Dr Ke is an employee of Pfizer Inc. All other authors have no conflict of interest.
Extracellular Nucleotides Act Through P2X7 Receptors to Activate NF-κB in Osteoclasts†
Article first published online: 5 JAN 2004
Copyright © 2004 ASBMR
Journal of Bone and Mineral Research
Volume 19, Issue 4, pages 642–651, April 2004
How to Cite
Korcok, J., Raimundo, L. N., Ke, H. Z., Sims, S. M. and Dixon, S. J. (2004), Extracellular Nucleotides Act Through P2X7 Receptors to Activate NF-κB in Osteoclasts. J Bone Miner Res, 19: 642–651. doi: 10.1359/JBMR.040108
- Issue published online: 2 DEC 2009
- Article first published online: 5 JAN 2004
- Manuscript Accepted: 20 NOV 2003
- Manuscript Revised: 9 NOV 2003
- Manuscript Received: 11 SEP 2003
Nucleotides, released in response to mechanical and other stimuli, act on P2 receptors in osteoclasts and other cell types. In vitro studies of osteoclasts from rabbits and P2X7 receptor-deficient mice revealed that P2X7 receptors couple to activation of the key transcription factor NF-κB.
Introduction: Osteoclasts express functional P2X4 and P2X7 receptors, which are ATP-gated cation channels. Knockout (KO) of the P2X7 receptor has revealed its role in regulating bone formation and resorption, but the underlying signals are not known. The transcription factor NF-κB plays a key role in the response of osteoclasts to RANKL and other cytokines. The aim of this study was to examine whether P2X receptors on osteoclasts signal through NF-κB.
Materials and Methods: Osteoclasts were isolated from neonatal rabbits or wildtype (WT) and P2X7 receptor KO mice. Immunofluorescence was used to detect the p65 subunit of NF-κB, which, on activation, translocates from the cytosol to the nuclei. The concentration of cytosolic free Ca2+ ([Ca2+]i) was monitored in single osteoclasts loaded with fura-2.
Results: In control samples, few rabbit osteoclasts demonstrated nuclear localization of NF-κB. Benzoyl-benzoyl-ATP (BzATP, a P2X7 agonist, 300 μM) induced nuclear translocation of NF-κB after 3 h in ∼45% of rabbit osteoclasts. In contrast, a low concentration of ATP (10 μM, sufficient to activate P2X4 and P2Y2, but not P2X7 receptors) did not induce nuclear translocation of NF-κB. Because BzATP activates multiple P2 receptors, we examined responses of osteoclasts derived from WT and P2X7 receptor KO mice. Treatment with BzATP for 30 minutes increased nuclear localization of NF-κB in osteoclasts from WT but not KO mice, showing involvement of P2X7 receptors. Both ATP (10 μM) and BzATP (300 μM) caused transient elevation of [Ca2+]i, indicating that rise of calcium alone is not sufficient to activate NF-κB. Pretreatment of rabbit osteoclasts with osteoprotegerin inhibited translocation of NF-κB induced by RANKL but not by BzATP, establishing that the effects of BzATP are independent of RANKL signaling.
Conclusion: These findings show that P2X7 nucleotide receptors couple to activation of NF-κB in osteoclasts. Thus, nucleotides, released at sites of inflammation or in response to mechanical stimuli, may act through NF-κB to regulate osteoclast formation and activity.
OSTEOCLAST DIFFERENTIATION AND activation are dependent on the presence of key signaling molecules, such as RANKL. RANKL, a TNF-related trimeric cell surface protein, is synthesized by stromal cells, osteoblasts, and activated lymphocytes.(1) RANKL binds to its receptor RANK, a member of the TNF receptor superfamily, to stimulate osteoclastogenesis and activate bone resorption. Alternatively, RANKL can bind to the decoy receptor osteoprotegerin (OPG), which prevents its signaling and thus blocks osteoclast formation and activation. Through interaction with TNF receptor-associated factors (TRAFs) 1, 2, 3, 5, and 6, RANK activates multiple signaling pathways, including the transcription factors activator protein-1 (AP-1) and NF-κB.(1)
NF-κB regulates the expression of a variety of genes involved in inflammation and immunity, cell proliferation, responses to stress, and apoptosis. NF-κB functions both as homo and heterodimers, the predominant form being p50/p65.(2) The essential role of NF-κB in osteoclast development was discovered in knockout (KO) mice lacking both the p50 and p52 subunits of NF-κB, in which osteoclasts failed to develop, resulting in severe osteopetrosis.(3)
ATP is an important signaling molecule that can be released into the extracellular environment by a number of mechanisms. ATP and ADP are stored within dense granules of platelets, where they reach a concentration of about 1 M.(4) During trauma, nucleotides are released from activated platelets, as well as the cytoplasm of damaged and dying cells.(5) ATP is also released by vesicular secretion from sympathetic nerves, acting as a cotransmitter with noradrenaline.(6) In a number of cell types, mechanical stimulation or fluid flow induces release of nucleotides.(7, 8) Consequently, nucleotides have been suggested to play a role in mechanotransduction in skeletal tissues.(9)
Extracellular nucleotides act through P2 receptors, classified as P2Y and P2X. Whereas P2Y receptors are G-protein-coupled receptors that in many cases induce release of Ca2+ from intracellular stores, P2X receptors are ligand-gated cation channels.(10) On activation, P2X receptors mediate cation fluxes, and in some cases, Ca2+ influx. Compared with other P2X receptors, the P2X7 receptor requires relatively high concentrations of ATP for activation (>100 μμ). Moreover, when divalent cations are present at low concentrations, several P2X receptors (P2X2, P2X4, and P2X7) form pores permeable to hydrophilic molecules as large as 900 Da.(11-13) It is not clear whether the receptors themselves form these pores or induce other proteins to form pores in the plasma membrane.
Characterization of KO mice has revealed an important role for the P2X7 receptor in controlling bone mass. KO mice exhibit a unique skeletal phenotype—diminished periosteal bone formation together with excessive trabecular bone resorption.(14) The precise mechanisms responsible for these effects remain unknown. However, functional P2X4 and P2X7 receptors have been demonstrated in osteoclasts,(14-17) and extracellular nucleotides have been shown to have multiple effects on osteoclast function. Low concentrations of ATP stimulate the formation and resorptive activity of rat osteoclasts in vitro, whereas high concentrations inhibit osteoclast formation.(18) The P2X7 receptor has been implicated in the formation of multinucleated cells by the fusion of murine macrophage-like cells(19) and has been suggested to play a role in the fusion of osteoclast precursors.(20) Furthermore, P2X7 receptor antagonists inhibit the formation of osteoclasts from human peripheral blood mononuclear cells in vitro.(17) On the other hand, P2X7 receptor KO mice possess multinucleated osteoclasts, indicating that this receptor is not essential for the fusion of osteoclast precursors in vivo.(14)
P2X7 receptors mediate a nonselective cation current and Ca2+ influx(21, 22); however, other signaling pathways activated by these receptors in osteoclasts remain largely unknown. In this study, we show that stimulation of rabbit osteoclasts with benzoyl-benzoyl-ATP (BzATP) induces activation of NF-κB. Although BzATP is a relatively potent P2X7 receptor agonist, it also activates other P2 receptors.(23) Consequently, we used osteoclasts isolated from P2X7 receptor KO mice to identify signaling pathways activated specifically by the P2X7 receptor. We demonstrated that BzATP activates NF-κB in osteoclasts isolated from wildtype (WT), but not KO, mice, indicating involvement of the P2X7 receptor. Furthermore, we demonstrated that nucleotide-induced activation of NF-κB is independent of RANKL signaling, consistent with the involvement of P2X7 receptors on osteoclasts rather than osteoblasts.
MATERIALS AND METHODS
Medium 199 (containing 25 mM HEPES and 26 mM HCO3−), HCO3−-free medium 199 (25 mM HEPES), heat-inactivated FBS, and antibiotic solution (penicillin 10,000 U/ml; streptomycin 10,000 μg/ml; amphotericin B 25 μg/ml) were from Invitrogen (Burlington, Canada). All nucleotides were purchased from Sigma (St Louis, MO, USA). Mouse monoclonal (sc-8008) and rabbit polyclonal antibodies (sc-109) against the p65 subunit of NF-κB were from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Mounting medium (Vecta-Shield), biotinylated goat anti-mouse IgG, and biotinylated goat anti-rabbit IgG were from Vector Laboratories (Burlingame, CA, USA). Fluorescein-conjugated streptavidin and fluorescent probes TOTO-3, SYTO-13, and fura-2-AM were from Molecular Probes (Eugene, OR, USA). U73122 was obtained from Calbiochem (La Jolla, CA, USA), dissolved in chloroform, aliquoted, evaporated under N2, and stored at −80°C. On the day of each experiment, U73122 was reconstituted in dimethyl sulfoxide and added to the physiological buffer bathing the cells. 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, USA). OPG (recombinant human 21-194) was purchased from Research Diagnostics (Flanders, NJ, USA).
Osteoclast isolation and culture
Osteoclasts were isolated from the long bones of neonatal New Zealand White rabbits and neonatal WT and P2X7 receptor KO mice. P2X7 receptor KO mice were generated by Solle et al.(24) Both WT and P2X7 receptor KO mice were of mixed genetic origin (129/Ola × C57BL/6 × DBA/2). Genotypes were confirmed using RT-PCR. Osteoclasts were isolated according to previously described procedures.(22) Briefly, rabbit and mice 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 HEPES and HCO3−, supplemented with 15% FBS and 1% antibiotic solution. Cells were suspended by repeated passage of the bone fragments through a pipette and plated on glass coverslips or 35-mm culture dishes. Rabbit osteoclasts were maintained at 37°C and 5% CO2 for 2 h after isolation, fresh culture medium was added, and cells were incubated at 37°C and 5% CO2 for 2-7 days before use. The majority of non-osteoclastic cells were removed from rabbit preparations using pronase E (0.001% in PBS with 0.5 mM EDTA) for ∼5 minutes at room temperature (22–25°C) with intermittent agitation.(25) To assess viability, cells were incubated in osteoclast culture medium containing ethidium bromide (2 μg/ml) for 20 minutes at 37°C. For the final 5-10 minutes, osteoclasts were exposed to the membrane-permeant, nucleic acid stain, SYTO-13 (1 μM), and fluorescence was assessed using an Axiovert S100 microscope (Carl Zeiss, Thornwood, NY, USA).
Mouse osteoclasts were incubated at 37°C in 5% CO2 for 1 h, gently washed with PBS to remove nonadherent cells and incubated in fresh culture medium for 1-5 h before use. Osteoclasts were identified by the presence of three or more nuclei and by their characteristic morphology under phase-contrast microscopy. For selected samples, nucleotide-induced changes in morphology were monitored by time-lapse video microscopy as described previously.(26) These procedures were approved by the Council on Animal Care of the University of Western Ontario.
NF-κB localization by immunofluorescence
Osteoclasts on glass coverslips were incubated with various test agents in osteoclast culture medium at 37°C, and at indicated times, washed in PBS (2×), fixed with 4% paraformaldehyde (10 minutes), washed in PBS (2 × 10 minutes), permeabilized with 0.1% Triton X-100 in PBS (10 minutes), washed in PBS (2 × 5 minutes), and blocked with 1% normal goat serum in PBS (blocking solution) for 1-2 h at room temperature. Osteoclasts were incubated overnight at 4°C with primary antibody to the p65 subunit of NF-κB (diluted 1:100 for rabbit or 1:200 for mouse in blocking solution). Coverslips were washed and incubated for 2 h at room temperature with biotinylated secondary antibody diluted to 1:100 in blocking solution, followed by washing and incubation with fluorescein-conjugated streptavidin (1:100 in PBS) for 30 minutes at room temperature. Nuclei were stained with TOTO-3 (2 μM), coverslips were washed and mounted on slides, and cells were observed using a Zeiss LSM 510 laser-scanning confocal microscope. All osteoclasts on each coverslip were examined (usually 30-50 osteoclasts/coverslip in mouse cultures and 200-250 osteoclasts/coverslip in rabbit cultures). Osteoclasts were rated positive for nuclear localization if NF-κB fluorescence labeling of one or more nuclei exceeded (in rabbit osteoclasts) or was equal (in mouse osteoclasts) to that of the cytoplasm.
Fluorescence measurement of cytosolic free Ca2+ concentration
[Ca2+]i of single rabbit osteoclasts loaded with fura-2 was monitored using microfluorimetric techniques. Cells on glass coverslips were incubated for 40 minutes at room temperature in HCO3−-free osteoclast culture medium containing 1.5 μM fura-2-AM. Coverslips were placed in a chamber mounted on the stage of a Nikon Diaphot inverted phase-contrast microscope and superfused at room temperature with buffer containing (in mM) 135 NaCl, 5 KCl, 10 glucose, 1 CaCl2, and 20 HEPES (adjusted to pH 7.4 with NaOH; 280-290 mOsm/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, Ontario, Canada) as described previously.(27) Test substances were applied locally to cells by pressure ejection from a micropipette. In some studies, cells were superfused with Ca2+-free buffer, supplemented with 0.5 mM EGTA.
Data are presented as representative images or traces or as percentages of total cells tested expressed as means ± SE. The sample size (n) indicates the number of osteoclasts for Ca2+ fluorescence determinations or the number of separate cell preparations for immunofluorescence studies. Data were analyzed by one-way ANOVA followed by a Bonferroni's post-test. Differences were accepted as statistically significant at p < 0.05. Error bars were omitted where they were smaller than the symbol.
NF-κB translocation in rabbit osteoclasts
To establish that immunofluorescence was an effective tool to assess translocation of NF-κB, rabbit osteoclasts were exposed to the potent activator of NF-κB, RANKL. Osteoclasts treated with RANKL (100 ng/ml, 30 minutes) showed strong nuclear localization of NF-κB (Fig. 1A), whereas vehicle-treated osteoclasts typically showed cytoplasmic localization (Fig. 1B). RANKL transiently increased the percentage of osteoclasts with nuclear localization of NF-κB, peaking at 30 minutes (41 ± 7% osteoclasts, p < 0.05 compared with 0 h). By 1 h, NF-κB localization was predominantly cytoplasmic and was not significantly different from 0 h (Fig. 1C). Vehicle-treated osteoclasts showed low levels of nuclear NF-κB up to 4 h (Fig. 1D). The time course for RANKL-induced activation of NF-κB in rabbit osteoclasts is consistent with that in osteoclasts isolated from rats and human giant cell tumors,(28, 29) indicating that immunofluorescence is an effective tool for examining NF-κB activation in rabbit osteoclasts.
Osteoclasts express functional P2X7 receptors.(14, 16, 17, 22) To investigate whether these receptors couple to NF-κB signaling, rabbit osteoclasts were exposed to BzATP—a nucleotide analog that activates P2X7 receptors with greater potency than ATP. BzATP (100 μM) increased the percentage of rabbit osteoclasts that exhibited nuclear translocation of NF-κB with a maximal effect at a 3-h exposure (Figs. 2A and 2C). However, BzATP is not specific for P2X7,(23) because it also activates a number of other P2 receptors including P2X4 and some P2Y receptors. Therefore, we examined responses to a low concentration of ATP (10 μM), sufficient to activate P2X4 and certain P2Y receptors, but not the relatively low-affinity P2X7 receptor. In contrast to the action of BzATP (100 μM), even prolonged exposure to ATP (10 μM) did not increase nuclear localization of NF-κB (Figs. 2B and 2D). Thus, BzATP, but not low concentrations of ATP, activate NF-κB in rabbit osteoclasts.
To investigate agonist specificity of NF-κB activation, rabbit osteoclasts were exposed to BzATP, adenosine, and a low concentration of ATP for 3 h (Fig. 3). In this series of experiments, 100 μM BzATP induced nuclear translocation of NF-κB in 22 ± 2% of osteoclasts. Treatment with 300 μM BzATP increased the percentage of osteoclasts with nuclear localization of NF-κB to levels comparable with those induced by RANKL (100 ng/ml, 30 minutes)—46% for BzATP and 41% for RANKL. In contrast, osteoclasts exposed to 10 μM adenosine (agonist at P1 receptors) or 10 μM ATP showed low levels of nuclear localization of NF-κB. Thus, the P2X7 receptor agonist BzATP, but not adenosine or a low concentration of ATP, induces nuclear localization of NF-κB in rabbit osteoclasts in a concentration-dependent manner.
Role of cytosolic calcium in activation of NF-κB
In Jurkat T cells, elevations of [Ca2+]i enhance the activation of NF-κB,(30) so we considered whether nucleotide-induced rise in [Ca2+]i is involved in the activation of NF-κB in osteoclasts. [Ca2+]i responses to BzATP were measured in fura-2-loaded rabbit osteoclasts. Repeated application of BzATP (300 μM, 1 mM Ca2+ buffer) induced a transient rise of [Ca2+]i (initial phase), followed by progressive development of increasing [Ca2+]i elevations (second phase; Fig. 4A). During the second phase, removal of extracellular Ca2+ abolished BzATP-induced [Ca2+]i elevations (Fig. 4A). In a different cell, repetitive stimulation with BzATP in Ca2+-free buffer induced the initial but not the second phase of the response (Fig. 4B), suggesting that the initial phase is caused by release of Ca2+ from intracellular stores and the second phase is caused by influx of Ca2+. Release of Ca2+ from stores indicates involvement of P2Y receptors that are coupled to activation of phospholipase C (PLC), production of inositol 1,4,5-trisphosphate, and subsequent Ca2+ release. Consistent with this interpretation, incubation of osteoclasts with the PLC inhibitor U73122 (1 μM, 10 minutes) inhibited the initial phase but not the second phase of the response to BzATP (Fig. 4C). The progressive development of the second phase of the Ca2+ response, its dependence on extracellular Ca2+, and its insensitivity to U73122 are consistent with Ca2+ influx through P2X7 receptors. These data provide evidence that BzATP induces Ca2+ responses mediated by both P2Y and P2X receptors in osteoclasts.
We next compared responses of rabbit osteoclasts to different agonists. Repetitive application of BzATP (300 μM) induced the initial and second phase Ca2+ responses described above (Fig. 5A), whereas ATP (10 μM) elicited only the initial phase that desensitized with successive applications (Fig. 5B). Adenosine (10 μM) did not induce any changes in [Ca2+]i (Fig. 5C). In summary, both BzATP and low concentrations of ATP induced an initial P2Y-mediated rise in [Ca2+]i; however, only BzATP induced the second phase Ca2+ response and activated NF-κB. These data indicate that the P2Y-mediated Ca2+ transient alone is not sufficient to activate NF-κB, suggesting the involvement of other signaling pathways.
NF-κB translocation in osteoclasts from WT and P2X7 receptor KO mice
To directly examine the role of the P2X7 receptor in activating NF-κB in osteoclasts, we compared cells isolated from WT and P2X7 receptor KO mice. Treatment of murine WT osteoclasts with BzATP (300 μM, 30 minutes), but not vehicle, increased nuclear localization of NF-κB (Figs. 6A and 6B). Importantly, KO osteoclasts treated with BzATP showed cytoplasmic localization of NF-κB (Fig. 6C). Time course studies revealed that BzATP transiently increased nuclear localization of NF-κB in WT osteoclasts to a maximum of 52 ± 5% at 30 minutes (Fig. 6D). Thus, mouse osteoclasts show much faster activation of NF-κB than rabbit osteoclasts. Treatment of KO osteoclasts with BzATP for up to 4 h did not induce significant NF-κB activation (Fig. 6E). The extent of NF-κB activation in mouse WT osteoclasts was dependent on BzATP concentration (Fig. 7). Half-maximal effects were observed at ∼100 μM BzATP, a potency consistent with that reported for P2X7-mediated responses in other systems.(31, 32) Together, these data establish that the P2X7 receptor is involved in activation of NF-κB in osteoclasts.
In some experiments, changes in osteoclast morphology in response to BzATP (300 μM) were assessed using time-lapse video microscopy. Within 5 minutes of BzATP application, WT osteoclasts showed marked retraction, followed by respreading and resumption of motility (n = 5 cells). Similar retraction was also observed in osteoclasts from P2X7 receptor KO mice, suggesting that BzATP-induced retraction is mediated by nucleotide receptors other than P2X7 (n = 2 cells).
Effect of BzATP on osteoclast viability
Under certain conditions, activation of P2X7 receptors induces formation of pores permeable to hydrophilic molecules of as large as 900 Da, which can lead to cell lysis.(11) To determine whether pore formation occurs in rabbit osteoclasts under the conditions used in this study, uptake of ethidium bromide (MW 314 Da) was assessed. Treatment of rabbit osteoclasts for up to 4 h with BzATP (100 μM) did not result in dye uptake (Figs. 8Ai and 8Aii). As a positive control, osteoclasts were treated with the membrane permeabilizing detergent digitonin (50 μg/ml) for 15 minutes; after this, the majority of cells showed ethidium bromide uptake (Fig. 8B). These observations indicate that osteoclasts remain viable for at least 4 h after activation of P2X7 receptors and that membrane integrity is not altered under conditions where we observe translocation of NF-κB.
Possible role of RANKL in P2X7 receptor-mediated NF-κB activation
Cells isolated from rabbit long bones contain marrow stromal cells and cells of the osteoblast lineage as well as osteoclasts. Both osteoclasts and osteoblasts express functional P2 receptors.(9, 20) Thus, extracellular nucleotides may act on osteoblasts to induce expression of RANKL, which can then bind to its receptor RANK on osteoclasts, activating resorption.(33) We examined the possibility that nucleotides activate NF-κB in osteoclasts indirectly through RANKL by treating some osteoclast preparations with the decoy receptor OPG (100 ng/ml) for 20 minutes before and during incubation with BzATP (100 μM) or RANKL (100 ng/ml). Treatment of osteoclasts with OPG did not significantly inhibit BzATP-induced nuclear translocation of NF-κB (Fig. 9A). In contrast, RANKL-induced nuclear translocation of NF-κB was significantly inhibited by OPG (p < 0.05; Fig. 9B), confirming its effectiveness. Thus, we conclude that BzATP induces nuclear translocation of NF-κB independently of RANKL.
In some cell types, the P2X7 receptor has been reported to play important roles in cytokine production and release, formation of multinucleated giant cells, and cell death.(19, 24, 34) In vivo, deletion of the P2X7 receptor leads to increased osteoclast number,(14) whereas in vitro studies have implicated P2X7 receptors in stimulation of osteoclast formation.(17) The signaling pathways underlying these opposing effects of P2X7 receptor activation are poorly understood. In this study, we have shown that P2X7 receptors couple to activation of NF-κB under conditions where osteoclast viability is not impaired.
A major obstacle in studying nucleotide receptor function is the lack of specific agonists and antagonists. BzATP was initially thought to be a selective agonist for the P2X7 receptor. However, it is now known that BzATP activates a number of other P2 receptors,(23) and in some cases, is even more potent at these receptors than at P2X7. Our results show that in osteoclasts, BzATP induces an initial Ca2+ transient caused by release of Ca2+ from intracellular stores, consistent with activation of P2Y receptors. Previous studies have demonstrated the presence of multiple P2Y receptors on osteoclasts, including P2Y1 and P2Y2.(27, 35, 36) Previous studies also indicate that BzATP activates P2X4 receptors in rabbit osteoclasts,(22) providing further evidence for multiple actions of BzATP. Antagonists used to characterize P2X7 receptors, such as pyridoxal-5-phosphate-6-azophenyl-2′,4′-disulfonic acid (PPADS) and oxidized ATP, lack both selectivity and potency for this receptor.(23) Therefore, genetically modified mice lacking the P2X7 receptor are useful for identifying P2X7 receptor-mediated effects. In this study, we showed activation of NF-κB in WT but not in KO mice, providing strong evidence that the effects of BzATP on NF-κB are mediated specifically by the P2X7 receptor.
Electrophysiological studies have previously provided evidence for the presence of functional P2X7 receptors on osteoclasts.(14, 21, 22) In addition, the presence of functional P2X7 receptors was recently established in osteoblasts.(14, 37) Bone cell preparations used in our studies consist of both osteoclasts and cells of the osteoblast lineage. Furthermore, ATP has been shown to induce expression of RANKL in osteoblastic cells.(33) Thus, we considered the possibility that BzATP acts on osteoblast P2 receptors to stimulate production of RANKL, which then activates NF-κB in osteoclasts. However, we found that the decoy receptor for RANKL, OPG, did not inhibit activation of NF-κB by BzATP. Thus, the P2X7 receptor activates NF-κB independently of RANKL signaling, consistent with BzATP acting directly on osteoclasts.
In rabbit osteoclasts, RANKL-mediated activation of NF-κB reached maximal levels at 30 minutes. This rapid effect is similar to that reported for activation of NF-κB by RANKL in rat and human osteoclasts.(28, 29) In contrast, BzATP-induced activation of NF-κB in rabbit osteoclasts was delayed, occurring at 3 h. In this regard, high concentrations of ATP act directly through P2X7 receptors on murine microglial cell lines to activate NF-κB with similar latency.(38) Although the latency for BzATP-induced translocation of NF-κB in rabbit osteoclasts might suggest an indirect mechanism, the time course for NF-κB activation in mouse osteoclasts was much faster. In mouse osteoclasts, maximal NF-κβ activation was observed at 30 minutes, similar to the time course for the effect of RANKL and consistent with an effect mediated directly by the P2X7 receptor. The reason for the difference in latency between mouse and rabbit osteoclasts is unclear.
Cytosolic Ca2+ plays an important role in regulating cellular activity and the pattern of Ca2+ elevation can determine the downstream cellular response. Cytosolic Ca2+ regulates a number of transcription factors including NF-κB.(30) In osteoclasts, RANKL acts to release Ca2+ from intracellular stores, which in turn, accelerates NF-κB activation.(28) Our data show that both BzATP and a low concentration of ATP cause an initial transient increase in [Ca2+]i; however, only BzATP activates NF-κB. These findings are consistent with previous results showing that transient elevation of Ca2+ alone is not sufficient to activate NF-κB in Jurkat T cells.(39) It is possible that activation of NF-κB depends on the pattern, amplitude or duration of the Ca2+ signal. It remains to be determined whether different patterns of Ca2+ elevation induced by P2X7 receptors are involved in the activation of NF-κB in osteoclasts. Alternatively, other pathways might provide complementary signals for activation of NF-κB in osteoclasts. For example, P2X7 receptors on microglial cells activated NF-κB through a mechanism dependent on reactive oxygen intermediates and proteases of the caspase family.(38)
The importance of P2X7 receptors in bone has been demonstrated in genetically modified mice lacking P2X7 receptors. P2X7 receptor KO mice exhibit excessive resorption associated with increased osteoclast surface and osteoclast number on the trabecular bone of the proximal tibial metaphysis.(14) Osteoclasts from P2X7 receptor KO mice seem morphologically normal, and the rate of osteoclast formation in cultures of bone marrow from KO mice is the same as that in marrow from WT mice. These data led the authors to suggest that P2X7 receptors promote osteoclast apoptosis.(14) In this regard, sustained activation of P2X7 receptors has been shown to induce apoptosis in several cell types, including macrophages.(40)
It is puzzling that the P2X7 receptor induces cell death yet also activates anti-apoptotic signals such as NF-κB. Paradoxical effects of P2X7 receptor activation have been noted previously. For example, transfection of lymphoid cells with the P2X7 receptor has been shown to stimulate their proliferation rather than to induce apoptosis.(41) Such observations have led to the suggestion that tonic low-level activation of P2X7 receptors has effects radically different than robust activation, which generally leads to cell death.(42) A similar situation may exist for osteoclasts, with low-level activation enhancing osteoclast formation,(17) an effect that may be mediated by the NF-κB signaling demonstrated in this study. On the other hand, robust activation would lead to osteoclast apoptosis through alternative signaling pathways. Like P2X7 receptors, TNF receptors induce opposing signals. In a number of cell types, TNFα activates both the caspase cascade promoting apoptosis and the NF-κB pathway, which suppresses the apoptotic signal.(43)
The roles of P2X7 receptors in diverse processes such as cell fusion, cytokine production and release, and apoptosis may be mediated through activation of multiple signaling pathways, including NF-κB. In vivo, nucleotides are released at sites of inflammation and in response to mechanical stimuli. Within the local bone microenvironment, ATP may then act through P2X7 receptors to regulate osteoclast number and activity.
This study was supported by the Canadian Institutes of Health Research (CIHR). JK is supported by a CIHR Doctoral Research Award. We thank Nattapon Panupinthu and Elizabeth Pruski for assistance in breeding and genotyping the P2X7 receptor-deficient mice and Dr T Michael Underhill for use of the epifluorescence microscope and imaging system. We gratefully acknowledge Dr Svetlana Komarova (The University of Western Ontario) and Dr Christopher A. Gabel (Pfizer Global Research and Development, Ann Arbor, MI, USA) for helpful advice on these studies.
- 12003 Osteoclast differentiation and activation. Nature 423:337–342., ,
- 22002 NF-κB at the crossroads of life and death. Nat Immunol 3:221–227.,
- 31997 Requirement for NF-κB in osteoclast and B-cell development. Genes Dev 11:3482–3496., , , , , , , , ,
- 42001 Nucleotide receptors: An emerging family of regulatory molecules in blood cells. Blood 97:587–600., , , , , , , ,
- 51984 Source and concentration of extracellular adenosine triphosphate during haemostasis in rats, rabbits and man. J Physiol 354:419–429.,
- 61995 Noradrenaline and ATP: Cotransmitters and neuromodulators. J Physiol Pharmacol 46:365–384.
- 72002 P2Y purinoceptors are responsible for oscillatory fluid flow-induced intracellular calcium mobilization in osteoblastic cells. J Biol Chem 277:48724–48729., , ,
- 81997 CFTR-independent ATP release from epithelial cells triggered by mechanical stimuli. Am J Physiol 272:C1058–C1066.,
- 92000 P2 purinergic receptors on osteoblasts and osteoclasts: Potential targets for drug development. Drug Dev Res 49:187–200.,
- 101998 Receptors for purines and pyrimidines. Pharmacol Rev 50:413–492.,
- 111996 The cytolytic P2Z receptor for extracellular ATP identified as a P2X receptor (P2X7). Science 272:735–738., , , ,
- 121999 Pore dilation of neuronal P2X receptor channels. Nat Neurosci 2:315–321., , , ,
- 131999 Neuronal P2X transmitter-gated cation channels change their ion selectivity in seconds. Nat Neurosci 2:322–330., , ,
- 142003 Deletion of the P2X7 nucleotide receptor reveals its regulatory roles in bone formation and resorption. Mol Endocrinol 17:1356–1367., , , , , , , , , , , , , ,
- 151999 P2X4 purinoceptors mediate an ATP-activated, non-selective cation current in rabbit osteoclasts. J Cell Sci 112:4425–4435., , , ,
- 162002 Intercellular calcium signaling occurs between human osteoblasts and osteoclasts and requires activation of osteoclast P2X7 receptors. J Biol Chem 277:7574–7580., , , , ,
- 172003 Blockade of the pore-forming P2X7 receptor inhibits formation of multinucleated human osteoclasts in vitro. Calcif Tissue Int 73:361–369., , ,
- 181998 ATP is a potent stimulator of the activation and formation of rodent osteoclasts. J Physiol 511:495–500., , , ,
- 191997 Spontaneous cell fusion in macrophage cultures expressing high levels of the P2Z/P2X7 receptor. J Cell Biol 138:697–706., , , , , , ,
- 202003 Regulation of bone resorption and formation by purines and pyrimidines. Trends Pharmacol Sci 24:290–297., ,
- 211994 Permeabilization of cells of hemopoietic origin by extracellular ATP4−: Elimination of osteoclasts, macrophages, and their precursors from isolated bone cell populations and fetal bone rudiments. Calcif Tissue Int 55:141–150., , , ,
- 222001 Activity-dependent development of P2X7 current and Ca2+ entry in rabbit osteoclasts. J Biol Chem 276:39107–39114., ,
- 232002 Molecular physiology of P2X receptors. Physiol Rev 82:1013–1067.
- 242001 Altered cytokine production in mice lacking P2X7 receptors. J Biol Chem 276:125–132., , , , , , ,
- 251992 Identification of osteopontin in isolated rabbit osteoclasts. Biochem Biophys Res Commun 186:911–917., , , , , , , , ,
- 262001 Transforming growth factor-β induces osteoclast ruffling and chemotaxis: Potential role in osteoclast recruitment. J Bone Miner Res 16:1237–1247., ,
- 272001 Activation of P2Y but not P2X4 nucleotide receptors causes elevation of [Ca2+]i in mammalian osteoclasts. Am J Physiol Cell Physiol 280:C1531–C1539., ,
- 282003 RANK ligand-induced elevation of cytosolic Ca2+ accelerates nuclear translocation of nuclear factor κ B in osteoclasts. J Biol Chem 278:8286–8293., , , ,
- 291999 Expression of functional RANK on mature rat and human osteoclasts. FEBS Lett 463:295–300., , , , , ,
- 301998 Calcium oscillations increase the efficiency and specificity of gene expression. Nature 392:933–936., ,
- 311998 Effects of antagonists at the human recombinant P2X7 receptor. Br J Pharmacol 124:1314–1320., ,
- 321997 ATP, a partial agonist for the P2Z receptor of human lymphocytes. Br J Pharmacol 122:911–917., ,
- 332002 Adenosine triphosphate stimulates human osteoclast activity via upregulation of osteoblast-expressed receptor activator of nuclear factor-κ B ligand. Bone 31:582–590., , , ,
- 341997 ATP-mediated cytotoxicity in microglial cells. Neuropharmacology 36:1295–1301., , , , , ,
- 352000 Expression of P2 receptors in bone and cultured bone cells. Bone 27:503–510., , , ,
- 361998 P2Y2 receptors are expressed by human osteoclasts of giant cell tumor but do not mediate ATP-induced bone resorption. Bone 22:195–200., , , ,
- 372001 Expression of a P2X7 receptor by a subpopulation of human osteoblasts. J Bone Miner Res 16:846–856., , ,
- 381997 Extracellular ATP activates transcription factor NF-κB through the P2Z purinoreceptor by selectively targeting NF-κB p65. J Cell Biol 139:1635–1643., , ,
- 391999 Protein kinase C and calcineurin synergize to activate IκB kinase and NF-κB in T lymphocytes. J Biol Chem 274:22923–22931., , ,
- 402000 Stress-activated protein kinase/JNK activation and apoptotic induction by the macrophage P2X7 nucleotide receptor. J Biol Chem 275:26792–26798., , ,
- 411999 Increased proliferation rate of lymphoid cells transfected with the P2X7 ATP receptor. J Biol Chem 274:33206–33208., , , , , ,
- 422002 P2X7 receptor expression in evolutive and indolent forms of chronic B lymphocytic leukemia. Blood 99:706–708., , , , , , , , ,
- 432001 Signal transduction by tumor necrosis factor and its relatives. Trends Cell Biol 11:372–377.,