EFA6A Enhances Glioma Cell Invasion through ADP Ribosylation

Factor 6/Extracellular Signal–Regulated Kinase Signaling

Ming Li,1Samuel Sai-ming Ng,

1Jide Wang,

4Lihui Lai,

6Suet Yi Leung,

2Michel Franco,

7

Ying Peng,5Ming-liang He,

3Hsiang-fu Kung,

3and Marie Chia-mi Lin

1

1Department of Chemistry, Open Laboratory of Chemical Biology, Institute of Molecular Technology for Drug Discovery and Synthesis, and2Department of Pathology, University of Hong Kong; 3Li Ka Shing Institute of Health Sciences and Centre for Emerging InfectiousDiseases, The Chinese University of Hong Kong, Shatin, Hong Kong, China; 4Department of Gastroenterology, Nanfang Hospital;5Department of Neurology, The Second Affiliated Hospital, Sun Yat-Sen University, Guangzhou, China; 6Institute of Molecularand Chemical Biology, East China Normal University, Shanghai, China; and 7Institut de Pharmacologie Moleculaire etCellulaire, Centre National de la Recherche Scientifique-Unite Mixte Recherche, Valbonne, France

Abstract

EFA6A, or Pleckstrin and Sec7 domain protein, is a member ofguanine nucleotide exchange factors for ADP ribosylationfactor 6 (ARF6). Whereas EFA6A is specifically expressed in thebrain, little is known about its function in glial cells or glioma.Here we show that elevated EFA6A expression is detectable inboth low-grade and high-grade human glioma tissues samples.To investigate the role of EFA6A in glioma carcinogenesis, wegenerated a human glioblastoma cell line which conditionallyoverexpresses EFA6A (U373-EFA6A). We showed that over-expression of EFA6A had no effect on cell proliferation,apoptosis, or cell cycle control. However, as shown by woundhealing and in vitro cell invasion assays, it significantlyenhanced the cell motility and invasiveness whereas silencingEFA6A by its dominant negative mutant EFA6A(E242K)produced opposite effects. We further showed that ARF6/extracellular signal–regulated kinase (ERK) signaling isrequired for the EFA6A-mediated cell invasion because bothEFA6A(E242K) and ARF6 dominant negative mutantARF6(T27N) markedly reduced the phosphorylated ERK leveland EFA6A-mediated invasive capacity. Consistently, mitogen-activated protein kinase/ERK kinase inhibitor U0126 couldabolish the EFA6A-induced cell invasion. These results suggestfor the first time a potential role of EFA6A/ARF6/ERK signalcascade in glioma cell migration and invasion. (Cancer Res2006; 66(3): 1583-90)

Introduction

EFA6A, or Pleckstrin and Sec7 domain protein, belongs to theexchange factor for ADP ribosylation factor 6 (ARF6) family ofguanine nucleotide exchange factors (GEF) that are conserved inmulticellular organisms throughout evolution (1–3). The EFA6protein family comprises a catalytic Sec7 domain bearing the GDP-GTP exchange activity, a PH domain responsible for membranelocalization, and a COOH-terminal region containing a putativecoiled-coil motif for actin cytoskeleton rearrangement (2). Inhumans, the EFA6 family consists of four members (EFA6A, EFA6B,EFA6C, and EFA6D; ref. 3). At the mRNA level, EFA6A, EFA6C, and

EFA6D expression is restricted to the brain (1, 2). In contrast,EFA6B is ubiquitously expressed except in the brain. Theseexpression profiles suggest that each member of the EFA6 familyhas distinct physiologic functions in different tissues and EFA6Amay have an important function in the brain.It has been shown that overexpression of the COOH-terminal

region of EFA6A and EFA6B induced the lengthening of microvilli-like membrane protrusions in fibroblastic cell lines (3). Both ofthese exogenously expressed proteins were found to be localized tostructures enriched in polymerized actin such as large membraneruffles and microspikes on the contact-free plasma membrane (2).In nonneuronal cells, EFA6A perturbed the membrane trafficking oftransferrin, suggesting that it might coordinate endocytosis withcytoskeletal rearrangements (2, 3). Moreover, it has also been foundto regulate tight junction formation in kidney cells (4). In neuronalcells, Sakagami et al. (5) have recently shown that EFA6A isexpressed at the somatodendritic regions of the rat hippocampalneurons and is involved in the regulation of dendritic growth.However, the functional role of EFA6A in the brain remains elusive.In the present study, we aimed to investigate the function of EFA6Ain glial cells or glioma.Human gliomas are the most common primary brain tumors.

They account for >40% of all central nervous system neoplasmsand have a median survival rate of <12 months (6, 7). The highlylethal nature of this tumor results from the acquisition of aninvasive phenotype that allows the tumor cells to infiltrate thesurrounding brain tissues (8). The mechanisms for this invasiveprocess, however, are poorly understood. In this study, we firstshowed that elevated EFA6A expression was detected in a panel ofhuman glioma patient tissue samples. We then examined the roleof EFA6A in glioma cell invasion. Using the human glioblastomacell line U373 as a model, we further showed that EFA6Asignificantly enhanced cell migration and invasiveness throughthe activation of ARF6 and extracellular signal–regulated kinase(ERK), suggesting for the first time a role of EFA6A in glioma cellmovement and invasion.

Materials and Methods

Tissue samples, cell lines, and cell culture conditions. Human glioma

tissue samples were collected from Department of Pathology, Queen Mary

Hospital (University of Hong Kong). Human glioblastoma cell lines (U87,U118, U373, SW1088, SW1783, and CCF-STTG1) were obtained from

American Type Culture Collection (Manassas, VA) and maintained in MEM

(U87, U118, U373, and SW1088), Leibovitz’s L-15 medium (SW1783), or RPMI1640 (CCF-STTG1) supplemented with 10% fetal bovine serum (FBS). LN-

308 was kindly provided by Dr. M.E. Hegi (University Hospital Lausanne,

Lausanne, Switzerland). Tetracycline-free FBS (Clontech, Palo Alto, CA) was

Note: M. Li and S.S. Ng contributed equally to this work.Requests for reprints: Marie Chia-mi Lin, Department of Chemistry, The

University of Hong Kong, Kadoorie Biological Science Building, Pokfulam Road,Hong Kong, China. Phone: 852-2299-0776; Fax: 852-2817-1006; E-mail: mcllin@hkusua.hku.hk.

I2006 American Association for Cancer Research.doi:10.1158/0008-5472.CAN-05-2424

www.aacrjournals.org 1583 Cancer Res 2006; 66: (3). February 1, 2006

Research Article

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

used for selection of double Tet-On transfectants and induction of EFA6Aexpression.

Measurement of the mRNA levels by semiquantitative RT-PCR. TotalRNA was extracted from various human tissue samples and human glioma

cells lines using Trizol reagent (Invitrogen, Carlsbad, CA). RT-PCR wasdone using the Superscript Preamplification System (Invitrogen). Hotstart

PCR conditions were as follows: 45 seconds at 94jC, 30 seconds at 55jC,1 minute at 72jC for 30 cycles [EFA6A and glial fibrillary acidic protein

(GFAP)] or for 26 cycles [glyceraldehyde-3-phosphate dehydrogenase(GAPDH)]. The primer pairs for EFA6A were sense 5V-gcctgactctttcagttgtg-3Vand antisense 5V-gaagtatcgctgggagaagt-3V; for GFAP were sense 5V-gagtcgc-tggaggaggagatc-3Vand antisense 5V-gggactcgttcgtgccgcgc-3V; and for GAPDH

were sense 5V-tgcctcctgcaccaccaact-3V and antisense 5V-cccgttcagctcagg-gatga-3V.

Plasmids, chemicals, and antibodies. The cDNAs encoding the

NH2-terminal vesicular stomatitis virus glycoprotein (VSVG)–tagged EFA6A

and its dominant negative mutant EFA6A(E242K) were cloned as previously

described (2). Hemagglutinin (HA)-tagged pXS-ARF6(T27N) was kindly

provided by Dr. J.G. Donaldson (National Heart, Lung, and Blood Institute,

NIH, Bethesda, MD). pGEX-4T1-GST-GGA3 (1-26) was a generous gift from

Dr. P. Chavrier (Centre National de la Recherche Scientifique, Institut Curie,

Paris, France). The full-length cDNAs encoding the VSVG-tagged EFA6A and

HA-tagged ARF6(T27N) were subcloned into the multiple cloning sites of

pTRE2hyg plasmid (Clontech) to generate pTRE2hyg-VSVG-EFA6A and

pcDNA3.1/Zeo-HA-ARF6(T27N) expression plasmids (Invitrogen), respec-

tively. The sequences were verified by DNA sequencing. pTRE-luc control

response plasmid was obtained from Clontech. Mouse monoclonal anti-

VSVG antibody (clone P5D4) was obtained from Roche Diagnostics Corp.

(Mannheim, Germany). Rabbit monoclonal anti-HA tag antibody was

bought from NeoMarkers (Fremont, CA). Mouse monoclonal anti-p-ERK1/2

and anti-ARF6 were bought from Santa Cruz Biotechnology (Santa Cruz,

CA). Mitogen-activated protein kinase/ERK kinase (MEK) inhibitor U0126

was obtained from Cell Signaling Technology, Inc. (Beverly, MA).

Generation of doxycycline-responsive gene inducible cell line.Human glioblastoma cell line U373 was transfected with pTet-On regulator

plasmid (Clontech) using Lipofectamine 2000 (Invitrogen). After selection

by G418 (500 mg/mL) for about 3 weeks, a stable U373 Tet-On cell line was

generated. This stable cell line was then transfected with either pTRE2hyg-VSVG-EFA6A or pTRE2hyg alone. After 2 days, the transfected cells were

selected with 200 mg/mL hygromycin B (Invitrogen) in the presence of 500

mg/mL G418. Twenty stable hygromycin-resistant cell lines were screened

for the expression of EFA6A on addition of doxycycline (1 mg/mL) by RT-PCR and Western blotting. Three stable hygromycin-resistant cell lines,

U373-EFA6A, which expressed EFA6A on doxycycline induction, were

chosen for subsequent experiments. Induction of luciferase activity in thecontrol pTRE2-Luc transfectants was also confirmed by a Luciferase assay

kit (Promega, Madison, WI).

Western blot analysis. Cells were washed twice with PBS and

solubilized in radioimmunoprecipitation assay lysis buffer [50 mmol/LTris-HCl (pH 7.4), 1% NP40, 0.25% Na-deoxycholate, 150 mmol/L NaCl,

1 mmol/L EDTA, 1 mmol/L phenylmethylsulfonyl fluoride, 1 mg/mL each of

aprotinin, leupeptin, and pepstatin, 1 mmol/L Na3VO4, 1 mmol/L NaF]. The

supernatants, which contained the whole-cell protein extracts, wereobtained after centrifugation of the cell lysates at 10,000 � g for 10

minutes at 4jC. The protein concentration was determined by bicincho-

ninic acid protein assay kit (Pierce, Rockford, IL). Twenty-microgramprotein samples were loaded on an SDS-PAGE gel and analyzed by Western

blotting.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay.3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assaywas done to assess the effect of EFA6A expression on cell proliferation.

U373-pTRE2hyg or U373-EFA6A cells (5.0 � 103) were plated in each well

of a 96-well plate. The cells were cultured in growth medium with or

without 1 Ag/mL doxycycline in a total volume of 100 AL. At various timesafter doxycycline treatment, 25 AL of sterile MTT dye (5 mg/mL; Sigma,

St. Louis, MO) were added and the cells were then incubated for 4 hours

at 37jC. After incubation, the MTT solution was removed and 200 AL of

DMSO were added and thoroughly mixed for 30 minutes. Spectrometricabsorbance was measured on a microplate reader at a wavelength of 570

nm with background subtraction at 660 nm (Spectra Max 340, Molecular

Devices, Sunnyvale, CA).

Cell cycle analysis. U373-pTRE2hyg and U373-EFA6A Tet-On cell lineswere cultured in medium with or without doxycycline for 2 days, harvested,

and washed with ice-cold PBS containing 0.1% glucose. The cells were then

fixed with 70% ethanol for 1 hour and incubated in 1 mL of PBS containing

50 mL/mL of propidium iodide (Sigma) and 66 units/mL RNase (Invitrogen)on ice for 30 minutes. DNA content analysis was done by FACScan with

CellQuest software (Becton Dickinson, San Jose, CA).

Wound healing assay. In vitro wound healing assay was carried out to

investigate the formation of membrane protrusion and cell migration. Equalnumbers of U373-pTRE2hyg or U373-EFA6A cells (1.0 � 105) were seeded

into six-well tissue culture plates. When the confluence reached 90%, a

single wound was created in the center of the cell monolayer by gentlyremoving the attached cells with a sterile plastic pipette tip. The debris was

removed by washing the cells with serum-free medium. Migration of the

cells into the wound was then observed at different time points. Cells that

migrated into the wounded area or cells with extended protrusion from theborder of the wound were visualized and photographed under an inverted

microscope. A total of nine areas were selected randomly from each well

under a 40� objective and the cells in three wells of either group were

quantified in each experiment.In vitro invasion assay. Cell invasiveness in vitro was reflected by the

ability of the cell to transmigrate a layer of extracellular matrix in Biocoat

Matrigel Invasion Chambers (Becton Dickinson Labware, Bedford, MA).U373-pTRE2hyg or U373-EFA6A cells were plated at a density of 3.0 � 104

per insert. Medium with 10% FBS was added to the lower chamber as a

chemoattracctant. For induction of EFA6A expression, doxycycline (1 mg/

mL) was added to the lower chamber. After incubation for 22 hours, cells onthe upper surface of the membrane were removed. Invasive cells which had

the ability to push themselves through the 8-Am pores and grow on the

lower surface were fixed with 100% methanol and stained with 1% toluidine

blue (Sigma) before counting under an inverted microscope (Leica, Solms,Germany). In all experiments, data were collected from triplicate chambers.

Glutathione S -transferase pull-down assay. The glutathione

S-transferase (GST) pull-down assay was done as described by Niederganget al. (9) by using a bait a fragment (1-226) of the Golgi-localized g ear–

containing ARF-binding protein 3 (GGA3) fused to GST. After washing

thrice in ice-cold PBS, the cells were lysed in 50 mmol/L Tris-HCl (pH 8.0),

100 mmol/L NaCl, 10 mmol/L MgCl2, 1% Triton X-100, 0.05% sodiumcholate, 0.005% SDS, 10% glycerol, 2 mmol/L DTT, and protease inhibitors

(Sigma). Lysates were incubated with 20 Ag of GST or 40 Ag of GST-GGA3(1-226) bound to glutathione-sepharose beads (Amersham Pharmacia Biotech,

Uppsala, Sweden). After 1 hour, the beads were washed thrice in lysis bufferand the proteins were eluted by boiling in 30 AL of sample buffer. Equal

amounts of proteins of each sample were analyzed by immunoblot by using

an anti-ARF6 antibody.

Statistical analysis. Results are expressed as the mean F SD. Statisticalanalyses were done by Student’s t test. P < 0.05 was considered statistically

significant.

Results

Expression profiles of EFA6A in glioma patient tissuesamples and cell lines. We first analyzed the mRNA expressionlevels of EFA6A in 9 low-grade glioma samples (grades 1-3), 15high-grade glioma tissue samples (grade 4), and 7 glioma cell lines(U87, U118, U373, SW1088, SW1783, LN-308, and CCF-STTG1) bysemiquantitative RT-PCR. As GFAP is a well-known marker ofastroglia in the brain, both GAPDH and GFAP were used as internalcontrols to compare the EFA6A expression levels in various gliomatissues. Representative RT-PCR results are presented in Fig. 1. Wefound that EFA6A mRNA level was up-regulated in 4 of 9 low-gradeglioma samples and in 9 of 15 high-grade glioma tissues (Fig. 1A

Cancer Research

Cancer Res 2006; 66: (3). February 1, 2006 1584 www.aacrjournals.org

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

and B). Among the seven glioma cell lines examined, EFA6A wasexpressed in five of them (U87, U118, U373, CCF-STTG1, andLN-308 cells).Establishment of stable transfectants of EFA6A under the

control of the doxycycline-responsive promoter. To investigatethe function of EFA6A in glioma carcinogenesis, a gain-of-functionstudy was done by generating a stable human glioblastoma U373-EFA6A cell line which conditionally overexpresses EFA6A oninduction by doxycycline. U373 Tet-On cells were first transfectedwith a plasmid encoding VSVG-tagged EFA6A (pTRE2hyg-VSVG-EFA6A) or a control empty plasmid (pTRE2hyg) and then selectedwith hygromycin B. Twenty stable hygromycin-resistant cell clones

were expanded and screened for efficient gene induction bydoxycycline (1 Ag/mL). Western blotting analysis using a mono-clonal antibody targeting the VSVG-tagged EFA6A showed thatthree of these stable cell lines, designated as U373-EFA6A, wereable to express a high level of EFA6A when cultured withdoxycycline. Results from the one which displayed the highestlevel of EFA6A expression on doxycycline induction are shown(Fig. 2A). VSVG-EFA6A expression in U373-EFA6A cells could bemaintained up to 4 days after doxycycline induction (Fig. 2B).Overexpression of EFA6A has no effect on cell proliferation

and cell cycle. We then studied whether EFA6A has tumor-relatedfunction by analyzing the cell proliferation and cell cycle using

Figure 1. Analysis of EFA6A expression levels in humanglioma tissues and glioblastoma/astrocytoma cell linesby semiquantitative RT-PCR. Representative results from9 low-grade glioma tissues (A ), 15 high-grade gliomatissues (B ), 7 human glioblastoma cell lines (U87, U118,U373, SW1088, SW1783, CCF-STTG1, and LN-308; C ).N1, N2, and N3, three representative adjacent normalbrain tissue samples. GFAP and GAPDH were donesimultaneously as internal controls for equal loading of thePCR products. The relative band intensities weremeasured by densitometry and normalized against therespective GFAP signals.

EFA6A Enhances Glioma Cell Invasion

www.aacrjournals.org 1585 Cancer Res 2006; 66: (3). February 1, 2006

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

MTT assay and flow cytometry, respectively. We found that the cellproliferation rate of U373-pTRE2hyg and U373-EFA6A cellsremained unchanged in the presence or absence of doxycyclineduring a time course of 4 days (Fig. 2C). In addition, lack of a pre-G1 peak or changes in the cell cycle profile of the doxycycline-treated U373-EFA6A cells indicated that no induction of apoptosiswas associated with EFA6A overexpression (Fig. 2D).Expression of EFA6A enhances U373 cell invasion and

migration. During metastasis, invasion of basement membraneby tumor cells is thought to be a critical event (10). To studywhether expression of EFA6A is associated with cell migration,wound healing assay was done. After doxycycline induction (1 Ag/mL) for 22 hours, the migration of U373-EFA6A cells was markedlyincreased whereas the control U373-pTRE2hyg cells did not showany significant changes (Fig. 3A and B).We then tested the effect of EFA6A on the invasion of U373 cells

by using an in vitro cell invasion assay. U373-EFA6A and the controlU373-pTRE2hyg cells were first treated with or without doxycy-cline. The cells were then placed in the upper chamber of transwellfilters containing a thin layer of reconstituted extracellular matrix(Matrigel) with 8-Am pores and allowed to migrate for 22 hours.Our results showed that the number of cells which penetrated the

Matrigel increased by >100% in the EFA6A-expressing cells ascompared with the control cells (Fig. 4A and B).ARF6 is a downstream effector of EFA6A for promoting

glioma cell invasion. It has been well established that ARF6 is aspecific downstream effector of EFA6A in vitro (2–5, 11). To studywhether glioma invasion stimulated by EFA6A acts via ARF6, apcDNA3.1/Zeo expression plasmid encoding HA-tagged ARF6 GTP-binding defective dominant negative mutant, pcDNA3.1/Zeo-HA-ARF6(T27N), was transfected into U373-EFA6A cells. A controltransfection was also done separately with pcDNA3.1/Zeo plasmidalone. After selection with zeocin (500 Ag/mL), the expression ofthe ARF6(T27N) protein was verified by Western blotting (Fig. 5A).As shown in Fig. 5B , ARF6(T27N) significantly suppressed theinvasiveness of U373-EFA6A cells, suggesting that ARF6 is anessential downstream effector of EFA6A for promoting glioma cellinvasion.EFA6A regulates ERK1/2 activation in U373 cell through

ARF6 signaling. It has been reported that ERK-dependentsignaling pathway is involved in glioma invasion (12, 13). Recently,Tague et al. (14) reported that ARF6 regulates melanoma cellinvasion through the activation of the MEK/ERK signalingpathway. Therefore, we also tested if EFA6A can regulate the

Figure 2. Effect of overexpression of EFA6A on gliomacell proliferation and cell cycle. A, Western blot analysis ofEFA6A expression using an antibody against VSVG inU373-pTRE2hyg and U373-EFA6A cells after inductionwith (+) or without (�) doxycycline (Dox ) for 24 hours.h-Actin was used as an internal control. B, expressionlevels of VSVG-EFA6A on doxycycline induction (1 mg/mL)during the 4-day experimental period. C, after addition ofdoxycycline, cell proliferation was monitored by MTT assayat various time points as indicated. MTT assay was donein quadruplicates and the mean absorbance values at570 nm were measured. 5 and n, U373-pTRE2hyg cellstreated with (+) or without (�) doxycycline, respectively.4 and E, U373-EFA6A cells treated with (+) or without (�)doxycycline, respectively (n = 6). D, cell cycle analysis bypropidium iodide staining assay. a and b, U373-pTRE2hygcells cultured in the absence (�) or presence (+) ofdoxycycline, respectively. c and d, U373-EFA6A cellscultured in the absence (�) or presence (+) of doxycycline,respectively (n = 6).

Cancer Research

Cancer Res 2006; 66: (3). February 1, 2006 1586 www.aacrjournals.org

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

invasiveness of U373 cells through the ERK1/2 pathway. Wefound that overexpression of EFA6A significantly increased thelevel of phosphorylated ERK1/2 (p-ERK1/2) in U373-EFA6A cells(Fig. 5C ). However, expression of the dominant negativeARF6(T27N) significantly inhibited the EFA6A-mediated ERK1/2activation. Both EFA6A and ARF6(T27N) expression had no effecton the total ERK1/2 levels.Inhibition of ERK signaling blocks U373 cell invasion

induced by EFA6A. We also investigated whether inhibitingERK1/2 phosphorylation can prevent EFA6A-induced glioma cellinvasion. Because MEK1/2 is an upstream activator of ERK1/2 (15),inhibition of MEK1/2 will also block ERK1/2 activation and hencethe cell invasion induced by EFA6A. To test this hypothesis, wepretreated U373-EFA6A cells with a specific MEK inhibitor (U0126;20 Amol/L) for 1 hour, and then exposed the cells to doxycycline(1 Ag/mL) in the presence of U0126 for 24 hours. Inactivation of

MEK abolished EFA6A-induced ERK1/2 activation (Fig. 5D, left).Importantly, MEK inhibitor U0126 also blocked U373 cell invasioninduced by EFA6A expression (Fig. 5D, right). Thus, our data showthat EFA6A regulates the invasion of U373 cells through ARF6/MEK/ERK signaling.The GEF-defective mutant EFA6(E242K) inhibited U373

invasion. Finally, we determined whether the dominant negativemutant EFA6A(E242K), which loses GEF activity for ARF6 (2),could inhibit glioma cell invasion. The pcDNA3.1 expressionplasmid alone or pcDNA3.1 carrying the GEF-defective mutantEFA6(E242K) was transfected into U373 cells. After selection withG418 (500 Ag/mL) for about 2 weeks, expression of EFA6A(E242K)was verified by Western blotting. To check whether the expressionof EFA6A(E242K) can lead to the loss of GEF activity in U373 cells,we also assessed the activation level of endogenous ARF6 byusing a pull-down assay based on the association of GTP-bound

Figure 3. Overexpression of EFA6A stimulates U373 cell migration. A, 22 hours after making the wounds, cells with extended membrane protrusion moved intothe wounded areas. a and b, U373-pTRE2hyg cells treated without (�) or with (+) doxycycline, respectively. c and d, U373-EFA6A cells treated without (�) or with (+)doxycycline, respectively. B, quantitative results for the migration of U373-EFA6A and U373-pTRE2hyg cells (n z 3). *, P < 0.05, statistically significant differencebetween U373-EFA6A cells and the control U373-pTRE2hyg cells. (�), Dox; (+), Dox.

Figure 4. Overexpression of EFA6A enhances the invasion of U373 cells. A, invasive U373 cells were able to invade through the Matrigel. a and b, U373-pTRE2hygcells treated without (�) or with (+) doxycycline, respectively. c and d, U373-EFA6A cells treated without (�) or with (+) doxycycline, respectively. B, quantitativeresults for the transmembrane ability of U373 cells (n z 3). *, P < 0.05, statistically significant difference between U373-EFA6A cells and the control U373-pTRE2hygcells. (�), Dox; (+), Dox.

EFA6A Enhances Glioma Cell Invasion

www.aacrjournals.org 1587 Cancer Res 2006; 66: (3). February 1, 2006

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

ARF6 and the GGA3 (9). In U373 cells expressing VSVG-EFA6A(E242K), the level of ARF6-GTP was significantly lower than thatof the control cells (Fig. 6A). Consistently, the expression ofEFA6A(E242K) also reduced the phosphorylated level of ERK1/2(Fig. 6B). The invasive ability of the pooled clones was alsoanalyzed by the in vitro cell invasion assay. As shown in Fig. 6C ,EFA6A(E242K) markedly inhibited U373 cell invasion by >4-fold.These results further verify that ARF6 is a downstream effector ofEFA6A and the GEF activity of EFA6A is essential for the U373 cellinvasion.

Discussion

In the present study, we found that elevated EFA6A mRNA levelis detectable both in low-grade and high-grade glioma tissuesamples. In vitro studies showed that EFA6A overexpression has noeffect on U373 cell proliferation, apoptosis, or cell cycle control butit significantly enhances cell motility and invasiveness. We furthershowed that ARF6 and ERK are essential for the EFA6A-stimulatedcell invasion. Taken together, our data provide the first evidencethat the EFA6A/ARF6/ERK signaling pathway is involved in gliomacell migration and invasion (Fig. 6D).Tumor cell invasion involves complex interactions between the

normal and malignant cells. It has been well established that thisdynamic process requires the concerted effects of various moleculesincluding proteolytic enzymes, growth factors, adhesion molecules,and extracellular matrix molecules (16, 17). However, the mecha-

nisms of glioma cell motility remain unclear. Studies in other celltypes, notably fibroblasts, have revealed that cell motility is depend-ent on the dynamic remodeling of the actin cytoskeleton (18). Tothis end, GTPases of the Rho family have been implicated asregulators for cell motility (19). More recently, it has become clearthat members of the ARF GTPase family also regulate cytoskeletalassembly andmay integrate with Rho proteins in this process (20, 21).EFA6A is mainly expressed in the brain and elevated expression

can be detected in f50% of the human glioma tissues tested. Wealso found that 5 of 7 (71.4%) glioma cell lines examined haveEFA6A expression. These findings led us to hypothesize that EFA6Amay play a role in glioma carcinogenesis. Our data showed thatoverexpression of EFA6A can promote glioma cell invasion and itseems that EFA6A exerts its effect by activating the ARF6/ERKpathway because blocking this pathway by a dominant negativemutant ARF6(T27N), MEK inhibitor U0126, or EFA6A(E242K) canabolish the cell invasiveness induced by EFA6A.Structurally, EFA6A comprises a catalytic Sec7 domain, a PH

domain, and a COOH-terminal coiled-coil motif. Each conserveddomain has its own function. Among the three conserved domains,the Sec7 domain bears the GDP-GTP exchange activity for ARF6(3), which is a member of the ARF family comprising the Ras-related, low molecular weight (f20 kDa) GTP-binding proteinsexpressed in all eukaryotes (22, 23). In mammals, there are sixARFs and many more ARF-like proteins. Like most GTPases,ARF6 alternates between its active GTP-bound and inactive GDP-bound conformations via activation by GEFs (24). This ARF6 small

Figure 5. EFA6A promotes glioblastoma U373 cells invasion through ARF6/ERK pathway. A, U373-EFA6A cells were transfected with either pcDNA3.1/Zeo-HA-ARF6(T27N) or pcDNA3.1/Zeo empty plasmid and then selected by zeocin. The expression of ARF6(T27N) protein in the pcDNA3.1/Zeo-HA-ARF6(T27N)–transfectedcells was validated by Western blotting using an antibody against HA. B, ARF6(T27N) significantly inhibited the U373 cell invasion induced by EFA6A (n z 3).

(�), Dox; (+), Dox. C, EFA6A activated ERK1/2 phosphorylation, which was inhibited by ARF6(T27N). Western blotting for total ERK1/2 was done in parallelas a control for equal loading of proteins. D, MEK inhibitor abolished glioblastoma U373 cells invasion induced by EFA6A. U373-EFA6A cells were pretreated withMEK inhibitor U0126 for 1 hour at the concentration of 20 Amol/L, and then the treated cells were exposed to doxycycline (1 Ag/mL) in the presence of U0126 for24 hours. Left, the expression of VSVG-tagged EFA6A protein U373-EFA6A cells was validated by Western blotting using an antibody against VSVG. ERK1/2phosphorylation was analyzed by Western blotting using an antibody against phosphorylated ERK1/2. Western blotting for total ERK1/2 was done in parallel as a controlfor equal loading of proteins. Right , the effect of U0126 on U373 cell invasion induced by EFA6A was studied by in vitro Matrigel transwell invasion assay asdescribed in Materials and Methods (n z 3). * and #, P < 0.05, statistically significant difference between the doxycycline-treated and non-doxycycline-treatedU373-EFA6A cells and between the doxycycline-treated and doxycycline/U0126–treated U373-EFA6A cells, respectively.

Cancer Research

Cancer Res 2006; 66: (3). February 1, 2006 1588 www.aacrjournals.org

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

G-protein cycle has been shown to be important for processescontrolling cell shapes and invasion, including endosome mem-brane trafficking, exocytosis, and actin rearrangements at the cellsurface (20, 25, 26).The molecular mechanism by which EFA6A promotes cell

invasion is not fully understood. Previous studies in fibroblasticcells have shown that activation of ARF6 by ARF nucleotidebinding site opener, another GEF for ARF6 and ARF1 (27),stimulates epithelial cell migration through downstream activationof both Rac1 and phospholipase D (21, 28). In the case of Madin-Darby canine kidney cells, this transformation requires the dualactivation of both Rac1 and phospholipase D, which presumablyfunction together to regulate actin structure and cell motility (21).In this study, we showed that EFA6A-mediated activation of ARF6/ERK is responsible, at least in part, for the migration and invasionof glioma cells. The existence of other mechanisms and thepotential cross-talks between various pathways in the preciouscontrol of glioma cell migration and invasion remain to bedetermined.Activation of ERK1/2 has been shown in other systems to be the

mechanism for promoting the production of matrix metallopro-teinases (MMP), such as MMP1 and MMP3 (29), which areimportant for cell proliferation, invasion, and neovascularization.With reference to our data in glioma U373 cells, blocking of theERK1/2 pathways by MEK1/2 inhibitor U0126 prevents cellinvasion induced by EFA6A/ARF6 signaling. This observationsuggests that activation of ERK1/2 is also responsible for gliomacell invasiveness. It has been reported that ERK1/2 directlyregulates the transcriptional activation of MMP1 and MMP3 (30).Supporting this notion, Westermarck et al. (31) showed that theactivation of ERK1/2 induces activator protein-1 transcription,

which in turn stimulates MMP1 promoter activity. Thus, it ispossible that EFA6A may enhance glioma cell invasion bystimulating the production of MMPs via activation of ERK1/2.Because ARF6 is a downstream effector for EFA6A and the

number of genes coding for ARF GEFs and ARF GTPase-activatingproteins in humans is significantly greater than that of genescoding for the ARF isoforms and ARF-like factors (23, 32), ARF6might be a bona fide factor for tumor invasion. In agreement withthis, Hashimoto et al. (33) recently reported that ARF6 is requiredfor breast cancer invasion and there is a direct correlation betweenthe level of ARF6 protein (but not mRNA level) and the invasivecapacity of the breast cancer cell lines. A more recent article alsoreported that ARF6 enhances melanoma cell invasion through theactivation of the MEK/ERK signaling pathway and that the ARF6GTPase cycle regulates ERK1/2 activation (14). Our data showingthat the EFA6A-stimulated glioma cell invasion requires theactivation of ARF6 and ERK1/2 further support the role ofEFA6A/ARF6/ERK signaling pathway in cancer cell invasion. Inthis regard, EFA6A, as well as its downstream effectors includingARF6, may be considered as a potential therapeutic target forpreventing tumor invasion and metastasis.

Acknowledgments

Received 7/11/2005; revised 10/27/2005; accepted 12/1/2005.Grant support: AoE scheme of University Grants Commission; Hong Kong

Research Grant Council grants HKU 7243/02 (M.C. Lin) and CUHK 7191/01M (H.F.Kung); Li Ka Shing Institute of Health Sciences (H.F. Kung); and ShanghaiMetropolitan Fund for Research and Development grant 04JC14096.

The costs of publication of this article were defrayed in part by the payment of pagecharges. This article must therefore be hereby marked advertisement in accordancewith 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. J.G. Donaldson and P. Chavrier for generously providing the cDNAs,and Dr. M.E. Hegi for the kind gift of LN-308 cell line.

Figure 6. The GEF-defective mutant EFA6A(E242K) inhibits U373invasion. A, U373 cells were transfected with the pcDNA3.1-VSVG-EFA6A(E242K) plasmid or pcDNA3.1 empty plasmid and then selected byG418 (500 mg/mL). The expression of EFA6A protein in pcDNA3.1-VSVG-EFA6A(E242K)–transfected U373 cells was validated by Western blottingusing an antibody against VSVG. The GST-GGA3(1-226) fusion proteinwas used to pull down ARF6-GTP in U373 cells expressing VSVG-EFA6A(E242K), with GST alone as a control. B, the effect ofEFA6A(E242K) on ERK1/2 phosphorylation was analyzed by Westernblotting using an antibody against phosphorylated ERK1/2. Westernblotting for total ERK1/2 was done in parallel as a control for equal loadingof proteins. C, effect of EFA6A(E242K) on U373 cell invasion was studiedby in vitro Matrigel transwell invasion assay as described in Materialsand Methods (n z 3). *, P < 0.05, statistically significant difference betweenthe pcDNA3.1-VSVG-EFA6A–transfected cells and the control pcDNA3.1–transfected cells. D, schematic diagram showing the proposed EFA6A/ARF6/ERK signal cascade in glioma cell migration and invasion. Pointedarrows, stimulation; blunted arrows, inhibition.

EFA6A Enhances Glioma Cell Invasion

www.aacrjournals.org 1589 Cancer Res 2006; 66: (3). February 1, 2006

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

Cancer Research

Cancer Res 2006; 66: (3). February 1, 2006 1590 www.aacrjournals.org

References

1. Perletti L, Talarico D, Trecca D, et al. Identification of anovel gene, PSD, adjacent to NFKB2/lyt-10, whichcontains Sec7 and pleckstrin-homology domains.Genomics 1997;46:251–9.

2. Franco M, Peters PJ, Boretto J, et al. EFA6, a sec7domain-containing exchange factor for ARF6, coordi-nates membrane recycling and actin cytoskeletonorganization. EMBO J 1999;18:1480–91.

3. Derrien V, Couillault C, Franco M, et al. A conserved C-terminal domain of EFA6-family ARF6-guanine nucleo-tide exchange factors induces lengthening of microvilli-like membrane protrusions. J Cell Sci 2002;115:2867–79.

4. Luton F, Klein S, Chauvin JP, et al. EFA6, exchangefactor for ARF6, regulates the actin cytoskeleton andassociated tight junction in response to E-cadherinengagement. Mol Biol Cell 2004;15:1134–45.

5. Sakagami H, Matsuya S, Nishimura H, Suzuki R, KondoH. Somatodendritic localization of the mRNA for EFA6A,a guanine nucleotide exchange protein for ARF6, in rathippocampus and its involvement in dendritic forma-tion. Eur J Neurosci 2004;19:863–70.

6. Feldman AR, Kessler L, Myers MH, Naughton MD. Theprevalence of cancer. Estimates based on the Connect-icut Tumor Registry. N Engl J Med 1986;315:1394–7.

7. Deen DF, Chiarodo A, Grimm EA, et al. Brain TumorWorking Group Report on the 9th InternationalConference on Brain Tumor Research and Therapy.Organ System Program, National Cancer Institute.Neurooncol 1993;16:243–72.

8. Chintala SK, Rao JS. Invasion of human glioma: roleof extracellular matrix proteins. Front Biosci 1996;1:d324–39.

9. Niedergang F, Colucci-Guyon E, Dubois T, Raposo G,Chavrier P. ADP ribosylation factor 6 is activated andcontrols membrane delivery during phagocytosis inmacrophages. J Cell Biol 2003;161:1143–50.

10. Nakada M, Okada Y, Yamashita J. The role of matrixmetalloproteinases in glioma invasion. Front Biosci2003;8:e261–9.

11. Macia E, Chabre M, Franco M. Specificities for thesmall G proteins ARF1 and ARF6 of the guaninenucleotide exchange factors ARNO and EFA6. J BiolChem 2001;276:24925–30.

12. Lakka SS, Jasti SL, Kyritsis AP, et al. Regulation ofMMP-9 (type IV collagenase) production and invasive-ness in gliomas by the extracellular signal-regulatedkinase and jun amino-terminal kinase signaling cas-cades. Clin Exp Metastasis 2000;18:245–52.

13. Lakka SS, Jasti SL, Gondi C, et al. Down-regulation ofMMP-9 in ERK-mutated stable transfectants inhibitsglioma invasion in vitro . Oncogene 2002;21:5601–8.

14. Tague SE, Muralidharan V, D’Souza-Schorey C. ADP-ribosylation factor 6 regulates tumor cell invasionthrough the activation of the MEK/ERK signalingpathway. Proc Natl Acad Sci U S A 2004;101:9671–6.

15. Cobb MH. MAP kinase pathways. Prog Biophys MolBiol 1999;71:479–500.

16. Friedl P, Wolf K. Tumour-cell invasion and migration:diversity and escape mechanisms. Nat Rev Cancer 2003;3:362–74.

17. Rao JS. Molecular mechanisms of glioma invasive-ness: the role of proteases. Nat Rev Cancer 2003;3:489–501.

18. Maidment SL. The cytoskeleton and brain tumourcell migration. Anticancer Res 1997;17:4145–9.

19. Schmitz AA, Govek EE, Bottner B, Van Aelst L. RhoGTPases: signaling, migration, and invasion. Exp CellRes 2000;261:1–12.

20. Boshans RL, Szanto S, van Aelst L, D’Souza-SchoreyC. ADP-ribosylation factor 6 regulates actin cytoskeletonremodeling in coordination with Rac1 and RhoA. MolCell Biol 2000;20:3685–94.

21. Santy LC, Casanova JE. Activation of ARF6 by ARNOstimulates epithelial cell migration through downstreamactivation of both Rac1 and phospholipase D. J Cell Biol2001;154:599–610.

22. Al-Awar O, Radhakrishna H, Powell NN, DonaldsonJG. Separation of membrane trafficking and actinremodeling functions of ARF6 with an effector domainmutant. Mol Cell Biol 2000;20:5998–6007.

23. Donaldson JG. Multiple roles for Arf6: sorting,structuring, and signaling at the plasma membrane.J Biol Chem 2003;278:41573–6.

24. Vetter IR, Wittinghofer A. The guanine nucleotide-binding switch in three dimensions. Science 2001;294:1299–304.

25. Chavrier P, Goud B. The role of ARF and Rab GTPasesin membrane transport. Curr Opin Cell Biol 1999;11:466–75.

26. Palacios F, Price L, Schweitzer J, Collard JG, D’Souza-Schorey C. An essential role for ARF6-regulatedmembrane traffic in adherens junction turnover andepithelial cell migration. EMBO J 2001;20:4973–86.

27. Frank S, Upender S, Hansen SH, Casanova JE. ARNOis a guanine nucleotide exchange factor for ADP-ribosylation factor 6. J Biol Chem 1998;273:23–7.

28. Palacios F, D’Souza-Schorey C. Modulation of Rac1and ARF6 activation during epithelial cell scattering.J Biol Chem 2003;278:17395–400.

29. Westermarck J, Holstorm TH, Ahonen M, Eriksson JE,Kahari VM. Enhancement of fibroblast collagenase-1(MMP-1) gene expression by tumor promoter okadaicacid is mediated by stress-activated protein kinasesJun N-terminal kinase and p38. Matrix Biol 1998;17:547–57.

30. Reunanen N, Li SP, Ahonen M, Foschi M, Han J,Kahari VM. Activation of p38-a MAPK enhancescollagenase-1 (matrix metalloproteinase (MMP)-1) andstromelysin-1 (MMP-3) expression by mRNA stabiliza-tion. J Biol Chem 2002;277:32360–8.

31. Westermarck J, Li SP, Kallunki T, Han J, Kahari VM.p38 mitogen-activated protein kinase-dependent acti-vation of protein phosphatases 1 and 2A inhibits MEK1and MEK2 activity and collagenase 1 (MMP-1) geneexpression. Mol Cell Biol 2001;21:2373–83.

32. Jackson CL, Casanova JE. Turning on ARF: the Sec7family of guanine-nucleotide-exchange factors. TrendsCell Biol 2000;10:60–7.

33. Hashimoto S, Onodera Y, Hashimoto A, et al.Requirement for Arf6 in breast cancer invasive activities.Proc Natl Acad Sci U S A 2004;101:6647–52.

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from

2006;66:1583-1590. Cancer Res   Ming Li, Samuel Sai-ming Ng, Jide Wang, et al.   Signaling

Regulated Kinase−Ribosylation Factor 6/Extracellular Signal EFA6A Enhances Glioma Cell Invasion through ADP

  Updated version

  http://cancerres.aacrjournals.org/content/66/3/1583

Access the most recent version of this article at:

   

   

  Cited articles

  http://cancerres.aacrjournals.org/content/66/3/1583.full#ref-list-1

This article cites 33 articles, 17 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/66/3/1583.full#related-urls

This article has been cited by 12 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

  .pubs@aacr.orgDepartment at

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

  Permissions

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

.http://cancerres.aacrjournals.org/content/66/3/1583To request permission to re-use all or part of this article, use this link

Research. on May 18, 2020. © 2006 American Association for Cancercancerres.aacrjournals.org Downloaded from