Aurora-A amplification associated with BRCA2 mutation in breast tumours

Cancer Letters 248 (2007) 96–102

www.elsevier.com/locate/canlet

Aurora-A amplification associated with BRCA2 mutationin breast tumours

Sigridur K. Bodvarsdottir a,b, Holmfridur Hilmarsdottir a,b,Valgerdur Birgisdottir a,b, Margret Steinarsdottir c, Jon G. Jonasson a,b,c,

Jorunn E. Eyfjord a,b,*

a Icelandic Cancer Society, Skogarhlid 8, 105 Reykjavik, Icelandb University of Iceland, Department of Medicine, Vatnsmyrarvegi 16, 101 Reykjavik, Iceland

c Landspitali University Hospital, Hringbraut, 101 Reykjavik, Iceland

Received 11 May 2006; accepted 13 June 2006

Abstract

Potential interaction of Aurora-A amplification and BRCA2 mutation was examined in breast tumours from BRCA2

999del5 mutation carriers (n = 20) and non-carriers (n = 41). Aurora-A amplification studied by FISH was significantlymore common in breast tumours from BRCA2 mutation carriers (p = 0.0005). Extensive Aurora-A amplification was alsodetected on metaphase chromosomes in three breast epithelial cell lines with the same BRCA2 mutation. In addition, sig-nificant association was found between Aurora-A amplification and TP53 mutations in non-BRCA2 mutation carriertumours (p = 0.007). These results suggest that breast tumours with mutations in BRCA2 or TP53 could be promising can-didates for Aurora-A targeted treatment.� 2006 Elsevier Ireland Ltd. All rights reserved.

Keywords: Breast cancer; Aurora-A; BRCA2; TP53; FISH

1. Introduction

Genomic instability is thought to be an earlytransition leading to epithelial carcinogenesis [1].Induction of tumorigenesis may be due to abnor-malities in chromosome segregation and cytokinesisas a consequence of tumour suppressor gene inacti-vation in addition to amplification of oncogenes.One such oncogene is the Aurora-A serine/threonine

0304-3835/$ - see front matter � 2006 Elsevier Ireland Ltd. All rightsdoi:10.1016/j.canlet.2006.06.003

* Corresponding author. Tel.: +354 540 1900; fax: +354 5401905.

E-mail address: jorunn@krabb.is (J.E. Eyfjord).

kinase (also known as STK15 and BTAK), locatedon chromosome 20q13 [2]. Aurora-A amplificationand overexpression are common events in breastcancer cells [2], causing centrosomal amplificationand aneuploidy [3]. Aurora-A overexpression andcentrosome amplification has been shown to be anearly event in rat mammary tumorigenesis [4].In human mammary tumorigenesis, Aurora-A

overexpression seems to correlate with transitionfrom in situ to invasive ductal carcinoma andmay therefore be more relevant in tumour initiationthan progression in breast carcinogenesis [5].Aurora-A amplification has been shown to override

reserved.

S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102 97

the mitotic spindle assembly checkpoint which sub-sequently leads to failure to complete cytokinesis,resulting in multinucleation [6,7].

Genomic instability in breast carcinomas hasbeen related to BRCA1 and BRCA2 familial breastcancers [8–10]. In the Icelandic population only onegerm-line mutation has been found in each gene, arare mutation in BRCA1, D1692N [11] and a morecommon mutation in BRCA2, 999del5 [12]. TheBRCA2 mutation is found in 6–7% of breast cancercases in Iceland [13]. This is an early truncationmutation leading to an unstable gene product [14].Absence of Brca2 in mouse embryonic fibroblasts(MEFs) leads to centrosome amplification andgenomic instability [15]. Neoplastic transformationand aneuploidy has been shown to occur throughinactivation of the mitotic spindle assembly check-point in an in vitro study of brca2 homozygous trun-cated murine cells [16] resulting in delay or failure tocomplete cytokinesis [17]. These observations arevery similar to changes seen in Aurora-A amplifiedtumour cells.

Other tumour suppressor genes, such as TP53,have also been shown to be strongly associated withcentrosomal amplification and chromosomal insta-bility [8,18]. TP53 interacts with Aurora-A and sup-presses its oncogenic activity by transactivationindependent function [19]. Induction of Aurora-Aoverexpression in MEFs prepared from transgenicmice led to aberrant mitosis and binucleated cellformation followed by apoptosis by increasedTP53 protein levels [20]. Apoptosis was significantlysuppressed by deletion of TP53 and significant alter-ation of centromeres and extra nuclei were detectedin TP53�/� MEF cells in the presence of Aurora-Aoverexpression [20,21]. On the other hand theAurora-A kinase activity seems to inactivate TP53function by phosphorylation [22,23]. Aurora-A

overexpression has been associated with mutationsin the TP53 gene in hepatocellular carcinomas [24]referring to in vivo cooperative effects that contrib-utes to tumour formation.

In light of common features of tumours withAurora-A amplification and BRCA2 germ linemutations, such as genomic instability and centro-some amplification, we set out to examine a possibleassociation of Aurora-A amplification with BRCA2.Aurora-A amplification was analysed by fluores-cence in situ hybridization (FISH) in paraffinembedded breast tumour sections from BRCA2

mutation carriers and non-carriers based on selec-tion of genomic instability. Tumours were also ana-

lysed for mutations in exons 5–8 of the TP53 gene.Furthermore, Aurora-A amplification was analysedon metaphase chromosomes from breast epithelialcell lines derived from carriers of the same BRCA2germ-line mutation.

2. Materials and methods

2.1. Breast tumour tissue selection

Selection of breast tumour samples for Aurora-A

amplification analysis by FISH was based on BRCA2

germ-line mutation status, allelic imbalance (AI) at theBRCA2 locus and high DNA index and/or knownkaryotype aberrations (Supplementary table 1). Of 61breast tumour sections from paraffin embedded and for-malin fixed tissues, 20 had BRCA2 999del5 germ-linemutation, additional 19 had AI at the BRCA2 locusand the remaining 22 tumours were without any detectedabnormalities in the BRCA2 gene. Permission from theNational Bioethics Committee and the Data ProtectionAuthority was obtained for the study (99/041V2S1; 99/111V1S1).

2.2. BRCA2 mutation analysis

BRCA2 exon 9 fragments were PCR-amplified fromperipheral blood DNA and run on 7.5% polyacrylamidegels for detection of the Icelandic founder mutation,999del5 [12,25]. Mutants were identified by the presenceof an extra band.

2.3. BRCA2 allelic imbalance detection

For AI detection two markers in the BRCA2 regionwere analyzed; markers D13S260 and D13S171 locatedin region 13q12, centromeric and telomeric to the BRCA2

gene, respectively. PCR was used to detect AI at polymor-phic microsatellite loci by comparing the allelic pattern oftumour and normal DNA. The PCR products wereresolved on 3 mm thick 8% (w/v) High Resolution Repro-Gel acrylamide/bisacrylamide (Amersham Pharmacia,UK), 1· TBE (0.1 M TRIS, 83 mM Boric acid, 1 mMEDTA), (Amersham Pharmacia, UK) using an automatedlaser fluorescent sequencer (ALF Express DNASequencer, Amersham Pharmacia, UK). The followingrunning parameters were used; 1500 V, 60 mA, 25 W,and 55 �C. The relative heights of the alleles were com-pared using an ALFwin Fragment analyzer 1.0 software(Amersham Pharmacia, UK). AI was defined if the differ-ence in relative height between normal and tumour allelewas more than 25% in one or both markers (Supplemen-tary figure 1). When no AI was detected, the sample wasdefined 1 and 0 when alleles were undistinguishablebecause of homozygosity of both microsatellite markers(Supplementary table 1).

98 S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102

2.4. TP53 mutation analysis

TP53 mutation analysis was carried out by polymerasechain reaction (PCR) amplification and constant denatur-ant gel electrophoresis (CDGE) on exons 5–8. Mutationswere confirmed by direct DNA sequencing. PCR condi-tions, CDGE and sequencing condition were as previouslydescribed [26].

2.5. Cell lines

Three mammary epithelial cell lines derived fromBRCA2 999del5 mutation carriers; BRCA2-999del5-1N,BRCA2-999del5-2N and BRCA2-999del5-2T [27], wereused for FISH analysis of Aurora-A amplification. TheN cell lines were derived from normal breast tissue adja-cent to tumour and the T cell line was isolated fromtumour tissue.

2.6. Fluorescence in situ hybridisation (FISH)

Paraffin embedded breast cancer tissue was sliced in4 lm sections for FISH analysis, which was performedusing two different probes simultaneously. For thedetection of Aurora-A amplification the clone RP5-1167H4 [EMBL:AL121914], from the Human BACClone Library (Sanger Centre, UK) was used whichspans the entire STK15/Aurora-A genomic region. Asa control for the ploidy level of chromosome 20 theprobe pZ20 for centromere 20 was used. The RP5-1167H4 clone was labelled with SpectrumOrange-dUTP(Vysis) and the pZ20 centromeric probe with fluoresce-in-12-dUTP (Enzo-Roche), by nick translation. Theslides were deparaffinized, boiled in microwave oventwice for 10 min each in 0.01 M citric acid buffer, pH6, cooled and incubated with 1500 U/ml pepsin at37 �C for 20 min. The slides were then dehydrated.Probes were diluted in t-DenHyb-2 hybridisation buffer(InSitus Biotechnologies, Albuquerque, NM, USA) asdescribed by the manufacturer. Sections and probeswere simultaneously denatured on a heated plate withlid placed on top of Perkin Elmer Cetus DNA ThermalCycle heat-block at 95 �C for 10 min followed by over-night hybridisation at 37 �C in a humid chamber. Slideswere washed three times for 5 min each in 0.1· SSC at60 �C. Fluorescence signals were scored in each sampleby counting the number of single-copy gene and centro-meric signals in average 120 (66–154) well-definednuclei. The cut-off for amplification was when theamplicon-reference ratio was above 1.5 or when theaverage oncogene copy number was higher than 4 (Sup-plementary table 1).

Metaphase chromosomes of cultured cell linesBRCA2-999del-1N, BRCA2-999del-2N and BRCA2-999del-2T were harvested as previously described [27]and spread on slides and FISH performed similar as

described above, excluding deparaffination and micro-wave pre-treatment, using c-DenHyb-2 hybridizationbuffer (Insitus) and lower denature temperature(85 �C). About 20 metaphases were analysed for eachcell line.

2.7. Statistics

Tables of absolute and relative frequencies of the dataare presented in the results section. Associations betweencategorical variables were examined using Fisher’s exacttest. All p values are 2-sided. The statistical packageGraphPad InStat version 3 (GraphPad Software, SanDiego, CA, USA) was used for statistical analysis.

3. Results

Aurora-A gene amplification was detected in 23 of 61(38%) breast tumours (Fig. 1A). Samples were selectedon base of BRCA2 mutation or high DNA index and/orcomplex karyotype (Supplementary table 1). Of the breasttumours from BRCA2 mutation carriers 14 of 20 (70%)had Aurora-A amplification, but only 9 of 41 (22%) fromnon-BRCA2 mutation carriers. This shows a highly signif-icant association of Aurora-A amplification with BRCA2

breast tumours (p = 0.0005; Odds Ratio (OR) = 8.3;95% Confidence Interval (CI) = 2.5–27.8; Table 1). Signif-icant association of Aurora-A amplification was not foundwith AI at the BRCA2 locus as compared to breasttumours without detected AI (p = 0.1274; OR = 4.154;95% CI = 0.7–24.0). However, the association betweenAurora-A amplification and BRCA2 tumours was evenstronger when compared with non-carrier tumours with-out AI at the BRCA2 locus, i.e. excluding those withBRCA2 AI (p = 0.0002; OR = 21.0; 95% CI = 3.6–120.4;Table 1).

Aurora-A copy number changes were also analysed inthree BRCA2 mammary epithelial cell lines carrying thesame mutation as the tumours tested. Extensive Aurora-

A amplification was detected in over 80% of cell meta-phases analysed, either as intrachromosomal amplifica-tion, chromosome 20 triploidy or extrachromosomalamplification (Fig. 1B–D). Intrachromosomal amplifica-tion of Aurora-A was also present in some of the chromo-some 20 triploid cells (Fig. 1C).

TP53 mutation analysis was carried out on DNAfrom the same breast tumours using CDGE on exons5–8 and mutations confirmed with direct DNA sequenc-ing. Mutations in TP53 were detected in 15 tumours of60 analysed (25%). TP53 mutations showed a significantassociation with Aurora-A amplification in tumoursfrom non-BRCA2 mutation carrier (p = 0.007;OR = 11.1; 95% CI = 1.8–67.8; Table 2). No associationwas found with tumours from BRCA2 mutation carriersbut numbers were too low for any conclusions to bedrawn.

Fig. 1. Aurora-A amplification in breast tumour and cell lines. Fluorescence in situ hybridization (FISH) analysis of Aurora-A

amplification (red signal) compared to centromere on chromosome 20 (green signal) (A) on paraffin embedded breast tumour tissue withAurora-A/centromere 20 ratio above 1.5. Intrachromosomal amplification of Aurora-A on chromosome 20 (white arrows) was detected inBRCA2 999del5 breast cell lines, either isolated from normal breast tissue adjacent to a tumour (B) BRCA2-999del5-N1 and (C) BRCA2-999del5-N2, or from breast tumour tissue (D) BRCA2-999del5-T2.

Table 1Association between Aurora-A amplification and BRCA2 mutation in breast tumours

Study group Aurora-A

amplificationp value ORa 95% CIb

Yes No

BRCA2 999del5 carriers 14 6Non-BRCA2 mutation carriers 9 32 0.0005 8.3 2.5–27.8

Non-BRCA2 mutation carriers without BRCA2 AIc 2 18 0.0002 21.0 3.6–120.4

a OR, Odds Ratio.b CI, Confidence interval.c AI, Allelic Imbalance.

S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102 99

4. Discussion

In the present study we describe previouslyunknown association between abnormalities in thecancer related genes, Aurora-A and BRCA2 inbreast tumours. Aurora-A amplification was detect-ed in 70% of tumours from BRCA2 mutation carri-ers but only in 22% of tumours from non-carriers.The association between Aurora-A amplification

and BRCA2 mutated breast tumours was highly sig-nificant (p = 0.0005). Extensive Aurora-A amplifica-tion was also demonstrated in three BRCA2

mutated cell lines, all of which had amplified Auro-

ra-A gene (Fig. 1B–D). In addition, Aurora-A

amplification was found to be more common intumours with AI at the BRCA2 locus (32%) as com-pared with tumours without AI (10%). It is impor-tant to note that these tumour samples were

Table 2Association between Aurora-A amplification and TP53 mutationin sporadic breast tumours

Study group TP53

mutationp value ORa (95% CIb)

Yes No

All breast tumoursAurora-A

amplification8 14

No Aurora-A

amplification8 31 0.14 2.5 (0.8–8.4)

Non-BRCA2 tumoursAurora-A

amplification6 2

No Aurora-A

amplification7 26 0.007 11.1 (1.8–67.8)

a OR = Odds Ratio.b CI = Confidence Interval.

100 S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102

selected to have genomic instability. The differencebetween amplification in tumours with and withoutAI at the BRCA2 locus was not significant but indi-cates a trend for Aurora-A amplification in tumourswith BRCA2 AI. One study published on Aurora-A

amplification in breast tumours analysed by FISHfound amplification in 12% of cases [3]. This is inagreement with our results of 10% Aurora-A ampli-fication in breast tumours with unaffected BRCA2.

The effect of Aurora-A amplification/overexpres-sion and BRCA2 mutations in breast tumour cells,such as aneuploidy, centrosomal amplification,G2/M transition induction and failures or delay incompleting cytokinesis appear to be similar [3,6,8,15–17]. These observations indicate that abnormali-ties in BRCA2 and Aurora-A may interact in earlystages of breast tumorigenesis. In BRCA2 mutationcarriers Aurora-A overexpression, which on its ownhas been demonstrated to be an early mammarytumorigenesis factor [4,5], may lead to earlytumorigenesis.

Frequencies of gains on chromosome 20q, wherethe Aurora-A gene is located, in BRCA2 relatedbreast tumours, have been reported to be 60-75%as compared with only 18-33% in sporadic breasttumours analysed by comparative genomic hybrid-ization (CGH) [28–30]. This difference of gains onchromosome 20q between BRCA2 related andnon-related breast tumours is in agreement with dif-ference of Aurora-A amplification between BRCA2

mutation carriers and non-carriers found in thepresent study by FISH.

Recent study showed that Aurora-A activity wasinhibited in response to double-strand breaks in

DNA through a Chk1-dependent pathway [31].Aurora-A inactivation prevents CDC25B phos-phorylation resulting in cell cycle arrest because oflack of CDK1-cyclin B1 activation [32]. Ectopicoverexpression of Aurora-A leads to bypass ofDNA damage response resulting in mitotic entry[6,31]. Aurora-A kinase activity is essential formitotic entry and becomes first evident at the cen-trosomes during G2 phase [33]. In such circum-stances CDK1 is activated at the centrosomes byAurora-A phosphorylation of CDC25B [34,35].CDK1 has indeed been shown to be among overex-pressed proteins in BRCA2 defective breast tumours[36]. Since BRCA2 responds to DNA damage itmight be expected that a small overdose of Auro-

ra-A expression could override cell cycle arrest whenonly one wild-type BRCA2 gene is expressed. Thiscould make BRCA2 germ-line mutation carriersmore prone to tumorigenesis caused by Aurora-Aoverexpression than non-BRCA2 mutation carriers.

Aurora-A amplification was found to be stronglyassociated with TP53 mutations in non BRCA2

breast tumours in the present study (p = 0.007).Association between Aurora-A and TP53 has pre-viously only been found in hepatocellular carcino-mas [24]. Aurora-A overexpression in transgenicMEFs led to binucleated cell formation followedby increased TP53 level and apoptosis. However,continuous Aurora-A overexpression did not entailmalignant tumour formation until TP53 was delet-ed resulting in suppressed apoptosis followed byneoplastic transformation [20]. Loss of TP53 inMEFs leads to abnormal amplification of centro-somes and failure of cytokinesis [18]. TP53 appearsto be involved both in the regulation of initiationof centrosome duplication in late G1 and suppres-sion of re-duplication in S-phase [37]. SomaticTP53 inactivation appears to be required fortumorigenesis caused by Aurora-A overexpression.As mentioned earlier BRCA2 is involved in main-taining the correct number of centrosomes in G2-M transition [15]. Absence of BRCA2 has beenshown to lead to centrosome amplification as wellas hampering cell division [15,17]. Our results sug-gest that BRCA2 germline mutations may alsoincrease the risk of Aurora-A associated tumorigen-esis probably both through abnormalities in DNAdamage response and control of cell division. Thismay be due to increased risk of Aurora-A amplifi-cation and/or selection for Aurora-A related path-ways leading to tumour formation in BRCA2

mutation carriers.

S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102 101

Aurora-A kinase is a promising target for breastcancer treatment because of its oncogenic behav-iour. Among inhibitors that are known to targetthe enzymatic activity of the Aurora-A kinase arequinazoline derivatives [38,39] and VX-680 [40].BRCA2 associated breast tumours and those withTP53 mutations might therefore be good candidatesfor such therapies because of high frequency ofAurora-A amplification as shown in the presentstudy.

Acknowledgements

The authors thank The Department of Pathologyand The Department of Genetics and MolecularMedicine at Landspitali University Hospital fortheir collaboration and The Icelandic Cancer Socie-ty Biobank for supplying samples. This work wassupported by the Icelandic Research Foundations(RANNIS), The memorial fund of Helga Jonsdottirand Sigurlidi Kristjansson, The memorial fund ofIngibjorg Gudjonsdottir Johnson and The IcelandicCancer Society.

Appendix A. Supplementary data

Supplementary data associated with this articlecan be found, in the online version, at doi:10.1016/j.canlet.2006.06.003.

References

[1] D.A. Nelson, T.T. Tan, A.B. Rabson, D. Anderson, K.Degenhardt, E. White, Hypoxia and defective apoptosisdrive genomic instability and tumorigenesis, Genes Dev. 18(17) (2004) 2095–2107.

[2] S. Sen, H. Zhou, R.A. White, A putative serine/threoninekinase encoding gene BTAK on chromosome 20q13 isamplified and overexpressed in human breast cancer celllines, Oncogene 14 (18) (1997) 2195–2200.

[3] H. Zhou, J. Kuang, L. Zhong, W.L. Kuo, J.W. Gray, A.Sahin, B.R. Brinkley, S. Sen, Tumour amplified kinaseSTK15/BTAK induces centrosome amplification, aneuploi-dy and transformation, Nat. Genet. 20 (2) (1998) 189–193.

[4] T.M. Goepfert, Y.E. Adigun, L. Zhong, J. Gay, D. Medina,W.R. Brinkley, Centrosome amplification and overexpres-sion of aurora A are early events in rat mammary carcino-genesis, Cancer Res. 62 (14) (2002) 4115–4122.

[5] A. Hoque, J. Carter, W. Xia, M.C. Hung, A.A. Sahin, S.Sen, S.M. Lippman, Loss of aurora A/STK15/BTAKoverexpression correlates with transition of in situ to invasiveductal carcinoma of the breast, Cancer Epidemiol. Biomark-ers Prev. 12 (12) (2003) 1518–1522.

[6] S. Anand, S. Penrhyn-Lowe, A.R. Venkitaraman, AURO-RA-A amplification overrides the mitotic spindle assembly

checkpoint, inducing resistance to Taxol, Cancer Cell 3 (1)(2003) 51–62.

[7] Y. Jiang, Y. Zhang, E. Lees, W. Seghezzi, AuroraAoverexpression overrides the mitotic spindle checkpointtriggered by nocodazole, a microtubule destabilizer, Onco-gene 22 (51) (2003) 8293–8301.

[8] S. Gretarsdottir, S. Thorlacius, R. Valgardsdottir, S. Gudl-augsdottir, S. Sigurdsson, M. Steinarsdottir, J.G. Jonasson,K. Anamthawat-Jonsson, J.E. Eyfjord, BRCA2 and p53mutations in primary breast cancer in relation to geneticinstability, Cancer Res. 58 (5) (1998) 859–862.

[9] Y. Miyoshi, K. Iwao, Y. Takahashi, C. Egawa, S. Noguchi,Acceleration of chromosomal instability by loss of BRCA1expression and p53 abnormality in sporadic breast cancers,Cancer Lett. 159 (2) (2000) 211–216.

[10] J.E. Eyfjord, S.K. Bodvarsdottir, Genomic instability andcancer: networks involved in response to DNA damage,Mutat. Res. 592 (1–2) (2005) 18–28.

[11] J.T. Bergthorsson, A. Jonasdottir, G. Johannesdottir, A.Arason, V. Egilsson, S. Gayther, A. Borg, S. Hakanson, S.Ingvarsson, R.B. Barkardottir, Identification of a novelsplice-site mutation of the BRCA1 gene in two breast cancerfamilies: screening reveals low frequency in Icelandic breastcancer patients, Hum. Mutat. Suppl. 1 (1998) S195–S197.

[12] S. Thorlacius, G. Olafsdottir, L. Tryggvadottir, S. Neuhau-sen, J.G. Jonasson, S.V. Tavtigian, H. Tulinius, H.M.Ogmundsdottir, J.E. Eyfjord, A single BRCA2 mutation inmale and female breast cancer families from Iceland withvaried cancer phenotypes, Nat. Genet. 13 (1) (1996) 117–119.

[13] S. Thorlacius, S. Sigurdsson, H. Bjarnadottir, G. Olafsdottir,J.G. Jonasson, L. Tryggvadottir, H. Tulinius, J.E. Eyfjord,Study of a single BRCA2 mutation with high carrierfrequency in a small population, Am. J. Hum. Genet. 60(5) (1997) 1079–1084.

[14] E.K. Mikaelsdottir, S. Valgeirsdottir, J.E. Eyfjord, T.Rafnar, The Icelandic founder mutation BRCA2 999del5:analysis of expression, Breast Cancer Res. 6 (4) (2004) R284–R290.

[15] A. Tutt, A. Gabriel, D. Bertwistle, F. Connor, H. Paterson,J. Peacock, G. Ross, A. Ashworth, Absence of Brca2 causesgenome instability by chromosome breakage and lossassociated with centrosome amplification, Curr. Biol. 9 (19)(1999) 1107–1110.

[16] H. Lee, A.H. Trainer, L.S. Friedman, F.C. Thistlethwaite,M.J. Evans, B.A. Ponder, A.R. Venkitaraman, Mitoticcheckpoint inactivation fosters transformation in cells lack-ing the breast cancer susceptibility gene, Brca2, Mol. Cell 4(1) (1999) 1–10.

[17] M.J. Daniels, Y. Wang, M. Lee, A.R. Venkitaraman,Abnormal cytokinesis in cells deficient in the breast cancersusceptibility protein BRCA2, Science 306 (5697) (2004)876–879.

[18] K. Fukasawa, T. Choi, R. Kuriyama, S. Rulong, G.F.Vande Woude, Abnormal centrosome amplification in theabsence of p53, Science 271 (5256) (1996) 1744–1747.

[19] S.S. Chen, P.C. Chang, Y.W. Cheng, F.M. Tang, Y.S. Lin,Suppression of the STK15 oncogenic activity requires atransactivation-independent p53 function, EMBO J. 21 (17)(2002) 4491–4499.

[20] D. Zhang, T. Hirota, T. Marumoto, M. Shimizu, N.Kunitoku, T. Sasayama, Y. Arima, L. Feng, M. Suzuki,M. Takeya, H. Saya, Cre-loxP-controlled periodic Aurora-A

102 S.K. Bodvarsdottir et al. / Cancer Letters 248 (2007) 96–102

overexpression induces mitotic abnormalities and hyperpla-sia in mammary glands of mouse models, Oncogene 23 (54)(2004) 8720–8730.

[21] P. Meraldi, R. Honda, E.A. Nigg, Aurora-A overexpressionreveals tetraploidization as a major route to centrosomeamplification in p53�/� cells, EMBO J. 21 (4) (2002) 483–492.

[22] H. Katayama, K. Sasai, H. Kawai, Z.M. Yuan, J. Bondaruk,F. Suzuki, S. Fujii, R.B. Arlinghaus, B.A. Czerniak, S. Sen,Phosphorylation by aurora kinase A induces Mdm2-medi-ated destabilization and inhibition of p53, Nat. Genet. 36 (1)(2004) 55–62.

[23] Q. Liu, S. Kaneko, L. Yang, R.I. Feldman, S.V. Nicosia, J.Chen, J.Q. Cheng, Aurora-A abrogation of p53 DNAbinding and transactivation activity by phosphorylation ofserine 215, J. Biol. Chem. 279 (50) (2004) 52175–52182.

[24] Y.M. Jeng, S.Y. Peng, C.Y. Lin, H.C. Hsu, Overexpressionand amplification of Aurora-A in hepatocellular carcinoma,Clin. Cancer Res. 10 (6) (2004) 2065–2071.

[25] K. Gudmundsdottir, S. Thorlacius, J.G. Jonasson, B.F.Sigfusson, L. Tryggvadottir, J.E. Eyfjord, CYP17 promoterpolymorphism and breast cancer risk in males and females inrelation to BRCA2 status, Br. J. Cancer 88 (6) (2003) 933–936.

[26] S. Gudlaugsdottir, V. Sigurdardottir, M. Snorradottir, J.G.Jonasson, H. Ogmundsdottir, J.E. Eyfjord, P53 mutationsanalysis in benign and malignant breast lesions: using needlerinses from fine-needle aspirations, Diagn. Cytopathol. 22 (5)(2000) 268–274.

[27] A.J. Rubner Fridriksdottir, T. Gudjonsson, T. Halldorsson,J. Bjornsson, M. Steinarsdottir, O.T. Johannsson, H.M.Ogmundsdottir, Establishment of three human breast epi-thelial cell lines derived from carriers of the 999del5 BRCA2Icelandic founder mutation, In Vitro Cell Dev. Biol. Anim.41 (10) (2005) 337–342.

[28] M. Tirkkonen, O. Johannsson, B.A. Agnarsson, H. Olsson,S. Ingvarsson, R. Karhu, M. Tanner, J. Isola, R.B. Barkar-dottir, A. Borg, O.P. Kallioniemi, Distinct somatic geneticchanges associated with tumor progression in carriers ofBRCA1 and BRCA2 germ-line mutations, Cancer Res. 57(7) (1997) 1222–1227.

[29] E.H. van Beers, T. van Welsem, L.F. Wessels, Y. Li, R.A.Oldenburg, P. Devilee, C.J. Cornelisse, S. Verhoef, F.B.Hogervorst, L.J. van’t Veer, P.M. Nederlof, Comparativegenomic hybridization profiles in human BRCA1 andBRCA2 breast tumors highlight differential sets of genomicaberrations, Cancer Res. 65 (3) (2005) 822–827.

[30] G. Jonsson, T.L. Naylor, J. Vallon-Christersson, J. Staaf, J.Huang, M.R. Ward, J.D. Greshock, L. Luts, H. Olsson, N.Rahman, M. Stratton, M. Ringner, A. Borg, B.L. Weber,Distinct genomic profiles in hereditary breast tumors iden-

tified by array-based comparative genomic hybridization,Cancer Res. 65 (17) (2005) 7612–7621.

[31] A. Krystyniak, C. Garcia-Echeverria, C. Prigent, S. Ferrari,Inhibition of Aurora A in response to DNA damage,Oncogene 25 (3) (2006) 338–348.

[32] M. Cazales, E. Schmitt, E. Montembault, C. Dozier, C.Prigent, B. Ducommun, CDC25B phosphorylation byAurora-A occurs at the G2/M transition and is inhibitedby DNA damage, Cell Cycle 4 (9) (2005) 1233–1238.

[33] T. Hirota, N. Kunitoku, T. Sasayama, T. Marumoto, D.Zhang, M. Nitta, K. Hatakeyama, H. Saya, Aurora-A aninteracting activator the LIM protein Ajuba are required formitotic commitment in human cells, Cell 114 (5) (2003) 585–598.

[34] S. Dutertre, M. Cazales, M. Quaranta, C. Froment, V.Trabut, C. Dozier, G. Mirey, J.P. Bouche, N. Theis-Febvre,E. Schmitt, B. Monsarrat, C. Prigent, B. Ducommun,Phosphorylation of CDC25B by Aurora-A at the centro-some contributes to the G2-M transition, J. Cell Sci. 117 (Pt12) (2004) 2523–2531.

[35] C.P. De Souza, K.A. Ellem, B.G. Gabrielli, Centrosomaland cytoplasmic Cdc2/cyclin B1 activation precedes nuclearmitotic events, Exp. Cell Res. 257 (1) (2000) 11–21.

[36] J. Palacios, E. Honrado, A. Osorio, A. Cazorla, D. Sarrio,A. Barroso, S. Rodriguez, J.C. Cigudosa, O. Diez, C.Alonso, E. Lerma, J. Dopazo, C. Rivas, J. Benitez, Pheno-typic characterization of BRCA1 and BRCA2 tumors basedin a tissue microarray study with 37 immunohistochemicalmarkers, Breast Cancer Res. Treat 90 (1) (2005) 5–14.

[37] P. Tarapore, H.F. Horn, Y. Tokuyama, K. Fukasawa,Direct regulation of the centrosome duplication cycle by thep53-p21Waf1/Cip1 pathway, Oncogene 20 (25) (2001) 3173–3184.

[38] C. Ditchfield, V.L. Johnson, A. Tighe, R. Ellston, C.Haworth, T. Johnson, A. Mortlock, N. Keen, S.S. Taylor,Aurora B couples chromosome alignment with anaphase bytargeting BubR1, Mad2, and Cenp-E to kinetochores, J. CellBiol. 161 (2) (2003) 267–280.

[39] N.M. Heron, M. Anderson, D.P. Blowers, J. Breed, J.M.Eden, S. Green, G.B. Hill, T. Johnson, F.H. Jung, H.H.McMiken, A.A. Mortlock, A.D. Pannifer, R.A. Pauptit, J.Pink, N.J. Roberts, S. Rowsell, SAR and inhibitor complexstructure determination of a novel class of potent andspecific Aurora kinase inhibitors, Bioorg. Med. Chem. Lett16 (5) (2006) 1320–1323.

[40] E.A. Harrington, D. Bebbington, J. Moore, R.K. Rasmus-sen, A.O. Ajose-Adeogun, T. Nakayama, J.A. Graham, C.Demur, T. Hercend, A. Diu-Hercend, M. Su, J.M. Golec,K.M. Miller, VX-680, a potent and selective small-moleculeinhibitor of the Aurora kinases, suppresses tumor growthin vivo, Nat. Med. 10 (3) (2004) 262–267.