Article Text

Download PDFPDF

Synovial macrophage M1 polarisation exacerbates experimental osteoarthritis partially through R-spondin-2
  1. Haiyan Zhang1,
  2. Chuangxin Lin1,
  3. Chun Zeng1,
  4. Zhenyu Wang1,
  5. Hua Wang2,
  6. Jiansen Lu1,
  7. Xin Liu1,
  8. Yan Shao1,
  9. Chang Zhao1,
  10. Jianying Pan1,
  11. Song Xu3,
  12. Yue Zhang1,4,
  13. Denghui Xie1,
  14. Daozhang Cai1,
  15. Xiaochun Bai1,4
  1. 1 Department of Orthopedics, Academy of Orthopedics - Guangdong Province, Orthopedic Hospital of Guangdong Province, The Third Affiliated Hospital of Southern Medical University, Guangzhou, China
  2. 2 Key Laboratory of Tropical Diseases and Translational Medicine of the Ministry of Education, and Hainan Provincial Key Laboratory of Tropical Medicine, Hainan Medical College, Haikou, China
  3. 3 Department of Orthopedics and Arthroplasty, Nanfang Hospital affiliated to Southern Medical University, Guangzhou, China
  4. 4 State Key Laboratory of Organ Failure Research, Department of Cell Biology, School of Basic Medical Science, Southern Medical University, Guangzhou, China
  1. Correspondence to Professor Daozhang Cai and Dr Xiaochun Bai, Department of Orthopedics, Academy of Orthopedics - Guangdong Province, Orthopedic Hospital of Guangdong Province, The Third Affiliated Hospital of Southern Medical University, Guangzhou, China; cdz{at}smu.edu.cn, baixc15{at}smu.edu.cn

Abstract

Objectives To investigate the roles and regulatory mechanisms of synovial macrophages and their polarisation in the development of osteoarthritis (OA).

Methods Synovial tissues from normal patients and patients with OA were collected. M1 or M2-polarised macrophages in synovial tissues of patients with OA and OA mice were analysed by immunofluorescence and immunohistochemical staining. Mice with tuberous sclerosis complex 1 (TSC1) or Rheb deletion specifically in the myeloid lineage were generated and subjected to intra-articular injection of collagenase (collagenase-induced osteoarthritis, CIOA) and destabilisation of the medial meniscus (DMM) surgery to induce OA. Cartilage damage and osteophyte size were measured by Osteoarthritis Research Society International score and micro-CT, respectively. mRNA sequencing was performed in M1 and control macrophages. Mice and ATDC5 cells were treated with R-spondin-2 (Rspo2) or anti-Rspo2 to investigate the role of Rspo2 in OA.

Results M1 but not M2-polarised macrophages accumulated in human and mouse OA synovial tissue. TSC1 deletion in the myeloid lineage constitutively activated mechanistic target of rapamycin complex 1 (mTORC1), increased M1 polarisation in synovial macrophages and exacerbated experimental OA in both CIOA and DMM models, while Rheb deletion inhibited mTORC1, enhanced M2 polarisation and alleviated CIOA in mice. The results show that promoting the macrophage M1 polarisation leads to exacerbation of experimental OA partially through secretion of Rspo2 and activation of β-catenin signalling in chondrocytes.

Conclusions Synovial macrophage M1 polarisation exacerbates experimental CIOA partially through Rspo2. M1 macrophages and Rspo2 are potential therapeutic targets for OA treatment.

  • osteoarthritis
  • synovitis
  • knee osteoarthritis
  • chondrocytes

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Osteoarthritis (OA) is a highly prevalent and degenerative joint disorder, causing pain and functional disability, and representing an enormous clinical and financial burden.1 OA is traditionally characterised by erosion of cartilage and subchondral bone sclerosis. It is now well established that OA is a disease which affects all joint tissues, characterised by progressive degeneration of the articular cartilage, vascular invasion of articular surface, subchondral bone remodelling, osteophyte formation and synovial inflammation (synovitis).2 Accumulating evidence suggests that synovial inflammation is correlative with the pathogenesis and progression of OA.3–5 Synovitis is significantly associated with OA severity,6 and 38% of patients with OA of grades 2–3 according to the Kellgren-Lawrence (KL) classification have infrapatellar synovitis, as do up to 83% with KL grade 4.7

Normal synovium consists of two distinct tissue layers. One is the intimal lining layer with two to three layers of macrophages and fibroblast-like synoviocytes; the other is the synovial sublining layer, composed of fibrous connective tissue and blood vessels, with few lymphocytes or macrophages.8 Macrophage accumulation in the intimal lining, reflecting mostly proliferative synovial tissue, is the principal morphological characteristic of synovitis.9 Macrophages can be broadly classified into classically activated M1 macrophages and alternatively activated M2 macrophages, in response to stimuli from their microenvironment.10 M1 macrophages are activated by interferon-γ and lipopolysaccharide (LPS) or tumour necrosis factor alpha (TNF-α), secreting large amounts of proinflammatory cytokines and mediators, such as TNF-α, interleukin (IL)-1, IL-6 and IL-12.11 M2 macrophages, also known as wound-healing macrophages, have been further divided into specific subtypes and have anti-inflammatory activity.12 13 Earlier studies have shown that macrophages accumulate and become polarised (M1 or M2) in the synovium and articular cavity during OA development, implying a correlation between macrophages and OA.14 15 However, a recent study demonstrated that macrophage depletion (with significantly fewer macrophages of both M1 and M2) increased inflammation and did not attenuate the severity of OA in obese macrophage Fas-induced apoptosis-transgenic mice.16 These studies indicate that OA is a complex disease that encompasses many pathways leading to degenerative conditions that progress to a common final outcome of end-stage OA, and some of these pathways involve inflammation especially macrophages. However, the role of macrophages, their polarisation in OA development and the underlying mechanisms are still unknown.

Emerging evidence reveals that the mammalian target of rapamycin (mTOR) pathway plays key roles in macrophage polarisation.17–19 mTOR is a central regulator of cellular metabolism and forms two distinct complexes, mTORC1 and mTORC2. Tuberous sclerosis complex 1/2 (TSC1/2) exhibits a selective guanosine triphosphatase-activating protein activity towards the small GTPase, Ras homologue enriched in brain (Rheb), which binds to and activates mTORC1.20 21 Several reports have shown that mice with mTORC1 constitutive activation in the myeloid lineage by deletion of TSC1 are refractory to IL-4-induced M2 polarisation, but mount a hyperinflammatory response to LPS, while mice with mTORC1 abruption have reduced M1 polarisation.17 18 Such divergence in M1 versus M2 activation is also observed in a model of TSC-deficient bone marrow-derived macrophages.22 Thus, mice with mTORC1 activation or inhibition in the myeloid lineage are desirable models to explore the role of macrophages and their polarisation in OA.

In this study, we used mice with Tsc1 or Rheb1 deletion in the myeloid lineage to generate collagenase-induced osteoarthritis (CIOA), a high synovial activation OA model.23 24 Destabilisation of the medial meniscus (DMM) surgery, a low synovial activation OA model, was also performed in mice with constitutive mTORC1 activation in the myeloid lineage (TSC1KO). We found that TSC1KO mice exhibited enhanced M1 macrophage polarisation and exacerbated cartilage degeneration and osteophyte formation in experimental OA. Conversely, deletion of Rheb1 in the myeloid lineage (Rheb1KO) enhanced synovial macrophage M2 polarisation and attenuated OA. We further identified R-spondin-2 (Rspo2) as a protein secreted by M1 macrophages which exhibits strong effects in OA development. Thus, M1 macrophages or Rspo2 represents novel therapeutic targets for OA treatment.

Materials and Methods

Detailed experimental procedures are described in supplemetary materials and methods (see online supplemetary file 9).

Supplementary file 9

Results

M1 but not M2-polarised macrophages accumulate in synovial tissue of patients with OA and OA mice

To explore the role of synovial macrophages in OA development, we first examined synovial inflammation and macrophage phenotypes in human OA synovial tissue. Consistent with previous results,25 26 high levels of synovial hyperplasia and abundant cell infiltration were observed in human OA synovial tissue, combined with significantly higher synovitis scores than that of normal controls (figure 1A,B). We further identified the accumulation and phenotypic characterisation of macrophages in OA synovial tissue. Compared with controls, marked elevation of F4/80 (macrophage marker)-positive cells was detected in both layers of OA synovial tissue, together with a significant increase of iNOS (M1-like macrophage marker)-positive cells predominantly in the intimal lining layer. In contrast, the proportion of cells positive for the M2-like macrophage marker CD206 in OA synovium showed a slight decrease (figure 1A,C). These results were further confirmed by qPCR analysis of digested synovial tissue samples from patients with OA and controls (see online supplementary figure 1A). These findings demonstrate that M1 but not M2-polarised macrophages accumulate in the synovium of OA joints from older patients and patients with OA as compared with young and normal controls.

Supplementary file 1

Figure 1

Macrophage polarisation in synovial tissue from patients with OA and OA mice. (A) Representative images of H&E staining (top), immunofluorescence of F4/80 and iNOS (middle), immunohistochemistry of CD206 (bottom) in normal and OA human synovial tissues. Scale bars: 50 µm, 200 µm. (B) Quantification of synovitis score in normal (n=7) and OA (n=9) human synovial tissues. Student’s t-test. **P<0.01. (C) Quantification of F4/80, iNOS and CD206-positive macrophages as a proportion of total macrophages in normal and OA human synovial tissues. Student’s t-test. **P<0.01. (D) Representative images of H&E staining (top), immunofluorescence of F4/80 and iNOS (middle), immunohistochemistry of CD206 (bottom) in synovium of controls and mice at 7 and 28 days after intra-articular injection of collagenase (CIOA surgery). Scale bar: 50 µm. (E) Quantification of synovitis score in synovium of CIOA mice. n=9 per group. One-way analysis of variance (ANOVA) and Tukey’s multiple comparison test. **P<0.01. (F) Quantification of F4/80, iNOS and CD206-positive macrophages as a proportion of total macrophages in synovium of CIOA mice. One-way ANOVA and Tukey’s multiple comparison test. *P<0.05; **P<0.01. Boxed area is enlarged in the bottom right corner. F4/80: green; iNOS: green; DNA: blue. Data are shown as mean±SEM. AC, articular cavity; CC, calcified cartilage; CIOA, collagenase-induced osteoarthritis; F, femur; HC, hyaline cartilage; L, synovial lining macrophages; M, meniscus; NS, not significant; OA, osteoarthritis; SL, synovial sublining macrophages.

We then assessed the accumulation of macrophages in synovium and carried out phenotypic characterisation of synovium near the patellofemoral junction in the CIOA mouse model (intra-articular injection of collagenase) (see online supplementary figure 1B). Similarly, little proliferation was observed in the synovium of control knees, but the synovium became hyperplastic and hypertrophic accompanied with degeneration and loss of structure of the cartilage of OA knees (figure 1D,E). Of note, collagenase injection consistently increased the number of macrophages in both lining layers of synovial tissue. The percentage of M1-like macrophages was increased at both 7 and 28 days after surgery. M2-like macrophages were present in both layers at 7 days after surgery, but existed only in some areas of the sublining layer at 28 days, with a significant decrease in number compared with controls (figure 1D,F). Taken together, these findings demonstrate that macrophages accumulate in OA synovial tissue, with enhanced M1-like and reduced M2-like polarisation, suggesting their potential role in the pathogenesis and development of OA.

Constitutive activation of mTORC1 in macrophages enhances their M1 polarisation in OA synovium

Mice with myeloid lineage specific activation of mTORC1, which present an enhanced M1 response and spontaneously develop M1-related inflammatory disorders, have been used as a model to investigate the role of polarised macrophages.17 However, the activity of mTORC1 in macrophages and its contribution to macrophage polarisation in OA synovial tissue are unknown. Interestingly, enhanced phosphorylation of S6 (S235/236) (a downstream effecter of mTORC1) and its colocalisation with F4/80 was detected in human OA synovial tissue (figure 2A,B). The polarised macrophages and activation of mTORC1 signalling were also detected in synovium of young and aged mice (see online supplementary figure 2). Furthermore, in the CIOA mouse model, macrophage mTORC1 activity was significantly enhanced when synovial hyperplasia was florid, while p-S6-positive macrophages were almost undetectable in control synovium (figure 2C). These data suggest that along with macrophage accumulation, synovial macrophage mTORC1 was activated during OA.

Supplementary file 2

Figure 2

mTORC1 signalling is activated in M1-polarised synovial macrophages from patients with OA and CIOA mice. (A) Representative images of coimmunostaining of F4/80 and pS6 in human synovium and quantitative analysis of pS6-positive macrophages compared with total macrophages. n=7 for human normal synovium, n=9 for human OA synovium. Student’s t-test. *P<0.05. Scale bar: 50 µm. (B) Western blot of pS6 in human synovial tissue. (C) Representative images of coimmunostaining of F4/80 and pS6 in synovium of controls and mice at 7 and 28 days after CIOA surgery, and quantitative analysis of pS6-positive macrophages compared with total macrophages. n=9 per group. Scale bar: 50 µm. One-way analysis of variance (ANOVA) and Tukey’s multiple comparison test. **P<0.01. (D) Representative images of coimmunostaining of F4/80 and pS6, immunofluorescence of iNOS and immunohistochemistry of CD206 in synovium of TSC1KO and control mice at 4 weeks after CIOA surgery, and quantitative analysis of pS6, iNOS and CD206-positive macrophages compared with total macrophages. Scale bar: 50 µm; n=10 per group. Boxed area is enlarged in the bottom right corner. Student’s t-test. *P<0.05; **P<0.01. F4/80: green; pS6: red; iNOS: green; DNA: blue. Data are shown as mean±SEM. AC, articular cavity; CIOA, collagenase-induced osteoarthritis; F, femur; L, synovial lining macrophages; M, meniscus; NS, not significant; OA, osteoarthritis; SL, synovial sublining macrophages.

In order to determine whether mTORC1 activation contributes to macrophage polarisation in synovium, we generated mice with conditional ablation of the Tsc1 gene in myeloid cells, using a Cre expression cassette under the control of the lysozyme proximal promoter (TSC1KO), in which the target gene was specifically ablated from macrophages and neutrophils. In gross appearance, TSC1KO mice exhibited a slightly decreased body weight and length but no significant difference compared with control. No discernible differences in the morphology or organisation of the articular cartilage or growth plates were apparent between the TSC1KO mice and the littermate control mice at 8 weeks of age (see online supplementary figure 3A–G). Notably, synovial macrophages demonstrated high activation of mTORC1, a significant increase of M1-like polarisation and a decrease of M2-like polarisation in synovium of TSC1KO mice 4 weeks after CIOA surgery compared with control mice (figure 2D).

Supplementary file 3

Synovial macrophage M1 polarisation exacerbates cartilage damage and osteophyte formation in collagenase-induced OA

We further examined the functional role of accumulated synovial M1 macrophages in the development of OA. In the CIOA mouse model, the anterior cruciate ligament (ACL) still existed with serious damage at the end of the experiment and TSC1KO mice showed increased synovitis score compared with control mice at both 4 and 8 weeks after surgery (see online supplementary figure 3H–J). Notably, TSC1KO mice demonstrated a significant higher OA score than controls. Four weeks after surgery, in TSC1KO mice the articular cartilage was substantially thinner, the area of surface fibrillation was larger and the chondrocytes showed an abnormal distribution compared with control mice. At 8 weeks after surgery, more severe cartilage degeneration (figure 3A) and increased MMP13 expression (figure 3B) were observed in TSC1KO mice compared with their littermate controls. Furthermore, TSC1KO mice also exhibited accelerated osteophyte formation at 8 weeks after surgery. We employed micro-CT analysis and three-dimensional (3D) modelling of knee joints to comprehensively evaluate the osteophytosis. Compared with controls, TSC1KO mice developed larger periarticular osteophytes with significantly increased volume and surface area of osteophytes at the end stage of OA (figure 3C). The acceleration of OA development in TSC1KO mice was further verified by DMM surgery (see online supplementary figure 4). These findings suggest that macrophage M1 polarisation exacerbates cartilage degeneration and osteophyte formation during OA progression in mice.

Supplementary file 4

Figure 3

Synovial macrophage M1 polarisation promotes osteoarthritis (OA) development in mice. (A) Safranin O/fast green staining of joints from TSC1KO and control mice at 4 (n=10) and 8 weeks (n=9) after intra-articular injection of collagenase, and quantitative analysis of Osteoarthritis Research Society International (OARSI) scale. Red boxed area on the left of the tibial plateau represents osteophyte formation. Scale bars: 50 µm, 100 µm; Student’s t-test. **P<0.01. (B) Representative images of immunohistochemical staining of MMP13 in articular chondrocytes of TSC1KO and control mice at 8 weeks after collagenase-induced osteoarthritis (CIOA) surgery, and quantitative analysis of MMP13-positive chondrocytes compared with total chondrocytes. Scale bar: 50 µm; n=9. Boxed area is enlarged in the bottom right corner. Student’s t-test. **P<0.01. (C) Micro-CT scan and three-dimensional reconstruction of the knee joint from TSC1KO and control mice at 4 weeks (n=7) and 8 weeks (n=8) after CIOA surgery, and quantitative analysis of osteophyte score and volume of region of interest (ROI). The ROI is marked in red for periarticular osteophytes; Student’s t-test. **P<0.01. Data are shown as mean±SEM. NS, not significant.

Synovial macrophage M2 polarisation prevents OA development

We next examined the role of M2 macrophages in OA development. Mice with conditional ablation of the Rheb1 gene, an upstream activator of mTORC1, in myeloid cells (Rheb1KO) were generated, in which M2 polarisation was enhanced while M1 polarisation was reduced in alveolar macrophages as we reported previously.27 There were no discernible differences either in gross appearance or in the organisation of the articular cartilage and growth plates between the Rheb1KO mice and littermate controls (see online supplementary figure 5A–G). As expected, Rheb1KO mice displayed inhibition of mTORC1 activity in synovial macrophages. In addition, there were far more CD206-positive cells in the synovium of Rheb1KO mice than in controls, together with a decreased proportion of iNOS-positive cells (figure 4A). Thus, disruption of Rheb1 increases synovial macrophage M2 polarisation but decreases M1 polarisation in mice.

Supplementary file 5

Figure 4

Synovial macrophage M2 polarisation prevents osteoarthritis (OA) development in mice. (A) Representative images of coimmunostaining of F4/80 and pS6, immunofluorescence of iNOS and immunohistochemistry of CD206 in synovium of Rheb1KO and control mice at 4 weeks after collagenase-induced osteoarthritis (CIOA) surgery, and quantitative analysis of pS6, iNOS and CD206-positive macrophages compared with total macrophages. Scale bar: 50 µm; n=11 per group. Student’s t-test. **P<0.01. (B) Safranin O/fast green staining of joints from Rheb1KO and control mice at 4 (n=11) and 8 (n=8) weeks after intra-articular injection of collagenase, and quantitative analysis of OARSI scale. Red boxed area on the left of the tibial plateau represents osteophyte formation. Scale bars: 50 µm, 100 µm; Student’s t-test. *P<0.05. (C) Representative images of immunohistochemical staining of MMP13 in articular chondrocytes of Rheb1KO mice at 8 weeks after CIOA surgery (n=8) and quantitative analysis of MMP13-positive chondrocytes compared with total chondrocytes. Scale bar: 50 µm; Student’s t-test. *P<0.05. Boxed area is enlarged in the bottom right corner. F4/80: green; pS6: red; iNOS: green; DNA: blue. Data are shown as mean±SEM. AC, articular cavity; F, femur; L, synovial lining macrophages; M, meniscus; NS, not significant; OARSI, Osteoarthritis Research Society International; SL, synovial sublining macrophages.

CIOA surgery was further performed on 8-week-old Rheb1KO mice and their littermate controls. Rheb1KO mice showed decreased synovial inflammation at 4 and 8 weeks after CIOA surgery along with injured ACL (see online supplementary figure 5H–J). No significant difference in knee OA Osteoarthritis Research Society International (OARSI) score between Rheb1KO and control mice was observed at 4 weeks after surgery. By 8 weeks after surgery, however, the OARSI score was significantly reduced in Rheb1KO mice (figure 4B). Accordingly, the expression of MMP13 in the cartilage of Rheb1KO mice was dramatically reduced compared with that of controls (figure 4C). Moreover, Rheb1KO mice developed much smaller osteophytes at 8 weeks after surgery (figure 4B). Together, these data suggest that synovial macrophage M2 polarisation prevents the development of collagenase-induced OA in mice.

M1-polarised macrophages produce inflammatory cytokines/enzymes and promote hypertrophic chondrocyte differentiation and maturation in vitro

To explore the mechanisms through which M1 macrophages promote OA progression, we performed mRNA sequencing to identify cytokines/enzymes involved in synovitis and cartilage degeneration by comparing the cultured macrophages from TSC1KO mice and their littermate controls (RNA sequencing data have been deposited in the NCBI SRA database under accession codes SRR6660735 and SRR6660734). Among 2279 mRNAs identified, 1153 mRNAs were upregulated and 1126 were downregulated in macrophages from TSC1KO mice compared with their littermate controls (see online supplementary table 1). The cytokines IL-1, IL-6 and TNF-α were upregulated in TSC1KO macrophages in response to LPS and have been previously shown to be produced by M1 macrophages and to contribute to OA development.28 The upregulation of these inflammatory cytokines was further confirmed by Q-PCR and ELISA (see online supplementary figure 6A,B).

Supplementary file 6

Supplementary file 7

Articular chondrocyte activation, involving aberrant proliferation and hypertrophic differentiation, is essential for OA initiation and progression.29 30 Therefore, we examined the effect of M1 macrophages on chondrogenesis and differentiation. ATDC5 chondroprogenitor cells were treated with conditioned medium (CM) from cultured macrophages from TSC1KO or control mice together with insulin, transferrin, selenium (ITS) to induce chondrogenesis. Decreased expression of Sox9, a master regulator of chondrogenesis, as well as the chondrocyte markers Col2a1 and Acan, was found in TSC1KO CM-treated ATDC5 cells. In contrast, CM from TSC1KO macrophages promoted terminal differentiation of chondrocytes, as evidenced by enhanced Col10a1 and Runx2 expression at day 14 after ITS treatment (figure 5A). Furthermore, toluidine blue staining to show matrix mineralisation further confirmed that TSC1KO macrophage CM promoted chondrocyte maturation (figure 5B). Taken together, these findings demonstrate that M1-polarised macrophages induced by mTORC1 produce high levels of inflammatory cytokines and promote hypertrophic chondrocyte differentiation and maturation, suggesting that it plays a critical role in cartilage degeneration during OA.

Figure 5

M1-polarised macrophages produce Rspo2 and promote hypertrophic chondrocyte terminal differentiation and maturation. (A) Quantitative PCR analysis of Sox9, Col2a1, Acan, Col10a1 and Runx2 in ATDC5 cells treated with conditioned medium (CM) from macrophage cultures from TSC1KO or control mice. n=6 per group. Student’s t-test. **P<0.01. (B) Representative images of toluidine blue staining in ATDC5 cells treated with CM from macrophage cultures from TSC1KO or control mice. (C) Representative images of immunohistochemistry of Rspo2 in synovium of TSC1KO and control mice at 4 weeks after collagenase-induced osteoarthritis (CIOA) surgery, with quantification of Rspo2-positive cells as a proportion of total cells. Scale bar: 50 µm; n=10 per group. Student’s t-test. **P<0.01. (D) Rspo2 concentrations assessed by ELISA in serum of TSC1KO and control mice; n=7 per group. Student’s t-test. *P<0.05. (E) Representative images of immunohistochemical staining of Lgr5, and immunofluorescence of β-catenin in articular chondrocytes of TSC1KO and control mice, with quantification of Lgr5 and β-catenin-positive chondrocytes as a proportion of total articular chondrocytes. Scale bar: 50 µm; n=8 per group. Student’s t-test.  **P<0.01. Boxed area is enlarged in the bottom right corner. β-catenin: green; DNA: blue. Data are shown as mean±SEM. F, femur; L, synovial lining macrophages; M, meniscus; NS, not significant; SL, synovial sublining macrophages.

M1-polarised macrophages produce Rspo2 to promote OA development in mice

We next sought to identify the M1 macrophage-derived cytokines/enzymes responsible for cartilage degradation and osteoblast formation during OA. In addition to IL-1, IL-6 and TNF-α, Rspo2, a cytokine upregulated 33.6-fold in TSC1KO mice, and further analysed by Q-PCR and ELISA (see online supplementary figure 6C,D), captured our attention. Rspo2 has been identified as an agonist for Wnt signalling and synergises with Wnt to activate β-catenin and facilitate chondrocytes and osteoblast differentiation,31 but its role in macrophage and OA development has not been reported. We confirmed that Rspo2 was expressed in synovial tissue (figure 5C) and serum (figure 5D) of TSC1KO mice. Furthermore, Lgr5, a receptor of Rspo2, was highly expressed in articular cartilage chondrocytes (figure 5E). Although no significant difference in the number of Lgr5-positive cells was detected between the articular cartilage of TSC1KO and control mice, TSC1KO macrophages produced much higher levels of Rspo2 than controls, manifested by a marked increase of Rspo2 mRNA in TSC1KO macrophages and Rspo2 protein in TSC1KO mouse serum. Importantly, β-catenin-positive cells were significantly enhanced in articular cartilage of TSC1KO mice compared with controls (figure 5E).

We then treated ATDC5 cells with or without anti-Rspo2 antibody to neutralise endogenous Rspo2 after stimulation with ITS and CM from TSC1KO macrophage culture. As expected, the effects of M1 macrophage CM on expression of Sox9, Col2a1 and Acan were rescued by Rspo2 antibody on day 5, and the upregulation of Col10a1 and Runx2 was significantly reduced on day 14. The stimulatory effect of M1 macrophages on matrix mineralisation was also abolished by treatment with anti-Rspo2 (see online supplementary figure 7A,B).

Supplementary file 8

To further identify the role of Rspo2 in development of the OA phenotype induced by M1 macrophages in vivo, anti-Rspo2 antibody was injected intra-articularly to TSC1KO mice to neutralise endogenous Rspo2 after CIOA surgery. Interestingly, the antibody against Rspo2 significantly reduced the destruction of the cartilage and the OARSI score compared with the vehicle-treated mice. Micro-CT analysis and 3D modelling of knee joints revealed that anti-Rspo2 attenuated osteophytosis (figure 6A). Moreover, intra-articular injection of rhRspo2 in C57 mice promoted cartilage degradation and osteophyte formation (figure 6B). Consistently, articular cartilage β-catenin-positive cells were enhanced by rhRspo2 and reduced by anti-Rspo2 antibody injection (figure 6C and online supplementary figure 7C). These data suggest that M1 macrophages promote cartilage degeneration and osteophyte formation partly through secretion of Rspo2.

Figure 6

M1-polarised macrophages exacerbate experimental osteoarthritis (OA) partially through Rspo2. (A) Safranin O/fast green staining, micro-CT scan and three-dimensional reconstruction of joints from TSC1KO mice at 8 weeks after intra-articular injection of collagenase with or without anti-Rspo2 antibody, and quantitative analysis of OARSI scale and volume of the ROI. The ROI is marked in red for periarticular osteophytes; n=8 per group. Student’s t-test. *P<0.05. (B) Safranin O/fast green staining, micro-CT scan and three-dimensional reconstruction of joints from mice at 8 weeks after intra-articular injection of collagenase with or without rhRspo2, and quantitative analysis of OARSI scale and volume of ROI. The ROI is marked in red for periarticular osteophytes; n=8 per group. Student’s t-test. *P<0.05. (C) Representative images of immunofluorescence of β-catenin in articular chondrocytes of mice described in (A) and (B), with quantification analysis of β-catenin-positive chondrocytes as a proportion of total articular chondrocytes. Scale bar: 50 µm; n=8 per group. One-way analysis of variance (ANOVA) and Tukey’s multiple comparison test. **P<0.01. (D) Model of Rspo2 secreted by M1-polarised macrophages in regulating hypertrophic chondrocyte terminal differentiation and osteophyte formation during OA. β-catenin: green; DNA: blue. Data are shown as mean±SEM. CIOA, collagenase-induced osteoarthritis; IL, interleukin; OARSI, Osteoarthritis Research Society International; ROI, region of interest; TNF, tumour necrosis factor.

Discussion

Using mouse models with enhanced M1 or M2-polarised macrophages, we identified for the first time the critical role of synovial macrophage M1 polarisation in the development of OA. We showed that synovial M1 macrophage polarisation was stimulated by mTORC1 activation and in turn exacerbated cartilage degeneration and osteophyte formation in experimental OA, while M2 polarisation was enhanced by mTORC1 inhibition and attenuated OA development. Furthermore, we found that M1 macrophages secreted Rspo2 to exacerbate experimental OA progression. Downregulating M1 macrophage polarisation or abolishing Rspo2 is thus a potential therapeutic target for OA treatment.

Recent studies have shown the essential role of synovial macrophages in the pathogenesis of rheumatoid arthritis (RA).32 33 Several studies have reported that depletion of macrophages attenuates RA,34 35 but the exact role of M1 and M2 macrophages in RA remains to be identified. M1 macrophage polarisation can be enhanced by activation of Notch, JNK and ERK1/2 signalling pathways, resulting in enhanced release of inflammatory factors such as TNF-α, IL-1β, IL-6 and IL-23 during RA development,36 37 while M2 macrophages exhibit an anti-inflammatory effect by releasing IL-10 and transforming growth factor-β1 during RA.38 However, little is known about the role of macrophages and their polarisation in OA. Macrophages are accumulated and polarised (M1 or M2) in the synovium and articular cavity during OA development, suggesting a correlation between macrophages and OA.25 A recent in vitro study by Utomo et al revealed that M1 but not M2 macrophages exert prominent effects on inflammation and degeneration of cultured cartilage explants.39 Furthermore, systemic depletion of macrophages reduces osteophyte formation in a collagen-induced arthritis model.40 However, in another recent study conditional macrophage depletion increased inflammation and did not inhibit the development of OA in obese mice.16 The role of macrophage polarisation and its regulatory mechanism in the pathogenesis and progression of OA remains unknown. Using mice with mTORC1 activation or inhibition in the myeloid lineage, this study establishes that macrophage M1 polarisation potentiates, while M2 polarisation attenuates, OA development in mice. We show that mTORC1 signalling and M1 macrophages are activated in the hyperplastic synovium of patients with OA and experimental OA mice. We also found that enhancement of M1 macrophages by mTORC1 activation in synovium efficiently increased the severity of pathology in CIOA, while M2 macrophages stimulated by mTORC1 inhibition protected mice against collagenase-induced cartilage degradation and osteophyte formation. Although TSC1KO mice also exhibited a significant acceleration of OA development in DMM model, we found that the effect of M1 macrophages was more prominent in an OA model with high synovial activation (CIOA), compared with low (DMM). These results suggest that the role of M1 macrophage in OA was related to synovitis. Nevertheless, our results showed that M1 and M2 macrophages always coexist in synovium and the imbalance of macrophage polarisation with a significant increase of M1 macrophages occurs during OA development. In these models, when M1 or M2 macrophages are stimulated, the proportion of opposite phenotypes is reduced. Enhancement of M2 macrophage numbers by mTORC1 inhibition attenuated OA along with a significant reduction in M1 macrophages. Moreover, mRNA sequence screening revealed that M1 macrophage-produced inflammatory cytokines and Rspo2 mediate its role in OA. We propose a model in which mTORC1-induced macrophage M1 polarisation stimulates production of inflammatory cytokines and Rspo2 secretion to promote expression of matrix-degrading enzymes, hypertrophic differentiation of articular chondrocytes and cartilage degeneration during OA development (figure 6D).

In OA, disease affects all joint tissues and therapy that is able to target the multiple pathological changes in joints is desirable. Accumulating evidence highlights the vital role of mTORC1 in the pathogenesis and progression of OA. Previous studies have established that activation of mTORC1 inhibits autophagy to promote articular chondrocyte apoptosis during OA development.41 42 Our previous study demonstrated the key role of the chondrocyte mTORC1 pathway in regulating cell proliferation and differentiation in OA.29 We also found that chondrocyte mTORC1 promoted OA partially through formation of vascular endothelial growth factor-A-stimulated subchondral H-type vessels.43 The current study demonstrates an important role of macrophage mTORC1 in OA synovitis. We have clearly demonstrated the important role of mTORC1 signalling in regulating synovial macrophage M1 polarisation during OA development. Our results showed that synovial macrophage mTORC1 activation occurred throughout OA development, combined with enhanced M1 macrophage polarisation. Hyperactive mTORC1 in synovial macrophages enforced cartilage degeneration in TSC1KO mice. In contrast, macrophage mTORC1 inactivation enhanced M2 polarisation and attenuated OA. Together, these studies suggest that multiple OA pathological joint compartments may benefit from targeting of mTORC1.

Although Rspo2 has been widely studied as an activator of Wnt/β-catenin signalling,44 its role in OA has not been reported. It has been shown that Rspo2 activates Wnt/β-catenin signalling to facilitate cultured chondrocyte hypertrophic differentiation during endochondral ossification.31 Another in vitro experiment in MC3T3-E1 cells (a preosteoblastic cell line) demonstrated that Rspo2 might be a positive regulator of bone metabolism.45 An interesting finding of this study was that high levels of Rspo2 were detected in synovial tissues and in the serum of TSC1KO mice. Importantly, abolishing Rspo2 with a neutralising antibody effectively reduced β-catenin in articular chondrocytes and attenuated the development of OA in mice. These data indicate that M1 macrophages secrete Rspo2 to induce terminal differentiation of articular chondrocytes, degradation of matrix proteoglycan and cartilage during OA development.

To conclude, our study found an association between M1 macrophages and Rspo2 on cartilage degeneration. Future studies are warranted to explore the underlying mechanisms and further management of the behaviour of polarised synovial macrophages seems a suitable approach to prevent OA.

References

View Abstract

Footnotes

  • Handling editor Josef S Smolen

  • Contributors HZ and XB conceived the ideas for experimental designs, analysed data and wrote the manuscript. CL and CZ conducted the majority of the experiments and helped with manuscript preparation. ZW, JL and HW performed microcomputed tomography analyses and provided suggestions for the project. XL and YS performed immunohistochemistry and immunofluorescence and confocal imaging. SX and YZ conducted cell cultures and western blot experiments. CZ and DX collected human tissue samples. XB and DC developed the concept, supervised the project, conceived the experiments and critically reviewed the manuscript. HZ, CL and CZ contributed equally to this work.

  • Funding This work was supported by grants from the National Natural Science Foundation of China (Grant No 81625015, 81530070, 81772406 and 31529002) and the State Key Development Program for Basic Research of China (2015CB553602).

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Human study was approved by the Ethics Committee of the Third Affiliated Hospital of Southern Medical University. All animal experiments were approved by the Southern Medical University Committee Animal Care and Use Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.