Mevalonate Pathway Provides Ubiquinone to Maintain ... · pyrimidine nucleotide biosynthesis and induced oxidative stress and apoptosis in p53-deﬁcient cancer cells in spheroid

CANCER RESEARCH | METABOLISM AND CHEMICAL BIOLOGY

Mevalonate Pathway Provides Ubiquinone to MaintainPyrimidine Synthesis and Survival in p53-DeficientCancer Cells Exposed to Metabolic Stress A C

Irem Kaymak1, Carina R. Maier1, Werner Schmitz1, Andrew D. Campbell2, Beatrice Dankworth1,Carsten P. Ade1, Susanne Walz3, Madelon Paauwe2, Charis Kalogirou4, Hecham Marouf1,Mathias T. Rosenfeldt5,6, David M. Gay2,7, Grace H. McGregor2,7, Owen J. Sansom2, and Almut Schulze1,6

ABSTRACT◥

Oncogene activation and loss of tumor suppressor functionchanges the metabolic activity of cancer cells to drive unrestrictedproliferation. Moreover, cancer cells adapt their metabolism tosustain growth and survival when access to oxygen and nutrientsis restricted, such as in poorly vascularized tumor areas. We showhere that p53-deficient colon cancer cells exposed to tumor-likemetabolic stress in spheroid culture activated the mevalonatepathway to promote the synthesis of ubiquinone. This was essentialto maintain mitochondrial electron transport for respiration andpyrimidine synthesis in metabolically compromised environments.Induction of mevalonate pathway enzyme expression in theabsence of p53 was mediated by accumulation and stabilization ofmature SREBP2. Mevalonate pathway inhibition by statins blockedpyrimidine nucleotide biosynthesis and induced oxidative stressand apoptosis in p53-deficient cancer cells in spheroid culture.

Moreover, ubiquinone produced by the mevalonate pathway wasessential for the growth of p53-deficient tumor organoids. Incontrast, inhibition of intestinal hyperproliferation by statins in anApc/KrasG12D-mutant mouse model was independent of de novopyrimidine synthesis. Our results highlight the importance of themevalonate pathway for maintaining mitochondrial electron trans-fer and biosynthetic activity in cancer cells exposed to metabolicstress. They also demonstrate that the metabolic output of thispathway depends on both genetic and environmental context.

Significance: These findings suggest that p53-deficient cancercells activate the mevalonate pathway via SREBP2 and promotethe synthesis of ubiquinone that plays an essential role inreducing oxidative stress and supports the synthesis of pyrimidinenucleotide.

IntroductionThe metabolic activity of cancer cells is controlled by genetic

alterations and by the tumor microenvironment. Under metabolicstress, defined by reduced access to nutrients and oxygen presentin poorly vascularized solid tumors, cancer cells need to adapttheir metabolic activity to maintain cell proliferation and survival.One important factor in the adaptation to metabolic stress is thehypoxia inducible factor (HIF), which is stabilized and activatedin the absence of oxygen, and promotes the uptake of glucose andits fermentation to lactate while reducing oxidative metabo-lism (1). However, poor access to the vascular network not only

reduces oxygen tension but also lowers the availability of serum-derived nutrients. Therefore, cancer cells need to undergo globalrewiring of their metabolic activity to be able to adapt to theseconditions.

The p53 tumor suppressor is a master regulator of cellular metab-olism (2, 3). It reduces glucose uptake (4) and alters glycolysis andmodulates the flux of metabolites into the pentose phosphatepathway (5–8). Conversely, p53 enhances mitochondrial metabolismby promoting the assembly of cytochrome c oxidase (complex IV) andincreasing respiration (9). It has been shown that p53 allows cancercells to adapt to nutrient deprivation, in particular, the absence ofthe amino acid serine and glutamine (10, 11). Thus, loss of p53function can increase the sensitivity of cancer cells toward metabolicstress, resulting in a selective vulnerability that could be exploitedtherapeutically.

In this study, we have investigated the role of p53 in the regulation ofmetabolic processes in colon cancer cells exposed to metabolic stress.To recreate the simultaneous reduction in oxygen and nutrientavailability found in tumors, we cultured cancer cells as multicellulartumor spheroids. Under these conditions, we find that p53-deficientcancer cells activate the expression of enzymes of the mevalonatepathway via the sterol regulatory element binding protein 2 (SREBP2).Moreover, inhibition of mevalonate pathway activity with statinsselectively induced apoptosis in p53-deficient cancer cells exposed tometabolic stress. This effect was mediated by reduced generation ofubiquinone (CoQ10), which p53-deficient cells require to maintaintricarboxylic acid (TCA) cycle activity, respiration, and the synthesis ofpyrimidine nucleotides. Our study thus reveals a novel link betweenthe regulation of isoprenoid synthesis and the modulation of electrontransfer mediated by ubiquinone in cancer cells. Mevalonate pathway

1Theodor-Boveri-Institute, Biocenter, W€urzburg, Germany. 2Cancer ResearchUKBeatson Institute, Glasgow,UnitedKingdom. 3ComprehensiveCancerCenterMainfranken, Core Unit Bioinformatics, Biocenter, University of W€urzburg,W€urzburg, Germany. 4Department of Urology, University Hospital W€urzburg,W€urzburg, Germany. 5Department of Pathology, University Hospital W€urzburg,W€urzburg, Germany. 6Comprehensive Cancer Center Mainfranken, W€urzburg,Germany. 7Institute of Cancer Sciences, University of Glasgow, Glasgow, UnitedKingdom.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

Corresponding Author: Almut Schulze, Division of Tumor Metabolism andMicroenvironment, German Cancer Research Center, Heidelberg, Germany.Phone: 4962-2142-3423; Fax: 4962-2142-3467; E-mail:[email protected]

Cancer Res 2020;80:189–203

doi: 10.1158/0008-5472.CAN-19-0650

�2019 American Association for Cancer Research.

AACRJournals.org | 189

on June 22, 2020. © 2020 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 19, 2019; DOI: 10.1158/0008-5472.CAN-19-0650

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-0650&domain=pdf&date_stamp=2019-12-27

http://crossmark.crossref.org/dialog/?doi=10.1158/0008-5472.CAN-19-0650&domain=pdf&date_stamp=2019-12-27

http://cancerres.aacrjournals.org/

activity is essential for p53-deficient cancer cells to proliferateand survive under the metabolic constraints of the tumormicroenvironment.

Materials and MethodsTissue culture and reagents

HCT116 p53-isogenic cells were obtained fromB.Vogelstein (JohnsHopkins University, Baltimore, MD) and HCT116 p21-isogeniccells from M. Dobbelstein (Georg-August University, G€ottingen,Germany). RKO p53-isogenic lines were a gift from K. Vousden(Beatson Institute, Glasgow, United Kingdom). All other cell lineswere from CRUK LRI Research Services, authenticated by shorttandem repeat profiling and used at low passage. Unless statedotherwise, cells were cultured in DMEM with 10% FBS (Gibco),4 mmol/L L-glutamine, and 1% penicillin–streptomycin, at 37�C ina humidified incubator at 5% CO2 and regularly tested for absenceof Mycoplasma. Etoposide, (R)-Mevalonic acid lithium salt,SB216732, CHIR99021, simvastatin, zoledronic acid monohydrate,coenzyme Q10, NAC, water-soluble cholesterol, uridine, lefluno-mide, and 5-FU were all from Sigma. MG132 and MK2206 werefrom Bertin Pharma, rapamycin from Cayman Chemicals, mevas-tatin and YM-53601 from Biomol, and nucleosides (EmbryoMax100�) from Merck-Milipore.

Spheroid formation, flow cytometry, and histologyFor spheroid formation, 10,000 cells/well were placed in 96-well

ultralow attachment plates (Corning CORN7007) followed by centri-fugation at 850 � g for 10 minutes. Spheroids were cultured for12–14 days, during which, medium was replaced every three days.

Monolayer and spheroid cells were incubated with 20 mmol/L BrdU(Sigma) for 24 hours, trypsinized, and fixed in 80% EtOH. Cellsincubated in 2 mol/L HCl with 0.5% Triton X-100 for 30 minutes atroom temperature, neutralized with Na2B4O7, and incubated withanti-BrdU-FITC antibodies (Biozol). Cells were washed and treatedwith RNAse A (24 mg/mL) and propidium iodide (54 mmol/L) for 30minutes. Analysis was performed on a BD FACSCanto II usingFACSDIVA software.

Spheroids were fixed with 3.7% paraformaldehyde, incubated in70% ethanol for 16 hours, mixed with low-melting agarose, andparaffin embedded. Four-micron–thick sections were deparaffinizedand rehydrated. Antigen retrieval was performed with citrate buffer(pH 6.0) in amicrowave oven for 6minutes. Sections were stainedwithanti-Ki67 (SP6, Thermo Fisher Scientific) and anti-cleaved caspase 3(Cell Signaling Technology) in PBS/1% BSA at 4�C and biotinylatedsecondary antibody (Vector Laboratories). Slides were developed with3,30-diaminobenzidine (Cell Signaling Technology) and counter-stained with Gilmore 3 hematoxylin. For TUNEL staining, sectionswere heated in citrate buffer (pH 6.0) for 2 minutes. TUNEL reactionswere developed for 1 hour (In Situ Cell Death Detection Kit, Sigma)and counterstained with Hoechst (Sigma). Archival tumor tissue (8)was stained with anti-Ki67 as above.

RNA sequencingRNA was extracted using RNeasy columns (Qiagen) including

DNase I digestion. mRNA was isolated using NEBNext Poly(A)mRNA Magnetic Isolation Module and library preparation was per-formed with NEBNext Ultra RNA Library Prep Kit for Illuminafollowing the manufacturer's instructions. Libraries were size-selected using Agencourt AMPure XP Beads (Beckman Coulter),followed by amplification with 12 PCR cycles. Library quantification

and size determination was performed with an Experion system (Bio-Rad) and libraries were sequenced with NextSeq500 (Illumina).

RNA sequencing (RNA-seq) data were analyzed as described in theSupplementary Information and are available at Gene ExpressionOmnibus (GSE124189).

RNA extraction and qRT-PCRTotal RNA was isolated using PeqGOLD Trifast, followed by

reverse transcription into cDNA usingM-MLV Reverse Transcriptase(Promega) and random hexamer primers. Real-time PCR wasperformed using Power-up SYBR Green Master Mix (Thermo FisherScientific) using Quantitect primers (Qiagen) or custom primers asfollows: human ACTB forward 50-GCCTCGCCTTTGCCGAT-30 andreverse 50-CGCGGCGATATCATCATCC-30; and human CDKN1Aforward 50-TCACTGTCTTGTACCCTTGTGC-30 and reverse 50-CGTTTGGAGTGGTAGAAA-30 (Sigma). All qPCR reactions wereperformed in technical duplicate on three biologically independentreplicate samples. Relative mRNA amounts were calculated using thecomparative Ct method after normalization to b-actin (ACTB).

Western blottingCells were lysed in lysis buffer (1% Triton X100, 50mmol/L Tris pH

7.5, 300 mmol/L NaCl, 1 mmol/L EGTA, 1 mmol/L DTT, 1 mmol/LNaVO4 with protease inhibitors for 30 minutes) for 30 minutes,cleared by centrifugation, and quantified using BCA assay (ThermoFisher). Nuclear extraction of SREBP2 was performed as describedpreviously (12). Proteins were separated on SDS-PAGE and blottedonto polyvinylidene difluoride membrane (Immobilon), blockedwith blocking solution (LI-COR), and incubated with primary andsecondary antibodies. Signals were detected on an Odyssey scannerand quantified using ImageJ. The following antibodies were used:p53 (DO-1), p21 (C-19), CCND1 (DSC-6; from Santa Cruz Bio-technology), HMGCS-1 (#ab155787), histone-3 (#ab1791, fromAbcam), SREBP-2 (1D2), ABCA1 (from Novus Biologicals),SREBP-2 (R&D Systems), GSK3 (4G-1E; from Milipore), PDK1,ACSS2, p-GSK3a/b (Ser21/9), S6 (5G10), p-S6 (Ser240/244), AKT,p-AKT (Ser473), PARP (from Cell Signaling Technology), b-actin(AC-15), FDFT1, and vinculin (from Sigma). Secondary antibodieswere from LI-COR Biosciences.

Stable isotope labeling and mass spectrometryMonolayer cells or spheroids were washed with PBS and medium

was replaced with either complete medium or glucose-free mediumwith 25 mmol/L [U13C]-glucose (Cambridge Isotope Laboratories).Cells were incubated for the indicated times, washed with cold154 mmol/L ammonium acetate, and snap frozen. For tissue extrac-tion, 150mgof frozen tissuewas homogenized in 3mLofH2Ousing anUltraTurrax.

For water-soluble metabolites, samples were extracted with ice-coldMeOH/H2O (80/20, v/v) containing 0.1 mmol/L lamivudine (Sigma)and separated by centrifugation. Supernatants were transferred to aStrata C18-E column (Phenomenex) that has been conditioned with1 mL of CH3CN and 1mL of MeOH/H2O (80/20, v/v). The eluate wasdried and dissolved in 50 mL of a mixture of CH3CN and 5 mmol/LNH4OAc (25/75, v/v).

For cholesterol and ubiquinone, samples were extracted with ice-coldMeOH/H2O (80/20, v/v) containing 1 mmol/L CoQ9 (Sigma) andseparated by centrifugation. Supernatants were extracted twice with0.4 mL of hexane, collected and taken to dryness under nitrogen at35�C. Samples were dissolved in 150 mL of hexane and transferred toStrata SI-1 columns (Phenomenex), washed with 750 mL hexane, and

Kaymak et al.

Cancer Res; 80(2) January 15, 2020 CANCER RESEARCH190




500 mL hexane/acetic acid ethyl ether (18/1 v/v). Ubiquinone waseluted with 0.5 mL hexane/acetic acid ethylester (9/1, v/v). Cholesterolwas fully eluted with 0.5 mL hexane/acetic acid ethylester (9/1, v/v).Eluates were dried under nitrogen at 35�C and dissolved in 50 mLiPrOH.

Metabolites were analyzed by LC-MS using setting provided inSupplementary Methods.

Seahorse assaysSpheroids were washed twice and transferred to XFe96 Spheroid

Microplates containing 160 mL of Seahorse XF Assay Mediumsupplemented with 25mmol/L D-glucose and 10 mmol/L sodiumpyruvate at pH 7.4. Oxygen consumption rates (OCR) were deter-mined using an XF96e Extracellular Flux Analyzer (Software Ver-sion 1.4; Agilent) following manufacturer's protocol. During theexperiment, 2 mmol/L oligomycin (Merck-Milipore), 0.5 mmol/LFCCP (Sigma), and 1 mmol/L rotenone/antimycin A (Sigma) wereinjected. OCR of at least 14 biologically independent samples wasnormalized to spheroid area.

Organoid cultureMouse small intestines were isolated from wild-type, VillinCreER

Apcfl/fl, or VillinCreERApcfl/flKrasG12D/þ mice sacrificed 3 days post-induction with tamoxifen, opened longitudinally, and washed withPBS. Crypts were isolated as described previously (13), mixed with20 mL Matrigel (BD Biosciences), and plated in 24-well plates inAdvanced DMEM/F12 (Thermo Fisher Scientific) supplementedwith 1% penicillin/streptomycin, 10 mmol/L HEPES, 2 mmol/Lglutamine, N2 (Thermo Fisher Scientific), B27 (Thermo FisherScientific), 100 ng/mL Noggin, and 50 ng/mL EGF (both fromPeproTech). Growth factors were added every two days. Experimentswere performed on two biologically independent samples. Genotypingwas performed using the following primers: p53fl/fl 50-GAAGACA-GAAAAGGGGAGGG-30 and 50-AAGGGGTATGAGGGACAAGG-30; KrasG12D 50-GTCTTTCCCCAGCACAGTGC-30, 50-CTCTTGCC-TACGCCACCAGCTC-30 and 50-AGCTAGCCACCATGGCTTGA-GTAAGT CTGCA-30.

MiceAll animal experiments were performed under UK Home Office

guidelines using project licenses 70-8645 or 70-8646. Experimentalprotocols were subject to theUniversity of Glasgow animal welfare andethical review board approval. VillinCreER Apcfl/fl and VillinCreER

Apcfl/fl KrasG12D/þ mice have been described previously (14). Forinduction of intestinal hyperproliferation, mice were given a singleintraperitoneal injection of 80 mg/kg tamoxifen on one occasion(VillinCreERApcfl/fl KrasG12D/þ), or on two consecutive days (Villin-CreERApcfl/fl). Mice were treated with a daily dose of 50 mg/kgsimvastatin in 0.5% methylcellulose/5% DMSO or vehicle or a dailydose of 35mg/kg leflunomide in 100 mL 0.15% carboxymethylcellulosevia oral gavage from one day post initial tamoxifen injection. For 2H2Otracing, mice were exposed to 8% 2H2O in their drinking water for4 days. Mice were given an intraperitoneal injection of 250 mL of cellproliferation reagent (RPN201, GE Healthcare/Amersham) 2 hoursprior to sacrifice and tissue sections were stained for BrdU as describedin Supplementary Methods.

Statistical analysisStatistical details for each experiment are stated in the figure

legends. Graphs were generated using GraphPad Prism 6.0 (GraphPadsoftware). Unless otherwise indicated, statistical significance wascalculated using the unpaired two-tailed Student t test.

ResultsSpheroid culture induces tumor-like transcriptional signaturesand leads to complex metabolic reprogramming

To determine the influence of conditions of the tumor microenvi-ronment on colon cancer cells, we used isogenic HCT116 lines that areeither wild-type (wt) for p53 or an isogenic derivative in which theTP53 gene had been deleted by homologous recombination (15). Thesecells do not express detectable levels of p53 and fail to induce p21 upontreatment with the DNA-damaging agent etoposide (SupplementaryFig. S1A). Both cell lines were cultured either as subconfluent mono-layers (MLC) for 48 hours or as large three-dimensional tumorspheroids (diameter >600 mm), thereby exposing cancer cells togradients of oxygen and nutrient depletion (16). Spheroid cultures(SPC) showed an overall reduction in proliferation compared withMLC, which was similar in both genotypes (Fig. 1A). However,staining for the proliferation marker Ki67 revealed that p53 wt SPCshow proliferation only in the outer regions (Fig. 1B), while SPC ofp53-deficient cells present Ki67 positivity throughout their cross-sections (Fig. 1B). Similarly, subcutaneous xenograft colon tumorsformed by p53 wt HCT116 cells displayed more heterogenous pro-liferation patterns compared with their p53-deficient counterparts(Fig. 1C). This suggests that p53 is required for cell-cycle arrestinduced by the nutrient and oxygen-depleted conditions found inSPC and tumors.

We next performed transcriptome analysis of p53 wt and deficientcells cultured as SPC,MLC, or xenograft tumors. Principal componentanalysis showed that global gene expression in SPC is more similar tothose in tumors rather than MLC (mainly in PC1 accounting for 80%of variance; Fig. 1D). Gene set enrichment analysis (GSEA) revealedreduced proliferation (Hallmark_E2F_targets) and induction of inter-feron response and hypoxia signatures as major transcriptional phe-notypes in both SPC and tumors compared withMLC (Fig. 1E and F).Moreover, analysis of the cell-cycle regulator cyclin D1 (CCND1) andthe HIF target pyruvate dehydrogenase kinase (PDK1) confirmedreduced proliferation and induction of hypoxia in SPC comparedwith MLC (Fig. 1G).

We next performed metabolomic analysis to determine differencesin metabolism between genotypes and culture conditions (Supple-mentary Fig. S1B). Stable isotope tracing using [U13C]-glucose showedthat SPC increases glucose-dependent lactate synthesis in both p53 wtand deficient cells (Fig. 1H). Time-course experiments revealed thatthe labeling of TCA cyclemetabolites as well as alanine, glutamate, andaspartate reached steady state more rapidly in SPC compared withMLC, as maximal labeling was reached much earlier (SupplementaryFig. S1C–S1E). Moreover, fractions of labeled metabolites werereduced, indicating that the contribution of precursors other thanglucose, most likely glutamine, to the TCA cycle is higher in SPCcompared withMLC.We also found evidence for pyruvate-dependentanaplerosis, asMþ3 isotopologues for succinate, fumarate, andmalatewere formed more rapidly in SPC compared with MLC (Supplemen-tary Fig. S1D). This pyruvate-dependent anaplerosis supported theproduction of aspartate, indicated by the high Mþ3 to Mþ2 ratio foraspartate in SPC (Fig. 1I).

While most metabolic differences between MLC and SPC werefound in both genotypes, the total levels of aspartate were higher inSPC from p53 wt cells compared with p53-deficient SPC and alsocompared withMLC (Fig. 1J). Aspartate is a precursor for pyrimidinenucleotide synthesis and thus essential for proliferation (17). Consis-tently, glucose-derived labeling of uridine monophosphate (UMP), acentral metabolite in pyrimidine biosynthesis, while overall reduced

Mevalonate Pathway Supports Ubiquinone Synthesis in Cancer

AACRJournals.org Cancer Res; 80(2) January 15, 2020 191




I

p53+

/+

p53+

/+

p53−

/−

p53−

/−

p53+

/+

p53+

/+

p53−

/−

p53−

/−

p53+

/+

p53+

/+

p53−

/−

p53−

/−0

200

400

600

800Lactate

M+0M+1M+2M+3

Pea

k in

tens

ity/p

rot

SPCMLC SPCMLC SPCMLC

0.0

0.2

PC

2 (1

8%)

-0.2

PC1 (80%)

Spheroids

Tumors

Monolayer

p53NULLp53WTp53NULLp53WTp53NULLp53WT

MLC

SPC

T

A

J

B

0.0

0.5

1.0

1.5

2.0

M+3

/M+2

Aspartate

0

50

100

150

200

250

Aspartate

M+0M+1M+2M+3M+4

Pea

k in

tens

ity/p

rot

E F

p53+

/+

p53−

/−

p53+

/+

p53−

/−

0

50

100%

Brd

U In

corp

orat

ion

D

MLC SPC

HALLMARK_E2F_TARGETS

MLC p53+/+

0.00.10.20.30.40.50.6

Enr

ichm

ent s

core

(ES

)

0.70.8

NES = 3.61q ≤ 0.05

MLC SPCp53+/+

0.0-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

BROWNE_INTERFERON_RESP_GENES

-0.6 NES = -3.22q ≤ 0.05

-0.7-0.8

MANALO_HYPOXIA_UP0.0

-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

MLC SPC

-0.6

p53+/+

-0.7

NES = -3.12q ≤ 0.05

MANALO_HYPOXIA_UP0.0

-0.1

-0.2

-0.3

-0.4

-0.5Enr

ichm

ent s

core

(ES

)

MLC Tp53+/+

NES = -1.93 q ≤ 0.05

Tp53+/+

0.0-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

-0.6

BROWNE_INTERFERON_RESP_GENES

NES = -2.28q ≤ 0.05

HALLMARK_E2F_TARGETS

MLC T

0.00.10.20.30.40.50.6

Enr

ichm

ent s

core

(ES

)

0.70.8

p53+/+

NES = 5.67q ≤ 0.05

Hp53+/+ p53−/−

MLC SPC

p53+/+ p53−/−

VCL

CCND1

PDK1

G

p53+/+ p53−/−

Ki67 Ki67

150 μm150 μm

Spheroids

50 μm 50 μm

p53+/+ p53−/−

Ki67 Ki67

Xenograft tumorsC

SPCp53−/−

0.0-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

-0.6 NES = -3.47q ≤ 0.05-0.7

-0.8

0.0-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

SPC

-0.6

p53−/−

NES = -3.32q ≤ 0.05

SPC

0.00.10.20.30.40.50.6

Enr

ichm

ent s

core

(ES

) 0.7

p53−/−

NES = 3.0 q ≤ 0.05

T

0.00.10.20.30.40.50.6

Enr

ichm

ent s

core

(ES

)

0.70.8

p53−/−

NES = 5.40q ≤ 0.05

0.0

-0.1

-0.2

-0.3

-0.4

-0.5

Enr

ichm

ent s

core

(ES

)

NES = -1.80 q ≤ 0.05

MLC Tp53−/−

0.0-0.1-0.2-0.3-0.4-0.5

Enr

ichm

ent s

core

(ES

)

-0.6

NES = -2.24 q ≤ 0.05

MLCSPC

MLC

MLCMLC

MLC

34 kDa

49 kDa

124 kDa

MLC Tp53−/−

Figure 1.

Spheroid cultures replicate tumor-like transcriptional profiles and show pyruvate-dependent anaplerosis. A, HCT116 p53þ/þ and p53�/� cells were cultured assubconfluent monolayer cultures (MLC) for 48 hours or as multilayered tumor spheroid cultures (SPC) for 14 days. Cells were incubated with BrdU for 24 hours andanalyzed by FACS. B, HCT116 p53þ/þ and p53�/� cells were cultured as SPC for 14 days, fixed, and embedded in paraffin. Histologic sections were analyzed forexpression of the proliferationmarker Ki67. Representative images of three spheroids analyzed per condition are shown.C,Analysis of proliferation in HCT116 p53þ/þ

and p53�/� xenograft tumor tissue from ref. 8 using Ki67. Representative images of tumors from 6 animals per group are shown. D, HCT116 p53þ/þ and p53�/� cellswere cultured as MLC for 48 hours or as SPC for 14 days. RNA was analyzed by RNA-seq and compared with RNA-seq data from HCT116 p53þ/þ and p53�/� cellsgrown as xenograft tumors (8). Principal component analysis showed overall higher similarity in gene expression signatures between tumors (T) and SPC comparedwith MLC. E, GSEA comparing MLC and SPC cultures of HCT116 p53þ/þ and p53�/� cells. Enrichment plots for HALLMARK_E2F_TARGETS, BROWNE_INTERFER-ON_RESPONSE_GENES, andMANALO_HYPOXIA_UP are shown.F,Enrichment plots for the samegene sets as inE comparingMLCand tumors (T) of HCT116 p53þ/þ

and p53�/� cells. G,Western blots showing levels of cyclin D1 (CCND1) and pyruvate dehydrogenase kinase 1 (PDK1) in HCT116 p53þ/þ or p53�/� cells grown as MLCor SPC. Vinculin is shown as loading control. H–J, HCT116 p53þ/þ and p53�/� cells were cultured as MLC or SPC and labeled for 16 hours with [U13C]-glucose. Cellswere extractedandmetaboliteswere analyzedby LC-MS.Data showmean�SEMof three independent biological replicates. Results from time-resolvedexperimentsare provided in Supplementary Fig. S1. H, Relative peak intensities for lactate. I, Ratios of Mþ3 and Mþ2 isotopologues for aspartate. J, Relative peak intensities ofaspartate.

Kaymak et al.





compared with MLC, was higher in p53-deficient SPC compared withtheir wt counterparts (Supplementary Fig. S1F), potentially reflectinghigher demand of nucleotides for proliferation.

Together, transcriptomic andmetabolic analyses demonstrated thatSPC induces hypoxic reprogramming of cellular metabolism in cancercells. However, oxidative reactions required to generate substrates foranabolic reactions (i.e., aspartate) are still supported throughanaplerosis.

Loss of p53 activates the mevalonate pathway via SREBP2We next compared gene expression signatures between p53 wt and

deficient cells under the different culture conditions. The majorsignatures associated with wt p53 status in all conditions were inflam-mation and IFNa response (Supplementary Fig. S2A). Signaturesassociated with p53 deficiency in MLC mapped to TGFb signaling,spermatogenesis, and cell cycle (Supplementary Fig. S2A, left). Incontrast, loss of p53 in SPC and xenograft tumors resulted in theinduction of genes associated with cholesterol homeostasis (Supple-mentary Fig. S2A, middle and right part), many of which are regulatedby the SREBP transcription factors (18).Moreover, SREBP target genes(Horton_SREBF_targets; ref. 19) showed strong enrichment in p53-deficient cells in SPC and xenograft tumors but not in MLC(Fig. 2A), suggesting that the combined effect of environment andloss of p53 leads to the upregulation of these genes. As cholesterolhomeostasis is preferentially regulated by SREBP2, rather than theclosely related SREBP1a or SREBP1c isoforms (19), we next inves-tigated expression of canonical SREBP2 target genes. This showedincreased expression of HMGCS1, MVD, HMGCR, DHCR7, andFDFT1 mRNA in HCT116 SPC compared with MLC, which wasfurther increased upon loss of p53, while expression of SREBF2mRNA was increased in SPC compared with MLC of p53-deficientcells (Fig. 2B). Moreover, HMGCS1, FDFT1, and ACSS2 showedincreased protein levels in p53-deficient SPC from a second isogeniccolon cancer cell line, RKO (Fig. 2C).

We also investigated whether expression of SREBP2 target genes isassociated with TP53mutation in human colorectal adenocarcinoma.Analysis of a The Cancer Genome Atlas (TCGA) dataset (20) revealedhigher expression of canonical SREBP2 targets in TP53-mutanttumors (Fig. 2D) and increased expression of HMGCS1 in high-grade colorectal adenocarcinoma (Fig. 2E). Moreover, two coloncancer cell lines expressing mutant TP53 (HT29 and DLD1) displayedstronger induction of HMGCS1 expression upon SPC compared withp53 wt cell lines (LS174T and LOVO; Fig. 2F), corroborating that lossof normal p53 function either through mutation or deletion increasesthe expression of mevalonate pathway genes.

Wild-type p53 was shown to inhibit mevalonate pathway genesthrough induction of the cholesterol transporter ABCA1 (21). Inagreement with this study, we found that ABCA1 mRNA expressionwas strongly reduced in p53-deficient cells both in MLC and SPC(Supplementary Fig. S2B). However, levels of ABCA1 protein werehigher in p53-deficient MLC and completely absent in SPC (Supple-mentary Fig. S2C). Interestingly, ABCA1 is a target for miRNA-33,which is encoded by an intron within the SREBF2 gene (22, 23). It istherefore possible that ABCA1 is repressed in SPC via a miRNA-dependent mechanism.

To address themechanism ofmevalonate pathway regulation in oursystem, we first confirmed that increased expression of HMGCS1protein in p53-deficient SPC is abolished upon shRNA-mediatedsilencing of SREBP2 (Supplementary Fig. S2D and S2E). We alsoestablished that p53-deficient SPC contain high levels of the 55 kDamature form of SREBP2 (Fig. 2G), which represents the active

transcription factor. Accumulation of mature SREBP2 and enhancedtarget expression was also observed in MLC of p53-deficient HCT116cells cultured in lipid-reduced medium (Supplementary Fig. S2F andS2G), a condition that induces SREBP processing (18). Nuclearaccumulation of mature SREBP2 is mediated by increased processingof the precursor or by stabilization of the mature protein. As SREBPprocessing is induced by mTORC1 (24, 25), we first investigated theactivity of this pathway. We found that phosphorylation of themTORC1 substrate p70S6K (indicated by the higher relative abun-dance of the upper band) and its downstream target S6 ribosomalprotein (S6RB), is strongly increased in SPC compared with MLC(Fig. 2H; Supplementary Fig. S2H). This was surprising as hypoxia, amajor feature of SPC, inhibits the mTORC1 pathway (26, 27). Indeed,exposure of HCT116 MLC to hypoxia decreased S6RB phosphoryla-tion and slightly reduced HMGCS1 expression (SupplementaryFig. S2I).

As increased mTORC1 activity was observed in SPC from bothgenotypes, we also addressedwhether loss of p53 alters protein stabilityof mature SREBP2. Treatment with the proteasome inhibitor MG132only increasedmature SREBP2 in p53 wt SPC, confirming that matureSREBP2 is more stable when p53 is absent (Fig. 2I). Mature SREBP2 isphosphorylated by glycogen synthase kinase 3 (GSK3), leading to itsubiquitination and degradation (28). We found an overall increase inGSK3 phosphorylation on serine 21 (GSK3a) and serine 9 (GSK3b) inSPC compared with MLC, with a further increase in p53-deficientcells (Fig. 2J), indicating reduced activity of the kinase upon p53 loss.Consistently, treatment of p53 wt SPC with GSK3 inhibitorsincreased levels of mature SREBP2 and restored expression ofHMGCS1 mRNA to the same level found in p53-deficient cells(Fig. 2K and L). Treatment with the mTORC1 inhibitor rapamycinreduced mature SREBP2 and HMGCS1 mRNA in p53-deficient SPC(Fig. 2K andM). However, this was independent of AKT, as treatmentwith MK2206 did not affect GSK3 or S6RB phosphorylation (Sup-plementary Fig. S2J).

Together, this suggests that loss of p53 in SPC induces nuclearaccumulation of mature SREBP2 through activation of mTORC1 andinhibition of GSK3.

Inhibition of mevalonate synthesis induces apoptosis inp53-deficient spheroids

We next tested whether the mevalonate pathway contributes tocancer cell survival in the metabolically compromised environment ofSPC. We used statins, a class of lipid-lowering drugs that inhibit theactivity ofHMGCR, the rate-limiting enzyme of the pathway (Fig. 3A).Statin treatment increased the expression of SREBP2 target genes, dueto inactivation of the negative feedback loop (29), and resulted inglobal downregulation of cell cycle and epithelial-to-mesenchymaltransition (EMT) expression signatures regardless of genotype (Sup-plementary Fig. S3A). Protein levels of the S-phase proteins cyclin A(CCNA1) and aurora kinase A (AURKA) were also reduced (Supple-mentary Fig. S3B), confirming that the mevalonate pathway contri-butes to proliferation (30) and disruption of tissue architecture (31).

When investigating the effect of mevalonate pathway inhibitors oncell viability, we found strong inhibition of proliferation in MLC,irrespective of genotype (Supplementary Fig. S3C). In contrast, in SPConly, p53-deficient cells were sensitive to mevastatin treatment, indi-cated by TUNEL staining, while p53 wt cells were largely resistantto this treatment (Fig. 3B and C). Mevastatin-induced apoptosis inp53-deficient cells was blocked by addition of mevalonate, the productof the HMGCR reaction (Fig. 3B and C), confirming the specificity ofthe inhibitor. Apoptotic cells positive for TUNEL and cleaved caspase-






0.0-0.1-0.2-0.3-0.4-0.5-0.6

Enr

ichm

ent s

core

(ES

)

-0.7

HORTON_SREBF_TARGETS

NES -2.64q ≤ 0.05

p53+/+ p53−/−

SpheroidsHORTON_SREBF_TARGETS

NES -1.8q = 0.057

0.0-0.1-0.2-0.3-0.4-0.5-0.6

Enr

ichm

ent s

core

(ES

)

p53+/+ p53−/−

TumorsA

B

mSREBP2

p53+/+ p53−/−

MLC SPCp53+/+ p53−/−

ACTIN

HMGCS1

p53+/+

p53−/−

GSK3α/β

p53+/+ p53−/−


ACTIN

P-GSK3α/β(Ser21/9)

M

J

G

HMGCS1

p53+/+ p53−/− p53+/+ p53−/−0

1

2

3

4*

CONCHIR99021

0

2

4

6

8

10HMGCS1

**

**

****

Rel

ativ

e m

RN

A e

xpre

ssio

n

MLC SPC0.0

0.5

1.0

1.5

2.0

2.5MVD

**

***

Rel

ativ

e m

RN

A e

xpre

ssio

n

0

2

4

6HMGCR

**

***

Rel

ativ

e m

RN

A e

xpre

ssio

n0

2

4

6DHCR7

*

*

**

Rel

ativ

e m

RN

A e

xpre

ssio

n

0

1

2

3

4

5FDFT1

*

*

****

Rel

ativ

e m

RN

A e

xpre

ssio

n

0

1

2

3

4SREBF2

n.s.

**

Rel

ativ

e m

RN

A e

xpre

ssio

n

L

Rel

ativ

e m

RN

A e

xpre

ssio

n

p53+/+ p53−/−


ACTIN

S6RB

0

1

2

3

4

5

Rel

ativ

e m

RN

A e

xpre

ssio

n

HMGCS1

HORTON_SREBF_TARGETS

p53+/+ p53−/−

Enr

ichm

ent s

core

(ES

)

0.0

-0.1

-0.2

0.1

0.2

Monolayer

mut wt

0

5

10

TP53 Status

z-S

core

/RE

SM

**D

pT1 pT2 pT3 pT4

02,

000

4,00

06,

000

8,00

010

,000

HMGCS1

Rel

ativ

e ex

pres

sion

(RP

KM

)

Tumor stage

E**

Kp53+/+ p53−/−

CON CHIRSBRAPA

mSREBP2

SPC

CON CHIRSBRAPA

VCL

P-S6RB(Ser240/244)

VCL

HMGCS1

p53

p53+

/+

p53−/−

RKO SPC

FDFT1

ACSS2

C

H

LS17

4T (W

T)

LOVO (W

T)

HT29 (R

273H

)

DLD1 (

S241F

)0

2

4

6

8

10

HMGCS1

Rel

ativ

e m

RN

A e

xpre

ssio

n MLCSPC

F

1.0 0.93 3.05 4.35

SPC

p53+

/+p5

3-/-

P-G

SK3/

GSK

3

1.0 1.01.06 2.25 2.88 0.59 0.87 0.86

p53+/+ p53−/−

− + MG132− +

VCL

mSREBP2

SPCI

1.0 2.32 1.0 0.84

CONRAPA

MLC SPC MLC SPC MLC SPC MLC SPC MLC SPC

NES 0.67q = 0.991

124 kDa

53 kDa

79 kDa

48 kDa

57 kDa

57 kDa

42 kDa

60–78kDa 42 kDa

32 kDa

32 kDa 60–78kDa

124 kDa

60–78kDa

124 kDa

42 kDa

0.0

0.5

1.0

1.5

2.0

2.5 *

51 kDa47 kDa51 kDa47 kDa

Figure 2.

Loss of p53 induces enzymes of the mevalonate pathway via activation of SREBP2. A, Enrichment plots for HORTON_SREBF_TARGETS (19) for HCT116 p53þ/þ andp53�/� cells cultured as MLC, SPC, or xenograft colon tumors. B, Expression of canonical mevalonate pathway genes and SREBF2 in HCT116 p53þ/þ or p53�/� cellsgrown as MLC or SPC. Data showmean� SEM of three independent biological replicates. � , P < 0.05; �� , P < 0.01; �� , P < 0.001; �� , P < 0.0001, unpaired two-tailedStudent t test. n.s., nonsignificant. C,Western blots showing levels of HMGCS1, FDFT1, ACSS2, and p53 in RKO p53þ/þ and RKO p53�/� cells grown as SPC. Vinculin isshown as loading control. D, Combined z-score for the expression of canonical mevalonate pathway genes (HMGCS1, HMGCR, MVD, DHCR7, ACSS2, FDFT1, andSREBF2) was calculated for tumors from the TCGA colorectal adenocarcinoma dataset. Mevalonate pathway signature values were compared between all p53 wt (n¼ 94) and p53-mutant (n ¼ 88) tumors. P ¼ 0.0051 was determined using an unpaired two-tailed Student t test. E, Expression of HMGCS1 in colorectaladenocarcinoma tumors from the TCGA dataset according to stage (PT1–4). F, Expression ofHMGCS1 in TP53wt andmutant colon cancer cell lines grown as MLC orSPC. G,Western blot analysis showing expression of HMGCS1 and mature SREBP2 in HCT116 p53þ/þ or p53�/� cells grown as MLC or SPC. Actin is shown as loadingcontrol. H,Western blots showing levels of phosphorylated ribosomal protein S6 (P-S6RB) and total ribosomal protein S6 expression (S6RB) in HCT116 p53þ/þ orp53�/� cells grown as MLC or SPC. (Continued on the following page.)

Kaymak et al.





3 were mainly found in the core regions, where cells are experiencingthe most severe oxygen and nutrient depletion (Fig. 3B; Supplemen-tary Fig. S3D), suggesting that the mevalonate pathway supports cellviability under metabolic stress.

As it has been shown that induction of p21 (CDKN1A) is requiredfor the p53-dependent remodeling of metabolism in response to serinedeprivation in colon cancer (10), we asked whether failure to inducep21 could be responsible for the induction of apoptosis by statins in

(Continued.) Actin is shown as loading control. I, SPC of HCT116 p53þ/þ or p53�/� cells were treated with 20 mmol/L MG132 or solvent for 1 hour. Levels of matureSREBP2 were detected byWestern blot analysis. Vinculin is shown as loading control. J,Western blots showing phosphorylation on GSK3a/b (serine 21/9) and totalGSK3a/b protein in HCT116 p53þ/þ or p53�/� cells grown as MLC or SPC. Numbers display P-GSK3/GSK3 signal ratio. Actin is shown as loading control. Graph showsmean� SEM ratios of biologically independent replicate SPC samples (n¼ 4). � , P� 0.05 was determined using a paired two-tailed Student t test. K, HCT116 p53þ/þ

or p53�/� cells grown as SPC were treated with 20 nmol/L of rapamycin (RAPA), 30 mmol/L of SB216763 (SB), or 10 mmol/L of CHIR99021 (CHIR) for 24 hoursandmature SREBP2was analyzed byWestern blot analysis. Vinculin is shown as loading control. Numbers show signal intensity formSREBP2 normalized to vinculin.L, Expression of HMGCS1mRNA in HCT116 p53þ/þ or p53�/� cells grown as spheroids and treated with 10 mmol/L of CHIR99021 (CHIR) for 72 hours. Data showmean � SEM of three independent biological replicates. � , P < 0.05, unpaired two-tailed Student t test. M, Effect of rapamycin on HMGCS1 expression inspheroid cultures of HCT116 p53þ/þ or p53�/� cells. Data are presented as mean of duplicate samples.

DMSOSIM

24 h

0

1

2

3

Rel

ativ

e m

RN

A e

xpre

ssio

n *

p53+/+ p53+/+p53−/−0

2

4

6

Rel

ativ

e m

RN

A e

xpre

ssio

n

***

*

p53−/−

DMSOSIM

72 h

D

DMSO SIM SIM + MVL

p21−/

−

110 μm 110 μm 110 μm

F

CDKN1A CDKN1A

Acetyl-CoA

HMG-CoA

Mevalonate

Farnesyl-PP

Squalene

Cholesterol

HMGCRStatins

A

p53+/+ p53−/−

DMSO SIM +MVL

SIM DMSO

p21

SIM +MVL

SIM

p21+/+ p21−/−

DMSO SIM +MVL

SIM DMSO

p21

SIM +MVL

SIM

E

G

150 μm

150 μm

150 μm

150 μm

150 μm

150 μm

DMSO MST MST + MVL

Purple=TUNELBlue=DAPI

C

p53+/

+p5

3−/−

B

DMSOMST

MST + MVL

DMSOMST

MST + MVL

0

100

200

300

400

TUN

EL-p

ositiv

e ce

lls/a

rea p53+/+

p53−/−**

VCL

VCL124 kDa

21 kDa

124 kDa

21 kDa

SPC

SPC

Figure 3.

The mevalonate pathway is essential for the survival of p53-deficient colon cancer cells. A, Diagram showing selected metabolites of the mevalonate pathway andHMGCR, themolecular target of statins. B,HCT116 p53þ/þ or p53�/� cells were grown as SPC and treatedwith 10 mmol/Lmevastatin (MST) or solvent (DMSO) eitheralone or in combination with 0.5 mmol/L mevalonate (MVL) for 72 hours. Spheroids were fixed and histologic sections were analyzed for the presence of apoptoticcells by TUNEL staining. Images show representative results of three spheroids analyzed per condition. C, Quantitation of data shown in B. Data are presented asmean� SEMof at least three spheroids analyzed per condition. �� , P <0.01, unpaired two-tailed Student t test.D,HCT116 p53þ/þ and p53�/� cells grown as SPCweretreatedwith 10 mmol/L simvastatin (SIM) or solvent (DMSO) for 24 or 72 hours. Expression of p21 (CDKN1A)mRNAwas determined by qPCR. Data showmean� SEMof three independent biological replicates. � , P < 0.05; �� , P < 0.001, unpaired two-tailed Student t test. E, HCT116 p53þ/þ and p53�/� cells were grown as SPC andtreated with 10 mmol/L simvastatin (SIM) or solvent (DMSO) either alone or in combination with 0.5 mmol/L mevalonate for 72 hours. Expression of p21 protein wasdetermined by Western blotting. Vinculin is shown as loading control. F, HCT116 p21�/� cells were grown as SPC and treated with 10 mmol/L simvastatin (SIM) orsolvent (DMSO) either alone or in combination with 0.5 mmol/L mevalonate (MVL) for 72 hours. Spheroids were fixed and histologic sections were analyzed for thepresence of apoptotic cells by TUNEL staining. Images show representative results of three spheroids analyzed per condition. G, HCT116 p21þ/þ or p21�/� cells weregrown as SPC and treated with 10 mmol/L simvastatin (SIM) or solvent (DMSO) either alone or in combination with 0.5 mmol/L mevalonate (MVL) for 72 hours.Expression of p21 protein was determined by Western blotting. Vinculin is shown as loading control.






p53-deficient cells. However, while short-term simvastatin treatment(24 hours) induced CDKN1A mRNA only in p53 wt cells, bothgenotypes increased CDKN1A mRNA and protein expression afterlonger statin exposure (72 hours; Fig. 3D and E). This indicatesthat mevalonate pathway inhibition induces p21 through a p53-independent mechanism. Furthermore, statin treatment of SPC ofp21-deficient HCT116 cells did not induce apoptosis (Fig. 3F and G),confirming that p21 is dispensable for statin resistance of p53 wt cells.

Mevalonate pathway inhibition blocks the production ofubiquinone

The mevalonate pathway facilitates the synthesis of isoprenoids,which are substrates for cholesterol synthesis, protein prenylation aswell as the synthesis of dolichol, heme A, and ubiquinone (Fig. 4A;ref. 32). Ubiquinone consists of a benzoquinone ring derived fromtyrosine linked to a tail comprising 10 (human) or 9 (mouse) iso-prenoid units and functions as electron transfer molecule between thecomplexes of the respiratory chain (33). To determine mevalonatepathway activity, we treated SPC of p53 wt and deficient cells with[U13C]-glucose and determined label incorporation into differentmetabolites. SPC of p53-deficient cells increase the incorporation oflabeled carbons into mevalonate, resulting in an overall increasedabundance of this metabolite (Fig. 4B and C). In addition, p53-deficient SPC displayed overall higher levels of acetyl-CoA, the sub-strate of the mevalonate pathway, without major difference in theproportional labeling of the Mþ2 fraction, which is generated fromcitrate (Supplementary Fig. S4A and S4B). However, despite theobserved increase in mevalonate synthesis, the amount of total andlabeled cholesterol was much lower in p53-deficient SPC comparedwith their wt counterparts (Fig. 4D and E). In contrast, the amount oftotal and labeled ubiquinone was higher in p53-deficient cells,demonstrating a rerouting of metabolites into the ubiquinonesynthesis pathway (Fig. 4F and G). Treatment with simvastatindecreased the amount of both metabolites and completely abolishedtheir glucose-derived labeling (Fig. 4D–F; Supplementary Fig. S4Cand S4D). Importantly, isotopolog peak distribution for cholesteroland ubiquinone were similar in both genotypes (Fig. 4E and G),indicating comparable contribution of glucose to the acetyl-CoApool (34).

To investigate whether a p53-dependent switch in the routing ofmetabolites in themevalonate pathway can also be observed in tumors,we determined the abundance of cholesterol, 7-dihydroxy cholesterol(7-DHC) and ubiquinone in xenograft colon tumors of p53 wt anddeficient HCT116 cells. In support of the results obtained in spheroidcultures, p53 wt colon tumors displayed somewhat higher levels ofcholesterol and 7-DHC compared with p53-deficient colon tumors. Incontrast, levels of ubiquinone were overall higher in p53-deficientcolon tumors, although this difference failed to reach significance(Fig. 4H). Together, these results suggest that loss of p53 altersmevalonate pathway flux to support the production of ubiquinone.

Inhibition of ubiquinone synthesis impairs TCA cycle andrespiration and results in oxidative stress

Ubiquinone is an essential component of the mitochondrial elec-tron transport chain (ETC) where it shuttles electrons betweenNADH-CoQ reductase (complex I) or succinate dehydrogenase (com-plex II) and CoQH2-cytochrome c reductase (complex III; Fig. 5A).Using stable isotope tracing with [U13C]-glucose, we found thatsimvastatin reduced labeled and unlabeled fractions of aspartate andmost TCA cycle metabolites in SPC from both genotypes (Fig. 5B;Supplementary Fig. S5A). However, simvastatin reduced basal and

maximal OCRs in SPC of p53-deficient cells, which was rescued bymevalonate addition. In contrast, SPC from p53wt cells only displayeda small reduction in maximal respiration upon statin treatment(Fig. 5C).

Reduced availability of oxygen as final electron acceptor of the ETCcan lead to electron leakage and the formation of reactive oxygenspecies (35). We therefore reasoned that inhibition of ubiquinonesynthesis could cause oxidative stress under the hypoxic conditions inSPC, which could lead to the induction of apoptosis. Indeed, replen-ishing spheroid cultures either with ubiquinone or the antioxidantN-acetyl-cysteine (NAC) was as effective as mevalonate in preventingstatin-induced apoptosis in SPC (Fig. 5D and E), while cell-permeablecholesterol had no effect (Supplementary Fig. S5B). In addition, theviability of statin-treated MLC was not restored by the addition ofubiquinone (Supplementary Fig. S5C), indicating that multiple pro-ducts of this pathway are needed to support the rapid proliferation ofcancer cells observed in MLC.

Production of ubiquinone by the mevalonate pathway supportspyrimidine nucleotide biosynthesis

Ubiquinone also functions as an electron acceptor for dihydroor-otate dehydrogenase (DHODH), an essential enzyme for the gener-ation of pyrimidine nucleotides forDNAandRNA synthesis (Fig. 6A).Stable isotope tracing showed higher incorporation of glucose-derivedcarbons into UMP in SPC of p53-deficient cells (Fig. 6B). This wasdetected in theMþ5 fraction, representing labeling via ribose, but alsoin the Mþ7/Mþ8 fractions, representing ribose plus either two orthree labeled carbons derived from aspartate (Fig. 6B). Treatmentwithstatins significantly lowered labeling and overall levels of UMP in SPCfrom both genotypes (Fig. 6B and C). This was restored by supple-menting statin–treated SPC with either mevalonate or ubiquinone(Fig. 6C), confirming that ubiquinone is rate-limiting for pyrimidinesynthesis under these conditions.Moreover, addition of nucleosides oruridine, which can readily be taken upby cells and used to replenish thenucleotide pool via the salvage pathway, was sufficient to block theinduction of apoptosis by statins in p53-deficient SPC (Fig. 6D;Supplementary Fig. S6A).

The antimetabolite drug 5-fluorouracil (5-FU), which is standard-of-care for advanced colorectal adenocarcinoma, exerts its effectmostly through inhibition of thymidylate synthase (TYMS; ref. 36).TYMS converts dUMP to dTMP for DNA synthesis, and 5-FUtreatment leads to DNA damage and cell death. We therefore inves-tigated whether statins alter 5-FU sensitivity of cancer cells under themetabolic constraints of SPC. Interestingly, while p53wtHCT116 cellsshowed remarkable resistance toward 5-FU, most likely due to the lowproliferation of these cells in this condition, the drug sensitized the cellsto simvastatin treatment (Fig. 6E). In contrast, p53-deficient cellsalready showed induction of apoptosis in response to statin alone,which was not further increased by 5-FU (Fig. 6E).

Collectively, these results demonstrate that ubiquinone productionby the mevalonate pathway is essential for pyrimidine biosynthesis incancer cells. Inhibition of ubiquinone synthesis blocks the viability ofp53-deficient cells under themetabolic constraints of SPC. In contrast,p53 wt cells are initially resistant to statin treatment but can besensitized by the antimetabolite 5-FU, which blocks dTMP synthesisand causes DNA and RNA damage.

The metabolic output of the mevalonate pathway depends onenvironmental context

To investigate the role of the mevalonate pathway under differentconditions resembling the tumor microenvironment, we used

Kaymak et al.





Acetyl-CoA

Mevalonate

Farnesyl-PP

Squalene

Cholesterol

Geranyl geranyl-PPDolicholHeme AUbiquinone (CoQ10)

A

0

50

100

150

Cholesterol

0

20

40

60

Ubiquinone (CoQ10)

UnlabeledLabeled

M+0M+1M+2M+3M+4M+5M+6M+7M+8M+9M+1

0M+1

1M+1

2M+1

3M+1

4M+1

5M+1

6M+1

7M+1

8M+1

9M+2

0M+2

1M+2

2M+2

3M+2

4M+2

5M+2

6M+2

7M+2

8M+2

9M+3

0M+3

1M+3

2M+3

3M+3

4M+3

5M+3

6M+3

7M+3

8M+3

9M+4

0M+4

1M+4

2M+4

3M+4

4M+4

5M+4

6M+4

7M+4

8M+4

9M+5

00

1

2

31020304050

Ubiquinone (CoQ10)

p53+/+ DMSOp53−/− DMSO

p53+

/+ DMSO

p53+

/+ SIM

p53−

/− DMSO

p53−

/− SIM

p53+

/+ DMSO

p53+

/+ SIM

p53−

/− DMSO

p53−

/− SIM

Pea

k in

tens

ity/p

rot

Pea

k in

tens

ity/p

rot

Pea

k in

tens

ity/p

rot

Glucose

UnlabeledLabeled

GF

ED

B

M+0 M+2 M+4 M+60.0

0.2

0.4

0.6p53+/+

p53−/−

0

100

200

300

400

7-DHC

p53+/+ p53−/− p53+/+ p53−/− p53+/+ p53−/−0

20

40

60

80

100

120

Pea

k in

tens

ity (x

10)/t

issu

e w

eigh

t7

Cholesterol

0.0

0.2

0.4

0.6

0.8

Pea

k in

tens

ity/ti

ssue

wei

ght

Ubiquinone (CoQ10)

*H

MevalonateC

*

012345

50

100

M+0M+1M+2M+3M+4M+5M+6M+7M+8M+9M+1

0M+1

1M+1

2M+1

3M+1

4M+1

5M+1

6M+1

7M+1

8M+1

9M+2

0M+2

1M+2

2M+2

3M+2

4M+2

5M+2

6M+2

7

p53+/+ DMSOp53−/− DMSO

Peak

inte

nsity

/pro

t

Cholesterol

Pea

k in

tens

ity/p

rot

p53+

/+

p53−

/−

0.0

0.5

1.0

1.5

Pea

k in

tens

ity/p

rot

UnlabeledLabeled

Mevalonate

Labeled

Pea

kint

ensi

ty (x

107 )/

tissu

e w

eigh

t

7-DHC

Figure 4.

Inhibition of mevalonate synthesis blocks the production of ubiquinone in colon cancer cells. A, Schematic showing the branching of the mevalonate pathway intocholesterol biosynthesis, the generation of isoprenoids for protein prenylation, and the synthesis of dolichol, hemeA, and ubiquinone (CoQ10).B andC,HCT116 p53þ/

þor p53�/� cellswere grownas SPCand labeledwith [U13C]-glucose for 16 hours before extraction andanalysis ofmevalonate isotopologues.Data showmean�SEMof three independent biological replicates. B, Relative peak intensities of labeled and unlabeled fractions for mevalonate. C, Relative peak intensities of individuallabeled fractions for mevalonate. D–G, HCT116 p53þ/þ or p53�/� cells were grown as SPC and treated with 10 mmol/L simvastatin (SIM) or solvent (DMSO) for72 hours. For the last 16 hours, cellswere labeledwith [U13C]-glucose before cellswere extracted andmetaboliteswere analyzedby LC-MS. Data showmean� SEMofthree independent biological replicates. D, Relative peak intensities of labeled and unlabeled fractions for cholesterol. E, Relative peak intensities of individualisotopologues for cholesterol. F, Relative peak intensities of labeled and unlabeled fractions for ubiquinone (CoQ10). G, Relative peak intensities of individualisotopologues for ubiquinone. H, Xenograft tumors from HCT116 p53þ/þ or p53�/� cells were extracted and levels of cholesterol, 7-dihydroxycholesterol (7-DHC),and ubiquinone (CoQ10) were determined by LC-MS. Data are shown as mean � SEM of six p53þ/þ and five p53�/� colon tumors. � , P < 0.05, unpaired two-tailedStudent t test.






organoid cultures of intestinal epithelial cells from mice carryingconditional alleles of Apc, Trp53, or KrasG12D together with Villin-CreERT2. Efficient recombination of the Trp53 and Kras locus wereconfirmed by PCR (Supplementary Fig. S7A). Placed in organoidculture medium, these cells grow as large cysts without any signs ofdifferentiation (13). Interestingly, while simvastatin only had a minoreffect on the growth of Apc-deficient organoids, Apc/p53 double-deficient cells showed a severe reduction in organoid growth, whichwas fully restored by mevalonate supplementation (Fig. 7A and B).Reduced organoid growth was accompanied by induction of PARPcleavage, a marker of apoptotic cell death (Supplementary Fig. S7B).Similar results were also obtained for Apcfl/fl/KrasG12D and Apcfl/fl/p53fl/fl/KrasG12 organoids (Fig. 7A and B), demonstrating that thedeletion of Trp53 sensitizes the organoids towardmevalonate pathwayinhibition. Moreover, inhibition of organoid growth in Apc/p53–deficient and Apcfl/fl/ p53fl/fl/KrasG12D/þ cells was robustly restored byaddition of either ubiquinone or nucleosides (Fig. 7C and D), con-firming that the provision of ubiquinone for nucleotide biosynthesis isan essential function of the mevalonate pathway in colorectal adeno-carcinoma tumor organoids.

We next assessed the ability of simvastatin to suppress intestinalhyperproliferation induced by acute deletion of Apc and activation ofKras in vivo. This was achieved by crossing mice carrying conditionalalleles of Apc or an activated allele of Kras (Apcfl/fl or Apcfl/fl;KrasG12D/þ) to mice bearing the VillinCreERT2 transgene (37). Afterinduction of CRE-dependent recombination, mice were treated for

4 days with simvastatin or vehicle and with D2O to assess cholesteroland ubiquinone synthesis in vivo (38). Histologic analysis ofBrdU-positive cells demonstrated that simvastatin had no effect onproliferation in Apc-deficient intestinal crypts, but blocked hyperpro-liferation induced by Kras activation (Fig. 7E and F). Deuteriumtracing revealed that the highly proliferating intestinal crypts inVillinCREERT2Apcfl/fl;KrasG12D/þmice exhibit high levels of cholesterolsynthesis, which was blocked by simvastatin (Fig. 7G). In contrast,ubiquinone synthesis was already lower in intestines from doublemutant mice and not further reduced by simvastatin (Fig. 7G),indicating that cholesterol rather than ubiquinone is the limitingmetabolite produced by the mevalonate pathway in this system. Thiswas further corroborated by the finding that hyperproliferation ofintestinal crypts in VillinCREERT2Apcfl/fl;KrasG12D/þ mice was insen-sitive to the DHODH inhibitor leflunomide (Supplementary Fig. S7Cand S7D), which has recently been shown to block growth of breastcancer cells (39). Together, these results indicate that de novo pyrim-idine synthesis is dispensable for KRAS-induced intestinal hyperpro-liferation and that the metabolic output of the mevalonate pathwaydepends on genetic factors and microenvironmental context.

DiscussionMetabolic gradients in tumors are likely to simultaneously limit

access to oxygen and nutrients, making adaptation by metaboliccompensation challenging (40). One potential response of cancer cells

A B

p53+

/+ DMSO

p53+

/+ SIM

p53−

/− DMSO

p53−

/− SIM

0

100

200

300

400

Aspartate

M+0M+1M+2M+3M+4

C

0 50 100 150 2000

100

200

300

400

500

Time (minutes)

OC

R (p

mol

/min

/ sph

eroi

d ar

ea)

HCT116 p53+/+

0 50 100 150 2000

100

200

300

400

500

Time (minutes)

OC

R (p

mol

/min

/sph

eoid

are

a)

HCT116 p53−/−

DMSOSIMSIM+MVL

DMSO SIM

p53+/

+p5

3−/−

SIM + CoQ10

150μm150μm150μm

150μm 150μm 150μm

D

DMSOSIM

SIM +

MVL

SIM +

NACDMSO

SIM

SIM +

MVL

SIM +

NAC0

100

200

300

400

500

TUN

EL-

posi

tive

cells

/are

a **** ******p53+/+

p53−/−

Purple=TUNELBlue=DAPI

SIM + NAC

150μm

150μm

***

E

Oligo FCCP R/A Oligo FCCP R/A

SIM + MVL

150μm

150μm

Pea

k in

tens

ity/p

rot

Q10

ATPSynthase

III IVIII

cyt C

e- e-

e-

e-

H2O1/2 O2+H+

e-

Matrix

Succinate FumarateNADHNAD+

CitrateAconitate

α-KG Malate

Oxaloacetate

Aspartate

NADH

Intermembrane Space

SIM +

CoQ10

SIM +

CoQ10

Figure 5.

Inhibition ofmevalonate synthesis blocks TCA cycle activity and cellular respiration and induces oxidative stress.A,Diagram showing the role of ubiquinone (CoQ10)in electron transport within the mitochondrial electron transport chain (ETC). B, HCT116 p53þ/þ or p53�/� cells were grown as SPC and treated with 10 mmol/Lsimvastatin (SIM) or solvent (DMSO) for 72 hours. For the last 16 hours, cells were labeledwith [U13C]-glucose before cellswere extracted andmetabolite levels wereanalyzed by LC-MS. Relative peak intensities of isotopologues for aspartate are shown. Data showmean� SEM of three independent biological replicates. C,HCT116p53þ/þ or p53�/� cellswere grown as SPC and treatedwith 10 mmol/L simvastatin (SIM) or solvent (DMSO) either alone or in the presence of 0.5mmol/Lmevalonatefor 72 hours. Oxygen consumption rates (OCR) were determined using the Seahorse Bioanalyzer. Oligomycin (oligo), FCCP, and rotenone/antimycin A (R/A) wereadded to determine ATP-dependent, maximal, and basal respiration. Data are presented as mean � SEM of at least 14 spheroids analyzed per condition. D, HCT116p53þ/þ or p53�/� cells were grown as SPC and treated with 10 mmol/L simvastatin (SIM) or solvent (DMSO) either alone or in combination with 0.5 mmol/Lmevalonate (MVL), 10mmol/L ubiquinone (CoQ10), or 5mmol/L N-acetylcysteine (NAC) for 72 hours. Spheroidswere fixed and histologic sectionswere analyzed forthe presence of apoptotic cells by TUNEL staining. Images show representative results of three spheroids analyzed per condition. E,Quantitation of data shown inD.Data are presented as mean � SEM of at least 3 spheroids analyzed per condition. �� , P < 0.001; �� , P < 0.0001, unpaired two-tailed Student t test.

Kaymak et al.





DMSO SIM

SIM +

NCL

DMSO SIM

SIM +

NCL0

100

200

300

400

TUN

EL-

posi

tive

cells

/are

a

DMSO SIM5-F

U

SIM +

5-FU

DMSO SIM5-F

U

SIM +

5-FU

0

100

200

300

400

TUN

EL-

posi

tive

cells

/are

a

**

*

* p53+/+

p53−/−p53+/+

p53−/−

D E**** *

0

10

20

30

UMP

M+0M+1M+2M+3M+4M+5M+6M+7M+8M+9

0

2

4

6

8

10

UMP

**

*

p53+

/+ DMSO

p53−

/− DMSO

p53+

/+ SIM

p53−

/− SIM

p53+

/+ DMSO

p53−

/− DMSO

p53+

/+ SIM

p53−

/− SIM

Pea

k in

tens

ity/p

rot

Pea

k in

tens

ity/p

rot

C

Pea

k in

tens

ity/p

rot Labeled

Labeled

0

1

2

3

4

5

*

n.s.****

*

****

p53+

/+ DMSO

p53+

/+ SIM

p53+

/+ SIM +

MVL

p53−

/− DMSO

p53−

/− SIM

p53−

/− SIM +

MVL

p53+

/+ DMSO

p53+

/+ SIM

p53+

/+ SIM +

MVL

p53−

/− DMSO

p53−

/− SIM

p53−

/− SIM +

MVL0

5

10

UMP UMP

Pea

k in

tens

ity/p

rot

B

***

M+0M+1M+2M+3M+4M+5M+6M+7M+8M+9

Q10

ATPSynthase

DHODH

Dihydroorotatate Orotate

UMP

III IVIII

cyt C

e- e-

e-

e-

H2O1/2 O2+H+

e-

e-

Intermembrane Space

Matrix

Carboamoylaspartate

A

p53+

/+ SIM +

CoQ10

p53−

/− SIM

+ CoQ

10

p53+

/+ SIM +

CoQ10

p53−

/− SIM

+ CoQ

10

Figure 6.

Mevalonate pathway activity is essential for pyrimidine nucleotide biosynthesis and survival in colon cancer cells.A,Diagram showing the role of ubiquinone (CoQ10)in the conversion of dihydroorotate to orotate during pyrimidine biosynthesis. B, HCT116 p53þ/þ or p53�/� cells were grown as SPC and treated with 10 mmol/Lsimvastatin (SIM) or solvent (DMSO) for 72 hours. For the last 16 hours, cells were labeled with [U13C]-glucose before metabolites were extracted and analyzedby LC-MS. Relative peak intensities (left) and total labeled fractions (right) for UMP are shown. Data show mean � SEM of three independent biologicalreplicates. �, P < 0.05; �� , P < 0.01, unpaired two-tailed Student t test. C, HCT116 p53þ/þ or p53�/� cells were grown as SPC and treated with 10 mmol/Lsimvastatin (SIM) or solvent (DMSO) for 72 hours either alone or in combination with 0.5 mmol/L mevalonate (MVL) or 10 mmol/L ubiquinone (CoQ10).For the last 16 hours, cells were labeled with [U13C]-glucose before metabolites were extracted and analyzed by LC-MS. Relative peak intensities (left)and total labeled fractions (right) for UMP are shown. Data show mean � SEM of three independent biological replicates. � , P < 0.05; �� , P < 0.01; �� , P < 0.001;�� , P < 0.0001, unpaired two-tailed Student t test; n.s., nonsignificant. D, HCT116 p53þ/þ or p53�/� cells were grown as SPC and treated with solvent (DMSO)or 10 mmol/L simvastatin (SIM) either alone or in combination with nucleosides (150 mmol/L each of cytidine, guanosine, adenosine, uridine, and 50 mmol/L ofthymidine ¼ NCL) for 72 hours. Spheroids were fixed and histologic sections were analyzed for the presence of apoptotic cells by TUNEL staining. Data arepresented as mean� SEM of at least three spheroids analyzed per condition. � , P < 0.05; �� , P < 0.0001, unpaired two-tailed Student t test. E, HCT116 p53þ/þ

or p53�/� cells were grown as SPCs and treated with solvent (DMSO), 10 mmol/L simvastatin (SIM), 10 mmol/L 5-FU, or a combination of the two for 72 hours.Spheroids were analyzed by TUNEL staining. Data are presented as mean � SEM of at least three spheroids analyzed per condition. � , P < 0.05; �� , P < 0.001,unpaired two-tailed Student t test.






A

B

DMSOSIM

SIM+N

UCL0

5,00

1,000

1,500

Apcfl/fl p53fl/fl

****

****

****

DMSOSIM

SIM+N

UCL0

5,00

1,000

1,500

Apcfl/fl p53fl/fl KrasG12D/+

****

********

DMSOSIM

SIM+M

EV0.0

0.5

1.0

1.5

Apcfl/fl

Org

anoi

d ar

ea p

er fi

eld

DMSOSIM

SIM+M

EV0.0

0.5

1.0

1.5

Apcfl/fl p53fl/fl

**

Org

anoi

d ar

ea p

er fi

eld

Org

anoi

d ar

ea p

er fi

eld

Org

anoi

d ar

ea p

er fi

eld

Apc

fl/fl p

53fl/

flA

pcfl/

fl p53

fl/fl

Kra

sG12

D/+

Control SimvastatinSimvastatin

+ CoQ10Simvastatin

+ nucleosidesC

Apc

fl/fl p

53fl/

flA

pc fl/

flControl Simvastatin

Simvastatin+ mevalonate

D

200 μm 200 μm

200 μm 200 μm

200 μm

200 μm 200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

200 μm

Apc

fl/fl p

53fl/

fl

Kra

sG12

D/+

Villi

nCR

EE

RT2

DMSOSIM

SIM+M

EV0.0

0.5

1.0

1.5Apcfl/fl p53fl/fl

KrasG12D/+

****

**

Org

anoi

d ar

ea p

er fi

eld

Villi

nCR

EE

RT2

0.0

0.5

1.0

1.5

DMSOSIM

SIM+M

EV

Org

anoi

d ar

ea p

er fi

eld

Apcfl/fl, KrasG12D/+

200 μm 200 μm 200 μmApc

fl/fl K

rasG

12D

/+

Villin CREER Apcfl/fl

Vehicle Simvastatin Vehicle SimvastatinVillin CREER Apcfl/fl KrasG12D/+

0

20

40

60

80

No.

of B

rdU

pos

cells

per h

alf c

rypt *

E F*

0.45

0.50

0.55

0.60

0.65

Frac

tion

enric

hmen

t

0.7

0.8

0.9Fr

actio

n en

richm

ent

Apcfl/f

l

Apcfl/f

l KrasG12

D/+

Apcfl/f

l + SIM

Apcfl/f

l KrasG12

D/+ + SIM

Apcfl/f

l

Apcfl/f

l KrasG12

D/+

Apcfl/f

l + SIM

Apcfl/f

l KrasG12

D/+ + SIM

Cholesterol Ubiquinone (CoQ9)G

* * **

Apcfl/f

l

Apcfl/f

l KrasG12

D/+

Apcfl/f

l + SIM

Apcfl/f

l KrasG12

D/+ + SIM

SIM+C

oQ10

SIM+C

oQ10

Figure 7.

Simvastatin reduces growth of p53-deficient tumor organoids and blocks proliferation in Apc/p53-deficient Kras-transformed intestinal crypts. A, Primary mouseintestinal cells derived from VillinCREERT2Apcfl/fl, VillinCREERT2Apcfl/fl;p53fl/fl, VillinCREERT2Apcfl/fl;KrasG12D/þ, or VillinCREERT2Apcfl/fl;p53fl/fl;KrasG12D/þ animals wereused to generate organoid cultures. Organoidswere treatedwith 10 mmol/L simvastatin (SIM) either alone or in combinationswith 0.5mmol/Lmevalonate (MEV) for48hours. Images show representativemicroscopic fields from three independent replicate cultures.B,Quantitation of data shown inA. Data are presented asmean�SEM of microscopic fields from three independent cultures. � , P < 0.05; �� , P < 0.01; ��, P < 0.0001, unpaired two-tailed Student t test. C, Apcfl/fl;p53fl/fl or Apcfl/fl;p53fl/fl;KrasG12D/þ organoids were treated with 10 mmol/L simvastatin either alone or in combination with 10 mmol/L ubiquinone (CoQ10) or nucleosides (150 mmol/Leach of cytidine, guanosine, adenosine, uridine, and 50 mmol/L of thymidine) for 48 hours. Images show representative microscopic fields from three independentreplicate cultures. D, Quantitation of data shown in C. Data are presented as mean � SEM of microscopic fields from three independent cultures. �� , P < 0.0001,unpaired two-tailed Student t test.E,VillinCREERT2;Apcfl/fl andVillinCREERT2;apcfl/fl;KrasG12D/þmicewere treatedwith a single intraperitoneal injection of 80mg/kgoftamoxifen on one occasion (VillinCreERT2Apcfl/flKrasG12D/þ) or on two consecutive days (VillinCreERT2Apcfl/fl). From one day post-induction, micewere treated with adaily dose of 50 mg/kg simvastatin (in 0.5% methylcellulose/ 5% DMSO). After four days, mice were sacrificed and intestinal mucosa was fixed, paraffin embedded,and histologic sections were stained for BrdU incorporation. Representative images are shown. F, Three intestinal crypts for each genotype and treatmentwere scored for BrdU-positive cells.G, Fraction of cholesterol and ubiquinone (CoQ9) containing deuteratedwater in intestinalmucosa from thedifferent genotypes.� , P < 0.05 unpaired two-tailed Student t test.

Kaymak et al.





to nutrient deprivation is cell-cycle arrest, which alleviates the met-abolic demand of nucleotide biosynthesis for DNA replication, allow-ing cancer cells to survive until nutrients become available, forexample, after formation of new blood vessels or engagement ofmetabolic symbiosis (41, 42). Using spheroid cultures (SPC) as model,we show here that p53 wt colon cancer cells respond to metabolicdeprivation by reducing proliferation. In contrast, p53-deficient colo-rectal adenocarcinoma cells are able to maintain proliferation in thespheroid center, where nutrient and oxygen supply is restricted.Contrary to monolayer cultures, gene expression signatures in SPCare characteristic of cell-cycle arrest and induction of hypoxia, similarto those found in tumor tissue. Moreover, stable isotope tracingshowed that SPC engage in hypoxic remodeling of their metabolism,with reduced glucose oxidation, enhanced lactate production, andincreased TCA cycle anaplerosis from pyruvate. Pyruvate anaplerosispromotes glutamine-independent growth of cancer cells (43) andsupports aspartate synthesis in succinate dehydrogenase (SDH)-deficient cancer cells (44). We found that SPC of p53-deficient coloncancer cells showed reduced aspartate levels, indicating its enhancedusage for pyrimidine biosynthesis.

Importantly, p53-deficient colorectal adenocarcinoma cells in SPCor grown as xenograft tumors increase expression of mevalonatepathway enzymes and upregulation of SREBP2 targets was observedin p53-mutant colorectal adenocarcinoma patient samples and celllines. Previous studies have shown that mutant p53 can bind toSREBP2 and promoting its transcriptional activity during disruptionof mammary tissue architecture (31), and that wt p53 represses themevalonate pathway through ABCA1-dependent inhibition ofSREBP2 processing (21). We demonstrate here that loss of p53 inSPC promotes expression of SREBP2 target genes by activatingmTORC1 signaling, which drives the processing of SREBP2 (24), andby limiting the activity of GSK3, which controls the phosphorylation-dependent degradation of mature SREBP2 (28). The combination ofmTORC1 activation and inhibition of GSK3 results in the accumu-lation of mature SREBP2 and increases the expression of its targetgenes.

Our study also shows that tumor-like metabolic stress alters thesensitivity of cancer cells toward mevalonate pathway inhibition. Inmonolayer, both genotypes were highly sensitive to mevalonate path-way inhibitors, most likely because cells require cholesterol for rapidproliferation (32). However, when exposed to metabolic stress, p53-proficient cells were largely resistant to statin treatment, while p53-deficient cancer cells showed induction of apoptosis. Cell death wasrestricted to the nutrient- and oxygen-deprived center of the spheroids,indicating that the mevalonate pathway provides essential metabolicfunctions under these conditions. Surprisingly, sensitivity towardmevalonate pathway inhibition was independent of p21, suggestingthat the protective effect of wt p53 is independent of its role astranscriptional inducer of this target. Indeed, it has been shown thatan acetylation-deficient form of p53 that is unable to induce p21expression retains important tumor-suppressive functions (45).

We also demonstrate that p53-dependent metabolic rewiring of themevalonate pathway supports the synthesis of ubiquinone, an impor-tant electron transport molecule of the ETC (46). Previous studiesindicate that nutrient deprivation increases dependency of cancer cellson ETC activity (47, 48), particularly for the generation of aspartate asprecursor for pyrimidine synthesis (49). DHODH, the enzyme con-verting dihydroorotate to orotate during UMP synthesis, requireselectron transfer via ubiquinone and has been shown to support thegrowth of respiration-deficient tumors (39). This suggests that ubi-quinone synthesis by the mevalonate pathway supports pyrimidine

synthesis, particularly when efficient electron transport is hampered bylow oxygen availability. Ubiquinone deprivation also induces oxidativestress, especially when demand for biosynthetic reactions that deliverelectrons to the ETC is high. We found that statin-induced cell deathwas prevented by antioxidants or by the addition of nucleosides oruridine, which allow cells to switch to the salvage pathway, suggestingthat reducing de novo pyrimidine synthesis prevents ROS formationand cell death. Furthermore, the antimetabolite 5-FU, which blocksdTMP synthesis and induces DNA and RNA damage, sensitizedp53 wt SPC to statin treatment. 5-FU may impose additional strainon pyrimidine biosynthesis and/or increase oxidative stress, both ofwhich would enhance the dependency of cancer cells on ubiquinone.While clinical trials combining statins with 5-FU in colorectal ade-nocarcinoma have produced some promising results (50), our studysuggests that p53 status could determine the outcome of mevalonatepathway inhibition in colorectal adenocarcinoma.

The dependence of p53-deficient cancer cells on mevalonate path-way activity was also confirmed in apc�/� intestinal tumor organoids,where deletion of p53, either alone or in combination with Krasactivation, induced sensitivity toward statin treatment. Addition ofubiquinone or nucleosides restored growth of p53-deficient tumororganoids in the presence of statins, suggesting that cells requiremevalonate pathway–derived ubiquinone to counteract oxidativestress and support biosynthetic reactions. Indeed, LGR5þ intestinalstem cells are enriched for gene expression signatures linked to purineand pyrimidine metabolism (51) and are highly dependent on mito-chondrial metabolism (52).

We also found that statins blockKras-dependent hyperproliferationin Apc-deficient intestinal crypts. However, in contrast to our findingsin SPC and organoids, this was associated with reduced cholesterolrather than ubiquinone synthesis. Cholesterol is required for mem-brane synthesis (32) and could be the major metabolic output of themevalonate pathway in rapidly proliferating tissues. Moreover, cellswithin the intestinal mucosa may not be exposed to metabolic dep-rivation, as they have access to the nutrient-rich contents of theintestinal lumen, including nucleosides generated by the degradationof diet-derived nucleic acids.

Together, our findings reveal a novel function of the mevalonatepathway in supporting the synthesis of ubiquinone for electrontransfer and pyrimidine biosynthesis in p53-deficient cancer cellsexposed to environmental stress. However, our results also show thatthe dependence on mevalonate pathway–derived metabolites isdetermined by environmental context. Mevalonate pathway inhi-bition may therefore be most effective under conditions of nutrientand oxygen deprivation. Beneficial effects of mevalonate pathwayinhibitors have already been demonstrated in several cancer entities,including colorectal adenocarcinoma (53, 54). The results of thisstudy indicate that mevalonate pathway inhibitors may need to becombined with treatments that induce metabolic stress, such asantiangiogenic therapy.

Disclosure of Potential Conflicts of InterestO.J. Sansom reports receiving a commercial research grant from AstraZeneca,

Novartis, and Cancer Research Technology. No potential conflicts of interest weredisclosed by the other authors.

Authors’ ContributionsConception and design: I. Kaymak, O.J. Sansom, A. SchulzeDevelopment of methodology: I. Kaymak, W. SchmitzAcquisition of data (provided animals, acquired and managed patients, providedfacilities, etc.): I. Kaymak, C.R. Maier, W. Schmitz, A.D. Campbell, B. Dankworth,C.P. Ade, M. Paauwe, H. Marouf, D.M. Gay, G.H. McGregor






Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): I. Kaymak, C.R. Maier, W. Schmitz, A.D. Campbell,B. Dankworth, C.P. Ade, S. Walz, M. Paauwe, C. Kalogirou, M.T. Rosenfeldt,D.M. Gay, A. SchulzeWriting, review, and/or revision of the manuscript: I. Kaymak, W. Schmitz,A.D. Campbell, C. Kalogirou, D.M. Gay, O.J. Sansom, A. SchulzeAdministrative, technical, or material support (i.e., reporting or organizing data,constructing databases): I. KaymakStudy supervision: M.T. Rosenfeldt, O.J. Sansom, A. Schulze

AcknowledgmentsWe thank B.Vogelstein (Johns Hopkins University, Baltimore, MD), K.Vousden

(The Francis Crick Institute, London), and M. Dobbelstein (University of G€ottingen,G€ottingen, Germany) for providing p53 and p21 isogenic colon cancer cell lines,respectively. We thank C. Sch€ulein-V€olk and U. Eilers for help with automated cell

counting. We also thank S. Janaki Raman and M.T. Snaebj€ornsson for criticallyreading the manuscript. This study was funded by grants from the German ResearchFoundation FOR2314 and SCHU2670-1 (to A. Schulze), the Graduate School of LifeSciences W€urzburg (to I. Kaymak), GRK2243 (to C.R. Maier), the Rosetrees Trust (toG. McGregor), the CRUK Grand Challenge (to M. Paauwe and D.M. Gay), and corefunding to the Beatson Institute from Cancer Research UK (A17196 to O.J. Sansom).

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

Received March 4, 2019; revised October 1, 2019; accepted November 14, 2019;published first November 19, 2019.

References1. DenkoNC. Hypoxia, HIF1 and glucosemetabolism in the solid tumour. Nat Rev

Cancer 2008;8:705–13.2. Floter J, Kaymak I, Schulze A. Regulation of metabolic activity by p53.

Metabolites 2017;7:pii:E21.3. Kruiswijk F, Labuschagne CF, VousdenKH. p53 in survival, death andmetabolic

health: a lifeguard with a licence to kill. Nat Rev Mol Cell Biol 2015;16:393–405.4. Kawauchi K, Araki K, Tobiume K, Tanaka N. p53 regulates glucose metabolism

through an IKK-NF-kappaB pathway and inhibits cell transformation. Nat CellBiol 2008;10:611–8.

5. Hitosugi T, Zhou L, Elf S, Fan J, KangHB, Seo JH, et al. Phosphoglyceratemutase1 coordinates glycolysis and biosynthesis to promote tumor growth. Cancer Cell2012;22:585–600.

6. BensaadK, TsurutaA, SelakMA,VidalMN,NakanoK, Bartrons R, et al. TIGAR,a p53-inducible regulator of glycolysis and apoptosis. Cell 2006;126:107–20.

7. Jiang P, Du W, Wang X, Mancuso A, Gao X, Wu M, et al. p53 regulatesbiosynthesis through direct inactivation of glucose-6-phosphate dehydrogenase.Nat Cell Biol 2011;13:310–6.

8. Ros S, Floter J, Kaymak I, Da Costa C, Houddane A, Dubuis S, et al. 6-Phosphofructo-2-kinase/fructose-2,6-biphosphatase 4 is essential for p53-nullcancer cells. Oncogene 2017;36:3287–99.

9. Matoba S, Kang JG, Patino WD, Wragg A, Boehm M, Gavrilova O, et al. p53regulates mitochondrial respiration. Science 2006;312:1650–3.

10. Maddocks OD, Berkers CR,Mason SM, Zheng L, Blyth K, Gottlieb E, et al. Serinestarvation induces stress and p53-dependent metabolic remodelling in cancercells. Nature 2013;493:542–6.

11. Tajan M, Hock AK, Blagih J, Robertson NA, Labuschagne CF, Kruiswijk F, et al.A role for p53 in the adaptation to glutamine starvation through the expression ofSLC1A3. Cell Metab 2018;28:721–36.

12. Wang X, Sato R, BrownMS, Hua X, Goldstein JL. SREBP-1, a membrane-boundtranscription factor released by sterol-regulated proteolysis. Cell 1994;77:53–62.

13. SatoT, StangeDE, FerranteM,Vries RG,Van Es JH,Vanden Brink S, et al. Long-term expansion of epithelial organoids from human colon, adenoma, adeno-carcinoma, and Barrett's epithelium. Gastroenterology 2011;141:1762–72.

14. Cammareri P, Vincent DF, Hodder MC, Ridgway RA, Murgia C, Nobis M, et al.TGFbeta pathway limits dedifferentiation following WNT and MAPK pathwayactivation to suppress intestinal tumourigenesis. Cell Death Differ 2017;24:1681–93.

15. Bunz F, Dutriaux A, Lengauer C, Waldman T, Zhou S, Brown JP, et al.Requirement for p53 and p21 to sustain G2 arrest after DNA damage. Science1998;282:1497–501.

16. Sutherland R, Freyer J,Mueller-KlieserW,Wilson R,Heacock C, Sciandra J, et al.Cellular growth and metabolic adaptations to nutrient stress environments intumor microregions. Int J Radiat Oncol Biol Phys 1986;12:611–5.

17. Sullivan LB, Gui DY, Hosios AM, Bush LN, Freinkman E, Vander Heiden MG.Supporting aspartate biosynthesis is an essential function of respiration inproliferating cells. Cell 2015;162:552–63.

18. Horton JD,Goldstein JL, BrownMS. SREBPs: activators of the complete programof cholesterol and fatty acid synthesis in the liver. J Clin Invest 2002;109:1125–31.

19. Horton JD, Shah NA,Warrington JA, Anderson NN, Park SW, BrownMS, et al.Combined analysis of oligonucleotide microarray data from transgenic andknockout mice identifies direct SREBP target genes. Proc Natl Acad Sci U S A2003;100:12027–32.

20. Cancer Genome Atlas Network. Comprehensive molecular characterization ofhuman colon and rectal cancer. Nature 2012;487:330–7.

21. Moon SH, Huang CH, Houlihan SL, Regunath K, Freed-Pastor WA, Morris JPIV, et al. p53 represses the mevalonate pathway to mediate tumor suppression.Cell 2019;176:564–80.

22. Gerin I, Clerbaux LA, Haumont O, Lanthier N, Das AK, Burant CF, et al.Expression of miR-33 from an SREBP2 intron inhibits cholesterol export andfatty acid oxidation. J Biol Chem 2010;285:33652–61.

23. Najafi-Shoushtari SH, Kristo F, Li Y, Shioda T, Cohen DE, Gerszten RE, et al.MicroRNA-33 and the SREBP host genes cooperate to control cholesterolhomeostasis. Science 2010;328:1566–9.

24. Porstmann T, Santos CR, Griffiths B, Cully M, Wu M, Leevers S, et al. SREBPactivity is regulated by mTORC1 and contributes to Akt-dependent cell growth.Cell Metab 2008;8:224–36.

25. Porstmann T, Griffiths B, Chung YL, Delpuech O, Griffiths JR, DownwardJ, et al. PKB/Akt induces transcription of enzymes involved in cholesteroland fatty acid biosynthesis via activation of SREBP. Oncogene 2005;24:6465–81.

26. Arsham AM, Howell JJ, Simon MC. A novel hypoxia-inducible factor-independent hypoxic response regulating mammalian target of rapamycin andits targets. J Biol Chem 2003;278:29655–60.

27. Liu L, Cash TP, Jones RG, Keith B, Thompson CB, Simon MC. Hypoxia-induced energy stress regulates mRNA translation and cell growth. Mol Cell2006;21:521–31.

28. Sundqvist A, Bengoechea-Alonso MT, Ye X, Lukiyanchuk V, Jin J, HarperJW, et al. Control of lipid metabolism by phosphorylation-dependent deg-radation of the SREBP family of transcription factors by SCF(Fbw7).Cell Metab 2005;1:379–91.

29. DeBose-Boyd RA, Ye J. SREBPs in lipid metabolism, insulin signaling, andbeyond. Trends Biochem Sci 2018;43:358–68.

30. ShammaA, Takegami Y,Miki T, Kitajima S, NodaM,Obara T, et al. Rb regulatesDNA damage response and cellular senescence through E2F-dependent sup-pression of N-ras isoprenylation. Cancer Cell 2009;15:255–69.

31. Freed-Pastor WA, Mizuno H, Zhao X, Langerod A, Moon SH, Rodriguez-Barrueco R, et al. Mutant p53 disrupts mammary tissue architecture via themevalonate pathway. Cell 2012;148:244–58.

32. Mullen PJ, Yu R, Longo J, Archer MC, Penn LZ. The interplay between cellsignalling and the mevalonate pathway in cancer. Nat Rev Cancer 2016;16:718–31.

33. Dallner G, Sindelar PJ. Regulation of ubiquinone metabolism. Free Radic BiolMed 2000;29:285–94.

34. Kamphorst JJ, Chung MK, Fan J, Rabinowitz JD. Quantitative analysis of acetyl-CoA production in hypoxic cancer cells reveals substantial contribution fromacetate. Cancer Metab 2014;2:23.

35. Brunelle JK, Bell EL, Quesada NM, Vercauteren K, Tiranti V, Zeviani M, et al.Oxygen sensing requires mitochondrial ROS but not oxidative phosphorylation.Cell Metab 2005;1:409–14.

36. Longley DB, Harkin DP, Johnston PG. 5-fluorouracil: mechanisms of action andclinical strategies. Nat Rev Cancer 2003;3:330–8.

37. elMarjou F, Janssen KP, Chang BH, LiM,Hindie V, Chan L, et al. Tissue-specificand inducible Cre-mediated recombination in the gut epithelium. Genesis 2004;39:186–93.

Kaymak et al.





38. McGregor G, Campbell A, Fey S, Tumanov S, Sumpton D, Rodriguez Blanco G,et al. Targeting the metabolic response to statin-mediated oxidative stressproduces a synergistic anti-tumor response. Cancer Res 2019;80.

39. Bajzikova M, Kovarova J, Coelho AR, Boukalova S, Oh S, Rohlenova K, et al.Reactivation of dihydroorotate dehydrogenase-driven pyrimidine biosynthesisrestores tumor growth of respiration-deficient cancer cells. Cell Metab 2019;29:399–416.

40. Eales KL, Hollinshead KE, Tennant DA. Hypoxia and metabolic adaptation ofcancer cells. Oncogenesis 2016;5:e190.

41. Pisarsky L, Bill R, Fagiani E, Dimeloe S, Goosen RW, Hagmann J, et al. Targetingmetabolic symbiosis to overcome resistance to anti-angiogenic therapy. Cell Rep2016;15:1161–74.

42. Allen E, Mieville P, Warren CM, Saghafinia S, Li L, Peng MW, et al. Metabolicsymbiosis enables adaptive resistance to anti-angiogenic therapy that is depen-dent on mTOR signaling. Cell Rep 2016;15:1144–60.

43. Cheng T, Sudderth J, Yang C, Mullen AR, Jin ES, Mates JM, et al. Pyruvatecarboxylase is required for glutamine-independent growth of tumor cells.Proc Natl Acad Sci U S A 2011;108:8674–9.

44. Cardaci S, Zheng L,MacKay G, van den Broek NJ,MacKenzie ED, Nixon C, et al.Pyruvate carboxylation enables growth of SDH-deficient cells by supportingaspartate biosynthesis. Nat Cell Biol 2015;17:1317–26.

45. Li T, Kon N, Jiang L, Tan M, Ludwig T, Zhao Y, et al. Tumor suppression in theabsence of p53-mediated cell-cycle arrest, apoptosis, and senescence. Cell 2012;149:1269–83.

46. Wang Y, Hekimi S. Understanding ubiquinone. Trends Cell Biol 2016;26:367–78.

47. Birsoy K, Possemato R, Lorbeer FK, Bayraktar EC, Thiru P, Yucel B, et al.Metabolic determinants of cancer cell sensitivity to glucose limitation andbiguanides. Nature 2014;508:108–12.

48. Zhang B, Wang J, Wang X, Zhu J, Liu Q, Shi Z, et al. Proteogenomic charac-terization of human colon and rectal cancer. Nature 2014;513:382–7.

49. Gui DY, Sullivan LB, Luengo A, Hosios AM, Bush LN, Gitego N, et al.Environment dictates dependence on mitochondrial complex I for NADþ andaspartate production and determines cancer cell sensitivity to metformin.Cell Metab 2016;24:716–27.

50. Lee J, JungKH, Park YS,Ahn JB, Shin SJ, Im SA, et al. Simvastatin plus irinotecan,5-fluorouracil, and leucovorin (FOLFIRI) as first-line chemotherapy in meta-static colorectal patients: a multicenter phase II study. Cancer ChemotherPharmacol 2009;64:657–63.

51. Lindeboom RG, van Voorthuijsen L, Oost KC, Rodriguez-Colman MJ, Luna-VelezMV, Furlan C, et al. Integrative multi-omics analysis of intestinal organoiddifferentiation. Mol Syst Biol 2018;14:e8227.

52. Rodriguez-ColmanMJ, ScheweM, Meerlo M, Stigter E, Gerrits J, Pras-Raves M,et al. Interplay between metabolic identities in the intestinal crypt supports stemcell function. Nature 2017;543:424–7.

53. Liu JC, Hao WR, Hsu YP, Sung LC, Kao PF, Lin CF, et al. Statins dose-dependently exert a significant chemopreventive effect on colon cancer inpatients with chronic obstructive pulmonary disease: a population-based cohortstudy. Oncotarget 2016;7:65270–83.

54. Sehdev A, Shih YC, Huo D, Vekhter B, Lyttle C, Polite B. The role of statins forprimary prevention in non-elderly colorectal cancer patients. Anticancer Res2014;34:5043–50.






2020;80:189-203. Published OnlineFirst November 19, 2019.Cancer Res Irem Kaymak, Carina R. Maier, Werner Schmitz, et al. Metabolic StressSynthesis and Survival in p53-Deficient Cancer Cells Exposed to Mevalonate Pathway Provides Ubiquinone to Maintain Pyrimidine

Updated version

10.1158/0008-5472.CAN-19-0650doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2019/11/19/0008-5472.CAN-19-0650.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/80/2/189.full#ref-list-1

This article cites 53 articles, 9 of which you can access for free at:

E-mail alerts related to this article or journal.Sign up to receive free email-alerts

Subscriptions

Reprints and

[email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

Permissions

Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/80/2/189To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-19-0650

http://cancerres.aacrjournals.org/content/suppl/2019/11/19/0008-5472.CAN-19-0650.DC1

http://cancerres.aacrjournals.org/content/80/2/189.full#ref-list-1

http://cancerres.aacrjournals.org/cgi/alerts

mailto:[email protected]

http://cancerres.aacrjournals.org/content/80/2/189


Documents

Mevalonate Pathway Provides Ubiquinone to Maintain ... · pyrimidine nucleotide biosynthesis and induced oxidative stress and apoptosis in p53-deﬁcient cancer cells in spheroid