The Metabolite R-2-Hydroxyglutarate (R2HG) Collaborates with HoxA9 to Induce Monocytic Leukemia

Mutations in the metabolic enzymes IDH1 and IDH2 are frequently found in several tumors including glioma and acute myeloid leukemia (AML). Mutant IDH produces R-2-hydroxyglutarate (R2HG), which induces histone- and DNA-hypermethylation through inhibition of epigenetic regulators, thus linking metabolism to tumorigenesis. However, it is unknown whether R2HG alone is sufficient to recapitulate the biologic effects of mutant IDH1 in vivo. Recently, we have shown that IDH1mut cooperates with HoxA9 and induces a monocytic leukemia in mice. In order to evaluate the effects of R2HG independently of the mutated IDH1 protein and to determine whether the effects are specific to the R-enantiomer of 2HG, we treated mice transplanted with HoxA9 immortalised bone marrow cells with R2HG, S-2-hydroxyglutarate (S2HG), alpha-ketoglutarate (aKG) and phosphate buffered saline (PBS). The mice in the metabolite cohorts received an intraperitoneal dose of 1 mg per day.

Mice treated with R2HG had higher engraftment levels at 16 and 20 weeks post transplantation than the mice treated with S2HG, αKG and PBS respectively (P<.01). High WBC counts (70±16 /nl) and lower platelet counts than in control mice were observed in the cohort receiving R2HG after 16 to 20 weeks of treatment, while the S2HG, αKG and PBS cohorts had normal blood counts even at 20 weeks (P<.05). Peripheral blood from R2HG treated mice revealed significantly more immature Mac1+Gr1- and less mature Mac1+Gr1+ cells at 12 and 16 weeks after treatment than S2HG, αKG and PBS treated mice (P<.001). In addition, the R2HG treated mice died with a median latency of 137 days post transplantation from monocytic leukemia, while mice treated with S2HG, αKG and PBS died with a median latency of 223, 202 and 184 days respectively (P<.001). Further, in order to assess whether R2HG alone was sufficient as a single hit to induce myeloproliferation, normal C57BL/6 mice (without HoxA9) were treated with R2HG, S2HG and PBS for eight months. No differences were observed for survival, blood counts, immunophenotype and frequencies of progenitor cells (lin-ckit+sca1+, CMP, GMP and MEP) between treatment groups and control. This data shows that the metabolite R2HG like the IDH1 mutant protein cooperates with HoxA9 to induce monocytic leukemia.

We next compared mice receiving transplants of HoxA9+IDH1mut cells with mice receiving HoxA9 cells that were then treated with R2HG. Both cohorts developed monocytic leukemia, albeit with different kinetics. The Hoxa9+IDH1mut mice died with a median latency of 83 days while the R2HG cohort died with a median latency of 137 days post transplantation (P<0.001). Also, while the former cohort developed severe leukocytosis, anemia and thrombocytopenia at 12 weeks, the R2HG treated mice had high WBC counts and lower platelet counts than control mice at 16 to 20 weeks after treatment. The faster disease kinetics in IDH1mut mice could be attributed to a significantly lower proportion of cells in G0/G1 and higher proportion of cells in S phase when compared to cells from mice treated with R2HG at 9 weeks after transplantation (P<.001). This resulted from a marked downregulation of cyclin-dependent kinase inhibitors (Cdkn) 1A (p21), 1B (p27), 2A (p16), and 2B (p15) in HoxA9+IDH1mut cells as compared to HoxA9 cells treated with R2HG or PBS.

In order to rule out an influence of differential R2HG levels on the disease progression between the two cohorts, levels of R2HG were quantified. IDH1mut expressing and R2HG treated bone marrow cells from mice had similarly high ratios of R2HG/S2HG. The quantified R2HG levels were also comparable to that of primary AML patient cells harbouring mutated IDH1. In unsupervised hierarchical clustering mutated cells clustered together with R2HG and separated from PBS treated mice. 69 of the top 100 enriched Gene Ontology gene sets of HoxA9+IDH1mut were also found in HoxA9+R2HG, suggesting largely redundant but also non-overlapping functions of the mutant IDH1 protein and the oncometabolite R2HG.

In summary, we show that R2HG, similar to the mutant IDH1 protein, promotes leukemogenesis in cooperation with HoxA9, although with delayed kinetics. Our data proves that R2HG acts as an oncometabolite in vivo in a murine model of leukemogenesis.