J Dent Res 84(5):390-406, 2005
© 2005 International and American Associations for Dental Research
REVIEW
CRITICAL REVIEWS IN ORAL BIOLOGY & MEDICINE |
Are Cementoblasts a Subpopulation of Osteoblasts or a Unique Phenotype?
D.D. Bosshardt
Department of Periodontology and Fixed Prosthodontics, School of Dental Medicine, University of Berne, Freiburgstrasse 7, CH-3010 Berne, Switzerland; dieter.bosshardt{at}zmk.unibe.
 |
ABSTRACT
|
|---|
Experimental studies have shown a great potential for periodontal regeneration. The limitations of periodontal regeneration largely depend on the regenerative potential at the root surface. Cellular intrinsic fiber cementum (CIFC), so-called bone-like tissue, may form instead of the desired acellular extrinsic fiber cementum (AEFC), and the interfacial tissue bonding may be weak. The periodontal ligament harbors progenitor cells that can differentiate into periodontal ligament fibroblasts, osteoblasts, and cementoblasts, but their precise location is unknown. It is also not known whether osteoblasts and cementoblasts arise from a common precursor cell line, or whether distinct precursor cell lines exist. Thus, there is limited knowledge about how cell diversity evolves in the space between the developing root and the alveolar bone. This review supports the hypothesis that AEFC is a unique tissue, while CIFC and bone share some similarities. Morphologically, functionally, and biochemically, however, CIFC is distinctly different from any bone type. There are several lines of evidence to propose that cementoblasts that produce both AEFC and CIFC are unique phenotypes that are unrelated to osteoblasts. Cementum attachment protein appears to be cementum-specific, and the expression of two proteoglycans, fibromodulin and lumican, appears to be stronger in CIFC than in bone. A theory is presented that may help explain how cell diversity evolves in the periodontal ligament. It proposes that Hertwig’s epithelial root sheath and cells derived from it play an essential role in the development and maintenance of the periodontium. The role of enamel matrix proteins in cementoblast and osteoblast differentiation and their potential use for tissue engineering are discussed.
KEY WORDS: cementoblast • osteoblast • periodontal regeneration • periodontal ligament • cementum • bone • enamel matrix proteins
|
INTRODUCTION
|
|---|
Periodontitis is an infectious and common disease in humans (Brown and Löe, 1993; Burt, 1993). Inadequate treatment may lead to early tooth loss. The goal of regenerative periodontal therapy is to restore the structure and function of the periodontium destroyed or lost due to periodontitis. Achieving periodontal regeneration may be a difficult task, because several tissues need to be reconstructed. The periodontium is a complex structural and functional unit consisting of 4 different components, i.e., the gingiva, the alveolar bone, the periodontal ligament, and the root cementum. A critical step toward achieving periodontal regeneration is the new attachment of connective tissue fibers to the previously contaminated root surface. Root cementum plays an important role in this process, because it invests and securely attaches the periodontal ligament fibers to the root surface.
Numerous studies applying regenerative techniques such as physical barriers (membranes) or growth/differentiation factors in selectively designed experiments have shown that new formation of cementum is feasible. However, predictability and quality of cementum regeneration in an everyday clinical situation appear to be low (Pitaru et al., 1994; McCulloch, 1995; Bartold et al., 2000; Grzesik and Narayanan, 2002). Ideally, the regenerated cementum should closely resemble the AEFC, because it contributes most to the attachment function (Schroeder, 1992; Bosshardt and Schroeder, 1996; Bosshardt and Selvig, 1997). In most periodontal regeneration studies, the quality of the attachment function is questionable, because the newly formed cementum is cellular, the numerical density of inserting fibers is low, and the interfacial tissue bonding appears to be weak.
The use of the terms ‘cementum-like’ or ‘bone-like’ in numerous experiments (e.g., Melcher et al., 1987; MacNeil and Thomas, 1993; Grzesik et al., 1998) and histopathological studies (e.g., El-Labban, 1990; Takeda et al., 2001; Bencharit et al., 2003) clearly reflects the problem of tissue classification. Enamel matrix proteins are supposed to induce the formation of AEFC predictably (Hammarström et al., 1997; Hammarström, 1997). In most studies, however, a bone-like mineralized tissue resembling CIFC is observed (e.g., Sculean et al., 1999, 2003; Yukna and Mellonig, 2000; Cochran et al., 2003; Donos et al., 2003; Bosshardt et al., 2005).
CIFC and bone share some structural features, but do cementoblasts and osteoblasts, therefore, belong to the same cell lineage? Cementum does not have the lamellar organization found in bone, is avascular in most species studied, is non-innervated, does not contain bone marrow, and does not undergo physiological remodeling. Our current understanding is that both cementoblast and osteoblast progenitors reside in the periodontal ligament (Pitaru et al., 1994; McCulloch, 1995; Beertsen et al., 1997a; MacNeil and Somerman, 1999; Cho and Garant, 2000; McCulloch et al., 2000). The periodontal ligament represents a cell renewal system in steady state. In a perivascular location, progenitor cells exhibiting some features of stem cells have been identified. It is not known whether cementoblasts and osteoblasts have a common precursor in the periodontal ligament, or whether distinct precursors exist. We do not know which molecular factors and local conditions direct a progenitor cell toward the osteoblastic or cementoblastic lineage. The aim of this review is to clarify whether cementoblasts represent a unique phenotype, or whether they are positional osteoblasts (MacNeil et al., 1998). The analysis will be performed on several levels: on a functional and structural basis of the extracellular matrix (section I); on a cell morphological basis (section II); on a biochemical basis of the extracellular matrix (section III); on the basis of functional cell regulation (section IV); on the basis of cell differentiation mechanisms (section V); and, finally, on the basis of cell origin within the periodontium (section VI).
|
(I) FUNCTION, STRUCTURE, AND STRUCTURE-FUNCTION RELATIONS
|
|---|
In addition to its mechanical support function, bone serves as the major reservoir of calcium (Weiner et al., 1999). Bone is present at various anatomical locations and is structurally adapted to meet its specific supportive requirements. Bone formation may occur directly through intramembranous ossification, indirectly through chondral ossification, or in growth cartilage through endochondral ossification. Chondral and endochondral ossification will not be discussed in this review article. ‘Woven bone’ forms mainly during embryonic development and in growing children, and is later replaced by lamellar bone. It can re-appear in repair and pathological conditions (Schenk, 1994). Woven bone is characterized by intertwined collagen fibrils that have no preferential fibril orientation. The interfibrillar spaces are comparatively wide (Bianco, 1992), and the mineral density and the number of large osteocytes are high (Fig. 1a
) (Schenk, 1994). ‘Lamellar bone’ possesses a much more complex structure, characterized by matrix layers of parallel collagen fibrils. The axes of collagen fibrils are parallel within a single lamella and opposing in adjacent lamellae. One lamellar unit is about 3–5 µm thick. Lamellar bone may be regarded as a complex rotated plywood-like structure (Weiner et al., 1999). ‘Primary parallel-fibered bone’ may be regarded as an intermediate type of bone. It is formed during earlier stages of bone formation as well as during periosteal and endosteal bone apposition. Its collagen fibrils are aligned parallel to the bone surface but lack a lamellar organization (Schenk, 1994). Bone lacking osteocytes, referred to as ‘acellular bone’, is the unique skeletal tissue of modern bony fish (Teleostei) (Smith and Hall, 1990). Mature bone consists of cortical (compact) (Fig. 1b) and cancellous (trabecular) (Fig. 1c) bone. Both undergo physiological remodeling. The fundamental structural unit of cortical bone remodeling is the osteon (Fig. 1b). It has a cylindrical structure with a central Haversian canal. Volkmann’s canals interconnect neighboring Haversian canals, and blood vessels traversing this system of canals keep the bone cells alive. The Haversian systems may be absent in very thin trabeculae of cancellous bone, because nourishment by diffusion through canaliculi is sufficient. A striking structural entity in bone is the ‘cement line’, which marks the interface between new and old bone (Fig. 1d). The cement line is formed during the reversal phase of the bone-remodeling cycle (Nanci, 1999).
View larger version (166K):
[in this window]
[in a new window]
|
Figure 1. Light-microscopic views of (a–c) undecalcified ground sections and (d–i) decalcified semi-thin sections stained with toluidine blue. (a) Woven bone formation; (b) cortical bone of the bovine mandible; (c) cortical (CB) and trabecular bone (TB) in the dog mandible; (d) alveolar bone in the rat mandible with cement lines (arrowheads); (e) acellular afibrillar cementum (AAC) at and near the dentino-cemental junction (DCJ) in a human tooth; (f) acellular extrinsic fiber cementum (AEFC) along the cervical root portion of a human tooth; and (g) cellular intrinsic fiber cementum (CIFC) as a repair tissue following root resorption in a human tooth. Note the reversal line (arrowheads) between dentin (D) and CIFC. (h) A soft-tissue inclusion with a central blood vessel in bovine cementum (C); and (i) an osteone-like structure with a central blood vessel (BV) within bovine cementum (C). ES, enamel space; PL, periodontal ligament.
|
|
Cementum is a heterogeneous mineralized tissue that covers the entire root surface of the tooth, and may also be found over the cervical enamel and line the apical wall of the root canal system. Cementum in humans consists of several types that differ from one another with respect to location, structure, function, rate of formation, proportional biochemical composition, and degree of mineralization (Bosshardt and Schroeder, 1996; Bosshardt and Selvig, 1997). One of the most important functions of cementum is to invest and anchor the principal periodontal ligament fibers, which span like a meshwork between the root and the alveolar bone, to the root. Other functions of cementum include participation in the maintenance of occlusal relationship (adaptation), repair of root defects after resorption or fracture, and protection of the pulp. ‘Acellular afibrillar cementum’ (AAC) covers minor areas of the cervical enamel, but may also be present along the cervical-most portion of the root dentin (Fig. 1e
). The AAC contains neither visible collagen fibrils nor embedded cells, and morphologically resembles the interfibrillar matrix of the AEFC (Bosshardt et al., 1998). The precise function of the AAC is unknown. The lack of collagen fibrils and the common presence of typical bone- and cementum-related noncollagenous proteins draw attention to a certain parallel between the AAC and the cement line in bone. However, the cement line is an interfacial matrix layer linking old with new bone, and does not reveal a layered appearance. Its formation is therefore restricted to a limited time period. The AAC will not be further discussed in this review. ‘Acellular extrinsic fiber cementum’ (AEFC) prevails on the cervical and middle root portions, but may extend further apically in anterior teeth (Fig. 1f). Under certain experimental conditions, matrix layers resembling the AEFC may also form along the inner wall of the alveolar bone (Beertsen and Everts, 1990). The cementum layer is cell-free and contains densely packed collagen fibers called Sharpey’s fibers. The very high number of principal periodontal ligament fibers inserting into the AEFC (approximately 30,000 fibers/mm2) reflects its important function in tooth attachment (Akiyoshi and Inoue, 1963; Kvam, 1973). The orientation of the Sharpey’s fibers is subject to changes throughout life, due to post-eruptive tooth movement. These changes in orientation are reflected by individual AEFC layers that are interfaced by growth, resting, or incremental lines. AEFC grows very slowly, but at a fairly constant rate. The slow rate of formation, the absence of cementocytes, and the densely aggregated and parallel-oriented Sharpey’s fibers account for the very uniform morphological appearance of the AEFC and make it a unique tissue. ‘Cellular intrinsic fiber cementum’ (CIFC) may be a component of cellular mixed stratified cementum, or may be present as a repair tissue filling resorptive and fracture defects of the root (Fig. 1g). As in bone, cells are trapped in the matrix. The collagen fibrils of CIFC are intrinsic, i.e., they do not protrude from the cementum into the periodontal ligament space. Thus, CIFC has no direct function in tooth attachment. Apart from unorganized collagenous matrix, a well-organized structural pattern—consisting of an alternate lamellar pattern resembling the twisted plywood structure of lamellar bone—has been observed (Chen, 1987; Matsuo and Yajima, 1990; Yamamoto et al., 1997, 2000). The number of cementocytes may be very low in structurally well-organized cementum portions. Subsequently, cementum without cementocytes has been referred to as ‘acellular intrinsic fiber cementum’ (AIFC) (Bosshardt and Schroeder, 1990). The rapid matrix deposition of CIFC may serve as a means to maintain the tooth in functional occlusion. Thus, local factors may have a stimulatory effect on cementoblast activity. Soft connective tissue and blood vessels within the CIFC matrix have not been observed in most species. In bovine teeth, however, structures resembling primitive osteons are present (Figs. 1h, 1i). ‘Cellular mixed stratified cementum’ (CMSC) is a mixture of AEFC and CIFC/AIFC. The stratification is derived from consecutively formed, alternating layers of CIFC/AIFC and AEFC that are unpredictably superimposed on one another. CMSC covers the apical one- to two-thirds of the roots and the furcations. The intrinsic part of CMSC may exert an adaptive function, while the extrinsic part may contribute to tooth anchorage to the surrounding alveolar bone.
From the above, we may conclude that, while the AEFC is truly a unique tissue, the CIFC and bone share the property of cellularity. Structurally and functionally, however, CIFC is not comparable with any bone type, at least in most species studied thus far.
|
(II) BONE CELLS, CEMENTUM CELLS, AND MATRIX FORMATION
|
|---|
Bone formation, maintenance, and resorption are regulated by osteoblasts, osteocytes, bone-lining cells, and osteoclasts. Heterogeneity among osteoblasts has been observed, suggesting that there is not one unique osteoblast phenotype (Candeliere et al., 2001). Osteoblasts are large, cuboidal cells that usually form a single layer that covers all periosteal or endosteal surfaces where bone formation is active (Marks and Popoff, 1988; Schenk, 1994) (Fig. 2a). Their nucleus is round, and their cytoplasm is filled with a prominent Golgi complex and abundant rough endoplasmic reticulum. Osteoblasts synthesize a mixture of molecules that are secreted into the extracellular milieu, where they constitute a seam of unmineralized bone matrix, the osteoid. Some osteoblasts become osteocytes by inversion of their own matrix secretion or by entrapment through neighboring osteoblasts (Marks and Popoff, 1988). The speed of matrix deposition may determine the number of embedded cementocytes (Qiu et al., 2002). Woven bone, which has a high number of osteocytes, is formed much more quickly than both parallel-fibered and lamellar bone (Ferretti et al., 1999). The osteocyte is trapped in the bone matrix in a lacuna (Fig. 2b). Neighboring osteocytes are interconnected by tiny cytoplasmic processes extending through a dense canalicular system, allowing for diffusion of metabolites and cell communication. In mammals, the critical transport distance is on the order of 100 µm (Ham, 1952). Osteocytes may actively participate in bone homeostasis through their involvement in bone turnover, ion exchange, and sensing of mechanical signals (Aarden et al., 1994; Noble and Reeve, 2000). The third cell type belonging to the osteoblast family, the bone-lining cell, is regarded as an inactive osteoblast. These cells are flat, have a reduced cytoplasmic armamentarium for protein synthesis and secretion (Fig. 2c), and are connected to one another and to the osteocytes.
View larger version (220K):
[in this window]
[in a new window]
|
Figure 2. Transmission electron micrographs illustrating (a) an osteoblast, (b) an osteocyte, (c) a bone-lining cell, (d) AEFC-forming cementoblasts, (e) a CIFC-forming cementoblast, and (f) a cementocyte. AEFC, acellular extrinsic fiber cementum; CIFC, cellular intrinsic fiber cementum; and EF, extrinsic fibers.
|
|
The AEFC is formed by a class of cells that morphologically resemble periodontal ligament fibroblasts (Fig. 2d) (Beertsen and Everts, 1990; Bosshardt and Schroeder, 1991, 1993). This does not imply that the AEFC-producing cells are periodontal ligament fibroblasts. A comparison of the gene expression profile of cells lining the AEFC and periodontal ligament cells away from the root surface clearly shows a difference (MacNeil et al., 1994, 1996, 1998; D’Errico et al., 1997). Unlike bone, CIFC, and AIFC, the AEFC does not have a comparable cementoid seam. However, the principal periodontal ligament fibers, or at least their cementum-related portion, may be regarded as equivalent to the cementoid. Among all cementum varieties, CIFC morphologically resembles bone most of all. The cementoblasts engaged in the formation of both CIFC and AIFC are large, cuboidal cells with a round, euchromatin-rich nucleus and the full cytoplasmic armamentarium required for protein synthesis and export (Fig. 2e) (Furseth, 1969; Bosshardt and Schroeder, 1990, 1992). As is the case with osteoblasts in bone, the cementoblasts producing the intrinsic cementum types first elaborate a cementoid seam that later mineralizes. And, like osteocytes in bone, the cementocytes are trapped in the cementum matrix (Fig. 2f). As in bone, rapid formation and a multipolar mode of matrix deposition are held responsible for the entrapment of cells in the cementum matrix (Bosshardt and Schroeder, 1992). Analogous to bone, a canalicular system may also be present in the CIFC matrix. In contrast to bone, cementocyte lacunae in the deeper portions of the cementum layer often appear empty (Cheng et al., 1996; Grzesik et al., 2000). This may have to do with the surpassing of the critical distance for exchange of metabolites. In most species, cementum lacks a vascular system that can keep cementocytes alive in deeper portions of the matrix. In addition, cellularity is said to be lower in CIFC than in bone (Cheng et al., 1996; Grzesik et al., 2000).
A striking difference between bone and cementum is that bone is found from the tip of the toe to the crown of the head, whereas cementum normally forms only around teeth or occasionally on materials that are in contact with teeth. In contrast, woven bone formation is not dependent on a hard substrate. In intramembranous ossification, woven bone arises from connective tissue serving as a template. A further characteristic of a mineralized tissue is its growth rate. AEFC grows throughout life with a rate ranging between 0.005 and 0.01 µm/day (Bosshardt and Schroeder, 1996; Beertsen et al., 1997b; Bosshardt and Selvig, 1997). Growth rates for CIFC in human teeth do not seem to be known. In monkeys, CIFC grows up to 30 times faster than does AEFC (Bosshardt and Schroeder, 1990). The linear appositional rate for human lamellar bone is 1–2 µm/day, human parallel-fibered bone forms 3–5 times faster, and woven bone possesses the fastest growth rate (Schenk and Buser, 1998).
Thus, while osteoblasts and CIFC-forming cementoblasts look very much the same, the morphology of the AEFC-forming cementoblasts is different. For proper functioning of bone, a system consisting of osteoblasts, osteocytes, bone-lining cells, and osteoclasts is required. In contrast, cementum seems to function without such a complex cell system, albeit signals from the adjacent periodontal ligament likely influence cementoblast function. There is no evidence that cementocytes have a function in tissue homeostasis. Concerning the speed of formation, significant differences exist between various bone types, CIFC, and AEFC.
|
(III) EXTRACELLULAR MATRIX MACROMOLECULES OF BONE AND CEMENTUM
|
|---|
The matrices of bone and cementum are biological composites consisting of water, mineral, collagens, and non-collagenous macromolecules, the latter usually being referred to as non-collagenous proteins (NCPs) (Table). Apart from having structural functions, the matrices of bone and cementum also serve as a reservoir for growth factors and cytokines. The biochemical composition of bone, cementum, and dentin has recently been reviewed (Robey, 1996; Bosshardt and Selvig, 1997; Butler, 1998, 2000; Saygin et al., 2000a; Sodek and McKee, 2000; Butler et al., 2003). It should be kept in mind that, for a given tissue, the relative amounts of matrix constituents may vary with anatomical location (Aerssens et al., 1997) and the speed of tissue formation (Nanci, 1999; Bosshardt and Nanci, 2000). Furthermore, the use of both various analytical techniques and different species renders a compositional matrix comparison quite difficult.
The collagens play important structural and morphogenic roles (Hay, 1991). In mineralized tissues, they interact with various NCPs and provide a scaffold for the accommodation of the mineral crystals (Christoffersen and Landis, 1991). In bone and cementum, collagen is the major constituent of the extracellular matrix, comprising about 80–90% of the organic matrix. The predominant collagen form is type I, but collagen types III, V, VI, and XII are also present. There is significant research regarding the nature, biosynthesis, gene structure and expression, tissue localization, and functions of the NCPs. More than 20 NCPs have been identified in the mineralized connective tissues of bone, cementum, and dentin. Functions of these NCPs may include matrix deposition, initiation and regulation of mineralization, and matrix remodeling (Robey, 1996; Butler, 1998; Fisher et al., 2001; Sodek et al., 2002). It is now evident that certain NCPs thought to be tissue-specific are present in many other tissues as well. Examples include osteopontin (OPN), dentin sialoprotein (DSP), dentin phosphoprotein (DPP), dentin matrix protein 1 (DMP1), and certain enamel matrix proteins (EMPs). It may thus be argued that bone, cementum, and dentin may have a common set of NCPs, and that the uniqueness of a tissue resides in the combination of its NCPs (Butler et al., 2003). The NCPs of bone, cementum, and dentin comprise glycoproteins, proteoglcycans, plasma-derived proteins, and other proteins that cannot be assigned to one of these three groups. In addition, several growth factors are sequestered in the mineralized matrix. It is beyond the scope of this review to discuss the functions of all of these NCPs. The aim of this section is to find out whether there are differences in the presence and fine distribution of NCPs in bone and cementum. The limited information available on osteoblast/osteocyte factor 45 (OF45), bone acidic glycoprotein-75 (BAG-75), vitronectin (VN), and osteoadherin (OSAD) does not allow one to make a definite conclusion regarding tissue specificity.
Bone sialoprotein (BSP) and OPN are two major NCPs of bone and cementum (Bronckers et al., 1994; MacNeil et al., 1996; D’Errico et al., 1997; Bosshardt et al., 1998; Nanci, 1999). Ultrastructurally, these two phosphorylated glycoproteins have a similar distribution pattern in bone and cementum. They fill in the interfibrillar spaces. Strong immunoreactivity is observed in the AEFC, because its matrix possesses large interfibrillar spaces (Fig. 3a). The overall denser collagen fibril aggregation in CIFC and bone leaves fewer interfibrillar spaces. Consequently, immunoreactions are weaker in bone (Fig. 3b) and CIFC (Fig. 3c), as compared with AEFC. A further determinant of the amount of BSP and OPN in bone and cementum appears to be the speed of formation (Nanci, 1999; Bosshardt and Nanci, 2000).
View larger version (192K):
[in this window]
[in a new window]
|
Figure 3. Transmission electron micrographs of (a) acellular extrinsic fiber cementum (AEFC), (b) bone, (c) cellular intrinsic fiber cementum (CIFC), and (d) cementoid formation in the cervical portion of a porcine tooth after immunocytochemistry for bone sialoprotein (BSP, b,c), osteopontin (OPN, a), and amelogenin (AMEL, d).
|
|
DMP1 is an acidic phosphorylated NCP that is expressed in odontoblasts (George et al., 1993), osteoblasts, and osteocytes (MacDougall et al., 1998; Feng et al., 2003). The DMP1 is localized in osteocytes and their pericellular bone matrix, but not in osteoblasts (Toyosawa et al., 2001). Gene expression has (MacDougall et al., 1998; Feng et al., 2003) or has not been detected in cementoblasts (D’Souza et al., 1997). The DMP1 gene is expressed in AEFC-forming cementoblasts and in cementocytes, whereas the protein is expressed in cementoblasts and cementocytes (Butler et al., 2002; Feng et al., 2003). The protein is particularly located in the pericellular matrix (Toyosawa et al., 2004).
DSP and DPP are expressed as a single mRNA transcript that encodes for a precursor protein termed ‘dentinsialophosphoprotein’ (DSPP). Using reverse-transcriptase polymerase chain-reaction and Western blotting, investigators have demonstrated low levels of DSPP transcripts and DSP, respectively, in bone (Qin et al., 2002). Protein and gene expression of DSP has been shown in bone, osteoblasts, osteocytes, CIFC, CIFC-forming cementoblasts, and cementocytes, but not in AEFC and associated cells (Baba et al., 2004). Lack of DSP expression in AEFC is a further indication of the uniqueness of this cementum type.
Osteonection (ON or SPARC) is another glycosylated protein present in bone (Termine et al., 1981). Immunostaining of the bone matrix is moderate to strong, while that of the cementum matrix is weak (Tung et al., 1985) or negative (Reichert et al., 1992). Cementoblasts forming both AEFC and CIFC and cementocytes reacted with the anti-ON antibody (Reichert et al., 1992). The association of ON with the collagen fibrils of the bone matrix is consistent with the proposed role of ON in linking mineral to collagen fibrils (Bianco et al., 1985, 1988). There are still limited data available on cementum.
Fibronectin (FN) and tenascin are multifunctional glycoproteins found in many tissues (Jones and Jones, 2000; Pankov and Yamada, 2002). In bone, FN and tenascin have been localized in the mineralized bone matrix (Carter et al., 1991; Nordahl et al., 1995). FN is found in the pericellular matrix of osteocytes (Steffensen et al., 1992). In mature cementum, both proteins have been immunolocalized at the site where the principal periodontal ligament fibers enter the mineralized portion of cementum, but not in the cementum layer itself (Lukinmaa et al., 1991). In another study, both glycoproteins were immunodetected in the cementum matrix (Zhang et al., 1993). Recently, it was shown that staining for AEFC was negative for FN and tenascin, whereas CIFC showed a moderate and strong reactivity for tenascin and FN, respectively (Matias et al., 2003a).
Osteocalcin (OC), also known as bone GLA (gamma-carboxyglutamic acid-containing) protein, is a very small protein that is present in abundance in bone, cementum, and dentin (Hauschka and Wians, 1989). In bone, immunolabeling was observed in the mineralized matrix and also in osteoid (weakly) (Camarda et al., 1987; Bronckers et al., 1994). While immunostaining for OC in AEFC was negative (Bronckers et al., 1994; Kagayama et al., 1997), another study showed a positive reaction (Tenorio et al., 1993). For CIFC, all three studies showed a positive reaction, albeit in the study by Tenorio and co-workers (1993), staining intensity in CIFC was weaker than in AEFC. Bronckers and co-workers (1994) interpreted their data as suggesting that cementoblasts producing CIFC express an osteoblast-like phenotype, while cementoblasts involved in the formation of AEFC express only a partial osteoblastic phenotype and are phenotypically related to periodontal ligament cells. Differences between these studies may be explained by the use of different antibodies and/or different tissue-processing protocols. The fact that the CIFC matrix was always labeled may be related to the embedding of cells in its matrix.
Cementum-derived attachment protein (CAP) is a 56- or 65-kDa collagenous protein that promotes the attachment of mesenchymal cells to the extracellular matrix (McAllister et al., 1990; Wu et al., 1996). CAP is found in the developing (Saito et al., 2001) and mature cementum matrix and cementoblasts, with no positive reaction being observed over the alveolar bone (Arzate et al., 1992a, 1998; Wu et al., 1996). Anti-CAP antibodies stain isolated cementum tumor cells in vitro (Arzate et al., 1992b, 1998). Immunolabeling in vitro is localized to the cell membrane and fibril-like structures, and, in vivo, is found in cells associated with calcified bodies. A fraction of osteoblastic cells derived from the alveolar bone immunoreacts with the anti-CAP antibody. It has been suggested that CAP is a marker molecule for cementogenesis (Arzate et al., 1992a), and that CAP is related to the development of the cementoblast phenotype (Pitaru et al., 1993). In vitro, cells bound to CAP can form a cementum-like tissue (Liu et al., 1997). After expansion in vitro, bovine dental follicle cells implanted into immunodeficient mice formed a matrix positive for the anti-CAP antibody 3G9, whereas bovine alveolar bone osteoblasts, treated the same way, formed a bone-like matrix that was negative for the same antibody (Handa et al., 2002a,b). These studies suggest that CAP may be a suitable marker to differentiate between bone and cementum.
Proteoglycans (PGs) play a role in tissue formation and mineralization (Embery et al., 2001). Chondroitin sulfate, dermatan sulfate, heparan sulfate, hyaluronate, and keratan sulfate are the glycosaminoglycans identified in the alveolar bone and in cementum (e.g., Bartold et al., 1990; Murahashi et al., 1990; Cheng et al., 1999). Among the PG species detected in association with bone and cementum are lumican, fibromodulin, versican, decorin, biglycan, and osteo-adherin (Sommarin et al., 1998; Wendel et al., 1998; Ababneh et al., 1999; Cheng et al., 1999; Matias et al., 2003b; Petersson et al., 2003; Ramstad et al., 2003; Waddington et al., 2003). Human CIFC reacts with antibodies against lumican, versican, decorin, and biglycan, whereas the staining for fibromodulin is negative (Ababneh et al., 1999). In sharp contrast are the results in AEFC, where none of these antibodies produced an immunoreaction (Ababneh et al., 1999). These observations corroborate those from other studies showing that most PG species in bone and cementum are located peripheral to cell lacunae and canaliculi, structures not present in AEFC. Cheng and co-workers (1996) and Grzesik and co-workers (2000) used human-cementum-derived cells for expansion in vitro, followed by transplantation into immunodeficient mice. They noted that cementocytes embedded in the cementum-like matrix that formed at an ectopic site were immunopositive for fibromodulin and lumican, whereas osteocyte-like cells in an ectopically formed bone-like matrix originating from transplanted human bone marrow stromal cells were not (Grzesik et al., 2000). These results are consistent with the in situ situation where fibromodulin and lumican are more abundant in cementum than in bone (Cheng et al., 1996), and lend strong support to the hypothesis that cementum-producing cells are different from osteoblasts.
Enamel matrix proteins (EMPs) had been considered an integral structural entity of the root (Lindskog and Hammarström, 1982; Lindskog, 1982a,b; Slavkin et al., 1989; Sasano et al., 1992). However, detectable expression of EMPs along the tooth root is a rather occasional event (Fong et al., 1996; Bosshardt and Nanci, 1998, 2000, 2004; Bosshardt et al., 1998) (Fig. 3d). The idea of an interfacial matrix layer consisting of EMPs originated probably from ultrastructural studies showing that cells of the Hertwig’s epithelial root sheath (HERS) possess cytoplasmic organelles indicative of protein synthesis and secretion (Bosshardt and Schroeder, 1996; Bosshardt and Nanci, 2004). Apart from an occasional secretion of EMPs, other studies suggest that HERS cells produce cementum-related proteins such as BSP, OPN, and fibrillar collagen (Somerman et al., 1992; Bosshardt et al., 1998; Bosshardt and Nanci, 2000, 2004). The rationale to use EMPs for periodontal regeneration is based on the hypothesis that a distinct EMP-containing interfacial matrix layer exists, and that EMPs trigger cementoblast differentiation. Growth/differentiation factors are highly potent substances present in small quantities in extracellular matrices. This does not mean that specific EMP species may not have a role in cell differentiation. This aspect will be discussed in the section on ‘cell differentiation’.
It is unclear whether EMPs are present in the cementum or bone matrix. Although it has been shown that molecules in the bone matrix share a common epitope with EMPs (Inai et al., 1993), no concrete data on the presence of EMPs in bone are available. EMPs may be present in cementum (Slavkin et al., 1989). However, dentin contaminations must always be considered in the biochemical analysis of cementum, and there is evidence that EMPs are present in the dentin matrix (Inai et al., 1991; Nanci and Smith, 1992; Nakamura et al., 1994; Karg et al., 1997; Veis et al., 2000; Oida et al., 2002).
Growth/differentiation factors may not only be produced by the local cells, but can also be released from the mineralized matrix. The matrices of bone and cementum are rich sources of growth/differentiation factors that modulate the activities of many cell types, including those originating from the periodontium (Cochran and Wozney, 1999; Saygin et al., 2000a; Grzesik and Narayanan, 2002). Growth/differentiation factors identified in cementum include members of the transforming growth factor (TGF)-ß superfamily, like TGF-ß1, and various bone morphogenetic proteins (BMPs; BMP-2, BMP-3, BMP-4, BMP-7), platelet-derived growth factors (PDGF) a and b, fibroblast growth factors (FGFs), insulin-like growth factors I (IGF-I) and II (IGF-II), epidermal growth factor (EGF), and cementum-derived growth factor (CGF). Most of these polypeptide growth/differentiation factors are also sequestered in the bone matrix (Hauschka et al., 1988; Solheim, 1998). However, CGF, a 14-kDa protein, appears to be unique to cementum. It is an IGF-I-like molecule that is weakly mitogenic to fibroblasts and reveals a synergistic potentiation in the presence of EGF (Ikezawa et al., 1998).
From this section, we can conclude that bone, CIFC, and AEFC have a similar biochemical composition. However, both the speed of tissue formation and the organization of the collagenous matrix may influence the distribution pattern of certain NCPs. In this respect, AEFC is distinctively different from both CIFC and bone. Another differential factor is the presence or absence of embedded cells. AEFC, which lacks cementocytes, does not exhibit the same distribution pattern for a variety of NCPs, as compared with CIFC and bone. Labeling for fibromodulin and lumican appears to be more pronounced in CIFC than in bone, and CAP appears to be exclusively present in the cementum matrix. Thus, fibromodulin, lumican, and CAP may be used as markers to differentiate between bone and CIFC.
|
(IV) REGULATION OF OSTEOBLAST AND CEMENTOBLAST FUNCTION
|
|---|
While there is a large body of information available on the regulation of bone formation, comparatively less information is available on the regulation of cementoblast function (Saygin et al., 2000a; Grzesik and Narayanan, 2002). This has also to do with the difficulties involved in obtaining a pure cementogenic cell lineage for experimental manipulations. Throughout life, the skeleton undergoes continuous remodeling, which serves the purpose of repair and mechanical adaptation. The consequences of an imbalance in the expression of signaling molecules are metabolic bone disorders or diseases like Paget’s disease, osteopetrosis, osteoporosis, arthritis, or bone loss in periodontitis (Suda et al., 1997; Mogi et al., 2004). Regulation of bone remodeling is under both systemic and local control. Local factors are operative in a paracrine and autocrine fashion, and osteoblasts, osteoclasts, and inflammatory/immune cells function as both sources and targets of signaling molecules. Numerous cytokines and growth factors have anabolic and/or catabolic effects on bone formation (Aubin, 2001; Harada and Rodan, 2003). Among these bone-regulatory molecules are parathyroid hormone (PTH), PTH-related peptide (PTHrP), calcitonin (CT), calcitriol (the active form of vitamin D), prostaglandin E2 (PGE2), growth hormone (GH), thyroid hormone, sex steroids (estrogen and testosterone), leptin, statins, interferon-
, tumor necrosis factor (TNF)-
, transforming growth factor (TGF)-, TGF-ß, BMPs, FGF, IGF-I, PDGFs, and interleukins (IL)-1, -6, -11, and -17.
A breakthrough in bone research with a potential for therapeutic interventions was the discovery of the RANK/RANKL/OPG system that is essential for bone regulation, and that links bone biology with immune cell biology (Grcevic et al., 2001; Hofbauer and Heufelder, 2001; Goldring, 2003). Osteoprotegerin (OPG), a member of the tumor necrosis factor receptor (TNFR) superfamily, has an important protective function for bone. It functions as a decoy receptor by binding to RANK ligand (RANKL). Thus, the prevented interaction between receptor activator of nuclear factor-kappaB (RANK) and RANKL suppresses both development and activity of osteoclasts. Cementum in human teeth does not undergo physiological remodeling, yet root resorption is a common finding in orthodontically treated teeth and pathologic processes like periodontitis. In this context, an interesting question is whether the RANK/RANKL/OPG system operates in cementoblasts as it does in osteoblasts. The expression of OPG and RANKL has been detected in periodontal ligament cells in vitro (e.g., Hasegawa et al., 2002; Sakata et al., 2002; Zhang et al., 2004), during physiological root resorption (Lossdörfer et al., 2002; Fukushima et al., 2003), and during experimental tooth movement (Oshiro et al., 2002). Whether cementoblasts are contributing to the production of OPG and RANKL is unknown. However, of interest in this context is that amelogenin-null mice display a reduced protection against root resorption, which is associated with an elevated expression of RANKL (Hatakeyama et al., 2003). Thus, the RANK/RANKL/OPG system seems to be involved in the regulation of resorption of both bone and cementum.
The PTH-dependent signaling pathways in bone cells have recently been reviewed (Swarthout et al., 2002). Cementoblasts express receptors for PTH/PTHrP (Tenorio and Hughes, 1996) and the PTHrP gene (Beck et al., 1995). PTH/PTHrP may have a regulatory role in cementogenesis (Robinson and Harvey, 1989; Takeuchi et al., 1989; Ouyang et al., 2000). The detection of PTH-binding sites in CIFC-forming cementoblasts and the lack of this signal in AEFC-forming cementoblasts support the contention of a phenotypic similarity between osteoblasts and CIFC-forming cementoblasts (Tenorio and Hughes, 1996). Since the major function of CIFC is to build up a thick mineralized layer in a short period of time, it can be expected that its cells respond to local factors in a much more pronounced way than do the AEFC-forming cementoblasts. Vitamin D- and calcium-deficient diet leads to increased resorptive root lesions and decreased mineralization of cementum (Bielaczyc and Golebiewska, 1997). Both osteoblasts and cementoblasts respond to vitamin D deficiency with an increase in BSP gene expression and a decreased OPN gene expression (Chen et al., 1999). The vitamin D receptor has been immunolocalized in some root-lining cells only during root development (Onishi et al., 2003). Controversial effects of exogenous PGE2 were observed in the development of resorptive root lesions in association with experimental tooth movement (Boekenoogen et al., 1996). This may have to do with the dual role of PGE2, which is also observed in bone (Kawaguchi et al., 1995). GH is a regulator of bone mass (Olney, 2003). While CIFC is highly responsive to GH, AEFC appears to be less affected (Clayden et al., 1994; Li et al., 2001; Smid et al., 2004). This difference may depend on the expression of the GH receptor, which is much more strongly expressed in CIFC- than in AEFC-forming cementoblasts (Zhang et al., 1992). In cementoblasts, IGF-I, PDGF-BB, and TGF-ß influence mitogenesis and phenotypic gene expression profiles (Saygin et al., 2000b). That cementoblasts carry the IGF1 receptor strengthens the proposed role for components of the IGF system in modulating cementoblast function (Götz et al., 2001).
Thus, while bone remodeling needs to be tightly coupled, protective mechanisms to prevent root resorption should exist. Both osteoblasts and CIFC-forming cementoblasts respond to many factors that regulate cell activity similarly. This does not prove that they belong to the same cell lineage. In contrast, AEFC-forming cementoblasts respond mildly to regulatory molecules.
|
(V) REGULATION OF OSTEOBLAST AND CEMENTOBLAST DIFFERENTIATION
|
|---|
The regulation of osteoblast differentiation has been reviewed (Karsenty, 2000a; Yamaguchi et al., 2000; Aubin, 2001; Katagiri and Takahashi, 2002). Attempts have been made to elucidate the differentiation pathway of the cementoblast cell lineage (Bosshardt and Schroeder, 1996; Bosshardt and Selvig, 1997; Saygin et al., 2000a; Grzesik and Narayanan, 2002). The influence of the local environment on cells during periodontal wound healing and regeneration has recently been reviewed (Grzesik and Narayanan, 2002).
Monoclonal antibodies have been generated as tools for studying the osteoblast lineage (Aubin and Turksen, 1996). Some of these antibodies show a selective reactivity against cells at different stages, i.e., against osteoprogenitors, pre-osteoblasts, osteoblasts, or osteocytes (Walsh et al., 1994; Aubin and Turksen, 1996; Bruder et al., 1997). The antibody E11 recognizes an antigen located at the cell surfaces of osteoblasts, pre-osteocytes, and young osteocytes (Wetterwald et al., 1996), and may be used as a marker for the late steps in the differentiation pathway of the osteoblastic lineage (Schulze et al., 1999). In the tooth, the E11 antibody reacts strongly with CIFC-forming cementoblasts and young cementocytes, but not with the AEFC-forming cementoblasts (Tenorio et al., 1993). A transcription factor essential for differentiation of osteoblasts from progenitor cells is Runx2 (runt-related transcription factor 2)/Cbfa1 (core-binding factor alpha 1) (Ducy, 2000; Karsenty, 2000b). Expression of Cbfa1 in fully differentiated cells suggests additional roles in osteoblast function. Cbfa1 expression is not restricted to cells of the osteogenic/chondrogenic lineage (Otto et al., 1997; Bronckers et al., 2001). It has been detected in odontoblasts, dental follicle cells, periodontal ligament cells, cementoblasts, cementocytes, in dental epithelial cells, and in some non-dental cell types. EGF is another signaling molecule implicated in regulating periodontal cell differentiation. The EGF receptor (EGFR) regulates the signal transduction from ligands such as EGF and TGF- (Wells, 1999). EGFR is down-regulated in the course of cell differentiation, and the EGF/EGFR system may act as a negative regulator of osteoblastic differentiation in the periodontal ligament (Matsuda et al., 1998; Chien et al., 2000). The role of EGF in the differentiation of cementoblasts is unclear (Cho et al., 1991; Cho and Garant, 1996; Sismanidou et al., 1996).
An understanding of the cell differentiation mechanisms in the periodontal ligament is imperative for the development of predictable regenerative techniques. It is important to understand that the situation in the periodontium is very special. The periodontal ligament represents a unique non-mineralizing connective tissue, and, at the same time, is the source for mineralized-tissue-forming cells (Karring et al., 1985; Pitaru et al., 1994; McCulloch, 1995; Beertsen et al., 1997a; McCulloch et al., 2000). The cell population of the periodontal ligament consists of fibroblastic subpopulations, osteoblasts, cementoblasts, endothelial cells, perivascular cells, and epithelial cells. The periodontal ligament harbors progenitor cells that can differentiate into periodontal ligament fibroblasts, osteoblasts, and cementoblasts. It is not known, however, whether there is one common progenitor cell type for all three specialized connective tissue cells, or whether distinct progenitor subpopulations exist. An enamel matrix derivative (EMD; Biora, Malmö, Sweden) is in clinical use to promote periodontal regeneration. The rationale for the use of this product is the assumption that EMPs, synthesized and secreted by HERS cells, are trigger molecules for the differentiation of dental follicle cells into AEFC-forming cementoblasts (Hammarström, 1997). Initial results from selective acute-type periodontal defect models (i.e., buccal bone dehiscence, and tooth extraction and re-implantation) in animals (Hammarström et al., 1997; Hammarström, 1997) and in one human tooth (Heijl, 1997) supported this assumption, but these experiments had little to do with periodontal regeneration in periodontitis-affected teeth. In sharp contrast to these findings are clinically more relevant studies. Histologic studies of EMD used in teeth affected by periodontitis show that the mineralized tissue formed on the treated root surface is predominantly cellular (e.g., Sculean et al., 1999, 2003; Yukna and Mellonig, 2000; Cochran et al., 2003; Donos et al., 2003; Bosshardt et al., 2005) (Figs. 4a, 4b). The mineralized tissue may be classified as cementum-like, but it could be bone-like as well. There is no doubt that it is relevant to know what type of tissue forms on the root under the influence of EMD. Bone undergoes physiologic remodeling, a process that ultimately leads to substitution of tooth-root substance by bone tissue, as is illustrated in ankylosed teeth.
View larger version (99K):
[in this window]
[in a new window]
|
Figure 4. (a) Light and (b) transmission electron micrographs showing the new mineralized tissue (NT) that formed on the root surface 5 wks after the application of Emdogain® in a human tooth affected by periodontitis. Note the high cellularity of the ‘regenerated’ tissue (arrowheads, a) and the large, cuboidal cell forming this tissue and exhibiting a cytoplasm filled with abundant rough endoplasmic reticulum (b). CF: collagen fibrils.
|
|
While evidence for a causal link between EMPs and cementogenesis is lacking, evidence is emerging that specific EMP variants have chondrogenic/osteogenic activities. The capacity of demineralized bone (Urist, 1965) and dentin matrix (Bang and Urist, 1967) to induce ectopic bone formation has been known for many years, and was attributed to the activity of molecules that were later referred to as BMPs. The observation that dentin matrix implants are even more potent than bone matrix implants in osteoinduction (Urist and Strates, 1971; Somerman et al., 1987; Kawai and Urist, 1989; Veis et al., 1989) could mean that there is a heightened content of one or more BMP variants in the dentin matrix or, alternatively, that additional chondrogenic/osteogenic proteins are present in dentin (Bessho et al., 1990, 1991; Amar et al., 1991). The active fraction of the chondrogenic/osteogenic proteins extracted from bovine dentin has been identified as a small splice product of the amelogenin gene (Nebgen et al., 1999). Thereafter, two specific cDNAs from a rat incisor tooth odontoblast pulp cDNA library were identified and the corresponding recombinant proteins, r[A+4] and r[A–4], produced (Veis et al., 2000). In vitro and in vivo, both amelogenin polypeptides showed chondrogenic and osteogenic activities, respectively. Demineralized enamel (Urist, 1971; Kawai and Urist, 1989) and amelogenin in combination with plaster of Paris, used as a carrier (Wang, 1993), also show bone-inductive activity. Experiments with EMD have not consistently shown osteogenic effects (Boyan et al., 2000; Schwartz et al., 2000; Shu et al., 2000; Jiang et al., 2001; Kawana et al., 2001; Dean et al., 2002; Ohyama et al., 2002; Sawae et al., 2002; Shimizu-Ishiura et al., 2002; Mizutani et al., 2003; Yoneda et al., 2003; He et al., 2004). The inconsistency of the results may have to do with the different in vitro and in vivo experimental designs, or with the EMD itself. It is a crude extract of enamel matrix consisting mainly of amelogenins (Gestrelius et al., 1997a, b; Hammarström, 1997; Maycock et al., 2002), but it might also contain BSP (Suzuki et al., 2001) and TGF-ß1 or TGF-ß-like substances (Kawase et al., 2001). EMD may also up-regulate the synthesis of TGF-ß1, IL-6, PDGF-AB, FGF-2, prostaglandin G/H synthase, PGE2, and OPG (Van der Pauw et al., 2000; Lyngstadaas et al., 2001; Dean et al., 2002; Mizutani et al., 2003; He et al., 2004), factors that may modulate cell differentiation and activity. Veis postulated that specific amelogenin splice products may function as potential epithelial-mesenchymal signaling molecules during tooth development (Veis, 2003). During crown formation, reciprocal epithelial-mesenchymal interactions induce the differentiation of papillary cells into odontoblasts, and those of inner enamel epithelial cells into ameloblasts. The absence of an osteogenic differentiation pathway may be due to the cell origins or the interactions with other signaling molecules, such as members of the TGF-ß superfamily, DSP, DPP, ameloblastin, FGFs, IGF-1, and/or IL-7 (Martin et al., 1998; Coin et al., 1999; MacDougall et al., 2000; Papagerakis et al., 2002, 2003). Veis’ group could also show that the amelogenin polypetides [A+4] and [A–4] have differential effects, one of which is the A-4-dependent up-regulation of CAP, a cementum-specific matrix macromolecule (Tompkins and Veis, 2002).
Thus, although the differentiation mechanisms of osteoblasts have been studied extensively, not as much is known about cementoblast differentiation. The use of EMPs as a therapeutic agent to induce mineralized tissue formation has attracted much attention in recent years. While there is no evidence for a causal connection between EMPs and cementogenesis, the chondrogenic/osteogenic potential of EMP constituents has been apparent for many years. In this regard, specific, small amelogenin polypeptides are very potent.
|
(VI) CELL ORIGINS OF OSTEOBLASTS AND CEMENTOBLASTS
|
|---|
In the periodontium, osteoblast precursors may originate from the dental follicle proper and/or the perifollicular mesenchyme (Ten Cate, 1997; Cho and Garant, 2000), but other sources—like pericytes, adipocytes, periosteum, and bone marrow stem cells—must also be considered (McCulloch et al., 1987; Devlin and Sloan, 2002). A major drawback for progress in the development of biologically active molecules for tissue engineering is the lack of knowledge about the location and origin of cemento-progenitors. If periodontal regeneration recapitulates tooth development, the origin of cementoblasts, periodontal fibroblasts, and osteoblasts during root formation needs to be known. The origin of cementoblast precursors and the molecular factors regulating their differentiation are not understood (Pitaru et al., 1994; Bosshardt and Schroeder, 1996; Bosshardt and Selvig, 1997; Saygin et al., 2000a). According to the classic theory, the dental follicle proper, and perhaps also the perifollicular mesenchyme, give rise to periodontal ligament fibroblasts, osteoblasts, and cementoblasts (Ten Cate, 1997; Cho and Garant, 2000). Most questionable is the theory about cementoblast origin, which is mainly based on transplantation and 3H-thymidine labeling studies (Bosshardt and Schroeder, 1996; Bosshardt and Selvig, 1997; Ten Cate, 1997). The detection of label in cementoblasts was interpreted as evidence of an origin from the dental follicle proper. However, as pointed out by Thomas and Kollar (1988) and Bosshardt and Selvig (1997), labeled cementoblasts could also originate from the HERS, since its cells also incorporated 3H-thymidine. An alternative is that cells of the HERS undergo an epithelial-mesenchymal transformation (Thomas and Kollar, 1988; Bosshardt, 1992; MacNeil and Thomas, 1993; Beck et al., 1995; Davideau et al., 1995; Thomas, 1995; Bosshardt and Schroeder, 1996; Webb et al., 1996; Bosshardt and Selvig, 1997; Bosshardt and Nanci, 1997, 1998, 2000, 2004; Bosshardt et al., 1998; Terling et al., 1998; Lezot et al., 2000). The origin of cementoblasts from the HERS would be a plausible explanation for the development of a cementoblast phenotype that is distinctly different from an osteoblast. Moreover, HERS may not only give rise to cementoblasts. According to a new hypothesis proposed here, HERS may also be the source of a special subpopulation of fibroblasts that contributes to the pool of periodontal ligament fibroblasts (Fig. 5). This hypothesis is based on morphological observations made in the region of the growing root in a large number of human and porcine teeth. Is there any circumstantial evidence supporting this hypothesis? The manner in which disintegrating and presumably transforming HERS cells occupy an increasingly widening space, starting from the root surface and extending into the forming periodontal ligament, has been described (Bosshardt and Nanci, 2004) and is illustrated in Fig. 5. Furthermore, the fact that periodontal ligament fibroblasts express cytokeratin 19 (Moxham et al., 1998), known to be expressed in HERS and the epithelial rests of Malassez (ERM) (Peters et al., 1995), and the presence of simplified desmosomes (Shore et al., 1981) and desmosomal proteins (Yamaoka et al., 1999) between periodontal ligament fibroblasts may be explained by the hypothesis that the HERS contains ancestral cells for a subpopulation of periodontal ligament fibroblasts.
View larger version (61K):
[in this window]
[in a new window]
|
Figure 5. Schematic drawing illustrating a hypothetical model of the contribution of the Hertwig’s epithelial root sheath (HERS) to the periodontal ligament cell populations. After its fragmentation, HERS gives rise to the epithelial rests of Malassez, but may also be the source of cementoblasts and other mesenchymal cells that populate the periodontal ligament. The illustration shows an early stage of root formation that is associated with the formation of acellular extrinsic fiber cementum.
|
|
What could be the consequences of this new theory for periodontal repair/regeneration? The compartmentalization of the periodontal ligament into subregions containing enriched subpopulations of specific cells committed to give rise to particular cell phenotypes would facilitate tissue homeostasis and regeneration. Within the cementum-related domain, three major cell types may be encountered: (1) the cementoblasts, which line the cementum surface; (2) the ERM, which form a fenestrated epithelial demarcation against the remainder of the periodontal ligament; and (3) the specific HERS-derived fibroblast subpopulation, which populates the cementum-related domain of the periodontal ligament. Thus, the following hypothetical scenarios are conceivable: (1) By recapitulating tooth developmental events (i.e., epithelial-mesenchymal transformation of HERS into cementoblasts), the ERM may directly give rise to new cementoblasts; 2) by recapitulating epithelial-mesenchymal interactions, the interaction of ERM with a specific periodontal ligament fibroblast subpopulation (either derived from the dental follicle proper or from HERS, or perivascular cells) may generate new cementoblasts; and (3) the special subpopulation of HERS-derived periodontal ligament fibroblasts may be the source of new cementoblasts. Is there any evidence to support one or more of these scenarios? Attempts to associate ERM with cementum repair have been made (Brice et al., 1991; Wesselink and Beertsen, 1993; Bosshardt and Schroeder, 1994), but failed to show a causal connection until the expression of BMP-2, OPN, ameloblastin, and proliferating cell nuclear antigen was shown in the ERM during reparative cementogenesis (Hasegawa et al., 2003). Thus, the ERM are likely involved in repair/regenerative cementogenesis. Numerous other studies suggest that the ERM represent more than a merely vestigial structure in the periodontal ligament (Kvam and Gilhuus-Moe, 1970; Sismanidou et al., 1996; Kagayama et al., 1998; Mouri et al., 2003; Yamashiro et al., 2003; Talic et al., 2004).
Thus, both osteoblasts and cementoblasts may originate from one ancestral source: the dental follicle proper. Alternatively, osteoblasts and cementoblasts may not have the same origin within the developing dental and periodontal tissues. The likely origin of osteoblasts from the dental follicle (with a possible contribution of the perifollicular mesenchyme) and the proposed origin of both cementoblasts and a subset of periodontal ligament fibroblasts from the HERS would, at least partly, explain cell diversity within the periodontal ligament. Furthermore, the establishment of clearly defined domains within the periodontal ligament would facilitate homeostasis, repair, and the guidance of tissue regeneration, so that cementum forms on the root surface and bone on the opposite side of the periodontal ligament.
|
CONCLUDING REMARKS
|
|---|
Several lines of evidence are presented to suggest that AEFC- and CIFC-forming cementoblasts are unique phenotypes. While AEFC truly represents a unique tissue, CIFC matrix resembles bone. CIFC, however, is morphologically, functionally, and biochemically different from any bone type. Consequently, the question arises as to how exactly cell diversity evolves in the developing periodontal ligament. In this context, it is imperative to determine, unequivocally, both the origin of cementoblasts during development and repair/regeneration and the nature of the cell differentiation factor(s), to propose experimental strategies aimed at developing more effective and predictable periodontal therapies. Compiling evidence suggests that HERS may be the source of cementoblasts. It has also been suggested that AEFC-forming cementoblasts and CIFC-forming cementoblasts may not be of the same origin. Does this now mean that the proposed epithelial-mesenchymal transformation of some HERS cells may apply to one cementoblast subtype only? If the theory about the proposed transformation of HERS is true, then, in my opinion, it applies to both AEFC- and CIFC-forming cementoblasts. This, in turn, raises questions about the molecular factors that discriminate between different cementoblast subtypes. The proposed function of EMPs in selectively guiding cell differentiation toward an AEFC-forming cementoblast has been contradicted by many histological studies. Nevertheless, specific EMPs most likely have a role in cell differentiation, as demonstrated for small amelogenin polypeptides, which have strong chondrogenic and osteogenic activity. Studies are needed to explore the role of EMPs in progenitor cell differentiation, so that this information can be used predictably to engineer bone or cementum or both.
Received May 7, 2004;
Accepted October 10, 2004
|
REFERENCES
|
|---|
Aarden EM, Burger EH, Nijweide PJ (1994). Functi
TOP
ABSTRACT
INTRODUCTION
