Oncogene (2002) 21, 6175 – 6183 ª 2002 Nature Publishing Group All rights reserved 0950 – 9232/02 $25.00 www.nature.com/onc

On the role of aurora-A in function

Ste´ phanie Dutertre1, Simon Descamps1 and Claude Prigent*,1

1Groupe Cycle Cellulaire, UMR 6061 Ge´ne´tique et de´veloppement, CNRS-Universite´ de Rennes I, IFR 97 Ge´nomique Fonctionnelle et Sante´, Faculte´ de Me´decine, 2 avenue du Pr Leon Bernard, CS 34317, 35043 Rennes cedex, France

Mammalian aurora-A belongs to a multigenic family of 1998). Aurora-B is a passenger involved mitotic serine/threonine comprising two other in and chromosome architecture (Adams et members: aurora-B and aurora-C. In this review we will al., 2000, 2001; Giet and Glover, 2001). Aurora-A is a focus on aurora-A that starts to localize to centrosome which function is not yet fully only in as soon as have been understood. We must say here that although mamma- duplicated, the protein is then degraded in early G1. lian encode for three different aurora kinases, Works in various organisms have revealed that the C. elegans and D. melanogaster genomes encodes for kinase is involved in centrosome separation, duplication aurora-A and aurora-B paralogue whereas the yeast and maturation as well as in bipolar spindle assembly genomes S. cerevisiae and S. pombe encode for only and stability. Aurora kinases are found in all organisms one (Chan and Botstein, 1993; Petersen in which their function has been conserved throughout et al., 2001). Although this is only speculation, it seems evolution, namely the control of chromosome segrega- that at least for yeast the functions of both the tion. In human, aurora-A has focused a lot of attention, mammalian aurora-A and -B is fulfilled by a unique since its overexpression has been found to be correlated kinase, suggesting that functions have diverged with the grade of various solid tumours. Ectopic kinase throughout evolution. One must then take carefully overexpression in any culture cell line leads to polyploidy the functions described for the yeast aurora kinases and centrosome amplification. However, overexpression and take into account the different localization of of aurora-A in particular cell lines such as NIH3T3 is aurora-A and aurora-B in mammalian cells. For sufficient to induce growth on soft agar. Those instance, the yeast aurora substrates found to date transformed cells form tumours when implanted in are the Ndc10p kinetochore protein, and the histone immunodeficient mice, indicating that the kinase is an H3 chromosomal protein whose localization corre- oncogene. sponds to the mammalian aurora-B localization. The Oncogene (2002) 21, 6175 – 6183. doi:10.1038/sj.onc. phosphorylation state of Ndc10p in yeast controls the 1205775 spindle attachment to the chromosome kinetochore (Biggins et al., 1999; Sassoon et al., 1999). Keywords: aurora; ; centrosome; kinase In addition, although the exact function of serine 10 phosphorylation in H3 during remains contro- versial, it seems to be necessary for chromosome segregation. In multicellular eukaryotes histone H3 Introduction have been demonstrated to be an aurora-B substrate by RNA-mediated interference both in C. elegans (Hsu et The mammalian aurora-A has been first identified al., 2000) and in D. melanogaster (Adams et al., 2001; within the amplicon 20q13 frequently detected in Giet and Glover, 2001). Aurora-A phosphorylates the human breast tumours (Tanner et al., 1994, 1995). related protein Eg5 required for both centro- Only a partial cDNA encoding a protein kinase was some separation and spindle assembly and stability, isolated and named BTAK for Breast Tumour just like aurora-A (Giet et al., 1999a). The kinase also Activated Kinase (Sen et al., 1997). Few years later, phosphorylates CPEB protein (Cytoplasmique Poly- a encoding a kinase homologue was adenylation Element Binding protein), and this identified and named aurora (Glover et al., 1995). phosphorylation induces of mRNA Many aurora kinases have then been found in various allowing translation during maturation organisms giving birth to as many different confusing (Mendez et al., 2000). This mechanism seems to occur names as publications, and Erich Nigg proposed the locally within the spindle pole in the vicinity of the nomenclature aurora-A, -B and -C (Nigg, 2001). centrosome (Groisman et al., 2000). Recently the Aurora-C expression seems to be restricted to human TACC3 (Transforming Acidic Coiled Coil spermatogenesis at the time when cells assemble the protein number 3) has also been found to be two meiotic spindles (Hu et al., 2000; Tseng et al., phosphorylated by aurora-A. The function of this phosphorylation in D. melanogaster is to recruit the protein to the centrosome (Giet et al., 2002). This *Correspondence: C Prigent; E-mail: [email protected] might in fact be one of the major roles of aurora-A; Aurora and centrosome S Dutertre et al 6176 recruiting proteins on the centrosome to fulfil a local function, insuring centrosome maturation for instance (Berdnik and Knoblich, 2002; Hannak et al., 2001). In order to better understand what could be the function of aurora-A, in the first part of this review we will describe in a very simplistic way the behaviour and the function of the centrosome during . We will then concentrate on the aurora-A kinase itself, its structure, localization and cell cycle behaviour. In a second part, we will focus on the four different functions attributed to aurora-A during cell cycle progression in various organisms. Finally, we will introduce the oncogenic potential of the kinase over- expression. Indeed, the main goal in current aurora-A research is to understand how its amplification leads to cell transformation.

Aurora-A localizes to the centrosome

The centrosome The centrosome in G1 can be schematically described as an organelle comprising three areas: the two centrioles, the connection between centrioles (the matrix) and the pericentriolar material (PCM) (Figure 1). The is made of nine triplets of micro- tubules assembled around a ‘cartwheel’ of unknown composition to form a cylinder. The two centrioles are different, one old ‘mother’ (with appendages) and one young ‘daughter’. The matrix is composed of various proteins that connect the two centrioles in a flexible Figure 1 Localization area of aurora-A at the centrosome. The top panel shows a very simplistic scheme of a centrosome with manner. The composition of the PCM varies along the its two centrioles (in blue); the young one (light blue) and the cell cycle progression with proteins coming and leaving old one (dark blue) with its appendages (in grey). The connections this very active area. In mammalian cells, the major between the centrioles are represented by a green area and the function of the centrosome has long been thought to be pericentriolar material in red correspond to aurora-A localization. only a organizing centre (MTOC) organiz- The lower panel shows an image taken in a cell showing one cen- trosome (two centrioles) stained with GFP-centrin (green) and an ing the interphase microtubule network and the bipolar anti-aurora-A antibody (red) mitotic spindle. However, the centrosome, at least one of the centrioles, is also implicated in cytokinesis. Indeed, mammalian cells lacking centrosome can assemble a bipolar spindle, but are not able to divide from the end of S phase to the beginning of the G1 (Bornens, 2002). phase of the next cycle (Figure 2a,b). The three aurora kinases that show different cellular localization have a conserved catalytic domain (C- Aurora-A term) and a non-catalytic domain (N-term) different in During the of the cell cycle aurora-A size and sequence. In the case of aurora-A, both GFP- colocalizes with g- in mammalian cells indicat- tagged domains localize to centrosomes. However, the ing that it is a centrosome protein (Gopalan et al., centrosome localization of the N-terminal domain 1997; Kimura et al., 1997; Roghi et al., 1998). The depends on the presence of microtubules (active protein is also found at the bipolar mitotic spindle mechanism) whereas the centrosome localization of poles. However, the immunodetection of aurora-A in the C-terminal domain is independent of microtubules cultured cells requires an absence of detergent indicat- (passive mechanism) (Giet and Prigent, 2001). Depoly- ing that the binding of the kinase to the centrosome is merization of the microtubules by nocodazole weak. At the electron microscopy level aurora-A is treatment does not remove aurora-A from the mainly found around the PCM (Roghi et al., 1998) and centrosome, indicating that its localization can be in GFP-centrin transformed cells aurora-A is clearly microtubule independent (Roghi et al., 1998). The detected in a quite large area around the two dots catalytic domain presumably localizes to centrosomes corresponding to the centrioles (Figure 1b and personal because of its affinity for the kinase substrates (Giet observation). The localization of aurora-A depends on and Prigent, 2001). However, its kinase activity is not the cell cycle: the protein is detected by indirect necessary for centrosome localization, since the kinase- immunofluorescence only on duplicated centrosomes, dead mutant localizes like the active one (Giet and

Oncogene Aurora and centrosome S Dutertre et al 6177

Figure 2 Aurora-A and centrosome during cell cycle. (a) Scheme of centriole duplication, centrosome behaviour and aurora-A localization during cell cycle progression. The old (mother) centriole is in dark blue and the young (daughter) one in light blue. Aurora-A localization area is in red. (b) Indirect immunolocalization of aurora-A (red) in GFP-centrin (green) transfected cells. Three cells in different cell cycle phases were selected : G1 (top panel), S/G2 (middle panel) and (lower panel). Enlarge- ment of the centrosome area is shown on the right panels. DNA stained with DAPI appears in blue

Prigent, 1998). Aurora-A is also thought to be a microtubule-associated protein and binds to taxol- stabilized microtubules in vitro (Roghi et al., 1998). Aurora-A is degraded by the proteasome and the ubiquitin dependent pathway (Honda et al., 2000; Walter et al., 2000). Its degradation occurs in early G1, and depends on a degradation box, the sequence RxxLxxG located in the C-terminal end of the protein, targeted by Fizzy-related protein (Arlot-Bonnemains et al., 2001; Castro et al., 2002) (Figure 3). Aurora-A kinase activity is activated by phosphorylation of a threonine residue within the activation loop of the kinase catalytic domain (Walter et al., 2000) (Figure 3). Although it seems that the activation depends on cdc2/cyclinB activity, the kinase that phosphorylates aurora-A remains to be found (Frank-Vaillant et al., Figure 3 Human aurora-A structure. The kinase catalytic do- main is shown in red. It contains an aurora kinase signature 2000). (xRxTxCGTx) in the activation loop between subdomains VII and VIII with a threonine whose phosphorylation activates the ki- nase. It also contains a degradation box (RxxLxG) in its C-term- Aurora-A and centrosome behaviour inal end that is targeted by Fizzy-related to induce aurora-A degradation by the ubiquitin dependent pathway and the protea- some. Aurora-A N-terminal domain (in blue) serves as a localiza- Aurora-A localizes to centrosomes immediately after tion domain to target the protein to the centrosome in a their duplication (end of S phase). The kinase microtubule dependent manner

Oncogene Aurora and centrosome S Dutertre et al 6178 accumulates there until early G1 when it is degraded. addition of an inactive recombinant protein kinase During all this period, centrosomes are not allowed to also leads to monopolar spindle (Roghi et al., 1998). re-duplicate, they must separate and maturate (end of Some of these monopolar spindles seemed to contain G2/), and they participate to bipolar spindle two centrosomes that failed to separate (Giet et al., assembly by nucleating both spindle and astral 1999b). In contrast, during the first mitosis in C. microtubules in mitosis. elegans, aurora-A is not required for centrosome separation (Schumacher et al., 1998), although in the absence of the kinase bipolar mitotic spindles fail to Centriole duplication assemble because centrosomes collapse together lead- Just like DNA replication, centriole duplication ing to monopolar spindle formation (Hannak et al., depends on the restriction point controlled at the 2001). molecular levels by cdk2/cyclinE kinase and the In Xenopus, the kinesin-related protein Eg5 is transcription factor E2F (Lacey et al., 1999; Matsu- phosphorylated by aurora-A (Giet et al., 1999). Eg5 moto et al., 1999; Meraldi et al., 1999), and begins a that has been implicated in centrosome separation little earlier than the G1/S transition before initiation (Blangy et al., 1995; Sawin et al., 1992; Whitehead and of DNA replication. Before the end of S phase, the cell Rattner, 1998; Whitehead et al., 1996) begins to contains four centrioles forming two unseparated localize precisely between the two duplicated centro- centrosomes (Doxsey, 2001). This is the time when somes decorated by aurora-A at the end of S phase aurora-A begins to decorate both centrosomes. Thus it (Giet et al., 1999). There is no direct proof that aurora- seems unlikely that aurora-A is directly involved in A phosphorylates Eg5 in order to trigger centrosome centriole duplication. separation, but inhibition of aurora-A or inhibition of However ectopic expression of aurora-A in any Eg5 leads to a failure to separate centrosomes both in cultured cells provokes centrosome amplification Xenopus (Roghi et al., 1998; Sawin et al., 1992) and in (Meraldi et al., 2002; Zhou et al., 1998). Surprisingly, Drosophila (Glover et al., 1995; Heck et al., 1993). S. the kinase activity is not required, because over- cerevisiae possesses only one aurora kinase (Ipl1p), expression of an aurora-A dead-kinase also leads to although the majority of ipl1 mutant cells show a centrosome amplification (Meraldi et al., 2002; our normal mitotic spindle, few cells exhibit a monopolar unpublished observation). How could aurora-A control spindle at the restrictive temperature (Kim et al., 1999). centrosome amplification? Centrioles start to duplicate In yeast, the spindle pole bodies (yeast centrosome) while the kinase is not yet detected in the cell. Aurora- separation requires the function of both motor proteins A could somehow participate in a cross-talk between Cin8p (Eg5 homologue) and Kip1p (Hoyt et al., 1992; the centrosome cycle and the nuclear cycle. The kinase Roof et al., 1992; Saunders and Hoyt, 1992). It could, for instance, control the ‘licence’ to duplicate by happens that IPL1 and CIN8 are genetically related localizing to centrosomes just after duplication to and that CIN8 gene becomes essential in ipl1 mutants activate an inhibitor of duplication. Duplication must (Geiser et al., 1997; Kim et al., 1999). occur once and only once during the cell cycle, re- duplication being allowed only after passage through Centrosome maturation mitosis. An increased level of active or inactive kinase in cells might be sufficient to deregulate this control. Once duplicated and separated, centrosomes must go through a maturation process that mainly consists of recruiting various proteins that contribute to the Centrosome separation structure of the centrosome, and that will also ensure After duplication, centrosomes must separate and centrosome function during mitosis and the next G1. migrate around the nuclear envelop to reach (in In C. elegans g-tubulin and at least two other prophase) opposite positions that will correspond to proteins from the PCM are recruited to the centrosome the bipolar mitotic poles. This movement depends on in an aurora-A dependent manner (Hannak et al., the presence of microtubules and is under control of 2001). In Drosophila D-TACC (Transforming Acidic molecular motors. Defects in centrosome separation Coiled Coil) and MSPS (Mini Spindle) do not localize necessarily lead to the formation of a monopolar anymore to centrosome in mutant embryos and spindle that is unable to ensure chromosome segrega- cultured cells in which aurora-A RNA-mediated tion, inducing abortive mitosis. First described in interference was performed (Giet et al., 2002). In Drosophila, mutation in aurora-A provokes the appari- Drosophila AurA37 mutant cells, aurora-A mutation tion of monopolar spindles due to centrosome lies in the kinase catalytic domain (subdomain III) with separation defect and bipolar spindle where one of a conserved arginine in position 201 replaced by a the poles is formed without centrosome (Glover et al., histidine residue. This mutation affects aurora-A 1995). In this case, the opposite pole often contains localization that is not detected anymore on the multiple centrioles suggesting that centrosomes failed centrosome. In the majority of the mutant cells, g- to separate and were amplified (Giet et al., 2002). A tubulin and centrosomin are not detected on the role of aurora-A in centrosome separation was centrosome, suggesting that aurora-A is required for confirmed in Xenopus egg extracts, in which inhibition the recruitment of these proteins on the centrosome of aurora-A during bipolar spindle formation by (Berdnik and Knoblich, 2002).

Oncogene Aurora and centrosome S Dutertre et al 6179 bad prognosis in adenopathy free breast (Isola Mitotic spindle et al., 1995). In gastric cancers, amplification of the In interphase cells, microtubule array is derived from a 20q13 region is detected by Southern blot in 5% (4/72) single centrosome. At the onset of mitosis, interphasic of the cases with amplification from 3.5 – 6.3-fold. In microtubule array breaks down and duplicated-centro- contrast, amplification of the 20q13 region is found by somes nucleate highly dynamic microtubules that FISH analysis in 34/72 of the cases. Amplification was assemble the bipolar spindle. Inhibition of aurora-A high in four cases (5%), moderated in 11 cases (15%) activity in Xenopus egg extracts that have already and low in 19 cases (28%) (Sakakura et al., 2001). The assembled such spindles leads to a destabilization of 20q13 amplification is also detected in 52% (41/79) of the structure (Giet and Prigent, 2000). From a bipolar primary colorectal tumours (Bischoff et al., 1998). spindle with aligned in the metaphase plate, two rosettes of microtubules surrounded by Aurora-A mRNA and protein overexpression chromosomes are generated. The whole mechanism seems to correspond to an inhibition of a stabilizing Aurora-A mRNA overexpression correlates with the factor that link microtubules together (Giet and presence of the 20q13 amplicon. In primary colorectal Prigent, 2000). A defect in Eg5 function would explain tumours, mRNA expression increases from 4 – 28-fold this phenotype, but other spindle stabilizing proteins in 54% (22/41) (Bischoff et al., 1998). Overexpression are also good candidates. The protein D-TACC, for also occurs in gastric cancer cell lines as well as in 18 of instance, is a centrosome protein, it associates with 35 primary gastric cancers (Sakakura et al., 2001). At microtubules and organizes and/or stabilizes centro- the immunohistochemical level aurora-A is over- some microtubules in early Drosophila embryos expressed in 94% (31/33) of primary invasive (Gergely et al., 2000a,b). D-TACC interacts with mammary carcinomas irrespective of their histopatho- mini-spindle (MSPS), a microtubule associated protein logical types. Moreover, the aurora-A staining in these also required for the integrity of the spindle (Cullen et cases is found diffusely in cytoplasm, and not restricted al., 1999; Lee et al., 2001). Interestingly, in Drosophila to centrosomes. Aurora-A is undetectable in normal aurora-A mutant cells, D-TACC and MSPS proteins mammary tissues or in necrotic cancer cells (Tanaka et are reduced at the poles of the mitotic spindle (Giet et al., 1999). Interestingly, aurora-A overexpression is al., 2002). Furthermore, aurora-A phosphorylates D- significantly associated with primary tumour invasive- TACC and both protein co-immunoprecipitates. The ness and accumulation (Takahashi et al., 2000). interaction is conserved in human in which aurora-A interacts with TACC3 (Giet et al., 2002). In addition to Aurora-A, ploidy and cancer a mis-localization of proteins, Drosophila aurora-A mutant shows a reduction in astral microtubules The more frequent genomic damage that occurs during associated with a reduction in the spindle length. D- cancer development is chromosomal instability that TACC could be considered as a regulatory subunit involves gains and losses of pieces of chromosomes or targeted to the centrosome by aurora-A and MSPS a whole chromosomes (Lengauer et al., 1998). A great catalytic subunit, which stabilizes microtubules. The number of published works show that is a same defects have been observed in C. elegans after good marker of tumour progression and prognosis. elimination of aurora-A by RNA interference (Schu- Bladder carcinoma are detected by measuring urine macher et al., 1998). Loss of aurora-A expression cytology DNA for instance (Klein et al., 1982). results in the assembly of a spindle with few and short Furthermore, in gastric carcinoma (Sasaki et al., microtubules. Centrosomes in these embryos also 1999) and in papillary thyroid carcinoma (Sturgis et display defects in microtubule nucleation activity al., 1999) aneuploidy is a marker of metastasis and in (Schumacher et al., 1998; Hannak et al., 2001). many malignancies aneuploidy is associated with a poor outcome (Abad et al., 1998; Magennis, 1997; Sasaki et al., 1999). A correlation between aurora-A Aurora-A overexpression, cancer and centrosome overexpression and aneuploidy exists in gastric cancer. amplification Indeed, clinical samples with aurora-A amplification and overexpression show aneuploidy and poor prog- Aurora-A gene lies in the 20q13 amplicon nosis (Sakakura et al., 2001). Aurora-A gene is located in the 20q13 chromosome region that is frequently amplified in several different Aurora-A, centrosome aberrations and cancer types of malignant tumours such as colorectal, breast and bladder cancers (Bischoff et al., 1998; Zhou et al., Usually cells become aneuploid because of aberrant or 1998). In breast cancer cells in particular, the 20q11- abortive mitosis. It can be a consequence of a q13 region is amplified in 40% of cancer cell lines as centrosome duplication cycle defect that would lead well as in 12 to 18% of primary tumours. The to an abnormal mitotic spindle assembly and a failure incidence increases to 60% in breast cancers with in correctly segregating chromosomes. Such errors familial predisposition due to a mutation of the might contribute to cancer appearance by generating BRCA2 gene (Tirkkonen et al., 1997). Moreover, a genetic instability, increasing the risk for a cell to amplification of the 20q13 region is associated to a gain involved in proliferation or to lose tumour

Oncogene Aurora and centrosome S Dutertre et al 6180 suppressor genes. Interestingly despite the fact that centrosome amplification is not the only centrosome abnormality found in solid tumours, it has become, just like polyploidy, a signature for many tumours. Indeed, centrosome aberrations are found in brain (Pihan et al., 1998; Weber et al., 1998), breast (Lingle et al., 1998; Lingle and Salisbury, 1999; Pihan et al., 1998), lung, colon, prostate (Pihan et al., 1998), pancreas (Sato et al., 1999), bile duct (Kuo et al., 2000), head and neck (Gustafson et al., 2000). Numbers of centrosome alterations are observed in tumours such as supernumerary centrioles, abnormal centrosome number or volume, centrosome protein phosphorylation modification, accumulation of an excess of pericentriolar material etc. (Lingle et al., 1998; Pihan et al., 1998). All those centrosome alterations might lead to bipolar mitotic spindle defects, chromosomal segregation deficiency and aneu- ploidy. Although no direct link has been found between aurora-A overexpression and centrosome abnormalities in cancer, aurora-A overexpression, centrosome amplification and aneuploidy are always associated. Furthermore, centrosome aberrations are known to lead to aneuploidy, suggesting that aurora-A overexpression which is responsible for centrosome amplification must participate in tumorigenesis in a particular genetic background.

Aurora-A ectopic overexpression in cell lines Many experiments have demonstrated the requirement of aurora-A in centrosome regulation. In Drosophila, Aurora-A mutants show several centrosome defects that were confirmed by RNA-mediated interference (RNAi) (Giet et al., 2002; Glover et al., 1995). Aurora- A overexpression in cell lines leads to centrosome amplification, multipolar spindle and polyploidy (Bischoff et al., 1998; Zhou et al., 1998) (Figure 4). Overexpression of wild type aurora-A or a constitu- tively active mutant of the kinase in Rat1 or NIH3T3 cells leads to a higher number of colonies growing in soft agar. That was not the case when an inactive mutant aurora-A was overexpressed. Moreover, Rat1 cells transformed with active aurora-A show tumour growth in nude mice, demonstrating that aurora-A is an oncogene (Bischoff et al., 1998). Transient expression of active aurora-A in the near diploid human breast epithelial cell line MCF10A leads to aberrant chromosome segregation and an increase in ploidy (Zhou et al., 1998). In contrast, a C-terminal- Figure 4 Overexpression of wild type GFP-aurora-A in deleted mutant of aurora-A induces no effect in the MCF10A cells. Top panel: interphase cell with centrosome ampli- ploidy state of transfected gastric cancer cells, but fication, the green dots correspond to aurora-A/centrosome loca- inhibits cell proliferation (Sakakura et al., 2001). lization. lower panel: a metaphase cell with a multipolar spindle. However, surprisingly, an inactive aurora-A kinase DNA stained with DAPI appears in blue produces virtually identical phenotypes to the wild-type active kinase when overexpressed in Chinese hamster ovary (CHO) cells (Meraldi et al., 2002). Both transform cells, and (2) that centrosome amplification transfected cells exhibit extra copies of centrosomes. is not sufficient to provoke transformation (Bischoff et However such inactive aurora-A is unable to transform al., 1998) (Figure 5). Thus, overexpression of the Rat1 or NIH3T3 cells, strongly indicating that (1) a aurora-A kinase is able to induce centrosome amplifi- gain of aurora-A kinase activity is necessary to cation, but its kinase activity is not necessary to obtain

Oncogene Aurora and centrosome S Dutertre et al 6181 this phenotype. Maybe the accumulation of the protein of aurora-A is transforming only in the presence of co- leads to a disorganization of the centrosome machinery amplification or mutation in additional genes that unrelated to its catalytic function but caused by a steric could be involved in a process supposed to eliminate accumulation that interferes with the centrosomal cells with aberrant DNA content, just as in p537/7 functions. cells (Meraldi et al., 2002). In contrast, the kinase activity of aurora-A is necessary to obtain the transformed phenotype by Aurora-A: a proliferation marker? overexpression in cultured cells. Thus centrosome amplification induced by aurora-A overexpression is The high correlation between aurora-A, high histo- not sufficient for its oncogenic activity. Under over- prognostic grade, aneuploidy makes the kinase a expression, aurora-A might phosphorylate substrates potential prognostic factor. In breast cancer less than unrelated to the centrosome function but involved in half of cancer cells are PCNA positive (Proliferating oncogenesis. Not all cell lines are transformed by Cell Nuclear Antigen) despite the fact that they are all aurora-A overexpression. Then, as suggested by aurora-A positive (490%) (Tanaka et al., 1999). In Bischoff et al. (1998), it is possible that overexpression primary colorectal tumours as well as in breast cancer

Figure 5 Overexpression of aurora-A, centrosome amplification and tumorogenesis. Overexpression of either the active or inactive aurora-A-GFP provokes centrosome amplification in any cell line. Duplicated (amplified) centrosomes appear in green. Only over- expression of the active kinase and in particular cell lines is sufficient to obtain a transformed phenotype (in red)

Oncogene Aurora and centrosome S Dutertre et al 6182 no correlation was found between aurora-A and Acknowledgements PCNA expression (Takahashi et al., 2000; Tanaka et We are grateful to Regis Giet for critical reading of the al., 1999). In contrast to PCNA that is a well- manuscript. We wish to thank Michel Bornens for established and routinely used proliferation marker, providing us with GFP-centrin transformed cell lines. aurora-A would rather be a malignancy marker that Work in the cell cycle group is supported by the CNRS, the ‘Ligue Nationale Contre le Cancer’ (LNCC), and the should also be useful in treatment choice. For instance, ‘Association pour la Recherche sur le Cancer’ (ARC). S the inhibition of the kinase activity could help to Dutertre is a fellow of the ARC and S Descamps is a fellow reduce tumour aggressiveness in addition to commonly of the LNCC. used chemotherapeutic drugs.

References

Abad M, Ciudad J, Rincon MR, Silva I, Paz-Bouza JI, Lopez Groisman I, Huang YS, Mendez R, Cao Q, Theurkauf W A, Alonso AG, Bullon A and Orfao A. (1998). Anal. Cell and Richter JD. (2000). Cell, 103, 435 – 447. Pathol., 16, 223 – 231. Gustafson LM, Gleich LL, Fukasawa K, Chadwell J, Miller Adams RR, Maiato H, Earnshaw WC and Carmena M. MA, Stambrook PJ and Gluckman JL. (2000). Laryngo- (2001). J. Cell Biol., 153, 865 – 880. scope, 110, 1798 – 1801. Adams RR, Wheatley SP, Gouldsworthy AM, Kandels- Hannak E, Kirkham M, Hyman AA and Oegema K. (2001). Lewis SE, Carmena M, Smythe C, Gerloff DL and J. Cell. Biol., 155, 1109 – 1116. Earnshaw WC. (2000). Curr. Biol., 10, 1075 – 1078. Heck MM, Pereira A, Pesavento P, Yannoni Y, Spradling Arlot-Bonnemains Y, Klotzbucher A, Giet R, Uzbekov R, AC and Goldstein LS. (1993). J. Cell. Biol., 123, 665 – 679. Bihan R and Prigent C. (2001). FEBS Lett., 508, 149 – 152. Honda K, Mihara H, Kato Y, Yamaguchi A, Tanaka H, Berdnik D and Knoblich JA. (2002). Curr. Biol., 12, 640 – Yasuda H, Furukawa K and Urano T. (2000). Oncogene, 647. 19, 2812 – 2819. Biggins S, Severin FF, Bhalla N, Sassoon I, Hyman AA and Hoyt MA, He L, Loo KK and Saunders WS. (1992). J. Cell. Murray AW. (1999). Genes Dev., 13, 532 – 544. Biol., 118, 109 – 120. Bischoff JR, Anderson L, Zhu Y, Mossie K, Ng L, Souza B, Hsu JY, Sun ZW, Li X, Reuben M, Tatchell K, Bishop DK, Schryver B, Flanagan P, Clairvoyant F, Ginther C, Chan Grushcow JM, Brame CJ, Caldwell JA, Hunt DF, Lin R, CS, Novotny M, Slamon DJ and Plowman GD. (1998). Smith MM and Allis CD. (2000). Cell, 102, 279 – 291. EMBO J., 17, 3052 – 3065. Hu HM, Chuang CK, Lee MJ, Tseng TC and Tang TK. Blangy A, Lane HA, d’Herin P, Harper M, Kress M and (2000). DNA Cell. Biol., 19, 679 – 688. Nigg EA. (1995). Cell, 83, 1159 – 1169. Isola JJ, Kallioniemi OP, Chu LW, Fuqua SA, Hilsenbeck S, Bornens M. (2002). Curr. Opin. Cell Biol., 14, 25 – 34. Osborne CK and Waldman FM. (1995). Am.J.Pathol., Castro A, Arlot-Bonnemains Y, Vigneron S, Labbe JC, 147, 905 – 911. Prigent C and Lorca T. (2002). EMBO Rep., 3, 457 – 462. Kim JH, Kang JS and Chan CS. (1999). J. Cell. Biol., 145, Chan CS and Botstein D. (1993). Genetics, 135, 677 – 691. 1381 – 1394. Cullen CF, Deak P, Glover DM and Ohkura H. (1999). J. Kimura M, Kotani S, Hattori T, Sumi N, Yoshioka T, Cell Biol., 146, 1005 – 1018. Todokoro K and Okano Y. (1997). J. Biol. Chem., 272, Doxsey S. (2001). Nat. Rev. Mol. Cell. Biol., 2, 688 – 698. 13766 – 13771. Frank-Vaillant M, Haccard O, Thibier C, Ozon R, Arlot- Klein FA, Herr HW, Sogani PC, Whitmore WFJ and Bonnemains Y, Prigent C and Jessus C. (2000). J. Cell Sci., Melamed MR. (1982). Cancer, 50, 389 – 395. 113, 1127 – 1138. Kuo KK, Sato N, Mizumoto K, Maehara N, Yonemasu H, Geiser JR, Schott EJ, Kingsbury TJ, Cole NB, Totis LJ, Ker CG, Sheen PC and Tanaka M. (2000). Hepatology, 31, Bhattacharyya G, He L and Hoyt MA. (1997). Mol. Biol. 59 – 64. Cell., 8, 1035 – 1050. Lacey KR, Jackson PK and Stearns T. (1999). Proc. Natl. Gergely F, Karlsson C, Still I, Cowell J, Kilmartin J and Raff Acad. Sci. USA, 96, 2817 – 2822. JW. (2000a). Proc. Natl. Acad. Sci. USA, 97, 14352 – Lee MJ, Gergely F, Jeffers K, Peak-Chew SY and Raff JW. 14357. (2001). Nat. Cell. Biol., 3, 643 – 649. GergelyF,KiddD,JeffersK,WakefieldJGandRaffJW. Lengauer C, Kinzler KW and Vogelstein B. (1998). Nature, (2000b). EMBO J., 19, 241 – 252. 396, 643 – 649. Giet R and Glover DM. (2001). J. Cell Biol., 152, 669 – 682. Lingle WL, Lutz WH, Ingle JN, Maihle NJ and Salisbury JL. Giet R, McLean D, Descamps S, Lee MJ, Raff JW, Prigent C (1998). Proc. Natl. Acad. Sci. USA, 95, 2950 – 2955. and Glover DM. (2002). J. Cell Biol., 156, 437 – 451. Lingle WL and Salisbury JL. (1999). Am.J.Pathol.,155, Giet R and Prigent C. (1998). Exp. Cell. Res., 244, 470 – 473. 1941 – 1951. Giet R and Prigent C. (2001). J. Cell. Sci., 114, 2095 – 2104. Magennis DP. (1997). Br. J. Biomed. Sci., 54, 140 – 148. Giet R and Prigent C. (2000). Exp. Cell. Res., 258, 145 – 151. Matsumoto Y, Hayashi K and Nishida E. (1999). Curr. Biol., Giet R, Uzbekov R, Cubizolles F, Le Guellec K and Prigent 9, 429 – 432. C. (1999a). J. Biol. Chem., 274, 15005 – 15013. Mendez R, Hake LE, Andresson T, Littlepage LE, Ruder- Giet R, Uzbekov R, Kireev I and Prigent C. (1999b). Biol. man JV and Richter JD. (2000). Nature, 404, 302 – 307. Cell., 91, 461 – 470. Meraldi P, Honda R and Nigg EA. (2002). EMBO J., 21, Glover DM, Leibovitz MH, Mc Lean DA and Parry H. 483 – 492. (1995). Cell, 81, 95 – 105. Meraldi P, Lukas J, Fry AM, Bartek J and Nigg EA. (1999). Gopalan G, Chan CS and Donovan PJ. (1997). J. Cell. Biol., Nat. Cell. Biol., 1, 88 – 93. 138, 643 – 656. Nigg EA. (2001). Nat. Rev. Mol. Cell. Biol., 2, 21 – 32.

Oncogene Aurora and centrosome S Dutertre et al 6183 Petersen J, Paris J, Willer M, Philippe M and Hagan IM. Takahashi T, Futamura M, Yoshimi N, Sano J, Katada M, (2001). J. Cell Sci., 114, 4371 – 4384. Takagi Y, Kimura M, Yoshioka T, Okano Y and Saji S. Pihan GA, Purohit A, Wallace J, Knecht H, Woda B, (2000). Jpn. J. Cancer Res., 91, 1007 – 1014. Quesenberry P and Doxsey SJ. (1998). Cancer Res., 58, Tanaka T, Kimura M, Matsunaga K, Fukada D, Mori H and 3974 – 3985. Okano Y. (1999). Cancer Res., 59, 2041 – 2044. Roghi C, Giet R, Uzbekov R, Morin N, Chartrain I, Le Tanner MM, Tirkkonen M, Kallioniemi A, Collins C, Guellec R, Couturier A, Doree M, Philippe M and Prigent Stokke T, Karhu R, Kowbel D, Shadravan F, Hintz M C. (1998). J. Cell Sci., 111, 557 – 572. and Kuo WL. (1994). Cancer Res., 54, 4257 – 4260. Roof DM, Meluh PB and Rose MD. (1992). J. Cell. Biol., Tanner MM, Tirkkonen M, Kallioniemi A, Holli K, Collins 118, 95 – 108. C, Kowbel D, Gray JW, Kallioniemi OP and Isola J. Sakakura C, Hagiwara A, Yasuoka R, Fujita Y, Nakanishi (1995). Clin. Cancer Res., 1, 1455 – 1461. M, Masuda K, Shimomura K, Nakamura Y, Inazawa J, Tirkkonen M, Johannsson O, Agnarsson BA, Olsson H, Abe T and Yamagishi H. (2001). Br.J.Cancer,84, 824 – Ingvarsson S, Karhu R, Tanner M, Isola J, Barkardottir 831. RB, Borg A and Kallioniemi OP. (1997). Cancer Res., 57, Sasaki O, Kido K and Nagahama S. (1999). Anal. Quant. 1222 – 1227. Cytol. Histol., 21, 85 – 88. Tseng TC, Chen SH, Hsu YP and Tang TK. (1998). DNA Sassoon I, Severin FF, Andrews PD, Taba MR, Kaplan KB, Cell. Biol., 17, 823 – 833. Ashford AJ, Stark MJ, Sorger PK and Hyman AA. (1999). Walter AO, Seghezzi W, Korver W, Sheung J and Lees E. Genes Dev., 13, 545 – 555. (2000). Oncogene, 19, 4906 – 4916. Sato N, Mizumoto K, Nakamura M, Nakamura K, Weber RG, Bridger JM, Benner A, Weisenberger D, Kusumoto M, Niiyama H, Ogawa T and Tanaka M. Ehemann V, Reifenberger G and Lichter P. (1998). (1999). Clin. Cancer Res., 5, 963 – 970. Cytogenet. Cell Genet., 83, 266 – 269. Saunders WS and Hoyt MA. (1992). Cell, 70, 451 – 458. Whitehead CM and Rattner JB. (1998). J. Cell Sci., 111, Sawin KE, LeGuellec K, Philippe M and Mitchison TJ. 2551 – 2561. (1992). Nature, 359, 540 – 543. Whitehead CM, Winkfein RJ and Rattner JB. (1996). Cell Schumacher JM, Ashcroft N, Donovan PJ and Golden A. Motil. Cytoskeleton, 35, 298 – 308. (1998). Development, 125, 4391 – 4402. Zhou H, Kuang J, Zhong L, Kuo WL, Gray JW, Sahin A, SenS,ZhouHandWhiteRA.(1997).Oncogene, 14, 2195 – Brinkley BR and Sen S. (1998). Nat. Genet., 20, 189 – 193. 2200. Sturgis CD, Caraway NP, Johnston DA, Sherman SI, Kidd L and Katz RL. (1999). Cancer, 87, 155 – 160.

Oncogene