Siglec-15 Regulates Osteoclast Differentiation by Modulating RANKL-Induced Phosphatidylinositol 3-Kinase/Akt and Erk Pathways in Association With Signaling Adaptor DAP12



Siglecs are a family of sialic acid–binding immunoglobulin-like lectins that regulate the functions of cells in the innate and adaptive immune systems through glycan recognition. Here we show that Siglec-15 regulates osteoclast development and bone resorption by modulating receptor activator of nuclear factor κB ligand (RANKL) signaling in association with DNAX-activating protein 12 kDa (DAP12), an adaptor protein bearing an immunoreceptor tyrosine-based activation motif (ITAM). Among the known Siglecs expressed in myeloid lineage cells, only Siglec-15 was upregulated by RANKL in mouse primary bone marrow macrophages. Siglec-15–deficient mice exhibit mild osteopetrosis resulting from impaired osteoclast development. Consistently, cells lacking Siglec-15 exhibit defective osteoclast development and resorptive activity in vitro. RANKL-induced activation of phosphatidylinositol 3-kinase (PI3K)/Akt and Erk pathways were impaired in Siglec-15–deficient cells. Retroviral transduction of Siglec-15–null osteoclast precursors with wild-type Siglec-15 or mutant Siglec-15 revealed that the association of Siglec-15 with DAP12 is involved in the downstream signal transduction of RANK. Furthermore, we found that the ability of osteoclast formation is preserved in the region adjacent to the growth plate in Siglec-15–deficient mice, indicating that there is a compensatory mechanism for Siglec-15–mediated osteoclastogenesis in the primary spongiosa. To clarify the mechanism of this compensation, we examined whether osteoclast-associated receptor (OSCAR)/Fc receptor common γ (FcRγ) signaling, an alternative ITAM-mediated signaling pathway to DAP12, rescues impaired osteoclastogenesis in Siglec-15–deficient cells. The ligands in type II collagen activate OSCAR and rescue impaired osteoclastogenesis in Siglec-15–deficient cells when cultured on bone slices, indicating that Siglec-15–mediated signaling can be compensated for by signaling activated by type II collagen and other bone matrix components in the primary spongiosa. Our findings indicate that Siglec-15 plays an important role in physiologic bone remodeling by modulating RANKL signaling, especially in the secondary spongiosa. © 2013 American Society for Bone and Mineral Research.


Osteoclasts are bone-resorbing cells with a pivotal role in physiologic bone remodeling. Osteoclasts are differentiated from hematopoietic precursors of the monocyte/macrophage lineage by stimulation with a tumor necrosis factor family cytokine, receptor activator of nuclear factor κB ligand (RANKL) and macrophage colony-stimulating factor (M-CSF). Because these two cytokines induce osteoclasts in vitro, they were thought to be sufficient for osteoclastogenesis, but recent studies demonstrated that osteoclastogenesis also requires immunoreceptor tyrosine-based activation motif (ITAM) signaling.[1-4]

The primary ITAM-bearing adaptors in osteoclast lineage cells are DNAX-activating protein (DAP) 12 and Fc receptor common γ chain (FcRγ). Mice doubly deficient in DAP12 and FcRγ exhibit severe osteopetrosis owing to the impaired development of osteoclasts, demonstrating that signaling through these two ITAM-bearing adaptors is essential for osteoclastogenesis.[3, 4] DAP12, but not FcRγ, rescues impaired osteoclastogenesis of DAP12/FcRγ doubly-deficient cells,[5] suggesting that DAP12 has a dominant role in osteoclastogenesis. The precise role of DAP12 in the development of osteoclasts, however, remains controversial. DAP12 impacts osteoclast development by modulating RANKL-mediated Ca2+ signaling, which enhances the induction of the master regulator of osteoclastogenesis, nuclear factor of activated T cells c1 (NFATc1), through an autoamplification mechanism.[3] The observation that DAP12-deficient macrophages undergo osteoclastogenesis but fail to form an actin ring[5] suggests that DAP12 promotes bone resorption, principally by modulating the cytoskeletal organization of osteoclasts.

Immunoreceptors that associate with ITAM-bearing adaptors might be key to understanding the function of DAP12 in osteoclast development and function. Because both DAP12 and FcRγ have minimal extracellular domains and are therefore incapable of sensing signals outside the cells, immunoreceptors associated with DAP12 or FcRγ are thought to regulate ITAM signaling. Several specific immunoreceptors have been identified in osteoclast lineage cells: DAP12-associated immunoreceptors (DARs) include triggering receptor expressed in myeloid cells 2 (TREM-2)[6] and signal-regulatory protein (SIRP) β1; FcRγ-associated immunoreceptors include osteoclast-associated receptor (OSCAR)[3, 7] and paired Ig-like receptor-A.[8] Ligands for these receptors and their exact function in the bone environment remain to be elucidated.

Recent screening of giant cell tumor-specific genes revealed a novel type of immunoreceptor, sialic acid binding Ig-like lectin (Siglec)-15, as an important factor for osteoclastogenesis.[9] The most recent study revealed an osteopetrotic phenotype of Siglec-15–deficient mice, suggesting that Siglec-15 is essential for osteoclast function in vivo.[10] Siglecs are cell-surface immunoglobulin-like lectins of vertebrates that recognize sialylated glycans and are involved in many physiologic processes, such as glycoprotein turnover, cellular trafficking, and pathogen recognition.[11, 12] Most Siglecs are expressed on immune cells and have an immunoreceptor tyrosine-based inhibitory motif in the cytoplasmic domain, and they negatively regulate the cells that express them. In contrast, Siglec-15 is associated with DAP12 and is therefore expected to positively regulate immune cells through ITAM-mediated signals.[13, 14] Although the physiologic function of Siglec-15 is unclear, it is considered to possess an essential biologic function in vertebrates because of the highly conserved sequence of the Siglec-15 gene throughout vertebrate evolution.

In the present study, we used gene-targeted mice to examine the effects of Siglec-15 deficiency on osteoclast development.

Materials and Methods

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).[14] Antibodies to DAP12 were purchased from Exalpha Biologicals Inc. (Shirley, MA, USA), antibodies to vav3 from Santa Cruz Biotechnology (Santa Cruz, CA, USA), mouse monoclonal antibodies for Syk from Abcam (Tokyo, Japan), anti-phosphotyrosine mAb 4G10 from Upstate (Billerica, MA, USA), antibodies for phospho-Erk, Erk, phospho-Akt, Akt, phospho-PI3K p85, PI3K p85, phospho-JNK, JNK, phospho-p38, p38, phospho-IκBα, IκBα, phospho-Syk, and Myc were purchased from Cell Signaling Technology (Tokyo, Japan). Recombinant human M-CSF and soluble RANKL were purchased from PeproTech EC, Ltd. (London, UK). Rat collagen I-coated plates were purchased from BD Falcon (Tokyo, Japan), and 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).

Generation of the Siglec-15 targeted allele

The Siglec-15 targeting vector was designed to replace exons 2 to 5 encoding most of the functional domains other than a part of the signal peptide of Siglec-15 with a neomycin-resistance gene (Neo) cassette (Supplemental Fig. S1A).{FIG S1} The targeting plasmid was linearized and electroporated into C57BL/6-derived embryonic stem cells. To generate chimeric mice, a targeted embryonic stem clone was injected into Balb/c blastocysts. These chimeric males were mated to C57BL/6 females, resulting in heterozygous F1 offspring. Heterozygous offspring were intercrossed to generate homozygous embryos. For genotyping analysis, genomic PCR was performed using a common primer (WT-R) in combination with primers discriminating wild-type Siglec-15 alleles (WT-F) and targeted alleles (Neo-F). Another primer pair (Neo-f or Neo-r) was used to confirm insertion of the Neo cassette (Supplemental Fig. S1B).

Micro-CT analysis

Left tibias and 5th lumbar vertebral bodies were scanned individually by micro-computed tomography (CT, R_mCT2; Rigaku, Tokyo, Japan) at a 10-µm isotropic resolution. Tibias were measured using a TRI/3D-BON (Ratoc System Engineering Co., Tokyo, Japan) in accordance with the guidelines described in Bouxsein and colleagues.[15] 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.

Histology and histomorphometry

Proximal tibias of 9-week-old mice were fixed in paraformaldehyde, decalcified in EDTA, and embedded in paraffin. Sections were stained for TRAP with methyl green counterstain to observe osteoclasts.

For histomorphometry analysis, tibias of 14-week-old mice were harvested, dehydrated in ethanol, and undecalcified sections were stained in Villanueva Bone Stain. Bone section preparation and histomorphometric analysis were performed at Ito Bone Histomorphometry Institute (Japan).

The numbers of osteoclasts/tissue volume, osteoclasts/bone surface, osteoclast surface/bone surface, and eroded surface/bone surface at the primary and secondary spongiosa were measured according to Parfitt and colleagues.[16] 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.[17] Briefly, the marrow cells were collected from femurs and tibias of 6- to 9-week-old mice. After removing the red blood cells, marrow cells were cultured on a suspension culture dish in the presence of 50 ng/mL M-CSF to enrich the CD11b+ (Mac1+) population.[18] After 3-day culture, the cells were washed twice with phosphate-buffered saline (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 RANKL for 3 days at 37°C in a 5% humidified CO2 incubator to generate osteoclasts. M-CSF and RANKL 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 I- or II-coated plastic culture wells or 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.

In vitro resorption assay of cultured osteoclasts

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

Real-time quantitative polymerase chain reaction (qPCR) of gene expression

Total RNA from cultured cells was isolated using the Qiagen RNeasy Mini Kit (Qiagen, Valencia, CA, USA). cDNA was synthesized from 1 µg total RNA using reverse transcriptase and oligo-dT primers for qPCR analyses. qPCR was performed as previously described.[19] cDNA samples were analyzed for both the genes of interest and the reference gene (glyceraldehyde-3-phosphate dehydrogenase). Primer sequences used for the experiment are shown in Supplemental Table S1.{TBL S1} The amount of mRNA expressed was normalized to the glyceraldehyde-3-phosphate dehydrogenase expression.


Cells were cultured on glass coverslips and fixed for 5 minutes 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, USA). 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, Kumamoto, Japan).

Flow cytometry

To analyze the kinetics of Siglec-15 expression on the plasma membrane during osteoclastogenesis, we quantitatively investigated the Siglec-15 content by flow cytometry. We prepared BMMs: 1) BMMs without RANKL stimulation and 2) BMMs stimulated with M-CSF and RANKL for 3 days. Mouse BD Fc Block (Becton Dickinson Co., San Jose, CA, USA) was used to block the Fc-mediated adherence of antibodies to mouse Fc receptors to reduce background staining before reacting these cells with the Siglec-15 antibody. These cells were incubated with the Siglec-15 antibody or control rabbit IgG, followed by staining with the Alexa Fluor 488-labeled secondary antibody (Invitrogen Molecular Probes). Cell surface expression of Siglec-15 was analyzed by flow cytometry (FACSCanto, BD Biosciences, San Jose, CA, USA).

Retroviral induction

Complementary DNAs were subcloned into pMX-puro (kindly provided by Toshio Kitamura, Tokyo University).[20] Retroviral vectors pMX Siglec-15 Myc-His were constructed by first inserting Siglec-15 cDNA into pcDNA3.1/myc-His A vector, and then a Siglec-15 myc-His sequence was transferred into pMX-Puro. pMX Siglec-15 mutants, R143A, and K273A were prepared with a QuikChange kit (Stratagene, La Jolla, CA) using point mutant primers. The resulting vectors were used to transfect Plat E cells using Lipofectamine LTX plus Reagent (Invitrogen), and then recombinant retroviruses were generated. Forty-eight hours after transfection, culture supernatants were harvested and used for infection. BMMs were infected with virus for 6 to 12 hours in the presence of M-CSF (30 ng/mL) and polybrene (4 μg/mL).

RNA interference against OSCAR

Short hairpin RNAs (shRNAs) were designed to target mouse OSCAR using an shRNA sequence designing tool published by Takara Bio Inc. (Otsu, Japan). The designed sequence was inserted into a pSIREN RetroQ vector (Takara Bio). The shRNA sequences used in the experiment were as follows.



The retrovirus was constructed and introduced into the BMMs as described above.

Tartrate-resistant acid phosphatase (TRAP) 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 minute 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 washed twice with ice-cold PBS. Cell lysates were then extracted using the PhosphoSafe Extraction Reagent (Novagen, Madison, WI, USA). Cell lysates were subjected to immunoblot or immunoprecipitation analyses using the indicated antibodies.

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 ± SD.

Study approval

The Ethics Review Committee for Animal Experimentation of Hokkaido University approved the experimental protocol.


Upregulation of Siglec-15 during osteoclastogenesis and its preferential distribution on the plasma membrane of migrating osteoclasts

We first examined the transcript levels of Siglec family proteins, including Siglec-1, -3, -5 (also known as Siglec-F), -15, and -H, which are expressed in mouse myeloid lineage cells,[12, 21] before and after RANKL stimulation. In mouse BMMs, Siglec-1, -3, -5, and -H expression was downregulated or unchanged after RANKL stimulation, and only Siglec-15 expression was markedly upregulated after RANKL stimulation (82-fold increase at day 3 versus day 0; Fig. 1A). We also confirmed increased expression of Siglec-15 protein during osteoclastogenesis (Fig. 1B).

Figure 1.

Upregulation of Siglec-15 during osteoclastogenesis and its unique distribution patterns in migrating and resorbing osteoclasts. Bone marrow macrophages (BMMs) were cultured in the presence of M-CSF and RANKL for the indicated number of days. (A) mRNA expression of Siglec-1, -3,-5, -15, and -H during osteoclast differentiation in mouse BMMs. (B) Siglec-15 protein expression was assessed by Western blotting. (C) Immunocytochemical study of Siglec-15 stained with anti-Siglec-15 antibody, followed by Alexa Fluor 488 goat anti-rabbit antibody in BMMs and osteoclasts. Nuclei were stained with DAPI and actin was stained with Alexa Fluor 633 phalloidin. After RANKL stimulation, Siglec-15 was expressed in the polykaryocytes with an actin ring and in some monocytes, and was more abundantly expressed in polykaryocytes with lamellipodia (white arrows). Scale bar = 100 μm. (D) Flow cytometric histograms showing Siglec-15 expression on the cell surface of BMMs during osteoclastogenesis. Left panel represents BMMs and right panel represents cells cultured with M-CSF and RANKL for 3 days. Red outlined histograms indicate control cells stained with rabbit IgG, and blue outlined histograms represent the Siglec-15 level measured by Alexa Fluor 488-labeled secondary antibody. Numbers indicate the percentage of cells in the indicated regions. (E, F) Confocal microscopic images of migrating polykaryons characterized by multiple nuclei and lamellipodia (E) and polykaryocytes with actin ring formation (F). Siglec-15 is distributed on the plasma membrane of migrating polykaryocytes, while it is expressed in the cytoplasm of polykaryocytes with an actin ring. Scale bar = 100 μm.

We then performed immunocytochemistry using the Siglec-15–specific antibody on BMMs cultured in the presence or absence of RANKL (Fig. 1C). After culture with M-CSF and RANKL for 3 days, certain subsets of cells became positive for Siglec-15. Migrating polykaryons, which were characterized by multiple nuclei and lamellipodia, expressed an especially high level of Siglec-15 on the plasma membrane (Fig. 1C, E). Polykaryocytes with an actin ring expressed Siglec-15 in the cytoplasm and sometimes on the plasma membrane (Fig. 1C, F). Some mononuclear cells expressed Siglec-15, but many of them only sparsely expressed Siglec-15 even at day 3 after RANKL stimulation.

Flow cytometric analysis was performed to confirm Siglec-15 expression in a heterogeneous cell population after RANKL stimulation (Fig. 1D). Before RANKL stimulation, low expression of Siglec-15 was detected on the BMM surface. RANKL stimulation induced a heterogeneous population of Siglec-15 high- and low-expressing cells. On day 3 after RANKL stimulation, the percentage of Siglec-15–positive cells increased from 0.6% to 14.1%. The population of Siglec-15–positive cells on day 3 after RANKL stimulation was smaller than we expected based on the results of immunocytochemistry, possibly because multinuclear giant cells, which express a high level of Siglec-15, were not included in the flow cytometric analysis because of their large size.

Siglec-15–deficient mice have increased bone mass

To investigate the physiologic role of Siglec15 in bone metabolism, we generated Siglec-15 deficient (Siglec-15−/−) mice. Siglec-15−/− mice are fertile, show no significant abnormalities in appearance, and grow normally until at least 12 months of age. To analyze possible abnormalities in skeletal tissues, we compared micro-CT scan images of tibias and 5th lumbar vertebral bodies in 9-week-, 6-month-, and 12-month-old wild-type (WT) and Siglec-15−/− mice. In Siglec-15−/− mice, there was a clear increase in the trabecular bone mass in both the tibias and 5th lumbar vertebral bodies (Fig. 2A). Quantitative measurements of the proximal metaphyseal region of the tibias revealed increases in the trabecular tissue volume (BV/TV; trabecular bone volume [BV] to total volume [TV]), trabecular number (Tb.N), and trabecular thickness (Tb.Th) in Siglec-15−/− mice compared with WT mice (Fig. 2B–D).

Figure 2.

Osteopetrotic phenotype in mice lacking Siglec-15. (A) Micro-CT of proximal tibias and 5th lumbar vertebral bodies of WT and Siglec-15−/− mice at 9 weeks of age. (B–D) Bone volume and microstructural indices of trabecular bone of the metaphyseal region of the tibias of WT and Siglec-15−/− mice (*p < 0.05). BV/TV, bone volume/total volume; Tb.N, trabecular number; Tb.Th, trabecular thickness.

Siglec-15–deficient mice exhibited mild osteopetrosis owing to impaired osteoclast development and function

We then investigated whether the mild osteopetrotic phenotype of Siglec-15−/− mice is the result of impaired osteoclast development. We first analyzed histologic sections of proximal tibias from 9-week-old WT and Siglec-15−/− mice (Fig. 3A). As expected, Siglec-15−/− mice had fewer osteoclasts compared with WT, and the appearance and number of osteoclasts varied by anatomic site (Fig. 3B, C). The number of multinucleated osteoclasts at the secondary spongiosa was much lower in Siglec-15−/− mice than in WT mice, whereas many multinucleated osteoclasts were observed at the primary spongiosa in Siglec-15−/− mice compared with WT mice. Further, Siglec-15−/− osteoclasts were small and did not spread on the bone surface in the secondary spongiosa. Consequently, fewer osteoclasts attached to the bone surface and eroded surface in Siglec-15−/− mice than in WT mice, and the difference was greater in the secondary spongiosa compared with the primary spongiosa. These results indicate that Siglec-15 deficiency leads to defective osteoclast development, but there is a compensatory mechanism against Siglec-15–mediated osteoclastogenesis at the primary spongiosa.

Figure 3.

Impaired osteoclast development and bone resorption in Siglec-15−/− mice. (A) Micrographs of the proximal tibia of WT and Siglec-15−/− mice 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. TRAP-positive multinuclear cells were observed in both WT and Siglec-15−/− mice, but TRAP-positive cells appeared smaller in size in Siglec-15−/− mice, especially at the secondary spongiosa, compared with WT mice. Scale bar = 100 μm. (B, C) Histomorphometric parameters were related to osteoclastic bone resorption at the primary and secondary spongiosa of the tibia. The number of osteoclasts/tissue volume (N.Oc/T.Ar) was similar between WT and Siglec-15−/− mice, whereas the numbers of osteoclasts/bone surface (N.Oc/B.Pm), osteoclast surface/bone surface (Oc.Pm/B.Pm), and eroded surface/bone surface (E.Pm/B.Pm) were lower in Siglec-15−/− mice compared with WT mice (*p < 0.05). The number of multinucleated osteoclasts/bone surface (N.Mu.Oc/B.Pm) at the primary spongiosa was similar between WT and Siglec-15−/− mice, whereas that at the secondary spongiosa was lower in Siglec-15−/− mice (*p < 0.05).

Because osteoblasts regulate bone mass by producing bone matrix, we examined whether the Siglec-15 gene is expressed during osteoblast differentiation and whether its deficiency increases bone formation in Siglec-15−/− mice. Dynamic bone formation parameters, including mineral apposition rate, bone formation rate/bone surface, and mineralized surface/bone surface, were decreased in Siglec-15−/− mice in both primary and secondary spongiosa, suggesting that bone formation is suppressed in Siglec-15−/− mice (Supplemental Fig. S2A). {FIG S2} The impaired bone formation in Siglec-15−/− mice is considered to be a secondary effect of impaired bone resorption because the Siglec-15 gene is not expressed in mouse primary osteoblasts and Siglec-15−/− cells and WT cells exhibit no differences in the alkaline phosphatase level in osteoblast cultures (Supplemental Fig. S2B).

Siglec-15 is required for the development of functional osteoclasts

To confirm that Siglec-15 is involved in osteoclast development, we compared the abilities to form osteoclasts and to resorb bone between Siglec-15–deficient BMMs and WT BMMs. Although TRAP-positive cells were formed from Siglec-15−/− cells, most of them were mononuclear and the number of TRAP-positive multinucleated cells was significantly lower in Siglec-15−/− cells than in WT cells (Fig. 4A). 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 10-day culture with M-CSF and RANKL was 8.9% ± 5.3% in Siglec-15−/− cells, whereas it was 77.5% ± 7.7% in WT cells (Fig. 4B). The induction of osteoclast marker genes, including TRAP, cathepsin K, integrin β3, and calcitonin receptor, on day 3 after RANKL stimulation was inhibited in Siglec-15−/− BMMs compared with WT BMMs, indicating that Siglec-15 is required for the development of functional osteoclasts in vitro (Fig. 4C).

Figure 4.

Siglec-15 deficiency results in abnormal osteoclast development and function in vitro. (A) Differentiation efficiency of bone marrow macrophages (BMMs) into tartrate-resistant acid phosphatase (TRAP)-positive multinuclear cells is suppressed in Siglec-15−/− cells compared with WT cells (*p < 0.05). BMMs were cultured in the presence of M-CSF and RANKL for 3 days, after which the cells were stained for TRAP. The number of TRAP-positive multinuclear cells (MNC; 3 or more nuclei per cell) in each well of the 48-well plates was counted. Scale bar = 100 μm. (B) BMMs were cultured in the presence of M-CSF and RANKL on bovine bone slices for 10 days. The percentage of substrate resorption (brown areas) was quantified. Scale bar = 100 μm. (C) Expression of osteoclast marker genes was examined by quantitative RT-PCR. BMMs cultured for 2 days and 3 days were used as preosteoclast (PreOC) and osteoclasts (OC), respectively. CTR = calcitonin receptor. (D) BMMs were cultured on bovine bone slices in the presence of M-CSF and RANKL for 4 days, after which the cells were stained for TRAP. (E) 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, whereas those lacking Siglec-15 fail to organize actin rings (white arrow). Scale bar = 100 μm.

Osteoclasts cultured on plastic wells are morphologically different from those cultured on bone slices. Therefore, to further investigate the morphologic abnormality of Siglec-15−/− osteoclasts, we cultured BMMs on bovine bone slices with M-CSF and RANKL. We performed TRAP staining or visualized actin cytoskeleton by staining the cells with phalloidin. As expected, TRAP-positive cells formed from Siglec-15−/− cells were mostly mononuclear and could not form actin rings, even in some multinucleated cells (Fig. 4D, E), indicating that Siglec-15 is involved in the differentiation of functional osteoclasts.

Siglec-15, in association with DAP12, regulates the 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), which is necessary for association with DAP12.[14] We therefore asked whether Siglec-15 signaling requires DAP12 in developing functional osteoclasts. To address this question, we constructed a point mutant of Siglec-15 (K273A), which results in the loss of DAP12-binding ability. Retroviral transduction of WT Siglec-15 dramatically rescued well-spread multinucleated osteoclast formation (Fig. 5A, B), which was not observed in a K273A mutant (data not shown), indicating that Siglec-15 in association with DAP12 regulates the development of functional osteoclasts.

Figure 5.

Retroviral transduction of Siglec-15 rescued impaired development of osteoclasts in Siglec-15−/− osteoclasts. (A) Osteoclasts generated after retroviral transduction with Siglec-15 or an empty pMX vector. Scale bar = 100 μm. (B) The number of TRAP-positive MNC was counted. Siglec-15 successfully rescued the impaired development of osteoclasts in Siglec-15−/− cells.

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 constructed 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 (data not shown).

RANKL-induced activation of PI3K/Akt and Erk pathway was impaired in Siglec-15–deficient cells

To clarify the role of Siglec-15 in osteoclast differentiation, we examined the downstream signaling of RANK in Siglec-15−/− preosteoclasts. Siglec-15 deficiency led to the inhibition of RANKL-induced phosphorylation of PI3K p85, Akt, and Erk (Fig. 6A). We then examined whether association of Siglec-15 with DAP12 is required for the downstream signaling of RANK by retroviral transduction of WT-Siglec-15 and K273A mutant to Siglec-15−/− cells. The impaired signaling of RANKL in Siglec-15−/− cells was rescued by the induction of WT-Siglec-15 but not by the K273A mutant, suggesting that association of Siglec-15 with DAP12 is necessary for the modulation of the RANKL signaling (Fig. 6B).

Figure 6.

Siglec-15, in association with DAP12, mediates RANKL-induced phosphorylation of PI3K, Akt, and Erk. (A) Preosteoclasts were starved for 1 hour in αMEM and then stimulated with 100 ng/mL RANKL. Phosphorylation of PI3K, Akt, Erk, JNK, p38, and IκBα was determined by immunoblot. (B) Siglec-15−/− preosteoclasts, transduced with either WT-Siglec-15 or K273 mutant, were starved for 1 hour in αMEM and then stimulated with 100 ng/mL RANKL. (C) Preosteoclasts were lifted with 0.25% Trypsin-EDTA and then maintained in suspension (Sus) or plated on osteopontin (OPN) for 30 minutes to activate the integrin signal. DAP12, Syk, and Vav3 tyrosine phosphorylation were determined by immunoprecipitation and immunoblot. DAP12, Syk, and Vav3 content in total cell lysates were used as loading control. (D) BMMs were cultured on type II collagen-coated dishes in the presence of M-CSF and RANKL for 4 days to generate giant cells. Osteoclasts were starved for 1 hour in αMEM and then stimulated with 100 ng/mL M-CSF for 5 minutes. Syk phosphorylation was detected by immunoblot, and activated forms of Rac and cdc42 were detected by GST pull-down assay.

The observation that Siglec-15 deficiency leads to cytoskeletal dysfunction also led us to examine the canonical downstream signaling pathways of DAP12. Contrary to our expectation, Siglec-15 deficiency did not attenuate osteopontin-induced integrin signaling (Fig. 6C) or M-CSF signaling (Fig. 6D), which trigger DAP12 signaling to promote the cytoskeletal rearrangement of osteoclasts.[22, 23]

Siglec-15 has little impact on NFATc1 induction through calcium signaling

DAP12 is essential for NFATc1 induction through calcium signaling.[3] To test whether Siglec-15 is required for calcium signaling, we observed Ca2+ oscillations in WT and Siglec-15 BMMs cultured for 72 hours in the presence of M-CSF and RANKL (data not shown). Although the oscillation frequency was slightly lower in Siglec-15−/− cells, the difference between WT and Siglec-15−/− cells was not statistically significant. Supporting this finding, nuclear translocation of NFATc1 was similar between WT and Siglec-15−/− cells on day 2 after RANKL stimulation. The NFATc1 mRNA expression level was slightly lower in Siglec-15−/− cells compared with WT cells on day 3 after RANKL stimulation. Western blotting analysis, however, indicated similar expression of NFATc1 protein between WT and Siglec-15−/− cells during osteoclastogenesis (data not shown). Hence, Siglec-15 is not crucial for calcium signaling during osteoclast development.

OSCAR/FcRγ signal rescues impaired osteoclast differentiation in the presence of bone matrix in Siglec-15−/− osteoclasts

The ability to form osteoclasts differs between in vitro and in vivo studies: Siglec-15−/− cells barely differentiate into mature multinuclear osteoclasts in vitro, whereas multinuclear osteoclasts are observed, especially at the primary spongiosa, in Siglec-15−/− mice. To explore the mechanism underlying this discrepancy, we examined the effect of possible compensatory factors on in vitro osteoclastogenesis and the function of Siglec-15−/− cells.

Another ITAM-bearing adaptor, FcRγ, compensates for DAP12 function in osteoclastogenesis.[3] Therefore, we tested whether FcRγ signals rescue impaired osteoclast development in Siglec-15−/− cells. The precise mechanism underlying the activation of FcRγ signals remains unclear, but the osteoblasts are expected to have ligands that stimulate FcRγ signaling.[24] Recent findings demonstrated that collagen binds to OSCAR and stimulates osteoclastogenesis through FcRγ signaling.[25] Therefore, Siglec-15−/− BMMs were co-cultured with WT mouse calvarial osteoblasts or cultured on type I or II collagen-coated culture wells.

In osteoblast co-culture, TRAP-positive multinuclear osteoclasts were formed from Siglec-15−/− cells, but these were much smaller than WT osteoclasts and did not spread as well as WT osteoclasts, indicating that osteoblast-mediated FcRγ signaling partially compensates for Siglec-15 deficiency (Fig. 7A). Although type I collagen could not rescue Siglec-15−/− osteoclastogenesis, type II collagen dramatically rescued multinucleated osteoclast formation in Siglec-15−/− BMMs (Fig. 7B, C). Most Siglec-15−/− osteoclasts induced by type II collagen, 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 from the abnormal appearance in TRAP staining, Siglec-15−/− osteoclasts induced by type II collagen failed to form actin rings when cultured on plastic wells, although they had the appearance of osteoclasts in brightfield images (Fig. 7E). To confirm the finding that type II collagen stimulates OSCAR and subsequently rescues giant cell formation in Siglec-15−/− BMMs, we knocked down OSCAR gene expression using retrovirally infected shRNA against OSCAR. Siglec-15−/− BMMs lacking OSCAR could not rescue multinucleation when cultured in type II collagen-coated culture wells (Fig. 7D).

Figure 7.

OSCAR/FcRγ signal rescues impaired osteoclast differentiation in the presence of bone matrix in Siglec-15−/− osteoclasts. (A–D) Osteoclastogenesis assay in the presence of possible OSCAR ligands. (A) BMMs co-cultured with WT osteoblasts for 7 days. TRAP-positive MNC (nuclei ≧3) were formed from Siglec-15−/− cells; however, these were smaller than WT osteoclasts and fewer in number than WT cells (*p < 0.05). (B, C) BMMs cultured on type I collagen- or type II collagen-coated culture wells for 3 days. Type II collagen induced TRAP-positive multinuclear cells from Siglec-15−/− cells comparable to that in WT cells, whereas type I collagen did not. (D) BMMs, knocked down with shRNA against OSCAR, cultured on type II collagen-coated culture wells. In the absence of OSCAR, type II collagen did not induce TRAP-positive multinuclear cells. (E) Cytoskeletal organization defects were not rescued by the FcRγ signal. Osteoclasts cultured on type II collagen-coated wells were immunostained with phalloidin and DAPI after observation with brightfield microscopy. Actin ring formation is impaired in Siglec-15−/− polykaryocytes on type II collagen. (F) TRAP-stained osteoclasts cultured on type II collagen-coated bovine bone slice. WT and Siglec-15−/− BMMs both differentiated to multinucleated osteoclasts. (G) 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. Scale bar = 100 μm (A–G).

We then performed an osteoclastogenesis assay on type II collagen-coated bovine bone slices to simulate the bone environment at the primary spongiosa. Siglec-15−/− osteoclasts restored cytoskeletal deficiency as well as multinuclear osteoclast formation when cultured in the presence of type II collagen and bone matrix. This indicates that OSCAR/FcRγ signaling, in concert with the signaling activated by some factors contained in bone matrix, rescues defective osteoclast development in Siglec-15−/− cells (Fig. 7F, G).


Siglecs are a family of sialic acid–binding immunoglobulin-like lectins that are mainly expressed by immune system cells.[12, 21] Each Siglec has a unique specificity for sialylated ligands and is therefore thought to mediate cell-cell interactions, signaling functions, and cell-pathogen interactions. Given that cell surfaces of immune cells are richly equipped with a complex mixture of glycans, it is conceivable that glycan recognition by Siglec-15 plays an important role in the development of osteoclasts, which are differentiated from hematopoietic precursors of the monocyte/macrophage lineage.

Our experiments using genetically manipulated mice revealed that Siglec-15 regulates the bone resorptive activity of osteoclasts. Although Siglec-15 is expressed in macrophages, dendritic cells, and osteoclasts,[9, 14, 26] its biologic function is poorly characterized as with most other Siglecs. Our data, similar to the previous report by Hiruma and colleagues,[10] show that although skeletal growth is normal, mild osteopetrosis is observed in Siglec-15−/− mice at all time points observed from 9 weeks to 1 year, indicating that Siglec-15 is essential for physiologic bone remodeling.

The results of the present study suggest that Siglec-15 is a key modulator for the development of functional osteoclasts, as reported previously.[9, 26] The bone resorptive activity of osteoclasts is attenuated, but it is not completely abrogated in Siglec-15−/− mice, suggesting that Siglec-15 is not critical but functions as a positive regulator for the bone resorptive activity of osteoclasts in the bone environment. Given that osteoclasts lacking Siglec-15 fail to spread on the bone surface in vivo and that osteoclast differentiation is compromised in Siglec-15−/− cells in vitro, it is likely that Siglec-15 regulates bone resorption by modulating key osteoclastogenic signaling. Of note, morphologic abnormalities of osteoclasts as well as the mild osteopetrosis phenotype in Siglec-15−/− mice are quite similar to those in mice lacking DAP12, which mediates key signaling related to the differentiation and cytoskeletal organization of osteoclasts.[3, 22, 27]

A previous study showed that Siglec-15 associates with the activating adaptor proteins DAP12 and DAP10 via its lysine residue in the transmembrane domain, implying an activating signaling potential.[14, 26] As expected, Siglec-15 co-precipitated with DAP12 in osteoclasts and disruption of the Siglec-15/DAP12 interaction failed to rescue the impaired development of functional osteoclasts in Siglec-15−/− cells, suggesting that Siglec-15 functions as a pairing partner of DAP12 in osteoclast development. It is unclear, however, why other DARs, such as TREM-2, SIRPb1, and myeloid DAP12-associating lectin-1, which are expressed on osteoclasts or their precursor cells, cannot compensate for Siglec-15 function. Given that the osteopetrotic phenotype has not been confirmed in mice lacking these DARs, Sigelc-15 may have a specific role in modulating DAP12-mediated signals in osteoclast development and function.

In fact, the known DAP12-mediated signaling was not compromised in Siglec-15−/− cells. DAP12-mediated signal transduction is initiated by the phosphorylation of two tyrosine residues of the DAP12 ITAM motif by c-src, which is activated by the engagement of integrin αvβ3 or M-CSF stimulation.[22, 27-29] A nonreceptor tyrosine kinase, Syk, is then recruited to the dually phosphorylated ITAM motif of DAP12, triggering the activation of Syk and its direct binding to the Vav3, phospholipase Cγ, PI3K, and SH2 domain-containing leukocyte protein 76, which govern the differentiation and cytoskeletal organization of osteoclasts.[3, 30-32] These DAP12-mediated signaling could be activated by integrin engagement and M-CSF stimulation in Siglec-15−/− cells comparable to WT cells. Although Ishida-Kitagawa and colleagues demonstrated that integrin signaling can induce the recruitment of Syk to the Siglec-15/DAP12 complex, our data suggest that other DARs compensate for the recruitment and phosphorylation of Syk to DAP12 in the absence of Siglec-15.

The results of this study show that the impaired development of osteoclasts in Siglec-15−/− cells is associated with reduced activation of PI3K p85, Akt, and Erk downstream of RANK. Akt is a downstream effector of PI3K,[33] and inhibition of either PI3K or Akt is reported to disrupt the actin ring formation of osteoclasts,[34, 35] suggesting that Siglec-15 regulates cytoskeletal organization of osteoclasts by modulating the PI3K/Akt signaling pathway. Another study using PI3K p85α-deficient (p85α−/−) mice showed that impaired osteoclast development in cells lacking PI3K p85α is associated with reduced activation of Akt and Erk.[36] P85α−/− osteoclasts revealed a significant reduction in the expression of several genes associated with the maturation and migration of osteoclasts, similarly to the Siglec-15−/− osteoclasts. Although the molecular mechanism of actions of Siglec-15 in RANKL signaling remain unclear, these data support the notion that Siglec-15, in association with DAP12, plays a specific role in modulating key osteoclastogenic signaling.

Interestingly, an alternative ITAM signaling, OSCAR/FcRγ signaling, compensates for osteoclastogenesis but not the actin ring formation of Siglec-15−/− cells when cultured on plastic wells. These results are consistent with previous findings that signaling via FcRγ compensates for defective osteoclastogenesis in DAP12−/− cells but not cytoskeletal disorganization of osteoclasts.[5] This supports the notion that the Siglec-15/DAP12 complex is responsible for both the cytoskeletal organization and differentiation of osteoclasts. Based on previous studies of FcRγ signaling,[3, 5, 25] we activated OSCAR/FcRγ signaling by osteoblasts and collagens. Osteoblasts and type II collagen have compensatory effects on osteoclastogenesis of Siglec-15−/− cells, whereas type I collagen has little effect. Given that type II collagen, but not type I collagen, has 6 prominent OSCAR-binding motifs,[25] this finding seems reasonable. Of note, Siglec-15−/− osteoclasts recover cytoskeletal disorganization as well as impaired differentiation when cultured on type II collagen-coated bone slices, suggesting that some factors in the bone matrix activate the signaling pathway related to cytoskeletal organization. Consequently, bone matrix proteins including type II collagen may compensate for Siglec-15–mediated signaling at the primary spongiosa of Siglec-15−/− mice.

It is worth mentioning that Siglec-15 localizes predominantly on the plasma membrane of migrating polykaryons, while it is distributed in the cytoplasm in polykaryons with actin ring formation. Although Siglec-15 is a type I transmembrane protein, it is localized inside macrophages/dendritic cells rather than on the plasma membrane.[14] Angata and colleagues speculated that Siglec-15 translocates to the cytoplasm based on the presence of a sequence motif (YENL) in Siglec-15, which conforms to the consensus endocytosis motif, YxxØ (where Ø stands for an amino acid residue with a bulky hydrophobic side chain).[14, 37] Given that endocytic activity and vesicular trafficking processes are essential for bone resorbing osteoclasts to establish and maintain cell polarization, Siglec-15 might be involved in a mechanism related to endocytosis.

The findings of the present study suggest a novel osteoclast activation mechanism that allows for the development of functional osteoclasts to be controlled by immunoreceptors/ITAM-bearing adaptors. Siglec-15 should recognize certain subsets of cells or certain matrix molecules responsible for osteoclast development by exploiting the structural diversity of sialylated glycans expressed on the cells or on the matrix molecule. Ligand occupancy of Siglec-15 somehow promotes differentiation and cytoskeletal organization of osteoclasts in association with DAP12. Given that myeloid DAP12-associating lectin-1, a C-type lectin, regulates osteoclast differentiation and function in association with DAP12,[38] glycan recognition may have an important role in developing functional osteoclasts.


All authors state that they have no conflicts of interest.


We thank Dr Toshio Kitamura for the plasmids and cell lines, and Dr Toshio Akazawa for the bovine bone slices. This project was supported in part by a Grant-in-Aid for Exploratory Research from the Ministry of Education, Culture, Sports, Science, and Technology of Japan 21659346 (NI) and a Grant-in-Aid for Research Activity Start-up from the Ministry of Education, Culture, Sports, Science, and Technology of Japan 23890008 (MT).

Authors' roles: Study design: MT and NI. Study conduct: YK, MKo, SM, TS, and MK. Data collection: YK, TS, and MKo. Data analysis: MT and KY. Data interpretation: MT, YK, SH, AM, and NI. TA provided essential reagents. Drafting manuscript: MT and YK. Revising manuscript content: MT and YK. Approving final version of manuscript: MT, YK, MK, SM,SH, TS, TA, MK, AM, and NI. MT takes responsibility for the integrity of the data analysis.