The present study offers data that are useful for the diagnosis and treatment of patients with collagen diseases and liver abnormalities.

Weinvestigatedthe specificity of various autoantibodies in antimitochondrial antibody (AMA)-negative patients with primary biliary cirrhosis (PBC). Weexamined sera from 144 patients with PBC, 17 of whomwere AMAnegative by indirect immunofluorescence. The AMAnegative group showed a significantly higher positivity for smooth muscle antibody, but not for antinuclear antibody, as compared with the AMA-positive group. IgG class anti-PDH by enzymelinked immunosorbent assay (ELISA) were detected in 13 % of the AMA-negative group. However, all PBC patients showed positive IgG, IgA, and/or IgM class anti-M2 to the four M2 proteins by immunoblotting. These results suggest that the detection of IgG and IgMclass anti-PDH and that of antibodies to the four M2proteins increases the positivity of this ELISA method, and that detection ofIgG, IgA, and IgM class anti-M2 to the four M2 proteins by immunoblotting is useful in diagnosing AMA-negative patients with PBC. (Internal Medicine 34: 496-501, 1995)

Autoimmune cholangitis (AIC) has been proposed as a distinct disease entity from primary biliary cirrhosis (PBC), without antimitochondrial antibody (AMA) and anti-M2 antibody but with a high titer of antinuclear antibody (ANA) in the serum. However, negativity for AMA and anti-M2 antibody was determined by different methods in different studies. We hypothesized that anti-M2 antibody negativity in AIC resulted from methodological differences, including selection of the immunoglobulin subclass of the autoantibody. Twenty-three patients compatible with AIC whose serum tested negative for AMA and positive for ANA (> or = 1:80) were compared with 71 AMA-positive PBC patients. Laboratory findings, histology, and the pattern of anti-M2 antibody assessed by immunoblotting were compared. Alkaline phosphatase, total bilirubin, total cholesterol, and IgM values were lower in patients with AIC (P < 0.05, 0.01, respectively). Anti-smooth muscle antibody was detected more frequently in patients with AIC (P < 0.01). However, anti-M2 antibody was detected using immunoblotting not only in PBC but also in AIC cases. IgA class alone, IgM class alone, or both IgA and IgM classes of anti-M2 antibody were detected in 13%, 17%, and 22% of AIC patients, respectively, whereas they were not detected in PBC patients (P < 0.05, P < 0.01, P < 0.01). IgG class anti-M2 was detected in all patients with PBC, whereas it was detected in 48% of patients with AIC (P < 0.01). Histological evaluation showed that the early stages of disease were found more frequently in AIC (78%) than in PBC patients (39%) (P < 0.01). Anti-M2 antibody was detected by immunoblotting in all AIC patients. Hence, AIC is not a distinct disease from PBC. For diagnosing AIC and/or PBC, anti-M2 antibody should be examined by the immunoblotting assay to detect not only IgG but also IgA and IgM subclasses.

Table 1 Causative drugs and constituents of groups. Group A: Patients who tested positive LST with SI over 1.81. Group B: Patients who tested negative LST with SI under 1.80. Group C: Normal controls.

Although antimitochondrial auto-antibodies are characteristically present in the serum of patients with primary biliary cirrhosis (PBC), there is a discrepancy between the positivity for antimitochondrial antibody (AMA) and that for anti-M2 auto-antibody. In an attempt to explain the discrepancy, this study investigates the relationship between the AMA titre, determined by indirect immunofluorescence, and immunoreactivity to four inner mitochondrial membrane proteins (M2 proteins) with molecular weights of 70, 50, 47, and 40 kDa in 129 patients with PBC. Antimitochondrial antibody positivity was identified in 114 (88%) of 129 patients with clinically and histologically confirmed PBC. There were no significant differences between the AMA-negative and AMA-positive groups in clinical characteristics or histologically determined disease stage. Immunoblot analysis showed that all patients had anti-M2 auto-antibodies to one or more of the four M2 proteins. Nine (60%) of the 15 AMA-negative patients had antibodies to only one M2 protein (either 70 or 47 kDa). In contrast, 34 (53%) of the 64 patients with high AMA titres (> or = 1:320) had antibodies to all four M2 proteins. There was a significant rank correlation between the AMA titre and the number of antibodies to M2 proteins (P < 0.01). These findings indicate that the AMA titre is not influenced by the immunogenicity of M2 proteins but by the number of M2 proteins that elicit an antibody response and that decreased immunoreactivity to M2 proteins may induce AMA negativity in PBC serum samples.