DNA methylation is one of the major epigenetic modifications in the vertebrate genome and is important for development, stem cell differentiation, and malignant transformation. DNA methylation is catalyzed by the DNA methyltransferase enzymes Dnmt1, Dnmt3a, and Dnmt3b. We have recently shown that Dnmt3a is essential for hematopoietic stem cell (HSC) differentiation. Ablation of Dnmt3a in hematopoietic cells (Mx1-CRE; Dnmt3a-KO) resulted in HSCs that could not sustain peripheral blood generation after serial transplantation, while phenotypically defined HSCs accumulated in the bone marrow. Recurrent somatic mutations in DNTM3A have been discovered in patients with a wide range of hematopoietic malignancies (AML, MDS, MPN, CML, T-ALL, T-cell lymphoma) suggesting a critical role for de novo DNA methylation in normal and leukemic hematopoiesis. As Dnmt3b is also highly expressed in HSCs and congenital mutations in DNMT3B can cause ICF (immunodeficiency, centromeric instability, and facial anomalies) syndrome, in this study we used a mouse model to investigate if Dnmt3b had distinct roles in HSCs.

We conditionally inactivated Dnmt3b in HSCs using the Mx1-CRE system (Dnmt3b-KO) and performed serial competitive transplantation. Loss of Dnmt3b had minimal functional consequences for adult HSC function even after three rounds of transplantation. However, combinatorial deletion of both Dnmt3a and Dnmt3b (Dnmt3ab-dKO) exacerbated the differentiation defect seen in Dnmt3a-KO HSCs, leading to a dramatic accumulation of mutant HSCs in the bone marrow (>50-fold), suggesting a synergistic effect resulting from simultaneous ablation of both de novo DNA methyltransferases. The accumulation of Dnmt3ab-dKO HSCs cannot be attributed to altered proliferation or apoptosis, but is due to an imbalance between self-renewal and differentiation. RNA-SEQ of the mutant HSCs revealed loss of transcriptional integrity in Dnmt3ab-dKO HSCs including increased expression of repetitive elements, inappropriate mRNA splicing, and over-expression of HSC-specific genes.

To examine the impact of loss of Dnmt3a and -3b on DNA methylation in HSCs, we performed Whole Genome Bisulfite Sequencing (WGBS). Ablation of both enzymes resulted in loss of DNA methylation that was much more extensive than that seen in the absence of Dnmt3a alone, while loss of Dnmt3b alone showed only minimal changes in DNA methylation compared to control HSCs. One puzzling finding was the observation that a subset of promoter CpG islands (CGIs) actually gained DNA methylation in Dnmt3a-KO HSCs. This CGI hypermethylation is a cancer methylome phenotype and was specific to Dnmt3a-KO HSCs (Figure 1A). The HSC transplant experiments suggest that Dnmt3a can compensate for Dnmt3b in HSCs, but Dnmt3b cannot reciprocate in the reverse situation. An explanation for increases in DNA methylation is that in the absence of Dnmt3a, abnormal function of Dnmt3b may lead to aberrant CGI hypermethylation as the hypermethylation was lost when both enzymes were conditionally inactivated. To confirm the mechanism, post-transplant Dnmt3ab-dKO HSCs were transduced with a retroviral vector encoding ectopic expression of Dnmt3b (MIG-Dnmt3b) or a control empty vector (MIG) and assessed for DNA methylation by bisulfite PCR. Using the promoter CGI of Praf2 as an example, enforced expression of Dnmt3b in Dnmt3ab-dKO HSCs resulted in increased DNA methylation at this loci compared to Dnmt3ab-dKO HSCs transduced with a control empty vector (MIG), control HSCs transduced with either MIG or MIG-Dnmt3b and untransduced HSCs (Figure 1B). It is possible that when Dnmt3b tries to compensate for Dnmt3a, the locus-specificity for targets is reduced, leading to aberrant DNA methylation patterns. Promoter CGI hypermethylation is a cancer phenotype observed in a wide range of tumors, including hematopoietic neoplasms driven by mutation in DNMT3A. Targeting DNMT3B in DNMT3A-mutation hematopoietic pathologies may be a therapeutic option for restoring normal DNA methylation and gene expression patterns.
Figure 1

Praf2 promoter DNA methylation. Open circle = unmethylated CpG, closed circle = methylated CpG. (A) DNA methylation patterns in control (Ctl), Dnmt3a-KO (3aKO), Dnmt3b-KO (3bKO) and Dnmt3ab-dKO HSCs (dKO). (B) Patterns in control and Dnmt3ab-dKO HSCs transduced with empty vector (MIG) or ectopic Dnmt3b, compared to untransduced HSCs.

Figure 1

Praf2 promoter DNA methylation. Open circle = unmethylated CpG, closed circle = methylated CpG. (A) DNA methylation patterns in control (Ctl), Dnmt3a-KO (3aKO), Dnmt3b-KO (3bKO) and Dnmt3ab-dKO HSCs (dKO). (B) Patterns in control and Dnmt3ab-dKO HSCs transduced with empty vector (MIG) or ectopic Dnmt3b, compared to untransduced HSCs.

Close modal
Disclosures:

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.

Sign in via your Institution