THE ANATOMICAL RECORD 248:339–345 (1997) Immuno-Scanning Electron Microscope Characterization of Large Tubules in Human Deciduous Dentin HIROKO AGEMATSU1*, TAKASHI SAWADA2, HIROKI WATANABE2, TAKAAKI YANAGISAWA2, AND YOSHINOBU IDE1 1Department of Anatomy, Tokyo Dental College, 1-2-2 Masago Mihamaku Chiba, Japan 2Department of Ultrastructural Science, Tokyo Dental College, 1-2-2 Masago Mihamaku Chiba, Japan ABSTRACT Background: This study was undertaken to elucidate the type and origin of collagen fibrils which construct the large tubules in deciduous coronal dentin by scanning electron microscope and anti-types I and III collagen antibody procedures. Methods: The studies were performed on human deciduous teeth. The teeth were fixed in 4% paraformaldehyde solution and then fractured either mesio-distally parallel to the long axis of the tooth or transversely perpendicular to the long axis of the tooth crown. The specimens were three-dimensionally observed employing the scanning electron microscope to distinguish the content of large tubules. Polyclonal antibodies of anti-type I and anti-type III collagen with 20 nm colloidal gold, and secondary electron imaging and backscatter electron imaging of highresolution field emission scanning electron microscopy were used to examine the types of collagen fibrils. Results: The large tubules extended from the vicinity of the incisal edge of the dentino-enamel junction to the pulp cavity. Inside the large tubules, fibers in compact bundles run parallel to the longitudinal axis of the tubules. The fiber bundles consisted of collagen fibrils which were 50-150 nm in diameter with typical cross striation. Immuno-scanning electron microscopy showed type I collagen-labelling gold particles and type III collagen-labelling gold particles to be abundant on the fibrils. Types I and III collagen-labelling gold particles were present on the banded collagen fibrils regardless of their diameter. Conclusions: It was found that type III collagen is present together with type I collagen on the fibrils constructing the large tubules of the human deciduous dentin. This immunohistochemical study suggested that the fibrils constructing the large tubules were derived from the von Korff fibers, and types I and III collagens formed copolymers. Anat. Rec. 248:339–345, 1997. r 1997 Wiley-Liss, Inc. Key words: large tubules; human deciduous dentin; type I collagen; type III collagen; immuno-scanning electron microscopy INTRODUCTION The dentin is penetrated by numerous dentinal tubules that extend through its entire thickness from the dentino-enamel junction (D.E.J.) to the pulp. The diameter of the dentinal tubules commonly range from 1 to 4 µm. Reports have shown that tubules which were clearly larger (5-50 µm) in diameter as compared to the ordinary dentinal tubules extend from the incisal edge of the D.E.J. to close vicinity of the pulp cavity in the coronal dentin of cow, red deer and human (Tronstad, 1972; Miller, 1981; Hals, 1983a,b; Hals and Olsen, 1984; Dyngeland and Fosse, 1986; Dyngeland, 1988; Agematsu et al., 1990). These tubules with large diameters have been referred to as giant tubules, large tubules, and microcanals. In this study, the term large tubules will be used. r 1997 WILEY-LISS, INC. In a histochemical study, Dyngeland et al. (1984) reported that in immature unerupted cow incisors, the large tubules were observed to communicate with the incisal pulp cavity via its wide openings and vital cells of the pulp were present in the large tubules. Furthermore, Dyngeland and Kvinnsland (1988) reported that in bovine dentin, large tubules consisted of a vascularized pulp portion and a large collagen filled portion. Contract grant sponsor: Ministry of Education, Science, and Culture (Japan); contract grant number: 08672093. *Correspondence to: Hiroko Agematsu, D.D.S., Department of Anatomy, Tokyo Dental College, 1-2-2 Masago Mihamaku Chiba, 261 Japan. Received 6 August 1996; accepted 10 February 1997. 340 H. AGEMATSU ET AL. Proceeding fracturing, these fragments were fixed again in 4% paraformaldehyde solution. Scanning Electron Microscopy Mesio-distally and transversely fractured dentin fragments were dehydrated in a series of ethanol, immersed in t-butyl alcohol, frozen, and dried in an Eiko ID-2, and gold-palladium coated using a cool sputter coater. The specimens were observed by a Hitachi S-800 FESEM operated at 15 kV. Immunostaining for Scanning Electron Microscopy Fig. 1. Diagrams of preparation procedure for exposing the large tubules. a: In the longitudinal plane. b: In the transversal plane. Recent observations (Agematsu et al., 1990) of the large tubules in the human deciduous dentin employing the scanning electron microscope revealed collagen fibers, but no odontoblast processes or other cells in the large tubules. It has become accepted in recent years that collagen is a significant component of the dentinal tubules. Dai et al. (1991) showed that intratubular collagen is a significant feature especially at the pulpal side of human dentin. However, it is uncertain whether or not the origin and distribution of collagen fibrils in the large tubules are similar to those of the dentinal tubules. The nature of the collagen fibrils within the large tubules have also not yet been defined in human teeth. In this study, we investigated the three-dimensional organization and distribution of collagen fibrils within the large tubules in human dentin using the highresolution field emission scanning electron microscopy (FESEM). Furthermore, we attempted to demonstrate the immunohistochemical characterization of collagen fibrils by the immunogold technique in order to reveal the nature of the collagen fibrils within the large tubules. MATERIALS AND METHODS Samples The experimental material consisted of 35 human deciduous incisors extracted from children aged 5-7. The teeth showed slight mobility and the degree of root resorption was 1⁄2 to 3⁄4 of its original length. All teeth were clinically caries free and with slight attrition. Immediately after extraction, the teeth were fixed in 4% paraformaldehyde solution in PBS at 4°C for 24 h. The mesial and distal surfaces of the teeth were ground down with a grindstone to expose the pulp cavity. Specimens were then frozen with liquid nitrogen and fractured mesio-distally parallel to the long axis of the tooth with a chisel (Fig. 1a). In several specimens, the incisal third was fractured transversely perpendicular to the long axis of the tooth with a chisel (Fig. 1b). For indirect immunostaining with antibodies using the colloidal gold procedure, the mesio-distally fractured dentin fragments were used. The labial fragments and lingual fragments were used respectively for type I collagen and type III collagen immunolabelling. Test specimens were rinsed in 0.01 M PBS (pH7.2) and immersed in PBS containing 1% chicken egg albumin (OVAL, Sigma) for 30 min. This was followed by incubation with rabbit polyclonal antiserum either to collagen type I (rabbit anti-bovine type I collagen, LSL Co. Ltd. Tokyo, 1:200) (Sasano et al., 1996; Kikuchi et al., 1996) or to collagen type III (rabbit anti-bovine type III collagen, LSL Co. Ltd. Tokyo, 1:80) (Kikuchi et al., 1996) for 12 h at 4°C. Controls were incubated either with PBS only, or with non-immune serum obtained from rabbits. All specimens were then rinsed well with PBS, followed by incubation with goat anti-rabbit IgG conjugated with 20 nm colloidal gold (BioCell Research Lab. UK, 1:50) for 12 h at 4°C. After immunostaining, the specimens were fixed in 1% glutaraldehyde buffered solution with PBS, dehydrated in a series of ethanol, and processed by sublimation. The specimens were then platinum coated (2-5 nm) with a cool sputter coater and examined by a Hitachi S-800 high-resolution FESEM for secondary electron imaging (SEI). In addition, immunogold particles were identified by the atomic number contrast which were obtained via backscattered electron imaging (BEI). RESULTS Scanning Electron Microscopic Findings The large tubules which were 10-20 times bigger in diameter as compared to the dentinal tubules were either round or ovoid in shape. They were aligned in a straight line from the mesial to the distal end, at close to the labio-lingual central portion of the coronal dentin (Fig. 2). The large tubules which are easily observed as dark grooves were directed toward the pulp cavity from the vicinity of the incisal edge of the D.E.J. Running parallel to the dentinal tubules, in the mesio-distally fractured dentin in which incisal dentinal exposure was minimum, they were more or less converged in the pulp vicinity. The distance between each large tubule was approximately 100-500 µm (Fig. 3). In its periphery, there was a wall structure similar to that of the peritubular dentin of the dentinal tubule. The inner axis of the large tubules having a diameter of approximately 20 µm was largely composed of fiber bundles that ran longitudinally and these small spherical bodies were present between these collagen fibers on the Fig. 2. Low magnification SEM micrograph of transverse fracture. The large tubules (LT) are arranged linearly in the mesio-distal direction. Bar scale indicates 100 µm. Fig. 3. Low magnification SEM micrograph of mesio-distal fracture. The large tubules (LT) are clearly observed as grooves extending from the incisal edge towards the pulp cavity. Bar scale indicates 200 µm. Fig. 4. The large tubule (LT) has a glossy wall (asterisks) similar to the peritubular dentin. The interior of the large tubule is composed of bundles of longitudinally oriented fibers and of numerous spherical shaped bodies (arrows). Bar scale indicates 20 µm. Fig. 5. Higher magnification of interior of the large tubule. Longitudinally oriented collagen fibrils (Col) with periodical cross striations and spherical shaped bodies (SB) which are made up of aggregates of regular parallelepipedical crystals (a) and finely granulated crystals (b) were also observed. Bar scales indicate 0.5 µm. 342 H. AGEMATSU ET AL. longitudinally fractured surface (Fig. 4). In addition, enlargements of the interior of the large tubules in the longitudinally fractured specimens revealed longitudinally oriented fiber bundles and lateral branches which have separated from the bundles. Under higher magnification, these fiber bundles appeared to consist of collagen fibrils that had a cross striation structure with a periodicity of approximately 60-70 nm and ranged from 50 to 150 nm in diameter (Fig. 5a,b). The spherical calcified bodies (0.5-2.5 µm in diameter) were composed of regular parallel-epipedal crystals (Fig. 5a) and granulated crystals (Fig. 5b) were also observed to be among the fibers. Scanning Electron Microscopic Findings on Immunostaining Immunostaining with type I collagen antibody revealed that gold particles were especially abundant on the collagen fibrils in the large tubules and appeared throughout its long axis. Immunogold labelling for type I collagen was observed on the collagen fibrils regardless of their diameter (Fig. 6). Under higher magnification, gold particles indicating type I collagen were clearly observed to be on the collagen fibrils which had approximately 60-70 nm periodicity and were 50-150 nm in diameter. Gold particles could not be found on the crystals which lie below the collagen fibrils (Fig. 7). Figure 8a and b show a higher magnification of SEI and BEI on the large tubule immunostained with collagen type I at the same location. The 20 nm immunogold particles attached on to the cross striation of the collagen fibrils were clearly identified by their shape and size, as observed by secondary electron imaging (Fig. 8a). Furthermore, owing to the atomic number contrast of BEI, the colloidal gold particles could easily be seen as bright particles (Fig.8b). Controls using normal rabbit serum resulted in minimum immunostaining and only few randomly distributed gold particles on the collagen fibrils within the large tubule were visible (Fig.9). Immunostaining with type III collagen antibody showed that immunogold labelling overlay the collagen fibrils within the large tubules (Fig. 10). The 20 nm colloidal gold particles coated with the platinum could be identified by SEI (Fig. 11a). This was further confirmed by visualizing the colloidal gold marker with atomic number contrast obtained through BEI (Fig. 11b). DISCUSSION Dyngeland and colleagues (Dyngeland, 1988; Dyngeland et al., 1984; Dyngeland and Fosse, 1986, 1987; Dyngeland and Kvinnsland, 1988) have made a series of investigations on the structure of the large tubules using the bovine dentin. However, reports on the human large tubule are few in number. In this study, our findings yielded some additional information on the formation and content of the large tubule in human dentin. Our observation which revealed that the large tubules were arranged linearly in the mesio-distal direction and only in the labio-lingual central portion of the incisal dentin is very interesting. The dentin below the incisal edge has characteristics slightly different from those of other dentinal regions (Tronstad, 1972; Mjor and Fejerskov, 1986). Agematsu et al. (1990) reported on the continuous vertical alignment of interglobular dentin under the incisal edge. Their findings are suggestive of dentinogenesis imperfecta of dentin beneath the incisal edge. Tronstad (1972) observed a slit which appeared to be located at the junction between the dentin developing from the buccal and from the lingual dentinal surfaces, and has assumed that the slit is probably caused by crowding of odontoblasts during dentinogenesis. At the initiation of dentin formation, odontoblasts differentiate at the D.E.J. and shift inward to where surface area is less. Consequently, we consider that the dentin formed on the labial side and on the lingual side collide, causing degeneration and disruption of odontoblasts during dentinogenesis at directly below the incisal center, and because of this phenomena, tubules of bigger diameter appear. As a result of electron microscopic observations on alterations occurring with increasing age in odontoblasts in the pulp horn directly below the cusp of the first molar in rats, Shimomura (1979) revealed that, in this region, cytoplasmic organelles in odontoblasts either degenerate or that the appearance of autophagic vacuoles result in the destruction or vanishing of some odontoblasts as dentinogenesis proceeds. From this report, we made the following hypothesis: during dentin formation, apoptosis of odontoblasts occur in the incisal dentin. Wright and Gantt (1985) reported that large tubules appeared more frequently in teeth with dentinogenesis imperfecta. On the basis of these findings, the authors speculated that the large tubules appear localized in the labio-lingual junction below the incisal edge for a reason; partial imperfect dentin formation occurs as a result of apoptosis induced by odontoblast concentration in this area. We observed the content of the large tubule in detail. As a result, odontoblast processes and other cell component were not found within the large tubule in any area. The majority of large tubules was seen to be filled with dense bundles of longitudinally oriented collagen fibrils with typical cross striations. Thus we attempted identification of collagen type using immuno-scanning electron microscopy which is useful in observing the localization of specific substances in three dimensions (Takata et al., 1988). We found the presence of type III collagen together with type I collagen on the collagen fibrils within the large tubule of human deciduous dentin. In general, normal mineralized human dentin is shown to be composed mainly of type I collagen and to lack type III collagen (Thesleff et al., 1979; Wang et al., 1980; Tung et al., 1985). But, type III collagen is present in human dentin where the organic matrix is newly formed in dentinogenesis imperfecta (Sauk et al., 1980) and is reparative dentin (Magloire et al., 1988). Nagata et al. (1992) described that positive staining for type III collagen was observed on the cross-banded fibril bundles running parallel to the dentinal tubules within the peritubular dentin or in vicinity of the dentinal tubules in the mice normal root dentin by immuno-electron microscopy. Recently, it has been reported that in normal human tooth, type III collagen is occasionally present in the peritubular dentin (Waltimo et al., 1994). We were able to detect type III collagen together with type I collagen only on the collagen fibrils within the large tubule. Fig. 6. Immunostaining for type I collagen coupled with 20 nm colloidal gold. Numerous immunogold particles (arrowheads) were attached to the banded collagen fibrils within the large tubule. Bar scale indicates 0.5 µm. Fig. 7. Higher magnification of collagen fibrils immunostaining for type I collagen coupled with 20 nm colloidal gold. Immunogold particles (arrowheads) are clearly observed on the banded collagen fibrils but are not observed on the crystals (arrows). Bar scale indicates 0.1 µm. Fig. 8. Higher magnifications of SEI (a) and BEI (b) of collagen fibrils immunostaining for type I collagen. Immunogold particles (arrowheads) detected on the collagen fibrils with SEI (a). Immunogold particles(arrowheads) are further identified by atomic number contrast with BEI (b). Bar scales indicate 0.2 µm. Fig. 9. Controls using normal rabbit serum. Immunogold particles can scarcely be observed on the collagen fibrils. Bar scale indicates 0.2 µm. Fig. 10. Immunostaining for type III collagen coupled with 20 nm colloidal gold. Immunogold particles(arrowheads) were attached to the banded collagen fibrils regardless of the fibril diameter. Bar scale indicates 0.2 µm. Fig. 11. SEI (a) and BEI (b) of collagen fibrils immunostaining for type III collagen. Immunogold particles (arrowheads) are detected on the collagen fibrils with SEI (a). Immunogold particles (arrowheads) are further identified by atomic number contrast with BEI (b). Bar scales indicate 0.2 µm. IMMUNO-SEM OF LARGE TUBULES IN HUMAN DENTIN It was reported that the presence of longitudinal collagen fibrils between odontoblasts, which run from the mineralized dentin to the odontoblast layer, were observed in pig molars (Bishop and Yoshida, 1992), and they regarded these fibrils as corresponding to the von Korff fiber. It has also been demonstrated that type III collagen is a major component of the fiber which runs among the odontoblasts and into the predentin (Ohsaki and Nagata, 1994). It is presumed that the fibrils which were observed within the large tubules in this study are similar to the von Korff fibers in distribution and collagen type. As a conclusion, we think that type III collagen positive fibrils within the large tubules were derived from von Korff fibers. It has been stated that type III collagen forms thinner fibrils as compared to type I collagen fibrils, and that type III collagen is generally considered to consist of reticular fibers (Romanic et al., 1991). However, Keene et al. (1987) suggested that type III collagen is present on all banded collagen fibrils regardless of diameter and that type III collagen is codistributed with type I collagen in most tissues. In this immunoscanning electron microscopic study, we found both type I and type III collagens on the banded collagen fibrils within the large tubules regardless of fibril diameter. We assume that type III collagen is codistributed together with type I collagen and that both of them probably form copolymers. ACKNOWLEDGMENTS This work was supported in part by grants-in-aid for scientific research (08672093: H.A.) from the Ministry of Education, Science, and Culture, Japan. LITERATURE CITED Agematsu, H., H. Watanabe, H. Yamamoto, M. Fukayama, T. Kanazawa, and K. Miake 1990 Scanning electron microscopic observations of microcanals and continuous zones of interglobular dentin in human deciduous incisal dentin. Bull. Tokyo Dental Coll., 31:163–173. Bishop, M.A. and S. Yoshida 1992 A permeability barrier to lanthanum and the presence of collagen between odontoblasts in pig molars. J. Anat., 181:29–38. Dai, X.-F., A.R. Ten Cate, and H. Limeback 1991 The extent and distribution of intratubular collagen fibrils in human dentine. Archs. Oral Biol., 36:775–778. Dyngeland, T., G. Fosse, and N.-P.B. Justesen 1984 Histochemical study of giant tubule content in dentin of unerupted cow incisors. Scand. J. Dent. Res., 92:177–182. Dyngeland, T. and G. Fosse 1986 Scanning electron microscopic, light microscopic and microradiographic study of giant tubules in bovine dentin. Scand. J. Dent. Res., 94:285–298. Dyngeland, T. and G. Fosse 1987 Scanning electron microscopic study of giant tubule content in bovine dentin. Scand. J. Dent. Res., 95:1–12. Dyngeland, T. 1988 Light microscopic and transmission electron microscopic study of initial phases of giant tubule formation in bovine dentin. Scand. J. Dent. Res., 96:479–488. Dyngeland, T. and S. Kvinnsland 1988 In vitro study of the giant tubule collagen formation in bovine dentin by [3H]-proline incorporation. Scand. J. Dent. Res., 96:317–323. 345 Hals, E. 1983a Observations on giant tubules in human coronal dentin by light microscopy and microradiography. Scand. J. Dent. Res., 91:1–7. Hals, E. 1983b A light microscopical and microradiographic study of coronal dentin in red deer with special reference to the occurrence of giant tubules. Scand. J. Dent. Res., 91:99–104. Hals, E. and H.C. Olsen 1984 Scanning electron and incident light microscopy of giant tubules in red deer dentin. Scand. J. Dent. Res., 92:269–274. Keene, D.R., L.Y. Sakai, H.P. Bächinger, and R.E. Burgeson 1987 Type III collagen can be present on banded collagen fibrils regardless of fibril diameter. J. Cell Biol., 105:2393–2402. Kikuchi, H., T. Sawada, and T. Yanagisawa 1996 Effects of a functional agar surface on in vitro dentinogenesis induced in proteolytically isolated, agar-coated dental papillae in rat mandibular incisors. Arch. Oral Biol., 41:871–883. Magloire, H., A. Joffre, and D.J. Hartmann 1988 Localization and synthesis of type III collagen and fibronectin in human reparative dentine: Immunoperoxidase and immunogold staining. Histochemistry, 88:141–149. Miller, J. 1981 Large tubules in dentin. J. Dent. Child., 48:269–271. Mjör, I.A. and O. Fejerskov 1986 Human Oral Embryology and Histology. Munksgaard, Copenhagen, pp. 90–130. Nagata, K., Y. Huang, Y. Ohsaki, T. Kukita, M. Nakata, and K. Kurisu 1992 Demonstration of type III collagen in the dentin of mice. Matrix, 12:448–455. Ohsaki, Y. and K. Nagata 1994 Type III collagen is a major component of interodontoblastic fibers of the developing mouse molar root. Anat. Rec., 240:308–313. Romanic, A.M., E. Adachi, K.E. Kadler, Y. Hojima, and D.J. Prockop 1991 Copolymerization of pN collagen III and collagen l. J. Biol. Chem., 266:12703–12709. Sasano, Y., M. Furusawa, H. Ohtani, I. Mizoguchi, I. Takahashi, and M. Kagayama 1996 Chondrocytes synthesize type I collagen and accumulate the protein in the matrix during development of rat tibial articular cartilage. Anat. Embryol., 194:247–252. Sauk, J.J., R. Gay, E.J. Miller, and S. Gay 1980 Immunohistochemical localization of type III collagen in the dentin of patients with osteogenesis imperfecta and hereditary opalescent dentin. J. Oral Pathol., 9:210–220. Shimomura, H. 1979 Electron microscope study on age changes of the pulpal horn areas of rat molars, with special reference to rearrangement of the odontoblasts. (in Japanese) Shikwa Gakuho, 79:179– 214. Takata, K., Y. Akimoto, K. Ogura, S. Yamagishi, and H. Hirano 1988 Colloidal gold label observed with a high resolution backscattered electron imaging in mouse lymphocytes. J. Electron Microsc., 37:346–350. Thesleff, I., S. Stenman, A. Vaheri, and R. Timpl 1979 Changes in the matrix proteins, fibronectin and collagen, during differentiation of mouse tooth germ. Dev. Biol., 70:116–126. Tronstad, L. 1972 Optical and microradiographic appearance of intact and worn human coronal dentine. Arch. Oral Biol., 17:847–858. Tung, P.S., C. Domenicucci, S. Wasi, and J. Sodek 1985 Specific immunohistochemical localization of osteonectin and collagen types I and III in fetal and adult porcine dental tissues. J. Histochem. Cytochem., 33:531–540. Waltimo, J., L. Risteli, J. Risteli, and P.-L. Lukinmaa 1994 Altered collagen expression in human dentin: Increased reactivity of type III and presence of type VI in dentinogenesis imperfecta, as revealed by immunoelectron microscopy. J. Histochem. Cytochem., 42:1593–1601. Wang, H.M., V. Nanda, L.G. Rao, A.H. Melcher, J.N.M. Heersche, and J. Sodek 1980 Specific immunohistochemical localization of type III collagen in porcine periodontal tissues using the peroxidaseantiperoxidase method. J. Histochem. Cytochem., 28:1215–1223. Wright, J.T. and D.G. Gantt 1985 The ultrastructure of the dental tissues in dentinogenesis imperfecta in man. Arch. Oral Biol., 30:201–206.