HLA associations, somatic loss of HLA expression, and clinical outcomes in immune aplastic anemia

Abstract

Immune aplastic anemia (AA) features somatic loss of HLA class I allele expression on bone marrow cells, consistent with a mechanism of escape from T-cell-mediated destruction of hematopoietic stem and progenitor cells. The clinical significance of HLA abnormalities has not been well characterized. We examined the somatic loss of HLA class I alleles and correlated HLA loss and mutation-associated HLA genotypes with clinical presentation and outcomes after immunosuppressive therapy in 544 AA patients. HLA class I allele loss was detected in 92 (22%) of the 412 patients tested, in whom there were 393 somatic HLA gene mutations and 40 instances of loss of heterozygosity. Most frequently affected was HLA-B*14:02, followed by HLA-A*02:01, HLA-B*40:02, HLA-B*08:01, and HLA-B*07:02. HLA-B*14:02, HLA-B*40:02, and HLA-B*07:02 were also overrepresented in AA. High-risk clonal evolution was correlated with HLA loss, HLA-B*14:02 genotype, and older age, which yielded a valid prediction model. In 2 patients, we traced monosomy 7 clonal evolution from preexisting clones harboring somatic mutations in HLA-A*02:01 and HLA-B*40:02. Loss of HLA-B*40:02 correlated with higher blood counts. HLA-B*07:02 and HLA-B*40:01 genotypes and their loss correlated with late-onset of AA. Our results suggest the presence of specific immune mechanisms of molecular pathogenesis with clinical implications. HLA genotyping and screening for HLA loss may be of value in the management of immune AA. This study was registered at clinicaltrials.gov as NCT00001964, NCT00061360, NCT00195624, NCT00260689, NCT00944749, NCT01193283, and NCT01623167.

© 2021 by The American Society of Hematology.

Figures

Graphical abstract

Figure 1.

Patients and clinical parameters studied. Correlations of clinical parameters with HLA genotype and HLA loss were tested in subgroups based on the availability of data.

Figure 2.

HLA class I allele loss. (A) Number of patients with HLA loss due to HLA gene mutations, 6p LOH, or both in a total of 412 subjects tested. (B) The number and proportion of patients who lacked individual HLA alleles (left); the total and mean number of clones with HLA gene mutations and 6p LOH among patients who lost respective HLA alleles (right). The number of patients tested for the presence of HLA loss is shown in parenthesis. (C) Combined clone size of cells with HLA loss (upper) and the clone size of individual HLA gene mutations and 6p LOH among cells with HLA loss (lower). (D) Positions and types of somatic inactivating mutations in HLA-A (n = 123) and HLA-B (n = 267). Hotspot mutations are noted on the figure.

Figure 3.

High-risk clonal evolution. Cumulative incidence of high-risk clonal evolution, defined as the acquisition of either chromosome 7 abnormalities, complex cytogenetics, myelodysplastic syndrome, or AML after institution of hATG-based IST, are shown according to the presence or absence of HLA-B*14:02 genotype (A), the presence or absence of HLA loss (B), age groups (C), and 3 risk groups of the prediction model for high-risk clonal evolution incorporating HLA-B*14:02 genotype, HLA loss, and age (D): a high-risk group, any HLA risk (HLA-B*14:02 genotype or HLA loss) present and aged 40 years or older; a low-risk group, no HLA risk present and aged less than 40 years; and an intermediate-risk group, not meeting the criteria for groups of high-risk nor low-risk.

Figure 4.

Clonal evolution of monosomy 7 from HLA class I allele-lacking clones. (A) A 56-year-old female at the time of severe AA diagnosis, who received hATG, CsA, and mycophenolate mofetil (MMF), and multiple rounds of salvage IST, including rabbit ATG, daclizumab and danazol, and alemtuzumab, without adequate response, had clonal evolution to monosomy 7 approximately 7.5 years after her initial diagnosis of AA and died 6 months later. Her blood, sampled 4 years after the institution of initial IST, contained a PNH clone and HLA-A0201-lacking cells (A0201-), consistent with the diagnosis of immune AA. Multiple clones with HLA-A*02:01 deactivating mutations constituted sorted A0201- cells (multiple clones with missense mutations also did A0201+ cells), from which chromosome 7 deletion was not detected by fluorescent in situ hybridization (FISH). After 3.5 years, virtually all hematopoiesis was replaced by a preexisting A0201- clone with an intron 1 mutation (intron 1mut; c.74-19A>G) that had acquired monosomy 7. (B) A 72-year-old man at the time of severe AA diagnosis was treated with hATG, CsA, and EPAG with a partial response at 6 months but had declining cytopenia which responded to reinitiation of CsA. About 2.5 years from the initial diagnosis and IST treatment, BM cytogenetics revealed monosomy 7, and myelodysplastic syndrome was diagnosed. The patient received symptomatic treatment (infrequent blood transfusions twice a year) and died of pneumonia 3.5 years later. HLA flow cytometry of cryopreserved cells revealed a gradual expansion of a clone partially lacking HLA-B4002 (B4002dim), which was attributed to an HLA-B*40:02 intron 1 mutation (c.74-9C>A). A small monosomy 7 clone that was not visible by conventional BM cytogenetics by Giemsa-banding was detected by FISH from the sorted B4002dim cells at least 2 years before clinical diagnosis of monosomy 7. NA, not assessed (due to deficient cryopreserved cells for analysis). See also supplemental Tables 8 and 9.