Article Text
Abstract
Background Interleukin (IL)-17A and Th17 cells are critically involved in T cell-mediated synovial inflammation. Besides IL-17A, Th17 cells produce IL-22. Recently, Th22 cells were discovered, which produce IL-22 in the absence of IL-17. However, it remains unclear whether IL-22 and Th22 cells contribute to T cell-mediated synovial inflammation. Therefore, we examined the potential of IL-22 and Th22 cells to induce synovial inflammation and whether IL-22 is required for T cell-mediated experimental arthritis.
Methods Peripheral and synovial Th17 and Th22 cells were identified and sorted from patients with rheumatoid arthritis (RA). Co-culture experiments of these primary T cell populations with RA synovial fibroblasts (RASF) were performed. The in vivo IL-22 contribution to synovial inflammation was investigated by inducing T cell-mediated arthritis in IL-22 deficient mice and wild-type mice.
Results Peripheral Th17 and Th22 cell populations were increased in patients with RA and present in RA synovial fluid. In T cell-RASF co-cultures, IL-22 in the presence of IL-17A had limited effects on IL-6, IL-8, matrix metalloproteinase-1 (MMP-1) and MMP-3 production. Furthermore, primary peripheral blood and synovial Th17 cells were more potent in the induction of these factors by RASF compared with Th22 cells. In line with this, similar synovial inflammation and disease severity was found between IL-22 deficient and wild-type mice in T cell-mediated experimental arthritis.
Conclusions These findings show that IL-17A/Th17 cell-mediated synovial inflammation is independent of IL-22 and Th22 cells. This implies that targeting IL-17A/Th17 cells, rather than IL-22/Th22 cells, should be the focus for treatment of T cell-mediated synovial inflammation.
- Cytokines
- Inflammation
- Synovitis
- T Cells
- Rheumatoid Arthritis
Statistics from Altmetric.com
Introduction
Rheumatoid arthritis (RA) is a chronic autoimmune disorder characterised by synovial inflammation and destruction of cartilage and bone.1 ,2 Proinflammatory cytokines, such as tumour necrosis factor α (TNF-α) and IL-17A, play a central role in RA.3 ,4 IL-17A producing T helper 17 (Th17) cells are implicated in the pathogenesis of RA and other T cell-mediated diseases.5–10 In patients with RA, elevated proportions of peripheral and synovial Th17 cells were found11–14 and the pathogenic role for IL-17A in arthritis has been shown in experimental mouse models.15 ,16 Furthermore, we have shown that Th17 cells from patients with RA, induce a proinflammatory loop upon interaction with RA synovial fibroblasts (RASF). This loop includes autocrine IL-17A production, which may explain the progression of an early inflammatory arthritis to a chronic destructive arthritis.14 ,17
Besides IL-17A, Th17 cells express the cytokine IL-22.18 ,19 Among target cells of IL-22 are mainly tissue epithelial cells. IL-22 is involved in antimicrobial defence and tissue regeneration.20–22 The IL-22 receptor is not expressed by immune cells and no effects of IL-22 have been identified in these cells. In addition to T cells, T cell receptor (TCR)γδ, lymphoid tissue inducer and natural killer (NK) cells also produce IL-22.23 ,24 Recently, the novel human Th22 cell subset was identified producing high levels of IL-22 in the absence of IL-17A production. Th22 cells are implicated in skin homeostasis and pathology25 and IL-22 production in T cells is dependent on the aryl hydrocarbon receptor and the transcription factor receptor tyrosine kinase-like orphan receptor C (ROR-C).26 ,27 Besides cytokine production, Th17 and Th22 cells can be characterised by a specific chemokine receptor expression pattern. Whereas Th17 and Th22 cells are negative for CXCR3 expression and positive for CCR6 and CCR4 expression, only Th22 cells are positive for the expression of CCR10.25 ,27 ,28
In patients with RA increased IL-22 expression levels or elevated proportions of IL-22 producing T cells were found in peripheral blood and in the inflamed synovium.11 ,29–31 The pathological role of IL-22 in RA has been suggested by an association of IL-22 serum levels with bone erosions. Also elevated Th22 and Th17 cell proportions were found in patients with RA compared with healthy controls.32–34 In addition, RASF have been shown to produce IL-22 and respond to IL-22 with increased proliferation and chemokine production.35 Moreover, IL-22 deficient (IL-22−/−) mice develop less severe collagen induced arthritis (CIA) compared with wild-type mice, indicating an involvement of IL-22 in the pathogenesis of CIA.36
However, the involvement of IL-22 in autoimmune disease development appears to be context dependent. In skin inflammatory disorders, such as psoriasis, IL-22 acts as a key mediator of pathogenic psoriatic skin features.37 Moreover, IL-22 produced by innate and adaptive immune cells is required for imiquimod induced skin inflammation in mice.38 In contrast, IL-22 is not required for the development of experimental autoimmune encephalomyelitis39and IL-22 is even protective for experimental inflammatory bowel disease.40 These discrepancies in the role of IL-22 in autoimmune disease development and the various IL-22 target cells and sources of IL-22 production prompted us to investigate the direct effect of IL-22 and Th22 cells in T cell-mediated synovial inflammation. Using human in vitro T cell-RASF co-cultures and murine T cell-mediated experimental arthritis, we show that IL-22 and Th22 cells do not contribute to IL-17/Th17 mediated synovial inflammation.
Methods
Subjects
For this report 10 healthy volunteers (eight women and two men, mean age±SD; 47.4±24.5), 8 treatment naive patients with early RA (six women and two men, mean age±SD; 49.7±13.7) and eight patients with established RA and active disease were studied. All patients fulfilled the American College of Rheumatology 1987 revised criteria for RA. Blood was obtained at the second visit after obtaining informed consent. Clinical and laboratory data of treatment naive patients with early RA are shown in online supplementary table S1. This study was embedded in the Rotterdam Early Arthritis Cohort Study and was approved by the Medical Ethics Committee of the Erasmus MC Rotterdam.
Animals
IL-22−/− mice on the C57BL/6 background21 were kindly provided by Dr Wenjun Ouyang, Genentech Inc., USA. Wild-type C57BL/6 mice were purchased from Harlan Laboratories B.V. (Horst, The Netherlands). Mice were kept under specific pathogen free conditions and provided with food and water ad libitum. Mice between 8–12 weeks of age were used for experiments. All experiments were approved by the Erasmus MC Animal Ethics Committee (DEC).
Flow cytometry and cell sorting
Monoclonal antibody preparations, intracellular cytokine detection, flow cytometry and cell sorting were described previously.14 ,41 The following human monoclonal antibodies (MoAb) were obtained from BD Biosciences (San Diego, California, USA): CD45RO, CCR6, CD4 and IFN-γ. IL-22 and IL-17A MoAb were obtained from eBioscience (San Diego, California, USA). Murine CD4, CD8, IL-17A and IFN-γ MoAb were obtained from BD Biosciences and TCRγδ MoAb from Biolegend Inc. (San Diego, California, USA). Samples were acquired on a FACScantoII flow cytometer (BD Biosciences) and analysed using FlowJo v7.6 research software (Tree Star Inc. Ashland, Oregon, USA). T cell populations were sorted from peripheral blood mononuclear cells (PBMC) or synovial fluid mononuclear cells (SFMC) using a FACSAria cell sorter (BD Biosciences).
Cell cultures
RASF isolation and subsequent culture has been described.14 10.0×103 RASF were co-cultured with sorted allogeneic 25.0×103 peripheral blood total CCR6+ T, Th17 or Th22 cells or 3.0–10.0×103 synovial fluid Th17 or Th22 cells. Cells were cultured for 72 h in Iscove's Modified Dulbecco's Media (IMDM, Lonza, Verviers, Belgium), supplemented with 10% fetal calf serum (Invitrogen, Carlsbad, California), 100 U/ml penicillin/streptomycin, 2 mM L-glutamin (Lonza) and 50 μM β-mercapto-ethanol (Merck, Darmstadt, Germany) and stimulated with soluble αCD3 and αCD28 (0.3 μg/ml and 0.4 μg/ml respectively, Sanquin, Amsterdam, The Netherlands). Cells were cultured in the absence or presence of 10 μg/ml neutralising IL-22 Moab (R&D systems, Minneapolis, Minnesota, USA) or 2 μM 6-formylindolo(3,2-b)carbazole (FICZ) (Enzo Life Sciences Inc., Farmingdale, New York, USA)
Cytokine measurements
Human IL-6, IL-8 and IFN-γ production was determined using ELISA (Invitrogen). Human IL-17A, IL-22, TNF-α, matrix metalloproteinase (MMP)-1, MMP-3 and murine IL-17 expression was measured using Duoset ELISA (R&D systems, Minneapolis, MN). ELISA was performed according to the manufacturer's instructions.
T cell-mediated arthritis
Methylated bovine serum albumin (mBSA, Sigma-Aldrich, St. Louis, Missouri, USA) 8 mg/ml was emulsified in an equal volume of complete Freund's adjuvant, containing 1 mg/ml heat-killed Mycobacterium tuberculosis (strain H37Ra; Difco Laboratories Inc., Detroit, Michigan, USA). At day −7, mice were immunised by intradermal injection of 100 μl mBSA/complete Freund's adjuvant emulsion into the tail base. On day 0, arthritis was induced by injecting mice intra-articularly (IA) with 60 μg mBSA in 6 μl 0.9% NaCl into both knee joints. The arthritis severity was scored macroscopically 7 days after IA injection on a scale of 0–2. Rear limbs were removed and prepared for histology. Sections were H&E stained as previously described.41 The analysis of murine IL-17A cytokine expression in synovial washouts was performed as described previously.41 To determine mBSA specific effector cell responses, draining lymph node cells were cultured in the presence or absence of 50 ug/ml mBSA for 3 days. IL-17A production was measured by ELISA. To measure DNA synthesis, cells were pulsed at day 2 with thymidine (3H) for 16–20 h, harvested and counted using standard methods.
Statistical analysis
Differences between experimental groups were tested with a two-sided paired t test or stated otherwise, using Prism software V.5.04 (GraphPad Software Inc. La Jolla, California, USA). p Values<0.05 were considered statistically significant.
Results
Th17 and Th22 cells are increased in peripheral blood of RA patients and present in the inflamed synovium
Recently, we have shown that the proportion of IL-17A and IL-22 producing CD4CD45RO+ (memory) CCR6+ T cells was increased in patients with early RA.14 ,29 Through a flow cytometry approach combining specific chemokine and cell surface receptor expression it is now possible to distinguish in more detail Th17 and Th22 cells, without the need to perform intracellular cytokine stainings.25 ,27 ,28 First, CD25−CCR6+ T cells were gated to exclude regulatory T cells. Subsequently, Th17 and Th22 cells were defined within this CCR6+ gate as CXCR3−CCR4+ whereby Th22 cells, but not Th17 cells were CCR10+ (figure 1A).
By following this gating strategy, significantly increased peripheral blood memory total CCR6+ T cell, Th17 cell and Th22 cell proportions were identified within the total memory T cell population of patients with early RA compared with age and sex matched healthy controls. The increase of Th17 and Th22 cell proportions (∼2.5-fold and ∼3.3-fold respectively) was even relatively larger than the total CCR6+ T cell population (∼1.6-fold) (figure 1B).
In addition, total memory CCR6+ T cell, Th17 cell and Th22 cell populations were present in matched PBMC and SFMC from patients with established RA and active disease. Moreover, total memory CCR6+ T cell, Th17 cell and Th22 cell proportions were similar in SFMC compared with PBMC (figure 2A). The presence of Th17 (IL-17A+IL-22±) and Th22 (IL-17A-IL-22+) cells in SFMC was also indicated by performing intracellular IL-17A, IL-22 and IFNγ stainings (figure 2B).
These data show that in addition to the total memory CCR6+ T cell population, Th17 and Th22 cells are increased in peripheral blood of patients with early RA and that Th17 and Th22 cells are present in synovial fluid.
IL-17A, but not IL-22 is upregulated in CCR6+ T cell-RASF cultures
The interaction of CCR6+ T cells and RASF induces a proinflammatory loop leading to increased autocrine IL-17A production.14 To verify whether this interaction leads to increased IL-22 production as well, CCR6+ T cell cultures with or without RASF were analysed for IL-17A, IL-22 and TNF-α production. Increased IL-17A producing CCR6+ T cell proportions and increased IL-17A protein levels were detected in CCR6+ T cell RASF co-cultures compared with CCR6+ T cell monocultures as shown before.14 In contrast, no difference was found in the fraction of IL-22 producing CCR6+ T cells as well as IL-22 and TNF-α protein levels (figure 3A,B). This clearly shows a specific induction of IL-17A, but not of IL-22 and TNF-α production by CCR6+ T cells upon RASF interaction.
IL-22 has limited effects on IL-6, IL-8, MMP-1 and MMP-3 production in CCR6+ T cell-RASF cultures
Upon interaction with CCR6+ T cells, RASF produce proinflammatory mediators such as IL-6, IL-8 and tissue destructive enzymes, such as MMP-1 and MMP-3. This production is largely dependent on IL-17A and TNF-α.14 However, the contribution of IL-22 to IL-6, IL-8, MMP-1 and MMP-3 induction is unclear. Therefore, CCR6+ T cell-RASF co-cultures were performed wherein IL-22 signalling was neutralised by anti-IL-22 antibodies, or wherein IL-22 production was induced by an aryl hydrocarbon receptor agonist, FICZ.27 IL-22 neutralisation had no effects on IL-17A, IL-6, IL-8, MMP-1 and MMP-3 production in CCR6+ T cell-RASF co-cultures (figure 4). Treatment of CCR6+ T cell cultures with FICZ resulted in a ∼4.7-fold induction of IL-22 expression. This had a slight, but not significant inhibitory effect on IL-17A production in CCR6+ T cell-RASF co-cultures. This effect on IL-17A was accompanied by a significant reduction of IL-6, but not of IL-8, MMP-1 and MMP-3 production. These findings show that IL-22 has limited effects on proinflammatory cytokine and MMP production in CCR6+ T cell-RASF co-cultures in the presence of IL-17A.
Th22 cells are less potent inducers of IL-6, IL-8 and MMP-1 production by RASF compared with Th17 cells
In addition, the effects of primary Th17 or Th22 cells from patients with early RA on RASF were investigated. Therefore, primary Th17 and Th22 cells were sorted according to the gating strategy as shown in figure 1A and cultured in the presence of RASF. To verify the phenotype of the sorted Th17 and Th22 cells, intracellular flow cytometric staining for IL-17A and IL-22 was performed. Th17 and Th22 cells produced IL-17A and IL-22, but IL-17A production was higher in Th17 cells (4.4% vs 1.3% in Th22) and IL-22 production was higher in Th22 cells (9.3% vs 2.0% in Th17) (figure 5A). These differences were reflected by protein expression levels in the culture supernatant. Th17 cells expressed higher levels of IL-17A compared with Th22 cells and IL-22 was expressed by Th17 and Th22 cells. Furthermore, slightly higher levels of IFN-γ and TNF-α were expressed by Th17 cells compared with Th22 cells (figure 5B).
Importantly, compared with peripheral blood Th17 cells, Th22 cells were less efficient in the induction of IL-6 and MMP-1 production by RASF (figure 5B). To verify whether this phenomenon is also true for synovial Th22 cells, synovial Th17 and Th22 cells were sorted and co-cultured with RASF. After co-culture, IL-17A and IFN-γ expression was restricted to Th17-RASF co-cultures, whereas Th17-RASF and Th22-RASF co-cultures expressed IL-22 and TNF-α. Compared with synovial Th17 cells, synovial Th22 cells are less potent in inducing IL-6 and MMP-3 production by RASF (figure 5C).
When taken together, compared with peripheral and synovial Th17 cells, Th22 cells are less efficient in inducing proinflammatory cytokine and MMP production by RASF.
IL-22 does not contribute to IL-17A/Th17 mediated synovial inflammation in antigen induced arthritis
To investigate whether IL-22 directly contributes to T cell-mediated synovial inflammation in vivo, the murine mBSA antigen induced arthritis (AIA) model was used. This model is largely T cell-mediated and dependent on IL-17A. Furthermore, T cells in this model can express IL-22, showing a possible involvement of IL-22.41 ,42 AIA was induced in IL-22 deficient (IL-22−/−) mice and wild-type mice. Knee swelling in time was measured and arthritis severity was macroscopically scored. No difference in knee swelling, macroscopic inflammation score, histology and IL-17A production between arthritic wild-type and IL-22−/− mice was detected, indicating that in the presence of IL-17A, IL-22 is not directly involved in T cell-mediated synovial inflammation in AIA (figure 6A–D). This was further supported by similar CD4, CD8 and TCRγδ T cell proportions, and similar proportions of IL-17A and IFN-γ producing CD4T cells in arthritic IL-22−/− and wild-type draining lymph nodes (figure 6E). In addition, similar mBSA specific proliferation and cytokine production responses were observed between arthritic IL-22−/− and wild-type draining lymph node cells (figure 6E,F).
Taken together, these findings show that lack of IL-22 has no in vivo contribution to IL-17A/Th17 mediated synovial inflammation in AIA.
Discussion
By using cultures with primary human T cells of treatment naive patients with early RA and T cell-mediated arthritis in IL-22−/− mice, we showed that IL-17A/Th17 mediated synovial inflammation is independent of IL-22. IL-22 produced by CCR6+ T cells had limited effects on RASF produced proinflammatory cytokines IL-6, IL-8 and tissue destructive enzymes MMP-1 and MMP-3. In addition, primary human peripheral or synovial Th22 cells were markedly less potent than Th17 cells in inducing IL-6, MMP-1 and MMP-3 production by RASF. Moreover, deficiency of IL-22 in vivo has no significant effect on IL-17A/Th17 mediated synovial inflammation in AIA.
Despite the limited observed effects of IL-22 and Th22 cells on IL-17A/Th17 mediated synovial inflammation, increased IL-22 levels or Th22 cell proportions are found in patients with early RA and are present in synovial fluid. Interestingly, Th22 cells correlate with elevated proportions of Th17 cells in patients with RA.32 This and the shared developmental programme and dependence on similar transcriptional regulators and cytokines, such as ROR-C and IL-23,19 may imply that IL-22 expression is commonly accompanied with IL-17A mediated function. This may be highly relevant in local mucosal immune responses against micro-organisms,19 ,23 ,24 whereas this is less relevant in non-mucosal environments, such as the joint synovium. We found increased IL-17A, but not IL-22 production by CCR6+ T cells in the presence of RASF. In patients with RA, increased levels of IL-22 or Th22 cells have been shown, suggesting additional signalling events or cellular interactions to induce IL-22 expression. In line with this, we found that addition of IL-23 or lipopolysaccharides activated monocytes were able to induce the induction of IL-22 in our system (data not shown).
On the other hand, IL-22R is expressed by RASF, and within synovial tissue IL-22 is produced by RASF,35 indicating a function for IL-22 in the inflamed synovium. Treatment of RASF with IL-22 has been shown to induce proliferation and expression of chemokines, such as CCL2.35 However, in our T cell-RASF co-cultures we were not able to identify an effect of IL-22 on the expression of CCL2 (data not shown). This lack of effect together with no inducing effects on the proinflammatory mediators IL-6 and IL-8 produced by RASF indicates that in comparison with IL-17A, which is a potent inducer of IL-6 and IL-8,14 IL-22 has limited effects in T cell-mediated synovial inflammation.
Moreover, the effect of IL-22 is likely dependent on the stage of synovial inflammation. In IL-1Ra−/− mice, IL-17A is already expressed in the early stages of inflammation, whereas IL-22 is mainly expressed in highly inflamed synovia.43 This may explain the mild effect on arthritis severity in IL-1Ra−/− mice after anti-IL-22 treatment compared with anti-IL-17A treatment.43 The expression of IL-22 in later stages of arthritis and our finding that IL-22 deficiency has no effect in T cell-mediated arthritis, shows that IL-22 has a secondary effect on synovial inflammation rather than a direct role in the induction of synovial inflammation.
It might be that IL-22 has a role in the induction of bone erosions in later stages of RA. This would be in line with the finding that IL-22 expression levels correlate with bone erosions and the induction of osteoclastogenesis by RASF induced receptor activator of nuclear factor-κB ligand (RANKL) production.12 ,34 ,44
The observation that the increase of IL-22 levels and Th22 cells in patients with RA correlates with elevated Th17 cells and progression of bone erosions,12 would argue for the use of IL-22 or Th22 cells as a biomarker in RA. For this purpose a combination of chemokine receptors and cytokines will be preferable to distinguish pure Th17 and Th22 cell populations (figure 1)
From the findings that synovial inflammation was not affected in IL-22−/− mice it can be concluded that local IL-22 produced by adaptive or innate immune cells have no direct contribution to the induction of T cell-mediated synovial inflammation. However, this is in disagreement with findings obtained by CIA in IL-22−/− mice, in which arthritis incidence and severity was lower compared with wild-type mice.36 On the other hand, our data are in line with other experimental autoimmune disease models, such as experimental autoimmune encephalomyelitis (EAE), in which IL-22 is not required for the induction of encephalomyelitis.39 A possible explanation for these differences is the underlying pathological mechanism in these experimental autoimmune models. Whereas AIA and EAE are largely dependent on cellular immune responses, CIA is dependent on humoral and cellular responses, suggesting a more important role of IL-22 in the humoral response and less in the cellular response.36 ,41 ,45 After CIA induction, high total IgG and collagen specific IgG levels were found in IL-22−/− mice,36 suggesting a role of IL-22 in IgG production by B cells. In this context, a current study performed in our laboratory (Corneth et al, unpublished), confirmed the data as published by Geboes et al,36 that IL-22 deficiency in the CIA model results in decreased arthritis incidence and severity. Importantly, we observed a critical role for IL-22 in germinal centre formation and terminal B cell differentiation. As IL-22 is expressed in lymphoid organs, the lower CIA severity in IL-22 deficient mice may be caused by intrinsic defects in germinal centre formation or maintenance and altered kinetics of collagen specific antibody production in IL-22−/− mice.
IL-22 serum levels in patients with RA correlated with serum titres of antibodies against citrinulated peptides.44 It would be of interest to investigate whether these antibodies are of high affinity and whether the kinetics of antibody production is altered in IL-22−/− mice.
Studying the role of IL-22 is complicated, because (1) IL-22 is expressed by multiple cells of the immune system, such as T cells, NK cells and lymphoid tissue inducer cells, and (2) IL-22 is expressed in different tissues, including the skin, gut and lymph nodes and (3) IL-22 has multiple context dependent functions, such as wound healing and microbial defence.1–2 The contrasting results of studies regarding IL-22, which show proinflammatory, protective or no effects of IL-22 in different diseases or disease models, may be inherent to these context dependent effects of IL-22. In psoriasis for example IL-22 can synergise with other proinflammatory cytokines to induce many of the pathogenic phenotypes from keratinocytes and exacerbate disease progression in psoriasis. In contrast, IL-22 plays a beneficial role in inflammatory bowel disease by enhancing barrier integrity and epithelial innate immunity of the intestinal tract.21 ,40 ,46 ,47 Moreover, IL-22 has protective roles in airway inflammation, and protects against liver pathology during malaria infection.48–50 When taken together, IL-22 does not contribute to the induction of IL-17A/Th17 mediated synovial inflammation. This implies that treatment of patients with early RA should focus on targeting IL-17A and Th17 cells, rather than on targeting IL-22.
Acknowledgments
We would like to thank H Vroman, M Brem, B Bartol, H Bouallouch-Charif and N Kops and the people from the EDC Animal Facility (Erasmus MC Rotterdam) for technical assistance at various stages of the project. We also thank Dr. Wenjun Ouyang for kindly providing the IL-22 deficient mice.
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 table
Footnotes
-
Handling editor Tore K Kvien
-
Contributors All authors concur with the resubmission. All authors have contributed to the work described in this manuscript by designing experiments, collecting and analysing data, writing and editing the revised manuscript.
-
Funding This project is partly supported by the Dutch Arthritis Association DAA 0801043.
-
Competing interests None.
-
Patient consent Obtained.
-
Ethics approval This study was embedded in the Rotterdam Early Arthritis Cohort Study (REACH) and approved by the Medical Ethics Committee of the Erasmus MC Rotterdam.
-
Provenance and peer review Not commissioned; externally peer reviewed.