Frequent Loss of <i>HLA-B*40:02</i> from Leukocytes in Male Patients with Hepatitis-Associated Aplastic Anemia

Background: Hepatitis-associated aplastic anemia (HAAA), a rare variant of idiopathic aplastic anemia (iAA), is a potentially life-threatening complication of acute hepatitis. Although it has been proposed that several hepatitis viruses are linked to the disease, the exact pathogenic factor underlying this disease is currently unknown. We previously reported somatic loss of HLA class I allele expression as a result of either loss of heterozygosity in chromosome 6p (6pLOH) or somatic loss-of-function mutations in hematopoietic stem progenitor cells (HSPCs) in a subset of patients with iAA. We also found that the missing HLA alleles in HLA class I allele-lacking leukocytes (HLA-LLs) were strikingly skewed toward certain HLA subtypes, which are associated with the risk of developing iAA. These findings provided perhaps the most robust evidence for the autoimmune nature of iAA, since the appearance of HLA-LLs suggests that HSPCs with somatic loss of HLA class I allele expression escape the attack of cytotoxic T lymphocytes that are specific for autoantigens presented by the missing HLA class I alleles. In this study, the correlation between this genomic alteration and the clinical features of HAAA was investigated, and the prevalence of certain high-risk HLA alleles was elucidated. Methods: A total of 19 patients with HAAA aged 1-53 (median: 25, a male/female ratio [M/F]: 1.7) and 205 patients with iAA aged 6-88 (median: 54, M/F: 1.1) were examined for the prevalence of somatic loss of HLA class I allele expression by analysis of allele-specific copy numbers using GeneChip 500K arrays and flow cytometry. In addition, HLA data from 33,733 patients, including 97 patients with HAAA aged 1-53 (median: 14, M/F: 2.2), 1,273 patients with iAA aged 0-75 (median: 18, M/F: 1.1), and 32,363 patients with other hematological diseases aged 0-82 (median: 57, M/F: 1.4) (18,329 with acute myeloid leukemia, 8,428 with acute lymphoblastic leukemia, and 5,606 with myelodysplastic syndromes), who received allogeneic bone marrow transplantation through the Japan Marrow Donor Program (JMDP) were collected from the Japanese Data Center for Hematopoietic Cell Transplantation for analysis of high-risk HLA allele frequency. Results: SNP array analysis showed the presence of 6pLOH(+) clones in 6 (31.6%), all of whom were male (P=0.04, vs. male patients with 6pLOH[-] HAAA), of the 19 HAAA patients, a significantly higher frequency compared with 28 (13.7%), 19 of whom were male (P=0.06, vs. male patients with 6pLOH[-] iAA), of the 205 iAA patients (P=0.048). The genetic loss of HLA haplotypes was explicitly confirmed in all 6 HAAA patients with 6pLOH by analyzing the cell surface expression of HLA-A proteins on leukocytes. Notably, the lost haplotype in all 6 6pLOH(+) HAAA patients carried HLA-B*40:02 at a significantly higher frequency than in patients with iAA (P=0.02), whereas only 12 (42.9%) of the 28 6pLOH(+) iAA patients carried HLA-B*40:02. Furthermore, whether HLA-B*40:02 is associated with an increase in the risk of developing HAAA was analyzed by evaluating a large dataset including 33,733 JMDP registrants. These analyses showed a significantly higher frequency of the HLA-B*40:02 allele in 97 patients with HAAA (25.8%) compared with that in 1,273 patients with iAA (15.2%, P=0.012) and those with other hematological diseases (14.3%, P=0.005), with odds ratios (ORs) of 1.94 (95% CI: 1.2-3.14) and 2.08 (95% CI: 1.32-3.29), respectively. It is noteworthy that the significantly higher frequency of HLA-B*40:02 in HAAA patients (26.9%) than in those with iAA (16.0%, P=0.04, OR: 1.93 [95% CI: 1.08-3.45]) and other hematological diseases (14.4%, P=0.007, OR: 2.18 [95% CI: 1.27-3.75]) was observed only in male patients, whereas the frequency was comparable among female patients with HAAA, iAA, and other hematological diseases when the frequency was compared by sex. Conclusions: This study demonstrated for the first time a higher frequency of 6pLOH, with the missing HLA alleles in HLA-LLs notably skewed toward only HLA-B*40:02 in male patients with HAAA. Our findings for HAAA not only shed light on the characteristics of HAAA patients, and the underlying immunopathological mechanisms, but also offer intriguing insights into the specific HLA allele, HLA-B*40:02, which is most closely associated with the pathogenesis of HAAA, and narrow down the candidate antigens presented by HLA-B*40:02 on HSPCs to those shared with hepatocytes.