Azacitidine and Ascorbate Inhibit the Competitive Outgrowth of Human TET2 Mutant HSPCs in a Xenograft Model of Pre-Leukemia

The TET2 gene is frequently mutated in pre-leukemic hematopoietic stem cells in human acute myeloid leukemia (AML) and encodes for an enzyme that catalyzes the conversion of DNA 5-methylcytosine to 5-hydroxymethylcytosine. Recent studies suggest that (i) the product of this reaction can be enhanced using high dose ascorbate, and (ii) formation of the substrate 5-methylcytosine can be blocked with azacitidine.

To understand the mechanisms of TET2 mutation-driven leukemogenesis, we developed two CRISPR/Cas9 approaches to disrupt the TET2 gene in primary human CD34⁺ HSPCs to mimic TET2-mutated pre-leukemia. First, in "Hit & Run," we use Cas9 with two single-guide RNAs (sgRNAs) to disrupt the TET2 gene within exon 3 (average indel frequencies=94.3%). Second, we using homology directed repair (HDR) of Cas9-mediated dsDNA breaks to disrupt the TET2 gene within exon 7 by inserting a GFP expression cassette to generate in vivo traceable cells. Thus, we have developed a tractable and cell-traceable model that recapitulates TET2-mutated pre-leukemia and clonal hematopoiesis.

First, we examined the effects of TET2 disruption on human erythroid differentiation in vitro by culturing bulk CD34⁺ cells for 10 days under conditions that promote erythroid differentiation. Both Hit & Run and HDR (GFP⁺) TET2 disruption decreased CD71⁺CD235⁺ erythroid differentiation compared to control cells. Exposure to high dose ascorbate partially rescued the erythroid defect in TET2-disrupted cells (Hit & Run, n=3 independent experiments, p<0.02). This underscores the importance of TET2 in promoting erythroid differentiation and suggests TET2 mutations can exert a myeloid lineage skewing sensitive to ascorbate.

Next, we investigated the effects of TET2 disruption on hematopoietic colony formation in methylcellulose. Both methods resulted in increased numbers of TET2-disrupted colonies compared to control (Hit & Run, n=4 independent experiments, p<0.0001; HDR, n=3 independent experiments, p<0.0001) and absence of erythroid BFU-E. Interestingly, analysis of indels in Hit & Run colonies showed that serial replating enriched for a 65 base pair deletion that results in a null allele, suggesting that TET2-disrupted cells outcompete normal HSPCs in vitro.

Next, we transplanted control or TET2-disrupted Hit & Run CD34⁺ cells into NSG mice. Primary transplantation at 4 months showed no statistical differences in either engraftment rate (human CD45⁺) or differentiation (T/ B/ Myeloid cells), although the frequency of TET2 indels increased gradually in CD33⁺ cells. Intriguingly, 36 weeks after secondary transplantation, we detected a marked expansion of human myeloid lineage cells (lymphoid=22.1%, myeloid=73.0%, Mann-Whitney U, p=0.0485) and a particular increase in a CMML-like CD33^highCD14⁺CD16^- population. Furthermore, preliminary data from tertiary transplantation (8 weeks after transplantation) indicates persistent myeloid skewing in the bone marrow in some mice and expansion of TET2-mutant cells, suggesting a CMML-like disease.

Finally, we used in vivo competition studies to determine if TET2-disrupted HSPCs are selectively targeted by azacitidine or ascorbate treatment compared to controls. NSG mice were intrafemorally transplanted with a one-to-one ratio of control and TET2-disrupted HSPCs, and 4 months later, these mice were treated with azacitidine (2.5mg/kg/dose, i.p. daily on days 1-5 of a 14-day cycle for 2 cycles) or ascorbate (4g/kg/dose, i.p. twice daily for a month). In PBS control treated mice, the percentage of TET2-disrupted cells increased from 29.3 to 71.6 over 4 weeks. Intriguingly, azacitidine slowed the expansion of TET2-disrupted cells in evaluable mice (delta increase of 42% in PBS vs 5% in azacitidine, p=0.036), but did not eradicate established TET2 pre-leukemia in all evaluable mice. Similarly, high dose ascorbate treatment slowed the rate of expansion to a lesser degree (delta increase of 42% in PBS vs 18.3% in ascorbate, p=0.14).

Our data show that TET2 disruption in primary human HSPCs blocked erythroid differentiation, increased colony formation and replating, and caused myeloid skewing and a CMML-like disease in vivo after an extended period of time. In this model, azacitidine or ascorbate treatment slowed expansion of TET2-mutant human pre-leukemic clones raising the intriguing possibility of preventing CHIP progression to de novo AML.