Search

Since certain autoimmune diseases, including myasthenia gravis and pemphigus vulgaris can be reproduced in mice by passive transfer of immunoglobulins from affected patients, we assessed whether this procedure could be optimised. Repeated injections of human IgG into mice during pregnancy induced tolerance to human IgG in the litter, and this persisted for at least 9 months. We show that three different human autoantibodies, to mitochondria, centromere and collagen, were retained in the serum of neonatally tolerized mice, but pathogenic effects of these particular autoantibodies were not demonstrable over the four week time scale of our experiments. However, our model should be applicable to studies on human autoantibodies which might damage the appropriate tissue in a heterologous species.

The reactivity of sera in liver diseases was investigated by an immunoradiometric assay for antibody to liver membrane antigens (LMAg). Serum reactivity was similar using glutaraldehyde treated hepatocytes from man and several animals, but the reactivity with human hepatocellular carcinoma cell lines was weak and irregular. F(ab)2 fragments from reactive sera were used in competitive inhibition experiments to ascertain whether anti-LMAg from different liver diseases reacted with the same or different antigenic determinants of the liver membrane. Sera from different patients with autoimmune CAH reacted predominantly with the same membrane components, as did sera from hepatitis B virus associated CAH (CAH-B); however sera from acute viral hepatitis A did not and, since such sera were reactive to LMAg in the assay, other membrane antigens are presumably involved. Enzymatic treatment of viable hepatocytes was performed to determine the nature of the reactive antigen(s): proteases or periodate resulted in reduced binding of reactive sera, but neuraminidase did not, suggesting that the LMAg recognized by CAH sera comprises species non-specific membrane glycoproteins. We conclude, on the basis of quantitative data from a immunoradiometric assay, that antibody to LMAg is more probably reactive than pathogenic, and that assessment of the antigenicity of specific membrane components is necessary.

Sera that contained autoantibodies to mitochondria (AMA) by immunofluorescence were examined by immunoblotting for reactivity with mitochondrial polypeptides from various mammalian species, yeast, and E. coli. Mitochondrial polypeptides were separated by polyacrylamide gel electrophoresis, were immobilized on nitrocellulose, and were exposed to sera. The sera tested included 18 AMA-positive sera from patients with primary biliary cirrhosis (PBC), two AMA-positive sera from patients without PBC, and 53 AMA-negative sera. All AMA-positive sera reacted with either one or the other, or usually both of two human mitochondrial polypeptides of 70 kilodalton (kD) and 45 kD. The 53 AMA-negative sera were not reactive with the 70 kD polypeptide, but six reacted with the 45 kD polypeptide. The reactivity of the 70 kD and the 45 kD polypeptide was destroyed by brief exposure to trypsin. The counterpart of the 70 kD reactive polypeptide in human mitochondria was a 65 to 70 kD polypeptide in rat and mouse mitochondria, and a 55 kD polypeptide in yeast and in E. coli. The apparent 45 kD polypeptide was similar in all mitochondrial preparations tested, but no counterpart could be identified in E. coli. Beef heart mitochondria were used to show that the reactive polypeptides were present in a semipurified preparation of the F1 portion of mitochondrial H+ ATPase; however, sera did not react with the beta subunit of ATPase, proposed as a candidate mitochondrial autoantigen. The present molecular characterization of two particular antigens should lead to the more precise identification of these antigens, and also to a clearer insight into the pathogenesis of PBC.

Autoantibodies to smooth muscle (ASMA), and to actin which is a major determinant of such reactivity, were measured in the serum of 94 patients with three defined categories of alcoholic liver disease, fatty liver, alcoholic hepatitis, and alcoholic cirrhosis, and in controls matched individually by age and sex with the patients. Autoantibody to monomeric G-actin was detected by an ELISA and autoantibody to polymeric F-actin by immunofluorescence staining of fibroblast stress fibers in cultured cells. Values for the ELISA were expressed as a percentage of the value for a strongly reactive standard serum. The mean value for antibody to G-actin in 40 patients with alcoholic cirrhosis (70 +/- 33%) was significantly greater than that for the matched controls (28 +/- 18%), but the mean value for the 30 patients with alcoholic hepatitis (46 +/- 16%) and the 24 with fatty liver (42 +/- 24%) did not differ significantly from the controls. High levels of reactivity with G-actin correlated significantly with HLA B7. ASMA was demonstrable to low titer in 11 of the 94 sera, and positive ASMA reactions by immunofluorescence correlated with high binding values to G-actin in the ELISA. Antibody to F-actin was found in only one serum and no controls. Thus in different liver diseases the reactivity of antibodies to monomeric G-actin and polymeric F-actin may differ, presumably because of specificity for different determinants of the actin molecule. Reactivity to G-actin may distinguish a group of alcoholic subjects in whom a predisposition to autoimmune reactivity is one of the determinants of progression of liver damage to cirrhosis.

The mononuclear cell infiltrate in the liver was analyzed, using a panel of monoclonal antibodies (MAbs) of known specificity, in 10 patients with auto-immune chronic active hepatitis and, for contrast, in 14 with other types of chronic parenchymal liver diseases. In all cases, the mononuclear cell (MNC) infiltrate in the liver consisted mostly of T lymphocytes. Helper (Th) cells were more frequent than suppressor/cytotoxic T cells (Tsc) in the portal tracts and cirrhotic scar tissue, while Tsc were more common in the hepatic parenchyma. The number of Tsc cells in the parenchyma was greatest in patients with histologically active CAH and least in patients with quiescent cirrhosis. "Plasmacytoid" cells with the morphology of plasma cells, a hallmark of the MNC infiltrate in auto-immune CAH, were more frequent in histologically active CAH than in quiescent cirrhosis. These plasmacytoid cells were T200 + ve, and hence of bone marrow origin, but the majority expressed neither membrane nor cytoplasmic immunoglobulin nor any lineage-specific marker antigens, and hence did not fulfil criteria for B lymphocytes; however, these cells were positive for OKT10 and HLA DR. No difference was evident between the MNC infiltrate in the liver in auto-immune CAH and that in the other acute or chronic liver diseases studied, including HBV-associated CAH; hence immunohistological studies do not point to any pathological processes uniquely responsible for the pattern of hepatocyte damage seen in auto-immune CAH.

A radioimmunoassay using a single cell suspension of glutaraldehyde-treated monkey hepatocytes was developed to measure antibody in human serum to liver membrane antigens. Hepatocytes were exposed to doubling dilutions of test sera in microtitre trays and binding of serum IgG was determined by exposure of the washed cells to 125I-labelled staphylococcal protein A. Specificity of binding was established according to several criteria including use of non-hepatic cells such as monkey kidney cells and various cell lines, IgG depletion of serum, absorption of binding activity by pre-exposure of serum to liver cells, and demonstration that binding was mediated by the F(ab) rather than Fc portions of immunoglobulin molecules. This radioimmunoassay distinguished autoimmune-type chronic active hepatitis (CAH) from other categories of liver disease. It should prove useful for diagnosis and classification of cases of CAH, and could facilitate identification of a liver-specific membrane antigen relevant to pathogenesis of autoimmune hepatitis.

An immunoradiometric assay for IgG class autoantibody to liver membrane antigens, based on serum binding to glutaraldehyde treated monkey hepatocytes, was used to examine sera from patients with chronic active hepatitis (CAH) and other acute and chronic liver diseases. All sera from normals and patients showed binding, up to a titre of 1/2,048. For comparison of assays, results were normalized by selecting two reference sera, one with a high degree of binding, and one from a healthy subject with a low degree of binding: at a dilution of 1/2,048, these sera were given binding values of 100% and 0%. The values for the binding of unknown sera at the same dilution were calculated from these two reference values. For 26 patients with autoimmune CAH, the mean (+/- s.d.) percentage binding value (70 +/- 33%) was significantly higher than the mean value for 26 healthy subjects (10 +/- 15%), and high binding values were significantly associated with biochemically active hepatitis. The mean percentage binding value was moderately increased for eight patients with HBsAg associated CAH (42 +/- 12%), 13 patients with alcoholic hepatitis with cirrhosis (37 +/- 25%) and 45 patients with acute viral hepatitis A (40 +/- 27%) or B (52 +/- 37%). At a cut-off binding value of 65%, the assay as a single diagnostic procedure was shown to have a 70% sensitivity and a 95% specificity for the diagnosis of autoimmune CAH. Better understanding of the pathogenetic significance of antibodies to liver membrane antigens in CAH and other liver diseases will depend upon biochemical analysis of the presumably multiple antigenic determinants on the hepatocyte membrane.

In 18 cases of chronic liver disease due to non-A, non-B hepatitis virus(es) in which the diagnosis was established by transmission, including chimpanzee inoculation in nine, sera were tested for the autoantibodies characteristically associated with autoimmune chronic active hepatitis. The frequency of autoantibodies to nuclear, smooth muscle, cytofilament, mitochondrial and liver membrane antigens was low, being not greater than that recorded for a normal population, and the few positive reactions obtained were at very low titre. These findings suggest that among cases of 'HBsAg negative' chronic hepatitis, those due to NANB infection are distinguishable from those due to autoimmune chronic hepatitis by negative serological tests for autoantibodies.

The frequency of serum autoantibodies including anti-albumin antibodies was investigated for patients with various categories of alcoholic liver disease (ALD). The 95 patients were grouped into 3 categories: fatty liver (25 cases), alcoholic hepatitis (29 cases), and alcoholic cirrhosis (41 cases), and each group was matched with healthy controls by age, sex, ethnic origin and socio-economic status. For all patients with ALD, there was a 5-fold greater frequency over the controls of positive tests for antinuclear antibody (ANA), but titres were low; the pattern of ANA reactions was speckled in 50% of the cases. For patients with alcoholic cirrhosis, there was an 8-fold increase in frequency over the controls in anti-smooth muscle antibody (ASMA), but titres were low, pointing to a possibility that autoimmunity might be one of several determinants of progression of alcoholic hepatitis to cirrhosis. For none of the groups was there an increase over the controls in the mean titre of sheep red cell agglutinins, nor were there increased haemagglutinin titres of antibody to bovine serum albumin or to human albumin, arguing against a general increase in antibody production in ALD. Greater knowledge of the frequency, significance and pathogenicity of reactive autoantibodies which occur in response to various types of tissue damage is required for interpretation of the increase in ANA and ASMA in alcoholic liver disease.