Objective: An association to variations in the dendritic cell immunoreceptor (DCIR) gene with rheumatoid arthritis (RA) was recently shown. However, protein expression of DCIR has so far not been assessed in a disease setting. In the present work, we aimed to determine the cellular and tissue distribution of this receptor in healthy controls and in patients with RA before and after local glucocorticoid administration.
Methods: DCIR mRNA expression was evaluated by quantitative PCR (n = 3) and protein expression by flow cytometry (n = 18), immunohistochemistry (n = 14) and double immunofluorescence (n = 5).
Results: DCIR protein was not detected in healthy synovia. By contrast, expression was abundant on cells from rheumatic joints in synovial fluid and in tissue. Following corticosteroid treatment this expression was downregulated. Interestingly, DCIR could be detected on natural killer (NK) cells and T cells, and CD4+ and CD8+, as well as on monocytes, B cells, dendritic cells and granulocytes. The frequency of DCIR+ T cells and the level of surface expression were increased in the rheumatic joint compared to blood. In synovial fluid the typical DCIR+ T cells were large activated cells, whereas blasted DCIR+ T cells were not detected in blood.
Conclusions: We demonstrate increased protein and mRNA expression of DCIR in RA, especially in the rheumatic joint. Expression was widespread and included a subpopulation of T cells. This suggests that the inflammatory synovial environment induces DCIR expression, and this may be related to synovial T cell function. Ligation of DCIR, or lack thereof, could contribute to the chronic inflammation characterising autoimmune diseases such as RA.
Statistics from Altmetric.com
Rheumatoid arthritis (RA) is a complex chronic inflammatory disease, where many arms of the immune system are active but where the pathogenesis is only partially understood. Apart from environmental factors, there is a substantial genetic impact on disease development from the human leukocyte antigen (HLA) and from other genomic regions that remain largely unknown.1–4 Identification of the genes underlying susceptibility to RA may provide insights into immune-mediated disease pathways, and this goal is therefore pursued in patient material and in experimental settings. We have employed such a comparative genetics approach with the aim to define risk genes in rodent models for subsequent evaluation in humans. In rat, we recently identified a cluster of type II lectin-like receptor genes that determines arthritis susceptibility,5 as well as many other clinical phenotypes, including levels of autoantibody and interleukin (IL)17 mRNA.6
The corresponding human gene cluster contains five genes, including the dendritic cell immunoreceptor (DCIR) gene, and we demonstrated association of the gene cluster and of DCIR with RA susceptibility.5 Hitherto, only DCIR has been targeted by homologous recombination in the mouse. It was observed that DCIR knockout (DCIR-KO) mice spontaneously develop autoimmune disease in old age, which also associated with elevated autoantibodies in serum; taken together, these facts suggest that DCIR is involved in the development of autoimmunity as a negative regulator.7 Future studies are warranted to determine if other genes than DCIR in the gene complex may also regulate autoimmunity. The immunoreceptor DCIR was originally identified as a dendritic cell-associated C-type lectin-like receptor, characterised by the likely functional features of pattern recognition via a carbohydrate recognition domain (CRD) and signalling through an immunoreceptor tyrosine-based inhibitory motif (ITIM).8 9 With no investigative leads from ligand characteristics, the DCIR receptor function remains enigmatic. To date, the receptor type has been implicated in cellular adhesion and migration, microbial pattern recognition, antigen uptake, T cell costimulation and signal transduction.10 11 That genetic data from three species point to the C-type lectin gene cluster and to DCIR as being important in autoimmunity and arthritis, provide incentive for investigation of the encoded receptor. Hitherto, protein expression of DCIR has not been demonstrated in any disease, including RA.
To put DCIR in perspective, it is crucial to determine (1) on which cell types it is expressed, (2) in which compartments and (3) under which circumstances ie, in health and disease. Previous reports describe expression on granulocytes and antigen-presenting cells.8 9 12 13 Here, we impartially assessed DCIR protein expression in synovial fluid and synovial tissue from rheumatic joints and peripheral blood from patients and healthy controls. We also determined the expression pattern of DCIR in synovial tissues before and after treatment with glucocorticoids to determine if disease amelioration is accompanied by DCIR downregulation.
For phenotypic cell analyses, paired peripheral blood and synovial fluid cell samples from 18 patients with RA were collected. Patient characteristics are summarised in table 1.
Additionally, three different immunostaining experiments were performed.
Firstly, synovia from five healthy controls (age range 21–43 years) and eight patients with RA (age range 30–74 years) were included in the study. Secondly, synovial biopsies before and 2 weeks after intrarticular corticosteroid injection of 40 mg of triamcinolone hexacetonide were obtained from six patients (five females and one male, median age 63 years, range 46–83) with RA. Of these patients, four received non-steroidal antirheumatic drugs (NSAIDs), three corticosteroids (maximum 10 mg prednisolone per day), three disease-modifying antirheumatic drugs (DMARDs), three methotrexate (MTX), one biological therapy (etanercept) and one azathioprine.14 All patients were recruited from the Rheumatology Clinic at Karolinska University Hospital in Stockholm. Lastly, synovias from five patients with RA (age range 46–83) were used for immunofluorescence staining. All these studies were performed after human ethics approval and informed consent was obtained from all contributing individuals.
Lymphocyte isolation and flow cytometric analysis
Most paired peripheral blood (PB) and synovial fluid (SF) lymphocytes were studied on unseparated cell samples, since previous studies on lectin receptors on lymphocytes have reported artefacts/upregulations following Ficoll isolation of mononuclear cells.15 Additionally, Ficoll separation could activate cells and thereby alter their expression of these markers. Antibodies used for flow cytometric analysis are displayed in table 2.
Two different DCIR antibodies were used, 111F8 and MAB1748 (for confirmation). Stained cells were immediately analysed on a FACSCalibur (BD Biosciences, Stockholm, Sweden), and evaluated with CellQuest, (V 3.3, BD Biosciences).
Peripheral blood was collected from two healthy individuals. Peripheral blood cells was stained with anti-CD3 proprotein convertase 5 (PC-5), anti-CD4 phycoerythrin (PE), unconjugated anti-DCIR, anti-CD14 PE and anti-CD19 fluorescein isothiocyanate (FITC) and Alexa Fluor 647-conjugated goat anti-mouse as secondary antibody. Sorting was performed on a MoFlo cell sorter (Dako, Glostrup, Denmark). To get pure T cell subsets we sorted CD3+ T cells into CD4+DCIR+, CD4+DCIR–, CD4–DCIR+ and CD4–DCIR– populations.
Sorted cells were pelleted and lysed for total RNA extraction (Qiagen total RNA extraction kit, Germany). Samples were incubated with DNase (Qiagen RNase-free DNase set) to avoid amplification/detection of genomic DNA. Reverse transcription was performed on total RNA. Amplification was performed using an ABI PRISM 7700 Sequence Detection System (Perkin Elmer, Norwalk, Connecticut, USA) with the Assay-on-Demand Gene Expression product Hs01087625_m1 (DCIR) and Hs99999901_S1 (18S rRNA) (Applied Biosystems, Foster City, California, USA). Samples were run in duplicate with primers and probes against 18S rRNA and DCIR mRNA in different wells. Samples without added cDNA served as negative controls. The relative amount of mRNA in each sample was calculated as the ratio between the target mRNA and the endogenous control.
Tissue preparation, immunostaining and analysis
Synovial biopsies from healthy controls and patients were obtained by knee arthroscopy from cartilage-proximal areas. Biopsy specimens were snap frozen in dry ice-cooled isopentane and stored at −70°C until sectioned. Serial cryostat sections (7 μm) were formaldehyde fixed and stored at −70°C. Synovial tissue expression of DCIR and CD3 was evaluated immunohistochemically using mouse monoclonal antibody anti-DCIR MAB1748 (R&D Systems, Abingdon, Oxford, UK) and monoclonal irrelevant IgG1 antibody (X 0931, Dakopatts, Copenhagen, Denmark), or in the therapy study using the mouse monoclonal antibodies anti-DCIR 111F8 and anti-CD3 IgG1 (BD Biosciences) and irrelevant IgG1 antibody (Dako). The staining procedure was as previously described.16 Immunohistochemical stainings were performed by overnight incubation with primary antibodies, listed in table 2. The following day, sections were incubated with secondary biotinylated horse anti-mouse antibody (Vector Laboratories, Peterborough, UK) followed by addition of streptavidin, Alexa Fluor 488 conjugate (Molecular Probes Inc, Eugene, Oregon, USA). Avidin/biotin blocking (Vector Laboratories) followed. Secondary biotinylated goat anti-rabbit (Vector Laboratories) was added followed by addition of Rhodamine Red-X-conjugated streptavidin (Jackson Immunoresearch Laboratories Inc, West Grove, Pennsylvania, USA). All antibodies and fluorophores were diluted in phosphate buffered saline (PBS)/0.1% saponin/0.1% bovine serum albumin (BSA). Matched isotype controls were used.
Stained synovial biopsy sections were evaluated using double-blind semiquantitative analysis.17 The scale used for DCIR and CD3 analysis (positive cells) was: 0 = none, 1 = minimal amounts, 2 = low amounts, 3 = moderate amounts and 4 = high amounts.
Analysis was performed at a magnification of 250×.
Anti-citrullinated protein antibody (ACPA) assay
For ACPA detection the Immunoscan RA assay (Euro-Diagnostica AB, Malmö, Sweden) was used. Diluted patient serum (1:50) was applied to the wells and bound antibody was detected by adding horseradish peroxidase (HRP)-labelled anti-human IgG, followed by incubation with substrate. Colour change was evaluated photometrically on an E-max (Molecular Devices, Chicago, Illinois, USA). Data was analysed with SOFTmax PRO V 3.1.1 (Molecular Devices).
The Wilcoxon signed rank test (WSR, t test) was used to compare two groups of paired data: either cellular DCIR expression before and after treatment, or cellular DCIR median fluorescence intensity (MFI) data from PB and SF. Kruskal–Wallis test (KW, one-way analysis of variance (ANOVA)) was used to compare three groups of data, analysing percentages of DCIR+ cell populations between controls, PB and SF from the flow cytometric material. Significance was defined at the 5% level.
The following gene designations were used: CLECSF6 (C-TYPE LECTIN SUPERFAMILY 6), CLEC4A (C-TYPE LECTIN DOMAIN FAMILY 4, MEMBER A), DCIR (DENDRITIC CELL IMMUNORECEPTOR), LLIR (LECTIN-LIKE IMMUNORECEPTOR).
DCIR is widely expressed in the rheumatic joint
By immunohistochemistry, no DCIR+ cells were present in synovial biopsies from the five healthy individuals examined (fig 1A). By contrast, DCIR expression was readily detected in inflamed synovia from all eight patients with RA that were investigated. We also observed a similar pattern in synovia from other inflammatory joint diseases (data not shown). Interestingly, DCIR was widely expressed on multiple cell populations located in infiltrates and sublining areas (fig 1B), as well as in the lining layer (fig 1C).
Synovial T cells displayed a pronounced DCIR expression
By flow cytometry, we detected DCIR on a variety of cell types in synovial fluid, including cell types not previously known to express this molecule. Strikingly, DCIR was expressed on a small but distinct population of T cells and natural killer (NK) cells in synovial fluid. A population of DCIR+ T cells could also be detected in the circulation from patients and controls, but to a much smaller extent and not in all samples analysed. This was true for the CD4+ and CD8+ T cell populations (fig 2B,C). Additionally, DCIR+ NK cells could also be observed more frequently in synovial fluid compared to peripheral blood of patients and controls (fig 2A). Consistently, the DCIR+ NK cells that could be detected were CD56dim rather than CD56bright suggesting them to be classical cytotoxic NK cells.18
In order to substantiate the unexpected finding of DCIR+ T cells in our flow cytometric analysis, we analysed DCIR mRNA expression on pure cell populations by RT-PCR. Blood samples were taken from healthy donors and as positive controls CD14+ monocytes, CD14– granulocytes and CD19+ B cells were isolated. For our T cell analyses, we subsorted DCIR positive and negative CD4+ and CD8+ cells. DCIR mRNA expression was confirmed in the DCIR+ T cell subpopulations to a level similar to that observed in monocytes, while DCIR-negative T cells lacked mRNA expression (fig 2D). Hence we confirmed the presence of DCIR on a subpopulation of T cells by two separate methods.
Synovial DCIR+ T cells were blasted/activated cells, while DCIR+ T cells in peripheral blood were small/resting
Next, we examined the relative quantity of DCIR on the cells by assessing the MFI in combination with defining the size of the DCIR+ cells based on flow cytometric scatter profile. MFI for DCIR+ CD4 and CD8 T cells were markedly higher in synovial fluid compared to peripheral blood (figs 1 and 3), (KW, p value = 0.0313). Although the DCIR+ T cells were a minor population, the cell surface expression levels were high, even higher than that of antigen-presenting cells (see Supplementary data). Additionally, DCIR+ T cells in peripheral blood were consistently small/resting cells, while most DCIR+ T cells in synovial fluid were blasted/activated T cells (fig 3).
In our patient cohort we also detected DCIR expression on cell populations previously demonstrated to be DCIR+ in healthy subjects, ie, DCs, monocytes, granulocytes and B cells (see Supplementary data). In synovial fluid, monocytes and granulocytes are abundant cell populations. Still, notably lower frequencies of synovial DCIR+ monocytes were seen compared to blood of our healthy donors (KW, p<0.001). Among polymorphonuclear cells (PMNs) there was a tendency to upregulation of DCIR in synovial fluid compared to healthy blood.
In contrast to monocytes and PMNs, B cells are much less prevalent in the inflamed joint, and here we observed markedly fewer DCIR+CD19+ cells in the synovial fluid compared to healthy blood (KW, p = 0.0062)
Low DCIR expression on DC populations
With regard to DCs, only a marginal fraction, median 2.5%, of myeloid-derived DC (mDC) in the joint were DCIR+ (fig 4A). Nevertheless, the frequency of cells expressing the mDC marker blood dendritic cell antigen (BDCA)3+ was notably enriched in synovial fluid (by a factor 10–50) compared to blood where this cell population was virtually absent. DCIR expression on plasmacytoid DC (pDC) was more abundant than on mDC. A median of 20% of the synovial pDCs expressed DCIR and a similar level was also detected in blood (fig 4B). However, the overall frequency of BDCA2+ pDC cells was low in peripheral blood and synovial fluid. Scattered DCIR+ pDC could also be demonstrated in synovial membrane specimens from patients with RA (data not shown).
DCIR is downregulated after successful corticosteroid treatment
Since no DCIR+ cells were found in healthy joints (fig 1A) we decided to also determine DCIR expression in rheumatic joints during remission. Six patients receiving intra-articular corticosteroid injections were biopsied at time of treatment and again 2 weeks later. DCIR+ cells were detected in all biopsies before treatment (fig 5). These cells displayed morphological similarities with several cell types including polymorphonuclear cells, fibroblasts, macrophages, endothelial cells and lymphoid cells. At 2 weeks post treatment, DCIR expression decreased (fig 5E,F) (WSR, p = 0.001), paralleled by a general reduction of CD3 positive cells (data not shown).14
To follow up on our discovery of a DCIR+ T cell population, we performed dual immunofluorescence stainings of synovia. Indeed, in joint tissue DCIR was also expressed on a fraction of the T cells, which represented less than 5% of the lymphocyte population (fig 5C). This is in accordance with our flow cytometric data of synovial fluid (fig 2). In fact, two of the patients in the biopsy cohort were also analysed for their synovial fluid composition (patients 5 and 18), confirming DCIR+ T cells in synovium and synovial fluid at the same time point.
Incited by recent genetic data indicating association of DCIR with RA, we aimed to determine on which cell types DCIR were expressed, in which compartments and under which circumstances, ie, in relation to disease status.
No DCIR expression was detected in healthy joints. By contrast, DCIR was expressed on multiple cell populations in inflamed synovia. This expression was downregulated during disease remission following local glucocorticoid treatment. Thus, we provide original evidence of abundant DCIR expression in affected tissues of a human inflammatory disease, and decreased expression accompanying disease amelioration.
Concerning different cell types expressing DCIR, we report abundant DCIR expression on granulocytes and professional antigen-presenting cells in RA, similar to studies of healthy subjects. Additionally, a surprising discovery was the expression of DCIR on a small but significant population of CD56+ NK cells and CD3+ T cells observed by flow cytometry. Since neither of these cell populations have previously been reported to express this molecule, we chose to confirm our observation by double immunofluorescence staining of synovial tissue and by RT-PCR of isolated T cells. Our analyses of DCIR+ T cells demonstrate that they are both of the CD4+ and CD8+ T cell lineages. Interestingly, DCIR+ T cells were much more common in synovial fluid than in blood. In addition, the relative level of DCIR expression was increased in synovial fluid compared to blood. An absolute difference between the two compartments was that DCIR+ T cells in blood were consistently small resting cells whereas most DCIR+ T cells in synovial fluid were large activated cells.
That expression of DCIR on activated T cells is restricted to joints is interesting considering that the intracellular part of the receptor contains an ITIM. Our data therefore suggest that DCIR ligation, or lack thereof, could modulate certain functions of activated T cells important for disease. We are currently performing experiments to further define the DCIR-expressing T cells in relation to categories of T cells that are of interest in rheumatic and autoimmune diseases, such as regulatory T cells19 20 and IL17-producing T cells.21–23 As it appears, signalling through intracellular ITIM on DCIR may inhibit or modify signal transduction through activating immunoreceptors that contain tyrosine-based activating motif(s) (ITAMs), or receptors that bind to adaptor proteins containing ITAMs.24 25 Molecules with ITAMs are expressed on CDIR-expressing cells. Thus, signalling through T cell receptors that associate with the ITAM-containing CD3 complex might be modulated, as well as ITAMs present on, eg, NK cells, ntigen presenting cells and B cells such as NK cell activating receptors, Fc receptors for IgG (FcγRIII, CD16) and IgE (Fc?RI) and the Igα/Igβ heterodimer.
Concerning myeloid dendritic cells (BDCA3+ mDC), we reproduce previous results that they are rare in blood and common in synovial fluid.26 We demonstrate that only a minor proportion of synovial mDCs express DCIR, but we do not presently know if they are activated in similarity to the DCIR+ T cells in the same compartment. Plasmacytoid dendritic cells (BDCA2+ pDC) where not dramatically increased in synovial fluid compared to blood, and the expression of DCIR appears to be decreased in joints. Overall, DC subpopulations in synovial fluid display limited signs of maturation/activation (data not shown).26 Decreased DCIR expression in joints was also observed for monocytes (CD14+) and B cells (CD19+). It is possible that the diminished expression of DCIR on pDC, monocytes and B cells is of clinical relevance since this may translate into reduced signalling via ITIM and impaired downtuning of critical cell activities.
Interestingly, previous studies on transcriptome profiling in arthritis reveals increased DCIR mRNA in mouse joints27 28 and human peripheral blood,29 thereby supporting our data. Herein, we provide detailed DCIR protein expression data and more importantly, we demonstrate that DCIR expression varies between cell types and compartments, which is important to consider for future studies.
Additionally, in a collaborative study DCIR has also been shown to associate with myocardial infarction (Professor Per Eriksson, Department of Medicine, Karolinska University Hospital, Sweden; personal communication) suggesting that this molecule could have implications also in other inflammatory settings.
In summary, we report a number of parameters that are important to consider when addressing DCIR expression in disease, and we demonstrate protein expression on many different leukocyte subpopulations, with novel data for NK cells and T cells. Precisely how this immunoreceptor contributes to arthritis is not known, since there is no known ligand for DCIR, but expression is suggested to be transient and ligation involves carbohydrate recognition (self or non-self). In the context of T cells, it is interesting that Th1, Th2 and Th17 cells have recently been demonstrated to display different glycosylation patterns, suggesting highly regulated interactions with lectin receptors.30 Therefore, with our results of increased DCIR expression in an inflammatory setting, together with the genetic association in rat, mouse and human, we view DCIR as a potential important marker, or even target, for RA and possibly even for inflammation in general.
We wish to thank Annika van Vollenhoven for performing the cell sortings, Eva Jemseby for keeping track of all the samples and Schering Plough for providing the 111F8 antibody.
Competing interests: None.
Funding: This study was supported by grants from the Swedish Medical Research Council, Professor Nanna Svartz Research Foundation, King Gustaf V’s 80-year Foundation, Börje Dahlin Foundation and Karolinska Institutet Foundations.
Ethics approval: All studies were performed after human ethics approval and informed consent was obtained from all contributing individuals.
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.