Trolox Prevents Osteoclastogenesis by Suppressing RANKL Expression and Signaling^,

Abstract

Excessive receptor activator of NF-κB ligand (RANKL) signaling causes enhanced osteoclast formation and bone resorption. Thus, down-regulation of RANKL expression or its downstream signals may be a therapeutic approach to the treatment of pathological bone loss. In this study, we investigated the effects of Trolox, a water-soluble vitamin E analogue, on osteoclastogenesis and RANKL signaling. Trolox potently inhibited interleukin-1-induced osteoclast formation in bone marrow cell-osteoblast coculture by abrogating RANKL induction in osteoblasts. This RANKL reduction was attributed to the reduced production of prostaglandin E₂ via a down-regulation of cyclooxygenase-2 activity. We also found that Trolox inhibited osteoclast formation from bone marrow macrophages induced by macrophage colony-stimulating factor plus RANKL in a reversible manner. Trolox was effective only when present during the early stage of culture, which implies that it targets early osteoclast precursors. Pretreatment with Trolox did not affect RANKL-induced early signaling pathways, including MAPKs, NF-κB, and Akt. We found that Trolox down-regulated the induction by RANKL of c-Fos protein by suppressing its translation. Ectopic overexpression of c-Fos rescued the inhibition of osteoclastogenesis by Trolox in bone marrow macrophages. Trolox also suppressed interleukin-1-induced osteoclast formation and bone loss in mouse calvarial bone. Taken together, our findings indicate that Trolox prevents osteoclast formation and bone loss by inhibiting both RANKL induction in osteoblasts and c-Fos expression in osteoclast precursors.

Bone development and remodeling depend on maintaining a delicate balance between bone resorption by osteoclasts and bone formation by osteoblasts (1). Excessive osteoclastic bone resorption plays a critical role in bone destruction in pathological bone diseases such as osteoporosis, rheumatoid arthritis, periodontal disease, and some metastatic cancers (2). Osteoclastic bone resorption entails several complicated processes as follows: commitment from hematopoietic progenitors, fusion of the cells, development of a ruffled border and a clear zone, and secretion of acids and lysosomal enzymes into a resorbing area (3).

Two molecules, macrophage colony-stimulating factor (M-CSF)3 and receptor activator of NF-κB (RANK) ligand (RANKL), which are mainly produced by osteoblasts and stromal cells, are essential for osteoclast formation from osteoclast precursors (4, 5). RANKL can sufficiently promote osteoclast development from osteoclast precursors in vitro in the presence of M-CSF (4, 6). Binding of RANKL to its receptor RANK causes receptor trimerization and recruits several adaptor molecules, such as tumor necrosis factor receptor-associated factor 6, which activates multiple downstream signaling pathways, including NF-κB, c-Jun N-terminal protein kinase (JNK), p38, ERK, and PI3K/AKT (7–9).

Several transcription factors, including PU.1, microphthalmia-associated transcription factor, NF-κB, c-Fos, and nuclear factor of activated T cells c1 (NFATc1, also known as NFAT2), have been shown to play a role in osteoclast development. PU.1 and microphthalmia-associated transcription factor act on early nonspecific differentiation along the osteoclast pathway (10), whereas NF-κB, c-Fos, and NFAT2 function downstream of RANKL signaling for osteoclast development. RANKL can activate the NF-κB pathway, and mice lacking the p50 and p52 NF-κB subunits develop osteopetrosis caused by the arrested generation of osteoclasts (11, 12). Stimulation of RANKL in osteoclast precursors induces a dramatic up-regulation of c-Fos, which is a member of the AP-1 transcription factor family, and c-Fos-deficient mice fail to form osteoclasts, with an elevated number of bone marrow macrophages (BMMs), and develop osteopetrosis (13, 14). Recently, it was shown that NF-κB functions as an upstream signal of c-Fos during RANKL-induced osteoclastogenesis. RANKL-induced c-Fos up-regulation is abolished in NF-κB p50/p52 double knock-out osteoclast precursors, and RANKL can induce osteoclast formation from NF-κB p50/p52 double knock-out osteoclast precursors when c-Fos is overexpressed (15). NFAT2 is also up-regulated in RANKL-stimulated osteoclast precursors through mechanisms dependent on NF-κB and c-Fos (15, 16). Furthermore, when overexpressed, NFAT2 can induce the differentiation of osteoclast precursors into mature osteoclasts even in the absence of RANKL (17). Thus, NFAT2 is thought to be a master transcription factor for osteoclast differentiation.

Trolox (6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid) is a hydrophilic derivative of α-tocopherol with a carboxylic group instead of a phytyl chain. Trolox has advantages over α-tocopherol because of its enhanced cell permeability. It has been shown to have beneficial effects on several types of cells. It prevents cisplatin-induced apoptosis in renal proximal tubular epithelial cells (18), inhibits singlet oxygen-induced DNA damage in human lymphoblast cells (19), and protects human red blood cells during photodynamic treatment (20). Moreover, Trolox selectively enhances arsenic-mediated apoptosis in several cancer cells, including myeloma and breast cancer cells (21). Recently, Quintanilla et al. (22) showed that Trolox protects the neurotoxicity mediated by amyloid β-peptide and hydrogen peroxide through a mechanism that involves the modulation of the Wnt signaling pathway. In animal models, Trolox improves ischemia-induced hepatic drug-metabolizing dysfunction (23) and enhances the anti-lymphoma effects of arsenic trioxide (24). Despite its various beneficial effects, little is known about the effects of Trolox on osteoclastic bone loss. In this study, we investigate the effects of Trolox and the molecular mechanisms of its action on osteoclast development in vitro and in vivo. Here we show that Trolox inhibits osteoclast formation by suppressing RANKL induction and its critical downstream target gene, c-fos.

EXPERIMENTAL PROCEDURES

Reagents and Materials—Penicillin, streptomycin, α-MEM, and fetal bovine serum were purchased from Invitrogen. Recombinant soluble human M-CSF, human RANKL, and mouse IL-1α were obtained from PeproTech. Cycloheximide, MG-132, PD98059, SB203580, SP600125, BAY11-7082, and LY294002 were purchased from Calbiochem. Trolox, α-tocopherol, and N-acetylcysteine were purchased from Sigma. Specific antibodies for phospho-ERK1/2, ERK, phospho-JNK1/2, JNK, phospho-p38, p38, phospho-AKT, AKT, phospho-IκBα, IκBα (Cell Signaling Technology), c-Fos, NFAT2, β-actin (Santa Cruz Biotechnology), cytosolic phospholipase A₂ (Chemicon International), cyclooxygenase-2 (COX-2) (BD Transduction Laboratories), and membrane-associated prostaglandin E synthase-1 (Cayman Chemical) were used for Western blotting experiments.

In Vitro Osteoclast Formation Assay—Murine osteoclasts were generated from bone marrow cells as described previously (25). In brief, mouse bone marrow cells were obtained from femurs and tibias of 5-week-old ICR mice and incubated in α-MEM complete media containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 μg/ml streptomycin on 10-cm culture dishes in the presence of M-CSF (10 ng/ml) overnight. Nonadherent cells were transferred to 10-cm bacterial culture dishes and further cultured in the presence of M-CSF (30 ng/ml) for 3 days. Adherent cells were used as BMMs after the nonadherent cells were washed out. To generate osteoclasts, BMMs (4 × 10⁴ cells/well) were cultured for 4 days with M-CSF (30 ng/ml) and RANKL (100 ng/ml) in 48-well (1 ml/well) tissue culture plates. To generate osteoclasts from the coculture of primary osteoblasts and bone marrow cells, primary osteoblasts from newborn ICR mouse calvariae were prepared as described previously (26). Bone marrow cells (3 × 10⁵ cells/well) and primary osteoblasts were cocultured in 48-well (1 ml/well) tissue culture plates for 6 days in α-MEM complete medium. The complete medium was changed every 3rd day. At the end of the culture period, the cells were fixed in 10% formalin for 10 min, permeabilized with 0.1% Triton X-100, and then stained for tartrate-resistant acid phosphatase (TRAP) by using the leukocyte acid phosphatase assay kit (Sigma).

RANKL and OPG Protein Expression in Osteoblasts—Primary osteoblasts (4 × 10⁵ cells/well) were pretreated with or without Trolox or vehicle for 24 h and then stimulated with IL-1 (10 ng/ml) for 24 h. The amounts of RANKL protein in cell lysates and osteoprotegerin (OPG) secretion in cell culture media were determined by using RANKL and OPG enzyme-linked immunosorbent assay (R & D Systems) according to the manufacturer's instructions.

Measurement of PGE₂ Production—The concentration of PGE₂ in the culture medium was measured by using an enzyme immunoassay (Amersham Biosciences). The antibody had the following cross-reactivity determined by comparing the bound/free ratios with several eicosanoids as follows: PGE₂, 100%; PGE₁, 25%; PGF_2α, 0.04%; and 6-keto-PGF_1α, <0.1%.

In Vitro Recombinant Cyclooxygenase-2 (COX-2) Activity— Human recombinant COX-2 activity was measured with a COX activity assay kit following the manufacturer's instructions (Cayman Chemical). In brief, human recombinant COX-2 was incubated for 15 min with or without Trolox in the presence of heme and was then incubated for an additional 5 min with colorimetric substrate N,N,N′,N′-tetramethyl-p-phenylenediamine and arachidonic acid solution. The peroxidase activity of COX was assayed colorimetrically by monitoring the appearance of oxidized N,N,N′,N′-tetramethyl-p-phenylenediamine at 590 nm.

Cell Viability Assay—The XTT assay was performed to examine the effects of Trolox on the viability of BMMs by using a cell proliferation kit (Roche Applied Science) according to the manufacturer's instructions. BMMs (1 × 10⁴ cells/well) were cultured with Trolox at various concentrations for 48 h in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) on 96-well plates (200 μl/well). Then the medium was discarded, and fresh medium was added with 50 μl of XTT solution (XTT labeling reagent + electron-coupling reagent). After a 6-h incubation, the plate was read at 450 nm (650 nm reference) by using a 96-well plate recorder.

Survival and Bone Resorption Assay—Mature osteoclasts were prepared from the coculture of bone marrow cells and primary osteoblasts as described previously (26). Briefly, bone marrow cells (1 × 10⁷ cells) and primary osteoblasts (1 × 10⁶ cells) were seeded on collagen gel-coated 10-cm culture dishes and cultured for 6–7 days in the presence of 10^-8 m 1,25-dihydroxyvitamin D₃ (Sigma) and 10^-6 m PGE₂ (Sigma). Cells were detached by treating with 0.2% collagenase (Invitrogen) at 37 °C for 10 min and were then replated on culture dishes following the experiments. Osteoblasts were removed by treatment with 0.1% collagenase at 37 °C for 30 min. The remaining cells were considered as enriched mature osteoclasts. The survival assay was performed as follows; mature osteoclasts were incubated with various concentrations of Trolox for 1 h and further cultured in the presence of RANKL (100 ng/ml). After 24 h, TRAP staining was performed to detect surviving osteoclasts.

For the bone resorption assay, the cocultured osteoclasts were replated on OAAS plates (Osteogenic Core Technologies Inc., Korea) coated with carbonated calcium phosphate and were permitted to settle for 12 h. Cells on OAAS plates were incubated with or without Trolox at the indicated doses for 1 h and were then further cultured in the presence of RANKL (100 ng/ml). After 24 h, the cells were removed, and total resorption pits were photographed and analyzed by using the Image Pro-Plus program, version 4.0 (Media Cybernetics).

Retroviral Gene Transduction—The retroviral vectors pMX-IRES-EGFP, pMX-c-Fos-IRES-EGFP, and pMX-constitutively active (CA)-NFAT2-IRES-EGFP were kindly provided by Dr. Nacksung Kim (University of Chonnam, Gwangju, Korea). Retrovirus packaging was performed by transient transfection of these pMX vectors into Plat-E retroviral packaging cells. After incubation in fresh medium for 2 days, culture supernatants of the retrovirus-producing cells were collected. For retroviral infection, nonadherent bone marrow cells were cultured in M-CSF (30 ng/ml) for 48 h. The medium was then removed and replaced with culture supernatants of pMX-IRES-EGFP, pMX-c-Fos-IRES-EGFP, and pMX-CA-NFAT2-IRES-EGFP virus-producing Plat-E cells together with Polybrene (6 μg/ml) and M-CSF (30 ng/ml) for 8 h. Infected cells were cultured in the presence of M-CSF for 1 day and were then further cultured with puromycin (2 μg/ml) and M-CSF for 2 days to remove uninfected cells. Stably infected cells were cultured with or without Trolox (500 μm) in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) for 4 days. Forced expressions of each construct and osteoclast formation were detected by using a fluorescence microscope and TRAP staining, respectively.

Western Blot Analysis—Cells were washed twice with ice-cold phosphate-buffered saline (PBS) and were lysed in lysis buffer containing 20 mm Tris-HCl, 150 mm NaCl, 1% Triton X-100, 0.2% deoxycholate, and protease and phosphatase inhibitors for 30 min on ice. Protein concentrations of cell lysates were determined by using the DC protein assay kit (Bio-Rad). Equal amounts of protein (30 μg/lane) were resolved by SDS-PAGE and were then transferred to a polyvinylidene difluoride membrane (Amersham Biosciences). The membrane was probed with the indicated primary antibody. Blots were developed by using horseradish peroxidase-conjugated secondary antibodies and were visualized by using a chemiluminescence technique (Amersham Biosciences).

Nuclear Fraction and Electrophoretic Mobility Shift Assay— Cells were washed twice with ice-cold PBS, lysed in C buffer (50 mm Tris-HCl (pH 8.0), 2 mm EDTA, 0.5% Nonidet P-40, 20% glycerol, and 0.5 mm phenylmethylsulfonyl fluoride) for 5 min on ice, and then microcentrifuged at 4000 rpm for 5 min. The pellet was lysed in N buffer (20 mm HEPES (pH 7.6), 420 mm NaCl, 2 mm EDTA, 1% Triton X-100, 20% glycerol, 25 mm β-glycerophosphate, and 0.5 mm phenylmethylsulfonyl fluoride) for 30 min on ice and was then microcentrifuged at 12,000 rpm for 15 min. Supernatants were used as nuclear extracts. Electrophoretic mobility shift assay was performed as described previously (27). Briefly, nuclear extracts (10 μg) were incubated with the reaction buffer (10 mm Tris-HCl, 50 mm KCl, 1 mm EDTA, 5% glycerol, 2 mm dithiothreitol, and 1 μg of poly(dI·dC)) containing ∼20,000 cpm of ³²P-labeled NF-κB-binding site oligomer (5′-AGTTGAGGGGACTTTCCCAGGC-3′; Santa Cruz Biotechnology) for 30 min at 20 °C. The DNA-bound NF-κB proteins were separated on 4% polyacrylamide gels. The gels were dried and subjected to autoradiography.

Relative Quantitation of mRNAs by Real Time Reverse Transcription-PCR—Total RNA was prepared by using an RNeasy mini kit (Qiagen) according to the manufacturer's instructions, and cDNA was synthesized from 2 μg of total RNA by reverse transcriptase (Superscript II Preamplification System; Invitrogen). Real time PCR was performed on a Prism 7500 sequence detection system with SYBR® Green PCR Master Mix (Applied Biosystems) and following the manufacturer's protocols. The 7500 sequence detector (Applied Biosystems) was programmed with the following PCR conditions: 40 cycles of 15-s denaturation at 95 °C and 1-min amplification at 60 °C. All reactions were run in triplicate and were normalized to the housekeeping gene Hprt. The evaluation of relative differences of PCR results was calculated by using the comparative cycle threshold (C_T) method. The following primer sets were used: mouse RANKL forward, 5′-TGGAAGGCTCATGGTTGGAT-3′, and reverse, 5′-CATTGATGGTGAGGTGTGCA-3′; mouse OPG forward, 5′-TGGAACCCCAGAGCGAAACA-3′, and reverse, 5′-GCAGGAGGCCAAATGTGCTG-3′; mouse Nos2 forward, 5′-CAGCTGGGCTGTACAAACCTT-3′, and reverse, 5′-TAGCAGGCCTCTGACGAAGTG-3′; mouse c-Fos forward, 5′-ACTTCTTGTTTCCGGC-3′, and reverse, 5′-AGCTTCAGGGTAGGTG-3′; mouse NFAT2 forward, 5′-CCGTTGCTTCCAGAAAATAACA-3′, and reverse, 5′-TGTGGGATGTGAACTCGGAA-3′; mouse Hmox1 forward, 5′-GGTGATGCTGACAGAGGAACAC-3′, and reverse, 5′-TAGCAGGCCTCTGACGAAGTG-3′; and mouse βHPRT forward, 5′-CCTAAGATGAGCGCAAGTTGAA-3′, and reverse, 5′-CCACAGGGACTAGAACACCTGCTA-3′.

Microarray Analysis—Total RNA was isolated from cells as described previously. Samples were digested with DNase I to remove any residual DNA contamination. The methods for preparation of cRNA and subsequent steps leading to hybridization of Affymetrix GeneChip® mouse ST 1.0 arrays, washing, and scanning were performed according to standard protocol (Affymetrix). Briefly, 1 μg of total RNA was converted to first-strand DNA by using SuperScript II reverse transcriptase (Invitrogen) and T7-Oligo(dT)₂₄ primer containing the T7 RNA polymerase sequence (Metabion). Double-stranded cDNA was then synthesized by using a SuperScript Choice system (Invitrogen). Biotinylated cRNA was synthesized from 1 μg of double-stranded cDNA and was purified and fragmented. The fragmented and biotinylated cRNAs were subjected to hybridization with mouse Affymetrix GeneChip® ST 1.0 arrays with use of a hybridization control kit (Affymetrix). The chips were washed and stained with streptavidin/phycoerythrin-biotinylated anti-streptavidin antibody (Affymetrix). The washing and staining were performed by using GeneChip Fluidics Station 450. The stained chips were scanned by using a GeneChip® scanner 3000 7G (Affymetrix). Differentially expressed genes were categorized according to the biological processes in which they are involved, and the molecular functions they code for by using the PANTHER data base (available on line).

Pulse-Chase Experiment—BMMs were stimulated with M-CSF and RANKL for 24 h in the presence or absence of Trolox and were then radiolabeled for 1 h with 200 μCi/ml of an l-[³⁵S]methionine/cysteine cell-labeling mix (Amersham Biosciences) in methionine/cysteine-free α-MEM complete media for pulse labeling. After incubation, the cells were washed three times with complete media and were then incubated for 40 min with unlabeled methionine (100 μg/ml) and cysteine (500 μg/ml) for the chase. Cell lysates were immunoprecipitated with anti-c-Fos antibody, and the immunoprecipitates were subjected to SDS-PAGE. The gels were dried and subjected to autoradiography.

IL-1-induced Mouse Calvarial Bone Loss—A mouse model of IL-1-induced bone loss was used as described previously (25). In brief, a collagen sponge treated with vehicle (PBS) or IL-1 (1.5 μg) was implanted over calvarial bone in groups of five mice (5-week-old male ICR mice). Mice were intraperitoneally administered with the vehicle (dimethyl sulfoxide) or Trolox (60 mg/kg of body weight) daily beginning on day -1. The mice were killed 7 days after the implantation, and whole calvariae were fixed in 4% paraformaldehyde and stained for TRAP. Three-dimensional images of calvarial bone were obtained by micro-computed tomography (micro-CT) scanning (SMX-90CT, Shimadzu, Japan). For histological analysis, calvarial tissues were fixed in 4% paraformaldehyde, decalcified in 12% EDTA, and then embedded in paraffin. After that, histological sections (5 μm) were prepared, stained for TRAP, and counterstained by using hematoxylin. All animal experiments were reviewed and approved by the Seoul National University School of Dentistry Animal Care Committee.

Statistical Analysis—All quantitative data are represented as the mean ± S.D. Each experiment was performed three or four times, and the results from one representative experiment are shown. Statistical analyses were performed by using one-way analysis of variance (SPSS statistical package, version 12, SPSS Inc.) with p < 0.01 considered significant.

RESULTS

Trolox Inhibits Osteoclast Formation in Bone Marrow Cell-Osteoblast Coculture—We first investigated the effects of Trolox on osteoclastogenesis in bone marrow cell-osteoblast coculture. With IL-1 treatment, TRAP-positive multinucleated osteoclasts were formed in the coculture on days 5 and 6. Trolox strongly inhibited the IL-1-induced osteoclast formation (Fig. 1A). In this culture system, osteoblasts support osteoclastogenesis by up-regulating the ratio of RANKL to OPG (1, 2). Thus, we assessed the expression of RANKL and OPG in osteoblasts by using real time quantitative PCR and enzyme-linked immunosorbent assay. IL-1 increased RANKL expression and decreased OPG expression in osteoblasts. Treatment with Trolox suppressed the IL-1-induced RANKL expression, with no marked change in OPG expression (Fig. 1, B and C). To determine whether the RANKL reduction in osteoblasts was solely responsible for the inhibitory effect of Trolox, we next investigated whether the inhibitory effect of Trolox could be reversed by exogenous addition of RANKL. As shown in Fig. 1A, the inhibitory effect of Trolox was somewhat prevented by treatment with exogenous RANKL.

FIGURE 1.

Effects of Trolox on IL-1-induced osteoclast formation in cocultures. A, mouse bone marrow cells and primary osteoblasts were cocultured in the presence of IL-1 (10 ng/ml) with or without Trolox (500 μm) and RANKL (100 ng/ml) for 6 days. After culturing, the generated osteoclasts were detected by TRAP staining, and TRAP-positive multinucleated cells containing three or more nuclei were counted as osteoclasts (^*, p < 0.01). B, primary osteoblasts were pretreated with Trolox or vehicle (dimethyl sulfoxide) for 24 h and were then stimulated with IL-1 (10 ng/ml) for 24 h. After culturing, total RNA was isolated, and the expression of mRNA for RANKL and OPG was analyzed by real time quantitative PCR with HPRT mRNA as an endogenous control. C, osteoblasts were treated as in B. The amounts of RANKL (^*, p < 0.01 versus untreated control; #, p < 0.01 versus group treated with IL-1 only) and OPG (^*, p < 0.01 versus untreated control) were determined by using enzyme-linked immunosorbent assay kits in cell lysates and in cell culture media, respectively.

Trolox Decreases IL-1-induced PGE₂ Synthesis by Inhibiting COX-2 Activity—We first examined the early signaling pathways involved in IL-1-induced RANKL induction and the effects of Trolox on these pathways in osteoblasts. Pretreatment of osteoblasts with PD98059 (a MEK1/2 inhibitor), SB203580 (a p38 inhibitor), and LY294002 (a PI3K/AKT inhibitor), but not SP600125 (a JNK inhibitor) and BAY11-7082 (an NF-κB inhibitor), strongly suppressed IL-1-induced RANKL expression in osteoblasts. This finding suggests that the ERK, p38, and PI3K/AKT signaling pathways are involved in the IL-1 induction of RANKL (supplemental Fig. 1A). Treatment with Trolox, however, failed to affect the IL-1-activated signaling pathways (supplemental Fig. 1B). Previous reports have shown that PGE₂ production in osteoblasts is required for IL-1-induced RANKL induction and osteoclast formation (25, 28). Thus, we examined whether Trolox affects IL-1-induced PGE₂ production. The stimulation of IL-1 increased the secretion of PGE₂ into the culture medium of osteoblasts, which was significantly inhibited by Trolox in a time- and dose-dependent manner (Fig. 2, A and B). PGE₂ increased RANKL mRNA levels without affecting OPG levels in osteoblasts. However, unlike for IL-1, induction of RANKL by PGE₂ was not affected by Trolox. Furthermore, the addition of PGE₂ rescued the IL-1-induced RANKL expression inhibited by Trolox (Fig. 2C). We then examined how Trolox inhibits IL-1-induced PGE₂ synthesis. As shown in Fig. 2, D and E, Trolox suppressed the enzyme activity of recombinant COX-2 in a dose-dependent manner without affecting the expression levels of cytosolic phospholipase A₂, COX-2, and membrane-associated prostaglandin E synthase-1, which are involved in PGE₂ synthesis in response to proinflammatory stimuli (28, 29).

FIGURE 2.

Effects of Trolox on IL-1-induced PGE₂ synthesis in osteoblasts. A, primary osteoblasts were pretreated with Trolox (500 μm) or vehicle (dimethyl sulfoxide) for 24 h and were then stimulated with IL-1 (10 ng/ml) for the indicated times. B, osteoblasts pretreated with Trolox (100–500 μm) were stimulated with IL-1 (10 ng/ml) for 24 h. The PGE₂ concentration (^*, p < 0.01 versus untreated control; #, p < 0.01 versus group treated with IL-1 only) in the culture medium was determined by enzyme immunoassay. C, primary osteoblasts were treated with or without IL-1 (10 ng/ml), Trolox (500 μm), and PGE₂ (1 nm) for 24 h. Expression of mRNA for RANKL and OPG was analyzed by real time PCR with HPRT mRNA as an endogenous control. D, osteoblasts were treated as in A. Western blotting was performed with the indicated antibodies. E, human recombinant COX-2 was incubated with the indicated Trolox doses, and the COX-2 activity assay was performed as described under “Experimental Procedures” (^*, p < 0.01 versus untreated control).

Trolox Inhibits Osteoclastogenesis via a Direct Action on Osteoclast Precursors—RANKL is essential and sufficient for the differentiation of osteoclast precursors into mature osteoclasts in the presence of M-CSF (2, 4, 5). Thus, we next examined the effects of Trolox on RANKL-induced osteoclast formation from BMMs. When BMMs were incubated with M-CSF (30 ng/ml) and RANKL (100 ng/ml) for 4 days, numerous TRAP-positive multinucleated osteoclasts were generated. Treatment of the same cultures with Trolox suppressed osteoclast formation in a dose-dependent manner (Fig. 3A). The results of the XTT assay showed that the anti-osteoclastogenic effect of Trolox was not attributed to cellular toxicity or cell proliferation (Fig. 3B). We next examined at which stage Trolox impairs osteoclast development. Because complete inhibition of osteoclast formation was achieved with 500 μm Trolox, this concentration was subsequently used unless otherwise noted. Trolox was added to osteoclast-generating cultures on different days (0–3), and TRAP staining was performed on day 4. Trolox effectively inhibited osteoclast formation only when added on the first 2 days of culture, which suggests that it affects early osteoclastogenesis (Fig. 3C). We also found that the inhibitory effect of Trolox was reversible. Withdrawal of Trolox on day 2 after its treatment rescued osteoclast formation (Fig. 3D). To determine the action of Trolox on mature osteoclasts, mature osteoclasts were seeded on OAAS plates and incubated with or without Trolox and RANKL for 24 h. As shown in Fig. 3E, Trolox failed to reduce RANKL-induced bone resorbing activity at the tested doses. It also did not affect RANKL-induced survival of mature osteoclasts (Fig. 3F).

FIGURE 3.

Effects of Trolox on RANKL-induced osteoclast formation in BMMs. A, BMMs were cultured in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) with or without Trolox for 4 days. After culturing, cells were fixed, and the number of TRAP-positive multinucleated cells was counted (^*, p < 0.01 versus untreated control). B, BMMs were cultured for 48 h with M-CSF (30 ng/ml) and RANKL (100 ng/ml) at the indicated doses of Trolox. Then cell viability was determined by use of the XTT assay. C, as in A, except the cells were treated with 500 μm Trolox on the indicated days (^*, p < 0.01 versus untreated control). D, BMMs were treated with (panels ii and iii) or with out (panel i) Trolox at the beginning of the culture period, and Trolox was continued (panel ii) or withdrawn (panel iii) on day 2. E, cocultured osteoclasts were replated on OAAS plates as described under “Experimental Procedures,” incubated with or without Trolox (100 and 500 μm) for 1 h, and then further cultured in the presence of RANKL (100 ng/ml) for 24 h. Resorbed pits were photographed (left), and the pit areas were analyzed (right) after osteoclasts were removed. F, purified mature osteoclasts from the coculture were incubated with Trolox (100 and 500 μm) for 1 h were then and further cultured in the presence of RANKL (100 ng/ml) for 24 h. After that, surviving osteoclasts were detected by TRAP staining (left), and the number of osteoclasts was counted (right).

Trolox Does Not Alter the Activation of MAPKs, AKT, and NF-κB Induced by RANKL—M-CSF is required for the survival, proliferation, and differentiation of osteoclast precursors by binding to its receptor, c-Fms (5). In addition, engagement of RANK by RANKL stimulation on osteoclast precursors permits multiple intracellular signaling cascades that lead to osteoclast differentiation (1–3). Because Trolox inhibits osteoclast formation by targeting an early stage of differentiation, we examined the mRNA expression of c-Fms and RANK in BMMs. The expression of both receptors was unchanged by Trolox treatment (data not shown). To define the molecular mechanism of the inhibitory effects of Trolox on osteoclastogenesis, we next examined the effects of Trolox on the early signaling pathways induced by RANKL in BMMs. As shown in Fig. 4A, activation of ERK, JNK, p38, and AKT was observed at 5 min after RANKL treatment, which was not affected by pretreatment with Trolox. We also examined the effect of Trolox on the NF-κB signaling pathway. RANKL stimulation led to the phosphorylation and degradation of IκBα within 5 min. Trolox did not dampen this process (Fig. 4B). Furthermore, Trolox did not impair the RANKL-stimulated DNA binding activity of NF-κB (Fig. 4C).

FIGURE 4.

Effects of Trolox on RANKL-induced early signaling in BMMs. A and B, BMMs were pretreated with Trolox for 3 or 24 h in the presence of M-CSF (30 ng/ml) and were then stimulated with RANKL (100 ng/ml) for the indicated time points. Whole-cell lysates were subjected to Western blotting with the indicated antibodies. C, BMMs were pretreated with Trolox or vehicle (dimethyl sulfoxide) for 24 h in the presence of M-CSF (30 ng/ml) and were then stimulated with RANKL (100 ng/ml) for 15 min. Cells were harvested, and nuclear extracts were prepared. The DNA binding activity of NF-κB was assessed by electrophoretic mobility shift assay. N.S., nonspecific band.

Trolox Down-regulates c-Fos Protein Levels Induced by RANKL— The c-Fos/c-Jun/NFAT2 pathway plays a critical and fundamental role in osteoclast development, and the lack of any of these three arrests osteoclastogenesis (30). c-Fos and c-Jun, which are components of the AP-1 transcription factor, regulate the expression of NFAT2 by binding to the NFAT2 promoter (14, 16). Thus, we investigated the effects of Trolox on their expression levels. As reported previously, RANKL stimulation increased the expression of c-Fos and NFAT2 in BMMs. Interestingly, Trolox did abolish the RANKL-induced c-Fos and NFAT2 protein expression, without marked change in c-Fos mRNA expression (Fig. 5, A and B). As shown in Fig. 5, C and D, this specific decrease in c-Fos protein expression was observed in a dose-dependent manner.

FIGURE 5.

Effects of Trolox on the RANKL-induced expression of c-Fos and NFAT2 in BMMs. A and B, BMMs were pretreated with Trolox or vehicle (dimethyl sulfoxide) in the presence of M-CSF for 24 h and were then stimulated with RANKL (100 ng/ml) for the indicated times. C and D, BMMs were pretreated with Trolox (100–500 μm) or vehicle for 24 h and were then stimulated with RANKL for 24 h. A and C, expression of mRNA for c-Fos and NFAT2 was analyzed by real time PCR using HPRT mRNA as an endogenous control. B and D, Western blotting was performed with the indicated antibodies. Actin served as an internal control.

Trolox Inhibits the Translation of c-Fos Protein in BMMs— c-Fos protein has a very short life span as the result of its rapid degradation via the ubiquitin-proteasome pathway (31). Thus, we investigated whether Trolox enhances the degradation of c-Fos protein. Addition of cycloheximide, an inhibitor of new protein synthesis, for 4 h abolished the increase in c-Fos protein levels induced by RANKL (Fig. 6A, lane 2 versus lane 4), which was recovered by co-treatment with MG132, a selective inhibitor of 26 S proteasome (Fig. 6A, lane 4 versus lane 6). By contrast, the Trolox-induced reduction of c-Fos protein was not influenced by MG132 treatment (Fig. 6A, lane 5 versus lane 7), which suggests that proteasome-mediated degradation is not related to the Trolox-induced reduction in c-Fos protein. Thus, we next checked the biosynthesis of c-Fos protein by using a pulse-chase experiment. As shown in Fig. 6B, biosynthesis of c-Fos protein was distinctly increased by RANKL stimulation, which was abrogated by Trolox treatment. To investigate whether the reduction in c-Fos protein mediated the anti-osteoclastogenic effect of Trolox, we overexpressed the c-fos gene in BMMs by using a retroviral system. Trolox suppression of osteoclastogenesis was efficiently overcome by the forced expression of c-fos as well as CA-NFAT2 (Fig. 6C).

FIGURE 6.

Trolox-induced inhibition of c-Fos translation in BMMs. A, BMMs were pretreated with or without Trolox in the presence of M-CSF for 24 h and were then stimulated with RANKL (100 ng/ml). After 20 h, 1 g/ml cycloheximide (CHX) and 7 μm MG132 (MG) were added to the cultures for 4 h before harvest. c-Fos protein levels were detected by Western blotting. B, BMMs were pretreated with or without Trolox for 24 h and were then stimulated with RANKL (100 ng/ml) for 23 h. The cells then were metabolically radiolabeled with l-[³⁵S]methionine/cysteine for 1 h and collected after a 30-min chase time. The cell lysates were subjected to immunoprecipitation with anti-c-Fos antibody, and the immunoprecipitates were resolved by SDS-PAGE and detected by autoradiography. C, BMMs were infected with retroviruses expressing pMX-IRES-EGFP (GFP-vector), pMX-c-Fos-EGFP (c-Fos), and pMX-CA-NFAT2-EGFP (CA-NFAT2). Infected cells were cultured with or without Trolox in the presence of M-CSF (30 ng/ml) and RANKL (100 ng/ml) for 4 days. After culturing, the cells were fixed; the ectopic expression of each construct was detected by a fluorescence microscope (upper), and the cells were stained for TRAP. TRAP-positive multinucleated osteoclasts were counted (lower;^*, p < 0.01 versus untreated control).

Trolox Prevents IL-1-induced Osteoclast Formation and Bone Loss in Vivo—Because Trolox inhibited not only RANKL expression but also RANKL-induced osteoclast formation in vitro, we examined the in vivo efficacy of Trolox for the treatment of osteoclastic bone loss by using a mouse model of calvarial bone destruction (25). Implantation of an IL-1-soaked collagen sponge on mouse calvarial bone resulted in dramatic osteoclast formation, whereas administration of Trolox suppressed this IL-1-induced osteoclast formation (Fig. 7, A and C). In line with this, the results from the micro-CT and histological sections also showed that the osteoclast formation and bone loss induced by IL-1 were notably prevented by Trolox treatment (Fig. 7, B and C).

FIGURE 7.

Effects of Trolox on IL-1-induced bone destruction in vivo. A–C, a collagen sponge treated with vehicle (PBS) or IL-1 (1.5 μg) was implanted over mouse calvaria. Trolox (60 mg/kg of body weight, intraperitoneal) or vehicle (dimethyl sulfoxide) was administered daily. The mice were killed 7 days after implantation. A, TRAP staining of whole calvaria. B, three-dimensional images of calvarial bone by micro-CT analysis. C, TRAP staining of histological sections of calvarial bone, with hematoxylin counterstaining.

DISCUSSION

Several pro-inflammatory cytokines, including IL-1, IL-6, and tumor necrosis factor-α, can augment osteoclast formation by both direct effects on its precursors and indirect effects via osteoblasts (32, 33). In particular, IL-1 is thought to be a critical mediator of the pathological bone destruction induced by both estrogen deficiency and inflammation. Mice lacking the type I IL-1 receptor are resistant to bone destruction after estrogen deficiency (34). Blocking IL-1 signaling can also reduce bone erosion and cartilage degradation in animal models of rheumatoid arthritis (35).

In this study, Trolox potently prevented IL-1-induced RANKL induction in osteoblasts and osteoclast formation in bone marrow cell-osteoblast coculture. Trolox inhibited IL-1-induced PGE₂ production by down-regulating COX-2 activity, and the addition of PGE₂ recovered the inhibitory effect of Trolox on IL-1-induced RANKL induction. These results suggest that the reduced PGE₂ synthesis is, at least in part, involved in the Trolox inhibition of RANKL expression in osteoblasts.

We also found that Trolox has direct effects on osteoclast precursors as well as indirect effects via osteoblasts. The time course analysis showed that Trolox reversibly suppresses RANKL-induced osteoclast formation by targeting an early stage of differentiation. Furthermore, Trolox did not affect the RANKL-induced survival and resorption activity of mature osteoclasts. Diaz et al. (21) reported that Trolox selectively enhances arsenic-mediated cytotoxicity in acute promyelocytic leukemia and other malignant cell lines. In this study, the inhibitory effect of Trolox on osteoclastogenesis was not derived from its cytotoxicity. Treatment of BMMs with Trolox at the tested doses did not affect cell proliferation. Furthermore, when Trolox was removed, osteoclast differentiation continued.

To gain insights into the molecular mechanism underlying the Trolox-induced anti-osteoclastogenic effects, we investigated the effects of Trolox on RANKL-induced signaling pathways. Stimulation of RANKL has been reported to activate three well known signaling pathways, including NF-κB, MAPKs, and PI3K/AKT. Furthermore, these signaling pathways have been reported to be involved in RANKL-induced osteoclastogenesis. Mice lacking the p50 and p52 NF-κB subunits generate severe osteopetrosis caused by failure of osteoclast formation (11, 12). JNK1 has been shown to be required for RANKL-induced efficient osteoclast formation from BMMs (36). Each specific inhibitor of MEK and p38 or loss of AKT by using small interfering RNA prevented RANKL-induced osteoclastogenesis (9, 37, 38). However, in our study, Trolox did not have any effects on these RANKL-induced signaling pathways.

We next checked the transcription factors c-Fos and NFAT2, which are known to play a critical role in RANKL-induced osteoclast development. The RANKL-induced expression of c-Fos and NFAT2 were dramatically down-regulated in a dose-dependent manner with Trolox pretreatment. By contrast, RANKL-induced c-Fos mRNA expression was not affected by Trolox. Our pulse-chase experiments showed that the reduction in c-Fos protein contributed to the suppression of c-Fos translation. Furthermore, the forced expression of c-Fos or CA-NFAT2 rescued the Trolox-induced inhibition of osteoclastogenesis, which suggests that down-regulation of c-Fos is responsible for the inhibitory effects of Trolox. Recently, Takayanagi et al. (39) reported that IFN-β interferes with the RANKL-induced expression of c-Fos protein but not mRNA in BMMs by inhibiting c-Fos protein synthesis. The inhibitory action of IFN-β on osteoclastogenesis is linked to the activation of a gene induction pathway mediated by IFN-stimulated transcriptional factor 3 (the heterotrimeric complex consisting of STAT1, STAT2, and IRF-9). The anti-osteoclastogenic activity of IFN-β is abrogated in BMMs from mice lacking STAT1 or IRF-9. These results prompted us to examine whether IFN-β mediates the Trolox-induced reduction in c-Fos expression and its anti-osteoclastogenic effects. However, we did not find any effects of Trolox on IFN-β expression. Moreover, the inhibitory effect of Trolox was not rescued in BMMs from mice lacking STAT1 (data not shown). 1,25-Dihydroxyvitamin D₃ has also been shown to inhibit RANKL-induced osteoclast differentiation by suppressing the translation of c-Fos protein in BMMs without affecting the RANKL-induced NF-κB and p38/JNK pathways (40). Thus, ours and the previous observations (40) suggest that decreasing RANKL-induced c-Fos expression, particularly its protein expression, may be a promising therapeutic target for excessive osteoclast development.

In line with its in vitro effects, Trolox markedly suppressed IL-1-induced calvarial bone loss as well as osteoclast formation in vivo. The in vivo effects were most likely the result of the suppression of both IL-1-induced RANKL expression in osteoclast-supporting cells and RANKL-induced c-Fos expression in osteoclast precursors.

Several studies have suggested that reactive oxygen species are involved in osteoclast differentiation and bone loss. Thiol antioxidants have been shown to play a crucial role in estrogen-deficiency bone loss. Administration of N-acetylcysteine or ascorbate, an antioxidant that increases tissue glutathione levels, inhibits bone loss after ovariectomy (41). Recently, it was reported that intracellular reactive oxygen species generation mediates not only RANKL-induced signaling pathways, including MAPKs and NF-κB, but also the development and function of osteoclasts (42, 43). However, Trolox did not affect RANKL-induced proximal signaling pathways in this study. Unlike Trolox, α-tocopherol (50–500 μm) did not affect RANKL-stimulated osteoclast formation or the protein induction of c-Fos and NFAT2 (supplemental Fig. 2, A and B). In addition, the powerful antioxidant N-acetylcysteine (1–10 mm) potentiated rather than inhibited RANKL-stimulated osteoclast formation and c-Fos up-regulation (supplemental Fig. 2, C and D). Thus, it seems unlikely that the suppressed reactive oxygen species generation itself is involved in the inhibitory effects of Trolox on c-Fos expression and osteoclast formation.

To identify overall changes in gene expression by Trolox treatment, we performed a microarray-based gene expression analysis in osteoblasts and BMMs. The gene expression profiles of osteoblasts were compared between groups treated with vehicle, IL-1, or IL-1 plus Trolox. In a similar fashion, the gene expression changes of BMMs treated with vehicle, RANKL, or RANKL plus Trolox were compared. Genes regulated by Trolox in osteoblasts and BMMs were categorized according to their biological processes and functions (supplemental Table 1 and 2). No single gene was altered more than 2-fold by Trolox in both cell types. Trolox-induced changes in gene expression in osteoblasts and BMMs were primarily attributable to suppression of IL-1 and RANKL regulation, respectively. For example, the expression levels of the Nos2 gene dramatically increased in response to IL-1 in osteoblasts but were suppressed by Trolox treatment (supplemental Table 1). In BMMs, the expression levels of the Hmox1 gene, which was previously shown to regulate osteoclast differentiation (44), decreased in response to RANKL, whereas Trolox reversed the decrease (supplemental Table 2). The suppressive effects of Trolox on the two genes were confirmed with real time PCR analyses (supplemental Fig. 3). However, we need further studies to determine whether the Trolox-induced changes in gene expression are involved in its suppressive effects on RANKL and c-Fos expression in osteoblasts and BMMs, respectively.

In summary, our findings clearly show that Trolox inhibits osteoclast formation via direct suppression of c-Fos expression in osteoclast precursors as well as through indirect inhibition of RANKL expression in osteoblasts. Trolox also prevented IL-1-induced osteoclastic bone loss. These results strongly suggest that Trolox may have therapeutic value for treating or preventing several bone diseases characterized by excessive bone destruction.