The most prominent autoantibodies in rheumatoid arthritis, ACPA (anti-citrullinated protein antibodies), were recently found to harbour a remarkably high amount of glycans in the variable domain. For these glycans to be expressed, ACPA B cells need to introduce glycosylation sites in the B cell receptor (BCR) through somatic hypermutation. DNA sequence analysis of the ACPA BCR repertoire showed that over 90% of the sequences contained a glycosylation site in the variable domain, whereas tetanus-specific B cells from RA patients showed no glycosylation sites. This striking observation has been confirmed on protein level using HILIC-UHPLC, showing that 93% of APCA-IgG bears one or more glycans in the variable domain.
Notably, we now demonstrated that increased ACPA variable domain glycosylation is exclusive to RA patients and healthy individuals transitioning to RA, while absent in ACPA+ individuals that remain healthy, suggesting a direct link between glycosylation and disease pathogenesis.
We furthermore showed that 1) introduction of these glycosylation sites was not due to random accumulation, 2) glycosylation sites are rarely located in the antigen binding site and 3) variable domain glycans do not improve affinity for autoantigens. This implies that introduction of ACPA variable domain glycans is driven by other selection mechanisms than classical antigen selection.
On DNA level, others have recently shown that increased variable domain glycosylation-sites are also found in the total BCR repertoire of patients with other autoimmune diseases than rheumatoid arthritis. Using HILIC-UHPLC and nano-LC-MS, we are currently investigating the glycosylation profile of these other autoantibody types. This would reveal whether these glycosylation-sites are (solely) derived from autoantibodies, and whether these sites are actually glycosylated.
Understanding the strong selection on variable domain glycosylated autoantibodies could be the key to understanding the breach of tolerance and one of the origins of autoimmunity.