the effect of the cyclin-dependent kinase inhibitor olomoucine on cell cycle kinetics

EXPERIMENTAL CELL RESEARCH 236, 4–15 (1997)ARTICLE NO. EX973700

The Effect of the Cyclin-Dependent Kinase Inhibitor Olomoucineon Cell Cycle Kinetics

Bert Schutte,*,1 Luc Nieland,* Manon van Engeland,* Mieke E. R. Henfling,*Laurent Meijer,† and Frans C. S. Ramaekers*

*Department of Molecular Cell Biology and Genetics, University of Maastricht, The Netherlands;and †Station Biologique, CNRS, Roscoff, France

homologous members [3–6]. These CDKs are con-The effect of the cyclin-dependent (CDK) inhibitors stantly expressed during the cell cycle, but their activ-

olomoucine and roscovitine on cell kinetics was stud- ity is tightly regulated by complex formation withied. To this end, nonsmall cell lung cancer (NSCLC) cyclins, a class of proteins which also consists of severalcell line MR65 and neuroblastoma cell line CHP-212 highly homologous members [5–7]. In contrast towere pulse labeled with bromodeoxyuridine (BrdUrd) CDKs, cyclins are periodically expressed during the dif-and chased in culture medium, to which various con- ferent phases of the cell cycle [7, 8]. For instance, thecentrations of olomoucine or roscovitine were added. mitotic cyclin B is expressed very late in the S-phaseA dose-dependent inhibition of the G1/S-phase and G2/ and during G2- and part of the M-phase, while cyclinM-/G1 transitions was observed. Furthermore, S-phase E is expressed during late G1 and early S-phase of theprogression was also inhibited in a dose-dependent cell cycle [9–13].manner. Similarly, roscovitine, another CDK inhibitor The complex formation per se does not result in acti-with a 10-fold higher efficiency for both CDK1 and

vation of the kinase. The activity of each complex isCDK2 as compared to olomoucine, showed the sameregulated by phosphorylation and dephosphorylationeffects at a 10-fold lower concentration. At the highestof specific amino acid residues of the CDK/cyclin com-tested doses both olomoucine (200mM) and roscovitineplex. For instance the cyclin B/CDK1 complex is turned(40 mM) induced a complete cell cycle block in bothon by phosphorylation of thr161 by the cyclin activatingcell lines, paralleled by the appearance of apoptotickinase (CAK, CDK7/cyclin H) [14, 15] and turned offfigures. In these cultures a decrease in CDK1 proteinby phosphorylation of thr14 (myt1 kinase) and tyr15level was found as shown by Western blotting. Bivari-(wee1 and mik1 kinase), two residues located in theate CDK1/DNA analysis confirmed these observationsATP-binding domain of the CDK1 kinase [16–19]. Acti-and showed that a subpopulation of cells with charac-

teristics of apoptosis became CDK1 negative. The pre- vation occurs by dephosphorylation of the thr14 andsented data suggest that cyclins and CDKs are in- tyr15 residues by the CDC25 phosphatase [20]. Bothvolved at an important nodal point shared by path- wee1 kinase and CDC25 phosphatase activities are un-ways regulating cellular proliferation and apoptosis. der control of internal triggers and depend on comple-q 1997 Academic Press tion of DNA synthesis. In this way the proper timing

of activation of the CDK/cyclin complexes is ensured.The rapid activation of the CDK/cyclin complex isbrought about through a self-amplification loop, re-INTRODUCTIONsulting in phosphorylation and activation of theCDC25-C by CDK-cyclin B [20] and phosphorylationThe carefully ordered progression through the celland inactivation of weel. At the same time the activecycle is tightly controlled by the sequential formation,CDK/cyclin complex is involved in activating compo-activation, and inactivation of a series of cell cycle regu-nents of the cyclin-ubiquitin ligase pathway [21], re-lating kinases [1, 2]. Key players in the regulation ofsulting in its inactivation. This inactivation of thethe cell cycle are the cyclin-dependent kinases (CDKs),CDK/cyclin complexes is brought about by rapid proteo-a family of serine/threonine kinases with several highlylytic breakdown of the cyclins [22], mediated throughthe ubiquitin degradation pathway [21, 23, 24].

1 To whom correspondence and reprint requests should be ad- Several critical checkpoints exists in the cell cycle,dressed at Department of Molecular Cell Biology and Genetics, all governed by different combinations of CDKs andUniversity of Maastricht, P.O. Box 616, 6200 MD Maastricht, The

cyclins [7]. In this way both the target specificity andNetherlands. Fax: /31-43-3670948. E-mail: [email protected]. the timing of activation of these complexes is regulated

40014-4827/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

AID ECR 3700 / 6i26$$$181 09-09-97 09:33:29 eca

5THE EFFECT OF OLOMOUCINE ON CELL CYCLE KINETICS

[25]. For instance, evidence exists that the CDK1/cyclin and a similar efficiency toward erk1 and the other ki-nases tested) on the cell kinetics of two cell lines. ToB complex is responsible for phosphorylation of the

lamins during mitosis [26], which results in the dissoci- this end cells were pulse labeled with the thymidineanalogue bromodeoxyuridine (BrdUrd) and chased ination of the nuclear lamina. On the other hand, the

CDK4/cyclin D complex seems to govern the G1 transi- normal culture medium supplemented with variousdoses of inhibitor. Evidence was found for a dose-depen-tion by phosphorylating the RB gene product [27, 28].

The CDK2/cyclin E kinase plays an essential role in dent inhibition of cell cycle phase transition rates inboth cell lines. At the highest doses tested, both olo-G1/S transition [10, 29]. Physical association of cyclin

A with transcription factor E2F has been found, which moucine and roscovitine induced an almost completearrest of cells in their specific cell cycle phase, whichimplicates a role of CDK/cyclin complexes in regulation

of transcription [30, 31]. Furthermore, biochemical evi- ultimately resulted in apoptosis. No evidence wasfound for a cell cycle phase-specific induction ofdence exists for a role of the CDK2/cyclin A complex

during DNA replication, since this complex is able to apoptosis by olomoucine or roscovitine.phosphorylate a replication factor in vitro [32]. Fur-thermore, immunocytochemical analysis revealed that MATERIALS AND METHODScyclin A colocalizes with replication sites [33, 34].

Cell lines. The human nonsmall cell lung cancer (NSCLC) cellMany putative target proteins have been identifiedline MR65, kindly supplied by Dr. Gropp (Philips Universitatsfor cyclin-dependent kinase complexes in cell-free sys-Klinik, Marburg, Germany), was cultured in Eagle’s modified mini-

tems [1, 2, 35]. However, little information exists on the mal essential medium (GIBCO, Paisley, Scotland), supplementedin vivo targets of the CDK/cyclin complexes. Recently, with 1% nonessential amino acid solution (GIBCO), 1% glutamine

(Serva, Heidelberg, Germany; No. 22942), 1% HEPES (GIBCO), 10%intrinsic CDK inhibitors have been identified, such asheat-inactivated newborn calf serum (GIBCO, 021-6010M), and 0.1%p21 CIP1/WAF1 and others [36, 37]. The expression of,gentamycin (AUV, Cuyck, The Netherlands). The human neuro-for instance, p21 is under the control of the p53 protein blastoma cell line CHP-212 [42] was cultured in Dulbecco’s modified

and functions to arrest the cells at the G1-phase of the medium (GIBCO), supplemented with 1% glutamine (Serva), 10%cell cycle in response to DNA damage [38]. By influenc- heat-inactivated newborn calf serum, and 0.1% gentamycin (AUV).

Cells were harvested during exponential growth by trypsinization.ing the transcription of these intrinsic inhibitors itBrdUrd pulse/chase labeling. The cell lines were pulse labeledwould be possible to elucidate the role of these CDK/

with 10 mM BrdUrd (Serva, Heidelberg, Germany) for 30 min, rinsedcyclin complexes in vivo and identify the target pro-twice in prewarmed PBS, and chased in prewarmed culture medium,

teins that are affected by these complexes. An alterna- supplemented with 5 mM deoxythymidine.tive approach would be to use extrinsic inhibitors of CDK inhibitor treatment. Olomoucine and roscovitine were dis-these kinases but so far the availability of specific in- solved in DMSO and added to the cell cultures after BrdUrd pulse

labeling, at various concentrations i.e., 10, 50, 100, and 200 mM forhibitors of these complexes has been limited.olomoucine and 1, 5, 10, 20, and 40 mM for roscovitine. The finalRecently, it was shown that a compound isolatedDMSO concentration in the culture medium was 2%. Control culturesfrom Aspergillus terreus, name butyrolactone I, was received 2% DMSO. At various time intervals cells were harvested

able to inhibit several CDK/cyclin complexes in vitro by collecting both the floating as well as the adherent cells. Partand in vivo [39]. The authors showed that butyrolac- of the cell suspension was centrifuged onto glass slides, while the

remainder of the suspension was fixed in 70% ethanol at 47C andtone I was able to inhibit the phosphorylation of Rbused for bivariate BrdUrd/DNA flow cytometric analysis.both in vitro and in vivo. Recently, Vesely et al. [40]

Immunocytochemistry. Incorporated BrdUrd was detected as de-described a new purine derivative, olomoucine (2-(hy-scribed previously [43]. Briefly, approximately 106 ethanol-fixed cellsdroxyethylamino)-6-benzylamino-9-methylpurine), were rinsed once in PBS and resuspended in 2 ml of 0.4 mg/ml pepsin

which acts as a potent inhibitor of the CDK1/cyclin B in 0.1 N HCl (Serva, Heidelberg, Germany). After 30 min at roomtemperature cells were pelleted, resuspended in 2 N HCl, and incu-and related kinases by competing for the ATP bindingbated for another 30 min at 377C. Cells were rinsed in 0.1 M boratedomain of the kinase [40, 41]. Of 35 protein kinasesbuffer, pH 8.5, and PBS/BSA (1 mg/ml BSA in PBS). Appropriatelytested, olomoucine inhibited CDK2/cyclin A, CDK2/diluted anti-BrdUrd antibody (clone IIB5, available form Euro-Diag-

cyclin E, CDK1/cyclin B, CDK5/p35, and erk1. The re- nostica B. V., Arnhem, The Netherlands) was added to the cell pellet,lated kinases CDK4/cyclin D and CDK6/cyclin D3 were resuspended in 100 ml PBS/BSA. After incubation for 1 h at room

temperature, the cells were rinsed twice in PBS/BSA. For visualiza-not inhibited or were less sensitive to olomoucine. Thistion, FITC-conjugated Fab2 fragments of rabbit anti-mouse Ignew compound showed no effect on other kinases(DAKO A/S, Glostrup, Denmark) antibody were added in a 1:10 dilu-tested, such as protein kinases A or C, cyclic AMP- or tion. After incubation for 45 min at room temperature samples were

cyclic GMP-dependent kinases, and many others, nor rinsed twice in PBS/BSA and the cells were finally resuspended inon several enzymes that are directly involved in DNA- 0.5 ml cold PBS supplemented with 100 mg/ml RNAse (Serva) and

20 mg/ml propidium iodide (PI; Calbiochem, La Jolla, CA). The sam-synthesis, such as DNA toposiomerases and DNA poly-ples were allowed to stand for 15 min on ice in the dark before flowmerases at doses less than 1 mM.cytometric analysis. In the negative control the primary antibodyIn this study we investigated in detail the effect of was omitted.

olomoucine and roscovitine (an olomoucine analogue For bivariate CDK1/DNA analysis cells were fixed in methanol for5 min at 0207C. After rinsing the cells in PBS/BSA the resuspendedwith 10-fold higher efficiency toward CDK1 and CDK2

AID ECR 3700 / 6i26$$$182 09-09-97 09:33:29 eca

6 SCHUTTE ET AL.

pellet was incubated for 1 h at RT with anti-cdc2 (SC-54, Santa CruzBiotechnologies Inc., 1:20), followed by rinsing twice in PBS/BSA.Monoclonal antibody binding was visualized by incubating the cellswith FITC-conjugated Fab2 fragments of rabbit anti-mouse Ig(DAKO A/S, Glostrup, Denmark) antibody in a 1:10 dilution. Afterincubation for 45 min at room temperature samples were rinsedtwice in PBS/BSA and the cells were finally resuspended in 0.5 mlcold PBS supplemented with 100 mg/ml RNAse (Serva) and 20 mg/ml propidium iodide (PI; Calbiochem, La Jolla, CA). The sampleswere allowed to stand for 15 min on ice in the dark before flowcytometric analysis.

Cells, which were cytocentrifuged onto glass slides, were fixed for5 min in cold methanol at 0207C, rinsed in PBS, stained with PI (1mg/ml in PBS) for 10 min at room temperature, and embedded inglycerol/DABCO/PI (9 parts of glycerol, 1 part of 0.2 M HCl, pH8.0, 0.02% NaN3, and 2% DABCO (1,4-di-azobicyclo-(2,2,2)-octane,Merck, Darmstadt, Germany; no. 803456), pH 8.0) containing 0.5 mg/ml PI. The slides were covered with a coverslip and sealed with nailpolish.

Flow cytometry. For flow cytometric analysis a FACSort (Becton–Dickinson, Sunnyvale, CA) equipped with a single Argon ion laserwas used. Excitation was done at 488 nm, and the emission filtersused were 515–545 BP (green; FITC), 572–588 BP (orange; PE), and600 LP (red; PI). A minimum of 10,000 cells per sample were ana- FIG. 1. Cytogram of a bivariate BrdUrd/DNA analysis of MR65lyzed and data stored in list mode. FITC signals were recorded as cells, showing DNA content (ordinate) versus BrdUrd immunofluo-logarithmic amplified data, while the PI signals were recorded as rescence (abcissa). The cells are pulse labeled with BrdUrd andlinear amplified data. For bivariate FITC/PI analysis no compensa- chased for 4 h in culture medium. Regions are set to quantitate eithertion was used. the mean DNA content of BrdUrd-positive/undivided cells (R2) and

Data analysis was performed with the standard Lysis and Cellfit BrdUrd-negative G1 and S-phase cells (R6) or to quantitate the num-software (Becton–Dickinson). As a standard procedure for all analy- ber of BrdUrd-positive/divided cells (R3), BrdUrd-negative G2/M-ses, data were gated on pulse-processed PI signals to exclude dou- phase cells (R4), and BrdUrd-negative G1-phase cells (R5). Theseblets and larger aggregates. data are used to calculate the progression through various cell cycle

For detailed cell kinetic analysis, five cell cycle compartments were compartments as outlined under Materials and Methods.identified as described by Higashikubo et al. [44]. Briefly, the follow-ing populations were quantified (see Fig. 1): (1) BrdU-positive undi-vided cells (Nlu); (2) BrdU-positive divided cells (Nld); (3) BrdU-nega-tive G2 cells (G20); (4) BrdU-negative G1 cells (G10); and (5) BrdU-

and 568 nm. Emission spectra were separated by the standard setsnegative G1 and S cells (G1S0). From populations 1 and 5 the meanof dichroic mirrors and barrier filters. All scans were recorded inDNA content was calculated and normalized to G1- and G2/M-phasephoton counting mode.position (Xlu and XG1S, respectively). From these, four parameters

Photography. Phase contrast images of cell cultures were takenreflecting the modes of cell progression through the cell cycle wereusing an Axiovert 35 M microscope (Zeiss, Oberkochen, Germany),determined:equiped with a Nikon camera. For this purpose an Agfa pan 25X

(i) relative movement of the cells through the S-phase (RM, ac- black and white film was used.cording to Begg et al. [45]); i.e., the mean DNA content of BrdUrd- Gel electrophoresis and immunoblotting. Cell samples were re-positive undivided population normalized to G1 and G2/M peak posi- suspended in 50 mM Tris–HCl, pH 7.4, containing 17% glycerol andtion Å (Xlu 0 XG1)/(XG2/M 0 XG1); 2 mM ATP and sonicated for 5 s at 07C. Samples were centrifuged

(ii) fraction of undivided cells among the BrdUrd-positive popula- at 47C and the supernatant and pellet fraction were split. These weretion (F / undivided) Å Nlu/(Nlu / Nld/2); boiled for 5 min in SDS-sample buffer and loaded onto denaturing

(iii) BrdUrd-negative G2/M fraction (FG2M0) Å NG2/M/NT; 12% polyacrylamide gels [46]. The equivalent of approximately 106

(iv) the position of negative cells excluding the initial G2/M cells cells was used per lane. Gels were run on the Mini-Protean II systemnormalized to G1 peak position (XGI) Å (XG1S 0 XG1)/(XG2/M 0 XG1). (Bio-Rad Laboratories) for approximately 45 min at 200 V.

Immunoblotting was performed according to the method of TowbinData analysis. The means of the four parameters were plottedet al. [47], using PVDF membranes (Schleicher and Schuell, Dassel,as a function of time after BrdUrd pulse labeling for each dose ofGermany). The blots were incubated with anti-cdc2 (sc-54, Santaolomoucine and roscovitine tested. To quantitatively assess cell cycleCruz Biotechnologies Inc.), washed with PBS, and subsequently incu-progression, linear portions of the kinetic curves were fitted by linearbated with a peroxydase-conjugated rabbit anti-mouse Ig antibodyregression analysis. For the BrdUrd-negative G2/M fractions the lin-(DAKO A/S) to detect primary antibody binding. Peroxydase reactiv-ear regression analysis was performed on the logarithmically trans-ity was detected by enhanced chemiluminescence (ECL-kit, Amer-formed data. The slopes of the regression lines were used to elucidatesham, Buckinghamshire, UK).the changes in the rate of cell cycle progression.

Confocal scanning laser microscopy. Cells cytocentrifuged ontoglass slides were analyzed using the MRC600 confocal scanning laser RESULTSmicroscope (Bio-Rad, Hemel Hempstead, UK), equipped with an air-cooled Argon–Krypton mixed gas laser and mounted onto an Axio- In a first series of experiments in which the cell linesphot microscope (Zeiss, Oberkochen, Germany). The laser scan mi-

MR65 and CHP212 were exposed to various concentra-croscope was used in the dual parameter setup, according to themanufacturers specification, using dual wavelength excitation at 488 tions of the CDK inhibitor olomoucine and harvested

AID ECR 3700 / 6i26$$$182 09-09-97 09:33:29 eca


FIG. 2. Cytograms of bivariate BrdUrd/DNA analyses of MR65 cells pulsed with BrdUrd and chased for 10 h in medium supplementedwith 0 (A), 50 (B), 100 (C), and 200 mM olomoucine (D). DNA content is shown on the ordinate and BrdUrd immunofluorescence on theabscissa. A clear dose-dependent inhibition of cell cycle traverse of the BrdUrd-positive and BrdUrd-negative cells can be seen. At a doseof 200 mM olomoucine the cells are completely blocked in the cell cycle and the appearance of cells in the sub G1 region is noticeable.

at different time intervals, no obvious changes in cell partment. In the control cultures (Fig. 2A) cells whichwere originally in the G1-phase of the cell cycle at thecycle phase distributions were observed (data not

shown). The cultures seem not to be blocked in a spe- time of pulse labeling, have progressed through S-phase in contrast to the cultures exposed to olomoucinecific phase of the cell cycle by olomoucine. In order to

investigate the cell cycle effects of this CDK inhibitor (Figs. 2B–2D). At the highest dose of olomoucine, i.e.,200 mM, the cells seem to be frozen in the cell cycle.in more detail, the cell lines were pulse labeled with

BrdUrd and chased in the presence of various concen- Furthermore, both in the BrdUrd-positive and -nega-tive compartment, cells with a sub-G1 DNA contenttrations of the drug. Figure 2 shows a series of bivariate

BrdUrd/DNA cytograms of MR65 cells, chased for 10 seem to accumulate.For a more detailed analysis the number of cells inh in the presence of 0, 50, 100, and 200 mM olomoucine.

In contrast to the control culture (Fig. 2A), in which the various regions (R2–R6) of the scattergram werequantitated as outlined under Materials and Methods.the BrdUrd-labeled cells have progressed to G2/M and

G1 phase of the cell cycle, cells exposed to olomoucine Figure 3A shows the number of BrdUrd-negative G2/M-phase cells (R5) as a function of time for the variouslagged behind in the cell cycle in a dose-dependent

manner. The number of BrdUrd-positive cells still in S- doses of olomoucine tested in MR65 cells. No dose ortime-dependent differences in exit rate from this cellphase were higher with increasing dose of the inhibitor.

Delayed cell cycle progression in the presence of olo- cycle compartment were observed up to a dose of 200mM olomoucine, at which the exit rate from the G2/moucine was also obvious when analyzing the number

of cells occupying the BrdUrd-negative S-phase com- M compartment declined. In contrast, for the BrdUrd-

AID ECR 3700 / 6i26$$$183 09-09-97 09:33:29 eca

8 SCHUTTE ET AL.

FIG. 3. The effect of various doses of olomoucine (m, 0; l, 10; *, 50; l, 100; and j, 200 mM ) on cell cycle phase transitions in MR65cells. A shows the percentage of BrdUrd-negative G2/M-phase cells (abcissa) versus time in the presence of olomoucine (ordinate). B depictsthe fraction of undivided cells among the BrdUrd-positive population (abcissa) versus time in the presence of olomoucine (ordinate). C andD show the relative DNA content of the BrdUrd-negative G1/S-phase cells and BrdUrd-positive/undivided cells, respectively, as function ofexposure to olomoucine. The exit rate from G2/M-phase, the entry rates to G1- and S-phase, and the rate of progression through the S-phase of the cell cycle are calculated from the linear portions of the graphs in A, B, C, and D, respectively.

positive cells, the exit rate from the G2/M compartment showed a delayed G2/M phase transition rate. A possi-ble explanation for this phenomenon could be the factdecreased in a dose-dependent fashion as is outlined

in Fig. 3B. Similarly, the S-phase entry rate also de- that it takes some time for the cells to acquire a suffi-cient high intracellular dose of the inhibitor to be effec-creased in a dose-dependent manner (Fig. 3C). No evi-

dence is found for either a delayed entry of cells into tive. In order to test this hypothesis cell cultures werepreincubated for 3 h with various doses of roscovitineG1- or S-phase in the presence of olomoucine, only the

rate at which the cells progress is affected by the CDK prior to BrdUrd pulse labeling. The cultures were sub-sequently chased for 3 h and harvested. The decreaseinhibitor. This is also the case for the progression of

cells through the S-phase of the cell cycle as depicted in the number of BrdUrd-negative G2/M phase cellswas normalized as compared to the control culture. Asin Fig. 3D. At the highest dose of olomoucine tested;

i.e., 200 mM the cells could be completely arrested in can be seen from Fig. 6 the decrease in the number ofBrdUrd-negative G2/M-phase cells was less as the dosethe S-phase.

The changes in cell cycle progression for both MR65 of roscovitine increased.In order to evaluate the CDK levels in MR65 cells,and CHP-212 cells were normalized and shown in Figs.

4A and 4B. Both cell lines showed similar dose-depen- incubated for various time periods in the presence ofboth inhibitors, cell lysates were prepared and sub-dent kinetic changes. The main effect of olomoucine is

on the exit rates from G1- and G2/M phase of the cell jected to polyacrylamide gel electrophoresis followed byWestern blotting. For both olomoucine and roscovitinecycle, while the progression rate through the S-phase

of the cell cycle was affected to a lesser extend. Incuba- treated cells a decrease in total CDK1 protein levelswere observed (Fig. 7). This decrease in the level oftion of MR65 cells with roscovitine showed exactly the

same results. Figure 5 shows the differential dose-de- CDK1 protein was not due to a nonspecific effect of cellcycle inhibition since no decrease in protein levels waspendent effects of both olomoucine and roscovitine on

G1-/S-phase transition and S-phase progression. How- found when the culture were exposed to hydroxyurea.This loss was confirmed by bivariate CDK1/DNA analy-ever, the ID50 values for roscovitin were 5- to 10-fold

lower as compared to olomoucine. sis (Fig. 8A) in which increasing amounts of CDK1-negative cells were detected when cells were culturedA contradictory finding for both CDK inhibitors was

that the cells that were originally in the G2/M phase in the presence of increasing amounts of roscovitine.Furthermore, the CDK1-negative cells showed charac-of the cell cycle at the time of BrdUrd administration

progressed normally while the BrdUrd-positive cells teristics of apoptosis, i.e., loss of DNA (Fig. 8A) and

AID ECR 3700 / 6i26$$$183 09-09-97 09:33:29 eca


The MR65 and CHP-212 cells showed different kinet-ics of induction of apoptosis at high olomoucine dose.Figure 10 shows that for MR65 cells the frequency ofapoptotic cells gradually increased, with up to 50%apoptotic cells after 10 h at a dose of 200 mM olomou-cine. In CHP-212 cultures the number of apoptotic fig-ures increased rapidly after a 3-h lag time.

DISCUSSION

The molecular basis of activation and deactivation ofCDKs is mainly elucidated in cell-free systems usingpurified components. In this way many target proteinsare identified which are phosphorylated by differentCDK/cyclin complexed in vitro [1, 2]. However, it isdifficult to correlate events occurring on a molecularscale (e.g., activation) with those that occur on a cellu-lar scale (e.g., the start of S-phase). Studies at the cellu-lar level are largely hampered by the lack of specificinhibitors. Recently, Kitagawa et al. [39], Losiewicz etal. [49], and Vesely et al. [40] described specific inhibi-tors of certain CDKs (see review by Meijer [50]).

In the present study a detailed analysis of the cellcycle effects of olomoucine was performed. To this endFIG. 4. The dose-dependent effects of olomoucine on cell cycle

kinetics of MR65 (A) and CHP-212 cells (B). Data of G2/M-exit rates BrdUrd pulse-labeled cells in culture were chased inof BrdUrd-negative (*), BrdUrd-positive cells (l), and S-phase (l) the presence of various concentrations of the drug. Itentry rates and S-phase progression (m) are compared to untreated became obvious that only the exit rates of cells fromcontrol cultures after normalization. A clear dose-dependent de-

the G1- and the G2/M-phase of the cell cycle were de-crease in the various traverse rates can be observed, except for thecreased in a dose-dependent manner. However, whenG2/M-exit rate as calculated from the BrdUrd-negative G2/M phase

cells. This parameter is, however, affected by a dose of 200 mM olo- the exit rate from the G2/M compartment was calcu-moucine in the case of MR65 cells. lated, based on the kinetics of the originally BrdUrd-

negative G2/M phase cells, no effect was seen. Thiscould be explained by the fact that a certain period oftime is needed for a cell to accumulate a sufficientlyaltered scatter characteristics (data not shown). Again

no loss of CDK1 immunoreactivity was observed when high intracellular dose of olomoucine to compete withATP for the binding domain of the kinase. Exposingcells were exposed to hydroxyurea (Fig. 8B).

When cell cultures, treated with highest dose of olo- the cells to the CDK inhibitor 3 h prior to the pulselabeling experiment resulted in a delayed progressionmoucine or roscovitine, were examined microscopically,

many cells showed membrane blebbing and detached of BrdUrd-negative cells through G2/M, similar to thatof BrdUrd-positive cells. No evidence was obtained forfrom the culture plastic. To examine these cells in more

detail, they were harvested at various time intervals a delay in entry to or exit from the various cell cyclephases, suggesting that olomoucine remained active forand cytocentrifuged onto glass slides. The slides were

stained with propidium iodide and the nuclear mor- prolonged periods of time or that the inhibiting effectwas not reversible. The latter explanation is less likely,phology was examined using confocal scanning laser

microscopy. Figures 9A–9D show phase contrast im- since evidence is provided by others [51] that the effectof olomoucine is easily reversible.ages of such MR65 cell cultures and confocal images of

nuclear morphology of the corresponding cytocentri- One may argue that delayed progression through onecell cycle phase may affect the calculations on phasefuge preparations (Figs. 9E–9H). With increasing time

of exposure more cells showed membrane blebbing and transitions through the other cell cycle phases. For in-stance the rate of increase in the relative DNA contentdetachment from the culture dish. Simultaneously, the

frequency of cells showing chromatin aggregation, typi- of G1/S cells unlabeled with BrdUrd does not reflect theG1/S transition alone, but is also affected by both thecal for apoptosis, increased. Some cells clearly showed

chromatin condensation at the nuclear periphery, subsequent rate of DNA synthesis and the rate of divi-sion of BrdUrd-negative G2/M-phase cells to yield newwhile in the majority of cells the nucleus was com-

pletely broken down and only fragments of condensed G1-cells. However, the latter rate is identical for allexperimental conditions except for the highest testedchromatin were seen.

AID ECR 3700 / 6i26$$$183 09-09-97 09:33:29 eca

10 SCHUTTE ET AL.

FIG. 5. The dose-dependent effects of olomoucine (l) and roscovitine (l) on cell cycle kinetics of MR65 cells. Data of S-phase entryrates (A) and S-phase progression (B) are normalized compared to untreated control cultures after normalization. Roscovitine shows identicaleffects on cell cycle kinetics as compared to olomoucine at a 5- to 10-fold lower concentration.

dose of inhibitor. Furthermore, the normalized data in cine having a 10-fold increased efficiency toward CDKsand a similar low efficiency toward erk1, showed ex-Fig. 5 shows that S-phase progression is only affected

in a modest way, indicating that the rate of increase actly the same cell cycle effects, makes it very likelythat the observed cell cycle effects are caused by directin the relative DNA content of G1/S cells unlabeled with

BrdUrd mainly reflects the G1/S transition alone. Simi- inhibition of CDKs in vivo rather than by inhibition ofMAP kinases. Taken together, the data presented herelarly, it can be shown that the division of BrdUrd-posi-

tive cells mainly reflects the G2/M transition and is are consistent with the hypothesis of a direct effect ofolomoucine and roscovitine on both CDK2 and CDK1,only affected to a limited extent by the rate of comple-

tion of S-phase. governing the G1- to S-phase and the G2/M- to G1-phaseof the cell cycle, respectively [2, 7].The finding that roscovitine, an analogue of olomou-

Interestingly, in both MR65 and CHP-212 cell linesa decreased S-phase progression rate was observed.Since this effect cannot be explained by a direct inhibi-tion of DNA-polymerases by olomoucine (IC50s forthese enzymes exceed 200 mM ), a plausible explanationmay be that CDKs play a direct role in DNA-synthesis.For example, Dutta et al. [32] reported that CDKs areable to phosphorylate RPA, a human cell DNA replica-tion factor, and activate replication. In concordancewith this finding, Vesely et al. [40] showed inhibition

FIG. 7. CDK1 expression in MR65 cells incubated in the presenceFIG. 6. Relative decrease in the number of BrdUrd-negative G2/M phase cells. MR65 cells were exposed to 5, 10, and 20 mM Roscovi- of roscovitin. Cell lysates of exponentially growing cells (lane 1), cells

treated for 7 h in the presence of 10 mM (lane 3), 25 mM (lane 4),tine 3 h prior to (solid bars) or immediately after (open bars) pulselabeling with BrdUrd (solid bars). The cell cultures were subse- and 50 mM (lane 5) roscovitine were separated by gel electrophoresis

and transfered on PVDF membranes. Western blots were probedquently chased, harvested at different time intervals, and subjectedto bivariate BrdUrd/DNA analysis. The number of BrdUrd-negative with anti-CDK1. A decrease in total CDK1 protein levels is evident

in the cells treated with roscovitine. No change in CDK1 proteinG2/M phase cells are quantified and the percentage decrease in num-bers is expressed compared to untreated cultures. levels are observed in the hydroxyurea-treated culture (lane 2).

AID ECR 3700 / 6i26$$$184 09-09-97 09:33:29 eca


FIG. 8. Flow cytometric analysis of CDK1 expression in MR65 cells after 7-h exposure to 50 mM Rosovitine. (A) Bivariate CDK1/DNAanalysis, showing DNA content (ordinate) versus CDK1 immunofluorescence (abcissa). Note the loss of DNA in the CDK1-negative cells.The line indicates the threshold for positive immunofluorescence and was set based on the negative control allowing less than 1% positiveimmunoreactivity. (B) Histograms of CDK1 immunofluorescence of control cultures (solid line), and cultures treated for 7 h with 50mM Roscovitine (dotted line) and 0.2 mM hydroxyurea (striped line) treated cultures. The marker indicates the threshold for positiveimmunofluorescence and was set based on the negative control allowing less than 1% positive immunoreactivity.

AID ECR 3700 / 6i26$$3700 09-09-97 09:33:29 eca

12 SCHUTTE ET AL.

AID ECR 3700 / 6i26$$3700 09-09-97 09:33:29 eca


exposure of cells to hydroxyurea did not result in de-creased levels of CDK1 protein levels.

To date, data on the role of CDKs during apoptosisare inconclusive. On the one hand, many authors pro-vide evidence for activation of CDKs during apoptosis.In a study of Meikrantz et al. [55] the appearance ofcondensed chromatin was accompanied by a 2- to 7-fold increase in cyclin A-associated histone H1 kinaseactivity, where both CDK1 and CDK2 were activated.The authors suggested that cyclin A targets activatedCDK1 and CDK2 to substrates necessary for chromatincondensation and other morphological changes duringboth apoptosis and mitosis. The authors also suggestthat the sensitizing effect of cycloheximide, which wasshown to occur in the MR65 NSCLC cell line [56], wouldbe due to its ability to increase cyclin A-associated ki-FIG. 10. Kinetics of induction of apoptosis by olomoucine in

MR65 (j) and CHP-212 (l) cells. While the numbers of apoptotic nase activity rather than to inhibit the synthesis offigures increased gradually in cultures of MR65 cells, a sudden in- apoptosis-suppressing proteins. Hoang et al. [57] alsocrease in the number of apoptotic figures were seen in CHP-212

provided evidence for the participation of cyclin A incultures after 2 h. However, this frequency remained constant at amyc-induced apoptosis. They showed that both in pro-level of approximately 50% after 4 h in the presence of olomoucine.liferating cells and cells undergoing apoptosis, thecyclin A (but not B, C, D1, and E) mRNA level waselevated in unsynchronized myc-overexpressing cellsof DNA synthesis in cell-free assays. Also, the p21 in- when compared with parental rat fibroblasts.hibitor of cylin-dependent kinases has been found in

On the other hand, recent evidence suggests thatassociation with PCNA, an auxillary protein for thePITSLRE kinase isoforms are cellular targets for cas-DNA-polymerase d. In addition, Cardoso et al. [33] andpase-3-mediated cleavage during TNF-induced apop-Sobczak-Thepot et al. [34] showed by means of immu-tosis [63]. Furthermore, Yoshida et al. [64] showed evi-nocytochemistry that cyclin A specifically localizes atdence that staurosporine-induced apoptosis is not ac-subcellular sites of DNA replication.companied by intracellular activation of CDK1. AlsoWhen cell cultures were grown in the presence of athe observation made by Guagdagno [60] seem to con-high dose of olomoucine, many cells detached from thetradict a need for CDK activation during apoptosis.culture disk. Based on morphological criteria, such asThey showed that the appearance of cyclin A messen-cell shrinkage, membrane blebbing, chromatin conden-ger RNA and protein in late G1 was dependent on cellsation, and loss of DNA, as evidenced by bivariateadhesion. It is also known that prevention of substrateBrdUrd/DNA flow cytometric analysis, these cellsattachment leads to apoptosis by a process named an-showed the characteristic features of apoptosis. In ad-oikis, a term derived from the Greek word for home-dition, in a separate series of experiments we were ablelessness [61].to show another characteristic feature of apoptosis,

In the presented study apoptosis was induced by in-namely loss of membrane assymetry [52], as measuredhibiting CDKs. An explanation for the apparent dis-by phosphatidylserine exposure at the outer cyto-crepancies of many studies might be that activationplasmic membrane, in combination with an intactor inhibition of CDKs per se is not associated withmembrane integrity [53].apoptosis, but untimely activation or deactivation ofThe presented data suggest a direct link betweenthese kinases, implying that cyclins and CDKs are in-CDK inhibition and induction of apoptosis, as recentlyvolved at an important nodal point shared by pathwayssuggested by Park et al. [54]. Exposing cells to highregulating cellular proliferation and apoptosis. In linedoses of CDK inhibitors resulted in loss of CDK1 pro-with this reasoning are the observations made by Lahtitein levels in a subpopulation of cells. This was not due

to some nonspecific effect of cell cycle inhibition, since et al., showing that ectopic expression of certain PIT-

FIG. 9. Phase contrast images (A–D) of MR65 cultures prior to harvesting and linear projections of stacks of confocal images (E–H) ofPI-stained cytocentrifuged preparations after trypsinization. The cells were harvested after culturing for various time intervals in thepresence of 200 mM olomoucine; i.e., 0 (A, D), 2 (B, E), 4 (C, F), and 6 h (D, H). After 2 h in the presence of olomoucine some cells showedmembrane blebbing (B) and many cells detached from the culture dish (C, D). As evidenced by the nuclear morphology the number of cellswith condensed chromatin increased with longer time of exposure to 200 mM olomoucine (F–H). Bar, 10 mm.

AID ECR 3700 / 6i26$$$184 09-09-97 09:33:29 eca

14 SCHUTTE ET AL.

21. Hershko, A., Ganoth, D., Sudakin, V., Dahan, A., Cohen, L. H.,SLRE kinase isoforms, distantly related members ofLuca, F. C., Ruderman, J. V., and Eytan, E. (1994) J. Biol.the CDK1 gene family, can induce apoptosis.Chem. 269, 4940–4946.It is conceivable that timely activation of CDK com-

22. Kubiak, J. Z., Weber, M., de Pennart, H., Winston, N. J., andplexes commits a cell to enter and complete the cell Maro, B. (1993) EMBO J. 12, 3773–3778.cycle and on the other hand inhibits the apoptotic pro- 23. Glotzner, M., Murray, A. W., and Kirschner, M. W. (1991) Na-gram by modifying regulatory molecules. The differ- ture 349, 132–138, 1991.ences observed in the induction of apoptosis between 24. Murray, A. (1995) Cell 81, 149–152.MR65 and CHP212 cells, while the kinetics of cell cycle 25. Peeper, D. S., Parker, L. L., Ewen, M. E., Toebes, M., Hall, F. L.,interference are identical, are in line with such reason- Xu, M., Zantema, A., van der Eb, A. J., and Piwnica-Worms, H.

(1993) EMBO J. 12, 1947–1954.ing assuming differences in cellular composition of26. Ward, G. E., and Kirschner, M. W. (1990) Cell 61, 561–577.apoptosis regulatory molecules such as bcl2 and bax.27. Ewen, M. E., Sluss, H. K., Sherr, C. J., Matsushime, H., Kato,In summary, the present study shows that the cell

J., and Livingston, D. M. (1993) Cell 73, 487–497.cycle effects of olomoucine and roscovitine, i.e., inhibi-28. Kato, J., Matsushime, H., Hiebert, S. W., Ewen, M. E., andtion of the G1/S- and G2/M-phase transition as well as

Sherr, C. J. (1993) Genes Dev. 7, 331–342.S-phase progression, are consistent with the assumed29. Resnitzky, D., and Reed, S. I. (1995) Mol. Cell. Biol. 15, 3463–specific inhibition of CDK1 and CDK2. At high dose, 3469.

olomoucine and roscovitine are able to completely block 30. Devoto, S. H., Mudryj, M., Pines, J., Hunter, T., and Nevins,cell cycle progression and to induce apoptosis. The J. R. (1992) Cell 68, 167–176.unique specificity of olomoucine and roscovitine for 31. Krek, W., Ewen, M. E., Shirodkar, S., Arany, Z., Kaelin Jr.,CDKs make these compounds useful new tools for W. G., and Livingston, D. M. (1994) Cell 78, 161–172.studying cell cycle regulation at the cellular level. 32. Dutta, A., and Stillman, B. (1992) EMBO J. 11, 2189–2199.

33. Cardoso, M. C., Leonhardt, H., and Nadal-Ginard, B. (1993) Cell74, 979–992.REFERENCES

34. Sobczak-Thepot, J., Harper, F., Florentin, Y., Zindy, F., Brechot,C., and Puvion, E. (1993) Exp. Cell Res. 206, 43–48.1. Hunter, T., and Pines, J. (1991) Cell 66, 1071–1074.

35. Nigg, E. A. (1993) Trends Cell Biol. 3, 296–301.2. Motokura, T., and Arnold, A. (1993) Biochem. Biophys. Acta36. Peter, M., and Herskowitz, I. (1994) Cell 79, 181–184.1155, 63–78.37. Hunter, T., and Pines, J. (1994) Cell 79, 573–582.3. Meyerson, M., Enders, G. H., Wu, C. L., Su, L. K., Gorka, C.,

Nelson, C., Harlow, E., and Tsai, L. H. (1992) EMBO J. 11, 38. Pines, J. (1994) Nature 369, 520–521, 1994.2909–2917. 39. Kitagawa, M., Okabe, T., Ogino, H., Matsumoto, H., Suzuki-

4. Bates, S., Bonetta, L., MacAllan, D., Parry, D., Holder, A., Dick- Takahashi, I., Kokubo, T., Higashi, H., Saitoh, S., Taya, Y.,son, C., and Peters, G. (1994) Oncogene 9, 71–79. Yasuda, H., Ohba, Y., Nishimura, S., Tanaka, N., and Oku-

yama, A. (1993) Oncogene 8, 2425–2432.5. Morgan, D. O. (1995) Nature 374, 131–134.40. Vesely, J., Havlicek, L., Strnad, M., Blow, J. J., Donella-Deana,6. Pines, J. (1995) Semin. Cancer Biol. 6, 63–72.

A., Pinna, L., Letham, D. S., Kato, J., Detivaud, L., Leclerc, S.,7. Sherr, C. J. (1993) Cell 73, 1059–1065.and Meijer, L. (1994) Eur. J. Biochem. 224, 771–786.8. Dalton, S. (1992) EMBO J. 11, 1797–1804.

41. Schulze-Gahmen, U., Brandsen, J., Jones, H. D., Morgan, D. O.,9. Pines, J., and Hunter, T. (1991) J. Cell Biol. 115, 1–17. Meijer, L., Vesely, J., and Kim, S. H. (1995) Proteins Struct.10. Koff, A., Giordano, A., Desai, D., Yamashita, K., Harper, J. W., Funct. Genet. 22, 378–391.

Elledge, S., Nishimoto, T., Morgan, D. O., Franza, B. R., and 42. Schlesinger, H. R., Gerson, J. M., Moorhead, P. S., Maguire, H.,Roberts, J. M. (1992) Science 257, 1689–1694. and Hummeler, K. (1976) Cancer Res. 36, 3094–3100.11. Dulic, V., Lees, E., and Reed, S. I. (1992) Science 257, 1958– 43. Schutte, B., Reynders, M. M. J., Van Assche, C. L. M. V. J.,

1961. Hupperets, P. S. J. G., Bosman, F. T., and Blijham, G. H. (1987)12. Resnitzky, D., Gossen, M., Burjard, H., and Reed, S. I. (1994) Cytometry 8, 372–376.

Mol. Cell. Biol. 14, 1669–1679. 44. Higashikubo, R., White, R. A., and Roti Roti, J. L. (1993) Cell13. Gong, J., Traganos, F., and Darzynkiewicz. Z. (1994) Cell Prolif. Prolif. 26, 337–348.

27, 357–371. 45. Begg, A. C., McNally, N. J., Schrieve, D. C., and Karchner, H.14. Tassan, J. P., Schultz, S. J., Bartek, J., and Nigg, E. A. (1994) (1985) Cytometry 6, 620–626.

J. Cell Biol. 127, 467–478. 46. Laemmli, U. K. (1970) Nature 227, 680–685.15. Solomon, M. J. (1994) Trends Biochem. Sci. 19, 496–500. 47. Towbin, H., Staehelin, T., and Gordon, J. (1979) Proc. Natl.16. Gu, Y., Rosenblatt, J., and Morgan, D. O. (1992) EMBO J. 11, Acad. Sci. USA 76, 4350–4354.

3995–4005. 48. Kitagawa, M., Higashi, H., Takahashi, I. S., Okabe, T., Ogino,17. Parker, L. L., and Piwnica-Worms, H. (1992) Science 257, 1955– H., Taya, Y., Nishimura, S., and Okuyama, A. (1994) Oncogene

1957. 9, 1549–2557.18. McGowan, C. H., and Russell, P. (1993) EMBO J. 12, 75–85. 49. Losiewicz, M. D., Carlson, B. A., Kaur, G., Sausville, E. A., and

Worland, P. J. (1994) Biochem. Biophys. Res. Commun. 201,19. Heald, R., McLoughlin, M., and McKeon, F. (1993) Cell 74, 463–474. 589–595.

50. Meijer, L. (1996) Chemical inhibitors of cyclin-dependent ki-20. Hoffmann, I., Clarke, P. R., Marcote, M. J., Karsenti, E., andDraetta, G. (1993) EMBO J. 12, 53–63. nases. In Progress in Cell Cycle Research (Meijer, L., Guidet,

AID ECR 3700 / 6i26$$$184 09-09-97 09:33:29 eca


S., and Tung, H. Y. L., Eds.), Vol. 1, pp. 381–363, Plenum Press, G. P. M., Ramaekers, F. C. S., and Schutte, B. (1995) Eur. J.Cell Biol. 68, 35–46.New York.

57. Hoang, A. T., Cohen, K. J., Barrett, J. F., and Bergstrom, D. A.51. Abrahams, R. T., Acquarone, M., Anderson, A., Asensi, A., Belle,(1994) Proc. Natl. Acad. Sci. USA 91, 6875–6879.R., Berger, F., Bergounioux, C., Brunn, G., Buquet-Fagot, C.,

Fagot, D., Glab, N., Goudeau, H., Goudeau, M., Guerrier, P., 58. Martin, S. J., Mc Gahon, A. J., Nishioka, W. K., La Face, D.,Houghton, P. J., Hendriks, H., Kloareg, B., Lippai, M., Marie, Guo, G., Th’ng, J., Bradbury, E. M., and Green, D. R. (1995)D., Maro, B., Meijer, L., Mester, J., Mulner-Lorillon, O., Poulet, Science 269, 106–107.S. A., Schierenberg, E., Schutte, B., Vaulot, D., and Verlhac, 59. Shi, L., Nishioka, W. K., Th’ng, J., Bradbury, E. M., Litchfield,M. H. (1995) Biol. Cell 83, 105–120. D. W., and Greenberg, A. H. (1994) Science 263, 1143–1145.

52. Koopman, G., Reutelingsperger, C. P. M., Kuijten, G. A. M., 60. Guadagno, T. M., Ohtsubo, M., Roberts, J. M., and Assoian,Keehnen, R. M. J., Pals, S. T., and van Oers, M. H. J. (1994) R. K. (1993) Science 262, 1572–1575.Blood 84, 1415–1420. 61. Ruoslahti, E., and Reed, J. C. (1994) Cell 77, 477–478.

53. Engeland, van M., Ramaekers, F. C. S., Schutte, B., Reuteling- 62. Lahti, J. M., Xiang, J., Heath, L. S., Campana, D., Kidd, V. J.sperger, C. P. M. (1996) Cytometry 24, 131–139. (1995) Mol. Cell. Biol. 15, 1–11.

54. Park, D. S., Farinelli, S. E., Greene, L. A. (1996) J. Biol. Chem. 63. Beyart, R., Kidd, V. J., Cornelis, S., Van de Craen, M., Dencker,271, 8161–8169. G., Lahti, J. M., Gururajan, R., Vandenabeele, P., Fiers, W.

55. Meikrantz, W., Gisselbrecht, S., Tam, S. W., and Schlegel, R. (1997) J. Biol. Chem. 272, 11694–11697.(1994) Proc. Natl. Acad. Sci. USA 91, 3754–3758. 64. Yoshida, M., Usui, T., Tsujimura, K., Inagaki, M., Beppu, T.,

Horinouchi, S. (1997) Exp. Cell Res. 232, 225–239.56. Tinnemans, M. M. F. J. Lenders, M. H. J. H. ten Velde,

Received January 16, 1997Revised version received June 10, 1997

AID ECR 3700 / 6i26$$$185 09-09-97 09:33:29 eca

the effect of the cyclin-dependent kinase inhibitor olomoucine on cell cycle kinetics

Documents