Volume 14 Number 9 September 2012 pp. 846–854 846

www.neoplasia.com

Inhibition of SIRT2 Potentiates theAnti-motility Activity of Taxanes:Implications for AntineoplasticCombination Therapies1,2

Katiuscia Bonezzi*,3, Dorina Belotti*,3,Brian J. North†, Carmen Ghilardi‡,Patrizia Borsotti*, Andrea Resovi*,Paolo Ubezio‡, Antonella Riva§,Raffaella Giavazzi*,‡, Eric Verdin†

and Giulia Taraboletti*

*Tumor Angiogenesis Unit, Department of Oncology,Mario Negri Institute for Pharmacological Research,Bergamo, Italy; †Gladstone Institute of Virology andImmunology, University of California at San Francisco,San Francisco, CA; ‡Department of Oncology, MarioNegri Institute for Pharmacological Research, Milan,Italy; §Indena S.p.A., Milan, Italy

AbstractTaxanes are potent inhibitors of cell motility, a property implicated in their antiangiogenic and antimetastatic activityand unrelated to their antiproliferative effect. The molecular mechanism of this anti-motility activity is poorly under-stood. In this study, we found that paclitaxel induced tubulin acetylation in endothelial and tumor cells, at concentra-tions that affected cell motility but not proliferation (10−8 to 10−9 M, for 4 hours). Induction of tubulin acetylationcorrelated with inhibition of motility but not proliferation based on a comparison of highly and poorly cytotoxic taxanes(paclitaxel and IDN5390) and tumor cell lines sensitive and resistant to paclitaxel (1A9 and 1A9 PTX22). Consistentwith the hypothesis that tubulin deacetylase activity might affect cell response to the anti-motility activity of taxanes,we found that overexpression of the tubulin deacetylase SIRT2 increased cell motility and reduced cell response tothe anti-motility activity of paclitaxel. Conversely, the SIRT2 inhibitor splitomicin reduced cell motility and potentiatedthe anti-motility activity of paclitaxel. The inhibitory effect was further potentiated by the addition of the HDAC6inhibitor trichostatin A. Paclitaxel and splitomicin promoted translocation into the nucleus—and hence activation—of FOXO3a, a negative regulator of cell motility. This study indicates a role for SIRT2 in the regulation of cell motilityand suggests that therapies combining sirtuin inhibitors and taxanes could be used to treat cell motility–based patho-logic processes such as tumor angiogenesis, invasion, and metastasis.

Neoplasia (2012) 14, 846–854

Address all correspondence to: Giulia Taraboletti, PhD, Tumor Angiogenesis Unit,Department of Oncology, Mario Negri Institute for Pharmacological Research, CentroAnna Maria Astori, Parco Scientifico Tecnologico Kilometro Rosso, Via Stezzano, 87,24126 Bergamo, Italy. E-mail: giulia.taraboletti@marionegri.it1This study was supported by grants from the Italian Ministry of Health, contract N.ROStrategici 11/07, the Italian Association for Cancer Research (AIRC), and FondazioneCariplo. G. Taraboletti and R. Giavazzi received research support from Indena S.p.A.A. Riva is employed by Indena S.p.A. The other authors declare no potential conflictof interest.2This article refers to supplementary materials, which are designated by Figures W1 toW6 and Tables W1 to W3 and are available online at www.neoplasia.com.3These authors contributed equally to this work.Received 26 April 2012; Revised 6 August 2012; Accepted 6 August 2012

Copyright © 2012 Neoplasia Press, Inc. All rights reserved 1522-8002/12/$25.00DOI 10.1593/neo.12728

IntroductionCell motility is a central event in a number of physiological and path-ologic processes. Agents that perturb the finely tuned microtubuleorganization, stability, and dynamics inhibit cell motility and havebeen proposed as therapeutic inhibitors of motility-based pathologies,including cancer metastasis and pathologic angiogenesis. Among these,the antineoplastic drug paclitaxel and other taxanes have anti-motilityactivity for tumor and endothelial cells, a property thought to con-tribute to their antimetastatic and antiangiogenic activity. This effectof tubulin-targeting compounds has opened new therapeutic oppor-tunities to target the tumor vascular compartment and the metastaticdissemination of tumor cells [1,2].

Neoplasia Vol. 14, No. 9, 2012 SIRT2 and Taxanes in Cell Motility Bonezzi et al. 847

The molecular mechanisms of the anti-motility activity of taxanesis still poorly understood [3]. Three lines of evidence indicate thatthe effect of taxanes on cell motility is distinct from their cytotoxicactivity. First, cell motility is inhibited by taxanes at concentrationslower than the cytotoxic concentrations. A short exposure to lowconcentrations of paclitaxel has no effect on cell proliferation orapoptosis but affects microtubule dynamics, cell motility, and morpho-genesis [4–7]. Second, tumor cells harboring the A364T mutationin β-tubulin are resistant to the antiproliferative activity of paclitaxelbut nonetheless remain sensitive to the anti-motility effect of taxanes[8]. Third, the seco-derivative taxane IDN5390 exhibits potent anti-motility and antiangiogenic activity but a lower cytotoxicity com-pared to paclitaxel [9,10], further supporting the concept that theantiproliferative and anti-motility activity of taxanes work throughdifferent pathways.Microtubule-dependent events potentially responsible for the in-

hibition of cell motility—rather than proliferation—include pro-motion of microtubule stability, alteration of microtubule dynamics,and induction of tubulin posttranslational modifications such asacetylation. Taxanes increase tubulin acetylation [11,12], althoughthe functional relevance of this event in inhibition of cell motility hasnot been clarified.Reversible posttranslational modifications of tubulin are considered

responsible for the structural and functional diversity of microtubulesubpopulations, characterized by a fine spatial and temporal regulation[12,13]. Acetylation occurs at the ɛ-amino group of Lys40 in theα-tubulin, and although other acetylation sites have been recentlyidentified [14], acetylation of this particular lysine has been correlatedwith changes in cell motility [12]. Tubulin acetylation is mainly regu-lated by a complex constituted by the class IIb histone deacetylaseHDAC6 and the NAD-dependent histone deacetylase SIRT2.Sirtuins are NAD+-dependent protein deacetylases and are impli-

cated in a variety of biologic processes including metabolic regulationand cellular response to a variety of physiological stresses. SIRT2 is pre-dominantly cytoplasmic but shuttles between the nuclear compart-ment and the cytoplasm during interphase, and it is mainly nuclearduring mitosis [15]. It deacetylates molecular targets both in the nucleusand cytoplasm, such as histone H4 [16] and tubulin [17], and is involvedin the regulation of cell cycle progression [18–20].HDAC6 and SIRT2 interact in the cytoplasm, where they bind

tubulin, colocalize with microtubules, and deacetylate tubulin[17,21,22]. HDAC6 and SIRT2 differ in their ability to deacetylatetubulin substrates, because SIRT2 deacetylates tubulin also as celllysate–derived heterodimers and taxol-stabilized microtubules whereasthese forms of tubulin are relatively resistant to HDAC6 [17,23].Acetylation of α-tubulin is a marker of microtubule stability,

although whether there is a causal relationship between acetylationand stability is still debated [12]. Tubulin acetylation/deacetylationparticipates in the regulation of cell motility: Overexpression ofHDAC6 decreases tubulin acetylation and increases cell motility infibroblast [21,24] and in breast cancer cells [25]. Conversely, reducedcell motility and increased tubulin acetylation are observed in fibro-blasts and transformed cells following reduction of HDAC6 activity,either by deacetylase inhibitors [such as trichostatin A (TSA) andtubacin] [24,26] or following knockdown of the deacetylase [26,27].Less is known on the role of SIRT2 in cell motility.Taxanes and tubulin deacetylases share the same target, tubulin,

and both induce changes in tubulin acetylation. Both tubulin acetyla-tion and binding of taxanes are predicted to occur in the lumen of

the microtubule structure [3,11,13], suggesting that these compoundsmodulate the transmission of signals from the lumen of the micro-tubules to the cytoskeleton.

Selective HDAC6 inhibitors synergize with taxanes, and combina-tion regimens that include both classes of compounds have proveneffective in preclinical studies and are currently in clinical trials incancer patients [28,29]. Inhibitors of sirtuins have also been proposedas antineoplastic agents [23,29], although the exact sirtuin involved(SIRT1 vs SIRT2), the mechanisms of action, and the possibilities ofcombinations with other agents have not been investigated.

This study was designed to investigate whether modification oftubulin acetylation, particularly through the modulation of SIRT2activity, might affect cell susceptibility to the anti-motility effect oftaxanes. We describe that modulation of SIRT2 activity affects cellmotility and response to the anti-motility activity of taxanes, suggestingthat specific inhibitors of SIRT2 might represent a new tool to poten-tiate the activity of taxanes in antineoplastic combination therapies.

Materials and Methods

ReagentsPaclitaxel and IDN5390 were provided by Indena S.p.A. (Milan,

Italy). Splitomicin [30], TSA, and sirtinol (Sigma, St Louis, MO) weredissolved in DMSO (1000× stock solution) and further diluted in testmedium immediately before the assay (control cells received the sameamount of DMSO). Nicotinamide (Sigma) was dissolved in water.

CellsPrimary cultures of endothelial cells (human umbilical vein endothelial

cells [HUVECs]) were isolated from umbilical cord veins and grown on1% gelatin-coated flasks in M199 supplemented with 10% FBS, 10%newborn calf serum, 20mMHepes, 2 mM glutamine, 6 U/ml heparin,50 μg/ml endothelial cell growth factor, penicillin, and streptomycin.Cells were used between the third and fifth passages. The 1A9 humanovarian carcinoma cell line and its paclitaxel-resistant variant, 1A9 PTX22[31], were obtained from A. Fojo (National Cancer Institute, NationalInstitutes of Health, Bethesda, MD) and grown in RPMI with 10%FBS and 2 mM glutamine. Stocks of cell lines, authenticated by short-tandem repeat profiling (AmpFlSTR Identifiler Plus PCR Amplifica-tion Kit; Applied Biosystems, Carlsbad, CA), were stored frozen inliquid nitrogen and used within 8 weeks after thawing.

Western Blot Analysis of Tubulin AcetylationEndothelial or tumor cells were plated in a 24-well plate in com-

plete medium. The next day, cells were treated with the indicatedcompound for 4 hours. After washing with phosphate-buffered saline(PBS), cells were lysed in 100 μl of lysis buffer [20 mM Tris-HCl(pH 7.6), 170 mM NaCl, 1 mM EDTA, 1 mM DTT, 1% Triton,5 μM TSA, and Roche Protease Inhibitor Cocktail Tablet] and centri-fuged at 14000 rpm for 10 minutes. Proteins were electrophoresed on a10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gel and transferred to a nitrocellulose membrane (Schleicher& Schuell, Dassel, Germany). After blocking overnight at 4°C withECL advance blocking agent, membranes were probed with anti-acetylated tubulin antibody (1:2000, T6793 clone 6-11B-1) or anti-total tubulin antibody (1:500, T9026), both antibodies from Sigma,in PBS/0.5% Tween, for 1 hour at room temperature, followed byperoxidase-conjugated anti-mouse IgG (Sigma) and Amersham ECLPlus Western Blotting Detection Reagent (GE Healthcare Europe

848 SIRT2 and Taxanes in Cell Motility Bonezzi et al. Neoplasia Vol. 14, No. 9, 2012

GmbH, Milan, Italy). After scanning (Duoscan T1200, AGFA), densi-tometry was analyzed with Gel-Pro Analyzer 3.1 (Media Cybernetics,Silver Spring, MD).

Immunofluorescence Analysis of Tubulin AcetylationEndothelial cells were plated on coverslips and incubated overnight.

After treatment with IDN5390 or paclitaxel (10 nM for 4 hours),cells were fixed with methanol at −20°C for 10 minutes. Coverslipswere washed in PBS and incubated in blocking solution (5% normalgoat serum in PBS) for 30 minutes at room temperature. Incubationwith antibodies against acetylated tubulin (1:1000, T6793, Sigma) inblocking solution was carried out for 1 hour at room temperature. Afterwashing, tetramethylrhodamine isothiocyanate (TRITC)-conjugatedrabbit anti-mouse IgG (1:300, Sigma) in blocking solution wasincubated for 1 hour. Nuclei were counterstained by 4′,6-diamidino-2-phenylindole (DAPI; 100 μg/ml, Sigma). Slides were analyzed by afluorescence microscope (OLYMPUS IX70).

Colocalization of SIRT2 with MicrotubulesEndothelial cells were plated on coverslips and incubated overnight.

After treatment with paclitaxel or IDN5390 (100 nM for 4 hours), cellswere fixed with methanol at −20°C for 10 minutes. Coverslips werethen washed in PBS and incubated in blocking solution (5% normalgoat serum in PBS) for 30 minutes at room temperature. Slides wereincubated with fluorescein isothiocyanate (FITC)–conjugated mono-clonal anti–β-tubulin (1:25, Sigma) and chicken anti-human SIRT2(1:50, [17]), in blocking solution for 1 hour at room temperature, fol-lowed by incubation with TRITC-conjugated rabbit anti-chicken IgG(1:50; Jackson ImmunoResearch, Newmarket, United Kingdom).Slides were analyzed with a Zeiss LSM 510 Meta confocal microscope.

Immunofluorescence Analysis of FOXO3aHUVECs were plated on coverslips and incubated overnight. After

treatment with paclitaxel (10 nM), splitomicin (1.2 mM), or TSA(1 μM) for 4 hours, cells were fixed with 3% paraformaldehyde for20 minutes, washed in PBS, and permeabilized with 0.1% Triton X-100in PBS. After blocking (in 5% goat serum in PBS) for 30 minutes, cellswere stained with antibodies against FOXO3a (1:50, 75D8; Cell Sig-naling Technology, Danvers, MA) in blocking solution for 40 min-utes at 37°C, followed by Alexa Fluor 546 anti-rabbit IgG (1:100;Invitrogen, Carlsbad, CA) for 45 minutes at room temperature. Nucleiwere counterstained by DAPI (100 μg/ml, Sigma). Slides were analyzedwith a Zeiss LSM 510 Meta confocal microscope.

Motility AssayChemotaxis was assessed using Boyden chambers and gelatin-coated

polycarbonate nucleopore filters (8-μm pore size). The supernatant ofNIH-3T3 cells was used as the attractant. Cells were resuspended inDulbecco’s modified Eagle’s medium/0.1% BSA at a concentrationof 1.5 × 106/ml and added to the upper compartment of the chamber.IDN5390, paclitaxel, or histone deacetylase inhibitors at the indicatedconcentration were added to the cells and incubated throughout theassay (4 hours). At the end of incubation, filters were stained withDiff-Quik (Marz-Dade, Dundingen, Switzerland) and the migratedcells were counted in 10 high-power fields.

Proliferation AssayCells (2–3 × 103 cells/well) were plated in a 96-well plate in complete

medium. After 24 hours, IDN5390, paclitaxel, or histone deacetylase

inhibitors at the indicated concentration were added and incubated for4 hours. Cells were then washed and incubated in culture medium foran additional 72 hours. In parallel plates, cells were exposed to thecompounds for the entire duration of the assay (72 hours). Cells werefixed and stained with 0.5% crystal violet in 20% methanol. The stainwas eluted with a 1:1 solution of ethanol/0.1 M sodium citrate, andthe absorbance at 540 nm was measured.

SIRT2 Stable TransfectionHuman wild-type green fluorescent protein (GFP)-SIRT2 and the

catalytically inactive mutant N168A plasmids were prepared as de-scribed [17]. Subconfluent 1A9 human ovarian carcinoma cells weretransfected using Lipofectin Reagent (Invitrogen) following manufac-turer’s protocol. For selection, cells were cultured in medium contain-ing 800 μg/ml of G418 (Invitrogen) and subsequently enriched inGFP-expressing cells by fluoresence-activated cell sorter (FACS). West-ern blot analysis of cell lysates with anti-SIRT2 antibody [17] confirmedthat the transfected cells produced similar amounts of wild-type andmutated proteins that were barely detectable in parental cells.

SIRT2 SilencingHUVECs were transiently transfected with siRNA for human

SIRT2 (mix of Hs_SIRT2_1_HP and Hs_SIRT2_4_HP) or HDAC6(Hs_HDAC6_6_HP, all from Qiagen, Hilden, Germany) by electro-poration. Forty-eight hours later, cells were subjected to a second roundof transfection, and after additional 48 hours, cells were tested for tran-script expression, tubulin acetylation, and motility. Silencing of thetranscripts was verified by real-time reverse transcription–polymerasechain reaction (PCR; TaqMan Gene Expression Assay Hs00247263-m1and Hs00195869-m1; Applied Biosystems), normalizing data on 18srRNA expression (Assay Hs99999901-s1; Applied Biosystems).

ResultsWe have investigated the activity of taxanes on endothelial and tumorcell motility, which represents critical events in the processes of, respec-tively, tumor angiogenesis and metastasis. To focus on events associ-ated with inhibition of cell motility rather than proliferation, we usedexperimental conditions in which taxanes inhibited cell motility with-out affecting proliferation [4,8]. We also used the seco-derivativetaxane IDN5390, which exhibits anti-motility activity and reducedcytotoxicity compared to paclitaxel ([9,10], summarized in Table W1).

Endothelial cells were exposed to paclitaxel or IDN5390 for 4 hours,and tubulin acetylation levels were assessed. Both PTX and IDN5390increased the levels of acetylated tubulin in treated HUVECs, usingeither immunofluorescence (Figure 1A) or Western blot analysis (Fig-ure 1B). Induction of tubulin hyperacetylation paralleled the inhibitionof cell motility but not of proliferation, which occurred at concentra-tions much higher than those required to inhibit motility (particularlyfor IDN5390), even when cells were treated with the taxane for longertimes or cytotoxicity was assessed by quantifying the mitotic index(Figures 1C and W1).

To confirm the separation between the effect of taxanes on cellproliferation and motility also in cancer cells, we used 1A9 PTX22 cells,which are resistant to the antiproliferative activity of taxanes but sen-sitive to their anti-motility effect, as shown by the previous finding thatpaclitaxel and IDN5390 induced a comparable reduction of motility inpaclitaxel-resistant 1A9 PTX22 cells and in the paclitaxel-sensitiveparental 1A9 cells ([8,9], summarized in Table W2). Interestingly,both the highly cytotoxic paclitaxel and the low cytotoxic IDN5390

Figure 1. Effect of paclitaxel and IDN5390 on tubulin acetylationand motility in endothelial and tumor cells. (A) Immunofluorescenceanalysis of acetylated tubulin in HUVECs untreated or treated with10 nM paclitaxel or IDN5390 for 4 hours (scale bar, 20 μm). (B) West-ern blot analysis of acetylated and total tubulin in HUVECs exposedfor 4 hours to increasing concentrations of paclitaxel or IDN5390.(C) Increased tubulin acetylation (black circles), decreased cell motil-ity (open squares), and unaffected cell proliferation (open diamonds,dotted line) after a 4-hour exposure to IDN5390. Tubulin acetylationis expressed as fold increase of untreated cells and was calculatedas the ratio of optical density values of acetylated tubulin versus totaltubulin. Migration and proliferation are expressed as percentage ofcontrol, untreated cells. (D) Western blot analysis of acetylatedand total tubulin comparing the response of sensitive 1A9 andpaclitaxel-resistant 1A9 PTX22 cells to a 4-hour exposure to paclitaxelor IDN5390.

Neoplasia Vol. 14, No. 9, 2012 SIRT2 and Taxanes in Cell Motility Bonezzi et al. 849

induced tubulin acetylation in 4 hours at concentrations sufficient toinhibit cell motility (∼30 nM) in both cell lines irrespective of theirsensibility to inhibition of proliferation (Figures 1D and W2).These findings support the hypothesis that tubulin acetylation might

be implicated in the anti-motility activity of taxanes. To test this pos-sibility, we first tested whether inhibitors of the tubulin deacetylasesHDAC6 and SIRT2 affected the motility of endothelial cells. As re-ported, TSA, a broad spectrum hydroxamic acid inhibitor of class Iand II HDAC (which include HDAC6), inhibited endothelial cellmotility, although inhibitors of sirtuins appeared more effective (Fig-ure 2A). Nicotinamide, sirtinol, and splitomicin inhibited endothelial

cell motility in a dose-dependent manner with IC50 of 47.9 ± 0.7 mM,43.3 ± 0.3 μM, and 1.6 ± 0.2 mM, respectively (not shown andFigure 2B). Inhibition of cell motility by splitomicin significantlycorrelated with induction of tubulin acetylation (P = .01; Figure 2, Band C).

Splitomicin was not as potent as TSA in promoting tubulin acetyla-tion (TSA induced a 26.3 ± 8.4 fold increase of acetylation, not shown),but it was more potent in inhibiting motility (Figure 2A).

Because splitomicin is not selective for SIRT2 and inhibits allsirtuins, we further tested the role of SIRT2 using siRNA knockdownin endothelial cells and SIRT2 stable overexpression in cancer cells.Silencing of SIRT2 (67% reduction measured by real-time PCR)was accompanied by decreased motility and increased tubulin acetyla-tion compared to mock-transfected cells (Figure 2D). Consistent withthe different potency of SIRT inhibitors and TSA in affecting cellmotility and tubulin acetylation, silencing of SIRT2 was more effectivein inhibiting cell motility than silencing of HDAC6 (62% reduction)although the latter was more effective in increasing tubulin acetyla-tion (Figure 2D). To investigate the role of SIRT2 overexpression oncell motility, we stably transfected 1A9 cells to overexpress wild-type(1A9-S2) or catalytically inactive SIRT2 (1A9-mutS2, Figure W3).Overexpression of SIRT2 did not relevantly affect cell proliferation(doubling time was 23.1 ± 3.4 and 21.9 ± 2.3 hours for 1A9-S2 and1A9-mutS2, respectively), morphology, or tubulin organization (Fig-ure W4). However, overexpression of SIRT2 markedly increased cellmotility and decreased the levels of tubulin acetylation (Figure 2E ),further confirming that catalytically active SIRT2 regulates cell motility.Interestingly, 1A9-mutS2 cells presented a slightly increased tubulinacetylation and decreased cell motility compared to parental 1A9 cells(not shown), suggesting that in this system the catalytically inactiveSIRT2 might act as an inhibitor of endogenous SIRT2.

Next, we tested the role of SIRT2 in the anti-motility activity oftaxanes. SIRT2-overexpressing 1A9-S2 cells were less responsive tothe anti-motility activity of both paclitaxel and IDN5390 than control1A9-mutS2 cells (Figure 3, A and B, and Table 1). Moreover, the dif-ference in tubulin acetylation between 1A9-S2 and 1A9-mutS2 cellswas maintained after taxane treatment. SIRT2-overexpressing cellsexposed for 4 hours to paclitaxel or IDN5390 showed lower tubulinacetylation compared to control cells (Figure 3, C and D). Again, thedifferential sensitivity in terms of cell motility was dissociated from theantiproliferative response: The antiproliferative activity of both taxaneswas unaffected by changes in SIRT2 activity in cells treated for either4 or 72 hours (Table 1).

To investigate whether SIRT2 might be a direct target of taxanes,we analyzed the effect of paclitaxel on SIRT2 cellular localization andenzymatic activity. Confocal microscopy analysis confirmed that SIRT2colocalizes with microtubules (Figures 3E and W5). Treatment of cellswith paclitaxel or IDN5390 did not affect SIRT2 colocalization withthe microtubules. Similar results were observed with HDAC6 (notshown). In addition, paclitaxel did not affect SIRT2 deacetylase activityon purified tubulin (Figure W6).

Next, we investigated the possibility that taxanes and SIRT2 in-hibitors might synergize. We tested the activity of a combination ofsplitomicin and paclitaxel on endothelial cell motility. Splitomicinpotentiated the anti-motility activity of paclitaxel (Figure 4A). TheIC50 of paclitaxel was 30.5 ± 6.4 when tested alone and 10.4 ±4.9 in the presence of splitomicin (1.2 mM). To determine whethersplitomicin and paclitaxel interacted synergistically, additively, orantagonistically, we calculated the combination index (CI) [32]. The

850 SIRT2 and Taxanes in Cell Motility Bonezzi et al. Neoplasia Vol. 14, No. 9, 2012

combination resulted synergistic (CI < 1) at the highest concentrationsof the two agents but not at lower concentrations, suggesting that thepotentiating effect of splitomicin was maximal at the most effectiveconcentrations of the compound (Table W3). Under these conditions(4-hour exposure), the combination of the two drugs did not produceincreased antiproliferative activity (Figure 4A), although a longer expo-sure (72 hours) to splitomicin inhibited cell proliferation (not shown),possibly through an inhibitory activity on SIRT1. A similar potentia-tion of paclitaxel anti-motility activity was also induced by the additionof 1 μM TSA (IC50 of paclitaxel in the presence of TSA was 13.3 ±2.7 nM, not shown). We therefore tested whether TSA might fur-ther potentiate the activity of paclitaxel/splitomicin. The combinationpaclitaxel/splitomicin/TSA was even more effective in inhibiting cellmotility (Figure 4B). When exposed for 4 hours to paclitaxel in com-bination with the two histone deacetylase inhibitors, endothelial cellmotility was almost completely abolished and tubulin acetylation washighly increased, whereas proliferation was not affected.

Finally, we investigated the effect of paclitaxel and splitomicin onthe transcription factor FOXO3a, a negative regulator of angiogenesis

Figure 2. Effect of SIRT2 modulation in cell motility. (A) Effect of diffewere exposed to splitomicin (2.4mM), sirtinol (50 μM), nicotinamide (75expressed as percentage of control migration in the absence of inhibitand **P≤ .001 compared to control (analysis of variance followed by Dsquares) and tubulin acetylation (black circles) by splitomicin. For motilifrom six experiments). Tubulin acetylation, evaluated by Western bloFigure 1 (mean of four experiments). (C) Effect of splitomicin on tubuin HUVECs exposed for 4 hours to splitomicin. (D) Migration (expressedtransiently transfected with SIRT2 or HDAC6 siRNA. Results are from on.01 compared to control (Mann-Whitney test). (E) Increased migrationoverexpress SIRT2 (1A9-S2) or catalytically inactive mutated SIRT2 (1A9-mutS2 cells (mean and SE of values from three experiments). *P

and cell motility. Because the nuclear localization of FOXO3a isan indicator of its activity, we investigated whether paclitaxel andsplitomicin affected FOXO3a cellular localization. Immunofluorescenceanalysis of untreated HUVECs showed FOXO3a mainly in the cyto-plasm (Figure 5). After treatment with paclitaxel (for 4 hours, in the sameconditions used in the motility assays), FOXO3a showed a predominantnuclear localization. Interestingly, splitomicin too induced translocationof FOXO3a to the nucleus, although some positivity remained in thecytoplasm. After exposure to a combination of paclitaxel and splitomicin,endothelial cells presented a higher localization of FOXO3a in thenucleus, which further increased when TSA was added to the com-bination (Figure 5).

Altogether these findings indicate that combination of SIRT2 in-hibitors with taxanes increases their anti-motility activity and pointto FOXO3a as a major mediator of this effect.

DiscussionInhibitors of motility can impact on the design of therapies forcell motility–based pathologic processes such as tumor angiogenesis,

rent histone deacetylase inhibitors on endothelial cell motility. CellsmM), or TSA (1 μM) for the duration of the assay (4 hours).Motility isors (mean and SE of values from two to four experiments). *P ≤ .01unnet test). (B) Concentration-dependent inhibition of motility (openty, data are the percentage of control motility (mean and SE of valuest, is expressed as fold increase of untreated cells, calculated as inlin acetylation: Western blot analysis of acetylated and total tubulinas percentage of control) and tubulin acetylation of endothelial cellse experiment representative of two independent transfections. *P≤and reduced tubulin acetylation in 1A9 cells stably transfected to

1A9-mutS2). Cell migration is expressed as percentage of control,< .01 compared to 1A9-mutS2 cells (Mann-Whitney test).

Figure 3. Effect of SIRT2 overexpression on cell sensitivity to taxanes. Inhibition of cell motility (A, B) and induction of tubulin acetylation(C, D) by paclitaxel (A, C) or IDN5390 (B, D) in cells overexpressing SIRT2 (1A9-S2, white symbols) or control, inactive SIRT2 (1A9-mutS2,black symbols). Cell migration is expressed as the percentage of control, untreated cells. (E) No effect of paclitaxel on the colocalization ofSIRT2 with microtubules. HUVECs were treated with paclitaxel (100 nM) for 4 hours, fixed, and processed for double immunofluorescencewith FITC-conjugated anti–β-tubulin and anti-human SIRT2 followed by TRITC-conjugated secondary antibody. Slides were analyzed with aconfocal microscope (scale bar, 10 μm).

Neoplasia Vol. 14, No. 9, 2012 SIRT2 and Taxanes in Cell Motility Bonezzi et al. 851

invasion, and metastasis. In this study, we have demonstrated thatthe modulation of SIRT2 tubulin deacetylase activity affects cellresponse to the anti-motility activity of taxanes.Paclitaxel increases the level of acetylated tubulin in endothelial and

tumor cells. Although other studies reported similar findings, this wasusually observed at high concentrations and/or long exposure times,often associated to a cytotoxic affect [33]. We provide three lines of

evidence indicating that the effect of paclitaxel on tubulin acetylationis associated with inhibition of cell motility rather than proliferation.First, increase in tubulin acetylation is observed in endothelial cellstreated with 10 to 30 nM paclitaxel for 4 hours, conditions associ-ated with inhibition of cell motility but not of proliferation. Second,the same inverse association between tubulin acetylation and cellmotility is also observed with the taxane IDN5390 that has the same

Table 1. Activity of Taxanes on 1A9 Cells Overexpressing Wild-type or Control, Mutated SIRT2.

Paclitaxel

IDN5390

Motility

Prol (72 hours) Prol (4 hours) Motility Prol (72 hours) Prol (4 hours)

1A9-S2

244.3 ± 141.1 4.6 ± 2.5 22.4 ± 14.5 143.5 ± 53.3 33.1 ± 15.3 241.9 ± 128.6 1A9-mutS2 24.5 ± 4.9 4.6 ± 2.1 34.4 ± 21.8 18.8 ± 5.9 41.9 ± 25.0 442.3 ± 49.1

Data are the IC50 values (in nM) and mean and SE of values from three experiments. Proliferation (prol) was assessed by exposing cells to the compounds for 4 hours (as in the motility assay) or 72 hours.

852 SIRT2 and Taxanes in Cell Motility Bonezzi et al. Neoplasia Vol. 14, No. 9, 2012

anti-motility activity as paclitaxel but substantially lower cytotoxic-ity. Third, induction of tubulin acetylation by both paclitaxel andIDN5390 parallels inhibition of cell motility in a tumor model oftubulin-mutated cells resistant to the antiproliferative activity of pacli-taxel. Altogether these findings indicate that tubulin acetylationinduced by a brief exposure to taxanes reflects their anti-motility activ-ity independently from their cytotoxic activity. Preliminary findings

Figure 4. Combination of paclitaxel with tubulin deacetylase in-hibitors. (A) Migration (diamonds) and proliferation (squares) ofendothelial cells exposed for 4 hours to paclitaxel alone (whitesymbols) or in combinationwith 1.2mMsplitomicin (black symbols).In these experiments, splitomicin alone caused 32% inhibition ofmotility and 12% inhibition of proliferation. (B) Effect of a triple com-bination of paclitaxel (PTX, 10 nM), splitomicin (Spl, 1.2 mM),and TSA (1 μM) on endothelial cell motility, tubulin acetylation, andproliferation. Motility and proliferation are expressed as percentageof control activity (untreated cells). Tubulin acetylation is expressedas fold increase of untreated cells and was calculated as the ratio ofoptical density values of acetylated tubulin versus total tubulin. P <.05 compared to control (*), control and paclitaxel alone (°), and con-trol and each inhibitor alone (**; analysis of variance followed byBonferroni test).

Figure 5. Confocal microscopy analysis of the effect of paclitaxeland tubulin deacetylase inhibitors on the localization of FOXO3ain endothelial cells. HUVECs were exposed to paclitaxel (PTX,10 nM), splitomicin (Spl, 1.2 mM), and TSA (1 μM) for 4 hours, pro-cessed, and immunostained for FOXO3a as detailed in Materialsand Methods. Scale bar, 10 μm.

indicate that tubulin acetylation was increased also by the tubulindestabilizer combretastatin (not shown), suggesting that induction oftubulin acetylation might be a feature common to different tubulin-targeting inhibitors of cell motility.

This finding prompted us to investigate whether modulation oftubulin acetylation might affect cell response to the anti-motilityactivity of taxanes, focusing in particular on the two identified tubulindeacetylases HDAC6 and SIRT2. Our data indicate that SIRT2 alsoplays an important role in cell motility, as specifically demonstratedby the findings that overexpression of SIRT2 decreased cell motil-ity, whereas down-regulation of SIRT2 (through siRNA) increasedcell motility. This differs from the finding of Pandithage et al., whoreported an anti-motility activity of SIRT2 in a scratch assay, because ofa de-adhesive activity of SIRT2 that caused cell detachment [34]. Theexperimental conditions used in the present study (three-dimensionalmigration in the Boyden chamber) are quite different and less de-pendent on cell adhesion and crawling on the substrate. Moreover,because SIRT2 modifications, such as phosphorylation, are involvedin regulating SIRT2 ability to control tubulin acetylation and motilityand are regulated in a cell cycle manner, it might be hypothesized thatthe mechanism and outcome of SIRT2 control on motility could varydepending on stage of the cell cycle, hence justifying different resultsobtained in different experimental conditions.

The molecular basis for the activity of HDAC6 and SIRT2 in cellmotility and in modulating the response to taxanes still needs to be

Neoplasia Vol. 14, No. 9, 2012 SIRT2 and Taxanes in Cell Motility Bonezzi et al. 853

clarified. The straight causal association between tubulin acetylationand motility, initially hypothesized, was apparently confuted by thefinding that modulation of HDAC6 activity by TSA or siRNA stronglyaffected tubulin acetylation but modestly modulated cell motility,whereas the opposite was observed with modulation of SIRT2. Theapparent different potency of HDAC6 on tubulin acetylation and cellmotility might depend on additional effects of HDAC6 in this celltype. It should also be considered that, although tubulin acetylationis commonly detected using the 6-11B-1 antibody that recognizeacetylated Lys40, additional acetylation sites on tubulin are present[14], which could be differentially regulated by HDAC6 and SIRT2and modulate motility in these cells.Both taxanes and inhibitors of histone deacetylases have been

reported to affect microtubule dynamic instability. In tumor cells,paclitaxel suppresses microtubule dynamics (reviewed in [3]), althoughin endothelial cells it surprisingly enhances dynamics at low concen-trations compatible with inhibition of cell motility [35]. HDAC6 inmelanoma cells regulates microtubule dynamic instability, possiblythrough the recruitment of kinesin-1 on microtubule and activationof c-Jun N-terminal kinase in the microtubule environment [36]. Ithas been proposed that enhancement of cell adhesion and inhibitionof motility by the histone deacetylase inhibitors TSA and tubacin isbecause of a decrease in microtubule dynamics that impairs the cells’ability to mediate the focal adhesion dynamics required for cell migra-tion [26]. Therefore, changes in microtubule dynamics might play arelevant role to the inhibition of motility by taxanes and inhibitors oftubulin deacetylases.The acetylation status of tubulin and tubulin deacetylases might

affect several molecular pathways involved in the regulation of cellmotility, including microtubule-associated protein (MAP) tau [37],p58 [24], p53, heat shock protein 90 (Hsp90), HOXA10 [38], p53[39], and the nuclear factor-kappaB (NF-κB) family of transcriptionfactor p65 [40] and cortactin [41]. Some of these molecules, such asNF-κB [42] and Hsp90 [43], are also targeted by taxanes and thereforerepresent possible convergence points between the pathways affectedby the two classes of compounds. Another possible candidate targetis FOXO3a, a member of the family of the forkhead transcription fac-tor FOXO, that negatively regulate cell motility [44,45]. We foundthat paclitaxel and splitomicin, alone and in combination, causedFOXO3a translocation to the nucleus, a requirement for its transcrip-tional activity. This finding is in agreement with studies showing thatpaclitaxel causes the translocation of FOXO3a to the nucleus of tumorcells [46]. By deacetylating FOXOs, SIRT1 and SIRT2 repress theirtranscriptional activity [47,48], preventing their inhibitory activityand stimulating cell motility and angiogenesis. In line with these evi-dences, our findings indicate that SIRT inhibitors might cooperatewith paclitaxel to derepress FOXO3a, restoring its anti-motility activity.Inhibitors of histone deacetylases are being developed as a new class

of antineoplastic agents. Although the exact mechanism of their activ-ity is still unclear, their biologic activities possibly depend on theirantiangiogenic effect [47,49,50]. Our finding that SIRT2 regulatesendothelial cell motility points to its potential as a new target for anti-angiogenic therapies.The therapeutic use of histone deacetylase inhibitors has been pro-

posed in combination with chemotherapeutic agents, including taxanes[51,52]. Our findings provide a rationale for the use of SIRT2-targetingcompounds in new approaches of combination therapies, particularlywith taxanes. Our results are in agreement with the described activityof cambinol, an inhibitor of SIRT1 and SIRT2, that enhances the effect

of TSA on tubulin acetylation and potentiates the cytotoxicity of PTXon lung cancer [53]. In addition, the farnesyl transferase inhibitorLonafarnib increases tubulin acetylation and synergizes with paclitaxel,through a mechanism requiring a functional tubulin deacetylase [33],further supporting the connection between modulation of tubulinacetylation and response to taxanes.

In conclusion, this study has demonstrated that inhibition oftubulin deacetylases enhances sensitivity to the anti-motility activityof paclitaxel. Preclinical and clinical findings have already demonstratedthat inhibitors of histone deacetylases represent a suitable addition totaxane-based therapies. This study points to the prospect of develop-ing inhibitors of sirtuins as new agents for combination therapies withtaxanes to treat motility-driven diseases.

AcknowledgmentsWe thank C. Manzotti from Indena S.p.A. for providing IDN5390and constructive discussion, Maria Rosa Bani for helpful commentsto the manuscript, and S. Gastoldi and F. Sangalli for assistance inconfocal microscopy analysis.

References[1] Pasquier E, Honore S, and Braguer D (2006). Microtubule-targeting agents in

angiogenesis: where do we stand? Drug Resist Updat 9, 74–86.[2] Giavazzi R, Bonezzi K, and Taraboletti G (2008). Microtubule targeting agents

and angiogenesis. In Microtubules as Targets for Cancer Therapies. T Fojo (Ed).Humana Press, Totowa, NJ. pp. 519–530.

[3] Jordan MA and Kamath K (2007). How do microtubule-targeted drugs work?An overview. Curr Cancer Drug Targets 7, 730–742.

[4] Belotti D, Vergani V, Drudis T, Borsotti P, Pitelli MR, Viale G, Giavazzi R, andTaraboletti G (1996). The microtubule-affecting drug paclitaxel has antiangiogenicactivity. Clin Cancer Res 2, 1843–1849.

[5] Hayot C, Farinelle S, De Decker R, Decaestecker C, Darro F, Kiss R, and VanDamme M (2002). In vitro pharmacological characterizations of the anti-angiogenicand anti-tumor cell migration properties mediated by microtubule-affecting drugs,with special emphasis on the organization of the actin cytoskeleton. Int J Oncol 21,417–425.

[6] Bijman MN, van Nieuw Amerongen GP, Laurens N, van Hinsbergh VW, andBoven E (2006). Microtubule-targeting agents inhibit angiogenesis at subtoxicconcentrations, a process associated with inhibition of Rac1 and Cdc42 activityand changes in the endothelial cytoskeleton. Mol Cancer Ther 5, 2348–2357.

[7] Ganguly A, Yang H, and Cabral F (2011). Class III β-tubulin counteracts theability of paclitaxel to inhibit cell migration. Oncotarget 2, 368–377.

[8] Belotti D, Rieppi M, Nicoletti MI, Casazza AM, Fojo T, Taraboletti G, andGiavazzi R (1996). Paclitaxel (Taxol®) inhibits motility of paclitaxel-resistanthuman ovarian carcinoma cells. Clin Cancer Res 2, 1725–1730.

[9] Taraboletti G, Micheletti G, Rieppi M, Poli M, Turatto M, Rossi C, Borsotti P,Roccabianca P, Scanziani E, Nicoletti MI, et al. (2002). Antiangiogenic andantitumor activity of IDN 5390, a new taxane derivative. Clin Cancer Res 8,1182–1188.

[10] Taraboletti G, Micheletti G, Giavazzi R, and Riva A (2003). IDN 5390: a newconcept in taxane development. Anticancer Drugs 14, 255–258.

[11] Perdiz D, Mackeh R, Pous C, and Baillet A (2010). The ins and outs of tubu-lin acetylation: more than just a post-translational modification? Cell Signal 23,763–771.

[12] Wloga D and Gaertig J (2010). Post-translational modifications of microtubules.J Cell Sci 123, 3447–3455.

[13] Westermann S and Weber K (2003). Post-translational modifications regulatemicrotubule function. Nat Rev Mol Cell Biol 4, 938–947.

[14] Choudhary C, Kumar C, Gnad F, Nielsen ML, Rehman M, Walther TC,Olsen JV, and Mann M (2009). Lysine acetylation targets protein complexesand co-regulates major cellular functions. Science 325, 834–840.

[15] North BJ and Verdin E (2007). Interphase nucleo-cytoplasmic shuttling andlocalization of SIRT2 during mitosis. PLoS One 2, e784.

[16] Vaquero A, Scher MB, Lee DH, Sutton A, Cheng HL, Alt FW, Serrano L,Sternglanz R, and Reinberg D (2006). SirT2 is a histone deacetylase with pref-erence for histone H4 Lys 16 during mitosis. Genes Dev 20, 1256–1261.

854 SIRT2 and Taxanes in Cell Motility Bonezzi et al. Neoplasia Vol. 14, No. 9, 2012

[17] North BJ, Marshall BL, Borra MT, Denu JM, and Verdin E (2003). Thehuman Sir2 ortholog, SIRT2, is an NAD+-dependent tubulin deacetylase.Mol Cell 11, 437–444.

[18] Dryden SC, Nahhas FA, Nowak JE, Goustin AS, and Tainsky MA (2003). Rolefor human SIRT2 NAD-dependent deacetylase activity in control of mitotic exitin the cell cycle. Mol Cell Biol 23, 3173–3185.

[19] Inoue T, Hiratsuka M, Osaki M, Yamada H, Kishimoto I, Yamaguchi S,Nakano S, Katoh M, Ito H, and Oshimura M (2007). SIRT2, a tubulin de-acetylase, acts to block the entry to chromosome condensation in response tomitotic stress. Oncogene 26, 945–957.

[20] North BJ and Verdin E (2007). Mitotic regulation of SIRT2 by cyclin-dependentkinase 1-dependent phosphorylation. J Biol Chem 282, 19546–19555.

[21] Hubbert C, Guardiola A, Shao R, Kawaguchi Y, Ito A, Nixon A, Yoshida M,Wang XF, and Yao TP (2002). HDAC6 is a microtubule-associated deacetylase.Nature 417, 455–458.

[22] Zhang Y, Li N, Caron C, Matthias G, Hess D, Khochbin S, and Matthias P(2003). HDAC-6 interacts with and deacetylates tubulin and microtubules in vivo.EMBO J 22, 1168–1179.

[23] Saunders LR and Verdin E (2007). Sirtuins: critical regulators at the crossroadsbetween cancer and aging. Oncogene 26, 5489–5504.

[24] Haggarty SJ, Koeller KM, Wong JC, Grozinger CM, and Schreiber SL (2003).Domain-selective small-molecule inhibitor of histone deacetylase 6 (HDAC6)-mediated tubulin deacetylation. Proc Natl Acad Sci USA 100, 4389–4394.

[25] Saji S, Kawakami M, Hayashi S, Yoshida N, Hirose M, Horiguchi S, Itoh A,Funata N, Schreiber SL, Yoshida M, et al. (2005). Significance of HDAC6 regu-lation via estrogen signaling for cell motility and prognosis in estrogen receptor-positive breast cancer. Oncogene 24, 4531–4539.

[26] Tran AD, Marmo TP, Salam AA, Che S, Finkelstein E, Kabarriti R, Xenias HS,Mazitschek R, Hubbert C, Kawaguchi Y, et al. (2007). HDAC6 deacetylationof tubulin modulates dynamics of cellular adhesions. J Cell Sci 120, 1469–1479.

[27] Gao YS, Hubbert CC, Lu J, Lee YS, Lee JY, and Yao TP (2007). Histone de-acetylase 6 regulates growth factor-induced actin remodeling and endocytosis.Mol Cell Biol 27, 8637–8647.

[28] Nolan L, Johnson PW, Ganesan A, Packham G, and Crabb SJ (2008). Willhistone deacetylase inhibitors require combination with other agents to fulfil theirtherapeutic potential? Br J Cancer 99, 689–694.

[29] Mai A and Altucci L (2009). Epi-drugs to fight cancer: from chemistry to cancertreatment, the road ahead. Int J Biochem Cell Biol 41, 199–213.

[30] Bedalov A, Gatbonton T, Irvine WP, Gottschling DE, and Simon JA (2001).Identification of a small molecule inhibitor of Sir2p. Proc Natl Acad Sci USA 98,15113–15118.

[31] Giannakakou P, Sackett DL, Kang YK, Zhan Z, Buters JT, Fojo T, andPoruchynsky MS (1997). Paclitaxel-resistant human ovarian cancer cells havemutant beta-tubulins that exhibit impaired paclitaxel-driven polymerization.J Biol Chem 272, 17118–17125.

[32] Naumova E, Ubezio P, Garofalo A, Borsotti P, Cassis L, Riccardi E, Scanziani E,Eccles SA, Bani MR, and Giavazzi R (2006). The vascular targeting propertyof paclitaxel is enhanced by SU6668, a receptor tyrosine kinase inhibitor, causingapoptosis of endothelial cells and inhibition of angiogenesis. Clin Cancer Res 12,1839–1849.

[33] Marcus AI, Zhou J, O’Brate A, Hamel E, Wong J, Nivens M, El-Naggar A, YaoTP, Khuri FR, and Giannakakou P (2005). The synergistic combination of thefarnesyl transferase inhibitor lonafarnib and paclitaxel enhances tubulin acetylationand requires a functional tubulin deacetylase. Cancer Res 65, 3883–3893.

[34] Pandithage R, Lilischkis R, Harting K, Wolf A, Jedamzik B, Luscher-Firzlaff J,Vervoorts J, Lasonder E, Kremmer E, Knoll B, et al. (2008). The regulationof SIRT2 function by cyclin-dependent kinases affects cell motility. J Cell Biol180, 915–929.

[35] Pasquier E, Honore S, Pourroy B, Jordan MA, Lehmann M, Briand C, andBraguer D (2005). Antiangiogenic concentrations of paclitaxel induce an increase

in microtubule dynamics in endothelial cells but not in cancer cells. Cancer Res65, 2433–2440.

[36] Zilberman Y, Ballestrem C, Carramusa L, Mazitschek R, Khochbin S, andBershadsky A (2009). Regulation of microtubule dynamics by inhibition of thetubulin deacetylase HDAC6. J Cell Sci 122, 3531–3541.

[37] Ding H, Dolan PJ, and Johnson GV (2008). Histone deacetylase 6 interactswith the microtubule-associated protein tau. J Neurochem 106, 2119–2130.

[38] Bae NS, Swanson MJ, Vassilev A, and Howard BH (2004). Human histonedeacetylase SIRT2 interacts with the homeobox transcription factor HOXA10.J Biochem 135, 695–700.

[39] Peck B, Chen CY, Ho KK, Di Fruscia P, Myatt SS, Coombes RC, Fuchter MJ,Hsiao CD, and Lam EW (2010). SIRT inhibitors induce cell death and p53acetylation through targeting both SIRT1 and SIRT2. Mol Cancer Ther 9,844–855.

[40] Rothgiesser KM, Erener S, Waibel S, Luscher B, and Hottiger MO (2010).SIRT2 regulates NF-κB-dependent gene expression through deacetylation ofp65 Lys310. J Cell Sci 123, 4251–4258.

[41] Zhang X, Yuan Z, Zhang Y, Yong S, Salas-Burgos A, Koomen J, Olashaw N,Parsons JT, Yang XJ, Dent SR, et al. (2007). HDAC6 modulates cell motilityby altering the acetylation level of cortactin. Mol Cell 27, 197–213.

[42] Bergstralh DT and Ting JP (2006). Microtubule stabilizing agents: their molecularsignaling consequences and the potential for enhancement by drug combination.Cancer Treat Rev 32, 166–179.

[43] Murtagh J, Lu H, and Schwartz EL (2006). Taxotere-induced inhibition of humanendothelial cell migration is a result of heat shock protein 90 degradation. CancerRes 66, 8192–8199.

[44] Srivastava RK, Unterman TG, and Shankar S (2010). FOXO transcription factorsand VEGF neutralizing antibody enhance antiangiogenic effects of resveratrol.Mol Cell Biochem 337, 201–212.

[45] Zhang H, Pan Y, Zheng L, Choe C, Lindgren B, Jensen ED, Westendorf JJ,Cheng L, and Huang H (2011). FOXO1 inhibits Runx2 transcriptional activityand prostate cancer cell migration and invasion. Cancer Res 71, 3257–3267.

[46] Sunters A, Fernandez de Mattos S, Stahl M, Brosens JJ, Zoumpoulidou G,Saunders CA, Coffer PJ, Medema RH, Coombes RC, and Lam EW (2003).FoxO3a transcriptional regulation of Bim controls apoptosis in paclitaxel-treatedbreast cancer cell lines. J Biol Chem 278, 49795–49805.

[47] Potente M, Ghaeni L, Baldessari D, Mostoslavsky R, Rossig L, Dequiedt F,Haendeler J, Mione M, Dejana E, Alt FW, et al. (2007). SIRT1 controls endo-thelial angiogenic functions during vascular growth. Genes Dev 21, 2644–2658.

[48] Zhao Y, Yang J, Liao W, Liu X, Zhang H, Wang S, Wang D, Feng J, Yu L, andZhu WG (2010). Cytosolic FoxO1 is essential for the induction of autophagyand tumour suppressor activity. Nat Cell Biol 12, 665–675.

[49] Ellis L, Hammers H, and Pili R (2009). Targeting tumor angiogenesis withhistone deacetylase inhibitors. Cancer Lett 280, 145–153.

[50] Shim JS, Matsui Y, Bhat S, Nacev BA, Xu J, Bhang HE, Dhara S, Han KC,Chong CR, Pomper MG, et al. (2010). Effect of nitroxoline on angiogenesis andgrowth of human bladder cancer. J Natl Cancer Inst 102, 1855–1873.

[51] Qian X, LaRochelle WJ, Ara G, Wu F, Petersen KD, Thougaard A, Sehested M,Lichenstein HS, and Jeffers M (2006). Activity of PXD101, a histone de-acetylase inhibitor, in preclinical ovarian cancer studies. Mol Cancer Ther 5,2086–2095.

[52] Ramalingam SS, Maitland ML, Frankel P, Argiris AE, Koczywas M, Gitlitz B,Thomas S, Espinoza-Delgado I, Vokes EE, Gandara DR, et al. (2010).Carboplatin and paclitaxel in combination with either vorinostat or placebofor first-line therapy of advanced non-small-cell lung cancer. J Clin Oncol 28,56–62.

[53] Heltweg B, Gatbonton T, Schuler AD, Posakony J, Li H, Goehle S, Kollipara R,Depinho RA, Gu Y, Simon JA, et al. (2006). Antitumor activity of a small-molecule inhibitor of human silent information regulator 2 enzymes. Cancer Res66, 4368–4377.

Supplementary Materials and Methods

Mitotic Index AnalysisExponentially growing HUVECs were exposed to paclitaxel or

IDN5390 for 4 or 24 hours. Cells were then washed with PBS, fixedwith cold methanol (20 minutes on ice), and stained with DAPI. Atleast 400 nuclei per experimental condition were analyzed by a fluo-rescence microscope (OLYMPUS IX70).

Western Blot Analysis of SIRT2 Expression in Transfected CellsLysates of parental and transfected 1A9 cells were electrophoresed

on a 10% SDS-PAGE gel and transferred to a nitrocellulose mem-brane (Schleicher & Schuell). After blocking overnight at 4°C withECL advance blocking agent, membranes were probed with anti-SIRT2 antibody (1:2000) for 1 hour at room temperature, followedby peroxidase-conjugated anti-chicken IgG (Sigma) and AmershamECL Plus Western Blotting Detection Reagent (GE Healthcare).

In Vitro Tubulin Deacetylation AssayPurified recombinant SIRT2-FLAG (0.284 μg) was incubated

with tubulin (10 μg; Pure, Cytoskeleton, Denver, CO) with or with-out IDN5390 or paclitaxel (10–1000 nM) for 2 hours at room tem-perature, after addition of NAD (1 mM, except in the negativecontrol), essentially as described [1]. Reaction was stopped by addingSDS-PAGE buffer. Samples were analyzed by Western blot withantibodies against acetylated tubulin or total tubulin. Data areexpressed as the ratio between optical density values of acetylatedtubulin and total tubulin.

Combination IndexCI was calculated as described [2] to determine whether the drugs

interacted synergistically, additively, or antagonistically. For eachcombination of drug concentrations (Dsplitomicin X, DpaclitaxelX), producing in combination the effect X, CI was calculated as fol-lows: CI = Dsplitomicin X/ICX, splitomicin + Dpaclitaxel X/ICX,paclitaxel, where ICX, splitomicin and ICX, paclitaxel are the con-centrations of each individual drug that would produce the effect Xif given alone. A confidence band was calculated around each meanusing a t distribution at the 90% probability level. Additivity wasclaimed when the value CI = 1 was inside the confidence band (withthe band included in the 0.8–1.2 interval), synergism when the CIwith its confidence band was <1, and antagonism when the CI withits confidence band was >1.

References[1] North BJ, Marshall BL, Borra MT, Denu JM, and Verdin E (2003). The

human Sir2 ortholog, SIRT2, is an NAD+-dependent tubulin deacetylase.Mol Cell11, 437–444.

[2] Naumova E, Ubezio P, Garofalo A, Borsotti P, Cassis L, Riccardi E, Scanziani E,Eccles SA, Bani MR, and Giavazzi R (2006). The vascular targeting propertyof paclitaxel is enhanced by SU6668, a receptor tyrosine kinase inhibitor, causingapoptosis of endothelial cells and inhibition of angiogenesis. Clin Cancer Res 12,1839–1849.

[3] Taraboletti G, Micheletti G, Rieppi M, Poli M, Turatto M, Rossi C, Borsotti P,Roccabianca P, Scanziani E, Nicoletti MI, et al. (2002). Antiangiogenic and anti-tumor activity of IDN5390, a new taxane derivative. Clin Cancer Res 8, 1182–1188.

[4] Belotti D, Rieppi M, Nicoletti MI, Casazza AM, Fojo T, Taraboletti G, andGiavazzi R (1996). Paclitaxel (Taxol®) inhibits motility of paclitaxel-resistanthuman ovarian carcinoma cells. Clin Cancer Res 2, 1725–1730.

Table W1. Activity of Paclitaxel and IDN5390 on Endothelial Cell Motility and Proliferation.

Figure W1. Mtaxel for 4 or 2motility. Mitonuclei compamotility wasrelevantly inc

MotilityIC50 (nM)

otility and mit4 hours. Motilitic index is thred to total. Irrinhibited by corease mitotic in

Proliferation (4 hours)IC50 (nM)

otic index of HUVECsty is expressed as perce percentage of mitotespective of the timencentrations of paclitdex.

Proliferation (72 hours)IC50 (nM)

IDN5390

43.7 ± 9.5 8341.7 ± 887.8 32.1 ± 7.5 Paclitaxel 29.0 ± 12.5 217.4 ± 117.4 1.8 ± 0.6

The table summarized previous findings on the comparison of the activity of paclitaxel andIDN5390 [3]. Cells were exposed to the compounds for 4 hours. Data are IC50 values (in nM)and mean and SE of values from at least three experiments.

exposed to pacli-entage of controlic (or abnormal)of exposure, cellaxel that did not

Table W2. Motility and Proliferative Response of Parental 1A9 and Paclitaxel-resistant 1A9PTX22 Cells to Paclitaxel and IDN5390.

Paclitaxel

IDN5390

MotilityIC50 (nM)

ProliferationIC50 (nM)

MotilityIC50 (nM)

ProliferationIC50 (nM)

1A9

40 ± 32 79 ± 55 69 ± 51 1905 ± 975 1A9 PTX22 30 ± 23 2650 ± 1050 47 ± 33 10207 ± 169

The table summarized previous findings on the activity of taxanes on responsive (1A9) and resistant(1A9 PTX22) cells [3,4]. Cells were exposed to the compounds for 4 hours. Data are IC50 values(in nM) and mean and SE of values from at least three experiments.

Figure W2. Inductionof tubulin acetylationbypaclitaxel and IDN5390in paclitaxel-sensitive 1A9 and paclitaxel-resistant 1A9 PTX22 cells.Data are the densitometric analysis of Western blot of Figure 1D(representative of four different experiments). Tubulin acetylation isexpressed as fold increase of untreated cells and was calculated asthe ratio of optical density values of acetylated tubulin versus totaltubulin. In 1A9 cells, induction of tubulin acetylation by paclitaxelincreased up to 10 nM and tended to decrease at higher concen-trations. This behavior might reflect a peculiar activity of paclitaxelon these cells, because this was not observed when 1A9 cellswere treated with IDN5390 (Figure 1D) nor when other cell types(HUVECs) were exposed to paclitaxel (Figure 1B). The baseline levelof tubulin acetylation in 1A9 PTX22 cells is slightly higher than in 1A9cells. For this reason, increment in tubulin acetylation—expressedas fold increase of untreated cells—is lower than that of 1A9 cells,although the overall level of taxane-induced acetylation is actuallycomparable or higher. Interestingly, the level of basal tubulin acetyla-tion in 1A9 PTX22 cells is paralleled by a slightly lower ability tomigrate in theBoyden chamber assay, compared to 1A9 cells, furthersupporting the correlation between tubulin acetylation and motility.

Figure W3. Western blot analysis of SIRT2 expression in parental1A9 cells and cells stably transfected to overexpress wild-type(1A9-S2) or catalytically inactiveGFP-SIRT2 (1A9-mutS2). (Upperpanel)Anti-SIRT2 antibody detected a band of approximately 75 kDa,corresponding to GFP-SIRT2. This band was similarly detectablein the transfected cells, whereas endogenous SIRT2 (43 kDa) wasessentially undetectable in parental cells. (Lower panel) Amido blackstaining of the nitrocellulose membrane to show comparable loadingand transfer.

Figure W4. Immunofluorescence analysis of themicrotubule cytoskeleton in parental 1A9 cells or cells transfected to overexpress wild-type(1A9-S2) or catalytically inactive SIRT2 (1A9-mutS2). Transfected SIRT2 proteins are fused to GFP (GFP-SIRT2). Cells where stained with anti-total tubulin antibody (T9026, Sigma). No relevant differences in cell morphology of cytoskeleton organization were observed in transfectedcells. Scale bar, 10 μm.

Figure W5. Confocal microscopy analysis showing that the colocalization of SIRT2 with microtubules was not affected by taxanes.HUVECs were treated with IDN5390 or paclitaxel (100 nM) for 4 hours, fixed, and processed for double immunofluorescence withFITC-conjugated anti–β-tubulin and anti-human SIRT2 followed by TRITC-conjugated secondary antibody. Scale bar, 10 μm.

Figure W6. Paclitaxel and IDN5390 do not inhibit SIRT2 activity. Thetubulin deacetylase activity of SIRT2 was evaluated in vitro, in thepresence of different concentrations of paclitaxel or IDN5390.Control−, samples containing SIRT2 but not NAD were used asa negative control (no SIRT2 activity, high acetylation). Data areexpressed as the ratio between optical density values of acetylatedtubulin and total tubulin.

Table W3. Effect of the Combination of Paclitaxel and Splitomicin on Endothelial CellMigration: CI.

Paclitaxel (nM)

Splitomicin (mM) CI (Mean and SE)

1.04

0.03 >3 5.27 0.14 0.82 ± 0.45 11.85 0.32 2.43 ± 0.93 26.67 0.71 0.60 ± 0.01 40.00 1.07 0.69 ± 0.09 60.00 1.60 0.50 ± 0.10 90.00 2.40 0.36 ± 0.18

HUVEC motility was tested in the presence of the indicated concentrations of paclitaxel and splito-micin, alone or in combination. Data were analyzed as described [2]. Additivity was claimed whenthe value of CI = 1 (with confidence band included in the 0.8–1.2 interval), synergism when CI < 1,and antagonism when CI > 1.