Of donors and dynamics Free

In this issue of Blood, Karpova et al describe how frequent blood donation subtly shapes the hematopoietic stem cell (HSC) landscape, evidenced by positive selection of HSC clones bearing particular mutations associated with distinct systemic stresses.¹ Whole blood donation is vital to ensure the availability of life-saving blood products and regularly relies on groups of healthy adults who routinely donate blood over their lifetime. The impact of frequent blood donation on the acquisition of, and selection for, mutated HSCs remains largely unknown and is of particular interest in the wake of increasing studies linking clonal hematopoiesis, inflammation, and aging.

Clonal hematopoiesis (CH) is a phenomenon typically associated with aging, whereby single HSC clones bearing a set of hallmark somatic mutations are dramatically overrepresented compared with their nonmutant counterparts.^2,3 Although CH is a common observation during aging, the outgrowth of clones has also been associated with an increased risk of developing a number of diseases, including cardiovascular disease and blood cancers.⁴ Mutations in TET2 and DNMT3A are amongst the most frequent in CH, with significant loss of HSC diversity observed beyond the seventh decade of life.⁵ In contexts where blood outputs are under substantial strain by an external force, such as frequent large volume phlebotomy, it is reasonable to postulate that the selection pressures driving CH may be altered.

In their study, Karpova and colleagues studied HSC mutational status in a cohort of 217 frequent whole blood male donors (>100 lifetime donations, 60 to 72 years of age) and an age- and sex-matched cohort of 212 healthy sporadic donors (<10 lifetime donations, 60 to 70 years of age). They reported that, on average, the rate of observed CH is highly similar between the frequent donor cohort and low-frequency donor controls. This similarity held both with respect to mutational frequencies and the types of genes mutated. Curiously, however, although TET2 and DNMT3A mutations were the most frequently detected mutations in both cohorts, specific types of DNMT3A mutations were enriched in blood from the frequent donor cohort. Although myeloid cancer–associated DNMT3A hotspot mutations (DNMT3A^R882) were detected at low levels in both cohorts, the authors report increased frequencies of DNMT3A frameshifts, premature stop codons, and structural variants in the frequent donor cohort, all of which they propose lead to reduced enzymatic activity. In contrast to the myeloid cell bias observed for DNMT3A^R882 mutations, these latter DNMT3A variants were detected in all cell lineages, including lymphocytes, indicating that the mutant HSCs are multipotent and actively contributing to hematopoiesis.

One of the most provocative findings was that specific DNMT3A mutations had distinct responses to erythropoietin (EPO), highlighting a new and functionally distinct class of DNMT3A variants. EPO is a hormone that stimulates bone marrow erythropoiesis and, along with changes in iron and hemoglobin concentrations, transient increases in EPO expression are a well-described consequence of blood donation.^6-8 Although the authors do not provide evidence of consistent EPO bursts across time, or explore whether individuals in the frequent donor cohort have higher EPO levels than controls, they do investigate response to EPO in an in vitro model system where they introduced 3 different DNMT3A mutations into primary human HSCs using CRISPR/Cas9. In this latter assay, DNMT3A-mutated clones preferentially expanded in EPO-rich medium compared with media supplemented with pro-inflammatory molecules such as interferon-γ or lipopolysaccharide. The authors suggest that these specific DNMT3A mutations driving reduced enzymatic activity provide HSCs with a selective advantage that manifests itself in the context of increased EPO exposure. Moreover, using a xenograft model infused with DNMT3A mutant human HSCs (DMNT3A^W305∗), the authors also show that frequent bleeding and increased human EPO exposure led to cells contributing to both the myeloid and lymphoid lineages.

In some respects, the results of this study are quite reassuring, providing the first evidence that frequent blood donation does not dramatically accelerate CH. However, complete donor safety with respect to HSC dynamics and other health metrics cannot be definitively concluded due to the relatively small cohort size and a number of other factors related to study design. First, as acknowledged by the authors in the Discussion, there may be frequent donors who develop disease who would not be captured in this healthy cohort. Second, it remains unclear whether sex differences may exist---for example, how would the male frequent donors compare with an older female frequent donor cohort? Third, the study does not address whether the specific CH mutations observed in frequent donors would be selected against if the individuals stopped donating blood. These deficiencies aside, the study is extremely thought-provoking and sets the stage for other potential questions to be asked in the future, especially related to the notion of specific selective pressures dictating the type of mutations that grow out. For example, it would be of interest to ask whether some women are more likely to be high EPO responders due to menstruation and, if so, whether a proportion of the population may be well suited to frequent blood donation because they have these mutant HSCs. It would also be of interest to work out the mechanisms of EPO sensitivity and determine whether methylation changes in such HSCs may be linked to the capacity for quickly responding to an increased need for blood production.

Either way, moving forward, it now becomes essential to study these phenomena in larger cohorts and, ideally, to perform these and related studies with more sensitive sequencing methods, such as targeted duplex sequencing, and identify other mutant clones that may have highly selective environmental triggers. Such studies would be particularly relevant for those clones harboring mutations in DNMT3A and TET2 and may also be related to the selection of DNMT3A mutations observed in other contexts such as gene therapy for sickle cell disease⁹ or after transplantation.¹⁰ Moreover, developing a catalog of CH mutations and their specific triggers may also help inform future therapeutic efforts in an array of diseases where the mature blood cell progeny bearing CH mutations may be influencing the immune cell microenvironment.

Conflict-of-interest disclosure: The authors declare no competing financial interests.