Académique Documents
Professionnel Documents
Culture Documents
SUMMARY
Identification of the factors critical to the tumor-initiating cell (TIC) state may open new avenues in cancer
therapy. Here we show that the metabolic enzyme
glycine decarboxylase (GLDC) is critical for TICs in
non-small cell lung cancer (NSCLC). TICs from primary NSCLC tumors express high levels of the oncogenic stem cell factor LIN28B and GLDC, which are
both required for TIC growth and tumorigenesis.
Overexpression of GLDC and other glycine/serine
enzymes, but not catalytically inactive GLDC, promotes cellular transformation and tumorigenesis.
We found that GLDC induces dramatic changes in
glycolysis and glycine/serine metabolism, leading
to changes in pyrimidine metabolism to regulate
cancer cell proliferation. In the clinic, aberrant activation of GLDC correlates with poorer survival in lung
2007). These findings underscore the urgent need for both combination therapies and also new approaches to treat cancerous
tumors. One such approach may be to target tumor-initiating
cells (TICs).
Data from leukemias, germ cell tumors, and a number of solid
tumors support the notion that cancers are maintained by
a subpopulation of self-renewing and evolving TICs. This is
also popularly known as the cancer stem cell (CSC) model
(Reya et al., 2001; Rosen and Jordan, 2009). Although the validity
of the CSC model is an issue of controversy in melanoma (Quintana et al., 2008, 2010; Boiko et al., 2010; Civenni et al., 2011),
many other solid tumors appear to follow the CSC model (Ishizawa et al., 2010). Recently it was proposed that at earlier stages
of tumorigenesis, rare TIC clones differentiate into nonmalignant
progeny to form the bulk of the tumor, whereas at advanced
stages, TIC clones constitute the bulk of the tumor (Boiko
et al., 2010). Studies with mouse models of lung cancer have
also begun to reconcile the connection between the evolving
genotype of TIC clones and the surface phenotype of TICs (Curtis et al., 2010). Thus accumulated findings suggest that targeting TICs may be a promising approach for eradicating tumors
early. However, progress in the targeting of TICs to improve
cancer therapy has been hindered by a lack of understanding
of the molecular pathways that are critical to TICs.
Recent studies have led to an emerging appreciation of the
importance of metabolic reprogramming in cancer (Hsu and Sabatini, 2008; Vander Heiden et al., 2009; Hanahan and Weinberg,
2011). Most recently, the embryonic isoform of pyruvate kinase
PKM2, in collaboration with phosphoglycerate mutase, was
found to regulate the shift from oxidative phosphorylation to glycolysis in cancer cells (Christofk et al., 2008; Vander
Heiden et al., 2010). These findings have led to a resurgence of
interest in the Warburg effectthe phenomenon whereby cancer
cells, like embryonic cells, preferentially use glycolysis even
under aerobic conditions (Warburg, 1956). Besides glycolysis,
an arm of metabolism that results in sarcosine production has
also been implicated in prostate cancer (Sreekumar et al.,
2009). These data suggest that metabolic reprogramming is
crucial for tumorigenesis, and much remains to be uncovered.
Here we show that glycine metabolism and the metabolic
enzyme glycine decarboxylase (GLDC) drive TICs and tumorigenesis in NSCLC. Using CD166 as a surface marker and
NOD/SCID Il2rg/ mice as xenotransplantation recipients, we
isolated lung TICs from a broad range of primary NSCLC tumors
(stages IIII). Primary lung TICs express high levels of LIN28B,
GLDC, and many other glycine/serine metabolism enzymes.
Both LIN28B and GLDC were required for lung TIC proliferation
and tumor growth. Overexpression of GLDC alone, and other
glycine/serine enzymes, promotes cellular transformation both
in vitro and in vivo. Metabolomic analysis shows that GLDC overexpression induces dramatic changes in glycolysis and glycine
metabolism, leading to changes in pyrimidine metabolism for
cancer cell proliferation. In human patients, aberrant upregulation of GLDC is significantly associated with higher mortality
from lung cancer, and aberrant GLDC expression is observed
in multiple cancer types. Our findings establish a link between
glycine metabolism and tumorigenesis and may provide novel
targets for advancing anticancer therapy.
260 Cell 148, 259272, January 20, 2012 2012 Elsevier Inc.
RESULTS
TICs in Lung Cancer
To assess the cellular heterogeneity within NSCLC, we obtained
freshly resected lung tumors from 36 human patients with a
broad range of stage IIII primary NSCLC (Table S1 available
online). Patient lung cancer cells were directly transplanted
subcutaneously into NOD/SCID Il2rg/ mice with Matrigel
(Quintana et al., 2008). Using this maximally sensitive assay,
we estimated by limiting dilution analysis that lung TICs exist
with a low frequency of 1 in 4 3 105 cells in unsorted NSCLC
tumor cells (n = 36 patients; Figure 1A), consistent with published
findings (Ishizawa et al., 2010).
To profile the surface phenotype of this subpopulation of lung
TICs, we fractionated the NSCLC tumors by fluorescenceactivated cell sorting (FACS; Figure S1A). After excluding
hematopoietic and endothelial cells (Lin), we tested a panel of
cell-surface markers, including CD166, CD44, CD133, and
EpCAM (Figure 1B). We found that CD166 was the most robust
marker for enriching the lung TIC subpopulation, compared to
CD133, CD44, or EpCAM, allowing us to reliably enrich lung
TICs by nearly 100-fold (Figures 1A and 1B). In 12 out of 12
NSCLC patient tumors (lung adenocarcinoma), the CD166+
Lin fraction contained cells that consistently initiated lung tumor
formation in vivo. In contrast, CD166Lin tumor cells generally
failed to initiate lung tumor formation even after 8 months of
observation, although they also expressed carcinoembryonic
antigen (CEA), a tumor-specific marker not expressed in normal
adult lung cells (Figures 1A, 1B, and S1B). Similar results were
observed in lung squamous cell carcinoma and large cell carcinoma (Figure S1C). Although CD166 expression varied across
the NSCLC tumors we examined, CD166 was consistently
higher in lung tumors than in normal adjacent lung tissues
(n = 25 patients; Figures S1D and S1E).
CD166+ lung TICs demonstrate a capacity for self-renewal
and differentiation in vivo. Serial transplantations showed that
only the CD166+ fraction was able to self-renew and initiate primary and secondary xenograft tumors (Figures 1A and S1F).
Upon transplantation, CD166+ lung TICs differentiated to form
xenograft tumors that phenocopy the complex cytoarchitecture of their parental patient tumors, sharing similar histological morphology by hematoxylin-eosin (H&E) staining and
similar tissue distributions of CD166, cytokeratin, E-cadherin,
vimentin, smooth muscle actin, and synaptophysin (Figures
1C, S1G, and S1H). Furthermore, we found that transplants
with more TICs grow more rapidly, suggesting that lung TIC
frequency is correlated with tumor growth rate (Figures 1D
and S1I).
The self-renewal capacity of CD166+ lung TICs is further
corroborated by in vitro assays. We tested the CD166+ fraction
for the ability to form tumor spheres, a widely used in vitro technique for assessing self-renewal capacity (Dontu et al., 2003).
Although both primary CD166+ and CD166 cells remained
viable in vitro, only primary CD166+ but not CD166 cells were
able to form compact self-renewing spheres (n = 9 patients;
Figures 1E, 1F, and S1J). Using immunofluorescence and flow
cytometry, we found that the lung tumor spheres retained high
levels of CD166 expression but undetectable CD133 expression
150,000
200,000
300,000
500,000
1,000,000
2,000,000
2/8
2/3
7/9
2/2
4/4
1/1
210,000
260,000
500,000
0/1
1/3
0/3
CD166+Lin-
CD166-Lin-
2/4
1/10
4/5
3/5
9/9
2/2
1,000
5,000
10,000
23,000
25,000
50,000
Enrichment
factor (x)
1/403,971
(1/255,949 1/637,598)
1/5,175
(1/2,837 1/9,441)
78
1.37E-23
1/3,297,091
(1/460,464 1/23,608,384)
0.1
4.82E-03
No. tumors/
No. injections
12/12
0/10
6/8
6/8
6/11
7/13
10/13
8/14
Population
CD166+
CD166CD44+
CD44CD133+
CD133EpCAM+
EpCAM-
Xenograft tumor
Patient tumor
P value
3)
Tumor volume (mm
(
Unsorted
LTICs
frequency
(95% CI)
H&E
No. tumors
/ No.
injections
CD166
Cells
pan-CK
400
300
AdC
200
100
0
14
15
16
17
18
19
20
21
Weeks
CD166+Lin-
AdC
CD
D166-Lin-
LCC
G
120
90
60
30
0
AdC
SCC
LCC
S/XX+/X-
P+/PN+/N-
Candidate
genes
4
Log2 ratio
Log2 ratio
Enrichment
c e
X+
S
-4
-4
N+/X-
P+/X-
X+/X-
N+/X-
S/X-
P+/X-
X+/X_
S/X-
D
Percentage of genes in pathway
0%
TNFAIP3
IGLL1
2
1
HSPA1A
N+/NRRAD
P+/P-
CXCL2
GADD45G
X+/X-
PPP1R15A
RASD1
MMP9
S/X+
RENBP
FOXK2
REEP1
-2
2%
4%
6%
8%
10%
Cell cycle
DNA replication
G i serine
Glycine,
i and threonine
i
metabolism
SMC6
PEA3
-1
TMEM118
0
GLDC
NPAS1
LIN28B
RG9MTD2
CCNB1
Pyrimidine metabolism
MAPK signaling pathway
-3
Experimental
Predicted
-4
-5
E
3P-Hydroxy pyruvate
PSAT1
Phosphoserine
PSPH
Serine
SHMT1
Glycine
GLDC
GCAT
L-2-Amino-acetoacetate
Liu et al., 2003; Levesque et al., 2007), as well as cell-cycle regulators like CCNB1 and GADD45G (Figure 2C). The highestranking genes were validated by qRT-PCR (Figure S2A). KEGG
pathway analysis of the lung TIC-gene profile showed that the
top enriched pathways were cell cycle, DNA replication,
glycine, serine, and threonine metabolism, pyrimidine metabolism, MAPK signaling pathway, and p53 signaling pathway (Figure 2D). Within the glycine, serine, and threonine
metabolism pathway, we found that glycine/serine metabolism
enzymes like GLDC, glycine C-acetyltransferase (GCAT), serine
hydroxymethyltransferase (SHMT1), phosphoserine phosphatase (PSPH), and phosphoserine aminotransferase (PSAT1)
were all upregulated in lung TICs (Figures 2E and S2BS2D). In
particular, GLDC was one of the most highly upregulated genes
in multiple analyses of lung TIC-enriched populations, at both the
mRNA and protein levels (Figures 2C and S2C). GLDC is a key
component of the highly conserved glycine cleavage system
in amino acid metabolism that catalyzes the breakdown of
glycine to form CO2, NH3, and 5,10-methylene-tetrahydrofolate
(CH2-THF) to fuel one-carbon metabolism (Kume et al., 1991).
GLDC Is an Oncogene that Promotes Tumorigenesis
and Cellular Transformation
High expression of GLDC and LIN28B in lung TIC-enriched populations, but not in CD166 lung cancer cells and CD166+ normal
lung cells, suggests that these two genes drive tumorigenicity in
lung TICs. To test this hypothesis, we knocked down GLDC and
LIN28B in lung tumor spheres with shRNAs (Figure S3A) and
compared their growth both in vitro and in vivo. We found that
both GLDC and LIN28B were necessary for cell proliferation in
sphere culture, as well as anchorage-independent colony formation in soft agar (Figures 3A and S3B). Importantly, tumorigenicity
was also significantly reduced upon knockdown of either GLDC
or LIN28B (Figures 3B and S3C). A549 lung adenocarcinoma
cells showed similar results (Figures S3DS3G). Our results
suggest that lung TICs and lung tumorigenesis are dependent
on GLDC. This led us to ask what oncogenes upregulate
GLDC. Because the E2F pathway upregulates many metabolic
genes during cell proliferation, we examined the expression of
GLDC over the course of the cell cycle in both normal human
lung fibroblasts (HLFs) and the transformed A549 cells after
synchronization by serum starvation. Our results showed that
GLDC is insensitive to cell-cycle progression in both normal
HLFs and transformed A549 cells, suggesting that GLDC is not
regulated by cell-cycle or E2F signals (Figure 3C). We then
examined GLDC levels in MCF10A cells after transformation by
oncogenic KRASG12D, PIK3CAE545K, and MYCT58A. Our results
show that all three oncogenes induce GLDC by 20-fold, suggesting that oncogene-induced GLDC transcription is common
to the cellular transformation process mediated by oncogenic
Ras, PI3K, and Myc (Figure 3D).
To test whether aberrant GLDC upregulation is sufficient
to drive cellular transformation, as has been shown for LIN28B
(Viswanathan et al., 2009), we overexpressed GLDC in NIH/
3T3 cells (Figure S3H). We found that GLDC overexpression
significantly increased colony formation by 3T3 cells under normal culture conditions (Figures 3E and 3F). To test for cellular
transformation in vitro, we cultured the 3T3 cells overexpressing
Tumor spheres
Number of cells
80,000
60,000
40,000
20,000
488 h
4h
8h
244 h
0.6
4h
166 h
2h
0h
0.8
A549
TS-Ctrl-sh
TS-28B-sh
TS-GD-sh
100,000
HLF
Ctrl-sh
GD-sh
28B h
28B-sh
cyyc
120,000
244 h
GLDC
GLDC
FOS
CDK1
CDK1
HSP90
*
0.4
**
0.2
HSP90
0
0
Days
MCF10A
3T3
30
100
3T3-GD
10
3T3-28B
**
**
75
50
25
3T3
3T3-GD
3T3-28B
80
No. of agar colonies / 55,000
3T3
20
3T3
3T3-GD
3T3-28B
125
No. of adherent colon
nies /
500 3T3
3T3
70
60
**
**
50
40
30
20
10
HLF
40,000
HLF-28B
20,000
0
0
140
No. of adherent colonies /
1,000 HL
LF
HLF
60,000
HL
LF-GD
Numbeer of cells
HLF
HLF-28B
120
100
80
60
40
20
0
Days
**
HLF-GD
80,000
K
HLF
HLF-GD
HLF-28B
**
100,000
35
No. of agar colon
nies / 5,000
HLF
30
25
HLF
HLF-GD
HLF-28B
**
**
20
15
10
5
0
g
Xenograft
L
Group
No. cells /
injection
No. tumors/
No. injections
200,000
0/12
CD166100,000
0/12
200,000
4/12
100,000
,
1/12
CD166- GD
Figure 3. GLDC and LIN28B Are Necessary and Sufficient for Malignant Growth
(A) Proliferation curve of tumor sphere (TS) cells with shRNA knockdown of either GLDC (GD-sh) or LIN28B (28B-sh).
(B) Quantitative mass analysis of xenograft tumors formed 30 days after transplanting 100,000 tumor sphere cells with either GLDC knockdown (GD-sh) or
LIN28B knockdown (28B-sh).
(C) Western blot analysis of endogenous GLDC during the cell cycle in synchronized normal human HLFs and transformed A549 cells. HLF or A549 cells were
serum-starved for 72 hr followed by release into serum-containing medium with samples collected at indicated time points. Expression of GLDC, FOS (early
264 Cell 148, 259272, January 20, 2012 2012 Elsevier Inc.
K754A
P769L
G771R
Vector
8/
8
0/
8
0/
8
0/
8
0/
8
0/
8
A549
H753P
WT
GLDC
No. tumors/
No. injections
R
Relative
RNA expression level
8/
8
100,000
10,000
1,000
100
10
1
Group
3T3-GLDC
3T3-PSAT1
3T3-PSPH
3T3-SHMT2
3T3-SHMT1
3T3-GCAT
3T3
No. tumors/
No. injections
6/6
3/6
2/6
1/6
0/6
0/6
0/6
GLDC-GFP
GLDC
-Actin
serum response), and CDK1 (E2F target) were tested. Normal growing, unsynchronized cells (Cyc) were used as a control. HSP90 was used as a loading control.
CDK1, cyclin-dependent kinase 1; HSP90, heat shock protein 90.
(D) Expression of endogenous GLDC in MCF10A cells transformed by oncogenic KRASG12D, PIK3CAE545K, MYC, or MYCT58A, by qRT-PCR.
(E) Colony formation assay in adherent conditions by seeding 500 3T3 cells overexpressing either GLDC (3T3-GD) or LIN28B (3T3-28B).
(F) Quantitative analysis of colony formation efficiency under adherent conditions as shown in (E).
(G) Quantitative analysis of soft agar colony formation by 5,000 3T3 cells overexpressing either GLDC (3T3-GD) or LIN28B (3T3-28B). Colonies were stained with
INT on day 28.
(H) Proliferation curve of HLF cells overexpressing GLDC (HLF-GD), LIN28B (HLF-28B), or the empty vector (HLF).
(I) Colony formation assay in adherent conditions seeding 1,000 HLF cells overexpressing GLDC (HLF-GD), LIN28B (HLF-28B), or the empty vector (HLF).
(J) Quantitative analysis of colony formation efficiency under adherent conditions as shown in (I).
(K) Quantitative analysis of soft agar colony formation by 5,000 HLF cells overexpressing either GLDC (HLF-GD), LIN28B (HLF-28B), or the empty vector (HLF).
(L) Tumor formation by CD166 lung tumor cells 3 months after overexpression of GLDC. CD166 tumor cells from xenografts were sorted by FACS and infected
with retrovirus expressing either the empty vector (CD166) or GLDC (CD166 GD), followed by transplantation into mice 24 hr after infection (n = 12 for each
group).
In all panels, error bars represent SEM. *p < 0.05, **p < 0.01. See also Figure S3.
Cell 148, 259272, January 20, 2012 2012 Elsevier Inc. 265
and HLF cells, with little effect on control 3T3 and HLF cells
(Figure 5E). Furthermore, methotrexate in combination with
GLDC shRNA killed transformed A549 cells much more effectively than either alone (Figure 5E), suggesting that a combination
of antifolates with a GLDC inhibitor could completely shut off
glycine catabolism to treat cancer cells more effectively. Using
these metabolic data, we constructed a model of how aberrant
GLDC expression might reprogram glycolysis and glycine
metabolism fluxes in cancer cells to promote cancer cell proliferation and tumorigenesis (Figure 5G).
Prognostic Significance of Aberrant GLDC Expression
in NSCLC Patients
To assess whether our experimental findings on GLDC are relevant to human lung cancer patients in the clinic, we used tissue
microarray immunohistochemistry to examine the prognostic
significance of GLDC expression, tumor size, tumor grade, and
cancer stage in clinical tumor samples from cohorts of NSCLC
patients (n = 143) (Figures 6A and S6; Table S2). Subdistribution
hazard ratio (SHR) analysis showed that patients with high or
grade 3+ GLDC expression have a 3-fold higher risk of lung
cancer mortality compared to patients with low or grade 0
GLDC expression, even when adjusted for cancer stage (SHR =
3.01, 95% confidence interval [CI]: 1.486.10, p = 0.002) (Figure 6B). Cumulative mortality analysis also showed that high
GLDC expression (grade 3+) is significantly associated with
higher cumulative incidence of mortality across 143 NSCLC
lung cancer patients, even when adjusted for cancer stage
(p = 0.005) (Figure 6C). CD166 expression was not significantly
associated with higher mortality in lung cancer patientswhich
was not unexpected given that only 1 in 5 3 103 CD166+ cells
are tumorigenic (Figures S6AS6C). Indeed, coimmunostaining
of lung tumors revealed that GLDC+ cells mostly form a subset
of CD166+ cells, and that not all CD166+ cells are GLDC+ (Figures
6D and S6D). LIN28B immunohistochemistry was also not significantly correlated with lung cancer patient mortality (Figures
S6AS6C), although western blots revealed that lung TICs
express a second LIN28B isoform that is indiscernible from
immunohistochemistry staining, thus rendering our LIN28B staining results inconclusive (Figure S6E). Our immunohistochemistry
results on clinical tumor samples are consistent with the idea that
lung TICs constitute the bulk of the tumors in late stages of
malignancy and demonstrate that aberrant activation of GLDC
is significantly associated with human mortality in NSCLC
patientsfurther supporting its role as a metabolic oncogene in
human NSCLC.
Aberrant GLDC Expression in Other Cancers
To check whether aberrant GLDC expression is specific to
NSCLC, we examined a variety of other cancers. Surprisingly
GLDC is also aberrantly upregulated in subsets of primary
tumors from other cancers, especially ovarian and germ cell
tumors (Figure 7A; Table S3). Further analysis of 606 human
cancer cell lines showed that 158 (26.1%) cancer cell lines overexpress GLDC, including lines from ovarian, germ cell, cervical,
lung, lymphoma, prostate, bladder, and colon cancer (Figure 7B;
Table S4). To test whether GLDC is also required for growth by
one of these GLDC-overexpressing cell lines, we knocked
B
4.00
4.00
1
2
3
4
1.00
1
0.50
Glycine-related metabolites
S
Sarcosine
i
Betaine aldehyde
Serine
Glycine
2.00
Fold change (Log2 scale)
Cells
Cells
0.25
0.13
3T3-GD/Ctrl
0.06
1
2
3
2 00
2.00
1.00
1
0.50
3T3-GD/Ctrl
HLF-GD/Ctrl
HLF-GD/Ctrl
0.03
Glycolysis
Glucose 11-phosphate
phosphate
Phosphoenolpyruvate
Pyruvate
A549-GD-sh/Ctrl
A549-GD-sh/Ctrl
0.25
0 02
0.02
2.5
2.5
3T3
A549-GD-sh
2.0
1.5
1.5
1.0
1.0
0.5
0.5
0.0
0.0
Lacta
ate OD
2.0
4.00
A549
3T3-GD
2.00
1
2
3
4
5
1 00
1.00
1
0.50
Pyrimidines
Thymidine
Deoxyuridine
Thymine
y
Cytosine
Uracil
3T3-GD/Ctrl
HLF-GD/Ctrl
24 Hr
A549-GD-sh/Ctrl
0
0.25
24 Hr
120
3T3
Agar colony number
F
60
40
20
0
0 0.1 0.2 0.5 1
5 10 20
Control
5 10 20
GLDC
100
0 uM Sarcosine
10 uM Sarcosine
80
60
40
20
0
Methotrexate ( M)
20
HLF
15
Glycolysis
10
5
Phosphoserine
0
0 0.1 0.2 0.5 1
10
Control
Methotrexate
10
GLDC
Serine
Tetrahydrofolate (THF)
Methotrexate ( M)
100
Betaine
Aldehyde
A549
Sarcosine
Glycine
GLDC
CH2-THF
C
Pyrimidine
metabolism
t b li
CO2
NH3
NADH
50
0
0
10 20 50
Control
10 20 50
GLDC shRNA1
Methotrexate ( M)
10 20 50
GLDC shRNA2
Cell 148, 259272, January 20, 2012 2012 Elsevier Inc. 267
GLDC
1+
2+
3+
AdC
C
C
SHR (95% CI)
P-value
0.005
0.514
0.700
0.002
GLDC
3+
0.8
Cumula
ative incidence
GLDC Intensity
1.0
1+
2+
0.6
04
0.4
P = 0.005
0.2
n=143
0
0
10
12
14
D
CD166
DAPI
GLDC
CD166
GLDC
DAPI
268 Cell 148, 259272, January 20, 2012 2012 Elsevier Inc.
3
2
1
Colon
Bladder
Esophagus
Prostate
-3
Lymphoma
Cervix
-2
Germ cell
-1
Brain
0
Ovary
Lung
-4
10000
Normalized GLDC intensity
1000
250
100
10
1
CACO2
C
Ctrl-sh
GD-sh
0.5
00
0.0
Relative LIN28B
expression
Ctrl-sh
1.0
0.5
Ctrl-sh
28B-sh
GD-sh
28B-sh
Ctrl-sh
GD-sh
28B-sh
800,000
Number of cells
Relative GLDC
expression
D
1.0
600,000
400,000
200,000
0
0
0.0
4
Days
down GLDC in CACO2 cells. Indeed we found that GLDC knockdown reduced their proliferation and tumorigenic potential upon
transplantation, suggesting that GLDC may act as an oncogene
in other cancer cells as well (Figures 7C7E). To examine the
possibility that GLDC is a housekeeping gene for cell proliferation, we also knocked down GLDC in normal HLFs (Figure S7A).
We found that HLF proliferation was unaffected by retroviral
knockdown of GLDC (Figures S7B and S7C). Furthermore we
observed that GLDC is highly expressed only in a few normal
tissues, including postmitotic liver cells, kidney cells, placenta
cells, and olfactory bulb neurons (Figure S7D). Altogether our
observations in both experimental and clinical settings suggest
Flow Cytometry
A list of antibodies used can be found in Table S5. Cells were FACS-sorted
using a FACSAria (BD). Flow cytometry was performed using a LSR II flow
cytometer, and data were analyzed with CELLQuest Pro software (BD).
Animals and Transplantation of Tumor Cells
NOD.Cg-Prkdcscid Il2rgtm1Wjl/SzJ mice (Jackson Laboratories) at 46 weeks
old were subcutaneously transplanted with single-cell suspensions in
serum-free medium and Matrigel (BD) (1:1).
Tumor Sphere Culture
Cells were grown in DMEM/F12 containing ITS (BD Biosciences) and supplemented with 50 ng/ml EGF and 20 ng/ml basic fibroblast growth factor (bFGF)
(Invitrogen), using nontreated cell culture plates (Nunc). Fresh medium was
replenished every 3 days.
(E.H.L.). We gratefully acknowledge assistance from Y.P.R. Yip, S.L. Long, and
J. Xu for collection of lung tissues. We are grateful to the Biopolis Shared Facilities Histopathology Laboratory staff for their support with immunohistochemistry and image analysis. We thank Z.M. Li for assistance with making transformed breast cell lines. We greatly acknowledge V. Gaddemane, C.S. Gan,
and T. Hennessy from Agilent Technologies, Singapore for their support in
acquiring and analyzing the mass spectrometry data for the differential analysis of the metabolites. We thank W.L. Tam and F. McKeon for comments
on the manuscript.
Received: February 22, 2011
Revised: August 11, 2011
Accepted: November 17, 2011
Published online: January 5, 2012
REFERENCES
Avril-Delplanque, A., Casal, I., Castillon, N., Hinnrasky, J., Puchelle, E., and
Peault, B. (2005). Aquaporin-3 expression in human fetal airway epithelial
progenitor cells. Stem Cells 23, 9921001.
Bester, A.C., Roniger, M., Oren, Y.S., Im, M.M., Sarni, D., Chaoat, M., Bensimon, A., Zamir, G., Shewach, D.S., and Kerem, B. (2011). Nucleotide deficiency promotes genomic instability in early stages of cancer development.
Cell 145, 435446.
Boiko, A.D., Razorenova, O.V., van de Rijn, M., Swetter, S.M., Johnson, D.L.,
Ly, D.P., Butler, P.D., Yang, G.P., Joshua, B., Kaplan, M.J., et al. (2010).
Human melanoma-initiating cells express neural crest nerve growth factor
receptor CD271. Nature 466, 133137.
Christofk, H.R., Vander Heiden, M.G., Harris, M.H., Ramanathan, A., Gerszten,
R.E., Wei, R., Fleming, M.D., Schreiber, S.L., and Cantley, L.C. (2008). The M2
splice isoform of pyruvate kinase is important for cancer metabolism and
tumour growth. Nature 452, 230233.
Chua, S.W., Vijayakumar, P., Nissom, P.M., Yam, C.Y., Wong, V.V.T., and
Yang, H. (2006). A novel normalization method for effective removal of systematic variation in microarray data. Nucleic Acids Res. 34, e38.
Civenni, G., Walter, A., Kobert, N., Mihic-Probst, D., Zipser, M., Belloni, B.,
Seifert, B., Moch, H., Dummer, R., van den Broek, M., and Sommer, L.
(2011). Human CD271-positive melanoma stem cells associated with metastasis establish tumor heterogeneity and long-term growth. Cancer Res. 71,
30983109.
Curtis, S.J., Sinkevicius, K.W., Li, D.A., Lau, A.N., Roach, R.R., Zamponi, R.,
Woolfenden, A.E., Kirsch, D.G., Wong, K.K., and Kim, C.F. (2010). Primary
tumor genotype is an important determinant in identification of lung cancer
propagating cells. Cell Stem Cell 7, 127133.
Dang, L., White, D.W., Gross, S., Bennett, B.D., Bittinger, M.A., Driggers, E.M.,
Fantin, V.R., Jang, H.G., Jin, S., Keenan, M.C., et al. (2009). Cancer-associated
IDH1 mutations produce 2-hydroxyglutarate. Nature 462, 739744.
Dontu, G., Abdallah, W.M., Foley, J.M., Jackson, K.W., Clarke, M.F., Kawamura, M.J., and Wicha, M.S. (2003). In vitro propagation and transcriptional
profiling of human mammary stem/progenitor cells. Genes Dev. 17, 1253
1270.
SUPPLEMENTAL INFORMATION
Eramo, A., Lotti, F., Sette, G., Pilozzi, E., Biffoni, M., Di Virgilio, A., Conticello,
C., Ruco, L., Peschle, C., and De Maria, R. (2008). Identification and expansion
of the tumorigenic lung cancer stem cell population. Cell Death Differ. 15,
504514.
Fine, J.P., and Gray, R.J. (1999). A proportional hazards model for the subdistribution of a competing risk. J. Am. Stat. Assoc. 94, 496509.
ACKNOWLEDGMENTS
Hanahan, D., and Weinberg, R.A. (2011). Hallmarks of cancer: the next generation. Cell 144, 646674.
Hennrick, K.T., Keeton, A.G., Nanua, S., Kijek, T.G., Goldsmith, A.M., Sajjan,
U.S., Bentley, J.K., Lama, V.N., Moore, B.B., Schumacher, R.E., et al.
(2007). Lung cells from neonates show a mesenchymal stem cell phenotype.
Am. J. Respir. Crit. Care Med. 175, 11581164.
Cell 148, 259272, January 20, 2012 2012 Elsevier Inc. 271
Hsu, P.P., and Sabatini, D.M. (2008). Cancer cell metabolism: Warburg and
beyond. Cell 134, 703707.
Hu, Y.F., and Smyth, G.K. (2009). ELDA: extreme limiting dilution analysis for
comparing depleted and enriched populations in stem cell and other assays.
J. Immunol. Methods 347, 7078.
Huang, W., Sherman, B.T., and Lempicki, R.A. (2009). Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat.
Protoc. 4, 4457.
Ishizawa, K., Rasheed, Z.A., Karisch, R., Wang, Q.J., Kowalski, J., Susky, E.,
Pereira, K., Karamboulas, C., Moghal, N., Rajeshkumar, N.V., et al. (2010).
Tumor-initiating cells are rare in many human tumors. Cell Stem Cell 7,
279282.
Jemal, A., Bray, F., Center, M.M., Ferlay, J., Ward, E., and Forman, D. (2011).
Global cancer statistics. CA Cancer J. Clin. 61, 6990.
Kume, A., Koyata, H., Sakakibara, T., Ishiguro, Y., Kure, S., and Hiraga, K.
(1991). The glycine cleavage system. Molecular cloning of the chicken and
human glycine decarboxylase cDNAs and some characteristics involved in
the deduced protein structures. J. Biol. Chem. 266, 33233329.
Kure, S., Kato, K., Dinopoulos, A., Gail, C., DeGrauw, T.J., Christodoulou, J.,
Bzduch, V., Kalmanchey, R., Fekete, G., Trojovsky, A., et al. (2006). Comprehensive mutation analysis of GLDC, AMT, and GCSH in nonketotic hyperglycinemia. Hum. Mutat. 27, 343352.
Levesque, B.M., Zhou, S.T., Shan, L., Johnston, P., Kong, Y.P., Degan, S., and
Sunday, M.E. (2007). NPAS1 regulates branching morphogenesis in embryonic lung. Am. J. Respir. Cell Mol. Biol. 36, 427434.
Liu, Y.R., Jiang, H.Y., Crawford, H.C., and Hogan, B.L.M. (2003). Role for ETS
domain transcription factors Pea3/Erm in mouse lung development. Dev. Biol.
261, 1024.
Locasale, J.W., Grassian, A.R., Melman, T., Lyssiotis, C.A., Mattaini, K.R.,
Bass, A.J., Heffron, G., Metallo, C.M., Muranen, T., Sharfi, H., et al. (2011).
Phosphoglycerate dehydrogenase diverts glycolytic flux and contributes to
oncogenesis. Nat. Genet. 43, 869874.
Murakami, Y., Hirata, H., Miyamoto, Y., Nagahashi, A., Sawa, Y., Jakt, M., Asahara, T., and Kawamata, S. (2007). Isolation of cardiac cells from E8.5 yolk sac
by ALCAM (CD166) expression. Mech. Dev. 124, 830839.
Nakai, T., Nakagawa, N., Maoka, N., Masui, R., Kuramitsu, S., and Kamiya, N.
(2005). Structure of P-protein of the glycine cleavage system: implications for
nonketotic hyperglycinemia. EMBO J. 24, 15231536.
Parsons, D.W., Jones, S., Zhang, X.S., Lin, J.C.H., Leary, R.J., Angenendt, P.,
Mankoo, P., Carter, H., Siu, I.M., Gallia, G.L., et al. (2008). An integrated
genomic analysis of human glioblastoma multiforme. Science 321, 1807
1812.
Possemato, R., Marks, K.M., Shaul, Y.D., Pacold, M.E., Kim, D., Birsoy, K.,
Sethumadhavan, S., Woo, H.K., Jang, H.G., Jha, A.K., et al. (2011). Functional
genomics reveal that the serine synthesis pathway is essential in breast
cancer. Nature 476, 346350.
Quintana, E., Shackleton, M., Sabel, M.S., Fullen, D.R., Johnson, T.M., and
Morrison, S.J. (2008). Efficient tumour formation by single human melanoma
cells. Nature 456, 593598.
272 Cell 148, 259272, January 20, 2012 2012 Elsevier Inc.
Quintana, E., Shackleton, M., Foster, H.R., Fullen, D.R., Sabel, M.S., Johnson,
T.M., and Morrison, S.J. (2010). Phenotypic heterogeneity among tumorigenic
melanoma cells from patients that is reversible and not hierarchically organized. Cancer Cell 18, 510523.
Reya, T., Morrison, S.J., Clarke, M.F., and Weissman, I.L. (2001). Stem cells,
cancer, and cancer stem cells. Nature 414, 105111.
Rosen, J.M., and Jordan, C.T. (2009). The increasing complexity of the cancer
stem cell paradigm. Science 324, 16701673.
Sabatini, F., Petecchia, L., Tavian, M., Jodon de Villeroche, V., Rossi, G.A., and
Brouty-Boye, D. (2005). Human bronchial fibroblasts exhibit a mesenchymal
stem cell phenotype and multilineage differentiating potentialities. Lab. Invest.
85, 962971.
Sequist, L.V., Joshi, V.A., Janne, P.A., Muzikansky, A., Fidias, P., Meyerson,
M., Haber, D.A., Kucherlapati, R., Johnson, B.E., and Lynch, T.J. (2007).
Response to treatment and survival of patients with non-small cell lung cancer
undergoing somatic EGFR mutation testing. Oncologist 12, 9098.
Sreekumar, A., Poisson, L.M., Rajendiran, T.M., Khan, A.P., Cao, Q., Yu, J.D.,
Laxman, B., Mehra, R., Lonigro, R.J., Li, Y., et al. (2009). Metabolomic profiles
delineate potential role for sarcosine in prostate cancer progression. Nature
457, 910914.
Tibbetts, A.S., and Appling, D.R. (2010). Compartmentalization of mammalian
folate-mediated one-carbon metabolism. Annu. Rev. Nutr. 30, 5781.
Vander Heiden, M.G., Cantley, L.C., and Thompson, C.B. (2009). Understanding the Warburg effect: the metabolic requirements of cell proliferation.
Science 324, 10291033.
Vander Heiden, M.G., Locasale, J.W., Swanson, K.D., Sharfi, H., Heffron, G.J.,
Amador-Noguez, D., Christofk, H.R., Wagner, G., Rabinowitz, J.D., Asara,
J.M., and Cantley, L.C. (2010). Evidence for an alternative glycolytic pathway
in rapidly proliferating cells. Science 329, 14921499.
Viswanathan, S.R., Powers, J.T., Einhorn, W., Hoshida, Y., Ng, T.L., Toffanin,
S., OSullivan, M., Lu, J., Phillips, L.A., Lockhart, V.L., et al. (2009). Lin28
promotes transformation and is associated with advanced human malignancies. Nat. Genet. 41, 843848.
Vogelstein, B., and Kinzler, K.W. (2004). Cancer genes and the pathways they
control. Nat. Med. 10, 789799.
Warburg, O. (1956). On the origin of cancer cells. Science 123, 309314.
Yang, D.H., and Moss, E.G. (2003). Temporally regulated expression of Lin-28
in diverse tissues of the developing mouse. Gene Expr. Patterns 3, 719726.
Zhu, H., Shah, S., Shyh-Chang, N., Shinoda, G., Einhorn, W.S., Viswanathan,
S.R., Takeuchi, A., Grasemann, C., Rinn, J.L., Lopez, M.F., et al. (2010). Lin28a
transgenic mice manifest size and puberty phenotypes identified in human
genetic association studies. Nat. Genet. 42, 626630.
Zhu, H., Shyh-Chang, N., Segre`, A.V., Shinoda, G., Shah, S.P., Einhorn, W.S.,
Takeuchi, A., Engreitz, J.M., Hagan, J.P., Kharas, M.G., et al; DIAGRAM
Consortium; MAGIC Investigators. (2011). The Lin28/let-7 axis regulates
glucose metabolism. Cell 147, 8194.