J Dent Res 85(7):596-607, 2006
© 2006 International and American Associations for Dental Research
REVIEW
CRITICAL REVIEWS IN ORAL BIOLOGY & MEDICINE |
Inflammation-induced Bone Remodeling in Periodontal Disease and the Influence of Post-menopausal Osteoporosis
U.H. Lerner
Department of Oral Cell Biology, Umeå University, Umeå SE-901 87, Sweden; Ulf.Lerner{at}odont.umu.se
 |
ABSTRACT
|
|---|
During physiological conditions, the skeleton is remodeled in so-called bone multi-cellular units. Such units have been estimated to exist at 1–2 x 106 sites in the adult skeleton. The number and activities of these units are regulated by a variety of hormones and cytokines. In post-menopausal osteoporosis, lack of estrogen leads to increased numbers of bone multi-cellular units and to uncoupling of bone formation and bone resorption, resulting in too little bone laid down by osteoblasts compared with the amount of bone resorbed by osteoclasts. Inflammatory processes in the vicinity of the skeleton, e.g., marginal and apical periodontitis, will affect the remodeling of the nearby bone tissue in such a way that, in most patients, the amount of bone resorbed exceeds that being formed, resulting in net bone loss (inflammation-induced osteolysis). In some patients, however, inflammation-induced bone formation exceeds resorption, and a sclerotic lesion will develop. The cellular and molecular pathogenetic mechanisms in inflammation-induced osteolysis and sclerosis are discussed in the present review. The cytokines believed to be involved in inflammation-induced remodeling are very similar to those suggested to play crucial roles in post-menopausal osteoporosis. In patients with periodontal disease and concomitant post-menopausal osteoporosis, the possibility exists that the lack of estrogen influences the activities of bone cells and immune cells in such a way that the progression of alveolar bone loss will be enhanced. In the present paper, the evidence for and against this hypothesis is presented.
KEY WORDS: osteoporosis • periodontitis • bone • estrogen • osteoclasts • osteoblasts
|
(1) INTRODUCTION
|
|---|
Bone tissue is a living organ which is continuously remodeled, not only during growth of the skeleton, but also throughout adult life. The basis for the remodeling process is the activities of bone-forming osteoblasts and bone-resorbing osteoclasts. Bone remodeling takes place at discrete bone multi-cellular units, where the process starts by osteoclasts resorbing old bone, followed by new bone formation by osteoblasts in the resorption lacunae. During physiological conditions, the two activities are coupled, i.e., the amount of bone formed by osteoblasts is equal to that resorbed by osteoclasts. In pathological processes, however, the two processes are uncoupled. In some diseases, the amount of bone resorbed exceeds that formed, and net bone loss will result. This is frequently seen in patients with inflammatory processes in the vicinity of the skeleton, as well as in most malignant tumors that metastasize to bone. In some patients, however, inflammatory processes or malignant tumors in the vicinity of the skeleton lead to an uncoupled remodeling process, in which more bone is formed than resorbed. In these situations, locally increased bone mass is seen in the skeleton. Histopathological studies in inflammatory conditions, and metastasizing tumors, have shown that both enhanced resorption and bone formation are observed in patients with bone loss. Similarly, increased bone formation as well as enhanced bone resorption can be observed in patients with increased bone mass. Thus, inflammatory processes and malignant tumors secrete factors/elements that influence both osteoblasts and osteoclasts, but it is the relative proportions of these activities that will determine whether bone loss or enhanced bone formation will develop.
In inflammation-induced bone remodeling, the activities of osteoblasts and osteoclasts are influenced not only by signaling molecules from immune cells, but also by systemic hormones, including sex hormones. Thus, estrogen receptors are expressed by bone cells and immune cells. It has been hypothesized that estrogen deficiency may influence remodeling in bone in sites with inflammatory processes in such a way that the resorptive phase will be enhanced, leading to a more progressive form of marginal periodontitis. Interestingly, the cytokines believed to be involved in post-menopausal osteoporosis (Lerner, 2006) and inflammation-induced remodeling (present paper) are very similar. In the present review, the cellular events in bone tissue close to inflammatory processes are described, followed by a discussion on the signaling molecules that have been suggested to be responsible for these events. In the latter part of the paper, the possibility that patients with post-menopausal osteoporosis will also exhibit bone loss in the jawbones is discussed. Finally, the evidence for and against the hypothesis that lack of estrogen will influence the progression of marginal periodontitis is summarized.
|
(2) BONE REMODELING IN INFLAMMATION-INDUCED BONE LOSS
|
|---|
Inflammatory processes, either within or in the vicinity of the skeleton, usually cause bone loss (inflammation-induced osteolysis). This clinical phenotype is observed not only in the alveolar bone surrounding the root of teeth at a site with marginal periodontitis, or at a site with apical periodontitis (periapical osteitis), but also in osteomyelitis, in the juxta-articular bone in patients with rheumatoid arthritis, and in the bone surrounding a loosened joint prosthesis. How an inflammatory process affects the skeleton at the cellular level has been studied much less as compared with how metabolic bone diseases like estrogen deficiency affect bone cells. One important reason for the lack of such information in patients with marginal peridontitis is the fact that it is impossible to obtain bone biopsies at periodontitis sites of humans suffering from this chronic disease. Furthermore, it is well-known from clinical studies that the loss of bone does nor occur linearly over a period of time, but usually progresses stepwise, making it even more difficult to gain insight into the pathogenesis of inflammation-induced bone loss based upon a series of biopsies at different time points. In apical periodontitis, however, histological studies are more easily performed, since, in some of the patients, the granulation tissue present at the apex of the root must be surgically removed, in addition to the conventional root filling. In these biopsies, which sometimes also include the surrounding bone tissue, osteoclasts on the surface of the jawbone can be seen (own unpublished observations). The important role of osteoclasts in bone loss in marginal periodontitis has been shown in experimentally induced marginal periodontitis in animals, including the hamster, rat, mouse, beagle dog, and the primate Macaca fascicularis (Biancu et al., 1995; Shibutani et al., 1997; Salvi et al., 1997; Paquette and Williams, 2000) (Fig. 1A
). Similarly, osteoclasts have been observed at the pannusbone interface and in subchondral bone in patients with rheumatoid arthritis (Bromley and Woolley, 1984; Gravallese et al., 1998), as well as in those with psoriatic arthritis (Ritchlin et al., 2003). On the bone tissue adjacent to the inflammatory pseudomembrane, which separates the joint prosthesis from bone, many Howship’s resorption lacunae and several osteoclasts have been described in specimens from patients who underwent revision arthroplasty (Atkins et al., 1997) (Fig. 1B). Thus, there is firm evidence in several diseases that the loss of bone seen near inflammatory processes can be attributed to increased bone resorption.
View larger version (121K):
[in this window]
[in a new window]
|
Figure 1. Inflammation-induced osteoclast formation. (A) The inflammatory process in the gingiva in periodontitis sites consists of mononuclear leukocytes and a large number of newly formed blood vessels. On the surface of the alveolar bone, several mononucleated bone-resorbing osteoclasts are present. The biopsy is from a mink with severe periodontal disease. (B) The inflammatory process in the pseudomembrane present between the prosthesis and femur consists of an infiltrate of mononuclear leukocytes. On the surface of the bone tissue in the femur from a patient re-operated due to loosening of a hip prosthesis, several osteoclasts and a large number of Howship’s resorption lacunae (dashed arrow) can be observed. The photo is a kind gift from Dr. Christopher Collins, University of Bristol, UK, and is reproduced with his permission. (magnification not defined)
|
|
Whether remodeling in inflammation-induced bone loss takes place in the same types of bone multi-cellular units as observed in physiological remodeling and post-menopausally induced remodeling is not known (Lerner, 2006). A characteristic feature in a loosened joint prosthesis is positive bone scans, both at the site where no radiological evidence of bone loss can yet be seen, and at sites with extensive localized bone loss (scalloping). The isotope used for these bone scans is incorporated into the skeleton during osteoblastic bone formation, indicating a high rate of bone formation at sites with high uptake. This suggests that, concomitantly with inflammation-induced bone resorption, there is substantially increased bone formation. In histological studies of loosened joint prostheses, not only can large numbers of osteoclasts be seen, but, in some areas, there is also new bone formation (Fig. 2A). Bone-resorbing intra-osseous malignant tumors (osteolytic tumors) are usually also revealed by the same types of bone scans, also suggesting that, in these cases, increased bone resorption is associated with a high rate of bone formation. In one study, this type of bone scan has been used in periodontitis patients. These individuals exhibited different levels of uptake of the isotope, probably reflecting different disease activities (Jeffcoat et al., 1991). Interestingly, treatment with the non-steroidal anti-inflammatory drug naproxen significantly reduced bone loss, as well as the uptake of the isotope, indicating that high uptake of the isotope reflected ongoing inflammation-induced bone remodeling.
View larger version (91K):
[in this window]
[in a new window]
|
Figure 2. Inflammation induced both osteoclast formation (osteolysis) and new bone formation (sclerosis). (A) On the surface of the femur from a patient with a loosened hip joint prosthesis, not only can osteoclasts (arrowhead) be seen, but also active osteoblasts (solid arrow), producing new wowen bone tissue (dashed line), easily distinguishable from the old lamellar bone (open arrow). The photo is a kind gift from Dr. Christopher Collins, University of Bristol, UK, and is reproduced with his permission. (B) In most patients with inflammation-induced apical periodontitis, a zone with bone loss (osteolytic lesions) is observed adjacent to the apices of the roots of the teeth (solid arrow to the left). In some patients, however, the inflammatory process leads to new bone formation and a sclerotic response (dashed arrows to the right) in the jawbones. (magnification not defined)
|
|
In some patients with inflammatory disease in the vicinity of the skeleton, the clinical phenotype, as assessed by conventional radiology, is that of a local increase of bone formation (inflammation-induced sclerosis). This can be observed in patients with apical periodontitis (Fig. 2B
), but not in patients with marginal periodontitis. However, with the more sensitive bone scans, increased isotope uptake has been observed in patients with marginal periodontitis (Jeffcoat et al., 1991), as mentioned above, indicating that increased bone formation is also part of the process at these sites. Moreover, in patients with rheumatoid arthritis, new bone formation and sclerosis can also be seen in peri-articular bone at affected joints. The observations that patients with untreated rheumatoid arthritis have not only increased urinary excretion of biochemical markers of bone resorption, but also enhanced serum levels of alkaline phosphatase and osteocalcin, further suggest that rheumatoid arthritis is also associated with increased anabolic activities in the skeleton (Suzuki et al., 1998). In agreement with these observations, it was recently shown that human tumor necrosis factor (TNF) transgenic mice exhibit not only osteoclastic bone resorption but also enhanced endosteal bone formation in the vicinity of B-lymphocyte-rich infiltrates in the bone marrow (Görtz et al., 2004). It is well-recognized, however, that patients with rheumatoid arthritis have decreased skeletal mass, not only in affected joints, but also in the skeleton in general (Haugeberg et al., 2003), and rheumatoid arthritis is considered a risk factor for secondary osteoporosis, independent of glucocorticoid use (Lerner, 2006). Some metastatic cancers are also associated with increased bone formation (sclerotic tumors), which can easily be observed with conventional radiology, particularly in patients with prostatic cancers. These observations indicate that inflammatory processes and malignant tumors stimulate not only osteoclastic resorption, but also osteoblastic bone formation. In some disease states, the clinical outcome will be that of bone loss, because the resorptive events dominate the formative events, whereas in other lesions the opposite is observed.
In discussions of the effects of inflammation on bone cells, it might be important to consider the fact that the inflammatory process is also necessary for a fracture to heal. It is not understood, however, how bone formation and resorption are coordinated during fracture healing, but uncoordinated in pathological processes such as periodontitis, rheumatoid arthritis, and a loosened joint prosthesis.
Evidence has been presented indicating that prostaglandins, as with parathyroid hormone (PTH; Hodsman et al., 2005), are able to stimulate not only bone resorption but also bone formation (Hartke and Lundy, 2001; Pilbeam et al., 2002; Raisz and Woodiel, 2003; Vrotsos et al., 2003). One potential stimulator of inflammation-induced bone formation, therefore, could be prostaglandin E2 (PGE2). Further support for this view is that inhibitors of prostaglandin production may affect fracture healing (Li et al., 2003; Harder and An, 2003). Inhibition of prostaglandin biosynthesis by agents such as non-steroidal anti-inflammatory drugs is also frequently used to prevent heterotopic bone formation in patients undergoing hip replacement surgery (Fransen and Neal, 2004). Most interestingly, it has recently been reported that T-lymphocytes express capacity for stimulating alkaline phosphatase activity and mRNA expression of the osteoblastic transcription factor cbfa1 (Runx2) and of osteocalcin in human bone marrow stromal cells (Rifas et al., 2003). It will be of great interest to characterize the molecules responsible for these effects.
|
(3) CYTOKINES STIMULATING BONE RESORPTION
|
|---|
The first demonstration that immune-competent cells could affect bone cells was the observation that peripheral blood mononuclear cells, when stimulated either with phytohemagglutinin or with dental plaque, released into the culture supernatant factor(s) which could stimulate bone resorption in organ-cultured bones (Horton et al., 1972). This activity was initially called osteoclast-activating factor, and, when eventually purified 13 years later, was found to be the cytokine interleukin-1ß (IL-1ß) (Dewhirst et al., 1985). Shortly thereafter, it was shown that the cytokines tumor necrosis factor-
(TNF-; Bertolini et al., 1986) and interleukin-6 (IL-6; Ishimi et al., 1990) were also able to stimulate bone resorption. In most systems, IL-6 requires the presence of the soluble IL-6 receptor to induce osteoclast formation and bone resorption (Tamura et al., 1993; Kotake et al., 1996; Palmqvist et al., 2002). In contrast to several other soluble cytokine receptors, the IL-6 soluble receptor is able to activate target cells that express the gp130 protein when it binds to its ligand. The relative importance of IL-6 signaling through the cell-membrane-bound receptor vs. the soluble receptor is unknown. IL-1, TNF-, and IL-6 have been found to be abundantly expressed in inflamed gingiva, and increased levels have been shown in the crevicular fluid from patients with periodontitis (Mogi et al., 1999; Boch et al., 2001; Graves and Cochran, 2003). Similarly, these cytokines are also highly expressed in inflamed synovium and in synovial fluid from patients with rheumatoid arthritis, and have been implicated in the pathogenetic mechanisms that lead to bone resorption in this disease (Choy and Panayi, 2001; Udagawa et al., 2002; Firestein, 2003; Walsh and Gravallese, 2004). Thus, the same types of cytokines have been implicated as stimulators of bone resorption in periodontitis, rheumatoid arthritis, and post-menopausal osteoporosis (Lerner, 2006). Similarly to osteoporosis, neutralization of IL-1 and TNF- by the soluble IL-1 receptor and the soluble TNF receptor has been found to decrease osteoclast formation and bone loss in periodontal disease (Assuma et al., 1998), as well as in experimentally induced arthritis (Choy and Panayi, 2001; Firestein, 2003; Walsh and Gravallese, 2004). The crucial role of TNF- in humans is illustrated by promising new data showing that treatment of rheumatoid arthritis patients with TNF- antagonists, with and without immunosuppressant drugs, results in substantial reductions of clinical symptoms as well as retardation of radiographic progression of bone loss, although it is not yet clear if the reduction of bone loss is due to the effect of TNF- on osteoclast formation, or if it is more a consequence of the reduction in inflammation (Walsh and Gravallese, 2004). Inhibition of IL-1 also leads to clinical improvement in patients with rheumatoid arthritis, including decreased progression of radiographically assessed bone loss. The effects, however, are less than those obtained by TNF- blockade, indicating that TNF- has a more prominent role. Interestingly, IL- has also been found to be an important stimulator of bone resorption in experimentally induced apical periodontitis (Wang and Stashenko, 1993).
Since the findings showing that IL-1, TNF-, and IL-6 can stimulate bone resorption, several other cytokines have also been found to stimulate osteoclast formation and bone resorption, both in vitro and in vivo. The group of bone-resorbing cytokines currently includes IL-1, IL-6, IL-11, IL-17, TNF-, leukemia inhibitory factor (LIF), and oncostatin M (OSM) (Martin et al., 1998; Horowitz and Lorenzo, 2002) (Fig. 3). IL-7 has also been found to regulate osteoclast formation, but the data are inconclusive, since evidence for IL-7 being either stimulatory or inhibitory has been presented (Lee et al., 2003; Toraldo et al., 2003). Chemokines are a large family of chemotactic cytokines, which are also expressed in inflammatory processes. These molecules may contribute to inflammation-induced bone resorption, since chemokines in both the CXC and CCR families can stimulate the recruitment of osteoclast progenitor cells and the fusion of these cells to form multi-nucleated osteoclasts (Yu et al., 2003, 2004). In addition to these cytokines, we have also shown that molecules produced in the kallikrein-kinin and coagulation cascades during inflammation can stimulate bone resorption. Thus, bradykinin and Lys-bradykinin (kallidin), as well as thrombin, can stimulate bone resorption in vitro (Lerner et al., 1987; Lerner and Gustafson, 1988). Kinins, acting through both the B1 and B2 bradykinin receptors, also synergistically potentiate the stimulatory effects of IL-1 and TNF- on bone (Lerner, 1997; Lerner and Lundberg, 2002).
View larger version (20K):
[in this window]
[in a new window]
|
Figure 3. Several cytokines and inflammatory mediators have been shown to be able to stimulate osteoclast formation and bone resorption, and have therefore been implicated in the pathogenesis of inflammation-induced bone resorption. Some of the cytokines, the two kinins bradykinin and kallidin, as well as thrombin, are stimulatory, whereas other cytokines are inhibitory. The stimulatory cytokines exert their effects not by affecting the osteoclast progenitor cells directly, but by stimulating the RANKL/OPG ratio in periosteal osteoblasts. The inhibitory cytokines cause their effects either indirectly, by affecting osteoblasts, or, in some cases, directly, by affecting the osteoclast progenitor cells. Chemokines are important for the recruitment of osteoclast progenitor cells to the inflammatory site and also for the fusion of these cells to multi-nucleated osteoclasts.
|
|
The effects of cytokines that stimulate osteoclast formation and bone resorption seem to be counteracted by other cytokines that inhibit the same processes (Fig. 3). Thus, IL-4, IL-10, IL-12, IL-13, IL-18, interferon-ß (IFN-ß), and IFN-
are all able to inhibit either osteoclast formation, periosteal bone resorption in mouse calvariae, or both (Horowitz and Lorenzo, 2002; Palmqvist et al., 2006). Several of these cytokines have also been found to inhibit bone loss in vivo. It is likely that it is the balance between stimulatory and inhibitory cytokines, together with the regulation of their receptors and signal-transducing mechanisms, which will determine the quantity of osteoclasts formed and their activity, and thus the degree of bone loss.
|
(4) THE ROLE OF THE RANKL-RANK-OPG SYSTEM IN CYTOKINE-STIMULATED BONE RESORPTION
|
|---|
Stimulation of the receptor activator of nuclear factor 
B (RANK), present on osteoclast progenitor cells, by RANK ligand (RANKL), expressed by stromal cells/osteoblasts and some other cell types, is the signal that determines the differentiation of osteoclast progenitor cells/macrophages into fully differentiated osteoclasts (Lerner, 2004, 2006). Stimulation of RANK can be decreased by osteoprotegerin (OPG), which binds to RANKL and inhibits the interaction between RANKL and RANK. Since no osteoclast can be formed unless RANK is activated, regulation of the expressions of RANKL and the soluble decoy receptor OPG is also important in inflammation-induced bone resorption, including periodontitis (Liu et al., 2003; Takayanagi, 2005). It is, therefore, not unexpected that increased RANKL expression has been found in inflamed gingiva from periodontitis patients (Crotti et al., 2003), in gingival crevicular fluid from patients with periodontitis (Mogi et al., 2004), as well as in synovium from patients with rheumatoid arthritis (Gravallese et al., 2000; Shigeyama et al., 2000; Romas et al., 2002a), and in the membranous tissue surrounding a loosened joint prosthesis (Horiki et al., 2004). Moreover, OPG protein expression in gingiva from periodontitis patients is decreased (Crotti et al., 2003), and, in line with this observation, OPG in gingival crevicular fluid has been found to be significantly lower in sites with periodontitis, compared with healthy sites (Mogi et al., 2004). The importance of RANK activation in periodontal disease is demonstrated by the finding that administration of OPG prevents bone loss in experimentally induced periodontitis in mice (Teng et al., 2000). Similarly, OPG prevents bone loss in experimentally induced arthritis (Romas et al., 2002b). In agreement with this observation, mice with deletions of the RANKL gene do not exhibit bone loss when arthritic lesions are induced experimentally by a serum transfer model, despite the presence of progressive inflammatory processes (Pettit et al., 2001). The importance of RANK-induced activation of the NFB pathway for inflammation-induced stimulation of osteoclast formation has also recently been shown by Jimi et al.(2004). Activation of this pathway is initiated by activation of the inhibitor-B kinase (IKK) complex (see Lerner, 2006, and Fig. 5 therein), consisting of two catalytic subunits (IKK and IKKß) and a regulatory component (IKK or NEMO). Disruption of the binding of NEMO to IKK/IKKß by a cell-permeable peptide (NBD peptide) inhibits IKK activity and subsequent phosphorylation of IB, and, as a result, the activation of the NFB pathway is decreased. Administration of the NBD peptide to mice with collagen-induced arthritis significantly reduced destruction of juxta-articular bone (Jimi et al., 2004).
View larger version (20K):
[in this window]
[in a new window]
|
Figure 5. Osteoclast formation adjacent to inflammatory processes is critically dependent on RANKL expression. In an inflammatory process, there are several cells and mechanisms which can theoretically enhance RANKL. One possibility is that RANKL expressed by T-lymphocytes is important for activation of RANK in macrophages/osteoclast progenitor cells (A). Another possibility is that pro-inflammatory cytokines, such as IL-1 and TNF-, stimulate RANKL expression in periosteal osteoblasts (B). In periodontitis, gingival fibroblasts may also be instrumental, since these can release M-CSF, important for expansion of the pool of cells which are macrophages/osteoclast progenitor cells, as well as RANKL-stimulating cytokines such as IL-1, IL-6, and TNF- (C). The gingival fibroblasts express constitutively very small amounts of RANKL, but the expression can be induced by cytolethal distending toxin (cdt) from Actinobacillus actinomycetemcomitans (Aa). Interestingly, gingival fibroblasts constitutively express substantial amounts of the RANKL inhibitor OPG. A fourth possibility in patients with localized aggressive periodontitis is that leukotoxin expressed by Actinobacillus actinomycetemcomitans plays an important role, due to its capacity to release large amounts of IL-1 from macrophages, which then is important to induce RANKL in osteoblasts (D).
|
|
The interdependent roles of IL-1, TNF-, and RANKL in osteoclast formation were recently demonstrated by Wei et al.(2005). It has been shown that TNF- can stimulate osteoclast formation in the absence of stromal cells/osteoblasts if a "constitutive level" of RANKL (not leading to stimulation of osteoclast formation by itself) is present (Lam et al., 2000). Furthermore, blockade of either IL-1 or TNF- leads to partial inhibition of peri-articular bone loss in experimentally induced arthritis, whereas blockade of both causes almost complete inhibition of bone loss (Zwerina et al., 2004). Wei et al.(2005) showed that TNF- causes enhanced RANKL expression in stromal cells, to a large extent because of induction of both IL-1 and IL-1 type I receptors (Fig. 4). In addition to IL-1-mediated stimulation of RANKL by TNF-, the activation of TNF- receptors in osteoclast progenitor cells leads to enhanced expression of RANK, IL-1, and stimulatory IL-1 receptor type I, and decreased expression of inhibitory IL-1 type II receptors. IL-1 can stimulate the formation of osteoclasts in the presence of "constitutive levels" of RANKL in a manner independent of TNF-. These observations may explain why inhibition of both IL-1 and TNF- is such an effective therapy to inhibit bone loss in inflammatory conditions.
Thus, the inhibition of RANKL-RANK signaling is an interesting possibility for the treatment of patients with excessive bone loss due to inflammation. However, since mice deficient in either RANKL or RANK not only develop osteopetrosis because of a lack of osteoclasts, but also have disturbances in lymph node development (Kong et al., 1999a; Dougall et al., 1999), and since RANKL-RANK signaling is also important for the cross-talk between T-lymphocytes and dendritic cells (see below), there is a risk that OPG treatment will also disturb the function of the immune system. The finding that mice deficient in OPG exhibit not only bone loss, because of increased osteoclast formation, but also calcifications in large blood vessel walls (Bucay et al., 1998), indicates that OPG may also have a role in the pathogenesis of arteriosclerosis.
|
(5) INVOLVEMENT OF T-CELLS IN INFLAMMATION-INDUCED BONE LOSS
|
|---|
RANKL is not exclusively expressed by stromal cells/osteoblasts, but was in fact initially described on T-cells as TRANCE (Anderson et al., 1997; Wong et al., 1997), a ligand activating a receptor on dendritic cells that enhanced the activation of NF-B, and therefore was called receptor activator or NF-B (RANK). In line with this observation, T-cells have also been found to support osteoclastogenesis (Horwood et al., 1999; Kotake et al., 2001), similar to stromal cells/osteoblasts. The possibility exists, therefore, that osteoclast formation in inflammatory conditions may be increased by RANKL-expressing T-cells (Fig. 5A), rather than by stromal cells/osteoblasts induced to express RANKL by bone-resorbing cytokines (Fig. 5B). Accordingly, Kong et al.(1999b) have shown that systemic activation of T-cells causes peri-articular bone loss, and that the loss of bone can be inhibited by the administration of OPG. Similarly, Teng et al.(2000) have demonstrated that transplantation of human peripheral blood lymphocytes, from patients with localized juvenile periodontitis or aggressive periodontitis, into NOD/SCID mice, followed by oral challenge with the bacterium Actinobacillus actinomycetemcomitans (a bacterium highly associated with aggressive periodontitis), resulted in activation of CD4+ T-cells in the gingiva as well as bone loss. The loss of alveolar bone was prevented by OPG treatment. However, the relative importance of RANKL expressed in T-cells, vs. stromal cells/osteoblasts, for the enhancement of osteoclast formation in inflammatory conditions like periodontitis and rheumatoid arthritis is not known.
Fibroblasts are abundant in gingival mucosa and have also been found to express several osteotropic cytokines. Besides expressing IL-1, TNF-, and IL-6, gingival fibroblasts also express OPG (Sakata et al., 1999) (Fig. 5C). In fact, the starting point for the discovery of the long-sought RANKL-RANK-OPG system was based on the finding that human embryonic lung fibroblasts (IMR-90) produce an inhibitor of bone resorption called osteoclast inhibitory factor (OCIF; Tsuda et al., 1997), which turned out to be the molecule that today is called OPG. With this molecule, its ligand (OPGL, now called RANKL) was cloned, which then led to the cloning of its cognate receptor, RANK. Gingival fibroblasts also produce M-CSF, but do not constitutively express RANKL, nor will IL-1 or TNF- stimulation result in RANKL expression (Palmqvist et al., unpublished observations). We have recently found, however, that the expression of RANKL mRNA and protein can be induced in gingival fibroblasts by a cytolethal distending toxin (cdt) expressed by the bacterium A. actinomycetemcomitans (Belibasakis et al., 2005a) (Fig. 5C). Induction of RANKL by cdt is not dependent on stimulation of cytokines such as IL-1, IL-6, or TNF-, or on PGE2 (Belibasakis et al., 2005b). The rapid loss of bone in patients with aggressive periodontitis may not be caused only by cdt increasing RANKL in gingival fibroblasts, since we recently have also found that leukotoxin, expressed by A. actinomycetemcomitans, induced abundant secretion of bioactive 17-kDa IL-1ß from human monocytes caused by stimulation of caspase-1-mediated activation of 31-kDa pro-IL-1ß (Kelk et al., 2005) (Fig. 5D). Interestingly, synovial fibroblasts also have the capacity to express RANKL (Gravallese et al., 2000). Thus, the possibility exists that osteoclast formation in inflammatory diseases such as periodontitis and rheumatoid arthritis may be caused by the expression of RANKL on either stromal cells/osteoblasts, T-cells, or fibroblasts, or by RANKL expressed by all cell types.
|
(6) PROSTAGLANDINS AS MEDIATORS OF INFLAMMATION-INDUCED BONE RESORPTION
|
|---|
In periodontal disease, it has long been known that treatment with non-steroidal drugs that inhibit prostaglandin production causes substantial inhibition of osteoclast formation and bone resorption in experimentally induced periodontal disease in a variety of different species, as well as in human periodontal disease (Salvi et al., 1997; Paquette and Williams, 2000). These observations are in line with the recent finding that the enzyme cyclo-oxygenase-2 (COX-2) is highly expressed in the gingiva from periodontitis patients (Zhang et al., 2003). Since it is not likely that systemic administration of these drugs to periodontitis patients will be ubiquitous, their topical administration has been considered. Inhibition of periodontitis-induced bone loss by non-steroidal anti-inflammatory drugs can be explained by the fact that PGE2 can stimulate bone resorption in vitro and in vivo. Less evidence exists, however, that these drugs affect the progression of bone loss in rheumatoid arthritis patients.
How does prostaglandin production come into the sequence of events leading to RANKL expression and osteoclast differentiation? Several of the cytokines stimulating RANKL expression and bone resorption also enhance the expression of COX-2 and prostaglandin production. This is true for IL-1 and TNF-, and we have recently observed that the long-known inflammatory mediator bradykinin synergistically potentiates IL-1- and TNF--induced expression of COX-2 and prostaglandin production in the human osteoblastic cell line MG-63 and in mouse calvarial bones, a phenomenon associated with enhanced RANKL expression in the calvarial bones (Brechter and Lerner, in preparation). These observations fit well with the findings that bradykinin potentiates IL-1-induced bone resorption in mouse calvariae (Lerner, 1991), and that PGE2 is an efficient stimulator of RANKL expression in osteoblasts (Li et al., 2002). Recently, it has been shown that the stimulatory effect by RANKL in spleen cell cultures can be potentiated by PGE2 (Ono et al., 2005), suggesting that prostaglandins are important not only for RANKL expression in stromal cells/osteoblasts, but also for the effect of RANKL on osteoclast progenitor cells.
Although much evidence points to an important role for prostaglandins as mediators of inflammation-induced bone resorption, the role of these arachidonic acid metabolites in inflammation-induced remodeling of the skeleton is complicated by the fact that prostaglandins can also stimulate bone formation (Pilbeam et al., 2002) and have also been implicated in inflammation-induced new bone formation, mostly during fracture healing (Harder and An, 2003; Li et al., 2003). It might be that PGE2, which acts via the cyclic AMP-protein kinase A pathway, similar to PTH, which also acts via the same pathway, may have dual effects in bone remodeling. At persistent high concentrations that stimulate RANKL expression in stromal cells/osteoblasts, PGE2 eventually causes enhanced bone resorption, but when PGE2 is exposed to osteoblasts intermittently, it may lead to increased bone-forming activity.
|
(7) HYPOTHETICAL MECHANISMS BY WHICH POST-MENOPAUSAL OSTEOPOROSIS MAY CONTRIBUTE TO BONE LOSS IN PERIODONTITIS
|
|---|
Osteoclast formation, locally in periodontitis and systemically in post-menopausal osteoporosis, shares many pathways in the pathogenetic mechanisms. This is particularly true with respect to the roles of the pro-inflammatory cytokines IL-1, TNF-, and IL-6 and the osteoclastogenic cytokine RANKL. However, the mechanisms triggering up-regulation of cytokine induction are different. In periodontitis, the up-regulation is caused primarily by an infection and the subsequent inflammatory response; in post-menopausal osteoporosis, it is caused by estrogen deficiency. Although there are no reasons to believe that periodontal disease is caused by estrogen deficiency, the possibility exists that post-menopausal osteoporosis may contribute to the progression of bone loss in periodontal disease. Hypothetically, two different mechanisms may be involved. First, if a reduction of bone mass due to systemic osteoporosis disease exists in the alveolar jawbones, it is possible that superimposed inflammation-induced bone resorption may lead to enhanced progression of bone loss in comparison with that in non-osteoporotic jawbones. Second, since estrogen inhibits the expression of the bone-resorbing cytokines IL-1, TNF-, and IL-6, it might be that larger amounts of these molecules are produced in an inflammatory process in post-menopausal women with estrogen deficiency, compared with an inflammatory process in women with normal estrogen levels. In the following sections, the existing evidence for these two possibilities will be discussed. Interested readers are also referred to previous reviews (Jeffcoat, 1998; Chesnut, 2001; Krall, 2001; Reddy, 2001; Wactawski-Wende, 2001).
|
(8) DOES OSTEOPOROSIS EXIST IN THE JAWBONES OF POST-MENOPAUSAL WOMEN?
|
|---|
As discussed extensively by von Wowern (2001), studies on osteoporosis in the jawbones should be based upon site-specific assessments of bone mineral density (BMD)/bone mineral content (BMC) and related to normal, population-based, and gender- and age-related reference values. Currently, there are no ideal devices available for this purpose, but with the new peripheral quantitative computed tomography (pQCT), devices that are being developed, it might be possible in the near future to obtain accurate assessments of bone in the cortical and trabecular bone, as well as in discrete sites of the jaws. With the currently available technology, only a weak correlation between BMD/BMC in different sites of the skeleton and jawbones has been found, although there are studies indicating a correlation between bone mass in the mandible and skeleton in general (Kribbs et al., 1983; von Wowern et al., 1994). There are also large variations of BMD in different areas of the jawbones, which makes it difficult to make comparative studies. In fact, there are no good studies, based on direct assessment of skeletal bone mass, to show conclusively that low BMD exists in the skeletons of women with post-menopausal osteoporosis, and that changes in jawbone BMD are correlated to changes in other well-recognized osteoporotic sites in the skeleton. However, a two-year longitudinal study has shown that alveolar bone density decreases more rapidly in post-menopausal women compared with women with normal BMD (Payne et al., 1999), indicating that bone tissue in the jaws is also affected by estrogen deficiency.
The observation that calcium and vitamin D supplementation in post-menopausal women also increases BMD in the jawbones (Kribbs, 1992; Hildebolt et al., 2004) indicates that bone remodeling in jawbones is affected by the same mechanisms causing decreased bone mass in other sites of the skeleton. Moreover, estrogen treatment has been found to prevent bone loss in alveolar bone (Payne et al., 1997), and to result in increased BMD to the same degree in the spine as in the jaws (Jacobs et al., 1996). Recently, Civitelli et al.(2002), in a randomized, double-blind, placebo-controlled, three-year prospective study, reported that hormone replacement therapy significantly increased alveolar bone mass as assessed by radiodensity measurements on bite-wing radiographs, compared with placebo treatment.
The pathogenetic mechanisms in corticosteroid-induced osteoporosis are different from those in post-menopausal osteoporosis. Nevertheless, it is interesting to note that corticosteroids decrease BMD in the radius and jawbone to similar extents (von Wowern et al., 1992).
Although analyses of BMD in jawbones have not proved that osteoporosis also exists in jawbones, analysis of the data obtained so far with different devices, together with the findings showing that jawbones respond to osteoporotic treatments similar to other sites in the skeleton, suggests that jawbones are also susceptible to osteoporosis.
|
(9) RELATIONSHIPS BETWEEN BMD/BMC AND PERIODONTITIS
|
|---|
Cross-sectional correlation studies, cross-sectional multivariate studies, and case-control studies have showed either no association between BMD and periodontal disease (Elders et al., 1992; Hildebolt et al., 1992, 1997; Weyant et al., 1999), weak association (Kribbs et al., 1989; Pilgram et al., 2002), or significant association (von Wowern, 1986; von Wowern et al., 1994; Mohammad et al., 1997). The observation that decreasing metacarpal BMD is associated with an increased proportion of individuals exhibiting periodontitis, both pre- and post-menopausally, indicates that there is an association between bone mass in the jawbones and periodontal disease (Inagaki et al., 2001). It has also been shown that decreased BMD of the trochanter, Ward’s triangle, and total regions of the femur is significantly associated with interproximal loss of alveolar bone and, to some extent, also with decreased clinical attachment level (Tezal et al., 2000). In contrast, a Swedish study reported that, in a small cohort of post-menopausal women (n = 19), BMD of the hip was not related to alveolar bone loss (Lundström et al., 2001).
In a recent study, osteoporosis was initially assessed by ultrasound densitometry of the heel, and the progression of periodontal disease was followed over three years by measurement of the loss of attachment level (Yoshihara et al., 2004). In that study on 179 individuals, all 70 years old, low initial BMD was significantly correlated with the longitudinal loss of attachment level in both females and males. In another two-year longitudinal study on 38 post-menopausal women, it was found that loss of alveolar bone height during the study was significantly associated with the initial BMD of the lumbar spine (Payne et al., 1999).
These studies suggest that systemic bone loss, due to uncoupled bone remodeling because of estrogen deficiency, also affects jawbones and is important for the progression of bone loss in periodontitis.
|
(10) RELATIONSHIPS BETWEEN BMC/BMD AND TOOTH LOSS
|
|---|
The loss of teeth is the end-point of periodontal disease and, therefore, is similar to osteoporotic fractures, a clinically important marker of the disease. However, teeth can be lost for reasons other than periodontal disease, e.g., teeth sometimes need to be extracted because of extensive caries lesions, or can be lost because of extensive trauma. Nevertheless, periodontal disease is one important factor causing tooth loss. Several studies have reported that decreased bone mineral density is associated with the number of lost teeth (Daniell, 1983; Kribbs, 1990; Klemetti et al., 1994; Krall et al., 1994, 1996; Bando et al., 1998; Taguchi et al., 1999; Inagaki et al., 2001). There are also reports in which no such association could be found (Elders et al., 1992; Hildebolt et al., 1997; Mohammad et al., 1997). In one study, the association between BMD and tooth loss was greater in men than in women (May et al., 1995). In a large study on 1171 post-menopausal Turkish women aged 40–86 yrs, it was found that total loss of teeth was significantly associated with low BMD in the lumbar spine but, in contrast, with high BMD of the femoral neck (Gur et al., 2003). Although analysis of most data obtained in cross-sectional studies supports the hypothesis that BMD/osteoporosis is associated with tooth loss, conclusive evidence has to await data from well-controlled prospective studies in large cohorts of both men and women. In a recent study in 145 edentulous and 253 dentate women, no association was found between the annual percentage change of BMD of the total hip during a two-year period and the initial loss of tooth attachment (Famili et al., 2005).
It has also been shown that treatment with calcium and vitamin D supplementation, which is not likely to have any effect on caries or extensive trauma, reduced the number of teeth lost (Krall et al., 2001).
In a cohort of 566 hip fracture patients, Åström et al.(1990) reported that the risk for a hip fracture increased with the number of teeth lost. In a recent study, using the end-points of both periodontal disease and osteoporosis in a large cohort (n = 567) of 70-year-old women, we found that total loss of teeth is related not only to BMD but also to the presence of osteoporosis (Österberg et al., 2004). Thus, loss of all teeth was significantly associated with skeletal fractures, with an odds ratio of 2.41. Logistic regression analysis showed that total loss of teeth was an independent significant factor for previous fracture, with an odds ratio of 2.37.
|
(11) ESTROGEN AND PERIODONTITIS
|
|---|
Treatment with estrogen is commonly used to prevent bone loss in osteoporotic women. Several studies have shown that estrogen treatment decreases the risk of tooth loss (Paganini-Hill, 1995; Grodstein et al., 1996; Krall et al., 1997), which may be due to the effects of estrogen on both cytokine production in inflamed gingiva and on bone cells in jawbones. Interestingly, it has been reported that a low estrogen level in the blood is associated with increased signs of inflammation in the gingiva (Norderyd et al., 1993) and increased bone loss around the teeth, as assessed by reduction of the clinical attachment level (Reinhardt et al., 1999).
|
(12) ASSOCIATIONS AMONG GENE POLYMORPHISMS, PERIODONTITIS, AND OSTEOPOROSIS
|
|---|
Single-nucleotide polymorphisms of a large number of genes—including those for the vitamin D receptor, estrogen receptor- (ER-), and pro-collagen (1) type I and different cytokines—have been associated with BMD and susceptibility to osteoporotic fractures (Ralston, 2002). Similarly, several studies have been published demonstrating the presence of associations between periodontal disease and genetic polymorphisms, mainly in the genes for pro- and anti-inflammatory cytokines (Kinane and Hart, 2003). However, these studies have generally been performed on a very small number of patients. Keeping this limitation in mind, it is nevertheless interesting to note that significant associations between periodontal disease and single-nucleotide polymorphisms have been reported in genes previously known to be important for bone mineral density and osteoporosis. Thus, there are several studies showing associations between periodontal disease and genetic polymorphism in the vitamin D receptor gene (Tachi et al., 2001, 2003; Sun et al., 2002; de Brito Junior et al., 2004). Polymorphisms in the ER- gene have been associated with the presence of periodontal disease (Zhang et al., 2004) and with the loss of teeth (Taguchi et al., 2001, 2003). In a large survey of 315 single-nucleotide polymorphisms in 125 candidate genes, many associations with periodontal disease were found, interestingly, not only in inflammatory cytokines, but also in genes encoding structural proteins in the periodontium, including pro-collagen (1) type I (Suzuki et al., 2004). Although all these data must be interpreted with caution and confirmed in much larger cohorts, it is interesting to note that some genes may be important not only for bone mineral density and osteoporosis, but also for periodontal disease.
|
(13) SUMMARY
|
|---|
Inflammatory processes in the vicinity of the skeleton result in stimulation of both bone-resorbing osteoclasts and bone-forming osteoblasts. The relative proportions of enhanced bone resorption and bone formation determine if an osteolytic or a sclerotic lesion will develop. Immune cells are capable of secreting a variety of cytokines with a capacity to stimulate the development and activity of bone-resorbing osteoclasts, including IL-1, IL-6, IL-11, IL-17, TNF-, LIF, and OSM. The stimulatory effects of these cytokines are decreased by several cytokines that inhibit osteoclastogenesis, such as IL-4, IL-10, IL-12, IL-13, IL-18, IFN-ß, and IFN-. In addition, kinins and thrombin, formed in the kallikrein-kinin and coagulation cascades, can also stimulate bone resorption and synergistically interact with IL-1 and TNF-. Thus, it is likely that it is the concerted actions of stimulatory and inhibitory signaling molecules in an inflammatory process that will determine the numbers of bone-resorbing osteoclast. The RANKL-RANK-OPG system needs to be activated for osteoclasts to develop, and most of the stimulatory cytokines described are able to enhance the RANKL/OPG ratio in osteoblasts. However, not only osteoblasts but also T-cells—and, under certain circumstances, even gingival fibroblasts—can express RANKL. Much less is known as to which molecules are responsible for stimulating osteoblasts in inflammatory processes, but PGE2 may be a candidate. It has been suggested that patients with marginal periodontitis and concomitant post-menopausal osteoporosis are at increased risk for a more progressive form of periodontitis. Interestingly, estrogen suppresses the expression of several of the cytokines suggested to be responsible for osteoclast stimulation in inflammatory conditions. There is also evidence that patients with post-menopausal osteoporosis have decreased bone mass in the jawbones. Therefore, estrogen deficiency may enhance the progression of marginal periodontitis, either by causing increased expression of osteotropic cytokines, or by decreasing the amount of alveolar jawbone. The data from clinical studies on the degree of periodontal disease in patients with concomitant periodontitis and post-menopauasal osteoporosis are inconclusive, and there is a need for well-controlled prospective studies in which the progression of periodontal bone loss is followed in relation to estrogen levels.
|
ACKNOWLEDGMENTS
|
|---|
Studies performed in the author’s laboratory have been supported by The Swedish Research Council, The Swedish Rheumatism Association, The Royal 80 Year Fund of King Gustav V, The Knut and Alice Wallenberg Foundation, the Swedish Foundation for Strategic Research, SalusAnsvar, Astra-Zenecca, Pharmacia-Upjohn, The Swedish Dental Association, Patentmedelsfonden, Anna-Greta Craaford Foundation, The County Council of Västerbotten, Umeå University and Centre for Musculoskeletal Research, and the National Institute for Working Life, Umeå, Sweden.
Received June 15, 2005;
Accepted November 23, 2005
|
REFERENCES
|
|---|
TOP
ABSTRACT
(1) INTRODUCTION
