Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling...

8
Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities to metabolic inhibitors Anneleen Daemen a , David Peterson b,1 , Nisebita Sahu b,1 , Ron McCord b,1 , Xiangnan Du c , Bonnie Liu c , Katarzyna Kowanetz b , Rebecca Hong c , John Moffat d , Min Gao c , Aaron Boudreau b , Rana Mroue b , Laura Corson c , Thomas OBrien c , Jing Qing c , Deepak Sampath c , Mark Merchant c , Robert Yauch b , Gerard Manning a , Jeffrey Settleman b , Georgia Hatzivassiliou c , and Marie Evangelista b,2 a Bioinformatics and Computational Biology, Genentech, South San Francisco, CA 94080; b Discovery Oncology, Genentech, South San Francisco, CA 94080; c Translational Oncology, Genentech, South San Francisco, CA 94080; and d Biochemical Pharmacology, Genentech, South San Francisco, CA 94080 Edited by Ronald A. DePinho, University of Texas M.D. Anderson Cancer Center, Houston, TX, and approved June 22, 2015 (received for review January 29, 2015) Although targeting cancer metabolism is a promising therapeutic strategy, clinical success will depend on an accurate diagnostic identification of tumor subtypes with specific metabolic require- ments. Through broad metabolite profiling, we successfully identified three highly distinct metabolic subtypes in pancreatic ductal adeno- carcinoma (PDAC). One subtype was defined by reduced proliferative capacity, whereas the other two subtypes (glycolytic and lipogenic) showed distinct metabolite levels associated with glycolysis, lipo- genesis, and redox pathways, confirmed at the transcriptional level. The glycolytic and lipogenic subtypes showed striking differences in glucose and glutamine utilization, as well as mitochondrial function, and corresponded to differences in cell sensitivity to inhibitors of glycolysis, glutamine metabolism, lipid synthesis, and redox balance. In PDAC clinical samples, the lipogenic subtype associated with the epithelial (classical) subtype, whereas the glycolytic subtype strongly associated with the mesenchymal (QM-PDA) subtype, suggesting functional relevance in disease progression. Pharmaco- genomic screening of an additional 200 non-PDAC cell lines vali- dated the association between mesenchymal status and metabolic drug response in other tumor indications. Our findings highlight the utility of broad metabolite profiling to predict sensitivity of tumors to a variety of metabolic inhibitors. metabolite profiling | metabolic subtypes in PDAC | glycolysis | lipid synthesis | biomarkers for metabolic inhibitors M etabolic reprogramming during tumorigenesis is an essential process in nearly all cancer cells. Tumors share a common phenotype of uncontrolled cell proliferation and must efficiently generate the energy and macromolecules required for cellular growth. The first example of metabolic reprogramming was dis- covered more than 80 y ago by Otto Warburg: tumor cells can shift from oxidative to fermentative metabolism in the course of onco- genesis (1). More recently, there has been a resurgence of interest in targeting cancer metabolism (24) because it may not only be ef- fective in inhibiting tumor growth, but may also provide a therapeutic window (5, 6). For example, inactivation of lactate dehydrogenase-A (LDHA), an enzyme that catalyzes the final step of aerobic gly- colysis, thereby reducing pyruvate to lactate, decreases tumorigenesis and induces regression of established tumors in mouse models of lung cancer driven by oncogenic KRAS or epidermal growth factor receptor (EGFR) while minimally affecting normal cell function (7). The finding that cancers have altered metabolism has prompted substantial investigation, both preclinically and in clinical trials, of several metabolically targeted agents, including those that elevate reactive oxygen species (ROS) or block glycolysis, lipid synthesis, mitochondrial function, and glutamine synthesis pathways (8). The identification of distinct metabolic reprogramming events or metabolic subtypes in cancer may inform patient selection for investigational metabolic inhibitors and in the selection of new therapeutic targets (9, 10). Just as tumors vary greatly in genomic alterations that impact signaling and regulatory pathways, meta- bolic transformation is also heterogeneous and dependent on tissue type, proliferation rate, and isoenzyme use (9, 11). In addition, the observed differences in the dependence on and utilization of the major nutrientsglutamine and glucoseare linked to oncogenic signaling and the genomic features of a cancer cell (12). Large-scale pharmacogenomic screening is a powerful method for identifying biomarkers of drug response and can accelerate the search for new cancer therapies (13, 14). In this study, we used broad baseline metabolite profiling in cell line models of pancreatic ductal adenocarcinoma (PDAC), a disease context previously associated with altered metabolism (1518), to iden- tify metabolic subtypes within PDAC and predict their sensitivity to various metabolic inhibitors. Results Baseline Metabolite Profiling Identifies Three Metabolic Subtypes in PDAC. We examined cell lines derived from naturally occurring tumors because they recapitulate many aspects of the tissue type Significance Targeting cancer metabolism requires personalized diagnostics for clinical success. Pancreatic ductal adenocarcinoma (PDAC) is characterized by metabolism addiction. To identify metabolic de- pendencies within PDAC, we conducted broad metabolite profiling and identified three subtypes that showed distinct metabolite profiles associated with glycolysis, lipogenesis, and redox path- ways. Importantly, these profiles significantly correlated with enriched sensitivity to a variety of metabolic inhibitors including inhibitors targeting glycolysis, glutaminolysis, lipogenesis, and re- dox balance. In primary PDAC tumor samples, the lipid subtype was strongly associated with an epithelial phenotype, whereas the glycolytic subtype was strongly associated with a mesenchymal phenotype, suggesting functional relevance in disease progression. Our findings will provide valuable predictive utility for a number of metabolic inhibitors currently undergoing phase I testing. Author contributions: A.D., G.H., and M.E. designed research; A.D., D.P., K.K., R.Y., G.M., J.S., G.H., and M.E. performed research; A.D., D.P., N.S., R. McCord, X.D., B.L., K.K., R.H., J.M., M.G., A.B., R. Mroue, L.C., T.O., J.Q., and M.E. contributed new reagents/analytic tools; A.D., D.P., N.S., R. McCord, X.D., B.L., K.K., R.H., J.M., M.G., A.B., R. Mroue, L.C., T.O., J.Q., D.S., M.M., G.H., and M.E. analyzed data; and A.D. and M.E. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. Freely available online through the PNAS open access option. 1 D.P., R. McCord, and N.S. contributed equally to this work. 2 To whom correspondence should be addressed. Email: [email protected]. This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. 1073/pnas.1501605112/-/DCSupplemental. www.pnas.org/cgi/doi/10.1073/pnas.1501605112 PNAS Early Edition | 1 of 8 CELL BIOLOGY PNAS PLUS Downloaded by guest on February 7, 2021

Transcript of Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling...

Page 1: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

Metabolite profiling stratifies pancreatic ductaladenocarcinomas into subtypes with distinctsensitivities to metabolic inhibitorsAnneleen Daemena, David Petersonb,1, Nisebita Sahub,1, Ron McCordb,1, Xiangnan Duc, Bonnie Liuc,Katarzyna Kowanetzb, Rebecca Hongc, John Moffatd, Min Gaoc, Aaron Boudreaub, Rana Mroueb, Laura Corsonc,Thomas O’Brienc, Jing Qingc, Deepak Sampathc, Mark Merchantc, Robert Yauchb, Gerard Manninga, Jeffrey Settlemanb,Georgia Hatzivassiliouc, and Marie Evangelistab,2

aBioinformatics and Computational Biology, Genentech, South San Francisco, CA 94080; bDiscovery Oncology, Genentech, South San Francisco, CA 94080;cTranslational Oncology, Genentech, South San Francisco, CA 94080; and dBiochemical Pharmacology, Genentech, South San Francisco, CA 94080

Edited by Ronald A. DePinho, University of Texas M.D. Anderson Cancer Center, Houston, TX, and approved June 22, 2015 (received for review January 29, 2015)

Although targeting cancer metabolism is a promising therapeuticstrategy, clinical success will depend on an accurate diagnosticidentification of tumor subtypes with specific metabolic require-ments. Through broad metabolite profiling, we successfully identifiedthree highly distinct metabolic subtypes in pancreatic ductal adeno-carcinoma (PDAC). One subtype was defined by reduced proliferativecapacity, whereas the other two subtypes (glycolytic and lipogenic)showed distinct metabolite levels associated with glycolysis, lipo-genesis, and redox pathways, confirmed at the transcriptional level.The glycolytic and lipogenic subtypes showed striking differences inglucose and glutamine utilization, as well as mitochondrial function,and corresponded to differences in cell sensitivity to inhibitors ofglycolysis, glutamine metabolism, lipid synthesis, and redox balance.In PDAC clinical samples, the lipogenic subtype associated with theepithelial (classical) subtype, whereas the glycolytic subtypestrongly associated with the mesenchymal (QM-PDA) subtype,suggesting functional relevance in disease progression. Pharmaco-genomic screening of an additional ∼200 non-PDAC cell lines vali-dated the association between mesenchymal status and metabolicdrug response in other tumor indications. Our findings highlight theutility of broad metabolite profiling to predict sensitivity of tumorsto a variety of metabolic inhibitors.

metabolite profiling | metabolic subtypes in PDAC | glycolysis |lipid synthesis | biomarkers for metabolic inhibitors

Metabolic reprogramming during tumorigenesis is an essentialprocess in nearly all cancer cells. Tumors share a common

phenotype of uncontrolled cell proliferation and must efficientlygenerate the energy and macromolecules required for cellulargrowth. The first example of metabolic reprogramming was dis-covered more than 80 y ago by Otto Warburg: tumor cells can shiftfrom oxidative to fermentative metabolism in the course of onco-genesis (1). More recently, there has been a resurgence of interestin targeting cancer metabolism (2–4) because it may not only be ef-fective in inhibiting tumor growth, but may also provide a therapeuticwindow (5, 6). For example, inactivation of lactate dehydrogenase-A(LDHA), an enzyme that catalyzes the final step of aerobic gly-colysis, thereby reducing pyruvate to lactate, decreases tumorigenesisand induces regression of established tumors in mouse models oflung cancer driven by oncogenic KRAS or epidermal growth factorreceptor (EGFR) while minimally affecting normal cell function(7). The finding that cancers have altered metabolism has promptedsubstantial investigation, both preclinically and in clinical trials, ofseveral metabolically targeted agents, including those that elevatereactive oxygen species (ROS) or block glycolysis, lipid synthesis,mitochondrial function, and glutamine synthesis pathways (8).The identification of distinct metabolic reprogramming events

or metabolic subtypes in cancer may inform patient selection forinvestigational metabolic inhibitors and in the selection of new

therapeutic targets (9, 10). Just as tumors vary greatly in genomicalterations that impact signaling and regulatory pathways, meta-bolic transformation is also heterogeneous and dependent on tissuetype, proliferation rate, and isoenzyme use (9, 11). In addition, theobserved differences in the dependence on and utilization of themajor nutrients—glutamine and glucose—are linked to oncogenicsignaling and the genomic features of a cancer cell (12).Large-scale pharmacogenomic screening is a powerful method

for identifying biomarkers of drug response and can acceleratethe search for new cancer therapies (13, 14). In this study, weused broad baseline metabolite profiling in cell line models ofpancreatic ductal adenocarcinoma (PDAC), a disease contextpreviously associated with altered metabolism (15–18), to iden-tify metabolic subtypes within PDAC and predict their sensitivityto various metabolic inhibitors.

ResultsBaseline Metabolite Profiling Identifies Three Metabolic Subtypes inPDAC. We examined cell lines derived from naturally occurringtumors because they recapitulate many aspects of the tissue type

Significance

Targeting cancer metabolism requires personalized diagnosticsfor clinical success. Pancreatic ductal adenocarcinoma (PDAC) ischaracterized by metabolism addiction. To identify metabolic de-pendencies within PDAC, we conducted broad metabolite profilingand identified three subtypes that showed distinct metaboliteprofiles associated with glycolysis, lipogenesis, and redox path-ways. Importantly, these profiles significantly correlated withenriched sensitivity to a variety of metabolic inhibitors includinginhibitors targeting glycolysis, glutaminolysis, lipogenesis, and re-dox balance. In primary PDAC tumor samples, the lipid subtypewas strongly associatedwith an epithelial phenotype, whereas theglycolytic subtype was strongly associated with a mesenchymalphenotype, suggesting functional relevance in disease progression.Our findingswill provide valuable predictive utility for a number ofmetabolic inhibitors currently undergoing phase I testing.

Author contributions: A.D., G.H., and M.E. designed research; A.D., D.P., K.K., R.Y., G.M., J.S.,G.H., and M.E. performed research; A.D., D.P., N.S., R. McCord, X.D., B.L., K.K., R.H., J.M.,M.G., A.B., R. Mroue, L.C., T.O., J.Q., and M.E. contributed new reagents/analytic tools;A.D., D.P., N.S., R. McCord, X.D., B.L., K.K., R.H., J.M., M.G., A.B., R. Mroue, L.C., T.O., J.Q.,D.S., M.M., G.H., and M.E. analyzed data; and A.D. and M.E. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.

Freely available online through the PNAS open access option.1D.P., R. McCord, and N.S. contributed equally to this work.2To whom correspondence should be addressed. Email: [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1501605112/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1501605112 PNAS Early Edition | 1 of 8

CELL

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 2: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

and genomic context of cancer (13, 14, 19, 20). Levels of 256metabolites were quantified in 38 pancreatic cancer cell lines (fivebiological replicates per cell line) in logarithmic growth phase usingmedia with physiological glucose and glutamine concentrations(Datasets S1–S3). We applied nonnegative matrix factorization(NMF) (21), a recently established approach for consensus clus-tering (22–24), to 153 metabolites with reproducible variation,allowing the capture of the strongest signal of metabolic de-pendency (SI Materials and Methods). This analysis revealed threestable and reproducible subtypes with adequate data coherence(Fig. S1 A and B). The metabolite profiles of the cell lines orderedby subtype are shown in Fig. 1A for metabolites with distinct in-tensities in at least one subtype compared with the other twosubtypes (F test, P < 0.05). These three subtypes provided a usefuland interpretable basis for further analysis.

Metabolic Characterization Reveals a Slow Proliferating, Glycolytic,and Lipogenic Subtype. The metabolite intensities within eachsubtype were then mapped to known, previously establishedmetabolic ontologies (Dataset S1 and SI Materials and Methods)(25). One subtype (34% of all lines) was especially low in aminoacids and carbohydrates (Fig. 1A, left subtype, and Fig. S1C). Celllines in this subtype had an average doubling time that was sig-nificantly higher (Fig. S1D) and were named the slow proliferatingsubtype. Doubling times for cell lines from the other two subtypeswere more similar (Fig. S1D); however, these two subtypes dis-played strikingly distinct metabolic profiles, independent of pro-liferation rate (SI Materials and Methods). Thus, these metabolicsubtypes have unique metabolic profiles that are independent ofgrowth rate.We further explored the metabolic differences between the two

subtypes with similar proliferation rates. One subtype (27% of alllines; Fig. 1A) exhibited, on average, elevated levels of variouscomponents of the glycolytic and serine pathways, mainly phos-phoenolpyruvate (PEP), glyceraldehyde-3-phosphate, lactate, andserine (Fig. 1 B and C and Fig. S1E), and was named the glycolyticsubtype. This subtype was also distinguished by lower levels ofmetabolites important for redox potential such as nicotinamideadenine dinucleotide (NAD) reduced (NADH), NAD phos-phate (NADP), NAD phosphate reduced (NADPH), glutathionedisulfide (GSSG), glutathione (GSH), and flavine adenine di-nucleotide (Fig. 1 B and C, Fig. S1F, and Dataset S4). In contrast,the other subtype (39% of all lines; Fig. 1A) was enriched forvarious lipid metabolites such as palmitic acid (C16:0), oleic acid(C18:cis[9]1), palmitoleic acid (C16:cis[9]1), and myristic acid(C14:0) (Fig. 1 B and D and Dataset S4), as well as mitochondrial[oxidative phosphorylation (OXPHOS)] metabolites important forthe electron transport chain such as coenzyme Q10 and coenzymeQ9 and components of the aspartate-malate shuttle such as as-partate and glutamate (Fig. S1G and Dataset S4), and was namedthe lipogenic subtype.

Differences Between Glycolytic and Lipogenic Subtypes Are ConfirmedTranscriptionally. We next determined whether differences in me-tabolite levels observed between the glycolytic and lipogenic subtypescould be explained by differences in the expression of genesknown to be associated with the metabolic ontologies (Dataset S5and SI Materials and Methods). Consistent with the differences inmetabolite levels, expression of genes associated with glycolysis andthe pentose phosphate pathway were found to be relatively ele-vated in cell lines from the glycolytic subtype (Fig. 1 E and F, Fig.S1 H and I, and Dataset S6). For example, most glycolytic linesdemonstrated higher expression of neuron-specific enolase [ENO2;adjusted P = 0.0016; Fig. 1 E and F], along with higher levels of itsproduct PEP, whereas other enolase homologs were not differen-tially expressed (Fig. S1J). We also noted that protein (and notmRNA) abundance of the lactate transporter, monocarboxylatetransporter 1 (MCT1) was elevated in the glycolytic lines compared

with the lipogenic lines (P < 0.05; Fig. 1E and Fig. S1K). In con-trast, cell lines within the lipogenic subtype were enriched for ex-pression of lipogenesis genes involved in cholesterol and de novolipid synthesis including 7-dehydrocholesterol reductase (DHCR7),stearoyl-CoA desaturase (SCD), and fatty acid synthase (FASN)(adjusted P < 0.1; Fig. 1 E and F, Fig. S1 H and L, and DatasetS6). Thus, PDAC-derived cell lines can be clustered by theirmetabolite profiles and these differences appear to be determinedin part by differences in gene expression.

Glycolytic and Lipogenic Subtypes Use Glucose and Glutamine in aDifferent Manner. The metabolic and transcriptional profiles sug-gested that these two subtypes may differ in their use of glucoseand glutamine, the most abundant carbon sources available tocancer cells. We predicted that the lipogenic subtype wouldpreferentially use glucose for the tricarboxylic acid (TCA) cycleand lipid synthesis, whereas the glycolytic subtype would use glu-cose more for aerobic glycolysis, and consequently, use more glu-tamine for TCA anaplerosis. 13C metabolic mass isotopomerdistribution analysis (MIDA) using either [U-13C5]glutamine or[U-13C6]glucose revealed a significant increase in the contribution of[U-13C6]glucose to TCA metabolites in representative cell linesfrom the lipogenic subtype relative to the glycolytic subtype (Fig.2A; P < 0.05). In contrast, representative glycolytic lines in-corporated [U-13C5]glutamine into TCA metabolites at signifi-cantly higher levels than lines from the lipogenic subtype (Fig. 2B;P< 0.05). Moreover, lipogenic cell lines incorporated 14C-glucose intolipid metabolites at a significantly higher level than cell lines fromthe glycolytic subtype (Fig. 2C; P < 0.01). Consistent with theseobservations, lipogenic lines showed on average higher O2 con-sumption (Fig. 2D; P < 0.01) and a greater mitochondrial content[Mitotracker and tetramethylrhodamine ethyl ester (TMRE) in-tensity] compared with glycolytic subtype lines (Fig. 2E; P < 0.01;Dataset S7). Thus, cell lines from the glycolytic and lipogenicsubtypes appear to use glucose and glutamine in a different manner.

Glycolytic and Lipogenic Cell Lines Show Distinct Sensitivity toVarious Metabolic Inhibitors in Vitro. Based on their distinct meta-bolic wiring, we predicted that glycolytic and lipogenic cell lineswould show differential sensitivity to inhibitors targeting aerobicglycolysis (oxamate and the LDHA inhibitor GNE-140) (26),glutaminolysis [bis-2-(5-phenylacetimido-1,2,4,thiadiazol-2-yl)ethylsulfide (BPTES)], and de novo lipid synthesis [FASN inhibitorGSK1195010 (27), SCD inhibitor (28), cerulenin, and orlistat].Indeed, as predicted, the glycolytic subtype was enriched for linesthat were sensitive to the LDHA inhibitor, oxamate, and BPTES,whereas the lipogenic subtype was enriched for lines that weresensitive to inhibitors targeting lipid synthesis (Fig. 3A and Fig.S2A; P < 0.05; Dataset S7 and SI Materials and Methods). More-over, glycolytic cell lines showed higher rates of fatty acid (FA)uptake (Fig. S2B) and increased sensitivity to media with reducedlipid content (Fig. S2C), suggesting these lines may be more re-liant on FA pathways for generating lipids.Maintaining redox balance is another key requirement for can-

cer cells (29). The differences in redox-related metabolites betweenglycolytic and lipogenic cell lines suggested that they may also showdifferential sensitivity to ROS-inducing agents or inhibitors of en-zymes or transporters important for maintaining glutathione syn-thesis and NADP/NADPH balance in cells. Indeed, we found thatcell lines within the glycolytic subtype showed enhanced sensitivityto a variety of such agents including inhibitors of gamma-gluta-mylcysteine synthetase [buthionine sulphoximine (BSO)],and the cystine transporter xCT {(S)-4-carboxyphenylglycine[(S)-4-CPG]} (Fig. 3B and Dataset S7).In addition to short-term (3 d) culture assays, we tested the

efficacy profile of LDHA inhibitor, oxamate, and the SCD in-hibitor in long-term (12 d) culture assays and observed similarresults (Fig. 3 A and C).

2 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1501605112 Daemen et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 3: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

Glucose

G6P

F6P

F1,6BP

Glyceraldehyde 3-P

3-Phosoglycerate

PEP

Ribulose-5-P G1P

2,6BP

Glycerol-3-P

Serine

Pyruvate Lactate

Glutamine

Glutamate

a-KG

Citrate Isocitrate Citrate

a-KG

Acetyl-CoA LipidsPDH

Oxaloacetate Aspartate

ENO2

MCT1

HMGCS1 MVD

FDFT1 DHCR7 SCD5DL DHCR24

DGAT1 SCD FASN ACAT2

PSPH

PDK1

p < 0.05 p < 0.1 p < 0.2

Glycolytic Lipogenic

Succinate Malate

Oxaloacetate

Fumarate

Cholines

Glycolysis / PPP

Carbohydrates

Other complex lipids

OXPHOS /mitochondria−related REDOX

Fatty acids

Amino acids

Sphingolipids

JG score

−5 0 5 10

Glycolyt

ic

Lipogenic

-1.0

-0.5

0.0

0.5

1.0

1.5

log2

Ole

ic a

cid

(RU

) ***

Glycolyt

ic

Lipogenic

-2.0

-1.0

0.0

1.0

2.0

log2

Pal

mito

leic

aci

d (R

U)

***

Glycolyt

ic

Lipogenic

-2.0

-1.0

0.0

1.0

2.0

3.0

log2

Myr

istic

aci

d (R

U)

***

Glycolyt

ic

Lipogenic

-2.0

-1.0

0.0

1.0

2.0

log2

Pal

mito

leic

aci

d (R

U)

***

Glycolyt

ic

Lipogenic

2

4

6

8

10

log2

SC

D (R

PKM

+1) *

Glycolyt

ic

Lipogenic

4

5

6

7

8

log2

FA

SN (R

PKM

+1)

*

Glycolyt

ic

Lipogenic

1

2

3

4

5

6

log2

EN

O2

(RPK

M +

1) ***

Glycolyt

ic

Lipogenic

4

5

6

7

8

log2

DH

CR

7 (R

PKM

+1)

**

Glycolyt

ic

Lipogenic

6.0

8.0

10.0

12.0

14.0

log2

PEP

(RU

) **

Glycolyt

ic

Lipogenic

0.0

2.0

4.0

6.0

log2

GSS

G (R

U) **

Glycolyt

ic

Lipogenic

-6.0

-4.0

-2.0

0.0

2.0

4.0

log2

GA

P (R

U) *

D

E

B

C

F

B

AsPC1

YAPCCFPAC

1KP

3TCC

PAN2

Hs 766TCapan

2Panc 02.03

Panc 04.03

Panc 03.27

Panc 05.04

KLM1

Panc 10.05

MIA Paca

2

KP2

PSN1HUP

T3PK

45HPL45PA

TU8988T

KP4PK

45PSW

1990PA

TU8902

HPACKCI

MO

H1

PANC1

SUIT2

PATU

8988S

HPAFII

SU.86.86PK

8PK

1DAN

GHUP

T4BxPC

3PK

59KP

3L

Car

bo-

hydr

ates

A

min

o ac

ids

REDOX

Fatty

ac

ids

Gly

coly

sis

PP

P

0

3.9

1.9

+2.

1

+4.

2

Slow Proliferating Glycolytic Lipogenic

A

Fig. 1. Identification of distinct metabolic subtypes in PDAC through baseline metabolite profiling. (A) Hierarchical clustering of identifiable metaboliteswith significant intensity differences between any of the three subtypes (F test, P < 0.05; 99 metabolites). Cell lines were grouped by subtype, with the orderper subtype defined by unsupervised clustering. Log2 intensity ratio data per metabolite are scaled across all cell lines to mean = 0 and SD = 1. Blue indicateslow scaled intensity, and yellow indicates high for each metabolite. Highlighted in gray are functionally related metabolites. Slow proliferating lines arelabeled in gray, glycolytic lines in purple, and lipogenic in cyan. (B) Relative enrichment of the eight metabolic ontology classes in the glycolytic and lipogenicsubtypes, represented by JG score (47). Positive scores represent ontologies enriched for metabolites with high intensities in the glycolytic subtype. SeeDataset S1 for a description and list of metabolites per ontology and Dataset S4 for the list of differentially expressed metabolites. (C) Normalized metaboliteintensity levels for metabolites involved in glycolysis/pentose phosphate and redox pathways that were differentially expressed between glycolytic andlipogenic lines. RU stands for relative unit, with intensity levels normalized to a reference pool of samples for metabolites from the Broad Profiling platform(Dataset S2) and to a universal 13C-labeled internal standard for metabolites from the Energy platform (Dataset S3). (D) Normalized metabolite intensity levelsfor metabolites involved in lipid synthesis that were differentially expressed between glycolytic and lipogenic lines. (E) Detailed metabolite map with genesdifferentially expressed between cell lines from the glycolytic vs. lipogenic subtype indicated with various shades of color depending on P value corrected formultiple testing. For MCT1, P value is based on protein expression level. We refer to Dataset S6 for a list of differentially expressed genes. (F) Expression levelsof ENO2, DHCR7, SCD, and FASN involved in glycolysis and lipid synthesis that were differentially expressed between glycolytic and lipogenic lines (DatasetS6). Asterisks denote a statistically significant difference by unpaired t test with Welch’s correction (*P < 0.05, **P < 0.01, ***P < 0.001).

Daemen et al. PNAS Early Edition | 3 of 8

CELL

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 4: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

Functional Confirmation of the Glycolytic and Lipogenic Subtype inVivo. To translate these findings in vivo and generate proof-of-concept findings for our two metabolic subtypes, we evaluatedxenografts of MIA Paca-2, a glycolytic cell line, and HPAC, alipogenic cell line, for their sensitivity to glycolysis vs. lipid syn-thesis inhibition. Because oxamate and LDHA inhibitors havepoor pharmacokinetics in mice (26), we inhibited glycolysis byengineering MIA Paca-2 and HPAC cells to express a doxycline(DOX)-inducible shRNA against LDHA. MIA Paca-2 xenografttumors treated with DOX showed undetectable levels of LDHA(Fig. 3D) and 68% tumor growth inhibition (TGI) compared withtumors expressing LDHA (Fig. 3E). In contrast, administration of

an SCD inhibitor showed no efficacy (Fig. 3E), although phar-macodynamic inhibition of SCD was seen (Fig. 3F). In contrast,HPAC xenograft tumors showed minimal sensitivity to LDHAknockdown (9% TGI; Fig. S2 D and E) but showed significanttumor growth inhibition to SCD inhibitor treatment (52% TGI)(30). Thus, glycolytic and lipogenic subtypes are functionally dis-tinct and show differential sensitivity to glycolytic and lipidbiosynthesis inhibition.

Glycolytic and Lipogenic Subtypes Are Associated with KnownSubtypes of PDAC, Driven by Mesenchymal Status. We next setout to determine how our defined metabolic subtypes associated

C Glycolyt

ic

Lipogenic

0

5

10

15

% M

alat

e (M

2) fr

om 1

3C-G

luco

se

Glycolyt

ic

Lipogenic

0

10

20

30

40

% C

itrat

e (M

2) fr

om 1

3C-G

luco

se

*

*

Glycolyt

ic

Lipogenic

0

2

4

6

8

10

% A

spar

tate

(M2)

fro

m 1

3C-G

luco

se

Glycolyt

ic

Lipogenic

0

5

10

15

20

25

% a

-KG

(M2)

from

13C

-Glu

cose

*

*

Glycolyt

ic

Lipogenic

0

5000

10000

15000

20000

25000

14C

-Glu

cose

into

Lip

ids

(CPM

/mg

prot

ein) **

Glycolyt

ic

Lipogenic

0.00

0.02

0.04

0.06

0.08

0.10

0.12

OCAR

(pm

oles

/min

/cel

l)

**

Glycoly

tic

Lipog

enic

0

200

400

600

800

1000

1200

TMR

E In

tens

ity/c

ell

Glycoly

tic

Lipog

enic

0

200

400

600

800

1000

1200

Mito

trac

ker I

nten

sity

/cel

l

MIA Paca-2HUP-T3SW 1990 SUIT-2HPACPA-TU-8988S

TMR

E M

itotr

acke

r

Glycolytic Lipogenicy y p g

** **

Glycolyt

ic

Lipogenic

10

20

30

40

% G

luta

mat

e (M

5) fr

om

13C

-Glu

tam

ine

Glycolyt

ic

Lipogenic

0

10

20

30

40

% M

alat

e (M

4) fr

om

13C

-Glu

tam

ine

Glycolyt

ic

Lipogenic

0

10

20

30

40

% a

-KG

(M5)

from

13

C-G

luta

min

e

* **

**

A

C

D

E

B

Fig. 2. Functional characterization of glycolytic and lipogenic subtypes. (A) Comparison of relative contribution of glucose oxidation to the TCA metabolites,determined by M2 labeling from [U-13C6]glucose for citrate, αKG, malate, and aspartate between glycolytic and lipogenic cell lines. (B) Comparison of relativecontribution of reductive glutamine metabolism to TCA metabolites, determined by M5 labeling from [U-13C5]glutamine for αKG and glutamate, and M4labeling for malate between glycolytic and lipogenic cell lines. (C) Comparison of relative contribution of glucose metabolism to de novo lipid synthesisbetween glycolytic and lipogenic cell lines. Cells were labeled with 1 μCi/mL D[U-14C] glucose for 6 h, and lipids were extracted. The incorporation of 14C intolipids was determined by scintillation counting. (D) Comparison of oxygen consumption rates (OCRs) between glycolytic and lipogenic cell lines. (E) Com-parison of relative mitochondria number (Mitotracker intensity per cell) and potential/fitness (TMRE per cell) between glycolytic and lipogenic cell lines. ForA–E, the mean and SD between cell lines belonging to the glycolytic subtype vs. lipogenic subtype is plotted where each cell line is shown as one dot,representing the mean of three replicates. Data are normalized to sample protein content (A–C) or cell number (D and E). Asterisks denote a statisticallysignificant difference by unpaired t test with Welch’s correction (*P < 0.05, **P < 0.01, ***P < 0.001).

4 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1501605112 Daemen et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 5: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

with primary PDAC tumor samples from patients. Three clinicalsubtypes of PDAC were recently identified through molecularprofiling of PDAC tumors: classical (characterized by high ex-pression of adhesion-associated and epithelial genes), quasi-mesen-chymal (QM-PDA, characterized by mesenchyme-associatedgenes), and exocrine-like (22). Because exocrine-like cell lineshave not been reported, we simplified the three-subtype PDACsignature to a 42-gene expression signature that distinguishesclassical from QM-PDA (22), and applied it to our cell line panel.We found that all cell lines within the glycolytic subtype associatedwith the quasimesenchymal subtype, whereas most lipogenic lines

were associated with the classical subtype (Fig. 4A; P = 0.0006;Dataset S7). Thus, our metabolite subtypes derived from pan-creatic cell lines strongly correlate with known subtypes ofPDAC tumors, with the glycolytic subtype strongly associatingwith mesenchymal features and the lipogenic subtype associ-ating with epithelial features.

Metabolic and Mesenchymal Markers Predict Response to Glycolyticand Glutaminolytic Inhibitors in PDAC and Other Tumor Types. Car-cinomas with mesenchymal features (including PDAC) tend to bemore aggressive and typically have an overall poorer prognosis

A C

D

F

B

E

Fig. 3. Glycolytic and lipogenic cell lines show distinct sensitivity to various metabolic inhibitors both in vitro and in vivo. (A) Comparison of IC50 values to variousmetabolic inhibitors between representative glycolytic vs. lipogenic cell lines in short-term (3 d) viability assays. Saturated IC50 values correspond to cell lines wherean IC50 was not reached at the maximum drug concentration. The mean and SD between cell lines belonging to the glycolytic vs. lipogenic subtype is plottedwhere each cell line is shown as one dot, representing the mean of three replicates. Asterisks denote a statistically significant difference by Mann–Whitney test(*P < 0.05, **P < 0.01, ***P < 0.001). (B) Comparison of IC50 values to various ROS-inducing agents between representative glycolytic vs. lipogenic cell lines in short-term (3 d) viability assays, similar to A. (C) Comparison of sensitivity to oxamate, LDHA, or SCD inhibitors between representative glycolytic vs. lipogenic cell lines inlonger-term (12 d), low seeding density growth assays. (D) Western blots showing 98% in vivo knockdown of LDHA levels in MIA Paca-2 xenografts administeredwith doxycycline (1 mg/mL) for 8 d vs. 5% sucrose. (E) In vivo knockdown of LDHA (n = 10 for each group) results in 68% TGI, 95% confidence interval [48, 83] inthe MIA Paca-2 shLDHA model of a glycolytic subtype tumor, whereas treatment with an SCD inhibitor (75 mg/kg, orally, BID) resulted in no significant change intumor volume. (F) Confirmed pharmacodynamic inhibition of lipid metabolism by SCD inhibitor. The SCD inhibitor reduces desaturation of palmitate and stearatein MIA Paca-2 shLDHA xenograft tumor tissues and in mouse liver and plasma (n = 5 per group). Data are presented as mean ± SD.

Daemen et al. PNAS Early Edition | 5 of 8

CELL

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 6: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

EMT High

EMT Low

100

1000

BSO

IC50

(

EMT High

EMT Low0

10

20

30

BPT

ES IC

50 (

M)

EMT High

EMT Low0

20

40

60

Oxa

mat

e IC

50 (m

M) **** **** ****

MIA Paca 2KP4

PSN1PA TU 8988T

PK 45PPK 45H

PL45SW 1990

DAN GHUP T3PANC 1

PA TU 8902BxPC 3

KP 2SU.86.86

PK 8SUIT 2PK 59KP 3L

KCI MOH1HPAC

HUP T4HPAF II

PA TU 8988S

Epithelial / Mesenchymal Score

2 1 0 1 2

GlycolyticLipogenic

Classical (Epithelial)

QMPDA (Mesenchymal)

Epithelial / Mesenchymal Score

Glycolysis / PPP

PPARA

Lipids

Amino acids

OXPHOS

JG score

5 0 5

PPARA

Glycolysis / PPP

Lipids

Amino acids

OXPHOS

JG score

15 10 5 0 5

Up in resistant

Up in resistant

LDHAi BPTES

Up in sensitive

JG score

Up in sensitive

JG score

Glycolytic Subtype Lipogenic Subtype Glucose Glucose

Pyruvate Lactate

Serine

Pentose Pathway

Mitochondria (TCA cycle)

Pyruvate Lipid synthesis

Mitochondria (TCA cycle)

Glutamine

Mesenchymal Tumors Sensitive to:

-Glycolytic, Glutamine inhibitors, ROS-inducing agents

Epithelial Tumors Sensitive to:

- Lipid inhibitors

Vim H

igh

Vim Low

0

10

20

30

BPT

ES IC

50 (

M)

Vim H

igh

Vim Low

0

100

200

300

400

500

Oxa

mat

e IC

50 (m

M)

Vim H

igh

Vim Low

100

1000

BSO

IC50

(nM

)

**** **** **** **** **** ****

ENO2 / Lipid Ave

ENO2

Lipid Ave

MVD

HMGCS1

FDFT1

DHCR7

DGAT1

0−4.

4

−2.

4

+1.

7

+3.

8

A

C

F

B

D

E

Fig. 4. Metabolic and mesenchymal markers predict response to glycolytic and glutaminolytic inhibitors in PDAC and other tumor types. (A) Epithelial/mes-enchymal score for the glycolytic and lipogenic cell lines based on a 42-gene set characteristic of the classical and QM-PDA subtypes (22). The score is defined as thedifference in average expression of QM-PDA vs. classical genes, with a positive score indicative of QM-PDA and a negative score of classical. Cell lines arecolored by metabolic subtype, with glycolytic lines in purple and lipogenic lines in cyan. All glycolytic cell lines are of the QM-PDA subtype, whereas lipogenic celllines are associated with the classical subtype (Fisher’s exact test, P = 0.0006). (B) Relative enrichment of the five curated metabolism gene sets in cell lines that aresensitive (positive JG score) or resistant (negative JG score) to LDHA inhibitor or BPTES in a pan-cancer panel of 204 and 167 nonpancreatic cell lines, respectively,after exclusion of cell lines with intermediate response. See Dataset S5 for a list of genes per gene set. (C) Metabolic dependency preference in the panel of ∼200nonpancreatic cell lines is based on the ratio of ENO2 expression to the average expression of five lipid genes, and labeled on top of the heatmap as glycolytic inpurple (ratio > third quartile), lipogenic in cyan (ratio < lower quartile), and undefined type in gray (ratio between lower and third quartile). Shown are expression(log2 RPKM + 1) of glycolytic gene ENO2, five lipid genes DGAT1, DHCR7, FDFT1, HMGCS1, and MVD, average expression of the five lipid genes (Lipid Ave), andthe ratio of ENO2 to average lipid expression (ENO2/Lipid Ave). Data from Dataset S8. (D) High expression of a pan-cancer EMT signature (EMT) associates withsensitivity to oxamate, BPTES, and BSO across a variety of tumor types. EMT low is defined by RPKM values < lower quartile, EMT high = RPKM values > thirdquartile. Asterisks denote a statistically significant difference by Mann–Whitney test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001). (E) High expression ofmesenchymal marker vimentin (Vim) associates with sensitivity to oxamate, BPTES, and BSO across a variety of tumor types. Vim low is defined by RPKM values <lower quartile, Vim high = RPKM values > third quartile. Asterisks as per D. (F) Model of preferential glucose and glutamine utilization in the glycolytic vs.lipogenic subtype.

6 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1501605112 Daemen et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 7: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

(22, 31, 32). Given the strong association between quasi-mesenchymalstatus and glycolytic dependency in the PDAC lines, we askedwhether this association might also exist in other tumor types.We screened ∼200 nonpancreatic cancer cell lines, representingvarious tumor types, for sensitivity to inhibitors of aerobic gly-colysis and glutaminolysis, as well as to ROS-inducing agents(Dataset S8). As in PDAC (Fig. 1E and Fig. S1H), we found thatcell lines most sensitive to the LDHA inhibitor, oxamate, andBPTES were associated with a glycolytic signature, whereas celllines that were most resistant to these inhibitors were associatedwith an OXPHOS signature (Fig. 4B and Fig. S3A). We nextassigned each cell line to a metabolic subtype (glycolytic vs.lipogenic) using the glycolytic and lipid genes that were mostdifferentially expressed in the PDAC metabolic subtypes (ad-justed P < 0.05). Using the ratio of expression of glycolytic geneENO2 to the average expression of lipid genes diacylglycerolO-acyltransferase 1 (DGAT1), DHCR7, farnesyl-diphosphatefarnesyltransferase 1 (FDFT1), 3-hydroxy-3-methylglutaryl-CoAsynthase 1 (HMGCS1), and mevalonate (diphospho) decarbox-ylase (MVD) clearly distinguished nonpancreatic cell lines bymetabolic dependency preference (Fig. 4C). A glycolytic pref-erence in nonpancreatic lines associated with sensitivity to LDHAinhibitor, oxamate, and BPTES (Fig. S3B; P < 0.05; Dataset S8). Inaddition, consistent with our findings in the PDAC tumors, mes-enchymal tumors [according to a pan-cancer epithelial-mesenchy-mal transition (EMT) signature (33) or vimentin] were moresensitive to the LDHA inhibitor, oxamate, BPTES, and ROS-inducing agents [BSO and (S)-4-CPG] across a variety of tumortypes (Fig. 4 D and E and Fig. S3C; P < 0.001; Dataset S8). Asimilar discrepant dependency was observed in the slow pro-liferating PDAC cell lines, with six lines more glycolytic and/ormesenchymal and six lines more lipogenic and/or epithelial, despitetheir slower proliferation (Fig. S3 D and E). Thus, mesenchymaltumors, regardless of indication, appear to share common metabolicvulnerabilities, and agents that block glycolysis, glutamine metabo-lism, or redox balance may be particularly effective. These resultssupport a model in which metabolic plasticity with regard to bio-energetic pathways is limited, and, consequently, unique metabolicdependencies exist in tumors that can be exploited for cancertherapy based on tumor subtype.

DiscussionUsing broad metabolite profiling, we successfully stratifiedPDAC-derived cell lines into discrete metabolic subtypes. Pre-vious metabolic profiling studies have been conducted in tumorsand in cell lines of the NCI-60 panel with different end points(9). However, this study is the first, to our knowledge, to suc-cessfully identify metabolic subtypes through profiling of a largenumber of samples within one tissue type and to demonstratethat each subtype is enriched for drug sensitivity to uniqueclasses of metabolic inhibitors.Although metabolic clustering accounted for a substantial frac-

tion of the drug response variation observed across cancer celllines, some heterogeneity in drug response within the lipogenicsubtype remained (see SI Text and Figs. S4 and S5 for a discussionon heterogeneity). Some cell lines were clearly “hard-wired” forlipogenesis and showed sensitivity to all lipid inhibitors tested,whereas the more refractory lines appeared to be capable ofswitching to alternative pathways, perhaps those involving fatty aciduptake. Further understanding of the nature and plasticity of met-abolic networks in these cancer cells will be required to more ac-curately predict their sensitivity to specific classes of metabolicinhibitors. In addition, although we successfully translated our invitro findings in vivo, additional factors within the tumor microen-vironment (tumor-stroma signaling, angiogenesis, and hypoxia) willinfluence sensitivity and adaptation to metabolic inhibition in vivo.Our study also identified PEP as one of the most differentially

expressed metabolites between glycolytic and lipogenic cell lines.

ENO2, which converts 2-phosphoglycerate (2-PG) to PEP, wasalso one of the most differentially expressed genes between thesetwo subtypes, suggesting that inhibitors of ENO2 may be par-ticularly effective against glycolytic tumors. Enolases act down-stream of phosphoglycerate mutase (PGAM1) and regulatepyruvate kinase (PK) M2 isoform (PKM2), genes that are par-ticularly active in glycolytic tumors and have recently attractedattention for their role in serine biosynthesis through regulationof 3-phosphoglycerate dehydrogenase (PHGDH) (34). ENO2 hasalso been proposed as a target in ENO1-deleted glioblastomas(35). Our findings further substantiate the biological rationalefor targeting ENO2 in a subset of cancers.Finally, we demonstrated that the observed metabolic sub-

types correlate with epithelial vs. (quasi)-mesenchymal cell statesboth in PDAC and other cancer types. We propose a model (Fig.4F) in which mesenchymal tumors are metabolically wired topreferentially use glucose for glycolysis and lactate productionand use glutamine for generating TCA metabolites, whereasepithelial tumors preferentially use glucose for the TCA cycleand de novo lipogenesis. Moreover, our analysis suggests thatmesenchymal tumors may be more vulnerable to ROS-inducingagents, potentially through differences in NADPH balance andantioxidant responses (36).Such differences in metabolic vulnerabilities between epithelial

and mesenchymal states could arise from the activation of signalingpathways associated with these states. For example, epithelialsubtypes have previously been shown to be enriched for activatingmutations in receptor tyrosine kinases (RTK) such as EGFR (37)and PI3K/AKT signaling pathways (23), leading to activation of themechanistic target of rapamycin (mTOR). mTOR increases bothprotein synthesis and lipogenesis through mechanisms includingenzyme phosphorylation and transcriptional activation of EIF1A(38) and SREBP1 (39–41). In contrast, mesenchymal states areassociated with increased c-Myc expression and HIF1A, which havebeen shown to drive a glycolytic profile (42, 43). Regardless of thenature or mechanism of action for the metabolic variation we ob-served, our data provide valuable predictive utility and therebyinform clinical evaluation of a variety of metabolic inhibitors suchas MCT and glutaminase inhibitors currently undergoing phase Itesting across a variety of tumor indications.

Materials and MethodsDetailed materials and methods can be found in SI Materials and Methods.All cell lines listed in Dataset S9 were grown in RPMI (without glucose,without glutamine) media (US Biological #R9011) supplemented with 6 mMglucose, 2 mM glutamine, 5% FBS, 100 μg/mL penicillin, and 100 U/mLstreptomycin. Metabolite profiling was performed as previously described(44). For flux analysis, cells were cultured for ∼18 h in RPMI with 10%(vol/vol) dialyzed FBS supplemented with either 3 mM D[U-13C]glucose or1 mM L[U-13C]glutamine. Data analysis was carried out with the MultiQuantsoftware. For short-term viability assays, cells were plated using optimalseeding densities in 384-well plates. The following day, cells were treatedwith LDHA inhibitor GNE-140 (26), oxamate (Sigma cat# O2751), SCDinhibitor (28), FASN inhibitor GSK1195010 (27), cerulenin, orlistat, BSO,S-4-CPG, aminooxyacetic acid (AOA), and BPTES (45), using a 6-pt dose titra-tion scheme. After 72 h, cell viability was assessed using the CellTiter-GloLuminescence Cell Viability assay. Absolute inhibitory concentration (IC)values were calculated using four-parameter logistic curve fitting and areaverages from a minimum of two independent experiments. For long-termgrowth assays, glycolytic cell lines (MIA Paca-2, SW 1990, PSN1, and HUP-T3)and lipogenic cell lines (PA-TU-8902, PK-8, KP-3L, and SUIT-2) were seeded ina 6-well dish at 3,000 cells per well overnight (RPMI, 5% serum, 2 mM glu-tamine) and then treated in media with indicated concentrations of oxamate,SCD inhibitor, or DMSO for 12 d at 37 °C and 5%CO2. Fatty uptake assays wereperformed using the Free Fatty Acid Uptake Assay Kit (ab176768) accordingto the manufacturer’s protocol. Reduced serum experiments were carriedout using 3% delipidated serum (SeraCare 502099) and 1% FBS (SeraCareCC5010-500). Seahorse Bioscience assays were used for oxygen consump-tion. All procedures involving animals were reviewed and approved by theInstitutional Animal Care and Use Committee (IACUC) at Genentech and

Daemen et al. PNAS Early Edition | 7 of 8

CELL

BIOLO

GY

PNASPL

US

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1

Page 8: Metabolite profiling stratifies pancreatic ductal ... · 7/21/2015  · Metabolite profiling stratifies pancreatic ductal adenocarcinomas into subtypes with distinct sensitivities

carried out in an AAALAC (Association for the Assessment and Accreditation ofLaboratory Animal Care) accredited facility. All statistical analyses were per-formed in R 3.0.0 (46). The optimal number of metabolic subtypes wasobtained with nonnegative matrix factorization, using the NMF package. TheDESeq2 package was used for differential expression analysis. Metabolic on-tology and gene set enrichment analyses were based on GSEAlm.

ACKNOWLEDGMENTS. We thank Richard Bourgon, Eva Lin, Billy Lam,Yihong Yu, and Arjan Gower for help with cell-based drug screens anddata analysis, Mandy Kwong for advice on 13C metabolic mass iso-topomer distribution analysis (MIDA), Allison Bruce for assistance withthe metabolic diagram, and Metanomics Health (Lisette Leonhardt,Ulrike Rennefarhrt, Oliver Schmitz, and Hajo Schiewe) for technical sup-port on metabolite profiling.

1. Warburg O, Posener K, Negelein E (1924) On the metabolism of carcinoma cells.Biochem Z 152:309–344.

2. Dang L, et al. (2009) Cancer-associated IDH1 mutations produce 2-hydroxyglutarate.Nature 462(7274):739–744.

3. Deberardinis RJ, Sayed N, Ditsworth D, Thompson CB (2008) Brick by brick: Metabo-lism and tumor cell growth. Curr Opin Genet Dev 18(1):54–61.

4. Hsu PP, Sabatini DM (2008) Cancer cell metabolism: Warburg and beyond. Cell 134(5):703–707.

5. Anastasiou D, et al. (2012) Pyruvate kinase M2 activators promote tetramer formationand suppress tumorigenesis. Nat Chem Biol 8(10):839–847.

6. Le A, et al. (2012) Glucose-independent glutamine metabolism via TCA cycling forproliferation and survival in B cells. Cell Metab 15(1):110–121.

7. Xie H, et al. (2014) Targeting lactate dehydrogenase–a inhibits tumorigenesis andtumor progression in mouse models of lung cancer and impacts tumor-initiating cells.Cell Metab 19(5):795–809.

8. Available at www.clinicaltrials.gov. Accessed December, 2014.9. Jain M, et al. (2012) Metabolite profiling identifies a key role for glycine in rapid

cancer cell proliferation. Science 336(6084):1040–1044.10. Sreekumar A, et al. (2009) Metabolomic profiles delineate potential role for sarcosine

in prostate cancer progression. Nature 457(7231):910–914.11. Hu J, et al. (2013) Heterogeneity of tumor-induced gene expression changes in the

human metabolic network. Nat Biotechnol 31(6):522–529.12. Dang CV (2010) Glutaminolysis: Supplying carbon or nitrogen or both for cancer cells?

Cell Cycle 9(19):3884–3886.13. Barretina J, et al. (2012) The Cancer Cell Line Encyclopedia enables predictive mod-

elling of anticancer drug sensitivity. Nature 483(7391):603–607.14. Garnett MJ, et al. (2012) Systematic identification of genomic markers of drug sen-

sitivity in cancer cells. Nature 483(7391):570–575.15. Ying H, et al. (2012) Oncogenic Kras maintains pancreatic tumors through regulation

of anabolic glucose metabolism. Cell 149(3):656–670.16. Son J, et al. (2013) Glutamine supports pancreatic cancer growth through a KRAS-

regulated metabolic pathway. Nature 496(7443):101–105.17. Zhang G, et al. (2013) Integration of metabolomics and transcriptomics revealed a

fatty acid network exerting growth inhibitory effects in human pancreatic cancer.Clin Cancer Res 19(18):4983–4993.

18. Commisso C, et al. (2013) Macropinocytosis of protein is an amino acid supply route inRas-transformed cells. Nature 497(7451):633–637.

19. McDermott U, et al. (2007) Identification of genotype-correlated sensitivity to selec-tive kinase inhibitors by using high-throughput tumor cell line profiling. Proc NatlAcad Sci USA 104(50):19936–19941.

20. Heiser LM, et al. (2012) Subtype and pathway specific responses to anticancer com-pounds in breast cancer. Proc Natl Acad Sci USA 109(8):2724–2729.

21. Brunet JP, Tamayo P, Golub TR, Mesirov JP (2004) Metagenes and molecular patterndiscovery using matrix factorization. Proc Natl Acad Sci USA 101(12):4164–4169.

22. Collisson EA, et al. (2011) Subtypes of pancreatic ductal adenocarcinoma and theirdiffering responses to therapy. Nat Med 17(4):500–503.

23. Cancer Genome Atlas Network (2012) Comprehensive molecular portraits of humanbreast tumours. Nature 490(7418):61–70.

24. Cancer Genome Atlas Research Network (2014) Comprehensive molecular charac-terization of gastric adenocarcinoma. Nature 513(7517):202–209.

25. Mutch DM, et al. (2009) Metabolite profiling identifies candidate markers reflectingthe clinical adaptations associated with Roux-en-Y gastric bypass surgery. PLoS One4(11):e7905.

26. O’Brien T, et al. (2014) Abstract 964: Inhibiting glycolysis with an LDHA inhibitor: Anew solution to an old problem. Cancer Res 74(19, Suppl):964.

27. Chaudhari AM, Hallman J, Laudeman CP, Musso DL, Parrish CA (2011) Azabenzimi-dazoles as fatty acid synthase inhibitors and their preparation and use for thetreatment of cancer. US Patent WO 2011066211 A1 (June 3, 2011).

28. Du X, et al. (2012) FGFR3 stimulates stearoyl CoA desaturase 1 activity to promotebladder tumor growth. Cancer Res 72(22):5843–5855.

29. Acharya A, Das I, Chandhok D, Saha T (2010) Redox regulation in cancer: A double-edged sword with therapeutic potential. Oxid Med Cell Longev 3(1):23–34.

30. Ashkenazi A, Du X, Qing J (2013) Methods of using scd1 antagonists. US PatentWO2013056148 A2 (April 18, 2013).

31. Thiery JP, Acloque H, Huang RY, Nieto MA (2009) Epithelial-mesenchymal transitionsin development and disease. Cell 139(5):871–890.

32. Peinado H, Olmeda D, Cano A (2007) Snail, Zeb and bHLH factors in tumour pro-gression: An alliance against the epithelial phenotype? Nat Rev Cancer 7(6):415–428.

33. Cheng WY, Ou Yang TH, Anastassiou D (2013) Development of a prognostic modelfor breast cancer survival in an open challenge environment. Sci Transl Med 5(181):181ra50.

34. Hitosugi T, et al. (2012) Phosphoglycerate mutase 1 coordinates glycolysis and bio-synthesis to promote tumor growth. Cancer Cell 22(5):585–600.

35. Muller FL, et al. (2012) Passenger deletions generate therapeutic vulnerabilities incancer. Nature 488(7411):337–342.

36. Gorrini C, Harris IS, Mak TW (2013) Modulation of oxidative stress as an anticancerstrategy. Nat Rev Drug Discov 12(12):931–947.

37. Yauch RL, et al. (2005) Epithelial versus mesenchymal phenotype determines in vitrosensitivity and predicts clinical activity of erlotinib in lung cancer patients. Clin CancerRes 11(24 Pt 1):8686–8698.

38. Lomakin IB, Steitz TA (2013) The initiation of mammalian protein synthesis and mRNAscanning mechanism. Nature 500(7462):307–311.

39. Gingras AC, Raught B, Sonenberg N (2001) Regulation of translation initiation byFRAP/mTOR. Genes Dev 15(7):807–826.

40. Luyimbazi D, et al. (2010) Rapamycin regulates stearoyl CoA desaturase 1 expressionin breast cancer. Mol Cancer Ther 9(10):2770–2784.

41. Menendez JA, Lupu R (2007) Fatty acid synthase and the lipogenic phenotype incancer pathogenesis. Nat Rev Cancer 7(10):763–777.

42. Kim JW, Gao P, Liu YC, Semenza GL, Dang CV (2007) Hypoxia-inducible factor 1 anddysregulated c-Myc cooperatively induce vascular endothelial growth factor andmetabolic switches hexokinase 2 and pyruvate dehydrogenase kinase 1. Mol Cell Biol27(21):7381–7393.

43. Yoo YG, Christensen J, Huang LE (2011) HIF-1α confers aggressive malignant traits onhuman tumor cells independent of its canonical transcriptional function. Cancer Res71(4):1244–1252.

44. van Ravenzwaay B, et al. (2007) The use of metabolomics for the discovery of newbiomarkers of effect. Toxicol Lett 172(1-2):21–28.

45. Shukla K, et al. (2012) Design, synthesis, and pharmacological evaluation of bis-2-(5-phenylacetamido-1,2,4-thiadiazol-2-yl)ethyl sulfide 3 (BPTES) analogs as glutaminaseinhibitors. J Med Chem 55(23):10551–10563.

46. R Development Core Team (2009) R: A Language and Environment for StatisticalComputing (R Foundation for Statistical Computing, Vienna).

47. Jiang Z, Gentleman R (2007) Extensions to gene set enrichment. Bioinformatics 23(3):306–313.

48. Gaujoux R, Seoighe C (2010) A flexible R package for nonnegative matrix factoriza-tion. BMC Bioinformatics 11:367.

49. Kim PM, Tidor B (2003) Subsystem identification through dimensionality reduction oflarge-scale gene expression data. Genome Res 13(7):1706–1718.

50. Wu TD, Nacu S (2010) Fast and SNP-tolerant detection of complex variants andsplicing in short reads. Bioinformatics 26(7):873–881.

51. Klijn C, et al. (2015) A comprehensive transcriptional portrait of human cancer celllines. Nat Biotechnol 33(3):306–312.

52. Possemato R, et al. (2011) Functional genomics reveal that the serine synthesispathway is essential in breast cancer. Nature 476(7360):346–350.

53. Mootha VK, et al. (2003) PGC-1alpha-responsive genes involved in oxidative phos-phorylation are coordinately downregulated in human diabetes. Nat Genet 34(3):267–273.

54. McCleland ML, et al. (2013) Lactate dehydrogenase B is required for the growth ofKRAS-dependent lung adenocarcinomas. Clin Cancer Res 19(4):773–784.

55. Subramanian A, et al. (2005) Gene set enrichment analysis: A knowledge-basedapproach for interpreting genome-wide expression profiles. Proc Natl Acad Sci USA102(43):15545–15550.

56. Cheng WY, Ou Yang TH, Anastassiou D (2013) Biomolecular events in cancer revealedby attractor metagenes. PLOS Comput Biol 9(2):e1002920.

57. Gray DC, et al. (2007) pHUSH: A single vector system for conditional gene expression.BMC Biotechnol 7:61.

58. Pinheiro J, Bates D, DebRoy S, Sarkar D, R Development Core Team (2009) nlme:Linear and nonlinear mixed effects models. R Package Version 3.1-96. Available atCRAN.R-project.org/package=nlme. Accessed July 8, 2015.

59. Benjamini Y, Hochberg Y (1995) Controlling the false discovery rate: A practical andpowerful approach to multiple testing. J R Stat Soc B 57(1):289–300.

8 of 8 | www.pnas.org/cgi/doi/10.1073/pnas.1501605112 Daemen et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 7,

202

1