TRAF2 Is Essential for TNF-α-Induced Osteoclastogenesis

Authors


  • The authors have no conflict of interest.

Abstract

TRAF2-deficient mice show embryonic lethality, and we developed a new in vitro differentiation system to show the function of TRAF2 in osteoclastogenesis, in which osteoclast progenitors are derived from the fetal liver of TRAF2-deficient mice. Using this system, we showed that TRAF2 is required for TNF-α-induced osteoclastogenesis.

Introduction: TNF receptor-associated factor 2 (TRAF2) is a signal transducer for RANK and for two TNF receptor isotypes, TNFR1 and TNFR2. Because TRAF2-deficient mice show embryonic lethality, it has remained unclear whether TRAF2 is crucial in RANKL- or TNF-α-induced osteoclastogenesis.

Materials and Methods: Osteoclast progenitors derived from fetal liver were cultured in the presence of monocyte macrophage colony-stimulating factor (M-CSF), and flow cytometry for characterization of surface markers on these cells was performed. To examine the involvement of TRAF2 in osteoclast differentiation, we cultured osteoclast progenitors from TRAF2-deficient and wildtype mice with soluble RANKL or TNF-α in the presence of M-CSF, and counted the number of TRACP+ multinucleate cells formed. c-jun N-terminal kinase (JNK) and NF-κB activation in osteoclast progenitors was examined by Western blot analysis and electrophoretic mobility shift assay, respectively. Nuclear factor of activated T cells (NFATc1) expression and activation were analyzed by RT-PCR and immunofluorescence staining, respectively. To examine whether TRAF2 overexpression induced osteoclastogenesis, TRAF2 was overexpressed in osteoclast progenitors form wildtype bone marrow by retrovirus infection.

Results and Conclusions: Osteoclast progenitors from normal fetal liver, which were cultured with M-CSF, expressed surface molecules c-fms, Mac-1, and RANK, and could differentiate into TRACP+ multinucleate cells in the presence of soluble RANKL or TNF-α. RANKL-induced osteoclastogenesis gave a reduction of 20% in the progenitors from TRAF2-deficient mice compared with that of the cells from littermate wildtype mice, whereas TNF-α-induced osteoclastogenesis was severely impaired in the cells from the TRAF2-deficient mice. Only a few TRACP+ multinucleate cells were formed, and TNF-α-mediated activation of JNK, NF-κB, and NFATc1 was defective. TRAF2 overexpression induced differentiation of osteoclast progenitors from wildtype mice into TRACP+ multinucleate cells. These results suggest that TRAF2 plays an important role in TNF-α-induced osteoclastogenesis.

INTRODUCTION

OSTEOCLASTS REGULATE BONE mass and calcium metabolism by resorbing bone.(1) These cells are derived from common progenitors in the macrophage lineage and are able to differentiate through different pathways according to the type of microenvironmental stimulus. Osteoclast differentiation is controlled by RANKL.(2–5) When the osteoclast activity is enhanced, various diseases such as rheumatoid arthritis occur, and the production of inflammatory cytokines such as TNF-α and interleukin-1 (IL-1) increases. TNF-α has the potential to induce osteoclast differentiation,(1,6,7) and TNF-α together with soluble RANKL strongly promotes osteoclast differentiation and activation in vitro.(7,8) Recently, TNF-α has become noteworthy as a target for the treatment of inflammatory bone diseases, and it is expected that understanding of molecules mediating TNF-α signals will give us clues to develop new therapies for such diseases.

TNF receptor-associated factors (TRAFs) are intracellular adapter proteins that are involved in signal transduction pathways initiated by a variety of TNF receptor family members and other cell surface receptors.(9) Out of the six known TRAFs, TRAF2, TRAF5, and TRAF6 were shown to activate NF-κB and c-jun N-terminal kinase (JNK).(10) TRAF6 interacts with RANK and plays a crucial role in RANKL-induced osteoclastogenesis. In our previous studies employing osteoclast progenitors from TRAF-deficient mice, TRAF6 was shown to be a key regulator of TNF-α-induced osteoclastogenesis in vitro,(11) and TRAF5 functioned in both RANKL- and TNF-α-induced osteoclastogenesis in vitro and in vivo.(12) On the other hand, although TRAF2 also associates with RANK and TNF receptors, no studies on osteoclastogenesis in TRAF2-deficient mice have been done.

Gene-targeting studies have revealed that some essential factors, such as RANK, RANKL, and TRAF6, are involved in the activation of NF-κB and JNK in vitro.(4,5,9) For investigation of the molecular mechanism underlying osteoclast differentiation, an in vitro system using osteoclast progenitors obtained from young mouse spleen or bone marrow has been established.(11,13) However, no in vitro stromal cell-free systems using the cells from fetal organs have yet been established, and TRAF2-deficient mice show embryonic lethality.(14) Therefore, it has remained unresolved as to whether TRAF2 participates in osteoclast differentiation.

Osteoclasts are derived from hematopoietic stem cells.(15–17) In young mammals, all blood cells mainly originate from a population of hematopoietic stem cells localized in the bone marrow.(18) In the mouse embryo, hematopoiesis initially happens in yolk sac blood islands. After the blood circulation starts at embryonic day 9, the site of hematopoiesis changes to the liver and then to the bone marrow. After embryonic day 13, the liver plays a major role in fetal hematopoiesis until embryonic day 19, when bone marrow hematopoiesis begins. In a previously published study using an in vitro co-culture system, osteoclast progenitors were found to be already present in the yolk sac at embryonic day 8, and osteoclasts could be derived from various tissues such as liver, spleen, thymus, and kidney at embryonic day 17.(19)

In this study, we show that we could obtain osteoclast progenitors from the mouse fetal liver at embryonic day 14.5 and that these cells could differentiate into TRACP+ cells. Therefore, this in vitro system made it possible for us to examine the role of TRAF2 in osteoclast differentiation. RANKL-induced osteoclastogenesis was slightly defective in osteoclast progenitors from TRAF2-deficient mice compared with that in the cells from wildtype littermates, whereas TNF-α-induced osteoclastogenesis was severely impaired. Thus, we have provided the first evidence showing that TRAF2 functions in osteoclast differentiation.

MATERIALS AND METHODS

Animals, plasmids, and reagents

TRAF2-deficient (−/−) and wildtype (+/+) mice used in this study were described previously,(14) and the genotypic analysis of these mice was performed.(14) Male ddy mice, 6-9 weeks of age, were used (Sankyo Laboratories, Tokyo, Japan). pMXs-IRES-EGFP was kindly donated by Dr Kitamura (University of Tokyo). pBluescript-FLAG-TRAF2 was generously provided by Dr Nakano (Juntendo University). After pBluescript-FLAG-TRAF2 was digested with EcoRI, a fragment of FLAG-TRAF2 was cloned into the EcoRI site of the expression plasmid pMXs-IRES-EGFP. Recombinant products of mouse TNF-α and mouse IL-1β were purchased from Pepro-Tech EC (London, UK). Recombinant mouse soluble RANKL (sRANKL) tagged with glutathione S-transferase was a gift from Dr Teitelbaum (Washington University). Rat monoclonal antibodies against c-fms (AFS98) and against RANK (1E6.66) were kindly donated by Dr Nishikawa (Kyoto University) and Dr Choi (Pennsylvania University), respectively. Rat monoclonal antibody against CD14 (rmC5-3) and biotin-conjugated rat antibody against Mac-1 (CD11b) were purchased from Pharmingen (San Diego, CA, USA). Hamster monoclonal antibodies against mouse TNF-Rs p55 and p75 were obtained from Genzyme (Cambridge, MA, USA). These antibodies were biotinylated using the enhanced chemiluminescence (ECL) protein biotinylation module (Amersham, Buckinghamshire, UK). Mouse monoclonal antibody against human NFATc1 (7A6) was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Fluorescein isothiocyanate (FITC)-conjugated streptavidin and horseradish peroxidase (HRP)-conjugated anti-rat immunoglobulin G (IgG) antibodies came from Molecular Probes (Eugene, OR, USA), and Cappel Research Products (Durham, NC, USA), respectively. Animal studies were conducted under a protocol approved by the Institutional Animal Use and Care Committee.

Primary cells from fetal liver and bone marrow

Monocyte macrophage colony-stimulating factor-dependent fetal liver-derived macrophage cell culture:

We used TRAF2−/− mice at embryonic day 14.5, because there were no developmental or morphological abnormalities up to this stage. Fetuses at embryonic day 14.5 were removed from the uterus and amniotic sacs. The fetal liver was aseptically removed from each embryo and minced using a sterile needle. The fetal liver cells were collected and washed with α-MEM, and red blood cells were removed by treatment with 0.747% NH4Cl-0.017% Tris-Cl (pH7.2) PBS solution. After being washed with α-MEM containing 10% FCS, the cells were cultured in α-MEM containing 10% FCS, 100 IU/ml penicillin G (Meiji Seika, Tokyo, Japan), 100 μg/ml streptomycin (Meiji Seika), and 35,000 units/ml of monocyte macrophage colony-stimulating factor (M-CSF) at 3 × 106 cells in a 10-cm suspension culture dish (Corning Costar Instruments, Corning, NY, USA). After 3 days in culture, the cells were washed vigorously with PBS twice to remove nonadherent cells and harvested by pipetting with 0.02% EDTA in PBS. The number of M-CSF-dependent fetal liver-derived macrophage (MDFLM) cell culture cells, 2-3 × 106 cells per mouse, was obtained at embryonic day 14.5.

M-CSF-dependent bone marrow-derived macrophage cell culture:

Bone marrow cells from adult mice were cultured in α-MEM containing 10% FCS in the presence of 35,000 units/ml of M-CSF supernatant in a 10-cm suspension culture dish as previously described.(13)

Flow cytometry

Cells were harvested and labeled with the desired antibodies. Nonspecific signals were calculated and attenuated by incubation with the fluorochrome-conjugated secondary antibody in the absence of the primary antibody. The stained cells were analyzed with a flow cytometer (FACSCalibur; Becton Dickinson, San Jose, CA, USA).

TRACP staining and bone resorption assay

Cultured cells were fixed with 10% formalin for 5 minutes, refixed with ethanol-acetone (50:50 vol/vol) for 1 minute, and incubated at room temperature for 20 minutes in 100 mM acetate buffer (pH 4.8) containing 0.01% naphtol AS-MX phosphate (Sigma Chemical, St Louis, MO, USA), 0.06% fast red violet LB salt (Sigma Chemical), and 50 mM sodium tartrate. Bone resorption assays were performed by culturing cells on calcium phosphate-coated discs (osteologic; BD Biosciences, Kingston, Canada) in the presence of M-CSF supernatant and cytokines for 6 days.

JNK kinase assay

MDFLM cells (5 × 106 cells) were stimulated with 50 ng/ml of sRANKL or TNF-α for 15 or 30 minutes, and JNK activity was examined using a JNK assay kit according to the manufacturer's protocol (New England Biolabs, Bevely, MA, USA). Briefly, after lysis of the cells, an amino-terminal c-jun fusion protein bound to glutathione-Sepharose beads was used to precipitate JNK from cell lysates. The kinase reaction was carried out using the c-jun fusion protein as a substrate in the presence of cold ATP. Phosphorylation of Ser-63 in the c-jun fusion protein was measured by Western blot analysis using a rabbit anti-phospho-c-jun polyclonal antibody followed by an HRP-conjugated goat anti-rabbit IgG polyclonal antibody. The blots were developed with an ECL (Amersham Pharmacia Biotech, Uppsala, Swegen). To strip bound antibodies mentioned above, we incubated the membranes in stripping buffer (62.5 mM Tris/HCl [pH 6.8], 100 mM 2-mercaptoethanol, 2% SDS) at 50°C for 30 minutes and detected the c-jun fusion protein with a rabbit anti-c-jun polyclonal antibody.

Electrophoretic mobility shift assay

Nuclear extracts were prepared as described by Dignam et al.(20) The sequence of the NF-κB-binding oligonucleotide used as a radioactive DNA probe was 5′-AGCTTGGGGACTTTCCGA-3′. The DNA binding reaction was performed at room temperature in a volume of 20 μl, which contained the binding buffer (10 mM Tris-HCl [pH 7.5], 1 mM EDTA, 4% glycerol, 100 mM NaCl, 5 mM dithiothreitol, and 100 mg/ml BSA), 3 μg of poly (dI-dC), 2 × 105 cpm of [32P]labeled probe, and 5 μg of nuclear proteins. After incubation for 15 minutes, the samples were electrophoresed on native 5% acrylamide gels prepared in 0.25× Tris borate-EDTA. The gels were dried and exposed to X-ray film.

Immunofluorescence staining

Cells were fixed in 4% paraformaldehyde for 30 minutes and treated with 0.1% Triton X for 5 minutes. After incubation with 5% BSA/PBS for 30 minutes, the cells were stained with rhodamine-conjugated phalloidin for actin (Molecular Probes) or with 2 μg/ml of mouse anti-NFATc1 monoclonal antibody (Santa Cruz Biotechnology) followed by 4 μg/ml of Alexa 488 anti-mouse IgG antibody (Molecular Probes), each for 60 minutes.

RT-PCR analysis

Total RNA from cultured cells was isolated by the guanidine/phenol method. cDNA were synthesized from 1 μg of total RNA using reverse transcriptase and oligo-dT primers in a volume of 10 μl, and the reaction mixture was finally adjusted to 50 μl with TE buffer (10 mM Tris-1 mM EDTA, pH 8.0) for PCR analysis. PCR was performed with 1 μl of cDNA reaction mixture using Taq polymerase and appropriate primers in a volume of 20 μl. For primers, nuclear factor of activated T cells (NFATc1) 5′-CAACGCCCTGACCACCGATAG-3′ and 5′-GGCTGCCTTCCGTCTCATAGT-3′(21); cathepsin K 5′-GCAGTATAACAGCAAGGTGG-3′ and 5′-TGAAAGCCCAACAGGAACCA-3′, MMP9 5′-GTATGGTCGTGGCTCTAAGC3′ and 5′-AAAACCCTCTTGGTCTGCGG-3′, and GAPDH 5′-ACTTTGTCAAGCTCATTTCC-3′ and 5′-TGCAGCGAACTTTATTGATG-3′ were used. Ten-microliter aliquots of PCR products were separated by electrophoresis on a 2% agarose gel and stained with ethidium bromide.

Virus infection

Virus-containing supernatants were prepared using the Phoenix-E packaging cells and the retrovirus vector, pMXs-TRAF2-IRES-EGFP or pMXs-IRES-EGFP. After MDBM cells were incubated with virus supernatants for 8 h, supernatants were removed, and the cells were cultured without sRANKL in the presence of M-CSF (35,000 units/ml).

RESULTS

Osteoclast differentiation of MDFLM cells

TRAF2 is an adapter molecule that mediates activation of the osteoclast differentiation factors such as NF-κB and JNK.(9) We previously reported that TRAF6 and TRAF5 were required for RANKL- and TNF-α-induced osteoclastogenesis in an in vitro stromal cell-free system using TRAF6- and TRAF5-deficient mice, in which osteoclast progenitors are derived from young mouse spleen or bone marrow.(11,12) However, it has been difficult to elucidate whether TRAF2 participates in osteoclastogenesis because TRAF2-deficient mice show embryonic lethality.(14) First, we obtained osteoclast progenitors (MDFLM cells) by culturing mouse fetal liver cells at embryonic day 14.5. By flow cytometry, these cells were shown to express the surface molecules Mac-1, c-fms, RANK, CD14, TNF-R1, and TNF-R2 (Fig. 1A), indicative of the typical phenotypes of osteoclast progenitors. When MDFLM cells were cultured with 50 ng/ml sRANKL in the presence of M-CSF for 5 days, they differentiated into TRACP+ mononuclear and multinucleate cells (MNCs; Fig. 1C). These TRACP+ cells formed actin rings and resorbed calcium phosphate-coated discs (Figs. 1E and 1G, respectively). TNF-α at 50 ng/ml induced differentiation of MDFLM cells into TRACP+ MNCs (Fig. 1D) that showed formation of actin rings together with bone resorption activity under 10 ng/ml IL-1β treatment (Figs. 1F and 1H, respectively). No TRACP+ cells were formed when sRANKL or TNF-α was not added (data not shown). Furthermore, we examined whether these TRACP+ cells expressed cathepsin K and MMP9 by RT-PCR, and we confirmed that MDFLM cells did not express mRNA of cathepsin K and MMP9, whereas these transcripts were detected in the cells stimulated with 50 ng/ml sRANKL or 50 ng/ml TNF-α plus 10 ng/ml IL-1β for 4 days (Fig. 1B). Thus, this in vitro stromal cell-free system allowed us to elucidate a particular molecule that functions in osteoclast differentiation.

Figure FIG. 1..

Characterization of wildtype MDFLM cells. MDFLM cells derived from day 14.5 wildtype fetal liver are positive for c-fms, Mac-1, CD14, TNF-R1, TNF-R2, and RANK. (A) For a negative control (dotted line), the staining was performed with only FITC-conjugated secondary antibody. Total RNA was extracted from MDFLM cells (lane 1), and these MDFLM cells were stimulated with 50 ng/ml sRANKL (lane 2) or 50 ng/ml TNF-α plus 10 ng/ml IL-1β (lane 3) for 4 days in the presence of M-CSF. (B) The expression of cathepsin K, MMP9, and GAPDH was determined by RT-PCR. In the presence of M-CSF (35,000 units/ml), MDFLM cells were induced by (C) 50 ng/ml sRANKL or (D) 50 ng/ml TNF-α to differentiate into TRACP+ MNCs. For actin ring formation, MDFLM cells were cultured with (E) 50 ng/ml sRANKL or (F) 50 ng/ml TNF-α in the presence of M-CSF (35,000 units/ml) for 5 days. Fixed cells were stained with rhodamine-conjugated phalloidin. Pit formation was assessed by culturing MDFLM cells on calcium phosphate-coated discs for 6 days with (G) 50 ng/ml sRANKL or (H) 50 ng/ml TNF-α plus 10 ng/ml IL-1β in the presence of M-CSF (35,000 units/ml).

Essential role of TRAF2 in TNF-α-induced osteoclastogenesis

TRAF2 is a common adapter protein between RANK and TNF receptors, and it plays an important role in the activation of NF-κB and JNK.(10) We next examined whether RANKL- or TNF-α-induced osteoclastogenesis requires TRAF2 by using MDFLM cells from TRAF2-deficient (TRAF2−/−) and those from littermate wildtype (TRAF2+/+) mice. We treated TRAF2+/+ or TRAF2−/− MDFLM cells with 50 ng/ml sRANKL or 50 ng/ml TNF-α in the presence of M-CSF for 5 days and counted the number of TRACP+ MNCs. The result showed that sRANKL strongly induced the formation of TRACP+ MNCs in TRAF2+/+ MDFLM cultures, whereas in TRAF2−/− MDFLM, the number of TRACP+ MNCs gave a reduction of 20% compared with that of TRAF2+/+ MDFLM (Fig. 2A), showing that TRAF2 is involved in the RANKL-induced osteoclastogenesis. In TNF-α-induced osteoclastogenesis, only a few TRACP+ MNCs were observed when TRAF2−/− MDFLM cells were stimulated with TNF-α (Fig. 2B). These results show that TRAF2 was prominently active in TNF-α-induced osteoclastogenesis; however, other signal pathways besides TRAF2 must also exist for TNF-α-induced osteoclastogenesis because of the appearance of a few TRACP+ cells in the TRAF2−/− cell cultures.

Figure FIG. 2..

Effect of sRANKL or TNF-α on the formation of TRACP+ MNCs by TRAF2+/+ and TRAF2−/− MDFLM cells. In the presence of M-CSF (35,000 units/ml), TRAF2+/+ or TRAF2−/− MDFLM cells were incubated for 5 days with (A) 50 ng/ml of sRANKL or (B) 50 ng/ml of TNF-α, and TRACP staining was performed. TRACP+ MNCs with more than three nuclei were counted. (A) The number of RANKL-induced TRACP+ MNCs from TRAF2−/− was about 80% of that found for TRAF2+/+ cells. (B) TRAF2−/− MDFLM cells were scarcely induced with TNF-α to differentiate into TRACP+ MNCs, and the mean number of TRACP+ MNCs from TRAF2−/− MDFLM cells was 15.4 ± 0.07. Results are expressed as the means ± SE of three cultures. Similar results were obtained from three independent experiments. *Significantly different from wildtype: (A) p < 0.05 and (B) p < 0.01, respectively.

JNK, NF-κB, and NFATc1 activation by TNF-α in MDFLM cells

To investigate the role of TRAF2 downstream of the RANK and TNF receptors, we analyzed JNK and NF-κB activation. After stimulation with 50 ng/ml sRANKL or 50 ng/ml TNF-α, JNK and NF-κB activation in TRAF2+/+ and TRAF2−/− MDFLM cells was examined. As shown in Fig. 3A, NF-κB was activated by sRANKL or TNF-α treatment at 15 minutes in both types of cells, as analyzed by electrophoretic mobility shift assay. JNK activation was detected by measuring the phosphorylation of the c-jun fusion protein. JNK activation by sRANKL or TNF-α was detected at 15 minutes in both types of cells (Fig. 3B). However, the levels of both NF-κB and JNK activation by TNF-α were decreased at 30 minutes in TRAF2−/− MDFLM cells (Figs. 4A and 4B), and the relative ratio of JNK activation was 4.1 at 30 minutes compared with 8.5 at 15 minutes (Fig. 4B). Thus, we next studied the cellular localization of NFATc1 proteins in TRAF2−/− cells, because it was recently reported that NFATc1 is essential for RANKL-induced osteoclastogenesis.(22) We performed immunofluorescence staining after stimulation with sRANKL or TNF-α for 5 days. In sRANKL stimulation, NFATc1 proteins were localized in nuclei of MNCs generated from TRAF2−/− as well as in those of MNCs from TRAF2+/+ MDFLM cells (data not shown). On the other hand, in TNF-α-induced MNCs from TRAF2−/− MDFLM cells, NFATc1 proteins were hardly detected, although it was localized in nuclei of TNF-α-induced MNCs from TRAF2+/+ MDFLM cells (Fig. 3C). We further examined TNF-α-induced NFATc1 transcription by RT-PCR at days 2 and 4 after TNF-α treatment and found that the level of NFATc1 mRNA was increased time dependently in TRAF2+/+ cells but not in TRAF2−/− cells (Fig. 3D).

Figure FIG. 3..

JNK, NF-κB, and NFATc1 activation by sRANKL or TNF-α in normal and in TRAF2-deficient MDFLM cells. (A) TRAF2+/+ or TRAF2−/− MDFLM cells were stimulated with 50 ng/ml of sRANKL or 50 ng/ml of TNF-α for 15 or 30 minutes, and activation of NF-κB was determined by EMSA. The probe only was loaded in lane c. (B) MDFLM cells from TRAF2+/+ or TRAF2−/− mice were stimulated with 50 ng/ml of sRANKL or 50 ng/ml of TNF-α for 15 or 30 minutes, and JNK activity was examined by using a JNK assay kit (top). Equal amounts of c-jun are shown (bottom). The induction level is shown after normalization of the signal strength with the amount of c-jun (below). JNK activity was measured by NIH image. (C) In the presence of M-CSF, TRAF2+/+ or TRAF2−/− MDFLM cells were incubated with 50 ng/ml TNF-α for 5 days, and localization of NFATc1 protein was determined by immunostaining. (D) In the presence of M-CSF, TRAF2+/+ or TRAF2−/− MDFLM cells were incubated with 50 ng/ml TNF-α for 2 or 4 days, and NFATc1 mRNA was determined by RT-PCR.

Figure FIG. 4..

Induction of osteoclastogenesis by TRAF2 overexpression. MDBM cells from wildtype mice were cultured without sRANKL in the presence of M-CSF for 4 days on 48-well plates after virus supernatants of only EGFP or both TRAF2 and EGFP were infected. (A) The number of TRACP+ MNCs with more than three nuclei was counted, and (B) TRACP staining or fluorescence microscopy to detect (C) EGFP expression is shown. (D) Pit formation was assessed by culturing MDBM cells from wildtype mice on calcium phosphate-coated discs for 6 days after TRAF2 infection.

Induction of osteoclastogenesis by TRAF2 overexpression in MDBM cells

It was reported that osteoclast progenitors differentiated into TRACP+ cells by TRAF6 overexpression in vitro.(23) Thus, we examined whether the overexpression of TRAF2 also induced differentiation of osteoclast progenitors into TRACP+ cells in vitro. We used the internal ribosomal entry site (IRES) system for the expression of TRAF2 and enhanced green fluorescent protein (EGFP) and the retrovirus infection. In this experiment, we used bone marrow-derived osteoclast progenitors (MDBM cells), because fetal liver derived osteoclast progenitors lost the differentiation ability after treatment of virus transfection. MDBM cells, osteoclast progenitors from bone marrow in adult mice, were infected only EGFP or both normal TRAF2 and EGFP in the presence of M-CSF. Although no TRACP+ cells were detected when overexpressed only EGFP, TRAF2 overexpression could induce the differentiation of MDBM cells into TRACP+ cells that resorbed calcium phosphate-coated discs (Fig. 4).

DISCUSSION

The embryonic lethality of TRAF2−/− mice has prevented the analysis of TRAF2 involvement in signal pathways by use of this established in vitro system. Therefore, we developed a new in vitro stromal cell-free system that allowed us to obtain osteoclast progenitors from fetal liver of 14.5-day-old mouse embryos. For the bone resorption by MNCs formed from TNF-α-treated MDFLM cells, IL-1β was required. This requirement has been often noted in other published experiment using in vitro cultures,(6) indicating that the osteoclast progenitors from fetal liver differentiate quite normally.

Among the six known TRAFs, TRAF6 has been shown to play a crucial role in RANKL-induced osteoclastogenesis.(24,25) In this study, osteoclast progenitors from TRAF2−/− mice were severely defective in their formation of TRACP+ MNCs when treated with TNF-α. This result suggests that TRAF2-dependent signal pathways are dominant in TNF-α-induced osteoclastogenesis. We and others have shown that the TNF-R1 signaling pathway is important for TNF-α-induced osteoclastogenesis.(6,7) TRAF2 is recruited to TNF-R1 indirectly through its interaction with the TNF receptor-associated death domain (TRADD). TRADD then recruits two downstream signaling adaptor molecules—Fas-associated death domain (FADD) and receptor interacting protein (RIP). Tada et al.(26) showed that TRAF5 as well as TRAF2 could interact with RIP physically, suggesting that TRAF5 participates in the TNF-α signal transduction, which coincides with our earlier observation that TRAF5 was required for TNF-α-induced osteoclastogenesis.(12) Previously, we also showed that TRAF6 was required for TNF-α-induced osteoclastogenesis,(7) although it has not been reported that TRAF6 is a component of both TNF-R complexes. Moreover, the results obtained from TRAF2 and TRAF5 double knockout mice suggest the possibility that TRAF6 mediates TNF-α signals.(26) From these observations, we consider that either or both TRAF5 and TRAF6 collaborate with TRAF2 in TNF-α-induced osteoclastogenesis as a consequence of TNF-α-mediated JNK and NF-κB activation.

The TNF-α-induced signal transduction of MDFLM cells from TRAF2−/− mice was different from those from TRAF2+/+ ones (i.e., JNK and NF-κB activation by TNF-α did not last for 30 minutes in MDFLM cells from TRAF2−/− animals, which may cause a defect in TNF-α-induced osteoclastogenesis, although other possibilities are taken into consideration). NFATc1 was not induced in TNF-α-treated TRAF2−/− cells, and MDFLM cells from TRAF2−/− scarcely differentiated into TRACP+ cells, suggesting that TNF-α-induced osteoclastogenesis also required expression of NFATc1 proteins. It was reported that NFATc1 is essential for RANKL-induced osteoclastogenesis(22) and that both activator protein 1 (AP-1) and NF-κB binding sites are present within the promoter region of the NFATc1 gene.(27,28) Therefore, it is possible that the initial induction of the NFATc1 gene is mediated by these transcription factors. Taken together, the data indicate that the disappearance of activated NF-κB and JNK, which was observed 30 minutes after TNF-α treatment, leads to a loss of upregulation of NFATc1 transcription, because TNF-α can not constantly activate the downstream signals of TNF receptors without TRAF2, thus indicating the specific property of TRAF2 for maintaining the activation signals.

Because it was reported that TRAF6 overexpression gave rise to osteoclastogenesis,(23) we examined whether the overexpression of TRAF2 also induced differentiation of osteoclast progenitors into TRACP+ cells. When MDBM cells from wildtype mice overexpressed TRAF2, TRACP+ mononuclear and MNCs were formed. This result also suggests that TRAF2 is involved in osteoclast differentiation. The plasma membranes of various cell types contain lipid rafts that are enriched in a variety of receptors and signaling molecules, suggesting that lipid rafts are necessary for the initiation of signal events. Recently, it was shown that the TNF receptor family and the TRAF family were localized in lipid rafts(29–31) and that overexpression of TRAF2 resulted in activation of JNK and NF-κB.(32,33) In the case of TRAF2 overexpression, TRAF2 may interact with components of either or both RANK and TNF receptor complexes in lipid rafts and then initiate signals. Although it remains to be clarified whether localization of TRAF2 to lipid rafts is important for initiation of its downstream signals when TRAF2 is overexpressed, such lipid raft formation may explain why overexpression of TRAF2 results in ligand-independent osteoclastogenesis.

Our results are the first to show that TRAF2 is involved in the RANK and TNF receptors signals in osteoclast differentiation. Further experiments are necessary to show the involvement of TRAF2 in the RANKL and TNF-α signals at the molecular level. Moreover, our recently developed in vitro stromal cell-free system using osteoclast progenitors from mouse fetal liver is a good tool to elucidate whether a molecule, whose deficient mice show embryonic lethality, participates in the osteoclast differentiation signaling pathway.

Acknowledgements

This work was supported by a grant-in-aid for scientific research from the Ministry of Education, Science, Sports, and Culture of Japan.

Ancillary