Voltar
Artigo Acesso aberto Revisado por pares
BIBTEX RIS

Tumor Necrosis Factor-α Induces Differentiation of and Bone Resorption by Osteoclasts

2000; Elsevier BV; Volume: 275; Issue: 7 Linguagem: Inglês

10.1074/jbc.275.7.4858

ISSN

1083-351X

Autores

Yoshiaki Azuma, Keisuke Kaji, Rei Katogi, Sunao Takeshita, Akira Kudō,

Tópico(s)

Bone and Joint Diseases

Resumo

Osteoclast progenitors differentiate into mature osteoclasts in the presence of receptor activator of NF-κB (RANK) ligand on stromal or osteoblastic cells and monocyte macrophage colony-stimulating factor (M-CSF). The soluble RANK ligand induces the same differentiation in vitro without stromal cells. Tumor necrosis factor-α (TNF-α), a potent cytokine involved in the regulation of osteoclast activity, promotes bone resorption via a primary effect on osteoblasts; however, it remains unclear whether TNF-α can also directly induce the differentiation of osteoclast progenitors into mature osteoclasts. This study revealed that TNF-α directly induced the formation of tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cells (MNCs), which produced resorption pits on bone in vitro in the presence of M-CSF. The bone resorption activity of TNF-α-induced MNCs was lower than that of soluble RANK ligand-induced MNCs; however, interleukin-1β stimulated this activity of TNF-α-induced MNCs without an increase in the number of MNCs. In this case, interleukin-1β did not induce TRAP-positive MNC formation. The osteoclast progenitors expressed TNF receptors, p55 and p75; and the induction of TRAP-positive MNCs by TNF-α was inhibited completely by an anti-p55 antibody and partially by an anti-p75 antibody. Our findings presented here are the first to indicate that TNF-α is a crucial differentiation factor for osteoclasts. Our results suggest that TNF-α and M-CSF play an important role in local osteolysis in chronic inflammatory diseases. Osteoclast progenitors differentiate into mature osteoclasts in the presence of receptor activator of NF-κB (RANK) ligand on stromal or osteoblastic cells and monocyte macrophage colony-stimulating factor (M-CSF). The soluble RANK ligand induces the same differentiation in vitro without stromal cells. Tumor necrosis factor-α (TNF-α), a potent cytokine involved in the regulation of osteoclast activity, promotes bone resorption via a primary effect on osteoblasts; however, it remains unclear whether TNF-α can also directly induce the differentiation of osteoclast progenitors into mature osteoclasts. This study revealed that TNF-α directly induced the formation of tartrate-resistant acid phosphatase (TRAP)-positive multinucleated cells (MNCs), which produced resorption pits on bone in vitro in the presence of M-CSF. The bone resorption activity of TNF-α-induced MNCs was lower than that of soluble RANK ligand-induced MNCs; however, interleukin-1β stimulated this activity of TNF-α-induced MNCs without an increase in the number of MNCs. In this case, interleukin-1β did not induce TRAP-positive MNC formation. The osteoclast progenitors expressed TNF receptors, p55 and p75; and the induction of TRAP-positive MNCs by TNF-α was inhibited completely by an anti-p55 antibody and partially by an anti-p75 antibody. Our findings presented here are the first to indicate that TNF-α is a crucial differentiation factor for osteoclasts. Our results suggest that TNF-α and M-CSF play an important role in local osteolysis in chronic inflammatory diseases. tumor necrosis factor-α interleukin monocyte macrophage colony-stimulating factor receptor activator of NF-κB RANK ligand soluble RANK ligand osteoprotegerin α-minimal essential medium M-CSF-dependent bone marrow macrophage tartrate-resistant acid phosphatase multinucleated cell Osteoclasts, which differentiate from hematopoietic stem cells (1.Scheven B.A. Visser J.W. Nijweide P.J. Nature. 1986; 321: 79-81Crossref PubMed Scopus (230) Google Scholar, 2.Ibbotson K.J. Roodman G.D. McManus L.M. Mundy G.R. J. Cell Biol. 1984; 99: 471-480Crossref PubMed Scopus (185) Google Scholar, 3.Burger E.H. Van der Meer J.W. van de Gevel J.S. Gribnau J.C. Thesingh G.W. van Furth R. J. Exp. Med. 1982; 156: 1604-1614Crossref PubMed Scopus (259) Google Scholar), have a crucial role not only in physiological bone remodeling (4.Kahn A.J. Partridge N.C. Hall B.K. Bone. 2. CRC Press, Boca Raton, FL1991: 119Google Scholar), but they also function in the local bone destruction that occurs in association with chronic inflammatory diseases (5.Mills B.G. Hall B.K. Bone. 2. CRC Press, Boca Raton, FL1991: 175Google Scholar). Tumor necrosis factor-α (TNF-α),1 which is a member of the TNF ligand superfamily and a multifunctional cytokine, regulates various cellular reactions such as proliferation, differentiation, maintenance of a differentiated phenotype, and apoptosis in various types of cells (6.Old L.J. Bonavida B. Granger G. Tumor Necrosis Factor: Structure, Mechanism of Action, Role in Disease and Therapy. Karger, Basel1990: 1-30Google Scholar, 7.Ashkenazi A. Dixit V.M. Curr. Opin. Cell Biol. 1999; 11: 255-260Crossref PubMed Scopus (1163) Google Scholar). In the bone microenvironment, TNF-α has multiple actions on bone cells (8.Mundy G.R. Roodman G.D. Bonewald L.F. Yoneda T. Sabatini M. Immunol. Ser. 1992; 56: 483-498PubMed Google Scholar);e.g. TNF-α inhibits DNA and collagen synthesis and osteocalcin gene expression in osteoblasts, but then it stimulates the synthesis of proteolytic enzymes such as plasminogen activators and matrix metalloproteinases and of cytokines such as interleukin (IL)-8, IL-6, and monocyte macrophage colony-stimulating factor (M-CSF) in these cells (9.Nanes M.S. McKoy W.M. Marx S.J. Endocrinology. 1989; 124: 339-345Crossref PubMed Scopus (84) Google Scholar, 10.Nanes M.S. Rubin J. Titus L. Hendy G.N. Catherwood B. Endocrinology. 1991; 128: 2577-2582Crossref PubMed Scopus (58) Google Scholar, 11.Kuno H. Kurian S.M. Hendy G.N. White J. deLuca H.F. Evans C.O. Nanes M.S. Endocrinology. 1994; 134: 2524-2531Crossref PubMed Scopus (42) Google Scholar, 12.Panagakos F.S. Kumar S. Inflammation. 1994; 18: 243-265Crossref PubMed Scopus (40) Google Scholar, 13.Chaudhary L.R. Spelsberg T.C. Riggs B.L. Endocrinology. 1992; 130: 2528-2534Crossref PubMed Scopus (126) Google Scholar, 14.Ishimi Y. Miyaura C. Jin C.H. Akatsu T. Abe E. Nakamura Y. Yamaguchi A. Yoshiki S. Matsuda T. Hirano T. Kishimoto T. Suda T. J. Immunol. 1990; 145: 3297-3303PubMed Google Scholar, 15.Passeri G. Girasole G. Manolagas S.C. Jilka R.L. Bone Miner. 1994; 24: 109-126Abstract Full Text PDF PubMed Scopus (42) Google Scholar, 16.Elford P.R. Felix R. Cecchini M. Trechsel U. Fleisch H. Calcif. Tissue Int. 1987; 41: 151-156Crossref PubMed Scopus (94) Google Scholar, 17.Felix R. Fleisch H. Elford P.R. Calcif. Tissue Int. 1989; 44: 356-360Crossref PubMed Scopus (88) Google Scholar). TNF-α is also a potent inducer of bone resorption (18.Bertolini D.R. Nedwin G.E. Bringman T.S. Smith D.D. Mundy G.R. Nature. 1986; 319: 516-518Crossref PubMed Scopus (1238) Google Scholar), which is a major action of activated macrophages (1.Scheven B.A. Visser J.W. Nijweide P.J. Nature. 1986; 321: 79-81Crossref PubMed Scopus (230) Google Scholar). It has been reported that TNF-α promotes bone resorption in vitro and in vivo by activating mature osteoclasts (19.Kitazawa R. Kimble R.B. Vannice J.L. Kung V.T. Pacifici R. J. Clin. Invest. 1994; 94: 2397-2406Crossref PubMed Scopus (290) Google Scholar, 20.Thomson B.M. Mundy G.R. Chambers T.J. J. Immunol. 1987; 138: 775-779PubMed Google Scholar, 21.Lerner U.H. Ohlin A. J. Bone Miner. Res. 1993; 8: 147-155Crossref PubMed Scopus (49) Google Scholar) or by stimulating proliferation and differentiation of osteoclasts precursors (22.van der Pluijm G. Most W. van der Wee-Pals L. de Groot H. Papapoulos S. Lowik C. Endocrinology. 1991; 129: 1596-1604Crossref PubMed Scopus (87) Google Scholar) or indirectly via a primary effect on osteoblasts (23.Pfeilschifter J. Chenu C. Bird A. Mundy G.R. Roodman G.D. J. Bone Miner. Res. 1989; 4: 113-118Crossref PubMed Scopus (408) Google Scholar). Diseases such as rheumatoid arthritis, aseptic loosening, and periodontal diseases have been associated with the accumulation of TNF-α and/or other proinflammatory cytokines such as IL-1 and IL-6, which likely mediate local bone destruction by stimulating osteoclast activity (24.Saxne T. Palladino Jr., M.A. Heinegard D. Talal N. Wollheim F.A. Arthritis Rheum. 1988; 31: 1041-1045Crossref PubMed Scopus (569) Google Scholar, 25.Xu J.W. Konttinen Y.T. Lassus J. Natah S. Ceponis A. Solovieva S. Aspenberg P. Santavirta S. Clin. Exp. Rheumatol. 1996; 14: 643-648PubMed Google Scholar, 26.Merkel K.D. Erdmann J.M. McHugh K.P. Abu-Amer Y. Ross F.P. Teitelbaum S.L. Am. J Pathol. 1999; 154: 203-210Abstract Full Text Full Text PDF PubMed Scopus (371) Google Scholar, 27.Nair S.P. Meghji S. Wilson M. Reddi K. White P. Henderson B. Infect. Immun. 1996; 64: 2371-2380Crossref PubMed Google Scholar). Most factors known to stimulate osteoclast formation (such as TNF-α, IL-1, IL-6, parathyroid hormone, vitamin D3, and prostaglandins) bind to receptors on stromal cells/osteoblastic cells rather than binding to receptors on osteoclast progenitors to induce the release of osteoclast-stimulating factors (28.Boyce B.F. Hughes D.E. Wright K.R. Xing L. Dai A. Lab. Invest. 1999; 79: 83-94PubMed Google Scholar). Recent studies demonstrate that two essential factors supplied by stromal cells/osteoblastic cells for the differentiation and maturation of osteoclast progenitors are M-CSF and RANK ligand (RANKL) (29.Takahashi N. Udagawa N. Suda T. Biochem. Biophys. Res. Commun. 1999; 256: 449-455Crossref PubMed Scopus (406) Google Scholar). In the presence of M-CSF, a genetically engineered soluble form of RANKL induced osteoclast differentiation of progenitors derived from mouse bone marrow cells, spleen cells, or human peripheral blood mononuclear cells in the absence of stromal cells/osteoblastic cells (30.Yasuda H. Shima N. Nakagawa N. Yamaguchi K. Kinosaki M. Mochizuki S. Tomoyasu A. Yano K. Goto M. Murakami A. Tsuda E. Morinaga T. Higashio K. Udagawa N. Takahashi N. Suda T. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 3597-3602Crossref PubMed Scopus (3642) Google Scholar, 31.Matsuzaki K. Udagawa N. Takahashi N. Yamaguchi K. Yasuda H. Shima N. Morinaga T. Toyama Y. Yabe Y. Higashio K. Suda T. Biochem. Biophys. Res. Commun. 1998; 246: 199-204Crossref PubMed Scopus (324) Google Scholar), suggesting that RANKL and M-CSF induce osteoclast formation by direct activation of osteoclast progenitors. Osteoprotegerin (OPG) ligand (RANKL)-deficient mice exhibit typical osteopetrosis with total occupation of the bone marrow space within endosteal bone (32.Kong Y.-Y. Yoshida H. Sarosi I. Tan H.-L. Timms E. Capparelli C. Morony S. Oliveira-dos-Santos A.J. Van G. Itie A. Khoo W. Wakeham A. Dunstan C.R. Lacey D.L. Mak T.W. Boyle W.J. Penninger J.M. Nature. 1999; 397: 315-323Crossref PubMed Scopus (2932) Google Scholar), demonstrating that RANKL is essential for osteoclast development. However, recently it has been shown that several inflammatory cytokines induce multinucleation of preosteoclasts and/or commitment to the osteoclast lineage in the absence of stromal cells/osteoblastic cells. IL-1 induced multinucleation of mononuclear prefusion osteoclasts and bone resorption activity in the absence of stromal cells/osteoblastic cells (33.Jimi E. Nakamura I. Duong L.T. Ikebe T. Takahashi N. Rodan G.A. Suda T. Exp. Cell Res. 1999; 247: 84-93Crossref PubMed Scopus (307) Google Scholar). IL-15 increased the number of mononuclear preosteoclast-like cells (34.Ogata Y. Kukita A. Kukita T. Komine M. Miyahara A. Miyazaki S. Kohashi O. J. Immunol. 1999; 162: 2754-2760PubMed Google Scholar). Lipopolysaccharide induced c-src mRNA expression in purified bone marrow macrophages, which is a specific marker of commitment to the osteoclast lineage (35.Abu-Amer Y. Ross F.P. Edwards J. Teitelbaum S.L. J. Clin. Invest. 1997; 100: 1557-1565Crossref PubMed Scopus (261) Google Scholar). These findings suggest that the direct actions of several cytokines on osteoclast progenitors contribute to osteoclast formation. Especially in some pathological conditions involving articular subchondral bones and/or synovial tissues as in rheumatoid arthritis patients and in the tissue surrounding failed implants in patients who received total joint replacement, several cytokines reacting to osteoclast progenitors may be involved in local osteolysis (24.Saxne T. Palladino Jr., M.A. Heinegard D. Talal N. Wollheim F.A. Arthritis Rheum. 1988; 31: 1041-1045Crossref PubMed Scopus (569) Google Scholar, 25.Xu J.W. Konttinen Y.T. Lassus J. Natah S. Ceponis A. Solovieva S. Aspenberg P. Santavirta S. Clin. Exp. Rheumatol. 1996; 14: 643-648PubMed Google Scholar, 26.Merkel K.D. Erdmann J.M. McHugh K.P. Abu-Amer Y. Ross F.P. Teitelbaum S.L. Am. J Pathol. 1999; 154: 203-210Abstract Full Text Full Text PDF PubMed Scopus (371) Google Scholar, 27.Nair S.P. Meghji S. Wilson M. Reddi K. White P. Henderson B. Infect. Immun. 1996; 64: 2371-2380Crossref PubMed Google Scholar). Here we show a novel activity for TNF-α, i.e. induction of osteoclastogenesis by a direct action on osteoclast progenitors, which action is mediated mainly by the p55 TNF receptor. We thus report for the first time that TNF-α is a crucial differentiation factor for osteoclasts in the presence of M-CSF. Male ddY mice, 6–9 weeks of age, were used (Sankyo Laboratories Co. Ltd., Tokyo, Japan). Recombinant products of mouse TNF-α, mouse IL-1β, human soluble RANKL (sRANKL), and human M-CSF were purchased from Pepro-Tech EC Ltd. (London, U. K.). Recombinant mouse OPG-Fc chimera was purchased from R&D Systems Inc. (Minneapolis, MN). Monoclonal hamster antibodies against mouse TNF receptors p55 and p75 and a rat monoclonal antibody against mouse IL-1 receptor type I were purchased from Genzyme Inc. (Cambridge, MA). Fluorescein isothiocyanate-conjugated streptavidin and anti-rat IgG were purchased from Santa Cruz (Santa Cruz, CA) and Cappel (Durham, NC), respectively. All other chemicals were purchased from Nacalai Tesque, Inc. (Kyoto, Japan) unless otherwise stated. The femurs and tibias of adult ddY mice were aseptically removed, and adherent soft tissues were dissected away. The bone ends were cut off with scissors, and the marrow cavity was flushed with α-MEM (ICN Biochemical Inc., Costa Mesa, CA). The marrow cells were collected, washed with α-MEM, and cultured in α-MEM containing 10% fetal calf serum (Dainippon Pharmaceuticals Co. Ltd., Tokyo, Japan), 100 IU/ml penicillin G (Meiji Chemical Co. Ltd., Tokyo, Japan), and 100 μg/ml streptomycin (Meiji Chemical Co. Ltd.) at 37 °C in 5% CO2. The mouse M-CSF was supplied from the culture supernatant of a selected high M-CSF-producing cell line, CMG14-12, 2S. Takeshita, K. Kaji, and A. Kudo, submitted for publication. or recombinant human M-CSF was used as a source of M-CSF. Osteoclastogenic cultures were obtained by use of two different protocols. Whole bone marrow cells were plated in a 48-well plate at a density of 1.5 × 105cells/cm2 and were maintained for 5 days in the presence of a 0.10 volume of CMG14-12 culture supernatant (final concentration, approximately 35,000 units/ml M-CSF). Whole bone marrow cells were cultured in α-MEM containing 10% fetal calf serum and a 0.10 volume of CMG14-12 culture supernatant at 5 × 106 cells in a 10-cm suspension culture dish (Corning Costar Inc., Corning, NY). After 3 days in culture, the cells were washed, harvested with 0.02% EDTA in phosphate-buffered saline, and seeded at 3 × 105cells into another 10-cm suspension culture dish. After a further 3 days in culture, the cells were harvested, plated at a density of 1.5 × 104 cells/cm2, and maintained for 5 days in the presence of a 0.10 volume of CMG14-12 culture supernatant or recombinant human M-CSF (100 ng/ml). We evaluated the expression of surface molecules on MDBM cells by flow cytometry. MDBM cells expressed CD9, FcγRII/III, c-Fms, Mac-1, Mac-2, F4/80, β1-integrin, β2-integrin, β3-integrin, and CD98. Because all positive signals were uniform, MDBM cells are likely to be a homogeneous population in contrast to bone marrow cells. 2S. Takeshita, K. Kaji, and A. Kudo, submitted for publication. Cultured cells were fixed with 10% formalin for 5 min. After that they were re-fixed with ethanol:cetone (50:50 v/v) for 1 min and incubated in acetate buffer (pH 4.8) containing naphthol AS-MX phosphate (Sigma Chemical Co., St. Louis, MO), fast red violet LB salt (Sigma), and 50 mm sodium tartrate at room temperature (36.Takahashi N. Yamana H. Yoshiki S. Roodman G.D. Mundy G.R. Jones S.J. Boyde A. Suda T. Endocrinology. 1988; 122: 1373-1382Crossref PubMed Scopus (702) Google Scholar). F-actin fibers were detected by staining with rhodamine-conjugated phalloidin solution (Wako Pure Chemicals, Co. Ltd., Osaka, Japan). After incubation with the staining solution, the cells were rinsed with phosphate-buffered saline for 10 min and then mounted with Permaflor (Lipshaw Immunon, Pittsburgh, PA) as described by Kanehisa et al. (37.Kanehisa J. Yamanaka T. Doi S. Turksen K. Heersche J.N. Aubin J.E. Takeuchi H. Bone. 1990; 11: 287-293Crossref PubMed Scopus (126) Google Scholar). Osteoclasts were characterized further by assessing their ability to form resorption pits on dentin slices, as described previously (38.Tamura T. Takahashi N. Akatsu T. Sasaki T. Udagawa N. Tanaka S. Suda T. J. Bone Miner. Res. 1993; 8: 953-960Crossref PubMed Scopus (90) Google Scholar). Briefly, the slices were cleaned by ultrasonication in 70% ethanol, and each slice was then placed into 48-well or 96-well plates (Corning Costar). MDBM cells (1 × 104 cells/well) in α-MEM containing 10% fetal calf serum were added into each well and incubated for different periods at 37 °C in a humidified atmosphere of 5% CO2in air. The slices were stained with Coomassie Brilliant Blue R (39.Kitamura K. Katoh M. Komiyama O. Kitagawa H. Matsubara F. Kumegawa M. Bone. 1993; 14: 829-834Crossref PubMed Scopus (14) Google Scholar). The number of resorption lacunae (pits) on each slice was counted under a light microscope. To estimate the efficiency of bone resorption activity further, we performed a calcium resorption assay using calcium phosphate-coated osteologic discs (Millenium Biologix, Ontario, Canada) (40.Davis J.E. Shapiro G. Lowenberg B.F. Cells Materials. 1993; 3: 245-256Google Scholar). MDBM cells (2.5 × 103 cells/well) were seeded onto the discs in 16-well plates (each well size was equivalent to that of typical 96-well plates). Each well was filled with 0.15 ml of α-MEM containing 10% fetal calf serum. The MDBM cells were incubated for different periods of time at 37 °C in a humidified atmosphere of 5% CO2 in air, after which the discs were examined by phase-contrast microscopy. These discs were washed in a solution of 6% NaClO and 5.2% NaCl for removal of the cells to observe the resorption pits more clearly. The number or area of resorption lacunae (pits) on each slice was measured under a light microscope. Cells were harvested and stained with biotin-labeled anti-TNF receptor monoclonal antibodies or anti-IL-1 receptor type I antibody. The stained cells were incubated further with fluorescein isothiocyanate-conjugated streptavidin or goat anti-rat IgG and analyzed by flow cytometry (FACS-Calibur instruments and software, Becton Deckinson Inc., San Jose, CA). To search for the dominant cytokines involved in local bone destruction, we examined the effects of proinflammatory cytokines such as IL-1 and TNF-α on osteoclast differentiation using the newly established osteoclast progenitors (MDBM cells) described under “Experimental Procedures” as well as bone marrow cells. As shown in Fig. 1 A, TNF-α strongly induced differentiation of MDBM cells into mature osteoclasts, TRAP-positive multinucleated cells (MNCs) in the presence of M-CSF. The number of MNCs formed was 473 ± 75 in the presence of TNF-α and M-CSF, whereas it was 842 ± 64 in the presence of sRANKL and M-CSF. However, no TRAP-positive MNCs were formed in the MDBM cell cultures treated with 50 ng/ml of IL-1β even in the presence of M-CSF. Additionally, IL-1β had no synergistic effect on MNC formation induced by TNF-α and M-CSF. The results demonstrate that TNF-α can induce the formation of TRAP-positive MNCs, though not to the extent found with sRANKL. Furthermore, TRAP-positive MNCs were also formed from bone marrow cells treated with TNF-α or sRANKL, but not when treated with IL-1β, in the presence of M-CSF (Fig. 1 B). The efficiency of induction of TRAP-positive MNCs by TNF-α was significantly lower in the bone marrow cell cultures than in the MDBM cell cultures even though the cell density in the former was 10-fold higher than that in the latter, although the efficiency of that by sRANKL was almost the same in both types of cultures. To examine the reason for the difference in efficiency of induction of TRAP-positive MNCs by TNF-α between the two culture systems, we investigated the effects of pretreatment with M-CSF. Pretreatment with M-CSF for 24 h markedly increased the number of TRAP-positive MNCs in the bone marrow cell cultures (TableI). In contrast, the efficiency of sRANKL-induced TRAP-positive MNC formation in the bone marrow cell cultures was not affected by this pretreatment (Table I). These results demonstrate that M-CSF has an important role in osteoclastogenesis induced by TNF-α.Table IEffects of M-CSF pretreatment on the TNF-α-induced TRAP-positive MNC formation in the bone marrow cell culturesTRAP-positive MNCs (>10 nuclei)No treatmentaTNF-α (50 or 250 ng/ml), sRANKL (50 ng/ml), and/or IL-1β (50 ng/ml) was added to the culture immediately after cell seeding in the presence of M-CSF (35,000 units/ml).PretreatmentbCells were pretreated with M-CSF (35,000 units/ml) for 24 h and then cultured under the same conditions as in footnotea.Ratio (pre-/no)TNF-α (50 ng/ml)8 ± 281 ± 1410.1TNF-α (250 ng/ml)31 ± 6130 ± 404.2TNF-α (50 ng/ml) + IL-1β (50 ng/ml)28 ± 874 ± 152.6sRANKL (50 ng/ml)811 ± 81712 ± 410.9Bone marrow cells (1 × 105 cells/well) were cultured for 5 days. TRAP-positive MNCs with more than 10 nuclei were counted. The results shown are the mean ± S.D. of three independent experiments.a TNF-α (50 or 250 ng/ml), sRANKL (50 ng/ml), and/or IL-1β (50 ng/ml) was added to the culture immediately after cell seeding in the presence of M-CSF (35,000 units/ml).b Cells were pretreated with M-CSF (35,000 units/ml) for 24 h and then cultured under the same conditions as in footnotea. Open table in a new tab Bone marrow cells (1 × 105 cells/well) were cultured for 5 days. TRAP-positive MNCs with more than 10 nuclei were counted. The results shown are the mean ± S.D. of three independent experiments. To test the effect of the TNF-α concentration in the culture, we treated MDBM cells with 0.4–250 ng/ml TNF-α. As shown in Fig.2, TNF-α induced the formation of TRAP-positive MNCs from MDBM cells in a dose-dependent manner at TNF-α concentrations of 10 ng/ml and above. The number of TRAP-positive MNCs in TNF-α-treated groups was about 50% of that in sRANKL-treated groups (Fig. 2). We used the pit formation assay on dentin slices and the calcium resorption assay on calcium phosphate-coated discs to examine the bone resorption activity of TNF-α-induced MNCs. MDBM cells were cultured in the presence of TNF-α and M-CSF on dentin slices or calcium-coated discs for 5–7 days. Resorption pits were observed on both dentin slices (Fig. 3 A) and calcium phosphate-coated discs (Fig. 3 B) in the cultures treated with TNF-α. But the area of resorption was approximately 10 times smaller than that in cultures treated with sRANKL (Fig. 3,panel A, a and c; panel B, b and c; and panel C), even though the number of TNF-α-induced TRAP-positive cells on dentin slices was about half of that of the sRANKL-induced cells (Fig. 3 A, d–f). Interestingly, IL-1 remarkably increased only the resorption activity but not the number of TRAP-positive MNCs in the cultures cotreated with TNF-α and M-CSF (Fig. 3, Ab and Bd). The level of bone resorption activity in the treatment with TNF-α plus IL-1β was close to that in the sRANKL-treated cultures, indicating that IL-1β stimulated the bone resorption activity of the TNF-α-induced MNCs (Fig. 3,panel A, a, b, d, ande; panel B, c and d; andpanel C). An anti-IL-1 receptor type I antibody markedly inhibited the bone resorption activity of TNF-α-induced MNCs which was stimulated by IL-1β (Fig. 3 C). OPG-Fc inhibited the bone resorption activity of sRANKL-induced MNCs but not of TNF-α-induced MNCs (data not shown). The morphological features of TNF-α-induced MNCs in the MDBM cell cultures and in the bone marrow cell cultures showed a smooth contour with a large number of nuclei, whose features were equivalent to those of the sRANKL-induced MNCs (Figs.4 A, d–f). The formation of ringed F-actin structures (actin rings) in osteoclasts is closely related to osteoclast function (41.Wesolowski G. Duong L.T. Lakkakorpi P.T. Nagy R.M. Tezuka K. Tanaka H. Rodan G.A. Rodan S.B. Exp. Cell Res. 1995; 219: 679-686Crossref PubMed Scopus (80) Google Scholar, 42.Jimi E. Nakamura I. Amano H. Taguchi Y. Tsurukai T. Tamura M. Takahashi N. Suda T. Endocrinology. 1996; 137: 2187-2190Crossref Scopus (188) Google Scholar, 43.Nakamura I. Takahashi N. Sasaki T. Jimi E. Kurokawa T. Suda T. J. Bone Miner. Res. 1996; 11: 1873-1879Crossref PubMed Scopus (94) Google Scholar). The band of F-actin-containing podosomes was recognized at the periphery of numerous TNF-α-induced TRAP-positive MNCs (Fig. 4 Aa). However, no large actin ring was clearly visible in these cells, despite the presence of actin deep inside of them (Fig.4 Aa). Although there was no difference in the number of MNCs with an actin ring between the TNF-α-treated cultures (Fig.4 Aa) and those cotreated with TNF-α and IL-1β (Fig.4 Ab), bone resorption activity was increased by the addition of IL-1β (Fig. 3 C). These results indicate that IL-1β stimulates a specific signal required for bone resorption rather than the rearrangement of cytoskeletal organization in TNF-α-induced MNCs, although IL-1β alone cannot induce the formation of TRAP-positive MNCs. The mean number of nuclei in the typical TNF-α-induced TRAP-positive MNCs (n = 200) was 22.0 ± 4.9 (Fig.4 B). There was no significant difference for the mean number between TNF-α-induced MNCs and sRANKL-induced MNCs (25.0 ± 8.49 in Fig. 4 Ac and B). It is well known that many cells, including macrophages, express two TNF receptors, p55 and p75 (44.Lewis M. Tartaglia L.A. Lee A. Bennett G.L. Rice G.C. Wong G.H. Chen E.Y. Goeddel D.V. Proc. Natl. Acad. Sci. U. S. A. 1991; 88: 2830-2834Crossref PubMed Scopus (589) Google Scholar). IL-1 recognizes cell surface receptors type I and type II but may elicit biological responses only through the type I receptor (45.Dinarello C.A. FASEB J. 1994; 8: 1314-1325Crossref PubMed Google Scholar). Flow cytometry indicated that MDBM cells expressed both of the TNF receptors and the type I receptor of IL-1 (Fig.5). To determine which TNF receptor mediates osteoclastogenesis of MDBM cells, we added an antagonistic antibody against mouse p55 or p75 to the MDBM cell cultures in the presence of TNF-α. The formation of TRAP-positive MNCs by TNF-α was inhibited completely by the anti-p55 antibody but only partially by anti-p75, using an optimal dosage (10 μg/ml) of both antibodies for inhibition after testing multiple concentration (Fig. 6). These results demonstrate that TNF-α induces osteoclastogenesis of MDBM cells mainly via the p55 TNF receptor. 50 ng/ml OPG-Fc partially blocked MNC induction by sRANKL but not that by TNF-α (Fig. 6), and then 100 ng/ml OPG-Fc completely blocked it (data not shown). IL-1β may stimulate the bone resorption activity of MNCs rather than multinucleation in TNF-α-induced osteoclastogenesis. Because IL-1β stimulated the bone resorption activity of TNF-α-induced MNCs (Fig.4 C) but did not increase the number of TNF-α-induced MNCs (Fig. 1 A), such IL-1β may transduce biological signals through the type I receptor (45.Dinarello C.A. FASEB J. 1994; 8: 1314-1325Crossref PubMed Google Scholar) in osteoclasts. Although an anti-IL-1 receptor type I antibody did not affect the formation of TRAP-positive MNCs in the presence of TNF-α and IL-1β (Fig. 6), it markedly inhibited the bone resorption activity of TNF-α-induced MNCs (Fig.3 C). The cause of chronic inflammatory diseases such as rheumatoid arthritis, aseptic loosening, and periodontal diseases has been associated with the accumulation of TNF-α and/or other proinflammatory cytokines, which likely mediate local bone destruction. For example, the concentrations of M-CSF, TNF-α, IL-6, and IL-1 in synovial fluid and/or in serum of rheumatoid arthritis patients are extremely high (24.Saxne T. Palladino Jr., M.A. Heinegard D. Talal N. Wollheim F.A. Arthritis Rheum. 1988; 31: 1041-1045Crossref PubMed Scopus (569) Google Scholar, 46.Smith J.B. Bocchieni M.H. Smith Jr., J.B. Sherbin-Allen L. Abruzzo J.L. Rheumatol. Int. 1990; 10: 131-134Crossref PubMed Scopus (19) Google Scholar, 47.Neidel J. Schulze M. Lindschau J. Inflamm. Res. 1995; 44: 217-221Crossref PubMed Scopus (90) Google Scholar, 48.Kahle P. Saal J.G. Schaudt K. Zacher J. Fritz P. Pawelec G. Ann. Rheum. Dis. 1992; 51: 731-734Crossref PubMed Scopus (178) Google Scholar, 49.Kanik K.S. Hagiwara E. Yarboro C.H. Schumacher H.R. Wilder R.L. Klinman D.M. J. Rheumatol. 1998; 25: 16-22PubMed Google Scholar). In aseptic loosening, IL-1, IL-6, and TNF-α are present in the tissue surrounding failed implants (25.Xu J.W. Konttinen Y.T. Lassus J. Natah S. Ceponis A. Solovieva S. Aspenberg P. Santavirta S. Clin. Exp. Rheumatol. 1996; 14: 643-648PubMed Google Scholar, 26.Merkel K.D. Erdmann J.M. McHugh K.P. Abu-Amer Y. Ross F.P. Teitelbaum S.L. Am. J Pathol. 1999; 154: 203-210Abstract Full Text Full Text PDF PubMed Scopus (371) Google Scholar,50.Goodman S.B. Huie P. Song Y. Schurman D. Maloney W. Woolson S. Sibley R. J. Bone Jt. Surg. Br. 1998; 80: 531-539Crossref PubMed Scopus (223) Google Scholar). It was also reported that TNF-α promoted bone resorption by activating mature osteoclasts or by stimulating proliferation and differentiation of osteoclasts indirectly via a primary effect on osteoblasts (20.Thomson B.M. Mundy G.R. Chambers T.J. J. Immunol. 1987; 138: 775-779PubMed Google Scholar, 21.Lerner U.H. Ohlin A. J. Bone Miner. Res. 1993; 8: 147-155Crossref PubMed Scopus (49) Google Scholar, 22.van der Pluijm G. Most W. van der Wee-Pals L. de Groot H. Papapoulos S. Lowik C. Endocrinology. 1991; 129: 1596-1604Crossref PubMed Scopus (87) Google Scholar, 23.Pfeilschifter J. Chenu C. Bird A. Mundy G.R. Roodman G.D. J. Bone Miner. Res. 1989; 4: 113-118Crossref PubMed Scopus (408) Google Scholar). However, it remains unclear whether this effect is via a direct action on osteoclast progenitors because of the small population of these progenitors in bone marrow and the absence of a suitable established cell line of progenitors. Recently we established an osteoclast progenitor (MDBM cell) culture system, which was obtained by treating mouse bone marrow cells with a high dose of M-CSF. The primary question addressed in this study was whether TNF-α could directly induce the differentiation from osteoclast progenitors into mature osteoclasts. The results demonstrated that TNF-α induced TRAP-positive MNCs in the presence of M-CSF in the MDBM cell cultures. TNF-α-induced MNCs showed the phenotypical and functional characteristics of osteoclasts including the expression of TRAP, a band formation of F-actin-containing podosomes, and the ability to produce resorption pits in bone. These results indicate that TNF-α can induce osteoclastogenesis by the direct action on osteoclast progenitors in the presence of M-CSF. It is possible to consider that TNF-α reacts on contaminating stromal cells in the MDBM cell cultures, thus we examined the level of contamination of stromal cells by reverse transcriptase polymerase chain reaction analyses for type I collagen and alkaline phosphatase genes. The level of contaminating stromal cells in the preparation of MDBM cells was lower than 1 in 1,000 cells (data not shown). Moreover, OPG-Fc could not block the formation of TNF-α-induced MNCs, indicating that RANKL was not produced from stromal cells, if any, to which TNF-α may react. The results demonstrate that no effective stromal cells were present in MDBM cells. TNF-α also induced TRAP-positive MNCs in the presence of M-CSF from bone marrow cells, but the efficiency of MNC formation was markedly lower than that in the MDBM cell cultures. However, M-CSF pretreatment for 24 h increased the efficiency of TNF-α-induced MNC formation, but not that of sRANKL-induced MNC formation, suggesting that M-CSF induces the TNF-α sensitivity to osteoclast progenitors to become reactive to TNF-α. M-CSF is a key regulator in osteoclastogenesis, mediating the proliferation and differentiation of osteoclast progenitors and the survival of mature osteoclasts (51.Jimi E. Shuto T. Koga T. Endocrinology. 1995; 136: 808-811Crossref PubMed Scopus (163) Google Scholar, 52.Kodama H. Nose M. Niida S. Yamasaki A. J. Exp. Med. 1991; 173: 1291-1294Crossref PubMed Scopus (264) Google Scholar, 53.Tanaka S. Takahashi N. Udagawa N. Tamura T. Akatsu T. Stanley E.R. Kurokawa T. Suda T. J. Clin. Invest. 1993; 91: 257-263Crossref PubMed Scopus (499) Google Scholar, 54.Hofstetter W. Wetterwald A. Cecchini M.C. Felix R. Fleisch H. Mueller C. Proc. Natl. Acad. Sci. U. S. A. 1992; 89: 9637-9641Crossref PubMed Scopus (123) Google Scholar, 55.Felix R. Hofstetter W. Wetterwald A. Cecchini M.G. Fleisch H. J. Cell. Biochem. 1994; 55: 340-349Crossref PubMed Scopus (79) Google Scholar). Because a high dose of M-CSF strongly stimulated proliferation of Colony-forming unit macrophage cells in the bone marrow (data not shown), the stimulation of MNC formation might be the result of the increase in the number of Colony-forming unit macrophage cells and/or up-regulation of the sensitivity toward TNF-α in the bone marrow. In contrast, M-CSF pretreatment did not affect sRANKL-induced MNC formation. Fahlman et al. (56.Fahlman C. Jacobsen F.W. Veiby O.P. McNiece I.K. Blomhoff H.K. Jacobsen S.E. Blood. 1994; 84: 1528-1533Crossref PubMed Google Scholar) reported that TNF-α potently enhanced in vitro the differentiation of macrophages from primitive murine hematopoietic progenitor cells in combination with stem cell factor and IL-7. These reports also suggested the possibility that TNF-α but not sRANKL interferes with the M-CSF signal, which is required for commitment to the osteoclast lineage from early hematopoietic progenitor cells because of the multifunction of TNF-α toward bone marrow cells. The bone resorption activity of TNF-α-induced MNCs was approximately 10-fold lower than that of sRANKL-induced MNCs, although the number of MNCs was about half in the former. However, IL-1β stimulated the bone resorption activity of TNF-α-induced MNCs without an increase in the number of MNCs. Because IL-1β did not induce the TRAP-positive MNC formation itself in the presence of M-CSF, IL-1β compensated for the insufficient ability of TNF-α in bone resorption. It was reported that IL-1 stimulated the bone resorption activity of isolated rat osteoclasts (45.Dinarello C.A. FASEB J. 1994; 8: 1314-1325Crossref PubMed Google Scholar) or multinucleation and pit-forming activity of mouse preosteoclasts (33.Jimi E. Nakamura I. Duong L.T. Ikebe T. Takahashi N. Rodan G.A. Suda T. Exp. Cell Res. 1999; 247: 84-93Crossref PubMed Scopus (307) Google Scholar). Jimi and co-workers reported that IL-1 prolonged the survival of purified mouse osteoclasts, enhancing their bone resorption activity (33.Jimi E. Nakamura I. Duong L.T. Ikebe T. Takahashi N. Rodan G.A. Suda T. Exp. Cell Res. 1999; 247: 84-93Crossref PubMed Scopus (307) Google Scholar, 57.Thomson B.M. Saklatvala J. Chambers T.J. J. Exp. Med. 1986; 164: 104-112Crossref PubMed Scopus (333) Google Scholar). In the present study, IL-1β coupled with TNF-α also stimulated bone resorption activity. An anti-IL-1 type I receptor antibody blocked the stimulation of bone resorption activity by IL-1β but not the induction of MNCs by cotreatment with TNF-α and IL-1β. Moreover, IL-1β treatment did not affect the number of MNCs which had actin ring, demonstrating that IL-1β stimulates a specific signal required for bone resorption rather than causes the rearrangement of the cytoskeletal organization in TNF-α-induced MNCs. The type I p55 TNF receptor has been thought to be the major biologically active form (58.Vandenabeele P. Declercq W. Beyaert R. Fiers W. Trends Cell Biol. 1995; 5: 392-399Abstract Full Text PDF PubMed Scopus (744) Google Scholar), but recent evidence indicates the type II p75 TNF receptor is also functional. Whereas the p75 TNF receptor mediates differentiation of early hematopoietic precursors, the p55 TNF receptor is active in later stages of the process (56.Fahlman C. Jacobsen F.W. Veiby O.P. McNiece I.K. Blomhoff H.K. Jacobsen S.E. Blood. 1994; 84: 1528-1533Crossref PubMed Google Scholar). Soluble TNF targets primarily the p55 receptor, but the membrane-associated TNF activates both receptors (59.Heehan K.C.F. Pinckard J.K. Arthur C.D. Dehner L.P. Goeddel D.V. Schreiber R.D. J. Exp. Med. 1995; 181: 607-617Crossref PubMed Scopus (152) Google Scholar). In our blocking experiment using the specific antibody to TNF receptors, the induction of TRAP-positive MNCs by TNF-α was completely inhibited by anti-p55 antibody and partially blocked by anti-p75 antibody. Using bone marrow macrophages from TNF receptor-disrupted mice (p55 −/−, p75 −/−, and both), Abu-Ameret al. (35.Abu-Amer Y. Ross F.P. Edwards J. Teitelbaum S.L. J. Clin. Invest. 1997; 100: 1557-1565Crossref PubMed Scopus (261) Google Scholar) also showed that lipopolysaccharide-induced osteoclastogenesis was mediated by TNF and was transmitted through the p55 TNF receptor but not through the p75 one. These results suggest that at least the receptor signaling via the p55 TNF receptor is required for TNF-α-induced osteoclastogenesis of MDBM cells. The role of the signal via p75 in the TNF-α-induced osteoclastogenesis of MDBM cells should be investigated further. OPG-Fc blocked MNC induction by sRANKL but not that by TNF-α in the MDBM cell cultures, and TNF-α and sRANKL synergistically induced MNCs in the presence of M-CSF in the MDBM cell cultures (data not shown). These results suggest that the intracellular signal through TNF-α which is involved in osteoclastogenesis of MDBM cells is independent of that of RANKL. Because several stimulators of osteoclast formation, including 1α,25-(OH)2D3, parathyroid hormone, prostaglandin E2, and IL-11, up-regulated expression of RANKL mRNA in mouse calvarial osteoblasts (30.Yasuda H. Shima N. Nakagawa N. Yamaguchi K. Kinosaki M. Mochizuki S. Tomoyasu A. Yano K. Goto M. Murakami A. Tsuda E. Morinaga T. Higashio K. Udagawa N. Takahashi N. Suda T. Proc. Natl. Acad. Sci. U. S. A. 1998; 95: 3597-3602Crossref PubMed Scopus (3642) Google Scholar), TNF-α might also stimulate the RANKL gene expression in osteoblastic cells. The RANK-RANKL signals may be involved in TNF-α-induced osteoclastogenesis in the bone marrow cell cultures. Osteoclasts play a crucial role in physiological bone remodeling (4.Kahn A.J. Partridge N.C. Hall B.K. Bone. 2. CRC Press, Boca Raton, FL1991: 119Google Scholar) and also function in the local bone destruction that occurs in association with chronic inflammatory diseases (5.Mills B.G. Hall B.K. Bone. 2. CRC Press, Boca Raton, FL1991: 175Google Scholar). RANKL is essential for osteoclast development and plays a critical role in physiological bone remodeling (32.Kong Y.-Y. Yoshida H. Sarosi I. Tan H.-L. Timms E. Capparelli C. Morony S. Oliveira-dos-Santos A.J. Van G. Itie A. Khoo W. Wakeham A. Dunstan C.R. Lacey D.L. Mak T.W. Boyle W.J. Penninger J.M. Nature. 1999; 397: 315-323Crossref PubMed Scopus (2932) Google Scholar). The RANK-RANKL signal is involved in local osteolysis in chronic inflammatory diseases; however, it remains unclear whether it is the absolute pathway. Diseases such as rheumatoid arthritis, aseptic loosening, and periodontal diseases have been associated with accumulation of inflammatory cytokines, which likely mediate local bone destruction by stimulating osteoclast activity (24.Saxne T. Palladino Jr., M.A. Heinegard D. Talal N. Wollheim F.A. Arthritis Rheum. 1988; 31: 1041-1045Crossref PubMed Scopus (569) Google Scholar, 25.Xu J.W. Konttinen Y.T. Lassus J. Natah S. Ceponis A. Solovieva S. Aspenberg P. Santavirta S. Clin. Exp. Rheumatol. 1996; 14: 643-648PubMed Google Scholar, 26.Merkel K.D. Erdmann J.M. McHugh K.P. Abu-Amer Y. Ross F.P. Teitelbaum S.L. Am. J Pathol. 1999; 154: 203-210Abstract Full Text Full Text PDF PubMed Scopus (371) Google Scholar, 27.Nair S.P. Meghji S. Wilson M. Reddi K. White P. Henderson B. Infect. Immun. 1996; 64: 2371-2380Crossref PubMed Google Scholar). We showed here for the first time that, first, TNF-α induces osteoclastogenesis by a direct action on osteoclast progenitors in the presence of M-CSF; second, the receptor signal mediated by the p55 TNF receptor is mainly involved in its effect; and third, IL-1 stimulates the bone resorption activity of TNF-α-induced osteoclasts. Our results suggest that the direct action of TNF-α and M-CSF on osteoclast progenitors contributes to osteoclast formation and local osteolysis in some pathological conditions such as rheumatoid arthritis and aseptic loosening. In conclusion, TNF-α is a crucial differentiation factor for osteoclasts. The present results strongly suggest that TNF-α and M-CSF play an important role in local osteolysis in chronic inflammatory diseases. We acknowledge gratefully the help of Dr. T. Ohta and Dr. K. Komoriya of Teijin Institute for Biomedical Research.

Referência(s)