OP0313 Molecular analysis of anti-citrullinated protein antibody variable regions indicates aberrant selection processes during acpa b-cell development

Background Anti-citrullinated protein antibodies (ACPA) represent the most specific biomarker in Rheumatoid Arthritis (RA) and have been associated with RA pathogenesis. ACPA-IgG are heavily N-glycosylated in the variable domain. Recently, we showed that>80% of ACPA-IgG clones harbour N-glycosylation sites in their variable regions that result from somatic hypermutation (SHM). The reason for this remarkable phenomenon is incompletely understood. Elucidation of its molecular basis might provide insights into mechanisms by which ACPA-expressing B cells breach tolerance. Objectives To understand the molecular origin of ACPA variable domain N-glycosylation based on B-cell receptor (BCR) sequence analyses. Methods ACPA-expressing B cells were isolated from peripheral blood of 12 ACPA-positive RA patients using CCP2-streptavidin tetramers and fluorescence activated cell sorting. Full-length immunoglobulin (Ig) transcripts of heavy chains (HC) and light chains (LC) were obtained using anchoring reverse transcription of Ig sequences and amplification by nested PCR. Sequences were analysed for the degree of SHM and the presence of N-glycosylation sites (defined as sequences encoding N-X-S/T (X ≠ Proline) in the protein backbone). Sites that required a single nucleotide mutation to be generated were defined as s-SHM sites, whereas sites requiring multiple mutations were defined as m-SHM sites. IgG sequences of 12 healthy donors were used as control. Results 67% of ACPA-Igκ LC and 47% of ACPA-Igλ LC contained ≥1 n-glycosylation sites compared to 82% of ACPA-IgG HC. Nucleotide mutation rates were similar for ACPA-Igκ LC and ACPA-Igλ LC (88.2%±5.7% and 87.3%±5.3% similar to germline, respectively) and lower compared to the mutation rate of ACPA-IgG HC (82.4%±5.8% similar to germline). The distribution of sites in ACPA-Igκ LC and ACPA-IgG HC was similar, with most sites located in framework region (FR) 3 (42% and 49%, respectively). In contrast, 65% of all N-glycosylation sites in ACPA-Igλ LC clones were located in FR1 and only 7% were located in FR3. Furthermore, 26% of all N-glycosylation sites in ACPA-IgG HC were m-SHM sites compared to 15% in IGHV-matched IgG clones derived from healthy donors. 28% and 44% of all N-glycosylation sites were m-SHM sites in ACPA-Igκ LC and ACPA-Igλ LC, respectively. No correlation was observed between the number of nucleotide mutations and the number of total N-glycosylation or m-SHM sites in ACPA clones. Conclusions Our analyses revealed an abundance of N-glycosylation sites in ACPA-IgG HC, ACPA-Igκ LC and ACPA-Igλ LC. N-glycosylation sites in ACPA are frequently m-SHM sites. Intriguingly, the generation of such sites requires multiple somatic mutations suggesting that m-SHM sites in specific positions in ACPA variable regions could be advantageous for the survival of ACPA-expressing B cells. This indicates that the introduction of N-glycosylation sites might be a selective process that could allow these B cells to escape from putative tolerance checkpoints. Disclosure of Interest None declared