Rheumatoid Arthritis Basic Science Studies
Fine specificity of the anti–citrullinated protein antibody response is influenced by the shared epitope alleles
In classic studies on the genetic background of antibody production, the major histocompatibility complex (MHC) has been shown to act as the most prominent immune response gene that controls the magnitude and the specificity of antibody production. The strongest genetic risk factor for rheumatoid arthritis (RA), the human MHC HLA–DRB1 shared epitope (SE) alleles, predisposes for antibodies against citrullinated proteins (ACPAs). ACPA levels are higher in SE-positive patients with RA than in SE-negative patients with RA. The aim of the present study was to determine whether SE influences not only the magnitude but also the specificity of the ACPA response.
In 2 cohorts of anti–citrullinated peptide 2–positive patients with RA, one from a study of recent-onset arthritis (n = 206) and the other from a treatment strategy study (n = 141), serum antibodies against a citrullinated peptide derived from vimentin (cVim) and antibodies against a citrullinated fibrinogen peptide (cFibr) were determined by enzyme-linked immunosorbent assay. HLA–DRB1 genotyping was performed.
In the first cohort, SE alleles were significantly associated with the presence of antibodies against cVim (odds ratio [OR] 4.95, 95% confidence interval [95% CI] 1.87–15.3) and were not significantly associated with the presence of antibodies against cFibr (OR 1.71, 95% CI 0.70–4.14). These results were replicated in the second cohort (OR 5.05, 95% CI 1.92–13.6 and OR 1.19, 95% CI 0.30–3.97, respectively).
In 2 cohorts of ACPA-positive patients with RA, SE alleles predisposed for the development of antibodies against cVim but not for the development of antibodies against cFibr. These data indicate that SE alleles act as “classic” immune response genes in the ACPA response, because they influence both the magnitude and the specificity of this RA-specific antibody response.
The most prominent genetic risk factors for rheumatoid arthritis (RA), the HLA–DRB1 shared epitope (SE) alleles, encode for a common amino acid sequence in the peptide-presenting part of the HLA class II molecule. These SE alleles have been described recently to be a risk factor for the development of antibodies against citrullinated proteins (ACPAs) rather than the development of RA (1, 2).
ACPAs have been postulated to play a pivotal role in the progression of RA, because they are highly specific and predictive for RA (3, 4), are associated with the extent of joint destruction (5), and have been shown to enhance disease severity in mice with experimental arthritis (6). It has been shown that ACPAs recognize a variety of citrullinated antigens, including citrullinated fibrinogen (cFibr) and citrullinated vimentin (cVim), which is also known as the Sa antigen (7). However, not all ACPA-positive sera will recognize all citrullinated antigens, as has been shown by analyzing the reactivity against different citrullinated peptide antigens (8).
In classic studies of the genetic determinants that influence antibody production in mice, a region (the immune response [Ir-1] region) that controlled the magnitude and specificity of antibody production was found (for review, see ref.9). Because the magnitude of the antibody response in first-generation offspring of parents producing high levels of antibodies and first-generation offspring of parents producing low levels of antibodies was comparable with the magnitude of response in the parent producing high levels of antibodies (10), it was concluded that this region influenced antibody production in a dominant manner. Moreover, it was observed that the ability of an animal to generate a proper antibody response against different model antigens strictly depended on the genetic variant located in the Ir-1 region, denoting that the immune response genes control antibody responses to different antigens (10, 11). Subsequently, the Ir-1 region was found to be similar to the H-2 region (12), the major histocompatibility complex region in mice that was originally identified by skin transplantation experiments. The human analog of this region is the HLA region.
In analogy with these classic studies, we recently reported that among ACPA-positive patients with RA, those without HLA–DRB1 SE alleles displayed lower levels of ACPA than did patients with 1 or 2 SE alleles (2). The number of SE alleles carried by patients did not influence the ACPA levels, indicating a dominant effect of the SE alleles on the level of circulating antibodies. The present study was designed to determine whether the specificity of the ACPA response, in addition to the magnitude of the response, is influenced by the presence of the HLA–DRB1 SE in patients with RA.
PATIENTS AND METHODS
The patients who were analyzed in this study were derived from the Leiden Early Arthritis Clinic (EAC) cohort (n = 206) and from the BeSt (Behandelstrategieën) study (n = 141). The Leiden EAC is an inception cohort of patients with recent-onset arthritis (symptom duration <2 years) that was started at the Department of Rheumatology of the Leiden University Medical Center in 1993 and is described in detail by van Aken et al (13). The BeSt study is a multicenter, randomized, controlled trial designed to compare the clinical efficacy and radiologic outcomes of 4 different treatment strategies in patients with early-onset RA (14).
All patients fulfilled the American College of Rheumatology (formerly, the American Rheumatism Association) 1987 revised criteria for the classification of RA (15) within 1 year of followup (EAC cohort) or at the time of inclusion (BeSt study). In the EAC cohort, 57% of the patients were ACPA positive. In the BeSt study, 61% of the patients were ACPA positive. Only patients who were positive for ACPA and for whom results of HLA–DRB1 genotyping were available were analyzed in this study.
The anti–cyclic citrullinated peptide 2 (anti–CCP-2) test (Immunoscan RA Mark 2; Euro-Diagnostica, Arnhem, The Netherlands) was used to determine whether patients were ACPA positive. A cutoff value of 25 units/ml was used, as described in the manufacturer's instructions.
Antibodies against both the citrullinated and the uncitrullinated form of a linear peptide derived from vimentin (VYATCitSSAVCitLCitSSVP and VYATRSSAVRLRSSVP) and of a linear peptide derived from fibrinogen (NEEGFFSACitGHRPLDKK and NEEGFFSARGHRPLDKK) were determined by enzyme-linked immunosorbent assay (ELISA). These peptides were selected from 2 sets of synthetic peptides that were generated; 1 was derived from the amino acid sequence of human fibrinogen α-chain and β-chain, and 1 was derived from that of human vimentin. The peptides that were synthesized contained at least 1 citrulline at a position of an arginine in the amino acid sequence of the respective proteins. The recognition of these peptides by several sera from patients with RA was determined, after which the peptides that were most frequently reactive with RA sera were selected for our study. For both of these peptides (cVim and cFibr), the corresponding “arginine variant” was synthesized as well and was used in parallel for the analyses. The specificity of the cVim and cFibr ELISAs was established by analysis of their recognition by 30 normal human sera as well as sera from 50 anti-CCP–negative patients with RA. Only 1 of these sera showed very low reactivity against cFibr (data not shown).
A signal higher than the mean optical density (OD) plus 2 SD for serum samples from 5 healthy control subjects that were included on each plate was considered positive. Citrulline-specific signals were defined as a positive signal against the citrullinated peptide and a negative signal against the uncitrullinated peptide, with a minimum difference of an OD value of 0.1. When a sample had a positive signal against both the citrullinated peptide and the uncitrullinated peptide, the sample was excluded from analyses, which was the case for 30 samples against fibrinogen (12 from the EAC cohort and 18 from the BeSt cohort) and 23 samples against vimentin (5 from the EAC cohort and 18 from the BeSt cohort).
Microtiter plates were coated with 10 μg/ml peptide diluted in phosphate buffered saline (PBS)/0.1% bovine serum albumin (BSA) at 4°C overnight. The coated plates were incubated with serum samples (100 μl/well) for 1 hour (diluted 100-fold in PBS/1% BSA/0.05% Tween 20). This and the subsequent incubation step were performed at 37°C in a humidified atmosphere and were followed by washing steps with PBS/0.05% Tween 20. Antibodies were detected after incubation for 1 hour with 100 μl/well rabbit anti-human IgG horseradish peroxidase–conjugated antibody (P0214; Dako, Glostrup, Denmark) (diluted 1:10,000 in PBS/1% BSA/0.05% Tween 20). Bound antibodies were visualized using 100 μl/well 3,3′,5,5′-tetramethylbenzidine solution (1:1 ratio) mixed with ureumperoxide as a substrate, followed by 100 μl/well 2M H2SO4 10 minutes later to stop the staining reaction. OD values were measured using an ELISA reader at 450 nm.
The HLA–DRB1 (sub)typing was performed using a polymerase chain reaction with specific primers and hybridization with sequence-specific oligonucleotides, as previously described (16). The SE alleles are DRB1*0101, *0102, *0104, *0401, *0404, *0405, *0408, *1001, and *1402.
Odds ratios (ORs) were calculated for the relative proportions of patients with antibodies (anti-cVim or anti-cFibr) among the SE-positive patients compared with the patients without SE alleles. First, the EAC cohort was analyzed. To replicate the data, ORs were calculated for the BeSt study. Subsequently, data from the 2 cohorts were pooled and analyzed with the help of a chi-square test to detect differences in proportions, to provide insight into the robustness of the results. ORs were reported with 95% confidence intervals (95% CIs), which excluded the value of 1 in case of statistical significance.
To investigate whether SE alleles are associated with the specificity of the ACPA response, the presence of antibodies against 2 citrullinated peptides derived from vimentin (cVim) and fibrinogen (cFibr) was determined in serum samples obtained from ACPA-positive patients with RA. These peptides were selected from a panel of vimentin-derived and fibrinogen-derived peptides based on the relatively high frequency of recognition by antibodies from patients with RA.
Among ACPA-positive patients with RA derived from the Leiden EAC cohort, 39% displayed antibodies against both cVim and cFibr, 7% displayed antibodies against only cVim, 36% displayed antibodies against only cFibr, and 19% had no antibodies against either cVim or cFibr. SE alleles were significantly associated with the presence of antibodies against cVim (OR 4.95, 95% CI 1.87–15.3) and not with the presence of antibodies against cFibr (OR 1.71, 95% CI 0.70–4.14) (Table 1). These data indicated a contribution of the HLA–DRB1 SE alleles in determining the fine specificity of the ACPA response.
Table 1. Presence or absence of antibodies against cVim and cFibr in ACPA-positive patients with RA, according to SE status*
| SE positive||86||81||4.95 (1.87–15.3)||124||38||1.71 (0.70–4.14)|
| SE negative||6||28||21||11|
| SE positive||62||30||5.05 (1.92–13.6)||80||14||1.19 (0.30–3.97)|
| SE negative||9||22||24||5|
To confirm and replicate these data, we subsequently performed similar analyses in another patient group consisting of ACPA-positive patients with RA who were derived from the BeSt study. Among these patients, 55% displayed antibodies against both cVim and cFibr, 2% displayed antibodies against only cVim, 28% displayed antibodies against only cFibr, and 15% had no antibodies against either cVim or cFibr. Comparable with what was observed in patients from the EAC cohort, in patients from the BeSt study, the presence of SE alleles was associated with the presence of antibodies against cVim (OR 5.05, 95% CI 1.92–13.6) and not with the presence of antibodies against cFibr (OR 1.19, 95% CI 0.30–3.97) (Table 1).
An analysis of both cohorts as a single group of patients in order to evaluate the robustness of the results yielded highly significantly more frequent detection of anti-cVim in patients with SE alleles (OR 4.44, 95% CI 2.28–8.75, P < 10−6) and no significantly higher frequency of anti-cFibr in SE-positive patients (OR 1.39, 95% CI 0.69–2.78, P = 0.31). Taken together, these findings indicate that SE alleles influence the specificity of the ACPA response.
The citrullinated CCP-2 peptide was selected for its ability to identify RA patients with high sensitivity and specificity. Many patients displaying antibodies against the CCP-2 peptide can also recognize other citrullinated peptides. However, they do not all recognize other citrullinated peptides to the same degree. In this study, we analyzed whether ACPA-positive (anti–CCP-2–positive) patients with RA, with and without SE alleles, differ in their recognition of 2 different citrullinated peptides (cVim and cFibr). In 2 separate cohorts, the presence of SE alleles was significantly associated with the presence of antibodies against cVim and not with the presence of antibodies against cFibr. When the data from these separate cohorts were analyzed together, the association with SE alleles was highly significant for the response against cVim (P < 10−6) and not for the response against cFibr (P = 0.31).
SE alleles have been demonstrated to be a risk factor for the development of ACPAs. Among patients displaying ACPAs, the level of ACPAs has been reported to be higher in patients with SE alleles (2). Although only 2 citrullinated peptides were analyzed in this study, our data are of interest because they show that SE alleles are associated with not only the magnitude but also the fine specificity of the ACPA response. Taken together, these data indicate that SE alleles act as “classic” immune response genes in the ACPA response.
The observation described above is also of interest because it points to the possibility that a peptide derived from vimentin, or a protein physically linked or structurally related to vimentin, is presented to T cells that are restricted by the HLA–DRB1 SE alleles (or the HLA–DQ alleles that are genetically linked to the SE alleles). In this model, citrullinated vimentin/protein complexes presented on, for example, apoptotic cells could be recognized by citrulline-specific B cells that would subsequently internalize this complex, process it, and present peptides from vimentin(-coupled protein) to T cells. These T cells could then provide help to the B cells, eventually resulting in the production of ACPAs. Although this scenario is highly speculative, it is intriguing to note that vimentin is a protein that is known to be citrullinated during apoptosis and is expressed on apoptotic cells (17). By that means, it may become visible for citrullinated vimentin–specific B cells.
In conclusion, in 2 cohorts of ACPA-positive patients with RA, we observed that the presence of SE alleles predisposed for the development of antibodies against a citrullinated vimentin peptide and not for the development of antibodies against a citrullinated fibrinogen peptide. SE alleles thereby act as “classic” immune response genes in the ACPA response, because they influence the magnitude as well as the fine specificity of this RA-specific antibody response.
Dr. Toes had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study design. Verpoort, Breedveld, Huizinga, Pruijn, Toes.
Acquisition of data. Verpoort, Cheung, Ioan-Facsinay, van der Helm-van Mil, de Vries-Bouwstra, Allaart, Drijfhout, Huizinga, Toes.
Analysis and interpretation of data. Verpoort, Ioan-Facsinay, van der Helm-van Mil, Allaart, Drijfhout, de Vries, Huizinga, Pruijn, Toes.
Manuscript preparation. Verpoort, de Vries-Bouwstra, de Vries, Huizinga, Pruijn, Toes.
Statistical analysis. Verpoort.
Production of reagents. Drijfhout.