Objective Cell surface glycosylation serves as a binding platform for endogenous immunoregulatory lectins such as galectins (Gal-s) or sialic acid-binding immunoglobulin-like lectins (Siglec-s). The expression on monocytes and serum levels of Siglec-1 correlate with disease activity in SLE. We have previously shown that activated T-cells from active lupus patients are resistant to the immunoregulatory activity of Gal-1 because of its impaired binding due to T-cell surface hypersialylation. In this study, we aimed at a detailed characterisation of the cell surface glycome composition and lectin binding in multiple white blood cell subsets.

Methods We collected peripheral blood mononuclear cells (PBMCs) from 18 new or relapsing SLE patients with active disease (mean age: 47.5/26–83/years, SLEDAI-2K:15.8/6–34/, anti-dsDNA level: 76.4/8–220/) IU/ml) and 18 healthy age-matched controls. Multicolour antibody panel was designed to identify a total of 17 peripheral immune subsets of T-, NK-, NKT-, B-cells and monocytes. The binding of 5 fluorochrome-labelled lectins – Gal-1, Gal-3, Siglec-1, Aleuria aurantia lectin (fucose-labelling) and Sambucus nigra agglutinin (SNA – sialic acid-labelling) – to each of these populations was assessed simultaneously by flow cytometry at resting state and following 72 hours of cell-activation. The differential distribution of cell subpopulations was further analysed with the flow self-organising map (FlowSOM) unsupervised algorithm.

Results Within both CD8^high and CD8^dim T-cells, there was a shift to highly activated, hyperglycosylated, in particular, hypersialylated cells with high Siglec-1 and Gal-1 binding in SLE patients as compared with controls. In CD4/CD8 double-negative T-cells, an activated subpopulation with the highest degree of glycosylation and lectin-binding was significantly more prevalent in lupus patients. Within CD56^high NK cells, two activated, but less glycosylated metaclusters with low Siglec-1 binding were enriched. Binding of all five lectins to memory B-cells and HLA-DR expression was lower in SLE than in controls, and plasmablasts bound less galectins in SLE. Intermediate monocytes were more prevalent in lupus patients, with an increase in a non-activated, CD14^highCD16^low, Siglec-1^medGal1^med subpopulation (p<0.05 for all data presented).

Conclusion Glycosylation and lectin-binding patterns, analysed in conjunction with the activation state, and broken down to immune cell subsets have revealed several distinct and potentially lupus-specific cell subpopulations.