Diffuse large B cell lymphoma in rheumatoid arthritis express polyreactive and autoreactive BCRs binding protein modifications

Background An important risk factor for diffuse large B-cell lymphoma (DLBCL) development in patients with rheumatoid arthritis (RA) is high cumulated disease activity. B cell intrinsic factors and autoimmune processes may contribute, and we hypothesize that BCR autoreactivity may drive lymphomagenesis and/or lymphoma cell survival. RA is primarily associated with anti-citrullinated protein autoantibodies but patients can also have other anti-modified protein autoantibodies, e.g. towards malondialdehyde acetaldehyde (MDA/MAA) adducts which correlate with disease activity and inflammation. Here, we investigated DLBCL properties from RA patients with concurrent lymphoma by B-cell receptor (BCR) next generation sequencing and re-expression. Methods In total 22 RA-DLBCL patients were included from the Swedish AUTO-LYMPHOMA cohort. Tumour DNA was extracted from fixed tissues and full length VH/VL fragments were amplified by multiplex PCR and sequenced using Illumina. Computational analysis assessed clonal variations. Dominating sequences were synthesized and expressed as recombinant monoclonal antibodies for antigen binding analysis. Results Complete VH/VL lymphoma tissue sequences were generated from nine RA-DLBCL patients. Clonotype diversity within the tumour was generally low and the dominant paired heavy and light chain BCR sequences were obtained for six patients. For three patients, amino acid variations in the DLBCL clonotype indicated intra-clonal evolution. Two of six patients had tumour BCRs with few VH and VL somatic hypermutations (≤ 5 mismatches) while three had high mutations (> 15 VH or VL mismatches). No bias in VH/VL usage was detected. Six lymphoma BCRs from five patients were expressed as recombinant monoclonal antibodies and their binding to autoantigens was assessed by ELISA and a solid-phase microarray (> 400 antigens). All patient-derived lymphoma BCRs were CCP2 negative, but five bound MDA/MAA-protein. Three of these displayed polyreactive binding by in-house assays of which two further displayed strikingly broad polyreactivity on the antigen array. Importantly, for one patient, the lymphoma BCR had high specificity to MDA/MAA-protein, and another patient DLBCL-BCR had binding only to methylated protein. Conclusions The data suggest a possible link between polyreactivity and anti-MDA/MAA autoreactivity in RA-DLBCL lymphomagenesis. We hypothesize that high inflammation burden could increase antigen load which may trigger polyreactive and modified-protein B cell responses and/or maintain lymphoma cell survival.