Biochimica et Biophysica Acta xxx (2013) xxx–xxx

BBAMCB-57479; No. of pages: 7; 4C: 2, 3, 5, 6

Contents lists available at SciVerse ScienceDirect

Biochimica et Biophysica Acta

j ourna l homepage: www.e lsev ie r .com/ locate /bba l ip

Cancer cells incorporate and remodel exogenous palmitate into structural andoncogenic signaling lipids☆

Sharon M. Louie a,1, Lindsay S. Roberts a,1, Melinda M. Mulvihill a, Kunxin Luo b, Daniel K. Nomura a,⁎a Program in Metabolic Biology, Department of Nutritional Sciences and Toxicology, University of California, Berkeley, Berkeley, CA 94720, USAb Department of Molecular and Cell Biology, University of California, Berkeley, Berkeley, CA 94720, USA

Abbreviations: FFA, Free fatty acid; QQQ-LC/Mchromatography-mass spectrometry; MAGL, monoacethylene glycol; EMT, epithelial-to-mesenchymal tramonitoring; AC, acyl carnitines; PC, phospholipids suLPC, lysophosphatidyl cholines; PA, phosphatidic aacids; PE, phosphatidyl ethanolamines; LPE, lysophophosphatidyl inositols; PG, phosphatidyl glycerolsTAG, triacylglycerols; DAGs, diacylglycerols; SM, sph1-phosphate; PAF, platelet activating factor; CPT, car☆ This article is part of a Special Issue entitledDysregulat⁎ Corresponding author. Tel.: +1 5106437258.

E-mail address: dnomura@berkeley.edu (D.K. Nomura1 These authors contributed equally.

1388-1981/$ – see front matter © 2013 Elsevier B.V. All rhttp://dx.doi.org/10.1016/j.bbalip.2013.07.008

Please cite this article as: S.M. Louie, et al., Clipids, Biochim. Biophys. Acta (2013), http://

a b s t r a c t

a r t i c l e i n f o

Article history:Received 17 May 2013Received in revised form 12 July 2013Accepted 12 July 2013Available online xxxx

Keywords:Cancer metabolismLipid signalingMetabolomicsLysophosphatidic acidCeramide-1-phosphatePlatelet activating factor

De novo lipogenesis is considered the primary source of fatty acids for lipid synthesis in cancer cells, even in thepresence of exogenous fatty acids. Here, we have used an isotopic fatty acid labeling strategy coupled withmetabolomic profiling platforms to comprehensivelymap palmitic acid incorporation into complex lipids in can-cer cells. We show that cancer cells and tumors robustly incorporate and remodel exogenous palmitate intostructural and oncogenic glycerophospholipids, sphingolipids, and ether lipids.We also find that fatty acid incor-poration into oxidative pathways is reduced in aggressive human cancer cells, and instead shunted into pathwaysfor generating structural and signaling lipids. Our results demonstrate that cancer cells do not solely rely onde novo lipogenesis, but also utilize exogenous fatty acids for generating lipids required for proliferationand protumorigenic lipid signaling. This article is part of a special issue entitled Dysregulated Lipid Metabolismin Cancer.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Heightened de novo lipogenesis is a fundamental hallmark of nearlyall cancers and is required for cellular transformation and cancer pro-gression [1]. Fatty acid synthase, the enzyme responsible for de novosynthesis of fatty acids, is upregulated across multiple types of humantumors and blocking FASN has been shown to attenuate cell prolifera-tion, tumorigenicity, and cancer malignancy [1]. Early studies, usingradioactivity-based methods measuring bulk lipids, have shown thatde novo synthesis of fatty acids from glucose and other carbon sourcesaccount for 93% of the total cellular lipid content in certain cancertypes [2]. Cancer cells are thus thought to rely almost solely on de novolipogenesis, rather than exogenous fatty acids for generation of cellularlipids [3]. In addition to lipogenic pathways that subserve cancer prolifer-ation, we have previously shown that aggressive human cancer cells alsoupregulate lipolytic pathways to mobilize free fatty acids to generate

S, triple quadrupole liquidylglycerol lipase; PEG, poly-nsition; SRM, single reactionch as phosphatidyl cholines;cids; LPA, lysophosphatidicsphatidyl ethanolamines; PI,; PS, phosphatidyl serines;ingomyelin; C1P, ceramide-nitine palmitoyl transferaseed LipidMetabolism in Cancer.

).

ights reserved.

ancer cells incorporate and rdx.doi.org/10.1016/j.bbalip.20

oncogenic signaling lipids that in-turn fuel aggressive features of cancer[4].We found that the tumorigenic impairments conferredby inactivatinga lipolytic enzyme monoacylglycerol lipase (MAGL) in cancer cells, couldbe rescued by exogenous fatty acids in situ or by high-fat diet feedingin vivo [4]. These results put forth the possibility that exogenous fattyacids, despite the dominant role of de novo fatty acid synthesis, may alsoplay an important role in cancer pathogenesis.

In this study, we investigated whether cancer cells are capable ofincorporating exogenous free fatty acids (FFA) and used advancedmetabolomic platforms to comprehensively understand how FFAsare remodeled within cancer cells, and whether this exogenousFFA-derived lipid metabolism is altered during cancer progression.

2. Materials and methods

2.1. Cell culture

C8161, MUM2C, 231MFP, MCF7, SKOV3, OVCAR3, PC3, and LNCaPcells were obtained from Benjamin Cravatt at The Scripps Research In-stitute or from ATCC. MCF10A, M2, M2T, and M4 cells were obtainedfrom Stefano Piccolo at the University of Padua [5]. Cells were culturedas previously described [4–6].

2.2. Isotopic fatty acid labeling of cancer cells and mice

Cancer cells were seeded (1.5 × 106 cells) and upon adherence, cellswere serum starved and treated with d0-palmitic acid or (7,7,8,8-d4)-palmitic acid (10 μM in 0.5% BSA) for 4 h. Cells were then washed

emodel exogenous palmitate into structural and oncogenic signaling13.07.008

untargeted metabolomicsbioinformatic analysis

of isotopic incorporation(XCMS)

rt

rt

rt

m/z 100

m/z 101

m/z 1200

lipidomeextraction

targeted metabolomicssingle reaction monitoring (SRM)

of representative lipidswith isotopic incorporation

mass spectrometer

m/z

ms1

m/z

ms2*

ms1 ms2*

rtd0-C16:0 FFA

d4-C16:0 FFA

FFA d4-FFA

FFA d4-FFA

FFA d4-FFA

O

HO

O

HO

DD

DD

+

+

Fig. 1.Metabolomic mapping of exogenously-derived isotopic FFA metabolism in cancer cells. Cells were treated with either d0-C16:0 FFA or d4-C16:0 FFA for 4 h. Lipids were extractedand analyzed by LC-MS using targeted SRM-based approaches and untargeted approaches. The large datasets resulting from untargeted metabolomics were analyzed by XCMS Online todetermine masses that were altered between d0- and d4-C16:0 FFA to identify isotopic-FFA-incorporated lipids.

2 S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

twice in phosphate-buffered saline (PBS) and harvested by scraping.Cells were collected on ice and centrifuged at 1000 ×g and cell pelletswere frozen at−80 °C until lipid extraction.

For isotopic fatty acid labeling of mouse tumor xenografts in vivo, M4cancer cells (2 × 106 cells)were subcutaneously injected into the flank ofimmune-deficient SCID mice and tumors were grown out to 800 mm3.Mice were treated with vehicle or d4-palmitic acid (100 mg/kg oral ga-vage in polyethylene glycol 300 (PEG300)) for 4 h. Mice were thensacrificed and tumors were removed and flash frozen.

2.3. Metabolomic analyses

Cell and tumor lipids were extracted as previously described [4].Briefly, cells and tumors were extracted in 2:1:1 chloroform:methanol:phosphate-buffered saline with inclusion of internal standards(10 nmoles of dodecylglycerol and 10 nmoles of pentadecanoic acid).The organic layer was collected and the remaining aqueous layer wasacidified with 0.1% formic acid and re-extracted in chloroform. Organiclayers were combined and dried down under a stream of nitrogen.Dried extracts were resolubilized in 120 μl of chloroform and an ali-quot (10 μl) was injected onto an Agilent 6430 triple quadrupole(QQQ)-liquid chromatography-mass spectrometry (LC-MS) instrument.

Targeted mass spectrometry analysis was performed as previouslydescribed [7]. Briefly, single-reaction monitoring (SRM) programs were

Fig. 2.Mapping exogenous isotopic FFA-derived lipidmetabolism inhuman cancer cells. (A–E) Shcancer cells. For the volcano plot, points that are to the left of the dotted line (P b 0.05) representlabeled cells. All points to the right of the dotted line (P b 0.05) represent ions that had significangroup, i.e. d4-incorporated lipids. In total, at least ~5000–10,000 ionswere detected and analyzedlabeled M4, 231MFP, C8161, SKOV3, or PC3 cells. The y-axis denotes fold-change between raw ibetween d0 versus d4-C16:0 FFA-labeled samples. For the ions that exhibited no background peathis value to be 1 to obtain a fold-change value compared to the raw integration values of d4-C16:change value by dividing the ion intensity for thed4-C16:0 FFA compared to d0-C16:0 FFA groupsaggressive (MCF7, OVCAR3, LNCaP, MUM2C) or non-transformed (MCF10A) cells compared tomap, relative levels of each d4-incorporated lipid metabolite are shown (darker blue shading ccolor-coded red for significantly higher, blue for significantly lower, and gray for unchanged d4-non-aggressive (MCF7, MCF10A, OVCAR3, LNCaP, and MUM2C, respectively) cells (P b 0.05). (Fmice bearing a tumor xenograft fromM4 cells. Mice were subcutaneously injected with 2 × 106

(polyethylene glycol 300 (PEG300)) or d4-C16:0 FFA (100 mg/kg in PEG) by oral gavage (4 h). TuFor A–F, thosemetaboliteswhere therewas a background peak for the d4-lipidm/z in the d0-C16:and d4-C16:0 FFA-treated groups. For all lipid shownhere, any backgroundpeak for a d4-lipid detenatural isotopic abundance of the lipid. We have only presented here the d4-incorporated lipids tgroup compared to the d0-C16:0 FFA-treated group. All data fromA–E is shown in Supplemental Tbiological replicates. Data in (F) aremean ± standard error of n = 4–6biological replicates. Signifitreatment.

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.2

derived from nonisotopic standards and databases. SRM programs forisotopic lipids were based on the ms2 fragments and optimized collisionenergies of nonisotopic standards. Metabolites were quantified by inte-grating the area under the curve and normalized against internal stan-dards and external standard curves.

Untargeted mass spectrometry analysis was performed by LC-MS inscanning mode collectingmass spectral data fromm/z 50 to 1200. Datafiles were extracted as m/zdata files and analyzed by XCMS Online(xcmsserver.nutr.berkeley.edu) to identify isotopic fatty acid incorpora-tion into cellular lipids [4,8]. Structures of metabolites from untargetedanalysis were identified based on database searches (METLIN [9]) andincorporation of d4-palmitate, as well as co-elution of metabolites withstandards within the same class of metabolites.

Lipidswere separated by reverse phase chromatographywith a LunaC5 column (Phenomenex) starting with 100% 95:5 water:methanolwith a gradient to 100% 60:35:5 isopropanol:methanol:water as previ-ously described. Formic acid (0.1%) with 50 mM ammonium formateor ammonium hydroxide (0.1%) was added for positive and negativeionizationmode, respectively.Metaboliteswere quantified by integratingthe area under the curve, normalizing to internal standards, and then cal-culating levels based on external standard curves with representativelipid standards from each lipid species. For those metabolites for whichthere was a background peak for the isotopic d4-lipid in the d0-C16:0FFA-treated group, we subtracted the average of the background from

ownon the left are all ions detected in 231MFP,M4, SKOV3, PC3, or C8161 aggressivehumanions that were not statistically altered in levels between d0-C16:0 FFA versus d4-C16:0 FFA-tly higher ion intensity in the d4-C16:0 FFA labeled group compared to d0-C16:0 FFA labeledbetween targeted anduntargeted analysis comparing d0-C16:0 FFA labeled to d4-C16:0 FFAntegrated values of isotopically-incorporated ions by either targeted or untargeted analysisk corresponding to the m/z of the d4-lipid in the d0-C16:0 FFA-treated cells, we considered0 FFA-treated cells. For the ions for which therewas a background peak, we obtained a fold-. The heat-map on the right shows relative levels of d4-C16:0 FFA-incorporated lipids in non-aggressive (231MFP, M2T, M4, SKOV3, PC3, C8161) or transformed (M2) cells. In the heat-orresponds to higher level of metabolite). The lipid designations next to the heat map aremetabolites in aggressive cancer cells (231MFP, M4, SKOV3, PC3, and C8161) compared to) Shown are lipids species that exhibited significant d4-C16:0 FFA incorporation in vivo inM4 cells and tumors were grown out to ~800–1000 mm3. Mice were treated with vehiclemors were harvested and lipids were extracted and analyzed by SRM-basedmetabolomics.0 FFA-treated cells, the average of the background ion intensitywas subtracted fromboth d0cted in d0-C16:0 FFA-treated cellswas assumed to either be a coeluting isobaricmetabolite orhat showed N5-fold significantly (P b 0.05) higher ion intensity in the d4-C16:0 FFA-treatedable 1 and certain lipids are quantified in Fig. 3. Data in (A–E) are average values of n = 4–6cance in (F) is represented as *P b 0.05 in d4-C16:0 FFA-treatedmice comparedwith vehicle-

emodel exogenous palmitate into structural and oncogenic signaling013.07.008

3S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

both d0- and d4-C16:0 FFA-treated groups. For all metabolites, the isoto-pic d4-lipid peak for the d0-C16:0 FFA-treated groupwas less than 20% ofthe d4-C16:0 FFA-treated group.

3. Results

3.1. Isotopic fatty acid tracing-basedmetabolomics reveals that cancer cellsincorporate exogenous fatty acids into structural and signaling lipids

To understand how cancer cells incorporate exogenous lipids andwhether this lipid metabolism is altered during cancer progression,

C16:0 FFAC16:0 ACC16:0/C18:1 PCC16:0 LPCC16:0/C18:1 PA C16:0/C16:0 PA C16:0/C18:0 PA C16:0 LPAC16:0/C16:0 PSC16:0/C18:1 PIC16:0/C18:1 CeramideC16:0/C18:1 SMC16:0/C18:1 C1PC16:0 PAF/C18:0 LPC C16:0/18:1 alkyl PA C16:0/18:1 alkyl PIC16:0/C14:1 alkyl PEC16:0/C12:0 alkyl PEC16:0/C14:0 alkyl PEC16:0/C20:2 alkyl PG

C16:0 FFAC16:0 ACC16:0/C16:0 PCC16:0/C18:1 PCC16:0 LPCC16:0/C18:1 PA C16:0 LPAC16:0/C16:0 PEC16:0/C16:1 PEC16:0/C22:4 PEC16:0/C22:6 PEC16:0/C16:0 PIC16:0/C16:1 PIC16:0/C18:1 PIC16:0/C20:1 PIC16:0/C18:1 PSC16:0/C18:2 PSC18:0/C22:6 PSC16:0/C20:1 PGC16:0/C22:6 PGC18:0/C22:6 PGC16:0/C16:0/C18:1 TAG C16:0/C18:1 CeramideC16:0/C18:1 SMC16:0/C18:1 C1PC16:0/C18:1 alkyl PCC16:0 PAF/C18:0 LPC C16:0 lysoPAFC16:0/C18:1 alkyl PA C16:0 alkyl LPAC16:0/C16:0 alkyl PEC16:1/C20:4 alkyl PEC16:0/C22:0 alkyl PEC16:0/C18:1 alkyl PIC16:1/C16:1 alkyl PSC16:1/C22:2 alkyl PS

d4-palmitate incorporation into breast cancer cells

10-910-810-710-610-510-410-310-210-1100100

101

102

103

104

105

106

p-value

fold

-cha

nge

825alkyl PI

663alkyl PA

259FFA

839PI

677PA

486lysoPAF

740PS

709SM

413LPA

500LPC 404

AC

528PAF/LPC

542ceramide

651PA

764PC

620C1P

765PG

679PAalkyl PE

624

654alkyl PE

652alkyl PE

MC

F7

231M

FP

- + - + d4-incorporated lipids

1100

101

102

103

104

105

106

107

fold

-cha

nge

d4-palmitate incorporation into ovarian cancer cells

10-1210-1010-810-610-410-2100100

101

102

103

104

105

106

p-value

fold

-cha

nge

OV

CA

R3

SK

OV

3

d4-FFA - + - + d4-incorporated lipidsFFAsand ACs

phospho-lipids

sphingo-lipids

etherlipids

neutral lipids

d4-FFA

867PI

663alkyl PA

399alkyl LPA

680alkyl PE

752PE

486lysoPAF

500LPC413LPA

811PI692

PE528

PAF/LPC

259FFA

404AC

542ceramide

839PI

817PI

620C1P

840-PS

779PG

677PA

825alkyl PI

804PS

764PS

722alkyl PS

858TAG

799PG

738PC

772PE 728

alkyl PE

764alkyl PE

750alkyl PC

764PS

696PE

766PS

827PG

709SM

FFAsand ACs

phospho-lipids

sphingo-lipids

etherlipids

incorporation into 231MFP cellular lipids

incorporation into SKOV3 cellular lipids

d4-incorporation into C8161 lipids

10-1410-1210-1010-810-610-410-2100100

101

102

103

104

105

106

107

108

p-value

fold

-cha

nge

C16:0 FFAC16:0 ACC16:0/C18:1 PCC16:0/C14:1 PCC16:0 LPCC16:0/C18:1 PEC16:0 LPEC16:0/C18:1 PAC16:0 LPAC16:0/C18:1 PSC16:0/C14:0 PSC16:0/C14:1 PSC16:0/C16:0 PSC16:0/C16:1 PSC16:0/C20:2 PSC16:0/C16:0 PIC16:0/C18:1 PGC16:0/C16:0/C18:1 TAG C16:0/C20:5 DAGC16:0/C18:1 CeramideC16:0/C18:1 SMC16:0/C18:1 C1PC16:0 PAF/C18:0 LPC C16:0/C14:0 alkyl PEC16:0/C18:1 alkyl PSC16:0/C18:2 alkyl PSC16:0/C18:1 alkyl PIC16:0/C22:2 alkyl PIC16:0/C22:1 alkyl PI

MU

M2C

C81

61

d4-FFA - + - + d4-incorporated lipidsFFAsand ACs

phospho-lipids

sphingo-lipids

etherlipids

neutral lipids

404AC

827alkyl PI

854TAG 620

C1P

750alkyl PS

636DAG

883alkyl PI

752alkyl PS 792

PS

709SM

740PS

500LPC652

alkyl PE

751PG

458LPE 677

PA817PI

413LPA

722PE

259FFA

528PAF/LPC

708PC 712

PS766

542ceramide

710PS 738

PS

764PC

d4-palmitate incorporation into melanoma cells

A B

C D

E F

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.20

we treated a panel of aggressive versus non-aggressive human cancercells from breast, ovarian, prostate, and melanoma cancers in situ withnonisotopic or isotopic palmitic acid (C16:0 free fatty acid (C16:0 FFA)),10 μMin 0.5% fatty-acid free BSA for 4 h). These aggressive human cancercells (231MFP, SKOV3, PC3, and C8161) have been previouslyshown to possess heightened motility, invasiveness, and in vivo tumorgrowth rates, compared with their non-aggressive counterparts (MCF7,OVCAR3, LNCaP, and MUM2C) [4,6]. We also profiled a human breastcancer progression model consisting of: 1) MCF10A nontransformedmammary epithelial cells; 2)MCF10A cells transformedwith the activat-ed HRAS (MCF10A-T1k cells or M2 cells); 3) M2 cells transduced with

C16:0 FFAC18:0 FFAC16:0 ACC16:0/C18:1 PCC16:0 LPCC16:0/C16:0 PAC16:0/C18:1 PAC18:0/C22:6 PAC16:0 LPAC16:0/C18:1 PEC18:0/C20:4 PEC16:0/C16:0 PIC16:0/C18:1 PIC16:0/C20:1 PIC16:0/C22:2 PIC16:0/C22:0 PIC16:0 LPIC16:0/C16:1 PSC16:0/C18:1 PSC16:0/C18:2 PSC18:0/C22:6 PSC16:0 CeramideC16:0/C18:1 SMC16:0/C18:1 C1PC16:0/C18:1 alkyl PCC16:0 PAF/C18:0 LPCC16:0/C18:1 alkyl PEC16:0/C18:1 alkyl PIC16:1/C20:0 alkyl PIC16:0/C18:1 alkyl PSC16:1/C22:2 alkyl PSC16:1/C22:4 alkyl PS

d4-palmitate incorporation into breast cancer cells

10-1010-910-810-710-610-510-410-310-210-100

p-value

458LPE

500LPC

528PAF

486lysoPAF 259

FFA

710PC

287FFA

404AC

542ceramide

616DAG

677PA828

TAG

839PI

766PS

827/855alkyl PI

800PE

854TAG

737alkyl PG

826TAG

709SM

620C1P

764PC

751PA

765alkyl PG

738PC

C16:0 FFAC18:0 FFAC16:0 ACC16:0/C18:1 PCC16:0/C14:0 PCC16:0/C16:0 PCC16:0 LPCC16:0 LPEC16:0/C18:1 PSC16:0/C18:1 PA C18:0/C22:6 PA C16:0/C18:1 PIC16:0/C16:0/C16:1 TAGC16:0/C16:0/C16:0 TAGC16:0/C16:0/C18:1 TAGC16:0/C18:1 DAGC16:0/C18:1 CeramideC16:0/C18:1 SMC16:0/C18:1 C1PC16:0 PAF/C18:0 LPCC16:0 lysoPAFC16:0/C18:1 alkyl PGC16:0/C20:1 alkyl PGC18:0/C20:4 alkyl PCC16:0/C18:1 alkyl PIC16:0/C20:1 alkyl PI

- + -+ - + - + d4-incorporated lipids

10A

M2

M2T

M4

FFAs and ACs

phospho-lipids

sphingo-lipids

etherlipids

neutrallipids

d4-FFA

LNC

aP

PC

3

d4-FFA - + - + d4-incorporated lipids

d4-palmitate incorporation into prostate cancer cells

10-1410-1210-1010-810-610-410-2100100

101

102

103

104

105

106

p-value

fold

-cha

nge

722PE

804alkyl PS

677PA

825alkyl PI

750alkyl PC817

PI893-PI

840PS

708alkyl PE

800PS

575LPI

500LPC

413LPA

752alkyl PS

828PE

883-PI

709SM

855alkyl PI

620C1P

766PS

738PS

764PS

764PC

528PAF/LPC

404AC287

FFA

651PA

751PA

259FFA

839PI

867PI

542ceramide

FFAs and ACs

phospho-lipids

sphingo-lipids

ether lipids

incorporation into M4 cellular lipids

incorporation into PC3 cellular lipids

C16:0 LPC

0

100

200

300

400

pmol

/g tu

mor

C16:0PAF/C18:0 LPC

0

20

40

60

80

100

pmol

/g tu

mor

C16:0/C18:1 C1P

pmol

/g tu

mor

0

5

10

15

d4-FFA - + d4-FFA - + d4-FFA - +

** *

d4-palmitate incorporation into M4 tumors

emodel exogenous palmitate into structural and oncogenic signaling13.07.008

4 S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

the constitutively activated transcription factor TAZ S89A (M2T cells)that have been previously shown to induce epithelial-to-mesenchymaltransition (EMT), poor breast cancer prognosis, and stem-cell-like fea-tures in breast cancer; and 4) M4 (or MCF10A-CA1a) cells that aremalignant derivatives of M2 cells through spontaneous malignantevolution in vivo. These cells are highly tumorigenic, metastatic, anddisplay increased stem-like features and an upregulation of TAZ [5].We then extracted the lipidome of these cells and quantitatively mea-sured isotopic incorporation into cellular lipids using a combination oftargeted SRM-based and untargeted discovery-basedmetabolomic pro-filing [10] to globally track the isotopic incorporation and remodeling ofexogenous fatty acids into cancer cells (Fig. 1). Our SRM methods in-cluded ~60 representative lipid species that could potentially incor-porate isotopic palmitate, including phospholipids, neutral lipids,sphingolipids, and ether lipids. Our untargeted methods collectedmass spectrometry data over a large mass range (m/z 50–1200)and subsequent datasets were analyzed by the XCMS Online software[8] to integrate all detectable ions (~5000–10,000 ions), and identifysignificant alterations in the metabolomes.We combined both targetedand untargeted data to gain a global understanding of exogenous FFA-derived lipid metabolism in cancer cells and mapped these data ontometabolic pathway maps.

Ourmetabolomic profiling of isotopic palmitic acid incorporation re-vealed that cancer cells robustly incorporate exogenous fatty acids intocancer cells, which are in-turn remodeled into acyl carnitines (AC),phospholipids such as phosphatidyl cholines (PC), lysophosphatidylcholines (LPC), phosphatidic acids (PA), lysophosphatidic acids (LPA),phosphatidyl ethanolamines (PE), lysophosphatidyl ethanolamines(LPE), phosphatidyl inositols (PI), phosphatidyl glycerols (PG) andphosphatidyl serines (PS), neutral lipids such as triacylglycerols(TAG) and diacylglycerols (DAGs), sphingolipids such as ceramide,sphingomyelin (SM), and ceramide-1-phosphate (C1P), and ether lipidssuch as alkyl PCs, alkyl PEs, alkyl PIs, alkyl PGs, alkyl PSs, alkyl PIs, plateletactivating factor (PAF), and lysoPAF (Fig. 2A–E; Supplemental Table 1).These incorporated lipids not only include structural lipids (e.g. PC, LPC,SM, PE, LPE, PS, PG, PI, alkyl PCs, alkyl PEs, alkyl PIs, alkyl PGs, alkyl PSs,and alkyl PIs) and lipid stores (e.g. TAGs and DAGs), but also signalinglipids such as LPA, ceramide, DAG, and C1P. Several of these signalinglipids, such as LPA, DAG, and C1P or their associated signaling pathwayshave been shown to promote cancer pathogenicity [11]. We also findthat C16:0 FFA also contributes to the generation of C18:0 FFA(stearic acid) and is incorporated into several C18:0 FFA-containinglipids. As such, with our targeted methods monitoring the m/z 184phosphocholine ms2 fragment, we acknowledge the possibility thatC16:0 PAF (1-O-hexadecyl-2-acetyl-sn-glycero-3-phosphocholine) maybe a combination of C16:0 PAF and C18:0 LPC (1-stearoyl-2-hydroxy-sn-glycero-3-phosphocholine).

3.2. Isotopic fatty acids are incorporated into oncogenic signaling lipids intumors in vivo

Wealsowanted to investigatewhether fatty acids can be incorporat-ed in vivo into tumor xenografts in mice. Mice bearingM4 tumors weretreated with d4-C16:0 FFA (100 mg/kg oral gavage, 4 h), and isotopicincorporation into tumor lipids was measured by mass spectrometry.Consistent with our in situ studies, we found that exogenous d4-C16:0FFA was incorporated into certain lipid species including LPC, PAF, andC1P (Fig. 2F). Our studies suggest that exogenous fatty acid-derivedlipids, which include oncogenic signaling lipids PAF and C1P, are foundin tumors or tumor-associated cells in vivo.

3.3. Fatty acid incorporation into structural and oncogenic signaling lipidsare heightened in aggressive cancer cells

We next wanted to understand alterations in lipid metabolism thatmay underlie cancer progression. We therefore compared isotopic

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.2

fatty acid incorporation across aggressive versus non-aggressive cancercells frommultiple cancer types, andfiltered for isotopic lipid levels thatwere commonly altered across three out of the five aggressive can-cer cells (231MFP, M4, PC3, SKOV3, and C8161) compared to theirnon-aggressive counterparts (MCF7, MCF10A, LNCaP, OVCAR3, andMUM2C). We intriguingly found a common signature of altered lipidmetabolism shared among aggressive cancer cells in which there arelower levels of isotopically labeledACs, and increased levels of isotopical-ly labeled phospholipids such as PA, PS, PC, and PI, sphingolipids such asceramide and SM, ether lipids such as alkyl PE and alkyl PC, aswell as on-cogenic signaling lipids PAF, LPA, and C1P (Figs. 2A-E, 3; SupplementalTable 1). While we believe that these changes are reflective of reducedor heightened fatty acid incorporation into these lipids, we note that inthis comparative analysis, we cannot formally distinguish between alter-ations in synthetic and degradation rates of each lipid. Using the KEGGpathway database [12] as a guide, fatty acid incorporation into cellularlipidswasmapped onto a pathway diagram.We find that FFA incorpora-tion and remodeling into phospholipid, sphingolipid, and ether lipidsis heightened across aggressive cancer cells indicating that aggressivecancer cells rely more heavily on exogenous FFAs for cancer cell lipids(Fig. 4).

4. Discussion

Taken together, our results reveal that cancer cells incorporate andutilize exogenous fatty acids not only for generation of cellular mem-branes for cell division, but also for synthesis of signaling lipids, suchas C1P, PAF, DAG, and LPA, that have been previously shown to fuel can-cer cell pathogenicity [11]. While recent studies have shown that carni-tine palmitoyltransferase (CPT)1A or CPT1C activity promotes cellsurvival, tumor growth, or cellular motility in certain types of cancercells and that blocking CPT may be a novel cancer therapeutic strategy[13], our data would suggest that CPT and fatty acid oxidation pathwaysare attenuated during cancer progression to shunt fatty acids from beta-oxidation pathways (i.e. carnitine palmitoyltransferase (CPT)-mediatedAC production) to generate more structural and oncogenic lipids. Theseresults are reinforced by our previous genomic profiling efforts showingthat the aggressive cancer cells used in this study possess lowerlevels of CPT expression compared to their non-aggressive counter-parts (Supplemental Fig. 1) [6]. Nonetheless, blocking CPT may bean attractive therapeutic strategy for combatting less aggressive orlow-grade tumors.

Of particular interest are the exogenous fatty acids that are incorpo-rated significantly more into the signaling lipids C1P, PAF, and LPAacross several types of aggressive human cancer cells compared withtheir less aggressive counterparts. C1P is formed by ceramide kinaseand has been shown to oppose the apoptotic effects of ceramide andpromote cell proliferation and survival through activating intracellularsignaling pathways such as MEK, ERK, PI3K/AKT, and JNK, activate in-flammatory responses by activating cytosolic phospholipase A2 for gen-erating pro-inflammatory prostaglandins, and stimulate cell migrationthrough stimulating a yet unknown extracellular Gi-coupled receptor11c.PAF, an inflammatory lipid that acts through PAF receptors and causesinflammation and platelet aggregation, has also been shown to be pro-duced by melanoma cancer cells and promote invasiveness and metas-tasis through stimulating cancer cell PAF receptors in an autocrinemechanism [14]. LPA is a potent oncogenic signaling lipid that actsthrough stimulating LPA receptors leading to activation of multipledownstream effector pathways including phospholipase C, PI3K-AKT,RAS-ERK, and RHO and RAC GTPases leading to proliferation, surviv-al, migration, invasion, and increased endothelial permeability11a.Increased incorporation of exogenous fatty acids into C1P, PAF, andLPA in aggressive cancer cells can thus potentially fuel cancer initia-tion, progression, and metastasis.

Beyond the generation of these oncogenic signaling lipids, wealso show that palmitic acid incorporation into complex lipids is

emodel exogenous palmitate into structural and oncogenic signaling013.07.008

d4-C16:0 AC

0

100

200

300pm

oles

/106 c

ells

d4-C16:0 AC

pmol

es/1

06 cel

ls

0

5

10

15

20

25

d4-C16:0 AC

pmol

es/1

06 cel

ls

0

200

400

600

800

1000

d4-C16:0 AC

pmol

es/1

06 cel

ls

0

50

100

150

d4-C16:0 PAF/d4-C18:0 LPC

pmol

es/1

06 cel

ls

0

10

20

30

d4-C16:0 PAF/d4-C18:0 LPCpm

oles

/106 c

ells

0

20

40

60

80

d4-C16:0 PAF/d4-C18:0 LPC

pmol

es/1

06 cel

ls

0

20

40

60

80

100

d4-C16:0 PAF/d4-C18:0 LPC

pmol

es/1

06 cel

ls

0

10

20

30

d4-C16:0 LPA

pmol

es/1

06 cel

ls

0.00

0.05

0.10

0.15

d4-C16:0 LPA

pmol

es/1

06 cel

ls

0.00

0.02

0.04

0.06

0.08

d4-C16:0 LPA

pmol

es/1

06 cel

ls

0.00

0.05

0.10

0.15

d4-C16:0 LPA

pmol

es/1

06 cel

ls

0.00

0.02

0.04

0.06

0.08

0.10

d4-C16:0/C18:1 C1P

pmol

es/1

06 cel

ls

0

1

2

3

4

5

d4-C16:0/C18:1 C1P

pmol

es/1

06 cel

ls

0.0

0.5

1.0

1.5

2.0

2.5

d4-C16:0/C18:1 C1P

pmol

es/1

06 cel

ls

0.0

0.5

1.0

1.5

2.0

2.5

d4-C16:0/C18:1 C1P

pmol

es/1

06 cel

ls

0

1

2

3

4

5

d4-FFA - + - +

aggr.

MCF7 231MFP

d4-FFA - + - +

aggr.

MCF7 231MFP

d4-FFA - + - +

aggr.

MCF7 231MFP

d4-FFA - + - +

aggr.

MCF7 231MFP

d4-FFA - + - +

aggr.

MUM2C C8161

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

OVCAR3 SKOV3

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

LNCaP PC3

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

d4-FFA - + - +

aggr.

MUM2C C8161 MUM2C C8161 MUM2C C8161

OVCAR3 SKOV3 OVCAR3 SKOV3 OVCAR3 SKOV3

LNCaP PC3 LNCaP PC3 LNCaP PC3

*

* * *

*

* * *

*

*

*

*

* *

Fig. 3. Aggressive cancer cells increase incorporation of fatty acids into oncogenic signaling lipids and reduce incorporation into oxidative pathways. Representative lipidswith significantfatty acid incorporation from Fig. 2 are quantitated and shown as bar graphs. In comparing d4-C16:0 FFA incorporation into aggressive (C8161, PC3, SKOV3, and 231MFP cells) comparedwith non-aggressive cancer cells (MUM2C, LNCaP, OVCAR3, and MCF7 cells), we find that there is reduced incorporation into C16:0 AC and increased incorporation into phospholipids,sphingolipids, and ether lipids, including the signaling lipids C1P, PAF/LPC, and LPA. Data are average values of n = 4–6 biological replicates and are presented asmean ± standard error.Significance is represented as *P b 0.05 comparing aggressive versus non-aggressive d4-FFA groups.

5S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

globally increased in aggressive cancer cells into glycerophospholipid,sphingolipid, and ether lipid pathways. While there have beenmany studies into the bioactive roles of glycerophospholipids andsphingolipids11b, the role of ether lipids in cancer remains relativelypoorly understood, despite its established correlation with aggressivecancers [15]. It will be of future interest to understand the role ofheightened ether lipid synthesis in cancer progression.

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.20

Previous studies have indirectly suggested that cancer cells uti-lize exogenous fatty acids for energy or membrane formation.Nieman et al. showed that ovarian cancer cells use lipids derivedfrom neighboring adipocyte stores in vitro by co-culture of ovariancancer cells and adipocytes [16]. Studies have also shown that adi-pose stromal cells transplanted into mice promote tumor growthby serving as perivascular adipocyte progenitors. Intratumoral

emodel exogenous palmitate into structural and oncogenic signaling13.07.008

OH

O

O

O

O

O

OP

O

OOH

N

O

O

HOO

P

O

OOH

N

NH

O

OP

O

OOH

N

OH

NH

O

OH

OH

NH

O

OP

O

OHOH

OH

sphingosine

ceramide(apoptotic)

SM

O

O

OH

O

N

O

O

HOO

P

O

OHOH

O

O

O

O

OP

O

OHOH

acyl glycerone phosphate

1-alkyl-sn-glycero-3-phosphate

2-acetyl-1-alkyl-sn-glycero-3-phosphate

2-acetyl-1-alkyl-sn-glycerol

PAF(oncogenic-PAF receptor)

O

O

OP

O

OOH

N

O

OH

OP

O

OOH

N

O

OO

O

OP

O

OHOH

OO

O

OP

O

OOH

NH2

O

O

O

O

OP

O

OOH

OH

O

NH2

O

O

O

O

OP

O

OOH

HO

OH

OH

OH

OH

C16:0 FFA

alkyl PA

alkyl PE

lyso PAF

LPC

PC

LPA(oncogenic-LPA receptors)

PA

PS

PI

AC (oxidation)

C1P(oncogenic-receptor unknown)

O

O

O

O

OH

DAG(oncogenic--protein kinase C)

OO

O

OP

O

OOH

NH2

PE

O

OO

O

OP

O

OOH

N

alkyl PC

CDP-DAG

2-acyl-1-alkyl-sn-glycerol

Fig. 4.Map of lipidmetabolism in aggressive cancer cells. The data gathered from isotopic tracing of d4-C16:0 FFA labeled cancer cells was compiled into ametabolic pathwaymap using theKEGG pathway database as a guide. d4-C16:0 FFA incorporation into the lipid structures is noted in red. The color of arrows andmetabolites notes increased (in red), decreased (in blue), orunchanged (in gray) levels of d4-lipid in three out of five comparisons of aggressive (231MFP, M4, SKOV3, PC3, and C8161) versus non-aggressive (MCF7, MCF10A, OVCAR3, LNCaP, andMUM2C, respectively) cancer cells. Metabolites and arrows in black were not detected in either targeted or untargeted analysis. LPA, C1P, DAG, and PAF are oncogenic signaling lipidsthat act through LPA receptors, unknown receptor, protein kinase C, and PAF receptors, respectively.

6 S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

adipocytes can also fuel tumor vascularization and cancer cell pro-liferation [17].

While we show here that cancer cells take up exogenous free non-esterified palmitic acid, we do not yet understand the mechanism forpalmitic acid uptake. Previous studies have shown that breast cancerand sarcoma cells expressing lipoprotein lipase and CD36, involved inlipoprotein-associated triglyceride lipolysis and fatty acid transport, re-spectively, treated with triglyceride-rich lipoproteins led to acceleratedcell proliferation [18]. These authors also found that providing lipopro-tein lipase to prostate cancer cells with triglyceride-rich lipoproteinsprevented the cytotoxic effects of fatty acid synthesis inhibition. The ex-pression of fatty acid binding proteins that are involved in fatty acid up-take and transport have also been associated with poor survival intriple-negative breast cancers [19]. Intriguingly, Kamphorst et al. re-cently demonstrated that under hypoxic conditions or Ras activation,cells switch from de novo lipogenic pathways to scavenging of serumfatty acids esterified to lysophospholipids to fuel membrane production[20]. Interestingly, this study also shows that this phenomenon is alsolinked to reduced glycolytic flux to acetyl-CoA and an increased flux ofglutamine to fatty acid synthesis. While these authors were also unableto ascertain themechanismof lysophopholipid import, they showyet an-othermechanism throughwhich cancer cells take up fatty acid sources. Itwill also be of future interest to determine the interplay between glyco-lytic and glutamine metabolism and fatty acid uptake and metabolismduring cancer progression.

Our results provide a potential alternate andmore directmechanismlinking obesity to increased incidence of cancer deaths bydirectly takingin exogenous fatty acids into structural and signaling lipids that candrive cancer pathogenicity. This mechanism adds to previous studieslinking obesity-induced inflammation, hyperinsulinemia and increasedinsulin growth factor signaling, and heightened adipokine signaling tocancer cell proliferation and malignancy [21].

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.2

In summary, we have used advanced metabolomic platforms toglobally map exogenous fatty acid incorporation and metabolism intocancer cells in situ and in vivo. We find a commonly dysregulated met-abolic signature of lipid metabolism that underlies aggressive humancancer cells where there is an overall increase in exogenous fatty acidincorporation that is redirected from oxidative pathways to the genera-tion of structural and signaling glycerophospholipids, sphingolipids, andether lipids. Targeting fatty acid uptake into cancer cells, in combinationwith inhibitors of key nodal lipid metabolism pathways may provide apotential alternate strategy for treating cancer.

Acknowledgments

We thank the members of the Nomura Research Group and the Luolaboratory for critical reading of themanuscript. Thisworkwas supportedby grants from the National Institutes of Health (R21CA170317,R01CA172667, R00DA030908 for SML, LSR, MMM, and DKN andR01CA101891 for KL) and the Searle Scholar Foundation (DKN).

Appendix A. Supplementary data

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.bbalip.2013.07.008.

References

[1] J.A. Menendez, R. Lupu, Fatty acid synthase and the lipogenic phenotype in cancerpathogenesis, Nat. Rev. Cancer 7 (10) (2007) 763–777.

[2] M. Ookhtens, R. Kannan, I. Lyon, N. Baker, Liver and adipose tissue contributions tonewly formed fatty acids in an ascites tumor, Am. J. Physiol. 247 (1 Pt 2) (1984)R146–R153.

[3] R.J. DeBerardinis, J.J. Lum, G. Hatzivassiliou, C.B. Thompson, The biology of cancer: met-abolic reprogramming fuels cell growth and proliferation, Cell Metab. 7 (1) (2008)11–20.

emodel exogenous palmitate into structural and oncogenic signaling013.07.008

7S.M. Louie et al. / Biochimica et Biophysica Acta xxx (2013) xxx–xxx

[4] D.K. Nomura, J.Z. Long, S. Niessen, H.S. Hoover, S.W. Ng, B.F. Cravatt,Monoacylglycerol lipase regulates a fatty acid network that promotes cancerpathogenesis, Cell 140 (1) (2010) 49–61.

[5] M. Cordenonsi, F. Zanconato, L. Azzolin, M. Forcato, A. Rosato, C. Frasson, M. Inui, M.Montagner, A.R. Parenti, A. Poletti, M.G. Daidone, S. Dupont, G. Basso, S. Bicciato, S.Piccolo, The Hippo transducer TAZ confers cancer stem cell-related traits on breastcancer cells, Cell 147 (4) (2011) 759–772.

[6] D.K. Nomura, D.P. Lombardi, J.W. Chang, S. Niessen, A.M. Ward, J.Z. Long, H.H. Hoover,B.F. Cravatt, Monoacylglycerol lipase exerts dual control over endocannabinoid andfatty acid pathways to support prostate cancer, Chem. Biol. 18 (7) (2011) 846–856.

[7] (a) D.K. Nomura, B.E.Morrison, J.L. Blankman, J.Z. Long, S.G. Kinsey,M.C.Marcondes,A.M.Ward, Y.K. Hahn, A.H. Lichtman, B. Conti, B.F. Cravatt, Endocannabinoid hy-drolysis generates brain prostaglandins that promote neuroinflammation, Sci-ence 334 (6057) (2011) 809–813;

(b) J.R. Piro, D.I. Benjamin, J.M. Duerr, Y. Pi, C. Gonzales, K.M. Wood, J.W. Schwartz,D.K. Nomura, T.A. Samad, A dysregulated endocannabinoid-eicosanoid networksupports pathogenesis in a mouse model of Alzheimer's disease, Cell Rep. 1 (6)(2012) 617–623.

[8] R. Tautenhahn, G.J. Patti, D. Rinehart, G. Siuzdak, XCMSOnline: aweb-basedplatformto process untargeted metabolomic data, Anal. Chem. 84 (11) (2012) 5035–5039.

[9] C.A. Smith, G. O'Maille, E.J. Want, C. Qin, S.A. Trauger, T.R. Brandon, D.E. Custodio, R.Abagyan, G. Siuzdak, METLIN: a metabolite mass spectral database, Ther. Drug Monit.27 (6) (2005) 747–751.

[10] (a) D.K. Nomura, M.M. Dix, B.F. Cravatt, Activity-based protein profiling for bio-chemical pathway discovery in cancer, Nat. Rev. Cancer 10 (9) (2010) 630–638;

(b) D.Medina-Cleghorn, D.K. Nomura, Chemical approaches to studymetabolic net-works, Pflugers Arch. 465 (3) (2013) 427–440.

[11] (a) G.B. Mills, W.H. Moolenaar, The emerging role of lysophosphatidic acid in can-cer, Nat. Rev. Cancer 3 (8) (2003) 582–591;

(b) M.P. Wymann, R. Schneiter, Lipid signalling in disease, Nat. Rev. Mol. Cell Biol. 9(2) (2008) 162–176;

(c) L. Arana, P. Gangoiti, A. Ouro, M. Trueba, A. Gomez-Munoz, Ceramide and cer-amide 1-phosphate in health and disease, Lipids Health Dis. 9 (2010) 15;

(d) A.B. Tsoupras, C. Iatrou, C. Frangia, C.A. Demopoulos, The implication of plateletactivating factor in cancer growth andmetastasis: potent beneficial role of PAF-inhibitors and antioxidants, Infect. Disord. Drug Targets 9 (4) (2009) 390–399.

[12] H. Ogata, S. Goto, W. Fujibuchi, M. Kanehisa, Computation with the KEGG pathwaydatabase, Biosystems 47 (1–2) (1998) 119–128.

[13] (a) A. Pacilli, M. Calienni, S. Margarucci, M. D'Apolito, O. Petillo, L. Rocchi, G.Pasquinelli, R. Nicolai, A. Koverech, M. Calvani, G. Peluso, L. Montanaro,Carnitine-acyltransferase system inhibition, cancer cell death, and preventionofmyc-induced lymphomagenesis, J. Natl. Cancer Inst 105 (7) (2013) 489–498;

Please cite this article as: S.M. Louie, et al., Cancer cells incorporate and rlipids, Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.20

(b) K. Zaugg, Y. Yao, P.T. Reilly, K. Kannan, R. Kiarash, J.Mason, P. Huang, S.K. Sawyer,B. Fuerth, B. Faubert, T. Kalliomaki, A. Elia, X. Luo, V. Nadeem, D. Bungard, S.Yalavarthi, J.D. Growney, A. Wakeham, Y. Moolani, J. Silvester, A.Y. Ten, W.Bakker, K. Tsuchihara, S.L. Berger, R.P. Hill, R.G. Jones, M. Tsao, M.O. Robinson,C.B. Thompson, G. Pan, T.W. Mak, Carnitine palmitoyltransferase 1C promotescell survival and tumor growth under conditions of metabolic stress, GenesDev. 25 (10) (2011) 1041–1051;

(c) L. Liu, Y.D.Wang, J.Wu, J. Cui, T. Chen, Carnitine palmitoyltransferase 1A (CPT1A): atranscriptional target of PAX3-FKHR andmediates PAX3-FKHR-dependent motili-ty in alveolar rhabdomyosarcoma cells, BMC Cancer 12 (2012) 154.

[14] (a) A. Fallani, L. Calorini, A. Mannini, S. Gabellieri, G. Mugnai, S. Ruggieri, Platelet-activating factor (PAF) is the effector of IFN gamma-stimulated invasivenessand motility in a B16 melanoma line, Prostaglandins Other Lipid Mediat. 81(3–4) (2006) 171–177;

(b) V.O. Melnikova, A.A. Mourad-Zeidan, D.C. Lev, M. Bar-Eli, Platelet-activatingfactor mediates MMP-2 expression and activation via phosphorylation ofcAMP-response element-binding protein and contributes to melanomame-tastasis, J. Biol. Chem. 281 (5) (2006) 2911–2922.

[15] F. Snyder, R. Wood, Alkyl and alk-1-enyl ethers of glycerol in lipids from normal andneoplastic human tissues, Cancer Res. 29 (1) (1969) 251–257.

[16] K.M. Nieman, H.A. Kenny, C.V. Penicka, A. Ladanyi, R. Buell-Gutbrod, M.R. Zillhardt,I.L. Romero, M.S. Carey, G.B. Mills, G.S. Hotamisligil, S.D. Yamada, M.E. Peter, K.Gwin, E. Lengyel, Adipocytes promote ovarian cancermetastasis and provide energyfor rapid tumor growth, Nat. Med. 17 (11) (2011) 1498–1503.

[17] Y. Zhang, A.C. Daquinag, F. Amaya-Manzanares, O. Sirin, C. Tseng, M.G. Kolonin, Stro-mal progenitor cells from endogenous adipose tissue contribute to pericytes and ad-ipocytes that populate the tumor microenvironment, Cancer Res. 72 (20) (2012)5198–5208.

[18] N.B. Kuemmerle, E. Rysman, P.S. Lombardo, A.J. Flanagan, B.C. Lipe, W.A. Wells, J.R.Pettus, H.M. Froehlich, V.A. Memoli, P.M. Morganelli, J.V. Swinnen, L.A. Timmerman, L.Chaychi, C.J. Fricano, B.L. Eisenberg, W.B. Coleman, W.B. Kinlaw, Lipoprotein lipaselinks dietary fat to solid tumor cell proliferation, Mol. Cancer Ther. 10 (3) (2011)427–436.

[19] R.Z. Liu, K. Graham, D.D. Glubrecht, D.R. Germain, J.R. Mackey, R. Godbout, Associa-tion of FABP5 expression with poor survival in triple-negative breast cancer: impli-cation for retinoic acid therapy, Am. J. Pathol. 178 (3) (2011) 997–1008.

[20] J.J. Kamphorst, J.R. Cross, J. Fan, E. de Stanchina, R. Mathew, E.P. White, C.B.Thompson, J.D. Rabinowitz, Hypoxic and Ras-transformed cells support growth byscavenging unsaturated fatty acids from lysophospholipids, Proc. Natl. Acad. Sci. U.S. A. 110 (22) (2013) 8882–8887.

[21] S.M. Louie, L.S. Roberts, D.K. Nomura, Mechanisms linking obesity and cancer,Biochim. Biophys. Acta (2013), http://dx.doi.org/10.1016/j.bbalip.2013.02.008.

emodel exogenous palmitate into structural and oncogenic signaling13.07.008