Background
Autoantibodies to citrullinated proteins (ACPA) are today a well-known and accepted feature of rheumatoid arthritis (RA) [
1,
2]. These autoantibodies have been linked to RA risk factors, most notably
HLA-DRB1 shared epitope (SE) alleles and cigarette smoking, and their presence predicts a more destructive disease process [
3‐
7]. However, despite the identification of several putative citrullinated autoantigens, including fibrinogen [
8], vimentin [
9], type II collagen [
10], α-enolase [
11] and histone 4 [
12], the specific in vivo ACPA targets triggering autoimmunity and driving disease remain obscure.
More recently, antibodies to carbamylated proteins containing homocitrulline (anti-CarP antibodies) were described in RA [
13]. Protein carbamylation, or homocitrullination, is an enzyme-independent post-translational modification of lysine residues by isocyanate, present in, for example, cigarette smoke [
14]. As smoking is a well-described risk factor for RA [
15,
16], it has been proposed that smoking could be linked to anti-CarP antibodies in RA via increased carbamylation and the subsequent production of anti-CarP antibodies [
17‐
19]. However, scientific data in support of this hypothesis has yet to be presented. Anti-CarP antibodies are specific for RA [
20] and reportedly distinct from ACPA, based on the detection of anti-CarP antibodies in ACPA-negative disease [
13,
21,
22]. However, in the Swedish Epidemiological Investigation of RA (EIRA) study and in the Dutch Early Arthritis Clinic (EAC) cohort, we recently showed that only 4–7 % of RA patients were anti-CarP antibody-positive in the absence of ACPA. Notably, there was no specific association between
HLA-DRB1 SE or smoking and anti-CarP antibodies, when the analyses were adjusted for the presence of ACPA [
21].
In addition, the widely used biochemical assay for detection of peptidylcitrulline, the so-called Senshu method [
23] where rabbit polyclonal antibodies bind chemically modified citrulline residues, was found to also detect homocitrulline [
24] and purified ACPA have been shown to bind not only citrullinated fibrinogen, but also carbamylated fibrinogen [
20].
The extent to which these two autoantibody specificities are cross-reactive, and the association between these antibodies and environmental and genetic risk factors for RA, has not been thoroughly explored, and as yet, only fibrinogen and more recently vimentin have been studied in this context [
13,
20,
21,
24‐
26]. Therefore, it is imperative that more work on specific antigens is performed in order to fully understand the relationship between ACPA and anti-CarP antibodies in the aetiopathology of RA [
19].
Citrullinated α-enolase has long been scrutinized as a potential target for ACPA in RA [
11,
27‐
34]. Antibodies to CEP-1, the immunodominant B cell epitope of citrullinated α-enolase [
27] are found in approximately 40 % of patients with RA, and have been associated with SE,
PTPN22 and smoking [
28,
32]. Hence, in the present study, we have investigated the antibody responses to citrullinated and carbamylated α-enolase and their relation to RA risk factors in the Swedish population-based case-control cohort EIRA.
Methods
Patients
The present study includes patients newly diagnosed with RA (cases) and age-matched, sex-matched and residential area-matched controls from the Swedish Epidemiological Investigation of RA (EIRA) cohort. Information on cigarette smoking (“ever smoker” or “never smoker”) was obtained via self-reported questionnaire at baseline [
16]. Genotyping of
HLA-DRB1 shared epitope (SE) alleles and the protein tyrosine phosphatase gene (
PTPN22 rs2476601) was performed on blood samples obtained within one week of the RA diagnosis [
5,
35]. Smoking and genetic data for the present study were retrieved from the EIRA database on 2784, 2235 and 2477 patients with RA and 4864, 1923 and 1936 controls, for smoking, SE and
PTPN22, respectively. For antibody purification, plasma samples from patients with RA with a strong anti-CEP-1 antibody response (n = 5) or a strong anti-CCP2 antibody response (n = 38) were collected at the Rheumatology Clinic, Karolinska University Hospital Solna, Stockholm, Sweden. Informed consent was obtained from participating patients and controls, and ethical approvals for the study was granted at the regional ethics review board at Karolinska Institutet, Stockholm, Sweden.
Antigens
Three cyclic peptides corresponding to amino acid 5-21 of full-length α-enolase were synthesized by Innovagen (Malmö, Sweden): the original CEP-1 peptide containing two citrulline residues (CEP-1) [
27]; the arginine-containing control peptide REP-1; and a version of CEP-1 containing homocitrulline in the place of citrulline, denoted carb-CEP-1 (Additional file
1: Table S1).
Recombinant human α-enolase, produced in-house, and purified human fibrinogen (Enzyme Research, South Bend, IN, USA) depleted of immunoglobulins, were citrullinated or carbamylated in vitro. Citrullination was performed for 2 h at 37 °C, at a protein concentration of 1 mg/ml, in peptidylarginine deiminase (PAD) buffer (100 mM Tris, 10 mM CaCl2, 5 mM dithiothreitol (DTT), pH 7.6) using 2 U/mg of rabbit skeletal muscle PAD2 enzyme (Sigma, St. Louis, MO, USA). The reaction was stopped by the addition of 20 mM ethylenediaminetetraacetic acid (EDTA), followed by extensive dialysis to calcium-free PBS. Carbamylated proteins were produced by incubating α-enolase and fibrinogen in PBS at 1 mg/ml in the presence of 100 mM potassium cyanate (KOCN) (Sigma) overnight at 37 °C, followed by extensive dialysis to calcium-free PBS. Successful citrullination and carbamylation were confirmed by mass spectrometry (data not shown). For a detailed description of the mass spectrometry analysis see Additional file
1: Supplementary methods.
Affinity purification of ACPA IgG
Plasma samples (n = 43) were centrifuged and diluted in PBS (1:5 v/v) before applied to Protein G HP columns (GE Healthcare) for whole IgG enrichment. To further purify CEP-1-specific IgG, REP-1 and CEP-1 peptides (1 mg/ml) were directly coupled to 1 ml NHS-Sepharose columns (GE Healthcare), and anti-CEP-1 IgG from five anti-CEP-1 antibody-positive serum samples was subsequently purified from whole IgG using the CEP-1 affinity column, after pre-absorption on the REP-1 column to remove non-citrulline-specific antibodies. Bound antibodies were eluted with 0.1 M glycine-HCl (pH 2.7) and directly neutralized with 1 M Tris (pH 9). Column flow-through (FT) fractions depleted of anti-CEP-1 IgG were collected in parallel. Microsep
TM UF Centrifugal Devices (Pall Life Science, Port Washington, NY, USA) were used in accordance with the manufacturer’s instructions, to concentrate the antibodies and to change the buffer into PBS. Anti-CCP2-reactive IgG from 38 anti-CCP2-positive RA serum samples were pooled after purification on CCP2-columns kindly donated by EuroDiagnostica AB, Malmö, Sweden, as previously described [
36].
Antibody detection using ISAC and ELISA
High-throughput anti-CEP-1, anti-REP-1 and anti-carb-CEP-1 antibody screening of serum samples from 2836 patients with RA from the EIRA cohort and 373 EIRA controls was accomplished using a custom-made microarray based on the ImmunoCAP immuno solid-phase allergen chip multiplex assay (ISAC) microarray system (Phadia AB, Uppsala, Sweden) containing the peptides of interest, as previously described [
25,
37]. This microarray also contains a large number of other citrullinated peptides derived from different proteins, including fibrinogen, vimentin and collagen type II, and their corresponding arginine-containing control peptides. Cut offs for antibody positivity were calculated based on the 98th percentile among the EIRA controls. A detailed description of the ISAC method is provided in Additional file
1.
For testing the reactivity of the affinity-purified anti-CEP-1 and FT IgG fractions, and for analysing the degree of cross-reactivity between double-positive (CEP-1
+/Carb-CEP-1
+) or single-positive (CEP-1
+/Carb-CEP-1
- and CEP-1
-/Carb-CEP-1
+) EIRA RA serum samples, peptide ELISAs detecting anti-CEP-1 and anti-Carb-CEP-1 IgG were used as previously described [
27,
28,
32] (see Additional file
1: Supplementary methods for details).
Cross-reactivity assay
Anti-CEP-1/anti-Carb-CEP-1 double-positive serum samples (n = 4), anti-CEP-1 single-positive (n = 4), and anti-Carb-CEP-1 single-positive serum samples (n = 4) were selected for the cross-reactivity experiment. Serum samples were diluted 1:100 in radioimmunoassay (RIA) buffer and incubated with 100 μg/ml of the CEP-1 or the Carb-CEP-1 peptide for 2 h at room temperature (RT). Following incubation, the absorbed serum was analysed using the same protocol as for the peptide ELISA described previously (see also Additional file
1).
Western blot
Citrullinated, carbamylated and unmodified proteins (100 ng/well) were separated on NuPAGE® Bis-Tris 4-20 % gels (Bio-Rad Laboratories, Hercules, CA, USA) and transferred to nitrocellulose membranes. Membranes were blocked with 5 % milk in Tris-buffered saline/0.05 % Tween and probed with a pool (n = 38) of purified anti-CCP2 IgG (or the corresponding CCP2 column FT IgG pool) at 2 μg/ml overnight at 4 °C, then washed with PBS/0.05 % Tween and incubated with HRP-conjugated goat anti-human IgG (Jackson ImmunoResearch, West Grove, PA, USA), diluted 1:10,000, for 1 h at RT. Bound antibody was detected using ECL chemiluminescence (GE Healthcare).
Statistics
Patients were divided into four different subsets according to the presence or absence of anti-CEP-1 and anti-carb-CEP-1 IgG. The odds ratio (OR) and 95 % confidence intervals (CI) for each RA subset, in relation to smoking, SE and PTPN22, were calculated separately through unconditional logistic regression models, adjusted for matching variables (age, gender and residential area). Exposed individuals were compared with unexposed individuals (smokers vs. non-smokers, carriers of any copy of SE vs. non-carriers, carriers of the PTPN22 risk allele vs. non-carriers). All analyses were implemented through SAS V.9.3. Statistical differences in antibody levels and number of ACPA fine specificities, between different subsets, were determined by the Mann-Whitney U test for independent groups. The same method was also used to determine the relationship between anti-carb-CEP-1 antibody levels and SE/smoking.
Discussion
This is the first report of an antibody response to carbamylated α-enolase in RA. Previous reports on anti-CarP antibodies in RA have focused mainly on carbamylated fibrinogen or the complex protein mixture of carbamylated fetal calf serum [
13,
20‐
22,
24,
25], and recently a report was published on antibodies against carbamylated vimentin (26). Our study suggests that the anti-CarP antibody response in RA can be explained by cross-reactive ACPA. This conclusion was particularly evident from immunoblotting experiments using affinity-purified anti-CCP2 IgG molecules. Purified ACPA bound not only citrullinated α-enolase and citrullinated fibrinogen, but also carbamylated α-enolase and -fibrinogen, while unmodified proteins were not targeted, and importantly, ACPA-depleted IgG was not able to recognize citrullinated or carbamylated epitopes. Notably, the two earliest reports linking carbamylation to the development of arthritis in mouse models already describe cross-reactivity between citrulline- and homocitrulline-containing epitopes [
17,
18].
Using a synthetic and artificial peptide based on the well-characterised CEP-1 epitope from citrullinated α-enolase [
27,
28,
32] but with homocitrulline in place of citrulline (amino acids 9 and 15), we showed that approximately 20 % of patients with RA in the EIRA cohort had antibodies to carb-CEP-1. While the carb-CEP-1 peptide most likely does not represent an in vivo antigenic target (amino acid 9 and 15 of unmodified α-enolase are arginines, not lysines), and any biological and mechanistic implications based on the peptide data therefore are limited, the observed cross-reactivity between anti-CEP-1 IgG and carb-CEP-1 suggests that antibodies to citrullinated α-enolase can also bind to homocitrulline-containing epitopes. The fact that the anti-carb-CEP-1 antibody-positive subset of patients was almost exclusively confined to the CEP-1-positive population, together with the observation that CEP-1 could block carb-CEP-1 reactivity much more efficiently than carb-CEP-1 could block CEP-1 reactivity, clearly supports our interpretation; that is, that the reactivity measured as anti-CarP-antibodies are for the most part represented by cross-reactive ACPA.
The group of patients with anti-carb-CEP-1 antibodies had higher anti-CEP-1 antibody levels, but also higher anti-CCP2 antibody levels, and a broader ACPA repertoire (than anti-CEP-1 antibody-positive patients without anti-carb-CEP-1 antibodies), suggesting a stronger ACPA response in general in this group of patients, with epitope spreading and more promiscuous antigen-recognition, i.e., also including epitopes containing homocitrulline, which is structurally very similar to citrulline. Our gene-environment association data suggest that this extended antibody reactivity is influenced by HLA-DRB1 SE alleles, but not by PTPN22 or smoking.
Recently, anti-CarP antibodies have also been described in small subsets of patients with non-RA early arthritis [
38], and in a large portion of patients with primary Sjögren’s syndrome [
39], where the presence of anti-CarP antibodies correlated with disease severity. Contrary to our conclusion, these reports seem to indicate that this class of autoantibodies could be more of a general marker for inflammation than cross-reactive ACPAs, which would be specific for RA.
Taken together, our data seem to suggest that cross-reactivity between ACPA and anti-CarP antibodies in RA is a common phenomenon. Here, we have described this cross-reactivity in the context of the RA candidate autoantigen α-enolase. However, supported by recent work from Scinocca and colleagues, demonstrating cross-reactivity between citrullinated and carbamylated fibrinogen [
20], we posit that this is also likely the case for other citrullinated/carbamylated antigens. In line with previous reports [
20,
24], this cross-reactivity is not complete or consistent, and indeed a small percentage (3 %) of RA cases in our study demonstrated reactivity exclusively to the carb-CEP-1 peptide and not CEP-1. However, this subset of patients was almost completely eliminated when using a more stringent cut off for positivity, casting doubt on the existence of specific anti-CarP antibodies.
Acknowledgements
We thank EIRA participants, research nurses and the EIRA study group, for their contributions; Professor Lars Klareskog, for establishing the EIRA study, and for support and scientific input; scientists previously involved in the generation of data for the EIRA database: Drs Leonid Padyukov, Patrick Stolt and Camilla Bengtsson; and Drs Per Matsson, Mats Nystrand and Thomas Schlederer (Thermo Fisher Scientific, Uppsala, Sweden) for their scientific support concerning the ISAC platform. This work was supported by grants from the Swedish Research Council, the Swedish Rheumatic Foundation and the EU-funded FP7 project TRIGGER (FP7-Health-2013-306029).
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (
http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Competing interests
Drs Hansson and Rönnelid are partners with Thermo Fisher Scientific within the Innovative Medicines Initiative, a public–private partnership between the EU and the European Federation of Pharmaceutical Industries (see
http://btcure.eu). Dr Mathsson-Alm works at Thermo Fisher Scientific on the Innovative Medicines Initiative project. Thermo Fisher Scientific contributes to this consortium with in-kind contributions for the development of the ISAC assay used in the current study. KL is co-inventor of patent US12/524,465, describing the diagnostic use of the CEP-1 epitope. No non-financial conflicts of interest exist.
Authors’ contributions
ER performed all ELISA and peptide absorption experiments, purified anti-CEP-1 and anti-CCP2 antibodies and together with KL wrote the manuscript and selected references. XJ performed statistical analysis of EIRA data and produced Tables
1 and 2. NK performed western blot experiments in Fig.
1. JY performed mass spectrophotometric analysis of modified proteins. AC was responsible for recruiting and obtaining serum samples from patients with RA for purification of ACPAs. LI developed the ELISA methods used. LM-A and MH performed and analyzed ISAC experiments. LA supervised the work of XJ and is responsible for administrating the EIRA study. JR provided scientific feedback, helped structure the study and performed some of the statistical analyses. All authors helped revise, read and approved the final manuscript.