[CANCER RESEARCH 50, 710-716, February 1, 1990|

Development of Polyploidization in Taxol-resistant Human Leukemia Cells inVitro1

John R. Roberts, David C. Allison,2 Ross C. Donehower, and Eric K. Rowinsky3

Departments of Surgery and Oncology, Johns Hopkins Hospital, Baltimore 21205 fj. R. R., E. K. R., R. C. D., D. C. A.), and Department of Surgery, Baltimore VAMedical Center ¡D.C. A.], Baltimore, Maryland 21218

ABSTRACT

The effects of taxol on antitubulin immunofluorescent staining patterns, cellular DNA content, and labeling with [3H|thymidine were measured for the taxol-sensitive III (ill and taxol-resistant K562 cell linesafter exposures for 0, 4, 12, and 24 h. Taxol caused a relative increasein the fraction of 4C interphase and metaphase cells in both lines althoughthe 4C interphase accumulation was greater for the resistant K562 line.Of the cells with S-phase DNA content, taxol-treated HL60 cells wereless likely to incorporate [3H|thymidine than taxol-treated K562 cells.However, a decrease in percentage of S-phase labeling for both linesrelative to control cells was seen. Finally, taxol induced the developmentof polyploid cells (cells with DNA contents greater than that of the 4C<;,-M peak) in the relatively taxol-resistant K562 cells, an effect not seenin the relatively taxol-sensitive 111.60line. After 24 h of taxol exposure70% of all K562 cells were polyploid while only 8% of the 111.61)cellswere polyploid. The capacity of K562 cells to generate polyploidy inresponse to taxol correlated with taxol resistance by previous assay andmay be a useful indicator of drug resistance.

INTRODUCTION

Taxol is a diterpenoid plant product that has demonstratedsignificant activity against metastatic melanoma (1) and otheradvanced solid tumors (2-6) in phase 1clinical trials. In contrastto other antimicrotubule chemotherapeutic agents, such as Col-cemid, vincristine, and vinblastine, that antagonize tubulin polymerization and induce the disassembly of microtubules intomonomers, taxol enhances tubulin monomer polymerizationand stabilizes tubulin polymers after they are formed, both invivo (7-11) and in vitro (12-20). Taxol also induced treatedcells to form microtubule "asters" or "bundles" when the cells

are viewed with antitubulin immunofluorescence (21). Workfrom this (22) and several other laboratories (23) have shownthat cells with microtubule asters are arrested in mitosis andthat cells with microtubule bundles are interphase cells.

The mechanisms of taxol cytotoxicity and resistance are notwell understood. Taxol treatment has been shown by flowcytometry to increase the proportion of cells with G2-M DNAcontent (24), and there has been one report of the developmentof polyploid cells in cultures after long-term taxol treatment(25). Nevertheless, as with the other antitubulin agents, clearevidence for a predominant mitotic or interphase cytotoxiceffect has yet to be provided. Recently, two of us showed thatcultured cells of the human promyelocytic leukemia K562 cellline and the human myelocytic leukemia HL60 cell line wereresistant and sensitive, respectively, to taxol in clonogenic assays (26). Surprisingly, cells from both the resistant K562 lineand the sensitive HL60 line formed microtubule asters andbundles with taxol treatment when viewed with antitubulin

Received 5/22/89; revised 10/17/89; accepted 10/30/89.The costs of publication of this article were defrayed in part by the payment

of page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1Supported by the Veterans Administration.2To whom requests for reprints should be addressed, at Department of

Surgery, Dispensary 681, Johns Hopkins Hospital, 600 N. Wolfe Street, Baltimore, MD 21205.

3Recipient of an American Cancer Society Career Development Award.

immunofluorescence (4). This finding demonstrates that thedifference in drug resistance between these two cell lines mustresult from intracellular differences in drug metabolism orresponse since enough drug enters cells of both lines to exert aconsiderable intracellular effect (aster and bundle formation).

We now report a study of the effects of taxol on the formationof asters and bundles, on cell cycle traverse, and on labeling ofS-phase cells with ['H]thymidine for the taxol-sensitive HL60

and the taxol-resistant K562 lines. Some interphase effects oftaxol were observed, including the finding that fewer S-phasecells of the HL60 line incorporated [3H]thymidine after taxoltreatment than did the S-phase cells of the K562 line. However,the most striking result of this study was the rapid developmentof polyploid cells in the taxol-resistant K.562 line, but not thetaxol-sensitive HL60 line, after taxol treatment.

MATERIALS AND METHODS

Computerized Microscopy. The microscopic system used in this studyhas been described previously in detail (27-30). Briefly, we used acomputerized stage controller system (Cell Tracker; Robertson Electronics, Albuquerque, NM), which controls Zeiss 0.5-^m scanningstages that were mounted onto a Zeiss fluorescent microscope and aVickers M85 absorption cytometer. The slide positions of cells werethen mapped and stored to allow sequential immunohistochemical(indirect antitubulin immunofluorescence) and histochemical (FeulgenDNA content and autoradiographic labeling) analyses to be performedon the same cells.

Cell Lines. HL60 (promyelocytic leukemia) and K562 (myelocyticleukemia) cells were maintained in RPMI 1640 with 10% fetal calfserum and 1% penicillin-streptomycin in 5% CC>2and 95% Oj at 37°Cin 10-mm2 Costar tissue culture flasks. For each replicate 5 x IO5cells/

ml were exposed to 10 JIM taxol for 0, 4, 12, and 24 h. Previouslyreported clonogenic survival studies in soft agar had demonstrated thatK562 and HL60 cells were relatively taxol resistant and sensitive,respectively (26), at this clinically attainable concentration (31). Studiesin our laboratory have demonstrated that the two cell lines haveapproximately equal doubling times under conditions of exponentialgrowth (18 h).

Labeling with |3H]Thymidine, Slide Preparation, and FluorescenceStaining. Cells were exposed to 0.25 /iCi/ml ['Hjthymidine (43 Ci/mmol; Amersham) for 30 min before harvest to label S-phase cells.Chicken erythrocytes were added as an internal standard for DNAcontent. The cells were then cytocentrifuged gently onto slides at 200rpm for 5 min, because we found that greater speeds damaged micro-tubules. The slides were immediately fixed in 1% paraformaldehyde at0°Cfor 20 min and were then placed in cold acetone (0°C)for 7 minto permeabilize the cell membranes. After a PBS4 wash, the cells wereincubated with a mouse monoclonal anti-tubulin antibody (New England Nuclear) at a 1:10 dilution for 2 h at 37°C,washed once in PBS,

and incubated with a 1:20 dilution of a goat anti-mouse immunoglob-ulin conjugated with fluorescein (Sigma, St. Louis) for 2 hours at 37°C.

After another PBS wash, the cell spots were covered with glycerolbuffered with Trizma base to pH 8.1, sealed under glass coverslips withnail polish, and then stored at 3°Cin the dark.

Fluorescence Microscopy. Following antitubulin immunofluorescence staining, microtubules were viewed with a Zeiss fluorescence

4 Abbreviations used are: PBS, phosphate-buffered saline; GAP, grain areaproportion (nuclear grain area divided by nuclear area).

710

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

Fig. 1. K562 cells after 12 h incubation with taxol. A, several cells with microtubule bundles and asters after staining of taxol-treated cells with indirectimmunofluorescence of tubulin; a and b, mark asters and bundles, respectively. B, several cells with microtubule asters after staining of taxol-treated cells with indirectimmunofluorescence for tubulin. b and n mark cells with microtubule bundles and with no configuration, respectively. All other cells possess microtubule asters.

microscope equipped with filters appropriate for fluorescein (Fig. \A).The positions of approximately 1000 cells in three replicate experiments from each cell line were mapped on coded slides for each timeinterval. We selected random fields to obtain unbiased cell cycle distributions. The cells were classified under fluorescence microscopy aseither asters bundles, or "null" cells, according to their microtubule

conformation, by an observer with no knowledge of the cell line or thetime period being studied. We chose a conservative threshold fordefinition of microtubule asters such that each cell with a microtubuleaster was found to have a metaphase nucleus.5 The mapped cell posi

tions and the corresponding morphological classifications were storedin computer memory so that the same cells could be relocated forsubsequent measurement of DNA content and [3H]thymidine labeling.

DNA Staining and Measurement. After antitubulin immunofluorescence microscopy, the cells were stained for measurement of cellularDNA content by the Feulgen reaction. Hydrolysis was performed in 4N HC1 for 60 min at 28°C,immediately followed by Feulgen staining

in Schiffs reagent (Sigma, St. Louis, MO) for 60 min at room temperature. The mapped cells were then relocated with the computerizedscanning system and cytophotometric measurements of cellular DNAcontent were made using a Vickers M85a microdensitometer (VickersInstruments, Maiden, MA). We used a xlOO achromatic lens with aOA-nm scanning spot. This instrument makes 50,000 point absorbancemeasurements over a slide area containing a single nucleus. Thesemeasurements are then integrated to calculate the total absorbancewhich is proportional to the Feulgen stain content per nucleus. Theabsorbances of all nuclei were normalized to the values obtained forchicken erythrocytes that were measured on the same slides (2.6 pgDNA/cell), as described previously (27-29). The cytophotometer measures nuclear area by counting the number of point absorbance measurements above a predetermined absorbance threshold. For Feulgen stainmeasurements, the absorbance threshold was set at 0.05 and the wavelength at 560 nm. The data shown are the pooled results of threereplicate experiments.

Autoradiography and Measurement of Autoradiographic Grain Areas.For autoradiography, the slides were immersed in NTB photographicemulsion (Kodak, Rochester, NY) and exposed for 14 days. The samecells which had previously been evaluated for microtubule morphologyand DNA content were relocated and the autoradiographic grain areasover each of these nuclei were then measured on the Vickers microdensitometer at a wavelength of 625 nm and an absorbance area thresholdof 0.2. Under these conditions the underlying Feulgen stain of thesenuclei is not detected (28). The autoradiographic grain area of each cell

' J. R. Roberts el al., unpublished findings.

was then divided by its nuclear area to correct for differences in nuclearsize (28). A cell was considered labeled if the grain area was greaterthan or equal to 4% of its nuclear area. With this threshold 97% of thecells with S-phase DNA contents were identified as labeled in control,untreated populations.

Statistics. Statistical evaluation was performed using unpaired, two-tailed Student's t test to compare the cell cycle proportions, fraction ofpolyploid cells, and percentage of labeled S-phase cells between treatedand untreated K562 and HL60 cells. Bonfrommi's rule for multiplecomparisons for Student's t test requires that the total (summed) P

value for all related comparisons must be «0.05in order for significanceto be attained; this rule was used where appropriate. Thus the P valueslisted in Table 2 involving comparison of proportions between cell lineswere each multiplied by 4 since, in each case, 4 related comparisonswere made.

RESULTS

Aster and Bundle Formation. Fig. 1 is a photomicrograph ofK562 cells that were treated with taxol at 10 /¿Mbefore harvestand before assay for microtubular morphology by antitubulinimmunofluorescence. The aster and bundle formations inducedby taxol are apparent. Fig. 2A shows the percentages of K562and HL60 cells showing either the aster or the bundle conformation as a function of time of taxol treatment. At time zero,no cells from either the K562 or the HL60 lines showed asteror bundle conformations. At 12 h, however, about 90% of cellsfrom each line possessed either microtubule asters or bundles(Fig. 2A). At 24 h significantly more of the resistant K562 cells(95%) were classified as either bundles or asters than thesensitive HL60 cells (82%; P < 0.05), although this effect wasrelatively small.

In these experiments we reconfirmed that cells from bothlines which formed bundles had DNA contents which rangedover all interphase (G0-Gi, S, and G2 phases) and that cellswhich formed asters were arrested mitotic figures. It was ofinterest, however, that 5,65, and 30% of the K562 asters presentafter 24 h of taxol treatment had 2C, 4C, or 8C DNA content,respectively (data not shown), suggesting that microtubule asters can persist (or reform) after karyokinesis.

Fig. IB shows the percentages of asters and bundles separately for both lines as a function of duration of taxol exposure.

711

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

100-r

) 510 15Time (hrs)2025A

90-]

75-

"601(tí

C«30-

K562 asters15-0

'7

/~—~~"~"''.¿•''

HL60asters*0

5 1015Time(hrs)•

' * '\20

25B

Fig. 2. Percentages of asters and bundles as a function of time. More than1000 cells were classified as asters, bundles, or null cells at each data point. A,change with time in the total percentage of cells expressing either of the micro-tubule forms; B, change with time in the percentage of each form in each cell line.In both cell lines there were more bundles than asters at all treatment times.

There were more bundles than asters (P < 0.001) for both celllines from 4 to 24 h of taxol treatment. The percentage of astersremained about 20% for both lines from 12 to 24 h (Fig. 2B).From 4 to 24 h, there was a slight decrease in the percentageof HL60 bundles. In contrast there was a gradual, but significant, increase in the percentage of K562 bundles during thesame period. These data cannot be compared to the previousstudy (26) because of the differences in slide preparation.

Labeling with [3H]Thymidine during Taxol Exposure. TheHL60 and K562 cell lines both have G0-G, DNA contents ofapproximately 11.5 pg of DNA (23). Therefore, for both ofthese lines, cells with DNA contents between 13.5 and 20 pgare away from the G0-Gi and G2-M peaks and should be in Sphase. That this is true is shown in Figs. 3A and 4/Õ,where itcan be seen that virtually all of the cells from both lines withS-phase DNA contents (between 13.5 and 20 pg) show [3H]-

thymidine labeling prior to taxol treatment and that these cellswere positioned between the nonlabeled G0-G| and G2-M peaks.After 24 h of taxol treatment we found polyploid K562 S-phasecells (23.5-40 pg of DNA) were present which also incorporated[3H]thymidine (Fig. 4D). Thus, at 24 h of taxol treatment thelabeling of the HL60 S-phase cells (13-20 pg) could be compared with two groups of K562 S-phase cells (13.5-20 and 23-

40 pg).The percentages of S-phase cells that incorporated [3H]thy-

midine decreased for both cell lines with increasing time oftaxol exposure (Table 1). This decrement was especially pronounced in the taxol-sensitive (HL60) line although the absolute fraction of K562 cells with S-phase DNA content wasreduced to a greater degree than the HL60 cells. Specifically,

the percentage of labeled S-phase cells was significantly greaterat 4, 12, and 24 h for K562 cells than for HL60 cells (Table 1).By 24 h, only 11.6% of HL60 cells with S-phase DNA contentswere labeled, a significantly smaller proportion than either the36.7% of cells from the first (13.5-20.5 pg) or the 52.1% ofcells from the second (23.5-40 pg) S-phase DNA contentintervals of the K562 cells (P < 0.02; Table 1). Thus, for alltime intervals of taxol treatment tested, the proportions of S-phase cells in the taxol-resistant K562 line that incorporated[3H]thymidine were higher than those for the taxol-sensitive

HL60 line (Table 1).We also compared the intensity of [3H]thymidine labeling for

those S-phase cells of both lines that had incorporated isotopeafter various times of taxol treatment. To do this, we calculatedthe average GAP value, or the average proportion of the areaof each labeled nucleus covered with autoradiographic grains,for each set of labeled S-phase cells at the different timeintervals. We determined that the GAP value is a valid indexof labeling intensity in these experiments because we found nosignificant difference between the nuclear areas for the S-phasecells of both lines at the different times of taxol exposure andbecause the cell cycle times and DNA contents of the stem lineswere similar (data not shown). We found that the GAP valuesof the labeled S-phase cells in both the K562 and HL60 linesdecreased with increasing time of taxol exposure (Table 1).Thus, although a higher percentage of resistant cells was labeledcompared to sensitive cells, the decreased labeling intensity forboth lines seems to be a nonspecific effect, in that the taxol-sensitive and the taxol-resistant lines are affected to a similardegree.

DNA Distributions of K562 and HL60 Cells after TaxolTreatment. The DNA distributions and the incorporation of[3H]thymidine of K.562 and HL60 cells after treatment with 10

UM taxol for 0, 4, 12, and 24 h are given in Figs. 3 and 4. TheDNA distributions are plotted as bar graphs, with the proportion of labeled cells in each 0.5 interval of DNA contentrepresented as a darkened bar. At 0 h, both HL60 and K562cells had unlabeled G0-G, and G2-M peaks (Figs. 3A and 4A)with almost all of the S-phase cells between these peaks incorporating [3H]thymidine (Table 1). With treatment, by 12 h

(Figs. 3C and 4C), cells from both lines had accumulated inG2-M at the expense of cells in G|-G0. This effect was morepronounced in the K562 cell line which demonstrated a markeddecrease (from 59 to 7%) in the percentage of cells in G0-Giand marked increase (from 10 to 55%) in the percentage ofcells in G2-M after 12 h of exposure to 10 fiM taxol (Table 2;Fig. 4C). Similar changes, but of lesser magnitude, were evidentfor HL60 cells, in which the proportion of cells in Gi-G0decreased from 61 to 33% and the proportion of cells in G2-Mincreased from 8 to 21% (Table 2; Fig. 3Ä).From 0 to 12 h,the proportion of K562 cells with polyploid DNA contents(DNA contents greater than G2-M) increased significantly from3 to 19% (P < 0.04; Table 2; Fig. 4C). There was no similarincrease in the proportion of polyploid cells in the HL60 cellline during the same period of exposure (4% versus 9%; Table2; Fig. 3C). By 24 h approximately 67% of the resistant K562cells had polyploid DNA contents compared to only 7% of thesensitive HL60 cells (P < 0.00004; Figs. 3D and 4D\ Table 2).

Fig. 5 is a plot of the percentage of polyploid cells for boththe sensitive HL60 and the resistant K562 cells with increasingduration of taxol exposure. At 12 h the difference in percentagesof K562 and HL60 cells with polyploid DNA content is alreadysignificant (P < 0.04) and becomes much greater at 24 h (P <0.00004).

712

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

30-

0>O

2CH

IO

<n

O

2.o

C0>o

I

20-

10

IO 30 30 40

DNA Content (pg/cell)

IO 20 30 40DNA Content (pg/cells) B

30

<-> 20

o,0

C0>uO)o.

01oo.o

0>o

20

20 30 40DNA Content (pg/cell)

20 30 40DNA Content (pg/cell)

Fig. 3. DNA distributions of HL60 human leukemia cells after treatment with taxol and incubation with [3H|thymidine. •¿�.labeled cells. A, time zero; B, 4; C, 12

h; Ö,24 h.

G2 to M to Go-Gi Transition. The etiology of the polyploidK562 cells is of some interest. These cells may be the result ofactivation of "quiescent" K562 cells that have 4C G0-G, DNA

content, or of a true transition of the diploid K562 stemline(2C-4C) into a tetraploid stemline (4C-8C). Although notconclusive, the data in this study strongly support the lattermechanism.

The numbers of K562 cells in the taxol-treated cultures stayedthe same, or increased slightly, from the 4-24 h of drug exposure. Given that 66% of the K562 cells were polyploid at 24 h,it is very unlikely that so many cells could be derived from arelatively small (if it exists) population of quiescent G0-G| K562cells with 4C DNA content.

Mitotic and <•¿�•¿�Interphase Block. Fig. 6A depicts the increasein percentage of cells with 4C or 8C DNA content (and meta-phase nuclear morphology) as a function of duration of taxolexposure for each line. The percentage of cells in 4C metaphaseincreased from 0 to 12 h of taxol exposure for both linesalthough to a greater degree for cells of the K562 line. From12 to 24 h the percentage of K562 cells with 4C DNA contentand metaphase morphology decreased while the percentage ofK562 cells with 8C DNA content and metaphase morphologyincreased. [The total percentage of cells in metaphase witheither 4C or 8C DNA content remained the same from 4through 24 h (Fig. 25).]

Fig. 6B illustrates the change with time of taxol exposure inthe percentages of interphase cells with 4C and 8C DNAcontents. In brief, these results demonstrate that the percentageof K562 cells with 4C DNA content and interphase morphology(as determined with Feulgen and Giemsa stain) increased dramatically after taxol treatment. From 0 to 12 h there was asteady increase in the proportion of interphase cells with 4CDNA content for the K562 line. Between 12 and 24 h theproportion of interphase K562 cells with 4C DNA contentdecreased as polyploidization developed and there was a corresponding increase in the proportion of interphase cells with8C DNA content, suggesting that taxol caused interphase K562cells from the tetraploid stemline to accumulate at 8C DNAcontent. Inspection of the K562 and HL60 DNA distributions(Figs. 3 and 4) reveal that most if not all of these 4C cellspresent at 12 h are derived from the diploid stemline. It therefore seems reasonable to conclude that these cells are theprogenitors of the polyploid cells in the K562 line and that adiploid to tetraploid stemline shift is occurring. Further, taxolcauses cells from these lines to accumulate both in G2 and inM (Figs. 2 and 6).

DISCUSSION

Although many studies have shown that the toxicity of theVinca alkaloids and other tubulin poisons correlates with the

713

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

Fig. 4. DNA distributions of K562 humanleukemia cells after treatment with taxol andincubation with (3H]thymidine. •¿�.labeled cells.

A, time zero: B, 4 h; C 12 h; D, 24 h.

r 20

K) 20 30 40

DNA Content (pg/cell)20 50

DNA Content (pq/cell B

20 30 40DNA Content (pg/cell)

20 30 40

DNA Content (pg/cell)

Table I Percentage and intensity ofS-phase labeling after taxol treatmentThe values given for "percentage of labeled S-phase cells" (A) denote percentages of S-phase cells the grain areas of which were »4%their nuclear areas. S-phase

cells were defined to have between 13.5 and 20.5 pg DNA/cell. P values are the comparison between cell lines. The difference in percentage of labeled S-phase cellsbetween the first (2C-4C) and second (4C-8C) K562 S-phase intervals at 24 h was not significant. B. Intensity of S phase labeling after taxol treatment. The "averagegap values" given are the averages of (grain area/nuclear area x 10) for labeled cells (gap value >0.4). The two S-phase intervals at 24 h in K562 cells arc depicted as2C-4C and 4C-8C as in text.

A.% of labeled S-phase cells B. Av. GAP valuesTime(h)

HL60 K562 P HL60K562098.2 ±2.2° 98.4 ±1.5 NS* 9.0 ±0.2 11.0 ±0.6

4 76.9 ±8.4 93.2 + 0.3 <0.05 7.4 ±0.2 9.3+1.012 50.5 + 2.0 74.9 + 8.5 <0.01 8.8 ±0.3 8.8 ±0.9

2C-4C 4C-8C 2C-4C 4C-8C 2C-4C4C-8C24

11.6 ±5.1 36.7 ±10.3 52.2 ±16.0 <0.02 <0.02 6.0 ±0.5 7.0 ±1.6 9.5 ±1.9P<0.01

<0.05NS

2C-4C4C-8CNS

NS"Mean ±SD.

* NS, not significant.

Table 2 Celi cycle percentages after taxol treatment"P 24/0" is the P value for comparison between 0 and 24 h within a cell line.

"P KH/12" and "P KH/24" are the P values for comparison between HL60 and

K562 cell lines at 12 and 24 h. respectively.

75 y

¿60

CelllineHL60K562•

Mean ±Time

(h)041224/>24/0041224P24/0PK.H/4PKH/12P

KH/24SD.%G,60.9

±3.8°54.2

±3.333.4±2.632.2±4.70<0.00458.7

±5.028.6±2.26.89±2.562.00±2.00<0.00005<0.0004<0.0004<0.0004%

S25.2

±1.617.8+6.335.9±3.618.1

±5.8<0.1028.3

±1.233.7±1.715.7

±2.564.0±3.0<0.005<0.02<0.04<0.08%

G2-M8.4

±0.6416.2±6.2021.3±2.142.5±5.2<0.00210.2

±10.030.6+0.6455.0+

10.229.2+4.70<0.005<0.02<0.04<0.08%

>G2-M4.3

±107.9±7.99.4+0.27.33±2.36NS*2.8

±1.56.4±11.019.2±10.066.8±2.47<0.00005<1.0(NS)<0.04<0.00004*

NS, not significant.

effects of these agents on the mitotic spindle (32, 33), anincreasing body of evidence suggests that some of the cytotox-icity of these drugs is due to effects on cells in interphase (34,35). Studies with vincristine and vinblastine have demonstratedthat HeLa cells treated in S phase ultimately died in arrested

10 15Time (hrs)

Fig. 5. Change in the percentage of polyploid cells with duration of taxolexposure, plotted as the mean for three replicate samples ±SD (bars). Thedifference between the resistant (K562) and sensitive (HL60) cells was barelysignificant at 12 h but highly significant at 24 h.

mitosis (35). If these same cells were treated in d, however,they died in S phase. Further, several authors have found thatboth vincristine and vinblastine block cells in mitosis at muchlower drug concentrations than are required for cell killing (34,35) suggesting that the cytotoxicity of these drugs may resultfrom effects other than mitotic block.

714

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

25 T

10 15

Time (hrs)

60 n

10 15Time (Hrs)

Fig. 6. A, change, with duration of taxol exposure, in the percentage of cellswith 4C and 8C DNA content and metaphase morphology on Feulgen staining.4C and 8C denote the different chromosome complements for each cell line. B,change, with duration of taxol exposure, in the percentage of cells with 4C and8C DNA content and interphase nuclear morphology on Feulgen staining.

We found in this study that taxol blocks cells in mitosis (Fig.2) as well as in the G2 phase of the cell cycle. Nevertheless,since taxol blocks progression to G2 and M for both the sensitiveHL60 and the resistant K562 lines, this effect is unlikely to bethe explanation for the difference in taxol cytotoxicity, unlessthe degree of block were greater for sensitive cells. In fact, agreater block for resistant cells would be one possible interpretation of our data.

Taxol treatment decreased the intensity of S-phase labelingwith ['H]thymidine for those HL60 and K562 cells that incor

porated sufficient isotope to be labeled. Since this effect wasseen for both cell lines, it is most likely also not responsible forthe differential sensitivity of these two lines. Besides, changesin [3H]thymidine labeling intensity do not necessarily reflectchanges in the rates of DNA synthesis in S-phase cells. Unla-beled or lightly labeled S-phase cells can fail to incorporatetritiated thymidine due to (a) shifts in intracellular nucleotidepools, (b) synthesis of DNA through pathways not requiringthe isotope, or (c) slowing or stopping of DNA synthesis duringS phase due to drug action.

The decreased proportions of labeled S-phase cells were,however, highly significant when the taxol-sensitive HL60 linewas compared to the taxol-resistant K562 line at all of thetimes tested (Table 1A). As described above for differences inintensity of DNA labeling, it is possible that the unlabeled S-phase cells have slowed or stopped DNA synthesis and that theHL60 line is more sensitive to this interphase effect of taxolthan the K562 line. Alternatively, unlabeled S-phase cells mayfail to incorporate tritiated thymidine due to shifts in intracel

lular nucleotide pools or to synthesis of DNA through metabolicpathways not requiring thymidine. Such a slowing of DNAsynthesis in the HL60 S-phase cells may explain why the K562cells show a greater decrement of G0-G, cells and greaterincrement of G2 and M cells with increasing duration of taxolexposure than did the taxol-sensitive HL60 cells (P < 0.005;

Table 2). Specifically, this is the effect that one would expectfor two cell lines blocked at G2 and M but with different ratesof DNA synthesis.

Alternatively, the "unlabeled S-phase cells" may, in reality,

not be S-phase cells at all, but rather products of taxol-induced

errors in mitotic segregation. The daughter cells of unequalmitotic separations that receive the greater amount of DNAwould "masquerade" as unlabeled cells in the S-phase com

partment. Such cells might be expected to occur with a higherfrequency in the taxol-sensitive HL60 line. We are currentlyconducting experiments to determine whether interphase tox-icity (slowing of DNA synthesis) or errors in mitotic segregationare responsible for the unlabeled cells with S-phase DNA content which are observed after taxol treatment.

The most striking finding of this study was the developmentof polyploidization in the taxol-resistant K562, but not in thetaxol-sensitive HL60 cell line. Fig. 5 depicts the marked differ

ence in the development of polyploid cells between K562 andHL60 cells treated with taxol. By 12 h the difference in fractionsof polyploid cells between sensitive (HL60) and resistant (K562)cells was already significant, and the difference had increasedby 24 h.

Polyploidization has been demonstrated to occur with vin-

cristine (34, 36, 37, 38), vinblastine (38), and maytansine (36)in various cell lines in culture, as well as by Colcemid in larvaeof Drosophila melanogaster (39). To our knowledge, however,there has been no previous demonstration that, after short-termtreatment with an antitubulin drug, the development of polyploid cells in human or mammalian cancer lines correlates withthe resistance of those lines to that drug in clonogenic assays.

The etiology of the polyploid K562 cells is of some interest.These cells result from activation of "quiescent" K562 cells that

have 4C G0-Gi DNA content or from a true transition of the

diploid K562 stemline (2C to 4C) into a tetraploid stemline(4C to 8C). Although not conclusive, the data in this studystrongly support the latter mechanism. The numbers of K562cells in the taxol-treated cultures stayed the same, or increasedslightly, from the 4-24 h of drug exposure. Given that 66% ofthe K562 cells were polyploid at 24 h, it is very unlikely that somany cells could be derived from a relatively small populationof quiescent Go-Gi K562 cells with 4C DNA content. In otherwords, this postulated small population of cells could notnumerically "overwhelm" the diploid cells, because there was

no evidence for extensive cell death in the diploid cell population.

The present data suggest that the polyploidization was dueto a shift of all, or most, of the K562 cells in the cultures froma diploid (2C-4C) to a tetraploid level (4C-8C; Figs. 3 and 6).The data do not allow discrimination of whether this shift atthe 4C DNA level involves a reversible transition throughmitosis (without division) or whether the cells go directly fromthe G2 phase of the diploid level of the G, phase of the tetraploidlevel. Regardless of mechanisms, if this is revealed to be ageneralized phenomenon, it is possible that the resistance of acancer to antitubulin agents may be predicted by the development of polyploid cells after treatment.

715

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

TAXOL-INDUCED POLYPLOIDIZATION

REFERENCES 22

1. Wiernik, P. H., Schwartz, E. L., Einzig, A., Strauman, J. J., Lipton, R. B.,and Duteher, J. P. Phase I trials of taxol given as a 24 hour infusion every 23.21 days: responses observed in metastatic melanoma. J. Clin. Oncol., 5:1232-1239, 1987.

2. Grem, J. L., Tutsch, K. D., Simon, K. J., Alberti, D. B., Willson, J. K.,Tormey, D. C., Swaminathan, S., and Trump, D. L. Phase I study of taxol 24.administered as a short iv infusion daily for 5 days. Cancer Treat. Rep., 71:1179-1184,1987. 25.

3. Donehower, R. C., Rowinsky, E. K., Grochow, L. B., Longnecker, S. M.,and Ettinger, D. S. Phase I trial of taxol in patients with advanced cancer. 26Cancer Treat Rep., 71: 1171-1177, 1987.

4. Wiernik, P. H., Schwartz, E. L., Strauman, J. J., Duteher, J. P., Lipton, R.B., and Paietta, E. Phase I clinical and pharmacokinetic study of taxol. 27Cancer Res., 47: 2486-2493, 1987.

5. Kris, M. G., O'Connell, J. P., Gralla, R. J., Wertheim, M. S., Parente. R.

M., Schiff, P. B., and Young, C. W. Phase I trial of taxol given as a 3 hour 28.infusion every 21 days. Cancer Treat. Rep., 70: 605-607, 1986.

6. Legha, S. S., Tenney, D. M., and Krakoff, I. R. Phase I study of taxol usinga 5-day intermittent schedule. J. Clin. Oncol., 4: 762-766, 1986.

7. Manfredi, J. J., and Horwitz, S. B. Taxol: An antimitotic agent with a new 29mechanism of action. Pharmacol. Ther., 25: 83-105, 1984.

8. Schiff, P. B., and Horwitz, S. B. Taxol stabilizes microtubules in mousefibroblast cells. Proc. Nati. Acad. Sci. USA, 77: 1561-1565, 1980.

9. Manfredi, J. J., Parness, J., and Horwitz, S. B. Taxol binds to cellular 30microtubules. J. Cell. Biol., 94: 688-696, 1982.

10. Bird, M. M. The effects of taxol on embryonic chick tectum maintained inculture: an electron microscope study. J. Ultrastruct. Res., 89: 123-135, 311984.

11. Wallin, M., Nordh, J., and Deinum, J. I. Changes in the hydrodynamicproperties of microtubules induced by taxol. Biochim. Biophys. Acta, 880:189-196, 1986. 32.

12. Black, M. M. Taxol interferes with the interaction of microtubule-associatedproteins with microtubules in cultured neurons. J. Neurosci., 7: 3695-3702,1987. 33.

13. Schiff, P. B.,Tant, J., and Horwitz, S. B. Promotion of microtubule assemblyin vitro by taxol. Nature (Lond.), 22: 665-667, 1979.

14. Hamel, E., del Campo, A. A., Lowe, M. C., and Lin, C. M. Interactions of 34taxol, microtubule associated proteins and guanine nucleotides in tubulinpolymerization. J. Biol. Chem., 256; 11887-11894, 1981.

15. Wani, M. C, Tayla, H. L., Wall, M. E., Coappan, P., and McPhail, A. T.Plant antitumor agents VI. The isolation and structure of taxol, a novel 35.antileukemic and antitumor agent from Taxis brevifolia. J. Am. Chem. Soc.,93:2325-2327, 1971.

16. Dustier, P. Microtubules. Sei. Am., 243: 66-76, 1980. 36.17. Turner, P. F., and Margolis, R. L. Taxol-induced bundling of brain-derived

microtubules. J. Cell. Biol., 99: 940-946, 1984.18. Collins, C. A., and Vallee, R. B. Temperature-dependent reversible assembly 37.

of taxol-treated microtubules. J. Cell. Biol., 105: 2847-2854, 1987.19. Wilson, L., Miller, H. P., Farrell, K. W., Snyder, K. B., Thompson, W. C.,

and Purick, D. L. Taxol stabilization of microtubules in vitro: dynamics oftubulin addition and loss at opposite microtubule ends. Biochemistry, 24: 38.5254-5262, 1985.

20. Vanier, M. T., and Launay, J. F. In vitro assembly of rat pancreas tubulin inthe presence of taxol. Biochim. Biophys. Acta, 871: 72-77, 1986.

21. Manfredi, J. J., and Horwitz, S. B. Taxol: an antimitotic agent with a new 39.mechanism of action. Pharmacol. Ther., 25: 83-105, 1984.

Roberts, J. R. Demonstration of the cell cycle positions of taxol-induced(asters and bundles) by sequential measurements of tubulin ¡mniunofluores-cence, DNA content, and autoradiographic labeling of taxol-sensitive and-resistant cells. J. Histochem. Cytochem., 37: 1659-1665, 1989.De Brabander, M., Geuens, G., Nuydens, R., Willebrords, R., and De Mey,J. Taxol induces the assembly of free microtubules in living cells and blocksthe organizing capacity of the centrosomes and kinetochores. Proc. Nati.Acad. Sci. USA. 78: 5608-5612, 1981.Schiff, P. B., and Horwitz, S. Taxol stabilizes microtubules in mouse fibroblast cells. Proc. Nati. Acad. Sci. USA, 77: 1561-1565, 1980.Koponen, M., Greider A., and Loor, F. The effects of cyclosporins on thecell cycle of T-lymphoid cell lines. Exp. Cell. Res., 140: 237-250, 1982.Rowinsky, E. K., Donehower, R. C., Jones, R. J., and Tucker, R. W.Microtubule changes and cytotoxicity in leukemia cell lines treated withtaxol. Cancer Res.. 48: 4093-4100, 1988.Allison, D. C., Ridolpho, P. F., Rasch, E. M., Rasch, R. W., and Johnson,T. S. Increased accuracy of absorption cytophotornetric DNA values bycontrol of stain intensity. J. Histochem. Cytochem., 29: 1219-1228, 1981.Allison, D. C., Meyne, J., Ridolpho, P. F., Böse,K., Chakerian, M., andRobertson, J. Computerized measurement of the DNA content, areas, andautoradiographic grains of the same nuclei. J. Histochem. Cytochem., 32:1197-1203,1984.Allison, D. C, Lawrence, G. N., Ridolpho. P. F., O'Grady, B. J., Rasch, R.

W., and Rasch. E. M. Increased accuracy and speed of absorption cytometricDNA measurement by automatic corrections for nuclear darkness. Cytome-try, 5:217-227, 1984.Allison, D. C., and Ridolpho, P. Use of a trypan blue assay to measure thedeoxyribonucleic acid content and radioactive labeling of viable cells. J.Histochem. Cytochem., 28: 700-703, 1980.Longnecker, S. M., Donehower, R. C.. Cates, A. E., Chen, T. L., Brundrett,R. B., Grochow, L. B., Ettinger, D. S., and Colvin. M. High-performanceliquid Chromatographie assay for taxol in human plasma and urine andpharmacokinetics in a phase I trial. Cancer Treat. Rep., 71: 53-59, 1987.Tucker, R. W., Owellen, R. J., and Harris, S. B. Correlation of cytotoxicityand mitotic spindle dissolution by vinblastine in mammalian cells. CancerRes., 37: 4346-4351, 1977.Krishan, A. Time-lapse and ultrastructure studies on the reversal of mitoticarrest induced by vinblastine sulfate in Earle's L cells. J. Nati. Cancer Inst.,41: 581-595, 1968.Mujagic, H., Conger, B. M., Smith, C. A., Occhipinti, S. J., Schuette, W.H., and Schackney, S. E. Schedule dependent of vincristine lethality insarcoma 180 cells following partial synchronization with hydroxyurea. Cancer Res., 43: 3598-3603, 1983.Madoc-Jones, H., and Mauro. F. Interphase action of vinblastine and vincristine: differences in their lethal action through the mitotic cycle of culturedmammalian cells. J. Cell. Physio]., 72: 185-196, 1968.Alabaster, O., and Cassidy, M. Flow microflurometric analysis of P388murine leukemia after administration of vincristine and maytansine in vivo.J. Nati. Cancer Inst., 60: 649-652, 1978.Mujagic. H., Chen, S., Geist, R., Occhipinti, S., Conger, B. M., Smith, C.A., Schuette, W. H., and Shackney, S. E. Effects of vincristine on cellsurvival, cell cycle progression, and mitotic accumulation in asynchronouslygrowing sarcoma 180 cells. Cancer Res., 43: 3591-3597, 1983.McGown, A. T., Poppiti, D. G., Swindell, R., and Fox, B. W. The effect ofVinca alkaloids in enhancing the sensitivity of a methotrexate-resistant(L1210/R7A) line, studied by flow cytometric and chromosome numberanalysis. Cancer Chemother. Pharmacol., 73:47-53, 1984.Held, L. I. Polyploidy and aneuploidy induced by Colcemid in Drosophilamelanogaster. Mutât.Res., 94: 87-101, 1982.

716

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from

1990;50:710-716. Cancer Res   John R. Roberts, David C. Allison, Ross C. Donehower, et al.  

in VitroLeukemia Cells Development of Polyploidization in Taxol-resistant Human

  Updated version

  http://cancerres.aacrjournals.org/content/50/3/710

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/50/3/710To request permission to re-use all or part of this article, use this link

Research. on August 27, 2018. © 1990 American Association for Cancercancerres.aacrjournals.org Downloaded from