Treg cells suppress osteoclast formationA new link between the immune system and bone.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 56, No. 12, December 2007, pp 4104–4112 DOI 10.1002/art.23138 © 2007, American College of Rheumatology Treg Cells Suppress Osteoclast Formation A New Link Between the Immune System and Bone Mario M. Zaiss,1 Roland Axmann,1 Jochen Zwerina,1 Karin Polzer,1 Eva Gückel,1 Alla Skapenko,2 Hendrik Schulze-Koops,2 Nikki Horwood,3 Andrew Cope,3 and Georg Schett1 Objective. To investigate whether Treg cells can suppress osteoclast differentiation, and to define a new potential link between the immune system and the skeleton. Methods. Regulatory CD4ⴙ,CD25ⴙ,Foxp3ⴙ T cells were isolated and purified from the spleen and cocultured with CD11bⴙ osteoclast precursor cells isolated from bone marrow. Osteoclastogenesis and bone erosion were assessed by tartrate-resistant acid phosphatase staining and pit resorption assay, respectively. In addition, Transwell experiments and cytokineblocking experiments were performed to define the mechanisms of interaction between Treg cells and osteoclasts. Results. CD4ⴙ,CD25ⴙ,Foxp3ⴙ T cells, but not CD4ⴙ,CD25ⴚ T cells, dose dependently inhibited macrophage colony-stimulating factor– and RANKLdependent osteoclast formation. Pit formation was in- hibited by up to 80% when Treg cells were added. The blockade of osteoclast formation was not based on the alteration of RANKL/osteoprotegerin balance but was essentially dependent on direct cell–cell contact via CTLA-4. Treg cell–mediated expression of transforming growth factor ␤, interleukin-4 (IL-4), and IL-10 contributed but was not essential to the inhibitory effect on osteoclastogenesis. Conclusion. These data show that CD4ⴙ,CD25ⴙ,Foxp3ⴙ Treg cells suppress osteoclast formation, provide a new link between the immune system and bone, and extend our knowledge on regulation of bone homeostasis by the immune system. Osteoclasts are multinucleated cells that resorb bone (1). They originate from cells of hematopoietic lineage, in particular from monocytes; upon challenge with specific signals, they undergo a series of differentiation steps to become mature osteoclasts. During the differentiation process, osteoclasts acquire specific markers such as tartrate-resistant acid phosphatase (TRAP), fuse to multinucleated giant cells, and polarize upon contact with bone. Essential signals for osteoclast differentiation are macrophage colony-stimulating factor (M-CSF) and RANKL. In a concerted action, these signals drive monocytes to become osteoclasts, and the absence of either one of these signals is sufficient to block osteoclast formation completely. This is highlighted by the osteopetrotic phenotype of op/op mice as well as that of RANKL⫺/⫺ mice (2,3). Osteoclastogenesis is highly dependent on the cellular microenvironment, which provides essential signals such as M-CSF and RANKL as well as costimulatory signals such as proinflammatory cytokines. Mesenchymal cells such as preosteoblasts or activated synovial fibroblasts, which are in close connection to cells of the osteoclast lineage, express M-CSF and RANKL and can Supported in part by an intramural ELAN grant from the University of Erlangen-Nuremberg and by training grant GK592 from the German Research Community (DFG) and the Sonderforschungsbereich 643 Project B08. Dr. Schett’s work was supported by a START Program award from the Austrian Ministry of Sciences. 1 Mario M. Zaiss, BSc, Roland Axmann, MD, Jochen Zwerina, MD, Karin Polzer, BSc, Eva Gückel, BSc, George Schett, MD: University of Erlangen-Nuremberg, Erlangen, Germany; 2Alla Skapenko, PhD, Hendrik Schulze-Koops, MD: University of Erlangen-Nuremberg, Erlangen, Germany, and Ludwig-Maximilian University of Munich, Munich, Germany; 3Nikki Horwood, MD, Andrew Cope, MD: Kennedy Institute of Rheumatology, London, UK. Dr. Schett has received consulting fees, speaking fees, and/or honoraria (less than $10,000 each) from Amgen, Schering-Plough, Abbott, UCB, and Roche. Address correspondence and reprint requests to Georg Schett, MD, Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nuremberg, Krankenhausstrasse 12, D-91054 Erlangen, Germany. E-mail: georg.schett@ uni-erlangen.de. Submitted for publication April 25, 2007; accepted in revised form August 31, 2007. 4104 ROLE OF REGULATORY T CELLS IN OSTEOCLAST FORMATION drive osteoclast formation (4,5). Stimulation of osteoclasts by cells such as preosteoblasts reflects the crosstalk of mesenchymal bone-forming cells with hematopoietic bone-resorbing cells, which is the principle of bone turnover. The other major cell type that stimulates osteoclast formation is T lymphocytes (6,7). T cell interaction with osteoclasts somehow reflects the immunologic principle of the crosstalk between T cells and antigen-presenting cells, which are of monocyte origin. Cells of monocyte lineage have a certain plasticity to differentiate into macrophages, myeloid dendritic cells, microglia, or osteoclasts, which depends on the cytokine milieu to which they are exposed (8). Stimulation of osteoclasts by activated Th1 cells has revolutionized the understanding of bone loss in immune-mediated inflammation, which is a hallmark of virtually every chronic inflammatory disease. Thus, conditions such as rheumatic diseases and inflammatory bowel diseases are characterized by an increased level of osteoclast formation (9,10). The fact that activated T cells express RANKL links immune activation with osteoclastogenesis and bone resorption. This concept has been further strengthened by showing that a certain inflammatory T cell subset, the interleukin-17 (IL-17)– producing Th17 cells, are also potent activators of osteoclast formation (7). Based on the fact that inflammatory T cell subsets drive osteoclast formation, we hypothesized that Treg cells might exert a different effect. We thus investigated the effects of CD25⫹,Foxp3⫹ Treg cells on osteoclast formation (11,12). Our experiments showed that Treg cells suppress osteoclast formation, thereby providing a new link between the immune system and bone. MATERIALS AND METHODS Isolation and characterization of Treg cells. Spleens were isolated from 5-week-old female C57BL/6 mice and homogenized through 70-m stainless steel mesh to archive a single-cell suspension. Erythrocytes in the cell suspension were lysed by NH4Cl treatment. CD25⫹ and CD25⫺,CD4⫹ T cell populations were isolated from the splenic cell suspension using the microbead-based Regulatory T Cell Isolation Kit (Miltenyi Biotec, Bergisch Gladbach, Germany), according to the manufacturer’s instructions. The purity of isolated populations was analyzed by flow cytometry. Isolated T cells were stained with fluorescein isothiocyanate (FITC)–conjugated anti-CD4 and phycoerythrin (PE)–conjugated anti-CD25 (BD Biosciences PharMingen, Hamburg, Germany). Sorted CD4⫹,CD25⫹ T cells were labeled with antibodies to CD4– FITC and CD25–PE and isolated by cell sorting using a MoFlo cytometer (DakoCytomation, Hamburg, Germany). 4105 For fluorescence-activated cell sorting analysis, sorted CD4⫹,CD25⫹ T cells were additionally intracellularly stained with streptavidin–Cy5–conjugated anti-Foxp3 (The Jackson Laboratory, Bar Harbor, ME). The functionality of isolated CD4⫹,CD25⫹ T cells was assessed by proliferation assay, as follows: CD25⫹ and CD25⫺ T cells (5 ⫻ 104/well) were cocultured or separately cultured for 3 days in 96-well roundbottomed plates (Corning, Corning, NY) in the presence of anti-CD3 monoclonal antibody (5 g/ml) (eBioscience, San Diego, CA) and 1 ⫻ 105 irradiated antigen-presenting cells. Incorporation of 3H-thymidine (1 Ci/well) during the last 18 hours of culture was measured by a liquid scintillation counter (1205 Betaplate; Wallac Pharmacia, Gaithersburg, MD). Isolation and culture of osteoclasts. Bone marrow was isolated from 5-week-old female C57BL/6 mice by flushing femoral bones with medium. Erythrocytes in the cell suspension were lysed as described above, and monocytes were then isolated from bone marrow–derived cell suspensions using CD11b microbeads (Miltenyi Biotec). The purity of isolated monocytes was controlled by flow cytometry using CD11b– FITC–labeled antibodies (Miltenyi Biotec). Monocytes were plated in 96-well plates (2.5 ⫻ 105/well) in the presence of 30 ng/ml M-CSF and 50 ng/ml RANKL in minimum essential medium (Gibco, Karlsruhe, Germany) supplemented with 10% heat-inactivated fetal calf serum (Gibco) and 1% penicillin/streptomycin (R&D Systems, Wiesbaden, Germany). Osteoclast differentiation was evaluated by staining for TRAP using a leukocyte acid phosphatase kit (Sigma-Aldrich, Poole, UK). Bone resorption was assessed on BioCoat Osteologic hydroxyapatite-coated slides (BD Biosciences PharMingen). Coculture of Treg cells and osteoclasts. Purified CD11b⫹ monocytes (2.5 ⫻ 105/well) and different numbers of activated CD4⫹,CD25⫹ T cells (5 ⫻ 104/well to 5 ⫻ 103/well) were cocultured in 96-well plates in the presence of 30 ng/ml M-CSF and 50 ng/ml RANKL (R&D Systems) for 4 days. T cells were activated with a soluble anti-CD3 monoclonal antibody (5 g/ml) for 1 hour before adding them to the cocultures. There was no difference between the use of platebound or soluble anti-CD3 antibody in the experiments. For Transwell experiments, a similar approach was used. Briefly, 96-well plates (0.4-m pore size; Corning) were loaded with 2.5 ⫻ 105 CD11b⫹ monocytes into the lower chambers and with various numbers of CD4⫹,CD25⫹ Treg cells into the upper chambers. Culture was performed in the presence of 30 ng/ml M-CSF and 50 ng/ml RANKL for 4 days. Osteoclast differentiation was evaluated by staining for TRAP using a leukocyte acid phosphatase kit (Sigma-Aldrich). In addition, calcified matrix resorption activity of the osteoclasts was tested on calcium hydroxyapatite–coated slides (BioCoat Osteologic; BD Biosciences), using a coculture setting identical to that described above. Von Kossa’s stain was used to visualize the resorption pits, and analysis was performed using an inverted phase-contrast microscope (Nikon, Melville, NY). Inhibition studies. For neutralization experiments, blocking monoclonal antibodies against IL-4 receptor (IL-4R) (clone mIL4R-M1; BD Biosciences), IL-10R (clone 1B1.3a; BD Biosciences), transforming growth factor ␤ receptor type II (TGF␤RII) (synthetic peptide specific for p75 TGF␤RII; Abcam, Cambridge, UK), and polyclonal antibody against CD152 (CTLA-4; Abcam), as well as recombinant CD152 4106 ZAISS ET AL Figure 1. Isolation, purification, and characterization of Treg cells. A, Scatter plot analysis of spleen cells from C57BL/6 mice (left), and fluorescence-activated cell sorting (FACS) analysis of unstained cells for gate validation of CD4⫹,CD25⫺ labeling (right). B, Dot plot analysis of FACS scan with labeling for CD4 and CD25 after isolation of cells with the respective microbeads (left), and labeling for Foxp3 and counterstaining for CD25 (right). C, Expression of Foxp3 mRNA in CD4⫹,CD25⫺ and CD4⫹,CD25⫹ T cells by real-time quantitative polymerase chain reaction, with ␣-actin as a reference gene. Mean and SD results from triplicate reactions are shown. D, Immunoblot of Foxp3 protein expression by CD4⫹,CD25⫺ and CD4⫹,CD25⫹ T cells. E, Analysis of suppressor activity of CD4⫹,CD25⫹ T cells in cocultures with CD4⫹,CD25⫺ responder T cells. The proliferation response was measured by 3H-thymidine incorporation. Values are the mean and SD results from triplicate cultures and are representative of 3 independent experiments. ⴱ ⫽ P ⬍ 0.001. SSC-H ⫽ side scatter height; FSC-H ⫽ forward scatter height. (CTLA-4) murine immunoglobulin (RDI, Flanders, NJ), were added to CD11b⫹ monocytes 1 hour before coculture with CD4⫹,CD25⫹ T cells. Osteoclast differentiation was evaluated by staining for TRAP, using a leukocyte acid phosphatase kit (Sigma-Aldrich). Real-time polymerase chain reaction (PCR) from monocytes. Real-time PCR was performed as follows: for Foxp3, sense 5⬘-AGGAGCCGCAAGCTAAAAGC-3⬘ and anti-sense 5⬘-TGCCTTCGTGCCCACTGT-3⬘; for RANKL, sense 5⬘-GAATCCTGAGACTCCATGAAAACG-3⬘ and antisense 5⬘-CCATGAGCCTTCCATCATAGCTGG-3⬘. Realtime PCR from monocytes after 4 days of cultivation, cocultured with or without purified CD4⫹,CD25⫹ T cells, was done for osteoclast gene expression, as follows: for matrix metalloproteinase 9 (MMP-9), sense 5⬘-CATTCGCGTGGATAAGGA-3⬘ and antisense 5⬘-TCACACGCCAGAAGAATTTG3⬘; for TRAP, sense 5⬘-CGACCATTGTTACCACATACG-3⬘ and antisense 5⬘-TCGTCCTGAAGATACTGCAGGTT-3⬘; for nuclear factor of activated T cells c1 (NFATc1), sense 5⬘-CCCGTTGCTTCCAGAAAATA-3⬘ and antisense 5⬘TCACCCTGGTGTVTCTTCCTC-3⬘; for osteoclastassociated receptor (OSCAR), sense 5⬘-TCGCTGATACTCC- AGCTGTC-3⬘ and antisense 5⬘-ATCCCAGGAGTCACAACTGC-3⬘; for cathepsin K, sense 5⬘-ATATGTGGGCCAGGATGAAAGTT-3⬘ and antisense 5⬘-TCGTTCCCCACAGGAATCTCT-3⬘; and for osteoprotegerin (OPG), sense 5⬘-AGCTGCTGAAGCTGTGGAA-3⬘ and antisense 5’GGTTCGAGTGGCCGAGAT-3⬘. As a reference gene, ␣-actin (sense 5⬘-TGTCCACCTTCCAGCAGATGT-3⬘ and antisense 5⬘-AGCTCAGTAACAGTCCGCCTAGA-3⬘) was used. Normalized gene expression was derived from the ratio of messenger RNA (mRNA) expression from the gene of interest to ␣-actin mRNA expression in each sample. Western blot analysis. Spleens were isolated from C57BL/6 (wild-type) mice, and single-cell suspensions were generated. CD4⫹,CD25⫺ and CD4⫹,CD25⫹ T cell populations were isolated from the splenic cell suspensions using the microbead-based Regulatory T Cell Isolation Kit (Miltenyi Biotec), as described above. T cell populations were lysed and separated on sodium dodecyl sulfate–polyacrylamide electrophoresis gels, transferred to nitrocellulose membranes, and stained with anti-mouse Foxp3 antibody (clone FJK-16s; eBioscience) and polyclonal rabbit anti-rat immunoglobulins (DakoCytomation). ROLE OF REGULATORY T CELLS IN OSTEOCLAST FORMATION 4107 Figure 2. Inhibition of osteoclast formation and bone resorption by Treg cells. A, Fluorescence-activated cell sorting analysis of purified bone marrow–derived monocytes after isolation by their expression of CD11b, via microbeads. B, Quantification of the results of coculture experiments with osteoclast precursors and various numbers of activated CD4⫹,CD25⫹ Treg cells or activated controls (CD4⫹,CD25⫺ T cells). C, Photomicrographs showing dose-dependent suppression of osteoclastogenesis by CD4⫹,CD25⫹ Treg cells activated with anti-CD3 antibody. Osteoclasts are labeled for tartrate-resistant acid phosphatase (TRAP) and appear as purple-colored multinucleated cells (original magnification ⫻ 20). D, Pit formation assay indicating the number of erosive sites upon coculture of osteoclast precursors with various numbers of CD4⫹,CD25⫹ Treg cells or controls (CD4⫹,CD25⫺ T cells). Values in B and D are the mean and SD results from triplicate cocultures and are representative of 4 independent experiments. FITC ⫽ fluorescein isothiocyanate; FSC-H ⫽ forward scatter height. ⴱ ⫽ P ⬍ 0.05 versus other doses and versus control. Enzyme-linked immunosorbent assay (ELISA). CD25⫹ and CD25⫺,CD4⫹ T cell populations were activated in 96-well plates precoated with anti-CD3e monoclonal antibody (5 g/ml; eBioscience) overnight. For detection of OPG, supernatants were harvested after 24 hours of culture, and OPG content was measured by a quantitative sandwich ELISA for murine OPG (R&D Systems), according to the manufacturer’s instructions. Statistical analysis. Results were analyzed by one-way analysis of variance followed by Tukey’s test or the Holm-Sidak test. RESULTS Isolation and characterization of Treg cells and osteoclast precursors. To allow coculture of Treg cells with osteoclasts, we isolated CD4⫹,CD25⫹ T cells from the spleens of healthy C57BL/6 mice. Spleens contained a significant proportion of CD4⫹ as well as CD8⫹ T cells. After isolation, the purity of CD4⫹,CD25⫹ T cells was ⬎90%, as shown in Figures 1A and B. In addition, permeabilized cells were assessed for the expression of the transcription factor Foxp3, which was high in ⬎90% of CD4⫹,CD25⫹ T cells (Figure 1B). Moreover, we analyzed the mRNA expression of Foxp3, showing that CD4⫹,CD25⫹ T cells, but not CD4⫹,CD25⫺ T cells, express large amounts of Foxp3 (Figure 1C). Similarly, immunoblot analysis showed expression of Foxp3 in CD4⫹,CD25⫹ T cells but only very weak expression in CD4⫹,CD25⫺ T cells (Figure 1D). Next, we analyzed the functional aspect of isolated Treg cells and showed that these cells suppress proliferation of CD4⫹,CD25⫺ T cells in coculture experiments, with results of proliferation assays identifying Treg cells as functionally competent (Figure 1E). Suppression of osteoclast formation by Treg cells. Osteoclast precursors were isolated from the bone marrow of the same mice and were identified according 4108 ZAISS ET AL Figure 3. Specific inhibition of osteoclast formation and expression of osteoclast-associated genes by activated Treg cells. A, Real-time polymerase chain reaction analysis of the expression of mRNA for matrix metalloproteinase 9 (MMP-9), tartrate-resistant acid phosphatase (TRAP), nuclear factor of activated T cells 1c (NFAT1c), osteoclast-associated receptor (OSCAR), cathepsin K (CK), and osteoprotegerin (OPG) in monocyte cultures exposed to macrophage colony-stimulating factor (M-CSF) and RANKL for 4 days and cocultivated with CD4⫹,CD25⫹ T cells (open bars) and CD4⫹,CD25⫺ T cells (solid bars). B, Osteoclast counts after monocytes were cocultured with activated or nonactivated CD4⫹,CD25⫹ T cells in the presence of M-CSF/RANKL. C, Osteoclast counts after monocytes were cocultured with activated CD4⫹ T cells (solid bars) or CD4⫹,CD25⫹ Treg (open bars) in the presence of M-CSF/RANKL. Control bars in B or C represent monocytes cocultured with irradiated CD4⫹,CD25⫺ T cells at a ratio of 1:5. Values are the mean and SD results from 3 independent experiments. to their light-scatter pattern and isolated by their expression of CD11b, which has been described as a marker for the monocyte population capable of differentiating into osteoclasts (Figure 2A). Next, we investigated whether CD4⫹,CD25⫹ Treg cells could suppress osteoclast formation. In the absence of any T cell subset, osteoclast formation was induced by cultivating CD11b⫹ osteoclast precursors with M-CSF and RANKL, leading to the formation of multinucleated TRAP-positive osteoclasts (Figure 2B). The addition of anti-CD3–activated CD4⫹,CD25⫹ Treg cells to the culture system dose dependently suppressed osteoclast formation. Even small numbers of Treg cells were sufficient to blunt osteoclast formation significantly, whereas maximal inhibition virtually completely abolished the formation of osteoclasts (Figures 2B and C). In contrast, addition of CD4⫹,CD25⫺ T cells did not affect osteoclast formation. Representative photomicrographs showing dose- dependent suppression of osteoclastogenesis by CD4⫹,CD25⫹ Treg cells activated by anti-CD3 antibody are presented in Figure 2C. Differentiated osteoclasts appeared as multinucleated giant cells that specifically express TRAP, whereas osteoclast precursors are mononuclear TRAP-positive cells. The addition of Treg cells dramatically inhibited formation of these cellular structures in a dose-dependent manner. We then investigated the functional consequences of impaired osteoclast formation on bone resorption. Whereas resorption pits were observed upon differentiation of osteoclasts, pit formation was strongly diminished by up to 80% when Treg cells were added (Figure 2D). This suggests that CD4⫹,CD25⫹ T cells suppress not only osteoclast formation but also functional bone resorption. Profiling of mRNA expression of osteoclastassociated genes showed significantly impaired expression of TRAP, OSCAR, cathepsin K, MMP-9, and ROLE OF REGULATORY T CELLS IN OSTEOCLAST FORMATION 4109 Figure 4. Cell contact–dependent inhibition of osteoclast formation by Treg cells. A, Relative expression of RANKL mRNA in CD4⫹,CD25⫹ Treg cells and CD4⫹,CD25⫺ T cells (by real-time quantitative polymerase chain reaction); ␣-actin was used as a reference gene. Normalized RANKL expression was derived from the ratio of RANKL mRNA expression to ␣-actin mRNA expression in each sample. Values are the mean results from triplicate reactions. B, Quantitative evaluation of osteoclast formation in Transwell plates, by coculturing osteoclast precursors and CD4⫹,CD25⫹ Treg cells and CD4⫹,CD25⫺ T cells in different chambers. Values are the mean and SD results from triplicate cocultures. C, Quantitative evaluation of osteoclast formation after neutralization experiments with antibodies (ab) against interleukin-4 receptor (IL-4R), IL-10R, and transforming growth factor ␤ receptor type II (TGF␤RII) in cocultures of osteoclast precursors with CD4⫹,CD25⫹ Treg cells. Values are the mean and SD results from 3 independent experiments. ⴱ ⫽ P ⬍ 0.01 versus no Tregs; ⴱⴱ ⫽ P ⬍ 0.01 versus antibody-treated groups. See Figure 3 for other definitions. NFATc1 in osteoclast precursor cells exposed to Treg cells (Figure 3A). This supports the findings that activated Treg cells suppress the differentiation of monocytic osteoclast precursors to osteoclasts. In contrast to activated Treg cells, nonactivated CD4⫹,CD25⫹ T cells failed to inhibit osteoclast formation except when they were added in very high concentrations (Figure 3B). There was a 20-fold difference in the inhibitory potential between nonactivated and activated Treg cells. Activated Treg cells at concentrations up to 1 per 100 osteoclast precursors inhibited osteoclast formation, suggesting that these cells are very potent in their inhibitor effect on the osteoclast. Unfractionated CD4 cells inhibited osteoclast formation only at high concentrations (1 per 5 osteoclast precursors), which exactly reflects the 5% fraction of Treg cells present within the CD4⫹ T cell pool (Figure 3C). Cell contact–dependent suppression of osteoclast formation by Treg cells. Considering the consistent inhibitory effect of Treg cells on osteoclast formation, we questioned the mechanisms by which these cells affect osteoclastogenesis. Interestingly, expression of RANKL was even higher in CD4⫹,CD25⫹ Treg cells than in CD4⫹,CD25⫺ T cells and was only partly compensated for by higher expression of OPG in these cells (Figure 4A). This suggests that Treg cells suppress osteoclast formation without interfering with the RANKL/RANK system, because increased RANKL expression would enhance rather than suppress osteoclast formation, which is not reflected by the functional data on osteoclast formation. We next addressed whether the suppressive effect is cell-contact dependent, using Transwell experiments. Importantly, the suppressive effect of Treg cells on osteoclast formation was completely abolished when cells were not allowed to interact directly by cell–cell contact (Figure 4B). This indicates that direct cell–cell contact is an essential prerequisite for Treg cells to inhibit osteoclastogenesis. To search for additional cytokine-mediated inhibitory effects, we next attempted to block osteoclast differentiation by neutralizing TGF␤, IL-4, and IL-10 (Figure 4C). The addition of any one of these antibodies partly rescued, to a significant level (P ⬍ 0.05), osteoclast formation in the presence of T cells. However, full rescue could not be achieved, suggesting that contactdependent mechanisms are crucial. Suppression of osteoclast formation by CTLA-4. Based on the pivotal role of cell contact in the suppressive effects of Treg cells on osteoclast formation, we hypothesized that CTLA-4 might be involved in this process. The addition of CTLA-4 to osteoclast precursor cells dose dependently suppressed osteoclast formation 4110 ZAISS ET AL Figure 5. Role of CTLA-4 in the suppressive effect of CD4⫹,CD25⫹ T cells on osteoclast formation. A, Dose-dependent suppression of osteoclast formation by CTLA-4. Different concentrations of CTLA-4 and anti–CTLA-4 antibody (ab) were added to purified bone marrow–derived monocytes stimulated with macrophage colony-stimulating factor (M-CSF)/RANKL. Values are the mean and SD results from 3 independent experiments. B, Representative photomicrographs of tartrate-resistant acid phosphatase (TRAP)–stained cultures of monocytes cultivated with M-CSF and RANKL as well as various doses of CTLA-4 and anti–CTLA-4 antibody (original magnification ⫻ 20). (Figure 5A). This process could be completely reversed by anti–CTLA-4 antibodies, regardless of whether CTLA-4 immunoglobulin or Treg cells were used to suppress osteoclast formation, indicating that the effect on osteoclast precursors is mediated by CTLA-4. Figure 5B shows representative results of osteoclast assays, indicating dose-dependent inhibition of osteoclast formation by CTLA-4 as well as its reversal by anti– CTLA-4 antibodies. DISCUSSION In this study, we describe a new link between the immune system and bone that addresses the regulatory site of interaction between these 2 organ systems. We show that regulatory CD4⫹,CD25⫹,Foxp3⫹ T cells suppress osteoclast formation in a cell contact– dependent manner, which suggests that these cells have a key function in terminating osteoclast-mediated bone resorption. This contrasts with the key role of inflammatory T cell subsets in stimulating osteoclast formation through expression of RANKL. Inflammation and bone loss are 2 frequently occurring and tightly linked disorders. The failure of the body to control inflammatory responses leads to profound bone loss, which increases skeletal fragility. All prototypes of inflammatory diseases such as rheumatoid arthritis, systemic lupus erythematosus, and inflammatory bowel disease are characterized by dramatic skeletal damage, which contributes to a high burden of disease due to fractures, bone erosion, functional disability, and even crippling (9,10). Bone loss is partly controlled by the hematopoietic cellular compartment, because osteoclasts, which are the primary bone-resorbing cells, emerge from monocyte precursors. Immune activation and bone resorption share several features. First, recruitment of immune cells from the bone marrow to sites of inflammation also affects the recruitment of osteoclast precursors to these sites. Importantly, tumor necrosis factor (TNF) is one of the key regulatory molecules for the trafficking of osteoclast precursors to these sites of inflammation (13). Second, inflammatory cytokines induce osteoclast differentiation, and many of them act via inflammatory T cell subsets, which drives osteoclast formation through the expression of M-CSF and RANKL (14). This interaction between T cells and osteoclast precursors and osteoclasts mimics the interaction of T cells with antigenpresenting cells. Finally, cytokines that have regulatory properties in inflammatory responses, such as TGF␤, IL-4, and IL-10, also negatively regulate osteoclast formation, suggesting that the negative regulation of immune activation follows principles similar to those involved in the regulation of bone resorption (15). Considering the intensive crosstalk between the immune system and mechanisms of bone resorption, it is of key interest whether bone resorption is subjected to direct negative regulation by mechanisms that regulate inflammatory responses. Considering the facts that severe bone loss develops only in the setting of uncontrolled chronic inflammatory diseases and that shortlived and self-limited immune activation does not entail pronounced skeletal effects, a direct link between anti- ROLE OF REGULATORY T CELLS IN OSTEOCLAST FORMATION 4111 Figure 6. Inhibition of osteoclast formation by Treg cells. Th17 cells and Th1 cells activate osteoclast formation by expressing RANKL, which engages RANK on the surface of mononuclear osteoclast (MO) precursors. Cytokines produced by Th17 cells (interleukin-17 [IL-17]) and Th1 cells (tumor necrosis factor) support expression of RANKL and osteoclast formation. Treg cells inhibit osteoclast formation by expressing CTLA-4, which binds to B7-1 and B7-2 on the surface of mononuclear osteoclast precursors, impairing their differentiation to osteoclasts. Cytokines such as IL-4, IL-10, and transforming growth factor ␤ (TGF␤) support this antiosteoclastogenic effect. inflammatory and antierosive mechanisms appears conceivable and is an attractive hypothesis in terms of understanding how bone erosion can be inhibited. For several reasons, Treg cells appear to be an attractive target cell that could represent a missing link between the immune system and bone. First, Treg cells are the best-defined cellular candidate for tearing down immune activation, and loss of these cells in mice leads to severe inflammatory disease. Next, as members of the T cells lineage, Treg cells can engage with osteoclasts and mimic the interaction between activated T cells and osteoclasts. Third, Treg cells express cytokines such as TGF␤, IL-4, and IL-10, which not only have antiinflammatory properties but also impair osteoclast formation. Our results clearly show that CD4⫹,CD25⫹,Foxp3⫹ Treg cells suppress osteoclast formation in a cell contact–dependent manner. In contrast to recent data (16), the suppressive effect of Treg cells on osteoclasts clearly depends on cell contact, and regulatory cytokines expressed by Treg cells do modulate this effect but do not substitute for direct cell contact. The suppressive effect of Treg is not based on a specific change in RANKL expression or OPG production by these cells, suggesting that interference with the RANKL–RANK interaction is not the cause of inhibition of osteoclast formation. Cell contact appears to be an essential factor in the suppression of osteoclasts by Treg cells. The molecular mechanism is based on CTLA-4, which is highly expressed by Treg cells and binds to monocytes through B7-1 and B7-2. CTLA-4 impairs osteoclast formation by binding to osteoclast precursors and inhibiting their differentiation in a dose-dependent manner. Moreover, neutralization of CTLA-4 completely restored the suppressive effect of Treg cells on osteoclasts. The fact that a single Treg cell is sufficient to block differentiation of 100 monocytic cells into the osteoclast lineage underlines the powerful role of this cell lineage in the regulation of bone resorption. The strength of this suppressive effect is presumably accomplished by cytokines, which add to the suppressive potential of Treg cells on osteoclast formation. In contrast to inflammatory Th1 and Th17 T cell subsets (Figure 6), which support osteoclast formation by cytokines such as IL-17 and TNF␣, the cytokine repertoire of Treg cells is clearly antiosteoclastogenic (6,7,14,15). IL-4 as well as IL-10 have been identified as inhibitors of osteoclast formation, and TGF␤ affects osteoblast rather than osteoclast activity. Thus, the 4112 ZAISS ET AL cytokine pattern produced by Treg cells is clearly supportive to allow suppression of osteoclast differentiation. In summary, our data show that a central regulatory mechanism of the immune system can inhibit osteoclast formation. This provides a new link between the immune system and bone and points out the cellular basics for a coordinated down-regulation of immune activation and bone resorption. This principle appears to be conceivable, because activation of the immune system and activation of bone resorption may go hand in hand. AUTHOR CONTRIBUTIONS Dr. Schett had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study design. Zaiss, Schulze-Koops, Cope, Schett. Acquisition of data. Zaiss, Zwerina, Gückel, Schett. Analysis and interpretation of data. Zaiss, Polzer, Skapenko, Schett. Manuscript preparation. Zaiss, Norwood, Schett. Statistical analysis. Zaiss, Axmann, Schett. REFERENCES 1. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000;289: 1504–8. 2. Yoshida H, Hayashi S, Kunisada T, Ogawa M, Nishikawa S, Okamura H, et al. The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene. Nature 1990;345:442–4. 3. Lacey DL, Timms E, Tan HL, Kelley MJ, Dunstan CR, Burgess T, et al. Osteoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 1998;93:165–76. 4. Gravallese EM, Manning C, Tsay A, Naito A, Pan C, Amento E, et al. Synovial tissue in rheumatoid arthritis is a source of osteoclast differentiation factor. Arthritis Rheum 2000;43:250–8. 5. Shigeyama Y, Pap T, Kunzler P, Simmen BR, Gay RE, Gay S. Expression of osteoclast differentiation factor in rheumatoid arthritis. Arthritis Rheum 2000;43:2523–30. 6. Kong YY, Feige U, Sarosi I, Bolon B, Tafuri A, Morony S, et al. Activated T cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 1999; 402:304–9. 7. Sato K, Suematsu A, Okamoto K, Yamaguchi A, Morishita Y, Kadono Y, et al. Th17 functions as an osteoclastogenic helper T cell subset that links T cell activation and bone destruction. J Exp Med 2006;203:2673–82. 8. Servet-Delprat C, Arnaud S, Jurdic P, Nataf S, Grasset MF, Soulas C, et al. Flt3⫹ macrophage precursors commit sequentially to osteoclasts, dendritic cells and microglia. BMC Immunol 2002;3: 15. 9. Schett G, Redlich K, Smolen JS. Inflammation-induced bone loss in the rheumatic diseases. In: Favus MJ, editor. Primer on the metabolic bone diseases and disorders of mineral metabolism. Sixth ed. Washington, DC: American Society for Bone and Mineral Research; 2006. p. 310–3. 10. Goldring SR. Inflammatory mediators as essential elements in bone remodeling. Calcif Tissue Int 2003;73:97–100. 11. Brunkow ME, Jeffery EW, Hjerrild KA, Paeper B, Clark LB, Yasayko SA, et al. Disruption of a new forkhead/winged-helix protein, scurfin, results in the fatal lymphoproliferative disorder of the scurfy mouse. Nat Genet 2001;27:68–73. 12. Sakaguchi S, Ono M, Setoguchi R, Yagi H, Hori S, Fehervari Z, et al. Foxp3⫹ CD25⫹ CD4⫹ natural regulatory T cells in dominant self-tolerance and autoimmune disease. Immunol Rev 2006;212: 8–27. 13. Li P, Schwarz EM, O⬘Keefe RJ, Ma L, Boyce BF, Xing L. RANK signaling is not required for TNF-mediated increase in CD11hi osteoclast precursors but is essential for mature osteoclast formation in TNF␣-mediated inflammatory arthritis. J Bone Min Res 2004;9:207–13. 14. Lam J, Takeshita S, Barker JE, Kanagawa O, Ross FP, Teitelbaum SL. TNF-␣ induces osteoclastogenesis by direct stimulation of macrophages exposed to permissive levels of RANK ligand. J Clin Invest 2000;106:1481–8. 15. Wei S, Wang MW, Teitelbaum SK, Ross FP. Interleukin-4 reversibly inhibits osteoclastogenesis via inhibition of NF-B and mitogen-activated protein kinase signaling. J Biol Chem 2002;277: 6622–30. 16. Kim YG, Lee CK, Nah SS, Mun SH, Yoo B, Moon HB. Human CD4⫹CD25⫹ regulatory T cells inhibit the differentiation of osteoclasts from peripheral blood mononuclear cells. Biochem Biophys Res Commun 2007;357:1046–52.