Bone 71 (2015) 217–226

Contents lists available at ScienceDirect

Bone journal homepage: www.elsevier.com/locate/bone

Original Full Length Article

Siglec-15 is a potential therapeutic target for postmenopausal osteoporosis Yusuke Kameda a, Masahiko Takahata a,⁎, Shintaro Mikuni b, Tomohiro Shimizu a, Hiroki Hamano a, Takashi Angata d, Shigetsugu Hatakeyama c, Masataka Kinjo b, Norimasa Iwasaki a a

Hokkaido University, Department of Orthopedic Surgery, School of Medicine, Sapporo, Japan Hokkaido University, Laboratory of Molecular Cell Dynamics, Faculty of Advanced Life Science, Sapporo, Japan Hokkaido University, Department of Biochemistry, School of Medicine, Sapporo, Japan d Institute of Biological Chemistry, Academia Sinica, 128 Academia Road, Section 2, Nankang, Taipei 11529, Taiwan b c

a r t i c l e

i n f o

Article history: Received 11 April 2014 Revised 16 October 2014 Accepted 23 October 2014 Available online 8 November 2014 Edited by: Hong-Hee Kim Keywords: Siglec-15 Osteoclast DAP12 Menopause Osteoporosis Sialic acid

a b s t r a c t Sialic acid-binding immunoglobulin-like lectin 15 (Siglec-15) is an immunoreceptor that regulates osteoclast development and bone resorption in association with an immunoreceptor tyrosine-based activation motif (ITAM) adaptor protein, DNAX-activating protein 12 kDa (DAP12). Although Siglec-15 has an important role in physiologic bone remodeling by modulating RANKL signaling, it is unclear whether it is involved in pathologic bone loss in which multiple osteoclastogenic factors participate in excessive osteoclastogenesis. Here we demonstrated that Siglec-15 is involved in estrogen deficiency-induced bone loss. WT and Siglec-15−/− mice were ovariectomized (Ovx) or sham-operated at 14 wk of age and their skeletal phenotype was evaluated at 18 and 22 wk of age. Siglec-15−/− mice showed resistance to estrogen deficiency-induced bone loss compared to WT mice. Although the number of tartrate-resistant acid phosphatase (TRAP)-positive osteoclasts increased after ovariectomy in both WT and Siglec-15−/− mice, the increase was lower in Siglec-15−/− mice than in WT mice. Importantly, osteoclasts in Siglec-15−/− mice were small and failed to spread on the bone surface, indicating impaired osteoclast differentiation. Because upregulated production of TNF-α as well as RANKL is mainly responsible for estrogen deficiency-induced development of osteoclasts, we examined whether Siglec-15 deficiency affects TNF-α-induced osteoclastogenesis in vitro. The TNF-α mediated induction of TRAP-positive multinucleated cells was impaired in Siglec-15−/− cells, suggesting that Siglec-15 is involved in TNF-α induced osteoclastogenesis. We also confirmed that signaling through osteoclast-associated receptor/Fc receptor common γ chain, which is an alternative ITAM adaptor to DAP12, rescues multinucleation but not cytoskeletal organization of TNF-α and RANKL-induced Siglec-15−/− osteoclasts, indicating that the Siglec-15/DAP12 pathway is especially important for cytoskeletal organization of osteoclasts in both RANKL and TNF-α induced osteoclastogenesis. The present findings indicate that Siglec-15 is involved in estrogen deficiency-induced differentiation of osteoclasts and is thus a potential therapeutic target for postmenopausal osteoporosis. © 2014 Elsevier Inc. All rights reserved.

Introduction Osteoclastogenesis requires immunoreceptor tyrosine-based activation motif (ITAM) signaling [1–4] based on findings that mice doubly deficient in DNAX-activating protein (DAP) 12 and Fc receptor common γ chain (FcRγ), which are the primary ITAM harboring adaptors in osteoclast lineage cells, exhibit severe osteopetrosis due to the impaired osteoclast development. Because both DAP12 and FcRγ have minimal extracellular domains, making then incapable of sensing signals outside ⁎ Corresponding author at: Hokkaido University, Department of Orthopedic Surgery, Graduate School of Medicine, Kita-15 Nishi-7, Kita-ku, Sapporo 060-8638, Japan. Fax: +81 11 706 6054. E-mail address: [email protected] (M. Takahata).

http://dx.doi.org/10.1016/j.bone.2014.10.027 8756-3282/© 2014 Elsevier Inc. All rights reserved.

the cells, immunoreceptors associated with DAP12 or FcRγ are thought to regulate ITAM signaling. Sialic acid binding Ig-like lectin (Siglec)-15 is a DAP12-associated immunoreceptor (DAR) that recognizes sialylated glycans and regulates osteoclast differentiation [5]. Although several other DARs, including triggering receptor expressed in myeloid cells 2 (TREM2) [6], signalregulatory protein β1 [7], and myeloid DAP12-associating lectin-1, have been identified in osteoclast lineage cells [8], Siglec-15 is likely a major immunoreceptor that regulates bone resorption based on the findings that Siglec-15 deficient mice have a mild but apparent osteopetrotic phenotype resulting from the impaired osteoclast development, while the osteopetrotic phenotype has not been confirmed in mice lacking other DARs [9,10]. We previously demonstrated that Siglec-15 is involved in physiologic bone remodeling by modulating

218

Y. Kameda et al. / Bone 71 (2015) 217–226

RANKL signaling [11]. Furthermore, the most recent study performed by Stuible et al. [12] demonstrated that treatment of young mice with Siglec-15 antibodies led to a significant increase in bone volume. However, it remains unclear whether Siglec-15 is involved in pathologic conditions in which cytokines other than RANKL participate in excessive osteoclastogenesis. Postmenopausal osteoporosis is a major metabolic bone disease associated with rapid bone loss and an increased fragility fracture risk. Because estrogen deficiency leads to increased bone remodeling in which resorption outstrips formation, osteoclasts are a principal target for the treatment of estrogen deficiency-induced bone loss. The molecular mechanisms underlying the upregulating effects of estrogen deficiency on osteoclastic bone resorption are complex, estrogen deficiency-induced bone loss is mainly due to upregulated production of TNF-α by activated T cells and the increased secretion of RANKL from osteoblasts/stromal cells [13–15]. TNF-α directly induces osteoclastogenesis independent of, and strongly synergistic with, RANKL, but it is not known if Siglec-15 has a pivotal role in postmenopausal osteoporosis [16,17]. In the present study, we examined whether Siglec-15 is involved in estrogen deficiency-induced bone loss using the ovariectomy model in mice lacking Siglec-15. In addition, we investigated whether Siglec-15 is involved in TNF-α induced osteoclastogenesis in vitro.

Materials and methods Mice The Ethics Review Committee for Animal Experimentation of Hokkaido University approved the experimental protocol. Mice were maintained under specific pathogen free conditions. Both WT and Siglec-15−/− had a C57BL/6 background. Female mice were ovariectomized (Ovx) or sham-operated (Sham) at 14 wk. of age. At 4 wk after surgery, some mice were killed and their tibias excised for histology. At 8 wk after surgery, the remaining mice were killed and both tibias and lumbar vertebrae were excised for micro-CT analysis.

Antibodies and reagents Polyclonal antibodies specific to mouse Siglec-15 were developed by one of the authors (TA) and provided by the National Institute of Advanced Industrial Science and Technology (Tsukuba, Japan) [18]. Antibodies for phospho-Erk, Erk, phospho-Akt, Akt, phospho-PI3K p85, PI3K p85, phospho-JNK, JNK, phospho-p38, p38, phospho-IκBα, and IκBα were purchased from Cell Signaling Technology (Tokyo, Japan). Recombinant human M-CSF and soluble RANKL were purchased from PeproTech EC, Ltd. (London, UK). Chicken collagen II was purchased from Sigma-Aldrich (Tokyo, Japan). Bovine bone slices were a kind gift from Dr. Toshiyuki Akazawa (Hokkaido Research Organization, Industrial Research Institute, Sapporo, Japan).

Histology and histomorphometry Proximal tibiae of either Ovx or sham mice were fixed in paraformaldehyde, decalcified in EDTA, and embedded in paraffin. Sections were stained for tartrate-resistant acid phosphatase (TRAP) with methyl green counterstain to observe osteoclasts. For histomorphometry analysis, tibiae of either Ovx or sham mice 2-wk after operation were harvested, dehydrated in ethanol, and undecalcified sections were stained in Villanueva Bone Stain. Bone section preparation was performed at Niigata Bone Science Institute (Japan). The numbers of osteoclasts/bone surface (N.Oc/B.Pm) and osteoclast surface/bone surface (Oc.Pm/B.Pm) at the secondary spongiosa were measured according to Dempster et al. [20]. Primary spongiosa was defined as the area 250 μm distal to the growth plate and secondary spongiosa was defined as the area from 250 μm to 1000 μm distal to the growth plate.

In vitro osteoclastogenesis Osteoclast differentiation from bone marrow cells was achieved as previously described [21]. Briefly, the marrow cells were collected from femurs and tibiae of mice. After removing the red blood cells, marrow cells were cultured on a suspension culture dish in 50 ng/ml M-CSF to enrich the CD11b + (Mac1 +) population [22]. After a 3-d culture, the cells were washed twice with PBS to remove nonadherent cells, and bone marrow macrophages (BMMs) were harvested. BMMs were resuspended and further cultured with 30 ng/ml M-CSF and 100 ng/ml TNF-α and/or 100 ng/ml RANKL with or without 500 ng/ml osteoprotegerin (OPG) for 5 days at 37 °C in a 5% humidified CO2 incubator to generate osteoclasts. M-CSF, TNF-α, RANKL, and OPG were used at these concentrations and the medium was changed every other day throughout the study unless otherwise described. In some experiments, BMMs were cultured on collagen II-coated plastic culture wells or on bovine bone slices. Co-culture of BMMs and primary osteoblasts derived from calvarial cells was performed in the presence of 50 μg/ml ascorbic acid, 10 nM 1,25-dihydroxyvitamin D3, and 1 μM dexamethasone. Mouse primary unfractionated bone marrow cells (UBMC) were flashed from mouse femurs and tibiae 2 wk after ovariectomy or sham operation, and cultured in the presence of 50 μg/ml ascorbic acid and 20 nM 1,25-dihydroxyvitamin D3 to generate osteoclasts.

In vitro resorption assay of cultured osteoclasts Cells were cultured in the presence of M-CSF and TNF-α and/or RANKL on bovine bone slices for 10 days. Cells were washed and then stained with 20 μg/ml peroxidase-conjugated wheat germ agglutinin (Sigma), followed by incubation with 3,3′-diaminobenzidine (0.52 mg/ml in PBS containing 0.1% H2O2). The area of resorption pits was then measured using ImageJ image analysis software (NIH).

Micro-CT analysis Left tibiae and the 5th lumbar vertebral bodies were scanned individually by micro-CT (R_mCT2; Rigaku, Tokyo, Japan) at a 10-μm isotropic resolution. Both were measured using a TRI/3D-BON (Ratoc System Engineering Co., Tokyo, Japan) in accordance with the guidelines described in Bouxsein et al. [19] For tibiae, a 1000-μm area of interest of 100 slices encompassing the region of the proximal metaphysis, starting from 300 μm distal to the growth plate, was used to assess trabecular bone morphology. For lumbar vertebral bodies, an area from the upper growth plate to the lower growth plate was used to assess trabecular bone morphology.

Real-time quantitative PCR (qPCR) of gene expression Total RNA from cultured cells was isolated using a Qiagen RNeasy Mini Kit (Qiagen, Valencia, CA). cDNA was synthesized from 1 μg total RNA using reverse transcriptase and oligo-dT primers for qPCR analyses. qPCR was performed as previously described [23]. cDNA samples were analyzed for both the genes of interest and the reference gene (GAPDH). Primer sequences used for the experiment were described previously [11]. The amount of mRNA expressed was normalized to the GAPDH expression.

Y. Kameda et al. / Bone 71 (2015) 217–226

Immunocytochemistry Cells were cultured on plastic wells or on bovine bone slices, and fixed for 5 min with 4% paraformaldehyde and then treated with 0.1% Triton X-100. After blocking with 5% bovine serum albumin, the cells were treated with polyclonal anti-Siglec-15 antibody followed by staining with Alexa Fluor-labeled secondary antibody (Invitrogen Molecular Probes, Carlsbad, CA). The cytoskeletal actin was stained using Alexa Fluor 633 phalloidin (Invitrogen Molecular Probes). The nuclei were visualized using 4′,6-diamidino-2-phenylindole reagent (Dojindo Laboratories, Japan). Retroviral induction Retroviral vectors pMX Siglec-15 Myc-His and pMX Siglec-15 mutants, R143A, and K273A, were prepared as previously described [11]. The resulting vectors were used to transfect Plat E cells using Lipofectamine LTX plus reagent (Invitrogen), and then recombinant retroviruses were generated. The culture supernatants were harvested

219

48 h after transfection and used for infection. BMMs were infected with virus for 6 to 12 h in the presence of M-CSF (30 ng/ml) and polybrene (4 μg/ml). Tartrate-resistant acid phosphatase staining Osteoclast generation was confirmed by TRAP staining. After aspirating the medium, cells were fixed with 4% formaldehyde containing acetone and citrate solution at room temperature for 1 min and stained for TRAP using a histochemical kit (Sigma-Aldrich) according to the manufacturer's instructions. Multinucleated osteoclasts were identified microscopically as TRAP-positive cells with at least three nuclei, and the number of cells in each well was quantified. Immunoblot analysis BMMs cultured for the indicated time in the presence of M-CSF and RANKL were prepared. The medium was removed and the cells were

Fig. 1. Siglec-15−/− mice are protected from estrogen deficiency-induced bone loss. A, C. Micro-CT of proximal tibiae (A) and the 5th lumbar vertebral bodies (C) of WT and Siglec-15−/− mice at 8-wk after ovariectomy or sham operation. B, D. Bone volume and microstructural indices of trabecular bone in the metaphyseal region of the tibiae and the vertebral bodies of WT and Siglec-15−/− mice. BV/TV, bone volume/total volume; Ct, cortical thickness; Ovx, ovariectomized; Tb.N, trabecular number; Tb.Th, trabecular thickness.

220

Y. Kameda et al. / Bone 71 (2015) 217–226

washed twice with ice-cold PBS. Cell lysates were then extracted using the PhosphoSafe Extraction Reagent (Novagen, Madison, WI). Cell lysates were subjected to immunoblot analyses using the indicated antibodies.

In vitro bone nodule formation assay UBMC were flashed from mouse femurs and tibiae 2 wk after ovariectomy or sham operation, seeded in 48-well plates in α-MEM with 10%

Fig. 2. Impaired osteoclast development and bone resorption following ovariectomy in Siglec-15−/− mice. A. Micrographs of the proximal tibiae of WT and Siglec-15−/− mice 4-wk after ovariectomy or sham operation stained with TRAP and methyl green. Upper panels show low magnification micrographs. Middle and lower panels show high magnification micrographs at the primary and secondary spongiosa, respectively. Although the number of TRAP-positive osteoclasts increased after ovariectomy in both WT and Siglec-15−/− mice, the increase was significantly smaller in Siglec-15−/− mice than in WT mice. Importantly, osteoclasts in ovariectomized Siglec-15−/− mice were small and failed to spread on bone surface, indicating impaired cytoskeletal organization of osteoclasts. Scale bar: 50 μm. B. Histomorphometric parameters were related to osteoclastic bone resorption at the secondary spongiosa of the tibiae. The number of osteoclasts was increased after ovariectomy in both WT and Siglec-15−/− mice, whereas the osteoclast contacting surface was increased in WT mice, but not in Siglec-15−/− mice. C. Mouse primary unfractionated bone marrow cells (UBMC) were flashed from mouse femurs and tibiae 2 wk after ovariectomy or sham operation, and cultured in the presence of 50 μg/ml ascorbic acid and 20 nM 1,25-dihydroxyvitamin D3. Scale bar: 50 μm. N.Oc/B.Pm, the numbers of osteoclasts/bone surface; N.Mo.Oc/B.Pm, the numbers of mononuclear osteoclasts/bone surface; N.Mu.Oc/B.Pm, the numbers of multinuclear osteoclasts/bone surface; Oc.Pm/B.Pm, osteoclast surface/bone surface; Ovx, ovariectomized.

Y. Kameda et al. / Bone 71 (2015) 217–226

FBS containing 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate at a density of 106/well per 1 ml. After 14 days, bone nodules were stained using a histochemical kit (Cosmo-Bio, Tokyo, Japan) according to the manufacturer's instructions. Statistical analysis Data of two-group comparisons were analyzed using a two-tailed Student's t test. Simultaneous comparisons of more than two groups were performed using ANOVA. A p value of less than 0.05 was considered statistically significant. The data are represented as mean ± SEM. Results Siglec-15−/− mice are protected from estrogen deficiency-induced bone loss Micro-CT evaluation of the proximal tibiae and the 5th lumber vertebral bodies was performed at 8 wk after ovariectomy or sham operation. Although trabecular bone loss was observed in both WT

221

and Siglec-15−/− mice, the percent differences in BV/TV and Tb.N between the sham and Ovx groups were smaller in Siglec-15−/− mice than in WT mice (Figs. 1A–D). The relative bone loss in proximal tibiae was 50.7% in WT and 20.2% in Siglec-15−/−, respectively, and that in the 5th lumber vertebral bodies was 27.4% in WT and 12.1% in Siglec-15−/− mice, respectively. Histologic analysis on paraffin embedded sections of proximal tibiae retrieved 4 wk after ovariectomy or sham operation was then performed to investigate the mechanism of resistance to ovariectomy in Siglec-15−/− mice in terms of bone resorption (Fig. 2A). As expected, the number of osteoclasts, including mononuclear and multinuclear TRAP-positive cells, was increased after ovariectomy in both WT and Siglec-15−/− mice, while the number of multinucleated osteoclasts was lower in Siglec-15−/− mice than in WT mice (Fig. 2B). Importantly, Siglec-15−/− osteoclasts were small and did not spread on the bone surface in the secondary spongiosa. These findings indicate that Siglec-15 blockade attenuates the development of functional osteoclasts induced by acute estrogen deficiency. To compare osteoclast differentiation capacities in ovariectomized mouse bone microenvironments, we isolated UBMC directly from Ovx

Fig. 3. Siglec-15 is required for the TNF-α induced development of functional osteoclasts. A. Differentiation efficiency of bone marrow macrophages (BMMs) into TRAP-positive multinuclear cells (MNCs) is suppressed in Siglec-15−/− cells, compared to WT cells (*p b 0.05). BMMs were cultured in the presence of M-CSF and TNF-α with or without RANKL for 5-d, after which the cells were stained for TRAP. B. BMMs were cultured in the presence of M-CSF and TNF-α with or without RANKL on a bovine bone slice for 10 days. The percentage of resorption of the substrate was significantly lower in Siglec-15−/− cells, compared to WT cells (*p b 0.05). C. Expression of osteoclast markers genes was examined by quantitative RT-PCR. BMMs cultured with M-CSF and TNF-α for 3 days and 5 days were used as preosteoclasts (PreOC) and osteoclasts (OC), respectively.

222

Y. Kameda et al. / Bone 71 (2015) 217–226

or sham-operated WT and Siglec-15−/− mice 2 wk after surgery. To generate osteoclasts, UBMC were cultured in the presence of 1,25dihydroxyvitamin D3. As expected, a significantly increased number of osteoclasts developed in ex vivo cultured UBMC from WT Ovx mice compared to WT sham mice. In contrast, Siglec-15−/− UBMC failed to form multinucleated osteoclasts and increased only slightly after ovariectomy (Fig. 2C). Siglec-15 is required for the TNF-α induced development of functional osteoclasts To investigate whether Siglec-15 is involved in acute bone loss after estrogen deficiency, we focused on TNF-α induced osteoclastogenesis based on reports indicating the importance of TNF-α in postmenopausal osteoporosis [13,24]. We compared the ability to form osteoclasts and resorb bone between Siglec-15−/− BMMs and WT BMMs in the presence of TNF-α. Although TRAP-positive osteoclasts were formed from Siglec15−/− cells, most of them were mononuclear and the number of TRAPpositive multinucleated cells was significantly lower in Siglec-15−/− cells than in WT cells. Culturing the cells in the presence of RANKL and TNF-α did not change this finding (Fig. 3A). To exclude the possibility that a miniscule amount of RANKL was produced by mesenchymal cells contaminating the BMMs, we added OPG to our culture system. Although WT cells differentiated into multinucleated osteoclasts, Siglec-15−/− cells failed to form well-spread osteoclasts, indicating the importance of Siglec-15 in TNF-α induced osteoclastogenesis (Sup. Fig. 1A). Although Siglec-15−/− cells formed some multinucleated osteoclasts, these were morphologically abnormal; they were small and did not spread on the culture dish. Consequently, the resorption pit area of the bovine bone slices after a 10-d culture with M-CSF and TNF-α was 4.0% ± 1.2% in Siglec-15−/− cells, whereas it was 29.9% ± 4.2% in WT cells (Fig. 3B). The combination of RANKL and TNF-α enhanced bone resorption in both cell types, but the resorption pit area was still significantly smaller in Siglec-15−/− cells (9.2% ± 2.5%) than in WT cells (52.4% ± 4.1%). Induction of osteoclast marker genes, including

TRAP, cathepsin K, calcitonin receptor, and integrin β3, on day 5 after TNF-α stimulation was inhibited in Siglec-15−/− BMMs compared to WT BMMs, indicating that Siglec-15 was required for the TNF-α induced development of functional osteoclasts in vitro (Fig. 3C). Osteoclasts cultured on plastic wells were morphologically different from those cultured on bone slices. Therefore, to further investigate the morphologic abnormalities of Siglec-15−/− osteoclasts, we cultured BMMs on bovine bone slices with M-CSF and TNF-α. We visualized the actin cytoskeleton by staining the cells with phalloidin. As expected, Siglec-15−/− osteoclasts were mostly mononuclear and could not form actin rings (Fig. 4), indicating that Siglec-15 was involved in the TNFα induced differentiation of functional osteoclasts. Osteoclast-associated receptor/FcRγ signaling rescues multinucleation, but not cytoskeletal organization of TNF-α induced Siglec-15−/− osteoclasts Another ITAM-bearing adaptor, FcRγ, compensates for DAP12 function in osteoclastogenesis [2]. Therefore, we tested whether FcRγ signaling rescues impaired osteoclast development in Siglec-15−/− cells. The precise mechanism underlying the activation of FcRγ signals remains unclear, but a recent study demonstrated that collagen binds to osteoclast-associated receptor (OSCAR) and stimulates osteoclastogenesis through FcRγ signaling [25]. Therefore, Siglec-15−/− BMMs were cultured on type II collagen-coated culture wells for TRAP staining and on bone slices for cytoskeletal evaluation. Type II collagen dramatically rescued TNF-α induced multinucleated osteoclast formation in Siglec-15−/− BMMs (Fig. 5A). Siglec-15−/− osteoclasts, however, did not form a ring-shaped boundary, but did form lamellipodia. We then analyzed the actin cytoskeleton with Alexa Fluor 633-conjugated phalloidin. As expected, Siglec-15−/− osteoclasts induced by TNF-α failed to form actin rings when cultured on type II collagen-coated bovine bone slices (Fig. 5B). As previously reported [11], however, the addition of RANKL restored cytoskeletal deficiency of Siglec-15−/− osteoclasts when cultured in the presence of type II collagen and bone matrix. This finding indicates that Siglec-15 is essential for TNF-α induced development of functional osteoclasts, and neither

Fig. 4. Impaired cytoskeletal organization of Siglec-15−/− osteoclasts in vitro. Osteoclasts generated on bovine bone slices were stained with anti-Siglec-15 antibody, Alexa Fluor 633 phalloidin (red), and DAPI. WT osteoclasts form an actin ring, while those lacking Siglec-15 fail to organize actin rings. The combination of RANKL and TNF-α failed to rescue the cytoskeletal organization of Siglec-15−/− osteoclasts. Scale bar: 100 μm.

Y. Kameda et al. / Bone 71 (2015) 217–226

223

Fig. 5. OSCAR/FcRγ signaling rescues multinucleation, but not cytoskeletal organization of TNF-α induced Siglec-15−/− osteoclasts. A. TRAP-positive multinuclear cells (MNCs) were formed from Siglec-15−/− cells when cultured on type II collagen-coated culture wells; however, these failed to form a ring-shaped boundary. B. Cytoskeletal organization defects were not rescued by FcRγ signaling. Osteoclasts cultured on type II collagen-coated bovine bone slices were immunostained with anti-Siglec-15 antibody, phalloidin, and DAPI. Actin ring formation was rescued in Siglec-15−/− osteoclasts in the presence of RANKL, but not TNF-α. Scale bar: 100 μm.

OSCAR/FcRγ signaling nor the signaling activated by some factors contained in bone matrix rescues the defective cytoskeletal organization of Siglec-15−/− cells. Siglec-15, in association with DAP12, regulates TNF-α induced development of functional osteoclasts Siglec-15 is a type-I transmembrane protein with two immunoglobulin-like domains and a transmembrane domain containing a lysine residue (K273) that is necessary for the association with DAP12 [18]. We previously demonstrated that Siglec-15 signaling requires DAP12 in RANKL-induced functional osteoclast development [11]. We therefore examined whether Siglec-15 also requires association with DAP12 in TNF-α induced osteoclastogenesis. To address this question, we retrovirally transduced a point mutant of Siglec-15

(K273A) to attenuate DAP12-binding ability in Siglec-15−/− cells. Retroviral transduction of WT Siglec-15 rescued the well-spread multinucleated osteoclast formation, which was not observed in the K273A mutant (Figs. 6A, C), indicating that Siglec-15 in association with DAP12 regulates TNF-α induced development of functional osteoclasts. To confirm whether ligand binding of Siglec-15 is needed to activate the downstream signaling of Siglec-15 required for the development of functional osteoclasts, we also transduced a point mutant of Siglec-15 (R143A), in which glycan-binding ability is lost. Retroviral transduction of Siglec-15 (R143A) could not rescue the impaired multinuclear osteoclast formation of Siglec-15−/− cells (Figs. 6A, C). We also performed osteoclastogenesis in the presence of both TNF-α and RANKL to confirm whether the combination of these two cytokines rescues osteoclast formation of Siglec-15−/− cells transduced with the R143A or K273A mutants. As expected, Siglec-15−/− cells could not

224

Y. Kameda et al. / Bone 71 (2015) 217–226

Fig. 6. Ligand binding and association with DAP12 is essential for Siglec-15 function in TNF-α induced osteoclastogenesis. A. Retroviral transduction of WT-Siglec-15 rescued the impaired development of osteoclasts in Siglec-15−/− osteoclasts, whereas R143A or K273 mutants did not (*p b 0.05 vs. control group). B. The combination of RANKL and TNF-α failed to rescue osteoclastogenesis of Siglec-15−/− cells (*p b 0.05 vs. control group). pMX-Puro, pMX-Puro plasmid control. Scale bar: 200 μm. C. Immunoprecipitates pulled down with anti-myc antibody were immunoblotted with anti-DAP12 or anti-myc. WT Siglec-15 and an R143A mutant lacking ligand binding ability associated with DAP12, while mutation of the conserved lysine residue of Siglec-15 to alanine (K273A mutant) abrogated the association between Siglec-15 and DAP12.

form well-spread osteoclasts, indicating that the association with DAP12 and glycan-binding ability were crucial for Siglec-15 function and cannot be compensated for, even in the presence of excessive osteoclastogenic cytokines (Figs. 6B, C). TNF-α-induced intracellular signaling is not altered in Siglec-15 deficient cells To clarify the role Siglec-15 plays in TNF-α induced osteoclast differentiation, we examined downstream signaling in Siglec-15−/− preosteoclasts. As Siglec-15 mediates RANKL-induced phosphorylation of the Erk and PI3K/Akt pathway [11], which is also activated by TNF-α, we confirmed whether signal transduction was attenuated in Siglec-15−/− cells. Unexpectedly, however, the known TNF-α mediated signals were not attenuated by Siglec-15 deficiency. We examined the phosphorylation of Erk, Akt, PI3K, Iκβα, JNK, and p38, but found no difference in any of them between the WT and Siglec-15−/− cells (Fig. 7). Osteoblast function is not altered in Siglec-15−/− mice To examine whether defects in Siglec-15 affect osteoblast-mediated bone formation in estrogen-deficient conditions, we performed calcein double-labeling on WT and Siglec-15−/− mice and compared dynamic bone formation parameters 2 wk after ovariectomy or sham operation. The mineral apposition rate (MAR) and bone formation rate/bone surface (BFR/B.Pm) were significantly lower in Siglec-15−/− mice compared to WT mice, but were not significantly increased after ovariectomy in either Siglec-15−/− or WT mice (Sup. Fig. 2A). This finding indicates that the preserved bone volume in Siglec-15−/− mice after ovariectomy is not due to upregulated bone formation, but rather to impaired osteoclast mediated bone resorption.

We previously reported that the Siglec-15 gene is not expressed in mouse primary osteoblasts and that bone formation ability is not altered in Siglec-15−/− osteoblasts [11]. To confirm whether this finding is maintained after ovariectomy, we flashed UBMC from WT and Siglec15−/− mice 2 wk after ovariectomy or sham operation and cultured them in α-MEM with 10% FBS containing 50 μg/ml ascorbic acid and 10 mM β-glycerophosphate. The bone nodule area in Siglec-15−/− UBMC was similar to that in WT UBMC and there was no significant difference in the bone nodule area between cells obtained from ovariectomized mice and sham-operated mice, indicating that Siglec-15 is not involved in osteoblast-mediated bone acquisition regardless of estrogen deficiency (Sup. Fig. 2B). Because osteogenic cells are important sources of RANKL or OPG to regulate osteoclast differentiation, we confirmed the expression of these genes in WT and Siglec-15−/− osteoblasts. As expected, expression of RANKL and OPG genes was similar between WT and Siglec-15−/− osteoblasts (Sup. Fig. 2C). To further confirm the function of Siglec15−/− osteoblasts to generate osteoclasts, we co-cultured WT BMMs with WT or Siglec-15−/− osteoblasts. WT BMMs developed osteoclasts similarly when cultured with either WT or Siglec-15−/− osteoblasts (Sup. Fig. 2D). Discussion The findings of the experiments presented here indicate that Siglec15 is involved in estrogen deficiency-induced bone loss as well as in physiologic bone remodeling. Although estrogen deficiency upregulated recruitment and induction of mononuclear osteoclasts, the terminal differentiation of osteoclasts was impaired in Siglec-15−/− mice. Mononuclear osteoclasts are thought to absorb bone matrix, but the bone resorptive capacity of mononuclear osteoclasts is much lower than

Y. Kameda et al. / Bone 71 (2015) 217–226

Fig. 7. TNF-α mediated signals are not attenuated in Siglec-15−/− cells. A. Preosteoclasts were starved for 1 h in αMEM and then stimulated with 100 ng/ml TNF-α. Phosphorylation of PI3K, Akt, Erk, JNK, p38, and IκBα was determined by immunoblot.

that of multinuclear giant osteoclasts. Consequently, relative bone loss after ovariectomy is much lower in Siglec-15−/− mice than in WT mice. These results suggest that Siglec-15 blockade protects against estrogen deficiency-induced bone loss. Our finding that Siglec-15 is involved in estrogen deficiency-induced bone loss suggests that Siglec-15 is essential for in vivo osteoclastogenesis, which is regulated by multiple osteoclastogenic factors, such as RANKL, TNF-α, and IL-6 [26]. Findings from our previous study [11] confirmed that Siglec-15 participates in estrogen deficiency-induced bone loss through modulating RANKL signaling. Therefore, we focused on the role of Siglec-15 in TNF-α induced osteoclastogenesis because upregulated production of TNF-α by activated T cells as well as increased secretion of RANKL from osteoblasts/stromal cells are mainly responsible for estrogen deficiency-induced bone loss [13–15]. In fact, Siglec15−/− BMMs failed to develop functional osteoclasts when cultured in the presence of TNF-α, indicating that Siglec-15 plays an important role in TNF-α induced osteoclastogenesis. This finding is consistent with a recent report that Siglec-15 neutralizing antibody inhibits the differentiation of TNF-α induced osteoclastogenesis in vitro [27]. The results of our in vitro experiments support the notion that Siglec-15 is involved in TNF-α induced osteoclastogenesis, although the precise mechanism of how Siglec-15 regulates TNF-α induced osteoclastogenesis remains unclear. Contrary to our expectation, the downstream signaling of TNF receptor 1, including the PI3K/Akt and MAP kinase pathways, was not attenuated in Siglec-15−/− cells, while Siglec-15 modulates RANKL-induced PI3K/Akt and Erk pathways in the development of osteoclasts [11]. Siglec-15 should recognize certain sialylated glycans and ligand occupancy of Siglec-15 somehow promotes TNF-α induced osteoclastogenesis; however, it is difficult to

225

identify the mechanistic actions of Siglec-15 on TNF-α induced osteoclast differentiation because a specific ligand for Siglec-15 has not yet been determined. Identification of such a ligand would promote further research, leading to a better understanding of the roles of Siglec-15 in TNF-α induced osteoclast differentiation. Of note is that Siglec-15 appears to be especially important for the cytoskeletal organization of osteoclasts. The signaling mediated by OSCAR/FcRγ, which is an alternative ITAM adaptor to DAP12, rescues multinucleation but not cytoskeletal organization of TNF-α and RANKL-induced Siglec-15−/− osteoclasts, indicating that the roles of Siglec-15/DAP12 and OSCAR/FcRγ in supporting osteoclast development do not completely overlap. This is not surprising because DAP12 has a dominant role in osteoclastogenesis as evidenced by the observation that DAP12, but not FcRγ, rescues impaired osteoclastogenesis of DAP12/FcRγ double-deficient cells [28]. Further studies are necessary, however, to elucidate the mechanisms of the Sigelc-15/DAP12 pathway in the cytoskeletal organization of osteoclasts. Although Siglec-15 has been shown to work as a DAR, the reported response to ovariectomy in DAP12−/− mice is inconsistent with that in Siglec-15−/− mice. Wu et al. [29] demonstrated that DAP12−/− mice show significant trabecular bone loss in the femur and tibia following ovariectomy similar to WT mice. A possible explanation for the discrepancy in response to ovariectomy between Siglec-15−/− mice and DAP12−/− mice is the conflicting function of DAP12. DAP12 was originally thought to transduce an activation signal, but recent studies demonstrated that DAP12 also transduces inhibitory signals [30,31]. Although the mechanism underlying the ability of DAP12 to switch between activating or inhibitory signaling is not completely understood, recent findings suggest that signaling downstream of DAP12 is dependent on the affinity and avidity of the ligands against DARs; highaffinity ligands cause activation signals, whereas low-affinity ligands generate inhibitory signals [30]. Given that the TREM2/DAP12 complex generates inhibitory signals [32,33], Siglec-15 deficiency may generate a condition in which inhibitory signaling is dominant. In conclusion, the findings of the present study indicate that Siglec15 is involved in estrogen deficiency-induced differentiation of osteoclasts and is therefore a potential therapeutic target for postmenopausal osteoporosis. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bone.2014.10.027. Disclosure All authors state that they have no conflict of interest. Acknowledgments We would like to thank Dr. Toshio Kitamura for the plasmids and cell lines, and Dr. Toshiyuki Akazawa for the bovine bone slices. This project was supported in part by a Grant-in-Aid for Research Activity Start-up from the Ministry of Education, Culture, Sports, Science, and Technology of Japan 23890008 (to M. Takahata). References [1] Kaifu T, Nakahara J, Inui M, Mishima K, Momiyama T, Kaji M, et al. Osteopetrosis and thalamic hypomyelinosis with synaptic degeneration in DAP12-deficient mice. J Clin Invest 2003;111:323–32. [2] Koga T, Inui M, Inoue K, Kim S, Suematsu A, Kobayashi E, et al. Costimulatory signals mediated by the ITAM motif cooperate with RANKL for bone homeostasis. Nature 2004;428:758–63. [3] Humphrey MB, Ogasawara K, Yao W, Spusta SC, Daws MR, Lane NE, et al. The signaling adapter protein DAP12 regulates multinucleation during osteoclast development. J Bone Miner Res 2004;19:224–34. [4] Mocsai A, Humphrey MB, Van Ziffle JA, Hu Y, Burghardt A, Spusta SC, et al. The immunomodulatory adapter proteins DAP12 and Fc receptor gamma-chain (FcRgamma) regulate development of functional osteoclasts through the Syk tyrosine kinase. Proc Natl Acad Sci U S A 2004;101:6158–63.

226

Y. Kameda et al. / Bone 71 (2015) 217–226

[5] Ishida-Kitagawa N, Tanaka K, Bao X, Kimura T, Miura T, Kitaoka Y, et al. Siglec-15 protein regulates formation of functional osteoclasts in concert with DNAX-activating protein of 12 kDa (DAP12). J Biol Chem 2012;287:17493–502. [6] Humphrey MB, Daws MR, Spusta SC, Niemi EC, Torchia JA, Lanier LL, et al. TREM2, a DAP12-associated receptor, regulates osteoclast differentiation and function. J Bone Miner Res 2006;21:237–45. [7] Dietrich J, Cella M, Seiffert M, Buhring HJ, Colonna M. Cutting edge: signal-regulatory protein beta 1 is a DAP12-associated activating receptor expressed in myeloid cells. J Immunol 2000;164:9–12. [8] Bakker AB, Baker E, Sutherland GR, Phillips JH, Lanier LL. Myeloid DAP12-associating lectin (MDL)-1 is a cell surface receptor involved in the activation of myeloid cells. Proc Natl Acad Sci U S A 1999;96:9792–6. [9] Klesney-Tait J, Turnbull IR, Colonna M. The TREM receptor family and signal integration. Nat Immunol 2006;7:1266–73. [10] Joyce-Shaikh B, Bigler ME, Chao CC, Murphy EE, Blumenschein WM, Adamopoulos IE, et al. Myeloid DAP12-associating lectin (MDL)-1 regulates synovial inflammation and bone erosion associated with autoimmune arthritis. J Exp Med 2010;207: 579–89. [11] Kameda Y, Takahata M, Komatsu M, Mikuni S, Hatakeyama S, Shimizu T, et al. Siglec15 regulates ssteoclast differentiation by modulating RANKL-induced phosphatidylinositol 3-kinase/Akt and Erk pathways in association with signaling adaptor DAP12. J Bone Miner Res 2013;28:2463–75. [12] Stuible M, Moraitis A, Fortin A, Saragosa S, Kalbakji A, Filion M, et al. Mechanism and Function of Monoclonal Antibodies Targeting Siglec-15 for Therapeutic Inhibition of Osteoclastic Bone Resorption. J Biol Chem 2014;289:6498–512. [13] Roggia C, Gao Y, Cenci S, Weitzmann MN, Toraldo G, Isaia G, et al. Up-regulation of TNF-producing T cells in the bone marrow: a key mechanism by which estrogen deficiency induces bone loss in vivo. Proc Natl Acad Sci U S A 2001;98: 13960–5. [14] Kimble RB, Bain S, Pacifici R. The functional block of TNF but not of IL-6 prevents bone loss in ovariectomized mice. J Bone Miner Res 1997;12:935–41. [15] Ammann P, Rizzoli R, Bonjour JP, Bourrin S, Meyer JM, Vassalli P, et al. Transgenic mice expressing soluble tumor necrosis factor-receptor are protected against bone loss caused by estrogen deficiency. J Clin Invest 1997;99:1699–703. [16] Fuller K, Murphy C, Kirstein B, Fox SW, Chambers TJ. TNFalpha potently activates osteoclasts, through a direct action independent of and strongly synergistic with RANKL. Endocrinology 2002;143:1108–18. [17] Azuma Y, Kaji K, Katogi R, Takeshita S, Kudo A. Tumor necrosis factor-alpha induces differentiation of and bone resorption by osteoclasts. J Biol Chem 2000;275: 4858–64. [18] Angata T, Tabuchi Y, Nakamura K, Nakamura M. Siglec-15: an immune system Siglec conserved throughout vertebrate evolution. Glycobiology 2007;17:838–46.

[19] Bouxsein ML, Boyd SK, Christiansen BA, Guldberg RE, Jepsen KJ, Muller R. Guidelines for assessment of bone microstructure in rodents using micro-computed tomography. J Bone Miner Res 2010;25:1468–86. [20] Dempster DW, Compston JE, Drezner MK, Glorieux FH, Kanis JA, Malluche H, et al. Standardized nomenclature, symbols, and units for bone histomorphometry: a 2012 update of the report of the ASBMR Histomorphometry Nomenclature Committee. J Bone Miner Res 2013;28:2–17. [21] Takeshita S, Kaji K, Kudo A. Identification and characterization of the new osteoclast progenitor with macrophage phenotypes being able to differentiate into mature osteoclasts. J Bone Miner Res 2000;15:1477–88. [22] Kaneda T, Nojima T, Nakagawa M, Ogasawara A, Kaneko H, Sato T, et al. Endogenous production of TGF-beta is essential for osteoclastogenesis induced by a combination of receptor activator of NF-kappa B ligand and macrophage-colony-stimulating factor. J Immunol 2000;165:4254–63. [23] Takahata M, Iwasaki N, Nakagawa H, Abe Y, Watanabe T, Ito M, et al. Sialylation of cell surface glycoconjugates is essential for osteoclastogenesis. Bone 2007;41:77–86. [24] Cenci S, Weitzmann MN, Roggia C, Namba N, Novack D, Woodring J, et al. Estrogen deficiency induces bone loss by enhancing T-cell production of TNF-alpha. J Clin Invest 2000;106:1229–37. [25] Barrow AD, Raynal N, Andersen TL, Slatter DA, Bihan D, Pugh N, et al. OSCAR is a collagen receptor that costimulates osteoclastogenesis in DAP12-deficient humans and mice. J Clin Invest 2011;121:3505–16. [26] Zhao R. Immune regulation of osteoclast function in postmenopausal osteoporosis: a critical interdisciplinary perspective. Int J Med Sci 2012;9:825–32. [27] Hiruma Y, Tsuda E, Maeda N, Okada A, Kabasawa N, Miyamoto M, et al. Impaired osteoclast differentiation and function and mild osteopetrosis development in Siglec-15-deficient mice. Bone 53: 87–93. [28] Zou W, Zhu T, Craft CS, Broekelmann TJ, Mecham RP, Teitelbaum SL. Cytoskeletal dysfunction dominates in DAP12-deficient osteoclasts. J Cell Sci 2010;123:2955–63. [29] Wu Y, Torchia J, Yao W, Lane NE, Lanier LL, Nakamura MC, et al. Bone microenvironment specific roles of ITAM adapter signaling during bone remodeling induced by acute estrogen-deficiency. PLoS One 2007;2:e586. [30] Turnbull IR, Colonna M. Activating and inhibitory functions of DAP12. Nat Rev Immunol 2007;7:155–61. [31] Hamerman JA, Tchao NK, Lowell CA, Lanier LL. Enhanced toll-like receptor responses in the absence of signaling adaptor DAP12. Nat Immunol 2005;6:579–86. [32] Turnbull IR, Gilfillan S, Cella M, Aoshi T, Miller M, Piccio L, et al. Cutting edge: TREM2 attenuates macrophage activation. J Immunol 2006;177:3520–4. [33] Hamerman JA, Jarjoura JR, Humphrey MB, Nakamura MC, Seaman WE, Lanier LL. Cutting edge: inhibition of TLR and FcR responses in macrophages by triggering receptor expressed on myeloid cells (TREM)-2 and DAP12. J Immunol 2006;177: 2051–5.