Autoantibody Clusters and SIGLEC1 are Predictive of Systemic Lupus Erythematosus Development

Objectives Up to one-third of patients with suspected SLE progress to definite disease, but reliable predictive markers are lacking. Previously, we developed a Lupus Lymphocyte Activation Score (LLAS) based on key cell subsets (transitional B cells, age-associated B cells, plasmablasts, Tph cells, and Tfh cells) observed in those with or at risk of developing SLE.[1] Here we examined the relationship between autoantibody profiles, blood sialic acid binding Ig like lectin 1 (SIGLEC1, a novel biomarker of macrophage type I interferon), LLAS, and systemic autoimmune rheumatic disease (SARD) development among those with suspected disease.

Methods Blood samples were collected from 45 patients with new onset (< 5 yrs), positive ANAs, and suspected SLE at Brigham and Women’s Hospital Lupus Center. At baseline, none met SLE or other SARD classification criteria. All received prednisone <10mg/day and no immunosuppressants. Using baseline samples, soluble SIGLEC1 was measured by ELISA and a comprehensive autoantibody profile was performed: ANA by indirect immunofluorescence assay on HEp-2 cells and patterns classified according to the International Consensus on ANA Pattern anti-cell (AC) nomenclature, SLE-related autoantibodies, including anti-DFS70 by a fully automated multi-analyte system using particle-based multi-analyte technology, and anti-C1q and anti-phospholipid antibodies by ELISA. To examine patterns of autoantibodies, we used hierarchical clustering and investigated their relationships with disease progression, LLAS (sum of standardized proportions of each of 5 lymphocyte subsets above), and SIGLEC1 levels.

Results Of 45 patients, 36 had multiple visits with a mean follow-up of 13.6 months. In follow-up, 3 patients were diagnosed with SLE and met 2012 SLICC or 2019 EULAR/ACR criteria; 4 were diagnosed with or suspected of new dermatomyositis or Sjögren disease. We identified 5 clusters of suspected SLE patients based on autoantibody profiles (Figure 1A). Cluster A (AC-4 nuclear speckled with multiple autoantibody reactivity including anti-RNP, -dsDNA, -anti-Ro60/SSA) was more likely to progress or develop a CTD including SLE (OR 1.94, 95% CI 0.42-3.46) than others (Figure 1B), particularly compared to Cluster C (no autoantibodies). Cluster A also had significantly higher SIGLEC1 vs Cluster E (Figure 1C). There was no difference in LLAS by autoantibody cluster (Figure 1D).

• Download figure
• Open in new tab
• Download powerpoint

Fig 1A) Five autoantibody clusters (A-E) based on baseline samples of 45 suspected SLE (L to R): Cluster A (n=11) nuclear speckled pattern (AC-4) with the greatest autoantibody reactivity including anti-RNP, anti-dsDNA, and anti-TROVE2/Ro60, Cluster B (n=4) nuclear large speckled pattern (AC-5), Cluster C (n=9) negative ANA (AC-0) with rare autoantibodies at low titers, Cluster D (n=9) isolated anti-DFS70 antibody with dense fine speckled (AC2) pattern, and Cluster E (n=12) nuclear speckled with mitotic plate staining (AC-30). Fig 1B) Cluster A (AC-4 nuclear speckled pattern and multiple autoantibody reactivity) had largest N with disease progression +/− definite SLE or other SARD at follow-up. *denotes a significant difference in proportion of patients who had disease progression +/− definite SLE or other CTD at follow-up (Cluster A vs. C, diff. 43.43%, 95%CI 0.08%-0.79%). Fig 1C) 5 autoAb clusters of suspected SLE patients and median SIGLEC1. Cluster A had higher median SIGLEC1 than others. Fig 1D) Lupus Lymphocyte Activation Score (LLAS) not statistically different across autoAb clusters

Conclusion Autoantibody profile characterized by nuclear speckled with multiple autoantibody reactivity, including anti-RNP, -dsDNA, and -Ro60/SSA, was predictive of SARD development. This cluster also had higher baseline SIGLEC1 vs those with monospecific DFS70 antibodies or no autoantibodies. Further analyses will assess whether LLAS, SIGLEC1, and autoantibodies are complementary, and when combined, could enhance the prediction of SLE development.

References [1.] Horisberger A. [Abstract]. Arthritis Rheumatol 2024;76 Suppl 9.