Autoimmune hepatitis type 2 associated with an unexpected and transient presence of primary biliary cirrhosis-specific antimitochondrial antibodies: a case study and review of the literature

Another intriguing finding of the present case is that PBC-specific AMA was present in conjunction with evidence of biliary epithelial cell destruction, diagnosed at the age of 3. The presence of AMA in the present case raises the question as to whether PBC has been reported among paediatric patients. The youngest child previously described with PBC was a six year old female [35] (Table2). The remaining cases of ‘paediatric’ PBC have been described in adolescents [1, 2]. Dahlan and colleagues reported two females aged 11 and 16 years, with confirmed PBC [2]. The first patient presented at 11 years of age with abdominal pain and raised aspartate aminotransferase (AST) (48 U/L; normal values <35 U/L), but all other liver biochemistry tests (including ALP) were normal (γGT was not tested) [2]. Her abdominal pain continued and liver function tests demonstrated a raised γGT (107 U/L) at the age of 15, when she was found to be positive for AMA by IIF (1:800), although the specificity of the AMA was not stated [2]. No other autoantibodies were detected [2]. Her serum IgG and IgA were normal but IgM was raised [2]. A liver biopsy at age 16 showed stage II PBC (damaged segmental bile ducts with portal and periportal lymphoid infiltrates), and liver function tests at age 18 demonstrated a cholestatic profile with raised ALP (660 U/L) and slight increase of total bilirubin (24 μmol/L) [2]. It was at this phase that she developed symptoms of pruritus, fatigue and weight loss [2]. She was transplanted at age 21 following worsening of symptoms and increasingly abnormal liver biochemistry despite ursodeoxycholic acid (UDCA) treatment [2]. Histology of the explanted liver demonstrated stage IV PBC. An intriguing aspect of that case is her strong family history of liver disease, as the mother presented at the age of 30 with overlapping features of PBC and AIH, and received a liver transplantation at the age of 34 [2]. Additionally, the grandmother and great-grandmother on the maternal side died of liver cirrhosis of unknown origin [2]. This is of interest given that the mother of the child in the current report is also AMA positive, suggesting a possible genetic predisposition. In the second case, a 16 year old female presented with Raynaud’s and Sicca syndromes as well as raised AST (163 U/L), and was found to be positive for AMA (1:160). She was also positive for ANA (1:320) but it is not clear whether these ANA were those specific for PBC. (i.e. with a multiple nuclear dot or a rim like membranous pattern) [2]. Serum IgG and IgA were normal, but IgM was raised [2]. Histology confirmed stage II PBC at the age of 17 years, and she was started on UDCA (2 g/day), with improvement of her symptoms and liver biochemical tests [2].

Immunological and histological features of PBC have been reported in younger children, but these extreme cases occur due to underlying genetic deficiencies such as IPEX syndrome or IL-2 receptor alpha (IL-2Rα) deficiency [37, 143]. IPEX syndrome is a congenital disorder of immune regulation caused by mutations in the FOXP3 gene, which is required for the suppressive function of naturally arising CD4 + CD25 + regulatory T cells [143‐145]. Tsuda et al. [37] note that patients with IPEX syndrome produce a variety of autoantibodies including AMA. Mutations in the FOXP3 gene (located on the centromeric region of the X chromosome) lead to decreased CD4 + CD25+ Tregs [63, 64, 146], and therefore failure to suppress the production of autoreactive T cells and multi-organ autoimmunity [37]. Tsuda and colleagues note that one individual in their study, an 11 year old male with IPEX, was positive for AMA, with no clinical or biochemical evidence of liver disease [37]. The AMA in that case was of the IgA isotype, with no AMA of the IgG isotype identified, and the plasma IgM and IgA levels were also raised [37]. It was not stated whether the AMA in that case was directed against PDC-E2, and it is unknown whether the child in that case eventually went on to develop PBC [37]. Aoki et al. [143] report the case of a male child born to consanguineous parents, who initially presented with recurrent infections at six months of age. At five years of age, liver biochemistry showed an elevated GGT, and a liver biopsy demonstrated mononuclear lymphocyte infiltration of the portal tracts [143]. These findings differ from the typical histological appearance of PBC, which includes non-suppurative cholangitis, ductopaenia, and sporadically non-caseating granuloma formation. The infant also had CD3+ CD25+ lymphocytopaenia leading to an abnormal CD4:CD8 ratio [143]. Infectious causes were ruled out, but antibody testing was positive for AMA, with reactivity to PDC-E2. Immunoblotting of peripheral lymphocytes showed the absence of the IL-2Rα. A thymic biopsy showed absent CD1a and increased Bcl-2 (an anti-apoptotic protein) expression, due to a four point mutation in FOXP3 leading to translational frameshift, which is similar to IPEX syndrome [143]. The authors suggested that the lack of CD4 + CD25+ Tregs led to a proliferation of autoreactive T cells which were not induced to apoptose due to an increased Bcl-2 [143]. In turn, these autoreactive T cells contributed to the development of PBC (or PBC-like pathology), likely via close interaction with B cells [143, 144]. The child in this case underwent myeloablative chemotherapy and allogenic stem cell transplantation, with complete resolution of his symptoms and no residual serological reactivity to recombinant PDC-E2 [143]. As mentioned, these cases do not represent true paediatric PBC, as the underlying pathological process was linked to a genetic deficiency.

The presence of AMA is highly suggestive of PBC, or predicts progression of an asymptomatic patient to overt PBC [13, 36]. However, Zamfir et al. [83] noted the presence of AMA in 10 of 900 patients presenting with extrahepatic disorders. Of those patients, nine were presenting with haematological disease and one with a dermatological condition [83]. Eight of the ten had high AMA titres, with the remaining two having low titres, but the presence of AMA was confirmed by immunoblot. The immunodominant antigen in all positive AMA cases confirmed by immunoblot was PDC-E2 [83]. It is not known if there was also reactivity to other PBC specific AMA antigens such as the E2 subunits of branched chain oxoacid dehydrogenase complex (BCOADC-E2) and oxoglutarate complex (OGDC-E2) [13, 26]. Interestingly three of the 10 patients were under the age of 20, at 19, 15 and 6 years of age [83]. All three of these individuals were male, and had significant AMA titres, as well as positive ANA of unknown specificity [83]. The presenting disorders among the three were haematological, autoimmune haemolytic anaemia, thrombocytopaenia and idiopathic thrombocytopaenic purpura, respectively [83]. It should be noted that in the case of the 15 year old, the AMA appearance by IIF were atypical (non-anti-M2), and the positivity was not confirmed by immunoblot. It is therefore reasonable to speculate that this case may represent a non-PBC related AMA or a false positive AMA test. The 19 and 6 year old cases had AMA titers of greater than 1/640, and the presence of AMA was confirmed by immunoblotting [83]. These patients were unusual not only due to their young age, but also because they were male. Of the remaining 7 AMA positive patients with extrahepatic disorders, six were female [83]. It has been suggested that these patients may eventually go on to develop PBC, but whether this occurred or not is unknown, as no liver biopsy or follow-up studies were performed.

Unlike the above study where no liver disease was reported in the presence of high AMA titres, Gregorio et al. [26] report a 12 year old girl with autoimmune hepatitis type 1 who was AMA anti-M2 positive. This patient initially presented with a four month history of jaundice, fatigue, anorexia and weight loss, and on examination was found to have hepatosplenomegaly and liver disease stigmata [26]. Her serum bilirubin and ALT were elevated, but all other liver biochemistry was normal. She had increased IgG and was positive for PBC-specific anti-M2 AMA (titre 1/640) [26]. The presence of anti-M2 AMA was confirmed by western blot analysis [26]. Immunological testing for ANA, anti-smooth muscle antibody, and anti-LKM-1 were negative [26]. Histological assessment of six liver biopsies taken over several years showed chronic hepatitis with varying degrees of inflammation, but no evidence of PBC [26]. Treatment with prednisolone and azathioprine improved her condition, and her transaminases normalised, but she had three episodes of spontaneous bacterial peritonitis and four episodes of relapse due to poor treatment compliance [26]. Twelve years after her initial diagnosis, she developed liver failure and died following an intraperitoneal haemorrhage. Liver tissue obtained at autopsy confirmed cirrhosis with persistent inflammation, but no features confirmatory of or compatible with PBC [26]. Over the twelve years since diagnosis, she remained AMA positive, with titers ranging from 1/40 to 1/1280. She did not develop Sicca syndrome, systemic sclerosis or any other rheumatological condition associated with AMA positivity or PBC [26]. Her low IgM and ALP, as well as the lack of histological PBC features, are atypical for “overlap” between autoimmune hepatitis and PBC seen in adults [13, 26, 142].

Hannam et al. [28] report two cases of neonatal liver disease, in which the transplacental passage of AMA occurred. The first case was that of a female infant born to a 32 year old mother with a three year history of pruritus and joint pain, who was AMA positive with normal liver function tests. The foetus was found to be hydropic, and required three intrauterine transfusions before delivery at 31 weeks by caesarean section [28]. The infant displayed deranged liver function tests, and was positive for AMA (titre 1/160) and ANA (1/20) [28]. These autoantibodies were present in the mother at the same titre, and both had the same antibody epitope recognition pattern to PDC-E2 and PDC-E3 binding protein (also known as PDC-X), recognising the major epitopic regions on these antigens [28]. Over several weeks the infant’s condition and liver function tests improved, and the autoantibodies became undetectable. The second case involved a male infant born to a 28 year old mother following an uneventful pregnancy. The infant presented at five weeks with poor weight gain, jaundice, pale stools and an erythematous rash [28]. Liver function tests were deranged, and an antibody screen was positive for AMA (titre 1/20) and ANA (1/10) [28]. The mother was also positive for the same antibodies at the same titre, but she was asymptomatic and had normal liver biochemistry [28]. Titres of 1/10 and 1/20 based on IIF using liver, kidney, stomach tissues are considered significant in children, though in adults titers are considered positive if they exceed 1/40. Antibodies in both mother and child had an identical epitope recognition pattern to PDC-E2 and PDC-E3 binding protein [28]. A liver biopsy of the infant demonstrated cholestasis with hepatitis and mild portal fibrosis [28]. Over the course of three months the infant's condition improved, liver biochemistry normalised, and the antibodies became undetectable [28]. In both cases the AMA were demonstrated to be maternal in origin, as in both infants they were of the placenta-crossing IgG1 and IgG 3 class, their concentrations declining over several weeks [28]. The epitope recognition pattern was also the same in both infant-mother pairs [28]. These cases are of interest given the transplacental passage of AMA, and its association with liver disease in the infants. Additionally, the AMA positivity was transient, akin to the current case report.

In the patient herein illustrated, AMA became undetectable five years after the initial presentation. The decline of AMA over time may be due to the effect of immunosuppressive treatment or may indeed reveal the transient nature of the AMA positivity. A transient appearance of AMA has been previously noted in North American patients with acute liver failure (ALF). AMA have been reported to be present in up to 40% of adult patients with ALF and no other signs of PBC [32]. One study examined 217 serum samples from 69 patients with ALF, with samples collected over a 24-month period [32]. Details as to how many patients were children were not given. Initial testing showed that 40.6% of these patients had AMA, with reactivity against the major mitochondrial antigens (PDC-E2, BCOADC-E2, and OGDC-E2) [32]. By 24 months, only one subject remained positive for AMA directed against OGDC-E2 [32]. Similar data have been reported in the series of ALF patients from King’s College Hospital in London [24]. In that report, 13/47 (28%) sera from ALF patients tested positive for AMA, using the highly sensitive and specific MIT-3 (mitochondrial antigen 3) enzyme-linked immunosorbent assay (ELISA) [24]. Of interest, these 13 cases tested negative for AMA using IIF, indicating that the sensitivity for the detection of AMA largely depends on the methodology used [24].

Unlike the reported cases of transient AMA, liver histology in this case showed biliary damage. In our experience, ALF in paediatric patients is not associated with the presence of AMA. During a large-scale autoimmune-screening program over a 7-year period, only 2 out of 3808 Italian children with liver diseases of a variety of causes tested positive for AMA.

In conclusion, we report a case of AIH-2 with PBC-specific AMA positivity, documented both by IIF and mitochondrial antigen-specific immunoassays. Investigation of this peculiar case may give insight into the mechanisms responsible for the breakdown of immunological tolerance.