Dr Roodman worked as a collaborator with MedImmune. All other authors have no conflict of interest
Screening a cDNA library enriched for genes expressed in OCLs identified ECF-L. ECF-L enhanced OCL formation without increasing RANKL levels. Anti-ECF-L inhibited RANKL-induced OCL formation. These results support a potent role of ECF-L in osteoclastogenesis.
Introduction: To investigate the molecular mechanisms that control osteoclastogenesis, we developed an immortalized osteoclast (OCL) precursor cell line that forms mature OCLs in the absence of stromal cells and used it to form pure populations of OCLs.
Materials and Methods: Polymerase chain reaction (PCR) selective cDNA subtraction was used to identify genes that are highly expressed in mature OCLs compared with OCL precursors employing OCL and OCL precursors derived from this cell line.
Results: Eosinophil chemotactic factor-L (ECF-L), a previously described chemotactic factor for eosinophils, was one of the genes identified. Conditioned media from 293 cells transfected with mECF-L cDNA, or purified ECF-L Fc protein, increased OCL formation in a dose-dependent manner in mouse bone marrow cultures treated with 10−10 M 1,25(OH)2D3. OCLs derived from marrow cultures treated with ECF-L conditioned media formed increased pit numbers and resorption area per dentin slice compared with OCLs induced by 1,25(OH)2D3 (p < 0.01). Addition of an antisense S-oligonucleotide to mECF-L inhibited OCL formation in murine bone marrow cultures treated only with 10−9 M 1,25(OH)2D3 compared with the sense S-oligonucleotide control. Time course studies demonstrated that ECF-L acted at the later stages of OCL formation, and chemotactic assays showed that mECF-L increased migration of OCL precursors. mECF-L mRNA was detectable in mononuclear and multinucleated cells by in situ hybridization. Interestingly, a neutralizing antibody to ECF-L blocked RANKL or 10−9 M 1,25(OH)2D3-induced OCL formation in mouse bone marrow cultures, although ECF-L did not induce RANKL expression.
Conclusions: These data show ECF-L is a previously unknown factor that is a potent mediator of OCL formation, which acts at the later stages of OCL formation and enhances the effects of RANKL.
Osteoclasts (OCLs) are multinucleated giant cells that resorb bone and are derived from cells in the monocytic lineage.(1) A number of factors that control osteoclastogenesis have been reported, including soluble cytokines and membrane bound factors on stromal cells and osteoblasts (e.g., RANKL).(2) The differentiation of OCLs requires the presence of marrow stromal cells and osteoblasts, and cell-to-cell contact between osteoblasts and hematopoietic cells is necessary for inducing differentiation of OCLs.(3)
RANKL is a critical osteoclastogenic factor that is produced by osteoblasts and marrow stromal cells in response to several osteotropic factors such as 1,25-dihydroxyvitamin D3 [1,25(OH)2D3],(4) parathyroid hormone (PTH),(5) and interleukin-11 (IL-11).(6) RANKL binds to its cognate receptor, RANK, which is found on OCLs and their precursors, and induces OCL formation and prolongs OCL survival. However, the genetic events controlling OCL formation from mononuclear precursors have not been fully elucidated.
To identify genes that regulate OCL differentiation, we developed an OCL precursor cell line (B/T cells) from mice doubly transgenic for the Bcl-XL and Tag genes,(7) and used it to form homogeneous populations of OCL. Using this precursor cell line and OCL derived from these cells, polymerase chain reaction (PCR)-selective cDNA subtraction hybridization was performed to identify genes that were upregulated in OCLs compared with their precursors. Using this differential screening approach, we identified ADAM 8 (a disintegrin and metalloproteinase) as an OCL stimulatory factor, which can increase mouse OCL formation and bone resorption.(8). We now report the identification and characterization of a novel osteoclastogenic cytokine, eosinophil chemotactic factor-L (ECF-L), which is overexpressed in OCLs. ECF-L was originally identified as a chemoattractant factor produced by mouse splenocytes that enhances chemotaxis of eosinophils, and attracts not only eosinophils but also T-lymphocytes and bone marrow cells.(9) ECF-L is expressed in spleen, bone marrow, lung, and heart. However, the role of ECF-L in osteoclastogenesis was previously unknown.
Chemokines have been characterized based on their chemotactic activity on leukocytic cells with little attention focused on their capacity to affect other cellular functions. Chemokines function as key mediators promoting the recruitment, proliferation, and activation of vascular and immune cells. Most chemokines have four characteristic cysteines and have been classified as α, β, γ, and δ chemokines. In general, the α chemokine subfamily members, including IL-8, chemoattract and activate neutrophils and T-cells, but not monocytes.(9) Many of the β chemokines, such as macrophage inflammatory protein-1α (MIP-1α), MIP-1β, RANTES, and MCP-1, act as chemoattractants and activators of monocytes, but not neutrophils.(10–12) MIP-1α, MIP-1β, and RANTES also are chemoattractants for T-lymphocytes.(13) Furthermore, RANTES, and to lesser extent MIP-1α, can act as chemoattracts for eosinophils.(14) Recently, we reported that the chemokine MIP-1α is a potent osteoclastogenic factor that acts directly on OCL precursors and enhances OCL formation and bone resorption.(15,16) In this study, we report the effects of ECF-L on OCL formation and/or activation that were previously unknown.
MATERIALS AND METHODS
RANKL (Immunex, Seattle, WA, USA) and 1,25(OH)2D3 (Teijin, Tokyo, Japan) were generously provided for these experiments. Restriction enzymes, Taq polymerase, fetal calf serum (FCS), and tissue culture media were purchased from Life Technologies (Grand Island, NY, USA). All other chemicals were obtained from Sigma (St Louis, MO, USA). Chemotactic assay plates were purchased from Corning Costar (Cambridge, MA, USA).
PCR-selective subtraction screening of OCL precursors and mature OCLs
OCL precursors and OCLs were prepared from B/T cells as previously described in detail.(17) PCR-selective cDNA subtraction screening for genes that were differentially overexpressed in mature OCLs rather than OCL precursors was performed as previously described,(8) using a PCR-based subtraction kit (K1804–1; Clontech, Palo Alto, CA, USA). Eighteen bands, which were overexpressed in mature OCLs, were reamplified by secondary PCR. After subcloning of the PCR products into the TA vector (Invitrogen, Carlsbad, CA, USA), the DNA sequences were determined and compared with the DNA sequence database in the National Center for Biotechnical Information (NCBI) to identify them.
Construction of a full-length murine ECF-L cDNA
The full-length ECF-L cDNA was generated by PCR using the mouse OCL cDNA as a template and specific primers sets for mECF-L [5′-ACACCATGGCCAAGCTCATT-3′ (sense) and 5′-TGCAGAATGCGCTGTGGAAA-3′ (antisense)]. The PCR conditions were 94°C for 30 s, 60°C for 30 s, 72°C for 2 minutes, and 40 cycles. The PCR product was subcloned into the TA vector and sequenced. The cDNA was digested with EcoRI and cloned into the mammalian expression vector pcDNA3 (Invitrogen) and transfected into 293 cells (a generous gift from Dr SV Reddy, University of Pittsburgh) to express mECF-L.
Murine OCL-like multinucleated cell formation and bone resorption assays
Mouse bone marrow cultures were performed as previously described to assess the effect of ECF-L on OCL formation/activation.(18) Briefly, freshly isolated mouse bone marrow cells (106/ml in α-MEM containing 10% FCS/well in 48-well plates) were cultured for 6–9 days. At the end of culture period, the cells were fixed and stained for TRACP, using a TRACP staining kit (A-367; Sigma) to identify OCL-like multinucleated cells (MNCs). MNCs were counted with an inverted microscope. Conditioned media from 293 cells transfected with the mECF-L cDNA or purified mouse ECF-L protein were added to mouse marrow cultures. The mECF-L cDNA was transiently transfected into 293 cells (4 × 105) grown in individual 35-mm wells, using the calcium phosphate method with a kit from Stratagene (La Jolla, CA, USA) according to the manufacturer's protocol. Twelve hours after the start of the cDNA transfection process, the cells were washed with 3 ml of serum-free DMEM and fed with 1.5 ml of the same media. Conditioned media were collected after 48 h and added to mouse marrow cultures for the first 2 days, days 2–4, days 4–6, or for the entire 7-day culture period. The number of TRACP+ MNCs formed was determined as previously described.(18) For pit formation assays, murine bone marrow cells were cultured on sperm whale dentin slices in 48-well plates. After 8 days of culture, the dentin slices were fixed and stained for TRACP, and the number of TRACP+ MNCs and mononuclear cells were scored. The cells on the dentin slices were removed gently by rubbing the slices between the thumb and first finger, and the number of bone resorption pits and area resorbed were measured by image analysis techniques as previously described.(19)
Human OCL formation assays
Nonadherent human bone marrow mononuclear cells were obtained from normal donors as described previously(16) and tested for their capacity to form OCL-like MNCs in long-term marrow cultures. These studies were approved by the Institutional Review Board of the University of Pittsburgh and the General Clinical Research Center at the University of Pittsburgh. The human ECF-L EST clone (AI934102) was identified by a homology search with mouse ECF-L cDNA and purchased from ATCC (Manassas, VA, USA). DNA sequence analysis was performed to confirm the identity of the hECF-L cDNA, and the insert cDNA was subcloned into the pcDNA3 mammalian expression vector. Conditioned media from 293 cells transfected with the hECF-L cDNA were added to marrow cultures weekly. At the end of the 3-week culture period, the number of MNCs that crossreacted with the 23c6 monoclonal antibody was determined. The 23c6 monoclonal antibody identifies OCL-like cells that express calcitonin receptors and resorb bone.(20)
Effects of antisense and sense oligonucleotides to mECF-L on OCL formation
To determine if native ECF-L was involved in OCL formation, we designed antisense (AS) and sense (SS) S-oligonucleotides (5′-AAGAATGAGCTTGGCCATGGTGTCTTCACG-3′ and 5′-CGTGAAGACACCATGGCCAAGCTCATTCTT-3′) that included the ATG and ribosome binding site of the mECF-L gene. The AS and SS oligonucleotides were added at varying concentrations to mouse bone marrow cultures stimulated with 10−9 M 1,25(OH)2D3. Every 2 days, one-half of the media was replaced with fresh media containing the oligonucleotide and 10−9 M 1,25(OH)2D3. At the end of the culture period, the cells were fixed and stained for TRACP activity, and the number of TRACP+ MNCs and mononuclear cells was determined.
In situ hybridization
In situ hybridization was performed according to the method of Nomura et al.(21) Digoxigenin (DIG)-labeled single-strand antisense and sense cRNA probes to mouse ECF-L were prepared using a DIG RNA labeling kit (Roche Diagnostics, Mannheim, Germany). Freshly isolated mouse bone marrow cells were cultured in Lab-Tek 4-chamber slides (Nalge Nunc International, Naperville, IL, USA) in the presence of 10−9 M of 1,25(OH)2D3. After 9 days of culture, the cells were fixed with 4% paraformaldehyde, rehydrated, and incubated with 2 μg/ml of proteinase K for 2 minutes at 37°C. The cells were then treated with 0.2 M HCl for 10 minutes at room temperature to minimize the nonspecific signals through quenching the intrinsic alkaline phosphatase activities. The slides were dehydrated with ethanol, and then hybridized with hybridization solution containing either a sense or antisense cRNA probe. The slides were washed with 2× SSC and then 0.2× SSC for 15 minutes at 50°C. The hybridization signals were detected with a DIG nucleic acid detection kit (Roche Diagnostics, Mannheim, Germany). The sense cRNA probe was used as a control. The slides were counterstained with 0.5% methyl green and imaged.
Expression and purification of mECF-L in E. coli
Recombinant mECF-L (rmECF-L) was expressed in the BL21 E. coli strain using the pET14b expression vector system (Novagen, Inc., Madison, WI, USA) according to the manufacturer's protocol. The nucleotide sequence encoding the mECF-L cDNA was amplified by PCR with sense primers (5′-CGAGGATCCGATGGCCAAGCTCATTCTTGTC-3′) and antisense primers (5′-CGAGGATCCTCAATAAGGGCCCTTGCAACT-3′) (underlined sequences represent BamHI site.) The PCR product was digested with BamHI site and then cloned into the pET 14b vector in frame with the 6xHis tag. The plasmid construct was transformed into the BL21 (DE3) E. coli, and the recombinant ECF-L was induced by treatment with 0.5 mM IPTG for 4 h. The cells were pelleted by centrifugation, washed with PBS, and resuspended in His buffer containing 8 M urea. After sonication and centrifugation, the supernatant was loaded onto Ni-NTA Superflow bulk resin (Qiagen, Valencia, CA, USA) and the 6xHis-r ECF-L fusion protein was eluted with a 50–100 mM imidazole gradient. The eluent was dialyzed against milliQ water and injected into rabbit to generate the anti ECF-L polyclonal sera.
Western blot analysis for mECF-L in conditioned media from mouse bone marrow cultures
Polyclonal antisera against rmECF-L were raised in rabbits as previously described(22) and used to determine mECF-L expression in murine bone marrow culture treated with 1,25(OH)2D3 by immunoblot analysis. Conditioned media from mouse bone marrow cultures or 293 cells transiently transfected with mECF-L cDNA were harvested and concentrated 30-fold using a Microcon YM-10 (Millipore Corp, Bedford, MA, USA) filter. Samples were subjected to SDS-PAGE and transferred to nitrocellulose membranes. After blocking, the membranes were incubated with polyclonal antibody to mECF-L at 1:2500 dilution for 1 h. Horseradish peroxidase-conjugated goat anti-rabbit IgG (Sigma) was used as a secondary antibody (1:10,000), and the blots were developed with an enhanced chemiluminescent (ECL) system (Pierce, Rockford, IL, USA) in Kodak X-ray films.
Neutralizing effects of mECF-L antisera on OCL formation in mouse bone marrow cultures
To confirm the role of endogenous ECF-L on OCL formation/activation, we tested the effects of mECF-L antisera on OCL formation. The anti mECF-L polyclonal antisera or rabbit preimmune antisera (1:1,000–1:10,000) were added to mouse bone marrow culture treated with 10−9 M 1,25(OH)2D3 or 20 ng/ml RANKL. Every 2 days, one-half of the media was replaced with fresh media containing the antisera. At the end of the culture periods, the cells were fixed and stained for TRACP activity, and the number of TRACP+ MNCs was determined.
Production and purification of recombinant ECF-L-Fc fusion protein
ECF-L cDNA was generated by PCR using T7 and antisense primer (5′-ATCGTAATCCATAAGGGCCCTTGCAACTTG-3′), and the EcoRV-digested PCR-product was fused with the Fc coding domain of human IgG1. The mECF-L-Fc construct was stably transfected into 293 cells using a CaPO4 mammalian transfection kit (Stratagene, La Jolla, CA, USA) according to the manufacturer's protocol. One hundred micrograms of mECF-L-Fc fusion protein were purified from 1 liter of 293 cells conditioned media by protein G affinity chromatography (Roche Diagnostics). The effects of purified mECF-L-Fc fusion protein on OCL formation/activation was tested in murine bone marrow cultures as described above.
RT-PCR and Western blot analysis of RANKL expression in mouse bone marrow cultures treated with mECF-L
Mouse bone marrow cells (1.2 × 107/well) were cultured with mECF-L conditioned media in 6-well plates for 2 days. Total RNA was extracted with RNA-BEE (Tel Test, Friendswood, TX, USA) according to the manufacturer's protocol, and the expression levels of mouse RANKL mRNA were determined by RT-PCR analysis. The PCR conditions were 94°C for 30 s, 58°C for 30 s, 72°C for 1 minute, and 28 cycles. PCR primer sequences for mouse RANKL are as follows: 5′-GAAGGTACTCGTAGCTAAGG-3′ (sense) and 5′-GGCTATGTCAGCTCCTAAAG-3′(antisense). GAPDH was used as an internal control using primer sequences 5′-ACCACAGTCCATGCCATCAC-3′ (sense) and 5′-TCCACCACCCTGTTGCTGTA-3′ (antisense).
Western blot analysis was used to determine the effects of ECF-L conditioned media on RANKL expression in mouse bone marrow cultures. Mouse bone marrow cells (106 cells) were cultured with 10−10M 1,25(OH)2D3 for 2 days in the presence and absence of 10% mECF-L conditioned media and mECF-L antibody at 1:1000 dilutions. At the end of the culture period, mouse bone marrow cells were lysed with 200 μl of SDS lysis buffer and subjected to Western blot analysis as described above. Anti-RANKL polyclonal antibody (R&D Systems, Minneapolis, MN, USA) was used as a primary antibody at 1:10,000 dilutions. After 1 h of incubation, HRP conjugated antigoat lgG (Santa Cruz Biotechnology, Inc, Santa Cruz, CA, USA) was hybridized as a secondary antibody and visualized with ECL on X-ray film. The blot was stripped with a buffer containing SDS and β-mercaptoethanol and reprobed with antibody to β-actin (Santa Cruz Biotechnology) as a control for equal loading. The bands were quantified with a densitometer, and the ratio of RANKL to β-actin was calculated.
Chemotaxis assays were performed using 24-well Transwell chambers (8 μm) (Corning Costar, Cambridge, MA, USA). Mouse bone marrow cells (2 × 106/200 μl) were plated in the upper well, and 400 μl of 10% FBS-αMEM containing 400 ng/ml of recombinant ECF-L-Fc protein were added to the lower well. Control Fc protein was added to the lower well for control cultures. ECF-L antisera were added to the lower well at 1:1000 dilution to confirm the effects of the ECF-L Fc protein on chemotaxis. The cells were incubated for 3–5 h, and then the upper wells were removed. To identify OCL precursors that migrated into the lower well, 25 ng/ml of hRANKL and 10 ng/ml of M-CSF were added to the lower well. The cells were cultured and stained for TRACP activity after 2 days, and the number of TRACP+ mononuclear cells present was determined.
All experiments were performed in quadruplicate, and the mean ± SE for the number of OCLs formed was determined. The means of individual treatment groups were compared using Student's t-test, and the results were considered significantly different for p < 0.01.
Detection of ECF-L in mature OCLs and effects of ECF-L conditioned media on OCL formation in mouse bone marrow cultures
Using PCR-selective subtraction hybridization, we detected approximately 200 bands that were overexpressed in mature OCLs compared with B/T precursor cells. Sequence analysis of the 68 bands that were most highly overexpressed in mature OCLs was performed as previously described.(8) We have reported the results of the first 50 bands.(8) DNA sequences of the other 18 bands were compared with the database in NCBI. Five bands were mouse complement component C3 of varying insert sizes, and one band was ECF-L.
As shown in Fig. 1, purified ECF-L Fc protein increased OCL-like TRACP+ MNC formation in a dose-dependent manner compared with control Fc protein. OCL-like MNC formation induced by ECF-L was enhanced by low concentrations of 1,25(OH)2D3. Human ECF-L conditioned media enhanced 23c6+ MNC formation in human bone marrow cultures in a dose-dependent pattern in the presence of low concentration of 1,25(OH)2D3 (10−10 M). As shown in Figs. 2A and 2B, purified ECF-L Fc protein (40–200 ng/ml) significantly increased the number of nuclei per TRACP+ MNC and enhanced the effects of 1,25(OH)2D3 (10−10 M) in mouse marrow cultures. The ratio of TRACP+ multinucleated cells to TRACP+ mononuclear cells in cultures treated with ECF-L conditioned media (10% vol/vol) versus control conditioned media was 0.5 ± 0.04 versus 0.3 ± 0.06 and was significantly increased (p = 0.03) compared with ratio of TRACP+ MNC to mononuclear cells in cultures treated with 10% conditioned media from empty vector transfected 293 cells (Fig. 2C).
To investigate the bone resorbing capacity of OCL-like cells induced by ECF-L, mouse bone marrow cells were cultured on dentin slices in the presence of 10−10 M 1,25(OH)2D3. As shown in Fig. 3, the number of pits and the resorption area per dentin slice were significantly increased by marrow cells treated with ECF-L conditioned media and 10−10 M 1,25(OH)2D3 compared with those treated with the empty vector (EV) conditioned media and 1,25(OH)2D3. However, the number of pits per OCL and area resorbed per OCL were not significantly increased in cultures treated with ECF-L.
Time course effects of mECF-L conditioned media on OCL formation in mouse bone marrow cultures
To determine whether ECF-L stimulated the proliferation or differentiation stage of OCL formation/activation, ECF-L conditioned media were added on days 0–2, 2–4, or for the entire 6 days of the culture in the presence of 10−10 M 1,25(OH)2D3, and TRACP+ MNC formation was determined. TRACP+ MNC formation was significantly increased compared with cultures treated with the empty vector control media when ECF-L conditioned media were present for the later stage (days 4–6) of the culture or for the entire culture period. ECF-L conditioned media did not increase MNC formation if present only during the early stages of the culture (Fig. 4).
Western blot analysis
To determine if murine bone marrow cells secrete ECF-L into their conditioned media, we performed Western blot analysis with conditioned media from mouse bone marrow cultures treated with 1,25(OH)2D3 or from 293 cells transfected with the ECF-L cDNA, using polyclonal rmECF-L antisera raised in rabbits. As shown in Fig. 5, a 43-kDa band was detected in conditioned media from mouse bone marrow treated with 10−9 M 1,25(OH)2D3 and 293 cells transiently transfected with ECF-L cDNA, but not in conditioned media from 293 cells transfected with the empty vector.
Effects of antisense S-oligonucleotide to mECF-L on OCL formation in mouse bone marrow cultures
To determine if endogenous ECF-L was playing a role in OCL formation/activation, we tested the effects of an antisense S-oligonucleotide to ECF-L on MNC formation in murine cultures treated with 1,25(OH)2D3 (10−9 M). Antisense S-oligonucleotide to ECF-L (5–25 nM) significantly inhibited OCL formation by about 40% in murine bone marrow cultures stimulated with 10−9 M 1,25(OH)2D3 compared with the control cultures treated with sense S-oligonucleotide (Fig. 6). High concentrations of S-oligonucleotide (more than 50 nM) were toxic to murine bone marrow cultures. To confirm if endogenous ECF-L stimulated the late stages of osteoclast formation, we tested the effects of adding antisense S-oligonucleotide for mECF-L at different time periods on OCL formation. Additon of 25 nM of antisense S-oligonucleotide for mECF-L to mouse bone marrow cultures stimulated with 10−9 M of 1,25(OH)2D3 significantly inhibited the ratio of TRACP+ MNC to mononuclear cells only at the later stages of culture compared with control cultures treated with the sense S-oligonucleotide (Fig. 7).
In situ hybridization
To identify the cells that express ECF-L, we performed in situ hybridization studies on mouse bone marrow cultures with a DIG-labeled cRNA probe to ECF-L. The expression of ECF-L mRNA was detected in monocytes and MNCs, but not in fibroblasts (data not shown). In contrast, no signal was detected in control cultures hybridized with sense RNA probes. Interestingly, MNCs that contained less than 5 nuclei strongly expressed ECF-L, while MNCs that contained more than 10 nuclei did not express ECF-L.
Effect of ECF-L antisera in TRACP+ MNC formation in mouse bone marrow cultures
To confirm the role of ECF-L on OCL formation, the effects of ECF-L polyclonal antisera were tested in mouse bone marrow cultures stimulated with 10−9 M 1,25(OH)2D3 or 20 ng/ml RANKL. OCL formation induced by 1,25(OH)2D3 (Fig. 8A) and RANKL (Fig. 8B) was dose-dependently inhibited about 60% by anti-ECF-L antibody at 1:10,000–1:1,000 dilution.
Effects of rECF-L-Fc fusion protein on OCL formation/activation in mouse bone marrow cultures
We tested the effects of rECF-L-Fc fusion protein on OCL formation in mouse bone marrow cultures. rECF-L-Fc fusion protein induced TRACP+ MNC formation in the presence of 10−10 M 1,25(OH)2D3 (Fig. 9A) or 2.5 ng/ml RANKL (Fig. 9B) in a dose-dependent manner (4–200 ng/ml) compared with the control Fc protein.
To determine if OCL formation induced by ECF-L occurred through the RANK-RANKL pathway, the effects of OPG and RANK-Fc on OCL formation stimulated with ECF-L were examined. OPG and RANK-Fc significantly inhibited OCL formation induced by ECF-L-Fc in the presence of 10−10 M 1,25(OH)2D3 (Fig. 9A) or 2.5 ng/ml RANKL (Fig. 9B). However, ECF-L did not enhance RANKL mRNA expression induced by 10−10 M 1,25(OH)2D3 compared with control cultures treated with empty vector conditioned media (Fig. 9C). Furthermore, Western blot analysis showed that the expression levels of RANKL in the presence of 10−10 M 1,25(OH)2D3 were not significantly enhanced by ECF-L CM or decreased by ECF-L antibody compared with the β-actin internal control (Fig. 9D).
To test the chemotactic effects of ECF-L on OCL precursors, we performed chemotactic assays. As shown in Fig. 10, recombinant ECF-L showed chemotactic activity for OCL precursors compared with control cultures. Neutralizing mECF-L with the ECF-L antisera blocked the chemoattract effects of ECF-L on OCL precursors.
PCR-selective subtraction is a powerful technique for identifying genes that are overexpressed in mature OCLs compared with OCL precursors. Using this technique, we detected mouse ECF-L. We confirmed by in situ hybridization that ECF-L mRNA was highly expressed in OCLs that had less than five nuclei and in marrow mononuclear cells.
ECF-L was first identified as a novel eosinophil chemotactic cytokine by Owhashi et al.(9) ECF-L possesses a CXC sequence near the NH2 terminus of the mature molecule, which is a typical motif shared with many chemokine family proteins. Sequence alignments revealed that ECF-L differs from other known eosinophil chemotactic cytokines such as IL-5,(23) RANTES,(24) eotaxin,(25) or ecalectin.(26) Comparisons of the deduced amino acid sequence with those contained in several databases revealed that ECF-L had a high homology with the chitinase family of 18 glycosyl hydrolases and vertebrate chitinase family proteins that do not show chitinase activity.(9) Although proteins of the chitinase family are detected in mammals, no chitinolytic activity has been detected,(6) and the actual physiological roles of the mammalian chitinases family proteins remain to be clarified.
Our study demonstrated that ECF-L enhanced OCL formation in the presence of low concentrations of osteotropic factors such as 1,25(OH)2D3 and RANKL in mouse bone marrow cultures. ECF-L did not activate preformed OCL because, as shown in Fig. 3, ECF-L did not increase the number of pits per OCL or the area resorbed per OCL, but rather increased resorption by increasing the number of OCL. Furthermore, ECF-L did not increase the survival of preformed osteoclasts or increase the expression of M-CSF in mouse marrow cultures (data not shown).
Time-course studies suggested that ECF-L acts at the later stages of OCL formation such as the cell fusion stage rather than inducing proliferation of OCL precursors. Blocking ECF-L expression in marrow cultures inhibits OCL formation, suggesting an important role for ECF-L in the later stages of osteoclastogenesis. ECF-L may also act as a chemoattractant for OCL precursors, as demonstrated by chemotactic assays with ECF-L.
ECF-L seems to play an important role in RANKL mediated osteoclastogenesis. ECF-L enhanced RANKL-induced OCL formation, and ECF-L antisera blocked OCL formation induced by RANKL. In addition, OPG or RANK-Fc inhibited ECF-L-enhanced OCL formation. However, the effects of ECF-L on OCL formation were not caused by increased expression of RANKL or RANK, because ECF-L did not increase RANKL levels in mouse bone marrow cultures treated with 1,25(OH)2D3, using either RT-PCR or Western blot analysis. Furthermore, ECF-L antisera did not affect the expression of RANK (data not shown). These data suggest that ECF-L requires RANKL to induce OCL formation, but is itself an important cofactor involved in RANKL-induced OCL formation, possibly through its chemotactic effects on OCL precursors. Interestingly, ECF-L did not enhance OCL formation induced by TNF-α, suggesting that it is a cofactor only for RANKL-induced OCL formation (data not shown).
Chemokines activate cells by binding to specific cell-surface receptors that belong to a superfamily of serpentine G-protein-coupled receptors, and the receptor binding profiles of various chemokines have been reviewed.(27,28) Votta et al.(29) showed that a novel chemokine, CK-β8 (recently designated as CCL23; and previously described as myeloid progenitor inhibitory factor-1 [MPIF-1]), was a chemotactic factor for human OCL precursors. CCR1 seems to be the primary receptor on monocytes and eosinophils through which CK-β8 signaling is transduced.(30,31) ECF-L has been reported to have a specificity similar to RANTES as a chemoattractant for eosinophils, T-lymphocytes, and bone marrow cells, and this result indicates that the receptor(s) for ECF-L is related to that for RANTES.(9) However, RANTES binds to multiple chemokine receptors (CCR1, CCR3, CCR4, and CCR5). Thus, the identification of receptor mediating the effects of ECF-L remains to be clarified.
In summary, ECF-L is a recently identified chemokine that is a chemoattractant for OCL precursors. ECF-L is highly expressed in OCL and mononuclear OCL precursors and enhances OCL formation induced by RANKL. However, ECF-L does not induce RANKL expression but seems to play an important role in RANK-induced OCL formation.
This study was supported by R01-AR41336 from NIAMS and funds from the MMRF.