POS0004 Pro-inflammatory anti-modified protein autoantibodies derived from bone marrow, synovial, and lung B cells preferentially target malondialdehyde-acetaldehyde cross-linked molecular structures

Background: Autoantibodies to malondialdehyde (MDA) proteins constitute a subset of anti-modified protein autoantibodies (AMPA) in rheumatoid arthritis (RA), which is not directly overlapping with citrulline reactivity. We have previously shown that IgM anti-MDA is prominent in the natural antibody repertoire at birth and may have homeostatic roles. Yet, IgG anti-MDA is elevated in RA and correlates with DAS28, CRP, IL6 and TNF-α [1]. Moreover, RA-derived anti-MDA monoclonal antibodies induce bone erosion and osteoclastogenesis by modulating cell metabolism [2]. MDA is an oxidation-associated reactive aldehyde that together with acetaldehyde (AA) mediates formation of a range of amino acid adducts including linear MDA-lysine, fluorescent MAA-lysine, and intramolecular cross-linking. MDA/MAA proteins carry multiple adduct-variants, and we here propose that distinct subsets of anti-MDA/MAA antibodies, defined based on the MDA-induced molecular structures recognized, have different pathogenic properties. Moreover, we have investigated the presence of class-switched MDA+ B cells in RA synovium, bone marrow, and bronchoalveolar lavage (BAL).

Objectives: Identification of RA MDA+ B cells from different compartments and distinct MDA/MAA reactivity patterns.

Methods: Human recombinant mAbs were generated from single-cell sorted B cells from BAL and bone marrow (n=148) and screened for reactivity to MDA and MAA modified antigens. MDA-cyanoborohydride and MDA/MAA-hapten proteins without cross-linking were used for characterization. Serology screening of IgG reactivity to MAA-BSA was performed as previously described [3].

Results: IgG MAA-reactivity is elevated in 49% of ACPA+ RA patients (n=278) compared to 14% of population controls (n=437). We did not detect any significant increase in IgG anti-MAA in ACPA+ RA-risk individuals at baseline,14% had high anti-MAA (n=263), compared to controls. But for risk-individuals that developed arthritis, anti-MAA levels increased at onset of arthritis with 26% displaying elevated levels (n=53). We identified IgG+ MDA+ cells among RA bone marrow plasma cells and BAL B cells from both early RA and ACPA+ RA-risk, which could be compared to three previously identified RA synovial clones. Seven out of eight investigated clones carried somatic hypermutations (SHM). Interestingly, all clones with SHM were found to specifically target MAA cross-linked structures rather than MDA- or MAA-hapten, while the germline-encoded synovial clone F04 bound strongly to linear MDA-lysine in proteins and peptides. Germline converting resulted in reduced MDA/MAA binding, suggesting different origin of the clones compared to F04. All binding was independent of protein backbone. Yet, only a subset of MAA-selective clones with SHM induced robust TNF-α expression in monocyte-derived macrophages and enhanced osteoclastogenesis.

Conclusion: MDA/MAA positive B cells can be identified in bone marrow and synovium in established RA and in the lungs of early and RA-risk individuals, but serum levels are only increased in conjunction with onset of arthritis and with inflammation. SHM and class-switching suggest affinity maturation and T-cell dependence of these responses. Distinct targeting of different MDA/MAA molecular structures was observed. Future studies will elucidate how MDA recognition patters are related to the anti-MDA inflammatory and bone-modulating properties and disease stages.

References: [1]Grönwall, C., et al J Autoimmun. 2017 84:29-45

[2]Sakuraba, K., et al J Autoimmun. 2022 133:102903

[3]Grönwall, C., et al Front Immunol. 202112:627986

Acknowledgements: NIL.

Disclosure of Interests: None Declared.