expression of a truncated active form of vdac1 in lung ... · 2017. 4. 13. · m. christiane...

Tumor and Stem Cell Biology

Expression of a Truncated Active Form of VDAC1 in LungCancer Associates with Hypoxic Cell Survival and Correlateswith Progression to Chemotherapy Resistance

M. Christiane Brahimi-Horn1, Danya Ben-Hail7, Marius Ilie2,3, Pierre Gounon4, Matthieu Rouleau5,V�eronique Hofman2,3, J�erome Doyen1, Bernard Mari6, Varda Shoshan-Barmatz7, Paul Hofman2,3,Jacques Pouyss�egur1, and Nathalie M. Mazure1

AbstractResistance to chemotherapy-induced apoptosis of tumor cells represents a major hurdle to efficient cancer

therapy. Although resistance is a characteristic of tumor cells that evolve in a low oxygen environment (hypoxia),the mechanisms involved remain elusive. We observed that mitochondria of certain hypoxic cells take on anenlarged appearance with reorganized cristae. In these cells, we found that a major mitochondrial proteinregulating metabolism and apoptosis, the voltage-dependent anion channel 1 (VDAC1), was linked to chemore-sistance when in a truncated (VDAC1-DC) but active form. The formation of truncated VDAC1, which had asimilar channel activity and voltage dependency as full-length, was hypoxia-inducible factor-1 (HIF-1)-dependentand could be inhibited in the presence of the tetracycline antibiotics doxycycline and minocycline, knowninhibitors of metalloproteases. Its formation was also reversible upon cell reoxygenation and associated with cellsurvival through binding to the antiapoptotic protein hexokinase. Hypoxic cells containing VDAC1-DC were lesssensitive to staurosporine- and etoposide-induced cell death, and silencing of VDAC1-DC or treatment with thetetracycline antibiotics restored sensitivity. Clinically, VDAC1-DC was detected in tumor tissues of patients withlung adenocarcinomas andwas foundmore frequently in large and late-stage tumors. Together, ourfindings showthat via induction of VDAC1-DC, HIF-1 confers selective protection from apoptosis that allows maintenance ofATP and cell survival in hypoxia. VDAC1-DC may also hold promise as a biomarker for tumor progression inchemotherapy-resistant patients. Cancer Res; 72(8); 2140–50. �2012 AACR.

IntroductionIt is well established that cells exposed to the limiting oxygen

microenvironment (hypoxia) of tumors acquire resistance tochemotherapy-induced apoptosis (1). However, the mechan-isms involved and the implication of the key factor of thehypoxic response, the hypoxia-inducible factor (HIF), have notbeen extensively investigated (2). We recently reported that

several types of cancer cells exposed to a hypoxic microenvi-ronment showed enlarged mitochondria with reorganizedcristae; a result of modifications to fusion/fission (3). Inaddition, we showed that these cells were resistant to chemo-therapy-induced apoptosis.

Because mitochondria regulate both metabolism and apo-ptosis (4–6) and that fusion/fission participates in apoptosis(7), we investigated whether certain mitochondrial proteinsimplicated in these processes play a role in resistance toapoptosis in a HIF-dependent manner.

The voltage-dependent anion channel (VDAC) regulatesmitochondrial import and export of Ca2þ and metabolitesincluding ATP and NADH and interacts with antiapoptoticproteins such as Bcl-2 andhexokinase in controlling the releaseof cytochrome c (8–11). Of note, screening by RNA interferenceidentified VDAC1 as a protein implicated in resistance tocisplatin-induced cell death (12). Inmammals VDAC is presentin 3 highly homologous isoforms: VDAC1, VDAC2, and VDAC3,andVDAC1 is composed of 19 amphipathicb strands that forma b barrel and of amobile N-terminala helix, located inside thepore (13). Through binding to VDAC1 hexokinase, the enzymethat catalyzes the first step of glycolysis is optimally positionedfor ATP capture (14) and hexokinase expression is increaseby HIF (15). Thus, these interactions influence the function of

Authors' Affiliations: 1Institute of Developmental Biology and CancerResearch, University of Nice, CNRS-UMR 6543, Centre Antoine Lacas-sagne; 2Human Tissue Biobank Unit/CRB and Laboratory of Clinical andExperimental Pathology, 3EA 4319, Faculty of Medicine, 4Centre Communde Microscopie Appliqu�ee, 5INSERM U898, University of Nice SophiaAntipolis, Nice; 6Institute of Molecular and Cellular Pharmacology,CNRS-UMR 6097, Valbonne, France; and 7Department of Life Sciencesand National Institute for Biotechnology in the Negev, Ben-Gurion Univer-sity of the Negev, Negev, Israel

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

Corresponding Authors: Nathalie M. Mazure, University of Nice, CentreAntoine Lacassagne, 33 avenue de valombrose, Nice 06189, France. Phone:0492031230; Fax: 0492031235; E-mail: [email protected] andM. Christiane Brahimi-Horn. E-mail: [email protected]

doi: 10.1158/0008-5472.CAN-11-3940

�2012 American Association for Cancer Research.

CancerResearch

Cancer Res; 72(8) April 15, 20122140

on April 12, 2017. © 2012 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst March 2, 2012; DOI: 10.1158/0008-5472.CAN-11-3940

http://cancerres.aacrjournals.org/

both hexokinase and VDAC in cell death and metabolism.However, the role of VDAC in metabolism and apoptosis inhypoxia is not known.

Materials and MethodsCell cultureLS174, PC3, HeLa, 786-O, SKMel, and A549 cells were grown

in Dulbecco's Modified Eagle's Medium (Gibco-BRL) supple-mented with 5% or 10% inactivated FBS as appropriate inpenicillin G (50 U/mL) and streptomycin sulfate (50 mg/mL).Dr. van de Wetering provided LS174 cells expressing thetetracycline repressor. A Bug-Box anaerobic workstation (Rus-kinn Technology Biotrace International Plc.) set at 1% oxygen,94% nitrogen, and 5% carbon dioxide was used for hypoxia.

Transfection of short interfering RNAThe 21-nucleotide RNAs were synthesized (Eurogentec).

siRNA sequences targeting SIMA (siCtl), and HIF-1a weredescribed previously (16). The short interfering RNA (siRNA)sequences targeting human VDAC1, VDAC2, VDAC3, andhexokinase II are given in the Supplementary Materials andMethods. HeLa cells were transfected with 40 nmol/L of siRNA24 hours before normoxia or hypoxia, as described (3).

Reconstitution of purified VDAC1 and VDAC1-DC into aplanar lipid bilayer, single-channel current recordingand data analysisVDAC1 and VDAC1-DC were purified from hypoxic HeLa

cells after solubilizing in lauryldimethylamine-oxide (LDAO)and chromatography on hydroxyapetite, as described (17).Elution with increasing Pi concentrations separated VDAC1and VDAC1-DC. The fractions containing either VDAC1 orVDAC1-DC were used for channel reconstitution into a planarlipid bilayer (PLB). A PLBwas prepared from soybean asolectindissolved in n-decane (50 mg/mL) and purified VDAC1 orVDAC1-DC was added to the cis chamber containing 1 mol/LNaCl and 10 m mol/L HEPES, pH, 7.4, unless otherwiseindicated. After one or a few channels were inserted into thePLB, currentswere recorded by voltage clampingwith a BilayerClamp BC-525B Amplifier (Warner Instruments), the currenttrace durationwas 4 or 10 seconds. Currentwasmeasuredwithrespect to the trans side of the membrane (ground). Thecurrent was digitized on-line with a Digidata 1200 interfaceboard and PCLAMP 6 software (Axon Instruments, Inc.).

Patients and tissue sample preparationForty-four patients who underwent surgery for lung adeno-

carcinoma between May 2007 and May 2010 at the PasteurHospital (Department of Thoracic Surgery, CHU de Nice, Nice,France) were selected. The patients received the necessaryinformation concerning the study and consent was obtained.The study obtained approval of the ethics committee (CHU deNice). The main clinical and histopathologic data are summa-rized in Supplementary Table S1. Morphologic classification ofthe tumors was assigned according to the World HealthOrganization (WHO) criteria (18). The tumors were stagedaccording to the international tumor-node-metastasis system

(19). Follow-up data for all the patients were collected regu-larly. The median follow-up was 21 months (3.8–38.2 months).Among these patients, 13 relapsed (29.5%) and 6 (13.6%) died.Protein and miRNA were extracted from the same tissuesample using the protocol AllPrep DNA/RNA/Protein fromQIAGEN.

StatisticsAll values are the means � SD of the indicated number of

determinations (n), and significant differences are based on theStudent t test and P values indicated. All categorical data usednumbers and percentages. Quantitative data were presentedusing the median and range or mean. Differences betweengroups were evaluated using the c2 test for categorical vari-ables and the Student t test for continuous variables. SPSS 16.0statistical software (SPSS Inc.) was used. All statistical testswere 2-sided, and P values less than 0.05 indicated statisticalsignificancewhereas P values between 0.05 and 0.10 indicated astatistical tendency.

ResultsHypoxic cells with enlarged mitochondria are resistantto chemotherapy and resistance implicatesmitochondrial proteins

We reported that certain tumor-derived cell lines exposed tohypoxia showed a tubular mitochondrial network (PC3,SKMel) whereas others showed enlargedmitochondria (LS174,HeLa, A549; ref. 3). All cells showed a mitochondrial trans-membrane potential (Dym) that was unchanged comparedwith normoxic cells but the latter group was resistant tostaurosporine (STS)-induced apoptosis. We now show thatwhen hypoxic LS174 cells with enlarged mitochondria weretreated with STS, the Dym decreased in normoxia butremained unaffected in hypoxia (Fig. 1A). In normoxia, LS174cells released mitochondrial cytochrome c when incubatedwith STS whereas hypoxic LS174 cells with enlarged mito-chondria did not (Fig. 1B). To address the implication ofantiapoptotic proteins of the Bcl-2 family in hypoxic resistanceto STS-induced apoptosis, we tested the effect of the BH3domain mimetic ABT-737, an inhibitor of Bcl-2 and Bcl-XL (20)on the apoptosis resistance of hypoxic cells. ABT-737 restoredapoptosis as induced by STS in hypoxic LS174 cells (Fig. 1C),suggesting that association with a BH3 domain protein isimplicated in resistance.

To better understand the molecular mechanisms behindresistance we compared the normoxic and hypoxic levels ofanti- and proapoptotic proteins of the Bcl-2 family (Fig. 1D).LS174 and PC3 cells incubated in hypoxia (72 hours) wereresistant or sensitive to STS-induced apoptosis, respectively.Bax and tBID were not or only slightly detected in LS174 cells(Fig. 1D) while the expression of Bak and Bcl-XL were slightlyenhanced in LS174 cells. Because Bcl-XL has been described tointeract with VDAC1 (21), we examined the level of VDAC. Weobserved hypoxic induction of a faster migrating SDS-PAGEform of VDAC in LS174 but not in PC3 cells (Fig. 1D). Immu-noblots of mitochondrial fractions confirmed mitochondrialorigin (Supplementary Fig. S1).

VDAC1-DC Mediates Hypoxic Chemoresistance

www.aacrjournals.org Cancer Res; 72(8) April 15, 2012 2141




The hypoxic induction of the formation of a smallerrelative molecular mass form of VDAC is dependent onHIF-1 activation

Because an additional VDAC form was observed in hypoxic-resistant cells with enlargedmitochondria, and not in sensitivecells, we focused on the hypoxic induction of this form.As HIF-1 is essential in adaptation to hypoxia, we checkedwhether HIF-1 was involved in the formation of this form.When HIF-1a was silenced hypoxic cells did not contain thefaster migrating form (Fig. 2A), but a normal mitochondrialmorphology was restored (data not shown). Similar resultswere obtained for LS174 and A549 cells (data not shown). ThusHIF-1 initiates hypoxia-induced VDAC.

Expression of VDAC isoforms is not induced at themRNAlevel by hypoxia and the hypoxia-mediated formof VDACis a C-terminal–truncated VDAC1

We quantified the mRNA expression of VDAC1, VDAC2,and VDAC3 in normoxia and hypoxia but did not observehypoxic induction of these isoforms (Fig. 2B). siRNA directedto the mRNA of the 3 isoforms gave knockdown of thecorresponding VDAC isoform (Fig. 2B). Knockdown wasconfirmed at the protein level and identified the differentisoforms (Fig. 2C). The top band corresponded to VDAC1,the intermediate band to VDAC3, and the bottom band to afaster migrating form of VDAC1. The identity of VDAC1 wasconfirmed with another VDAC1-specific antibody, but

Figure 1. Anti- and proapoptoticproteins in the blockade ofcytochrome c release. A, flowcytometric analysis of Dy of LS174cells in normoxia (N) or hypoxia (H)without (�) or with (þ)staurosporine (STS); 3experiments. STS was added for18 hours at 1 mmol/L. B,immunofluorescence tocytochrome c (Cyto. c) of LS174cells in normoxia or hypoxia withSTS, bar 7.3mm. The percentage ofapoptotic (apop.) cells is given. C,caspase activity of cells incubatedin normoxia or hypoxia with orwithout the Bcl-2 and Bcl-XL

inhibitor ABT-737 (10 mmol/L) forthe last 4 hours and without orwith STS treatment for the last4 hours (n ¼ 8, 2 experiments).D, immunoblot of anti- andproapoptotic proteins and VDACsin LS174 andPC3 cells in normoxiaor hypoxia. HK, hexokinase; DAPI,40, 6—diamidino-2-phenylindole;RLU, relative luciferase units.

Brahimi-Horn et al.

Cancer Res; 72(8) April 15, 2012 Cancer Research2142




directed to the N-terminus, and both forms were silencedwith siRNA (Fig. 2D).We considered the possibility that the fastmigrating VDAC1

resulted from alternative splicing or hypoxia-mediated trans-lation by internal ribosome entry but did not find any evidenceto support either possibility (Supplementary Fig. S2).Finally, the faster migrating VDAC1 was not detected with a

VDAC1 antibody directed to the C-terminus (Fig. 2D), suggest-ing that the C-terminus of the protein was truncated (VDAC1-DC). Doxycycline, a second-generation tetracycline that hascytoprotective and metal chelator effects, was found to dimin-ish the formation of VDAC1-DC (Fig. 2E). Because metalchelators increase the stability of HIF-1a we examined thelevel of HIF-1a but found only a slight increase when the fastmigrating form of VDAC1 was significantly decreased (60%, 60

mg/mL doxycycline; Fig. 2E). Minocycline, another tetracyclineantibiotic, which exerts uncoupling and inhibiting effects onmitochondrial respiration (22), also inhibited formation (Fig.2F), and partially restored normal mitochondrial morphology(Fig. 2G). Because tetracycline is an inhibitor of matrix metal-loproteases we tested a number of protease inhibitors, but theydid not inhibit formation of VDAC1-DC (data not shown). Thepossibility of posttranslational cleavage of VDAC1, asdescribed previously (23), is the most likely explanation forthe appearance of VDAC1-DC.

VDAC1-DC is associated with cell survival in hypoxiaWe then hypothesized that VDAC1-DC could be involved in

hypoxic resistance to apoptosis. Silencing of VDAC1 andVDAC1-DC decreased the number of enlarged mitochondria

Figure 2. Hypoxia induced a HIF-1–dependent novel form of VDAC1. A,induction in hypoxia of a fastermigrating SDS-PAGE form of VDACis dependent onHIF-1a. Immunoblotof HIF-1a and VDACs in HIF-1asilenced HeLa cells in normoxia orhypoxia in the absence (�) orpresence (þ) of HIF-1a siRNA. B,expression of themRNA of VDAC1-3in normoxia and hypoxia in HeLacells. Expression of VDAC1, 2, and 3after transfection with control (siCtl)or VDAC1 (siVDAC1), VDAC2(siVDAC2), or VDAC3 (siVDAC3.1 orsiVDAC3.2) siRNA, results arerepresentative of 2 different siRNAfor each isoform. C, immunoblot toHIF-1a and VDAC1 (ab15895) incontrol (siCtl) or VDAC1 (siVDAC1),VDAC2 (siVDAC2), or VDAC3-(siVDAC3.1 or siVDAC3.2) silencedHeLa cells in hypoxia. Hypoxia-induced fast migrating VDAC1. D,immunoblot using antibodies againstthe N- or C-terminus of VDAC1 inHeLa cells incubated in normoxia orhypoxia transfected or not withsiRNA. E, immunoblot of HIF-1a andVDAC1 (ab15895) of LS174 cellsincubated in hypoxia. Doxycyclinewas added for the first 24 hours ofhypoxia. F, immunoblot of HIF-1aand VDAC1 (ab15895) of HeLA cellsin hypoxia for 48 hours. Minocyclinewas added for the 48 hours ofhypoxia. G, immunofluorescence tocytochrome c of HeLa cells inhypoxia without (þDMSO) or withminocycline. DMSO, dimethylsulfoxide.






and restored the tubular mitochondrial morphology (Fig. 3A).To evaluate the sensitivity to an apoptotic stimulus of nor-moxic and hypoxic LS174 cells, we determined the caspase-3and -7 activity in cells exposed to STS, an inducer of mito-chondrion-dependent apoptosis, and to etoposide, a topo-isomerase II inhibitor used in cancer therapy (Fig. 3B). Thecaspase activity was the same in cells in normoxia or hypoxia,indicating that there was no induction of cell death in hypoxia.Silencing of VDAC1 in hypoxia partially reestablished thesensitivity of hypoxic LS174 cells to apoptosis (Fig. 3B).

Two additional cell lines were examined for enlarged mito-chondria, VDAC1-DC, and resistance to STS-induced apopto-sis. SKMel cells did not show any of these features whereasA549 cells showed all of them (Supplementary Fig. S3), as forHeLa cells (Figs. 2 and 3 and Supplementary Fig. S4). We thenquestioned which form of VDAC1 (full-length or truncated)was responsible for triggering resistance. In hypoxia, the levelof VDAC1 decreased by around 50% (Fig. 3C) whereas VDAC1-DC increased by around 50%, which supported posttransla-tional cleavage of VDAC1. The silencing of vdac1 in normoxic

Figure 3. Hypoxic induction ofVDAC1-DC in cells with enlargedmitochondria show protectionfrom stimuli-induced apoptosis. A,immunofluorescence tocytochrome c in LS174 cellssilenced for VDAC1 in hypoxia.Arrows indicate fragmentedmitochondria; bar, 7.3 mm. Thepercentage of enlargedmitochondria is given. B, caspaseactivity of control and VDAC1-silenced LS174 cells (n ¼ 8; 2experiments; �, P < 0.001) innormoxia or hypoxia without (�) orwith (þ) staurosporine (STS) oretoposide (Etop) for the last4 hours. C, immunoblot of VDAC1in HeLa cells in normoxia orhypoxia. Histograms show thequantification of the bands. D,immunofluorescence tocytochrome c of normoxic HeLacells transfected with siRNA toVDAC1. E, immunoblots of HIF-1aor VDAC1 ofHeLa cells transfectedwith increasing concentrations ofsiVDAC1. F, immunoblots to HIF-1a or VDAC1 of HeLa cellsincubated first in hypoxia and thenreoxygenated. Histograms showquantification of the signal toVDAC1 and VDAC1-DC. G,expression of pVDAC1-5kDaDC(0.5 and 5.0 mg plasmid) innormoxia in HeLa cells andendogenous VDAC1-DCexpression in hypoxia. H, caspaseactivity of HeLa cells, in normoxia,transfected with a control plasmidor a plasmid expressing a 5 kDa C-terminal–truncated VDAC1 with orwithout STS treatment. Thepercentage of cells with enlarged(enlarg.) mitochondria (mito.) isgiven. I, caspase activity (%) ofHeLa cells incubated in normoxiaor hypoxia with or withoutdoxycycline (Dox) or minocycline(MC) and without or with STS forthe last 4 hours (n ¼ 8, 2experiments). RLU, relativeluciferase units.

Brahimi-Horn et al.





cells was associated with a change in mitochondrial morphol-ogy, as visualized with anticytochrome c (Fig. 3D) and adecrease in VDAC1 in hypoxia (Fig. 3E). However, these cellsshowed an increase in apoptosis with STS (Fig. 3B), suggestingthat the decrease in VDAC1 in normoxia was not responsiblefor the protection against apoptosis. Moreover, silencing ofHIF-1a (þTet) in hypoxia with STS (Fig. 2A) restored sensi-tivity to apoptosis (data not shown). We thus concluded thatresistance to apoptosis was associated with VDAC1-DC. Toevaluate this further, cells were placed in hypoxia and thenreoxygenated. The level of VDAC1-DC was maintained for 4hours after reoxygenation, then progressively decreased after 8hours and disappeared at 48 hours (Fig. 3F). As expected, thelevel of VDAC1 was inversely proportionate to that of VDAC1-DC. We showed previously that during the first 24 hours ofreoxygenation, cells were protected from apoptosis (3). Tran-sient exogenous overexpression of a small form of VDAC1truncated by 5 kDa in the C-terminus, pVDAC1-5kDaDC (Fig.3G), in cells exposed for 4 hours to STS showed a slightresistance to apoptosis (Fig. 3H). Finally, in the presence ofdoxycycline (Fig. 2E) orminocyclin (Fig. 2F), hypoxic cells wereno longer protected from STS (Fig. 3I). Taken together, theseresults show that both the enlarged morphology of mitochon-dria and VDAC1-DC participate in protection against apopto-sis in hypoxic LS174, A549, and HeLa cells exposed to STS oretoposide.

VDAC1-DC has the same channel activity and voltagedependency as VDAC1 and binds Bcl-XL

VDAC1 and VDAC1-DC proteins were purified from hypoxicHeLa cells (Fig. 4A) and their channel activity was examinedfollowing reconstitution into a PLB. The current through lipidbilayer–reconstituted VDAC1 (fraction 12) or VDAC1-DC (frac-tion 22) in response to a voltage step from 0 to �10 or to �40mV (Fig. 4B and C) was the same for the 2 proteins. At�10mV,the channel conductance of both proteins was the same (30pA). At a higher voltage of �40 mV, the full-length channelshowed 2 major conducting states with higher occupancy atthe closed substate (S1), whereas VDAC1-DC showed higheropen-state occupancy (O) in comparison to the occupancy oflow-conducting substates (S1, S2; Fig. 4C). Both channelsshowed similar but not identical voltage-dependent conduc-tance. At the high voltages, VDAC1-DC showed slightly higherconductance than VDAC1 (Fig. 4D), in agreement with thesingle-channel experiments [Fig. 4B(ii) and C(ii)]. The volt-age sensitivity of VDAC1-DC suggests the presence of theN-terminus, conferring voltage gating of the channel (8).VDAC1-DC showed similar Ca2þ conductance to VDAC1,but at higher voltages spent a longer time in its open state, asreflected in the decreased voltage sensitivity (Fig. 4E).For example, at þ40 mV the Ca2þ conductance VDAC1-DCwas about 1.4-fold higher than that of VDCA1 (Fig. 4E).VDAC1-DC, like VDAC1, interacted with purified Bcl-XL(DC)and decreased its channel conductance (Fig. 4F). Similarresults were obtained with hexokinase I from rat brain (datanot shown). These results suggest that the C-terminaldomain is not required for the interaction of these anti-apoptotic proteins with VDAC1.

VDAC1-DC forms a complex with hexokinase II and isassociated with cell survival in hypoxia

As hexokinase II is a major player in maintaining the highlymalignant state of cancer cells (24), we focused on interactionbetween hexokinase II and VDACs in hypoxia. Immunopreci-pitates with antihexokinase II contained VDAC1, VDAC3, andVDAC1-DC (Fig. 5A). In addition, the hexokinase II expressionlevel was substantially increased in hypoxia (Fig. 5B). Silencingof more than 90% of the hypoxia-inducible expression ofhexokinase II decreased considerably the level of VDAC1-DC(Fig. 5B). Silencing of hexokinase II in normoxia did not affectthe level of VDAC1. Conversely, silencing of VDAC1 in nor-moxia and hypoxia decreased slightly the level of hexokinase II(Fig. 5B). These results were confirmed by immunofluores-cence; no or little labeling was observed with anti-VDAC incells silenced for either hexokinase II or VDACs (Fig. 5C). Inaddition, a more intense and punctate immunofluorescencewas observed with anti-VDAC in VDAC1-DC–containing cellsincubated in hypoxia than in normoxia (Supplementary Fig.S5). Clotrimazole (CTM) and bifonazole (BFN) induce apopto-sis by detaching hexokinase from mitochondria (17, 25). Bothagents increased mortality to a similar extent to that for VDACsilencing (Fig. 5D) and the mortality was enhanced in cells inhypoxia in their presence. This suggested that VDAC1-DCinteracted with hexokinase II, as did purified hexokinase I,which decreased VDAC1-DC channel conductance. To betterunderstand the role of VDAC1 and hexokinase II in cell survivalin hypoxia, we silenced VDAC1 or hexokinase II and tested cellproliferation/death and ATP and lactate production in nor-moxia and hypoxia. Hypoxia does not kill cells (26), but it slowsproliferation, as shown by a 3-fold decrease in the area ofcolonies of cells after 10 days in hypoxia (Fig. 5E). Transientsilencing of VDAC1 (siVDAC1) in hypoxia hadno impact on cellsurvival but affected proliferation (P < 0.01), whereas silencingof hexokinase II (siHKII) strongly inhibited survival (Fig. 5E). Ashexokinase II and VDAC1 form a complex and silencing ofhexokinase II decreased the level of VDAC1 in cells (Fig. 5B), wehypothesized that hexokinase II interfered with ATP transportand thereby its production via its interaction with VDAC1-DC.Cells produced almost 2 times more ATP in hypoxia (Fig. 5F).Silencing of VDAC1 decreased hypoxic but not normoxicproduction of ATP, suggesting that VDAC1-DC influenced ATPproduction (Supplementary Fig. S6). This could reflect theimpact of VDAC1 silencing on hexokinase II expression. Asexpected the silencing of hexokinase II in hypoxia decreasedATP. Lactate production, which reflects ATP synthesis viaglycolysis, was increased in hypoxia and diminished withVDAC1 or hexokinase II silencing (Fig. 5G). Taken together,these results confirm that VDAC1 is involved in energy homeo-stasis and points to VDAC1-DC as an essential actor in bothglycolysis and mitochondrial energy production in hypoxiaprobably through interaction with hexokinase.

VDAC1-DC is present in tissues of patients with lungadenocarcinoma and is more frequently detected inlate-stage rather than in early-stage tumors

Because we detected both VDAC1 and VDAC1-DC inhypoxic A549 lung carcinoma cells (Supplementary Fig. S3),






we tested for VDAC1 and VDAC1-DC in lung adenocarci-nomas tumor tissue from 44 patients. Tumor tissues weredivided into 2 groups: stage IA and IB (n ¼ 25) and stageIIIA and IIIB (n ¼ 19). The clinical characteristics of thepatients are listed in supplementary Table S1. The level of

VDAC1-DC was determined in corresponding controlmatched healthy (C) and tumor (T) tissue of patients withlung cancer (Fig. 6A). The tumor tissue, but not healthytissue, contained VDAC1-DC and the level of VDAC1-DC instage III was several fold higher than in stage I (see also

Figure 4. VDAC1-DCchannel activity and binding of Bcl-XL-DC are identical to that of VDAC1. A, VDAC1 and VDAC1-DCpurification fromHeLa cells identifiedby immunoblotting (anti-VDAC1; Calbiochem); fractions 21 or 22 were used. B and C, channel activity of bilayer-reconstituted purified VDAC1 or VDAC1-DC.Currents through bilayer-reconstituted VDAC1 or VDAC1-DC in response to a voltage step from 0 to�10 mV [B(i) and C(i)] or to �40 mV [B(ii) and C(ii)] wererecorded. The dashed lines indicate the zero current level. The total current amplitude histogram traces (in the same recording), showing the relativeoccupancy of the open state (O) and closed substate (S) or, for VDAC1-DC, of 2 or more substates (S1 and S2) during a 4-second recording are shown (B andC). D, currents through theVDAC1 (*) or VDAC1-DC (o) channelswere recorded in the presence of 1mol/LNaCl and in response to a voltage step from0mV tovoltages between�60 toþ60mV. Relative conductance was determined as the ratio of conductance at a given voltage (G) to themaximal conductance (G0).The results are representative of 9 similar experiments in which the value of each voltage represents the average of 3 to 6 swipes. E, currents through VDAC1(*) or VDAC1-DC (o) as recorded in the presence of 0.2 mol/L CaCl2 and in response to a voltage step from 0 mV to voltages between �60 to þ60 mV. Theresults are the average of 2 similar experiments with 3 swipes for each voltage. F, VDAC channel conductance was recorded before and 10 minutes after theaddition of purified Bcl-XL-DC to the cis chamber. A representative experiment of 3 similar experiments is shown.

Brahimi-Horn et al.





below). In addition, electron micrographs of mitochondriaof tumor and matched normal patients' tissues showedenlarged mitochondria in only tumor samples (Supplemen-tary Fig. S7).

The expression of carbonic anhydrase IX (CAIX; refs. 27, 28)and of miR-210 (ref. 29; Fig. 6B), 2 HIF-induced gene products,was analyzed to confirm the hypoxic status of the tissues (Fig.6A andB). Positivity for CAIXwas about 76%and 71% for tumor

Figure 5. Hypoxic VDAC1-DC forms a complexwith hexokinase II and cell survival in hypoxia requires VDAC1-DCand hexokinase II. A, coimmunoprecipitationof endogenous VDAC1 with hexokinase II. Lysates of hypoxic LS174 cells were immunoprecipitated with an antihexokinase II antibody. VDAC proteins in thelysates and the immunoprecipitates are shown (arrow). Cytochrome c was used as a control. IB, immunoblot; IP, immunoprecipitate. B, immunoblot ofhexokinase II and VDAC in control (siCtl), VDAC (siVDAC1), or hexokinase II (siHKII) transfected HeLa cells in normoxia or hypoxia. Histograms show theintensities normalized toa-tubulin for VDAC1 andVDAC1-DC (left) and hexokinase II (right). C, immunofluorescence of VDAC1 inHeLa cells silenced for VDAC(siVDAC1) or hexokinase II (siHKII) in hypoxia. Bar, 7.3 mm. D, percentage of mortality (Trypan blue exclusion) of HeLa cells incubated in normoxia or hypoxia.Cells were transfected with or without a siRNA to VDAC1 and incubated or not with clotrimazole (CTM-50 mmol/L) or bifonazole (BFN-50 mmol/L) for the lasthour of normoxia or hypoxia. E, viability assay of HeLa cells in control (siCtl), VDAC (siVDAC1), or hexokinase II (siHKII) silenced cells in normoxia or hypoxia.Cells were transiently transfected twice with the indicated siRNA (40 nmol/L), then seeded, and incubated in normoxia or hypoxia for 10 days before staining.The colony number and area was quantified with ImageJ software. Three experiments with 2 different sets of siRNA. The Student t test (�, P < 0.01). F, ATPproduction of HeLa cells silenced (siVDAC1) or not (siCtl) for VDAC1 or hexokinase II (siHKII; n ¼ 8, 3 experiments; �, P < 0.001) in normoxia or hypoxia. G,lactate production of HeLa cells silenced (siVDAC1) or not (siCtl) for VDAC1 or hexokinase II (siHKII; n¼ 2, 3 experiments; �, P < 0.01) in normoxia or hypoxia.DAPI, 40, 6-diamidino-2-phenylindole.






tissues from stage I and stage III patients, respectively (Fig.6A). Only rare control tissues showed minimal CAIX expres-sion. The quality control of the miRNA in the extracts wasconfirmed by the level of miR-21. Significant relative expres-sion of miR-210 was observed in tumors (Fig. 6B). The bandintensity for VDAC1 and VDAC1-DC was determined withGeneTools software from Syngene, and the ratio of theintensity of VDAC1-DC to VDAC1 was evaluated for early-stage I (A and B) and late-stage III (A and B) tumors (Fig. 6Aand C). The overall ratio for the early- and late-stagesshowed a significant increase when comparing healthycontrol and tumor tissue (Fig. 6A, C, and D). When patienttissues were subgrouped into tissue showing either a low or

high ratio (arbitrary threshold of 0.25) the number ofpatients showing a high ratio was slightly increased forstage III tumors (11 of 19 vs. 11 of 25), that is, 57.8% ofstage III tumors were positive compared with 44% of stage Itumors (Fig. 6C). The difference between the matchedhealthy and tumor tissue was substantially higher for bothearly- and late-stage positive tumors (subgroup high ratio).In particular, VDAC1-DC was detected more frequently inlarger tumors (41.7% in T1 and T2 tumors vs. 75% in T3 andT4 tumors, P ¼ 0.08, statistical tendency; Fig. 6E, Table 1)and higher lung adenocarcinoma stage—patients not oper-ated on (83.3% stage IIIB vs. 42.1% in other stages, P ¼ 0.06,statistical tendency; Fig. 6F, Table 1). The level of necrosis of

Figure 6. The ratio of VDAC1-DC toVDAC1 was higher in tumor tissuethan in control tissue andproportionately more stage IIItumors had a high ratio. A,representative immunoblots ofVDAC in tissue extracts of healthycontrol (C) and tumor (T) tissuefrom 6 individual patients with lungcancer with either early-stage I orlate-stage III tumors. CAIX was anindicator of hypoxia in tumors. B,the fold induction of miR-21 andmiR-210 was determined forcontrol and tumor tissues frompatients with stage I and III tumors.The level of miR-21 indicated thequality of the miRNA in extractswhereas miR-210 is a hypoxia-induced miRNA. C, ratio of theintensity of the immunoblot signalof VDAC1-DC to VDAC1determined with GeneToolssoftware (Syngene) for healthy (C)and tumor (T) tissue from patientswith lung cancer with either stage Ior III tumors. D, ratio of theimmunoblot signal to VDAC1-DCto VDAC1 for each individualpatient for healthy (C) and tumor (T)tissue. Patients with lung cancerwith stage I tumors (left graph) andstage III tumors (right graph). E,evaluation of the number ofpatients with lung cancer eithernegative or positive for VDAC1-DC,respectively (VDAC1-DC neg.) and(VDAC1-DC pos.), with small- (T1and T2) or large- (T3 and T4) sizedtumors. F, evaluation of the numberof patients with lung cancer eithernegative or positive for VDAC1-DC,respectively (VDAC1-DC neg.) and(VDAC1-DC pos.), with stages IA–IB–IIIA or stage IIIB tumors.

Brahimi-Horn et al.





the patients' tissues was neither high nor substantiallydifferent between stages I and III, 6% and 12%, respectively.

DiscussionHerein, we showed that hypoxia induces the appearance of a

C-terminal truncated form of VDAC1. The mechanism regu-lating formation was HIF dependent and the truncated formpossessed channel activity, interacted with Bcl-XL and hexo-kinase I, both of which protect against apoptosis.Interaction of VDACs with Bcl-2 family members is

implicated in translocation of metabolites across the mito-chondrial outer membrane (21). Nonetheless, it has beenreported that the 3 isoforms are dispensable for mitochon-drial-dependent cell death, but this was showed in acellular and environmental context that was neither malig-nant nor hypoxic (30). In addition, VDAC2 but not VDAC1has been shown to inhibit Bak-mediated mitochondrialapoptosis (31). It is possible to hypothesize that changesin the expression of Bcl-2 proteins are implicated directlyin resistance. Cytoprotection of lung cancer cells to cis-platin correlated with suppression of activation of Bax butnot Bak by cisplatin (12). However, in our study the Baxprotein was not detected in LS174 cells. In addition, thereexists an intricate crosstalk between the machineries ofmitochondrial dynamics (fusion and fission), thus mor-phology, and apoptosis (32, 33). Both antiapoptotic (Bcl-2) and proapoptotic (Bak and Bax) proteins interact withproteins involved in mitochondrial fusion (mitofusins) andfission (dynamin-related GTPases). Thus, we may speculatethat modifications in the expression of Bak and Bcl-XL

correlate with the morphologic alterations observed. Inaddition, the increases in Bak and Bcl-XL in hypoxic LS174cells correlated with the morphologic alteration in hypoxia.

It may also be hypothesized that the increase in theexpression of Bcl-XL and the modification of the openconfiguration of the VDAC1-DC channel by Bcl-XL inhibitsmitochondrial ATP/ADP exchange, which favors ATP pro-duction through glycolysis. A shift toward glycolysis is acharacteristic of hypoxic cancer cells and may explainsurvival and thus resistance when confronted with a poten-tially lethal agent. The observed change in organization ofthe cristae of the mitochondria may also rupture theinteraction between VDAC and the adenine nucleotidetranslocase thereby leading to a change in VDAC-mediatedATP transport. In addition, if VDAC oligomerization isresponsible for cytochrome c release in apoptosis (11), achange in its conformation may block cytochrome c in themitochondrial intermembrane space and thus diminishapoptosis.

The notion that resistance of hypoxic regions of tumors tochemotherapy (34) is associated with hypoxic VDAC1-DC andBcl-XL is supported by reports showing that protection ofHepG2 cells against etoposide-induced apoptosis was HIF-1a dependent (35) and that Bcl-XL is induced by HIF (36).

Because we detected VDAC1-DC in tumor tissue of patients(50%) and that the frequency of positivity for VDAC1-DC washigher in late-stage tumors than in early-stage tumors, webelieve that VDAC1-DC represents a product of tumor pro-gression. Gene expression of VDAC1 has been reported topredict poor outcome in early-stage non–small cell lung cancer(37). In addition, VDAC1 was shown to be upregulated inprednisolone-sensitive acute lymphoblastic leukemia cells butnot in resistant cells (38).

In conclusion, our results point to modifications in mito-chondrial dynamics and production of VDAC1-DC as asurvival response in hypoxic cancer cells that resist apopto-sis. Because agents that promote apoptosis may hold ther-apeutic benefit, these results may have important repercus-sions for combating cancer cell resistance to chemotherapy.A synthetic lethality approach targeting RAS tumor cellsidentified a small-molecule inhibitor of VDAC2 that inducedchanges in mitochondrial morphology and cell death (39).We propose that VDAC1-DC may be a potential biomarker tostratify patients with respect to tumor progression and thatthe VDAC1-DC/hexokinase complex may be a cancer spe-cific target for therapy.

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

AcknowledgmentsThe authors are thankful to N. Pons for helping to extract RNA/proteins from

patients' tissues and to J. Hickman for giving ABT737.

Grant SupportThis research was supported by grants from the LNCC (Equipe labellis�ee), the

ANR, the INCA, METOXIA (FP7-EU program), ARC, and Canceropole PACA. Thelaboratory is funded by the Centre Antoine Lacassagne, CNRS, and INSERM.

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

Received December 5, 2011; revised February 3, 2012; accepted February 15,2012; published OnlineFirst March 2, 2012.

Table 1. Comparison of the characteristics ofthe patients and their tumors with VDAC1-DCdetection

VDAC1-DC detection c2 (P value)

GenderFemale 5 (50%) P ¼ 0.87Male 16 (47.1%)

Age, y<60 13 (52%) P ¼ 0.51�60 8 (42.1%)

Tumor sizeT1 and T2 15 (41.7%) P ¼ 0.08T3 and T4 6 (75%)

Node stageN0 14 (50%) P ¼ 0.69N1 and N2 7 (43.8%)

Initial stageIA, IB, IIIA 16 (42.1%) P ¼ 0.06IIIB 5 (83.3%)






References1. WilsonWR,HayMP. Targeting hypoxia in cancer therapy. Nat RevCan

2011;11:393–410.2. Semenza GL. Defining the role of hypoxia-inducible factor 1 in cancer

biology and therapeutics. Oncogene 2010;29:625–34.3. Chiche J, Rouleau M, Gounon P, Brahimi-Horn MC, Pouyssegur J,

Mazure NM. Hypoxic enlarged mitochondria protect cancer cells fromapoptotic stimuli. J Cell Physiol 2010;222:648–57.

4. Kroemer G, Galluzzi L, Brenner C. Mitochondrial membrane permea-bilization in cell death. Physiol Rev 2007;87:99–163.

5. Galluzzi L, Kroemer G. Mitochondrial apoptosis without VDAC. NatCell Biol 2007;9:487–9.

6. Kroemer G, Pouyssegur J. Tumor cell metabolism: cancer's Achilles'heel. Cancer Cell 2008;13:472–82.

7. Suen DF, Norris KL, Youle RJ. Mitochondrial dynamics and apoptosis.Genes Dev 2008;22:1577–90.

8. Abu-Hamad S, Arbel N, Calo D, Arzoine L, Israelson A, Keinan N, et al.The VDAC1 N-terminus is essential both for apoptosis and the pro-tective effect of anti-apoptotic proteins. J Cell Sci 2009;122:1906–16.

9. Abu-Hamad S, Zaid H, Israelson A, Nahon E, Shoshan-Barmatz V.Hexokinase-I protection against apoptotic cell death is mediated viainteraction with the voltage-dependent anion channel-1: mapping thesite of binding. J Biol Chem 2008;283:13482–90.

10. Arzoine L, ZilberbergN,Ben-RomanoR,Shoshan-BarmatzV. Voltage-dependent anion channel 1-based peptides interact with hexokinaseto prevent its anti-apoptotic activity. J Biol Chem 2009;284:3946–55.

11. Keinan N, Tyomkin D, Shoshan-Barmatz V. Oligomerization of themitochondrial protein voltage-dependent anion channel is coupled tothe induction of apoptosis. Mol Cell Biol 2010;30:5698–709.

12. TajeddineN,Galluzzi L,KeppO,HangenE,Morselli E, Senovilla L, et al.Hierarchical involvement of Bak, VDAC1 and Bax in cisplatin-inducedcell death. Oncogene 2008;27:4221–32.

13. Hiller S, Abramson J,Mannella C,Wagner G, Zeth K. The 3D structuresof VDAC represent a native conformation. Trends Biochem Sci2010;35:514–21.

14. Pastorino JG, Hoek JB. Regulation of hexokinase binding to VDAC.J Bioenerg Biomembr 2008;40:171–82.

15. Natsuizaka M, Ozasa M, Darmanin S, Miyamoto M, Kondo S, KamadaS, et al. Synergistic up-regulation of Hexokinase-2, glucose transpor-ters and angiogenic factors in pancreatic cancer cells by glucosedeprivation and hypoxia. Exp Cell Res 2007;313:3337–48.

16. Berra E, Benizri E, Ginouves A, Volmat V, Roux D, Pouyssegur J. HIFprolyl-hydroxylase 2 is the key oxygen sensor setting low steady-statelevels of HIF-1alpha in normoxia. Embo J 2003;22:4082–90.

17. Shoshan-Barmatz V, De Pinto V, Zweckstetter M, Raviv Z, Keinan N,Arbel N. VDAC, a multi-functional mitochondrial protein regulating celllife and death. Mol Aspects Med 2010;31:227–85.

18. Travis WD, Brambilla E, M€uller-Hermelink HK, Harris CC. WHO histo-logical classification of tumors of the lung World Health OrganizationClassification of tumours. In: Pathology and genetics of tumours of thelung, pleura, thymus and heart. Vol. 10. Lyon, France: IARC; 2004.

19. Mountain CF. Revisions in the International System for Staging LungCancer. Chest 1997;111:1710–7.

20. Hickman JA, Hardwick JM, Kaczmarek LK, Jonas EA. Bcl-xL inhibitorABT-737 reveals a dual role for Bcl-xL in synaptic transmission. JNeurophysiol 2008;99:1515–22.

21. Vander Heiden MG, Li XX, Gottleib E, Hill RB, Thompson CB, Colom-bini M. Bcl-xL promotes the open configuration of the voltage-depen-dent anion channel and metabolite passage through the outer mito-chondrial membrane. J Biol Chem 2001;276:19414–9.

22. Cuenca-LopezMD, Karachitos A,Massarotto L, Oliveira PJ, Aguirre N,GalindoMF, et al. Minocycline exerts uncoupling and inhibiting effectson mitochondrial respiration through adenine nucleotide translocaseinhibition. Pharmacol Res 2012;65:120–8.

23. Ozaki T, Yamashita T, Ishiguro S. Mitochondrial m-calpain plays a rolein the release of truncated apoptosis-inducing factor from the mito-chondria. Biochima Biophys Acta 2009;1793:1848–59.

24. Mathupala SP, Ko YH, Pedersen PL. Hexokinase II: cancer's double-edged sword acting as both facilitator and gatekeeper of malignancywhen bound to mitochondria. Oncogene 2006;25:4777–86.

25. Penso J, Beitner R. Clotrimazole and bifonazole detach hexokinasefrom mitochondria of melanoma cells. Eur J Pharmacol 1998;342:113–7.

26. Bellot G, Garcia-Medina R, Gounon P, Chiche J, Roux D, PouyssegurJ, et al. Hypoxia-induced autophagy is mediated through hypoxia-inducible factor induction of BNIP3 andBNIP3L via their BH3domains.Mol Cell Biol 2009;29:2570–81.

27. Chiche J, Ilc K, Laferriere J, Trottier E, Dayan F, Mazure NM, et al.Hypoxia-inducible carbonic anhydrase IX and XII promote tumor cellgrowth by counteracting acidosis through the regulation of the intra-cellular pH. Cancer Res 2009;69:358–68.

28. Ilie M, Mazure NM, Hofman V, Ammadi RE, Ortholan C, Bonnetaud C,et al. High levels of carbonic anhydrase IX in tumour tissue and plasmaare biomarkers of poor prognostic in patients with non-small cell lungcancer. Br J Cancer 2010;102:1627–35.

29. Puissegur MP, Mazure NM, Bertero T, Pradelli L, Grosso S, Robbe-Sermesant K, et al. miR-210 is overexpressed in late stages of lungcancer and mediates mitochondrial alterations associated with mod-ulation of HIF-1 activity. Cell Death Differ 2011;18:465–78.

30. Baines CP, Kaiser RA, Sheiko T, Craigen WJ, Molkentin JD. Voltage-dependent anion channels are dispensable for mitochondrial-depen-dent cell death. Nat Cell Biol 2007;9:550–5.

31. Cheng EH, Sheiko TV, Fisher JK, Craigen WJ, Korsmeyer SJ. VDAC2inhibits BAK activation and mitochondrial apoptosis. Science2003;301:513–7.

32. Brooks C, Dong Z. Regulation of mitochondrial morphological dynam-ics during apoptosis by Bcl-2 family proteins: a key in Bak? Cell Cycle2007;6:3043–7.

33. Rolland SG, Conradt B. New role of the BCL2 family of proteins in theregulation of mitochondrial dynamics. Curr Opin Cell Biol 2010;22:852–8.

34. Brown JM,WilsonWR. Exploiting tumour hypoxia in cancer treatment.Nat Rev Cancer 2004;4:437–47.

35. Sermeus A, Cosse JP, Crespin M, Mainfroid V, de Longueville F,Ninane N, et al. Hypoxia induces protection against etoposide-induced apoptosis: molecular profiling of changes in gene expressionand transcription factor activity. Mol Cancer 2008;7:27.

36. Chen N, Chen X, Huang R, Zeng H, Gong J, MengW, et al. BCL-xL is atarget gene regulated by hypoxia-inducible factor-1{alpha}. J BiolChem 2009;284:10004–12.

37. GrillsC, JitheshPV,Blayney J, ZhangSD, Fennell DA.Gene expressionmeta-analysis identifies VDAC1 as a predictor of poor outcome in earlystage non-small cell lung cancer. PLoS One 2011;6:e14635.

38. Jiang N, Kham SK, Koh GS, Suang Lim JY, Ariffin H, Chew FT, et al.Identification of prognostic protein biomarkers in childhood acutelymphoblastic leukemia (ALL). J Proteomics 2011;74:843–57.

39. Yagoda N, von Rechenberg M, Zaganjor E, Bauer AJ, Yang WS,Fridman DJ, et al. RAS-RAF-MEK-dependent oxidative cell deathinvolving voltage-dependent anion channels. Nature 2007;447:864–8.

Brahimi-Horn et al.





2012;72:2140-2150. Published OnlineFirst March 2, 2012.Cancer Res M. Christiane Brahimi-Horn, Danya Ben-Hail, Marius Ilie, et al. Progression to Chemotherapy ResistanceAssociates with Hypoxic Cell Survival and Correlates with Expression of a Truncated Active Form of VDAC1 in Lung Cancer

Updated version

10.1158/0008-5472.CAN-11-3940doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2012/03/02/0008-5472.CAN-11-3940.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/72/8/2140.full.html#ref-list-1

This article cites 38 articles, 13 of which you can access for free at:

Citing articles

/content/72/8/2140.full.html#related-urls

This article has been cited by 4 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

[email protected]

To request permission to re-use all or part of this article, contact the AACR Publications Department at



http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-11-3940

http://cancerres.aacrjournals.org/content/suppl/2012/03/02/0008-5472.CAN-11-3940.DC1

http://cancerres.aacrjournals.org/content/72/8/2140.full.html#ref-list-1

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

mailto:[email protected]

mailto:[email protected]


expression of a truncated active form of vdac1 in lung ... · 2017. 4. 13. · m. christiane...

Documents