Escherichia coli and Proteus mirabilis Inhibit the Perinuclear But Not the Circulating Antineutrophil Cytoplasmic Antibody Reaction

Abstract

Background: Perinuclear antineutrophil cytoplasmic antibodies (P-ANCA) are found in 48%-83% of serum samples from patients with ulcerative colitis (UC). Their pathogenic role and initiating stimuli are unknown. In contrast to patients with vasculitides and ANCA reactivities, the antibodies in UC patients do not react with myeloperoxidase (MPO) or proteinase 3 (PR3). The aim of the present study was to investigate whether bacterial species of the intestinal tract and other sources could interfere with P-ANCA in sera from patients with UC. Methods: Seventeen P-ANCA-positive and anti-MPO-negative serum samples from patients with UC were tested with Escherichia coli 014 and Staphylococcus aureus Wood 46. Six of these serum samples with different P-ANCA titres were selected to test further the influence of 15 different gram-negative or gram-positive bacterial strains. Six anti-MPO positive P-ANCA, 5 anti-PR3 positive C-ANCA, and 10 antinuclear antibody (ANA)-positive serum samples were used as controls. The antineutrophil cytoplasmic antibodies (ANCAs) were analysed by an indirect immunofluorescence method (IIF) on ethanol-fixed neutrophils, and the ANAs were tested by IIF on HEp-2 cells or rat liver tissues. The bacteria used in the experiments were either live or killed by formalin or glutaraldehyde fixation or heated at 80°C for 30 min. The test was first performed as a bacterial absorption test with sedimented organisms and then at various temperatures with the supernatant from suspension of live bacteria. Results: Both MPO-positive and MPO-negative P-ANCA reactivity was abolished by absorption of patient sera with live E. coli and Proteus mirabilis but not with bacteria representing members of 10 other species, suggesting that antibody reactivity was absorbed away. However, continued experiments indicated that the inhibition of P-ANCA was not due to classic antigen-antibody interactions but rather to decomposition of the antigenic substrate of the neutrophils by factors present in the supernatants of live E. coli and P. mirabilis. The activity of the supernatant was temperature-dependent, with strong activity at room temperature and 37°C, no activity at 0°C, and abolished by mild heat treatment (56° or 60°C). No activity was shown in the supernatants from bacteria treated with formaldehyde or glutaraldehyde. Conclusions: Soluble material from live E. coli and P. mirabilis has the capacity to decompose the antigenic substrate of neutrophils responsible for both MPO-positive and MPO-negative P-ANCA, most probably brought about through enzymatic activity. Anti PR3-positive C-ANCA were not affected, which suggests substrate specificity of the proposed enzymatic activity.