Background/Purpose: Severe lupus tubulointerstitial nephritis (TIN) is prognostic of renal failure and characterized by tertiary lymphoid organogenesis. We have previously demonstrated that the dominant target of monoclonal antibodies (mAbs) cloned from B cells isolated from TIN biopsies was vimentin and not nuclear antigens. Probing bovine lens vimentin coated arrays confirmed anti-vimentin antibody (AVA) reactivity and serum IgG AVA titers correlated with TIN severity. Vimentin is an abundant secreted antigen in inflamed tissue. Citrullinated vimentin comprises a low proportion of the antigen, yet high titers of anti-citrullinated protein antibodies (ACPAs) are specific features of rheumatoid arthritis. However, it is not known if in situ selected AVAs in TIN target citrullinated vimentin. Furthermore, the origin of AVAs, and how they are selected in situ, is not clear. 

Methods: Recombinant human vimentin was made in E.coli and used as substrate by ELISA. Eight TIN mAbs (with high vimentin reactivity) from six different TIN biopsies were probed for reactivity with unmodified and in vitro citrullinated vimentin. Predicted germ-line reversions of the respective TIN mAbs, 23 mAbs from healthy naïve and 20 mAbs from anergic B-cells, were also assayed for vimentin reactivity. Lupus serum samples from 101 subjects were screened for IgM, IgG, IgA and IgE AVAs. Other routinely assayed autoantibodies were titrated using commercially available ELISAs.

Results: In vitro citrullination of human vimentin, confirmed by mass spectrometry, inhibited TIN mAb binding. Germline reversion of seven of eight somatically hypermutated TIN mAbs diminished but did not ablate AVA activity. Vimentin immunoreactivity was, surprisingly, a common feature of the naïve and anergic repertoires. Furthermore, the range of relative affinities between reverted AVAs and anergic antibodies with AVA activity was similar. IgM and IgA AVAs could be readily detected in the serum of some patients with SLE. High titers of one AVA isotype did not categorically determine high titers of another isotype.

Conclusion: These studies demonstrated that the AVAs associated with SLE and the ACPAs associated with RA are different. It also appeared that AVAs are selected in situ from preexisting repertoires of AVA and are common through the spectrum of antibody isotypes associated with SLE. These data suggest that loss of tolerance to vimentin, and selection of high affinity and potentially pathogenic AVAs, is a common feature of SLE.