Topoisomerase II inhibition prevents anaphase chromatid segregation in mammalian cells independently of the generation of DNA strand breaks

During replication the double-stranded DNA molecules of sister chromatids inevitably become concatenated. These concatenations must be separated in preparation for anaphase chromatid segregation; Cook (1991) has recently reviewed the structural complexities of this decatenation. The only known eukaryotic enzyme capable of decatenating double-stranded DNA is DNA topoisomerase II (EC 5.99.1.3); see Osheroff et al. (1991) for a review of its mechanism.

In budding and fission yeast, temperature-sensitive Top2 mutants have demonstrated the essential role of this enzyme in decatenating sister chromatids before anaphase commences (di Nardo et al., 1984; Holm et al., 1989; Uemura and Yanagida, 1986; Uemura et al., 1987). When grown at the non-permissive temperature, Top2 mutants of Schizosaccharomyces pombe are unable to complete the final stages of chromosome condensation (Uemura et al., 1987). If Top2 cells are shifted to the non-permissive temperature after condensation is complete, the cells attempt anaphase but fail to separate their chromatids. The cells do, however, continue with septation, though full cytokinesis is prevented by chromatin bridges joining the two daughters (Uemura and Yanagida, 1986; Uemura et al., 1987). When cytokinesis is blocked, as in the Top2-cdc11 double mutant, the cells re-enter interphase with a single nucleus (Uemura and Yanagida, 1986). Yeast topoisomerase II is therefore required for chromosome condensation and chromatid segregation, but is not needed for progression through the later stages of mitosis.

In higher eukaryotes, definitive proof of mitotic topoisomerase II functions is not yet possible, since topoisomerase II mutants have not been isolated. As a substitute, specific topoisomerase II inhibitors such as the non-intercalating epipodophyllotoxins, etoposide (VP-16-123) and teniposide (VM-26), or intercalating inhibitors such as m-AMSA, have been employed. Treating mammalian cells with such inhibitors produces a G₂ delay (Kalwinsky et al., 1983; Tobey et al., 1990). Events in cell division are also affected in a manner parallel to yeast Top2 mutations. Wright and Shatten (1990) found that meiotic chromosome condensation and segregation could be inhibited with teniposide in Spisula solidissima (surf clam) oocytes. Similarly, Downes et al. (1991) showed that etoposide inhibits mitotic segregation in mammalian cells (HeLa and PtK₂), but like the aforementioned yeast Top2 mutations, does not prevent cytokinesis. More recently, Shamu and Murray (1992) elegantly showed that etoposide inhibits chromatid separation in an in vitro system derived from sperm nuclei in a Xeno pus egg extract.

Data obtained with topoisomerase II inhibitors have not been totally unequivocal, however. Sumner (1992) has shown that low doses of a variety of topoisomerase II inhibitors do not prevent the separation of chromatids that occurs in human lymphocytes arrested for a prolonged period in mitosis with Colcemid. In this system, it is not possible to employ high doses of topoisomerase inhibitors, such as have been found to be necessary to prevent anaphase disjunction in other cells and cell-free systems (Downes et al., 1991; Shamu and Murray, 1992), since such concentrations are toxic over prolonged periods; the low doses employed may not be adequate to inhibit topoisomerase II fully in the lymphocyte system. Alternatively, Sumner suggests that effects seen in other systems may be a consequence of DNA damage induced by high doses of the inhibitors used.

Such damage is an inevitable consequence of the mode of action of all the topoisomerase II inhibitors used by previous investigators (etoposide, teniposide, m-AMSA). All these act by arresting topoisomerase II at the transition state, when it is covalently linked to DNA at sites of DNA double-strand breaks, thus generating what are known as cleavable complexes. This prevents the topoisomerase II-mediated re-ligation reaction, which is the last step of decatenation; and ultimately double-stranded DNA breaks result. Such DNA damage is known to induce G₂ delay, and so the G₂ delay seen in response to topoisomerase II inhibitors could be explained by a cell cycle arrest in response to DNA damage (Lock and Ross, 1990). More speculatively, the effects of topoisomerase II inhibitors at anaphase could be analogous to the disturbed segregation that has long been known to occur in mitotic mammalian cells after X-irradiation (Perk, 1941).

However, a different class of topoisomerase II inhibitors has now become available: the bisdioxopiperazines, a class of antitumor agents which were originally developed in the 1970s in the ICRF laboratories, Lincoln’s Inn Fields, London. Since many antineoplastic drugs have chelating properties, bisdioxopiperazines were designed as derivatives of the calcium chelator EDTA, with increased ability to penetrate cell membranes. ICRF-193 is the meso-dimethyl derivative of ICRF-159, which was itself produced from ICRF-154, the bis-cyclic imide of EDTA (Creighton et al., 1969). Early attempts to characterise the cytotoxicity of these agents showed that ICRF-159 prevented complete mitosis, resulting in the formation of tetraploid cells. Creighton and Birnie (1970) found that ICRF-159 reduced the gross rate of DNA synthesis in cultured cells; and Creighton (1979) showed that progression through S-phase and mitosis is delayed. ICRF-159 also delays BHK-21S cells in G₂ (White and Creighton, 1976; White and Creighton, 1977).

The mechanism of action of the ICRF bisdioxopiperazines has recently become clearer. Tanabe et al. (1991) showed that ICRF-193 inhibits calf thymus topoisomerase II, as measured by decatenation of Crithidia fasciculata kinetochore DNA, without effect on topoisomerase I or DNA polymerases α and β. The drug inhibited topoisomerase II without generating topoisomerase II-DNA cleavable complexes, and suppressed the formation of etoposide-induced cleavable complexes. Ishimi et al. (1992) also demonstrated total inhibition of topoisomerase II-mediated decatenation of replicated P4 phage DNA (in HeLa extract) by 12.5 μM ICRF-193. Ishida et al. (1991) found that in RPMI 8402 human leukaemic cells, ICRF-154 does not form cleavable complexes but does cause G₂ delay. They also observed prophase blocked cells with very tangled chromatin after treating asynchronous cells with 10 μM ICRF-193 for 15 hours.

We have investigated the action of ICRF-193 on mitotic events in HeLa and PtK₂ cells, arguing that if topoisomerase II is required for mitotic decatenation, ICRF-193 should have effects similar to those of etoposide (Downes et al., 1991).

ICRF-193 stock, a generous gift from Dr Andrew Creighton, was dissolved at 2 mg/ml in DMSO. Etoposide was dissolved as a 34 mM stock in DMSO. The membrane-permeant Ca²⁺ chelator BAPTA-AM (bis-(o-aminophenoxy)-N,N,N′,N′-tetraacetic acid, acetoxymethyl ester) was dissolved at 20 mM in DMSO. All stocks were stored at −20°C. In each experiment, drug-free controls containing DMSO only showed no effects. [Methyl-³H]thymidine (45.5 Ci/mmol) was obtained from Amersham International.

HeLa (human adenocarcinoma) and PtK₂ cells (Potorus tri dactylis, rat kangaroo) were grown as monolayers in Eagle’s minimal essential culture medium supplemented with 10% foetal calf serum, penicillin and streptomycin at 37°C, in an atmosphere of 5% CO₂.

Logarithmically growing HeLa were incubated with 2 mM thymidine for 24 h to arrest cells in S-phase. Thymidine-containing medium was removed and cells were allowed to continue through S-phase for 6 hours, after which a 9 h automatic nitrous oxide treatment (80 psi) was employed to achieve a mitotic block (Rao, 1968; Downes et al., 1987). Mitotic populations were isolated by shake-off and mitotic purity analysed after cytocentrifugation. For the clonal survival and cleavable complex assay, populations with over 97% mitotic cells were used.

Mitotic HeLa (isolated as described above) were treated for 1 h with ICRF-193, or DMSO only for control, washed thoroughly with warm PBS, then incubated with fresh medium for 7 days. Each dose was tested at least in triplicate.

Logarithmically growing HeLa were prelabelled for 48 h with 0.05 μCi/ml [³H]thymidine, then arrested in mitosis as above. Samples, consisting of 2×10⁵ mitotic cells in medium, were incubated for 15 min with ICRF-193 or etoposide; or for 15 min with ICRF-193 then with added etoposide for a further 15 min (DMSO concentration never >0.1%). Cells were pelleted by centrifugation, and medium removed, then cleavable complexes of inhibited topoisomerase II and DNA were prepared by a modification (Squires et al., 1991) of the method of Liu et al. (1983). Cells were lysed in alkaline sucrose solution (5% w/v sucrose, 10 mM Na₂EDTA, 0.1 M NaOH, 0.15 M NaCl) for 25 min on ice. Following neutralisation with 0.15 M K₂HPO₄, lysates were sonicated to reduce the size of the DNA. DNA cross-linked to protein was precipitated with 100 mM KCl, 1% SDS and 0.5 mg herring sperm DNA. The samples were vortexed, incubated on ice for 10 min, then the cross-linked precipitates were isolated as described by Liu et al. (1983). Radioactive counts in the precipitated samples were expressed as percentages of the total radioactivity in aliquots taken from each lysate before potassium/SDS precipitation.

Suspensions of mitotic HeLa were placed in poly-L-lysine-coated wells of Flow multiwell immunoassay slides (approximately 30 μl/well) in normal medium or medium containing ICRF-193 or BAPTA-AM (DMSO never >0.5%). Slides were incubated under normal conditions for 10-90 min, then fixed in −20°C methanol for 4 min. After being air dried, cells were stained with toluidine blue and mounted with DPX (Analar); or for laser confocal microscopy, cells were stained with propidium iodide (50 mg/ml with 50 mg/ml DNase-free ribonuclease, 10 min) and mounted with P-phenylamine diamine (1 mg/ml in 90% glycerol/10% PBS). Photographs were taken on Kodak Technical Pan film with a Nikon FE2 camera. Passage through anaphase/telophase was examined in samples of 200-400 cells for each drug concentration, accumulating data from three experiments.

A heated (37°C) and CO₂-equilibrated (5%) chamber surrounding a Nikon Diaphot inverted microscope was used to study living PtK₂ cells during cell division. The DMSO concentration never exceeded 0.1%. Cells were filmed with a Hitachi Denshi KP-C500E/K video camera and a Sony VO-9850P videocassette recorder. Still photographs were taken at intervals on Ilford pan F film.

Fig. 1 shows clonal survival for mitotic HeLa, treated with ICRF-193 for 1 hour (beginning immediately after the cells were released from a nitrous oxide metaphase block), relative to control cells treated with medium containing DMSO only (DMSO never >0.1%). At higher concentrations the potency of ICRF-193 levels off. HeLa are a particularly aneuploid cell line consisting of near-triploid and near-tetraploid populations. The lethality of ICRF-193 may only reflect killing of the cells with fewer chromosomes, since these cells should be more vulnerable to the effects of nondisjunction. Alternatively, mitotic cells that respond to ICRF-193 by generating only one post-mitotic daughter may survive at high doses of ICFR-193. It is also possible that the nitrous oxide synchrony affects a sub-population of cells, rendering them especially sensitive to topoisomerase II inhibition. We do not believe that the ICRF-193 resistant population corresponds to cells that had begun anaphase early, since HeLa cells released from a nitrous oxide block (which disorganises microtubules) require about 30 minutes to reform their metaphase plates and begin anaphase, while the ICRF-193 was added immediately after release.

The generation of topoisomerase II-DNA cleavable complexes by etoposide and ICRF-193 has been investigated in mitotic HeLa (Fig. 2a). In our assay we find that a 15 minutes incubation with 1-3 μM etoposide traps around 40% of the [³H]thymidine-labelled DNA. Under the same conditions ICRF-193 (up to 7 μM) fails to increase the radioactivity in the precipitates above background levels. We conclude, therefore, that ICRF-193 does not covalently link topoisomerase II molecules to DNA. We have also been unable to detect DNA strand breaks in mitotic HeLa treated with ICRF-193 (data not shown).

When a prior 15 minute incubation with ICRF-193 is followed by a 15 minute incubation with etoposide, there is a substantial suppression of etoposide-induced cleavable complexes (Fig. 2b). This is consistent with ICRF-193 inhibiting topoisomerase II and thus preventing etoposide from blocking active transition states. These results are in complete agreement with previous work in cell free systems (Tanabe et al., 1991; Ishimi et al., 1992) and with RPMI 8402 human leukaemic cells (Ishida et al., 1991), in which the mechanism of action of ICRF-193 has already been more thoroughly determined.

To quantify the effects of ICRF-193 on chromosome segregation, mitotic HeLa were incubated for various times on poly-L-lysine-coated multiwell slides. After fixation and staining, the frequency of chromosomes lagging in anaphase and of abortive anaphases (or of cytokinesis attempted in the absence of anaphase segregation) were scored. Examples of such aberrant anaphases are shown in Fig. 3. To confirm that the material defined as chromatin contained DNA, parallel slides were stained with propidium iodide for fluorescence microscopy (Fig. 3e and i). Fig. 4 collates data on anaphase disruption from three separate experiments. Clearly, very low concentrations of the drug (175-350 nM) are sufficient to increase the frequency of lagging figures, and very considerable inhibition of segregation is apparent at 1.75-7 μM ICRF-193 (50-75% abortive anaphases).

These effects of ICRF-193 parallel those already reported for etoposide, but cannot easily be attributed to DNA damage induced by the drug. However, ICRF-193 is a derivative of a calcium-chelating agent, and may act on intracellular calcium levels as well as inhibiting topoisomerase II. The role of cellular calcium ions in the control of mitotic events is disputed, but there is no doubt that changes in levels of free calcium can correlate with, though not necessarily cause, mitotic events (Hepler, 1989; Preston et al., 1991; Zhang et al., 1992). One might argue that ICRF-193 opposes anaphase progression by sequestering intracellular calcium. To test this hypothesis, we have treated mitotic HeLa, on release from arrest at metaphase by nitrous oxide, with the potent calcium chelator BAPTA, in the esterified form that can pass through cell membranes. HeLa possess adequate intracellular esterases to convert the molecules taken up into free BAPTA, which is trapped in the cells (Preston et al., 1991). No inhibition of anaphase segregation by BAPTA-AM, at concentrations up to 40 μM, was seen (data not shown). We therefore doubt that the anaphase inhibition caused by ICRF-193 is attributable to any calcium chelating activity it may possess.

The effects of inhibiting topoisomerase II by ICRF-193 during mitosis (without prior synchronisation) were studied in more detail by time-lapse video analysis. For this, PtK₂ cells were used, since they do not round up during mitosis and hence can be observed easily. ICRF-193 (1.4 μM) was given to selected prophase or metaphase cells and the effects on cycle progression monitored. Several video sequences were recorded and still photographs were taken of four mitotic cells treated with ICRF-193. Eight other mitotic cells were observed during the time-lapse analysis, which were not recorded photographically. The sequence of still photographs in Fig. 5 shows typical responses by a metaphase cell (cell 1, on the right) and a prophase cell (cell 2, on the left) to ICRF-193. Cell 2 took 16 minutes to reach metaphase; metaphase itself lasting for 18 minutes. In the abortive anaphases which ensued, chromatid separation did not occur completely. Cell 1 began anaphase after 12 minutes and after 30 minutes a highly abnormal cytokinesis was attempted. In control cells, cytokinesis usually follows complete chromatid separation; this clearly has not happened in this case. Anaphase was accompanied by apoptotic blebbing typical of abnormal cell division (Johnson et al., 1975), and during the next 10 minutes half of the cell was destroyed. The chromatin of the intact daughter cell began to decondense after 45 minutes. Cell 2 also undergoes an abnormal anaphase, although in this case the chromatin appears to have separated into several portions. In the PtK₂ time-lapse studies of Downes et al. (1991) control cells took an average of 27 minutes to progress from early metaphase to the onset of cytokinesis; cells treated with 60 μM etoposide during prophase took an average of 41 minutes. These data are comparable with the results we have obtained using 1.4 μM ICRF-193: metaphase to cytokinesis taking approximately 30 minutes (cell 1, which was already in metaphase when the cell was identified) and 44 minutes (cell 2). We note that no effect of ICRF-193 on the adjacent interphase cells can be observed.

This work establishes that inhibiting mitotic topoisomerase II activity in mammalian cells with ICRF-193 prevents anaphase chromatid segregation, without inflicting measurable damage on DNA. This effect is the same as that previously reported for agents such as etoposide, which binds to topoisomerase II and arrests its action after it has created DNA strand breaks; or m-AMSA, which intercalates into DNA and prevents the resealing of DNA strand breaks formed by topoisomerase II (Downes et al., 1991). The common element in the action of all these agents is topoisomerase II inhibition. The effects of ICRF-193 show that topoisomerase II activity is required for anaphase segregation in mammalian cells, as it is in yeast, in conditions where interpretation is not complicated by side effects of DNA damage. We cannot formally exclude the possibility that ICRF-193 inflicts a small number of DNA lesions in mitotic mammalian cells, too few to be distinguished from background in our assays. However, other workers have consistently failed to detect ICRF-193-induced DNA damage in other systems. And with other inhibitors of topoisomerase II, which act by creating cleavable complexes and DNA strand breaks, considerable amounts of damage can be inflicted without much retarding mitotic progression (Downes et al., 1991; Shamu and Murray, 1992). This work leaves open the question of whether the previously reported action of ICRF bisdioxopiperazines in delaying progression through G₂ into mitosis is also a direct effect of topoisomerase II inhibition.

The action of ICRF-193 in inhibiting segregation (Figs 4-6) is consistent with previous work using etoposide. To quantify the effects of ICRF-193 on chromosome segregation during anaphase, we have scored lagging chromosomes and aborted anaphases. ICRF-193 is far more potent than etoposide, significantly increasing the frequency of anaphase laggards at 350 nM and causing over 40% of anaphases to abort at 1.75 μM. Concentrations of 60-80 μM or 10-30 μM etoposide would be required to inhibit topoi-somerase II to this extent in HeLa cells (Downes et al., 1991) or in the Xenopus extract system of Shamu and Murray (1992), respectively.

The exact topoisomerase II target of ICRF-193 is not known. In mammalian cells there are two kinds of topoi-somerase II; the 170 and 180 kDa proteins, encoded by genes on chromosomes 17 and 3, respectively. Expression of the 170 kDa protein increases markedly in G₂ and mitosis (Woessner et al., 1991) and it has been identified as an integral component of mitotic chromosomes prepared from primary chicken fibroblasts or from Chinese hamster cells (Earnshaw et al., 1985; Earnshaw and Heck, 1985; Charron and Hancock, 1990). In Escherichia coli, segregation of replicated, catenated DNA, and subsequent separation of the circular supercoiled chromosomes at cell division, are known to be effected by a special enzyme, topoisomerase IV, which like the bacterial topoisomerase II is a class 2 topoisomerase capable of decatenation (Adams et al., 1992). It would be satisfyingly symmetrical if the mammalian 170 kDa protein turned out to be analogous to the bacterial topoisomerase IV.

The 170 kDa species is known to be inhibited preferentially by merbarone, a topoisomerase II inhibitor, which appears to act in a similar way to ICRF-193, without generating DNA breaks (Drake et al., 1989). The relative effects of ICRF-193 on the 170 and 180 kDa species of topoisomerase II have not been investigated, nor is it known whether merbarone inhibits mitosis. We would predict that in mitotic cells the target of ICRF-193 would be the 170 kDa species, and that merbarone should likewise inhibit anaphase segregation.

In time-lapse studies of PtK₂ cells, we observe that cells given ICRF-193 in prophase appear to complete prophase without delay: we have not observed problems in the late stages of chromatin condensation such as are seen in the Top2 yeast mutants (di Nardo et al., 1984; Uemura et al., 1987; Holm et al., 1989); and in in vitro systems (Wood and Earnshaw, 1990; Adachi et al., 1991; Hirano and Mitchison, 1991). Time-lapse studies using etoposide to inhibit topoisomerase II similarly failed to provide evidence for topoisomerase II function in prophase (Downes et al., 1991). However, Ishimi et al. (1992) demonstrated that ICRF-193 blocks the late, decatenatory stages of SV40 chromosome replication; and Ishida et al. (1991) observed extremely tangled chromatin (which inhibited cycle progression through prophase) in prophase after lymphocytes were incubated with 10 μM ICRF-193 for 15 hours. We interpret these data as showing that decatenation is largely completed in S-phase and G₂. Short pulses of ICRF-193 or etoposide during prophase presumably fail to perturb condensation since sufficient decatenation has already taken place.

The time-lapse studies with PtK₂ confirm the effects of ICRF-193 in unsynchronised cells. This eliminates the possibility that the effects seen in synchronised mitotic HeLa are due to the nitrous oxide treatment which destablises the microtubules of the spindle apparatus during synchrony; such destabilisation could arguably result in aberrant anaphase segregation in some cells after the block is released. When given in prophase or metaphase to unsynchronised PtK₂ cells, ICRF-193 prevents anaphase chromosome segregation. We suppose that residual catenations between sister chromatids, which are not enough to prevent chromosome condensation, are present at least until metaphase. Although the period from full metaphase until the onset of cytokinesis is prolonged, cycle progression is not blocked and cytokinesis continues, albeit aberrantly. Hence, as in yeast, mammalian cells need topoisomerase II for chromatid segregation, but neither topoisomerase II activity nor chromatid segregation are necessary for the terminal stages of cell division. Previous work with etoposide left this conclusion uncertain (although probable) since etoposide-induced DNA strand breaks could conceivably have caused aberrant cycle progression, with uncoupling of segregation and cytokinesis.

In a very recent paper, Sumner (1992) reported that inhibitors of topoisomerases do not block the separation of human lymphocyte chromatids during prolonged colcemid arrest. He correctly points out that the inhibitors of topoi-somerase II previously used have a variety of non-specific effects, including the production of DNA damage. This criticism does not apply to the topoisomerase II inhibitor ICRF-193. Sumner used very low concentrations of etoposide and other drugs in an attempt to limit the affects of DNA damage; unfortunately these low concentrations, which had little or no effect on chromatid segregation, may not have adequately inhibited topoisomerase II either.

We have here confirmed that topoisomerase II action is required for anaphase chromatid segregation in mammalian cells. In the absence of segregation, cytokinesis is not prevented. Hence, there is apparently no mammalian ‘decatenation complete?’ checkpoint for exiting from mitosis.

We thank Charles ffrench-Constant for use of the time-lapse equipment, Tony Mills for instruction in confocal microscopy, Paul Smith for etoposide, Tim Cheek for BAPTA-AM, and especially Dr Andrew Creighton, Department of Rheumatology, St. Bartholomew’s Hospital, London, for a generous gift of ICRF-193 and information about its history. We are grateful to the Cancer Research Campaign, of which R.T.J. is a Fellow, for continued support, and to the Medical Research Council for a studentship to D.J.C.