RANK Ligand-induced Elevation of Cytosolic Ca 2 (cid:1) Accelerates Nuclear Translocation of Nuclear Factor (cid:2) B in Osteoclasts*

RANK ligand (RANKL) induces activation of NF (cid:2) B, enhancing the formation, resorptive activity, and survival of osteoclasts. Ca 2 (cid:1) transduces many signaling events, how-ever, it is not known whether the actions of RANKL involve Ca 2 (cid:1) signaling. We investigated the effects of RANKL on rat osteoclasts using microspectrofluorimetry and patch clamp. RANKL induced transient elevation of cytosolic free Ca 2 (cid:1) concentration ([Ca 2 (cid:1) ] i ) to maxima 220 n M above basal, resulting in activation of Ca 2 (cid:1) -dependent K (cid:1) current. RANKL elevated [Ca 2 (cid:1) ] i in Ca 2 (cid:1) -containing and Ca 2 (cid:1) -free media, and responses were prevented by the phospholipase C inhibitor U73122. Suppression of [Ca 2 (cid:1) ] i elevation using the intracellular Ca 2 (cid:1) chelator 1,2-bis( O -aminophen-oxy)ethane- N,N,N (cid:1) , N (cid:1) -tetraacetic acid (BAPTA) abolished the ability of RANKL to enhance osteoclast survival. Using immunofluorescence, NF by repeated passage though a pipette and plated on glass coverslips or 35-mm culture dishes. Rat osteoclasts were incubated at 37 °C in 5% CO 2 for 1 h, then gently washed with phosphate-buffered saline (PBS) to re-move nonadherent cells and incubated in fresh medium for at least 1 h before use. To quantify survival, we counted the number of rat osteoclasts in culture dishes at time 0 (time of addition of RANKL or vehicle) and at 24 h. Rabbit osteoclasts were maintained at 37 °C in 5% CO 2 for 2 h after isolation, then fresh culture medium was added and cells were incubated at 37 °C in 5% CO 2 for 2 to 7 days before use. The majority of nonosteoclastic cells were removed from rabbit preparations using Pronase E (0.001% in PBS with 0.5 m M EDTA) for (cid:4) 5 min at room temperature (22–25 °C) with intermittent agitation (modified from Ref. 20). Osteoclasts were identified by the presence of three or more nuclei, and by their characteristic morphology under phase-contrast micros- copy. Rat osteoclast precursors were identified as large (20–40 (cid:4) m in diameter) mononucleated cells that were generally circular in outline, possessed a region of granular cytoplasm, and often exhibited a broad lamellipod. These mononucleated cells all stained strongly with neutral red, and 92% stained positive for the osteoclast marker tartrate-resis-tant acid phosphatase. These procedures were approved by the Council on Animal Care of the University of Western Ontario.

RANK 1 ligand (RANKL) is a member of the tumor necrosis factor superfamily that plays an essential role in osteoclasto-genesis, as well as the activation and survival of mature osteoclasts. This factor is expressed on osteoblasts, stromal cells, B-lymphoid lineage cells, and activated T-cells as a transmembrane ligand and it also exists in a biologically active soluble form (1)(2)(3). RANKL acts through its receptor RANK, which is expressed on osteoclast precursors, mature osteoclasts, as well as dendritic cells (4). Osteoprotegerin (OPG) is a soluble decoy receptor, which binds RANKL and blocks its interaction with RANK (4).
Signaling through RANK involves the recruitment of cytosolic tumor necrosis factor receptor-associated factors (TRAFs) 1, 2, 3, 5, and 6, which in turn activate multiple signaling pathways (5)(6)(7). For example, the association of RANK with TRAF2 induces activation of c-Jun N-terminal kinase, which leads to phosphorylation of c-Jun and activation of AP-1 (7)(8)(9). TRAF6 has been implicated in activation of the nonreceptor tyrosine kinase c-Src and the transcription factor NFB (10,11).
NFB transcription factors are dimers of the five mammalian NFB proteins: p65 (RelA), RelB, c-Rel, p50 (NFB1), and p52 (NFB2). NFB regulates the expression of a large number of genes involved in cell survival as well as in cellular responses to inflammation and stress (12,13). Typically, NFB exists as a heterodimer of p50 and p65 (12). NFB is retained in the cytoplasm complexed with inhibitory proteins IBs. RANK signaling involves activation of NFB-inducing kinase, leading to activation of IB kinases (IKK) ␣ and ␤, which in turn phosphorylate serine residues on IB, targeting it for degradation in the proteasome (5,11,14). IB degradation exposes the NFB nuclear localization sequence, permitting its nuclear import. Within the nucleus, NFB acts in concert with other transcription factors to regulate gene expression, with termination of the signal caused by binding of IB (15). NFB is essential for osteoclastogenesis, as disruption of both p50 and p52 subunits of NFB leads to an osteopetrotic phenotype, because of impaired osteoclast differentiation (16).
Interaction of RANKL with RANK is crucial for osteoclast function, however, there are gaps in our understanding of the signaling events leading to activation of NFB in response to RANKL. Although interaction of RANK with TRAF6 is necessary and sufficient to activate NFB, dominant negative forms of TRAF molecules are unable to completely block NFB activation, suggesting that a TRAF-independent pathway is also involved (5,11). Because Ca 2ϩ -sensitive effectors such as calcineurin and protein kinase C (PKC) mediate NFB activation in T lymphocytes and monocytic cell lines (17,18), we considered the possible role of Ca 2ϩ in the activation of NFB by RANKL in osteoclasts.
We tested the hypothesis that RANKL signaling in osteoclasts involves elevation of [Ca 2ϩ ] i , and examined the role of cytosolic Ca 2ϩ in cell survival and activation of NFB. Classical biochemical approaches for studying osteoclasts are limited because of difficulty in isolating cells in sufficient number and purity. Furthermore, osteoclasts are terminally differentiated, and therefore do not proliferate in culture. We overcame these restrictions by studying authentic osteoclasts using single-cell techniques: microspectrofluorimetry and patch clamp to study changes in [Ca 2ϩ ] i and membrane currents, and immunofluorescence to assess nuclear translocation of NFB. We report that RANKL stimulates phospholipase C (PLC) leading to release of Ca 2ϩ from intracellular stores, transient elevation of [Ca 2ϩ ] i , and activation of Ca 2ϩ -dependent K ϩ current. The effect of RANKL on osteoclast survival was found to be dependent on elevation of [Ca 2ϩ ] i . Moreover, nuclear translocation of NFB was slowed when elevation of Ca 2ϩ was suppressed or when calcineurin or PKC were inhibited. Thus, phospholipase C and Ca 2ϩ signaling are revealed to be important regulators of NFB activation and osteoclast survival.

EXPERIMENTAL PROCEDURES
Osteoclast Isolation and Culture-Osteoclasts were isolated from the long bones of neonatal Wistar rats or neonatal New Zealand White rabbits as described previously (19). Briefly, long bones were dissected free of soft tissue and cut with a scalpel to release bone fragments into 2-3 ml of osteoclast culture medium that consisted of Medium 199 buffered with 25 mM HEPES and HCO 3 Ϫ (Invitrogen, Burlington, Ontario) supplemented with 15% heat-inactivated fetal bovine serum and 1% antibiotic solution (penicillin, 10,000 units/ml; streptomycin, 10,000 g/ml; amphotericin B, 25 g/ml). Cells were suspended by repeated passage though a pipette and plated on glass coverslips or 35-mm culture dishes. Rat osteoclasts were incubated at 37°C in 5% CO 2 for 1 h, then gently washed with phosphate-buffered saline (PBS) to remove nonadherent cells and incubated in fresh medium for at least 1 h before use. To quantify survival, we counted the number of rat osteoclasts in culture dishes at time 0 (time of addition of RANKL or vehicle) and at 24 h. Rabbit osteoclasts were maintained at 37°C in 5% CO 2 for 2 h after isolation, then fresh culture medium was added and cells were incubated at 37°C in 5% CO 2 for 2 to 7 days before use. The majority of nonosteoclastic cells were removed from rabbit preparations using Pronase E (0.001% in PBS with 0.5 mM EDTA) for ϳ5 min at room temperature (22-25°C) with intermittent agitation (modified from Ref. 20). Osteoclasts were identified by the presence of three or more nuclei, and by their characteristic morphology under phase-contrast microscopy. Rat osteoclast precursors were identified as large (20 -40 m in diameter) mononucleated cells that were generally circular in outline, possessed a region of granular cytoplasm, and often exhibited a broad lamellipod. These mononucleated cells all stained strongly with neutral red, and 92% stained positive for the osteoclast marker tartrate-resistant acid phosphatase. These procedures were approved by the Council on Animal Care of the University of Western Ontario.
Test Substances-Soluble RANKL (murine recombinant 158 -316) and OPG (human recombinant 21-194 fused at the N terminus to the Fc domain of human IgG1) were kindly provided by Amgen (Thousand Oaks, CA) and RANKL (human recombinant 151-316 fused at the N terminus to a linker peptide and a FLAG tag) was purchased from Alexis Corp. (San Diego, CA). U73122 and U73343 were obtained from Calbiochem (La Jolla, CA), dissolved in chloroform, aliquoted, evaporated under N 2 , and stored at Ϫ80°C. On the day of each experiment, U73122 and U73343 were reconstituted in dimethyl sulfoxide and added to the physiological buffer or medium bathing the cells. BAPTAacetoxymethyl ester (AM) and calcein blue-AM were obtained from Molecular Probes (Eugene, OR) and stock solutions were prepared in dimethyl sulfoxide. Osteoclasts were loaded with the BAPTA or calcein blue by incubation in loading medium (HCO 3 Ϫ -free Medium 199 buffered with HEPES supplemented with 15% fetal bovine serum and 1% antibiotic solution) containing 50 M BAPTA-AM or calcein blue-AM for 10 min at room temperature. Cyclosporin A, FK506, bisindolylmaleimide I and V were obtained from Calbiochem and stock solutions were prepared in dimethyl sulfoxide.
Fluorescence Measurement of Cytosolic Free Ca 2ϩ Concentration-[Ca 2ϩ ] i of single rat osteoclasts and osteoclast precursors was monitored using microfluorimetric techniques. Cells on glass coverslips were loaded with 1.5 M fura-2-AM (Molecular Probes) for 40 min at room temperature in loading medium. Coverslips were then placed in a chamber mounted on the stage of a Nikon Diaphot inverted phasecontrast microscope, and superfused at room temperature with physi-ological buffer containing (in mM): NaCl, 130; KCl, 5; glucose, 10; MgCl 2 , 1; CaCl 2 , 1; HEPES, 20; adjusted to pH 7.4 with NaOH; 280 -290 mOsmol/liter. The ratio of fluorescence emission at 510 nm with alternate excitation wavelengths of 345 and 380 nm was measured using a Deltascan illumination system (Photon Technology International, London, ON, Canada) as described previously (21). RANKL was applied locally to cells by pressure ejection from a micropipette. In some studies, cells were superfused with a Ca 2ϩ -free physiological buffer, supplemented with 0.5 mM EGTA.
NFB Localization by Immunofluorescence-Osteoclasts on glass coverslips were incubated with or without RANKL in osteoclast culture medium at 37°C and at the indicated times fixed with 4% paraformaldehyde (10 min), washed in PBS (3ϫ 5 min); permeabilized with 0.1% Triton X-100 in PBS (10 min), washed in PBS (3ϫ 5 min); and blocked with 1% normal goat serum in PBS (NGS) for 1-2 h at room temperature. Monoclonal antibody to p65 (catalog number sc-8008, Santa Cruz Biotechnology, Santa Cruz, CA) was diluted 1:100 in NGS and applied overnight at 4°C, followed by washing in PBS and incubation for 2 h at room temperature with biotinylated goat anti-mouse IgG (Vector Laboratories Inc., Burlingame, CA) diluted 1:200 in NGS. After incubation (2 h, room temperature) with fluorescein-conjugated streptavidin (Vector Laboratories Inc.) (1:100 in NGS), coverslips were washed, mounted on slides with Vecta-Shield (Vector Laboratories Inc.), and examined using a Zeiss LSM 510 laser-scanning confocal microscope. We assessed localization of fluorescent label in all osteoclasts on each coverslip (usually 40 -70 cells/coverslip). Osteoclasts were rated positive for nuclear localization if fluorescence intensity of one or more nuclei exceeded that of the cytoplasm.
Statistical Analyses-Data are presented as representative traces, as percentages of total cells tested, or as mean Ϯ S.E. with sample size (n) indicating the number of osteoclasts for Ca 2ϩ fluorescence or electrophysiology studies, or the number of separate cell preparations for immunofluorescence studies. Differences were assessed by one-way analysis of variance for correlated samples, followed by a Tukey or Bonferroni test and accepted as statistically significant at p Ͻ 0.05. Sigmoid curves were fit by nonlinear regression using Prism (GraphPad Software, Inc., San Diego, CA). Error bars were omitted where they were smaller than the symbol.

RANKL Induces Elevation of Cytosolic Free
Ca 2ϩ -Rat osteoclasts were loaded with fura-2, and Ca 2ϩ was monitored by microspectrofluorimetry. Osteoclasts had basal [Ca 2ϩ ] i of 154 Ϯ 4 nM (n ϭ 118, mean Ϯ S.E.). Osteoclasts responded to soluble RANKL with elevation of [Ca 2ϩ ] i , which typically peaked and then declined slowly, even in the continued presence of RANKL. Upon washout of RANKL, [Ca 2ϩ ] i returned promptly to basal levels (Fig. 1A). Multiple [Ca 2ϩ ] i transients could be elicited by successive applications of RANKL (Fig. 1B), although the subsequent responses were slightly decreased in amplitude. No responses were observed when osteoclasts were stimulated with vehicle (n ϭ 25). Moreover, OPG blocked the ability of RANKL to induce [Ca 2ϩ ] i elevations in osteoclasts that were responsive to multiple applications of RANKL alone (n ϭ 3).
The proportion of osteoclasts responding to RANKL with elevation of [Ca 2ϩ ] i was dependent on the concentration of RANKL (Fig. 1C). [Ca 2ϩ ] i elevations were elicited by concentrations of RANKL as low as 10 pg/ml. The maximum proportion of osteoclasts (ϳ60%) responded to RANKL at 10 -100 ng/ml, with half-maximal effects at ϳ0.1 ng/ml. When the amplitudes of the RANKL-induced Ca 2ϩ transients were quantified, similar concentration dependence was observed ( Fig.  1D). At concentrations of 10 -100 ng/ml, RANKL elevated [Ca 2ϩ ] i to peaks of 220 Ϯ 30 nM above basal (based on 15 responsive osteoclasts of 24 tested). We also assessed changes in Ca 2ϩ upon application of RANKL to rat osteoclast precursors. Even at concentrations of 1 g/ml, RANKL caused elevation of [Ca 2ϩ ] i in only 3 of 27 osteoclast precursors tested, whereas 11 of 17 multinucleated osteoclasts, tested in the same preparations, responded with elevation of Ca 2ϩ . As a negative control, we tested the responses of spindle-shaped stromal cells and found that none of the 12 cells tested responded to RANKL. Thus, a proportion of osteoclast precursors responded to RANKL with elevation of [Ca 2ϩ ] i , although the percentage of responsive precursors was significantly lower than that of mature osteoclasts. All subsequent studies were performed using multinucleated osteoclasts.
We next investigated the source of Ca 2ϩ contributing to RANKL-induced elevation of [Ca 2ϩ ] i in osteoclasts. RANKL elicited Ca 2ϩ elevations of comparable amplitude in Ca 2ϩ -containing and Ca 2ϩ -free extracellular solutions, consistent with release of Ca 2ϩ from intracellular stores (Fig. 2, A and B, n ϭ 5). Ca 2ϩ release from stores often involves PLC-mediated production of inositol 1,4,5-trisphosphate. We have shown previously that the PLC inhibitor U73122 blocks P2Y nucleotide receptor-mediated elevation of [Ca 2ϩ ] i in osteoclasts (21 ] i elevations in rat osteoclasts. Single rat osteoclasts were loaded with fura-2, bathed in physiological buffer, and [Ca 2ϩ ] i was monitored by microspectrofluorimetry. A, RANKL (100 ng/ml) was applied to cells for 60 s, as indicated by the bar below the Ca 2ϩ trace. B, illustrated is the response of one osteoclast to 2 successive stimulations with RANKL (10 ng/ml), where Ca 2ϩ transients diminished in amplitude upon successive stimulation. C, the percentage of osteoclasts responding to the single application of RANKL increased with increasing RANKL concentration (10 Ϫ3 to 10 2 ng/ml, applied locally, n ϭ 12 osteoclasts for each concentration, except 10 Ϫ3 ng/ml, where n ϭ 6). Elevations in [Ca 2ϩ ] i Ͼ25 nM above basal were considered to be responses. D, amplitudes of Ca 2ϩ transients were quantified as maximum elevations above basal levels. The curve illustrates dependence of amplitude on RANKL concentration. Data are mean Ϯ S.E. of three to eight responsive osteoclasts for concentrations of RANKL Ն10 Ϫ2 ng/ml.

FIG. 2. RANKL causes release of Ca 2؉ from intracellular stores.
A, to examine the source of Ca 2ϩ contributing to the response, RANKL was applied locally to rat osteoclasts in the presence or absence of extracellular Ca 2ϩ . Illustrated are responses of one osteoclast to RANKL (100 ng/ml, 60 s applications) superfused with physiological buffer containing 1 mM Ca 2ϩ . Where indicated, cells were superfused with nominally Ca 2ϩ -free buffer containing 0.5 mM EGTA (0 Ca 2ϩ ). Dotted lines indicate when the recording was interrupted for 60 s to allow buffer exchange. B, histogram illustrates the amplitude of Ca 2ϩ transients during applications of RANKL (100 ng/ml) in the presence and absence of extracellular Ca 2ϩ (means Ϯ S.E., n ϭ 5 cells). Withdrawal of Ca 2ϩ did not significantly affect the amplitude of the response. C, osteoclasts were stimulated twice with RANKL (100 ng/ml) before application of test compounds to ensure responsiveness to repeated stimulation with RANKL. Cells were then treated for 10 min with vehicle (dimethyl sulfoxide, 0.1%), U73122, or U73343 (1 M in the bath). Traces illustrate responses of separate osteoclasts challenged with RANKL (100 ng/ml) following treatment with the indicated test compound. D, histogram illustrates the amplitude of RANKL-induced Ca 2ϩ transients following treatment with the indicated test compound. Data are expressed as a percentage of response to RANKL in the same cell prior to treatment. Data are means Ϯ S.E. of 5 cells for each condition. U73122 significantly inhibited RANKL-induced elevations of [Ca 2ϩ ] i when compared with responses elicited by RANKL in osteoclasts treated with vehicle or U73343 (p Ͻ 0.05). exhibits this current (22), we used patch clamp techniques to monitor the effects of RANKL on membrane currents of rabbit osteoclasts, which all demonstrate the current (25). Cells were held at Ϫ30 mV and voltage ramp commands were applied every 2 s. RANKL (100 ng/ml) evoked outward current after a delay of ϳ10 s (n ϭ 6 out of 14 osteoclasts tested) with inward current apparent at Ϫ100 mV (Fig. 3A). Currentvoltage (I-V) relationships were determined from the voltage ramp commands. Basal current prior to application of RANKL was dominated by the inwardly rectifying K ϩ current Kir2.1 that has been identified previously in osteoclasts (Fig. 3B, Control). Subtraction of the control current from that recorded at the peak of the response to RANKL showed that the RANKL-induced current was linear and reversed close to Ϫ70 mV, indicating K ϩ -selective current (Fig. 3C). A similar, linear K ϩ current has been shown previously to closely follow elevations of [Ca 2ϩ ] i in osteoclasts (22). Thus, RANKL-induced current likely represents activation of Ca 2ϩdependent K ϩ channels because of rise of [Ca 2ϩ ] i . Hence, the voltage-clamp data independently confirm that RANKL induces elevation of [Ca 2ϩ ] i in osteoclasts.
Role of Ca 2ϩ in Osteoclast Survival-It was shown previously that RANKL prolongs osteoclast survival in vitro (26). We investigated the role of Ca 2ϩ in this process using the intracellular Ca 2ϩ chelator BAPTA. To establish conditions for effective buffering of Ca 2ϩ by BAPTA, we used ATP, which activates P2Y nucleotide receptors on osteoclasts leading to release of Ca 2ϩ from intracellular stores and reproducible elevation of [Ca 2ϩ ] i (21). Osteoclasts were stimulated with ATP (100 M) to ensure their responsiveness, then treated with different concentrations of BAPTA-AM and rechallenged with ATP. We established that loading with 50 M BAPTA-AM for 10 min at room temperature was optimal for suppressing elevation of [Ca 2ϩ ] i induced by ATP (Fig. 4A, n ϭ 8). We then confirmed that, under these conditions, BAPTA was effective in preventing RANKL-induced elevation of [Ca 2ϩ ] i (Fig. 4B).
To examine the role of [Ca 2ϩ ] i in osteoclast survival, cells were treated with BAPTA-AM or vehicle. The medium was changed and osteoclasts were incubated with RANKL (100 ng/ml) or vehicle at 37°C for 24 h. The number of osteoclasts per dish at 24 h was expressed as a percentage of the initial number of osteoclasts in the same dish. As expected, RANKL significantly increased the number of osteoclasts that survived 24 h (Fig. 4C, n ϭ 4 independent experiments). BAPTA suppressed this effect of RANKL, but did not affect osteoclast survival in the absence of RANKL (Fig. 4C). Thus, elevation of [Ca 2ϩ ] i appears to be necessary for RANKL to promote osteoclast survival.
Effect of PLC Inhibitor on RANKL-induced Nuclear Translocation of NFB-NFB is one of the major downstream effectors of RANK signaling and activation of NFB enhances cell survival in many cell types (13,26). Therefore, we investigated the possible involvement of the PLC/Ca 2ϩ pathway in the nuclear translocation of NFB in osteoclasts. Activation of NFB was assessed at the single-cell level using immunofluorescence to monitor the spatial distribution of the p65 subunit of NFB. In the majority of untreated rat or rabbit osteoclasts, NFB was located in the cytoplasm for the duration of the experiment (Fig. 5A, left). In a proportion of osteoclasts, RANKL induced redistribution of NFB to the nuclei, most often to all the nuclei within a single osteoclast (Fig. 5A, right). Nuclear translocation of NFB was rapid (within 7 min after addition of RANKL) and transient, reversing within 60 min (Fig. 5B). Maximal translocation was evident by 15 min, with 52 Ϯ 4% of osteoclasts exhibiting nuclear localization of NFB (n ϭ 12 independent experiments), compared with 4 Ϯ 1% in untreated osteoclasts (n ϭ 7 independent experiments).
To examine a role of the PLC signaling pathway, we determined the kinetics of RANKL-induced NFB translocation in osteoclasts treated with the PLC inhibitor U73122, which prevents [Ca 2ϩ ] i elevations induced by RANKL. U73122 markedly delayed translocation of NFB from the cytoplasm to the nuclei in response to RANKL. In U73122-treated cells, maximum translocation was delayed until 30 min following addition of RANKL (versus 15 min in parallel samples treated with RANKL alone) (Fig. 5B). U73122 did not significantly affect the maximum proportion of osteoclasts exhibiting nuclear translocation of NFB in response to RANKL (100 ng/ml) (47 Ϯ 6% in control versus 44 Ϯ 5% in U73122-treated, based on parallel samples from six independent experiments). U73343, a closely related analog of U73122, which does not inhibit PLC or block RANKL-induced elevation of [Ca 2ϩ ]i (Fig. 2, C and D), had no significant effect on RANKL-induced translocation of NFB (Fig. 5C). Thus, we provide evidence that RANKL signaling through PLC affects the kinetics of NFB translocation.
Effect of BAPTA on RANKL-induced Nuclear Translocation of NFB-We next examined whether chelation of intracellular Ca 2ϩ using BAPTA affected the kinetics of RANKL-induced translocation of NFB. Using loading conditions established above, we found that BAPTA delayed nuclear translocation of NFB induced by RANKL (Fig. 6, data based on parallel samples from seven independent experiments). In BAPTA-loaded osteoclasts, maximum translocation was observed 30 -60 min following treatment with RANKL (100 ng/ml) versus 15-30 min in cells treated with RANKL alone. Loading of cells with BAPTA significantly reduced the proportion of osteoclasts exhibiting nuclear localization of NFB at 15 min, whereas the proportion of cells exhibiting nuclear localization at 60 min was significantly increased (as indicated by asterisks in Fig. 6). Furthermore, BAPTA reduced the maximum proportion of osteoclasts exhibiting nuclear translocation of NFB in response to treatment with RANKL (41 Ϯ 6% for BAPTA-loaded osteoclasts versus 57 Ϯ 5% for control osteoclasts). In the absence of RANKL, BAPTA did not significantly affect NFB distribution (Fig. 6).
Hydrolysis of BAPTA-AM results in release of small molecular weight products because of the degradation of the acetoxymethyl ester (AM) moieties. To determine whether these degradation products might affect NFB translocation, we ex-  05%, vehicle) for 10 min. The medium was then changed and cells were incubated with or without RANKL (100 ng/ml) at 37°C for the indicated times. The number of osteoclasts exhibiting nuclear localization of NFB was expressed as a percentage of maximum translocation observed in each experiment. Loading with BAPTA, but not calcein blue, significantly delayed RANKL-induced nuclear translocation of NFB. Data are means Ϯ S.E. of seven independent experiments, except for AM control where n ϭ 3. 100% value was 59 Ϯ 4% (of total number of osteoclasts). *, indicates significant difference of BAPTA/RANKL compared with RANKL alone, p Ͻ 0.05. amined cells treated with calcein blue-AM (a compound that bears the same AM modification as BAPTA-AM, but is ineffective as a Ca 2ϩ chelator at physiological [Ca 2ϩ ] i ). Loading cells with calcein blue did not affect the kinetics or degree of RANKL-induced NFB translocation (Fig. 6, n ϭ 3), arguing against possible nonspecific effects of the AM degradation products. Taken together, these data indicate that RANKL-induced elevation of [Ca 2ϩ ] i accelerates nuclear translocation of NFB.
Role of Calcineurin and Protein Kinase C in RANKL-induced Nuclear Translocation of NFB-We examined the role of a Ca 2ϩ -regulated effector known to contribute to activation of NFB in other systems, the Ca 2ϩ -calmodulin-dependent phosphatase, calcineurin. The calcineurin inhibitor, cyclosporin A (1 M), applied 30 min before addition of RANKL, suppressed the initial NFB translocation (7 min), with no significant effect at later times (15-30 min) (Fig. 7, data based on parallel samples from eight independent experiments). The structurally distinct calcineurin inhibitor, FK506, had similar effects. FK506 (10 nM) significantly suppressed RANKL-induced NFB translocation at 7 min to 67 Ϯ 7% of control, but had no significant effect at 15 min (94 Ϯ 7% of control) (n ϭ 8). These findings are consistent with RANKL-induced elevation of [Ca 2ϩ ] i causing transient activation of calcineurin, which in turn accelerates activation of NFB.
PLC activation leads to the generation of 2 second messengers, Ca 2ϩ and diacylglycerol, both of which contribute to activation of PKC. In other systems, PKC can activate IB kinases, inducing translocation of NFB (12). In osteoclasts, the PKC inhibitor bisindolylmaleimide I (100 nM) applied 30 min before addition of RANKL suppressed NFB translocation at 7 and 15 min (Fig. 7, data based on parallel samples from seven independent experiments), with no significant effect at 30 min. In contrast, the control compound, bisindolylmaleimide V (100 nM) that does not inhibit PKC, had no significant effect on RANKL-induced NFB translocation (107 Ϯ 3 and 107 Ϯ 4% of control at 7 and 15 min, respectively, n ϭ 6). These findings are consistent with a role for PKC, activated following stimulation of phospholipase C, in mediating the effects of RANKL on NFB.
The effects of cyclosporin A and bisindolylmaleimide I were additive at 7 min, however, the calcineurin inhibitor had no additional effect at 15 min (Fig. 7, data based on parallel samples from seven independent experiments). Similarly, NFB translocation at 7 min in samples treated with FK506 together with bisindolylmaleimide I was 54 Ϯ 6% of control, significantly less than in samples treated with bisindolylmaleimide I or FK506 alone (77 Ϯ 3 and 67 Ϯ 7%, respectively, n ϭ 7). Like cyclosporin A, FK506 had no additional effect at 15 min (translocation in the presence of FK506 and bisindolylmaleimide I was 78 Ϯ 4% of control versus 71 Ϯ 5% in osteoclasts treated with bisindolylmaleimide I alone, n ϭ 7). Taken together, these data indicate that calcineurin and PKC are downstream effectors of RANK that accelerate activation of NFB. DISCUSSION We demonstrate that RANKL induces transient elevation of [Ca 2ϩ ] i in osteoclasts because of activation of PLC and release of Ca 2ϩ from intracellular stores. The rise of [Ca 2ϩ ] i stimulates Ca 2ϩ -dependent K ϩ current, accelerates nuclear translocation of NFB, and promotes osteoclast survival, indicating key functional roles for PLC and Ca 2ϩ in RANK signaling in osteoclasts.
Half-maximal effects of RANKL on [Ca 2ϩ ] i were observed at ϳ0.1 ng/ml, with maximal actions at 10 -100 ng/ml. These findings are in keeping with the concentration dependence reported for the induction of osteoclastogenesis by RANKL (27,28), suggesting that elevation of [Ca 2ϩ ] i is mediated by RANK. Moreover, OPG prevented RANKL-induced elevation of [Ca 2ϩ ] i , ruling out possible nonspecific effects. Repeated application of RANKL elicited multiple Ca 2ϩ transients indicating lack of short term desensitization. In vivo, RANK signaling is thought to be mediated primarily by interaction of osteoclasts and their precursors with cells expressing RANKL (stromal cells, osteoblasts, and lymphocytes). Lack of desensitization would allow osteoclasts to receive multiple signals from neighboring cells, giving rise to temporal and spatial summation of RANKL signals.
Our data demonstrate that RANK signaling in osteoclasts involves PLC. PLC-␥ is activated by growth factor receptor tyrosine kinases or nonreceptor tyrosine kinases linked to cytokine receptors (29). In this regard, RANK and TRAF6 interact with, and activate the nonreceptor tyrosine kinase c-Src (10), which may serve to recruit PLC-␥ to the RANK-signaling complex in osteoclasts, as was shown for endothelial cells (30). Targeted disruption of c-src results in an osteopetrotic phenotype because of compromised resorptive function of osteoclasts (31). Interestingly, osteoclastogenesis is not impaired in c-src knockout mice, indicating a different requirement for c-Src signaling in osteoclasts and their precursors. In keeping with these differences, we observed that RANKL-induced Ca 2ϩ signaling is more prominent in mature osteoclasts than in precursors. Furthermore, TRAF6-deficient mice have been reported to display osteopetrosis because of nonfunctional osteoclasts (32), also suggesting critical differences between RANKL signaling in precursors and mature osteoclasts. However, osteoclastogenesis was found to be impaired in another TRAF6 knockout model (33), leaving open the question of the precise role for TRAF6-dependent pathways in RANKL-induced osteoclastogenesis and activation of resorption.
As in other cell types, cytosolic Ca 2ϩ likely plays important roles in regulating a number of osteoclast functions. We have shown that RANKL activates the intermediate conductance provide independent evidence to support RANKL-induced elevation of [Ca 2ϩ ] i . In vivo, the membrane hyperpolarization resulting from activation of IK Ca would modulate the activity of electrogenic ion transport systems and increase the driving force for influx of Ca 2ϩ , a vital process in T cell signaling (34). Similarly, IK Ca may play an important role in osteoclast regulation, especially considering that IK Ca channels are selectively expressed in rat osteoclasts having the morphological characteristics of actively resorbing cells (22). Ca 2ϩ also regulates cytoskeletal organization through Ca 2ϩ -dependent actin-binding proteins, such as gelsolin, that play critical roles in osteoclast motility (35). Therefore, RANKL-induced alterations in [Ca 2ϩ ] i could have multiple downstream effects in osteoclasts. In this regard, we found that elevation of [Ca 2ϩ ] i is necessary for RANKL to prolong osteoclast survival. Although previous reports have suggested that elevation of [Ca 2ϩ ] i in osteoclasts leads to inhibition of resorption (36), more recent studies have implicated Ca 2ϩ -dependent pathways in promoting osteoclast formation (37). In osteoclasts, as in other systems, the effects of [Ca 2ϩ ] i elevation likely depend upon the pattern, amplitude, and duration of the Ca 2ϩ signal as well as its interactions with other signaling pathways.
NFB is a key transcription factor that promotes cell survival in many systems (13). In the present study, we have shown that RANKL-induced Ca 2ϩ signaling accelerates nuclear translocation of NFB. Several mechanisms may be considered. Elevation of [Ca 2ϩ ] i alone appears to be insufficient to activate NFB in osteoclasts. Nucleotides, such as ATP (10 -100 M), which bind to the P2Y class of G protein-coupled receptors causing even greater release Ca 2ϩ from stores (21), do not activate NFB in osteoclasts. 2 Similarly, elevation of [Ca 2ϩ ] i alone is insufficient to activate NFB in immune cells (18,38). Therefore, Ca 2ϩ appears to act in concert with other canonical signaling pathways to accelerate activation of NFB in osteoclasts.
RANK signaling involves a cascade of events, beginning with the recruitment of TRAFs, leading to activation of NFB-inducing kinase and the IKK complex (5,11,14). Enhanced activity of the enzyme complex would increase phosphorylation of IB, accelerating translocation of NFB to the nucleus. Ca 2ϩ may act to increase activity of the IKK complex, and indeed, IKK␤ is reported to be activated synergistically by two Ca 2ϩdependent mediators of T cell receptor signaling, calcineurin and PKC (18). We show in osteoclasts that RANK signaling involves these same mediators.
Calcineurin is a serine/threonine phosphatase that is stimulated by Ca 2ϩ . The immunosuppressant drugs cyclosporin A and FK506 bind distinct immunophilins and inhibit the activity of calcineurin (38). We found that both of these inhibitors delayed the initial phase of NFB activation in osteoclasts, without affecting maximal activation. Similar actions of distinct inhibitors support the interpretation that their effects are mediated by calcineurin. The actions of calcineurin inhibitors were restricted to early time points, which might be explained by the fact that elevation of [Ca 2ϩ ] i in response to RANKL is transient, giving rise to only brief activation of calcineurin. Our findings may have relevance to previously observed inhibitory effects of cyclosporin A on osteoclastic resorption in vitro (39).
Stimulation of PLC also leads to increased activation of PKC. To investigate the involvement of PKC, we used bisindolylmaleimide I, a selective inhibitor of the conventional and novel PKC isoforms, and carried out control studies using the inactive analog, bisindolylmaleimide V. Inhibition of PKC also caused delay in NFB activation, again with no marked change in maximal activation. In contrast to the calcineurin blockers, PKC inhibition resulted in a greater delay in NFB translocation, which could reflect the time course for production of diacylglycerol. Therefore, it would appear that early stages of NFB activation involve the convergence of several signaling pathways. Moreover, the effects of calcineurin and PKC blockers were additive, and similar in amplitude to the effects of BAPTA, all supporting a critical role for Ca 2ϩ in controlling the latency of NFB activation.
Because the concurrence of multiple transcription factors regulates expression of genes, the kinetics of their activation is of critical importance. In this regard, the temporal pattern of NFB activation has recently been shown to generate specificity in gene expression (40). In concert with other transcription factors, NFB controls cell survival and the expression of genes encoding cytokines and adhesion molecules (12,13,41). RANK also activates AP-1, which requires cooperative interactions with other transcription factors and coactivators, including NFB, to achieve stimulus-specific regulation of transcription (42,43). In this regard, c-Jun and c-Fos have been shown to interact directly with the p65 subunit of NFB, enhancing transactivation via both the B and AP-1 response elements (44). In addition, NFB interacts with the cancer-amplified transcription coactivator ASC-2 and other transcriptional regulators to control gene expression (45). Therefore, the relative activities of NFB and other transcription factors at any point of time will determine the composition of multipartner transcription complexes and consequently gene expression. Thus, distinct responses will depend upon the kinetics of activation of NFB and other transcription factors stimulated by RANKL or by other signaling molecules acting on the osteoclast. Whether the dependence of osteoclast survival on [Ca 2ϩ ] i elevation is because of changes in the kinetics of NFB translocation or because of the activation of other Ca 2ϩ -dependent pathways is yet to be determined. Nevertheless, the cross-talk between NFB and Ca 2ϩ signaling demonstrated here provides a novel mechanism for the temporal regulation of NFB activity and gene expression in osteoclasts and other cell types.