Abstract
The majority of palindromic CpG dinucleotides in mammalian DNA are symmetrically methylated on cytosine (5mC). Erythroid differentiation is associated with replication-dependent genome-wide loss of ~30-50% of all 5mC (1). One potential mechanism of demethylation involves Tet dioxygneases, which oxidize 5 methylcytosine (5mC) to 5-hydroxy-methylcytosine (5hmC), thereby interfering with maintenance methylation by DNMT1 (2). Tet2 and Tet3 are the two major Tet proteins expressed in hematopoietic cells. Tet2 is one of the most commonly mutated genes in hematological malignancies. Germline deletion of Tet2 results in increased hematopoietic stem cell self-renewal, erythroid and myeloid hyperplasia and myeloid and lymphoid malignancies (3, 4). Germline deletion of Tet3 is embryonic lethal (5). Here we investigated the potential role of Tet2 and Tet3 in the global loss of DNA methylation during erythroid differentiation, using mouse genetic models.
We bred Tet2fl/fl and Tet3fl/fl mice onto the Rosa26-CreERT2 background, enabling acute deletion in vivo of either Tet2, Tet3 or both genes in adult and fetal mice. Within 4 to 10 days of deletion, we noted a rapid decline in total bone marrow cellularity in all genotypes; bone-marrow erythroblasts decreased to 15% of control value in Tet2-deleted (Tet2del) mice (p<0.0001), and to 43% (p=0.05) and 20% (p<0.0001) of controls in the Tet3 del and Tet2/Tet3 double deleted mice (DKO), respectively. These findings were associated with decreased reticulocyte counts in peripheral blood, and a decline in hematocrit, in Tet2del (36% vs. 46% in controls, p<0.004) and the DKO (31%, p=0.03). Bone-marrow CFU-e progenitors, but not BFU-e progenitors, also decreased in number. These losses partially improved over the next 2 to 4 weeks in mice deleted for either Tet2 or Tet3, but not in the DKO mice, in which anemia persisted. Megakaryocyte CFU-Mk progenitors increase in the DKO but are unaffected in the single deletions. Lymphoid lineage cells (CD4+, CD8+ and CD19+) showed similar transient losses, but the numbers of granylocytic and monocytic progenitors in the bone marrow was not affected. Similarly, fetal liver cell number decreased in all genotypes within 4 days of deletion, compared with controls, in all genotypes, and most severely in the DKO.
We used three approaches to determine the potential effect of Tet proteins on erythroid DNA demethylation: Mass spectrometry, to measure global levels of 5mC and 5hmC in fetal liver erythroblasts; 5hMe-DIP-seq (5hyrdoxymethyl-DNA immune precipitation- sequencing), to determine the genomic distribution of hydroxymethylation in erythroblasts; and pyrosequencing for DNA methylation/hydroxymethylation at specific genomic loci including repetitive LINE_1 elements. We found that 5hmC is highly enriched in previously identified erythroid enhancers. Tet2 accelerates, but is not essential, for DNA demethylation of specific erythroid enhancers. Neither Tet2 nor Tet3 are required for global DNA demethylation.
Taken together, we conclude that Tet2 and Tet3 are required during erythropoiesis, where they play partially redundant roles, possibly contributing to the changing chromatin environment during activation of erythroid terminal differentiation. Their acute deletion leads to transient loss of erythroblast viability and to anemia, but this is followed by a process of adaptation and chronically compensated erythropoiesis. These results raise the possibility that the adaptation required for the rapid partial recovery in the erythroid and lymphoid lineages following acute Tet2 loss may contribute, over the long term, to the development of myeloid and lymphoid malignancies.
1. J. R. Shearstone et al., Global DNA demethylation during mouse erythropoiesis in vivo. Science334, 799-802 (2011).
2. M. Ko et al., Impaired hydroxylation of 5-methylcytosine in myeloid cancers with mutant TET2. Nature468, 839-843 (2010).
3. M. Ko et al., TET proteins and 5-methylcytosine oxidation in hematological cancers. Immunol Rev263, 6-21 (2015).
4. J. An et al., Acute loss of TET function results in aggressive myeloid cancer in mice. Nature communications6, 10071 (2015).
5. T. P. Gu et al., The role of Tet3 DNA dioxygenase in epigenetic reprogramming by oocytes. Nature477, 606-610 (2011).
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.