Tumor and Stem Cell Biology Cancer Research Aurora-A Is … · 2010. 6. 1. · Corresponding Authors: Giorgio Stassi, University of Palermo, Via Liborio Giuffre′, 5 Palermo, 90127,

Published OnlineFirst May 11, 2010; DOI: 10.1158/0008-5472.CAN-09-3953

Tumor and Stem Cell Biology
Cancer
Research

Aurora-A Is Essential for the Tumorigenic Capacity andChemoresistance of Colorectal Cancer Stem Cells

Patrizia Cammareri1, Alessandro Scopelliti1,2, Matilde Todaro2, Vincenzo Eterno1,2, Federica Francescangeli5,Mary Pat Moyer6, Antonino Agrusa4, Francesco Dieli3, Ann Zeuner5, and Giorgio Stassi1,2

Abstract

Authors' ARicovero ePavia, Ital3BiopathoUniversityOncology, IsSan Antonio

Note: SuppOnline (http

CorresponLiborio Gi6553211; Fgiorgio.stastere Scient10 Pavia,0382-5920

doi: 10.115

©2010 Am

www.aacr

D

Colorectal cancer stem cells (CR-CSC) are responsible for the generation and maintenance of intestinaltumors and are highly resistant to conventional chemotherapeutic agents. Aurora-A, a serine-threonine kinaseinvolved in mitosis regulation, plays multiple key functions in tumor initiation and progression. We found thatAurora-A is overexpressed in primary colorectal tumor cells, in the CR-CSC fraction, and in stem cell–deriveddifferentiated cells, compared with normal colon tissue. Aurora-A expression was functionally linked tocentrosome amplification in CR-CSC, as indicated by the decrease in cells with multiple centrosomes thatfollowed Aurora-A silencing. Knockdown of Aurora-A resulted in growth inhibition of CR-CSC, alteration ofcell cycle kinetics, and downregulation of the expression levels of antiapoptotic Bcl-2 family members, stronglysensitizing to chemotherapy-induced cell death. Moreover, Aurora-A silencing compromised the ability toform tumor xenografts in immunocompromised mice and reduced the migratory capacity of CR-CSC. Alto-gether, these results indicate that Aurora-A is essential for CR-CSC regeneration and resistance to cytotoxicstimuli and suggest that therapies directed against Aurora-A may effectively target the stem cell population incolorectal cancer. Cancer Res; 70(11); 4655–65. ©2010 AACR.

Introduction

Emerging evidence suggests the existence of a functionalheterogeneity in different cell populations shaping the tumormass. A small subpopulation of cells within the tumor masscalled “cancer stem cells” (CSC) or “cancer-initiating cells”was shown to sustain tumor growth upon injection intoimmunocompromised mice, in contrast to more differentiatedcells that are nontumorigenic (1). CSCs were first identifiedin human hematologic malignancies (2) and later in severalsolid tumors (3–7). Colorectal CSC (CR-CSC) isolated fromhuman tumors can be enriched in vitro through culture innonadherent conditions in the presence of basic fibroblastgrowth factor (bFGF) and epidermal growth factor (EGF),and subsequently propagated as “colospheres.”When injected

ffiliations: 1Cellular and Molecular Oncology, Istituto DiCura a Carattere Scientifico Fondazione Salvatore Maugeri,y; Departments of 2Surgical and Oncological Sciences,logy and Biomedical Methodologies, and 4GENURTO,of Palermo, Palermo, Italy; 5Department of Hematology andtituto Superiore di Sanita, Rome, Italy; and 6INCELL Corporation,, Texas

lementary data for this article are available at Cancer Research://cancerres.aacrjournals.org/).

ding Authors: Giorgio Stassi, University of Palermo, Viauffre′, 5 Palermo, 90127, Italy. Phone: 011-0039-091-ax: 011-0039-091-6553238; E-mail: [email protected],[email protected]; or Ann Zeuner, Istituto Di Ricovero e Cura a Carat-ifico Fondazione Salvatore Maugeri, Via Salvatore Maugeri,27100, Italy. Phone: 011-0039-0382-592065; Fax: 0039-61; Email: [email protected].

8/0008-5472.CAN-09-3953

erican Association for Cancer Research.

journals.org

Researcon April 26cancerres.aacrjournals.org ownloaded from

into immunocompromised mice, sphere-derived CSCs retainthe ability to generate tumor xenografts with morphologicstructure and antigenic profile that reproduce those of theparental human tumors (1, 7).The question remains open of whether CSCs originate

from normal tissue stem cells or from partially differentiatedprogenitors, acquiring unlimited self-renewal potential asa consequence of genetic and/or epigenetic alterations.Because the tumorigenic process requires the accumulationof multiple mutational events, stem cells seem to be the idealcandidates to support this model due to their long life span(8, 9). Genomic instability (GIN) is a key molecular step inneoplastic transformation. The mechanisms underlying GINhave been investigated in embryonic stem cells, in which ithas been shown that the mitotic-spindle checkpoint, al-though functional, does not initiate apoptosis, thus contrib-uting to karyotypic instability (10). However, mechanisms ofGIN in CSC represent a largely unexplored field. It wasrecently shown that embryonic germ cells transplanted inthe testis of severe combined immunodeficient mice couldgenerate a highly metastatic CSC population characterizedby new nonclonal genomic rearrangements (11). GIN canbe broadly classified into microsatellite instability and chro-mosomal instability. Chromosomal instability is observed in∼85% of colon cancers occurring as the consequence ofmutation or amplification of several genes such as hBUB1,BRCA1, BRCA2, and Aurora-A (12).Aurora-A (also known as STK15/BTAK) is a member of a

serine/threonine kinase family involved in mitosis entry,formation of spindle bipolarity, control of centrosome matu-ration, and segregation during mitosis (13). Repression of

4655

h. , 2021. © 2010 American Association for Cancer

http://cancerres.aacrjournals.org/

Cammareri et al.

4656


Aurora-A by RNA interference delays mitotic entrance inseveral cancer cells (14–16), whereas its overexpression leadsto centrosome amplification, cytokinesis inhibition, andaneuploidy (17, 18).The Aurora-A gene maps to chromosome 20q13.2, a region

frequently amplified in several human epithelial tumors(19, 20). Overexpression of Aurora-A is sufficient to transformrodent fibroblast cell lines giving rise to aneuploid cells (18).Moreover, it was recently shown that Aurora-A overexpres-sion in mouse mammary epithelium induced malignanttransformation preceded by centrosome amplification, chro-mosome tetraploidization, and premature sister chromatidsegregation (21). This evidence establishes Aurora-A as anoncogene that causes neoplastic transformation throughthe induction of GIN.Aurora-A is also a key regulatory component of the p53

pathway as it phosphorylates p53 at Ser215 and Ser315,abrogating p53 DNA binding/transactivation activity andleading to p53 proteolytic degradation after MDM2-mediatedubiquitination (22, 23). Consequently, Aurora-A overexpressionabrogates wild-type p53 functions, such as growth arrest and/or apoptosis, entailing a checkpoint-response deregulation andthus contributing to GIN and malignant transformation.Due to its role as a spindle checkpoint regulator, Aurora-A

has been implicated in conferring resistance to chemotherapyin cancer cells. Increased expression of Aurora-A confersresistance to apoptosis induced by chemotherapeutic agents(24, 25), whereas interference with Aurora-A expressionsensitizes tumor cells to antineoplastic treatments (26, 27).In this study, we show that Aurora-A is essential to maintain

CR-CSC tumorigenicity, chemoresistance, and migratory abili-ty, indicating it as an attractive target for the development ofnew therapeutic approaches for colorectal cancer treatment.

Materials and Methods

Tumor specimens and cellsColon cancer specimens were obtained from patients under-

going surgical resection, in accordance with ethical standardsof the Institutional Committee of University of Palermo on hu-man experimentation. Normal colon specimens were obtainedfrom peritumoral tissue. Enzymatic digestion was performedusing collagenase (1.5 mg/mL; Invitrogen) and hyaluronidase(20 μg/mL; Sigma-Aldrich) as previously described (1, 28).The digested tissue was cultured either in adherent conditionsin the presence of DMEM supplemented with 10% fetal bovineserum (FBS) to obtain primary tumor cells, or in nonadherentconditions using ultralow adhesion flasks (Corning, Inc.) in thepresence of serum-free cell medium supplemented with EGF(20 ng/mL) and bFGF (10 ng/mL; both from Sigma-Aldrich)to promote the growth of undifferentiated spheres, which took1 to 2 months. Only sphere cultures validated for a CR-CSCphenotype (CD133+CD29+CK20−) and for the ability to xeno-graft in immunocompromised mice were considered as CR-CSC and were used for subsequent studies. To achieve thein vitro differentiation of CR-CSC into sphere-derived adherentcells (SDAC), spheres were cultured in adherent conditions in

Cancer Res; 70(11) June 1, 2010

Researcon April 26cancerres.aacrjournals.org Downloaded from

the presence of DMEM supplemented with 10% FBS for atleast 7 days. Magnetic cell separation of CD133+/CD133− cellswas performed on fresh tumor digest using microbeads conju-gated with anti-CD133/1 (Miltenyi Biotec).

Western blotCell pellets were resuspended in ice-cold NP40 lysis buffer

and Western blot analyses were performed as previously de-scribed (29). Membranes were incubated with antibodiesagainst Aurora-A (Cell Signaling Technology, Inc.), Mcl-1,Bcl-xL, MDM2, p53 (Santa Cruz Biotechnology, Inc.), Bcl-2(Upstate Biotechnology), and β-Actin (Calbiochem). Densi-tometry analysis was performed by Scion image (Scion). Re-sults were expressed as protein/β-actin absorbance ratio.

Immunohistochemistry and immunofluorescenceImmunohistochemistry was performed on 5-μm-thick

paraffin-embedded sections of human CRC tissue or tumorsxenografts. Sections were exposed to antibodies against cyto-keratin 7, cytokeratin 20 (DAKO Corp.), or isotype-matchedcontrols at appropriate dilutions. Sections were processed aspreviously described (29). For H&E staining, dewaxed sectionswere stained in hematoxylin for 5minutes, washed inwater, andthen exposed for 1 minute to eosin. For Alcian blue staining,sections were stained with Alcian blue 8GX (Sigma-Aldrich)and counterstained with nuclear fast red (Lab Vision, Inc.).Immunofluorescence was performed on 5-μm-thick opti-

mum cutting temperature–embedded cryosections of humannormal and cancer tissues, on primary cancer cells and SDAC(grown on polylisine-coated glass coverslips), and on CR-CSC(cytospun on microscope slides). Tissue sections were fixedin ice-cold acetone, whereas cells were fixed in 2% parafor-maldehyde, permeabilized with 0.2% Triton-X, and blockedwith 0.5% bovine serum albumin. Samples were exposed toantibodies against γ-Tubulin (Sigma-Aldrich), Aurora-A(CST), CD133 (Miltenyi Biotec), CD29 (Calbiochem), cytoker-atin 20 (DAKO), epithelial antigen (Ber-EP4, DAKO), andβ-Catenin (Santa Cruz), or isotype-matched controls. Then,cells were treated with fluorochrome-conjugated anti-mouseor anti-rabbit antibodies (Invitrogen) plus RNase A (200 ng/mL,Sigma-Aldrich) and counterstained using Toto-3 iodide (Invi-trogen). For centrosome counting, at least five randomlyselected regions for slides were analyzed and a minimumof 500 nuclei was counted for each sample. Samples wereanalyzed on a Nikon confocal microscope equipped with×40 oil objectives and EZ-C1 software.

Ploidy determinationCR-CSCs were treated with 0.2 μg/mL colcemid (Sigma-

Aldrich) for 6 hours, dissociated by trypsinization, swollen in75mmol/L KCl, fixedwith 3:1methanol/acetic acid, and droppedonto glass microscope slides. Slides were dried and stained with3% Giemsa. Chromosome numbers were evaluated using aBX60 Olympus microscope under a ×100 oil objective.

Cell cycle analysisDissociated CR-CSCs were fixed overnight in ice-cold 70%

ethanol. Cells were resuspended in PBS containing 50 μg/mL

Cancer Research



Aurora-A in Colorectal Cancer Stem Cells


propidium iodide (Sigma-Aldrich), 3.8 mmol/L sodium citrate,and 10 μg/mL RNase. Samples were analyzed by a FACSCali-bur cytometer and CellQuest Software (Becton Dickinson).

Real-time PCRTotal RNA was obtained using the Rneasy Mini kit (Qiagen

GmbH) and retrotranscribed using the High-Capacity cDNAArchive kit (Applied Biosystems) according to the manufac-turer's instructions. Quantitative PCR analysis was per-formed with the ABI PRISM 7900HT (Applied Biosystem).The probes and primer sets (all from Applied Biosystems)used were as follows: Hs01582072_m1 (Aurora-A) ,Hs00195591_m1 (Snail), Hs00195591_s1 (Twist), and Hu glyc-eraldehyde-3-phosphate dehydrogenase. Relative quantita-tion of gene expression was calculated on triplicatereactions using the ΔΔCt method. Data processing wasperformed using the ABI PRISM SDS software v2.1 (AppliedBiosystems).

Cell viabilityCell viability was evaluated on primary cancer cells, CR-CSC

and SDAC-transduced PLK0.1-shAurora-A, or control shRNAand treated with 50 μg/mL 5-fluorouracil (5-FU) and/or100 μmol/L oxaliplatin (both from Sigma-Aldrich) for 24hours. Viability was assessed using the CellTiter Aqueous As-say kit (Promega Corp.) according to the manufacturer's in-structions. Cell death was analyzed with orange acridine (50μg/mL) and ethidium bromide (1 μg/mL) staining.

Lentiviral transductionGene transfer of primary tumor cells, CR-CSC, and SDAC

was performed with the PLK0.1 lentiviral vector (Sigma-Aldrich) and sequenced-verified short hairpin RNA (shRNA)for Aurora-A or an equivalent scrambled sequence. Lentiviralsupernatants were collected 48 hours after calcium phosphatetransfection of HEK-293T packaging cell line and centrifugedat 141,000 g at 4°C for 90 minutes to obtain a virus-containingpellet that was resuspended in DMEM or stem cell medium.Cells (1 × 105; either primary, CR-CSC, or SDAC) were centri-fuged for 40 minutes at 800 rpm with 1 mL of viral superna-tant (PLK0.1-shAurora-A or control shRNA) in the presence of8 μg/mL polybrene (Sigma-Aldrich). After three cycles of in-fection, cells stably expressing the transduced sequences wereselected through exposure to puromycin (Sigma-Aldrich).

Invasion assayDissociated CR-CSCs (1.5 × 104) were plated on solidified

growth factor–depleted Matrigel (BD) diluted 1:3 in serum-free medium in 8-μm pore size Transwell (Corning). 3T3stem–conditioned medium was used as chemoattractant.Chambers were incubated at 37°C for 48 hours.

Generation of tumor xenograftsFive- to 6-week-old female athymic (nu+/nu+) mice were

obtained from Charles River Laboratories and maintained ac-cording to the institutional guidelines for animal experimenta-tion. CR-CSCs (5 × 105) transduced with PLK0.1-shAurora-Aor control shRNA were suspended in 100 μL of PBS/Matrigel

www.aacrjournals.org


and injected s.c. Tumor mass size was measured weekly forup to 8 weeks according to the formula: (π/6) * larger diam-eter * (smaller diameter)2.

Statistical analysisStatistical significance was determined by ANOVA (one

way or two way) with the Bonferroni posttest using Graph-Pad Prism (GraphPad Software, Inc). Results were consideredstatistically significant when P values were <0.05. Asterisksindicate P < 0.05 (*) and P < 0.01 (**).

Results

Aurora-A is expressed in human colon CSCCR-CSCs were previously characterized by our and other

groups through surface expression of CD133 and absenceof differentiation markers such as CK20 (1, 7, 30). Colon tu-mor specimens (Table 1) were phenotypically characterizedby H&E and Alcian blue staining, and analyzed byimmunohistochemistry for the expression of CK7, CK20,and CD133 (Fig. 1A). CR-CSCs were characterized throughthe CD133+CD29+CK20− phenotype (Fig. 1A) and further val-idated through the capacity to form colon tumor xenograftsin immunodeficient mice (data not shown; ref. 1).RNA analysis performed on five different samples showed

that Aurora-A was barely detectable in normal colon tissue,whereas it was clearly expressed in primary tumor cells,CR-CSC, SDAC, CD133+, and CD133− colon tumor popula-tions (Fig. 1B). Immunoblot analysis revealed a high expres-sion of Aurora-A in tumor specimens compared with normaltissue (Fig. 1C, top), as subsequently confirmed in matchedsamples derived from five different patients (Fig. 1C, bottom).Among samples derived from the same patient, the highestAurora-A expression was displayed by SDAC, possibly reflect-ing their high proliferative rate. Further immunofluorescenceanalysis showed that Aurora-A expression was barely detect-able in normal specimens, whereas a strong signal wasdetected in cancer tissue sections and in all tumor cell popu-lations (Table 1; Fig. 1D). A diffuse cytoplasmic immunore-activity was observed in primary tumor cells, whereas inCR-CSC and SDAC, the expression was also nuclear and moreintense in dividing cells. Centrosome localization of Aurora-Ain G2/M cells also revealed mitotic spindle alterations(Fig. 1D). Altogether, these observations point to a deregula-tion of Aurora-A expression in colon cancer. In addition, thepresence of Aurora-A in CR-CSC and in CD133+ cells suggeststhat it plays an unpredicted role in the CR-CSC compart-ment. Because it was hypothesized that CSC could derivefrom normal stem cells, we investigated the expression ofAurora-A and CD133 both on normal and tumoral tissues.Accordingly (31), in normal colon, Aurora-A was expressedonly in a small population (3.6 ± 0.6%) located in proximityto colon cells with the high proliferative index (Supplemen-tary Fig. S1A, top). The rare CD133+ population presentin the normal colon was mostly negative for Aurora-A(Supplementary Fig. S1A, top), whereas Aurora-A was de-tected in CD133+ and CD133− epithelial cell populations

Cancer Res; 70(11) June 1, 2010 4657



Cammareri et al.

4658


in colon cancer (Supplementary Fig. S1A, bottom). To verifyif Aurora-A was able to transform normal colon epithelialcells, we transduced the NCM460 normal colon cell line withAurora-A and assessed their ability to form colonies in semi-solid culture and xenografts in immunocompromised mice(Supplementary Fig. S1B–E). In contrast to what we observedin other cellular systems (17, 18, 21), Aurora-A expression wasnot capable per se to induce cell transformation, suggesting atissue-specific influence of this gene on growth regulation.Because loss-of-function mutations of p53 may induce a

phenotype analogue to Aurora-A overexpression such as cen-trosome amplification, GIN, and oncogenic transformation(32, 33), we analyzed p53 status in CRC samples to selectspecimens in which the effects of Aurora-A overexpressionwere not masked by those of p53 mutations. We foundwild-type p53 in 8 of the 15 Aurora-A–overexpressing CR-CSClines analyzed (53%; Supplementary Fig. S2A; Table 1) and weselected such samples for further analyses.

Aurora-A overexpression is responsible for centrosomeamplification in CR-CSCAurora-A has been reported to regulate centrosome num-

ber and genomic stability (15, 18). To investigate whether



altered centrosome numbers were found in CR-CSC overex-pressing Aurora-A, we evaluated the number of centrosomesin primary tumor cells, CR-CSC, and SDAC (Fig. 2A, left). Weobserved supernumerary centrosomes in 20 ± 5% of prima-ry colon cancer cells, in 15 ± 6% of CR-CSC, and in 19 ±2% of SDAC (Fig. 2A, right), whereas a regular number ofcentrosomes was observed in normal colon sections (datanot shown).Transduction of the Aurora-A–silencing construct, but not

of a control scrambled shRNA, resulted in the efficientknockdown of Aurora-A expression at the RNA and proteinlevels in all the treated tumor cell populations (Fig. 2B).γ-Tubulin immunostaining of cells expressing Aurora-A shRNArevealed a significant reduction in the number of cells withmore than two centrosomes and a simultaneous increase ofcells with one centrosome (Fig. 2C), indicating that centro-some amplification is a reversible process linked to Aurora-Aexpression levels.Cytogenetic analysis revealed a high percentage of aneu-

ploidy (≥90%) in all cell populations analyzed (primary cells,CR-CSC, and SDAC derived from the same patient; Fig. 2D,left). Freshly isolated CD133+ cells showed a percentage ofaneuploidy comparable with that observed in CR-CSC, thus

Table 1. Case description

Patient
Age/sex Site Dukes
h. , 2021. © 20

Aurora-A protein expression*

Cancer Re

10 American Association for Cancer

P53†

P1
56/F Left D + Mut P2 62/F Recto-sigmoid A + Mut P3 65/M Right D + Wt P4 61/M Right C − — P5 59/F Left D − — P6 52/F Left C + Wt P7 55/M Recto-sigmoid D − — P8 62/M σ D + Wt P9 70/F Rectum C + Wt P10 64/M Left D − — P11 67/F Recto-sigmoid A − — P12 76/F σ C + Mut P13 57/M Recto-sigmoid C + Mut P14 67/M Recto-sigmoid C + Wt P15 61/M σ C − — P16 52/M Rectum B − — P17 72/M σ C + Wt P18 55/F Left C + Mut P19 68/M Recto-sigmoid B + Wt P20 59/M Rectum C − — P21 58/F Left D − — P22 68/M Recto-sigmoid B + Mut P23 57/M σ B + Wt P24 52/M Rectum B − — P25 59/M Rectum C + Mut
*+ indicates immunohistochemical expression of Aurora-A in at least 20% of tumor cells.†Mut, Mutated p53. Wt, wild-type p53.

search




excluding any effect of in vitro long-term culture. (Fig. 2D,left). Then, we investigated whether Aurora-A silencing wasable to affect the number of chromosomes present in hyper-diploid colon tumor cells. However, we did not find a signif-



icant decrease in the number of cells with chromosomalalterations following Aurora-A knockdown. (Fig. 2D, right).These observations suggest that although Aurora-A over-expression may be involved in the initial development of

Figure 1. Aurora-A is expressed in human CRC specimens and tumor cell subpopulations. A, immunohistochemical staining with H&E, CK7, and CK20,and Alcian Blue and immunofluorescence analysis of CD133 on colorectal cancer tissue sections (top) and CD133, CD29, and CK20 on CR-CSC (bottom).Nuclei of immunohistochemical sections were revealed by hematoxylin, whereas nuclei of immunofluorescence analyses were counterstained withToto-3 (pseudocolored blue). B, mRNA expression of Aurora-A on peritumoral tissue (normal), cultured primary colon tumor cells (primary), CR-CSC,SDAC, CD133+, and CD133− cells freshly purified from cancer tissue. HeLa cells were used as positive control. Columns, mean of five independentexperiments performed with five matched sets of samples, each derived from a different patient; bars, SD. C, immunoblot analysis of Aurora-A on samplesas above. Columns, mean of five independent experiments, expressed as Aurora-A/β-Actin absorbance ratio; bars, SD. D, confocal microscopyanalysis of Aurora-A on cryostat sections of normal (normal) and tumor (parental) colon tissue, cultured primary colon tumor cells (primary), CR-CSC, andSDAC. Arrowheads, supernumerary centrosomes. One representative experiment of 20 performed with cells derived from different patients is shown.

Cancer Res; 70(11) June 1, 2010 4659



Cammareri et al.

4660


Figure 2. Aurora-A controls centrosome numbers in CR-CSC. A, immunofluorescence analysis of cultured primary colon tumor cells (primary), CR-CSC,and SDAC stained with anti–γ-Tubulin antibody. Arrowheads, supernumerary centrosomes. One representative experiment of five performed is shown(left). Percentage of cells with one (1 centr), two (2 centr), or more than two centrosomes (>2 centr) assessed in five different cultures of primary colon tumorcells (primary), CR-CSC, and SDAC (right). B, mRNA (left) and protein (right) levels of Aurora-A on primary colon tumor cells (primary), CR-CSC andSDAC transduced with control shRNA (SCR), or Aurora-A shRNA (sh-A). Columns, mean of the results obtained from five independent experimentsperformed with cells from different patients; bars, SD. C, percentage of cells with one (1 centr), two (2 centr), or more than two centrosomes (>2 centr)in primary colon tumor cells (primary), CR-CSC, and SDAC transduced as in B. *, P < 0.05. D, representative Giemsa-stained metaphases obtained fromcultured primary colon tumor cells (primary), CR-CSC, CD133+ freshly purified cells (CD133+), and SDAC (left). Quantification of chromosome numbers,expressed as hyperdiploid (hyper), diploid (diploid), and hypodiploid (hypod) present in five samples of primary colon tumor cells (primary), CR-CSC, andSDAC transduced as in B (right).

Cancer Res; 70(11) June 1, 2010 Cancer Research

Research. on April 26, 2021. © 2010 American Association for Cancercancerres.aacrjournals.org Downloaded from




aneuploidy, once chromosome alterations have been estab-lished, it becomes dispensable for their maintenance.

Knockdown of Aurora-A alters the cell cycle kineticsof CR-CSC and inhibits their growth in vitroTargeting Aurora-A by means of chemical inhibitors or

RNA interference has been shown to reduce tumor cellproliferation in several cancer types (14, 34). We found thatAurora-A silencing effectively inhibited the growth of all thetumor cell populations examined, indicating that Aurora-Ais essential for the proliferation of colon cancer cells includingthe CR-CSC fraction (Fig. 3A). Furthermore, we observed amodest increase Aurora-A–silenced CR-CSC in G2/M (from12–19%), a simultaneous decrease of cells in G0/G1 (from 69–53%), and an increase of the sub-G1 population (from 2–12%;Fig. 3B and C). These results suggest that Aurora-A silencingcan induce G2 arrest in part of CR-CSC and apoptosis inanother part of the population, possibly reflecting a hetero-geneity in p53 levels within the same CR-CSC population.In support of this hypothesis, we found that the expression

levels of p53 were increased in CR-CSC transduced withAurora-A shRNA compared with control shRNA (Supplemen-tary Fig. S2B). Additional flow cytometry analysis revealedthat CR-CSC (which are 100% positive for the epithelial antigenBer-EP4) show a heterogeneous expression of p53 and suchexpression was increased by 55 ± 13.5% following Aurora-Aknockdown (Supplementary Fig. S2C).

Aurora-A silencing decreases antiapoptotic proteinexpression and sensitizes CR-CSCs to chemotherapy-induced deathWehave previously reported thatCD133+ CR-CSCs arewidely

resistant to chemotherapeutic drugs (1). Because Aurora-Aoverexpression correlates to chemotherapy resistance in cancercell lines (22, 35), we investigated whether Aurora-A silencing



was able to overcome CR-CSC resistance to chemotherapy.Aurora-A silencing resulted in a slight decrease of basal cellviability in the three populations examined. As expected, con-trol CR-CSCs were highly resistant to chemotherapeuticdrugs, whereas the viability of primary cells and SDAC wasmoderately affected. Importantly, chemotherapy treatment ofAurora-A–silenced CR-CSC resulted in a significant increasein cell death compared with control CR-CSC, with a 75% reduc-tion in cell viability obtained with the combination of bothdrugs (Fig. 4A and B). A similar sensitization to drug-induceddeath by Aurora-A knockdown was observed in primary tumorcells and SDAC (Fig. 4A). To investigate whether resistant cellswould emerge from chemotherapy treatment despite Aurora-Aknockdown as a possible consequence of increased p53 expres-sion, we exposed CR-CSC that survived chemotherapy to asecond round of drug treatment. Although Aurora-A–silencedCR-CSC died after 48 hours of drug restimulation, CR-CSCtransduced with control shRNA showed only a modest effecton cell viability (Supplementary Fig. S2D), indicating thatAurora-A–mediated p53 inhibition is essential for CR-CSCchemoresistance. Altogether, these results show that Aurora-A silencing enhances the combined cytotoxic effect of 5-FUand oxaliplatin, thus representing a valid approach to sensitizeCR-CSC to conventional treatment. We found a strong de-crease in MDM2, Mcl-1, and Bcl-2 expression upon Aurora-Asilencing, whereas Bcl-xL levels were moderately decreased(Fig. 4C). These findings suggest that Aurora-A overexpressionmay confer resistance to apoptosis both by promoting MDM2-dependent p53 degradation and by upregulating survivalfactors involved in the mitochondrial apoptotic pathway.

Aurora-A regulates the migratory ability and thetumorigenic capacity of CR-CSCIt has been hypothesized that CSCs are the only cells able to

migrate from primary tumors to generate distal lesions (36).Because Aurora-A overexpression was shown to enhance the

Figure 3. Aurora-A knockdown inhibits CR-CSC growth and cell cycle kinetics. A, proliferation of primary colon tumor cells (primary), CR-CSC and SDACtransduced with control shRNA (SCR), or Aurora-A shRNA (sh-A). Points, mean SD of five different experiments; bars, SD. The difference between thegrowth of control and Aurora-A–silenced CR-CSC was statistically significant with P < 0.05. B, percentage of CR-CSC transduced as in A in different phasesof cell cycle, as determined by propidium iodide staining. The results shown represent the mean of five different experiments. C, representative cell cycleprofile of CR-CSC transduced as above. Insets, an acridine orange/ethidium bromide staining of transduced CR-CSC (green, viable cells; red, dead cells).

Cancer Res; 70(11) June 1, 2010 4661



Cammareri et al.

4662


migratory capacity of cancer cell lines (37), we investigated itsinvolvement in cell motility of CR-CSC. Aurora-A silencing re-duced the migratory capacity of CR-CSC after only 24 hours(Fig. 5A). Quantitative real-time PCR analysis showed thatAurora-A knockdown reduced the expression of Twist and Snail(Fig. 5B), two genes involved in epithelial-mesenchimal transition,suggesting that Aurora-A may modulate CR-CSC invasivenessthrough the regulation of key factors responsible for epithelial-mesenchimal transition. Finally, we investigated the role ofAurora-A in the production of tumor xenografts by CR-CSC. Nudemice were s.c. inoculated with CR-CSC transduced with controlshRNA or Aurora-A shRNA. Control samples generated a palpabletumor within 2 weeks from injection in five of five mice. Tumorxenografts reached an average volume of 250mm3 in 8weeks andreproduced the same histopathology of the human colorectaltumors of origin (Fig. 5C and D, bottom). Notably, no tumorformation was observed after the injection of equal numbers ofCR-CSC transduced with Aurora-A shRNA (Fig. 5C and D), indi-cating that Aurora-A is essential for tumor formation by CR-CSC.

Discussion

Aurora kinases are critical for cell division and have beenclosely linked to tumorigenesis and cancer susceptibility.



Aurora-A is overexpressed in colorectal tumors (17, 38–40)and is strongly related to chromosomal instability. We havefound Aurora-A overexpression in 60% of colon tumor speci-mens examined, a percentage compatible with that observedpreviously (17, 39). Then, we have analyzed Aurora-A expres-sion in different cell populations within colon cancer andfound that Aurora-A was overexpressed in primary tumorcells, in CSC, in CD133+/- fractions, and in SDAC comparedwith nontumor tissue.Although Aurora-A did not show a transforming ability

per se in normal colon cells, the presence of an aberrantAurora-A expression in the CR-CSC fraction suggests its con-tribution to neoplastic transformation by facilitating centro-some amplification and GIN with a subsequent occurrence ofadditional oncogenic mutations. In line with this hypothesis,several studies have shown a correlation between Aurora-Aoverexpression and aneuploidy (15, 18). In particular, arecent analysis of Aurora-A expression in chromosomal-unstable versus microsatellite-unstable sporadic colorectalcancers showed that higher Aurora-A gene copy numberswere found only in chromosomal-unstable colorectal carci-nomas (41). Accordingly, we found that 90% of Aurora-A–overexpressing colon tumor specimens displayed a markedaneuploidy that could not be reverted by Aurora-A silencing.

Figure 4. Aurora-A knockdown sensitizes CR-CSC to chemotherapy-induced cell death. A, cell viability of primary colon tumor cells (primary), CR-CSCand SDAC transduced with control shRNA (SCR), or Aurora-A shRNA (sh-A) and treated for 24 h with 50 μg/mL 5-FU, 100 μmol/L oxaliplatin, or acombination of both drugs. Columns, mean of five independent experiments; bars, SD. **, P < 0.01 between control and Aurora-A–silenced CR-CSC.B, immunofluorescence staining with orange acridine/ethidium bromide of CR-CSC treated as in A. One representative experiment of five performedwith cells from different patients is shown. C, immunoblot analysis (left) and relative band quantification (right) of MDM2, Bcl-xL, Mcl-1, and Bcl-2 onCR-CSC treated as in A. Columns, mean of five independent experiments performed with cells from different patients; bars, SD.

Cancer Research





However, we have observed a dynamic balance betweenAurora-A expression and centrosome amplification in coloncancer cell populations, as Aurora-A silencing was able torestore normal centrosome numbers. According to studiesby Katayama and colleagues (22), we found that Aurora-Aregulates the expression levels of MDM2 in CR-CSC. IncreasedAurora-A levels in CR-CSC could therefore contrast with theG1 checkpoint arrest by inactivating p53, allowing aneuploidcells to proceed during the S-phase of the cell cycle.Our observation that Aurora-A controls MDM2 expression

also supports a general model in which Aurora-A levels con-tribute to determine p53 status in cells with wild-type p53.An initial overexpression of Aurora-A would determine p53activation as a consequence of GIN, although later, its con-stitutive overexpression would promote p53 degradation,thus creating a favorable environment for the establishmentof additional genetic alterations. In accordance, Aurora-Asilencing provoked an increased p53 stabilization in CR-CSC.Our finding that Aurora-A is expressed in the stem cell frac-

tion of colon tumors has important therapeutic implications.The CSC population is particularly resistant to conventionaltherapies due to high expression of ATP-binding cassette pro-teins, to hyperactive DNA repair pathways, or to decreasedproduction of reactive oxygen species (42–44). Additionally,CR-CSCs have been shown to possess an autocrine and/or



paracrine mechanism of interleukin-4 production that conferscell death resistance through the upregulation of antiapopto-tic molecules (1). Our results show that Aurora-A silencingdecreases the expression of antiapoptotic Bcl-2 family mem-bers (Bcl-2, Mcl-1, and, to a minor extent, Bcl-xL) in CR-CSC.Accordingly, we found that Aurora-A silencing sensitized CR-CSC to chemotherapy-induced cytotoxicity, at least in part bystabilizing p53 levels. Although the molecular mechanismsthat link Aurora-A to antiapoptotic protein expression andchemoresistance remain to be fully clarified, Aurora-A thera-peutic inhibition may increase the efficacy of chemotherapeu-tic drugs toward the stem cell population, possibly preventingtumor relapse and metastatic dissemination.Overexpression of Aurora-A has been correlated with the

occurrence of metastasis, whereas Aurora-A knockdowninhibited the proliferation and invasion of human esophagealcancer cells (16, 45). Our finding that Aurora-A knockdowncompromised the migratory capacity of CR-CSC in vitromay have important implications given that CSC may playa central role in metastasis formation (36). Besides suggest-ing a nonmitotic role of Aurora-A in CSC, these observationssuggest that pharmacologic Aurora-A inhibition may reducethe incidence of metastasis in cancer patients.Finally, we found that Aurora-A silencing completely

abrogates the ability of CR-CSC to form tumor xenografts,

Figure 5. Aurora-A silencing inhibits thein vitro migratory capacity and in vivotumorigenicity of CR-CSC. A, migrationassay of CR-CSC transduced with controlshRNA (SCR) or Aurora-A shRNA (sh-A).Columns, mean of five independentexperiments performed with cells fromdifferent patients; bars, SD. B, mRNAexpression of Twist and Snail of CR-CSCtreated as in A. Columns, mean of fiveindependent experiments performed withcells from different patients; bars, SD.C, in vivo growth of CR-CSC transduced asin A expressed as the size of subcutaneoustumor xenografts derived from the injectionof 5 × 105 colosphere-derived cells. Points,mean obtained with five mice for each group(SCR and sh-A); bars, SD. D, representativeanalysis of tumor xenografts obtained as inC, showing a xenograft-bearing nude mouse8 wk after CR-CSC injection (top) and H&Eand CK20 staining of xenograft-derivedparaffin-embedded sections.

Cancer Res; 70(11) June 1, 2010 4663



Cammareri et al.

4664


highlighting its essential role in supporting CR-CSC tumori-genicity in vivo.Although the overexpression of Aurora-A is not restricted

to CR-CSC, our results support its involvement in the exclu-sive properties of CR-CSC such as resistance to convention-al drugs and tumorigenic potential. Recently, Cicalese andcolleagues (46) showed that p53 stabilization in tumormammospheres switches stem cell division from symmetri-cal to asymmetric, representing a potential mechanism oftumor suppression. By regulating p53 stability, Aurora-Acould indirectly contribute to tumor growth in vivo throughan increase of CSC self-renewal. Preclinical studies indicatethat Aurora-A inhibitors lead to tumor stabilization and/orregression (47, 48), in accordance with a possible role ofAurora-A in driving colon cancer progression from the stemcell compartment. Future clinical studies will definitelyclarify the therapeutic potential of Aurora-A inhibition incounteracting progression and metastasis of colon tumors.



Disclosure of Potential Conflicts of Interest

G. Stassi: commercial research support, Apogenix GmbH; honoraria fromspeakers bureau, American Society of Clinical Oncology. The other authorsdisclosed no potential conflicts of interest.

Acknowledgments

We thank APOGENIX GmbH for the materials and criticism.

Grant Support

Programmi di Ricerca Scientifica di Rilevante Interesse Nazionale 2007 prot.2007TE8NFY (G. Stassi), Programma Straordinario di Ricerca Oncologica MoH(G. Stassi), and Associazione Italiana per la Ricerca sul Cancro (G. Stassi,M. Todaro, and A. Zeuner).

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

Received 10/27/2009; revised 03/03/2010; accepted 03/23/2010; publishedOnlineFirst 05/11/2010.

References
1. Todaro M, Alea MP, Di Stefano AB, et al. Colon cancer stem cells
dictate tumor growth and resist cell death by production of interleukin-4.Cell Stem Cell 2007;1:389–402.

2. Bonnet D, Dick JE. Human acute myeloid leukemia is organized as ahierarchy that originates from a primitive hematopoietic cell. Nat Med1997;3:730–7.

3. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF.Prospective identification of tumorigenic breast cancer cells. ProcNatl Acad Sci U S A 2003;100:3983–8.

4. Singh SK, Clarke ID, Terasaki M, et al. Identification of a cancer stemcell in human brain tumors. Cancer Res 2003;63:5821–8.

5. Clarke MF, Dick JE, Dirks PB, et al. Cancer stem cells-perspectiveson current status and future directions: AACR workshop on cancerstem cells. Cancer Res 2006;66:9339–44.

6. Dalerba P, Dylla SJ, Park IK, et al. Phenotypic characterization ofhuman colorectal cancer stem cells. Proc Natl Acad Sci U S A 2007;104:10158–63.

7. Ricci-Vitiani L, Lombardi DG, Pilozzi E, et al. Identification andexpansion of human colon-cancer-initiating cells. Nature 2007;445:111–5.

8. Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem cells, cancer,and cancer stem cells. Nature 2001;414:105–11.

9. Hiyama E, Hiyama K. Telomere and telomerase in stem cells. Br JCancer 2007;96:1020–4.

10. Mantel C, Guo Y, Lee MR, et al. Checkpoint-apoptosis uncoupling inhuman and mouse embryonic stem cells: a source of karyotpic insta-bility. Blood 2007;109:4518–27.

11. Conway AE, Lindgren A, Galic Z, et al. A pluripotency and self re-newal program controls the expansion of genetically unstable cancerstem cells in pluripotent stem cell-derived tumors. Stem Cells 2009;27:18–28.

12. Grady WM, Markowitz S. Genomic instability and colorectal cancer.Curr Opin Gastroenterol 2000;16:62–7.

13. Fu J, Bian M, Jiang Q, Zhang C. Roles of Aurora kinases in mitosisand tumorigenesis. Mol Cancer Res 2007;5:1–10.

14. Hata T, Furukawa T, Sunamura M, et al. RNA interference targetingaurora kinase a suppresses tumor growth and enhances the taxanechemosensitivity in human pancreatic cancer cells. Cancer Res2005;65:2899–905.

15. Lentini L, Amato A, Schillaci T, Di Leonardo A. Simultaneous Aurora-A/STK15 overexpression and centrosome amplification inducechromosomal instability in tumour cells with a MIN phenotype.BMC Cancer 2007;7:212.

16. Wang X, Dong L, Xie J, Tong T, Zhan Q. Stable knockdown of Aurora-A

by vector-based RNA interference in human esophageal squamouscell carcinoma cell line inhibits tumor cell proliferation, invasion andenhances apoptosis. Cancer Biol Ther 2009;8.

17. Bischoff JR, Anderson L, Zhu Y, et al. A homologue of Drosophilaaurora kinase is oncogenic and amplified in human colorectalcancers. EMBO J 1998;17:3052–65.

18. Zhou H, Kuang J, Zhong L, et al. Tumour amplified kinase STK15/BTAK induces centrosome amplification, aneuploidy and transfor-mation. Nat Genet 1998;20:189–93.

19. Sen S, Zhou H, Zhang RD, et al. Amplification/overexpression of amitotic kinase gene in human bladder cancer. J Natl Cancer Inst2002;94:1320–9.

20. Aust DE, Muders M, Kohler A, et al. Prognostic relevance of 20q13gains in sporadic colorectal cancers: a FISH analysis. Scand JGastroenterol 2004;39:766–72.

21. Wang X, Zhou YX, Qiao W, et al. Overexpression of aurora kinase Ain mouse mammary epithelium induces genetic instability precedingmammary tumor formation. Oncogene 2006;25:7148–58.

22. Katayama H, Sasai K, Kawai H, et al. Phosphorylation by aurorakinase A induces Mdm2-mediated destabilization and inhibition ofp53. Nat Genet 2004;36:55–62.

23. Liu Q, Kaneko S, Yang L, et al. Aurora-A abrogation of p53 DNAbinding and transactivation activity by phosphorylation of serine215. J Biol Chem 2004;279:52175–82.

24. Anand S, Penrhyn-Lowe S, Venkitaraman AR. AURORA-A amplifica-tion overrides the mitotic spindle assembly checkpoint, inducingresistance to Taxol. Cancer Cell 2003;3:51–62.

25. Scharer CD, Laycock N, Osunkoya AO, et al. Aurora kinase inhibitorssynergize with paclitaxel to induce apoptosis in ovarian cancer cells.J Transl Med 2008;6:79.

26. Sun C, Chan F, Briassouli P, Linardopoulos S. Aurora kinase inhibi-tion downregulates NF-κB and sensitises tumour cells to chemother-apeutic agents. Biochem Biophys Res Commun 2007;352:220–5.

27. VanderPorten EC, Taverna P, Hogan JN, Ballinger MD, FlanaganWM,Fucini RV. The Aurora kinase inhibitor SNS-314 shows broadtherapeutic potential with chemotherapeutics and synergy withmicrotubule-targeted agents in a colon carcinoma model. MolCancer Ther 2009;8:930–9.

28. Cammareri P, Lombardo Y, Francipane MG, Bonventre S, Todaro M,Stassi G. Isolation and culture of colon cancer stem cells. MethodsCell Biol 2008;86:311–24.

29. Stassi G, Todaro M, Zerilli M, et al. Thyroid cancer resistance tochemotherapeutic drugs via autocrine production of interleukin-4and interleukin-10. Cancer Res 2003;63:6784–90.

Cancer Research





30. O'Brien CA, Pollett A, Gallinger S, Dick JE. A human colon cancercell capable of initiating tumour growth in immunodeficient mice.Nature 2007;445:106–10.

31. Burum-Auensen E, De Angelis PM, Schjolberg AR, Kravik KL,Aure M, Clausen OP. Subcellular localization of the spindle proteinsAurora A, Mad2, and BUBR1 assessed by immunohistochemistry.J Histochem Cytochem 2007;55:477–86.

32. Bischoff FZ, Yim SO, Pathak S, et al. Spontaneous abnormalities innormal fibroblasts from patients with Li-Fraumeni cancer syndrome:aneuploidy and immortalization. Cancer Res 1990;50:7979–84.

33. Fukasawa K, Choi T, Kuriyama R, Rulong S, Vande Woude GF.Abnormal centrosome amplification in the absence of p53. Science1996;271:1744–7.

34. Shimomura T, Hasako S, Nakatsuru Y, et al. MK-5108, a highlyselective Aurora-A kinase inhibitor, shows antitumor activity aloneand in combination with docetaxel. Mol Cancer Ther 2010;9:157–66.

35. Yang H, He L, Kruk P, Nicosia SV, Cheng JQ. Aurora-A inducescell survival and chemoresistance by activation of Akt through ap53-dependent manner in ovarian cancer cells. Int J Cancer2006;119:2304–12.

36. Li F, Tiede B, Massague J, Kang Y. Beyond tumorigenesis: cancerstem cells in metastasis. Cell Res 2007;17:3–14.

37. Guan Z, Wang XR, Zhu XF, et al. Aurora-A, a negative prognosticmarker, increases migration and decreases radiosensitivity in cancercells. Cancer Res 2007;67:10436–44.

38. Nishida N, Nagasaka T, Kashiwagi K, Boland CR, Goel A. High copyamplification of the Aurora-A gene is associated with chromosomalinstability phenotype in human colorectal cancers. Cancer Biol Ther2007;6:525–33.

39. Lam AK, Ong K, Ho YH. Aurora kinase expression in colorectaladenocarcinoma: correlations with clinicopathological features, p16expression, and telomerase activity. Hum Pathol 2008;39:599–604.



40. Baba Y, Nosho K, Shima K, et al. Aurora-A expression is indepen-dently associated with chromosomal instability in colorectal cancer.Neoplasia 2009;11:418–25.

41. Lassmann S, Danciu M, Muller M, et al. Aurora A is differentiallyexpressed and regulated in chromosomal and microsatellite instablesporadic colorectal cancers. Mod Pathol 2009;22:1385–97.

42. Dean M, Fojo T, Bates S. Tumour stem cells and drug resistance. NatRev Cancer 2005;5:275–84.

43. Bao S, Wu Q, McLendon RE, et al. Glioma stem cells promote radio-resistance by preferential activation of the DNA damage response.Nature 2006;444:756–60.

44. Diehn M, Clarke MF. Cancer stem cells and radiotherapy: newinsights into tumor radioresistance. J Natl Cancer Inst 2006;98:1755–7.

45. Wang XX, Liu R, Jin SQ, Fan FY, Zhan QM. Overexpression of Aurora-Akinase promotes tumor cell proliferation and inhibits apoptosis inesophageal squamous cell carcinoma cell line. Cell Res 2006;16:356–66.

46. Cicalese A, Bonizzi G, Pasi CE, et al. The tumor suppressor p53regulates polarity of self-renewing divisions in mammary stem cells.Cell 2009;138:1083–95.

47. Tentler JJ, Bradshaw-Pierce EL, Serkova NJ, et al. ENMD-2076exerts antiangiogenic and antiproliferative activity against humancolorectal cancer (CRC) xenograft models. Presented at 20thEORTC-NCI-AACR symposium, Molecular Targets and CancerTherapeutics; 2001 Oct 21–24, 2008; Geneva, Switzerland.

48. Bastos BR. Diamond J, Hansen R, et al. An open-label, dose es-calation, safety, and pharmacokinetic study of ENMD-2076 admin-istered orally to patients with advanced cancer. Presented atASCO Annual Meeting 2009 (abstract 3520). Available from: http://www.asco.org/ASCOv2/Meetings/Abstracts?&vmview=abst_detail_view&confID=65&abstractID=34778.

Cancer Res; 70(11) June 1, 2010 4665



2010;70:4655-4665. Published OnlineFirst May 11, 2010.Cancer Res Patrizia Cammareri, Alessandro Scopelliti, Matilde Todaro, et al. Chemoresistance of Colorectal Cancer Stem CellsAurora-A Is Essential for the Tumorigenic Capacity and

Updated version

10.1158/0008-5472.CAN-09-3953doi:

Access the most recent version of this article at:

Material

Supplementary

http://cancerres.aacrjournals.org/content/suppl/2010/05/10/0008-5472.CAN-09-3953.DC1

Access the most recent supplemental material at:

Cited articles

http://cancerres.aacrjournals.org/content/70/11/4655.full#ref-list-1

This article cites 45 articles, 15 of which you can access for free at:

Citing articles

http://cancerres.aacrjournals.org/content/70/11/4655.full#related-urls

This article has been cited by 8 HighWire-hosted articles. Access the articles at:

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

Subscriptions

Reprints and

[email protected] at

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

Permissions

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

.http://cancerres.aacrjournals.org/content/70/11/4655To request permission to re-use all or part of this article, use this link

Research. on April 26, 2021. © 2010 American Association for Cancercancerres.aacrjournals.org Downloaded from


http://cancerres.aacrjournals.org/lookup/doi/10.1158/0008-5472.CAN-09-3953

http://cancerres.aacrjournals.org/content/suppl/2010/05/10/0008-5472.CAN-09-3953.DC1

http://cancerres.aacrjournals.org/content/70/11/4655.full#ref-list-1

http://cancerres.aacrjournals.org/content/70/11/4655.full#related-urls

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

mailto:[email protected]

http://cancerres.aacrjournals.org/content/70/11/4655


Documents

Tumor and Stem Cell Biology Cancer Research Aurora-A Is … · 2010. 6. 1. · Corresponding Authors: Giorgio Stassi, University of Palermo, Via Liborio Giuffre′, 5 Palermo, 90127,