Use and significance of anti-CCP autoantibodies in rheumatoid arthritis

One of the most common autoimmune diseases is rheumatoid arthritis (RA), affecting 0.5–1% of the population. This systemic disease is marked by chronic inflammation of synovial joints, which leads to destruction of cartilage and bone and eventually to disability of the patient [1]. Though not directly life-threatening, RA severely affects the quality of life of a patient and also has major economic consequences for society. Therefore, every attempt should be made to prevent the erosive processes to occur. Currently, the classification of RA relies mainly on the criteria described by the American College of Rheumatology (ACR) [2]. These criteria, originally formulated 50 yr ago and last adjusted in 1987, are based mainly on clinical parameters. Since these parameters are often only sufficiently fulfilled when the damaging effects of the inflammatory process are already in progress, this set of criteria is not very suitable for the early diagnosis of RA [3]. The key to early recognition of autoimmunity lies within the humoral immune system. Since blood samples are taken from clinic-visiting (pre-)patients, screening for serological RA-indicators can be performed quite easily. In the ACR criteria for RA one serological marker is included: rheumatoid factor (RF). The RF autoantibody system, directed against the Fc part of immunoglobulin (Ig) G molecules, has had a central role in the diagnosis and prognosis of RA during recent decades [4] because RF can be detected in the majority of RA patients. However, it is becoming more and more clear that the presence of RF is not restricted to patients with RA, but that it can also be detected in subsets of patients suffering from other diseases and even in a percentage of healthy (especially elderly) individuals [5]. The resulting lack of specificity for RA can lead to confusion and unwanted treatment. The shortcomings of the RF test have kept the search for more specific RA markers alive. Most autoantibody systems described during recent decades have failed to mature into mainstream tests for RA because of low sensitivity, lack of specificity or technical inconvenience, as reviewed previously [6–8]. The only antibody system that combines good sensitivity with superior specificity for RA is that targeting citrullinated epitopes. This review will focus mainly on the diagnostic potential of the second-generation anti-cyclic citrullinated peptide (CCP) test (CCP2) for RA. Next, we will address the prognostic ability of the anti-CCP test. The presence of these antibodies early in disease development opens a window of opportunity for early custom-tailored treatment of RA. Finally, we will review the effect of disease treatment on anti-CCP levels and briefly go into the putative functional role of these antibodies in the chronic aspects of RA. The autoantibody system most specific for RA known to date is that directed to citrullinated antigens. The citrulline moiety, which is the essential part of the antigenic determinant in these antigens, is post-translationally generated by peptidylarginine deiminases (PAD; EC 3.5.3.15) [9]. RA autoantibodies against citrullinated antigens have been detected and used for diagnostic purposes for many decades via the well-known antiperinuclear factor (APF) [10] and antikeratin antibody (AKA) tests [11]. Supported by the fact that APF and AKA share many features, as reviewed previously [12], and are reactive with native filaggrin, these autoantibodies are now designated as antifilaggrin antibodies (AFA) [13]. Using the laborious and inconvenient immunofluorescence assay, roughly 50% of RA sera can be scored AFA-positive. A key finding was the discovery that the reactivity of these AFA was completely dependent on the presence of citrulline residues (present in mature filaggrin but not in profilaggrin) [14, 15]. Since then, two approaches for detecting autoantibodies to citrullinated epitopes have been taken: a protein-based and a peptide-based approach. Screening for citrulline-specific RA reactivity has been performed with several proteins, including both purified naturally occurring citrullinated proteins and in vitro-citrullinated proteins. For these purposes, mainly filaggrin, fibrinogen and myelin basic protein have been used. Although most of these proteins are arginine-rich, there is obviously a limit to the number of citrullinated epitopes associated with a certain protein. A complication of the use of natural antigens is that it is difficult to obtain reasonable amounts in sufficient purity in a reproducible way. Batch-to-batch variation also compromises standardization when in vitro citrullination is used to generate the antigen. Insufficient purity of the antigen lowers the specificity of the test because reactivities directed to other components (e.g. the PAD enzyme, the non-citrullinated part of the antigen, other contaminants) may be detected as well. The use of proper (non-citrullinated) controls is therefore very important, as indeed has been noted by Vittecoq and colleagues [16]. Interestingly, autoantigenicity of the PAD protein has recently been described, but this reactivity is not specific for RA [17]. However, despite these limitations several studies have successfully improved the sensitivity of AFA detection while maintaining specificity levels [18]. Using various technologies, assays with various sources of filaggrin have been developed, allowing sensitivities up to 60% [16, 19, 20]. Using citrullinated fibrinogen, Nielen and colleagues [21] reported a similar sensitivity (56%) for a cohort of early arthritis patients. The use of synthetic citrullinated peptides for anti-citrullinated protein antibody detection can overcome many of the complicating factors of the protein-based approach. Synthetic peptide production and purification is cheap and easily standardized, and via peptides one can synthesize an unlimited pool of defined epitopes. Furthermore, citrulline residues can be incorporated during synthesis of the peptides, leading to a homogeneous preparation of citrullinated molecules. A major breakthrough came with the development of an enzyme-linked immunosorbent assay (ELISA) that used filaggrin-based citrullinated peptides [14]. The reactivity of RA sera was completely dependent on the citrulline residue(s) present, since the same peptides in which the citrulline was replaced by another amino acid were not antigenic. The variation in reactivity patterns against different citrullinated peptides clearly showed that the anti-CCP response in RA is polyclonal. When a filaggrin-based cyclic peptide (cyclization increased the sensitivity) was applied in the first generation anti-CCP (CCP1) test, a sensitivity of 68% was obtained with very high specificity for RA (98%) [22]. Though better than the protein-based methods, the sensitivity was not as high as that of the routinely used RF test. Because filaggrin is not present in the synovium, dedicated libraries of citrulline-containing peptides were screened with RA sera to select for superior epitopes. This culminated in the CCP2 test, which displays a sensitivity of up to 80% without loss of specificity. Diagnostic markers of disease ideally fulfil three requirements: (i) good sensitivity, to detect a high percentage of patients; (iii) good specificity, to limit false-positive results as much as possible; and (iii) early presence, to facilitate early diagnosis. Over the last decade many studies have investigated the diagnostic performance of the anti-CCP test. Those using the CCP1 test have been reviewed by van Boekel and colleagues [6]. Increasing data on the improved [23] second-generation anti-CCP test show that the CCP2 test result is a good diagnostic parameter for (early) RA. The first large cohort studies of anti-CCP2 as a diagnostic marker showed that anti-CCP2 combines RF-like sensitivity with almost absolute specificity for RA [5, 24, 25]. These multicentre studies showed that anti-CCP2 antibodies, just like RF, are present in about 80% of established RA patients. In the healthy control group and the non-RA disease controls, the CCP2 test was only positive in maximally 1 and 5%, respectively. The corresponding percentages of the RF (over 10% of healthy controls and more than 20% of disease controls) were markedly higher [5, 25]. Several recent independent studies confirmed these sensitivity/specificity data for CCP2 [26–30]. Using a cohort of 549 RA patients, the study of Suzuki and colleagues [26] clearly showed higher discriminative ability for the CCP2 test than for RF test. The reported sensitivity of 65% (at 96% specificity) by Dubucquoi and colleagues [27] was found in a group of RA patients that included several patients with recent onset of disease. The sensitivity in their established RA patient group was 77%. Overall, the observed intercohort variations in these studies might be explained by differences in the characteristics of the patients that were included. One study reported a somewhat lower specificity for RA of about 90% with the CCP2 test [31], but this could have been caused by the fact that the juvenile RA (JRA) group included in this study might have contained several adult RA patients with an early onset of disease, as discussed previously [32]. The JRA patients had a high average age (31 yr) and often longstanding disease (21 yr on average), mostly resulting in erosions (87%). A striking example of the diagnostic performance of the CCP2 test is the recent study by de Rycke and colleagues [28]. Setting the specificity values for RA at 98.5% resulted in a sensitivity for RF of only 12.8% compared with 73.7% for anti-CCP2. Taken together, these studies show that the anti-CCP2 test at least equals the RF level for sensitivity, but combines this with far better specificity. The fact that around 40% of RF-seronegative patients appear to be anti-CCP-positive substantiates the additional diagnostic potential of CCP [29, 33]. The anti-CCP test also enables clinicians to effectively distinguish RA patients from other arthritic diseases in cases where the RF is not always discriminative. One of the first examples of such a role in differential diagnosis comes from patients with erosive systemic lupus erythematosus (SLE). Mediwake and colleagues [34] showed that anti-CCP (in this case CCP1) can be used to distinguish RA patients from SLE patients who present with erosive polyarthritis, which is often accompanied by RF seropositivity. Another disease that can readily be misdiagnosed because it often reveals RA-like arthropathies is chronic hepatitis C virus (HCV) infection, which is often accompanied by a positive RF. Wener et al. [35] reported a good discriminative ability of anti-CCP2 over RF in a group of randomly selected HCV patients (44% RF+, none CCP2+). These data were confirmed by Bombardieri and colleagues [36]. Whereas RF was detected in 15% of the HCV patients (37% in case of joint involvement), no anti-CCP2 positivity was seen in these patients. The value of anti-CCP for use in differential diagnosis was also shown when comparing RA patients with polymyalgia rheumatica patients [37]. Taken together, these data clearly outline the diagnostic strength of the anti-CCP test for RA. Recently, some papers have reported anti-CCP positivity in arthritic diseases that share certain features with RA [38–40]. In palindromic arthritis (PR), a percentage of anti-CCP positivity (56%) similar to that of patients with early RA (55%) was found [38]. PR is a relapsing from of arthritis that shares many features of RA. In addition to the observed anti-CCP reactivity, many of these sera are also RF-positive (30–60%). Moreover, as is the case for RA, PR is associated with the presence of the MHC shared epitope (odds ratio of 2.9) [41]. The current assumption is that PR is an abortive form of RA that in a percentage of patients may develop into RA. The presence of RF in PR has already been shown to be associated with the development of chronic disease [42]. Regarding the prognostic significance of anti-CCP positivity in PR and the possible development of RA, additional studies with larger cohorts and longer follow-up are needed. In patients with primary Sjögren's syndrome, anti-CCP reactivity was detectable in 8% of the 134 patients tested [39]. For another RA-like disease, psoriatic arthritis, the group of de Rycke [40] found anti-CCP positivity in 8% of the 192 patients tested. This reactivity, confirmed with an additional test using home-made citrullinated peptides, was higher than could be ascribed to simple overlap with RA. These studies make clear that, especially in RA-like arthritic disease entities, further clinical follow-up is needed to establish whether the anti-CCP positivity predicts the development of RA. Because RA patients at their first visit to the clinician often do not fulfil the criteria for the diagnosis/classification of RA, an early detectable, highly predictive marker would greatly help the clinician in reaching a diagnosis. Obviously, the sensitivity and specificity of such a marker should be as high as possible. Recently, two studies, both making use of dated samples from RA patients who were former blood donors, reported the presence of anti-CCP antibodies prior to the appearance of the first clinical symptoms of arthritis [43, 44]. Samples from 72 blood donors were characterized by Nielen and colleagues [43] for positivity of IgM-RF or anti-CCP1. Both serological markers were detectable long before the disease became clinically overt. In some patients, anti-CCP1 was found up to 14 yr prior to the first clinical symptoms of disease. The same was found for 39% of the patients 5.3 yr (median) before the first visit to the clinic. IgM-RF was also found in predisease samples, but not as far back (up to 10 yr) and in a smaller percentage of patients (23% at a median of 3.3 yr). The second study, with a similar set-up, detected anti-CCP2 and RF up to 10 yr before clinical disease in predisease blood samples of 83 RA patients. Anti-CCP2 positivity gradually increased in the years prior to the first clinical symptoms and had reached a positivity of 70% at the time the patients visited the rheumatology clinic for the first time. The sensitivity for detecting RF autoantibodies in these predisease samples was slightly less than for anti-CCP2 [44]. From these studies it is clear that the production of anti-CCP and RF autoantibodies is an early process in RA development, and that their presence is predictive for the development of this disease. Recent data from several longitudinal studies confirm the predictive ability of anti-CCP2 for RA development [16, 45, 46]. van Gaalen and colleagues [45] used serological markers to predict which of the patients attending an early arthritis clinic, who were classified as undifferentiated arthritis (UA), would progress to RA within the next few years. Follow-up data of 318 patients with UA clearly showed the predictive potential of anti-CCP autoantibodies for the development of RA. After 1 yr of follow-up, 75% of the UA patients who were anti-CCP2-positive at baseline had already progressed to RA. This percentage increased to 93% after 3 yr (odds ratio 38). Of the UA group who were anti-CCP2 negative at baseline, only 25% were classified RA after 3 yr. Similar results were reported by Vittecoq and colleagues [16]. In their cohort of 314 early arthritis patients, 90% of the anti-CCP2-positive patients were classified as RA patients at the 1-yr follow-up. Thus, the combined early presence and predictive ability of anti-CCP may find an important clinical application in the design of treatment strategies. Various studies have addressed the prognostic value of anti-CCP antibodies. Though the anti-CCP test has only recently become widely available, several studies have already demonstrated its ability to predict the erosiveness of developing RA. Most of these studies used the CCP1 test and the results of these have been discussed previously [32, 47]. An increasing number of studies with anti-CCP2 confirm the prognostic potential. First, in a cohort of 379 early RA patients, Forslind and colleagues [48] showed that anti-CCP2 positivity at baseline (55%) predicts radiological damage and progression at 2 yr follow-up. Similar results were obtained by Kastbom and colleagues [49]. In their study, CCP2 positivity at baseline predicted disease activity at the 3-yr follow-up. In addition, Rönnelid et al. [50] showed that anti-CCP2-positive early RA patients developed worse clinical disease and greater radiological damage within a few years in comparison with patients. The prognostic ability of the anti-CCP test is often by other disease parameters. to the presence of the shared epitope MHC molecules, this was reported by independent studies of et al. van Gaalen et al. de Rycke et al. and et al. The study by de Rycke and colleagues also showed that anti-CCP2 RF, not with the presence of [28]. and colleagues reported that the of anti-CCP2 and RF positivity was the for the development of RA in a patient group with very early Anti-CCP2 levels are also associated with the development of bone erosions in RA the other et al. reported that there was no the of anti-CCP antibodies and of early RA. The from these studies is that a positive anti-CCP test to predict the development of erosive RA and that this predictive value that of RF. with additional data (e.g. the presence of shared with this may lead to a very high that erosive RA is the most for the treatment of RA at present are targeting the like can disease activity and progression of RA However, such as an increased of and the of autoantibodies (e.g. antibodies or have also been reported The of autoantibody levels during the of treatment may some the of the treatment and the role of these antibodies in the disease In the several papers have reported effects on including in patients with with or without in and colleagues autoantibody during treatment with and Though treatment resulted in a in disease activity no in the percentages of patients who were positive for anti-CCP2 or IgM-RF were of RF, on the other were of These results were confirmed by several other studies, follow-up the studies In the study of and colleagues anti-CCP by the anti-CCP1 not during of follow-up, 60% of the patients had a clinical IgM-RF levels were somewhat in the first of follow-up of and and colleagues and de Rycke and colleagues also observed a in RF levels but not of In to the studies, and colleagues find a but in anti-CCP levels at of treatment in the patients with clinical This also observed for RF, was dependent on since in anti-CCP levels were not observed after treatment with can result in a in both anti-CCP and RF in about 50% of patients In this last study, treatment with a of RF. Recently, Rönnelid and colleagues [50] showed that treatment with but not other resulted in a in anti-CCP but this only in the first of follow-up and not with clinical of longitudinal studies on anti-CCP during RA treatment of longitudinal studies on anti-CCP during RA treatment From these studies, it is clear that the anti-CCP autoantibody system is different from the RF system with to response to treatment. the functional of anti-CCP in the progression of RA other approaches at this humoral Because anti-CCP autoantibodies are specific for RA, one can about the of the disease. that we to more about protein citrullination and the PAD for the of peptidylarginine to The of PAD show a [9]. has been shown that PAD are present in the and that their activity is at the and levels In addition, these high about higher than present in the of a Interestingly, citrullination in during when many by or one can detect citrullinated proteins, both in of inflammation and in the of RA and non-RA patients. the presence of citrullinated proteins in most cases not lead to the generation of anti-citrullinated protein antibodies This might be to the of the has been known for some time that there is a RA and certain and These the shared epitope and colleagues showed that the of to citrulline increased the of a peptide for to and can lead to of in These results that the production of anti-citrullinated protein antibodies is dependent on the presence of certain for RA. The fact that the observed ratio of for the of anti-CCP and is over two the value of the of the corresponding also that these factors are associated recently it was shown that patients who were for and for anti-CCP antibody with an increased of joint destruction compared with patients only one or none of these these results that the production of these antibodies the inflammation in RA. These data have to a simple for the development of RA, in 1 and discussed previously In individuals with a for RA, an inflammation of the in leads to of inflammatory that PAD After these will via and will be by However, when there is or a in the system, some may become citrullinated proteins and PAD The PAD can synovial proteins, such as In of individuals this will be the of the In the of individuals who are to present of the citrullinated proteins to the via certain MHC molecules, a response to citrullinated antigens will be resulting in the of immune which will the inflammatory process by of and will the synovium, will be and a of The of this process over a number of eventually accompanied by or that will lead to a chronic inflammation which can develop into the disease we as rheumatoid In this anti-CCP to the of joint inflammation and to the and of RA. of RA for RA, reviewed by and are mainly and to the disease. these are to the of and erosive obtained over recent years that early in the disease have the potential the of it is to RA patients before joint damage The fact that the CCP2 test is now widely will up these studies, as shown by the of data on CCP in the about the of citrullinated the application of screening and use in assays will further the diagnostic potential of these a from the for and are and of as a to and