mr-detectable metabolic consequences of mitogen-activated protein kinase kinase (mek) inhibition

MR-detectable metabolic consequences ofmitogen-activated protein kinase kinase (MEK)inhibitionAlessia Lodi†, Sarah M. Woods and Sabrina M. Ronen*

Metabolic reprogramming is increasingly being viewed as a hallmark of cancer. Accordingly, metabolic readouts canserve as biomarkers of response to therapy. The goal of this study was to investigate some of the MRS-detectablemetabolic consequences of mitogen-activated protein kinase kinase (MEK) inhibition. We investigated PC3 prostatecancer, MCF-7 breast cancer and A375 melanoma cells, and determined that, consistent with previous studies, MRS-detectable levels of phosphocholine decreased significantly in all cell lines (to 63%, 50% and 18% of the control, re-spectively) following MEK inhibition with U0126. This effect was mediated by a decrease in the expression of cholinekinase α, the enzyme that catalyzes the phosphorylation of choline. In contrast, the impact of MEK inhibition onglycolysis was cell line dependent. A375 cells, which express mutant BRAF, demonstrated significant decreases inglucose uptake (to 36% of control) and lactate production (to 42% of control) in line with positron emission tomog-raphy data. In contrast, in PC3 and MCF-7 cells, increases in glucose uptake (to 198% and 192% of control, respec-tively) and lactate production (to 177% and 212% of control, respectively) were observed, in line with a previoushyperpolarized 13C MRS study. This effect is probably mediated by the activation of the phosphoinositide3-kinase pathway and AMP-activated protein kinase. Our findings demonstrate the value of translatable non-invasive MRS methods for the provision of information on cellular metabolism as an indication of the activationof potential feedback loops following MEK inhibition. Copyright © 2014 John Wiley & Sons, Ltd.Additional supporting information may be found in the online version of this article at the publisher’s web site.

Keywords: MRS; MEK; U0126; metabolism

INTRODUCTION

Metabolic reprogramming is increasingly being recognized as akey aspect of carcinogenesis and disease progression. It is medi-ated by signaling via key oncogenic signaling pathways and,conversely, response to therapies specifically targeting thesesignaling pathways is associated with the normalization of cellmetabolism (1–3). These observations have fueled the develop-ment of imaging approaches that can monitor metabolism andthus improve the early detection of disease, provide timely andefficient evaluation of response to treatment and lead to theidentification of therapeutic resistance.

Specifically, the Warburg effect – elevated glucose uptake andits aerobic metabolism to lactate – is observed in most tumortypes, and can be imaged in vivo by monitoring of the uptakeof the glucose analog [18F]-2-fluoro-2-deoxy-D-glucose (FDG)using positron emission tomography (PET). FDG-PET is routinelyused in the clinic to detect the presence of tumor, and for theevaluation of the early response and treatment outcome (4,5).Methods based on MRS and MRSI can also be used to monitorthe Warburg effect. 1H MRS, which detects steady-state metabo-lite levels, can be used to probe the elevated levels of tumorlactate associated with increased glycolysis (6). In addition, 13CMRS can be used to monitor metabolic fluxes and to probe themetabolic fate of 13C-labeled metabolites, including glucose.However, the relatively low sensitivity of 13C MRS has limitedits application in the clinic (7,8). In recent years, with thedevelopment and optimization of dissolution dynamic nuclear

polarization (DNP) methods, and the outstanding signal en-hancement that can be achieved with hyperpolarization of 13C-labeled compounds, the use of 13C MRS for the detection of

* Correspondence to: S. M. Ronen, Radiology and Biomedical Imaging, ByersHall, 1700 4th St., San Francisco, CA 94158, USA.E-mail: [email protected]

A. Lodi, S. M. Woods, S. M. RonenRadiology and Biomedical Imaging, University of California San Francisco, SanFrancisco, CA, USA

† Present address: Department of Nutritional Sciences, University of Texas atAustin, Austin, TX, USA

Abbreviations used: AMPK, AMP-activated protein kinase; CHKA, choline ki-nase α; DMEM, Dulbecco’s modified Eagle’s medium; DMSO, dimethyl sulfoxide;DNP, dynamic nuclear polarization; EDTA, ethylenediaminetetraacetic acid;ERK, extracellular signal-regulated kinase; FDG, [18F]-2-fluoro-2-deoxy-D-glucose; FDR, false discovery rate; GAPDH, glyceraldehyde 3-phosphate dehy-drogenase; GPC, glycerophosphocholine; HK2, hexokinase 2; HSP90, heat shockprotein 90; IKK, I kappa beta kinase; MAPK, mitogen-activated protein kinase;MDPA, methylenediphosphonic acid; MEK, mitogen-activated protein kinase ki-nase; PC, phosphocholine; PCA, principal component analysis; p-ERK, phosphor-ylated extracellular signal-regulated kinase; PET, positron emission tomography;PFKFB3, 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3; PI3K, phosphoi-nositide 3-kinase; RIN, RNA Integrity Number; RIPA, radioimmunoprecipitation assay;SAM, significance analysis of microarrays; SDS-PAGE, sodium dodecylsulfate-polyacrylamide gel electrophoresis; SLC2A1, solute carrier family 2 (facilitatedglucose transporter), member 1; tCho, total choline, comprising choline, PC andGPC; TMSP, sodium 3-(trimethylsilyl)propionate-2,2,3,3-d4; UCSF, University ofCalifornia San Francisco; VEGFR, vascular endothelial growth factor receptor.

Research article

Received: 3 January 2014, Revised: 8 March 2014, Accepted: 9 March 2014, Published online in Wiley Online Library: 2 April 2014

(wileyonlinelibrary.com) DOI: 10.1002/nbm.3109

NMR Biomed. 2014; 27: 700–708 Copyright © 2014 John Wiley & Sons, Ltd.

700

in vivo metabolism has generated increasing interest. In particu-lar, hyperpolarized pyruvate has been used in cancer modelsin vitro and in vivo to detect the production of lactate at theend of the glycolytic pathway and to monitor the response totherapy in a variety of cancers. A clinical trial recently performedat the University of California San Francisco (UCSF) has also dem-onstrated the utility of this approach in the clinic (9–19).Choline metabolism is another metabolic pathway that is al-

tered in cancer and has been used to detect disease and to mon-itor response to treatment (20). PET tracers based on 11C- and18F-labeled choline and choline derivatives have been devel-oped and evaluated for the in vivo radiological diagnosis of met-astatic and recurrent tumors, as well as for the evaluation oftumor response to drug treatment (21–23). 1H MRS can be usedto detect steady-state choline-containing metabolite levels [totalcholine (tCho), comprising choline, phosphocholine (PC) andglycerophosphocholine (GPC)], and the concentration of PCand tCho metabolites has been shown to be associated with ma-lignancy in cancer models as well as in patients, including inbreast, brain, prostate and other cancers. Conversely, these me-tabolites typically decrease in response to therapy (20,24–27).In the case of treatment aimed at the inhibition of the Ras/mi-

togen-activated protein kinase (Ras/MAPK) pathway, inhibitionof Ras signaling in fibroblasts, and response of breast and colo-rectal cancer models to mitogen-activated protein kinase kinase(MEK), inhibition has been associated with a decrease in PClevels (24,26). However, its effect on the glycolytic pathway ismore variable. In cells that harbor the BRAF mutation, treatmentwith BRAF and BRAF/vascular endothelial growth factor recep-tor-2 (BRAF/VEGFR-2) inhibitors led to a decrease in FDG-PET-detected glucose uptake in melanoma cells (28,29), and a de-crease in extracellular lactate levels was also detected using 1HMRS following MEK inhibition (30). In contrast, we have found re-cently that the production of lactate is increased in prostate andbreast cancer cells treated with the MEK inhibitor U0126 (11).The goal of this work was therefore to perform a detailed investiga-tion of the metabolic alterations associated with inhibition of theMAPK pathway, with a particular emphasis on the glucose andcholine metabolic pathways that can be imaged. To this end, wecharacterized the metabolic consequences of treatment with theMEK inhibitor U0126 in prostate cancer, breast cancer and mela-noma cells. We found that, in MEK-inhibited mutant BRAF mela-noma, a decrease in glycolytic flux occurred. In contrast, MEKinhibition in prostate and breast cancer cells resulted in an increasein glucose uptake and lactate production, probably mediated byactivation of the phosphoinositide 3-kinase (PI3K) pathway and/or AMP-activated protein kinase (AMPK). However, in all cell lines,the levels of PC decreased following U0126 treatment, mediatedby a decrease in choline kinase α (CHKA) expression.

EXPERIMENTAL DETAILS

Cell culture and treatments

PC3 (prostate), MCF-7 (breast) and A375 (melanoma) cancer celllines were obtained from the American Type Culture Collectionvia the UCSF Cell Culture Facility (San Francisco, CA, USA). Allexperiments were performed within 6–12 months of cell lineauthentication (using short tandem repeat analysis, PowerPlex 1.2System, Promega, Madison, WI, USA) and mycoplasma testing(MycoAlert™ mycoplasma detection kit, Lonza, Rockland, ME,USA). PC3 and MCF-7 cells were cultured in low-glucose Dulbecco’s

modified Eagle’s medium (DMEM) and A375 cells in high-glucoseDMEM. In all cases, medium was supplemented with 10%heat-inactivated fetal bovine serum, 2 mM L-glutamine, 100 units/mL of penicillin and 100 μg/mL of streptomycin, and cells werecultured in a humidified chamber at 37 °C in 5% CO2 in air.

Cells were incubated with U0126 (LC Laboratories, Woburn,MA, USA) at doses of 50 μM for PC3 and 25 μM for MCF-7 andA375 cells. All treatments were performed over 48 h withmatching DMSO solvent controls (1: 1000 in the culturemedium), and U0126 was replenished every 24 h. Treatmentdoses were determined as described previously (31) using theWST-1 cell proliferation assay (Roche, Indianapolis, IN, USA),and doses were selected such that they induced a 50% decreasein cell viability after 48 h of treatment. In all cases, inhibition ofthe target at the selected drug dose was confirmed by probingfor phosphorylated extracellular signal-regulated kinase (p-ERK)levels using Western blotting (see below).

Cell extraction and MRS data acquisition and analysis

Cell extracts were prepared using a dual-phase extraction, as de-scribed previously (24,31–33). For each sample, a parallel flaskwas trypsinized and used to determine cell numbers. At the timeof extraction, 1 mL of the extracellular medium was alsocollected, snap frozen and stored at�80 °C until 1H MRS analysis.

MR spectra of cell extracts and growth medium were acquiredon a 600-MHz scanner (Varian/Agilent, Palo Alto, CA, USA). 1HMRS data were acquired using a 90° pulse, relaxation delay of3 s, spectral width of 8 kHz and either presat or excitationsculpting (34) for suppression of the water signal. Sodium3-(trimethylsilyl)propionate-2,2,3,3-d4 (TMSP; Cambridge IsotopeLaboratories, Tewksbury, MA, USA) was used as reference. Fullyrelaxed spectra (90° pulse and 60 s relaxation delay) were alsoobtained in order to determine saturation correction factors.ACD/Spec Manager version 9.15 software (Advanced ChemistryDevelopment, Toronto, Ontario, Canada) was used for theanalysis of spectra and for peak area integration. Peakassignments were based on previous literature and availabledatabases, such as HMDB (http://www.hmdb.ca/) and BML(www.bml-nmr.org). Metabolites were quantified by correctingfor saturation and normalizing to the reference and cell number.Proton-decoupled (Waltz-16) 31P MR spectra were acquired usinga 30° pulse, relaxation delay of 3 s and spectral width of 8 kHzfollowing the addition of ethylenediaminetetraacetic acid (EDTA)(pH 8; final concentration, 10 mM). Methylenediphosphonic acid(MDPA; Sigma-Aldrich, St. Louis, MO, USA) was used as referenceand metabolites were quantified as above.

Multivariate statistical analysis (principal component analysis,PCA) was used to identify the main treatment-induced metabolicchanges in the 1H MR spectra, as described previously (31).Difference spectra were obtained by subtracting the average ofall spectra acquired on control cells from the average spectrumof treated cells (n = 8 in both conditions).

Cell perfusion system

For MRS studies of live cells, experiments were performed usinga cell perfusion MR-compatible bioreactor system, as describedpreviously (11,12,32,35). Briefly, cells were seeded, allowed to ad-

here and grow on Biosilon® microcarrier beads (Nunc, Roskilde,Denmark) for approximately 24 h and treated for approximately48 h prior to the MRS study. In addition to the cells required for

MRS DETECTION OF MEK INHIBITION

NMR Biomed. 2014; 27: 700–708 Copyright © 2014 John Wiley & Sons, Ltd. wileyonlinelibrary.com/journal/nbm

701

the MRS studies, one extra aliquot of cells was seeded ontomicrocarrier beads at the same density, treated in the sameway and used, following trypsinization, to obtain a cell count. Im-mediately prior toMRS studies, the cell-bearing beads were loadedinto the bioreactor system. A perfusion system consisting of oneinflow and two outflow lines enabled the circulation of growthmedium to a 10-mm MR tube at the bottom of which the cells onbeads were immobilized. The system was under a 5% CO2/95% airover-pressure and placed in the MR probe at 35 °C. Glucose andcholine in the growth medium (100 mL, continuously circulated at1.5 mL/min) were completely replaced with 1-13C-glucose and1,2-13C-choline, respectively. To maintain treatment, either theappropriate dose of U0126 or DMSOwas also added to themedium.

Live cell studies and MRS data acquisition and analysis

All MRS studies on live cells were performed on a 500-MHz spec-trometer (Varian/Agilent). Proton-decoupled (Waltz-16) 13C MRspectra were acquired every hour for a period of up to 20 h usinga 60° pulse and a relaxation delay of 6 s. Cells were continuouslyperfused with 13C-labeled medium. Inverse-gated fully relaxedspectra (90° pulse and 60 s relaxation delay) were also obtainedto determine the saturation correction factors for metabolitequantification. Prior to and at the end of each 13C MRS experi-ment, 31P MRS spectra were acquired to confirm cell viabilityusing a pulse–acquire scheme with a 30° pulse, relaxation delayof 3 s, spectral width of 8 kHz and composite pulse 1H decouplingduring acquisition. ACD/Spec Manager version 9.15 software wasused for peak area integration. Metabolite concentrations weredetermined as above by normalizing each metabolite level to areference compound of known concentration (inorganic phos-phate in the perfusion medium for 31P data, and glucose at theonset of the experiment for 13C studies), correcting for saturationeffects and normalizing to cell number.

Microarray analysis of gene expression

Total cellular RNA was isolated using the RNeasy Mini Kit (Qiagen,Germantown, MD, USA), according to the manufacturer’s instruc-tions. RNA quality was determined by Bioanalyzer (Agilent, SantaClara, CA, USA), considering RNA Integrity Number (RIN) values of≥8.0 as acceptable (for all samples, RIN = 10). Microarray hybridiza-tion was performed at the UCSF Genomics Core Laboratories(UCSF/Gladstone, San Francisco, CA, USA) using the Human Gene1.0 ST, and analysis was carried out by fluorescence detection usingthe Agilent GeneArray Scanner (Agilent). Data acquisition wasperformed using Micro Array Suite 5.0 software (Affymetrix, SantaClara, CA, USA). Microarray experiments were performed with fourrepeats for each condition. Analysis of significance was performedusing significance analysis of microarrays (SAM) (36). Microarraydata are available through the ArrayExpress public repository(ArrayExpress accession number: E-MTAB-2142) in compliance withstandards of the Microarray Data Gene Expression Society.

Quantitative real-time polymerase chain reaction analysis ofgene expression

Total cellular RNA was isolated using the RNeasy Mini Kit andquantified using a NanoDrop ND1000 fluorospectrometer(NanoDrop Technologies, Wilmington, DE, USA). Reversetranscription was performed using the QuantiTect ReverseTranscription kit (Qiagen). Real-time polymerase chain reactionwas performed on the resulting cDNA on a Taqman 7900

(Applied Biosystems, Grand Island, NY, USA). Expression of SLC2A1[solute carrier family 2 (facilitated glucose transporter), member 1;Hs00892681_m1], HK2 (hexokinase 2; Hs00606086_m1),PFKFB3 (6-phosphofructo-2-kinase/fructose-2,6-biphosphatase3; Hs00998700_m1), CHKA (Hs03682798_m1) and SLC44A1(Hs00223114_m1; all gene-specific primer/probe sets fromApplied Biosystems) was examined using Assays-on-Demand(Applied Biosystems) and normalized to the expression ofglyceraldehyde 3-phosphate dehydrogenase (GAPDH) or the18S ribosomal subunit (Integrated DNA Technologies,Coralville, IA, USA).

Western blotting

Whole-cell extracts were obtained [radioimmunoprecipitation assay(RIPA) buffer], separated on 4–20% sodium dodecylsulfate-poly-acrylamide gel electrophoresis (SDS-PAGE) gels (Bio-Rad, Hercules,CA, USA) and probed with the following: rabbit antibodies againstp44/42 MAPK (ERK1/2), phospho-p44/42 MAPK (p-ERK1/2), β-actin(as loading control), Akt, p-Akt, AMPK, p-AMPK (all fromCell Signaling,Danvers, MA, USA), mouse antibodies against CHKA (Sigma-Aldrich,St. Louis, MO, USA) and horseradish peroxidase-linked secondaryantibodies (Cell Signaling or Abcam, Cambridge, MA, USA).Immunocomplexeswere visualized using Pierce ECLWestern BlottingSubstrate (Pierce Biotechnology, Rockford, IL, USA). Quantification ofbands was performed using the gel analysis submenu in ImageJ(http://rsbweb.nih.gov/ij/index.html), normalized to the loadingcontrol and reported as a percentage of the control.

Statistics

All data are reported as means ± standard deviation. Statistical sig-nificance was determined using Student’s t-test with p < 0.05considered to be significant.

RESULTS

Untargeted metabolic profiling of MEK inhibition in PC3prostate cancer cells

PC3 prostate cancer cells were treated with 50 μM U0126 for 48h, a treatment dose that was found to effectively inhibit MEK sig-naling, as illustrated by the decrease in ERK phosphorylation(Fig. 1A, n = 4), and which induced a decrease in cell numberto 54 ± 8% of untreated cells. To identify the metabolic changesassociated with MEK inhibition in PC3 cells, we first performed anuntargeted 1H MRS-based metabolomics study and analyzed theentire 1H MR-detectable metabolome using PCA. The scores plotobtained from PCA is illustrated in Fig. 1B and indicates verygood separation of the treated and untreated samples. Becausewe had previously investigated the metabolomic consequencesof treatment with PI3K and heat shock protein 90 (HSP90) inhibi-tion (31), we also questioned whether inhibition of MAPK signal-ing in response to U0126 resulted in a specific metabolomicssignature that could be clearly distinguished from the responseto other targeted therapies. Our findings clearly demonstratethat the response to each of the three therapies leads to a dis-tinct metabolic signature, although, when compared with othertherapies, treatment with U0126 induced a more limited meta-bolic response [illustrated in Fig. S1 (Supporting information)by the larger separation of the control group from theLY294002 and 17-AAG treatment groups relative to its separa-tion from the U0126 treatment group].

A. LODI ET AL.

wileyonlinelibrary.com/journal/nbm Copyright © 2014 John Wiley & Sons, Ltd. NMR Biomed. 2014; 27: 700–708

702

We then performed an in-depth analysis of the specificmetabolic changes induced in PC3 prostate cancer cells byU0126 treatment (n = 8). The difference spectrum between theaverage data of control and treated cells (Fig. 1C) indicates thatonly a few metabolites change significantly following treatment[false discovery rate (FDR)-corrected p values are shown]. In par-ticular, lactate and fumarate (signal resonates at 6.5 ppm and isnot shown in Fig. 1C) were found to accumulate intracellularlyin U0126-treated PC3 cells, whereas PC, GPC, aspartate andsuccinate levels all decreased significantly (Fig. 1D).

Treatment-induced metabolic flux changes in MEK-inhibitedPC3 prostate cancer cells

Next, we focused on the investigation of the metabolic pathwaysthat are associated with lactate and PC – two metabolites thatshow significant changes with treatment and can be readilyimaged in vivo. As intracellular lactate was elevated in U0126-treated cells, we investigated the treatment-induced modulationof glucose metabolism. Using 13C MRS, we longitudinallymonitored glucose consumption and its metabolic fate in livePC3 cells (n = 6). 13C MR spectra acquired sequentially showedprogressive consumption of glucose from the medium and itsconversion into lactate (Fig. 2A). Significant increases in bothglucose consumption (to 198% of the control from 270 ± 45 to534 ± 69 fmol/cell/h; p = 0.01) and lactate production (to 177%of control from 208 ± 39 to 369 ± 53 fmol/cell/h; p = 0.003) were

observed in cells undergoing treatment with U0126 whencompared with controls (Fig. 2B).

In an effort to understand the underlyingmechanism for our ob-servations, we investigated the expression of genes coding forproteins involved in the regulation of the glycolytic pathway (37)using microarray analysis. Most notably, we found that HK2increased by 1.6-fold (q = 0.001), PFKFB3 increased by two-fold(q = 0.001) and SLC2A1 increased by 1.95-fold (q = 0.001). Wefurther confirmed these findings by probing mRNA levels usingquantitative real-time polymerase chain reaction in control andU0126-treated PC3 cells (n = 4). HK2 increased by 1.6-fold, whereasPFKFB3 and SLC2A1 increased by more than two-fold (Fig. 2C).

The increase in glycolysis observed in our cells can be medi-ated by several factors. In particular, signaling via the PI3K path-way influences the expression of glycolytic enzymes (38), and itis well established that inhibition of the MAPK pathway can in-duce a compensatory effect that involves the activation of PI3Ksignaling (39,40). We therefore probed the Akt status in our cells(n = 4) and observed a drug-induced increase in the levels ofp-Akt to 150 ± 19% of the control (p = 0.04; Fig. 3A), providinga possible explanation for our findings. Activation of AMPK canalso lead to an increase in glycolytic flux (41). Using 31P MRS,we probed the levels of phosphorus-containing intracellular me-tabolites in control and U0126-treated PC3 cells (Fig. 3B; n = 4);we found that ATP levels decreased from 7.7 ± 0.8 fmol/cell incontrols to 5.7 ± 0.6 fmol/cell (p = 0.03) in treated cells, whereasADP levels remained unchanged (AMP levels were below the

Figure 1. Metabolic changes induced by mitogen-activated protein kinase (MAPK) inhibition with U0126 in PC3 prostate cancer cells. (A) Inhibition ofextracellular signal-regulated kinase (ERK) phosphorylation as probed by Western blotting following 48 h of treatment with U0126. C, untreated; U,U0126-treated. (B) Scores plot for the principal component analysis performed on the 1H MRS datasets acquired on cell extracts (n = 8) of untreatedand U0126-treated PC3 cells. (C) 1H MR spectrum (zoomed in section 0.8–4.2 ppm) illustrating the difference between control and U0126-treated cells(average spectrum acquired on U0126-treated PC3 cells minus average spectrum of control cells). Color bar indicates false discovery rate (FDR)-corrected p value calculated at each point of the spectrum (fumarate, which resonates at 6.5 ppm, is not included in the represented spectral region).(D) Intracellular concentration (fmol/cell) of metabolites altered by treatment with U0126. Asp, aspartate; Fum, fumarate; GPC, glycerophosphocholine;Lac, lactate; PC, phosphocholine; Suc, succinate. *p < 0.01; **p < 10�4; ***p < 10�5.



703

detection level). This resulted in an increase in the ADP to ATP ratio(ADP/ATP) in treated cells to 255 ± 116% of the control (p = 0.04;Fig. 3C), an effect that has been shown to promote the activation ofAMPK (42). We confirmed that AMPK was activated by probingp-AMPK levels (n = 3), and observed an increase in p-AMPK in treatedcells to 249 ± 90% (p = 0.009) of the control (Fig. 3A). This could alsoexplain the increase in glycolytic flux observed in our treated cells.

In line with previous findings, the 1H MRS data (Fig. 1D) alsoshowed a decrease in PC levels from 14.5 ± 1.2 to 8.6 ± 1.5fmol/cell. 31P MRS (Fig. 3B) confirmed these findings. PC levelsdecreased in U0126-treated cells from 12.8±1.8 to 8.4 ± 1.0fmol/cell (p = 0.03). Similarly, GPC levels also decreased

significantly after treatment, in line with the 1H MRS results, from1.6 ± 0.3 to 0.7 ± 0.2 fmol/cell (p = 0.03). To further investigatethe underlying mechanism for the modulation in PC levels, wedetermined the rates of de novo PC synthesis by perfusing livePC3 cells with medium containing 1,2-13C-choline and using 13CMRS to monitor the buildup of 1,2-13C-PC (Fig. 2A). In linewith the decrease in total PC levels, the initial rate of its denovo synthesis decreased to 60% of the control (p = 0.04)in U0126-treated PC3 cells from 1.04 ± 0.18 to 0.62 ± 0.12fmol/cell/h (Fig. 4A). The accumulation of 1,2-13C-PC reacheda plateau after approximately 14 h, at which point the sizesof the labeled PC pools in untreated (12.8 ± 2.2 fmol/cell)

Figure 2. Effect of treatment with U0126 on glucose metabolism and glycolysis-related gene expression. (A) Representative 13C MR spectra acquired se-quentially over 16 h in live PC3 cells continuously perfused with medium containing 1-13C-glucose and 1,2-13C-choline. (B) Metabolic rates (fmol/cell/h) ofglucose consumption and lactate production. (C) Changes in mRNA levels for selected genes related to glycolysis. HK2, hexokinase 2; PFKFB3,6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3; SLC2A1, solute carrier family 2 (facilitated glucose transporter), member 1; *p ≤ 0.01; **p < 0.001.

Figure 3. Effect of treatment with U0126 on signaling and phosphorus-containing metabolites. (A) Changes in p-Akt, Akt, p-AMPK and AMPK(AMP-activated protein kinase) protein levels in untreated (C) and U0126-treated (U) PC3 cells (β-actin shown as loading control). (B) Representative31P MR spectra of control and treated PC3 cells. GPC, glycerophosphocholine; GPE, glycerophosphoethanolamine; PC, phosphocholine; PCr,phosphocreatine; PE, phosphoethanolamine; Pi, inorganic phosphate; UDPS, uridine diphosphate sugars. (C) Changes in ADP/ATP levels. *p < 0.05.

A. LODI ET AL.


704

and treated (7.9 ± 1.7 fmol/cell) cells were in line with thesteady-state total intracellular PC levels determined fromthe 1H and 31P MRS metabolic profiling.Choline is actively transported to the cytoplasm from the ex-

tracellular space by choline transporters, and is then intracellu-larly converted to PC by CHK (32). In an effort to understandthe factors that lead to the reduction in PC synthesis, we moni-tored the expression of choline transporters and CHKA, the iso-form that is typically overexpressed and active in cancer. Asignificant decrease in the expression of CHKA, at both the levelof protein (62 ± 5% of the control, p = 0.002, n = 3; Fig. 4B) andmRNA (65 ± 8% of the control, p = 0.005, Fig. 4C) was observed intreated PC3 cells. In addition, a decrease in the mRNA levels ofthe SLC44A1 choline transporter was observed (74 ± 3% of the

control, p = 0.001, Fig. 4C), although the change in SLC44A1 pro-tein levels did not reach statistical significance (data not shown).

Treatment-induced modulation of metabolism in MCF-7 cells

To further explore whether the metabolic modulations inducedby treatment with U0126 in PC3 prostate cancer cells could beextended to other cancer models, we investigated the effectsof drug treatment on metabolism in MCF-7 breast cancer cells.Cells were treated with 25 μM U0126, which induced a decreasein cell number to 52 ± 7% of the control after 48 h.

We have shown previously that, following treatment withU0126, intracellular lactate accumulation increases to 206 ± 42%of the control (p = 0.001, n = 5) (11), in line with our findings in

Figure 4. Effect of treatment with U0126 on phosphocholine (PC) metabolism. (A) Rate of de novo PC synthesis (fmol/cell/h) in live PC3 cells. (B) Cho-line kinase α (CHKA) protein levels in control (C) and U0126-treated (U) cells (β-actin shown as loading control). (C) U0126-induced changes in mRNAlevels of CHKA and the choline transporter SLC44A1. *p < 0.05; **p ≤ 0.005.

Figure 5. Effect of treatment with U0126 in MCF-7 cells. (A) Metabolic rate of glucose consumption and lactate production in untreated (black) andU0126-treated (white) MCF-7 cells. (B) ADP/ATP levels. (C) p-AMPK (AMP-activated protein kinase) and choline kinase α (CHKA) protein levels as probedby Western blotting. (D) Intracellular concentration of PC. (E) Rate of de novo PC synthesis in live cells. *p < 0.05; **p < 0.005.



705

PC3 cells. Consistent with this observation, and similar to ourfindings in PC3 cells, 13C MRS of MCF-7 cells perfused with 1-13C-glucose-containingmedium also showed elevated rates of glucoseconsumption (up to 192% of the control from 362 ± 125 to 696 ±239 fmol/cell/h, p = 0.02, n = 8) and lactate production (up to212% of the control from 295 ± 99 to 625 ± 230 fmol/cell/h,p = 0.02) in cells receiving U0126 treatment (Fig. 5A). When consid-ering themechanism throughwhich this effect could bemediated,the mRNA levels of HK2 increased significantly (404 ± 241% of thecontrol, p = 0.006, n = 4), but, in contrast with our findings in pros-tate cancer cells, we did not observe any significant changes in p-Akt levels (n = 4). However, similar to the results obtained in PC3cells, profiling of 31P MRS-detectable metabolites indicated thatthe intracellular levels of ATP decreased following treatment(from 10.1 ± 1.9 fmol/cell in the control to 5.5 ± 1.3 fmol/cellin treated cells, n = 5), resulting in an increase in ADP/ATP (from0.24 ± 0.03 to 0.40 ± 0.02, p = 0.002; Fig. 5B). Moreover, p-AMPKlevels increased significantly to 178 ± 42% of the control (p =0.01) in treated MCF-7 cells (Fig. 5C).

Metabolic profiling using 1H and 31P MRS also indicated thatintracellular PC levels decreased significantly following treat-ment to approximately 50% (p < 0.05) of the control (Fig. 5D;PC decreased from 19.5 ± 3.8 to 9.9 ± 3.5 fmol/cell and from21.0 ± 4.7 to 10.4 ± 4.6 fmol/cell based on the 1H and 31P spectra,respectively). The rate of de novo PC synthesis also decreased to47% of the control (from 1.8 ± 0.8 to 0.8 ± 0.3 fmol/cell/h, p =0.02) in U0126-treated MCF-7 cells (Fig. 5E), as did protein levelsof CHKA (63 ± 12% of the control, p = 0.04; Fig. 5C), although thechange in CHKA mRNA levels was not significant. No significantchange was observed in the expression of SLC44A1.

Treatment-induced modulation of metabolism in A375 cells

In the case of melanoma cells, inhibition of the MAPK pathwayby treatment with B-Raf and B-Raf/VEGFR-2 inhibitors has beenshown to induce a substantial decrease in FDG-PET-detectedglucose uptake in the A375 model which harbors the V600EBRAF mutation (28,29). Furthermore, a 1H MRS study showedthat extracellular lactate decreases in BRAF-dependent cells,but not in BRAF-independent cells (30). We therefore decidedto investigate the melanoma cell line A375. Treatment with 25

μM U0126 for 48 h induced a decrease in A375melanoma cell num-ber to 49 ± 5% of the control. Using 1H MRS on A375 cellular ex-tracts, we determined that intracellular lactate accumulationdecreased dramatically to 27 ± 4% of the control (p = 0.01, n = 3)in cells treated with U0126 from 27.2 ± 3.9 to 7.1 ± 0.3 fmol/cell(Fig. 6A). A treatment-induced decrease in glucose uptake andlactate production was also confirmed using 1H MRS on the condi-tioned culture medium: glucose uptake from the medium de-creased to 36.0 ± 7.2% of the control, and extracellular lactateaccumulation decreased to 42.3 ± 3.5% of the control. Importantly,levels of AMPK, p-AMPK, Akt and p-Akt remained unchanged(Fig. 6B, n = 3), highlighting the fact that the decrease in glycolysisobserved in A375 cells, in contrast with the activation of glycolysisobserved in PC3 and MCF-7 cells, reflects a range of different mo-lecular consequences associated with MAPK signal inhibition inBRAF-dependent cells, as described previously by others (30).In the case of PC, a treatment-induced decrease to 18% of the

control (from 10.0 ± 1.1 to 1.8 ± 0.1 fmol/cell, p = 0.005) was ob-served, in line with our results in PC3 and MCF-7 cells, as well asprevious reports in breast and colon cancer cells (26). The de-crease in PC was associated with a decrease in CHKA proteinlevels to 46 ± 7% (p = 0.008) of the control (Fig. 6B).

DISCUSSION

Deregulated cell metabolism is increasingly being recognized asa hallmark of cancer. Consequently, novel metabolic biomarkersand methods for their non-invasive imaging are being investi-gated and developed. A possible approach to identify such novelMRS-detectable biomarkers is based on the comprehensive pro-filing of the metabolome to obtain broad-based metabolic signa-tures associated with the presence of disease or response totreatment. Such studies are typically performed on cell extractsor tissue biopsies (31,33,43,44). Here, we used this approachand first investigated the 1H MRS-detectable consequences oftreatment with U0126 on cell extracts. When considering indivi-dual cell lines, the metabolomics signature associated with re-sponse to U0126 treatment differed from that observedpreviously in the same cells following response to treatmentwith the PI3K inhibitor LY294002 or the HSP90 inhibitor 17-AAG (31), highlighting the potential value of metabolomics inthe identification of a metabolic signature that is specifically as-sociated with response to a particular therapy. Such a signaturecould become useful as higher magnetic field strengths areincreasingly being used in the clinic, providing improvedsensitivity and resolution for the detection of a wide range ofmetabolites in vivo. However, currently, the translation ofmetabolomics signatures to the clinical setting using non-inva-sive in vivo MRS is limited by the relatively low sensitivity ofMRS. Nonetheless, the metabolomics approach can serve toidentify translatable biomarker candidates. In our studies,because the modulation of PC and lactate was observed, and be-cause tCho and lactate are the most readily detectable metabo-lites in the 1H spectrum in vivo, we focused on the metabolicpathways associated with these metabolites, and investigatedglycolysis and choline metabolism.When considering glucose metabolism, the metabolic sequelae

of MAPK inhibition that were associated with the inhibition of cellproliferation in our cells varied depending on the cell line investi-gated. We observed an increase in steady-state intracellular lactateand an increase in glycolytic flux and lactate production in PC3

Figure 6. Effect of treatment with U0126 in A375 cells. (A) Lactate peak in-tensity in 1H MR spectra acquired on untreated (black) and U0126-treated(gray) A375 cell extracts. (B) Protein levels of AMPK (AMP-activated proteinkinase), p-AMPK, Akt, p-Akt and choline kinase α (CHKA) (β-actin shown asloading control). C, untreated; U, U0126-treated.

A. LODI ET AL.


706

prostate cancer andMCF-7 breast cancer cells, but a decrease in in-tracellular lactate, glucose uptake and lactate production in A375melanoma cells. Our findings inmelanoma cells are consistent withprevious publications, which reported that, following treatmentwith a BRAF inhibitor, FDG-PET-detectable glucose uptake de-creased in melanoma cells and in tumors harboring the V600EBRAF mutation (28,29). Similarly, an MRS-detectable decrease inlactate levels was observed following MEK inhibition in cells thatwere BRAF dependent (30). Our observations in PC3 and MCF-7cells are consistent with a previous study in which we showed a13C-MRS-detectable increase in the amount of hyperpolarizedlactate produced from exogenous pyruvate in treated PC3 cells rel-ative to controls. Furthermore, PC3 and MCF-7 cells investigated inthis study are wild-type for BRAF, and thus our findings in thesecells are also consistent with previous work showing a decreasein FDG uptake and lactate production following MAPK pathwayinhibition in BRAF mutant cells, but not in BRAF wild-type cells(independent of RAS status), and the finding that FDG uptake in-creased in a BRAF-mutant tumor that had developed resistanceto treatment (28,30).AMPK and PI3K are known to play a significant role in the con-

trol of glucose homeostasis (45). Our findings indicate that, in thecase of PC3 and MCF-7 cells, PI3K and/or AMPK were activatedfollowing MAPK inhibition. Activation of the PI3K pathwayfollowing MAPK inhibition is well established as a feedback path-way that probably contributes to the lack of clinical response inpatients treated with MAPK inhibitors, particularly when their tu-mors do not harbor a BRAF mutation (40). Activation of the PI3Kpathway downstream of MAPK inhibition would be expected toincrease glycolytic flux, as observed in our studies. Activation ofAMPK following MAPK inhibition is less well established. How-ever, depletion of ATP following ERK inhibition has beenreported in macrophages and astrocytes (46,47), and AMPK acti-vation could also be mediated by I kappa beta kinase (IKK),which is up-regulated following MEK inhibition (48,49). Impor-tantly, however, in the A375 melanoma cell line, in which wesaw a decrease in glycolysis, the activation of PI3K and AMPK sig-naling was not observed. Thus, an increase in glycolysis couldpotentially serve as an indicator of the activation of feedbackpathways that limit the efficacy of MAPK inhibitors.In contrast with our findings regarding glycolysis, a decrease

in intracellular PC levels was consistently associated with MEK in-hibition, and was observed in all three cell lines investigated here(PC3, MCF-7 and A375). Our study demonstrated that this effectwas mediated by a decrease in de novo PC synthesis. PC synthe-sis involves the transport of choline into the cell via cholinetransporters, and choline phosphorylation by CHK (20). The de-crease in choline transporter expression observed in PC3 cellscould contribute to the decrease in PC. PC levels could also belower because of a decrease in CHK activity, as the activity of thisenzyme depends on ATP levels, and ATP decreased in ourtreated cells (50). Furthermore, in PC3 and MCF-7 cells, the de-crease in the PC level is similar to that in CHKA, but, in A375 cells,the decrease in PC is much greater than the treatment-inducedmodulation of CHKA, suggesting that other factors are affectedby U0126. However, when considering the data in all three celllines, the consistent finding was a decrease in the expressionof CHKA, and this is probably the most significant factor contrib-uting to our findings. Our findings are also consistent with previ-ous reports showing an MRS-detectable decrease in PC levelsfollowing treatment with U0126 in other breast and colon cancermodels (26), and following treatment with Ras inhibitors in

mouse fibroblasts (24), as well as other studies linking MAPK sig-naling to choline metabolism (20).

Further studies are needed to confirm our findings in vivo.Nonetheless, imaging of PC could serve to provide informationon the inhibition of MAPK signaling. Choline-containing metabo-lites (mainly PC, choline and GPC) are readily detectable in vivoby 1H MRS as the so-called tCho peak (20) and, although a de-crease in PC has also been reported in response to other thera-peutic approaches, tCho could serve as a non-specific indicatorof drug delivery and drug target inhibition following treatmentwith MAPK pathway inhibitors. More significant, however, mightbe the imaging of glycolysis. In BRAF-dependent melanoma, in-hibition of MAPK signaling has been particularly successful, andFDG-PET has proved to be a useful tool for the assessment ofthe response. In other cancer types, the clinical response hasbeen more limited, probably because of multiple active feedbackpathways requiring combination therapies (40,48). Our observa-tions, combined with previous studies, suggest that imaging ofglycolysis, be it by FDG-PET, 1H MRS or hyperpolarized 13C MRSof pyruvate (11,12,28–30), could serve to detect the activationof such feedback pathways, predicting therapeutic resistanceand potentially providing an important tool to improve the im-plementation of personalized cancer care.

Acknowledgement

The authors acknowledge National Institutes of Health (NIH)funding R01 CA130819 and R01 CA154915.

REFERENCES1. Ward PS, Thompson CB. Metabolic reprogramming: a cancer hallmark

even Warburg did not anticipate. Cancer Cell 2012; 21: 297–308.2. Schulze A, Harris AL. How cancer metabolism is tuned for prolifera-

tion and vulnerable to disruption. Nature 2012; 491: 364–373.3. Mullen AR, DeBerardinis RJ. Genetically-defined metabolic

reprogramming in cancer. Trends Endocrinol. Metab. 2012; 23:552–559.

4. Juweid ME, Cheson BD. Positron-emission tomography and assess-ment of cancer therapy. N. Engl. J. Med. 2006; 354: 496–507.

5. Rohren EM, Turkington TG, Coleman RE. Clinical applications of PETin oncology. Radiology 2004; 231: 305–332.

6. Mellon EA, Lee SC, Pickup S, Kim S, Goldstein SC, Floyd TF, Poptani H,Delikatny EJ, Reddy R, and Glickson JD. Detection of lactate with aHadamard slice selected, selective multiple quantum coherence,chemical shift imaging sequence (HDMD-SelMQC-CSI) on a clinicalMRI scanner: application to tumors and muscle ischemia. Magn.Reson. Med. 2009; 62: 1404–1413.

7. Wijnen JP, Van der Graaf M, Scheenen TW, Klomp DW, de Galan BE,Idema AJ, Heerschap A. In vivo 13C magnetic resonance spectros-copy of a human brain tumor after application of 13C-1-enriched glu-cose. Magn. Reson. Imaging 2010; 28: 690–697.

8. de Graaf RA, Mason GF, Patel AB, Behar KL, Rothman DL. In vivo1H-[13C]-NMR spectroscopy of cerebral metabolism. NMR Biomed.2003; 16: 339–357.

9. Chaumeil MM, Ozawa T, Park I, Scott K, James CD, Nelson SJ, RonenSM. Hyperpolarized 13C MR spectroscopic imaging can be used tomonitor Everolimus treatment in vivo in an orthotopic rodent modelof glioblastoma. Neuroimage 2012; 59: 193–201.

10. Dafni H, Larson PEZ, Hu S, Yoshihara HAI, Ward CS, Venkatesh HS, WangC, Zhang X, Vigneron DB, Ronen SM. Hyperpolarized 13C spectroscopicimaging informs on hypoxia-inducible factor-1 and Myc activity down-stream of platelet-derived growth factor receptor. Cancer Res. 2010; 70:7400–7410.

11. Lodi A, Woods SM, Ronen SM. Treatment with the MEK inhibitorU0126 induces decreased hyperpolarized pyruvate to lactate conver-sion in breast but not in prostate cancer cells. NMR Biomed. 2013; 26:299–306.



707

12. Ward CS, Venkatesh HS, Chaumeil MM, Brandes AH, VanCriekinge M,Dafni H, Sukumar S, Nelson SJ, Vigneron DB, Kurhanewicz J, JamesCD, Haas-Kogan DA, Ronen SM. Noninvasive detection of targetmodulation following phosphatidylinositol 3-kinase inhibition usinghyperpolarized 13C magnetic resonance spectroscopy. Cancer Res.2010; 70: 1296–1305.

13. Brindle K. Watching tumours gasp and die with MRI: the promise ofhyperpolarised 13C MR spectroscopic imaging. Br. J. Radiol. 2012; 85:697–708.

14. Nelson S, Kurhanewicz J, Vigneron D, Larson P, Harzstarck A, FerroneM,van Criekinge M, Chang J, Bok R, Park I, Reed G, Carvajal L, Crane J,Ardenkjaer-Larsen J, Chen A, Hurd R, Odegardstuen L, Tropp J. Proofof concept clinical trial of hyperpolarized C-13 pyruvate in patientswithprostate cancer. Proceedings of the 20th Annual Meeting and ExhibitionISMRM, Melbourne, Vic., Australia, 2012; 274.

15. Day S, Kettunen M, Gallagher F, Hu D-E, Lerche M, Wolber J, GolmanK, Ardenkjaer-Larsen J, Brindle K. Detecting tumor response to treat-ment using hyperpolarized 13C magnetic resonance imaging andspectroscopy. Nat. Med. 2007; 13: 1382–1387.

16. Kohler S, Yen Y, Wolber J, Chen A, Albers M, Bok R, Zhang V, Tropp J,Nelson S, Vigneron D, Kurhanewicz J, Hurd R. In vivo 13 carbon met-abolic imaging at 3T with hyperpolarized 13C-1-pyruvate. Magn.Reson. Med. 2007; 58: 65–69.

17. Witney T, Kettunen M, Day S, Hu D, Neves A, Gallagher F, Fulton S,Brindle K. A comparison between radiolabeled fluorodeoxyglucose up-take and hyperpolarized 13C-labeled pyruvate utilization as methodsfor detecting tumor response to treatment. Neoplasia 2009; 11: 574–582.

18. Day S, Kettunen M, Krishna M, Mitchell J, Lizak M, Morris H,Matsumoto S, Koretsky A, Brindle K. Detecting response of rat C6 gli-oma tumors to radiotherapy using hyperpolarized [1-13C] pyruvateand 13C magnetic resonance spectroscopic imaging. Magn. Reson.Med. 2011; 65: 557–563.

19. Park I, Bok R, Ozawa T, Phillips J, James C, Vigneron D, Ronen S,Nelson S. Detection of early response to temozolomide treatmentin brain tumors using hyperpolarized 13C MR metabolic imaging.J. Magn. Reson. Imaging 2011; 33: 1284–1290.

20. Glunde K, Bhujwalla Z, Ronen S. Choline metabolism in malignanttransformation. Nat. Rev. Cancer 2011; 11: 835–848.

21. Rice SL, Roney CA, Daumar P, Lewis J. The next generation of posi-tron emission tomography radiopharmaceuticals in oncology.Semin. Nucl. Med. 2011; 41: 265–282.

22. MonazzamA, Razifar P, Ide S, Rugaard JensenM, Josephsson R, BlomqvistC, Langström B, Bergström M. Evaluation of the Hsp90 inhibitorNVP-AUY922 in multicellular tumour spheroids with respect to effectson growth and PET tracer uptake. Nucl. Med. Biol. 2009; 36: 335–342.

23. Nimmagadda S, Glunde K, Pomper M, Bhujwalla Z. Pharmacody-namic markers for choline kinase down-regulation in breast cancercells. Neoplasia 2009; 11: 477–484.

24. Ronen SM, Jackson LE, Beloueche M, Leach MO. Magnetic resonancedetects changes in phosphocholine associated with Ras activationand inhibition in NIH 3T3 cells. Br. J. Cancer 2001; 84: 691–696.

25. Beloueche-Babari M, Jackson LE, Al-Saffar NM, Eccles SA, Raynaud FI,Workman P, Leach MO, Ronen SM. Identification of magnetic reso-nance detectable metabolic changes associated with inhibition ofphosphoinositide 3-kinase signaling in human breast cancer cells.Mol. Cancer Ther. 2006; 5: 187–196.

26. Beloueche-Babari M, Jackson LE, Al-Saffar NM, Workman P, LeachMO, Ronen SM. Magnetic resonance spectroscopy monitoring ofmitogen-activated protein kinase signaling inhibition. Cancer Res.2005; 65: 3356–3363.

27. Venkatesh HS, Chaumeil MM, Ward CS, Haas-Kogan DA, James CD,Ronen SM. Reduced phosphocholine and hyperpolarized lactate pro-vide magnetic resonance biomarkers of PI3K/Akt/mTOR inhibition inglioblastoma. Neuro-Oncol. 2012; 14: 315–325.

28. Baudy A, Dogan T, Flores-Mercado J, Hoeflich K, Su F, van Bruggen N,Williams S. FDG-PET is a good biomarker of both early response and ac-quired resistance in BRAFV600 mutant melanomas treated withvemurafenib and theMEK inhibitor GDC-0973. EJNMMI Res. 2012; 2: 22.

29. Tseng J, Stuart D, Aardalen K, Kaplan A, Aziz N, Hughes NP, GambhirSS. Use of DNA microarray and small animal positron emission to-mography in preclinical drug evaluation of RAF265, a novel B-Raf/VEGFR-2 inhibitor. Neoplasia 2011; 13: 266–275.

30. Falck Miniotis M, Arunan V, Eykyn TR, Marais R, Workman P, LeachMO, Beloueche-Babari M. MEK1/2 inhibition decreases lactate inBRAF-driven human cancer cells. Cancer Res. 2013; 73: 4039–4049.

31. Lodi A, Ronen SM. Magnetic resonance spectroscopy detectablemetabolomic fingerprint of response to antineoplastic treatment.PLoS One 2011; 6: e26155.

32. Brandes A, Ward C, Ronen SM. 17-allyamino-17-demethoxygelda-namycin treatment results in a magnetic resonance spectroscopy-detectable elevation in choline-containing metabolites associatedwith increased expression of choline transporter SLC44A1 and phos-pholipase A2. Breast Cancer Res. 2010; 12: R84.

33. Su JS, Woods SM, Ronen SM. Metabolic consequences of treatmentwith AKT inhibitor perifosine in breast cancer cells. NMR Biomed.2012; 25: 379–388.

34. Hwang TL, Shaka AJ. Water suppression that works – excitationsculpting using arbitrary wave-forms and pulsed-field gradients.J. Magn. Reson. 1995; 112: 275–279.

35. Ronen SM, Rushkin E, Degani H. Lipid metabolism in T47D humanbreast cancer cells: 31P and 13C-NMR studies of choline and ethanol-amine uptake. Biochim. Biophys. Acta 1991; 1095: 5–16.

36. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarraysapplied to the ionizing radiation response. Proc Natl Acad Sci USA2001; 98: 5116–5121.

37. Dang CV. Role of aerobic glycolysis in genetically engineered mousemodels of cancer. BMC Biol. 2013; 11: 3.

38. Plas D, Thompson C. Akt-dependent transformation: there is more togrowth than just surviving. Oncogene 2005; 24: 7435–7442.

39. Hoeflich KP, O’Brien C, Boyd Z, Cavet G, Guerrero S, Jung K, JanuarioT, Savage H, Punnoose E, Truong T, Zhou W, Berry L, Murray L, AmlerL, Belvin M, Friedman LS, Lackner MR. In vivo antitumor activity ofMEK and phosphatidylinositol 3-kinase inhibitors in basal-like breastcancer models. Clin. Cancer Res. 2009; 15: 4649–4664.

40. Pratilas CA, Solit DB. Targeting the mitogen-activated protein kinasepathway: physiological feedback and drug response. Clin. CancerRes. 2010; 16: 3329–3334.

41. Hardie DG. AMP-activated/SNF1 protein kinases: conserved guard-ians of cellular energy. Nat. Rev. Mol. Cell Biol. 2007; 8: 774–785.

42. Mihaylova MM, Shaw RJ. The AMPK signalling pathway coordinatescell growth, autophagy and metabolism. Nat. Cell Biol. 2011; 13:1016–1023.

43. Bathen TF, Sitter B, Sjøbakk TE, Tessem MB, Gribbestad IS. Magneticresonance metabolomics of intact tissue: a biotechnological tool incancer diagnostics and treatment evaluation. Cancer Res. 2010; 70:6692–6696.

44. Swanson M, Zektzer A, Tabatabai Z, Simko J, Jarso S, Keshari K,Schmitt L, Carroll P, Shinohara K, Vigneron D, Kurhanewicz J. Quanti-tative analysis of prostate metabolites using 1H HR-MAS spectros-copy. Magn. Res. Med. 2006; 55: 1257–1264.

45. Schultze SM, Hemmings BA, Niessen M, Tschopp O. PI3K/AKT, MAPKand AMPK signalling: protein kinases in glucose homeostasis. ExpertRev. Mol. Med. 2012; 14: e1.

46. Yung HW, Wyttenbach A, Tolkovsky AM. Aggravation of necrotic deathof glucose-deprived cells by the MEK1 inhibitors U0126 and PD184161through depletion of ATP. Biochem. Pharmacol. 2004; 68: 351–360.

47. Monick MM, Powers LS, Barrett CW, Hinde S, Ashare A, Groskreutz DJ,Nyunoya T, Coleman M, Spitz DR, Hunninghake GW. Constitutive ERKMAPK activity regulates macrophage ATP production and mitochon-drial integrity. J. Immunol. 2008; 180: 7485–7496.

48. Gioeli D, Wunderlich W, Sebolt-Leopold J, Bekiranov S, Wulfkuhle JD,Petricoin EF, 3rd, Conaway M, Weber MJ. Compensatory pathwaysinduced by MEK inhibition are effective drug targets for combinationtherapy against castration-resistant prostate cancer. Mol. CancerTher. 2011; 10: 1581–1590.

49. Criollo A, Senovilla L, Authier H, Maiuri MC, Morselli E, Vitale I, KeppO, Tasdemir E, Galluzzi L, Shen S, Tailler M, Delahaye N, Tesniere A,De Stefano D, Younes AB, Harper F, Pierron G, Lavandero S, ZitvogelL, Israel A, Baud V, Kroemer G. The IKK complex contributes to the in-duction of autophagy. EMBO J. 2010; 29: 619–631.

50. Jacobs R, Lingrell S, Dyck J, Vance D. Inhibition of hepatic phospha-tidylcholine synthesis by 5-aminoimidazole-4-carboxamide-1-b-4-ribofuranoside is independent of AMP-activated protein kinaseactivation. J. Biol. Chem. 2007; 282: 4516–4523.

SUPPORTING INFORMATIONAdditional supporting information may be found in the onlineversion of this article at the publisher’s web site.

A. LODI ET AL.


708

mr-detectable metabolic consequences of mitogen-activated protein kinase kinase (mek) inhibition

Documents