By homology to triggering receptor expressed by myeloid cells-2, we screened the mouse expressed sequence tag database and isolated a new single Ig domain receptor, which we have expressed and characterized. The receptor is most similar in sequence to the human CMRF-35 receptor, and thus we have named it CMRF-35-like molecule (CLM)-1. By screening the mouse genome, we determined that CLM-1 was part of a multigene family located on a small segment of mouse chromosome 11. Each contains a single Ig domain, and they are expressed mainly in cells of the myeloid lineage. CLM-1 contains multiple cytoplasmic tyrosine residues, including two that lie in consensus immunoreceptor tyrosine-based inhibitory motifs, and we demonstrate that CLM-1 can associate with Src-homology 2 containing phosphatase-1. Expression of CLM-1 mRNA is down-regulated by treatment with receptor activator of NF-κB ligand (RANKL), a cytokine that drives osteoclast formation. Furthermore, expression of CLM-1 in the osteoclastogenic cell line RAW (RAW.CLM-1) prevents osteoclastogenesis induced by RANKL and TGF-β. RAW.CLM-1 cells fail to multinucleate and do not up-regulate calcitonin receptor, but they express tartrate-resistant acid phosphatase, cathepsin K, and β3 integrin, suggesting that osteoclastogenesis is blocked at a late-intermediate stage. Thus, we define a new family of myeloid receptors, and demonstrate that the first member of this family, CLM-1, is an inhibitory receptor, able to block osteoclastogenesis.
Cells of the myeloid lineage are the major cellular component of the innate immune system. They lack a highly diverse Ag receptor and yet are able to specifically recognize and destroy a wide range of microbes. Additionally, they orchestrate the adaptive immune response both directly via Ag presentation and indirectly via secretion of a wide variety of chemokines and cytokines. The requirement for control of myeloid cell function is demonstrated by the central role played by myeloid cells in inflammatory pathologies such as endotoxemia, asthma, and osteoporosis. Their activity is in part regulated by families of Ig-like receptors, including the human Ig-like transcript receptors (1), the mouse paired Ig receptors (orthologues of human Ig-like transcripts) (2), signal-regulatory proteins (SIRPs)4 (3, 4), and triggering receptors expressed by myeloid cells (TREMs) (5, 6, 7). These receptor families each contain both inhibitory and activating members: inhibitory members recruit Src homology 2-containing phosphatases (SHPs) via immunoreceptor tyrosine-based inhibitory motif (ITIM) motifs in their cytoplasmic domains; activating members associate with immunoreceptor tyrosine-based activation motif (ITAM)-containing adaptor molecules via a positively charged residue in their transmembrane domains (8).
In addition to their role in the immune system, myeloid cells also play a central role in tissue remodeling during embryonic development and in adult tissue. In bone, a specialized myeloid cell type, the osteoclast, is essential for bone resorption, which occurs during bone growth and repair as well as during normal bone turnover. Osteoclasts are multinucleated giant cells that form from myeloid precursors under the action of receptor activator of NF-κB ligand (RANKL) and M-CSF (9, 10, 11, 12, 13). They possess specialized machinery for the destruction of both mineral and organic bone matrix (14). Osteoporosis results from a pathological overactivity of osteoclasts, leading to excessive bone resorption. A relatively small increase in bone resorption can, over time, result in osteoporosis, while decreased osteoclast activity results in osteopetrosis. Thus, there is a requirement for tight control over osteoclast differentiation and activation (12, 15).
In this work, we describe the cloning of a new murine Ig-like receptor, CMRF-35-like molecule-1 (CLM-1), that shows homology to the human CMRF-35 receptor. We demonstrate that CLM-1 is part of a multigene family on mouse chromosome 11. The expression of the CLM receptors is essentially restricted to myeloid cell lines. Unlike other myeloid receptor families that possess only ITIM-containing inhibitory receptors and adaptor-associated activating receptors, the CLM family demonstrates a diversity of transmembrane and cytoplasmic domains, indicating that a number of distinct signaling mechanisms are used by different members of the CLM family. The cytoplasmic domain of CLM-1 contains ITIMs. In accordance with this, we show that CLM-1 associates with SHP-1, which mediates the inhibitory function of ITIM-containing receptors. RT-PCR analysis demonstrates that CLM-1 mRNA is decreased in the osteoclastogenic cell line RAW following treatment with the cytokine RANKL. Furthermore, expression of CLM-1 in RAW prevents the formation of multinucleated osteoclasts in vitro following stimulation of the cells with RANKL and TGF-β. Thus, CLM-1 is a novel inhibitory receptor expressed by myeloid cells that can control osteoclast differentiation.
Materials and Methods
Reagents and cell lines
The M2 anti-FLAG mAb was purchased from Sigma-Aldrich (St. Louis, MO). The MT2 mouse macrophage cell line was provided by M. McKichan (University of California, San Francisco, CA). All other cell lines are from the American Type Culture Collection (Manassas, VA). Cell lines other than MT2 were grown in RPMI 1640 supplemented with 10% heat-inactivated FBS, 25 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. MT2 cells were grown in IMDM supplemented with 10% heat-inactivated FBS, 25 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. RAW 264.7 (RAW) cells were transfected with expression vectors encoding FLAG-tagged CLM-1 or FLAG-tagged Ly-49A under the control of the SRα promoter, by using Fugene (Roche, Basel, Switzerland), according to the manufacturer’s instructions. Stable transfectants were selected in complete RPMI containing G418.
Isolation of CLM cDNA and genomic mapping
GenBank-expressed sequence-tagged database (dbEST) was searched with the amino acid sequence of the Ig domain of TREM-2, using the tblastn algorithm. An EST corresponding to CLM-1 was identified, and this construct was obtained from the American Type Culture Collection and confirmed by sequencing. For genomic localization, the Celera Discovery System mouse genome database and the National Center for Biotechnology Information mouse genome were searched with the cDNA sequence of CLM-1. Predicted CLM family gene sequences were used to screen dbEST, and further ESTs were identified, obtained from American Type Culture Collection, and confirmed by sequencing. Sequences for CLM-1–9 have been submitted to GenBank (accession nos. AY457047–AY457055).
RNA was prepared from a variety of cell lines, including macrophage, T cell, B cell, and nonhemopoietic cell lines, using RNAqueous (Ambion, Austin, TX), according to the manufacturer’s instructions. Total RNA (20 μg) was resolved on denaturing agarose gels and transferred to nylon membrane using standard procedures. CLM-1 DNA was labeled with Strip-EZ PCR kit (Ambion) and [32P]dATP. Membranes were incubated overnight at 42°C with probe in Ultra-hyb hybridization solution (Ambion), and were then washed in 0.25× SSC, 0.1% SDS at 42°C before exposure to x-ray film.
Pervanadate treatment, immunoprecipitation, and Western blotting
Twenty million cells were stimulated with pervanadate by resuspension in serum-free RPMI 1640 medium supplemented with 10 mM sodium vanadate and 0.6% hydrogen peroxide (pervanadate) and incubation for 10 min at 37°C. Cells were pelleted and rapidly lysed by suspension in ice-cold lysis buffer (1% digitonin, 0.12% Triton X-100, 150 mM NaCl, 20 mM triethanolamine, pH 7.8, 2.5 mM CaCl2, 1 mM MgSO4). After 2-h incubation at 4°C, the insoluble fraction was removed by centrifugation at 25,000 × g for 20 min at 4°C. Supernatants were incubated with 10 μg of M2 anti-FLAG Ab. Immune complexes were recovered by incubation with 10 μl of 10% protein G-Sepharose for 60 min at 4°C, followed by centrifugation at 500 × g for 3 min at 4°C. Pellets were washed four times in 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS) buffer (5 mM CHAPS, 150 mM NaCl, 50 mM Tris, pH 8.0). The proteins were solubilized in 2× loading buffer and were boiled for 5 min before resolution by SDS-PAGE using 12% polyacrylamide gels. Samples were transferred to polyvinylidene difluoride membrane (Millipore, Bedford, MA) using a semidry blotting apparatus. Membranes were blocked for 1 h in TBS containing 0.1% Tween 20 and 2% BSA, and then incubated for 2 h in the same solution containing either 4G10 anti-phosphotyrosine Ab or Ab to SHP-1. Membranes were incubated for a further 1 h with HRP-conjugated rabbit anti-mouse or donkey anti-rabbit Ab, washed extensively with TBS-Tween, and developed with Supersignal chemiluminescent substrate (Pierce, Rockford, IL) before exposure to film.
RNA isolation and RT-PCR
Results. Briefly, 0.4 ml of reagent was added to the well and then vigorously pipetted. The solution was transferred to an Eppendorf tube, and 0.2 ml chloroform/ml TRIzol reagent was added and mixed, after which the phases were separated by centrifugation for 15 min at 12,000 × g at 4°C. The supernatant was removed, and an equal volume of isopropanol was added. RNA was precipitated by centrifugation at 12,000 × g at 4°C for 10 min. The RNA pellet was washed with 1 ml of 75% ethanol and briefly centrifuged at 7,500 × g at 4°C for 5 min. The pellet was air dried and resuspended in 1 mM EDTA. Five micrograms of the RNA were used for the reverse-transcription reaction. For this, 1 μg of oligo(dT) (Amersham Pharmacia Biotech, Piscataway, NJ), 5 μg of RNA, and H2O up to 20 μl were briefly heated to 95°C, then equilibrated at 42–55°C. Thirty microliters of RT-Mix (1 μl RNase inhibitor (Roche), 10 μl 5× reverse-transcriptase buffer (Roche), 2 μl 10 mM dNTPs (Roche), 1 μl avian myeloblastosis virus-reverse transcriptase (Roche), and 16 μl H2O) were added, and the reaction was incubated at 42°C for 1 h. RNA was denatured by adding 0.5 μl 0.5 M EDTA and 2 μl 1 M NaOH, and boiling for 3 min. Two microliters of 1 M HCl, 5.5 μl 3 M NaOAc, and 145 μl 100% ethanol were then added, and the cDNA was precipitated by centrifugation for 10 min at 15,000 × g. The pellet was washed with 70% ethanol, air dried, and resuspended in 50 μl 10 mM Tris, pH 8.0, 1 mM EDTA. Two microliters of cDNA were used per 30 μl PCR with 3 μl 10× TAQ polymerase buffer (Roche), 0.5 μl TAQ polymerase (Roche), and 1 μl 10 mM dNTPs (Roche). A PerkinElmer (Wellesley, MA) GeneAmp Cycler was used with the following program: 10 min at 94°C, 30 s at 50°C, 30 s at 72°C, 30 s at 94°C, for 30 cycles. Primers used were: GAPDH, 5′-ACCACAGTCCATGCCATCAC, 3′-TCCACCACCCTGTTGCTGTA; RANK, 5′-AAGATGGTTCCAGAAGACGGT, 3′-CATAGAGTCAGTTCTGCTCGGA; calcitonin receptor, 5′-ACCGACGAGCAACGCCTACGC, 3′-GCCTTCACAGCCTTCAGGTAC; integrin β3, 5′-CTGGTAAAAACGCCGTGAAT, 3′-CGGTCATGAATGGTGATGAG; cathepsin K, 5′-ACGGAGGCATCGACTCTGAA, 3′-GATGCCAAGCTTGCGTCGAT; OSCAR, 5′-TCATCTGCTTGGGCATCATA, 3′-ACAAGCCTGACAGTGTGGTG; CLM-1, 5′-TCCAAGTACCCATTACAGTGCC, 3′-CAGGACTGCAGAGATGACTGG; CLM-2, 5′-CTGACGGGCCCTGGCTCTG, 3′-CTAAAGTCTGTTCACCCAG; CLM-4, 5′-TGTGTCCCACTGCATGGC, 3′-TCACTGGTTCTCATAACAG; CLM-5, 5′-CAGGATTCAGTCACAGGTC, 3′-CTAGGCAACAGGACTATG; CLM-8, 5′-CTGCATGGTCCCAGCACC,3′-TCACAGGTAAAGGTCAGAG; CLM-9, 5′-CTGAAGGGTCCAAAGAG, 3′-TTACACAGAGATGAACTC.
Osteoclast differentiation and TRAP staining
2 to mouse IgG (ICN Pharmaceuticals, Costa Mesa, CA) was added as a cross-linker to a final concentration of 10 μg/ml. After 5–7 days in culture, cells were fixed with 10% formaldehyde in PBS for 10 min. Plates were then washed twice in PBS, incubated for 30 s in 50% acetone/50% ethanol, and washed with PBS. Cells were stained for tartrate-resistant acid phosphatase (TRAP) by using a kit from Sigma-Aldrich, following the manufacturer’s instructions. Multinucleated (>2 nuclei) TRAP-positive cells were then counted by light microscopy.
Osteoclast resorption assay
A total of 1 × 104 RAW cells was plated on dentine discs (IDS, Boldon, U.K.) and treated with RANKL and TGF-β, as above. Medium was changed every 3 days. Dentine discs were fixed with 3.7% formaldehyde for 30 min, and cells were removed with a tissue before staining with 1% toluidine blue in 0.5% sodium tetraborate solution. Resorption pits were analyzed by light microscopy.
Real-time quantitative PCR
RNA was prepared as described above. Totals of 500, 250, 125, or 62.5 ng of RNA was used for reverse transcription and real-time PCR performed on each RT preparation in order to ensure linearity of the RT reaction. The RT reaction was performed in a final volume of 100 μl with a final concentration of: 1 × PCR buffer, 7.5 mM MgCl2162, 0.2-0.9 μM each primer, 200 μM each dNTP, 100–200 nM probe and 0.025 U/μl Tag Gold (ABI, Foster City, CA). PCR was conducted using the following cycle parameters: 95°C 12 min for 1 cycle, (95°C 20 sec, 60°C 1 min), for 40 cycles. Analysis was carried out using the sequence detection software supplied with the ABI 7700. The software calculates the threshold cycle (Ct) for each reaction and this was used to quantitate the amount of starting template in the reaction. The Ct values for each set of three reactions were averaged for all subsequent calculations. A difference in Ct values (ΔCt) was calculated for each gene by taking the mean Ct of gene of interest and subtracting the mean Ct for a control gene for each cDNA sample. Relative expression levels were expressed as a percentage and calculated as 100 × (2−ΔCt).
Mouse CLM-1 is a single Ig domain transmembrane protein
We have previously described the cloning of mouse TREM-2 and TREM-3, which are single Ig domain, DAP12-associated activating receptors expressed by myeloid cells. Using the Ig domain of TREM-2, we searched GenBank mouse EST database to find cDNAs for homologous proteins. A novel mouse EST was identified, which showed homology to mouse TREM-2 (21% identical), but which was more closely related to the human CMRF-35 receptors (31% identical) (17, 18, 19, 20). We therefore named this protein CLM-1. The amino acid sequence of CLM-1 shows conservation of the Ig domain cysteines and the adjoining consensus residues, as well as two internal Ig domain cysteine residues that are shared by TREM-2, NKp44, and CMRF-35 receptors. The cytoplasmic domain of CLM-1 contains multiple tyrosine residues, including two that fall into consensus ITIM motifs, a third that lies in a YxxM consensus motif that has been associated with activation signaling (via association with phosphatidylinositol 3′-kinase), and a fourth that lies in a consensus motif for SLAM-associated protein binding (TxYxxI) (Fig. 1⇓A). Thus, CLM-1 could potentially perform multiple functions in the cell depending on which of these residues is phosphorylated.
CLM-1 is part of a single Ig domain gene complex
By using the Celera Discovery System mouse genome database and the National Center for Biotechnology Information mouse genome, we localized the gene for CLM-1 to chromosome 11. Further searching with CLM-1 sequence revealed eight putative genes with homologous extracellular domains in close proximity to CLM-1 (Fig. 1⇑B). CLM-9 lies somewhat distal to the other receptors, but it again contains a single Ig domain and shows homology to CMRF-35, justifying its inclusion in the CLM family. By using the predicted gene sequences, we isolated ESTs corresponding to several of these CLM family members and performed full-length sequencing. Remarkably, the signaling domains of these receptors showed great diversity (Fig. 1⇑C). Thus, CLM-2, CLM-4, and CLM-6 all possess a transmembrane lysine residue, which could permit association with adaptor proteins such as DAP12. Indeed, CLM-4 corresponds to a recently characterized receptor, DigR1, which associates with DAP12 (21), and we have confirmed that association (data not shown). CLM-5 appears to be orthologous to the human CMRF-35A receptor (19). Like CMRF-35A, CLM-5 contains a short cytoplasmic domain and a negatively charged glutamate residue in its transmembrane domain, which has unknown function, but which may allow association with an adaptor molecule. CLM-8 corresponds to the human CMRF-35H/IRp60 receptor (20). CMRF-35H contains both consensus ITIM and ITAM motifs. In CLM-8, the tyrosine residues from these motifs have been conserved, but important adjacent residues have been altered, making it difficult to predict whether CLM-8 will provide inhibitory or activating functions or both. CLM-9 contains a single tyrosine residue that does not lie in a consensus motif for binding by any common signaling protein, so again its function is unknown (Fig. 1⇑C). Thus, the CLM receptor family is not simply divided into those expressing ITIMs and those binding DAP12. Instead, it appears to include receptors with diverse mechanisms for cell signaling.
CLM receptors are expressed in myeloid cells
To examine the expression of CLM-1, we performed Northern blot analysis using a panel of cell lines (Fig. 2⇓A). CLM-1 expression could be detected in the monocyte cell lines P388D1, MT2, and RBL-5, and in the dendritic cell line DC2.4. Expression could also be detected to a lesser extent in the monocyte line RAW, although the major mRNA species expressed was distinct from those expressed by the other cell lines (Fig. 2⇓A). Expression could not be detected in any of the other cell lines analyzed. Expression of CLM-1 in these lines was confirmed by RT-PCR analysis (Fig. 2⇓B). This analysis also revealed expression of CLM-1 in the B cell lymphoma A20, and indeed, extended exposure of the Northern blot revealed a faint band in this lane (data not shown). We then used RT-PCR analysis to examine expression of the other CLM receptors in this series of hemopoietic cell lines (Fig. 2⇓B). The dendritic cell line DC2.4 expressed all of the CLM family members examined, while other myeloid cell lines showed different patterns of CLM expression. At least one member of the CLM family was expressed by every myeloid cell examined, including the mastocytoma line P815, which expressed the DAP12-associated CLM-4. No expression of CLMs could be found in the T cell lymphomas EL4 and YAC-1, nor in the NKT cell lymphoma C1498 (22). Low level expression could be detected in the B cell lymphoma A20 (CLM-8), and in the T cell lymphoma BW5147 (CLM-2), suggesting that CLM family receptors may also be expressed in lymphocyte subsets (Fig. 2⇓B).
CLM-1 is a monomeric cell surface protein that associates with SHP-1
To assess the function of CLM-1 in myeloid cells, we transfected a FLAG-tagged CLM-1 construct into the RAW monocyte cell line to create the RAW.CLM-1 line (Fig. 3⇓A). Immunoprecipitation of FLAG-tagged CLM-1 from RAW.CLM-1 cells and analysis under reducing and nonreducing conditions demonstrated that mature CLM-1 is a monomeric protein that runs at a molecular mass of ∼60 kDa under both reducing and nonreducing conditions (Fig. 3⇓B). To analyze associated proteins, cells were treated with pervanadate, and CLM-1 was again immunoprecipitated. Even in the absence of pervanadate treatment, there was low level phosphorylation of CLM-1 (Fig. 3⇓C, upper panel). Following pervanadate treatment, CLM-1 became hyperphosphorylated, and a clear shift in migration was observed (Fig. 3⇓C, upper panel). Interestingly, the pattern of CLM-1-associated tyrosine-phosphorylated proteins also changed upon hyperphosphorylation of the receptor; following pervanadate treatment, the ∼220-kDa protein coprecipitated from unstimulated cells was replaced by an ∼190-kDa protein (Fig. 3⇓C, upper panel). The identities of these proteins remain to be determined. Blotting with anti-SHP-1 Ab revealed the association of SHP-1 with CLM-1 following pervanadate treatment (Fig. 3⇓C, lower panel). We could not detect association of SHP-2, Src-homology 2 containing inositol 5′ phosphatase (SHIP), or phosphatidylinositol 3′-kinase p85 subunit with CLM-1 following pervanadate treatment (data not shown).
CLM-1 inhibits osteoclast formation
Despite the association of CLM-1 with SHP-1, we were unable to find an inhibitory effect of CLM-1 ligation by using standard assays for SHP-1-mediated inhibition, including redirected inhibition of lysis by an NK cell line or FcεRI-mediated serotonin release from RBL-2H3 mast cells following transfection of both lines with CLM-1 (data not shown). However, by using semiquantitative PCR to analyze the regulation of CLM-1, we noted that CLM-1 was significantly down-regulated in RAW cells following only 24-h treatment with the osteoclastogenic cytokine RANKL (Fig. 4⇓A). Other CLMs were not affected by RANKL (data not shown). Although TGF-β inhibits osteoclastogenesis in the presence of stromal/osteoblast cells, using purified preosteoclast populations in vitro TGF-β enhances osteoclast formation in the presence of RANKL (23), and TGF-β did not prevent down-regulation of CLM-1 transcripts by RANKL (Fig. 4⇓A). Additionally, treatment of C57BL/6 bone marrow macrophages with M-CSF and RANKL greatly reduced levels of CLM-1, as detected by RT-PCR (Fig. 4⇓B), suggesting that down-regulation of CLM-1 is a general effect of RANKL during osteoclast development and is not specific to the RAW cell line. To strengthen this observation, we prepared cDNA from C57BL/6 bone marrow cells treated for 5 days with M-CSF (Mθ) or with M-CSF and RANKL (OC). We used this cDNA for real-time PCR analysis of CLM-1 expression, using GAPDH or L19 ribosomal protein as reference cDNA. Regardless of the reference cDNA used, the real-time PCR confirmed that CLM-1 was down-regulated upon differentiation of myeloid cells into osteoclasts (Fig. 4⇓C). We, therefore, wondered whether CLM-1 could have an inhibitory role in osteoclast formation. RAW cells are known to form multinucleated osteoclasts in response to RANKL and TGF-β (23). In accordance with this, RAW cells treated for 5 days with RANKL and TGF-β formed multinuclear osteoclasts that stained positively for TRAP (Fig. 5⇓, A and B) and formed resorption pits in dentine slices (Fig. 5⇓C). RAW cells transfected with the NK-inhibitory receptor Ly-49A also showed robust formation of multinucleated osteoclasts in response to RANKL and TGF-β (Fig. 5⇓A). Remarkably, however, RAW.CLM-1 cells were markedly impaired in their ability to form multinuclear osteoclasts following this same treatment (Fig. 5⇓, A and B). Although these mononucleated RAW.CLM-1 cells were TRAP positive (Fig. 5⇓B), they were unable to resorb bone (Fig. 5⇓C). This was not a clone-specific effect, because all RAW.CLM-1 clones examined displayed this same phenotype (data for two independent clones shown in Fig. 5⇓A). Cross-linking of CLM-1 with anti-FLAG Ab did not significantly enhance this inhibitory effect (data not shown). To further elucidate the effect of CLM-1 on osteoclast formation, we examined the expression of various osteoclast markers in RAW and RAW.CLM-1 cells following 5 days of treatment with RANKL and TGF-β. As shown in Fig. 6⇓, for RAW cells this treatment led to increased transcripts for the early osteoclast markers cathepsin K and OSCAR, as well as the late osteoclast markers calcitonin receptor and integrin β3. Similar treatment of RAW.CLM-1 cells led to increased transcripts for cathepsin K (24), OSCAR (25), and integrin β3 (26, 27), but it failed to up-regulate expression of the late osteoclast marker calcitonin receptor (28) (Fig. 6⇓). Thus, CLM-1 expression allows expression of some early osteoclast markers, but it prevents the expression of the late osteoclast marker, calcitonin receptor; prevents multinucleation; and prevents acquisition of bone-resorptive ability. Thus, it appears to block osteoclast progression at an intermediate to late stage of differentiation.
We describe in this study a new family of receptors expressed by mouse myeloid cells, the CLM family. The CLM family appears to be orthologous to the human CMRF-35 family on human chromosome 17, of which two members have been identified, CMRF-35A (19) and CMRF-35H (20), together with a pseudogene, CMRF-35J (18). CMRF-35A contains a transmembrane glutamic acid residue, and has unknown function, while CMRF-35H/IRp60 contains both ITIM and ITAM motifs, and has been shown to associate with SHP-1 and SHP-2 (29). We have studied one member of the mouse CLM family, CLM-1, which is not orthologous to either CMRF-35A or CMRF-35H. CLM-1 contains ITIM motifs, and we show that it associates with SHP-1. Furthermore, we demonstrate that transcripts for CLM-1 are down-regulated by the osteoclastogenic cytokine RANKL in both the osteoclastogenic monocyte cell line, RAW, and C57BL/6 bone marrow cells. Furthermore, we show that expression of CLM-1 in RAW cells effectively inhibits in vitro osteoclast formation.
The members of the CLM receptor family each contain a single extracellular Ig domain. We have previously described the existence of a single Ig domain myeloid receptor gene complex, the TREM complex, on mouse chromosome 17 (6). Like TREMs, the CLM receptors are preferentially expressed on myeloid cells, but they lie in a separate gene complex on mouse chromosome 11. Among the different cell lines examined, each demonstrated a different pattern of CLM expression. It is tempting to speculate that the differential expression of CLM receptors by myeloid cell lines in vitro reflects heterogenous expression by myeloid cells in vivo. Such a situation would allow functional diversity among subsets of myeloid cells in a manner analogous to the well-established functional heterogeneity imparted by differential killer Ig-like receptor expression on human NK cells (30), or Ly-49 expression on mouse NK cells (31).
Both SHIP-deficient mice (32) and SHP-1-deficient motheaten mice (33, 34) show increased numbers of osteoclasts, and these osteoclasts have increased activity, suggesting that both phosphatases play a role in the control of osteoclast differentiation and activation. In this study, we show that CLM-1 can inhibit the formation of osteoclasts, and that in cells treated with pervanadate it associates with SHP-1. In the absence of pervanadate treatment, we were unable to detect SHP-1 binding, even though CLM-1 was clearly phosphorylated (Fig. 3⇑C). Although we were unable to detect association of SHIP with CLM-1, this may reflect a lack of sensitivity of the reagents we used. It is also possible that SHIP or other molecules do associate with CLM-1 under certain conditions. It is unlikely that under normal physiological conditions all tyrosine residues would be phosphorylated, leaving the possibility for binding of different molecules in different phosphorylation states. Indeed, CLM-1-associated phosphotyrosine proteins could be detected that disappeared upon pervanadate treatment. We therefore believe that CLM-1 may play a number of different roles in the physiology of myeloid cells, depending upon the stimulus and selective phosphorylation of cytoplasmic tyrosine residues.
Expression of CLM-1 in RAW cells was able to inhibit osteoclastogenesis without exogenous receptor stimulation. However, immunoprecipitation of CLM-1 from RAW.CLM-1 cells demonstrated that CLM-1 was normally phosphorylated to some degree in these cells. The ligand for CLM-1 is currently unknown, but it is possible that RAW cells also express this ligand either as a cell surface protein, or as a secreted autocrine factor. Alternatively, it is possible that CLM-1 becomes phosphorylated indirectly following activation of unrelated receptors. For instance, SIRP-α becomes phosphorylated after treating cells with either epidermal growth factor or insulin despite the fact that these mitogens do not bind to SIRP-α (3).
Much progress has been made in recent years in delineating the progression of osteoclast differentiation. It is known that both M-CSF and RANKL are essential for the promotion of osteoclastogenesis (9, 10) (RAW cells are M-CSF independent and thus form osteoclasts in response to RANKL alone). Treatment of bone marrow cells with M-CSF and RANKL leads to expression of early osteoclast markers TRAP (35), cathepsin K (24), and OSCAR (25), then the late markers αvβ3 integrin (26, 27) and calcitonin receptor (28), and eventually to cell fusion and acquisition of the multinuclear phenotype. Expression of CLM-1 in the osteoclastogenic line RAW did not prevent expression of early osteoclast markers, but it did prevent expression of the calcitonin receptor, and it severely inhibited multinucleation and bone resorption. Thus, CLM-1 arrests osteoclast differentiation at a late-intermediate stage of osteoclastogenesis.
The process of monocyte-monocyte fusion, which is required for the multinucleation of osteoclasts, is incompletely understood. Recently, however, two receptors, SIRPα (36) and CD44 (37), have been implicated in the fusion process. Indeed, one of the cloning strategies used to identify SIRPα used Abs that blocked monocyte-monocyte fusion (36). Furthermore, a fusion protein containing the extracellular IgV domain of SIRPα is able to block monocyte fusion (36, 38). Interestingly, SIRPα associates with both SHP-1 and SHP-2 (3, 39, 40). Thus, although SIRPα and CLM-1 each associate with SHP-1, they appear to have opposing effects on osteoclastogenesis. In addition to SHP-1, SIRPα is known to bind Pyk-2, as well as SKAP55hom, through which it associates with Fyb (SLAP130) (41). Thus, SIRPα has additional pathways through which it may influence macrophage adhesion (and fusion). Similarly, CLM-1 has additional cytoplasmic motifs, not found in SIRPα, that may modulate the signals provided by this receptor during osteoclast differentiation. Alternatively, distinct cellular locations may explain the opposing effects of SIRPα and CLM-1 on monocyte fusion.
As mentioned previously, osteoclastogenesis is driven by a combination of M-CSF and RANKL. However, both of these cytokines are present throughout the body, which requires an explanation for the bone-restricted localization of osteoclasts. One possibility is that a bone-restricted receptor such as OSCAR is required to impart an essential cosignal for osteoclastogenesis (25). An alternative possibility would be that an inhibitory receptor is normally expressed by macrophages outside the bone environment, preventing inappropriate development of osteoclasts. Our data demonstrate that CLM-1 would fit the requirements for such a receptor. Expression of CLM-1 in RAW cells is able to inhibit in vitro osteoclastogenesis. Furthermore, the osteoclastogenic cytokine RANKL reduced expression of CLM-1 in both RAW cells and C57BL/6 bone marrow macrophages. Thus, down-regulation of CLM-1 may be an important step on the osteoclast differentiation pathway.
↵1 This work was supported by the Veterans Administration and by Grant R01 CA87922-01A1 (to W.E.S.) from the National Institutes of Health.
↵2 D.-H.C. and M.B.H. contributed equally to this work.
↵3 Address correspondence and reprint requests to Dr. Michael R. Daws, Department of Immunology Number 111R, Veterans Administration Medical Center, 4150 Clement Street, San Francisco, CA 94121. E-mail address:
↵4 Abbreviations used in this paper: SIRP, signal-regulatory protein; CLM, CMRF-35-like molecule; EST, expressed sequence tag; SHP, Src-homology 2-containing phosphatase; SHIP, Src-homology 2 containing inositol 5′ phosphatase; ITAM, immunoreceptor tyrosine-based activation motif; ITIM, immunoreceptor tyrosine-based inhibitory motif; RANKL, receptor activator of NF-κB ligand; TRAP, tartrate-resistant acid phosphatase; TREM, triggering receptor expressed by myeloid cells.
- Received November 25, 2002.
- Accepted October 15, 2003.
- Copyright © 2003 by The American Association of Immunologists