12-O-tetradecanoylphorbol-13-acetate (TPA) Inhibits Osteoclastogenesis bySuppressing RANKL-Induced NF-�B Activation

CATHY WANG,1 JAMES H STEER,2 DAVID A JOYCE,2 KIRK HM YIP,1

MING H ZHENG,1 and JIAKE XU1

ABSTRACT

The mechanism by which TPA-induced PKC activity modulates osteoclastogenesis is not clear. Using aRAW264.7 cell culture system and assays for NF-�B nuclear translocation, NF-�B reporter gene activity, andMAPK assays, we demonstrated that TPA inhibits osteoclastogenesis through the suppression of RANKL-induced NF-�� activation.

Introduction: The protein kinase C (PKC) pathway has been suggested to be an important regulator of osteoclasticbone resorption. The role of PKC in RANKL-induced osteoclastogenesis, however, is not clear. In this study, weexamined the effects of 12-O-tetradecanoylphorbol-13-acetate (TPA), a PKC activator, on osteoclastogenesis andstudied its role in RANKL-induced signaling.Materials and Methods: RANKL-induced RAW264.7 cell differentiation into osteoclast-like cells was used to assessthe effect of TPA on osteoclastogenesis. Assays for NF-�B nuclear translocation, NF-�B reporter gene activity,protein kinase activity, and Western blotting were used to examine the effects of TPA on RANKL-induced NF-��,c-Jun N-terminal kinase (JNK), and MEK/ERK and p38 signal transduction pathways.Results: We found that TPA inhibited RANKL-induced RAW264.7 cell differentiation into osteoclasts in a dose-dependent manner. Time course analysis showed that the inhibitory effect of TPA on RANKL-induced osteoclas-togenesis occurs predominantly at an early stage of osteoclast differentiation. TPA alone had little effect on NF-��activation in RAW264.7 cells, but it suppresses the RANKL-induced NF-�� activation in a dose-dependent fashion.Interestingly, the suppressive effect of TPA on RANKL-induced NF-�� activation was prevented by a conventionalPKC inhibitor, Go6976. Supershift studies revealed that the RANKL-induced DNA binding of NF-�� complexesconsisted of C-Rel, NF-�B1 (p50), and RelA (p65). In addition, TPA induced the activation of JNK in RAW264.7 cellsbut had little effect on RANKL-induced activation of JNK. TPA also inhibited RANKL-induced activation of ERKbut had little effect on p38 activation.Conclusion: Given that NF-�B activation is obligatory for osteoclast differentiation, our studies imply that inhibition ofosteoclastogenesis by TPA is, at least in part, caused by the suppression of RANKL-induced activation of NF-�� duringan early stage of osteoclastogenesis. Selective modulation of RANKL signaling pathways by PKC activators may haveimportant therapeutic implications for the treatment of bone diseases associated with enhanced bone resorption.J Bone Miner Res 2003;18:2159–2168

Key words: RANKL, osteoclastogenesis, 12-O-tetradecanoylphorbol-13-acetate, protein kinase C, NF-�B

INTRODUCTION

OSTEOCLASTS, THE MULTINUCLEATED giant cells that resorbbone, are derived from the monocyte-macrophage lin-

eage cells.(1) The protein kinase C (PKC) pathway has been

suggested to be an important regulator of osteoclastic boneresorption.(2) 12-O-tetradecanoyl phorbol-13-acetate (TPA),an activator of PKC, decreased osteoclastic bone-resorbingactivity.(3) A synthetic peptide fragment, PKC(530-558),which activates PKC, significantly inhibits osteoclasticbone resorption, and this effect is abolished in the presenceof a synthetic peptide fragment PKC(19-36), which inhibitsThe authors have no conflict of interest.

1Department of Orthopaedics, University of Western Australia, Nedlands, Western Australia, Australia.2Department of Pharmacology, University of Western Australia, Nedlands, Western Australia, Australia.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 18, Number 12, 2003© 2003 American Society for Bone and Mineral Research

2159

PKC, suggesting an important physiological role for thePKC pathway in osteoclast function.(4) In addition, syn-thetic diacylglycerols, analogs of the physiological activa-tors of PKC, also inhibited bone resorption, providing fur-ther evidence for an important role of the PKC pathway inthe regulation of osteoclast activity.(5) However, to date, themechanisms of PKC activity in the signaling pathwaysduring osteoclastogenesis are not well established.

The receptor activator of NF-kB ligand (RANKL/OPGL/ODF/TRANCE), a member of the tumor necrosis factor(TNF) receptor-ligand family, is directly involved in thedifferentiation of monocyte-macrophages into oste-oclasts.(1,6,7) Mice with a disrupted RANKL gene show alack of osteoclasts, severe osteopetrosis, and a defect intooth eruption, indicating that RANKL is essential for os-teoclast differentiation.(8) In addition, RANKL was able toinduce mature osteoclast activation in vitro(9) and osteoclastpolarization and hypercalcemia in vivo.(9,10) RANKL-induced osteoclastogenesis is mediated through its receptor,RANK. The interaction of RANKL with RANK leads to therecruitment of members of the TNF receptor associatedfactor (TRAF) adapter proteins and the activation of theNF-�� and c-Jun N-terminal kinase (JNK) signalingpathways.(11–14) In addition, MEK/ERK and p38 MAP ki-nase pathways have been shown to play an important role inRANKL-induced osteoclastogenesis.(15,16) Selective modu-lation of RANKL signaling pathways may have importanttherapeutic implications for the treatment of bone diseasesassociated with enhanced bone resorption, such as osteopo-rosis, osteoarthritis, and cancer-induced bone loss.

In this study, we examined the effects of TPA onRANKL-induced osteoclastogenesis and potential mecha-nisms by which TPA interferes with RANKL-activated sig-naling pathways. We showed that TPA inhibits RANKL-induced RAW cell differentiation into osteoclasts, as well asthe expression of osteoclast specific genes. We demon-strated that TPA suppresses the RANKL-induced NF-�Bactivation and nuclear translocation of DNA binding-competent NF-�B complexes but had little effect onRANKL-induced activation of JNK. TPA also preventedRANKL-induced activation of ERK but had little effect onp38 activation. Given NF-�B’s essential role in osteoclastdifferentiation,(17–19) our studies imply that inhibition ofosteoclastogenesis by TPA is, at least in part, caused by thesuppression of RANKL-induced activation of NF-�B.

MATERIALS AND METHODS

RAW264.7 cells were obtained from the American TypeCulture Collection (Rockville, MD, USA). PKC inhibitorsGo6976, GF109203X, and rottlerin were purchased fromCalbiochem (Alexandria, Australia). [�32P]ATP and en-hanced chemiluminescence (ECL) system were from Am-ersham Pharmacia Biotech. cDNA-Jun-(1-135) was ex-pressed as glutathione S-transferase (GST) fusion proteinsand purified by glutathione-Sepharose chromatography foruse as substrates for JNK assays. Recombinant rat GST-rRANKL was expressed and purified as previously de-scribed.(10) TPA, LPS, recombinant TNF-�, interleukin(IL)-1�, and Goat anti-mouse IgG-conjugated horseradish

peroxidase were obtained from Sigma (Sydney, Australia).Polyclonal antibodies to NF-�B1 (p50), Rel-A (p65), c-Rel,and B-Rel, and antibodies to phospho-specific p44/42 ERKand p38 were obtained from Santa Cruz Biotechnology(Santa Cruz, CA, USA).

RAW cells culture and in vitro osteoclastogenesisassay

RAW264.7 cells were cultured in �-modified essentialmedium (�-MEM) (Biosciences Pty Ltd.) supplementedwith 10% fetal calf serum (FCS), 2 mM L-glutamine, and100 U/ml penicillin/streptomycin. For osteoclast culture,RAW cells were seeded in a 6-well plate to a density of 5 �104 cells/well and cultured for 5–7 days in full � -MEM inthe presence of 100 ng/ml of GST-rRANKL.(10) To examinethe effect of TPA (Sigma, Sydney, Australia) on osteoclas-togenesis, RAW264.7 cells were plated in a 96-well plate(1 � 103 cells/well) and cultured in complete medium in thepresence of GST-rRANKL, TPA, or RANKL plus TPA.Cultures were fed every 2–3 days by replacing with freshmedium. After 5–7 days, cultures were fixed for 10 minutesat room temperature with 4% paraformaldehyde in PBS andwashed four times with PBS. The fixed cells were stainedfor TRACP using the Diagnostic Acid Phosphatase kit(Sigma) according to the manufacturer’s instructions, andTRACP activity was calculated as arbitrary densitometryunits. For the bone resorption assay, RAW264.7 cells werecultured on dentine slices in a 96-well plate and fixed. Boneresorption pits were viewed using a scanning electron mi-croscope.

RNA isolation and reverse transcriptase-polymerasechain reaction

Total RNA was isolated from RAW cells treated withGST-rRANKL according to the manufacturer’s instructions(Qiagen, Victoria, Australia). For reverse transcriptase-polymerase chain reaction (RT-PCR), single-strandedcDNA was prepared from 2 �g of total RNA using RT withan oligo-dT primer. Two microliters of each cDNA wassubjected to PCR amplification using specific primers. Asan internal control, the single-stranded cDNA was PCR-amplified for 25 cycles using 36 B4 forward primer: 5�-TCATTGTGGGAGCAGACA-3�; and reverse primer: 5�-TCCTCCGACTCTTCCTTT-3�. For the amplification ofmouse calcitonin receptor, forward primer: 5�-TGGTTGA-GGTTGTGCCCA-3�; and reverse primer: 5�-CTCGTG-GGTTTGCCTCATC-3� were used, and PCR-amplificationwas carried out for 35 cycles (94°C, 40 s; 62°C, 40 s; and72°C, 40 s). For the amplification of cathepsin K, forwardprimer: 5�-GGGAGAAAAACCTGAAGC-3�; and reverseprimer: 5�-ATTCTGGGGACTCAGAGC-3� were used.PCR amplification was carried out for 30 cycles (94°C, 40 s;55°C, 40 s; and 72°C 40 s).

Transient and stable transfection of luciferase reportergene

To examine the NF-�B activation in RAW cells,RAW264.7 cells were transfected with a luciferase reportergene. The 3kB-Luc-SV40 reporter, which contains three

2160 WANG ET AL.

NF-�B sites from the interferon gene upstream of the lucif-erase coding region, has been described previously.(20,21)

RAW264.7 cells (5 � 106) were transfected with 0.5 �g ofplasmid DNA using a DEAE-dextran method (AmershamPharmacia Biotech) as previously described.(21) Cells wereplated out in 24-well plates at a density of 1 � 105 cells/welland treated with RANKL, TPA, or RANKL plus TPA, andsubsequently, transfected cells were harvested for assay ofluciferase activity. For stable transfection, the 3kB-Luc-SV40 reporter construct (20 �g) and pcDNA3.1 (2 �g)vectors were transfected into RAW264.7 cells using electro-poration with the following conditions (280 V and 960 �F).The transfected cells were selected with 400 �g/ml of G418(Gibco BRL, Life Technologies, Melbourne Australia). Theresulting stable cell line was used to investigate NF-�Bactivation by RANKL and TPA. Firefly luciferase expres-sion was measured using the Promega Luciferase AssaySystem according the manufacturer’s instructions (Pro-mega, Sydney, Australia).

EMSA

RAW cells were treated with RANKL, TPA, or RANKLplus TPA for various periods. Unexposed cells served ascontrols. Nuclear extracts were prepared from 1 � 107 cellsas previously described.(22,23) EMSA was carried out asdescribed previously.(21) Briefly, nuclear proteins (4 �g)were preincubated for 10 minutes at room temperature with0.5 �g of poly(dI-dC) (Amersham Pharmacia Biotech) in abinding buffer (4% Ficoll, 20 mM HEPES [pH 7.9], 1 mMEDTA, 1 mM dithiothreitol, 50 mM KCl, and 0.05% IGE-PAL CA-630) to give a final reaction volume of 10 �l. Adouble-strand oligonucleotide probe NF-�B with overhang-ing 5�G, 5�-gggcatgggaatttccaactc-3�, which had been fill-inlabeled with [�-32P]dCTP (Amersham Pharmacia Biotech)using Klenow fragment of E. coli DNA polymerase I (Pro-mega), was added. In competition experiments, unlabeledoligonucleotide probes (incorporating either a consensusNF-�B binding sequence or a nonbinding mutant sequence)were added at 100-fold molar excess. After 10 minutes ofincubation, samples were loaded onto a 4% polyacrylamidegel containing 0.25 � Tris-Borate-EDTA buffer, which hadbeen pre-run for 2 h in the same buffer. For supershiftexperiments, polyclonal antibodies (1 �l per biding reac-tion) to NF-�B1 (p50), Rel-A (p65), c-Rel, and B-Rel(Santa Cruz Biotechnology) were added. Protein/DNAcomplexes were separated at 150 V for 90 minutes. Gelswere exposed to Kodak X-ray film using a single intensi-fying screen.

JNK activity assay

The in vitro JNK activity assay was performed as previ-ously described.(24) In brief, RAW cells were treated withRANKL, TPA, or RANKL plus TPA for various times. Thetreated cells were lysed in ice-cold Buffer A (20 mMHEPES, 2.5 mM MgCl2, 0.1 mM EDTA, 20 mM�-glycerophosphate, and 100 mM NaCl, pH 7.7, supple-mented with 0.05% [vol/vol] Triton X-100, 1% [vol/vol]NP-40, 500 �M dithiothreitol, 100 �M sodium orthovana-date, 20 �g/ml leupeptin, 100 �g/ml phenylmethylsulfonylfluoride, and 20 �g/ml aprotinin). JNK activity was mea-

sured by incubation in Buffer A supplemented with 20 �MATP, 1 �Ci of [�-32P] ATP, and a protein substrate (5 �gof GST-c-Jun-1–135) from Dr Marie A Bogoyevitch (CellSignaling Laboratory, Department of Biochemistry and Mo-lecular Biology, School of Biomedical and Chemical Sci-ences, the University of Western Australia). The reactionwas performed for 30 minutes at 30°C, and the phosphor-ylated substrate was separated by SDS-PAGE and visual-ized by autoradiography.

Western blotting analysis of active ERK and p38

Proteins were separated by SDS-PAGE and electroblottedonto Hybond-P (PVDF) membranes (Bio-Rad). Membraneswere blocked with 5% (wt/vol) nonfat milk powder inTBST (10 mM Tris, pH7.5, 150 mM NaCl, 0.1% [vol/vol]Tween 20) and probed with primary antibodies in the block-ing solution. The membranes were washed three times withTBS. Horseradish peroxidase–conjugated secondary anti-bodies were diluted 1/5000 in 1% (wt/vol) nonfat milkpowder in TBST. The membranes were developed using theECL system (Amersham Pharmacia Biotech).

RESULTS

TPA inhibits differentiation of RAW264.7 cells intoosteoclast-like cells

RANKL is a key cytokine for the induction of osteoclas-togenesis. The differentiation of bone marrow cells orspleen cells into osteoclasts requires macrophage-colonystimulating factor (M-CSF) and RANKL, whereas RAWcells require only RANKL to form bona fide bone-resorbingosteoclasts.(10,12) Therefore, RAW cells were used to studythe direct effect of TPA on RANKL-induced osteoclastformation and the relevant signaling pathways. RAW cellswere treated for 5 days with different concentrations of TPAin the presence or absence of RANKL. When RAW cellswere treated with RANKL alone, they formed osteoclast-like cells that display TRACP activity, express calcitoninreceptor mRNA, and were capable of bone resorption (datanot shown). In contrast, when RAW cells were treated withTPA alone, neither TRACP activity nor calcitonin receptormRNA or bone resorption was evident. Although RANKLis capable of inducing osteoclastogenesis in the presence ofTPA, TPA decreases RANKL-induced osteoclastogenesisas demonstrated by TRACP activity in the culture. Theeffect of TPA on RANKL-induced osteoclastogenesis wasin a dose-dependent manner (Fig. 1). Interestingly, it seemsthat the TRACP� mutinucleated osteoclast-like cells in-duced by RANKL in the presence of TPA (Figs. 1C and 1D)exhibited smaller sizes compared with those in the absenceof TPA (Fig. 1E). In short, these results indicate that TPAinhibits RANKL-induced osteoclastogenesis.

Next, we examined whether the inhibitory effect by TPAon RANKL-induced osteoclastogenesis occurs at the earlystage of osteoclast differentiation. RAW cells were pre-treated with TPA for 24 h before the addition of RANKL ortreated with TPA at the early time (at days 1–2) or late timecourse (at days 3–4) in the presence of RANKL. The degreeof osteoclast formation was examined by TRACP stainingat the end of a 5-day culture. The results showed that TPA,

2161INHIBITION OF OSTEOCLASTOGENESIS AND NF-�B BY TPA

at a dose of 10 ng/ml, inhibits osteoclastogenesis induced byRANKL in the pretreatment and early and late treatmentgroups. The greatest reduction was evidenced in the earlytreatment group (Fig. 2). At lower doses of 1 ng/ml, TPAinhibited osteoclastogenesis in the early treatment group buthad minimum effects on the pretreatment and late groups.Thus, these data demonstrated that the inhibitory effect ofTPA occurred predominantly but not exclusively during theearly stage of RANKL-induced osteoclastogenesis.

To further investigate the inhibitory effects of TPA onRANKL-induced osteoclastognesis, RAW cells were treatedwith RANKL in the absence or presence of TPA at varyingdoses, and total RNA was isolated. Semiquantitative RT-PCRwas performed using primers for calcitonin receptors and ca-

thepsin K, markers of osteoclast differentiation. The resultsshowed that TPA reduces gene expression of calcitonin recep-tors and cathepsin K, in a dose-dependent fashion, duringRANKL-induced osteoclastogenesis (Fig. 3).

TPA suppresses the RANKL-induced activation of NF-�B

RANKL-induced osteoclast differentiation involves acti-vation of two major signaling pathways, the JNK andNF-�B pathways.(11–14) Activation of PKC by TPA has alsobeen shown to involve the JNK and NF-kB pathways inseveral cell types.(25–30) To examine the effect of TPA onthe RANKL-induced activation of NF-�B, RAW cells weretransiently transfected with an NF-�B–driven luciferasereporter gene construct, 3�B-Luc-SV40, and then treated

FIG. 1. TPA inhibits RANKL-induced osteoclastogenesis .RAW264.7 cells were cultured inthe presence of RANKL, TPA, orRANKL plus TPA. After 5 days,the cells were fixed with 4% para-formaldehyde and stained forTRACP activity. (A) Light micros-copy images showing the effect ofTPA on RANKL-induced oste-oclast formation with morphologi-cal changes. RAW cells were (A)left untreated or (B) treated with 10ng/ml of TPA, (C) 10 ng/ml ofTPA plus RANKL, (D) 1 ng/ml ofTPA plus RANKL, or (E) RANKLalone. (F) A graph representing to-tal TRACP activity calculated asarbitrary densitometry units.

FIG. 2. TPA inhibits RANKL-induced oste-oclastogenesis predominantly during the earlystage. RAW cells were pretreated with TPA for24 h before the addition of RANKL, treated withTPA at the early time (at days 1–2) or late timecourse (at days 3–4), or left untreated and cul-tured in the presence of RANKL for a total of 5days. The treated cells were fixed with 4% para-formaldehyde and stained for TRACP activity.TRACP� multinucleated (�5 nuclei) cells werecounted. Values are expressed as mean � SEfrom quadriplicate wells (*p � 0.001).

2162 WANG ET AL.

with RANKL, TPA, or RANKL plus TPA. The treated cellswere harvested at various time-points, and luciferase activ-ities were measured. RANKL alone induced NF-�B activa-tion (21.5 � 0.55-fold at 6 h, Fig. 4A). In comparison,treatment of TPA alone resulted in a low level of NF-�Bactivation in RAW cells (1.86 � 0.12-fold after 6 h),whereas TPA decreased the RANKL-induced NF-�B acti-vation at all time-points tested.

To examine the dose dependence of the inhibitory effectof TPA on NF-�B activation by RANKL, a stable RAW cellline transfected with a 3�B-Luc-SV40 expression constructwas established. The transfected RAW cells were treatedwith RANKL in the presence of varying doses of TPA (1,10, and 50 ng/ml). TPA was shown to suppress the NF-�Bactivation by RANKL in a dose-dependent fashion (Fig.4B). To examine if TPA also inhibited the activation ofNF-�B induced by stimuli other than RANKL in RAWcells, we tested the effect of TPA on LPS, TNF-�, andIL-1�–induced luciferase activity in RAW cells stably ex-pressing the 3kB-Luc-SV40 reporter gene. The results dem-onstrated that TPA suppressed the induction of NF-�B–

mediated transcription by LPS, TNF-�, and IL-1� atdifferent degrees compared with the control (Fig. 4C).

To further verify the effect of TPA on RANKL-inducedNF-�B activation, nuclear translocation of DNA bindingcapability of NF-�B was examined by EMSA as describedin the Materials and Methods section. RAW cells weretreated with RANKL, TPA, or RANKL plus TPA for var-ious time periods. Nuclear extracts were prepared and in-cubated with [�-32P]-labeled oligonucleotide containing aconsensus NF-�B binding site. As shown in Fig. 5A, TPAdecreased the abundance of DNA-binding NF-�B com-plexes in nuclei of RANKL-stimulated RAW cells. Thisimplies that TPA inhibits at a step before DNA binding ofNF-�B complexes.

To determine the subunits present in the RANKL-activated NF-�B complexes, an EMSA was performed us-ing anti-NF-�B1 (p50), Rel-A (p65), C-Rel, and B-Relantibodies. In the presence of anti-NF-�B1 (p50), Rel-A(p65), and C-Rel, supershift bands of the specific DNA-bound proteins were detected (Fig. 5B). In contrast, nosupershift signal was detected when B-Rel antibody wasadded. These results suggest that RANKL-induced DNAbinding of NF-�B complexes in RAW cells consists ofNF-�B1 (p50), Rel-A (p65), and c-Rel, but not B-Rel.

To further confirm that the suppression of RANKL-induced NF-kB activity induced by TPA was caused by theactivation of a PKC isoform, we tested the effect of the PKCinhibitors, Go9676, GF109203X, and rottlerin. As shown inFig. 6, Go9676 (5 �M), an inhibitor of conventional PKCisoforms, prevented the suppressive effect of TPA onRANKL-induced NF-�B transcriptional activation. On theother hand, the PKC �-specific inhibitor, rottlerin (5 �M),had no effect on NF-�B–mediated transcriptional activity,whereas the broad spectrum PKC inhibitor GF109203X (5�M) partially prevented TPA’s suppression of RANKL-induced NF-�B transcriptional activity (data not shown).These results indicate that TPA-induced activation of con-ventional PKC isoforms is involved in the inhibition ofRANKL-induced NF-�B activation.

TPA has no effect on RANKL-induced activation ofJNK

The activation of JNK is one of the major signalingevents induced by interaction of RANKL with RANK.(12)

To examine the effect of TPA on JNK activation induced byRANKL, RAW cells were treated with RANKL, TPA, orRANKL plus TPA for 0–30 minutes. The c-Jun kinaseassay was performed using recombinant GST-c-Jun proteinand [�-32P]ATP. Phosphorylated GST-c-Jun proteins wereseparated by SDS-PAGE and visualized by autoradiogra-phy. The result showed that RANKL and TPA each inducedthe activation of JNK. Unlike the situation with the activa-tion of the NF-�B pathway, TPA had no or little effect onRANKL-induced activation of JNK (Fig. 7A).

TPA suppresses RANKL-induced activation of the ERKbut not the p38 pathway

The MEK/ERK and p38 MAP kinase pathways have beenshown to play an important role in RANKL-induced oste-

FIG. 3. TPA dose-dependently reduced RANKL-induced expression ofosteoclast genes. 1.5 � 105 RAW246.7 cells seeded into 6-well plates wereincubated in the presence or absence of 100 ng/ml of RANKL for 5 dayswith various doses of TPA. Total cellular RNA was extracted and cDNAwas synthesized using 2 �g of total RNA with oligo-dT. PCR amplifica-tion was performed using specific primers for calcitonin receptor, cathep-sin K, and 36B4 genes. (A) PCR products were separated on 1.2% ofagarose gels and measured by densitometry. The levels of (B) calcitoninreceptor (CR) and (C) cathepsin K (CaK) mRNA are shown as the ratio ofcalcitonin receptor or cathepsin K to 36B4.

2163INHIBITION OF OSTEOCLASTOGENESIS AND NF-�B BY TPA

oclastogenesis.(15,16) To examine the effect of TPA onRANKL-induced activation of ERK and p38, Western blot-ting analysis was performed using antibodies to the phos-phorylated forms of ERK and p38. RANKL and TPA eachinduced the phosphorylation of ERK and p38. However,whereas TPA prevented RANKL-induced activation ofERK, it had little or no effect on RANKL-induced activa-tion of p38 (Fig. 7B).

DISCUSSION

Many cytokines such as IL-1, IL-6, and TNF-� havestimulatory effects on the formation of osteoclasts in vivo orin co-culture systems, but they are not capable of directlyinducing the formation of osteoclasts from their precursors.Knockout mice lacking IL-1, IL-6, or TNF-� or their re-ceptors do not exhibit osteopetrosis,(31–33) further suggest-ing that these cytokines enhance osteoclast formation indi-rectly either by increasing the production of M-CSF(34) orby potentiating the osteoclastogenic effects of RANKL.(35)

In contrast, RANKL is a key inducer of osteoclastogen-esis,(6,8) and intracellular mechanisms that inhibit RANKL-induced osteoclastogenesis are therefore of major interestbecause they offer possible means of managing bone dis-eases associated with enhanced osteoclastic bone resorp-tion.

PKC is a serine/threonine protein kinase involved inmany cellular responses. Although the analysis of PKCactivity has provided crucial insights to its biological func-tion in many systems,(36) its role in the differentiation ofosteoclast progenitor cells into osteoclasts remains unclear.Previous studies have shown that calcitonin or TPA, which

FIG. 5. TPA inhibits RANKL-induced nuclear translocation ofNF-�B in RAW cells. (A) RAW cells were treated with medium alone(none), LPS (1 �g/ml, positive control), RANKL (200 ng/ml), TPA (20ng/ml), or RANKL and TPA for 0, 2, or 18 h. Nuclear extracts wereprepared from 1 � 107 cells, and EMSA was carried out using [32P]labeled NF-�B oligonucleotide probe. (B) Supershift assays usinganti-NF-�B1 (p50), Rel-A (p65), C-Rel, and B-Rel antibodies deter-mined that RANKL-induced DNA binding of NF-�B complexes inRAW cells consists of NF-�B1 (p50), Rel-A (p65), and C-Rel, but notB-Rel.

FIG. 4. TPA inhibits RANKL, LPS, IL-1B,and TNF-induced NF-�B–dependent transcrip-tion in RAW cells. (A) RAW cells, transientlytransfected with the 3kB-Luc-SV40 reportergene, were treated with medium alone (vehicle),RANKL (200 ng/ml), TPA (20 ng/ml), orRANKL and TPA, and the luciferase activity inlysates was determined at various times up to72 h. Each point is the mean � SE from tripli-cate wells. (B) TPA suppresses RANKL-induced activation of NF-�B in a dose-dependent fashion. RAW cells, stablytransfected with the 3kB-Luc-SV40 reportergene, were treated with various doses of TPA (0,1, 10, and 50 ng/ml) in the presence or absenceof RANKL (200 ng/ml). The luciferase activitieswere measured 24 h after treatment. Each pointis the mean � SE from triplicate wells. (C)RAW cells stably transfected with the 3�B-Luc-SV40 reporter gene were plated into a 24-wellculture dish 18 h before stimulation for 8 h withTNF � (10 ng/ml), LPS (1 �g/ml), or IL-1� (10ng/ml). Each bar represents the luciferase activ-ity of cells stimulated in the presence or absenceof TPA (20 ng/ml; mean � SE from triplicatewells).

2164 WANG ET AL.

increases PKC activity, inhibits bone pit formation invitro.(2,37,38) The effect of PKC activation on osteoclasto-genesis, however, remains controversial. Here we have usedRAW264.7 cells culture for the study of TPA inhibition onRANKL-induced osteoclastogenesis and signaling path-ways. The ability of the soluble form of RANKL alone toinduce differentiation of RAW cells into osteoclasts in theabsence of supportive cells allowed us to test whether TPAacts directly on RANKL-induced osteoclastogenesis in theabsence of stromal cells. Our studies showed that TPA actsdirectly on osteoclast precursor cells and prevents the dif-ferentiation of osteoclast precursors into osteoclasts. Itseems that the suppressive effect of TPA takes place pre-dominantly during the early stage of RANKL-induced os-teoclastogenesis. Apart from a direct effect of TPA onosteoclast precursors demonstrated in this study, there arealso reports of the indirect effects of PKC activators throughstromal cells to modulate osteoclastogenesis by stimulatingthe expression of RANKL and OPG mRNAs in primaryosteoblasts.(39,40)

The interaction of RANKL with RANK results in acascade of intracellular events, including the activation ofNF-�� and the protein kinase JNK.(11,12,41) NF-�� signal-ing has been shown to play an important role in osteoclas-togenesis.(17) NF-�B p50�/� and p52�/� double knockoutmice exhibit severe osteopetrosis caused by failure of oste-oclast formation.(18,19) NF-�B is activated by RANKL bothin RAW264.7 cells and in monocytes(6,12,42,43) and is requiredin vivo for osteoclast formation.(18) More recent studiesindicate that p50 and p52 expression are essential forRANK-expressing osteoclast precursors to differentiate intoTRACP� osteoclasts in response to RANKL and otherosteoclastogenic cytokines.(44) Therefore, suppression ofNF-�B activation would play an important role in osteoclastformation. Interestingly, in the present studies, TPA wasshown to reduce RANKL-induced NF-�B activity simulta-neously with inhibition of RANKL-induced differentiationof RAW264.7 cells into osteoclasts. TPA alone has little effect

on the activation of NF-�B in RAW264.7 cells, consistent withprevious findings,(45) but in contrast with enhanced NF-�Bactivity in cell types such as JB6 mouse epidermal cell line(29)

and the human T-cell leukemia Jurkat cells.(30) In addition,TPA has inhibitory effects on LPS-, TNF-�–, and IL-1�–induced activation of NF-�B–mediated transcription but atdifferent degrees compared with the control.

RAW264.7 cells express eight PKC isoforms from threegroups: conventional PKCs (�, �I, �II), novel PKCs (�, �,�), and atypical PKCs (� and ).(46,47) TPA induces theactivation of the �, �I, �II, �, �, and � isoforms in RAWcells, whereas pretreatment for 2–24 h could result in down-regulation of PKC �, �I, �II, and � expression.(46) Becausea 24-h preincubation with TPA only partially preventsRANKL-induced osteoclastogenesis (Fig. 2), it is not pos-sible to conclude whether it is the initial activation orsubsequent downregulation of PKC isoforms that is respon-sible for TPA’s inhibition of RANKL-induced osteoclasto-genesis. Interestingly, we have shown that a conventionalPKC inhibitor, Go6976, prevents the inhibitory effect ofTPA on RANKL-induced activation of NF-�B in RAWcells, indicating that the activation of one or more conven-tional PKC isoforms is likely to be involved in NF-�Bsuppression.

cDNA-Jun phosphorylation also plays an essential role inosteoclastogenesis.(48) The signaling pathways mediated byRANK/TRAF6/JNK/AP-1/C-Fos seem to be critical for os-teoclastogenesis, because TRAF6-(49,50) and c-fos–deficientmice(51,52) develop osteopetrosis. Using mice lacking JNK1or JNK2, it has been demonstrated that JNK1, but not JNK2,is specifically activated by the osteoclast-differentiating fac-tor RANKL, and activation of JNK1 is required for efficientosteoclastogenesis from bone marrow–derived mono-cytes.(48) Furthermore, estrogen-mediated suppression ofJNK1 and transcription factors c-Jun and c-Fos were alsofound to decrease the ability of RANKL to induce oste-oclastogenesis.(53,54) Activation of JNK is likely to play animportant role early in the process of osteoclast differenti-ation. We have found that TPA induces the activation ofJNK in RAW cells, which is in keeping with findings inother cell types.(25–28) However, TPA does not lead to theinhibition of RANKL-induced JNK activity in RAW cells,suggesting that inhibition of osteoclastogenesis might beindependent of TPA’s action on c-Jun signaling. Alterna-tively, in addition to activating the JNK pathway, TPAmight lead to activation of factors that inhibit RANKL-induced c-Jun transcriptional activity. TPA is able to acti-vate both classical (�, �, �) and novel (�, �, ) PKCisozymes in several cell types,(36) but predominantly theclassical PKC isozymes in RAW cells.(46) Further investi-gation of the PKC isozymes that regulate RANKL-inducedosteoclast formation might provide insight into the cross-interaction of these signaling pathways.

The MEK/ERK specific inhibitors, U0126 and PD98059,have been shown to accelerate the differentiation ofRAW264.7 cells into osteoclast-like cells, indicating that theMEK/ERK pathway(15,16) negatively regulates osteoclasto-genesis, whereas the p38 pathway positively regulates os-teoclastogenesis. We have demonstrated that TPA preventsRANKL-induced activation of ERK and has no effect on

FIG. 6. Go9676 blocks the inhibitory effect of TPA on RANKL-induced NF-�B–dependent transcription in RAW cells. RAW cellsstably transfected with the 3�B-Luc-SV40 reporter gene were platedinto a 24-well culture dish 18 h before stimulation for 8 h with vehicle(bar 1), 2-TPA (10 ng; bar 2), vehicle � G09676 (5 �M; bar 3),RANKL4 (100 ng; bar 4), RANKL (100 ng) � TPA (10 ng; bar 5), andRANKL (100 ng) � TPA (10 ng) � G09676 (5 �M; bar 6). Each baris the mean � SE from triplicate wells.

2165INHIBITION OF OSTEOCLASTOGENESIS AND NF-�B BY TPA

p38 activation, outcomes that might be expected to confer apositive effect on osteoclastogenesis. Our finding that TPAhas an inhibitory effect on RANKL-induced osteoclastogen-esis suggests that this effect of TPA is unlikely to be causedby effects on the MEK/ERK and p38 pathways but ratherbecause of its effect on NF-�B activation.

In summary, our findings demonstrated that TPA sup-presses RANKL-induced RAW264.7 cell differentiation intoosteoclasts and the expression of osteoclast-specific genes.It seems that the inhibitory effect takes place predominantlyduring the early stage of osteoclast differentiation inducedby RANKL. Suppression of NF-�B activation accounts, atleast in part, for the mechanism of action.

ACKNOWLEDGEMENTS

This work was supported in part by the National Healthand Medical Research Council of Australia. The authorsthank John M Papadimitriou for reading the manuscript.

REFERENCES

1. Teitelbaum SL 2000 Bone resorption by osteoclasts. Science 289:1504–1508.

2. Su Y, Chakraborty M, Nathanson MH, Baron R 1992 Differentialeffects of the 3�, 5�-cyclic adenosine monophosphate and proteinkinase C pathways on the response of isolated rat osteoclasts tocalcitonin. Endocrinology 131:1497–1502.

3. Mano M, Arakawa T, Mano H, Nakagawa M, Kaneda T, KanekoH, Yamada T, Miyata K, Kiyomura H, Kumegawa M, Hakeda Y2000 Prostaglandin E2 directly inhibits bone-resorbing activity ofisolated mature osteoclasts mainly through the EP4 receptor. CalcifTissue Int 67:85–92.

4. Moonga BS, Dempster DW 1998 Effects of peptide fragments ofprotein kinase C on isolated rat osteoclasts. Exp Physiol 83:717–725.

5. Moonga BS, Stein LS, Kilb JM, Dempster DW 1996 Effect ofdiacylglycerols on osteoclastic bone resorption. Calcif Tissue Int59:105–108.

6. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T,Elliott R, Colombero A, Elliott G, Scully S, Hsu H, Sullivan J,Hawkins N, Davy E, Capparelli C, Eli A, Qian YX, Kaufman S,Sarosi I, Shalhoub V, Senaldi G, Guo J, Delaney J, Boyle WJ 1998Osteoprotegerin ligand is a cytokine that regulates osteoclast dif-ferentiation and activation. Cell 93:165–176.

7. 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 forosteoprotegerin/osteoclastogenesis-inhibitory factor and isidentical to TRANCE/RANKL. Proc Natl Acad Sci USA 95:3597–3602.

FIG. 7. TPA has little effect on RANKL-induced phosphorylation of JNK and p-38, butinhibits phosphorylation of ERK. RAW cellswere treated with vehicle alone (mock), RANKL(200 ng/ml), TPA (20 ng/ml), or RANKL andTPA for 0, 15, and 30 minutes, and whole cellextracts were analyzed for (A) JNK activity byautoradiography of [32P] phosphorylated GST-c-Jun(1-135) separated on SDS-PAGE gels. (B)Phosphorylated forms of ERK and p38 by West-ern blot analysis. Whole cell extracts were pre-pared after 0, 15, or 30 minutes of stimulation,and proteins were separated by SDS-PAGE andtransferred to PVDF membranes. Membraneswere probed sequentially with antibodies top-ERK, p-p38, and Tubulin (in that order), andbands were detected by ECL.

2166 WANG ET AL.

8. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C,Morony S, Oliveira-dos-Santos AJ, Van G, Itie A, Khoo W,Wakeham A, Dunstan CR, Lacey DL, Mak TW, Boyle WJ, Pen-ninger JM 1999 OPGL is a key regulator of osteoclastogenesis,lymphocyte development and lymph-node organogenesis. Nature397:315–323.

9. Burgess TL, Qian Y, Kaufman S, Ring BD, Van G, Capparelli C,Kelley M, Hsu H, Boyle WJ, Dunstan CR, Hu S, Lacey DL 1999The ligand for osteoprotegerin (OPGL) directly activates matureosteoclasts. J Cell Biol 145:527–538.

10. Xu J, Tan JW, Huang L, Gao XH, Laird R, Liu D, Wysocki S,Zheng MH 2000 Cloning, sequencing, and functional character-ization of the rat homologue of receptor activator of NF-kappaBligand. J Bone Miner Res 15:2178–2186.

11. Zhang YH, Heulsmann A, Tondravi MM, Mukherjee A, Abu-Amer Y 2001 Tumor necrosis factor-alpha (TNF) stimulatesRANKL-induced osteoclastogenesis via coupling of TNF type 1receptor and RANK signaling pathways. J Biol Chem 276:563–568.

12. Hsu H, Lacey DL, Dunstan CR, Solovyev I, Colombero A, TimmsE, Tan HL, Elliott G, Kelley MJ, Sarosi I, Wang L, Xia XZ, ElliottR, Chiu L, Black T, Scully S, Capparelli C, Morony S, ShimamotoG, Bass MB, Boyle WJ 1999 Tumor necrosis factor receptorfamily member RANK mediates osteoclast differentiation andactivation induced by osteoprotegerin ligand. Proc Natl Acad SciUSA 96:3540–3545.

13. Galibert L, Tometsko ME, Anderson DM, Cosman D, Dougall WC1998 The involvement of multiple tumor necrosis factor receptor(TNFR)- associated factors in the signaling mechanisms of recep-tor activator of NF-kappaB, a member of the TNFR superfamily.J Biol Chem 273:34120–34127.

14. Darnay BG, Haridas V, Ni J, Moore PA, Aggarwal BB 1998Characterization of the intracellular domain of receptor activator ofNF-kappaB (RANK). Interaction with tumor necrosis factorreceptor- associated factors and activation of NF-kappab and c-JunN-terminal kinase. J Biol Chem 273:20551–20555.

15. Hotokezaka H, Sakai E, Kanaoka K, Saito K, Matsuo K, Kitaura H,Yoshida N, Nakayama K 2002 U0126 and PD98059, specificinhibitors of MEK, accelerate differentiation of RAW264.7 cellsinto osteoclast-like cells. J Biol Chem 277:47366–47372.

16. Matsumoto M, Sudo T, Saito T, Osada H, Tsujimoto M 2000Involvement of p38 mitogen-activated protein kinase signalingpathway in osteoclastogenesis mediated by receptor activator ofNF-kappa B ligand (RANKL). J Biol Chem 275:31155–31161.

17. Boyce BF, Xing L, Franzoso G, Siebenlist U 1999 Required andnonessential functions of nuclear factor-kappa B in bone cells.Bone 25:137–139.

18. Franzoso G, Carlson L, Xing L, Poljak L, Shores EW, Brown KD,Leonardi A, Tran T, Boyce BF, Siebenlist U 1997 Requirement forNF-kappaB in osteoclast and B-cell development. Genes Dev11:3482–3496.

19. Iotsova V, Caamano J, Loy J, Yang Y, Lewin A, Bravo R 1997Osteopetrosis in mice lacking NF-kappaB1 and NF-kappaB2. NatMed 3:1285–1289.

20. Akama KT, Albanese C, Pestell RG, Van Eldik LJ 1998 Amyloidbeta-peptide stimulates nitric oxide production in astrocytesthrough an NF-kappaB-dependent mechanism. Proc Natl Acad SciUSA 95:5795–5800.

21. Steer JH, Kroeger KM, Abraham LJ, Joyce DA 2000 Glucocorti-coids suppress tumor necrosis factor-alpha expression by humanmonocytic THP-1 cells by suppressing transactivation throughadjacent NF-kappa B and c-Jun-activating transcription factor-2binding sites in the promoter. J Biol Chem 275:18432–18440.

22. Li YC, Ross J, Scheppler JA, Franza BR Jr 1991 An in vitrotranscription analysis of early responses of the human immunode-ficiency virus type 1 long terminal repeat to different transcrip-tional activators. Mol Cell Biol 11:1883–1893.

23. Joyce D, Bouzahzah B, Fu M, Albanese C, D’Amico M, Steer J,Klein JU, Lee RJ, Segall JE, Westwick JK, Der CJ, Pestell RG1999 Integration of Rac-dependent regulation of cyclin D1 tran-scription through a nuclear factor-kappaB-dependent pathway.J Biol Chem 274:25245–25249.

24. Barr RK, Kendrick TS, Bogoyevitch MA 2002 Identification of thecritical features of a small peptide inhibitor of JNK activity. J BiolChem 277:10987–10997.

25. Engedal N, Korkmaz CG, Saatcioglu F 2002 C-Jun N-terminalkinase is required for phorbol ester- and thapsigargin- induced

apoptosis in the androgen responsive prostate cancer cell lineLNCaP. Oncogene 21:1017–1027.

26. Zhou H, Lin A, Gu Z, Chen S, Park NH, Chiu R 2000 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced c-Jun N-terminalkinase (JNK) phosphatase renders immortalized or transformedepithelial cells refractory to TPA-inducible JNK activity. J BiolChem 275:22868–22875.

27. Lin SY, Liang YC, Ho YS, Tsai SH, Pan S, Lee WS 2002Involvement of both extracellular signal-regulated kinase and c-junN- terminal kinase pathways in the 12-O-tetradecanoylphorbol-13-acetate- induced upregulation of p21(Cip1) in colon cancer cells.Mol Carcinog 35:21–28.

28. Mauro A, Ciccarelli C, De Cesaris P, Scoglio A, Bouche M,Molinaro M, Aquino A, Zani BM 2002 PKCalpha-mediated ERK,JNK and p38 activation regulates the myogenic program in humanrhabdomyosarcoma cells. J Cell Sci 115:3587–3599.

29. Nomura M, Ma W, Chen N, Bode AM, Dong Z 2000 Inhibition of12-O-tetradecanoylphorbol-13-acetate-induced NF-kappaB activa-tion by tea polyphenols, (-)-epigallocatechin gallate and theafla-vins. Carcinogenesis 21:1885–1890.

30. Ariga A, Namekawa J, Matsumoto N, Inoue J, Umezawa K 2002Inhibition of tumor necrosis factor-alpha -induced nuclear translo-cation and activation of NF-kappa B by dehydroxymethylep-oxyquinomicin. J Biol Chem 277:24625–24630.

31. Vargas SJ, Naprta A, Glaccum M, Lee SK, Kalinowski J, LorenzoJA 1996 Interleukin-6 expression and histomorphometry of bonesfrom mice deficient in receptors for interleukin-1 or tumor necrosisfactor. J Bone Miner Res 11:1736–1744.

32. Lorenzo JA, Naprta A, Rao Y, Alander C, Glaccum M, Widmer M,Gronowicz G, Kalinowski J, Pilbeam CC 1998 Mice lacking thetype I interleukin-1 receptor do not lose bone mass after ovariec-tomy. Endocrinology 139:3022–3025.

33. Poli V, Balena R, Fattori E, Markatos A, Yamamoto M, Tanaka H,Ciliberto G, Rodan GA, Costantini F 1994 Interleukin-6 deficientmice are protected from bone loss caused by estrogen depletion.EMBO J 13:1189–1196.

34. Cenci S, Weitzmann MN, Gentile MA, Aisa MC, Pacifici R 2000M-CSF neutralization and egr-1 deficiency prevent ovariectomy-induced bone loss. J Clin Invest 105:1279–1287.

35. Lam J, Takeshita S, Barker JE, Kanagawa O, Ross FP, TeitelbaumSL 2000 TNF-alpha induces osteoclastogenesis by direct stimula-tion of macrophages exposed to permissive levels of RANK li-gand. J Clin Invest 106:1481–1488.

36. Carter CA 2000 Protein kinase C as a drug target: Implications fordrug or diet prevention and treatment of cancer. Curr Drug Targets1:163–183.

37. Murrills RJ, Stein LS, Horbert WR, Dempster DW 1992 Effects ofphorbol myristate acetate on rat and chick osteoclasts. J BoneMiner Res 7:415–423.

38. Teti A, Colucci S, Grano M, Argentino L, Zambonin Zallone A1992 Protein kinase C affects microfilaments, bone resorption, and[Ca2�]o sensing in cultured osteoclasts. Am J Physiol 263:C130–C139.

39. Brandstrom H, Bjorkman T, Ljunggren O 2001 Regulation ofosteoprotegerin secretion from primary cultures of human bonemarrow stromal cells. Biochem Biophys Res Commun 280:831–835.

40. Takami A, Takahashi N, Udagawa N, Miyaura C, Suda K, Woo JT,Martin TJ, Nagai K, Suda T 2000 Intracellular calcium and proteinkinase C mediate expression of receptor activator of nuclear factor-kappaB ligand and osteoprotegerin in osteoblasts. Endocrinology141:4711–4719.

41. Yamamoto A, Miyazaki T, Kadono Y, Takayanagi H, Miura T,Nishina H, Katada T, Wakabayashi K, Oda H, Nakamura K,Tanaka S 2002 Possible involvement of IkappaB kinase 2 andMKK7 in osteoclastogenesis induced by receptor activator of nu-clear factor kappaB ligand. J Bone Miner Res 17:612–621.

42. Wong BR, Josien R, Lee SY, Sauter B, Li HL, Steinman RM, ChoiY 1997 TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantlyexpressed in T cells, is a dendritic cell-specific survival factor. JExp Med 186:2075–2080.

43. Jimi E, Akiyama S, Tsurukai T, Okahashi N, Kobayashi K, Uda-gawa N, Nishihara T, Takahashi N, Suda T 1999 Osteoclastdifferentiation factor acts as a multifunctional regulator in murineosteoclast differentiation and function. J Immunol 163:434–442.

2167INHIBITION OF OSTEOCLASTOGENESIS AND NF-�B BY TPA

44. Xing L, Bushnell TP, Carlson L, Tai Z, Tondravi M, Siebenlist U,Young F, Boyce BF 2002 NF-kappaB p50 and p52 expression isnot required for RANK-expressing osteoclast progenitor formationbut is essential for RANK- and cytokine- mediated osteoclasto-genesis. J Bone Miner Res 17:1200–1210.

45. Vincenti MP, Burrell TA, Taffet SM 1992 Regulation of NF-kappaB activity in murine macrophages: Effect of bacterial lipopolysac-charide and phorbol ester. J Cell Physiol 150:204–213.

46. Lin WW, Chen BC 1998 Distinct PKC isoforms mediate theactivation of cPLA2 and adenylyl cyclase by phorbol ester inRAW264.7 macrophages. Br J Pharmacol 125:1601–1609.

47. Larsen EC, DiGennaro JA, Saito N, Mehta S, Loegering DJ,Mazurkiewicz JE, Lennartz MR 2000 Differential requirement forclassic and novel PKC isoforms in respiratory burst and phagocy-tosis in RAW 264.7 cells. J Immunol 165:2809–2817.

48. David JP, Sabapathy K, Hoffmann O, Idarraga MH, Wagner EF2002 JNK1 modulates osteoclastogenesis through both c-Junphosphorylation- dependent and -independent mechanisms. J CellSci 115:4317–4325.

49. Kim HH, Lee DE, Shin JN, Lee YS, Jeon YM, Chung CH, Ni J,Kwon BS, Lee ZH 1999 Receptor activator of NF-kappaB recruitsmultiple TRAF family adaptors and activates c-Jun N-terminalkinase. FEBS Lett 443:297–302.

50. Lomaga MA, Yeh WC, Sarosi I, Duncan GS, Furlonger C, Ho A,Morony S, Capparelli C, Van G, Kaufman S, van der Heiden A,Itie A, Wakeham A, Khoo W, Sasaki T, Cao Z, Penninger JM,Paige CJ, Lacey DL, Dunstan CR, Boyle WJ, Goeddel DV, MakTW 1999 TRAF6 deficiency results in osteopetrosis and defectiveinterleukin-1, CD40, and LPS signaling. Genes Dev 13:1015–1024.

51. Wang ZQ, Ovitt C, Grigoriadis AE, Mohle-Steinlein U, Ruther U,Wagner EF 1992 Bone and haematopoietic defects in mice lackingc-fos. Nature 360:741–745.

52. Johnson RS, Spiegelman BM, Papaioannou V 1992 Pleiotropiceffects of a null mutation in the c-fos proto-oncogene. Cell 71:577–586.

53. Srivastava S, Toraldo G, Weitzmann MN, Cenci S, Ross FP,Pacifici R 2001 Estrogen decreases osteoclast formation by down-regulating receptor activator of NF-kappa B ligand (RANKL)-induced JNK activation. J Biol Chem 276:8836–8840.

54. Shevde NK, Bendixen AC, Dienger KM, Pike JW 2000 Estrogenssuppress RANK ligand-induced osteoclast differentiation via astromal cell independent mechanism involving c-Jun repression.Proc Natl Acad Sci USA 97:7829–7834.

Address reprint requests to:Jiake Xu, MD, PhD

Department of OrthopaedicsSchool of Surgery and Pathology

University of Western AustraliaQEII Medical Centre

2nd Floor M BlockNedlands, 6009 WA, Australia

E-mail: jiakexu@cyllene.uwa.edu.au

Received in original form February 19, 2003; in revised form June2, 2003; accepted July 23, 2003.

2168 WANG ET AL.