Targeting cyclin B1 inhibits proliferation and sensitizes breast ...

56MethodsCell culture, reagents and cell synchronizationCervical cancer cell line HeLa and breast cancer cell lines MCF-7, BT-474, SK-BR-3 andMDA-MB-231 were obtained from DSMZ (Braunschweig). Fetal calf serum (FCS) waspurchased from PAA laboratories (Cölbe). Opti-MEM I, oligofectamine, glutamine,penicillin, streptomycin and trypsin were obtained from Invitrogen (Karlsruhe). Taxol wasfrom Mayne Pharma (Haar). Cells were synchronized to G1/S boundary by a doublethymidineblock. Briefly, cells were treated with 2 mM thymidine (Sigma-Aldrich,of protein expression were carried out at 24 h, 48 h, 72 h and 96 h and proliferation assayswere performed at 24 h, 48 h and 72 h after siRNA transfection in MCF-7 cells.For chemotherapeutic treatment, MCF-7 cells were at first transfected with siRNA and 4 hlater followed by treatment of taxol (3 ng/ml). For irradiation, 6 h post transfection cells wereexposed to a single dose of 8 Gy at room temperature by a linear accelerator (SL 75/5, Elekta,Crawley, UK) with 6 MEV photons/100 cm focus-surface distance and a dose rate of 4.0Gy/min. 48 h after transfection of siRNAs cells were harvested for proliferation assay, cellcycle analysis and apoptosis evaluation.Taufkirchen) for 16 h, released into fresh medium for 8 h and subjected again to thymidine forfurther 16 h. To obtain prometaphase arrest, after initial thymidine incubation and 8hreleasecells were exposed to 50 ng/ml nocodazole (Sigma-Aldrich) for 14 h.Transfection of siRNA and the combined treatment with drugs or irradiationFour siRNAs targeting cyclin B1 (NCBI accession number of cyclin B1: NM 031966) weresynthesized by Dharmacon Research, Inc. (Lafayette), referred to as siRNA1-4. siRNA1against cyclin B1 corresponds to positions 340-360 of the cyclin B1 open reading frame,siRNA2 to positions 476-496, siRNA3 to positions 776-796 and siRNA4 to positions 1302-1322. Control siRNA targeting green fluorescent protein (siGFP) was also purchased fromDharmacon. All siRNAs were 21 nucleotides in length and contained symmetric 3’ overhangsof two deoxythymidines.Cells were transfected with siRNA using transfection reagent oligofectamine, according to themanufacturer’s instructions (Invitrogen). In brief, one day prior to transfection, cells wereseeded without antibiotics to a density of 50-60%. In all experiments cells were transfectedwith siRNA1-4 or siGFP at a concentration of 10 nM. Cells were harvested 48 h after siRNAtreatmentfor cell cycle evaluation, Western blot analysis and kinase assay. The time kineticsWestern blot analysis and kinase assay in vitroCell lysis was performed in RIPA buffer (50 mM Tris-HCl pH 8.0, 150 mM NaCl, 1% NP-40,0.5% Na-desoxycholate, 0.1% SDS, 1 mM Na3VO4, 1 mM phenylmethylsulphonyl-fluoride(PMSF), 1 mM Dithiothreitol (DTT), 1 mM NaF, and protease inhibitor cocktail Complete(Roche, Mannheim)). Total protein was separated by using 12% sodium dodecyl sulfatepolyacrylamidegel electrophoresis (SDS-PAGE) and then transferred to Immobilon-Pmembranes (Millipore, Bedford, MA). Membranes were exposed to corresponding antibodiesfor 1 h in PBS containing 5% slim milk, washed with phosphate-buffered saline (PBS)containing 0.2 % Tween-20, incubated subsequently with secondary antibodies for 1 h.Finally, the protein bands were visualized with the enhanced chemiluminescence reagent(ECL, Pierce, Rockford). Mouse monoclonal antibodies against cyclin B1 (1:5,000), Cdk1(1:2,000), anti-mouse secondary antibodies (1:4,000) and anti-rabbit secondary antibodies(1:4,000) were purchased from Santa Cruz (Heidelberg). Rabbit polyclonal antibodies againstPARP (poly(ADP-ribose) polymerase, 1: 1000) were from Cell Signaling Technology(Beverly). Mouse monoclonal antibodies against β-actin (1:200,000) were obtained fromSigma-Aldrich. Western blots were quantified by applying a Kodak gel documentation system7(model 1D 3.5) and standardized with loading control. For kinase assays in vitro, antibodiesagainst cyclin B1 (Santa Cruz, Heidelberg) were used for immunoprecipitation from 600 μgof cellular extracts. 0.5 μg histone H1 (Calbiochem, Darmstadt) served as substrate for eachreaction. Kinase assays were performed as previously described [29].8(Zeiss, Oberkochen). The colony number in control sample was referred as 100 % byquantification.As to experiments in vivo, HeLa cells were treated with siRNA3 or siGFP and harvested after48 h. HeLa cells (1×10 6 ) were resuspended in 300 μl of 0.9% NaCl and subcutaneouslyinjected into both flanks of nude mice. Each group contained 4 mice. Three weeks afterCell proliferation, cell cycle analysis and apoptosis assayCell viability was assessed by trypan blue staining. The proliferation rate of cells wasdetermined at indicated time points by counting cell numbers with a hemacytometer. Allexperiments were performed in triplicate. For cell cycle analysis, cells were harvested,washed with PBS and fixed in 70% chilled ethanol at 4°C for 30 min, then treated with 1mg/ml of RNase A (Sigma-Aldrich) and stained with 100 μg/ml of propidium iodide for 30min . The DNA content of 10,000 cells was determined with a fluorescent-activated cell sorterFACScan (Becton Dickinson Biosciences, Heidelberg). The data were analysed with cellcycle analysis software ModFit LT 2.0 (Verity Software House, Topsham, ME). Most of theexperiments were performed in triplicate. Indirect immunofluorescence staining forinoculation the tumor sizes were measured every 3-4 days using callipers and the tumorvolumes were calculated according to a standard formula: π/6 × length × width 2 . The tumorvolumes within the group were represented by the mean value. All mice were properly treatedin accordance with the guidelines of the local animal committee.Statistic analysisFor assays in vitro, Student’s t-tests were used to evaluate the significance of differencebetween control cells and siRNAs-treated cells. Differences were considered as statisticallysignificant when p< 0.05. With xenograft mouse model, the significant difference between thesiGFP-treated group and siRNA3-treated group was analyzed by Mann-Whitney U test.subcellular cyclin B1 localization and DNA were carried out as previously described [29].Apoptosis was assessed using Vybrant TM apoptosis assay kit according to the manufacturer’sinstructions (Molecular Probes, Leiden).Colony formation assay and in vivo experimentsMCF-7 cells were treated with siRNA1-3 or siGFP for 48 h and harvested for colonyformation assays. Briefly, cells were seeded in 24 well-plates at a density of 200 cells/wellinto culture medium containing 0.3% agar (Roth, Karlsruhe) overlaying 0.5% agar. Cells werecultured at 37°C with 5% CO2, and colonies were counted 4 weeks later using a microscope

9ResultsCancer cell lines lend themselves as useful models to further our understanding ofgynecological cancers such as breast and cervical cancer. To study the function of cyclin B1in breast cancer cells we selected cancer cell lines MCF-7, BT-474, SK-BR-3 and MDA-MB-231, as they represent the best characterized cell lines for breast cancer research. While MCF-7 and BT-474 express both estrogen receptor (ER) and progesterone receptor (PgR), SK-BR-3and MDA-MB-231 cells lack those receptors [30]. Unlike SK-BR-3 and BT-474, which areHer-2/neu-positive, MCF-7 cells only exhibit a basal level of Her-2/neu [30]. In addition,MCF-7 cells express high amounts of markers typical of the luminal epithelia phenotype ofbreast cells, whereas BT-474 and SK-BR-3 cells exhibit a weakly luminal epithelia-likephenotype. Distinct from the two phenotypes above, MDA-MB-231 represents a highlyinvasive “mesenchymal-like“ breast cancer cell line by expressing a high level of vimentin.In addition, the cervical cancer cell line HeLa was selected, because it represents the mostextensively studied cancer cell line thus far.Specific downregulation of cyclin B1 with siRNAWe were at first interested whether breast cancer cell lines are similarly sensitive to specificdownregulation of cyclin B1 by siRNA. As shown in Fig. 1A, while the protein level ofcyclin B1 in MDA-MB-231 was almost undetectable after treatment with siRNA1 or siRNA3,33% and 16% of cyclin B1 were still detectable in MCF-7 cells after treatment with siRNA1and siRNA3, respectively, relative to the protein level of cyclin B1 in control cells. In contrastto cyclin B1, β-actin was not affected. The protein level of Cdk1, the catalytic partner ofcyclin B1, hardly changed, indicating cyclin B1 knockdown by siRNAs was specific.Treatment of BT-474 and SK-BR-3 cells with siRNA1 and siRNA3 also reduced cyclin B1levels, albeit to a lower extent as compared to MDA-MB-231 and MCF-7 cells.10staining with monoclonal specific antibodies against cyclin B1 (data not shown). Furthermore,in consistence with downregulation of cyclin B1 protein, the kinase activity of Cdk1/cyclinB1 was decreased to 28% in cellular extracts from siRNA1-treated BT-474 cells as comparedto siGFP-treated cells (Fig. 1B). Similar results were also obtained in MCF-7 cells aftersiRNA administration (data not shown).Proliferation is inhibited in breast cancer cells with reduced cyclin B1As Cdk1/cyclin B1 is essential for the initiation of mitosis and required for cell division wesubsequently studied the impact of cyclin B1 downregulation on cell proliferation rate.Expectedly, a reduced proliferation was observed in all four breast cancer cell lines after 48 htreatment with siRNAs as compared to control cells treated with siGFP (Fig. 2). In particular,MCF-7 cells exhibited a strong inhibition of proliferation, followed by MDA-MB-231, BT-474 and SK-BR-3 cells after 48 h siRNA1-3 treatment against cyclin B1 (Fig. 2A-D, upperpanels). Analyses of cell cycle distribution displayed an accumulation of cells in G2/M phase,suggestive of a G2/M arrest, after treatment with siRNA1-3 targeting cyclin B1 (Fig. 2A-D,lower panels).Time kinetics of siRNA treatment in MCF-7 cellsIn order to study the effect of cycin B1 knockdown on proliferation, we subjected MCF-7cells to more detailed time dependent analysis. MCF-7 cells were treated with siRNA1 orsiRNA3 and harvested for Western blot analysis and proliferation assay at the indicated timepoints. The reduction of cyclin B1 protein levels were evident 24 h after siRNA1/3 treatmentas compared to control cells (Fig. 3A). This effect increased dramatically with time and after96 h cyclin B1 protein became undetectable. In line with the reduction of cyclin B1, theDownregulation of cyclin B1 was also corroborated by indirect immunofluorenscenceproliferation rate of MCF-7 cells was clearly inhibited at 24 h with siRNA3 treatment and theinhibition became more striking with longer exposure to siRNA1 (Fig. 3B).11investigated the combination of knockdown of cyclin B1 with irradiation. Irradiationenhanced the G2/M population and induced more apoptosis in cyclin B1 reduced MCF-712cells, compared to control cells (data not shown).Suppression of cyclin B1 renders cells more susceptible to taxolTaxol (Paclitaxel ® ), a taxane frequently used in multidrug regiments for the therapy of severalImpaired colony-forming ability and inhibited tumor growthsolid tumors, binds to the β-subunit of tubulin,thereby impairing the dynamics ofAnchorage independent cell growth is one of the hallmarks of malignant tumor cells. Givenmicrotubules by promoting their polymerization, leading to mitotic arrest and apoptosis [31].In order to explore cyclin B1 knockdown as a possible combination with taxol, we transfectedMCF-7 cells with siRNA3 and followed by further incubation with taxol. Cells wereharvested 48 h posttransfection. As depicted in Fig. 4A, cyclin B1 protein level was stronglyreduced after siRNA3 treatment. Cyclin B1 level was also decreased after treatment withtaxol at a low dosage of 3 ng/ml, possibly due to the induction of apoptosis, and almostdisappeared when cells were pre-treated with siRNA3 (Fig. 4A, upper panel). Indeed, PARP(poly(ADP-ribose) polymerase) was cleaved in cells treated with taxol and more stronglycleaved when siRNA3 was used together with taxol (Fig. 4A, middle panel), indicating thatthe combined treatment triggers more robust apoptotic response. In addition, proliferation wasinhibited to a higher extent after the combined treatment (Fig. 4B). The results were furthercorrelated with the cell cycle analysis: a G2/M population was more prominent in MCF-7cells after the combined treatment, compared to cells exposed to siRNA3 alone (Fig. 4C). Thedata indicate that the combined therapy activates more strongly caspase-3 independentapoptotic pathways in MCF-7 cells as the caspase-3 gene is deleted in MCF-7 cells [32].the notion of cyclin B1 deregulation being involved in neoplastic transformation andassociated with malignancy grade of tumors, we wondered whether knockdown of cyclin B1protein might translate into reduced colony-forming ability in MCF-7 cells. As shown in Fig.5A, the colony numbers of MCF-7 cells treated with siRNA1-3 were strongly reduced, inparticular, in MCF-7 cells treated with siRNA3, compared to controls.Cervical carcinoma HeLa cells express a high level of cyclin B1 (data not shown). In cellculture, only 25-30% HeLa cells were left after 48 h treatment with siRNA3 targeting cyclinB1 (data not shown). To further address whether cyclin B1 is required for aggressive growthof tumors in vivo, a xenograft experiment with HeLa cells was performed in nude mice. Micewere inoculated with HeLa cells treated 48 h with siRNA3 targeting cyclin B1 or siGFP. Asshown in Fig. 5B, tumor growth of HeLa cells treated with siRNA3 prior to inoculation waseffectually retarded, in comparison with the growth of control HeLa tumors. The data suggestthat cyclin B1 is indeed required in vivo for promoting proliferation of tumor cells and thereduction of cyclin B1 slows down the tumor growth.Similar results were also obtained in MDA-MB-231 cells: while siRNA3 or taxol alonereduced proliferation by 33% and 31%, respectively, relative to siGFP treatment, thecombination of siRNA3 with taxol resulted in a 65% reduction of proliferation (data notshown). Thus, the combined action of cyclin B1 knockdown together with taxol enhancesantiproliferative andproapoptotic responses in breast cancer cells. Furthermore, we

1314DiscussionThe prognosis of breast and cervical cancer patients has been improved during recent years,related partly to sophisticated surgery, radiotherapy and adjuvant systemic therapy. Despitethese advances, these cancers remain major clinical problems by causing considerablemorbidity and mortality in women worldwide. Apart from the standard approaches, novelpotent molecular agents for anticancer therapy are in great demand.In this communication we show that the knockdown of cyclin B1, the regulatory subunit ofCdk1, inhibited cell proliferation and induced apoptosis in various breast and cervical cancercell lines. Importantly, siRNA mediated cyclin B1 knockdown in combination withchemotherapeutical agent taxol, enhanced the antiproliferative effect on breast cancer cells.Interestingly, the reduction of cyclin B1 in MCF-7 cells impaired colony-forming ability, ahallmark of malignancy in tumor cells. Moreover, while control HeLa cells wereprogressively growing, the tumor growth of HeLa cells treated with siRNA targeting cyclinB1 prior to inoculation was strongly inhibited in nude mice, indicating cyclin B1 isindispensable for tumor growth in vivo. Taken together, the data strengthen the notion ofcyclin B1 being required for the survival and proliferation of breast and cervical cancer cellsand depletion/downregulation of cyclin B1 inhibits proliferation of cancer cells in vitro aswell as in vivo.Recent genetic evidence demonstrates that Cdk1 is the only Cdk sufficient to drive themammalian cell cycle because embryos from Cdk1 - / - mice fail to develop to the morula andblastocyst stages, whereas mouse embryos lacking all interphase Cdks (Cdk2, Cdk3, Cdk4and Cdk6) undergo organogenesis and develop to midgestation [33]. These data underscorethat Cdk1 is essential for cell cycle regulation and a major force driving cell proliferation.Cyclin B1, the regulatory subunit of Cdk1, controls the activity of Cdk1 as it associates withand thereby activates Cdk1, regulates its nuclear translocation and passively mediates itsinactivation when cyclin B1 is degraded at anaphase transition. Cyclin B1 is fundamental forcell proliferation. Uncontrolled expression of cyclin B1 is associated with neoplastictransformation and gynecological cancer development [5,9,11,34,35]. Overexpression ofcyclin B1 is believed to confer therapy resistance [8,11]. Thus, targeting cyclin B1, leadingconsequently to the inactivation of Cdk1, could be a promising specific strategy for cell cycleintervention against breast and cervical cancer.In this work, as a proof-of-concept, the RNA interference was used to downregulate/depletecyclin B1 and a clear antiproliferative effect was observed in all cancer cell lines studied.Among the breast cancer cell lines investigated, MCF-7 cells exhibited the strongestinhibitory effect on cell proliferation after cyclin B1 siRNA treatment, followed by MDA-MB-231, SK-BR-3 and BT-474 cells (Fig. 2), which possibly correlates with the cyclin B1level in exponential growing status of each cell line (data not shown). Although the proteinlevel of cyclin B1 in MDA-MB-231 cells was nearly undetectable after siRNA1 or siRNA3transfection, the inhibitory impact was moderate (Fig. 2D), suggesting the proliferation ofMDA-MB-231 cells is not necessarily dependent on the normal level of cyclin B1 and thelittle amount of remaining cyclin B1 might be sufficient for the survival of MDA-MB-231cells. Finally, SK-BR-3 and BT-474 cells were also not as sensitive to siRNA treatment (Fig.2B and 2C) as HeLa or MCF-7 cells. This could be due to the cellular context of SK-BR-3and BT-474 cells, e.g. Her-2/neu+, which very often leads to a hormone-independentproliferation of cells. Thus, unlike in MCF-7 cells, targeting Her-2/neu or other factorspromoting G1/S transition could be more effective for inhibiting cell cycle progression in SK-BR-3 and BT-474 cells, which has been shown by our previous study [36]. On that account,specific targeting of oncogene(s) in individual cancer cell lines, like Her-2/neu in SK-BR-3and BT-474 cells, or cyclin B1 in MCF-7, could improve breast cancer therapy. Collectively,downregulation/depletion of cyclin B1 worked effectively in all gynecological cancer celllines tested. However, only in some cell lines, such as MCF-7 and HeLa, cyclin B115knockdown resulted in a strong proliferative inhibition, most likely because proliferation inthose cell lines is more dependent on high cyclin B1 levels as compared to other cell lines.Taxane drugs represent the most important class of anticancer agents and are integrated in16as well as in vivo and sensitizes breast cancer cells to taxol. The data indicate that thecombination of reducing cyclin B1 with chemotherapeutic drugs could be a new strategy formolecular intervention in a subset of breast cancers.multidrug-regiments for the therapy of several solid tumors including gynecological cancers.Despite their relevant contribution in ameliorating the quality of life and overall survival ofcancer patients, drug resistance and site-effects hamper its wide usage. Therefore, it isdesirable to find new ways of lowering drug dosage without losing effectiveness to limit sideeffectsand possibly also to slow down drug resistance. In this work, cyclin B1 siRNA incombination with taxol, blocking entry into mitosis and targeting the transition of metaphaseto anaphase, respectively, demonstrated a high efficacy in inhibiting proliferation of MCF-7cells. The data suggest that specific targeting of cyclin B1 could sensitize some gynecologicalcancer cells, like MCF-7 and MDA-MB-231 cells, to conventional chemotherapeutic agentslike taxol, thereby reducing their side-effects by lowering their dosage.Taken together, the data from this work further strengthen the notion that cyclin B1 could bean attractive target for potential anticancer therapy. Inhibiting cyclin B1 function incombination with chemotherapeutic drugs could reinforce the antiproliferative effect in asubset of cancers. As RNA interference still faces the major challenge of systematic delivery[37], an alternative strategy could be small molecule inhibitors targeting cyclin B, as itscrystal structure is recently published [38]. In parallel to Cdk inhibitors, which have beenextensively under clinical investigations, small molecule inhibitors against cyclin B1 couldopen up a new door for specific molecular cancer therapy by interfering with its proteinstability, binding capacity to Cdk1 or its subcellular localization.Competing interestsThe authors declare that they have no competing interests.Authors’ contributionsIA and AK conducted cell cycle analyses, proliferation assays, Western blot analyses,apoptosis assays and mouse xenograft experiments in vivo. RY performed the kinetics ofMCF-7 cells and soft-agar assays. FR is involved in the combination therapy and assays. MKand RG coordinated this project. KS co-supervised this study and supported the manuscriptwriting. JY designed and supervised this study, and drafted the manuscript. All the authorsread and approved the final manuscript.AcknowledgementsThis work was supported by the Deutsche Krebshilfe (#107594). We gratefully acknowledgethe help of Dr. F. Eckerdt (Howard Hughes Medical Institute and Department ofPharmacology, University of Colorado, School of Medicine), who improved and modified themanuscript. We thank also Dr. S. Kappel for critical reading the manuscript. We are gratefulto Ms H. Beschmann and Ms S. Diehl, Department of Dermatology, University of Frankfurt,for supporting apoptotic analysis.ConclusionsThis work demonstrates that cyclin B1 is required for survival and proliferation of breast andcervical cancer cells. Downregulation of cyclin B1 inhibits proliferation of tumor cells in vitro

References1. Weigelt B, Peterse JL, 't Veer LJ: Breast cancer metastasis: markers and models.Nat Rev Cancer 2005, 5:591-602.2. Schiffman M, Castle PE, Jeronimo J, Rodriguez AC, Wacholder S: Humanpapillomavirus and cervical cancer. Lancet 2007, 370:890-907.3. Wang A, Yoshimi N, Ino N, Tanaka T, Mori H: Overexpression of cyclin B1 inhuman colorectal cancers. J Cancer Res Clin Oncol 1997, 123:124-127.4. Zhao M, Kim YT, Yoon BS, Kim SW, Kang MH, Kim SH, Kim JH, Kim JW, ParkYW: Expression profiling of cyclin B1 and D1 in cervical carcinoma. Exp Oncol2006, 28:44-48.5. Kawamoto H, Koizumi H, Uchikoshi T: Expression of the G2-M checkpointregulators cyclin B1 and cdc2 in nonmalignant and malignant human breastlesions: immunocytochemical and quantitative image analyses. Am J Pathol 1997,150:15-23.6. Soria JC, Jang SJ, Khuri FR, Hassan K, Liu D, Hong WK, Mao L: Overexpressionof cyclin B1 in early-stage non-small cell lung cancer and its clinical implication.Cancer Res 2000, 60:4000-4004.7. Banerjee SK, Weston AP, Zoubine MN, Campbell DR, Cherian R: Expression ofcdc2 and cyclin B1 in Helicobacter pylori-associated gastric MALT and MALTlymphoma : relationship to cell death, proliferation, and transformation. Am JPathol 2000, 156:217-225.8. Hassan KA, Ang KK, El Naggar AK, Story MD, Lee JI, Liu D, Hong WK, Mao L:Cyclin B1 overexpression and resistance to radiotherapy in head and necksquamous cell carcinoma. Cancer Res 2002, 62:6414-6417.9. Rudolph P, Kuhling H, Alm P, Ferno M, Baldetorp B, Olsson H, Parwaresch R:Differential prognostic impact of the cyclins E and B in premenopausal andpostmenopausal women with lymph node-negative breast cancer. Int J Cancer2003, 105:674-680.10. Nozoe T, Korenaga D, Kabashima A, Ohga T, Saeki H, Sugimachi K: Significance ofcyclin B1 expression as an independent prognostic indicator of patients withsquamous cell carcinoma of the esophagus. Clin Cancer Res 2002, 8:817-822.11. Suzuki T, Urano T, Miki Y, Moriya T, Akahira J, Ishida T, Horie K, Inoue S, SasanoH: Nuclear cyclin B1 in human breast carcinoma as a potent prognostic factor.Cancer Sci 2007, 98:644-651.12. EgloffAM,VellaLA,FinnOJ: Cyclin B1 and other cyclins as tumor antigens inimmunosurveillance and immunotherapy of cancer. Cancer Res 2006, 66:6-9.13. Suzuki H, Graziano DF, McKolanis J, Finn OJ: T cell-dependent antibodyresponses against aberrantly expressed cyclin B1 protein in patients with cancerand premalignant disease. Clin Cancer Res 2005, 11:1521-1526.1714. Innocente SA, Abrahamson JL, Cogswell JP, Lee JM: p53 regulates a G2checkpoint through cyclin B1. Proc Natl Acad Sci U S A 1999, 96:2147-2152.15. Passalaris TM, Benanti JA, Gewin L, Kiyono T, Galloway DA: The G(2) checkpointis maintained by redundant pathways. MolCellBiol1999, 19:5872-5881.16. Taylor WR, DePrimo SE, Agarwal A, Agarwal ML, Schonthal AH, Katula KS, StarkGR: Mechanisms of G2 arrest in response to overexpression of p53. MolBiolCell1999, 10:3607-3622.17. MacLachlan TK, Dash BC, Dicker DT, El Deiry WS: Repression of BRCA1through a feedback loop involving p53. J Biol Chem 2000, 275:31869-31875.18. Yin XY, Grove L, Datta NS, Katula K, Long MW, Prochownik EV: Inverseregulation of cyclin B1 by c-Myc and p53 and induction of tetraploidy by cyclinB1 overexpression. Cancer Res 2001, 61:6487-6493.19. Zoubine MN, Weston AP, Johnson DC, Campbell DR, Banerjee SK: 2-methoxyestradiol-induced growth suppression and lethality in estrogenresponsiveMCF-7 cells may be mediated by down regulation of p34cdc2 andcyclin B1 expression. Int J Oncol 1999, 15:639-646.20. Perks CM, Gill ZP, Newcomb PV, Holly JM: Activation of integrin and ceramidesignalling pathways can inhibit the mitogenic effect of insulin-like growth factor I(IGF-I) in human breast cancer cell lines. Br J Cancer 1999, 79:701-706.21. Schroeder MD, Symowicz J, Schuler LA: PRL modulates cell cycle regulators inmammary tumor epithelial cells. Mol Endocrinol 2002, 16:45-57.22. Cho NH, Kang S, Hong S, An HJ, Choi YH, Jeong GB, Choi HK: Elevation ofcyclin B1, active cdc2, and HuR in cervical neoplasia with human papillomavirustype 18 infection. Cancer Lett 2006, 232:170-178.23. Shen M, Feng Y, Gao C, Tao D, Hu J, Reed E, Li QQ, Gong J: Detection of cyclinb1 expression in g(1)-phase cancer cell lines and cancer tissues by postsortingWestern blot analysis. Cancer Res 2004, 64:1607-1610.24. Collecchi P, Santoni T, Gnesi E, Giuseppe NA, Passoni A, Rocchetta M, Danesi R,Bevilacqua G: Cyclins of phases G1, S and G2/M are overexpressed in aneuploidmammary carcinomas. Cytometry 2000, 42:254-260.25. Jin P, Hardy S, Morgan DO: Nuclear localization of cyclin B1 controls mitoticentry after DNA damage. JCellBiol1998, 141:875-885.26. Park M, Chae HD, Yun J, Jung M, Kim YS, Kim SH, Han MH, Shin DY:Constitutive activation of cyclin B1-associated cdc2 kinase overrides p53-mediated G2-M arrest. Cancer Res 2000, 60:542-545.27. Yuan J, Yan R, Kramer A, Eckerdt F, Roller M, Kaufmann M, Strebhardt K: CyclinB1 depletion inhibits proliferation and induces apoptosis in human tumor cells.Oncogene 2004, 23:5843-5852.1828. Yuan J, Kramer A, Matthess Y, Yan R, Spankuch B, Gatje R, Knecht R, KaufmannM, Strebhardt K: Stable gene silencing of cyclin B1 in tumor cells increasessusceptibility to taxol and leads to growth arrest in vivo. Oncogene 2006, 25:1753-1762.29. Yuan J, Eckerdt F, Bereiter-Hahn J, Kurunci-CsacskoE,KaufmannM,StrebhardtK:Cooperative phosphorylation including the activity of polo-like kinase 1 regulatesthe subcellular localization of cyclin B1. Oncogene 2002, 21:8282-8292.30. Lacroix M, Leclercq G: Relevance of breast cancer cell lines as models for breasttumours: an update. Breast Cancer Res Treat 2004, 83:249-289.31. Hernandez-Vargas H, Palacios J, Moreno-Bueno G: Telling cells how to die:docetaxel therapy in cancer cell lines. Cell Cycle 2007, 6:780-783.32. Simstein R, Burow M, Parker A, Weldon C, Beckman B: Apoptosis,chemoresistance, and breast cancer: insights from the MCF-7 cell model system.Exp Biol Med (Maywood ) 2003, 228:995-1003.33. Santamaria D, Barriere C, Cerqueira A, Hunt S, Tardy C, Newton K, Caceres JF,Dubus P, Malumbres M, Barbacid M: Cdk1 is sufficient to drive the mammaliancell cycle. Nature 2007, 448:811-815.34. Zhao M, Kim YT, Yoon BS, Kim SW, Kang MH, Kim SH, Kim JH, Kim JW, ParkYW: Expression profiling of cyclin B1 and D1 in cervical carcinoma. Exp Oncol2006, 28:44-48.35. Cho NH, Kang S, Hong S, An HJ, Choi YH, Jeong GB, Choi HK: Elevation ofcyclin B1, active cdc2, and HuR in cervical neoplasia with human papillomavirustype 18 infection. Cancer Lett 2006, 232:170-178.36. Faltus T, Yuan J, Zimmer B, Kramer A, Loibl S, Kaufmann M, Strebhardt K:Silencing of the HER2/neu gene by siRNA inhibits proliferation and inducesapoptosis in HER2/neu-overexpressing breast cancer cells. Neoplasia 2004, 6:786-795.37. Bumcrot D, Manoharan M, Koteliansky V, Sah DW: RNAi therapeutics: a potentialnew class of pharmaceutical drugs. Nat Chem Biol 2006, 2:711-719.38. Petri ET, Errico A, Escobedo L, Hunt T, Basavappa R: The crystal structure ofhuman cyclin B. Cell Cycle 2007, 6:1342-1349.1920Figure legendsFigure 1 - Reduced cyclin B1 protein and kinase activity of Cdk1/cyclin B1A. Upper panel: Breast cancer cells MCF-7, BT-474, SK-BR-3 and MDA-MB-231 cells weretreated with cyclin B1 siRNA1, siRNA3, GFP siRNA (GFP) or without treatment (C) for 48h. Cells were harvested and cellular lysates were prepared for Western blot analyses withantibodies targeting cylin B1, Cdk1 and β-actin. The later served as loading control. Lowerpanel: Quantification of cyclin B1 levels from Western blots (upper panel), normalized to β-actin. B. Kinase assay of Cdk1/cyclin B1 in vitro. BT-474 cells were treated with cyclin B1siRNA1 or siGFP or without treatment (C). 24 h later cells were lysed and Cdk1/cyclin B1complex was immunoprecipitated by using antibodies against cyclin B1 from cellularextracts. Normal IgG served as negative control. Kinase assays were carried out with theprecipitates in the presence of histone H1 as substrate.Figure 2 – Inhibited proliferation after downregulation of cyclin B1MCF-7 (A), BT-474 (B), SK-BR-3 (C) and MDA-MB-231 cells (D) were treated withsiRNA1-3, or siGFP or oligofectamine alone (OF). After 48 h cells were harvested for cellnumber counting (upper panels in Fig. 3 A-D) and for cell cycle analysis (lower panels in Fig.3 A-D). Cells without any treatment (C) served as control. The results of cell numbers areexpressed as mean ± SD (n = 3) and statistically analysed. *P < 0.05, **P < 0.01.Figure 3 - Time kinetics in MCF-7 cellsMCF-7 cells were treated with cyclin B1 siRNA1, siRNA3, siGFP or without treatment (C).Treated cells were harvested at different time points as indicated for Western blot analyses

21with antibodies against cyclin B1, Cdk1 and β-actin (A) and for cell number determinations(B). The results of cell numbers are expressed as mean ± SD (n =3).Figure 4 – More sensitive to taxol in MCF-7 cells with reduced cyclin B1A. MCF-7cellsweretransfectedwithsiRNA3orsiGFPand4hlaterfollowedwithtreatmentAMCF-7BT-474SK-BR-3MDA-MB-231cyclin B1of taxol (C+, siGFP+ and siRNA3+) or without taxol (C, siGFP and siRNA3). 48 h aftertransfection cells were harvested for Western blot analyses with antibodies as indicated. β-actin served as loading control. PARP: poly (ADP-ribose) polymerase. B. Cells were treatedas in A and cell numbers were counted. The results are expressed as mean ± SD (n =3)andstatistically analysed. *P < 0.05. C. MCF-7 cells were treated as in A and cell cycle wasanalyzed (upper panel) and distribution of cell cycle population was quantified (lower panel).Figure 5 - Impaired colony-forming ability and reduced tumor growthA. Colony formation assay. MCF-7 cells were treated with siRNA1-3 or siGFP. 48 h aftertransfection cells were transferred to soft-agar plates for further incubation. 4 weeks later thecolony numbers were scored with a microscope. The colony number of control cells wasregarded as 100% for calculation. The data were expressed as mean ± SD (n =3)andanalysed by Student’s t-test. *P < 0.05, **P < 0.01. B. Xenograft experiment in nude mice.1×10 6 of HeLa cells treated with siRNA3 or control siGFP were subcutaneously injected intoeach flank of nude mice. Each group contained 4 mice. Tumor sizes were measured every 2-3days and tumor volumes were calculated. The data were statistically analysed by Mann-cyclin B1 relativeprotein levelB1,61,20,80,40Figure 1MCF-7CsiGFPsiRNA1siRNA3CsiGFPsiRNA1siRNA3relative kinase activity(%)16012080400CsiGFPsiRNA1siRNA3CCBT-474CsiGFPsiRNA1siRNA3siGFPCsiGFPsiRNA1siRNA3SK-BR-3siGFPsiRNA1siRNA3siRNA1CBT-474MDA-MB-231IgGsiGFPsiRNA1siRNA3CsiGFPsiRNA1siRNA3β-actinCdk1Whitney U test. *P < 0.05, **P < 0.01.A10 4 cells/well%201510506040200MCF-7∗ ∗∗C OF siGFP siRNA1 siRNA2 siRNA3MCF-7C OF siGFP siRNA1 siRNA2 siRNA3BG0/G1SG2/M10 4 cells/well%302520151080604020050BT-474∗C OF siGFP siRNA1 siRNA2 siRNA3BT-474C OF siGFP siRNA1 siRNA2 siRNA3AB10.0CsiGFPCsiRNA1siRNA3siGFP24 h 48 h72 h 96 hCsiRNA1siRNA3siGFPCsiRNA1siRNA3siGFPsiRNA1siRNA3cyclin B1β-actinCdk1C10 4 cells/well161284SK-BR-3∗D10 4 cells/well302010∗MDA-MB-231∗10 4 cells/well7.55.02.524 h 48 h 72 h060%40C OF siGFP siRNA1 siRNA2 siRNA3SK-BR-3G0/G1SG2/M%06040C OF siGFP siRNA1 siRNA2 siRNA3MDA-MB-2310Figure 3CsiGFPsiRNA1siRNA3CsiGFPsiRNA1siRNA3CsiGFPsiRNA1siRNA320200Figure 2C OF siGFP siRNA1 siRNA2 siRNA30C OF siGFP siRNA1 siRNA2 siRNA3

ACsiGFPsiRNA3C+siGFP+siRNA3+cyclin B1PARPcleaved PARPA30MCF-7BC10 4 cells/well403020100CsiGFPβ-actin- - - + + + taxol (3 ng/ml)MCF-7siRNA3 C+ siGFP+ siRNA3+∗Btumor volume (x10 3 mm)colony number201004003002001000∗ ∗ ∗∗C OF siGFP siRNA1 siRNA2 siRNA3siGFPsiRNA3∗∗ P

ARTICLE IN PRESSBiochemical and Biophysical Research Communications xxx (2009) xxx–xxxContents lists available at ScienceDirectBiochemical and Biophysical Research Communicationsjournal homepage: www.elsevier.com/locate/ybbrcNon-anti-mitotic concentrations of taxol reduce breast cancer cell invasivenessTruong-An Tran a,b,1 , Ludovic Gillet a,b,1 , Sébastien Roger a,b , Pierre Besson a,b ,Edward White c , Jean-Yves Le Guennec a,b, *a Inserm, U921, 37000 Tours, Franceb Université François-Rabelais, 37000 Tours, Francec Institute of Membrane and Systems Biology, University of Leeds, LS2 9JT LEEDS, UKarticleinfoabstractArticle history:Received 9 December 2008Available online xxxxKeywords:TaxolTubulinCancer invasivenessNa V 1.5MetastasisMDA-MB-231MDA-MB-468Taxol is widely used in breast cancer chemotherapy. Its effects are primarily attributed to its anti-mitoticactivity. Microtubule perturbators also exert antimetastatic activities which cannot be explained solelyby the inhibition of proliferation. Voltage-dependent sodium channels (Na V ) are abnormally expressedin the highly metastatic breast cancer cell line MDA-MB-231 and not in MDA-MB-468 cell line. InhibitingNa V activity with tetrodotoxin is responsible for an approximately 0.4-fold reduction of MDA-MB-231 cellinvasiveness. In this study, we focused on the effect of a single, 2-h application of 10 nM taxol on the twocell lines MDA-MB-231 and MDA-MB-468. At this concentration, taxol had no effect on proliferation after7 days and on migration in any cell line. However it led to a 40% reduction of transwell invasion of MDA-MB-231 cells. There was no additive effect when taxol and tetrodotoxin were simultaneously applied.Na V activity, as assessed by patch-clamp, indicates that it was changed by taxol pre-treatment. We concludethat taxol can exert anti-tumoral activities, in cells expressing Na V , at low doses that have no effecton cell proliferation. This effect might be due to a modulation of signalling pathways involving sodiumchannels.Ó 2008 Elsevier Inc. All rights reserved.Taxol and its derivatives are widely used in the treatment ofbreast cancers [1]. The mechanism for their beneficial effects isthought to be the stabilization of microtubules during mitosiscausing a blockade of cell division [2]. This blockade can also involveparticular potassium channels, EAGs, known to be regulatedby the cytoskeleton and involved in the control of cell proliferation[3]. However, the effect of taxol on cell invasiveness, an importantparameter linked to the capacity of cells to produce metastases,has not been evaluated in breast cancer cells.Voltage-dependent sodium channels are known to be involvedin breast and lung cancer cell invasiveness [4,5]. These proteinshave been found in biopsies from metastatic tissues and are thusthought to be involved in the development of secondary tumours[6]. In breast cancer, the isoform expressed is Na V 1.5 [4,7]. Thischannel is expressed in the breast cancer cell line MDA-MB-231and not in MDA-MB-468 [4]. It promotes cell invasiveness and islikely to be controlled by intracellular regulators such as proteinkinases and/or cytoskeletal elements [4,8]. Depending on the cell* Corresponding author. Address: Inserm U921, N2C, Université François-Rabelais,Faculté de Médecine, 10 Bd Tonnellé, 37032 Tours, France. Fax: +33 2 47 36 6226.E-mail address: Jean-Yves.LeGuennec@Univ-Tours.Fr (J.-Y Le Guennec) .1 These authors contributed equally to this work.type in which it is expressed, Na V 1.5 is, or is not, sensitive to microtubuleperturbators [9].In the present study, we investigated whether stabilizing microtubulesby taxol regulates the invasiveness of MDA-MB-231 andMDA-MB-468 cells. Cell invasiveness was assessed by measuringtranswell migration on Matrigel Ò -coated inserts. The effect of taxolon Na V activity was evaluated in patch-clamp. The proportions offree vs polymerized tubulin were assessed by Western blot. Wefound that pre-treating the cells with a low concentration of taxol(10 nM) slightly increased microtubule polymerization. The effectsof TTX and taxol on cell invasivity were not additive. We concludethat taxol is responsible for the reduced invasion of MDA-MB-231cells through a mechanism interfering with Na V 1.5 signalling atdoses which do not interfere with cell mitosis.Materials and methodsCell culture. The human breast cancer cells MDA-MB-231 andMDA-MB-468 were obtained from the ATCC (Manassas, Virginia,USA) and cultured in Dulbecco’s modified Eagle’s medium (DMEM,Lonza, France) supplemented with 5% FCS at 37 °C in a humidified5% CO 2 incubator. For all experiments, cells were pre-treated for2 h with concentrations of taxol ranging from 10 nM (low concentration)to 10 lM (high concentration). Taxol was then removedprior to performing the different assays (proliferation, migration0006-291X/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.bbrc.2008.12.073Please cite this article in press as: T. -A. Tran et al., Non-anti-mitotic concentrations of taxol reduce breast cancer cell invasiveness, Biochem.Biophys. Res. Commun. (2009), doi:10.1016/j.bbrc.2008.12.073

ARTICLE IN PRESS2 T.-A. Tran et al. / Biochemical and Biophysical Research Communications xxx (2009) xxx–xxxand invasion). Drugs and chemicals were purchased from Sigma–Aldrich (France) and Latoxan (France).Cell survival and proliferation. Cells were seeded at 4 10 4 cellsper well in 12 wells of a 24-well plate for a given condition (includinguntreated controls) on two separate experiments and weregrown for a total of 5 days. The culture medium was changed everyother day. Growth and viability of cells were measured as a wholeby the tetrazolium salt assay [10] as previously described [4]. Cellproliferation was expressed as formazan 570 nm absorbance andexpressed as a ratio of treated to control values.Migration and in vitro invasion assay. To evaluate migration andinvasion, we performed transwell migration assays using inserts aspreviously described [4]. In these assays, we evaluated the capacityof cells to move through holes smaller than the cell diameter undera chemotactic gradient. To distinguish between migration andinvasion, inserts were either devoid of, or coated with Matrigel Ò .Therefore, in invasion assays, cells had to digest Matrigel Ò beforemigrating through the pores. In our experiments, we evaluated24 h migration and invasion. Cells were pre-treated for 2 h withthe different concentrations of taxol, washed and then seeded intothe inserts and were not in contact with taxol for the duration ofthe test. Migration was analyzed in 24-well plates with 8 lm poresize polyethylene terephtalate membrane cell culture inserts (Becton–Dickinson,France). The upper compartment was seeded with4 10 4 cells. The lower compartment was filled with culture mediumsupplemented with 10% FCS (fetal calf serum) as a chemoattractant.After 24 h at 37 °C, remaining cells were removed fromthe upper side of the membrane and cells that had migrated andwere attached to the lower side were stained with haematoxylinand counted on the whole insert, using a light microscope at200 magnification [4].In vitro invasion was assessed using the same inserts and thesame protocol as above but with the membrane covered with afilm of Matrigel Ò (Becton–Dickinson, France), mimicking the extracellularmatrix. Migration and invasion assays were performed intriplicate in eight separate experiments. For comparison betweenconditions, results obtained for migration and invasion were normalizedto the control condition.Polymerization of microtubules. The evaluation of the polymerizationof microtubules was performed by Western blot analysis[11]. Specific primary monoclonal antibodies were directed towarda-tubulin (Sigma–Aldrich, France). MDA-MB-231 cells, in theirsubconfluent state, were collected by trypsinization. After centrifugation(1000g for 5 min), the culture medium was replaced by amicrotubule stabilizing buffer [11,12]. After further centrifugations(two at 30,000g for 1 h 30 min), the free and polymerized tubulinfractions were collected in the supernatants [11,12]. The total proteinconcentrations of cell lysates were determined using thebicinchoninic acid method (Bio-Rad, USA) (triplicate experiments).Protein samples (20 lg per lane) were diluted in reducing (DTT)and denaturing (SDS) sample buffer, boiled for 3 min, separatedby SDS–PAGE on 10% gels [13], then transferred onto a PVDF membrane(Millipore, USA). After saturation for 2 h in 5% non-fat milkTBS solution containing 0.5% Tween 20 and 5% non-fat milk(TBST5), the membrane was incubated overnight at 4 °C with theprimary antibody (1/1000) in a 2% non-fat milk TBST5 solution.The membrane was further incubated for 1 h at room temperaturewith a goat anti-mouse (1/1000 in TBST5 solution) horseradishperoxidase-conjugated secondary antibody (Sigma–Aldrich,France). Finally, immunoblots were revealed with the ECL immunodetectionsystem (Amersham Biosciences, UK). Pre-stainedmolecular masses were from Bio-Rad. After washing to removeECL, the membrane was incubated (2 h at room temperature) withb-actin specific monoclonal antibodies (Santa Cruz, USA) to correctfor protein quantity in each well. Immunoblots were revealed withthe ECL immunodetection system a second time.Electrophysiology. Patch pipettes were pulled from borosilicateglass to a resistance of 4–6 MX. Currents were recorded, in rupturedwhole-cell configuration, under voltage-clamp mode at room temperatureusing an Axopatch 200 B patch-clamp amplifier (AxonInstrument, USA). Analogical signals were filtered at 5 kHz, and sampledat 10 kHz using a 1322A Digidata converter. Cell capacitanceand series resistances were electronically compensated by about60%. The P/2 sub-pulse correction of cell leakage and capacitancewas used to study Na + current (I Na ). Sodium currents were recordedby depolarizing the cells from a holding potential of 100 mV to amaximal test pulse of 5 mV for 30 ms every 500 ms. The protocolAMDA-MB-231Relative proliferation (%)BMDA-MB-231Relative migration/invasionCRelative invasion10080604020****** **0 ****** ** ********** **0 1 2 3 4 5 6 71.00.80.60.40.20.01.00.80.60.40.20.0ControlControlControlTaxol 10 nMTaxol 100 nMTaxol 1 µMTaxol 10 µMDays after treatment**Taxol10 nM∗∗**MDA-MB-231**Taxol100 nM10 nM taxol**MigrationInvasion**Taxol1 µMMDA-MB-468Fig. 1. Effect of taxol on breast cancer cell proliferation, migration and invasivity.Taxol, at the concentrations of 10 nM, 100 nM, 1 lM and 10 lM, was applied for 2 hthen washed off. Effect of taxol, on; (A) cell proliferation, mean of threeexperiments. (B) Cell migration (line filled bars) and invasion (hatched bars) meanof three and four experiments in triplicate, respectively. (C) Effect of 10 nM taxol onMDA-MB-231 and MDA-MB-468 cell invasiveness. Each bar represents the mean ofeight experiments performed in triplicate. ** p < 0.05 vs control (one-way ANOVA).Please cite this article in press as: T. -A. Tran et al., Non-anti-mitotic concentrations of taxol reduce breast cancer cell invasiveness, Biochem.Biophys. Res. Commun. (2009), doi:10.1016/j.bbrc.2008.12.073

ARTICLE IN PRESST.-A. Tran et al. / Biochemical and Biophysical Research Communications xxx (2009\) xxx–xxx 3used to build sodium current–voltage (I Na –V) relationships was asfollows: from a holding potential of 100 mV, the membrane wasstepped to potentials from 90 to +60 mV, with 5-mV increments,for 50 ms at a frequency of 2 Hz. Availability–voltage relationshipswere obtained by applying 50 ms prepulses using the I Na –V curveprocedure followed by a depolarizing pulse to 5 mV for 50 ms. Inthis case, currents were normalized to the amplitude of the test currentwithout a prepulse (i.e., from 100 mV). Conductance throughNa + channels (g Na ) was calculated as already described [4]. The conductance–voltageand availability–voltage relationships were fittedwith a Boltzman equation to get the EC 50 and slope of the relationship.Current amplitudes were normalized to cell capacitance andexpressed as current density (pA/pF).Solutions. The physiological saline solution (PSS) had the followingcomposition (in mM): NaCl 140, KCl 4, MgCl 2 1, CaCl 2 2, D-glucose11.1 and Hepes 10, adjusted to pH 7.4 with 1 M NaOH. Theintrapipette solution had the following composition (in mM):zK-glutamate, 125, KCl 20, CaCl 2 0.37, MgCl 2 1, Mg–ATP 1, EGTA1, Hepes 10, adjusted to pH 7.2.Statistics. Data are described as means ± standard error of themean (n = number of cells). Kruskal–Wallis one-way ANOVA onranks followed by a Student–Newmann–Keuls test were used tocompare cell proliferation, migration and invasion in control conditionsor in the presence of different taxol concentrations. Twoways repeated measures ANOVA on ranks followed by a Student–Newmann–Keulstest was used to compare availability andconductance relationship. p < 0.05 was considered as statisticallysignificant.ResultsWe evaluated MDA-MB-231 cell viability over 7 days in controlconditions or after 2 h pre-treatment with taxol (10 nM, 100 nM,1 lM and 10 lM). As shown in Fig. 1A, taxol reduced relative cellA1.0MDA-MB-231Relative invasion0.80.60.40.2******0.0ControlTTXTaxolTTX + TaxolB4036200 pA150 pAI NaCurrent (pA/pF)32282420161285 ms5 ms4ControlTaxolCActivationControl1.0 *Taxol***0.8*0.6*0.4*0.2*DAvailability1.00.80.60.40.2ControlTaxol0.0-80 -60 -40 -20 0 20Voltage (mV)0.0-120 -100 -80 -60 -40 -20 0Voltage (mV)Fig. 2. Involvement of Na V activity in taxol-sensitive cell invasiveness. (A) Effect of 10 nM taxol and 30 lM TTX on MDA-MB-231 cell invasiveness. At this concentration TTXinhibits Na V 1.5 [4]. Inhibitors were applied separately or in combination as indicated on the graph. Each bar represents the mean of eight experiments performed in triplicate.** p < 0.05 vs control. (B) The amplitude of I Na was measured as the difference between the peak and the end-pulse current when the cell was depolarized to 5 mV for 50 msfrom a holding potential of 100 mV. The individual values obtained from control (left, n = 12 cells) and taxol (right, n = 12 cells) experiments are given and the bars indicatethe median. Representative currents in both conditions are given at the top of the panel. (C) Taxol induced a significant leftward shift of the activation properties of Na V 1.5. Incontrol conditions (filled squares), V 1/2 = 19.8 ± 0.1 mV (n = 12 cells) and slope = 6.1 ± 0.1. After the treatment with taxol (empty squares), V 1/2 = 29.3 ± 0.1 mV and slope5.9 ± 0.1 (n = 12 cells). (D) Taxol had no effect on the availability of the Na V 1.5. In control conditions (filled squares), V 1/2 = 62.3 ± 0.2 mV (n = 12 cells) and width = 9.2 ± 0.2.After the treatment with taxol (empty squares), V 1/2 = 64.7 ± 0.2 mV and width = 9.2 ± 0.2 (n = 12 cells). * p < 0.05.Please cite this article in press as: T. -A. Tran et al., Non-anti-mitotic concentrations of taxol reduce breast cancer cell invasiveness, Biochem.Biophys. Res. Commun. (2009), doi:10.1016/j.bbrc.2008.12.073

ARTICLE IN PRESS4 T.-A. Tran et al. / Biochemical and Biophysical Research Communications xxx (2009) xxx–xxxnumber for concentrations above 10 nM. A 2-h pre-treatment with10 nM taxol had no effect on cell proliferation.On this basis, we evaluated the effects of 10 nM taxol on the cancercell properties involved in the development of metastases. Thetwo parameters evaluated were cell migration and invasion.Fig. 1B shows that migration was not affected by 10 nM taxol whileit was significantly reduced by higher concentrations (22.6% ± 0.5inhibition at 100 nM and 30.1% ± 1.9 at 1 lM). For invasion, 10 nMtaxol was sufficient to reduce cell invasion by about 40%. Higher concentrationsof taxol had no greater effect on cell invasion.Since the Matrigel Ò invasion properties of MDA-MB-468 cellsare similar to those of MDA-MB-231 cells [4], we evaluated the effectof 10 nM taxol on the two cell lines invasivity. As seen onFig. 1C, 10 nM taxol (which had no effect on MDA-MB-468 cell proliferationand migration, data not shown) reduced MDA-MB-231cells invasiveness while it did no affect MDA-MB-468 cellsinvasiveness.Invasiveness is under the control of Na V activity in MDA-MB-231 but not MDA-MB-468 cells [4]. Inhibition of Na V 1.5 by tetrodotoxinis known to reduce cancer cell invasiveness without affectingcell migration [4]. To evaluate if the effect of taxol involves Na V 1.5,we measured the effects of 30 lM TTX and 10 nM taxol, alone andin combination. Fig. 2A shows that 30 lM TTX, known to block Na Vactivity [4] reduced cell invasiveness by about 40% as did 10 nMtaxol. Adding 30 lM TTX after the pre-treatment with taxol hadadditional effect upon the reduction of invasiveness. This suggeststhat taxol might act through a regulation of Na V 1.5 signalling.Patch-clamp experiments were then performed to evaluate theeffects of 2 h of pre-treatment with 10 nM taxol on I Na . We havepreviously shown that maximum I Na current was obtained whencells were depolarized from 100 to 5mV[4]. We thus comparedthe amplitudes of the current elicited by a similar voltage-clampprotocol in control conditions with those of cells pre-treated withtaxol. Fig. 2B shows that the distribution of the maximal currentamplitude was not affected by taxol. Since we did not observe areduction of the current, we checked other functional parametersof Na V 1.5, namely activation and availability. A significant leftwardshift of the activation was observed with taxol (Fig. 2C) while it hadno effect on channel availability (Fig. 2D). This result indicates thatpolymerization of microtubules regulates the channel gating.To evaluate the polymerization of microtubules, we performedWestern blot analysis using antibodies directed against a-tubulinin conditions allowing the separation of free and polymerizedforms of this microtubule subunit (Fig. 3A). Fig. 3B summarisesthe results obtained in three separate experiments. While about40% of a-tubulin subunits were in a polymerized form in the controlconditions, this proportion increased in a dose-dependentmanner when cells were pre-treated with 100 nM and 1 lM taxol,indicating that a unique short application of a low dose of taxol issufficient to give a measurable effect on microtubule polymerization.At 10 nM, there was a tendency for an increased tubulin polymerizationwhich was not statistically significant.DiscussionTaxol and its derivatives, such as docetaxel, are used in breastcancer chemotherapy [1]. The classically accepted mechanism ofaction is the stabilization of the mitotic spindle, leading to an inhibitionof cell proliferation. In this study, we showed that a unique,short, 2 h long, application of low doses of taxol, ineffective uponcell proliferation, had significant anti-invasive effects on the cancercell line MDA-MB-231. This cell line expresses Na V 1.5 in contrastto another invasive cell line, MDA-MB-468 [4]. In patients, it hasbeen reported that the presence of mRNA coding for Na V 1.5 arepositively related to lymph node metastasis [14]. Thus, the effectsof low concentrations of taxol on cell invasiveness are likely to prevent/reducethe formation of metastases. This finding probably accountsfor most of the effects of taxol since it is known that areduced invasive capacity of cancer cells is associated with a reducedtumour growth in situ. We propose that taxol may reduceinvasion by interfering with Na V 1.5 signalling pathways. Indeed,we found that the reduction of invasiveness induced by taxolwas not additive with the effect of TTX. The reduction of cell migrationobserved for concentrations of taxol above 10 nM are in thesame range as the reduction of cell proliferation. The toxicologicaleffects of taxol, more than anti-migrative effects, can explain thisobservation.It has been found that gating properties of the endogenous cardiacchannels Na V 1.5 were affected by taxol [9]. This finding is inline with our results. The lack of further I Na reduction after taxoltreatment suggests that the involvement of Na V activity in thecells’ invasive properties is rather complex and does not involvesolely a direct ionic or electrical effect, taxol might affect proteinsother than Na V 1.5 and which are involved in the signalling pathwayregulated by Na V 1.5 activity. For example, it is known thatperturbation of microtubules by colchicine can modulate the activityof sodium channels in rat cardiac cells through perturbation ofGTP homeostasis [15]. It is thus possible that this kind of effect occursand modulates Na V 1.5 activity, or other participants in theNa V 1.5 signalling pathway, to produce a reduced invasiveness.In conclusion, taxol might exert some anti-cancer propertiesthrough mechanisms other than its well-known anti-mitotic activity.A reduction of cell invasiveness can be observed in cell expressingNa V at concentrations of taxol (10 nM) having no effect on cellmitosis. At this concentration, no significant global tubulin polymerizationwas found. However, a localised, sub-membrane increasein polymerization may occur and sensitivity of membraneproteins such like Na V 1.5 to local perturbations of the cytoskeletoncould explain the observed effects upon cell invasiveness. It seemsthat this effect involves the Na V 1.5 signalling pathways but theprecise mechanisms remain to be determined. This will be doneA Free PolymerizedBkDa755037Tubulin relative proportion1.00.90.80.70.60.50.40.30.20.10.0ControlControlTaxol 1000 nMTaxol 100 nMTaxol 10 nMFree-tubulin**Polymerized-tubulinControlTaxol 1000 nMTaxol 100 nMTaxol 10nM**10 100 1000Taxol (nM)α-Tubulinβ-ActinFig. 3. Effect of treating cancer cells for 2 h with taxol on the relative concentrationof free and polymerized a-tubulin. (A) Example of gels obtained for free andpolymerized a-tubulin. (B) Relative proportion of normalized free and polymerizeda-tubulin in control conditions and after pre-treatment with 10, 100 and 1000 nMtaxol, Tubulin levels were normalized to b-actin. Mean results obtained from threedifferent experiments. ** p < 0.05 vs control.Please cite this article in press as: T. -A. Tran et al., Non-anti-mitotic concentrations of taxol reduce breast cancer cell invasiveness, Biochem.Biophys. Res. Commun. (2009), doi:10.1016/j.bbrc.2008.12.073

ARTICLE IN PRESST.-A. Tran et al. / Biochemical and Biophysical Research Communications xxx (2009\) xxx–xxx 5only when we will have a better understanding of the molecularlink between Na V and invasiveness.AcknowledgmentsWe thank the Université François-Rabelais de Tours for fundingthe visit of EW to Inserm U921. This work was financed in part by agrant from the Ligue Contre le Cancer de la Région Centre.References[1] G. Dang, C. Hudis, Adjuvant taxanes in the treatment of breast cancer: nolonger at the tip of the iceberg, Clin. Breast Cancer 7 (2006) 51–58.[2] S. Horwitz, Taxol (Paclitaxel): mechanisms of action, Ann. Oncol. 5 (1994)S3–S6.[3] J. Camacho, A. Sanchez, W. Stühmer, L. Pardo, Cytoskeletal interactionsdetermine the electrophysiological properties of human EAG potassiumchannels, Pflügers Arch. 441 (2000) 167–174.[4] S. Roger, P. Besson, J.-Y. Le Guennec, Involvement of a novel fast inward sodiumcurrent in the invasion capacity of a breast cancer cell line, Biochim. Biophys.Acta 1616 (2003) 107–111.[5] S. Roger, J. Rollin, A. Barascu, P. Besson, P.-I. Raynal, S. Iochmann, M. Lei, P.Bougnoux, Y. Gruel, J.-Y. Le Guennec, Voltage-gated sodium channelspotentiate the invasive capacities of human non-small-cell lung cancer celllines, Int. J. Biochem. Cell Biol. 39 (2007) 774–786.[6] S. Roger, M. Potier, C. Vandier, P. Besson, J.-Y. Le Guennec, Voltage-gatedsodium channels: new targets for epithelial cancer therapy?, Curr Pharm. Des.12 (2006) 3681–3695.[7] S. Judé, S. Roger, E. Martel, P. Besson, S. Richard, P. Bougnoux, J.-Y. Le Guennec,Dietary long-chain omega-3 fatty acids of marine origin: a comparison of theirprotective effects on coronary heart disease and breast cancers, Prog. Biophys.Mol. Biol. 90 (2006) 299–325.[8] S. Roger, P. Besson, J.-Y. Le Guennec, Influence of the whole cell patch-clampconfiguration on electrophysiological properties of the voltage-dependentsodium current expressed in MDA-MB-231 breast cancer cells, Eur. Biophys. J.33 (2004) 274–279.[9] V. Maltsev, A. Undrovinas, Cytoskeleton modulates coupling betweenavailability and activation of cardiac sodium channel, Am. J. Physiol. 273(1997) H1832–H1840.[10] T. Mosmann, Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays, J. Immunol. Methods 65(1983) 55–63.[11] H. Tsutsui, K. Ishihara, G. Cooper 4th, Cytoskeletal role in the contractiledysfunction of hypertrophied myocardium, Science 260 (1993) 682–687.[12] M. Takahashi, H. Shiraishi, Y. Ishibashi, K. Blade, P. McDermott, D. Menick, D.Kuppuswamy, G. Cooper IV, Phenotypic consequences of b 1 -tubulin expressionand MAP 4 decoration of microtubules in adult cardiocytes, Am. J. Physiol.Heart Circ. Physiol. 285 (2003) H2072–H2083.[13] U.K. Laemmli, Cleavage of structural proteins during the assembly of the headof bacteriophage T4, Nature 227 (1970) 680–685.[14] S. Fraser, J. Diss, A. Chioni, M. Mycielska, H. Pan, R. Yamaci, F. Pani, Z. Siwy, M.Krasowska, Z. Grzywna, W. Brackenbury, D. Theodorou, M. Koyuturk, H. Kaya,E. Battaloglu, M. De Bella, M. Slade, R. Tolhurst, C. Palmieri, J. Jiang, D.Latchman, R. Coombes, M. Djamgoz, Voltage-gated sodium channel expressionand potentiation of human breast cancer metastasis, Clin. Cancer Res. 11(2005) 5381–5389.[15] D. Motlagh, K. Alden, B. Russell, J. Garcia, Sodium current modulation by atubulin/GTP coupled process in rat neonatal cardiac myocytes, J. Physiol. 540.1(2002) 93–103.Please cite this article in press as: T. -A. Tran et al., Non-anti-mitotic concentrations of taxol reduce breast cancer cell invasiveness, Biochem.Biophys. Res. Commun. (2009), doi:10.1016/j.bbrc.2008.12.073

Oncogene (2008), 1–12& 2008 Macmillan Publishers Limited All rights reserved 0950-9232/08 $32.00www.nature.com/oncORIGINAL ARTICLEAutotaxin protects MCF-7 breast cancer and MDA-MB-435 melanomacells against Taxol-induced apoptosisN Samadi 1 , C Gaetano 2 , IS Goping 2 and DN Brindley 21Department of Laboratory Medicine and Pathology, University of Alberta, Edmonton, Alberta, Canada and 2 Departmentof Biochemistry (Signal Transduction Research Group), University of Alberta, Edmonton, Alberta, CanadaAutotaxin (ATX) promotes cancer cell survival, growth,migration, invasion and metastasis. ATX converts extracellularlysophosphatidylcholine (LPC) into lysophosphatidate(LPA). As these lipids have been reported toaffect cell signaling through their own G-protein-coupledreceptors, ATX could modify the balance of this signaling.Also, ATX affects cell adhesion independently of its catalyticactivity. We investigated the interactions of ATX,LPC and LPA on the apoptotic effects of Taxol, which iscommonly used in breast cancer treatment. LPC had nosignificant effect on Taxol-induced apoptosis in MCF-7breast cancer cells, which do not secrete significant ATX.Addition of incubation medium from MDA-MB-435melanoma cells, which secrete ATX, or recombinat ATXenabled LPC to inhibit Taxol-induced apoptosis ofMCF-7 cells. Inhibiting ATX activity blocked this protectionagainst apoptosis. We conclude that LPC has nosignificant effect in protecting MCF-7 cells against Taxoltreatment unless it is converted to LPA by ATX. LPAstrongly antagonized Taxol-induced apoptosis throughstimulating phosphatidylinositol 3-kinase and inhibitingceramide formation. LPA also partially reversed theTaxol-induced arrest in the G2/M phase of the cell cycle.Our results support the hypothesis that therapeuticinhibition of ATX activity could improve the efficacy ofTaxol as a chemotherapeutic agent for cancer treatment.Oncogene advance online publication, 15 December 2008;doi:10.1038/onc.2008.442Keywords: ceramides; chemotherapy; chemoresistance;lysophosphatidate; lysophosphatidylcholine; phosphatidylinositol3-kinaseIntroductionCorrespondence:Professor DN Brindley, Department of Biochemistry(Signal Transduction Research Group), University of Alberta,Edmonton, Alberta, Canada T6G 2S2.E-mail:david.brindley@ualberta.caReceived 24 July 2008; revised 8 October 2008; accepted 1 November2008Breast cancer is the most common malignancy amongwomen in North America and approximately one-thirdof these women develop metastases and die (Jemal et al.,2006). Dysregulation of normal mechanisms of apoptosisplay an important role in the pathogenesis andprogression of breast cancer. Importantly, the efficacy ofchemotherapy can be compromised by the survivalsignals that tumor cells receive (Krajewski et al., 1999).There is a strong association of autotaxin (ATX)expression with breast cancer cell survival, growth,migration, invasion and metastasis (Nam et al., 2000,2001; Umezu-Goto et al., 2002; Yang et al., 2002; Hamaet al., 2004). ATX was originally isolated from humanmelanoma A2058 cells (Stracke et al., 1992) and itgenerates lysophosphatidate (LPA) from circulatinglysophosphatidylcholine (LPC). Although the involvementof LPA and ATX in the invasiveness of breastcancer has been studied (Yang et al., 2002), relativelylittle is known about how ATX might confer chemoresistance.First, the substrate of ATX, LPC, hasbeen postulated to be an extracellular signaling lipidby acting on G2A and GPR4 (Kabarowski et al., 2001;Zhu et al., 2001; Rikitake et al., 2002; Lin and Ye, 2003;Radu et al., 2004; Kim et al., 2005). Unsaturated LPC issecreted by the liver (Brindley, 1993) and saturated LPCis produced by circulating lecithin:cholesterol acyltransferasein high-density lipoproteins (Aoki et al.,2002). LPC is present in blood at up to 200 mM(Moolenaar et al., 2004). ATX could, therefore, regulatecell activation through changing signaling by LPCversus LPA. Secondly, ATX decreases the adhesion ofoligodendrocytes to the extracelllular matrix through anon-catalytic mechanism involving its C-terminus, andthis facilitates morphological remodeling (Dennis et al.,2005). This suggests that ATX is a matrix-cellularprotein that signals through integrin-dependent focaladhesion assembly and consequently cell interactionswith the extracellular matrix (Fox et al., 2004). This andother non-catalytic effects of ATX could contribute toits association with the aggressiveness of cancer cells.Autotaxin provides a major route for generatingextracellular LPA, which is present at up to 20 mM inblood and extracellular fluid (Moolenaar et al., 2004;Yue et al., 2004). LPA is produced by activated plateletsto facilitate wound healing and is secreted by cancer cells(Fang et al., 2000; Radeff-Huang et al., 2004). ExtracellularLPA has been implicated in the etiology ofhuman cancer, as it stimulates cell growth, proliferation,differentiation, motility and survival (Mills and Moolenaar,2003; Brindley, 2004). Diverse actions of LPA aremediated by at least six G-protein coupled receptors

2present on the cell surface:LPA 1 /EDG2, LPA 2 /EDG4,LPA 3 /EDG7, LPA 4 /GPR23/p2y9, LPA 5 /GRP92 andLPA 6 /p2y5 (Lee et al., 2006; Pasternack et al., 2008).The expression of these receptors is cell specific,enabling different cells to respond in a unique mannerthrough signaling pathways that are activated by G i ,G s ,G q and G 12/13 . These signaling cascades includephosphatidylinositol 3-kinase (PI3K) and Akt, theextracellular signal-regulated kinase (ERK) pathway,and the small G-protein, RhoA, that mediates cellsurvival (Ye et al., 2002; Tigyi and Parrill, 2003; Radeff-Huang et al., 2004). Activation of LPA receptors alsodecreases the abundance of the p53 tumor suppressor,thereby promoting cell survival and cell cycle progression(Murph et al., 2007).Lysophosphatidate levels are high in ascites fluid andplasma of patients with ovarian tumors (Xu et al., 1998;Fang et al., 2002; Mills and Moolenaar, 2003). LPApromotes ovarian tumor development through increasedcyclin D expression (Chappell et al., 2001). LPA alsoincreases vascular endothelial growth factor production,which stimulates angiogenesis (So et al., 2005). In a coloncancer cell line, LPA increases the synthesis of macrophagemigration inhibitory factor, which promotestumor growth (Sun et al., 2003). However, the effect ofLPA on cell survival and apoptosis varies among celltypes. LPA mediates survival of ovarian cancer cells,macrophages, fibroblasts and neonatal cardiac myocytes,while promoting apoptosis in hippocampal neurons andPC12 cells (Fang et al., 2002; Ye et al., 2002).Taxol is a microtubule-stabilizing agent that interfereswith spindle microtubule dynamics causing cell cyclearrest and apoptosis through interaction with b-tubulin(Bergstralh and Ting, 2006). This causes lateral polymerizationand microtubule stability (Snyder et al.,2001). Taxol is widely used for treating metastatic andearly-stage breast cancer, with benefits in overall anddisease-free survival. However, resistance to Taxol iscommon and there is a need to identify patients who willrespond to treatment (McGrogan et al., 2008).In this study, we established that ATX protects MCF-7 breast cancer cells and MDA-MB-435 melanoma cellsagainst Taxol-induced apoptosis through its productionof LPA from the LPC that bathes most cells. LPA isstrongly antagonistic to Taxol-induced apoptosis,whereas LPC has no significant effect. From our results,we predict that identifying patients who express highATX activity and then inhibiting this activity couldprovide an important adjuvant for improving theefficacy of Taxol in cancer treatment.ResultsLysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alExpression of ATX in MCF-7 and MDA-MB-435 cancercellsATX expression in the MCF-7 and MDA-MB-435 cellsused in our experiments was compared. Real-timereverse transcriptase–PCR (RT–PCR) analysis showedthat MDA-MB-435 cells expressed relatively high levelsof ATX mRNA compared with MCF-7 cellsRelative fluorescence105Relative ATX mRNA6420MCF-7mRNAMDA-MB-435MCF-7MediumMDA-MB-435Western blotkDa MDA-MB-435 MCF-725015010075ATX activity00 25 50 75 100 125Time (minutes)(Figure 1a). MDA-MB-435 cells secreted ATX proteininto the serum-free medium, whereas ATX was notdetectable in the equivalent medium concentrated fromMCF-7 cells (Figure 1b). ATX activity secreted byMDA-MB-435 cells was about 24 times higher(Po0.001) than that from MCF-7 cells (Figure 1c).Therefore, MCF-7 cells provide a system for studyingapoptosis in the absence of ATX and for examining theeffects of introducing ATX.LPA protects MCF-7 cells from Taxol-induced apoptosisWe next established conditions for studying Taxolinducedapoptosis and how LPC and LPA modify this†rATXFigure 1 Differential expression of autotoxin (ATX) in MDA-MB-435 and MCF-7 cells. In (a), mRNA for ATX was determinedby real-time RT-PCR and expressed relative to that of glyceraldehyde3-phosphate dehydrogenase (GAPDH). (b) A westernblot obtained by loading 40 mg of total protein from threeindependent samples of concentrated serum-free medium obtainedfrom either MDA-MB-435 or MCF-7 cells is shown. A standard ofrecombinant rATX is shown on the right. In (c), 20 concentratedserum-free medium from MCF-7 and MDA MB-435 cells wasassayed for ATX activity using the FS-3 compound. Fluorescencewas measured at excitation and emission wavelengths of 485 and538 nm, respectively. The results are shown as the average ofrelative fluorescence activity±s.e.m. from three independentexperiments. w Po0.001.Oncogene

Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et al3Figure 2 Protective effect of lysophosphatidate (LPA) on Taxolinducedapoptosis in MCF-7 cells. (a) The percentage of apoptoticnuclei from MCF-7 cells that were incubated with 0–200 nM Taxolfor 24 h (&) or48h(m) is shown. (b) The percentage of viable cellsrecovered from the wells expressed relative to untreated cells isshown. (c and d) The percentage of apoptotic nuclei and viablecells, respectively, that were recovered after adding differentconcentration of LPA with 50 nM Taxol and incubating for 24 hincubation is shown. Results are expressed as means±s.e.m. fromthree independent experiments.process. Apoptosis was detected both by morphologicalexamination using 4,6-diamidino-2-phenylindole (DAPI)staining and by measuring mitochondrial tetramethlyrhodamineethyl ester (TMRE) uptake. We used mediumsupplemented with 10% charcoal-treated fetal bovineserum (FBS) to remove lipid mediators includingLPC and LPA. This treatment also removed ATX, aswe could not detect significant ATX activity in mediacontaining 10% charcoal-treated FBS (results notshown).Treating MCF-7 cells with 50 nM Taxol produced anoptimum increase in the percentage of recovered apoptoticnuclei from 0.8% to about 12% after 24 h and 26%after 48 h (Figure 2a), and a similar effect was observedwith MDA-MB-435 cells (results not shown). Thisprovides part of the picture, as treatment with 50 nMTaxol for 24 and 48 h decreased the number of cellsattached to the dishes by about 60 and 70%, respectively(Figure 2b). We could not recover most of these cellsfrom the medium presumably because of completefragmentation. These results are compatible with thoseof Sunders et al. (1997), who concluded that Taxolinduces both apoptotic and nonapoptotic death inMCF-7 cells.The next experiments determined whether LPA canprotect against apoptosis induced by 50 nM Taxol over24 h. LPA (0.5 mM) alone had no significant effect on thepercentage of apoptotic nuclei, but it produced aFigure 3 Effect of lysophosphatidate (LPA) on Taxol-inducedchange in mitochondrial potential. (a) The percentage of tetramethlyrhodamineethyl ester (TMRE)-negative MCF-7 cells thatwere obtained after incubated with 0–50 nM Taxol for 24 h (&) or48 h (m) is shown. (b) The percentage of TMRE-negative cellsobtained after adding different concentrations of Taxol (0–50 nM)with 0–20 mM LPA is shown. The results are shown as means±s.e.m. of three experiments.maximum decrease in the percentage of apoptotic cellsremaining on the plates in the presence of Taxol. Lesseffect was observed at 5 and 10 mM LPA (Figure 2c).However, 0.5–10 mM LPA increased the recovery ofviable cells from 49±4.3 to 68±6.1% (Figure 2d).The second method used flow cytometry to detectapoptotic cells by TMRE-negative staining. Optimumapoptosis in the recovered cells was detected at 20 nMTaxol and this increased from 24 to 48 h (Figure 3a).We chose to use 48 h incubations, which increasedthe population of apoptotic cells to about 55%. LPA(5–20 mM) had a significant protective effect againstdifferent concentrations of Taxol. Maximum protectionwas obtained with 5 mM LPA against 20 nM Taxol inwhich the percentage of apoptotic cells decreased(Po0.05) from 54±2 to33±3% (Figure 3b).Isobologram analysis was used to evaluate theinteractions between Taxol and LPA at the IC 50(inhibitory concentration 50%). All LPA concentrationsused produced an antagonistic effect on Taxol-inducedapoptosis with combination index values of >1(Table 1). This antagonism was particularly apparentat 10 and 20 mM LPA (Po0.01). For example, 50 nMTaxol was required when 20 mM LPA was present toproduce equivalent apoptosis to that obtained withOncogene

4Table 1Taxol(nM)Effect of LPA on Taxol-induced change in mitochondrialpotentialCombinationdoseLPA(mM)FractionalapoptosisLysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alTaxol dose(nM)(alone)Dose–responseindexCombinationindex5 1 0.31 4.6 0.9 22.210 5 0.32 5.0 0.5 23120 10 0.38 8.0 0.4 163050 20 0.36 7.2 0.2 1490The dose–response interactions between different concentrations ofTaxol and LPA in Figure 3b were assessed by isobologram analysis.7.2 nM Taxol in the absence of LPA. Thus, 20 mM LPAproduced a 6.9-fold decrease at IC 50 in Taxol sensitivity.ATX protects MCF-7 and MDA-MB-435 cells againstTaxol-induced apoptosis by producing LPAWe next studied whether ATX protects MCF-7 cellsagainst Taxol-induced apoptosis by changing the balanceof cell activation by LPC versus LPA. MCF-7 cells wereincubated with 10% charcoal-treated FBS and concentratedmedium obtained from 10 6 MDA-MB-435 cells, orMCF-7 cells in the presence or absence of LPC or LPA.Concentrated medium from MDA-MB-435 cells showedan average of 32 times more ATX activity compared withequivalent medium from MCF-7 cells.Concentrated medium from MDA-MB-435 cellssupplemented with 10 mM LPC decreased percentageapoptosis measured by DAPI staining from 7.4±1.3% to4.4±0.6% (88% of the effect of 0.5 mM LPA) (Figure 4a).The percentage of viable cells recovered also increased(49±3% to 77±7%; Po0.05) (Figure 4b). Concentratedmedium from MDA-MB-435 cells appeared todecrease apoptosis (6.1±0.8%) in the absence of addedLPC or LPA, but this was not statistically significant.LPC in the absence of conditioned medium had noprotective effect (Figure 4a).We also used an ATX inhibitor, VPC8a202 (Cui et al.,2007), to determine whether the effect of concentratedmedium depended upon ATX activity. VPC8a202 (1 mM)decreased ATX activity by 87% as expected (results notshown) and it abolished the protective effect of LPC plusconcentrated media from MDA-MB-435 cells (Figure 4a).There was no increase in the percentage of viable cellsafter applying the ATX inhibitor alone (Figure 4b).The TMRE assay also showed a significant decreasein apoptosis (Po0.01) in cells treated with concentratedmedium from MDA-MB-435 cells when 200 mM LPCwas added (Figure 4c). Concentrated medium alone orLPC alone had no significant protective effect(Figure 4c). VPC8a202 (1 mM) abolished the protectiveeffect of concentrated media with LPC. As a control, theeffect of VPC8a202 was shown to be specific, as it didnot alter the protective effect of LPA on Taxol-inducedapoptosis (Figures 4b and c).To confirm that the effects of concentrated mediumcells depended on ATX activity, we used recombinantATX, which protected against Taxol-induced apoptosisand increased the number of viable cells recoveredwhen LPC was present (Figures 4d and e). BlockingATX activity with 1 mM of the inhibitors VPC8a202, orS32826 (Ferry et al., 2008), or heating at 70 1C abolishedthe protective effect of recombinant ATX. Furthermore,concentrated medium from MCF-7 cells that containedno significant ATX activity failed to provide anysignificant protection with LPC against Taxol-inducedapoptosis (Figures 5a–c).We also showed that the endogenous ATX producedby MDA-MB-435 cells protects these cells againstTaxol-induced apoptosis when LPC is present, as thiseffect is blocked by the ATX inhibitors, VPC8a202 orS32826 (Figure 6). The inhibitors had no effect of theprotection afforded by LPA.LPA protects against Taxol-induced apoptosis byincreasing PI3K/Akt and decreasing ceramide productionWe examined the roles of the PI3K-Akt and MAPK/ERK kinase-ERK pathways on apoptosis by usingLY294002 and PD98059, respectively. LY294002(50 mM) almost completely abrogated the effects ofLPA in protecting against apoptosis (Figures 7aand b). PD98059 (40 mM) showed no significant effecton the LPA protection against Taxol-induced apoptosis,although it blocked the activation of p42/44 mitogenactivatedprotein kinase (MAPK) (Figures 8a–c). Also,phospho-42/44p MAPK and total 42/44p MAPKshowed a significant increase over 12 h in response toLPA. There was no significant effect of Taxol alone onthe expression of phospho-42/44p MAPK and total42/44p MAPK.Lysophosphatidate (5 mM) increased (Po0.05) theexpression of both phospho-Akt and total Akt by abouttwofold over 12 h (Figure 8a) and the ratio of phospho-Akt/totalAkt (Figure 8d). Taxol (30 nM) alonehad no significant effect. When Taxol was added withLPA (5 mM), there was still a 50% increase in phospho-Akt and total Akt compared with nontreated cells(Figure 8a) and an increase in the phospho-Akt/totalAkt ratio (Figure 8d). Applying 50 mM LY294002abolished the effect of LPA on the elevation of activatedand total Akt.To determine the protective mechanism of LPA againstTaxol-induced apoptosis, we measured ceramide mass.There was a 2.9-fold increase in ceramide accumulationin Taxol-treated cells (Po0.05) after 24 h of incubation.LPA (0.5 mM) decreased (Po0.05) the Taxol-inducedincrease in ceramides to 1.9-fold. LY294002 (50 mM)blocked the effect of LPA in decreasing ceramideaccumulation by about 80% (Figure 9).Interactions of LPA versus Taxol on the cell cycleTaxol-induced apoptosis is accompanied by increasedaccumulation of cells in the G2/M phase of the cell cycle(McGrogan et al., 2008). Under our conditions, thisrepresented an increase from about 16 to 43% (Po0.05)(Figures 10a and c), which is compatible with previousstudies (Spankuch et al., 2007; Demidenko et al., 2008).Oncogene

Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et al5Figure 4 Autotaxin protects against Taxol-induced apoptosis in MCF-7 cells. Combinations of concentrated medium from MDA-MB-435 cells (medium/435) or recombinant autotoxin (ATX) (50 ng/ml), with lysophosphatidate (0.5 mM for 4,6-diamidino-2-phenylindole (DAPI) staining and 5 mM for tetramethlyrhodamine ethyl ester (TMRE) assay), lysophosphatidylcholine (200 mM) andthe ATX inhibitors, VPC8a202 (1 mM) and S32826 (1 mM), were added with 50 nM (a, b, d, e)or30nM Taxol (c). (a, b) The percentage offragmented nuclei and the relative number of viable cells on the dishes as determined by DAPI staining after a 24 h incubation isshown. (c) The percentage of TMRE-negative cells after a 48 h incubation is shown. (d, e) The percentage of fragmented nuclei and therelative number of viable cells on the dishes as determined by DAPI staining after a 24 h incubation with 50 ng/ml recombinant ATX(rATX), rATX heated for 10 min at 70 1C to inhibit ATX activity >96% and the ATX inhibitors, VPC8a202 (1 mM) and S32826(1 mM), is shown. The results are shown as means±s.e.m. of three experiments. *Po0.05.Adding 5 mM LPA decreased the cells in G2/M to 28%and this represented reversal of the Taxol-induced effectof about 44% (Po0.05) (Figures 10c and d). AddingLY294002 blocked most of the LPA effect in reducing(Po0.05) the cells trapped in G2/M in the presence ofTaxol (Figures 10d and h). As controls, we showed that50 nM LY294002 or 5 mM LPA alone or in combinationhad no significant effect on the cells in G2/M phase(Figures 10b, e and f). These results demonstrated thedominant role of the PI3K pathway in LPA protectionagainst Taxol-induced apoptosis.DiscussionEffective chemotherapy depends on successful inductionof apoptosis and defects in apoptotic signaling are amajor cause of drug resistance (Melet et al., 2008).Chemotherapeutic resistance results in poor responsesand reduced overall survival in patients with metastaticbreast cancer (McGrogan et al., 2008).We used doses of Taxol, LPA and LPC up to 50 nM,20 mM and 200 mM, respectively, with MCF-7 cells. TheseTaxol concentrations are used for the treatment ofbreast cancer patients (Marchetti et al., 2002), and theLPA and LPC concentrations are compatible with thosein the circulation and tumor microenvironment (Moolenaaret al., 2004; Yue et al., 2004). We showed thatTaxol-induced killing is strongly antagonized by LPAon the basis of measurement of nuclear fragmentation,loss of mitochondrial potential and the recovery ofviable cells. LPA (0.5 mM) partially blocked the Taxolinducedincrease in the percentage of apoptotic nuclei,but we needed 5 mM LPA to antagonize the effect ofOncogene

6Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alFigure 6 Autotaxin protects against Taxol-induced apoptosis inMDA-MB-435 cells. Lysophosphatidate (0.5 mM) or lysophosphatidylcholine(200 mM) and autotaxin inhibitors, VPC8a202 (1 mM)and S32826 (1 mM), were added with 50 nM Taxol. (a, b) Thepercentage of fragmented nuclei and the relative number of viablecells on the dishes as determined by 4,6-diamidino-2-phenylindolestaining after a 24 h incubation, respectively, is shown. Results areshown as means±s.e.m. of three experiments. *Po0.05.Figure 5 The effect of concentrated medium from MCF-7 cells onTaxol-induced apoptosis. Combinations of concentrated mediumfrom MCF-7 cells (medium/MCF-7), together with lysophosphatidate(LPA) (0.5 mM), or lysophosphatidylcholine (200 mM) wereadded with 50 nM Taxol (a, b) or 30nM Taxol (c). (a, b) Thepercentage of apoptotic cells and the relative number of viable cellson the dishes as determined by 4,6-diamidino-2-phenylindolestaining after a 24 h incubation is shown. (c) The percentageof tetramethlyrhodamine ethyl ester-negative cells after a 48 hincubation where 5 mM LPA was used. The results are shown asmeans±s.e.m. of three experiments. *Po0.05.Taxol in increasing the proportion TMRE-negativecells. The discrepancies suggest that cells that wereprotected at low LPA concentrations, where no differencesin TMRE staining were observed, probably diedby a mitochondria-independent and nonapoptotic(necrotic) mechanism. However, another major factoris that the DAPI and TMRE measurements determineonly what happened to cells remaining on the dishes.We, therefore, do not have information concerningthe mechanisms of cell death of more than 50% of cellsthat were not recovered. Consequently, it is significantthat 0.5–10 mM LPA increased the number of cells thatwere recovered after Taxol treatment and that were thenreflected in the percentage of apoptotic cells recordedin the DAPI and TMRE staining assays.Accumulating evidence shows that LPA promotessurvival of ovarian cancer cells (Frankel and Mills,1996; Goetzl et al., 1999a), renal proximal tubular cells(Levine et al., 1997), lymphoblastoma T cells (Goetzlet al., 1999b), macrophages (Koh et al., 1998), neonatalSchwann cells (Weiner and Chun, 1999), fibroblast(Fang et al., 2000), mesangial cells (Inoue et al., 2001)and colon cancer cells (Rusovici et al., 2007) bypreventing apoptosis. Different cell types expressdifferent LPA receptors, which have specific roles, andhence the effects of LPA signaling vary from cell to cell.In contrast to the survival role of LPA 1 in Schwann cells(Weiner and Chun, 1999; Contos et al., 2000), overexpressionof LPA 1 in ovarian cancer cells inducedapoptosis and anoikis (Furui et al., 1999), whereas LPA 2appeared to mediate LPA-stimulated survival of ovariancancer cells (Goetzl et al., 1999a). In T lymphoblasts,both LPA 1 and LPA 2 promoted LPA-induced survival(Goetzl et al., 1999a). LPA also prevents apoptosis incolon cancer cells through LPA 2 receptor (Rusoviciet al., 2007).LPA 2 is the predominant LPA receptor in MCF-7cells (Chen et al., 2007). This is compatible with ourfinding that 1 mM VPC51299 (an LPA 1/3 receptorantagonist) failed to block the LPA protection againstOncogene

Figure 7 Phosphatidylinositol 3-kinase, but not extracellularsignal-regulated kinase activation, is required for lysophosphatidate(LPA)-mediated protection against Taxol-induced apoptosis.(a) The effects of 50 mM LY294002 and 40 mM PD9805 on thepercentage of apoptotic nuclei in MCF-7 cells treated for 24 h inthe presence or absence of 50 nM Taxol and 0.5 mM LPA are shown.(b) The effects of 50 mM LY294002 and 40 mM PD9805 onpercentage of tetramethlyrhodamine ethyl ester-negative cellsafter treatment in the presence or absence of 30 nM of Taxol with5 mM LPA are shown. Results are shown as means±s.e.m. of threeexperiments. *Po0.05.Taxol-induced apoptosis in MCF-7 cells (results notshown). Unfortunately, no effective specific agonist forLPA 2 receptors has been developed.Activation of PI3K-Akt protects primary lymphocyticleukemia cells, ovarian cancer cells and intestinalepithelial cells from apoptosis, whereas G i -mediatedactivation of ERK is necessary for the survival offibroblasts (Fang et al., 2000; Gauthier et al., 2001;Baudhuin et al., 2002; Dufour et al., 2004; Harnoiset al., 2004; Hu et al., 2005). In this study, inhibition ofPI3K blocked the protective effect of LPA on Taxolinducedapoptosis, whereas blocking ERK activationwith PD98059 was without any effect.Lysophosphatidylcholine is present in plasma andserum at >100 mM in a predominantly albumin-boundform (Croset et al., 2000) and it is also detected in mediafrom tumor cells (Umezu-Goto et al., 2002). However,in our work, LPC alone did not antagonize Taxolinducedapoptosis unless concentrated medium fromMDA-MB-435 cells or recombinant ATX was added.This protective effect was blocked by VPC8a202 andS32826, two ATX inhibitors (Cui et al., 2007; Ferryet al., 2008), or by heat inactivation of ATX activity.The MDA-MB-435 medium, compared with thatproduced by MCF-7 cells, contained abundant ATX.Our work establishes that LPC is inactive in protectingagainst Taxol-induced apoptosis unless it is converted toLPA by ATX. We also showed in our unpublished workLysophosphatidate inhibits Taxol-induced apoptosisN Samadi et althat LPC only stimulates the migration of MDA-MB-231 breast cancer cells in the presence of active ATX.These conclusions are compatible with the view that theputative LPC receptors, G2A and GPR4, are protonsensingreceptors whose activity is negatively regulatedby LPC (Tomura et al., 2005). Many reported biologicaleffects of LPC are probably mediated through ATX andLPA production.In vivo, ATX might also hydrolyze sphingosylphosphorylcholineto sphingosine 1-phosphate (Clair et al.,2003). LPA and sphingosine 1-phosphate show overlappingbiological activities on cell survival by actingon distinct receptor families. However, the physiologicalsignificance of ATX in producing sphingosine1-phosphate is not established.Binding of Taxol to microtubules aberrantly engagesthe mitotic checkpoint, effecting sustained mitotic arrest(McGrogan et al., 2008). Ceramide levels regulatesustained mitotic arrest, apoptosis and mitotic slippage(Asakuma, 2003), and elevated ceramide concentrationsare a known component of Taxol-mediated cell death(Charles et al., 2001; Sietsma et al., 2002; Swanton et al.,2007). Sustained ceramide elevation in the endoplasmicreticulum coordinately activates the stress responses andinactivates anti-apoptotic Akt, thus leading to apoptosis(Swanton et al., 2007). We showed that LPA attenuatedthe Taxol-induced increase in ceramides and subsequentapoptosis through PI3K/Akt activation.In conclusion, our work demonstrates that theabundant extracellular lipid, LPC, protects MCF-7breast cancer cells and MDA-MD-435 melanoma cellsagainst Taxol-induced apoptosis, but only after itsconversion to LPA by ATX. LPA itself showed strongantagonism against the apoptotic effects of Taxol. Thiswork identifies that inhibition of ATX activity couldimprove the treatment of cancers by Taxol and possiblyother chemotherapeutic agents in patients who expresshigh ATX activity. It is, therefore, hoped that biologicallystable ATX inhibitors will be developed to test thishypothesis in vivo.Materials and methodsMaterialsRPMI1640 medium, penicillin and streptomycin were obtainedfrom GIBCO (Carlsbad, CA, USA). Trypsin-ethylenediaminetetraaceticacid, TMRE, goat anti-mouse IgG, ATX primersand superscript II reverse transcriptase were obtained fromInvitrogen (Carlsbad, CA, USA). DAG kinase, PD98059 anddithiothreitol were obtained from EMD Chemicals Inc.(Gibbstown, NJ, USA). Fatty acid-free bovine serum albumin,paclitaxel (Taxol), activated charcoal (Norit), Hoechst33258,ATP, diethylenetriaminepentaacetic acid (DETAPEC), pertussistoxin, oleoyl-L-a-lysophosphatidic acid, sodium salt (LPA),oleoyl-L-a-LPC, ceramide and monoclonal anti-GAPDH werepurchased from Sigma-Aldrich (Oakville, ON, Canada).VPC51299 and VPC8a202 were provided by Drs KR Lynchand TL Macdonald (University of Virginia, Charlottesville,VA, USA). LY294002 was obtained from Biomol Int.(Plymouth Meeting, PA, USA). S32826 was provided byDr JA Boutin (Institut de Recherches Servier). [ 32 P]LPA was7Oncogene

8Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alPhosphoAktTotal AktP-44p MAPKP-42p MAPKTotal 44p MAPKTotal 42p MAPKGAPDHP-p42 MAPK/Totalp42 MAPK0.100.050.00***Figure 9 The role of Taxol and lysophosphatidate (LPA) incontrolling ceramide concentrations. Ceramide mass was determinedin MCF-7 cells after treating with 50 nM Taxol in thepresence or absence of 0.5 mM of LPA or 50 mM LY294002 for 24 h.The mass of ceramide (pmol) was normalized to the concentrationof phospholipid phosphate (nM) in each individual lipid extract.The results are expressed as means±s.e.m. of eight individualdishes for each treatment. *Po0.05.P-p44 MAPK/Totalp44 MAPKP-Akt/Total Akt0.60.40.20.00.40.30.20.1*** * *0.0Taxol - + - + - + - + - + - +- - + + - - + + - - + +LPALY294002PD98059- - - - - - - - + + + +- - - - + + + + - - - -Figure 8 The effects of lysophosphatidate (LPA) on phosphatidylinositol3-kinase (PI3K)/AKT and mitogen-activated proteinkinase (MAPK) pathways. MCF-7 cells were incubated for 12 hwith Taxol (30 nM) and LPA (5 mM) in the presence of 50 mMLY294002 (PI3K inhibitor) and 40 mM PD98059 (MAPK/extracellularsignal-regulated kinase kinase inhibitor). Results in (a)show the western blot analyses for phospho- and total Akt,phospho- and total ERK (p42 and p44 MAPK) and glyceraldehydephosphate dehydrogenase (GAPDH), which was used as a loadingcontrol. The results are representative of three independentexperiments and they were quantified by densitometric analysisusing a Licor Imaging System. (b) The densitometric ratio ofphospho-p42MAPK/total p42MAPK is shown; (c) phosphop44MAPK/totalp44MAPK ratio is shown and in (d), phospho-Akt/total Akt is shown. The results are shown as means±s.e.m. ofthree independent experiments. *Indicates the difference (Po0.05)from the nontreated cells.synthesized as before (Jasinska et al., 1999). Rabbit polyclonalantibody against ATX and recombinant ATX were gifts fromDr T Clair, through the Intramural Research Program ofNIH, National Cancer Institute, Center for Cancer Research**(Bethesda, MD, USA). Goat anti-rabbit IgG was obtainedfrom Rockland (Gilbertsville, PA, USA). Polyclonal anti-Akt,phospho-Akt (S473) and monoclonal phospho-p44/42 MAPkinase were obtained from Cell Signaling Technology (Danvers,MA, USA). Polyclonal anti-MAP kinase (sc-93) wasfrom Santa Cruz Biotechnology (Santa Cruz, CA, USA). FS-3compound for the ATX assay was from Echelon Company(San Jose, CA, USA). FBS was from Medicorp Inc. (Montre´ al,PQ, Canada). RNAqueous kit, DNA-free kit and Syber Greenbuffer mix were obtained from Ambion (Austin, TX, USA).Cell cultureThe MCF-7 and MDA-MB-435 cells were obtained from theAmerican Type Culture Collection. Cells were maintainedin RPMI1640 medium with 10% FBS and 1% penicillin/streptomycin in a humidified atmosphere with 5% CO 2at 37 1C.Quantification of apoptotic nuclei using DAPI stainMCF-7 cells (2 10 4 cells/well) were grown in 96-well plates.Cells (about 70% confluent) were washed with phosphatebufferedsaline and treated in serum-free media for 24 h. Cellswere then fixed in 4% paraformaldehyde for 20 min andpermeabilized in 0.1% Triton X-100 for 15 min. Fixed cellswere stained with 1 mg/ml Hoechst33258 for 15 min afterevaporation of the fixing solution. Nuclear morphology of thecells was observed by fluorescence microscopy. Triplicatesamples were prepared for each treatment and at least 300 cellswere counted in random fields for each sample and apoptoticnuclei were identified (Sakashita et al., 2007).Measurement of LPA breakdownSix-well plates were seeded with 2 10 5 cells and these weregrown to 70% confluence. The half-life of LPA wasdetermined after adding 2 ml of medium containing 32 P-labeledLPA (100,000 d.p.m.) with 1 or 20 mM nonradioactive LPA andmeasuring 32 P i - and 32 P-labeled LPA during 24 h (Pilquil et al.,2006). The half-life of LPA was about 10 h and this wasindependent of the LPA concentration (results not shown).Mitochondrial membrane potential (TMRE) assayBreakdown of the mitochondrial membrane potential (DCm)was determined using TMRE staining. MCF-7 cells (2 10 5 )Oncogene

Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alCounts1801501209060300M1Non-treated cells73%11%CountsM216% 40M113%M300 200 400 600 800 10000 200 400 600 800 1000FL2-AFL2-A20016012080M2LPA (5 µM)M375%12%Sub G1/G1SG2/M9Taxol (30 nM)Taxol (30 nM) + LPA (5 µM)Counts1801501209060300M1M2M352%5%43%0 200 400 600 800 1000FL2-ACounts1200M1M2M360%12%28%0 200 400 600 800 1000FL2-ALY-294002 (50 µM)LY-294002 (50 µM) + LPA (5 µM)15012069%15012072%Counts906030M1M2M311%23%Counts906030M1M28%20%00 200 400 600 800 1000FL2-A0M30 200 400 600 800 1000FL2-ATaxol (30 nM) + LY-294002 (50 µM)Taxol (30 nM)+ LY-294002 (50 µM)+ LPA (5 µM)Counts1501209060300M1M2M350%9%41%0 200 400 600 800 1000FL2-ACounts1501209060300M1M2M353%8%39%0 200 400 600 800 1000FL2-AFigure 10 Effects of lysophosphatidate (LPA) and Taxol on cell cycle progression of MCF-7 breast cancer cells. MCF-7 cells wereincubated for 48 h with 30 nM Taxol and 5 mM LPA. Then the cells were collected and stained with propidium iodide (PI) beforeanalysis by fluorescence-activated cell sorting. DNA histograms were measured using CellQuest software and the percentage of G0/G1,S and G2/M cells were calculated. Results are representative of two independent experiments. The action of LPA in partially reversingthe Taxol-induced G2/M arrest was also confirmed in a single experiment with a 24 h incubation.were grown to confluence in six-well plates over 48 h inRPMI1640 medium with 10% FBS. The medium was changedto RPMI1640 medium with 10% of charcoal-treated FBS.This removed >95% of the LPA as assessed after adding 32 P-labeled LPA to the FBS (results not shown). Cells wereincubated for an additional 48 h with LPA-depleted FBS withTaxol, or LPA as indicated. LPA concentrations were maintainedby adding half of the original concentration of LPA every10 h on the basis of the calculated half-life of LPA. Cells werethen tripsinized, collected in 96 V-bottomed plate and exposed toTMRE (500 nM) for30minat371C. Excess dye was removedwith phosphate-buffered saline containing 0.1% fatty acid-freeBSA and cell-associated fluorescence was measured with aCalibur Flow Cytometer (Becton Dickinson, San Jose´ , CA,USA) using Cell Quest software (Ramjaun et al., 2007).ATX expressionmRNA expression Total RNA was extracted using theRNAqueous kit, according the manufacture’s instruction.DNA-free kit was also applied to remove contaminatingDNA from RNA preparation. Total RNA was treated withOncogene

10Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alsuperscript II reverse transcriptase. Real-time RT-PCR wasperformed with 25 ml of master mix containing 2 SyberGreen buffer mix and forward and reverse primers (Invitrogen).The internal control was the constitutively expressedhousekeeping human glyceraldehyde 3-phosphate dehydrogenase(GAPDH). Primers for human ATX were as follows:sense, 5 0 -ACAACGAGGAGAGCTGCAAT-3 0 ; antisense,5 0 -AGAAGTCCAGGCTGGTGAGA-3; and for humanGAPDH were as follows:sense, 5 0 -ACAGTCAGCCGCATCTTCTT-3 0 ; antisense, 5 0 -GACAAGCTTCCCGTTCTCAG-3 0 .Samples were assayed in triplicate on the 7500 Real Time PCRSystem (Applied Biosystems).Western blot analysis Serum-free medium from MD-MB-435cells was concentrated to 200 ml per 10 6 cells by centrifugationusing a Centricon Millipore’s Ultracel YM cellulose membrane(Millipore Corporation, Bedford, TX, USA) and analysedusing 10% SDS–polyacrylamide gel electrophoresis. Nitrocellulosemembranes were treated with Odyssey blocking bufferfor 24 h and incubated with 1:5000 dilution of ATX antibodyfor 60 min followed by four washes with phosphate-bufferedsaline þ 0.1% Tween-20. Blots were incubated in 1:10 000dilution of goat anti-rabbit IgG for 45 min. After washing fourtimes, ATX was visualized at 800 nm with the OdysseyInfrared Scanner.ATX activity assay ATX activity was assayed in 96-wellplates using 1 mM FS-3 as a substrate. Fluorescence was read at15 min intervals by a Synergy2 system (BioTek, Winooski, VT,USA) with excitation and emission wavelengths of 485 and538 nm, respectively.Ceramide mass assayCells (1 10 6 ) were plated in 10 cm culture dishes and grown toconfluence in RPMI1640 medium with 10% FBS. After 48 h,the medium was changed to RPMI1640 medium with 10% ofcharcoal-treated FBS, and cells were incubated for 48 h withthe indicated concentrations of Taxol, LPA and LY294002.After washing with phosphate-buffered saline, cell lysateswere collected in methanol and the lipids were extracted(Martin et al., 1997). Ceramide concentrations were measuredby conversion to ceramide 1-phosphate, using the diacylglycerolkinase and [g- 32 P]ATP and by comparison to standardsof ceramide. Results were normalized to the phospholipidconcentration of the extract (Martin et al., 1997).Statistical analysis and determination of the interaction of Taxolwith LPA on apoptosisThe interaction was evaluated by the isobologram technique(CalcuSyn software, Biosoft, Cambridge, UK). The experimentaliso-effect points were the concentrations (relative toIC 50 concentrations) of the two agents that when combinedkill 50% of the cells. A quantitative index of these interactionswas provided by the isobologram equation combinationindex ¼ (a/A) þ (b/B), where A and B represent the Taxoland LPA concentrations that alone produce a fixed levelof inhibition (IC 50 ), a and b represent the concentrationsrequired for the same effect when they were combined. CIvalues of o1 indicate synergy; 1 represents addition and 41indicate antagonism.Statistical analysis was performed by analysis of variancewith a Kruskal–Wallis post hoc test.AbbreviationsATX, autotaxin; DAPI, 4,6-diamidino-2-phenylindole; FBS,fetal bovine serum; LPA, lysophosphatidate; LPC, lysophosphatidylcholine;MAPK, mitogen-activated protein kinase;PI3K, phosphatidylinositol 3-kinase; TMRE, tetramethlyrhodamineethyl ester.AcknowledgementsWe thank Mr M Czernick and Mr J Dewald for excellenttechnical support in the fluorescence-activated cell sortinganalysis and preparing labeled LPA, respectively. We thankDr M Sawyer for help with the isobologram analysis,Dr E Posse de Chaves for advice on DAPI staining andDrs F Bamforth and GS Cembrowski for their support. Wethank Dr T Clair for providing recombinant ATX and theATX antibody. Drs KR Lynch and TL Macdonald donatedthe VPC8a202 and VPC51299 and Dr JA Boutin gave usS32826. DNB is a Medical Scientist of the Alberta HeritageFoundation for Medical Research. ISG is a recipient of aRecruitment Award from the Alberta Cancer Board/AlbertaCancer Foundation. NS is a recipient of scholarships fromIranian Ministry of Health and the Bell McLeod EducationalFund, University of Alberta. The work was supported by theCanadian Institutes of Health Research (MOP 81137).ReferencesAoki J, Taira A, Takanezawa Y, Kishi Y, Hama K, Kishimoto T et al.(2002). Serum lysophosphatidic acid is produced through diversephospholipase pathways. J Biol Chem 277:48737–48744.Asakuma J, Sumitomo M, Asano T, Asano T, Hayakawa M (2003).Selective Akt inactivation and tumor necrosis actor-relatedapoptosis-inducing ligand sensitization of renal cancer cells by lowconcentrations of paclitaxel. Cancer Res 63:1365–1370.Baudhuin LM, Cristina KL, Lu J, Xu Y. (2002). Akt activationinduced by lysophosphatidic acid and sphingosine-1-phosphaterequires both mitogen-activated protein kinase kinase and p38mitogen-activated protein kinase and is cell-line specific.Mol Pharmacol 62:660–671.Bergstralh DT, Ting JP. (2006). Microtubule stabilizing agents:theirmolecular signaling consequences and the potential for enhancementby drug combination. Cancer Treat Rev 32:166–179.Brindley DN. (1993). Hepatic secretion of lysphosphatidylcholine:anovel transport system for polyunsaturated fatty acids and choline.J Nutr Biochem 4:442–449.Brindley DN. (2004). Lipid phosphate phosphatases and relatedproteins:signaling functions in development, cell division, andcancer. J Cell Biochem 92:900–912.Chappell J, Leitner JW, Solomon S, Golovchenko I, Goalstone ML,Draznin B. (2001). Effect of insulin on cell cycle progression inMCF-7 breast cancer cells. Direct and potentiating influence. J BiolChem 276:38023–38028.Charles AG, Han TY, Liu YY, Hansen N, Giuliano AE, Cabot MC.(2001). Taxol-induced ceramide generation and apoptosis in humanbreast cancer cells. Cancer Chemother Pharmacol 47:444–450.Chen M, Towers LN, O’Connor KL. (2007). LPA2 (EDG4) mediatesRho-dependent chemotaxis with lower efficacy than LPA1 (EDG2)in breast carcinoma cells. Am J Physiol Cell Physiol 292:C1927–C1933.Clair T, Aoki J, Koh E, Bandle RW, Nam SW, Ptaszynska MMet al. (2003). Autotaxin hydrolyzes sphingosylphosphorylcholine toproduce the regulator of migration, sphingosine-1-phosphate.Cancer Res 63:5446–5453.Oncogene

Contos JJ, Fukushima N, Weiner JA, Kaushal D, Chun J. (2000).Requirement for the lpA1 lysophosphatidic acid receptor genein normal suckling behavior. Proc Natl Acad Sci USA 97:13384–13389.Croset M, Brossard N, Polette A, Lagarde M. (2000). Characterizationof plasma unsaturated lysophosphatidylcholines in human and rat.Biochem J 345(Part 1):61–67.Cui P, Tomsig JL, McCalmont WF, Lee S, Becker CJ, Lynch KR et al.(2007). Synthesis and biological evaluation of phosphonate derivativesas autotaxin (ATX) inhibitors. Bioorg Med Chem Lett 17:1634–1640.Demidenko ZN, Kalurupalle S, Hanko C, Lim CU, Broude E,Blagosklonny MV. (2008). Mechanism of G1-like arrest by lowconcentrations of paclitaxel:next cell cycle p53-dependent arrestwith sub G1 DNA content mediated by prolonged mitosis.Oncogene 27:4402–4410.Dennis J, Nogaroli L, Fuss B. (2005). Phosphodiesterase-Ialpha/autotaxin (PD-Ialpha/ATX):a multifunctional protein involved incentral nervous system development and disease. J Neurosci Res 82:737–742.Dufour G, Demers MJ, Gagne D, Dydensborg AB, Teller IC,Bouchard V et al. (2004). Human intestinal epithelial cell survivaland anoikis. Differentiation state-distinct regulation and roles ofprotein kinase B/Akt isoforms. J Biol Chem 279:44113–44122.Fang X, Schummer M, Mao M, Yu S, Tabassam FH, Swaby R et al.(2002). Lysophosphatidic acid is a bioactive mediator in ovariancancer. Biochim Biophys Acta 1582:257–264.Fang X, Yu S, LaPushin R, Lu Y, Furui T, Penn LZ et al. (2000).Lysophosphatidic acid prevents apoptosis in fibroblasts via G(i)-protein-mediated activation of mitogen-activated protein kinase.Biochem J 352(Part 1):135–143.Ferry G, Moulharat N, Pradere JP, Desos P, Try A, Genton A et al.(2008). S32826:a nanomolar inhibitor of autotaxin. Discovery,synthesis and applications as a pharmacological tool. J PharmacolExp Ther 327:809–819.Fox MA, Alexander JK, Afshari FS, Colello RJ, Fuss B. (2004).Phosphodiesterase-I alpha/autotaxin controls cytoskeletal organizationand FAK phosphorylation during myelination. Mol CellNeurosci 27:140–150.Frankel A, Mills GB. (1996). Peptide and lipid growth factors decreasecis-diamminedichloroplatinum-induced cell death in human ovariancancer cells. Clin Cancer Res 2:1307–1313.Furui T, LaPushin R, Mao M, Khan H, Watt SR, Watt MA et al.(1999). Overexpression of edg-2/vzg-1 induces apoptosis and anoikisin ovarian cancer cells in a lysophosphatidic acid- independentmanner. Clin Cancer Res 5:4308–4318.Gauthier R, Harnois C, Drolet JF, Reed JC, Vezina A, Vachon PH.(2001). Human intestinal epithelial cell survival:differentiationstate-specific control mechanisms. Am J Physiol Cell Physiol 280:C1540–C1554.Goetzl EJ, Dolezalova H, Kong Y, Hu YL, Jaffe RB, Kalli KRet al. (1999a). Distinctive expression and functions of the type 4endothelial differentiation gene-encoded G protein-coupled receptorfor lysophosphatidic acid in ovarian cancer. Cancer Res 59:5370–5375.Goetzl EJ, Kong Y, Mei B. (1999b). Lysophosphatidic acid andsphingosine 1-phosphate protection of T cells from apoptosis inassociation with suppression of Bax. J Immunol 162:2049–2056.Hama K, Aoki J, Fukaya M, Kishi Y, Sakai T, Suzuki R et al. (2004).Lysophosphatidic acid and autotaxin stimulate cell motility ofneoplastic and non-neoplastic cells through LPA1. J Biol Chem 279:17634–17639.Harnois C, Demers MJ, Bouchard V, Vallee K, Gagne D, Fujita Net al. (2004). Human intestinal epithelial crypt cell survival anddeath:complex modulations of Bcl-2 homologs by Fak, PI3-K/Akt-1, MEK/Erk, and p38 signaling pathways. J Cell Physiol 198:209–222.Hu X, Haney N, Kropp D, Kabore AF, Johnston JB, Gibson SB.(2005). Lysophosphatidic acid (LPA) protects primary chroniclymphocytic leukemia cells from apoptosis through LPA receptorLysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alactivation of the anti-apoptotic protein AKT/PKB. J Biol Chem280:9498–9508.Inoue CN, Nagano I, Ichinohasama R, Asato N, Kondo Y, Iinuma K.(2001). Bimodal effects of platelet-derived growth factoron rat mesangial cell proliferation and death, and the role oflysophosphatidic acid in cell survival. Clin Sci (Lond) 101:11–19.Jasinska R, Zhang QX, Pilquil C, Singh I, Xu J, Dewald J et al. (1999).Lipid phosphate phosphohydrolase-1 degrades exogenous glycerolipidand sphingolipid phosphate esters. Biochem J 340(Part 3):677–686.Jemal A, Siegel R, Ward E, Murray T, Xu J, Smigal C et al. (2006).Cancer statistics, 2006. CA Cancer J Clin 56:106–130.Kabarowski JH, Zhu K, Le LQ, Witte ON, Xu Y. (2001). Lysophosphatidylcholineas a ligand for the immunoregulatory receptorG2A. Science 293:702–705.Kim KS, Ren J, Jiang Y, Ebrahem Q, Tipps R, Cristina K et al.(2005). GPR4 plays a critical role in endothelial cell function andmediates the effects of sphingosylphosphorylcholine. FASEB J 19:819–821.Koh JS, Lieberthal W, Heydrick S, Levine JS. (1998). Lysophosphatidicacid is a major serum noncytokine survival factor formurine macrophages which acts via the phosphatidylinositol3-kinase signaling pathway. J Clin Invest 102:716–727.Krajewski S, Krajewska M, Turner BC, Pratt C, Howard B, ZapataJM et al. (1999). Prognostic significance of apoptosis regulators inbreast cancer. Endocr Relat Cancer 6:29–40.Lee CW, Rivera R, Gardell S, Dubin AE, Chun J. (2006). GPR92 as anew G12/13- and Gq-coupled lysophosphatidic acid receptor thatincreases cAMP, LPA5. J Biol Chem 281:23589–23597.Levine JS, Koh JS, Triaca V, Lieberthal W. (1997). Lysophosphatidicacid:a novel growth and survival factor for renal proximal tubularcells. Am J Physiol 273:F575–F585.Lin P, Ye RD. (2003). The lysophospholipid receptor G2A activates aspecific combination of G proteins and promotes apoptosis. J BiolChem 278:14379–14386.Marchetti P, Urien S, Cappellini GA, Ronzino G, Ficorella C. (2002).Weekly administration of paclitaxel:theoretical and clinical basis.Crit Rev Oncol Hematol 44(Suppl):S3–S13.Martin A, Duffy PA, Liossis C, Gomez-Munoz A, O’Brien L, StoneJC et al. (1997). Increased concentrations of phosphatidate,diacylglycerol and ceramide in ras- and tyrosine kinase (fps)-transformed fibroblasts. Oncogene 14:1571–1580.McGrogan BT, Gilmartin B, Carney DN, McCann A. (2008).Taxanes, microtubules and chemoresistant breast cancer. BiochimBiophys Acta 1785:96–132.Melet A, Song K, Bucur O, Jagani Z, Grassian AR, Khosravi-Far R.(2008). Apoptotic pathways in tumor progression and therapy. AdvExp Med Biol 615:47–79.Mills GB, Moolenaar WH. (2003). The emerging role of lysophosphatidicacid in cancer. Nat Rev Cancer 3:582–591.Moolenaar WH, van Meeteren LA, Giepmans BN. (2004). Theins and outs of lysophosphatidic acid signaling. Bioessays 26:870–881.Murph MM, Hurst-Kennedy J, Newton V, Brindley DN,Radhakrishna H. (2007). Lysophosphatidic acid decreases thenuclear localization and cellular abundance of the p53 tumorsuppressor in A549 lung carcinoma cells. Mol Cancer Res 5:1201–1211.Nam SW, Clair T, Campo CK, Lee HY, Liotta LA, Stracke ML.(2000). Autotaxin (ATX), a potent tumor motogen, augmentsinvasive and metastatic potential of ras-transformed cells. Oncogene19:241–247.Nam SW, Clair T, Kim YS, McMarlin A, Schiffmann E, Liotta LAet al. (2001). Autotaxin (NPP-2), a metastasis-enhancing motogen, isan angiogenic factor. Cancer Res 61:6938–6944.Pasternack SM, von Kugelgen I, Aboud KA, Lee YA, Ruschendorf F,Voss K et al. (2008). G protein-coupled receptor P2Y5 and its ligandLPA are involved in maintenance of human hair growth. Nat Genet40:329–334.11Oncogene

12Lysophosphatidate inhibits Taxol-induced apoptosisN Samadi et alPilquil C, Dewald J, Cherney A, Gorshkova I, Tigyi G, English Det al. (2006). Lipid phosphate phosphatase-1 regulates lysophosphatidate-inducedfibroblast migration by controlling phospholipaseD2-dependent phosphatidate generation. J Biol Chem281:38418–38429.Radeff-Huang J, Seasholtz TM, Matteo RG, Brown JH. (2004). Gprotein mediated signaling pathways in lysophospholipid inducedcell proliferation and survival. J Cell Biochem 92:949–966.Radu CG, Yang LV, Riedinger M, Au M, Witte ON. (2004). T cellchemotaxis to lysophosphatidylcholine through the G2A receptor.Proc Natl Acad Sci USA 101:245–250.Ramjaun AR, Tomlinson S, Eddaoudi A, Downward J. (2007).Upregulation of two BH3-only proteins, Bmf and Bim, during TGFbeta-induced apoptosis. Oncogene 26:970–981.Rikitake Y, Hirata K, Yamashita T, Iwai K, Kobayashi S, Itoh Het al. (2002). Expression of G2A, a receptor for lysophosphatidylcholine,by macrophages in murine, rabbit, and human atheroscleroticplaques. Arterioscler Thromb Vasc Biol 22:2049–2053.Rusovici R, Ghaleb A, Shim H, Yang VW, Yun CC. (2007).Lysophosphatidic acid prevents apoptosis of Caco-2 colon cancercells via activation of mitogen-activated protein kinase andphosphorylation of Bad. Biochim Biophys Acta 1770:1194–1203.Sakashita F, Osada S, Takemura M, Imai H, Tomita H, Nonaka Ket al. (2007). The effect of p53 gene expression on the inhibition of cellproliferation by paclitaxel. Cancer Chemother Pharmacol 62:379–385.Saunders DE, Lawrence WD, Christensen C, Wappler NL, Ruan H,Deppe G. (1997). Paclitaxel-induced apoptosis in MCF-7 breastcancercells. Int J Cancer 70:214–220.Sietsma H, Dijkhuis AJ, Kamps W, Kok JW. (2002). Sphingolipids inneuroblastoma:their role in drug resistance mechanisms. NeurochemRes 27:665–674.Snyder JP, Nettles JH, Cornett B, Downing KH, Nogales E. (2001).The binding conformation of Taxol in beta-tubulin:a model basedon electron crystallographic density. Proc Natl Acad Sci USA 98:5312–5316.So J, Wang FQ, Navari J, Schreher J, Fishman DA. (2005). LPAinducedepithelial ovarian cancer (EOC) in vitro invasion andmigration are mediated by VEGF receptor-2 (VEGF- R2). GynecolOncol 97:870–878.Spankuch B, Kurunci-Csacsko E, Kaufmann M, Strebhardt K. (2007).Rational combinations of siRNAs targeting Plk1 with breast cancerdrugs. Oncogene 26:5793–5807.Stracke ML, Krutzsch HC, Unsworth EJ, Arestad A, Cioce V,Schiffmann E et al. (1992). Identification, purification, and partialsequence analysis of autotaxin, a novel motility-stimulating protein.J Biol Chem 267:2524–2529.Sun B, Nishihira J, Suzuki M, Fukushima N, Ishibashi T, Kondo Met al. (2003). Induction of macrophage migration inhibitory factorby lysophosphatidic acid:relevance to tumor growth and angiogenesis.Int J Mol Med 12:633–641.Swanton C, Marani M, Pardo O, Warne PH, Kelly G, Sahai E et al.(2007). Regulators of mitotic arrest and ceramide metabolism aredeterminants of sensitivity to paclitaxel and other chemotherapeuticdrugs. Cancer Cell 11:498–512.Tigyi G, Parrill AL. (2003). Molecular mechanisms of lysophosphatidicacid action. Prog Lipid Res 42:498–526.Tomura H, Mogi C, Sato K, Okajima F. (2005). Proton-sensing andlysolipid-sensitive G- protein-coupled receptors:a novel type ofmulti-functional receptors. Cell Signal 17:1466–1476.Umezu-Goto M, Kishi Y, Taira A, Hama K, Dohmae N, Takio Ket al. (2002). Autotaxin has lysophospholipase D activity leading totumor cell growth and motility by lysophosphatidic acid production.J Cell Biol 158:227–233.Weiner JA, Chun J. (1999). Schwann cell survival mediated by thesignaling phospholipid lysophosphatidic acid. Proc Natl Acad SciUSA 96:5233–5238.Xu Y, Shen Z, Wiper DW, Wu M, Morton RE, Elson P et al. (1998).Lysophosphatidic acid as a potential biomarker for ovarian andother gynecologic cancers. Jama 280:719–723.Yang SY, Lee J, Park CG, Kim S, Hong S, Chung HC et al. (2002).Expression of autotaxin (NPP-2) is closely linked to invasiveness ofbreast cancer cells. Clin Exp Metastasis 19:603–608.Ye X, Ishii I, Kingsbury MA, Chun J. (2002). Lysophosphatidic acidas a novel cell survival/apoptotic factor. Biochim Biophys Acta 1585:108–113.Yue J, Yokoyama K, Balazs L, Baker DL, Smalley D, Pilquil C et al.(2004). Mice with transgenic overexpression of lipid phosphatephosphatase-1 display multiple organotypic deficits withoutalteration in circulating lysophosphatidate level. Cell Signal 16:385–399.Zhu K, Baudhuin LM, Hong G, Williams FS, Cristina KL,Kabarowski JH et al. (2001). Sphingosylphosphorylcholine andlysophosphatidylcholine are ligands for the G protein- coupledreceptor GPR4. J Biol Chem 276:41325–41335.Oncogene

Neuroscience Letters 443 (2008) 17–22Contents lists available at ScienceDirectNeuroscience Lettersjournal homepage: www.elsevier.com/locate/neuletTaxol induces oxidative neuronal cell death by enhancing the activity ofNADPH oxidase in mouse cortical culturesHong Jeon Jang a,b , Shinae Hwang b , Kyu Yong Cho a , Do Kyung Kim c , Kee-Oh Chay d , Jong-Keun Kim b,∗a Department of Neurosurgery, Kwangju Christian Hospital, Gwangju 503-715, Republic of Koreab Department of Pharmacology, Chonnam National University Medical School, 5 Hak-Dong, Gwangju 501-746, Republic of Koreac Department of Oral Physiology and The second stage of BK21, Chosun University College of Dentistry, Gwangju 501-759, Republic of Koread Department of Biochemistry, Chonnam National University Medical School, Gwangju 501-746, Republic of KoreaarticleinfoabstractArticle history:Received 23 May 2008Received in revised form 16 July 2008Accepted 18 July 2008Keywords:TaxolOxidative stressNADPH oxidaseNeuronal deathWe examined the involvement of oxidative stress in neuronal cell death induced by taxol, a microtubulestabilizinganti-cancer drug and investigated whether NADPH oxidase plays a role in taxol-inducedneuronal cell death in mouse cortical cultures. Cell death was assessed by measuring lactate dehydrogenasein the bathing media after 24-h exposure to taxol. Taxol (30–1000 nM) induced theconcentration-dependent neuronal death with apoptotic features. The neuronal death induced by taxolwas significantly attenuated not only by anti-apoptotic drugs such as z-VAD-fmk and cycloheximide butalso by antioxidants such as trolox, ascorbic acid and tempol. Vinblastine, a microtubule-depolymerizinganti-cancer drug, also induced neuronal death. The neuronal cell death induced by vinblastine was alsoattenuated by z-VAD-fmk, but not by antioxidants and NADPH oxidase inhibitors. Exposure the corticalcultures to taxol for 80 min formed neurite beadings visualized by fluorescence immunocytochemistry fortubulin. Treatment with either trolox or apocynin, an NADPH oxidase inhibitor, did not affect formationof the neurite beadings. RT-PCR and Western blot analysis revealed that exposure to taxol increased theexpression of p47 phox and gp91 phox and induced translocation of the p47 phox to the membrane in corticalcultures. Exposure to taxol markedly increased cellular 2,7-dichlorofluorescin diacetate fluorescence, anindicator for reactive oxygen species. Apocynin and trolox markedly inhibited the taxol-induced increaseof the fluorescence. Moreover, treatment with NADPH oxidase inhibitors or suppression of gp91 phox bysiRNA significantly attenuated the taxol-induced neuronal death. These results indicate that taxol inducesoxidative neuronal apoptosis by enhancing the activity of NADPH oxidase.© 2008 Elsevier Ireland Ltd. All rights reserved.Taxol has been widely used as an anti-cancer drug for ovarian,breast, lung and prostate cancer [4]. It inhibits the normal functionof microtubules, which causes the failure of mitosis and the death ofproliferating cells, and also affects other cellular functions such asintracellular signaling, organelle transport and cellular locomotion[8].Taxol induces apoptosis in cortical neurons by a mechanism distinctfrom that in non-neuronal cells. In contrast to non-neuronalcells, expression of wild type Bcl-2 in cortical neurons protectsagainst taxol-induced apoptosis [7]. Moreover, taxol induces activationof N-terminal c-Jun protein kinase (JNK) and phosphorylationof Bcl-2 in cancer cells [4], whereas it induces neither JNK activationnor phosphorylation of Bcl-2 in cortical neurons [7].NADPH oxidase is a superoxide-producing enzyme that consistsof membrane (gp91 PHOX and p22 PHOX ) and cytosolic (p47 PHOX ,∗ Corresponding author. Fax: +82 62 232 6974.E-mail address: ckkim@jnu.ac.kr (J.-K. Kim).p67 PHOX , and p40 PHOX ) components [3,5]. Although NADPH oxidasewas originally characterized as part of the host defense machineryin phagocytes, all of the subunits of NADPH oxidase are alsoexpressed in cortical neurons [10,15] and are involved in ROSinducedoxidative cell death in some neurodegenerative diseases[16,18].In the present study, we examined the effects of some antioxidantson neuronal cell death induced by taxol and investigatedwhether NADPH oxidase plays a role in taxol-induced neuronaldeath.4-(2-Aminoethyl)benzenesulfonylfluoride (AEBSF), apocynin,tempol and N-benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone(z-VAD-fmk) were obtained from Calbiochem Corporation(San Diego, CA, USA). Fetal bovine serum and horse serum werefrom Hyclone (Logan, UT, USA). Media for cell culture were purchasedfrom Gibco BRL (Rockville, MD, USA) and other reagentswere from Sigma (St Louis, MO, USA).Mixed cortical cell cultures, containing both neurons andastroglia, were prepared from fetal ICR mice at 14–15 days of0304-3940/$ – see front matter © 2008 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/j.neulet.2008.07.049

18 H.J. Jang et al. / Neuroscience Letters 443 (2008) 17–22gestation as described previously [14]. Briefly, dissociated corticalcells were plated onto a previously established astroglial cell monolayerat 3 hemispheres per 24-well plate (Primaria, Falcon, USA)in plating medium (Modified Eagle medium supplemented with21 mM glucose, 26.5 mM sodium bicarbonate, 2 mM glutamine, 5%fetal bovine serum, and 5% horse serum). Cytosine arabinoside(10 M) was added 5–6 days after plating to halt the growth ofnon-neuronal cells.Treatments with taxol and other drugs were performed at DIV12–14 in serum-free MEM (minimal essential medium with Earle’ssalts). The stock solutions of drugs were made at 100× concentrationof working solution in appropriate solvent and diluted in MEMimmediately before experiments. In addition to morphologicalassessment under a phase-contrast microscopy, overall neuronalcell injury in mixed cortical cultures was quantitatively assessedby measuring lactate dehydrogenase (LDH) activity released fromdamaged cells into the bathing medium [11]. Each LDH value, aftersubtracting the mean background value in sham-washed controls(=0), was scaled to the mean value in positive control culturestreated with 500 M NMDA for 24 h (=100), which induced nearcomplete neuronal death in the absence of glial damage.SYTOX Green (Invitrogen, Carlsbad, CA, USA) staining was usedfor morphological evaluation of nuclei. After fixing the cells with4% paraformaldehyde for 50–60 min at room temperature, the cellswere permeabilized with 0.5% Triton X-100 for 10 min and treatedwith 1 M SYTOX Green stock for 15 min.Changes of microtubule network after exposure to taxol werevisualized by fluorescence immunocytochemistry for tubulin. Afterfixation with 4% paraformaldehyde for 30 min and blocking, cellswere labelled with rabbit anti-tubulin antibody (Sigma) at 1:80dilution at 4 ◦ C overnight. After washes, signals were visualizedby Cy3-conjugated affinity purified anti-rabbit IgG (1:200,Jackson ImmunoResearch, West Grove, PA) using epifluorescencemicroscopy.Intracellular free radical generation was measured using 2,7-dichlorofluorescin diacetate (DCF) (Invitrogen, Carlsbad, CA, USA)[15]. Briefly, cells were loaded with 10 M DCF for 30 min and thentreated with taxol alone or in combination with apocynin or trolox.After treatments, ROS generation was monitored at 0.5, 1, 2, 4, 6,and 8 h at 37 ◦ C in a fluorescent multiwell plate reader (FluoroskanAscent FL, Labsystems) with excitation at 485 nm and emission at530 nm.RNA was prepared with TRIZOL and reverse-transcribed intocDNA with random primer and M-MLV reverse transcriptase(Invitrogen). PCR was performed under the following conditions:denaturing for 1 min at 94 ◦ C; annealing for 1 min atFig. 1. Taxol-induced concentration-dependent neuronal apoptosis and effect of antioxidants on neuronal cell death induced taxol or vinblastine in mixed cortical cultures.(A) Concentration-dependent neuronal death in mixed cortical cultures. Neuronal death was assessed by LDH release. Each point and bar represents mean ± S.E.M. from 4to 12 wells. (B) Photomicrographs of Sytox green nuclear staining showing chromatin condensation and nuclear fragmentation (arrows) induced by 8-h exposure to 300 nMtaxol (Tax) and 100 nM vinblastine (Vin) in mouse cortical cultures. Sham-wash control (Sham). Calibration bar: 50 m. (C) Inhibitory effect of 1 g/mL cycloheximide (+CHX),100 M z-VAD-fmk (+ZVAD), 100 M trolox, 100 M ascorbic acid, and 5 mM tempol on taxol-induced neuronal death. Each column and bar is the mean ± S.E.M. from 12to 16 wells. *P < 0.05, compared with taxol-treated control group. (D) Effect of 100 M z-VAD-fmk (+ZVAD), 100 M trolox, 100 M ascorbic acid, 500 M apocynin (+Apo)and 100 nM DPI on the 100 nM vinblastine (Vin)-induced neuronal death in mixed cortical cultures. Each column and bar is the mean ± S.E.M. from 8 to 12 wells. *P < 0.05,compared with vinblastine-treated group.

H.J. Jang et al. / Neuroscience Letters 443 (2008) 17–22 1958 ◦ C for p47 phox , 57 ◦ C for gp91 phox and HPRT; extensionfor 2min at 72 ◦ C (35 cycles for p47 phox and gp91 phox ; 30cycles for HPRT). PCR products were run on 1.5% agarosegels, visualized with ethidium bromide and quantitativelymeasured with an image analyzer. Primer sequences wereas follows: 5 ′ -CAGCCAGCACTATGTGTACA-3 ′ and 5 ′ -GAACTCGTA-GATCTCGGTGAA-3 ′ for p47 phox (91 bp); 5 ′ -GGTTTATGATGAT-GGGCCTAA-3 ′ and 5 ′ -GCACTGGAACCC CTGAGAAA-3 ′ , for gp91 phox(158 bp); 5 ′ -GTAATGATCAGTCAACGGGGGAC-3 ′ and 5 ′ -CCAGCAA-GCTTGCAACCTTAACCA-3 ′ for HPRT (200 bp).For knockdown of gp91 phox expression, Stealth TM siRNA ofgp91 phox (Invitrogen) was used. The sense and antisense sequencesof gp91 phox were CCAUCCACACAAUUGCA CAUCUCUU and AAGA-GAUGUGCAAUUGUGUGGAUGG, respectively. At DIV 6, the corticalcultures were transfected with 60 pmol siRNA of gp91 phox or negativecontrol siRNA using the GeneSilencer siRNA transfectionreagent (Gene Therapy Systems, San Diego, CA) as described previously[1]. In brief, siRNA and GeneSilencer were each diluted to25 L with media and then mixed together for 10 min. Cells werefed 500 L of fresh culture medium and overlaid with the transfectionmix. After 24 h, the cells were fed with 500 L of culturemedium. Transfected cells were used for experiments at DIV 13–14.Cortical cells were lysed in lysis buffer and centrifuged at13,000 rpm for 15 min. The pellet was discarded, and the supernatantwas used for protein quantification. Fractionation of cytosoland membrane was performed as described previously [14]. Equalamounts of protein from total cell lysates or membrane or cytosolicfraction were electrophoresed on 12% SDS–PAGE and transferredto nitrocellulose membranes using Towbin buffer. The membraneswere then incubated with anti-p47 phox or anti-gp91 phox antibodies(Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) overnightat 4 ◦ C. Immunoreactive proteins were detected using enhancedchemiluminescence protocol (Amersham Life Science).Results are expressed as mean ± S.E.M. Significant differencesbetween means were determined by analysis of variance withTukey multiple comparisons post hoc analysis using Instat software(Graphpad Software Inc., CA, USA). P < 0.05 was consideredstatistically significant.Mixed cortical cultures were exposed to selected concentrationsof taxol for 24 h. The neurotoxicity of taxol began to appear at 30 nMwith about 10% neuronal death, and neuronal death increased to33%, 46% and 72% with increasing concentrations of 100, 300 and1000 nM, respectively (Fig. 1A). In our culture systems, taxol barelyinjured astrocytes in the 30–1000 nM concentration range (datanot shown). Taxol induced chromatin condensation and nuclearfragmentation in cortical cultures (Fig. 1B). In addition to the morphologicalapoptotic features, treatment with anti-apoptotic drugs,such as cycloheximide (1 g/mL, a protein synthesis inhibitor) andz-VAd-fmk (100 M, a pan-caspase inhibitor), also significantlyattenuated the taxol-induced neuronal death (TIND) (Fig. 1C).These data demonstrate that taxol induces neuronal apoptosis ina concentration-dependent manner in mixed cortical cultures, andthis result is in agreement with the report by Figueroa-Masot etal. [7]. To examine the involvement of oxidative stress in TIND, weinvestigated the effect of some antioxidants, such as trolox, ascorbicacid and tempol, on TIND. All three antioxidants significantlyinhibited TIND (Fig. 1C). These results indicate that oxidative stressis involved in the process of TIND. Vinblastine is a microtubuledisruptinganticancer drug like taxol, but it is different from taxolin that it causes depolymerization of microtubule [2]. Vinblastine(100 nM) induced about 50% neuronal cell death in corticalcultures. Vinblastine also induced chromatin condensation andFig. 2. Fluorescence imaging showing taxol-induced alteration of microtubule network and effects of trolox and apocynin on the changes of microbule network in corticalneurons. (A) Control cortical neurons. (B) At 80 min in presence of 300 nM taxol. Note taxol-induced neurite beadings (arrows). (C) Co-treated with 100 M trolox and taxolfor 80 min. (D) Co-treated with 500 M apocynin and taxol for 80 min. Calibration bar: 50 m.

20 H.J. Jang et al. / Neuroscience Letters 443 (2008) 17–22nuclear fragmentation (Fig. 1B). To examine whether oxidativestress is specific for TIND or not, we investigated the effects ofz-VAD-fmk and antioxidants on the vinblastine-induced neuronaldeaths. Treatment with z-VAD-fmk significantly attenuated thevinblastine-induced neuronal death, but antioxidants such as troloxand ascorbic acid or NADPH oxidase inhibitors such as apocynin andDPI failed to inhibit the vinblastine-induced death (Fig. 1D). Thesefindings indicate that oxidative stress is uniquely involved in TINDbut not in vinblastine-induced deaths.To delineate whether the oxidative stress is involved in upstreamor downstream of microtubule inhibition by taxol, we observedthe effects of taxol on microtubule network using fluorescenceimmunocytochemistry for tubulin. After treating neurons withtaxol for 80 min, neurite beadings were identified (Fig. 2B). So weinvestigated the effect of trolox and apocynin on the formation ofneurite beadings by taxol. Treatment with either trolox or apocynindid not affect formation of the neurite beadings (Fig. 2C andD). These results suggest that oxidative stress is downstream eventof microtubule inhibition by taxol.Based on our findings and reports that all subunits of NADPHoxidase are expressed in cultured cortical cells [10,15] and thatactivation of this enzyme is involved in ROS-induced oxidative celldeath in neurodegenerative diseases [9,18], we hypothesize thatactivation of NADPH oxidase may also be involved in TIND in mixedcortical cultures. First, we examined p47 phox and gp91 phox expressionat the mRNA level by RT-PCR analysis and at the protein levelby western blot. Whereas cortical cultures expressed low levels ofp47 phox and gp91 phox mRNA, 2-h exposure to taxol substantiallyincreased the mRNA levels for p47 phox and gp91 phox in cortical cultures(Fig. 3A). Western blot assays confirmed that taxol exposureincreased the expressions of both p47 phox and gp91 phox after 3 h(Fig. 3B) and induced translocation of the p47 phox to the membranein cortical cultures, a signature event for the activation of NADPHoxidase (Fig. 3C).Next, we examined whether the enhanced expression of NADPHoxidase contributed to ROS generation after exposure to taxol.Exposure of mixed cortical cultures to taxol for 3 h markedlyincreased cellular DCF fluorescence, an indicator for superoxide,compared with sham-washed controls. Treatment with eitherapocynin, an NADPH oxidase inhibitor [19], or trolox, a vitamin Eanalog, markedly reduced the DCF fluorescence (Fig. 4A). We monitoredthe changes in DCF fluorescence for 8 h after exposure to taxol.As shown in Fig. 4B, taxol markedly increased the fluorescence forup to 2 h and then gradually increased the fluorescence for up to6 h. Treatment with either apocynin or trolox significantly reducedthe increase of DCF fluorescence.The next question was whether ROS generated by NADPH oxidasecontributed to TIND. Treatment with known NADPH oxidaseFig. 3. Induction and activation of p47 phox and gp91 phox in mixed cortical cultures exposed to taxol. (A) RT-PCR for p47 phox and gp91 phox in sham-washed control cultures(S) and in sister cultures after 2-h exposure to 300 nM taxol (T). The panels are representative of three independent experiments. HPRT was used as a house-keeping gene.Bars denote densitometer readings of RT-PCR signals expressed as folds of respective control values (mean ± S.E.M.; n = 3). *P < 0.05, compared with respective controls. (B)Western blots for p47 phox and gp91 phox in sham-washed control cultures (S) and in sister cultures after 3-h exposure to 300 nM taxol (T). The panels are representative ofthree independent experiments. (C) Western blot for p47 phox in the cytosolic fraction and the membrane fraction of cortical cultures, 3 h after sham wash (S) and 300 nMtaxol exposure (T).

H.J. Jang et al. / Neuroscience Letters 443 (2008) 17–22 21Fig. 4. Reduction of taxol-induced ROS generation by apocynin or trolox. (A) DCF fluorescencein cortical cultures 3 h after sham wash (Sham) or 3 h exposure to 300 nMtaxol without (Tax) or with addition of 500 M apocynin (+Apo) or 100 M trolox(+Tro). (B) Relative ROS generation was analyzed by measuring DCF fluorescence incortical cultures exposed to 300 nM taxol without () or with 500 M apocynin (),100 Mtrolox(○) for the indicated times. Each point and bar is mean ± S.E.M. from6 to 12 wells. *P < 0.05, compared with taxol-treated group.inhibitors, such as apocynin, AEBSF [6] or DPI [12], significantlyreduced the neuronal cell death (Fig. 5A). Furthermore, suppressingthe expression of gp91 phox with siRNA also significantly attenuatedTIND (Fig. 5B).Taxol mainly causes peripheral neuropathy in humans and itsCNS toxicity is rare and transient [20]. In the present study, wedemonstrated that TIND is apoptotic based on morphological features,such as chromatin condensation and nuclear fragmentation,and intervention with anti-apoptotic drugs in mouse cortical cultures.Although taxol is known to induce apoptotic cell death inproliferating cells, taxol also induces apoptotic cell death in nonproliferatingcells like cortical neurons and cardiomyocytes [7,17].Interestingly, it has been suggested that taxol may be useful fortreating Alzheimer’s disease due to its stabilization of microtubules[13].The attenuation of TIND by treatment with antioxidants indicatesthat oxidative stress is involved in this process. Increasedoxidative stress is an important underlying factor for the pathogenesisof a number of neurodegenerative diseases, includingAlzheimer disease (AD), multiple sclerosis (MS) and stroke. Thepresent study has demonstrated that NADPH oxidase is expressedin mixed cortical cultures and exposure of the cortical cultures totaxol induces the expression of this enzyme. Several lines of evi-Fig. 5. Attenuation of taxol-induced neuronal death by NADPH oxidase inhibitorsand knockdown of gp91 phox by siRNA. (A) Effect of 500 M apocynin, 50 M AEBSFand 100 nM DPI on the 300 nM taxol-induced neuronal death in mixed corticalcultures. Each column and bar is the mean ± S.E.M. from 8 to 12 wells. *P < 0.05,compared with taxol-treated group. (B) Effect of knockdown of gp91 phox with siRNAon the 300 nM taxol-induced neuronal death in mixed cortical cultures. Each columnand bar is the mean ± S.E.M. from 8 to 12 wells. *P < 0.05, compared withtaxol-treated control group.dence indicate that NADPH oxidase participates in the process ofneuronal death in some pathologic conditions. For example, activationof NADPH oxidase was also observed in Alzheimer’s diseasebrain [16], and NADPH oxidase participates in the neuronal deathinduced by beta-amyloid peptide [9], in zinc-induced neuronaldeath [15] and in BDNF-induced neuronal death in cultured neurons[10].In the present study of mixed cortical cultures, the contributionby NADPH oxidase to TIND was directly supported by findingsthat antioxidants and inhibitors of NADPH oxidase attenuate bothROS generation and TIND. Furthermore, suppression of NADPH oxidasewith siRNA for gp91 phox also significantly reduced TIND. Thisagain supports the hypothesis that oxidative stress through activationof NADPH oxidase is a major mechanism in TIND. Consistentwith these findings, a recent study showed that taxol killed cancercells by generating ROS through enhancing the activity of NADPHoxidase [3].Vinblastine and taxol are involved in microtubule-disruptinganticancer drug, but they have opposite action on microtubule polymerization[8]. In present study, vinblastine also induced apoptotic

22 H.J. Jang et al. / Neuroscience Letters 443 (2008) 17–22neuronal cell deaths which were not attenuated by treatment withantioxidants or NADPH oxidase inhibitors. These results indicatethat oxidative stress is uniquely involved in TIND and suggest thatmechanisms of activating NADPH oxidase in TIND may be relatedto its mode of action on the microtubule polymerization.Although mechanisms underlying the activation of NADPH oxidasemust still be delineated, we report for the first time that taxolcan generate ROS through the expression and activation of NADPHoxidase and cause neuronal cell death in mixed cortical cells. ThisTIND in cortical cultures may be a useful model for neurodegenerativedisorders that are characterized by increased oxidative stresssuch as AD, MS and stroke.References[1] M. Aarts, K. Iihara, W.L. Wei, Z.G. Xiong, M. Arundine, W. Cerwinski, J.F. Mac-Donald, M. Tymianski, A key role for TRPM7 channels in anoxic neuronal death,Cell 115 (2003) 863–877.[2] J. Alexandre, Y. Hu, W. Lu, H. Pelicano, P. Huang, Novel action of paclitaxel againstcancer cells: bystander effect mediated by reactive oxygen species, Cancer Res.67 (2007) 3512–3517.[3] B.M. Babior, NADPH oxidase: an update, Blood 93 (1999) 1464–1476.[4] M.V. Blagosklonny, T. Fojo, Molecular effects of paclitaxel: myths and reality (acritical review), Int. J. Cancer 83 (1999) 151–156.[5] F.R. DeLeo, M.T. Quinn, Assembly of the phagocyte NADPH oxidase: molecularinteraction of oxidase proteins, J. Leukoc. Biol. 60 (1996) 677–691.[6] V. Diatchuk, O. Lotan, V. Koshkin, P. Wikstroem, E. Pick, Inhibition of NADPHoxidase activation by 4-(2-aminoethyl)-benzenesulfonyl fluoride and relatedcompounds, J. Biol. Chem. 272 (1997) 13292–13301.[7] X.A. Figueroa-Masot, M. Hetman, M.J. Higgins, N. Kokot, Z. Xia, Taxol inducesapoptosis in cortical neurons by a mechanism independent of Bcl-2 phosphorylation,J. Neurosci. 21 (2001) 4657–4667.[8] S. Honore, E. Pasquier, D. Braguer, Understanding microtubule dynamics forimproved cancer therapy, Cell. Mol. Life Sci. 62 (2005) 3039–3056.[9] A. Jana, K. Pahan, Fibrillar amyloid-beta peptides kill human primary neuronsvia NADPH oxidase-mediated activation of neutral sphingomyelinase,Implications for Alzheimer’s disease, J. Biol. Chem. 279 (2004) 51451–51459.[10] S.H. Kim, S.J. Won, S. Sohn, H.J. Kwon, J.Y. Lee, J.H. Park, B.J. Gwag, Brain-derivedneurotrophic factor can act as a pronecrotic factor through transcriptionaland translational activation of NADPH oxidase, J. Cell. Biol. 159 (2002)821–831.[11] J.Y. Koh, D.W. Choi, Quantitative determination of glutamate mediated corticalneuronal injury in cell culture by lactate dehydrogenase efflux assay, J. Neurosci.Methods 20 (1987) 83–90.[12] Y. Li, M.A. Trush, Diphenyleneiodonium, an NAD(P)H oxidase inhibitor, alsopotently inhibits mitochondrial reactive oxygen species production, Biochem.Biophys. Res. Commun. 253 (1998) 295–299.[13] R. Nelson, Microtubule-stabilising drugs may be therapeutic in AD, Lancet Neurol.4 (2005) 83–84.[14] K.M. Noh, Y.H. Kim, J.Y. Koh, Mediation by membrane protein kinase C of zincinducedoxidative neuronal injury in mouse cortical cultures, J. Neurochem. 72(1999) 1609–1616.[15] K.M. Noh, J.Y. Koh, Induction and activation by zinc of NADPH oxidase in culturedcortical neurons and astrocytes, J. Neurosci. 20 (2000) RC111.[16] S. Shimohama, H. Tanino, N. Kawakami, N. Okamura, H. Kodama, T. Yamaguchi,T. Hayakawa, A. Nunomura, S. Chiba, G. Perry, M.A. Smith, S. Fujimoto, Activationof NADPH oxidase in Alzheimer’s disease brains, Biochem. Biophys. Res.Commun. 273 (2000) 5–9.[17] E. Skobel, H. Kammermeier, Relation between enzyme release and irreversiblecell injury of the heart under the influence of cytoskeleton modulating agents,Biochim. Biophys. Acta 1362 (1997) 128–134.[18] G.Y. Sun, L.A. Horrocks, A.A. Farooqui, The roles of NADPH oxidase and phospholipasesA2 in oxidative and inflammatory responses in neurodegenerativediseases, J. Neurochem. 103 (2007) 1–16.[19] Q. Wang, K.D. Tompkins, A. Simonyi, R.J. Korthuis, A.Y. Sun, G.Y. Sun, Apocyninprotects against global cerebral ischemia-reperfusion-induced oxidativestress and injury in the gerbil hippocampus, Brain Res. 1090 (2006)182–189.[20] C.G. Ziske, B. Schöttker, M. Gorschlüter, U. Mey, R. Kleinschmidt, U. Schlegel, T.Sauerbruch, I.G. Schmidt-Wolf, Acute transient encephalopathy after paclitaxelinfusion: report of three cases, Ann. Oncol. 13 (2002) 629–631.

BRAIN RESEARCH 1239 (2008) 216– 225available at www.sciencedirect.comwww.elsevier.com/locate/brainresResearch ReportDifferential sensitivity of human glioblastoma LN18(PTEN-positive) and A172 (PTEN-negative) cells to Taxolfor apoptosisRan Zhang a , Naren L. Banik a , Swapan K. Ray b, ⁎a Department of Neurosciences, Medical University of South Carolina, Charleston, SC 29425, USAb Department of Pathology, Microbiology and Immunology, University of South Carolina School of Medicine, 6439 Garners Ferry Road,Building 2, Room C11, Columbia, SC 29209, USAARTICLE INFOABSTRACTArticle history:Accepted 21 August 2008Available online 4 September 2008Keywords:ApoptosisCytochrome cGlioblastomaPTENTaxolVEGFGlioblastoma is the most malignant brain tumor in humans and an average survival ofglioblastoma patients hardly exceeds 12 months. Taxol is a plant-derived anti-cancer agent,which has been used in the treatments of many solid tumors. Deletion or mutation ofphosphatase and tension homolog located on chromosome ten (PTEN) occurs in more than80% of glioblastomas. We examined the sensitivity of human glioblastoma LN18 (PTENpositive)and A172 (PTEN-negative) cells to Taxol for induction of apoptosis. Wright stainingshowed morphological features of apoptosis after treatment with different doses of Taxol for24 h. Significant amount of apoptosis occurred in LN18 cells after treatment with 25 nM Taxol,while in A172 cells only after treatment with 50 nM Taxol. Western blotting with an antibodythat could specifically detect activation or phosphorylation of Akt (p-Akt) did not show any p-Akt in LN18 cells but an increase in p-Akt in A172 cells. Activation of Akt in A172 cells could bereversed by pre-treatment of the cells with the phosphatidylinositol-3-kinase (PI3K) inhibitorLY294002, indicating involvement of PI3K activity in this process. Apoptosis occurred with anincrease in Bax:Bcl-2 and mitochondrial release of cytochrome c into the cytosol leading toactivation of mitochondria-dependent caspase cascade. Taxol did not cause upregulation ofvascular endothelial growth factor (VEGF), a key mediator of angiogenesis, in LN18 cells butsubstantial upregulation of VEGF in A172 cells. After treatment with Taxol, increases in p-Aktand VEGF could maintain survival and angiogenesis, respectively, in PTEN-negativeglioblastoma. As a single chemotherapy, Taxol might be more efficacious in PTEN-positiveglioblastoma than in PTEN-negative glioblastoma. Thus, our study showed differentialsensitivity of PTEN-positive and PTEN-negative glioblastoma cells to Taxol.© 2008 Elsevier B.V. All rights reserved.1. IntroductionTaxol is one of the most powerful anti-cancer compoundsamong all chemotherapeutic drugs (Gligorov and Lotz, 2004). Itis a plant-derived anti-cancer agent that has been efficacious invarious solid tumors in preclinical studies (Rose, 1992; Geneyet al., 2005; Das et al., 2008) as well as in clinical studies(Utsunomiya et al., 2006; Veltkamp et al., 2007). Taxol has beenwidely used for induction of apoptosis in many solid tumors(Gan et al., 1998; Gagandeep et al., 1999) including glioblastoma⁎ Corresponding author. Fax: +1 803 733 3192.E-mail address: raysk8@gw.med.sc.edu (S.K. Ray).0006-8993/$ – see front matter © 2008 Elsevier B.V. All rights reserved.doi:10.1016/j.brainres.2008.08.075

BRAIN RESEARCH 1239 (2008) 216– 225217(Cahan et al., 1994; Das et al., 2008). It activates an intrinsicpathway of apoptosis via mitochondrial releaseof cytochrome c into the cytosol leading to formation ofapoptosome for activation of caspases (Jordan and Wilson,2004). Taxol is not so effective for treatment of recurrentglioblastoma even though it has shown great effectiveness intreatment of high-grade glioblastoma (Glantz et al., 1999;Rosenthal et al., 2000). However, the anti-cancer efficacy ofTaxol in glioblastoma can be significantly increased incombination with other therapeutic agents (Son et al., 2006;Karmakar et al., 2008) and also by increasing its transportacross the blood-brain barrier (Fellner et al., 2002; Koziaraet al., 2004). Recently, there is an excitement about the use oftemozolomide for treatment of some glioblastomas thatcontain epigenetic silencing of the DNA repair gene O 6 -methylguanine methyltransferase (Hegi et al., 2005). Even theuse of temozolomide alone may fail (Huang et al., 2008) orshow only moderate efficacy (Yang et al., 2006) in recurrentand progressive glioblastomas. Temozolomide may alsopromote angiogenesis in glioblastomas (Fisher et al., 2007).So, research continues to advocate Taxol as an importantchoice for treatment of glioblastoma and to identify themolecular markers that can modulate (increase or decrease)Taxol efficacy in glioblastoma. Apoptosis machinery doesexist in glioblastomas (Ray et al., 2002). But recurrentglioblastomas may lack or harbor molecular markers of cellsurvival. Identification of such molecular markers thatmodulate cell survival in recurrent glioblastomas may helpdesign rational therapeutics for their treatment.Phosphatase and tensin homolog located on chromosometen (PTEN), which encodes a cytoplasmic enzyme with bothprotein and lipid phosphatase activity, is frequently mutatedor deleted at chromosome 10q23 in malignant glioblastomas(Sano et al., 1999; Ohgaki and Kleihues, 2007). Recent studiesindicate that mutation and deletion of PTEN account for morethan 80% of human glioblastomas. Glioblastoma is the mostprevalent and malignant brain tumor in humans and theaverage survival time of glioblastoma patients is less than12 months even after treatment with the currently availablebest therapeutic regimens (Jiang et al., 2007; Cloughesy et al.,2008). Angiogenesis is a crucial feature for the growth andprogression of glioblastoma (Hanahan and Folkman, 1996;Kleihues et al, 1999). In comparison with other malignancieselsewhere in the body, glioblastoma is a notorious tumor dueto its high capability of angiogenesis. Vascular endothelialgrowth factor (VEGF) plays a critical role in tumor angiogenesisand the level of expression of VGEF correlates with the degreeof malignancy of glioblastomas (Fischer et al., 2005; Reardonet al., 2008). Currently, it is a popular theory that attacking thetumor vascular bed instead of tumor cells themselves is aneffective anti-tumor strategy due to low cytotoxicity and lackof drug resistance problem in this approach (Steiner et al.,2004). Transfection of PTEN to glioblastoma cells coulddecrease the secretion of VEGF (Gomez-Manzano et al., 2003).Because of the modest anti-tumor activity of Taxol in someglioblastomas, we wanted to explore any potential role for thetumor suppressor PTEN in Taxol sensitivity in glioblastomasusing human glioblastoma LN18 (PTEN-positive) and A172(PTEN-negative) cells. Our findings suggest that the dramaticdifference in Taxol sensitivity for apoptosis in these two celllines is due to their PTEN status. Presence of PTEN makesglioblastoma cells sensitive to Taxol, which also reduces theexpression of VEGF in PTEN-positive cells but not in PTENnegativecells. Notably, VEGF is an important angiogenic factorand the primary focus of anti-angiogenic interventions. Ourstudy indicates that Taxol is a promising therapeutic agent forboth induction of apoptosis and inhibition of angiogenic factorin PTEN-positive glioblastoma cells.Fig. 1 – Western blotting to assess the levels of PTENexpression and activation of Akt in LN18 and A172 cells.(A) Representative Western blots showing the levels of PTENexpression. (B) Densitometric determination of the levels ofPTEN expression. (C) Representative Western blotsshowing activation or phosphorylation of Akt (p-Akt).(D) Densitometric determination of the levels of p-Akt.Treatments (24 h): Control, 25 nM Taxol, 50 nM Taxol, and20 μM LY294002 (1 h pre-treatment)+50 nM Taxol.Significant difference from control value was indicated by*p

218 BRAIN RESEARCH 1239 (2008) 216– 2252. Results2.1. Levels of PTEN expression in human glioblastomaLN18 and A172 cell linesOverexpression of the tumor suppressor PTEN in malignantcells contributes to apoptosis and suppression of survivalsignaling. We examined the levels of PTEN expression in LN18and A172 cells after treatments with 25 nM and 50 nM Taxol(Fig. 1). The results demonstrated that treatments of LN18 cellswith 25 nM and 50 nM Taxol induced PTEN expression (Fig. 1A)significantly (p

BRAIN RESEARCH 1239 (2008) 216– 22521925 nM and 50 nM Taxol significantly (p

220 BRAIN RESEARCH 1239 (2008) 216– 225active 37 kDa caspase-9 in LN18 and A172 cells after Taxoltreatments (Fig. 5B). Further, colorimetric assay confirmedsignificant (p

BRAIN RESEARCH 1239 (2008) 216– 225221(Fig. 7B) and caspase-3-specific 120 kDa SBDP (Fig. 7C),respectively, after Taxol treatments.2.6. Changes in expression of VEGF and bFGFVEGF is a key factor for changing tumor microenvironmentand tumor angiogenesis. It has been reported previously thatthe expression of VEGF in glioblastoma is regulated in a PI3Kdependentmanner (Maity et al., 2000). Because PTEN controlsPI3K activity, the level of PTEN may alter expression of VEGFand any other angiogenic factor. Expression of bFGF alsocontributes to tumor angiogenesis. We used Western blottingto examine any influence of PTEN status in changing levels ofthe potential angiogenic factors (VEGF and bFGF) in LN18 andA172 cells after Taxol treatments (Fig. 8). Western blotting (Fig.8A) showed that 50 nM Taxol caused a significant decrease inexpression of VEGF (Fig. 8B) and no change in expression ofbFGF (Fig. 8C) in PTEN-positive LN18 cells. On the other hand,Taxol treatments increased the levels of VEGF significantly(Fig. 8B) and bFGF non-significantly (Fig. 8C) in the PTENnegativeA172 cells.Fig. 8 – Western blotting to estimate the levels of VEGF andbFGF in LN18 and A172 cells. Treatments (24 h): Control,25 nM Taxol, and 50 nM Taxol. (A) Representative Westernblots to show the changes in the levels of VEGF and bFGF.Expression of β-actin was examined to ensure almostuniform protein loading in all lanes. (B) Bar diagrams to showdensitometric determination of the levels of VEGF. (C) Bardiagrams to show densitometric determination of the levelsof bFGF. Significant difference from control value wasindicated by *p

222 BRAIN RESEARCH 1239 (2008) 216– 225treatment (Fig. 5). We performed Western blotting andcolorimetric assay to observe increases in active 37 kDacaspase-9 fragment and caspase-9 activity, respectively,after Taxol treatments. The low dose of Taxol was sufficientto cause the highest increases in caspase-9 activation andactivity in LN18 cells while Taxol dose-dependently increasedthe caspase-9 activation and activity in A172 cells (Fig. 5). Wealso examined the increases in active 20 kDa caspase-3fragment and caspase-3 activity following Taxol treatments(Fig. 6). The low and high doses of Taxol were most efficaciousin LN18 and A172 cells, respectively, for increasing thecaspase-3 activation and activity. Similarly, the low and highdoses of Taxol were the most effective in LN18 and A172 cells,respectively, for proteolysis of 270 kDa α-spectrin to generatethe calpain-specific 145 kDa SBDP and caspase-3-specific120 kDa SBDP (Fig. 7). These results demonstrated that bothlow and high doses of Taxol in LN18 and A172 cells caused thehighest increases in proteolytic activities of calpain andcaspase-3. Therefore, the presence of PTEN in LN18 cellsappeared to help increase the sensitivity to Taxol for activationof proteolytic activities for apoptosis.Because VEGF and bFGF are the critical angiogenic factors forpromoting growth in glioblastoma, we examined the changes inthe levels of VEGF and bFGF in LN18 and A172 cells after Taxoltreatments (Fig. 8). Taxol treatments significantly decreasedexpression of VEGF in LN18 cells but increased in A172 cells. Ithas previously been shown that transfection of PTEN into theglioblastoma U87MG cells using retroviral vector could suppresstumor angiogenesis due to decrease in VEGF at the mRNA levelin vitro as well as in vivo (Wen et al., 2001; Abe et al., 2003).When compared with the PTEN-positive cells, the PTENnegativecells showed high secretion of VEGF protein (Zundelet al., 2000; Edwards et al., 2008). Integrin-linked kinase (ILK) is atherapeutic target in cancer treatment. Studies demonstratedthat the PTEN-positive cells were sensitive to ILK inhibitorsthan the PTEN-negative cells and also suggested that ILKinhibitors decreased VEGF secretion due to induction of PTENexpression (Edwards et al., 2008). Our data indicated that Taxolcould decrease expression of VEGF significantly and bFGF tosome extent in PTEN-positive LN18 cells but increased theirlevels in PTEN-negative A172 cells.In conclusion, our current study demonstrated differentialanti-cancer effects of Taxol in human glioblastoma LN18(PTEN-positive) and A172 (PTEN-negative) cells. Activation ofthe cell survival PI3K/Akt signaling pathway and lastingexpression of the anti-apoptotic Bcl-2 in PTEN-negativeglioblastoma cells could confer resistance to Taxol chemotherapy.This study implies that the absence of PTEN inrecurrent glioblastomas may help avoid apoptosis and promoteangiogenesis leading to overall poor prognosis of theseglioblastomas even after treatment with a powerful chemotherapeuticdrug such as Taxol.4. Experimental procedures4.1. MaterialsHuman glioblastoma cell lines LN18 and A172 were purchasedfrom the American Type Culture Collection (ATCC, Rockville,MD, USA). We obtained polyclonal IgG antibodies against PTENfrom Cell Signaling Technologies (Danvers, MA, USA), cytochromec from BD Biosciences (San Jose, CA, USA), and Bax,Bcl-2, VEGF, and bFGF from Santa Cruz Biotechnology (SantaCruz, CA, USA). Monoclonal IgG antibody against α-spectrinwas received from Affiniti (Exeter, UK).4.2. Cell cultureHuman glioblastoma LN18 and A172 cell lines were separatelygrown in DMEM supplemented with 10% fetal bovine serum(FBS) and 1% penicillin and streptomycin. Cells were seeded in75-cm 2 flasks and incubated at 37 °C in a fully-humidifiedatmosphere with 5% CO 2 . Once the cells were 80% confluent,they were starved in DMEM with 1% FBS for 24 h andmaintained in this low serum condition in the course of alltreatments. Cells were treated with 25 nM and 50 nM Taxol for24 h. After the treatments, cells were processed for assessmentsof residual cell viability, morphological and biochemicalfeatures of apoptosis, apoptosis-related proteins byWestern blotting, and caspase activities by colorimetricassays.4.3. Cell viability assayAfter the treatments of cells with Taxol, trypan blue dyeexclusion test was performed to evaluate the residual cellviability (Ray et al., 1999; 2006). Viable cells maintainedmembrane integrity and did not take up trypan blue. Cellswith compromised cell membranes took up trypan blue andtherefore were counted as dead. At least 600 cells in eachtreatment were counted in four different fields and thenumber of viable cells was calculated as percentage of thetotal cell population.4.4. Wright staining for morphological featuresof apoptosisCells from all treatments were harvested and washed in PBS,pH 7.4, and sedimented onto the microscopic slides using theCentra CL2 centrifuge (IEC, Needham Heights, MA, USA) at1000 rpm for 5 min. Cells were fixed for Wright staining, as wereported previously (Das et al., 2006). Cellular morphology wasexamined under the light microscopy to examine amount ofapoptosis. Cells were considered apoptotic with the characteristicssuch as reduction in cell volume, condensation of thechromatin, and/or the presence of cell membrane blebbing.About 600 cells were counted in each treatment and thepercentage of apoptotic cells was calculated.4.5. ApopTag peroxidase assay for detection of apoptoticDNA fragmentationFor detection of DNA fragmentation as a biochemical markerof apoptosis, glioblastoma cells on the microscopic slides weresubjected to the TdT-mediated dUTP nick-end labeling(TUNEL) using the ApopTag peroxidase assay kit (Intergen,Purchase, NY, USA), as we reported (Ray et al., 2006). In TUNELstaining, we used 3,3′-diaminobenzidine (DAB) as a peroxidasesubstrate. Cells with DNA fragmentation produced a brown

BRAIN RESEARCH 1239 (2008) 216– 225223product from oxidative polymerization and cyclization of DABin the course of the ApopTag peroxidase assay. ApopTagpositivecells were brown in a pale green background and wereconsidered as apoptotic cells. Experiments were conducted intriplicate and the percentage of ApopTag-positive cells wasdetermined by counting the brown cells from randomlyselected fields under the light microscope.4.6. Protein extraction and Western blottingAfter the treatments, cells were lysed in a buffer composed of50 mM Tris–HCl, pH 7.4, 0.1 mM phenylmethylsulfonyl fluoride(PMSF), and 5 mM EGTA for extraction of cellular proteins.Concentration of total proteins was determined colorimetricallyusing Coomassie-Plus protein assay reagent (Pierce,Rockford, IL, USA). The samples were mixed with an equalvolume of 2× loading buffer [125 mM Tris–HCl, pH 6.8, 4%sodium dodecyl sulfate (SDS), 20% glycerol, 200 mM 1,4-dithio-DL-threitol (DTT), and 0.02% bromophenol blue], boiled for5 min, and loaded (40 μg/lane) onto the 4–20% gradient gels forthe SDS-polyacrylamide gel electrophoresis (SDS-PAGE). AfterSDS-PAGE, the gels were blotted to Immunobilon-P nylonmembrane. The blots were blocked in 5% non-fat milk, 0.1%Tween, Tris–HCl, pH 7.8, for 2 h at room temperature. Then theblots were incubated with a specific primary IgG antibody for2 h at room temperature or overnight at a cold room followedby alkaline horseradish peroxidase-conjugated secondary IgGantibody for 1 h. Blots were developed using the enhancedchemiluminescence (ECL) or ECL-Plus reagents (AmershamPharmacia, Buckinghamshire, UK). The ECL autoradiogramswere scanned on a PowerLook Scanner (Umax Technologies,Fremont, CA, USA) using Photoshop software (Adobe Systems,Seattle, WA, USA) and the optical density (OD) of each bandwas determined using Quantity One software (Bio-Rad,Hercules, CA, USA).4.7. Analysis of mitochondrial release of cytochrome c intothe cytosolCells from each treatment were harvested, washed once withice-cold PBS, and gently lysed for 1 min in 50 μl ice-cold lysisbuffer (250 mM sucrose, 1 mM EDTA, 0.05% digitonin, 25 mMTris–HCl, pH 6.8, 1 mM DTT, 1 μg/ml leupeptin, 1 μg/mlpepstatin, 1 μg/ml aprotinin, 1 mM benzamidine, and 0.1 mMPMSF) following the standard procedure (Piqué et al., 2000).Lysates were centrifuged at 12,000 ×g in cold (4 °C) for 3 minto obtain a pellet (the fraction containing mitochondria) andsupernatant (the cytosolic extract without mitochondria).Pellet and supernatant were analyzed by Western blottingusing cytochrome c antibody.4.8. Colorimetric assays for measurement of caspase-9and caspase-3 activitiesMeasurements of caspase-9 and caspase-3 activities in cellswere performed with the commercially available colorimetricassay kits (Sigma Chemical, St. Louis, MO, USA). Proteolyticactivities of caspase-9 and caspase-3 were assayed based onthe release of p-nitroanilide (p-NA) from LEHD-p-NA andDEVD-p-NA, respectively. The concentration of the p-NAreleased from the substrate was calculated from the absorbanceat 405 nm.4.9. Statistical analysisResults were analyzed using StatView software (AbacusConcepts, Berkeley, CA, USA) and compared using one-wayanalysis of variance (ANOVA) with Fisher's post hoc test.Data were presented as mean±standard deviation (SD) ofseparate experiments (n ≥ 3). A significant difference fromcontrol value was indicated by ⁎(p

224 BRAIN RESEARCH 1239 (2008) 216– 225Fellner, S., Bauer, B., Miller, D.S., Schaffrik, M., Fankhänel, M.,Spruss, T., Bernhardt, G., Graeff, C., Färber, L., Gschaidmeier, H.,Buschauer, A., Fricker, G., 2002. Transport of paclitaxel (Taxol)across the blood-brain barrier in vitro and in vivo. J. Clin. Invest.110, 1309–1318.Fischer, I., Gagner, J.P., Law, M., Newcomb, E.W., Zagzag, D., 2005.Angiogenesis in gliomas: biology and molecularpathophysiology. Brain Pathol. 15, 297–310.Fisher, T., Galanti, G., Lavie, G., Jacob-Hirsch, J., Kventsel, I.,Zeligson, S., Winkler, R., Simon, A.J., Amariglio, N., Rechavi, G.,Toren, A., 2007. Mechanisms operative in the antitumoractivity of temozolomide in glioblastoma multiforme. Cancer J.13, 335–344.Gagandeep, S., Novikoff, P.M., Ott, M., Gupta, S., 1999. Paclitaxelshows cytotoxic activity in human hepatocellular carcinomacell lines. Cancer Lett. 136, 109–118.Gan, Y., Wientjes, M.G., Au, J.L., 1998. Relationship betweenpaclitaxel activity and pathobiology of human solid tumors.Clin. Cancer Res. 4, 2949–2955.Geney, R., Chen, J., Ojima, I., 2005. Recent advances in the newgeneration taxane anticancer agents. Med. Chem. 1,125–139.Glantz, M.J., Chamberlain, M.C., Chang, S.M., Prados, M.D.,Cole, B.F., 1999. The role of paclitaxel in the treatment ofprimary and metastatic brain tumors. Semin. Radiat. Oncol. 9,27–33.Gligorov, J., Lotz, J.P., 2004. Preclinical pharmacology of thetaxanes: implications of the differences. Oncologist 9 (Suppl. 2),3–8.Gomez-Manzano, C., Fueyo, J., Jiang, H., Glass, T.L., Lee, H.Y.,Hu, M., Liu, J.L., Jasti, S.L., Liu, T.J., Conrad, C.A., Yung, W.K.,2003. Mechanisms underlying PTEN regulation of vascularendothelial growth factor and angiogenesis. Ann. Neurol. 53,109–117.Hanahan, D., Folkman, J., 1996. Patterns and emergingmechanisms of the angiogenic switch during tumorigenesis.Cell 86, 353–364.Hegi, M.E., Diserens, A.C., Gorlia, T., Hamou, M.F., de Tribolet, N.,Weller, M., Kros, J.M., Hainfellner, J.A., Mason, W., Mariani, L.,Bromberg, J.E., Hau, P., Mirimanoff, R.O., Cairncross, J.G.,Janzer, R.C., Stupp, R., 2005. MGMT gene silencing and benefitfrom temozolomide in glioblastoma. N. Engl. J. Med. 352,997–1003.Huang, F., Kavan, P., Guiot, M.C., Markovic, Y., Roberge, D., 2008.When temozolomide alone fails: adding procarbazine insalvage therapy of glioma. Can. J. Neurol. Sci. 35, 192–197.Jiang, Z., Pore, N., Cerniglia, G.J., Mick, R., Georgescu, M.M.,Bernhard, E.J., Hahn, S.M., Gupta, A.K., Maity, A., 2007.Phosphate and tensin homologue deficiency in glioblastomaconfers resistance to radiation and temozolomide that isreversed by the protease inhibitor nelfinavir. Cancer Res. 67,4467–44473.Jordan, M.A., Wilson, L., 2004. Microtubules as a target foranticancer drugs. Nat. Rev. Cancer 4, 253–265.Kargiotis, O., Rao, J.S., Kyritsis, A.P., 2006. Mechanisms ofangiogenesis in gliomas. J. Neurooncol. 78, 281–293.Karmakar, S., Banik, N.L., Ray, S.K., 2008. Combination of all-transretinoic acid and paclitaxel-induced differentiation andapoptosis in human glioblastoma U87MG xenografts in nudemice. Cancer 112, 596–607.Kaur, B., Tan, C., Brat, D.J., Post, D.E., Van Meir, E.G., 2004.Genetic and hypoxic regulation of angiogenesis in gliomas.J. Neurooncol. 70, 229–243.Kleihues, P., Ohgaki, H., 1999. Primary and secondaryglioblastomas: from concept to clinical diagnosis.Neuro-oncology 1, 44–51.Koziara, J.M., Lockman, P.R., Allen, D.D., Mumper, R.J., 2004.Paclitaxel nanoparticles for the potential treatment of braintumors. J. Control Release 99, 259–269.Maity, A., Pore, N., Lee, J., Solomon, D., O'Rourke, D.M., 2000.Epidermal growth factor receptor transcriptionallyup-regulates vascular endothelial growth factor expression inhuman glioblastoma cells via a pathway involvingphosphatidylinositol 3′-kinase and distinct from that inducedby hypoxia. Cancer Res. 60, 5879–5886.Merighi, S., Benini, A., Mirandola, P., Gessi, S., Varani, K., Leung, E.,Maclennan, S., Baraldi, P.G., Borea, P.A., 2007. Hypoxia inhibitspaclitaxel-induced apoptosis through adenosine-mediatedphosphorylation of bad in glioblastoma cells. Mol. Pharmacol.72, 162–172.Ohgaki, H., Kleihues, P., 2007. Genetic pathways to primary andsecondary glioblastoma. Am. J. Pathol. 170, 1445–1453.Parsons, R., 2004. Human cancer, PTEN and the PI-3 kinasepathway. Semin. Cell Dev. Biol. 15, 171–176.Piqué, M., Barragán, M., Dalmau, M., Bellosillo, B., Pons, G., Gil, J.,2000. Aspirin induces apoptosis through mitochondrialcytochrome c release. FEBS Lett. 480, 193–196.Pore, N., Lius, A., Haas-Kogan, D.A., O'Rourke, D.M., Maity, A., 2003.PTEN mutation and epidermal growth factor receptoractivation regulate vascular endothelial growth factor (VEGF)mRNA expression in human glioblastoma cells bytransactivating the proximal VEGF promoter. Cancer Res.63, 236–241.Ray, S.K., Wilford, G.G., Crosby, C.V., Hogan, E.L., Banik, N.L., 1999.Diverse stimuli induce calpain overexpression and apoptosisin C6 glioma cells. Brain Res. 829, 18–27.Ray,S.K.,Patel,S.J.,Welsh,C.T.,Wilford,G.G.,Hogan,E.L.,Banik, N.L., 2002. Molecular evidence of apoptotic death inmalignant brain tumors including glioblastoma multiforme:upregulation of calpain and caspase-3. J. Neurosci. Res. 69,197–206.Ray, S.K., Karmakar, S., Nowak, M.W., Banik, N.L., 2006. Inhibitionof calpain and caspase-3 prevented apoptosis and preservedelectrophysiological properties of voltage-gated andligand-gated ion channels in rat primary cortical neuronsexposed to glutamate. Neuroscience 139, 577–595.Reardon, D.A., Desjardins, A., Rich, J.N., Vredenburgh, J.J., 2008.The emerging role of anti-angiogenic therapy for malignantglioma. Curr. Treat. Options Oncol. 9, 1–22.Rong, Y., Post, D.E., Pieper, R.O., Durden, D.L., Van Meir, E.G.,Brat, D.J., 2005. PTEN and hypoxia regulate tissue factorexpression and plasma coagulation by glioblastoma. CancerRes. 65, 1406–1413.Rose, W.C., 1992. Taxol: a review of its preclinical in vivo antitumoractivity. Anticancer Drugs 3, 311–321.Rosenthal, M.A., Gruber, M.L., Glass, J., Nirenberg, A., Finlay, J.,Hochster, H., Muggia, F.M., 2000. Phase II study ofcombination taxol and estramustine phosphate in thetreatment of recurrent glioblastoma multiforme. J. Neurooncol.47, 59–63.Sano, T., Lin, H., Chen, X., Langford, L.A., Koul, D., Bondy, M.L.,Hess, K.R., Myers, J.N., Hong, Y.K., Yung, W.K., Steck, P.A., 1999.Differential expression of MMAC/PTEN in glioblastomamultiforme: relationship to localization and prognosis. CancerRes. 59, 1820–1824.Son, M.J., Song, H.S., Kim, M.H., Kim, J.T., Kang, C.M., Jeon, J.W.,Park, S.Y., Kim, Y.J., Groves, M.D., Park, K., Kim, J.H., Nam,D.H., 2006. Synergistic effect and condition of pegylatedinterferon alpha with paclitaxel on glioblastoma. Int. J.Oncol. 28, 1385–1392.Steiner, H.H., Karcher, S., Mueller, M.M., Nalbantis, E., Kunze, S.,Herold-Mende, C., 2004. Autocrine pathways of the vascularendothelial growth factor (VEGF) in glioblastoma multiforme:clinical relevance of radiation-induced increase of VEGF levels.J. Neurooncol. 66, 129–138.Utsunomiya, H., Akahira, J., Tanno, S., Moriya, T., Toyoshima, M.,Niikura, H., Ito, K., Morimura, Y., Watanabe, Y., Yaegashi, N.,2006. Paclitaxel–platinum combination chemotherapy for

BRAIN RESEARCH 1239 (2008) 216– 225225advanced or recurrent ovarian clear cell adenocarcinoma: amulticenter trial. Int. J. Gynecol. Cancer 16, 52–56.Veltkamp, S.A., Rosing, H., Huitema, A.D., Fetell, M.R., Nol, A.,Beijnen, J.H., Schellens, J.H., 2007. Novel paclitaxelformulations for oral application: a phase Ipharmacokinetic study in patients with solid tumours. CancerChemother. Pharmacol. 60, 635–642.Wen, S., Stolarov, J., Myers, M.P., Su, J.D., Wigler, M.H., Tonks,N.K., Durden, D.L., 2001. PTEN controls tumor-inducedangiogenesis. Proc. Natl. Acad. Sci. U. S. A. 98, 4622–4627.Yang, S.H., Kim, M.K., Lee, T.K., Lee, K.S., Jeun, S.S., Park, C.K.,Kang, J.K., Kim, M.C., Hong, Y.K., 2006. Temozolomidechemotherapy in patients with recurrent malignant gliomas.J. Korean Med. Sci. 21, 739–744.Zhang, R., Banik, N.L., Ray, S.K., 2007. Combination ofall-trans retinoic acid and interferon-gamma suppressedPI3K/Akt survival pathway in glioblastoma T98G cellswhereas NF-κB survival signaling in glioblastoma U87MGcells for induction of apoptosis. Neurochem. Res. 32,2194–2202.Zundel, W., Schindler, C., Haas-Kogan, D., Koong, A., Kaper, F.,Chen, E., Gottschalk, A.R., Ryan, H.E., Johnson, R.S., Jefferson, A.B., Stokoe, D., Giaccia, A.J., 2000. Loss of PTEN facilitatesHIF-1-mediated gene expression. Genes Dev. 14, 391–396.