Cancer-Associated IDH Mutations: Biomarker and Therapeutic Opportunities

KE Yen, MA Bittinger, SM Su and VR Fantin

Molecular Oncology, Agios Pharmaceuticals, Cambridge, MA, USA

The discovery of somatic mutations in the isocitrate far reaching implications of metabolic alterations in dehydrogenase (IDH) enzymes through a genome-wide malignant cells, and their potential contribution to mutational analysis in glioblastoma represents a milestone initiation and development of cancer. In addition to event in cancer biology. The nature of the heterozygous, their intermediary role in energy and biomass producing point mutations mapping to arginine residues involved in reactions, certain privileged metabolites have key roles the substrate binding inspired several research teams to as allosteric regulators and cofactors of enzymes investigate their impact on the biochemical activity of involved in fundamental processes including chromatin these enzymes. Soon, it became clear that the mutations remodeling, mitochondrial respiration, angiogenesis and identified impaired the ability of IDH1 and IDH2 to migration. Furthermore, the same molecular mechan- catalyze the conversion of isocitrate to a-ketoglutarate isms leading to uncoupling of cell proliferation and (aKG), whereas conferring a gain of a novel enzymatic survival from growth factor signaling and other extra- activity leading to the reduction of aKG to the metabolite cellular cues, also contribute to metabolic rewiring. D2-hydroxyglutarate (D-2HG). Across glioma as well as It has long been recognized that the aberrant activity several hematologic malignancies, mutations in IDH1 and of proto-oncogenes and tumor suppressors affect the IDH2 have shown prognostic value. Several hypotheses expression and activity of nutrient transporters and implicating the elevated levels of D-2HG and tumorigenesis, metabolic enzymes (Dang and Semenza, 1999; Kroemer and the therapeutic potential of targeting mutant IDH and Pouyssegur, 2008; Dang et al., 2009; Vander Heiden enzymes will be discussed. et al., 2009). However, the first examples of mutations in Oncogene (2010) 29, 6409–6417; doi:10.1038/onc.2010.444; genes coding for mitochondrial enzymes involved in published online 25 October 2010 metabolism have only come to light in the past few years. Those include mutations in fumarate hydratase Keywords: IDH; mutations; cancer (FH) and succinate dehydrogenase (SDH) in hereditary leiomyomatosis and renal cell carcinoma as well as paragangliomas (Baysal et al., 2000; Tomlinson et al., Introduction 2002; Brandon et al., 2006; King et al., 2006; Kaelin, 2009). A mutation in the isocitrate dehydrogenase 1 Malignant transformation is most often a multistep (IDH1) gene was first discovered during a sequencing process in which a series of genetic, epigenetic and post- effort across colorectal cancers (Sjo¨blom et al., 2006). transcriptional events ultimately confer emerging neo- More recently, somatic mutations in IDH1 were plastic cells with a growth and survival advantage identified through a genome wide mutation analysis in (Vogelstein and Kinzler, 1993; Hanahan and Weinberg, glioblastoma (Parsons et al., 2008). Following this 2000; Jones and Baylin, 2002). It is becoming increas- pivotal study, high throughput deep sequencing revealed ingly clear that metabolic adaptation is an integral the presence of mutations in either IDH1 or its aspect of this process (Gillies and Gatenby, 2007; mitochondrial counterpart IDH2 in 470% of grade Kroemer and Pouyssegur, 2008). Fast proliferating II–III gliomas and secondary glioblastomas (GBMs; cancer cells meet their elevated energetic and anabolic Yan et al., 2009). Since these initial reports, mutations in B demands by preferentially ‘feasting’ on certain carbohy- IDH1 and IDH2 have been identified in 16–17% of drates, lipids and amino acids and resorting to unique patients with acute myeloid leukemia, and spotted in a metabolic strategies. In fact, the elevated glucose avidity variety of other malignancies at lower frequencies (Abbas of hematologic malignancies and solid tumors has been et al., 2010; Paschka et al., 2010). Unlike the loss-of- widely exploited for diagnostic purposes. But only more function mutations in FH and SDH genes, the IDH1 and recently researchers have begun to appreciate the IDH2 mutations described to date confer a neoactivity to their enzyme products (Dang et al., 2009; Mardis et al., 2009; Gross et al., 2010; Ward et al., 2010). This review Correspondence: Dr VR Fantin, Molecular Oncology, Agios Pharma- will attempt to summarize our current understanding of ceuticals, 2nd Floor, 38 Sidney Street, Cambridge, MA 02139, USA. E-mail: [email protected] the recently identified mutations in IDH1 and IDH2 and Received 21 June 2010; revised 18 August 2010; accepted 19 August provide several potential molecular mechanisms linking 2010; published online 25 October 2010 them to malignant transformation. Cancer-associated IDH mutations KE Yen et al 6410 Biochemistry of IDHs peroxisomal localization sequence and is localized in both peroxisomes and cytosol (Geisbrecht and Gould, The family of IDHs consists of three members: IDH1, 1999; Yoshihara et al., 2001). IDH2 and IDH3 are IDH2 and IDH3. All three enzymes catalyze the mitochondrial isoforms of IDH (Figure 1a). Unlike the þ oxidative decarboxylation of isocitrate to produce CO2 NADP -dependent enzymes, the IDH3 holoenzyme is a and a-ketoglutarate (aKG). These reactions require the heterotetrameric protein consisting of two subunits of presence of a divalent ion (Mg2 þ or Mn2 þ ) and the IDH3A and one subunit each of IDH3B and IDH3G. cofactors NADP þ (IDH1 and IDH2) or NAD þ (IDH3) To date, there have been no reports of cancer-associated as electron acceptors, generating nicotinamide adenine mutations in any of the IDH3 subunits. In yeast, dinucleotide phosphate (NADPH) or NADH respec- peroxisomal IDH is important for maintaining NADPH tively. The NADP þ -dependent family members, IDH1 levels that are essential for the metabolism of branched and IDH2, are homodimeric enzymes that share chain fatty acids (Henke et al., 1998; van Roermund significant structural similarity. The human proteins et al., 1998). Although NAD þ -dependent IDH3 is are 69% identical and crystallography studies have believed to be the principal IDH enzyme in the tri- shown that the overall quaternary structure is nearly carboxylic acid cycle, the exact roles of IDH1 and IDH2 identical between the two proteins (Ceccarelli et al., in cellular metabolism have been less well defined. 2002; Xu et al., 2004). One fundamental difference The oxidation of isocitrate to aKG by IDH is widely between IDH1 and IDH2 is the subcellular localization viewed as one of the irreversible enzymatic reactions in of the two proteins. IDH1 contains a C-terminal the central metabolism. This conversion involves a large

Figure 1 Metabolic reactions catalyzed by the IDH family. (a) IDH1 catalyzes the NADP þ -dependent decarboxylation of isocitrate to aKG in the peroxisome and cytoplasm, whereas IDH2 mediates the same reaction in the mitochondria. Also, within the mitochondrial compartment, heterotetrameric IDH3 catalyzes the conversion of isocitrate to aKG in a NADP þ -dependent manner. (b) Mutations in IDH1 and IDH2 impart these enzymes the ability to catalyze the reduction of aKG to 2HG. On the basis of the chemical similarity between these two metabolites, it is possible that the activity of aKG-dependent enzymes is affected by the elevated level of 2HG in tumors harboring IDH1 and IDH2 mutations. In this model, these events may trigger alterations in angiogenesis, epigenetic state, extracellular matrix dynamics, and in turn may affect proliferation, survival and invasive properties of these tumors. Cit, citrate; Fum, fumarate; Mal, malate; OAA, oxaloacetate; Suc, succinate.

Oncogene Cancer-associated IDH mutations KE Yen et al 6411 free energy change associated with the production of CO2. Recently, we demonstrated that mutant IDH1 and On binding of isocitrate and NADP þ ,IDH1shiftsfrom IDH2 are not catalytically inactive enzymes, but rather an open conformation to a catalytically active closed possess novel enzymatic activities, reconciling the conformation (Xu et al., 2004). The first step in the proposed gain-of-function model based on the hetero- enzymatic reaction is the oxidation of isocitrate to zygous nature of the point mutations (Dang et al., 2009). oxalosuccinate with the concomitant reduction of NADP þ Untargeted metabolic profiling by liquid chromatogra- to NADPH. Enzyme-bound oxalosuccinate then under- phy-mass spectrometry identified one metabolite, goes decarboxylation to form aKG that is then released as 2-hydroxyglutarate (2HG), that was elevated over the enzyme transitions back to the open conformation in 100-fold in U87-MG cells expressing mutant IDH1 preparation for another round of catalysis. relative to the same cells expressing wild-type IDH1. The chemical similarity of 2-HG to aKG, in combina- tion with observations from isotope labeling experi- Neoactivity of IDH1/IDH2 mutant enzymes ments in whole cells showing carbon flow from aKG to 2HG, led our group to immediately hypothesize that As mentioned earlier, somatic mutations in IDH1 and mutant IDH1 could catalyze a novel enzymatic reaction IDH2 were initially identified in low-grade gliomas and in which the reduction of aKG to 2HG is coupled to secondary GBMs. Those reports established the pre- the oxidation of NADPH to NADP þ . Indeed, our sence of a heterozygous mutation affecting Arg132 in studies established that purified mutant protein effi- IDH1as well as the equivalent residue, Arg172, in IDH2, ciently catalyzed the proposed reduction of aKG albeit at lower frequency. Since Arg132 together with to 2HG, while being unable to synthesize isocitrate Arg100 and Arg109 are part of a catalytic arginine triad (Dang et al., 2009). There are two possible enantiomers involved in isocitrate binding, it was immediately of 2HG because of the presence of a chiral center in hypothesized that these mutations would affect IDH1 this molecule, and mutant IDH1 produces exclusively enzymatic activity (Yan et al., 2009; Figure 2). Initial R(–)-2HG (also commonly called D-2HG) as established biochemical studies established that mutant IDH1/ by chemical derivatization of the product. Supporting IDH2 were unable to efficiently catalyze the oxidative X-ray crystallography studies with IDH1R132H decarboxylation of isocitrate, consistent with the muta- showed that within the active site, aKG and NADPH tions resulting in a loss of enzymatic activity. A are positioned as expected for the reduction of the subsequent report suggested that a single mutant subunit substrate to the expected enantiomer. Overall, the within the enzyme dimer could exert a dominant negative enzymatic characterization of mutant enzyme, suggests effect (Zhao et al., 2009). In this case, purified hetero- that the enzymatic gain-of-function is likely driven by dimers of wild-type and mutant IDH1 were unable to both the loss in affinity for isocitrate as well as efficiently catalyze the conversion of NADP þ to NADPH significant increases in affinity for aKG and NADPH when isocitrate was provided as a substrate. (Dang et al., 2009).

Figure 2 Localization of mutated amino acid residues in IDH1. Active site of wild-type IDH1, showing in green amino acid residues Arg 132, Gly97, Val71 and Gly123 reported to be mutated in cancer. Amino acids Arg 100 and Arg 109 (in blue) together with Arg 132 (corresponding to Arg172 on IDH2) form the catalytic triad that coordinates isocitrate (red). The cofactor NADP þ is depicted in cyan. In AML, the frequently mutated Arg140 on IDH2 corresponds to Arg100 on IDH1.

Oncogene Cancer-associated IDH mutations KE Yen et al 6412 Supporting the in vitro biochemical data and the of tumor development suggesting that the generation of observations in engineered cell lines, D-2HG levels were D-2HG may somehow function as an oncometabolite dramatically elevated in malignant gliomas harboring and promote formation and progression of tumors IDH1 and IDH2 mutations. In a panel of human (Watanabe et al., 2009; Pardanani et al., 2010b). secondary GBMs samples, tumors with IDH1 mutations contained up to B100-fold higher levels of D-2HG than tumors harboring wild-type IDH1 (Dang et al., 2009). I. Cancers of the central nervous system Two subsequent studies in acute myeloid leukemia As introduced in previous sections, in low-grade gliomas (AML) also found elevated levels of D-2HG in cells and secondary GBMs, heterozygous IDH1 and IDH2 with mutant IDH1 and IDH2 as compared with cells mutations occur at a single arginine residue, Arg132 of with wild-type alleles (Gross et al., 2010; Ward et al., IDH1, and the analogous Arg172 residue of IDH2. 2010). To date IDH1R132, IDH2172 and IDH2R140 Mutations in IDH1 and IDH2 are mutually exclusive, mutant enzymes of IDH1 and IDH2 have shown the suggesting that they independently confer a growth ability to synthesize D-2HG (Dang et al., 2009; Gross advantage to mutant cells (Hartmann et al., 2009). et al., 2010; Ward et al., 2010). The fact that IDH1 and Proliferation of poorly differentiated gliomas with IDH2 mutant enzymes possess the ability to catalyze the mixed astrocytic and oligodendroglial features is key partial reverse reaction, suggests that this gain-of- to the progression of the disease and their pathogenesis function activity is the relevant consequence of IDH may be related to a block in differentiation rather than mutations in cancer. merely just an increase in cell proliferation (Ward et al., Loss-of-heterozygosity at the IDH locus is very rarely 2010). This observation raises the possibility that D-2HG observed in tumors with IDH1 mutations (Zhao et al., may interfere with processes involved in cellular 2009). It is likely that wild-type enzymatic activity is differentiation, and ongoing studies will shed light on essential for tumor initiation or progression, possibly by this question. Patients that harbor these mutations tend potentiating 2HG production by forming heterodimers to be of younger age and have a better clinical prognosis. with mutant subunits. Although heterodimers of mutant Understanding the link between IDH1/IDH2 muta- and wild-type IDH subunits have been purified and tions and other common genetic events in gliomas may initially characterized (Zhao et al., 2009), the enzymatic provide further insights into their contribution to activity was only indirectly monitored by measuring pathogenesis (Yan et al., 2009; Verhaak et al., 2010). steady state NADP/NADPH ratios. It will be important In general, TP53 mutations are associated more fre- to evaluate the ability of these heterodimers to produce quently with low-grade astrocytomas and secondary D-2HG. If both subunits of the heterodimers retain their GBMs compared with primary GBMs. Within these respective enzymatic activities, it is possible that the lower grade tumors, TP53 mutations are more frequent heterodimeric enzyme is more efficient than the mutant across diffuse astrocytomas (74%), anaplastic astro- homodimeric enzyme in catalyzing the partial reverse cytomas (65%) and secondary GBMs (62%), whereas reaction. The close proximity of the two active sites they are less frequent in oligodendrogliomas and could lead to substrate channeling or local concentra- anaplastic oligodendrogliomas, with an overall preva- tion effects that would drive the production of D-2HG lence of 16–9%, respectively. Conversely, deletions in 1p from aKG. Alternatively, if cells heterozygous for IDH and 19q are found more often in oligodendroglial mutations produce sufficient levels of D-2HG to tumors than in astrocytic tumors. Interestingly, in these drive tumorigenesis, irrespective of the presence of two tumor types IDH1/IDH2 mutations are found at heterodimeric enzyme, there may simply be no selective relatively similar frequency (60–100%; Yan et al., 2009). pressure to drive loss-of-heterozygosity. It is equally Unlike the clear association between TP53 loss, 1p19q possible that cancer cells require wild-type IDH enzy- loss and tumor lineage, IDH1/IDH2 mutations may matic activity in order to remain viable. As future contribute to the pathogenesis of both astrocytic and studies define the role of mutant enzyme and D-2HG in oligodendroglial tumors. Noteworthy, sequential biop- tumorigenesis, the exact role of wild-type IDH in these sies from multiple patients showed no cases in which an tumors will likely become apparent. IDH1 mutation occurred after the acquisition of a TP53 mutation or loss of 1p/19q suggesting that acquisition of IDH1 mutations are early events in gliomagenesis and may affect a common glial precursor cell (Hartmann Incidence of IDH1/IDH2 mutations and correlations et al., 2009; Watanabe et al., 2009). to other genetic alterations High-grade primary GBMs are often associated with activating mutations or amplification of epidermal Following the initial discovery of IDH1 and IDH2 growth factor receptor and PTEN (phosphatase and mutations in low-grade gliomas and secondary GBMs, tensin homolog) deletion (Voelzke et al., 2008; Huang subsequent sequencing efforts revealed alterations in et al., 2009). Analysis of the relationship between genetic these two genes across additional cancers including alterations in epidermal growth factor receptor and acute myeloid leukemia, prostate, B-acute lymphoblas- IDH1/IDH2 in GBMs revealed that amplification of tic leukemia and others (Yan et al., 2009). In gliomas EGFR and IDH1/IDH2 mutations are mutually ex- and AML, mutations in IDH1 and IDH2 were mutually clusive events (Sanson et al., 2009). Similarly, genomic exclusive of one another and occur at very early stages profiling of 200 GBMs and two normal brain samples

Oncogene Cancer-associated IDH mutations KE Yen et al 6413 has defined four subtypes of tumors with a common and 19% of the cases respectively. More recently, unlike morphologic diagnosis of GBM. Consistent with earlier gliomas, two patients were found to harbor both an reports, IDH1 mutations are most commonly found in IDH1 and IDH2 mutation (Marcucci et al., 2010; the proneural class along with TP53 mutations, whereas Paschka et al., 2010). Unbiased sequencing has also PTEN, neurofibromin 1 (NFI) and epidermal growth revealed the presence of 12 non-synonymous recurring factor receptor alterations are more commonly associated mutations (including IDH1) in more than one AML with classical and mesenchymal subclasses (Verhaak genome. The data suggest that these alterations may be et al., 2010). Finally, hypermethylation of a large number important for the pathogenesis of the tumor and indeed of loci has revealed the existence of a glioma–CpG island all these mutations were retained in the dominant clone methylator phenotype that belongs to the proneural (Mardis et al., 2009). Similar to that in glioma, samples subgroup of gliomas (Noushmehr et al.,2010).In from AML patients harboring IDH1 and IDH2 muta- particular, methylation of the 06-Methylguanine-DNA tions showed elevated levels of D-2HG (Gross et al., Methyltransferase (MGMT) gene promoter confers a 2010; Ward et al., 2010). better prognosis for patients and is often associated with To gain further insights into potential cooperating IDH1 and IDH2 mutations (Sanson et al.,2009; events in AML, several groups have examined the Noushmehr et al., 2010). Whether this is a direct correlation between IDH1/IDH2 mutations and other consequence of the activity of mutant IDH or a surrogate common genetic alterations in AML (Lalisang et al., marker for epigenetic changes in tumors harboring 1997; Marcucci et al., 2010). Mutations in the phospho- IDH1/IDH2 mutation remains to be determined. protein NPM1 are found in 25–35% of adult de novo IDH1 mutations (R132H) have also been found in up AML and are associated with a wide spectrum of to 33% of rare, aggressive adult supratentorial primitive morphological subtypes. Recent studies across large neuroectodermal tumors of the cerebral hemispheres patient cohorts revealed that IDH1/IDH2 mutations are (Balss et al., 2008; Hayden et al., 2009). The fact that strongly associated with NPM1 mutation. Activating none of the supratentorial primitive neuroectodermal FLT3/ITD mutations are also common in AML (30– tumors arising in the infratentorial compartment (that 35%) and result in increased proliferation and reduced is, medulloblastomas) contain IDH1 mutations further susceptibility to apoptosis in hematopoietic cells. Both supports the idea that the supratentorial and infra- IDH1 and IDH2 mutations are seen in AMLs and tentorial compartments have distinct mechanisms of subgroup analysis of AML patients have shown an disease pathogenesis despite having nearly identical adverse effect of IDH mutations on overall survival histopathology. These findings may help to stratify among patients with normal karyotype AML (Paschka patients with supratentorial primitive neuroectodermal et al., 2010). Retrospective clinical trial data has tumors most likely to respond to targeted therapy revealed that the presence of IDH1 mutations is directed toward IDH1 mutations once available. Note- associated with a worse prognosis in younger cytogen- worthy, these tumors primarily affect children and etically normal AML patients with mutated NPM1 interestingly, the presence of the IDH1 mutation was without FLT3-ITD and patients with IDH2 mutations found exclusively in the adult patient tumors (X18 years had a significantly lower complete response rate of age) in two independent studies suggesting that this (Marcucci et al., 2010; Paschka et al., 2010). Interest- mutation does not have a significant role in childhood ingly a unique microRNA signature previously found to tumors of this type (Hayden et al., 2009). block myeloid cell and stem cell differentiation was also observed among these patients (Marcucci et al., 2010). Additional studies will help understand whether they II. Acute myeloid leukemia may contribute to block differentiation of these cell The first study on IDH mutations in AML, reported the populations and impact disease progression. identification on IDH1R132 alterations in B8% of the 188 cases analyzed (Mardis et al., 2009). An association with normal cytogenetics was noted. Subsequent III. Myelodysplastic syndrome (MDS) and sequencing efforts revealed the existence of mitochon- myeloproliferative neoplasm (MPN) drial IDH2 mutations (Lalisang et al., 1997; Gross et al., MDS and MPN are clonal disorders of hematopoiesis 2010; Ward et al., 2010). Interestingly IDH2 mutations characterized by the proliferation of myeloid cells. are significantly more common in AML than gliomas. Patients with MDS and MPN have a significantly Two groups reported the identification of a novel higher risk of developing secondary AML. Importantly, leukemia-associated mutation in IDH2, resulting in the mutations in IDH enzymes have been linked to this substitution of Arg140 for glutamate in the catalytic leukemic transformation, an event that is associated arginine triad (Green and Beer, 2010; Ward et al., 2010; with a poor clinical prognosis (Abdel-Wahab et al., Figure 2). In a small cohort of western patients, IDH1 2010). Sequence analysis has shown that IDH mutations mutations (R132 H, R132C and R132G) were observed occur in B5% MDS, 8.8% of MPN and 9.7% of in 6/78 samples (7.7%), whereas IDH2 mutations were secondary AML patients and were always heterozygous seen at a higher frequency of 12/78 (15.4%; Ward et al., and mutually exclusive of one another (Kosmider et al., 2010). Subsequently, within a more homogenous group 2010). Of note, mutations in IDH2 were restricted comprised of adult de novo cytogenetically normal AML to MPN, whereas the IDH1R132H mutation was patients, mutations in IDH1 and IDH2 were found in 14 detected in MDS and secondary AML patient samples

Oncogene Cancer-associated IDH mutations KE Yen et al 6414 (Kosmider et al., 2010). Previous studies have shown (Murugan et al., 2010; Figure 2). Unlike Arg132 that mutations in JAK2 (V617F), TET2 and ASXL1 are involved in substrate coordination, both residues are also commonly found in these diseases however, were located in the vicinity of the catalytic domain region typically wild type in patient samples that contained and their impact on IDH1 enzyme activity remains mutations in either IDH1 or IDH2 as noted above unclear. Interestingly, follicular thyroid cancer is a (Abdel-Wahab et al., 2010; Kosmider et al., 2010). precursor to anaplastic thyroid cancer suggesting that Finally, recent data indicates that IDH1 mutations in IDH1 mutations could be an early event in tumorigen- MDS are associated with a shorter overall survival and a esis and have a role in the progression of the disease. In higher rate of transformation into AML (Thol et al., addition, IDH1 mutations have been observed in a few 2010). In blast phase MPN, the presence of IDH human cancer cell lines. For example, a V71I mutation mutations predicted worse survival (Pardanani et al., has been found in a plasma cell myeloma line RPMI- 2010a). In addition, studies both in MPN as well as in 8226, and G97D mutations have been found in the high-risk MDS with del (5q) suggest that acquisitions of colorectal cancer cell lines DLD-1 and HCT15 both IDH1 and IDH2 mutations are early events and have a derived from the same tumor, as well as in a cell line role in disease progression (Pardanani et al., 2010a, b). derived from glioma, HGG153 (Catalogue of Somatic Mutations in Cancer; Bleeker et al., 2009). One cell line MZ7-mel derived from a malignant melanoma, also IV. IDH mutations in other malignancies contains a V294M mutation in IDH2 (Catalogue of In addition to cancers of the central nervous system Somatic Mutations in Cancer). Future studies will and AML, IDH1 mutations have been identiﬁed at elucidate whether these mutations are present in primary low frequency in prostate cancer (2.7%) and B-acute tumors and whether they affect the biochemical activity lymphoblastic leukemias (1.7%), suggesting that IDH1 of IDH1 and IDH2. mutations may have a role in the pathogenesis of both tumors types (Kang et al., 2009; Table 1). Follow-up studies across many more patients will clarify the D-2HG as a potential biomarker of cancers harboring prevalence of mutations in IDH enzymes in these cancers. IDH1 and IDH2 mutations In addition, analysis of 365 pheochromocytomas and paragangliomas for alterations in codons 132 and 172 of Elevated levels of D-2HG are a common feature of IDH1 and IDH2, respectively, identiﬁed one case of a human tumors with IDH1 and IDH2 mutations. heterozygous IDH1R132C mutation (Gaal et al.,2010). Because this metabolite is a direct product of mutant A series of less characterized mutations in IDH1 have IDH1 and IDH2, it provides a means to track the been reported in thyroid and colorectal cancers. The two activity of mutant IDH enzymes. Our initial studies most aggressive subtypes of thyroid cancer, anaplastic suggest that D-2HG may represent a biomarker with thyroid cancer and follicular thyroid cancer, both potential clinical utility. Across a panel of glioma harbor novel mutations in IDH1 corresponding to the samples, we observed an absolute correlation between predicted G123R and V71I amino acid substitutions elevated intracellular D-2HG levels and the presence of

Table 1 IDH1 and IDH2 mutations in hematologic malignancies and solid tumors Cancers IDH1 IDH2 References

R132H R172K R132C R172M R132S R172S Balss et al., 2008; Parsons et al., 2008; Dang et al., 2009; CNS cancers (gliomas, sPNET) R132G R172G Hartmann et al., 2009; Hayden et al., 2009; Yan et al., 2009 R132L R172W R132V R132H R140Q R132G R172K Mardis et al., 2009; Green and Beer, 2010; Chou et al., 2010; Aute myeloid leukemia, myeloid dysplastic R132C R140W Gross et al., 2010; Ho et al., 2010; Kosmider et al., 2010; syndrome and myeloid proliferative neopalsms R132S R172G Marcucci et al., 2010; Ward et al., 2010 R132L R140L R132H et al. Prostate cancer R132C Kang , 2009 Acute B-lymphoblastic leukemia R132H Kang et al., 2009 Paragangliomas R132C Gaal et al., 2010 R132H Colon cancer R132C Sjo¨blom et al., 2006; Kang et al., 2009; Bleeker et al., 2009 V71I Follicular and anaplastic thyroid cancer Murugan et al., 2010 G123R

Abbreviations: CNS, central nervous system; IDH, isocitrate dehydrogenase; sPNET, supratentorial primitive neuroectodermal tumor.

Oncogene Cancer-associated IDH mutations KE Yen et al 6415 IDH mutations (Dang et al., 2009). Similarly, in AML enantiomers are associated with pathological acidurias. intracellular D-2HG levels correlated with IDH1/IDH2 Indeed, elevated D-2HG and L-2HG in urine, plasma mutational status (Gross et al., 2010). Because D-2HG is and CSF characterize this group of rare neurometabolic excreted from cells, the presence of this metabolite in disorders with inherited mutations in D-2HGDH as well serum from patients with IDH mutant AML was as L-2HGDH. examined. In this case, high levels of D-2HG in serum Characterization of the molecular events triggered by also correlated with the presence of IDH mutations elevated D-2HG has revealed that this metabolite has an (Gross et al., 2010). These results underscore the excitotoxic effect on neurons. Activation of N-methyl-D- potential utility of 2HG as serum biomarker. In parallel, aspartic acid receptors by D-2HG results in disruption of Ward et al. (2010) determined relative levels of Ca2 þ homeostasis and an increase in reactive oxygen intracellular 2HG in a panel of 18 AML clinical species levels of primary neuronal cultures (Kolker samples, and based on metabolite levels, the presence et al., 2002). In addition, several observations provide of IDH mutations was identified with 100% accuracy. the framework to begin to link D-2HG and tumorigen- Additionally, the novel IDH2(R140Q) mutation in esis (Figure 1b). D-2HG exerts a direct inhibitory effect AML was identified through this screening approach, on adenosine-50-triphosphate synthase contributing to further demonstrating the utility of D-2HG as a mitochondrial dysfunction. Whether these observations biomarker of IDH1/2 mutations. in primary neurons are relevant to the malignant state Although these initial results hold great promise, remains to be determined. It is tempting to speculate clearly, a number of important questions await investi- that 2HG-driven mitochondrial dysfunction may con- gation. In particular, the correlation of serum D-2HG stitute selective pressure that promotes metabolic levels and disease burden as well as the potential of adaptation and a shift towards aerobic glycolysis, which D-2HG levels as a predictive marker of therapeutic could ultimately confer a growth advantage in a response to treatment with standard of care deserve hyperproliferative cellular state. careful evaluation. It has been reported that D-2HG may The dicarboxylic acids aKG and D-2HG are very be tractable via magnetic resonance spectroscopy similar whereby a ketone group and hydroxyl group on (Aydin et al., 2003; Sener, 2003). If that were the the second carbon functionally distinguish these two case, non-invasive detection of D-2HG by magnetic molecules. On the basis of their chemical similarity, resonance spectroscopy (MRS) could help to diagnose several groups have proposed that D-2HG could and guide therapy before the surgery in patients with modulate the activity of aKG-dependent enzymes or less accessible tumors such as gliomas. As additional interfere with aKG-dependent processes linked to clinical data emerges on efficacy of specific therapeutic oncogenesis (Frezza et al., 2010; Gross et al., 2010). strategies in gliomas with IDH1 or IDH2 mutations, This hypothesis is also in part rooted in our current early diagnosis before surgery may have significant understanding of the impact of alterations in the therapeutic impact. Additionally, it will be important to tricarboxylic acid cycle enzymes and tumor suppressors establish whether this metabolite constitutes a pharmaco- SDH and FH (Pollard et al., 2005; Selak et al., 2005; dynamic biomarker to monitor the activity of inhibitors Kaelin, 2009). In this case, loss of SDH and FH targeting mutant IDH enzymes once available. results in elevated levels of succinate and fumarate, respectively, which in turn interfere with the activity of the superfamily of aKG-dependent dioxygenases The metabolic product of somatic mutant IDH enzymes as that catalyze a wide range of biochemical reactions a modulator of cancer-relevant processes (Loenarz and Schofield, 2008). Potential targets of D-2HG include histone and DNA demethylases that The five carbon dicarboxylic acid D-2HG is normally regulate chromatin remodeling, prolyl-hydroxylases found in human urine, and is an intermediate in the involved in HIF1a degradation as well as prolyl and synthesis of 5-aminolaevulinate and porphyrin for heme lysyl hydroxylases that mediate collagen biosynthesis. synthesis, as well as metabolism of hydroxy-lysine Thus, the intriguing possibility that the elevated levels of (Lindahl et al., 1967; Chalmers et al., 1980). Our current D-2HG, the metabolic product of mutant IDH1 and understanding of the metabolic origin and fate of IDH2 enzymes, could impact epigenetic state and gene D-2HG in mammalian cells is rather limited. This transcription, angiogenesis as well as extracellular metabolite is produced by the action of hydroxyacid matrix dynamics is actively being explored by many oxoacid transhydrogenase enzyme (HOT) in the liver, groups. Interestingly, a recent correlation between CpG kidney and brain of mammals. In this reaction, the island methylation phenotype and IDH1 mutational formation of D-2HG from a-KG is coupled to oxidation status has been noted (Noushmehr et al., 2010). This of g-hydroxybutyrate to succinic semialdehyde (Kaufman observation adds to the hypothesis linking D-2HG and et al., 1988). In addition, the promiscuous activity of epigenetic rewiring. metabolic enzymes is an alternative source of 2HG Finally, because D-2HG is excreted by malignant cells (Rzem et al., 2007). Under physiologic conditions the it is possible that this metabolite may mediate non-cell cellular levels of D-2HG rate are maintained in check by autonomous effects, impacting the tumor stroma, the activity of the enantiomer selective D-2-hydroxyglutaric normal surrounding tissue and even distal metastasis. dehydrogenase (D-2HGDH). Alterations in the balance Availability of selective inhibitors of mutant IDH between production and metabolism of both 2HG enzymes and of relevant genetically engineered mouse

Oncogene Cancer-associated IDH mutations KE Yen et al 6416 cancer models will enable the study of cell autonomous of a genetic biomarker for patient stratification and of a and paracrine effects of D-2HG. metabolic biomarker to track target activity makes mutant IDH1 and IDH2 unique opportunities for therapeutic intervention and drug development. Multiple Concluding remarks research teams are actively working to define whether interfering with D-2HG production results in a ther- The discovery of cancer-associated mutations in IDH1 apeutic benefit. Furthermore, understanding the pre- and IDH2 constitutes the first example of gain-of- valence of specific mutations in particular disease types function somatic genetic alterations in metabolic en- (Table 1) will provide valuable insight into the mechan- zymes. The discovery that all mutants characterized to ism of disease progression associated with mutant date possess the ability to catalyze the reduction of aKG IDH1/IDH2. Finally, the utility of D-2HG as a biomarker to D-2HG has profound implications. From a pragmatic may offer a unique opportunity for clinicians and stand point, the metabolite D-2HG provides a biomar- patients. For example, in AML, the ability to detect ker of mutant enzyme activity with potential clinical D-2HG in serum and the potential to image D-2HG in applicability. In addition, based on the similarities gliomas via MRS may provide a means to track the between aKG and D-2HG, a series of hypotheses have progression and/or relapse of disease without painful been proposed to molecularly link IDH mutations and invasive procedures. Although the ultimate diagnosis tumorigenesis. Ongoing studies will help determine will come via genetic confirmation of the presence or whether elevated D-2HG levels interfere with mitochon- absence of IDH1 and IDH2 mutations, plasma D-2HG drial activity and with aKG-dependent enzymes, and may serve as a pharmacodynamic biomarker and whether mutations in IDH, SDH and FH enzymes may potentially as a response biomarker, once inhibitors represent another example of evolutionary convergence. become available. On the basis of genetic and prognostic correlations, mutant IDH1 and IDH2 have emerged as potential therapeutic targets. In addition, replacement of Arg 132 by smaller side chains or hydrophobic amino acids Conflict of interest suggests that the identification of selective inhibitors of mutant IDH1 may be feasible. Clearly, the availability The authors are employees of Agios Pharmaceuticals.

References

Abbas S, Lugthart S, Kavelaars FG, Schelen A, Koenders J, Zeilemaker Dang CV, Semenza GL. (1999). Oncogenic alterations of metabolism. A et al. (2010). Acquired mutations in the genes encoding IDH1 and Trends Biochem Sci 24: 68–72. IDH2 both are recurrent aberrations in acute myeloid leukemia Dang L, White DW, Gross S, Bennett BD, Bittinger MA, Driggers (AML): prevalence and prognostic value. Blood 116: 2122–2126. EM et al. (2009). Cancer-associated IDH1 mutations produce 2- Abdel-WahabO,ManshouriT,PatelJ,HarrisK,YaoJ,HedvatCet al. hydroxyglutarate. Nature 462: 739–744. (2010). Genetic analysis of transforming events that convert chronic Frezza C, Tennant DA, Gottlieb E. (2010). IDH1 mutations in myeloproliferative neoplasms to leukemias. Cancer Res 70: 447–452. gliomas: when an enzyme loses its grip. Cancer Cell 17: 7–9. Aydin K, Ozmen M, Tatli B, Sencer S. (2003). Single-voxel MR Gaal J, Burnichon N, Korpershoek E, Roncelin I, Bertherat J, Plouin spectroscopy and diffusion-weighted MRI in two patients with l-2- PF et al. (2010). Isocitrate dehydrogenase mutations are rare in hydroxyglutaric aciduria. Pediatr Radiol 33: 872–876. pheochromocytomas and paragangliomas. J Clin Endocrinol Metab Balss J, Meyer J, Mueller W, Korshunov A, Hartmann C, von 95: 1274–1278. Deimling A. (2008). Analysis of the IDH1 codon 132 mutation in Geisbrecht BV, Gould SJ. (1999). The human PICD gene encodes a brain tumors. Acta Neuropathol 116: 597–602. cytoplasmic and peroxisomal NADP(+)-dependent isocitrate dehy- Baysal BE, Ferrell RE, Willett-Brozick JE, Lawrence EC, Myssiorek drogenase. J Biol Chem 274: 30527–30533. D, Bosch A et al. (2000). Mutations in SDHD, a mitochondrial Gillies RJ, Gatenby RA. (2007). Hypoxia and adaptive landscapes in complex II gene, in hereditary paraganglioma. Science 287: 848–851. the evolution of carcinogenesis. Cancer Metastasis Rev 26: 311–317. Bleeker FE, Lamba S, Leenstra S, Troost D, Hulsebos T, Vandertop Green A, Beer P. (2010). Somatic mutations of IDH1 and IDH2 in the WP et al. (2009). IDH1 mutations at residue p. R132 (IDH1(R132)) leukemic transformation of myeloproliferative neoplasms. N Engl J occur frequently in high-grade gliomas but not in other solid Med 362: 369–370. tumors. Hum Mutat 30: 7–11. Gross S, Cairns RA, Minden MD, Driggers EM, Bittinger MA, Jang Brandon M, Baldi P, Wallace DC. (2006). Mitochondrial mutations in HG et al. (2010). Cancer-associated metabolite 2-hydroxyglutarate cancer. Oncogene 25: 4647–4662. accumulates in acute myelogenous leukemia with isocitrate dehy- Ceccarelli C, Grodsky NB, Ariyaratne N, Colman RF, Bahnson BJ. drogenase 1 and 2 mutations. J Exp Med 207: 339–344. (2002). Crystal structure of porcine mitochondrial NADP+-depen- Hanahan D, Weinberg RA. (2000). The hallmarks of cancer. Cell 100: dent isocitrate dehydrogenase complexed with Mn2+ and isocitrate. 57–70. Insights into the enzyme mechanism. JBiolChem277: 43454–43462. Hartmann C, Meyer J, Balss J, Capper D, Mueller W, Christians A Chalmers RA, Lawson AM, Watts RW, Tavill AS, Kamerling JP, Hey et al. (2009). Type and frequency of IDH1 and IDH2 mutations E et al. (1980). D-2-hydroxyglutaric aciduria: case report and are related to astrocytic and oligodendroglial differentiation and biochemical studies. J Inherit Metab Dis 3: 11–15. age: a study of 1010 diffuse gliomas. Acta Neuropathol 118: 469–474. Chou WC, Hou HA, Chen CY, Tang JL, Yao M, Tsay W et al. (2010). Hayden JT, Fruhwald MC, Hasselblatt M, Ellison DW, Bailey S, Distinct clinical and biological characteristics in adult acute myeloid Clifford SC. (2009). Frequent IDH1 mutations in supratentorial leukemia bearing isocitrate dehydrogenase 1 (IDH1) mutation. primitive neuroectodermal tumors (sPNET) of adults but not Blood 115: 2749–2754. children. Cell Cycle 8: 1806–1807.

Oncogene Cancer-associated IDH mutations KE Yen et al 6417 Henke B, Girzalsky W, Berteaux-Lecellier V, Erdmann R. (1998). mutation without FLT3 internal tandem duplication. J Clin Oncol IDP3 encodes a peroxisomal NADP-dependent isocitrate dehydro- 28: 3636–3643. genase required for the beta-oxidation of unsaturated fatty acids. Pollard PJ, Briere JJ, Alam NA, Barwell J, Barclay E, Wortham NC J Biol Chem 273: 3702–3711. et al. (2005). Accumulation of Krebs cycle intermediates and over- Ho PA, Alonzo TA, Kopecky KJ, Miller KL, Kuhn J, Zeng R et al. expression of HIF1alpha in tumours which result from germline FH (2010). Molecular alterations of the IDH1 gene in AML: a and SDH mutations. Hum Mol Genet 14: 2231–2239. Children’s Oncology Group and Southwest Oncology Group study. Rzem R, Vincent MF, Van Schaftingen E, Veiga-da-Cunha M. (2007). Leukemia 24: 909–913. L-2-hydroxyglutaric aciduria, a defect of metabolite repair. J Inherit Huang PH, Xu AM, White FM. (2009). Oncogenic EGFR signaling Metab Dis 30: 681–689. networks in glioma. Sci Signal 2: re6. Sanson M, Marie Y, Paris S, Idbaih A, Laffaire J, Ducray F et al. (2009). Jones PA, Baylin SB. (2002). The fundamental role of epigenetic events Isocitrate dehydrogenase 1 codon 132 mutation is an important in cancer. Nat Rev Genet 3: 415–428. prognostic biomarker in gliomas. J Clin Oncol 27: 4150–4154. Kaelin Jr WG. (2009). SDH5 mutations and familial paraganglioma: Selak MA, Armour SM, MacKenzie ED, Boulahbel H, Watson DG, somewhere Warburg is smiling. Cancer Cell 16: 180–182. Mansfield KD et al. (2005). Succinate links TCA cycle dysfunction Kang MR, Kim MS, Oh JE, Kim YR, Song SY, Seo SI et al. (2009). to oncogenesis by inhibiting HIF-alpha prolyl hydroxylase. Cancer Mutational analysis of IDH1 codon 132 in glioblastomas and other Cell 7: 77–85. common cancers. Int J Cancer 125: 353–355. Sener RN. (2003). L-2 hydroxyglutaric aciduria: proton magnetic Kaufman EE, Nelson T, Miller D, Stadlan N. (1988). Oxidation of resonance spectroscopy and diffusion magnetic resonance imaging gamma-hydroxybutyrate to succinic semialdehyde by a mitochondrial findings. J Comput Assist Tomogr 27: 38–43. pyridine nucleotide-independent enzyme. J Neurochem 51: 1079–1084. Sjo¨blom T, Jones S, Wood LD, Parsons DW, Lin J, Barber TD et al. King A, Selak MA, Gottlieb E. (2006). Succinate dehydrogenase and (2006). The consensus coding sequences of human breast and fumarate hydratase: linking mitochondrial dysfunction and cancer. colorectal cancers. Science 314: 268–274. Oncogene 25: 4675–4682. Thol F, Weissinger EM, Krauter J, Wagner K, Damm F, Kolker S, Pawlak V, Ahlemeyer B, Okun JG, Horster F, Mayatepek E Wichmann M et al. (2010). IDH1 mutations in patients with et al. (2002). NMDA receptor activation and respiratory chain myelodysplastic syndromes are associated with an unfavorable complex V inhibition contribute to neurodegeneration in d-2- prognosis. Haematologica (e-pub ahead of print 21 May 2010; hydroxyglutaric aciduria. Eur J Neurosci 16: 21–28. doi:10.3324/haematol.2010.025494). Kosmider O, Gelsi-Boyer V, Slama L, Dreyfus F, Beyne-Rauzy O, Tomlinson IP, Alam NA, Rowan AJ, Barclay E, Jaeger EE, Kelsell D Quesnel B et al. (2010). Mutations of IDH1 and IDH2 genes in early et al. (2002). Germline mutations in FH predispose to dominantly and accelerated phases of myelodysplastic syndromes and MDS/ inherited uterine fibroids, skin leiomyomata and papillary renal cell myeloproliferative neoplasms. Leukemia 24: 1094–1096. cancer. Nat Genet 30: 406–410. Kroemer G, Pouyssegur J. (2008). Tumor cell metabolism: cancer0s van Roermund CW, Hettema EH, Kal AJ, van den Berg M, Tabak Achilles’ heel. Cancer Cell 13: 472–482. HF, Wanders RJ. (1998). Peroxisomal beta-oxidation of polyunsa- Lalisang RI, Wils JA, Nortier HW, Burghouts JT, Hupperets PS, turated fatty acids in Saccharomyces cerevisiae: isocitrate dehydro- Erdkamp FL et al. (1997). Comparative study of dose escalation genase provides NADPH for reduction of double bonds at even versus interval reduction to obtain dose-intensification of epirubicin positions. EMBO J 17: 677–687. and cyclophosphamide with granulocyte colony-stimulating factor Vander Heiden MG, Cantley LC, Thompson CB. (2009). Under- in advanced breast cancer. J Clin Oncol 15: 1367–1376. standing the Warburg effect: the metabolic requirements of cell Lindahl G, Lindstedt G, Lindstedt S. (1967). Metabolism of 2-amino- proliferation. Science 324: 1029–1033. 5-hydroxyadipic acid in the rat. Arch Biochem Biophys 119: 347–352. Verhaak RG, Hoadley KA, Purdom E, Wang V, Qi Y, Wilkerson MD Loenarz C, Schofield CJ. (2008). Expanding chemical biology of et al. (2010). Integrated genomic analysis identifies clinically 2-oxoglutarate oxygenases. Nat Chem Biol 4: 152–156. relevant subtypes of glioblastoma characterized by abnormalities Marcucci G, Maharry K, Wu YZ, Radmacher MD, Mrozek K, in PDGFRA, IDH1, EGFR, and NF1. Cancer Cell 17: 98–110. Margeson D et al. (2010). IDH1 and IDH2 gene mutations identify Voelzke WR, Petty WJ, Lesser GJ. (2008). Targeting the epidermal novel molecular subsets within de novo cytogenetically normal acute growth factor receptor in high-grade astrocytomas. Curr Treat myeloid leukemia: a cancer and leukemia group B study. J Clin Options Oncol 9: 23–31. Oncol 28: 2348–2355. Vogelstein B, Kinzler KW. (1993). The multistep nature of cancer. Mardis ER, Ding L, Dooling DJ, Larson DE, McLellan MD, Chen K Trends Genet 9: 138–141. et al. (2009). Recurring mutations found by sequencing an acute Ward PS, Patel J, Wise DR, Abdel-Wahab O, Bennett BD, Coller HA myeloid leukemia genome. N Engl J Med 361: 1058–1066. et al. (2010). The common feature of leukemia-associated IDH1 and Murugan AK, Bojdani E, Xing M. (2010). Identification and functional IDH2 mutations is a neomorphic enzyme activity converting alpha- characterization of isocitrate dehydrogenase 1 (IDH1) mutations in ketoglutarate to 2-hydroxyglutarate. Cancer Cell 17: 225–234. thyroid cancer. Biochem Biophys Res Commun 393: 555–559. Watanabe T, Nobusawa S, Kleihues P, Ohgaki H. (2009). IDH1 Noushmehr H, Weisenberger DJ, Diefes K, Phillips HS, Pujara K, Berman mutations are early events in the development of astrocytomas and BP et al. (2010). Identification of a CpG island methylator phenotype oligodendrogliomas. Am J Pathol 174: 1149–1153. that defines a distinct subgroup of glioma. Cancer Cell 17: 510–522. Xu X, Zhao J, Xu Z, Peng B, Huang Q, Arnold E et al. (2004). Pardanani A, Lasho TL, Finke CM, Mai M, McClure RF, Tefferi A. Structures of human cytosolic NADP-dependent isocitrate (2010a). IDH1 and IDH2 mutation analysis in chronic- and blast- dehydrogenase reveal a novel self-regulatory mechanism of activity. phase myeloproliferative neoplasms. Leukemia 24: 1146–1151. J Biol Chem 279: 33946–33957. Pardanani A, Patnaik MM, Lasho TL, Mai M, Knudson RA, Finke C Yan H, Parsons DW, Jin G, McLendon R, Rasheed BA, Yuan W et al. et al. (2010b). Recurrent IDH mutations in high-risk myelodysplas- (2009). IDH1 and IDH2 mutations in gliomas. N Engl J Med 360: tic syndrome or acute myeloid leukemia with isolated del(5q). 765–773. Leukemia 24: 1370–1372. Yoshihara T, Hamamoto T, Munakata R, Tajiri R, Ohsumi M, Parsons DW, Jones S, Zhang X, Lin JC, Leary RJ, Angenendt P et al. Yokota S. (2001). Localization of cytosolic NADP-dependent (2008). An integrated genomic analysis of human glioblastoma isocitrate dehydrogenase in the peroxisomes of rat liver cells: multiforme. Science 321: 1807–1812. biochemical and immunocytochemical studies. J Histochem Cyto- Paschka P, Schlenk RF, Gaidzik VI, Habdank M, Kronke J, Bullinger chem 49: 1123–1131. L et al. (2010). IDH1 and IDH2 mutations are frequent genetic Zhao S, Lin Y, Xu W, Jiang W, Zha Z, Wang P et al. (2009). Glioma- alterations in acute myeloid leukemia and confer adverse prognosis derived mutations in IDH1 dominantly inhibit IDH1 catalytic in cytogenetically normal acute myeloid leukemia with NPM1 activity and induce HIF-1alpha. Science 324: 261–265.

Oncogene