Objective: Interleukin (IL)23, composed of a p19 and a p40 subunit, is suggested to play key roles in rheumatoid arthritis (RA), dependent on the promotion and proliferation of IL17-producing T helper (Th)17 cells. However, previous studies on IL23 expression in human tissues were based on the p19 subunit only. We aimed to study the expression and regulation of IL23 subunits p19 and p40 in RA compared to patients with osteoarthritis (OA).
Methods: The expression of p19 and p40 in synovial tissues was analysed by in situ hybridisation and immunohistochemistry. IL23 in RA and OA synovial fluids and sera was determined by ELISA. Toll-like receptor (TLR)-dependent induction of p19, p40 and bioactive IL23 was determined in RA synovial fibroblasts (RASF), monocytes and monocyte-derived dendritic cells (MDDCs) by real-time PCR and reverse transcriptase (RT)-PCR, Western blot and functional assays.
Results: The p19 subunit was abundantly expressed in RA but not in OA synovial tissues. p19 was most prominently expressed by RASF in the synovial lining layer and at the site of invasion, but no heterodimeric IL23 was detected at these sites. Correspondingly, soluble IL23 was not detectable or found at very low levels in synovial fluids and sera of patients with RA. By in vitro experiments, we confirmed that TLR-activated RASF expressed p19 but not p40, in contrast to monocytes, which produced IL23 following TLR stimulation.
Conclusion: The TLR-dependent induction of p19 but not p40 in RASF and the abundant expression of p19 along with the low or undetectable levels of IL23 in patients with RA provides strong evidence that p19 does not necessarily indicate the presence of IL23, as has been proposed to date.
Statistics from Altmetric.com
Interleukin 23 (IL23) is a heterodimeric cytokine composed of a p19 and a p40 subunit, which is shared with IL12. The p19 subunit has been identified by computational screening of the IL6 helical cytokine family. Purified p19 has no biological activity in vitro, whereas p19 combined with the p40 subunit forms the heterodimeric and biologically active cytokine IL23.1 The major source of IL23 are macrophages and dendritic cells. Toll-like receptor (TLR) agonists and interactions with T cells (CD40/CD40L interaction) were shown to trigger the induction of bioactive IL23.2–4 IL23 is related to IL12 but has functionally distinct properties. IL12 is composed of the p40 and the p35 subunit, which has an overall sequence identity of approximately 40% to the p19 subunit. In contrast to IL12, IL23 does not promote the development of interferon (IFN)γ producing T helper (Th)1 CD4+ cells, but is one of the essential factors required for the maintenance and survival of a recently described IL17 producing CD4+ T cell subset, named Th17.5 Expression of IL17 has been detected in sera and tissues of patients with various autoimmune diseases, including rheumatoid arthritis (RA), multiple sclerosis and systemic lupus erythaematosus.6–8
RA is a chronic inflammatory disease that is characterised by the destruction of articular cartilage and bone. RA synovial fibroblasts (RASF) play a major role in the destruction of the joint by secreting matrix-degrading enzymes. Moreover, a hallmark of RA is synovial hyperplasia that is caused by the proliferation of resident RASF and by the accumulation of inflammatory immune cells including B cells, T cells and macrophages. IL23 is suggested to be essential in autoimmune inflammation in joints as p19 and p40 deficient mice exhibit reduced severity of collagen-induced arthritis.9 In humans, levels of p19 as well as p40 were shown to be elevated in synovial fluids and sera of patients with RA compared to osteoarthritis (OA).10 11 These data suggested high expression of IL23 in joints of patients with RA even though the presence of heterodimeric IL23 has not been ascertained so far.
There is mounting evidence for an activation of TLR signalling pathways in RA. We and others have shown in previous reports elevated TLR2, 3 and 4 levels in synovial tissues of patients with RA compared to OA. Furthermore, RASF stimulated with TLR2, 3 and 4 ligands produced high amounts of cytokines, chemokines and matrix metalloproteinases that are characteristically found in joints of patients with RA. Regarding the induction of IL23 in RASF, it has been demonstrated that the p19 subunit is upregulated by IL1β or IL17.10 12 However, whether RASF have the potential to produce heterodimeric IL23 after TLR activation remains to be determined.
The purpose of our study was to analyse the presence of heterodimeric IL23 in RA and OA synovial tissues by immunohistochemistry and to determine levels of soluble IL23 in RA and OA synovial fluids and sera by enzyme-linked immunoabsorbent assay. Furthermore, we analysed the induction of IL23 subunits, p19 and p40, in RASF, human primary blood monocytes and monocyte-derived dendritic cells (MDDCs) after stimulation with TLR2, 3 and 4 ligands.
MATERIALS AND METHODS
Patients and tissue preparation
Synovial tissue specimens were obtained during synovectomy or joint replacement surgery from patients with RA and OA, after informed consent had been obtained (Department of Orthopedic Surgery, Schulthess Clinic, Zurich, Switzerland). RASF and OA synovial fibroblasts (OASF) were isolated from synovial tissues, digested by collagenase, and used after passages 4 to 8 as described.13 To obtain tissue sections, synovial specimens were fixed in paraformaldehyde and embedded in paraffin. Sera and synovial fluids from patients with RA and OA were collected, centrifuged and stored at −80° until analysis. Before analysis, synovial fluid samples were pretreated for 1 h at 37°C with 1 mg/ml of hyaluronidase (Fluka, Buchs, Switzerland). All patients with RA fulfilled the American College of Rheumatology (ACR) criteria for the classification of RA.14
In situ hybridisation
IL23 sense and IL23 antisense probes for in situ hybridisation (ISH) were prepared according to methods previously described. The primer sequences were as follows: upper primer 5′-CTATCAGGGAGCAGAGAAG-3′; and lower primer, 5′-ACTAGTGGGACACATGGAT-′3.15 ISH was performed as described by Kriegsmann et al.16
After the detection of IL23 mRNA in RA synovial tissues by ISH, subsets of synovial cells expressing p19 were analysed by double labelling the p19 mRNA stained sections with monoclonal mouse anti-human CD68 or anti-human vimentin antibodies (2 μg/ml, Dako, Glostrup, Denmark), respectively. Bound mouse primary antibodies were detected using horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibodies (Jackson ImmunoResearch Europe, Soham, UK). HRP labelled cells were visualised using 3-amino-9-ethylcarbazol (AEC) substrate chromogen (Dako).
In order to analyse IL23 protein expression, consecutive synovial tissue sections derived from patients with RA or OA were stained for p19 or/and p40. The catalysed signal amplification (CSA) system for mouse primary antibodies was used according to the manufacturers instruction (Dako). Briefly, tissue sections were deparaffinised and pretreated with target retrieval solution. Endogenous avidin-binding activity was blocked with the avidin/biotin blocking system and endogenous peroxidase was blocked with 1.5% H2O2. After blocking unspecific IgG binding with protein block solution, sections were incubated for 15 min with mouse anti-human p19 antibodies (10 μg/ml, BioLegend, San Diego, California, USA) or mouse anti-human p40 antibodies (10 μg/ml, AbD Serotec, Germany). Sections were then incubated with biotinylated link antibodies, streptavidin–biotin complex solution, amplification reagent and streptavidin peroxidase. HRP-labelled cells were visualised using 3,3′-diaminobenzidine tetrahydrochloride (DAB) or Histo-green substrate chromogen. In control experiments, isotype matched mouse IgG was used instead of the primary antibodies. Tissues were counterstained with haematoxylin.
Stimulation assays with RASF
RASF were cultured in Dulbecco modified Eagle medium (DMEM) (Gibco, Basel, Switzerland) supplemented with 10% foetal calf serum (FCS) and stimulated with the following agents: polyinosinic/polycytidylic acid (poly(I-C), 20 μg/ml; Invivogen, San Diego, California, USA), lipopolysaccharide (LPS) from Escherichia coli (100 ng/ml; List Biologicals, Campbell, California, USA), palmitoyl-3-cysteine-serine-lysine-4 (bLP, 300 ng/ml, Invivogen).
Isolation and stimulation of monocytes and generation of monocyte-derived dendritic cells
Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats of healthy volunteers with Ficoll-Paque PLUS (Amersham Biosciences, Uppsala, Sweden) gradient centrifugation. Peripheral blood monocytes were positively separated from PBMCs with CD14 microbeads (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturers protocol. The purity of the CD14+ cell fraction, as assessed by flow cytometry, was consistently over 90%. CD14+ cells were transferred in 12-well plates, cultured in RPMI-1640 supplemented with 5% FCS and stimulated with bLP (300 ng/ml), poly(I-C) (10 μg/ml) or LPS (100 ng/ml) for 24 h.
For the generation of MDDCs, CD14+ cells were cultured for 6 days in RPMI-1640, supplemented with 10% FCS, IL4 (500 U/ml; R&D Systems, Abington, UK) and granulocyte monocyte-colony stimulating factor (GM-CSF, 800 U/ml; R&D Systems). Fresh complete culture medium was added after 3 days. At day 6 the immature MDDCs (iMDDCs) were harvested. To generate mature MDDCs, iMDDCs were transferred to new 12-well culture plates in fresh complete culture medium and stimulated for 24 h with bLP (300 ng/ml), poly(I-C) (10 μg/ml) or LPS (100 ng/ml). Expression of cell surface markers (CD14, CD86, HLA-DR and CD83) on iMDDCs and mMDDCs was measured by fluorescence activated sell sorter (FACS) analysis in order to confirm corresponding phenotypes.
Isolation and stimulation of CD8+ T cells
PBMCs were isolated from buffy coats of healthy volunteers with Ficoll-Paque PLUS (Amersham Biosciences) gradient centrifugation. T Cells were positively separated from PBMCs with CD8 microbeads (Miltenyi Biotec) according to the manufacturers protocol. CD8+ cells were transferred in 96-well plates, which were precoated with anti-CD3 antibodies (BD Pharmingen, Erembodegem, Belgium) for 24 h at 5 μg/ml. CD8+ T cells were cultured in RPMI-1640 supplemented with 5% FCS in presence of anti-CD28 antibodies (BD Pharmingen) at 5 μg/ml. CD8+ cells were stimulated either with recombinant human IL23 (R&D Systems) at 1 ng/ml, 10 ng/ml, 100 ng/ml or with supernatants of RASF and monocytes stimulated with TLR ligands bLP, poly(I-C) and LPS as described above.
Total RNA from cultured RASF, monocytes and MDDCs was isolated with the RNeasy MiniPrep Kit including treatment with RNase-free DNase (Qiagen, Basel, Switzerland) and reverse transcribed using random hexamers and multiscribe reverse transcriptase (both Applied Biosystems, Rotkreuz, Switzerland). Non-reverse transcribed samples were used as negative controls. Quantification of p19 and p40 mRNA was performed by TaqMan RT-PCR using the ABI Prism 7700 Sequence Detection system (Applied Biosystems). The following validated TaqMan gene expression assays were used: IL23A (p19) (Hs00372324_m1) and IL12B (p40) (Hs00233688_m1). The endogenous control 18S cDNA was used for correcting the results with the comparative threshold cycle (Ct) method for relative quantification as described by the manufacturer.
Conventional real-time PCR
Total RNA was isolated as described above for reverse transcriptase (RT) PCR and reverse transcribed using oligo(dT) and moloney murine leukaemia virus (MuLV) reverse transcriptase (both Invitrogen, Basel, Switzerland). Conventional PCR was performed on a GenAmp PCR System 9700 (Applied Biosystems) with the following primer pairs and protocols. p19: forward primer 5′-CTATCAGGGAGCAGAGAAG-3′, reverse primer 5′-ACTAGTGGGACACATGGAT-3′; p40 forward primer 5′-ATGTCGTAGAATTGGATTGG-3′, reverse primer 5′-AGGTGAAACGTCCAGAATAA-3′; β microglobulin forward primer 5′-AAGATTCAGGTTTACTCACGTC-3′, reverse primer 5′-TGATGCTGCTTACATGTCTCG-3′; 5 min 94°C; 30 cycles 30 s at 94°C, 30 s at Tm (Tm = 54°C for p19, 54°C for p40, 56°C for β microglobulin), 30 s at 72°C; and a final elongation of 5 min with 72°C. Reaction products were separated on a 1% agarose gel and signals were visualised using ethidium bromide. As a negative control, PCR was carried out in the absence of cDNA for each set of primers.
ELISA and enzyme immunoassay (EIA)
IL23 heterodimer was detected using a human IL23 ELISA Kit (eBioscience, San Diego, California, USA) according to the manufacturers instructions. IL17 was measured with a human IL17 DuoSet kit (R&D Systems) following the manufacturers instructions. Absorption was measured at 450 nm and data were analysed using Revelation V 4.22 software (Dynex Technologies, Denkendorf, Germany).
RASF were stimulated for 24 h with TLR ligands before cells were lysed in Laemmli buffer. Proteins were separated on 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel and blotted on Protran nitrocellulose transfer membrane (Schleicher & Schüll, Dassel, Germany). Membranes were probed with anti-p19 antibodies (0.5 μg/ml, BioLegend) and detected with HRP-conjugated second antibodies using the enhanced chemiluminescence (ECL) Western blotting detection system (Amersham). Blots were stripped and reprobed with monoclonal mouse anti human α-Tubulin (Sigma, Basel, Switzerland) antibodies to confirm similar loading of the gels.
Abundant IL23 p19 mRNA expression in RA synovial tissues
To investigate whether IL23 might play a role in the pathogenesis of RA we analysed the presence and distribution of the IL23 subunit p19 in the rheumatoid joint. Expression of p19 was assessed by in situ hybridisation with an anti-p19 mRNA probe using paraffin embedded RA synovial tissue sections (n = 5). We observed p19 mRNA positive cells throughout the synovium. The most marked expression of p19 mRNA was found in the synovial lining layer as well as at sites of invasion into cartilage (fig 1A,C). Control sections hybridised with sense- p19 mRNA probe showed no signal (fig 1B,D). Furthermore, double staining with cell type specific markers revealed that p19 positive cells expressed the fibroblast marker vimentin or the macrophage marker CD68 (fig 1E,F), documenting strong p19 mRNA expression by RA synoviocytes.
Differential p19 and p40 protein expression in RA and OA synovial tissues
IL23 is a heterodimer consisting of the p19 and the p40 subunit, the latter being shared with IL12. In order to examine whether p19 protein expression colocalises with p40 protein expression we performed immunohistochemical analysis for p19 and p40 on consecutive RA synovial tissue sections (n = 9). In RA synovial tissues p19 protein was abundantly expressed and the distribution corresponded to the presence of p19 mRNA as detected by in situ hybridisation (fig 2A). In contrast to the broad expression of the p19 subunit, p40 protein expression was found to be restricted to a few cell clusters scattered in the synovium (fig 2B). These results revealed that most p19 expressing synoviocytes do not coexpress the p40 subunit. Control sections incubated with matched isotype antibodies showed no signal (fig 2C,F). In addition we analysed the expression of p19 and p40 in synovial tissues derived from patients with OA. Subunits p19 and p40 were both not, or very weakly, expressed in OA synovial tissues (n = 8) (fig 2D,E). Thus, the subunits p19 and p40 might be implicated in the pathogenesis of RA but not OA.
To address the question whether heterodimeric IL23 is present in joints of patients with RA we performed immunohistochemical double stainings for p19 and p40 on RA synovial tissue sections (n = 6). The majority of synoviocytes in the synovial lining layer as well as in the sublining were single positive for p19 (fig 2G,H). In addition we also detected p40 single positive cells (fig 2H). However, within some p40 positive cell clusters we found p19/p40 double positive cells (fig 2H insert). Control sections incubated with matched isotype antibodies showed no signal (fig 2J). These results demonstrate that the majority of cells did not express p40 and therefore no heterodimeric IL23.
Activated dendritic cells are known to produce heterodimeric IL23. Therefore, to analyse whether p19/p40 double positive cells in the synovium might be dendritic cells we performed double stainings with the dendritic cell marker CD83 and p40, as p40 seemed to be the limiting factor for the potential production of IL23. p40 Positive cells coexpressed the dendritic cell marker CD83 (fig 3A), suggesting that dendritic cells in the synovium have the capability to coexpress both subunits of IL23
Levels of soluble IL23 in RA and OA synovial fluids and sera
The concentration of the shared IL12/IL23 subunit p40 was reported to be elevated in synovial fluids of patients with RA compared to controls11 17 and high levels of soluble p19 protein in RA synovial fluid and serum samples were demonstrated recently by Kim et al.10 By immunohistochemical analysis we detected only a small number of p19/p40 double positive cells in RA synovial tissues. To assess the concentrations of soluble IL23 in synovial fluids and sera from patients with RA and OA we performed ELISA specific for the IL23 heterodimer (fig 4). In only 6 out of 37 synovial fluids derived from patients with RA and in only 2 out of 21 synovial fluids of patients with OA we were able to detect concentrations of IL23 above the detection limit (range: 16–29 pg/ml IL23). Similarly, we also detected only small amounts of IL23 in 10 out of 37 RA sera and in 9 out of 29 OA sera (range: 16–114 pg/ml IL23). No statistically significant difference of IL23 levels between RA and OA synovial fluids and sera were observed.
Differential regulation of p19 and p40 expression by TLR ligands
As RA synoviocytes express elevated levels of TLR2, 3 and 4, we assessed TLR dependent regulation of the IL23 subunits p19 and p40 in different cell types. RASF, monocytes or MDDCs were stimulated with the TLR2 ligand bLP, the TLR3 ligand poly(I-C) and the TLR4 ligand LPS. At 24 h following stimulation we determined the presence and the induction of the two IL23 subunits by conventional RT-PCR as well as by quantitative real-time PCR. RASF did not express p19 mRNA constitutively (fig 5A). However after stimulation with TLR2, 3 and 4 ligands p19 mRNA was found to be induced in RASF (fig 5A,B). The most prominent induction was seen after TLR3 activation with poly(I-C) (24.5 (2.1)-fold upregulation relative to unstimulated cultures, n = 5). Conversely, p40 mRNA was neither expressed in unstimulated nor in TLR ligand stimulated RASF. In two RASF cultures derived from individual patients the induction of p19 protein by TLR ligands was confirmed on protein levels by Western blot analysis (fig 5C). As a positive control we used monocytes and MDDCs as it has been demonstrated that these cells have the capability to produce both subunits of IL23. In primary monocytes both IL23 subunits were inducible to similar levels by bLP as well as by LPS (fig 5A,B). However no constitutive mRNA expression was detected for p19 nor p40 (fig 5A). Because monocytes do not express TLR3, neither p19 nor p40 mRNA was detected after stimulation with poly(I-C). MDDCs constitutively expressed the p40 subunit (fig 5A). In addition p40 mRNA was strongly upregulated following stimulation with bLP and LPS as well as with poly(I-C) (fig 5A,B). The p19 subunit was not expressed in unstimulated MDDCs but was induced after stimulation with the TLR2, 3 and 4 ligands. Thus, TLR activated monocytes and MDDCs have the potential to induce both subunits necessary for the production of bioactive IL23, whereas TLR activated RASF express the p19 subunit only.
Activated monocytes but not RASF secrete bioactive IL23
It has been demonstrated that stimulation of CD8+ T cells with rhIL23 results in the production of IL17.18 To assess whether monocytes or RASF secrete biologically active IL23, we stimulated polyclonally activated CD8+ T cells with supernatants of TLR ligand stimulated monocytes and RASF and measured IL17 production. Stimulation of activated CD8+ T cells with rhIL23 served as a positive control. IL17 production by activated CD8+ T cells was dose dependently upregulated by rhIL23 (fig 6A). Similarly, the addition of supernatants of TLR2 activated monocytes to CD8+ T cell cultures resulted in an 1.4 (0.2)-fold upregulation of IL17 and supernatants of TLR4 activated monocytes in a 3.6 (0.2)-fold upregulation of IL17 production by activated CD8+ T cells (fig 6B). By contrast, treatment of CD8+ T cells with supernatants of TLR2, 3 or 4 activated RASF had no effect on the IL17 production (fig 6C). These results confirmed that RASF stimulated with TLR ligands, although expressing the p19 subunit, do not produce bioactive IL23 in contrast to TLR activated monocytes.
In the current study we demonstrate that p19 is expressed and induced independently of p40 in RASF in contrast to MDDCs and monocytes. In addition we show that heterodimeric IL23 is present at unexpectedly low levels in joints of patients with RA.
Recent reports have indicated the involvement of IL23 in several autoimmune diseases including colitis, psoriasis and arthritis. Treatment with anti-p19 antibodies in mice inhibited central nervous system autoimmune inflammation, prevented active colitis and exacerbated lyme arthritis.19–21 In the collagen-induced arthritis mouse model, p19 and p40 are essential for the development of joint inflammation.9 However, no studies regarding the role of IL23 in RA have been published to date. We show that IL23 is barely detectable in joints of patients with RA. Only a small number of p19/p40 double positive cells were detected in RA synovial tissues. In accordance with this finding we could detect heterodimeric IL23 in only 5 out of 28 patients. This result was unexpected as high p40 and high p19 levels were described in earlier reports.7 10 Interestingly, p40 seemed to be the limiting factor for the production of IL23 as most synoviocytes were single positive for p19. In addition we found p40 positive/p19 negative cells, presumably secreting the related cytokine IL12, which is composed of the p40 and p35 subunits. In accordance IL12 positive cells have been described in the sublining layer of the RA synovium.22
IL23 was shown to be expressed by activated dendritic cells, monocytes, keratinocytes and microglia.18 23–25 The activity of inflammatory cells can be influenced by the activation of TLRs that interact with microbial or endogenous ligands. In previous reports we have demonstrated that TLR2, 3 and 4 expression is elevated in RA synovial tissues.26–29 Endogenous TLR ligands, such as heat shock proteins and necrotic cells, are present in the chronically inflamed joints and may lead to a sustained activation of TLR signalling pathways in RA. In the current study we show differential expression and induction of the IL23 subunits by TLR2, 3 or 4 ligands in RASF, monocytes and MDDCs. We found that TLR activation induced p19 but not p40 protein expression in RASF whereas in MDDCs and monocytes both IL23 subunits were upregulated. Stimulation of preactivated CD8+ T cells with rhIL23 as well as with supernatants of TLR activated monocytes but not TLR activated RASF led to an increase in the production of IL17. Therefore, we assume that activated dendritic cells and monocytes or macrophages are the main producers of heterodimeric IL23 in the RA synovium whereas RASF express p19 independently of p40. This is the first study showing differential expression of IL23 subunits in the synovium. Consequently p19 expression does not necessarily correlate with the presence of bioactive IL23 as has been assumed in previous reports.10 30
Ubiquitous transgenic expression of p19 in mice results in a phenotype of systemic inflammation, impaired growth and premature death.31 Additionally, high levels of TNFα and IL1β were detected in the circulation of p19 transgenic mice. It has been suggested that the development of multiorgan inflammation in p19 transgenic mice is dependent on the dimerisation of p19 with p40 and is therefore dependent on the heterodimer IL23. However, the p40 subunit was not detectable in serum of p19 transgenic mice. These findings, together with our data showing a large number of p19 positive/p40 negative synoviocytes in the RA synovial lining layer and at sites of invasion, gives further evidence that p19 may be implicated in the RA pathogenesis independently of its dimerisation with p40. The p19 protein shares homology with members of the IL6/IL12 family, which includes IL6, oncostatin M, granulocyte-colony stimulating factor (G-CSF) and p35. It has been speculated that additional heterodimeric complexes might exist in the IL12 family, which comprises IL12, IL23 as well as IL27. For example combinations of Epstein–Barr virus-induced gene 3 (EBI3) and p35 as well as EBI3 and p19 have been described.32 Whether p19 might have additional binding partners other than p40 in vivo needs further investigation.
In conclusion we report abundant expression of p19 but not heterodimeric IL23 by RASF at sites of invasion and in vitro upon stimulation by TLR ligands. The differential expression of p19 and p40 suggests that p19 does not necessarily indicate the presence of active IL23 and gives evidence for a p40 independent involvement of p19 in the pathogenesis of RA
We thank Maria Comazzi and Ferenc Pataky for excellent technical assistance and Beat Simmen for providing synovial tissues.
Competing interests: None declared.
Funding: This work was supported by the Swiss National Fund Grant 3200B0-105923 (to DK).
Ethics approval: Ethics approval and informed consent was obtained.
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.