okadaic acid-induced apoptosis in neuronal cells: evidence for an abortive mitotic attempt

Journal of NeurochemistryLippincott—Raven Publishers, Philadelphia© 1998 International Society for Neurochemistry

Okadaic Acid-Induced Apoptosis in Neuronal Cells:Evidence for an Abortive Mitotic Attempt

*tRony Nuydens, *Mirjam de Jong, *Gerd Van Den Kieboom, tCara Heers,

§Gwenda Dispersyn, *Frans Cornelissen, *Roger Nuyens, *~Marce1Borgers,and *Hugo Geerts

* Department of Cell Physiology, Janssen Research Foundation, Beerse; § Department of Biochemistry, University of

Antwerp, Antwerp, Belgium; tDepartment of Molecular Cell Biology and Genetics, Maastricht University, Maastricht,The Netherlands; and ~Preclinical Development, Janssen Research Foundation, Neuss, Germany

Abstract: There is increasing evidence that apoptosis inpostmitotic neurons is associated with a frustrated at-tempt to reenter the mitotic cycle. Okadaic acid, a spe-cific protein phosphatase inhibitor, is currently used inmodels of Alzheimer’s research to increase the degree ofphosphorylation of various proteins, such as the microtu-bule-associated protein tau. Okadaic acid induces pro-grammed cell death in the human neuroblastoma celllines TR14 and NT2-N, as evidenced by fragmentation ofDNA and attenuation of this process by protein synthesisinhibitors. In differentiated TR14 cells, okadaic acid in-creases the fraction of cells in the S phase, induces theappearance of cyclin B1 and cyclin D1 markers of the cellcycle, and triggers a time-dependent increase in DNAfragmentation after release of a thymidine block. Fullydifferentiated NT2-N cells are forced to enter the mitoticcycle as shown by DNA staining. Chromatin condensa-tion and chromosome formation are initiated, but the cellsfail to complete their mitotic cycle. These data suggestthat okadaic acid forces differentiated neuronal cells intothe mitotic cycle. This pattern of cyclin up-regulation andcell cycle shift is compared with apoptosis induced byneurotrophic factor deprivation in differentiated rat pheo-chromocytoma PC12 cells. Key Words: Okadaic acid—Apoptosis—Neuronal cells—Mitotic cycle—Alzheimer’sdisease model—Nerve growth factor deprivation.J. Neurochem. 70, 1124—1133 (1998).

Apoptosis of neuronal cells plays a major role duringdevelopment and fine-tuning of the nervous system.The amount of neurotrophic factors released by thetarget tissue regulates the cell survival and the innerva-tion density of this particular tissue. There is, however,increasing evidence that this form of cell death mayalso contribute to some chronic neurological diseasestates (for a review, see Bredesen, 1995). In thesesituations of chronic neurodegeneration it may attenu-ate the inflammatory reaction to the surrounding cellsand tissues. As an example it has been shown thatduring Alzheimer’ s disease, neurons display the typical

characteristics of programmed cell death and as suchcontribute to the neuronal cell loss observed duringprogression of the disease (Su et al., 1994; Lassmannet al., 1995).

Recently, it has been argued that apoptosis in differ-entiated neuronal cells is a consequence of the cellularattempts to escape the G0 status and reenter the cellcycle (Park et al., 1996). Indeed, the programmed celldeath in neuronal cells is accompanied by numerouscharacteristics normally associated with proliferatingcells: up-regulation of p34cdc2 (Brooks et al., 1993)and induction of cyclin ~and cyclin D1 (Freeman etal., 1994) to levels normally seen during mitosis inproliferative cells.

One of the hallmarks in AD is the appearance ofhighly phosphorylated tau isoforms (Grundke-Iqbal etal., 1986) in paired helical filaments. This might leadto the dissociation between tau and microtubules andsubsequent cytoskeletal instability (Alonso et al.,1994). Recently, it was shown that in Chinese hamsterovary cells transfected with the neuronal microtubule-associated protein tau, this protein is hyperphosphor-ylated during mitosis as indicated by the use of phos-phorylation-sensitive antibodies and autoradiography(Pope et a!., 1994; Preuss et al., 1995). Whenapoptosis is induced in neuronal cells by sodium buty-rate, tau also becomes hyperphosphorylated (Nuydenset al., 1995a).

Aberrant tau phosphorylation can be induced inneu-roblastoma cells by treatment with okadaic acid

Received July 14, 1997; revised manuscript received October 6,1997; accepted October 6, 1997.

Address correspondence and reprint requests to Dr. H. Geertsat Department of Cell Physiology, Janssen Research Foundation,Turnhoutseweg 30, 2340 Beerse, Belgium.

Abbreviations used: DMEM, Dulbecco’ s minimal essential me-dium; MTT, 3- (4,5-dimethylthiazol-2-yl )-2,5-diphenyltetrazoliumbromide; NGF, nerve growth factor; OD, optical density; P1, propid-ium iodide; XTT, 2,3-his (2-methoxy-4-nitro-5-sulfophenyl-2H-tet-razolium-5-carboxanilide inner salt.

1124

OKADAIC ACID-INDUCED NEURONAL APOPTOSIS 1125

(Vandermeeren eta!., 1993; Sautiere eta!., 1994; Shea,1996). At nanomo!ar concentrations okadaic acid is aspecific inhibitor of protein phosphatase 2A activity.The first evidence for an effect of okadaic acid on thecell cycle came from the observation that low concen-trations of this inhibitor interfered with the normalprogression of LLC-PK cells through mitosis byblocking the ce!!s in a metaphase-!ike state. After pro-longed incubation, mu!tipo!ar spind!es formed, and themetaphase plate was disrupted (Vandre and Wi!!s,1992).

In this study we first show that okadaic acid inducesapoptosis in human neurob!astoma TR14 cel!s, in bothproliferative and differentiated states, and in fully dif-ferentiatedNT2-N neurons, which can serve as a mode!for human postmitotic neurons. We compared this pro-cess with another well-documented example of pro-grammed cell death, such as nerve growth factor(NGF) deprivation in fully differentiated PC 12 cells.We further investigate the relation between neuronalapoptosis and the reappearance of mitosis-relatedevents, within the context of the cell cycle/apoptosismode!.

MATERIALS AND METHODS

Cell culturePC12 cells were grown in Dulbecco’s minimal essential

medium (DMEM) supplemented with 7.5% fetalbovine se-rum and7.5% horse serum. For the actual experiments, PC12cells were plated on poly-L-!ysine-coated covers!ips. After1 day the cells were placed in DMEM supplemented with1% horse serum and 50 ng/ml NGF (Sigma) for at least 7days, and the medium was changed every 3—4 days. Forthe deprivation experiments medium was changed either toDMEM alone or to DMEM plus NGF. TR14 neuroblastomacells were a generous gift of Dr. G. Carmeliet (Centre forHuman Genetics-KU, Leuven, Belgium) and grown inDMEM/F12 medium supplemented with 10% fetal calf se-rum. Differentiation was induced by treating thece!!s during4 days with 1 mM dibutyryl cyclic AMP in serum-free me-dium.

Synchronized cultures were obtainedafter adouble thymi-dine block with 2 mM thymidine, for 16 and 18 h, respec-tively, separated by an 8-h recovery period.

NT2-N human neuroblastomas were differentiated as de-scribed (Pleasure et al., 1992). In brief, 2 x 106 hNT2 cellswere treated in T75 flasks with 10

1uM retinoic acid for 4weeks. The cells were differentially harvested and replatedin DMEM with 10% fetal calf serum and supplementedwith1 ~aMcytosine arabinoside, 10 ~M fluorodeoxyuridine, and10 ,uM uridine.

Cell survivalCel! surviva! was measured using 3- (4,5-dimethylthiazol-

2-yl )-2,5-diphenyltetrazolium bromide (MTT; in PC12cells) or 2,3-bis ( 2-methoxy-4-nitro-5-sulfophenyl) -2H-tet-razolium-5-carboxanilide inner salt (XTT; in TR14 cells).This is based on the conversion of the substrate to a water-soluble formazan product by mitochondrial dehydrogenasesof viable cells only (Mosmann, 1983). At appropriate times,a prewarmed solution of 0.3 mg/mi XTT or 0.5 mg/mi MTT

in Ca2~-and Mg2~-free phosphate-buffered saline was

added to the culture medium. The plates were further incu-bated for 2 h at 37°Cand shaken for 10 mm. Optical density(OD) was read at 450 nm using 650 nm as a reference ona plate reader (Molecular Dynamics). In each 96-well platetwo determinations of OD at each concentration were per-formed.

Time-lapse video microscopyAn MXR (Adimec, Eindhoven, The Netherlands) camera

was mounted on top of an Axiovert 35 (Zeiss, Oberkochen,Germany) microscope. The video output was connected tothe VINO analogue video input port on an Indy Workstation(Silicon Graphics, U.S.A.). Image processing was per-formed using the SCIL software package (University of Am-sterdam, Amsterdam, The Netherlands). A computer-drivenmotorized stage (Merzhauser, Germany) was used to recordseveral sets of (x, y, z) coordinates corresponding to thefields of interest.

Movies were composed by driving the stage to the fields,executing an autofocus routine, grabbing the image, and ap-pending it to the movie file of the current field in the SGImovie format. The COSMO hardware JPEG (de)compres-sion board allowed real-time display of the movies (25frames/s). Movies can be analyzed by reading each framein turn and extracting the desired features with the SCIL-Image software.

Cell staging by flow cytometryAt the appropriate time points, cells were mechanica!ly

detached from the substrate by scraping, centrifuged, fixedwith 70% ethanol for 1 h at 4°C,and rinsed in phosphate-buffered saline. The cells were then pelleted and incubatedwith RNase and propidium iodide (P1). The stained cellswere analyzed on a Cytoron flowcytometer (Ortho Pharma-ceuticals, U.S.A.). The various fractions of the cell cyclewere calculated with theModFit cell cycle analysis program.

ImmunocytochemistryAt the appropriate time points, the celis were fixed with

paraforma!dehyde for 10 mm followed by methanol at—20°C overnight. The following antibodies were used forimmunodetection: cyclin B

1 (clone2H 1 -H6; Ca!biochem),cyclin D1 (clone Gl24-326; Pierce), p34cdc2 (Santa CruzBiotechnology), and AT8 (Innogenetics). Goat anti-mouse—cyanine 3 (GAM-Cy3 ) -labeled antibodieswere usedas secondary antibodies. Double labeling for tubulin wasperformed using an affinity-purified polyclonal antibody ora monoclonal anti-tubulin (Sigma Biosciences). In that casegoat anti-rabbit—fluorescein isothiocyanate (GAR-FITC ) -

labeled antibodies were used. Nuclei were counterstainedusing 4,6-diamidino-2-phenylindole, and the preparationswere mounted in Gelvatol containing 100 mg/ml diazabi-cyc!o[2.2.2]octane to avoid bleaching.

Quantitative light microscopyThe preparations were observed in epifluorescence mode

on an Axiovert 35 microscope equipped with a Merzhauserscanning stage and MXRi CCD camera and were digitizedby an Indy Workstation (Silicon Graphics). Image analysiswasperformed with the SCIL package on the Silicon Graph-ics workstation.

In each image the total number of cells and the degreeof epitope expression were determined by fully automaticthresholding. In brief, cells and rieurites were detected by aHave edge-detection routine followed by a holeclose opera-

J. Neurochem., Vol. 70, No. 3, 1998

1126 R. NUYDENS ET AL.

FIG. 1. lmmunocytochemistry of AT8 in fully differentiated NT2-N cells observed with fluorescein isothiocyanate microscopy. In controlconditions (a), a moderate signal in the cell body is observed. No clear staining of axons is visible. Application of 50 nM okadaic acidfor 24 h (b) increases the ATB signal, suggesting an increased phosphorylation of tau proteins at Ser202/Thr205. Note the morepronounced increase in the cell body versus the axons and the distal-to-proximal gradient of AT8 label.

tion. This ensured correction for the varying background.The resulting binary image, containing only the cell bodiesand neurites, was labeled, and nonspecific small objects, i.e.,those below a critical size, were exc!uded. From this labeledimage, the fluorescence intensity of each individual (cellu-lar) object n was reconstructed by ANDING the originalimage with a binary image, derived from the labeled imagecontaining only pixels belonging to that cell n. From thislist of cellular objects, parameters such as mean intensitycan be calculated.

Analysis of DNA fragmentationDNA labeling was performed on living cells by means of

the SYTO-9 probe (Molecular Probes, Eugene, OR, U.S.A.),and the preparations were observed on an Axiop!an micro-scope using classical fluorescein isothiocyanate optics (~.< 505 nm; band-pass emission 520 < X~< 560 nm),whereas P1 labeling was visualized using rhodamine filters(X~.< 560 nm; emission 590 nm < Xe,,,). This made itpossible to distinguish clearlybetween intact cellmembranesand intact nuclei versus mitotic and versus fragmented nuclei(apoptotic cells). At least 300 cells were evaluated for eachcondition.

Statistical analysisStatistical analysis was carried out with Dunnett’s t test

and the Pearson x2 test, where appropriate, using the statisti-cal software package JMP.

RESULTS

Induction of neuronal apoptosisIn differentiated NT2-N neurob!astoma cel!s, oka-

daic acid, at a concentration of 50 nM, increases thedegree of tau protein phosphory!ation as illustrated byAT8 immunocytochemistry (Fig. 1). Similar resultshave been obtained in cultures of differentiated TR14cells. In these cells the phosphatase inhibitor okadaicacid showed a biphasic dose-dependent effect on cell

survival as measured by XTT conversion (Fig. 2). Atconcentrations beginning at 50 nM and up, cellulartoxicity dose-dependently increased. For instance, at100 nM, cell survival was reduced to 51 ±5% (n= 36), whereas application of the protein synthesisinhibitor cycloheximide at 350 ,uM increased cell sur-vival to 69 ±7% (n = 36; p <0.01). Manual countingshowed that treatment of TR14 cells with 1 ,uM okadaicacid actually increased the fraction of fragmented nu-clei from 5 to 27% in 24 h. However, at subtoxicconcentrations (between 1 and 10 nM), a substantialincrease in measured XTT signal was observed, espe-cially in the differentiated cultures. As expected, theinactive form of okadaic acid, norokadaic acid, did notshow any toxicity at all.

FIG. 2. Dose-dependent effect of okadaic acid (OA) and its in-active isoform, norokadaic acid (NorOA), on the survival of dif-ferentiated TR14 cells after a 24-h exposure (treated with 1 mMdibutyryl cyclicAMP for 3 days before treatment). Data are nor-malized to OD values in the absence of OA. °p< 0.05 withDunnett’s t test.



TABLE 1. Effect of okadaic acid treatment for 24 h onsurvival of both nond(fferentiated (in serum-containing

medium) and differentiated (treated for 3 days with1 mM dibutyryl cyclic AMP) TR14 cells

Concentration ofokadaic acid

(nM)Nondifferentiated

cellsDifferentiated

cells

0 100 1000.1 101.8 ±8.8 103 ±8.20.5 100.4 ±6.29 112 ±11.61 103.8 ±7.3 117.2 ±12.25 112.7 ±12.3 139.3 ±21.5”

10 130 ±8.5” 158.3 ±36.8”50 84.3 ±30.3 64.9 ±16.1”

100 62.6 ±13.7” 57.1 ±25.9”500 62.1 ±6.5” 59.1 ±16.4”

Data are mean ±SEM values from at least sixXTT measurements of viability.

p < 0.05 by Dunnett’ s t test versus solvent control.

experiments for

phase-contrast video microscopy. Addition of a toxicokadaic acid concentration to TR14 cells induced anactivation of the overall ce!!u!ar activity when com-pared with control cultures: ruffling activity, increasedmobility, and growth cone activity. Another strikingphenomenon observed in the time-lapse experimentson TR14 cells was the high number of cells unsuccess-fully entering the mitotic cycle (Fig. 4). Sometimesthis resulted in the formation of multipolar spindlesand subsequently the formation of multinucleated cells.Higher concentrations of okadaic acid induced a rapidrounding of the cells, and this was accompanied bythe formation of pro- and prometaphase cells. Thechromosomes did not, however, align to form a propermetaphase plate. These cells quite often assumed ana-phase-!ike configurations but without chromatid sepa-ration. Finally, the chromatin recondensed, and frag-mented nuclei typical for apoptotic cells were formed.In some cells the irregular spindle formation led to theformation of multinucleated cells.

TRI4 cells maintained in the presence of serum andin the absence of any dibutyryl cyclic AMP wereslightly less sensitive to the effects of okadaic acid(Table 1).

In PC12 cells the removal of NGF caused 30% celldeath within 24 h (Farine!!i and Greene, 1996). In ourPC12 system, survival as measured using the MTTassay—was reduced to 75% after a 24-h deprivationof NGF. Manual counting of nuclear morphology re-vealed that the fraction of apoptotic nuclei increasedfrom 0.4 to 6.5% after 8 h of NGF deprivation.

Apoptosis and mitosis in synchronized TR14 cellsIf apoptosis and mitosis have some mechanisms in

common, then these events should show a temporalcorrelation. To study this hypothesis, we b!ocked pro-liferative TR14 cells in S phase by a 16-h thymidineblock, followed by an 8-h recovery period and finallyan 18-h thymidine block. One hour after this last thy-midine block, okadaic acid (iO-~M) or solvent wasadded to the cultures. At different time points the cellswere fixed and permeabilized, and the percentages ofapoptotic or mitotic cells were determined microscopi-cally using PT as DNA label. Ce!!s were consideredas mitotic when the chromosome pairs were readilydiscernible. Apoptotic cells were characterized bymeans of nuclear fragmentation. In control cultures thepercent mitotic TR14 cells linearly increased starting 2h after removal of the last thymidine block. In culturestreated with okadaic acid the fraction of apoptotic cellstime-dependently increased as well, with a similar lin-ear relationship as in the control conditions. A typicalexample is given in Fig. 3.

Altered chromatid separation during meta- andanaphase

To investigate further the spatial and temporal orga-nization in TR14 cells during progression throughapoptosis, we followed these cells with time-lapse

Sequence of events during apoptosis in NT2-Ncells

In fully differentiated NT2-N cells treated with 0.1p~Mokadaic acid, similar events occurred as observedin TR14 cells. Table 2 shows the effect of i0~ Mokadaic acid on the nuclear morphology of fully differ-entiated NT2-N neurons. In this experiment, livingcells with intact nuclei are Plsnegative and have a uni-form SYTO staining of the cell nuclei. Apoptotic nu-clei showed a fragmented nuclear SYTO staining andwere P1-negative (because they have still an intactplasma membrane), whereas necrotic cells were P1-positive (suggesting that their plasma membrane isdeficient). Okadaic acid increased the fraction of apo-ptotic cells from 2 to 18%, an increase that was par-tially attenuated by the broad-spectrum kinase inhibitorstaurosporine (Table 2).

FIG. 3. Time-dependent increase in apoptotic DNA fragmenta-tion in synchronized TR14 cells treated with 1 jiM okadaic acid(left axis; •) and mitotic figures in untreated synchronized TR14cells (right axis; 0), both after release of the blocking agentthymidine at t = 0. Both the percent cells with mitotic figuresand with DNAfragmentation follow a linear time course, with thelatter being five times as fast.

.1. Neurochem., Vol. 70, No. 3, 1998


FIG. 4. a: Time-lapse sequence on the effects of okadaic acid on fully differentiated TR1 4 cells. Time is indicated in the lower rightcorner (h:min). Application of low doses of okadaic acid (10 nM) induces abnormal cell division. A cell is shown (arrow) entering themitotic phase at 09:37 that finally results in a multipolar division some 5 h later. b: When the concentration of okadaic acid is increasedup to 100 nM (after 60:00 h), the cells first become very active, increase their cell volume, and finally round up and die throughapoptosis, a process completed in a few minutes (from 68:45 to 68:49).



TABLE 2. Effect of 0.1 jiM okadaic acid on nuclearmorphology offully differentiated NT2-N cells

Condition Intact nuclei Apoptotic Necrotic

Control 97.8% 2.15% 0%Staurosporine 94.8% 3.8% 1.2%Okadaic acid” 74.4% 18.2% 7.4%Okadaic acid and

staurosporine6 89% 8.8% 2.1%

The fraction of cells (measured on at least 200 cells) in each ofthree categories is given. Intact nuclei are considered to be thosethat are P1-negative and that have a uniform SYTO staining, apo-ptotic nuclei show a fragmented nuclear SYTO signal and are PT-negative, whereas necrotic cells are P1-positive.

“p < 0.05 versus control by y2 test.~‘p< 0.05 versus okadaic acid by x2 test.

In another series of experiments we treated cellswith0.1 jiM okadaic acid, fixed them at different times,and stained the DNA with 4,6-diamidino-2-pheny!in-dole. In these preparations we counted the number ofcells that assumed mitosis-like configurations. Duringthe first 8 h there was a small increase, from 0% incontrol cultures to 5% in okadaic-acid treated cultures;after 24 h, however, the number of “mitotic cells”increased up to 25%. Figure 5 illustrates the differentstages of chromosomedistribution during the apoptoticprocess up to the formation of the fragmented nuclei.

Cell cycle distribution during apoptosisIn the next series of experiments, using DNA analy-

sis, we compared the relative fraction of cells in differ-ent stages of the cell cycle in okadaic acid-treatedTR 14 cells with that in PC 12 cells after NGF depriva-tion (Fig. 6). Flow cytometry data clearly showed thaton removal of NGF a substantial number of PC 12 cellsleave G

0/G1 and reinitiate DNA synthesis. Within 8 hafter deprivation the percentage of cells in G0/G1 wasdecreased from 78 to 67%, with a compensatory in-crease in S phase from 10 to 22%. After 24 h thefraction of cells in G0/G1 phase decreased to 58%,with a concomitant further increase in S phase to 29%.Similarly, analysis of the fraction of TR14 cells indifferent stages of the cell cycle revealed that okadaicacid treatment increased the fraction of cells in S phasefrom 36 to 52.4% and reduced the fraction of cells inG0/G~phase from 50.6 to 35.4% (Fig. 5).

Expression of cell cycle-related epitopesIf the onset of the apoptotic process is related to

initiation of mitosis, then this should also be reflectedinthe expression levels of certain G2/M phase-associatedproteins. Usingquantitative fluorescence microscopy weinvestigated the effect of okadaic acid in TR14 neuro-blastoma ce!!s and NGF deprivation in PC12 cells onthe levels of cyclin B1, p34°’~

2,and cyclin D1.

Visual scoring of >250 PC12 cells in each conditionrevealed that the fraction of p34 cdc2 positive cells in-creased from 2.5 to 6% at 8 h of deprivation and to

6.2% at 24 h of NGF deprivation. In the case of okadaicacid-treated TR14 cells (0.05 jiM), the fraction ofcyclin B1-positive cells increased from 2.2 to 5.2% at24 h, and that of p34”°

2-positivecells increased from1.8 to 7.3% at 24 h. Moreover, automatic image analy-sis showed that the levels of cell cycle marker proteins,such as cyclin B

1, cyclin D1, and p34cdc2 were notaffected after 8 h of okadaic acid treatment, but espe-cially cyclin D1 levels and to a lesser extent cyclin B1and p34~

2levels increased after 24 h of treatment(Tab!e 3). For instance, cyclin D

1 levels increased by30%, similar to the 35% increase in fraction of cellsin the S phase as assessed by DNA flow cytometry.The comparison between manual scoring and auto-matic image analysis is illustrated by the followingexample, based on the expression levels of cyclin B1in control versus okadaic acid-treated cells (mean in-tensity, 123.7 vs. 143.0). Based on the cumulativehistogram of the intensities of the individual cells anddefining an intensity threshold for a positive cell asmean intensity plus 2 SD (143.0 + 2 X 28.1 = 199),we find 2.7% positive cells in control situations and5.7% positive cells in okadaic acid conditions. Thiscompares favorably with the 2.2 and 5.2%, respec-tively, obtained by manual counting.

DISCUSSION

This report shows that in human neuroblastomacells, aberrant phosphorylation of tau can be inducedby application of the phosphatase inhibitor okadaicacid and is associated with a form of programmed celldeath. In addition, it is shown that this form of neuronalcell death is accompanied by a frustrated attempt toenter mitosis. The effect is observed both in prolifera-tive and in fully differentiated neuronal cells.

In differentiated PC12 cells the process of pro-grammed ce!l death is initiated when differentiatedcells are deprived of NGF (Park et a!., 1996), whereasin human TR14 and fully differentiated NT2-N cellsthis can be initiated by application of okadaic acid.Using these three different cell systems allowed us tocompare species, human versus rat, but more impor-tantly the differences between proliferative and differ-entiated cells. It also showed that the artificial okadaicacid model does share some features with the morephysiological model of NGF deprivation. In both NGFdeprivation and okadaic acid treatment there is a sub-stantial dose-dependent loss of cellular viability over24 h. Furthermore, the attenuation of cytotoxicity bythe protein synthesis inhibitor cycloheximide and theappearance of DNA fragmentation and membraneb!ebs suggest that okadaic acid-induced cytotoxicity isapoptosis. Similar results have been found in variouscultured retinoblastoma cells (Inomata et a!., 1996).

It is intriguing that okadaic acid at subtoxic concen-trations induces a dose-dependent increase of the bio-chemical parameters indicative of cell viability, whichmight be explained by an increasing number of cells



FIG. 5. Illustration of the different stages of chromatin distribution and DNA morphology monitored with the membrane-permeableDNA probe SYTO-9 in fully differentiated NT2-N cells treated for 24 h with 100 nM okadaic acid. Fluorescence images show thatchromosomes are actually formed (a), but they fail to separate (b), leading to reaggregation (c) and the formation of fragmentednuclei (d).

forced to enter the cell cycle. This is corroborated bythe observation in our time-lapse experiments thatsome ce!!s actually were able to escape the mitoticblock and went on to form multinucleated cells. Inaddition, the effect is less pronounced on nondifferenti-ated TR14 cells than in their more differentiated coun-terpart. This could be partly explainedby the consider-ation that the effect of the forced mitosis induced byokadaic acid is masked by the higher normal basalproliferative rate in the nondifferentiated condition. Al-ternatively, cells in the proliferative phase may havetheir machinery for cell cycle-related events moreready. A similar observation has been reported in pro-liferating LLC-PK cells where okadaic acid was docu-mented to induce a mitotic block in a metaphase-likestate within 6 h after the initial treatment (Vandre andWills, 1992).

Differentiated neurob!astoma cells, such as the dibu-tyry! cyclic AMP-treated TR14 cells (Rupniak et a!.,1984) or the retinoic acid-treated NT2-N cells (Plea-

sure et a!., 1992), can be considered almost synchro-nized as they either are mostly in G0 or G1 phase.However, it is not excluded that they maintain someproliferative capacities and that this system might bereactivated by okadaic acid. This question was ad-dressed by synchronizing proliferating TR14 cells inS phase through addition of excess thymidine. Re-moval of the thymidine block allowed the cells to pro-ceed through the cell cycle, and addition of okadaicacid at the time of thymidine removal induced a !inearincrease in the number of apoptotic ce!ls. A time-de-pendent increase in the number of fragmented nucleipara!!e!ed the appearance of mitotic ce!!s in contro!conditions. This indicated that inhibition of phospha-tase activity actually seems to enhance steadily theprogression of the cell through the mitotic cycle, aprogression that eventually ended up in an apoptoticprocess. Again, this observation suggests that mitosisand apoptosis may use simi!ar processes.

This report a!so documents that both in NGF-de-



FIG. 6. Fraction of cells in different phases of thecell cycle as measured by flow cytometry: Pci 2 cells(a) in the presence of NGF and (C) after NGF depri-vation; TR14 cells (b) in control conditions and (d)in the presence of 100 nM okadaic acid (OA). Notethat with both NGF deprivation and OA treatmentconditions a shift from cells in G0/G1 phase to Sphase takes place, whereas the fraction of cells inthe G2/M phase essentially remains unchanged.Both experiments were sampled after 24 h.

prived PC12 cells and in okadaic acid-treated TR14cells, the induction of apoptosis is accompanied bythe expression of specific M phase-related epitopes,especially cyclin D1, suggesting again that cellsseemed to be forced into a mitotic cycle. The inductionof cyclin D1 is prominent at subtoxic concentrationsof okadaic acid.

Similar cell cycle arrest induced by okadaic acid hasbeen documented in the human myeloid !eukemic celllines HL-60 and U937 (Ishida et a!., 1995). This ex-tends earlier observations of Gao and Ze!enka (1995),who showed that cyclin D1 and cyclin B are up-regu-lated after NGF deprivation. Brooks et a!. (1993)

showed that under these conditions p34 cdc2 levels alsowere increased. In lymphocytes it was shown that theactivation of p34°’°

2was necessary for the progressionof apoptosis after exposure to a lymphocyte granuleprotease (Shi et al., 1994). Flow cytometric analysis~f the cell stages in PC 12 cells indicated that inductionof apoptosis was preceded by reinitiation of the cellcycle in these cells normally considered as blocked inG

0—G1 phase. The induction of programmed cell deathapparently forces the cells to enter S phase and startDNA synthesis, as evidenced by the decreased cellnumber in G0—G1. This result is in line with the obser-vation that blocking the cell cycle before G1 /S phase

TABLE 3. Effect of okadaic acid (OA) treatment (50 nM) on induction of the cell cycle-relatedproteins cyclin B1, ~34~d~2 and cyclin D, by means of quantitative immunocytochemistry

Condition in TR14 cells Cyclin B1 p34~2 Cyclin D

1

Control 24 h 123.7 ±1.42 (n = 301) 156.9 ±2.15 (n = 181) 101.6 ±1.85 (n = 107)OA 24 h 143.0 ±2.01 (n = 193)” 165.4 ±2.72 (n = 1l1)~’ 132.8 ±2.13 (n = 185)”

Data are mean ±SEM fluorescence intensities per cell after background correction (no. of cells).“p < 0.01, b~ < 0.05 by Student’s t test versus control at 24 h.



prevents the death of trophic factor-deprived PC 12cells, whereas agents that act at different points duringthe cell cycle did not show any long-term protectiveeffects (Farinelli and Greene, 1996). Induction ofcyclin D1 has been documented to be a hallmark ofmitogenic cells entering the mitotic cell cycle, whereit has been implicated in the G1 checkpoint (for review,see Darzynkiewicz et al., 1996). It is interesting thatboth the morpho!ogica! observations of chromosomeformation in okadaic acid-treated NT2-N cells and theappearance of the cyclin B1 marker suggest that somecells at least are able to pass this G1 checkpoint. There-fore, the data imply that nontoxic concentrations ofokadaic acid actually force differentiated neuro-blastoma cells into mitosis and that bifurcation into anapoptotic pathway actually may take place after the G1checkpoint.

We documented earlier that features typical of cyto-kinesis were observed in TR14 cells when these cellswere exposed to hyperstimulating medium, leading ul-timately to apoptosis (Nuydens et al., 1995a,b).

Taken together with previous results, this reportshows that on induction of apoptosis neurona! cellsresume DNA synthesis and attempt to enter the cellcycle. During M phase, synchronization is lost, andchromosome separation is not completed. At this pointthe metaphase checkpoint is probably overridden, andfragmented nuclei are formed.

One of the key players in this process is the microtu-bular cytoskeleton. An important regulatory mecha-nism for the meta- to anaphase transition is the chemi-cal and mechanical interaction between microtubulesand the kinetochores (Gorbsky, 1995). If the stabilityof the microtubules is in any way impaired, properchromosomealignment is prohibited, and consequentlymetaphase to anaphase transition is delayed. One ofthe factors influencing microtubule stability is thephosphorylation state of microtubule-associated pro-teins. It has been shown that aberrant phosphory!ationof tau diminished its stabilizing effect on microtubules(Alonso et al., !994; Shea, 1996). In human platelets,application of okadaic acid induces a dramatic frag-mentation of microtubules (Yano et al., 1995). Theapplication of okadaic acid to microtubules changesthe dynamic instability of microtubules from an in-terphase to a mitosis-like state by reducing the rescuefrequency (Gliksman eta!., 1992). In Chinese hamsterovary cells, okadaic acid reduces the number of micro-tubules, whereas the remaining microtubules have atendency to rearrange into an aster-like pattern (Thy-berg and Moskalewski, 1992). A particularly interest-ing role, especially during mitosis, is attributed to thecyclin B1 /p34cdc2 complex. This M phase-promotingfactor mediates chromatin condensation and break-down of the nuclear envelope, and it is targeted tomicrotubules through the interactionwith microtubule-associated protein 4 (Ookata et al., 1995). This sug-gests that proper microtubule functioning is essentialfor progression through mitosis. Diminished stability

of spindle microtubules through aberrant phosphory!a-tion of microtubule-associated proteins might in turndisturb the mitotic forces that control a cell divisioncheckpoint (Li and Nicklas, 1995). For example, it isdocumented that inHeLa cells the activity of a kinesin-related protein is regulated through p34 cdc

2~mediatedphosphorylation (Blangy et al., 1995) and as such isrequired for the formation of normal bipolar spindles.Alternatively, disturbing the phosphorylation—dephos-phorylation equilibrium influences the degradation ofthe cyclin B

1/p34°~2complex, necessary to proceed

properly through the mitotic cycle (Yamashita et al.,1990). Especially in chronic neurodegenerative dis-ease evidence has been provided that neurons diethrough an apoptotic pathway (Su et al., 1994). Inthis regard it is important to note that apoptosis-linkedgene-3 is quite homologous to the familial Alzheimer’ sdisease gene preseni!in-2 (Vito et a!., 1996). In addi-tion, the A!zheimer-specific antibody TG-3, directedagainst a modified form of tau, actually stains a nuclearprotein only in mitotic cells (Vincent et a!., 1996).The mechanisms that trigger these postmitotic neu-ronal cells to undergo apoptosis are not fully under-stood. It has been argued that the apoptotic process inseveral cell types shares some features with the normalmitotic cycle.

The identification of the trigger that initiates G~/S transition may therefore be useful for developingtherapeutic agents that inhibit apoptosis in neuronalcells.

Acknowledgment: We gratefully acknowledge the exper-tise and help of Dr. Marina Cools in the experiments withthe flow cytometer. The discussions with Dr. Bert Schutte(Maastricht University) together with his critical suggestionsare highly appreciated.

REFERENCES

Alonso A., Zaidi T., Grundke-Iqbal I., and Iqbal K. (1994) Role ofabnormally phosphorylated tau in the breakdown of microtu-bules in Alzheimer’ s disease. Proc. Nati. Acad. Sd. USA 91,5562—5566.

Belmont L., Hyman A., Sawin K., and Mitchison T. (1990) Real-time visualisation of cell-cycle dependent changes in microtu-bule dynamics in cytoplasmic extracts. Cell 62, 579—589.

Blangy A., Lane H., d’Herin P., Harper M., Kress M., and Nigg E.(1995) Phosphorylation by p34cdc2 regulates spindle associa-tion of human Eg5, akinesin-related motor essential for bipolarspindle formation in vivo. Cell 83, 1159—1169.

Bradford M. M. (1976) A rapid and sensitive method for the quanti-tation of microgram quantities of protein utilizing the principleof protein—dye binding. Anal. Biochem. 72, 238—245.

Bredesen D. (1995) Neural apoptosis. Ann. Neurol. 38, 839—851.Brooks S. F., Gibson A., and Rubin L. L. (1993)Apoptosis induced

by NGF-withdrawal from differentiated PC12 cells involvesactivation of p34cdc2 kinase. Soc. Neurosci. Abstr. 19, 885.

Darzynkiewicz Z., Gong J., Juan G., Ardelt B., and Traganos F.(1996) Cytometry of cyclin proteins. Cytometry 25, 1—13.

Farinelli S. and Greene L. (1996) Cell cycle blockers mimosine,ciclopirox and deferoxamine prevent the death of PC12 cellsand postmitotic sympathetic neurones after removal of trophicsupport. J. Neurosci. 16, 1150—1162.

Ferrigno P., Langan T., and Cohen P. (1993) Protein phosphatase



2A1 is the major enzyme in vertebrate cell extracts that dephos-phorylates several physiological substrates for cyclin-dependentprotein kinases. Mol. Biol. Cell 4, 669—677.

Freeman R. S., Estus S., and Johnson E. M. Jr. (1994) Analysisof cell cycle-related gene expression in postmitotic neurons:selective induction of cyclin Dl during programmed cell death.Neuron 12, 343—355.

Gao C. and Zelenka P. (1995) Induction of cyclin B and Hl kinaseactivity in apoptotic PC12 cells. Exp. Cell Res. 219, 612—618.

Gliksman N., Parson S., and Salmon E. (1992) Okadaic acid inducesinterphase to mitotic-like microtubule dynamic instability byinactivating rescue. J. Cell Biol. 119, 1271—1276.

Goedert M. and Jakes R. (1990) Expression of separate isoforms ofhuman tau protein: correlation with the tau pattern in brain andeffects on tubulin polymerisation. EMBO J. 9, 4225—4230.

Gorbsky G. (1995) Kinetochores, microtubules and the metaphasecheckpoint. Trends Biol. Sci. 5, 143—147.

Grundke-Iqbal I., Iqbal K., Tung Y., Quinlan M., Wisniewski H.,and Binder L. (1986) Abnormal phosphorylation of the micro-tubule-associated protein tau in Alzheimer cytoskeletal pathol-ogy. Proc. Nati. Acad. Sci. USA 83, 4913—4917.

Inomata M., Saijo N., Kawashima K., Kaneko A., Fujiwara Y.,Kunikane H., and Tanaka Y. (1996) Induction of apoptosis incultured retinoblastoma cells by the protein phosphatase inhibi-tor, okadaic acid. J. Cancer Res. Clin. Oncol. 121, 729—738.

Ishida Y., Furukawa Y., Decaprio J., Saito M., and Griffin J. (1995)Treatment of myeloid leukemic cells with the phosphatase in-hibitor okadaic acid induces cell cycle arrest at either Gl IS orG2/M depending on dose. J. Cell. Physiol. 150, 484—492.

Lassmann H., Bancher C., Breitschopf H., Wegiel J., Bobinski M.,Jellinger K., and Wisniewski H. (1995) Cell death in Alzhei-mer’s disease evaluated by DNA fragmentation in situ. ActaNeuropathol. (Berl.) 89, 35—41.

Li X. and Nicklas R. B. (1995) Mitotic forces control a cell-cyclecheckpoint. Nature 373, 630—632.

Mosmann T. (1983) Rapid colorimetric assay for cellular growth andsurvival: application to proliferation and cytotoxicity assays. J.Immunol. Methods 65, 55—63.

Nuydens R., Heers C., Chadarevian A., De Jong M., Nuyens R.,Cornelissen F., and Geerts H. (l995a) Sodium butyrate inducestau hyperphosphorylation and programmed cell death in humanneuroblastoma. Brain Res. 688, 86—94.

Nuydens R., De Jong M., Nuyens R., Cornelissen F., and GeertsH. (1995b) Neuronal kinase stimulation leads to aberrant tauphosphorylation and neurotoxicity. Neurobiol. Aging 16, 465—477.

Ookata K., Hisanaga S., Bulinski J., Murofushi H., Aizawa H., ItohT., Hotani H., Okumura E., Tachibana K., and Kishimoto T.(1995) Cyclin B interaction with microtubule-associated pro-tein 4 (MAP4) targets p34cdc2 kinase to microtubules and isa potential regulator of M-phasemicrotubule dynamics. J. CellBiol. 128, 849—862.

Park D., Farinelli S., and Greene I. (1996) Inhibitors of cyclin-dependent kinases promote survival of post-mitotic neuronaldifferentiated PC12 cells and sympathetic neurones. J. Biol.Chem. 271, 8161—8169.

Pleasure S., Page C., and Lee V. (1992) Pure postmitotic polarisedhuman neurones derived from Ntera2 cells provide a systemfor expressing exogenous proteins in terminally differentiatedneurones. J. Neurosci. 12, 1802—1815.

Pope W., Lambert M., Leypold B., Seupaul R., Sletten L., KrafftG., and Klein W. (1994) Microtubule-associated protein tau is

hyperphosphorylated during mitosis in the human neuro-blastoma cell line SH-SY5Y. Exp. Neurol. 126, 185—194.

Preuss U., Doring F., Illenberger S., and Mandelkow E. M. (1995)Cell cycle-dependent phosphorylation and microtubule bindingof tau protein stably transfected into Chinese hamster ovarycells. Mol. Biol. Cell 6, 1397—1410.

Rupniak T., Rein G., Powell J., Ryder T., Carson S., Povey S., andHill B. (1984) Characteristics of a new human neuroblastomacell line which differentiates in response to cyclic adenosine3 ‘:5’ monophosphate. Cancer Res. 44, 2600—2607.

Sautiere P. E., Caillet M. L., Wattew A., and Delacourte A. (1994)Detection ofAlzheimer-type tau proteins in okadaic acid-treatedSKNSH-SY 5Y neuroblastoma cells. Neurodegeneration 3,53—60.

Shea T. (1996) Induction of abnormal phosphorylation and tau by-perphosphorylation in human neuroblastoma cells by colchicineand okadaic acid: evidence that microtubule disruption contri-butes to Alzheimer’s disease. Neurosci. Res. Commun. 19, 27—36.

Shea T. and Fischer I. (1996) Phosphatase inhibition in humanneuroblastoma cells alters tao antigenicity and renders it incom-petent to associate with exogenous microtubules. FEBS Lett.380, 63—67.

Shi L., Nishioka W., Th’ng J., Bradbury M., Litchfield D., andGreenberg A. (1994) Premature P34cdc2 activation requiredfor apoptosis. Science 263, 1143—1145.

Shim N., Gotoh Y., and Nishida E. (1995) Microtubule-severingactivity in M-phase. Trends Biol. Sci. 5, 283—286.

Sontag E., Nunbhakdi-Craig V., Bloom G., and Mumby M. (1995)A novel pool of protein phosphatase 2A is associated withmicrotubules and is regulated during the cell cycle. J. Cell Biol.128, 1131—1144.

Su J., Anderson A., Cummings B., and Cotman C. (1994) Immuno-chemical detection for DNA fragmentation in neurones in ADbrain. Neuroreport 5, 2529—2533.

Thyberg J. and Moskalewski 5. (1992) Disorganisation of the Golgicomplex and the cytoplasmic microtubule system in CHO cellsexposed to okadaic acid. J. Cell Sci. 103, 1167—1175.

Vandermeeren M., Lubke U., Six J., and Cras P. (1993) The phos-phatase inhibitor okadaic acid induces a phosphorylated helicalfilament epitope in human LA-N-S neuroblastoma cells. Neu-rosci. Lett. 153, 57—60.

Vandre D. and Wills V. (1992) Inhibition of mitosis by okadaicacid: possible involvement of a protein phosphatase 2A in thetransition from metaphase to anaphase. J. Cell Sd. 101, 79—91.

Vincent I., Rosada M., and Davies P. (1996) Mitotic mechanismsin Alzheimer’s disease? J. Cell Biol. 132, 413—425.

Vito P., Lacana E., and d’Adamio L. (1996) Interfering withapoptosis: Ca-binding protein ALG-3 and Alzheimer’s diseasegene ALG-3. Science 271, 521—524.

Yamashita K., Yasuda H., Pines J., Yasumoto K., Nishitani H.,Ohtsubo M., Hunter T., Sugimura T., and Nishimoto T. (1990)Okadaic acid, a potent inhibitor of type 1 and type 2A proteinphosphatases, activatescdc2IHl kinase and transiently inducesa premature mitosis-like state in BHK21 cells. EMBO J. 9,433 1—4338.

Yano Y., Sakon M., Kambayashi J.-i., Kawasaki T., Senda T., Ta-naka K., Yamada F., and Shibata N. (1995) Cytoskeletal reor-ganisation of human platelets induced by the protein phospha-tase inhibitor okadaic acid and calyculin A. Biochem. J. 307,439—449.


okadaic acid-induced apoptosis in neuronal cells: evidence for an abortive mitotic attempt

Documents