Expansion of a liver‐infiltrating cytotoxic T‐lymphocyte clone in concert with the development of hepatitis‐associated aplastic anaemia

Hepatitis-associated aplastic anaemia (HAA) is a variant of acquired aplastic anaemia (AA) in which an episode of hepatitis precedes AA by a period of weeks or months (Hagler et al, 1975). In AA, cytotoxic T-lymphocytes (CTLs) play a central role in bone marrow (BM) destruction, and oligoclonal expansion of CTLs correlates with disease activity (Young et al, 2006). Immune-mediated mechanisms have also been postulated in the pathogenesis of fulminant hepatitis (FH). However, the exact immunopathophysiological relationship between FH and AA in HAA remains unclear. Previous reports have presented occasional cases in which cytopenia of more than one linage was observed, even at the onset of FH (Honkaniemi et al, 2007). Taken together, we hypothesized that common CTLs recognizing similar target antigens against both liver and BM cells exist during the early period of FH; subsequently, a selective proliferation of clones that are highly tropic to BM cells cause HAA at later stages. To prove this hypothesis, we have tracked common T-lymphocytes that recognize similar target antigens both against liver and BM at the onset of FH and AA from the patient with HAA. A 1-year-old boy was transferred to our hospital in April 2009 for treatment of acute liver failure, with markedly elevated bilirubin and transaminase levels (aspartate aminotransferase [AST], 4229 iu/l; alanine aminotransferase [ALT], 3583 iu/l; total bilirubin, 227·4 μmol/l). On admission, he showed marked jaundice and grade II hepatic encephalopathy. With intensive treatment, the encephalopathy and laboratory findings gradually improved; two months later they had recovered to normal range. Microscopic examination of a liver specimen showed focal necrosis and syncytial giant cells, suggesting viral infections. However, we could not identify any causative pathogens such as hepatitis A, B, or C; herpes simplex virus; rubella; cytomegalovirus; or Epstein-Barr virus. Immunohistochemical staining of liver specimens (liver-FH) revealed that CD8+ T-lymphocytes were predominant and represented more than 95% of all lymphocytes, which was consistent with a diagnosis of FH. Three months after the onset of FH, transaminase levels gradually elevated again (AST, 231 iu/l; ALT, 167 iu/l). Thereafter, the patient progressed to pancytopenia (white blood cell [WBC] count, 1·85 × 109/l; neutrophil count, 0·17 × 109/l; platelet count, 94 × 109/l; haemoglobin [Hb], 66 g/l; reticulocytes, 46 × 109/l). A diagnosis of HAA was made based on a BM biopsy in July 2009 without any evidence of malignancy or chromosomal abnormalities. In addition, we continued to observe focal necrosis and syncytial giant cells in liver biopsy samples obtained at the onset of AA (Liver-AA) but to a much lesser degree compared with those of FH. Immunosuppressive therapy (IST) using ciclosporin (CsA), antithymocyte globulin (ATG), and methylprednisolone was initiated. Two months after starting IST, the pancytopenia had almost resolved (WBC count, 2·40 × 109/l; neutrophil count, 0·96 × 109/l; platelet count, 147 × 109/l; Hb, 117 g/l; reticulocytes, 29 × 109/l). CDR3 size distribution was analysed as described previously (Lu et al, 2004; Wada et al, 2005). To determine whether each individual profile was normal or skewed, we used a previously reported complex scoring system (Konno et al, 2004). The number of T-cell receptor (TCR) Vβ (TRBV) subfamilies with a skewed CDR3 size pattern, as defined by a complex score less than 4, was determined for each subject. TCR CDR3 cloning and sequencing was performed as described previously (Mizuno et al, 2004). Significant skewing of the CDR3 size distribution was only observed in TRBV13 in BM and PB specimens at the onset of AA (Figs 1A–B), suggesting that the T-lymphocyte TRBV13 clones possessing TRBV13 were deeply involved in the pathogenesis of AA. In contrast, 14 and 17 TRBVs, including TRBV13, showed a highly skewed pattern in liver samples at the onset of FH and AA, respectively (Fig 1C, TRBV3-1, 6-6, 11-2, 13, 15, 18, 19, 20-1, 24-1, 25-1, 27, 28, 29-1, 30; Fig 1D, TRBV2, 3-1, 6-3, 6-5, 6-6, 9, 10-1, 11-2, 12-3, 13, 14, 15, 18, 19, 27, 29-1, 30). Subsequently, CDR3 nucleotide sequences of predominant products were determined in TRBV13 from each sample (Fig 2) to identify the same clones more precisely. A TRBV13 clone with identical CDR3 nucleotide sequence predominated (clone 1) in the BM and PB (34% and 60% within TRBV13, respectively). Surprisingly, this clone constituted a minor but definitive one in the liver at the onset of FH (16% within TRBV13). In contrast, initially predominant clones in the liver (clone 3; 47% within TRBV13) contracted to a minor subpopulation in the BM and PB when the patient developed AA (17% and 10% within TRBV13, respectively). We could not directly identify the pathological antigens that drive specific T-lymphocyte proliferation because of technological limitations of our approach using size distribution and sequence of CDR3 regions. These results, however, suggest that common T-lymphocytes recognizing similar target antigens against both liver and BM cells clones exist during the early period of FH, and subsequent selective proliferation of clones that are highly tropic to BM cause HAA at later stages. This patient had a very good response to IST. Two months after starting IST, the pattern of CDR3 size distribution was just recovering to normal (Gaussian distribution) and initially predominant clones in both the liver and BM almost disappeared in the TRBV13 in PB (Fig 2). However, an extremely skewed pattern of the CDR3 region was still observed and abnormal clone 1 still occupied 71% within the TRBV13 in the BM. Eight months after starting IST, CDR3 size distribution recovered to Gaussian distribution and CTL clones against both the liver and BM almost disappeared in both the BM and PB (Fig 2). These data suggest that it would take a relatively long time to completely recover immunological parameters even in IST good responders. In conclusion, this is the first report to track CTL clones infiltrating target organs of HAA over the course of the disease. This work was supported by a grant from the Ministry of Health, Labour, and Welfare of Japan, Tokyo. YI wrote the manuscript. YI, RK, SM, HM, and RA performed the laboratory work. YI, TW and TT provided interpretation and analysis of data. RN and SK design the research study. AY is the principal investigator and takes primary responsibility for the paper. We have no conflict of interest to declare.