Objective: A pathogenic role for granulocyte-macrophage colony stimulating factor (GM-CSF) and interleukin (IL)17 in rheumatoid arthritis (RA) has been suggested. In previously published work, the therapeutic potentials of GM-CSF and IL17 blockade in arthritis have been described. In the present study, the simultaneous blockade of both pathways in a mouse model for chronic arthritis was investigated to identify whether this double blockade provides a superior therapeutic efficacy.
Methods: A chronic relapsing arthritis was induced in C57Bl/6 wild type (WT) and C57Bl/6 genetically deficient for IL17 receptor (IL17R knockout (KO)) mice by intra-articular injection of Streptococcal cell wall (SCW) fragments into knees on days 0, 7, 14 and 21. Treatments (intraperitoneal) were given weekly starting on day 14. Animals were analysed for inflammation, joint damage and a range of inflammatory mediators.
Results: Joint swelling and cartilage damage were significantly reduced in the IL17R KO mice and in WT mice receiving anti-GM-CSF neutralising mAb 22E9 compared to isotype control antibodies. The therapeutic effect was significantly more pronounced in mice where IL17 and GM-CSF pathways were inhibited (eg, IL17R KO mice treated with 22E9 mAb). Tumour necrosis factor (TNF)α blockade had essentially no effect.
Conclusion: Our data further support the therapeutic potentials of GM-CSF and IL17 blockade in a RA model that is no longer responsive to an established TNFα antagonist, moreover, our results suggest that concomitant inhibition of both pathways may provide the basis for a highly effective treatment of chronic RA in patients that are resistant to treatment by TNFα inhibitors.
Statistics from Altmetric.com
Granulocyte-macrophage colony stimulating factor (GM-CSF), initially identified as a haematopoietic growth factor, has more recently been shown to be an important cytokine in inflammation and autoimmunity. Elevated levels of GM-CSF mRNA or protein have been measured at a variety of inflammatory sites in patients who are allergic and psoriatic, and patients who are arthritic and asthmatic.1–6
Numerous in vivo studies have shown over the past few years that blockade of GM-CSF via neutralising antibodies can prevent or even cure proinflammatory diseases in various models of inflammation including models for arthritis7 8 experimental autoimmune encephalitis,9 psoriasis,10 and lung disease.11–13 GM-CSF plays an important role in innate immunity by stimulating the proliferation and activation of mature neutrophils and macrophages.14–16 In addition, a key role for GM-CSF has been demonstrated in antigen-presentation by governing differentiation and maturation of dendritic cells.17–19 In vivo, GM-CSF has been reported to preferentially induce type 1 proinflammatory cytokines by human peripheral blood mononuclear cells (PBMCs), T cells and antigen-presenting cells (APCs).20 21
Interleukin 17 (IL17) is a family of cytokines of the acquired immune system, consisting of six members, IL17A to IL17F. The receptor for IL17A, which is further on referred to as IL17R, is abundantly expressed by all cells of the immune system, and stimulation of various cell types with IL17 can induce the expression of other cytokines such as IL1β, tumour necrosis factor (TNF)α and IL6, and the chemokines IL8 and macrophage inflammatory protein 1α (MIP-1α). In contrast to its receptor, IL17 is mainly produced by the recently discovered Th17 cell,22–24 and its expression has been frequently related to infection and autoimmunity.25–30
Rheumatoid arthritis (RA) results from complex interactions between many cell types and factors belonging to the innate and acquired arms of the immune system.31 32 Previously published reports have hypothesised a pathogenic role for GM-CSF in RA. In support for this hypothesis are the findings (i) that GM-CSF is produced in synovium of patients with RA3 5 and that elevated levels of this cytokine can be measured in their synovial fluid;1 (ii) that treatment with a neutralising anti-GM-CSF monoclonal antibody (mAb) decreases disease severity in the collagen-induced mouse model for arthritis (CIA);7 (iii) that GM-CSF-deficient mice have a reduced susceptibility to disease induction by collagen and methylated bovine serum albumin (mBSA);33 34 (iv) that injection of recombinant GM-CSF to CIA mice exacerbates disease;35 and (v) that patients with RA receiving GM-CSF after chemotherapy show flares of arthritis severity.36
Similarly, IL17 appears implicated in RA pathology because IL17 levels are elevated in RA synovium and synovial fluid,37–39 and IL17 blockade reduces joint inflammation and destruction during arthritis in experimental models.40 41 In addition, mice genetically deficient of IL17 show suppressed collagen-induced arthritis,42 and when crossed to IL1Ra−/− mice, IL17−/− mice completely lack the spontaneous onset of polyarthritis usually seen in Balb/c mice deficient for the IL1 receptor antagonist.43
One of the models used for investigating human RA-like disease in mice is the Streptococcal cell wall (SCW) arthritis model. In this model, an acute disease and a chronic relapsing arthritis can be induced by intra-articular injection of bacterial cell wall fragments into one knee joint of mice. An acute arthritis, in which the innate immunity plays the major pathogenic role, is obtained by a single injection of SCW fragments into knee joints of naive mice.44 45 By repeated intra-articular injection of SCW fragments, a chronic relapsing model is established where mediators of acquired immunity gradually take over the initial dominance of the innate response. At that point, T cells and the T cell cytokine IL17 gradually become more important46 47 and the chronic arthritis becomes fully independent of TNFα.
In the present study, we explored the therapeutic efficacy of GM-CSF neutralisation in the TNFα-independent chronic SCW arthritis model. In addition, we studied the effect of blocking innate and adaptive immunity via the GM-CSF and IL17 pathways by neutralising GM-CSF in IL17R-deficient mice.
MATERIALS AND METHODS
Male C57Bl/6 mice were obtained from Charles River (Sulzfeld, Germany). IL17R deficient mice were kindly provided by J Peschon, Amgen, Seattle, Washington, USA, and were backcrossed to C57Bl/6 background for >10 generations. The mice were housed in filter top cages, and water and food were provided ad libitum. The mice were used at an age of 10 to 12 weeks. All animal procedures were approved by the institutional ethics committee.
SCW preparation and induction of SCW arthritis
Streptococcus pyogenes T12 organisms were cultured overnight in Todd–Hewitt broth. Cell walls were prepared as previously described.44 The resulting 10 000 g supernatant was used throughout the experiments. The preparation contained 11% muramic acid. Unilateral arthritis was induced by intra-articular injection of 25 μg SCW (rhamnose content) in 6 μl phosphate-buffered saline (PBS) into the right knee joint of naive mice, as described.48 To create a chronic SCW-induced arthritis, intra-articular injections into the right knee joint were performed at days 0, 7, 14 and 21. These repeated injections result in a chronic arthritis. As a control, PBS was injected into the left knee joint.
Reagents and treatment protocol
During chronic SCW arthritis, mice were randomly assigned to one of the different antibody treatments. GM-CSF was neutralised using rat mAb 22E9 (MM500CS, Perbio Science, Bonn, Germany). Etanercept (Enbrel; Wyeth Pharma, Münster, Germany) was used for TNFα blockade. Several studies have reported the effectiveness of this human soluble TNF receptor Fc fusion protein in different mouse models, including CIA.49–53 Rat IgG2a isotype control (BLD-400516-bulk, Biozol Diagnostica, Eching, Germany) and Humira (Abbott, Wiesbaden-Delkenheim, Germany) were used as isotype controls. IL1β was neutralised with the rat anti-mouse IL1β mAb 1400.24.17 (MM425, Perbio Science, Bonn, Germany). All treatments were administered intraperitoneally as 300 μg doses were given 4 times: (i) 2 h prior to the third reactivation (day 14), (ii) on day 17, (iii) 2 h prior to the fourth reactivation (day 21) and (iv) on day 24 after initial disease induction.
Measurement of joint swelling
Joint swelling during SCW arthritis was quantified by 99mTc uptake in the arthritic knee joint. This method was shown earlier to correlate well with histological findings.54 This validated method measures by external γ radiation counting the accumulation of radioisotope at the site of inflammation due to local increased blood flow and tissue swelling. Briefly, 6–8 mice per treatment group were sedated with chloralhydrate and subsequently injected intraperitoneally with 20 μCi 99mTc. After 30 min, γ radiation was assessed by use of a collimated NaI scintillation crystal with the knee in a fixed position. The severity of swelling is expressed as the ratio of the 99mTc uptake in the right (inflamed) over the left (control) knee joint. All values exceeding 1.10 were considered as joint swelling.
Cytokine and chemokine measurements
Levels of several cytokines and chemokines, including IL1β, IL6, TNFα, regulated upon activation, normal T cell expressed and secreted (RANTES), keratinocyte chemoattractant (KC) and MIP-1α, were determined in patellae washouts. Patellae with surrounding synovial tissue were isolated from inflamed knee joints for five mice per group and cultured in RPMI 1640 medium containing 0.1% BSA (200 μl/patella) for 1 h at room temperature, as previously described.55 Thereafter, supernatants were harvested and centrifuged for 5 min at 1000 g. Cytokine and chemokine levels were determined using the Luminex multianalyte technology. We used the Bio-Plex system from BioRad (Munich, Germany) in combination with multiplex cytokine and chemokine kits. Cytokines and chemokines were measured in 50 μl of patellae washout medium. The sensitivity of the multiplex kits was <10 pg/ml.
Mice were killed by cervical dislocation on day 28. Whole knee joints were removed and fixed in 4% formaldehyde for 7 days before decalcification in 5% formic acid and processing for paraffin embedding. Tissue sections (7 μm) were stained with haematoxylin and eosin (H&E) or safranin O/fast green (SO). Histopathological changes in the knee joints of 6–8 mice per group were scored in the patella/femur region on 5 semiserial sections spaced 140 μm apart. Scoring was performed on coded slides by two separate observers, using the following parameters. In the H&E stained slides the amount of cells infiltrating the synovial lining was scored from 0–3, depending on the degree of influx of proinflammatory cells into the synovial tissue and synovial cavity (0 = no cells, 1 = mild cellularity, 2 = moderate cellularity, 3 = maximal cellularity). Cartilage damage was scored in the SO stained slides on a scale from 0–3, ranging from a score of 0 for fully stained, healthy cartilage to a score of 3 for completely eroded cartilage surfaces.
Differences between experimental groups were tested using the Mann–Whitney U test and using GraphPad Prism V. 4 software (GraphPad, San Diego, California, USA). Significance readouts were grouped as follows: * = 0.05>p>0.01, ** = 0.01>p>0.001 and *** = p<0.001.
Systemic GM-CSF neutralisation decreases joint swelling in the chronic SCW model
During the chronic phase of SCW arthritis in C57Bl/6 mice, the effect on joint swelling after treatment with biologicals neutralising GM-CSF (mAb 22E9), TNFα (etanercept) or IL1β (mAb 1400.24.17) was investigated on days 15, 16, 22, 23 and 28 by differential uptake of 99mTc into the knee joints. Results are expressed as the ratio of 99mTc uptake between the arthritic SCW-injected knee and the PBS-injected control knee.
Systemic administration of the GM-CSF-neutralising antibody potently and significantly decreased joint swelling on days 16, 22, 23 and 28 with p values of 0.018, 0.004, 0.004 and 0.002, respectively (fig 1A). As previously reported,8 ILβ neutralisation also decreased joint swelling, although a significant reduction in 99mTc uptake of knees vs control knees was only seen on days 22 (p = 0.011) and 23 (p = 0.001) (fig 1B). As expected, TNFα blockade with etanercept, which is able to neutralise human as well murine TNFα,49 51–53 56 had no effect on joint swelling in the chronic SCW model (fig 1C). In contrast, etanercept was previously shown to be active in the acute phase of this disease model.8 Neutralisation of GM-CSF during chronic SCW arthritis thus appeared to be more potent than neutralisation of IL1β, and its effect was sustained until day 28 (ie, 4 days after the last administration of the antibody). A second independent study confirmed the efficacy of GM-CSF neutralisation on decreasing joint swelling in the chronic SCW model (data not shown).
GM-CSF neutralisation reduces inflammatory cell influx to synovium and cartilage damage
Histopathological sections from joints of the different groups of mice were prepared after termination of the experiment on day 28. The extent of inflammatory cell influx into synovium and assessment of cartilage damage were independently scored by two investigators on blinded H&E-stained and SO-stained tissue sections.
All three treatments, GM-CSF neutralisation with mAb 22E9, IL1β neutralisation with mAb1400.24.17 and TNFα blockade with etanercept, were efficacious at significantly reducing the influx of inflammatory cells into the synovium (fig 2A). TNFα blockade, although significantly effective, appeared less potent than GM-CSF or IL1β neutralisation with p values vs controls of 0.042, 0.004, 0.001, respectively. Furthermore, despite a reduction of inflammatory cell influx in knee joints of etanercept-treated mice, cartilage integrity was not preserved (fig 2B). In contrast, GM-CSF neutralisation significantly protected from cartilage damage (p = 0.02; mAb 22E9 vs isotype control mAb) (fig 2B). As previously reported by our group,45 57 IL1β neutralisation was very potent at protecting cartilage from damage (p = 0.004, anti-IL1β vs control; fig 2B).
The impact of the various treatments on cartilage integrity are illustrated in fig 3, showing microphotographs of Safranin O/fast green staining of knee joints from one representative mouse for each of the three treatment groups. The robust cartilage staining and good tissue preservation observed in the mAb 22E9-treated mouse (fig 3A) highlights the effect of GM-CSF neutralisation on protecting cartilage integrity. In contrast, cartilage from the mouse receiving the isotype control antibody (fig 3D) shows destructive erosions and reduced staining intensity demonstrating loss of proteoglycan, one of the major components of articular cartilage. Similarly, loss of proteoglycan and increased cartilage damage is seen in the etanercept-treated mouse (fig 3C). This is consistent with earlier studies in the chronic SCW model of arthritis showing independence of TNFα.57 IL1β is known to have a prominent destructive effect on cartilage in experimental models of arthritis.58–60 Accordingly, neutralisation of IL1β by an antibody has a pronounced protective effect on cartilage in our present study (fig 3B).
GM-CSF neutralisation reduces production of IL1β and KC in knee joints
In order to better understand the protective effect of GM-CSF and its relationship to IL1β we investigated concentrations of various cytokines and chemokines in patella washouts. Only the arthritic (right) knees were analysed as levels in the non-affected control knees (left) have been repeatedly found to be below the limit of detection in previous experiments.61
GM-CSF neutralisation with mAb 22E9 resulted in a significant reduction of local IL1β in comparison to the levels detected in joints from mice receiving the isotype control antibody treatment (p = 0.042; 22E9-treated vs control) (fig 4). TNFα blockade with etanercept had no effect on the levels of IL1β in joints (fig 4) whereas, and as expected, in mice having received IL1β-neutralising mAb, levels of IL1β were close to base line. Levels of the chemokine KC (mouse GRO-α) were significantly reduced in the arthritic knee joints by all three treatments (p = 0.005 for 22E9 vs control; p<0.001 for etanercept vs control; p = 0.007 for anti-IL1β vs control). Local levels of IL6 and RANTES were not influenced by any of the treatments investigated (data not shown). Levels of IL2, TNFα and GM-CSF were below the detection limits of the assays (eg, <10 pg/ml).
GM-CSF neutralisation in the absence of IL17 signalling potentiates the protective effects on cartilage destruction
GM-CSF neutralisation decreased joint swelling and protected cartilage from damage with an efficacy similar to that observed with IL1β neutralisation. Subsequently, similar studies with anti-GM-CSF mAb in chronic SCW arthritis were performed in mice deficient for IL17R. As previously reported,46 IL17R-deficiency results in suppressed joint swelling and cartilage destruction during chronic SCW arthritis (fig 5A). Combined targeting of GM-CSF and IL17 signalling in this arthritis model resulted in a strong, enhanced suppression of joint swelling (fig 5A). Although anti-GM-CSF treatment as well as IL17R-deficiency resulted in reduced cell influx, combined targeting did not result in significantly less joint inflammation (fig 5B). Interestingly, however, proteoglycan depletion and cartilage damage (chrondocyte death and erosion) were markedly reduced in anti-GM-CSF treated IL17R deficient mice. (fig 5B–H). These results indicate that the protective effect on cartilage of anti-GM-CSF can be further enhanced by the additional targeting of the T cell cytokine IL17.
The present study demonstrates for the first time that during a chronic, TNFα-independent experimental arthritis, anti-GM-CSF therapy effectively reduces joint inflammation and cartilage damage, and that combined blockade of GM-CSF and IL17 signalling results in a highly efficient protection from joint destruction.
The chronic relapsing SCW mouse model of arthritis is characterised by a severe destruction of joints as is typical in later stages of chronic RA in humans. In contrast to what is observed in the CIA mouse model and the acute SCW model of arthritis, TNFα neutralisation is no longer effective in controlling chronic SCW arthritis in which IL1β appears to play the major pathogenic role.57 The TNFα independence and a key role for IL1β in cartilage destruction in chronic SCW arthritis have been confirmed in the present study.
GM-CSF blockade was studied for the first in this particular model and found to have a profound inhibitory effect on joint swelling and cartilage destruction in SCW-injected knees when doses of 300 μg antibody were administered intraperitoneally in the chronic phase of disease. This suggests that an anti-GM-CSF antibody in mice at a dose, which is equivalent to an antibody dose of approximately 1 mg/kg in humans (after allometric correction), is sufficient to correct GM-CSF levels in arthritic knee joints in this experimental model. Whether the anti-GM-CSF antibody in mice acted following penetration and local neutralisation of the cytokine in the affected joint, or, via creation of a peripheral cytokine sink, or both, requires further testing. The therapeutic efficacy of GM-CSF neutralisation in the chronic arthritis model was profound. Joint swelling was better controlled by anti-GM-CSF than by anti-IL1β treatment while TNFα blockade was ineffective. Aberrant TNFα production may still play some role in chronic SCW arthritis because its neutralisation had an effect on influx of inflammatory cells and KC chemokine levels. However, the role of TNFα in driving this chronic disease is diminished as opposed to the acute phase of the disease, and in contrast to other mouse models of arthritis. With respect to cartilage protection, anti-GM-CSF and anti-IL1β treatments were very effective. Interdependence between the actions of GM-CSF and IL1 has been reported previously in another model of arthritis.34 In this model of IL1-induced arthritis following mBSA injection, GM-CSF plays a preponderant pathogenic role. Absence of GM-CSF as in GM-CSF KO mice, or by GM-CSF neutralisation in WT animals, markedly reduced arthritis.34 During the chronic SCW arthritis, however, GM-CSF seems to act upstream of IL1β, since its neutralisation reduced IL1β levels in the arthritic joints. This reduced IL1 production by activated macrophages and other GM-CSF-stimulated immune cells might also explain why anti-GM-CSF treatment had a protective effect on cartilage in our model. In the acute SCW model, we also found that anti-GM-CSF antibody could reduce IL1β levels, while the TNFα blocker etanercept could not.8 In the CIA mouse model of RA, GM-CSF blockade reduced the levels of IL1β and TNFα in a very significant way.7
While GM-CSF expression is acutely induced in various immune cells by proinflammatory cytokines such as TNFα and IL1β through activation of transcription factor nuclear factor (NF)κB and others,16 62 63 the hierarchy of cytokines appears to flip in later stages of inflammation, with GM-CSF taking over control of TNFα and IL1β production, and perhaps of other cytokines and chemokines. Simultaneously to the inhibition of TNFα and IL1β in arthritic tissue, GM-CSF blockade also has the potential to reduce the activity and survival of GM-CSF-dependent immune cells, such as granulocytes, neutrophils, macrophages.34 It is conceivable that GM-CSF not only directly induces IL1β and TNFα expression, but also causes a coordinated antiapoptotic action and a continuous activation of multiple cells of the innate immune system, thereby indirectly enhancing IL1β and TNFα production. Such an effect on cell cycling and survival has been demonstrated in the mBSA arthritis model in which GM-CSF neutralisation in vivo resulted in markedly reduced overall cellularity as well as number of cycling cells in the arthritic joints.34
In addition to blocking GM-CSF during chronic SCW arthritis in WT animals, experiments were also performed in IL17R-deficient mice. IL17 is produced by Th17 cells, which can simultaneously produce TNFα and GM-CSF.64 In the presence of TNFα, IL17 triggers synoviocytes to produce GM-CSF,65 suggesting a role for IL17 upstream of GM-CSF. By contrast, GM-CSF-treated bone marrow cells stimulated with lipopolysaccharide (LPS) produce IL23,66 which is an important survival factor for IL17-producing Th17 cells. Combined blocking of IL17 and GM-CSF has not yet been studied in vitro or in vivo. The present study is first to show that simultaneous blockade of GM-CSF and IL17 pathways resulted in superior suppression of joint swelling and increased protection of cartilage destruction relative to blockade of single pathways. This strong effect on cartilage might be explained by a synergy between IL17 and (GM-CSF-induced) IL1β, since these two cytokines have previously shown synergy on cytokine production by synovium from patients with RA67 and on prostaglandin E2 (PGE2) and NO production in osteoarthritic cartilage.68 Further studies are required to more fully understand the role of GM-CSF and its relation to IL17 in chronic stages of RA. Nevertheless, the present and previous studies7–9 13 16 34 62 69 make a strong point that neutralisation of GM-CSF may have therapeutic potential in human patients with RA also in patients that are no longer, or have initially not been responsive to TNFα blockade. In addition, our study demonstrates that anti-GM-CSF in combination with anti-IL17 treatment, may have a profound therapeutic effect in RA as well as in other autoimmune and inflammatory disease settings.
The authors would like to thank Birgitte Oppers-Walgreen and Cora Arndtz for performing histology.
Competing interests: CPZ and PAB are full-time employees of Micromet AG.
Ethics approval: All animal procedures were approved by the institutional ethics committee.
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.