Abstract
Objective. To examine fine specificity of anticitrullinated peptide antibodies (ACPA) in relation to responsiveness to anti-tumor necrosis factor (anti-TNF) agents in rheumatoid arthritis (RA).
Methods. Samples from 450 patients with RA treated with anti-TNF agents were analyzed for antibodies to citrullinated α-enolase, vimentin, and fibrinogen peptides. The Disease Activity Score-28 was measured at baseline and 6 months.
Results. Both anti-cFib antibodies and the number of citrullinated peptides recognized were associated with a poorer response. These findings were not significant following stratification for anti-cyclic citrullinated peptide 2 antibodies.
Conclusion. The presence of any ACPA rather than individual ACPA specificities was associated with a poorer response to anti-TNF agents. We suggest that this reflects distinctive differences in the pathogenesis of ACPA-positive and negative RA.
The second-generation cyclic citrullinated peptide assay (CCP2) is widely used for the diagnosis of rheumatoid arthritis (RA). However, the peptides it contains were selected from a library in order to maximize diagnostic sensitivity. Consequently it may fail to differentiate subsets of patients with antibodies to particular in vivo citrullinated proteins1,2.
One of the most important subsets in clinical practice consists of nonresponders to anti-tumor necrosis factor (anti-TNF) agents. In the Biologics in Rheumatoid Arthritis Genetics and Genomics Study Syndicate (BRAGGSS) cohort, the presence of anti-CCP antibodies was associated with a poorer response to anti-TNF agents3. However, a model comprising both clinical and serological predictors could account for only a small proportion of variance in drug response. As the anti-CCP assay is a generic test for anticitrullinated protein/peptide antibodies (ACPA) this does not account for the possibility that antibodies to specific citrullinated peptides could have greater predictive value. A number of different citrullinated antigens have been reported as reactive with RA sera4, but only a few have been mapped to immunodominant peptides and their specificity and sensitivity reproducibly examined in several laboratories. These include citrullinated α-enolase (CEP-1, amino acids 4–21)1,5,6,7,8, vimentin (cVim, amino acids 60–75)2,8,9,10,11,12, and fibrinogen ß chain (cFib, amino acids 36–52)8,10,11,12. The apparent preferential association of anti-CEP-1 and cVim antibodies with important genetic and environmental risk factors for RA1,2,10 led us to investigate whether these also segregated with clinical response.
MATERIALS AND METHODS
Patient selection, recruitment, and sample collection were as described3. Serum samples from 450 patients with RA enrolled in BRAGGSS were available for analysis. Informed consent and ethical approval were obtained.
Serum samples were analyzed for antibodies to CEP-1 (CKIH-ACit-EIFDSCitGNPTVEC), cVim (CVYATCitSSAVCitLCitSSVPC), and cFib (CNEEGFFSACitGHRPLDKKC) peptides by in-house ELISA. Cysteine residues were added at the amino- and carboxy-termini of each peptide to facilitate cyclization, and antibodies to the corresponding arginine-containing control peptides were measured in parallel to ensure positive results were specific for the citrullinated peptide. Anti-CCP2 antibodies were analyzed using a commercial ELISA (Diastat, Axis-Shield, Dundee, UK).
The primary outcome measure was change in 28-joint Disease Activity Score (DAS28) between baseline and 6 months. Associations with the magnitude of change were tested by linear regression. The European League Against Rheumatism (EULAR) response criteria were a secondary outcome measure and associations were tested by ordinal logistic regression.
RESULTS
Baseline characteristics are detailed in Table 1 and are comparable to those previously reported for the wider British Society for Rheumatology Biologics Register population3.
Baseline demographics of the 450 patients with rheumatoid arthritis included in this study. Values indicate number (%) or mean (SD).
Antibodies to CEP-1, cVim, and cFib were positive in 39%, 42%, and 75% of samples, respectively. In comparison, 84% were positive for anti-CCP antibodies and 87% for rheumatoid factor. Only 30 sera (7%) had antibodies to CCP2 but not to 1 or more of cFib, cVim, and CEP-1. Conversely, 19 sera (5%) had antibodies to 1 or more of these 3 peptides, but not to CCP.
Antibodies to cFib, but not CEP-1 or cVim, were associated with a smaller reduction in DAS28 scores at 6 months, indicating a poorer outcome (Table 2). This was statistically significant when analyzed by linear regression (coefficient 0.37; p = 0.02) and controlling for baseline DAS28, concomitant use of disease-modifying antirheumatic drugs (DMARD), sex, Health Assessment Questionnaire results, and smoking status, which were all found to be predictive of response in this cohort.
Change in DAS28 scores and EULAR response criteria following 6 months of anti-TNF therapy in anti-CEP-1, anti-cVim, and anti-cFib antibody-positive and negative patients with RA. Values indicate number (%) or mean (SD). Linear regression adjusted for baseline DAS28, concomitant DMARD, sex, HAQ, and smoking.
There was also a trend toward a poorer EULAR response with anti-cFib antibodies and the same adjustments (OR 0.68 for response, 95% CI 0.42, 1.02; p = 0.06), but not with anti-CEP-1 (OR 0.90) or anti-cVim antibodies (OR 1.01).
Patients with antibodies to both CCP2 and cFib (n = 303) were compared to those with antibodies to CCP2 but not cFib (n = 49) in a linear regression model with the same adjustments and with the double antibody-negative patients as referent. Coefficients were 0.55 (95% CI 0.13, 0.97; p = 0.01) and 0.31 (95% CI −0.24, 0.8; p = 0.27), respectively, which did not differ significantly, and the adjusted R2 for this model did not differ from that with anti-CCP2 alone (both 0.15). The number of positive assays incorporating citrullinated peptides was associated with response, with a coefficient of 0.13 (p = 0.03). Thus the group with 4 positive assays had a mean DAS28 score at 6 months that was 0.57 higher than the group with none (Figure 1). However, there was no association with peptide number when the analysis was restricted to patients with anti-CCP2 antibodies (coefficient 0.02; p = 0.78). These results did not differ according to the anti-TNF inhibitor used (data not shown).
Relationship between change in DAS28 scores after 6 months of anti-tumor necrosis factor treatment and number of positive antibody tests for citrullinated peptides.
DISCUSSION
Recently, attention has focused on the “fine specificity” of the ACPA response1,2,8,10,11. We have demonstrated that antibodies to cFib are associated with a poorer response to anti-TNF agents at 6 months in patients with RA. In contrast, antibodies to CEP-1 and cVim, which have recently been shown to be strongly associated with the HLA-DRB1 shared epitope (SE)1,2, were not. These results are consistent with neither the SE nor PTPN22 predicting response to anti-TNF agents3, with the same being true for most other RA susceptibility genes13. However, the association with anti-cFib antibodies might simply reflect their high frequency in patients with anti-CCP2 antibodies, with the key serological determinant of response being positivity to ACPA, rather than ACPA specificity. Increasing evidence suggests important differences between the pathogenesis of ACPA-positive and ACPA-negative RA. Such evidence includes different HLA associations and other genetic and environmental risk factors that segregate only with ACPA-positive disease. ACPA are also associated with greater disease severity, joint erosions, and possibly responsiveness to methotrexate in early disease (reviewed by Klareskog, et al14). In accord with this, no statistically significant difference was shown between anti-CCP2-positive patients with and those without anti-cFib antibodies. In addition, we found that the number of peptides recognized by RA sera was associated with response. Antibodies to a higher number of citrullinated peptides are associated with the onset of RA8,11, and it is possible that this might reflect a more mature immune response that is harder to inhibit with anti-TNF agents. Again, however, this finding was not significant when restricted to patients with anti-CCP2 antibodies, indicating no additional predictive value.
Limitations to our study have been discussed in detail3. Of note, serum samples were not taken prior to treatment with anti-TNF agents and antibody status might have changed subsequent to this. However, anti-CCP2 antibodies do not show a large magnitude of change following anti-TNF inhibitor therapy15 and a change in status from positive to negative is unusual3. While we have not been able to demonstrate that antigen-specific ACPA testing shows promise in predicting treatment response, consistent with recent findings for other clinical outcomes16,17, 2 possibilities have not been excluded. First, it remains possible that there is an unidentified citrullinated autoantigen with a more marked association with poor response, since research into specific ACPA has been relatively recent. Second, our results may have been confounded by differences in dose of concomitant DMARD16 or the ability of the anti-CCP assay to detect multiple specificities. The latter possibility could be investigated with the multiplex assays for ACPA that are currently in development. However, our findings in this cohort suggest that the presence of any ACPA rather than individual or multiple ACPA specificities is associated with a poorer response to anti-TNF agents. The distinctive difference therefore lies in the differing pathogenesis of ACPA-positive and ACPA-negative RA.
Acknowledgment
We are grateful to Laura Gibbons for help with statistical analysis.
APPENDIX
List of study collaborators: The Biologics in Rheumatoid Arthritis Genetics and Genomics Study Syndicate (BRAGGSS): Dr. Kimme Hyrich, Prof. John Isaacs, Prof. Ann Morgan, Prof. Gerry Wilson. Basingstoke and North Hampshire NHS Foundation Trust: Dr. R.K. Moitra, Dr. P.J. Prouse, Dr. D.J. Shawe. Cambridge University Hospitals NHS Foundation Trust: Dr. A.J. Crisp, Prof. J.S.H. Gaston, Dr. F.C. Hall, Dr. B.L. Hazleman, Dr. J.R. Jenner, Dr. M.S. Lillicrap, Dr. A. Ostor, Dr. B. Silverman, Dr. C. Speed. Central Manchester University Hospitals NHS Foundation Trust: Prof. I.N. Bruce, Dr. K. Hyrich, Prof. A. Barton, Dr. P. Ho, Dr. R. Gorodkin. County Durham and Darlington NHS Foundation Trust: Dr. D. Armstrong, Dr. A.J. Chuck, Dr. S. Hailwood, Dr. N. Kumar. Derby Hospitals NHS Foundation Trust: Dr. L.J. Badcock, Dr. C.M. Deighton, Dr. S.C. O’Reilly, Dr. N. Raj, Dr. M.R. Regan, Dr. G.D. Summers, Dr. R.A. Williams. Doncaster and Bassetlaw Hospitals NHS Foundation Trust: Dr. J.R. Lambert, Dr. R. Stevens, Dr. C. Wilkinson. Gateshead Health NHS Foundation Trust: Dr. J. Hamilton, Dr. C.R. Heycock, Dr. C.A. Kelly, Dr. V. Saravanan. Hereford Hospitals NHS Trust: Dr. D.H. Rees, Dr. R.B. Williams. Leeds Teaching Hospitals NHS Trust: Dr. S. Bingham, Prof. P. Emery, Dr. A. Morgan, Prof. H.A. Bird, Prof. P.G. Conaghan, Dr. C.T. Pease, Dr. R.J. Wakefield. Mid-Staffordshire NHS Foundation Trust: Dr. S.V. Chalam, Dr. D. Mulherin, Dr. T. Price, Dr. T. Sheeran, Dr S. Venkatachalam. Norfolk and Norwich University Hospital NHS Foundation Trust: Dr. K. Gaffney, Prof. A.J. Macgregor, Dr. T. Marshall, Dr. P. Merry, Prof. D.G.I. Scott. Northumbria Healthcare NHS Foundation Trust: Dr. F.N. Birrell, Dr. P.R. Crook. Pennine Acute Hospitals NHS Trust: Dr. B. Harrison, Dr. M. Pattrick, Dr. H.N. Snowden, Dr. A.P. Bowden, Dr. E.E. Smith, Dr. P. Klimiuk, Dr. D.J. Speden. Peterborough and Stamford Hospitals NHS Foundation Trust: Dr. N.J. Sheehan, Dr. N.E. Williams, Dr. S. Dahiya. Portsmouth Hospitals NHS Trust: Dr. R.G. Hull, Dr. J.M. Ledingham, Dr. F. Mccrae, Dr. M.R. Shaban, Dr. A.L. Thomas, Dr. S.A. Young Min. Queen Mary’s Sidcup NHS Trust: Dr. A.N. Bamji, Dr. N.T. Cheung. Sandwell and West Birmingham Hospitals NHS Trust: Prof. C.D. Buckley, Dr. D.C. Carruthers, Dr. R. Elamanchi, Dr. P.C. Gordon, Dr. K.A. Grindulis, Dr. F. Khattak, Dr. K. Raza, Dr. D. Situnayake. Sheffield Teaching Hospitals NHS Foundation Trust: Dr. M. Akil, Dr. R. Amos, Dr. D.E. Bax, Dr. S. Till, Dr. G. Wilson, Dr. J. Winfield. South Tees Hospitals NHS Foundation Trust: Dr. F. Clarke, Dr. J.N. Fordham, Dr. M.J. Plant, Dr. S. Tuck, Dr. S.K. Pathare. Southampton University Hospitals NHS Trust: Dr. C.J. Edwards, Dr. N.K. Arden, Dr. R.D. Armstrong, Dr. A. Calogeras, Prof. C. Cooper, Dr. B.K.S. Davidson, Dr. E.M. Dennison. St. Helens and Knowsley Hospitals NHS Trust: Dr. V.E. Abernethy, Dr. A.R. Clewes, Dr. J.K. Dawson, Dr. M. Lynch. The Dudley Group of Hospitals NHS Foundation Trust: Prof. G. Kitas, Dr. J.P. Delamere, Dr. N. Erb, Dr. R. Klocke, Dr. A.J. Whallett. The Newcastle upon Tyne Hospitals NHS Foundation Trust: Mr. P. Crook, Dr. H.E. Foster, Dr. B. Griffiths, Dr. I.D. Griffiths, Dr. M.L. Grove, Prof. J.D. Isaacs, Dr. L. Kay, Dr. W.F. Ng, Dr. A. Myers, Dr. P.N. Platt, Dr. D.J. Walker. University Hospital Birmingham NHS Foundation Trust: Dr. S. Bowman, Dr. P. Jobanputra, Dr. R.W. Jubb, Dr. E.C. Rankin. University Hospital of North Staffordshire NHS Trust: Dr. P.T. Dawes, Dr. C.M. Dowson, Dr. A. Hassell, Prof. E.M. Hay, Dr. S. Kamath, Dr. J. Packham, Dr. R.S. Sandhu, Dr. M.F. Shadforth. University Hospitals of Morecambe Bay NHS Trust: Dr. M. Bukhari, Dr. W.N. Dodds, Dr. J.P. Halsey, Dr. W.S. Mitchell. West Suffolk Hospitals NHS Trust: Dr. D.T. O’Reilly. Whipps Cross University Hospital NHS Trust: Dr. S.P. Donnelly, Dr. D. Doyle, Dr. A. Hakim, Dr. J.G. Lanham, Dr. H.I.S. Tahir, Dr. J. Rathi. Worcestershire Acute Hospitals NHS Trust: Dr. A. Rai, Dr. I.F. Rowe, Dr. A. Prabu, Dr. C.A.M. Buckley.
Footnotes
-
Supported by Arthritis Research UK; the collaboration was part of the European Union initiative Autocure, within the sixth framework program.
- Accepted for publication January 12, 2012.
REFERENCES
In this issue
