Secreted Frizzled-Related Protein-1 Inhibits RANKL-DependentOsteoclast Formation

Karl D Hausler,1,2 Nicole J Horwood,1,2 Yoshiro Chuman,3 Jane L Fisher,1 Jennifer Ellis,3 T John Martin,1 Jeffrey S Rubin,3

and Matthew T Gillespie1

ABSTRACT: We determined that sFRP-1 mRNA was differentially expressed by osteoblast/stromal cell linesand that sFRP-1 neutralizing antibodies and siRNA complementary to sFRP-1 coding sequence enhanced,while recombinant sFRP-1 inhibited, osteoclast formation. In studying the mechanism of action for sFRP-1,we found that sFRP-1 could bind recombinant RANKL. These results suggest potential cross-talk betweenWnt and RANKL pathways.

Introduction: Osteoclast formation in normal bone remodeling requires the presence of osteoblast lineage cells thatexpress RANKL and macrophage-colony-stimulating factor (M-CSF), which interact with their cognate receptors onthe osteoclast precursor. We identified secreted Frizzled-related protein-1 (sFRP-1), which is known to bind to Wntand inhibit the Wnt signaling pathway, as an osteoblast-derived factor that impinges on osteoclast formation andactivity.Materials and Methods: Differential display of mRNA from osteoblast lineage cell lines established sFRP-1 to behighly expressed in an osteoclast supporting cell line. sFRP-1 expression in bone was determined by in situhybridization, and the effects of sFRP-1 on osteoclast formation were determined using a neutralizing antibody,siRNA, for sFRP-1 and recombinant protein.Results: In situ hybridization revealed sFRP-1 mRNA expression in osteoblasts and chondrocytes in murine bone.sFRP-1 mRNA expression could be elevated in calvarial primary osteoblasts in response to prostaglandin E2 (PGE2)or interleukin (IL)-11, whereas many other osteotropic agents (e.g., IL-1, IL-6, calcitrol, parathyroid hormone) werewithout any effect. In vitro assays of osteoclast formation established sFRP-1 to be an inhibitor of osteoclastformation. Neutralizing antibodies against sFRP-1 enhanced TRACP� mononuclear and multinuclear osteoclastformation (3- and 2-fold, respectively) in co-cultures of murine osteoblasts with spleen cells, whereas siRNAcomplementary to sFRP-1 coding sequence significantly enhanced osteoclast formation in co-cultures of KUSA O(osteoblast/stromal cell line) and bone marrow cells, cultured in the presence of PGE2 and 1,25(OH)2 vitamin D3.Recombinant sFRP-1 dose-dependently inhibited osteoclast formation in osteoblast/spleen co-cultures, RANKL �M-CSF–treated splenic cultures, and RANKL-treated RAW264.7 cell cultures, indicating a direct action of sFRP-1on hematopoietic cells. Consistent with this, sFRP-1 was found to bind to RANKL in ELISAs.Conclusion: sFRP-1 is expressed by osteoblasts and inhibits osteoclast formation. While sFRP-1 activity mightinvolve the blocking of endogenous Wnt signaling, our results suggest that, alternatively, it could be because of directbinding to RANKL. This study describes a new mechanism whereby osteoblasts regulate osteoclastogenesis throughthe expression and release of sFRP-1.J Bone Miner Res 2004;19:1873–1881. Published online on August 16, 2004; doi: 10.1359/JBMR.040807

Key words: osteoclast, RANKL, secreted Frizzled-related protein, Wnt

INTRODUCTION

OSTEOCLASTS ARE MULTINUCLEATED cells responsible forresorbing bone and are derived from mononuclear pre-

cursors of the monocyte-macrophage lineage. The differen-tiation of monocyte-macrophage lineage cells into oste-oclasts during normal bone development is programmed by

cells of the osteoblast lineage. Critically, membrane-boundRANKL and macrophage-colony-stimulating factor (M-CSF), both of which are expressed by osteoblastic cells,promote osteoclast differentiation through interaction withtheir cognate signaling receptors (RANK and c-fms, respec-tively).(1,2) This process is regulated by an array of factorsthat are expressed by osteoblasts, stromal cells, fibroblasts,and lymphocytes. Principally, the secreted decoy receptor ofRANKL, osteoprotegerin (OPG), inhibits osteoclast forma-The authors have no conflict of interest.

1Bone, Joint, and Cancer Unit, St Vincent’s Institute of Medical Research, Fitzroy, Victoria, Australia; 2These authors contributedequally; 3National Cancer Institute, National Institute of Health, Bethesda, Maryland, USA.

JOURNAL OF BONE AND MINERAL RESEARCHVolume 19, Number 11, 2004Published online on August 16, 2004; doi: 10.1359/JBMR.040807© 2004 American Society for Bone and Mineral Research

1873

tion both in vitro and in vivo.(3) The requirement forRANKL, RANK, and OPG in the control of osteoclastformation has been well documented.(4) In particular, thephenotypes of genetically altered mice that are either defi-cient in or overexpress these factors attest to their funda-mental importance. RANK�/� and RANKL�/� mice dis-play severe osteopetrosis because of a complete lack ofosteoclasts, whereas OPG�/� mice are osteoporotic becauseof an excess of osteoclasts.(5,6) In addition, an absence oflymph node development and reduced numbers of matureT-cells and splenic B-cells were noted in RANKL�/� mice,implicating RANKL as a regulator of lymph node organo-genesis and of lymphopoiesis.(5) These latter actions high-light the importance of the RANKL pathway in extra-skeletal tissues and its fundamental requirement inembryogenesis and organogenesis.

Wnt signaling has a major impact on many aspects ofembryonic development.(7,8) Wnt proteins regulate cell pro-liferation, morphology, motility, and fate,(7,8) and operatethrough seven-pass transmembrane Frizzled receptors thattransmit signals in non-canonical pathways.(9) Frizzled re-ceptors transduce signals in the �-catenin/canonical Wntsignaling pathway; only this process requires a ternarycomplex that includes a member of the low-density lipopro-tein receptor–related family, either LRP5 or LRP6, whichfunctions as a Wnt co-receptor. Recently, LRP5 has beenshown to have a critical role in maintaining BMD, becausemice deficient in LRP5 have low bone mass, low bodyweight, and abnormal eye vascularization.(10) In contrast,mice and humans expressing an LRP5 mutation (G171V)exhibited increased trabecular and cortical BMD, as well asenhanced skeletal strength.(11,12) Similar to the RANKLpathway that can be perturbed by OPG, a decoy receptor forRANK, the interaction of Wnts with Frizzleds can beblocked by secreted Frizzled-related proteins (sFRPs). ThesFRPs (1–5) each contain a module that is homologous tothe Wnt-binding, cysteine-rich domain (CRD) of the Friz-zleds.(13,14) Through the CRD, and perhaps other portions ofthe protein, sFRPs have been shown to act as decoy recep-tors for Wnts, inhibiting Wnt–Frizzled interactions.(4,15–18)

Notably, several sFRPs are expressed by cells withinbone.(19–21)

We previously reported that several bone marrow–derived stromal cell lines from a transgenic mouse immor-talized with a temperature-sensitive variant of the SV40large T antigen differ in their osteoclast-inductive ability.The use of these lines permitted the identification of inter-leukin (IL)-18 as an inhibitor of osteoclast formation.(22,23)

We have extended our analyses of these cell lines to identifyother osteoblastic genes that are involved in the regulationof osteoclastogenesis. Using differential display PCR(ddPCR)(22,23) to compare the mRNA populations betweenosteoclast-inductive and -noninductive cell lines, we iden-tified sFRP-1 as a product of osteoblastic stromal cells.Using neutralizing antibodies to sFRP-1, as well as recom-binant sFRP-1, we established that sFRP-1 blocks osteoclastformation. Furthermore, sFRP-1 binds directly to RANKL,and this interaction may, in part, be responsible for itsinhibition of osteoclastogenesis.

MATERIALS AND METHODS

Animals, cell lines, and chemicals

Newborn (0–1 day old) C57BL/6J mice and 6- to9-week-old male C57BL/6J mice were purchased from Mo-nash University Animal Services Centre (Clayton, Austra-lia). The murine stromal cell lines, tsJ2, tsJ10, and tsJ14,were generated by transfection with a retroviral vector ex-pressing a temperature-sensitive variant of the immortaliz-ing gene of SV40 (tsA58).(23,24) The murine stromal celllines MC3T3-E1,(25) KUSA O,(26) and ST2(25) were gener-ated as described. The murine macrophage-like cell lineRAW 264.7 that differentiates into osteoclasts(27) was ob-tained from American Type Culture Collection (Manassas,VA, USA). 1�,25(OH)2 vitamin D3 [1�,25(OH)2D3] waspurchased from Wako Pure Chemical (Osaka, Japan). Pros-taglandin E2 (PGE2) was obtained from Sigma Chemical (StLouis, MO, USA). M-CSF was a gift from Genetics Insti-tute (Andover, MA, USA).

Expression and purification of soluble RANKL

The core biologically active ectodomain of murine RANKLwas expressed in Escherichia coli as a glutathioneS-transferase fusion protein (GST-fusion protein). cDNA en-coding a soluble form of murine RANKL (amino acids 158–316) was cloned into pGEX-4T-1 (Amersham Pharmacia Bio-tech, Piscataway, NJ, USA) downstream of GST using SalIand NotI restriction endonucleases: the RANKL expressionconstruct was a kind gift of Dr FP Ross (Washington Uni-versity, St Louis, MO, USA). After isopropylthio-�-D-galactosidase–mediated (IPTG-mediated) induction of proteinexpression, GST-RANKL protein was extracted from cell ly-sates essentially as previously described.(28) Eluted GST-RANKL protein was dialyzed against PBS and sterilized, andprotein concentration was ascertained using a BCA proteinassay (Pierce). For some ELISA experiments, soluble RANKLwas purchased from Peprotech (Rocky Hill, NJ, USA).

Recombinant sFRP-1 and antiserum

Recombinant sFRP-1 was prepared as previously de-scribed.(18) Rabbit polyclonal antiserum was raised againstfull-length, purified recombinant sFRP-1 and fractionatedon protein-G resin.(18)

ddPCR

Total RNA (1 �g) was extracted from cell lines, reversetranscribed, and used for PCR essentially as described.(29,30)

ddPCR was performed essentially as described.(22) The oli-gonucleotides used for ddPCR were ddPCR T12VA (5�-TTTTTTTTTTTT[A,C,G]A-3�) and DDMR-2 (5�CTTGA-TTGCC-3�). Oligonucleotides for amplification of sFRP-1cDNA were sFRP-1–15 (sense strand oligonucleotide5�GTTGTGGCTTTTGCATTGCAC3�, nucleotides 1303–1324, GenBank accession no. NM_013834) and sFRP-1–6(antisense strand oligonucleotide 5�AGCCTTGGGCAGT-CAACGACG3�, nucleotides 2249–2229), and sFRP-1–9(antisense strand oligonucleotide 5�TGTTGAAAACTAG-TAGCTG3�, nucleotides 2029–2011) served as an internaloligonucleotide probe for hybridization studies as de-

1874 HAUSLER ET AL.

scribed.(29,30) Murine GAPDH primers were GAPDH-1,GAPDH-2, and GAPDH-4 as described previously.(22)

Osteoclast formation assays

Osteoblast/spleen cell co-culture assay: Osteoblasticcells were prepared from the calvaria of newborn mice bysequential digestion with 0.1% collagenase (WorthingtonBiochemical) and 0.2% dispase (Godo Shusei, Tokyo, Ja-pan). Spleen cells were obtained from adult mice by disag-gregation through a wire sieve.(31) Co-cultures of murinecalvarial osteoblasts and spleen cells were performed aspreviously described.(22,31,32) After 7 days of culture, adher-ent cells were fixed and stained for TRACP, a marker ofmature osteoclasts. The number of TRACP� multinucleatedcells was scored as described.(33) To test the effects ofendogenous sFRP-1 in osteoclast co-culture assays, the ex-pression of sFRP-1 by an osteoblast lineage cell line (KUSAO) was knocked down using small interfering RNA(siRNA). The synthesis and purification of siRNA comple-mentary to sFRP-1 coding sequence was performed usingSilencer siRNA construction kit (Ambion, Austin, TX,USA) following the manufacturer’s instructions. The result-ing double-stranded siRNA (10 nM) was transfected intoKUSA O cells using Fugene6 transfection agent (RocheDiagnostics, Indianapolis, IN, USA). The sequences of the21-mer siRNA oligonucleotides previously established to bespecific for sFRP-1(34) were 5�-AACTTCTTGGGGACA-ATCTTC-3� and 5�-AAGAAGATTGTCCCCAAGAAG-3�. KUSA O cells were transfected with siRNA for 24 h andco-cultured with bone marrow cells, essentially as previ-ously described for osteoblast and bone marrow co-cultures.(22) In parallel with co-culture assays, total cellularRNA was isolated from KUSA O cells and KUSA O cellstransfected with siRNA, in the presence of either1�,25(OH)2D3 or 1�,25(OH)2D3 and PGE2, after 7 days ofculture. Total RNA (5 �g) was reversed transcribed, andPCR was performed within the linear range of amplificationusing either sFRP-1–6 and sFRP-1–15 oligonucleotides orGAPDH-2 and GAPDH-4 oligonucleotides.

Spleen, bone marrow, and RAW264.7 cell cultures: Os-teoclast formation assays involving either spleen or bonemarrow cells were performed by culturing cells in thepresence of M-CSF (25 ng/ml) and RANKL (100 ng/ml).RAW264.7 cells were cultured in the presence of RANKL(100 ng/ml) alone. All treatment groups were tested inquadruplicate, and fresh culture media were provided onday 3. Osteoclast formation was evaluated after 7 days ofculture by staining for TRACP and counting the number ofTRACP� multinucleated cells per well. Spleen cells wereobtained from adult mice as described above and seeded(105 cells/well) in 48-well plates (Corning Glass, Corning,NY, USA) with 0.2 ml of �-MEM (GIBCO BRL, Gaith-ersburg, MD, USA) containing 10% FBS (Cytosystems,Castle Hill, New South Wales, Australia). Adult bone mar-row cells were obtained by transversely cutting the epiph-yses of the femur and tibia and flushing out the medullarycavity with 5 ml of PBS using a 25-gauge needle. Thissuspension of bone marrow cells was gently agitated andallowed to briefly settle. The supernatant was collected in afresh 30-ml tube and centrifuged at 2000 rpm for 5 minutes.

The pelleted cells were resuspended in media and culturedin 48-well plates (105 cells/well) in the presence of M-CSF(25 ng/ml) and RANKL (100 ng/ml).

sFRP-1/RANKL ELISA binding assays

sFRP-1 diluted in 0.02% NaN3/PBS was incubated in96-well Falcon ELISA plates (300 ng/50 �l/well) for 2 h at37°C. After decanting, all wells were filled with 4% bovineserum albumin (BSA)/0.02% NaN3/PBS and incubated foran additional 2 h at 37°C. After five washes with TAPS(0.05% Tween-20 in 0.02% NaN3/PBS), wells were incu-bated with 50-�l aliquots of soluble RANKL (0–10 �g/ml)in 1% BSA/0.01% NaN3/PBS overnight at room tempera-ture. After five washes with TAPS, 50 �l/well of RANKLrabbit polyclonal antibody (Peprotech) diluted in 1% BSA/TAPS to a final concentration of 1 �g/ml was incubated inwells for 2 h at 37°C. Another five washes in TAPS werefollowed by a 2-h treatment at 37°C with 1:400 dilution ofconjugated alkaline phosphatase-goat anti-rabbit IgG(Sigma). After a final set of five washes with TAPS, 2mg/ml p-nitrophenolphosphate (Sigma) in carbonate buffer(0.1 M Na2CO3, 1 mM MgCl2, pH 9.8) was added. Absor-bance at 405 nm was measured using an ELISA plate reader(Bio-Rad).

In situ hybridization

A 946-bp fragment of murine sFRP-1 corresponding to thepoorly conserved 3� untranslated region was generated byRT-PCR using the oligonucleotides sFRP-1–6 (sense, 5�-AGCCTTGGGCAGTCAACGACG-3�, nucleotides 2229–2249, GenBank U88566) and sFRP-1–15 (antisense, 5�-GTTGTGGCTTTTGCATTGCAC-3�, nucleotides 1303–1324, GenBank U88566) and RNA derived from mousecalvarial osteoblasts as template. This cDNA was purifiedusing the PCR Wizard column (Promega, Madison, WI, USA)and cloned into pCR-Script Amp SK(�) vector (Stratagene,La Jolla, CA, USA). The resultant plasmid was linearized withBsaI restriction enzyme and transcribed with T7 polymerase togenerate an antisense riboprobe for in situ hybridization.(35)

Paraffin sections, mounted on slides pretreated with3-aminoprophyltriethoxysilane, were dewaxed with xyleneand rehydrated sequentially in decreasing concentrations ofethanol, from 100% to 30%, before rinsing in DEPC-treatedwater. Deproteinization was carried out with 0.2 M HCl for 20minutes at room temperature, followed by digestion with pro-teinase K at 2 �g/ml in 0.1 M Tris buffer, pH 8, and 50 mMEDTA, pH 8, for 30 minutes at 37°C, followed by 2 mg/ml ofglycine in PBS for 5 minutes. The tissues were fixed in 4%paraformaldehyde in PBS for 15 minutes. Slides were rinsed inPBS between each pretreatment, and all procedures were car-ried out at room temperature. Pre-hybridization was performedat 37° C for 1 h in hybridization buffer, which consisted of50% deionized formamide, 5� SSC (SSC contains 0.15 MNaCl and 0.15 M sodium citrate, pH 7.0), 2% block reagent,0.1% N-lauroyl-sarcosine, and 2% sodium dodecylsulphate.The DIG-labeled probe was diluted to a final concentration of0.25 ng/�l, and 30 �l was applied to each slide. Slides wereprotected with coverslips and incubated for 16–18 h at 42°C ina humidified chamber. Coverslips were removed after thistime, and the slides were incubated in 2� SSC for 15 minutes.

1875sFRP-1 INHIBITS OSTEOCLAST FORMATION

Excess probe was removed by incubation with 25 �g/mlRNase A in 2� SSC for 30 minutes at 37° C. Slides wereincubated in 2� SSC for 15 minutes, 1� SSC for 30 minutes,and 0.1� SSC for 30 minutes at 37°C. Hybridized probe wasdetected with alkaline phosphatase–coupled anti-DIG anti-body according to the manufacturer’s instructions. Sectionswere counterstained with nuclear fast red. To ensure validity ofthe signal, a negative control was used wherein the section wastreated with 150 �g/ml RNase A in 2� SSC (2 h; 37°C) beforepre-hybridization and subsequent hybridization with the anti-sense riboprobe as described above. Probe sensitivity wasconfirmed with additional experiments performed with a hy-bridization temperature of 65°C, although 42°C was generallyused to prevent the bone sections from lifting off the slide.

RESULTS

Detection of sFRP-1 expression by differential displayPCR

To identify genes expressed by osteoblast/stromal cellsthat may influence osteoclast formation, we made use of aseries of highly related and stable cell lines derived fromH-2KbtsA58 transgenic mice that were capable (tsJ2 andtsJ10) or incapable (tsJ14) of supporting osteoclast forma-tion in vitro.(22,36) The cell lines tsJ2 and tsJ10 stimulatedosteoclast formation after hydrocortisone (10�6 M) or PGE2

(10�7 M) and 1�,25(OH)2D3 (10�8 M) treatment. However,in the absence of hydrocortisone or PGE2 and1�,25(OH)2D3, the tsJ2 and tsJ10 cell lines were unable topromote osteoclast differentiation. In contrast, the tsJ14 cellline was unable to support osteoclast formation even in thepresence of hydrocortisone or PGE2 and 1�,25(OH)2D3.Because they are genetically very closely related, these celllines are ideally suited for the identification of factors thatmay influence osteoclast formation by differential display ofexpressed transcripts. We previously used this approach toidentify IL-18 as an osteoclast inhibitor.(22) In this study, weidentified an mRNA species that was highly expressed byosteoblast lineage cell lines under conditions that supportedosteoclast formation [Fig. 1A, lanes 1 and 3; tsJ10 cellstreated with hydrocortisone, and tsJ2 cells treated with 1�,25(OH)2D3 and PGE2, respectively] but poorly expressedby cells that failed to stimulate osteoclast formation [Fig.1A, lanes 2 and 4; tsJ14 treated either with hydrocortisoneor PGE2 and 1�,25(OH)2D3]. The 234-bp differentiallyexpressed PCR fragment was extracted, cloned intopCRScript, sequenced, and identified as a fragment of mu-rine sFRP-1, corresponding to nucleotides 2444–2659(GenBank database U88566).

The differential expression of sFRP-1 mRNA in osteo-blast lineage cell lines was confirmed by semiquantitativeRT-PCR using sFRP-1–specific oligonucleotide primers(Fig. 1B). sFRP-1 mRNA levels were found to be higher inthe osteoclast-inductive cells than the noninductive cells:10-fold higher in the hydrocortisone-treated tsJ10 cells (lane1) and 54-fold higher in 1�,25(OH)2D3- and PGE2-treatedtsJ2 cells (lane 3) than the osteoclast noninductive cells[tsJ14 treated with hydrocortisone (lane 2) or with PGE2

and 1�,25(OH)2D3 (lane 4)]. In addition, sFRP-1 andRANKL mRNA levels were determined by semiquantita-

tive RT-PCR using murine calvarial osteoblasts and severalosteoblast lineage cell lines with varying capacities to sup-port osteoclast formation (Fig. 2). Consistent with resultsfrom differential display PCR, sFRP-1 mRNA levels wereelevated in osteoblast lineage cell lines having the capacityto support osteoclast formation (tsJ2 and KUSAO) com-pared with lines incapable of supporting osteoclast forma-tion (tsJ14 and MCT3E1). Note that the expression level ofsFRP-1 was relatively low in primary osteoblasts (see be-low). The abundance of RANKL mRNA in this panel of cell

FIG. 1. Identification of sFRP-1. (A) An example of a ddPCR gel.Lanes correspond to RNA from the different sources: (1)hydrocortisone-treated tsJ10 cells, (2) hydrocortisone-treated tsJ14cells, (3) 1�,25(OH)2D3- and PGE2-treated tsJ2 cells, and (4)1�,25(OH)2 D3- and PGE2-treated tsJ14 cells. The PCR fragmentidentified as sFRP-1 is indicated by the arrow on the left. Theosteoclast-supporting activity (OSA) of these cell lines is indicatedbelow the gel: � (supportive) or � (nonsupportive). (B) Semiquanti-tative RT-PCR analysis of sFRP-1 mRNA. PCR products for RNAisolated from different sources were reversed transcribed with oli-go(dT) and PCR performed with the primers sFRP-1–15 and sFRP-1–6for 24 cycles, which was in the log-linear range of amplification. Lanescorrespond to RNA extracted from the cells and treatments described inA. Resultant PCR products were electrophoresed, transferred to a nylonmembrane, and hybridized with a [�-32P]–labeled internal detectionoligonucleotide for sFRP-1 (sFRP-1–9). Similar amplifications forGAPDH with GAPDH-2 and GAPDH-4 for 20 cycles were performed,and products were detected with [�-32P]–labeled GAPDH-1 as previ-ously described.(22)

1876 HAUSLER ET AL.

lines was included for comparison with sFRP-1, becauseRANKL is essential for normal osteoclast formation.(4,37)

Regulation and distribution of sFRP-1 expression

Osteoclast formation is regulated by a number of cyto-kines and growth factors that principally act through theosteoblast to influence RANKL and OPG production.(4,36,37)

We examined the expression of sFRP-1 by primary murinecalvarial osteoblasts in response to several agents. Notably,10�7 M PGE2 and 10�8 M IL-11 increased mRNA levels ofsFRP-1 beginning at 2 and 8 h, respectively (Figs. 2B and2C). In contrast, dexamethasone, 1�,25(OH)2D3, PTH,IL-1, IL-4, IL-10, and IL-18 were without effect (data notshown).

We determined the tissue distribution of sFRP-1, spe-cifically during bone development by in situ hybridiza-tion. In the newborn (day 1) murine long bone, terminal

chondrocytic differentiation has occurred, while endo-chondral ossification proceeds with cartilage mineraliza-tion. This is accompanied by the formation of a distinctgrowth plate and primary ossification center. At this stageof long bone development, mRNA for sFRP-1 was ex-pressed in osteoblasts, chondrocytes of the growth plate,and to a lesser extent, among hypertrophic chondrocytes(Figs. 3A and 3C); no signal was observed with a senseprobe or after RNase treatment (Fig. 3B and data notshown).

FIG. 2. Semiquantitative RT-PCR analysis of sFRP-1, RANKL, andGAPDH mRNA. Total RNA was isolated from primary osteoblasts(POb), tsJ2, KUSA O, ST2, tsJ14, and MCT3E1 cell lines, reversetranscribed with oligo(dT), and subjected to PCR within the log-linearrange of amplification for either sFRP-1 (24 cycles), RANKL (28cycles), or GAPDH (20 cycles). Resultant PCR products were agarosegel electrophoresed, transferred to nylon membrane, and hybridizedwith �-32P–labeled internal detection oligonucleotide specific to eachof the cDNA products. RT-PCR analysis was repeated in triplicate. Theosteoclast-supporting activity (OSA) of these cell lines is indicatedbelow the gel: � (supportive) or � (nonsupportive). (B and C) Semi-quantitative RT-PCR analysis of sFRP-1 mRNA. Primary osteoblastswere treated with (B) 10�7 M PGE2 and (C) 10�8 M IL-11. A timecourse was performed with total RNA extracted from the treatedcultures at 0, 0.5, 1, 2, 4, 8, 12, and 24 h. Total RNA were reversetranscribed with oligo(dT) and subjected to PCR within the log-linearrange of amplification for either sFRP-1 (25 cycles) or GAPDH (20cycles). Resultant PCR products were agarose gel electrophoresed,transferred to nylon membrane, and hybridized with �-32P–labeledinternal detection oligonucleotide specific to each of the cDNA prod-ucts. RT-PCR analysis was repeated in triplicate.

FIG. 3. Expression of sFRP-1 mRNA in developing endochondralskeletal tissues of the mouse. (A) In the neonate tibia, sFRP-1 mRNAtranscripts (dark blue/purple color) were detectable in some of theproliferating and pre-hypertrophic chondrocytes within the transitionalzone, and modest expression was noted in the hypertrophic chondro-cytes. (B) No specific hybridization signal was observed when a sensesFRP-1 riboprobe or a RNase-treated negative control was used (coun-terstain: nuclear fast red). (C) sFRP-1 mRNA transcripts were detectedin osteoblasts (arrow) within the cancellous bone of the proximalepiphysis.

1877sFRP-1 INHIBITS OSTEOCLAST FORMATION

Regulation of osteoclast formation by sFRP-1antiserum and recombinant protein

To test the possibility that sFRP-1 affects osteoclast for-mation, we examined the impact of sFRP-1 antiserum on theformation of TRACP� mononucleated and multinucleatedcells (MNCs) in co-cultures of murine calvarial osteoblastsand splenic cells. In the absence of any supplements, fewTRACP� cells were detected. The combination of PGE2

and 1�,25 (OH)2D3 stimulated the formation of a smallnumber of TRACP� MNCs and a moderate number ofsimilarly stained mononucleated cells (Fig. 4A). Simulta-neous addition of sFRP-1 antiserum (final concentration,�1 �g/ml) with these supplements resulted in at least adoubling of the number of TRACP� MNCs and mononucle-ated cells (Fig. 4A), suggesting that the antiserum wasblocking the action of an endogenous inhibitor; preimmuneor unrelated antisera did not effect osteoclast formation. Insupport of this view, recombinant sFRP-1 elicited a dose-dependent inhibition of osteoclast formation when murinespleen cells were cultured in the presence of RANKL (100ng/ml) and M-CSF (25 ng/ml; Fig. 4B). Similar results wereobtained when the assay was performed only with adherentspleen cells (devoid of nonadherent cells such as lympho-cytes; data not shown), and with the RAW264.7 cell line, amouse macrophage-like line was used, which differentiatesinto osteoclasts in the presence of RANKL (100 ng/ml) andabsence of M-CSF (Fig. 4C). In all these assay systems,sFRP-1 inhibited the formation of TRACP� MNCs with ahalf-maximal effect at �28 nM (�1 �g/ml) and completeinhibition at 280 nM (Figs. 4B and 4C).

Regulation of osteoclast formation by siRNA

To determine the contribution of endogenous sFRP-1expression to influence osteoclast formation, we usedsiRNA complementary to sFRP-1 coding sequence usingvalidated siRNA to sFRP-1.(34) When KUSA O cells weretransfected with siRNA for sFRP-1, endogenous sFRP-1mRNA expression was downregulated 2-fold after 7 days ofculture (Fig. 5, and data not shown). When KUSA O cellstransfected with siRNA for sFRP-1 were cultured with bonemarrow cells in the presence of either 1�,25(OH)2D3 (10�8

M) or 1�,25(OH)2D3 and PGE2 (10�7 M), sFRP-1 mRNAlevels were 50% of control cells after 7 days of culture (Fig.5). In parallel cultures to those used to determine mRNAlevels, TRACP� MNCs were counted. Inhibition of KUSAO sFRP-1 mRNA expression resulted in increased oste-oclast formation relative to control co-cultures, although thedifference was only significant in cultures treated with1�,25(OH)2D3 and PGE2.

RANKL binding to sFRP-1 in ELISA experiments

sFRPs bind Wnts and typically block their activity.Therefore, despite the lack of evidence that Wnt signalingstimulates osteoclastogenesis, one might surmise that dis-ruption of Wnt activity is the mechanism of sFRP-1 inhib-itory activity in osteoclast formation assays. However, an-other possibility came to our attention after the screening ofa phage-displayed peptide library to identify sFRP-1 bind-ing motifs. We detected and validated a peptide motif that

conferred binding to sFRP-1 and resembled a sequencepresent in RANKL (unpublished observations). ELISA ex-periments indicated that RANKL and sFRP-1 bound eachother in a specific manner (Fig. 6). This interaction mayexplain or contribute to the ability of sFRP-1 to inhibitosteoclast formation.

DISCUSSION

We report the identification of a new inhibitor of oste-oclast formation, sFRP-1, which is a member of the secreted

FIG. 4. Effect of anti-sFRP-1 antibody and sFRP-1 on osteoclastformation. (A) Murine splenic and primary osteoblastic cells wereco-cultured with and without 1�,25(OH)2D3 (10�8 M) and PGE2 (10�7

M) in the presence of an anti-sFRP-1 antibody. For negative andpositive controls, co-cultures were performed in the absence and pres-ence of 1�,25(OH)2D3 and PGE2, respectively. After culture for 7 days,TRACP� cells (mononucleated [Mono] and multinucleated [MNC])were counted. Data are expressed as the means � SE of quadruplicatecultures and are representative of three similar experiments. (B) Theeffects of sFRP-1 (dose response) on osteoclast (TRACP� MNC)formation in cultured adult murine splenic cells. Unfractionated C57/BL6 spleen cells were treated with RANKL (50 ng/ml) and M-CSF (25ng/ml) for 7 days. (C) Dose response to sFRP-1 on osteoclast formationfrom RAW264.7 cells stimulated with RANKL (100 ng/ml) for 7 days.Medium and mediators were changed at days 3 and 7. Cultures wereperformed in the presence and absence of RANKL and M-CSF forpositive and negative controls, respectively. Data are expressed as themeans � SE of quadruplicate cultures from three separate experiments.Statistical analysis was performed using a two-tailed t-test comparingdata to the positive control wells. *p � 0.05 and **p � 0.01.

1878 HAUSLER ET AL.

Frizzled-related protein family that has structural featuressimilar to the extracellular domain of the Frizzled family ofWnt receptors. The previously reported action for sFRP-1was to perturb Wnt signaling by a mechanism reminiscentof decoy receptors.(38) We established that sFRP-1 mRNAwas differentially expressed in cells of the osteoblast lin-eage, that its expression by osteoblasts was elevated underconditions conducive for osteoclast formation, and thatPGE2 or IL-11 was capable of enhancing its expression.

Osteoblasts and cells of the immune system produce arepertoire of cytokines and growth factors that can eitherenhance or inhibit osteoclast formation. Notably, the twoprincipal factors promoting osteoclast formation areM-CSF and RANKL.(4,37) Other locally produced factorscan also influence these processes, including OPG,granulocyte/macrophage-colony-stimulating factor (GM-CSF), TNF, osteoclast inhibitory lectins, and IL-1�, IL-4,IL-10, IL-12, IL-13, and IL-18.(4,27,37,39) Many of thesefactors are expressed constitutively and are refractive toregulation by exogenous stimuli or their regulation isinverse to that of RANKL, such as observed withOPG.(36) The finding that sFRP-1 is expressed by theosteoblast and that PGE2 or IL-11 elevated sFRP-1 ex-

pression along with that of RANKL initially suggestedthat sFRP-1 might have a positive impact on osteoclastformation. However, a neutralizing antibody againstsFRP-1 enhanced, whereas recombinant sFRP-1 inhib-ited, in vitro osteoclast formation, indicating that sFRP-1was an inhibitor of osteoclast formation, a new functionfor sFRP-1. Presumably its elevated expression in osteo-blasts that support osteoclast formation represents a com-pensatory mechanism to limit osteoclastogenesis. Thisnotion is supported by our finding that osteoclast forma-tion is increased when endogenous sFRP-1 mRNA inosteoclast inductive KUSA O cells (osteoblast/stromalcell line) is inhibited by siRNA. This finding is in agree-ment with a recent report comparing ex vivo osteoclastformation from bone marrow cells from 25- to 26-week-old female sFRP-1�/� mice compared with wildtypemice, with sFRP-1�/� cultures having increased oste-oclast formation.(40)

The ability of sFRP-1 to bind RANKL provides a possi-ble novel explanation for a mechanism of action. sFRP-1binding to RANKL could prevent RANK–RANKL interac-tions, in a manner akin to OPG, thereby providing anelegant mechanism for interaction of the Wnt and RANKLpathways. Alternatively, sFRP-1 binding to RANKL mayallow Wnt signaling to proceed, which may influence oste-oclast formation. However, the possibility remains that theability of sFRP-1 to inhibit osteoclastogenesis is caused byits disruption of endogenous Wnt signaling. This wouldimply a previously unrecognized role for Wnts in osteoclastformation that is potentially regulated by RANKL. Therelatively high concentrations of sFRP-1 required to inhibitosteoclast formation (Ki � 30 nM) are consistent withamounts needed to block Wnt-dependent �-catenin stabili-zation.(18)

This report contains direct evidence indicating a role forsFRP-1 in regulating the differentiation of cells involved inbone formation and remodeling. Potential involvement ofsFRPs in bone development might be inferred from the fact

FIG. 6. sFRP-1 binds to RANKL. Soluble RANKL was incubated atthe indicated concentrations in ELISA wells precoated with sFRP-1 orBSA. Bound RANKL was measured using RANKL antiserum andsecondary reagents, and color development was at 405 nm.

FIG. 5. Effect of inhibiting endogenous sFRP-1 by siRNA on oste-oclast formation. The synthesis and purification of siRNA for sFRP-1were performed using Silencer siRNA construction kit (Ambion) fol-lowing the manufacturer’s instruction. The resulting double-strandedsiRNA (10 nM) was transfected into KUSA O cells using Fugene6transfection reagent (Roche Diagnostics). Transfected cells were cul-tured for a further 7 days in the presence of 1�,25(OH)2D3 (10�8 M)or 1�,25(OH)2D3 and PGE2 (10�7 M). Total cellular RNA was ex-tracted from these cells, and nontransfected control cells and the levelsof sFRP-1 mRNA were determined by RT-PCR. Samples were nor-malized against GAPDH. sFRP-1 transcripts were reduced 2-fold in thesiRNA transfected cells, relative to nontransfected control cells. Inparallel with this experiment, siRNA transfected KUSA O cells andnontransfected control cells were co-cultured with bone marrow cellsfor 7 days in the presence of 1�,25(OH)2D3 or 1�,25(OH)2D3 andPGE2. Inhibition of sFRP-1 mRNA in KUSA O cells increased oste-oclast formation, although this was only significant in cultures treatedwith 1�,25(OH)2D3 and PGE2.

1879sFRP-1 INHIBITS OSTEOCLAST FORMATION

that genes encoding sFRPs have only been detected invertebrates. Moreover, various sFRPs, including sFRP-1 asdescribed in this study, are expressed in bone. For example,sFRP-3 (FRZB) transcripts are expressed in the developinglong bones in a graded manner, decreasing from the chon-droblasts of the epiphyses to the primary ossification cen-ter.(19) Strong hybridization signals were detected forsFRP-1 transcripts in the transitional chondrocytes of de-veloping long bones, with weaker signal in regions of pro-liferating and hypertrophic chondrocytes. FrzB2/sFRP-4 isexpressed in chondrocytes undergoing apoptosis.(20) Theredundancy of sFRP family members may complicate ef-forts to define their respective contributions to the regula-tion of bone formation.

A role for Wnt signaling in the development, differenti-ation, and maintenance of tissues related to the skeleton,particularly cartilage and bone, has recently been exempli-fied by the identification of loss of function mutations inLRP5, which result in a loss of BMC, whereas gain offunction mutations in LRP5 result in increased bonemass.(10–12) These LRP5 effects have been attributed todefects in osteoblast function. As LRP5 is a Wnt co-receptor, it is not surprising that Frizzleds and Wnts are alsoexpressed in bony tissues and have been implicated in var-ious aspects of development, such as chondrogenesis,(41–46)

limb initiation,(47) and joint formation.(48) Furthermore, therecent finding that mice deficient for sFRP-1 have increasedaccrual of trabecular bone argues a fundamental role forsFRP-1 to modulate osteoblast function, which may resultfrom enhanced canonical Wnt signaling,(40) in addition tothe roles that we describe. Recent evidence suggests that adownstream target of the non-canonical Wnt/Ca�� path-way, NFATc1 (nuclear factor activated T-cells),(49) is im-portant in osteoclast formation RANKL and has been shownto induce NFATc1 expression through TRAF6 and c-Fossignaling pathways.(50) In addition, retroviral delivery ofNFATc1 into bone marrow macrophages (osteoclast precur-sor cells) induced osteoclast formation in the absence ofRANKL, suggesting that NFATc1 is a target of RANKsignaling that may mediate terminal differentiation of oste-oclasts.(50) It is therefore conceivable that sFRP-1 mayprevent NFATc1 expression, either by directly blockingRANKL/RANK signaling, Wnt signaling, or a combinationof both processes.

The finding that sFRP-1 is expressed by the osteoblast,can bind to RANKL, is an inhibitor of osteoclastogenesis,and under certain circumstances, may be regulated intandem with RANKL provides new evidence for thecontrol of osteoclast formation and suggests a link be-tween the NF-�B and Wnt pathways. We provide evi-dence for cross-talk between these pathways in bone,although both pathways are operative in several differentorgans (for example, lymph node, brain, breast, and co-lon). It remains to be established whether other membersof the sFRP family can inhibit osteoclast formation, thepotential cross-talk between these pathways in normalphysiological bone formation, pathophysiological condi-tions of bone (arthritis, cancer and osteoporosis), and atextraskeletal sites.

ACKNOWLEDGMENTS

This work was supported by National Health and MedicalResearch Council, Australia, Program Grant 003211 to TJMand MTG.

REFERENCES

1. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T,Elliott R, Colombero A, Elliott G, Scully S, Hsu H, Sullivan J,Hawkins N, Davy E, Capparelli C, Eli A, Qian YX, Kaufman S,Sarosi I, Shalhoub V, Senaldi G, Guo J, Delaney J, Boyle WJ 1998Osteoprotegerin ligand is a cytokine that regulates osteoclast dif-ferentiation and activation. Cell 93:165–176.

2. Yasuda H, Shima N, Nakagawa N, Yamaguchi K, Kinosaki M,Mochizuki S, Tomoyasu A, Yano K, Goto M, Murakami A, TsudaE, Morinaga T, Higashio K, Udagawa N, Takahashi N, Suda T1998 Osteoclast differentiation factor is a ligand forosteoprotegerin/osteoclastogenesis-inhibitory factor and is identi-cal to TRANCE/RANKL. Proc Natl Acad Sci USA 95:3597–3602.

3. Simonet WS, Lacey DL, Dunstan CR, Kelley M, Chang MS, LuthyR, Nguyen HQ, Wooden S, Bennett L, Boone T, Shimamoto G,DeRose M, Elliott R, Colombero A, Tan HL, Trail G, Sullivan J,Davy E, Bucay N, Renshaw-Gegg L, Hughes TM, Hill D, PattisonW, Campbell P, Sander S, Van G, Tarpley J, Derby P, Boyle WJ1997 Osteoprotegerin: A novel secreted protein involved in theregulation of bone density. Cell 89:309–319.

4. Suda T, Takahashi N, Udagawa N, Jimi E, Gillespie MT, MartinTJ 1999 Modulation of osteoclast differentiation and function bythe new members of the tumor necrosis factor receptor and ligandfamilies. Endocr Rev 20:345–357.

5. Kong YY, Yoshida H, Sarosi I, Tan HL, Timms E, Capparelli C,Morony S, Oliveira-dos-Santos AJ, Van G, Itie A, Khoo W,Wakeham A, Dunstan CR, Lacey DL, Mak TW, Boyle WJ, Pen-ninger JM 1999 OPGL is a key regulator of osteoclastogenesis,lymphocyte development and lymph-node organogenesis. Nature397:315–323.

6. Anderson DM, Maraskovsky E, Billingsley WL, Dougall WC,Tometsko ME, Roux ER, Teepe MC, DuBose RF, Cosman D,Galibert L 1997 A homologue of the TNF receptor and its ligandenhance T-cell growth and dendritic-cell function. Nature 390:175–179.

7. Smalley MJ, Dale TC 1999 Wnt signalling in mammalian devel-opment and cancer. Cancer Metastasis Rev 18:215–230.

8. Wodarz A, Nusse R 1998 Mechanisms of Wnt signaling in devel-opment. Annu Rev Cell Dev Biol 14:59–88.

9. Dale TC 1998 Signal transduction by the Wnt family of ligands.Biochem J 329:209–223.

10. Kato M, Patel MS, Levasseur R, Lobov I, Chang BH, Glass DA II,Hartmann C, Li L, Hwang TH, Brayton CF, Lang RA, Karsenty G,Chan L 2002 Cbfa1-independent decrease in osteoblast prolifera-tion, osteopenia, and persistent embryonic eye vascularization inmice deficient in Lrp5, a Wnt coreceptor. J Cell Biol 157:303–314.

11. Little RD, Recker RR, Johnson ML 2002 High bone density due toa mutation in LDL-receptor-related protein 5. N Engl J Med347:943–944.

12. Little RD, Carulli JP, Del Mastro RG, Dupuis J, Osborne M, FolzC, Manning SP, Swain PM, Zhao SC, Eustace B, Lappe MM,Spitzer L, Zweier S, Braunschweiger K, Benchekroun Y, Hu X,Adair R, Chee L, FitzGerald MG, Tulig C, Caruso A, Tzellas N,Bawa A, Franklin B, McGuire S, Nogues X, Gong G, Allen KM,Anisowicz A, Morales AJ, Lomedico PT, Recker SM, VanEerdewegh P, Recker RR, Johnson ML 2002 A mutation in theLDL receptor-related protein 5 gene results in the autosomal dom-inant high-bone-mass trait. Am J Hum Genet 70:11–19.

13. Rattner A, Hsieh JC, Smallwood PM, Gilbert DJ, Copeland NG,Jenkins NA, Nathans J 1997 A family of secreted proteins containshomology to the cysteine-rich ligand-binding domain of frizzledreceptors. Proc Natl Acad Sci USA 94:2859–2863.

14. Jones SE, Jomary C 2002 Secreted Frizzled-related proteins:Searching for relationships and patterns. Bioessays 24:811–820.

15. Leyns L, Bouwmeester T, Kim SH, Piccolo S, De Robertis EM1997 Frzb-1 is a secreted antagonist of Wnt signaling expressed inthe Spemann organizer. Cell 88:747–756.

1880 HAUSLER ET AL.

16. Wang S, Krinks M, Lin K, Luyten FP, Moos M Jr 1997 Frzb, asecreted protein expressed in the Spemann organizer, binds andinhibits Wnt-8. Cell 88:757–766.

17. Bafico A, Gazit A, Pramila T, Finch PW, Yaniv A, Aaronson SA1999 Interaction of frizzled related protein (FRP) with Wnt ligandsand the frizzled receptor suggests alternative mechanisms for FRPinhibition of Wnt signaling. J Biol Chem 274:16180–16187.

18. Uren A, Reichsman F, Anest V, Taylor WG, Muraiso K, BottaroDP, Cumberledge S, Rubin JS 2000 Secreted frizzled-relatedprotein-1 binds directly to Wingless and is a biphasic modulator ofWnt signaling. J Biol Chem 275:4374–4382.

19. Hoang B, Moos M Jr, Vukicevic S, Luyten FP 1996 Primarystructure and tissue distribution of FRZB, a novel protein related toDrosophila frizzled, suggest a role in skeletal morphogenesis.J Biol Chem 271:26131–26137.

20. James IE, Kumar S, Barnes MR, Gress CJ, Hand AT, Dodds RA,Connor JR, Bradley BR, Campbell DA, Grabill SE, Williams K,Blake SM, Gowen M, Lark MW 2000 FrzB-2: A human secretedfrizzled-related protein with a potential role in chondrocyte apo-ptosis. Osteoarthritis Cartilage 8:452–463.

21. Baranski M, Berdougo E, Sandler JS, Darnell DK, Burrus LW2000 The dynamic expression pattern of frzb-1 suggests multipleroles in chick development. Dev Biol 217:25–41.

22. Udagawa N, Horwood NJ, Elliott J, Mackay A, Owens J, OkamuraH, Kurimoto M, Chambers TJ, Martin TJ, Gillespie MT 1997Interleukin-18 (interferon-gamma-inducing factor) is produced byosteoblasts and acts via granulocyte/macrophage colony-stimulating factor and not via interferon-gamma to inhibit oste-oclast formation. J Exp Med 185:1005–1012.

23. Chambers TJ, Owens JM, Hattersley G, Jat PS, Noble MD 1993Generation of osteoclast-inductive and osteoclastogenic cell linesfrom the H-2KbtsA58 transgenic mouse. Proc Natl Acad Sci USA90:5578–5582.

24. Owens JM, Gallagher AC, Chambers TJ 1996 Bone cells requiredfor osteoclastic resorption but not for osteoclastic differentiation.Biochem Biophys Res Commun 222:225–229.

25. Udagawa N, Takahashi N, Akatsu T, Sasaki T, Yamaguchi A,Kodama H, Martin TJ, Suda T 1989 The bone marrow-derivedstromal cell lines MC3T3–G2/PA6 and ST2 support osteoclast-likecell differentiation in cocultures with mouse spleen cells. Endocri-nology 125:1805–1813.

26. Umezawa A, Maruyama T, Segawa K, Shadduck RK, Waheed A,Hata J 1992 Multipotent marrow stromal cell line is able to inducehematopoiesis in vivo. J Cell Physiol 151:197–205.

27. Horwood NJ, Elliott J, Martin TJ, Gillespie MT 2001 IL-12 aloneand in synergy with IL-18 inhibits osteoclast formation in vitro.J Immunol 166:4915–4921.

28. Lam J, Nelson CA, Ross FP, Teitelbaum SL, Fremont DH 2001Crystal structure of the TRANCE/RANKL cytokine reveals deter-minants of receptor-ligand specificity. J Clin Invest 108:971–979.

29. Southby J, Murphy LM, Martin TJ, Gillespie MT 1996 Cell-specific and regulator-induced promoter usage and messenger ri-bonucleic acid splicing for parathyroid hormone-related protein.Endocrinology 137:1349–1357.

30. Traianedes K, Findlay DM, Martin TJ, Gillespie MT 1995 Mod-ulation of the signal recognition particle 54-kDa subunit (SRP54)in rat preosteoblasts by the extracellular matrix. J Biol Chem270:20891–20894.

31. Takahashi N, Yamana H, Yoshiki S, Roodman GD, Mundy GR,Jones SJ, Boyde A, Suda T 1988 Osteoclast-like cell formation andits regulation by osteotropic hormones in mouse bone marrowcultures. Endocrinology 122:1373–1382.

32. Horwood NJ, Udagawa N, Elliott J, Grail D, Okamura H, Kuri-moto M, Dunn AR, Martin T, Gillespie MT 1998 Interleukin 18inhibits osteoclast formation via T cell production of granulocytemacrophage colony-stimulating factor. J Clin Invest 101:595–603.

33. Udagawa N, Takahashi N, Katagiri T, Tamura T, Wada S, FindlayDM, Martin TJ, Hirota H, Taga T, Kishimoto T 1995Interleukin(IL)-6 induction of osteoclast differentiation depends on IL-6 re-ceptors expressed on osteoblastic cells but not on osteoclast pro-genitors. J Exp Med 182:1461–1468.

34. Han X, Amar S 2004 Secreted frizzled-related protein 1 (SFRP1)protects fibroblasts from ceramide-induced apoptosis. J Biol Chem279:2832–2840.

35. Kartsogiannis V, Zhou H, Horwood NJ, Thomas RJ, Hards DK,Quinn JM, Niforas P, Ng KW, Martin TJ, Gillespie MT 1999Localization of RANKL (receptor activator of NF kappa B ligand)mRNA and protein in skeletal and extraskeletal tissues. Bone25:525–534.

36. Horwood NJ, Elliott J, Martin TJ, Gillespie MT 1998 Osteotropicagents regulate the expression of osteoclast differentiation factorand osteoprotegerin in osteoblastic stromal cells. Endocrinology139:4743–4746.

37. Khosla S 2001 Minireview: The OPG/RANKL/RANK system.Endocrinology 142:5050–5055.

38. Kawano Y, Kypta R 2003 Secreted antagonists of the Wnt signal-ling pathway. J Cell Sci 116:2627–2634.

39. Zhou H, Kartsogiannis V, Quinn JM, Ly C, Gange C, Elliott J, NgKW, Gillespie MT 2002 Osteoclast inhibitory lectin, a family ofnew osteoclast inhibitors. J Biol Chem 277:48808–48815.

40. Bodine PV, Zhao W, Kharode YP, Bex FJ, Lambert AJ, Goad MB,Gaur T, Stein GS, Lian JB, Komm BS 2004 The Wnt antagonistsecreted frizzled-related protein-1 is a negative regulator of trabec-ular bone formation in adult mice. Mol Endocrinol 18:1222–1237.

41. Xu L, Tan L, Goldring MB, Olsen BR, Li Y 2001 Expression offrizzled genes in mouse costochondral chondrocytes. Matrix Biol20:147–151.

42. Fischer L, Boland G, Tuan RS 2002 Wnt signaling during BMP-2stimulation of mesenchymal chondrogenesis. J Cell Biochem 84:816–831.

43. Stott NS, Jiang TX, Chuong CM 1999 Successive formative stagesof precartilaginous mesenchymal condensations in vitro: Modula-tion of cell adhesion by Wnt-7A and BMP-2. J Cell Physiol180:314–324.

44. Bergwitz C, Wendlandt T, Kispert A, Brabant G 2001 Wntsdifferentially regulate colony growth and differentiation of chon-drogenic rat calvaria cells. Biochim Biophys Acta 1538:129–140.

45. Hartmann C, Tabin CJ 2000 Dual roles of Wnt signaling duringchondrogenesis in the chicken limb. Development 127:3141–3159.

46. Fischer L, Boland G, Tuan RS 2002 Wnt-3A enhances bonemorphogenetic protein-2-mediated chondrogenesis of murineC3H10T1/2 mesenchymal cells. J Biol Chem 277:30870–30878.

47. Kawakami Y, Capdevila J, Buscher D, Itoh T, Rodriguez EstebanC, Izpisua Belmonte JC 2001 WNT signals control FGF-dependentlimb initiation and AER induction in the chick embryo. Cell104:891–900.

48. Hartmann C, Tabin CJ 2001 Wnt-14 plays a pivotal role in induc-ing synovial joint formation in the developing appendicular skel-eton. Cell 104:341–351.

49. Saneyoshi T, Kume S, Amasaki Y, Mikoshiba K 2002 The Wnt/calcium pathway activates NF-AT and promotes ventral cell fate inXenopus embryos. Nature 417:295–299.

50. Takayanagi H, Kim S, Koga T, Nishina H, Isshiki M, Yoshida H,Saiura A, Isobe M, Yokochi T, Inoue J, Wagner EF, Mak TW,Kodama T, Taniguchi T 2002 Induction and activation of thetranscription factor NFATc1 (NFAT2) integrate RANKL signalingin terminal differentiation of osteoclasts. Dev Cell 3:889–901.

Address reprint requests to:Matthew Gillespie, PhD

St. Vincent’s Institute of Medical Research9 Princes Street

Fitzroy, Victoria 3065, AustraliaE-mail: mgillespie@sui.edu.au

Received in original form October 11, 2003; in revised form April6, 2004; accepted May 27, 2004.

1881sFRP-1 INHIBITS OSTEOCLAST FORMATION