Antibodies to citrullinated proteins are the most specific serological marker for rheumatoid arthritis (RA). They are associated with severity of disease and may occur years before clinical manifestations. It is unclear whether antibodies to citrullinated proteins react with any citrullinated protein or whether there are specific citrullinated proteins driving the autoimmune response. In this study we use the HL60 cell line as a substrate for identifying autoantigens in myelomonocytic cells, the major cell type present in the rheumatoid joint.
Citrullinated autoantigens were sought by immunoblotting with a screening panel of four RA and one healthy control sera, using lysates of HL60 cells, untreated or differentiated to monocytes or granulocytes. Each lysate was incubated with or without peptidyl arginine deiminase in vitro. Reactive bands were identified using two-dimensional gel electropheresis and MALDI TOF microsequencing. IgG antibodies to purified α-enolase, incubated with or without peptidyl arginine deiminase, were detected by immunoblotting with 53 RA and 40 healthy control serum samples. Immunohistochemistry of synovial sections and immunoblotting of lysed synoviocytes was performed using anti-α-enolase antibody diluted to 1:400. Macrophages stimulated with lipopolysaccharide or tumour necrosis factor were derived from primary monocytes, differentiated with monocyte colony-stimulating factor.
By immunoblotting, a band of 47 kDa in the deiminated lysates of HL60s differentiated to monocytes and granulocytes reacted selectively with the four RA sera (Fig. 1, sera A–D), but not with the control serum. It was identified as citrullinated α-enolase by microsequencing and confirmed by immunoblotting of two-dimensional subcellular fractions with an anti-α-enolase antibody and RA sera. In a larger panel of sera, 24 (46%) of the RA samples reacted with citrullinated α-enolase, of which seven also reacted with the uncitrullinated form. Six (15%) normal sera reacted with both forms. α-Enolase was demonstrated in the subsynovium and vascular endothelial cells in all synovial membrane biopsies. This was confirmed by immunoblotting where a protein in RA synoviocyte lysates, co-migrating with purified α-enolase, reacted with anti-α-enolase antibody. Stimulation of macrophages with lipopolysaccharide or tumour necrosis factor upregulated and induced secretion of α enolase. Secretion was inhibited with monensin.
Figure 1. Reactivity of rheumatoid arthritis and healthy serum with non-citrullinated/citrullinated lysates of HL60 monocytes. PAD, peptidyl arginine deiminase.
In this study, nearly one-half of the RA sera reacted with citrullinated α-enolase. This is a high frequency given the relative insensitivity of immunoblotting. We suggest that using a more refined ELISA-based assay, both diagnostic sensitivity and specificity will increase. The finding of antibody, together with abundant expression of antigen in the joint (which is amplified with inflammatory stimuli) makes citrullinated α-enolase a candidate autoantigen for driving the chronic immune response in RA.
This work was supported by the Arthritis Research Campaign (arc).