Reduces the Serum Level of Receptor Activator of NF- κB Ligand Derived from Osteoblasts

This information is current as Yuko Nakamichi, Nobuyuki Udagawa, Yasuhiro Kobayashi, of October 1, 2021. Midrori Nakamura, Yohei Yamamoto, Teruhito Yamashita, Toshihide Mizoguchi, Masahiro Sato, Makio Mogi, Josef M. Penninger and Naoyuki Takahashi J Immunol 2007; 178:192-200; ; doi: 10.4049/jimmunol.178.1.192 http://www.jimmunol.org/content/178/1/192 Downloaded from

References This article cites 43 articles, 11 of which you can access for free at:

http://www.jimmunol.org/content/178/1/192.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 October 1, 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 © 2007 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Osteoprotegerin Reduces the Serum Level of Receptor Activator of NF-␬B Ligand Derived from Osteoblasts1

Yuko Nakamichi,* Nobuyuki Udagawa,† Yasuhiro Kobayashi,* Midrori Nakamura,† Yohei Yamamoto,† Teruhito Yamashita,* Toshihide Mizoguchi,* Masahiro Sato,† Makio Mogi,‡ Josef M. Penninger,§ and Naoyuki Takahashi2*

Osteoprotegerin (OPG) is a decoy receptor for receptor activator of NF-␬B ligand (RANKL). We previously reported that OPG deficiency elevated the circulating level of RANKL in mice. Using OPG؊/؊ mice, we investigated whether OPG is involved in the shedding of RANKL by cells expressing RANKL. Osteoblasts and activated T cells in culture released a large amount of RANKL in the absence of OPG. OPG or a soluble form of receptor activator of NF-␬B (the receptor of RANKL) suppressed the release of RANKL from those cells. OPG- and T cell-double-deficient mice showed an elevated serum RANKL level equivalent to that of ؊/؊ ؊/؊

OPG mice, indicating that circulating RANKL is mainly derived from bone. The serum level of RANKL in OPG mice was Downloaded from ␣ increased by ovariectomy or administration of 1 ,25-dihydroxyvitamin D3. Expression of RANKL mRNA in bone, but not ␣ ؊/؊ or spleen, was increased in wild-type and OPG mice by 1 ,25-dihydroxyvitamin D3. These results suggest that OPG suppresses the shedding of RANKL from osteoblasts and that the serum RANKL in OPG؊/؊ mice exactly reflects the state of bone resorption. The Journal of Immunology, 2007, 178: 192–200.

he maintenance of bone mass requires a strict balance be- RANKL from cells expressing RANKL (7, 8). Osteoblasts and T http://www.jimmunol.org/ tween bone formation by osteoblasts and bone resorption by cells express not only metalloproteinases but also tissue inhibitor T osteoclasts. The differentiation and activation of osteoclasts of metalloproteinases (TIMPs) (9–11). Osteoblasts in culture are tightly regulated by osteoblasts (1, 2). Osteoblasts express two hardly release RANKL, but activated T cells in culture release a cytokines essential for osteoclast differentiation: receptor activator of large amount of RANKL in a soluble form (6, 12). Therefore, the 3 NF-␬B ligand (RANKL) and M-CSF. RANKL is a type II mem- soluble RANKL produced by activated T cells is believed to be brane protein that is a member of the TNF family (3, 4). The involved in bone resorption induced by inflammation (13, 14). expression of RANKL in osteoblasts is up-regulated by osteotropic However, it is not known why activated T cells but not osteoblasts ␣ ␣ factors such as 1 ,25-dihydroxyvitamin D3 (1 ,25(OH)2D3) and easily release RANKL in a soluble form. by guest on October 1, 2021 parathyroid hormone (5). Osteoclast precursors express RANK Osteoprotegerin (OPG) is a member of the TNFR family that (the receptor of RANKL), recognize RANKL through cell-to-cell acts as a decoy receptor for RANKL to block the RANKL-RANK contact with osteoblasts, and differentiate into osteoclasts in the interaction (1, 2). In vitro studies have shown that the soluble form presence of M-CSF (5). of RANK, the extracellular domain of RANK, similarly inhibits RANKL has been shown to be proteolytically released from the RANKL-induced osteoclast differentiation (15). So far, 29 proteins plasma membrane (6–8). Recent studies have suggested that membrane-type matrix metalloproteinases (MT-MMPs) and TNF- have been identified as TNFR family members, all of which con- ␣-converting enzyme (TACE) are responsible for the release of tain cysteine-rich domains (CRDs) (16). The CRD is the most highly conserved region and is responsible for the ligand binding. The CRDs of OPG exhibit high to the CRDs *Institute for Oral Science, Matsumoto Dental University, Nagano, Japan; †Depart- of three members of the TNFR family: RANK, ment of Biochemistry, Matsumoto Dental University, Nagano, Japan; ‡Department of Pharmacology, School of Dentistry, Aichi Gakuin University, Aichi, Japan; and §In- (DcR3), and (DR6) (16, 17). The CRDs have been stitute of Molecular Biotechnology of the Austrian Academy of Sciences, Vienna, classified into several groups on the basis of their distinct crystal- Austria lized structure modules (see Fig. 2C) (16). The modular organiza- Received for publication October 7, 2005. Accepted for publication September tion in OPG is more similar to that in DR6 or DcR3 than that in 29, 2006. RANK. DcR3, which lacks the transmembrane domain, is a decoy 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 receptor for FasL (18). DcR3 induced osteoclastogenesis of mono- with 18 U.S.C. Section 1734 solely to indicate this fact. cytes and macrophages through TNF-␣ induction (19). DR6 was 1 This work was supported in part by Grants-in-Aid 16390535, 07791336, and identified by a homology search of an expressed sequence tag da- 17390497. tabase (20). The ligand for DR6 has not yet been identified. 2 Address correspondence and reprint requests to Dr. Naoyuki Takahashi, Institute for Ϫ/Ϫ Oral Science, Matsumoto Dental University, 1780 Gobara, Hiro-oka, Shiojiri, Nagano OPG-deficient (OPG ) mice exhibit severe osteoporosis due 399-0781, Japan. E-mail address: [email protected] to enhanced osteoclast differentiation and function (21, 22). Ϫ/Ϫ 3 Abbreviations used in this paper: RANKL, receptor activator of NF-␬B ligand; OPG mice also show accelerated bone resorption and forma- ␣ ␣ 1 ,25(OH)2D3,1 ,25-dihydroxyvitamin D3; MT-MMP, membrane-type matrix met- tion (21, 23). Juvenile Paget’s disease, an autosomal recessive os- alloproteinase; TACE, TNF-␣-converting enzyme; TIMP, tissue inhibitor of metal- loproteinase; OPG, osteoprotegerin; CRD, cysteine-rich domain; DcR3, decoy recep- teopathy, is characterized by rapidly remodeling woven bone. A tor 3; DR6, death receptor 6; WT, wild type; OVX, ovariectomized; siRNA, small homozygous deletion of the OPG gene was found in patients with interfering RNA; TAPI-2, TNF-␣ protease inhibitor-2. juvenile Paget’s disease and the serum level of free RANKL was Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$2.00 markedly elevated in one such patient (24). Serum levels of www.jimmunol.org The Journal of Immunology 193

Ϫ Ϫ RANKL were markedly elevated in OPG / mice as well. How- tioned medium was collected from cultures of osteoblasts and CD4ϩ T ever, it was recently reported that the serum concentration of cells and subjected to the ELISAs. RANKL-OPG complex is markedly higher than that of free RANKL or OPG in healthy human adults (25). This raises the RT-PCR analysis possibility that the presence of OPG may interfere with the quan- For semiquantitative RT-PCR analysis, total RNA was extracted using tification of RANKL level in ELISA. TRIzol reagent (Invitrogen Life Technologies). First-strand cDNA was synthesized from total RNA with oligo(dT) primer and Revertra Ace In the present study, we verified the above possibility using 12–18 Ϫ/Ϫ Ϫ/Ϫ (Toyobo) and subjected to PCR amplification with Ex Taq polymerase wild-type (WT), RANKL-deficient (RANKL ), and OPG (Takara). The specific primer pairs and temperature conditions were as Ϫ/Ϫ mice. We confirmed that the serum level of RANKL in OPG follows: for mouse RANKL cDNA amplification, forward: 5Ј-CGCTC mice is actually 20-fold higher than that of WT mice. We also TGTTCCTGTACTTTCGAGCG-3Ј/reverse: 5Ј-TCGTGCTCCCTCCTTT investigated how OPG and other OPG-related receptors are in- CATCAGGTT-3; denaturation at 94°C for 2 min followed by 32 cycles of 94°C for 30 s, 58°C for 45 s, 72°C for 1 min. For mouse OPG cDNA volved in the shedding of RANKL by cells expressing RANKL. amplification, forward: 5Ј-TGGAGATCGAATTCTGCTTG-3Ј/reverse: Our results suggest that OPG plays key roles in the shedding of 5Ј-TCAAGTGCTTGAGGGCATAC-3; denaturation at 94°C for 2 min fol- RANKL in mice. The physiological importance of this new role of lowed by 34 cycles of 94°C for 30 s, 50°C for 45 s, 72°C for 1 min. For OPG is discussed here. mouse TACE cDNA amplification, forward: 5Ј-AGCATGGACTCAG CATCTGTT-3Ј/reverse: 5Ј-TGCAGCTTGAATGAGGCCG-3; denatur- ation at 94°C for 2 min followed by 30 cycles of 94°C for 30 s, 50°C for Materials and Methods 30 s, 72°C for 1 min. For mouse 24-hydroxylase cDNA amplification, Mice forward: 5Ј-AAGATGATGGTGACCCCCGT-3Ј/reverse: 5Ј-AGCCCAG

Ϫ/Ϫ CACTTGGGTATTTA-3; denaturation at 94°C for 2 min followed by 30 OPG (C57BL/6) mice (22) and control WT mice were purchased from Downloaded from Ϫ Ϫ cycles of 94°C for 30 s, 58°C for 45 s, 72°C for 1 min. For mouse GAPDH CLEA Japan. RANKL / (C57BL/6) mice were generated in one of our Ј Ј nu/nu cDNA amplification, forward: 5 -ACCACAGTCCATGCCATCAC-3 / laboratories (26). C57BL/6 nude (Fox n1 ) mice were purchased from reverse: 5Ј-TCCACCACCCTGTTGCTGTA-3Ј; denaturation at 94°C Taconic Farms (27). All procedures for animal care were approved by the for 2 min followed by 20 or 22 cycles of 94°C for 30 s, 58°C for 45 s, animal management committee of Matsumoto Dental University. 72°C for 1 min. Cell cultures Northern blot analysis

Calvarial osteoblasts were prepared as described previously (28). Osteo- http://www.jimmunol.org/ ␣ Total RNA was prepared from bone, intestine, kidney, thymus, and spleen blasts were cultured in -MEM with 10% FBS in the presence or absence Ϫ/Ϫ Ϫ8 ␣ ϩ of 12-wk-old WT or OPG mice using TRIzol (Invitrogen Life Tech- of 10 M1 ,25(OH)2D3 (Wako). For T cell activation, CD4 T cells ␮ were isolated from the spleens of 8-wk-old WT mice using a MACS CD4ϩ nologies). Total RNA (20 g) was separated by electrophoresis on a 1.2% T cell isolation (Miltenyi Biotec). Purified CD4ϩ T cells were seeded agarose 2% formaldehyde gel and transferred onto a nylon membrane into 96-well anti-mouse CD3 Ab-coated plates or control plates (BD (Roche). The nylon membrane was then hybridized at 68°C in DIG-Easy Pharmingen) at 1 ϫ 105 cells/well and were cultured in RPMI 1640 me- Hyb buffer (Roche) with a digoxigenin-labeled antisense RNA probe en- dium with 10% FBS in the presence or absence of 2 ␮g/ml anti-mouse compassing the complete protein-coding sequence of mouse RANKL. Af- CD28 Ab (BD Pharmingen). ter hybridization, the membrane was washed at room temperature for 5 min in 2ϫ SSC, 0.1% SDS once, then washed at 68°C for 15 min in 0.1ϫ SSC, Fc-chimeric receptors and protease inhibitors 0.1% SDS twice. The specific bands for RANKL mRNA were immuno- stained with alkaline phosphatase-conjugated anti-digoxigenin Ab (Roche) by guest on October 1, 2021 The following Fc chimeric proteins of TNFR family members were used: and CDP-Star reagent (New England Biolabs). OPG-Fc (mouse OPG (Glu22)-Leu401) fused to the Fc region of human IgG1 (Pro100-Lys330) via a polypeptide linker), RANK (mouse RANK OPG knockdown by siRNA and re-expression by OPG 30 213 1 350 (Gln -Pro ))-Fc, DR6 (human DR6 (Met -Leu ))-Fc, DcR3 (human expression plasmid in calvarial osteoblasts DcR3 (Val24)-His300)-Fc, and control Fc. All recombinant Fc chimeric pro- teins were purchased from R&D Systems. The following protease inhibi- RNA interference-mediated knockdown of OPG protein expression was tors were used: aprotinin (Trasylol; Bayer), pepstatin (Peptide Institute), performed using Silencer predesigned small interfering RNA (siRNA) E-64 (Calbiochem), the hydroxamate inhibitor TAPI-2 (Calbiochem) (29), (Ambion). The specific RNA oligonucleotides for targeting OPG were as a thiadiazole urea inhibitor (MMP-3 inhibitor III; Calbiochem) (30), a syn- follows: sense, 5Ј-CGUGUGUUCCGGAAACAGATT-3Ј/antisense, 5Ј-UC thetic cyclic peptide inhibitor (H-Cys-Thr-Thr-His-Trp-Gly-Phe-Thr-Leu- UGUUUCCGGAACACACGTT-3Ј. The scrambled siRNA duplex was Cys-OH) for gelatinase (MMP2/9 inhibitor III; Calbiochem) (31), recom- used as a negative control (Ambion). WT calvarial osteoblasts were plated binant mouse TIMP-1 (R&D Systems), and recombinant mouse TIMP-2, in 12-well dishes at 105 cells/well for 24 h before transfection. Transfection -3 (Calbiochem). of the siRNA was accomplished with the HVJ Envelope Vector kit Geno- mONE Neo (Ishihara Sangyo) according to the manufacturer’s instruc- In vivo experiments tions. Forty-eight hours posttransfection, osteoblasts were collected for RNA isolation and the conditioned media were collected for the quantifi- To determine the origin of serum RANKL, OPG- and T cell-double-defi- Ϫ/Ϫ Ϫ Ϫ Ϫ Ϫ cation of OPG protein by ELISA. OPG re-expression in OPG osteo- cient (OPG / /Fox n1nu/nu) mice were created by cross-breeding OPG / blasts was accomplished by transient transfection of the pCMV-OPG mice and athymic nude (Fox n1nu/nu) mice (27). The double-deficient mice which encompassed the complete protein-coding sequence of mouse OPG. at the age of 8 days were sacrificed and blood was collected. To examine Ϫ Ϫ Transfection of the OPG expression plasmid was performed using the HVJ the relationship between age and serum RANKL level, WT and OPG / Envelope Vector kit GenomONE-Neo. Seventy-two hours posttransfec- mice at different ages (from 8-days old to 18-wk old) were sacrificed and tion, osteoblasts were collected for RNA isolation and the conditioned blood was collected. To examine the effects of estrogen deficiency on the Ϫ Ϫ media were collected for the quantification of OPG protein by ELISA. serum level of RANKL, 12-wk-old female WT and OPG / mice were ovariectomized (OVX) and sacrificed 4 wk later. To determine the effects ␣ ␣ Biotinylation of RANKL and analysis for RANKL disappearance of 1 ,25(OH)2D3 on the serum level of RANKL, 50 pM 1 ,25(OH)2D3 diluted in propylene glycol (0.1 ml/head/day) were injected i.p. into 10- in the culture medium of osteoblasts wk-old WT and OPGϪ/Ϫ mice daily for 2 days. On day 3, the mice were One microgram of mouse RANKL (R&D Systems) was dissolved in 0.5 ml sacrificed and blood was collected. Femora, tibiae, spleen, and thymus ␮ were also removed from each mouse. The diaphyses of femora and tibiae of 40 mM NaHCO3 (pH 8.0) containing 14 l of 5 mg/ml sulfosuccinimi- were prepared as bone samples. These tissues were used for semiquanti- dyl-D-biotin (Biotinylation kit; Dojin Laboratories), and incubated at 25°C tative RT-PCR analysis. for 2 h. Free sulfosuccinimidyl-D-biotin was removed from the reaction mixture using a gel-filtration column according to the manufacturer’s in- Ϫ/Ϫ Measurements of RANKL and OPG structions (Dojin Laboratories). WT and OPG osteoblasts were plated in a 96-well plate at 104 cells/well. After culture for 24 h, the medium was Serum levels of RANKL and OPG were determined by performing the changed to the fresh medium containing 50 ng/ml biotinylated RANKL respective ELISAs (Quantikine M ELISA kit; R&D Systems). The condi- (100 ␮l/well). After incubation for 48 h, the culture medium was collected 194 THE REGULATION OF SERUM RANKL LEVELS BY OPG

FIGURE 1. Evidence that OPG deficiency markedly elevated the circulating level of RANKL in mice. A, Serum concentrations of RANKL and OPG in WT, OPGϪ/Ϫ, and RANKLϪ/Ϫ mice at the age of 4 wk. Serum obtained from WT, OPGϪ/Ϫ, and RANKLϪ/Ϫ mice was subjected to RANKL and OPG Significantly different between the two groups; p Ͻ 0.01. n.d., Not detectable. B, Effects ,ء .ELISAs. Data are expressed as the mean Ϯ SD of four mice of OPG-Fc on the measurement of serum RANKL levels of OPGϪ/Ϫ mice. Serum obtained from OPGϪ/Ϫ mice was incubated with increasing concen- trations of OPG-Fc for 2 days. Concentrations of RANKL in the serum were determined by ELISA. C, Effects of rRANKL on the measurement of serum OPG levels of RANKLϪ/Ϫ mice. Serum obtained from RANKLϪ/Ϫ mice was incubated with increasing concentrations of RANKL for 2 days. Concen- trations of OPG in the serum were determined by ELISA. Downloaded from and subjected to SDS-PAGE, followed by blotting to a clear blot mem- Four RANKL mRNA species were observed in thymus and spleen: brane-P (Atto). The biotinylated RANKL was detected with streptavidin- a major transcript of 2.4 kb and three minor transcripts of 10.4, 6.0, HRP (Roche) using the ECL Plus detection system (GE Healthcare). and 0.5 kb (Fig. 2). In agreement with the previous finding (32), Results kidney and intestine scarcely expressed RANKL mRNA. Bone exclusively expressed the 2.4-kb transcript. The expression levels

OPG deficiency results in markedly elevated circulating levels Ϫ Ϫ http://www.jimmunol.org/ and patterns of RANKL transcripts in OPG / mice were quite of RANKL in mice similar to those in WT mice (see Fig. 5A). We previously reported that OPGϪ/Ϫ mice showed marked ele- vated serum levels of RANKL compared with WT mice (23). Regulation of RANKL mRNA expression in osteoblasts However, it was reported that the serum concentration of the RANKL is highly expressed in osteoblasts and lymphoid cells RANKL-OPG complex is extremely high in healthy human adults (14). We examined whether OPG is involved in the shedding of (25). This raises the possibility that the presence of OPG may RANKL from osteoblasts. Osteoblasts obtained from OPGϪ/Ϫ interfere with the quantification of RANKL in ELISA. We first mice spontaneously released RANKL even in the absence of any measured the concentrations of RANKL and OPG in serum from

osteotropic factors. The concentration of RANKL was signifi- by guest on October 1, 2021 WT, OPGϪ/Ϫ, and RANKLϪ/Ϫ mice in the respective ELISA kits ␣ cantly increased following the addition of 1 ,25(OH)2D3 (Fig. Ϫ/Ϫ (Fig. 1A). The serum level of RANKL in OPG mice was 20- 3A). RT-PCR analysis showed that the mRNA expression of fold higher than that in WT mice. In contrast, the serum level of RANKL in osteoblasts was also up-regulated by the addition of OPG in RANKLϪ/Ϫ mice was equivalent to that in WT mice. This ␣ ␣ 1 ,25(OH)2D3, concomitantly with the up-regulation of 1 ,25 indicated that the presence of RANKL in serum did not interfere (OH)2D3-24-hydroxylase mRNA expression (Fig. 3B). The ex- with OPG measurement. To examine whether OPG interferes with pression of TACE, which has been proposed to be an enzyme the quantification of RANKL in this ELISA, the serum obtained responsible for the shedding of RANKL, was not regulated by from OPGϪ/Ϫ mice was incubated with increasing concentrations ␣ 1 ,25(OH)2D3. Northern blot analysis confirmed that WT and of mouse rOPG-Fc for 2 days and subjected to RANKL measure- OPGϪ/Ϫ osteoblasts exclusively expressed a 2.4-kb transcript of ment (Fig. 1B). RANKL measurements were barely affected by the presence of physiological concentrations of rOPG (2–5 ng/ml), but 70% interference was observed at 10 ng/ml OPG. When OPGϪ/Ϫ mouse serum was incubated with a high concentration of OPG-Fc (200 ng/ml), the observed RANKL level was almost equivalent to that in WT mouse serum (Fig. 1B). RANKLϪ/Ϫ mouse serum was also incubated with increasing concentrations of mouse rRANKL for 2 days and subjected to OPG measurements. The quantification of serum OPG was not affected even at a high concentration of rRANKL (50 ng/ml) (Fig. 1C). These results suggest that the se- rum concentrations of OPG and RANKL in WT mice shown in Fig. 1A closely reflect the true values.

Effects of OPG gene inactivation on expression levels and patterns of RANKL transcripts in vivo OPG withdrawal may affect the expression level of RANKL FIGURE 2. Effects of OPG gene inactivation on expression levels and mRNA or induce a short RANKL mRNA transcript encoding a patterns of RANKL transcripts. Total RNA was isolated from bone, intes- soluble form of RANKL. The expression of RANKL mRNA in tine, kidney, thymus, and spleen of 12-wk-old WT and OPGϪ/Ϫ mice, and Ϫ/Ϫ various tissues from WT and OPG mice was analyzed by analyzed by Northern blot hybridization to detect RANKL transcripts. The Northern blotting using an antisense RNA probe encompassing the positions of 28S and 18S rRNA are noted. The lower panel of ethidium complete protein-coding sequence of mouse RANKL (Fig. 2). bromide (EtBr) staining shows the amount of RNA in each lane. The Journal of Immunology 195

RANKL mRNA, in which expression was up-regulated by the ad- ␣ ␣ dition of 1 ,25(OH)2D3 (Fig. 3C). Treatment with 1 ,25(OH)2D3 did not induce other sizes of RANKL transcripts in osteoblasts. Changes in OPG production in osteoblasts may alter the expres- sion of RANKL. We performed both OPG knockdown by siRNA in WT osteoblasts and OPG re-expression by an OPG expression plasmid in OPGϪ/Ϫ osteoblasts. Transfection of 250 pM siRNA for the OPG sequence into WT osteoblasts led to an ϳ70% reduc- tion in the OPG protein (Fig. 3D). However, RANKL mRNA ex- pression was not affected by the OPG knockdown. Transfection of an OPG expression plasmid in OPGϪ/Ϫ osteoblasts dose-depen- dently increased OPG secretion into the culture medium, but had no effect on RANKL mRNA expression (Fig. 3E). Thus, RANKL mRNA expression was not affected by the changes in OPG pro- duction. OPG has been shown to bind to cell surface heparan sul- fates and to be internalized (33). Tat et al. (34) also reported that OPG accelerated internalization of RANKL. We examined the possibility that RANKL internalization is impaired in OPGϪ/Ϫ os- teoblasts and soluble RANKL accumulates in the culture medium. Biotinylated soluble RANKL was added for 48 h to WT and Downloaded from OPGϪ/Ϫ osteoblast cultures, and the loss of biotinylated RANKL in the culture medium was determined by SDS-PAGE, followed by Western blotting (Fig. 3F). Similar levels of biotinylated RANKL were detected in cultures of WT and OPGϪ/Ϫ osteoblasts (Fig. 3F). These results suggest that the elevated RANKL level in cul- ture medium of OPGϪ/Ϫ osteoblasts is due to the shedding from http://www.jimmunol.org/ the osteoblasts.

Inhibition of the shedding of RANKL from osteoblasts by OPG and RANK The amino acid sequence of the CRDs in OPG shows high ho- mology to that of the CRDs in three TNFR members: RANK, DcR3, and DR6 (Fig. 4A) (16). We next examined the effects of

OPG and these OPG-related receptors on the shedding of RANKL by guest on October 1, 2021

the osteoblasts, and analyzed for the expression of RANKL, TACE, 24- hydroxylase, and GAPDH mRNAs by semiquantitative RT-PCR. ID no., The identification number of the mice examined. C, Northern blot analysis of RANKL mRNA expression in WT and OPGϪ/Ϫ osteoblasts. WT and OPGϪ/Ϫ osteoblasts were cultured for 16 h in a 6-cm dish (3 ϫ 105 cells) ␣ Ϫ8 in the presence or absence of 1 ,25(OH)2D3 (10 M). Total RNA was isolated from the osteoblasts and analyzed by Northern blot hybridization. The positions of 28S and 18S rRNA are noted. The lower panel of ethidium bromide (EtBr) staining shows the amount of RNA in each lane. D, Effects of OPG knockdown by siRNA on RANKL mRNA expression in WT os- teoblasts. WT osteoblasts cultured in a 12-well plate were transfected with various amounts of siRNA targeting OPG (50 Ϫ500 pM/well) or a scram- bled control siRNA. After 48 h posttransfection, cells were harvested and analyzed by semiquantitative RT-PCR for OPG and RANKL mRNA levels (upper panel). Concentrations of OPG in the culture medium were deter- mined by ELISA (lower panel). E, Effects of OPG re-expression by an OPG expression plasmid on RANKL mRNA expression in OPGϪ/Ϫ os- teoblasts. OPGϪ/Ϫ osteoblasts cultured in a 12-well plate were transfected with various amounts of an OPG expression plasmid (0–60 ␮g/well) or a FIGURE 3. Regulation of RANKL mRNA expression in osteoblasts. A, GFP expression plasmid as a negative control. Seventy-two hours post- Spontaneous release of RANKL into culture medium from OPGϪ/Ϫ os- transfection, cells were harvested and analyzed by semiquantitative RT- teoblasts. Calvarial osteoblasts from OPGϪ/Ϫ mice were cultured for 72 h PCR for OPG and RANKL mRNA levels (upper panel). Concentrations of in 6-cm diameter dishes (3 ϫ 105cells) in the presence or absence of OPG in the culture medium were determined by ELISA (lower panel). F, ␣ Ϫ8 1 ,25(OH)2D3 (10 M). Concentrations of RANKL in the culture me- Effects of OPG on RANKL loss in culture medium of osteoblasts. WT and dium were determined by ELISA. Data are expressed as the mean Ϯ SD of OPGϪ/Ϫ osteoblasts were plated in a 96-well plate at 1 ϫ 104 cells/well -Significantly different between the two groups; p Ͻ 0.01. and cultured in ␣-MEM containing 10% FBS (0.1 ml/well). After incuba ,ء .four cultures B, Expression of RANKL, TACE, and 24-hydroxylase mRNAs in OPGϪ/Ϫ tion for 24 h, the medium was changed to the fresh medium containing ␣ Ϫ/Ϫ osteoblasts treated with or without 1 ,25(OH)2D3. OPG osteoblasts biotinylated recombinant soluble RANKL (50 ng/ml). After incubation for were cultured for 16 h in 6-cm diameter dishes (3 ϫ 105cells) in the pres- 48 h, the culture media were subjected to SDS-PAGE, followed by West- ␣ Ϫ8 ence or absence of 1 ,25(OH)2D3 (10 M). Total RNA was isolated from ern blotting. Biotinylated RANKL was detected with streptavidin-HRP. 196 THE REGULATION OF SERUM RANKL LEVELS BY OPG Downloaded from http://www.jimmunol.org/

FIGURE 5. Effects of OPG and RANK on the shedding of RANKL by activated T cells. A, Release of RANKL from activated T cells into the culture medium. CD4ϩ T cells were isolated from spleens of WT mice, and ␣

cultured in the presence or absence of soluble anti-CD28 Ab ( CD28) in by guest on October 1, 2021 96-well plates (1 ϫ 105 cells/well) that had been precoated with anti-CD3 Ab (␣CD3). After incubation for 96 h, the culture medium was collected, and concentrations of RANKL were determined by ELISA. n.d., not de- tectable. B, Expression of RANKL, OPG, TACE, and GAPDH mRNAs in FIGURE 4. Effects of OPG and OPG-related receptors on the shedding of T cells cultured with or without an anti-CD28 Ab and/or anti-CD3 Ab. Ϫ Ϫ RANKL by OPG / osteoblasts. A, The modular structures of the CRDs of CD4ϩ T cells (1 ϫ 105 cells/well) were cultured for 16 h in 96-well plates OPG and OPG-related receptors. The amino acid sequence of the CRDs precoated with anti-CD3 Ab (␣CD3). Some cultures were treated with of mouse OPG (mOPG) shows high homology to that of the CRDs of anti-CD28 Ab (␣CD28). Total RNA was isolated from T cells and ana- mouse RANK (mRANK), human DR6 (hDR6), and human DcR3 lyzed for the expression of RANKL, OPG, TACE, and GAPDH mRNAs (hDcR3). The crystallized modules A1, A2, B1, and B2 in the CRDs are by semiquantitative RT-PCR. C, Evaluation of interference by OPG and pattern-coded as shown in the inset. B, Evaluation of interference by OPG RANK in the RANKL assay. RANKL at 500 pg/ml was measured in an and RANK in the RANKL assay. RANKL at 50 pg/ml was measured in an ELISA system in the presence of increasing concentrations of OPG-Fc and ELISA system in the presence of increasing concentrations of OPG-Fc and RANK-Fc. The results are expressed as the ratio of the observed values to RANK-Fc. The results are expressed as the ratio of the observed values to the expected value (500 pg/ml). D, Effects of OPG-Fc and RANK-Fc on the expected value (50 pg/ml). C, Effects of OPG-Fc, RANK-Fc, DcR3-Fc, the shedding of RANKL by activated T cells. CD4ϩ T cells were cultured Ϫ Ϫ and DR6-Fc on the shedding of RANKL from OPG / osteoblasts. in an anti-CD3 Ab-coated dish and costimulated with anti-CD28 Ab for Ϫ Ϫ OPG / osteoblasts were cultured for 96 h in 24-well plates (3 ϫ 104 96 h. Various concentrations of Fc-chimeric proteins were added to the T cells/well) in the presence of various concentrations of OPG-Fc and OPG- cell cultures. Concentrations of RANKL in the culture medium were de- related receptor-Fcs. Concentrations of RANKL in the culture medium termined by ELISA, and corrected according to the interference curves were determined by ELISA, and corrected according to the interference shown in C. Data are expressed as the mean Ϯ SD from four cultures. Significantly different from the control cultures treated with IgG-Fc; p Ͻ ,ء .curves shown in B. Data are expressed as the mean Ϯ SD of four cultures .Significantly different from the control cultures treated with IgG-Fc; p Ͻ 0.001 ,ء 0.001. n.d., not detectable.

suggesting that OPG-Fc similarly behaves like a native protein. by OPGϪ/Ϫ osteoblasts. Before measuring the effects of these mol- OPG-Fc at concentrations higher than 2 ng/ml and RANK-Fc at 50 ecules on the shedding of RANKL, the interference by OPG, ng/ml strongly interfered with the assay of RANKL at 50 pg/ml RANK, DcR3, and DR6 in the RANKL assay was evaluated in the (Fig. 4B). In contrast, DR6-Fc and DcR3-Fc even at 1,000 ng/ml presence of RANKL at 50 pg/ml, because the concentration of barely interfered with the RANKL assay (data not shown). We RANKL in the culture medium of OPGϪ/Ϫ osteoblasts was 50– therefore examined the effects of OPG-Fc at 0.5–2 ng/ml and 100 pg/ml (Fig. 3A). It was reported that OPG exists and acts as a RANK-Fc at 1–10 ng/ml on the shedding of RANKL from homodimeric form (35). OPG-Fc also exists as a homodimer (35), OPGϪ/Ϫ osteoblasts (Fig. 4C). The concentrations of RANKL in The Journal of Immunology 197

FIGURE 6. Effects of various types of protease inhibitors on the shedding of RANKL by OPGϪ/Ϫ osteoblasts. Effects of (A) inhibitors of serine, aspartate, and cysteine proteases (aprotinin, pepstatin A, and E-64), (B) inhibitors of metalloproteinases (TAPI-2, MMP2/9 inhibitor III, and MMP3 inhibitor III), and (C) tissue inhibitor of metalloproteinases (TIMP-1, TIMP-2 and TIMP-3) on the shedding of RANKL were examined in cultures of OPGϪ/Ϫ osteoblasts. OPGϪ/Ϫ osteoblasts were cultured for 96 h in 24-well plates (3 ϫ 104 cells/well) in the presence or absence of several types of protease inhibitors at the concentrations indicated. Concentrations of RANKL in the culture medium were determined by ELISA. Data are expressed as the mean Ϯ .p Ͻ 0.01 ,ء ;Significantly different from the control cultures ,ء .SD from four cultures Downloaded from the culture medium were corrected on the basis of the respective the presence or absence of several kinds of protease inhibitors (Fig. interference curves shown in Fig. 4B. OPG-Fc dose-dependently 6). The release of RANKL from OPGϪ/Ϫ osteoblasts was not in- suppressed the release of RANKL from OPGϪ/Ϫ osteoblasts. The hibited by aprotinin (a serine protease inhibitor), pepstatin A (an concentration of RANKL was decreased to an undetectable level aspartate protease inhibitor), or E-64 (a cysteine protease inhibi- in the presence of 2 ng/ml OPG. RANK-Fc weakly inhibited the tor), when these agents were added at 10 times higher concentra- http://www.jimmunol.org/ shedding of RANKL compared with OPG-Fc. Neither DR6-Fc nor tions than those used for the ordinary preparation of cell lysates DcR3-Fc inhibited the shedding of RANKL by OPGϪ/Ϫ osteo- (Fig. 6A). Mouse osteoblasts are known to express MMP-2, blasts even at higher concentrations (Fig. 4C). MMP-3, MMP-13, and MT1-MMP (9, 10). TNF-␣ protease in- hibitor-2 (TAPI-2), a broad spectrum inhibitor of MMPs, MT- Inhibition of the shedding of RANKL from activated T cells by MMPs, and TACE (29), strongly inhibited the shedding of OPG and RANK RANKL, whereas MMP2/9 inhibitor III (31) and MMP3 inhibitor Activated CD4ϩ T cells have been shown to release excess III (30) did not (Fig. 6B). The biological activity of MMPs is amounts of RANKL (13). Therefore, we hypothesized that acti- suppressed by TIMPs, which are also expressed by osteoblasts. by guest on October 1, 2021 vated T cells release RANKL, because hemopoietic cells, includ- TIMP-2 and TIMP-3 inhibited the release of RANKL, but TIMP-1 ing T cells, barely produce OPG. When T cells obtained from WT did not (Fig. 6C). MT-MMPs are selectively inhibited by TIMP-2 mice were activated by plate-bound anti-CD3 Ab together with and TIMP-3 but not by TIMP-1 (36–38). TACE is selectively soluble anti-CD28 Ab, the concentration of RANKL in the culture inhibited by TIMP-3 but not TIMP-1 or TIMP-2 (39). These find- medium was markedly elevated (Fig. 5A). RT-PCR analysis con- ings suggest that MT-MMPs and TACE, but not serine proteases, firmed that activated T cells strongly expressed RANKL mRNA aspartate proteases, or cysteine proteases are involved in the shed- Ϫ Ϫ and weakly expressed OPG mRNA (Fig. 5B). The expression of ding of RANKL in OPG / osteoblasts. TACE mRNA was also up-regulated by T cell activation (Fig. 5B). We then examined whether OPG-Fc and RANK-Fc could suppress the release of RANKL by activated T cells. The OPG and RANK in the RANKL assay was evaluated in the presence of RANKL at 500 pg/ml, because the conditioned medium of activated T cells contained nearly 500 pg/ml RANKL (Fig. 5A). OPG-Fc at higher than 20 ng/ml and RANK-Fc at higher than 50 ng/ml strongly interfered with the RANKL assay at 500 pg/ml RANKL (Fig. 5C). Accordingly, OPG-Fc and RANK-Fc at concentrations up to 20 and 50 ng/ml, respectively, were used in the experiments with T cell cultures. OPG-Fc and RANK-Fc dose-dependently suppressed the shedding of RANKL by activated T cells (Fig. 5D), but DR6-Fc and DcR3-Fc did not (data not shown). The inhibitory effect of OPG-Fc on the shedding of RANKL was stronger than that of RANK-Fc in cultures of activated T cells. FIGURE 7. Effects of T cell deficiency on serum RANKL levels in Involvement of MT-MMPs and TACE in the shedding of RANKL OPGϪ/Ϫ mice. Serum concentrations of RANKL in OPG-and T cell dou- by osteoblasts ble-deficient mice. OPG-and T cell double-deficient (OPGϪ/Ϫ/Fox n1nu/nu) mice were produced by cross-breeding of OPGϪ/Ϫ mice and athymic nude It is difficult to characterize the enzymes involved in the shedding (Fox n1nu/nu) mice. Blood was collected from WT (OPGϩ/ϩ/Fox n1ϩ/ϩ) of RANKL from osteoblasts because WT osteoblasts spontaneously mice, athymic nude (OPGϩ/ϩ/Fox n1nu/nu) mice, OPGϪ/Ϫ (OPGϪ/Ϫ/Fox release a large amount of OPG, which inhibits the shedding of n1ϩ/ϩ) mice, and OPG- and T cell double-deficient (OPGϪ/Ϫ/Fox n1nu/nu) RANKL. To identify the proteases responsible for the shedding of mice at the age of 8 days. Serum concentrations of RANKL were measured Ϫ Ϫ RANKL from osteoblasts, OPG / osteoblasts were cultured in by ELISA. Data are expressed as the mean Ϯ SD from four mice. 198 THE REGULATION OF SERUM RANKL LEVELS BY OPG

FIGURE 8. Origin of serum RANKL in OPGϪ/Ϫ mice. A, Effects of OVX on serum RANKL levels in WT and OPGϪ/Ϫ mice. Female WT and OPGϪ/Ϫ mice at the age of 12 wk were OVX or sham operated. Blood was collected from those mice 4 wk after surgery. Serum con- centrations of RANKL were measured by ELISA. Data -Sig ,ء .are expressed as the mean Ϯ SD from four mice nificantly different from the sham-operated controls; p Ͻ 0.01. B, Age-dependent changes in serum RANKL levels in OPGϪ/Ϫ mice. Serum concentrations of RANKL were measured by ELISA. Data are expressed as the mean Ϯ ␣ SD of four mice. C, Effects of 1 ,25(OH)2D3 administra- tion on serum RANKL levels in WT and OPGϪ/Ϫ mice. WT and OPGϪ/Ϫ mice at the age of 10 wk were injected ␣ i.p. with 1 ,25(OH)2D3 (50 pM/head/day) for 2 days. On day 3, blood was collected from the mice by heart punc- ,ء ;Significantly differences between two groups ,ء .ture Ͻ ␣ p 0.001. D, Effects of 1 ,25(OH)2D3 on RANKL and OPG mRNA expression in bone, thymus, and spleen were determined by semiquantitative RT-PCR. Total RNA was

analyzed for the expression of RANKL and OPG mRNAs Downloaded from by semiquantitative RT-PCR. Data represent typical RANKL mRNA expression in four individual WT and OPGϪ/Ϫ mice (upper) and OPG mRNA expression in six individual WT mice (lower). ID no., The identification number of mice examined. http://www.jimmunol.org/

Tissues responsible for the elevated RANKL in OPG Ϫ/Ϫ mice Discussion Ϫ/Ϫ Both OPGϪ/Ϫ osteoblasts and activated T cells released a large We previously reported that OPG mice showed marked ele- amount of RANKL in the culture medium and OPG-Fc effectively vated serum levels of RANKL compared with WT mice (23). suppressed the release of RANKL from both types of cells. We However, the recent study demonstrated the possibility that the examined which tissues, bone cells or T cells, were responsible for presence of OPG interferes with the quantification of the serum the elevated RANKL in OPGϪ/Ϫ mice. OPG- and T cell-double- level of RANKL in humans (25). Using serum obtained from WT, Ϫ Ϫ Ϫ Ϫ deficient (OPGϪ/Ϫ/Fox n1nu/nu) mice were created by cross-breed- RANKL / , and OPG / mice, we examined this possibility. 1) Ϫ Ϫ by guest on October 1, 2021 ing of OPGϪ/Ϫ mice and athymic nude (Fox n1nu/nu) mice (Fig. 7). The serum level of OPG in RANKL / mice was equivalent to Serum levels of RANKL in OPG- and T cell-double deficient that in WT mice. 2) RANKL measurements were barely affected (OPGϪ/Ϫ/Fox n1nu/nu) mice were as high as those in OPGϪ/Ϫ by the presence of physiological concentrations of rOPG (2–5 ng/ Ϫ Ϫ (OPGϪ/Ϫ/Fox n1ϩ/ϩ) mice. Estrogen-deficiency caused by OVX ml). 3) Experiments of mixture of serum from OPG / mice and induces a high turnover state of bone with increased bone formation WT mice showed that circulating OPG in WT mice did not inter- and resorption, and that the bone turnover rate decreases with age in fere with the RANKL assay (data not shown). These results sug- mice (40). Circulating levels of RANKL in OPGϪ/Ϫ mice were sig- gest that the circulating level of RANKL is actually elevated in Ϫ Ϫ nificantly increased by OVX (Fig. 8A). Serum levels of RANKL in OPG / mice. OPGϪ/Ϫ mice were gradually decreased with age (Fig. 8B). These Northern blot analysis showed that the expression levels and Ϫ Ϫ results suggest that the high levels of serum RANKL detected in patterns of RANKL transcripts in several tissues in OPG / mice OPGϪ/Ϫ mice were derived not from T cells but from bone tissues. were similar to those in WT mice. Neither OPG knockdown by ␣ When 1 ,25(OH)2D3 (50 pM/head/day) was administered i.p. to WT siRNA in WT osteoblasts nor OPG re-expression by an OPG ex- Ϫ Ϫ and OPGϪ/Ϫ mice for 2 days, serum concentrations of RANKL in pression plasmid in OPG / osteoblasts alters RANKL mRNA ␣ ␣ WT mice were significantly increased in response to 1 ,25(OH)2D3 expression. Treatment of osteoblasts with 1 ,25(OH)2D3 up-reg- administration but were still lower than those in untreated OPGϪ/Ϫ ulated the expression of the 2.4-kb RANKL transcript but did not ␣ Ϫ/Ϫ mice. Administration of 1 ,25(OH)2D3 to OPG mice further el- induce the other sizes of RANKL transcripts. In addition, disap- Ϫ Ϫ evated the increased concentration of serum RANKL (Fig. 8C). Con- pearance of biotinylated RANKL in culture medium of OPG / ␣ comitantly, 1 ,25(OH)2D3 induced hypercalcemia in mice with both osteoblasts was similar to that of WT osteoblasts. These results genotypes (data not shown), suggesting that serum levels of RANKL suggest that neither the transcriptional events nor the decrease in reflected bone resorption. In contrast to RANKL, serum concentra- RANKL degradation is involved in the elevated RANKL level in ␣ Ϫ/Ϫ Ϫ/Ϫ tions of OPG in WT mice were not affected by 1 ,25(OH)2D3 ad- serum in OPG mice and in culture medium of OPG osteo- ministration (data not shown). Semiquantitative RT-PCR showed that blasts. Thus, it was proposed that OPG may prevent the shedding the thymus and spleen of WT and OPGϪ/Ϫ mice expressed RANKL of RANKL-expressing cells. mRNA, but the expression levels were much lower than those in bone Using OPGϪ/Ϫ mice, we investigated 1) whether OPG inhibits (Fig. 8D, upper). The expression of RANKL mRNA in bone but not the shedding of RANKL, 2) which proteases are involved in the ␣ thymus or spleen was up-regulated by 1 ,25(OH)2D3 administration. shedding of RANKL, 3) which organ is responsible for the ele- OPG mRNA was highly expressed in bone but not in thymus or vated serum level of RANKL, and 4) whether serum RANKL lev- ␣ Ϫ/Ϫ spleen in WT mice (Fig. 8D, lower). 1 ,25(OH)2D3 administration els reflect the state of bone resorption. OPG osteoblasts spon- failed to down-regulate the OPG mRNA expression in bone (Fig. 8D, taneously released RANKL into the culture medium. Activated T lower). cells obtained from WT mice also released a large amount of The Journal of Immunology 199

RANKL. WT osteoblasts but not activated T cells secreted OPG in 3. Lacey, D. L., E. Timms, H. L. Tan, M. J. Kelley, C. R. Dunstan, T. Burgess, the culture medium. Soluble RANK also inhibited the shedding of R. Elliott, A. Colombero, G. Elliott, S. Scully, et al. 1998. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 93: Ϫ/Ϫ RANKL in OPG osteoblasts. Similarly, the release of RANKL 165–176. from activated T cells was inhibited by the addition OPG or sol- 4. Yasuda, H., N. Shima, N. Nakagawa, K. Yamaguchi, M. Kinosaki, S. Mochizuki, A. Tomoyasu, K. Yano, M. Goto, A. Murakami, et al. 1998. Osteoclast differ- uble RANK. These results suggest that activated T cells release entiation factor is a ligand for osteoprotegerin/osteoclastogenesis-inhibitory fac- RANKL, because they do not produce OPG, and that the binding tor and is identical to TRANCE/RANKL. Proc. Natl. Acad. Sci. USA 95: of OPG and soluble RANK to RANKL is important for the sup- 3597–3602. 5. Suda, T., N. Takahashi, N. Udagawa, E. Jimi, M. T. Gillespie, and T. J. Martin. pression of the shedding of RANKL. MT-MMPs and TACE were 1999. Modulation of osteoclast differentiation and function by the new members Ϫ/Ϫ suggested to be involved in the shedding of RANKL in OPG of the tumor necrosis factor receptor and ligand families. Endocr. Rev. 20: osteoblasts. The expression of TACE mRNA as well as RANKL 345–357. 6. Nakashima, T., Y. Kobayashi, S. Yamasaki, A. Kawakami, K. Eguchi, H. Sasaki, mRNA was up-regulated in activated T cells, indicating that TACE and H. Sakai. 2000. Protein expression and functional difference of membrane- is especially important for the shedding of RANKL in activated T bound and soluble receptor activator of NF-␬B ligand: modulation of the expres- cells. Neither DR6 nor DcR3 showed any effect on the shedding of sion by osteotropic factors and cytokines. Biochem. Biophys. Res. Commun. 275: 768–775. RANKL by osteoblasts or activated T cells. These results suggest 7. Schlondorff, J., L. Lum, and C. P. Blobel. 2001. Biochemical and pharmacolog- that the binding of OPG or soluble RANK to the membrane-asso- ical criteria define two shedding activities for TRANCE/OPGL that are distinct ciated RANKL in osteoblasts and T cells protects RANKL against from the tumor necrosis factor ␣ convertase. J. Biol. Chem. 276: 14665–14674. 8. Lum, L., B. R. Wong, R. Josien, J. D. Becherer, H. Erdjument-Bromage, attack by some proteases. J. Schlondorff, P. Tempst, Y. Choi, and C. P. Blobel. 1999. Evidence for a role In agreement with the previous study (32), the RANKL tran- of a tumor necrosis factor-␣ (TNF-␣)-converting enzyme-like protease in shed- scripts were found in bone, thymus, and spleen but not in intestine ding of TRANCE, a TNF family member involved in osteoclastogenesis and

J. Biol. Chem. Downloaded from Ϫ/Ϫ dendritic cell survival. 274: 13613–13618. and kidney in WT and OPG mice. Activated T cells released a 9. Uchida, M., M. Shima, T. Shimoaka, A. Fujieda, K. Obara, H. Suzuki, Y. Nagai, large amount of RANKL. These findings raised the possibility that T. Ikeda, H. Yamato, and H. Kawaguchi. 2000. Regulation of matrix metallo- Ϫ/Ϫ proteinases (MMPs) and tissue inhibitors of metalloproteinases (TIMPs) by bone not all but some of the soluble RANKL found in OPG mice resorptive factors in osteoblastic cells. J. Cell. Physiol. 185: 207–214. Ϫ/Ϫ originated from T cells. However, our experiments using OPG 10. Breckon, J. J., S. Papaioannou, L. W. Kon, A. Tumber, R. M. Hembry, and T cell-deficient mice suggested that T cells are not involved in G. Murphy, J. J. Reynolds, and M. C. Meikle. 1999. Stromelysin (MMP-3) syn- Ϫ/Ϫ thesis is up-regulated in estrogen-deficient mouse osteoblasts in vivo and in vitro. the elevated levels of RANKL in serum in OPG mice. The J. Bone Miner. Res. 14: 1880–1890. Ϫ/Ϫ http://www.jimmunol.org/ serum level of RANKL in OPG mice was closely correlated 11. Stetler-Stevenson, M., A. Mansoor, M. Lim, P. Fukushima, J. Kehrl, G. Marti, ␣ K. Ptaszynski, J. Wang, and W. G. Stetler-Stevenson. 1997. Expression of matrix with the state of bone resorption: OVX and 1 ,25(OH)2D3 admin- Ϫ/Ϫ metalloproteinases and tissue inhibitors of metalloproteinases in reactive and neo- istration increased the serum level of RANKL in OPG mice. plastic lymphoid cells. Blood 89: 1708–1715. The expression of RANKL mRNA in bone but not thymus or 12. Weitzmann, M. N., S. Cenci, L. Rifas, J. Haug, J. Dipersio, and R. Pacifici. 2001. spleen was up-regulated by 1␣,25(OH) D administration. Quinn T cell activation induces human osteoclast formation via receptor activator of 2 3 nuclear factor ␬B ligand-dependent and -independent mechanisms. J. Bone et al. (41) reported that cultured fibroblastic cells from skin, lung, Miner. Res. 16: 328–337. and respiratory diaphragm expressed RANKL mRNA in response 13. Kong, Y. Y., U. Feige, I. Sarosi, B. Bolon, A. Tafuri, S. Morony, C. Capparelli, to 1␣,25(OH) D . Therefore, the possibility that tissues other than J. Li, R. Elliott, S. McCabe, et al. 1999. Activated T cells regulate bone loss and 2 3 joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 402: bone such as skin and lung are responsible for the elevated serum 304–309. by guest on October 1, 2021 Ϫ Ϫ level of RANKL in OPG / mice cannot be excluded. 14. Theill, L. E., W. J. Boyle, and J. M. Penninger. 2002. RANK-L and RANK: T OPGϪ/Ϫ mice exhibit stimulated osteoclastic bone resorption cells, bone loss, and mammalian evolution. Annu. Rev. Immunol. 20: 795–823. Ϫ/Ϫ 15. Jimi, E., S. Akiyama, T. Tsurukai, N. Okahashi, K. Kobayashi, N. Udagawa, (21–23). Cross-sections of femora of OPG mice showed ex- T. Nishihara, N. Takahashi, and T. Suda. 1999. Osteoclast differentiation factor cessive osteoclastic bone resorption even in cortical bones (22, acts as a multifunctional regulator in murine osteoclast differentiation and func- 23). OPG deficiency in humans results in juvenile Paget’s disease tion. J. Immunol. 163: 434–442. 16. Bodmer, J. L., P. Schneider, and J. Tschopp. 2002. The molecular architecture of characterized by rapidly bone remodeling (24). Thus, OPG is a the TNF superfamily. Trends Biochem. Sci. 27: 19–26. critical regulator of bone homeostasis in humans as well as mice. 17. Bridgham, J. T., and A. L. Johnson. 2003. Characterization of chicken TNFR Transgenic mice overexpressing soluble RANKL in the liver were superfamily decoy receptors, DcR3 and osteoprotegerin. Biochem. Biophys. Res. Commun. 307: 956–961. generated and exhibited accelerated bone resorption (42). We in- 18. Pitti, R. M., S. A. Marsters, D. A. Lawrence, M. Roy, F. C. Kischkel, P. Dowd, Ϫ Ϫ vestigated whether circulating RANKL in OPG / mice is active A. Huang, C. J. Donahue, S. W. Sherwood, D. T. Baldwin, et al. 1998. Genomic for inducing osteoclastogenesis. Serum concentrations of RANKL amplification of a decoy receptor for in lung and colon cancer. Nature Ϫ/Ϫ 396: 699–703. in OPG mice treated with 20-epi-1a, a highly potent analog of 19. Yang, C. R., J. H. Wang, S. L. Hsieh, S. M. Wang, T. L. Hsu, and W. W. Lin. ␣ ϳ 1 ,25(OH)2D3, was increased up to 30 ng/ml (43). Serum ob- 2004. Decoy receptor 3 (DcR3) induces osteoclast formation from monocyte/ tained from 20-epi-1a-treated OPGϪ/Ϫ mice was able to induce macrophage lineage precursor cells. Cell Death Differ. 11: S97–S107. 20. Pan, G., J. H. Bauer, V. Haridas, S. Wang, D. Liu, G. Yu, C. Vincenz, tartrate-resistant acid phosphatase-positive cells in mouse bone B. B. Aggarwal, J. Ni, and V. M. Dixit. 1998. Identification and functional char- marrow macrophage cultures in the presence of M-CSF (data not acterization of DR6, a novel -containing TNF receptor. FEBS Lett. shown). These results suggest that soluble RANKL in serum is 431: 351–356. 21. Bucay, N., I. Sarosi, C. R. Dunstan, S. Morony, J. Tarpley, C. Capparelli, involved in the high bone turnover. The present study has estab- S. Scully, H. L. Tan, W. Xu, D. L. Lacey, et al. 1998. osteoprotegerin-deficient lished the new concept that OPG is a potent suppressor of the mice develop early onset osteoporosis and arterial calcification. Genes Dev. 12: release of RANKL. OPG appears to determine the precise sites of 1260–1268. 22. Mizuno, A., N. Amizuka, K. Irie, A. Murakami, N. Fujise, T. Kanno, Y. Sato, bone resorption by inhibiting the shedding of RANKL. N. Nakagawa, H. Yasuda, S. Mochizuki, et al. 1998. Severe osteoporosis in mice lacking osteoclastogenesis inhibitory factor/osteoprotegerin. Biochem. Biophys. Res. Commun. 247: 610–615. Disclosures 23. Nakamura, M., N. Udagawa, S. Matsuura, M. Mogi, H. Nakamura, H. Horiuchi, The authors have no financial conflict of interest. N. Saito, B.Y. Hiraoka, Y. Kobayashi, K. Takaoka, et al. 2003. Osteoprotegerin regulates bone formation through a coupling mechanism with bone resorption. Endocrinology 144: 5441–5449. References 24. Whyte, M. P., S. E. Obrecht, P. M. Finnegan, J. L. Jones, M. N. Podgornik, 1. Boyle, W. J., W. S. Simonet, and D. L. Lacey. 2003. Osteoclast differentiation W. H. McAlister, and S. Mumm. 2002. Osteoprotegerin deficiency and juvenile and activation. Nature 423: 337–342. Paget’s disease. N. Engl. J. Med. 347: 175–184. 2. Takahashi, N., N. Udagawa, M. Takami, and T. Suda. 2002. Cells of bone: os- 25. Hofbauer, L. C., M. Schoppet, P. Schuller, V. Viereck, and M. Christ. 2004. teoclast generation. In Principles of Bone Biology, 2nd Ed, Vol. 1. Effects of oral contraceptives on circulating osteoprotegerin and soluble RANK J. P. Bilezikian and L. G. Raisz, eds. Academic Press, New York, pp. 109–126. ligand serum levels in healthy young women. Clin. Endocrinol. 60: 214–219. 200 THE REGULATION OF SERUM RANKL LEVELS BY OPG

26. Kong, Y. Y., H. Yoshida, I. Sarosi, H. L. Tan, E. Timms, C. Capparelli, 35. Schneeweis, L. A., D. Willard, and M. E. Milla. 2005. Functional dissection of S. Morony, A. J. Oliveira-dos-Santos, G. Van, A. Itie, et al. 1999. OPGL is a key osteoprotegerin and its interaction with receptor activator of NF-␬B ligand. regulator of osteoclastogenesis, lymphocyte development and lymph-node orga- J. Biol. Chem. 280: 41155–41164. nogenesis. Nature 397: 315–323. 36. Will, H., S. J. Atkinson, G. S. Butler, B. Smith, and G. Murphy. 1996. The 27. Nehls, M., D. Pfeifer, M. Schorpp, H. Hedrich, and T. Boehm. 1994. New mem- soluble catalytic domain of membrane type 1 matrix metalloproteinase cleaves ber of the winged-helix protein family disrupted in mouse and rat nude mutations. the propeptide of progelatinase A and initiates autoproteolytic activation: regu- Nature 372: 103–107. lation by TIMP-2 and TIMP-3. J. Biol. Chem. 271: 17119–17123. 28. Sato, N., N. Takahashi, K. Suda, M. Nakamura, M. Yamaki, T. Ninomiya, 37. Butler, G. S., H. Will, S. J. Atkinson, and G. Murphy. 1997. Membrane-type-2 Y. Kobayashi, H. Takada, K. Shibata, M. Yamamoto, et al. 2004. MyD88 but not matrix metalloproteinase can initiate the processing of progelatinase A and is TRIF is essential for osteoclastogenesis induced by lipopolysaccharide, diacyl regulated by the tissue inhibitors of metalloproteinases. Eur. J. Biochem. 244: lipopeptide, and IL-1␣. J. Exp. Med. 200: 601–611. 653–657. 29. Hooper, N. M., E. H. Karran, and A. J. Turner. 1997. secre- 38. Shimada, T., H. Nakamura, E. Ohuchi, Y. Fujii, Y. Murakami, H. Sato, M. Seiki, tases. Biochem. J. 321(Pt. 2): 265–279. and Y. Okada. 1999. Characterization of a truncated recombinant form of human 30. Jacobsen, E. J., M. A. Mitchell, S. K. Hendges, K. L. Belonga, L. L. Skaletzky, membrane type 3 matrix metalloproteinase. Eur. J. Biochem. 262: 907–914. L. S. Stelzer, T. J. Lindberg, E. L. Fritzen, H. J. Schostarez, T. J. O’Sullivan, 39. Amour, A., P. M. Slocombe, A. Webster, M. Butler, C. G. Knight, B. J. Smith, et al. 1999. Synthesis of a series of stromelysin-selective thiadiazole urea matrix P. E. Stephens, C. Shelley, M. Hutton, V. Knauper, et al. 1998. TNF-␣ converting metalloproteinase inhibitors. J. Med. Chem. 42: 1525–1536. enzyme (TACE) is inhibited by TIMP-3. FEBS Lett. 435: 39–44. 31. Koivunen, E., W. Arap, H. Valtanen, A. Rainisalo, O. P. Medina, P. Heikkila, 40. Harada, S., and G. A. Rodan. 2003. Control of osteoblast function and regulation C. Kantor, C. G. Gahmberg, T. Salo, Y. T. Konttinen, et al. 1999. Tumor tar- of bone mass. Nature 423: 349–355. geting with a selective gelatinase inhibitor. Nat. Biotechnol. 17: 768–774. 41. Quinn, J. M., N. J. Horwood, J. Elliott, M. T. Gillespie, and T. J. Martin. 2000. 32. Kartsogiannis, V., H. Zhou, N. J. Horwood, R. J. Thomas, D. K. Hards, Fibroblastic stromal cells express receptor activator of NF-␬B ligand and support J. M. Quinn, P. Niforas, K. W. Ng, T. J. Martin, and M. T. Gillespie. 1999. osteoclast differentiation. J. Bone Miner. Res. 15: 1459–1466. Localization of RANKL (receptor activator of NF ␬B ligand) mRNA and protein 42. Mizuno, A., T. Kanno, M. Hoshi, O. Shibata, K. Yano, N. Fujise, M. Kinosaki, in skeletal and extraskeletal tissues. Bone 25: 525–534. K. Yamaguchi, E. Tsuda, A. Murakami, et al. 2002. Transgenic mice overex- 33. Standal, T., C. Seidel, O. Hjertner, T. Plesner, R. D. Sanderson, A. Waage, pressing soluble osteoclast differentiation factor (sODF) exhibit severe osteopo- M. Borset, and A. Sundan. 2002. Osteoprotegerin is bound, internalized, and rosis. J. Bone Miner. Metab. 20: 337–344. degraded by multiple myeloma cells. Blood 100: 3002–3007. 43. Sato, M., Y. Nakamichi, N. Sato, M. Nakamura, T. Ninomiya, H. Nakamura, Downloaded from 34. Tat, S. K., M. Padrines, S. Theoleyre, S. Couillaud-Battaglia, D. Heymann, M. Shimizu, H. Ozawa, N. Takahashi, and N. Udagawa. 2005. Highly potent ␣ F. Redini, and Y. Fortun. 2006. OPG/membranous-RANKL complex is internal- analogs of 1 ,25-dihydroxyvitamin D3 induce osteoclastogenesis and hypercal- ized via the clathrin pathway before a lysosomal and a proteasomal degradation. cemia in osteopetrotic op/op mice, but not in osteoclast-absent c-fos deficient Bone 39: 706–715. mice. J. Bone Miner. Res. 20: S82 (Abstract). http://www.jimmunol.org/ by guest on October 1, 2021