Objectives Anticitrullinated protein antibodies (ACPA) are specifically associated with rheumatoid arthritis (RA) and produced in inflamed synovial membranes where citrullinated fibrin, their antigenic target, is abundant. We showed that immune complexes containing IgG ACPA (ACPA-IC) induce FcγR-mediated tumour necrosis factor (TNF)-α secretion in macrophages. Since IgM rheumatoid factor (RF), an autoantibody directed to the Fc fragment of IgG, is also produced and concentrated in the rheumatoid synovial tissue, we evaluated its influence on macrophage stimulation by ACPA-IC.
Methods With monocyte-derived macrophages from more than 40 healthy individuals and different human IgM cryoglobulins with RF activity, using a previously developed human in vitro model, we evaluated the effect of the incorporation of IgM RF into ACPA-IC.
Results IgM RF induced an important amplification of the TNF-α secretion. This effect was not observed in monocytes and depended on an increase in the number of IgG-engaged FcγR. It extended to the secretion of interleukin (IL)-1β and IL-6, was paralleled by IL-8 secretion and was not associated with overwhelming secretion of IL-10 or IL-1Ra. Moreover, the RF-induced increased proinflammatory bioactivity of the cytokine response to ACPA-IC was confirmed by an enhanced, not entirely TNF-dependent, capacity of the secreted cytokine cocktail to prompt IL-6 secretion by RA synoviocytes.
Conclusions By showing that it can greatly enhance the proinflammatory cytokine response induced in macrophages by the RA-specific ACPA-IC, these results highlight a previously undescribed, FcγR-dependent strong proinflammatory potential of IgM RF. They clarify the pathophysiological link between the presence of ACPA and IgM RF, and RA severity.
- Rheumatoid Arthritis
- Rheumatoid Factor
Statistics from Altmetric.com
In the affected joints of patients with rheumatoid arthritis (RA), activated resident and infiltrating cells contribute to synovitis notably by producing cytokines among which tumour necrosis factor (TNF)-α, mainly produced by synovial tissue macrophages, has a pivotal role.1 RA is also associated with the production of rheumatoid factors (RF), autoAbs targeting the Fc portion of IgG discovered around 75 years ago. Although not highly specific for RA, RF of the IgM class (IgM RF) is present in the serum of about 80% of patients. Moreover, it is produced abundantly by plasma cells in inflamed rheumatoid synovial tissues2–7 and IgM RF-containing immune complexes (IC) accumulate in RA joints.4 ,8–10 In addition, the more recently discovered IgG autoAbs called anticitrullinated protein antibodies (ACPA) are present in the serum of about 75% of patients with established disease and have a high (>95%) diagnostic specificity.11 Citrullyl residues, generated by post-translational deimination of arginyl residues (‘citrullination’), are essential components of the epitopes ACPA recognise.12–15 ACPA are also produced and concentrated in inflamed rheumatoid synovial membranes.16–18 Moreover, citrullinated proteins are abundant therein, and among them citrullinated fibrin, that is present in interstitial deposits and thus constitutes a prominent target for the local formation of disease-specific IC.19
The hypothesis of a contribution of ACPA and RF in the RA pathophysiology notably stems from their early appearance,20 frequently before the onset of symptoms,21 ,22 and from their definitely established link to disease aggressiveness.23–28 To verify the arthritogenic potential of ACPA, we developed a human in vitro model in which monocyte–derived macrophages were stimulated with ACPA-IC generated by immunocapture of ACPA from RA sera by in vitro citrullinated human fibrinogen (C-FBG) immobilised on culture wells, in the absence of an exogenous supply of complement components. Supporting the hypothesis of pathophysiological role for ACPA in RA, induction of TNF-α secretion by the ACPA-IC was observed and the major contribution of the activating FcγRIIa was established.29 Despite the ancient consensus regarding the participation of RF-containing IC in the RA synovitis, the pathophysiological RF targets and effector pathways involved have not been extensively investigated. IgM RF-containing IC are thought to mediate tissue injury by activating the classical pathway of complement.30–32 Regarding FcR-dependent responses, only one report concerns mononuclear phagocytes.9 It shows that IC containing IgM RF and IgG RF promote prostaglandin E production by monocytes, inhibitable by staphylococcal protein A that interferes with FcγR-mediated responses.
In the present study, as it constituted a factor potentially modulating their FcR-mediated inflammatory effect, the influence of IgM RF on the macrophage response to ACPA-IC was evaluated using the in vitro model developed previously.
IgMs with and without RF reactivity
The virtually monoclonal IgM with RF reactivity (IgM RF), termed ‘HUL’, was purified from the serum of a patient with type II mixed cryoglobulinemia, as previously described.33 A human myeloma-derived (monoclonal) IgM, whose absence of RF activity was verified by measurement of binding to human or to rabbit IgG Fc fragments (multiplexed FIDIS Rheumatoid assay, Biomedical Diagnostics, Marne la Vallée, France), was purchased from Athens Research and Technology (Athens, Georgia, USA). In all IgM solutions, contamination by endotoxin, checked using a LAL assay (QCL-1000; Cambrex Bioscience, Verviers, Belgium), was less than 0.15 EU/µg.
Monocyte purification and macrophage differentiation were described previously.29 Stimulation by IC was performed essentially following a published method.29 Briefly, 96-well plates were coated with C-FBG and blocked with PBS containing 2% BSA (PBS-BSA). IC were then generated by incubation with 100 µL/well of an ACPA-positive IgG fraction (ACPA+ IgG), prepared as described already from a pool of RA sera with high ACPA titers29 and used at 2.5 mg/mL after dilution in PBS-BSA containing 2 M NaCl (PBS-BSA-NaCl) supplemented or not with HUL at the indicated final concentration. After being washed, each well received 50 000 cells suspended in a serum-free medium (Macrophage-SFM, Gibco, Cergy Pontoise, France). In some experiments, instead of being co-incubated with the ACPA+ IgG, HUL (25 µg/mL) was allowed to incorporate into the IC during a subsequent incubation stage, followed or not by another incubation with 100 µL/well of ACPA− IgG, prepared from a pool of sera from patients with non-RA rheumatic diseases,29 at 2.5 mg/mL in PBS-BSA-NaCl, all incubations being preceded and followed by washing. Alternatively, cells were cultured in wells previously coated overnight at 4°C with 50 µL/well of HUL at 20 µg/mL in PBS that, prior to cell addition and after blocking with PBS-BSA, was incubated either with ACPA+ IgG or ACPA− IgG (at 2.5 mg/mL in PBS-BSA-NaCl) or with buffer alone. Negative controls were included in all assays, consisting of culture on uncoated wells or on C-FBG-coated wells previously incubated with either PBS-BSA-NaCl or with ACPA− IgG. Controls in which the RF-negative control IgM replaced HUL were also included in most experiments. After 24 h culture, supernatants were harvested and stored at −30°C until measurement of cytokine concentrations. TNF-α was assayed by ELISA (eBiosciences, Hatfield, UK). Alternatively, TNF-α, interleukin (IL)-1β, IL-6, IL-8, IL-10 and IL-1Ra were simultaneously assayed using a multiplexed assay (Procarta Cytokine Profiling Kit, Panomics-Affymetrix Inc., Santa Clara, California, USA) read with a Luminex 100 IS flow cytometer (Luminex, Austin, Texas, USA). All conditions were tested in triplicate and results correspond to the means of the three measurements. For the calculation of amplification or cytokine ratios, values <1 pg/mL were replaced by 1 pg/mL. Cytokine secretions obtained in different culture conditions were compared using Wilcoxon's or Friedman's test.
IgM RF enhances the macrophage TNF-α secretion induced by ACPA-IC
To appraise how incorporation of IgM RF into ACPA-IC would influence the TNF-α response of macrophages to these IC, monocyte-derived macrophages of healthy blood donors were stimulated with IC containing these two autoAb families. They were formed by incubating immobilised C-FBG simultaneously with ACPA+ IgG and with a human monoclonal IgM RF (IgM RF ‘HUL’). After 24 h culture, TNF-α secretion was compared with that induced by ACPA-IC formed in the absence of any IgM. Figure 1A shows the response of macrophages from one blood donor to IC formed in the presence of increasing doses of HUL. Under control conditions where no IC formed, that is, when C-FBG was incubated in buffer alone or with an ACPA-negative (ACPA−) IgG fraction in the absence or presence of HUL at 3.1, 12.5 or 50 µg/mL, no or very low TNF-α secretion occurred. In contrast, a significant TNF-α secretion occurred when ACPA-IC formed, that is, when C-FBG was incubated in the presence of ACPA+ IgG. Moreover, when such IC formed in the presence of HUL at 12.5 and 50 µg/mL, the TNF-α secretions multiplied by 2.6 and 6.2, respectively. Such dose-dependent amplifications of TNF-α secretion were replicated using various increasing doses of HUL during IC formation with the macrophages from three additional donors (see online supplementary figure S1). Then, the TNF-α secretions obtained from the macrophages from eight different donors were compared when IC were formed in the absence or in the presence of HUL consistently used at 25 µg/mL (figure 1B). With HUL, the TNF-α secretion was significantly enhanced compared with that induced by IC formed in its absence (p<0.01, median amplification ratio=3.0), and such significant amplification was replicated using another cryoglobulinemia-derived monoclonal IgM RF termed ‘Alt’ also at 25 µg/mL (see online supplementary figure S2). With HUL, the TNF-α secretion was also significantly enhanced compared with that induced by IC formed in the presence of a control IgM devoid of RF activity (p<0.01, median amplification ratio=3.6), which did not differ from the TNF-α secretion obtained in the absence of any IgM, ruling out an effect of the IgM RF that would be independent of its capacity to associate with IgG (figure 1B).
IgM RF enhances TNF-α secretion in conjunction with IgG, and this occurs in macrophages but not in monocytes
To get an insight into the mechanisms underlying amplification of IC-induced responses by IgM RF, we verified that, on its own, HUL did not prompt TNF-α secretion (figure 2A). When macrophages from five different donors were cultured in wells on which HUL had been passively adsorbed, virtually no TNF-α secretion occurred. Conversely, TNF-α concentration could reach high values when the adsorbed HUL had been previously incubated with IgG, whether these were ACPA+ or ACPA− IgG, showing that IgM RF induces significant TNF-α secretion only when associated with IgG. As expected, insignificant TNF-α secretions occurred when the macrophages were cultured in wells coated with the control IgM, previously incubated or not with IgG, be they ACPA+ or ACPA−.
Most probably, during simultaneous incubation of C-FBG with the ACPA+ IgG and the IgM RF, the IC formed not only include C-FBG and the ACPA immunocaptured by this autoantigen but also additional IgG immunocaptured by the IgM RF. Both the IgM RF and these additional IgG possibly playing a role in the amplification of TNF-α secretion, their contributions were separately assessed with the macrophages from 10 different donors (figure 2B). When HUL was introduced into the IC after washing out the IgG unconjugated to C-FBG, the TNF-α secretion did not significantly vary compared with that induced when incubating the preformed IC with buffer alone. However, when further incorporation of IgG into the IC was allowed by subsequent incubation with human IgG (in these assays by incubation with ACPA− IgG), the TNF-α secretion was systematically enhanced compared with that induced in the absence of such IgG (incubation with buffer alone) and with that induced in the absence of HUL (median amplification ratio=18 and 24, respectively; p<0.01 in both comparisons). Therefore, when all reagents are simultaneously incubated during IC formation, the increase in the IgG content of the IC arising from the presence of HUL decisively contributes to the amplification of TNF-α secretion.
Finally, using monocyte and macrophage pairs from the same individuals allowed confirming that differentiation into macrophages greatly influenced the sensitivity to IC that did not contain IgM RF29 and to show that this also held true for IC formed in its presence since monocytes exhibited extremely low TNF-α responses to both IC types (see online supplementary figure S3).
IgM RF drives the cytokine secretion induced by ACPA-IC towards an even more pronounced proinflammatory profile
TNF-α, IL-1β, IL-6, IL-8, IL-1Ra and IL-10 were simultaneously assayed in culture supernatants from the macrophages of six different blood donors stimulated with ACPA-IC formed in the absence or presence of HUL (figure 3A). As equally observed previously for TNF-α, for IL-1β, IL-6 and IL-10, virtually no basal secretion occurred when macrophages were cultured on uncoated wells (median=1, 1 and 3 pg/mL, respectively), whereas substantial amounts of IL-8 and IL-1Ra were spontaneously secreted (median=1105 and 2407 pg/mL, respectively) (not shown). However, after culture in wells coated with C-FBG not previously incubated with any Ig, none of the cytokine secretions measured significantly differed from secretions in the basal condition. As a result, the median elevations in secretion of each cytokine were relatively low (figure 3A). Conversely, following stimulation with IC only including C-FBG and ACPA, for all cytokines, specific cytokine secretion, calculated by subtracting background secretion (obtained after culture under conditions where ACPA− IgG replaced ACPA+ IgG), was significantly enhanced compared with what was specifically obtained upon culture on C-FBG (median ratio=5.0, 2.8, 11, 13.8, 6762 and 44.4 for TNF-α, IL-1β, IL-6, IL-8, IL-1Ra and IL-10, respectively; p<0.05 for all cytokines). Moreover, following stimulation with IC formed in the presence of HUL, not only the specific TNF-α secretion was systematically increased, as expected (median amplification ratio=26.4, p<0.05), but also those of IL-1β, of IL-6 and of IL-10 were significantly upregulated (median ratio=4.8, 17.5 and 6.9, respectively; p<0.05 for all comparisons). An increase in the production of IL-8 or IL-1Ra was noted for two donors, but it was not systematic and, overall, was not significant (median ratio=1.8 and 1.3, respectively). Interestingly, the TNF-α:IL-10 ratio systematically increased with IC formed in the presence of HUL (p<0.05), while the IL-1Ra:IL-1β ratio decreased for all donors but one and the difference shortly missed significance (p=0.0625) (figure 3B). Finally, macrophage supernatants obtained after stimulation with ACPA-IC formed in the absence of RF dose-dependently enhanced basal IL-6 secretion by RA fibroblastic synoviocytes. Moreover, this IL-6-inducing capacity was enhanced (median amplification ratio=4) after stimulation with ACPA-IC formed in the presence of HUL. That a blocking anti-TNF-α antibody inhibited up to ≈45% of the IL-6 production induced by the corresponding macrophage supernatants (undiluted median concentration 6 ng/mL) while it inhibited up to ≈75% of the IL-6 induced by TNF-α at 5 ng/mL (see online supplementary figure S4) indicated the major, but not exclusive, contribution of TNF-α in their bioactivity.
After having demonstrated the capacity of IC containing ACPA to induce macrophage secretion of TNF-α,29 we speculated that this potential would be modulated by incorporation of IgM RF into the IC. With the monocyte-derived macrophages from more than 40 healthy blood donors and different monoclonal IgM with RF activity from patients with cryoglobulinemia, this study demonstrated that incorporation of IgM RF into ACPA-IC can substantially enhance their capacity to induce TNF-α.
This is of particular importance since TNF-α is known to regulate many pathophysiological processes and has been proved to play a central role in RA.34 However, it was also important to check the induction of other cytokines also known to play significant roles in the RA synovitis. That ACPA-IC induced a considerable enhancement of secretion of the chemotactic and proangiogenic cytokine IL-8 and that equally high secretion levels were obtained after stimulation with IC incorporating IgM RF suggest that both types of IC can significantly contribute to joint infiltration by inflammatory cells and to angiogenesis in the synovial membrane. In addition, IC incorporating ACPA and IgM RF prompted significant secretions of IL-1β, also indicating that they can contribute to joint damage notably via the release of matrix metalloproteinases that this cytokine stimulates. Also noteworthy is the significant induction of IL-6 by ACPA-IC and its enhancement in the presence of IgM RF. However, IL-1Ra counterbalances the effects of IL-1β and the data showed that ACPA-IC importantly raised the level of IL-1Ra. Moreover, IL-10, the autocrine inhibitor of macrophage activation, is also induced by ACPA-IC. Nevertheless, interestingly, after incorporation of IgM RF into the IC, the balance between IL-1β and IL-1Ra tended to shift towards the inflammatory mediator and even if IL-10 secretion was enhanced, it did not prevent the parallel upregulation or sustained expression of the inflammatory mediators TNF-α, IL-1β, IL-6 and IL-8 it is known to downregulate. Furthermore, that RF increased the ability of the ACPA-IC-induced macrophage cytokine cocktail to prompt important IL-6 secretion by RA synoviocytes confirms that RF increases the proinflammatory bioactivity of the cytokine response to ACPA-IC, notably its capacity to induce a cytokine playing a major role in the synovitis, joint damage and autoantibody production.
That conjugation of IgM RF with IgG was necessary for the induction of TNF-α secretion indicates a major role of FcγRs in the macrophage response to IgM RF-containing IC. In the IC formed during simultaneous incubation of C-FBG with the ACPA+ IgG and the IgM RF, fixation of additional IgGs on the free antigenic sites of the IgM RF generates IC with a size and a density of IgG largely higher than those of the ACPA-IC. Therefore, the increased potency for inducing TNF-α secretion by macrophages most probably arises from a higher degree of occupancy of their FcγR. Moreover, that the IC formed during the successive incubation of C-FBG with ACPA+ IgG then with the IgM RF do not prompt a significantly increased TNF-α secretion compared with ACPA-IC suggests that, at least in our experimental conditions where ACPA are immobilised following their interaction with C-FBG, the cross-linking of ACPA by the multivalent IgM RF, though likely to enhance the avidity of the interaction between ACPA and FcγR, exerts a minimal direct contribution in the cytokine-inducing signal that transduces through these receptors. Moreover, even if a direct interaction of the IgM RF with a receptor for IgM that could stabilise the overall interaction of cells with IC containing ACPA and RF cannot be excluded,35 ,36 in view of our results, it is unlikely that such receptor acts as a signal transducing molecule able to directly prompt TNF-α secretion or to boost activation via FcγR.
That monocytes proved refractory to induction of TNF-α secretion by IgM RF-containing IC provides a clue as to why the IgM RF-containing IC present into the circulation of RA patients remain fairly innocuous as long as they circulate at a relatively low concentration. Conversely, when they accumulate in tissues, where they are notably met by macrophages endowed with a higher capacity of cytokine response, they are more likely to cause inflammation and the extra-articular manifestations to which they are associated.37 However, it should be stressed that our study only investigated the response of M-CSF-differentiated macrophages, taken as prototypic synovial tissue macrophages38 ,39 while macrophages can be phenotypically and functionally heterogeneous,40 and recent data suggest that different subpopulations of macrophages with different Fcγ receptor expression profiles coexist in the RA synovial membrane.41 ,42 It would therefore be interesting to compare the cytokine responses of differentially polarised macrophages to ACPA-IC. Assessing how these IC modulate spontaneous cytokine secretion by RA synovial tissue cell suspensions could even be more informative. Moreover, the in vitro model deliberately only investigated FcR-mediated pathways by performing macrophage activation in the absence of exogenous supply of complement components. However, it was shown that ACPA and RF could induce complement activation suggesting that IC containing ACPA and RF can also promote inflammation through activation of complement receptors.30 ,43 ,44 Consistent with this hypothesis, a comparison of in vitro complement activation by immobilised ACPA-IC formed in the absence and in the presence of HUL demonstrated an important increase in the production of the terminal complement component C5-b9 when RF is present and therefore an increased capacity for complement activation (see online supplementary figure S5). Finally, though beyond the scope of the present study, given the presence of soluble ACPA-containing and/or RF-containing IC in the synovial fluid of RA patients and that divergent results have been reported concerning the inflammatory effects of soluble versus insoluble IC in monocyte-macrophages,45–48 exploring macrophage responses to soluble IC containing ACPA and/or RF could bring information concerning their inflammatory potential in this compartment.
Interestingly, some degree of competition can exist between FcγRs and IgM RFs for accessibility to their respective binding sites on the Fc portion of IgG,49 possibly counterbalancing the enhancing effect IgM RF has on signalling through FcγRs. Obviously the level of competition is expected to be influenced by the fine antigenic specificity of the related RF. Variations in this antigenic specificity have been reported not only for polyclonal RF or IgM RF clones derived from RA patients, but also for monoclonal IgM RF cryoglobulins.50–52 Therefore, though it requires cumbersome purification procedures, it would be very interesting to study the capacity of IgM RF isolated from the serum of individual RA patients to amplify the IC-induced inflammatory response.
The presence of ACPA within RA synovial membrane, where they form IC with citrullinated autoantigenic peptides, makes them a target for the IgM RF also present in these tissues. ACPA mainly belong to the IgG1 isotype that is recognised by all the so far described RA-derived IgM RF.53 Moreover, IgM RF have been reported to exhibit enhanced avidity towards agalactosylated forms of IgG,54 which are elevated in RA patients and associated with disease activity.55 A recent glycan profiling of ACPA isolated from RA sera and synovial fluids showed that circulating IgG1 ACPA tended to be more agalactosylated compared with total serum IgG1 and that ACPA IgG1 from the joints were more heavily agalactosylated than their circulating counterparts.56 Interestingly, we noted that some IgG-bound IgM RF is carried over in the IgG fractions obtained from RA serum pools (our unpublished observations). Moreover, we observed that the TNF-α response of macrophages to ACPA-IC formed from IgG fractions obtained following a protocol that minimises such contamination by IgM RF were lower than those obtained using a standard protocol (data not shown). Therefore, it is already clear that in vivo interactions form between ACPA and IgM RF that can contribute to enhance the TNF-α responses of macrophages to ACPA-IC. This indirectly shows the inflammatory potential of RA-derived IgM RF. Within RA joints, ACPA could constitute a privileged target of IgM RF and the studied IC, containing the two autoantibodies suspected to play a critical role in the disease, are probably permanently formed in the inflammatory synovial tissue. Shedding light on their role in the RA pathophysiology, the present study shows that they most probably contribute to the synovitis persistence.
The technical assistance of R. Llobera, M.-F. Isaïa, M.-P. Henry, C. Gohory, E. Parra and L. Ceccato is gratefully acknowledged. We also thank A. Desquesnes (Phenotype Analysis service of the Anexplo Platform in Life Sciences of Toulouse Genopole) for performing multiplex cytokine assays, F. Fortenfant (“Laboratoire d’Immunologie”, Rangueil Hospital, Toulouse) for performing RF assays and F.-E. L’Faqihi-Olive, V. Duplan-Eche and D. Destrade (Cytometry and Cell Sorting service of the Toulouse RIO Imaging platform) for assistance in routine flow cytometry analyses of monocytes and macrophages.
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 supplement
Handling editor Tore K Kvien
LL and FA, and MS and GS contributed equally to this work and are co-first and co-last authors, respectively.
This work was presented in part at the 30th and 31st European Workshop for Rheumatology Research and subsequently appeared in abstract form in Annals of the Rheumatic Diseases (69S2:A5 and 70S2:A38, respectively).
Contributors All authors contributed to conception or design of the study, acquisition, analysis or interpretation of data and writing the article or revising it.
Funding This study was supported by grants from the Toulouse III University, the “CNRS”, the “INSERM”, the “Fondation Arthritis” and the “Société Française de Rhumatologie”.
Competing interests Two authors (MS and GS) are co-inventors of patents that are licensed to companies selling ACPA (CCP) assays. However, the financial relation with the companies selling the assays is with Toulouse University and not directly with the authors. The work was not commissioned by any of these companies.
Patient consent Obtained.
Ethics approval The serum samples collections used in this study have been obtained from informed and consenting patients and have been declared and approved by the ‘Comité de protection des personnes (CPP) Sud Ouest et Outre-Mer II (Toulouse, France) DC2008-463’ in compliance with a French decree (2007–1220) that rules sampling, conservation and preparation of human body products and elements used for scientific purposes. The synovial tissue samples from which synoviocytes were derived also have been obtained in compliance with this decree from informed and consenting patients. Blood samples are anonymously obtained exclusively for research purposes within the frame of an agreement with the French blood agency, ‘Etablissement Français du Sang’.
Provenance and peer review Not commissioned; externally peer reviewed.
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.