IgA is the predominant isotype of anti‐β2 glycoprotein I antibodies in rheumatoid arthritis

Abstract Background The aim of this study was to determine the frequency of anti‐cardiolipin antibodies (aCL) and anti‐β2 glycoprotein I antibodies (aβ2GPI) among Tunisian patients with rheumatoid arthritis (RA). Methods Ninety RA patients with positive anti‐cyclic citrullinated antibodies (anti‐CCP) and 90 healthy blood donors (HBD) were studied. aCL and aβ2GPI of isotype IgG, IgA and IgM were detected by ELISA. Result The frequency of antiphopholipid antibodies (aPL) (aCL and/or aβ2GPI) was significantly higher in patients with RA than in HBD (35.5% vs 11.1%, P = .0001). The frequencies of aCL and aβ2GPI were significantly higher in patients than in healthy subjects (15.5% vs 5.5%, P = .04 and 32.2% vs 11.1%, P = .0005 respectively). aβ2GPI‐IgA were significantly more frequent in patients than in the control group (26.7% vs 7.8%, P = .0007). In patients, aβ2GPI‐IgA were significantly more frequent than aβ2GPI‐IgG (26.7% vs. 6.7%, P = .0003) and aβ2GPI‐IgM (26.7% vs 5.6%, P = .0001). In RA patients, the frequency of aβ2GPI was significantly higher than that of aCL (32.2% vs 15.5%, P = .008). aβ2GPI‐IgA was significantly more frequent than aCL‐IgA (26.7% vs 4.4%, P = .00005). The average titer of anti‐CCP in aPL positive patients was significantly higher than in aPL negative patients (170.6 ± 50 RU/mL vs 147.7 ± 51 RU/mL, P = .04). Significant correlation was found between aβ2GPI‐IgA and anti‐CCP (r = .235, P = .026). Conclusions aPL and particularly aβ2GPI‐IgA are frequent in RA and are correlated with anti‐CCP.


| INTRODUC TI ON
Rheumatoid arthritis (RA) is a chronic, inflammatory joint disease of autoimmune nature. The ethiopathogenic mechanisms involved are complex and include gut dysbiosis. 1 RA is characterized by autoantibodies production (rheumatoid factor (RF) and anti-citrullinated protein antibody (ACPA)). RA can lead to accumulating joint damage and irreversible disability. 2 Antiphospholipid antibodies (aPL) are a heterogeneous group of antibodies that have been associated with thrombotic or obstetrical events in patients with antiphospholipid syndrome (APS). 3 These antibodies can occur not only in APS but also in a variety of autoimmune, malignant, and infectious diseases. 4 In fact, the definition of clinically significant aPL positivity is not well established. 3 The most commonly detected aPL antibodies are lupus anticoagulant, anti-cardiolipin antibodies (aCL) and anti-β2 glycoprotein I (aβ2GPI). In RA, several studies determined the frequency of aCL and aβ2GPI. [5][6][7][8][9][10][11] However, to our knowledge, aβ2GPI-IgA has been determined in only three studies. [7][8][9] Furthermore, the frequency of aPL antibodies is not known in RA in Tunisia. So, the aim of our study is to evaluate the frequency of aCL (IgG, IgA, IgM) and aβ2GPI (IgG, IgA, IgM) in a cohort of RA patients without looking for APS.

| Patients
In our retrospective study, sera of 90 RA patients, with positive anticyclic citrullinated antibodies (anti-CCP), were included from the database of our Immunology laboratory. Sera were collected between 2017 and 2018 from four hospitals in the center of Tunisia. Patients were diagnosed with RA according to American College of Rheumatology/ European League Against Rheumatism (ACR/EULAR). 12 Sera of sex-matched 90 healthy blood donors (HBD) served as normal controls. All sera of control group were tested for anti-CCP and RF.
All sera were stored at −80°C until the use. Ethical committee of our hospital gave approval for this study.

| aCL assays
Serum samples were evaluated for aCL-IgG, IgA, and IgM by using a commercial enzyme-linked immunosorbent assay (ELISA) (Orgentec Diagnostika ® ) as we have described it previously. 13 Results were expressed as arbitrary units with a cutoff of positivity of 10 U/mL for IgA and IgG and 7 U/mL for IgM following the manufacturer's instructions.

| aβ2GPI assays
The determination of aβ2GPI IgG, IgA, and IgM were carried out with a commercial ELISA (Orgentec Diagnostika ® ) using a purified human β2GPI as we have described it previously. 13 Results were expressed as arbitrary units with a cutoff for positivity of 8 U/mL following the manufacturer's instructions.

| RF assays
Serum samples were evaluated for IgG, IgA, and IgM-FR by using a commercial ELISA (Orgentec Diagnostika ® ) as we have described it previously. 14 Results were expressed as arbitrary units following the manufacturer's instructions.

| Anti-CCP assays
Anti-CCP was detected by using a commercially available secondgeneration ELISA (Euroimmun ® ) as we have described it previously. 15 Results were expressed as arbitrary units with a cutoff for positivity of 5 RU/mL according to the manufacturer's instructions.

| Statistical analysis
The comparison of frequencies of aPL was performed using Chisquare or Fisher's test. The variables were tested for normality using the Kolmogorov-Smirnov test. To compare the mean titer of anti-CCP between positive and negative aPL patients, we used a parametric Student's t test. Correlation study between aβ2GPI-IgA and anti-CCP was done by calculating Spearmans's correlation coefficient. A P-value <.05 was considered significant.

| RE SULTS
The characteristics of patients and normal controls are presented in Table 1. aCL and aβ2GPI frequencies are summarized in Table 2. The frequency of having any type of aPL (aCL and/or aβ2GPI) was significantly higher in patients with RA than in HBD (35.5% vs 11.1%, P = .0001).

Control group (n = 90)
Distribution of titers of aCL and aβ2GPI in positive aPL patients is presented in Figure 2.

| Frequencies of aCL-IgG, IgA, and IgM
The frequency of aCL (IgG, IgA, or IgM) was significantly higher in RA patients than in controls (15.5% vs 5.5%, P = .04).

| Frequency of aPL according to sex
In RA patients, the frequency of aPL was not statistically different between females and males (32.1% and 41.2%, respectively) (

| Association between aPL and RA antibodies
The average titer of anti-CCP in aPL positive patients was significantly higher than in aPL negative patients (170.6 RU/mL ± 50 vs 147.7 ± 51 RU/mL, P = .04) (Figure 1).
No significant difference was found in the average titer of RF (IgG, IgA, or IgM) between positive and negative aPL patients.
Significant correlation was found between titers of aβ2GPI-IgA and titers of anti-CCP (r = .235, P = .026). and also in Afro-Americans. 20 Alessandri et al 21  there is a similarity between these two diseases. Interestingly, dysbiosis of gut microbiota was described not only in RA 1,22 but also in APS. 23 This dysbiosis induces not only protein citrullination 22  **Comparison between aβ2GPI-IgG and aβ2GPI-IgA (P = .0003).
a conformational change of β2GPI, that exposes a cryptic epitopes in domain I of β2GPI 23 and therefore aβ2GPI synthesis. 24 In our RA patients, we found a high frequency of aβ2GPI and a correlation between aβ2GPI-IgA and anti-CCP. So, the question arises: are aβ2GPI-IgA implicated in the pathogenesis of arthritis in RA? During RA, gut microbiota dysbiosis may cause the activation of innate-like T cells, which can be skewed toward a pro-inflammatory state and contribute to inflamed joint tissue. 25 Aberrant epigenetic changes (histone modifications, DNA methylation, and miRNAs) are implicated in inflammatory joints of RA. 26 Moreover, phospholipid transfer protein is highly expressed in joints and its activity in synovial fluid is elevated and correlated with pro-inflammatory cytokines (Il-1 β, Il-6) and, therefore, it may directly trigger inflammation. 27 Surprisingly, during joint inflammation, enzymatically activated β2GPI is transformed from closed conformation to an open hockey stick-like conformation. The resulting aβ2GPI is responsible for cartilage degradation of phospholipid bilayers and, therefore, boundary-lubricating ability is deactivated. 28 Moreover, through multiple mechanisms, aPL activity results not only in vasculopathy, thrombosis, and pregnancy complications but also in inflammation. 3 So, could aβ2GPI be both the cause and the consequence of the inflammation in the synovial joints?
Gut microbiota dysbiosis 1,22 is associated with an intestinal barrier dysfunction. 29 Alterations in gut permeability may allow intraluminal compounds entry the mucosal site, and this may be a trigger cause of an autoimmune reaction. In the gut, there is not only microbiota but also mycobiota (fungal community). 30 Two major genera in the mycobiota were Candida and Saccharomyces. 31 Because of a leaky gut, Saccharomyces cerevisiae arrives to the mucosa and induces the synthesis of antibodies to Saccharomyces cerevisiae named ASCA. 30 ASCA has been described in RA. 32 Interestingly, cross-reactive epitopes on β2GPI and the phosphopeptidomannan part of the cell wall of Saccharomyces cerevisiae have been described. 33 In the same way, we have previously demonstrated Our study presents some limitations: 1-It is a retrospective one, so we do not have data on clinical manifestations and correlation between aβ2GPI-IgA and any clinical feature of RA could not be studied. 2-Our study lacks an experimental demonstration on a possible pathogenic mechanism of aβ2GPI in RA.

| CON CLUS ION
In conclusion, we found a significantly higher frequency of aβ2GPI in RA patients in comparison to the healthy subjects and we tried to explain why these antibodies are produced in RA. We could hypothesize, as said Hippocrates "all disease starts in the gut", that RA begins in the gut by: (a) Microbiota which induces joint inflammation, protein citrullination, aβ2GPI synthesis, and intestinal barrier dysfunction. (b) Mycobiota which induces synthesis of antibodies (ASCA) who recognize self antigens such as β2GPI and citrillunated proteins. In Tunisia, stress, 36 smoking, 37 and high prescription of antibiotics 38 trigger gut microbiota dysbiosis and high bread consumption trigger a mycobiota rich in Saccharomyces cerevisiae. All these factors combined with a high frequency of consanguineous marriage 39 could explain the high frequency of RA in our country.

CO N FLI C T O F I NTE R E S T
None of the authors have conflicts of interest to declare.