The balance between self-renewal and differentiation of hematopoietic stem cells (HSCs) is strictly regulated to sustain blood production throughout adult life. De novo DNA methyltransferase 3-alpha (DNMT3A) is one of the major epigenetic regulators that is essential for efficient HSC differentiation. DNMT3A mutations are prevalent in myeloid diseases that include acute myeloid leukemia (AML; ~22%) and myelodysplastic syndrome (MDS; ~10%) where they act as initiating events However, the precise molecular mechanisms of how DNMT3A regulates normal hematopoiesis and its mutations prime HSCs for leukemic formation are unclear. Although DNMT3A is described as a DNA methyltransferase enzyme, the lack of consistent correlation between changes in DNA methylation and differential gene expression in Dnmt3a-null HSCs in mouse models, and AML patients with DNMT3A mutations undermine the conventional understanding of DNMT3A's canonical role in hematopoietic cells. Hence, we hypothesized that DNMT3A may have novel functions outside of DNA methylation that regulate HSC fate decisions.

To answer this question, we first ectopically expressed GFP-labeled Dnmt3a constructs (wild-type Dnmt3a, Dnmt3aE752A; complete DNA methylation dead, and Dnmt3aR832A; reduced DNA methylation target recognition) and empty vector (negative control) in Dnmt3a-null (Vav-Cre: Dnmt3afl/fl = Dnmt3a-/- in hematopoiesis) bone marrow (BM) cells. The result showed that similar to restoring wild-type Dnmt3a, ectopic expression of Dnmt3aE752A as well as Dnmt3aR832A showed a rescue effect of decreased engraftment of transduced cells in the peripheral blood as well as reduced HSC numbers in the BM. Analysis of DNA methylation by whole-genome bisulfite sequencing (WGBS) in transduced cells showed this phenotypic and functional rescue of the Dnmt3a-/- phenotype occurred in the absence of restored DNA methylation patterns. To study the importance of Dnmt3a-mediated DNA methyltransferase activity in a more physiological system, we generated knock-in mice that have one copy of either wild-type Dnmt3a, Dnmt3aE752A, or Dnmt3aR832A (CAGG-Cre-ER T2 = ER T2-Cre: Dnmt3afl/+, Dnmt3afl/E752A, and Dnmt3afl/R832A) to be compared to the Dnmt3a-null group (ER T2-Cre: Dnmt3afl/-). These mice contain one allele with loxP-flanked Dnmt3a that is deleted by tamoxifen-inducible Cre-mediated recombination and one allele of either wild-type Dnmt3a, Dnmt3aE752A, Dnmt3aR832A, or germline knockout Dnmt3anull. 5-weeks post-tamoxifen (~93% floxed allele recombination), competitive transplantation of 250 phenotypically defined test HSCs against with 2.5x10 5 congenic competitor BM cells was performed.

Dnmt3afl/R832A recipients had higher engraftment (35.6 % +/- 6.1) than Dnmt3afl/+ (28.5% +/- 7.2) and Dnmt3afl/- (10.7% +/- 2.79), while Dnmt3afl/E752A had slightly higherengraftment (12.5% +/- 3) than Dnmt3afl/-. Analysis of the BM 18 weeks post-transplant showed that Dnmt3afl/E752A and Dnmt3afl/R832A HSCs phenocopied the HSC self-renewal potential phenotype of heterozygous Dnmt3a fl/+HSCs (Fig. 1). The absolute count of donor-derived HSCs per mouse after the transplant were: ER T2-Cre control (675.7 +/- 299.3), Dnmt3afl/+ (1870 +/- 961.4), Dnmt3afl/- (3546 +/- 1019), Dnmt3afl/E752A (1130 +/- 362.7), and Dnmt3afl/R832A (1184 +/- 344.5) (mean +/- S.E.M.). While the described clonal expansion of Dnmt3a-null HSCs was observed, HSCs with one copy of full-length Dnmt3a but devoid of its methyltransferase capacity mimicked the heterozygous state rather than the homozygous loss-of-function. This is the first evidence to suggest that DNMT3A potentially regulates HSCs by non-canonical (DNA methylation independent) mechanisms. DNA methylation analysis by WGBS is ongoing to determine if Dnmt3afl/E752A and Dnmt3afl/R832A HSCs show a methylome comparable to Dnmt3a-null HSCs whilst having the functional potential of Dnmt3a-heterozygous HSCs, which will be complemented with other molecular analyses including gene expression. Our study opens new avenues for investigations into the molecular mechanisms of DNMT3A function in HSC biology, which could ultimately benefit clinical practice by identifying new therapeutic approaches for the patients with DNMT3A mutations.

Figure 1
Figure 1.
View largeDownload slide

Disclosures

No relevant conflicts of interest to declare.

© 2021 by The American Society of Hematology
2021
Sign in via your Institution