Suppressors of Cytokine Signaling-1 and -3 Regulate Osteoclastogenesis in the Presence of Inflammatory Cytokines

This information is current as Masanobu Ohishi, Yumiko Matsumura, Daisuke Aki, of September 27, 2021. Ryuichi Mashima, Koji Taniguchi, Takashi Kobayashi, Toshio Kukita, Yukihide Iwamoto and Akihiko Yoshimura J Immunol 2005; 174:3024-3031; ; doi: 10.4049/jimmunol.174.5.3024

http://www.jimmunol.org/content/174/5/3024 Downloaded from

References This article cites 35 articles, 11 of which you can access for free at: http://www.jimmunol.org/content/174/5/3024.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 27, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2005 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Suppressors of Cytokine Signaling-1 and -3 Regulate Osteoclastogenesis in the Presence of Inﬂammatory Cytokines1

Masanobu Ohishi,*† Yumiko Matsumura,* Daisuke Aki,* Ryuichi Mashima,* Koji Taniguchi,* Takashi Kobayashi,* Toshio Kukita,‡ Yukihide Iwamoto,† and Akihiko Yoshimura2*

Bone metabolism and the immune system have a correlative relationship, and both are controlled by various common cytokines, such as IFNs and ILs, produced in the bone microenvironments. The suppressor of cytokine signaling-1 (SOCS1) and SOCS3 are negative regulators of such cytokines. Although SOCSs are shown to be induced during osteoclast differentiation, their physiological roles in osteoclast differentiation and function have not been clarified. Thus, we examined the roles of SOCS1 and SOCS3 in osteoclastogenesis using SOCS1- and SOCS3-deficient mice. IFN-␥-mediated inhibition of osteoclast differentiation from bone marrow-derived monocytes (BMMs) was strongly enhanced in SOCS1-deficient BMMs, but was diminished in SOCS1-overex- ؉ ؊ pressing BMMs. Moreover, LPS-induced osteoclastogenesis and bone destruction in vivo were suppressed in SOCS1 / mice Downloaded from compared with those in wild-type mice, suggesting that SOCS1 antagonizes the inhibitory effect of IFN-␥ on osteoclastogenesis. SOCS3 did not alter the inhibitory effect of IFNs in osteoclastogenesis in both gain and loss of functional assays; however, the suppressive effect of IL-6 on osteoclast differentiation was greater in SOCS3-deficient BMMs than in wild-type BMMs in vitro. In addition, IL-6 significantly prevented LPS-induced bone destruction in SOCS3-deficient mice, although it failed in wild-type mice in vivo. In SOCS3-deficient BMMs, expression levels of TNF-receptor-associated factor-6 and I␬B were drastically reduced and receptor activator of the NF-␬B ligand-induced I␬B phosphorylation was severely impaired in the presence of IL-6. These data http://www.jimmunol.org/ suggest that both SOCS1 and SOCS3 regulate osteoclastogenesis by blocking the inhibitory effect of inflammatory cytokines on receptor activator of the NF-␬B ligand-mediated osteoclast differentiation signals. Selective suppression of SOCS1 and SOCS3 in osteoclast precursors may be a possible therapeutic strategy for inflammatory bone destruction. The Journal of Immunology, 2005, 174: 3024–3031.

balance between bone resorption and bone formation by osteoprotegerin ligand (7), produced by activated T cells stimu- bone-resorbing osteoclasts and bone-forming osteo- lates its cognate receptor, RANK, expressed on osteoclast

A blasts is critical for the maintenance of bone strength and precursors (8). RANKL, a member of the TNF family, is essential by guest on September 27, 2021 integrity (1). Hyperactivation and/or accumulation of osteoclasts for the differentiation of monocyte/macrophage precursors to os- cause bone destruction in pathological conditions, such as autoim- teoclasts (9). Binding of RANKL to its receptor, RANK, results in mune arthritis, periodontitis, postmenopausal osteoporosis, Paget’s the recruitment of the TNF-receptor-associated factor (TRAF) disease, and bone tumors (2, 3). Enhanced osteoclastic bone re- family of proteins, including TRAF6, which is an essential adapter sorption causes severe bone damage, leading to progressive joint molecule for osteoclastogenesis, leading to the activation of destruction in autoimmune arthritis, in which receptor activator of NF-␬B and JNK (9). In addition, RANKL induces the expression 3 the NF-␬B ligand (RANKL), (4) also known as TNF-related ac- of c-Fos, an essential transcription factor for osteoclastogenesis, by tivation-induced cytokine (5)/osteoclast differentiation factor (6)/ an unknown mechanism (10). A recent study by Takayanagi et al. (11) demonstrated that IFN-␥ secreted from activated T cells counterbalances the action *Division of Molecular and Cellular Immunology, Medical Institute of Bioregulation, †Department of Orthopedic Surgery, Faculty of Medicine, and ‡Oral Cellular and of RANKL, which contributes to maintaining bone homeostasis. Molecular Biology, Department of Oral Biological Sciences, Faculty of Dental Sci- Moreover, it was demonstrated that IFN-␤ inhibits osteoclastogen- ences, Kyushu University, Fukuoka, Japan esis by interfering with the RANKL-induced expression of c-Fos Received for publication August 31, 2004. Accepted for publication December (10). IFN-␥ is shown to induce ubiquitination and proteasomal 13, 2004. degradation of TRAF6 (11), whereas IFN-␤ induces the suppres- 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 sion of c-Fos expression by a posttranscriptional control mecha- with 18 U.S.C. Section 1734 solely to indicate this fact. nism(s) (10). Therefore, the modulation of RANK signaling by 1 This work was supported by special grants-in-aid from the Ministry of Education, inﬂammatory cytokines is important for bone remodeling to main- Science, Technology, Sports, and Culture of Japan; the Takeda Science Foundation; tain bone homeostasis in both physiologic and pathologic states. the Kato Memorial Trust for Nanbyo Research; the Haraguchi Memorial Foundation; the Yamanoucih Foundation for Research on Metabolic Disorders; and the Uehara How IFN-␥ induces degradation of TRAF6 and how IFN-␤ in- Memorial Foundation. duces c-Fos reduction remain to be determined. 2 Address correspondence and reprint requests to Dr. Akihiko Yoshimura, Division of IFNs use the JAK/STAT pathways for their signal transduction Molecular and Cellular Immunology, Medical Institute of Bioregulation, Kyushu Uni- (12). A group of cytokine-inducible factors, the suppressor of cy- versity, 3-1-1 Maidashi, Higashi-ku, Fukuoka 812-8582, Japan. E-mail address: [email protected] tokine signaling (SOCS) family, which mediate negative feedback 3 Abbreviations used in this paper: RANKL, receptor activator of NF-␬B ligand; mechanisms of cytokine signaling by inhibiting the JAK/STAT BMM, bone marrow-derived monocyte; IRES, internal ribosomal entry site; KO, pathway, has recently been discovered (12). Among them, SOCS1 knockout; MNC, multinuclear cell; RANK, receptor activator of NF-␬B; SOCS, suppressor of cytokine signaling; TRAF, TNF receptor-associated factor; TRAP, tartrate- and SOCS3, negative feedback regulators for STAT1 and STAT3 resistant acid phosphatase; WT, wild type. pathways, respectively, inhibit JAK tyrosine kinase activity in

Copyright © 2005 by The American Association of Immunologists, Inc. 0022-1767/05/$02.00 The Journal of Immunology 3025 slightly different manners; SOCS1 directly binds to the activation Retroviral gene transduction loop of JAKs through the Src homology 2 domain, whereas We constructed a retrovirus vector carrying SOCS1 and SOCS3 cDNA SOCS3 directly binds to cytokine receptors, including gp130, a with an internal ribosomal entry site (IRES)-enhanced GFP (pMX-SOCS1/ signal transducer of IL-6-related cytokines. The Src homology 2 SOCS3-IRES-EGFP). pMX-IRES-EGFP (empty) was used as a control domain of SOCS3 has been shown to bind to Y757 of gp130, vector. Retrovirus packaging was performed as previously described (25). which is also known as a binding site for the protein tyrosine Bone marrow cells obtained from 5- to 8-wk-old ddy mice were cultured for 24 h in the presence of M-CSF (R&D Systems; 100 ng/ml). Cells were phosphatase 2 (Src homology 2 domain-containing phosphatase-1) vigorously washed by pipetting, and adherent cells were incubated in a (13). These two molecules contain a similar kinase inhibitor region virus stock medium containing polybrene (Roche) at 4 ␮g/ml for 4 h. at the N terminus, which is essential for JAK inhibition (13). Anal- Infected cells were cultured for 2 days in the presence of M-CSF (100 ysis of SOCS1- and SOCS3-deficient macrophages indicated that ng/ml) and subjected to in vitro osteoclastogenesis. SOCS1 negatively regulates the IFNs/STAT1 (12) as well as TLR In vitro osteoclastogenesis (14, 15) signaling pathways, whereas SOCS3 inhibits the IL-6/ In vitro osteoclastogenesis using BMMs was performed as previously de- gp130 signaling pathways (16–18). Through examination of scribed (26). Bone marrow cells were collected from the femurs and tibias knockout mice, it has been reported that IL-6 and gp130 are in- of 5- to 8-wk-old mice. These cells were suspended in ␣MEM containing volved in regulation of osteoclast development in pathologic (19) 10% FBS and cultured in 48-well plates (1.5 ϫ 105 cells/well) in the and physiologic states (20). Interestingly, the absence of the Socs3 presence of M-CSF (100 ng/ml) for 3 days. The cells were then washed vigorously by pipetting. Adherent cells were cultured for another 3 days in gene in macrophages resulted in modulation of the character of the presence of both M-CSF (100 ng/ml) and RANKL (PeproTech; 100 IL-6 function. For example, in SOCS3-deficient macrophages, ng/ml) with or without IFN-␤ (Calbiochem), IFN-␥ (PeproTech), or IL-6 IL-6 functions like IL-10 (16) and IFNs (17, 18), both of which are (PeproTech). The cells were fixed with formalin, acetone, and acetate and Downloaded from known as negative regulators of osteoclastogenesis (10, 11, 21). stained for tartrate-resistant acid phosphatase (TRAP) using a leukocyte acid phosphatase kit (Sigma-Aldrich) according to the manufacturer’s pro- However, it is unknown whether IL-6 functions as a negative reg- tocol. The number of TRAP-positive multinuclear and/or mononuclear ulator of osteoclastogenesis, particularly in the absence of SOCS3. cells was determined by counting. In vitro osteoclastogenesis using spleen Hayashi et al. (22) reported that bone marrow-derived osteoclast cells was performed as follows. Spleen cells were collected from 2- to ϩ ϩ Ϫ Ϫ progenitors constitutively express IFN-␤, which is further up-reg- 3-wk-old SOCS1 / mice or SOCS1 / mice. These cells were cultured in ϫ 6 ulated by RANKL. IFN-␤ intrinsically inhibits the differentiation 96-well plates (1 10 cells/well) in the presence of M-CSF (25 ng/ml) http://www.jimmunol.org/ and RANKL (50 ng/ml) for 4 days. After fixation, the cells were stained of osteoclasts. In addition, RANKL simultaneously induces the for TRAP. expression of SOCS1 and SOCS3, although their putative roles in osteoclastogenesis have not been confirmed (22). It has also been Quantification of NO release demonstrated that SOCS3 is induced by TGF-␤, which antago- BMMs were cultured for 3 days in the presence of M-CSF, RANKL, and nizes the effect of IFN-␤ (23). These findings suggest that SOCS1 IFN-␥ as described above, then stimulated with TNF-␣ (PeproTech; 100 and SOCS3 suppress the intrinsic inhibitory effect of IFN-␤ on ng/ml) for 24 h. Culture supernatant was collected, and accumulation of nitrite was measured using a nitrate/nitrite assay kit (BioDynamics RANK signaling. However, no evidence supporting these propos- Laboratory). als has been provided using knockout mice. In the present study we by guest on September 27, 2021 used SOCS1- and SOCS3-deficient mice and demonstrated that Endotoxin-induced bone resorption SOCS1 and SOCS3 regulate osteoclastogenesis induced by Mice (8–12 wk old) were administered a local calvarial injection of LPS RANKL in a pathological state. These genes are especially impor- (Sigma-Aldrich) at 25 mg/kg. Some SOCS3fl/fl and LysM-Cre:SOCS3fl/fl tant for suppression of the inhibitory effect of exogenous IFNs and mice received an s.c. injection of IL-6 at the same site. After 5 days, the mice were killed, and histological sections of the calvarial bones were IL-6. Therefore, SOCS1 and SOCS3 are potentially critical factors prepared as described previously (10). The specimens were stained for for regulating bone remodeling during inflammation. TRAP and counterstained with hematoxylin. Bone destruction was evaluated by calculating the percentage of erosive surface per total bone surface (10). The osteoclast number per millimeter of trabecular bone surface was Materials and Methods counted as described previously (10). Mice Immunohistochemistry SOCS1ϩ/Ϫ, SOCS1Ϫ/Ϫ, IFN-␥Ϫ/ϪSOCS1ϩ/ϩ, IFN-␥Ϫ/ϪSOCS1Ϫ/Ϫ, Immunohistochemical staining for SOCS3 in calvarial sections was per- SOCS3fl/fl and LysM-Cre:SOCS3fl/fl mice have been described (14, 16, 24). formed using anti-SOCS3 Ab (IBL). Briefly, calvarial sections were first All mice were kept in specific pathogen-free conditions in Station for Col- treated with proteinase K. After blocking with goat serum for 30 min, the laborative Research. sections were incubated with rabbit anti-SOCS3 Ab for 30 min, followed by incubation with Alexa 546-labeled anti-rabbit IgG (Molecular Probes) for 30 min. The sample sections were further stained for TRAP as de- Confirmation of Socs3 deletion in LysM-Cre:SOCS3fl/fl bone scribed above. Immunohistochemical staining for RelA (NF-␬B p65 subunit) in BMMs marrow-derived monocytes (BMMs) was performed as follows. The cells were fixed with 70% ethanol for 5 Floxed Socs3 gene deletion was evaluated by PCR using genomic DNAs min. After fixation, the cells were treated with 0.1% Triton X-100 for 3 obtained from SOCS3fl/fl and LysM-Cre:SOCS3fl/fl BMMs. For the detec- min, followed by incubation with rabbit anti-RelA Ab (Santa Cruz Bio- tion of the floxed Socs3 allele (SOCS3fl) and the deleted Socs3 allele technology). Then they were incubated with Alexa 546-labeled (SOCS3del), the following primers were used: primer A, 5Ј-GCGG anti-rabbit IgG. Ј Ј GCAGGGGAAGAGACTGTCTGGGGTTG-3 ; primer B, 5 -GGCGCAC Western blotting GGAGCCAGCGTGGATCTGCG-3Ј; and primer C, 5Ј-AGTCCGCTT GTCAAAGGTATTGTCCCAC-3Ј. The band of the SOCS3fl locus (380 Total cells were lysed in lysis buffer A (20 mM HEPES (pH 7.3), 150 mM bp) was obtained with primers A and B. The band of the SOCS3del locus NaCl, 0.5% Nonidet P-40, 50 mM NaF, 10 mM sodium pyrophosphate, 1 (250 bp) was obtained with primers A and C. mM DTT, 1 mM sodium vanadate, 1 mM PMSF, and 1% aprotinin). Total We also evaluated Socs3 deletion in LysM-Cre:SOCS3fl/fl BMMs by cell extracts were resolved by SDS-PAGE, and proteins were detected by RT-PCR. cDNAs were obtained from BMMs cultured for5hinthepres- immunoblotting using anti-TRAF6 Ab (Santa Cruz Biotechnology), anti- ence of M-CSF (40 ng/ml) and RANKL (50 ng/ml). RT-PCR was per- phosphoserine-I␬B (Cell Signaling), anti-I␬B (Cell Signaling), anti-c-Fos formed using following primers: SOCS3 forward, 5Ј-GGGTGGCAAA Ab(OncogeneScience),anti-phosphotyrosine-STAT1(CellSignaling),anti- GAAAAGGAG-3Ј; and SOCS3 reverse, 5Ј-GTTGAGCGTCAAGA phosphotyrosine-STAT3(CellSignaling),anti-STAT1(CellSignaling),anti- CCCAGT-3Ј. STAT3 (Cell Signaling), and anti-Erk Ab (Santa Cruz Biotechnology). 3026 SOCS AND OSTEOCLASTOGENESIS

RT-PCR for osteoclast-specific genes of SOCS1 and SOCS3 on exogenous IFN-␥ and IFN-␤. As shown BMMs infected with retrovirus were cultured for 3 days in the presence of in Fig. 1, C and D, SOCS1, but not SOCS3, overexpression con- M-CSF (100 ng/ml) and RANKL (100 ng/ml) with or without IFN-␥. cDNA ferred partial resistance to both IFN-␥ and IFN-␤. SOCS1-over- was obtained, and RT-PCR was performed as previously described (22). expressing BMMs could differentiate into TRAP-positive MNCs, even in the presence of 5 U/ml IFN-␥ or 10 U/ml IFN-␤, whereas Statistical methods control and SOCS3-overexpressing BMMs failed to do so (Fig. 1, The significance of differences between group means was determined by C and D). Notably, SOCS1-overexpressing BMMs could generate Student’s t test. many TRAP-positive mononuclear cells, even in the presence of IFNs, whereas control and SOCS3-overexpressing BMMs could Results not generate any TRAP-positive cells (Fig. 1B). Resistance to IFNs Effects of SOCS1 and SOCS3 overexpression on osteoclast by SOCS1 overexpression was confirmed by monitoring the ex- differentiation pression of osteoclast-specific genes including TRAP, calcitonin To determine the effects of SOCS1 and SOCS3 on osteoclast dif- receptor, and cathepsin K, as shown in Fig. 1E. These markers ferentiation, we overexpressed SOCS1 and SOCS3 in BMMs us- were detectable in SOCS1-overexpressing BMMs even in the pres- ing a pMX retroviral vector system. Comparable levels of retro- ence of IFN-␥, suggesting that SOCS1, but not SOCS3, confers viral transduction were observed among pMX-empty, -SOCS1, resistance to exogenous IFNs on osteoclast progenitors by sup- and -SOCS3 (ϳ60%; Fig. 1A). Consistent with previous reports, pressing IFN signaling. SOCS1 overexpression inhibited IFN-␥- generation of osteoclasts (TRAP-positive multinuclear cells induced STAT1 activation and TRAF6 degradation (data not

(MNCs)) from control BMMs (pMX-empty) occurred within 3 shown). Downloaded from days of culture with RANKL, but was strongly suppressed by IFN-␥ and IFN-␤ (Fig. 1, B–D). Neither SOCS1 nor SOCS3 over- Effect of SOCS1 gene deletion on osteoclastogenesis expression caused an increase in the number of RANKL-induced, To conﬁrm the physiological relevance of the effect of SOCS1, we TRAP-positive MNCs, suggesting that overexpression of SOCS1 examined RANKL-induced osteoclast differentiation using BMMs or SOCS3 does not affect RANKL-induced osteoclastogenesis un- from control (IFN-␥Ϫ/ϪSOCS1ϩ/ϩ) and SOCS1-deﬁcient (IFN- Ϫ/Ϫ Ϫ/Ϫ der normal conditions (Fig. 1, B–D). We then examined the effects ␥ SOCS1 ) mice. Similar numbers of TRAP-positive MNCs http://www.jimmunol.org/ by guest on September 27, 2021

FIGURE 1. SOCS1, but not SOCS3, overexpression conferred resistance to suppression of osteoclast differentiation by IFNs. A, BMMs obtained from ddy mice were infected with the indicated retrovirus. Forty-eight hours after gene transduction, GFP-positive cells were monitored. B, BMMs infected with the indicated retrovirus were cultured in the presence of M-CSF (100 ng/ ml) and RANKL (100 ng/ml) with or without IFN-␥ or IFN-␤ for 3 days. Cells were then stained for TRAP and photographed (B), or the numbers of TRAP-positive multinucleated cells (TRAP (ϩ) MNCs) were scored (C and D). All values are the mean Ϯ SEM of triplicate .p Ͻ 0.05 ,ء .cultures from three different experiments E, Total RNA was isolated from the infected BMMs cultured with or without IFN-␥ (10 U/ml), and RT-PCR was performed for TRAP, calcitonin receptor (CTR), and cathepsin K (Cat K), as described in Materials and Methods. The Journal of Immunology 3027 were obtained from wild-type (WT; IFN-␥ϩ/ϩSOCS1ϩ/ϩ) and BMMs in both SOCS1Ϫ/Ϫ and SOCS1ϩ/ϩ BMMs (Fig. 2C), sug- IFN-␥-deficient (IFN-␥Ϫ/ϪSOCS1ϩ/ϩ) bone marrow (Fig. 2A). gesting that the suppression of osteoclast differentiation by IFN-␥ The numbers of osteoclasts differentiated from SOCS1-deficient was not due to a cytotoxic effect of IFN-␥. IFN-␥ is known to BMMs were less than those from control BMMs. This indicates induce macrophage activation. Therefore, we examined NO syn- that SOCS1 might be an intrinsic regulator of osteoclastogenesis. thesis induced by TNF-␣ in BMM cultures in the presence of However, SOCS1 seems not to be indispensable for regulating IFN-␥. TNF-␣ induced NO production more efficiently in osteoclastogenesis in vivo under normal conditions, because bone SOCS1Ϫ/Ϫ than in SOCS1ϩ/ϩ BMMs after IFN-␥ pretreatment histomorphometric analysis showed no difference in the number of (Fig. 2D). In addition, the numbers of multinucleated osteoclasts osteoclasts between SOCS1-deficient and control mice (data not differentiated from spleen cells of IFN-␥ϩ/ϩSOCS1Ϫ/Ϫ mice in shown). which IFN-␥ levels are shown to be high (27) were less than those Next we examined the effect of IFN-␥ on osteoclast differenti- from IFN-␥ϩ/ϩSOCS1ϩ/ϩ mice (Fig. 2E), whereas spleens of ation of SOCS1-deficient BMMs. As shown in Fig. 2B, SOCS1- IFN-␥ϩ/ϩSOCS1Ϫ/Ϫ mice contained more macrophages than deficient BMMs were much more sensitive to IFN-␥ than control those from IFN-␥ϩ/ϩSOCS1ϩ/ϩ mice (28). Taken together, these BMMs. Osteoclastogenesis was almost completely inhibited in data suggest that SOCS1-deficient BMMs were more sensitive to SOCS1Ϫ/Ϫ BMMs in the presence of 0.1 U/ml IFN-␥, whereas it the negative effect of IFN-␥ for osteoclast differentiation and to the was inhibited by only 50% in SOCS1ϩ/ϩ BMMs at the same con- positive effect of IFN-␥ for macrophage differentiation. We also centration. IFN-␥ had little effect on the total cell number of compared the sensitivity to IFN-␤-induced suppression of osteoclastogenesis between control and SOCS1Ϫ/Ϫ BMMs. No signif-

icant difference was observed between them, which indicates that Downloaded from SOCS1 is not involved in negative regulation of IFN-␤ signaling in osteoclast precursors (Fig. 2F).

SOCS1 deficiency suppressed endotoxin-induced bone resorption To determine the effect of SOCS1 on osteoclastogenesis in vivo, we used an established model of endotoxin-induced bone resorp- http://www.jimmunol.org/ tion, in which T cells and IFN-␥ play an essential regulatory role (10, 11). We administered a local calvarial injection of LPS to 7- to 8-wk-old Socs1 heterozygous (SOCS1ϩ/Ϫ) and WT (SOCS1ϩ/ϩ) mice. Because SOCS1Ϫ/Ϫ mice show very severe inflammation in various organs and die at ϳ3 wk of age (12, 27, 28), it is not possible to assess LPS-induced bone resorption in vivo. In contrast, IFN-␥Ϫ/ϪSOCS1Ϫ/Ϫ mice are not adequate for ␥ this model, because IFN- is not present. Therefore, we compared by guest on September 27, 2021 SOCS1ϩ/Ϫ and SOCS1ϩ/ϩ mice. Five days after calvarial injection of LPS, we measured the percentage of bone surface covered by osteoclasts (eroded surface) and the osteoclast number per millimeter of trabecular bone surface (osteoclast number). As shown in Fig. 3A, a notable suppression of osteoclast formation and bone destruction induced by LPS was observed in mice with haploin- sufficiency of the Socs1 gene. Quantitative analyses evaluated by eroded surface and osteoclast number revealed that the LPS-induced osteoclast development and activation were significantly suppressed in SOCS1ϩ/Ϫ mice (Fig. 3, B and C). These findings indicate that SOCS1 is an important regulator of bone remodeling in inflammation sites.

Effects of Socs3 gene deletion on osteoclast differentiation FIGURE 2. Hyper-responsiveness of SOCS1-deficient BMMs to ϩ ϩ ϩ ϩ Ϫ Ϫ ϩ ϩ SOCS3 expression is also up-regulated by RANKL in BMMs (22). IFN-␥. BMMs from IFN-␥ / SOCS1 / mice, IFN-␥ / SOCS1 / mice, and IFN-␥Ϫ/ϪSOCS1Ϫ/Ϫ mice were cultured in the presence of M-CSF and To determine the physiological role of SOCS3 in osteoclastogen- RANKL, and the cells were stained for TRAP after fixation. The numbers esis, we used a conditional knockout (KO) approach, because of TRAP-positive MNCs were counted (A). BMMs were cultured for 3 SOCS3-deficient mice die during embryonic development as a re- days with different concentrations of IFN-␥ (B) and IFN-␤ (F) in the pres- sult of placental defects (29). To generate a monocyte-specific de- ence of RANKL (100 ng/ml) and M-CSF (100 ng/ml). The numbers of letion, a mouse line in which the Cre cDNA was knocked-in at the TRAP-positive MNCs were assessed to evaluate the sensitivity of BMMs lysozyme M gene locus (LysM-Cre mice) (24) was crossed with to IFNs. C, Total cell numbers were counted. D, Influence of SOCS1 de- SOCS3fl/fl mice. Deletion of the Socs3 gene in BMMs was con- ficiency on nitrite production. After a 3-day culture of BMMs under the firmed by genomic PCR and RT-PCR (Fig. 4, A and B). BMMs ␣ conditions described in A, TNF- was added to the culture medium, and from LysM-Cre:SOCS3fl/fl mice (SOCS3-KO) showed a 250-bp cells were further incubated for 24 h. Nitrite production in the culture band corresponding to the Cre-mediated deletion of the Socs3 medium was measured. E, The numbers of TRAP-positive MNCs differ- ϩ/ϩ Ϫ/Ϫ gene, whereas no deletion band was observed in the DNA from entiated from the spleen cells of SOCS1 mice and SOCS1 mice fl/fl were also counted. B and F, The numbers of TRAP-positive MNCs in the SOCS3 mice (SOCS3-WT; Fig. 4A). In contrast, BMMs from cultures without IFNs are valued as 100%. All values are the mean Ϯ SEM SOCS3-WT exhibited a 380-bp band corresponding to the floxed p Ͻ Socs3 gene, whereas the floxed gene could not be amplified from ,ءء ;p Ͻ 0.01 ,ء .of triplicate cultures from three pairs of female mice 0.05. SOCS3-KO BMMs (Fig. 4A). SOCS3 mRNA was induced by 3028 SOCS AND OSTEOCLASTOGENESIS Downloaded from http://www.jimmunol.org/

FIGURE 4. SOCS3 gene deletion in BMMs from LysM-Cre:SOCS3fl/fl ϩ ϩ ϩ Ϫ mice. A, PCR analysis of genomic DNA from BMMs. The PCR product of FIGURE 3. LPS-induced bone resorption in SOCS1 / and SOCS1 / ϩ ϩ ϩ Ϫ the WT SOCS3 allele is 280 bp, and that from the SOCS3fl allele is 380 mice. A, Histology of calvarial bone of SOCS1 / and SOCS1 / mice bp. A band of 250 bp indicates Cre-mediated deletion of the Socs3 gene in injected with saline or LPS (TRAP and hematoxylin staining). Higher mag- LysM-Cre:SOCS3fl/fl-derived BMMs. B, mRNA expression levels of nifications of the encircled area of LPS-injected calvaria are shown in the SOCS3 in RANKL-stimulated BMMs. BMMs from both SOCS3fl/fl and lowest panels. Osteoclast-resorbing bone surfaces are pointed out by ar- LysM-Cre:SOCS3fl/fl mice were incubated with RANKL (100 ng/ml) for rows. B, The percentage of erosive surface per bone surface (eroded sur- 5 h. SOCS3 mRNA was detected by RT-PCR analysis. C, SOCS3 deletion face) and the numbers of osteoclasts per millimeter of trabecular bone by guest on September 27, 2021 in osteoclasts in vivo. Specimens of calvarias were immunohistochemically surface (osteoclast number) were scored. All values are the mean Ϯ SEM ϩ Ϫ ϩ ϩ stained for SOCS3 as described in Materials and Methods. Then, the same . / p Ͻ 0.01, SOCS1 / vs SOCS1 ,ء .from five mice per group samples were stained for TRAP for osteoclast detection. SOCS3-KO, LysM-Cre:SOCS3fl/fl; SOCS3-WT; SOCS3fl/fl. RANKL treatment within5hinBMMs from SOCS3-WT, although it could not be detected in those from SOCS3-KO mice (Fig. 4B). Furthermore, we confirmed that SOCS3 protein was not and I␬B levels were not much affected by IL-6 pretreatment, detected in osteoclasts of SOCS3-KO mice in vivo (Fig. 4C). whereas the expression levels of TRAF6 and I␬B, but not that of The numbers of TRAP-positive MNCs differentiated from c-Fos, were significantly reduced in the presence of IL-6 in SOCS3-deficient BMMs were significantly reduced compared SOCS3-deficient BMMs (Fig. 6B). Moreover, phosphorylation of with those from WT BMMs (Fig. 5A). This indicates that SOCS3 I␬B was diminished by IL-6 pretreatment in SOCS3-deficient may play a role in osteoclast differentiation induced by RANKL, BMMs. Consistent with these data, RANKL-induced nuclear although the mechanism is unclear. We did not see any differences translocation of RelA (p65), a subunit of NF-␬B, was inhibited by in the inhibitory effect of IFN-␤ and IFN-␥ on osteoclast differ- IL-6 in SOCS3-deficient BMMs (Fig. 6C). These data indicate that entiation between WT and SOCS3-deficient BMMs (Fig. 5, B and RANKL signaling was severely suppressed by IL-6 in SOCS3- C). Thus, SOCS3, unlike SOCS1, does not play a major regulatory deficient BMMs, resulting in the suppression of osteoclast role in IFN-␥ sensitivity. Other studies as well as ours have shown differentiation. that IL-6-gp130 signaling is specifically enhanced by the Socs3 gene deletion in macrophages (16–18). This prompted us to ex- IL-6 prevents LPS-induced bone resorption in SOCS3-deficient amine the effect of IL-6 on osteoclastogenesis. As shown in Fig. mice 5D, IL-6 did not affect osteoclast differentiation of WT BMMs at Next we examined the role of SOCS3 in vivo using LysM-Cre: a low concentration (0.1 ng/ml), but partially suppressed it at SOCS3fl/fl mice. Unlike osteoclastogenesis in vitro, bone density higher concentrations (Ն1 ng/ml). In contrast, IL-6 strongly sup- and mass were not significantly different between SOCS3-KO and pressed the osteoclast differentiation of SOCS3-deficient BMMs control SOCS3-WT mice under normal conditions in vivo. To as- even at low IL-6 concentrations and completely blocked it at sess the role of SOCS3 in the regulation of inflammatory bone higher concentrations. destruction, we used LPS-induced bone destruction model (Fig. 7). To define the molecular mechanism of this phenomenon, we LPS-induced bone destruction was less severe, but not signifi- examined the levels of TRAF6, c-Fos, and I␬B as well as the cantly different, in SOCS3-KO mice compared with SOCS3-WT phosphorylation of I␬B, STAT1, and STAT3 (Fig. 6). IL-6-in- mice. However, when exogenous IL-6 was s.c. administered at the duced phosphorylation of STAT3 as well as STAT1 was enhanced same site of LPS injection, bone destruction in SOCS3-KO mice in SOCS3-KO BMMs (Fig. 6A). In WT BMMs, TRAF6, c-Fos, was strongly suppressed, which was not observed in control The Journal of Immunology 3029 Downloaded from http://www.jimmunol.org/

FIGURE 5. SOCS3-deficient BMMs are hyper-responsive to IL-6, but not to IFNs. A, BMMs derived from both SOCS3-WT and SOCS3-KO mice were cultured for 3 days in the presence of RANKL (100 ng/ml) and M-CSF (100 ng/ml). Cells were fixed and stained for TRAP. The numbers p Ͻ 0.05. B–D, BMMs were ,ء .of TRAP-positive MNCs were counted cultured for 3 days with various concentrations of IFN-␤ (B), IFN-␥ (C), or IL-6 (D) in the presence of RANKL (100 ng/ml) and M-CSF (100 ng/ml). The number of TRAP-positive MNCs in the cultures without IFNs and FIGURE 6. Effects of IL-6 on both STAT and TRAF6-NF-␬B activa- IL-6 is valued as 100%. tion in SOCS3-KO BMMs. A, Prolonged phosphorylation of STAT1 and by guest on September 27, 2021 STAT3 induced by IL-6 in SOCS3-deficient BMMs. BMMs from SOCS3-WT and SOCS3-KO mice were cultured with M-CSF (100 ng/ml) SOCS3-WT littermates. These results are consistent with in vitro and RANKL (100 ng/ml) with or without IL-6 (1 ng/ml). Whole cell ex- experiments showing that IL-6 strongly suppressed the osteoclast tracts were immunoblotted with the indicated Abs. B, Suppression of ␬ ␬ differentiation of SOCS3-deficient BMMs. TARF6 and I B expression level and RANKL-induced I B phosphorylation by IL-6 in SOCS3-deficient BMMs. Whole cell extracts from BMMs cultured as described in A were immunoblotted with the indicated Abs. C, Discussion Inhibition of NF-␬B nuclear translocation by IL-6 in SOCS3-KO BMMs. In this study we tried to clarify the physiological and pathological SOCS3-WT and SOCS3-KO BMMs were stimulated with RANKL with or roles of SOCS1 and SOCS3 expression in osteoclastogenesis. We without IL-6. The cells were immunohistochemically stained with anti- have found that neither SOCS1 nor SOCS3 overexpression in RelA Ab as described in Materials and Methods. BMMs impacts RANKL-induced osteoclastogenesis. Although the number of multinucleated TRAP-positive osteoclasts induced by RANKL was reduced in both SOCS1- and SOCS3-deficient tion of IL-6 in SOCS3-conditional KO mice (Fig. 7). These data BMMs in vitro, the numbers of osteoclasts were not significantly indicate that SOCS1 and SOCS3 are necessary for osteoclastogen- different between IFN-␥Ϫ/ϪSOCS1ϩ/ϩ and IFN-␥Ϫ/ϪSOCS1Ϫ/Ϫ esis by blocking inflammatory cytokine signaling at inflammation mice in vivo (data not shown). Moreover, the bone density and sites. However, because SOCS1 or SOCS3 deficiency did not re- mass of SOCS3fl/fl and LysM-Cre:SOCS3fl/fl mice were almost sult in complete resistance to LPS-induced bone destruction, it is identical (data not shown). Thus, neither Socs1 nor Socs3 gene possible that they compensate each other. disruption impacts on osteoclastogenesis under physiological con- Bone destruction is a pathological hallmark of several chronic ditions in vivo. The mechanism of reduced osteoclast differentia- inflammatory diseases, including rheumatoid arthritis and peri- tion from SOCS1- and SOCS3-deficient BMMs in response to odontitis. In these disorders, the recruitment of macrophage lin- RANKL (Figs. 2A and 5A) remains to be solved. eage cells to the site of inflammation and the action of local os- Importantly, our present data indicate that SOCS1 and SOCS3 teoclastogenic cytokines associated with the inflammatory process play critical roles in osteoclastogenesis in pathological situations initiate the bone destruction. Inflammation-induced bone loss of both in vitro and in vivo. SOCS1- and SOCS3-deficient BMMs this sort results from elevated numbers of osteoclasts (3, 8, 11). were more sensitive to the suppressive effect of IFN-␥ (Fig. 2B) The differentiation of osteoclasts is dependent on signals from and IL-6 (Fig. 5D), respectively, than the control BMMs in vitro. RANK, which is stimulated by its ligand, RANKL, expressed on LPS-mediated calvarial bone destruction and osteoclast induction osteoblasts (5–7), fibroblasts (30), and activated T cells (8). Up- were suppressed in SOCS1ϩ/Ϫ in vivo (Fig. 3). Furthermore, LPS- regulation of RANKL expression on osteoblasts and synovial fi- induced bone destruction was strongly blocked by coadministra- broblasts by inflammatory cytokines such as TNF-␣, IL-1, and 3030 SOCS AND OSTEOCLASTOGENESIS

of the negative feedback regulation of osteoclastogenesis (10). The number of osteoclasts was notably greater in the bones of IFN-␤ receptor-deficient mice than in those of WT mice (10). Because SOCS1 overexpression inhibits IFN-␤ signaling, and SOCS1 deficiency resulted in low efficiency of osteoclast differentiation, we expected to see a reduction of osteoclasts in SOCS1-deficient mice. However, there was little difference in the number of osteoclasts between SOCS1-deficient mice and controls. Moreover, no difference in the suppressive effect of exogenous IFN-␤ was seen between SOCS1-deficient BMMs and SOCS1ϩ/ϩ BMMs (Fig. 2F). The reason for this is not clear at present. The SOCS1 protein level induced by RANKL may not be high enough to block IFN-␤ signaling. Additional study is necessary to define the role of SOCS1 in the regulation of autocrine IFN-␤. The molecular mechanism of the suppression of RANK signaling by inflammatory cytokines has not been clarified. Takayanagi and his colleagues proposed that IFN-␥ induces the ubiquitination and degradation of TRAF6 by a STAT1-dependent mechanism, and that

IFN-␤ suppresses c-Fos induction by an ISGF3-dependent mecha- Downloaded from nism (10). However, little is known about how STAT1 induces TRAF6 ubiquitination and how ISGF3 suppresses c-Fos expression. In WT cells, IL-6 has less effect on osteoclastogenesis than SOCS3-KO cells probably because IL-6/gp130 cannot transmit suf- ﬁcient signals in the presence of SOCS3. In this study we found that

IL-6 functions against RANK signaling, such as IFN-␥ and IFN-␤ in http://www.jimmunol.org/ the absence of SOCS3. Because the c-Fos level was not reduced by IL-6 in SOCS3-deficient BMMs, the mechanism of the effect of IL-6 FIGURE 7. Suppression of LPS-induced bone destruction by IL-6 in on SOCS3-deficient BMMs may be different from the IFN-␤- SOCS3-KO mice. LPS was injected into calvaria of SOCS3-WT and mediated effect. Instead, we found a strong suppression of TRAF6 SOCS3-KO mice with or without s.c. injection of IL-6 at the same site. and I␬B expression levels and RANKL-induced I␬B phosphorylation Samples were collected 5 days after the injection and then stained with by IL-6 in SOCS3-deficient BMMs. This phenomenon is similar to TRAP and hematoxylin (A). The results of bone morphometric analysis that observed in osteoclast precursors treated with IFN-␥ (11). IL-6 (eroded surface and osteoclast number per millimeter) are shown in B and ␬ p Ͻ 0.01; may inhibit the TRAF6-NF- B pathway in the absence of SOCS3 ,ء .C. All values are the mean Ϯ SEM from five mice per group by guest on September 27, 2021 p Ͻ 0.05. through hyper-STAT1 activation (Fig. 6). This is consistent with an ,ءء IFN-␥-like effect of IL-6 in the absence of SOCS3 (17, 18). In addition, NF-␬B activation may be down-regulated by STAT3, be- IL-6 (31) contributes to osteoclastogenesis. However, direct ad- cause the anti-inflammatory cytokine IL-10 suppresses the TLR- ministration of inflammatory cytokines, such as IFN-␥ (Fig. 2) and mediated NF-␬B pathway through a STAT3-dependent mechanism IL-6 (Fig. 5), to osteoclast progenitors was found to suppress os- (34, 35). IL-10 was shown to be a potent inhibitor of osteoclastogen- teoclast differentiation. Therefore, inflammatory cytokines regulate esis (21). We showed that IL-6 functions like IL-10 in the absence of osteoclastogenesis both positively and negatively. It is likely that SOCS3 in macrophages (16). Therefore, we propose that IL-6 func- the balance between the induction of RANKL and the inhibition of tions like IFN-␥ and/or IL-10, which negatively regulate osteoclasto- RANK signaling by inflammatory cytokines controls the develop- genesis, in SOCS3-deficient BMMs. Therefore, IL-6/gp130 signaling ment of bone destruction in inflammatory diseases. As shown in probably suppresses RANKL-induced NF-␬B activation by multiple this study, SOCS1 and SOCS3 are important regulators of RANK mechanisms in the absence of SOCS3. In other words, SOCS3 signaling in osteoclasts and might contribute to this balance at determines the nature of cytokine function. The molecular mecha- inflammation sites. nisms of suppression of TRAF6-NF-␬B activation by STAT1 and In chronic synovitis of RA, the balance may be skewed in favor STAT3 are still open questions. of RANKL expression even in the presence of high levels of inflammatory cytokines, including IL-6. Other studies as well as ours Acknowledgments have reported that SOCS1 (32) and SOCS3 (33) are highly ex- We thank Y. Kawabata for technical assistance, Y. Nishi for manuscript pressed in arthritic joints. Such high SOCS1 and SOCS3 expres- preparation, M. C. Walsh for critical reading and review of the manuscript, sion in osteoclast precursors may inhibit inflammatory cytokine and Drs. H. Takayanagi and T. Watanabe for participating in discussions. signaling, thereby canceling the suppressing effect of these inflam- Disclosures matory cytokines on RANK signaling. There is a possibility that The authors have no financial conflict of interest. the protection from bone damage in SOCS1- and SOCS3-deficent mice is partly due to the elevated levels of endogenous IFN-␥ or References IL-6. However, we could not detect significant differences in se- 1. Manolagas, S. C. 2000. Birth and death of bone cells: basic regulatory mecha- rum levels of these cytokines after LPS challenge. Therefore, we nisms and implications for the pathogenesis and treatment of osteoporosis. En- think that this possibility is unlikely, although we could not ex- docr. Rev. 21:115. 2. Roodman, G. D. 1996. Advances in bone biology: the osteoclast. Endocr. Rev. clude high levels of those cytokines in inflammation sites. 17:308. In our view, SOCS1 and SOCS3 are not involved in the homeo- 3. Rodan, G. A., and T. J. Martin. 2000. Therapeutic approaches to bone diseases. ␤ Science 289:1508. static regulation of osteoclasts. This is surprising, because IFN- , 4. Anderson, D. M., E. Maraskovsky, W. L. Billingsley, W. C. Dougall, which is rapidly up-regulated by RANKL, is reported to be a part M. E. Tometsko, E. R. Roux, M. C. Teepe, R. F. DuBose, D. Cosman, and The Journal of Immunology 3031

L. Galibert. 1997. A homologue of the TNF receptor and its ligand enhance T-cell 21. Hong, M. H., H. Williams, C. H. Jin, and J. W. Pike. 2000. The inhibitory effect growth and dendritic-cell function. Nature 390:175. of interleukin-10 on mouse osteoclast formation involves novel tyrosine-phos- 5. Wong, B. R., J. Rho, J. Arron, E. Robinson, J. Orlinick, M. Chao, S. Kalachikov, phorylated proteins. J. Bone Miner. Res. 15:911. E. Cayani, F. S. Bartlett III, W. N. Frankel, et al. 1997. TRANCE is a novel 22. Hayashi, T., T. Kaneda, Y. Toyama, M. Kumegawa, and Y. Hakeda. 2002. Reg- ligand of the tumor necrosis factor receptor family that activates c-Jun N-terminal ulation of receptor activator of NF-␬B ligand-induced osteoclastogenesis by en- kinase in T cells. J. Biol. Chem. 272:25190. dogenous interferon-␤ (IFN-␤) and suppressors of cytokine signaling (SOCS): ␤ 6. Yasuda, H., N. Shima, N. Nakagawa, K. Yamaguchi, M. Kinosaki, S. Mochizuki, the possible counteracting role of SOCSs in IFN- -inhibited osteoclast forma- A. Tomoyasu, K. Yano, M. Goto, A. Murakami, et al. 1998. Osteoclast differ- tion. J. Biol. Chem. 277:27880. entiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory fac- 23. Fox, S. W., S. J. Haque, A. C. Lovibond, and T. J. Chambers. 2003. The possible ␤ tor and is identical to TRANCE/RANKL. Proc. Natl. Acad. Sci. USA 95:3597. role of TGF- -induced suppressors of cytokine signaling expression in oste- 7. Lacey, D. L., E. Timms, H. L. Tan, M. J. Kelley, C. R. Dunstan, T. Burgess, oclast/macrophage lineage commitment in vitro. J. Immunol. 170:3679. R. Elliott, A. Colombero, G. Elliott, S. Scully, et al. 1998. Osteoprotegerin ligand 24. Clausen, B. E., C. Burkhardt, W. Reith, R. Renkawitz, and I. Forster. 1999. is a cytokine that regulates osteoclast differentiation and activation. Cell 93:165. Conditional gene targeting in macrophages and granulocytes using LysMcre mice. Transgenic Res. 8:265. 8. Kong, Y. Y., U. Feige, I. Sarosi, B. Bolon, A. Tafuri, S. Morony, C. Capparelli, 25. Sasaki, A., K. Inagaki-Ohara, T. Yoshida, A. Yamanaka, M. Sasaki, J. Li, R. Elliott, S. McCabe, et al. 1999. Activated T cells regulate bone loss and H. Yasukawa, A. E. Koromilas, and A. Yoshimura. 2003. The N-terminal trun- joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature cated isoform of SOCS3 translated from an alternative initiation AUG codon 402:304. under stress conditions is stable due to the lack of a major ubiquitination site, 9. Boyle, W. J., W. S. Simonet, and D. L. Lacey. 2003. Osteoclast differentiation Lys-6. J. Biol. Chem. 278:2432. and activation. Nature 423:337. 26. Kobayashi, K., N. Takahashi, E. Jimi, N. Udagawa, M. Takami, S. Kotake, 10. Takayanagi, H., S. Kim, K. Matsuo, H. Suzuki, T. Suzuki, K. Sato, T. Yokochi, N. Nakagawa, M. Kinosaki, K. Yamaguchi, N. Shima, et al. 2000. Tumor ne- H. Oda, K. Nakamura, N. Ida, et al. 2002. RANKL maintains bone homeostasis crosis factor ␣ stimulates osteoclast differentiation by a mechanism independent ␤ through c-Fos-dependent induction of interferon- . Nature 416:744. of the ODF/RANKL-RANK interaction. J. Exp. Med. 191:275. 11. Takayanagi, H., K. Ogasawara, S. Hida, T. Chiba, S. Murata, K. Sato, 27. Marine, J. C., D. J. Topham, C. McKay, D. Wang, E. Parganas, D. Stravopodis, A. Takaoka, T. Yokochi, H. Oda, K. Tanaka, et al. 2000. T-cell-mediated regu- A. Yoshimura, and J. N. Ihle. 1999. SOCS1 deﬁciency causes a lymphocyte- lation of osteoclastogenesis by signalling cross-talk between RANKL and IFN-␥. dependent perinatal lethality. Cell 98:609. Downloaded from Nature 408:600. 28. Alexander, W. S., R. Starr, J. E. Fenner, C. L. Scott, E. Handman, N. S. Sprigg, 12. Yasukawa, H., A. Sasaki, and A. Yoshimura. 2000. Negative regulation of cy- J. E. Corbin, A. L. Cornish, R. Darwiche, C. M. Owczarek, et al. 1999. SOCS1 tokine signaling pathways. Annu. Rev. Immunol. 18:143. is a critical inhibitor of interferon ␥ signaling and prevents the potentially fatal 13. Nicholson, S. E., D. De Souza, L. J. Fabri, J. Corbin, T. A. Willson, J. G. Zhang, neonatal actions of this cytokine. Cell 98:597. A. Silva, M. Asimakis, A. Farley, A. D. Nash, et al. 2000. Suppressor of cytokine 29. Takahashi, Y., N. Carpino, J. C. Cross, M. Torres, E. Parganas, and J. N. Ihle. signaling-3 preferentially binds to the SHP-2-binding site on the shared cytokine 2003. SOCS3: an essential regulator of LIF receptor signaling in trophoblast giant receptor subunit gp130. Proc. Natl. Acad. Sci. USA 97:6493. cell differentiation. EMBO J. 22:372.

14. Kinjyo, I., T. Hanada, K. Inagaki-Ohara, H. Mori, D. Aki, M. Ohishi, H. Yoshida, 30. Takayanagi, H., H. Iizuka, T. Juji, T. Nakagawa, A. Yamamoto, T. Miyazaki, http://www.jimmunol.org/ M. Kubo, and A. Yoshimura. 2002. SOCS1/JAB is a negative regulator of LPS- Y. Koshihara, H. Oda, K. Nakamura, and S. Tanaka. 2000. Involvement of re- ␬ induced macrophage activation. Immunity 17:583. ceptor activator of nuclear factor B ligand/osteoclast differentiation factor in 15. Nakagawa, R., T. Naka, H. Tsutsui, M. Fujimoto, A. Kimura, T. Abe, E. Seki, osteoclastogenesis from synoviocytes in rheumatoid arthritis. Arthritis Rheum. S. Sato, O. Takeuchi, K. Takeda, et al. 2002. SOCS-1 participates in negative 43:259. regulation of LPS responses. Immunity 17:677. 31. Nakashima, T., Y. Kobayashi, S. Yamasaki, A. Kawakami, K. Eguchi, H. Sasaki, and H. Sakai. 2000. Protein expression and functional difference of membrane- 16. Yasukawa, H., M. Ohishi, H. Mori, M. Murakami, T. Chinen, D. Aki, T. Hanada, bound and soluble receptor activator of NF-␬B ligand: modulation of the expres- K. Takeda, S. Akira, M. Hoshijima, et al. 2003. IL-6 induces an anti- sion by osteotropic factors and cytokines. Biochem. Biophys. Res. Commun. inflammatory response in the absence of SOCS3 in macrophages. Nat. Immunol. 275:768. 4:551. 32. Egan, P. J., K. E. Lawlor, W. S. Alexander, and I. P. Wicks. 2003. Suppressor of 17. Lang, R., A. L. Pauleau, E. Parganas, Y. Takahashi, J. Mages, J. N. Ihle, cytokine signaling-1 regulates acute inflammatory arthritis and T cell activation. R. Rutschman, and P. J. Murray. 2003. SOCS3 regulates the plasticity of gp130 J. Clin. Invest. 111:915. by guest on September 27, 2021 signaling. Nat. Immunol. 4:546. 33. Shouda, T., T. Yoshida, T. Hanada, T. Wakioka, M. Oishi, K. Miyoshi, 18. Croker, B. A., D. L. Krebs, J. G. Zhang, S. Wormald, T. A. Willson, S. Komiya, K. Kosai, Y. Hanakawa, K. Hashimoto, et al. 2001. Induction of the E. G. Stanley, L. Robb, C. J. Greenhalgh, I. Forster, B. E. Clausen, et al. 2003. cytokine signal regulator SOCS3/CIS3 as a therapeutic strategy for treating in- SOCS3 negatively regulates IL-6 signaling in vivo. Nat. Immunol. 4:540. flammatory arthritis. J. Clin. Invest. 108:1781. 19. Balto, K., H. Sasaki, and P. Stashenko. 2001. Interleukin-6 deficiency increases 34. Riley, J. K., K. Takeda, S. Akira, and R. D. Schreiber. 1999. Interleukin-10 inflammatory bone destruction. Infect. Immun. 69:744. receptor signaling through the JAK-STAT pathway: requirement for two distinct 20. Kawasaki, K., Y. H. Gao, S. Yokose, Y. Kaji, T. Nakamura, T. Suda, K. Yoshida, receptor-derived signals for anti-inflammatory action. J. Biol. Chem. 274:16513. T. Taga, T. Kishimoto, H. Kataoka, et al. 1997. Osteoclasts are present in gp130- 35. Moore, K. W., R. de Waal Malefyt, R. L. Coffman, and A. O’Garra. 2001. In- deficient mice. Endocrinology 138:4959. terleukin-10 and the interleukin-10 receptor. Annu. Rev. Immunol. 19:683.