Selected cyclic citrullinated peptides derived from the sequence of mutated and citrullinated vimentin (MCV) are targeted by different antibodies subclasses in patients with rheumatoid arthritis in Russian patients O. Derganova1,2, L. Martinez-Gamboa1, K. Egerer1,3, H. Bang4, G. Fredenhagen4, D. Roggenbuck5,6, I. Esaulenko2, G.R. Burmester1, T. Chernykh2, E. Feist1
1Department of Rheumatology and Clinical Immunology, Charite-Universitätsmedizin, Berlin, Germany; 2Voronezh N.N. Burdenko State Medical Academy, Voronezh, Russia; 3Labor Berlin GmbH, Berlin, Germany; 4Orgentec Diagnostika GmbH, Mainz, Germany; 5Medipan GmbH, Dahlewitz/Berlin, Germany; 6Faculty of Science, Lausitz University of Applied Sciences, Senftenberg, Germany. Abstract Objective Antibodies against citrullinated antigens (ACPA) represent one rheumatoid arthritis (RA) classification criteria. Recently, mutated and citrullinated vimentin (MCV), containing approx. 45 potentially citrullinated sites, was characterised as another modified autoantigenic RA target. Therefore, we wanted to screen, select and validate predominant MCV autoantigenic epitopes (called here MCE) as possible new diagnostic targets.
Methods MCV-derived peptides with citrullinated sites were screened in healthy controls and patients. Based on this, twelve selected MCE were used for validation of ACPA isotypes (IgA/IgG/IgM) with ELISA in early RA (ERA, <12 months) and established RA (>12 months) Russian patients. Sensitivity of MCE reactivity was compared to commercially available ELISAs for anti-CCP IgG, anti-MCV IgG, and anti-RF IgA/IgM/IgG.
Results Anti-MCE IgG/IgA//IgM antibodies were observed in 64.1%, 23.1%, and 15.4% ERA, and 63.9%, 26.7%, and 13.1% established RA patients, respectively. Anti-MCV IgG was present in 64.1% ERA and 55.0% RA patients. Furthermore, anti-CCP IgG and RF IgG/IgA/IgM were detectable in up to 76.9%, 71.8%, 71.8%, and 38.5% ERA, and 80.1%, 72.3%, 67.5%, and 43.0% RA patients. Anti-CCP IgG single positivity was observed in 7.7% ERA and 6.3% RA patients. Only one RA patient was anti-MCE single positive.
Conclusion MCV autoantigenic epitopes were emulated by cyclic citrullinated MCV-derived peptides and recognised by all autoantibody-Ig subclasses in RA. Tested MCE were recognized more frequently by IgG as the original MCV antigen. High antibody prevalence against CCP epitopes suggests a strong CCP-linkage to RA pathogenesis in the investigated Russian cohort.
Key words rheumatoid arthritis, citrullinated peptides, ACPA, modified vimentin
Clinical and Experimental Rheumatology 2014; 32: 622-629. Selected MCV-derived epitopes in RA / O. Derganova et al.
Olga Derganova, MD, PhD* Introduction far the CCPs have been the most specific Lorena Martinez-Gamboa* In rheumatoid arthritis (RA), as a autoantigenic structures available for the Karl Egerer, PhD chronic autoimmune disease with ero- serological diagnosis of RA (14, 15). Holger Bang, PhD sive arthritis and systemic involvement, In the past years, additional proteins Gert Fredenhagen, Dirk Roggenbuck, MD, Prof. certain post-translational protein modi- (e.g. fibrinogen, vimentin and alpha- Igor Esaulenko, MD, PhD fications can lead to the formation of au- enolase) have been identified as modi- Gerd Rüdiger Burmester, MD, Prof. toantigens (1). The so far best described fied autoantigens in the pathophysiol- Tatiana Chernykh, MD, PhD modifications include citrullination and ogy of RA. Nowadays these different Eugen Feist, MD mutation of targeted proteins such as humoral autoimmune responses, in- *These authors contributed equally vimentin, fibrinogen, collagen type cluding abs to CCP, are summarised to this study. 2, and alpha-enolase (2-6). Recently, as ACPA. These abs provide superior Please address correspondence proteins with homocitrullin residues (a specificity compared to RF, and also and reprint requests to: carbamylated lysine residue) have also contribute to the prediction of course Eugen Feist, MD, been described as autoantigens in pa- and outcome of disease (16). The sen- Charité Universitätsmedizin Berlin, Department of Rheumatology tients with RA (7). In this context, rab- sitivity of ACPA in the diagnosis of RA and Clinical Rheumatology, bits immunised with homocitrullinated is 70-80%, with a very high specificity Chariteplatz 1, albumin generated antibodies strongly of more than 90% (15). Importantly, the 10117 Berlin, Germany. cross react with commercial cyclic cit- sensitivity of anti-CCP IgG for patients E-mail: [email protected] rullinated peptide (CCP) but weakly with early RA (ERA) is also high, with Received on July 5, 2013; accepted in with mutated citrullinated vimentin approximately 60% (17, 18). revised form on January 22, 2014. (MCV). It has been shown that protein Recently, MCV was isolated from © Copyright Clinical and citrullination could be influenced by synovial fluid of patients with RA and Experimental Rheumatology 2014. environmental factors such as smoking characterised as a new promising au- and that it is closely linked to a genetic toantigen (6). Anti-MCV abs as well as background of the shared epitope (8). few other ACPAs, including abs to third Thus, multiple coincidental and cross- generation CCP, are now used for diag- linked factors are involved in the patho- nostic purposes in the clinic (19). In this genesis of RA leading to the specific study, we have identified and selected autoimmune response observed. antigenic epitopes of MCV (MCE) to Early diagnosis of RA is essential but generate cyclic peptides thereof. These often hindered by a slow onset of unspe- selected antigens were used for char- cific symptoms such as fatigue, loss of acterisation of autoantibody immuno- appetite, weakness, morning stiffness, globulin (Ig) subclass reactivities in arthralgias, and mono- or oligoarthritis, comparison to standard anti-MCV, anti- which can be also seen in many other CCP, and RF assays in RA patients. diseases. Therefore, to improve early diagnosis of RA, new classification cri- Methods teria have recently been developed (9, Patients and controls 10). Considering the relative weight of For standardisation and validation the different available diagnostic pa- of the self-developed ELISA test for rameters, specific laboratory markers, evaluation of reactivity against selected namely seropositivity for rheumatoid MCV peptides, samples from healthy factor (RF) and/or anti-citrullinated controls (n=437, mean age 38 years ± protein antibodies (ACPA), strongly standard deviation (SD) = 13.5 years) contribute to the early identification of were tested. Furthermore, samples from patients with RA (11). individuals with different diseases (e.g. Competing interests: For detection of citrullin-specific anti- rheumatoid arthritis, n=147; arthral- H. Bang and G. Fredenhagen are bodies (abs), anti-CCP assays with dif- gia, n=10; Sjögren’s syndrome, n=10; employees of Orgentec Diagnostika ferent performances are used as a stand- systemic lupus erythematosus (SLE), GmbH, Mainz, Germany; ard worldwide (12). In this context, n=23; polymyalgia, n=1; psoriasis ar- D. Roggenbuck is employee of Medipan the employed citrullinated peptides are thritis, n=28; reactive arthritis, n=5; GmbH, Dahlewitz/Berlin, Germany; artificial and arranged in a cyclic struc- spondylarthritis, n=4; spondylitis anky- the other co-authors have declared no ture. Furthermore, the sequence of CCP losans (Morbus Bechterew), n=1) were competing interests. This work was not supported by any used for the first generation test is- de collected and included for analysis as external financial sources. rived from filaggrine, and it has not been disease controls. In order to examine Test kits were kindly provided by shown that this antigen is expressed in if anti-MCE are detected in infectious Orgentec and Medipan. the synovial tissue (13). Nevertheless, so diseases, additional individuals with
623 Selected MCV-derived epitopes in RA / O. Derganova et al. hepatitis C virus infection (anti-HCV Table I. Clinical characteristics of Russian patients with RA. positive, n=31) were tested by ELISA Clinical characteristics ERA Established RA for anti-MCE (Orgentec) and anti-CCP (n=39) n=191) (Medipan). For further validation and evaluation of autoantibody reactivity Activity of disease (DAS 28 ESR) remission (DAS28<2.6) 0 1 (0.5%) against cyclic epitopes of MCV, we in- low (2.6
624 Selected MCV-derived epitopes in RA / O. Derganova et al. substrate (Organon-Teknika, Boxtel, struments Inc, USA). The assay was showed increased binding to MCV pep- The Netherlands) for 15 min, and the standardised by using positive and neg- tides containing citrulline in contrast to reaction was terminated by adding 50 ative controls. Furthermore, a standard those containing unmodified arginine μl stop solution (0.5 mol/l sulfuric acid) serum was used for calculation of an- residues. In fact, strong ab reactivities to each well. Finally, absorbance at 450 tibody titers (U/ml) by 4-Parameter-Fit were only observed against peptides nm was determined using an ELISA with lin-log coordinates for OD. Nor- with citrulline containing residues. reader (Rainbow Reader, Tecan, Germa- mal range (<20 U/ml) was established ny). All steps were carried out at room by testing blood donors (n=437, with Reactivity against MCV peptide temperature. Background optical densi- mean + 3-fold SD for IgG ab 7.3+3.8 variants ty (OD) values were obtained by adding U/ml, for IgA ab 5.4+2.4 U/ml and for Subsequently, sera from patients with each serum to a well without protein. A IgM ab 8.7+3.6 U/ml). RA, disease controls, and healthy vol- positive serum was defined by an OD To assess the imprecision of the anti- unteers reactive with MCV peptides value more than twice background OD MCE peptide ELISA, anti-MCE posi- were assayed by ELISA on selected (23). Each serum sample was tested in tive sera (one low value sample [L], biotinylated peptide variants. Of note, duplicate. To further define the assay one medium value sample [M] and one none of the patients from the disease characteristics, 21 normal human sera high value sample [H]) were assayed in control groups were positive for the test were assayed in accordance with the five independent runs on one day (inter- (cut-off 20 U/ml), and in the healthy instructions for use. In all subjects with assay) or in a single run (intra-assay). control group, only one donor (0.22%) RA, anti-MCV and RF abs were also For within-run imprecision, L, M and H was positive (26.3 U/ml), while in the determined with commercially avail- samples were measured in six replicates RA group 123 patients (83.7%) were able ELISA test systems according to on one solid phase. The imprecision data tested positive. Peptides missing five manufacturer’s procedure (Orgentec). were calculated using analysis of vari- COOH-terminal residues were still re- ance. The intra-assay coefficient of vari- active with the same number of sera as Reactivity against cyclic MCV-derived ation (CV) was 3.2, 3.1, and 2.3% for the original peptide, although in most peptides in ELISA the mean concentrations of 70.1, 142.3, cases the OD value was lower. Interest- MCV-derived peptides showing a good and 827.3 U/ml, respectively, whereas ingly, reactivity of these sera was en- reaction in the epitope ELISA screen- inter-assay CV was 4.6, 3.4, and 3.3% hanced by cyclisation of the same pep- ing were selected and synthetised as for mean concentrations of 62.1, 186.3, tide via S-S-bridge from the introduced biotinylated large scale peptide vari- and 712.6 U/ml, respectively. N- and C-terminal cystein residues ants (MCE). Single MCE peptides were (data not shown). This can be explained coupled to microtiterplates (Maxisorb, Statistical analysis by an influence of antigen conformation Nunc, Denmark) via streptavidin and Sensitivity was calculated exclu- on antibody binding. A total of 12 pre- blocked with 1% bovine serum albu- sively employing the RA disease con- dominant epitopes were defined by an min (BSA) in PBS for 30 min at room trol group. Specificity was calculated antibody reactivity of OD >0.5. Char- temperature. For blank controls, wells against both the rheumatic disease con- acterisation of fine specificity of the were coupled with 2 μg/ml biotinylated trol and healthy control groups. Valida- ACPAs against MCV resulted mainly in BSA/well. Antibodies against MCE cy- tion of the MCE test and comparison identification of epitopes corresponding clic epitopes of IgG, IgM, and IgA sub- with CCP and MCV tests were made to earlier described dominant vimentin classes were detected by ELISA (Or- with the Russian RA group. Means, peptide sequences for HLA binding and gentec). Sera were diluted 100-fold in standard deviations, p-values and cor- stimulation of T cell proliferation (24- 1x PBS with 1% BSA/0.05% Tween 20 relation analyses were calculated us- 26). Furthermore, there was a positive and incubated for 30 min at room tem- ing GraphPad Prism software (version correlation between the level of ACPAs perature. After washing three times with 4.00). p-values of <0.05 were consid- and the number of recognised epitopes. PBS/0.05% Tween-20, 100 μl of anti- ered to be statistically significant. No single peptide was sufficient to cov- human IgG conjugated to peroxidase er the reactivity corresponding to the (Dianova, Germany) diluted 1:15.000 Results whole recombinant protein, but rather was added to the wells. After incuba- Epitope mapping of the MCV antigen a mixture of at least six MCV related tion (15 min at room temperature), To evaluate the effect of arginine cit- peptides was required for this. These plates were washed again three times, rullination on the antigenicity of MCV six peptides, which are presented in Ta- and bound antibodies were detected and to localise relevant epitopes on lin- ble II, were used for further analyses. with 3,3-5,5- tetramethyl-benzidine ear polypeptides, a panel of anti-MCV (Organon-Teknika) as a substrate. After IgG positive sera (n=142) was tested Autoantibody reactivities against 15 minutes, the reaction was stopped for reactivity with linear 17mer pep- MCE and other established antigens by adding 100 μl 2 M sulfuric acid/ tides covering the complete amino acid in patients with RA well. Plates were read at a wavelength sequence of MCV, based on the MCV Sera from patients with early and estab- of 450/620 nm in a BioTek microwell sequence as reported elsewhere (6). lished RA were used for further evalu- photometer (Synergy HT, BioTek In- All tested anti-MCV IgG positive sera ation of selected MCE by ELISA. In-
625 Selected MCV-derived epitopes in RA / O. Derganova et al. creased IgG, IgA and, IgM to MCE were Table II. Mutated and citrullinated vimentin (MCV)-related peptides required for covering observed in 25 (64.1%), 9 (23.1%), and the reactivity of the whole recombinant protein. 6 (15.4%) patients with ERA, as well MTP Sequence Corresponding vimentin region and substitution as in 122 (63.9%), 51 (26.7%), and 25 position (13.1%) patients with established RA, respectively (Table III). In compari- B1 SSSSYXXMFGXPGTASX Sequence 7 - 24, Arg12/13-Citr, Gly16-Citr son, anti-MCV IgG were positive in 25 E1 SXPSSSXSYVTTSTXTY Sequence 22 - 41, Arg23 - Citr, Arg28 - Citr., Arg35 -Citr (64.1%) patients with ERA, and in 105 H1 TYSLGSALXPSTSXHLY Sequence 37 - 54, Arg45-Citr, Arg50 - Citr, Ser51-His (55.0%) patients with established RA. C2 PGXVYATXSSAVXLXSS Sequence 57 - 74, Gly59-Citr, Arg64 - Citr., Arg69-Citr., With respect to the other autoantibody Arg71-Citr reactivities, anti-CCP IgG (Medipan) G5 NASLAXLDLEXKVESLQ Sequence 211 - 228, Arg217 -Citr, Arg222- Citr and RF IgG, IgA, and IgM (Orgentec) B8 DALXQAKQESTEYXXQV Sequence 307 - 324, Arg310-Citr., Arg320/321 - Citr were elevated in 30 (76.9%), 28 (71.8%), Six predominant MCV related peptides are required for covering the reactivity of the whole MCV 28 (71.8%), and 15 (38.5%) of patients recombinant protein. Bold “X” in the sequences denotes citrulline; bold and underlined “X” indicates with ERA, and in 153 (80.1%), 138 mutated glycine residue in arginine and citrullinated. (72.3%), 129 (67.5%), and 82 (43.0%) of patients with established RA, respec- Table III. Distribution of RA-associated autoantibody subclasses in early (ERA) and estab- tively. Subsequently, the observed low lished RA in the Russian cohort. prevalence of RF IgM measured with ERA (0–12 months) n=39 Established RA (>12 months) n=191 the first ELISA assay was confirmed by using a second commercially available Positive samples Titer Positive samples Titer absolute Mean, U/ml absolute Mean, U/ml ELISA kit (Medipan), revealing positive (relative) numbers (relative) numbers results in only 10 (25.6%) of patients with ERA, and in 37 (19.4%) of patients CCP IgG (negative <30 U/ml) with established RA. These results are 30 (76.9%) 2169.5 153 (80.1%) 1951.1 also shown in Table III. However, for MCE IgA (negative <20 U/ml) further interpretation, only the results for 9 (23.1%) 99.8 51 (26.7%) 318.2 RF obtained by the Orgentec ELISA kits MCE IgG (negative <20 U/ml) were considered. Interestingly, 43.6% 25 (64.1%) 687.4 122 (63.9%) 563.5 of the patients with ERA and 43.9% of MCE IgM (negative <20 U/ml) patients with established RA expressed 6 (15.4%) 378.4 25 (13.1%) 124.2 anti-CCP ab titres of more than 1000 U/ RF IgA (negative <20 U/ml) ml. Of note, the antibody titres for anti- 28 (71.8%) 338.3 129 (67.5%) 364.7 MCE IgG and for anti-MCV IgG were RF IgG (negative <20 U/ml) comparable in individual patients. 28 (71.8%) 359.3 138 (72.3%) 390.6 A parallel seropositivity for all inves- RF IgM (ELISA, from Orgentec; negative <20 U/ml) tigated autoantibodies (anti-CCP IgG, 15 (38.5%) 363.9 82 (43.0%) 228.9 anti-MCV IgG, and anti-MCE IgG) RF IgM (ELISA, from Medipan; negative <20 U/ml) was detectable in 43.6% of samples 10 (25.6%) 28.2 37 (19.4%) 44.5 from patients with ERA, and in 41.9% MCV IgG (negative <20 U/ml) of patients with established RA. In 25 (64.1%) 724.9 105 (55.0%) 532.5 comparison, seronegative results for all autoantibodies (anti-CCP IgG, anti- Shown are the absolute and relative (in parenthesis) numbers, as well as mean titers of positive samples tested for the different parameters (CCP, MCV, MCE, RF) and immunoglobulin (Ig) isotypes (A, G, MCV IgG, and anti-MCE IgG) was M) in a Russian cohort of patients with early (ERA) or established rheumatoid arthritis. Cut-off value present in 12.8% of samples with ERA for CCP was 30, for all other parameters always 20 Units/ml. CCP: cyclic citrullinated peptides; MCV: and 16.8% with established RA. Sin- mutated and citrullinated vimentin; MCE: selected epitopes of MCV; RF: rheumatoid factor. gle positive results for anti-CCP IgG without detectable anti-MCV IgG and tive anti-CCP IgG and anti-MCE IgG, CCP IgG, determination of anti-MCV anti-MCE IgG were observed in 7.7% but negative anti-MCV IgG was found IgG identified an additional propor- of patients with ERA, and in 6.3% of in 15.4% of patients with ERA, and in tion of 10% of ACPA positive cases patients with established RA. 21.5% of patients with established RA. with ERA and 2.6% of patients with The combination of double positive In summary, from all anti-MCE abs established RA. In the same context, anti-CCP IgG and anti-MCV IgG, but positive samples, which were negative anti-MCE IgG ab were detectable in negative anti-MCE IgG was detectable for anti-CCP IgG and for anti-MCV 5% of anti-CCP IgG seronegative pa- in 10.3% of patients with ERA, and IgG, only 0.5% of patients with RA tients with ERA and 0.5% of patients in 10.5% of patients with established were detected in our cohort (Table IV). with established RA. A schematically RA. The combination of double posi- Of note, in patients with negative anti- summary of IgG positive cumulative
626 Selected MCV-derived epitopes in RA / O. Derganova et al.
Table IV. Summary of combinatory analysis of RA-associated CCP/MCV/MCE auto-anti- Discussion bodies. Since mutated and citrullinated vimen- Parameter ERA (n=39) Established RA tin (MCV) is another modified auto- absolute (n=191) absolute antigen in RA, containing approx. 45 (relative) numbers (relative) numbers potentially citrullinated sites, the aim of this study was to screen and select pre- CCP+ / MCV+ / MCE+ 17 (43.6%) 80 (41.9%) CCP- / MCV- / MCE- 5 (12.8%) 32 (16.8%) dominant MCV autoantigenic epitopes, CCP+ / MCV -/ MCE- 3 (7.7%) 12 (6.3%) which we called MCE, as possible new CCP+ / MCV +/ MCE- 4 (10.3%) 20 (10.5%) diagnostic targets, and to validate these CCP+ / MCV -/ MCE+ 6 (15.4%) 41 (21.5%) selected epitopes through analysis of Ig CCP- / MCV +/ MCE+ 2 (5.1%) 0 (0%) CCP- / MCV -/ MCE+ 0 (0%) 1 (0.5%) subclass reactivity against them in de- CCP- / MCV +/ MCE- 2 (5.1%) 5 (2.6%) tail. For this, we used a novel ELISA CCP- / MCV- / MCE- IgG & MCE IgM- / MCE IgA- 4 (10.3%) 29 (14.7%) system for the detection of abs against CCP- / MCV- / MCE- IgG & MCE IgM ± / MCE IgA ± 1 (2.6%) 3 (1.6%) MCE in a cohort of Russian patients MCE IgG- / IgA+ / IgM+ 1 (2.6%) 0 (0%) MCE IgG- / IgA+ / IgM- 2 (5.1%) 3 (1.6%) with confirmed diagnosis of early or es- MCE IgG- / IgA- / IgM+ 2 (5.1%) 4 (2.1%) tablished RA in direct comparison with detection of the original MCV antigen Shown are the absolute and relative (in parenthesis) numbers of samples according to the combinations and standard anti-CCP IgG assay. Taken of CCP/MCV/MCE tested in a Russian cohort of patients with early (ERA) or established rheumatoid arthritis. CCP: cyclic citrullinated peptides; MCV: mutated and citrullinated vimentin; MCE: selected together, we were able to show that the epitopes of MCV. autoantigenic epitopes of MCV could be emulated by using cyclic citrullinat- ed vimentin peptides. Antibody reactivity against MCE in- cluded all Ig subclasses, with a pre- dominance of IgG. Of note, the citrul- linated cyclic vimentin peptides were recognised as autoantigens by abs from patients with early and established RA. The binding of anti-MCV IgG to the peptides was significantly influenced by citrullination of arginine residues as well as by cyclic formation of the epitopes. The main autoantibody subclass of an- ti-MCE abs was identified as IgG, fol- lowed by IgA and IgM subclasses. According to our results, detection of anti-CCP IgG provided a higher sen- Fig. 1. Cumulative positive results of IgG CCP, MCV and MCE in patients with early and established sitivity compared to anti-MCV IgG arthritis. ab as well as anti-MCE IgG in Rus- Approximately 77% of early arthritis patients (ERA) and 80% of established RA patients were positive for CCP IgG. Additionally in CCP IgG negative patients, determination of anti-MCV IgG identified sian patients with early and established 10% ACPA further positive patients with ERA and 2.6% patients with established RA, and anti-MCE RA. Of note, the combination of anti- IgG ab were detectable in further 5% of ERA and 0.5% of patients with established RA. CCP IgG and anti-MCV IgG showed a higher sensitivity than the combination results for these different parameters is ies are not more frequent in patients of anti-CCP IgG and anti-MCE IgG, presented in Figure 1. Taken together, with hepatitis C compared to anti-CCP especially in patients with early RA. anti-MCV IgG and anti-MCE IgG were antibodies. Taken together, the standard anti-CCP found to be positive simultaneously in In this cohort, no significant correla- ELISA provided the highest sensitiv- 15% of anti-CCP IgG negative patients tion was observed between the dif- ity for diagnostic purposes compared to with ERA and 3% with established RA, ferent autoantibody results and clini- both the anti-MCV as well as anti-MCE respectively. cal (DAS28, functional class) as well tests in this cohort of Russian ERA and Notable, none of the patients with hepa- as radiological stage (according to RA patients. The observed differences titis C virus infection (anti-HCV posi- Steinbrocker). However, the number are remarkable and have not been re- tive) showed a positive result for anti- of completely seronegative cases was ported in a similar setting so far. In an- MCE, while 3 samples (9.7%) were very low for ERA (8%) as well as es- other study, a rather low sensitivity of positive against CCP. Although the ana- tablished RA (6%), which does not al- 53.3% for anti-MCV abs has also been lysed number of samples was low, these low appropriate comparison to the se- reported in a cohort of Russian patients results indicate that anti-MCE antibod- ropositive group. with ERA (27). For full explanation of
627 Selected MCV-derived epitopes in RA / O. Derganova et al. these results, the molecular details of Another study described that several Several studies provided strong evi- the antibody binding have to be eluci- citrullinated antigens are recognised dence that ACPA positivity is associated dated in the future with respect to the only by IgG, whereas other are also with presence of RF and joint erosions at published cross reaction of homocitrul- recognised by IgM ACPA, suggesting time of diagnosis, and that this autoanti- line containing proteins/peptides with that the IgM ACPA response is more bodies correlate with radiologic progres- ACPAs. Although the comparability be- restricted than that of IgG ACPA in the sion in RA (40-42). Of note, we were tween anti-CCP and anti-MCV tests for same patient, and that not all citrulli- not able to show any significant correla- diagnosis and prognosis of RA has been nated antigens are able to activate new tion between the different autoantibody shown by different studies with a ten- B cells, despite concurrent recognition reactivities, nor with clinical as well as dency for a higher predictive value and by IgG ACPA (39). In our cohort, the radiological findings. In part, this can specificity of anti-CCP antibodies (6, prevalence of anti-MCE IgA was higher be explained by the very low number of 28-30), it should be also considered that than anti-MCE IgM in ERA and estab- completely seronegative cases for ERA the anti-MCV assay might provide infe- lished RA patients, which might be of as well as established RA, which does rior diagnostic sensitivity compared to relevance for the immunological mech- not allow for appropriate comparison anti-CCP at least in some European RA anisms of the disease. to the seropositive group. Furthermore, cohorts. In this context, a lower sensitiv- Surprisingly, the prevalence of RF the used Steinbrocker’s score is also ity for anti-MCV was also reported re- IgM was confirmed to be very low in a rather non-sensitive method for the cently in cohorts of Italian patients with the investigated cohort of Russian pa- evaluation of radiological findings (43). RA (31). In contrast, previous results tients with ERA as well as with estab- However, a lack of correlation between from other European RA cohorts includ- lished RA by using two independent antibody reactivity to MCV and clinical ing own data showed a higher sensitiv- ELISA tests. Even the comparison with as well as radiological findings has also ity of anti-MCV compared to anti-CCP the other RF subclasses revealed that been reported in ERA patients from a antibodies in RA (6, 17, 32). Since our prevalence of RF IgG was equal to IgA Russian cohort (27). results revealed a higher sensitivity for but higher than IgM in ERA, whereas anti-CCP in the present study, it is pos- prevalence of RF IgG was higher than Conclusions sible that different epitopes are linked to IgA, and IgA higher than IgM in es- So far, the combination of ACPA and a stronger degree to the pathogenesis of tablished RA. In fact, only 3 (7.7%) of RF has provided the best performance RA in Russian patients. This could also patients with ERA and 4 (2.1%) of pa- for the diagnosis of RA. In our study, explain the unusual high antibody titers tients with established RA were posi- the combined analyses of anti-CCP IgG, detected here. tive for RF IgM without detectable RF anti-MCV IgG and anti-MCE abs was It has been reported that a broader IgG or IgA. This low frequency of IgM sufficient to characterise the majority ACPA recognition profile in patients RF accompanied by a high frequency of patients (92% ERA and 94% of RA) with arthralgia seems to be associated of IgG and IgA in this RA population as seropositive. Therefore, for anti-CCP with a subsequent progression to ar- is of special interest. Although the rea- IgG seronegative patients it can be rec- thritis, while a serum conversion from son for this unusual finding is unclear, ommended to test further ACPA, namely ACPA positive to negative and vice ver- it seems that immunoglobulin class anti-MCV IgG as suggested elsewhere. sa was observed rather rarely (33-35). switch in IgM RF producing B cells/ Of note, the autoantigenic properties of However, no significant epitope spread- plasma cells is somehow enhanced to the MCV antigen are also sufficiently ing of ACPAs has been described after produce IgG and IgA isotypes in this displayed by using cyclic citrullinated disease onset. Furthermore, an extended cohort. Interestingly, this effect does vimentin epitopes. However, the MCE ACPA reactivity profile was not associ- not appear to correlate with duration of selected here did not provide an addi- ated with joint damage after a follow-up disease, since IgM RF frequency was tional advantage or improved diagnostic of seven years (36). In this RA cohort, almost equal low both in ERA and in performance compared to the CCP an- we observed a higher sensitivity for established RA patients. Here, it can tigen. In this context, it is important to anti-MCV IgG, anti-MCE IgG and IgM be speculated that an increased Ig class note that ACPAs can recognise different as well as for RF IgA in ERA compared switching in these B cells is influenced antigens, including citrullinated fibrino- to established RA. This can only be par- by the genetic background (e.g. pres- gen, alpha enolase, and collagen, but it tially explained by the geographic ori- ence of shared epitopes), the immuno- is the citrulline moiety which forms the gin of the cohort. Another striking find- logical context (e.g. a prominent acti- antibody binding site. Thus, the amino ing in our study was the high prevalence vation of the mucosal associated im- acids surrounding the citrulline are im- of ACPA of the IgA subclass. Recently, mune response), or by environmental portant for binding of ACPA with differ- a predictive value for the development factors (e.g. smoking, therapeutic strat- ent fine specificities. Therefore, it might of RA has been reported for anti-CCP egy). The influence of these or other be possible to optimise the MCV pep- IgG and IgA, while the prevalence of possible factors on RF formation and tide sequences identified here, to end up anti-CCP IgG and IgA subclass reactivi- maintenance in the Russian population with better overall sensitivity, which is ties was also higher compared to IgM in should be elucidated in a further longi- being investigated in a further ongoing different studies (37, 38). tudinal study. project.
628 Selected MCV-derived epitopes in RA / O. Derganova et al.
References BERGLIN E et al.: Antibodies against cyclic Rheum 2009; 60: 2232-41. 1. NIJENHUIS S, ZENDMAN AJ, VOSSENAAR citrullinated peptide and IgA rheumatoid fac- 30. DEJACO C, KLOTZ W, LARCHER H, DUFTNER ER, PRUIJN GJ, van VENROOIJ WJ: Autoanti- tor predict the development of rheumatoid C, SCHIRMER M, HEROLD M: Diagnostic val- bodies to citrullinated proteins in rheumatoid arthritis. Arthritis Rheum 2003; 48: 2741-9. ue of antibodies against a modified citrullinat- arthritis: clinical performance and biochemi- 17. MATHSSON L, MULLAZEHI M, WICK MC et ed vimentin in rheumatoid arthritis. Arthritis cal aspects of an RA-specific marker. Clin al.: Antibodies against citrullinated vimentin Res Ther 2006; 8: R119. Chim Acta 2004; 350: 17-34. in rheumatoid arthritis: higher sensitivity and 31. BARTOLONI E, ALUNNO A, BISTONI O et al.: 2. SEBBAG M, MOINARD N, AUGER I et al.: extended prognostic value concerning future Diagnostic value of anti-mutated citrullinated Epitopes of human fibrin recognized by the radiographic progression as compared with vimentin in comparison to anti-cyclic citrulli- rheumatoid arthritis-specific autoantibodies antibodies against cyclic citrullinated pep- nated peptide and anti-viral citrullinated pep- to citrullinated proteins. Eur J Immunol 2006; tides. Arthritis Rheum 2008; 58: 36-45. tide 2 antibodies in rheumatoid arthritis: An 36: 2250-63. 18. WHITING PF, SMIDT N, STERNE JA et al.: Sys- Italian multicentric study and review of the 3. LUNDBERG K, KINLOCH A, FISHER BA et al.: tematic review: accuracy of anti-citrullinated literature. Autoimmun Rev 2012; 11: 815-20. Antibodies to citrullinated alpha-enolase pep- Peptide antibodies for diagnosing rheumatoid 32. URSUM J, NIELEN MM, van SCHAARDEN- tide 1 are specific for rheumatoid arthritis and arthritis. Ann Intern Med 2010; 152: 456-64; BURG D et al.: Antibodies to mutated citrul- cross-react with bacterial enolase. Arthritis W155-466. linated vimentin and disease activity score in Rheum 2008; 58: 3009-19. 19. WILLEMZE A, TROUW LA, TOES RE, HUIZ- early arthritis: a cohort study. Arthritis Res 4. MENARD HA, LAPOINTE E, ROCHDI MD, INGA TW: The influence of ACPA status and Ther 2008; 10: R12. ZHOU ZJ: Insights into rheumatoid arthritis characteristics on the course of RA. Nat Rev 33. MJAAVATTEN MD, van der HEIJDE DM, UH- derived from the Sa immune system. Arthritis Rheumatol 2012; 8: 144-52. LIG T et al.: Should anti-citrullinated protein Res 2000; 2: 429-32. 20. PREVOO ML, van ‘t HOF MA, KUPER HH, antibody and rheumatoid factor status be re- 5. BURKHARDT H, SEHNERT B, BOCKERMANN van LEEUWEN MA, van de PUTTE LB, van assessed during the first year of followup in R, ENGSTROM A, KALDEN JR, HOLMDAHL R: RIEL PL: Modified disease activity scores recent-onset arthritis? A longitudinal study. J Humoral immune response to citrullinated col- that include twenty-eight-joint counts. Devel- Rheumatol 2011; 38: 2336-41. lagen type II determinants in early rheumatoid opment and validation in a prospective lon- 34. van de STADT LA, van der HORST AR, de arthritis. Eur J Immunol 2005; 35: 1643-52. gitudinal study of patients with rheumatoid KONING MH et al.: The extent of the anti-cit- 6. BANG H, EGERER K, GAULIARD A et al.: Mu- arthritis. Arthritis Rheum 1995; 38: 44-8. rullinated protein antibody repertoire is asso- tation and citrullination modifies vimentin to 21. HOCHBERG MC, CHANG RW, DWOSH I, ciated with arthritis development in patients a novel autoantigen for rheumatoid arthritis. LINDSEY S, PINCUS T, WOLFE F: The Ameri- with seropositive arthralgia. Ann Rheum Dis Arthritis Rheum 2007; 56: 2503-11. can College of Rheumatology 1991 revised 2011; 70:128-33. 7. SHI J, KNEVEL R, SUWANNALAI P et al.: criteria for the classification of global func- 35. van der WOUDE D, RANTAPAA-DAHLQVIST Autoantibodies recognizing carbamylated tional status in rheumatoid arthritis. Arthritis S, IOAN-FACSINAY A et al.: Epitope spread- proteins are present in sera of patients with Rheum 1992; 35: 498-502. ing of the anti-citrullinated protein antibody rheumatoid arthritis and predict joint damage. 22. STEINBROCKER O, TRAEGER CH, BATTER- response occurs before disease onset and is Proc Natl Acad Sci USA 2011; 108: 17372-7. MAN RC: Therapeutic criteria in rheumatoid associated with the disease course of early 8. WEGNER N, LUNDBERG K, KINLOCH A et al.: arthritis. J Am Med Assoc 1949; 140: 659-62. arthritis. Ann Rheum Dis 2010; 69: 1554-61. Autoimmunity to specific citrullinated proteins 23. AUGER I, BALANDRAUD N, RAK J, LAMBERT 36. WILLEMZE A, BOHRINGER S, KNEVEL R et gives the first clues to the etiology of rheuma- N, MARTIN M, ROUDIER J: New autoantigens al.: The ACPA recognition profile and sub- toid arthritis. Immunol Rev 2010 ; 233: 34-54. in rheumatoid arthritis (RA): screening 8268 grouping of ACPA-positive RA patients. Ann 9. ALETAHA D, NEOGI T, SILMAN AJ et al.: 2010 protein arrays with sera from patients with Rheum Dis 2012; 71: 268-74. Rheumatoid arthritis classification criteria: an RA. Ann Rheum Dis 2009; 68: 591-4. 37. KOKKONEN H, MULLAZEHI M, BERGLIN E et American College of Rheumatology/Europe- 24. VERPOORT KN, CHEUNG K, IOAN-FACSINAY al.: Antibodies of IgG, IgA and IgM isotypes an League Against Rheumatism collaborative A et al.: Fine specificity of the anti-citrulli- against cyclic citrullinated peptide precede initiative. Arthritis Rheum 2010; 62: 2569-81. nated protein antibody response is influenced the development of rheumatoid arthritis. Ar- 10. ALETAHA D, NEOGI T, SILMAN AJ et al.: 2010 by the shared epitope alleles. Arthritis Rheum thritis Res Ther 2011; 13: R13. rheumatoid arthritis classification criteria: an 2007; 56: 3949-52. 38. LAKOS G, SOOS L, FEKETE A et al.: Anti- American College of Rheumatology/Europe- 25. IOAN-FACSINAY A, WILLEMZE A, ROBINSON cyclic citrullinated peptide antibody isotypes an League Against Rheumatism collaborative DB et al.: Marked differences in fine specific- in rheumatoid arthritis: association with dis- initiative. Ann Rheum Dis 2010; 69: 1580-8. ity and isotype usage of the anti-citrullinated ease duration, rheumatoid factor production 11. NEOGI T, ALETAHA D, SILMAN AJ et al.: protein antibody in health and disease. Arthri- and the presence of shared epitope. Clin Exp The 2010 American College of Rheumatol- tis Rheum 2008; 58: 3000-8. Rheumatol 2008; 26: 253-60. ogy/European League Against Rheumatism 26. HILL JA, SOUTHWOOD S, SETTE A, JEVNI- 39. SUWANNALAI P, WILLEMZE A, van TOORN L classification criteria for rheumatoid arthri- KAR AM, BELL DA, CAIRNS E: Cutting edge: et al.: The fine specificity of IgM anti-citrul- tis: Phase 2 methodological report. Arthritis the conversion of arginine to citrulline allows linated protein antibodies (ACPA) is different Rheum 2010; 62: 2582-91. for a high-affinity peptide interaction with from that of IgG ACPA. Arthritis Res Ther 12. COENEN D, VERSCHUEREN P, WESTHOVENS the rheumatoid arthritis-associated HLA- 2011; 13: R195. R, BOSSUYT X: Technical and diagnostic per- DRB1*0401 MHC class II molecule. J Im- 40. MUSTILA A, KORPELA M, HAAPALA AM et al.: formance of 6 assays for the measurement of munol 2003; 171: 538-41. Anti-citrullinated peptide antibodies and the citrullinated protein/peptide antibodies in the 27. SIZOVA L: Diagnostic value of antibodies to progression of radiographic joint erosions in diagnosis of rheumatoid arthritis. Clin Chem modified citrullinated vimentin in early rheu- patients with early rheumatoid arthritis treated 2007; 53: 498-504. matoid arthritis. Hum Immunol 2012; 73: 389- with FIN-RACo combination and single dis- 13. SCHELLEKENS GA, de JONG BA, van den 92. ease-modifying antirheumatic drug strategies. HOOGEN FH, van de PUTTE LB, van VEN- 28. SYVERSEN SW, GOLL GL, van der HEIJDE D Clin Exp Rheumatol 2011; 29: 500-5. ROOIJ WJ: Citrulline is an essential constitu- et al.: Prediction of radiographic progression 41. FEIST E, EGERER K, BURMESTER GR: [Au- ent of antigenic determinants recognized by in rheumatoid arthritis and the role of anti- toantibody profile in rheumatoid arthritis]. rheumatoid arthritis-specific autoantibodies. bodies against mutated citrullinated vimentin: Zeitschrift fur Rheumatologie 2007; 66: 212- J Clin Invest 1998; 101: 273-81. results from a 10-year prospective study. Ann 4, 216-8. 14. van VENROOIJ WJ, VOSSENAAR ER, ZEND- Rheum Dis 2010; 69: 345-51. 42. AGRAWAL S, MISRA R, AGGARWAL A: Au- MAN AJ: Anti-CCP antibodies: the new rheuma- 29. van der LINDEN MP, van der WOUDE D, IO- toantibodies in rheumatoid arthritis: asso- toid factor in the serology of rheumatoid arthri- AN-FACSINAY A et al.: Value of anti-modified ciation with severity of disease in established tis. Autoimmun Rev 2004; 3 (Suppl. 1): S17-19. citrullinated vimentin and third-generation RA. Clinical rheumatology 2007; 26: 201-4. 15. PRUIJN GJ, WIIK A, van VENROOIJ WJ: The anti-cyclic citrullinated peptide compared 43. KAYE JJ, FUCHS HA, MOSELEY JW, NANCE use of citrullinated peptides and proteins for with second-generation anti-cyclic citrulli- EP, Jr., CALLAHAN LF, PINCUS T: Problems the diagnosis of rheumatoid arthritis. Arthri- nated peptide and rheumatoid factor in pre- with the Steinbrocker staging system for radi- tis Res Ther 2010; 12: 203. dicting disease outcome in undifferentiated ographic assessment of the rheumatoid hand 16. RANTAPAA-DAHLQVIST S, de JONG BA, arthritis and rheumatoid arthritis. Arthritis and wrist. Invest Radiol 1990; 25: 536-44.
629