Objective To investigate the impact of systemic inhibition of interleukin 6 (IL-6) or signal transducer and activator of transcription (Stat3) in an experimental model of osteoarthritis (OA).
Methods Expression of major catabolic and anabolic factors of cartilage was determined in IL-6-treated mouse chondrocytes and cartilage explants. The anti-IL-6-receptor neutralising antibody MR16-1 was used in the destabilisation of the medial meniscus (DMM) mouse model of OA. Stat3 blockade was investigated by the small molecule Stattic ex vivo and in the DMM model.
Results In chondrocytes and cartilage explants, IL-6 treatment reduced proteoglycan content with increased production of matrix metalloproteinase (MMP-3 and MMP-13) and a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS-4 and ADAMTS-5). IL-6 induced Stat3 and extracellular signal-regulated kinase (ERK) 1/2 signalling but not p38, c-Jun N-terminal kinase or Akt. In the DMM model, Stat3 was activated in cartilage, but neither in the synovium nor in the subchondral bone. Systemic blockade of IL-6 by MR16-1 alleviated DMM-induced OA cartilage lesions, impaired the osteophyte formation and the extent of synovitis. In the same model, Stattic had similar beneficial effects on cartilage and osteophyte formation. Stattic, but not an ERK1/2 inhibitor, significantly counteracted the catabolic effects of IL-6 on cartilage explants and suppressed the IL-6-induced chondrocytes apoptosis.
Conclusion IL-6 induces chondrocyte catabolism mainly via Stat3 signalling, a pathway activated in cartilage from joint subjected to DMM. Systemic blockade of IL-6 or STAT-3 can alleviate DMM-induced OA in mice.
Statistics from Altmetric.com
Interleukin 6 (IL-6) blockade has been shown to be effective in various musculoskeletal disorders, in particular rheumatoid arthritis (RA).1 Although in vitro and animal studies have suggested that IL-6 is a major cytokine in the pathogenesis of osteoarthritis (OA), the effect of systemic IL-6 inhibition on cartilage degradation, a hallmark feature of OA, has not been investigated yet in OA animal models, nor in humans.
IL-6 is a pleiotropic proinflammatory cytokine involved in many physiological and pathological processes.2 It belongs to a family of cytokines known as gp130 cytokines. IL-6 binds first to its non-signalling specific receptor (IL-6R) and then to a common subunit (gp130), which triggers two main signalling pathways: signal transducer and activator of transcription (STAT) and extracellular signal-regulated kinase (ERK).3 Stat1, 3 and 5 are all activated on IL-6 stimulation, but Stat3 is the main signalling factor downstream of IL-6.4
IL-6 serum levels are strongly associated with obesity and metabolic syndrome, two major risk factors for knee OA.5 Data from a prospective cohort showed that higher serum levels of IL-6 and body mass index were independent predictors of incident radiographic knee OA.6 Within the joint, high levels of IL-6 are found in knee OA synovial fluid and surrounding tissues at the early stages of the disease7 ,8 and are associated with progression of knee OA after meniscectomy.9 Many joint tissues can produce IL-6: fibroblast-like synoviocytes, chondrocytes and subchondral bone osteoblasts release high levels of this cytokine after different stimulations, such as mechanical loading.10–12 The infrapatellar fat pad from patients with knee OA can stimulate the secretion of IL-6 by synoviocytes.13 In bovine chondrocytes, IL-6 increases matrix metalloproteinase 1 [Mmp1], 3 and 13, and a disintegrin and metalloproteinase with thrombospondin motifs 4 [Adamts4] and 5 mRNA levels.14 Intra-articular injections of IL-6 reproduced OA-like cartilage lesions in mice, and IL-6 knockout (KO) mice develop less severe posttraumatic OA lesions.15 The role played by IL-6 in age-related OA is controversial. In humans, a low innate production of IL-6 is associated with the absence of OA in the elderly.16 By contrast, it has been shown that male IL-6 KO mice develop more severe OA on ageing, yet a finding not observed in female mice.17
More recently, a positive feedback loop was identified between calcium-containing crystals and IL-6 leading to cartilage destruction.18 Despite these data, the mechanisms by which IL-6 elicits its effects in cartilage remain poorly known.19
In this study, we investigated the impact of IL-6 on cartilage metabolism and assessed subsequently the efficacy of IL-6/Stat3 pathway blockade in an experimental model of OA. We found that IL-6 induced Stat3 and ERK1/2 pathways in chondrocytes and upregulated the production of the main proteases involved in OA pathogenesis (MMP-3 and MMP-13, and ADAMTS-4 and ADAMTS-5), thereby leading to cartilage degradation. With a mouse model of OA induced by destabilisation of the medial meniscus (DMM), systemic inhibition of IL-6 with a neutralising monoclonal antibody or Stat3 with a non-peptidic small molecule decreased the severity of OA lesions. Our findings provide strong evidence for a pivotal role for IL-6 in OA via Stat3 signalling and open up new therapeutic perspectives.
Materials and methods
Culture of primary articular chondrocytes and cartilage explants
Primary articular chondrocytes were isolated from femoral heads, femoral condyles and tibial plateaus of newborn mice20 and expanded in Dulbecco's modified Eagle's medium (DMEM; Gibco Life Technologies) with 10% decomplemented fetal calf serum (FCS), 2% L-glutamine and 1% penicillin/streptomycin until confluence. Confluent monolayers of chrondrocytes were then stimulated with 10, 50 or 100 ng/mL IL-6 (R&D Systems) and 50 ng/mL IL-6R (R&D Systems), with or without 1, 2 or 5 µmol/L Stattic (sc-202818, Santa Cruz Biotechnology), a non-peptidic small molecule that selectively inhibits Stat3,21 or 10, 50 or 100 µmol/L PD98059 (Gibco), a mitogen-activated protein kinase kinase inhibitor,22 for 24 hours. At the end of culture, RNA and protein were extracted and medium was collected. All experiments were performed in duplicate.
Cartilage explants were obtained from the femoral heads of 10-week-old mice as described,23 cultured in FCS-free DMEM for 24 hours, then stimulated with 500 ng/mL IL-6 and 500 ng/mL IL-6R, with or without 5 µmol/L Stattic or 100 µmol/L PD98059, for 72 hours (n=3 explants per condition). The medium was collected and explants were prepared for cryosection. For additional details, see online supplementary methods.
OA was induced in 10-week-old male C57BL/6 mice by DMM of the right knee as described.24 Sham operation was performed on the controlateral knee by incision of the cutaneous and muscular planes at baseline. Sham surgery was also performed on the right knee from a separate group of mice.25 In a first experiment, mice received intraperitoneal injections of a rat anti-mouse IL-6R monoclonal antibody (MR16-1, 0.5 mg once a week, kindly provided by Chugai Pharmaceutical Co.) (n=10) or phosphate-buffered saline (PBS) as a control (n=10) for 6 weeks from the day after the DMM. In a second experiment, mice received Stattic (25 mg/kg every other day) by oral gavage (n=10) or 1% Tween 80 PBS as a control (n=8) for 6 weeks from the day after the DMM. The local animal ethics committee and the French Ministry of Higher Education and Research approved all animal protocols (approval no. apafis#665-2015051311034087-v1). For additional details, see online supplementary methods.
A Mann-Whitney test was used to determine statistical significance. p<0.05 was considered statistically significant. Data are expressed as mean±SEM from at least three experiments unless otherwise indicated. Statistical analyses involved use of StatView (V.5.0; SAS Institute, Cary, North Carolina, USA).
IL-6 increases cartilage catabolism by activating proteolytic enzymes
First, we examined the overall effects of IL-6 on proteoglycan content from in vitro chondrocytes cultures. Alcian blue staining showed that the glycosaminoglycan content in the extracellular matrix (ECM) layer decreased upon IL-6 stimulation (−36.6±12.8%; p<0.05). In addition, IL-6 induced a significant increase of glycosaminoglycan release to the culture medium (+17.0±2.3%, p<0.05) (figure 1A). We next confirmed these results in an organotypic model. In cartilage explants, IL-6 triggered a loss of Safranin O staining in the superficial layer of cartilage (figure 1B, upper part), along with a significant release of proteoglycans to the culture medium (+65.6±38.6%, p<0.05) (figure 1B). Of note, the loss of Safranin O staining paralleled an increase in NITEGE staining (figure 1B, lower part), attesting to aggrecanases activation following IL-6 stimulation.
To clarify the mechanisms by which IL-6 led to proteoglycan breakdown and triggered aggrecanases activity in cartilage, we investigated its effect on the expression of ADAMTS as well as MMPs and major anabolic and phenotypic genes in primary mouse chondrocytes. After 24 hours, IL-6 dose-dependently increased the mRNA levels of Mmp3, Mmp13, Adamts4 and Adamts5 (figure 2A). ADAMTS-4 and ADAMTS-5 were secreted in their activated form on IL-6 stimulation (molecular weight of 90 kDa and 73 kDa, respectively). IL-6 also enhanced the secretion of MMP-3 and MMP-13 in the culture medium (figure 2B), the highest stimulation being observed with MMP-3. Furthermore, IL-6 slightly increased Timp1 but not Timp3 mRNA levels (figure 2A). Therefore, the proteolytic burden of IL-6 was not counteracted by metalloprotease inhibitors.
By contrast, IL-6 stimulation did not modulate the mRNA levels of the anabolic genes Col2a1 and Acan, nor the cartilage hypertrophy marker gene Vegf (figure 2C). Furthermore, IL-6 had no effect on either the Ptgs2 (COX2-encoding gene) mRNA levels, nor the secretion of prostaglandin2 (18.9±3.4 vs 20.1±3.6 pg/mL; p=NS) or the NO production (1.24±0.2 vs 1.14±0.1 µmol/L; p=NS) in the culture medium. Thus, IL-6 had no effect on the production of inflammatory mediators (figure 2C–E). Therefore, the IL-6-induced cartilage degradation is likely mediated by the production and activation of several ECM-degrading enzymes.
Systemic IL-6 blockade alleviates osteoarthritic cartilage lesions
We next investigated the effect of the systemic inhibition of IL-6 by a rat anti-mouse IL-6R neutralising monoclonal antibody (MR16-1) in the DMM mouse model of OA.23 MR16-1 was administered by intraperitoneal injection once a week for 6 weeks beginning the day after DMM surgery. OA lesions were less severe with 6-week MR16-1 than control treatment (n=10 per group), as assessed by the Osteoarthritis Research Society International (OARSI) score (3.95±0.44 vs 5.80±0.51; p=0.007; figure 3A, B). Results were similar when a sham-operated right knee from a separate group (n=4) was used as control (see online supplementary figure S1). The operated knees from MR16-1-treated mice had smaller osteophytes (osteophyte size score: 1.30±0.12 vs 1.65±0.09, p=0.02; figure 3C and see online supplementary figures S2A and S2B) and milder synovial inflammation as compared with controls (synovitis score: 2.03±0.09 vs 2.70±0.28, p=0.02; figure 3D, E). By contrast, MR16-1 had no effect on the subchondral bone remodelling as assessed by the bone volume/tissue volume (BV/TV) (figure 3F).
Stat3 signalling mediates the catabolic effects of IL-6 on cartilage
We then investigated the signalling pathway involved in the catabolic effects of IL-6. Chondrocytes were challenged or not with IL-6 for 15 min and lysates underwent western blot analysis. IL-6 induced the phosphorylation of Stat3 and ERK1/2 but had no effect on the phosphorylation of p38, JNK or Akt (figure 4A).
Stat3 is the most specific actor of IL-6 signalling.4 To confirm that the increased production of proteases was mediated by Stat3, IL-6-treated chondrocytes were cultured with increasing concentrations (1–5 µM) of Stattic, a non-peptidic small molecule that specifically inhibits Stat3.21 Stattic dose-dependently inhibited Stat3 phosphorylation (figure 4B). We compared the effects of Stattic with those of PD98059, an ERK1/2 inhibitor (figure 4C). In cartilage explants, Stattic counteracted both the IL-6-induced loss in proteoglycans and Stat3 phosphorylation in the superficial layer of cartilage explants (figure 4D). Importantly, Stattic decreased immunostaining for MMP-3, MMP-13, ADAMTS-4, ADAMTS-5 and NITEGE in IL-6-stimulated explants (figure 4D). Moreover, Stat3 blockade mitigated the release of proteoglycans in the supernatant (−23.6±7.3%, p<0.05) (figure 4E). Stattic also suppressed IL-6-induced chondrocyte apoptosis as assessed by terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay (see online supplementary figure S3). By contrast, the inhibition of ERK1/2 signalling had little effect on Safranin O staining, as well as on IL-6-induced proteases secretion and proteoglycan release to the culture medium (figure 4D, E). Thus, Stat3 is most likely the main signalling pathway that mediates the deleterious effects of IL-6 on cartilage.
Stat3 blockade is chondroprotective
To investigate the role of Stat3 in OA, we administered Stattic (25 mg/kg every other day, n=10) or PBS (n=8) to mice by oral gavage for 6 weeks, beginning the day after DMM surgery.
As compared with controls, the severity of OA cartilage lesions was significantly reduced in Stattic-treated mice (OARSI score: 2.57±0.29 vs 4.50±0.32 for control treatment, p=0.0001) (figure 5A, B). Same results were observed when comparing with a sham-operated right knee (n=4) from a separate control group (see online supplementary figure S4). Stattic also decreased the size of osteophytes as compared with PBS (1.20±0.05 vs 1.88±0.17, p=0.0006; figure 5C and see online supplementary figures S2C and S2D). However, Stattic had no significant effect on the extent of synovitis nor on the subchondral bone remodelling (figure 5D–F).
The proportion of phosphorylated Stat3 (phospho-Stat3)-positive chondrocytes increased 6 weeks after DMM (34.5±4.4 vs 16.0±2.3%; p=0.001; figure 5G, H), demonstrating the activation of Stat3 pathway in OA. Remarkably, administration of Stattic significantly inhibited the DMM-phosphorylated Stat3 in chondrocytes as compared with control treatment (14.3±2.5% vs 34.5±4.4%; p=0.001). In addition, Stattic decreased DMM-induced NITEGE staining in cartilage (figure 5I), so reduced aggrecanases activity paralleled the inhibition of phosphorylated Stat3.
In line with histological findings suggesting that the effects of Stattic were limited to cartilage and osteophyte in the DMM model, Stat3 activation occurred mostly in articular cartilage after DMM, but not in the synovium nor in the subchondral bone cells (see online supplementary figure S5). These results confirm that the chondroprotective effect of Stattic, administered orally in the DMM model, is mediated via inhibition of Stat3 activation in cartilage.
OA is considered a low-grade inflammatory disease involving the whole joint rather than a simple wear and tear of the cartilage.26 Indeed, beyond the cartilage breakdown and the subchondral bone remodelling, synovial inflammation is an important hallmark of OA.12 IL-6 levels are almost 100-fold higher in OA synovial fluid7 than in serum from obese patients and/or patients with OA,6 ,27 which suggests a local production of IL-6 in knee OA. Here, we confirmed in two different models that IL-6 activates chondrocytes to synthesise MMPs and aggrecanases, thereby degrading cartilage ECM. By using a specific small molecule inhibitor, Stattic, we demonstrated that the catabolic effects of IL-6 on cartilage are mostly mediated by the transcription factor Stat3. Thus, the increased expression of matrix-degrading enzymes we observed with IL-6 likely occurred at the transcriptional level. The involvement of Stat3 in IL-6-induced cartilage catabolism seems to be specific since little effect was observed after blockade of ERK1/2 signalling.
IL-6 was previously reported to induce Mmp1, Mmp3 and Mmp13 and Adamts-4 and Adamts-5 mRNA levels in bovine chondrocytes.14 Our NITEGE immunohistochemistry experiments of femoral head cartilage explants extended these results and showed that these IL-6-induced proteases are active and able to cleave aggrecan. Albeit statistically significant, the Adamts4 and Adamts5 mRNA levels’ induction was modest in our model, suggesting that IL-6 might also regulate aggrecanases activity through posttranslational processing by proprotein convertases, such as furin or furin-like enzymes.28
IL-6 KO has been previously shown to reduce DMM-induced OA in mice,15 and in humans, a low innate capacity of production of IL-6 is associated with the absence of OA in the elderly,16 a finding not observed in mice.17 Here, we show for the first time that the systemic use of anti-IL-6R monoclonal antibody has protective effects in an experimental OA model, thus providing the relevance for IL-6 therapeutic targeting in OA. We hypothesise that both synovitis and angiogenesis, common findings in the DMM model,25 ,29 allowed for the MR16-1 and Stattic to reach the joint cavity and to exert their effects on cartilage. This finding supports ongoing clinical investigations of the efficacy of tocilizumab, a humanised anti-IL-6R monoclonal neutralising antibody, for severe subsets of patients with OA (ClinicalTrials.gov NCT02477059), as has been done with tumour necrosis factor alpha (TNFα) blockers.30
In our study, the protective effect of Stat3 blockade on cartilage destruction in the DMM model appeared greater than that observed with the neutralising anti-IL-6R antibody. A similar difference was previously observed in the effects of IL-6 blockade upstream or downstream of its interaction with its gp130 receptor,18 so Stat3 may be activated independent of IL-6 by other gp130 cytokines. Notably, several studies have pointed to the role of oncostatin M (OSM) in synovial inflammation or osteophyte-like bone proliferation in mouse models.31 ,32 Moreover, a proinflammatory gene signature induced by IL-6 in murine chondrocytes was recently found to be closer to that of OSM than any other gp130 cytokine.33 The Stat3 activation in OA cartilage may result from a synergistic action between several gp130 cytokines, especially IL-6 and OSM.
In the DMM model, Stat3 signalling was strongly activated in cartilage, but not in the synovium nor in the subchondral bone, a finding which may explain that we did not observe a significant effect of Stattic on synovitis or bone remodelling. Yet, we cannot exclude that our time point for histological assessment (6 weeks) was not the most appropriate to detect changes in bone remodelling. Indeed, we previously showed that the BV/TV decreased early after DMM, at week 4 but not at weeks 6 and 9.23 To the difference of Stattic, MR16-1 significantly decreased the extent of synovitis, which occurred early in the DMM model.29 ,34 This suggests that MR-16 might act through other pathways than Stat3 in the synovium tissue. Indeed, in RA synoviocytes, IL-6 can enhance p65 nuclear factor κB (NF-κB) activation and both Toll-like receptor ligand-induced inflammatory cytokine and chemokine production,35 which are known to be also involved in the genesis of OA synovitis.36
DMM-induced joint instability may result in experimental OA by the increased and sustained mechanical stress on cartilage and bone. The main transcription factor known to regulate biomechanical signals in chondrocytes is NF-κB,37 which is known to have an important role in OA.38 IL-6 is a direct target gene of NF-κB in human chondrocytes,39 and a key mediator of several chondrocyte procatabolic factors involved in OA.15 ,40–44 Our results suggest that the IL-6/Stat3 pathway could be another important mediator of catabolic and inflammatory effects of mechanical stress in OA-related cartilage damage, acting downstream of numerous procatabolic cytokines or transcriptional factors in chondrocytes. Furthermore, IL-6 might act synergistically with other proinflammatory mediators to promote OA cartilage destruction, as there is evidence for a crosstalk between IL-6 family cytokines and IL-1 or TNFα in various joint tissues.45 ,46
In conclusion, we showed that Stat3 is activated in cartilage from joint subjected to DMM and appears as the main signalling pathway involved in IL-6-induced cartilage breakdown. Our results open new therapeutic perspectives with the evidence that pharmacological inhibition of the IL-6/Stat3 pathway could reduce OA progression. Indeed, because the inhibition of IL-6 or Stat3 protected against experimental murine OA, these factors may represent relevant therapeutic targets for OA. As it has been previously shown with epidermal growth factor receptor signalling in a rat model of OA,47 we show here that systemic blockade of different actors of a pleiotropic signalling pathway ameliorates joint degradation and we therefore provide a proof of concept. These findings need to be further investigated with other inhibiting strategies (other pharmacological inhibitors, silencing RNA or gene KO) in different models of OA before considering their application in human OA. Regardless, they bring important insights into OA pathogenesis by demonstrating the pivotal role of the IL-6/Stat3 pathway in OA.
The authors thank Caroline Marty for technical assistance with histology and immunohistochemistry experiments. They also thank Dr Tadamitsu Kishimoto, Chugai Pharmaceutical Co. and Chugai Pharma France for kindly providing MR16-1 antibody.
Handling editor Tore K Kvien
Contributors AL, WB, TF-B, MC-S, EH and PR contributed to the conception of the work. AL, CC, JM, H-KE, EH and PR contributed to the acquisition and interpretation of data. All authors and contributors contributed to drafting the work and critical revision of its content. All authors and contributors gave their final approval for the submitted version of the manuscript and agreed to be accountable for all aspects of this work.
Funding This work was supported by ART Viggo and the French Society of Rheumatology.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
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.