Blimp1-mediated repression of negative regulators isrequired for osteoclast differentiationKeizo Nishikawaa,b, Tomoki Nakashimaa,b,c, Mikihito Hayashia,b, Takanobu Fukunagaa,b, Shigeaki Katod,e,Tatsuhiko Kodamaf, Satoru Takahashig, Kathryn Calameh, and Hiroshi Takayanagia,b,c,1

aDepartment of Cell Signaling, Tokyo Medical and Dental University, Tokyo 113-8549, Japan; bGlobal Center of Excellence Program, International ResearchCenter for Molecular Science in Tooth and Bone Diseases, Tokyo 113-8549, Japan; cTakayanagi Osteonetwork Project, Exploratory Research for AdvancedTechnology, Japan Science and Technology Agency, Tokyo 113-8549, Japan; dInstitute of Molecular and Cellular Biosciences, Graduate School of Medicine,University of Tokyo, Tokyo 113-0032, Japan; eKato Nuclear Complex, Exploratory Research for Advanced Technology, Japan Science and Technology Agency,Saitama 332-0012, Japan; fResearch Center for Advanced Science and Technology, Department of Molecular Biology and Medicine, University of Tokyo, Tokyo153-8904, Japan; gInstitute of Basic Medical Sciences and Laboratory Animal Resource Center, University of Tsukuba, Tsukuba 305-8575, Japan; andhDepartment of Biochemistry and Molecular Biophysics, Columbia University College of Physicians and Surgeons, New York, NY 10032

Edited by Anjana Rao, Harvard Medical School, Boston, MA, and approved December 23, 2009 (received for review November 7, 2009)

Regulation of irreversible cell lineage commitment depends on adelicate balance between positive and negative regulators, whichcomprise a sophisticated network of transcription factors. Recep-tor activator of NF-κB ligand (RANKL) stimulates the differentia-tion of bone-resorbing osteoclasts through the induction ofnuclear factor of activated T cells c1 (NFATc1), the essential tran-scription factor for osteoclastogenesis. Osteoclast-specific robustinduction of NFATc1 is achieved through an autoamplificationmechanism, in which NFATc1 is constantly activated by calciumsignaling while the negative regulators of NFATc1 are suppressed.However, it has been unclear how such negative regulators arerepressed during osteoclastogenesis. Here we show that B lym-phocyte-induced maturation protein-1 (Blimp1; encoded byPrdm1), which is induced by RANKL through NFATc1 during osteo-clastogenesis, functions as a transcriptional repressor of anti-osteoclastogenic genes such as Irf8 and Mafb. Overexpression ofBlimp1 leads to an increase in osteoclast formation, and Prdm1-deficient osteoclast precursor cells do not undergo osteoclast dif-ferentiation efficiently. The importance of Blimp1 in bone homeo-stasis is underscored by the observation that mice with anosteoclast-specific deficiency in the Prdm1 gene exhibit a highbone mass phenotype caused by a decreased number of osteo-clasts. Thus, NFATc1 choreographs the determination of cell fatein the osteoclast lineage by inducing the repression of negativeregulators as well as through its effect on positive regulators.

B lymphocyte induced maturation protein 1 | nuclear factor of activated Tcells c1 | osteoclastogenesis | transcription | bone

Osteoclasts are multinucleated cells of monocyte/macrophagelineage that degrade bone matrix. The maintenance of bone

homeostasis is dependent on the balance between bone-resorb-ing osteoclasts and bone-forming osteoblasts (1, 2). Excessivebone resorption by osteoclasts is often associated with diseasesaccompanied by pathological bone loss, including osteoporosisand rheumatoid arthritis (3–6); thus research into the mecha-nisms of osteoclast differentiation is both biologically andclinically important.Receptor activator of NF-κB ligand (RANKL) is a key cyto-

kine that stimulates the osteoclast precursor cells to fuse andundergo differentiation into osteoclasts (7, 8). Macrophage col-ony-stimulating factor (M-CSF) also is an essential cytokine forosteoclastogenesis, which is important for the survival of theosteoclast lineage and the expression of RANK, the receptor forRANKL (9). RANKL binding to RANK results in the recruit-ment of the TNF receptor-associated factor 6 (TRAF6), leadingto the activation of NF-κB and JNK (10). RANK also stimulatesthe induction of Fos and thus activates the activator protein-1(AP-1) complex containing c-Fos (11). NF-κB induces the initialinduction of nuclear factor of activated T cells c1 (Nfatc1), theexpression of which is autoamplified by NFATc1 binding to its

own promoter in cooperation with c-Fos (12). This autoampli-fication mechanism enables a robust induction of Nfatc1 specificto osteoclasts (13).During osteoclastogenesis, NFATc1 is constantly activated by

calcium signaling, which is dependent on Ig-like receptors suchas osteoclast-associated receptor (OSCAR) and triggeringreceptor expressed in myeloid cells-2 (TREM2) (1, 2). Thesereceptors are associated with the adaptor molecules, DNAX-activating protein 12 (DAP12) and Fc receptor common γ sub-unit (FcRγ), both of which contain the immunoreceptor tyrosine-based activation motif (ITAM) in their cytoplasmic domain (14,15). Spleen tyrosine kinase (Syk) is recruited to the tyrosine-phosphorylated ITAM and makes a complex with the Tec familykinases activated by RANK, which efficiently induces phospho-lipase Cγ (PLCγ) phosphorylation leading to the activation ofcalcium signal through inositol triphosphate (16).NFATc1 functions as the key regulator of osteoclast differ-

entiation (13, 17, 18) by inducing fusogenic genes such as Tm7sf4[encoding dendritic cell-specific transmembrane protein (DC-STAMP)] (19, 20) and Atp6v0d2 (20, 21) in addition to a numberof genes (such as Acp5, Calcr, and Itgb3) expressed in maturepolarized osteoclasts that are involved in the regulation of bone-resorbing activity (1). Recent reports indicate that NFATc1activity is negatively regulated by other transcription factors suchas IFN regulatory factor-8 (IRF-8) and v-maf musculoaponeur-otic fibrosarcoma oncogene family, protein B (MafB) duringosteoclastogenesis (22, 23). The expression of such negativeregulators was observed to be repressed through osteoclasto-genesis. This repression is consistent with the notion that highNFATc1 activity is a prerequisite for efficient osteoclastogenesis(12), but the mechanism by which the expression of these anti-osteoclastogenic regulators is repressed during RANKL-inducedosteoclastogenesis has remained obscure.To identify the negative transcriptional regulators in osteoclast

differentiation, we performed a genome-wide screening ofRANKL-inducible transcription factors and found that B lym-phocyte-induced maturation protein-1 (Blimp1) (encoded byPrdm1) is crucial for the repression of anti-osteoclastogenicgenes such as Irf8 andMafb. Blimp1 is a transcriptional repressorthat plays crucial roles in the differentiation and/or function of

Author contributions: K.N., T.N. and H.T. designed research; K.N., M.H., T.F., S.K., T.K., S.T.and C.K. performed research; K.N. analyzed data; and K.N., T.N. and H.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Data deposition: GeneChip data are available for download from the Genome NetworkPlatform (http://genomenetwork.nig.ac.jp/).1To whom correspondence should be addressed. E-mail: taka.csi@tmd.ac.jp.

This article contains supporting information online at www.pnas.org/cgi/content/full/0912779107/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.0912779107 PNAS | February 16, 2010 | vol. 107 | no. 7 | 3117–3122

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

various kinds of cells such as macrophages and lymphocytes (24).In particular, Blimp1 is necessary for B cells to undergo terminaldifferentiation into plasma cells through direct repression oftranscription factors such as Pax5, Bcl6, and Myc, which arespecifically important in the lineage commitment and pro-liferation of B cells (24, 25).Here we show that Blimp1 is induced by RANKL through

NFATc1 and functions as a global repressor of negative regu-lators of NFATc1 during osteoclastogenesis. We also providegenetic evidence that Blimp1 is important for bone homeostasisby its promotion of osteoclastogenesis. These results reveal apreviously undescribed aspect of the NFATc1-mediated tran-scriptional program of osteoclastogenesis, the repression ofnegative regulators by an NFATc1 target gene.

Results and DiscussionGenome-Wide Screening of Transcription Factors Involved inOsteoclastogenesis. To explore the transcription factors involvedin osteoclast differentiation, we performed a genome-widescreening of mRNAs expressed in bone marrow-derived mono-cyte/macrophage precursor cells (BMMs) stimulated byRANKL. We clustered the data of 1,675 transcription factors(26) using k-means algorithm into 22 groups (27). Four clustersof genes exhibited a common expression pattern characterized bya steady increase (by more than 2-fold at each time point afterRANKL treatment) (Fig. 1A, Upper). Clusters I, II, III, and IVcontained 72, 54, 48, and 7 transcription factors, respectively(Fig. 1A, Upper). From these four clusters, we finally selected 13genes that are highly expressed in mature osteoclasts (an averagedifference >500 at 3 days) (Fig. 1A, Lower). Among the 13 genesselected were Nfatc1, Ppargc1b, and Fosl2, which are known tofunction as transcriptional activators in osteoclast differentiation(13, 28, 29). Blimp1 also has been classified with this group oftranscription factors but is known to be a transcriptionalrepressor required for plasma cell function (24).We confirmed that both the protein and mRNA expression of

Prdm1 increased markedly during osteoclastogenesis (Fig. 1 Band C). Prdm1 was dramatically induced in BMMs stimulated byRANKL in the presence of M-CSF (Fig. 1C). It has beenreported that Prdm1 expression increases with the macrophagedifferentiation induced by phorbol-12-myristate 13-acetate (30),but M-CSF treatment alone did not induce the Prdm1 expressionin this time course. To test whether Prdm1 induction by RANKLis dependent on NFATc1, we investigated the Prdm1 expressionin the presence of a calcineurin inhibitor, cyclosporin A. Asexpected, Prdm1 induction was greatly suppressed by the treat-ment with cyclosporin A, suggesting that Prdm1 expression isregulated by NFATc1 (Fig. 1D). We performed a computationalsearch on the sequence from 5 kb upstream to 5 kb downstreamof the transcription start site of the Prdm1 gene and found anNFATc1-binding site in a region conserved in the human andmouse sequences (Fig. 1E). Indeed, ChIP analysis showed thatNFATc1 binds directly to the site in BMMs stimulated withRANKL for 2 days but not in unstimulated BMMs (Fig. 1E).These results suggest that Prdm1 is a direct transcriptional targetof NFATc1 during osteoclastogenesis.

Repression Activity of Blimp1 Is Important for OsteoclastDifferentiation. Although it has been reported previously thatBlimp1 functions as a transcriptional repressor in other cell types(24, 25), transcription factors can function as either a positive or anegative transcriptional regulator in a context-dependent manner(31). To investigate whether Blimp1 acts as a transcriptionalactivator or repressor during osteoclast differentiation, we con-structed Blimp1 variants that constitutively function as a tran-scriptional activator or repressor (32). We fused either thetransactivation domain of herpes simplex virus VP16 or therepressor domain of Drosophila engrailed to Blimp1 (AD-Blimp1

and RD-Blimp1, respectively) (Fig. 2A) and overexpressed thesevariants in BMMs using the retrovirus vector (Fig. 2B) to test theireffects on osteoclast formation. The introduction of RD-Blimp1into BMMs by retroviral transfer led to increased osteoclast dif-ferentiation, whereas the overexpression of AD-Blimp1 inhibited

E

C

B

D

A

Fig. 1. Increased expression of Prdm1 during osteoclastogenesis. (A) Gen-eChip analysis of transcription factor expression during osteoclast differ-entiation. Among the 22 clusters, clusters I to IV were characterized by asteady increase. The gene expression increased mainly in the early (I),intermediate (II), or late (III) phase, or continuously (IV). The log-transformedvalue of the mRNA expression of each gene is depicted by a black line, andthe average of the black lines is shown in red. (B) Immunoblot analysis todetect the expression of Blimp1 protein in BMMs stimulated with RANKL inthe presence of M-CSF. (C) mRNA expression of Prdm1 in BMMs culturedwith or without RANKL. (D) Effect of cyclosporin A (CsA; 0.8 μM) on theexpression of Prdm1 in BMMs stimulated with RANKL for 2 days. **, P < 0.01.(E) Recruitment of NFATc1 to the Prdm1 promoter region. Schematic view ofthe NFATc1-binding site in the regulatory region of Prdm1 (expanded, LowerLeft). DNA sequences conserved between the mouse and human sequenceswere identified using VISTA Genome Browser (http://genome.lbl.gov/vista/index.shtml). Arrowheads indicate the primer set used for ChIP. The sitescolored red represent the NFATc1-binding site. ChIP assay and representativeresults (Lower Right).

3118 | www.pnas.org/cgi/doi/10.1073/pnas.0912779107 Nishikawa et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

osteoclast formation (Fig. 2 C and D). These results suggest thatBlimp1 functions mainly as a transcriptional repressor in thepromotion of osteoclast differentiation.

High Bone Mass Phenotype Induced by Osteoclast-Specific Deficiencyof Prdm1. Because Prdm1-deficient mice are embryonically lethal(25), to investigate the role of Blimp1 in osteoclasts in vivo, wecrossed Prdm1flox/– mice (25) with CtskCre/+ mice (33) to disruptthe Prdm1 gene specifically in the osteoclast lineage (Prdm1ΔOC/–).Microcomputed tomography clearly indicated that the bonevolume was greatly enhanced in Prdm1ΔOC/– mice at the age of 12and 32 weeks (Fig. 3A). The osteoclast number was significantlyreduced in the metaphyseal region, and the bone marrow wasabnormally filled with trabecular bone in Prdm1ΔOC/– mice (Fig.3 B and C). Bone morphometric analysis indicated an increase inbone volume associated with a reduced osteoclast number and adecrease in the indicators of osteoclastic bone resorption (Fig.3C). Osteoblastic parameters were slightly decreased inPrdm1ΔOC/– mice (Fig. 3C), possibly because of an osteoclast–osteoblast coupling mechanism. These results collectively suggestthat the increase in bone mass in Prdm1ΔOC/– mice was caused byimpaired osteoclastic bone resorption owing to a defect inosteoclast differentiation. Having normal tooth eruption, thePrdm1ΔOC/– mice exhibit an incomplete osteopetrotic phenotype,but the results nevertheless suggest that Blimp1 plays a criticallyimportant role in osteoclast differentiation.

Impaired Osteoclastogenesis in Prdm1-Deficient Cells. In vitroosteoclast differentiation was evaluated by counting the multi-nucleated cells (MNCs) positive for the osteoclast marker tar-trate-resistant acid phosphatase (TRAP) after stimulation ofBMMs with RANKL in the presence of M-CSF. The number ofTRAP-positive MNCs was markedly reduced in the Prdm1ΔOC/–

cells compared with the Prdm1flox/flox cells (Fig. 4A). Quantitative

genomic PCR analysis indicates that the Prdm1 gene was excisedefficiently 2 days after RANKL treatment (Fig. 4B). Consistently,the protein and mRNA expression of Prdm1 was diminished inPrdm1ΔOC/– cells 2 days after RANKL treatment (Fig. 4 C andD).These results indicate that the Prdm1 gene was disrupted in themiddle stage of osteoclast differentiation (24–48 h after RANKLstimulation in the 72-h period of differentiation). There was nosignificant difference in the number of CD11b+ cells in the BMMs(Fig. 4E), suggesting that Prdm1ΔOC/– cells contain a normalnumber of osteoclast precursor cells. RANKL-stimulated induc-tion of the osteoclast-specific genes Nfatc1, Acp5, and Ctsk wasseverely reduced in Prdm1ΔOC/– cells (Fig. 4D).

A B C

D

-2

Fig. 2. Blimp1 functions as a transcriptional repressor in the promotion ofosteoclast differentiation. (A) Schematic view of WT Blimp1, RD-Blimp1, andAD-Blimp1. (B) mRNA expression of RD-Blimp1 and AD-Blimp1 after infec-tion with pMX-RD-Blimp1-IRES-EGFP (pMX-RD-Blimp1) and pMX-AD-Blimp1-IRES-EGFP (pMX-AD-Blimp1), respectively. (C) Effects of retroviral expressionof RD-Blimp1 or AD-Blimp1 on the osteoclastogenesis from BMMs stimu-lated with RANKL. *, P < 0.05; **, P < 0.01. (D) Osteoclast formation in BMMstransduced with pMX-RD-Blimp1 or pMX-AD-Blimp1.

C

B

A

Fig. 3. Prdm1ΔOC/− mice exhibit a high bone mass phenotype. (A) Micro-computed tomography (μCT) analysis of the femurs of Prdm1flox/flox andPrdm1ΔOC/– mice (Upper photograph: longitudinal view; Lower photograph:axial view of the metaphyseal region). The parameters are based on the μCTanalysis of the metaphyseal region. *, P < 0.05; **, P < 0.01; N.S., not sig-nificant. (B) Histological analysis of the proximal tibia of Prdm1flox/flox andPrdm1ΔOC/− mice (Left, toluidine blue staining, purple marks cartilage; Cen-ter, von Kossa staining, black indicates mineralized bone; Right, TRAPstaining, red marks osteoclasts). (C) Parameters for osteoclastic boneresorption and osteoblastic bone formation in the bone morphometricanalysis of Prdm1flox/flox and Prdm1ΔOC/– mice at the age of 32 weeks. *, P <0.05; **, P < 0.01; N.S., not significant.

Nishikawa et al. PNAS | February 16, 2010 | vol. 107 | no. 7 | 3119

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

To evaluate the effect of Blimp1 overexpression, we con-structed a retroviral vector carrying the HA-tagged Blimp1 gene.We transduced the Prdm1ΔOC/– cells with the viral vector andevaluated the TRAP-positive MNC formation. Impaired osteo-clast formation in Prdm1ΔOC/– cells was significantly restored bythe overexpression of Blimp1 (Fig. 4F), demonstrating thatBlimp1 has an ability to promote osteoclastogenesis.

Blimp1 Represses Expression of the Irf8 and Mafb Genes inOsteoclasts. What are the target(s) of Blimp1-mediated repres-sion in the promotion of osteoclastogenesis? To identify theputative Blimp1-regulated genes, we performed a genome-widescreening of mRNAs during osteoclastogenesis using WT andPrdm1ΔOC/ΔOC BMMs (Fig. 5A, Left and Center). We first

DB

C E

F

A

-2-2

xx

Fig. 4. Impaired osteoclastogenesis in Prdm1ΔOC/– cells. (A) Osteoclast dif-ferentiation in Prdm1flox/flox and Prdm1ΔOC/– BMMs stimulated with RANKL.**, P < 0.01. (B) Quantitative genomic PCR analysis to detect the recombi-nation efficiency in Prdm1flox/flox and Prdm1ΔOC/– BMMs stimulated byRANKL. (C) Protein expression of Blimp1 in Prdm1flox/flox and Prdm1ΔOC/–

BMMs stimulated with RANKL. (D) Real-time RT-PCR analysis of mRNAs forNfatc1, Acp5, Ctsk, and Prdm1 in Prdm1flox/flox and Prdm1ΔOC/– BMMsstimulated with RANKL. **, P < 0.01. (E) Expression of CD11b in thePrdm1flox/flox and Prdm1ΔOC/– BMMs (flow cytometry). (F) Effect of retroviralexpression of HA-tagged Blimp1 on osteoclastogenesis on Prdm1ΔOC/– BMMsstimulated with RANKL. **,P < 0.01.

A

B

C

E

D

Fig. 5. Blimp1 regulates osteoclast differentiation through the down-regulation of multiple negative regulators. (A) Comprehensive search forthe Blimp1-regulated genes during osteoclastogenesis. The expression of365 genes decreased more than 2-fold every 24 h during osteoclast dif-ferentiation (Left). Of these genes, we identified the expression of 116genes to be sufficiently recovered (>120%) in Prdm1-deficient cells (Cen-ter). Prdm1-deficient cells were generated by retroviral introduction of Crerecombinase into Prdm1flox/flox BMMs. A computational search for theBlimp1-binding site in the region conserved between the human andmouse sequences within 500 kb upstream of the transcriptional start site of116 genes using the ECR Browser (http://ecrbrowser.dcode.org/) showedthat 60 genes, including eight transcription factors, have a putative Blimp1-binding site (Right). The number in parentheses indicates the number ofBlimp1-binding sites. (B) Recruitment of Blimp1 to the Irf8 and Mafb reg-ulatory regions. Schematic view of the Blimp1-binding site in the regulatoryregion of Irf8 and Mafb (Upper). BMMs cultured with RANKL and M-CSF for2 days were used. Arrowheads indicate the primer sets used for ChIP. Sitescolored red represent Blimp1-binding sites. ChIP assay and representativeresults (Lower). (C) Real-time RT-PCR analysis of Irf8 and Mafb mRNAs inPrdm1flox/flox and Prdm1ΔOC/– BMMs stimulated with RANKL for 2 days. *, P< 0.05; **, P < 0.01. (D) Real-time RT-PCR analysis of Irf8 andMafbmRNAs inHA-tagged Blimp1-expressed BMMs stimulated with RANKL for 2 days. **, P< 0.01. (E) A model of Blimp1-mediated regulation of osteoclast differ-

3120 | www.pnas.org/cgi/doi/10.1073/pnas.0912779107 Nishikawa et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

selected genes the expression of which was decreased more than2-fold every 24 h during osteoclast differentiation in WT cells(Fig. 5A, Left). We further selected genes in which the reducedexpression was clearly normalized by the Prdm1 deficiency (Fig.5A, Center). We computationally searched the 5′-flanking regionof these genes for the Blimp1-binding site to identify 60 geneshaving a putative Blimp1-binding site (Fig. 5A, Right). The eighttranscription factors were included in these genes (Fig. 5A,Right). In this study, we focused on Irf8 and Mafb because thesegenes are reported to be negative regulators of osteoclast dif-ferentiation (22, 23). We identified nine Blimp1-binding sites inthe 5′-flanking sequence of Irf8 and three Blimp1-binding sites inthe 5′-flanking sequence of Mafb. A cluster of Blimp1-bindingsites is located ≈5.5 kb upstream from the transcription start siteof the Irf8 gene. Indeed, ChIP analysis showed that Blimp1bound to the sequence containing the cluster in the 5′-flankingsequence of Irf8 (Fig. 5B). Blimp1 also bound to one of the threeBlimp1-binding sites located ≈16 kb upstream from the tran-scription start site of the Mafb gene (Fig. 5B). These resultssuggest that Blimp1 represses the expression of Irf8 and Mafb bydirectly binding to their regulatory regions. Consistent with thisnotion, the expression of Irf8 and Mafb was increased inPrdm1ΔOC/– cells (Fig. 5C), and the forced expression of Blimp1decreased the expression of Irf8 and Mafb significantly (Fig. 5D).These results suggest that the Irf8 and Mafb genes are among thedirect repressive targets of Blimp1 during osteoclastogenesis.To investigate whether Blimp1 promotes osteoclastogenesis by

the down-regulation of Irf8 and Mafb, we generated retroviralvectors carrying Irf8 or Mafb shRNA. However, the introductionof Irf8 and/or Mafb shRNAs into Prdm1ΔOC/– BMMs had only aminimal rescue effect (Fig. S1), suggesting that Blimp1 exerts itsfunction through the regulation of multiple genes, including Irf8andMafb. Pax5 expression is known to be repressed by Blimp1 inplasma cells (24), and Pax5−/− mice exhibit an osteopenic phe-notype because of an increased number of osteoclasts (34),suggesting that Pax5 is a candidate target of Blimp1 in osteo-clasts. However, the contribution of Pax5 to the Blimp1-medi-ated regulation of osteoclastogenesis may be minimal, becausethe expression of Pax5 is negligible during osteoclastogenesisbased on the GeneChip data. Bcl6 also is repressed by Blimp1,and Blimp1 is negatively regulated by BCL6 in plasma cells (24).Although the shRNA-mediated knockdown of Bcl6 did notnormalize the decreased osteoclastogenesis in Prdm1ΔOC/– cells(Fig. S1), we cannot rule out the possibility that Bcl6 may alsocontribute to the Blimp1-mediated regulation of osteoclasto-genesis in cooperation with Irf8 and Mafb.Cell lineage commitment is strictly regulated to maintain the

homeostasis of the biological systems of the multicellularorganisms. To this end, cell differentiation signals usually stim-ulate the negative-feedback regulatory pathways to correct anyexcess. During osteoclastogenesis, however, NFATc1 activity andexpression are maintained at an extremely high level comparedwith other cell types (12). Therefore, the expression of moleculesthat inhibit NFATc1 activity needs to be repressed. Blimp1,which is induced by the RANKL–NFATc1 axis, represents therepressor of anti-osteoclastogenic genes (Fig. 5E). It is likely thatNFATc1 functions as the crucial regulator of osteoclastogenesisboth by inducing osteoclastogenic genes and by acting as arepressor of negative regulatory genes.

Accumulating evidence indicates that immune and bone cellsshare important regulatory molecules, including receptors, sig-naling molecules, enzymes, and transcription factors. Thesefindings have shed light on the new interdisciplinary researchfield of osteoimmunology (1, 2). In particular, an unexpectedlyhigh number of factors that regulate osteoclast differentiationwere identified originally in immune cells (1, 2). Blimp1 also hasbeen well characterized as a repressor of genes involved in earlyB-cell differentiation that enable B cells to undergo terminaldifferentiation into plasma cells (24). The function of Blimp1 inosteoclasts is analogous to its role in plasma cells, in that Blimp1represses the genes involved in macrophage development such asIrf8 and Mafb (35, 36), thus converting monocyte/macrophageprecursor cells into terminally differentiated osteoclasts.Because Blimp1 acts as a global repressor of multiple anti-

osteoclastogenic genes, it is possible that inhibitors of Blimp1function may serve as antiresorptive agents by inducing variousfactors affecting osteoclastogenesis. In depth understanding ofthe gene regulatory programs mediated by Blimp1 should helpprovide a molecular basis for future therapeutic strategies inbone disease.

Materials and MethodsMice and Analysis of Bone Phenotype. Prdm1flox/– and CtskCre/+ mice weregenerated previously and have been described elsewhere (25, 33). Bothstrains were backcrossed with C57/BL6 mice for more than six generations.All mice were born and maintained under specific-pathogen–free con-ditions. All animal experiments were performed with the approval of theAnimal Study Committee of Tokyo Medical and Dental University and con-formed to relevant guidelines and laws. CT scanning was performed using aScanXmate-A100S Scanner (Comscantechno). Three-dimensional micro-structural image data were reconstructed and structural indices were cal-culated using TRI/3D-BON software (RATOC Systems). Bone morphometricanalysis was performed as described (14, 16).

Cell Cultures. The method for in vitro osteoclast differentiation was describedpreviously (13, 14, 16). In brief, bone marrow cells cultured with 10 ng mL−1

M-CSF (R&D Systems) for 2 days were used as BMMs, which were furthercultured with 50 ng mL−1 RANKL (Peprotech) in the presence of M-CSF (10 ngmL−1) for 3 days.

Flow Cytometric Analysis. Two days after stimulation with RANKL, cells werestained with the PE-conjugated anti-CD11b antibody (BD Biosciences) orcontrol rat IgG. Stained cells were analyzed with a flow cytometer (FACS-Canto II; BD Biosciences).

Quantitative PCR Analysis. Total RNA and cDNA were prepared by ISOGEN(WakoChemicals)andSuperScriptIII reversetranscriptase(Invitrogen)accordingto the manufacturer’s instructions. Genomic DNA was obtained by phenol/chloroformextractionandethanol precipitation.Quantitative PCR analysiswasperformed with a LightCycler (Roche) using SYBR Green (Toyobo). The primersequences used in real-time RT-PCR analysis were Actb, 5′-CTTCTACAATG-AGCTGCGTG-3′ and 5′-TCATGAGGTAGTCTGTCAGG-3′; Acp5, 5′-GGGAAAT-GGCCAATGCCAAAGAGA-3′and5′-TCGCACAGAGGGATCCATGAAGTT-3′;Ctsk,5′-AGGCAGCTAAATGCAGAGGGTACA-3′ and 5′-ATGCCGCAGGCGTTGTTCTT-ATTC-3′; Nfatc1, 5′-GGTAACTCTGTCTTTCTAACCTTAAGCTC-3′ and 5′-GTGAT-GACCCCAGCATGCACCAGTCACAG-3′; Prdm1, 5′-TGCTTATCCCAGCACCCC-3′and 5′-CTTCAGGTTGGAGAGCTGACC-3′; Mafb, 5′-AACGGTAGTGTGGAGGAC-3′ and 5′-TCACAGAAAGAACTCAGGA-3′; Irf8, 5′-AAGGTCACCGTGGTCCTTAG-3′ and 5′-GGAAAGCCTTACCTGCTGAC-3′; engrailed, 5′-TCGAATTCAAGG-TATGGCCCTGGAG-3′ and 5′-GGGAATTCGCTCGAGAGGGATCCCAGAGCA-3′;VP16, 5′-GGGAATTCGGATCCCATCGATATGACCGGA-3′ and 5′-GGGAATTC-TTCTAGAGGCTCGAGCCCACCG-3′.

The primer sequences used in quantitative genomic PCR analysis werePrdm1 third exon, 5′-gctgcccccaagtctagc-3′ and 5′-cttcaaactcggcctctgtcc-3′;Prdm1 sixth intron/seventh exon, 5′-aagaggcttgcatctgcatt-3′ and 5′-tgtgg-cttctctcctgtgtg-3′.

GeneChip Analysis. GeneChip analysis was performed as described (13, 14, 16).Briefly, the total RNAs were used for cDNA synthesis by reverse transcriptionfollowed by synthesis of biotinylated cRNA through in vitro transcription.After cRNA fragmentation, hybridization with mouse genome 430 2.0 array

entiation. During osteoclastogenesis, the critical transcription factorNFATc1 is induced by RANKL and ITAM signals. NFATc1 is known to regu-late various osteoclastogenic genes. Anti-osteoclastogenic transcriptionfactors such as IRF-8 and MafB inhibit osteoclast differentiation mainlythrough the suppression of NFATc1 activity. Blimp1, which also is inducedby the RANKL-NFATc1 axis, represents the repressor of anti-osteoclasto-genic genes.

Nishikawa et al. PNAS | February 16, 2010 | vol. 107 | no. 7 | 3121

MED

ICALSC

IENCE

S

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020

(Affymetrix) was performed. The main part of the dataset has been depos-ited and can be obtained from the Genome Network Platform (http://genomenetwork.nig.ac.jp/). K-means clustering analysis of gene expressionon their Euclidean distance was performed with the Cluster 3.0 (27).

ChIP Assay. ChIP assay was performed by the ChIP Assay Kit (Upstate)according to the manufacturer's instructions with minor modifications. Inbrief, BMMs stimulated with RANKL for 2 days were fixed with form-aldehyde. After quenching with glycine, cells were lysed and sonicated.Sonicated chromatin was incubated with either rabbit anti-Blimp1 poly-clonal antibody serum (37) or anti-NFATc1 antibody (Santa Cruz). We usedrabbit serum or rabbit IgG (Santa Cruz) as a control. Immunoprecipitationwas performed using protein A Dynabeads (Invitrogen). After extensivewashing, protein-DNA crosslinks were reversed, and the precipitated DNAwas treated with proteinase K before phenol chloroform extraction andethanol precipitation. The primer sequences were Prdm1, 5′-AAACGTGTGGGTACGACCTT-3′ and 5′-AGCGCTGGTTTCTACTGAGG-3′; Irf8,5′-TCCCTCCCTCCTTCTCCTTA-3′ and 5′-AAGCCCTGAGTGCACAGACT-3′; Mafbsite 1, 5′-CAGAGGCAAAGGAACAGCTC-3′ and 5′-GCACCCTGGACTGTAAC-CAT-3′; Mafb site 2, 5′-CCTTGCCTTGTGTCCTGAAG-3′ and 5′-GAGGGGG-TATGAAGGAGAGG-3′; Mafb site 3, 5′-GCGTCAGCTTCCCAGTAGAA-3′ and 5′-CGACTGGGAGCGTAGTTAGC-3′.

Plasmid Construction and Retroviral Gene Transfer. A retroviral vector, pMX-HA-Blimp1, was constructed by inserting DNA fragments encoding HA andBlimp1 into a retroviral expression vector, pMx-IRES-GFP (38). The retroviralvectors pMX-RD-HA-Blimp1 and pMX-AD-HA-Blimp1 were constructed byinserting DNA fragments encoding engrailed or VP16 (32) into pMX-HA-Blimp1. A retroviral vector, pMX-Cre, has been described elsewhere (39).Retroviral packaging was performed by transfecting the plasmids into Plat-Ecells using FuGENE 6 (Roche) as described (38). Retroviral particles were usedfor infection after concentration by centrifugation at 8,000 × g for 16 h.After 12-h inoculation, BMMs were stimulated with RANKL.

Knockdown Analysis. We used the RNAi system (BD Biosciences) and con-structed the Irf8 and Mafb RNAi-Ready pSIREN vector according to the man-ufacturer's protocol. Briefly, annealed oligonucleotides (Irf8 RNAi sense:gatcCCGCTGAAATGGAGTTAGTTTCTCGAGAAACTAACTCCATTTCAGCGGTTT-TTG; Irf8 RNAi antisense: aattCAAAAACCGCTGAAATGGAGTTAGTTTCTCGA-GAAACTAACTCCATTTCAGCGG; Mafb RNAi sense: gatcGCTGAGTCTTTGTTTGGGTTTCTCGAGAAACCCAAACAAAGACTCAGCTTTTTG; Mafb RNAi anti-sense: aattCAAAAAGCTGAGTCTTTGTTTGGGTTTCTCGAGAAACCCAAACAAAGACTCAGC; Bcl6 RNAi sense: gatcGCTGTCAAAGAGAAGGCTTTACTCGAGTAAAGCCTTCTCTTTGACAGCTTTTTG;Bcl6RNAi antisense: aattCAAAAAGCTGTCAAAGAGAAGGCTTTACTCGAGTAAAGCCTTCTCTTTGACAGC) were ligatedinto a pSIREN vector. After 12-h inoculation with retroviruses carrying thepSIREN vector, BMMswere stimulatedwith RANKL followed by selectionwith2.5 μg mL−1 puromycin (Invitrogen).

Immunoblot Analysis. Immunoblot analysis was performed using antibodiesagainst Blimp1 (NOVUS), HA (Abgent), and Laminb1 (Santa Cruz) as previouslydescribed (40). Anti-Laminb1 antibody was used as an internal control.

Statistical Analysis. Statistical analysis was performed using the unpaired two-tailed Student's t test. All data are expressed as the mean ± SEM.

ACKNOWLEDGMENTS. We thank T. Ando, Y. Kunisawa, M. Shinohara, M.Oh-hora, A. Izumi, and A. Hirota for discussion and technical assistance. Wealso thank M. Kobayashi for the gift of pCS2-En and pCS2-VP16. This workwas supported in part by a grant to the Exploratory Research for AdvancedTechnology Takayanagi Osteonetwork Project from Japan Science andTechnology Agency; Grants-in-Aid for Creative Scientific Research andYoung Scientist (A and Start-up) from the Japan Society for the Promotionof Science (JSPS); a Grant-in-Aid for Challenging Exploratory Research fromthe JSPS; grants for the Genome Network Project and Global Center ofExcellence Program from the Ministry of Education, Culture, Sports, Scienceand Technology of Japan; and by grants from the Tokyo BiochemicalResearch Foundation, the Life Science Foundation of Japan, the YokoyamaFoundation for Clinical Pharmacology, and the Takeda Science Foundation.

1. Takayanagi H (2007) Osteoimmunology: Shared mechanisms and crosstalk betweenthe immune and bone systems. Nat Rev Immunol 7:292–304.

2. Lorenzo J, Horowitz M, Choi Y (2008) Osteoimmunology: Interactions of the bone andimmune system. Endocr Rev 29:403–440.

3. Goldring SR, Gravallese EM (2000) Mechanisms of bone loss in inflammatory arthritis:Diagnosis and therapeutic implications. Arthritis Res 2:33–37.

4. Kong YY, et al. (1999) Activated T cells regulate bone loss and joint destruction inadjuvant arthritis through osteoprotegerin ligand. Nature 402:304–309.

5. Takayanagi H, et al. (2000) T-cell-mediated regulation of osteoclastogenesis bysignalling cross-talk between RANKL and IFN-γ. Nature 408:600–605.

6. Rodan GA, Martin TJ (2000) Therapeutic approaches to bone diseases. Science 289:1508–1514.

7. Theill LE, Boyle WJ, Penninger JM (2002) RANK-L and RANK: T cells, bone loss, andmammalian evolution. Annu Rev Immunol 20:795–823.

8. Xing L, Schwarz EM, Boyce BF (2005) Osteoclast precursors, RANKL/RANK, andimmunology. Immunol Rev 208:19–29.

9. Ross FP, Teitelbaum SL (2005) α/β3 and macrophage colony-stimulating factor:Partners in osteoclast biology. Immunol Rev 208:88–105.

10. Kobayashi N, et al. (2001) Segregation of TRAF6-mediated signaling pathwaysclarifies its role in osteoclastogenesis. EMBO J 20:1271–1280.

11. Wagner EF, Eferl R (2005) Fos/AP-1 proteins in bone and the immune system. ImmunolRev 208:126–140.

12. Asagiri M, et al. (2005) Autoamplification of NFATc1 expression determines itsessential role in bone homeostasis. J Exp Med 202:1261–1269.

13. Takayanagi H, et al. (2002) Induction and activation of the transcription factorNFATc1 (NFAT2) integrate RANKL signaling in terminal differentiation of osteoclasts.Dev Cell 3:889–901.

14. Koga T, et al. (2004) Costimulatory signals mediated by the ITAM motif cooperatewith RANKL for bone homeostasis. Nature 428:758–763.

15. Mócsai A, et al. (2004) The immunomodulatory adapter proteins DAP12 and Fcreceptor γ-chain (FcRγ) regulate development of functional osteoclasts through theSyk tyrosine kinase. Proc Natl Acad Sci USA 101:6158–6163.

16. Shinohara M, et al. (2008) Tyrosine kinases Btk and Tec regulate osteoclastdifferentiation by linking RANK and ITAM signals. Cell 132:794–806.

17. Winslow MM, et al. (2006) Calcineurin/NFAT signaling in osteoblasts regulates bonemass. Dev Cell 10:771–782.

18. Aliprantis AO, et al. (2008) NFATc1 in mice represses osteoprotegerin duringosteoclastogenesis and dissociates systemic osteopenia from inflammation incherubism. J Clin Invest 118:3775–3789.

19. Yagi M, et al. (2007) Induction of DC-STAMP by alternative activation anddownstream signaling mechanisms. J Bone Miner Res 22:992–1001.

20. Kim K, Lee SH, Ha Kim J, Choi Y, Kim N (2008) NFATc1 induces osteoclast fusion via up-regulation of Atp6v0d2 and the dendritic cell-specific transmembrane protein (DC-STAMP). Mol Endocrinol 22:176–185.

21. Lee SH, et al. (2006) v-ATPase V0 subunit d2-deficient mice exhibit impaired osteoclast

fusion and increased bone formation. Nat Med 12:1403–1409.22. Zhao B, et al. (2009) Interferon regulatory factor-8 regulates bone metabolism by

suppressing osteoclastogenesis. Nat Med 15:1066–1071.23. Kim K, et al. (2007) MafB negatively regulates RANKL-mediated osteoclast

differentiation. Blood 109:3253–3259.24. Martins G, Calame K (2008) Regulation and functions of Blimp-1 in T and B

lymphocytes. Annu Rev Immunol 26:133–169.25. Shapiro-Shelef M, et al. (2003) Blimp-1 is required for the formation of immunoglobulin

secreting plasma cells and pre-plasma memory B cells. Immunity 19:607–620.26. Kanamori M, et al. (2004) A genome-wide and nonredundant mouse transcription

factor database. Biochem Biophys Res Commun 322:787–793.27. Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of

genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868.28. Bozec A, et al. (2008) Osteoclast size is controlled by Fra-2 through LIF/LIF-receptor

signalling and hypoxia. Nature 454:221–225.29. Ishii KA, et al. (2009) Coordination of PGC-1β and iron uptake in mitochondrial

biogenesis and osteoclast activation. Nat Med 15:259–266.30. Chang DH, Angelin-Duclos C, Calame K (2000) BLIMP-1: Trigger for differentiation of

myeloid lineage. Nat Immunol 1:169–176.31. Alvarez M, Rhodes SJ, Bidwell JP (2003) Context-dependent transcription: All politics

is local. Gene 313:43–57.32. Kessler DS (1997) Siamois is required for formation of Spemann’s organizer. Proc Natl

Acad Sci USA 94:13017–13022.33. Nakamura T, et al. (2007) Estrogen prevents bone loss via estrogen receptor alpha

and induction of Fas ligand in osteoclasts. Cell 130:811–823.34. Nutt SL, Heavey B, Rolink AG, Busslinger M (1999) Commitment to the B-lymphoid

lineage depends on the transcription factor Pax5. Nature 401:556–562.35. Tamura T, Nagamura-Inoue T, Shmeltzer Z, Kuwata T, Ozato K (2000) ICSBP directs

bipotential myeloid progenitor cells to differentiate into mature macrophages.

Immunity 13:155–165.36. Moriguchi T, et al. (2006) MafB is essential for renal development and F4/80

expression in macrophages. Mol Cell Biol 26:5715–5727.37. Tunyaplin C, et al. (2004) Direct repression of prdm1 by Bcl-6 inhibits plasmacytic

differentiation. J Immunol 173:1158–1165.38. Morita S, Kojima T, Kitamura T (2000) Plat-E: An efficient and stable system for

transient packaging of retroviruses. Gene Ther 7:1063–1066.39. Sato S, et al. (2005) Essential function for the kinase TAK1 in innate and adaptive

immune responses. Nat Immunol 6:1087–1095.40. Nishikawa K, et al. (2003) Self-association of Gata1 enhances transcriptional activity in

vivo in zebra fish embryos. Mol Cell Biol 23:8295–8305.

3122 | www.pnas.org/cgi/doi/10.1073/pnas.0912779107 Nishikawa et al.

Dow

nloa

ded

by g

uest

on

Sep

tem

ber

26, 2

020