• prostaglandin G/H synthase;
  • cyclo-oxygenase 2;
  • prostaglandin E2;
  • osteolysis;
  • inflammation;
  • aseptic loosening


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

Aseptic loosening is a major complication of prosthetic joint surgery and is manifested as chronic inflammation, pain, and osteolysis at the bone implant interface. The osteolysis is believed to be driven by a host inflammatory response to wear debris generated from the implant. In our current study, we use a selective inhibitor (celecoxib) of cyclo-oxygenase 2 (COX-2) and mice that lack either COX-1 (COX-1−/−) or COX-2 (COX-2−/−) to show that COX-2, but not COX-1, plays an important role in wear debris-induced osteolysis. Titanium (Ti) wear debris was implanted surgically onto the calvaria of the mice. An intense inflammatory reaction and extensive bone resorption, which closely resembles that observed in patients with aseptic loosening, developed within 10 days of implantation in wild-type and COX-1−/− mice. COX-2 and prostaglandin E2 (PGE2) production increased in the calvaria and inflammatory tissue overlying it after Ti implantation. Celecoxib (25 mg/kg per day) significantly reduced the inflammation, the local PGE2 production, and osteolysis. In comparison with wild-type and COX-1−/− mice, COX-2−/− mice implanted with Ti had a significantly reduced calvarial bone resorption response, independent of the inflammatory response, and significantly fewer osteoclasts were formed from cultures of their bone marrow cells. These results provide direct evidence that COX-2 is an important mediator of wear debris-induced osteolysis and suggests that COX-2 inhibitors are potential therapeutic agents for the prevention of wear debris-induced osteolysis.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

PROSTAGLANDINS COMPRISE a group of short fatty acid derivatives that are the most abundant eicosanoids in bone where their production is highly regulated. The role of prostaglandins in bone metabolism appears to be complex, because they have been shown to have both stimulatory and inhibitory effects on bone formation and bone resorption.(1) Prostaglandin synthesis is controlled by three different enzymes: phospholipase A2's, which release arachidonic acid (AA) from the cell membrane; the cyclo-oxygenases, which catalyze the oxygenation and further reduction of AA to form prostaglandin H2 (PGH2); and isomerases, which convert PGH2 to individual prostaglandins. Among these, the cyclo-oxygenases are the most important rate-limiting enzymes in the pathway. The identification of a second isoform of cyclo-oxygenase (cyclo-oxygenase 2 [COX-2]) in the prostaglandin synthetic pathway has expanded our knowledge of the function of this group of fatty acid derivatives that serve as microenvironmental hormones.(2–6) In general, COX-1 is the predominantly constitutive form that is expressed throughout the body and presumably provides certain homeostatic functions. In contrast, COX-2 is the inducible form that is increased in response to inflammation, growth, or other physiological stimuli and is believed to mediate pain perception and modulate inflammatory responses and mitosis.(7)

Although there is strong evidence that COX-1 and COX-2 represent separate prostaglandin biosynthesis systems(7–9) and that each enzyme has distinct and sometimes overlapping functions,(10) the role of the two cyclo-oxygenases in bone metabolisms is still unclear. Recent studies suggest that COX-2 mediates increased prostaglandin production in bone in response to tumor necrosis factor α (TNF-α), interleukin-1β (IL-1β), IL-6, IL-11, IL-17, and other osteotropic factors,(11–15) and many factors that either stimulate or inhibit bone resorption affect COX-2 expression in osteoblasts or stromal cells.(1) These studies suggest that COX-2 may have a role distinct from that of COX-1 in bone resorption.

Aseptic loosening is a major clinical problem that causes debilitating pain and often requires revision surgery.(16) Currently, it is estimated that up to 20% of total hip arthroplasties will show evidence of osteolysis within 10 years of implantation.(16) Perhaps the most important cause of osteolysis is the inflammatory reaction to particulate wear debris, in which particles (<10 μM) of polyethylene, metal, and bone cement debris are phagocytosed by macrophages. Proinflammatory cytokines, such as TNF-α, IL-1β, and IL-6, that are subsequently released on phagocytosis of wear particles by macrophages have been identified as the mediators of the osteolysis induced by this wear debris.(16–19) In addition to these factors, increased prostaglandin production has been reported in cultured human peritoneal macrophages exposed to wear debris such as titanium (Ti) and in tissues around loose prosthetic components.(20) Recently, COX-2 has been detected immunohistologically in macrophages laden with wear debris in the pseudomembrane at the bone-implant interface,(21) and, thus, COX-2 has been implicated as a potential drug target in this disease.

To investigate the role of COX-2 in wear debris-induced osteolysis, we used the mouse calvaria bone resorption model in which wear debris particles were implanted surgically over mouse calvarial bones for 5 days or 10 days,(17) and the subsequent inflammatory and resorptive responses were quantified using histomorphometry.(22) A COX-2 selective inhibitor and knockout mice that lack either COX-1 or COX-2 were used in this study. Our results indicated that the selective COX-2 inhibitor celecoxib reduced inflammation and its associated bone resorption in this calvarial model, at doses that did not inhibit COX-1 activity. Furthermore, COX-2−/− mice had significantly less bone resorption in response to Ti than wild-type littermates and COX-1−/− mice. To elucidate the mechanism by which COX-2 contributes to this wear debris-induced osteolysis, we examined its role in proinflammatory cytokine production and osteoclastogenesis in vitro. We found that COX-2 was not required for the production of TNF-α, IL-1β, or IL-6 by macrophages in response to Ti, but that PGE2 synthesis was COX-2 dependent. We also found that bone marrow cells derived from COX-2−/− mice were deficient in their ability to form osteoclasts in response to inflammatory signals in vitro.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

Metal particles

Pure Ti particles were obtained from Johnson Mathey Chemicals (Ward Hill, MA, USA). These particles were shown to be free of endotoxin by Limulus assay (BioWhittaker, Walkersville, MD, USA). Thirty grams of particles was suspended in phosphate-buffered saline (PBS) at a concentration of 2 × 108 particles/ml. Analysis using a Coulter counter showed that the particles were between 2.0 ± 1 μm in diameter.

Experimental animals

All animal studies were conducted in accordance with principles and procedures approved by the University of Rochester Institutional Review Board. Male and female mice, 6–8 weeks old, were used in the experiments. The age- and sex-matched CBAxBL6 mice used in the drug studies were purchased from Jackson Laboratory (Bar Harbor, ME, USA). COX-1−/− and COX-2−/− mice(23,24) originally were obtained from the breeding colony in the University of North Carolina. They were of hybrid C57BL/6Jx129/ola genetic background, intercrossed for about 15–20 generations. Mice were genotyped using polymerase chain reaction (PCR) as described.(25)

Mouse calvarial resorption model

Eight-week-old female mice were anesthetized by peritoneal injection of 80 mg/kg of ketamine and 5–7 mg/kg of xylazine. A 1-cm midline sagittal incision was made above the calvaria. Ti (30 mg) particles were placed evenly on top of the calvarial bones leaving the periosteum intact. After implantation, the skin was closed with 4–0 Ethilon suture. Mice were given celecoxib (10 mg/kg or 25 mg/kg) daily by gavage. At day 10, mice were killed by cervical dislocation and calvaria were removed, fixed in 10% formalin, decalcified in 10% EDTA, dehydrated in graded alcohols, and embedded in paraffin. Representative sections were cut and stained with hematoxylin and eosin (H&E), as described previously.(22)

Immunocytochemical staining

Calvaria harvested at day 0, 1, 3, 5, 7, 10 were fixed in 10% formalin and decalcified in EDTA, as described. COX-2 expression was examined by immunocytochemical staining using an antibody against COX-2 (Cayman Chemical, Ann Arbor, MI, USA). Briefly, tissue sections were deparaffinized and then soaked in 3% hydrogen peroxide for 10 minutes. Normal goat serum, used to block the nonspecific staining and primary antibody at 1:200, were applied on the slides, which were incubated at 4°C overnight and washed with PBS. The sections were then incubated with biotinylated secondary antibody followed by streptavidin-peroxidase complex, and the binding of antibody was visualized with aminoethyl carbazole (AEC; Zymed, S. San Francisco, CA, USA), according to the supplier instructions.


Histomorphometry was performed, as described previously.(22) Briefly, after decalcification, the occipital bone was removed by cutting immediately behind and parallel to the lambdoid suture. Between six and twenty 3-μm-thick nonconsecutive coronal sections (150-μm intervals between sections) were cut for each specimen and stained using H&E. Sections at comparable levels from each sample were used for histomorphometric analysis. The histomorphometry analysis on the tissue sections we performed interactively with the Osteometrics (Atlanta, GA, USA), as we have described previously,(22) to quantify the following parameters as the percentage of total calvarial bone area or bone/bone marrow interface length: (1) the percentage of newly formed woven bone area over total bone area (TBA), (2) the number of osteoclasts within the marrow cavity and on the periosteal surface over TBA (expressed by per mm2 of the TBA), (3) the percentage of the eroded surface over the total bone surface (expressed by per mm of the total bone/bone marrow interface), and (4) the percentage of area of periosteal inflammation over that of Ti-treated mice. (Periosteal inflammation refers to Ti-induced inflammatory/fibrotic tissue overlying the calvarial bone. It is expressed as the area of the soft tissue on top of the calvarium per mm bone surface.)

Tartrate-resistant acid phosphatase staining

Bone sections and bone marrow cultures also were stained for tartrate-resistant acid phosphatase (TRAP) as described previously(26) to identify osteoclasts. Briefly, bone marrow cultures and bone sections were incubated in 0.2 M acetate buffer, pH 4.5–5.0, containing 0.01 M tartaric acid and 2% naphthol AS-BI phosphate (dissolved at 20 mg/ml in ethylene glycol monomethylether) for 15 minutes (marrow culture) or 30 minutes (tissue sections) at 37°C. The slides were then transferred to a second solution consisting of the same buffer and concentration of tartaric acid with a volume of 4% sodium nitrite for 5 minutes at room temperature. This treatment causes red cytoplasmic staining of cells expressing TRAP. Harris' hematoxylin was used as a nuclear counterstain.

Isolation and culture of bone marrow precursor cells for osteoclastogenesis assays

Bone marrow cells were isolated from 5- to 8-week-old mice as described previously.(27) Mice were killed by cervical dislocation. Femora and tibias were removed aseptically and dissected free of adherent soft tissue. The bone ends were cut, and bone marrow cells were flushed from marrow cavity by slowing injecting modified essential medium (MEM) at one end of the bone using a sterile 21G needle. The marrow suspension was dispersed gently by pipetting several times to obtain a single cell suspension. The bone marrow cells are washed twice, plated on culture dishes and incubated for about 2 h. These cells were then washed again and plated on to 96-well plates. Cells were cultured in phenol red-free MEM containing 10% fetal bovine serum (Hyclone, Laboratories, Salt Lake, UT, USA) for 8 days in the presence of 10−8 M 1,25-dihydroxyvitamin D3 [1,25(OH)2D3] and/or 10 ng/ml IL-1β or 20 ng/ml TNF-α, 5 μM celecoxib and/or 10−7 M PGE2. Cultures were fed with fresh medium every other day. At day 8 cultures were stained for TRAP activity and counterstained with hematoxylin to identify osteoclasts, as described previously.(28) TRAP-positive (TRAP+) multinucleated cells (more than three nuclei) were counted using light microscopy. The TRAP+ multinucleated cells from this bone marrow culture model have been shown to express calcitonin receptor, can form resorption pits on bone wafers, (28) and thus are considered osteoclasts.

PGE2 assay and thromboxane B2 assay

To measure PGE2 production, calvaria (bone plus the soft tissue membrane) were homogenized in the presence of 10 μM indomethacin acidified to pH 4.0 with 1N HCl, and then subjected to c18 reversed-phase chromatography (J.T. Baker, Inc., Philipsburgh, NJ, USA). PGE2 was assayed using an enzyme-linked immunosorbent assay (EIA) kit purchased from Cayman Chemicals. For thromboxane B2 (TXB2), whole blood from mice treated for 2 h with celecoxib was obtained by cardiac puncture and collected in heparinized tubes. Calcium ionophore A23187 (20 μg/ml final concentration; Sigma, Inc., St. Louis, MO, USA) was used to stimulate TXA2 production by incubation with blood for 10 minutes at 37°C, after which the samples were centrifuged (1000g, 10 minutes, at 4°C). The supernatant was acidified to pH 4.0 with 1N HCl and then subjected to the c18 reversed-phase chromatography (J.T. Baker, Inc.). TXB2, the breakdown product of TXA2, was quantified using EIA kit purchased from Cayman Chemicals.

Quantification of cytokine production

Raw264.7 mouse macrophage cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Sigma, St. Louis, MO, USA), and 107 Ti particles were added to cultures in the presence or absence of the COX-2 inhibitor (celecoxib). The supernatant was collected after 16 h and PGE2 was measured by competitive enzyme immunoassay kit (Cayman Chemicals). IL-6, IL-1β, and TNF-α levels were quantified using ELISA (BD Pharmingen, San Diego, CA, USA).

Statistical analysis

Data are expressed as means ± SEM. And statistical significance was determined by one-way analysis of variance (ANOVA). A p value < 0.05 was considered statistically significant.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

Celecoxib specifically inhibits COX-2-mediated local PGE2 production in response to Ti wear debris in vivo

We and others have shown previously that surgical implantation of Ti wear debris on the mouse calvaria can induce a pronounced inflammatory response accompanied by osteolysis in the calvarial bone.(17,29,30) The inflammatory infiltrate of variable intensity is observed in the soft tissues overlying the calvaria and also in some of the marrow cavities. This fibroinflammatory tissue is laden with Ti particles, fibroblast-like cells, monocytes/macrophages, and lymphocytes. The most intense bone resorption typically occurs around the sagittal suture line and in marrow cavities in the parietal bone between the posterior lambdoid and anterior frontal sutures. Osteoclastogenesis peaks on day 5. Between day 7 and 21, new bone formation takes place and some of the resorption sites are filled with new woven bone that can be distinguished readily from the preexisting lamellar bone because of its distinctive staining pattern. This model reproducibly shows a 40–70% calvarial erosion surface after Ti stimulation.(29)

To examine a possible role of COX-2 in this bone resorption model, we first measured local PGE2 production in the calvaria with and without Ti implantation. The COX-2 selective inhibitor celecoxib (25 mg/kg) was administered to mice daily for 7 days after Ti administration, and then the calvaria were harvested to determine the production of local PGE2. Four mice in each group were included in this experiment. As shown in Fig. 1, PGE2 levels in the tissue increased significantly from 13.9 ± 1.9 μg/g to 24.0 ± 1.1 μg/g of tissue on Ti stimulation, and celecoxib significantly inhibited the production of PGE2 down to 16.1 ± 3.0 μg/g of tissue.

thumbnail image

Figure FIG. 1.. Inhibition of Ti-induced PGE2 production by celecoxib. Mouse calvaria were harvested on day 7 after surgery without Ti implantation (sham), after Ti implantation (Ti), or after Ti implantation and daily administrations of 25 mg/kg celecoxib. Tissues were homogenized in the presence of 10 μM indomethacin and the supernatants were collected and analyzed for PGE2 concentration as described in the Materials and Methods section. The data represent the mean ± SEM of four mice in each group. PGE2 production is expressed as micrograms per gram of the tissue.

Download figure to PowerPoint

It has been reported that some COX-2 selective inhibitors also may inhibit COX-1, particularly at higher doses.(31) Because it is known that blood levels of TXB2 are derived from platelet COX-1 and reflect the activity of this enzyme in vivo,(32) we measured TXB2 blood levels in mice treated with celecoxib to determine its specificity at the doses used in this study. In this experiment, we included both COX-1−/− and COX-2−/− mice as controls. As expected, COX-1−/− mice had very low levels of TXB2 whereas COX-2−/− and wild-type mice had considerably higher levels of TXB2 (Fig. 2), confirming that COX-1 is the major source of TXB2 in blood. The blood level of TXB2 in mice given celecoxib remained at the same level as the control wild-type mice, showing that celecoxib did not inhibit COX-1 at the dose used in the experiments and that the inhibition of PGE2 production by the drug was predominantly caused by the inhibition of COX-2.

thumbnail image

Figure FIG. 2.. TXB2 levels in mouse blood. All mice were implanted with Ti for 10 days. One group of mice received celecoxib daily at the dose indicated. On day 10, 2 h after the mice were given the drug, whole blood was collected and TXB2 levels were measured as described in the Materials and Methods section. Each value represents the mean ± SEM of three mice in each group. ***p < 0.001.

Download figure to PowerPoint

Immunocytochemical staining was performed on paraffin embedded tissue sections to examine COX-2 expression in the mouse calvaria in response to Ti (Fig. 3). This study showed that COX-2 staining was minimal in sham control mice, but its expression was significantly induced as early as day 3 and continued to day 10 after Ti implantation. COX-2 staining was observed in inflammatory cells on the surface of the calvaria and its induction paralleled the inflammatory and bone resorption responses (data not shown). Interestingly, we found that COX-2 immunoreactivity was strongest in osteoblasts on previously resorbed surfaces and on the periosteal bone surface (Figs. 3B and 3C). Further study using consecutive sections revealed that TRAP+ osteoclasts in the bone resorption lacunae did not stain positively for COX-2 (data not shown). These data indicate that activated osteoblasts are the primary bone cells that express COX-2 during high bone turnover in this model.

thumbnail image

Figure FIG. 3.. COX-2 expression is induced in mouse calvaria implanted with Ti. Immunohistochemistry for COX-2 was performed as described in the Materials and Methods section. (A and B) Photomicrographs of COX-2 immunostaining in calvaria (A) without Ti and (B) with Ti at day 10 (magnification ×20). The reddish-brown color indicates positive staining, and the arrows indicate COX-2-positive osteoblasts. (C) Photomicrograph of a section adjacent to panel B stained with COX-2-specific antibodies at higher magnification (×40) shows the COX-2 immunoreactivity localizes in the cytoplasm of osteoblasts actively involved in matrix synthesis along previously resorbed surfaces. The arrowhead points to an osteoclast negative for COX-2.

Download figure to PowerPoint

Selective inhibition of COX-2 reduces inflammation and bone resorption in wear debris-induced osteolysis

Assuming that COX-2 is the primary enzyme involved in producing high levels of PGE2 during inflammation, we hypothesized that a COX-2 selective inhibitor would be efficacious in preventing bone resorption induced by Ti wear debris. To test this hypothesis, celecoxib (10 or 25 mg/kg per day) was given to the mice via gavage for 10 days after the implantation of the particles. All the mice tolerated this drug at all doses used. On day 10 calvaria were removed and prepared for histomorphometric analyses. Representative histological sections are illustrated in Fig. 4, and the statistical data are summarized in Fig. 5. At doses of 10 mg/kg per day and 25 mg/kg per day, celecoxib significantly reduced the Ti-induced area of periosteal inflammation by 57 ± 3% and 63.5 ± 3%, respectively. The drug also inhibited osteoclastic bone resorption by 60 ± 3% and 69 ± 2%, respectively, and reduced the number of TRAP+ cells per square millimeter TBA by 36 ± 2.5% and 43 ± 4%, respectively. The exaggerated bone resorption in this model is always followed by extensive new bone formation at the same site after day 5 and is considered to be an indicator of resorption that has taken place.(22) Therefore, we measured the new woven bone formation and found that the celecoxib-treated mice had much less new bone formation, which is consistent with inhibition of resorption and osteoclastogenesis (Fig. 5). We did not find a statistical difference between the group treated with 10 mg/kg and 25 mg/kg celecoxib; therefore, the lower dose is effective and could be used in future studies.

thumbnail image

Figure FIG. 4.. Effects of celecoxib on wear debris-induced osteolysis. Histology of the sagittal suture and parietal bone lateral to the sagittal suture of mice (A and D) without Ti, (B and E) Ti treated, and (C and F) Ti treated with daily administration of 25 mg/kg of celecoxib are shown. All sections were stained with H&E and photographed at ×20 magnification. (B and E) Bone resorption was observed in the sagittal suture and parietal bones in mice treated with Ti. New bone formation (arrows) was evident on day 10 after Ti implantation. (C and F) Mice treated with celecoxib exhibited reduced resorption, new bone formation, and inflammatory responses to Ti.

Download figure to PowerPoint

thumbnail image

Figure FIG. 5.. Effects of Ti on indices of inflammation and bone resorption in the calvaria of mice with or without celecoxib. Mouse calvaria were harvested at day 10 from untreated (sham) or Ti-treated mice. The number of mice used in each group is indicated in parenthesis. Osteoclast number is expressed per square millimeter TBA. Resorption surface is expressed as a percentage of the total bone/marrow interface. New bone formation is expressed as percentage of TBA. Periosteal inflammation is expressed as the percentage of the area of the soft tissue on top of the calvarium per millimeter bone surface divided by that of the control (Ti treated alone). ANOVA test was used to show the statistical differences between the Ti- and celecoxib-treated groups. **p < 0.02.

Download figure to PowerPoint

COX-2 but not COX-1 knockout mice display reduced bone resorption in response to Ti wear debris

To further identify the role of COX-2 in wear debris-induced osteolysis, we evaluated the response to Ti in COX-1 and COX-2 knockout mice.(23,24) Six- to eight-week-old mice of different genotypes from the same litters were used to examine inflammatory and osteolytic responses in calvaria 10 days after Ti implantation. Sections from comparable levels in the calvaria from all the mice were used to perform histomorphometry analysis as described in the Materials and Methods section. Representative histological sections are presented in Fig. 6 and the statistical histomorphometry data are shown in Fig. 7. Consistent with previous reports on inflammatory responses in COX-2−/− mice,(24) the periosteal inflammatory reaction to Ti in the calvaria of COX-2−/− mice was similar to that observed in the wild-type littermates (p > 0.05; Fig. 7A). However, the COX-2−/− mice displayed a significantly reduced bone resorption response, as indicated by reduction of osteoclast number, resorption surface, and new bone formation in response to Ti when compared with their wild-type littermates (Fig. 7; paired t-test, p < 0.05). The response in COX-1−/− mice was not significantly different from that in the wild-type mice. Because the osteoclastogenenic response in this model peaks at day 5, and at this time new bone formation has just begun at the site of the previous bone resorption, we conducted another experiment and measured the osteoclastogenesis and bone resorption response at day 5. As shown in Fig. 7B, COX-2−/− mice again had significantly reduced osteoclastogenesis and bone resorption as indicated by reductions in osteoclasts by 61 ± 4%, and total bone resorption surface by 50 ± 1%. In contrast to COX-2−/− mice, COX-1−/− mice exhibit a similar response as the wild-type mice. Taken together, these data provide strong evidence to show that COX-2 is involved directly in wear debris-induced, inflammation-driven bone resorption, independent of the inflammatory response in vivo.

thumbnail image

Figure FIG. 6.. COX-2−/− mice have less bone resorption in response to Ti wear debris than wild-type or COX-1−/− mice. Histology of both the sagittal suture and the parietal bone lateral to the midsagittal suture of (A, D, and G) wild type, (B, E, and H) COX-2−/−, and (C, F, and I) COX-1−/− mice are shown. Sections shown in panels A-F were stained with H&E and photographed at ×20 magnification. Bone resorption was observed in both sagittal suture and parietal bones. New bone formation (arrows in panels D and F) was evident on day 10 after Ti implantation. Although COX-2−/− mice form an inflammatory membrane in response to Ti, all of them had reduced bone resorption compared with wild-type littermates and COX-1−/− mice. In another experiment, calvaria were harvested on day 5 after Ti treatment. Sections were stained for TRAP activity to examine osteoclastogenesis. The arrows in panels G-I indicate TRAP+ osteoclasts. (H) COX-2−/− mice exhibited mitigated osteoclastogenesis on Ti implantation (magnification ×20), in comparison with (G) wild-type mice or (I) COX-1−/− mice.

Download figure to PowerPoint

thumbnail image

Figure FIG. 7.. Effects of Ti wear debris on the calvaria of COX-1−/−, COX-2−/−, and wild-type mice. Calvaria were harvested on (A) day 10 or (B) day 5 after Ti implantation. Tissue sections were prepared as described in the Materials and Methods section. Osteoclast number is expressed per square millimeter TBA. Resorption surface is expressed as a percentage of the total bone/marrow interface. The ANOVA test was used to show statistical differences (p < 0.05) followed by paired Student's t-test to show the statistical significance between wild type and the COX-2−/− group (*p < 0.05). No significant differences were found between wild type and the COX-1−/− (p > 0.05).

Download figure to PowerPoint

Inflammatory cytokine-induced osteoclastogenesis in vitro is dependent on COX-2

In our current study, we observed a reduction in bone resorption and osteoclastogenesis in mice treated with celecoxib and in COX-2−/− mice. To examine whether COX-2 is involved in modulating osteoclastogenesis in vitro, bone marrow culture assays were carried out to examine the differentiation of osteoclasts in the presence or absence of COX-2. As illustrated in Fig. 8, osteoclastogenesis from wild-type bone marrow cells (white bars) was induced by 1,25(OH)2D3 and further enhanced by TNF-α or IL-1β by about 2-fold, indicating that TNF-α or IL-1β synergized with vitamin D3 to produce more osteoclasts in vitro. Celecoxib at 5 μM significantly inhibited TRAP+ multinucleated cell formation by 65–90%. This inhibition was restored completely by addition of exogenous 10−7 M PGE2. Notably, bone marrow cultures from COX-2 knockout mice formed 60–85% fewer osteoclasts in response to vitamin D3, TNF-α, and/or IL-1β compared with those from wild-type or COX-1−/− mice. Again, the addition of PGE2 at 10−7 M to the cultures increased TRAP+ multinucleated cell numbers to the same level seen in the wild-type cells treated with vitamin D3. However, PGE2 seems to show no further promotion of the formation of TRAP+ multinucleated cells in the presence of IL-1β and TNF-α. In contrast to COX-2−/− cultures, we found that COX-1−/− bone marrow cells formed similar numbers of osteoclasts in response to vitamin D3 and cytokines as the wild-type cells. These results show that COX-2 is required for maximal osteoclastogenesis induced by vitamin D3 and proinflammatory cytokines such as IL-1 and TNF-α in vitro.

thumbnail image

Figure FIG. 8.. COX-2-dependent PGE2 is required for osteoclastogenesis in vitro. Bone marrow cells were collected as described in the Materials and Methods section. Vitamin D3 (10−8 M), IL-1β (10 ng/ml), and TNF-α (10 ng/ml) were added as indicated into marrow cultures to induce osteoclast formation. Celecoxib (5 μM) and PGE2 (10−7 M) also were used as indicated. Each value represents the average number of TRAP+ multinucleated cells (more than three nuclei) from 4 wells.

Download figure to PowerPoint

Celecoxib selectively inhibits PGE2 but not proinflammatory cytokine production by macrophages in response to Ti wear debris

The cytokines TNF-α, IL-1β, and IL-6, have been identified as important mediators for wear debris-induced osteolysis.(16–19,33) Macrophages activated by Ti are responsible for the production of these resorptive factors.(34) To further elucidate the mechanism by which COX-2 contributes to wear debris-induced osteolysis, we examined the production of these factors by macrophages stimulated with Ti in the presence or absence of celecoxib. Raw 264.7 macrophages were stimulated in culture with 5 × 106 Ti particles/ml for 16 h with or without various doses of celecoxib. The PGE2, TNF-α, IL-1β, and IL-6 production in the culture media were then measured by ELISA. In the absence of celecoxib, Ti induced PGE2 production from 0.67 ± 0.18 ng/ml to 7.5 ± 0.26 ng/ml, TNF-α production from 86 ± 3.5 ng/ml to 1000 ± 31 ng/ml, IL-1β production from 33.7 ± 18.9 ng/ml to 1000 ± 35 ng/ml, and IL-6 production from 6 ± 0.16 μg/ml to 141 + 13.0 μg/ml, respectively. As shown in Fig. 9, the increased production of PGE2 was inhibited by celecoxib in a dosage-dependent manner whereas cytokine production was not significantly affected (analysis of variance [ANOVA], p > 0.05). These results show that PGE2 production from macrophages stimulated with Ti is dependent on COX-2, whereas proinflammatory cytokine synthesis is not.

thumbnail image

Figure FIG. 9.. The effects of celecoxib on macrophage production of inflammatory mediators in response to Ti wear debris. Raw 264.7 macrophage cells were exposed to Ti particles for 16 h. Supernatants were then analyzed for PGE2, IL-1β, IL-6, and TNF-α by ELISA. The induction of these factors by Ti is expressed as a percentage of the amount present in the cultures without celecoxib. Each measurement is an average value from four wells. ANOVA tests show no significant difference among control and treated groups for the production of IL-1β, TNF-α, or IL-6 (p > 0.05). In contrast, PGE2 production was significantly inhibited by celecoxib beginning at 1 μM (p < 0.05)

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

Aseptic loosening of prostheses remains the primary cause of early implant failure and represents the most common complication of total joint arthroplasty.(33,35) The current understanding of its etiology ascribes a central role for macrophages in the development of the disease. Phagocytosis of wear debris from implants by macrophages leads to the release of various inflammatory mediators that initiate a chronic inflammatory response and eventually result in progressive bone loss. Identification of mediators in this disease process is of paramount importance, because it could lead to the development of therapeutic agents targeted to the prevention of aseptic loosening. Along these lines, potential treatments of osteolysis induced by implant wear debris have already been envisaged, such as the inhibition of TNF-α,(29) IL-1,(19) and matrix metalloprotease activity.(36)

Prostaglandins have long been suspected as one of the important mediators of aseptic loosening, based on the known roles of prostaglandins in bone metabolism(1) and the abundance of PGE2 in periprosthetic membranes of failed implants.(20,37) Nonspecific COX-2 inhibitors have been shown to reduce resorption in preclinical animal models.(37,38) However, clinical trials of short-term prophylactic treatment with indomethacin did not show a significant beneficial effect in patients with prosthetic implants.(39,40) Two recent breakthroughs prompt us to reexamine the role of cyclo-oxygenases in orthopedic implant loosening. The first is the generation of mice specifically deficient in COX-1 and COX-2.(23,24) Together with new advances in mouse models that mimic the human pathological conditions,(17,29) these animals permit the evaluation of the role of these enzymes in wear debris-induced osteolysis in vivo. More importantly, specific COX-2 inhibitors have been developed and have already been shown to have protective effects on bone destruction in rat models of arthritis.(41–43) These new COX-2 inhibitors (i.e., celecoxib [Celebrex] and Rofecoxib [Vioxx]) are now widely prescribed for patients with rheumatoid and osteoarthritis.(41,44) Because the destructive effects of these diseases on bone and joint cartilage are the most common indication for total joint arthroplasty, it seems most appropriate that these drugs be investigated for their effects on wear debris-induced osteolysis.

Here, we used a selective inhibitor of COX-2, celecoxib, and COX-1−/− and COX-2−/− mice to show a direct role of COX-2 in wear debris-induced osteolysis in vivo. Although the mouse calvaria model we used in the study has discrepancies from clinical aseptic loosening of prosthetic implants (e.g., no implant, no mechanical load, and flat bones containing a periosteum are studied instead of long bones in which the periosteum is removed), several laboratories have found this model to be useful in analyzing the molecular and cellular events involved in wear debris-induced osteolysis. Our findings showed that pharmacologic inhibition of COX-2, at doses that did not inhibit COX-1 activity (Fig. 2), dramatically inhibited Ti-induced PGE2 production, inflammation, osteoclastogenesis, and bone resorption in vivo (Figs. 1, 4, and 5). These results were corroborated by our studies with COX-2−/− mice, which had reduced osteoclastogenesis and bone resorption in vivo in response to Ti (Figs. 6 and 7). Moreover, the resorption data obtained from the COX-1−/− mice (Figs. 6, 7, and 9) did not show any major differences compared with wild-type controls. As has been shown in several inflammatory models,(24,31,45) we found that celecoxib significantly reduced periosteal inflammation (p < 0.01), while no differences in inflammation were observed between the COX-2−/− and wild-type mice. Because inflammation is a complex biological response orchestrated by a myriad of biological and chemical mediators with which prostaglandins play a synergistic role, it is possible that COX-2 mutant mice may have developed a COX-2-independent inflammatory response. However, our finding indicates that maximal induction of osteoclastogenesis and bone resorption requires the function of COX-2, which cannot be compensated by the presence of COX-1 or other factors.

The reduced bone resorptive response in COX-2−/− mice in the presence of intense inflammation directly supports COX-2 as an important mediator in bone resorption and indicates that COX-2 is involved directly in the modulation of osteoblast and osteoclast function. A recent report from Okada et al.(46) shows that COX-2 is required for vitamin D3 and parathyroid hormone (PTH)-induced osteoclastogenesis in vitro and in vivo. In the current study, we observed a significant reduction in osteoclastogenesis in COX-2−/− mice in vivo and in vitro in response to inflammatory signals. In bone marrow cultures, we observed a strong inhibitory effect of celecoxib on osteoclastogenesis induced by vitamin D3 and the inflammatory cytokines IL-1β and TNF-α and this inhibition was not further reduced in the marrow cells derived from COX-2−/− mice (Fig. 8). We also noticed that the synergistic effects between vitamin D3 and the cytokines IL-1β and TNF-α cannot be restored in COX-2−/− cells by the addition of exogenous PGE2, indicating that other prostaglandins generated by COX-2 may be necessary for the synergistic action of resorptive cytokines and vitamin D3.

PGE2 has been studied extensively in the past, and like many resorptive cytokines has pleiotropic effects on bone metabolism. Prostaglandins can act directly on osteoblasts to regulate their activity and also regulate osteoclastic bone resorption indirectly via their interaction with osteoblasts. Recently, PGE2 has been shown to stimulate expression of osteoclast differentiation factor (ODF)/receptor activator of NFκB ligand (RANKL)/osteoprotegerin ligand (OPGL)/TNF-related activation cytokine (TRANCE) in osteoblasts.(47–49) Wani et al. further showed that PGE2 strongly synergizes with ODF in osteoclast differentiation, cell spreading, and fusion, suggesting that local PGE2 can function as a cofactor with RANKL in modulation of osteoclast function.(50) Based on these findings it is therefore conceivable that COX-2, which we find is induced strongly at the bone inflammation interface, produces increased amount of prostaglandins that may further potentiate and amplify inflammation and bone resorption induced by Ti particles in vivo. This can occur directly by increasing bone resorption, or indirectly through the regulation of osteoclastogenic factors like ODF and inhibitors like osteoprotegerin (OPG).

Another mechanism by which COX-2 is involved in osteoclastogenesis is in response to inflammatory signals, such that COX-2 may regulate the production of proinflammatory cytokines.(43,51) Because both proinflammatory cytokine and prostaglandins are produced by macrophages in response to wear debris(18) and both are potent stimulators of osteoclastic bone resorption, we performed in vitro experiments to examine cytokine production in the presence of celecoxib. We found that pharmacologic inhibition of COX-2 in macrophages exposed to Ti can inhibit completely PGE2 production without significantly affecting TNF- α, IL-1β, or IL-6 production (Fig. 9), suggesting that inhibition of cytokine production is not the mechanism for the inhibition of bone resorption by celecoxib. In view of the fact that TNF-α, IL1β, and IL-6 can stimulate COX-2 expression in vitro, it is possible that the osteoclast stimulating effects of these cytokines in this model are mediated partially through COX-2. Further studies are underway to resolve this issue.

Overall, this study provides direct evidence for an important role of COX-2, that is distinct from that of COX-1, in inflammation-mediated bone loss. COX-2 contributes to increased bone loss in inflammatory conditions in which the bone resorption is accelerated after the induction of this enzyme. Selective inhibition of COX-2 may prevent or reduce bone loss in inflammation-induced bone disease.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References

We thank Dr. Kerry O'Banion and Dr. Alice Pentland for critical reading of the manuscript. We also thank Jennifer Harvey for her assistance with histological work. This work was supported by a Research Grant from the Orthopedic Research Education Foundation and a National Health Service Award AR45971.


  1. Top of page
  2. Abstract
  7. Acknowledgements
  8. References
  • 1
    Kawaguchi H, Pilbeam CC, Harrison JR, Raisz LG 1995 The role of prostaglandins in the regulation of bone metabolism. Clin Orthop 313:3646.
  • 2
    Hershman HR 1994 Regulation of prostaglandin synthase-1 and prostaglandin synthase 2. Cancer Metastasis Rev 13:241.
  • 3
    Simmons DL, Levy DB, Yannoni Y, Erikson RL 1989 Identification of a phorbol ester-repressible v-src-inducible gene. Proc Natl Acad Sci USA 86:11781182.
  • 4
    Kujubu D, Herschman H 1992 Dexamethasone inhibits mitogen induction of the TIS10 prostaglandin synthase/cyclooxygenase gene. J Biol Chem 267:79917994.
  • 5
    O'Banion MK, Sadowski HB, Winn V, Young DA 1991 A serum and glucocorticoid-regulated 4-kilobase mRNA encodes a cyclooxygenase-related protein. J Biol Chem 266:23261.
  • 6
    O'Banion MK, Winn V, Young DA 1992 cDNA cloning and functional activity of a glucocorticoid-regulated inflammatory cyclooxygenase. Proc Natl Acad Sci USA 89:4888.
  • 7
    Herschman HR 1996 Prostaglandin synthase 2. Biochim Biophys Acta 1229:125140.
  • 8
    Reddy ST, Herschman HR 1997 Prostaglandin synthase-1 and prostaglandin synthase-2 are coupled to distinct phospholipases for the generation of prostaglandin D2 in activated mast cells. J Biol Chem 272:32313237.
  • 9
    Reddy ST, Herschman HR 1994 Ligand-induced prostaglandin synthesis requires expression of the TIS10/PGS-2 prostaglandin synthase gene in murine fibroblasts and macrophages. J Biol Chem 269:1547315480.
  • 10
    Langenbach R, Loftin CD, Lee C, Tiano H 1999 Cyclooxygenase-deficient mice. A summary of their characteristics and susceptibilities to inflammation and carcinogenesis. Ann NY Acad Sci 889:5261.
  • 11
    Min YK, Rao Y, Okada Y, Raisz LG, Pilbeam CC 1998 Regulation of prostaglandin G/H synthase-2 expression by interleukin-1 in human osteoblast-like cells. J Bone Miner Res 13:10661075.
  • 12
    Tokushima T, Sato T, Morita I, Murota S 1997 Involvement of prostaglandin endoperoxide H synthase-2 in osteoclast formation induced by parathyroid hormone. Adv Exp Med Biol 433:307309.
  • 13
    Tai H, Miyaura C, Pilbeam CC, Tamura T, Ohsugi Y, Koishihara Y, Kubodera N, Kawaguchi H, Raisz LG, Suda T 1997 Transcriptional induction of cyclooxygenase-2 in osteoblasts is involved in interleukin-6-induced osteoclast formation. Endocrinology 138:23722379.
  • 14
    Kotake S, Udagawa N, Takahashi N, Matsuzaki K, Itoh KS, Ishiyama S, Saito S, Inoue K, Kamatani N, Gillespie MT, Martin TJ, Suda T 1999 IL-17 in synovial fluids from patients with rheumatoid arthritis is a potent stimulator of osteoclastogenesis. J Clin Invest 103:13451352.
  • 15
    Morinaga Y, Fujita N, Ohishi K, Zhang Y, Tsuruo T 1998 Suppression of interleukin-11-mediated bone resorption by cyclooxygenases inhibitors. J Cell Physiol 175:247254.
  • 16
    Jacobs JJ, Sumner DR, Galante JO 1993 Mechanisms of bone loss associated with total hip replacement. Orthop Clin North Am 24:583590.
  • 17
    Merkel KD, Erdmann JM, McHugh KP, Abu-Amer Y, Ross FP, Teitelbaum SL 1999 Tumor necrosis factor-alpha mediates orthopedic implant osteolysis. Am J Pathol 154:203210.
  • 18
    Blaine TA, Pollice PF, Rosier RN, Reynolds PR, Puzas JE, O'Keefe RJ 1997 Modulation of the production of cytokines in titanium-stimulated human peripheral blood monocytes by pharmacological agents. The role of cAMP-mediated signaling mechanisms. J Bone Joint Surg Am 79:15191528.
  • 19
    al Saffar N, Revell PA 1994 Interleukin-1 production by activated macrophages surrounding loosened orthopaedic implants: A potential role in osteolysis. Br J Rheumatol 33:309316.
  • 20
    Goodman SB, Chin RC, Chiou SS, Schurman DJ, Woolson ST, Masada MPA 1989 Clinical-pathologic-biochemical study of the membrane surrounding loosened and nonloosened total hip arthroplasties. Clin Orthop 244:182187.
  • 21
    Hukkanen M, Corbett SA, Batten J, Konttinen YT, McCarthy ID, Maclouf J, Santavirta S, Hughes SP, Polak JM 1997 Aseptic loosening of total hip replacement. Macrophage expression of inducible nitric oxide synthase and cyclo-oxygenase-2, together with peroxynitrite formation, as a possible mechanism for early prosthesis failure. J Bone Joint Surg Br 79:467474.
  • 22
    Boyce BF, Aufdemorte TB, Garrett IR, Yates AJ, Mundy GR 1989 Effects of interleukin-1 on bone turnover in normal mice. Endocrinology 125:11421150.
  • 23
    Langenbach R, Morham SG, Tiano HF, Loftin CD, Ghanayem BI, Chulada PC, Mahler JF, Lee CA, Goulding EH, Kluckman KD 1995 Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration. Cell 83:483492.
  • 24
    Morham SG, Langenbach R, Loftin CD, Tiano HF, Vouloumanos N, Jennette JC, Mahler JF, Kluckman KD, Ledford A, Lee CA 1995 Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 83:473482.
  • 25
    Zhang X, Morham SG, Langenbach R, Young DA 1999 Malignant transformation and antineoplastic actions of nonsteroidal anti-inflammatory drugs (NSAIDs) on cyclooxygenase-null embryo fibroblasts. J Exp Med 190:451459.
  • 26
    Hughes DE, Dai A, Tiffee JC, Li HH, Mundy GR, Boyce BF 1996 Estrogen promotes apoptosis of murine osteoclasts mediated by TGF-beta. Nat Med 2:11321136.
  • 27
    Mundy GR, Roodman GD, Bonewald LF, Oreffo RO, Boyce BF 1991 Assays for bone resorption and bone formation. Methods Enzymol 198:502510.
  • 28
    Takahashi N, Yamana H, Yoshiki S, Roodman GD, Mundy GR, Jones SJ, Boyde A, Suda T 1988 Osteoclast-like cell formation and its regulation by osteotropic hormones in mouse bone marrow cultures. Endocrinology 122:13731382.
  • 29
    Child LM, Goater JJ, Sanz I, Ritchlin C, Looney R, Zwillich SJ, Schwarz EM 1999 Efficacy of the soluble TNF-a inhibitor (Enbrel) to prevent prosthetic wear debris-induced osteolysis. Arthritis Rheum 42(Suppl.):S154.
  • 30
    Schwarz EM, Lu AP, Goater JJ, Benz EB, Kollias G, Rosier RN, Puzas JE, O'Keefe RJ 2000 TNFa/NFkB signaling in periprosthetic osteolysis. J Orthop Res 18:472480.
  • 31
    Wallace JL, Bak A, McKnight W, Asfaha S, Sharkey KA, MacNaughton WK 1998 Cyclooxygenase 1 contributes to inflammatory responses in rats and mice: Implications for gastrointestinal toxicity. Gastroenterology 115 :101109.
  • 32
    Smith CJ, Zhang Y, Koboldt CM, Muhammad J, Zweifel BS, Shaffer A, Talley JJ, Masferrer JL, Seibert K, Isakson PC 1998 Pharmacological analysis of cyclooxygenase-1 in inflammation. Proc Natl Acad Sci USA 95:1331313318.
  • 33
    Stea S, Visentin M, Granchi D, Melchiorri C, Soldati S, Sudanese A, Toni A, Montanaro L, Pizzoferrato A 1999 Wear debris and cytokine production in the interface membrane of loosened prostheses. J Biomater Sci Polym Ed 10:247257.
  • 34
    Blaine TA, Rosier RN, Puzas JE, Looney RJ, Reynolds PR, Reynolds SD, O'Keefe RJ 1996 Increased levels of tumor necrosis factor-alpha and interleukin-6 protein and messenger RNA in human peripheral blood monocytes due to titanium particles. J Bone Joint Surg Am 78:11811192.
  • 35
    Horowitz SM, Algan SA, Purdon MA 1996 Pharmacologic inhibition of particulate-induced bone resorption. J Biomed Mater Res 31:9196.
  • 36
    Takagi M, Santavirta S, Ida H, Ishii M, Mandelin J, Konttinen YT 1998 Matrix metalloproteinases and tissue inhibitors of metalloproteinases in loose artificial hip joints. Clin Orthop 352:3545.
  • 37
    Goodman SB, Chin RC, Chiou SS, Lee JS 1991 Suppression of prostaglandin E2 synthesis in the membrane surrounding particulate polymethylmethacrylate in the rabbit tibia. Clin Orthop 221:300304.
  • 38
    Goodman SB, Chin RC, Magee FP 1992 Prostaglandin E2 production by the membrane surrounding loose and fixated cemented tibial hemiarthroplasties in the rabbit knee. Clin Orthop 288:283287.
  • 39
    Kjaersgaard-Andersen P, Schmidt SA, Pedersen NW, Kristensen SS, Pedersen P 1989 Early radiolucencies following cemented total hip replacement. Influence of postoperative treatment with indomethacin. Orthopedics 12:401406.
  • 40
    Trnka HJ, Zenz P, Zembsch A, Easley M, Ritschl P, Salzer M 1999 Stable bony integration with and without short-term indomethacin prophylaxis. A 5-year follow-up. Arch Orthop Trauma Surg 119:456460.
  • 41
    Chan CC, Boyce S, Brideau C, Charleson S, Cromlish W, Ethier D, Evans J, Ford-Hutchinson AW, Forrest MJ, Gauthier JY, Gordon R, Gresser M, Guay J, Kargman S, Kennedy B, Leblanc Y, Leger S, Mancini J, O'Neill GP, Ouellet M, Patrick D, Percival MD, Perrier H, Prasit P, Rodger I 1999 Rofecoxib [Vioxx, MK-0966; 4-(4′-methylsulfonylphenyl)-3-phenyl-2-(5H)-furanone]: A potent and orally active cyclooxygenase-2 inhibitor. Pharmacological and biochemical profiles. J Pharmacol Exp Ther 90:551560.
  • 42
    Masaki M, Matsushita M, Wakitani K 1998 Inhibitory effects of JTE-522, a novel prostaglandin H synthase-2 inhibitor, on adjuvant-induced arthritis and bone changes in rats. Inflamm Res 7:187192.
  • 43
    Anderson GD, Hauser SD, McGarity KL, Bremer ME, Isakson PC, Gregory SA 1996 Selective inhibition of cyclooxygenase (COX)-2 reverses inflammation and expression of COX-2 and interleukin 6 in rat adjuvant arthritis. J Clin Invest 97:26722679.
  • 44
    Simon LS, Lanza FL, Lipsky PE, Hubbard RC, Talwalker S, Schwartz BD, Isakson PC, Geis GS 1998 Preliminary study of the safety and efficacy of SC-58635, a novel cyclooxygenase 2 inhibitor: Efficacy and safety in two placebo-controlled trials in osteoarthritis and rheumatoid arthritis, and studies of gastrointestinal and platelet effects. Arthritis Rheum 41:15911602.
  • 45
    Gavett SH, Madison SL, Chulada PC, Scarborough PE, Qu W, Boyle JE, Tiano HF, Lee CA, Langenbach R, Roggli VL, Zeldin DC 1999 Allergic lung responses are increased in prostaglandin H synthase-deficient mice. J Clin Invest 104:721732.
  • 46
    Okada Y, Lorenzo JA, Freeman1 AM, Tomita M, Morham GS, Raisz LG, Pilbeam C 2000 Prostaglandin G/H synthase-2 is required for maximal formation of osteoclast-like cells in culture. J Clin Invest 105:823832.
  • 47
    Tsukii K, Shima N, Mochizuki S, Yamaguchi K, Kinosaki M, Yano K, Shibata O, Udagawa N, Yasuda H, Suda T, Higashio K 1998 Osteoclast differentiation factor mediates an essential signal for bone resorption induced by 1 alpha,25-dihydroxyvitamin D3, prostaglandin E2, or parathyroid hormone in the microenvironment of bone. Biochem Biophys Res Commun 246:337341.
  • 48
    Yasuda H, Shima N, Nakagawa N, Yamaguchi K, Kinosaki M, Mochizuki S, Tomoyasu A, Yano K, Goto M, Murakami A, Tsuda E, Morinaga T, Higashio K, Udagawa N, Takahashi N, Suda T 1998 Osteoclast differentiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL. Proc Natl Acad Sci USA 95:35973602.
  • 49
    Nakagawa N, Yasuda H, Yano K, Mochizuki Si, Kobayashi N, Fujimoto H, Shima N, Morinaga T, Chikazu D, Kawaguchi H, Higashio K 1999 Basic fibroblast growth factor induces osteoclast formation by reciprocally regulating the production of osteoclast differentiation factor and osteoclastogenesis inhibitory factor in mouse osteoblastic cells. Biochem Biophys Res Commun 265:158163.
  • 50
    Wani MR, Fuller K, Kim NS, Choi Y, Chambers T 1999 Prostaglandin E2 cooperates with TRANCE in osteoclast induction from hemopoietic precursors: Synergistic activation of differentiation, cell spreading, and fusion. Endocrinology 140:19271935.
  • 51
    Votta BJ, Bertolini DR 1994 Cytokine suppressive anti-inflammatory compounds inhibit bone resorption in vitro. Bone 15:533538.