Introduction The anticyclic citrullinated peptide 2 (anti-CCP2) assay is a generic test for antibodies to citrullinated proteins, among which there is a subset of about 50% with antibodies to citrullinated enolase peptide 1 (CEP-1). The anti-CEP-1 positive subset is strongly associated with the HLA-DRB1 shared epitope and its interaction with smoking.
Objective To investigate whether anti-CEP-1 antibodies may be helpful in predicting outcome.
Methods Anti-CEP-1 and anti-CCP2 antibodies were measured in two prospective cohorts of patients (Karolinska n=272, Norfolk Arthritis Register (NOAR) n=408) with early rheumatoid arthritis (RA). Outcomes measured were C-reactive protein, erythrocyte sedimentation rate, visual analogue scales for pain and global assessment of disease activity, Health Assessment Questionnaire, physician's assessment, swollen and tender joint counts and radiological progression.
Results Anti-CCP2 antibodies were present in 57% and 50%, and anti-CEP-1 in 27% and 24% of the Karolinska and NOAR cohorts, respectively. Importantly, no statistically significant differences in clinical outcomes were demonstrated between the anti-CEP-1−/CCP2+ and the anti-CEP-1+/CCP2+ subsets in either cohort, or in radiological outcomes in the Karolinska cohort.
Conclusion Although antibodies to specific citrullinated proteins may have distinct genetic and environmental risk factors, the similarity in clinical phenotype suggests that they share common pathways in the pathogenesis of joint disease in RA.
Statistics from Altmetric.com
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.
Rheumatoid arthritis (RA) is a heterogeneous disease with marked differences in the severity of clinical and radiological outcomes. Antibodies to citrullinated proteins/peptides show disease specificity and are associated with more severe disease. The current ‘gold standard’ is the anticyclic citrullinated peptide 2 (anti-CCP2) assay which detects antibodies that are associated with the most important genetic and environmental risk factors for rheumatoid arthritis (RA) identified to date, including the HLA-DRB1 shared epitope (SE) and smoking.1 However, neither the peptides in the CCP2 test nor the filaggrin peptide in the CCP(1) assay are thought to occur in the human joint. Furthermore, even among patients with anti-CCP2 antibodies there is heterogeneity, with differing rates of radiological progression.2 3 One explanation might be that the anti-CCP2 assay identifies antibodies to different citrullinated autoantigens,4 which vary in pathological importance.5 One of these citrullinated autoantigens is α-enolase. Antibodies to the immunodominant epitope, citrullinated α-enolase peptide 1 (CEP-1), occur in approximately half of anti-CCP2 positive patients, and cross-reactivity of antibodies with CCP2 is only partial.6 We have recently demonstrated that a subset of patients with RA with both anti-CCP2 and anti-CEP-1 antibodies has a very much stronger association with the SE and its interaction with smoking than anti-CCP2 antibody positive but anti-CEP-1 negative patients.7 The purpose of this study was to determine whether anti-CEP-1 antibodies also define a clinical subset among anti-CCP2 positive patients, by investigating outcomes in two independent, prospective early RA cohorts.
Patients and methods
Two hundred and seventy-two patients with early RA of <12 months' duration and fulfilling the 1987 American College of Rheumatology (ACR) classification criteria, were recruited at Karolinska University Hospital, Stockholm between January 1995 and October 2000 as previously described.8 The Norfolk Arthritis Register (NOAR) is a primary care-based inception cohort of patients with inflammatory arthritis that enrols patients in a defined geographical area of the UK with swelling in two or more joints for 4 weeks or longer.9 Data from 408 subjects, recruited between January 2000 and November 2004 and fulfilling 1987 ACR classification criteria for RA within 3 - years of enrolment, were available for analysis. The anti-CCP2 assay only became established in ro utine clinical practice towards the end of this time period and therefore does not constitute a bias for either cohort. Evaluations for all subjects included baseline joint counts, erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), visual analogue scales for pain and global assessment of disease activity and the Health Assessment Questionnaire for measuring disability. A 28-joint count was used for the Karolinska cohort and 28- and 51-joint counts in NOAR. Owing to incomplete ESR data, the DAS28 (CRP) formula was used for NOAR. Physician assessment on a 0–4 scale was also obtained for the Karolinska cohort. Additional clinical measurements were made at 3 months and 1, 2, 3 and 5 years for the Karolinska cohort. More limited follow-up data collected in the NOAR cohort consisted of 51-joint counts and Health Assessment Questionnaire at 1, 2, 3 and 5 years with repeat CRP at 5 years. Both studies were approved by local research ethics committees and all patients gave informed consent.
Anti-CCP2 and anti-CEP-1 antibodies
Serum samples collected at baseline were used for this study. The presence of anti-CCP2 antibodies was determined using second-generation assays: Immunoscan RA (Eurodiagnostica, Malmö, Sweden) for the Karolinska cohort and DIASTAT (Axis-Shield, Dundee, UK) for NOAR. For measurement of anti-CEP-1 antibodies, the same ELISA as in our previous study was used for both cohorts (see online supplementary text).7
For the Karolinska cohort, radiographs of the hands and feet were taken at baseline and repeated at 1 and 2 years. A modified Larsen score which excluded grade 1 and had a maximum score of 160 was used as previously described.8
Comparisons between groups were analysed with the Mann–Whitney test, or a two-sample t-test for normally distributed data. For differences in proportions, the χ2 or Fisher's exact tests were used.
In the Karolinska cohort, 57% were anti-CCP2 antibody positive and 27% had anti-CEP-1 antibodies. This compared with 50% and 24%, respectively, in the NOAR cohort. Most patients with anti-CEP-1 antibodies were also anti-CCP2 positive (92% in Karolinska, 85% in NOAR). Serum from six patients in the Karolinska cohort gave higher optical density values with the arginine-containing control peptide than with the citrulline-containing CEP-1 peptide, and were excluded from further analysis.
Given the strong gene–environment interaction associated with the subset of patients with RA positive for both anti-CEP-1 and anti-CCP2 antibodies, we divided the cohorts into three subpopulations: anti-CEP-1−/CCP2−, anti-CEP-1−/CCP2+ and anti-CEP-1+/CCP2+. A fourth subpopulation, anti-CEP-1+/CCP2−, was too small (n=5 and n=15 in the Karolinska and NOAR cohorts, respectively) for statistical analysis. Baseline characteristics are shown in table 1. In the NOAR cohort, there were trends towards higher CRP, DAS28 (CRP) and presence of rheumatoid factor in the anti-CEP-1+/CCP2+ subset in comparison with the anti-CEP-1−/CCP2+ subset, although these did not reach statistical significance. Otherwise little difference was observed between the two anti-CCP2+ subsets.
Clinical characteristics over time
For the Karolinska cohort, nine different measures of disease activity were available. No statistically significant difference was demonstrated between the anti-CEP-1−/CCP2+ and the anti-CEP-1+/CCP2+ subsets over the 5 years of follow-up (figure 1). Both these subsets showed higher scores for physician's assessment, swollen and tender joint counts and DAS28 than the anti-CEP-1−/CCP2− group. Similarly, in the NOAR cohort there was no statistically significant difference between the anti-CEP-1−/CCP2+ and the anti-CEP-1+/CCP2+ subsets (online supplementary figure S1), although the trend towards higher CRP in the anti-CEP-1+/CCP2+ group persisted at 5 years. Both the anti-CCP2+ subsets had a lower tender joint count over time than the anti-CEP-1−/CCP2− group in the NOAR cohort. The low prevalence of rheumatoid factor in the anti-CEP-1−/CCP2− subset may have influenced doctors and led them to treat less aggressively (see below), so explaining the higher tender joint count.
Median baseline Larsen scores were 5, 4 and 3.5 for the anti-CEP-1−/CCP2−, anti-CEP-1−/CCP2+ and anti-CEP-1+/CCP2+ subsets, respectively, in the Karolinska cohort. The change in Larsen score between 1 and 2 years was significantly greater for both anti-CCP2 positive subsets than for the anti-CEP-1−/CCP2− subset (figure 2). However, there was no statistically significant difference between the two anti-CCP2+ subsets.
Use of disease-modifying antirheumatic drugs
Treatment in RA is titrated to disease activity and, as it is possible that this may have masked differences in disease severity between subsets, we analysed the use of disease-modifying antirheumatic drugs. In the NOAR cohort (online supplementary figure S2), there were no statistically significant differences between the anti-CEP-1−/CCP2+ and anti-CEP-1+/CCP2+ subsets in the percentage who used methotrexate, sulfasalazine, hydroxychloroquine or steroids at baseline, but there was a trend towards higher methotrexate use at 1 (81% vs 67%; p=0.05) and 5 (92%, vs 78%; p=0.09) years in the anti-CEP-1+/CCP2+ subset compared with the anti-CEP-1−/CCP2+ subset. Patients negative for both antibodies used significantly less methotrexate (48.5% at 5 years) than both the anti-CCP2+ groups. No clear patterns of differing disease-modifying antirheumatic drug usage were observed in the Karolinska cohort (data not shown). Information on dosage was not available for either cohort.
It is well established that, at a population level, anti-CCP2 antibodies predict a more severe disease course. Yet at an individual level there are differing outcomes, such as divergent rates of radiographic progression, with some patients not developing erosions at all.2 A subset of patients with antibodies to both CEP-1 and CCP2 has a particularly strong association with the SE and its interaction with smoking.7 We therefore investigated whether this serologically defined subset predicted clinical outcomes in two independent cohorts of patients with early disease.
Importantly, we demonstrated no statistically significant differences in clinical or radiological outcomes between the CEP-1−/CCP2+ and CEP-1+/CCP2+ subsets.
It is difficult to exclude the possibility of a confounding effect from treatment. In the NOAR cohort there were trends towards higher CRP, DAS28 (CRP) and also greater use of methotrexate in the anti-CEP-1+/CCP2+ subset than in the anti-CEP-1−/CCP2+ subset; however, these were not reflected in the Karolinska cohort so must be interpreted with caution. Nevertheless, it remains possible that statistically significant differences might be shown in a much larger cohort.
Our findings seemingly contrast with the preferential segregation of important genetic and environmental risk factors with the anti-CEP-1+/CCP2+ subset.7 One explanation might be that antibodies to citrullinated autoantigens share a common pathogenic pathway that results in joint disease. Part of this pathway may involve epitope spreading to other citrullinated autoantigens,10 11 with the consequence that an assay like the CCP2 test, detecting multiple antibody specificities, would have greater value in predicting downstream clinical outcomes than an assay detecting only one. An alternative consideration is that the SE and smoking may be risk factors for antibodies to other citrullinated autoantigens such as vimentin.12 However, even if these other antibody specificities are of greater pathological importance, one would expect overlap with the anti-CEP-1+/CCP2+ subset, given the similarly strong gene–environment associations. Other immunological variables that may govern the diversity of outcomes seen among patients with anti-CCP2 antibodies include antibody levels,13 number of antibody isotypes14 and number of citrullinated peptides recognised.11 15 Consequently, we propose that anti-CEP-1 antibodies, with their important risk factor associations, are a useful probe for aetiological subsets, but one may have to look outside the joint for clinical associations.
JR and PJV contributed equally to this work and are joint senior authors.
Funding Arthritis Research UK, the Swedish Rheumatism Association and the Swedish Research Council. This study is part of the EU funded research project Autocure, within the 6th Framework Programme.
Competing interests A patent for the diagnostic use of the CEP-1 peptide (patent application number: WO0890360, published on 31 July 2008) is jointly held by two of the authors (PJV and KL) and funded by Imperial Innovations, Imperial College London.
Ethics approval This study was conducted with the approval of the Norwich research ethics committee and Northern Stockholm ethical committee.
Provenance and peer review Not commissioned; externally peer reviewed.