Dnmt3a Mutation Confers a Selective Advantage Specifically to Cells within the Long-Term Hematopoietic Stem Cell (LT-HSC) Compartment

The mechanisms by which specific somatic mutations promote the competitive advantage of hematopoietic stem and progenitor cells (HSPCs), particularly in the context of clonal hematopoiesis (CH), is poorly understood. CH in the healthy aging population has been suggested to originate from hematopoietic stem cells (HSCs), supported by data showing that somatic mutations occur at similar variant allele frequencies (VAF) across mature hematopoietic cells of all hematopoietic lineages, as well as HSPCs. However, this data could also be consistent with CH-associated somatic mutations conferring long-term self-renewal potential onto multipotent progenitor cells.

We recently developed an inducible knockin mouse model (Dnmt3a^R878H/+Mx1-Cre^tg/o) to recapitulate the hotspot mutation DNMT3A^R882H observed in human CH and AML (Loberg et al., Leukemia 2019) in order to empirically determine the cell type(s) within the HSPC compartment that gain a competitive advantage following acquisition of a CH-associated somatic mutation. In this model, we observed that the Dnmt3a^R878H mutation expands multiple populations within the HSPC compartment with distinct kinetics. At 6 weeks post-induction of the mutation, we observed significant expansion of multipotent progenitor populations MPP2 and MPP3, both in frequency and in total cell number, compared to Mx1-Cre controls. At 6 months post-induction of the mutation, we observed significant expansion of long-term (LT)-HSCs, short-term (ST)-HSCs, and MPP3 cells. Furthermore, serial colony-forming unit (CFU) assays revealed that Dnmt3a^R878H-mutant LT-HSCs, ST-HSCs, and MPP3 cells, but not MPP4 cells, had increased replating capacity compared to their control counterparts. This data suggested that several distinct populations of HSPCs might be capable of gaining an enhanced competitive advantage with the Dnmt3a^R878H mutation. To determine which of the expanded Dnmt3a^R878H-mutant HSPC populations gained a selective advantage in vivo, we performed competitive transplantation studies of purified subpopulations of HSPCs. Upon transplant, Dnmt3a^R878H-mutant LT-HSCs demonstrated a competitive engraftment advantage compared to control LT-HSCs. In contrast, Dnmt3a^R878H-mutant ST-HSCs and MPP3 cells showed significantly impaired engraftment compared to their control counterparts. These data definitively demonstrate that LT-HSCs are the cells driving Dnmt3a^R878H-mutant CH.

To uncover the molecular mechanisms by which this mutation confers a competitive advantage specifically to LT-HSCs, we performed RNA-seq on Dnmt3a^R878H-mutant and Mx1-Cre control LT-HSCs, ST-HSCs, MPP3, and MPP4 cells at 6 weeks post-induction of the mutation. Our analysis revealed that LT-HSCs had the greatest number of transcriptional alterations when comparing Dnmt3a^R878H-mutant to control cells (412 genes), while ST-HSCs, MPP3 and MPP4 cells had few significantly differentially expressed genes when comparing Dnmt3a^R878H-mutant to control cells (15, 3 and 4 genes, respectively). Of the 412 genes significantly differentially expressed in LT-HSCs, 396 were downregulated and 16 were upregulated in Dnmt3a^R878H-mutant versus control cells. Downregulated genes were enriched in signatures of mature hematopoietic cell differentiation and function, including heme metabolism, immune system processes, and inflammatory response, all consistent with a mechanism of impaired differentiation capacity of Dnmt3a^R878H-mutant LT-HSCs. Upregulated genes included several associated with lipid biosynthesis, metabolism and glycosylation (Vldlr, Gcnt2, Fut8, Uprt, Scd), genes relevant to HSC function (Angpt1, Kit), and the tumor suppressor Rasa1, indicating that these mechanisms and processes may directly contribute to the selective advantage of Dnmt3a^R878H-mutant LT-HSCs.

Collectively, our data demonstrate that the LT-HSC compartment specifically and selectively gains a competitive advantage upon acquiring a somatic mutation that commonly occurs in CH and AML. Furthermore, Dnmt3a^R878H does not confer long-term in vivo self-renewal potential onto ST-HSCs or multipotent progenitor cells. Given the cell-context specificity of Dnmt3a^R878H mutation, mechanistic studies focused on the processes highlighted above should be performed specifically within LT-HSCs to identify therapeutic targets to prevent CH.