Introduction
Rheumatoid arthritis (RA) is an autoimmune inflammatory disease characterised by autoantibody production and immune complex (IC) formation. Common autoantibodies are rheumatoid factor (RF) and those against citrullinated peptides (CCPs) [
1]. Approximately 70% of all RA patients display rheumatoid factor and/or anti-CCP antibodies, and the presence of anti-CCP antibodies can be detected in serum several years before disease debut [
2]. Most autoantibodies are of the IgG isotype, which have the potential to activate Fc gamma receptors (FcγRs) on leukocytes, such as macrophages, neutrophils, dendritic cells and B cells. Cross-linking of FcγRs by IgG-ICs leads to cellular effector functions such as phagocytosis, antibody-dependent cellular toxicity and release of inflammatory cytokines.
Three different classes of FcγRs have been identified in humans so far; FcγRI (CD64), FcγRII (CD32) and FcγRIII (CD16). Furthermore, FcγRII and FcγRIII exist in two isoforms, a and b, which carry out divergent functions. FcγRI is a high affinity receptor that binds monomeric IgG as well as IgG-ICs, while FcγRII and FcγRIII are low affinity receptors that predominantly bind IgG-ICs. FcγRI, FcγRIIa, FcγRIIIa and FcγRIIIb are activating receptors. FcγRI and FcγRIIIa consist of an α-chain with three and two Ig-domains respectively, which is connected with a cytoplasmic signalling subunit, the γ-chain. The γ-chain is responsible for intracellular signalling via its immunoreceptor tyrosine based activation motif (ITAM). FcγRIIa is a single chain receptor that contains an ITAM-motif in the cytoplasmic tail. FcγRIIb is an inhibitory receptor that is structurally similar to FcγRIIa, but has an immunoreceptor tyrosine based inhibitory motif in the cytoplasmic domain. FcγRIIb has been shown to have an important negative regulatory function on Fc receptor activation [
3].
The involvement of FcγRs in experimental arthritis has been thoroughly investigated, and it is now clear that activating FcγRs are essential for the development of disease. Thus, mice lacking the common γ-chain or FcγRIII are protected from collagen-induced arthritis (CIA) as well as other experimental models of arthritis [
4‐
8]. Consequently, FcγRIIb deficiency in mice leads to increased susceptibility to CIA [
9,
10]. These findings emphasize the importance of FcγRs in the pathogenesis of experimental arthritis, which may also be true for arthritis in humans. A reported gene polymorphism of FcγRIIIa has been correlated with RA [
11‐
13] as this polymorphism changes the receptor affinity for different IgG-subclasses [
14,
15]. The FcγRIIIA 158 V/F allele variant has been especially associated with the risk of developing RA [
16], although conflicting data exist [
17]. Recently, it was also reported that there is an association between rheumatoid factor and the FcγRIIIa 158 V/F allele in RA patients [
18] and that a functional variant of FcγRIIb is associated with increased joint destruction in RA but not disease susceptibility [
19]. Moreover, several studies have shown that the percentage of FcγRIII positive monocytes is increased in peripheral blood of RA patients [
20,
21] and that the expression levels of FcγRI, FcγRII and FcγRIII on RA monocytes are increased compared to healthy individuals [
22‐
24], while FcγRIIb expression is unaffected [
25].
It has previously been hard to obtain knowledge about FcγR expression in healthy synovial tissue for comparison with FcγR expression in RA patients. Synovia from trauma patients or osteoarthritis patients has often been used as control material [
26,
27] and only limited studies have been done with healthy synovium [
28]. Therefore, we have in this study investigated the expression of the different FcγRs using synovial tissue from healthy volunteers in comparison with RA synovia. We were particularly interested in investigating expression of the inhibitory FcγRIIb as this receptor has not previously been studied due to the lack of a specific antibody against it. Here, by using a novel FcγRIIb specific antibody, GB3, we demonstrate a pronounced FcγRIIb expression in the synovial inflammation in RA, which is in sharp contrast to the lack or weak staining of FcγRIIb in healthy synovia. Additionally, we could detect FcγRII and FcγRIII, but not FcγRI, in healthy synovial tissue. In the RA synovia, the expression of all activating FcγRs was significantly increased, regardless of disease duration. The synovial FcγRI expression could be reduced by intraarticular glucocorticoid treatment.
Materials and methods
Subjects
Synovial tissues from 26 RA patients, 12 men and 14 women, were obtained through the rheumatology clinic at the Karolinska Hospital, Solna, Sweden. Characteristics of the patients are presented in Table
1. Healthy synovial biopsies were obtained from ten volunteers, four men and six women, who did not display any arthritic symptoms at the time of the biopsy. An additional group of nine patients with either RA (three men and two women), oligoarthritis (two men) or polyarthritis (two women) was studied before and after treatment with an intraarticular injection of 40 mg of glucocorticoids (triamcinolone hexacetonide; Lederspan, Wyeth Lederle, Solna, Sweden) in the knee. The biopsies were taken before and 9 to 15 days after treatment. The characteristics of these patients are presented in Table
1. All patients included in the glucocorticoid study displayed a decreased synovial inflammation two weeks after treatment as synovial vascularity and hypertrophy were reduced as assessed by arthroscopic evaluation (data not shown). The ethics committee at the Karolinska hospital approved all experiments on human tissue and informed consent was obtained from all study subjects.
Table 1
Characteristics of patients
1 | Early RA | RA | - | M/64 | 9 mo | None |
2 | Early RA | RA | + | M/63 | 3 mo | NSAID |
3 | Early RA | RA | + | F/77 | 1.5 mo | None |
4 | Early RA | RA | ND | M/20 | 6 mo | NSAID |
5 | Early RA | RA | + | M/40 | 11 mo | None |
6 | Early RA | RA | + | M/54 | 11 mo | None |
7 | Early RA | RA | + | F/62 | 10 mo | NSAID |
8 | Early RA | RA | + | M/75 | 18 mo | NSAID |
9 | Early RA | RA | + | M/82 | 12 mo | NSAID |
10 | Early RA | RA | + | M/51 | 11 mo | NSAID |
11 | Early RA | RA | + | F/76 | 12 mo | Salazopyrin |
12 | Late RA | RA | - | F/47 | 20 yr | MTX |
13 | Late RA | RA | + | F/67 | 40 yr | MTX |
14 | Late RA | RA | + | F/62 | 12 yr | MTX, NSAID |
15 | Late RA | RA | + | F/65 | 30 yr | MTX, NSAID, Pred |
16 | Late RA | RA | + | F/63 | 2 yr | MTX, NSAID, Pred |
17 | Late RA | RA | + | F/83 | 11 yr | MTX, Pred |
18 | Late RA | RA | + | M/38 | 22 yr | Enbrel, MTX |
19 | Late RA | RA | + | F/72 | 44 yr | Cyclosporine, NSAID |
20 | Late RA | RA | + | F/59 | 18 yr | Salazopyrin, Glucosamine |
21 | Late RA | RA | + | M/73 | 20 yr | Paracetamol |
22 | Late RA | RA | + | F/59 | 11 yr | MTX, Remicade |
23 | Late RA | RA | + | M/69 | 40 yr | MTX, Pred, Cyclosporine |
24 | Late RA | RA | + | M/53 | 7 yr | Pred, NSAID |
25 | Late RA | RA | + | F/55 | 4 yr | Remicade |
26 | Late RA | Juvenile RA | + | F/35 | 20 yr | MTX, NSAID |
27 | Glu-study | RA | + | M/53 | 7 yr | Pred, NSAID |
28 | Glu-study | RA | + | M/63 | 3 mo | NSAID |
29 | Glu-study | RA | - | F/35 | 20 yr | MTX, NSAID |
30 | Glu-study | Oligoarthritis | - | M/21 | 4 mo | None |
31 | Glu-study | Oligoarthritis | - | M/53 | 10 yr | None |
32 | Glu-study | Polyarthritis | - | F/43 | 12 yr | Pred, NSAID |
33 | Glu-study | Polyarthritis | + | F/73 | 2 yr | MTX |
34 | Glu-study | RA | ND | M/20 | 6 mo | NSAID |
35 | Glu-study | RA | + | F/55 | 4 yr | Remicade |
Synovial tissue
The synovial membrane biopsies were obtained by an arthroscopic technique as previously described [
29] and were taken, when possible, from synovitis adjacent to the cartilage-pannus junction. Synovial tissue was also obtained at articular surgery in late RA cases. The synovial tissues were divided into three groups: I, healthy (
n = 10); II, early RA, diagnosed for less than 18 months (
n = 11); and III, late RA, diagnosed for more than 18 months (
n = 15). The biopsies from glucocorticoid treated patients (
n = 9) were grouped into before and after treatment.
Tissue preparation
The tissue was snapfrozen in liquid isopentane chilled with dry ice and stored at -70°C until sectioned. Frozen biopsies were embedded in OCT compound (TissueTek, Sakura Finetek, Zoeterwoude, The Netherlands) and sectioned into 7 μm serial sections using a cryostat onto SuperFrost Plus slides (Menzel-Gläser, Braunschweig, Germany). The slides were air dried for 30 minutes, then fixed for 20 minutes at 4°C with 2% (volume/volume) formaldehyde (Sigma, St Louis, MO, USA) in PBS, pH 7.4, then washed in PBS and left to air dry and stored at -70°C until use.
Staining of cell lines
The GB3 (mouse IgG1) monoclonal antibody (mAb), specific for human FcγRIIb, was generated by immunization of mice with a recombinant CD32b coupled cyclic 126-SKKFSRSDPNFSG-138 peptide, including N-acetyl-glucosaminylated asparagine at position 135. The cyclic peptide represents a loop in the Fc-fragment binding region of CD32 that carries unique residues for the inhibitory b-form of the receptor and should facilitate a specific immune response against the CD32b unique glycosylation site at Asn135 of the receptor. Mice were sacrificed and splenic B-cells were fused with immortalized cells using standard protocols. The propagation of approximately 700 hybridoma clones resulted in the GB3 clone. To test the specificity of the GB3 mAb, the monocytic U937 cell line (kind gift from Prof. Lars Hellman, Uppsala University, Sweden) and the B cell lymphoma Raji cell line (kind gift from Dr Fredrik Öberg, Uppsala University, Sweden) were stained with the GB3 mAb, mouse IgG2b anti-FcγRIIa (clone IV.3; kind gift from Dr Johan Rönnelid, Uppsala University, Sweden), mouse IgG1 pan anti-human FcγRII (clone KB61, DAKO, Glostrup, Denmark) or isotype controls (mouse IgG1, DAKO and mouse IgG2b, Sigma) for 30 minutes at 4°C. The cells were then washed in 1% BSA in PBS and a phycoerythrin-(PE-)conjugated rabbit anti-mouse IgG secondary antibody (Biosite, Täby, Sweden) was added to the cells and incubated for 30 minutes at 4°C. After further washing, the cells were re-suspended in 1% BSA in PBS and analysed using a FACScan (Becton Dickinson, Mountain View, CA, USA).
ELISA
To confirm the specificity of the GB3 mAb, enzyme-linked immunosorbent assay (ELISA) was used. Microtiter plates (Immunolon 2 HB, Dynex Technologies Inc., Chantilly, VA, USA) were coated with 2 μg per well of recombinant soluble (s)FcγRIIb or recombinant sFcγRIIa (both R&D systems, Minneapolis, Minnesota, USA) diluted in 0.5% BSA in PBS and incubated overnight at 4°C in a humid chamber. The plate was then washed with 0.05% Tween 20 (Merck, Schuchardt, Germany) in PBS. The epitope specificity of the GB3 mAb was determined by titrating the antibody five times in each step, with a starting concentration of 5 μg/ml in 0.5% BSA, and then serially added to the plate in duplicates and incubated for 2 h at room temperature (RT). The plate was washed and 50 μl of sheep anti-mouse IgG conjugated to alkaline phosphatase (Sigma) was added per well and incubated for 2 h at RT. The plate was washed and 50 μl per well of p-nitrophenyl phosphate substrate (1 mg/ml; Sigma) in ethanolamine buffer were added and the plate incubated in the dark. The absorbance value was determined at 405 nm in an ELISA reader (Molecular Devices Corporation, Sunnyvale, CA, USA).
Immunohistochemistry
Slides were thawed and washed in PBS with 0.1% saponin (PBS/Sap, pH 7.4) for 10 minutes. Any endogenous peroxidase was blocked using 1% H2O2 in PBS/Sap for 1 h at RT in dark. Sections were then washed repeatedly in PBS/Sap. The primary antibody, diluted in PBS/Sap, was added and left at RT over night in the dark. After several washing steps, the sections were incubated with 1% normal horse serum in PBS/Sap for 15 minutes. The serum was thereafter removed and biotinylated horse anti-mouse secondary antibody, diluted in 1% normal horse serum, was added for 30 minutes at RT. Sections were washed and avidin-biotin-complex (ABC-elite, Vectastain elite kit, Vectorlab, Burlingame, CA, USA) added for 45 minutes according to the manufacturer's instructions. Subsequently, after washing, any positive staining was developed in diaminobenzidine (DAB; DAB substrate kit, Vectorlab) for 7 minutes according to the manufacturer's instructions. Finally, the sections were counterstained with haematoxylin (Histolab, Gothenburg, Sweden), rinsed with tap water, dehydrated in alcohol and mounted with Mountex (Histolab).
Immunofluorescent staining
Formaldehyde fixed sections were washed in PBS/Sap and incubated with 0.1% BSA in PBS/Sap for 30 minutes at RT. If necessary, 20% normal human serum was added for 30 minutes at RT to block any non-specific binding of antibody before incubation with 0.1% BSA in PBS/Sap. Fluorescently labelled or unlabelled primary antibody was then diluted in PBS/Sap with 0.1% BSA and added for 90 to 120 minutes at RT followed by washing in either PBS or PBS/Sap. For non-conjugated antibodies, biotinylated or fluorescently labelled secondary antibody was added for 30 minutes followed by washing in PBS/Sap. Then streptavidin conjugated to either alexa-red or green was added for one hour when needed. Sections were then washed in PBS, dried and mounted in PBS/glycerol or Mowiol (Calbiochem, San Diego, CA, USA).
Primary antibodies
A commonly used marker for macrophages is the CD68 antigen. However, this marker can also be found on fibroblasts and other leukocytes depending on the mAb clone used for detection [
30,
31]. So to avoid this and exclusively investigate mature monocytes/macrophages as a source of FcγR expression, we used a mAb against CD163. The CD163 antigen is known to be exclusively expressed by mature peripheral blood monocytes and macrophages [
32,
33]. The antibodies used were thus PE-conjugated anti-CD163 (clone 215927; R&D-systems), non-conjugated anti-CD163 (clone Ber-MAC3; DAKO), FITC-conjugated and non-conjugated anti-CD64 (clone 10.1; BD Pharmingen, San Diego, CA, USA), FITC-conjugated and non-conjugated anti-CD32 (clone KB61; DAKO), non-conjugated anti-CD32b (clone GB3), FITC-conjugated and non-conjugated anti-CD16 (clone DJ130c) (DAKO), non-conjugated anti-CD3 (clone SK7; Becton Dickinson), non-conjugated anti-CD19 (clone HD37; DAKO) and non-conjugated anti-CD20 (clone L27; Becton Dickinson). All antibodies were of mouse IgG1 isotype and an irrelevant mouse IgG1 (DAKO) was used as negative control. As an additional specificity control, the GB3 mAb was absorbed by incubating the GB3 mAb with recombinant human FcγRIIb protein at a 1:1 ratio overnight before adding it to the tissue sections.
Secondary antibodies
Secondary antibodies included biotinylated horse anti-mouse IgG (Vector, Burlingame, CA, USA), FITC-conjugated F(ab')2-fragment of rabbit anti-mouse immunoglobulins (DAKO) and PE-conjugated rabbit anti-mouse IgG secondary antibody (Biosite).
Microscopic analysis
The immunohistochemical staining was analysed in a Polyvar II light microscope (Reichert-Jung, Vienna, Austria) and evaluated by two independent observers (SEM and SK). Both observers were blinded to the tissue identity and staining. An arbitrary scale was used to identify the amount of positively stained area of the whole tissue section, where 0 = 0% positive tissue area, 1 = 1% to 20% positive tissue area, 2 = 21% to 50% positive tissue area, 3 = 51% to 80% positive tissue area and 4 = 81% to 100% positive tissue area. The staining pattern was also noted as well as presence of vessels and lymphocyte infiltrates. The immunofluorescence staining was analysed using a Leica DMRXA2 fluorescence microscope (Leica Microsystems, Cambridge, UK).
Statistical analyses
The Mann-Whitney rank sum test was used on unpaired immunohistochemistry data, the Wilcoxon signed rank test on paired immunohistochemistry data and correlations were determined using Spearman's rank correlation.
Discussion
These data demonstrate that the inflammation in RA synovium is characterised by a pronounced expression of the inhibitory FcγRIIb, which suggests it has a role in counteracting the effects of activating FcγRs in RA synovia. Specific FcγRIIb expression patterns in RA or healthy synovia have not previously been described; nor has the expression of activating FcγRs in healthy synovia been described fully. Synovium from trauma patients has been reported to express FcγRI, II and III, and the FcγRII and III expression has been shown to be significantly lower in these patients compared to RA patients, while their FcγRI expression does not differ from that of RA synovia [
27]. This is in contrast to our findings where we could not identify any FcγRI expression in healthy synovium. Thus, it appears that FcγRI is expressed as a consequence of a general inflammation in the joint, as we also observed FcγRI expression in synovial tissue from osteoarthritis patients (data not shown), although not to the same extent as seen in RA patients. The fact that FcγRI is absent from healthy synovial tissue but present in RA synovium and significantly decreased by administration of glucocorticoids indicates that FcγRI has a significant inflammatory role in the pathogenesis of arthritis. This is interesting and in line with experimental studies where FcγRI deficiency in mice leads to a reduced uptake of IgG-ICs and to decreased IC-induced inflammation [
36]. It has also been reported that up-regulation of FcγRI leads to increased cartilage destruction in arthritic mice [
37]. Similar to FcγRI, FcγRIIb was absent from, or only weakly expressed in, healthy synovia, whereas RA synovium clearly expressed FcγRIIb. This indicates a need for FcγRIIb to control the stimulatory activity of the ITAM-containing receptors in the RA joint.
Although the majority of RA macrophages expressed FcγRIIb, it was also evident that some macrophages did not. This may point towards a small subpopulation of FcγRIIb negative macrophages that may have extraordinary inflammatory capacities. We also observed FcγRIIb expression in lymphocyte infiltrates of RA synovium, which further suggests that infiltrating B cells may also be regulated by this receptor. In contrast to the near absence of FcγRIIb expression in healthy synovia, we clearly observed positive FcγRII staining with the pan anti-FcγRII mAb, most likely as a result of the presence of the activating FcγRIIa in healthy synovia. FcγRIII is also expressed in healthy synovia, ready to bind ICs caught in the joint. In RA synovium both FcγRII and FcγRIII expression was significantly increased, as has also previously been observed [
26‐
28]. However, no association with disease duration and the degree of FcγR expression in RA patients could be shown, as both early and late patient groups expressed similar amounts of FcγRs. This suggests that the FcγRs are important during the whole disease course and not only in the induction phase of RA.
The importance of FcγRII and FcγRIII in arthritis has also been emphasized in animal models of RA. Transgenic mice expressing human FcγRIIa [
38] develop CIA much earlier than wild-type mice and normally arthritis resistant mice become susceptible to arthritis when expressing FcγRIIa [
39]. Furthermore, the induction of CIA is dependent on FcγRIII and, in particular, FcγRIII positive macrophages [
4,
40]. Studies of RA monocytes/macrophages have also stressed the significance of FcγRII and FcγRIII in disease pathogenesis. Thus, FcγRII and FcγRIII are up-regulated on peripheral blood monocytes and FcγRIII expression is also enhanced on synovial fluid macrophages [
21,
22].
The importance of FcγRII in RA was also noted in this study by the decrease in FcγRII expression (albeit not significant) after glucorticoid treatment. The decrease in FcγRII as well as FcγRI expression after local glucocorticoid injection was not due to a reduced number of macrophages in the tissue, as no significant difference in CD163 expression was found after treatment. In agreement, a previous study, where the same patient material was included, also found that the amount of CD163 and CD68 positive cells were not affected by intraarticular glucocorticoid treatment [
41]. This indicates that FcγRs are down regulated from the cell surface by glucocorticoid treatment, which may help to explain some of the improvement seen in RA patients upon treatment with it, in addition to its reported suppressive effects on cytokines [
41]. These results are in line with an earlier report that indirectly showed that FcγR expression is decreased on peripheral blood monocytes from autoimmune hemolytic anemia patients after systemic administration of corticosteroid, measured by radiolabelled IgG-binding [
42]. In a more recent paper, FcγRI and II were shown to be decreased on peripheral blood monocytes one month after systemic therapy with daily low glucocorticoid doses [
22]. Recent studies with other anti-rheumatic drugs have demonstrated that methotrexate treatment could reduce the expression of FcγRI and IIa on peripheral blood monocytes from RA patients, while the expression of FcγRIII was unaffected [
43]. We did not see a clear reduction in FcγRIII expression after glucocorticoid treatment, which might indicate that FcγRIII expression is hard to modify using anti-rheumatic drugs. It is difficult to speculate on how relevant the reduction of FcγRs is for the physiological outcome of glucocorticoid treatment, since patients receiving an intraarticular injection of glucocorticoids experience an almost instant effect, but positive anti-inflammatory effects in the joint are seen several weeks after treatment.
Macrophages are present in the synovial lining layer of healthy synovium and the amount of macrophages in joints of RA patients is correlated with disease activity [
44‐
46]. We identified the presence of FcγRI, II, IIb and III on RA synovial macrophages and, in addition, expression of FcγRI, FcγRII and FcγRIII revealed by DAB staining was significantly correlated with the expression of the macrophage marker CD163. This indicates that macrophages are likely to be involved in the IC-mediated damage in RA via their FcγR expression and they may also be responsible for antigen presentation to T cells, as macrophages were observed in close proximity to CD3 positive T cells in the RA synovium. The FcγR-positive macrophages were often localised perivasculary together with CD3-positive T cells, most likely as a result of recent extravasation. It is possible that presentation of antigen taken up via FcγRs on the macrophages may activate T cells to secrete cytokines. This could, in turn, lead to further activation of the macrophages and, thus, production of inflammatory cytokines, resulting in a continuous inflammatory state in RA joints.
Conclusion
Our findings demonstrate that expression of the inhibitory FcγRIIb, as well as of the activating FcγRI, FcγRII and FcγRIII, is increased in RA synovium, regardless of disease duration. The importance of FcγRI and FcγRIIb in the synovial inflammation of RA patients is further highlighted by the fact that healthy synovia lack FcγRI expression and substantially lack FcγRIIb expression. Furthermore, anti-inflammatory drugs, such as glucocorticoids, suppress FcγRI expression after local administration in the joint. These results clearly point towards a central role for the FcγRs in the synovial inflammation of RA.
Acknowledgements
The authors acknowledge the invaluable help of Erik af Klint, MD, for collecting biopsies from early RA patients and healthy individuals, Andre Stark, associate professor, for collecting material from late RA patients and Dimitrios Makrygiannakis, MD, for tissue sectioning healthy biopsies. This study was supported by The Swedish Medical Research Council, The Swedish Rheumatism Association, King Gustaf V's 80 years Foundation, Börje Dahlin Foundation, The Clas Groschinsky Memorial Foundation, Åke Wiberg Foundation and Freemason "Barnhuset" in Stockholm.
Competing interests
Dr Uwe Jacob declares a commercial interest in SuppreMol GmbH. He is one of the cofounders and shareholders of SuppreMol, which develops FcγR agonists and antagonists for clinical use. Dr Uwe Jacob is also listed as inventor on patent applications of SuppreMol regarding the specific FcγRIIb antibodies. The other authors of this paper declare no potential conflicting financial interests.
Authors' contributions
SK and A-KU were the initiators of the study. SK, SEM, A-KU and ME planned the experiments together. SEM performed all experiments. ME did the tissue sectioning of the biopsies. Immunohistochemistry experiments were performed by SEM with guidance and technical help from ME. SEM and SK did the analysis and evaluation of the immunohistochemistry data. UJ developed the GB3 anti-FcγRIIb specific mAb and contributed with valuable ideas for the study. All authors have read and commented on the text of this paper.