Credit: IMAGE 100

Two recent studies show that mutations in the metabolic gene isocitrate dehydrogenase 2 (IDH2) can drive the development of leukaemia and sarcoma in vivo. Monoallelic point mutations in IDH1 or IDH2 have been associated with various cancer types, which suggests that these mutations may be drivers of tumorigenesis; however, the study of IDHs has been hindered by a lack of in vivo models.

IDH enzymes normally convert isocitrate to α-ketoglutarate (αKG). However, cancer-associated IDH mutations result in an enzyme that converts αKG to 2-hydroxyglutarate (2HG). 2HG can competitively inhibit the targets of αKG, which have DNA and histone demethylase activity. Thus, 2HG production leads to changes in gene expression that have previously been reported to result in impaired differentiation.

Lu and colleagues investigated IDH2 mutations in sarcomas, which are mesenchymal tumours that show poor differentiation and that frequently contain mutated IDH2. The authors used non-transformed multipotent 10T mouse cells that expressed the mutant form of IDH2, and they observed that the cells failed to differentiate into mesenchymal lineages and that this failure was associated with high levels of 2HG. Moreover, the treatment of cells expressing IDH2-mutant proteins with a DNA methyltransferase inhibitor restored the ability of the cells to functionally differentiate into mature cell types. Subcutaneous injection of 10T cells that expressed mutant IDH2 into mice resulted in a striking increase of sarcoma growth, whereas no sarcomas developed in mice that were injected with cells expressing wild-type IDH2 or vector alone. Furthermore, the tumours were poorly differentiated with no detectable expression of mature cell markers, which mimics the immunohistochemical profile of human sarcomas. The sarcomas also had increased levels of 2HG, as seen in cell culture studies.

these studies confirm that, in vivo, IDH2 mutations cause inhibition of differentiation, altered DNA methylation and hyper-proliferation

In another study, Chen and colleagues described a novel mouse model of mutant IDH2-driven acute myeloid leukaemia (AML), and they used this model to test possible treatments. The authors investigated whether IDH2 mutations would cooperatewith mutations in two other known AML-associated genes: Flt3 and Nras. Mice that carry Flt3 mutations develop chronic myelomonocytic leukaemia, which is a milder form of leukaemia that does not progress to AML, whereas mice that carry Nras mutations develop a myeloproliferative disorder. Haematopoietic stem and progenitor cells (HSPCs) were isolated from Flt3- or Nras-mutant mice and transduced with vectors that contained human mutant IDH2 or control vectors. Mice that were transplanted with HSPCs that contained mutant IDH2 and Nras, or mutant IDH2 and Flt3, developed aggressive AML. These mice had increased proliferation of leukaemic cells and these leukaemic cells expressed only progenitor markers, which suggests that these cells had failed to differentiate, mimicking AML in humans. Similar to the result observed in the Lu study, these cells exhibited high levels of 2HG and altered DNA methylation. In addition, when leukaemic cells that were derived from mice with IDH2-driven AML were treated with a small-molecule inhibitor for a specific IDH2 mutation (AGI-6780), in vitro, AML cells ceased proliferating and differentiated. This demonstrates that mutant IDH2 is required for continued tumour maintenance. The authors also showed that mice with IDH2-driven AML responded to treatment with a bromodomain-containing protein 4 (BRD4) inhibitor in vivo, which is another AML-associated therapeutic target, providing a potential new avenue to target leukaemias that have IDH2 mutations.

Together, these studies confirm that, in vivo, IDH2 mutations cause inhibition of differentiation, altered DNA methylation and hyperproliferation in different tumour contexts. In addition, they confirm speculation that IDH2 is required for tumour maintenance rather than simply for tumour initiation, and they provide models to investigate possible therapeutics for IDH2-driven tumours