met signaling regulates glioblastoma stem...

Tumor and Stem Cell Biology

MET Signaling Regulates Glioblastoma Stem Cells

Kyeung Min Joo1,3, Juyoun Jin1,2, Eunhee Kim5, Kang Ho Kim1,2, Yonghyun Kim1, Bong Gu Kang1,2,Youn-Jung Kang1,2, Justin D. Lathia5, Kwang Ho Cheong6, Paul H. Song6, Hyunggee Kim4,Ho Jun Seol1,2, Doo-Sik Kong2, Jung-Il Lee2, Jeremy N. Rich5, Jeongwu Lee5, and Do-Hyun Nam1,2

AbstractGlioblastomas multiforme (GBM) contain highly tumorigenic, self-renewing populations of stem/initiating

cells [glioblastoma stem cells (GSC)] that contribute to tumor propagation and treatment resistance. However,our knowledge of the specific signaling pathways that regulate GSCs is limited. TheMET tyrosine kinase is knownto stimulate the survival, proliferation, and invasion of various cancers including GBM. Here, we identified adistinct fraction of cells expressing a high level ofMET in human primaryGBM specimens thatwere preferentiallylocalized in perivascular regions of human GBM biopsy tissues and were found to be highly clonogenic,tumorigenic, and resistant to radiation. Inhibition of MET signaling in GSCs disrupted tumor growth andinvasiveness both in vitro and in vivo, suggesting that MET activation is required for GSCs. Together, our findingsindicate that MET activation in GBM is a functional requisite for the cancer stem cell phenotype and a promisingtherapeutic target. Cancer Res; 72(15); 3828–38. �2012 AACR.

IntroductionGlioblastoma multiforme (GBM) is the most common and

lethal primary brain tumor with a median survival of 14.6months despite maximal therapy (1). As included in classifi-cation criteria for GBMs, excessive and unstable blood vesselformation and tumor necrosis associated with hypoxia are keypathologic characteristics of GBMs (2). GBMs heavily infiltrateinto the neighboring brain parenchyma and are almost uni-formly resistant to standard therapeutic regimes such asirradiation and chemotherapy (3). These biologic character-istics of GBMs are major reasons of lethality and need to betargeted for therapy.

Cancer stem cell hypothesis posits that a subpopulationof cancer cells is highly enriched with tumorigenic potential(3–7). Comparedwith bulk tumor cells, glioblastoma stem cells

(GSC) survive better against irradiation and chemotherapies,thereby contributing to therapeutic resistance and tumorrecurrence (8, 9). In addition, GSCs frequently reside in peri-vascular and hypoxic regions, actively promoting angiogenesisand facilitating the survival in harsh environment (10–12).However, the underlyingmolecular pathways that govern theseprocesses in GSCs are poorly understood.

The MET receptor tyrosine kinase regulates cell growthand motility. Hepatocyte growth factor (HGF)/scatter factoris a cognate ligand for MET signaling (13, 14). MET pathwayactivation induces glioma cell proliferation, survival, andmigration (15, 16). We and others reported that MET over-expression is associated with poor prognosis and tumorinvasiveness in GBM patients (17, 18). Large-scale genomicstudies confirmed frequent MET pathway activation andgenomic amplification of MET in GBMs, indicating thataberrant activation of MET is an important genetic eventin GBMs (19–21).

As hypoxia directly induces MET expression and HGFinduces expression of VEGF in glioma cells, MET signalingmay be particularly important for tumor cell survival inhypoxia and co-option with angiogenesis (15, 22, 23). Inaddition, it was reported that MET signaling is a key mech-anism to maintain stem cell niche in brain (24). A recent studyreported that MET signaling enhances GSC populations, sug-gesting a link between MET signaling and GSCs (25). Theprecise roles of MET signaling operated in GSCs, however,remain unclear (25).

On the basis of this background, we hypothesized that METsignaling promotes self-renewal and therapeutic resistance ofGSCs. By extensive in vivo and in vitro studies using a largenumber of freshly isolated patient-derived GBM cells, weprovide evidence suggesting that MET signaling plays criticalroles in GSC maintenance, migration, and resistance toradiation.

Authors' Affiliations: 1Cancer StemCell ResearchCenter, 2Department ofNeurosurgery, Samsung Medical Center and Samsung BiomedicalResearch Institute, SungkyunkwanUniversity School ofMedicine; 3Depart-ment of Anatomy, Seoul National University College of Medicine; 4Schoolof Life Sciences and Biotechnology, Korea University, Seoul, Korea;5Department of Stem Cell Biology and Regenerative Medicine, LernerResearch Institute, Cleveland Clinic, Cleveland, Ohio; and 6TherapeuticAntibody Group, Bio Lab, Samsung Advanced Institute of Technology,Samsung Electronics, Gyeonggi-do, Korea

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

Corresponding Authors: Jeongwu Lee, Department of Stem Cell Biologyand Regenerative Medicine, Lerner Research Institute, Cleveland Clinic,9500 Euclid Avenue, Cleveland, OH 44195. Phone: 216-444-9834; Fax:216-636-5454; E-mail: [email protected]; andDo-HyunNam,CancerStemCellResearch Center, Department of Neurosurgery, Samsung Medical Centerand Samsung Biomedical Research Institute, Sungkyunkwan UniversitySchool of Medicine, Seoul, Korea. E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-11-3760

�2012 American Association for Cancer Research.

CancerResearch

Cancer Res; 72(15) August 1, 20123828

on June 19, 2018. © 2012 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst May 22, 2012; DOI: 10.1158/0008-5472.CAN-11-3760

http://cancerres.aacrjournals.org/

Materials and MethodsIsolation of primary GBM cells and establishment ofshort-term cultured GSCsFollowing informed consent, glioblastoma specimens were

obtained from patients undergoing surgery at the SamsungMedical Center in accordance with the Institutional ReviewBoards. Tumor specimens were classified as GBM based onWorld Health Organization criteria by examination of pathol-ogists (2). Tumor specimens were enzymatically dissociatedinto single cells and cultured, following the procedures previ-ously reported (26, 27).

Fluorescence-activated cell sorting analysis and sortingPatient GBM-derived tumor cells were dissociated into single

cells and labeled with the following antibodies; anti-MET (R&DSystems), anti-CD133–PE (Miltenyi Biotec), and anti-CD15–FITC (MMA clone, BD Biosciences; ref. 28). For MET stainingandsorting, onemillioncellswere labeledwith2mgofantibodiesat 4�C for 10 minutes. Antibodies against mouse immunoglob-ulin conjugated to PE or fluorescein isothiocyanate (FITC) wereused as antibody isotype controls (BD Biosciences). The stainedcells were analyzed on the FACS Aria (BD Biosciences). Cell-Quest Acquisition and Analysis software (BD Biosciences) wasused to acquire and quantify the fluorescence signal distribu-tions and intensities from individual cells.

AntibodiesThe following antibodies were used as primary antibodies:

MET (Zymed and Santa Cruz for detection of total METprotein; Cell Signaling for detection of phosphorylated METproteins), Nestin, Sox2 (R&D Systems), GFAP (Dako), TuJ1(Covance), O4 (Chemicon), a-tubulin and b-actin (Sigma), andAKT and Erk (Cell Signaling).

Tumorsphere forming limiting dilution assayTumor cells were dissociated into single-cell suspensions,

sorted for MET expression, and then plated into 96-well plateswith various seeding densities (1–500 cells per well dependingon the experiment, more than 30 wells for each condition).

Intracranial tumor cell injection intoNOD/SCID Il2rg�/�

miceUnsorted GBM cells or the sorted METhigh and METlow/�

tumor cells were resuspended in 5 mL of HBSS and injectedstereotactically into the striatum of adult nonobese diabetic/severe combined immunodeficiency mice lacking interleukin-2 gamma receptor (NOD/SCID Il2rg�/�) mice by using astereotactic device (Kopf instruments; coordinates: 2 mmanterior, 2 mm lateral, and 2.5 mm depth from the dura). Micewith neurologic signs were killed for the analysis of tumorhistology and immunohistochemistry. All mice experimentswere carried out according to the guidelines of the Animal Useand Care Committees.

In vivo delivery of MET siRNA-polyelectrolyte–micellecomplexesFor the conjugation of MET siRNA or nontargeting siRNA

with polyethylene glycol 5000 (PEG5K), we followed the

procedures as previously described (29). PEG-MET siRNAor PEG-control siRNA (0.5 mg/kg body weight) complexeswere delivered through tail vein injections twice a week for4 weeks.

Determination of intracranial tumor volumesTo estimate the size of intracranial tumor, we carried

immunohistochemical analysis using an anti-human nucleispecific antibody (Chemicon) on the serial brain sections.Staining images were analyzed by ImageJ software (NIH), andthe number of immunopositive cells and the total number ofcells in a given field were calculated. The volume of theintracranial tumor was estimated by the largest width2 �length � 0.5.

Statistical analysisAll values are shown as mean � SD. For group comparison,

paired t test (2 group comparison) or ANOVA (more than 3group comparison) were used. Kaplan–Meier survival analysiswas done using Prism 4.0 software.

ResultsMET-positive tumor cells are preferentially located inperivascular regions of human GBM specimens andcoexpress GSC markers

We reasoned that components of a GSC regulator pathwaywould be preferentially expressed in perivascular regions andhypoxic edges, the proposed in vivo niches for GSCs (10). Byimmunohistochemical analysis using human GBM patientspecimens (n ¼ 93), we determined the expression patternsand intensities of various growth factor receptors, oncogenes,and stem cell–associated proteins.

High frequency of cells positive for MET (more than 10%of tumor cells) in GBM specimens was positively correlatedwith the shorter progression-free survival (PFS) and overallsurvival (OS) of the patients, suggesting a potential role ofMET overexpression in malignancy (Fig. 1A and B). We alsodetermined whether high expression of MET mRNA in GBMspecimens portends poor patient survival by the analysis ofThe Cancer Genome Atlas (TCGA) database (Supplemen-tary Fig. S1; ref. 19). Patients with high MET mRNA expres-sion have shorter PFS and OS, however, the prognosticvalue was marginal although significant (SupplementaryFig. S1).

We found that MET protein expression displayed inter- andintratumoral heterogeneity. Interestingly, cells stronglystained for MET were predominantly located near bloodvessels and hypoxic edges (Fig. 1C and D). To quantify theproximity between MET-positive cells and endothelial cells inprimary GBM specimens, we carried out dual immunofluo-rescence for CD31 (endothelial marker) and MET using thefrozen sections derived from 4 GBM patient specimens (Fig.1E–G). MET-positive cells were located closely to the nearestblood vessels, compared with MET-negative cells (Fig. 1G). Inaddition, a subset of tumor cells express both GSC enrichmentmarkers CD133/CD15 and MET (Supplementary Fig. S2;refs. 7, 28).

MET Signaling in Glioblastoma Stem Cells

www.aacrjournals.org Cancer Res; 72(15) August 1, 2012 3829




FE

μ

CD31 MET DAPI CD31 MET DAPIG

A B

TumorDistance from the nearest vessel (μm S.D)

T testMEThigh METlow/-

541 18.2 ( 13.7) 67.3 (

1.0

0.8

0.6

0.4

0.2

0

1.0

0.8

0.6

0.4

0.2

0

Progression-free survival

0 200 400

Days

600 800 0 200 400

Days

600 800 1,000

P = 0.007 P = 0.015

c-MET-low

c-MET-high

c-MET-low

c-MET-high

Overall survival

Cum

ula

tive s

urv

ival ra

te

Cum

ula

tive s

urv

ival ra

te

28.8) <0.0001

609 28.5 ( 21.3) 77.0 ( 37.4) <0.0001

586 18.2 ( 12.7) 35.9 ( 16.5) <0.0001

381 19.2 ( 8.1) 27.1 ( 14.8) 0.01

/ /

C D

Figure1. ExpressionofMET inprimary humanGBMspecimens. A andB,Kaplan–MeierPFS (A) andOS (B) graphsof patientswith high (red) and lowexpressionof MET (blue). GBM sections from 93 patients were stained by using MET antibodies and categorized by the frequency of MET-stained cells (>10% oftumor cells as high expression group). P values were determined by log-rank test: P < 0.007 for A and P <0.015 for B. C and D, representativemicrophotographs of immunohistochemical staining of MET protein in paraffin sections of GBM patient specimens. Immunopositive cells were visualized bybrown DAB staining. Arrows indicate the stained cells. Ve indicates blood vessels, and Nec indicates the area of necrosis. Scale bar represents 20 mm.E and F, representative microphotographs showing dual immunofluorescence analysis using MET antibody (green) and CD31 (red) in frozen sectionsof GBMpatient specimens. 40, 6-Diamidino-2-phenylindole (DAPI) was used to visualize nuclei. Scale bar represents 100 mm.G,measurement of the averagedistance between MET-positive cells and blood vessels.

Joo et al.

Cancer Res; 72(15) August 1, 2012 Cancer Research3830




To obtain more quantitative results, we determined theexpression level of MET in the freshly dissociated cellsobtained from 23 different GBM patients by fluorescence-activated cell sorting (FACS) analysis (Fig. 2A). The percen-tages of MET-positive cells vary between the samples (3.7%–76.4% of bulk cells; SD 16.1%). In 15 of 23 samples, thepercentages of MET-positive cells are between 10% to 30% ofthe bulk cells. The cells dissociated from nontumor epilepticbrain specimens did not have any distinct MET-positive pop-ulation (data not shown). To determine whether there aredistinct subpopulations expressing both MET and CD133 and/or CD15, we carried out dual FACS staining on the freshlyisolated GBM cells derived from 14 different patients (Fig. 2Aand Supplementary Table S1). CD133- and/or CD15-positivecells were enriched (2- to 7- fold) in MET-positive subpopu-lation compared with the bulk tumor cells. Conversely, MET-positive cells were enriched in CD133þ or CD15þ populations(Supplementary Table S1).To corroborate these findings, we extended our analysis to

the previously characterized GSCs derived from human GBM

specimens (26, 28). GSCs and their derivatives used in thisstudy are described in supplementary information (Supple-mentary Table S2). We fractionated these short-term culturedGSCs by FACS and carried out immunoblot analysis (Fig. 2B).MET protein was highly expressed in CD133þ (or in CD15þ)subpopulations compared with the corresponding negativepopulations in all 3 GSCs (Fig. 2B). Total MET and activatedMET protein (Y1234) were preferentially expressed in GSC-enriched population compared with GSC-depleted cells (Fig.2C). Collectively, these data indicated that MET-positive GBMcells have activated MET signaling and that they significantlyoverlap with CD133þ and/or CD15þ GBM populations.

METhigh cells sustain in vitro growth and are highlyclonogenic compared with METlow/� cells

To determine whether METhigh GBM cells have cellularcharacteristics associatedwithGSCphenotypes, we used FACSusing an antibody recognizing the extracellular domain ofMETprotein and prospectively isolated METhigh and METlow/�

subpopulations from the patient GBM specimen 448. Weoperationally defined METhigh cells as the cells with top 10%to 20% highest intensities out of total MET-positive cells. Afterconfirming the purity and viability of these sorted populationsby FACS reexamination, we first determined the growth kinet-ics of each subpopulation by culturing these cells in thestandard stem cell culture condition. METhigh cells readilyformed neurosphere-like aggregates and continued to prolif-erate, whereas METlow/� cells failed to sustain active cellgrowth in vitro (Fig. 3A–C).

Clonogenic growth as neurospheres is an in vitro indicator ofself-renewal in normal neural stem/progenitor cells (NSC) andGSCs. To determine the clonogenic potentials of METhigh andMETlow/� populations, we carried out limiting dilution assays(LDA) using both 448 freshly isolated GBM cells and 206 short-term cultured GSCs (Fig. 3D). FACS-sorted METhigh andMETlow/� cells were plated into 96-well plates with variousseeding densities and allowed to grow. Neurospheres werereadily formed by METhigh cells but not by METlow/� cells (Fig.3D). In addition, we carried out a competitive growth assays inwhich METhigh and METlow/� cells were differentially labeledwith GFP and red fluorescence protein (RFP), respectively,mixed at various ratios, and then cultured for additional 2weeks. Even when the initial fraction of METhigh GFP cells was10%, more than 85% of the resultant cells after the culture wereGFP positive (Fig. 3E). Taken together, these data indicatedthat METhigh GBM cells are highly clonogenic and sustain invitro growth of total population.

To determine whether the HGF/MET signaling pathway isfunctional in GSCs, we examined the activation status of METdownstream effectors (Supplementary Fig. S3A). GSCs dis-played basal activation of MET, suggesting the presence of anautocrine loop with further response to the addition of exog-enous HGF. Treatment with exogenous HGF further increasedactivated MET (Y1234 and Y1349), AKT (S473), and ERK (16).Consistent with this, we detected high levels of the secretedHGF protein in GSC-conditioned media (7 of 11 GSCs) byELISA assays (Supplementary Fig. S3B). Finally, we determinedwhether HGF could induce the proliferation of GBM cells by

18.5 % 1.4 %

77.8 % 2.2 %

10.1 % 3.5 %

86.3 % 0.2%

14.7 % 5.2 %

72.0 % 8.1 %

9.3 % 0.5 %

89.2 % 1.0 %

9.1 % 0.8 %

88.6% 1.6 %

17.3 % 0.7 %

96.8 % 0.8 %

6.4 % 1.0 %

91.5 % 1.1 %

4.7 % 2.7 %

88.0 % 4.6 %

5.6 % 1.9 %

92.3 % 0.2 %

14.4 % 5.9 %

76.7 % 3.0 %

10.2 % 10.1 %

73.0 % 6.7 %

8.9 % 8.1 %

82.2 % 0.8 %

ME

T-A

PC

CD15-FITC

In vitro GSCs Xenograft tumor

CD

133

+

211 308 905 308

MET

pAKT

AKT

SOX2

Actin

MET

pMET

pAKT

AKT

SOX2

Actin

CD

133

+

CD

133

-

1C

D33

-

CD

15

+

CD

15

+

CD

15

-

CD

15

-

CD133-FITC CD15-FITC

ME

T-A

PC

CD

133-A

PC

538

049

719

776

CB

A

Figure 2. Expression of MET in primary human GBM specimens andGBM-derivedGSCs. A, flowcytometry data ofMET-, CD133-, andCD15-positive cells in freshly isolated patient GBM cells. B and C, immunoblotanalysis using CD133 and/or CD15 � cell lysates from short-termcultured GSCs (B) and GSC-derived xenograft tumors (C). The numbersrepresent the designated names for each GSC. Actin was used as aloading control.






carrying out LDA. Similar to EGF, HGF increased the clono-genic growth of GBM cells (Supplementary Fig. S3C). Collec-tively, these data indicated the presence of active HGF/METsignaling in GSCs.

METhigh GBM cells are highly tumorigenic in orthotopictransplantation models

Tumorigenicity in vivo is the gold standard for determiningcancer stemness. We injected METhigh and METlow/- cells into

the brains of highly immunodeficient NOD/SCID Il2rg�/�miceand monitored tumor formation (30). METhigh cells from 4GBMs (448 and 464, freshly isolated patient GBM cells; 822 and206, short-term cultured GSCs) generated tumors more effi-ciently than their correspondingMETlow/� cells (Fig. 4A andB).Of note, all of the above GBM cells express HGF (data notshown). To estimate the relative enrichment of tumorigenicityin METhigh cells compared with METlow/� cells, we carried outin vivo tumorigenicity titration assays (Fig. 4C). Varying num-bers ofMEThigh andMETlow/� cells isolated fromprimary GBMspecimen (559 and 905) were injected into brains of NOD/SCIDIl2rg�/� mice. Two of 5 mice injected with 100 METhigh cellsdeveloped tumors and so did all themice (n¼ 5) that had 1,000METhigh cells (median survival of 88 days). In contrast, all themice that received 100 METlow/� cells and 4 of 5 mice injectedwith 1,000 METlow/� cells showed no sign of tumor develop-ment 6 months after the injection (Fig. 4C). Kaplan–Meiersurvival analysis further showed significant survival differ-ences between animals receiving these subpopulations (Fig.4D and E).

Similar to the parental GBM tumor, 448 METhigh cell–derived xenograft tumors harbored a mixture of METhigh cellsand METlow/� cells, indicating the reconstitution of tumorheterogeneity (ref. 31; Supplementary Fig. S4). In addition,immunohistochemical analysis on the xenograft tumorsrevealed that METhigh cells were preferentially located nearthe blood vessels and infiltrating edges of tumors (Fig. 4F–H).Robust expression of tumor-derived HGF was also detected inxenograft sections by immunohistochemical analysis, suggest-ing an active HGF/MET signaling in these tumors (Supple-mentary Fig. S5).

METhigh cells are efficient in tumor formation regardlessof CD133 expression

Our data showed that some of METhigh cells coexpressCD133 (Fig. 2 and Supplementary Table S1). To determinewhether MET expression alone can enrich GSCs, we fraction-ated GBM cells into 4 subpopulations (CD133þ/METhigh,CD133�/METhigh, CD133þ/METlow/�, and CD133�/METlow/�

cells) and determined their tumor formation efficiencies. BothCD133þ/METhigh and CD133�/METhigh cells, derived from thefreshly isolated GBM cells (060) and GBM xenografts (559 and464), were highly efficient in tumor formation, whereasCD133þ/METlow/� and CD133�/METlow/� cells were not(Fig. 5A). In vitro clonogenic potential of each subpopulationcorrelated with in vivo tumorigenicity of the correspondingcells (Fig. 5B). These data suggested that MET, but not CD133,is a major determinant for GSC enrichment at least in theseGBMs.

MET inhibition decreases the survival, migration, andclonogenicity of GSCs

To interrogate the role of MET in GSC biology, we inhibitedMETby short hairpin RNA (shRNA)–mediated knockdown andevaluated its effects on GSCs. To ensure the specific knock-down of MET, we used 2 independent sequences of shRNAdirected againstMETmRNA. BothMET shRNA constructs ledto a significant reduction in MET protein compared with the

80

70

60

50

40

30

20

10

0

100%90%80%70%60%50%40%30%20%10%

0%

100%90%80%70%60%50%40%30%20%10%

0%

1009080706050403020100

-10 40 90 140

No of initial cells/well Ratio of input cells100 90 50 10 0

0 10

Culture after sorting (d)

448

Culture after sorting (d)

4

3

2

1

0

No

of c

ells

afte

r cu

lture

(E

6)

% o

f wel

ls w

ithou

t sph

eres

20 0 10 20

Sub-G1 G0/G1

Sub-G1

G0/G1

G2/M

G2/M

S

S

METhigh

METhigh

METhigh

METlow/-

METlow/-

METlow/-

448

905

903METhigh

METlow/-METhigh

METlow/-

% o

f to

tal p

op

ula

tio

n

ED

C

A B

Figure 3. Characterization of METhigh and METlow/� cells from freshlyisolated GBM specimens and GBM-derived GSCs. A, representativemicrophotographs of METhigh and METlow/� GBM cells cultured for aweek. B, cell-cycle analysis ofMEThigh andMETlow/� populations directlyisolated from 448, the freshly isolated patient GBM cells. Eachsubpopulation was seeded at the density of 5,000 cells/mL media andcultured for 2weeks. After 2weeks of in vitro culture, cells were harvestedfor cell-cycle analysis. Error bars represent SD (conducted in triplicates).�, P < 0.05. C, cell-cycle analysis of METhigh and METlow/� GBM cellsisolated from short-term cultured GSC 905. Each subpopulation wassorted on day 0 and cultured for additional 20 days. At the indicated timepoints, we carried out cell-cycle analysis of these populations. D,limiting dilution sphere forming assays to determine the clonogenicpotentials of METhigh and METlow/� subpopulations from 448 GBM cells.Cells were plated into 96-well plates with various seeding densities(2–50 cells per well, 30 wells per each condition). E, competitiveproliferation assay todetermine relative proliferationpotentials ofMEThigh

and METlow/� cells isolated from the freshly isolated patient GBM cells(903). GFP-transduced METhigh cells and RFP-transduced METlow/�

GBM cells were mixed at various ratios (% of GFP input cells are shownin x axis) and cultured. Two weeks later, the number of total cells wascounted and the ratio of GFP- and RFP-positive cells was determinedby FACS analysis.

Joo et al.





nontargeting shRNA control (Supplementary Fig. S6). First, wedetermined the cell growth/survival ofMET knockdown GBMcells. Compared with the control, MET knockdown cellsshowed a significant decrease in the proliferation index (Sup-plementary Fig. S6B and C). MET knockdown cells were largelyinG0/G1 phasewithout notable increase in sub-G1 populations,suggesting that cell-cycle arrest is a main contributor of the

decreased cell proliferation. MET knockdown significantlydecreased the clonogenicity of various GSCs, determined byLDA (Supplementary Fig. S6D and E). We also tested whetherpharmacologic inhibition of MET signaling can decrease GSCclonogenicity. GSCs treated with SU11274, a widely used METkinase inhibitor (25, 32), were less efficient in tumor-sphereformation compared with the control (Fig. 6A).

METhigh

METlow/-

559No. of mice with tumor/total

mice (median survival)

Cell No.

injectedMEThigh METlow/-

10,000 5/5 (8.0 wks) 4/5 (15.6 wks)

1,000 5/5 (12.6 wks) 1/5

100 2/5 0/5

905No. of mice with tumor/total

mice (median survival)

Cell No.

injectedMEThigh METlow/-

500,000 N.D. 5/5 (10.5 wks)

100,000 5/5 (7.6 wks) 2/6

10,000 4/5 (9.1 wks) 0/5

1,000 3/5 (12.7 wks) 0/5

METhigh META

C D

E

F G H

B

448 464

Survival (d)

Survival (d)

822 206

Tu

mo

r vo

lum

e (

mm

3)

Perc

en

t su

rviv

al (%

)P

erc

en

t su

rviv

al (%

)

high

METlow/-

METlow/-

559 METhigh

559 METlow/-

905 METhigh

905 METlow/-

200

150

100

50

0

100

80

60

40

20

0

0

0 50 100 150

100

80

60

40

20

00 50 100 150

Figure 4. Tumorigenic potential of METhigh and METlow/� cells from freshly isolated GBM specimens and GBM-derived GSCs. A, representativemicrophotographs of the brain sections of mice injected with either METhigh and METlow/� cells derived from short-term cultured GSC 822. Bar represents2 mm. Ten thousand cells of either METhigh or METlow/� cells were injected into the brains of NOD/SCID Il2rg�/� mice. Mice were sacrificed 2 monthsafter injection, and their brains were sectioned and stained with hematoxylin and eosin (H&E). B, tumor formation assay of METhigh and METlow/�

subpopulations from freshly isolatedGBMcells (448 and 464) and short-term culturedGSCs (822 and 206). Tumor sizewasmeasured from the brain sectionsof tumor cell injected mice (3 mice per each group). �, P < 0.01. C, in vivo tumorigenicity titration assay of METhigh and METlow/� subpopulations fromfreshly isolated GBM cells (559) and short-term cultured GSCs (905). The number of tumor-bearing mice and tumor cell injected mice are shown. Mediansurvival of each group is shown in parenthesis. D and E, Kaplan–Meier survival plot of mice injected with METhigh and METlow/� cells from freshlyisolated GBM 559 cells (1,000 cells injected) and short-term cultured GSC 905 (100,000 cells injected). P value was determined by log-rank test. P < 0.01 inboth 559 and 905 cells. F–H, representative microphotographs of the brain sections of 448 METhigh cell–derived xenograft tumors. Many tumor cellswere found in corpus callosum (CC) in H&E section (F) and stained with anti-PCNA antibody (G) and anti-MET antibody (H). Immunopositive cells werevisualized by brown DAB staining. Arrows indicate strongly stained cells. Bars represent 100 mm.






HGF is an effective chemokine for glioma cell migration(13, 33). To test whether MET signaling mediates invasivegrowth of GSCs, we carried out transwell-based in vitromigration assays (Supplementary Fig. S7A and B). We platedcontrol shRNA- or MET shRNA–transduced GBM cells inserum-free media and added serum (or HGF)-containingmedia on the other side of the transwell membranes tostimulate migration.MET knockdown decreased the numberof migrating/invading cells up to 70% compared with thecontrol (Supplementary Fig. S7A and B). In addition, wecarried out ex vivo brain slice assays that simulate in vivo

migration behavior much better than transwell assays(Supplementary Fig. S7C). RFP-expressing GSCs were trea-ted with a chemical MET inhibitor PHA665752 for one day,mixed with GFP-expressing GSCs treated with the vehiclecontrol, and then implanted into the cortex region of 300-mmthick brain slices. Three days after, more than 90% of themigrated cells were GFP positive, suggesting that METinhibition decreases glioma cell invasion (SupplementaryFig. S7C and D). Together, these data showed that METsignaling is required for the self-renewal, proliferation, andmigration of GSCs.

In vivo MET targeting via liposome-conjugated siRNAincreases the survival of tumor bearing mice

By using GSC-derived tumors as a clinically relevant modelsystem, we wanted to test whether MET targeting in vivo couldprovide therapeutic benefit. Polyelectrolyte complex micellesare stable in vivo and efficient for intracranial delivery of siRNAbecause of the highly lipophilic nature of brain tissues (29). Weconjugated MET siRNA or nontargeting siRNA with polyeth-ylene glycol 5000 (PEG5K) and mixed with small liposomes(29). Intracranial tumorswere generated from3 freshly isolatedGBMcells (559 and 578, 5 for each group; 464, 4 for each group).These cells express both MET and HGF at high levels, deter-mined by Western blot analysis and HGF ELISA (data notshown). Once tumors were established in themouse brains, weinitiated the administration of either PEG-MET siRNA or PEG-control siRNA (0.5 mg/kg body weight) intravenously twice aweek for 4 weeks. Compared with the tumors from the controlsiRNA-treated mice, tumor cells in MET siRNA-treated micerevealed a significant decrease in MET staining positivity (Fig.6B). After 4 weeks of treatment, we monitored the size oftumors by MRI (Fig. 6C) and brain histology section (Fig. 6Dand Supplementary Fig. S8). All ofMET siRNA–treated groupsshowed a significant decrease in tumor volumes, indicating theefficacy of in vivo MET targeting in intracranial GSC-derivedtumors (Fig. 6E).

Irradiation induces MET upregulation and METtargeting decreases the clonogenicity of surviving GBMcells

Irradiation treatment decreases tumor burden and pro-long the survival of GBM patients. However, patients even-tually succumb because of tumor recurrence. Given theobservations that GSCs preferentially survive radiation andthat MET targeting disrupted GSC self-renewal and clono-genicity, we hypothesized that combination of radiation andMET inhibition in GSCs might be an effective therapeuticapproach (9, 34, 35). Cells of a short-term cultured GSC (822)were irradiated in vitro with doses of 5 or 10 Gy andharvested 2 days later. We found about a 3-fold increase inthe number of MET-positive cells and upregulation of METprotein in irradiated cells compared with nonradiated con-trol (Fig. 7A and B). To determine the clonogenic potentialsof METhigh and METlow/� GBM cells after radiation treat-ment, we sorted METhigh and METlow/- cells from the freshlyisolated 448 GBM cells, exposed them to radiation, andimmediately seeded at the clonal density for LDA assays.

GBM Cell No.

Subpopulations

(No. of mice with tumor/total mice)

CD133+/

METhigh

CD133-

/METhigh

CD133+/

METlow/-

CD133-

/METlow/-

060 10,000 3/3 3/3 0/4 0/4

559 1,000 5/5 4/5 N.D. 0/5

464 10,000 5/5 5/5 3/5 1/5

464 1,000 4/4 4/4 0/4 0/4

A

B

0

20

40

60

80

100

0 100 200 300 400 500

060CD133+/MET high

CD133+/MET low/-

CD133-/METhigh

CD133-/METlow/-

No. of initial cells

% o

f w

ells w

/o s

ph

ere

s

0

20

40

60

80

100

0 100 200 300 400 500

090

No. of initial cells

% o

f w

ells w

/o s

ph

ere

s

Figure 5. Clonogenicity and tumorigenicity of GBM subpopulationsfractionated based on MET and CD133 expression. A, in vivotumorigenicity titration assay of CD133þ/METhigh, CD133�/METhigh,CD133þ/METlow/�, and CD133�/METlow/� subpopulations derivedfrom freshly isolatedGBMcells (060) and xenograft tumor–derived GSCs(559 and 464). The number of tumor-bearingmice and tumor cell injectedmice are shown. Ten thousand tumor cells were intracranially injectedinto the brains of SCID mice and animal survival was monitored over5 months. B, limiting dilution sphere forming assays to determine theclonogenic potentials of each subpopulations from 060 and 090, thefreshly isolated patient GBM cells. Cells were plated into 96-wellplates with various seeding densities (1–500 cells per well, 24 wellsper each condition).

Joo et al.





METhigh cells remained highly clonogenic after radiationtreatment, whereas irradiated METlow/� cells were signifi-cantly impaired in the ability to generate subsequent colo-nies (Fig. 7C). To determine whether MET is directlyinvolved in the clonal regrowth after radiation, we carriedout the similar clonogenic assays by MET knockdown ortreatment with SU11274. Cells with MET inhibition werehardly able to generate spheres after radiation treatment,suggesting that MET inhibition significantly sensitizes GBMcells to radiation (Fig. 7D and Supplementary Fig. S9).The above data raise the possibility that MET upregulation

after radiation can be a prognostic biomarker for the GBMpatients who have received radiation therapy. We acquired 14matched sets of pre- and postirradiated (recurrent) tumorspecimens from the same GBM patients (Supplementary TableS3). All the patients were diagnosed as de novo primary GBMsand received radiation as a front-line therapy after initial

surgical resection. We determined the expression level of METprotein by immunohistochemical analysis on the sections frompre- and postradiated GBM specimens. In 4 sets, MET expres-sion in postradiated GBM specimens was significantlyincreased (>2-fold increase in the number of MET-positivecells) compared with the untreated specimens. The remaining10 sets showed either unchanged or modest changes in METexpression. Notably, the patients showing the significant METinduction hadworse prognosis (median survival of 55 days afterrecurrence) than theother group (median survival of 189 days;P< 0.01; Fig. 7E), suggesting that MET upregulation after radi-ation is associated with aggressive growth of recurrent tumors.

DiscussionGSCs remains controversial because of unresolved ques-

tions with regard to the frequency of these cells, the surface

Figure 6. Effects of MET inhibitionon in vitro clonogenicity andin vivo tumorigenicity of GSCs.A, limiting dilution sphere formingassays to determine the clonogenicpotentials of various GSCs aftertreatments with SU11274 (2 mmol/L).B, representative microphotographsof immunohistochemistrydetermining MET expressionin the brains of control siRNA-or MET siRNA-treated mice.Immunopositive cellswere visualizedby brown DAB staining. Arrowsindicate strongly stained cells. Scalebar represents 100 mm. C and D, thebrains from control siRNA or METsiRNA–treated mice were examinedby MRI images (C) and histologysections (D). White arrowheadsindicate tumor. Bars represent 2mm.E, tumor volume in the brains wasdetermined by measuring the size ofthe tumor in the serial H&E sections.Error bars represent SD.

0

20

40

60

80

100

120

0 500 1,000

controlSU11274 2 μμmol/L

464267

0

20

40

60

80

100

120

-20 30 80

controlSU11274 2 μmol/L

0

20

40

60

80

100

120

0 50 100

controlSU11274 2 μmol/L

745

%o

f w

ells w

/o s

ph

ere

s

No. of initial

cells/well

No. of initial

cells/well

No. of initial

cells/well

B

A

C

D

E

Control

140

120

100

80

60

40

20

0

140

120

100

80

60

40

20

0

P < 0.01

P < 0.001

Tu

mo

r v

olu

me

(m

m3)

Tu

mo

r v

olu

me

(m

m3)

siRNA

MET

siRNA

Control

siRNA

MET

siRNA

464

578

578

Control siRNA MET siRNA








markers by which they can be identified, and the nature of thecell(s) of origin (31, 36). Previous reports have identified a seriesof GSC enrichment markers, including CD133, CD15, CD44,and A2B5 (7, 28, 37, 38). Although useful for the prospectiveisolation of putative GSCs, it is unclear whether these markershave distinct functions pertinent to GSC phenotypes (39, 40).Therefore, our screening approach in search for a GSC regu-lator was focused on the protein expression pattern in patient-

derived GBM specimens, based on the hypothesis that GSCsare highly enriched in perivascular regions of human GBMs insitu (41).

In this study, we identified a distinct fraction of cellsexpressing high level of MET in various human primary GBMsand showed that these subpopulations have key characteristicsof GSCs. Through extensive in vivo limiting dilution tumorformation assays, we have shown that METhigh GBM cells are

822

ShCount

ShMET345

ShMET3310

% of METhigh cells after radiation

c-MET FITC-A

A

C

D

E

B

Co

un

ts1

00

80

60

40

20

0

100 101 102 103

104

Cont 5 Gy 10 Gy

822

MET

METhight contMEThight 5 Gy

METlow/- 5 Gy

448

Nonradiated: 14%

10 Gy : 43%

5 Gy : 22%

No. of initial cells/ well

0

100

90

80

70

60

50

40

30

20

10

0

100

80

60

40

20

0

100

80

60

40

20

0

% o

f w

ells w

/o s

ph

ere

s

% o

f w

ells w

/o s

ph

ere

s

% o

f w

ells w

/o s

ph

ere

s

50 100 150 200


0 50 100 150

1.0

0.8

0.6

0.4

0.2

0.0

200


P < 0.01

0 100 200 300 400 500

0 100 200 600

Progression-free survival (d)

Fra

cti

on

su

rviv

al

300 400 500

METlow/- Cont

pAKT

AKT

β-Actin

5 Gy822

10 Gy

Figure 7. Roles of MET in GBMradiation response. A, flowcytometry analysis of MET (FITClabeled, x-axis) in GSCs afterirradiation. GSC 822 cells wereirradiated in vitro either with 5 or 10Gy, and the survived cells wereexamined for MET expression at 2days after irradiation. The verticalline depicts the cut-off for METhigh

cells. Numbers indicatepercentage of METhigh cells in eachcondition. B, immunoblot analysisof lysates from the cells used in A.C, limiting dilution sphere-formingassay to determine clonogenicpotentials of METhigh and METlow/�

GBM 448 cells with 5 Gy or without(cont) irradiation. The sorted cellswere mock treated or irradiatedand immediately seeded withvarious seeding densities (2–50cells per well, 30 wells per eachcondition). D, limiting dilutionsphere forming assay to determineclonogenic potentials of METknockdown in GBM 822 cells afterirradiation. Higher seedingdensities (5–500 cells per well)were used. Three independentexperiments were carried out, andthe data from a representativeexperiment were shown. E,Kaplan–Meier survival plot of thepatients with MET upregulation(shown as red) or without METupregulation (shown as black) inpostradiated tumors comparedwith untreated GBM specimens.P value was determined by log-rank test: P < 0.01.

Joo et al.





highly enriched with tumor stem/propagating populations,supporting that MET is a key regulator of GSCs. We showedthe importance of MET in GSC biology by showing that METinhibition disrupts the clonogenicity, radioresistance, andtumorigenicity of GSCs. Co-option with angiogenesis andtumor cell survival in hypoxia are critical for tumor growth,and MET signaling is known to play important roles on theseprocesses. Therefore, high MET expression in GSCs can beviewed as an adaptive response of cancer cells to their micro-environment (42, 43).Numerous small molecule inhibitors and antagonistic anti-

bodies targeting theHGF/METpathway are at advanced stagesof clinical development (44–48). Understanding of the molec-ular determinant(s) of response and resistance to HGF/METtargeting therapeutic agents is one of the most critical unmetneeds in clinical research. A few of biomarkers to predicttherapeutic response have been proposed, including genomicamplification ofMET, HGF levels, orMET expression (total andphosphorylated MET), and PTEN status (33, 49, 50). Robust-ness of these biomarkers in GBM needs further validation. Forexample, genomic MET amplification seems to predict sensi-tivity to MET inhibition in gastric cancers but less likely inGBMs (33). A recent paper reported that HGF autocrine statusmay correlate withMET activity in GBMand predict sensitivityto MET inhibitors (33). We have determined the correlationbetween tumor-derived HGF and sensitivity toMET inhibition.Our ongoing studies so far indicated that high HGF-producingGBM cells seem to be more sensitive to MET inhibitioncompared with the little or no HGF-expressing GBM cells,consistent with the previous report (data not shown; ref. 33).Further studies are warranted to determine the role of HGF inGSC biology, the effects of HGF targeting on GSC self-renewaland tumorigenicity, and the correlation betweenHGF andGSCsensitivity to MET inhibition in vivo.

In conclusion, we show that MET is an enrichment markerfor GSCs and a functional requisite for cancer stem pheno-type. Not only will these data provide a clue for a morethorough understanding of cancer stem cell biology but alsothey further implicate MET as a promising therapeutictarget.

Disclosure of Potential Conflicts of InterestNo potential conflicts of interest were disclosed.

Authors' ContributionsConception and design: K.M. Joo, J. Lee, D.-H. NamDevelopment of methodology: K.M. Joo, J. Jin, B.G. Kang, Y.-J. Kang, H. Kim, J.Lee, D.-H. NamAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.):K.M. Joo, J. Jin, E. Kim, K.H. Kim, Y. Kim, B.G. Kang, J.D.Lathia, H.J. Seol, D.-S. Kong, J.-I. Lee, J.N. Rich, D.-H. NamAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K.M. Joo, J. Jin, Y. Kim, J.D. Lathia, H.J. Seol, J. Lee, D.-H. NamWriting, review, and/or revision of the manuscript: K.M. Joo, Y. Kim, P.H.Song, J.N. Rich, J. LeeAdministrative, technical, or material support (i.e., reporting or orga-nizing data, constructing databases): K. M. Joo, J. Jin, K.H. Kim, H. KimStudy supervision: K.H. Cheng, D.-H. Nam

Grant SupportThis work was supported by the National Research Foundation of KOREA

(NRF) grant funded by the Korea government (MEST; No.20090093731; D.-H.Nam), a grant of the Korea Healthcare Technology R&D Project, Ministry forHealth & Welfare Affairs, Republic of Korea (A092255; D.-H. Nam), the grant forthe Future-based Technology Development Program (2010-0020232) funded byNRF of the MEST, Republic of Korea (D.-H. Nam), Cleveland Clinic Foundationgrant (J. Lee), and CaseWestern Reserve University/Cleveland Clinic CTSCGrantUL1 RR024989 and RES50-5510 (J. Lee).

The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 18 U.S.C. Section 1734 solely to indicate thisfact.

Received November 14, 2011; revised April 18, 2012; accepted May 3, 2012;published OnlineFirst May 22, 2012.

References1. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn

MJ, et al. Radiotherapy plus concomitant and adjuvant temozolomidefor glioblastoma. N Engl J Med 2005;352:987–96.

2. Louis DN, Ohgaki H, Wiestler OD, Cavenee WK, Burger PC, Jouvet A,et al. The 2007 WHO classification of tumours of the central nervoussystem. Acta Neuropathol 2007;114:97–109.

3. Sanai N, Alvarez-Buylla A, Berger MS. Neural stem cells and the originof gliomas. N Engl J Med 2005;353:811–22.

4. Rosen JM, Jordan CT. The increasing complexity of the cancer stemcell paradigm. Science 2009;324:1670–3.

5. Reya T, Morrison SJ, ClarkeMF,Weissman IL. Stem cells, cancer, andcancer stem cells. Nature 2001;414:105–11.

6. Visvader JE, Lindeman GJ. Cancer stem cells in solid tumours: accu-mulating evidence and unresolved questions. Nat Rev Cancer2008;8:755–68.

7. Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, Hide T, et al.Identification of human brain tumour initiating cells. Nature 2004;432:396–401.

8. Liu G, Yuan X, Zeng Z, Tunici P, Ng H, Abdulkadir IR, et al. Analysis ofgeneexpressionandchemoresistanceofCD133þcancer stemcells inglioblastoma. Mol Cancer 2006;5:67.

9. BaoS,WuQ,McLendonRE,HaoY,ShiQ,HjelmelandAB, et al.Gliomastem cells promote radioresistance by preferential activation of theDNA damage response. Nature 2006;444:756–60.

10. Li Z, Bao S,WuQ,Wang H, Eyler C, Sathornsumetee S, et al. Hypoxia-inducible factors regulate tumorigenic capacity of glioma stem cells.Cancer Cell 2009;15:501–13.

11. Bao S, Wu Q, Sathornsumetee S, Hao Y, Li Z, Hjelmeland AB, et al.Stem cell-like glioma cells promote tumor angiogenesis through vas-cular endothelial growth factor. Cancer Res 2006;66:7843–8.

12. Calabrese C, Poppleton H, Kocak M, Hogg TL, Fuller C, Hamner B,et al. A perivascular niche for brain tumor stem cells. Cancer Cell2007;11:69–82.

13. Birchmeier C, Birchmeier W, Gherardi E, Vande Woude GF. Met,metastasis, motility and more. Nat Rev Mol Cell Biol 2003;4:915–25.

14. Boccaccio C, Comoglio PM. Invasive growth: a MET-driven geneticprogramme for cancer and stem cells. Nat RevCancer 2006;6:637–45.

15. Abounader R, Laterra J. Scatter factor/hepatocyte growth factor inbrain tumor growth and angiogenesis. Neuro Oncol 2005;7:436–51.

16. Trusolino L, Bertotti A, Comoglio PM. MET signalling: principles andfunctions in development, organ regeneration andcancer. Nat RevMolCell Biol 2010;11:834–48.

17. KongDS, SongSY, KimDH, Joo KM, Yoo JS, Koh JS, et al. Prognosticsignificance of c-Met expression in glioblastomas. Cancer 2009;115:140–8.

18. Nabeshima K, Shimao Y, Sato S, Kataoka H, Moriyama T, Kawano H,et al. Expression of c-Met correlates with grade of malignancy in






human astrocytic tumours: an immunohistochemical study. Histopa-thology 1997;31:436–43.

19. TCGA. Comprehensive genomic characterization defines human glio-blastoma genes and core pathways. Nature 2008;455:1061–8.

20. VerhaakRG,HoadleyKA,PurdomE,WangV,Qi Y,WilkersonMD, et al.Integrated genomic analysis identifies clinically relevant subtypes ofglioblastoma characterized by abnormalities in PDGFRA, IDH1, EGFR,and NF1. Cancer Cell 2010;17:98–110.

21. Phillips HS, Kharbanda S, Chen R, Forrest WF, Soriano RH, Wu TD,et al. Molecular subclasses of high-grade glioma predict prognosis,delineate a pattern of disease progression, and resemble stages inneurogenesis. Cancer Cell 2006;9:157–73.

22. Pennacchietti S, Michieli P, Galluzzo M, Mazzone M, Giordano S,Comoglio PM. Hypoxia promotes invasive growth by transcription-al activation of the met protooncogene. Cancer Cell 2003;3:347–61.

23. Corso S, Migliore C, Ghiso E, De Rosa G, Comoglio PM, Giordano S.Silencing the MET oncogene leads to regression of experimentaltumors and metastases. Oncogene 2008;27:684–93.

24. Nicoleau C, Benzakour O, Agasse F, Thiriet N, Petit J, Prestoz L, et al.Endogenous hepatocyte growth factor is a niche signal for subven-tricular zone neural stem cell amplification and self-renewal. StemCells 2009;27:408–19.

25. Li Y, Li A, GlasM, Lal B, YingM,SangY, et al. c-Met signaling induces areprogramming network and supports the glioblastoma stem-likephenotype. Proc Natl Acad Sci U S A 2011;108:9951–6.

26. Lee J, KotliarovaS, Kotliarov Y, Li A, SuQ,DoninNM, et al. Tumor stemcells derived from glioblastomas cultured in bFGF and EGF moreclosely mirror the phenotype and genotype of primary tumors thando serum-cultured cell lines. Cancer Cell 2006;9:391–403.

27. Joo KM, Nam DH. Prospective identification of cancer stem cells withthe surface antigen CD133. Methods Mol Biol 2009;568:57–71.

28. SonMJ,Woolard K, NamDH, Lee J, Fine HA. SSEA-1 is an enrichmentmarker for tumor-initiating cells in human glioblastoma. Cell Stem Cell2009;4:440–52.

29. Kim SH, Jeong JH, Lee SH, Kim SW, Park TG. LHRH receptor-mediated delivery of siRNA using polyelectrolyte complex micellesself-assembled from siRNA-PEG-LHRH conjugate andPEI. BioconjugChem 2008;19:2156–62.

30. Ito M, Hiramatsu H, Kobayashi K, Suzue K, Kawahata M, Hioki K, et al.NOD/SCID/gamma(c)(null)mouse: an excellent recipientmousemodelfor engraftment of human cells. Blood 2002;100:3175–82.

31. Magee JA, Piskounova E, Morrison SJ. Cancer stem cells: impact,heterogeneity, and uncertainty. Cancer Cell 2012;21:283–96.

32. Sattler M, Pride YB, Ma P, Gramlich JL, Chu SC, Quinnan LA, et al. Anovel small molecule met inhibitor induces apoptosis in cells trans-formed by the oncogenic TPR-MET tyrosine kinase. Cancer Res2003;63:5462–9.

33. Xie Q, Bradley R, Kang L, Koeman J, Ascierto ML, Worschech A, et al.Hepatocyte growth factor (HGF) autocrine activation predicts sensi-

tivity to MET inhibition in glioblastoma. Proc Natl Acad Sci U S A2012;109:570–5.

34. Lal B, Xia S, Abounader R, Laterra J. Targeting the c-Met pathwaypotentiates glioblastoma responses to gamma-radiation. Clin CancerRes 2005;11:4479–86.

35. De Bacco F, Luraghi P, Medico E, Reato G, Girolami F, Perera T,et al. Induction of MET by ionizing radiation and its role in radio-resistance and invasive growth of cancer. J Natl Cancer Inst2011;103:645–61.

36. Chen J, McKay RM, Parada LF. Malignant glioma: lessons fromgenomics, mouse models, and stem cells. Cell 2012;149:36–47.

37. Anido J, Saez-Borderias A, Gonzalez-Junca A, Rodon L, Folch G,CarmonaMA, et al. TGF-beta receptor inhibitors target theCD44(high)/Id1(high) glioma-initiating cell population in human glioblastoma. Can-cer Cell 2010;18:655–68.

38. Ogden AT, Waziri AE, Lochhead RA, Fusco D, Lopez K, Ellis JA, et al.Identification of A2B5þCD133- tumor-initiating cells in adult humangliomas. Neurosurgery 2008;62:505–14; discussion 514–05.

39. Bidlingmaier S, Zhu X, Liu B. The utility and limitations of glycosylatedhuman CD133 epitopes in defining cancer stem cells. J Mol Med2008;86:1025–32.

40. Nguyen LV, Vanner R, Dirks P, Eaves CJ. Cancer stem cells: anevolving concept. Nat Rev Cancer 2012;12:133–43.

41. Lathia JD, Gallagher J, Heddleston JM, Wang J, Eyler CE, MacswordsJ, et al. Integrin alpha 6 regulates glioblastoma stem cells. Cell StemCell 2010;6:421–32.

42. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation.Cell 2011;144:646–74.

43. Hjelmeland AB, Lathia JD, Sathornsumetee S, Rich JN. Twisted tango:brain tumor neurovascular interactions. Nat Neurosci 2011;14:1375–81.

44. Yap TA, de Bono JS. Targeting the HGF/c-Met axis: state of play. MolCancer Ther 2010;9:1077–9.

45. Comoglio PM, Giordano S, Trusolino L. Drug development of METinhibitors: targeting oncogene addiction and expedience. Nat RevDrug Discov 2008;7:504–16.

46. Appleman LJ. MET signaling pathway: a rational target for cancertherapy. J Clin Oncol 2011;29:4837–8.

47. Feng Y, Thiagarajan PS, Ma PC. MET signaling: novel targeted inhi-bition and its clinical development in lung cancer. J Thor Oncol2012;7:459–67.

48. Gherardi E, Birchmeier W, Birchmeier C, Vande Woude G. TargetingMET in cancer: rationale and progress. Nat Rev Cancer 2012;12:89–103.

49. Li Y, Guessous F, DiPierro C, Zhang Y, Mudrick T, Fuller L, et al.Interactions between PTEN and the c-Met pathway in glioblastomaand implications for therapy. Mol Cancer Ther 2009;8:376–85.

50. Comoglio PM, Giordano S, Trusolino L. Drug development of METinhibitors: targeting oncogene addiction and expedience. Nat RevDrug Discov 2008;7:504–16.


Joo et al.




2012;72:3828-3838. Published OnlineFirst May 22, 2012.Cancer Res Kyeung Min Joo, Juyoun Jin, Eunhee Kim, et al. MET Signaling Regulates Glioblastoma Stem Cells

Updated version

10.1158/0008-5472.CAN-11-3760doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2012/05/22/0008-5472.CAN-11-3760.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/72/15/3828.full#ref-list-1

This article cites 50 articles, 10 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/72/15/3828.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected]

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

Permissions

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

.http://cancerres.aacrjournals.org/content/72/15/3828To request permission to re-use all or part of this article, use this link



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-11-3760

http://cancerres.aacrjournals.org/content/suppl/2012/05/22/0008-5472.CAN-11-3760.DC1

http://cancerres.aacrjournals.org/content/72/15/3828.full#ref-list-1

http://cancerres.aacrjournals.org/content/72/15/3828.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/72/15/3828


met signaling regulates glioblastoma stem...

Documents