DOI: 10.1126/scitranslmed.3002693, 116ra4 (2012);4 Sci Transl Med

, et al.Ovidiu C. AndronesiSpectroscopyVivo Spectral-Editing and 2D Correlation Magnetic Resonance

-Mutated Glioma Patients by InIDHDetection of 2-Hydroxyglutarate in

 Editor's Summary

   

and genetic makeup.procedures and help doctors not only predict cancer outcomes but also effectively treat tumors on the basis of grade

one that would avoid invasive clinical−−in vivo brain imaging for genotyping cancer patients is a possibility (also in this issue), the authors show thatet al.With these results and those in the companion study by Elkhaled

tissue biopsies.. The method was further validated ex vivo inIDH mutations, but not in healthy volunteers with wild-type IDH1with

however, two-dimensional MRS allowed the authors to see the presence of 2HG in the brains of two glioma patients metabolites, such as glutamate and glutamine, precludes detection with traditional one-dimensional spectroscopy;patients with glioma using magnetic resonance spectroscopy (MRS) imaging of 2HG. The similarity of 2HG to other

mutations inIDH1 developed a strategy to detect et al.such mutations could be used for prognosis. Andronesi gliomas, suggesting thatIDH1gene mutations have a greater 5-year survival rate than do patients with wild-type

IDH12-hydroxyglutarate (2HG). This mutation has been found in up to 86% of grade II to IV gliomas. Patients with Mutations in the enzyme isocitrate dehydrogenase (IDH) lead to the accumulation of the metabolite

imaging the brain to identify a glioma gene mutation that is correlated with patient survival.''see'' inside your head; but, for gliomas, Andronesi and coauthors have found it to be beneficial by noninvasivelyat the time of surgery, with results available only after several weeks. Normally, it is a good thing that people can't

unless you can sample the diseased tissue itself via biopsy. This invasive procedure is typically performed−−predictGliomas are diffuse brain tumors that are difficult to diagnose, with outcomes that are nearly impossible to

Spectroscopy Gets Inside Your Head

http://stm.sciencemag.org/content/4/116/116ra4.full.htmlcan be found at:

and other services, including high-resolution figures,A complete electronic version of this article

http://stm.sciencemag.org/content/suppl/2012/01/09/4.116.116ra4.DC1.html can be found in the online version of this article at: Supplementary Material

http://stm.sciencemag.org/content/scitransmed/4/116/116ra5.full.html http://stm.sciencemag.org/content/scitransmed/4/116/116ps1.full.html

can be found online at:Related Resources for this article

http://www.sciencemag.org/about/permissions.dtl in whole or in part can be found at: article

permission to reproduce this of this article or about obtaining reprintsInformation about obtaining

is a registered trademark of AAAS. Science Translational Medicinerights reserved. The title NW, Washington, DC 20005. Copyright 2012 by the American Association for the Advancement of Science; alllast week in December, by the American Association for the Advancement of Science, 1200 New York Avenue

(print ISSN 1946-6234; online ISSN 1946-6242) is published weekly, except theScience Translational Medicine

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

R E S EARCH ART I C L E

BRA IN IMAG ING

Detection of 2-Hydroxyglutarate in IDH-MutatedGlioma Patients by In Vivo Spectral-Editing and 2DCorrelation Magnetic Resonance SpectroscopyOvidiu C. Andronesi,1* Grace S. Kim,1 Elizabeth Gerstner,2 Tracy Batchelor,2 Aria A. Tzika,1,3

Valeria R. Fantin,4 Matthew G. Vander Heiden,5,6 A. Gregory Sorensen1

9, 2

012

Mutations in the gene isocitrate dehydrogenase 1 (IDH1) are present in up to 86% of grade II and III gliomas andsecondary glioblastoma. Arginine 132 (R132) mutations in the enzyme IDH1 result in excess production of themetabolite 2-hydroxyglutarate (2HG), which could be used as a biomarker for this subset of gliomas. Here, we useoptimized in vivo spectral-editing and two-dimensional (2D) correlationmagnetic resonance spectroscopy (MRS)methods to unambiguously detect 2HG noninvasively in glioma patients with IDH1 mutations. By comparison,fitting of conventional 1D MR spectra can provide false-positive readouts owing to spectral overlap of 2HG andchemically similar brain metabolites, such as glutamate and glutamine. 2HG was also detected using 2D high-resolutionmagic angle spinningMRSperformedex vivo on a separate set of gliomabiopsy samples. 2HGdetectionbyin vivo or ex vivo MRS enabled detailedmolecular characterization of a clinically important subset of human gliomas.This has implications for diagnosis as well as monitoring of treatments targeting mutated IDH1.

arch

on

Mst

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

INTRODUCTION

Isocitrate dehydrogenase 1 (IDH1) is an intracellular enzyme that cat-alyzes the oxidative decarboxylation of isocitrate to a-ketoglutarate inthe cytoplasm and in peroxisomes. Recent genomic studies have iden-tified heterozygous pointmutations in arginine 132 (R132) of the IDH1enzyme (1, 2). These mutations result in a neomorphic activity leadingto overproduction and accumulation of the R (also D) enantiomer of themetabolite 2-hydroxyglutarate (2HG) in 68 to 86% of grade II to III as-trocytic and oligodendroglial tumors, aswell as grade IV secondary glio-blastoma, having higher frequency in young patients (3–5). Gliomapatients with mutations in the gene IDH1 have a greater 5-year sur-vival rate than patients with wild-type IDH1 gliomas (93% versus 51%)when correcting for age (3), suggesting that IDH1mutations represent aclinically distinct subset of patients. In addition to glioma, mutations inIDH1have also been found in patientswith acutemyelogenous leukemiaand various other tumors, but at lower frequency than in glioma (6).

The full impact of the R132 mutation is not yet fully understood,but a major consequence of mutating this residue in IDH1 is a gain-of-function enzymatic activity favoring reduction of a-ketoglutarateto 2HG (7). This neomorphic activity leads to the accumulation of 2HG,a metabolite usually present in low levels in vivo as an error product ofnormal metabolism. Analogous mutations in the mitochondrial IDH2isoform also result in 2HG production, but IDH2mutations are foundless frequently than IDH1 in various tumors, including gliomas (4).

2HG is a small biomolecule that has been shown ex vivo to identifyIDH1/2-mutant tumors in humans (8). In transfectedU87MG glioblas-toma cell cultures, the intracellular concentration of 2HG can increase

1Athinoula A. Martinos Center for Biomedical Imaging, Department of Radiology, Mas-sachusetts General Hospital, Harvard Medical School, Boston, MA 02114, USA. 2PappasCenter for Neuro-Oncology, Massachusetts General Hospital, Harvard Medical School,Boston, MA 02114, USA. 3NMR Surgical Laboratory, Department of Surgery, MassachusettsGeneral Hospital and Shriners Burn Institute, Harvard Medical School, Boston, MA 02114,USA. 4Agios Pharmaceuticals, Cambridge, MA 02139, USA. 5Koch Institute for IntegrativeCancer Research, Massachusetts Institute of Technology, Cambridge, MA 02139, USA.6Dana-Farber Cancer Institute, Boston, MA 02115, USA.*To whom correspondence should be addressed. E-mail: ovidiu@nmr.mgh.harvard.edu

www.Scien

more than 100-fold (7), up to 5 to 35 mmol/g (or 5 to 35mM, assumingtissue density of 1.05 g/ml), after introduction of R132-mutated IDH1in the U87MG genome. Accumulation of 2HG to similar levels as inU87MG cell cultures was measured in human glioma biopsy sampleswith IDH1R132 mutations (7, 9). This high concentration of 2HG (5 to35mM) is suitable for detection by in vivomagnetic resonance spectros-copy (MRS). Because the sensitivity threshold of in vivoMRS is roughly1 mM, 2HG is not expected to be visible under normal conditions, but2HGmight becomemeasurable upon local accumulation owing to IDHmutation. Thus, the presence or absence of 2HG in theMR spectrum ofglioma patients could effectively genotype tumors as being positive ornegative for IDH1 or IDH2 mutations.

The presence of the S (also L) enantiomer of 2HG (L-2HG) has beensuggested using MRS in vivo in patients with hydroxyglutaric aciduria(10, 11). The detection challenge arises from the fact that the 2HGspectrum is largely overlapping with glutamate (Glu) and glutamine(Gln)—both abundant brain metabolites that have a similar five-spinsystem. Peaks in the region of 2.6 to 2.4 ppm (parts per million) thatwere previously indicated (10, 11) for L-2HG are shared with Glu,Gln, and also by N-acetyl-L-aspartate (NAA). With the limited in vivospectral resolution (0.1 ppm) present in most clinical settings, theseoverlapping species are not easily resolved using conventional one-dimensional (1D) MRS, especially if spectral fitting was not used, as inthese earlier reports (10, 11). Spectral fitting programs (12, 13) try tomodel the in vivo MR spectrum as a combination of individual spectra(basis set) from all detectable metabolites. This approach might fail forsome metabolites at clinically available fields when there is severeoverlap, as it is known for g-aminobutyric acid (GABA) (14), or aswe show here for 2HG.

2D correlation spectroscopy (COSY) (15) can potentially differenti-ate the overlapping metabolite spectra, because correlating two chemi-cal shifts of coupled spins creates specific patterns of signals (cross peaks)for each metabolite that are better separated in the plane of the 2Dspectrum than single spectral lines in a 1D spectrum. The 2D COSY ex-ploits the idea that there is less likelihood for two metabolites to have

ceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ra4 1

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

two identical shifts, even if theymight share a common chemical shift inthe 1D spectrum. In particular, the cross peaks involving Ha protons of2HG appear in a region of the 2D COSY spectra where no other me-tabolite is found in healthy tissue or tumors without IDH mutations.Hence, although in 1D spectra the signals of 2HG appear in a regionwhere other metabolites normally contribute, in 2D COSY spectra thecross peaks involving Ha protons of 2HG can be uniquely identified.Alternatively, spectral editing of 1DMRS, such as J-difference spectros-copy (14), can be tuned to detect a specific metabolite by removing thecontribution of unwanted overlappingmetabolites. The spectral-editingexperiment can be easier to run on clinical scanners but offers limitedmetabolite information, whereas, on the other hand, the 2D COSY re-tains the full spectral information at the expense of complexity of theexperiment.

Here, we show that 2HG can be detected in glioma patients usingan optimized in vivo adiabatic 2D COSY method, developed previ-ously for studying brain metabolism (16), or by spectral-editing MRS.We also find that fitting conventional 1D spectra might provide false-positive results. In vivo measurements were compared with ex vivohigh-resolution magic angle spinning (HR-MAS) 2D MRS and liquidchromatography–mass spectrometry (LC-MS) of glioma biopsysamples. Results from brain phantoms, two glioma patients harboringthe IDH1R132 mutation, and eight control cases, including primary glio-blastoma (n = 4) and healthy volunteers (n = 4) with wild-type IDH1,demonstrate that noninvasive detection of 2HGusing 1D spectral-editingand 2D correlation MRS is feasible and may allow stratification ofpatients on the basis of IDH1 mutation.

RESULTS

Spectroscopic detection of 2HG in phantomsWe performed phantom experiments at 3 T on clinical scanners to es-tablish that 2HG can be distinguished from other metabolites by local-ized 2D correlation MRS as well as localized spectral-editing 1D MRS.2HG was added to a phantom containing a mixture of brain metabo-lites, and a recently developed 2D LASER-COSY sequence (16) withimproved in vivo performance based on localized adiabatic selectiverefocusing (LASER) was used as described in Materials and Methods.An adiabatic spectral-editing sequence (MEGA-LASER) was newly de-signed here specifically for the purpose of 2HGdetection (Materials andMethods and figs. S1 and S3). A series of phantomswith a range of 2HGconcentrations expected to be present in IDH1-mutant tumors were alsoinvestigated to test the sensitivity limit of MRS. Assignments of 2HG

www.Scien

(17) and other metabolites, such as myo-inositol (Myo), choline (Cho),NAA, Glu, Gln, andGABA, weremade (Fig. 1) according to publishedliterature values (18).

Fig. 1. 2D LASER-COSY and 1D MEGA-LASER spectra from brain phantoms3

at 3 T, with 3 × 3 × 3 cm voxels used in all measurements. (A) Overlay of 2D

LASER-COSY spectra from a phantom containing a mixture of normal brainmetabolites (red contours) and a phantom where 2HG was added to themixture of normal brain metabolites (blue contours). The Ha-Hb cross peakof 2HG is at 4.02/1.91 (d2/d1) ppm. (B) Overlay of 1D MEGA-LASER from thesame phantoms. The position of the Ha peak of 2HG at 4.02 ppm lines withthe cross peak in the 2D spectrumabove (dashed line). a.u., arbitrary units. (C)Intensity of 2HG signal in 2D LASER-COSY and 1D MEGA-LASER at different2HGconcentrations. Error bars represent 1 SDof two independentmeasures,with signal intensity normalized (Inorm) to the maximum intensity. Other me-tabolites shown: choline (Cho), g-aminobutyric acid (GABA), glutamate (Glu),lactate (Lac), myo-inositol (Myo), and N-acetyl-L-aspartate (NAA).

ceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ra4 2

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgw

nloa

ded

from

Figure 1A shows the overlay of 2D LASER-COSY spectra recordedin a phantom containing a mixture of 2HG and brainmetabolites (bluecontours), and a phantom that contains only normal brain metabolites(red contours). The normal brain metabolites are at physiological con-centrations in both phantoms. The Ha-Hb cross peak of 2HG locatedat 4.02/1.91 (d2/d1) ppm is well separated from other metabolites, in-cluding the chemically similar metabolite Glu, with Ha-Hb correlationlocated at 3.75/2.12 (d2/d1) ppm. Detailed information of the overlapbetween 2HG and other metabolites can be gleaned from spectra simu-lations (fig. S2). The strongly coupled five-spin system of 2HG,Glu, andGln is very similar, and a large overlap is observed in the 2.6-to 2.0-ppmregion for Hb and Hg protons. Additionally, GABA overlaps 2HG be-tween 2.0 to 1.8 ppm and 2.4 to 2.2 ppm.As expected from the chemicalstructure, the largest separation between 2HG, Glu, and Gln is noticedforHa protons owing to attached hydroxyl and aminomoieties at Ca on2HG and Glu/Gln, respectively. This Ha separation can be exploited inspectral-editing MRS.

Figure 1B shows the edited 1D spectra obtained with the MEGA-LASER sequence on the same phantom as Fig. 1A. The Hamultipletsignal of 2HG at 4.02 ppm (Fig. 1B, blue) is aligned with the 2HG crosspeak from the 2D LASER-COSY spectrum. The signal at 4.02 ppm ismissing in the brain phantom that does not contain 2HG (Fig. 1B, red).In addition to 2HG, themultiplet signals of Glu at 3.75 ppm andGABAat 3.01 ppm are co-edited, and their multiplets can be better observedin the inset of Fig. 1B. By comparison, in conventional 1D spectra ob-tained with LASER, the Ha proton of 2HG is largely overlapped by thestrong Hb peak of myo-inositol at 4.05 ppm (fig. S2). For in vivo spec-troscopy, which typically has lower spectral resolution owing to sus-ceptibility anisotropy of tissues, the Ha proton of 2HG might beobscured more even by the neighboring peaks of lactate (4.09 ppm)and both creatine and phosphocreatine (3.91 ppm).

MEGA-LASER showed excellent localization when compared toMEGA-PRESS in fig. S3, with no contamination of lipid signal fromoutside the voxel. The echo time (TE) ofMEGA-LASERwas optimizedaround the value of 1/2J (J, scalar coupling) formaximizingHa signal of2HG. The maximum was found for TE = 75 ms.

Calibrationmeasurements weremade for a series of 2HG phantomswith concentrations in the range of 0 to 16mM (Fig. 1C). A strong cor-relation (R = 0.992) was found between 2D LASER-COSY cross peak

www.Scien

volume and 2HG concentration. A sensitivity limit of 2 mM 2HG wascalculated for 2D LASER-COSY, with a voxel of 27 cm3, measurementtime of 12.8 min, and a minimum signal-to-noise ratio (SNR) of 5, forreliable identification of the Ha cross peaks. Similar strong correlation(R= 0.995) was found between 2HG concentration and the area of 2HGpeak in spectral-editing MEGA-LASER (Fig. 1C). The SNR of 5 for2 mM2HG concentration and 27 cm3 voxel can be reached byMEGA-LASER with a shorter acquisition time of 5 min.

2HG detection in intact brain biopsiesTo ensure that 2HG measurements were possible in human tissue,we measured brain biopsy samples (n = 10) before in vivo experiments(Table 1). Biopsies were used because they contain the full set of metab-olites and tumor metabolic profiles that are hard to replicate in phan-toms. 2D spectra were obtained with HR-MAS conditions at 14 T frombrain biopsies representing varied pathologies and IDH1 mutationstatus (Fig. 2). The 1H-1H2DTOBSY [total through-bond spectroscopy(18)] spectra of an IDH1-mutated anaplastic astrocytoma containedwell-resolved and separated 2HG cross peaks involving the correlations ofHawithHb (4.02/1.91 ppm) andHg (4.02/2.24 ppm) (Fig. 2A). The 2HGcross peaks from biopsy spectrum overlaid entirely with the correspond-ing2HGcrosspeaksof thephantomspectrum(Fig. 2A). Projections alongd1 and d2 spectral dimensions through the 2HG cross peaks of the ana-plastic astrocytoma biopsy are shown along axes of the 2DTOBSY. 2HGsignals are not present in the 2DTOBSY spectra from both primary glio-blastoma (Fig. 2B) and nontumor control (Fig. 2C) tissues, which wereboth wild-type IDH1. Notably, our 2HG findings from HR-MAS mea-surements are based on the single biopsy that had mutant IDH1.

In addition to 2HG, large qualitative and quantitative differences areeasily observed among different biopsy samples, most notably the pres-ence of lipids, L-fucose, and b-glucose, as well as the absence of glutathione(GSH) in glioblastoma; the increased GPC (glycerophosphocholine)–to–PC (phosphocholine) ratio in anaplastic astrocytoma comparedto nontumor control biopsy; and a decreased GPC–to–PC ratio in glio-blastoma compared to nontumor control biopsy (Fig. 2). Similar find-ings have been previously reported regarding increased lipids (19)and the presence of L-fucose in glioblastoma (20), and increased GPCin low-grade glioma versus increased PC in high-grade glioma (21–23).For comparison, 1D HR-MAS spectra acquired on the anaplastic

Do

Table 1. Brain biopsies from tumor or epileptic foci (n = 10) analyzed ex vivo with HR-MAS and LC-MS. Figure numbers are given for representativesubjects. wt, wild-type.

Patient ID

Diagnosis IDH1 mutation status

ceTranslationalMedicine.org 11 Jan

Age/gender

uary 2012 Vol 4 Issue 116 116ra

Figure

1

Anaplastic astrocytoma R132H 54/M 2A, S4

2

Anaplastic astrocytoma wt 31/F

3

Primary glioblastoma wt 23/F

4

Primary glioblastoma wt 40/F

5

Primary glioblastoma wt 50/M

6

Primary glioblastoma wt 52/F

7

Primary glioblastoma wt 52/M 2B

8

Nontumor wt 8/M

9

Nontumor wt 9/F

10

Nontumor wt 17/M 2C

4 3

R E S EARCH ART I C L E

www.Scien

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

astrocytoma biopsy are shown in fig. S4. Because no tissue is destroyedduringHR-MASmeasurements, further assays are possible, such as his-tology, genomics, or LC-MS, to characterize the tumors. LC-MS wasperformed on the same biopsies and 2HG levels were measured to be151.58 ng of 2HG per milligram of tissue (wet weight) for IDH1R132Hanaplastic astrocytoma, 2.39 ng/mg for wild-type IDH1 glioblastoma,and 1.79 ng/mg for wild-type IDH1 nontumor control. The 2HG levelin IDH1R132H anaplastic astrocytoma was 1.02 mmol/g, which is anorder of magnitude above the lower sensitivity limit (0.1 mmol/g) ofHR-MAS (24), whereas the wild-type IDH1 tissues had almost 100-foldless 2HG (0.01 to 0.015 mmol/g), concordant with previous results (7).The 2HG levels in wild-type IDH1 biopsies are <0.1 mmol/g (detectionthreshold of HR-MAS) and hence not visible in Fig. 2, B and C.

In vivo 2HG detection by spectral-editing and 2Dcorrelation MRSAfter confirming MRS detection of 2HG in biopsy tissue ex vivo, weperformedMRS spectroscopy in vivo in a separate set of human subjects(n = 10) (Table 2). The results obtained on biopsies were important toidentify the 2HG peaks that have the best chances to be detected in vivoand helped us select the appropriate in vivo methods. Two gliomapatients (n = 2) with known IDH1 mutations (R132C and R132H),as well as two control groups, including primary glioblastoma pa-tients without IDH1 mutations (n = 4) and healthy, nontumor volun-teers (n = 4), were investigated (Table 2). Single voxels were prescribedon the basis of fluid attenuation inversion recovery (FLAIR) image ab-normalities in tumor patients and magnetization prepared by rapidacquisition of gradient echo (MEMPRAGE) images in volunteers.2D LASER-COSY, 1D MEGA-LASER, and 1D LASER spectra wereacquired from the tumor patients and volunteers.

2DLASER-COSYresults fromonepatientwithanaplastic astrocytoma—confirmedby tumorDNAsequencing tohaveR132Cmutation of IDH1—are shown in Fig. 3A. A 27-cm3 voxel (3 × 3 × 3 cm3) was placed on theFLAIR images to include most of the solid tumor located in the sple-nium of the corpus callosum and the tail of the left hippocampus. TheHa-Hb cross peak of 2HG was present in the 2D LASER-COSYspectrumat 4.02/1.91 ppm (d2/d1), with d2 and d1 projectionswell abovethe baseline noise level (Fig. 3A). Similar 2HG projections can be ob-served in the phantom spectrum (fig. S2A). Cross peaks of several othermetabolites can be identified (Fig. 3A). Results from LCModel fittingof 1D LASER spectra are shown in fig. S5. Considering a basis set of

Fig. 2. HR-MAS spectra recorded at 14 T ex vivo on biopsy tissue from1 1

patients with and without IDH1 mutation. H- H 2D TOBSY spectra are

shown for all biopsies (the minimum contour levels were set five timesthe noise level). (A) For anaplastic astrocytoma biopsy tissue with IDH1R132mutation (n= 1), the spectra are shown in green-blue contours. The phan-tom is shown in red-yellow. Projections along d1 and d2 show the 2HG crosspeaks, outlined by a red rectangle. (B and C) Spectra for wild-type IDH1patients: primary glioblastoma (B) (n = 1) and nontumor (C) (n = 1). The re-gion where 2HG cross peaks would be expected is outlined by a red rec-tangle. For all 2D TOBSY brain spectra, several other metabolites can beidentified. Amino acids: alanine (Ala), aspartate (Asp), histidine (His), iso-leucine (Ile), leucine (Leu), lysine (Lys), proline (Pro), serine (Ser), andthreonine (Thr). Membrane phospholipid–related compounds: ethanol-amine (Etn), glycerol (Glr), glycerophosphocholine (GPC), glycero-phosphoethanolamine (GPE), phosphocholine (PC), and phosphoethanolamine(PE). Sugars: L-fucose (lFuc) and b-glucose (bGlc). Miscellaneous: glutathione(GSH), lipids (Lip), and taurine (Tau).

ceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ra4 4

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

spectra composed of the 20 most abundant metabolites in additionto 2HG, the fitting algorithm estimated the contribution of eachmetab-olite so that the computed spectrumoverlaps as best as possible with themeasured spectrum (fig. S5).

An example of 2D LASER-COSY from a primary glioblastoma pa-tient (wild-type IDH1 by tumorDNA sequencing) is shown (Fig. 3B). Aslightly bigger voxel (3.5 × 3.5 × 3.5 cm3) was chosen owing to the ex-tension of the tumor into the left occipital lobe. The 2D LASER-COSYdoes not contain any 2HG cross peak in the Ha region outlined by thegreen rectangle. Fitting methods applied to conventional 1D MRS (fig.S6) erroneously suggest the large presence of 2HG within confidencelimits for goodness of fit (16% Cramer-Rao lower bounds), and thatthe level of 2HG is higher than NAA or GPC. These latter metabolitesare both present in the 2D LASER-COSY spectrum; therefore, if thecomputed 1DMRS results in fig. S6 were true, the 2HG should be alsovisible in the 2D spectrum in Fig. 3B. This contradiction suggests that, inthis case, the fitted 1D MRS result represents a false positive. Fittingprograms, such as LCModel, assume that the composite spectrum canbe obtained by a unique combination of individual metabolite spectra.However, this is known to fail in vivo for somemetabolites because ofadverse combination of lower resolution and severe overlap of weakermetabolite signals by stronger metabolite signals. The most known ex-ample is erroneous GABA measurement by fitting conventional 1Dspectra (25). On the other hand, 2D LASER-COSY ismore in line withthe genetic analysis that showed no IDH1 mutations in this patient.

Figure 3C shows data froma healthy volunteerwithwild-type IDH1.A 27-cm3 voxel was placed in the white matter of the left occipital lobesimilar to the patient tumor positions. 2HG is absent, as expected, fromthe 2D LASER-COSY spectrum with the Ha cross peak region outlinedin green. 2D spectral quality is indicated by projections through the Gluand Gln cross peak [Glu + Gln, 3.75/2.12 (d2/d1) ppm]. The fitting ofthe 1DMRS from the healthy volunteer is shown in fig. S7. TheCramer-Rao lower bound (23%) for 2HG fit is only slightly above the acceptedlimit (20%) for goodness of fit. However, 2HG was not expected to befound in a healthy control (see also fig. S9 where the LCModel fits 2HGwith 17% Cramer-Rao lower bounds in the healthy contralateral hem-isphere of the glioblastoma patient).

Results obtained with the spectral-editing 1D MEGA-LASER se-quence are presented in Fig. 4. Overlay of spectra acquired from tumorand the healthy contralateral side in a secondary glioblastoma patient

www.Scien

with IDH1R132H mutation is shown in Fig. 4A. The Ha multiplet signalof 2HG around 4.02 ppm is found only in the tumor spectrum and noton the healthy side. The multiplets of Glu and Gln (Glu + Gln), andGABA [+macromolecules (MM)] are present in both voxels. LCModelfitting of the corresponding nonedited 1D LASER spectra showed 2HGin tumor (Cramer-Rao lower bounds 15%) (fig. S8) andalso in thehealthyvoxel (Cramer-Rao lower bounds 17%) (fig. S9), contrary to expectations.

Spectra from the control primary glioblastoma patient (wild-typeIDH1) (Fig. 4B) and from the healthy volunteer (wild-type IDH1)(Fig. 4C) do not contain any 2HG, but they do show Glu + Gln andGABA + MM peaks. The Glu + Gln peaks in tumor voxels seem tobe shifted slightly (0.01 ppm) upfield compared to healthy side spectra(Fig. 4, A and B). The shift can be caused by different pH conditions intumors compared to healthy brain tissue, and by different Glu and Glnrelative contributions. There is no shift for Glu + Gln peaks betweenright and left sides in the healthy volunteer (Fig. 4C). No shift wasobserved for the GABA + MM peaks in any subject (Fig. 4).

Quantification of 2HG from in vivo spectral-editing and2D correlation MRSQuantitative analysis of 2D LASER-COSY, 1DMEGA-LASER, and 1DLASER spectra was performed using the ratio of 2HG to the sum ofGlu + Gln for reasons outlined in Materials and Methods. Volumes ofcross peaks at 4.02/1.91 ppm for 2HG and 3.75/2.12 ppm for Glu + Glnwere used to calculate the 2HG/(Glu + Gln) ratio from 2D LASER-COSY (Fig. 5A). Areas of peaks at 4.02 ppm for 2HG and 3.75 ppm forGlu + Gln were used to estimate the 2HG/(Gln + Glu) ratio from 1DMEGA-LASER (Fig. 5B). For 1D LASER, the values fitted by LCModel(Fig. 5, A and B) were used to calculate the ratio.

The 2HG/(Glu + Gln) ratios are plotted for phantoms (n = 2),mutant IDH1R132H glioma patients (n = 2), wild-type IDH1 glioblasto-ma patients (n = 4), and healthy volunteers (n = 4) (Fig. 5). In the caseof phantoms, there is very good agreement between 2D correlation(LASER-COSY) MRS, 1D spectral-editing (MEGA-LASER) MRS, and1D conventional (LASER)MRS for the 2HG/(Glu + Gln) ratio, whichwas close to 0.4, as expected from their respective concentrations:2HG (3 mM), Glu (7.5 mM), and Gln (0 mM). Because of similar spinsystems, the ratio of cross peak volumes was determined by their con-centrations (assuming similar T2 and T1 times), without the need tocorrect for number of protons and buildup rates.

Table 2. Human subjects (n = 10) scanned with in vivo MRS. Figure numbers are given for representative subjects. wt, wild-type.

Patient ID

Diagnosis IDH1 mutation status

ceTranslationalMedicine.org 11 J

Age/gender

anuary 2012 Vol 4 Issue 116 11

Figure

1

Anaplastic astrocytoma R132C 39/F 3A, S5

2

Secondary glioblastoma R132H 41/F 4A, S8, S9

3

Primary glioblastoma wt 40/M 4B

4

Primary glioblastoma wt 44/M

5

Primary glioblastoma wt 52/F 3B, S6

6

Primary glioblastoma wt 70/F

7

Healthy volunteer wt 24/F

8

Healthy volunteer wt 30/F 3C, S7

9

Healthy volunteer wt 32/M 4C

10

Healthy volunteer wt 34/M

6ra4 5

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

In the case of mutant IDH1R132 glioma patients, estimation of2HG/(Glu + Gln) ratio showed slight differences, which, however, arenot statistically significant (P=0.28). 2DLASER-COSYand1DMEGA-

www.ScienceTranslationalMedicine.org 1

LASER found an average ratio of 1.27,whereas LCModel fitting estimated anaverage ratio of 1.11 (Fig. 5). In the wild-type IDH1 primary glioblastoma andnontumor controls, there was a signif-icant difference (P = 0.03) between the2HG/(Glu + Gln) ratios obtained byLCModel fitting (ratios 0.33 and 0.57),and 2D LASER-COSY (ratio 0.04) and 1DMEGA-LASER (ratio 0.03), respectively.

DISCUSSION

The discovery that mutated IDH1/2 ingliomas can be correlated with survivalbenefit (3) has generated interest in usingthis mutation for diagnostic and prog-nostic purposes. 2HG is ametabolite thataccumulates in humangliomas that harborIDH1mutations. Here, we preliminarilyshow that 2HGcanbe detected unambig-uously and noninvasively by localized 2Dcorrelation and 1D spectral-editing MRSin patients with mutated IDH1.

In vivo MRS detection of 2HG ingliomas has been suggested previously(7). Our results show that 2D LASER-COSY and 1D MEGA-LASER can reli-ably identify 2HG. The sensitivity of 2D

LASER-COSY was about 2 mM (or 10 mg) for a 3 × 3 × 3 cm3 voxel,using a 13-min in vivo acquisition time and a minimum SNR of 5. Thesame sensitivity can be achieved by 1D MEGA-LASER in 5-min scan

Fig. 3. 2D LASER-COSY spectra in vivo in

human subjects at 3 T. (A) An anaplastic as-trocytoma patient with IDH1R132C. The 2DLASER-COSY shows at 4.02/1.91 ppm theHa-Hb cross peak of 2HG. Projections alongboth spectral dimensions through2HGcrosspeak indicate the SNR and spectral quality.The single voxel (3 × 3 × 3 cm3, red rectan-gle) was placed on the FLAIR images to in-clude most of the tumor abnormality. (B) Aprimary glioblastoma patient (wt-IDH1).The 2D LASER-COSY does not contain any2HG cross peak in the Ha-Hb region outlinedby the green rectangle. Projections throughGlu + Gln cross peak indicate spectral qual-ity. The single voxel (3.5 × 3.5 × 3.5 cm3, redrectangle) was placed on the FLAIR imagesto include most of the tumor abnormality.(C) A healthy volunteer (wt-IDH1). 2HG isnot found in the Ha-Hb region of 2D LASER-COSY outlined by the green rectangle. Pro-jections through Glu + Gln indicate spectralquality. The single voxel (3 × 3 × 3 cm3, redrectangle) was placed on the MEMPRAGEimages in the white matter of the occipitallobe, in a region similar to tumor locationsfrom patients in (A) and (B).

1 January 2012 Vol 4 Issue 116 116ra4 6

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

for the same voxel size. This is sufficient for the range (5 to 35 mM) of2HG concentrations reported in IDH1-mutant tumors. Although thevoxels used seem to be pretty large, several aspects besides maximizingsensitivity may justify this choice. First, gliomas are very infiltrative tu-mors with ill-definedmargins, and active tumor exceeds the contours ofthe T1-weighted postcontrast images, which are mostly used to reporttumor diameters or volumes. Second, IDH1 mutations seem to beuniformly expressed in tumorswhenpresent (26), so a large tumor vol-

www.ScienceTranslationalMedicine.org

ume could be included in the voxel. Fi-nally, our method can separate or removethe contribution of normal metabolite;hence, the inclusion of healthy tissue, whichwe showed does not contain 2HG, does notalter 2HG estimation. Further improve-ments in spatial resolution and multivoxelacquisitions of 2D LASER-COSY (27) orspectral-editing MRS (28) are possible. Inaddition, the sensitivity of the ex vivo HR-MAS approach is 1 mM and may thereforebe used as a nondestructive method formore detailed metabolite profiling in tu-mor samples.

Relative quantification of in vivo MRSdata indicated that a 2HG/(Glu + Gln)ratio of >1 could be specific for IDH1mu-tations. Moreover, existing data indicatethat 2HG and Glu might be inversely pro-portional: a slight decrease of Glu togetherwith a large increase of 2HG in IDH1R132H(9) compared with a slight increase of Glu(20) with virtually undetectable 2HG (7) inwild-type IDH1 gliomas. These results sug-gest that the 2HG/(Glu + Gln) ratio mighthave increased dynamic range for detectingIDH1mutations compared with either me-tabolite alone.

Comparing 2D correlation and 1Dspectral-editing MRS, each method hasits own strengths and limitations. For ex-ample, all metabolites are preserved andidentified by two well-defined chemicalshifts in 2D COSY, whereas in 1D spectralediting, there is only one well-definedchemical shift in addition to a range ofpossible chemical shifts given by the band-width of the selective pulse (fig. S1). Con-versely, spectral-editing experiments areeasier to run and may require shorter scantimes or smaller voxel sizes to detect thesame concentration.

In addition to using 2HG as a bio-marker, there is mounting interest in de-ciphering the biological mechanisms thatlink IDH mutations, 2HG production,and tumorigenesis. 2HG might act as anoncometabolite by competitive inhibitionofa-ketoglutarate–dependentdioxygenases(29). This includes inhibition of histone

demethylases and 5-methlycytosine hydroxylases, leading to genome-wide alterations in histone and DNAmethylation, as well as inhibitionof hydroxylases resulting in up-regulation of hypoxia-inducible fac-tor 1 (HIF-1) (30). Hence, a large interest exists from pharmacologicalcompanies and research groups to develop inhibitors of mutatedIDH1R132. The ability to objectively and noninvasively follow the ef-fects of these compounds in animals and patients is a prerequisite forsuccessful drug development.

Fig. 4. 1D MEGA-LASER spectra in vivo in human subjects at 3 T. In all subjects, two voxels (3 × 3 × 3 cm3

each) were placed in both brain hemispheres, symmetrically from the middle line. (A) A secondary glio-

blastoma patient with IDH1R132H mutation. (B and C) The spectra from subjects with wt-IDH1: primaryglioblastoma (B) and healthy volunteer (C). MM denotes contamination of GABA signal with macro-molecule signal.

11 January 2012 Vol 4 Issue 116 116ra4 7

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

The importance of a reliablemethod for a novel noninvasivemethodof detecting 2HG in vivo is underscored by the fact that no report existsabout increased D-2HG in the blood, cerebrospinal fluid, or urine of gli-oma patients with IDH1mutations. This situation is different from hy-droxyglutaric aciduria metabolic disorders, which show high levels ofL-2HG in body fluids. Therefore, other than tumor biopsy, no assay cur-rently exists to probe 2HG in IDH-mutated gliomas. Although a biopsymight be necessary for the initial diagnosis,multiple serial brain biopsiesare generally not feasible. Moreover, MRS methods could map 2HGdistribution and identify 2HG hot spots, guiding biopsy proceduresto increase the chances for correct typing of the tumor, because it isknown that biopsy based on conventional computed tomography(CT) or magnetic resonance imaging (MRI) is suboptimal and subjectto undergrading (26).

The in vivo 2D correlation and 1D spectral-editing MRS methodsthat we demonstrated can be repeated noninvasively, without harmfuleffects to patients, and might facilitate preclinical or clinical studies ofnew therapies, as well as assist with initial diagnostic workup.With fur-ther validation in humans, this approach could even allow moleculartyping of IDH-mutant tumor using MRI investigations, which are al-ready included in most patients’ diagnostic workup, resulting in cost-effective and rapid genotyping of IDH mutations.

MATERIALS AND METHODS

Biopsy collection for HR-MAS and LC-MSBiopsies (n = 10, Table 1) were collected at the time of surgery, snap-frozen in liquid nitrogen, and subsequently stored in a freezer at −80°Cuntil HR-MAS measurement. Informed consent was obtained beforesurgery for biopsy collection. Biopsies were obtained from seven gliomapatients: five primary glioblastoma (wild-type IDH1) and two anaplasticastrocytoma (one patient with IDH1R132H and one patient withwild-typeIDH1). In addition, nontumor healthy control biopsies were obtainedfrom three patients who had been surgically treated for epilepsy.

Selection of human subjectsPatients and healthy volunteers listed in Table 2 were scanned withinformed consent approved by the Internal Review Board at our insti-

www.ScienceTranslationalMedicine.org 1

tution. Patients were diagnosed as part ofstandard diagnostic practice by neuro-pathologistswho examined formalin-fixed,paraffin-embedded samples from thesubjects that had been stained with hema-toxylin and eosin (H&E). In total, 10subjects (2 mutant IDH1R132 glioma pa-tients, 4 wild-type IDH1 glioma patients,and 4 wild-type IDH1 healthy volunteers)were scanned with in vivo MRS.

Acquisition of in vivoMR spectroscopyAll in vivoMR scans were performed on3 T Tim Trio scanners (Siemens) with ahead 32-channel phased array for receiveand body radio-frequency coil for trans-mit. Single voxel spectroscopy was per-formed with recently optimized 1D LASER(31) and the 2D LASER-COSY (16) se-

quences. In addition, a newly designed 1D MEGA-LASER was usedfor spectral editing (fig. S1). The same LASER module was used for lo-calization in all sequences because of sharp excitationmargins,minimalchemical shift displacement error, reduced lineshape modulation,insensitivity to B1 inhomogeneity or flip-angle errors. Low-power,gradient offset independent adiabaticity wurst modulated [GOIA-W(16,4) (31)] pulseswere usedwith 3.5-msduration, 20-kHzbandwidth,and 0.817-kHzmaximum B1 field amplitude. Typical voxel sizeswere27 cm3 (3 × 3×3 cm3) or 42.8 cm3 (3.5 × 3.5 × 3.5 cm3), in the case of largetumors. A repetition time (TR) of 1.5 s was used for all acquisitions.

For 1D LASER and 2D LASER-COSY, a TE of 45 ms was used. 1DLASER spectra were collected with 128 averages (acquisition time of3.2 min), and the 2D LASER-COSY spectra were acquired with 64 t1 in-crements (10-ppm d1 spectral window), 8 averages per t1 increment, and4 dummy scans for the first t1 (acquisition time of 12.8 min). The d2 di-rectly acquired spectral dimension was set to 1.25 kHz (~10 ppm), andthe free induction decay (FID) had 512 points in all experiments.

The 1DMEGA-LASER spectra were acquiredwith TE of 75ms, and200 averages were collected (acquisition time of 5min). In all sequences,water suppression was performed with WET scheme (32). Automaticshimming of the single voxels was performed with FASTESTMAP (33)to ensure linewidths of 6 to 12 Hz in human subjects. Anatomical MRimages were collected to guide the position ofMRS voxels. For patients,the preferred modality was axial fluid-attenuated inversion recovery(FLAIR) acquired with 10-s TR, 70-ms TE, 5-mm slice thickness (1-mmgap), 0.6 × 0.45mm2 in-plane resolution, 23 slices, and 384× 512matrix(imaging time, 3.03 min).

For healthy volunteers, a multiecho MEMPRAGE (34) volumetricacquisition was performed, with 1-mm isotropic voxels, TR = 2.53 s,TE1/TE2/TE2/TE3/TE4 = 1.64/3.5/5.36/7.22 ms, inversion time TI =1.2 s (imaging time, 6.1 min). Voxels on healthy volunteers were placedin similar regions as observed on patients to match coil sensitivityprofile and regional metabolic differences.

Processing, analysis, and quantification of in vivoMR spectroscopyRawdata were exported from the Siemens scanners for subsequent pro-cessing and analysis. The 1D LASER data (FID) were processed and

Fig. 5. Signal intensity ratios of 2HG to the sum of glutamate and glutamine (Glu + Gln). (A and B) Ratiosare shown for all phantom and in vivo human spectra: 2D correlation MRS (LASER-COSY) (A), 1D spectral-

editedMRS (MEGA-LASER) (B), and 1D conventional MRS (LASER) (A and B). Ratios are given as averages ±1 SD (n = 2 for phantoms and IDH1R132 patients; n = 4 for wt-IDH1 subjects).

1 January 2012 Vol 4 Issue 116 116ra4 8

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rg

quantified with LCModel (12) using a GAMMA-simulated basis setfor LASER (Supplementary Methods). Fourier transform (FT), phasecorrection, and baseline correction were performed as part of theLCModel processing. For 1DMEGA-LASER data, the FT, phase cor-rection, baseline correction, and line fitting were done in jMRUI (13).For 2DLASER-COSY, the FIDs of all 64 t1 increments were imported inMatlab (TheMathWorks). Processing steps included (i) FT along t2, (ii)linear prediction forward to 128 points in t1 using the ITMPMmethod(35), (iii) FT along t1, and (iv) square-sine window function in bothd1 and d2 dimensions to improve cross peaks and reduce diagonal ring-ing and baseline distortion. The 2D spectra were displayed as contourlevels in magnitude mode, with the first contour level chosen five timesthe floor noise level as estimated from SD of noise floor in a signal-freespectral region (0.5 to 0 ppm/0.5 to 0 ppm, d1/d2).

Aminimum SNR of 5 was considered for reliable identification ofcross peaks from the noise. This was decided on the basis of the seriesof 2HG phantoms. At 1 mM, the Ha cross peaks had an SNR of ~2.5,which was considered insufficient to distinguish them from noise.Metabolites were assigned on the basis of the literature (17, 18, 36)values for their nuclear magnetic resonance (NMR) parameters, andcross peak volumes were integrated inMatlab. For quantification, the2HG/(Glu + Gln) ratio was chosen for the following reasons: (i) 2HG,Glu, andGln have a similar five-spin system; hence, the buildup of theirCOSY cross peaks and spectral-edited peaks is similar; (ii) Glu and Glnare largely present in both tumors and healthy brain, yielding clearlyresolved cross peaks and spectral-edited peaks; (iii) the absolute quan-tification based on internal water cannot be used in tumors where watercontent varies largely; and (iv) 2HG quantification relative to Glu andGln is preferred over quantification relative to creatine (37), because cre-atine does not have cross peaks or spectral-edited peaks, and may varywith disease. Thus, 2HG/(Glu +Gln) ratio enabled direct comparison ofthe three in vivoMRSmethods. Additional details are given in the Sup-plementary Material.

Dow

nloa

ded

from

SUPPLEMENTARY MATERIAL

www.sciencetranslationalmedicine.org/cgi/content/full/4/116/116ra4/DC1Materials and MethodsFig. S1. Pulse sequence diagram for the 1D MEGA-LASER spectral editing experiment.Fig. S2. Phantom experiments and simulations for 2D LASER-COSY and 1D LASER at 3 T.Fig. S3. Optimization of the 1D MEGA-LASER spectral editing on phantoms at 3 T.Fig. S4. 1D HR-MAS spectra recorded at 14 T and 3-kHz MAS on a biopsy sample from onepatient with R132H IDH1 anaplastic astrocytoma.Fig. S5. LCModel fitting of the 1D LASER spectrum from the R132C IDH1 anaplastic astrocytomapatient.Fig. S6. LCModel fitting of the 1D LASER spectrum from a wild-type IDH1 primary glioblastomapatient.Fig. S7. LCModel fitting of the 1D LASER spectrum from a wild-type IDH1 healthy volunteer.Fig. S8. LCModel fitting of the 1D LASER spectrum from the tumor voxel of the R132H IDH1secondary glioblastoma patient.Fig. S9. LCModel fitting of the 1D LASER spectrum from the healthy side voxel of the R132HIDH1 secondary glioblastoma patient.References

REFERENCES AND NOTES

1. D. W. Parsons, S. Jones, X. Zhang, J. C. H. Lin, R. J. Leary, P. Angenendt, P. Mankoo, H. Carter,I. M. Siu, G. L. Gallia, A. Olivi, R. McLendon, B. A. Rasheed, S. Keir, T. Nikolskaya, Y. Nikolsky,D. A. Busam, H. Tekleab, L. A. Diaz Jr., J. Hartigan, D. R. Smith, R. L. Strausberg, S. K. N. Marie,S. M. O. Shinjo, H. Yan, G. J. Riggins, D. D. Bigner, R. Karchin, N. Papadopoulos, G. Parmigiani,

www.Scien

B. Vogelstein, V. E. Velculescu, K. W. Kinzler, An integrated genomic analysis of human glio-blastoma multiforme. Science 321, 1807–1812 (2008).

2. Cancer Genome Atlas Research Network, Comprehensive genomic characterization defineshuman glioblastoma genes and core pathways. Nature 455, 1061–1068 (2008).

3. H. Yan, D. W. Parsons, G. Jin, R. McLendon, B. A. Rasheed, W. Yuan, I. Kos, I. Batinic-Haberle,S. Jones, G. J. Riggins, H. Friedman, A. Friedman, D. Reardon, J. Herndon, K. W. Kinzler,V. E. Velculescu, B. Vogelstein, D. D. Bigner, IDH1 and IDH2 mutations in gliomas. N. Engl.J. Med. 360, 765–773 (2009).

4. N. K. Kloosterhof, L. B. Bralten, H. J. Dubbink, P. J. French, M. J. van den Bent, Isocitratedehydrogenase-1 mutations: A fundamentally new understanding of diffuse glioma? LancetOncol. 12, 83–91 (2011).

5. F. E. Bleeker, N. A. Atai, S. Lamba, A. Jonker, D. Rijkeboer, K. S. Bosch, W. Tigchelaar, D. Troost,W. P. Vandertop, A. Bardelli, C. J. Van Noorden, The prognostic IDH1R132mutation is associatedwith reduced NADP+-dependent IDH activity in glioblastoma. Acta Neuropathol. 119, 487–494(2010).

6. E. R. Mardis, L. Ding, D. J. Dooling, D. E. Larson, M. D. McLellan, K. Chen, D. C. Koboldt,R. S. Fulton, K. D. Delehaunty, S. D. McGrath, L. A. Fulton, D. P. Locke, V. J. Magrini, R. M. Abbott,T. L. Vickery, J. S. Reed, J. S. Robinson, T. Wylie, S. M. Smith, L. Carmichael, J. M. Eldred,C. C. Harris, J. Walker, J. B. Peck, F. Du, A. F. Dukes, G. E. Sanderson, A. M. Brummett, E. Clark,J. F. McMichael, R. J. Meyer, J. K. Schindler, C. S. Pohl, J. W. Wallis, X. Shi, L. Lin, H. Schmidt,Y. Tang, C. Haipek, M. E. Wiechert, J. V. Ivy, J. Kalicki, G. Elliott, R. E. Ries, J. E. Payton,P. Westervelt, M. H. Tomasson, M. A. Watson, J. Baty, S. Heath, W. D. Shannon, R. Nagarajan,D. C. Link, M. J. Walter, T. A. Graubert, J. F. DiPersio, R. K. Wilson, T. J. Ley, Recurringmutations found by sequencing an acute myeloid leukemia genome. N. Engl. J. Med.361, 1058–1066 (2009).

7. L. Dang, D. W. White, S. Gross, B. D. Bennett, M. A. Bittinger, E. M. Driggers, V. R. Fantin,H. G. Jang, S. Jin, M. C. Keenan, K. M. Marks, R. M. Prins, P. S. Ward, K. E. Yen, L. M. Liau,J. D. Rabinowitz, L. C. Cantley, C. B. Thompson, M. G. Vander Heiden, S. M. Su, Cancer-associatedIDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739–744 (2009).

8. K. E. Yen, M. A. Bittinger, S. M. Su, V. R. Fantin, Cancer-associated IDH mutations: Biomarkerand therapeutic opportunities. Oncogene 29, 6409–6417 (2010).

9. Z. J. Reitman, G. Jin, E. D. Karoly, I. Spasojevic, J. Yang, K. W. Kinzler, Y. He, D. D. Bigner,B. Vogelstein, H. Yan, Profiling the effects of isocitrate dehydrogenase 1 and 2 mutationson the cellular metabolome. Proc. Natl. Acad. Sci. U.S.A. 108, 3270–3275 (2011).

10. R. N. Sener, L-2 hydroxyglutaric aciduria: Proton magnetic resonance spectroscopy anddiffusion magnetic resonance imaging findings. J. Comput. Assist. Tomogr. 27, 38–43(2003).

11. S. M. Goffette, T. P. Duprez, M. C. Nassogne, M. F. Vincent, C. Jakobs, C. J. Sindic, L-2-Hydroxyglutaric aciduria: Clinical, genetic, and brain MRI characteristics in two adultsisters. Eur. J. Neurol. 13, 499–504 (2006).

12. S. W. Provencher, Estimation of metabolite concentrations from localized in vivo protonNMR spectra. Magn. Reson. Med. 30, 672–679 (1993).

13. A. Naressi, C. Couturier, J. M. Devos, M. Janssen, C. Mangeat, R. de Beer, D. Graveron-Demilly,Java-based graphical user interface for the MRUI quantitation package. MAGMA 12, 141–152(2001).

14. M. Mescher, H. Merkle, J. Kirsch, M. Garwood, R. Gruetter, Simultaneous in vivo spectralediting and water suppression. NMR Biomed. 11, 266–272 (1998).

15. R. R. Ernst, 2-Dimensional spectroscopy. Chimia 29, 179–183 (1975).16. O. C. Andronesi, S. Ramadan, C. E. Mountford, A. G. Sorensen, Low-power adiabatic sequences

for in vivo localized two-dimensional chemical shift correlated MR spectroscopy.Magn. Reson.Med. 64, 1542–1556 (2010).

17. D. Bal, A. Gryff-Keller, 1H and 13C NMR study of 2-hydroxyglutaric acid and its lactone.Magn. Reson. Chem. 40, 533–536 (2002).

18. V. Govindaraju, K. Young, A. A. Maudsley, Proton NMR chemical shifts and coupling con-stants for brain metabolites. NMR Biomed. 13, 129–153 (2000).

19. K. S. Opstad, B. A. Bell, J. R. Griffiths, F. A. Howe, Toward accurate quantification of metab-olites, lipids, and macromolecules in HRMAS spectra of human brain tumor biopsies usingLCModel. Magn. Reson. Med. 60, 1237–1242 (2008).

20. S. Ramadan, O. C. Andronesi, P. Stanwell, A. P. Lin, A. G. Sorensen, C. E. Mountford, Use ofin vivo two-dimensional MR spectroscopy to compare the biochemistry of the humanbrain to that of glioblastoma. Radiology 259, 540–549 (2011).

21. J. P. Usenius, P. Vainio, J. Hernesniemi, R. A. Kauppinen, Choline-containing compounds inhuman astrocytomas studied by 1H NMR spectroscopy in vivo and in vitro. J. Neurochem.63, 1538–1543 (1994).

22. J. Sabatier, V. Gilard, M. Malet-Martino, J. P. Ranjeva, C. Terral, S. Breil, M. B. Delisle, C. Manelfe,M. Tremoulet, I. Berry, Characterization of choline compounds with in vitro 1H magneticresonance spectroscopy for the discrimination of primary brain tumors. Invest. Radiol. 34,230–235 (1999).

23. V. Righi, J. M. Roda, J. Paz, A. Mucci, V. Tugnoli, G. Rodriguez-Tarduchy, L. Barrios, L. Schenetti,S. Cerdán, M. L. García-Martín, 1H HR-MAS and genomic analysis of human tumor biopsiesdiscriminate between high and low grade astrocytomas. NMR Biomed. 22, 629–637 (2009).

ceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ra4 9

R E S EARCH ART I C L E

on

Mar

ch 9

, 201

2ce

mag

.org

24. O. Beckonert, M. Coen, H. C. Keun, Y. Wang, T. M. D. Ebbels, E. Holmes, J. C. Lindon,J. K. Nicholson, High-resolution magic-angle-spinning NMR spectroscopy for metabolicprofiling of intact tissues. Nat. Protoc. 5, 1019–1032 (2010).

25. J. Near, R. Simpson, P. Cowen, P. Jezzard, Efficient g-aminobutyric acid editing at 3T withoutmacromolecule contamination: MEGA-SPECIAL. NMR Biomed. 10.1002/nbm.1688 (2011).

26. M. Kunz, N. Thon, S. Eigenbrod, C. Hartmann, R. Egensperger, J. Herms, J. Geisler, C. la Fougere,J. Lutz, J. Linn, S. Kreth, A. von Deimling, J. C. Tonn, H. A. Kretzschmar, G. Pöpperl, F. W. Kreth,Hot spots in dynamic 18FET-PET delineate malignant tumor parts within suspected WHOgrade II gliomas. Neuro Oncol. 13, 307–316 (2011).

27. O. C. Andronesi, B. A. Gagoski, E. Adalsteinsson, A. G. Sorensen, Correlation chemical shiftimaging with low-power adiabatic pulses and constant-density spiral trajectories. NMRBiomed. 10.1002/nbm.1730 (2011).

28. H. Zhu, R. A. Edden, R. Ouwerkerk, P. B. Barker, High resolution spectroscopic imaging ofGABA at 3 Tesla. Magn. Reson. Med. 65, 603–609 (2011).

29. W. Xu, H. Yang, Y. Liu, Y. Yang, P. Wang, S. H. Kim, S. Ito, C. Yang, P. Wang, M. T. Xiao, L. X. Liu,W. Q. Jiang, J. Liu, J. Y. Zhang, B. Wang, S. Frye, Y. Zhang, Y. H. Xu, Q. Y. Lei, K. L. Guan, S. M. Zhao,Y. Xiong, Oncometabolite 2-hydroxyglutarate is a competitive inhibitor of a-ketoglutarate-dependent dioxygenases. Cancer Cell 19, 17–30 (2011).

30. S. C. Williams, M. A. Karajannis, L. Chiriboga, J. G. Golfinos, A. von Deimling, D. Zagzag,R132H-mutation of isocitrate dehydrogenase-1 is not sufficient for HIF-1a upregulationin adult glioma. Acta Neuropathol. 121, 279–281 (2011).

31. O. C. Andronesi, S. Ramadan, E. M. Ratai, D. Jennings, C. E. Mountford, A. G. Sorensen,Spectroscopic imaging with improved gradient modulated constant adiabaticity pulseson high-field clinical scanners. J. Magn. Reson. 203, 283–293 (2010).

32. R. J. Ogg, P. B. Kingsley, J. S. Taylor, WET, a T1- and B1-insensitive water-suppressionmethod for in vivo localized 1H NMR spectroscopy. J. Magn. Reson. B 104, 1–10 (1994).

33. R. Gruetter, I. Tkác, Field mapping without reference scan using asymmetric echo-planartechniques. Magn. Reson. Med. 43, 319–323 (2000).

34. A. J. van der Kouwe, T. Benner, D. H. Salat, B. Fischl, Brain morphometry with multiechoMPRAGE. Neuroimage 40, 559–569 (2008).

35. Y. Y. Lin, P. Hodgkinson, M. Ernst, A. Pines, A novel detection–estimation scheme for noisyNMR signals: Applications to delayed acquisition data. J. Magn. Reson. 128, 30–41 (1997).

36. M. C. Martínez-Bisbal, L. Martí-Bonmatí, J. Piquer, A. Revert, P. Ferrer, J. L. Llácer, M. Piotto,O. Assemat, B. Celda, 1H and 13C HR-MAS spectroscopy of intact biopsy samples ex vivoand in vivo 1H MRS study of human high grade gliomas. NMR Biomed. 17, 191–205 (2004).

www.Scienc

37. M. A. Thomas, N. Hattori, M. Umeda, T. Sawada, S. Naruse, Evaluation of two-dimensionalL-COSY and JPRESS using a 3 T MRI scanner: From phantoms to human brain in vivo.NMR Biomed. 16, 245–251 (2003).

Acknowledgments: We would like to acknowledge J. A. Iafrate and D. N. Louis from the De-partment of Pathology of Massachusetts General Hospital for assistance with SNaPshot anal-ysis for IDH mutation and useful comments on our results. We also acknowledge P. M. Blackfrom the Department of Neurosurgery of Brigham and Women’s Hospital for access to biopsiesthat were analyzed by ex vivo HR-MAS spectroscopy. We thank M. Malgorzata and M. Garwoodfrom Center for Magnetic Resonance Research at University of Minnesota for help with LCModelbasis set for LASER excitation. Funding: This work was funded by grants from NIH (R01 1200-206456, S10RR013026, S10RR021110, and S10RR023401). O.C.A. was also supported by a KL2Medical Research Investigator Training (MeRIT) award from Harvard Catalyst, The Harvard Clinicaland Translational Science Center (NIH Award #KL2 RR 025757). Author contributions: O.C.A.provided conceptual design, obtained measurements, analyzed the data, and drafted the man-uscript. G.S.K. obtained experimental measurements. E.G. and T.B. recruited patients and providedclinical guidance and manuscript review. A.A.T. provided support for biopsy measurements andmanuscript review. V.R.F. performed LC-MS measurements, literature review, and manuscriptediting. M.G.V.H. provided expertise on IDH1mutations and 2HG, literature review, and manuscriptediting. A.G.S. initiated the project, funding support, and manuscript review. Competing interests:The content is solely the responsibility of the authors and does not necessarily represent theofficial views of Harvard Catalyst, Harvard University and its affiliated academic health carecenters, the National Center for Research Resources, or the NIH. O.C.A. and A.G.S. have appliedfor a patent for the 2D COSY-LASER method that is used in the paper, U.S. Patent ApplicationSerial No. 13/237,799.

Submitted 23 May 2011Accepted 23 November 2011Published 11 January 201210.1126/scitranslmed.3002693

Citation: O. C. Andronesi, G. S. Kim, E. Gerstner, T. Batchelor, A. A. Tzika, V. R. Fantin,M. G. Vander Heiden, A. G. Sorensen, Detection of 2-hydroxyglutarate in IDH-mutatedglioma patients by in vivo spectral-editing and 2D correlation magnetic resonancespectroscopy. Sci. Transl. Med. 4, 116ra4 (2012).

n

eTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ra4 10

stm

.sci

eD

ownl

oade

d fr

om

P E R S P E C T I V E

www.ScienceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ps1 1

METABOLISM, UNHINGED

Whether mutations in metabolic pathways contribute to pathogenesis of cancer is still controversial. In the 1920s, Warburg � rst raised the hypothesis that aberrant metabo-lism, speci� cally cellular respiration, was the origin of cancer. � e relatively recent discov-ery that the genes encoding isocitrate dehy-drogenase enzyme isoforms 1 (IDH1) and 2 (IDH2)—cytoplasmic and mitochondrial, respectively—are recurrently mutated in can-cer represents one of the biggest advances in cancer biology in the era of high-throughput sequencing. � ese metabolic enzymes, which are essential to cellular respiration, have sub-sequently unveiled an amazing complex bi-ology associated with the dysregulation and functional consequences of metabolites in cancer. In this issue of Science Translational

Medicine, two studies show that IDH muta-tion status could be assessed noninvasively, which would give preoperative diagnostic and prognostic clues essential in glioma management (1, 2).

In 2008, large-scale sequencing e� orts in more than 200 glioblastoma tumors re-discovered IDH1 and IDH2 as recurrently mutated in a minority of primary gliomas and a large majority of secondary gliomas (3, 4). � ese mutations occur in the highly conserved residue arginine 132 (R132) in the active site of IDH1 and mostly concern arginine to histidine (R→H) amino acid substitutions (codon CGT→CAT change).

Less frequently, CGT→TGT (substitution of arginine to cysteine), CGT→GGT (argi-nine to glycine), and CGT→AGT (arginine to serine) mutations are present. Extraordi-narily high rates of spontaneous mutations in IDH1 have been reported in adult dif-fuse gliomas of World Health Organization (WHO) grades II and III of astrocytic and oligodendroglial lineages (4–6) and, in lower frequency, mutations in the gene encoding mitochondrial nicotinamide adenine di-nucleotide phosphate (NADP)–dependent isocitrate dehydrogenase (IDH2) (4). IDH1 R132 mutations are present in 55 to 80% of grade II and III oligodendrogliomas and as-trocytomas, but rarely in primary glioblas-toma or in a variety of other primary brain tumors, including pilocytic astrocytoma (7). Although TP53 mutations and 1p/19q loss are mutually exclusive in glioma, IDH1 mu-tations are present in TP53-mutated and in 1p/19q-codeleted tumors. � is, in combina-tion with the � ndings in patients with multi-ple biopsies in which there were no cases that acquired IDH1 mutations a� er the acquisi-tion of either TP53 mutations or combined 1p/19q loss, suggests that the occurrence of IDH1 mutations is an early event in the tu-morigenesis of di� use glioma (6, 8).

IDH enzymes catalyze the oxidative decarboxylation of isocitrate to produce α-ketoglutarate (α-KG)—a reaction that is accompanied by the reduction of NADP+ to NADPH. Recently, it has been shown that tumor-derived IDH1 mutations impair the enzyme’s a� nity for its substrate and dominantly inhibit wild-type IDH1 activity through the formation of catalytically inac-tive heterodimers (9). � us, mutant IDH1, by reducing formation and increasing con-sumption of α-KG, would lead to an in-

creased level of hypoxia-inducible factor sub-unit 1α (HIF-1α), a transcription factor that facilitates tumor growth and for which stabil-ity is regulated by α-KG (9). In a recent study, HIF-1α and its targets were found to be up-regulated in cultured cells harboring IDH1 mutation. Because IDH1 has also been shown to participate in a glucose-sensing pathway along with pseudo-hypoxia induction, IDH mutations would favor a metabolic shi� in tumoral cells toward glycolysis and hence increased glucose consumption (10, 11). A metabolic imaging study explored glycolytic metabolism in human cancer, showing that 18-� uorodeoxyglucose (FDG) uptake corre-lated with glioma malignancy (12) and thus a shi� toward a glycolytic metabolism in glial tumors. Together, these data suggest that IDH-mutated tumors display an increase in both FDG uptake and expression of hypoxic biomarkers. Yet, two recent studies did not � nd any association between IDH mutation and hypoxic biomarker expression (HIF-1α and its targets) or FDG uptake in patients with WHO grade II to IV gliomas (13, 14). � ese � ndings suggest that other biochemi-cal and cellular pathways are a� ected by mu-tations in IDH1 and IDH2 and that alterna-tive products generated by mutated IDH1 or IDH2 enzymes might a� ect entirely di� erent pathways. Further investigation is needed to establish whether IDH1 and IDH2 are func-tionally important to tumor formation and tumor progression.

Finally, IDH1 mutations in gliomas not only result in a loss of function but also con-fer an enzymatic gain of function. Recently, unbiased metabolite pro� ling revealed that expression of IDH mutant in glioma cell lines led to production of 2-HG (10). Further studies revealed that R132 (IDH1) and R172 (IDH2) mutants reduce α-KG to 2-HG, while converting NADPH to NADP+ (10, 11, 15). � e fact that neomorphic enzyme activity is a shared feature of the IDH1 and IDH2 mu-tations points to the importance of these en-zymes in cancer, which has led to speculation that 2-HG is an oncometabolite and potential biomarker in gliomas. Indeed, IDH-mutated gliomas are known to harbor a better prog-nosis than their wild-type counterparts, with a higher frequency of IDH mutations in low-grade than in high-grade gliomas (16).

SHIFTING THE PARADIGM IN GLIOMA

DETECTION

IDH mutation is a relevant biomarker in glioma management for both diagnostic and prognostic considerations. However, deter-

B R A I N I M A G I N G

Magnetic Resonance Metabolic Imaging of Glioma

Philippe Metellus1,2* and Dominique Figarella-Branger2,3

*To whom correspondence should be addressed. E-mail: philippe.metellus@mail.ap-hm.fr

1Department of Neurosurgery, Hôpital de la Timone, APHM, 13005 Marseille, France. 2Aix-Marseille Univer-sity, INSERM UMR 911, Centre de Recherche en Oncolo-gie et Oncopharmacologie (CRO2), 13005 Marseille, France. 3Department of Neuropathology, Hôpital de la Timone, APHM, 13005 Marseille, France.

P E R S P E C T I V E

2-Hydroxyglutarate (2-HG) is a potential oncometabolite involved in gliomagenesis that has been identi� ed as an aberrant product of isocitrate dehydrogenase (IDH)–mutated glial tumors. Recent genomics studies have shown that heterozygous mutation of IDH genes 1 and 2, present in up to 86% of grade II gliomas, is associated with a favorable outcome. Two reports in this issue describe both ex vivo and in vivo methods that could noninvasively detect the presence of 2-HG in glioma patients. This approach could have valuable implications for diagnosis, prognosis, and strati� cation of brain tumors, as well as for monitoring of treatment in glioma patients.

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

P E R S P E C T I V E

www.ScienceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ps1 2

mination of IDH status requires tumor tissue genetic analysis and is thus known only a� er a surgical procedure, such as a biopsy. � us, a noninvasive way to assess the presence of IDH mutation directly or indirectly would be useful in a clinical setting (Fig. 1). Hence, Capper et al. recently evaluated whether 2-HG in preoperative serum samples from patients with glioma correlated with the IDH1 and IDH2 mutation status and wheth-er there was an association between 2-HG levels and tumor size (17). In contrast to the strong accumulation of the 2-HG observed in the serum of acute myeloid leukemia (AML) patients with mutated IDH1 and IDH2, the authors failed to show any asso-ciation between 2-HG serum levels and IDH

tumor status (17). � ey also did not � nd any correlation between tumor size measured by magnetic resonance imaging (MRI) and 2-HG serum levels in glioma patients.

In this issue of Science Translational Med-

icine, Elkhaled, Jalbert, and colleagues report on a new magnetic resonance spectroscopy (MRS) method that could detect and quanti-fy the levels of 2-HG in recurrent, low-grade gliomas ex vivo (2). � e researchers found a strong concordance (86.4%) between IDH-mutant tissue samples and the presence of 2-HG as detected by proton (1H) high-resolution magic angle spinning (HR-MAS) spectroscopy. � e spectral editing and more sophisticated data acquisition methods used in this study, including variants of the two-dimensional (2D) experiment, appeared to be useful in separating these components and verifying individual resonances (2). Fo-cusing on the unobstructed spectral feature created by the α-proton actually improvesthe speci� city and accuracy of detecting 2-HG in an in vivo setting. Moreover, 2-HG production is found in any IDH1 or IDH2

mutation type (R132H, R132C, R132S, or R172) and thus constitutes a relevant bio-marker of tumor IDH status.

� e MRS methodology described by Elkhaled, Jalbert, and colleagues allowed for e� ective and clear delineation of 2-HG from neighboring metabolites, such as γ-aminobutyric acid (GABA), glutamine (Gln), and glutamate. � e levels of these—and several others (choline-containing spe-cies, lactate, and glutathione) known to be associated with malignant tissue—were found to correlate with 2-HG abundance when we used MRS (2), which suggests that it may be useful as a noninvasive biomarker. Finally, the authors showed that 2-HG levels correlated with histopathology (2): 2-HG

abundance was substantially lower in grade II than in grade IV glioma tissue samples. However, a� er normalizing by average cel-lularity, this di� erence no longer existed. � is implies that in vivo levels of 2-HG may be able to contribute not only to the classi-� cation of glioma but also to characterizing the spatial extent of in� ltrative lesions.

� ese � ndings raised the possibility of assessing IDH mutation status by using new MRS imaging paradigms. However, ex vivo results from Elkhaled, Jalbert, et al. needed in vivo validation. In the same is-sue, Andronesi et al. reported successful in vivo detection of 2-HG in two glioma patients with IDH1-mutated tumors, using similar MRS-based methods (2D LASER-correlation spectroscopy) (1). To ensure that 2-HG measurements were possible in human tissue, they � rst analyzed 2-HG sig-nals by using HR-MAS conditions in brain biopsies representing varied pathologies and IDH mutation statuses. � ey showed substantial di� erences in 2-HG signals in an IDH1-mutated anaplastic astrocytoma com-

pared with primary glioblastoma and non-tumor controls, which were both wild-type IDH1. Accordingly, in vivo experiments in human patients and volunteers showed the 2-HG cross-peak present in two IDH-mutated gliomas, whereas the metabolite was undetectable in wild-type IDH1 glio-blastoma patients and healthy volunteers. � ese major � ndings provide evidence that IDH mutation status can be assessed non-invasively by spectroscopic imaging (1). � e fact that no report exists about detectable increases in 2-HG in the blood, cerebrospi-nal � uid, or urine in glioma underscores the relevance and importance of such � ndings.

� e possibility to assess IDH mutation status preoperatively by using a noninva-sive MRS imaging–based technique could help distinguish di� erent glioma subtypes: astrocytomas from ependymomas and oli-gotumors from other brain tumors with oligodendroglioma-like morphology. It also could help in distinguishing gliomas from nonneoplastic central nervous sys-tem (CNS) lesions and therapy-induced

Fig. 1. Magnetic resonance metabolic imaging of a brain tumor. Detection of a surrogate marker for IDH mutation (2-HG) in a patient with a suspected glioma during a routine. The spectra identifi ed would assist in diagnosis, prognosis, and monitoring of these tumors. From left to right: a metabolic map of a brain tumor, a representative 1D metabolic spectrum recorded in a brain tumor

patient, and IDH1 direct sequencing showing an R123H mutation (CGT→CAT).

CR

ED

ITS

: (M

AIN

PA

NE

L) ©

BA

RIS

SIM

SE

K/I

ST

OC

KP

HO

TO

.CO

M; (T

OP

PA

NE

LS

) M

AX

IME

GU

YE

AN

D Y

AN

LE

FU

R

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

P E R S P E C T I V E

www.ScienceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ps1 3

changes. Furthermore, monitoring levels of 2-HG could inform whether or not thera-pies targeting the IDH1 pathway are hav-ing an e� ect. � e knowledge of IDH status in deep-seated (insula) tumors would assist in developing the appropriate therapeutic strategy, considering the surgical challenge in these cases (18).

Several other techniques have been de-veloped to detect IDH1 mutations, including DNA-based approaches and immunohisto-chemistry (19). However, these approaches depend on time-consuming extraction of nucleic acids; moreover, instrumentation is not available in all clinics. An antibody has been designed that speci� cally recog-nizes and binds the IDH mutant protein (20). � is has resulted in the generation of a highly speci� c antibody against IDH1 that is used routinely in neuropathology. Also, be-cause the mutant enzymes result in a gain of function to catalyze the reduction of α-KG to 2-HG, a recent study devised a gas chro-matography/mass spectrometry (GC/MS) method to detect 2-HG in formalin-� xed para� n-embedded glioma tissue samples (21). However, these approaches still rely on the analysis of tumor tissue samples.

Although the � ndings reported by Elkhaled, Jalbert, et al. (2) and Andronesi et al. (1) are promising, from a neuroonco-logical perspective, some limitations should be stressed. Indeed, the MRS methodol-ogy used in these studies is not available in all clinical radiology settings. Also, in vivo proof-of-principal data will need to be re-produced in a larger cohort. Last, in the study by Andronesi and coauthors, the 2-HG levels detected in the two patients were com-pared with matched glioma samples but not with the actual level of 2-HG measured in the patients’ tumors. To validate such an ap-proach, the comparison between 2-HG spec-troscopic measurements and actual 2-HG levels in tumor tissue by using liquid chro-matography/mass spectrometry will need to be performed in the same patient. With ad-ditional investigation, the compelling meth-od reported could be routinely applied in a clinical setting (Fig. 1).

DIAGNOSIS, PROGNOSIS, AND

PREDICTION

Histological diagnosis of gliomas is subject to substantial interobserver variation; thus, IDH1 mutation status can provide an objec-tive marker to identify the glioma subtype (22), such as distinguishing between prima-ry and secondary glioblastoma multiforme

(16). � e low incidence of IDH1 mutation in pilocytic astrocytomas also allows clini-cally relevant di� erentiation between dif-fuse and pilocytic astrocytomas (16). Fur-thermore, IDH1 mutations can have a role in distinguishing anaplastic astrocytomas from primary glioblastoma multiforme, as-trocytomas from ependymomas, oligoden-droglial tumors from other brain tumors with oligodendroglioma-like morphology, and also glioma from nonneoplastic CNS lesions and therapy-induced changes (16). � erefore, testing of the IDH status is rel-evant for diagnostic considerations in pri-mary brain tumors.

Whereas the strong prognostic signi� -cance of IDH mutations in high-grade glio-mas has been � rmly established (3–5), their importance in grade II gliomas is still under debate. In 2010, a retrospective study, which included a large group of both astrocytomas and oligotumors, the authors reported no in-� uence of IDH1 mutations on survival (8). Yet, several recent studies found IDH muta-tion to be associated with a favorable out-come in this population (5, 7, 18, 23). Never-theless, because patients in these studies were managed heterogeneously, whether IDH mu-tations are predictive or prognostic markers has not been widely investigated. � e impact of IDH1 and IDH2 mutations on the bene� t of adjuvant procarbazine, 1-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea, and vincristine chemotherapy in anaplastic oligodendrog-lial tumors has been studied in a prospective randomized phase III trial (European Orga-nization for Research and Treatment of Can-cer study 26951), and the investigators found that IDH1 mutation did not predict outcome (7). However, the potential impact of this ge-netic change should be investigated further with other treatment regimens.

� e neomorphic mutant IDH enzymes not only provide an attractive molecule for targeted therapy but also generate 2-HG, a new metabolite that may serve as a bio-marker for diagnosis, disease strati� cation, and prognosis. As a discovery with broad implications, IDH mutations have dramati-cally altered our understanding of cancer biology with the introduction of a new un-conventional type of cancer gene. Metabolic enzymes, much like kinases and transcrip-tion factors, have now regained their right-ful place on the growing “oncogene hit list” revealed by the “next gen” sequencing era. Recent � ndings (24, 25) suggest that 2-HG accumulation in IDH-mutated tumors may lead to genome-wide histone and DNA

methylation alterations. Indeed, these stud-ies provided a mechanistic link between these genetic mutations and the associated widespread epigenetic modi� cations. Spe-ci� cally, 2-HG has been shown to function as a competitive inhibitor of various α-KG–dependent dioxygenases, including histone demethylases and members of the ten-elev-en–translocation family of 5-methylcytosine hydroxylases (26). � e recent technical ad-vances concerning genome-wide sequencing and large-scale epigenetic pro� ling methods will certainly provide important new insights into the complex interplay between genetic and epigenetic alterations in various glioma subtypes and malignancy grades. MRS-based methods dedicated to noninvasively monitoring 2-HG levels could thus provide a means to better characterize genetic and epigenetic modi� cations in gliomas and to monitor the activity of eventual inhibitors of IDH and perhaps epigenetic therapies.

REFERENCES AND NOTES 1. O. C. Andronesi, G. S. Kim, E. Gerstner, T. Batchelor, A. A.

Tzika, V. R. Fantin, M. G. Vander Heiden, A. G. Sorensen,

Detection of 2-hydroxyglutarate in IDH-mutated glioma

patients by in vivo spectral-editing and 2D correlation

magnetic resonance spectroscopy. Sci. Transl. Med., 4,

116ra4 (2012).

2. A. Elkhaled, L. E. Jalbert, J. J. Phillips, H. A. I. Yoshihara,

R. Parvataneni, R. Srinivasan, G. Bourne, M. S. Berger, S.

M. Chang, S. Cha, S. J. Nelson, Magnetic resonance of 2-

hydroxyglutarate in IDH1-mutated low-grade gliomas.

Sci. Transl. Med., 4,116ra5 (2012).

3. D. W. Parsons, S. Jones, X. Zhang, J. C. Lin, R. J. Leary, P.

Angenendt, P. Mankoo, H. Carter, I.-M. Siu, G. L. Gallia, A.

Olivi, R. McLendon, B. A. Rasheed, S. Keir, T. Nikolskaya, Y.

Nikolsky, D. A. Busam, H. Tekleab, L. A. Diaz Jr., J. Hartigan,

D. R. Smith, R. L. Strausberg, S. K. N. Marie, S. M. O. Shinjo,

H. Yan, G. J. Riggins, D. D. Bigner, R. Karchin, N. Papado-

poulos, G. Parmigiani, B. Vogelstein, V. E. Velculescu, K. W.

Kinzler, An integrated genomic analysis of human glio-

blastoma multiforme. Science 321, 1807–1812 (2008).

4. H. Yan, D. W. Parsons, G. Jin, R. McLendon, B. A. Rasheed,

W. Yuan, I. Kos, I. Batinic-Haberle, S. Jones, G. J. Riggins,

H. Friedman, A. Friedman, D. Reardon, J. Herndon, K. W.

Kinzler, V. E. Velculescu, B. Vogelstein, D. D. Bigner, IDH1

and IDH2 mutations in gliomas. N. Engl. J. Med. 360, 765–

773 (2009).

5. M. Sanson, Y. Marie, S. Paris, A. Idbaih, J. Laff aire, F. Du-

cray, S. El Hallani, B. Boisselier, K. Mokhtari, K. Hoang-

Xuan, J. Y. Delattre, Isocitrate dehydrogenase 1 codon

132 mutation is an important prognostic biomarker in

gliomas. J. Clin. Oncol. 27, 4150–4154 (2009).

6. T. Watanabe, S. Nobusawa, P. Kleihues, H. Ohgaki, IDH1

mutations are early events in the development of astro-

cytomas and oligodendrogliomas. Am. J. Pathol. 174,

1149–1153 (2009).

7. M. J. van den Bent, H. J. Dubbink, Y. Marie, A. A. Brandes,

M. J. Taphoorn, P. Wesseling, M. Frenay, C. C. Tijssen, D.

Lacombe, A. Idbaih, R. van Marion, J. M. Kros, W. N. Din-

jens, T. Gorlia, M. Sanson, IDH1 and IDH2 mutations are

prognostic but not predictive for outcome in anaplastic

oligodendroglial tumors: A report of the European Or-

ganization for Research and Treatment of Cancer Brain

Tumor Group. Clin. Cancer Res. 16, 1597–1604 (2010).

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

P E R S P E C T I V E

www.ScienceTranslationalMedicine.org 11 January 2012 Vol 4 Issue 116 116ps1 4

8. Y. H. Kim, S. Nobusawa, M. Mittelbronn, W. Paulus, B.

Brokinkel, K. Keyvani, U. Sure, K. Wrede, Y. Nakazato, Y.

Tanaka, A. Vital, L. Mariani, R. Stawski, T. Watanabe, U. De

Girolami, P. Kleihues, H. Ohgaki, Molecular classifi cation

of low-grade diff use gliomas. Am. J. Pathol. 177, 2708–

2714 (2010).

9. S. Zhao, Y. Lin, W. Xu, W. Jiang, Z. Zha, P. Wang, W. Yu, Z.

Li, L. Gong, Y. Peng, J. Ding, Q. Lei, K.-L. Guan, Y. Xiong,

Glioma-derived mutations in IDH1 dominantly inhibit

IDH1 catalytic activity and induce HIF-1α. Science 324,

261–265 (2009).

10. L. Dang, D. W. White, S. Gross, B. D. Bennett, M. A. Bit-

tinger, E. M. Driggers, V. R. Fantin, H. G. Jang, S. Jin, M. C.

Keenan, K. M. Marks, R. M. Prins, P. S. Ward, K. E. Yen, L. M.

Liau, J. D. Rabinowitz, L. C. Cantley, C. B. Thompson, M. G.

Vander Heiden, S. M. Su, Cancer-associated IDH1 muta-

tions produce 2-hydroxyglutarate. Nature 462, 739–744

(2009).

11. P. S. Ward, J. Patel, D. R. Wise, O. Abdel-Wahab, B. D. Ben-

nett, H. A. Coller, J. R. Cross, V. R. Fantin, C. V. Hedvat, A. E.

Perl, J. D. Rabinowitz, M. Carroll, S. M. Su, K. A. Sharp, R.

L. Levine, C. B. Thompson, The common feature of leuke-

mia-associated IDH1 and IDH2 mutations is a neomor-

phic enzyme activity converting alpha-ketoglutarate to

2-hydroxyglutarate. Cancer Cell 17, 225–234 (2010).

12. O. De Witte, F. Lefranc, M. Levivier, I. Salmon, J. Brotchi, S.

Goldman, FDG-PET as a prognostic factor in high-grade

astrocytoma. J. Neurooncol. 49, 157–163 (2000).

13. P. Metellus, C. Colin, D. Taieb, E. Guedj, I. Nanni-Metellus,

A. M. de Paula, C. Colavolpe, S. Fuentes, H. Dufour, M.

Barrie, O. Chinot, L. Ouafi k, D. Figarella-Branger, IDH

mutation status impact on in vivo hypoxia biomarkers

expression: New insights from a clinical, nuclear imaging

and immunohistochemical study in 33 glioma patients.

J. Neurooncol. 105, 591–600 (2011).

14. S. C. Williams, M. A. Karajannis, L. Chiriboga, J. G. Golfi -

nos, A. von Deimling, D. Zagzag, R132H-mutation of

isocitrate dehydrogenase-1 is not suffi cient for HIF-1α

upregulation in adult glioma. Acta Neuropathol. 121,

279–281 (2011).

15. S. Gross, R. A. Cairns, M. D. Minden, E. M. Driggers, M. A.

Bittinger, H. G. Jang, M. Sasaki, S. Jin, D. P. Schenkein, S.

M. Su, L. Dang, V. R. Fantin, T. W. Mak, Cancer-associated

metabolite 2-hydroxyglutarate accumulates in acute

myelogenous leukemia with isocitrate dehydrogenase

1 and 2 mutations. J. Exp. Med. 207, 339–344 (2010).

16. C. Hartmann, J. Meyer, J. Balss, D. Capper, W. Mueller, A.

Christians, J. Felsberg, M. Wolter, C. Mawrin, W. Wick, M.

Weller, C. Herold-Mende, A. Unterberg, J. W. Jeuken, P.

Wesseling, G. Reifenberger, A. von Deimling, Type and

frequency of IDH1 and IDH2 mutations are related to

astrocytic and oligodendroglial diff erentiation and age:

A study of 1,010 diff use gliomas. Acta Neuropathol. 118,

469–474 (2009).

17. D. Capper, M. Simon, C.-D. Langhans, J. G. Okun, J. C.

Tonn, M. Weller, A. V. Deimling, C. Hartmann for the Ger-

man Glioma Network, 2-Hydroxyglutarate concentra-

tion in serum from patients with gliomas does not cor-

relate with IDH1/2 mutation status or tumor size. Int. J.

Cancer A. (2011).

18. P. Metellus, B. Coulibaly, C. Colin, A. M. de Paula, A.

Vasiljevic, D. Taieb, A. Barlier, B. Boisselier, K. Mokhtari, X.

W. Wang, A. Loundou, F. Chapon, S. Pineau, L. Ouafi k, O.

Chinot, D. Figarella-Branger, Absence of IDH mutation

identifi es a novel radiologic and molecular subtype of

WHO grade II gliomas with dismal prognosis. Acta Neu-

ropathol. 120, 719–729 (2010).

19. M. Preusser, D. Capper, C. Hartmann; Euro-CNS Research

Committee, IDH testing in diagnostic neuropathology:

Review and practical guideline article invited by the

Euro-CNS research committee. Clin. Neuropathol. 30,

217–230 (2011).

20. D. Capper, H. Zentgraf, J. Balss, C. Hartmann, A. von De-

imling, Monoclonal antibody specifi c for IDH1 R132H

mutation. Acta Neuropathol. 118, 599–601 (2009).

21. F. Sahm, D. Capper, S. Pusch, J. Balss, A. Koch, C. D. Lang-

hans, J. G. Okun, A. von Deimling, Detection of 2-hydrox-

yglutarate in formalin-fi xed paraffi n-embedded glioma

specimens by gas chromatography/mass spectrometry.

Brain Pathol. 22, 26–31 (2012).

22. N. K. Kloosterhof, L. B. Bralten, H. J. Dubbink, P. J. French,

M. J. van den Bent, Isocitrate dehydrogenase-1 muta-

tions: A fundamentally new understanding of diff use

glioma? Lancet Oncol. 12, 83–91 (2011).

23. H. J. Dubbink, W. Taal, R. van Marion, J. M. Kros, I. van

Heuvel, J. E. Bromberg, B. A. Zonnenberg, C. B. Zon-

nenberg, T. J. Postma, J. M. Gijtenbeek, W. Boogerd, F. H.

Groenendijk, P. A. Smitt, W. N. Dinjens, M. J. van den Bent,

IDH1 mutations in low-grade astrocytomas predict sur-

vival but not response to temozolomide. Neurology 73,

1792–1795 (2009).

24. W. Xu, H. Yang, Y. Liu, Y. Yang, P. Wang, S. H. Kim, S. Ito, C.

Yang, P. Wang, M. T. Xiao, L. X. Liu, W. Q. Jiang, J. Liu, J. Y.

Zhang, B. Wang, S. Frye, Y. Zhang, Y. H. Xu, Q. Y. Lei, K. L.

Guan, S. M. Zhao, Y. Xiong, Oncometabolite 2-hydroxy-

glutarate is a competitive inhibitor of α-ketoglutarate-

dependent dioxygenases. Cancer Cell 19, 17–30 (2011).

25. M. E. Figueroa, O. Abdel-Wahab, C. Lu, P. S. Ward, J. Patel,

A. Shih, Y. Li, N. Bhagwat, A. Vasanthakumar, H. F. Fer-

nandez, M. S. Tallman, Z. Sun, K. Wolniak, J. K. Peeters, W.

Liu, S. E. Choe, V. R. Fantin, E. Paietta, B. Löwenberg, J. D.

Licht, L. A. Godley, R. Delwel, P. J. Valk, C. B. Thompson,

R. L. Levine, A. Melnick, Leukemic IDH1 and IDH2 muta-

tions result in a hypermethylation phenotype, disrupt

TET2 function, and impair hematopoietic diff erentiation.

Cancer Cell 18, 553–567 (2010).

26. B. Malzkorn, M. Wolter, M. J. Riemenschneider, G. Reifen-

berger, Unraveling the glioma epigenome: From molec-

ular mechanisms to novel biomarkers and therapeutic

targets. Brain Pathol. 21, 619–632 (2011).

Competing interests: The authors have no confl icts of inter-

est to declare.

Citation: P. Metellus, D. Figarella-Branger, Magnetic resonance

metabolic imaging of glioma. Sci. Transl. Med. 4, 116ps1 (2012)

10.1126/scitranslmed.3003591

on

Mar

ch 9

, 201

2st

m.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om