Background Perforin- and complement-induced cell lysis has recently been implicated in the generation of citrullinated autoantigens in rheumatoid arthritis (RA). By activating PADs in target cells, immune-mediated pore-forming pathways induce abnormal protein citrullination in the RA joint. Here, we demonstrate a novel approach to identify unique autoantibody reactivities to citrullinated proteins generated by the perforin pathway.
Objectives To identify novel perforin-induced autoantigens in RA, and to characterize the antibody response to citrullinated hnRNP A1 (cit-hnRNP A1) in patients with RA.
Methods 293T cells transiently transfected to express human PAD2 or PAD4 were incubated with perforin from cytotoxic granule contents (GC) to induce PAD activation. Cellular citrullination was confirmed by anti-modified citrulline (AMC) immunoblotting. Inducible autoantigens generated during perforin-mediated cytolysis were identified by immunoblotting using RA patient sera. Antigen identity was determined by 2D gel electrophoresis/mass spectrometry (MS). Recombinant hnRNP A1 was expressed in E. coli. Patients with RA (n=196) from a longitudinal cohort study (ESCAPE RA) and healthy controls (n=56) were assayed for IgG antibodies to unmodified and in vitro citrullinated hnRNP A1 by ELISA. Positivity was defined as an arbitrary unit (AU) above the 98th percentile of controls. Anti-cit-hnRNP A1 antibody specificity was determined by subtracting reactivities measured against unmodified hnRNP A1 from reactivities against cit-hnRNP A1.
Results 293T cells expressing PAD2 or PAD4 were used to study PAD-isotype specific autoantigen citrullination. A 34-kDa autoantigen inducible in both PAD2- and PAD4-transfected cells was distinctly targeted by RA patient sera (Fig. 1A; GC: perforin-treated cells, NT: untreated controls). MS identified the modified protein as cit-hnRNP A1, and citrullination by both rPAD2 and rPAD4 was confirmed using purified hnRNP A1. Antibody reactivity against cit-hnRNP A1 was detected in 70 (36%) of RA patients vs. only 1 (2%) of controls (p<0.0001: Figure 1B; HC: healthy controls), and anti-cit-hnRNP A1 specific antibodies were confirmed in 65 (33%) of RA patients. In contrast, the low-titer reactivities against unmodified hnRNP A1 observed in patients with RA did not differ significantly from controls (7% vs. 2%; p=0.32) (Fig. 1B). Anti-cit-hnRNP A1 positivity in RA was associated with markers of erosive disease as measured using the Sharp-van der Heijde score (median 12 vs. 6 units, respectively for antibody positive vs. negative; p=0.019) and total erosion score (median 6 vs. 2 units, respectively; p=0.008) despite equivalent therapeutic intervention. Interestingly, this marker uniquely associated with elevated levels of IL-6 (58% higher in antibody positive vs. negative; p=0.01).
Conclusions These studies identify hnRNP A1 as a citrullinated autoantigen in RA generated during perforin-mediated cytotoxicity. Anti-cit-hnRNP A1 antibodies may serve as a novel biomarker associated with mechanisms of joint damage in a subset of RA.
Disclosure of Interest None declared