Interleukin-4 Reversibly Inhibits Osteoclastogenesis via Inhibition of NF- (cid:1) B and Mitogen-activated Protein Kinase Signaling*

To define the molecular mechanism(s) by which inter-leukin (IL)-4 reversibly inhibits formation of osteoclasts (OCs) from bone marrow macrophages (BMMs), we examined the capacity of this T cell-derived cytokine to impact signals known to modulate osteoclastogenesis, which include those initiated by macrophage colony-stimulating factor (M-CSF), receptor for activation of NF- (cid:1) B ligand (RANKL), tumor necrosis factor (TNF), and IL-1. We find that although pretreatment of BMMs with IL-4 does not alter M-CSF signaling, it reversibly blocks RANKL-dependent activation of the NF- (cid:1) B, JNK, p38, and ERK signals. IL-4 also selectively inhibits TNF signaling, while enhancing that of IL-1. Contrary to previous reports, we find that MEK inhibitors dose-depend-ently inhibit OC differentiation. To identify more proximal signals mediating inhibition of OC formation by IL-4, we used mice lacking STAT6 or SHIP1, two adapter proteins that bind the IL-4 receptor. IL-4 fails to inhibit RANKL/M-CSF-induced osteoclastogenesis by BMMs derived from STAT6-, but not SHIP1-, knockout mice. Consistent with this observation, the inhibitory effects of IL-4 on RANKL-induced NF- (cid:1) B and mitogen-activated

Bone-resorbing osteoclasts, cells formed by the fusion of mononuclear progenitors of the monocyte/macrophage lineage, play an essential role in regulating bone morphogenesis and remodeling. It is now clear that two molecules produced by bone marrow mensenchymal cells are required for osteoclastogenesis, macrophage colony-stimulating factor (M-CSF) 1 and receptor for activation of nuclear factor-B (NF-B) (RANK) ligand (RANKL) (1,2). M-CSF binding to its receptor, c-Fms, on osteoclast precursors provides signals required for their survival and proliferation. RANKL, a member of the tumor necrosis factor (TNF) cytokine superfamily, exists in both soluble and membrane-bound forms, and its function is critical to the differentiation and fusion of precursors into mature osteoclasts (3)(4)(5)(6)(7). TNF␣ and IL-1 are two other cytokines that impact osteoclastogenesis while being incapable of replacing either of the its two key regulators, M-CSF and RANKL. Recent genetargeting studies show that the commitment of mononuclear precursors to mature osteoclasts involves transcription factors such as AP-1, NF-B, and p38␣ (8 -11).
Interleukin-4 (IL-4) is an immunoregulatory protein produced mainly by activated T lymphocytes. This Th2 cytokine functions as a growth and/or differentiation factor for a wide variety of bone marrow-derived hematopoietic progenitor cells (12). The hematopoietic origin of osteoclasts has led to the hypothesis that IL-4 may play an important role in the regulation of bone metabolism. Indeed, IL-4 inhibits the in vitro bone resorption stimulated by various agents, and the inhibitory effect of IL-4 on bone resorption is abolished by a monoclonal anti-IL-4 antibody (13). IL-4 also inhibits spontaneous and parathyroid hormone-related protein-stimulated osteoclast formation and modulates development of humoral hypercalcemia malignancy in intact mice (14). Furthermore, IL-4 inhibits in vitro osteoclast formation by affecting the commitment of its precursors to the osteoclast (15). Although these observations establish IL-4 as a potent anti-bone resorption factor, the mechanism(s) by which IL-4 inhibits osteoclastogenesis remain poorly understood.
The availability of RANKL permits generation of both human and murine osteoclasts from purified myeloid precursors without stromal cell/osteoblast co-culture (3,4,16). The present study was designed to investigate the molecular mechanism(s) mediating IL-4 inhibition of osteoclastogenesis. We find that IL-4 inhibits formation of osteoclasts (OCs) when murine OC precursors, namely bone marrow macrophage (BMMs), are treated with M-CSF and RANKL. The cytokine arrests multiple aspects of RANKL-induced signaling including the following: 1) preventing IB␣ phosphorylation and thus NF-B activation, and 2) blockade of the JNK, p38, and ERK MAP kinase pathways. In addition, IL-4 selectively inhibits TNF signaling and enhances that induced by IL-1 in BMMs. These inhibitory events of IL-4 on RANK signaling are reversible and require signal transducer and activator of transcription-6 (STAT6). In contrast, exposure of BMMs to IL-4 fails to impact those components of M-CSF-induced intracellular signaling relating to osteoclastogenesis, namely activation of the MAP kinase and Akt pathways.

EXPERIMENTAL PROCEDURES
Reagents-Polyclonal anti-IB␣, polyclonal anti-RANK, and polyclonal anti-TRAF6 antibodies were purchased from Santa Cruz Biotech-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: M-CSF, macrophage colony-stimulating factor; RANK, receptor for activation of nuclear factor B; NF-B, nuclear factor B; RANKL, RANK ligand; TNF, tumor necrosis factor; IL-4, interleukin-4; IL-1, interleukin-1; OC, osteoclast; STAT, signal transducer and activator of transcription; JNK, c-Jun N-terminal kinase; ERK, extracellular signal-regulated kinase; TRAF, TNF receptorassociated factor; IKK, IB kinase; TRAP, tartrate-resistant acid phosphatase; ICAM, intracellular adhesion molecule; TLR, Toll-like receptor; MAP, mitogen-activated protein; MKK, MAP kinase kinase; RT, reverse transcription; EMSA, electrophoretic mobility shift assay; SHIP, SH2-containing inositol phosphatase; PPAR␥1, peroxisome proliferator-activated receptor ␥1. nology (Santa Cruz, CA). Monoclonal anti-IKK␣ and monoclonal anti-IKK␤ antibodies were from Upstate Biotechnology, Inc. (Lake Placid, NY). All other antibodies and MEK inhibitors were from Cell Signaling Technology (Beverly, MA). Protease inhibitor mixtures were from Calbiochem. The bicinchoninic acid kit for protein determination and enhanced chemiluminescence kits were obtained from Pierce Endogen (Rockford, IL). Recombinant murine M-CSF, IL-4, TNF␣, and IL-1␣  were from R & D (Minneapolis, MN). Murine RANKL was expressed in our laboratory as described previously (17). All other chemicals were obtained from Sigma.
Cell Culture-Wild type C57BL/6 were purchased from Harlan Industries (Indianapolis, IN). STAT6-and SHIP1-deficient mice were kindly provided by Dr. James Ihle (St. Jude Children's Research Hospital, Memphis, TN) and Dr. Keith Humphries (British Columbia Cancer Agency, Vancouver, British Columbia, Canada), respectively. BMMs were isolated from whole bone marrow of 4 -6-week-old mice and incubated in tissue culture dishes at 37°C in 5% CO 2 . After 24 h in culture, the non-adherent cells were collected and layered on a Ficoll-Paque gradient, and the cells at the gradient interface were collected. The cells were replated at 65,000/cm 2 in ␣-minimal essential medium and supplemented with 10% heat-inactivated fetal bovine serum at 37°C in 5% CO 2 in the presence of recombinant mouse M-CSF (10 ng/ml). Cells were stimulated with cytokines (M-CSF, RANKL, TNF, or IL-1) on day 4 or 5. In some experiments (for MEK1/2, MKK3/6, Akt activation, and M-CSF signaling), cells were cultured in serum and M-CSF-free medium for 24 h before stimulation.
Characterization of Osteoclasts-BMMs were cultured in 48-or 96-well cell culture dishes in the presence of M-CSF (10 ng/ml) and RANKL (40 ng/ml) and other cytokines as appropriate, and medium was changed on day 3. Osteoclast-like cells were characterized by staining for tartrate-resistant acid phosphatase (TRAP). The numbers of osteoclasts were assessed by TRAP solution assay as described previously (18).
Immunoblotting-Cytokine-treated or control monolayers of BMMs were washed twice with ice-cold phosphate-buffered saline. Cells were lysed in the buffer containing 20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 3 VO 4 , 1 mM NaF, and 1ϫ protease inhibitor mixture. Fifty g of cell lysates were boiled in the presence of SDS sample buffer (0.5 M Tris-HCl, pH 6.8, 10% (w/v) SDS, 10% glycerol, 0.05% (w/v) bromphenol blue) for 5 min and subjected to electrophoresis on 8% SDS-PAGE. Proteins were transferred to nitrocellulose membranes using a semi-dry blotter (Bio-Rad, Hercules, CA) and incubated in blocking solution (5% non-fat dry milk in TBS containing 0.1% Tween 20) for 1 h to reduce nonspecific binding. Membranes were then exposed to primary antibodies overnight at 4°C, washed three times, and incubated with secondary goat anti-mouse or rabbit IgG horseradish peroxidase-conjugated antibody for 1 h. Membranes were washed extensively, and enhanced chemiluminescence detection assay was performed following the manufacturer's directions.
Electrophoretic Mobility Shift Assay-EMSA was performed as described previously (19). In brief, cells were lifted from the dish after treatment with 5 mM EDTA and 5 mM EGTA in phosphate-buffered saline, resuspended in hypotonic lysis buffer A (10 mM HEPES, pH 7.8, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol, 0.5 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, 5 g/ml leupeptin), and incubated on ice for 15 min, and then Nonidet P-40 was added to a final concentration of 6.4%. Nuclei were pelleted and resuspended in nuclear extraction buffer B (20 mM HEPES, pH 7.8, 420 mM NaCl, 1.2 mM MgCl 2 , 0.2 mM EDTA, 25% glycerol, 0.5 mM dithiothreitol, 0.5 mM 4-(2-aminoethyl) benzenesulfonyl fluoride, 5 g/ml pepstatin A, 5 g/ml leupeptin) and rotated for 30 min at 4°C. The samples were then centrifuged, and nuclear proteins in the supernatant were transferred to fresh tubes. Nuclear extracts (3 g) were incubated with an end-labeled double-stranded oligonucleotide probe containing the sequence 5Ј-AAACAGGGGGCTTTCCCTCCTC-3Ј derived from the B3 site of the TNF promoter (20). The reaction was performed in a total of 20 l of binding buffer (20 mM HEPES, pH 7.8, 100 mM NaCl, 0.5 mM dithiothreitol, 1 l of poly(dI-dC), and 10% glycerol) for 20 min at room temperature. The samples were fractionated on a 4 -20% TBE gel (NOVEX, San Diego, CA) and visualized by exposing dried gel to film.
RNA Extraction and Amplification by Reverse Transcriptase-PCR-RNA was isolated using RNeasy kits (Qiagen, Inc., Valencia, CA) and reverse-transcribed using 1 g of total RNA, 0.5 g of (dT) 15 (15,21) have reported that IL-4 inhibits osteoclast formation when whole marrow is cultured with 1,25(OH) 2 D 3 and dexamethasone. Thus, we first determined whether IL-4 inhibits osteoclastogenesis induced by M-CSF and RANKL. As seen in Fig. 1, IL-4 inhibits formation of TRAP-positive multinucleated OCs when BMMs are exposed to optimal concentrations of M-CSF and RANKL. The effect occurs at the lowest dose of the cytokine used in these studies, indicating the potent nature of the inhibition.

IL-4 Inhibits Osteoclastogenesis from Purified OC Precursors-We and others
IL-4 Targets Only Early OC Precursors-To determine at which stage IL-4 inhibits osteoclastogenesis, IL-4 was added to OC-generating cultures on days 0 -4, and TRAP staining was performed on day 5. We find that the cytokine is effective only when present during the first 2 days of culture, suggesting it targets early OC precursors (Fig. 2).
IL-4 Inhibition of OC Formation Is Reversible-To determine whether IL-4 inhibition of osteoclastogenesis is reversible, the cytokine was added at the beginning of OC-generating culture and withdrawn on day 3 or day 5 by medium change. Withdrawal of IL-4, up to 5 days after its addition, restores OC differentiation. Thus, inhibition of OC formation by IL-4 is reversible (Fig. 3). IL

Interleukin-4 and Osteoclastogenesis
activate JNK1, as reported previously (22), it activates JNK2, ERK1/2, and Akt, but not p38 or NF-B (Fig. 4). Exposure of cells to IL-4 fails to impact these aspects of M-CSF-induced signaling (Fig. 4). Since generation of osteoclasts from myeloid precursors requires the activity of two cytokines, M-CSF and RANKL (1, 2), the mechanism by which the anti-inflammatory cytokine decreases osteoclastogenesis is by blunting RANKLstimulated signaling.
IL-4 Fails to Alter Cellular Levels of RANK or TRAF6 -Activation of RANKL signaling involves binding of the cytokine to its unique receptor RANK, a member of the TNF receptor superfamily. To exclude the possibility that IL-4 impairs RANKL signaling by decreasing expression of RANK and/or TRAF6, a member of the TRAF superfamily that binds to RANK (23) and is critical for osteoclastogenesis (24,25), we performed Western blot analysis of BMMs exposed to IL-4 for 3 days. As can be seen in Fig. 5, levels of both the receptor and its immediate effector are unchanged with IL-4 treatment.
While pretreatment of IL-4 for 1 h does not significantly affect any RANKL-induced signals, exposure to the cytokine for 24 h dampens phosphorylation of IB␣, thus decreasing degradation of this protein (Fig. 6A). Furthermore, the extent of inhibition is greater when cells are exposed for 3 days (Fig. 6B). Documenting that IL-4 inhibition of osteoclast formation is reversible, 24 h after withdrawal of the cytokine, after 3 days of exposure, completely normalizes RANKL-induced IB␣ phosphorylation. Cytokine-induced phosphorylation of IB␣ is mediated by a large complex in which the active components are IKK␣ and IKK␤ (27). To exclude the possibility that IL-4 de-  creases levels of IKK␣ and/or IKK␤, lysates from RANKL-treated BMMs pre-exposed to vehicle or IL-4 were subjected to Western blot analysis with IKK-specific antibodies. As can be seen in Fig.  6C, levels of neither IKK␣ nor and IKK␤ are altered by IL-4.
The fact that IL-4 pretreatment of BMMs decreases RANKLstimulated phosphorylation of IB␣ should dampen nuclear translocation of NF-B (27). To confirm this, we performed EMSA for NF-B by using nuclear extracts from BMMs pretreated with IL-4 or vehicle and then exposed to RANKL. We find a 70% decrease in binding of nuclear levels NF-B to a consensus oligonucleotide (Fig. 7A), which should diminish transcription of NF-B-dependent genes. To test this hypothesis we used RT-PCR to measure mRNA levels of three NF-Bdependent early response genes. Consistent with our prediction, IL-4 pretreatment decreases both basal and RANKLstimulated mRNA levels of IB␣, ICAM-1, and TLR-2 (Fig. 7B), all of which are transcriptionally activated by NF-B (28 -30).
Pretreatment with IL-4 for 24 h also partially inhibits p38 activation, without significantly altering that of JNK or ERKs (Fig. 8A). Longer exposure to the cytokine (3 days) almost entirely blocks activation of JNK, p38, and ERK. Once again the alterations in MAP kinase activation, which mirror the effect of IL-4 on osteoclastogenesis, are reversed by withdrawal of the cytokine for 24 h (Fig. 8B). Turning to effectors of the SAPK/JNK ERK and p38 pathways, IL-4 decreases RANKLstimulated activation of SEK1/MKK4, which is upstream of JNK, MEK1/2, the kinases responsible for ERK activation, and MKK3/6, which enhance p38 activity (Fig. 8C).
RANKL blocks apoptosis by phosphorylating Akt/protein kinase B on serine 473 and threonine 308 through a signaling complex containing c-Src and TRAF6 (23). To determine whether IL-4 alters RANKL-induced activation of the Akt pathway, we examined phosphorylation of Akt using antibodies specific for phosphothreonine 308. As seen in Fig. 9, IL-4 alone stimulates Akt phosphorylation but does not alter subsequent RANKL-induced Akt activation in BMMs.
We next asked whether the cytokine alters the signaling pathways induced by TNF, a molecule that synergizes with RANKL in stimulating osteoclastogenesis (7), and IL-1, an inflammatory cytokine that induces mature osteoclast activation and, like RANKL, also signals via TRAF6 (31)(32)(33). We find that IL-4 selectively inhibits TNF-induced signaling, because it decreases JNK, p38, and Akt activation but fails to inhibit that of ERKs and NF-B (Fig. 10A). Interestingly, exposure of BMMs to IL-4 enhances IL-1-induced NF-B and MAP kinase activation (Fig. 10B). IL-1-induced Akt activation in BMMs was not detectable (data not shown). Thus, the results in Fig.  10, combined with those in Fig. 4, reflect the specificity of IL-4 inhibition of RANKL-induced signaling in relation to osteoclast differentiation. ERK Activation Is Required for Osteoclastogenesis-As shown in Fig. 8A and by others (26), RANKL induces ERK activation, raising the possibility that activation of the ERK pathway plays a role in osteoclastogenesis. In contrast to a previous report (26), and consistent with our observation that IL-4 inhibits RANKL-induced ERK activation, we find that both PD98059, a MEK1-specific inhibitor, and U0126, which blunts both MEK1 and MEK2, dose-dependently inhibit OC formation (Fig. 11, A and B). Withdrawal of the MEK inhibitors substantially rescues OC formation (Fig. 11C), demonstrating that, in the doses used, the inhibitors are non-toxic and ERK activation is required for osteoclastogenesis.
IL-4 Inhibits Osteoclastogenesis in a STAT6-dependent Manner-IL-4 signal transduction is initiated by binding of the cytokine to its heterodimeric membrane receptor, comprising IL-4R␣, which binds the cytokine with high affinity and the ␥C chain, a protein common to other members of the IL-4 receptor family (12). Receptor activation is followed by phosphorylation of both subunits, by Janus kinase, an event that initiates signaling by providing docking sites for adapter molecules or proteins capable of transducing signals. To date five adapters/ transducers have been identified that bind the activated IL-4 receptor complex as follows: signal transducer and activator of transcription-6 (STAT6), members of the insulin receptor substrate family, and three phosphatases including SH2-containing inositol phosphatase-1 (SHIP-1), SHP-1, and SHP-2 (12). The latter two molecules are recruited to the Fc␥RIIb immunoglobulin receptors on the immunoregulatory tyrosine-based inhibitory motif, whose related sequence is also found in the C terminus of IL-4R␣, leading to dephosphorylation of the receptors and a range of proteins (12). SHIP1, a protein expressed almost exclusively in hematopoietic cells, removes the 5Ј-phosphate group on phosphatidylinositol 3,4,5-P 3 , and thus negatively regulates phosphatidylinositol 3-kinase (12). Activation of Akt by phosphatidylinositol 3,4,5-P 3 is known to be a downstream effector of RANK signaling (23). We used BMMs from mice null for STAT6 and SHIP1 to determine which of these IL-4 receptor-bound adapters mediates the inhibitory effect of the cytokine on osteoclastogenesis and RANKL-induced signaling. Whereas IL-4 inhibits OC differentiation in wild type (Fig.  1) and SHIP1 knockout mice (Fig. 12A), the inhibitory effect is completely abrogated when the target cells lack STAT6 (Fig.  12A). Consistent with this observation, the inhibitory effects of IL-4 on RANKL-induced NF-B, JNK, p38, and ERK MAP FIG. 8. IL-4, when present for at least 3 days, reversibly inhibits MAP kinase activation pathways. BMMs were maintained with IL-4 (5 ng/ml) for 1 or 24 h (A) or 3 days (B and C) and then stimulated with RANKL as indicated. In parallel, IL-4 was withdrawn for 24 h before stimulation from cells pretreated with the cytokine for 3 days (ϩ3Ϫ), followed by RANKL stimulation (B). Cells were lysed and analyzed by Western blotting.

Interleukin-4 and Osteoclastogenesis
kinases activation are STAT6- (Fig. 12B) not SHIP1-dependent (Fig. 12C). Similarly, neither RANKL-induced NF-B activation (by EMSA) nor expression of NF-B-dependent genes IB␣, ICAM-1, and TLR-2 (by RT-PCR) are inhibited by IL-4 in STAT6-deficient BMMs (data not shown). DISCUSSION IL-4 is a T cell cytokine impacting both osteoclasts and osteoblasts. Mice overexpressing IL-4 develop systemic bone loss akin to type 2 osteoporosis, in man, characterized by attenuated remodeling activity in which osteoblast and osteoclast recruitment is suppressed (34). Consistent with this observation, ovariectomy or parathyroid hormone-related proteinstimulated bone resorption is blunted by IL-4 in vivo (14,35). The cytokine also decreases isotope release from 45 Ca-labeled fetal mouse long bones exposed osteoclastogenic factors (13). Finally, IL-4 decreases formation of osteoclasts from cultured marrow (15,21). Because these assay systems are complex, containing many cell types, these experiments did not permit careful dissection of the mechanism by which IL-4 suppresses osteoclastogenesis. The availability of recombinant M-CSF and RANKL, which together induce osteoclast formation by purified BMMs, now make it possible to explore this issue.
To identify potential molecular targets of IL-4 in osteoclast precursors, we initially examined signal transduction stimulated separately by M-CSF and RANKL, the two cytokines whose combined functions are necessary and sufficient for basal osteoclastogenesis (3,4,16). Treatment of BMMs with M-CSF results in activation of MAP kinases and Akt but not NF-B, and pretreatment of the cells with IL-4 fails to alter the degree of signal transduction. Given this finding, we focused our attention on both proximal and more distal components of RANKL/RANK signaling. In respect to proximal signaling, the cytokine does not impact expression of either RANK, the specific receptor for RANKL, or TRAF6, a RANK-binding adapter molecule required for osteoclastogenesis. By using mice lacking transcription factor STAT6 or the 5Ј-phosphatidylinositol 3,4,5-P 3 -phosphatase SHIP1, both of which transduce signals from the IL-4 receptor (12), we demonstrate that IL-4 inhibition of RANKL-induced osteoclastogenesis is STAT6-but not SHIP1dependent. This observation is in keeping with pathway specificity immediately downstream of the IL-4 receptor.
Distal events in RANK signaling include activation of the NF-B complex, the JNK, p38, and ERK MAP kinases, and the phosphatidylinositol 3-kinase/Akt axis. Each of these pathways plays a key role in osteoclast differentiation and/or function.
Thus, mice lacking the p50 and p52 subunits of NF-B or c-Fos, a member of the AP-1 family of transcription factors and a downstream target of SAPK/JNK, generate no osteoclasts (8 -10). Similarly, TNF treatment of bone marrow cell cultures from mice lacking the p38␣ gene yield fewer osteoclasts than do wild type controls (11). Finally, SHIP1-deficient mice are markedly osteopenic and have a 2-fold increase in OC number in their long bones. 2 Given the role of NF-B and phosphatidylinositol 3-kinase/Akt and the JNK, p38, and ERK MAP kinases FIG. 12. IL-4 inhibition of osteoclastogenesis and NF-B and MAP kinase activation is STAT6-dependent but not SHIP1-dependent. A, BMMs were purified from STAT6-or SHIP1-deficient mice, and IL-4 was added at the initiation of osteoclast-generating culture at 5 ng/ml. TRAP stain was performed on day 4. B and C, BMMs from wild type and STAT6-deficient mice (B) or SHIP1 heterozygous and deficient mice (C) were maintained without/with IL-4 (5 ng/ml) for 3 days, stimulated with RANKL as indicated, and lysed. Western blotting was performed with indicated antibodies.

Interleukin-4 and Osteoclastogenesis
in osteoclastogenesis, we examined the effects of IL-4 on each of these RANKL-induced signaling pathways.
Pretreatment of BMMs with IL-4 for 24 h blunts RANKLmediated NF-B and p38 MAP kinase activation, while 3 days of IL-4 treatment virtually eliminates RANKL-induced NF-B, JNK, p38, and ERK activation. The fact that altered RANKLmediated signaling, by osteoclast precursors, requires prolonged IL-4 exposure stands in contrast to the capacity of this T cell-derived cytokine to rapidly alter gene expression (36 -39). We propose that this apparent inconsistency may reflect a phenotypic change in macrophages following protracted IL-4 treatment. Support for this hypothesis comes from the observation that human and murine macrophages exposed to IL-4, in the absence of RANKL, differentiate into foreign body giant cells with an immune phenotype different from that of osteoclasts (40,41).
Our studies also reveal that IL-4 attenuates ERK, JNK, and p38 activity by inhibiting their proximal effectors SEK1/ MKK4, MEK1/2, and MKK3/6, respectively. However, the means by which IL-4 decreases NF-B activation are presently unclear, because we observe no changes in levels of key elements of the immediate upstream kinase complex, namely IKK␣ and IKK␤.
Osteoclastogenesis is increased in a range of pathophysiological conditions, such as postmenopausal osteoporosis, implant osteolysis, alveolar bone loss, and rheumatoid arthritis (42)(43)(44)(45)(46). In these circumstances the presence of molecules such as TNF-␣ and IL-1, alone or in combination, contribute to the increased numbers of osteoclasts seen at sites of bone resorption. We and others have examined the mechanisms by which these molecules impact formation and/or activity of osteoclasts (7,31,32). RANKL and RANK are members of a signaling superfamily, of which the archetypal components are TNF-␣ and its receptors, p55 and p75 (47). In contrast to other reports (52,53), we found recently that TNF-␣ alone is completely inactive in stimulating osteoclast formation both in vivo and in vitro. Rather, it is a potent synergistic agonist of RANKL. Furthermore, this synergy involves enhanced activation of the same intracellular signaling pathways stimulated by RANKL (7). The differential capacity of IL-4 to block activation of signals by the closely related molecules, TNF and RANKL, provides evidence for receptor-mediated signaling specificity. Finally, contrary to the report that IL-4 down-regulates both the p55 and p75 TNF receptors in transformed myeloid cells (48), thereby totally abrogating TNF-induced signaling, we find that in primary BMMs the T cell-derived cytokine inhibits some, but not all, components of TNF-induced signaling.
IL-1 stimulates fusion and survival of committed osteoclasts (31, 32) but is incapable of replacing RANKL in their generation. Although RANKL and IL-1 initiate signaling by binding to different receptor family members, they share the same adapter signaling molecule, TRAF6, and stimulate similar signaling pathways such as NF-B and MAP kinases (33,49). While the mechanism remains unclear, our observations reveal that exposure of BMMs to IL-4 enhances IL-1-induced signaling. Since mature osteoclasts activate NF-B and MAP kinase in response to IL-1 (32,50), it seems possible that only more differentiated BMMs are capable of responding to the cytokine.
IL-4 was recently reported to decrease osteoclast number by activating peroxisome proliferator-activated receptor ␥1 (PPAR␥1) (51), in osteoclast precursors. It seems unlikely, however, that IL-4 regulation of PPAR␥1 fully explains the capacity of the cytokine to inhibit RANKL-induced osteoclast differentiation. For example, while IL-4 does not stimulate PPAR␥1 transcripts in RAW264.7 cells, the cytokine continues to inhibit their osteoclastogenic response to RANKL. In addition, a PPAR␥1 antagonist neither reverses the inhibitory effect of low levels of IL-4 on osteoclast formation nor blocks IL-4 suppression of RANKL-induced NF-B activation. In summary, we provide the first evidence that IL-4 specifically decreases basal formation of osteoclasts by inhibiting RANKL induction of both the NF-B complex and MAP kinases.