We read with interest the article by Leung et al.,1 who investigated the presence of antimitochondrial antibodies (AMAs), the hallmark of primary biliary cirrhosis, in patients with acute liver failure (ALF). This is a condition frequently characterized by oxidative stress,2 which is a possible trigger for AMA production3; this is the rationale of the study. As controls, antibodies to glycoprotein 210 kDa (gp210), speckled protein 100 kDa (sp100), centromere, chromatin, soluble liver antigen (SLA), tissue transglutaminase, and deaminated gliadin peptides were also studied.
AMAs were sought by a highly sensitive and specific triple-hybrid (mitochondrial antigen 3 [MIT3]) enzyme-linked immunosorbent assay (ELISA) containing the immunodominant mitochondrial epitopes of E2 subunits of pyruvate dehydrogenase, branched chain 2-oxo-acid dehydrogenase, and 2-oxo-glutarate dehydrogenase and were found in 28/69 (40.6%) patients, 9 (13%) of whom had acetaminophen-related ALF. The only additional antibody reactivity present in a significant percentage (57.1%) was that against anti–tissue transglutaminase, whereas other autoantibodies such as those against SLA were practically absent when testing was performed with a commercially available ELISA.1
Investigating an autoimmune component of ALF, we recently described autoantibody reactivity in an ALF series in the United Kingdom comprising 73 patients4 and, in apparent contrast to Leung et al.,1 found virtually no AMAs (1/73, 1%) through testing with immunofluorescence but did find anti-SLA antibodies in 22% through testing with a sensitive radioligand assay.5 Because discrepancies between the two studies may be due to methodological differences, we retested 47 of our ALF patients in whom sera were available both by conventional indirect immunofluorescence6 and by an MIT3 ELISA similar to that used by Leung et al. Although all sera remained negative for AMA by immunofluorescence, 13/47 (28%) were positive by the MIT3 assay. Conversely, when for the detection of anti-SLA we used an ELISA kit similar to that used by Leung et al. which is less sensitive than a radioligand assay, we also found practically no antibody reactivity. Our results indicate the influence of the methodological approach on the results obtained in comparable patient cohorts. We can anticipate anti-SLA to be present in the American series if a sensitive radioligand assay is used for their detection. We found no relation between the presence of anti-MIT3 AMAs and patient age, sex, biochemical parameters of liver injury, immunoglobulin response mode of presentation of ALF, or outcome of illness.
Both Leung et al.'s study1 and our study show that autoantibody reactivity is frequent in a condition such as ALF characterized by extensive release of autoantigens subsequent to massive liver damage. This autoreactivity is, however, transient because, on repeat testing, AMAs in Leung et al.'s study and anti-SLA in our own cohort disappear.
In conclusion, reactivities to antigens specific to autoimmune liver diseases are present in the course of ALF, although their detection depends on the methodology used. These specific autoimmune reactivities are likely to arise from the release of antigens from damaged hepatocytes and stimulation of lymphocytes that have escaped negative deletion, but they do not, however, progress to chronic autoimmune liver disease, probably because of the intervention of regulatory mechanisms after the acute phase of the disease is over. A deficiency of such mechanisms characterizes both autoimmune hepatitis and primary biliary cirrhosis.6–8
In view of the discrepancy in results when different techniques are used, standardization of methodological approaches for autoantibody detection is warranted to avoid confusing the clinician.