Identification of novel autoantibodies in Sjögren's disease

Abstract

Introduction: The diagnosis of Sjögren's disease (SjD) in patients without autoantibodies against Ro/SSA is a major challenge. We aimed to identify novel autoantibodies in SjD that may facilitate the diagnostic procedure for Ro/SSA negative SjD.

Methods: IgG and IgA autoantibody reactivity of 94 potential candidate autoantigens for SjD, selected from a discovery screen of 1,629 human antigens coupled to Luminex beads and prior knowledge about potential biological relevance, were examined in serum of SjD patients (n=347) using Luminex and ELISA technology. Healthy (HC, n=118) and non-Sjögren's sicca syndrome (NSS, n=44) individuals served as controls. To assess disease specificity, the novel autoantibodies were also measured in serum of patients with Rheumatoid Arthritis (RA, n=50), Systemic Lupus Erythematosus (SLE, n=49), and Systemic Sclerosis (SSc, n=37).

Results: 45 novel autoantibodies were significantly (p ≤ 0.05) more prevalent in SjD than in HC and were detected in up to 19% of the SjD cohort. The most common autoantibodies were against CCL4, M5, TMPO and OAS3. Some of the novel autoantibodies were associated with extraglandular disease manifestations, such as anti-TONSL or anti-IL6 with pulmonary involvement. We have developed a three and five marker panel for the detection of Ro/SSA negative patients, consisting of anti-FNBP4, anti-SNRPC, anti-CCL4, anti-M3 and anti-KDM6B, which had a sensitivity of up to 46% with a specificity of 95% (SjD vs. HC). Both panels discriminate these patients from HC, whereas the three-marker more effectively differentiates between Ro/SSA negative patients and NSS.

Discussion: Novel autoantibodies will facilitate the diagnosis of Ro/SSA negative patients with SjD, in particular our predictive panel will be useful in the diagnosis and differentiation of these patients from healthy and NSS individuals in a clinical context. In addition, the autoantibodies may also be useful for risk stratification of extraglandular manifestations.

Keywords: Sjögren’s disease; autoantibodies; biomarkers; connective tissue diseases; extraglandular manifestations; seronegative patients; sicca syndrome.

Figure 1

Heatmap of binarized IgG and IgA autoantibody reactivity of 72 antigens in primary Sjögren’s disease (SjD, n=347), separated into Ro/SSA positive (n=257) and negative (n=90) SjD patients. Positivity of individual patients for single or both isotypes is indicated by color code. Patients were ordered by unsupervised hierarchical cluster analysis and prevalence of each antigen in all SjD patients in percent have been added as bar chart.

Figure 2

Combination of complementary autoantibodies into a panel increases diagnostic capability for identification of Ro/SSA negative SjD patients (n=90). (A) Co-prevalence heatmap of common IgG autoantibodies in seronegative SjD patients. The numbers in the cells represent the percentage of positive patients, for individual antibodies (diagonal) and the co-prevalence of two-marker-combinations. (B) Heatmap of binarized autoantibody reactivity of selected IgG antigens in Ro/SSA positive (n=257) and negative (n=90) SjD patients as well as healthy controls (n=118). The heatmap color is related to the binary outcome in antibody measurement for each patient. Patients within groups were ordered by unsupervised hierarchical cluster analysis and top annotation shows SjD panel prediction outcome, with black indicating a SjD diagnosis. (C) Box and whisker plots showing the log2 MFI of panel marker (FNBP4, SNRPC, CCL4, M3, KDM6B) in individual sample groups (HC, Ro/SSA-, Ro/SSA+). Horizontal lines indicate cut-off for data binarization..

Figure 3

Association of autoantibodies in primary Sjögren’s disease patients (SjD, n=289) with clinical characteristics. Forest plot of logistic regression results for antibodies with significant association (p-value ≤ 0.05) and meaningful effect size (|coefficient| ≥ 0.45).