Cre-loxP-controlled periodic Aurora-A overexpression induces mitotic abnormalities and hyperplasia in mammary glands of mouse models

Article metrics

Abstract

Aurora-A, a serine/threonine mitotic kinase, was reported to be overexpressed in various human cancers, and its overexpression induces aneuploidy, centrosome amplification and tumorigenic transformation in cultured human and rodent cells. However, the underlying mechanisms and pathological settings by which Aurora-A promotes tumorigenesis are largely unknown. Here, we created a transgenic mouse model to investigate the involvement of Aurora-A overexpression in the development of mammary glands and tumorigenesis using a Cre-loxP system. The conditional expression of Aurora-A resulted in significantly increased binucleated cell formation and apoptosis in the mammary epithelium. The surviving mammary epithelial cells composed hyperplastic areas after a short latency. Induction of Aurora-A overexpression in mouse embryonic fibroblasts prepared from the transgenic mice also led to aberrant mitosis and binucleated cell formation followed by apoptosis. The levels of p53 protein were remarkably increased in these Aurora-A-overexpressing cells, and the apoptosis was significantly suppressed by deletion of p53. Given that no malignant tumor formation was found in the Aurora-A-overexpressing mouse model after a long latency, additional factors, such as p53 inactivation, are required for the tumorigenesis of Aurora-A-overexpressing mammary epithelium. Our findings indicated that this mouse model is a useful system to study the physiological roles of Aurora-A and the genetic pathways of Aurora-A-induced carcinogenesis.

Introduction

Chromosomal instability and aneuploidy are remarkable hallmarks of human cancers (Lengauer et al., 1998; Cahill et al., 1999). In most cancers, high rates of chromosome gains/losses leading to aneuploidy have been observed. The causes of aneuploidy involve failure in various critical mitotic events, including centrosome separation, chromosome alignment, chromosome segregation and completion of cytokinesis. The error-free mitosis that is important to genomic integrity is regulated by phosphorylation reactions driven by several evolutionarily conserved serine/threonine kinases, known as mitotic kinases. Mitotic kinases include cyclin-dependent kinase 1 (Cdk1), Polo-related, NimA-related, Aurora-related and Warts-related kinases (Nigg, 2001).

Mutations in genes encoding Aurora-related kinases induce abnormal mitotic phenotypes from yeast to higher eukaryotes (Bischoff and Plowman, 1999; Nigg, 2001). The founding members of this family are Ipl1p from Saccharomyces cerevisiae and Aurora from Drosophila melanogaster. Ipl1 mutants display missegregation of chromosomes and aneuploidy in S. cerevisiae (Chan and Botstein, 1993; Francisco and Chan, 1994). In Drosophila, mutations of Aurora alleles cause a mitotic arrest with characteristics of circular monopolar spindles around large centrosome and result in pupal lethality (Glover et al., 1995). In mammals, three members of this kinase family, Aurora-A, -B and –C, were identified. Recently, observations have revealed that Aurora-A kinase activity is required for various events during mitosis, such as G2–M transition, centrosome separation, chromosome alignment and cytokinesis (Hirota et al., 2003). Given that not only elevated expression of Aurora-A but also depletion of Aurora-A leads to mitotic failure and multinucleation, it is speculated that the proper timing and amplitude of Aurora-A expression is important for accurate chromosome segregation and fidelity of chromosome transmission (Meraldi et al., 2002; Anand et al., 2003; Marumoto et al., 2003).

The human Aurora-A gene is mapped to chromosome 20q13, a region frequently amplified in many human cancers (Kallioniemi et al., 1994; Schlegel et al., 1995). The amplification of Aurora-A gene has been found in approximately 12% of primary breast tumors and in many human tumor cell lines (Sen et al., 1997; Zhou et al., 1998), and Aurora-A mRNA and protein are frequently overexpressed in various human cancers, including breast, colorectal, pancreatic, ovarian and gastric cancers (Bischoff et al., 1998; Tanaka et al., 1999; Miyoshi et al., 2001; Sakakura et al., 2001; Gritsko et al., 2003; Li et al., 2003). Furthermore, overexpression of Aurora-A was reported to transform Rat1 fibroblasts, and these cells formed micronuclei that represented chromosome instability and grew as tumors in nude mice (Bischoff et al., 1998). Given that its overexpression leads to centrosome amplification, aneuploidy and tumorigenic transformation in mammalian cells, Aurora-A is considered to be a potential oncogene. Moreover, in recent studies of cancer predisposition, human Aurora-A and its mouse homologue were identified as candidate tumor-susceptibility genes (Ewart-Toland et al., 2003). These evidences strongly suggest that Aurora-A plays a role in the development of human malignant tumors. However, the underlying mechanisms and pathological settings by which Aurora-A promotes tumorigenesis are understood poorly.

In this study, we have generated a transgenic mouse model to investigate the involvement of Aurora-A overexpression in the development of mammary glands and tumorigenesis. Using a Cre-loxP system, we achieved the conditional expression of Aurora-A specifically in the mammary epithelium of adult mice during pregnancy and lactation. Elevated Aurora-A expression resulted in mitotic failure, leading to p53-dependent postmitotic G1 arrest and apoptosis. P53 function is potentially a crucial factor for suppressing Aurora-A-induced tumorigenesis.

Results

Mammary gland-specific expression of Aurora-A by a Cre-loxP system in transgenic mice

To study the function of Aurora-A in mouse mammary gland development, we used a Cre-loxP-mediated gene-switch approach. We initially generated transgenic mice carrying a construct of pCAG-loxP-CAT-loxP-Aurora-A (designated as CAG (chicken beta-actin)-CAT (chloramphenicol acetyltransferase)-Aurora-A mice), which was designed to induce human Aurora-A expression by Cre-mediated recombination. The CAG-CAT-Aurora-A transgenic mice developed normally without any abnormalities of the mammary gland or other tissues. To induce Aurora-A expression specifically in the mammary gland, the CAG-CAT-Aurora-A transgenic mice were bred with transgenic mice that carried Cre genes under the control of the Wap (whey acidic protein) gene promoter. The WAP-Cre transgene is expressed almost exclusively in mammary epithelial cells during pregnancy and lactation (Wagner et al., 1997). In the CAG-CAT-Aurora-A;Wap-Cre transgenic mice, Cre-mediated excision removes the CAT gene from the construct, allowing expression of the Aurora-A gene (Figure 1a). PCR analyses revealed that Wap-Cre-mediated recombination was detected in the mammary glands of late pregnant and lactating CAG-CAT-Aurora-A;Wap-Cre mice, while no recombination was found in other tissues, including tail, skin, lung, heart, kidney and spleen (Figure 1b). No recombination was detected in the mammary glands of virgin CAG-CAT-Aurora-A;Wap-Cre mice. However, a very low level of recombination was detected in brain tissue.

Figure 1
figure1

Structure of the pCAG-loxP-CAT-loxP-Aurora-A transgene and Wap-Cre-mediated recombination in CAG-CAT-Aurora-A;Wap-Cre mice. (a) Plasmid construction and production of transgenic mice. The conditional transgenic Aurora-A transgene consists of a CAG promotor, a loxP-flanked CAT gene, followed by human Aurora-A cDNA. Wap-Cre-mediated recombination excises the floxed CAT gene, resulting in the expression of Aurora-A. Arrows represent primers used in PCR recombination analysis. (b) PCR analysis of Wap-Cre-mediated recombination in CAG-CAT-Aurora-A;Wap-Cre mice. Top: The panels are the 750 bp P1/P2 Aurora-A PCR product in different tissues from 10-day lactating and 7-week-old virgin CAG-CAT-Aurora-A;Wap-Cre mice. Bottom: PCR recombination analysis. The nonrecombined P3/P2 PCR product is 2.8 kb (arrowhead) and the recombined is 1.2 kb (arrow). T, tail; Sk, skin; L, lung; H, heart; K, kidney; Sp, spleen; M, mammary gland; B, brain

The Cre-mediated Aurora-A expression in mammary glands was confirmed by RT–PCR and Western blot analyses (Figure 2a and b). Aurora-A mRNA and protein were expressed specifically in the mammary glands of lactating mice, but not in virgin mice. The immunohistochemical analysis also revealed that Aurora-A was specifically expressed in mammary gland epithelial cells of CAG-CAT-Aurora-A;Wap-Cre mice for a long period of time (Figure 2c). These data indicate that the CAG-CAT-Aurora-A;Wap-Cre mice are suitable for analysing role of Aurora-A in mammary gland development and tumorigenesis.

Figure 2
figure2

Wap-Cre-mediated Aurora-A expression in the mammary gland. (a) Expression of Aurora-A mRNA in CAG-CAT-Aurora-A;Wap-Cre mice. Aurora-A mRNA was detected in tissues taken from 7-week-old virgin and 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice by RT–PCR analysis with a pair of primers, P1 and P2, that yield a 750 bp product. Aurora-A mRNA was found specifically in the mammary gland (M) but not in the tail (T) of a 10-day lactating mouse. Control (β-actin) RT–PCR was performed in the same reaction mixtures. (b) Expression of Aurora-A protein in the mammary glands of CAG-CAT-Aurora-A;Wap-Cre mice by Western blot analysis. The proteins were taken from mammary glands of two 7-week-old virgin (lanes 1 and 2, control) and two 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice (lanes 3 and 4). α-tubulin was used as a loading control. Aurora-A overexpression was detected in lactating mice following the Cre activity (lanes 3 and 4). (c) Immunohistochemical analyses for Aurora-A protein expression in the mammary glands of a 10-day lactating CAG-CAT-Aurora-A mouse and CAG-CAT-Aurora-A;Wap-Cre mice at 10 days (10d), 30 days (30d), 90 days (90d) and 180 days (180d) after the first lactation. We stained the paraffin-embedded mammary gland sections using anti-human Aurora-A antibodies at a dilution of 1 : 1500. Scale bars, 100 μm

Aurora-A overexpression results in the formation of multinucleated cells in mammary gland

To investigate the effect of Aurora-A overexpression in mammary gland development, we compared histopathological findings between the CAG-CAT-Aurora-A (control) and the CAG-CAT-Aurora-A;Wap-Cre mice after the first pregnancy. In both control and CAG-CAT-Aurora-A;Wap-Cre mice, normal morphological changes due to pregnancy were observed. These included an increased number of acini per lobule, vacuoles within the epithelial cell cytoplasm, secretory material within distended lumens and a hobnail appearance. These changes were not seen in most mice 12 months after the first pregnancy, when lobular regression was observed. The CAG-CAT-Aurora-A;Wap-Cre mice at 3 months after the first pregnancy frequently showed focal hyperplastic lesions having prominent secretions, a slight loss of cellular cohesion and foci of cellular aggregates (Figure 3a), although these appeared to be different from the hyperplastic alveolar nodules found in MMTV-infected mice (Cardiff and Wellings, 1999). Although hyperplastic foci were observed frequently in mammary glands that overexpressed Aurora-A, malignant tumor formation has not been found in 14 CAG-CAT-Aurora-A;Wap-Cre mice or 10 control CAG-CAT-Aurora-A mice over long periods of follow-up (longer than 15 months after the first pregnancy).

Figure 3
figure3

Hyperplastic changes and binucleated cell formation in the mammary glands of CAG-CAT-Aurora-A;Wap-Cre mice. (a) Histological sections of mammary gland tissues from a 5-month-old lactating CAG-CAT-Aurora-A mouse (control, left) and a lactating CAG-CAT-Aurora-A;Wap-Cre mouse 3 months after the first pregnancy (right). The arrow indicates the area of papillary hyperplasia. Scale bars, 50 μm. (b) Hematoxylin and eosin staining of a paraffin-embedded mammary gland section from a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse. Scale bar, 20 μm. Note that cell aggregates and binucleated epithelial cells were frequently observed. (c) Confocal scans of mammary gland sections from a 10-day lactating CAG-CAT-Aurora-A mouse (control, left) and a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse (right). The tissues were stained with anti-E-cadherin antibodies (green), which highlight cell–cell contact regions. DNA was stained with ToTo-3 (blue). The multinucleated epithelial cells are indicated by arrowheads. Scale bars, 20 μm. (d) Graph showing the increased percentage of multinucleated cells in mammary glands of 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice contrasted to CAG-CAT-Aurora-A mice. Values shown are the means from three animals per experimental group in which over 1000 cells were counted

Histopathological analysis of 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice frequently identified foci of cell aggregates and binucleated cell formation (Figure 3b).To confirm whether binucleated cells increase in mammary glands overexpressing Aurora-A, we examined the mammary epithelium using immunofluorescent analysis for E-cadherin counterstained with ToTo-3 to label DNA (Figure 3c). We observed that a significant population of the multinucleated mammary epithelial cells was generated in 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice (11.9±2.0%), which is 13-fold greater than control (CAG-CAT-Aurora-A) mice (0.9±1.0%). Interestingly, most of the multinucleate cells contained two nuclei. The cells containing three nuclei or more are a very small portion of the whole (Figure 3d). These results suggest that upregulation of Aurora-A may result in cytokinesis failure, leading to the formation of cells with two nuclei in vivo.

Elevated Aurora-A expression induces apoptosis

Histopathological analysis of the lactating CAG-CAT-Aurora-A;Wap-Cre mice revealed that chromatin appears condensed beneath the nuclear membrane in a significant number of mammary gland cells (data not shown), suggesting that Aurora-A overexpression induces apoptosis in mammary epithelial cells. To verify this hypothesis, we performed TdT-mediated dUTP-biotin nick-end labeling (TUNEL) assays (Figure 4a–c). Given that mammary epithelial cells are known to undergo programmed cell death at the end of lactation (Lund et al., 1996), we investigated transgenic mice that were in the early phase of lactation. In 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice, a significant number of apoptotic cells (14.9±2.3%) was observed among mammary epithelial cells while a very small number of apoptotic cells (0.4±0.1%) were detected in control mice (Figure 4d). These findings indicate that overexpression of Aurora-A leads to not only mitotic failure but also to apoptosis in mammary epithelium.

Figure 4
figure4

Aurora-A overexpression induces apoptosis in mammary epithelial cells. (a–c) TUNEL staining for apoptotic cells of mammary gland tissues from a 10-day lactating CAG-CAT-Aurora-A mouse (a, control) and a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse (b). A higher magnification of the section of (b) is shown in (c). Scale bars, 50 μm. (d) Graph showing the increased percentage of apoptotic epithelial cells in mammary glands of CAG-CAT-Aurora-A;Wap-Cre mice in contrast to CAG-CAT-Aurora-A mice. Values shown are the means from three animals per experimental group in which over 1000 cells were counted

Induction of Aurora-A overexpression leads to mitotic abnormalities and apoptosis in mouse embryonic fibroblasts (MEFs)

The effects of Aurora-A overexpression on mitotic progression and cell death that we observed in mammary epithelial cells of CAG-CAT-Aurora-A;Wap-Cre mice were confirmed using immortalized MEFs established from CAG-CAT-Aurora-A mice. The growth rate of CAG-CAT-Aurora-A MEFs was indistinguishable from that of MEFs derived from normal C57BL/6 mice (data not shown). To induce Aurora-A overexpression, CAG-CAT-Aurora-A MEFs were infected with AxCANCre (Kanegae et al., 1995), an adenovirus that encodes the Cre enzyme with an artificial nuclear localization signal. The Aurora-A protein was first detected at 12 h after the virus infection, and their protein levels peaked at 48 h after infection (Figure 5a). Induction of Aurora-A overexpression resulted in significantly increased numbers of cells having two nuclei (Figure 5b). The number of CAG-CAT-Aurora-A MEFs with two nuclei increased to more than 10% at 72 h after virus infection (Figure 5c). Staining by propidium iodide showed frequent abnormal mitotic events after Aurora-A overexpression, including chromosome misalignment and chromosome missegregation (Figure 5d).

Figure 5
figure5

Mitotic abnormalities and apoptosis in Aurora-A-overexpressing MEFs. (a) MEFs from CAG-CAT-Aurora-A mice were infected with adenoviral luciferase (Luc) or AxCANCre (Cre) virus. Aurora-A overexpression was observed 24 h after virus infection by Western blot analysis. (−), untreated. α-tubulin was used as a loading control. (b) Confocal microscopic analysis shows that multinucleated cells were observed at 48 h after AxCANCre (Cre) virus infection in MEFs. The MEFs were stained with anti-α-tubulin antibodies (green), counterstained by propidium iodide (red). Scale bars, 20 μm. (c) The number of nuclei of control MEFs (from normal C57BL/6 mice) and MEFs from CAG-CAT-Aurora-A mice, before and 72 h after virus infection (Luc or Cre), were counted by propidium iodide staining. Values shown are the percentage of binucleated cells from three independent experiments in which over 300 cells were counted. (d) Chromosome misalignment (upper panels) and chromosome missegregation (lower panels) were observed in CAG-CAT-Aurora-A MEFs at 48 h after AxCANCre virus infection. (e) Aurora-A overexpression induces hyperploidy and apoptosis in MEFs from CAG-CAT-Aurora-A mice. MEFs were infected with AxCANCre (Cre) and adenoviral luciferase (Luc). DNA contents were measured by FACScan analysis at 0, 48 and 72 h after virus infection. Percentages of cells at G1/S, G2/M and hyperploid cells were calculated. Apoptosis was scored using propidium iodide-based FACScan analysis to quantitate cells with sub-G1 DNA content. Data are means±s.d. of values from three independent experiments

Fluorescence-activated cell sorting (FACS) analysis revealed that the proportion of cells at G1/S was reduced while G2/M cells were increased up to approximately 50% for Aurora-A-overexpressing MEFs compared with control cells (CAG-CAT-Aurora-A MEFs infected with adenoviral luciferase) at 48 h after virus infection (Figure 5e). Furthermore, overexpression of Aurora-A induced a marked reduction in the proportion of MEFs at G2/M that was also concomitant with an increase in the size of the sub-G1 population at 72 h after virus infection, suggesting that a large number of cells at G2/M phase underwent cell death as a result of the prolonged expression of Aurora-A (Figure 5e). Based on these findings, we hypothesized that the binucleated cells induced by Aurora-A overexpression tend to undergo apoptosis. To address this issue, the fate of Aurora-A-overexpressing binucleated MEFs was followed by the time-lapse microscopy. At 72 h after AxCANCre infection, most of the binucleated CAG-CAT-Aurora-A MEFs died (Supplementary movie 1). Taken together, these data indicate that overexpression of Aurora-A induces the formation of binucleated cells and subsequently leads to apoptosis.

Overexpression of Aurora-A induces p53-dependent apoptosis

It was shown previously that disruption of chromosome segregation during mitosis or failure of cytokinesis activates a p53-dependent checkpoint that normally acts in postmitotic G1 to arrest tetraploid cells. Thus, we speculated that mitotic errors induced by Aurora-A overexpression may activate the postmitotic checkpoint, resulting in the accumulation of cells arrested in tetraploid G1, followed by cell death. To test this hypothesis, we analysed p53 expression in CAG-CAT-Aurora-A;Wap-Cre mice. Levels of p53 protein detected by Western blot analysis were remarkably increased in mammary epithelium of the lactating CAG-CAT-Aurora-A;Wap-Cre mice, but not in other tissues of the CAG-CAT-Aurora-A;Wap-Cre mice and the mammary glands of lactating CAG-CAT-Aurora-A and Wap-Cre mice (Figure 6a). Immunohistochemical analysis also showed that p53 is accumulated in nuclei of the Aurora-A-overexpressing mammary epithelial cells, especially in the binucleated cells (Figure 6b). Moreover, when Aurora-A overexpression was induced in the CAG-CAT-Aurora-A MEFs by AxCANCre infection, the nuclear accumulation of p53 was specifically detected in the binucleated cells (Figure 6c). Based on these results, we speculated that p53 is responsible for the increased apoptosis in the mammary glands that overexpress Aurora-A. To test this hypothesis, we generated CAG-CAT-Aurora-A;Wap-Cre mice carrying null alleles for p53 (CAG-CAT-Aurora-A;Wap-Cre;p53−/− mice). TUNEL staining of the mammary glands of the 10-day lactating mice revealed that apoptosis was significantly suppressed in the Aurora-A-overexpressing transgenic mice that lack p53 (Figure 6d and e). These data indicate that p53 plays a crucial role in postmitotic G1 arrest and subsequent apoptosis of the Aurora-A-overexpressing normal cells.

Figure 6
figure6

Aurora-A overexpression induces p53-dependent apoptosis. (a) Western blot analysis of p53 in tissues from a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse, a 10-day lactating CAG-CAT-Aurora-A mouse and a 10-day lactating Wap-Cre mouse. α-tubulin was used as a loading control. H, heart; K, kidney; Sp, spleen; L, liver; M, mammary gland; C, colon. (b) Immunohistochemical analysis for p53 in the mammary gland of a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse. A frozen tissue section was immunostained by anti-p53 antibody. P53 accumulation was observed in binucleated mammary epithelial cells (Arrows). (c) Specific p53 accumulation in Aurora-A-overexpressing binucleated MEF. MEFs from CAG-CAT-Aurora-A mice were infected with AxCANCre virus and were then subjected to immunofluorescence analysis for p53 at 72 h after infection. DNA was labeled by 4′,6-diamidino-2-phenylindole (DAPI). (d) TUNEL staining for apoptotic cells of mammary gland tissues from a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mouse (upper) and a 10-day lactating CAG-CAT-Aurora-A;Wap-Cre;p53−/− mouse (lower). Scale bars, 50 μm. (e) Suppression of apoptosis in CAG-CAT-Aurora-A;Wap-Cre;p53−/− mice. The percentage of TUNEL-positive apoptotic cells in mammary glands was significantly lower in 10-day lactating CAG-CAT-Aurora-A;Wap-Cre;p53−/− mice than in CAG-CAT-Aurora-A;Wap-Cre mice. Values shown are the means from three animals per experimental group in which over 1000 cells were counted

Discussion

Mitotic kinases are key regulators of mitotic progression, and the timing of expression and activity of those kinases are regulated precisely during the cell cycle to maintain genomic integrity in mammalian cells. Aurora-A kinase is localized mainly at the centrosome and mitotic apparatus in normally proliferating cells, and the expression and activity of Aurora-A are controlled to peak during late G2 to M phase, similarly to other mitotic kinases (Bischoff et al., 1998; Marumoto et al., 2002; Hirota et al., 2003; Marumoto et al., 2003). However, immunohistochemical analyses of clinical samples have revealed that various epithelial cancer cells overexpress Aurora-A, which was stained diffusely in the cytoplasm of both interphase and mitotic cells (Tanaka et al., 1999; Gritsko et al., 2003; Li et al., 2003). In addition to these pathological observations, evidence that overexpression of Aurora-A over-rides the cell cycle checkpoint (Marumoto et al., 2002; Anand et al., 2003) and induces transformation in immortalized rodent fibroblasts (Bischoff et al., 1998; Zhou et al., 1998) suggests that dysregulation of Aurora-A expression and activity is a direct cause of sporadic malignant tumors. To address this issue, we developed a transgenic mouse that conditionally overexpresses Aurora-A in adult mammary glands using the Cre-loxP recombination system under the control of the Wap promoter.

Mammary glands that overexpress Aurora-A showed the following findings: (1) in 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice, the numbers of binucleated cells and cell aggregates were increased significantly compared to control (CAG-CAT-Aurora-A) mice; (2) in 10-day lactating CAG-CAT-Aurora-A;Wap-Cre mice, apoptosis of mammary epithelial cells was increased; and (3) although hyperplastic lesions frequently were found in mammary glands of the CAG-CAT-Aurora-A;Wap-Cre mice 3 months after pregnancy, no pathological findings indicating malignant transformation have been found. The conditional expression system employed in the present study provides insight into the importance of Aurora-A function in mitosis, apoptosis and tumorigenesis.

Downregulation of Aurora-A is required for completion of mitosis

It has been shown previously that Aurora-A activation is required for multiple events in mitosis, such as G2–M transition, centrosome maturation, centrosome separation, metaphase chromosome alignment and cytokinesis (Zhou et al., 1998; Hirota et al., 2003; Marumoto et al., 2003). We have reported that microinjection of anti-Aurora-A antibodies into metaphase cells that had completed centrosome separation and metaphase chromosome alignment subsequently fail to complete cytokinesis, suggesting that Aurora-A needs to be active for regulating cytokinesis even after the metaphase–anaphase transition (Marumoto et al., 2003). However, the present study showed that overexpression of Aurora-A also induces cytokinesis failure, resulting in binucleated cell formation in both mammary epithelial cells and cultured MEFs. This finding is consistent with the recent observation, using time-lapse analysis, that MEFs transfected with Aurora-A frequently fail to undergo cytokinesis (Anand et al., 2003). Therefore, the activation and the subsequent inactivation of Aurora-A may be required for completion of cytokinesis. The levels of Aurora-A protein are regulated by proteasome-mediated degradation. Aurora-A is reportedly degraded in vitro in late mitosis/early G1 by the Cdh1/Fizzy-related form of the anaphase-promoting complex/cyclosome (APC/C) (Honda et al., 2000; Castro et al., 2002). Thus, the rapid degradation of Aurora-A at late anaphase–telophase may be an important event for completion of cytokinesis. The molecular mechanism by which Aurora-A regulates cytokinesis is currently under investigation. Our laboratory is attempting to identify certain specific substrates that are phosphorylated by Aurora-A at metaphase–anaphase transition and dephosphorylated toward the end of M phase.

Aurora-A overexpression induces p53-dependent postmitotic G1 checkpoint and subsequent apoptosis in mammary epithelial cells and MEFs

We have shown here that induction of Aurora-A overexpression elicits binucleated cell formation and that those cells have p53 accumulation. It was previously reported that the replication of DNA in tetraploid cells that have entered the subsequent G1 phase without cell division is usually blocked by p53- and pRB-dependent cell cycle arrest, which is referred to as the postmitotic G1 checkpoint (Borel et al., 2002; Margolis et al., 2003). Therefore, cytokinesis failure induced by Aurora-A overexpression may trigger the p53-dependent postmitotic G1 checkpoint to avoid further cell cycle progression. However, recent reports have indicated that cytokinesis failure alone is not sufficient to activate the postmitotic ‘tetraploidy’ checkpoint (Uetake and Sluder, 2004) and that activation of spindle checkpoint is required for this postmitotic checkpoint activation (Vogel et al., 2004). It is thus possible that Aurora-A overexpression may induce not only cytokinesis failure but also spindle checkpoint activation, leading to the activation of p53-dependent postmitotic G1 checkpoint.

We have also observed that Aurora-A overexpression induces apoptosis in both mammary epithelial cells and cultured MEFs. Several studies have indicated that expression of activated oncogene products, such as myc, generally induce apoptotic cell death in MEFs (Evan et al., 1992; Hermeking and Eick, 1994; Wagner et al., 1994; Baudino et al., 2003). Myc-overexpressing MEFs proliferated less well and had a considerably higher apoptotic index (10–15% TUNEL positive) (Zindy et al., 1998). Apoptotic cell death induced by activated myc expression in normal cells has been demonstrated to be p53 dependent. Similarly to the experience in myc-overexpressing cells, we have found that p53 levels were increased markedly in Aurora-A-overexpressing mammary epithelial cells and MEFs, and that apoptosis was suppressed significantly in the mammary tissues that lack p53 expression. These observations suggest that Aurora-A overexpression induces p53-dependent apoptosis. Recent studies have shown that Wap-Cre-mediated conditional mutation of BRCA1 in mammary epithelial cells results in increased apoptosis, with a low frequency of tumor formation which is accelerated in a p53-null background (Xu et al., 1999). Taken together, activation of p53 appears to play an important role in prevention of oncogene- and antioncogene-associated tumorigenesis.

Aurora-A, p53 and tumorigenesis

Overexpression of Aurora-A promotes changes in ploidy and generates focal hyperplastic lesions in mammary epithelium. This implies that overfunction of Aurora-A induces mitotic abnormalities and consequent chromosome instability, contributing to tumorigenesis. However, no malignant tumor formation was found in mammary glands overexpressing Aurora-A over 15 months of observation, indicating that not only Aurora-A overexpression but also further changes are required for tumor formation.

Recent studies have revealed direct relationships between Aurora-A and p53. It was reported that p53 interacts with Aurora-A and suppresses the oncogenic activity of Aurora-A, such as centrosome amplification and cellular transformation (Chen et al., 2002). These data may partly explain why Aurora-A overexpression alone does not induce the tumor formation in a short period of the time of observation. On the other hand, another recent study has demonstrated that the Aurora-A kinase directly phosphorylates p53 at Ser 315, facilitating Mdm2-mediated ubiquitination and destabilization of p53 in cancer cell lines, such as H1299 and MCF7 (Katayama et al., 2004). This finding suggests that Aurora-A overexpression suppresses the p53 tumor suppressor function, which seemingly contradicts our data. These apparently conflicting observations regarding the role of Aurora-A in p53 expression might be due to differences between normal and cancer cells. Alternatively, they may be attributable to differences in the mode of p53 regulation between normal and cancer cells.

Based on our observations, we postulate that p53 is an important factor that inhibits tumor progression in Aurora-A-overexpressing mammary glands. Loss of the p53 function can facilitate genetic instability by leading cells to over-ride the apoptotic pathway and continue through subsequent cell cycles and is required for Aurora-A to induce tumorigenesis. Further work may benefit from a focus on proving the correlation of p53 and Aurora-A in mammary tumorigenesis by long-term follow-up of CAG-CAT-Aurora-A;Wap-Cre;p53−/− mice.

Materials and methods

Plasmid construction and production of CAG-CAT-Aurora-A transgenic mice

The transgene vector pCAG-CAT-Aurora-A, which contains a CAG gene promotor-loxP-CAT gene-loxP-Aurora-A region, was constructed from pCAG-loxP-CAT-loxp-lacZ by replacing the lacZ gene with human Aurora-A cDNA. The pCAG-CAT-lacZ plasmid was a gift from Kimi Araki (Araki et al., 1995). The construct was purified by electrophoresis and elution from NACS PREPAC (BRL) and used for microinjection. The CAG-CAT-Aurora-A transgenic mice were produced as described elsewhere (Hogan et al., 1994).

Generation of CAG-CAT-Aurora-A;Wap-Cre transgenic mice and CAG-CAT-Aurora-A;Wap-Cre;p53−/− mice

CAG-CAT-Aurora-A mice (strain C57BL/6) were mated with Wap-Cre mice to generate CAG-CAT-Aurora-A;Wap-Cre double transgenic mice. The F1 offspring that carried the two transgenes were genotyped by PCR using following primers. The primers for detecting Aurora-A were P1 (5′-IndexTermAAAGAGCAAGCAGCCCCTGC-3′) and P2 (5′-IndexTermGAATTCAACCCGTGATATTCTT-3′), and yielded a 750 bp product. The primers for detecting Cre were Cre1 (5′-IndexTermAGGTTCGTTCACTCATGGA-3′) and Cre2 (5′-IndexTermTCGACCAGTTTAGTTACCC-3′), and yielded a 235 bp product. The PCR cycling profile consisted of 35 cycles of 30 s at 95°C, 30 s at 56°C and 45 s at 72°C. CAG-CAT-Aurora-A;Wap-Cre female mice were mated at 8 weeks of age and kept with males for continuous mating. Cre-mediated recombination was detected by PCR analysis. Another P3 primer (5′-IndexTermCTGCTAACCATGTTCATGCC-3′) was used to detect Cre-mediated recombination. The recombined P3/P2 PCR product was 1.2 kb and the nonrecombined 2.8 kb.

CAG-CAT-Aurora-A;Wap-Cre transgenic mice carrying null alleles for p53 were typically generated from crosses of CAG-CAT-Aurora-A;p53+/− mice and Wap-Cre;p53+/− mice. Mice carrying a null allele of p53 were obtained (Tsukada et al., 1993) and back-crossed seven generations into a C57BL/6 background prior to crosses with the CAG-CAT-Aurora-A and Wap-Cre transgenic lines.

Histology and immunofluorescence

For conventional histological analysis, mammary glands were fixed in 10% phosphate-buffered formaldehyde for 24 h and embedded in paraffin. Thick sections (3 μm) were stained with hematoxylin and eosin using standard techniques. For cryosections, tissues were fixed with 4% paraformaldehyde for 6 h, then transferred to a 30% sucrose solution for 24 h before 8-μm sections were prepared. Immunohistochemical analysis was performed following the manufacturer's manual (Nichirei, HISTOFINE SAB-PO(R) kit). Immunofluoresence was performed as follows: sections were washed once with PBS, permeabilized with 0.5% Triton X-100 in PBS for 1 h, blocked for 1 h with 5% BSA in PBS and incubated with primary antibodies in 0.3% BSA in PBS. A rat monoclonal antibody to E-cadherin (Takara) was used at a dilution of 1 : 1000. Fluorescein isothiocyanate (FITC)-conjugated secondary antibodies (Biosource) were used as secondary antibodies. Slides were counterstained with ToTo-3 (Molecular Probes) prior to being mounted under glass coverslips and analysed by confocal microscopy (FV300, Olympus).

Additional materials and methodology including references can be found in ‘Supplementary materials and methods’.

Abbreviations

CAG:

chicken beta-actin

CAT:

chloramphenicol acetyltransferase

TUNEL:

TdT-mediated dUTP-biotin nick-end labeling

MEFs:

mouse embryonic fibroblasts

FITC:

fluorescein isothiocyanate

FACS:

fluorescence-activated cell sorting

References

  1. Anand S, Penrhyn-Lowe S and Venkitaraman AR . (2003). Cancer Cell, 3, 51–62.

  2. Araki K, Araki M, Miyazaki J and Vassalli P . (1995). Proc. Natl. Acad. Sci. USA, 92, 160–164.

  3. Baudino TA, Maclean KH, Brennan J, Parganas E, Yang C, Aslanian A, Lees JA, Sherr CJ, Roussel MF and Cleveland JL . (2003). Mol. Cell, 11, 905–914.

  4. Bischoff JR, Anderson L, Zhu Y, Mossie K, Ng L, Souza B, Schryver B, Flanagan P, Clairvoyant F, Ginther C, Chan CS, Novotny M, Slamon DJ and Plowman GD . (1998). EMBO J., 17, 3052–3065.

  5. Bischoff JR and Plowman GD . (1999). Trends Cell Biol., 9, 454–459.

  6. Borel F, Lohez OD, Lacroix FB and Margolis RL . (2002). Proc. Natl. Acad. Sci. USA, 99, 9819–9824.

  7. Cahill DP, Kinzler KW, Vogelstein B and Lengauer C . (1999). Trends Cell Biol., 9, M57–M60.

  8. Cardiff RD and Wellings SR . (1999). J. Mammary Gland Biol. Neoplasia, 4, 105–122.

  9. Castro A, Arlot-Bonnemains Y, Vigneron S, Labbe JC, Prigent C and Lorca T . (2002). EMBO Rep., 3, 457–462.

  10. Chan CS and Botstein D . (1993). Genetics, 135, 677–691.

  11. Chen SS, Chang PC, Cheng YW, Tang FM and Lin YS . (2002). EMBO J., 21, 4491–4499.

  12. Evan GI, Wyllie AH, Gilbert CS, Littlewood TD, Land H, Brooks M, Waters CM, Penn LZ and Hancock DC . (1992). Cell, 69, 119–128.

  13. Ewart-Toland A, Briassouli P, de Koning JP, Mao JH, Yuan J, Chan F, MacCarthy-Morrogh L, Ponder BA, Nagase H, Burn J, Ball S, Almeida M, Linardopoulos S and Balmain A . (2003). Nat. Genet., 34, 403–412.

  14. Francisco L and Chan CS . (1994). Cell Mol. Biol. Res., 40, 207–213.

  15. Glover DM, Leibowitz MH, McLean DA and Parry H . (1995). Cell, 81, 95–105.

  16. Gritsko TM, Coppola D, Paciga JE, Yang L, Sun M, Shelley SA, Fiorica JV, Nicosia SV and Cheng JQ . (2003). Clin. Cancer Res., 9, 1420–1426.

  17. Hermeking H and Eick D . (1994). Science, 265, 2091–2093.

  18. Hirota T, Kunitoku N, Sasayama T, Marumoto T, Zhang D, Nitta M, Hatakeyama K and Saya H . (2003). Cell, 114, 585–598.

  19. Hogan B, Beddington R, Costantini F and Lacy E . (1994). Manipulating the Mouse Embryo: A Laboratory Manual 2nd edn. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY.

  20. Honda K, Mihara H, Kato Y, Yamaguchi A, Tanaka H, Yasuda H, Furukawa K and Urano T . (2000). Oncogene, 19, 2812–2819.

  21. Kallioniemi A, Kallioniemi OP, Piper J, Tanner M, Stokke T, Chen L, Smith HS, Pinkel D, Gray JW and Waldman FM . (1994). Proc. Natl. Acad. Sci. USA, 91, 2156–2160.

  22. Kanegae Y, Lee G, Sato Y, Tanaka M, Nakai M, Sakaki T, Sugano S and Saito I . (1995). Nucleic Acids Res., 23, 3816–3821.

  23. Katayama H, Sasai K, Kawai H, Yuan ZM, Bondaruk J, Suzuki F, Fujii S, Arlinghaus RB, Czerniak BA and Sen S . (2004). Nat. Genet., 36, 55–62.

  24. Lengauer C, Kinzler KW and Vogelstein B . (1998). Nature, 396, 643–649.

  25. Li D, Zhu J, Firozi PF, Abbruzzese JL, Evans DB, Cleary K, Friess H and Sen S . (2003). Clin. Cancer Res., 9, 991–997.

  26. Lund LR, Romer J, Thomasset N, Solberg H, Pyke C, Bissell MJ, Dano K and Werb Z . (1996). Development, 122, 181–193.

  27. Margolis RL, Lohez OD and Andreassen PR . (2003). J. Cell Biochem., 88, 673–683.

  28. Marumoto T, Hirota T, Morisaki T, Kunitoku N, Zhang D, Ichikawa Y, Sasayama T, Kuninaka S, Mimori T, Tamaki N, Kimura M, Okano Y and Saya H . (2002). Genes Cells, 7, 1173–1182.

  29. Marumoto T, Honda S, Hara T, Nitta M, Hirota T, Kohmura E and Saya H . (2003). J. Biol. Chem., 278, 51786–51795.

  30. Meraldi P, Honda R and Nigg EA . (2002). EMBO J., 21, 483–492.

  31. Miyoshi Y, Iwao K, Egawa C and Noguchi S . (2001). Int. J. Cancer, 92, 370–373.

  32. Nigg EA . (2001). Nat. Rev. Mol. Cell. Biol., 2, 21–32.

  33. Sakakura C, Hagiwara A, Yasuoka R, Fujita Y, Nakanishi M, Masuda K, Shimomura K, Nakamura Y, Inazawa J, Abe T and Yamagishi H . (2001). Br. J. Cancer, 84, 824–831.

  34. Schlegel J, Stumm G, Scherthan H, Bocker T, Zirngibl H, Ruschoff J and Hofstadter F . (1995). Cancer Res., 55, 6002–6005.

  35. Sen S, Zhou H and White RA . (1997). Oncogene, 14, 2195–2200.

  36. Tanaka T, Kimura M, Matsunaga K, Fukada D, Mori H and Okano Y . (1999). Cancer Res., 59, 2041–2044.

  37. Tsukada T, Tomooka Y, Takai S, Ueda Y, Nishikawa S, Yagi T, Tokunaga T, Takeda N, Suda Y, Abe S, Mastuo I, Ikawa Y and Aizawa S . (1993). Oncogene, 8, 3313–3322.

  38. Uetake Y and Sluder G . (2004). J. Cell Biol., 165, 609–615.

  39. Vogel C, Kienitz A, Hofmann I, Muller R and Bastians H . (2004). Oncogene, 23, 6845–6853.

  40. Wagner AJ, Kokontis JM and Hay N . (1994). Genes Dev., 8, 2817–2830.

  41. Wagner KU, Wall RJ, St-Onge L, Gruss P, Wynshaw-Boris A, Garrett L, Li M, Furth PA and Hennighausen L . (1997). Nucleic Acids Res., 25, 4323–4330.

  42. Xu X, Wagner KU, Larson D, Weaver Z, Li C, Ried T, Hennighausen L, Wynshaw-Boris A and Deng CX . (1999). Nat. Genet., 22, 37–43.

  43. Zhou H, Kuang J, Zhong L, Kuo WL, Gray JW, Sahin A, Brinkley BR and Sen S . (1998). Nat. Genet., 20, 189–193.

  44. Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Sherr CJ and Roussel MF . (1998). Genes Dev., 12, 2424–2433.

Download references

Acknowledgements

We thank Dr Kimi Araki (Kumamoto University) for providing pCAG-CAT-lacZ plasmid; Mr Takenobu Nakagawa (Kumamoto University) for technical assistance; Drs Izumu Saito and Yumi Kanegae (University of Tokyo) for providing adenoviral luciferase and AxCANCre virus; members of the Saya lab for valuable suggestions; and members of the Gene Technology Center at Kumamoto University for their contributions to the technical assistance. This work was supported by the Research for the Future program of the Japan Society for the promotion of Science and by a grant for Cancer Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (to HS).

Author information

Correspondence to Hideyuki Saya.

Additional information

Supplementary Information accompanies the paper on Oncogene website (http://www.nature.com/onc).

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Keywords

  • Aurora-A
  • mitosis
  • mammary gland
  • p53
  • Cre-loxP

Further reading