Rheumatoid factors: what's new?

Rheumatoid arthritis (RA) is a classic example of an autoimmune disorder, with chronic inflammation of the synovial membrane, and deterioration of cartilage and bone in the affected joints. The resultant pain, loss of function and permanent disability are also associated with increased morbidity and mortality. Although the symmetrical joint deformities are a predominant characteristic, the disease has many systemic features. Since the pivotal studies of Waaler [1], serum rheumatoid factor (RF) estimation has been a basic diagnostic aid, varying in popularity with changes in scientific trends and with considerable controversy over its significance. Yet 60 yr later it is still recommended as a prognostic indicator of disease activity and progression in the adult and juvenile forms of arthritis [2, 3]. Science moves in paradigms and for several decades T-lymphocytes have dominated the literature, but with the devotion of persistent researchers, such as Jo Edwards, the B-lymphocyte has moved again to centre stage in the aetiology and progression of RA. Moreover, a recent Delphi panel of senior rheumatologists concluded that the ‘highest value ever’ for serum RF levels was still the best laboratory indicator of disease severity [4]. Changes in serum RF levels may be used as indicators of disease activity, but they are also proving useful in monitoring responses to therapy, especially biological disease modifiers [5]. More recently, the therapeutic genetically engineered chimeric monoclonal antibody rituximab was found under trial, in an open-label multicentre study, to provide significant improvement of symptoms in RA patients who were non-responders to methotrexate [6]. The target for rituximab is CD20, a cell surface protein that is present on stages of differentiation from pre-B to mature B-lymphocytes, but not stem cells or end-stage immunoglobulin-secreting plasma cells [7]. This therapy therefore selectively depletes the CD20+ B-lymphocyte subpopulation. Rituximab has also been effective in resistant RA when given with haematopoietic stem cell transplantation, providing additional evidence that B-lymphocyte depletion leads to reduced disease activity [8]. Intriguingly, this therapy has worked much better for RF-positive patients. The apoptosis of B-lymphocytes, a therapeutic effect of rituximab, was significant in RF-positive patients compared with those who were RF-negative (Wilcoxon matched pairs, P = 0.0156). This finding was further emphasized in a recent study in which anti-B-lymphocyte therapy failed to have an impact on symptoms in RA patients who were RF-negative [9]. Importantly, the therapy acts on the rapidly dividing and mature B-lymphocyte populations whilst leaving the precursor B-lymphocyte untouched. The occurrence of seronegative RA has frequently proved a stumbling block to the idea that RF is pivotal in the pathogenesis of RA. The evidence that rituximab has significant therapeutic effects only in RF-positive patients adds credence to the role of RF-producing B-lymphocytes in the pathogenesis of RA and would suggest that the mechanism of disease in RF-positive patients differs from that in patients who are RF-negative. This idea that RF-negative RA is different is further supported by genetic studies of a group of RF-negative supposed RA patients who did not have the characteristic shared HLA epitope allele found in RF-positive RA, but instead possessed a separate apoptosis-associated gene (PD-1.3A) [10]. An important antigenic target for RFs is the Cγ2–Cγ3 interface in the Fc region on the immunoglobulin molecule [11], thus making RFs the ultimate autoantibody, since there is no shortage of IgG antigen to form immune complexes. In addition, other potential RF epitopes have been identified via epitope mapping studies [12–14]. The target sites do not seem to have any special autoimmune potential as the main epitopes for autoantibodies appear to be adjacent to, or overlap with, those recognized by antibodies from other species when immunized with human IgG. For instance, epitopes defined by murine anti-IgG antibodies bind at or adjacent to RF antigenic hot spots [15, 16] (Fig. 1). Molecular models of the Fc region of IgG highlighting the antigenic regions (dark residues) identified by (a) monoclonal antibodies raised by deliberate immunization with IgG [15, 16] and (b) polyclonal and monoclonal rheumatoid factors [12–14]. Certainly, monomeric IgG is a poor activator of RF+ B-lymphocytes compared with immune-complexed IgG: the structure of these immune complexes appears to be significant, as the antibodies need to be bound to strictly ordered antigens rather than irregularly arranged, to promote efficient induction of RFs [17]. A question to be asked: ‘Is the target for RF the normal immunoglobulin molecule or is it an altered molecule that is the important inducer and binder of RF?’ Enzyme modification of antigen has already been established as an important step in coeliac disease—tissue transglutaminase enzyme modification of gliadin is a vital part of the autoimmune process [18]. Perhaps enzyme modification of self antigen may be a significant step in the pathogenesis of other autoimmune diseases. In the case of RA, the detection of agalactosyl immunoglobulins is a characteristic feature. The absent galactose within the Fc region of IgG allows access to RF-binding epitopes that are either exposed, or revealed as a consequence of conformational changes [19]. Importantly, antibodies to these agalactosyl IgG molecules are found to be raised in RA and their estimation is considered a more sensitive indicator of disease activity than conventional RFs [20, 21]. This mirrors the earlier findings of Soltys et al. [22], who found that the monoclonal RFs that bound better to agalactosyl IgG composed the high-affinity set of RFs in RA. Other RF-binding epitopes have been shown to be independent of the oligosaccharides [22]. Proteins may be modified by the binding of glucose [e.g. in diabetes, where haemoglobin is glycated (HbA1c) and HbA1c levels are used as an indicator of glycaemic control]. Likewise, IgG also may be modified, and autoantibodies against advanced glycated end (AGE)-damaged IgG have also been detected and shown to be highly significantly associated with RFs (P<0.0001) and with swollen joint counts (P<0.01) [23]. The microenvironment of the synovial membrane becomes an ‘ectopic’ secondary lymphoid organ, a site of isotype switching and somatic hypermutation for the RF-positive B-lymphocytes [24]. Williams et al. [25] have shown in the autoimmune murine model that somatic hypermutation in autoimmune response happens outside the conventional germinal centres. B-lymphocytes that mutate to autoreactive specificities will have escaped the normal tolerance self-censoring mechanisms that are present in germinal centres. Recent evidence has found that the RFs in RA patients with inflamed joints RA are produced by terminally differentiated CD20−CD38+ plasma cells in the synovial fluid that are generated by germinal centre-like structures within the synovium [26]. Thus, when giving anti-CD20 therapy (rituximab), which targets the B-lymphocytes at an earlier developmental stage, one would expect a lag phase between the start of treatment and the onset of disease remission, since the CD20−CD38+ plasma cells are not eliminated by the treatment, but die eventually of old age. Naturally occurring RFs are detectable in healthy individuals, thus leading to doubt regarding the significance of RFs in disease. These natural RFs are generally polyreactive IgM antibodies of low affinity produced by CD5+ B-lymphocytes. They are considered to be part of the normal response, e.g. to lipopolysaccharide or Epstein–Barr virus (EBV) [27, 28], and contribute to the immune complex clearance mechanism by the formation of larger immune complexes that are phagocytosed [29]. The fact that in healthy subjects the RF-positive B lymphocytes coexist with IgG antigen in a non-autoimmune disease state would suggest the existence of a tolerance mechanism. A possibility could be the inhibition of class-switching, so that IgM/IgD RF-positive cells do not switch to being the IgG RF-positive cells that produce high-affinity and pathogenic antibodies via somatic hypermutation [30]. In order to induce B-lymphocytes to switch from producing IgM RFs to secreting IgG RFs or IgA RFs, T-cell help is crucial. It is noteworthy that Lang et al. [31] have shown that the response to IgG is HLA-DR-restricted and mediated by T-helper cells (CD4+) that are stimulated by IgG, for the initiation and propagation of the immune mechanism of RA [31]. However, it is surprising that this has not been followed up and the experiments repeated to confirm this hypothesis. If the autoimmune mechanism is both a T- and a B-lymphocyte response to IgG, this is highly significant for the development of tolerance-inducing therapies. It has also been proposed that autoreactive T-lymphocytes are stimulated by the presence of IgG aggregates within the joints. Recent evidence has indicated the presence of T-lymphocyte responses to IgG molecules that have been attacked by oxygen free radicals, a pathological mechanism prevalent within the diseased joint [32]. Rheumatoid factors may be considered somewhat analogous in function to the soluble form of Fc receptors (FcγRs) in specifically cross-linking with immune complexes. FcγRs are normally expressed on selective cell types and, once bound to their targets, the resultant leucocyte activation induces the effector functions of cytokine release, antibody-dependent cytotoxicity and phagocytosis with respiratory bursts. Studies in murine models of FcγR deficiency with collagen-induced arthritis indicate a role for these Fc receptors in inflammation and joint destruction [33]. Studies of allelic forms of FcγR on NK cells in patients with RA and healthy controls have suggested that cellular activation mediated by binding to these receptors could be influential in the disease process of RA [34]. The function of the soluble forms of FcγR remains unclear, although Wines et al. [35] have demonstrated inhibition of RF binding to IgG immune complexes by the soluble forms of FcγRIIa. As to whether the physiological concentrations of soluble FcγR in the synovium are sufficiently high to influence RF binding has yet to be confirmed. Surprisingly, all the IgM RFs that have been examined so far have shown unusual molecular structures compared with normal IgM molecules. Analysis of comparative crystallographic data of IgM and IgM RF has exposed differences in the shape, where the IgM RF has an asymmetrical F(ab)2 region signifying structural differences distinctive from IgM [36]. Likewise, by extrapolating the crystallographic data of the monoclonal human rheumatoid factor, Sutton et al. [37] proposed that some RFs are able to simultaneously bind both antigen (environmental or other autoantigens) and the Fc region of IgG via the conventional antigen-binding site and an adjacent binding site. Accordingly, this could lead to high levels of RF synthesis wherever these antibodies are stimulated by an external antigen, and such evidence would point to a possible infectious or environmental trigger for RF-positive RA. A reasonable predictor of RA risk may be estimated from an analysis of serological and HLA genotype screening. Family studies would point to a genetic influence, but environmental triggers are also fundamental. Toll-like receptors are key for the induction of innate response to many environmental pathogens. There is increasing evidence to suggest that the inappropriate stimulation of such receptors is important in autoimmunity (see recent review of Rifkin et al. [38]). Moreover, Toll-like receptors are expressed in the synovium of patients with clinically active RA [39]. The crucial importance of Toll-like receptors in RF production has been demonstrated using two forms of immune complex. Rifkin et al. [40] showed that in autoimmune mice IgG2a immune complexes were able to stimulate RF production by RF B-lymphocytes. Subsequent work has revealed that such immune complexes contain a nucleosome antigen, and it is this antigen that was key because IgG2a immune complexes made with other antigens had no effect [41]. Detailed analysis of this system has showed that RF+ B-lymphocytes require two signals to be triggered—one via the B-lymphocyte receptor (i.e. for the IgG as an antigen) and the other via the Toll-like receptors (stimulated by the nucleosome antigen). Interestingly, a similar role for Toll-like receptors in stimulating RF+ B-lymphocytes has also recently been found in bacterial infections [42] (Toll-like receptors have also being shown to be immunomodulators and pathogenic costimulators of autoreactive B-lymphocytes [43]). This is especially relevant for chronic disease, given the finding of Lanzavecchia's group, who showed that the immunology memory response could be maintained by activating B-lymphocytes via Toll-like receptors alone [44]. Several viral agents, e.g. EBV, parvovirus B19 and more latterly human endogenous retroviruses (HERVs), have been implicated in the aetiology of chronic autoimmune disorders [45, 46] via excess immune complex production. In particular, HERVs may contribute to the pathogenesis of disease, possibly by acting in concert with other infectious agents [47], or as potential transducing agents that enable other environmental/hormonal triggers to alter the immune system [48]. EBV is a classic example that infects and stimulates B-lymphocytes and may also impact on T-helper cells by transactivating an endogenous viral superantigen [49]. It is noteworthy that RFs targeting the membrane-bound IgG3 have also been shown to induce both EBV and B-lymphocyte activation [50]. Patients with active RA have a lower than normal cellular immune response to an EBV protein that is important for its replication, thereby allowing the virus to replicate, and results in a high EBV load [51], although therapy could also have an impact on this. The virus can continue to replicate even in the face of high antibody responses. This leads to a high antigen load, which, if accompanied by a significantly high antibody response, could easily result in a high level of immune complex formation analogous to an autoimmune response with the resulting stimulation of RF [52]. Mixed cryoglobulinaemia is now known to be associated with hepatitis C infection, and Sansonno et al. [53] have shown that these cryoglobulins are formed of hepatitis C virus (HCV) core protein plus IgG antibody, which are then cryoprecipitated by IgM RF. In such patients, B-lymphocyte depletion with rituximab therapy has led to clinical remission of hepatitis C-associated type II mixed cryoglobulinaemia with glomerulonephritis, and signs of systemic vasculitis, where the high levels of RF contribute to the pathology [54]. There are rudimentary data on certain risk factors. Environmental pollutants, silica and blood transfusion have been implicated, but long-term smoking is recognized as a serious factor in RA patients who are RF-positive. It is proposed that high levels of pollutants like smoking and fossil fuels increase the risk of respiratory tract infections, and infectious agents become antigenic targets that trigger autoimmune responses. Long-term smoking, i.e. for more than 20 yr, carries a significant risk of RF-positive RA and the ability to generate detectable RF levels in advance of the clinical presentation of RA. Alternatively, smoking cessation could also reduce the risk of RA [55]. An example of gene–environment interactions is also shown in the significant risk of RA with the shared epitope on HLA-DR that influences at-risk male and female patients who are also smokers [56]. Likewise, particularly in female RA patients with a polymorphism of the enzyme glutathione S-transferase, Mattey et al. [57] have hypothesized that a lack of the enzyme that detoxifies smoke-derived by-products has the potential to damage IgG in smokers and could encourage RF production. Coffee consumption is another suggested risk factor for RA, and its mechanism has been linked with the production of RF [58]. In the Iowa Women's Health Study, the link with RA was with decaffeinated coffee [59]. Since tobacco and significant coffee importation hailed from the New World, could this provide the basis for the hypothesis that the origins of RA were in North America [60]? The increased incidence of RA reported in several American Indian and Alaskan populations is well documented and would indicate a strong genetic influence [61]. Population-based studies of the Pima Indians in Arizona have exposed the high incidence of raised RF levels and the subsequent development of RA. However, the decline revealed in recent secular and birth cohort data would implicate environmental influences in early life as significant predictors of the lifelong probability of being RF-positive [62]. Certain HLA-DR4 alleles are associated with RA in white populations in Western Europe and the USA, whilst different alleles are associated with RA in Jewish and Japanese population studies, suggesting that there are interactions of HLA with other genes [63]. Wu et al. [64] suggested possible genetic differences between patients with early-onset and those with late-onset RA. They showed a significant association between novel RANKL (receptor activator of NF-κB ligand; a significant ligand that must be controlled in order to maintain bone homeostasis and prevent bone resorption) genetic polymorphisms and early-onset RA (18–20 yr younger) and patients who were positive for the shared epitope containing DRB1*04 alleles (P = 0.0004) have the alleles with to their or effects in RA RF is detected in other autoimmune such as and systemic In of are with a level of IgA RFs in male compared with female patients of RF-positive B-lymphocyte is to be in the aetiology of the disorders that in of patients with However, it be that patients may be for other e.g. antibodies and antibodies There is a of in the early onset of RA which, if it is identified and if treatment is could lead to a better the early and that in the stage of RA would in the of therapeutic Moreover, subjects with raised RF levels who are are at significantly increased risk of RA are for serum RF but have been A modification of the when compared with at to be more sensitive so could lead to earlier detection of a of by the of RFs, result and lead to in Likewise, a raised B level has an early disease association in RA patients who are RF-positive A diagnostic indicator with high for RA is antibodies by and although their presence not other autoimmune such as The antibodies appear in response to an epitope modified by a controlled which to have been found in synovial sites in more than of RA compared with of controls It is that only one antigen is the immune target but it would be to the of this enzyme on IgG and subsequent binding of RF. The Delphi panel from their of indicators of RA disease severity [4]. have suggested that in patients who have had for more than both and RF are predictors of patients to RA some suggest that antibodies are more sensitive than RFs in the of RA and juvenile arthritis Yet there is evidence as to whether a in and RF levels is detected monoclonal antibody some studies in both have found a in RF levels with levels and yet have shown that further autoantibodies may be e.g. and antibodies It is suggested that antibodies are particularly in the of RA, but RF are better predictors of disease severity a indicator is the useful predictor of joint damage is a of i.e. IgA IgM protein and cartilage protein of IgA RFs and antibodies may be predictors of RA in the of clinical signs and symptoms by and may be useful in the disease of RA patients who are RF-negative or where the may be In and RF are antibody different on the mechanisms of disease that to result in RA has been yet again to The results in which effective responses in RF-positive patients with rituximab against pre-B to mature or monoclonal antibody strong to RF-positive RA being a separate disease from RF-negative RA. this in is being on the role of RF-producing B-lymphocytes in the pathological Accordingly, by to these it would be a to the development of an effective and therapy rather than a on B-lymphocytes. The B-lymphocyte has to centre stage in the aetiology and progression of RA, and RF with are still in the and of the different of RA patients. The have no of for Health and of of in of of and for and of