Rheumatoid arthritis (RA) was one of the first diseases to be classified as autoimmune, based on the prevalence of an autoantibody, rheumatoid factor, which was first identified in the 1940s. However, the diagnostic value and pathogenic significance of rheumatoid factor have been constrained by its striking nonspecificity for RA, in view of its presence in many other chronic diseases and in large numbers of apparently healthy individuals, especially the elderly. Beginning in the 1960s, other autoantibody specificities were described in RA, primarily in immunofluorescence assays. These were termed “anti–perinuclear factor” or “anti-keratin” Abs, and a subset of these were found to react with fillagrin. These autoantibodies were proposed to be more specific for RA than was rheumatoid factor, but the nature of the epitope recognized was unclear.
In a landmark paper published in 1998 in the Journal of Clinical Investigation, Schellekens et al. (1) from the laboratory of Walther J. van Venrooij in Nijmegen showed that the unconventional amino acid citrulline was an essential component of the determinant recognized by anti-fillagrin autoantibodies in RA. This paper, thus far cited well over 500 times, has opened a new era in the diagnosis and classification of RA and has led to important new insights into RA pathogenesis and the role of posttranslational modifications of proteins in autoantigenicity. Citrullination of a variety of proteins is emerging as an essential component of inflammation in a variety of diseases; thus, both alteration of protein functions by citrullination and the interactions of anti-citrulline autoantibodies with these proteins now command increased attention in efforts to understand the etiology and pathogenesis of RA. Citrullination and autoantibodies to citrulline-containing proteins also provided early biomarkers of preclinical RA that now allow testing of strategies to prevent RA.
What is citrullination?
Citrulline is an amino acid that is not encoded in the genome, but is instead created by posttranslational modification of arginine residues in proteins, catalyzed by the enzyme peptidyl arginine deiminase (PAD), which exists in several isoforms (2). A wide variety of proteins can be citrullinated, including histones, cell surface receptors, and extracellular connective tissue components. Citrullination turns out to be essential for many important biological processes, for example neutrophil NETosis, which depends on citrullination of histones.
What did Schellekens and colleagues show in their 1998 Journal of Clinical Investigation paper?
Schellekens et al. (1) synthesized a variety of peptides corresponding to potential immunogenic epitopes of fillagrin, substituting citrulline for arginine in some of the peptides. Using immunoblotting, they showed that most of the anti-fillagrin reactivity in RA sera depended on the presence of citrulline instead of arginine. They then affinity purified citrulline-reactive autoantibodies from these RA sera and demonstrated that these autoantibodies were reactive in the immunofluorescence assays that had been used to describe “anti–perinuclear factor” and “anti-keratin” specificities. Moreover, they showed that anti-citrulline autoantibodies were present in 76% of RA sera but in <5% of sera from control groups, including patients with other autoimmune diseases (1). Multiple citrullinated peptide epitopes of fillagrin were recognized by the RA sera, foreshadowing subsequent demonstrations that a wide array of citrullinated proteins is recognized by RA-associated autoantibodies.
The impact of anti-citrullinated protein Abs on the diagnosis and management of RA
ACPA and anti-CCP are abbreviations commonly used to refer to anti-citrullinated protein Abs that are measured in widely available clinical laboratory assays. ACPAs can react with a variety of citrullinated proteins, such as vimentin, fibrinogen, α enolase, fillagrin, and type II collagen (3, 4). Most clinical assays currently in use employ cyclic citrullinated peptides as the target Ag, which may lead to a small percentage of false-negative results (5, 6). The level of sensitivity and specificity of anti-citrulline positivity reported in the 1998 Journal of Clinical Investigation paper (1) has held up well during the subsequent 17 years. Occasionally, tests for ACPAs are positive in patients with tuberculosis (7), but the specificity for RA has led to the incorporation of ACPAs into criteria for the classification of RA, which can also be used for early diagnosis of RA (8). ACPA positivity predicts more severe disease as defined by greater destruction of cartilage and bone (9–12), and RA patients are now commonly subdivided into those who are ACPA+ versus ACPA−. ACPA is superior to rheumatoid factor as a predictor of RA severity in the joints, but rheumatoid factor is more tightly associated with extra-articular manifestations such as rheumatoid nodules (13–15). Increased mortality in RA is most prominent in patients with both rheumatoid factor and ACPAs (16). Given all of this information, the presence of ACPAs is viewed as an indication for more aggressive management of patients with RA (17).
Citrullination and ACPAs in the pathogenesis of RA
Citrullination of proteins is widespread in inflamed tissue, for example in synovial tissue in both RA and other arthritides (18). Thus, proinflammatory effects of citrullination that are independent of ACPAs could influence RA and many other diseases. Indeed, in the mouse RA model of collagen-induced arthritis, inhibition of PAD reduces disease severity even though collagen-induced arthritis is not dependent on ACPAs (19). Inhibition of PAD also reduces vascular damage in a murine model of systemic lupus (20).
In another experimental system, it was shown that the chemokine CXCL5, when citrullinated, acquires the ability to bind to CXCR1 and CXCR2, which enables it to attract monocytes and enhance joint inflammation (21). Additionally, citrullination of calreticulin enhances the ability of this molecule to function as a receptor for the RA-associated MHC allele that is termed the shared epitope. This binding interaction is a novel mechanism for the shared epitope to enhance joint inflammation and damage that is distinct from the Ag-presenting function of MHC molecules (22).
Whereas ACPAs contribute to joint inflammation by forming immune complexes and engaging Fc receptors (23), citrullination can also contribute specifically to the pathogenesis of RA through unique mechanisms that depend on ACPAs binding to their citrullinated targets. For example, ACPAs that bind to citrullinated vimentin on the surface of osteoclasts activate the osteoclasts to promote bone resorption (24). This mechanism may explain the observation that osteopenia is present in ACPA+ individuals who do not yet have clinical RA (25). ACPAs also affect NETosis of neutrophils, a process that is enhanced in RA. NETs contain citrullinated Ags, and ACPAs not only bind to NETs but also enhance the rate of NETosis. NETs can directly activate synovial fibroblasts to secrete proinflammatory cytokines (26). Patients with RA can also have autoantibodies to PAD, which, remarkably, activate the PAD enzyme to augment citrullination (27). Taken together, these and other observations reveal a complex and interacting series of mechanisms through which targeted citrullination of proteins enhances inflammatory pathways, and the presence of ACPAs confers unique additional augmentation of proinflammatory and tissue-destructive processes.
ACPAs in the earliest stages of RA: clues to etiology and prevention?
In patients with early undifferentiated arthritis, ACPAs are a powerful predictor of progression to frank RA (28). Moreover, ACPAs are found in serum up to 10 years before the onset of RA (29), at a time when synovitis and citrullination of synovial proteins are not evident (30). In pre-RA, ACPAs already exhibit a pattern of decreased galactosylation that has been found in total IgG in active RA (31). The presence of ACPAs is associated with the shared epitope HLA-DR allele that is the most important genetic risk factor for RA (32), and is also associated with two other established risk factors for RA, smoking and periodontitis (33). These observations have led to hypotheses that the formation of ACPAs provides a link between the risk factors for RA and the pathway that ultimately leads to clinical arthritis. Notably, smoking can lead to formation of citrullinated proteins in the lung, the presence of ACPAs is associated with lung inflammation in RA, and ACPAs can be measured in sputum not only in those with RA, but even in individuals at risk for RA who do not yet have ACPAs in their serum (34–38). Periodontitis, which is bidirectionally associated with RA, is driven by the bacterium Porphyromonas gingivalis, which uniquely among oral flora expresses PAD and thus generates local protein citrullination, leading to formation of ACPAs (33, 39–41). These findings give important momentum to the notion that RA begins in organs and tissues other than the joints, such as the lung or periodontal tissue, with citrullination and ACPA formation being early and critical events in its pathogenesis. Given that ACPAs can be a very early biomarker for pre-RA and are present in some first-degree relatives of RA patients (who are known to have an increased risk for developing RA) (42), trials for prevention of RA can now be designed around the measurement of ACPAs and ACPA-triggered inflammatory pathways.
The discovery of ACPAs by van Venrooij’s group was a landmark event that has changed our approach to the diagnosis, classification, and management of RA, an autoimmune disease that has enormous impact on millions of people. It also has opened up new understanding of the pathogenesis of RA and may be getting us closer to grasping the etiology of RA and strategies for prevention. Moreover, conversion of arginine to citrulline turns out not to be the only posttranslational protein modification that engenders autoimmunity: carbamylation (43) and formation of malondialdehyde–acetaldehyde adducts (44) are other examples of posttranslational modifications that are recognized by recently defined RA-associated autoantibodies. Important questions remain for immunologists regarding how posttranslationally modified proteins are handled and recognized during immune repertoire selection and at peripheral checkpoints for establishment or maintenance of tolerance to these determinants.
Abbreviations used in this article:
anti-citrullinated protein Ab
peptidyl arginine deiminase
The author has no financial conflicts of interest.