Dentinogenesis

Critical Reviews http://cro.sagepub.com/ in Oral Biology & Medicine Dentinogenesis Anders Linde and Michel Goldberg CROBM 1993 4: 679 DOI: 10.1177/10454411930040050301 The online version of this article can be found at: http://cro.sagepub.com/content/4/5/679 Published by: http://www.sagepublications.com On behalf of: International and American Associations for Dental Research Additional services and information for Critical Reviews in Oral Biology & Medicine can be found at: Email Alerts: http://cro.sagepub.com/cgi/alerts Subscriptions: http://cro.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav >> Version of Record - Jan 1, 1993 What is This? Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Critical Reviews in Oral Biology and Medicine, 4(5):679-728 (1993) Dentinogenesis Anders Linde Department of Oral Biochemistry, Faculty of Odontology, University of Goteborg, Sweden Michel Goldberg Laboratoire des Matrices Extracellulaires et Biomineralisations, Faculte de Chirurgie Dentaire, Universit6 Rene Descartes, Paris, France ABSTRACT: The formation of dentin, dentinogenesis, comprises a sophisticated interplay between several factors in the tissue, cellular as well as extracellular. Dentin may be regarded as a calcified connective tissue. In this respect, as well as in its mode of formation, it is closely related to bone. Using dentinogenesis as an experimental model to study biomineralization provides several practical advantages, and the results may be extrapolated to understand similar processes in other tissues, primarily bone. After describing dentin structure and composition, this review discusses items such as the morphology of dentinogenesis; the dentinogenically active odontoblast, transport, and concentrations of mineral ions; the constituents of the dentin organic matrix; and the presumed mechanisms involved in mineral formation. KEY WORDS: dentin, calcification, odontoblast, calcium, collagen, proteoglycan, phosphoprotein. I. INTRODUCTION Dentin, the most voluminous mineralized tissue of the tooth, may be considered a connective tissue. As for connective tissues in general, the constituents of the extracellular space primarily give the tissue its functional characteristics. In the case of dentin, the extracellular matrix has been modified so that it contains a mineral phase, thus making it possible for the tissue to fulfill the functional requirements placed on it. This review describes the morphology and biochemical mechanisms comprising dentinogenesis, the formation of dentin. Dentinogenesis involves a chain of different mechanisms such as cell differentiation and interactions, the synthesis of an organic matrix, and the eventual formation of mineral crystals in this extracellular matrix. As a basis for these considerations, a description of dentin morphology will be given, and the different chemical constituents of dentin will be discussed in some detail. The subject of dentinogenesis has been discussed earlier to different extents in reviews and textbooks. An excellent, standard textbook in oral histology is that of Ten Cate (1989), now in its third edition. The two volumes published in 1984, edited by Linde, provide an extensive treatise on dentin and dentinogenesis but suffer in part from being a decade old (Linde, 1984a). The monograph of Goldberg et al. Abbreviations used: BM, basement membrane; Ca-ATPase, Ca2+-activated adenosine triphosphatase; EDTA, ethylenediaminetetraacetic acid; EGTA, ethyleneglycol-bis(P-aminoethylether)NN-tetraacetic acid; GAG, glycosaminoglycan; Gla, y-carboxyglutamate; Gla-protein, y-carboxyglutamate-containing proteins of the osteocalcin type; MGP, matrix Glaprotein; NCP, noncollagenous protein; pCa, negative logarithm of the calcium ion activity; PG, proteoglycan; PP-H, highly phosphorylated dentin phosphoprotein (phosphophoryn); RER, rough endoplasmic reticulum. 1045-4411/93/$.50 by CRC Press, Inc. © 1993 679 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. (1987b) concentrates on the morphology of dentin and dentinogenesis, whereas a detailed description of tooth morphogenesis and differentiation can be found in the review by Ruch (1987). The most recent review of the structure and calcification of dentin is that of Linde (1992). II. GENERAL ASPECTS The formation of dentin is a highly regulated and well-controlled process, to which several cellular and extracellular constituents contribute. Interaction between various factors takes place at the stage of differentiation and morphogenesis but also occurs when minerals form in the dentin extracellular matrix. The later parts of dentinogenesis, that is, the calcification process, display several similarities with osteogenesis and cementogenesis. In the case of dentin, bone, and cementum formation, a layer of cells form a collagenous organic matrix in which an inorganic calcium phosphate is deposited in the form of mineral crystals. The difference in mode of formation between these mesodermally derived tissues, on one hand, and formation of the ectodermally derived dental enamel, on the other, is great. The enamel is devoid of any fibrous component at all developmental stages, and the proteinaceous matrix during amelogenesis is quite different in its general character from the matrices of dentin, bone, and cementum. The fully formed enamel is a highly crystalline tissue, containing very little organic material; it lacks cellular elements. Compared with the study of osteogenesis, dentinogenesis has several advantages as a model for experimental studies of biomineralization mechanisms. The morphology of the odontoblast/predentin/dentin region is distinct with different cell types well separated from each other. The cells responsible for the formation of the tissue, the odontoblasts, as well as the nonmineralized precursor tissue, the predentin, can be dissected out in virtually pure fractions for analysis (Linde, 1972, 1973). Another difference between dentin and bone is that dentin does not participate in the calcium homeostasis of the organism. In contrast to bone, dentin is normally not remodeled; no resorptive processes normally occur in the tissue. The only occasion when physiological resorption of dentin is seen is in conjunction with the shedding of the deciduous teeth. Because of their intimate functional relationship throughout the life of the tooth, dentin and the dental pulp are often discussed together in textbooks. The two tissues share a common ancestry, the dental papilla, and the development of the two are closely interrelated. In spite of this, there are no direct chemical similarities between the pulp and dentin, and the two tissues are formed by different cells with quite different properties. III. DENTIN STRUCTURE From a phylogenetic point of view, dentin has been regarded as being derived from bone (0rvig, 1967). The most primitive form of dentin, referred to as osteodentin, displays vascular channels with associated laminae, together known as denteons. Cells, being intermediates between osteoblasts and odontoblasts, are totally included in the tissue, and they are not polarized. During further evolution of the tissue, the cells gradually seem to become polarized, although still being totally embedded in the mineralized tissue. Finally, with the evolution to mesodentin or orthodentin, the cells become highly polarized and are not located inside the mineralized tissue; only a cell process extends into the tissue within the dentinal tubules (De Ricql6s, 1979). A. General Composition The mineral content of human dentin is somewhat higher than that of bone. The mineral phase constitutes about 70% on a weight basis, whereas some 20% is organic material. On a volume basis, which perhaps provides a more relevant picture, the mineral and the organic phases account for about 50 and 30%, respectively, the remaining portion being water. The constituents are not evenly distributed in the tissue; the structure of dentin varies quite 680 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. considerably at the microscopic level. As discussed later, there are also reports of considerable compositional differences between dentin in different parts of the tooth. The dentin matrix contains macromolecular constituents that are characteristic components of many connective tissues. The compo- sition of dentin is akin also to other mineralized connective tissues such as bone and cellular cementum. One could ask whether there are characteristics in the chemical composition of dentin and bone that make it possible to distinguish the two, or if the definition of dentin is altogether morphological. As shown later, a congruency in chemical details exists, but there are also distinct differences in that some components are specific for dentin. It is important to realize that differences in structure and chemistry exist between dentin from different species. Because most biochemical studies have been performed on bovine and rat dentin, available data mainly concern these species; comparatively little has been done on the biochemistry of human dentin. In terms of protein composition of the organic matrix, dentin species seem to fall into two main groups, the continuously growing rodent teeth forming one category, whereas teeth with dentinogenesis during a limited period of time, such as bovine, porcine, and human dentin, obviously form another group. Other types of dentin have not been characterized enough to permit any classification yet. Care should thus be taken when attempting to extrapolate findings from one species to the other. B. Types of Dentin Dentin is a complex structure, comprised of several morphologically different types of calcified tissue. Primary dentin constitutes by far the major part of the dentin mass and is produced at a relatively high rate during formation of the tooth. After root formation has progressed as far as to allow the tooth to become functional, the odontoblasts continue to form secondary dentin on the pulpal aspect of the primary dentin, although at a much reduced rate. As a response to varying external stimuli such as chemical irritants, caries, restorative procedures, attrition, or other trauma, tertiary dentin may form later during the life of the tooth. The first layer of primary dentin, which is deposited peripherally in the dental papilla beneath the enamel organ at the onset of dentin formation, is called mantle dentin. The thickness of the mantle dentin varies in humans between 5 and 30 jm; in the rat incisor it is <2 gim. In the root, a similar superficial layer is observed. The rest of the primary dentin is usually referred to as circumpulpal dentin. The circumpulpal dentin may be divided into intertubular dentin and peritubular dentin. The intertubular dentin is the main secretion product of the odontoblasts during dentinogenesis, constituting the largest volume of primary dentin, whereas the peritubular dentin forms a less voluminous sheath of mineralized tissue in the periphery of dentinal tubules. It may be absent in the most pulpal part of the circumpulpal dentin. In the continuously erupting rat incisor as well as in the most peripheral part of human dentin, only intertubular dentin is found. A layer of predentin is located on the pulpal aspect of dentin, between the mineralized tissue and the loose connective tissue of the dental pulp. This is an unmineralized organic matrix, primarily comprising collagen, and without any inclusion of cell bodies. Predentin varies considerably in width, about 10 to 30 jim, depending on species and location. It is present during dentinogenesis but exists also throughout the lifetime of the tooth. C. Morphology of Dentin The shape of the interface between the mineralized dentin and predentin may vary from fairly linear to mineralized globular protrusions. The presence of such calcospherites is believed to reflect the pattern of dentin mineralization. Furthermore, incremental lines may reflect the rhythmicity in dentin formation. Local areas of mineral defects are the interglobular dentin of the coronal circumpulpal dentin and the socalled granular layer of Tomes beneath the root cementum. 681 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Dentin is a highly permeable tissue, both to fluid and molecular flow as well as to microbial invasion, due to numerous dentinal tubules. These radiate from the dental pulp throughout the entire dentin. The tubules are 1 to 3 im in diameter, describe an S-shaped curvature, and are densely packed. A considerable volume of dentin is composed of the lumina of these dentinal tubules; it can be calculated that about 15,000 tubules per square millimeter are present in the outer dentin, 25,000/mm2 in the central part, and as many as 55,000/mm2 near the pulp. The dentinal tubules display lateral branches as well as a terminal branching in the periphery of dentin. The first layer of dentin, the mantle dentin, which forms when dentin forms in the tooth germ, lacks any tubular structures. The organic matrix of mantle dentin is somewhat more irregular in structure and seems to differ in chemical composition, compared to circumpulpal dentin. In histological sections, the mantle dentin can be recognized by its stainability with cationic dyes, such as alcian blue, presumably due to a richness in proteoglycans (PGs). The mantle dentin is produced by polarizing young odontoblasts, which have not yet developed their junctional complexes, but may also contain constituents of other origin. It has also been reported as being slightly less mineralized than circumpulpal dentin. In the main portion of the primary dentin, the intertubular dentin, a fibrous network of collagen with deposited mineral crystals, can be seen in the electron microscope (Figures 1-4, 6). Plate-shaped crystals, 2 to 3 nm thick and 60 nm long, are located either along the subjacent collagen fibers or in the spaces between these fibers. High-resolution technique indicates that these crystals are formed by two flat parallelepipedic plates. Although the fully grown crystallites at the surface of the collagen fibers, which are 100 to 120 nm in diameter, hide the structure of the organic polymer, its periodic banding is known to be related to the organization of the mineral phase. The dense packing within dentin makes it unfeasible to establish clearly whether crystals are also located inside the collagen fibers. More than likely, however, crystallites are both located at the fiber surface and inside the holes or vacant spaces between the subunits of the fibers, as in many other mineralized collagen-rich tissues. Histochemical staining of sections with cationic dyes gives evi- dence for the presence of polyanionic components in close association with the periodic striations of the collagen fibers (Figure 6). In human teeth, the peritubular dentin forms a well demarcated and highly mineralized sheath around the dentinal tubule, 0.5 to 1 pLm in thickness (Figures 1-3). In reality, peritubular dentin is intratubular, in that it is deposited on the inner surface of the lumen of larger tubules by the odontoblasts subsequent to the formation of intertubular dentin. In humans, peritubular dentin starts to develop inside dentinal tubules at some distance away from the pulp chamber. Peritubular dentin is considered to be absent from rodent teeth. In a few unusual cases, as in elephant and opossum molars, peritubular dentin formation precedes intertubular dentin formation. Ultrastructurally, the peritubular dentin has a more compact appearance than the intertubular dentin. High-resolution technique has revealed the crystals to be parallelepipedic structures with the dimensions 36 x 26 x 10 nm (Schroeder and Frank, 1985). The mineral phase seems to comprise mainly hydroxyapatite but is also rich in magnesium and contains carbonate, accounting for its high solubility. The organic matrix of the peritubular dentin is virtually devoid of fibrous structures, adding to the difficulties to study peritubular dentin after demineralization. However, preparations using nonaqueous demineralization technique have allowed the visualization of carbohydrate-rich components (Goldberg et al., 1980). The formation of peritubular dentin is a continuous process, but its mode of formation is still unknown. A major reason is that no good experimental model has been identified for such studies. Thus, it is not known whether peritubular dentin is the result of an active secretion from the odontoblast processes into the lumen of the tubules or if a more passive deposition of matrix components occurs. It is clear, however, that peritubular dentin formation occurs only in the presence of a viable odontoblast process and differs from sclerotic processes in so-called dead tracks. Compared with primary dentin, secondary dentin is deposited at a much slower rate, essentially during the whole life of the tooth. Its structure varies from the primary dentin mainly by the accentuated curvature of its dentinal tubules. In roots of mature human teeth, the mesial and distal 682 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 1. Peritubular dentin (PT) around the lumen (LU) of a dentinal tubule and intertubular dentin (IT) in a human tooth. No staining. (x54,000.) FIGURE 2. Peritubular dentin (PT) along the dentinal tubule lumen and intertubular dentin (IT) in a young human premolar. No staining. (x82,000.) FIGURE 3. Postembedding demineralization allows preservation and staining of GAGs as a dense network in the peritubular dentin (PT). In the intertubular dentin (IT), GAGs appear as round dots, 15-20 nm in diameter, after staining with alcian blue. (x82,000.) FIGURE 4. Intertubular human dentin demineralized with EDTA. The sections were stained with uranyl acetate and lead citrate. Thick collagen fibers are clearly seen. The spaces located between the collagen fibers are empty. (x82,000.) FIGURE 5. Intratubular mineralization (ITM) due to sclerotic processes. The lumen of the tubule will eventually be filled with such crystals. Peritubular dentin (PT) differs in appearance from such mineralizations. (x54,000.) FIGURE 6. Human dentin, fixed, demineralized with EDTA, and further processed for electron microscopy. The sections were stained with alcian blue and counterstained with uranyl acetate and lead citrate. (x54,000.) 683 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. walls of the pulp are covered with a sclerotic dentin, deprived of dentinal tubules. Occlusal attrition, trauma, carious decay, dental restorations, and in general any external nox- may induce the formation of tertiary dentin, by some referred to as reparative dentin or irritation dentin. It varies widely in shape and appearance. In some cases, proximal to ious stimulus a so-called calciotraumatic line, a true tubular orthodentin is formed. In other cases, an atubular dentinal structure or osteodentin is formed at the pulp wall. Tertiary dentin is formed beneath those dentinal tubules directly affected by the stimulus, thus being localized to certain areas of the pulp/ dentin interface. The origin of the cells forming tertiary dentin is still a matter of investigation. It seems that both an activation of already-existing odontoblasts as well as a differentiation of new odontoblasts may be involved, depending on the character of the stimulus. Intratubular mineral deposits, except for the peritubular dentin, do not constitute true dentin components (Figure 5). They may result from several causes, one being re-precipitation occurring inside the lumen after dissolution of calcium and phosphate in conjunction with caries. Various apatitic and nonapatitic forms of mineral are found in such structures. D. Cells Dentin is formed by odontoblasts, which line the pulpal aspect of dentin, or rather predentin, throughout the life of the tooth (Figures 7, 8). These cells are responsible for the formation of dentin and the production of its constituents. Odontoblasts thus line the surface of the tissue they have formed in the same way as the osteoblasts in bone. During active dentinogenesis, the odontoblasts are columnar and reveal the characteristics of actively synthesizing and secreting cells. At later functional stages, they take on a more quiescent appearance with a lowered height and a much reduced density of organelles. Whereas bone contains cells, osteocytes, embedded in the tissue, dentin is cellular only in the sense that the odontoblasts have processes penetrating the predentin and dentin (Figures 7, 9, 10). The odontoblast process lacks major organelles but contains an abundance of longitudi- nally arranged filaments and microtubules. Thin lateral branchings of the odontoblast processes, crossing the dense peritubular dentin, allow lat- eral connections between different odontoblast processes and supposedly play a role in the nutrition of dentin. Because each dentinal tubule, in principle, is occupied by one process and thus is related to one individual cell, the density of tubules in dentin gives an idea of the number of odontoblasts present at the periphery of the pulp in young teeth. The extent of the odontoblast processes in the dentinal tubules in the mature tooth is a matter of debate. Several investigators are of the opinion that the odontoblast processes in human, cat, and rat teeth are limited to the pulpal half of dentin (Garant, 1972; Holland, 1976; Thomas, 1983). In contrast, immunocytochemical studies have given evidence that tubulin-, vimentin-, and actin-containing structures extend to the dentino-enamel junction (Sigal et al., 1984a,b). Whether intact odontoblast processes are really present throughout the tubules or only cytoskeletal remnants is still open for discussion. It has been pointed out that the process is difficult to fix properly and retracts easily, if special care is not taken (Lafleche et al., 1985). There is evidence for the presence of odontoblast processes also in the outer third of human root dentin (Frank and Steuer, 1988). Aging reduces the number of odontoblasts as the pulp gradually retracts. Proximal to the layer of odontoblast cell bodies, smaller less differentiated cells, the Hohl cells, are present. This area is rich in terminal innervation, and there is also a well-developed capillary vascular network; some capillaries even penetrate between the cell bodies in the odontoblast layer. IV. MORPHOLOGY OF DENTINOGENESIS A. Differentiation of Odontoblasts Formation of the tooth begins with a series of ectodermal-mesenchymal interactions that are instrumental in morphogenesis and differentiation. Reciprocal cell and matrix interactions are believed to play a key role for the terminal differentiation of odontoblasts and ameloblasts. As these early stages fall outside the scope of this review, 684 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 7. Semithin section of rat incisor. The pulp (P) is seen on the right, containing fibroblasts and blood vessels. The subodontoblastic Hohl cells (HC) are located beneath the odontoblasts (0). Capillaries (CP) may penetrate between the odontoblast cell bodies. Odontoblast processes extend into the predentin (PD) and the dentin (D). This section was used for autoradiographic investigation 4 h after [3H]serine injection. Labeling is seen both in predentin and at the mineralization front at the dentin/predentin interface. Toluidine blue. (x1000.) FIGURE 8. Odontoblasts (0) and predentin (PD) in rat incisor observed after ruthenium hexamine trichloride fixation and staining. Junctional complexes (arrowheads) form a tight seal, separating the cell body compartment from the predentin zone (x8100.) FIGURE 9. Ruthenium hexamin trichloride staining of rat incisor. Small electron-dense dots are seen in predentin in close association with collagen fibers. These aggregates visualize the GAGs located in this area. In this undemineralized section, the electron density of the mineralized dentin (D) does not allow observation of any further details. Odontoblast processes (0) penetrate the dentin inside tubules. (x67,500.) FIGURE 10. Thin section of the odontoblast, predentin, and dentin area of rat molar, conventionally fixed. An unmyelinated neuronal axon (arrow head) is seen in predentin near the dentin edge, in close association with an odontoblast process. (x6000.) 685 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. the reader is referred to Ruch (1987, 1990) and Slavkin (1988) for a detailed treatise of this interesting and dynamic area. Here, only some of the morphological features, seen at the onset of dentinogenesis, will be highlighted. A basement membrane (BM) is located between the epithelial cells and the dental papilla throughout the differentiation of the tooth bud, before dentin and enamel formation (Figure 13). At early stages, the dental papilla cells differentiate and become preodontoblasts, which are located randomly in relation to the BM. When the preodontoblasts withdraw from the mitotic cycle, they become polarizing odontoblasts (Figures 11, 12, 14). These group and establish a palisade-like structure (Figure 15), involving the formation of intercellular junctions. They finally acquire their terminal polarization and become dentinogenically active, secretory cells. Like other BMs, the dental BM consists of collagen type IV, associated with noncollagenous molecules such as laminin, fibronectin, and heparan sulfate PG. Basement membranes are three-layered structures, but even after rapid freezing and freeze substitution, it has been difficult to demonstrate a three-laminated structure in the inner dental BM (Goldberg and EscaigHaye, 1986). During the course of development, holes or discontinuities appear in the BM, allowing heterotypic cell contacts between presecretory ameloblasts and polarizing odontoblasts. The thickness of the BM increases from the early stages to the time when a distinct predentin is apparent. After this, the overall thickness decreases, and gradually the BM can no longer be distinguished. These BM modifications are believed to be causally related to successive steps in odontogenesis. The terminal differentiation of odontoblasts, as well as ameloblasts, seems to be dependent on specific cell membrane/cytoskeleton interactions with BM constituents and other extracellular components (Ruch, 1987, 1990). At the end of the mitotic cycle, preodontoblasts appear as rounded or oval cells with a few protrusions. They display minute contacts, either between these lateral or apical processes or between the protrusions and the bodies of adjacent cells. Many cells are binucleate, suggesting that they are at the end of mitosis. At this stage, the cells reveal little polarity. The nucleus to cytoplasmic ratio is high, and each nucleus contains one or two nucleoli. In the cytoplasm, numerous free ribosomes can be seen, often grouped in a rosette form. The rough endoplasmic reticulum (RER) is scarce with short cisternae and only at the beginning of its development. The Golgi apparatus is already formed but at an initial functional stage. Few lysosomal structures are observed. Between the preodontoblasts or early polarizing odontoblasts, the extracellular matrix is comprised of large collagen fibers, 70 to 100 nm in diameter, displaying a regular distribution of periodic striations and thinner fibrils, altogether forming a network (Figures 15, 16). Close to the BM, fibrils appear at right angles to the BM, giving rise to a general orientation in relation to the BM/predentin junction (Figure 16). In the polarizing odontoblast, the RER develops actively, and free ribosomes are numerous in the cytosol as well. A few electron-lucent vesicles and electron-dense lysosome-like structures can be seen. With the development of the cytoplasm, the nucleus-to-cytoplasmic ratio decreases. Polarizing odontoblasts are elongated cells with well-developed centrioles, basal bodies, and prominent ciliary structures. During differentiation of the odontoblasts, one of the centriole pair is ciliated (Sasano and Kagayama, 1987). The cilia (Figure 17) are also seen in mature odontoblasts but to a lesser extent. Intercellular junctions are formed at this stage at some distance from the site where extracellular material has accumulated between the cell processes. This early predentin prevents the odontoblasts to establish intercellular contacts in their distal third. Minute contacts and development of gap- and desmosome-like junctions are determinants in the establishment of the different characteristics of the apical and baso-lateral domains. Microfilaments reinforce the formation of the junctional complexes; vimentin has been shown to accumulate at the apical pole during differentiation (Lesot et al., 1982). Actin can be demonstrated in the terminal web of young odontoblasts, its stainability becoming gradually more intense as the thickness of the predentin-dentin increases (Nishikawa and Kitamura, 1986). At the end of odontoblast po- Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. a' I~I--- ~ I~*- -3 Cllx-r ~ - - I ··--r··u-···XI· FIGURE 11. Early stage of dentinogenesis. Preodontoblasts are still dividing in this area, as seen from the presence of a mitotic cell (arrow). A basement membrane separates the presecretory ameloblasts (A) from the dental papilla. Saponin-permeabilized, tannic acid-glutaraldehyde fixed. (x10,800.) FIGURE 12. Dividing preodontoblast at the end of the mitosis. Saponin-tannic acid/glutaraldehyde. (x13,500.) FIGURE 13. Early stage of predentin formation. Collagen fibers, decorated with fibrils and electron-dense dots, are deposited in this area (arrow). A relationship is seen between this material and the lamina fibro-reticularis of the basement membrane (BM), separating the connective tissue from the epithelial presecretory ameloblasts (A). Saponin-tannic acid/glutaraldehyde. (x54,000.) larization, the junctional complexes have caused the establishment of individual tissue compartments, in the sense that the predentin extracellu- space has become isolated from that of the pulp. Intercellular diffusion could be expected to be hampered by this barrier. lar 687 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 14. Early stage of postmitotic polarizing odontoblast (0). These cells are not linked with junctions. A first hint of predentin is seen, containing few fibers and associated fibrils and granules. A basement membrane (arrow) separates the presecretory ameloblasts (A) from the dental papilla. Saponin-tannic acid/glutaraldehyde. (x6700.) FIGURE 15. Polarizing odontoblasts (0) are now at right angles with the basement membrane (arrow). Intercellular collagen fibers, the so-called von Korff fibers (arrowheads), are present between elongating cells, which have established minute contacts but are not yet at the stage of junctional complex formation. The extracellular material in the early predentin is increasing in density. Saponin-tannic acid/glutaraldehyde. (x13,500.) FIGURE 16. Detail of Figure 15 (lower part on the right). (0) odontoblast, (A) presecretory ameloblast. Predentin collagen fibers are inserted at right angles with the lamina fibro-reticularis of the basement membrane. This extracellular material and intercellular collagen fibers (arrowhead) may be implied in cell shaping. Saponin-tannic acid/glutaraldehyde. (x27,000.) FIGURE 17. Microtubules (arrowheads) in the odontoblast cytoplasm play a role in the polarization of the cells. Microtubules are also grouped in order to form a cilium (C). Saponin-tannic acid/glutaraldehyde. (x67,500.) 688 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. B. Initial Dentin Formation Young dentinogenic odontoblasts are actively involved in the synthesis, secretion, and reabsorption of matrix components. Whereas synthesis occurs in the cell bodies, exocytotic and endocytotic activities take place primarily in the cell processes. These processes form complex structures with a main trunk, about 0.5 to 1 pm in diameter, which quite often may divide into two or more considerably thinner lateral branchings (0.1-0.2 pm) (Goldberg and Escaig, 1981). During the initial formation of the nonmineralized mantle dentin matrix, a number of concomitant events take place. The young, newly differentiated odontoblasts secrete collagen and noncollagenous components (Figure 18). Bundles of large collagen fibers are grouped in globular structures, and, because of the addition of new fibers, the collagen becomes packed more tightly. Before the onset of mineral formation, large amorphous plaques of osmiophilic material are seen at the enamel-dentin junction. This disappears at more mature stages. On the dentin side of the enamel-dentin junction in the rat incisor, there is special affinity for cationic dyes and phospholipid-staining reagents (Lorm6e et al. (1989). Relatively few investigators have attempted to specifically elucidate the constituents of mantle dentin matrix. Immunostaining for collagen type I shows that it is present in rodent mantle dentin (Lesot et al., 1981); also collagen type Vis present at an early stage (Bronckers et al., 1986). Fibronectin can be demonstrated in the organic matrix of the forming mantle dentin of rat incisor but is absent within intertubular dentin after mineralization (Linde etal., 1982; Connor etal., 1984). In rat teeth, Gla proteins of the osteocalcin type can be demonstrated by immunostaining in newly differentiated odontoblasts as well as in predentin before initiation of mineralization (Bronckers etal., 1987; Gorter de Vries etal., 1987); the staining in mantle dentin is reportedly restricted to mineralized globular patches (Camarda et al., 1987). Also dentin phosphoprotein seems to be absent from mantle dentin (Nakamura et al., 1985; Takagi and Sasaki, 1986; Rahima et al., 1988). Enamel proteins are expressed in presecretory ameloblasts at an early stage but have also been reported to be present to some extent in the forming mantle dentin (Slavkin et al., 1988). Although the presence of amelogenin mRNA in presecretory ameloblasts was ascertained by Shimokawa et al. (1991), labeling of the forming mantle dentin with antibodies to amelogenins could not be con- firmed. Electron microscope findings indicate that, instead of occurring in the extracellular matrix, the initial crystal formation takes place inside 0.10.2-pm membrane-bound so-called matrix vesicles, located in the mantle dentin matrix (Figure 19) (Bernard, 1972; Eisenman and Glick, 1972; Katchburian, 1973; Bonucci, 1984). Only after this stage (Figure 20) can mineral be seen in association with the collagen fibers in the tissue. No vesicles are seen at later stages, that is, during circumpulpal dentin formation. Such vesicles are believed to constitute the site of initial mineralization in other calcified tissues. Whereas matrix vesicles have been observed in rather few numbers during mantle dentin mineralization, as is the case during initial bone formation, similar vesicles occur in large quantities during cartilage mineralization. Although the mechanism for mineral formation in vesicles is not clear, it may be presumed that they provide a protected micro-environment for crystallization and that they function as vehicles for nucleating macromolecules, enzymes, or other essential components (Bonucci, 1984). Studies of cartilage and bone have established that the vesicles take their origin by budding off from the cell plasma membrane; during mantle dentin formation, they are presumably derived from the odontoblasts. Among their characteristics, they display a high activity of alkaline phosphatase, are constituted of phospholipids, and contain annexin II (Wuthier, 1988). The importance of matrix vesicles for the onset of mineralization is still a matter of dispute. Loci of crystal initiation appear and also develop in dentin at some distance from the matrix vesicles. Some confusion may arise between the matrix vesicles and lateral branchings of odontoblast processes owing to similar appearance when sectioned transversely, thus providing a heterogeneous population of structures seen in the electron microscope that may or may not be matrix vesicles. 689 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 18. Collagen fibers accumulate in the forming predentin. Thick fibers are grouped, forming nodules. Thinner fibrils are also seen at right angles with the basement membrane, which separates presecretory ameloblasts (A) from the dental pulp. Ruthenium hexamine trichloride-glutaraldehyde. (x21,600.) FIGURE 19. Eventually, the basement membrane disappears. Remnants of the BM; thin fibrils; thicker collagen fibers, where periodic striation is clearly seen; GAGs stained positively by the cationic ruthenium dye; and matrix vesicles (MV) can be seen in the forming unmineralized predentin. At this stage of early dentin formation, crystals appear first inside matrix vesicles (arrows) and then protrude from the vesicles. (A) presecretory ameloblasts. Ruthenium hexamine trichlorideglutaraldehyde. (x108,000.) FIGURE 20. Predentin (PD) and the first layer of mineralized dentin (D). Odontoblasts (0) are now polarized cells, and their cell bodies are connected to each other by junctional complexes. Odontoblast processes are seen in the predentin zone, and a thin layer of mineralizing mantle dentin has been formed by an accumulation of calcospherites (C). Ameloblasts (A) are now initiating enamel secretion. Ruthenium hexamine trichloride-glutaraldehyde. (x5400.) 690 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. For a number of years, the existence of interodontoblastic fibers during the early phase of dentin formation, the von Korfffibers, has been a matter of debate. These are argyrophilic corkscrew-like fibers, located between the cell bodies, fanning out into the predentin and penetrating the dentin, where they are embedded together with the collagen fibers secreted by the odontoblasts. A reappraisal of the nature of these fibers, earlier identified only on the basis of silver staining at the light microscopy level, was made by Ten Cate (1978). No corresponding fibrous structures could be demonstrated by electron microscopy, an opinion supported by Magloire et al. (1982), who were unable to demonstrate any immunolabeling using anticollagen antibodies. Other studies, however, found evidence for the presence of collagen fibers in the odontoblast intercellular spaces (Figure 15), but failed to observe any continuity between pulpal fibers and those in dentin (Goldberg and Escaig, 1981; Szabo et al., 1985). Such a continuity has, however, been found by others (Wigglesworth etal., 1986), and evidence for a continuity between the intercellular bundles and the fibers of predentin has been noted (Bishop etal., 1991; Salomon et al., 1991). Scanning electron microscope investigations have indicated that the von Korff fibers may be found only in the root of the rat molar (Salomon et al., 1991). The mantle dentin, resulting from this initial phase of dentin formation, does not contain dentinal tubules, only sometimes thin canaliculi. Mineralized globular structures, about 2 gtm in diameter and either isolated or in groups, can be seen embedded in a network of interglobular dentin in the crown mantle dentin. The equivalent is also found in the root; the Tomes granular layer in the root dentin displays thin canaliculi and poorly fused globules, rendering to the tissue its granular appearance. C. Formation of Circumpulpal Dentin Circumpulpal dentinogenesis, like osteogenesis, occurs by two simultaneous processes: formation of the organic predentin and its subsequent mineralization at the mineralization front (Chart 1). From a functional point of view, circumpulpal dentinogenesis may schematically be viewed as comprising three compartments associated with each other. One is the cell compartment, comprised of cell bodies and cell processes; the others are the predentin and the mineralized dentin. In the odontoblast cell bodies, biosynthesis occurs. Exocytosis and endocytosis are the major events occurring in the cell processes, penetrating the predentin and further into the dentin tubules. The second, the predentin, compartment is composed of a 15- to 20-im-wide extracellular matrix layer, where the collagen molecules aggregate to form collagen fibrils. Predentin is thus a zone for the formation and maturation of the collagen meshwork of the dentinal matrix; the width of predentin represents the time allotted for arranging collagen molecules in the form of a fibrous web. The composition of predentin differs from that of dentin matrix, in that some components are added at or just in advance of the mineralization front, whereas other predentin constituents may be metabolized (Chart 1). The secretory activity of the odontoblasts regulates the formation of circumpulpal dentin. The rate of deposition of dentin is related to several factors. The cells are more active when they are young and less during aging. The functional stage of the tooth is also of importance. Finally, the anatomical location in the tooth, in the coronal or radicular portion, is also of importance. In the rat incisor, dentinogenic odontoblasts actively secrete predentin and dentin then become old cells and degenerate in about 2 months. In noncontinuously growing teeth, odontoblasts have a high activity when they are young, which decreases gradually but is still present at later stages. Deposition of circumpulpal dentin in the crowns of human molars occurs at a higher rate in the furcation region than at the occlusal aspect. A slower apposition occurs on the lateral walls of the pulp chamber. In rat molars, in contrast, labeled precursors were found to be incorporated at a higher extent in lateral areas of the dentin than in the occlusal part, the lowest incorporation being detected in the furcation area (Salomon et al., 1990). In human dentin, incremental lines, referred to as von Ebner lines, are separated by approximately 20 gam. The average rate of growth of human and monkey dentin is some 4 gm/d, whereas it is in the range of 16 pm/d or higher for small animals. Dentin located between two incremental lines thus needs several days to be depos- 691 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. ited. The presence of intermediate extra lines, each 4 Am, may be explained by results demonstrating daily rhythms of deposition (Kawasaki etal., 1980). From studies using fluorochromes such as tetracycline, it is obvious that the secretory activity of the odontoblast varies considerably during its life cycle as well as with the type of tooth and the position within the tooth (Kawasaki et al., 1977). D. Predentin and the Mineralization Front Predentin is a constant feature in sound teeth. It has a rather constant thickness, about 15 to 20 pm, limited at its proximal side by the cell bodies of the odontoblasts, firmly connected by their distal junctional complexes, and at its distal side by the mineralization front. Thus, predentin is organized as a more or less sealed compartment. Collagen is exocytosed near the cell bodies and becomes associated in fibrils, on which interand intramolecular stabilizing cross-links form. Mineral formation in the collagenous web occurs at the advancing mineralization front, which moves forward at a rate varying between 4 and 20 gm or more per day, depending on species. In addition, PGs constitute the amorphous ground substance of the predentin. The functions of phospholipids, also detected in this area, are not known. DENTIN r4- t I .- PP-H,PG MINERALIZATION FRONT . Gla-protein ----------_ t . I I C :R I rr I Metabolism of PG t PREDENTIN ZONE Formation of collagen network Secreionof coenand -t t ODONTOBLASTS CHART 1. Schematic drawing of the odontoblast-predentin region during dentinogenesis (Linde, 1989). Macromolecules are synthesized within the odontoblast cell bodies. The constituents of predentin, primarily collagen and proteoglycan (PG), are secreted close to the odontoblast cell bodies (lower set of arrows) to form the extracellular, nonmineralized predentin matrix; some of this PG is metabolized in predentin. Uptake into the cells of degradation products from such metabolism in predentin seems to occur along the odontoblast process membrane (arrows). Several noncollagenous components are transported within the odontoblast process and secreted at the mineralization front, i.e., in predentin, just before the advancing mineralized dentin, where mineral formation is initiated. Among these are the highly phosphorylated phosphoprotein (PP-H), y-carboxyglutamatecontaining proteins of the osteocalcin type (Gla-protein), as well as a second pool of PG. 692 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. The main trunk of odontoblast processes and lateral branches thereof penetrate the predentin zone. Along the plasma membrane of the processes, a pericellular area, about 150 nm in width, has specific properties. In this area, collagen and PGs are secreted. Endocytotic events also occur, leading to the formation of coated pits, and consequently of coated vesicles. In the predentin itself, collagen fibril formation is seen. In the area adjacent to the cell bodies, thin fibrils, 20 nm in diameter, can be observed. In the central predentin, the collagen fibers are 40 nm thick, while becoming 55 to 75 nm in the distal part near the mineralization front. This clearly demonstrates that lateral aggregation of new collagen occurs in predentin, leading to the formation of thick fibers that in turn mineralize. The space between the collagen fibers is gradually reduced, primarily because ofthe increased diameter of the fibers. In the rat incisor, the collagen fibers are grossly parallel to the mineralization front (Figures 36-39). In the molar, fan-shaped columns of longitudinally sectioned collagen fibers between dentin and odontoblasts can also be seen. The order that exists in the predentin is regulated by the functional activity ofthe odontoblasts, which becomes evident when using microtubule inhibitors such as vinca alkaloids (Goldberg et al., 1987b) (Figure 40). In humans, only intertubular dentin is found close to the mineralization front. The formation of peritubular dentin starts at some distance away from the predentin/dentin junction. Only in two species, elephant and opossum (Figure 22), is peritubular dentin found at the mineralization front protruding into the predentin (Boyde and Jones, 1972; Acevedo and Goldberg, 1987). The appearance of the mineralization front varies, depending on which part of the tooth is observed as well as the activity of dentin formation. Globular structures are present at the onset of dentin formation, when dentinogenesis occurs at the highest rate. When dentin formation becomes slower, there is a decrease in calcospherite formation, and these structures are hardly seen when dentin apposition is less active. The formation of globular structures seems to be independent of the distribution of the odontoblast processes. The formation of FIGURE 21. Scanning electron microscopic view of the mineralization front in the rat incisor. Globular structures, 15-25 gm in diameter, are seen. Clearly, these formations are unrelated to the distribution of the odontoblast processes. Only intertubular dentin is present in rat dentin. (x540.) FIGURE 22. Scanning electron microscopic view of the mineralization front in the molar of opossum. Calcospherites, 15-20 jlm in diameter, are seen. This represents one of the very few species in which peritubular dentin formation occurs before the formation of intertubular dentin. (x 1 00.) 693 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. peritubular dentin is unrelated to such configurations. In the continuously growing rat incisor, which contains only intertubular dentin, the appearance of the dentin surface also varies. Close to the apex of the tooth, numerous calcospherites, about 15 to 25 pm in diameter, are seen in the labial and lingual portions (Figure 21); only flat surfaces appear in the lateral portions of the incisor. Globular dentin is observed mainly during early dentin formation. The assumption that globules may be because of the concentric formation of dentin around matrix vesicles is incorrect. Ultrastructural observations of collagen organization in predentin and dentin have shown that there are whorls of collagen fibers, organized in a concentric pattern, in predentin as well as in dentin in the absence of initial matrix vesicles. Such whorls are known to occur spontaneously even in the absence of cells and are believed to be because of self-organizing properties of the collagen fibers (Bouligand et al., 1985). Interglobular spaces, which constitute defects in dentin where nonmineralized residual material accumulates, are often enhanced in pathologic dentin. In X-linked hypophosphatemic vitamin D-resistant rickets, for example, there is a lack of fusion between the calcospherites. This leads to the formation of an outer dentin layer, similar in some ways to the granular Tomes layer but considerably thicker and present mainly in the outer third of the tooth (Shellis, 1983). X-ray diffraction analysis shows that larger crystals are present, probably because of the presence of unmineralized spaces around the crystal nuclei (Abe et al., 1991). E. Root Dentin Formation At the end of crown formation, and eventually related to the onset of tooth eruption, root dentin formation begins. From the junction between the inner and the outer enamel epithelia in the enamel organ, at the cervical loop, Hertwig's epithelial root sheath starts to develop. This is a two-layered structure, consisting of an inner and an outer epithelial layer, which initiates the differentiation of odontoblasts at the periphery of the dental papilla. These cells become polarized and start to secrete a predentin, which in turn will give rise to dentin. The predentin is considerably more narrow in the root than in the crown. The collagen fibers are in their general orientation parallel to the mineralization front and to the long axis of the developing root. Differences in the incorporation of precursors as well as differences in composition in some matrix components also reflect substantial and important variations between crown and root dentin. In rodent incisors, the labial portion of dentin, which is covered with enamel, has been regarded as analogous to crown dentin in noncontinuously growing teeth, whereas the lingual and lateral portions, covered with cementum, are regarded as a root analogue. Thus, Beertsen and colleagues found differences in reactivity to bisphosphonate, in amino acid incorporation, as well as differences with respect to phosphoprotein distribution (Beertsen et al., 1985; Beertsen and Niehof, 1986; Steinfort et al., 1989). In bovine incisors, phosphoproteins, present in the crown, could not be detected in the root (Takagi et al., 1988). Another example of the differences between crown and root dentin is the radioactive labeling with [3H]proline in rat teeth. In the molar, little conversion of the predentin into dentin was found in the root during 24 h, whereas most of the precursor was found within dentin during the same period in the crown (Salomon and Goldberg, unpub- lished). granular Tomes layer is located in the periphery of the root dentin. During root dentinogenesis, the Hopewell-Smith hyaline layer, believed to be a derivative of the epithelial root sheath (Lindskog and Hammarstr6m, 1982), is deposited at the dentin/cementum interface. Dislocation of Hertwig's sheath allows the deposiThe tion of cementum at the surface of the root dentin, initially as an afibrillar, then as a fibrillar, acellular structure. Later, when at least half of the root has formed, cellular cementum is formed. F. Vascularization and Innervation Blood vessels are not present in dentin. Arteries, penetrating through the apex of the tooth, migrate coronally in the center of the pulp and 694 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. divide into terminal arteries that lead to a network of capillaries beneath the odontoblast layer. Capillaries extending between the odontoblast cell bodies are a frequent observation. Regulation of pulpal hemodynamics is under the control of sympathic nerves. The pulp capillaries are fenestrated, which is of importance for the diffusion of precursors and other nutrients in the proximity of the baso-lateral plasma membrane of the odontoblasts. The existence of nervous structures in the subodontoblastic region, the so-called Raschkow's plexus, was first established by the light microscope on the basis of silver staining. A small number of unmyelinated axons may pass between the odontoblasts, extending a short distance into dentinal tubuli, some 100 to 150 gm (Byers, 1980; Byers et al., 1987; Ramieri etal., 1990). Inside the tubuli they may be seen in close association with odontoblast processes (Frank, 1968; Tsukada, 1987). It has been suggested that odontoblast activities might be regulated by sympathic nerve fibers both during development as well as at later stages (Avery et al., 1984; Chiego et al., 1987). In contrast, others have denied any influence of the trigeminal innervation during dental development (Lumsden and Buchanan, 1986). For example, no difference was found in the dentin uptake of [3H]proline after sympathectomy in rats (Herskovits and Singh, 1986). G. Aging and Replacement of the Odontoblasts The life cycle of the human odontoblasts is not well known. In the rat, in contrast, a good description of differentiation, maturation, and aging has been given (Takuma and Nagai, 1971; Sasaki et al., 1982a). On aging, the odontoblasts gradually become filled with lysosomal vacuoles and degenerate. In the rat incisor, the rate of odontoblast translocation in the incisal direction has been reported to be about 0.6 mm/d (Warshawsky, 1979). This implies that in these teeth odontoblasts differentiate, are active secretory cells, and degenerate within 30 to 40 d. Human odontoblasts have been calculated to have a period of dentinogenic activity of approximately 750 d. During this period, about 4 gm of dentin is deposited per day. After this, the odontoblasts constitute a nonrenewing static cell population, which in principle survives as long as the tooth. Changes in the shape of the nucleolus as well as degenerative changes have been reported in the less active, aging odontoblast (Couve, 1986). When reparative processes are induced experimentally, there is some evidence that new odontoblasts participate (Nakashima, 1990). Also in grafted coronal tissue of molar tooth germs, generation of new odontoblasts occurs (H6ritier et al., 1989), which demonstrates the potential for cell differentiation into new odontoblasts. Two groups of cells are candidates. First are the Hohl cells, subjacent to the odontoblasts. Second, undifferentiated pulp cells may migrate through the pulp parenchyma and turn into new odontoblastlike cells (Fitzgerald et al., 1990). The question as to the mechanisms regulating this differentiation still remains open. V. THE DENTINOGENICALLY ACTIVE ODONTOBLAST A. Ultrastructure of the Cell Dentinogenically active odontoblasts are tall polarized cells, the bodies of which are grouped at the periphery of the pulp, having long processes that are inserted into the dentinal tubules. Depending on the plane of section and the stage of development, one single layer of cell bodies is seen, or they constitute a multilayered structure. Fenestrated capillaries are present adjacent to the cell bodies, mainly in the subodontoblastic layer near the basal plasma membrane of the odontoblasts, but also in the intercellular spaces. The cell bodies are 20 to 40 gm tall and 3 to 5 gm wide. Schematically, three zones can be recognized. The proximal or basal zone contains mitochondria near the basal plasma membrane, a nucleus whose nucleoli vary in shape with the cellular synthetic activity, and lateral parts where a well-developed RER is seen. The central zone (Figure 23) displays in the supranuclear area of its central part a prominent Golgi apparatus with large rounded cistemae and many vesicles at its periphery (Figure 25). This is observed together 695 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 23. Odontoblast cell body. Nucleus (N), rough endoplasmic reticulum (RER), the Golgi apparatus (GO) and mitochondria (M) are seen in the cell. (x40,000.) FIGURE 24. Odontoblasts are ciliated (C) cells even at a mature 25. Detail of the Golgi stage. Bundles of microfilaments can also be seen in the cytoplasm. (x54,000.)andFIGURE multivesiculated bodies (MVB) apparatus area (GO). Vesicles, containing abacus-like structures (arrowhead), are present nearby. (x54,000.) FIGURE 26. After rapid-freezing and freeze-substitution fixation, the mitochondria (M) in the distal cell bodies of odontoblasts contain numerous electron-dense mineral inclusions. The electron-lucent areas correspond to material either released or lost during sectioning. (PD) predentin. (x40,000.) with a population of secretory vesicles, lysosomes, and multivesicular bodies. Again, in the lateral portion of the central zone, rows of RER cisternae are present. The distal zone can be subdivided into a secretory vesicle-rich RER-filled area and a distal part, where the RER is abruptly inter- 696 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. erupted. Numerous mitochondria are present at the junction between the cell body and the process (Figures 26, 28). Large lysosome-like electron- lucent vacuoles can be seen, loaded with membraneous residues. The cell bodies are connected to one another in the distal part, close to the predentin, by junctional complexes (Goldberg et al., 1981) (Figures 27-29, 31), which isolate the cell bodies distinctly from predentin and form a stable barrier (Boyde et al., 1978) (Figures 28, 30). A few tight junctions have been recognized in old cells in the rat incisor as well as in human and Kitten teeth (Sasaki et al., 1982b; Iguchi et al., 1984; Calle, 1985). However, these tight junctions are arranged to form small maculae or faciae occludentes, rather than beltlike zonulae. The question about the permeability ofthe interodontoblastic junctional complexes is still not fully resolved. Some researchers have given evidence for the penetration of different tracer molecules such as [I25]albumin (Kinoshita, 1979), La3+ ions (Tanaka, 1980), horseradish peroxidase (Sasaki et al., 1984), and fatty acid bound to albumin (Goldberg etal., 1989) into predentin and dentin, whereas others are of the opinion that these junctions provide a tight seal between the odontoblasts (Jones and Boyde, 1984; Bishop, 1985). These discrepancies may at least in part be explained on methodological grounds, but the functional stage of the odontoblasts also is clearly of importance. The transport pathways for ions into dentin mineral is of special interest in this respect (Section VI.D). The presence of cytoskeletal proteins in the odontoblast cell bodies has been ascertained (Nishikawa and Kitamura, 1987). This was confirmed by immunocytochemical visualization of a-tubulin, present throughout the cytoplasm (Sawada et al., 1989), whereas myosin, a-actinin, and tropomyosin were detected mainly at the distal end of the cell bodies (Nishikawa and Sasa, 1989). Anti-actin immuno-gold labeling (Figure 31) has been demonstrated in the cytosol near the RER and beneath the plasma membrane, around the mitochondria and vesicles (Goldberg et al., 1991). Microtubules are parallel with the long axis of the cells. Intermediate filaments, detected as vinculin-containing structures, are also present along the long axis of the cells (Lesot et al., 1982). In addition, groups of microtubules are present in the cilia, which re- main at advanced secretory stages of the odontoblast (Figure 24). Peripherally to the distal junctional complex, the odontoblast process arises, crosses through the predentin zone, and penetrates into the dentinal tubules. The processes consist of one main trunk, which may divide and occupy two tubules, and several thinner lateral branches (Figures 32, 33). Processes never contain organelles involved in macromolecular synthesis (Figures 28,32), only a number of vesicles involved in exocytosis and endocytosis, trapped in a network of microtubules and intermediate filaments located in the central portion of the processes and interconnected with microfilaments. A subplasmalemmal actinrich undercoat is also evident (Figures 31, 35). Small mitochondria may be present in the processes (Garant, 1972). The plasma membrane of the processes is coated externally with a cationic dye-stainable material, presumably PG. A periplasmatic area, 150 to 200 nm wide along the processes, seen as a ladderlike structure after rapid freezing and freeze substitution, allows extracellular assembly of secreted material as well as accumulation of components being processed for reinternalization (Goldberg and Septier, 1986; Goldberg et al., 1987b). The lateral branches do not contain many vesicles but display a high content of intermediary filaments (Figure 34). The branches extend between the lumina of dentin tubules, but it is not known whether they establish contact with other cell processes. Biosynthetic Pathway Autoradiographic studies have been employed B. The to establish the biosynthetic pathway in dentinogenically active odontoblasts. In the rat incisor, [3H]proline, a major component of collagen, is found in the RER after 2 min, in the Golgi region after 10 min, and in secretory vesicles between 20 and 30 min after injection. Between 90 min and 4 h, it can be found extracellularly in predentin (Frank, 1970; Weinstock and Leblond, 1974). The procollagen formed is thus processed from one cellular compartment to another and packed inside secretory vesicles in order to be secreted into the predentin; thereafter it loses its amino- and carboxy-terminal extensions and aggregates into collagen fibrils. The secretion of collagen occurs 697 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Cytosolic microfilaments, intermediate filaments, and microfilaments associated with the junccomplexes (mainly gap and desmosome-like junctions) can be observed in the mature odontoblasts. Saponin-tannic acid fixation. (x40,500.) FIGURE 28. Organelles, involved in biosynthesis, and numerous mitochondria are located in the odontoblast cell bodies (upper part of figure). The odontoblast cell bodies are tightly connected by junctional complexes. Odontoblast processes, containing cytoskeletal proteins and vesicles, protrude into the predentin. (xl16,400.) FIGURE 29. Freeze-fractured replica. En face view of a gap junction (G), located between two odontoblasts. A desmosome (D) is seen in the upper part of the figure. (x230,000.) FIGURE 30. Scanning electron microscope view of odontoblast cell bodies (0), firmly attached to one another in their distal part. Odontoblast processes inserted into the predentin (PD). (x400.) FIGURE 31. Anti-actin-gold immunolabeling. The junctional complexes are labeled (arrows). The cytosol is also partly stained but to a lesser degree than the subplasmalemmal undercoat (arrowheads). (x27,000; (kindly provided by Dr. A. C. Acevedo.) FIGURE 27. tional 698 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURE 32. Section of an odontoblast process (OP) of rat incisor. Microfilaments, bridging intermediate filaments, can be seen in the central part. These accumulate along the inner face of the plasma membrane, forming the subplasmalemmal undercoat. Coated vesicles (CV) and vesicles with various degrees of electron density are also seen. Transverse sections of lateral branchings (LB) can be seen in predentin. (x40,000.) FIGURE 33. Freezefractured replica of an odontoblast process (OP) in predentin. (x84,000.) FIGURE 34. Section of an odontoblast process in predentin, showing the emergence of a lateral branching (LB). Coated vesicles (CV) can be seen together with cytoskeletal proteins, primarily intermediate filaments and microfilaments. (x66,000.) FIGURE 35. Immunocytochemical labeling with an anti-actin antibody, conjugated with colloidal gold. In the main trunk of the odontoblast process (OP) the subplasmalemmal undercoat is strongly labeled, whereas in the center of the process the labeling is considerably lower. Dense labeling is seen throughout a lateral branching (arrowhead). This, together with the appearance in Figure 34, suggests that there is a functional difference between the two parts of the processes. (x54,000.) 699 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. at only one site, the proximal predentin adjacent to the cell bodies. Type I procollagen has been demonstrated by immunolocalization in the odontoblasts (Cournil et al., 1979), the reaction is intense in the predentin close to the odontoblast cell bodies, and the intensity decreases gradually in the distal part of predentin (Karim et al., 1979). There is some difference in opinion between authors who state that abacus-like vesicles, containing bundles of collagenous structures with periodic striations, contain the procollagen, which is secreted (Frank, 1970; Weinstock and Leblond, 1974), and those who, on the basis of the presence of acid phosphatase in these vesicles, suggest that such vesicles are not secretory but lysosomal in character (Goldberg etal., 1987b; Leblond, 1989). A possible explanation is the finding that a considerable portion of newly synthesized collagen in cells may never be secreted but is degraded intracellularly (Bienkowski et al., 1978; Berg etal., 1980). As labeled precursors for phosphoproteins, the route for [3H]serine and [33P]phosphate can be traced through the odontoblasts in a similar manner as [3H]proline. Whereas [33P]phosphate is well within the mineralized dentin 90 min after injection, [3H]serine appears in the proximal predentin as well as in a band located in the dentin at the mineralization front 4 h after injection (Weinstock and Leblond, 1973; Inage and Toda, 1988). The comparison between proline and serine labeling suggests that these precursors in part follow different secretory pathways. Both are constituents of collagen, even though the latter, like phosphate, would reflect mainly phosphoprotein synthesis and transport. From these and other studies, it can be concluded that phosphoprotein secretion does not occur into the predentin matrix, close to the odontoblast cell bodies. There thus seems to be a transport inside the odontoblast processes and a secretion just in advance of the mineralization front (Chart 1). Fucose is a terminal carbohydrate constituent of dentin noncollagenous glycoproteins. [3H]fucose can be detected in the Golgi apparatus 5 to 10 min after injection but is absent from the RER. By 35 min, silver grains are observed in the odontoblast processes, and after 4 h labeling it is seen in the predentin and in the dentin at the mineralization front (Weinstock et al., 1972). From the labeling pattern, it seems likely that secretion occurs mainly at the dentin mineralization front (Chart 1), as is the case with dentin phosphoprotein. The pathway for PG synthesis has been elucidated in part. Radiolabeling of the Golgi apparatus can be seen 5 min after injection of 35S04 (Weinstock and Young, 1972). At 20 min, the label is seen over the secretory granules within the basal portion of the processes, and at this time the predentin adjacent to the cell bodies also begins to be labeled. After 4 h, silver grains become aligned along the dentin side of the predentin-dentin junction but also spread in the predentin extracellular matrix (Sundstrom, 1971). Goldberg and Escaig (1985) compared 35SO4 and [3H]glucosamine incorporation in the rat incisor using different morphological techniques. Aldehyde fixation and demineralization were found to remove between 22 and 67% of the silver grains scored when cryotechnique was used. Again, the secretion was found to occur at two levels. It may be concluded that two pools of PG seem to be secreted by the odontoblasts. It is thus clear that secretion of dentin constituents occurs at different locations during dentinogenesis (Chart 1). The procollagen molecule, together with a pool of PG, are secreted in the proximal part of predentin, close to the cell bodies, whereas Gla-proteins, phosphoproteins, acidic glycoproteins, and a second pool of PGs are secreted at the dentin-predentin interface, the mineralization front. Electron-lucent and electron-dense vesicles, coated and uncoated, are present inside odontoblast processes. Pinched fragments of the plasma membrane, rich in unsaturated fatty acids but cholesterol-poor and covered with membrane-associated particles, represent the initial site of endocytotic events. Transport of such pinched fragments or internalized vesicles may be observed if the fixative procedure is adequate. In contrast, secretory vesicles may be presumed to be released from odontoblast processes, as from other cells, within a few milliseconds. They are thus difficult to demonstrate because of 700 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. the time needed for good fixation, especially with cells organized as a palisade-like structure and linked by junctional complexes. What is seen inside the odontoblast processes after conventional fixation most likely does not represent secretory vesicles but rather endocytotic material. Cryotechnique is, unfortunately, not a remedy due to the limited depth of freezing and to the formation of ice crystals that damage the tissue. From the few observations that exist in spite of these technical constraints, it seems that secretory vesicles are much larger than the small oval vesicles regularly observed in the odontoblast processes. Proteolytic activities are present during dentinogenesis and may be involved in the breakdown of some organic matrix components, presumably being reintemalized and further degraded in the lysosomal system in the odontoblast processes and distal part of the cell bodies. Proteolytic activities have been detected both by gelatin-film substrate technique (Betti and Katchburian, 1982) and biochemically (Smith and Smith, 1984). One specific protease is cathepsin D, which has been detected in odontoblasts as well as in predentin (Linde and Persliden, 1977; Nygren et al., 1979, Andujar et al., 1989). By immunotechnique, dentin collagenase has been demonstrated in human predentin (Dumas et al., 1989); it is also noteworthy that a collagenase inhibitor has been found in bovine dentin, located mainly along the wall of dentinal tubules (Hoshino et al., 1986). Metalloproteinases, active in a pH range of 6 to 9 and activated by calcium, have also been detected in porcine dentin and have been suggested to degrade noncollagenous proteins during dentinogenesis (Fukae et al., 1991). Taken together, there are clear indications that certain components, such as noncollagenous proteins and PGs, are degraded to a certain degree in predentin and may also explain the presence of the lysosome-vacuolar system in the odontoblast processes far away from the odontoblast cell bodies. The microtubules of the cytoskeleton have a well-known role in intracellular transport of secretory granules. Experiments using drugs such as colchicine and vinca alkaloids have given further evidence for this transport during the formation of dentin. Colchicine does not affect the incorporation of [3H]proline into odontoblasts, but the secretion of the predentin matrix is inhibited, and the intracellular retention of the radiolabeled precursor is increased (Nogueira et al., 1988; Ishizeki et al., 1989). Vinblastine modifies cell membrane permeability and interacts with the junctional complexes, and the polarity of the cell is impaired (Goldberg et al., 1981). As a consequence, secretory granules accumulate in the cell bodies in the area between the Golgi apparatus and the place where secretion usually occurs, which shows that microtubules not only play a role in cell polarity but also are involved in intracellular transport of secretory products. In the rat incisor, injection of vinblastine induces within 4 h a disorganization of the architecture of the collagen fibers in the predentin. In control rats they are grossly parallel to the mineralization front (Figures 36-39) but lose this orientation in treated rats (Figure 40). This leads to the conclusion that, whereas collagen fibrillogenesis is clearly a self-regulated process, the orientation of collagen fibers in predentin is regulated by cellular activities. VI. TRANSPORT AND CONCENTRATIONS OF MINERAL IONS For obvious reasons, it is of fundamental importance to understand how calcium and phosphate, the ions constituting the inorganic phase, are transported from blood to the site of mineral formation, their route of transportation, and how this transport is regulated. In the past, a relatively small number of investigations have been devoted to these questions in biomineralization research. For dentinogenesis, calcium is essentially the only ion for which some data are available, and virtually no studies deal with phosphate. One reason is that phosphate, in addition to being present as inorganic ions, is also a constituent of organic molecules, thus complicating the picture. During dentinogenesis, calcium is transferred from the well-developed vascular network in the subodontoblastic area, at the proximal end of the odontoblasts, across the odontoblast cell layer (Chart 2) to be incorporated in the mineral phase at the interface between the nonmineralized 701 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. FIGURES 36 AND 37. When sections are made along the long axis of the rat incisor, the collagen fibers in predentin are essentially parallel with this axis, with each other, and with the dentin mineralization front (arrows). (O) odontoblast; (JC) junctional complex; (PD) predentin; (OP) odontoblast process. (Figure 36: x40,000; Figure 37: x54,000.) FIGURE 38. Transverse section, made at right angles with the long axis of the rat incisor and stained with phosphotungstic acid-chromic acid (PTA). Most collagen fibers in the predentin can be seen in cross-section. (x13,400.) FIGURE 39. Section cut at right angles with the long axis of the rat incisor. Ruthenium hexamine trichloride-glutaraldehyde (RHT). Collagen fibers are cut transversely. RHT-positive material is seen in the form of small granules in the predentin, in close association with the surface of the collagen fibers. Larger aggregates are present in the periodontoblastic space, along the plasma membrane. The plasma membrane is also stained. This suggests the presence of either cell coat material, integral membrane PGs, or nonspecific lipid staining. (x54,000.) FIGURE 40. The spatial organization of collagen fibers in the rat incisor predentin is disrupted 4 h after a single injection of vinblastin sulfate. The direction of sectioning is the same as in Figures 38 and 39. The distribution of collagen fibers is multidirectional, and they are thicker than in untreated rats. Fewer vesicles are seen inside the odontoblast processes, suggesting that most cell activities are interrupted by vinblastin. (x54,000.) 702 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. DENTIN 10 0 0 I, F000 DO II PREDENTIN 000 00 Ca- TPase change horetic uniporter G 1h3 Electro -I Na/Ca exchange A 3_ Na/Ca - a- ATFri. I rvas B . ^% - Ca channel CHART 2. Schematic representation, showing different calcium ion transport pathways in dentinogenically active odontoblasts. Ca2+ ions are transferred from the blood across the odontoblast cell layer, either through the cells, between the cells, or by both pathways (left). If the Ca2+ ions migrate between the cells by diffusion, they would have to penetrate through the intercellular junctions between the odontoblasts and into the predentin layer (arrows). If, on the other hand, Ca2+ ions are actively transported by intracellular pathways, some extrusion mechanism from the odontoblast process and into the predentin space has to be postulated (arrows upper left). The cell to the right illustrates the transmembraneous Ca2+-transporting mechanisms shown to be present in odontoblasts. (From Linde and Lundgren, 1990.) predentin and the mineralized dentin, the miner- alization front. Little was known earlier about the metabolism of calcium during osteogenesis, the majority of data in the literature pertaining mainly to the role of Ca2+ ions as intracellular second messengers, but during the last years the role of 703 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. the osteoblasts in calcium transport has been further elucidated. Before describing what is known about calcium metabolism in dentinogenesis, it is essential to point out a fundamental difference between the techniques used to detect this element. Histochemical stains as well as electron probe methodology, for example, detect total calcium in a specimen, that is, free ions as well as calcium in complexed form, whereas ion specific electrode methodologies and fluorescent probes measure Ca2+ ion activities. The free, ionized Ca2+ is metabolically easily mobilized and seems thus to be the portion of main importance for considerations when discussing mineral induction and growth. A. Calcium in Predentin Nicholson and collaborators (1977) demonstrated early that there is an extracellular enrichment of calcium and phosphorous in rat incisor predentin, compared with rat tail tendon, that in its general composition is a connective tissue resembling predentin matrix, suggesting the existence of some mechanisms that concentrate mineral ions to the site of mineral formation. The findings also implied that the calcium in predentin was in fact bound to some macromolecular constituent. Systematic measurements along the dentin mineralization front further showed regions with a considerably higher calcium enrichment. Nanoliter quantities of a fluid phase were aspirated from rat predentin by Larsson et al. (1988) and analyzed for different elements, including calcium, by means of electron probe microtechnique, thus measuring total calcium. When compared with serum as well as pulpal and cartilage extracellular fluid, lower total concentrations of calcium below 1 mM, were found in predentin. In contrast, phosphorous values were considerably elevated, around 4 to 5 mM. This finding is in contrast to more recent results, obtained by calcium-specific microelectrode methodology, thus revealing free ionized ion activity (Lundgren et al., 1992). The Ca2+ ion activity in predentin (pCa [negative logarithm of calcium ion activity] = 2.9), as measured in situ in rat incisors, was found to be significantly higher than that in the dental pulp (pCa = 3.4) from the same teeth. The almost three times higher calcium activity extracellularly in predentin, compared with the dental pulp, clearly implies the existence of an ion-concentrating mechanism across the odontoblast layer in the direction of the mineralization front. The most plausible explanation is that the odontoblasts are instrumental in this process. B. Odontoblast Calcium Transport Mechanisms In most cell types, cytoplasmic Ca2+ activities are in the micromolar range or below, whereas the total calcium concentration in extracellular fluids, including blood plasma, is around 3 mM, about half or less of which is in ionized form (Carafoli, 1987). A Ca2+ concentration gradient, approximately three orders of magnitude, thus exists over the cell membrane directed into the cells. In most cell types, a large portion of the intracellular calcium is not in free form and is complexed to intracellular organic ligands such as Ca2+-binding proteins, capable of complexing and storing larger amounts of Ca2+. Studies of Ca2+-transporting systems in cells in general have to a large extent been concerned with the maintenance of a low, steady-state Ca2+ activity in the cytosol. Several systems, capable of translocating Ca2+ ions across the different cellular membranes and with different kinetic properties, have been identified and characterized in various cell types. Some of these systems seem to be designed for slower movements of bulk amounts of Ca2+, whereas others are capable of a precise and rapid adjustment of Ca2+ concentrations (Carafoli, 1987). Little is actually known about possible mechanisms for a high capacity, unidirectional, trans-membranous extracellular extrusion of Ca2+ ions, as would be expected in the distal portion of odontoblasts in order to bring such ions to the site of mineral formation. For the analysis of intracellular calcium homeostasis, ion-specific electrode techniques as well as fluorescence measurements have been employed to study Ca2+ influx/efflux cycling in suspensions of microsomes and mitochondria, obtained by subcellular fractionation of rat incisor odontoblasts (Lundgren and Linde, 1987, 704 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. 1988). The steady-state free Ca2+ activities maintained by mitochondria and microsomes in the odontoblast cytoplasm were determined to be pCa = 6.2 to 6.4 and pCa = 6.4 to 6.6, respectively, and these levels were buffered on repeated additions of Ca2+ and EGTA. The intracellular buffering capacity at these pCa levels was further demonstrated in incubations of whole odontoblasts whose plasma membranes had been permeabilized by digitonin (Lundgren and Linde, 1987). The maintained steady-state pCa level was found to be 6.4 to 6.6, thus implying that this is the level of intracellular Ca2+ activity in rat incisor odontoblast cytoplasm, somewhere around 0.5 pM. For a number of years it has been known that high activities of a Ca2+-activated ATPase (CaATPase), which is different from the nonspecific alkaline phosphatase, exist in dentinogenically active odontoblasts (Linde and Magnusson, 1975; Granstrom and Linde, 1976; Granstrom et al., 1979). This Ca-ATPase activity seems to be located in the membranes of vesicles in the distal portion of the cell body, derived from the Golgi complex, as well as in the odontoblast plasma membrane (Granstrom et al., 1978; Lundgren and Linde, 1988). Microsomal fractions, prepared from rat incisor odontoblasts by subcellular fractionation, display an ATP-dependent intravesicular accumulation of Ca2+, demonstrating the involvement of a transport Ca-ATPase (Granstrom and Linde, 1981; Granstrom, 1984; Lundgren and Linde, 1987). Another transport mechanism for Ca2+ extrusion through the odontoblast plasma membrane is a Na+lCa2+ exchanger (Lundgren and Linde, 1988). In addition, this mechanism has been demonstrated to be also present in isolated odontoblast mitochondria. Because of the impressive, inwardly directed Ca2+ ion gradient over the plasma membrane, it may seem that cells like odontoblasts would not need any specialized gating system for the uptake of Ca2+ ions. Obviously, there is a need for some cellular control over this process, and several types of calcium channels have also been identified in various cell types (Carafoli, 1987). In rat incisor odontoblasts, calcium channels of the L-type and possibly also of the N- and Ttypes have been identified recently (Lundgren and Linde, 1992 and unpublished). It can thus be concluded that a large number of the Ca2+-transporting systems, earlier identified in other cell types from noncalcifying tissues, have been demonstrated to exist in dentino- genically active odontoblasts (Chart 2). Their relative importance in this cell type is less understood at present, and no unique Ca2+ ion transport char- acteristics have been identified that make the odontoblast stand out as a cell specifically engaged in mineral formation. C. Ca2+-Binding Proteins in Odontoblasts Ca2+-binding proteins are supposed to be of importance for the calcium regulation between different intracellular compartments. This regulation seemingly occurs in interactions with cytoskeletal proteins and some membrane components. The relationship between this regulation and the general function of odontoblasts in mineral formation has not been clearly established. Several Ca2+-binding proteins have been shown to be present in dentinogenically active odontoblasts. They belong to two major groups, the E-F hand-containing family of proteins and the annexins. Among the first group, calmodulin, the vitamin D-dependent 28-kDa calbindin, and parv- albumin have been demonstrated (Hubbard et al., 1981; Celio etal., 1984; Kardos and Hubbard, 1984; Taylor, 1984). The 28-kDa calbindin is present in human as well as rat odontoblasts (Magloire et al., 1988; Berdal et al., 1989). In rat incisor odontoblasts, the highest concentration of calmodulin can be seen in the cistemae of the RER (Goldberg et al., 1987a). It is also present in the nucleus and in association with different vesicular structures. Mitochondria are not labeled. In the cell processes, a lower concentration can be seen, compared with the cell bodies. Regarding the annexin family, annexins I and II do not seem to be present in odontoblasts, whereas annexins HI to VI have been demonstrated. Annexin V, being the major component, was found to be cytosolic and interacted weakly with the membranes of the cell (Goldberg et al., 1990). In the odontoblast cell body, labeling for annexin VI was seen in the cytosol, nucleus, and 705 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. mitochondria. In the cell processes, the codistribution of annexin VI and actin may suggest that the two are involved in exocytotic and endocytotic events (Goldberg et al., 1991). Related to the intracellular complexing of calcium to proteins is the finding of numerous intra-mitochondrial granules (Figure 26), seen after rapid-freezing and freeze substitution, especially at the junction between the odontoblast cell body and its process (Goldberg and Escaig, 1984b). It may be presumed that, if anything, mitochondria function as a buffering storage for calcium, rather than providing a direct pathway for transfer of Ca2+ ions to dentin. D. Calcium Dynamics during Dentinogenesis A central question is whether the Ca2+ ions, aimed for the mineralization front, are transported through the odontoblasts, between these cells, or along both pathways (Chart 2). Of importance for understanding this is whether the predentin zone, and thus also the mineralization front, could be regarded as being in a totally unrestricted ionic equilibrium with the body fluids. If so, there could be a free influx of the ionic constituents of the dentin mineral, and mineral induction and growth would be solely regulated by extracellular factors. The odontoblasts form a continuous cell layer, but it is not yet clear whether this layer is freely permeable extracellularly. As discussed above (Section V.A), tracers such as La3+ ions and insulin have been employed to resolve this question, unfortunately with irreconcilable results. Recent studies demonstrate, however, that these discrepant results might at least in part be explained by methodological factors such as fixation before analysis (Lundgren and Linde, 1992). The presence of calcium in odontoblasts and their processes has been disclosed mainly by techniques that demonstrate total calcium. Calcium is accumulated in the distal cell body as well as in the odontoblast process (Boyde and Reith, 1977). Calcium deposits inside organelles also have been demonstrated in the cell body as well as the odontoblast process (Reith, 1976; Appleton and Morris, 1979). Autoradiography at different time intervals after injection of 45Ca2+ has been used to provide a dynamic picture of the calcium influx, but published data are conflicting. Arguments for both an intracellular and an intercellular transport route through the odontoblast layer have been presented (Fromme et al., 1972; Munhoz and Leblond, 1974; Nagai and Frank, 1974). The time between injection and the appearance of radioactivity within the dentin ranged between 30 s (Munhoz and Leblond, 1974) and 2 h (Fromme et al., 1972). More recent data, obtained by using an autoradiographic technique aimed at avoiding Ca2+ ion redistribution in the tissue during processing as well as by radiochemical measurements, support a transport route through the odontoblasts and along their processes with the appearance of calcium in the dentin mineral phase 10 to 15 min after i.v. injection (Lundgren and Linde, 1992). It has been found that by disturbing odontoblast microtubules involved in intracellular transport processes by colchicine and by specifically blocking odontoblast calcium uptake channels by nifedipine and neomycin, the Ca2+ ion transport into dentin mineral was strongly impaired (Lundgren and Linde, 1992). This finding may be taken as an indication that transcellular Ca2+ ion transport mechanisms have a major role during dentinogenesis. Together, the studies referred to indicate that some mechanisms exist for concentrating calcium to the site of mineral formation during dentinogenesis and that the odontoblasts may be involved in this process. If this is so, regulatory mechanisms would be easier to envision, being cell-physiological in character, than if no cellular involvement took place. Whether an extra- cellular route exists remains to be further clarified. pH at the Mineralization Front It is known from in vitro experiments that the rate of mineral formation as well as the type of calcium/phosphate mineral formed are highly dependent on factors like pH, the concentrations of calcium and phosphate, the presence of other physiological ions, and charged macroE. molecules. 706 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. In order to characterize the conditions during mineral formation, the pH at the dentin mineralization front within predentin in situ in the rat incisor has been determined by ion-specific micro-electrode technique and found to be 7.0 (Lundgren et al., 1992). The exact value in such measurements may vary with the experimental conditions, but the result shows that no pH extremes are involved. VII. THE ORGANIC MATRIX OF DENTIN Characteristic for the formation of calcified tissues is the prior formation of an extracellular organic framework that becomes mineralized. The role of serving as a template as well as its assumed regulatory capacity for mineral formation has caused the nonmineral extracellular substance of calcified tissues to be referred to as the organic matrix. During dentinogenesis, the odontoblasts thus first synthesize the predentin, which is primarily collagenous in character. In this, the mineral crystals are deposited at some distance away from the cells, that is, a certain time lag after the synthesis of that specific matrix portion. Almost 90% of the dentin organic matrix is collagen, whereas the remainder consists of noncollagenous proteins and PG with a predominantly anionic character and, in addition, some lipid-containing components. Dentin collagen was characterized in detail relatively early (Butler, 1984). During the last 2 decades a major interest has been devoted to the noncollagenous macromolecules in dentin (Linde, 1984, 1989). This may be explained by the specific chemical characteristics of those constituents, conferring on them important functions in processes such as induction and regulation of mineral formation, cell attachment, and collagen fibrillogenesis. The organic matrix of circumpulpal dentin is, with the exception of a minor serum protein content, produced entirely by the odontoblasts. A. Collagens As in bone, the collagen in dentin is primarily of type I. Type I collagen is also predominant in other connective tissues, but whereas type III collagen is a conspicuous constituent in many soft connective tissues, type III is virtually absent from dentin. For example, one third or more of the total collagen in the dental pulp is of type III (Tsuzaki etal., 1990). Several reports give evidence for the absence of collagen type III from predentin (Cournil et al., 1979; Dodd and Carmicael, 1979; Munksgaard and Moe, 1980). It has been reported, however, that odontoblasts in unerupted porcine teeth display immunostaining for type III collagen, whereas dentin and predentin are negative (Tung et al., 1985). A decade ago, the observation was made that odontoblasts synthesize small amounts of collagen type V (Sodek and Mandell, 1982). Further evidence for the presence of this collagen type in rodent dentin has been obtained by immunohistochemical technique (Bronckers etal., 1986). In one study on human dentin, no immunostaining for collagen type V could be discerned, whereas staining for pro-collagen type III and collagen type VI was found to be prominent in predentin (Becker et al., 1986). Collagen of the type I genetic species is made up of three a chains, two of which are identical, giving the composition [al(I)]2a2 (Butler, 1984; Miller and Gay, 1987). Each chain comprises approximately a 1000 amino acid residues. The central portion, about 95% of the total collagen molecule, is triple helical in structure. In this part of the molecule, the individual a chains have an amino acid sequence with glycine in every third position, a prerequisite for helix formation. Another characteristic feature is that the amino acids, proline and hydroxyproline, together account for approximately one fourth of the residues. In the tissue, the rodlike collagen molecules are aggregated in a longitudinally staggered manner to form fibrils. This arrangement creates alternating spatial areas of the fibril with overlapping molecules (overlap zones) and with gaps (hole zones), thus accounting for the cross-striation seen in the electron microscope. Type V collagens exist in a number of tissues, where they constitute only a minor portion. Molecules of the type V system are either homotrimeric or heterotrimeric molecules, comprised of three unique a chains. The heterotrimer [al(V)]2a2(V) is the most prevalent and appears to be the only type V collagen in bone (Miller and Gay, 1987). 707 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. The exact characteristics of type V collagen in dentin are not known. Collagen is synthesized by the odontoblasts mechanisms similar to those in other tissues by and secreted into predentin, where the collagen molecules are arranged into fibers (Reith, 1967; Leblond, 1989). The collagenous network is less dense adjacent to the odontoblast cell bodies but reaches its full density closer to the predentin/ dentin interface. The synthesis occurs on ribosomal complexes in the RER, yielding procollagen chains (Butler, 1984). Relative to the individual a chains, procollagen chains have extensions at their amino-terminal as well as their carboxy-terminal ends, together accounting for approximately one third of the molecule. The a chains are posttranslationally modified by hydroxylation and glycosylation reactions; some of the prolyl and lysyl residues are converted to hydroxyproline and hydroxylysine, respectively, and some carbohydrate is attached. Activities of prolyl hydroxylase have been demonstrated in the odontoblast layer (Takei et al., 1983). Three procollagen chains assemble to form a procollagen molecule with a triple-helical structure in the central part of the molecule, and the molecules are packaged into membrane-bound vesicular structures and exocytosed into predentin (Weinstock and Leblond, 1974; Leblond, 1989). By means of specific procollagen peptidases the amino- and carboxy-terminal extensions of procollagen are excised, and the procollagen molecules thus become converted to collagen. Extracellularly, the collagen molecules spontaneously aggregate in a staggered manner to form collagen fibrils. By the subsequent formation of covalent crosslinks between the collagen molecules, the formed fibrils become stabilized (Eyre, 1987). Lysyl oxidase, an enzyme involved in the initiation of this process, has been demonstrated in the odontoblast/predentin layer (Numata and Hayakawa, 1986). Of the two borohydride-reducible bifunctional cross-links, dihydroxylysinonorleucine (DHLNL) and hydroxylysinonorleucine (HLNL), the former constitutes the major cross-link in dentin and bone (Mechanic et al., 1971). In addition, pyridinoline, which is trifunctional and nonreducible, is known to be a major cross-link (Eyre and Oguchi, 1980; Kuboki et al., 1981). Deoxypyridinoline, a minor cross-link, seems to be unique to bone and dentin (Eyre et al., 1984; Robins and Duncan, 1987). It has been reported that cross-link location differs between dentin and bone collagens (Kuboki and Mechanic, 1982). In mineralized human dentin, the content of pyridinium cross-links has been found to increase with age, concomitant with a decrease in the amounts of reducible cross-links (Walters and Eyre, 1983). Dentin from unerupted bovine incisors contains about one reducible cross-link per collagen molecule, whereas rat incisor dentin contains twice this amount. In contrast, bovine dentin contains twice as many pyridinium crosslinks, about 0.2 per collagen molecule, compared with rat dentin collagen (Linde and Robins, 1988). Because pyridinium cross-links are supposed to be maturation products of the bifunctional crosslinks, a possible interpretation of these findings is that the collagen in rat dentin is less mature than that of bovine dentin. There is evidence in the literature that a large amount of the collagen synthesized by dentinogenically active odontoblasts is of type I trimer, that is, a molecule with the composition [al(I)]3. This has been demonstrated in vitro for rat incisors and mouse molars (Munksgaard et al., 1978a; Lesot, 1981) as well as in vivo for rat incisor odontoblasts (Sodek and Mandell, 1982). This collagen is known to be present in minor amounts in other tissues, but the reason for its synthesis in dentinogenesis and its mode of removal from the tissue need to be clarified. B. Noncollagenous Macromolecules the noncollagenous proteins of dentin and bone are so and PGs (NCPs) strongly associated with the mineral phase in the tissue that they are extractable only after demineralization (Linde et al., 1980). A minor PG fraction can, however, be extracted from the intact tissue. On the other hand, although the quantitatively major portion of the NCPs is extracted on demineralization with, for example, EDTA, some components like the matrix Gla protein (MGP) and bone morphogenetic protein (BMP) need additional procedures to be solubilized. In general, 708 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. 1. Dentin Phosphoprotein The major NCP component in all dentin species is phosphoprotein. Several types of proteins with highly different degrees of phosphorylation can be isolated from dentin (Linde et al., 1980; Butler et al., 1983; Linde, 1988). The denotion dentin phosphoprotein, however, usually refers to the highly phosphorylated phosphoprotein species (PP-H), which by some authors is referred to as phosphophoryn. In dentin, the PP-H fraction constitutes half or more of the NCPs. No PP-H has been demonstrated in bone matrix, thus showing a clear difference in chemical composition between bone and dentin. The content of phosphoprotein in dentin with a low or intermediate degree of phosphorylation (sometimes referred to as PP-L and PP-M) varies, depending on species (Linde, 1988). Some of these other phosphorylated molecules have been identified as phosphorylated NCPs isolated and characterized from bone. PP-H seems to be transported directly to the dentin mineralization front after its synthesis (Chart 1). As demonstrated by biochemical analysis of dissected bovine predentin (Jontell and Linde, 1983) and immunohistochemistry on teeth from different species (MacDougall et al., 1985; Nakamura et al., 1985; Gorter de Vries etal., 1986; Rahima et al., 1988), PP-H is absent from predentin. In organ culture, PP-H is synthesized by odontoblasts (Munksgaard etal., 1978b). As described above (Section V.B), Weinstock and Leblond (1973) demonstrated that [33P]phosphate and [3H]serine in vivo, representing newly synthesized PP-H, were taken up in dentin of rat incisors within a short period after injection, whereas the collagen constituent [3H]proline and some of the [3H]serine, representing in part also collagen, were distributed throughout the predentin extracellular matrix and appeared within dentin only later. The difference in secretory timing between PP-H and collagen has also been confirmed by biochemical technique (DiMuzio and Veis, 1978). PP-H is the most acidic protein known, with a pI of 1.1 for rat PP-H (Jonsson etal., 1978). This is due to the very high phosphate content, 26% in rat and 20% in bovine PP-H, as well as the unique amino acid composition (Linde, 1988). Serine, including phosphoserine, accounts for about 50% of the amino acid residues and aspartic for about 40%. Thus, >80% of the amino acid residues carry negatively charged phosphate or carboxyl groups. Two closely related PP-H molecules can be demonstrated in rat incisor dentin, whereas only one exists in bovine dentin (Linde et al., 1980; Linde, 1988). The finding of two distinct N-terminal sequences for PP-H provides additional evidence for the existence of two PP-H species in rat incisor (Butler et al., 1983). The bovine PP-H is characterized by a somewhat lower phosphate content and a higher lysine content. Because of their unusual composition, PP-H molecules show a pronounced nonideal behavior with standard methods for determining molecular weights (Jontell et al., 1982). Thus, values for PP-H in the literature range between 30 and 100 kDa for rat PP-H and between 35 and 158 kDa for bovine PP-H. By means of sedimentation equilibrium ultracentrifugation of dephosphorylated PP-H, a value of 38 kDa could be calculated for rat incisor dentin PP-H (Jontell et al., 1982). Bovine PP-H seems to be a larger molecule than rat PP-H (Stetler-Stevenson and Veis, 1983). Rat and bovine PP-H molecules are thus different in some respects. However, their general chemical characteristics are similar and, as a result, it could be expected that their respective biological functions should be the same. Even though PP-H is readily solubilized on dentin demineralization, a small portion remains firmly associated with the insoluble collagen. It has been argued that this is covalently linked to collagen and that such a conjugate is formed extracellularly, at the mineralization front, from precursors following different secretory pathways (Maier et al., 1983). It has been suggested that this PP-H fraction has a specific function in mineral formation (Veis et al., 1981), even though it is clear that the amount of PP-H firmly associated with the collagen matrix is much less than was first believed. Investigations by alternative methodology have confirmed the existence of a minute but definite amount of phosphoprotein, strongly bound to collagen in bovine and rat dentin (M. Yamauchi, personal communication; Linde, unpublished). Because of its pronounced polyanionic character, PP-H could 709 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. be expected to bind electrostatically to collagen, which has been substantiated experimentally (Stetler-Stevenson and Veis, 1986). It may thus be debated whether the finding of components, with an apparent amino acid composition of PP-H added to that of collagen, can be interpreted as proof for a covalent linkage between the two (Linde et al., 1981). It has been shown that bovine root dentin contains only half the amount of phosphoprotein, compared with crown dentin from the same teeth (Takagi et al., 1988). Similarly, the content of PP-H in "enamel-related dentin" of rat incisors, that is, the dentin portion on the convex side of the tooth, covered with enamel, was about four times higher than that of "cementumrelated dentin" (Steinfort etal., 1989). The somewhat puzzling problem with these data is that, in contrast to the bovine tooth, where the root is formed subsequently to the crown, the cementum-related and enamel-related portions of the rat incisor dentin are formed simulta- neously. PP-H has been demonstrated to bind Ca2+ ions with a strong affinity (Zanetti et al., 1981; Stetler-Stevenson and Veis, 1987). As judged by nuclear magnetic resonance (NMR), this binding is not localized to specific sites; the Ca2+ ions evidently have a high mobility on the protein surface (Cookson et al., 1980). The concomitant binding of calcium and phosphate ions to PP-H has also been studied by ultrafiltration technique (Marsh, 1989). The protein-mineral ion complex was described as a protein with two different ligands, Ca2+ ions and calcium phosphate clusters having a stoichiometry of about Ca.5PO4, that is, similar to that of an amorphous calcium phosphate. In metastable in vitro solutions, PP-H affects the conversion of amorphous calcium phosphate to hydroxyapatite (Nawrot et al., 1976; Termine and Conn, 1976) and inhibits crystal growth (Termine etal., 1980). Relatively high concentrations of PP-H are needed to exert these inhibitory functions. A strong Ca2+ ion binding of a matrix macromolecule is often interpreted as indicative for a role in mineral formation, and in spite of the above findings it has been postulated for a number of years that phosphoproteins are instrumental in in vivo mineral induction. The first experi- mental evidence for such a function came when it was shown that PP-H at very low concentrations, when immobilized by a stable support, might induce hydroxyapatite formation from calcium phosphate solutions at physiological concentrations in vitro as well as in vivo (Lussi et al., 1988; Linde and Lussi, 1989, Linde etal., 1989; Lussi and Linde, 1993). The possible function of phosphoproteins in mineral induction and growth during dentinogenesis will be discussed in Section VIII.B). 2. Proteoglycans comprise another major portion of noncollagenous components in dentin. They are macromolecules with a number of carbohydrate side chains, glycosaminoglycans (GAGs), covalently bound to a protein core (for reviews, see Heineg&rd and Sommarin, 1987a,b). The side chains are made up of repeating disaccharide units, each consisting of one uronic acid and one N-acetyl-hexosamine. The identity of the different GAGs, and to a large extent also PGs the functional characteristic of the PGs, are defined by the identity of their respective uronic acid and hexosamine residues. The protein core length, and the number, size, and identity of the GAG chains vary between different types of PGs. In PGs, the GAG side chains are sulfated; the location of the sulfate group on the galactosamine makes, for example, the difference between chondroitin-4-sulfate and chondroitin6-sulfate. The nonsulfated GAG hyaluronate (hyaluronic acid) exists in tissues as a highmolecular-weight free carbohydrate chain. Some PG species, although not those in dentin, are capable of forming very large molecular aggregates with hyaluronate and so-called linkproteins. In contrast to the major PG species isolated from cartilage, which contain many GAG side chains and have molecular weights of a million or more, the PGs in mineralized dentin and bone belong to the class of small PGs with only one or two GAG side chains and a molecular weight around 75 kDa (Rahemtulla et al., 1984). They are referred to as decorin (PG-S2; one GAG chain) and biglycan (PG-S1; two GAG chains), respectively, and have a protein core 710 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. size of 36 to 38 kDa (Heinegard and Oldberg, 1989). The GAG side chains in dentin have been identified as galactosaminoglycans, primarily chondroitin-4-sulfate and chondroitin6-sulfate. According to Hjerpe et al. (1983), the former accounts for about 80% in human dentin, whereas the latter constitutes some 14%. In addition, small amounts of a highly hybridized dermatan sulfate have been demonstrated. There is evidence that the PG associated with the dentin mineral contains solely chondroitin-4sulfate (Rahemtulla et al., 1984). Small amounts of keratan sulfate also seem to be present in rat incisor dentin (Steinfort et al., 1990). In micro-dissected porcine predentin, the presence of hyaluronate was demonstrated as well as considerable amounts of a heterogeneous galactosaminoglycan population, presumably in the form of PG (Linde, 1973). In addition to collagen, these constitute the only other major macromolecules in predentin. However, the characteristics of PG in predentin differ from dentin PG in that PGs of different sizes can be found. Thus, a class of PG in predentin seems to have a molecular size considerably greater than those of dentin and bone PGs (Linde et al., unpublished), but the exact chemical nature of this PG remains to be analyzed. Different cationic dyes have been used to visualize PG distribution in the electron microscope, for example, alcian blue, cuprolinic blue, and ruthenium red. After rapid freezing and freeze-substitution fixation, PGs in predentin appear as an amorphous gel, located between and around the collagen fibers (Goldberg and Escaig, 1984a). By using a hyaluronidase-gold complex technique, a more intense staining is found in the proximal zone in predentin, close to the odontoblasts (Chardin etal., 1990). In dentin, PGs are found in close association with the collagen fibers. At higher critical electrolyte concentrations, staining is confined to the dentin edge near the mineralization front. The type and distribution of specific GAGs and PGs in predentin and dentin of rat incisors have been examined by immunohistochemistry. In predentin, staining of the protein core of small and large PGs as well as PGs with chondroitin-4-sulfate/dermatan sulfate, keratan sulfate, and unsulfated chondroitin was quite homogeneous and strong, whereas staining for PG with chondroitin-6-sulfate was relatively weak (Takagi et al., 1990a). Only the protein core of small PG and chondroitin-4-sulfate/ dermatan sulfate PG could be demonstrated in dentin. The PG distribution in mineralized dentin was found to be very heterogeneous in that an intensely reactive component was found in the dentinal tubules, whereas the intertubular dentin was virtually unstained. The functions of PGs in different tissues may vary considerably due to their wide range in chemical composition. The properties of PGs are determined by their substantial negative charge density, their relatively large hydrodynamic size, and the identity of their protein cores. Most investigations of PG function have utilized PGs isolated from cartilage. It may be expected that the large chondroitin sulfate- and keratan sulfate-containing PGs from hyaline cartilage may have properties quite different from those of the small chondroitin sulfate-containing PGs of dentin. Interactions between PGs and collagen have been considered to be of functional significance in tissues (Scott, 1988). PG-collagen complexes have been postulated to promote collagen fiber formation by serving as nuclei. On the other hand, if present after the initial nucleation phase, PGs may significantly retard fiber formation (Obrink, 1973). It has also been shown that PGs stabilize assembled collagen fibers (Snowden, 1982). Decorin is able to bind to collagen in vitro with a dissociation constant of 108 and to influence in vitro collagen fibrillogenesis (Hedbom and HeinegArd, 1989). Thus, one possible function for PGs in dentinogenesis may be simply to affect or even control the organization of the collagenous network being formed in the predentin. Cations such as calcium bind electrostatically to PGs, although rather nonspecifically (Blumenthal, 1981). This ion binding is a prerequisite for the demonstrated capacity of PG, isolated from dentin, to induce hydroxyapatite formation in vitro at physiological pH and ionic conditions (Linde et al., 1989) as well as in vivo (Lussi and Linde, 1993) when immobilized on a solid support (Section VIII.B). In contrast, when in solution, PGs as well as PG aggregates inhibit mineral formation in vitro (Howell and Pita, 1976; 711 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Blumenthal, 1981; Chen et al., 1984). Chondroitin sulfate may inhibit mineralization even in boneforming cultures (Tenenbaum and Hunter, 1987). In the tissue, the diffusion-inhibitory properties of PGs in relation to calcium and phosphate may be of some significance (Maroudas, 1975). It has been suggested that it is the large molecular extension of PGs in solution that is primarily responsible for their ability to regulate mineral formation (Blumenthal, 1981), but again, this statement may pertain mainly to the large cartilage PGs. The turnover of a considerable portion of PGs during dentinogenesis, as demonstrated morphologically by Sundstrom (1971) and chemically by Rahemtulla et al. (1981) and Prince et al. (1984), may be of importance for the course of dentinogenesis. Removal of a significant pool of PGs would affect their putative inhibition of mineralization or influence collagen fibrillogenesis in predentin. 3. y-Carboxyglutamate-Containing Proteins The unusual amino acid ycarboxyglutamate (abbreviated Gla) was first identified in proteins participating in the blood coagulation cascade. It is formed from glutamic acid residues by a post- translational modification reaction that is vitamin K-dependent. Two major groups of Gla-containing proteins are present in bone and dentin, the Gla-containing proteins of the osteocalcin type and the matrix Gla protein (MGP). The former are usually referred to as "bone Gla-protein" or "osteocalcin" in the literature; here they will be denoted Gla-proteins. The proteins in mineralized tissues are essentially structurally unrelated to the blood coagulation factors. Gla-proteins (of the osteocalcin type) are major components in all bone species, as well as in rat incisor dentin (Linde et al., 1982), and are easily extracted from the tissue following demineralization. However, Gla-protein is virtually absent from dentin from developing permanent bovine and human teeth (Gorter de Vries et al., 1988b; Linde, 1988). In fact, the Gla-protein fraction from both dentin and bone comprises a group of Gla-containing constituents that can be separated by chromatography (Linde etal., 1982, 1983). Depending on the species, the Gla-proteins are small, containing <50 amino acid residues, with a Mr slightly <6 kDa. Gla-proteins contain three Gla-residues and one disulfide bridge. Characteristically, the Gla-proteins of dentin and bone also contain one hydroxyproline residue. There is a high degree of sequence homology between species. In particular, the region around the Gla-residues and the disulfide bridge has been well conserved. Gla-proteins have been shown by immunohistochemical technique to be synthesized and secreted by dentinogenically active odontoblasts (Linde and Hansson, 1983). In vitro biosynthetic studies, using rat incisor odontoblast-dentin specimens (DiMuzio et al., 1983) and rat molar tooth germs (Finkelman and Butler, 1985), have given additional evidence. Dentin Gla-proteins apparently are synthesized by the odontoblasts before formation of the first mineral (Finkelman and Butler, 1985, Bronckers etal., 1987), whereas opinions differ as to its appearance in developing bone (Price et al., 1981; Bronckers et al., 1987). Gla-proteins are absent from predentin during dentinogenesis, as determined by biochemical methods (Jontell and Linde, 1983) and by immunohistochemical technique (Linde and Hansson, 1983; Bronckers et al., 1985; Gorterde Vries et al., 1987). Gla-protein can, in fact, be demonstrated within intracellular vesicular structures in the odontoblast process (Linde and Hansson, 1983; Gorter de Vries et al., 1988a,b), suggesting a direct intracellular transport to the dentin mineralization front (Chart 1). Several studies have been performed to elucidate possible physiological functions of Gla-proteins. It has been demonstrated that 1,25dihydroxy-vitamin D3 stimulates the synthesis of Gla-protein, suggesting that Gla-protein may have some specific function in relation to this hormone (Price and Baukol, 1980, Beresford et al., 1984). Some findings indicate a role for Gla-protein in hard tissue resorption (Lian et al., 1984), which is not readily consonant with the high content of Gla-protein in rodent incisor dentin but not in human dentin, where the former never undergoes any physiological resorption. In proteins, Gla binds cations such as calcium at its bidentate binding site, however, with a rather low specificity and affinity (Williams, 1977; Svard 712 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. et al., 1986). It has also been suggested that the Gla-proteins, being primarily anionic in character, influence the formation of mineral during dentinogenesis. Dentin Gla-protein is, however, an ineffective nucleator for mineral (Linde et al., 1989; Lussi and Linde, 1993). Gla-proteins bind strongly and reversibly to hydroxyapatite but not to amorphous calcium phosphate; opinions differ as to whether the Gla residues as such are essential for this binding (Price et al., 1976; Poser and Price, 1979; Hauschka and Wians, 1989). Glaproteins inhibit seeded hydroxyapatite growth in vitro as well as precipitation of mineral from metastable calcium phosphate solutions (Price etal., 1976; Romberg etal., 1986); removal of the y-carboxylate by thermal decarboxylation significantly reduces this inhibition (Romberg et al., 1986; van de Loo et al., 1987). In contrast to the Gla-proteins of the osteocalcin type, which are easily extractable from dentin after demineralization, the MGP is characteristically insoluble in buffer and is extractable from the collagen matrix only by denaturing agents such as urea or guanidine chloride. So far, MGP has been analyzed only in bone, but it has been reported that MGP is present in dentin at comparable levels as in bone in relatively minor amounts (Price et al., 1985). In contrast to the Gla-proteins, MGP is also present in cartilage (Hale et al., 1988). MGP is a 15-kDa, 79-residue protein, containing four to five Gla-residues and, like the Glaproteins, one disulfide bridge. In contrast to the other Gla-proteins in bone and dentin, it is devoid of hydroxyproline. There is no immunological cross-reactivity between MGP and bone Gla-protein (Price et al., 1983). An interesting observation is the distinct difference in timed accumulation of MGP and Gla-proteins during the early formation of bone. In the rat, MGP levels were found to be about the same at any stage of development of the bone, whereas Gla-protein contents were very low in newly mineralized bone (Otawara and Price, 1986). 4. Other Acidic Noncollagenous Proteins In recent years, in addition to the quantitatively predominating groups discussed above, a number of dentin protein NCP components have been isolated and characterized. Many are rich in aspartic acid, glutamic acid, sialic acid, and carbohydrate, and are generally grouped together as acidic glycoproteins. The major NCP in bone, osteonectin, was so named because of its apparent affinity for collagen and hydroxyapatite (Termine et al., 198 la). It contains relatively high amounts of aspartic and glutamic acids, as well as 10% carbohydrate (Sato et al., 1985), about 0.5% organic phosphorous, and was originally reported to have a molecular weight of about 30 kDa (Termine et al., 198 lb; Romberg et al., 1985). Osteonectin from porcine bone has also been characterized extensively (Domenicucci etal., 1988). It differs from the bovine species in that it seems not to be phosphorylated. Osteonectin binds strongly to hydroxyapatite (Romberg et al., 1985) and has been shown to be a potent inhibitor of hydroxyapatite-seeded crystal growth (Romberg et al., 1986) but is apparently devoid of any specific binding to collagen (Otsuka et al., 1984). Osteonectin has been reported to be present in bovine dentin in considerable amounts, 4 to 6% of the total protein extractable from the tissue (Termine et al., 198 lb), and is also present in porcine dentin (Domenicucci etal., 1988). The presence of osteonectin in porcine dentin has been studied by immunohistochemistry (Tung et al., 1985). The predentin was found to stain with greater intensity than the intertubular dentin, and intense immunostaining of the odontoblasts was seen. In contrast, osteonectin seems to be absent from rat incisor dentin (Sodek et al., 1986). Osteonectin is, in fact, a widespread component in the extracellular matrix of noncalcifying tissues, although its concentration is relatively low in tissues other than dentin and bone (Wasi et al., 1984; Tung et al., 1985; Malaval et al., 1987). As demonstrated by cDNA probes, osteonectin mRNA is synthesized by cultured fibroblasts (Young et al., 1986). Another 43-kDa glycoprotein, SPARC, which was first identified as a component of a specific embryonic basement membrane in mice (Mason et al., 1986), has been shown to be identical to osteonectin. From their respective cDNAs, a 92% sequence homology between bovine osteonectin and mouse SPARC 713 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. has been predicted, and high levels of SPARC mRNA can be demonstrated in osteoblasts and odontoblasts by in situ hybridization technique (Holland et al., 1987). Thus, this strong evidence indicates that osteonectin and SPARC are identical and that expression of this gene is not confined to mineralized tissues. It is, at the present stage of knowledge, difficult to envision any specific role for this protein in the calcification process. Osteopontin, also referred to as bone sialoprotein I or 44-kDa bone phosphoprotein, is a prominent phosphorylated glycoprotein in bone matrix. It has an -Arg-Gly-Asp- receptor binding sequence, which is a characteristic for proteins with a capacity for aiding in cell attachment and spreading. Recent evidence points to a role for this protein in the attachment of osteoclasts during bone resorption. Immunohistochemical results indicate that it is present in predentin, but its presence in dentin has not been elucidated completely (Mark et al., 1988). Two major NCPs, a 95-kDa glycoprotein and a 60-kDa glycoprotein, have been isolated from rat incisor dentin and partially characterized (Butler et al., 1985). These contain 34 and 20% carbohydrate, respectively, and are rich in aspartic, serine, glutamic, and glycine. The 95-kDa glycoprotein contains about 10% sialic acid. The function of these NCPs is not yet known. Bronckers et al. (1989) found positive immunostaining for the 95-kDa glycoprotein in odontoblasts, dentin, and predentin. There are indications for a higher content of the 95-kDa glycoprotein in cementumrelated dentin of rat incisors, compared with the enamel-related dentin (Steinfort etal., 1989). Another minor constituent, a 68-kDa protein, present in rat and rabbit dentin, but also in bone and enamel, has been shown by Lesot and coworkers (1988) to have common antigenic determinants with keratins. Dentin also contains serum proteins. Albumin is liberated from rat dentin only after demineralization (Butler et al., 1981). It is not clear whether albumin is synthesized by the odontoblasts, because it has been shown that radioactively labeled albumin, injected intravenously into rabbits, is incorporated into developing dentin (Kinoshita, 1979). a2HS-glycoprotein, a plasma protein of about 50 kDa, has been shown to be present both in bone and in dentin (Leaver et al., 1977). Immunohistochemical results from bovine and human teeth suggest that a2HS-glycoprotein is present mainly in the peritubular dentin, whereas the intertubular dentin was found to be virtually unstained (Takagi et al., 1990b). The significance of the presence of serum proteins in dentin is not clear. 5. Growth Factors Dentin extracellular matrix also contains proteins that are growth factors or growth-related factors, that is, having the capacity to influence processes such as cell recruitment and differentiation, amplify cellular synthetic activities, and thus enhance tissue growth and repair. Thus, rabbit incisor dentin NCP extracts have been found to affect epithelial as well as mesenchymal cells from tooth germs in vitro (Lesot et al., 1986). The mesenchymal cells were found to elongate, polarize, and increase their metabolic activity, when cultured on filters coated with NCP fractions. Dentin matrix, like that of bone, contains bone morphogenetic protein (BMP), which induces the differentiation of new osteogenic cells when implanted in soft tissue (Bang and Urist, 1967). Attempts have been made to isolate this from rat dentin (Conover and Urist, 1981), although the characterization of such a protein in dentin is far from complete. Studies on the morphogenetic properties of dentin NCP extracts in vivo have also been reported (Smith et al., 1990). The extracellular matrix of bone contains a number of polypeptide growth factors, many of which are synthesized by osteoblasts. This has led to the analysis of such factors in dentin. Human dentin NCP extracts stimulate bone cell proliferation in vitro, and quantitation of growth factors in such extracts demonstrated the presence of transforming growth factor 3 (TGF-P), insulin-like growth factor-I (IGF-I), and insulin-like growth factor-II (IGF-II) (Finkelman etal., 1990). All three growth factors were present in concentra- tions lower than those in human bone. Bovine dentin NCP extracts have been shown to cause chondrocyte-like colony formation from mesenchymal cells in an agarose gel culture system, and evidence was given that TGF-PI was at least in part responsible for this activity (Harada et al., 1990). Also, rat incisor dentin contains a factor that alters the phenotypic expression and stimu- 714 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. lates chondrogenesis in fibroblast-like cells in vitro (Veis et al., 1990). Lipids Lipids comprise a quantitatively minor part of the dentin organic constituents, <2% of the organic mass. In general, the lipid composition of dentin is not unique; phospholipids, cholesterol, C. cholesterol esters, and triacylglycerols account for some 90% of the contents (Wuthier, 1984). Of the phospholipids, phosphatidylcholine comprises about half, followed by phosphatidylethanolamine, sphingomyelin, and diphosphatidyl-glycerol. About two thirds of the dentin lipids are extractable before demineralization, whereas one third are tightly complexed to the mineral in the tissue. This portion comprises primarily acidic phospholipids, such as phosphatidylserine, phosphatidylinositol, and phosphatidic acid. In predentin, the lipid content has been reported to be higher than that in mature dentin (Ellingson et al., 1977). It must be borne in mind, when discussing lipids, that some of the lipids extractable from the tissue are regular cell membrane components, rather than genuine matrix constituents. The first indication that lipids may be involved in dentin mineralization is the finding by Irving (1959) of an intense sudanophilia at the predentin-dentin interface, presumably due to an accumulation of acidic phospholipids (Shapiro etal., 1966). Irving's (1959) finding that a rachitogenic diet in rats led to an impairment of dentin mineralization, in addition to a perturbation of the sudanophilia, was taken as indirect evidence for a functional role of lipids in dentinogenesis. The disturbance was restored on vitamin D repletion. Dentinogenesis has also been shown to be adversely affected by a deficiency in essential fatty acids (Prout and Tring, 1973). Earlier histochemical investigations were carried out on conventionally fixed material; later studies have employed methods better suited to preserve lipids in the tissue. Using malachite greenaldehyde fixation, a network of granules and filaments was demonstrated in the spaces between the collagen fibers in predentin, and granules or needlelike structures coating the collagen fibers in dentin (Goldberg and Septier, 1985). In contrast to Irving's results, the predentin-dentin junction was not especially stained. By rapid freezing and freeze-substitution fixation, a network of branched fibrillar structures, disappearing on methanol extraction, was seen between the collagen fibers in the predentin; similar results have been obtained by a phospholipase-gold technique (Goldberg et al., 1993). A gradient between the proximal part of the predentin and the distal part, close to the mineralization front, where the enzyme-gold labeling was the highest, could be visualized. No difference was, however, detectable between the mineralization front and other parts of the dentin. A number of possible functions of phospholipids in biomineralization have been discussed. Although there are some discrepancies between existing in vitro studies, it is clear that proteolipids and complex acid phospholipids may influence the process of mineral formation. Phospholipids bind calcium ions with moderate affinity and selectivity. This binding is enhanced in the presence of phosphate ions, resulting in the formation of calcium-acidic-phospholipid phosphate complexes (Cotmore et al., 1971). These can be isolated from calcifying but not from noncalcifying tissues (Boskey, 1978). Boskey and co-workers (Boskey and Posner, 1977; Boskey, 1989) have shown such complexes to be able to enhance mineral formation in vitro. Proteolipids and complexed acidic phospholipids have also been shown to be capable of inducing hydroxyapatite in vivo (Raggio etal., 1986). It has been postulated that acidic phospholipids may act to promote initial mineral formation; they also may control the proliferation of these mineral crystals (Boskey and Dick, 1991). D. Comments on the Matrix It is obvious that even though dentin is related to bone and to soft connective tissues in its general composition, the composition of the dentin organic matrix is in some ways unique. Dentin is essentially devoid of fibronectin and collagen type III, both of which are ubiquitous constituents of soft connective tissues. Like bone, dentin contains macromolecules that are characteristic for mineralized tissues such as collagen type I, PG, Gla-proteins, and MGP. It seems clear, however, that PP-H is a molecule unique for dentin. It is important to realize that major chemical differences exist between dentin from different 715 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. species; the composition of the organic matrix of dentin may even differ quite considerably between different parts of the tooth. The collagen cross-linking differs between rat and bovine dentin. PP-Hs constitute the major NCP fraction in both rat and bovine dentin. The rat incisor PP-H fraction, however, consists of two molecular species, whereas only one has been identified in bovine dentin; these PP-Hs also differ in composition. Gla-protein of the osteocalcin type is a major fraction in rat incisor dentin but is seemingly absent from bovine and human dentin. Furthermore, osteonectin/SPARC seems to be a major component of bovine and porcine dentin but is reportedly absent from rat incisor dentin. It should also be pointed out that virtually nothing is known about the matrix constituents of secondary and tertiary dentin. A characteristic feature for dentinogenesis, as for osteogenesis, is the prior formation of the organic matrix, which after a certain time lag becomes mineralized. Because the composition of the predentin matrix in some respects differs from that of the mineralized dentin, such as the absence of PP-H from the former, an analysis and comparison between the two may provide valuable insights into the specific functions of the respective constituents. In this respect, the study of dentinogenesis offers an advantage over osteogenesis because, in contrast to osteoid, the predentin is easily dissected out for analysis. In the following, some of these differences will be discussed. In addition to collagen type I, predentin contains PG and, in some species, osteonectin/ SPARC. PP-H is clearly absent from predentin. This finding, together with the above-mentioned radioisotope experiments, suggests that newly synthesized PP-H is directly transported within the odontoblast processes through the predentin zone and secreted just in advance of the mineralization front (Chart 1). The alternative explanation, that PP-H would diffuse rapidly by an extracellular route through the collagenous matrix, seems less likely in view of its extended molecular configuration. In bovine predentin, the content of reducible collagen cross-links is almost twice that of dentin. Only minute amounts of nonreducible cross-links are present in predentin, whereas the nonreducible pyridinium cross-links can be found in mineral- ized dentin. The observed differences in crosslinking between predentin and dentin of the same teeth may indicate some alterations within the area of mineralization and, in addition, clearly identify predentin as a zone for organization of the collagen web before mineralization. During dentinogenesis, Gla-proteins of the osteocalcin type are not present in predentin but can be demonstrated within intracellular vesicular structures in the odontoblast processes, indicating a direct intracellular transport to the dentin mineralization front. Two pools of sulfated PGs, with different metabolic characteristics, seem to be synthesized by the odontoblasts (Chart 1), one pool being incorporated into mineralized dentin, presumably by a direct transport mechanism within the odontoblast process. The other pool remains in predentin for an extended time and seems to be metabolized there by PG-degrading enzymes (Sundstrom, 1971; Rahemtulla et al., 1981; Prince et al., 1984). The concept that GAG degradation and removal accompany proteolysis of PG core protein during dentin mineralization is supported by immunohistochemical data (Takagi et al., 1990a). Some researchers, however, have been unable to demonstrate any partial loss of PGs. PGs are known, from in vitro experiments, to be able to influence extracellular collagen fibrillogenesis as well as to inhibit mineral formation. It may thus be speculated that the role of predentin PGs would be to exert some regulatory function in collagen fibril formation while inhibiting mineral formation. Because the highest concentrations of osteonectin/SPARC in nonmineralized tissues are observed in collagen-rich tissues, it may be that this protein also has some role in organizing the extracellular matrix. A main reason that the composition of the mineralized dentin matrix is different from that of predentin seems to be that some NCPs such as Gla-protein, PP-H, and PG are added at or just before the mineralization front, presumably subsequent to an intracellular transport along the odontoblast process, followed by exocytosis. Thus an obvious subject matter for future research is the role of these NCPs in the main process occurring at this tissue level, the formation of hydroxya- patite crystals. 716 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. VIII. MINERAL FORMATION A. Dentin Mineral The mineral of dentin is, as in bone, hydroxyapatite, Cao1(OH)2(PO4)6. The dentin mineral contains, however, more than trace amounts of carbonate, and its crystallinity is far from perfect (Posner and Tannenbaum, 1984). The mineral crystals in dentin are generally intimately associated with the collagenous matrix, in that they are largely arranged with their c-axes parallel to the collagen fibers. B. Mechanisms of Mineral Formation It is generally held that induction of mineral crystals during the mineralization of calcified tissues such as dentin is brought about by heterogeneous nucleation. This term implies that there are macromolecules in the organic matrix that display a specific stereochemical arrangement of reactive groups, possessing an electrical charge or other properties that lower the energy barrier, so that a solid phase of calcium phosphate mineral can be formed from a solution that would otherwise be stable (Landis et al., 1977). The stereochemical geometry and the charge distribution are supposed to mimic certain crystal planes of the crystal to be nucleated. Mineral nuclei formed at these nucleation sites would then grow and fuse to form mineral crystals. In addition, several molecules in the tissue are inhibitory to mineral induction and crystal growth, and it may be that the removal of such molecules is of importance. Charged macromolecules of the organic matrix may have this property, but smaller molecules such as pyrophosphate and nucleotides also have been discussed in this context. The PG of predentin is one such candidate; in predentin it would prevent precocious formation of mineral and, possibly, also influence collagen fibrillogenesis. On enzymatic metabolization, mineral formation would be permitted and space would be provided in the interfibrillar matrix for the forming crystals. It has also been discussed that alkaline phosphatase, an enzyme classically related to the calcification process (Granstrom et al., 1979), would have a function in disarming inhibitory phosphate com- pounds. Because the hydroxyapatite crystals in mineralized tissues are found in a specific relationship with the collagen fibers, much research has been devoted to find some distinctive chemical properties of collagen from mineralized tissues that would confer a nucleating function to this predominant matrix constituent. The collagens derived from dentin and bone differ, however, very little from soft tissue collagen in their general chemical composition. There seem, on the other hand, to be differences in the intermolecular cross-linking geometry (Mechanic et al., 1987). Even though cross-link location differs between dentin and bone collagens (Kuboki and Mechanic, 1982), it has been suggested that the cross-linking pattern would be of central importance for the mineralization of calcified tissues (Butler, 1984). One such possibility is that mineralization would physically inhibit the formation of pyridinium cross-links and, conversely, that such cross-links would inhibit mineralization by stabilizing a shortened intermolecular distance between the collagen chains (Banes et al., 1983; Mechanic et al., 1984). In the case of dentinogenesis, the finding of a considerably higher content of pyridinoline in mineralized dentin, compared with predentin, speaks against this hypothesis (Linde and Robins, 1988). Most authors now seem to favor the idea that the collagen in the matrix functions mainly as a means for orientation and a stable support for the mineral crystals and NCPs (Glimcher, 1989). Because of their unique chemical characteristics, the noncollagenous bone and dentin components have instead come into focus as being responsible mainly for the induction and regulation of mineral formation. The concept of phosphoproteins of bone and dentin being involved in the mineralization process has been discussed frequently in earlier reviews (Linde, 1984b; Veis, 1985; Glimcher, 1989). Because of their pronounced anionic characNCPs such as PP-H and PG have an affinity ter, for Ca2+ ions, suggesting that they may function as hydroxyapatite nucleators in vivo. For quite some time, NCPs have been known to exert an inhibitory effect on mineral induction and growth in vitro. Direct evidence that dentin NCPs have the capacity to nucleate apatite at physiological 717 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. ionic conditions has been provided only recently (Lussi et al., 1988; Linde et al., 1989). Available evidence shows that, in order to nucleate a mineral phase, the polyanionic macromolecules need to be immobilized by some solid support; they are then inductive in quite minute amounts. Free in solution, however, they exert their well-known inhibitory function, but this demands relatively high molecular concentrations. Polyanionic proteins exist in various calcifying systems in nature, such as bones, teeth, pathological calcifications, and mollusc shells. It is conceivable that the use of polyanionic macromolecules is a general motif for biomineralization in nature. As discussed above, PP-H and one of the PG pools obviously bypass predentin and are transported directly to the site of mineral formation (Chart 1), thus giving circumstantial evidence for some function for these molecules in mineral formation. Gla-proteins can be localized inside odontoblast processes, also suggesting a direct intracellular transport. The finding that bovine dentin, in contrast to rat dentin, contains no or very little Gla-protein makes it, however, difficult to envision any specific role for this protein in the mineralization process. PG and PP-H, covalently linked to a solid substratum, are capable of inducing an apatitic mineral phase in vitro at conditions similar to those in vivo and where spontaneous mineral precipitation would not occur (Lussi etal., 1988; Linde et al., 1989). To inhibit this mineral formation, several orders of magnitude higher concentrations of PP-H in solution are needed. It has been shown recently that immobilized PP-H and PG are able to induce mineral under actual in vivo conditions (Lussi and Linde, 1993). It has further been shown that PP-H, in order to induce mineral, does not necessarily need to be immobilized by covalent attachment to a solid support; immobilization within a gel structure might be sufficient (Boskey et al., 1990). Dissected bovine predentin has been found to be incapable of mineral induction in vitro and in vivo, whereas demineralized dentin, still containing some macromolecular phosphate, has this capacity (Linde and Lussi, 1989; Lussi and Linde, 1993). In addition to being mineral nucleators, it could be speculated that molecules like PP-H and PG might also function as regulators of mineralization rate and crystal size. The PP-H or PG, first released into the matrix, could promote the formation of the initial mineral crystals, whereas the additional accumulation of NCPs could participate in the regulation of the extent of crystal formation (Boskey et al., 1990). The strong affinity binding of calcium ions to PP-H, with the ions not localized but highly mobile on the surface of the molecule, may also bring about a facilitated calcium ion diffusion that would ensure a rapid formation of calcium phosphate mineral in a hydroxyapatite phase (Cookson et al., 1980; Linde, 1984). Based on these findings, it can be concluded that polyanionic NCPs such as PP-H and PG may be responsible for the induction and regulation of mineralization during dentinogenesis. From a quantitative point of view, the minor fraction of PP-H, strongly associated with dentin collagen, would suffice for this purpose. The fact that other strongly polyanionic proteins such as phosvitin from egg yolk are also capable of mineral induction in vitro (Linde et al., 1989) demonstrates that mineral induction by polyanions may be a relatively nonspecific process and implies that other regulating factors certainly are crucial. The compartmentalization in the tissue is of a great significance as well. In addition to playing a role for mineral formation during dentinogenesis, NCPs in the dentin matrix may be of importance for the remineralization of the tissue during the caries process. It has thus been suggested that removal of the soluble NCPs, especially PP-H, may enhance the remineralization potential of root caries lesions (Clarkson et al., 1991). IX. SUMMARY AND CONCLUSIONS Dentin, the most voluminous mineralized tissue of the tooth, may be considered a connective tissue, and its formation, dentinogenesis, is a highly regulated and well-controlled process. A sophis- ticated interaction between various factors, cellular as well as extracellular, takes place beginning at the stage of differentiation and morphogenesis of the tooth germ and subsequently when mineral formation occurs in the dentin extracellular matrix. In several respects, dentinogenesis provides a superior model for experimental studies of 718 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. biomineralization mechanisms. Differentiation phenomena have been studied extensively using tooth development as an experimental model. In fact, much ofthe current knowledge in these fields is derived from studies of the tooth. The constituents of the organic matrix of dentin have been studied in detail, and the specific roles of these molecules are beginning to unravel. In terms of protein composition, dentin species fall into two main groups, the continuously growing rodent teeth and teeth that are formed during a delimited time period, such as bovine and human dentin. Care should thus be taken when attempting to extrapolate findings from one species to the other. Another important area in which dentinogenesis provides a good system for investigations concerns the dynamics of mineral ions such as calcium and phosphate. Some knowledge about the cellular handling of such ions has been acquired, but the picture is far from complete. Very little is known about how dentinogenesis is regulated. The study of dentinogenesis may provide knowledge that can be extrapolated to understand the formation of calcifying tissues in gen- eral. In addition, some information of relevance for the caries process has been provided recently. The last few years have seen an increased knowledge about the cell biology of odontoblasts. With the advent of newer methods of research, including morphological, biophysical, immunochemical, and molecular biological techniques, several additional pieces of the puzzle may soon fall into place. ACKNOWLEDGMENTS Studies by the authors, referred to in this review, were supported by the Faculty of Odontology at the University of G6teborg, University Ren6 Descartes (Paris V), the Fondation de France, the INSERM, the Ministere de l'Education Nationale, and the Swedish Medical Research Council. NOTE ADDED IN PROOF Please note that the article was conceived in 1992. Covering of the literature thus extends essentially through 1991. early REFERENCES Abe, K., T. Oshima, S. Sobue, and Y. Moriwaki: The Crystallinity of Human Deciduous Teeth in Hypo- phosphataemic Vitamin D-Resistant Rickets. Archs. Oral Biol. 34:365-372 (1989). Acevedo, N. and M. Goldberg: Le D6veloppement PostNatal des Structures Dentaires Chez Didelphis albiventris. J. Biol. Buccale 15:23-35 (1987). Andujar, M. B., D. J. Hartmann, G. Caillot, G. Ville, and H. Magloire: Immunolocalization of Cathepsin D in Dental Tissues: Matrix 9:397-404 (1989). Appleton, J. and D. C. Morris: An Ultrastructural Investigation of the Role of the Odontoblast in Matrix Calcification Using the Potassium Pyroantimonate Osmium Method for Calcium Localization. Archs. Oral Biol. 24:467-475 (1979). Avery, J. K., C. F. Cox, and D. J. Chiego: Structural and Physiologic Aspects of Dentin Innervation. In: Dentin and Dentinogenesis. pp. 19-46. Vol. 1, (A. Linde, Ed.) CRC Press, Boca Raton, FL (1984). Banes, A. J., M. Yamauchi, and G. J. Mechanic: Nonmineralized and Mineralized Compartments of Bone. The Role of Pyridinoline in Nonmineralized Collagens. Biochem. Biophys. Res. Commun. 113:975981 (1983). Bang, G. and M. R. Urist: Bone Induction in Excavation Chambers in Matrix of Decalcified Dentin. Arch. Surg. 94:781-789 (1967). Becker, J., D. Schuppan, H. Benzian, T. Bals, E. G. Hahn, C. Cantaluppi, and P. Reichart: Immunohistochemical Distribution of Collagens Types IV, V and VI and of Pro-Collagens Types I and HI in Human Alveolar Bone and Dentine. J. Histochem. Cytochem. 34:1417-1429 (1986). Beertsen, W., A. Niehof, and V. Everts: Effects of 1-Hydroxyethylidene-l,l-Biphosphonate (HEBP) on the Forma- tion ofDentin and the Periodontal Attachment Apparatus in the Mouse. Am. J. Anat. 174:83-103 (1985). Beertsen, W. and A. Niehof: Root-Analogue Versus CrownAnalogue Dentin: A Radioautographic and Ultrastructural Investigation of the Mouse Incisor. Anat. Rec. 215:106-118 (1986). Berdal, A., A. Nanci, N. Balmain, M. Thomasset, A. Brehier, P. Cuisinier-Gleizes, and H. Mathieu: Immunolocalization and Dosing of Calbindins during Amelogenesis in Normal and Vitamin D-Deficient Rats. In: Tooth Enamel V, pp 154-162. (R. W. Fearnhead, Ed.) Yokohama, Florence (1989). Beresford, J. N., J. A. Gallagher, J. W. Poser, and R. G. G. Russell: Production of Osteocalcin by Human Bone Cells In Vitro: Effects of 1,25-Dihydroxyvitamin D3, 24,25-Dihydroxyvitamin D3, PTH and Glucocorticoids. Metabol. Bone Dis. Rel. Res. 5:229-234 (1984). Berg, R. A., M. L. Schwartz, and R. G. Crystal: Regulation of the Production of Secretory Proteins: Intracellular Degradation of Newly Synthesized Defective Collagen, Proc. Natl. Acad. Sci. U.S.A. 77:47464750 (1980). 719 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Bernard, G. W.: Ultrastructural Observations of Initial Calcification in Dentine and Enamel. J. Ultrastruct. Res. 41:1-17 (1972). Betti, F. and E. Katchburian: Proteolytic Activity of Developing Dentine of Rat Tooth Germs Revealed by the Gelatin-Film Substrate Technique. Archs. Oral Biol. 27:891-896 (1982). Bienkowski, R. S., B. J. Baum, and R. G. Crystal: Fibroblasts Degrade Newly Synthesized Collagen within the Cells Before It Is Secreted. Nature 276:413-416 (1978). Bishop, M. A.: Evidence for Tight Junctions between Odon- toblasts in the Rat Incisor. Cell Tissue Res. 239:137140 (1985). Bishop, M. A., M. Malhotra, and S. Yoshida: Interodontoblastic Collagen (Von Korff Fibers) and Circumpulpal Dentin Formation: An Ultrathin Serial Section Study in the Cat. Am. J. Anat. 191:67-73 (1991). Blumenthal, N. C.: Mechanism of Proteoglycan Inhibition of Hydroxyapatite Formation. In: The Chemistry and Biology of Mineralized Connective Tissues, pp. 509-515. (A. Veis, Ed.) Elsevier/North-Holland, New York (1981). Bonucci, E.: Matrix Vesicles: Their Role in Calcification. In: Dentin and Dentinogenesis, pp. 135-154. Vol. 1 (A. Linde, Ed.) CRC Press, Boca Raton (1984). Boskey, A. L.: The Role of Calcium-Phospholipid-Phosphate Complexes in Tissue Mineralization. Metab. Bone Dis. Rel. Res. 1:137-148 (1978). Boskey, A. L.: Phospholipids and Calcification. In: Calcified Tissues, pp. 215-243. (D. W. L. Hukins, Ed.) Macmillan, London (1989). Boskey, A. L. and B. L. Dick: The Effect of Phosphatidylserine on In Vitro Hydroxyapatite Growth and Proliferation. Calcif. Tissue Int. 49:193-196 (1991). Boskey, A. L., M. Maresca, S. Doty, B. Sabsay, and A. Veis: Concentration-Dependent Effects of Dentin Phosphophoryn in the Regulation of In Vitro Hydroxyapatite Formation and Growth. Bone Mineral 11:55-65 (1990). Boskey, A. L. and A. S. Posner: The Role of Synthetic and Bone-Extracted Ca-Phospholipid-PO4 Complexes in Hydroxyapatite Formation. Calcif. Tissue Res. 23:251(1977). Bouligand, Y., J.-P. Denefle, J.-P. Lachaire, and M. Maillard: 258 Twisted Architectures in Cell-Free Assembled Collagen Gels: Study of Collagen Substrates Used for Cultures. Biol. Cell 54:143-162 (1985). Boyde, A. and S. J. Jones: Scanning Electron Microscopic Studies of the Formation of Mineralized Tissues. In: Developmental Aspects of Oral Biology. pp. 243-274. (H. C. Slavkin and L. A. Bavetta, Eds.) Academic Press, New York (1972). Boyde, A. and E. J. Reith: Quantitative Electron Probe Analysis of Secretory Ameloblasts and Odontoblasts in the Rat Incisor. Histochemistry 50:347-354 (1977). Boyde, A., E. J. Reith, and S. J. Jones: Intercellular Attachments between Calcified Collagenous Tissue Forming Cells in the Rat. Cell Tiss. Res. 191:507-512 (1978). Bronckers, A. L. J. J., S. Gay, M. T. Dimuzio, and W. T. Butler: Immunolocalization of y-Carboxyglutamic AcidContaining Proteins in Developing Molar Tooth Germs of the Rat. Collagen Rel. Res. 5:17-22 (1985). Bronckers, A. L. J. J., S. Gay, R. D. Finkelman, and W. T. Butler: Developmental Appearance of Gla Proteins (Osteocalcin) and Alkaline Phosphatase in Tooth Germs and Bones of the Rat. Bone Mineral 2:361-373 (1987). Bronckers, A. L. J. J., S. Gay, D. M. Lyaruu, R. E. Gay, and E. J. Miller: Localization of Type V Collagen with Monoclonal Antibodies in Developing Dental and Peridental Tissues of the Rat and Hamster. Collagen Rel. Res. 6:1-13 (1986). Bronckers, A. L. J. J., D. M. Lyaruu, and J. H. M. Woltgens: Immunohistochemistry of Extracellular Matrix Proteins during Various Stages of Dentinogenesis. Conn. Tissue Res. 22:65-70 (1989). Butler, W. T.: Dentin Collagen: Chemical Structure and Role in Mineralization. The Morphology of Dentin and Dentinogenesis. In: Dentin and Dentinogenesis. pp. 3753. Vol. 2, (A. Linde, Ed.) CRC Press, Boca Raton (1984). Butler, W. T., M. Bhown, M. T. Dimuzio, W. C. Cothran, and A. Linde: Multiple Forms of Rat Dentin Phosphoproteins. Archs. Biochem. Biophys. 225:178-186 (1983). Butler, W. T., M. Bhown, M. T. Dimuzio, and A. Linde: Noncollagenous Proteins of Dentin: Isolation and Partial Characterization of Rat Dentin Proteins and Proteoglycans Using a Three-Step Preparative Method. Collagen Rel. Res. 1:187-199 (1981). Butler, W. T., S. Sato, F. Rahemtulla, C. W. Prince, M. Tomana, M. Bhown, M. T. Dimuzio, and A. L. J. J. Bronckers: Glycoproteins of Bone and Dentin. In: The Chemistry and Biology of Mineralized Tissues. pp. 107112. (W. T. Butler, Ed.) EBSCO Media, Birmingham, AL (1985). Byers, M. R.: Development of Sensory Innervation in Dentin. J. Comp. Neurol. 191:413-427 (1980). Byers, M. R., M. V. Narhi, and W. K. Dong: Sensory Innervation of Pulp and Dentin in Adult Dog Teeth as Demonstrated by Autoradiography. Anat. Rec. 218:207-215 (1987). Calle, A.: Intercellular Junctions between Human Odontoblasts. A Freeze-Fracture Study after Demineralization. Acta Anat. 122:138-144 (1985). Camarda, A. J., W. T. Butler, R. D. Finkelman, and A. Nanci: Immunocytochemical Localization of y-Carboxyglutamic Acid-Containing Proteins (Osteocalcin) in Rat Bone and Dentin. Calcif. Tissue Int. 40:349-355 (1987). Carafoli, E.: Intracellular Calcium Homeostasis. Ann. Rev. Biochem. 56:395-433 (1987). Celio, M. R., A. W. Norman, and C. W. Heizmann: Vitamin D-Dependent Calcium-Binding-Protein and Parvalbumin Occur in Bones and Teeth. Calcif. Tissue Int. 36:129-130 (1984). Chardin, H., I. Londono, and M. Goldberg: Visualization of Glycosaminoglycans in Rat Incisor Extracellular Ma- 720 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. trix Using a Hyaluronidase-Gold Complex. Histochem. J. 22:588-594 (1990). Chen, C.-C., A. L. Boskey, and L. C. Rosenberg: The Inhibitory Effect of Cartilage Proteoglycan on Hydroxyapatite Growth. Calcif. Tissue Int. 36:285-290 (1984). Chiego, D. J., J. K. Avery, and R. M. Klein: Neuroregulation of Protein Synthesis in Odontoblasts of the First Molar of the Rat after Wounding. Cell Tissue Res. 248:119123 (1987). Clarkson, B. H., F. F. Feagin, S. P. Mccurdy, J. H. Sheetz, and R. Speirs: Effects of Phosphoprotein Moieties on the Remineralization of Human Root Caries. Caries Res. 25:166-173 (1991). Connor, N. S., J. E. Aubin, and A. H. Melcher: The Distribution of Fibronectin in Rat Tooth and Periodontal Tissues. J. Histochem. Cytochem. 32:565572 (1984). Conover, M. A. and M. R. Urist: Dentin Matrix Bone Morphogenetic Protein. In: The Chemistry and Biology of Mineralized Connective Tissues. pp. 597-606. (A. Veis, Ed.) Elsevier/North-Holland, New York (1981). Cookson, D. J., B. A. Levine, R. J. Williams, M. Jontell, A. Linde, and B. De Bernard: Cation Binding by the Rat-Incisor-Dentine Phosphoprotein. A Spectroscopic Investigation. Eur. J. Biochem. 110:273-278 (1980). Cotmore, J. M., G. Nichols, and R. E. Wuthier: Phospho- lipid-Calcium Phosphate Complex: Enhanced Calcium Migration in the Presence of Phosphate. Science 172:1339-1341 (1971). Coumil, I., C. P. Leblond, J. Pomponio, A. R. Hand, L. Sederlof, and G. R. Martin: Immunohistochemical Lo- calization of Procollagens. I. Light Microscopic Distribution of Procollagen I, Il and IV Antigenicity in the Rat Incisor Tooth by the Indirect PeroxidaseAntiperoxidase Method. J. Histochem. Cytochem. 27:1059-1069 (1979). Couve, E.: Ultrastructural Changes during the Life Cycle of Human Odontoblasts. Archs. Oral Biol. 31:643-651 (1986). De Ricqles, A.: Quelques Remarques sur l'histoire Evolutive des Tissus Squelettiques Chez les Vert6bres et Plus Particulierement Chez les T6trapodes. Ann. Biol. 18:135 (1979). Dimuzio, M. T., M. Bhown, and W. T. Butler: The Biosynthesis of Dentine y-Carboxyglutamic Acid-Containing Proteins by Rat Incisor Odontoblasts in Organ Culture. Biochem. J. 216:249-257 (1983). Dimuzio, M. T. and A. Veis: The Biosynthesis of Phosphophoryns and Dentin Collagen in the Continuously Erupting Rat Incisor. J. Biol. Chem. 253:68456852 (1978). Dodd, C. M. and D. J. Carmichael: The Collagenous Matrix of Bovine Predentine. Biochim. Biophys. Acta 577:117124 (1979). Domenicucci, C., H. A. Goldberg, T. Hofmann, D. Isenman, S. Wasi, and J. Sodek: Characterization of Porcine Osteonectin Extracted from Foetal Calvariae. Biochem. J. 253:139-151 (1988). Dumas, J., H. Chardin, D. Septier, and M. Goldberg: Immu- nocytochemical Localization of Collagenase in Human Teeth. J. Dent Res. 68:641 (1989). Eisenman, D. R. and P. L. Glick: Ultrastructure of Initial Crystal Formation in Dentin. J. Ultrastruct. Res. 41:18(1972). Ellingson, J. S., M. Smith, and L. R. Larson: Phospholipid Composition and Fatty Acid Profiles in Bovine Predentin. Calcif. Tissue Res. 24:127-133 (1977). Eyre, D.: Collagen Cross-Linking Amino Acids. Methods Enzymol. 144:115-139 (1987). Eyre, D. R., T. J. Koob, and K. P. Van Ness: Quantitation of Hydroxypyridinium Crosslinks in Collagen by HighPerformance Liquid Chromatography. Anal. Biochem. 137:380-388 (1984). Eyre, D. R. and H. Oguchi: The Hydroxypyridinium Crosslinks of Skeletal Collagens: Their Measurement, Properties and a Proposed Pathway of Formation. Biochem. Biophys. Res. Commun. 92:403-410 (1980). Finkelman, R. D. and W. T. Butler: Appearance of Dentin y-Carboxyglutamic Acid-Containing Proteins in Developing Rat Molars In Vitro. J. Dent. Res. 64:1008-1015 (1985). Finkelman, R. D., S. Mohan, J. C. Jennings, A. K. Taylor, S. Jepsen, and D. J. Baylink: Quantitation of Growth 28 Factors IGF-I, SGF/IGF-II, and TGF-P in Human Dentin. J. Bone Mineral Res. 5:717-723 (1990). Fitzgerald, M., D. J. Chiego, and D. R. Heys: Autoradiographic Analysis of Odontoblast Replacement Following Pulp Exposure in Primate Teeth. Archs. Oral Biol. 35:707-715 (1990). Frank, R. M.: Attachment Site between the Odontoblast Process and the Intradentinal Nerve Fibre. Archs. Oral Biol. 13:833-834 (1968). Frank, R. M.: Autoradiographie de la Dentinog6nese en Microscopie llectronique a l'Aide de la Proline Tritiee Chez le Chat. Archs. Oral Biol. 15:583-596 (1970). Frank, R. M. and P. Steuer: Transmission Electron Microscopy of the Human Odontoblast Process in Peripheral Root Dentine. Archs. Oral Biol. 33:91-98 (1988). Fromme, H., H. J. Hohling, and H. Riedel: Elektronmikroskopische Studien fiber die Dentinbildung. II. Mitteilung: Autoradiographische Untersuchungen zur Funktion der Odontoblasten. D. zahnirztl. Z. 27:6-13 (1972). Fukae, M., I. Kaneko, T. Tanabe, and M. Shimizu: Metalloproteinases in the Mineralized Compartments of Porcine Dentine as Detected by Substrate-Gel Electrophoresis. Archs. Oral Biol. 36:567-573 (1991). Garant, P. R.: The Organization of Microtubules within Rat Odontoblast Processes Revealed by Perfusion Fixation with Glutaraldehyde. Archs. Oral Biol. 17:1047-1058 (1972). Glimcher, M. J.: Mechanism of Calcification: Role of Col- lagen Fibrils and Collagen-Phosphoprotein Complexes In Vitro and In Vivo. Anat. Rec. 224:139-153 (1989). Goldberg, M. and F. Escaig: Odontoblastes: Collagene Dans la Predentine et la Dentine de l'Incisive de Rat. Etude par Cryofracture. Biol. Cell 40:203-216 (1981). 721 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Goldberg, M. and F. Escaig: The Appearance in TEM of Proteoglycan in Predentine is Fixation Dependent. J. Microsc. 134:161-167 (1984a). M. and F. Escaig: Improved Preservation of Intramitochondrial Granules in Rat Incisor Odontoblasts by Rapid-Freezing and Freeze-Substitution Fixation. Archs. Oral Biol. 29:295-301 (1984b). Goldberg, M. and F. Escaig: Incorporation of [35S]Sulfate and [3H]Glucosamine into Glycosaminoglycans in Rat Incisor Predentine and Dentin: Comparison by Autoradiography of Fixation by Rapid-Freezing, Freeze Substitution, and Aldehyde Fixation. Calcif. Tissue Int. 37:511-518 (1985). Goldberg, M., F. Escaig, J. Feinberg, and S. Weinman: Ultrastructural Localization of Calmodulin in Rat Incisor Ameloblasts and Odontoblasts during the Early Stages of Development. Adv. Dent. Res. 1:227-235 (1987a). Goldberg, M. and F. Escaig-Haye: Is the Lamina Lucida of the Basement Membrane a Fixation Artefact? Eur. J. Cell Biol. 42:365-368 (1986). Goldberg, M., F. Escaig, and D. Septier: Etude Par Cryofracture et sur Coupes Fines des Jonctions Intercellulaires Entre les Odontoblastes de l'Incisive de Rat. Modifications Induites par la Vinblastine. J. Biol. Buccale 9:141-153 (1981). Goldberg, M., J. Feinberg, S. Lecolle, M. A. Kaetzel, D. Rainteau, J. Lessard, J. R. Dedman, and S. Weinman: Co-Distribution of Annexin VI and Actin in Secretory Ameloblasts and Odontoblasts of Rat Incisor. Cell Tissue Res. 263:81-89 (1991). Goldberg, M., J. Feinberg, D. Rainteau, S. Lecolle, M. A. Kaetzel, J. R. Dedman, and S. Weinman: Annexins I-VI in Secretory Ameloblasts and Odontoblasts of Rat Incisor. J. Biol. Buccale 18:289-298 (1990). Goldberg, M., S. Lecolle, and H. Chardin: Electron Microscopic Visualization of Phospholipids in Rat Incisor Predentin, Dentin and Forming Enamel Using a Phospholipase A2-Gold Complex. Submitted for publication. Goldberg, M., S. Lecolle, D. Septier, J. P. Carreau, and J. V. Ruch: Lipids in Presecretory and Secretory Ameloblasts Their Role in Basement Membrane Formation. In: Enamel V, pp. 51-58. (R. W. Fearnhead, Ed.) Yokohama, Florence (1989). Goldberg, M., M. Molon Noblot, and D. Septier: Effets de Deux M6thodes de D6min6ralisation sur la Preservation des Glycoprot6ines et des Prot6oglycanes dans les Dentines Intercanaliculaires et P6rcanaliculaires chez le Cheval. J. Biol. Buccale 8:315-330 (1980). Goldberg, M. and D. Septier: Improved Lipid Preservation by Malachite Green-Glutaraldehyde Fixation in Rat Incisor Predentine and Dentine. Archs. Oral Biol. 30:717-726 (1985). Goldberg, M. and D. Septier: Visualization of Proteoglycans and Membrane-Associated Components in Rat Incisor Predentine and Dentine Using Ruthenium Hexammine Trichloride. Archs. Oral Biol. 34:205-212 (1986). Goldberg, M., D. Septier, and F. Escaig-Haye: Glycoconjugates in Dentinogenesis and Dentin. In: Progr. Histochem. Cytochem. pp. 1-113. 17: No 2, Fischer Goldberg, Verlag, Stuttgart (1987b). Gorter De Vries, I., D. Coomans, and E. Wisse: Ultrastructural Localization of Osteocalcin in Rat Tooth Germs by Immunogold Staining. Histochemistry 89:509-514 (1988a). Gorter De Vries, I., D. Coomans, and E. Wisse: Immunocytochemical Localization of Osteocalcin in Human and Bovine Teeth. Calcif. Tissue Int. 43:128-130 (1988b). Gorter De Vries, I., E. Quartier, A. Van Steirteghem, P. Boute, D. Coomans, and E. Wisse: Characterization and Immunocytochemical Localization ofDentine Phosphoprotein in Rat and Bovine Teeth. Archs. Oral Biol. 31:57-66 (1986). Gorter De Vries, I., E. Quartier, P. Boute, E. Wisse, and D. Coomans: Immunocytochemical Localization of Osteocalcin in Developing Rat Teeth. J. Dent. Res. 66:784-790 (1987). Granstr6m, G.: Further Evidence of an Intravesicular Ca2+Pump in Odontoblasts from Rat Incisors. Archs. Oral Biol. 29:599-606 (1984). Granstrom, G., M. Jontell, and A. Linde: Separation of Odontoblast Ca2+-ATPase and Alkaline Phosphatase. Calcif. Tissue Int. 27:211-217 (1979). Granstrom, G. and A. Linde: A Comparison of ATP-Degrading Enzyme Activities in Rat Incisor Odontoblasts. J. Histochem. Cytochem. 24:1026-1032 (1976). Granstrom, G. and A. Linde: ATP-Dependent Uptake of Ca2+ by a Microsomal Fraction from Rat Incisor Odontoblasts. Calcif. Tissue Int. 33:125-128 (1981). Granstr6m, G., A. Linde, and H. Nygren: Ultrastructural Localization of Alkaline Phosphatase in Rat Incisor Odontoblasts. J. Histochem. Cytochem. 26:359-368 (1978). Hale, J. E., J. D. Fraser, and P. A. Price: The Identification of Matrix Gla Protein in Cartilage. J. Biol. Chem. 263:5820-5824 (1988). Harada, K., S. Oida, S. Sasaki, and S. Enomoto: ChondrocyteLike Colony Formation of Mesenchymal Cells by Dentin Extracts in Agarose Gel Culture. J. Dent. Res. 69:1555-1559 (1990). Hauschka, P. V. and F. H. Wians: Osteocalcin-Hydroxyapatite Interaction in the Extracellular Organic Matrix of Bone. Anat. Rec. 224:180-188 (1989). Hedbom, E. and D. Heinegard: Interactions of a 59-kDa Connective Tissue Matrix Protein with Collagen I and Collagen I. J. Biol. Chem. 264:6898-6905 (1989). Heinegard, D. and A Oldberg: Structure and Biology of Cartilage and Bone Matrix Noncollagenous Macromolecules. FASEB J. 3:2042-2051 (1989). Heinegard, D. and Y. Sommarin: Proteoglycans: An Overview. Methods Enzymol. 144:305-319 (1987a). Heinegard, D. and Y. Sommarin: Isolation and Characterization of Proteoglycans. Methods Enzymol. 144:319-372 (1987b). Heritier, M., M. Dangleterre, and Y. Bailliez: Ultrastructure of a New Generation of Odontoblasts in Grafted Coronal Tissues of Mouse Molar Tooth Germs. Archs. Oral Biol. 34:875-883 (1989). Herskovits, M. S. and I. J. Singh: Autoradiography of (3H)Proline Uptake by Fibroblasts of the Periodontal Ligament and Odontoblasts of the Rat Incisor after 722 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Guanethidine-Induced Sympathectomy. Archs. Oral Biol. 31:415-417 (1986). Hjerpe, A., C. A. Antonopoulos, B. Engfeldt, and B. Wikstr6m: Analysis of Dentine Glycosaminoglycans Using High-Performance Liquid Chromatography. Calcif. Tissue Int. 35:496-501 (1983). Holland, G. R.: The Extent of the Odontoblast Process in the Cat. J. Anat. 121:133-149 (1976). Holland, P. W. H., S. J. Harper, J. H. McVey, and B. L. In Vivo Expression of mRNA for the Ca2+-Binding Protein SPARC (Osteonectin) Revealed by In Situ Hybridization. J. Cell Biol. 105:473-482 M. Hogan: (1987). Hoshino, T., J.-I. Kishi, T. Kawai, K. Kobayashi, and T. Hayakawa: Immunoelectron Microscopic Localization of Collagenase Inhibitor in Bovine Dentin. Collagen Rel. Res. 6:303-312 (1986). Howell, D. S. and J. C. Pita: Calcification of Growth Plate Cartilage with Special Reference to Studies on Micropuncture Fluid. Clin. Orthoped. 118:208-229 (1976). Hubbard, M. J., M. P. Bradley, T. B. Kardos, and I. T. Forrester: Calmodulin-Like Activity in a Mineralizing Tissue: The Rat Molar Tooth Germ. Calcif. Tissue Int. 33:545-548 (1981). Iguchi, Y., T. Yamamura, T. Ichikawa, S. Hashimoto, T. Horiuchi, and M. Shimono: Intercellular Junctions in Odontoblasts of the Rat Incisor Studied with FreezeFracture. Archs. Oral Biol. 29:487-497 (1984). Inage, T. and Y. Toda: Phosphoprotein Synthesis and Secretion by Odontoblasts in Rat Incisors as Revealed by Electron Microscopic Radioautography. Am. J. Anat. 182:369-380 (1988). Irving, J. T.: A Histological Staining Method for Site of Calcification in Teeth and Bone. Archs. Oral Biol. 1:8996 (1959). Ishizeki, K., Y. Sakakura, T. Nawa, and Y. Harada: Morphological Evidence of the Function of Intracellular Collagen Fibrils in the Embryonic Mouse Molar Odontoblasts Induced by Colchicine Administration. Acta Anat. 134:133-140 (1989). Jones, S. J. and A. Boyde: Ultrastructure of Dentin and Dentinogenesis. In: Dentin and Dentinogenesis. pp. 81-134. Vol. I, (A. Linde, Ed.) CRC Press, Boca Raton (1984). Jonsson, M., S. Fredriksson, M. Jontell, and A. Linde: Isoelectric Focusing of the Phosphoprotein of RatIncisor Dentin in Ampholine and Acid pH Gradients. Evidence for Carrier Ampholyte-Protein Complexes. J. Chromatogr. 157:235-242 (1978). Jontell, M. and A. Linde: Non-Collagenous Proteins of Predentine from Dentinogenically Active Bovine Teeth. Biochem. J. 214:769-776 (1983). Jontell, M., H. Pertoft, and A. Linde: Disagreement in Molecular Weight Determinations of Dentin Phosphoprotein. Biochim. Biophys. Acta 705:315-320 (1982). Kardos, T. B. and M. J. Hubbard: Regulatory Role of Calcium in Odontogenesis. In: Tooth Enamel IV. pp. 131135. (R. W. Fearnhead and S. Suga, Eds.) Elsevier, Amsterdam (1984). Karim, A., I. Cournil, and C. P. Leblond: Immunohistochemical Localization of Procollagens. II. Electron Microscopic Distribution of Procollagen I Antigenicity in the Odontoblasts and Predentin of Rat Incisor Teeth by a Direct Method Using Peroxidase Linked Antibodies. J. Histochem. Cytochem. 27:1070-1083 (1979). Katchburian, E.: Membrane-Bound Bodies as Initiators of Mineralization of Dentine. J. Anat. 116:285-302 (1973). Kawasaki, K., S. Tanaka, and T. Ishikawa: On the Incremental Lines in Human Dentine as Revealed by Tetracycline Labeling. J. Anat. 123:427-436 (1977). Kawasaki, K., S. Tanaka, and T. Ishikawa: On the Daily Incremental Lines in Human Dentine. Archs. Oral Biol. 24:939-943 (1980). Kinoshita, Y.: Incorporation of Serum Albumin Into the Developing Dentine and Enamel Matrix in the Rabbit Incisor. Calcif. Tissue Int. 29:41-46 (1979). Kuboki, Y. and G. L. Mechanic: Comparative Molecular Distribution of Cross-Links in Bone and Dentin Collagen. Structure-Function Relationships. Calcif. Tissue Int. 34:306-308 (1982). Kuboki, Y., M. Tsuzaki, S. Sasaki, C. F. Liu, and G. L. Mechanic: Location of the Intermolecular Cross-Links in Bovine Dentin Collagen, Solubilization with Trypsin and Isolation of Cross-Link Peptides Containing Dihydroxylysinonorleucine and Pyridinoline. Biochem. Biophys. Res. Commun. 102:119-126 (1981). Lafleche, R. G., R. M. Frank, and P. Steuer: The Extent of the Human Odontoblast Process as Determined by Transmission Electron Microscopy: The Hypothesis of a Retractable Suspensor System. J. Biol. Buccale 13:293305 (1985). Landis, W. J., M. C. Paine, and M. J. Glimcher: Electron Microscopic Observations of Bone Tissue Prepared Anhydrously in Organic Solvents. J. Ultrastruct. Res. 59:1-30 (1977). Larsson, P. A., D. S. Howell, J. C. Pita, and L. N. Blanco: Aspiration and Characterization of Predentin Fluid in Developing Rat Teeth by Means of a Micropuncture and Micro-Analytical Technique. J. Dent. Res. 67:870875 (1988). Leaver, A. G., M. Thomas, and I. B. Holbrook: Glycoproteins of Human Dentine. Calcif. Tissue Res. 22(Suppl.): 347-349 (1977). Leblond, C. P.: Synthesis and Secretion of Collagen by Cells of Connective Tissue, Bone, and Dentin. Anat. Rec. 224:123-138 (1989). Lesot, H.: Collagen Type I Trimer Synthesis by Cultured Embryonic Mouse Molars. Eur. J. Biochem. 116:541546 (1981). Lesot, H., J. M. Meyer, J. V. Ruch, K. Weber, and M. Osborn: Immunofluorescent Localization of Vimentin, Prekeratin and Actin during Odontoblast and Ameloblast Differentiation. Differentiation 21:133-137 (1982). Lesot, H., M. Osman, and J. V. Ruch: Immunofluorescent Localization of Collagens, Fibronectin and Laminin during Terminal Differentiation of Odontoblasts. Dev. Biol. 82:371-381 (1981). 723 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Lesot, H., A. J. Smith, J. B. Matthews, and J. V. Ruch: An Extracellular Matrix Protein of Dentine, Enamel, and Bone Shares Common Antigenic Determinants with Keratins. Calcif. Tissue Int. 42:53-57 (1988). Lesot, H., A. J. Smith, J.-M. Meyer, A. Staubli, and J. V. Ruch: Cell-Matrix Interactions: Influence of Noncollagenous Proteins from Dentin and Cultured Dental Cells. J. Embryol. Exp. Morphol. 96:195-209 (1986). Lian, J. B., M. Tassinari, and J. Glowacki: Resorption of Implanted Bone Prepared from Normal and WarfarinTreated Rats. J. Clin. Invest. 73:1223-1226 (1984). Linde, A.: A Method for the Biochemical Study of Enzymes in the Odontoblastic Layer during Dentinogenesis. Archs. Oral Biol. 17:1209-1212 (1972). Linde, A.: Glycosaminoglycans ofthe Odontoblast-Predentine Layer in Dentinogenically Active Porcine Teeth. Calcif. Tissue Res. 12:281-294 (1973). Linde, A.: Dentin and Dentinogenesis, Vols. I and I, CRC Press, Boca Raton (1984a). Linde, A.: Noncollagenous Proteins and Proteoglycans in Dentinogenesis. In: Dentin and Dentinogenesis. pp. 55-92. (A. Linde, Ed.) CRC Press, Boca Raton (1984b). Linde, A.: Differences between Non-Collagenous Protein Content of Rat Incisor and Permanent Bovine Dentin. Scand. J. Dent. Res. 96:188-198 (1988). Linde, A.: Dentin Matrix Proteins: Composition and Possible Functions in Calcification. Anat. Rec. 224:154166 (1989). Linde, A.: Structure and Calcification of Dentin. In: Calcification in BiologicalSystems. pp. 269-311. (E. Bonucci, Ed.) CRC Press, Boca Raton (1992). Linde, A., M. Bhown, and W. T. Butler: Noncollagenous Proteins of Dentin. J. Biol. Chem. 255:5931-5942 (1980). Linde, A., M. Bhown, and W. T. Butler: Non-Collagenous Proteins of Rat Dentin. Evidence That Phosphoprotein Is Not Covalently Bound to Collagen. Biochim. Biophys. Acta 667:341-350 (1981). Linde, A., M. Bhown, W. C. Cothran, A. Hoglund, and W. T. Butler: Evidence for Several y-Carboxyglutamate Acid-Containing Proteins in Dentin. Biochim. Biophys. Acta 704:235-239 (1982). Linde, A. and H.-A. Hansson: Localization of Gla Proteins during Calcification. In: Calcium-Binding Proteins 1983. pp. 65-66. (B. de Bernard, G. L. Sottocasa, G. Sandri, E. Carafoli, A. N. Taylor, T. C. Vanaman, and R. J. P. Williams, Eds.) Elsevier, Amsterdam (1983). Linde, A., S. Johansson, R. Jonsson, and M. Jontell: Localization of Fibronectin during Dentinogenesis in Rat Incisor. Archs. Oral Biol. 27:1069-1073 (1982). Linde, A., M. Jontell, T. Lundgren, B. Nilson, and U. Svanberg: Noncollagenous Proteins of Rat Compact Bone. J. Biol. Chem. 258:1698-1705 (1983). Linde, A. and A. Lussi: Mineral Induction by Polyanionic Dentin and Bone Proteins at Physiological Ionic Conditions. Conn. Tissue Res. 21:197-203 (1989). Linde, A., A. Lussi, and M. A. Crenshaw: Mineral Induction by Immobilized Polyanionic Proteins. Calcif. Tissue Int. 44:286-295 (1989). Linde, A. and B. C. Magnusson: Inhibition Studies of Alkaline Phosphatases in Hard Tissue-Forming Cells. J. Histochem. Cytochem. 23:342-347 (1975). Linde, A. and B. Persliden: Cathepsin D Activity in Isolated Odontoblasts. Calcif. Tissue Res. 23:33-38 (1977). Linde, A. and S. P. Robins: Quantitative Assessment of Collagen Crosslinks in Dissected Predentin and Dentin. Collagen Rel. Res. 8:443-450 (1988). Lindskog, S. and L. Hammarstrom: Formation of Intermediate Cementum. III. 3H-Tryptophan and 3H-Proline Uptake into the Epithelial Root Sheath of Hertwig In Vitro. J. Craniofac. Gen. Dev. Biol. 2:171-177 (1982). Lormee, P., S. Lecolle, D. Septier, D. Le Denmat, and M. Goldberg: Autometallography for Histochemical Visualization of Rat Incisor Polyanions with Cuprolinic Blue. J. Histochem. Cytochem. 37:203-208 (1989). Lumsden, A. G. S. and J. A. G. Buchanan: An Experimental Study of Timing and Topography of Early Tooth Development in the Mouse Embryo with an Analysis of the Role of Innervation. Archs. Oral Biol. 31:301-311 (1986). T. and A. Linde: Regulation of Free Ca2+ by Subcellular Fractions of Rat Incisor Odontoblasts. Archs. Oral Biol. 32:463-468 (1987). Lundgren, T. and A. Linde: Na+/Ca2+ Antiports in Membranes of Rat Incisor Odontoblasts. J. Oral Pathol. 17:560-563 (1988). Lundgren, T. and A. Linde: Calcium Ion Kinetics during Dentinogenesis. Effects of Disrupting Odontoblast Cellular Transport Systems. Bone Min. 19:31-44 (1992). Lundgren, T. and A. Linde: A Dihydropyridine-Sensitive Calcium Uptake in Rat Incisor Odontoblasts. Submitted for publication (1993b). Lundgren, T., U. Nannmark, and A. Linde: Calcium Ion Activity and pH in the Odontoblast-Predentin Region: Ion-Selective Microelectrode Measurements. Calcif. Tissue Int. 50:134-136 (1992). Lussi, A., M. A. Crenshaw, and A. Linde: Induction and Inhibition of Hydroxyapatite Formation by Dentine Phosphoprotein In Vitro. Archs. Oral Biol. 33:685-691 Lundgren, (1988). Lussi, A. and A. Linde: Mineral Induction In Vivo by Dentin Proteins. Caries Res. in press (1993). MacDougall, M., M. Zeichner-David, and H. C. Slavkin: Production and Characterization of Antibodies Against Murine Dentin Phosphoprotein. Biochem. J. 232:493- 500 (1985). Magloire, H., A. Joffre, J. Azerad, and D. E. M. Lawson: Localization of 28 kDa Calbindin in Human Odontoblasts. Cell Tissue Res. 254:341-346 (1988). Magloire, H., A. Joffre, J. A. Grimaud, D. Herbage, M. L. Couble, and C. Chavrier: Distribution of Type III Collagen in the Pulp Parenchyma of the Human Developing Tooth. Histochemistry 74:319-328 (1982). Maier, G. D., J. H. Lechner, and A. Veis: The Dynamics of Formation of a Collagen-Phosphophoryn Conjugate in 724 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Relation to the Passage of the Mineralization Front in Rat Incisor Dentin. J. Biol. Chem. 258:1450-1455 (1983). Malaval, L., B. Founier, and P. D. Delmas: Radioimmunoassay for Osteonectin. Concentrations in Bone, Nonmineralized Tissues, and Blood. J. Bone Min. Res. 2:457-465 (1987). Mark, M. P., C. W. Prince, S. Gay, R. L. Austin, and W. T. Butler: 44-kDal Bone Phosphoprotein (Osteopontin) Antigenicity at Ectopic Sites in Newborn Rats: Kidney and Nervous Tissues. Cell Tissue Res. 251:23-30 (1988). Maroudas, A.: Biophysical Chemistry of Cartilaginous Tissues with Special Reference to Solute and Fluid Transport. Biorheology 12:233-240 (1975). Marsh, M. E.: Binding of Calcium and Phosphate Ions to Dentin Phosphophoryn. Conn. Tissue Res. 21:205-211 (1989). Mason, I. J., A. Taylor, J. G. Williams, H. Sage, and B. L. M. Hogan: Evidence from Molecular Cloning That SPARC, a Major Product of Mouse Embryo Parietal Endoderm, Is Related To an Endothelial Cell "Culture Shock" Glycoprotein of M, 43,000. EMBO J. 5:1465-1472 (1986). Mechanic, G. L., A. J. Banes, M. Henmi, and M. Yamauchi: Possible Collagen Structural Control of Mineralization. In: The Chemistry and Biology of Mineralized Tissues. pp. 98-102. (W. T. Butler, Ed.) EBSCO Media, Birmingham, AL (1984). Mechanic, G., P. M. Gallop, and M. L. Tanzer: The Nature of Crosslinking in Collagens from Mineralized Tissues. Biochem. Biophys. Res. Commun. 45:644-653 (1971). Mechanic, G. L., E. P. Katz, M. Henmi, C. Noyes, and M. Yamauchi: Locus of a Histidine-Based, Stable Trifunctional, Helix to Helix Collagen Cross-Link: Stereospecific Collagen Structure of Type I Skin Fibrils. Biochemistry 26:3500-3509 (1987). Miller, E. J. and S. Gay: The Collagens: An Overview and Update. Meth. Enzymol. 144:3-41 (1987). Munhoz, C. O. G. and C. P. Leblond: Deposition of Calcium Phosphate Into Dentin and Enamel As Shown by Radioautography of Sections of Incisor Teeth Following Injection of 45Ca Into Rats. Calcif. Tissue Res. 15:221235 (1974). Munksgaard, E. C. and D. Moe: Types of Collagen in an Extract of Odontoblasts and Dentine from Developing Bovine Teeth. Archs. Oral Biol. 25:485-489 (1980). Munksgaard, E. C., M. Rhodes, R. Mayne, and W. T. Butler: Collagen Synthesis and Secretion by Rat Incisor Odontoblasts in Organ Culture. Eur. J. Biochem. 82:609-617 (1978a). Munksgaard, E. C., W. S. Richardson, and W. T. Butler: Biosynthesis of Phosphoprotein by Rat Incisor Odontoblasts in In Vitro Culture. Archs. Oral Biol. 23:583-585 (1987b). Nagai, N. and R. M. Frank: Electron Microscopic Autoradiography of 45Ca during Dentinogenesis. Cell Tissue Res. 155:513-523 (1974). Nakamura, 0, E. Gohda, M. Ozawa, I. Senba, H. Miyazaki, T. Murakami, and Y. Daikuhara: Immunohistochemi- cal Studies with a Monoclonal Antibody on the Distri- bution ofPhosphophoryn in Predentin and Dentin. Calcif. Tissue Int. 37:491-500 (1985). Nakashima, M.: The Induction of Reparative Dentine in the Amputed Dental Pulp of the Dog by Bone Morphogenetic Protein. Archs. Oral Biol. 35:493-497 (1990). Nawrot, C. F., D. J. Campbell, J. K. Schroeder, and M. Van Valkenburg: Dental Phosphoprotein-Induced Formation of Hydroxyapatite during In Vitro Synthesis of Amorphous Calcium Phosphate. Biochemistry 15:3445-3449 (1976). Nicholson, W. A. P., B. A. Ashton, H. J. H6hling, P. Quint, J. Schreiber, K. Ashton, and A. Boyde: Electron Micro- probe Investigations Into the Process of Hard Tissue Formation. Cell Tissue Res. 177:331-345 (1977). Nishikawa, S. and H. Kitamura: Localization of Actin during Differentiation of the Ameloblast, Its Related Epithelial Cells and Odontoblasts in the Rat Incisor Using NBDPhallacidin. Differentiation 30:237-243 (1986). Nishikawa, S. and H. Kitamura: Microtubules, Intermediate Filaments, and Actin Filaments in the Odontoblast of Rat Incisor. Anat. Rec. 219:144-151 (1987). Nishikawa, S. and S. Sasa: Immunofluorescent Localization of Myosin, Alpha-Actinin and Tropomyosin in Odontoblast Microfilament Bundles of the Rat Incisor. Cell Struct. Funct. 14:407-414 (1989). Nogueira, T. O., T. M. T. Zom, T. Stene, and H. S. Koppang: Autoradiographic Investigation ofEffects of High-Dose Colchicine on Dentinogenesis in Rat Incisors. Acta Anat. 131:45-49 (1988). Numata, Y. and T. Hayakawa: Lysyl-Oxidase Activity in the Odontoblast Predentine Layer Isolated from Bovine Incisor Teeth. Archs. Oral Biol. 31:67-68 (1986). Nygren, H., B. Persliden, H.-A. Hansson, and A. Linde: Cathepsin D: Ultra-Immunohistochemical Localization in Dentinogenesis. Calcif. Tissue Int. 29:251-256 (1979). Obrink, B.: The Influence of Glycosaminoglycans on the Formation of Fibers from Monomeric Tropocollagen In Vitro. Eur. J. Biochem. 34:129-137 (1973). 0rvig, T.: Phylogeny of Tooth Tissues: Evolution of Some Calcified Tissues in Early Vertebrates. In: Structural and Chemical Organization of Teeth. pp. 45-110. (A. E. W. Miles, Ed.) Academic Press, London (1967). Otawara, Y. and P. A. Price: Developmental Appearance of Matrix Gla Protein during Calcification in the Rat. J. Biol. Chem. 261:10828-10832 (1986). Otsuka, K., K.-L. Yao, S. Wasi, P. S. Tung, J. E. Aubin, J. Sodek, and J. D. Termine: Biosynthesis of Osteonectin by Fetal Porcine Calvarial Cells In Vitro. J. Biol. Chem. 259:9805-9812 (1984). Poser, J. W. and P. A. Price: A Method for Decarboxylation of y-Carboxyglutamic Acid in Proteins. Properties of the Decarboxylated y-Carboxyglutamic Acid Protein from Calf Bone. J. Biol. Chem. 254:431-436 (1979). Posner, A. and P. J. Tannenbaum: The Mineral Phase of Dentin. In: Dentin and Dentinogenesis. pp. 17-36. Vol. 2 (A. Linde, Ed.) CRC Press, Boca Raton (1984). Price, P. A. and S. A. Baukol: 1,25-Dihydroxyvitamin D3 Increases Synthesis of the Vitamin K Dependent Bone 725 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Protein by Osteosarcoma Cells. J. Biol. Chem. 255:90639068 (1980). Price, P. A., J. W. Lothringer, S. A. Baukol, and A. H. Reddi: Developmental Appearance of the Vitamin K-Dependent Protein of Bone during Calcification. J. Biol. Chem. 256:3781-3783 (1981). Price, P. A., A. S. Otsuka, J. W. Poser, J. Kristaponis, and N. Raman: Characterization of a y-Carboxyglutamic AcidContaining Protein from Bone. Proc. Natl. Acad. Sci. U.S.A. 73:1447-1451 (1976). Price, P. A., M. R. Urist, and Y. Otawara: Matrix Gla Protein, a New y-Carboxyglutamic Acid-Containing Protein Which Is Associated with the Organic Matrix of Bone. Biochem. Biophys. Res. Commun. 117:765-771 (1983). Price, P. A., M. K. Williamson, and Y. Otawara: Characterization of Matrix Gla Protein. A New Vitamin K-Dependent Protein Associated with the Organic Matrix of Bone. In: The Chemistry and Biology of Mineralized Tissues. pp. 159-163. (W. T. Butler, Ed.) Ebsco Media, Birmingham, AL (1985). Prince, C. W., F. Rahemtulla, and W. T. Butler: Incorporation of [35S]sulphate Into Glycosaminoglycans by Mineralized Tissues In Vivo. Biochem. J. 224:941-945 (1984). E. E. S. and F. C. Tring: Dentinogenesis in Incisors of Essential Fatty Acid Deficient Rats. J. Dent. Res. 52:462-466 (1973). Raggio, C. L., B. D. Boyan, and A. L. Boskey: In Vivo Hydroxyapatite Formation Induced by Lipids. J. Bone Min. Res. 1:409-415 (1986). Rahemtulla, F., C. W. Prince, and W. T. Butler: Isolation and Partial Characterization of Proteoglycans from Rat Incisors. Biochem. J. 218:877-885 (1984). Rahemtulla, F., C. W. Prince, B. Caterson, J. E. Christner, J. R. Baker, and W. T. Butler: Studies on the Chemistry and Metabolism of Dentin Proteoglycans Using an In Vivo Labelling Technique. In: The Chemistry and Biology of Mineralized Connective Tissues. pp. 389-393. (A. Veis, Ed.) Elsevier/North-Holland, New York Princeprout, (1981). Rahima, M., T.-G. Tsay, M. Andujar, and A. Veis: Localiza- tion of Phosphophoryn in Rat Incisor Dentin Using Immunocytochemical Techniques. J. Histochem. Cytochem. 36:153-157 (1988). Ramieri, G., G. C. Amselmetti, F. Baracchi, G. C. Panzica, C. Viglietti-Panzica, R. Modica, and J. M. Pollak: The Innervation of Human Teeth and Gingival Epithelium as Revealed by Means of an Antiserum for Protein Gene Product 9.5 (PGP 9.5). Am. J. Anat. 189:146-154 (1990). Reith, E. J.: Collagen Formation in Developing Molar Teeth of Rats. J. Ultrastruct. Res. 21:383-414 (1967). Reith, E. J.: The Binding of Calcium within the Golgi Saccules of the Rat Odontoblast. Am. J. Anat. 147:267-272 (1976). Robins, S. P. and A. Duncan: Pyridinium Crosslinks of Bone Collagen and Their Location in Peptides Isolated from Rat Femur. Biochim. Biophys. Acta 914:233239 (1987). Romberg, R. W., P. G. Werness, P. Lollar, B. L. Riggs, and K. G. Mann: Isolation and Characterization of Native Adult Osteonectin. J. Biol. Chem. 260:2728-2736 (1985). Romberg, R. W., P. G. Werness, B. L. Riggs, and K. G. Mann: Inhibition of Hydroxyapatite Crystal Growth by Bone-Specific and Other Calcium-Binding Proteins. Biochemistry 25:1176-1180 (1986). Ruch, J. V.: Determinisms of Odontogenesis. Cell Biol. Rev. 14:1-81 (1987). Ruch, J. V.: Odontoblasts: Developmental Aspects. In: Dynamic Aspects of the Dental Pulp. pp. 29-50. (R. Iuoki, T. Kudo, and L. Olgart, Eds.) Chapman and Hall, London (1990). Salomon, J. P., S. Lecolle, M. Roche, D. Septier, and M. Goldberg: A Radioautographic Comparison of the In Vivo 3H-Proline and 3H-Serine Incorporation in the Pulpal Dentine of Rat Molars: Variations According to the Different Zones. J. Biol. Buccale 18:307-312 (1990). Salomon, J. P., D. Septier, and M. Goldberg: Ultrastructure of Interodontoblastic Fibres in the Rat Molar. Archs. Oral Biol. 36:171-176 (1991). Sasaki, T., I. Ishida, and S. Higashi: Ultrastructure and Cytochemistry of Old Odontoblasts in Rat Incisors. J. Electron Microsc. 31:378-388 (1982a). Sasaki, T., K. Nakagawa, and S. Higashi: Ultrastructure of Odontoblasts in Kitten Tooth Germs as Revealed by Freeze-Fracture. Archs. Oral Biol. 27:897-904 (1982b). Sasaki, T., H. Tominaga, and S. Higashi: Endocytic Activity of Kitten Odontoblasts in Early Dentinogenesis. Thin Section and Freeze Fracture Study. J. Anat. 138:485492 (1984). Sasano, Y. and M. Kagayama: Change of Microtubular Arrangement Around Centrioles in Rat Incisor Odontoblasts during Cell Differentiation. J. Craniofac. Genet. Dev. Biol. 7:284-294 (1987). Sato, S., F. Rahemtulla, C. W. Prince, M. Tomana, and W. T. Butler: Acidic Glycoproteins from Bovine Compact Bone. Conn. Tiss. Res. 14:51-64 (1985). Sawada, T., T. Yamamoto, T. Yanagisawa, S. Takuma, H. Hasegawa, and K. Watanabe: Immunochemistry of the Distribution of a-Tubulin in Rat Incisor Odontoblasts and Ameloblasts. Acta Histochem. Cytochem. 22:651658 (1989). Schroeder, L. and R. M. Frank: High-Resolution Transmission Electron Microscopy of Adult Human Peritubular Dentine. Cell Tissue Res. 242:449-451 (1985). Scott, J. E.: Proteoglycan-Fibrillar Collagen Interactions. Biochem. J. 252:313-323 (1988). Shapiro, I. M., R. E. Wuthier, and J. T. Irving: A Study of the Phospholipids of Bovine Dental Tissues. I. Enamel and Dentin. Archs. Oral Biol. 11:501-512 (1966). Shellis, R. P.: Structural Organization of Calcospherites in Normal and Rachitic Human Dentine. Archs. Oral Biol. 28:85-95 (1983). Shimokawa, H., T. Inage, K. Wakao, M. Kobayashi, Y. Toda, H. Shigeta, T. Okamura, H. Okada, and S. Sasaki: Studies on Amelogenin Gene Expression by In Situ Hybridization. J. Dent. Res. 70:311 (1991). 726 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Sigal, M. J., J. E. Aubin, A. R. Ten Cate, and S. Pitaru: The Odontoblast Process Extend to the Dentinoenamel Junction: An Immunocytochemical Study of Rat Dentine. J. Histochem. Cytochem. 32:872-877 (1984a). Sigal, M. J., J. E. Pitaru, J. E. Aubin, and A. R. Ten Cate: A Combined Scanning Electron Microscopy and Immunofluorescence Study Demonstrating That the Odontoblast Process Extends to the Dentinoenamel Junction in Human Teeth. Anat. Rec. 210:453-462 (1984b). Slavkin, H. C.: Gene Regulation in the Development of Oral Tissues. J. Dent. Res. 67:1142-1149 (1988). Slavkin, H. C., C. Bessem, P. Bringas, M. Zeichner-David, A. Nanci, and M. L. Snead: Sequential Expression and Differential Function of Multiple Enamel Proteins during Fetal, Neonatal, and Early Postnatal Stages of Mouse Molar Organogenesis. Differentiation 37:26-39 (1988). Smith, A. and G. Smith: Proteolytic Activity of Rabbit Incisor Dentine. Archs. Oral Biol. 29:1049-1050 (1984). Smith, A. J., R. S. Tobias, C. G. Plant, R. M. Browne, H. Lesot, and J. V. Ruch: In Vivo Morphogenetic Activity of Dentine Matrix Proteins. J. Biol. Buccale 18:123129 (1990). Snowden, J. M.: The Stabilization of In Vivo Assembled Collagen Fibrils by Proteoglycans/Glycosaminoglycans. Biochim. Biophys. Acta 703:21-25 (1982). Sodek, J., C. Domenicucci, P. Zung, F. Kuwata, and S. Wasi: Biosynthesis, Properties and Distribution ofOsteonectin. In: Cell Mediated Calcification and Matrix Vesicles. pp. 135-141. (S. Y. Ali, Ed.) Elsevier, Amsterdam (1986). Sodek, J. and S. M. Mandell: Collagen Metabolism in Rat Incisor Predentine In Vivo: Synthesis and Maturation of Type I, al(I) Trimer, and Type V Collagens. Biochemistry 21:2011-2015 (1982). Steinfort, J., T. Van Den Bos, and W. Beertsen: Differences between Enamel-Related and Cementum-Related Dentin in the Rat Incisor with Special Emphasis on the Phosphoproteins. J. Biol. Chem. 264:2840-2845 (1989). Steinfort, J., T. Van Den Bos, and W. Beertsen: Proteoglycans in Rat Incisor Dentin: Composition, Synthesis and Possible Degradation. Ph.D. thesis, chapter 5, University of Amsterdam (1990). Stetler-Stevenson, W. G. and A. Veis: Bovine Dentin Phosphophoryn: Composition and Molecular Weight. Biochemistry 22:4326-4335 (1983). Stetler-Stevenson, W. G. and A. Veis: Type I Collagen Shows a Specific Binding Affinity for Bovine Dentin Phosphophoryn. Calcif. Tissue Int. 38:135-141 (1986). Stetler-Stevenson, W. G. and A. Veis: Bovine Dentin Phosphophoryn: Calcium Ion Binding Properties of a High Molecular Weight Preparation. Calcif. Tissue Int. 40:97-102 (1987). Sundstrom, B.: New Aspects on the Utilization of Inorganic Sulphate during Dentin Formation. Histochemie 26:6166 (1971). Svard, M., T. Drakenberg, T. Andersson, and P. Femlund: Calcium Binding to Bone y-Carboxyglutamic Acid Protein from Calf Studied by 43Ca NMR. Eur. J. Biochem. 158:373-378 (1986). Szabo, J., I. Szabo, and K. Trombitas: Inter-Odontoblastic Fibers in Human Dentine Observed by Scanning Electron Microscopy. Archs. Oral Biol. 30:161-165 (1985). Takagi, M., H. Hishikawa, Y. Hosokawa, A. Kagami, and F. Rahemtulla: Immunohistochemical Localization of Glycosaminoglycans and Proteoglycans in Predentin and Dentin of Rat Incisors. J. Histochem. Cytochem. 38:319-324 (1990a). Takagi, Y., H. Nagai, and S. Sasaki: Difference in Noncollagenous Matrix Composition between Crown and Root Dentin of Bovine Incisor. Calcif. Tissue Res. 42:97-103 (1988). Takagi, Y. and S. Sasaki: Histological Distribution of Phosphophoryn in Normal and Pathological Human Dentins. J. Oral Pathol. 15:463-467 (1986). Takagi, Y., H. Shimohawa, M. Suzuki, H. Nagai, and S. Sasaki: Immunohistochemical Localization of o2HS Glycoprotein in Dentin. Calcif. Tissue Int. 47:4045 (1990b). Takei, T., Y. Numata, and T. Hayakawa: Prolyl-Hydroxyl- ase Activity in Odontoblasts from Bovine Incisor and Premolar Teeth. Archs. Oral Biol. 28:1067-1068 (1983). Takuma, S. and N. Nagai: Ultrastructure of Rat Odontoblasts in Various Stages of Their Development and Maturation. Archs. Oral Biol. 26:993-1011 (1971). Tanaka, T.: The Origin and Localization of Dentinal Fluid in Developing Rat Molar Teeth Studied with Lanthanum as a Tracer. Archs. Oral Biol. 25:153-162 (1980). Taylor, A. N.: Tooth Formation and the 28,000-dalton Vitamin D-Dependent Calcium-Binding Protein: An Immunocytochemical Study. J. Histochem. Cytochem. 32:159-164 (1984). Ten Cate, A. R.: A Fine Structural Study of Coronal and Root Dentinogenesis in the Mouse: Observations on the So-Called "von KorffFibres" and Their Contribution to Mantle Dentin. J. Anat. 125:183-197 (1978). Ten Cate, A. R.: Oral Histology. Development, Structure and Function, 3rd ed. Mosby, St. Louis (1989). Tenenbaum, H. C. and G. K. Hunter: Chondroitin Sulfate Inhibits Calcification of Bone Formed In Vitro. Bone Mineral 2:43-51 (1987). Termine, J. D., A. B. Belcourt, K. M. Conn, and H. K. Kleinmann: Mineral and Collagen-Binding Proteins of Fetal Calf Bone. J. Biol. Chem. 256:10403-10408 (1981a). Termine, J. D. and K. M. Conn: Inhibition of Apatite Formation by Phosphorylated Metabolites and Macromolecules. Calcif. Tissue Res. 22:149-157 (1976). Termine, J. D., E. D. Eanes, and K. M. Conn: Phosphoprotein Modulation of Apatite Crystallization. Calcif. Tissue Int. 31:247-251 (1980). Termine, J. D., H. K. Kleinman, S. W. Whitson, K. M. Conn, M. L. McGarvey, and G. R. Martin: Osteonectin, a Bone-Specific Protein Linking Mineral to Collagen. Cell 26:99-105 (1981b). Thomas, H. F.: The Effect of Various Fixatives on the Extent of the Odontoblast Process in Human Dentine. Archs. Oral Biol. 28:465-469 (1983). 727 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. Tsukada, K.: Ultrastructure of the Relationship between Odontoblast Processes and Nerve Fibers in Dentinal Tubules of Rat Molar Teeth. Archs. Oral Biol. 32:8792 (1987). Tsuzaki, M., M. Yamauchi, and G. L. Mechanic: Bovine Dental Pulp Collagens: Characterization of Types m and V Collagen. Archs. Oral Biol. 35:195-200 (1990). Tung, P. S., C. Domenicucci, S. Wasi, and J. Sodek: Specific Immunohistochemical Localization of Osteonectin and Collagen Types I and Ill in Fetal and Adult Porcine Dental Tissues. J. Histochem. Cytochem. 33:531-540 (1985). Van De Loo, P. G. F., B. A. M. Soute, J. M. Van Haarlem, and C. Vermeer: The Effect of Gla-Containing Proteins on the Precipitation ofInsoluble Salts. Biochem. Biophys. Res. Commun. 142:113-119 (1987). Veis, A.: Phosphoproteins of Dentin and Bone. Do They Have a Role in Matrix Mineralization? In: The Chemistry and Biology of Mineralized Tissues. pp 170-176. (W. T. Butler, Ed.) EBSCO Media, Birmingham, AL (1985). Veis, A., B. Sires, J. Clohisy, B. Sabsay, and S. Amar: Rat Incisor Dentine Contains a Factor Which Alters the Phenotypic Expression and Stimulates Chondrogenesis in Fibroblast-Like Cells In Vitro. Biomaterials 11:3537 (1990). Veis, A., W. Stetler-Stevenson, Y. Takagi, B. Sabsay, and R. Fullerton: The Nature and Localization of the Phosphorylated Proteins of Mineralized Dentin. In: The Chemistry and Biology of Mineralized Connective Tissues. pp. 377-387. (A. Veis, Ed.) Elsevier/North-Holland, New York (1981). Walters, C. and D. R. Eyre: Collagen Crosslinks in Human Dentin: Increasing Content of Hydroxypyridinium Residues with Age. Calcif. Tissue Int. 35:401405 (1983). Warshawsky, H.: Radioautographic Studies on Amelogenesis. J. Biol. Buccale 7:105-126 (1979). Wasi, S., K. Otsuka, K.-L. Yao, P. S. Tung, J. E. Aubin, J. Sodek, and J. D. Termine: An Osteonectin-Like Protein in Porcine Periodontal Ligament and its Synthesis by Periodontal Ligament Fibroblasts. Can. J. Biochem. Cell Biol. 62:470-478 (1984). Weinstock, M. and C. P. Leblond: Radioautography Visual- ization of the Deposition of a Phosphoprotein at the Mineralization Front in the Dentin of the Rat Incisor. J. Cell Biol. 56:838-845 (1973). Weinstock, M. and C. P. Leblond: Synthesis, Migration and Release of Precursor Collagen by Odontoblasts as Visualized by Radioautography after [3H]Proline Administration. J. Cell Biol. 60:92-127 (1974). Weinstock, A., M. Weinstock, and C. P. Leblond: Autoradiographic Detection of 3H-Fucose Incorporation Into Glycoprotein by Odontoblasts and its Deposition at the Site of the Calcification Front in Dentin. Calcif. Tiss. Res. 8:181-189 (1972). Weinstock, A. and R. W. Young: 35S Uptake by the Golgi Apparatus of Odontoblasts and the Migration of the Label to the Mineralization Front of Dentin. J. Cell Biol. 55:276a (1972). Wigglesworth, D. J., G. A. Longmore, I. M. Kuc, and C. Murdoch: Early Dentinogenesis in Mice: von Korff Fibres and Their Possible Significance. A Preliminary Study by Light and Electron Microscopy. Acta Anat. 127:151-160 (1986). Williams, R. J. P.: Calcium Chemistry and its Relation to Protein Binding. In: Calcium-Binding Proteins and Calcium Function. pp. 3-12. (R. H. Wasserman, R. A. Corradino, E. Carafoli, R. H. Kretsinger, D. H. MacLennan, and F. L. Siegel, Eds.) Elsevier/NorthHolland, New York (1977). Wuthier, R. E.: Lipids in Dentinogenesis. In: Dentin and Dentinogenesis. pp. 93-106. Vol. 2 (A. Linde, Ed.) CRC Press, Boca Raton (1984). Wuthier, R. E.: Mechanisms of Matrix Vesicle-Mediated Mineralization of Cartilage. ISI Atlas Sci.: Biochem. 231-241 (1988). Young, M. F., M. E. Bolander, A. A. Day, C. I. Ramis, P. Gehron-Robey, Y. Yamada, and J. D. Termine: Osteonectin mRNA: Distribution in Normal and Transformed Cells. Nucleic Acids Res. 14:4483-4497 (1986). Zanetti, M., B. De Bernard, M. Jontell, and A. Linde: Ca2+- Binding Studies of the Phosphoprotein from Rat-Incisor Dentine. Eur. J. Biochem. 113:541-545 728 Downloaded from cro.sagepub.com by guest on June 28, 2012 For personal use only. No other uses without permission. View publication stats (1981).