In this issue of Blood, Kulasekararaj et al describe a subgroup of acquired aplastic anemia (AA) that is characterized by myelodysplastic syndrome (MDS)–like somatic mutations and is at high risk for progression to MDS.1
AA is a prototype of acquired bone marrow failure caused by immune-mediated destruction of hematopoietic stem/progenitor cells,2 whose prognosis has been dramatically improved during the past 3 decades by the development of immunosuppressive therapies (ISTs), especially for those who are not eligible for allogeneic hematopoietic stem cell transplantation (allo-HSCT).3 However, in some cases (up to 15%), even successful ISTs are complicated by the development of MDS and/or acute myeloid leukemia (AML).4 In fact, differential diagnosis between AA, especially of nonsevere forms, and hypoplastic MDS has been a long-standing issue in clinical hematology due to close similarities in their clinical presentation, both showing hypoplastic bone marrow that hampers sufficient morphological evaluation for dysplasia, cytogenetic and other evidence of clonal hematopoiesis, and sizable responses to ISTs.5
Using next-generation sequencing of targeted exons, Kulasekararaj et al explored somatic mutations in bone marrow samples from 150 cases with AA and investigated their impact on progression to MDS as well as correlation to disease duration and telomere attrition.1 Through an initial screening of 835 known genes in 57 discovery cases, followed by focused sequencing of detected targets in 93 extended cases, a total of 32 somatic mutations typically seen in MDS and other myeloid malignancies were detected in 29 of 150 patients (19%), where the predominant mutational targets included ASXL1, DNMT3A, and BCOR. Transformation to MDS occurred in 17 cases, which included 11 of the 29 mutation (+) cases, or 7 of 12 ASXL1-mutated, 3 of 8 DNMT3A-mutated, and 1 of 6 BCOR-mutated cases. Importantly, somatic mutations were significantly associated with longer disease duration, that is, time from diagnosis to sample collection (37 vs 8 months, P < .04), shorter telomere lengths (median telomere-to-single copy gene ratio length, 0.9 vs 1.1, P < .001), and a higher rate of progression to MDS/AML (38% [11 of 29] vs 5.0% [6 of 121], P < .001).
Clonal hematopoiesis in AA has long been discussed based on the presence of cytogenetic abnormalities,6 skewed X-chromosome inactivation in female patients,7 appearance of varying degrees of blood cells having paroxysmal nocturnal hemoglobinuria phenotypes,2 and more recently, recurrent uniparental disomy (UPD) in the 6p arm commonly involving the class I HLA locus.8 It has been speculated that the clonal hematopoiesis may be derived from some “bottleneck” effect caused by hematopoietic repopulation from a severely reduced number of hematopoietic stem cells,2 or represent escaped hematopoiesis from autoimmunity, especially in 6pUPD(+) cases8 and/or premalignant hematopoiesis as has been demonstrated recently in AML.9 The current study, together with a previous report by Lane et al,10 demonstrated that clonal hematopoiesis in AA frequently accompanies somatic mutations commonly seen in MDS/AML and also presented the first implication that these “MDS/AML-like” somatic mutations predict a substantial risk for progression to MDS/AML (∼40%). The findings not only provide intriguing insight into the relationship between clonal hematopoiesis and malignant transformation in AA, but also have significant clinical relevance in terms of choice of therapies, such as early use of allo-HSCT for mutation (+) cases. On the other hand, a number of important issues are raised with regard to the origin and chronological behavior of these mutated clones that frequently heralded clinically relevant MDS. The higher frequency of mutations in patients with longer disease duration and the trend of larger clone sizes and shorter telomere lengths thereof may indicate clonal dominance of the mutated clones over time. However, to identify the exact origin of mutations and their temporal behavior, analysis of carefully fractionated cells using serially collected samples from the diagnosis to overt MDS would be needed. Also, the entire picture of clonal hematopoiesis in AA and its pathogeneic/clinical significance could be better delineated through more exhaustive detection of somatic mutations in an unbiased way, using whole genome/exome sequencing.
Conflict-of-interest disclosure: The author declares no competing financial interests.