OPEN

Citation: Oncogenesis (2017) 6, e329; doi:10.1038/oncsis.2017.34

[Link]/oncsis

ORIGINAL ARTICLE
Inhibition of malic enzyme 1 disrupts cellular metabolism and
leads to vulnerability in cancer cells in glucose-restricted
conditions
S Murai1, A Ando2, S Ebara1, M Hirayama2, Y Satomi2 and T Hara1

Malic enzyme 1 (ME1) regulates one of the main pathways that provide nicotinamide adenine dinucleotide phosphate (NADPH),
which is essential for cancer cell growth through maintenance of redox balance and biosynthesis processes in the cytoplasm. In this
study, we found that ME1 inhibition disrupted metabolism in cancer cells and inhibited cancer cell growth by inducing senescence
or apoptosis. In glucose-restricted culture conditions, cancer cells increased ME1 expression, and tracer experiments with labelled
glutamine revealed that the flux of ME1-derived pyruvate to citrate was enhanced. In addition, cancer cells showed higher
sensitivity to ME1 depletion in glucose-restricted conditions compared to normal culture conditions. These results suggest that in a
low-glucose environment, where glycolysis and the pentose phosphate pathway (PPP) is attenuated, cancer cells become
dependent on ME1 for the supply of NADPH and pyruvate. Our data demonstrate that ME1 is a promising target for cancer
treatment, and a strategy using ME1 inhibitors combined with inhibition of glycolysis, PPP or redox balance regulators may provide
an effective therapeutic option.

Oncogenesis (2017) 6, e329; doi:10.1038/oncsis.2017.34; published online 8 May 2017

INTRODUCTION dependent isoform, ME 1 (ME1); a mitochondrial NAD+-dependent

In malignant tumours, oncogenic gene alterations such as point isoform, ME 2 (ME2); and a mitochondrial NADP+-dependent
mutations, translocations and gene amplification and deletion are isoform, ME 3 (ME3).8,9 Different from the mitochondrial ME2 and
recurrently observed, and the oncogenes and their downstream ME3, ME1 produces NADPH in the cytoplasm. It has been reported
signals have been targeted for cancer therapy.1 In recent decades, that ME1 produces NADPH at levels as high as those produced by
it has become clear that cancer cells alter their cellular metabolism G6PD in the PPP shunt, which is a major cellular source of
to adapt to gene and microenvironment alterations to sustain NADPH.10 ME1 levels are used as a prognosis and predictive
growth and survival.2 Therefore, targeting the reprogramming of marker for radiation therapy in cancer.11,12 Therefore, ME1 is
cellular metabolism to treat cancers has currently come under the considered a promising target for cancer therapy and has been
spotlight.3 evaluated as such.13,14
Cancer-associated metabolic reprogramming affects In this study, we explored the use of ME1 as a target of cancer
gene expression, cellular differentiation and tumour microenvir- treatment through testing the vulnerability of cancer cells to
onment, and these characteristics were recently summarised by ME1 inhibition by analysing the resulting alterations in gene
Pavlova NN et al.4 as the six hallmarks of cancer metabolism. One expression and metabolite levels.
hallmark corresponds to the use of glycolysis and tricarboxylic
acid (TCA) intermediates for biosynthesis and nicotinamide
adenine dinucleotide phosphate (NADPH) production.
RESULTS
NADPH functions as a reducing agent in cells and has critical
roles in macromolecular biosynthesis and redox homoeostasis. ME1 localises in the cytosol and is expressed in various cancer cell
NADPH is produced by metabolic enzymes such as glucose-6- lines
phosphate dehydrogenase (G6PD) and 6-phosphogluconate To identify the localisation of MEs in cancer cells, we conducted
dehydrogenase (6PGD) of the pentose phosphate pathway immunocytochemistry. U2OS cells were transfected with ME1,
(PPP), malic enzymes (MEs), isocitrate dehydrogenases (IDHs) ME2 or ME3 cDNAs and immuno-stained with each antibody;
and enzymes in one-carbon-tetrahydrofolate oxidation pathways. mitochondria were stained with Mito-Tracker. As previously
These NADPH-producing enzymes have been evaluated as reported,8,9 ME1 localised in the cytoplasm, whereas ME2 and
potential therapeutic targets of cancer.5–7 ME3 localised in the mitochondria (Figure 1a). We conducted
MEs regulate cellular energy, redox balance and biomolecular TaqMan analysis to investigate the expression profile of MEs in
synthesis by converting the TCA cycle intermediate malate into cells of various cancer types. We found that most cancer cell
the TCA carbon source pyruvate and NADPH. In mammalian cells, lines express higher levels of ME1 than of ME2 and ME3, although
three isoforms of MEs have been identified: a cytosolic NADP+- some cell lines predominantly expressed ME2 and ME3

1
Oncology Drug Discovery Unit, Pharmaceutical Research Division, Takeda Pharmaceutical Company, Kanagawa, Japan and 2Integrated Technology Research Laboratories,
Pharmaceutical Research Division, Takeda Pharmaceutical Company, Kanagawa, Japan. Correspondence: S Murai or Dr T Hara, Pharmaceutical Research Division, Takeda
Pharmaceutical Company Limited, 26-1, Muraoka-Higashi 2-chome, Fujisawa, Kanagawa 251-8555, Japan.
E-mail: [Link]@[Link] or [Link]@[Link]
Received 21 November 2016; revised 17 February 2017; accepted 27 March 2017
 Role of malic enzyme 1 in cancer cells
S Murai et al
2
14
12 ME1

Relative Expression (% of actin)

ME1 ME2 ME3 10 ME2
DAPI DAPI DAPI 8
ME3
6
4
2
0

T84
RKO

ES-2
WiDr
SW48
HT-29
DLD-1

HCT-8

SW403

SW480
SW620
SW837
SW948

CaOV3

SKOV3
RKOE6
HCT-15

LS1034

SW1116
SW1417
LS-174T
HCT-116
Colo 201
Colo 205

OVCAR3
NCI-H508
NCI-H747
Colo 320DM

RKO-AS45-1
Colo 320 HSR
ME1 ME2 ME3

Colon Ovary

12
Mitochondria Mitochondria Mitochondria
10

Relative Expression (% of actin)

ME1
8 ME2

6 ME3

4
Merge Merge Merge
2

PANC-1
Calu6

Hs 766T
Calu1

ChaGoK1

SHP-77

SKMES1

Wi38

YAPC
NCI-H292

NCIH446

NCI-H460

NCI-H520

NCI-H596

NCI-H661

AsPC-1

BxPC-3

Capan-1

HPAF-II

Mia PaCa-2
Capan-2

CFPAC-1

HuP-T4
A427

A549

NCI-H69

NCI-H82

SW900

SU.86.86
Lung Pancreas

Figure 1. ME1 localises in the cytosol and is expressed in cancer cell lines. (a) U2OS cells were transfected with ME1, ME2 or ME3 cDNA plasmid
and, 48 h later, cells were fixed and stained with DAPI, anti-ME1, anti-ME2 and anti-ME3 antibodies, and Mito-tracker, which specifically stains
mitochondria. Nucleus (blue), ME1, ME2 or ME3 (green), mitochondria (red) and the merged image are shown. (b) Basal ME1, ME2 and ME3
mRNA expression levels in colorectal, ovarian, lung and pancreatic cancer cell lines were determined by TaqMan PCR. Expression levels were
normalised by expression of β-actin in each cell line and the relative expression is shown in the bar graph (n = 3).

250
si ME2-1
Relative Total Area (µm^2)

N.T. si ME1-1 si Cont

200

150

100
si Cont si ME1-2 si ME2-2 si ME1-1

50

0
si .T.
M t
M 1
M 2
M 1
-2
si Con
si E1-
si E1-
si E2-
E2
N

200
120
Relative Cell Proliferation (%)

N.T. 60
SA-b-Gal Positive Cell (%)

100 160 si Control

50
si ME1-1
80
120 si ME1-2 40
Counts

60
30
80
40 20
40
20 10

0 0 0
0 200 400 600 800 1000
t

-1

-2

si .T.
si ont

si 1-1

-2
on

E1

E1

E1
N
C

FSC-H
M

M
si

si

si

Figure 2. ME1 knockdown induces cellular senescence and suppresses cell growth. (a,b) HCT116 cells were reversely transfected with control,
ME1, or ME2 siRNAs and 1000 cells were seeded on six-well plates. Then, cells were cultured for 11 days for colony formation assay. Colonies
were fixed and stained with Crystal violet, and colony area of each well was quantified by GelCount and relative total area is shown as a bar
graph (n = 3, mean with s.d.). (c) Light microscopy images of colonies of HCT116 cells transfected with si control and si ME1-1. (d) PC3 cells
were transfected with ME1 siRNAs and viability was determined 96 h after transfection. Relative cell proliferation is shown as a bar graph
(n = 3, mean with s.d.). (e) PC3 cells were transfected with ME1 siRNAs and cultured for 96 h. Cell size forward scatter (FSC) was detected by
FACS analysis. (f) PC3 cells were transfected with ME1 siRNAs for 9 days and β-galactosidase was then stained. Cell (outer white line) and its
nucleus (inner white line) are shown in the upper photograph. β-galactosidase was stained in blue and is shown in the lower photograph.
(g) β-galactosidase-positive cells were manually counted and the percentage of stain-positive (SA-b-Gal positive) cells are shown in the
bar graph. FACS, fluorescence-activated cell sorting.

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
3
Malate (m+4) Fumarate (m+4) Pyruvate (m+3)
Glucose
150 * 800 8
#

Malate Pyruvate Lactate 600 6

100

Peak Area

Peak Area
Peak Area
ME1
400 4

Pyruvate 50
Mitochondria
200 2

Ac-CoA
Citrate 0 0 0
synthetase Reductive

-1

-1

-1
Oxaloacetate

.T

.T

.T
on

on

on
E1

E1

E1
carboxylation

N
C

C
Malate

M
si

si

si
Citrate Citrate

si

si

si
Fumarase IDH2/3 Citrate (m+2) Citrate (m+4) Citrate (m+6)
*
Fumarate 40 150 40
Isocitrate Isocitrate #
Succinate
IDH1 30 30
SDH α-ketoglutarate 100

Peak Area

Peak Area
Peak Area
α-ketoglutarate
GLS2 20 20
GLS1
Glutamate Glutamate 50
Oxidative 10 10
glutaminolysis
Glutamine [U- N, U- C] Glutamine 0 0 0

-1

-1

-1
.T

.T

.T
on

on

on
E1

E1

E1
N

N
C

C
M

M
si

si

si
si

si

si
Figure 3. [U-13C, U-15N] L-glutamine isotope profiling analysis revealed malate metabolism is suppressed by si ME-1. (a) Schematic
representation of glutamine isotope profiling. Blue and purple circles indicate a carbon isotope, metabolized through oxidative glutaminolysis
and reductive carboxylation, respectively, pink circles a nitrogen isotope and white circles non-labelled carbons. (b) [U-13C, U-15N]
L-glutamine profiling analysis was performed in the HCT116 cell line. Cells were cultured for 72 h after siRNA transfection and labelled for 24 h
in McCoy's 5A media with 2 mM [U-13C, U-15N] L-glutamine. Cell pellets were then collected for flux analysis. Peak area of each metabolite is
shown in the dot graph (n = 3, t-test; *P o0.01; #P o0.05).

(Figure 1b). Therefore, we examined the effects of ME1 inhibition ME1 knockdown disturbs malate levels and its downstream
on cancer cell proliferation and metabolism. metabolism in the HCT116 cell line
To examine the metabolic alterations resulting from ME1
ME1 knockdown suppresses cancer cell growth and induces depletion, we conducted a tracer experiment with [U-13C,
senescence U-15N] L-glutamine in the HCT116 cell line. The flow of the
To examine the effects of ME1 or ME2 inhibition on HCT116 cell isotopic glutamine tracer is shown in Figure 3a. Incorporated
growth, we performed knockdown experiments using siRNAs. [U-13C, U-15N] L-glutamine is metabolised to fumarate (m+4) and
We excluded ME3 siRNA because ME3 mRNA was not detected malate (m+4) in mitochondria through oxidative glutaminolysis
in HCT116 cells (Figure 1b). Both ME1 and ME2 mRNA and and TCA cycle reactions. A portion of malate (m+4) becomes
protein levels were suppressed by siRNAs (Supplementary Figures oxaloacetate (m+4) and citrate (m+4) through consecutive TCA
S1A and S1B). ME1 depletion suppressed colony formation in cycle reactions in mitochondria. Another portion of malate (m+4)
clonogenic assays (Figures 2b and c), whereas ME2 depletion did is exported to the cytoplasm from mitochondria and is converted
not affect HCT116 cell growth. Some colonies of cells transfected into pyruvate (m+3) by ME1. Pyruvate (m+3) is converted into
with ME1-1 siRNA included enlarged cells (Figure 2d), which is a acetyl-CoA (m+2) in mitochondria, which then generates citrate
hallmark of senescent cells.15 To confirm these observations, we (m+2) or citrate (m+6) (Figure 3a). In the experiment, we used
repeated the experiments with the PC3 cell line. We first examined ME1-1 siRNA, which suppressed ME1 expression but not cell
if ME1 depletion affected PC3 cell growth. ME1 expression was growth 96 h after transfection, as a control to exclude metabolic
suppressed by the two siRNAs in PC3 cell line (Supplementary changes caused by cell growth inhibition. Malate (m+4) and
Figure S1D). In this condition, we found that ME1 depletion fumarate (m+4) were found to be accumulated after ME1
suppressed PC3 cell proliferation 96 h after siRNA reverse- knockdown. Although pyruvate (m+3) levels were not reduced,
transfection (Figure 2d). Next, we conducted clonogenic assays both citrate (m+2) and citrate (m+6) levels decreased, whereas
with the PC3 cell line. PC3 cells were reversely transfected with citrate (m+4) levels did not change (Figure 3b and Supplementary
siRNAs and cultured for nine days. Fluorescence-activated cell Figure S2A). These results suggest that the supply of pyruvate
sorting analysis revealed that the number of enlarged PC3 cells (m+3) and subsequently of acetyl-CoA (m+2) are reduced after
increased after transfection with ME1 siRNAs (Figure 2e), which ME1 knockdown without affecting TCA cycle reactions. We
is consistent with microscopic observation of colonies conclude that ME1 knockdown suppresses malate levels and its
(Supplementary Figure S1E). We investigated the expression of downstream metabolism in the HCT116 cell line.
SA-b-Gal, which is another hallmark of senescence, in the PC3
colonies and found some SA-b-Gal positive cells (Figure 1f, blue ME1 inhibition induces metabolic reprograming and disturbs
staining). We manually counted SA-b-Gal stain-positive and redox homeostasis in cancer cell lines
-negative PC3 cells and calculated the percentage of stained To investigate the effect of ME1 inhibition on cellular metabolism,
cells. The results show that the number of SA-b-Gal stain-positive we conducted metabolomics analysis in HCT116 cell lines after
cells clearly increased after ME1 knockdown (Figure 1g). These ME1 knockdown. We detected some metabolite level changes
results suggest that ME1 depletion induces cellular senescence 48 h or 72 h after siRNAs transfection. Consistent with isotope
and suppresses cancer cell growth. profiling experiments (Figure 3b), malate and fumarate levels

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
4
Malate Ribose-5-Phosphate Glucose-6-Phosphate Pyruvate Lactate
#
#
*
#
15 * * 100 # 40 600 15000 *
* *
*
90 *
#
* 30 *

Peak Area
Peak Area

Peak Area

Peak Area

Peak Area
10 80 400 # 10000
* *
70 20

5 60 200 5000
10
50

0 40 0 0 0

si ME1-1
si ME1-2

si ME1-1
si ME1-2
si Cont

si Cont
si ME1-1
si ME1-2

si ME1-1
si ME1-2
si Cont

si Cont

si Cont

si Cont
si ME1-1
si ME1-2

si ME1-1
si ME1-2

si Cont

si Cont
si ME1-1
si ME1-2

si ME1-1
si ME1-2

si Cont

si Cont
si ME1-1
si ME1-2

si ME1-1
si ME1-2
48 h 72 h 48 h 72 h 48 h 72 h 48 h 72 h 48 h 72 h

4 Glucose (+) Glucose (-)

si Cont
Relative mRNA Expression

3.5

C911 ME1-1

C911 ME1-1
si ME1-1

si Control

si Control
si ME1-1

si ME1-1
3 si ME1-2

N.T.

N.T.
2.5

2
ME1
1.5
HO-1
1

0.5 CDKN1A
0
β-actin
E1

E2

E3

1
M H2

M D1

C FD2

C 1A

C 1B

2A

-1
53

A1

B
H

H
6P

O
TP
M

KN

KN

KN

H
ID

ID

PD
6P

H
TH

TH
G

PD
D

Figure 4. ME1 knockdown disturbs metabolic homeostasis and induces oxidative stress response in cancer cells. (a–c) Metabolomics analysis
was conducted 48 and 72 h after ME1 siRNAs transfection in HCT116 cells. Peak area of malate, ribose-5-phosphate, glucose-6-phosphate,
pyruvate and lactate are shown in the dot graph (n = 3, t-test; *P o0.01; #P o0.05). (d) Gene expressions of NADPH-producing enzymes,
p53-related enzymes, a redox-related gene, and glycolysis-relating enzymes were detected by TaqMan PCR 48 h after ME1 siRNAs transfection
in PC3 cells (n = 3). (e) ME1, HO-1 and β-actin were detected by western blotting 96 h after control siRNA, ME1 siRNAs, or ME1 C911
transfection in HCT116 cells (ME1 is indicated by a black arrow). After siRNA transfection, cells were cultured in 2 g/l glucose McCoy's 5A media
for 72 h, cultured in media with/without glucose for further 24 h, and then the cell lysate was collected for western blotting.

increased after ME1 depletion (Figure 4a and Supplementary not (Figure 4e). These results suggest that ME1 inhibition disturbs
Figure S2B). Intriguingly, the levels of glucose-6-phosphate (G6P) redox homeostasis by regulating glucose metabolism.
and ribose-5-phosphate (R5P), which are metabolites in PPP, also In the H460 cell line, which shows high levels of ME1 expression
increased after ME1 depletion (Figure 4b). Furthermore, pyruvate (Figure 1b), HO-1 expression was enhanced and cell growth was
and lactate levels increased significantly after ME1 inhibition inhibited by ME1 inhibition as observed in HCT116 and PC3 cells
(Figure 4c). These results suggest that glycolysis and PPP are (Supplementary Figures S2E and S2F), although we did not
enhanced as a result of ME1 inhibition in HCT116 cells. Because observe senescence-like cells even in a clonogenic assay (data not
ME1 inhibition induced cellular metabolic reprograming, we shown). On the other hand, caspase-3,7 activity was augmented
speculated that ME1 depletion induces some cellular response. by ME1 knockdown in H460 cell lines, indicating that ME1
Therefore, we investigated the expression of NADPH-producing inhibition induced apoptosis in H460 cells (Supplementary
enzymes and senescence-related genes in HCT116 and PC3 cells. Figure S2G). These results suggest that ME1 inhibition disturbs
ME1 depletion increased CDKN1A and HMOX-1 (HO-1) mRNA cellular metabolism and redox balance in several cancer cell lines
levels (without affecting expression of other NADPH-producing but the specific phenotype, such as senescence or apoptosis,
enzymes in the PC3 cell line, Figure 4d), and the same result was varies in each cell line.
observed in ME1-depleted HCT116 cells (Supplementary Figures
S2C and S2D). To verify that CDKN1A and HO-1 expression were Glucose depletion increases reductive carboxylation and enhances
induced by ME1 inhibition, we examined CDKN1A and HO-1 ME1-dependent glutamine metabolism in HCT116 cells
protein expression in HCT116 cells using C911ME1 siRNA,16 which We found that ME1 inhibition affected glucose metabolism, and
is similar in sequence to ME1 siRNA but was designed to remove that glucose concentration in the medium affected HO-1
gene knockdown ability, as a negative control. We also examined expression induced by ME1 inhibition in HCT116 cells. Therefore,
whether the presence of glucose could affect expression we investigated whether glucose availability affected malate
of CDKN1A and HO-1. The results show that CDKN1A and metabolism and expression. We first conducted cell growth assays
HO-1 protein levels increased after ME1 inhibition and the in glucose-restricted conditions. HCT116 cells were cultured in
phenotypes were reverted by C911 (Figure 4e), suggesting that RPMI1640 media with 0, 0.2, 0.5, 1, and 2 g/l glucose, and cell
ME1 knockdown stimulates CDKN1A and HO-1 expression in viability was measured after 24, 48 and 72 h (normal medium
HCT116 cells. Furthermore, we found that HO-1 expression was contains 2 g/l glucose). Compared to 2 g/l glucose, HCT116 cell
enhanced in glucose-depleted condition but that of CDKN1A was growth was suppressed after 48 h in 0.2 g/l, slightly suppressed

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
5

Glutamate
Labeled Glutamine (% of pool)
100 50 50

Malate
700 2g/L glucose
45 2g/L glucose 2g/L glucose
90 45
2 g/L
80 Glucose free 40 Glucose free 40 Glucose free
1 g/L
70 35 35

(% of pool)
(% of pool)

Labeled Glutamine
600 0.5 g/L 60 30 30

Labeled Glutamine
0.2 g/L * 50 25 25
40 20 20
Relative Cell Growth

Glc free
500 * 30
20
15
10
15
10
10 5 5
400 0 0 0
*

+0
+1
+2
+3
+4
+5
+6
+7

+0

+1

+2

+3

+4

+5

+6

+0

+1

+2

+3

+4
m
m
m
m
m
m
m
m

m
300

Pyruvate
120 45 120

Lactate
Labeled Glutamine Citriate
2g/L glucose 40 2g/L glucose 2g/L glucose
100 100
Glucose free 35 Glucose free Glucose free
200

(% of pool)

(% of pool)
(% of pool) 80 30 80

Labeled Glutamine
Labeled Glutamine

25
60 60
100 20
40 15 40
10
0 20 20
5
0 24 48 72 0 0 0
Hours
+0

+1

+2

+3

+0

+1

+2

+3

+4

+5

+6

+0

+1

+2

+3
m

m
ME1 ME2 G6PD 6PGD IDH1 IDH2
9 1.8 2.5 2.5 3.5 1.8
Relative mRNA Expression

Relative mRNA Expression

Relative mRNA Expression

Relative mRNA Expression

Relative mRNA Expression

Relative mRNA Expression

8 1.6 3 1.6
7 1.4 2 2 1.4
2.5
6 1.2 1.2
1.5 1.5 2
5 1 1
4 0.8 1.5 0.8
1 1
3 0.6 0.6
1
2 0.4 0.5 0.5 0.4
1 0.2 0.5 0.2
0 0 0 0 0 0
0 24 48 72 0 24 48 72 0 24 48 72 0 24 48 72 0 24 48 72 0 24 48 72
Hours Hours Hours Hours Hours Hours

2 1 0.5 0.2 [g/L]

Figure 5. Glucose depletion increases ME1-dependent glutamine metabolism and reduces expression of NADPH-producing enzymes other
than ME1 in HCT116 cells. (a) HCT116 cells were cultured in glucose-depleted or -limited media (2, 1, 0.5 or 0.2 g/l glucose or glucose-free) for
72 h. Cell viability was determined at 24, 48, and 72 h using CellTiter Glo. Relative cell growth, normalised by cell viability at Day 0, is shown in
the line graph (n = 3, t-test; *Po 0.01). (b) HCT116 cells were cultured in media with or without glucose for 24 h and labelled with [U-13C,
U-15N] L-glutamine isotope for an additional 24 h. Then, isotope profiling analysis was conducted. Percentage of the pool of metabolites and
isotopes is shown in the bar graph (2 g/l glucose in black and glucose-free in white; n = 3, mean with s.d.). (c) HCT116 cells were cultured in
media with 2, 1, 0.5 or 0.2 g/l glucose for 72 h and mRNA expression levels of NADPH-producing enzymes at 0, 24, 48 and 72 h were
determined by TaqMan PCR and are shown in the line graph (n = 3).

after 72 h in 0.5 g/l, and suppressed in less than 24 h in glucose- ME1 decreased in a glucose concentration-dependent manner,
free (0 g/l) medium (Figure 5a). whereas ME1 expression levels remained unchanged regardless of
To monitor metabolic alterations in cells in glucose-depleted glucose concentration. These results suggest that HCT116 cells
conditions, we performed a tracer experiment with [U-13C, U-15N] depend in part on ME1 in glucose-restricted conditions (Figure 5c).
L-glutamine in the presence or absence of glucose in the medium. We conclude that glucose depletion induces reductive carboxyla-
No difference in glutamine incorporation or malate (m+4) content tion instead of glycolysis and enhances ME1-dependent glutamine
was found; however, pyruvate (m+3), citrate (m+6) and lactate metabolism in HCT116 cells.
(m+3) levels increased (Figure 5b), suggesting that malate
(m+4) metabolism is enhanced in glucose-depleted conditions. ME1 knockdown synergistically suppress HCT116 cell growth in
Furthermore, malate (m+0) and citrate (m+0) levels decreased glucose-depleted conditions
in glucose-free conditions, suggesting that glucose-derived Our experiments suggested that malate metabolism and
metabolites are reduced in cells. Malate (m+3) levels increased glycolysis are coordinately regulated to maintain cellular
in glucose-free conditions, which may be due to activation of redox homeostasis. We next examined whether ME1 inhibition
reductive carboxylation of glutamine. Reductive carboxylation synergistically suppressed cell growth in glucose-depleted
produces citrate (m+5) from glutamine (m+7) and results in conditions. We initially established shRNA tetracycline-inducible
production of malate (m+3) through TCA cycle reactions HCT116 cell clones, which expressed different shRNA sequences
(Figure 3a). Production of malate (m+3) results in an increase (sh control, sh ME1-1 or sh ME1-2) when induced by doxycycline.
in downstream metabolites such as pyruvate (m+2) and lactate ME1 mRNA levels were suppressed 48 h after doxycycline
(m+2) (Figure 5b). treatment in HCT116 clones (Figure 6a). We also examined ME1
To examine the effects of glucose-restricted conditions on the protein expression in sh control, sh ME1-1 and sh ME1-2 clones
expression of NADPH-producing enzymes, mRNA expression cultured for 48 h or 72 h with or without doxycycline in 2 g/l or
levels of ME1, ME2, G6PD, 6PGD, IDH1 and IDH2 were determined 0.5 g/l glucose. ME1 expression was suppressed by ME1 shRNA
in HCT116 cells cultured in glucose-restricted media for 24, 48 and regardless of glucose concentration (Supplementary Figures S3A
72 h. Expression levels of NADPH-producing enzymes except and S3B). In contrast, ME2 mRNA expression levels remained

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
6
1.2
Dox - Dox -
1.2 3 *
Dox + Dox +
1 *
* *
Relative ME1 Expression
1.2 1 * 2.5
*

Relative glucose conc.

Relative lactate conc.

0.8
1

Relative Cell Number

0.8 2
0.6 0.8
0.6 1.5
0.6
0.4
0.4 1
0.4
0.2 0.2 0.5
0.2

0 0 0 0
- Dox - Dox - Dox - Dox - Dox - Dox

sh Cont sh ME1-1 sh ME1-2 sh Cont sh ME1-1 sh ME1-2

Malate (m+4) Citrate (m+2) Lactate (m+0) NADPH

*
25000 5000 1500000 20000 # #
*
*
*
20000 4000
* 15000
* 1000000
Peak Area

Peak Area

Peak Area

Peak Area
15000 3000
10000
10000 2000
500000
5000
5000 1000

0 0 0 0
Dox

Dox

Dox

Dox

Dox

Dox

Dox

Dox
-

-
sh ME1-1 sh ME1-2 sh ME1-1 sh ME1-2 sh ME1-1 sh ME1-2 sh ME1-1 sh ME1-2

sh Cont sh ME1-1 sh ME1-2

120 120 120
*
100 100 100 *
*
Reltive Cell Number
Reltive Cell Number

Reltive Cell Number

80 80 80
*

60 60 60

40 40 40

20 20 20

0 0 0
- Dox - Dox - Dox - Dox - Dox - Dox
2 g/L 0.5 g/L 2 g/L 0.5 g/L 2 g/L 0.5 g/L

Figure 6. ME1 knockdown by shRNA induces metabolic reprograming and synergistically suppresses cancer cell growth in glucose-depleted
conditions. HCT116 cell clones (sh ME1-1 and sh ME1-2), which are stably expressing tetracycline-inducible ME1 shRNA, were established.
(a) Control shRNA, sh ME1-1 or sh ME1-2 clones were cultured in the media with or without 2 μg/ml doxycycline for 48 h, and ME1 expression
was detected by TaqMan PCR. ME1 expression was normalised by GAPDH expression and relative ME1 expression is shown in the bar
graph (n = 3). (b) Cell numbers were determined 72 h with or without 2 μg/ml doxycycline. Relative cell number is shown in the bar graph
(n = 3, mean with s.d., t-test; *Po 0.01) (c) Control shRNA, sh ME1-1 or sh ME1-2 were cultured in the media with or without 2 μg/ml
doxycycline for 96 h and, D-glucose and L-lactate concentration in the media were measured. Relative glucose or lactate concentration are
shown in the bar graph (n = 3, mean with s.d., t-test; *Po 0.01). (d,e) [U-13C, U-15N] L-glutamine isotope profiling analysis was conducted in sh
ME1-1 and sh ME1-2 clones. Clones were cultured in the media with 2 g/l glucose with or without doxycycline for 72 h and labelled for
additional 24 h in the glucose-depleted media with or without doxycycline. Then, metabolites and isotopes were measured. Peak area of
labelled metabolites are shown in the dot graphs (n = 3, t-test; *P o0.01; #Po0.05). (f) Three clones were cultured for 24 h in 2 g/l medium
with or without doxycycline, and additionally cultured for 72 h in medium with 2 or 0.5 g/l glucose with or without doxycycline. Then, cell
numbers were determined by cell counter (n = 3, mean with s.d.; t-test; *P o0.01).

unchanged after doxycycline treatment (Supplementary in D-glucose levels and increase in L-lactate levels were observed
Figure S3C). Cell proliferation was partially suppressed by ME1 after ME1 shRNA induction (Figure 6c), suggesting that glycolysis
shRNA induction in HCT116 clones (Figure 6b). We determined is activated by ME1 inhibition in these clones. We next conducted
D-glucose and L-lactate concentrations in the media in which sh a tracer experiment with [U-13C, U-15N] L-glutamine in HCT116
control, sh ME1-1 or sh ME1-2-transfected cells were cultured for clones under glucose-depleted conditions. We initially determined
96 h with or without doxycycline. Statistically significant decrease if doxycycline affected cellular metabolism in the sh control clone,

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
7
and found that doxycycline did not affect metabolite levels (m+6) levels, but not in citrate (m+4) levels, indicates that
in glucose-sufficient or -depleted conditions (Supplementary the malate-pyruvate-acetyl-CoA flux is also suppressed by
Figure S3E). In sh ME1-1 and sh ME1-2 clones, accumulation of ME1 knockdown. Although pyruvate (m+3) reduction was
malate (m+4) and lactate (m+0) and reduction of citrate (m+2) expected after ME1 knockdown, no significant changes
were observed in both clones after ME1 knockdown (Figure 6d in pyruvate levels were observed. This may be due to pyruvate
and Supplementary Figure S3D), indicating that malate (m+3) being immediately converted to acetyl-CoA and the
metabolism is suppressed and glycolysis is activated in these malate-to-pyruvate conversion being rate-limiting in cells. Basal
clones. We also observed significant reduction in NADPH levels pyruvate levels would thus be kept low, and a reduction would be
after ME1 knockdown in sh ME1 clones (Figure 6e), suggesting difficult to detect.
that redox balance is disturbed in these clones. To investigate the Pyruvate accumulation was observed in our metabolomics
effect of ME1 depletion on cell growth in glucose-depleted analysis (Figure 4c). We also observed lactate accumulation by
conditions, we conducted cell growth assays with these clones ME1 siRNA (Figure 4c) and ME1 shRNAs (Figure 6d), which may be
under these conditions. HCT116 clones were treated with or caused by upregulation of glycolysis. Increase of R5P and G6P
without doxycycline and cultured for 24 h in normal culture levels supports this hypothesis (Figure 4b). We also found that
conditions. Clones were then cultured for 96 h in normal (2 g/l) or glucose levels decreased and lactate levels increased in the media
0.5 g/l glucose medium. ME1 knockdown significantly suppressed after ME1 knockdown in HCT116 cells (Figure 6c), suggesting that
cell growth in glucose-restricted compared to normal culture the Warburg Effect is enhanced by ME1 inhibition in HCT116 cells.
conditions, indicating that glucose depletion and ME1 inhibition We demonstrated that ME1 knockdown induced HO-1 expression,
have a synergistic effect in HCT116 cell clones (Figure 6f and a marker of oxidative stress (Figures 4d and e). Cancer cells may
Supplementary Figure S4C). Similar results were observed in an enhance glycolysis after ME1 depletion to compensate for the
additional clone (Supplementary Figures S4A and S4B). In contrast, reduction in NADPH by activating PPP in the cytoplasm and may
glucose-restricted conditions did not synergistically suppress adapt to an oxidative state for protection. HO-1 expression
cell growth of a non-ME1 silencing control clone (Figure 6f and after ME1 inhibition was higher in glucose-depleted compared
Supplementary Figure S4C). These results suggest that ME1 to glucose-sufficient conditions (Figure 4e), suggesting that
knockdown synergistically inhibits HCT116 cell growth in glycolysis may relieve the oxidative state in HCT116 cells.
glucose-depleted condition. Alternatively, because pyruvate is known to have antioxidative
function,19,20 cancer cells may manage the oxidative state
by increasing pyruvate levels in addition to the activation of the
DISCUSSION PPP pathway.
NADPH production is regulated by multiple metabolic Although both si ME1-1 and si ME1-2 suppressed ME1
pathways to maintain cellular energy, reducing equivalents and expression at mRNA and protein levels (Supplementary Figures
biomolecules. Cancer cells regulate NADPH homeostasis through S1A and S1B) and both enhanced malate accumulation
multiple pathways such as the PPP and the flux through ME1 (Figure 4a), which is a target engagement marker of ME1
and IDH1 in the cytoplasm. Recently, stepwise oxidation of inhibition, si ME1-2 showed stronger cell growth inhibition than
one-carbon-tetrahydrofolate species was also reported to have si ME1-1. We believe that si ME1-2 has an off-target effect in
a role in cellular NADPH production.17 addition to ME1 inhibition. Enhanced accumulation of pyruvate
ME1 metabolism is known to be regulated by well-known and lactate by si ME1-2 compared to si ME1-1 may be caused by
oncogenes or tumour suppressors such as K-ras, myc or p53.7,13 this off-target effect.
Furthermore, ME1 is reported to be a potential prognostic or We demonstrated that ME1 inhibition induces metabolic
sensitivity marker of radio therapy.11,12 Our immunocytochemis reprogramming and abrogates redox balance in cancer cells,
study revealed that ME1 localised in the cytosol, whereas ME2 and both being related to glucose metabolism. Therefore, we
ME3 localised in the mitochondria, which is consistent with investigated if glucose concentration in the media could affect
previous findings.8,9 We found that ME1 is highly expressed in cellular malate metabolism and ME levels. Isotope profiling
most cancer cell lines compared to ME2 and ME3, although some analysis indicated that cellular metabolic balance is changed
cell lines predominantly expressed ME2 or ME3. Taken together, from glycolysis-dependent to glutaminolysis-dependent in
these findings suggest that ME1 has an important function in glucose-depleted conditions (Figure 5b). Malate (m+0) levels
the growth and survival of cancer cells and that it could be decreased in glucose-free conditions because of the reduction in
a target for cancer therapy. glucose-derived metabolites. Decrease in citrate (m+0) levels
The ME1 knockdown experiments described here revealed that supports this hypothesis. The increase in malate (m+3) levels in
ME1 depletion does not induce acute cell growth inhibition, glucose-free conditions may be due to activation of reductive
but suppresses cancer cell growth gradually by inducing carboxylation of glutamine. Reductive carboxylation produces
senescence in HCT116 and PC3 cell lines or apoptosis in H460 citrate (m+5) from glutamine (m+7) and results in the production
cell line. Partial cell growth inhibition was also observed in other of malate (m+3) through TCA cycle reactions. Increase of malate
cancer cell lines (Supplementary Table 1). We demonstrated that (m+3) results in increase of pyruvate (m+2) and lactate (m+2).
ME1 knockdown increased the expression of HMOX-1 (HO-1), Malate (m+4) levels are similar between normal glucose and
which is an oxidative stress marker regulated by Nrf2,18 glucose-free conditions. However, pyruvate (m+3) and citrate
suggesting that ME1 depletion disturbs metabolic and redox (m+6) were observed only in cells cultured in glucose-free media,
balance in cancer cells. We speculate that ME1 depletion suggesting that ME1 activity could be induced in cancer cells
suppresses cancer cell growth by abrogating metabolic and redox under glucose-depleted condition. ME1 expression did not
balance in the cancer cells, and the specific phenotype such as decreased in glucose-depleted conditions, although the levels
senescence, apoptosis, or others, depends on cellular context in of other NADPH-producing enzymes decreased (Figure 5c),
each cell line. We need further studies to identify cell fate which suggests that cancer cells depend in part on ME1 in
determinants in ME1-suppressed cancer cells, which could be glucose-depleted conditions. Here, we demonstrated that HCT116
sensitivity markers for a ME1 inhibitor in clinical settings. cells increase its dependency on ME1 metabolism in glucose-
L-glutamine isotope profiling analysis revealed the metabolic depleted conditions.
alterations in cancer cells by ME1 knockdown. Accumulation of From these results, we speculated that ME1 inhibition
malate (m+4) and fumarate (m+4) implies inhibition of the malate would synergistically inhibit cancer cell growth in glucose-
flux by ME1 knockdown. Reduction in citrate (m+2) and citrate depleted conditions. To address the hypothesis, we established

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
8
ME1 shRNA-inducible HCT116 cell lines, which showed similar (Cat. No. 22400-089, GIBCO/ThermoFisher Scientific, Waltham, MA, USA)
metabolic changes to HCT116 cells transfected with ME1 siRNA with 10% FBS (Hyclone/GE healthcare, Chicago, IL, USA). Glucose-free or
under ME1 shRNA induction in glucose-depleted conditions glutamine-free RPMI1640 medium (Cat. No. 11879020, Cat. No. 42401-018,
(Figure 6d). Expression of ME1 mRNA in sh ME1 clones were GIBCO/ThermoFisher Scientific) and Glucose- and glutamine-free
RPMI1640 medium (obtained based on the components of RPMI1640 by
lower than that of sh control clones, which may be due to leak
Funakoshi, Tokyo, Japan) were used for isotope profiling analysis and
of ME1 shRNA in the clones. These results may also explain growth assays. Tetracycline-free FBS (Clonetech/Takara Bio USA, Inc.,
the higher glucose consumption and higher lactate production Mountain view, CA, USA) were used with shRNA-expressing cell clones, and
in ME1 shRNA clones than those in sh control clones (Figures 6a dialysed FBS (Cat. No. 26400-044, GIBCO/ThermoFisher Scientific) were
and c). We demonstrated that combining ME1 inhibition with used for isotope profiling analysis. H460 cells were cultured in DMEM
glucose depletion synergistically inhibited cancer cell growth. medium (Cat. No. 11965-092, GIBCO/ThermoFisher Scientific).
We did not observe any cells with senescence-like or apoptotic
phenotypes in these cells. Senescence-like phenotype similar to siRNAs, shRNAs and reagents
that shown in Figure 2d could be observed if shRNA-expressing SilencerSelect siRNAs, SilencerSelect Negative Control No. 2 (439080846),
cells were cultured for a much longer period in normal KIF11 (s7903), ME1 (s8637, s8639), and ME2 (s8640, s8642) were purchased
media. However, cells died earlier in glucose-depleted than in from Ambion/ThermoFisher Scientific (Waltham, MA, USA). C911 siRNA was
glucose-sufficient conditions, suggesting that another cell killing ordered and synthesised by Ambion. The C911 siRNA sense and antisense
mechanism was activated by combining ME1 inhibition and sequences are 50 -CCAGGUUCAAUGAGUAGUAtt-30 and 50 -UACUACUCAU-
glucose depletion. The oxidative state, indicated by a reduction UGAACCUGGat-30 , respectively. Doxycycline hyclate was purchased from
in NADPH levels, may result in cell growth inhibition. Additional Sigma-Aldrich (Cat. No. D9891-5G, St Louis, MO, USA) and crystal violet
studies such as rescue of growth inhibition by antioxidants from Wako Chemicals GmbH (Cat. No. 038-04862, Neuss, Germany).
pTRIPZ-inducible lentiviral shRNAs for non-silencing control and ME1 were
would test this hypothesis.
purchased from Open Biosystems/GE healthcare, (Cat. No. RHS4743, Clone
An in vivo efficacy study is needed to verify whether IDV2THS_151673, V3THS_322622, Chicago, IL, USA).
ME1 inhibition is effective as a tumour growth suppressor. It is
known that glucose levels are depleted within the tumour,
although the tumour-induced angiogenesis, which provides siRNA transfection
nutrients (including glucose) to support rapid growth, may lead For siRNA experiments, cells were reversely transfected with siRNAs using
to higher levels of intra-tumour glucose than expected. Therefore, Lipofectamine RNAiMAX transfection reagent (Invitrogen, Waltham, MA,
USA) according to the manufacturer's recommendations. Individual siRNAs
the efficacy of ME1 inhibition on tumour growth suppression
were transfected at a final concentration of 10 nM.
probably depends on the tumour microenvironment, and in vivo
experiments are necessary to clarify the issue. We hypothesise
that a treatment strategy combining an ME1 inhibitor with Clonogenic assay
an inhibitor of glycolysis, such as 2-DG, or with an inhibitor of HCT116 cells were seeded in six-well plates at 1000 cells/well and were
glucose transporter, such as GLUT-4 inhibitor, could result in reversely transfected with siRNAs. After siRNA transfection, cells were
a synergic action against the tumour regardless of tumour cultured for 11 days with medium changed every 3 days. Colonies were
washed twice with phosphate-buffered saline, fixed with 4% formaldehyde
microenvironment. for 15 min, and stained with crystal violet. Colony area was quantified by
ME1 inhibition may be toxic in some tissues, as ME1 is GelCount Tumour Colony Counter (Oxford Optronix, Abingdon, UK).
ubiquitously expressed. Tumours sensitive to ME1 inhibition
should therefore be selected. For example, ME1 expression is
known to be regulated by Nrf2, and tumours with mutated Cell proliferation, cell count, caspase-3/7 and senescence analyses
Keap1 or Nrf2 genes could be dependent on ME1 and thus be Cell viability was determined using the CellTiter-Glo luminescence assay
sensitive to ME1 inhibition. Alternatively, as described above, (Promega, Madison, WI, USA) according to the manufacturer's suggested
protocol. Luminescent signals were detected using an ARVO MX1420
a glucose-depleted microenvironment could make tumours microplate reader (PerkinElmer, Wellesley, MA, USA). Cell numbers were
sensitive to ME1 inhibition. determined using a TC10 automated cell counter (Bio-Rad Laboratories,
In this study, we demonstrated that ME1 is an essential enzyme Hercules, CA, USA). In general, cancer cells were reversely transfected with
that regulates cellular metabolism and redox balance. We also siRNA and cell proliferation or cell numbers were determined 96 h after
showed that, in glucose-depleted conditions, cancer cells become siRNA transfection. Caspase-3/7 activity was determined using the
dependent on the ME1 flux to produce NADPH and pyruvate and Caspase-Glo 3/7 Assay (Promega). Cellular senescence was assessed
to manage redox homeostasis, suggesting that these cells become using the SA-b-Gal Staining kit from Cell Biolabs (Cat. No. CBA-230,
vulnerable to reduced ME1 activity. Taken together, our results San Diego, CA, USA).
demonstrate that ME1 inhibition suppresses cancer cell growth
and induces apoptosis or senescence depending on the cellular Glucose and lactate assays
context, consistent with previous reports,7,13,14 and that tumours Glucose and L-lactate concentrations in the medium were determined
in a nutrient-limited microenvironment are a sensitive target for using LabAssay Glucose (Wako Chemicals GmbH) and L-Lactate Assay Kit
ME1 inhibition. No ME1 inhibitors are currently being evaluated in (Colorimetric) (Abcam, Cambridge, UK), respectively.
clinical trials, although a few small-molecule ME2 inhibitors
were discovered in pre-clinical trials.21–23 A treatment strategy Establishment of stable pTRIPZ ME1 cell lines
combining ME1 inhibitors with inhibition of glycolysis, NADPH- HCT116 cells were transfected with pTRIPZ vectors and selected with
producing enzymes, or redox-regulating enzymes would provide 1 μg/ml puromycin 48 h after transfection. Cells were cultured in medium
an effective therapeutic option for some types of cancer. Further with 1 μg/ml puromycin for 14 days and grown colonies were picked up.
work is required in animal models to assess if such a strategy Then, stably shRNA-expressing HCT116 cell clones were established.
would be effective in tumours.
Metabolomics and isotope profiling analyses
MATERIALS AND METHODS After siRNA or shRNA treatment, 0.5–1.0 × 106 cells were suspended in 1 ml
of 100% methanol and centrifuged for 5 min at 15 000 g at 4 °C. The
Cell lines and culture supernatant (200 μl) of cell extracts was dried by nitrogen stream and then
Human cancer cell lines were purchased from American Type derivatised by a two-step reaction: oximation and trimethylsilylation.
Culture Collection (Manassas, VA, USA). HCT116 cells, HCT116 shRNA- The reaction mixture (1 μl) was injected into an Agilent 7890 A/5975C gas
expressing stable cell clones and PC3 cells were cultured in RPMI1640 chromatography-mass spectrometer (GC/MS) (Agilent Technologies,

Oncogenesis (2017), 1 – 9
 Role of malic enzyme 1 in cancer cells
S Murai et al
9
Santa Clara, CA, USA) equipped with a J&W Scientific HP-5MS column 11 Wen D, Liu D, Tang J, Dong L, Liu Y, Tao Z et al. Malic enzyme 1 induces epithelial-
(30 m × 0.25 mm i.d., film thickness = 0.25 μm, Agilent Technologies), and mesenchymal transition and indicates poor prognosis in hepatocellular
molecules were separated in a temperature gradient from 60 °C to 325 °C carcinoma. Tumour Biol 2015; 36: 6211–6221.
at 10 °C/min. The GC/MS data were analysed using ChemStation software 12 Chakrabarti G. Mutant KRAS associated malic enzyme 1 expression is a predictive
(Agilent Technologies), and the peak areas of target molecules were marker for radiation therapy response in non-small cell lung cancer. Radiat Oncol
exported to a spreadsheet for further analysis. For stable isotope tracer 2015; 10: 145–152.
analysis, the abundances of mass isotopologues were quantified by 13 Son J, Lyssiotis Ca, Ying H, Wang X, Hua S, Ligorio M et al. Glutamine supports
subtracting the abundance of natural isotopes from the raw mass pancreatic cancer growth through a KRAS-regulated metabolic pathway. Nature
spectra.24 2013; 496: 101–105.
14 Zheng FJ, Ye H, Bin, Wu MS, Lian YF, Qian CN, Zeng YX. Repressing malic enzyme
1 redirects glucose metabolism, unbalances the redox state, and attenuates
Statistical analyses migratory and invasive abilities in nasopharyngeal carcinoma cell lines. Chin J
All data represented the results from three independent experiments. The Cancer 2012; 31: 519–531.
statistical differences between groups were determined using a two-tailed 15 Rodier F, Campisi J. Four faces of cellular senescence. J. Cell Biol. 2011; 192:
Student’s t-test. 547–556.
16 Buehler E, Chen YC, Martin S. C911: a bench-level control for sequence specific
siRNA off-target effects. PLoS ONE 2012; 7: e51942.
CONFLICT OF INTEREST 17 Fan J, Ye J, Kamphorst JJ, Shlomi T, Thompson CB, Rabinowitz JD. Quantitative
The authors declare no conflict of interest. flux analysis reveals folate-dependent NADPH production. Nature 2014; 510:
298–302.
18 Furfaro AL, Traverso N, Domenicotti C, Piras S, Moretta L, Marinari UM et al. The
ACKNOWLEDGEMENTS Nrf2/HO-1 axis in cancer cell growth and chemoresistance. Oxid Med Cell Longev
2016; 2016: 1–14.
We thank Yuichi Hikichi, Hiroshi Miyake and Christopher Claiborne for their guidance
19 Tejero-Taldo MI, Caffrey JL, Sun J, Mallet RT. Antioxidant properties of pyruvate
during the course of this work and the Takeda Pharmaceutical Company for financial
mediate its potentiation of [beta] -adrenergic inotropism in stunned myocardium.
support.
J Mol Cell Cardiol 1999; 31: 1863–1872.
20 Dobsak P, Courderot-Masuyer C, Zeller M, Vergely C, Laubriet A, Assem M et al.
Antioxidative properties of pyruvate and protection of the ischemic rat heart
REFERENCES during cardioplegia. J Cardiovasc Pharmacol 1999; 34: 651–659.
1 Zhang J, Yang PL, Gray NS. Targeting cancer with small molecule kinase inhibitors. 21 Hsieh J-Y, Li S-Y, Tsai W-C, Liu J-H, Lin C-L, Liu G-Y et al. A small-molecule
Nat Rev Cancer 2009; 9: 28–39. inhibitor suppresses the tumor-associated mitochondrial NAD(P)+-dependent
2 Cairns R, Harris I, Mak T. Regulation of cancer cell metabolism. Nat Rev Cancer malic enzyme (ME2) and induces cellular senescence. Oncotarget 2015; 6:
2011; 11: 85–95. 20084–20098.
3 Vander Heiden MG. Targeting cancer metabolism: a therapeutic window opens. 22 Hsieh JY, Liu JH, Yang PC, Lin CL, Liu GY, Hung HC. Fumarate analogs act as
Nat Rev Drug Discov 2011; 10: 671–684. allosteric inhibitors of the human mitochondrial NAD(P)+-dependent
4 Pavlova NN, Thompson CB. The emerging hallmarks of cancer metabolism. Cell malic enzyme. PLoS ONE 2014; 9: e98385.
Metab 2016; 23: 27–47. 23 Wen Y, Xu L, Chen F, Gao J, Li J, Hu L et al. Discovery of a novel inhibitor of NAD(P)
5 Lin R, Elf S, Shan C, Kang H-B, Ji Q, Zhou L et al. 6-Phosphogluconate +-dependent malic enzyme (ME2) by high-throughput screening. Acta Pharmacol
dehydrogenase links oxidative PPP, lipogenesis and tumour growth by inhibiting Sin 2014; 35: 674–684.
LKB1-AMPK signalling. Nat Cell Biol 2015; 17: 1484–1496. 24 Fan TWM, Lane AN, Higashi RM, Farag MA, Gao H, Bousamra M et al. Altered
6 Dang L, Yen K, Attar EC. IDH mutations in cancer and progress toward develop- regulation of metabolic pathways in human lung cancer discerned by (13)C stable
ment of targeted therapeutics. Ann. Oncol. 2016; 27: 599–608. isotope-resolved metabolomics (SIRM). Mol Cancer 2009; 8: 41.
7 Jiang P, Du W, Mancuso A, Wellen KE, Yang X. Reciprocal regulation of p53 and
malic enzymes modulates metabolism and senescence. Nature 2013; 493:
689–693. Oncogenesis is an open-access journal published by Nature Publishing
8 Loeber G, Dworkin MB, Infante A, Ahorn H. Characterization of cytosolic malic Group. This work is licensed under a Creative Commons Attribution 4.0
enzyme in human tumor cells. FEBS Lett 1994; 344: 181–186. International License. The images or other third party material in this article are included
9 Pongratz RL, Kibbey RG, Shulman GI, Cline GW. Cytosolic and mitochondrial malic in the article’s Creative Commons license, unless indicated otherwise in the credit line; if
enzyme isoforms differentially control insulin secretion. J Biol Chem 2007; 282: the material is not included under the Creative Commons license, users will need to
200–207. obtain permission from the license holder to reproduce the material. To view a copy of
10 DeBerardinis RJ, Mancuso A, Daikhin E, Nissim I, Yudkoff M, Wehrli S et al. Beyond this license, visit [Link]
aerobic glycolysis: transformed cells can engage in glutamine metabolism that
exceeds the requirement for protein and nucleotide synthesis. Proc Natl Acad Sci
USA 2007; 104: 19345–19350. © The Author(s) 2017

Supplementary Information accompanies this paper on the Oncogenesis website ([Link]

Oncogenesis (2017), 1 – 9