Induction of polyploidy and apoptosis after exposure to high ...

Mutagenesis vol.14 no.5 pp.513–520, 1999

Induction of polyploidy and apoptosis after exposure to high concentrations of the spindle poison nocodazole

Berlinda Verdoodt, Ilse Decordier, Karen Geleyns, Mońica Cunha, Enrico Cundari and Micheline Kirsch-Volders1 Laboratory for Cell Genetics, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussels, Belgium

The proportions of aneuploid/polyploid versus euploid cells formed after treatment with spindle poisons like nocodazole are of course dependent on the relative survival of cells with numerical chromosome aberrations. This work aimed at studying the survival of polyploid cells formed after treatment with a nocodazole concentration sufficient to significantly decrease tubulin polymerization (0.1 µg/ml). First, normal primary lymphocytes were analysed and the following complementary chromosomal parameters were quantified: mitotic index, frequency of abnormal mitoses, polyploid metaphases and apoptotic cells. The results clearly indicate a positive correlation between abnormal mitotic figures, apoptosis and the induction of polyploidy. They therefore led to a single cell approach in which both apoptosis and polyploidy induction could be scored in the same cell. For this purpose, actively proliferating cells are required and two human leukaemic cell lines were used, KS (p53-positive) and K562 (p53-negative), which have a near-triploid karyotype. Cells were separated into an apoptotic and a viable fraction by means of annexinV staining and flow cytometry. In KS, treatment with nocodazole induced a similar fraction of hexaploid cells in both the viable and apoptotic fraction, but no dodecaploid cells were ever observed. In contrast, a population of dodecaploid cells (essentially viable) was clearly observed in the K562 cell line. The results in KS, as compared with K562, confirm that wild-type p53 can prevent further cycling of polyploid cells by blocking rereplication. The most probable explanation for these data is that not only the mitotic spindle but also interphase microtubules are sensitive to nocodazole treatment. Our data thus strongly suggest that besides the G1/S checkpoint under the control of p53, the G2/M transition may be sensitive to depolymerization of microtubules, possibly under the control of Cdc2, Bcl-2, Raf-1 and/or Rho.

Introduction Polyploidy, which corresponds to a change to an exact multiple of the haploid number of chromosomes (e.g. 4n and 8n), can be induced via two main routes: rereplication of the DNA in the absence of an intervening mitosis or in the absence of a functional spindle or premature exit from mitosis to the next G1 phase without having completed chromatid migration to the poles (for a review see Kirsch-Volders et al., 1998). The consequences of polyploidy may be directly dependent on the ploidy status, as for abnormal sperm cells or oocytes, which 1To

are responsible for spontaneous abortions, increased transcription rate in particular tissues (liver and trophoblast) or predisposition to oncogenic transformation (Pathak et al., 1994). Since polyploidization in tumours is often followed by aneuploidy, which is in turn associated with high grade invasive tumours and poor prognosis (Sandberg, 1977; Giaretti, 1994; Segers et al., 1994; Verdoodt et al., 1994), the mechanisms leading to polyploidy and the survival of polyploid cells have recently been attracting increasing interest. Polyploidy can be induced by certain chemicals; these compounds can often, but not always, also induce aneuploidy (for a review see Aardema et al., 1998). The methodologies to assess polyploidy are similar to those available to detect aneuploidy (for a review see Kirsch-Volders et al., 1998). Both events can therefore be studied in parallel. An example of chemicals that can induce both aneuploidy and polyploidy are the so-called spindle poisons, compounds that interfere with the formation of the metaphase spindle. Nocodazole, which influences microtubule turnover (Jordan et al., 1992; Vasquez et al., 1997) and which has been used for chemotherapy, belongs to this group of products. In isolated human lymphocytes analysed with the in vitro cytokinesis blocked micronucleus test, Elhajouji et al. (1995, 1997, 1998) demonstrated that nocodazole did induce aneuploidy and at higher doses also polyploidy. The proportions of aneuploid/polyploid versus euploid cells observed are of course also dependent on the relative survival of cells with numerical chromosome aberrations. Until now, induction of programmed cell death specifically in these aneuploid/polyploid cells during the following interphase and in the next mitosis was not our object of study. Since our earlier data (Cundari et al., 1996; Cundari et al., submitted for publication) suggested that aneuploid cells induced by lower nocodazole concentrations survive relatively well, in this work we studied the survival of polyploid cells formed after treatment with a higher nocodazole concentration (0.1 µg/ml). Our aim was also to try to identify the signals responsible for apoptosis after nocodazole treatment and to determine the phases of the cell cycle in which it is induced. In a first approach, normal primary human lymphocytes were studied to avoid any bias due to the aneuploid character of transformed permanent cell lines. Complementary chromosomal parameters were quantified: frequency of abnormal mitotic figures and apoptotic cells during the first and second mitotic wave and frequencies of polyploid metaphases at the second mitosis. The results clearly indicate a positive correlation between abnormal mitotic figures, apoptosis and polyploidy. This positive correlation between apoptosis and abnormal metaphases or polyploidy in lymphocytes led to a single cell approach in which both events can be scored in the same cell. For this purpose, actively proliferating cells are required and two human leukaemic cell lines were used, KS and K562.

whom correspondence should be addressed. Tel: 132 2 629 34 23; Fax: 132 2 629 27 59; Email: [email protected]

© UK Environmental Mutagen Society/Oxford University Press 1999

513

B.Verdoodt et al.

K562 is often used as a model cell line for the study of apoptosis (see for example Kaufmann et al., 1993; McGahon et al., 1994; Durrieu et al., 1998; Esteve et al., 1998). KS was derived from K562 by means of infection with H1 parvovirus. This virus preferentially kills tumoral cells; cells surviving infection with H1 have a suppressed malignancy. KS was found to re-express wild-type p53 (Telerman et al., 1993); the gene is present in K562, but is not expressed. These specific cell lines were selected as there are indications that p53 has a role in the prevention of aneuploidy and polyploidy. Cells in which a functional spindle cannot be formed normally arrest in metaphase (Rieder et al., 1994; Sluder et al., 1994). This arrest is, however, only temporary and cells eventually exit metaphase to become aneuploid or polyploid (reviewed in Rudner and Murray, 1996). It has already been found that p53-negative cells become aneuploid or polyploid at a higher frequency when treated with spindle poisons (Cross et al., 1995; Minn et al., 1996; Di Leonardo et al., 1997), suggesting a role for p53 in the regulation of polyploid cell propagation via activation of a post-mitotic checkpoint (Lanni and Jacks, 1998). In both cell lines, the following chromosomal parameters were analysed: frequencies of polyploid metaphases in the presence of nocodazole and frequencies of polyploidy versus aneuploidy in apoptotic and viable interphase cells during up to three cell cycles of nocodazole treatment. The results confirm that, in the presence of nocodazole, p53 can block the rereplication of DNA in polyploid cells. Moreover, it was demonstrated that polyploid and euploid cells are present in similar proportions in viable and apoptotic cells before rereplication. This suggests that apoptosis is triggered before mitotic segregation has taken place. The possibility that the absence of polymerized tubulins at the G2/M transition can also induce apoptosis, possibly under the control of Cdc2, Bcl2, Raf-1 and/or Rho, is also discussed. Materials and methods Cell culture Human peripheral blood samples were obtained from healthy donors of ,30 years of age. Lymphocytes were isolated using Ficoll-Pacque (Pharmacia Biotech, Uppsala, Sweden) and were cultured at a concentration of 0.53106 cells/ml in Ham’s F-10 medium with 25 mM HEPES buffer and l-glutamine (Gibco BRL, Paisley, UK), supplemented with 15% fetal calf serum (Gibco BRL), 1% penicillin/streptomycin (Gibco BRL) and 2% phytohaemagglutinin HA16 (PHA; Murex Biotech Ltd, Dartford, UK). For the chromosome number counting, colcemid (GIibco BRL) was added at 0.2 µg/ml culture 1.5 h before terminating the cultures. The human chronic myelogenous leukaemia cell line K562 (Lozzio and Lozzio, 1975) and its subclone KS (Telerman et al., 1993) were grown in RPMI 1640 medium (Life Technologies, Paisley, UK) supplemented with 10% heat-inactivated fetal calf serum (Gibco BRL) in an atmosphere with 5% CO2 and 100% humidity at 37°C. These cell lines are near-triploid in untreated cultures (Lozzio and Lozzio, 1975). Nocodazole, dissolved in dimethylsulfoxide (DMSO), was added to the lymphocyte cultures 24 h after the start of culture. The final DMSO concentration did not exceed 0.5%. Primary lymphocyte cultures were harvested after a total culture time of 48 or 72 h, corresponding to 24 and 48 h of nocodazole exposure; exponentially growing cultures of K562 and KS were treated for 24, 48 and 72 h. Only 0.5% DMSO was added to control cultures. Cells were fixed with 3:1 methanol/acetic acid. Slides were prepared by dropping the fixed cell suspension onto them; they were stored at –20°C until use. For chromosome counting and classification of mitotic figures, slides were stained in 5% Giemsa solution (Merck) in So¨rensen buffer, pH 6.8 (Prosan, Gent, Belgium). Scattered chromatin as defined in Kirsch-Volders (1986) actually corresponds to apoptotic nuclei. Cells were counted at a magnification of 12503 on a Zeiss transmitted light microscope.

514

TUNEL assay The In Situ Cell Death Detection kit AP (Boehringer-Mannheim, Mannheim, Germany) was used according to the instructions of the manufacturer, both for the primary lymphocytes and for both cell lines. Slides were counterstained with ethidium bromide, to be visualized with epifluorescence microscopy at a magnification of 4003. Annexin staining and flow sorting Cells were collected by centrifugation and resuspended in 100 µl annexin labelling solution, consisting of 2% annexin-V-FLUOS (BoehringerMannheim) and 0.1 µg/ml propidium iodide in HEPES buffer (10 mM HEPES/ NaOH, 140 mM NaCl, 5 mM CaCl2, pH 7.4). Cells were incubated in the dark in this solution for 15 min. When used for counting aneuploid cells, they were spread onto slides and air dried, but not otherwise fixed. If cells were going to be used for flow sorting, they were resuspended in HEPES buffer and no propidium iodide was added to the labelling solution. Cells were separated into apoptotic (annexin-positive) and viable (annexin-negative) populations, using a FACstar PLUS flow cytometer (Becton Dickinson, Oxford, UK). In situ hybridization The pUC1.77 probe (Cooke and Hindley, 1979), which hybridizes to the satellite III pericentromeric region of chromosome 1, was used to detect chromosome 1. To label the centromere of chromosome 17 the probe D17Z1 was used; this probe was obtained from ATCC (Rockville, MD). Probes were labelled by nick-translation with digoxigenin-11-dUTP (Boehringer-Mannheim) or biotin-11-dUTP (Boehringer-Mannheim). The labelling reaction took place in a total volume of 50 µl for 1 µg of probe with 1.5 nmol of dATP, dCTP and dGTP and 3 nmol of digoxigenin-11-dUTP or the same amount of biotin-11-dUTP, adding DNase I/DNA polymerase I. Products were acquired in the form of a kit (Nick Translation System; Gibco BRL). The reaction was carried out at a temperature of 15°C for 2 h. Fluorescence in situ hybridization (FISH) was essentially carried out as described by Viegas-Pequignot et al. (1989). Briefly, slides were treated with 0.01% RNase in 23 SSC (0.3 M sodium chloride, 0.03 M sodium citrate) for 60 min at 37°C, rinsed in 23 SSC, then treated with 50 µg/ml pepsin in 10 mM HCl for 10 min at 37°C and finally dehydrated in a graded ethanol series (50, 70 and 98%) and air dried. Aliquots of 40 ng of both probes, dissolved in 20 µl of a hybridization mix containing 50% formamide, were used per slide; slides and probe were denatured together for 4 min at 90°C on a hotplate. Hybridization took place overnight at 37°C. After hybridization, slides were washed twice for 7 min in 23 SSC, 50% formamide and twice for 2 min in 23 SSC, both at 43°C, after which they were rinsed in blocking buffer (0.5% blocking reagent in 43 SSC; BoehringerMannheim). The slides were then incubated first with avidin–FITC (Vector Laboratories), at 1/200 dilution, then with monoclonal mouse anti-digoxigenin (Boehringer-Mannheim) at a dilution of 1/250 together with anti-avidin (Vector Laboratories) at a dilution of 1/100, and finally again with avidin–FITC, at 1/200 dilution with Texas Red®-conjugated sheep anti-mouse antibody (Amersham, Little Chalfont, UK) at a dilution of 1/20. All incubations took place at 37°C for 30 min. In between incubations, slides were washed in 43 SSC, 0.05% Tween 20. After dehydration in a graded alcohol series, the nuclei were counterstained with DAPI at a concentration of 1 µg/ml. To score the FISH slides, the criteria as defined by Hopman et al. (1990) were applied: spots were counted on non-overlapping intact nuclei and secondary spots, which give a much weaker signal, were ignored. Cells with more than nine spots for either chromosome were counted as one category, as it was difficult to distinguish acurately between such high numbers of spots within one cell. Statistics For statistical analysis of the counts of apoptotic cells and abnormal and polyploid metaphases Fisher’s exact test was used. For the comparisons of the numbers of FISH signals in the cell lines either Fisher’s exact test or the χ2 test were used, depending on the number of categories. The significance of correlation coefficients is determined by a test based on the normal distribution.

Results Positive correlations between abnormal mitotic figures, polyploidy and apoptosis after in vitro nocodazole treatment in human primary lymphocytes Induction of abnormal mitotic figures Treatment of PHA-stimulated lymphocytes for 24 or 48 h with nocodazole at concentrations of 0.04 and 1.0 µg/ml induced a

Nocodazole, polyploidy and apoptosis

Fig. 1. Changes in the fraction of normal and abnormal mitotic figures after nocodazole treatment for 24 and 48 h in primary human lymphocytes. , normal mitoses; , C mitoses; , anaphase lagging and dislocated chromosomes; , chromatin masses. The columns indicate the mean value of both cultures, the lines the values for the two separate cultures.

Table I. Mitotic figures counted on the total number of cells in nocodazoletreated primary lymphocytes Treatment

Mitotic index

Normal mitoses (% of total no. of cells)

C mitoses (% of total no. of cells)

24 24 24 48 48 48

12.45 22.65 22.02 6.90 28.00 36.10

7.67 13.50 12.05 4.73 10.30 13.00

0.47 3.90 4.71 0.55 9.25 11.40

h, h, h, h, h, h,

control 0.04 µg/ml 0.1 µg/ml control 0.04 µg/ml 0.1 µg/ml

statistically significant increase in the fraction of abnormal metaphases and anaphases, as compared with the total number of mitotic cells (Figure 1). In each case, two parallel cultures from the same donor were counted, which gave similar results, although the difference between the two cultures was always larger at the highest nocodazole dose. Per culture, 1000–1800 cells and 68–384 mitoses were counted. The increase in the fraction of abnormal metaphases and anaphases was time dependent; the differences between the 24 and 48 h treatments were highly significant for both the treatment with 0.04 (P , 0.0001) and with 0.1 µg/ml nocodazole (P , 0.0001). Differences in effect between the two nocodazole concentrations were less clear and only significant for the shortest treatment time (P 5 0.0022). The high fraction of chromatin masses in the control cultures after 24 h treatment is surprising. These might represent (early) apoptotic cells or be an indication of toxicity. More advanced techniques would be required to determine their exact origin. The increased mitotic index (Table I) with the longer treatment time was mainly due to an increase in the fraction of C mitoses. More surprising is that the number of normallooking mitoses also increased. Induction of apoptosis On the same slides that were used for the counting of normal and abnormal mitotic figures, apoptotic cells were also scored as scattered chromatin. A clear time-dependent increase in the frequency of apoptotic cells was found by scoring the Giemsa

stained slides (Figure 2). No pronounced differences were observed between the two nocodazole concentrations studied (0.04 and 0.1 µg/ml), although the differences were statistically significant (P , 0.0001 for both treatment times). However, the treatment duration had a much stronger effect on the frequency of apoptotic cells. Even in the controls, more apoptosis was observed after 72 than after 48 h in culture. The spontaneous induction of apoptosis in cultures of longer duration may be due to the fact that this is a more or less artificial system, as compared with the human body, their normal environment. The correlation between the frequencies of abnormal metaphases/anaphases and those of apoptotic nuclei on the same Giemsa stained slides was quite high (r2 5 0.665, P . 0.05), but not significant. Similar effects were seen when TUNEL staining was used to detect apoptotic cells on separate slides prepared from the same donor (Figure 2). Here, 1500–2000 cells were scored per slide. Assessments of apoptosis with Giemsa or TUNEL staining correlated very well (r2 5 0.963, P , 0.01), although the fraction of apoptotic cells detected by TUNEL staining tended to be higher in the controls and with the shorter treatment. The differences between the frequency of apoptosis detected by TUNEL and Giemsa staining may be due to detection of different stages of aneuploidy; TUNEL is likely to detect earlier stages which are not yet grossly morphologically abnormal. For the TUNEL staining, the correlation with the fraction of abnormal mitoses was somewhat weaker (r2 5 0.608, P . 0.05) than with Giemsa staining. Polyploidy induction in colcemid-blocked human lymphocytes In a separate experiment, the frequencies of polyploid metaphases were analysed after in vitro treatment of human lymphocytes with nocodazole (Figure 3). Chromosome counting was performed on metaphases blocked with colcemid 48 or 72 h after stimulation with PHA and continuous treatment with nocodazole (0.04 and 0.1 µg/ml) for 24 and 48 h, respectively. The data from this experiment clearly show that no polyploidy was observed in untreated primary lymphocytes: at 24 h treatment, of 130 metaphases from two parallel cultures, none was found to be polyploid; at 48 h treatment, no polyploid metaphases were seen either. The percentage of polyploid 515

B.Verdoodt et al.

Fig. 2. Frequency of apoptotic cells in cultures of normal lymphocytes treated with nocodazole, as detected after Giemsa staining ( ) and by TUNEL staining ( ). The columns indicate the mean value of both cultures, the lines the values for the two separate cultures.

Fig. 3. Frequencies of polyploid metaphases in nocodazole-treated metaphases; comparison between 24 and 48 h treatment. The columns indicate the mean value of the two cultures, the lines the values for the separate cultures.

metaphases obtained after nocodazole treatment was higher after 48 (3.5 and 8%, respectively) than after 24 h (0.5 and 5%, respectively) treatment. The difference in frequency of polyploid metaphases between the two treatment times was significant for 0.04 µg/ml nocodazole only (P 5 0.0335). The fractions of polyploid metaphases do not appear very high, but it must be taken into account that there was a strong accumulation of cells in metaphase in these cultures, pointing to a delay in the cell cycle. Moreover, stimulation with PHA does not induce all cultured lymphocytes to divide. A high interculture variability was observed at the higher concentration, for both treatment times. Except for one octoploid cell, all polyploid cells were tetraploid. Genetic identity of apoptotic and non-apoptotic cells in human cell lines expressing or not expressing p53 Polyploidy does not always trigger apoptosis, but loss of p53 allows further divisions of polyploid human leukaemic cells treated with nocodazole To identify the genetic content of apoptotic versus viable cells, the copy numbers of chromosomes 1 and 17 were determined by FISH in both the KS and the K562 cell lines after 24, 48 and 72 h treatment with 0.1 µg/ml nocodazole. Before applying FISH, the cells were separated into apoptotic and viable cells by means of flow sorting after annexin staining. Annexin 516

staining is based on a reversal of cell membrane asymmetry and detects early apoptotic cells. This is important, as in later stages of apoptosis we observed that the DNA is too degraded to obtain meaningful FISH signals. As a control, cells of both cell lines that had been cultured without nocodazole for all studied treatment durations were also subjected to FISH for chromosome 1. The variation in the distribution of hybridization signals between cultures was considerable and appeared somewhat larger in the p53-negative cell line K562 than in the p53-positive KS cells. Nevertheless, the results show that the major cell population is trisomic for chromosome 1 in both cell lines (58–89% of cells) but that tetra- and pentasomic cells are not rare (5.44–21.20 and 2.30– 8.90% of cells, respectively). The distribution of the number of spots per nucleus did not change in any clear direction with culture time, but varied rather a lot between cultures, indicating karyotypic instability in these cell lines. There were no indications of a spontaneous increase in ploidy. These data confirm the results of Lozzio and Lozzio (1975), who also found these cell lines to be near-triploid. A doubling of the chromosome number thus leads to hexaploid cells. An extra control culture was not added to the experiments where cells were separated by FACS, as the fraction of apoptotic cells was always too low to give meaningful results (unpublished results).


Fig. 4. Fluorescence in situ hybridization for chromosomes 1 and 17 in viable ( ) and apoptotic ( ) cells of the KS and K562 cell lines. Cells were treated with 0.1 µg/ml nocodazole for 24, 48 or 72 h. The x-axis indicates the number of spots for both chromosomes; different, different number of spots for the two chromosomes (aneuploid cells) in the same nucleus.

Interestingly, the proportion of cells treated with nocodazole with different numbers of spots for chromosome 1 versus chromosome 17 was very low and never exceeded 15% (Figure 4). This type of signal indicates aneuploid rather than polyploid cells. Moreover, the difference in the number of signals for both chromosomes was never greater than two. It was also independent of the treatment duration or of whether the cells were viable or apoptotic. These data indicate that at the applied concentration, nocodazole essentially induces polyploidy. In KS, treatment with nocodazole shows an induction of hexaploid cells at the three time points considered, but no dodecaploid cells were observed after any of the treatments (Figure 4). This was the case for both viable and apoptotic cells. The ratio of hexaploid to triploid cells did not change

much over time in the viable cells. In the apoptotic cells, the fraction of hexaploid cells was clearly higher after 48 (P , 0.0001) and 72 h (P 5 0.0006) than after only 24 h treatment. Here the cells with four and five spots have been included in the triploid population, as such cells also occur in non-treated cultures of these cell lines. In contrast, a population of dodecaploid cells (essentially viable) was clearly observed at 48 and 72 h in the K562 cell line (Figure 4). A small fraction of dodecaploid apoptotic cells was only found at 72 h. The fraction of hexaploid cells was highest after 48 h treatment, but then decreased, concomitant with the increase in the fraction of dodecaploid cells; the fraction of triploid cells decreased to a lower level than in KS at 48 and 72 h treatment. These data show that polyploid cells 517

B.Verdoodt et al.

Fig. 5. Frequencies of polyploid metaphases in nocodazole-treated metaphases; comparison between 24, 48 and 72 h treatment in the two cell lines: K562; , KS. The bars indicate the mean value of the two cultures.

did not preferentially accumulate in the apoptotic subpopulation, as would be expected if polyploidy constituted a strong signal for the induction of apoptosis. Survival of polyploid metaphases in the human p53 nonexpressing leukaemia cell line K562 after in vitro treatment with nocodazole for 72 h Chromosome counting performed on metaphases blocked with colcemid 48 or 72 h after stimulation with PHA and continuous treatment with nocodazole (0.04 and 0.1 µg/ml) clearly showed (Figure 5) that polyploidy is present in non-treated cells of these cell lines, between 2 and 9% in K562 and between 6 and 15% in KS. Polyploidy in these cell lines means a multiple of the triploid chromosome complement, as these cells are near-triploid in control conditions. In the treated cells the percentage of polyploid metaphases is dependent on the cell line, as well as on the treatment. In K562, the p53 nonexpressing cell line, the maximum frequency of polyploid metaphases was very high (40%) after the longest nocodazole treatment; polyploidy induction is time but not concentration dependent. In KS, the p53-expressing cell line, the highest frequency of polyploid metaphases was 23%; here also induction depends on treatment duration but not on nocodazole concentration. For these experiments, 100 metaphases were counted per slide. Discussion The major reasons to consider that aneuploidy and polyploidy are compatible with normal cell life and do not always induce apoptosis are found in the survival of aneuploid individuals and polyploid tissues, like the liver (Sateer et al., 1988) in humans. On the other hand, the idea that the induction of aneuploidy or polyploidy could trigger apoptosis emerged from different types of considerations and observations: (i) in a given tissue the maintenance of a stable karyotype is the goal of mitotic division and spontaneous segregation errors should be eliminated; (ii) in mammalian cell cultures in vitro apoptotic cells were observed after treatment with nocodazole (Wahl et al., 1996; Cundari et al., submitted for publication); (iii) the contradictory results described by the ECETOC working party on chemically induced aneuploidy in which, after treatment with spindle inhibitors, almost no polyploidy was observed after the first mitotic cycle in primary cells but quite 518

,

clearly occurs in permanent cell lines (Aardema et al., 1998); (iv) our own quantitative data on the respective frequencies of chromosome non-disjunction, chromosome loss and polyploidy induced in vitro in human lymphocytes by spindle inhibitors (Elhajouji et al. 1995, 1997, 1998). Given these earlier observations (Elhajouji et al. 1995, 1997, 1998), our aim was to define which genomic changes induced by nocodazole treatment lead to apoptosis, and in which phase(s) of the cell cycle. With this objective in mind, we first studied the correlations between the induction of polyploidy, apoptosis and abnormal divisions in normal primary human lymphocytes, to avoid any bias due to the aneuploid character of most transformed cell lines. We found a positive correlation between the frequencies of abnormal mitotic figures and of apoptotic nuclei. Whether these abnormal divisions were all lethal or whether they might in some cases lead to polyploid cells was investigated by classical chromosome counting in metaphase. Our results indicate that polyploid cells are indeed formed and survive, since a low number of polyploid metaphases was already seen after 48 h in culture and a higher percentage was found at 72 h. Moreover, an increase in the mitotic index was also observed in the presence of nocodazole, which suggests that no strong selective pressure was operating against mitotic cells. The positive correlation observed between apoptosis, abnormal metaphases and polyploidy in human primary lymphocytes led us to the choice of a system in which apoptosis and polyploidy can be assessed in the same cell. We decided to do a sorting of apoptotic versus non-apoptotic cells. However, this methodology requires actively proliferating cells in relatively large amounts. We therefore selected two cell lines of leukaemic origin: K562 (Lozzio and Lozzio, 1975) and its derivative KS (Telerman et al., 1993). KS expresses the wildtype p53 gene, whereas K562 does not. K562 has been used before by different laboratories as a model for the study of the role of p53 in apoptosis (see for example Kaufmann et al., 1993; McGahon et al., 1994; Durrieu et al., 1998; Esteve et al., 1998). As the induction of polyploid metaphases in these cell lines did not seem to depend on the nocodazole concentration, only the higher concentration (0.1 µg/ml) was used for the comparison of apoptotic and viable interphase cells. In KS, the cell line which expresses wild-type p53, at 24 h most


Fig. 6. Some of the genes involved in the induction of apoptosis after treatment with spindle poisons.

viable cells already showed six spots for both chromosomes 1 and 17, indicating that a first round of polyploidization had taken place. At both later time points a higher fraction of trisomic viable cells was seen than at 24 h; no cells with more than nine spots for either chromosome were ever observed in this cell line. Of the apoptotic cells, a relatively high fraction was still trisomic, although the major part was hexasomic. This would indicate that these cells do not progress to higher levels of polyploidy than one doubling of the chromosome number. The K562 cell line, on the other hand, does not express p53. In this cell line, few trisomic cells undergo apoptosis at 24 h; the great majority of apoptotic cells were hexasomic, which remains the case for all further time points. The viable cells in this line evolved from mainly hexasomic cells at 24 h to an increasing fraction of dodecasomic cells at 48 and 72 h. This would indicate that, although some cells die after a first attempted division with a damaged spindle, other cells are apparently able to undergo a second division cycle in these circumstances and to survive further. After 24 h nocodazole treatment, only a weak increase in the frequency of polyploid metaphases could be seen and only at the highest concentration in both cell lines. This is not surprising, as these cells are not synchronized and have a cell cycle time of ~19 h. Only a few cells will have had time to reach a second metaphase in the 24 h period, especially as a delay in mitosis is to be expected after treatment with a spindle poison. The fraction of polyploid metaphases clearly increased with treatment time, as increasing numbers of cells had had time to go through more than one metaphase. The frequency of polyploid metaphases did not seem to depend on the nocodazole concentration in either cell line, however. From these data it would seem that both concentrations are about as effective in disturbing the function of the metaphase spindle, although neither is able to completely inhibit the polymerization of tubulin (Jordan et al., 1992). To correctly interpret the data on the frequency of polyploidy in the apoptotic versus the non-apoptotic cells, it is important to realise that the cells which were analysed after 48 h treatment derive from the non-apoptotic cells observed after 24 h treatment; the same applies for the cells analysed after

72 h treatment, which derive from the surviving cells studied after 48 h treatment. Our results show that after 24 h treatment polyploid cells either survive or undergo apoptosis, independent of the presence of p53 protein. This suggests that p53 is not the only source of an apoptotic signal. Interestingly, the distribution of spots in apoptotic cells at each time point mimics that of the viable cells at the preceding time. In a second phase, the cells which survived after 24 h treatment will again either survive longer or undergo apoptosis. In the absence of the p53 gene product these cells can continue to cycle, as is indicated by the appearance of dodecasomic cells. This clearly confirms that wild-type p53 can block further cycling of polyploid cells by blocking rereplication. The distribution of FISH signals after 72 h treatment confirms the results observed after 48 h treatment. If the observation that p53 can block rereplication of the DNA in polyploid cells is not new (Minn et al., 1996), these results are the first to demonstrate that the fraction of polyploid cells is almost the same in apoptotic as in non-apoptotic cells during the hours which precede the first rereplication cycle. The most probable explanation for induction of apoptosis by nocodazole before the passage of polyploid cells through rereplication is that not only spindle microtubules but also interphase microtubules are sensitive to nocodazole treatment. Sensitivity to this spindle poison appears to be greatest at initiation of mitosis, when the microtubules progress from a stable, less dynamic interphase organization to the more dynamic organization of the mitotic spindle (Lieuvin et al., 1994). Our data thus strongly suggest that besides the G1/S checkpoint under the control of p53 (for a review see Wahl et al., 1997), the G2/M transition may be sensitive to depolymerization of microtubules. This seems to correspond to earlier data (Donaldson et al., 1994; Haldar et al., 1997) on the effects of taxol on the induction of apoptosis, where G2/M cells were most sensitive to the effects of this chemical. Apoptosis induced by spindle poisons is likely independent of p53 function: KS and K562 underwent apoptosis with about equal frequencies after nocodazole treatment (Cundari et al., submitted for publication). Other genes have been found to be involved in the apoptotic pathway after spindle poison treatment (Figure 6); these include Cdc2 and some of its regulators, genes of the Bcl-2 family, the c-erbB-2 oncogene, Raf-1 and Rho (Donaldson et al., 1994; Minn et al., 1996; Blagosklonny et al., 1997; Huang et al., 1997; Esteve et al., 1998; Yu et al., 1998). An important factor appears to be the inappropriate activation of Cdc2, i.e. when the cell is not ready to enter mitosis. This activation can be prevented (possibly indirectly) through up-regulation of p21Waf1; c-erbB-2 is able to produce this effect (Yu et al., 1998). On the other hand, modulation of the effects of genes of the Bcl-2 family in response to spindle poison treatment has also been observed. Raf-1, a kinase that is activated by Ras, can in this case phosphorylate Bcl-2. This phosphorylation inactivates the anti-apoptotic action of Bcl-2 (Blagosklonny et al., 1997). Another effector in this pathway is Rho, a Rasrelated GTPase. It has been found to promote apoptosis in response to serum starvation; this effect is counteracted by Bcl-2. Finally, up-regulation of the anti-apoptotic Bcl-2 family member Bcl-xL can prevent the induction of apoptosis after nocodazole treatment (Minn et al., 1996). Apart from providing a better insight into the mechanisms of apoptosis, these results may also aid the development of antitumour agents, as two of the most successful anti-mitotic 519

B.Verdoodt et al.

agents used in chemotherapy (the vinca alkaloids and the taxanes) exert their effect through suppression of the dynamics of the microtubules. Acknowledgements The authors wish to thank Prof. P. Caillet-Fauquet and Dr M. Tuynder, of the Radiobiology Laboratory of the Universite´ Libre de Bruxelles, for the generous gift of the KS and K562 cell lines. This work was supported by EU research programme ENV4-CT97-0471.

References Aardema,M.J., Albertini,S., Arni,P., Henderson,L.M., Kirsch-Volders,M., Mackay,J.M., Sarrif,A.M., Stringer,D.A. and Taalman,R.D. (1998) Aneuploidy: a report of an ECETOC task force. Mutat. Res., 410, 3–79. Blagosklonny,M.V., Giannakakou,P., El-Deiry,W.S., Kingston,D.G.I., Higgs,P.I., Neckers,L. and Fojo,T. (1997) Raf-1/bcl-2 phosphorylation: a step from microtubule damage to cell death. Cancer Res., 57, 130–135. Cooke,H.J. and Hindley,J. (1979) Cloning of human satellite III DNA, different components are on different chromosomes. Nucleic Acids Res., 6, 3177–3197. Cross,D.M., Sanchez,C.A., Morgan,C.A., Schimke,M.K., Ramel,S., Idzerda,R.L., Raskind,W.H. and Reid,B.J. (1995) A p53-dependent mouse spindle checkpoint. Science, 267, 1353–1356. Cundari,E., Tuynder,M., Elhajouji,A., Caillet-Fauquet,P., Mousset,S. and Kirsch-Volders,M. (1996) Induction of apoptosis by mitotic spindle poisons (abstract). In 26th EEMS Annual Meeting–Workshop on Chromosome Instability and Cell Cycle Control, Rome, September 3–7, 1996. Di Leonardo,A., Khan,S.H., Linke,S.P., Greco,V., Seidita,G. and Wahl,G.M. (1997) DNA rereplication in the presence of mitotic inhibitors in human and mouse fibroblasts lacking either p53 or pRb function. Cancer Res., 57, 1013–1019. Donaldson,K.L., Goolsby,G.L., Kiener,P.A. and Wahl,A.F. (1994) Activation of p34cdc2 coincident with taxol-induced apoptosis. Cell Growth Differ., 5, 1041–1050. Durrieu,F., Belloc,F., Lacoste,L., Dumain,P., Chabrol,J., Dachary-Prigent,J., Morjani,H., Boisseau,M.-R., Reiffers,J., Bernard,Ph. and Lacombe,F. (1998) Caspase activation is an early event in anthracycline-induced apoptosis and allows detection of apoptotic cells before they are ingested by phagocytosis. Exp. Cell, Res., 240, 165–175. Elhajouji,A., Van Hummelen,P. and Kirsch-Volders,M. (1995) Indications for a threshold of chemically-induced aneuploidy in vitro in human lymphocytes. Environ. Mol. Mutagen., 26, 292–304. Elhajouji,A., Tibaldi,F. and Kirsch-Volders,M. (1997) Indication for thresholds of chromosome non-disjunction versus chromosome lagging induced by spindle inhibitors in vitro in human lymphocytes. Mutagenesis, 12, 133–140. Elhajouji,A., Cunha,M. and Kirsch-Volders,M. (1998) Spindle poisons can induce polyploidy by mitotic slippage and micronucleate mononucleates in the cytokinesis-block assay. Mutagenesis, 13, 193–198. Esteve,P., Embade,N., Perona,R., Jimeńez,B., del Peso,L., Leoń,J., Arends,M., Miki,T. and Lacal,J.C. (1998) Rho-regulated signals induce apoptosis in vitro and in vivo by a p53-independent, but Bcl2-dependent pathway. Oncogene, 17, 1855–1869. Giaretti,W. (1994) A model of DNA aneuploidization and evolution in colorectal cancer. Lab. Invest., 71, 904–910. Haldar,S., Basu,A. and Croce,C. (1997) Bcl2 is the guardian of microtubule integrity. Cancer Res., 57, 229–233. Hopman,A.H.N., Ramaekers,F.C.S. and Vooijs,G.P. (1990) Interphase cytogenetics of solid tumours. In Polak,J.M. and McGee,J. (eds), In Situ Hybridisation. Principles and Practice. Oxford University Press, New York, NY, pp. 165–186. Huang,T.-S., Shu,C.-H., Yang,W.K. and Whang-Peng,J. (1997) Activation of CDC25 phosphatase and CDC2 kinase involved in GL331-induced apoptosis. Cancer Res., 57, 2974–2978. Jordan,M.A., Thrower,D. and Wilson,L. (1992) Effects of vinblastine, podophylotoxin and nocodazole on mitotic spindles. J. Cell Sci., 102, 401–416. Kaufmann,S.H., Desnoyers,S., Ottaviano,Y., Davidson,N.E. and Poirier,G.G. (1993) Specific proteolytic cleavage of poly(ADP-ribose) polymerase: an early marker of chemotherapy-induced apoptosis. Cancer Res., 53, 3976–3985. Kirsch-Volders,M. (1986) Differential staining of chromosomes and spindle cannot be used as an assay to determine the effect of cancer promotors on primary cultures of human fibroblasts. Mutat. Res., 171, 177–183.

520

Kirsch-Volders,M., Cundari,E. and Verdoodt,B. (1998) Towards a unifying model for the metaphase/anaphase transition. Mutagenesis, 13, 321–335. Lanni,J.S. and Jacks,T. (1998) Characterization of the p53-dependent postmitotic checkpoint following spindle disruption. Mol. Cell. Biol., 18, 1055–1064. Lieuvin,A., Labbe´,J.-C., Doreé,M. and Job,D. (1994) Intrinsic microtubule stability in interphase cells. J. Cell Biol., 124, 985–996. Lozzio,C.B. and Lozzio,B.B. (1975) Human chronic myelogenous leukemia cell-line with positive Philadelphia chromosome. Blood, 45, 321–334. McGahon,A., Bissonnette,R., Schmitt,M., Cotter,K., Green,D. and Cotter,G. (1994) BCR-ABL maintains resistance of chronic myelogenous leukemia cells to apoptotic cell death. Blood, 83, 1179–1187. Minn,A.J., Boise,L.H. and Thompson,C.B. (1996) Expression of Bcl-xL and loss of p53 can cooperate to overcome a cell cycle checkpoint induced by mitotic spindle damage. Genes Dev., 10, 2621–2631. Pathak,S., Dave,B.J. and Gagos,S. (1994) Chromosome alterations in cancer development and apoptosis. In Vivo, 8, 843–850. Rieder,C.L., Schultz,A., Cole,R., and Sluder,G. (1994) Anaphase onset in vertebrate somatic cells is controlled by a checkpoint that monitors sister kinetochore attachment to the spindle. J. Cell Biol., 127, 1301–1310. Rudner,A.D. and Murray,A.W. (1996) The spindle assembly checkpoint. Curr. Opin. Cell Biol., 8, 773–780. Saeter,G., Lee,C.Z., Schwarze,P.E., and Seglen,P.O. (1988) Changes in ploidy distributions in human liver carcinogenesis. J. Natl Cancer Inst., 80, 1480–1485. Sandberg,A.A. (1977) Chromosome markers and progression in bladder cancer. Cancer Res., 37, 2950–2956. Segers,P., Haesen,S., Amy,J.J., De Sutter,P., Van Dam,P. and Kirsch-Volders,M. (1994) Detection of premalignant stages in cervical smears with a biotinylated probe for chromosome 1. Cancer Genet. Cytogenet., 75, 120–129. Sluder,G., Miller,F.J., Thompson,E.A. and Wolf,D.E. (1994) Feedback control of the metaphase–anaphase transition in sea urchin zygotes: role of maloriented chromosomes. J. Cell Biol., 126, 198–198. Telerman,A., Tuynder,M., Dupressoir,T., Robaye,B., Sigaux,F., Shaulian,E., Oren,M., Rommelaere,J. and Amson,R. (1993) A model for tumor suppression using H-1 parvovirus. Proc. Natl Acad. Sci. USA, 90, 8702– 8706. Vasquez,R.J., Howell,B., Yvon,A.M., Wadsworth,P. and Cassimeris,L. (1997) Nanomolar concentrations of nocodazole alter microtubule dynamic instability in vivo and in vitro. Mol. Biol. Cell, 8, 973–985. Verdoodt,B., Castelain,Ph., Bourgain,C. and Kirsch-Volders,M. (1994) Aneuploidy for chromosome 1 and overall DNA-content in benign and malignant breast disease. Cancer Genet. Cytogenet., 78, 53–63. Viegas-Pequignot,E., Jeanpierre,M., Dutrillaux,A.M., Gerbault-Seureau,M., Muleris,M. and Dutrillaux,B. (1989) Detection of 1q polysomy in interphase nuclei of human solid tumors with a biotinylated probe. Hum. Genet., 81, 311–314. Wahl,A.F., Donaldson,K.L., Fairchild,C., Lee,F.Y.F., Foster,S.A., Demers,G.W., and Galloway,D.A. (1996) Loss of normal p53 function confers sensitization to taxol by increasing G2/M arrest and apoptosis. Nature Med., 2, 72–79. Wahl,G.M., Linke, S.P., Paulson,T.G. and Huang,L.C. (1997) Maintaining genetic stability through TP53 mediated checkpoint control. Cancer Surv., 29, 183–219. Yu,D., Jing,T., Liu,B., Yao,J., Tan,M., McDonnell,T.J. and Chung,M.-C. (1998) Overexpression of ErbB2 blocks taxol-induced apoptosis by upregulation of p21Cip1, which inhibits p34Cdc2 kinase. Mol. Cell, 2, 581–591. Received on March 8, 1999; accepted on May 18, 1999