Combined Effect Of Dnmt3a Loss-Of-Function and Idh2 neomorphic mutation Promotes Hematopoietic Malignancy

In recent studies on acute myeloid leukemia (AML), genes involving DNA methylation (DNMT3A, IDH1/2, TET2) were identified as frequently mutated. Around 17% of AML patients were found to have a mutation in isocitrate dehydrogenase-1 or -2 (IDH1/2). In these patients, 2-hydroxylglutarate (2-HG), which is generated from a neomorphic activity of mutant IDH1/2, accumulates and inhibits DNA hydroxylmethylation mediated by TET family proteins. Interestingly, mutations in the de novo DNA methyltransferase 3A (DNMT3A) and IDH1/2 mutations co–occur in a statistically significant portion of AML patients. In current known mouse models, neither Dnmt3a knockout (KO) and IDH1/2 mutation alone initiates overt hematopoietic malignancy, leading to our hypothesis that Dnmt3a and IDH1/2 mutations may act synergistically to initiate hematopoietic malignancy.

To create a Dnmt3a knockout IDH1/2 mutant double mutant mouse model, we have employed a transplantation approach to create Idh2^R140Q (one of the abundant IDH2 mutation in AML patients) overexpressing stem cells on a Dnmt3a KO (Dnmt3a^–/–) and wild-type (WT) background by retroviral transduction. Idh2^WT overexpression was also used as a control group. With a latency of 180 days after transplantation, a hematopoietic disease with a median survival of 197 days was observed in the Dnmt3a^–/––Idh2^R140Q group. Anemia, thrombocytopenia and monocytosis were observed in morbid Dnmt3a^–/––Idh2^R140Q mice. The pathological examination of Dnmt3a^–/––Idh2^R140Q mice showed myelodysplasia in one or more lineages, an accumulation of less differentiated myeloid progenitors in the bone marrow and pronounced extramedullary hematopoiesis in the spleen. Moreover, approximately 20% of Dnmt3a^–/––Idh2^R140Q mice developed AML. Together, these features led to the diagnosis of MDS/MPN (Myelodysplastic Syndrome / Myeloproliferative Neoplasms) with high transformation rate to AML. In comparison, WT–Idh2^R140Q mice also developed less severe MDS/MPN characterized by myeloid differentiation bias and extramedullary hematopoiesis without lethality in one year after bone marrow transplantation. The MDS/MPN of Dnmt3a^–/––Idh2^R140Q developed into MPD after secondary transplantation with Lin- cKit+ progenitors, while the same progenitors from the control genotypes did not cause hematopoietic diseases in secondary transplantation .

The profiling of 2-HG with gas chromatography mass spectrometry (GC-MS) in the serum of morbid mice transplanted with Dnmt3a^–/––Idh2^R140Q showed an 80-fold increase compared to normal mouse serum, while the 2-HG content in mice transplanted with WT–Idh2^R140Q cells was 10-fold higher than that of normal mouse serum. This suggests that Dnmt3a loss-of-function can promote the synthesis of 2-HG by the Idh2^R140Q mutation. The metabolomic profiling on cKit+ bone marrow cells identifies 43 metabolites differentially present in malignant Dnmt3a^–/––Idh2^R140Q cells compared with groups of other genotypes. Furthermore, the unsupervised cluster analysis shows Dnmt3a^–/––Idh2^R140Q cells have a distinct metabolome profile compared with cells of other genotypes. This suggests a synergistic effect on metabolome between the two genetic backgrounds. The essential amino acid and glycolysis pathway metabolites– are among the most enriched differential present metabolites. In addition, the glutamine anaplerosis pathway is highly upregulated in the Dnmt3a^–/––Idh2^R140Q group indicated by the increase of a-ketoglutarate and glutamate. In contrast, the WT – Idh2^R140Q and Dnmt3a KO groups have no alteration in glutamine anaplerosis pathway metabolites, indicating a synergy on 2-HG synthesis between Dnmt3a knockout and Idh2^R140Q. Moreover, an excess of glutamine during in vitro culture significantly promotes the colony forming ability Dnmt3a^–/––Idh2^R140Q HSPCs, while there’s no effect in Dnmt3 Idh2^WT and WT–Idh2^R140Q HSPCs.

In summary, our research shows for the first time, Dnmt3a loss-of-function promotes 2-HG synthesis with mutant Idh2 and strongly aggravates the phenotype induced by Idh2 mutation. Our research also shows the synergistic effect on the metabolome of two genetic backgrounds. These data likely explain the high frequency of co-mutations between DNMT3A and IDH1/2 that promotes aggressive AML in patients.