You are here

  • Home
  • Archive
  • Volume 73, Issue 6
  • The fibrin-derived citrullinated peptide β60–74Cit60,72,74 bears the major ACPA epitope recognised by the rheumatoid arthritis-specific anticitrullinated fibrinogen autoantibodies and anti-CCP2 antibodies

Article Text

Article menu
Basic and translational research
Extended report
The fibrin-derived citrullinated peptide β60–74Cit60,72,74 bears the major ACPA epitope recognised by the rheumatoid arthritis-specific anticitrullinated fibrinogen autoantibodies and anti-CCP2 antibodies
  1. M Cornillet1,
  2. M Sebbag1,
  3. E Verrouil2,3,
  4. A Magyar4,
  5. F Babos4,
  6. A Ruyssen-Witrand2,
  7. F Hudecz4,
  8. A Cantagrel2,
  9. G Serre1,3,
  10. L Nogueira1,3
  1. 1‘Epidermis Differentiation and Rheumatoid Autoimmunity’ Laboratory, UMR CNRS 5165, INSERM U 1056, Toulouse III University, Toulouse, France
  2. 2Rheumatology Centre, Toulouse University Hospital, Toulouse, France
  3. 3Cell Biology and Cytology Laboratory, Toulouse University Hospital, Toulouse, France
  4. 4Research Group of Peptide Chemistry, Department of Organic Chemistry, Hungarian Academy of Sciences, Eötvös Loránd University, Budapest, Hungary
  1. Correspondence to Dr Leonor Nogueira, ‘Epidermis Differentiation and Rheumatoid Autoimmunity’ Laboratory, UMR 5165 CNRS—U 1056 INSERM—Toulouse III University and Cell Biology and Cytology Laboratory, Toulouse University Hospital, Toulouse, 31059, France; Nogueira.l{at}chu-toulouse.fr

Abstract

Objectives To evaluate the proportions of rheumatoid arthritis (RA) sera containing anticitrullinated proteins autoantibodies (ACPA) reactive to α36–50Cit38,42 and/or β60–74Cit60,72,74, two peptides identified as bearing the immunodominant epitopes of their major target, citrullinated fibrin. To analyse the relationships of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies with autoantibodies reactive to the complete citrullinated human fibrinogen molecule (AhFibA) and with anti-CCP2 antibodies.

Methods  617 sera from 181 patients with established RA and 436 with non-RA rheumatic diseases were tested by ELISA for AhFibA, anti-CCP2, anti-α36–50Cit38,42, anti-β60–74Cit60,72,74 autoantibodies, and by nephelometry for rheumatoid factor (RF). Diagnostic indexes, correlations and concordances between tests were analysed. Crossreactivity of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies was assessed in competition experiments.

Results At a diagnostic specificity of 95%, the diagnostic sensitivity of AhFibA (83%) was significantly higher than that of all other tests. The diagnostic sensitivity of anti-β60–74Cit60,72,74 (71%) was significantly higher than that of anti-α36–50Cit38,42 autoantibodies (51%) but similar to that of anti-CCP2 (74%). Titres of RF, anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies were weakly correlated with each other, whereas titres of anti-β60–74Cit60,72,74 were strongly correlated with those of AhFibA (r=0.633) and anti-CCP2 (r=0.634). Anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 mainly corresponded to two non-crossreactive subfamilies of ACPA. More than 90% of AhFibA-positive or anti-CCP2-positive sera recognised the α36–50Cit38,42 and/or the β60–74Cit60,72,74 peptide.

Conclusions Autoantibodies reactive to α36–50Cit38,42 and β60–74Cit60,72,74 form two distinct, non-overlapping subfamilies of ACPA that, together, cover practically all the ACPA reactivity to citrullinated fibrinogen and to CCP2 antigens. In established RA, anti-β60–74Cit60,72,74 autoantibodies show diagnostic indexes similar to those of anti-CCP2.

  • Ant-CCP
  • Autoantibodies
  • B cells
  • Rheumatoid Arthritis
  • Autoimmunity

https://doi.org/10.1136/annrheumdis-2012-202868

Statistics from Altmetric.com

    Request Permissions

    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.

    Introduction

    In rheumatoid arthritis (RA), the most common autoimmune disease, several factors, such as female gender (sex ratio of 1/3), human leukocyte antigen polymorphism and smoking, are accepted as predisposing to the disease but, to date, the disease pathophysiology remains unclear.1 ,2 However, one biological feature that is most probably of prime importance, is the presence of two families of autoantibodies, rheumatoid factor (RF) and anticitrullinated proteins antibodies (ACPA), in the RA sera.2

    RF corresponds to autoantibodies directed against the Fc portion of Immunoglobulin G (IgGs). IgM-RF, the most commonly detected isotype, is found in 70–80% of RA sera, but with a weak diagnostic specificity of 80–90%.3 By contrast, ACPA, IgG directed against epitopes generated on the target antigens by post-translational deimination (citrullination) of arginyl residues,4 ,5 are the most specific serological markers of RA (diagnostic specificity >98%).2 ,3 ,6 ACPAs are present in up to 80% of RA sera and appear early in the disease course, even years before the onset of symptoms.7 Moreover, they are associated with more severe disease and predict the progression of undifferentiated arthritis to RA.8 ,9 Therefore, they constitute a very good diagnostic biomarker for RA and were recently included in the American College of Rheumatology (ACR)/The European League Against Rheumatism classification criteria beside RF.10 ACPAs are most commonly detected with the second-generation anti-CCP2 immunoassays, the immunosorbent of which is described as a mixture of cyclic citrullinated synthetic peptides that do not match any human protein sequence, while several proteins have been shown to be in vivo targets of ACPA.11–13 Among them, citrullinated forms of the α-chains and β-chains of fibrin have been identified as major targets in the RA synovial tissue.11 Accordingly, the ELISA detection of autoantibodies to in vitro citrullinated human fibrinogen (AhFibA-ELISA) has proved to have high diagnostic performance for RA.9 ,14

    RF and ACPA are both produced by plasma cells of rheumatoid synovial tissues15 ,16 where citrullinated fibrin is located and, together, they most probably contribute to the tissue inflammation although the precise role of each antibody is not clearly understood. In accordance with the hypothesis of a contribution of ACPA to RA synovitis, we demonstrated that, in vitro, immune complexes containing ACPA and citrullinated fibrinogen trigger tumor necrosis factor (TNF)α secretion by macrophages.17

    To further explore the autoimmune reactivity to citrullinated fibrin, we mapped the epitopes targeted by ACPA using 15-mer peptides.18 We identified two peptides, α36–50Cit38,42 and β60–74Cit60,72,74, derived from the α-chains and β-chains of fibrin respectively, as bearing the immunodominant epitopes. A large majority of AhFibA-positive sera were either only reactive to α36–50Cit38,42 (α-sera) or only to β60–74Cit60,72,74 (β-sera) or to both these peptides (αβ-sera).18 ,19 Then, we refined the epitope mapping and identified two 4 amino acids (aa) or 5-aa-long sequences that defined two epitopes (α′- and α-epitope) on α36–50Cit38,42 and one 5-aa β-epitope on β60–74Cit60,72,74.19 By analysing individual sera, we found that a large majority of α-sera recognised only the α-epitope, and that β-sera recognised only the β-epitope, whereas αβ-sera had a less restricted profile recognising the α′-epitope in addition to the α- and β-epitopes.

    The aim of the present study was to analyse how anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies associate in RA sera, and to explore their relationships with one another and with AhFibA, anti-CCP2 and RF, in a large series of patients. We demonstrated that anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies define two distinct, non-overlapping subfamilies of ACPA that, together, encompass more than 90% of the reactivity of RA sera, to citrullinated fibrinogen and also to CCP2 antigens. In addition, we showed that β60–74Cit60,72,74 bears the major epitope recognised by AhFibA and anti-CCP2.

    Patients and methods  

    Patients and serum samples

    Serum samples from 617 patients attending the Rheumatology Centre of the Toulouse teaching hospital were stored at −80°C until assayed in research protocols. Among the patients, 181 had an established RA diagnosed according to the 1987 ACR criteria, and 436 suffered from non-RA rheumatic diseases. Details are given in the online supplementary table.

    Antibody measurements

    RF was measured by nephelometry (IMMAGE Immunochemistry System, Beckman Coulter, Fullerton, California, USA) and anti-CCP2 by ELISA (Immunoscan RA, mark 2; Eurodiagnostica, Arnhem, The Netherlands), both in accordance with the manufacturer's recommendations. IgG-autoantibodies to in vitro citrullinated human fibrinogen (AhFibA), their subclasses, and IgG-autoantibodies to the fibrin-derived citrullyl (Cit)-containing peptides, α36–50Cit38,42 (α-chain aa 36–50) and β60–74Cit60,72,74 (β-chain aa 60–74), referred to as citrullinated peptides, were detected by ELISA according to previously described protocols.18–20 β60–74Cit60,72,74 was synthesised with a terminal carboxamide instead of a free carboxylic group.

    ELISA-based competition assays

    Among RA patients’ sera, three α-sera, three β-sera and three αβ-sera were chosen. None of them was reactive towards the non-citrullinated counterparts of the peptides. Sera were diluted so that they exhibited similar reactivity (optical density (OD)≈1) toward the tested peptide and incubated for 1 h at 4°C with increasing concentrations of the competing citrullinated peptide or with its non-citrullinated counterpart as a control. Then, the reactivity to 36–50Cit38,42 or β60–74Cit60,72,74 was measured.18

    Statistical analyses

    Data analyses were performed using MedCalC software for windows (Broekstraat 52 B-9030 Mariakerke, Belgium). The Kruskal–Wallis or Mann–Whitney U tests were used to compare the median differences. Correlations of antibody titres were tested by the Spearman's rank test. Diagnostic indexes were compared using McNemar's χ2 test. Accurate  CIs were computed by the method described by McKenzie.21  p Value adjustment for multiple comparisons was done by the Holm correction method (sequential Bonferroni).22 p Values ≤0.05 were considered significant.

    Results

    Diagnostic indexes of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies

    Serum reactivity to α36–50Cit38,42 and β60–74Cit60,72,74 was measured comparatively to that of citrullinated fibrinogen (AhFibA; figure 1A). The titres of anti-α36–50Cit38,42 and those of anti-β60–74Cit60,72,74 autoantibodies were significantly higher in RA than in the control sera (p<10-3; figure 1A). No difference was found between men and women, and positive control sera were not associated with a particular diagnosis.

    Figure 1
    Figure 1

    (A) AhFibA, anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibody titres in 181 rheumatoid arthritis (RA) and 436 control sera. AhFibA titres were considered as the OD obtained on citrullinated fibrinogen. Anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibody titres were considered as the difference between the OD obtained on the citrullinated form and that obtained on the non-citrullinated form of each peptide (ΔOD), possibly leading to negative values. (B) Receiver Operating Characteristic curves of the detection of AhFibA, anti-CCP2, RF, rheumatoid factor, anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies, obtained with 181 RA and 436 control sera. Only the part of the curves corresponding to a diagnostic specificity >90% is shown. The 95% and 98.5% specificity thresholds are indicated by dashed lines. (C) Proportions of each IgG-AhFibA subclass, calculated using previously established titration curves20 in three groups of RA sera: sera containing autoantibodies to α36-50Cit38,42 only (α-sera), to β60–74Cit60,72,74 only (β-sera) or to α36–50Cit38,42 and to β60–74Cit60,72,74 (αβ-sera). Boxes represent the 25th–75th percentiles, dots within the boxes represent the medians, and bars extend to the 10th and 90th percentiles.

    The diagnostic indexes of the detection of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies are compared with those of AhFibA, of anti-CCP2 and of RF in the receiver operating characteristic curves presented in figure 1B. AhFibA, anti-CCP2 and anti-β60–74Cit60,72,74 autoantibodies exhibited very high diagnostic values, while RF and anti-α36–50Cit38,42 autoantibodies were less sensitive. The diagnostic sensitivity of AhFibA-ELISA (76.2% and 83.4% at 98.5% and 95% specificity, respectively) was significantly higher than that of all the other tests (p<0.05; table 1). Moreover, the diagnostic sensitivity of anti-CCP2 and that of anti-β60–74Cit60,72,74 autoantibodies did not differ significantly. Finally, the diagnostic indexes of RF and anti-α36–50Cit38,42 autoantibodies were weaker than those of all the other tests (p<0.05). Although the diagnostic sensitivity of RF did not differ from that of anti-α36–50Cit38,42 autoantibodies at 95% specificity, it was significantly lower for a high specificity of 98.5% (p<10−3).

    View this table:
    Table 1

    Diagnostic sensitivities (95% CI) of AhFibA, anti-CCP2, anti-β60–74Cit60,72,74, anti-α36–50Cit38,42 and RF detection for thresholds giving 95% or 98.5% of specificity

    The diagnostic sensitivities of the combined detection of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies, using the 98.5% and 95% specificity thresholds, were 72.4% (95% CI 65.3 to 78.7%) and 79.6% (95% CI 72.9 to 85.2%) respectively, positive sera being considered as those containing autoantibodies directed towards one or both peptides. However, because positive control sera reacting to α36–50Cit38,42 were different from those reacting to β60–74Cit60,72,74, using these thresholds, the diagnostic specificity of the combination decreased to 97.3% or 90%, respectively. We thus evaluated the diagnostic sensitivities of AhFibA, anti-CCP2 or anti-β60–74Cit60,72,74 autoantibodies, for a specificity of 97.3%. Interestingly, the sensitivity of the combination (72.4%, 95% CI 65.3% to 78.7%) was not significantly different from that of AhFibA (79.6%, 95% CI 72.9% to 84.8%), but did not increase sufficiently to be significantly higher than that of anti-CCP2 (69.1%, 95% CI 61.8% to 75.7%) or anti-β60–74Cit60,72,74 (67.4%, 95% CI 60.1% to 76.8%).

    Taking advantage of the determination of IgG-AhFibA subclasses conducted previously,20 we compared the proportions of each AhFibA subclass in sera with anti-α36–50Cit38,42, in sera with anti-β60–74Cit60,72,74 autoantibodies and in sera with both autoantibodies (figure 1C). No difference could be observed between these three groups of sera (p=ns for all comparisons), IgG1 being the major subclass in all groups (p<10−3).

    Anti-β60–74Cit60,72,74 and to a lesser extent anti-α36–50Cit38,42 autoantibody titres are correlated with those of AhFibA and anti-CCP2 in RA sera

    AhFibA and anti-CCP2 titres correlated strongly with each other (r=0.679; p<10−3, figure 2A and table 2). In addition, the titres of autoantibodies to β60–74Cit60,72,74 and to α36–50Cit38,42 correlated significantly with those of AhFibA. This is not surprising, given that the reactivity to α36–50Cit38,42 and to β60–74Cit60,72,74 are part of the reactivity to the complete citrullinated fibrinogen molecule. However, anti-β60–74Cit60,72,74 autoantibodies correlated more strongly with AhFibA than anti-α36–50Cit38,42 autoantibodies did. Similarly, anti-β60–74Cit60,72,74 autoantibodies also correlated strongly with anti-CCP2, the correlation coefficient being almost identical to that observed with AhFibA. By contrast, anti-α36–50Cit38,42 autoantibody titres only correlated weakly with those of anti-CCP2 autoantibodies (table 2).

    View this table:
    Table 2

    Correlation of the titres of AhFibA, of anti-CCP2, anti-α36–50Cit38,42, anti-β60–74Cit60,72,74 autoantibodies and of RF, computed for the 181 RA sera

    Figure 2
    Figure 2

    Correlations of the titres of the various families of anticitrullinated proteins autoantibodies in rheumatoid arthritis  sera (n=181) (A) Anti-CCP2 versus AhFibA (B) Anti-α36–50Cit38,42 versus anti-β60–74Cit60,72,74 (C) Anti-α36–50Cit38,42 titres added to anti-β60–74Cit60,72,74 titres versus AhFibA (D) Anti-α36–50Cit38,42 titres added to anti-β60–74Cit60,72,74 titres versus anti-CCP2. The indicated r value corresponds to the computed Spearman's rank correlation coefficients (95% CI).

    The relationship between anti-α36–50Cit38,42 and β60–74Cit60,72,74 autoantibodies is illustrated in figure 2B. Indeed, most of the dots are located along the axes and correspond to sera that are highly reactive with one peptide and weakly reactive with the other. In addition, their titres correlated only very weakly with each other (r=0.229), and only 8/181 (4.4%) RA sera showed a high reactivity to both peptides (ΔOD>1). Interestingly, the values obtained by summing the titres of anti-α36–50Cit38,42 and β60–74Cit60,72,74 autoantibodies were found to be strongly correlated with the titres of AhFibA (figure 2C; r=0.693; p<10−3) and also with those of anti-CCP2 (figure 2D; r=0.579 ; p<10−3). However, the correlation coefficients were similar to those obtained when comparing titres of β60–74Cit60,72,74 autoantibodies with those of AhFibA (r=0.633) or those of anti-CCP2 titres (r=0.634).

    Finally, RF correlated significantly but only weakly with AhFibA or anti-CCP2 and neither anti-α36–50Cit38,42 nor anti-β60–74Cit60,72,74 autoantibodies exhibited a strong correlation with RF (table 2).

    Anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies largely overlap with the ACPA reactivity to citrullinated fibrinogen and to CCP2 antigens

    In most sera, the presence or absence of AhFibA and anti-CCP2 was concordant. At 98.5% specificity, 572/617 (93%) sera were either positive for both tests or negative for both. In the same way, the presence of autoantibodies to either α36–50Cit38,42 or β60–74Cit60,72,74 was most often concordant with the presence of AhFibA. For example, at a specificity of 98.5%, 576/617 (93%) sera were either (A) positive for AhFibA and positive for anti-α36–50Cit38,42 and/or anti-β60–74Cit60,72,74 autoantibodies or (B) negative for all the autoantibodies. Similarly, the presence of anti-α36–50Cit38,42 and/or anti-β60–74Cit60,72,74 autoantibodies was also concordant with that of anti-CCP2. Indeed, 559/617 (90%) RA sera either were positive for anti-CCP2 and also positive for autoantibodies to at least one of the two peptides or were negative for all three autoantibody reactivities.

    Among RA sera, at 95% specificity, 16/181 (9%) sera contained only anti-α36–50Cit38,42 autoantibodies (α-sera) whereas 52/181 (29%) contained only anti-β60–74Cit60,72,74 autoantibodies (β-sera) and the majority 76/181 (42%) contained both subfamilies of autoantibodies (αβ-sera). Similar results were observed at 98.5% specificity (table 3). Considering the AhFibA-positive RA sera, the distribution was: 14/151 (9%) α-sera, 47/151 (31%) β-sera and 75/151 (50%) αβ-sera (table 3). In other words, 90% (136/151) of the AhFibA-positive RA sera contained autoantibodies to one, the other or both peptides, and the anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies were present separately in 61/151 (40%) sera. A similar distribution was also found within the anti-CCP2-positive RA sera: 12/134 (9%) α–sera, 43/134 (32%) β-sera and 69/134 (51%) αβ-sera. Moreover, 93% (124/143) of the anti-CCP2-positive RA sera contained autoantibodies to one, the other or both peptides (table 3).

    View this table:
    Table 3

    Number (%) of RA sera containing autoantibodies to α36–50Cit38,42 and/or to β60–74Cit60,72,74, among the AhFibA or anti-CCP2-positive or negative sera, at the 98.5% or 95% specificity thresholds

    Anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 comprise two non-crossreactive subfamilies of autoantibodies

    To investigate whether anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 were distinct subfamilies of autoantibodies, we performed competition assays (figure 3). First, using α-sera or β-sera, we tested the ability of α36–50Cit38,42 or β60–74Cit60,72,74 to inhibit the binding of autoantibodies to the same or the other peptide bound to the ELISA plate. The reactivity to α36–50Cit38,42 was not inhibited with 100 µg/ml of β60–74Cit60,72,74, whereas it was inhibited by up to 80% with a concentration of α36–50Cit38,42 as low as 4 µg/ml (figure 3A). No significant inhibition was observed with the non-citrullinated peptides. Similar results were observed with β60–74Cit60,72,74. The reactivity of sera to β60–74Cit60,72,74 was not inhibited by α36–50Cit38,42, whereas it was inhibited by up to 80% with a concentration of β60–74Cit60,72,74 as low as 4 µg/ml (figure 3B). Then, we confirmed that both peptides were able to inhibit the binding of autoantibodies to the same plate-bound peptide using αβ-sera (figure 3C, D). However, at high concentration, β60–74Cit60,72 also weakly inhibited the reactivity of αβ-sera to α36–50Cit38,42 (figure 3C). In the same way, at very high concentration, α36–50Cit38,42 weakly inhibited the serum reactivities to β60–74Cit60,72,74 (figure 3D). Therefore, although, a limited crossreactivity was observed in αβ-sera, autoantibodies to α36–50Cit38,42 and β60–74Cit60,72,74 were largely non-crossreactive.

    Figure 3
    Figure 3

    ELISA-based competition assays. Inhibition of the detection by ELISA (ΔOD) of anti-α36–50Cit38,42 or anti-β60–74Cit60,72,74 autoantibodies in (A) three α-sera, (B) three β-sera and (C and D) three αβ-sera. Plates were coated with α36–50Cit38,42 (A and C) or with β60–74Cit60,72,74 (B and D) and competition assays were performed with increasing amounts of the indicated soluble citrullinated (Cit) or non-citrullinated (NC) competing peptide. Results are expressed as the proportion of the baseline ΔOD inhibited by each concentration of competing peptide (% of inhibition). Each serum was tested in duplicate and results were averaged. Each graph shows means and SDs obtained with the three sera.

    Discussion

    In the present work, we have analysed the relationship between the autoantibodies directed against two peptides, α36–50Cit38,42 and β60–74Cit60,72,74, previously identified as containing the immunodominant epitopes of citrullinated fibrin18 ,19 with AhFibA and anti-CCP2 antibodies.

    We first evaluated their diagnostic performances. AhFibA showed a significantly higher diagnostic performance than anti-CCP2 (83% vs 74% at 95% specificity), although both tests had shown similar diagnostic performance in previous cohorts.9 ,14 Some variability may depend on cohort composition, disease duration and thresholds for positivity.3 ,6 Because our cohort was composed of patients with long-standing RA, we suggest that AhFibA assay allowed the detection of autoantibodies that appeared lately in the disease course and that were not detectable in CCP2 assays. This needs to be confirmed.

    Autoantibodies to α36–50Cit38,42 and β60–74Cit60,72,74 were detected in 50% and 70% of RA sera, respectively. Because their titres did not correlate with each other, and about 40% of RA sera contained autoantibodies directed towards only one of those two peptides, we hypothesised that they were distinct subfamilies of autoantibodies. Using competition assays, we showed that the autoantibodies present in α-sera and β-sera were not crossreactive, while in αβ-sera, some autoantibodies were weakly crossreactive. On the α36–50Cit38,42 and β60–74Cit60,72,74 peptides, we had previously demonstrated that α-sera recognised only the α-epitope of α36–50Cit38,42 and β-sera recognised only the β-epitope of β60–74Cit60,72,74 whereas αβ-sera recognised an α′-epitope in addition to the α- and β-epitopes.19 Our present data show that anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies mainly correspond to two non-crossreactive subfamilies of autoantibodies directed against the α-epitope or β-epitope and include a small fraction of weakly crossreactive autoantibodies that are present only in the αβ-sera and are directed against the α′-epitope.

    Finally, 90% of AhFibA-positive sera contained anti-α36–50Cit38,42, or anti-β60–74Cit60,72,74 or both subfamilies of autoantibodies. Interestingly, the sum of the titres of both antibodies correlated strongly with the titres of AhFibA. Thus, the association of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies almost completely summarise the ACPA reactivity to the complete citrullinated fibrinogen molecule.

    In this cohort of patients, the diagnostic sensitivity of anti-β60–74Cit60,72,74 autoantibodies was similar to that of anti-CCP2 (71% vs 74%). In addition, the proportions of anti-CCP2-positive sera reacting to α36–50Cit38,42 and/or β60–74Cit60,72,74 was similar to that observed for AhFibA: 93% of anti-CCP2-positive sera contained autoantibodies to α36–50Cit38,42 and/or anti-β60–74Cit60,72,74. Thus, anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 autoantibodies also almost completely overlap with the ACPA reactivity to the CCP2 antigens.

    RA serum reactivity to α36–50Cit38,42 and/or β60–74Cit60,72,74 defines three groups of RA-patients (α-, β- and αβ-positive) that might present clinical or genetic differences. Analysing those differences was beyond the scope of the present work. However, since, for example, in pemphigus vulgaris 23 IgG subclasses were shown to be associated with clinical features and IgG4-ACPA were shown to preferentially decrease during anti-TNF treatment of RA,24 we investigated whether there was any difference in the distribution of IgG-AhFibA subclasses within RA sera. We did not find any evidence that ACPA subclass distribution was biased according to serum reactivity to α36–50Cit38,42 and/or to β60–74Cit60,72,74.

    β60–74Cit60,72,74 was recognised by a much higher proportion of RA sera than α36–50Cit38,42. This is of particular interest in connection with two recent works in which the citrullination sites targeted by anti-CCP2 positive sera either on in vitro citrullinated fibrinogen or on peptides extracted from RA synovial tissues, were identified by mass spectrometry.25 ,26 Indeed, in both cases, peptides containing citrullinated forms of the β60–74 peptide were identified, supporting the biological relevance of β60–74Cit60,72,74.

    Beside α36–50Cit38,42 and β60–74Cit60,72,74, additional peptides have been shown to be detected by RA sera. Among them, the fibrinogen-derived peptides, β36–52Cit44 (β-chain aa 36–52) and α27–43Cit35 (α-chain 27–43), were shown to be detected by 26–45% and 8–14% of RA sera, respectively.27–30 Peptides derived from other proteins were also studied. The α-enolase-derived CEP-1 peptide was recognised by 14% to 43% of RA sera,27–31 the Epstein Bar Virus EBNA35–58 peptide by 45%32 and the two vimentin peptides, Vim 2–17 and Vim 59–74, by 2.7–11% and 26–38%, respectively.28–30 The differences between the diagnostic performance of autoantibodies to those peptides probably arise from differences in the cohorts analysed and in the thresholds used for positivity, and also in ELISA protocols, such as the precise nature of the immunodetected antigen (peptide length and sequence), its structure (cyclic or linear peptide), and its immobilisation method (biotinylated or not). Diagnostic performances are thus very difficult to compare. Moreover, concerning the fibrinogen-derived peptides, the biological significance of serum reactivity to citrullyl residues located on thrombin-cleaved portions of fibrinogen, and thus, absent from citrulinated fibrin deposits of the synovium, is not clear.11

    Finally, a degree of crossreactivity toward the various peptides probably also exists. Several partial investigations have addressed this question without drawing definite conclusions.33–35 However, only some RA sera are reactive with each of these peptides. The present study is the first in which autoantibodies directed towards few (only two) peptides derived from a single autoantigen can account for almost all the ACPA reactivity. Therefore, we hypothesise that the epitopes borne by other peptides are molecular mimics of those present on α36–50Cit38,42 or β60–74Cit60,72,74. Crossreactivity of anti-α36–50Cit38,42 and anti-β60–74Cit60,72,74 antibodies with all other autoantigens deserves to be addressed with carefully conducted competition assays. Very recently, two monoclonal ACPA have been described that present distinct crossreactivity patterns to five other autoantigens.36 Interestingly, both of them were reactive to β60–74Cit60,72,74, which we found to bear the major epitope on citrullinated fibrin.

    In conclusion, the peptide β60–74Cit60,72,74 bears the major epitope recognised by the RA specific autoantibodies to citrullinated fibrinogen and by anti-CCP2. Moreover, in this cohort of established RA, the anti-β60–74Cit60,72,74 autoantibodies show a diagnostic performance similar to that of anti-CCP2. However, both autoantibodies to α36–50Cit38,42 and to β60–74Cit60,72,74, mainly composed of non-crossreactive subfamilies of ACPA, are required to reach the ACPA reactivity to the whole molecule of citrullinated fibrinogen or to CCP2 antigens.

    Acknowledgments

    We thank M-F Isaïa, A Legue, E Parra for their excellent technical assistance and Dr C Vincent for his help in the analysis of data.

    References

    1. ↵
      1. Klareskog L,
      2. Ronnelid J,
      3. Lundberg K,
      4. et al
      . Immunity to citrullinated proteins in rheumatoid arthritis. Annu Rev Immunol 2008;26:651–75.
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Sebbag M,
      2. Clavel C,
      3. Nogueira L,
      4. et al
      . Autoimmune response to posttranslationally modified (citrullinated) proteins: prime suspect in the pathophysiology of rheumatoid arthritis. Chichester: John Wiley and Sons, Ltd, 2009;279–308.
    3. ↵
      1. Nishimura K,
      2. Sugiyama D,
      3. Kogata Y,
      4. et al
      . Meta-analysis: diagnostic accuracy of anti-cyclic citrullinated peptide antibody and rheumatoid factor for rheumatoid arthritis. Ann Intern Med 2007;146:797–808.
      OpenUrlCrossRefPubMedWeb of Science
    4. ↵
      1. Girbal-Neuhauser E,
      2. Durieux JJ,
      3. Arnaud M,
      4. et al
      . The epitopes targeted by the rheumatoid arthritis-associated antifilaggrin autoantibodies are posttranslationally generated on various sites of (pro)filaggrin by deimination of arginine residues. J Immunol 1999;162:585–94.
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Schellekens GA,
      2. de Jong BA,
      3. van den Hoogen FH,
      4. et al
      . Citrulline is an essential constituent of antigenic determinants recognized by rheumatoid arthritis-specific autoantibodies. J Clin Invest 1998;101:273–81.
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Whiting PF,
      2. Smidt N,
      3. Sterne JA,
      4. et al
      . Systematic review: accuracy of anti-citrullinated Peptide antibodies for diagnosing rheumatoid arthritis. Ann Intern Med 2010;152:456–64; W155–66.
      OpenUrlCrossRefPubMedWeb of Science
    7. ↵
      1. Nielen MM,
      2. van Schaardenburg D,
      3. Reesink HW,
      4. et al
      . Specific autoantibodies precede the symptoms of rheumatoid arthritis: a study of serial measurements in blood donors. Arthritis Rheum 2004;50:380–6.
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Forslind K,
      2. Ahlmen M,
      3. Eberhardt K,
      4. et al
      . Prediction of radiological outcome in early rheumatoid arthritis in clinical practice: role of antibodies to citrullinated peptides (anti-CCP). Ann Rheum Dis 2004;63:1090–5.
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Nicaise-Roland P,
      2. Nogueira L,
      3. Demattei C,
      4. et al
      . Autoantibodies to citrullinated fibrinogen compared with anti-MCV and anti-CCP2 antibodies in diagnosing rheumatoid arthritis at an early stage: data from the French ESPOIR cohort. Ann Rheum Dis 2013;72:357–62.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Neogi T,
      2. Aletaha D,
      3. Silman AJ,
      4. et al
      . The 2010 American College of Rheumatology/European League Against Rheumatism classification criteria for rheumatoid arthritis: Phase 2 methodological report. Arthritis Rheum 2010;62:2582–91.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Masson-Bessière C,
      2. Sebbag M,
      3. Girbal-Neuhauser E,
      4. et al
      . The major synovial targets of the rheumatoid arthritis-specific antifilaggrin autoantibodies are deiminated forms of the alpha- and beta-chains of fibrin. J Immunol 2001;166:4177–84.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Tilleman K,
      2. Van Steendam K,
      3. Cantaert T,
      4. et al
      . Synovial detection and autoantibody reactivity of processed citrullinated isoforms of vimentin in inflammatory arthritides. Rheumatology 2008;47:597–604.
      OpenUrlAbstract/FREE Full Text
    13. ↵
      1. Kinloch A,
      2. Tatzer V,
      3. Wait R,
      4. et al
      . Identification of citrullinated alpha-enolase as a candidate autoantigen in rheumatoid arthritis. Arthritis Res Ther 2005;7:R1421–9.
      OpenUrlCrossRefPubMedWeb of Science
    14. ↵
      1. Vander Cruyssen B,
      2. Cantaert T,
      3. Nogueira L,
      4. et al
      . Diagnostic value of anti-human citrullinated fibrinogen ELISA and comparison with four other anti-citrullinated protein assays. Arthritis Res Ther 2006;8:R122.
      OpenUrlCrossRefPubMed
    15. ↵
      1. Wernick RM,
      2. Lipsky PE,
      3. Marban-Arcos E,
      4. et al
      . IgG and IgM rheumatoid factor synthesis in rheumatoid synovial membrane cell cultures. Arthritis Rheum 1985;28:742–52.
      OpenUrlCrossRefPubMedWeb of Science
    16. ↵
      1. Masson-Bessière C,
      2. Sebbag M,
      3. Durieux JJ,
      4. et al
      . In the rheumatoid pannus, anti-filaggrin autoantibodies are produced by local plasma cells and constitute a higher proportion of IgG than in synovial fluid and serum. Clin Exp Immunol 2000;119:544–52.
      OpenUrlCrossRefPubMedWeb of Science
    17. ↵
      1. Clavel C,
      2. Nogueira L,
      3. Laurent L,
      4. et al
      . Induction of macrophage secretion of tumor necrosis factor alpha through Fcgamma receptor IIa engagement by rheumatoid arthritis-specific autoantibodies to citrullinated proteins complexed with fibrinogen. Arthritis Rheum 2008;58:678–88.
      OpenUrlCrossRefPubMedWeb of Science
    18. ↵
      1. Sebbag M,
      2. Moinard N,
      3. Auger I,
      4. et al
      . Epitopes of human fibrin recognized by the rheumatoid arthritis-specific autoantibodies to citrullinated proteins. Eur J Immunol 2006;36:2250–63.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Iobagiu C,
      2. Magyar A,
      3. Nogueira L,
      4. et al
      . The antigen specificity of the rheumatoid arthritis-associated ACPA directed to citrullinated fibrin is very closely restricted. J Autoimmunity 2011;37:263–72.
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Chapuy-Regaud S,
      2. Nogueira L,
      3. Clavel C,
      4. et al
      . IgG subclass distribution of the rheumatoid arthritis-specific autoantibodies to citrullinated fibrin. Clin Exp Immunol 2005;139:542–50.
      OpenUrlCrossRefPubMedWeb of Science
    21. ↵
      1. McKenzie DP,
      2. Vida S,
      3. Mackinnon AJ,
      4. et al
      . Accurate confidence intervals for measures of test performance. Psychiatry Res 1997;69:207–9.
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Zhang J,
      2. Quan H,
      3. Ng J,
      4. et al
      . Some statistical methods for multiple endpoints in clinical trials. Control Clin Trials 1997;18:204–21.
      OpenUrlCrossRefPubMedWeb of Science
    23. ↵
      1. Dhandha MM,
      2. Seiffert-Sinha K,
      3. Sinha AA
      . Specific immunoglobulin isotypes correlate with disease activity, morphology, duration and HLA association in Pemphigus vulgaris. Autoimmunity 2012;45:516–26.
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Bos WH,
      2. Bartelds GM,
      3. Vis M,
      4. et al
      . Preferential decrease in IgG4 anti-citrullinated protein antibodies during treatment with tumour necrosis factor blocking agents in patients with rheumatoid arthritis. Ann Rheum Dis 2009;68:558–63.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. van Beers JJ,
      2. Raijmakers R,
      3. Alexander LE,
      4. et al
      . Mapping of citrullinated fibrinogen B-cell epitopes in rheumatoid arthritis by imaging surface plasmon resonance. Arthritis Res Ther 2010;12:R219.
      OpenUrlCrossRefPubMed
    26. ↵
      1. Hermansson M,
      2. Artemenko K,
      3. Ossipova E,
      4. et al
      . MS analysis of rheumatoid arthritic synovial tissue identifies specific citrullination sites on fibrinogen. Proteomics Clin Appl 2010;4:511–18.
      OpenUrlPubMedWeb of Science
    27. ↵
      1. Ioan-Facsinay A,
      2. Willemze A,
      3. Robinson DB,
      4. et al
      . Marked differences in fine specificity and isotype usage of the anti-citrullinated protein antibody in health and disease. Arthritis Rheum 2008;58:3000–8.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. van der Woude D,
      2. Rantapaa-Dahlqvist S,
      3. Ioan-Facsinay A,
      4. et al
      . Epitope spreading of the anti-citrullinated protein antibody response occurs before disease onset and is associated with the disease course of early arthritis. Ann Rheum Dis 2010;69:1554–61.
      OpenUrlAbstract/FREE Full Text
    29. ↵
      1. Hansson LM,
      2. Schlederer T,
      3. Israelsson L,
      4. et al
      . Validation of a multiplex chip-based for the detection of autoantibodies against citrullinated peptides. Arthritis Res Ther. 2012;14:R201.
      OpenUrlCrossRefPubMed
    30. ↵
      1. Willemze A,
      2. van der Woude D,
      3. Ghidey W,
      4. et al
      . The interaction between HLA SE alleles and smoking and its contribution to autoimmunity against several citrullinated antigens. Arthritis Rheum 2011;63:1823–32.
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Lundberg K,
      2. Kinloch A,
      3. Fisher BA,
      4. et al
      . Antibodies to citrullinated alpha-enolase peptide 1 are specific for rheumatoid arthritis and cross-react with bacterial enolase. Arthritis Rheum 2008;58:3009–19.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Anzilotti C,
      2. Merlini G,
      3. Pratesi F,
      4. et al
      . Antibodies to viral citrullinated peptide in rheumatoid arthritis. J Rheumatol 2006;33:647–51.
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Ioan-Facsinay A,
      2. el-Bannoudi H,
      3. Scherer HU,
      4. et al
      . Anti-cyclic citrullinated peptide antibodies are a collection of anti-citrullinated protein antibodies and contain overlapping and non-overlapping reactivities. Ann Rheum Dis 2010;70:188–93.
      OpenUrlPubMedWeb of Science
    34. ↵
      1. van de Stadt LA,
      2. van der Horst AR,
      3. de Koning MH,
      4. et al
      . The extent of the anti-citrullinated protein antibody repertoire is associated with arthritis development in patients with seropositive arthralgia. Ann Rheum Dis 2011;70:128–33.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Lundberg K,
      2. Bengtsson C,
      3. Kharlamova N,
      4. et al
      . Genetic and environmental determinants for disease risk in subsets of rheumatoid arthritis defined by the anticitrullinated protein/peptide antibody fine specificity profile. Ann Rheum Dis 2013;72:652–8.
      OpenUrlAbstract/FREE Full Text
    36. ↵
      1. van de Stadt LA,
      2. van Schouwenburg PA,
      3. Bryde S,
      4. et al
      . Monoclonal anti-citrullinated protein antibodies selected on citrullinated fibrinogen have distinct targets with different cross-reactivity patterns. Rheumatology (Oxford) 2013;52:631–5.
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      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:

    Footnotes

    • Handling editor Tore K Kvien

    • Contributors MC performed laboratory work, ELISAs validation, data analysis and wrote the manuscript. LN planned the study, participated to data analysis and wrote the manuscript. MS and GS participated to data analysis and participated to the manuscript writing. EV, ARW and AC provided sera and patient diagnosis. AM, FB and FH participated to the manuscript drafting. All authors read, commented upon and approved the final manuscript.

    • Funding The study was supported by research grants from the ‘Université Toulouse III’, the ‘Centre National de la Recherche Scientifique’ (CNRS), and the ‘Institut National de la Santé et de la Recherche Médicale’ (INSERM).

    • Competing interests None.

    • Ethics approval Comité de protection des personnes (CPP) Sud Ouest et Outre-Mer II (Toulouse, France) DC2008-463.

    • Provenance and peer review Not commissioned; externally peer reviewed.