A 42-year-old man was admitted to our hospital because of elevated liver enzymes (aspartate aminotransferase, 642 IU/L [normal range: 12-37]; alanine aminotransferase, 788 IU/L [normal range: 7-45]; alkaline phosphatase, 605 IU/L [normal range: 124-367]; γ-glutamyl transpeptidase, 180 IU/L [normal range: 6-30]; and total bilirubin, 8.6 mg/dL [normal range: 0.3-1.2]). His serum immunoglobulin G (IgG) concentration was 5622 mg/dL (normal range: 870-1700), and anti-nuclear antibody titer (1:20480), anti-double-stranded DNA (>400 IU/mL), and smooth muscle antibody titer (1:40) were all abnormal. Infection with hepatitis A, B, and C; cytomegalovirus; and Epstein-Barr virus were excluded, and no drug use was noted. Ultrasonography, abdominal computed tomography, and magnetic resonance imaging showed no abnormalities of the extrahepatic bile ducts or pancreas. The first liver biopsy showed changes associated with typical autoimmune hepatitis (AIH); liver parenchyma was collapsed with broad fibrous septa containing entrapped hepatocytes, and lymphoplasmacytic infiltration with interface activity was seen (Fig. 1A; hematoxylin and eosin [H&E] staining, magnification ×200). Hepatocytes showed rosetting in numerous places (Fig. 1B; H&E staining, magnification ×400). Lobular inflammation was evident with giant cell change of hepatocytes (Fig. 1C; H&E staining, magnification ×400), but no biliary epithelial changes were found. The patient fulfilled the criteria for definite AIH by the International Autoimmune Hepatitis Group and was administered corticosteroids at 60 mg/day, which led to improvement of laboratory findings. Prior to treatment, however, the patient's serum IgG4 concentration was 642 mg/dL (normal: ≤ 135) in a stored serum sample, and immunostaining of liver tissue showed abundant plasma cells with strong immunohistochemical reactivity to IgG4 in a portal tract (Fig. 1D; IgG4 immunostaining, magnification ×400). A second liver biopsy performed 7 months afterward showed remaining portal sclerosis, but lobular distortion and portal inflammation were ameliorated, and serum alanine aminotransferase and IgG4 concentrations were normalized. IgG4-positive plasma cells were scarce in portal tracts (data not shown).
In an earlier report, a strong and unexpected association was seen between serum IgG4 concentration and IgG4-bearing plasma cell infiltration in the liver of a case with type 1 AIH, raising the possibility of a new disease entity termed IgG4-associated AIH.1 Raised serum IgG4 concentration and IgG4-bearing plasma cell infiltration have a high sensitivity and specificity for the diagnosis of IgG4-related diseases.2-4 Similar to the present case, histological findings in the liver of patients with IgG4-associated AIH showed bridging fibrosis, portal inflammation with abundant plasma cell infiltration, interface hepatitis, and lobular hepatitis. More interestingly, giant cell change and rosette formation were obvious as well. These two cases imply that IgG4-related inflammatory processes can occur in the hepatic parenchyma similarly to those in the pancreatobiliary system, and such cases may resemble AIH both clinically and pathologically. On the contrary, Chung et al. described IgG4-associated AIH as patients with AIH who had IgG4-positive plasma cells in the liver.5 Because no cases showed high serum IgG4 in their cohort, we believe they are different from our two representative patients and thus should not be classified as an IgG4-related disease. Koyabu et al. recently reported that an IgG4/IgG1-bearing plasma cell ratio of >1 in the liver is specific for IgG4-related diseases.6 In our patient, the IgG4/IgG1 ratio was >1 (data not shown) and consistent with their findings, which provides further evidence of our case as an IgG4-related disorder. Because IgG4-associated AIH is clearly an IgG4 hepatopathy, this disease should be differentiated from classical AIH. Detection of IgG4 and assessment of liver histology using IgG4 immunostaining may be useful for distinguishing IgG4-related diseases from classical AIH.