of type III collagen in crossbanded fibrils in the mouse periodontal ligament. In the present study, type III collagen was found in the crossbanded fibrils, supporting the presence of type III collagen fibrils with cross-banding.
Acknowledgements A part of this study was supported by Grant-in-Aid for Scientific Research from the Ministry of Education of Japan (Project 61571860).
References Amenta, P. S., Gay, S., Vaheri, A. and Martinez-Hernandez, A.: The extracellular matrix is an integrated unit: Ultrastructural localization of collagen type I, III, IV, VI, fibronectin, and laminin in human term placenta. Collagen ReI. Res. 6: 125 -152, 1986. Andujar, M. B., Hartmann, D.J., Emonard, H. and Magloire, H.: Distribution and synthesis of type I and type III collagens in developing mouse molar tooth root. Histochemistry 88: 131-140, 1988. Andujar, M. B., Couble, P., Couble, M. L. and Magloire, H.: Differential expression of type I and type III collagen genes during tooth development. Development 111: 691-698, 1991. Araki, S.: Ultrastructral changes in rat-incisor odontoblasts and dentin caused by administration of sodium fluoride. Shika Gakuho 89: 49-91,1989. Armitage, P. M. and Chapman,]. A.: New fibrous long spacing form of collagen. Nature 229: 151-152,1971. Becker,J., Schuppan, D., Benzeian, H., Bals, T., Hahn, E. G., Cantaluppi, C. and Reichart, P.: Immunohistochemical distribution
of collagens type IV, V and VI and of pro-collagens types I and III in human alveolar bone and dentine. J. Histochem. Cytochem. 34: 1417-1429,1986. Chapman,].A. and Armitage, P.M.: An analysis of fibrous longspacing forms of collagen. Conn. Tissue Res. 1: 31-37, 1972. Cho, M. L., Lee, Y. L. and Garant, P. R.: Immunocytochemical localization of extracellular matrix components in beagle periodontium: I. Collagen type I and type III in healthy gingival connective tissue. J. Period. Res. 22: 313-319, 1987. Chung,E. and Miller, E.].: Collagen polymorphism: Characterization of molecules with the chain composition [a1(III)h in human tissues. Science 183: 1200-1201,1974. Cournil, I., Lebrond, C. P., Pomponio,]., Hand, A. R., Sederlof, L. and Martin, G. R.: Immunohistochemical localization of procollagens. I. Light microscopic distribution of procollagen I, III and IV antigenicity in the rat incisor tooth by the indirect peroxidase-anti-peroxidase method. J. Histochem. Cytochem. 27: 1059-1069,1979. Engel, D., Schroeder, H. E., Gay, R. and Clagett,].: Fine structure of cultured human gigival fibroblasts and demonstration of simultaneous synthesis of types I and III collagen. Archs. Oral BioI. 25: 283-296, 1980. Fleischmajer, R., Tuderman, L., Raisher, L., Wiestner, M., Perlish,J. S. and Graves, P. N.: Ultrastructural identification of extension aminopropeptides of type I and type III collagens in human skin. Proc. Natl. Acad. Sci. USA 78: 7360-7364, 1981. Gage, I. P.: Dentinogenesis imperfecta. A new perspective. Austr. Dent. j. 30: 285 - 290,1985. Holzmann, B., McIntyre, B. W. and Weissman, I. L.: Identification of a murine Peyer's patch-specific lymphocyte homing receptor as an integrin molecule with an a chain homologous to VLA-4a. Cell 56: 37-46, 1989. Huang, Y. H., Ohsaki, Y. and Kurisu, K.: Distribution of type I and type III collagen in the developing periodontal ligament of mice. Matrix 11: 100-108,1991. Johnson,G.D. and Araujo,G.M.C.N.: A simple method of reducing the fading of immunofluorescence during microscopy. J. Immunol. Methods 42: 349-350, 1981. Keene, D. R., Sakai, L. Y., Bachinger, H. P. and Burgeson, R. E.: Type III collagen can be present on banded collagen fibrils regardless of fibril diameter. J. Cell Bioi. 105: 2393-2402, 1987. Keene, D. R., Sasaki, L. Y. and Burgeson, R. E.: Human bone contains type III collagen, type IV collagen, and Fibrillin: Type III collagen is present on specific fibers that may mediate attachment of tendons, ligaments and periosteum to calcified bone cortex. J. Histochem. Cytochem. 39: 56-69, 1991. Kuhn, K.: The classical collagens. In: Types I, II and III. In: Structure and Function of Collagen Types. ed. by Mayne, R. and Burgeson, R.E., Academic Press, New York, 1987, pp. 1-42. Linsenmayer, T. F.: Collagen. In: Cell Biology of Extracellular Matrix, ed. by Hay,E.D., Plenum Press, New York, 1981, pp.5-37. Lukinmaa, P. L., Ranta, H., Ranta, K., Peltonen, L. and Hietanen,].: Demineralization of dentin with EDTA in organic solvent: Immunofluorescence of collagen in ostegenesis imperfecta and normal teeth. Collagen Rei. Res. 5: 505 - 512,1985. Lukinmaa, P. L.: Immunofluorescent localization of type III collagen and the N-terminal propeptide of type III procollagen in dentin matrix in osteogenesis imperfecta. J. Craniofac. Genet. Develop. Bioi. 8: 235-243, 1988. Magloire, H., Joffre, A. and Hartman, D.].: Localization and synthesis of type III collagen and fibronectin in human reparative dentine. Histochemistry 88: 141-149,1988.
Type III Collagen in Dentin Miake, Y. and Takuma, S.: Unusual collagen aggregates induced in rat incisor dentin after vinblastine administration. Calcif. Tissue Int. 37: 501-510, 1985. Miller, E.].: Biochemical characteristics and biological significance of the genetically-distinct collagens. Mol. Cell Biochem. 13: 165-192,1976. Sauk,].]., Gay, R., Miller, E.J. and Gay, S.: Immunohistochemical localization of type III collagen in the dentin of patients with osteogenesis imperfecta and hereditary opalescent dentin. J. Oral Path. 9: 210- 220,1980. Slot,J. W. and Geuze, H.J.: Gold markers for single and double immunolabelling of ultrathin cryosection. In: Immunolabelling for Electron Microscopy, ed. by Polak,]. M. and Varndell, I. M., Elsevier Science Publishing, Amsterdam, 1984, pp. 129-142. Slot,J. W. and Geuze, H.].: A new method of preparing gold pobes for multiple labeling cytochemistry. Eur. J. Cell Bioi. 38: 87-93,1985. Takita, K., Ohsaki, Y., Nakata, M. and Kurisu, K.: Immuno-
455
fluorescence localization of type I and type III collagen and fibronectin in mouse dental tissues in late development and during molar eruption. Archs. Oral BioI. 32: 273-279, 1987. Tung, P. S., Domenicucci, C., Wasi, S. and Sodek,].: Specific immunohistochemical localization of osteonectin and collagen types I and III in fetal and adult porcine dental tissues. J. Histochem. Cytochem. 33: 531-540, 1985. Warshawsky, H.: The presence of atypical collagen fibrils in EDTA decalcified predentin and dentin of rat incisors. Archs. Oral Bioi. 17: 1745-1754,1972. Weinstock, M.: Centrosymmetrical cross-banded structures in the matrix of rat incisor predentin and dentin. J. Ultrastruct. Res. 61: 218-229, 1977. Dr. Kengo Nagata, Second Department of Anatomy, Faculty of Dentistry, Kyushu University, Maidashi, Higashi-ku, Fukuoka 812,Japan.
