RA Sera show no specificity for Antibodies to Glucose-6-Phosphate Isomerase
Glucose-6-phosphate isomerase (GPI) is a ubiquitous enzyme in the cytoplasm of all cells. In the K/BxN mouse, GPI is the target of arthritogenic antibodies in the presence of a particular class II allele of the major histocompatibility complex, resulting in a chronic, progressive and destructive arthritis. Prior studies have found anti-GPI antibodies in the serum of patients with rheumatoid arthritis (RA) as well as GPI deposits in sections of inflamed synovium. In this study, Matsumoto et al (Arthritis Rheum 48:944, 2003) evaluate serum samples from patients with different arthritic and autoimmune diseases to look for prevalence of anti-GPI antibodies in this disease set.
Methods: Serum samples from 811 patients with 12 different conditions in several independent cohorts were assayed for antibodies to GPI by enzyme-linked immunosorbent assay (ELISA). A standardized serum pool from arthritic K/BxN mice was used as a control allowing comparison between experiments. Two forms of GPI antigen were tested for reactivity in the ELISA, a recombinant human GPI expressed in E coli and a native GPI purified from rabbit liver.
Results: Reactivity to GPI was detected in the serum of many individuals, however no specificity was found for RA compared with other autoimmune diseases, including PsA, SLE, JRA, Crohns, Sjogrens, sarcoidosis, viral, undifferentiated, and spondyloarthropathy. Among the RA patients, no association was observed with rheumatoid factor status. However, slightly higher positive GPI titers were observed in RA patients with active disease compared to patients with relatively controlled RA.
Conclusion: Although GPI may be a target of autoimmune attack in some patients, no disease-specific association was observed in this study. Specifically, no clear prevalence or specificity of anti-GPI antibodies was found in patients with RA.
Editorial Comment: These data contradict the results published in 2001 by Schaller et al (Nature Immunol 2:746, 2001) that showed a high correlation between the presence of anti-GPI antibodies and RA. In the current study by Matsumoto et al, a broad spectrum of patients was tested and, in the arthritis groups, a broad spectrum of disease activity was represented. Furthermore, they used two separate sources of GPI and presented detailed quality control data. Therefore, their negative results showing no specificity of anti-GPI autoantibodies for human RA would seem to be definitive. The authors do not rule out the possibility that the presence of anti-GPI antibodies could identify a very specific subgroup of RA patients such as those with Feltys syndrome.