SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Objective

Immune activation triggers bone loss. Activated T cells are the cellular link between immune activation and bone destruction. The aim of this study was to determine whether immune regulatory mechanisms, such as naturally occurring Treg cells, also extend their protective effects to bone homeostasis in vivo.

Methods

Bone parameters in FoxP3-transgenic (Tg) mice were compared with those in their wild-type (WT) littermate controls. Ovariectomy was performed in FoxP3-Tg mice as a model of postmenopausal osteoporosis, and the bone parameters were analyzed. The bones of RAG-1–/– mice were analyzed following the adoptive transfer of isolated CD4+CD25+ T cells. CD4+CD25+ T cells and CD4+ T cells isolated from FoxP3-Tg mice and WT mice were cocultured with monocytes to determine their ability to suppress osteoclastogenesis in vitro.

Results

FoxP3-Tg mice developed higher bone mass and were protected from ovariectomy-induced bone loss. The increase in bone mass was found to be the result of impaired osteoclast differentiation and bone resorption in vivo. Bone formation was not affected. Adoptive transfer of CD4+CD25+ T cells into T cell–deficient RAG-1–/– mice also increased the bone mass, indicating that Treg cells directly affect bone homeostasis without the need to engage other T cell lineages.

Conclusion

These data demonstrate that Treg cells can control bone resorption in vivo and can preserve bone mass during physiologic and pathologic bone remodeling.

During life, our skeleton is subjected to a process of continuous remodeling. This process allows the skeleton not only to gain peak bone mass during adolescence and to individually model bone architecture during adulthood, but also to maintain bone mass during aging. Bone loss results in a reduced quantity and quality of bone, leading to osteopenia and osteoporosis, which are considered to be major health problems because they enhance the risk of fractures.

The mechanism of bone remodeling that leads to the degradation of bone is a net imbalance between bone formation and bone resorption. With regard to local regulation of bone remodeling, the RANKL/RANK/osteoprotegerin (OPG) system plays the most prominent role (1). Initiation of osteoclastogenesis largely depends on the direct local interaction between osteoclast precursor cells with cells from the osteoblast lineage that are the local source of macrophage colony-stimulating factor (M-CSF) and RANKL. Strikingly, bone resorption mediated by osteoclasts (2, 3) is enhanced and is not compensated by sufficient osteoblast-mediated bone formation (4). Several factors can aggravate this imbalance and speed up bone loss. The most prominent are older age, low body mass, postmenopausal state, and genetic variables.

Over the last several years, tight molecular and cellular links between immune activation and bone loss have been identified (5, 6). The first molecular link that was discovered was RANKL. Indeed, RANKL, the essential factor for osteoclast formation, which is expressed by osteoblasts, is also expressed by activated T cells (7–9). RANKL expression is induced by cytokines, such as tumor necrosis factor α (TNFα), interleukin-1 (IL-1), IL-6, and IL-17. At the same time, some of these cytokines shut down pathways of bone formation and can create a profound imbalance in bone homeostasis, which further aggravates bone loss (10, 11).

Initial consideration was directed toward Th1 cells as the immune cell that drives osteoclast formation. The exact role of Th1 cells in vivo, however, is unclear. Indeed, Th1 cells produce high amounts of interferon-γ (IFNγ), which can directly suppress osteoclast formation (12), but which has also been reported to indirectly stimulate osteoclast differentiation in vivo (13). Th17 cells were recently described as being the major cellular component in the induction of osteoclast formation, since IL-17 induces not only RANKL, but also other osteoclastogenic factors, such as TNFα and IL-1 (14). So far, however, the picture of immune-driven bone loss is quite unilateral and does not take into account the mechanisms of immune regulation that naturally inhibit immune activation.

Treg cells are characterized by their function, which is the suppression of immune activation, but they do not constitute a homogenous cell type. However, the overwhelming knowledge on Treg cells is derived from those cells characterized by the expression FoxP3, a member of the forkhead family of transcription factors (15–18). The important role of FoxP3 in the development of Treg cells is shown by a loss-of-function mutation in humans with immune dysregulation, polyendocrinopathy, and enteropathy, X-linked (IPEX) syndrome, as well as in scurfy mice, which develop a fatal autoimmune lymphoproliferative disease due to the absence of Treg cells (19, 20). Conversely, overexpression of FoxP3 in mice leads to the generation of functional Treg cells from non-Treg cells (21). Treg cells have been shown to inhibit a variety of animal models of autoimmune disease (22–24) and are thus considered a major component of immune regulation (25, 26).

Despite extensive studies on the role of Th1 cells in bone resorption and the discovery of the potential of Th17 cells to indirectly drive osteoclast formation, the role of Treg cells in bone homeostasis has not been characterized. However, we and other investigators recently showed that the addition of Treg cells to monocyte cultures inhibited their differentiation into osteoclasts in vitro, indicating that Treg cells could in fact suppress, rather than activate, bone resorption (27, 28).

We therefore hypothesized that naturally occurring Treg cells not only balance immune activation, but also extend their role in bone homeostasis, thus representing a central cellular mechanism for maintaining bone mass. To investigate this hypothesis, we analyzed the relevance of increased Treg cell numbers in mice to bone homeostasis in vivo.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Animals.

The FoxP3-transgenic (Tg) mice (strain 2826) have been described previously (20). All animal experiments were performed with the approval of the local ethics authorities. Sex-matched mutant and wild-type (WT) littermate control mice maintained on a C57BL/6 background were maintained under sterile conditions.

Micro–computed tomography (micro-CT).

Micro-CT images of tibias were acquired with a laboratory cone-beam micro-CT scanner developed at the Institute of Medical Physics for ultra high resolution imaging (29). It uses a microfocus x-ray tube (Hamamatsu Photonics) and a 2-dimensional cooled CCD detector array (1,024 × 1,024 elements, 19 μm pitch; Photometrics) with a dynamic range of 16 bits. A fiberoptic taper enlarges the sensitive input area of the CCD by a factor of 3. The detector and the sample stage can be linearly translated independently of each other and with respect to the x-ray source, providing variable magnification of the object. For the current project, the following acquisition parameters were used: voltage 40 kV, x-ray current 250 μA, exposure time 5,000 msec/projection, 720 projections, matrix 1,024 × 1,024 elements, voxel size in the reconstructed image 9 μm. Images were analyzed using a plug-in programmed for the Amira 4.1.2 imaging platform (Mercury). The following histomorphometric parameters were calculated: bone volume/total volume, trabecular thickness, and trabecular number.

Bone histomorphometric and immunohistologic analyses.

Histomorphometric analysis was performed on methacrylate-embedded undecalcified plastic sections stained with von Kossa's stain and with Goldner's stain for bone. Histologic and immunohistologic analyses were performed as described previously (30, 31). All quantifications were performed by digital image analysis (OsteoMeasure; OsteoMetrics).

Flow cytometry.

Single-cell suspensions isolated from the spleen, tibia, and femur were stained with fluorochrome-conjugated antibodies in phosphate buffered saline (PBS) containing 5% (weight/volume) bovine serum albumin (Sigma-Aldrich) and were then analyzed with a FACSCalibur instrument (Becton Dickinson). Monoclonal antibodies to CD4 (GK1.5), CD11b (M1/70), CD25 (PC61), annexin V, and bromodeoxyuridine were obtained from BD Biosciences. Intracellular FoxP3 staining was performed according to the manufacturer's instructions using anti-mouse FoxP3 antibody (FJK-16s; eBioscience).

Measurement of serum cytokines.

Cytokines in serum/plasma were measured using a mouse FlowCytomix kit (Bender MedSystems) according to the manufacturer's instructions and then analyzed with a FACSCalibur instrument. Serum levels of RANKL (R&D Systems), OPG (R&D Systems), osteocalcin (Nordic Bioscience), and C-terminal telopeptide α1 chain of type I collagen (CTX-I) (RatLaps; Nordic Bioscience) were measured by enzyme-linked immunosorbent assay according to the manufacturer's instructions.

Isolation and characterization of Treg cells.

Isolation of Treg cells and analysis of their purity and functionality were performed as described previously (28). Spleens were isolated from 5-week-old female mice and were homogenized through a 70-μm stainless steel mesh to obtain single-cell suspensions. Erythrocytes were lysed by treatment with NH4Cl. CD25+ and CD25–CD4+ T cell populations were isolated from the spleen cell suspensions using a microbead-based regulatory T cell isolation kit (Miltenyi Biotec) according to the manufacturer's instructions.

The purity of isolated populations was analyzed by flow cytometry. Isolated T cells were stained with fluorescein isothiocyanate–conjugated anti-CD4 or phycoerythrin-conjugated anti-CD25 (PharMingen). For analysis by fluorescence-activated cell sorting, sorted CD4+CD25+ T cells were also stained intracellularly with streptavidin–Cy5–conjugated anti-FoxP3 (The Jackson Laboratory).

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 were cultured separately for 3 days in 96-well round-bottomed plates (Corning) in the presence of anti-CD3 monoclonal antibody (5 μg/ml; eBioscience) and 1 × 105 irradiated antigen-presenting cells. 3H-thymidine (1 μCi/well) was added during the last 18 hours of the culture, and incorporation was measured with a liquid scintillation counter (1205 Betaplate; Wallac Pharmacia).

For preparation of the adoptive Treg cell transfer model in RAG-1–/– mice, sex-matched littermates were injected intravenously twice within 6 weeks with 1 × 106 isolated WT CD4+CD25+ T cells or with PBS as a control.

Isolation and culture of osteoclasts and coculture with Treg cells.

Bone marrow was isolated from 12-week-old mice by flushing bones with medium. Single-cell suspensions were obtained as described above. CD11b+ cell populations were isolated from the cell suspensions using positive microbead-based isolation according to the manufacturer's instructions. CD11b+ cells were seeded at a density of 1 × 106 cells/well in 24-well plates or a density of 2.5 × 105/well in 96-well plates.

To generate osteoclasts, 30 ng/ml of M-CSF and 50 ng/ml of RANKL (both from R&D Systems) were added to the plates. Cultures were incubated in triplicate, with a complete medium change on day 3. After 5 days of culture, tartrate-resistant acid phosphatase (TRAP) staining for the evaluation of osteoclast differentiation was performed, using a leukocyte acid phosphatase kit from Sigma-Aldrich. Osteoclasts were identified as TRAP+ cells with 3 or more nuclei.

Purified CD11b+ monocytes (2.5 × 105/well) and increased numbers of activated CD4+CD25+ T cells (50,000–12,500 cells/well) were cocultured in 96-well plates in the presence of M-CSF and RANKL. Before the cocultures, T cells were activated for 1 hour with a soluble anti-CD3 monoclonal antibody (5 μg/ml) and 1 × 105 irradiated antigen-presenting cells.

Inhibition studies.

For neutralization experiments, CD4+CD25+ T cells were incubated for 1 hour at 37°C with different concentrations of anti–CTLA-4 monoclonal antibody (clone UC10-4B9; eBioscience) prior to the coculture experiments with CD11b+ monocytes. Osteoclast differentiation was evaluated by TRAP staining.

Ovariectomy.

At 6 weeks of age, mice were either sham-operated or were ovariectomized. Mice were analyzed at 12 weeks of age. One tibia was excised for histologic assessment, and the other for micro-CT imaging. Surgically removed ovaries were examined histologically to verify successful ovariectomy.

Quantitative reverse transcription–polymerase chain reaction (RT-PCR).

Total RNA was isolated from cells using TRIzol (Invitrogen). Total RNA (1 μg) was used for the first-strand complementary DNA synthesis (Amersham Biosciences), which was then used for SYBR Green–based quantitative RT-PCR, which was performed in triplicate according to the manufacturer's instructions. The following RT-PCR primer sequences were used: for FoxP3, 5′-AGG-AGC-CGC-AAG-CTA-AAA-GC-3′ (sense) and 5′-TGC-CTT-CGT-GCC-CAC-TGT-3′ (antisense); for TRAP, 5′-CGA-CCA-TTG-TTA-GCC-ACA-TAC-G-3′ (sense) and 5′-TCG-TCC-TGA-AGA-TAC-TGC-AGG-TT-3′ (antisense); and for retinoic acid–related orphan receptor γt (RORγt), 5′-CCA-CTG-CAT-TCC-CAG-TTT-CT-3′ (sense) and 5′-CGT-AGA-AGG-TCC-TCC-AGT-CG-3′ (antisense). Normalized gene expression values for each sample were calculated as the ratio of expression of messenger RNA (mRNA) for the gene of interest to the expression of mRNA for β-actin.

Statistical analysis.

All statistical analyses were performed with Student's t-test or with one-way analysis of variance followed by Tukey's test. Data are presented as the mean ± SEM except where indicated otherwise. All experiments included 10 mice per group, except where indicated otherwise. P values less than 0.05 were considered significant.

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Increased bone mass in FoxP3-Tg mice.

We first investigated whether mice that overexpressed FoxP3, the central transcription factor of Treg cells, showed an overt bone phenotype. Micro-CT of tibial bones from sex-matched WT and FoxP3-Tg littermates revealed a significantly increased bone mass (bone volume/total volume [BV/TV]) in FoxP3-Tg mice (Figures 1A and B). The phenotype was the result of a higher number of bony trabeculae as well as an increased trabecular thickness (Figure 1B). Decreased osteoclasts numbers and osteoclast-covered bone surface indicated impaired bone resorption in FoxP3-Tg mice (Figure 1C). The impaired bone resorption was confirmed by reduced levels of CTX-I (Figure 1D). The decrease in osteoclast numbers was not the result of an altered ratio of circulating RANKL and OPG in the mice (Figure 1D). TRAP expression, a marker of osteoclast differentiation, was also decreased in the bones of FoxP3-Tg mice (Figure 1E).

thumbnail image

Figure 1. Increased bone density due to decreased bone resorption in FoxP3-transgenic (Tg) mice as compared with wild-type (WT) littermate controls. A, Micro–computed tomography images of trabecular bone from the tibia of WT and FoxP3-Tg littermate mice. Bars = 0.25 mm. B, Structural parameters of tibias from male mice at week 12 (n = 7 mice per group). Bone volume/total volume (BV/TV), trabecular number, and trabecular (Trab.) thickness were measured. C, Quantitative histomorphometry of osteoclast numbers normalized to the trabecular bone perimeter (OcN/mm) and the osteoclast surface normalized to the bone surface (OcS/BS). D, Enzyme-linked immunosorbent assay (ELISA) for serum levels of C-terminal telopeptide α1 chain of type I collagen (CTX-I) as a marker of bone destruction (RatLaps ELISA) and for serum levels of RANKL and osteoprotegerin (OPG) (n = 7 mice per group). E, Quantitative reverse transcription–polymerase chain reaction analysis of tartrate-resistant acid phosphatase (TRAP) expression in the tibias of FoxP3-Tg mice and WT mice. F, Quantitative histomorphometry of osteoblast numbers normalized to the trabecular bone perimeter (ObN/mm), the osteoblast surface normalized to the bone surface (ObS/BS), and the osteoid volume/total volume (OV/TV). G, Mineral apposition rate (MAR) in trabecular bone, as analyzed by double Calcein labeling (n = 6 mice per group). A labeled section from each group of mice is shown at the right. Bars = 2.5 μm. H, ELISA for serum levels of the bone formation marker osteocalcin. Values in B–H are the mean and SEM (n = 10 mice per group, except where indicated otherwise). ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001.

Download figure to PowerPoint

In contrast, bone formation parameters, such as osteoblast numbers, osteoblast covered bone surface, and osteoid volume, were unchanged in FoxP3-Tg mice (Figure 1F). Consistent with this finding, there was no change in the mineral apposition rate (Figure 1G) or in the circulating osteocalcin level in the serum of FoxP3-Tg mice (Figure 1H).

The low number of osteoclasts in FoxP3-Tg mice was not due to a decreased number of osteoclast precursors in the bone marrow, as shown by the similar amounts of CD11b+ monocytes in the bone marrow of WT and FoxP3-Tg mice (Figure 2A). It was also not due to an impaired ability of monocytes to differentiate into osteoclasts (Figure 2B) or to a change in Th17 cells, as indicated by the similar level of RORγt expression (Figure 2C). Among the cytokines that are known to regulate osteoclastogenesis, IFNγ was the only one that was slightly increased in the serum of FoxP3-Tg mice (Figure 2D). The known phenotype of FoxP3-Tg mice, with high levels of FoxP3 mRNA expression (Figure 2E) and increased numbers of Treg cells (Figure 2F), was also detectable in the bone marrow, especially at the epiphysis along the trabecular bone (Figures 2G and H), where osteoclast precursors are committed to differentiating into bone-resorbing osteoclasts.

thumbnail image

Figure 2. Accumulation of Treg cells at sites of bone remodeling in FoxP3-transgenic (Tg) mice as compared with wild-type (WT) littermate controls. A, Flow cytometry of osteoclast progenitors (CD11b+) in bone marrow cells gated on cells with high values for forward and side scatter on the dot plots. B, In vitro osteoclast differentiation assay in the presence of 30 ng/ml of macrophage colony-stimulating factor (M-CSF) and 50 ng/ml of RANKL using CD11b+ bone marrow cells from WT and FoxP3-Tg mice. Osteoclasts were identified as tartrate-resistant acid phosphatase (TRAP)–positive cells with ≥3 nuclei. C, Quantitative reverse transcription–polymerase chain reaction (RT-PCR) analysis of retinoic acid–related orphan receptor γt (RORγt) expression (Expr.) in bone marrow cells (n = 5 mice per group). D, Serum levels of interferon-γ (IFNγ) in WT and FoxP3-Tg mice. E, Quantitative RT-PCR analysis of FoxP3 expression in bone marrow cells. F, Flow cytometry of CD25+FoxP3+ bone marrow cells (n = 6 mice per group). G, FoxP3-stained immunohistology sections of the tibial epiphysis from WT and FoxP3-Tg mice shown at different resolutions. Arrows indicate FoxP3+ cells. Bars = 100 μm at top and 10 μm at bottom. H, Histochemical comparison of FoxP3+ cells in the tibias of WT and FoxP3-Tg mice, as measured in the epiphysis (Ep.) and in the bone shaft (Bs.). Values in A–F and in H are the mean and SEM (n = 10 mice per group, except where indicated otherwise). ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001.

Download figure to PowerPoint

Partial protection from ovariectomy-induced bone loss in FoxP3-Tg mice.

We next prepared an ovariectomy model in WT and FoxP3-Tg mice to examine whether an increase in Treg cells could positively regulate postmenopausal bone loss. Ovariectomy led to a significant decrease in the bone mass of WT mice, as shown by the 8.83% decrease in BV/TV measured by micro-CT analysis, whereas FoxP3-Tg mice were less affected, having only a 3.88% decrease (Figures 3A and B). The bone mass in ovariectomized FoxP3-Tg mice remained at the same level as in sham-operated WT mice (Figures 3A and B). Moreover, the number of bone trabeculae was virtually maintained in FoxP3-Tg mice after ovariectomy, but was strongly diminished and was the major contributor to bone loss in ovariectomized WT mice (Figures 3A and B).

thumbnail image

Figure 3. Partial protection from bone loss after ovariectomy in FoxP3-transgenic (Tg) mice. A, Micro–computed tomography slice images and 3-dimensional reconstruction of trabecular bone from the tibia of wild-type (WT) littermate control and FoxP3-Tg mice that were either sham-operated or ovariectomized (ovx). Bars = 0.25 mm. B, Structural parameters of trabecular bone from the tibias of female WT and FoxP3-Tg mice at week 12. Bone volume/total volume (BV/TV), trabecular number, and trabecular thickness were measured. The BV/TV analysis showed a difference in mean bone loss of 8.83% in WT mice as compared with 3.88% in FoxP3-Tg mice. C, Quantitative histomorphometry of osteoclast numbers normalized to the trabecular bone perimeter (OcN/mm) and the osteoclast surface normalized to the bone surface (OcS/BS). D, Enzyme-linked immunosorbent assay (ELISA) for serum levels of C-terminal telopeptide α1 chain of type I collagen (CTX-1) as a marker of bone destruction (RatLaps ELISA) (n = 5 mice per group). E, Quantitative histomorphometry of osteoblast numbers normalized to the trabecular bone perimeter (ObN/mm) and the osteoblast surface normalized to the bone surface (ObS/BS). F, ELISA for serum levels of the bone formation marker osteocalcin (n = 5 mice per group). G, Histochemical comparison of FoxP3+ cells in the tibias of WT and FoxP3-Tg mice, as measured in the epiphysis (Ep.) and in the bone shaft (Bs.). Values in B–G are the mean and SEM (n = 10 mice per group, except where indicated otherwise). ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001.

Download figure to PowerPoint

Protection of FoxP3-Tg mice from dramatic bone loss was achieved by the prevention of enhanced osteoclast-mediated bone resorption after ovariectomy, which was evident in WT mice, in which a significant increase in osteoclast numbers and osteoclast-covered bone surface were observed (Figure 3C). The increased osteoclast numbers and the increased osteoclast-covered bone surface were less pronounced in FoxP3-Tg mice (Figure 3C). Consistent with these findings, levels of the serum marker of bone resorption (CTX-I) remained low in FoxP3-Tg mice, but increased significantly in WT mice, after ovariectomy (Figure 3D), which further suggests that the increased Treg cell numbers in FoxP3-Tg mice also control the enhanced osteoclastogenesis in the postmenopausal state.

Analysis of parameters of bone formation, such as osteoblast numbers and osteoblast-covered bone surface (Figure 3E) or serum osteocalcin levels (Figure 3F), revealed no differences among the groups of FoxP3-Tg and WT mice that had been ovariectomized or sham-operated. Importantly, immunohistochemical analysis of the tibia revealed an increased accumulation of FoxP3-expressing lymphocytes next to bony trabeculae in the epiphysis of sham-operated and ovariectomized FoxP3-Tg mice (Figure 3G).

Increased bone density in RAG-1–deficient mice after adoptive transfer of Treg cells.

Our in vitro experiments clearly indicated that Treg cells could directly inhibit osteoclastogenesis (28). We therefore prepared a Treg cell adoptive transfer model to address whether Treg cells could indeed regulate bone homeostasis through direct engagement of osteoclast precursors independently of other T cell subsets in vivo. Therefore, 1 × 106 sorted CD4+CD25+ T cells were injected twice within a period of 6 weeks into lymphocyte-deficient RAG-1–/– mice. The sorted CD4+CD25+ T cells were analyzed by flow cytometry for FoxP3 expression as described previously (28), confirming that ∼90% of the injected CD4+CD25+ T cells were FoxP3+ Treg cells (data not shown). Six weeks after the first Treg cell transfer into RAG-1–/– mice, their spleens were analyzed by flow cytometry and RT-PCR for the presence of CD4+CD25+FoxP3+ T cells.

As expected, FoxP3 mRNA was absent from the spleens of control RAG-1–/– mice and was only detected in RAG-1–/– mice into which Treg cells had been transferred (Figure 4A). Flow cytometry confirmed the efficiency of the transfer (Figure 4B). Importantly, analyses for other T cell subsets, such as CD4+CD25+IFNγ+ for Th1 cells, CD4+CD25+IL-4+ for Th2 cells, and CD4+CD25+IL-17+ for Th17 cells, were not detectable by flow cytometry at the end of the experiment.

thumbnail image

Figure 4. Increased bone density in lymphocyte-deficient RAG-1–/– mice after adoptive transfer of Treg cells. A, Quantitative reverse transcription–polymerase chain reaction analysis of FoxP3 expression (Expr.) in spleen cells from RAG-1–/– mice that did or did not undergo adoptive transfer of Treg cells. B, Flow cytometry analysis of spleen cells from the 2 groups of RAG-1–/– mice. C, Enzyme-linked immunosorbent assay (ELISA) for serum levels of C-terminal telopeptide α1 chain of type I collagen (CTX-1) as a marker of bone destruction (RatLaps ELISA) in the 2 groups of RAG-1–/– mice. D, ELISA for serum levels of the bone formation marker osteocalcin. E, Quantitative histomorphometry of osteoclast numbers normalized to the trabecular bone perimeter (OcN/mm) and the osteoclast surface normalized to the bone surface (OcS/BS). F and G, ELISA for serum levels of osteoprotegerin (OPG) (F) and RANKL (G) in the 2 groups of RAG-1–/– mice. H, Structural parameters of tibias from RAG-1–/– mice at week 12. Bone volume/total volume (BV/TV), trabecular number, and trabecular thickness were measured. I, Micro–computed tomography images of trabecular bone from the tibia of RAG-1–/– mice that did or did not undergo adoptive transfer of Treg cells (left) and tartrate-resistant acid phosphatase–stained sections of tibias from the same mice (right). Bars = 100 μm. Values in A–H are the mean and SEM (n = 10 mice per group). ∗ = P < 0.05; ∗∗ = P < 0.01.

Download figure to PowerPoint

The transfer of Treg cells resulted in decreased bone resorption, as measured by serum levels of CTX-I, while bone formation, as assessed by serum levels of osteocalcin, remained unaffected (Figures 4C and D). Decreased osteoclast numbers and decreased osteoclast-covered bone surface indicated the presence of impaired bone resorption in RAG-1–/– mice into which Treg cells had been transferred (Figures 4E and I), whereas there was no change in osteoblast numbers or osteoblast-covered bone surface (data not shown). Levels of both OPG and RANKL were slightly increased in the serum of RAG-1–/– mice into which Treg cells had been transferred (Figures 4F and G), but no major change in the ratio of RANKL to OPG occurred.

Micro-CT analysis of the tibias of age- and sex-matched RAG-1–/– mice that had undergone Treg cell transfer and untreated control littermates revealed significantly increased bone mass (BV/TV) and trabecular numbers in the RAG-1–/– mice that had undergone Treg cell transfer (Figures 4H and I). These data clearly established that Treg cells could directly block osteoclastogenesis in vivo without engaging T cells.

Normal suppressive effects of Treg cells from FoxP3-Tg mice on osteoclast differentiation.

The observations of impaired osteoclastogenesis in vivo but unchanged differentiation in vitro and an unchanged pool of osteoclast precursors clearly pointed to a role of an altered T cell environment in FoxP3-Tg mice. This change in the microenvironment in FoxP3-Tg mice is reflected by 1) an increase in the expression of mRNA for FoxP3 in the bone marrow (Figure 2E), 2) a higher proportion of FoxP3-expressing lymphocytes in bone marrow or spleen (Figure 2F), and 3) a denser accumulation of FoxP3-expressing lymphocytes next to bony trabeculae in FoxP3-Tg mice (Figure 2G), even in the ovariectomized FoxP3-Tg mice (Figure 3F). These observations suggested that Treg cells locally affect osteoclastogenesis.

To test this hypothesis, cocultures of osteoclast precursors isolated from the bone marrow and splenic T cells derived from WT and FoxP3-Tg mice were performed. A lower concentration of CD4+ T cells from FoxP3-Tg mice was needed to efficiently suppress osteoclastogenesis (Figure 5A). This effect is the result of a higher fraction of Treg cells in FoxP3-Tg mice, and not the result of an enhanced suppressive potential, as indicated by identical suppressive effects of purified Treg cells from WT and FoxP3-Tg mice on osteoclast differentiation (Figure 5B).

thumbnail image

Figure 5. Absence of enhanced suppressive capacity of Treg cells from FoxP3-transgenic (Tg) mice. A and B, In vitro osteoclast differentiation assay of wild-type (WT) CD11b+ monocytes cocultured with activated CD4+ T cells (A) or CD4+CD25+FoxP3+ T cells (B) isolated from WT and FoxP3-Tg mice at the indicated ratios, as determined by tartrate-resistant acid phosphatase (TRAP) staining. C, WT osteoclast coculture assay with Treg cells at the indicated ratios. The decrease in Treg cells per well was compensated by the addition of CD4+CD25–T cells to each well to keep the total amount of T cells constant. D, WT osteoclast coculture assay with Treg cells at the indicated ratios. Total bone marrow cells were used as the osteoclast precursors instead of sorted CD11b+ monocytes. E, Coculture of WT CD4+CD25+FoxP3+ T cells preincubated with a blocking anti–CTLA-4 monoclonal antibody or with isotype control before coculture with CD11b+ WT monocytes. F, In vitro osteoclast differentiation assay of CD11b+ monocytes from WT mice or CD80/86–/– mice cocultured with activated CD4+CD25+FoxP3+ T cells from WT mice at the indicated ratios. G, Enzyme-linked immunosorbent assay (ELISA) for interferon-γ (IFNγ) in supernatants from Treg cells cocultured with CD11b+ monocytes at a 1:5 ratio, as shown in the previous experiments. Supernatants were taken every 24 hours from each well. Values are the mean and SEM (n = 10 mice per group). ∗ = P < 0.05; ∗∗ = P < 0.01.

Download figure to PowerPoint

In order to confirm that the effects of adding Treg cells were due to the specific inhibition of osteoclast formation and not simply to a crowding out of the osteoclast precursors, the total numbers of cells was kept constant in additional coculture experiments. These experiments showed very similar results (Figure 5C). Moreover, virtually identical results were obtained when cocultures were performed with total bone marrow cells instead of with CD11b+ sorted osteoclast precursors in order to mimic the situation in vivo (Figure 5D).

Considering that cell–cell contact is essential for Treg cell–mediated suppression of osteoclastogenesis (28, 32), we hypothesized that molecules involved in monocyte–T cell interactions may play an important role in the suppression of bone resorption by Treg cells. Blockade of CTLA-4 by neutralizing antibody fully abolished the inhibitory effect of Treg cells on osteoclastogenesis (Figure 5E). Further coculture experiments using CD11b+ monocytes from either WT or CD80/86–/– mice and Treg cells from WT mice showed a significant reduction of Treg cell–induced suppression of osteoclast differentiation of the CD80/86–/– monocytes as compared with monocytes isolated from WT mice (Figure 5F). Direct cell–cell contact of Treg cells and monocytes was also essential for the secretion of IFNγ (Figure 5G), since no IFNγ was detectable in the culture supernatants when monocytes were challenged with CTLA-4 alone (data not shown).

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

Herein, we show that FoxP3-Tg mice have increased bone density and are partially protected from postmenopausal osteoporosis. These data suggest a novel bone protection mechanism by the immune system. Our conclusions are based on the following experimental observations. First, bone mass was increased in FoxP3-Tg mice with increased Treg cell numbers. Second, these mice were also partially protected from estrogen depletion–induced osteoporosis. Third, increased bone mass was seen in T cell–deficient mice following adoptive transfer of Treg cells.

Immune activation, particularly the initiation of Th1 and Th17 responses, has been recognized as a key mechanism of bone loss. Th1 cells that can produce RANKL were first identified as the immune cells driving osteoclastogenesis and bone destruction in the context of autoimmune diseases (33). The same cells, however, secrete large amounts of IFNγ, which blocks osteoclast differentiation (12), and were found to directly inhibit, rather than stimulate, osteoclastogenesis in vitro (14). In vivo, IFNγ-secreting T cells were shown to indirectly stimulate osteoclast differentiation (13). More recently, Th17 cells, a proinflammatory T cell subset, were also found to indirectly promote osteoclast differentiation by stimulating the supportive activity of osteoblasts in response to IL-17 secretion (14). Whether the balance of the immune system, with its activating, but yet regulatory, aspects, also applies to bone homeostasis has been poorly characterized so far (34). The results of our present study clearly indicate that Treg cells preserve bone mass, therefore emerging as an immune protector from bone loss.

How do Treg cells regulate bone homeostasis? The number of osteoclasts in vivo as well as the serum parameters of bone resorption were consistently blunted in mice overexpressing FoxP3, as well as in RAG-1–deficient mice following Treg cell transfer. Moreover, no change in the osteoblast numbers, mineral apposition rate, or osteocalcin levels in the serum were observed in FoxP3-Tg mice. These data and the fact that Treg cells were conspicuously localized along the trabecular bone clearly suggest that Treg cells work locally on bone homeostasis and through osteoclasts.

One possible way that osteoclast formation is affected in FoxP3-Tg mice might be through intrinsically impaired osteoclast formation. This hypothesis is highly unlikely, since monocytes from WT and FoxP3-Tg mice did not express FoxP3 (data not shown), which is rather confined to the T cell lineage (15). In addition, direct osteoclast differentiation assays without T cells did not show any difference in the osteoclastogenic differentiation potential among monocytes from WT mice and those from FoxP3-Tg mice. That FoxP3 expression on Treg cells is crucial for the regulatory effect on osteoclasts is evident from coculture experiments of Treg cells and monocytes, which showed a dose-dependent blockade of osteoclast formation by Treg cells.

There are three potential mechanisms through which Treg cells could principally affect osteoclasts. First, the ratio of RANKL to OPG could be shifted in favor of OPG by Treg cells, which impairs the ability of RANKL to engage RANK and thereby prevents osteoclast formation. This has been ruled out by previous studies showing an even higher expression of RANKL in Treg cells as compared with activated T cells (28). In addition, even if the serum levels of OPG and RANKL are not necessarily informative of local concentrations in the bone marrow, no clear change in the ratio of circulating OPG to RANKL could be detected in the blood of FoxP3-Tg mice or in RAG-1–/– mice into which Treg cells had been transferred.

Second, Treg cells could shift the cytokine milieu from a pro-osteoclastogenic cytokine pattern, with TNFα, IL-1, IL-6, or IL-17, to an anti-osteoclastogenic pattern, with more transforming growth factor β (TGFβ), IL-10, or IFNγ. Our previous coculture experiments using Treg cells and osteoclasts did not favor this hypothesis, since blockade of IL-10 or TGFβ only slightly reversed the effects of Treg cells on osteoclasts (28). Kim et al (27), who independently reported a suppressive effect of Treg cells on human osteoclast differentiation, emphasized the role of soluble cytokines as a source of suppression. Zauli et al and other investigators (28, 32, 35–38) showed that TRAIL, a member of the TNF superfamily, might be a potential link between the 2 studies (27, 28). Our results from the present study showed a mild, but significant, increase in the level of circulating IFNγ in FoxP3-Tg mice, and we found increased IFNγ levels in the supernatants from cocultures of Treg cells with osteoclast progenitors, whereas no IFNγ was detected when osteoclast precursors were challenged with CTLA-4. These data suggested that IFNγ was not needed for the inhibition of osteoclastogenesis by Treg cells. Thus, cytokines, the secretion of which depends on the direct interaction of Treg cells with the osteoclast precursors, may only partly account for Treg cell inhibition of osteoclastogenesis.

Whether IFNγ is the mediator of Treg cell–mediated osteoclast inhibition is a subject of controversy. Indeed, while IFNγ is able to directly block osteoclastogenesis in vitro (12), systemic long-term injection of IFNγ induces bone resorption (13), a property that has proved useful in the treatment of severe osteopetrosis in human (39). Differences between the doses, the duration of the treatment, as well as the local concentration in the bone may explain the different effects of IFNγ on bone homeostasis.

Third, Treg cells could inhibit osteoclastogenesis by suppressing activated T cells. The fact that Treg cells are able to inhibit osteoclastogenesis in T cell–deficient mice strongly favors a direct effect of Treg cells on bone without involvement of other T cells. Thus, the most attractive concept is that Treg cells directly impair osteoclast differentiation via cell-to-cell contact. The observation that FoxP3+ Treg cells accumulate at sites of intense bone remodeling, such as the epiphysis of long bones, also favors the concept that Treg cells need cell-to-cell contact and act locally on bone. We observed a mild increase in the number of FoxP3+ Treg cells in the epiphysis of ovariectomized FoxP3-Tg mice. This phenomenon was not seen in the WT controls. Although we do not fully understand the basis of this difference, this observation also favors a local action of Treg cells on bone. This interpretation is also supported by recent studies suggesting that Treg cells act on cells of the innate immune system in order to control arthritis (40, 41).

In summary, these data open a new perspective in the interactions between the immune system and the skeleton. The central idea of osteoimmunology that inflammatory T cell subsets, such as activated Th1 cells and Th17 cells, drive bone loss is now enriched by the concept that immune homeostasis is tightly linked to bone homeostasis. Thus, a key mechanism of immune regulation, the naturally occurring FoxP3+ Treg cells, transduces its favorable role in immune balance to skeletal homeostasis. This may drive new concepts of a “bone protection system” and shape immunologic tools to maintain bone mass.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. 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 conception and design. Zaiss, Nimmerjahn, Voll, Schett, David.

Acquisition of data. Zaiss, Sarter, Hess, Engelke, Böhm, Schett.

Analysis and interpretation of data. Zaiss, Engelke, Nimmerjahn, Voll, Schett, David.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES

We thank Richard A. Kroczek (Molecular Immunology Department, Robert-Koch Institute, Berlin, Germany) for kindly providing the CD80/86–/– mice, we thank B. Roy, C. Stoll, I. Schmidt, C. Zech, and M. Loibl for technical assistance, and we thank Steven Ziegler (Benaroya Research Institute at Virginia Mason, Seattle, WA) for kindly providing the FoxP3-Tg mice and FoxP3-deficient mice.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. AUTHOR CONTRIBUTIONS
  7. Acknowledgements
  8. REFERENCES