Article Text
PDF
Abstract
Objective Under both physiological and pathological conditions, bone volume is determined by the rate of bone formation by osteoblasts and bone resorption by osteoclasts. Excessive bone loss is a common complication of human IBD whose mechanisms are not yet completely understood. Despite the role of activated CD4+ T cells in inflammatory bone loss, the nature of the T cell subsets involved in this process in vivo remains unknown. The aim of the present study was to identify the CD4+ T cell subsets involved in the process of osteoclastogenesis in vivo, as well as their mechanism of action.
Design CD4+ T cells were studied in IL10−/− mice and Rag1−/− mice adoptively transferred with naive CD4+CD45RBhigh T cells, representing two well-characterised animal models of IBD and in patients with Crohn's disease. They were phenotypically and functionally characterised by flow cytometric and gene expression analysis, as well as in in vitro cocultures with osteoclast precursors.
Results In mice, we identified bone marrow (BM) CD4+ T cells producing interleukin (IL)-17 and tumour necrosis factor (TNF)-α as an osteoclastogenic T cell subset referred to as Th17 TNF-α+ cells. During chronic inflammation, these cells migrate to the BM where they survive in an IL-7-dependent manner and where they promote the recruitment of inflammatory monocytes, the main osteoclast progenitors. A population equivalent to the Th17 TNF-α+ cells was also detected in patients with Crohn's disease.
Conclusions Our results highlight the osteoclastogenic function of the Th17 TNF-α+ cells that contribute to bone loss in vivo in IBD.
- BONE MINERAL DENSITY
- IBD
- IMMUNOLOGY
- T LYMPHOCYTES
Statistics from Altmetric.com
Significance of this study
What is already known on this subject?
Patients with IBD have an increased risk for bone loss and fractures.
Th1 and Th17 cells are involved in the pathogenesis of IBD.
In vitro, Th17 cells are able to increase the differentiation of osteoclasts induced by exogenous receptor activator of nuclear factor κβ ligand (RANK-L) and macrophage colony-stimulating factor (M-CSF).
We have previously reported that CD4+ T cells contribute to the differentiation of osteoclasts in vivo in an osteopetrotic mouse model.
What are the new findings?
The identification of bone marrow (BM) CD4+ T cells producing interleukin (IL)-17 and tumour necrosis factor (TNF)-α, but not interferon-γ, as osteoclastogenic T cells, as demonstrated by their capacity to induce osteoclast differentiation without addition of exogenous RANK-L and M-CSF, in two well-characterised models of IBD.
Th17 TNF-α+ cells induce the expression of osteoclastogenic factors by BM mesenchymal stromal cells and the recruitment of inflammatory monocytes that are osteoclast precursor cells in the BM.
IL-7 is essential for the survival of osteoclastogenic Th17 TNF-α+ cells in the BM.
The presence of an equivalent Th17 TNF-α+ population in patients with Crohn's disease.
How might it impact on clinical practice in the foreseeable future?
The present study highlights that the BM Th17 TNF-α+ population corresponds to an osteoclastogenesis-inducer population that maintains a vicious circle linking inflammation and bone destruction. These cells, therefore, represent a potential target for innovative immunotherapeutic strategies against chronic intestinal inflammation and bone loss.
Because bone loss is associated with the pathology of other chronic inflammatory diseases, these results may also have broader implications.
Introduction
Inflammation is a module of host defence immune responses against pathogens or danger signals. Persistence of these responses results in chronic inflammatory diseases that are often associated with bone destruction even when the inflammatory site is not the bone, as in IBD or Crohn’s disease (CD).1–3 Indeed, bone loss has been reported in more than 40% of patients with IBD and remains a major extraintestinal cause of morbidity leading to an impaired quality of life and productivity.2 The prevalence of osteopenia and osteoporosis in patients with IBD ranges from 22%–77% to 17%–41%, respectively, depending on the populations, location or study design.4 ,5
Osteoclasts (OCL) are multinucleated cells from the monocyte/macrophage lineage that degrade bone matrix and dynamically remodel the skeleton.6 Bone remodelling is a highly regulated process involving complex interactions between bone-forming osteoblasts (OBL) and bone-resorbing OCLs. These interactions require cell–cell contact, production of cytokines and the generation of coupling factors during bone resorption.7 A precise regulation of this process is a prerequisite for normal bone homeostasis and an imbalance is often linked to metabolic bone diseases in humans, such as osteoporosis and inflammatory bone loss. The generation of OCLs is supported by OBLs that produce macrophage colony-stimulating factor (M-CSF) and receptor activator of nuclear factor-κ β ligand (RANK-L), the two essential signals for OCL differentiation from myeloid cells, such as inflammatory monocytes, macrophages or dendritic cells.8 ,9 It has become clear that systemic inflammation induces high levels of circulating inflammatory immune cells that can interact with bone cells through multiple pathways and that, in turn, synergise to cause severe bone destruction, independently of other possible causes such as vitamin D and calcium deficiency, or steroid treatment.10 However, the immunological link between chronic intestinal inflammation and bone destruction is still poorly understood.
Among these immune cells, activated CD4+ T cells have been involved in bone destruction associated with chronic inflammation.11 We have shown that presence of CD4+ T cells contributes to the differentiation of OCL in vivo in the pathological osteopetrotic oc/oc model of mice that are defective in OCL resorptive function.9
In vitro differentiated CD4+ Th1 and Th2 cells have been shown to inhibit osteoclast formation through their canonical production of interferon (IFN)-γ and interleukin (IL)-4, respectively.12 By contrast, in vitro differentiated Th17 cells act as an osteoclastogenic helper T cell subset resulting from their production of IL-17.12 IL-17 was associated with increased osteoclastogenesis in chronic inflammation. IL-17 is abundant in rheumatoid arthritis synovial fluid and stimulate osteoclastogenesis by inducing RANK-L expression on OBLs.13 IL-17-deficient mice are resistant to bone destruction induced by lipopolysaccharide.12 Otherwise, Oostlander et al3suggested a particular role for IL-17 in osteoclastogenesis in patients with CD. However, despite the implication of CD4+ T cells in inflammatory bone loss, the nature of the T cell subsets involved in this process and how they act in vivo remains unknown.
The aim of the present study was to identify and characterise such osteoclastogenic Th cells responsible for bone destruction during chronic intestinal inflammation using two well-characterised models of IBD with bone destruction, as well as blood from patients with CD. The results revealed that CD4+ T cells producing IL-17- and tumour necrosis factor (TNF)-α- but not IFN-γ, are able to recruit osteoclast progenitor cells to the bone marrow (BM) and to induce osteoclastogenesis, and to link chronic inflammation and bone destruction thereby maintaining this vicious circle in IBD.
Materials and methods
Mice
BALB/c IL10−/− and C57BL/6 IL10−/− mice were obtained from the Département de Cryopréservation, Distribution, Typage et Archivage animal (CDTA, Orléans, France). Clinical signs of IBD were determined by weight loss, hunching and diarrhoea which was generally observed in about 60% of mice, both male and female, irrespective of their genetic background, 4 months after birth, at which time the animals were sacrificed and analysed. For each experiment, three groups of mice without (IL-10−/−) or with clinical signs of IBD (IBD IL-10−/−), as well as control wild-type (WT) mice, consisting of five to six animals each, were studied.
C57BL/6 Rag1−/− mice were obtained from Charles River laboratories (Wilmington, USA). WT BALB/c and C57BL/6 mice were obtained from Janvier (Le Genest-St-Isle, France). Mice were bred and maintained in the animal facility of the Centre Mediterranéen de Médecine Moléculaire (Nice, France) in accordance with the general guidelines of the Direction des Services Vétérinaires.
Lymphocytes isolation
CD4+ T cells were isolated from the spleen, mesenteric lymph nodes (MLN) and BM as previously described,14 using the negative selection Kit CD4+ according to manufacturer's protocol (Life technologies, St Aubain, France).
CD4+ IL-17 and TNF-α-producing cells referred to as Th17 TNF-α+ cells, were enriched from BM of IBD IL-10−/− mice using above mentioned kit (Life technologies, St Aubain, France) and stimulated for 2 h with phorbol myristate acetate (PMA) (10 ng/mL) and ionomycin (1 µg/mL). CD4+ T cells secreting IL-17 were enriched with IL-17-secreting assay according to the manufacturer's instructions (Miltenyi Biotec, Paris, France) and sorted using FACSAria flow cytometer (BD Biosciences, Le Pont le Claix, France).
Patients and human T cells
Peripheral blood mononuclear cells (PBMC) were isolated from freshly collected, heparinised peripheral blood of healthy voluntary donors and patients with CD (Department of gastroenterology, CHU of Nice, France) by centrifugation over Ficoll-Hypaque. The protocol was approved by the local ethics committee of the Hospital Center of Nice University, and all 15 patients and 10 control subjects who were included in this study provided a signed informed consent.
PBMC were stained with a fluorescein isothiocyanate-conjugated anti-CD4 monoclonal antibody (mAb) and a phycoerythrin-conjugated anti-CCR6 mAb and CCR6+ and CCR6− T cells were purified using a FACSAria cell sorter that were cloned by limiting dilution and selected T cell clones were used in subsequent experiments. Cloning and culture procedure, as well as characterisation of the T cell clones, based on their cytokine production profile have been described previously.15
Statistical analysis
All results were expressed as mean±SD. The level of statistical significance was determined by one-way analysis of variance followed by Dunnett's t test for the experiments with the 3 groups and Student t test for the two experimental groups. p Values <0.05 were considered significant.
Supplementary material
Protocols for adoptive transfer experiment, in vitro Th17 TNF-α+ differentiation, in vitro OCL differentiation and real-time quantitative PCR are provided in online supplemental materials.
Results
Only IL-10-deficient (IL-10−/−) mice with chronic intestinal inflammation develop bone destruction
IL-10−/− mice gradually develop a chronic intestinal inflammation that shares histopathological features with IBD in humans. They also develop severe bone destruction with age as previously described.16 In agreement with this notion, microscanner analysis showed a bone destruction and a dramatic decrease in trabecular thickness, trabecular number, bone surface density and in the trabecular bone volume per tissue volume only in IBD IL-10−/− mice, as compared with healthy IL-10−/− and control mice (figure 1A, B). No significant differences in the incidence of these parameters were observed between healthy IL-10−/− and WT mice, indicating that bone destruction in IL-10−/− mice is dependent on the development of intestinal inflammation rather than on the mere absence of IL-10. Similar results were obtained irrespective of the genetic background (C57BL/6 or BALB/c) or the sex of the IL-10−/− mice (data not shown).
Bone marrow (BM) CD4+ T cells induce osteoclast (OCL) differentiation leading to osteopenia, in interleukin (IL)-10−/− mice with IBD. (A) Longitudinal and cross-sectional μCt reconstruction of the distal femur of IL-10−/− mice afflicted with IBD (IBD IL-10−/−), IL-10−/− mice that had not yet developed IBD (healthy IL-10−/−) and control littermates (WT, wild type). (B) Morphological analysis by μCt of the femur from WT, healthy IL-10−/−, and IBD IL-10−/− mice. Trabecular indices were computed from μCt scans. BV/TV: trabecular bone volume per tissue volume. The data are presented as the mean±SD of 6 mice per group. One-way analysis of variance followed by a Dunnett's test (*p<0.05 and **p<0.01 versus healthy IL-10−/−). (C) Purified BM CD4+ T cells were added to WT monocytes cocultured with WT osteoblasts generated from mouse calvaria, or to the mouse osteoclastic cell line RAW264.7 in the presence of phytohaemagglutinin (1 µg/mL) during 8 days. Multinucleated cells (OCLs) were analysed for their tartrate-resistant acid phosphatase activity. Coculture of osteoblasts and WT monocytes in the presence of 10−8 M VitD3 and 10−8 M dexametasone, or RAW264.7 cells treated with 50 ng/mL receptor activator of nuclear factor κβ ligand served as a positive control. A representative result of three independent experiments with 5 mice per group and with similar results is shown.
To analyse the ability of BM CD4+ T cells to induce OCL differentiation, CD4+ T cells were isolated from the BM of IBD IL-10−/−, healthy IL-10−/− or control mice and their osteoclastogenic effect was tested in vitro in coculture with WT monocytes (OCL precursors) and with WT OBLs generated from mouse calvaria, as well as in coculture with the mouse osteoclastic cell line RAW264.7. The formation of OCLs was evidenced by the presence of multinucleated cells (≥3 nuclei) expressing the tartrate-resistant acid phosphatase (TRAP), a common marker of OCLs.
As shown in figure 1C, after 8 days of coculture, only BM CD4+ T cells from IBD IL-10−/− mice, but not those from healthy IL-10−/− or WT mice, were able to induce OCL formation in both monocyte and RAW264.7 cell differentiation models. It is of note that the induction of OCL differentiation by this population of CD4+ T cells did not require addition of RANK-L and M-CSF. These results suggested the presence of an osteoclastogenic CD4+ T cell population in the BM of IBD IL-10−/− mice that might be involved in increased OCL differentiation in vivo.
Accumulation of Th17 TNF-α+ cells in the BM of IL-10−/− mice with IBD
To determine the nature of this potentially osteoclastogenic CD4+ Th cell subset present in IBD IL-10−/− mice, CD4+ T cells were isolated from the spleen, MLN, lamina propria (LP) and BM of control, healthy IL-10−/− or IBD IL-10−/− mice, and their cytokine production profile was determined by intracellular cytokine analysis by flow cytometry. This analysis revealed in IBD IL-10−/−, but not healthy IL-10−/− and WT mice, the emergence of a CD4+ T cell population, expressing IL-17A and TNF-α (figure 2A, B). Although a significant fraction of these CD4+ T cells isolated from the LP, MLN and spleen from IBD IL-10−/− mice also produced IFN-γ, production of this cytokine by BM Th17 TNF-α+ cells was undetectable (figure 2C). In subsequent experiments, BM Th17 TNF-α+ cells of IBD IL-10−/− mice were enriched using IL-17 secreting assay to further analyse their phenotype. We have shown that about 26% of the Th17 TNF-α+ cells, but none of the CD4+IL-17− T cells expressed the membrane-bound form of RANK-L (mRANK-L) (figure 2D). Additionally, quantitative PCR on the purified cells showed that the Th17 TNF-α+ cells expressed Il-22, Rorγt, M-csf, Ccr6 and Cxcr4 at much higher levels than CD4+IL-17− cells, whereas, in contrast, these latter cells strongly expressed T-bet and IFN-γ (figure 2E). Collectively, these results confirmed that the Th17 TNF-α+ cells are related to the Th17 cell population. Because of their selective expression of M-CSF and mRANK-L, but not IFN-γ described as an inhibitor of OCL differentiation,17 ,18 these cells represent a good candidate for an OCL-inducer Th cell subset.
Characterisation of bone marrow (BM) CD4+ T cells from IBD interleukin (IL)-10−/− mice. (A) Analysis of cytokine expression of CD4+ T cells from lamina propria (LP), mesenteric lymph nodes (MLN), spleen (SPL) and BM of wild type (WT), healthy IL-10−/− and IBD IL-10−/− mice was assessed by intracellular flow cytometry after in vitro restimulation with phorbol myristate acetate and ionomycin. (B) Frequency of Th17 tumour necrosis factor (TNF)-α+ cells in the BM of WT, healthy IL-10−/− and IBD IL-10−/− mice. One-way analysis of variance followed by a Dunnett's test (**p<0.01 versus healthy IL-10−/−). (C) Intracytoplasmic interferon (IFN)-γ staining on Th17 TNF-α+ cells from IBD IL-10−/− mice. The experiment was repeated more than 3 times with similar results. (D) FACS analysis of membrane-bound receptor activator of nuclear factor κβ ligand (mRANK-L) on sorted BM Th17 TNF-α+ cells from IBD IL-10−/− mice after enrichment with IL-17-secreting assay. (E) Gene expression of Il-17, Il-22, Tnfα, Ifnγ, M-csf, Rorγt, T-bet, Ccr6 and Cxcr4 was analysed by quantitative PCR on sorted Th17 TNF-α+ cells and CD4+IL-17− T cells. The data are presented as fold induction of gene expression in Th17 TNF-α+ cells versus CD4+IL-17− T cells. Data are representative of three independent experiments.
BM Th17 TNFα+ cells are potent inducers of OCL differentiation
The osteoclastogenic capacity of BM T cells was tested in the two models of OCL differentiation described above. In both systems, only the Th17 TNF-α+ cells induced the differentiation of multinucleated TRAP+ cells (figure 3A) able to resorb mineralised matrix, thus corresponding to true OCLs (figure 3B). These results underlined the osteoclastogenic potential of the Th17 TNF-α+ cells. As expected, osteoprotegerin inhibits OCL differentiation whatever the culture system (figure 3C). In coculture between monocytes, OBLs and Th17 TNF-α+ cells, addition of anti-TNFα or anti-IL17 neutralising antibodies inhibited OCL differentiation, which is consistent with the known effect of IL-17 and TNFα on OBL expression of RANK-L (figure 3C). When using coculture of RAW cells and Th17 TNF-α+ cells, addition of anti-IL17 antibodies had only a weak effect on OCL differentiation, whereas anti-TNF-α was a potent inhibitor of this differentiation (figure 3C). These results indicate that IL-17 has not a major direct effect on OCL precursors.
Bone marrow (BM) Th17 tumour necrosis factor (TNF)-α+ cells represent an osteoclastogenic T cell subset. (A) Sorted BM Th17 TNF-α+ cells were cultured with wild type (WT) monocytes and WT osteoblasts (OBL) generated from mouse calvaria, or with RAW264.7 cells in the presence of phytohaemagglutinin (PHA) (1 µg/mL) for 8 days. Osteoclasts (OCLs) were analysed for their tartrate-resistant acid phosphatase (TRAP) activity. (B) Resorption activity was assessed on a hydroxy-apatite-coated disc for cells cultured as indicated above (C) Enumeration of TRAP-positive multinucleated OCLs. Th17 TNFα+ cells were added to WT monocytes cocultured with WT OBLs generated from mouse calvaria (gray bar), or to the mouse osteoclastic cell line RAW264.7 in the presence of PHA (1 µg/mL) (black bar) and in the presence or absence of neutralising anti-IL-17, anti-TNF-α monoclonal antibodies (10 μg/mL each) or osteoprotegerin (OPG) (100 ng/mL), during 8 days as indicated. (D) Rag1−/− mice were transferred with 4.105 CD4+CD45RBhigh T cells from WT or healthy IL-10−/− donors and sacrificed when recipients developed clinical signs of disease (approximately 8 weeks after transfer). IL-17A and TNF-α expression in BM CD4+ T cells was assessed by intracellular flow cytometry after in vitro restimulation with phorbol myristate acetate and ionomycin. Data are representative of the results from two independent experiments.
Additionally, we sought to demonstrate the presence of this population in the well-characterised model of IBD induced by the transfer of naive CD4+CD45RBhigh T cells into syngenic Rag1−/− mice. Intraperitoneal injection of CD45RBhigh T cells from WT or IL-10−/− mice into Rag1−/− mice induced, after 8 weeks, the development of clinical signs of colitis19 ,20 as well as bone destruction.21 Analysis of intracytoplasmic cytokine expression showed that BM Th17 TNF-α+ cells are induced whatever the origin of injected T cells (from IL-10−/− or WT mice) (figure 3D), indicating that the differentiation of these cells in vivo is independent from the presence or absence of IL-10. Additionally, these observations further strengthened the link between inflammation, bone destruction and the activity of Th17 TNF-α+ cells.
IL-7 is essential for the survival of osteoclastogenic Th17 TNF-α+ cells
The BM is a preferential site for maintenance of memory T cells.22 ,23 Flow cytometric analysis of BM Th17 TNF-α+ cells from IBD IL-10−/− mice revealed that they are CD44hiCD62L− (data not shown) and IL-7R+ (figure 4A) but that they do not express Ly6C, a marker of BM dormant memory T cells. These results suggested that the BM Th17 TNFα+ cells correspond to memory cells with an effector phenotype.
Interleukin (IL)-7 is essential for the osteoclastogenic Th17 tumour necrosis factor (TNF)-α+ cell survival. (A) Phenotypical analysis by flow cytometry of bone marrow (BM) Th17 TNFα+ cells. (B) Intracellular flow cytometry of IL-17A and TNF-α in enriched BM CD4+ T cells from IBD IL-10−/− mice. T cells were stimulated with CD3/CD28 beads in the presence of IL-2 (2 ng/mL) or IL-7 (10 ng/mL) for 7 days. (C) Representative flow cytometry of IL-17A and TNF-α expression after repeated stimulation of naive T cells under Th17 cell-polarising conditions in the presence of IL-7. (D) Intracellular flow cytometry of TNF-α and interferon (IFN)-γ and FACS analysis of IL-7 receptor (IL-7R) and Ly6C on purified Th17 TNFα+ cells differentiated in vitro. A representative result of three independent experiments. (E) Purified Th17 TNFα+ cells were added to wild type (WT) monocytes co-cultured with WT osteoblasts (OBL) generated from mouse calvaria in the presence of PHA (1 µg/mL) during 8 days. Multinucleated cells (OCLs) were analysed for their tartrate-resistant acid phosphatase (TRAP) activity. (F) The gene expression of Il-7 was analysed by quantitative PCR on sorted CD45−Ter119−CD3−CD19−CD31− stromal cells from WT, healthy IL-10−/− and IBD IL-10−/− mice. Data represent pooled results from three independent experiments. One-way analysis of variance followed by a Dunnett's test (*p<0.05 versus healthy IL-10−/−).
To further analyse the role of IL-7 on the survival of the Th17 TNFα+ cells, CD4+ T cells isolated from the BM of IBD IL-10−/− mice were treated with IL-7, or IL-2 as negative control. After 7 days of IL-7 treatment, the frequency of the Th17 TNF-α+ cells was fourfold increased as compared with treatment with IL-2 (figure 4B). We next explored whether Th17 TNF-α+ cells could be generated and expanded in vitro in the presence of IL-7. Repeated stimulation and subsequent culture of naive T cells under Th17 cell-polarising conditions in the presence of IL-7 resulted in the generation of a CD4+ T cell population consisting of about 60% of Th17 TNF-α+ cells (figure 4C). These in vitro-generated cells shared the same phenotypical and functional characteristics than those isolated ex vivo in that they were IL-7R+, but did not produce IFN γ and did not express Ly6C at their cell surface (figure 4D). Moreover, they were able to induce OCL differentiation from monocytes (figure 4E), as well as from RAW264.7 cells (data not shown). In order to determine whether the survival of osteoclastogenic T cells induced by stromal cells could be mediated by IL-7 in vivo, the presence of IL-7 transcripts was determined in BM stromal cells purified as described.24 The results showed an increased expression of IL-7 in purified CD45− Ter119− CD3− CD19− CD31− stromal cells from IBD IL-10−/− mice, as compared with cells from healthy IL-10−/− and control mice (figure 4F). Collectively, these findings suggested that the osteoclastogenic Th17 TNF-α+ cell population is likely to reside in the BM of mice with IBD in an IL-7-dependent manner.
Th17 TNFα+ cells promote the recruitment of inflammatory monocytes, OCL precursors, during IBD
The increase in OCL differentiation leading to bone destruction in IBD necessitates a higher recruitment of OCL precursors into the BM. Thus, OCL differentiation was performed in vitro from equal numbers of total BM cells from control, healthy IL-10−/− and IBD IL-10−/− mice.
Enhanced osteoclastogenesis was only obtained with BM cells from IBD IL-10−/− mice, which was consistent with the characteristic bone loss of these animals (figure 5A). Increased osteoclast differentiation potential in IBD IL-10−/− mice could result from an increased number of precursor cells. Interestingly, flow cytometric analysis of CD11bhiLy6Chi inflammatory monocytes and CD11b−/lo Ly6Chi cells (representing OCL precursors recently described by Charles et al25) showed that these populations were increased in the BM of IBD IL-10−/− mice, as compared with healthy IL-10−/− and control mice (figure 5B).
Th17 tumour necrosis factor (TNF)-α+ cells promote a recruitment of inflammatory monocytes, into bone marrow (BM) during IBD. (A) Tartrate-resistant acid phosphatase (TRAP) staining of osteoclasts differentiation of total BM cells from wild-type (WT), healthy interleukin (IL)-10−/− and IBD IL-10−/− mice, One-way analysis of variance (ANOVA) followed by a Dunnett's test (*p<0.05 and **p<0.01 versus healthy IL-10−/−). (B) Representative flow cytometry analysis of CD11bhiLy6Chi and CD11b−/loLy6Chi cells from BM cells of WT, healthy IL-10−/− and IBD IL-10−/− mice, and their frequency. One-way ANOVA followed by a Dunnett's test (*p<0.05 and **p<0.01 versus healthy IL-10−/−). Data represent pooled results from two independent experiments with 5 mice per group. (C) The gene expression of RANK-L, Mcp1 and Mip1α, was analysed by quantitative PCR on sorted WT CD45− Ter119− CD3− CD19− CD31− stromal cells cocultured 2 days with sorted Th17 TNFα+ cells or CD4+IL-17− T cells. Data represent pooled results from three independent experiments *p<0.05 was calculated with a Student t test. (D) 106 cells of sorted in vitro differentiated Th17 TNFα+ cells or CD4+IL-17− T cells were transferred into Rag1−/− mice previously immunised intraperitoneally twice with 50 µg of ovalbumin protein. Seven days later, flow cytometry analysis of CD11b+Ly6C+ cells from BM cells of Rag1−/− mice were performed and frequency of these cells are indicated. *p<0.05 was calculated with a Student t test. Data represent pooled results from two independent experiments.
BM mesenchymal stromal cells (BM MSC) have been reported to be a major producer of soluble factors, including chemokines. We analysed, therefore, the expression of chemokines known to attract inflammatory monocytes, by purified BM MSCs from control mice and cocultured with sorted Th17 TNF-α+ cells. These stromal cells expressed higher levels of Mcp1, Mip1α and Rank-l compared with stromal cells in coculture with sorted CD4+IL-17− T cells (figure 5C). Confirming the role of Th17 TNFα+ cells in the recruitment of inflammatory monocytes, adoptive transfer of these cells into Rag1−/− mice significantly increased the percentage of CD11b+Ly6C+ monocytes in the BM after 7 days (figure 5D). Altogether, these results indicate that the recruitment of CD11b+Ly6C+ inflammatory monocytes in BM during IBD is associated with the presence of Th17 TNFα+ cells that induce Mcp1 and Mip1α expression by BM MSCs.
Identification of Th17 TNF-α+ cells in patients with Crohn's disease
We finally extended our analysis of osteoclastogenic Th17 TNF-α+ cells to human patients affected with CD. Classical human Th17 cells are known to express the CCR6 chemokine receptor.26 Thus, CD4+ T cells were purified from PBMCs of healthy subjects and patients with CD according to their CCR6 expression. CCR6+ and CCR6− cells were assessed by flow cytometry for their ability to produce IL-17 and TNF-α in response to stimulation with PMA and ionomycin. As shown in figure 6, Th17 TNF-α+ cells are included only in the CCR6+ fraction from PBMCs of patients with CD. By contrast, peripheral blood CD4+ CCR6− T cells from both healthy subjects and patients with CD were not capable of producing IL-17 (figure 6A,B). To evaluate the osteoclastogenic effect of human Th17 TNF-α+, we have cloned these cells from the peripheral blood of a patient with CD (figure 6C). None of these T cell clones did produce IFN-γ (data not shown). Coculture of three of these clones with RAW264.7 cells resulted in the induction of giant multinucleated TRAP+ OCLs (≥9 nuclei) (figure 6D). Furthermore, unlike Th1 cells, the Th17 TNF-α+ cell clones induced osteoclast differentiation from allogeneic monocytes (figure 6E).
Identification of Th17 tumour necrosis factor (TNF)-α+ T cells in Crohn's disease (CD). (A) Cytokine expression of CD4+ T cells from peripheral blood of healthy donor or patients with CD was assessed by intracellular flow cytometry after in vitro restimulation with phorbol myristate acetate and ionomycin. (B) Frequency of Th17 TNFα+ CCR6+ T cells in peripheral blood from healthy donor or patients with CD. *p<0.05. (C) Intracytoplasmic analysis of CD4+ T clones isolated from the peripheral blood of a patient with CD. (D) Human Th17 TNF-α+ T cell clones were cultured with RAW264.7 cells in the presence of phytohaemagglutinin (1 µg/mL) for 8 days. (E) The same clones were also cultured with allogeneic monocytes in the presence or not of anti-CD3/CD28 beads for 8 days. Th1 cells were used as controls. Osteoclasts were analysed for their tartrate-resistant acid phosphatase (TRAP) activity. Representative TRAP staining of clone 1. Similar results were obtained with the other Th17 TNF-α+ cell clones (n=3). (F) The gene expression of Mcp1 and Mip1α was analysed by quantitative PCR on primary human osteoblast (hOST) cells cocultured 2 days with supernatants from each T cell clone. Data represent pooled results from two independent experiments *p<0.05 was calculated with a Student t test.
Similarly, human osteoblasts (hOST) cultured in the presence of supernatants from these T cell clones expressed high levels of Mcp1 and Mip1α (figure 6F). These results suggested that the human Th17 TNF-α+ cells probably have a similar mechanism of action than the murine equivalent population.
Discussion
The mechanisms linking inflammation with bone destruction are still incompletely understood, in particular, with respect to the role of CD4+ T cells. In the present study, we have identified, for the first time, the Th17 TNFα+ T cells as potent inducers of OCL differentiation in the context of CD in mouse and human. Using two well-characterised models of IBD that develop bone loss, we showed that BM CD4+ T cells producing IL-17 and TNF-α, but not IFN-γ, induce OCL differentiation in the absence of exogenous osteoclastogenic factors. These cells are associated with the development of inflammation, have a Th17 gene expression signature, express M-CSF and mRANK-L and have an effector memory phenotype. Therefore, their presence and activity are associated with the development of osteopenia in an inflammatory environment.
Many studies suggest that Th17 cells exhibit flexibility of function and can modify their phenotype, including the acquisition of IFN-γ production, during induction of colitis after adoptive transfer of CD4+CD45RBhi T cells into immunodeficient hosts.20 Our results extend these studies by demonstrating, in the lymphoid organs and the intestine of IBD IL-10−/− mice, the presence of IL-17A and IFN-γ producing-CD4+ T cells that are absent in WT and healthy IL-10−/− mice. Furthermore, we showed here the emergence Th17 TNF-α+ cells that do not produce IFN-γ in the lymphoid organs, the intestine and mainly the BM of IBD IL-10−/− mice which are prominent compared with WT and healthy IL-10−/− mice. Together, these results underlined the inflammatory nature of the Th17 TNF-α+ cell population in the BM.
A picture emerged from the literature (reviewed27) to define what we believe to be osteoclastogenic T cells: first, osteoclastogenic T cells should not produce a large amount of IFN-γ. Second, they should trigger local inflammation and the production of inflammatory cytokines, including TNF-α, that induce RANK-L expression on mesenchymal stromal cells. Third, they should express RANK-L and might directly participate in increased osteoclastogenesis. Our study indicates that in IBD IL-10−/− mice with advanced bone destruction, the Th17 TNF-α+ cell population represents the long-sought-after osteoclastogenic T cell subset that fulfils all the criteria mentioned above.
In this study, we also identified a potential role of IL-7 in the survival of these osteoclastogenic Th17 TNF-α+ cells. First, we have shown that after 7 days of IL-7 treatment the frequency of Th17 TNFα+ cells was increased more than fourfold. This cytokine also strongly enriched the differentiation of these cells in vitro under Th17 cell polarisation-inducing conditions from naive T cells. These results are consistent with those of Watanabe et al, who demonstrated that, in the colitis model induced by adoptive transfer of CD4+CD45RBhi T cells into Rag1−/− mice, IL-7 is essential for the persistence of colitogenic memory CD4+ T cells in the BM.22 Furthermore, the accumulation of memory CD4+ T cells in the BM was decreased significantly in IL-7−/− Rag1−/− mice reconstituted with colitogenic memory CD4+ T cells from IBD mice.22 In line with this report, our results show an increased expression of IL-7 in purified BM MSCs from IBD IL-10−/− mice compared with cells from healthy IL-10−/− and control mice, suggesting a role of BM MSCs expressing IL-7 in the maintenance of the Th17 TNFα+ cells. In this sense, more recently, Nemoto et al have demonstrated that BM MSCs are a major source of IL-7 and play a pathological role in IBD by forming the niche for colitogenic CD4 memory T cells in the BM of Rag1−/− mice transferred by naive CD4+ T cells.28 Collectively, these findings suggest a role of IL-7 in the survival and maintenance of the osteoclastogenic Th17 TNFα+ cells.
Our observations in IBD IL-10−/− mice showed that enhanced osteoclastogenesis is associated with the accumulation of CD11bhiLy6Chi inflammatory monocytes and CD11b−/lo Ly6Chi cells that have been described as OCL precursors that accumulate in rheumatoid arthritis.25 Interestingly, our mechanistic analysis revealed that, on one hand, the osteoclastogenic Th17 TNFα+ cells promote a recruitment of inflammatory monocytes representing OCL precursors in the BM during inflammation, suggesting a possible role in osteoclastogenesis. This local recruitment was possible by the increased expression of MCP-1 and MIP1α by stromal cells. In agreement with this, Powrie's group has shown an excessive accumulation of inflammatory myeloid cells, especially monocytes, in the colon and in the BM of colitic mice.29 Indeed, haematopoietic progenitor cells regulate this myeloid accumulation in the BM during colitis.29 On the other hand, the Th17 TNFα+ cells are able to induce the expression of osteoclastogenic factors, such as RANK-L, by BM MSCs. Altogether, the homing of Th17 TNFα+ cells in BM allows to establish a vicious circle of bone destruction by inducing the expression of the osteoclastogenic factors and promoting the recruitment of OCL precursors.
In humans, the expression of CCR6 has been tightly associated with Th17.15 In light of this observation, we have explored the presence of Th17 TNFα+ cells in the blood of healthy donors and patients with CD. Our results showed an association between the presence of Th17 TNFα+ cells and CD. Furthermore, Th17 TNFα+ cell clones from patients with CD were able to induce OCL differentiation.
Collectively, our study provides the first detailed phenotypical and functional characterisation of the osteoclastogenic Th17 TNFα+ cells linking chronic intestinal inflammation and bone destruction. This subset of osteoclastogenic Th17 TNFα+ cells exists in vivo in mice and in humans and can be induced and enriched in vitro in the presence of IL-7. We believe that dissecting how and when the Th17 TNFα+ cells migrate into the BM could improve our general understanding of the exacerbation of inflammation in IBD. Finally, our results highlight that BM Th17 TNFα+ cells could be a potential target for innovative immunotherapeutic strategies against chronic intestinal inflammation.
Acknowledgments
We thank Dr F Lezot and Dr D Heymann for their help in the μCt analysis (histomorphometry facility, UMR 957 Medical Faculty, Nantes University, Nantes, France).
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
LI and AB contributed equally.
Contributors TC researched data, reviewed the manuscript and contributed to discussions, LI and AB performed experiments and analysed data for revisions, EB-B, JP and GAB did experiments and analysed the data. HY and MR contributed to discussions and reviewed the manuscript, NA provided and processed blood from patients, XH contributed to discussions and reviewed the manuscript. CB-W designed the research and reviewed the manuscript. AW designed the research, researched data, wrote and reviewed the manuscript.
Funding This work was supported by grants from the Arthritis Fondation, the Fondation pour la Recherche Médicale (Equipe FRM DEQ20130326467), the Agence Nationale pour la Recherche and the Société Française de Rhumatologie.
Competing interests None.
Ethics approval The local ethics committee of the Hospital center of Nice University, Nice, France.
Provenance and peer review Not commissioned; externally peer reviewed.