Drs. Hirayama, Dai, and Abbas contributed equally to this work.
Inhibition of inflammatory bone erosion by constitutively active STAT-6 through blockade of JNK and NF-κB activation
Version of Record online: 2 SEP 2005
Copyright © 2005 by the American College of Rheumatology
Arthritis & Rheumatism
Volume 52, Issue 9, pages 2719–2729, September 2005
How to Cite
Hirayama, T., Dai, S., Abbas, S., Yamanaka, Y. and Abu-Amer, Y. (2005), Inhibition of inflammatory bone erosion by constitutively active STAT-6 through blockade of JNK and NF-κB activation. Arthritis & Rheumatism, 52: 2719–2729. doi: 10.1002/art.21286
- Issue online: 2 SEP 2005
- Version of Record online: 2 SEP 2005
- Manuscript Accepted: 17 JUN 2005
- Manuscript Received: 18 MAR 2005
- NIH. Grant Numbers: R01-DE-13754, R01-AR-47443, R01-AR-049192
- Arthritis Foundation Arthritis Investigator award
- Shriners Hospital for Children
NF-κB and JNK signaling pathways play key roles in the pathogenesis of inflammatory arthritis. Both factors are also activated in response to osteoclastogenic factors, such as RANKL and tumor necrosis factor α. Inflammatory arthritis and bone erosion subside in the presence of antiinflammatory cytokines such as interleukin-4 (IL-4). We have previously shown that IL-4 inhibits osteoclastogenesis in vitro through inhibition of NF-κB and JNK activation in a STAT-6–dependent manner. This study was undertaken to investigate the potential of constitutively active STAT-6 to arrest the activation of NF-κB and JNK and to subsequently ameliorate the bone erosion associated with inflammatory arthritis in mice.
Inflammatory arthritis was induced in wild-type and STAT-6–null mice by intraperitoneal injection of arthritis-eliciting serum derived from K/B×N mice. Bone erosion was assessed in the joints by histologic and immunostaining techniques. Cell-permeable Tat-STAT-6 fusion proteins were administered intraperitoneally. Cells were isolated from bone marrow and from joints for the JNK assay, the DNA-binding assays (electrophoretic mobility shift assays), and for in vitro osteoclastogenesis.
Activation of NF-κB and JNK in vivo was increased in extracts of cells retrieved from the joints of arthritic mice. Cell-permeable, constitutively active STAT-6 (i.e., STAT-6-VT) was effective in blocking NF-κB and JNK activation in RANKL-treated osteoclast progenitors. More importantly, STAT-6-VT protein significantly inhibited the in vivo activation of NF-κB and JNK, attenuated osteoclast recruitment in the inflamed joints, and decreased bone destruction.
Our findings indicate that the administration of STAT-6-VT presents a novel approach to the alleviation of bone erosion in inflammatory arthritis.
Inflammatory arthritis is complicated by osteolytic responses that weaken the structure of bone and, in severe cases, lead to disability. These responses require the recruitment and activation of several types of cells to the site of inflammation, most notably, bone-resorbing osteoclasts and T lymphocytes (1–3). The loss of bone and its architectural integrity are the result of massive recruitment of osteoclasts to sites of inflammation (4–6), as evidenced by the fact that expression of RANK is required for bone erosion associated with arthritis (7, 8).
Osteoclasts are hematopoietic cells originating from the monocyte/macrophage lineage. Differentiation of these cells is primarily controlled by the stromal cell–derived and T cell–derived proteins RANKL and macrophage colony-stimulating factor (M-CSF) (9). Binding of RANKL to its receptor (RANK) prompts the differentiation of bone marrow macrophages into multinucleated mature osteoclasts through activation of a key signaling cascade that culminates with activation of the NF-κB and JNK pathways (10–12). The central role of these 2 pathways in osteoclastogenesis and inflammatory osteolysis has been described. In this regard, genetic manipulation studies in mice have shown that certain NF-κB subunits and JNK-1 are essential for osteoclastogenesis and that they regulate bone resorption and modulate inflammatory bone responses (6, 13, 14). In a serum-transfer model of inflammatory arthritis, activation of NF-κB and JNK appears to be central to the progression of focal bone erosion in the inflamed joints of the extremities of arthritic mice. In those studies, direct administration of a dominant-negative form of IκB, which lacks amino-terminal phosphorylation sites, was shown to significantly block NF-κB activation and attenuate osteoclast recruitment and bone erosion (15, 16).
Induction of the JNK pathway by RANKL and proinflammatory cytokines indicates a vital role of JNK in inflammatory osteolysis. Studies have established that dominant-negative forms of various MAP kinases and selective inhibitors of the MAP kinase pathways inhibit osteoclastogenesis and/or reduce osteoclast survival (14, 17–20). Specifically, genetic studies with JNK-1 provide evidence that the expression and phosphorylation of c-Jun modulate osteoclastogenesis (14). In light of this information, it is not surprising that the effect of antiinflammatory immune cell–derived cytokines, such as interleukin-4 (IL-4) and interferon-γ (IFNγ), which are also known for their potent antiosteoclastogenic activity, target the RANK/RANKL pathway. Several studies have outlined unique pathways by which cells and products of the immune cells regulate osteoclastogenesis and dampen progressive bone resorption (21, 22). Whereas IFNγ targets upstream modulators of the RANK pathway, such as tumor necrosis factor receptor−associated factor 6 (TRAF6), causing its degradation (23), the T cell–derived cytokine IL-4 is sufficient to block osteoclastogenesis by interfering with key osteoclast signaling pathways, including NF-κB and MAP kinases. This process depends entirely upon activation of the IL-4–responsive STAT, namely, STAT-6 (22).
IL-4 ligation to its receptor activates protein tyrosine kinases of the JAK family, which, in turn, leads to phosphorylation, dimerization, and nuclear translocation of the transcription factor STAT-6, one of the signal transducers and activators of transcription (24–27). In the nucleus, STAT-6 binds to specific regions of promoters of target genes. Several studies have indicated that STAT-6 is an essential component of Th2 lymphocyte responses and is critical for many IL-4 functions (28). STAT-6–null mice are viable and are not distinguishable from their wild-type counterparts. However, they fail to mount an IL-4–induced antiosteoclastogenic response, as we have established previously (22).
To further examine the mechanisms by which IL-4/STAT-6 affects osteoclastogenesis and arthritic bone erosion, we examined the effect of a constitutively active STAT-6 on osteoclastogenesis and bone erosion in a serum-transfer model of inflammatory arthritis. Our findings indicate that the JNK pathway is significantly activated in RANKL-stimulated osteoclast progenitors and in histologic sections derived from the joints of arthritic mice. Furthermore, activation of JNK coincides with massive recruitment and differentiation of osteoclasts. More interestingly, we found that the constitutively active form of STAT-6 protein significantly reduces the activation of JNK, inhibits osteoclast recruitment, and alleviates bone erosion.
MATERIALS AND METHODS
Antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Recombinant murine IL-4 and M-CSF were purchased from R&D Systems (Minneapolis, MN). RANKL was purchased from PeproTech (Rocky Hill, NJ). The enhanced chemiluminescence (ECL) kit was obtained from Pierce (Rockford, IL). All other chemicals were obtained from Sigma (St. Louis, MO).
STAT-6–knockout mice and BALB/c control mice were purchased from The Jackson Laboratory (Bar Harbor, ME). KRN-TCR (K/B) transgenic mice on a B6 background were kindly provided by Drs. D. Mathis and C. Benoist (Harvard Medical School, Boston, MA). K/B×N mice were generated by breeding KRN-TCR mice with NOD mice (Taconic, Germantown, NY). All mice were housed in a controlled barrier facility at Washington University, and experiments were performed in compliance with the regulations of the Institutional Animal Care and Use Committee.
Administration of arthritogenic serum.
Serum was obtained from K/B×N mice (6–12 weeks old), pooled, and stored in aliquots at −70°C until used. A single dose of 0.15 ml of serum per mouse was selected as being optimal for the induction of arthritis in 100% of the mice (16). BALB/c mice ages 3–4 weeks were injected intraperitoneally with either phosphate buffered saline (PBS), serum (0.2 ml), Tat-TK fusion protein, or Tat-STAT-6 fusion protein (fusion proteins administered at 1 mg/kg of body weight; established as the optimal dose) alone or with a combination of serum and each Tat fusion protein. Tat fusion proteins were injected twice (first and third days of the experiment) over the entire course of each in vivo experiment. Five mice were evaluated for each experimental condition, and the experiments were repeated at least 3 times.
Osteoclast precursors in the form of bone marrow macrophages were isolated from whole bone marrow of 4–6-week-old mice and incubated in tissue culture plates (at 37°C in an atmosphere of 5% CO2) in the presence of 10 ng/ml of M-CSF (29). After 24 hours in culture, nonadherent cells were collected and layered on a Ficoll-Hypaque gradient. Cells at the gradient interface were then collected and plated in α-minimum essential medium, supplemented with 10% heat-inactivated fetal bovine serum (at 56°C for 30 minutes) in the presence of 10 ng/ml of M-CSF, and then plated (at 37°C in an atmosphere of 5% CO2) according to each experimental condition.
Immunopurification of osteoclast precursors.
Ficoll-Hypaque gradient–purified macrophages were further purified with CD11b (PharMingen, San Diego, CA) antibody and with Dynabeads (Dynal, New Hyde Park, NY) according to the manufacturers' instructions.
Generation of osteoclasts.
Purified bone marrow macrophages were cultured at 1 × 106 cells/ml in the presence of 10 ng/ml of M-CSF and 20 ng/ml of RANKL for 4 days. On day 2, cultures were supplemented with M-CSF and RANKL. On day 4, osteoclast cultures were fixed and stained for expression of tartrate-resistant acid phosphatase (TRAP).
Knee and ankle joints were excised, and the skin and soft tissues were removed. Joints were immediately snap-frozen in liquid nitrogen for further analysis. For cellular and nuclear isolation, extracts were prepared from mouse paws according to the method described by Han et al (30). Briefly, whole mouse joints and paws were flash-frozen in liquid nitrogen, crushed using a stainless steel mortar and pestle, then pulverized using a CertiPrep Freezer Mill 6570 (Spex CertiPrep; Metuchen, NJ) under liquid nitrogen to achieve a fine powder of ∼2 × 1 mm. Equal amounts were then subjected to lysis or nuclear isolation as described elsewhere (31).
For histologic assessments, intact limbs were preserved in 10% buffered formalin (24 hours), deskinned, and subjected to a decalcification process using 10% EDTA, pH 7.0, for 7 days, with gentle rocking. Solution was replaced daily. Decalcified bones were then dehydrated in graded alcohol, cleared through xylene, and embedded in paraffin. Paraffin blocks were sectioned longitudinally. Five-micron sections were then stained with hematoxylin and eosin or were stained histochemically for TRAP to identify osteoclasts.
Histologic sections embedded on slides were subjected to immunostaining using primary antibodies, as indicated for each experiment. Antibodies were detected with horseradish peroxidase–conjugated secondary antibody according to standard procedures.
Crude cell lysates were boiled for 5 minutes in the presence of 2× sodium dodecyl sulfate (SDS) sample buffer (0.5M Tris HCl, pH 6.8, 10% [weight/volume] SDS, 10% glycerol, 0.05% [w/v] bromphenol blue, and distilled water) and subjected to electrophoresis on 10% SDS–polyacrylamide gels (32). Proteins were transferred to nitrocellulose membranes using a semidry blotter (Bio-Rad, Richmond, CA) and incubated in blocking solution (10% skim milk prepared in PBS containing 0.05% Tween 20) to reduce nonspecific binding. Membranes were washed with PBS–Tween buffer and exposed to primary antibodies (1 hour at room temperature), washed again 4 times, and incubated with the respective secondary horseradish peroxidase–conjugated antibodies (1 hour at room temperature). Membranes were washed extensively (5 times for 15 minutes each), and an ECL detection assay was performed according to the manufacturer's directions.
Electrophoretic mobility shift assay.
Nuclear fractions were prepared as previously described (31, 33). Briefly, monolayers of osteoclast precursors grown in a 100-mm2 tissue culture dish were washed twice with ice-cold PBS. Cells were lifted from the dish by treatment with 5 mM EDTA and 5 mM EGTA in PBS. Cells were then resuspended in hypotonic lysis buffer A (10 mM HEPES, pH 7.8, 10 mM KCl, 1.5 mM MgCl2, 0.5 mM dithiothreitol, 0.5 mM 4-[2-aminoethyl]benzenesulfonyl fluoride [AEBSF], and 5 μg/ml of leupeptin) and incubated on ice for 15 minutes. Nonidet P40 was added to a final concentration of 0.64%. Nuclei were pelleted and the cytosolic fraction was carefully removed.
The nuclei were then resuspended in nuclear extraction buffer B (20 mM HEPES, pH 7.8, 420 mM NaCl, 1.2 mM MgCl2, 0.2 mM EDTA 25% glycerol, 0.5 mM dithiothreitol, 0.5 mM AEBSF, 5 μg/ml of pepstatin A, and 5 μg/ml of leupeptin), vortexed for 30 seconds, and then rotated for 30 minutes at 4°C. The samples were then centrifuged, the nuclear proteins in the supernatant were transferred to fresh tubes, and the protein content was measured using a standard bicinchoninic acid assay kit (Pierce).
Nuclear extracts (10 μg) were incubated with an end-labeled double-stranded oligonucleotide probe containing the sequence 5′-AAA-CAG-GGG-GCT-TTC-CCT-CCT-C-3′ (34) derived from the κB3 site of the tumor necrosis factor α (TNFα) promoter. The reaction was performed in a total of 20 μl of binding buffer (20 mM HEPES, pH 7.8, 100 mM NaCl, 0.5 mM dithiothreitol, 1 μg of poly[dI-dC], and 10% glycerol) for 30 minutes at room temperature. Samples were then fractionated on a 4% polyacrylamide gel and were visualized by exposing the dried gel to film.
The pTat construct and protein coupling.
STAT-6 constructs were cloned into the pTat-HA bacterial expression vector previously described by Nagahara and Dowdy (35–37), which contains a 6-histidine tag for purification, a hemagglutinin tag for detection, followed by the Tat transduction domain, and finally, the STAT-6 coding sequence. To generate the VTAA point mutations, the following primers were used: Stat6-VT_F, 5′-GGCTTTATTAGTAAGCAATATGCCGCTAGCCTTCTCCTCAATGAGC-3′ (forward) and STAT-6-VT_R, 5′-GCTCATTGAGGAGAAGGCTAGCGGCATATTGCTTACTAATAAAGCC-3′ (reverse). Similarly, tyrosine 641 was mutated into phenylalanine using the primers STAT6Y641F_F, 5′-GGAAGGACGGGAGGGGTTTTGTCTCTACTACTA-TCAAGATG-3′ (forward) and STAT6Y641F_R, 5′-CATCTTGATAGTAGTAGAGACAAAACCCCTCCCGTCCTTCC-3′ (reverse). The resultant plasmids, pTat-STAT-6-WT, pTat-STAT-6-VT, and pTat-STAT-6-Y641F (pTat-STAT-6-Y/F) were sequenced to confirm the fusion frame and were then used to transform BL21 Escherichia coli (Pharmacia, Piscataway, NJ) to produce the Tat-fused STAT-6 proteins. Protein expression was induced with IPTG.
Following 2–3 hours of induction, the cells were sonicated in 8M urea, and the Tat-coupled STAT-6 proteins were purified on a nickel–Sepharose column (Qiagen, Chatsworth, CA), and then applied to an ion-exchange column (Mono Q) in 4M urea. Eluted proteins were desalted, detoxified, and quantified. Cellular transduction of these and other Tat-fused proteins has been documented elsewhere (22, 35–37). The coupled Tat-STAT-6 proteins were then added to cultured cells without the aid of transfection agents (22).
Extracts prepared from the various cells and tissues were suspended in lysis buffer containing 40 mM Tris HCl, pH 8.0, 500 mM NaCl, 0.1% Nonidet P-40, 6 mM EDTA, 6 mM EGTA, 5 mM β-glycerophosphate, 5 mM NaF, 1 mM Na3VO4, pH 10.0, and protease inhibitor (Roche Molecular Biochemicals, Indianapolis, IN). The cell lysates (500 μg) were subjected to immunoprecipitation with the relevant antibody (anti-JNK or isotype-matched IgG) and γ-bind Sepharose (25 μl). After the immunoprecipitates were washed, kinase assays were performed at 30°C for 30 minutes, using buffer containing 50 mM Tris HCl, pH 8.0, 100 mM NaCl, 10 mM MgCl2, 1 mM dithiothreitol, 10 μM ATP, 5 mM β-glycerophosphate, 5 mM NaF, 1 mM Na3VO4, pH 10.0, 5 μCi of γ32P-ATP, and 5 μg of c-Jun. Phosphorylation of c-Jun was then detected by immunoblotting.
Results are reported as the mean ± SEM for each group. Student's t-test was used to analyze statistical differences between the control group and the treated group. Significant P values are indicated where appropriate.
Attenuation of osteoclastogenesis by inhibition of JNK pathways.
Similar to NF-κB, the JNK pathway is also essential for osteoclastogenesis and mediates inflammatory responses (14, 38, 39). We found that the pharmacologic inhibitor of JNK, SP600125, blocks JNK activation and inhibits osteoclastogenesis in vitro (Figure 1). Inhibition of osteoclastogenesis was dose-dependent (Figure 1b, panels A–C) and reversible (Figure 1b, panels E and F), thereby excluding cell toxicity and providing evidence that the JNK pathway is an important mediator of osteoclastogenesis. SP600125 was JNK-specific, since it did not affect the activation of other MAP kinase pathways, including p38 and ERK-1/2 (results not shown).
Inhibition of NF-κB and JNK activation in vitro by STAT-6-VT, a constitutively active form of STAT-6.
In recent studies, we have established that STAT-6 is essential for IL-4 inhibition of NF-κB, JNK, and osteoclastogenesis, and that administration of exogenously expressed STAT-6 protein restores the IL-4 antiosteoclastogenic effect in STAT-6–null cells (22). In the current study, we generated an active form of STAT-6 in which 2 amino acids (at positions 547/548VT) were substituted for alanine residues, as described in Materials and Methods. This form of STAT-6 has been shown to bind DNA and to activate transcription in the absence of IL-4 stimulation (40).
Consistent with this notion, we found that whereas IL-4 failed to inhibit RANKL induction of NF-κB DNA-binding activity in STAT-6–null cells, the addition of STAT-6-VT protein blocked such activation of the transcription factor to levels indistinguishable from those of the control (Figure 2a; compare lane 4 with lane 1). This level of inhibition surpassed the moderate level seen in the presence of the STAT-6-WT (wild-type) protein. In contrast, a tyrosine-mutated form of STAT-6, in which tyrosine 641 was substituted with phenylalanine (STAT-6-Y/F), failed to affect NF-κB activation (Figure 2a). These observations indicate that STAT-6-VT is a superior inhibitor of RANKL-induced NF-κB activity, whereas tyrosine 641–mutated STAT-6 is virtually inactive.
Similar studies were conducted to examine the effect of the various forms of STAT-6 on JNK phosphorylation. We found that whereas STAT-6-Y/F failed to block, and STAT-6-WT only moderately inhibited RANKL-induced phosphorylation of JNK, STAT-6-VT significantly blocked the phosphorylation of the MAP kinase (Figure 2b). Together, these data indicate that STAT-6-VT is a potent inhibitor of NF-κB and JNK activation pathways in the absence of exogenous induction (of STAT-6) by IL-4.
STAT-6-VT, a potent inhibitor of osteoclastogenesis in vitro.
Having established that STAT-6-VT is a potent inhibitor of NF-κB and JNK pathways, both of which are essential for osteoclast formation, we examined whether this inhibition led to the arrest of osteoclastogenesis. To this end, STAT-6–null osteoclast precursors were placed with RANKL and M-CSF in the absence or presence of STAT-6-VT, STAT-6-WT, STAT-6-Y/F, or IL-4. Wild-type and STAT-6–null cells were treated with IL-4 as controls for osteoclast inhibition or the lack of it, respectively. As shown in Figure 3a, IL-4 inhibited osteoclast formation in wild-type (Figure 3a, panel C) and failed to do so in STAT-6–null cell cultures (Figure 3a, panel B), as expected. Furthermore, consistent with its effect on NF-κB and JNK activation, STAT-6-VT significantly blocked (Figure 3a, panel E), STAT-6-WT moderately inhibited (Figure 3a, panel D), and STAT-6-Y/F failed to inhibit (Figure 3a, panel F) osteoclastogenesis in STAT-6–null cells. The differences between the various treatments are depicted quantitatively in Figure 3b.
To further validate the inhibition of osteoclast activity, we performed a resorption assay in which osteoclasts were differentiated over calcium phosphate–coated plates in the presence or absence of IL-4 or different forms of STAT-6 (Figure 3c). Consistent with the findings described above, STAT-6-VT significantly inhibited bone resorption compared with RANKL treatment alone (Figure 3c, panels B and E), as evidenced by a lack of resorbed surfaces in the latter condition. Moreover, STAT-6-WT significantly, but incompletely, inhibited osteoclast activity (Figure 3c, panel D), whereas STAT-6-Y/F failed to affect the resorptive activity of osteoclasts (Figure 3c, panel F). Thus, STAT-6-VT appears to act as a potent inhibitor of osteoclastogenesis in vitro through direct targeting of osteoclast precursors. In contrast, the partial inhibition of osteoclastogenesis by wild-type STAT-6 is due to a lack of optimal activation in the absence of IL-4, whereas the failure of tyrosine-mutated STAT-6 to exert an osteoclast inhibitory effect is reflective of its inactive state.
Attenuation of NF-κB and JNK activation in joint synovium from mice with inflammatory arthritis and alleviation of bone erosion by STAT-6-VT.
Both the transcription factor NF-κB and the MAP kinase JNK have been implicated in the pathogenesis of inflammatory arthritis and osteoclastogenesis. Furthermore, bone erosion that complicates inflammatory arthritis is mediated by heightened activation of osteoclasts. Thus, we reasoned that STAT-6-VT may inhibit in vivo the activation of these pathways and may possibly attenuate the subsequent osteoclastogenesis, without bone erosion. To address this possibility, we used a serum-transfer model of inflammatory arthritis. Crude cell populations were extracted from the joint synovial tissue of control and arthritic mice and from their STAT-6–treated counterparts. Nuclear extracts and crude lysates were prepared to examine NF-κB activation and JNK phosphorylation, respectively.
Consistent with the in vitro findings, STAT-6-VT significantly inhibited, whereas STAT-6-WT moderately inhibited, the NF-κB activation observed in arthritic mice (Figure 4a). Similar results were obtained for JNK activation (Figure 4b). Furthermore, immunostaining of histologic sections for the presence of phospho-JNK clearly indicated that the elevated levels of phospho-JNK observed in arthritic mice (Figure 4c, panel B) were significantly inhibited by the in vivo administration of STAT-6-VT (Figure 4c, panel F) and STAT-6-WT (Figure 4c, panel D) proteins. In contrast, STAT-6-Y/F failed to reduce JNK phosphorylation in arthritic joints (Figure 4c, panel H). Staining of sections from control and arthritic mouse joints with control IgG are shown for comparison (Figure 4c, panels I and J).
The fact that STAT-6-VT mediates the attenuation of in vitro osteoclastogenesis through inhibition of NF-κB and JNK activation taken together with the evidence of inhibition of NF-κB and JNK activation in arthritic joint tissues suggests that the STAT-6-VT protein may abrogate the in vivo bone erosion seen in this model. Therefore, histologic sections were obtained from the various treatment groups and were stained for TRAP reactivity to highlight osteoclasts. There were ample intensely stained osteoclasts in sections of bone from arthritic mice as compared with the normal baseline osteoclast content in sections from the control mice (Figure 5, compare panel A with panel E). Interestingly, STAT-6-VT protein strongly inhibited osteoclast recruitment in arthritic mice (Figure 5, compare panel G with panel E). Consistent with our previous observations, STAT-6-WT partially blocked osteoclast recruitment, and STAT-6-Y/F failed to do so. Closer examination of these sections further revealed that STAT-6-VT was superior to other STAT-6 proteins in protecting the integrity of the bones, significantly reducing focal bone erosion. Together, these findings show that STAT-6-VT is a strong inhibitor of NF-κB and phospho-JNK and of inflammatory osteoclast recruitment in vivo, and significantly alleviates the focal bone erosion associated with inflammatory arthritis.
As is the case with all bone lesions, osteoclasts are responsible for focal bone erosion and destruction of the bone matrix associated with inflammatory arthritis. This is evidenced by recent studies indicating that the in vivo absence of osteoclasts in RANK-null mice protects them from inflammatory bone erosion (8, 41). Numerous previous studies have implicated NF-κB and JNK in the pathogenesis of inflammatory arthritis (6, 7, 13, 39, 42). These factors are believed to facilitate the cartilage and bone destruction associated with this disease. Osteoclasts are the principal bone-resorbing cells, and they require functional NF-κB and JNK pathways for their differentiation and function (43, 44). These observations led us in previous studies to disable the NF-κB pathway using various IκBα dominant-negative mutants and inactivation of the upstream IKK complex signal. The use of these approaches enabled the successful inhibition of osteoclastogenesis in vitro and, more interestingly, the blockade of focal bone erosion in the serum-transfer model of inflammatory arthritis (15, 16, 45). Successful inhibition of osteoclastogenesis was achieved in the present studies by direct blockade of JNK activation in osteoclast precursor cultures, using the synthetic inhibitor SP600125 (46).
To better understand the molecular mechanisms of antiinflammatory intervention, we have shown in previous studies that the antiinflammatory cytokine IL-4 is a potent inhibitor of osteoclastogenesis, which dampens the activation of the NF-κB and JNK pathways in a STAT-6–dependent manner. In fact, a bacterially expressed Tat-fused STAT-6 protein was sufficient to block NF-κB activation and restore IL-4 inhibition of osteoclastogenesis by STAT-6–null osteoclast progenitors (22). This constricted the antiosteoclastic effect of IL-4 to the induction of STAT-6, thus providing a specific target molecule and sparing other pleiotropic effects of IL-4. However, the need for the presence of IL-4 to achieve complete induction of STAT-6 hampered our efforts to restrict other undesired effects of the cytokine.
Traditional STAT-6 activation entails IL-4–dependent tyrosine phosphorylation of the transcription factor. However, an IL-4–independent, active form of STAT-6 has recently been described (40). This form, called STAT-6-VT, harbors 2 amino acid changes in the SH2 domain, undergoes tyrosine phosphorylation, binds DNA, and activates transcription in the absence of IL-4. Our studies show that unlike the wild-type or tyrosine 641–mutated forms of STAT-6, the STAT-6-VT protein readily inhibited NF-κB and JNK activation and arrested osteoclastogenesis in the STAT-6–null background without the need for IL-4. Moderate inhibition by STAT-6-WT protein is attributed to hypophosphorylation or dephosphorylation of the protein in the absence of IL-4, a scenario that does not hold with the STAT-6-VT protein because of conformational changes and presumed inaccessibility of phosphatases (40). This is further supported by the observation that tyrosine 641–mutated STAT-6 did not inhibit NF-κB and JNK activation or block osteoclastogenesis.
Assertion of the biologic function of STAT-6-VT came from earlier studies using forced expression of the CD2:STAT-6-VT transgene in mice. However, expression of STAT-6-VT in these mice was under the control of the CD2 locus, thus restricting its expression to the lymphoid population and providing no insight into its role in osteoclastogenesis. Nevertheless, these studies confirmed the notion that STAT-6-VT is a sufficient substitute for limited IL-4 functions, while not affecting other IL-4–transmitted signals that are deemed to be STAT-6–independent (47).
The highlight of our studies is the ability of STAT-6-VT to prevent activation of the NF-κB and JNK pathways in inflammatory arthritis, which further translates into attenuation of osteoclastogenesis and bone erosion in the joints of arthritic mice. The mechanisms of joint destruction involve activation of the NF-κB and JNK pathways. Selective targeting of either pathway, as we and others have done, shows significant sparing effects on joint destruction (16, 30). In this regard, we have shown that levels of pro-osteoclastogenic and inflammatory cytokines are elevated in the synovial tissue of mice with inflammatory arthritis and coincide with increased activation of NF-κB (16), c-Fos (results not shown), and JNK (as shown in the present study). These findings are consistent with the report that high levels of JNK-1 expression were found in synovial tissue from patients with rheumatoid arthritis (39). More interestingly, administration of a specific JNK inhibitor in an animal model of rheumatoid arthritis was shown to protect the animals from bone destruction (48), suggesting that JNK is a key regulator of inflammatory osteolysis.
Although the precise mechanism by which STAT-6-VT impedes these pathways remains unclear, it is plausible to speculate that it inhibits a common upstream activator situated at the upstream bifurcation point of the MAP kinase and NF-κB activation pathways. This idea is not uncommon, since both signaling events are transmitted by RANK and TNF receptor activation, converge at the level of TRAF6 and TRAF2 (10, 12), and are regulated by upstream MEKKs and IKKs (49–51). Specifically, previous studies have shown that cytokine-mediated degradation of TRAF6 correlates with decreased NF-κB and JNK activation, leading to the suppression of osteoclastogenesis by IFNγ, another T cell–derived factor (23). Whether STAT-6-VT regulates TRAFs in general and TRAF6 in particular remains to be elucidated. Alternatively, STAT-6-VT may be acting at multiple levels to inhibit either pathway separately.
Recent evidence indicates that c-Jun cooperates with nuclear factor of activated T cells (NF-AT) protein to mediate osteoclastogenesis (52, 53). The paradigm emerging from these advances indicates that RANKL induces and facilitates the formation of ternary complexes, including activator protein 1/c-Jun and NF-AT proteins, to transactivate target genes. Moreover, expression of dominant-negative forms of either factor is sufficient to attenuate osteoclastogenesis. Thus, it is possible that STAT-6 inhibition of osteoclastogenesis and bone erosion may be the outcome of inhibiting NF-AT/AP-1 complex formation and DNA binding. Further studies will be required to investigate this possibility.
In summary, this study provides new evidence that constitutively active STAT-6, in the form of STAT-6-VT, is a potent inhibitor of the NF-κB and JNK signaling pathways, both of which are extensively involved in osteoclastogenesis and inflammatory responses that occur in inflammatory arthritis. These observations position NF-κB and JNK as primary mediators of joint destruction in inflammatory arthritis, and as such, they are viable targets for therapeutic intervention.