Abstract 3200Poster Board III-137Hepatitis-associated aplastic anemia (HAA) is a subset of acquired AA that is highly responsive to immunosuppressive therapy. The target antigens of the immune system attack in HAA are thought to be a protein shared by both liver and hematopoietic stem cells, since it is usually associated with severe hepatitis of unknown etiology. Screening sera from patients with HAA for the presence of antibodies (Abs) recognizing liver cell-derived proteins may be useful in identifying novel auto-antigens in AA. To test this hypothesis, sera from HAA patients were examined using immunoblotting with a lysate of a hepatocellular carcinoma cell line Huh7 and subsequent peptide mass fingerprinting. Methods and ResultsThe serum of a patient with typical HAA (a 23 year-old male) possessing a small population of paroxysmal nocturnal hemoglobinuria (PNH)-type cells was used for Western blotting (WB) with the lysates of Huh7. A distinct band of 70 kDa protein was revealed. The same band was revealed when the culture supernatant of Huh7 cells was subjected to WB. The peptide mass fingerprinting of the 70 kDa band identified this protein to be heat shock protein (HSP) 72. HSP72 is a stress-inducible protein and extracellular HSP72 enhances the cytotoxicity of CD4+ T cells and NK cells. An examination of the sera from HAA patients, idiopathic acquired AA (IAA) patients and healthy individuals with WB revealed the anti-HSP72 Abs to be detected in 10 of 12 (83%) HAA patients and in 57 of 80 (71%) IAA patients while it was detected only in 8 of 59 (14%) healthy individuals. The prevalence of anti-HSP72 Abs in AA was markedly higher than that of anti-kinectin Abs (39%), anti-PMS1 Abs (10%), anti-DRS-1 Abs (38%) or anti-moesin Abs (37%) reported previously. Anti-HSP72 Abs were frequently detectable both in patients with IAA possessing PNH-type cells (63%) and in patients without PNH-type cells (86%), a finding contrasting to the higher prevalence of anti-DRS-1 Abs and anti-moesin Abs in patients with PNH-type cells than in those without PNH-type cells reported previously. Although anti-HSP72 Abs were detectable in the sera of patients with rheumatoid arthritis and systemic lupus erythematosus, the prevalence was 15% (4 of 27) and 20% (1 of 5), respectively. In contrast to a previous report that detected anti-HSP72 Abs in 24% of patients with chronic hepatitis C, WB failed to detect the Abs in the sera of 4 patients with autoimmune hepatitis and 5 with hepatitis B or C. Ten patients with HAA were treated with immunosuppressive therapy, and 7 of the 8 responders expressed anti-HSP72 Abs. The quantification of the gene expression level of HSP72 by blood cells using real-time PCR demonstrated that the HSP72 mRNA levels were markedly higher in myeloid leukemia cell lines as well as CD34+ cells isolated from 3 healthy individuals in comparison to that in lymphoid or monocytoid leukemia cell lines. HSP72/GAPDH ratios of PBMCs and CD34+ cells from 3 healthy individuals, K562, KH88, OUN-1 were 0.51, 1.31, 1.02, 0.07 and 0.09 respectively. Other leukemia cell lines such as Daudi, Molt-4 and THP-1 did not display detectable levels of HSP72 mRNA. The cell surface expression of HSP72 was examined in various kinds of leukemia cell lines and CD34+ bone marrow (BM) cells derived from 3 healthy individuals using Ab to HSP72 (Clone C92F3A-5) because previous studies demonstrated heat-inducible expression of HSP72 by K562. Flow cytometry detected cell surface HSP72 on immature CML cell lines such as K562 but not on CD34+ BM cells, acute promyelocytic leukemia cell lines such as NB-4 and HL-60, and lymphoid leukemia cell lines such as Molt-4 and Daudi. Exposure to 42°C for 2 h increased the HSP72 expression on K562 cells and Molt-4 cells but not on CD34+ cells. ConclusionAnti-HSP72 Ab is the most prevalent auto-Ab in AA among the auto-Abs previously detected. Given the increased expression of HSP72 by immature myeloid cells as well as stress-inducible cell surface expression of the molecule, immune responses to HSP72 may thus play an essential role in the pathogenesis of HAA and IAA. DisclosuresNo relevant conflicts of interest to declare.
Read full abstract