JBMR
ORIGINAL ARTICLE
Bone Sialoprotein Deficiency Impairs Osteoclastogenesis and Mineral Resorption In Vitro Maya Boudiffa, 1,2,4 Nde´ye Marie`me Wade-Gueye, 1,2,4 Alain Guignandon, 1,2,4 Arnaud Vanden-Bossche, 1,2,4 Odile Sabido , 1,3,4 Jane E Aubin , 5 Pierre Jurdic , 6 Laurence Vico , 1,2,4 Marie He´le`ne Lafage-Proust , 1,2,4 and Luc Malaval1,2,4 1
Universite´ de Lyon, Saint-Etienne, France INSERM U890/IFR143, F42023, Saint-Etienne, France 3 Centre Commun de Cytome´trie en Flux, Faculte´ de Me´decine, Saint-Etienne, France 4 Universite´ Jean Monnet, Saint-Etienne, France 5 Department of Molecular Genetics, University of Toronto, Toronto, Ontario, Canada 6 UMR 5161 CNRS/ENS, Unite´ Mixte INRA/CNRS 1237, Ecole Normale Superieure and IFR 128, Lyon, France 2
ABSTRACT Bone sialoprotein (BSP) and osteopontin (OPN) belong to the small integrin-binding ligand N-linked glycoprotein (SIBLING) family, whose members interact with bone cells and bone mineral. Previously, we showed that BSP knockout (BSP/) mice have a higher bone mass than wild type (BSPþ/þ) littermates, with very low bone-formation activity and reduced osteoclast surfaces and numbers. Here we report that approximately twofold fewer tartrate-resistant acid phosphatase (TRACP)–positive cells and approximately fourfold fewer osteoclasts form in BSP/ compared with BSPþ/þ spleen cell cultures. BSP/ preosteoclast cultures display impaired proliferation and enhanced apoptosis. Addition of RGD-containing proteins restores osteoclast number in BSP/ cultures to BSPþ/þ levels. The expression of osteoclast-associated genes is markedly altered in BSP/ osteoclasts, with reduced expression of cell adhesion and migration genes (aV integrin chain and OPN) and increased expression of resorptive enzymes (TRACP and cathepsin K). The migration of preosteoclasts and mature osteoclasts is impaired in the absence of BSP, but resorption pit assays on dentine slices show no significant difference in pit numbers between BSPþ/þ and BSP/ osteoclasts. However, resorption of mineral-coated slides by BSP/ osteoclasts is markedly impaired but is fully restored by coating the mineral substrate with hrBSP and partly restored by hrOPN coating. In conclusion, lack of BSP affects both osteoclast formation and activity, which is in accordance with in vivo findings. Our results also suggest at least some functional redundancy between BSP and OPN that remains to be clarified. ß 2010 American Society for Bone and Mineral Research. KEY WORDS: BSP; OSTEOCLAST; RESORPTION; SIBLING
Introduction
M
ultinucleated bone-resorbing osteoclasts are unique polykaria originating from the hematopoietic lineage. In response to certain stimuli [receptor activator for nuclear factor kb ligand (RANKL), macrophage colony-stimulating factor (MCSF)], osteoclast precursors commit and become tartrateresistant acid phosphatase (ACP5, TRACP)–positive mononucleated osteoclasts [preosteoclasts (POCs)]. POCs migrate and fuse to form mature multinucleated osteoclasts (OCs). OCs express other specific markers in addition to TRACP, including calcitonin receptor (CTR), cathepsin K (CatK), and matrix
metalloproteinase 9 (MMP9), and are able to degrade the mineralized bone matrix.(1) Bone sialoprotein (BSP) and osteopontin (OPN), which belong to the small integrin-binding N-linked glycoprotein (SIBLING) family(2,3) are noncollagenous RGD-containing bone matrix proteins. They are highly expressed by hypertrophic chondrocytes and osteoblasts, as well as osteoclasts,(4–7) and their functional roles in these cells are only beginning to be understood.(8) In vitro studies of bone resorption suggest that RGD-containing proteins such as OPN and BSP enhance osteoclast attachment to bone through their interaction with the aVb3 integrin.(9)
Received in original form October 28, 2009; revised form April 21, 2010; accepted July 7, 2010. Published online September 1, 2010. Address correspondence to: Luc Malaval, PhD, INSERM U890, Universite´ Jean Monnet, Faculte´ de Me´decine, 15. rue A. Pare´, 42023 St Etienne, Cedex 02, France. E-mail:
[email protected] Journal of Bone and Mineral Research, Vol. 25, No. 12, December 2010, pp 2669–2679 DOI: 10.1002/jbmr.245 ß 2010 American Society for Bone and Mineral Research
The December 2010 issue of Journal of Bone and Mineral Research was published online on 23 Nov 2010. A pagination error was subsequently identified. This notice is included to indicate that the pagination is now correct and authoritative [20 January 2011].
2669
The generation of mice lacking OPN (OPN/) provided valuable information on the role of this abundant SIBLING protein in bone resorption: OPN/ mice display a higher bone mass with a normal bone-formation rate (BFR) and more numerous but poorly resorbing OCs.(10) In vitro studies showed that OPN/ osteoclasts are hypomotile and less active in bone resorption than wild-type OCs.(10) Recently, we showed that adult BSP knockout mice (BSP/) display a higher trabecular bone mass despite a reduced BFR and lower osteoclast surfaces and number, suggesting a dominant role of impaired bone resorption in the bone phenotype observed.(11) Raynal and colleagues(12) showed that addition of exogenous BSP promotes bone resorption when mature osteoclasts are seeded onto dentine slices, most likely through a BSP/aVb3 integrin interaction. However, when exogenous BSP was added to cocultures of osteoblasts and spleen cells, the number of OCs generated was dose-dependently decreased.(12) Valverde and colleagues(13) reported a synergistic effect of BSP and RANKL on the survival of mature osteoclasts and showed that overexpression of BSP leads to a higher number of multinucleated OCs with higher resorptive ability.(6) Notably, we showed previously that fewer POCs and OCs form in cultures of BSP/ spleen or bone marrow cells.(11) While these results clearly indicate a role for BSP in OC formation and activity, the mechanisms underlying these effects remain unclear. In this study, we used the BSP/ mouse model to investigate further the role of BSP in osteoclastogenesis and resorption. We report that endogenous BSP regulates POC number, motility, fusion, and expression of OC-related genes and that multinucleated OC formation and activity can be restored by exogenous human recombinant BSP (hrBSP) but only partly by hrOPN.
Materials and Methods Generation of spleen-derived osteoclasts The production of BSP knockout mice on a 129/CD1 background and their phenotype have been described extensively in ref. (11). Briefly, exons II-III of the mouse Bsp gene were replaced by a PGKneo cassette that created a null allele in mouse R1embryonic stem cells kindly provided by Dr. Andras Na¨gy.(14) BSP/ and BSPþ/þ mice were euthanized by cervical dislocation, spleens were removed, and OC precursors were isolated by centrifugation using the Lympholyte cell separation medium (Cedarlane, TEBU-BIO, Le Perray en Yvelines, France) for 20 minutes at 2500 rpm. Cells then were washed twice with aMEM, counted, and plated in 24-well plates at 25 104 cells per well in a differentiation medium containing 50 ng/mL of RANKL and 20 ng/mL of M-CSF (Peprotech, France). Cells were grown for 3 or 7 days, and the medium was changed every other day. On day 3 or day 7, cells were fixed with 2% paraformaldehyde (PFA), washed with phosphate-buffered saline (PBS), and incubated with a mixture of 2 mg/mL of Naphtol AS-TR Phosphate (SigmaAldrich, St Quentin Fallavier, France) and 5 mg/mL of Fast Violet B Salt (Sigma-Aldrich) for 1 hour at 40 8C. TRACPþ cells were counted under a light microscope. Mononucleated TRACPþ cells on day 3 were designated POCs, and TRACPþ cells on day 7 with three or more nuclei were designated OCs. Size distribution of
2670
Journal of Bone and Mineral Research
OCs (number of nuclei/cell) was normalized to the total number of TRACPþ cells on day 7. In some experiments, 5 or 10 ng/mL of bovine serum albumin (BSA, ICN Biomedicals, Inc., USA), human fibronectin (hFN, BD Bioscience, Mountainview, CA, USA), and human recombinant (hr) BSP or hrOPN (ImmunDiagnostik AG, Bensheim, Germany) were added to the culture medium throughout the differentiation process. For survival and resorption assays with OCs, the proportion of TRACPþ cells with three or more nuclei in BSPþ/þ and BSP/ day 7 cultures was first determined, and the number of cells plated was adjusted to ensure equal densities of OCs in the assay wells.
Flow cytometry Spleen cells pooled from three BSPþ/þ and three BSP/ mice were immunolabeled according to standard procedures using the following antibodies: anti-CD11b/Mac1/aMb2 integrin (coupled with PerCP-Cy5.5; BD Pharmingen, Le Pont de Claix, France) and anti-CD115/c-Fms (coupled with phycoerythrin; BD Pharmingen) for 1 hour at 4 8C. Using a fluorescence-activated cell-sorting (FACS) DiVa cytometer (BD Biosciences, Mountainview, CA, USA), the viable cell population and the monocyte subset were gated based on forward scatter, and the proportions of CD11bþ and CD115þ cells were determined. The experiment was done twice independently.
mRNA extraction and real-time polymerase chain reaction (PCR) Total RNA was extracted from spleen cell cultures grown for 3 and 7 days with M-CSF and RANKL using TRI Reagent (SigmaAldrich) according to the manufacturer’s instructions. Samples were reverse transcribed, and 400 ng of cDNA (RNA equivalent) was amplified through RT-PCR using the LightCycler FastStart DNA MasterPLUS SYBR Green I PCR Master Mix (Roche Applied Science, Meylan, France) on the LightCycler carousel-based system (Roche Applied Science). Expression of mRNA of the genes of interest was normalized to expression of cyclophilin (cylophilin mRNA expression was not changed by the differentiation state of the cells or the conditions used). PCR primer sequences are listed in Table 1.
Cell migration assay Cell migration assays were performed using Transwell plates (Corning, USA). POC cultures (on day 3) or OC cultures (on day 7) were detached with Accutase (Sigma-Aldrich), plated in the upper chambers (2 104 cells/chamber), and allowed to attach for 2 hours. Then 100 ng/mL of human recombinant vascular endothelial growth factor A (hrVEGF-A), human recombinant parathyroid hormone–related protein (hrPTHrP; Sigma-Aldrich), murine recombinant macrophage colony-stimulating factor (mrM-CSF), or mrRANKL was added or not (controls) to the lower chambers, and the cells were allowed to migrate for 20 hours. The membranes then were fixed with 2% paraformaldehyde (PFA) and washed with PBS, and the cells were scraped from their lower surface using cotton buds. The cells remaining on the upper surface were stained with Hoechst 33258 (10 mg/mL) for 10 minutes in the dark, and then the membranes were washed twice with PBS, removed carefully from the insert, BOUDIFFA ET AL.
Table 1. PCR Primers in the 5’!3’ Direction Gene Cylcophilin A RANK c-Fms c-Fos NFATc1 CTR Integrin aV Integrin b3 CD44 DC-STAMP Atp6V0D2 BSP OPN CatK TRAP MMP9 MMP2
Sequencea
PrimerBank ID
F: GGTGACTTTACACGCCATAATG R: GGCTTCCACAATGTTCATGCC F: CCAGGACAGGGCTGATGAGA R: TGGCTGACATACACCACGATGA F: GAGAGCATCTTTGACTGCGTC R: ACAGGCTGGGCCATTTGGTA F: CGGGTTTCAACGCCGACTAC R: AAAGTTGGCACTAGAGACGGACAGA F: GGTCTTCCGAGTTCACATCC R: GCCTTCTCCACGAAAATGAC F: GTCTTGCAACTACTTCTGGATGC R: AAGAAGAAGTTGACCACCAGAGC F: CTGCACGGCAGATACAGAGA R : CCACAGCCCAAAGTGTGAACA F: CGCATCCCATTTGCTAGTGTT R: AATGCCTGCCAGTCTTCCAT F: CCTGCCGCTACGCAGGTGTA R: GGAGGTGTTGGACGTGACGA F: GGGCACCAGTATTTTCCTGA R: CAGAACGGCCAGAAGAATGA F: GTTGCTATCCAGGACTCGGA R: GTCATGTAGGTGAGAAATGT R: AAAGTGAAGGAAAGCGACGA F: GTTCCTTCTGCACCTGCTTC F: TCTGATGAGACCGTCACTGC R: TCTCCTGGCTCTCTTTGGAA F: CTCTCGGCGTTTAATTTGGGAG R: CCCTGGTTCTTGACTGGAGTAACG F: GACAAGAGGTTCCAGGAGACC R: GACAAGAGGTTCCAGGAGACC F: TGAATCAGCTGGCTTTTGTG R: GTGGATAGCTCGGTGGTGTT F: GGAGACAAGTTCTGGAGATA R: GTGTGCAGCGATGAAGATGA
NM_008907.1 NM_009399.3 NM_001037859.2 NM_010234.2 NM_016791 NM_007588.2 NM_008402 NM_016780 NM_009851.2 NM_029422.2 NM_175406.2 NM-008318 NM_009263 NM_007802.3 NM_007388.2 NM_013599.2 NM_008610
a
F ¼ forward; R ¼ reverse.
and mounted. Nonmigrated cells on the upper side of the membranes were counted using the ImageJ software (http:// rsb.info.nih.gov/ij/), clusters of 3 or more nuclei being taken as OCs, and the values were normalized to values in control wells.
Adhesion test At 20-minute intervals after seeding, spleen cell cultures were washed twice with PBS to rinse off unattached cells, fixed with 2% PFA, stained with toluidine blue, and counted under a light microscope.
Survival assay Day 7 cultures were washed once with PBS and detached with Accutase (Sigma-Aldrich). Twenty-four-well plates were seeded with appropriate amounts of cell suspension so that the same number of OCs was plated per well. The plates were cultivated for 2 (day 0), 24, 48, or 72 hours. Cultures then were fixed with 2% BSP DEFICIENCY IMPAIRS OSTEOCLASTOGENESIS
PFA, washed with PBS, and stained for TRACP activity (as described earlier) and the OCs were counted under a light microscope.
Immunolabeling and assays for proliferation and apoptosis Spleen cells were seeded onto 12-mm-diameter glass coverslips and cultured for 3 and 7 days with M-CSF and RANKL. At the times indicated, some coverslips were fixed with 2% PFA, permeabilized in methanol, and preincubated with 2% BSA in PBS for 30 minutes at room temperature to minimize nonspecific staining and then for 1 hour with primary antibodies [rabbit polyclonal antimouse OPN (Abcam, Cambridge, UK) and rabbit polyclonal antimouse BSP (kindly provided by Dr C Chenu, Royal Veterinary College, London, UK), produced as described in ref. (12), diluted 1:200 in PBS, 2% BSA (PBS/BSA)]. After three washes in PBS, the cells were incubated for 1 hour at room temperature with Alexa 488–labeled goat antirabbit secondary Journal of Bone and Mineral Research
2671
antibody (1:200) and 10 mg/mL of Hoechst 33258 (1:1000) in PBS/ BSA. After rinsing, the coverslips were mounted on glass slides with Fluokeep (Argene, Verniolle, France) and observed under a Leica DMRB microscope (equipped with a CoolSnapfx camera; Roper Scientific, Lisses, France). In some experiments, cells were permeabilized with 0.1% Triton X-100 for 3 minutes and then stained with phalloidin rhodamine (Invitrogen, Cergy-Pontoise, France) to label the actin cytoskeleton. For proliferation assays, 3-day-old spleen cell cultures, supplemented or not with 5 ng/mL of hrBSP, were fixed, permeabilized, and immunolabeled for the proliferation marker Ki67 (Sigma) using the horseradish peroxidase Vectastain kit (AbCys, Paris, France) according to the manufacturer’s instructions and the DAB substrate kit (AbCys). Other coverslips were washed twice with cold PBS and then incubated with 0.5 mM Apopercentage dye (Apopercentage, Biocolor, Carrickfergus, UK) in fresh medium for 30 minutes. After rinsing, the coverslips were mounted on glass slides, observed, and imaged. As a positive control, some wells were incubated with 40 mM CaCl2 for 1.5 hours prior to dye addition. Image analysis and cell counting were performed with the ImageJ software.
Mineral substrate resorption assays BSPþ/þ and BSP/ OC cultures were washed once with PBS, and the cells were detached with Accutase (Sigma-Aldrich). Cultures grown in parallel were fixed and stained with TRACP, and the OCs were counted. Appropriate amounts of cell suspension were plated either on dentin slices or on Osteologic disks (BD BIOcoat, Franklin Lakes, NJ, USA) to achieve a density of 500 OCs/cm2. In some experiments, Osteologic slides were precoated overnight at 48C with 100 mg/mL each of hrBSP, hrOPN, or hrBSP þ hrOPN and rinsed in culture medium before plating. Protein assays (microBCA, Thermo Fisher Scientific, Eragny, France) done on recovered coating solutions indicated a 95% coating efficiency. After 2 or 3 days of culture, cells were lysed with distilled water for 2 hours at room temperature, and then dentin slices were stained with toluidine blue and Osteologic discs with the Von Kossa mineral staining procedure. The number and mean area of resorption pits were assessed with the ImageJ software.
Statistical analysis Data are presented as mean SEM; N is indicated in the figure legends. For statistical testing, the Mann-Whitney U test was used whenever two conditions were compared and for RT-PCR data, and one-way ANOVA with the Bonferroni post-test was performed for other multicondition experiments using the Instat software (Version 3.00, Graphpad Software, San Diego, CA, USA).
Results Preosteoclast and multinucleated osteoclast formation is impaired in the absence of BSP In response to RANKL and M-CSF, OC precursors commit to the OC lineage and differentiate to TRACPþ POCs. As previously published,(11) the number of POCs formed in BSP/ spleen cell cultures was up to 1.7 times lower than in BSPþ/þ cultures (Fig. 1A). POCs migrate and fuse to form mature OCs. A lower
2672
Journal of Bone and Mineral Research
Fig. 1. Osteoclastogenesis in the absence of BSP. (A) Number of POCs formed from BSPþ/þ (þ/þ) and BSP/ (/) spleen cells after 3 days of culture, (B) micrographs, and (C) numbers of OCs formed from þ/þ and / spleen cells after 7 days of culture. (D) Size distribution (number of nuclei per cell) of OCs. Values are mean SEM of 12 wells. p < .05; p < .001 versus matched þ/þ.
number of mature OCs (Fig. 1B, C) containing fewer nuclei (Fig. 1D) formed in BSP/ compared with BSPþ/þ cultures.
Expression of osteoclast markers is markedly deregulated in the absence of BSP By flow cytometric analysis of freshly isolated spleen cells, we found no difference in the percent of cells expressing CD11b, a marker of the monocyte population, between BSPþ/þ and BSP/ mice (8.6% and 8.9%, respectively, data not shown). However, the percent of CD115/c-Fmsþ cells was lower in BSP/ versus BSPþ/þ splenocytes (0.2% versus 0.5%, data not shown). Realtime PCR analysis of RNA isolated from cells undergoing osteoclastogenesis in vitro revealed that lack of BSP altered expression of some but not all of the genes involved in early differentiation and maturation of OCs (Fig. 2). For example, although expression of mRNA for the OC master gene, NFATc1, was not affected, that of RANKL and M-CSF receptors (RANK and c-Fms, respectively) and their downstream effector, c-Fos, was reduced 2 to 5 compared with that in wild-type POCs (Fig. 2A). av but not the b3 integrin subunit also was downregulated in POCs (Fig. 2A). Interestingly, OPN gene expression was upregulated dramatically in BSP/ POC cultures (Fig. 2B), BOUDIFFA ET AL.
Fig. 2. Real-time PCR analysis of gene expression during osteoclastogenesis in BSP/ versus BSPþ/þ mouse spleen cell cultures. The histogram bars in part A show relative mRNA expression in BSP/ POC and OC cultures, expressed as a percentage of BSPþ/þ culture values (100%, gray area in the graph); clear and dark-shaded bars, respectively, indicate overexpressed and underexpressed genes. (B) OPN expression in POCs and OCs as well as (C) MMP9 and MMP2 expression in OCs from BSPþ/þ and BSP/ mice are expressed in absolute values as percent of the internal control, cyclophilin A. All values are mean SEM of 3 wells. p < .05; p < .001; p < .0001 versus matched BSPþ/þ.
whereas that of CD44 was 5 lower compared with wild-type POC cultures (Fig. 2A). In mature OCs, mRNA for calcitonin receptor (CTR), avb3, OPN (Fig. 2B), and the V-ATPase proton pump subunit ATP6V0d2 were reduced in BSP/ OCs, whereas CatK and TRACP were upregulated, as well as the dendritic cell– specific transmembrane protein (DC-STAMP) (Fig. 2A). Strikingly, MMP2 and MMP9 mRNAs were increased dramatically in BSP/ OC cultures (Fig. 2C).
only in an hrVEGF gradient and an mrRANKL gradient, respectively (data not shown). The migration ability of BSP/ POCs (Fig. 3A, B) and OCs (Fig. 3A, C) was reduced to 2 to 3 that of BSPþ/þ cells.
Migration ability is reduced in BSP/ preosteoclasts and mature osteoclasts in vitro
To determine the basis of the lower number of POCs and OCs in BSP/ cultures, we first investigated the ability of wild-type and mutant spleen cells to attach to the substrate because initial cell adhesion is a limiting factor for OC differentiation in vitro.(15) We found no significant differences in spleen cell attachment between genotypes throughout the adhesion time (Fig. 4A). In
We used the Transwell system to investigate the ability of POCs and OCs to migrate in a gradient of hrVEGF-A, mrM-CSF, hrPTHrP, or mrRANKL. We found that wild-type POCs and OCs migrate BSP DEFICIENCY IMPAIRS OSTEOCLASTOGENESIS
BSP/ preosteoclast cultures display reduced proliferation and increased apoptosis, a defect rescued by the addition of exogenous rhBSP
Journal of Bone and Mineral Research
2673
Fig. 3. Migration ability of BSP/ POCs and OCs. Micrographs of Hoechst 33258–stained nuclei (A) and quantification of BSPþ/þ (þ/þ) and BSP/ (/) POCs (B) and OCs (C) remaining in the upper chamber of transwells after 20 hours of migration in a gradient of VEGF-A or RANKL, respectively. Values are mean SEM of 4 transwells. p < .05; p < .001 versus þ/þ.
contrast, we found half the number of proliferating cells, as assessed by Ki67 labeling, in early BSP/ versus BSPþ/þ cultures (Fig. 4B, C); addition of exogenous hrBSP restored BSP/ cell proliferation to BSPþ/þ levels (Fig. 4B, C). Given that it has been suggested that BSP enhances OC survival,(13) we next assessed cell survival in BSP/ POC and OC cultures. The percent of apoptotic cells in BSP/ cultures was 4 higher than in BSPþ/þ cultures at day 3 (Fig. 4D, E), a phenotype rescued by addition of exogenous hrBSP (Fig. 4D, E). In cultures of OCs replated at the same density, OC numbers decreased with time to the same extent in both genotypes (Fig. 4F).
Addition of exogenous RGD-containing proteins rescues osteoclast formation but not preosteoclast numbers or their gene expression profile in BSP/ spleen cell cultures To assess whether addition of RGD-containing proteins restores osteoclastogenesis in BSP/ cultures to BSPþ/þ levels, 5 and 10 ng/mL of hFN, rhBSP, or rhOPN was added to the culture medium throughout the differentiation process. Addition of 5 ng/mL (83 nM) of rhBSP or hFN or 10 ng/mL (>400 nM) of hrOPN to BSP/ cultures did not rescue POCs (monocytic TRACPþ) cell numbers (Fig. 5B) but restored OC numbers (Fig. 5A)
2674
Journal of Bone and Mineral Research
to BSPþ/þ levels. Similarly, addition of rhBSP to BSP/ POCs cultures normalized c-Fms expression but not RANK ; c-Fos, aV integrin chain, and CD44 expression remained decreased, and OPN expression remained dramatically increased (Fig. 5C).
BSP and OPN expression and actin organization in BSPþ/þ and BSP/ osteoclasts Immunolabeling and RT-PCR analysis confirmed the cytoplasmic expression of BSP in OCs from wild-type cells and the perinuclear and membrane labeling of OPN(2) (Fig. 6A). Rhodamine phalloidin labeling of the actin cytoskeleton showed that while neither cell area nor relative podosome area differed between genotypes (Fig. 6B, C), podosomes were more numerous and smaller in BSP/ versus wild-type OCs (Fig. 6B, C).
The resorptive activity of BSP/ osteoclasts is normal on dentin but much reduced on pure hydroxyapatite, a defect rescued by exogenous BSP but only partly by OPN As reported previously,(11) no significant difference in resorptive capacity (the number or the mean area of resorption pits) was observed between BSPþ/þ and BSP/ OCs on dentin (Fig. 7A, B). However, when plated onto pure hydroxyapatite, BSP/ OCs displayed only half the pit number and pit mean area of BSPþ/þ BOUDIFFA ET AL.
Fig. 4. Proliferation, survival, and apoptosis in BSP/ POC cultures. (A) Number of adherent BSPþ/þ (þ/þ) and BSP/ (/) spleen cells 20 and 40 minutes after seeding. (B) Percentage of proliferating þ/þ and / cells, as assessed by KI67 immunostaining (C), and (D) percentage of apoptotic þ/þ and / cells, as assessed by Apopercentage dye staining (E), of spleen cells grown for 3 days with RANKL and M-CSF with or without the addition of hrBSP. ( F) Kinetics of þ/þ and / OC survival over 72 hours of culture following plating at equal density. Values are mean SEM of 4 to 6 wells/ coverslips. p < .05 versus genotype-matched 20 minutes; #p < .001 versus þ/þ ; $p < .05 versus untreated /; §p < .05 versus genotype-matched 24 hours.
OCs (Fig. 7C, D). rhBSP but not rhOPN totally rescued pit number when BSP/ OCs were plated onto either rhBSP- or rhOPNcoated Osteologic disks, whereas rhOPN coating rescued the mean pit area as efficiently as rhBSP (Fig. 7C, D). When OCs were plated onto disks coated with both rhBSP and rhOPN, both pit number and mean pit area were rescued, but no additive effect was observed (Fig. 7C, D).
Discussion Although BSP/ and OPN/ mice display different bone phenotypes,(11,16) the expression of both proteins appears to be required for proper OC formation and/or activity. Previous work from our laboratory showed that under normal conditions in BSP DEFICIENCY IMPAIRS OSTEOCLASTOGENESIS
vivo, the bones of BSP/ mice display fewer OCs, with overall bone parameters indicating a reduced resorption activity.(11) The in vitro results described here confirm that lack of BSP inhibits both POC production and OC formation in spleen and bone marrow cell cultures.(11) Recently, Valverde and coworkers(6) found that OC surfaces and numbers were doubled in the trabecular bone of mice overexpressing BSP and that the number of OCs and their resorptive activity in vitro were increased in transgenic bone marrow cultures. Taken together, these results confirm the important role of BSP in OC differentiation and activity. We now show in addition that although the number of monocyte progenitors in spleen cells is the same in mutant and wild-type mice, the lack of BSP affects proliferation, apoptosis, migration, cytoskeletal structure, and pattern of marker gene expression in cells of the osteoclast Journal of Bone and Mineral Research
2675
Fig. 5. Effect of addition of exogenous BSA, hFN, hrBSP, or hrOPN on osteoclastogenesis in BSPþ/þ (þ/þ) or BSP/ (/) spleen cell cultures. (A) Number of OCs and (B) number of POCs formed in spleen cell cultures supplemented or not with 5 or 10 ng/mL of BSA, hFN, hrBSP, or hrOPN, as indicated. Values are mean SEM of 12 wells. (C) RT-PCR analysis of gene expression in / versus þ/þ POC cultures following treatment with exogenous BSP; mRNA levels are expressed as a percentage of þ/þ culture values (100%, gray area in the graph). Light and dark-shaded bars, respectively, indicate overexpressed and underpexpressed genes. Values are mean SEM of 3 wells. p < .05; p < .001; p < .0001 versus þ/þ.
lineage, all factors bearing on multinucleated OC formation and activity. It has been shown previously that in response to M-CSF, the RANKL receptor (RANK) is upregulated in OC precursor cells, making them sensitive to RANKL.(17) Binding of RANKL leads to TRAF6 activation,(18) which enhances c-Fos and NF-kB expression.(19,20) These molecules activate the expression and autoamplification of the OC master gene NFATc1, which, in turn,
2676
Journal of Bone and Mineral Research
enhances expression of OC marker genes (including TRACP, CatK, CTR, and MMP9).(21,22) On the other hand, c-Fms binding of M-CSF leads to recruitment of c-Src and activation of Rho GTPases, which upregulate c-Fos expression and participate in actin cytoskeleton reorganization.(23) We found that c-Fms and RANK mRNA expression levels are lower, along with their downstream signaling molecule c-Fos, in BSP/ versus BSPþ/þ POC cultures, and flow cytometry confirmed that a lower number of c-Fmspositive cells is present in cells recovered from BSP/ versus BSPþ/þ spleens. We also found decreased expression of aV integrin mRNA, suggesting an impaired activity of the aVb3 complex. The aVb3 signaling pathway is known to interact with the c-Fms signaling pathway, both promoting cell differentiation.(24) While impairment of these pathways could block the progression of BSP/ precursors along the osteoclastogenic lineage, gene expression of NFATc1, which is downstream of cFos, is not affected by the lack of BSP. This suggests activation of a parallel signaling pathway, and indeed, DAP12 outside-in signaling has been shown to interact with the RANK signaling pathway, thus upregulating the intracellular Ca2þ concentration, which enhances the autoamplification of NFATc1,(25) consistent with the fact that BSP/ cultures still form some mature-looking OCs. Once committed, POCs proliferate, migrate, and fuse to generate mature multinucleated OCs. Binding of M-CSF to its receptor, c-Fms, promotes cell proliferation; the proliferation defect of BSP/ spleen cells in culture is consistent with the decreased expression of c-Fms mRNA, and its rescue by the addition of hrBSP may contribute to the increased proliferation in hrBSP-treated BSP/ cultures. aVb3 integrin is implicated in POC migration, as evidenced by the abrogation of mature OC formation when POCs are treated with echistatin, a strong aVb3 agonist.(26) Suzuki and colleagues(27) showed that intracellular OPN and CD44 are also involved in cell migration. The impaired migration of BSP/ POCs thus could reflect the lower expression of CD44 and the aV integrin chain not compensated by enhanced OPN production. M-CSF is also known to promote cell survival. Indeed, it activates the Mitf transcription factor, which then stimulates transcription of Bcl-2, a major antiapoptotic molecule.(28) The elevated apoptosis observed in BSP/ POC cultures may result from the downregulation of c-Fms mRNA expression in these cells, which is rescued, along with cell survival, by exogenous hrBSP. Furthermore, in BSP/ OCs, c-Fms mRNA expression is not altered, which may explain why lack of BSP had no effect on OC survival. We also observed that BSP/ OCs have a lower number of nuclei per cell, suggesting a possible fusion defect, which may result from the impaired migration of POCs, and/or a defect in cell-cell interactions. Indeed, PCR analysis revealed that expression of DC-STAMP and ATP6V0d2 mRNA is differentially regulated in BSP/ cells, depending on the differentiation stage (Fig. 2A). These proteins are required for cell-cell fusion during osteoclastogenesis,(29,30) and their differential regulation is likely to impair this process.(31) Taken together, our results indicate that BSP is an important regulator of osteoclastogenesis because its absence impairs this process at multiple levels, even though multinucleated OCs still form in BSP/ cultures. Interestingly, while some other consequences of the lack of BSP, namely, alterations in BOUDIFFA ET AL.
Fig. 6. Actin cytoskeleton organization in BSP/ OCs. (A) Micrographs of BSPþ/þ (þ/þ) and BSP/ (/) OCs immunostained for BSP and OPN and RTPCR analysis of BSP expression in BSPþ/þ POC (day 3) and OC (day 7) cultures. (B) Micrographs of the actin cytoskeleton organization in þ/þ and / OCs. (C) Quantitative analysis of cell area and podosome area, number, and size in actin-labeled þ/þ and / OCs. Values are mean SEM calculated from 15 micrographs. p < .05 versus matched þ/þ OCs.
proliferation and apoptosis in spleen cell cultures, as well as the decrease in OC number, can be compensated by the addition of exogenous hrBSP or other RGD-containing proteins, suggestive of an outside-in action through avb3, POC number and phenotype are not normalized by these factors. This suggests that either an endogenous autocrine pool of BSP or some specific posttranslational modification(32) or a specific conformation or integration into a multimolecular complex is not restored by exogenous hrBSP and is required for at least some functions of this protein. Further analyses are required to explore these hypotheses. As shown previously,(11) BSP/ OC activity on dentin slices is normal, suggesting that resorption is not affected by the lack of endogenous BSP. Because dentin contains many molecules and factors that could rescue OC activity, including BSP itself, OPN, and other SIBLINGs, we tested the resorptive ability of OCs on Osteologic disks, which contain only calcium hydroxyapatite (HAP). As expected both from in vivo observations and our other results in this study, in vitro resorption by BSP/ OCs was impaired on this substrate. Valverde and colleagues(13) showed that BSP and RANKL synergistically protect mature OCs from apoptosis. However, we found that the rate of death of mature OCs was the same in BSPþ/þ and BSP/ cultures, and thus this cannot explain the poorer resorptive activity on a BSP-free BSP DEFICIENCY IMPAIRS OSTEOCLASTOGENESIS
substrate. Because of its role in resorption, we compared the organization of the actin cytoskeleton in BSP/ and BSPþ/þ OCs seeded onto glass coverslips. Podosomes were higher in number but smaller in size in BSP/ OCs than in wild-type OCs. Pfaff and Jurdic(33) showed earlier that aVb3 integrin is recruited into and participates in the integrity of podosomes. The reduced expression of aVb3 in BSP/ OCs may explain the differences seen in podosome organization and the lower resorptive capacity on a BSP-free substrate. We also show here that BSP/ OCs are hypomotile, which also could affect their activity. Indeed, Nakamura and colleagues(34) showed that aVb3 not only plays a role in maintaining the integrity of the sealing zone but also regulates OC activity by modulating migration. While hrBSP but not hrOPN rescues resorption pit number when BSP/ OCs are plated onto coated hydroxyapatite, both proteins rescue the mean pit area, which suggests a specific role for BSP but also redundancy and cooperation between BSP and OPN proteins in OC activity. In summary, we have shown that BSP expression is involved in both OC formation and activity. Addition of exogenous BSP only partially rescues BSP/ OC differentiation, suggesting that endogenous autocrine BSP or some local context for interprotein interactions is required for at least some functions of BSP. Journal of Bone and Mineral Research
2677
Fig. 7. Resorption activity of BSP/ OCs. (A) Number and area and (B) micrographs of resorption pits observed on dentin slices plated for 72 hours with BSPþ/þ (þ/þ) and BSP/ (/) OCs; values are mean SEM of 6 dentin slices. (C) Number and mean area and (D) micrographs of resorption pits observed on Osteologic disks plated with þ/þ and / OCs and coated or not with 100 ng/mL of either hrBSP, hrOPN, or both (B þ O). Values are mean SEM of 4 disks. p < .001 versus þ/ þ .
2678
Journal of Bone and Mineral Research
BOUDIFFA ET AL.
Disclosures The authors state that they have no conflicts of interests.
Acknowledgments INSERM U890/IFR143 and UMR 5161 CNRS/ENS are part of the Re´seau d’Etude des Tissus Mine´ralise´s (RETIM) network. We thank Dr N Takahashi (Matsumoto Dental University, Nagagano, Japan) for the generous gift of dentin slices; Dr C Faure, A Perrier, and C Domenget for advice and protocols; N Laroche, S Peyroche, and Dr M-T Linossier for their excellent technical support; and Dr D Denhardt for his advice on the manuscript. This work was funded by the Canadian Institutes of Health Research (FRN83704 to JEA) and by the French Institut National de la Sante´ et de la Recherche Me´dicale (INSERM). Dr NM Wade-Gueye gratefully acknowledges scholarship funding from the Socie´te´ Franc¸aise de Rhumatologie (SFR).
References 1. Rodan GA, Martin TJ. Therapeutic approaches to bone diseases. Science. 2000;289:1508–1514. 2. Sodek J, Ganss B, McKee MD. Osteopontin. Crit Rev Oral Biol Med. 2000;11:279–303.
15. Miyamoto T, Arai F, Ohneda O, Takagi K, Anderson DM, Suda T. An adherent condition is required for formation of multinuclear osteoclasts in the presence of macrophage colony-stimulating factor and receptor activator of nuclear factor kappa B ligand. Blood. 2000;96: 4335–4343. 16. Rittling SR, Matsumoto HN, McKee MD, et al. Mice lacking osteopontin show normal development and bone structure but display altered osteoclast formation in vitro. J Bone Miner Res. 1998;13:1101–1111. 17. Arai F, Miyamoto T, Ohneda O, et al. Commitment and differentiation of osteoclast precursor cells by the sequential expression of c-Fms and receptor activator of nuclear factor kappaB (RANK) receptors. J Exp Med. 1999;190:1741–1754. 18. Lomaga MA, Yeh WC, Sarosi I, et al. TRAF6 deficiency results in osteopetrosis and defective interleukin-1, CD40, and LPS signaling. Genes Dev. 1999;13:1015–1024. 19. Grigoriadis AE, Wang ZQ, Cecchini MG, et al. c-Fos: a key regulator of osteoclast-macrophage lineage determination and bone remodeling. Science. 1994;266:443–448. 20. Franzoso G, Carlson L, Xing L, et al. Requirement for NF-kappaB in osteoclast and B-cell development. Genes Dev. 1997;11:3482–3496. 21. Kim K, Kim JH, Lee J, et al. Nuclear factor of activated T cells c1 induces osteoclast-associated receptor gene expression during tumor necrosis factor-related activation-induced cytokine-mediated osteoclastogenesis. J Biol Chem. 2005;280:35209–35216. 22. Asagiri M, Sato K, Usami T, et al. Autoamplification of NFATc1 expression determines its essential role in bone homeostasis. J Exp Med. 2005;202:1261–1269.
3. Ganss B, Kim RH, Sodek J. Bone sialoprotein. Crit Rev Oral Biol Med. 1999;10:79–98.
23. Sakai H, Chen Y, Itokawa T, Yu KP, Zhu ML, Insogna K. Activated c-Fms recruits Vav and Rac during CSF-1-induced cytoskeletal remodeling and spreading in osteoclasts. Bone. 2006;39:1290–1301.
4. Bianco P, Fisher LW, Young MF, Termine JD, Robey PG. Expression of bone sialoprotein (BSP) in developing human tissues. Calcif Tissue Int. 1991;49:421–426.
24. Faccio R, Takeshita S, Zallone A, Ross FP, Teitelbaum SL. c-Fms and the alphavbeta3 integrin collaborate during osteoclast differentiation. J Clin Invest. 2003;111:749–758.
5. Arai N, Ohya K, Kasugai S, et al. Expression of bone sialoprotein mRNA during bone formation and resorption induced by colchicine in rat tibial bone marrow cavity. J Bone Miner Res. 1995;10:1209–1217.
25. Asagiri M, Takayanagi H. The molecular understanding of osteoclast differentiation. Bone. 2007;40:251–264.
6. Valverde P, Zhang J, Fix A, et al. Overexpression of bone sialoprotein leads to an uncoupling of bone formation and bone resorption in mice. J Bone Miner Res. 2008;23:1775–1788. 7. Tu Q, Zhang J, Fix A, et al. Targeted overexpression of BSP in osteoclasts promotes bone metastasis of breast cancer cells. J Cell Physiol. 2009;218:135–145. 8. Fisher LW, Fedarko NS. Six genes expressed in bones and teeth encode the current members of the SIBLING family of proteins. Connect Tissue Res. 2003;44(Suppl 1):33–40. 9. Ross FP, Chappel J, Alvarez JI, et al. Interactions between the bone matrix proteins osteopontin and bone sialoprotein and the osteoclast integrin alpha v beta 3 potentiate bone resorption. J Biol Chem. 1993;268:9901–9907. 10. Chellaiah MA, Kizer N, Biswas R, et al. Osteopontin deficiency produces osteoclast dysfunction due to reduced CD44 surface expression. Mol Biol Cell. 2003;14:173–189. 11. Malaval L, Wade-Gueye NM, Boudiffa M, et al. Bone sialoprotein plays a functional role in bone formation and osteoclastogenesis. J Exp Med. 2008;205:1145–1153.
26. Nakamura I, Tanaka H, Rodan GA, Duong LT. Echistatin inhibits the migration of murine prefusion osteoclasts and the formation of multinucleated osteoclast-like cells. Endocrinology. 1998;139:5182–5193. 27. Suzuki K, Zhu B, Rittling SR, et al. Colocalization of intracellular osteopontin with CD44 is associated with migration, cell fusion, and resorption in osteoclasts. J Bone Miner Res. 2002;17:1486–1497. 28. Hershey CL, Fisher DE. Mitf and Tfe3: members of a b-HLH-ZIP transcription factor family essential for osteoclast development and function. Bone. 2004;34:689–696. 29. Miyamoto T. The dendritic cell-specific transmembrane protein DCSTAMP is essential for osteoclast fusion and osteoclast bone-resorbing activity. Mod Rheumatol. 2006;16:341–342. 30. Lee SH, Rho J, Jeong D, et al. v-ATPase V0 subunit d2-deficient mice exhibit impaired osteoclast fusion and increased bone formation. Nat Med. 2006;12:1403–1409. 31. Kim K, Lee SH, Ha Kim J, Choi Y, Kim N. NFATc1 induces osteoclast fusion via up-regulation of Atp6v0d2 and the dendritic cell-specific transmembrane protein (DC-STAMP). Mol Endocrinol. 2008;22:176– 185.
12. Raynal C, Delmas PD, Chenu C. Bone sialoprotein stimulates in vitro bone resorption. Endocrinology. 1996;137:2347–2354.
32. Qin C, Baba O, Butler WT. Post-translational modifications of sibling proteins and their roles in osteogenesis and dentinogenesis. Crit Rev Oral Biol Med. 2004;15:126–136.
13. Valverde P, Tu Q, Chen J. BSP and RANKL induce osteoclastogenesis and bone resorption synergistically. J Bone Miner Res. 2005;20:1669– 1679.
33. Pfaff M, Jurdic P. Podosomes in osteoclast-like cells: structural analysis and cooperative roles of paxillin, proline-rich tyrosine kinase 2 (Pyk2) and integrin alphaVbeta3. J Cell Sci. 2001;114(Pt 15):2775–2786.
14. Nagy A, Rossant J, Nagy R, Abramow-Newerly W, Roder JC. Derivation of completely cell culture-derived mice from early-passage embryonic stem cells. Proc Natl Acad Sci U S A. 1993;90:8424–8428.
34. Nakamura I, Pilkington MF, Lakkakorpi PT, et al. Role of alpha(v)beta(3) integrin in osteoclast migration and formation of the sealing zone. J Cell Sci. 1999;112(Pt 22):3985–3993.
BSP DEFICIENCY IMPAIRS OSTEOCLASTOGENESIS
Journal of Bone and Mineral Research
2679
