Search

In immune-mediated acquired aplastic anemia (AA), the presence of an HLA allele, which is highly overrepresented or lost due to somatic mutations, may represent a specific immune pathophysiology and a clinical manifestation. HLA-B*14:02 is one of the most overrepresented class I alleles in AA and is also frequently affected by a somatic loss of expression; the inherited B*14:02 genotype was correlated with high-risk clonal evolution in two independent cohorts in the U.S. (Babushok DV et al. Blood Adv 2017; Zaimoku Y et al. manuscript in preparation). In contrast, HLA-B*14:02 is virtually absent in Japanese, in whom somatic mutations of AA have frequently been detected in HLA-B*40:02, B*54:01, and A*02:06, and occasionally in A*02:01, A*02:07, A*31:01, B*13:01, B*40:01, B*40:03, B*44:03, B*55:02, and B*56:01 (Mizumaki H et al. Haematologica 2021). A class II allele HLA-DRB1*15 is highly overrepresented in AA across various ethnic groups, including those in the U.S. and Japanese. This retrospective study in the Japanese population aimed to explore the clinical significance of disease-associated non-B*14:02 HLA class I and II alleles in AA. A total of 423 enrolled patients with AA (very severe [n = 81], severe [n = 266], transfusion dependent non-severe [n = 76]; median age 60 [range, 1-86] years) had undergone genotyping for HLA-A, HLA-B, HLA-C, and HLA-DRB1 at 2-field resolution. The HLA allele frequencies in these patients were compared to those in a Japanese HLA haplotype dataset (n = 19183; Ikeda N et al. Tissue Antigens 2016). The most overrepresented allele in AA was HLA-DRB1*15:02, followed by DRB1*15:01, B*40:02, and A*02:06 (Table); DRB1*13:02 and B*44:03, which are in linkage disequilibrium, were markedly underrepresented, consistent with a well-known protective role of DRB1*13 against autoimmune diseases. HLA-DRB1*15:02 was also significantly correlated with age and its frequency among patients aged The overall response rate to anti-thymocyte globulin-based immunosuppressive therapy at 6 months was 63% (139 of 220 treated and evaluable patients). A trend for a higher response was observed in patients harboring mutation-related HLA-B alleles (except for minor alleles B*13:01, B*40:03, and B*55:02) and the highly overrepresented or protective HLA-DRB1 alleles, but not in the HLA-A alleles (Figure D). A multivariate logistic regression revealed that the combination of the presence of any favorable alleles in HLA-B (odds ratio 3.6, P < 0.0001) or in HLA-DRB1 (odds ratio 2.3, P = 0.00085) was significantly and independently associated with a hematologic response; the tendencies for a lower or higher response in very severe disease and the presence of paroxysmal nocturnal hemoglobinuria clone did not reach statistical significance. Further, there was likely an additive effect when two favorable alleles coexisted in HLA-B or HLA-DRB1 (Figure E); the copy number of the favorable HLA-B and HLA-DRB1 alleles stratified the response rate to four groups: three or four copies, 95% (19 of 20); two copies, 72% (61 of 85); one copy, 59% (50 of 85); and zero copy, 30% (9 of 30). Only eight patients displayed clonal evolution to monosomy 7, myelodysplastic syndrome, or acute myeloid leukemia after immunosuppression without significant overrepresentation or underrepresentation of the pathogenic HLA alleles. Using a large dataset of homogeneous Japanese population with high-resolution HLA typing, we revealed, for the first time, a strong relationship between disease-associated (overrepresented, inactivated, or protecting) HLA alleles and the responsiveness to immunosuppressive therapy. Figure 1 Figure 1. Disclosures Takamatsu: Bristol-Myers Squibb: Honoraria, Research Funding; SRL: Consultancy; Adaptive Biotechnologies, Eisai: Honoraria; Janssen: Consultancy, Honoraria, Research Funding. Yamazaki: Novartis Pharma: Honoraria; Kyowa Kirin: Honoraria; Kyowa Kirin: Research Funding. Nakao: Symbio: Consultancy; Kyowa Kirin: Honoraria; Novartis Pharma: Honoraria; Alexion Pharma: Research Funding.

[Background] HLA-class I allele-lacking (HLA[-]) leukocytes are detected in approximately 30% of patients with acquired aplastic anemia (AA), and are thought to represent the involvement of cytotoxic T lymphocyte attack against hematopoietic stem cells (HSCs) in the development of AA, based on the high response rate to immunosuppressive therapy (IST) in patients with such aberrant leukocytes. Similar to glycosylphosphatidylinositol-anchored protein (GPI-AP)-deficient (GPI[-]) leukocytes in patients with paroxysmal nocturnal hemoglobinuria (PNH), HLA(-) leukocytes in AA patients are often clonal or oligoclonal and expand to account for more than 50% of the total leukocytes. Despite such overwhelming proliferation, somatic mutations in driver genes as well as telomere shortening that portend clonal evolution are rarely detected in HLA(-) granulocytes, suggesting the genetic stability of HLA(-) HSCs and the persistence of the immune pressure on HSCs that favors expansion of HLA(-) HSCs (Imi, et al. Blood Adv). However, recent studies from the United States have shown a higher incidence of clonal evolution to secondary myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) in AA patients with HLA(-) leukocytes than in those without such leukocytes, a finding inconsistent with the results of our previous study. Given the high prevalence of HLA(-) leukocytes in AA patients, it is critical to determine whether or not the presence of the aberrant leukocytes is associated with clonal evolution. We therefore addressed this issue by studying the prognosis of a large number of AA patients with or without HLA(-) leukocytes who had been followed for a long term period. We also studied the clonal composition of granulocytes in AA patients with HLA(-) cells, wherein aberrant clones other than HLA(-) cells might be responsible for clonal evolution to MDS/AML. [Methods] We retrospectively analyzed the clinical characteristics of 633 AA patients and peripheral blood samples were examined for the presence of HLA(-) leukocytes using a high-sensitivity flow cytometry (FCM) assay, droplet digital PCR, single-nucleotide polymorphism arrays, or next generation sequencing (NGS) between 2010 and 2020. GPI(-) cells were detected using a high-sensitivity FCM assay as previously described. [Results] HLA(-) granulocytes were detected in 127 (20.1%) of the 633 patients with a median clone size of 16.9% (range, 0.04%-100%); the aberrant granulocytes accounted for greater than 50% of the total granulocytes in 29 (22.8%) of 127 patients. Eighty-nine (70.0%) of the 127 patients possessed aberrant clones other than HLA(-) clones, which included 0.005% to 91.6% GPI(-) cells (n=86), del(13q) cells (n=3), t(1;10) cells (n=1), t(9;13) cells (n=1), inv12 cells (n=1), and trisomy 8 cells (n=1). The prevalence of GPI(-) cells was not significantly different between patients with and without HLA(-) cells (67.7% vs 65.4%). Eighty-five of 102 (83.3%) patients with HLA(-) cells responded to IST, whereas 231 of 318 (72.6%) without HLA(-) cells responded (p90% of granulocytes, suggesting that these few escape clones were enough to sustain the hematopoietic function of the patients. The prognosis survey revealed no clonal evolution to MDS/AML in any of the 127 AA patients with HLA(-) leukocytes after a follow-up period of the median 5 years. In contrast, 15 of 234 (6.4%) patients without HLA(-) cells who were trackable evolved to MDS/AML during a median 5 year follow-up. [ Conclusions] The presence of HLA(-) leukocytes and concomitant aberrant clones was not associated with clonal evolution to MDS/AML in Japanese AA patients, even in those possessing a large (>50% of the total granulocyte) HLA(-) cell population. The discrepancy between our results and the data from the United States may be due to the difference in the race and mechanism underlying HLA loss. These data suggest that HSC clones that escape immune attack, such as HLA(-) and GPI(-) clones, are healthy enough to support hematopoiesis for a long term in AA patients. Disclosures Ishiyama: Novartis: Honoraria; Alexion: Research Funding. Yamazaki:Novartis: Honoraria; Kyowa Kirin: Honoraria, Research Funding. Ogawa:Eisai Co., Ltd.: Research Funding; Sumitomo Dainippon Pharma Co., Ltd.: Research Funding; Chordia Therapeutics, Inc.: Membership on an entity's Board of Directors or advisory committees, Research Funding; Asahi Genomics Co., Ltd.: Current equity holder in private company; Otsuka Pharmaceutical Co., Ltd.: Research Funding; KAN Research Institute, Inc.: Membership on an entity's Board of Directors or advisory committees, Research Funding. Nakao:Alexion: Research Funding; Kyowa Kirin: Honoraria; Novartis: Honoraria; Symbio: Consultancy.

[Background] Acquired aplastic anemia (AA) is thought to be caused by cytotoxic T lymphocyte (CTL) attack specific to antigens presented by class I HLA alleles on hematopoietic stem cells (HSCs) although the target antigens on HSCs are still unknown. HLA alleles that are responsible for the auto-antigen presentation, which can be inferred from the presence of leukocytes that lack particular HLA class I alleles (HLA-lacking leukocytes [HLA-LLs]), may provide useful information on the identification of autoantigens in AA. We previously reported that deep sequencing of HLA class I genes of HLA-LLs revealed various loss-of-function mutations in alleles, including HLA-A*02:06 and B*40:02, suggesting that limited HLA-class I alleles are involved in the autoantigen presentation of AA (ASH 2018). Intriguingly, 60% of patients possessing HLA-LLs due to 6pLOH or several inactivating mutations in different HLA-A or HLA-B genes shared a nonsense mutation in exon 1 (Exon1mut) of which allelic frequency was very low (range 1.0-37.3%, median 4.5%). We hypothesized that the nonsense mutation, which efficiently lacks the corresponding HLA-allelic expression, might have been overlooked due to its low VAFs, and if we could establish a highly sensitive assay for detecting Exon1mut and determine the HLA alleles that undergo the mutation for a large number of AA patients, we might be able to define all HLA alleles that are involved in autoantigen presentation of AA. [Objectives/Methods] To test this hypothesis, we developed a highly sensitive droplet digital PCR (ddPCR) assay for precisely detecting Exon1mut in the peripheral blood (PB) of AA patient. In brief, the exon 1 regions of HLA-A and HLA-B alleles were amplified using two different sets of primer pairs that are complementary to the consensus sequences of the HLA-A and HLA-B alleles. The amplicons were subjected to a ddPCR assay using TaqMan probes complementary to wild-type (WT) and mutant-specific (MT) sequences, which were labeled with different fluorochromes (6-FAM for MT and HEX for WT). Peripheral blood leukocytes from 363 patients with AA (mean 64 [range, 11-93]) years of age, 134 with severe AA and 229 with non-severe AA; 173 males and 190 females; 84 6pLOH[+] and 279 6pLOH[-]) were subjected to the ddPCR assay. All blood samples were analyzed for 6pLOH by SNP array-based methods or a ddPCR assay as previously described. The HLA allele that underwent Exon1mut was determined by targeted deep sequencing using a unique molecular identifier (UMI), which enabled us to detect variant calling at a VAF as low as 0.1%, or deduced from the alleles contained in the lost haplotype, which are known to be the frequently lost alleles due to 6pLOH. [Results] Using 2 different ddPCR mixtures for HLA-A and HLA-B, the presence of Exon1mut was evaluable in all 363 AA patients. The sensitivity of the ddPCR assay for detecting Exon1mutwas 0.07%. 6pLOH was detected in 84 (23.1%) of the 363 AA patients. Ninety-nine (27.3%) of the 363 patients with (55 [65.5%] of 84) or without (44 [15.8%] of 279) 6pLOH were found to be positive for Exon1mut. The median allele frequency of Exon1mut in DNA from the Exon1mut(+) patients was 0.6% (range, 0.074% to 21.3%). In 17 patients whose blood samples were serially available, Exon1mutwas persistently detected in 13 and disappeared in 4 patients for 10-77 months (Figure 1). Among 43 different HLA-A and HLA-B alleles carried by the Exon1mut(+) patients, those with Exon1mutcould be identified by targeted deep sequencing in 54 patients. In 13 of the remaining 42 patients with Exon1mut, the Exon1mut-involved HLA alleleswere deduced from the alleles contained in the lost haplotype due to 6pLOH. These were 12 alleles and included A*02:06 (n=11), A*31:01 (n=3), B*13:01 (n=2), B*40:01 (n=3), B*40:02 (n=26), B*40:03 (n=1), B*54:01 (n=6), A*02:01 (n=2), A*02:07 (n=1), B*44:03 (n=1), B*55:02 (n=2) and B*56:01 (n=1) (Figure 2). The last five infrequent alleles were newly identified as "risk alleles" using the Exon1mutdetection. Two-hundred and twenty (92%) of 239 patients with PNH-type cells possessed at least 1 of the 12 alleles, while 103 (85%) of 121 patients without PNH-type cells did (P=0.045). [Conclusions] The Exon1mutdetection assay identified 12 HLA-alleles that are closely and exclusively involved in the autoantigen presentation of AA in Japanese patients. Similarity analyses of their antigen-presentation motifs may help to identify autoantigen peptides in AA. Disclosures Nakao: Takeda Pharmaceutical Company Limited: Honoraria; SynBio Pharmaceuticals: Consultancy; Ono Pharmaceutical: Honoraria; Novartis Pharma K.K: Honoraria; Bristol-Myers Squibb: Honoraria; Kyowa Kirin: Honoraria; Alaxion Pharmaceuticals: Honoraria; Ohtsuka Pharmaceutical: Honoraria; Daiichi-Sankyo Company, Limited: Honoraria; Janssen Pharmaceutical K.K.: Honoraria; Celgene: Honoraria; Chugai Pharmaceutical Co.,Ltd: Honoraria.

[Background] Olfactomedin 4 (OLFM4), a member of the olfactomedin-related protein family, is constitutively expressed in bone marrow (BM) cells and many gastrointestinal organs and is involved in a variety of cellular functions, including proliferation, differentiation, apoptosis, and cell adhesion in tumor cells. Previous studies have suggested that OLFM4 supports the survival of leukemic stem cells derived from iPS cells of patients with chronic myeloid leukemia. Our recent study revealed a somatic mutation of OLFM4 in HLA allele-lacking granulocytes of a patients with acquired aplastic anemia (AA) in long-term remission (Blood advances, 2(9):1000-1012, 2018). The OLFM4 gene is located at chromosome 13q14.3, which is a commonly deleted region in AA patients with del(13q) who show a good response to immunosuppressive therapy and a high prevalence of increased glycosylphosphatidylinositol-anchored protein(GPI-AP)-deficient cells (Haematologica, 97(12):1845-9, 2012). There may be common mechanisms underlying the preferential commitment between GPI-AP- and del(13q) hematopoietic stem and progenitor cells (HSPCs), like insensitivity to inhibitory cytokines such as TGF-β, in immune-mediated BM failure. To confirm this hypothesis, we studied the effect of OLFM4 knockout on the erythroid differentiation of erythroid leukemia cell lines induced by TGF-β. [Methods] We established OLFM4 knockout (KO) K562 and TF-1 cells using a CRISPR-Cas9 system. OLFM4-knockdown (KD) cells were also prepared using siRNA to validate the results of OLFM4-KO cells. The OLFM4 mRNA and protein levels were determined using quantitative polymerase chain reaction, flow cytometry (FCM), Western blotting, immunocytochemistry, and immunofluorescence methods. The erythroid differentiation was assessed by measuring the expression of glycophorin A (GPA) with FCM, GATA-1 protein expression using Western blotting, and iron staining of the cells. [Results] The OLFM4 KO cells showed slower proliferation than wild-type (WT) cells. Both OLFM4-KO cells and OLFM4-KD cells showed a higher GPA expression than WT cells (median fluorescence intensity [MFI] of K562: 2924 and 2143 vs. 1469 and TF-1: 950 and 870 vs. 694, respectively). OLFM4-KO cells showed erythroid morphology, an elevated expression of GATA-1, and positivity for iron granules, suggesting that OLFM4 KO promoted the erythroid differentiation of K562 and TF-1 cells in RPMI1640 containing 10% fetal bovine serum (Figure 1). When WT cells were cultured in a serum-free culture medium (Steampro34) with or without TGF-β (6 ng/ml) for 8 days, the GPA expression was induced in both TF-1 (MFI: 4358 vs. 883), and K562 cell lines (33440 vs. 25655). The OLFM4 protein levels in these cell lines were significantly decreased by the TGF-β treatment in a dose-dependent manner, suggesting that TGF-β directly downregulated the OLFM4 expression in WT cells; the relative expression of OLFM4 was 1, 0.64, and 0.12 while that of TF-1 was 1, 0.12, and 0.05 at 0, 2, and 6 ng/ml of TGF-β, respectively (Figure 2). [Conclusion] OLFM4 prevents K562 and TF-1 cells from differentiating into erythroid cells in response to TGF-β. The erythroid differentiation of these leukemic cells may be mediated by the downregulation of OLFM4 induced by TGF-β. Haploinsufficiency of OLFM4 due to either a loss of function mutation or del(13q) may be related to the mechanisms underlying the preferential commitment of the mutant HSPCs to erythroid cells in patients with immune-mediated BM failure where TGF-β is abundantly present. Disclosures Yoroidaka: Ono Pharmaceutical: Honoraria. Nakao:Takeda Pharmaceutical Company Limited: Honoraria; Novartis Pharma K.K: Honoraria; Kyowa Kirin: Honoraria; Bristol-Myers Squibb: Honoraria; Janssen Pharmaceutical K.K.: Honoraria; Daiichi-Sankyo Company, Limited: Honoraria; Ohtsuka Pharmaceutical: Honoraria; Alaxion Pharmaceuticals: Honoraria; Ono Pharmaceutical: Honoraria; Celgene: Honoraria; Chugai Pharmaceutical Co.,Ltd: Honoraria; SynBio Pharmaceuticals: Consultancy.

To identify HLA alleles closely involved in the autoantigen presentation in acquired aplastic anemia (AA), we studied the HLA allelic loss frequencies of 312 AA patients, including 43 patients with loss of heterozygosity of 6p chromosome (6pLOH). An analysis of the HLA alleles contained in the lost haplotype revealed HLA-B*40:02 to be the most frequently lost allele. When we examined 28 AA (12 6pLOH[+] and 16 6pLOH[-]) patients with HLA-B*40:02 for the presence of leukocytes lacking HLA-B4002 (B4002-) using a new monoclonal antibody specific to this allele, B4002- granulocytes were detected not only in all 6pLOH(+) patients but also in 9 (56%) of the 16 6pLOH(-) patients. Furthermore, 10 (83%) of the 12 6pLOH(+) patients possessed 1.0% to 78% B4002- granulocytes that retained the HLA-A allele on the same haplotype (B4002-A+), suggesting the frequent coexistence of granulocytes that underwent mutations restricted to HLA-B*40:02 with 6pLOH(+) (B4002-A-) granulocytes. Deep sequencing of the HLA-B*40:02 of sorted B4002-A+ granulocytes revealed various somatic mutations, such as frameshift, nonsense, and splice site mutations, in all 15 patients studied. Surprisingly, missense mutations in the α-3 domain of HLA-B*40:02 that are not involved in the antigen presentation were detected exclusively in the B4002+ granulocytes of 3 patients possessing B4002- granulocytes. The markedly high prevalence of leukocytes lacking HLA-B4002 as a result of either 6pLOH or structural gene mutations, or both, suggests that antigen presentation by hematopoietic stem/progenitor cells to cytotoxic T cells via the HLA-B allele plays a critical role in the pathogenesis of AA.

[Background] Leukocytes that lack HLA class I alleles derived from hematopoietic stem progenitor cells (HSPCs) that undergo copy number neutral loss of heterozygosity of the short arm of chromosome 6 (6pLOH) or HLA allelic mutations are often detected in acquired aplastic anemia (AA) patients. The presence of HLA class I allele-lacking leukocytes provides compelling evidence that CTLs are involved in the development of AA, but the precise mechanisms underlying HLA missing and clonal hematopoiesis by such HLA(-) HSPCs are unknown. Our recent study showed that, among several HLA-class I alleles that are likely to be lost as a result of 6pLOH, HLA-B*40:02 is the most frequently lost allele in Japanese AA patients. The study also showed that B*54:01 was one of three HLA-alleles that were most likely to be possessed by 6pLOH(+) patients (29% [5/17]) when only patients not carrying HLA-B*40:02 were analyzed. These results prompted us to study the role of HLA-B*54:01 in the pathogenesis of AA in a larger number of patients. [Method] To identify HLA class I alleles other than HLA-B*40:02 that are closely involved in the auto-antigen presentation in AA, we studied leukocytes of 549 AA patients for the presence of 6pLOH as well as HLA alleles that are lost due to 6pLOH. To gain insight into the mechanism underlying clonal hematopoiesis by HLA-B*54:01-lacking HSPCs, we studied HSPCs derived from induced pluripotent stem cells (iPSCs) that were generated from an AA patient possessing B5401-lacking monocytes. We also investigated the association between male AA patients possessing B*54:01 and CAG microsatellites of androgen receptor (AR) gene which are related to transactivation of the AR gene. [Results] 6pLOH was detected in 91 (16.6%) of the total patients and in 48 (10.4%) of the 462 patients not possessing B*40:02. Among the HLA alleles possessed by the 48 patients, B*54:01 was the most frequent (23%). 6pLOH was detected in 17 (34%) of 50 patients possessing B*54:01, and the incidence was markedly higher in males (15/24, 62.5%) than in female patients (2/26, 7.7%, P Disclosures Yoshida: Nihon-shinyaku: Research Funding. Akashi:Novartis pharma: Research Funding; Bristol-Myers Squibb: Research Funding, Speakers Bureau; MSD: Research Funding; Eisai: Research Funding; Asahi-kasei: Research Funding; Ono Pharmaceutical: Research Funding; Pfizer: Research Funding; Celgene: Research Funding, Speakers Bureau; sanofi: Research Funding; Chugai Pharma: Research Funding; Otsuka Pharmaceutical: Research Funding; Astellas Pharma: Research Funding; Taiho Pharmaceutical: Research Funding; Kyowa Hakko Kirin: Research Funding, Speakers Bureau; Eli Lilly Japan: Research Funding. Nakao:Novartis: Honoraria; Kyowa Hakko Kirin Co., Ltd.: Honoraria; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria.

[Background] Hematopoietic stem progenitor cells (HSPCs) with PIGA mutations are thought to acquire a survival advantage over normal HSPCs under immune attack against HSPCs and produce glycosylphosphatidylinositol-anchored protein-deficient (GPI[-]) cells in patients with acquired aplastic anemia (AA). Various underlying mechanisms of the survival advantage of PIGA-mutated HSPCs have been proposed; however, it remains still unclear how PIGA-mutated HSPCs are immunologically selected in AA. Approximately 15% of AA patients with increased GPI(-) cells possess another aberrant leukocyte subset that lacks the expression of the HLA-class I allele due to a copy number-neutral loss of heterozygosity of the HLA haplotype, which occurs in the short arm of chromosome 6 (6pLOH) as a result of uniparental disomy, or HLA allelic mutations. The presence of HLA-class I allele-lacking leukocytes (HLA-LLs) is considered to be the most compelling evidence to support the involvement of cytotoxic T lymphocytes (CTLs) in the development of bone marrow failure. Charactering GPI(-) leukocytes and platelets in AA patients with HLA-LLs may provide an insight into the mechanism underlying the immune selection of PIGA-mutated HSPCs. [Patients and Methods] We investigated the presence of GPI(-) leukocytes, erythrocytes, and platelets in 63 patients with AA using high-sensitivity flow cytometry (FCM). For the platelet analysis, platelet rich plasma (PRP) was obtained by centrifuging anticoagulated blood at 1000 rpm for 7 minutes with the brake turned off. Thirty microliters of PRP was incubated with monoclonal antibodies specific to CD55-PE, CD59-PE, CD41a-APC and HLA-A2 or A24-FITC for 20 minutes at room temperature in the dark. To prevent doublets, samples were diluted 1 to 100 in PBS and filtered with mesh immediately before the FCM analysis. Thirty of the 63 patients were heterozygous for the HLA-A allele with A24 and A2, and thus the presence of both HLA-LLs and HLA-A allele-lacking platelets could be evaluated by FCM. The lack of the HLA-A allele due to 6pLOH or allelic mutations in all HLA-LL(+) patients was confirmed by a droplet digital PCR or deep sequencing. [Results] Increased GPI(-) granulocytes, which accounted for 0.01-99.8% of the total granulocytes, were detected in 37 (58.7%) patients while HLA-A24 or A2-lacking granulocytes accounted for 0.39-98.3% of the total granulocytes in 20 (66.7%) of the 30 patients. Eight patients possessed both GPI(-) cells and HLA-LLs. In all 8 of these patients, the two aberrant cell populations were mutually exclusive. The analyses of different cell lineages revealed HLA-A allele-lacking cells in all lineages of cells, including granulocytes (Gs), monocytes (Ms), T cells (Ts), B cells (Bs), NK cells (NKs), and platelets (Ps) in 7 of the 8 patients; the remaining one patient had the GMTP pattern. In contrast, the lineage diversity of GPI(-) in the 8 patients was more restricted; GMTBNKP was only detected in 2 patients; the combinations in the other 6 patients were GT (n=1), GMBNKP (n=2), GMTNKP (n= 1) and GMTBP (n= 2). In Case 1, GPI(-) cells were not detected in T cells while HLA-A24(-) cells were detected in all lineages of cells including T cells (Figure 1). The limited lineage diversity of GPI(-) cells was also evident in 6 patients who did not possess HLA-LLs (GMP, GMBP, GMBNKP, GMTNKP, GMTBP) with GPI(-) granulocytes>10% while the GMTBNKP pattern was common in 10 HLA-LL(+) patients who did not possess GPI(-) cells, regardless of their percentage of HLA-A allele-lacking granulocytes. Longitudinal follow-up of 5 patients over a period of 8-27 years showed a decline in the percentage of GPI(-) granulocytes (39.2 to 0.00%, 11.4 to 0.04%, 3.50 to 0.30%, 1.77 to 0.00% and 0.79 to 0.11%) and a reciprocal increase in the percentage of HLA-A allele-lacking granulocytes (80.0 to 95.2%, 92.0 to 99.1%, 24.0 to 24.4%) in 3 patients who had been placed under observation; in two patients (Cases 2 and 3) whose GPI(-) granulocyte percentages had been >10%, the PNH clones were completely replaced by HLA-LL clones during 6 and 8 years, respectively (Figure 2). [Conclusions] The limited diversity of the blood cell lineage and spontaneous decline of GPI(-) cells that coexisted with HLA-LLs suggest that GPI(-) cells are derived from the PIGA-mutated hematopoietic progenitor cells that were allowed to proliferate as a bystander in the environment where the CTL attack against HSPCs is taking place. Disclosures Nakao: Novartis: Honoraria; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria; Kyowa Hakko Kirin Co., Ltd.: Honoraria.

Background: Dysregulation of T-cell mediated immunity is responsible for acquired pure red cell aplasia (PRCA). Although STAT3 mutations are frequently detected in patients with T-cell large granular lymphocytic leukemia (T-LGLL), which is often complicated by PRCA and is also reported to be associated with acquired aplastic anemia (AA) and myelodysplastic syndrome (MDS), whether STAT3-mutated T cells are involved in the pathophysiology of PRCA and other types of bone marrow failure (BMF) remains unknown. Methods: Patients with AA, AA-paroxysmal nocturnal hemoglobinuria (AA-PNH) syndrome, MDS or PRCA were enrolled in this study. We performed STAT3 mutation analyses of the peripheral blood mononuclear cells using an allele-specific PCR (AsPCR) to detect STAT3 Y640F or D661Y mutations and amplicon sequencing using primers covering entire coding region of STAT3. CD4+ T cells, CD8+ T cells or granulocytes were sorted, if possible, and also subjected to the analyses. The T-cell receptor (TCR) Vβ repertoire was analyzed using whole blood samples by flow cytometry. Results: A total of 124 patients including PRCA (n=42), AA (n=54), AA-PNH (n=7) and MDS (n=21) were enrolled. The subtypes of PRCA were as follows: idiopathic, n= 15, T-LGLL-associated, n=13; thymoma or thymic cancer-associated, n=7; autoimmune disease-associated, n=5; adverse drug reactions, n=1; and human parvovirus B19 infection complicated by T-LGLL, n=1. As an initial step in the STAT3 mutational analysis, we screened all patients with an AsPCR. Among the 42 patients with PRCA, 3 (10%) of 29 patients without T-LGLL and 6 (46%) of 13 patients with T-LGLL were positive for mutations. In contrast, none of the patients with AA, MDS or AA-PNH were positive (P value= 0.000031). Then, we examined MNC-derived DNA from 73 patients (PRCA, n=42 and AA, n=31) using amplicon sequencing. In this sequencing analysis, the median depth of coverage was 6,219x (range; 1,065-14,188). No STAT3 mutations were detected in 31 AA patients. In contrast, 15 (36%) PRCA patients possessed STAT3 mutations. The variant allele frequency of STAT3 mutations ranged from 0.0057 to 0.489. In all of the 7 patients studied, the STAT3 mutations were restricted to sorted CD8+ T cells. Three patients were negative for STAT3 mutations in MNCs but found to be positive when sorted CD8+ T cells were analyzed. The prevalence of STAT3 mutation in idiopathic, thymoma-associated, autoimmune disorder-associated and T-LGLL-associated PRCA was 33% (5/15), 29% (2/7), 20% (1/5), and 77% (10/13), respectively. In total, STAT3 mutations were detected in 8 of 29 (28%) PRCA patients without T-LGLL and 10 of 13 (77%) PRCA patients with T-LGLL. When TCRVβ repertoires of CD8+ T cells sorted from 3 STAT3 mutation(+) PRCA patients without T-LGLL were analyzed, skewed TCRVb repertoires were evident in all patients, and STAT3 mutations were detected in skewed TCRVβ fractions from 2 patients whose samples were available for cell sorting. The STAT3-mutation(+) patients were younger (median age of 63 years vs 73 years, P= 0.026) and less responsive to cyclosporine (CsA) (46% [6/13] vs 100% [8/8], P= 0.0092) in comparison to STAT3-mutation(-) patients. Of note, 4 of 8 STAT3 mutation(+) patients who had been refractory to CsA were treated with CY and all of them responded well, whereas none of 8 STAT3 mutation(-) patients required secondary treatment with CY owing to a sustained response to CsA. Discussion: This study is the first to reveal frequent STAT3 mutations in PRCA patients, even in thouse without T-LGLL, using a large cohort of patients. The presence of STAT3-mutated CD8+ T cells may be unique background of PRCA, irrespective of disease etiology. Poor response to CsA in STAT3 mutation(+) patients suggests that STAT3-mutated CD8+ T cells may be less sensitive to the inhibitory effects of CsA than non-mutated CD8+ T cells. In contrast, our analyses failed to detect STAT3 mutations in any of 52 AA patients, suggesting that STAT3 mutation(+) T cells had little impact on the development of AA in Japanese patients. Conclusion: STAT3 mutations are frequently detected in the CD8+ T cells of PRCA patients, regardless of the presence of T-LGLL. The identification of STAT3 mutations may be useful for appropriately managing patients with PRCA. Figure. Figure. Disclosures Nakao: Novartis: Honoraria; Kyowa Hakko Kirin Co., Ltd.: Honoraria; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria.

[Background] Acquired aplastic anemia (AA) is a rare syndrome characterized by pancytopenia and bone marrow hypoplasia. The cytotoxic T lymphocyte (CTL) attack against autologous hematopoietic stem progenitor cells (HSPCs) is thought to be responsible for bone marrow failure in the majority of AA cases; however, little is known about the target antigens of the CTLs. HLA class I-allele lacking leukocytes (HLA-LL) due to copy-number neutral loss of heterozygosity in the short arm of chromosome 6 (6pLOH) or somatic loss-of-function mutations in HLA class I genes are detected in approximately 20% of patients with newly diagnosed AA, and the presence of HLA-LL represents compelling evidence to support that CTLs specific to HSPCs are involved in the development of AA. Our recent studies using single nucleotide polymorphism array (SNP-A) genotyping and droplet digital polymerase chain reaction (ddPCR) revealed that HLA-B*40:02 is the most frequently lost among all class I alleles that are lost as a result of 6pLOH (Zaimoku, et al. Blood 2017). Various somatic loss-of-function mutations in B*40:02 revealed by deep sequencing in the study substantiated the important role of HLA-B4002 in the autoantigen presentation of AA. However, in the other 6pLOH(+) AA patients who did not possess HLA-B4002, which accounted for 20% of the total AA cases involving patients possessing HLA-LL, the allele in the missing haplotype that was responsible for the autoantigen presentation was largely unknown because the lost fragment of chromosome 6p usually contained 2 or more HLA class I alleles. [Objectives/Methods] To identify class I alleles other than HLA-B*40:02 that are critically involved in the auto-antigen presentation of AA, we screened a total of 624 patients for the presence of HLA-LL using monoclonal antibodies specific to class I HLA alleles, SNP-A, and ddPCR, and performed targeted deep sequencing of HLA class I genes by using SeqCap EZ Choice pobes (Roche) and MiSeq sequencer (Illumina). The paired fractions, including granulocytes that lacked an HLA-A allele and granulocytes that retained the HLA-A allele, as well as CD3+ T cells, were sorted using monoclonal antibodies specific to HLA-A alleles with a BD FACSAria Fusion system (BD Biosciences), and were subjected to DNA extraction. All DNA samples of granulocytes and control cells (CD3+ T cells or buccal mucosa cells) were prepared for targeted deep sequencing. [Results] One hundred and fourteen patients were found to be positive for HLA-LL and 62 (54.4%) of the 114 HLA-LL(+) patients did not carry B*40:02 (severe, n=30; non-severe, n=32; male, n=38; female, n=24; median age, 62 [range, 6-93] years). Apart from B*40:02 (45.6%), A*02:06 (24.6%) was the second-most frequent HLA class I allele in the lost haplotype. The targeted deep sequencing of 20 patients with HLA-LL revealed 6pLOH alone in 11 patients, and somatic loss-of-function mutations plus 6pLOH in 9 patients; none of the patients were positive for somatic loss-of-function mutations alone. Of note, somatic loss-of-function mutations were found in only 5 alleles (A*02:06 in four, B*40:01 in two, B*40:03, A*31:01, and B*54:01 in one each) out of 27 different alleles contained in the lost haplotype. Among the 9 patients with somatic loss-of-function mutations, the median number of mutations per patient was 1 (range, 1-2); these included a missense mutation (n=1), frameshift deletions (n=3) and nonsense mutations (n=7) (Figure). Four patients had a breakpoint of 6pLOH in between the HLA-A and C loci; their lost alleles were A*02:06 (n=2) and A*31:01 (n=2), and the occurrence of 6pLOH in the four patients was therefore attributed to the two HLA-A alleles. Sixty-six percent of the HLA-LL(+) B*40:02(-) patients had at least one of the five alleles in the lost haplotypes. The frequencies of each "high risk" allele found in patients possessing HLA-LL are summarized in Table. [Conclusions] In addition to B*40:02, five class I alleles including HLA-A*02:06, A*31:01, B*54:01, B*40:03 and B*40:01 are thought to play an essential role in the auto-antigen presentation by the HSPCs of Japanese AA patients. The frequencies of the six class I alleles in general Japanese population are much higher than those in the general Caucasian populations but similar to the frequencies in East Asian populations. The higher frequencies of the six alleles in comparison to Caucasian countries may account for the higher incidence of AA in East Asia. Disclosures Takamatsu: Ono: Research Funding; Bristol-Myers Squibb: Research Funding; Janssen: Honoraria; Celgene: Honoraria, Research Funding. Nakao:Novartis: Honoraria; Alexion Pharmaceuticals, Inc.: Consultancy, Honoraria; Kyowa Hakko Kirin Co., Ltd.: Honoraria.

[Background] The frequent loss of heterozygosity of the HLA haplotype in the short arm of chromosome 6 (6pLOH) in leukocytes is thought to offer compelling evidence of cytotoxic T cell (CTL) involvement in the development of acquired aplastic anemia (AA) because it represents the escape of hematopoietic stem/progenitor cells (HSPCs) with 6pLOH from the attack of CTLs that are specific to autoantigens presented by the lacked HLA class I allele. Although our previous study suggested that HLA-B*40:02 is the major allele involved in this phenomenon, the exact role of B*40:02 remained unclear because 6pLOH involving this allele is always associated with a lack of HLA-A and C alleles in the haplotype, and the presence of B*40:02-missing leukocytes were unable to be shown due to the lack of monoclonal antibodies (mAbs) specific to B61, the HLA-B antigen that corresponds to B*40:02. We recently succeeded in generating a mAb specific for HLA-B61 that enabled us to explore the role of B*40:02 in the development of AA. [Methods] Using the new mAb, we examined peripheral blood samples of 28 AA (12 with 6pLOH and 16 without 6pLOH) patients carrying this allele for the presence of B61(-) leukocytes using flow cytometry. HLA genes were enriched by sequence capture, a hybridization-based gene enrichment method, from genomic DNA of sorted B61(-) granulocytes, and were subjected to deep sequencing using an NGS (MiSeq). B61(+) granulocytes or T cells were used as controls. Potential mutations responsible for the B61-missing were identified when 10 or more variant reads were found only in B61(-) granulocytes. Thereafter, HLA-B alleles carrying those mutations were determined taking advantage of the nearest allele-specific SNPs. [Results] Among the 12 6pLOH(+) patients, 10 (83%) possessed 0.5%-60% B61-missing granulocytes that were not lacking HLA-A, in addition to 12% to 99% 6pLOH(+) granulocytes that lacked both B61 and an HLA-A allele on the same haplotype (Figure 1). B61(-) granulocytes that accounted for 0.5%-99% of the total granulocytes were detected in 9 (56%) of the 16 6pLOH(-) patients. The prevalence of missing B61 in the 28 AA patients was 21/28 (75%), much more frequent than those of the 3 other major alleles (A*02:01, 32%; A*02:06, 30%; A*24:02, 6%). B61(-) granulocytes were available for mutation analyses of HLA-B alleles in 15 of the 19 patients who possessed B61(-) granulocytes. The mean coverage of HLA-B gene was 426x. In total, 43 somatic mutations of HLA-B were identified in B61(-) granulocytes, all of which were present in B*40:02 but not in any of the other HLA-B alleles. Median variant allele frequency was 4.8% (range, 1.0% - 43%) and the number of mutations in each patient was 1 to 6 (Figure 2). Thirty-nine mutations were exonic while 4 were intronic. Exonic mutations included frameshift insertions (n=12), frameshift deletions (n=16), non-frameshift deletions (n=2), nonsense mutations (n=7), a missense mutation (n=1) and a start codon mutation (n=1). All four intronic mutations were considered to be a splice site mutation; two mutations deactivated 5' and 3' splice sites, whereas the other two were single base substitutions within intron 3, making alternative 5' splicing site with strong consensus sequence: GGC [A>G] TGAGT and TTC [C>G] TGAGT. Surprisingly, missense mutations in the alpha-2 and alpha-3 chain-coding region of HLA-B*40:02 were detected exclusively in the B61(+) granulocytes of two patients possessing B61(-) granulocytes, suggesting the inability of the mutant HSPCs to interact with CTLs. Variant allele frequencies of the two missense mutation were 40% and 45%, respectively. As a result of the mutation, virtually all granulocytes of the two patients were affected by B*40:02 mutations that allowed the HSPCs to escape the T cell attack. [Conclusions] The markedly high prevalence of leukocytes lacking HLA-B*40:02 as a result of either or both 6pLOH or structural gene mutations clearly indicates that antigen presentation by HSPCs to CTLs via the HLA-B allele plays a critical role in the pathogenesis of AA. Disclosures Takamatsu: Celgene: Honoraria; Janssen Pharmaceuticals: Honoraria. Nakao:Alexion Pharmaceuticals: Honoraria, Research Funding.

[Background] Clonal hematopoiesis is currently known to be common in patients with acquired aplastic anemia (AA). One of the most common abnormalities underlying clonal hematopoiesis in AA patients is copy-number neutral loss of heterozygosity (LOH) in the short of 6 chromosome (6pLOH) caused by acquired uniparental disomy. Hematopoietic stem/progenitor cells (HSPCs) having undergone 6pLOH are thought to evade attack by cytotoxic T lymphocytes (CTLs) specific to auto-antigens by lacking particular HLA-A alleles. These HSPCs then produce HLA class I allele-lacking [HLA(-)] leukocytes to support hematopoiesis in patients with AA patients in remission. Our recent study showed that HLA(-) granulocytes are detected in about 24% of newly-diagnosed AA patients, and the aberrant granulocytes often account for more than 95% of the total granulocytes and persist for many years. The sustainability of 6pLOH(+) HSPC clones suggests that these HSPCs may suffer from secondary somatic mutations that confer a proliferative advantage on them over normal HSPCs. Alternatively, 6pLOH(+) HSPCs may persist and continue to support hematopoiesis according to their inherent sustainability, just like the PIGA mutant HSPCs we previously described (Katagiri et al. Stem Cells, 2013). To test these hypotheses, we determined the sequences of genes associated with the clonal expansion of HSPCs in HLA(-) granulocytes. [Patients and Methods] Eleven AA patients whose percentages of HLA(-) granulocytes ranged 6.4%-99.8% (median 94.2%) of the total granulocyte population were chosen for this study. The patients (male/female, 5/6 and age 27-79 [median 53] years) had been diagnosed with severe (n=5) or non-severe (n=6) AA 2-25 [median 12.5] years earlier, and 7 and 4 patients achieved complete response and partial response, respectively after treatments with cyclosporine (CsA) alone (n=4), CsA+antithymocyte globulin (ATG, n=3), CsA+anabolic steroids (AS, n=2), AS+romiplostim (n=1), and AS alone (n=1). The lineage combinations of HLA(-) cells were granulocyte, monocytes, B cells and T cells (GMBT) in 6, GMB in 4 and GM in 1. HLA(-) and normal [HLA(+)] granulocytes were sorted from the blood leukocytes of the 11 patients and the DNA of each cell population as well as that of buccal mucosa cells was subjected to target sequencing of 61 myelodysplastic syndrome (MDS)-related genes with MiSeq. DNA samples from 5 patients including 4 patients whose HLA(-) cell percentages were greater than 95% were further analyzed by whole-exome sequencing (WES) using HiSeq. The percentage of 6pLOH(+) cells in the total granulocytes or sorted HLA(-) granulocytes were estimated using digital droplet PCR or deep sequencing of HLA alleles. [Results] Target sequencing of 8 of the 11 patients revealed somatic mutations in the HLA(-) granulocytes of 3 patients. HLA(-) granulocytes-specific mutations were found in DNMT3A, PRR5L, SMC3A, and LRCH1 (Table). The variant allele frequencies (VAF) of these mutations were far lower (5.1%-20%) than those of HLA(-) granulocytes that accounted for 95% of sorted cells. WES revealed 22 non-synonymous and 9 synonymous mutations in the HLA(-) granulocytes from 4 of the 5 patientsthat included 3 new patients and 2 patients whose samples were negative for mutations revealed by the target sequencing. The VAF of these mutations ranged from 20.7-52.5% (median 44.1%, Table). Very-high VAFs of several mutant genes suggested that these mutations occurred simultaneously with or soon after the occurrence of 6pLOH. A patient who achieved remission after romiplostim therapy without ATG showed various gene mutations that were thought to have occurred after 6pLOH. Despite of their highly biased hematopoiesis supported by single or few clones, recurrent or MDS-related oncogenic mutations were not detected in any of the 11 patients. Of note, the percentages of 6pLOH(+) cells in the sorted HLA(-) granulocytes were ≤75% (36.7%, 46%, 74%, and 75%) in 4 patients, indicating the presence of granulocytes lacking HLA-A alleles through mechanisms other than 6pLOH. [Conclusions] HLA(-) HSPCs caused by 6pLOH or other unknown mechanisms support long-term hematopoiesis without the development of oncogenic driver mutations that are associated with clonal hematopoiesis of MDS; as such, clonal hematopoiesis by 6pLOH(+) HSPCs may not portend a poor prognosis. Disclosures Nakao: Alexion Pharmaceuticals: Honoraria, Research Funding.

A T-cell attack against hematopoietic stem cells (HSCs) is believed to be the principal mechanism underlying the development of acquired aplastic anemia (AA). The presence of leukocytes that lack HLA class I alleles as a result of the copy number-neutral loss of heterozygosity of the HLA haplotype due to uniparental disomy in the short arm of chromosome 6 (6pLOH) is compelling evidence for the involvement of cytotoxic T lymphocytes (CTLs) in the HSC destruction; this is based on the high response rate to immunosuppressive therapy (IST) in 6pLOH(+) patients (Katagiri, et al. Blood, 2011). However, target cells of the putative CTLs have not yet been characterized because of the small number of patients analyzed via flow cytometry (FCM) in our previous study. FCM can be substituted for SNP arrays by detecting HLA-A allele-lacking leukocytes (HLA-LLs) caused by 6pLOH. To gain insight into the CTL target responsible for the development of AA, we examined a total of 223 patients (213 idiopathic and 10 hepatitis-associated AA; 93 severe and 130 non-severe AA) for the presence of HLA-LLs and determined the lineage combinations of the aberrant leukocyte population. Of the 223 patients, 145 (65.0%) were heterozygous for the HLA-A allele and could be assessed for the presence of HLA-LLs by FCM. Eighteen (25.4%; 10 with severe AA and 8 with non-severe AA) of the 71 pre-treatment patients, and 26 (35.1%; 13 with severe AA and 13 with non-severe AA) of the 74 post-treatment patients were found to be positive for HLA-LLs. The lineage combinations of HLA-LLs in the 44 HLA-LL(+) patients were granulocytes (Gs), monocytes (Ms), B cells (Bs), and T cells (Ts, GMBT) in 13, GMB in 16 and GM in 11 patients. Surprisingly, HLA-LLs were found in Bs alone in three patients, and in one patient, the lineage combination pattern was TB (Figure). The presence of 6pLOH was confirmed via deep sequencing of isolated Bs from one of the three Bs alone patients. These lineage combination patterns were not observed to change for 1-40 months in 22 of 23 patients whose blood samples were available for follow-up analyses. In one patient with the GMB pattern, HLA-LLs decreased from 11.7% before treatment to 0% after 12 months of ATG therapy. The response rate to IST in GMBT patients (3/4, 75%) and in patients with GMB or GM (9/10, 90%) were similarly higher than in patients without HLA-LLs (22/39, 56%) or in patients who were homozygous for the HLA-A allele (23/36, 64%). Two of the three Bs alone patients showed complete responses at the time of sampling three years after ATG therapy and 20 years after CsA therapy, and another patient had a secondary myelodysplastic syndrome two years after response to ATG. The TB alone patient developed AA 20 years earlier, but had not been treated until recently because there was no need for blood transfusions, and is now improving in response to eltrombapag. This study revealed that the targets of putative CTLs in more than half of AA patients are hematopoietic progenitor cells with limited differentiation but long-lasting capacity, and in some patients, they are lymphoid progenitor cells that do not contribute to hematopoiesis. This suggests that CTL attack against non-HSCs including lymphoid precursors could trigger BM failure. Consistent with our previous report, the bystander effects caused by the immune response to non-HSCs such as myelosuppressive cytokines may play a major role in the development of AA. Disclosures No relevant conflicts of interest to declare.