IDH1 and IDH2 mutations in myeloid neoplasms – Novel paradigms and clinical implications

In the World Health Organization (WHO) classification of tumors of hematopoietic and lymphoid tissues1 myeloid neoplasms include myeloproliferative neoplasms (MPN), myelodysplastic syndromes (MDS), myelodysplastic/myeloproliferative neoplasms (MDS/MPN), and acute myeloid leukemia (AML). In the last few years there have been major advances in our understanding of the molecular bases of these disorders, and molecular genetic data are increasingly being used for diagnosis, risk assessment and definition of treatment strategies.2,3 These data now include information on mutations in the IDH1 and IDH2 genes.

IDH1 and IDH2 encode the enzymes isocitrate dehydrogenase 1 and 2, respectively. The essential information about these genes and their products is reported in Table 1.

Somatic mutations of IDH1 and IDH2 in malignant gliomas

In 2008, though a genome-wide analysis Parsons et al.4 identified somatic mutations at codon 132 of IDH1 in approximately 12% of patients with glioblastoma multiforme, the most common and fatal type of brain cancer. In a subsequent study, these authors detected somatic mutations that affected amino acid 132 of IDH1 in more than 70% of gliomas.5 In most cases, arginine 132 was mutated to histidine (R132H). Some tumors without mutations in IDH1 had mutations affecting the analogous amino acid (R172) of the IDH2 gene, strongly indicating a role of mutations in the NADP-dependent isocitrate dehydrogenase genes in the pathogenesis of these malignancies. Overall, brain tumors with IDH1 or IDH2 mutations represented a distinctive subgroup of low-grade and secondary gliomas with a better outcome compared to that of tumors with wild-type IDH genes.

A causal relationship between acquired error in cellular metabolism and malignant transformation

Since a single copy of the gene – IDH1 or IDH2 – is mutated in human gliomas, Dang and co-workers6 hypothesized that the mutations do not result in a simple loss of function. They did an elegant study to determine the impact of the IDH1 (R132H) mutation on cellular metabolism, and showed that it resulted in the production of 2-hydroxyglutarate.6 Since overproduction of this metabolite is associated with a high risk of brain tumors in patients with inborn metabolic errors, the authors concluded that the accumulation of excess 2-hydroxyglutarate in vivo contributes to the formation and malignant progression of gliomas, establishing a link between abnormal metabolism and malignancy.7 The altered metabolic pathway associated with mutant IDH1 has also been shown to contribute to tumor growth by activating hypoxia-inducible factor-1α.8 These observations have important clinical implications, as patients with initial or low-grade forms of glioma may benefit from the therapeutic inhibition of 2-hydroxyglutarate production.9

Table 1.The genes IDH1 and IDH2 and their products.*

Table 2.Recent studies on the clinical significance of somatic mutations of IDH1 and IDH2 in acute myeloid leukemia.

Somatic mutations of IDH1 and IDH2 in acute myeloid leukemia

In a study of whole genome sequencing, Mardis et al.10 analyzed the leukemic genome in a patient with de novo cytogenetically normal AML with minimal maturation. They identified 12 somatic mutations within the coding sequences of genes, including IDH1 (R132C). Interestingly, somatic mutations at codon 132 of IDH1 were found in 15 of 187 additional AML genomes: all these mutations were heterozygous and strongly associated with normal cytogenetic status.

In the last few months, several studies have been published on somatic mutations of IDH1 and IDH2 in AML, and their main findings are summarized in Table 2.11–20

Based on data reported in Table 2, it can be concluded that:

• somatic mutations of IDH1 and IDH2 are found mainly, although not exclusively, in cytogenetically normal AML, in which they are – with a few exceptions - mutually exclusive. On average, about 30% of patients with cytogenetically normal AML carry a mutant IDH1 or IDH2 gene, and in this category IDH2 mutations are more common that IDH1 mutations;
• the vast majority of somatic mutations of IDH1 and IDH2 involve residues R132 of IDH1, and R140 or R172 of IDH2. This allows a quick and sensitive screening for IDH1 and IDH2 mutations,21 providing clinicians with an important diagnostic instrument;
• the prognostic significance of somatic mutations of IDH1 and IDH2 in AML is currently under investigation, but the available evidence indicates that they may be associated with an intermediate to high genetic risk.

Metabolic abnormalities in acute myeloid leukemia associated with somatic mutations of IDH1 and IDH2 and potential novel therapeutic perspectives

Two recent studies showed that AML cells bearing heterozygous IDH1 or IDH2 mutations accumulate 2-hydroxyglutarate.22,23 This finding suggests that 2-hydroxyglutarate is an onco-metabolite that plays a role not only in the pathogenesis of gliomas but also in leukemic transformation, and might have therapeutic implications for the treatment of AML. In fact, blocking the accumulation of 2-hydroxyglutarate through the inhibition of mutant IDH enzymes could represent a therapeutic goal: for detailed information about this issue, the reader is referred to a very comprehensive review article by Dang et al.9 A few small molecules capable of inhibiting IDH enzymes have already entered preclinical studies or clinical development.

Figure 1.Schematic representation of the molecular bases of myeloid neoplasms and hypothesized role of IDH1/IDH2 mutations.

Somatic mutations of IDH1 and IDH2 in myeloproliferative neoplasms, myelodysplastic syndromes and secondary acute myeloid leukemia

Green and Beer24 searched for mutations in IDH1 and IDH2 in patients with AML that had evolved from JAK2-mutated MPN, and found somatic mutations in five of 16 patients: three involved IDH1 (R132) and two IDH2 (R140). These mutations were not present in 180 unselected patients with polycythemia vera or essential thrombocythemia in chronic phase. Pardanani et al.25 detected IDH1 (R132) or IDH2 (R140) mutations in seven of 34 patients with blast-phase MPN, and in three of 166 patients with chronic-phase MPN (these latter patients had primary myelofibrosis). In a study reported in this issue of Haematologica, Andrulis et al.26 investigated bone marrow samples from 160 patients with chronic MPN using an antibody highly specific for the IDH1 (R132H) mutation, and found three positive patients. The mutation could be confirmed by DNA sequencing in only one of these three individuals, which may have been because of a low mutant allele burden. This underscores the need for sensitive assays to detect IDH1 or IDH2 mutations in patients with low mutation loads.

Also in this issue of Haematologica, Thol et al.27 report findings of a study on IDH1 or IDH2 mutations in patients with MDS or AML arising from MDS. Among 193 MDS patients, seven (3.6%) had a heterozygous mutation in IDH1 codon 132, while no IDH2 mutation was detected in any subject. Patients carrying mutated IDH1 had a high rate of leukemic transformation and a poor event-free and overall survival. Among 53 AML patients with a previous history of MDS, four patients had mutations in codon 132 of IDH1 and four had mutations in codon 140 of IDH2: thus, 15% of cases of AML arising from MDS had IDH1 or IDH2 mutations. Kosmider et al.28 recently reported findings of a study on IDH1 and IDH2 mutations in early and accelerated phases of MDS and MDS/MPN. The frequencies of these mutations were 5% in MDS, 8.8% in MDS/MPN and in 9.7% in secondary AML.

Thus, the available evidence suggests that somatic mutations in IDH1 and IDH2 may represent a mechanism of progression to AML in MPN and MDS as schematically represented in Figure 1, although this awaits prospective validation. Other recently identified mechanisms of disease progression in these disorders include somatic mutations of ASXL1,29 inactivating mutations of the histone methyltransferase gene EZH2,30,31 and deletion of the IKZF1 gene.32

Conclusions

The remarkable achievements described in this perspective article have been made in the last 2–3 years, which indicates how fast our understanding of the molecular bases of myeloid neoplasms is evolving. It is to be hoped that these achievements will not only make a molecular classification of myeloid neoplasms feasible, but will also allow novel targeted therapies to be developed.

Footnotes

• Mario Cazzola is Professor of Hematology at the University of Pavia Medical School, Pavia, Italy.
• The author wishes to thank Konstanze Döhner and Peter Paschka for their critical reading of this paper.
• Financial and other disclosures provided by the author using the ICMJE (www.icmje.org) Uniform Format for Disclosure of Competing Interests are available with the full text of this paper at www.haematologica.org.
• ( Related Original Articles on pages 1668 and 1754 and related Letter on page 1797)