The European Journal of Orthodontics Advance Access originally published online on May 10, 2006
The European Journal of Orthodontics 2006 28(3):221-240; doi:10.1093/ejo/cjl001
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The tissue, cellular, and molecular regulation of orthodontic tooth movement: 100 years after Carl Sandstedt
Department of Oral Sciences, Faculty of Dentistry, University of Otago, Dunedin, New Zealand
Address for correspondence Professor Murray C. Meikle, Faculty of Dentistry, University of Otago, P.O. Box 647, Dunedin, New Zealand. E-mail: murray.meikle{at}dent.otago.ac.nz
 |
Summary |
|---|
Top
Summary
IntroductionThe first experimental investigation of orthodontic tooth movement was published by Sandstedt in 1904–1905. After 100 years, there is a good understanding of the sequence of events at both tissue and cellular levels and now the current focus of research is at the molecular level. The techniques of reverse transcription–polymerase chain reaction and in situ hybridization to detect mRNAs of interest have revolutionized tooth movement studies and an expanding list of antibodies and enzyme-linked immunosorbent assays directed against human and animal proteins will facilitate their identification in tissue sections and/or culture supernatants. Nevertheless, although this technology has greatly simplified research for the clinical and laboratory investigator, message is not always translated into protein, and the presence of a protein does not necessarily mean it is biologically active.
In vivo and in vitro methods have been widely used in tooth movement studies. However, data from in vitro models, in which the mechanical stimulus can be carefully controlled (tension versus compression; intermittent versus continuous), should be correlated with in vivo data from animal models. The current evidence suggests that downstream from the initial mechanotransduction event at focal adhesions which link the extracellular matrix to the cytoskeleton, mechanically induced remodelling is mediated by a complex feedback mechanism involving the synthesis of cytokines such as interleukin-1 (IL-1), IL-6, and receptor activator of nuclear factor k B ligand by cells of the osteoblast and/or fibroblast lineages. These in turn act in an autocrine/paracrine fashion to regulate the expression of transcription factors, cytokines, growth factors, enzymes, and structural molecules involved in the differentiation, proliferation, and function of mesenchymal and other cell types. Contrary to the impression gained from the literature, tooth movement is not confined to events within the periodontal ligament. Orthodontic tooth movement involves two interrelated processes: (1) deflection or bending of the alveolar bone and (2) remodelling of the periodontal tissues.
|
Introduction |
|---|
It is now 100 years since a three-part article on the theory of tooth movement was published (Sandstedt, 1904, 1905
). It seems appropriate therefore to review the progress in our understanding of the tissue, cellular, and molecular mechanisms involved in orthodontic tooth movement since that time. The literature is now extensive (for previous reviews, see Reitan, 1951; Davidovitch, 1991; Sandy et al., 1993), which creates difficulties when attempting to present a coherent narrative. To add some structure to the subject, it will be addressed largely chronologically and from the biological viewpoint, with a focus on how knowledge has evolved with the discovery of new molecules and the introduction of new experimental techniques. This approach shows that, in common with the rest of the biological sciences, the methodology of tooth movement research has become progressively more reductionist, making the subject less accessible to the clinician.
The significance of mechanical stimuli in the maintenance and structure of skeletal tissues has been recognized since the middle of the 19th century (von Meyer, 1867; Wölff, 1892), and is the traditional starting point for any discussion of bone remodelling and tooth movement. The skeleton is continuously remodelled during life by osteoclasts resorbing old bone and osteoblasts subsequently forming new bone. Bone resorption and bone formation are therefore said to be coupled, a process of renewing the skeleton while maintaining its structural integrity (Frost, 1963). Remodelling is orchestrated by cells of the osteoblast lineage and involves a complex network of cell–cell and cell–matrix interactions involving systemic hormones, locally produced cytokines, growth factors, many of which are sequestrated within the bone matrix, as well as the mechanical environment of the cells.
Orthodontic and orthopaedic theory and practice have much in common. Both involve the remodelling of bone and both require a thorough understanding of bone biology, particularly the relationship between mechanical stress and the various cell types in bone. However, tooth movement is a more complicated process requiring changes in the periodontal ligament (PDL) as well as the supporting alveolar bone, tissues with different cell populations and remodelling characteristics. During locomotion and other forms of physical activity, the skeleton is loaded intermittently. Alveolar bone is also loaded intermittently during mastication, but is subject to a continuous deformation or strain during orthodontic tooth movement, and if the applied force is of sufficient magnitude, bending of the alveolus will occur. In the skeleton, the relationship between mechanical stress and bone resorption is poorly understood and the application of mechanically induced strain appears to be primarily an osteogenic stimulus (H
rt et al., 1971; Lanyon and Baggott, 1976). Furthermore, the osteogenic response of bone to external loading appears to be due to a reactivation of quiescent bone-lining cells (Pead et al., 1988; Chambers et al., 1993), and not dependent on a preliminary phase of bone resorption (Chow et al., 1998). Alveolar bone in contrast undergoes both resorption and deposition during orthodontic tooth movement, the extent of which is dependent upon the magnitude, direction, and duration of the applied force.
|
Tissue reaction to orthodontic tooth movement |
|---|
The idea that orthodontic tooth movement is dependent on the resorption and deposition of the bone of the socket dates back at least to 1839 (Harris, 1839
). It was not until the turn of the 20th century, however, that the original histological investigation that forms the foundation of our present knowledge was carried out by Sandstedt (1904, 1905) on dogs and published shortly after his death (Persson, 2005).
Laying the foundations: Sandstedt and Oppenheim
In Sandstedt's experimental model, a labial arch was bent to engage the six maxillary incisors of a dog and inserted into horizontal tubes attached to bands on the canines. The appliance was activated over a 3-week period by screws distal to the buccal tubes and during that time the crowns of the incisors were moved lingually by 3 mm. Sandstedt found that bone was deposited on the alveolar wall on the tension side of the tooth with both heavy and light forces, and that the newly formed bone spicules followed the orientation of the periodontal fibre bundles. On the pressure side, with light forces, alveolar bone was resorbed directly by numerous multinucleate osteoclasts in Howship's lacunae (Figure 1). With heavy forces, the periodontal tissues were compressed, leading to capillary thrombosis, cell death, and the production of localized cell-free areas of what he called hyalinization (owing to its glasslike appearance resembling hyaline cartilage in histological sections). At these sites, osteoclastic resorption of the adjacent alveolar wall did not take place directly, but was initiated by a process referred to by Sandstedt as ‘undermining resorption’ from the neighbouring marrow spaces.
|
Oppenheim (1911, 1930
) published the results of experimental work carried out on the primary teeth of monkeys. His observations proved to be substantially different from those of Sandstedt as far as the response of the bone to compression was concerned. He found that where a tooth had been tipped labially, the original bone disappeared completely from the labial surface and was replaced by new bone. He concluded that... bone tissue, be it compact or cancellated, reacts to pressure by a transformation of its entire architecture; this takes place by resorption of the bone present and deposition of new bone tissue; both processes occur simultaneously. Deposition finally preponderates over resorption.
Schwarz (1932) attempted to explain the difference between the findings of Sandstedt and Oppenheim by the fact that Oppenheim had killed his animals several days after the appliance had been last activated and did not therefore see the acute effect of the applied force, only a stage of regeneration after the force had been exhausted. It seems more likely that what Oppenheim was describing was the response of the periosteal and endosteal bone surfaces to the bending of the labial alveolar plate. The labial alveolar bone is particularly thin in his illustrations and would have been easily deformed in a monkey in the primary dentition (Figure 2). As discussed later, when bones are subjected to continuous mechanical deformation, concave surfaces are characterized by osteogenesis, and convex surfaces by bone resorption.
|
Whatever the explanation Oppenheim had lectured at the Angle School of Orthodontics and his research reported in the American Orthodontist, a journal published by the Angle Alumni Society (Oppenheim 1911, 1912
). His theory of bone transformation also supported Angle's (1907) non-extraction philosophy and the widespread belief at the time that orthodontic appliances could ‘grow bone’. Oppenheim's previous research on monkeys was reported again in 1930 together with the original photomicrographs and the findings of a new series of experiments carried out on dogs (Oppenheim, 1930). He also published articles on root resorption and repair in extracted human premolars following experimental tooth movement (Oppenheim, 1935, 1936). As a result, it was Oppenheim's work that appeared in mainstream English language orthodontic textbooks up until the 1950s and beyond. Sandstedt's early death and the lack of an English translation of his work did not help. He consequently became the forgotten man of tooth movement research until partially rescued from obscurity in the paper by Schwarz (1932).
Reitan and the post-war period
Following the work of Sandstedt and Oppenheim, reports by other investigators appeared in the literature (Gottlieb and Orban, 1931; Schwarz, 1932; Breitner, 1940), but it was not until the 1950s that tooth movement studies attracted wider attention. The leading histomorphological investigator of the period was the Norwegian orthodontist Kaare Reitan whose classic memoir was published in 1951 (Reitan, 1951).
Reitan (1957, 1964) made extensive use of human material, particularly premolars that were destined for orthodontic extractions. His work highlighted the complexity of the tissue response to orthodontic treatment depending upon (1) the type (continuous versus intermittent) and magnitude of the force applied, (2) the mechanics involved (tipping versus bodily movement), and (3) the variation in tissue reaction between individual patients. He observed that during the initial stages of a tipping movement, cell-free or hyalinized areas were frequently created with a continuous force of 30 g. The time taken to remove such tissue by undermining resorption varied from 2 to 4 weeks and occasionally longer depending on the length of the root. Intermittent forces of 70–100 g were also found to produce hyalinization. Cell-free areas were more common in tipping than in bodily movements presumably because, in the latter, the force was more evenly distributed along the root–bone interface (Figure 3). These experiments showed that even with an applied force as low as 30 g, some degree of hyalinization and root resorption appeared inevitable.
|
In a study of rotated teeth in young dogs, Reitan (1959)
found that while the principal fibres of the PDL were rearranged or remodelled within 28 days, even after a retention period of 232 days, some of the free gingival fibres remained displaced and stretched. He concluded that rotational relapse was caused primarily by a contraction of the supra-alveolar gingival fibres (which contain elastic or oxytalin fibres) and advised the over-rotation and/or transection (pericision) of these fibres during retention to ensure tooth stability.
The concept of differential force
Storey (1955a,b,c) who, in addition to analysing the histological changes produced by torsion springs applied to rodent and largomorph incisors, also carried out a series of experiments with canine retraction springs to determine what force levels should be used in clinical practice (Storey and Smith, 1952; Smith and Storey, 1952). In an investigation involving nine patients, they found that movement of the canine teeth into premolar extraction sites occurred rapidly when the value of the applied force was in the range of 150–250 g (5–9 ounces); however, below 150 g, the canines did not move significantly. When the springs were activated to apply forces in the range of 400–600 g (14–21 ounces), the anchor teeth (molars and second premolars) moved forward, with the canines remaining relatively stationary.
These experiments gave rise to the differential force concept and the idea that there is an optimum range of force values that will produce the maximum rate of tooth movement. Subsequent research, however, found that the rate of canine retraction using the forces recommended by Storey and Smith (1952) and Smith and Storey (1952) was highly variable between individual patients (Hixon et al., 1970; Boester and Johnston, 1974). This does not necessarily invalidate the concept as implied at the time, but indicates that the optimal force will be different for each patient. In any event, the magnitude of the applied force is just one of the many variables affecting the rate of tooth movement.
Autoradiographic investigations
One of the key technical developments during the 1960s as far as histomorphometry is concerned was the introduction of autoradiography. Tritium-labelled molecules, in particular, enabled, for the first time, changes in cell proliferation and metabolic activity to be measured with reasonable accuracy. Using a rat molar tooth movement model, Baumrind (1969) and Baumrind and Buck (1970) reported that cell proliferation (measured by 3H-thymidine incorporation) and metabolic activity (3H-uridine incorporation) were increased, and protein synthesis (3H-proline incorporation) was decreased, on both the ‘tension’ and ‘pressure’ sides of the PDL. This led them to question whether significant differences between the two sides actually existed.
Roberts and Jee (1974), Roberts et al. (1974), and Smith and Roberts (1980) used the same model to study the cellular kinetics of 3H-thymidine incorporation into PDL cells at tension sites. Smith and Roberts (1980) reported that over a time course of 20 hours, a continuous force produced a three-stage proliferative response. An interesting finding was a burst of mitotic activity within 2 hours, suggesting that the initial effect of mechanical strain was to allow G2-blocked cells to enter the cell cycle and undergo mitosis, as well as G1-blocked cells to synthesize DNA. The application of mechanical stress to the rat model also suggested that under strained conditions the cells of the PDL were primarily osteogenic (Roberts and Chase, 1981). This observation was supported by subsequent in vitro studies. PDL fibroblasts are functionally heterogeneous and contain a subpopulation of cells able to produce the osteoblast-related matrix proteins osteopontin, alkaline phosphatase (ALP), and bone sialoprotein (BSP; Lekic et al., 2001; Murakami et al., 2003).
Hyalinization and root resorption
The Scandinavian tradition of movement research was continued by Kvam (1972) and Rygh (1972a,b), with particular emphasis on the cellular and tissue changes on the compression side. It had been realized by Sandstedt that hyalinization was related to changes in vasculature, and Gianelly (1969) also showed that resorption was dependent on the maintenance of patent vascular channels. Rygh (1972b) studied the ultrastructural changes in blood vessels in both human and rat material and found (1) packing of erythrocytes in dilated blood vessels within 30 minutes, (2) fragmentation of erythrocytes after 2–3 hours, and (3) disintegration of blood vessel walls and extravasation of their contents after 1–7 days. He also observed necrotic changes in PDL fibroblasts such as dilatation of the endoplasmic reticulum and mitochondrial swelling within 30 minutes, followed by rupture of the cell membrane and nuclear fragmentation after 2 hours; cellular and nuclear fragments remained within hyalinized zones for several days (Rygh, 1972a). There was no difference in the response of rat and human tissues apart from timing; changes observed after 2 days in humans were seen after 2 hours in rats.
Kvam (1972) and Rygh (1972a,b) showed that root resorption is a side-effect of the cellular activity associated with the removal of the necrotic hyalinized tissue. Scanning electron microscopy of premolar root surfaces following the application of a 50 g force to the crown in a lateral direction revealed resorption cavities extending into the dentine, but where the hyalinized tissue remained intact, the root surface was unaffected (Figure 4; Kvam, 1972). Most recently, tartrate-resistant acid phosphatase (TRAP) staining in a rat tooth movement model has highlighted the involvement of TRAP-positive macrophages and multinucleate giant cells in the removal of hyalinized tissue (Brudvik and Rygh, 1994), and clearly demonstrated that on reaching the adjacent root surface, TRAP-positive cells continue to remove the cementum and subjacent dentine-producing resorption lacunae (Figure 5).
|
|
|
The pressure–tension hypothesis |
|---|
The idea that pressure and tension sites are generated within the PDL is firmly embedded in the orthodontic subconscious and it continues to play a key role in organizing our ideas, as well as advancing our understanding of a complex biological process. However, there are two major conceptual problems associated with the hypothesis. First, does stretching of the principal fibre bundles generate tension and second, can differential pressures be developed within the tissues of the periodontium?
Does stretching of the principal fibres of the PDL generate tension?
A persistent dogma of the orthodontic literature is that the collagen fibres of the PDL are stretched during tooth movement. Tension is thereby generated in the fibres responsible for the cellular response, particularly the stimulation of osteogenesis at the cortical bone surface into which the fibres are inserted. This belief seems to have been partly due to the traditional textbook representation of a tooth suspended in its socket by the PDL, which overemphasized the role of collagen in tooth support, and also due to a lack of knowledge of the structure and function of the ligamental proteoglycans and other non-collagenous structural molecules that are not easily visualized by conventional histological processing.
Direct measurements of the experimental intrusion of teeth (Parffitt, 1960; Bien and Ayres, 1965; Picton, 1965) suggested that, when mechanically loaded, the periodontal tissues behave as a viscoelastic gel which flows when subjected to a steady force but ‘bounces’ when a load is briefly applied and then removed. The damping of the mechanical forces acting on a tooth was originally attributed to three distinct but interacting fluid systems: (1) the vascular system, (2) cells and periodontal fibres, and (3) the interstitial fluid continuum (Bien, 1966). While these fluid systems undoubtedly play a part, the shock-absorbing function of the ligament is more likely to result from the ability of the hydrophilic proteoglycan molecules to form a strongly hydrated space-filling gel, whose displacement is limited by the collagen fibre network and lamina dura of the alveolar bone. Momentary pressures created by the sudden forces of mastication are therefore of different physiological significance to the prolonged pressures of orthodontic appliances, since biting forces in the neighbourhood of 1500 g/cm2 do not crush the periodontal membrane and impact the tooth through bone (Bien, 1966).
To test the hypothesis that tension generated in the collagen fibres of the PDL provides the stimulus for osteogenesis, Heller and Nanda (1979) disrupted collagen metabolism and function in rats by the systemic administration of the lathyritic agent ß-aminoproprionitrile; this inhibits the intermolecular cross-linking of the polypeptide chains of the collagen molecule. They found that in lathyritic rats the histological response of the bone to orthodontic tooth movement appeared to be normal. This suggests that when a tooth is continuously loaded during orthodontic treatment, it is unlikely that the principal fibres of the PDL undergo significant tensile strain, or transfer forces directly to the alveolar bone via Sharpey's fibres.
Can differential pressures be generated within the PDL?
Baumrind (1969) proposed that since the PDL appears to behave as a continuous hydrostatic system, any force delivered to it will, in accordance with Pascal's law, be transmitted equally to all regions of the ligament. But is the analogy appropriate? This depends on how one defines a fluid. To what extent can the supporting structures of a tooth (cells, collagen, proteoglycans, blood vessels, tissue fluids) be regarded as a fluid? And can the lamina dura of the tooth socket with its numerous vascular perforations be regarded as a closed vessel? Apart from minor adjustments following normal occlusal loading (and perhaps even then), the evidence from tooth movement experiments (localized vascular stasis, hyalinization, direct versus undermining bone resorption) would seem to overwhelmingly support the hypothesis that differential pressures can be generated within the periodontium.
|
The role of bone bending in orthodontic tooth movement |
|---|
Baumrind (1969)
also observed that the crown of the first molar was displaced, on average, 10 times more than the average reduction in PDL width on the pressure side, suggesting that bone deforms more readily than the PDL. Chapter 6 in Angle's Seventh Edition (Angle, 1907, p. 132) begins with the following paragraph:When a force is exerted upon the teeth to be moved two principal changes take place in the alveolar process. First, a bending of the process; second, absorption of the process in advance of the moving teeth and deposition of bone behind it. These changes vary greatly: according to the age of the patient, in different patients of the same age, in the direction of movement and also in the rapidity of movement.
The role of bone bending in orthodontic tooth movement was then ignored for the next 50 years, apparently even by Angle.
Bone deflection studies
Laboratory investigations that demonstrated unequivocally that bone deflection accompanied tooth displacement (Mühlemann, 1954; Mühlemann and Zander, 1954; Picton, 1965) were not undertaken with orthodontics in mind, but were physiological studies of tooth support. Experiments of Mühlemann and Zander (1954) with Macaca mulatta (rhesus) monkeys found that forces of 50–100 g were required to initiate labial or lingual displacement of a tooth from its physiological rest position and deformation of the alveolar bone started when forces greater than 100 g were applied to the teeth. Picton (1965) measured the distortion of alveolar bone with horizontal and axial forces in adult monkeys and noted that bone displacement started in response to forces less than 100 g and occurred in a linear manner up to 1 kg. Horizontal forces of more than 50 g caused the labial and lingual plates to be displaced in the same direction as the applied force; however, intrusive forces of the same magnitude caused dilatation of the socket, making it difficult to explain tooth support in terms of a tensional mechanism.
Although most bone deflection studies have been carried out on laboratory animals, experiments in humans have shown that the interseptal bone can also be bent. Measurements of interseptal bone deformation adjacent to tooth extraction sites were undertaken in two 12-year-old patients by Grimm (1972). Apart from variations in the rate and magnitude of tooth movement, bending of the alveolar crest was detectable with forces of less than 50 g and reached 35 µm with a load of 1.5 kg. Somewhat surprisingly, he recorded a 5 to 30 µm initial deflection of the crestal bone in the opposite direction to the tooth movement; this effect was transitory, however, and reversed to a slow positive trend after 1 minute.
The cellular and molecular response to bone deflection
Osteocytes are sensitive to mechanical deformation, and the idea that the cytoplasmic syncytium may detect changes in mechanical loading and therefore act as mechanoreceptors in bone has a long history (Pritchard, 1956). Experimental studies using both in vitro and in vivo models indicate that osteocytes are sensitive to stress applied to intact bone and may indeed function as mechanoreceptors. Load-related increases in glucose-6-phosphate dehydrogenase, 3H-uridine, c-fos, and insulin-like growth factor-I expression have been detected in osteocytes within 6 hours (Skerry et al., 1989; El Haj et al., 1990; Lean et al., 1996), demonstrating that intermittent loading at physiological strain magnitude produces rapid changes in metabolic activity.
Recently, Pavlin et al. (2001), Pavlin and Gluhak-Heinrich (2001), and Gluhak-Heinrich et al. (2003) focussed on the temporal expression of osteoblast- and osteocyte-associated genes following the loading of alveolar bone in a mouse molar tooth movement model. They found the response of osteoblast-associated genes to deformation in cells at the PDL–bone interface to be 10- to 20-fold greater than the increase in their number, suggesting that differentiation and increased cell function were the initial responses to loading the bone. ALP and BSP gene expression were detected after 24 hours, followed by the concomitant stimulation of osteocalcin and collagen type I expression from 24 to 48 hours (Pavlin et al., 2001). Most interesting, however, was the response of the osteocytes; the expression of dentine matrix protein-1 mRNA in the osteocytes of alveolar bone was increased twofold as early as 6 hours after loading, on both the formative and resorptive sides of the tooth (Gluhak-Heinrich et al., 2003).
Stress-generated electrical effects in bone
When an external load is applied to a long bone, deformation occurs, thus producing changes in the surface curvature (Figure 6). The external surface of the left cortex elongates in tension, while the internal surface shortens in compression; opposite effects are produced on the internal and external surfaces of the right cortex (Epker and Frost, 1965). Increasing concavity has consistently been shown to be associated with electronegativity and bone formation. In contrast, increasing convexity is associated with electropositivity and bone resorption (Bassett and Becker, 1962).
|
Stress-generated electrical potentials have been recorded with surface electrodes in a dog mandible following the application of a mechanical force to the teeth (Gillooly et al., 1968
), and Zengo et al. (1973) proposed that electrical potentials are responsible for regulating osteogenesis and bone resorption in the course of orthodontic tooth movement (Figure 7). During the 1960s and 1970s, the regulation of bone remodelling by piezoelectrical effects enjoyed considerable support within the orthopaedic and orthodontic communities. However, the problems from a biological point of view are whether electrical phenomena are sufficiently discriminatory to be able to regulate the metabolic activity of cell types as diverse as osteoblasts and osteoclasts, which function in close proximity, and also the fact that piezoelectricity does not require the presence of living cells; dead bone displays the same effects, which appear to be generated by shearing forces acting on the collagen fibres of the bone matrix. The more that is learnt about cell–cell and cell–matrix interactions, the more it is likely that stress-generated electrical potentials represent a by-product of deformation, a physical phenomenon that provided a plausible explanation for bone remodelling before the discovery of growth factors, cytokines, and other locally produced biochemical mediators. Nevertheless, while piezoelectricity as an explanation for the regulation of bone remodelling may have been incorrect, Figure 7 highlights the manner in which the alveolar plates are deformed when a tooth is loaded horizontally, as well as the sites at which bone deposition and resorption take place.
|
|
Surrogate models of tooth movement |
|---|
It is clear from the preceding discussion that by the 1970s, numerous well-documented studies of the histology of tooth movement had been undertaken in animals. Animal experimentation had limitations, however, and if significant progress were to be made in understanding the cellular and molecular mechanisms involved, then different experimental methods would be required. This provided the impetus for the development of in vitro cell and tissue culture systems designed to answer questions not readily accessible from animal studies conducted in vivo. Some of the experiments investigating cell signalling and mechanotransduction in which osteoblasts and PDL fibroblasts have been mechanically deformed in culture will be described later. This section focuses on the use of cranial sutures as a model of tooth movement.
Melcher and Turnbull (1972) and Yen and Melcher (1978) developed an organ culture system for mouse molar explants, but the technical difficulties of isolating and maintaining an explant composed of teeth and bone in vitro and applying force to the PDL were formidable. Meikle et al. (1979) developed organ culture systems in which mechanical force could be applied to rabbit cranial sutures under controlled conditions (Figure 8). The aim was to produce a simple experimental model that would mimic the forces to which the periodontal and other sutural articulations of the craniofacial skeleton are exposed during orthodontic treatment, and was based on the assumption that remodelling mechanisms are similar for all fibrous joints.
|
A particular advantage of the model was that both synthesis and degradation could be quantified contemporaneously using biochemical methods. Culturing explants in medium containing 3H-proline, for example, resulted in a two- to threefold increase in suture protein synthesis and a twofold increase in collagen synthesis that could be detected after 6 hours. There was also a significant increase in 3H-thymidine incorporation, and a threefold increase in the DNA content of the sutures after 48 hours, indicating a stimulation of cell proliferation and an increase in the cell population (Meikle et al., 1979, 1984
). While non-stressed sutures synthesized type I collagen, in stressed sutures, 20 per cent of the newly synthesized collagen was type III (Meikle et al., 1982). Type III collagen is found in connective tissues with a high turnover and suggests that the biomechanical environment of a cell is an important determinant of the collagen type synthesized.
Surprisingly, cells under tensile strain did not preferentially synthesize structural proteins. Mechanical deformation also stimulated the synthesis of enzymes such as the matrix metalloproteinases (MMPs), collagenase, and stromelysin responsible for their specific hydrolysis, as well as down-regulating tissue inhibitor of metalloproteinase (TIMP) expression (Figure 9; Meikle et al., 1980, 1984; Green et al., 1990). These findings suggested that degradation of the extracellular matrix (ECM) was stimulated by fibroblasts under a continuous tensile strain. The functional significance of such a response is not clear, but it seems likely that increased matrix turnover is necessary to facilitate cell division and accommodate an increase in the cell population.
|
) Denotes active enzyme. *P < 0.05, **P < 0.01. (Right) Immunolocalization of collagenase in a stressed coronal suture showing positive cells with diffuse fluorescence of the surrounding extracellular matrix indicative of the presence of active enzyme.Although rendered largely redundant by the increasing availability of cDNA probes and antibodies to human, mouse, and rat proteins, recent in vivo studies in rats (Takahashi et al., 2003
) and in vitro studies with PDL fibroblasts (Bolcato-Bellemin et al., 2000) have confirmed the above findings. Despite requiring a measure of lateral thinking, calvarial sutures continue to provide valuable information regarding the growth and remodelling of fibrous joints. In a mouse sagittal suture model for example, Ikegame et al. (2001) found that tensile mechanical stress increased bone morphogenetic protein-4 expression in preosteoblasts and fibroblasts, which in turn activated the expression of the osteoblast-specific transcription factor core-binding factor alpha-1, thereby inducing osteoblast differentiation. |
Mechanical strain activates multiple cell signalling pathways |
|---|
One area in which in vitro methods continue to make a significant impact is in the expanding field of cell signalling. The earliest experiments showed that the immediate response of cells to mechanical strain was the generation of prostaglandins (PGs) and the second messengers 3',5'cyclic adenosine monophosphate (cAMP; Harell et al., 1977
; Yeh and Rodan, 1984) and inositol phosphates (Sandy et al., 1989). Changes in intracellular [Ca2+] were later shown to occur via stretch-activated ion channels (Davidson et al., 1990; McDonald et al., 1996). Little was known at the time, however, about integrin receptors other than their role in cell adhesion. Activation of the cAMP and inositol phosphate signalling pathways is likely to lie downstream from the initial mechanoreception event at focal adhesions, where integrin receptors link the ECM to the cytoskeleton (Wang et al., 1993; DeMali et al., 2003).
The role of PGs
The first evidence that PGs might be involved in mechanical force transduction came from the cell culture studies of Hong et al. (1976), who observed that perturbing cell membranes by detaching cells mechanically from culture dishes caused an increase in PG synthesis. They suggested that mechanical distortion acts by changing the conformation of the cell membrane, thereby exposing membrane phospholipids to the action of phospholipases to release free arachidonic acid. These findings, plus research showing that changes in intracellular cyclic nucleotides followed mechanical deformation (Rodan et al., 1975; Davidovitch and Shanfield, 1975), provided the impetus for an in vitro study by Harell et al. (1977). They deformed petri dishes with orthodontic screws cemented to the base on which osteoblast-like cells had been cultured and suggested the following sequence of events: (1) activation of phospholipase A2 with the subsequent release of arachidonic acid resulted in a rapid (threefold) synthesis of PGE2 at 5 minutes, (2) this in turn activated adenylate cyclase and a transient increase in intracellular cAMP which peaked at 15 minutes, and (3) an increase in intracellular [Ca2+] and a stimulation of DNA synthesis. The ability to mimic or inhibit these changes by exogenous PGE2 or indomethacin [a cyclooxygenase-1 (COX-1) inhibitor] provided further evidence that the transduction of mechanical deformation into a biological response was mediated by PG production (Harell et al., 1977; Somjen et al., 1980). Other in vitro models have implicated PGs as mediators of mechanical stress in a number of cell types, including mechanically deformed osteoblasts (Yeh and Rodan, 1984), gingival fibroblasts (Ngan et al., 1988), and PDL fibroblasts (Ngan et al., 1990).
In an interesting practical application of the findings of Harell et al. (1977), Yamasaki et al. (1980) investigated the role of PGs in orthodontic tooth movement initially with the rat molar model. They found that indomethacin produced a dose-related inhibition in the appearance of osteoclasts and bone resorption that was limited to the first 12 hours of tooth movement. This explains why tooth movement is slower in patients who consume large quantities of non-steroidal anti-inflammatory drugs (NSAIDs). Injections of PGE1 and PGE2 into the gingival tissues near the upper first molar also stimulated osteoclast formation and bone resorption. This led to further work that showed local PG injections could increase the rate of tooth movement (Yamasaki et al., 1982, 1984).
Leukotrienes
A rabbit model showed that the COX inhibitor flurbiprofen significantly reduced osteoclast numbers, but interestingly not the degree of tooth movement (Sandy and Harris, 1984). This paradoxical observation suggested that PGs alone did not account for the bone remodelling associated with tooth movement. Leukotrienes (LTs) are also metabolites of arachidonic acid (but produced via the lipoxygenase pathway) and could have accounted for this discrepancy since they are potent stimulators of bone resorption (Meghji et al., 1988). It was subsequently shown that AA861, a selective inhibitor of the LT pathway, caused a significant inhibition of LTB4 and a reduction in tooth movement in rats (Mohammed et al., 1989).
The cAMP pathway
The second messenger system classically associated with mechanical force transduction is cAMP. The first evidence for the involvement of the cAMP pathway in mechanical signal transduction was provided independently by Rodan et al. (1975) and Davidovitch and Shanfield (1975). Rodan et al. (1975) showed that a compressive force of 60 g/cm2 applied to 16-day-old chick tibia in vitro inhibited the accumulation of cAMP in the epiphyses, as well as in cells isolated from the proliferative zone of the growth plate. The effect was mediated by an enhanced uptake of Ca2+ which inhibited membrane-associated adenyl cyclase activity (Bourret and Rodan, 1976).
Davidovitch and Shanfield (1975) sampled alveolar bone from compression and tension sites surrounding orthodontically tipped canines in cats. They found that cAMP levels initially decreased, followed by an increase after 1–2 days that remained elevated to the end of the experimental period of 28 days. They suggested that the initial decrease at the compression sites was due to necrosis of PDL cells, and at the tension sites to a rapid increase in the cell population; the elevation in cAMP observed 2 weeks after the initiation of treatment was probably a reflection of increased bone remodelling activity. Subsequently, Davidovitch et al. (1976), in a study on the cellular localization of cAMP in the same model, found an increase in the number of cAMP-positive cells in areas of the PDL where bone resorption or deposition subsequently occurred. Osteocytes in the adjacent alveolar bone, however, appeared to be relatively unaffected by the mechanical force.
The phosphoinositide pathway
The phosphoinositide (PI) pathway had been investigated in the 1950s, but the importance of membrane phospholipids as cellular second messengers was not fully appreciated until the 1980s, when Streb et al. (1983) showed that products of inositol lipid breakdown caused a release of intracellular Ca2+. The PI pathway accounts for many of the changes seen in mechanically deformed tissues, including the elevation of intracellular [Ca2+] from the endoplasmic reticulum and increased DNA synthesis.
PGE2 and parathyroid hormone (PTH) elevate inositol phosphates as well as cAMP in mouse calvarial osteoblast cultures; however, the effect of PTH on inositol phosphate mobilization was not direct, being mediated indirectly by PGs in an autocrine/paracrine fashion (Farndale et al., 1988). Evidence that mechanical stress was also PG-mediated made it unlikely that only one pathway would be selectively mobilized, and subsequently, Sandy et al. (1989) found that both cAMP and inositol phosphates were elevated in mouse osteoblasts following intermittent mechanical strain. Activation of cAMP and inositol phosphates by PGE2 has been shown to induce the expression of transcription factors such as Erg-1 (early growth response gene-1) and the proto-oncogenes c-jun and c-fos (referred to as immediate early genes). These activate or repress the transcription of late-acting genes required for bringing about changes in phenotype. Erg-1 is involved in cell proliferation and has been shown to be elevated in rat osteoblasts by mechanical stretch within 15 minutes (Dolce et al., 1996).
Signal transduction by integrins and focal adhesions
Cells in culture are not uniformly attached to their substrate but are ‘tack welded’ at specialized sites of cell attachment called focal adhesions. Focal adhesions are sites at which integrin receptors physically link actin-associated cytoskeletal proteins (talin, vinculin, 
-actinin, paxillin) with the ECM (Figure 10; Sastry and Burridge, 2000), as well as with adhesion molecules on the surface of adjacent cells. Integrins are therefore ideally sited to mediate the transmission of bidirectional forces across the cell membrane (Ingber, 1991). Integrins that are not bound to ligands in the ECM are generally distributed diffusely over the cell surface and appear not to be linked to the actin cytoskeleton; association with the actin cytoskeleton is induced by binding to the ECM (Wang et al., 1993). Integrins are composed of structurally distinct and ß subunits that combine to generate heterodimeric receptors with unique binding properties; for example, 2ß1 integrin binds to collagen and laminin and vß3 to vitronectin. Integrins bind to adhesive and structural proteins in the ECM via the peptide sequence RGD (arginine–glycine–aspartate). Integrin-mediated adhesive interactions play key roles in cell migration, proliferation, and differentiation, and also regulate intracellular signal transduction pathways that control adhesion-induced (outside–in) changes in cell physiology (Wang et al., 1993; Clarke and Brugge, 1995). Integrins thus function both as cell adhesion molecules and intracellular signalling receptors.
|
The numerous signalling pathways activated by integrins are beyond the scope of this discussion, but the low-molecular weight guanosine triphosphatases, Rab and Rho, as well as mitogen-activated protein (MAP) kinase subtypes that are components of integrin-mediated signalling have been shown to be altered in mechanically stretched PDL fibroblasts (Basdra et al., 1995
) and osteoblasts (Peverali et al., 2001). MAP kinase phosphorylates and activates phospholipase A2, leading to the hydrolysis of glycerophospholipids to yield arachidonic acid. Furthermore, integrin-dependent increases in intracellular [Ca2+] appear to involve either inositol triphosphate-induced Ca2+ mobilization from the endoplasmic reticulum or the transport of extracellular Ca2+ through mechanosensitive ion channels (Clarke and Brugge, 1995).
Evidence is accumulating making it increasingly likely that the previously reported changes in cell signalling in response to mechanical deformation are downstream of events mediated by integrins at focal adhesions. In a study designed to determine whether cells sense mechanical signals through generalized membrane distortion or through integrins, Meyer et al. (2000) found that mechanical stress altered the cAMP signalling cascade and downstream gene transcription through signals generated by activated integrin receptors, and in a G-protein-dependent manner. In an earlier report (Wilson et al., 1995), the proliferative response of cultured vascular smooth muscle cells to cyclic mechanical strain was found to be dependent on interactions between integrins and specific matrix proteins (collagen, fibronectin, vitronectin) since the response to strain was neutralized by antibodies to ß3 and vß5 integrins. Carvalho et al. (1996) have further shown that mechanical stimulation of bone cells results in a marked increase in ß1 mRNA expression after 30 minutes (v expression did not change). Firm attachment to the ECM at focal adhesions is evidently necessary to enable mechanical distortion to be recognized by (1) the cytoskeleton and intracellular signalling pathways and (2) mechanosensitive ion channels, phospholipids, and G-protein-coupled receptors in the cell membrane.
|
Cytokines |
|---|
One of the most significant advances in connective tissue biology during the 1980s was the demonstration that many cytokines, in addition to mediating the host immunological response to exogenous antigens, were also produced by connective tissue cells such as fibroblasts and osteoblasts and involved in normal physiological turnover and bone remodelling. Cytokines are low-molecular weight proteins (mw < 25 kDa) produced by cells that regulate or modify the action of other cells in an autocrine (acting on the cell of origin) or paracrine (acting on adjacent cells) manner. The definition includes the interleukins (ILs), tumour necrosis factors (TNFs), interferons, growth factors, and colony-stimulating factors. A major difficulty in understanding cytokine biology is the sheer number and complexity of these factors. Another problem is that several factors such as IL-1 and TNF exhibit overlapping biological activities (redundancy) and many have multiple biological effects (pleiotropy).
The role of cytokines in cell–cell signalling in bone
Cells of the osteoblast lineage play a pivotal role in bone remodelling, a process that involves interactions between osteoblasts and osteoclasts, systemic hormones, cytokines, and growth factors. Systemic hormones and mechanical stimuli influence the process via their ability to control the synthesis and/or action of cytokines. Since bone remodelling occurs at discrete sites throughout the skeleton, osteoblast-derived cytokines are ideally placed to regulate or modify the action of other cell types in bone. The first cytokine shown to play a role in bone turnover was IL-1, which in addition to activating lymphocytes also proved to be a potent bone resorptive agent (Gowen et al., 1983; Heath et al., 1985). Not long afterwards, the TNFs were shown to stimulate bone resorption and inhibit bone formation in vitro (Bertolini et al., 1986), and subsequent studies have implicated numerous others.
Experiments in which bone marrow had been cultured in the presence of bone resorptive agents such as PTH, 1,25(OH)2 vitamin D3, or IL-1 had shown that osteoclast formation and function was dependent upon the presence of osteoblasts/stromal cells. This suggested that cells of the osteoblast lineage produced soluble factors that were involved in osteoblast–osteoclast signalling. After many attempts to purify, characterize, and sequence such an osteoclast-activating factor using biochemical methods, a cell-surface protein that was found to be identical to receptor activator of nuclear factor 
B ligand (RANKL, a member of the membrane-associated TNF ligand family) was simultaneously discovered by two research teams (Lacey et al., 1998; Yasuda et al., 1998) using molecular techniques. RANKL and its receptor RANK expressed on osteoclasts and their precursor cells turned out to be the molecular determinants of osteoclast formation and function (Figure 11).
|
Cell–cell signalling by RANKL is essential for the induction of osteoclast differentiation. PTH and other systemic hormones as well as cytokines, such as IL-1, TNF-
, and IL-6, stimulate bone resorption by their ability to up-regulate RANKL expression by osteoblasts/stromal cells. Another cytokine osteoprotegerin (OPG; Simonet et al., 1997) also produced by osteoblasts/stromal cells acts as an inhibitor of osteoclast function by competing with RANKL for the membrane receptor RANK. Evidence is emerging which suggests that RANKL and OPG produced by PDL fibroblasts and osteoblasts play important roles in regulating connective tissue turnover and bone resorption during orthodontic tooth movement.
Cytokines as mediators of mechanically induced bone remodelling
Early research into the effects of tensile strain on cellular metabolism showed, with few exceptions, that whatever was measured always seemed to increase. The