Vol. 7, 83-90, January 1996 Cell Growth & Differentiation 83

Topoisomerase II Inhibitors Affect Entry into Mitosis andChromosome Condensation in BHK Cells1

Hilary Anderson and Michel Roberge2

Department of Biochemistry and Molecular Biology, University of British

Columbia, Vancouver, British Columbia, V6T 1Z3 Canada

AbstractDNA topoisomerase II (topo II) is required at mitosis inyeast for high chromosome condensation and forchromosome segregation. Recent studies on intactmammalian cells using topo II inhibitors that do notstabilize cleavable complexes also suggest arequirement for topo II for complete chromosomecondensation and perhaps also for entry into mitosis.We have investigated the effects of merbarone, ICRF-187, and aclarubicin, three topo II inhibitors that do notstabilize the cleavable complex, on entry into mitosisand on chromosome condensation in BHK and intsBN2 cells. We have compared their effects with thoseof etoposide, a topo II inhibitor that stabilizes thecleavable complex. All inhibitors induced aconcentration-dependent G2 delay or arrest that couldbe overcome with fostriecin or okadaic acid or byinactivation of RCCI in tsBN2 cells. Mitoticchromosomes from cells treated with etoposidewere extensively fragmented, whereas mitoticchromosomes from cells treated with merbarone,ICRF-187, or aclarubicin were intact but elongated andtangled. These results provide strong evidence thattopo II activity is required in chromosomecondensation for final coiling of the chromatids. Ourresults also indicate that protein phosphatases andRCCI play a role in G2 delay induced by all inhibitors,whether they do or do not stabilize the cleavablecomplex.

IntroductionThe enzyme DNA topo lI� can alter DNA topology by tran-siently introducing a double-strand break in a DNA segmentand passing another DNA segment through the break beforeresealing it (reviewed in Refs. 1 and 2). Topo II is the mostabundant component of the chromosome scaffold and im-

Received 8/9/95; revised 10/2/95; accepted 10/30/95.The costs of publication of this article were defrayed in part by thepayment of page charges. This article must therefore be hereby markedadvertisement in accordance with 1 8 U.S.C. Section 1 734 solely to mdi-cate this fact.1 This work was supported by a grant from the Medical Research Councilof Canada. M. R. is a Scholar of the Medical Research Council of Canada2 To whom requests for reprints should be addressed. Phone: (604) 822-2304; Fax (604) 822-5227.3 The abbreviations used are: topo II, topoisomeraso II; VP-16, etoposide;SPCC, S-phase premature chromosome condensation; VM-26, tenipo-side.

munolocalizes to the core of the long axis of the chromatids

of metaphase chromosomes (3), where it could play a role inanchoring loops of DNA to the scaffold. Genetic studies in

yeast have shown that topo II is not required for cells to entermitosis but is required for disentangling sister chromatids

and for the additional chromosome condensation (hypercon-densation) that occurs when cells are treated with microtu-

bule inhibitors (4-8).No topo II conditional mutants are available for mammalian

cells, but other approaches have been devised to examinethe role of topo II in mitosis. Mitotic extracts from chicken

cells or Xenopus eggs are able to convert interphase nucleito mitotic chromosomes. Chromosome condensation was

observed when nuclei containing high levels of topo II wereadded to mitotic extracts immunodepleted of topo II, while

no condensation was observed with nuclei containing little orno topo II (9-11). Adding purified topo II back to immunode-pleted extracts reestablished their ability to condense chro-

mosomes (1 0). These studies showed that topo II is required

for chromosome condensation but did not establish whetherthe requirement is early or late in the condensation process.

Whether its role is structural or enzymatic is still unclear (11,12).

The role of topo II in mitosis has also been examined inliving cells by using topo II inhibitors. A large number of topo

II inhibitors act by stabilizing the cleavable complex, a co-valent topo Il-DNA intermediate in which the DNA strands arecut (13, 14). These inhibitors block entry of cells into mitosis(1 5-1 7). This could be because topo II activity is required for

entry into mitosis (15, 16) or because the stabilized cleavablecomplex is recognized as DNA damage (1 5, 18). Cells re-

spond to DNA damage by activating a checkpoint that ar-rests cells in G2, presumably to allow time for DNA repair

before entry into mitosis (1 7, 19-21).More recently, drugs have been identified that inhibit topo

II at an earlier stage in its reaction and do not stabilizecleavable complexes in vivo. They include merbarone (22,

23), aclarubicin (23, 24), and ICRF-187 (25). ICRF-187 is themore readily soluble and clinically useful D-isomer of theracemic ICRF-159. Other compounds in this family include

the mother compound ICRF-1 54 and the derivatives ICRF-1 93 and MST-1 6, all of which inhibit topo II without stabilizing

cleavable complexes. The effects of such inhibitors on entryinto mitosis are still unclear. Merbarone causes G2 arrest inGEM and HeLa cells (23, 26). IGRF-154, ICRF-159, andIGRF-1 93 are reported to have caused G2 arrest in several

mammalian cell lines, including HeLa and CHO cells (27-29),whereas in other studies, IGRF-1 93 caused G2 delay in HeLa

cells but not in CHO cells (30), and IGRF-187 and IGRF-159

did not cause G2 arrest in Ptkl cells (31).

A B

Merbarone

COOCH,

H

CHr�

cH,o_.J:c?._%oc:l,

VP-16

ICRF- 187

20 20.�� .s0

-�015 ,, �150 / 0�0 �/ .2 rn.�� ,

� -0 1 � � Li 5

tlme(h)

C D

Aclarubicin

Fig. 1. Chemical structures of the four topo II inhibitors used in thisstudy.

.� 20

.�

0 15

2 10E

U �, , ,

0 1 2 3 4

time(h)

40

�30

0

time (h)tlme(h)4

Fig. 2. Percentage of mitotic BHK cells at different times after treatmentat 0 h with different concentrations of topo II inhibitors and at 2 h without(0, 0, L�) or with (#{149},U, A) 100 pi�i fostriecin. A: 0, #{149},no topo II inhibitors;D, U, 10 � VP-16. B: L�, A, 10 �t�i morbarono; 0, U, 100 �M morbarone.

C: �, A, 10 �M ICRF-187; 0, U, 100 �M ICRF-187. D: 0, #{149},no topo IIinhibitors; & A, 0.1 �M aclarubicin; 0, U, 1 �M aclarubicin.

84 Topoisomerase II and Mitosis

treatment with protein phosphatase inhibitors or temperature

In the study of Gorbsky (31) in which G2 arrest did notoccur and cells were able to enter mitosis in the presence ofICRF-187 or ICRF-159, chromosomes were long, incom-pletely condensed, and entangled, suggesting that topo II isrequired for the final Stages of condensation. Although IGRF-193 dramatically reduced entry into mitosis of RPMI 8402cells, the few cells that did enter mitosis had a similar ap-pearance (28). However, these could be regarded as anunrepresentative sample of cells. In two other cases in whichG2 delay did occur, this was overcome experimentally so thatmany cells entered mitosis (29, 30). Ishida et a!. (30) treatedG2-synchronized tsBN2 cells with ICRF-1 93. The tsBN2 cellline is a temperature-sensitive mutant of the BHK line andhas a mutation in the RCC1 gene. The RCC1 protein isthought to be part of the mitotic entry checkpoint, whichprevents entry into mitosis until DNA is fully replicated ordamaged DNA is repaired; at the nonpermissive tempera-

ture, the RCC1 protein is inactivated, and cells can entermitosis even with unreplicated or damaged DNA (32, 33).Cells that entered mitosis after shifting to the nonpermissivetemperature in the presence of ICRF-193 had long and tan-

gled chromosomes (30). In the second study, Downes et aL(29) showed that G2 delay induced by ICRF-1 93 in IndianMuntjac cells can be ovemden by caffeine to produce im-perfectly condensed, entangled chromosomes. Thus, the

results published with mammalian cells, using inhibitors ofthe ICRF series, suggest that topo II is required at a late stagein chromosome condensation.

Results obtained with a single inhibitor must be interpreted

with caution because they could result from action on anunrecognized target. Therefore, we used three structurallyunrelated topo II inhibitors that do not stabilize the cleavablecomplex, as well as VP-i 6 that does stabilize the cleavablecomplex, to study the importance of topo II on entry intomitosis in BHK cells. We also used two different methods,

shift of tsBN2 cells, that overcome G2 arrest elicited by thetopo II inhibitors to examine the role of topo II in chromosomecondensation.

Results

VP-16; Merbarone, Aclarubicin, and ICRF-187 CauseG2 Delay or ArrestFig. 1 shows the structure of the four topo II inhibitors used.VP-i 6 belongs to the class of topo II inhibitors that stabilizethe cleavable complex (1 3). Merbarone, IGRF-1 87 and ada-rubicin have unrelated structures and do not stabilize thecleavable complex (22-25). We first determined whether thetopo II inhibitors that do not stabilize the cleavable complexalso induce G2 delay in BHK cells.

Unsynchronized cycling cells were treated with Golcemidat 0 h to trap cells in mitosis. Samples were taken at 0, 2, and4 h, chromosome spreads were prepared, and the percent-age of mitotic cells was determined. The percentage of mi-totic cells rose linearly at a rate of 3.4% per hour in oneexperiment (Fig. 2A) and 4.7% per hour in another experi-ment (Fig. 2D). Additional cell samples were treated withColcemid and with different topo II inhibitors at 0 h. These,and all subsequent experiments, were also carried out withthe drug solvents DMSO, ethanol, or water, and these had noeffect on entry into mitosis or chromosome condensation(results not shown).

The results of a typical experiment using VP-i 6, merbar-one, and IGRF-i87 are shown in Fig. 2, A-C, and of another

B

25In

.� 200

:� 150

.� 10

a05

pI

‘A‘/

I,

I,‘I

‘I

lime(h)

C

1 234

time (h)

D

30.5?0000

E

C,)

a)C)

C).�

0

Fig. 4. Percentage of mitotic tsBN2 cells at different times after treat-mont of a synchronized 32#{176}Cpopulation at 0 h with different topo IIinhibitors (D; all inhibitors showed 0% mitotic cells) and at 1 h without (0,0) or with (+, #{149}, A, #{149},Y) temperature shift to 40.5#{176}C.0, #{149},no topo IIinhibitors; +, 10 �u�i VP-16; A, 100 � morbarone; #{149},100 pi�i ICRF-187;V. 1 �M aclarubicin.

20

10

)0

fime (h)

0 f �time(h)

4

Fig. 3. Percentage of mitotic BHK cells at different times after treatmentat 0 h with different concentrations of topo II inhibitors and at 2 h without(0, 0, L�) or with (#{149},U, A) 0.5 �u�i okadaic acid. A: 0, #{149},no topo IIinhibitors; 0, #{149},10 pi� VP-16. B: L�, A, 10 �.w merbarone; 0, �, 100 �merbarone. C: L�, A, 10 �M ICRF-187; 0, #{149},100 �M ICRF-187. D: 0, #{149},no topo II inhibitors; L�, A, 0.1 �i�vi aclarubicin; 0, #{149},1 �.w aclarubicin.

Cell Growth & Differentiation 85

25

ID 20� A

�15 /

llme(h)

experiment using aclarubicin in Fig. 2D. Ten �M VP-i6 com-pletely blocked entry into mitosis (Fig. 2A), as expected. Onehundred p.M merbarone also completely blocked entry intomitosis (Fig. 2B). A i 0-fold lower concentration of merbarone(iO p.M) did not cause any detectable G2 delay (Fig. 2B). Onehundred �.LM ICRF-i87 delayed entry into mitosis, permittingonly 5.5% of cells to enter mitosis over the 4-h period, whilea i 0-fold lower concentration of IGRF-i 87 had little or noeffect (Fig. 2C). Similarly, i p.M aclarubicin induced a delay ofentry into mitosis, permitting only 4.2% of cells to entermitosis (Fig. 2D). A i 0-fold lower concentration of aclarubicincaused a less marked G2 delay (Fig. 2D). Thus, the threeinhibitors that do not stabilize the cleavable complex caninduce G2 arrest or significant G2 delay in a dose-dependentmanner.

G2 Delay Induced by Topo II Inhibitors Is Partially orCompletely Overcome by Fostriecin or Okadaic AcidG2 delay caused by treatment with VP-i 6 and its congenerVM-26 can be overcome with agents that interfere with themitotic entry checkpoint (29, 34). To investigate whether G2delay induced by topo II inhibitors that stabilize or do notstabilize the cleavable complex operates by a similar mech-anism, we first determined whether it can be overcome byfostriecin and okadaic acid, two structurally unrelated proteinphosphatase inhibitors that can override the checkpoint (34).

Cycling BHK cells were treated with Golcemid and with orwithout topo II inhibitors at 0 h. After 2 h, when G2 delay was

observed, the cells were additionally treated with fostriecin orokadaic acid for 2 more hours, and the percentage of mitoticcells was determined. One hundred �LM fostriecin partiallyovercame the G2 arrest caused by VP-i 6, merbarone, andIGRF-i 87 (Fig. 2, A-C). Fostnecin overcame the G2 delaycaused by 0.1 �M aclarubicin but not the G2 arrest caused by

1 �LM aclarubicin (Fig. 2D). Okadaic acid (0.5 �.LM) completelyovercame the G2 delay caused by all four topo II inhibitors(Fig. 3).

These experiments show that G2 delay caused by VP-i 6,merbarone, IGRF-i 87, and aclarubicin can be overcome ex-perimentally and is, therefore, not due to an inability of thecells to enter mitosis but to an active response of the cells.They suggest that G2 delay induced by the different drugs iscaused by a common checkpoint that requires protein phos-phatases.

G2 Delay Induced by Topo II Inhibitors Is Overcomeby Inactivation of RCCIThe mitotic entry checkpoint can also be overcome by atemperature shift in the tsBN2 cell line (32). We next deter-mined whether the G2 delay induced by VP-16, ICRF-187,

merbarone, and aclarubicin can also be overcome by tem-perature shift in the tsBN2 cell line.

Cycling tsBN2 cells maintained at the permissive temper-ature of 32#{176}Cwere synchronized to G2 phase by isoleucinedeprivation and aphidicolin treatment, followed by releaseinto normal medium (see “Materials and Methods”). At thistime, defined as 0 h, the cells were treated or not with topoII inhibitors. One h later, some cultures were shifted to thenonpermissive temperature of 40.5#{176}C.All cultures were in-cubated for an additional 3 h. Samples were collected at 0,1 , and 4 h, chromosome spreads were prepared, and thepercentage of mitotic cells was determined.

Fig. 4 shows that control cultures had only 3.7% mitoticcells after 4 h at 32#{176}C(note that Colcemid was not used in

86 Topoisomerase II and Mitosis

Fig. 5. Morphology of mitotictsBN2 cells after treatment of asynchronized 32CC populationwith different topo II inhibitorsfollowed 1 h later by temperatureshift to 40.5CC for 3 h. A, B, andC, typical control mitotic cellswith no topo II inhibitors for com-parison with subsequent rowstreated with 10 �M VP-i 6 (D, E,and F), 1 00 MM merbarone (G, H,and I), 100 �.LM ICRF-187 (J, K,and L), or 1 ,.LM aclarubicin (M, N,and 0). Each row shows therange of morphologies ob-served. Only A was not sub-jected to temperature shift. Bar,20 Mm.

these experiments), but this rose to i 1 .3% after temperature

shift. No mitotic cells were observed in any of the cultures

that had been treated with topo II inhibitors and maintained

at 32#{176}C.However, 3 h after temperature shift, all treatedcultures had mitotic cells: 2i.3% for VP-i6, 15.1% for mer-barone, i7.8% for ICRF-i87, and 9.8% for aclarubicin.These results show that G2 delay caused by the two classes

of topo II inhibitors requires RCC1 and suggest that these

drugs activate a checkpoint involving RCC1 in addition toprotein phosphatases.

Topo II Activity Is Required for Final Stages ofChromosome CondensationThe observation that G2 delay or arrest caused by the dif-ferent topo II inhibitors can be overcome provided a means

of studying the role of topo II in chromosome condensation

in living mammalian cells. We treated cells with the different

topo II inhibitors, overcame the resulting G2 delay by tem-

perature shift of tsBN2 cells or by treatment of BHK cells with

protein phosphatase inhibitors, and then examined the mor-

phology of their chromosomes by fluorescence microscopy.

tsBN2 Cells. tsBN2 cells were synchronized to G2,treated or not with topo II inhibitors for 1 h at 32#{176}C,and then

forced into mitosis by shift to the nonpermissive temperature

for 3 h as for Fig. 4. Mitotic tsBN2 cells at the nonpermissive

temperature are either metaphase-like, showing condensed

chromosomes with visible sister chromatids (Fig. 58) similarto normal mitotic cells at the permissive temperature (Fig.

5A), or are SPCC-like (Fig. 5C). The latter exhibit a morphol-ogy characteristic of premature chromosome condensation

from cells with incompletely replicated DNA in which the

chromatin appears “pulverized” (32).

Mitotic cells obtained by temperature shift after treatment

with VP-i 6 do not have normal mitotic chromosomes. They

are all SPCC-like, exhibiting fine clouds of condensed chro-

matin fragments (Fig. 5, D, E, and F). Merbarone, ICRF-i87,

Table 1 Induction of SPCC-like morphology and condensedchromosomes in tsBN2 cells treated with topo II inhibitors prior to shiftto .40.5�Ca

C -

Fig. 6. Morphology of mitotic BHK cells after treatment of a cyclingpopulation with different topo II inhibitors followed 2 h later with fostriecinfor an additional 2 h. A and B, typical control mitotic cells with no topo IIinhibitors for comparison with subsequent rows treated with 1 0 MM VP-i 6(C and D), 100 MM merbarone (E and F), 100 MM ICRF-187 (G and H), or

1 MM aclarubicin (I and J). Each row shows the range of morphologiesobserved. Only A was not treated with fostriecin. Bar, 20 �m.

Cell Growth & Differentiation 87

Topo II inhibitor % SPCC-like % chromosomes

- 4.3 7.0

VP-16 21.3 0

Merbarone 11.0 4.2

ICRF-187 3.1 14.7Aclarubicin 1.3 8.4

a Cells were treated with the topo II inhibitors at 0 h, shifted to 40.5CC at

2 h, and harvested at 4 h. Mitotic spreads were prepared and SPCC-likemorphology and condensed mitotic chromosomes were scored.

or aclarubicin treatment followed by temperature shift re-

suIted in mitotic cells with condensed chromosomes (Fig. 5,

G, H, J, K, M, and N) or with an SPCC-like morphology (Fig.

5, I, L, and 0). Typical examples of condensed chromo-

somes are shown in Fig. 5, H, K, and N, where they are long

with tangled sister chromatids. Examples of the most con-

densed chromosomes observed are shown in Fig. 5, G, J,

and M; even these were considerably longer than controls

(Fig. 5, A and B).

The percentage of cells with condensed chromosomes or

with an SPCC-like morphology after the different treatments

is shown in Table 1 . Although the cells were synchronized to

G2 at the time of drug treatment and temperature shift, the

presence of SPCC-like cells in addition to metaphase-like

cells in control cultures indicates that synchronization is im-

perfect with some cells still in S phase. Temperature shift

alone produced 4.3% SPCC-Iike cells, representing the per-

centage of cells driven into mitosis from S phase, and 7.0%

cells with condensed chromosomes, representing those

driven into mitosis from G2.

Pretreatment with VP-i 6 resulted in 21 .3% SPCC-like

cells and none with condensed chromosomes. Of these,

about 4.3% result from imperfect synchronization and are

cells that were in S phase at the time of treatment. The

remaining 17% represent cells driven into mitosis from G2

and provide a visual demonstration that VP-16 treatment

results in severely broken DNA. No intact chromosomes are

seen because no long stretches of DNA remain uncut by

topo II.

Cells treated with ICRF-i 87 and aclarubicin showed no

increase in the percentage of SPCC-like cells over controls,

indicating that these drugs produce little or no DNA damage.

No breaks were detected in the condensed chromosomes

(Fig. 5). Cells treated with merbarone showed 1 1 .0% SPCC-

like cells, suggesting that merbarone causes some DNA

damage. However, 4.2% metaphase-like cells with appar-

ently unbroken chromosomes were also observed.

BHK Cells. Cycling BHK cells were also treated or not

with topo II inhibitors for 2 h and then induced to enter

mitosis by treatment with fostriecin or okadaic acid in the

presence of Colcemid for 2 h. Cells were collected at the end

of the experiment, and chromosome spreads were prepared.

Results obtained with okadaic acid were indistinguishable

from those obtained with fostriecin and are not further de-

scribed here.

E

�l i4�4�

B �

‘C..

:11*D

F

G �

-�

�‘I

. � 1_.I�

r�r;*

ri�q#{231}�

Almost all mitotic cells obtained after treatment with Col-

cemid alone show highly condensed chromosomes typical of

metaphase (Fig. 6A). After treatment with fostriecin and Col-

cemid, most mitotic cells are also metaphase-like (Fig. 6B)

and are indistinguishable from those treated with Colcemid

alone. SPCC-Iike cells are observed rarely in these nonsyn-

chronized cells.

Cells treated with VP-16 and then fostriecin never look

metaphase-like; they are SPCC-like with patches of con-

densed chromatin (Fig. 6, C and D). As described above for

tsBN2 cells, these represent cells driven into mitosis from G2

but with severely damaged DNA.

Mitotic cells obtained after treatment with merbarone (Fig.

6, E and F), ICRF-1 87 (Fig. 6, G and H), or aclarubicin (Fig. 6,

I and J) and then fostriecin have clearly condensed chromo-

somes. However, as described above for tsBN2 cells, these

88 Topoisomeraso II and Mitosis

chromosomes are never as condensed as the metaphasechromosomes seen in controls (Fig. 6, A and B), even for themost condensed examples observed (Fig. 6, E, G, and I), andtheir sister chromatids are tangled. SPCC-like cells wereseen rarely.

These results show that chromosomes of cells forced toenter mitosis in a variety of ways after treatment with avariety of topo II inhibitors remain tangled and are unable toachieve complete condensation.

DiscussionThe topo II inhibitors used in this study have quite differentstructures (Fig. i) and different mechanisms of inhibition.

VP-i 6 has been studied extensively and inhibits topo II ad-tivity by stabilizing the cleavable complex so that religationdoes not occur (13). Merbarone, ICRF-i 87, and aclarubicinreduce the formation of stabilized cleavable complexes byVP-i 6 or the related compound VM-26 and, therefore, inhibittopo II at an earlier stage in its action (22-25, 35). ICRF-i 93,a close relative of ICRF-1 87, stabilizes the enzyme in theform of a closed protein clamp by inhibiting its ATPaseactivity (36), presumably converting unbound topo II to a

form that is unable to bind DNA, and DNA-bound topo II toa form that is unable to leave the DNA. Aclarubicin has beenshown to prevent topo II from performing its initial noncova-lent DNA-binding reaction (37). Aclarubicin also inhibits theeffect of the topo I inhibitor camptothecin, suggesting it alsoinhibits topo I (38). Aclarubicin intercalates DNA, whereas theother inhibitors do not.

Topoisomerase II Inhibition and G2 Delay. We showhere that all four topo II inhibitors induce G2 arrest or delay.Ten � VP-i 6 completely blocked entry into mitosis duringthe 4 h following treatment. The ability of VP-i 6 to induce G2arrest has been documented in many cell lines (1 5-i 7, 39).One hundred �tM merbarone also completely blocked entryinto mitosis of BHK cells during the 4 h, in agreement withstudies in GEM cells (23) and HeLa cells (26). One hundredjA.M ICRF-i 87 did not have such a pronounced effect asVP-i 6 or merbarone, reducing the rate of entry into mitosisrather than completely blocking it in BHK cells. In studieswith a similar time course, ICRF-i 59 and ICRF-i 87 did notappear to cause G2 arrest in Ptki cells (3i). An early study

with ICRF-i 59 showed that it caused G2 arrest in humanperipheral blood lymphocytes (27). ICRF-i 93 produced a G2delay in six different cell lines including CHO cells in onestudy (29) and a G2 delay in HeLa cells but not in CHO cellsin another study (30). Ten �i�i aclarubicin also produced a G2delay in BHK cells. This was the highest concentration wewere able to use without extreme toxic effects. There havebeen no comparable studies of the short-term effect of ada-rubicin on entry into mitosis in other cell lines.

There is no obvious explanation for the contradictory re-suIts on G2 arrest obtained with ICRF compounds. In mam-malian cells, topo II is a homodimer composed of M, i 70,000(also called a) or Mr 1 80,000 subunits. The Mr 1 70,000 formis maximally expressed in proliferating cells at G2-M,whereas expression of the M� 1 80,000 form is relatively con-stant through the cell cycle. Merbarone preferentially inhibitsthe Mr i 70,000 form rather than the Mr 180,000 form in vitro

(35). Merbarone produced a complete G2 block in GEM cells,but in mutant sublines which express mutant topo Ia, cellsescaped this block, entered mitosis, and had abnormally

condensed chromosomes (23). It is possible that the Ptki

cells used by Gorbsky (3i) have a topo II that is less sensitiveto ICRF i 87 or IGRF-i 59. However the discrepancy betweenthe results obtained with the same cell line CHO with thesame drug IGRF-i93 in the studies of Downes et a!. (29) andIshida et aL (30) is not explainable in this way.

Cells respond to DNA-damaging agents by activating acheckpoint which prevents entry into mitosis, presumably toallow time for DNA repair(i 7, 19-21). The cleavable complexstabilized by agents such as VP-i 6 is thought to be recog-nized as a form of DNA damage by the cell, resulting in G2block (1 5, i 8). Fluorescence microscopy of VP-i 6-treatedcells driven into mitosis provides a visual demonstration thatthe DNA is indeed extensively cut (Figs. 5 and 6). Therefore,our results and those of others showing that topo II inhibitorsthat do not stabilize the cleavable complex nevertheless

induce G2 arrest or delay were unexpected.Do the inhibitors that do not stabilize the cleavable com-

plex nevertheless induce DNA damage? This question hasbeen addressed using a number of different techniques, all ofwhich reveal extensively broken DNA in the presence ofVP-i6 or VM-26.

Chen and Beck (26) used microscopy to detect whetherDNA has been cut by determining whether it can be electro-phoresed from lysed agarose-embedded cells to form halosaround the nucleus. No clear halos were seen in merbarone-treated HeLa cells, indicating that this drug produced fewDNA breaks (26). An alkaline elution assay showed that mer-barone produces a small but consistent number of protein-concealed DNA breaks in Li 2i 0 cells (22). We show, usingfluorescence microscopy of mitotic spreads, that merbarone

produces a greater percentage of SPCC-Iike tsBN2 cells

than does temperature shift alone, suggesting that it causesDNA damage in some cells. However, we also observedmany cells with a set of apparently intact chromosomes,suggesting that merbarone caused little or no DNA damagein these cells.

In an early study, treatment of human peripheral lympho-cytes with about 2 p.M ICRF-i 59 for 3 h produced variouslesions in i 9 of i 00 cells examined: 19 gaps, 3 breaks, oneexchange, and one minute (27). ICRF-i 87 up to i mrn did notresult in DNA fragmentation in the alkaline elution assay (25),and ICRF-i93 showed no increase in DNA breaks in HeLacells using an alkaline unwinding assay (29). Background forthis assay is assessed at about 0.2 single-strand breaks!megabase, or about 200 breaks!celI. Similarly, we observed

many BHK or tsBN2 cells treated with ICRF-i 87 with appar-ently intact metaphase chromosomes (Figs. 5 and 6).

Aclarubicin is reported to not generate DNA-protein com-

plexes in intact cells (23). Aclarubicin at up to 100 �M pro-duced no increase in the percentage of DNA concealed byprotein over background in human small cell lung cancer celllines (24) and did not increase the percentage of SPCC-likecells or show broken metaphase chromosomes in this study.

In summary, there is evidence that merbarone can pro-duce DNA breaks, but no evidence that either ICRF-i 87 and

Cell Growth & Differentiation 89

related compounds or adlarubicin do so. It is possible thatthe techniques used are not sensitive enough to detect a

small number of DNA breaks. The sensitivity of the mamma-han checkpoint to DNA breaks is not known, but the yeastcheckpoint can be activated by a single double-strand break(40). It is also possible that topo II inhibitors that do notstabilize the cleavable complex may instead induce lesionsnot detected by the techniques described above but which

can induce G2 arrest. In addition, we note that the strength ofthe G2 delay observed in this study (VP-i 6 � merbarone>

ICRF-i 87 = adlarubicin) seems to correlate with the propen-

sity of the different drugs to generate DNA strand breaks(VP-i6 > merbarone > ICRF-i87 = adlarubicin) rather thanwith their topo II inhibitory potency; all inhibit topo II strongly

in vivo at the concentrations used (22-25, 35). It is possiblethat ICRF-i 87 and aclarubicin cause G2 delay through the

generation of an undetectably small, but sufficient, numberof DNA breaks.

Our results are also compatible with the recent proposalthat there exists a topo Il-dependent mitotic entry checkpointdistinct from the DNA-damage checkpoint (29). These au-thors suggest that the main function of this checkpoint is toensure sufficient decatenation of replicated DNA before acell attempts mitosis. Such a checkpoint does not exist inyeast (7). However, decatenation is generally believed to takeplace during mitosis rather than before mitosis (8, 30). In-deed, it is difficult to conceive how decatenation of entan-

gled duplicated chromosomes could take place efficientlyduring interphase, without spatial separation of the sisterchromatids. Decatenation may be possible only at mitosis,when the process of chromosome condensation could helpprovide this spatial separation (8). Our observation that G2delay caused by all four inhibitors can be overcome withfostriecin, okadaic acid, or inactivation of RCC1 suggeststhat if a distinct decatenation checkpoint exists in mamma-ian cells, it seems to operate via the same biochemical

pathway as the checkpoint that monitors DNA damage andreplication. Whether a decatenation checkpoint exists inmammalian cells will be resolved when topo II activity can bemanipulated genetically rather than biochemically.

Topoisomerase II and Chromosome Condensation.Topo II is required for high chromosome condensation in

yeast (7). In mammalian cells, work with ICRF compoundssuggests that this is also the case (29-31). Because of theuncertainties associated with the use of a single inhibitor,we examined this question in more detail using differenttopo II inhibitors and different ways to force treated cellsto enter mitosis. We show that in all cases, entry intomitosis in the presence of inhibitors that do not stabilizethe cleavable complex results in chromosomes that arevisibly condensed but which remain long and entangled.The observation that three structurally unrelated topo IIinhibitors produce the same effect on chromosome con-densation constitutes a strong argument for the effect tobe mediated by topo II inhibition rather than by an unrec-ognized target. Thus, it is now well established that topo IIactivity is required for full chromosome condensation inhigher eukaryotes.

According to our current understanding of chromatin andchromosome structure (reviewed in Refs. 41 and 42), thecondensation of interphase chromatin into mitotic chromo-

somes involves structural transitions at two levels of packing.lnterphase chromatin is composed of regions of condensedheterochromatin in the form of a 30-nm fiber and regions ofless condensed euchromatin composed of 10-nm fiber orloose 30-nm fiber. The i 0-nm fiber corresponds to the com-pletely decondensed “beads on a string” configuration,whereas the 30-nm fiber may be a coil of tightly packednucleosomes. A first step in chromosome condensation isprobably the transition of all i 0-nm fiber and loose 30-nmfiber into 30-nm fiber. A second transition is readily visible asnuclei progress through prophase and involves the shorten-

ing of the long chromatids into the short and thick chroma-tids of metaphase chromosomes, which most likely occursby coiling of the chromatids (43). Cells entering mitosis in thepresence of VP-i 6 show dots of DNA staining more similar inintensity to that of metaphase chromosomes than that ofinterphase chromatin. This is a strong indication that chro-

matin condensation can happen without topo II activity. Cellsforced to enter mitosis in the presence of inhibitors that donot stabilize the cleavable complex also show chromosomesthat are clearly condensed but which are defective in chro-matid coiling. These results indicate that topo II activity isprobably not required for formation of the 30-nm fiber butthat it is required for coiling of the chromatids in the last stepof chromosome condensation.

Daughter DNA duplexes are intertwined as a result of DNAreplication. It is likely that most of the intertwines remain untilthe time the cell enters mitosis because chromosomes ofcells entering mitosis in the presence of topo II inhibitors arenot only elongated but also tangled (Refs. 30 and 31 and thisstudy). It is not yet known whether topo II activity is requiredfor the coiling process or whether it is required only for sisterchromatid decatenation, the chromatids being unable to coilcompletely when intertwined.

Materials and MethodsCell Culture and SynchronizatIon. BHK 21 cells were grown as mono-layers in DMEM supplemented with 10% fetal bovine serum and antibi-

otics and maintained at 37”C in humidified 10% CO2. tsBN2 cells weregrown in the same way but maintained at 32#{176}C.tsBN2 cells (44) weresynchronized to G2 by growing in isoleucine-free medium for 24 h, fol-

lowed by 13 h in normal medium containing 2.5 �g/ml aphidicolin (Sigma

Chemical Co.), and then 5 h in normal medium.

Drug Treatments. Colcemid (Sigma) was from a 0.5-mg/mI stock in

ethanol at -20”C. VP-16 (Bristol-Meyers) was from a l000x stock inDMSO, merbarone (National Cancer Institute) from a stock in DMSO,ICRF-187 (Adria-SP, Inc.) from a stock in 0.2 M HCI, and aclarubicin

(Sigma) from a fresh stock in H2O.Fostriecin (NSC 339638, 94528)was obtained from the National Cancer

Institute as vials containing 25 mg fostriecin, 39 mg ascorbic acid (as an

antioxidant), and NaOH to neutralize to pH 7. It was used from a freshstock solution in H2O. Okadaic acid (GIBCO) was from a 0.5-mM stock in

DMSO.Fluorescence Microscopy. Cells were collected by trypsinization,

swelled in hypotonic medium (75 m�i KCI), fixed with methanol:acetic acid(3:1), spotted onto microscope slides, stained with bisbenzimide, and

observed using a Zeiss standard microscope as in Guo et a!. (45). Cells

were photographed on Kodak Tmax 400 film using a Zeiss Axiophot

microscope.

90 Topoisomerase II and Mitosis

AcknowledgmentsWe thank Tom Cavalier-Smith for use of his photomicroscope.

References1 . Anderson, H. J., and Roberge, M. DNAtopoisomerase II: a review of itsinvolvement in chromosome structure, DNA replication, transcription andmitosis. CoIl Biol. Int. Rep., 16: 717-724, 1992.

2. Roca, J. The mechanisms of DNA topoisomorases. Trends BiochemSci., 20: 156-160, 1995.

3. Gassor, S. M., Laroche, T., Falquet, J., Boy do laTour, J., and Laemmli,U. K. Metaphase chromosome structure: involvement of topoisomerase II.J. Mol. Biol., 188: 613-629, 1986.

4. Uomura, T., and Yanagida, M. Isolation of type I and II DNA topo-isomerase mutants from fission yeast: single and double mutants showdifferent phenotypes in cell growth and chromatin organization. EMBO J.,3: 1737-1744, 1984.

5. DiNardo, S., Voelkel, K., and Sternglanz, R. DNA topoisomerase IImutants of Saccharomyces cerevisiae: topoisomerase II is required forsegregation of daughter molecules at the termination of DNA replication.Proc. NatI. Acad. Sci. USA, 81: 2616-2620, 1984.

6. HoIm, C., Goto, T., Wang, J. C., and Botstein, D. DNA topoisomerasoII is required at the time of mitosis in yeast. Cell, 41: 553-563, 1985.7. Uemura, T., Ohkura, H., Adachi, Y., Morino, K., Shiozaki, K., andYanagida, M. DNA topoisomerase II is required for condensation andseparation of mitotic chromosomes in S. pombe. Cell, 50: 917-925, 1987.

8. HoIm, C. Coming undone: how to untangle a chromosome. Cell, 77:955-957, 1994.

9. Wood, E. R., and Eamshaw, W. C. Mitotic chromatin condensation invitm using somatic cell extracts and nuclei with variable levels of endo-genous topoisomorase II. J. Cell Biol., 1 1 1: 2839-2850, 1990.10. Adachi, Y., Luke, M., and Laommli, U. K. Chromosome assembly invitro: topoisomerase II is required for condensation. Cell, 64: 137-148,1991.

1 1 . Hirano, T., and Mitchison, T. J. Topoisomerase II does not play ascaffolding rob in the organization of mitotic chromosomes assembled inXenopus egg extracts. J. Cell Biol., 120: 601-612, 1993.

12. Vassetzky, Y. S., Dang, 0., Benedetti, P., and Gasser, S. M. Topo-isomerase II forms multirnors in vitro: effects of metals, (3-glycerophos-phato, and phosphorylation of its COOH-terminal domain. Mol. Cell. Biol.,14: 6962-6974, 1994.

1 3. IJu, L F. DNA topoisomorase poisons as antitumor drugs. Annu. Rev.Biochem., 58: 351-375, 1989.

14. Zhang, H., DArpa, P., and Liu, L F. A model for tumor cell killing bytopoisomorase poisons. Cancer Cells, 2: 23-27, 1990.

15. Robergo, M., Th’ng, J., Hamaguchi, J., and Bradbury, E. M. Thetopoisomerase II inhibitor VM-26 induces marked changes in histone Hikinase activity, histones Hi and H3 phosphorylation, and chromosomecondensation in G2 and mitotic BHK cells. J. Cell Biol., 1 11: 1753-1762,1990.

16. Chan’on, M., and Hancock, R. DNA topoisomoraso II is required forformation of mitotic chromosomes in Chinese hamster ovary cells: studiesusing the inhibitor 4’-demethylopipodophyllotoxin 9-(4,6-O-thenylidene-f3-o-glucopyranosido). Biochemistry, 29: 9531-9537, 1990.

17. Lock, R. B., and Ross, W. E. Inhibition of p34Cdc2 kinaso activity byetoposido or irradiation as a mechanism of G2 arrest in Chinese hamsterovary cells. Cancer Res., 50: 3761-3766, 1990.

1 8. Lock, R. B. Inhibition of p34cdc2 kinase activation, p34cdc2 tyrosinedophosphorylation, and mitotic progression in Chinese hamster ovarycells exposed to etoposide. Cancer Res., 52: 1817-1822, 1992.

19. Li, J. J., and Deshaies, R. J. Exercising self-restraint: discouragingillicit acts of S and M in eukaryotos. Cell, 74: 223-226, 1993.

20. Murray, A., and Hunt, T. The Cell Cycle: an Introduction, pp., 139-144. New York: W. H. Freeman and Company, 1993.

21 . Hartwell, L, Weinort, T., Kadyk, L, and Garvik, B. Cell cycle check-points, genomic integrity, and cancer. Cold Spring Harbor Symp. Quant.Biol., 59: 259-263, 1994.

22. Drake, F. H., Hofmann, G. A., Mong, S-M., O’Leary Bartus, J.,Hortzberg, R. P., Johnson, R. K., Mattorn, M. R., and Mirabelli, C. K. Invitro and intracellular inhibition of topoisomoraso II by the antitumor agentmerbarono. Cancer Res., 49: 2578-2583, 1989.

23. Chon, M., and Beck, W. T. Toniposide-resistant CEM cells, whichexpress mutant DNA topoisornerase Ila when treated with non-complex-stabilizing inhibitors of the enzyme, display no cross-resistance and revealaberrant functions of the mutant enzyme. Cancer Res., 53: 5946-5953,1993.

24. Jensen, P. B., Sorensen, B. S., Dornant, E. J. F., Sohested, M., Jensen,P. S., Vindolov, L, and Hansen, H. H. Antagonistic effect ofaclarubicin on thecytotoxicity of etoposide and 4’-9-acridinylarnino)methanesulfon-manisididein human small cell lung cancer cell lines and on topoisomerase Il-mediatedDNA cleavage. Cancer Res., 50: 331 1-3316, 1990.

25. Sohested, M., Jensen, P. B., Sorensen, B. S., HoIm, B., Friche, E., andDomant, E. J. F. Antagonistic effect of the cardioprotector (+)-1 ,2-bis(3,5-dioxopiperazmnyl-1-yI)propane (ICRF-i87) on DNA breaks and cytotoxicityinduced by the topoisomeraso II directed drugs daunorubicin and etopo-side (VP-i6). Biochem. Pharmacol., 46: 389-393, 1993.

26. Chen, M., and Beck, W. T. Differences in inhibition of chromosomeseparation and G2 arrest by DNA topoisomerase II inhibitors merbaroneand VM-26. Cancer Res., 55: 1509-1516, 1995.27. Sharpo, H. B. A., Field, E. 0., and Hellmann, K. Mode of action of thecytostatic agent “ICRF 159.” Nature (Lond.), 226: 524-526, 1970.

28. Ishida, R., Miki, T., Narita, T., Yui, R., Sato, M., Utsumi, K. R., Tanabe,K., and Andoh, T. Inhibition of intracellular topoisomerase II by antitumourbis(2,6-dioxopiperazine) derivatives: mode of coIl growth inhibition dis-tinct from that of the cleavable complex-forming type inhibitors. CancerRes., 51: 4909-4916, 1991.

29. Downes, C. S., Clarke, D. J., Mullinger, A. M., Gim#{233}nez-Abi#{225}n,J. F.,Creighton, A. M., and Johnson, R. T. A topoisomerase Il-dependent G2checkpoint in mammalian cells. Nature (Lond.), 372: 467-470, 1994.

30. Ishida, R., Sato, M., Narita, T., Utsumi, K. R., Nishimoto, T., Morita, T.,Nagata, H., and Andoh, T. Inhibition of DNA topoisomerase II by ICRF-i 93induces polyploidization by uncoupling chromosome dynamics from othercell cycle events. J. Cell Biol., 126: 1341-1351 , 1994.

31 . Gorbsky, G. J. Cell cycle progression and chromosome segregationin mammalian cells cultured in the presence of the topoisomorase IIinhibitors ICRF-i87 [(+)-1 ,2-bis(3,5-dioxopiperazmnyl-i-yI)propane; ADR-529] and ICRF-159 (Razoxane). Cancer Res., 54: 1042-1048, 1994.

32. Nishimoto, T. The “BN2,” a regulator for the onset of chromosomecondensation. BioEssays, 9: 121-124, 1988.

33. Robergo, M. Checkpoint controls that couple mitosis to completion ofDNA replication. Trends Cell Biol., 2: 277-281 , 1992.

34. Roberge, M., Tudan, C. T., Hung, S. M. F., Harder, K. W., Jirik, F. R.,and Anderson, H. J. Antitumor drug fostriecin inhibits the mitotic entrycheckpoint and protein phosphatasos 1 and 2A. Cancer Res., 54: 6115-6121, 1994.

35. Drake, F. H., Hofmann, G. A., Bartus, H. F., Mattem, M. R., Crooke, S. T.,and Mirabolli, C. K. Biochemical and pharmacological properties of p1 70 andp180 forms of topoisomoraso II. Biochemistry, 28: 8154-8160, 1989.

36. Roca, J., Ishida, R., Berger, J. M., Andoh, T., and Wang. J. C.Antitumor bisdioxopiperazines inhibit yeast DNA topoisomerase II by trap-ping the enzyme in the form of a closed clamp. Proc. NatI. Acad. Sci. USA,91: 1781-1785, 1994.

37. Sorensen, B. S., Smnding, J., Andersen, A. H., Alsner, J., Jensen, P. B.,and Wostergaard, 0. Mode of action of topoisomerase Il-targeting agentsat a specific DNA sequence: uncoupling the DNA binding, cleavage andreligation events. J. Mol. Biol., 228: 778-786, 1992.

38. Sorensen, B. S., Jensen, B. P., Sehested, M., Jensen, P. 5., Kjeldsen,E., Nielsen, 0. F., and Alsner, J. Antagonistic effect of aclarubicin oncamptothecin induced cytotoxicity: role of topoisomerase I. Biochem.Pharmacol., 47: 2105-2110, 1994.

39. Rao, A. P., and Rao, P. N. The cause of G2 arrest in Chinese hamster

ovary cells treated with anticancer drugs. J. Natl. Cancer Inst., 57: 1139-1143, 1976.

40. Sandell, L L, and Zakian, V. A. Loss of a yeast telomere: arrest,recovery, and chromosome loss. Cell, 75: 729-739, 1993.

41 . van Holde, K. E. Chrornatin, pp. 289-354. New York: Springer-Veriag,1989.42. Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K., and Watson, J. D.Molecular Biology of the Cell, pp. 346-356. New York: Garland Publish-ing, Inc., 1994.

43. Boy do Ia Tour, E., and Laemmli, U. K. The metaphase scaffold ishelically folded: sister chromatids have predominantly opposite helicalhandedness. Cell, 55: 937-944, 1988.

44. Ishida, R., Takahashi, T., and Nishimoto, T. Chromosomes of G2arrested cells are easily analyzed by use of the “tsBN2” mutation. CellStruct. Funct., 10: 417-420, 1985.

45. Guo, X. W., Th’ng, J. P. H., Swank, R. A., Anderson, H. J., Tudan, C.,Bradbury, E. M., and Roberge, M. Chromosome condensation induced byfostriecin does not require p34cdc2 kinase activity and histone Hi hyper-phosphorylation but is associated with enhanced histone H2A and H3phosphorylation. EMBO J., 14: 976-985, 1995.