Levels of Anti-Citrullinated Protein Antibodies and Rheumatoid Factor, Including IgA Isotypes, and Articular Manifestations in Ulcerative Colitis and Crohn’s Disease

Systemic presence of arthritis autoantibodies (AAb) is specific for rheumatoid arthritis (RA). AAb initiation might be triggered by chronic mucosal inflammation, such as in inflammatory bowel disease (IBD). We assessed the prevalence of anti-citrullinated protein antibodies (ACPA) and rheumatoid factor (RF) in ulcerative colitis (UC) and Crohn’s disease (CD) patients, with regard to the prevalence of joint complaints in AAb+ versus AAb− IBD patients. RA patients and healthy subjects (HC) served as controls. Serum was collected from 226 UC, 165 CD and 86 RA patients, and 36 HCs. One-hundred-and-ten UC (48.7%) and 76 CD (46.1%) patients were seropositive for at least one autoantibody, compared to 4 (13.9%) HCs and 81 (94.2%) RA patients. Eighty-three (37%) UC and 52 (32%) CD patients were seropositive for the anti-cyclic citrullinated protein antibody (anti-CCP2) of the immunoglobulin A type (IgA anti-CCP2), compared to 1 (2.8%) HC and 64 (74%) RA patients. RF of the immunoglobulin G type (IgG RF) and IgA RF seropositivity in UC and CD patients was comparable to HCs and low compared to RA patients. Arthralgia was reported by 34 (18.7%) UC and 50 (33.1%) CD patients, but presence of arthralgia was not increased in AAb+ patients. AAbs are frequently present in IBD patients, supporting the hypothesis that inflammation of intestinal mucosa induces low systemic levels of ACPA.


Introduction
Inflammatory bowel disease (IBD) is characterized by inflammation of the intestinal mucosa. Ulcerative colitis (UC) and Crohn's disease (CD) are the main subtypes. UC and CD differ in the location and nature of the disease [1,2]. UC is usually characterized by involvement of the colon with acute and chronic inflammation limited to the mucosa, while inflammation can be transmural in any

Patients
A total of 391 IBD patients (226 patients with UC and 165 with CD) participating in the Dutch IBD Pearl Chain Institute Biobank Initiative were included. All patients gave their informed consent before study enrollment. Patients were included when there was a confirmed diagnosis for UC or CD. Patients were examined by a gastroenterologist at the outpatient clinic of the University Medical Center Groningen, who asked about the presence of arthralgia. The diagnosis of arthritis and spondylarthropathies needed to be confirmed by a rheumatologist. Blood was withdrawn at study enrollment and serum was stored at −80 • C until further use. Clinical and demographic information, including age, gender, diagnosis, duration of disease, body mass index (BMI), and smoking habits were standardized and recorded at the time of blood withdrawal. Serological parameters, such as pANCAs, anti-ASCA (IgA and IgG), C-Reactive Protein (CRP) levels and erythrocyte sedimentation rate (ESR) were routinely measured. CRP was measured by turbidimetry with a lower limit of detection ≤3 mg/L. An HC group (n = 36), consisting of persons without systemic disease [24], was included to define cut off-levels for autoantibody levels. Additionally, 86 RA patients were included as a reference group for serological measurements. The study was conducted with the approval of the Medical Ethics Committee of the University Medical Center Groningen.

Laboratory Measurements
Anti-cyclic citrullinated protein antibody (anti-CCP2) levels were measured using the commercial anti-CCP2 ELISA (Eurodiagnostica). The manufacturer's protocol was adjusted to measure specifically low IgG anti-CCP2 levels as earlier described [16]. Samples were diluted 1:10 instead of 1:50 in dilution buffer. Seropositivity was defined as >2SD above the mean of HC (2.2 U/mL).
The specific ACPA response of the 20 IBD patients with the highest IgG anti-CCP2 levels was measured with ELISA by testing the reactivity against 4 well-known citrullinated antigens, which are often recognized by ACPA from RA patients. These well-known citrullinated antigens were two peptides from fibrinogen (Fib1, β-chain amino acids 36-52, NEEGFFSACitGHRPLDKK and Fib2, β-chain amino acids 60-74, CitPAPPPISGGGYCitACit), one peptide from α-enolase (Eno1, KIHACitEIFDSCitGNPTVE), and one peptide from vimentin (Vim1, VYATCitSSAVCitLCitSSV). Citrulline-specific IgG reactivity was determined by measuring the difference in reactivity against the citrullinated and native form of the peptides, with the cut-off defined as the difference in optical density (∆OD) >2SD above the mean of HCs, as previously described [24].
IgA anti-CCP2 levels were measured using a modified anti-CCP2 assay. Sera were diluted 1:50 using the dilution buffer provided by the manufacturer. Bound human IgA was detected by the secondary antibody horseradish peroxidase (HRP)-conjugated polyclonal goat anti-human IgA (SouthernBiotech, Birmingham, AL, USA), diluted in phosphate-buffered saline (PBS) with 1% bovine serum albumin (BSA) and 0.05 % Tween-20 (Sigma-Aldrich, St. Louis, MO, USA). The color reaction was performed using tetramethylbenzidine (Sigma-Aldrich) and hydrogen peroxide. A pool of sera from 4 RA-patients with high IgA anti-CCP2 levels served as a calibrator for the standard curve expressed in arbitrary units per milliliter (AU/mL) and starting at 200 AU/mL. Seropositivity was defined as >2 SD above the mean of HCs.
IgM and IgA RF levels in serum were measured by a validated in-house ELISA assay [25]. Levels were expressed in international units per milliliter (IU/mL), and seropositivity was defined as >10 IU/mL for IgM RF and >25 IU/mL for IgA RF.

Statistical Analysis
Statistical analysis was performed with GraphPad Prism software (version 5.00 for Windows, GraphPad Software, La Jolla, CA, USA) and IBM SPSS Statistics for Windows software (version 20.0, IBM, Armonk, NY, USA). For group comparisons, a Mann-Whitney U test was used for continuous variables, and Fisher's exact test or Chi-square test for categorical variables. Values of p < 0.05 were considered to be significant.

General Characteristics
The demographic and clinical characteristics of patients and healthy controls are depicted in Table 1. UC patients had a significantly higher BMI than CD patients (p = 0.001), while the CD-cohort was comprised of more (ever) smokers than the UC cohort (p = 0.009). The disease duration was similar in both groups.

Autoantibody Seropositivity
Overall, IgG anti-CCP2 levels were low when the diagnostic value in RA diagnosis was taken into account (25 U/mL). In contrast to the RA patients, none of the IBD patients displayed levels above the diagnostic cut-off value of RA, with the highest level being 13.6 U/mL in the IBD patients. However, when seropositivity was defined using a lower cut-off, >2SD above the mean of HCs, 29 UC patients (13%) and 28 CD patients (17%) were seropositive for IgG anti-CCP2, while 2 (5.6%) HCs and 74 (86%) RA patients were seropositive. Additionally, the specific IgG response against four citrullinated peptides (Fib1, Fib2, Eno1 and Vim1) was measured in 20 sera from IBD patients with the highest IgG anti-CCP2 response, all being higher than 4 U/mL. None of these IBD patients showed citrulline-specific reactivity against any of these four peptides. Therefore, citrulline-specific reactivities against these peptides were not measured in any of the other IBD patients, as no positive results were to be expected.
IgA anti-CCP2 levels are generally not used in RA diagnosis, therefore no diagnostic cut-off is known. When applying a similar cut-off, as was used for IgG anti-CCP2, >2SD above the mean of HCs, 83 UC patients (37%) and 52 CD patients (32%) were seropositive for IgA anti-CCP2, while only 1 (2.8%) HC and 64 (74%) RA patients were.
In total, 110 UC patients (48.7%) and 76 CD patients (46.1%) were seropositive for at least one arthritis autoantibody, of which the majority were positive for IgA anti-CCP2. The autoantibody levels are shown in Figure 1, while seropositivity in the groups is depicted in Figure 2.
reactivities against these peptides were not measured in any of the other IBD patients, as no positive results were to be expected.
IgA anti-CCP2 levels are generally not used in RA diagnosis, therefore no diagnostic cut-off is known. When applying a similar cut-off, as was used for IgG anti-CCP2, >2SD above the mean of HCs, 83 UC patients (37%) and 52 CD patients (32%) were seropositive for IgA anti-CCP2, while only 1 (2.8%) HC and 64 (74%) RA patients were.
In total, 110 UC patients (48.7%) and 76 CD patients (46.1%) were seropositive for at least one arthritis autoantibody, of which the majority were positive for IgA anti-CCP2. The autoantibody levels are shown in Figure 1, while seropositivity in the groups is depicted in Figure 2.

Rheumatic Manifestations and Arthritis Autoantibody Status
Next, IBD patients who were seropositive for any of the arthritis autoantibodies (AAb+) were compared to arthritis autoantibody negative (AAb−) IBD patients on clinical manifestations ( Table  2). Over 90% of AAb+ patients were seropositive for IgA and/or IgG anti-CCP2. AAb+ and AAb− patients did not differ in smoking behavior, disease duration or disease activity. AAb+ positive patients were significantly younger than AAb− patients (UC p = 0.016, CD p = 0.017). AAb+ UC patients were significantly younger at the disease onset than AAb− UC patients (p = 0.021). CRP

Rheumatic Manifestations and Arthritis Autoantibody Status
Next, IBD patients who were seropositive for any of the arthritis autoantibodies (AAb+) were compared to arthritis autoantibody negative (AAb−) IBD patients on clinical manifestations (Table 2). Over 90% of AAb+ patients were seropositive for IgA and/or IgG anti-CCP2. AAb+ and AAb− patients did not differ in smoking behavior, disease duration or disease activity. AAb+ positive patients were significantly younger than AAb− patients (UC p = 0.016, CD p = 0.017). AAb+ UC patients were significantly younger at the disease onset than AAb− UC patients (p = 0.021). CRP levels higher than 3 mg/L were seen in 21% of AAb− UC patients and 22% of AAb+ UC patients, while this level was 18% in AAb− CD patients and 38% in AAb+ CD patients. ESR (p = 0.009) and CRP (p = 0.008) levels were significantly increased in AAb+ CD patients compared to AAb− CD patients.  The presence of arthralgia was reported in 34 (18.7%) UC-patients and 50 (33.1%) CD-patients. Remarkably, in AAb+ IBD patients, the presence of arthralgia was not increased. Articular manifestations other than arthralgia were reported in few patients not allowing for any statistical analysis. Treatment in UC and CD consisted of (a combination of) immunosuppressive drugs, mesalazine, steroids, azathioprine, methotrexate, and anti-Tumor Necrosis Factor (TNF) agents. Anti-CCP2 and/or RF seropositivity was not associated with any of these therapies.

Discussion
Inflamed mucosal surfaces (e.g., intestinal tissue, lungs and periodontium) have been hypothesized as sites where the formation of ACPA initiates. The lungs [17,19] and periodontium [26,27] have already been studied as potential sites for AAb generation, while the intestinal area is still to be assessed as a site where autoantibody generation starts. The current study revealed that AAbs, especially ACPA of both IgA and IgG subclasses, are present in serum from IBD patients. ACPAs, especially of the IgA subclass, are increased in IBD patients compared to HCs, but do not exceed the diagnostic level for RA [24]. IgM RF and IgA RF seropositivity was comparable to that observed in the general population [28]. With regard to IBD patients with arthralgia, AAb+ IBD patients were significantly younger; the implications of this observation is not straightforward and was unexpected, as AAb positivity is more prevalent among the aged [29]. The higher ESR and CRP levels in AAb+ than in AAb− CD patients confirm the hypothesis that low positive ACPA levels are associated with inflammation. Although CD and UC are both inflammatory diseases, CRP is a less reliable marker of inflammation and disease activity in patients with UC, perhaps except for severe, extensive colitis [30].
Environmental and genetic factors are thought to play a role in the development of ACPA. Environmental factors such as smoking were shown as risk factors for developing ACPA-positive RA [15,31,32]. Smoking is considered to induce ACPA via the citrullination of proteins in the lungs by activating the enzyme peptidyl arginine deiminase (PAD). Our results showed that the AAb− and AAb+ IBD patients did not differ in smoking behavior. The presence of citrullinated proteins in colonic biopsies from IBD patients [12,33] might be on the basis of the higher systemic ACPA levels in IBD patients, similarly as smoking is able to induce citrullination in the lungs [34].
Genetic factors for ACPA development in RA, especially human leukocyte antigen DR beta 1 (HLA-DRB1) shared epitope (SE) alleles, are thought to play a role in the epitope spreading and rise of ACPA levels after the initiation of ACPA production, which is HLA-DRB1 SE independent [35]. In addition, bacterial antigens are hypothesized to contribute to ACPA development via molecular mimicry of self-antigens presented by predisposing HLA-DQ molecules [36]. However, HLA-associated genetic risk factors involved in IBD [37,38] are not involved in development of ACPA+ RA [39]. Therefore, we can speculate that due to the absence of these HLA-related risk factors, ACPA initiation in IBD is not followed by a maturation of the ACPA response and spreading of epitopes, which can be measured in the context of RA development. This is also reflected by the negative results on the reactivity against the four citrullinated peptides, although this assay is also less sensitive than the anti-CCP2 ELISA.
The results of our study show that extra-intestinal manifestations in IBD patients are not associated with arthritis autoantibody levels in serum. The presence of arthralgia was reported in a high number of patients (34 (18.7%) UC patients and 50 (33.1%) CD patients). Arthralgia was not clearly defined; therefore, it could have been that enthesitis and fibromyalgia were also in part responsible for these large numbers. Notwithstanding, our study confirms the results of other studies that found no association between articular manifestations and ACPA in IBD patients [21,22,40].
IBD is a condition which is particularly linked to gut microbiota, especially to gammaproteobacteria and the presence of Escherichia coli and fusobacteria [41][42][43][44]. The link between RA and the human microbiome is less clear [45]. The oral anaerobe bacterium Porphyromonas gingivalis has been hypothesized to play a role in RA development by creating citrullinated bacterial and/or human antigens with its own PAD-enzyme (PPAD) [46]. However, oral microbiome studies [47,48] did not find associations with P. gingivalis in the oral microbiome and the presence of RA. Thus, enzymatic activity of PPAD is probably not a major virulence mechanism during the early stages of inflammatory arthritis [49]. Studies on gut-microbiota in treatment-naive RA patients point towards an increased prevalence of Prevotella copri in early RA patients [50] as well as Lactobacillus salivarius [48]. Zhang et al [48] studied the relationship between the gut microbiome and autoantibody levels and found that Haemophilus spp. were negatively correlated with anti-CCP, while the presence of Prevotella spp. positively correlated with RF. Whether gut-microbiota also play a role in articular manifestations and autoantibody levels in IBD patients remains a topic for future research.
A limitation of the current study was that not all UC and CD patients were seen by a rheumatologist. Only in case of a suspect of arthritis and spondylarthropathies was the patient seen by a rheumatologist. Another limitation of the study was that we could not relate arthritis autoantibodies to gut microbiome data, which would have been very interesting, since such an association has been suggested in several studies.

Conclusions
In conclusion, arthritis autoantibodies are present in serum from IBD patients, which supports the hypothesis that the inflammation of intestinal mucosa induces low systemic levels of ACPA, although