Multiple HLA-loss clones are required for hematologic regeneration in immune aplastic anemia

Abstract Background In patients with immune aplastic anemia, hematopoietic stem cells deficient in specific HLA class I alleles can evade cytotoxic T-lymphocyte recognition and regenerate sustained hematopoiesis in the absence of standard anti-thymocyte globulin (ATG)-based immunosuppressive therapy, while HLA-intact clones continue to be eliminated by T cells (Zaimoku et al., ASH 2024). This study aimed to evaluate the clonal architecture of hematopoiesis successfully regenerated via HLA-loss clones in the absence of ATG. Methods A total of 13 patients with aplastic anemia harboring HLA-loss clones who had not received ATG were analyzed. Among them, nine patients achieved a complete response (CR)—defined as a neutrophil count >1 × 10⁹/L, hemoglobin >10 g/dL, and platelet count >100 × 10⁹/L—accompanied by expansion of HLA-deficient cells. In contrast, four patients did not achieve CR despite the presence of HLA-loss clones: two remained in partial response, and two progressed to overt PNH clonal expansion. Cell subpopulations—including HLA-deficient cells, PNH-type cells, residual wild-type cells, and HLA-intact and non–PNH-type T cells (used as controls)—were isolated using FACS Aria Fusion, employing HLA allele-specific monoclonal antibodies and fluorescently labeled proaerolysin. Genomic DNA was extracted from the sorted subpopulations for downstream genomic analyses. Hybridization probe-based targeted sequencing was performed to detect somatic mutations in HLA genes, genes associated with myeloid neoplasms, and genes involved in immune escape, as well as copy number alterations. The assay was specifically optimized for the detection of copy-neutral loss of heterozygosity on chromosome 6 (6pLOH) and its associated breakpoints. Additionally, whole-exome sequencing was conducted using the HLA-loss cell fraction in five patients who achieved CR, and using both the HLA-loss and PNH-type cell fractions in two patients who developed PNH clonal expansion in the presence of HLA-loss clones. Results In the nine CR patients, HLA-deficient cells consisted of a median of 3 (range, 1–5) clones with distinct HLA gene mutations or 6pLOH breakpoints. Five of them, with a median of 3 (range, 2–5) HLA-loss clones, had no additional somatic mutations in non-HLA genes. In two patients who achieved CR through expansion of a single HLA-loss clone, HLA-deficient cells also harbored additional loss-of-function mutations—in HLA-DQB1, SETBP1, and U2AF1 in one patient, and in TET2 in the other. In the remaining two CR patients, who achieved hematologic recovery via two and four distinct HLA-loss clones, additional mutations were observed: DNMT3A and TET2 as founder mutations in one, and a subclonal DNMT3A mutation in the other. In contrast, all four patients who did not achieve CR had HLA-deficient cells composed of a single clone without additional non-HLA gene mutations. Among the two non-CR patients who progressed to PNH, PNH-type cells were driven by one and two distinct dominant PIGA mutations, respectively. The majority of these expanded PNH clones also harbored multiple nonsynonymous mutations, including nonsense mutations in genes such as CELF3, KCNS3, and ZNF488. Conclusion These findings demonstrate that successful hematologic regeneration in aplastic anemia via HLA-loss clones, in the absence of ATG, is frequently driven by multiple distinct hematopoietic stem cell clones with different HLA gene mutations or 6pLOH breakpoints. In contrast, patients with a single HLA-loss clone lacking additional somatic alterations failed to achieve CR. These results suggest that multiple escape clones or additional driver mutations are necessary to achieve complete hematopoietic regeneration.