sodium valproate induces apoptosis in the rat hepatoma cell line, fao

Toxicology 192 (2003) 219–227

Sodium valproate induces apoptosis in the rathepatoma cell line, FaO

Anna Phillips, Tabitha Bullock, Nick Plant∗

School of Biomedical and Molecular Sciences, University of Surrey, Guildford, Surrey GU2 7XH, UK

Received 7 February 2003; received in revised form 1 April 2003; accepted 25 July 2003

Abstract

Sodium valproate (VPA) is clinically employed as an anti-convulsant and, to a lesser extent, mood stabiliser. While the incidenceof toxicity associated with the clinical use of valproate is low, serious hepatotoxicity makes up a significant percentage. Ratstreated with high doses of sodium valproate are subject to hepatotoxicity, and the study of the molecular mechanisms underlyingthis phenomenon may shed further light on the human situation.

Exposure to sodium valproate results in the down regulation in rat liver of several transcripts whose products are involvedin cellular energy homeostasis, resulting in time-dependent fluctuations in cellular ATP, possibly resulting in cell death. Tofurther examine this, classical markers of apoptosis were examined in the rat hepatoma cell line FaO following sodium valproateexposure. Concentrations greater than 300�M sodium valproate resulted in a transient wave of apoptosis, as assessed bychromatin condensation and DNA fragmentation assay. Analysis indicated that Fas-ligand and caspase-11 expression wereincreased at the transcriptome level, while caspase-3 was activated at the proteome level during the exposure period. These datademonstrates that sodium valproate causes cell death through apoptosis in a rat liver cell line, and provides information on thepossible molecular mechanisms underlying this phenomenon in vivo.© 2003 Elsevier Ireland Ltd. All rights reserved.

Keywords: Sodium valproate; Hepatotoxicity; Apoptosis; Caspase-11

1. Introduction

Sodium valproate (VPA) is a branched-chain satu-rated fatty acid (2-propylpentenoic acid) that is widelyused in the treatment of epilepsy at plasma concen-trations of 280–690�M (Loscher, 1999; Penry andDean, 1989). The anticonvulsant action of VPA hasbeen demonstrated to be due to a combined phar-

∗ Corresponding author. Tel.:+44-1483-686412;fax: +44-1483-576978.

E-mail address: [email protected] (N. Plant).

macological effect of increased�-aminobutyric acid(GABA) levels (Loscher and Vetter, 1985), attenuationof N-methyl-d-aspartate (NMDA) receptor-mediatedexcitation (Zeise et al., 1991) and decreased repetitivefiring of action potentials (McLean and Macdonald,1986).

In addition to the use of VPA as an anticonvul-sant, it is being increasingly prescribed for a numberof other conditions, including treatment of generalisedand partial seizures, bipolar and schizoaffective disor-ders and even the prophylactic treatment of migraine(Loscher, 1999), although use as an anticonvulsant is

0300-483X/$ – see front matter © 2003 Elsevier Ireland Ltd. All rights reserved.doi:10.1016/S0300-483X(03)00331-7

220 A. Phillips et al. / Toxicology 192 (2003) 219–227

still the main clinical role. In contrast to the well de-scribed mode of action for anticonvulsant treatmentthe mechanisms underlying many of these alternativetherapeutic targets remain largely undefined.

As the therapeutic use of valproate increases, andin particular its use in chronic, and sometimes pro-phylactic, treatment an increasing need is emerging tofully understand the molecular mechanisms of any ad-verse side-effects associated with VPA treatment, thusallowing more accurate risk assessment. Fortunately,toxicity associated with the clinical use of valproate islow, particularly when compared to other anticonvul-sant/mood stabilising drugs. However, of the reportedadverse effects, fatal hepatotoxicity makes up a sig-nificant percentage. Numerous reports detail hepaticfailure following valproate dosing, and in the largeststudy Jeavons reported 60 cases of death due to hep-atic failure within the first 6 months of valproate treat-ment (Jeavons, 1984). The observed hepatotoxicity isidiosyncratic, and presents mainly in young children(<2-year-old) who are on VPA as part of a polyphar-macy treatment schedule (Serrano et al., 1999). Thisselectivity towards young children may either be dueto altered pharmacokinetics/pharmacodynamics in theyoung, or merely reflect that many individuals arefirst exposed to VPA at a young age, when epilepsyfirst manifests itself. Due to the idiosyncratic modeof toxicity seen in man, delineation of the molecularmechanisms has been complicated, although evidencesuggests that disruption of energy homeostasis islikely to play an important role. High dose exposureof rats to VPA results in histologically similar hepa-totoxicity to that seen in humans, and hence has beensuggested as a suitable model. Clear differences doexist between the two systems, with, for example,glutathione depletion playing a more prominent rolein rodent toxicity than in man (Tang et al., 1996).However, study of the high dose rat model will allowfor the formation of hypothesis, which can then betested to see if they extrapolate to the human situation.

Previously, gene hunting technology has been usedto study the changes in gene expression caused by VPAexposure in both rat cerebral cortex (Wang et al., 1999)and liver (Plant et al., 2002) in an attempt to explainboth its therapeutic and adverse effects. In the latterstudy, we have demonstrated that subacute exposurein both rats and the FaO rat liver cell line resulted in atransient decrease in expression of a number of genes

whose products are involved in cellular ATP produc-tion, including aldolase B, succinate dehydrogenaseand enolase; these transcriptome effects were also mir-rored at the level of the proteome.McLaughlin et al.(2000)have previously demonstrated that VPA expo-sure results in perturbation of�-oxidation, potentiallythrough the action of the 2-n-propyl-4-pentanoic acidmetabolite (Bjorge and Baillie, 1985; Kesterson et al.,1984). Taken together, these disruptions in processesinvolved in cellular energetics may be expected to alterthe ATP flux within liver cell, and such was observedto be the case in FaO cells (Plant et al., 2002). Fol-lowing 48 h of exposure, ATP levels were transientlyraised within FaO cells, which we hypothesised maybe a marker for onset of VPA-mediated cell death.

In the present study, we extend these observationsby examining the overall mode of death induced byVPA exposure in FaO cells, as well as the molecularmechanisms that may underlie this phenomenon. Wedemonstrate that cell death occurs via apoptotic mech-anisms at all but the highest concentration tested. Inaddition, by examination of the transcripts for severalgenes associated with apoptotic cell death we beginto delineate the molecular mechanisms which underlieVPA-induced apoptosis.

2. Materials and methods

2.1. Chemicals

Sodium valproate was purchased from Sigma(Poole, UK). All other reagents were of molecular bi-ology grade and purchased from Sigma (Poole, UK)unless otherwise stated.

2.2. Cell culture and dosing

All cell culture media and supplements werepurchased from Gibco BRL (Paisley, UK). FaO cells,derived from a rat hepatoma, were obtained from theEuropean Collection of Animal Cell Cultures (ECACCNo. 89042701, Porton Down, UK). FaO cells werecultured in Hams F10:DMEM (1:1) supplementedwith 10% foetal bovine serum (FBS) and 100�g/mlgentamycin.

Cells were seeded at a density of 0.5× 106 cells/mlin 24-well plates for Q-PCR analysis and 25 cm2 flasks

A. Phillips et al. / Toxicology 192 (2003) 219–227 221

for all other analysis and allowed to attach for 24 h.Following attachment cells were exposed to varyingconcentrations of VPA (0, 100, 200, 300, 400, 500,600 or 700�M) for the required time (24, 48, 72, 96,120 or 144 h) and processed as described below.

2.3. Cell viability analysis

FaO cells were cultured in the presence of varyingconcentrations of VPA as described above for 144 h.Cell viability was then assessed by Trypan blue ex-clusion. The percentage of cells staining blue werecounted in 10 random fields selected from triplicaterepeat flasks, representing approximately 3000 cellsin total.

2.4. DNA fragmentation analysis

FaO cells were cultured in the presence of vary-ing concentrations of VPA as described above for 72 hand a DNA fragmentation assay carried out as pre-viously described by (Ray et al., 1993). Briefly, cellswere washed with 1x PBS, harvested by trypsinisationand lysed on ice for 30 min (10 mM Tris–HCl, pH 8.0,10 mM EDTA, 1% Nonidet P-40). The resulting lysatewas centrifuged at 1500× g for at 4◦C for 5 min andthe supernatant incubated with RNase A (5�g/ml) at37◦C for 1 h, followed by incubation at 50◦C for 2 hwith proteinase K (200�g/ml) and 1% SDS. Follow-ing extraction with phenol/chloroform/isoamyl alco-hol (25:24:1), fragmented DNA was precipitated withpropan-2-ol and resuspended in UHP water. Fragmen-tation products were then separated on 2% Metaphoragarose (Flowgen, Lichfield, UK) and visualised un-der UV illumination with ethidium bromide staining.

2.5. Morphological analysis of apoptotic cells

FaO cells were exposed to varying concentrationsof VPA for the appropriate time and then fixed (75%ethanol, 25% glacial acetic acid) at−20◦C for 70 minand then allowed to air dry.

For staining, fixed cells were first rehydrated withPBS, and then stained with Hoechst 33258 (8�g/mlin UHP water) at room temperature for 10 min on arocking platform. At the end of the staining period,flasks were washed in UHP water, covered in moun-tant (20 mM citric acid pH 5.5, 50 mM Na2HPO4,

50% glycerol) and flasks visualised at 350–460 nm.For quantitation, the percentage apoptotic nuclei wascounted in 10 random fields selected from triplicaterepeat flasks, representing approximately 3000 cellsin total.

2.6. Quantitative RT-PCR analysis

Following exposure of FaO cells to VPA cells wereharvested by the trypsinization and total RNA ex-tracted using the RNeasy® Mini kit (Qiagen, Crawley,UK) according to the manufacturer’s instructions. To-tal RNA quantitation was carried out using RiboGreenRNA Quantitation kit (Molecular Probes).

An amount of 1�g of total RNA was then treatedwith RNase-free DNase I (Promega, Southampton,UK) at 37◦C for 30 min to remove any contaminat-ing genomic DNA, followed by enzyme inactivationat 65◦C for 10 min prior to cDNA synthesis. cDNAwas produced using Superscript II (Invitrogen, Paisley,UK) according to the manufacturer’s protocol. cDNAwas stored at−20◦C prior to quantitative 5′-nucleaseassay (Q-PCR) analysis.

Q-PCR reactions were set up using FAM reporterdye/TAMRA quencher dye labelled probes in con-junction with appropriate primer sets, all designedusing the Primer Express Software and purchasedfrom Applied Biosystems (Warrington, UK). TaqManUniversal PCR Mastermix was purchased from Ap-plied Biosystems, and reactions set up according tothe manufacturer’s instructions, with the exceptionthat all volumes were halved to give a final reactionvolume of 25�l. Q-PCR reactions were carried outusing an ABI7000 SDS instrument and quantitationcarried out using the ABI proprietary software againsta standard curve generated from rat genomic DNA.

2.7. Caspase-3 activity

Caspase-3 activity was measured using theApoAlert fluorescent assay system (BD Biosciences,Oxford, UK), according to the manufacturers in-structions. Briefly, FaO cells were cultured in thepresence of 600�M VPA as described above. Aftervarying periods, cells were collected, lysed, reactionbuffer added and then incubated with the fluorescentcaspase-3 substrate (DEVD-AFC) at 37◦C for 1 h.Following this incubation sample fluorescence was


measured with an excitation wavelength of 400 nmand emission wavelength of 505 nm.

2.8. Statistics

For chromatin condensation quantitation, statisti-cal significance was assessed using one-way ANOVA,followed by Bonferroni–Dunn post-hoc analysis. Sta-tistical comparison of quantitative PCR analysis andcaspase-3 activity assay, treated versus control cells,was analysed by Student’st-test.

3. Results

3.1. VPA induces apoptotic cell death in rat FaO cells

Initially, we examined the effect of VPA on FaOviability and survival. Cells were exposed to VPA for144 h and cell viability assessed by the Trypan blueexclusion test. As can be seen fromFig. 1, while un-treated cells retained approximately 95% viability af-ter 144 h in culture, exposure to concentrations of VPAgreater than 300�M, resulted in a significant decreasein cell viability. This was maximal at 700�M VPA,

0 100 200 300 400 500 600 7000.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

100.00

Sodium valproate (uM)

*

****

**

Fig. 1. Sodium valproate causes a dose-dependent reduction in ratliver FaO cell viability. Rat liver FaO cells were exposed to varyingconcentrations of sodium valproate for 144 and then cell viabilityassessed by Trypan blue exclusion assay, as detailed in materialsand methods. Each data point represents the mean of four inde-pendent determinations, and error bars standard error of the mean.Statistically significant differences, compared to control treated

cells, were determined by Student’st-test, where P < 0.05 and

P < 0.01.

with only 40% of the cells being viable following144 h. Such a decrease in cell viability is indicativeof marked VPA-induced cell death but gives no infor-mation on the mode this cell death manifests as, andthis was examined next.Fig. 2A shows DNA frag-mentation, indicative of apoptosis, occurring in cul-tures treated with >400�M VPA for 72 h, althoughthis fragmentation was much reduced in the top dosegroup, 700�M VPA. Such data suggested that the re-duction in cell viability observed in FaO cells fol-lowing exposure to concentrations >300�M VPA wasapoptotic, although the top dose group was not conclu-sive. To further characterise this apoptotic response,we examined the time course of chromatin conden-sation, using Hoechst 33258 staining and histologicalevaluation. Exposure of cells to 400, 500 and 600�MVPA caused a dose and time dependent increase inchromatin condensation which peaked after 72 h ex-posure to VPA (Fig. 2B). As was observed with theDNA fragmentation assay, this effect was absent inthe 700�M dose group, possibly reflecting a switchto necrotic cell death at the maximal dose used in thisstudy. As the maximal apoptotic response, as deter-mined by DNA fragmentation and chromatin conden-sation, was achieved following exposure of FaO cellsto 600�M VPA subsequent studies were undertakenwith this dose.

Cellular responses to toxic insult are often thenet result of several physiological processes that aremechanistically linked. In the case of apoptosis apotential partner is DNA synthesis, with oppositeregulation of these two processes being seen fol-lowing exposure to a variety of xenobiotics, suchas the peroxisome proliferators (Bayly et al., 1993).p21/Waf1 is a cyclin-dependent kinase inhibitor capa-ble of blocking cell cycle progression. Hence, levelsof this factor are a good marker of cellular replication(Noda et al., 2002). As can be seen fromFig. 3, theexpression level of p21/Waf1 is unchanged followingexposure of FaO cells to 600 mM VPA, suggestingthat in this instance apoptosis occurs independentlyfrom inhibition of cellular division.

3.2. Effect of VPA on the expression of genesassociated with apoptosis

There exist several, interconnected, pathwayswithin cells that may produce an apoptotic response


Fig. 2. Sodium valproate induces apoptosis in rat liver FaO cells. Rat liver FaO cells were exposed to varying concentra-tions of sodium valproate for between 24 and 144 h. Panel (A) shows DNA fragmentation occurring after 72 h of VPA ex-posure, with nucleosomal laddering indicated by arrows, whereas panel (B) shows chromatin condensation assessed by Hoechst33258 staining after different amounts of VPA and time, as described in materials and methods. Each data point represents themean of four independent determinations, and error bars standard error of the mean. Statistically significant differences, com-

pared to control treated cells, were determined by one-way ANOVA, with Bonferroni–Dun post-hoc analysis, whereP < 0.05

and P < 0.01.

following toxic stimuli (Hengartner, 2000). To ex-amine the activation of these we used Q-PCR anal-ysis to measure the transcriptome levels of mRNAspecies whose products have been associated with theevents of apoptosis. Initially, we examined the ex-pression levels of three members of the TNF family,TNF�, Fas and Fas-ligand, which may act to trans-mit external stimuli into cells (Zhou et al., 2002).As can be seen fromFig. 4A, Fas-ligand levels are

induced in a time-dependent manner, with increasedlevels occurring between 48 and 120 h of 600�MVPA exposure. By comparison, expression levels ofTNF� and Fas remain unchanged (Fig. 4A) duringthe exposure period. Next, the expression of two‘death associated proteins’ was examined: Bcl2L8(Bad) is a pro-apototic regulator while Daxx is anadaptor that exerts both apoptotic and anti-apoptoticeffects on cells.Fig. 4B shows that the expression


0 24 48 72 96 120 1440.00

50.00

100.00

150.00

200.00

250.00

Exposure (Hours)

Fig. 3. Sodium valproate does not affect p21/Waf 1 expression inrat liver FaO cells. Rat liver FaO cells were exposed to 600�Msodium valproate for between 24 and 144 h and then total RNAextracted as described in materials and methods. Expression levelsof p21/Waf1 were measured by quantitative PCR and compared torat genomic DNA standards. Each data point represents the meanof four independent determinations.

of both of these ‘death associated proteins’ in FaOcells is unchanged by exposure to 600�M VPAfor up to 144 h. Finally, the levels of transcripts forthree of the major co-ordinators of apoptosis, thecaspases, were examined. Caspase-11 transcript lev-els were shown to be significantly elevated in FaOcells exposed to 600�M VPA for between 24 and120 h (Fig. 4C). In contrast, expression levels ofthe two effector caspases, caspase-3 and caspase-6,in FaO cells were unaffected by VPA exposure(Fig. 4C).

3.3. Effect of VPA on caspase-3 activity

Effector caspase activity, such as caspase-3,can be modulated at both the transcriptional andpost-transcriptional level, and indeed the majoritymay occur at the latter stage. To complement tran-scriptional analysis of caspase-3 mRNA levels, en-zyme activity of caspase-3 was next measured inFaO cells exposed to 600�M VPA for periods up to144 h. As can been seen fromFig. 5, caspase-3 activ-ity shows a dose-dependent increase following VPAexposure. This increase is maximal at 72 h of expo-sure, concurrent with the maximal levels of chromatincondensation observed.

0 24 48 72 96 120 1440.00

50.00

100.00

150.00

200.00

250.00

Exposure (hours)

TNFa

FasL

FasRc

0 24 48 72 96 120 1440.00

50.00

100.00

150.00

200.00

250.00

Exposure (hours)

Bcl2L8

Daxx

**

*

0 24 48 72 96 120 1440.00

50.00

100.00

150.00

200.00

250.00

Exposure (hours)

Caspase 11

Caspase 3

Caspase 6

(A)

(B)

(C)

Fig. 4. FaO cell exposure to sodium valproate changes in the ex-pression of transcripts associated with apoptosis. Rat liver FaOcells were exposed to 600�M sodium valproate for between 24and 144 h and then total RNA extracted as described in materi-als and methods. Expression levels of transcripts associated withapoptosis were then measured by quantitative PCR and comparedto rat genomic DNA standards. (A) Represents Fas, Fas-ligandand TNF�, (B) the death associated proteins Bcl2L8 and Daxxand (C) caspase-3, caspase-6 and caspase-11. Each data point rep-resents the mean of four independent determinations. Statisticallysignificant differences, compared to control treated cells, were de-

termined by Student’st-test, where P < 0.05 and P < 0.01.


0 24 48 72 96 120 1441000

1200

1400

1600

1800

2000

2200

Exposure (hours)

0

600uM

**

**

**

**

Fig. 5. Sodium valproate causes an increase in caspase-3 activityin rat liver FaO cells. Rat liver FaO cells were exposed to 600�Msodium valproate for between 24 and 144 h and then caspase-3activity measured as described in materials and methods. Eachdata point represents the mean of four independent determinations.Statistically significant differences, compared to control treated

cells, were determined by Student’st-test, where P < 0.01.

4. Discussion

Our previous studies suggested that VPA hepato-toxicity in the rat may be due to an imbalance incellular energetics, caused by the down regulationof several transcripts whose products are key reg-ulators of glycolysis, the Krebs cycle and electrontransport chain (Plant et al., 2002). Such a hypoth-esis is also supported by the fact that�-oxidationis also disturbed in rat liver following VPA expo-sure (McLaughlin et al., 2000), primarily throughthe formation of the 2-n-propyl-4-pentenoic acidmetabolite (Kesterson et al., 1984). The importanceof this metabolite in human toxicity has been ques-tioned, with differing pharmacodynamics betweenman and rat leading to reduced production in man.However, even without disruption of�-oxidation, thestress placed on cellular energy homeostasis by thedown regulation of other enzymes involved in cellularenergy homeostasis, such as aldolase B and succi-nate dehydrogenase may still be sufficient to causecellular ATP fluctuations and precipitate cell death.Indeed, in rat FaO cells ATP concentrations weretransiently elevated following exposure to 500 and600�M VPA. Such an imbalance could lead to themitochondrial dysfunction and consequent cell death.

To further investigate such a mechanism, we havenow examined the mode of cell death. Exposure ofFaO cells to concentration of VPA >400�M resultedin both DNA fragmentation and chromatin conden-sation, classical indicators of apoptosis. Hence, itis likely that VPA causes cell death in hepatocytesthrough induction of apoptosis, in a similar fashionto the way it causes B-cell apoptosis (Kawagoe et al.,2002). However, it is interesting to note that at themaximum dose studied, 700�M, neither DNA frag-mentation or chromatin condensation were observed,yet cell viability was reduced to approximately 40%control levels. This would suggest that a thresh-old exists for the cellular damage, beyond whichcells are destroyed via necrotic rather than apoptoticmechanisms.

Following the demonstration that VPA-induces rathepatocyte death through apoptosis, it is pertinent toexamine the underlying molecular mechanisms thatcause this apoptotic response. To examine this ques-tion we have used Q-PCR to measure the transcrip-tome levels of several mRNA species, whose proteinproducts are associated with apoptosis. Measurementat the level of the transcriptome carries the caveat thatchanges at this level may not translate to alteration inprotein activity. Indeed, evidence exists to suggest thattranscriptional activation may not be obligate for ini-tiation of apoptosis, with protein localisation and/orcleavage events being more important (Demeret et al.,2003). However, much evidence exists to show thattranscriptional activation does proceed apoptosis inmany systems (Earnshaw et al., 1999; Hoffmann et al.,2003; MacLachlan and El-Deiry, 2002; Michaelsonand Leder, 2003).

Fas, Fas-ligand and TNF� are all members of thetumour necrosis factor family of proteins, and playa role in cellular responses to stimuli, including theinduction of apoptosis. Fas, a type I membranebound glycoprotein, is often expressed constitu-tively in cells, whereas levels of Fas-ligand, a typeII membrane protein, are often increased at thetranscriptome level by toxic insult (Maher et al.,2002). These two factors may then associate, alongwith FADD to form a complex that signals cleav-age of pro-caspase-8 to active caspase-8 and theinitiation of the caspase cascade that ultimately re-sults in apoptosis. Therefore, constitutive expres-sion of Fas, and induction of Fas-ligand levels as


observed herein are in agreement with VPA stim-ulating caspase-mediated apoptosis. It should benoted however that Fas-ligand mRNA levels reachmaximal levels after the peak of chromatin con-densation is observed. Two explanations exist forsuch an observation. First, changes in Fas-ligandlevels may be independent of, or caused by, theapoptotic response and may play no direct role inits progression. Second, apoptosis progressed viaFas-ligand may cause increases in Fas-ligand ex-pression, leading to an increased potential for re-sponse. Further experiments would be required toconfirm which of these two hypotheses are correct.VPA has previously been shown to prevent induc-tion of TNF� levels in human monocytes, putativelythrough inhibition of NF-�B activation (Ichiyamaet al., 2000). We now demonstrate that TNF� isalso refractory to VPA-mediated induction in rathepatocytes.

Exposure of FaO cells to 600�M VPA also hadno transcriptional effect on two of the so-called‘death associated proteins’; Bcl2L8 (Bad) and Daxx.Bcl2L8 is a pro-apoptotic signal (Hengartner, 2000)whereas Daxx is an adapter protein that has beenassociated with both apoptotic (Wu et al., 2002) andanti-apoptotic events (Michaelson and Leder, 2003),and hence, it might be expected to see changes inexpression of one or both of these transcripts dur-ing apoptosis. However, while the products of thesetranscripts are associated with modulating apopto-sis events, it is probably their localisation that iskey rather than their levels. Many proteins foundin the mitochondria, including several ‘death asso-ciated proteins’, leak out into the cytosol duringthe initiation of apoptosis and thereby elicit theireffects (Hengartner, 2000). Thus, our demonstra-tion that VPA treatment does not cause a transcrip-tional response in these two genes does not pos-itively exclude them from a role in VPA-inducedapoptosis.

Finally, we examined the transcript levels of threecaspases in FaO cells during VPA exposure. Asdiscussed above, increases in Fas-ligand expressionsuggest that apoptosis within hepatocytes is cas-pase dependent, and such a hypothesis is furtherstrengthened by the increased transcript levels ofcaspase-11 observed herein. Caspase-11 has beendemonstrated to play an obligate role in the activation

of effector caspases such as caspase-3 (Kang et al.,2000) and caspase-1 (Wang et al., 1998), and hencemay represent a key regulator in caspase-dependentapoptosis.

Interestingly, levels of caspase-3 and caspase-6were not altered in FaO cells by VPA exposure, re-maining at control levels throughout the exposureperiod. Two possible explanations may exist for thisapparent contradiction. First, activation of caspase-3or caspase-6 may occur solely at the proteome level inthis system, as opposed to the transcriptome, throughthe cleavage of zymogen molecules to form active cas-pase molecules. Alternatively, activation of caspase-11may result in the activation of other caspases, such ascaspase-1, and apoptosis may therefore be essentiallyindependent of caspase-3 or caspase-6. Measurementof caspase-3 activity by cleavage of the fluorescentsubstrate DEVD-AFC shows that control exists atthe proteome level, as caspase-3 activity increasesupon exposure to VPA. The increase in caspase-3activity closely mirrors the observed chromatin con-densation, with a maximal activity being achievedafter 72 h of exposure. While such data strongly sug-gests that VPA-mediated apoptosis of FaO cells iscaspase-3-dependent it should be noted thatKawagoeet al. (2002)report that VPA may induce apoptosisin human B-cells through both caspase-dependentand caspase-independent mechanisms. These authorsdemonstrated that MV411 cells exposed to VPAunderwent apoptosis, as measured by DNA fragmen-tation, cytochromec release and phosphatidylserineexternalisation, and that this was accompanied byactivation of caspase-3, caspase-8 and caspase-9 incells. However, inclusion of zVAD-FMK, a caspaseinhibitor, in the incubations prevented caspase ac-tivation but not apoptosis. Hence, it is clear thatVPA-induced apoptosis probably occurs via multiple,complementary pathways.

In summary, we have shown that VPA causes cel-lular death in the rat liver FaO cell line via apoptosisin a time- and dose-dependent manner, and that thisevent corresponds to the previously identified peak incellular ATP levels. Expression of caspase-11 was in-creased at the transcriptome level and caspase-3 at theenzyme activity level, suggesting a role of caspasesin VPA-mediated apoptosis in FaO cells. Such infor-mation provides further avenues of research for theidiosyncratic, low dose toxicity, observed in humans.


References

Bayly, A.C., et al., 1993. Non-genotoxic hepatocarcingenesisin vitro: the FaO hepatoma line responds to peroxisomeproliferators and retains the ability to undergo apoptosis. J.Cell Sci. 104, 307–315.

Bjorge, S.M., Baillie, T.A., 1985. Inhibition of medium-chain fattyacid beta-oxidation in vitro by valproic acid and its unsaturatedmetabolite, 2-n-propyl-4-pentenoic acid. Biochem. Biophys.Res. Commun. 132 (1), 245–252.

Demeret, C., et al., 2003. Transcription-independent triggering ofthe extrinsic pathway of apoptosis by human papillomavirus18 E2 protein. Oncogene 22 (2), 168–175.

Earnshaw, W.C., et al., 1999. Mammalian caspases: structure,activation, substrates, and functions during apoptosis. Annu.Rev. Biochem. 68, 383–424.

Hengartner, M.O., 2000. The biochemistry of apoptosis. Nature407 (6805), 770–776.

Hoffmann, A., et al., 2003. Transcriptional activities of the zincfinger protein zac are differentially controlled by DNA binding.Mol. Cell. Biol. 23 (3), 988–1003.

Ichiyama, T., et al., 2000. Sodium valproate inhibits production ofTNF-alpha and IL-6 and activation of NF-kappaB. Brain Res.857 (1–2), 246–251.

Jeavons, P., 1984. Non-dose related side effects of valproate.Epilepsia 25 (S1), S50–S55.

Kang, S.J., et al., 2000. Dual role of caspase-11 in mediatingactivation of caspase-1 and caspase-3 under pathologicalconditions. J. Cell Biol. 149.3, 613–622.

Kawagoe, R., et al., 2002. Valproic acid induces apoptosis inhuman leukemia cells by stimulating both caspase-dependentand caspase-independent apoptotic signaling pathways. Leuk.Res. 26 (5), 495–502.

Kesterson, J.W., et al., 1984. The hepatotoxicity of valproic acidand its metabolites in rats. I. Toxicologic, biochemical andhistopathologic studies. Hepatology 4 (6), 1143–1152.

Loscher, W., 1999. Valproate: a reappraisal of its pharmacodynamicproperties and mechanisms of action. Prog. Neurobiol. 58 (1),31–59.

Loscher, W., Vetter, M., 1985. In vivo effects of aminooxyaceticacid and valproic acid on nerve terminal (synaptosomal) GABAlevels in discrete brain areas of the rat. Biochem. Pharmacol.34, 1747–1756.

MacLachlan, T.K., El-Deiry, W.S., 2002. Apoptotic threshold islowered by p53 transactivation of caspase-6. Proc. Natl. Acad.Sci. U.S.A. 99 (14), 9492–9497.

Maher, S., et al., 2002. Activation-induced cell death: thecontroversial role of Fas and Fas-ligand in immune privilegeand tumour counterattack. Immunol. Cell. Biol. 80 (2), 131–137.

McLaughlin, D.B., et al., 2000. Apparent autoinduction ofvalproate�-oxidation in humans. Br. J. Clin. Pharmacol. 49 (5),409–415.

McLean, M., Macdonald, R., 1986. Sodium valproate, but notethosuximide, produces use- and voltage-dependent limitationof high frequency repetitive firing of action potentials of mousecentral neurons in cell culture. J. Pharmacol. Exp. Ther. 237,1001–1011.

Michaelson, J.S., Leder, P., 2003. RNAi reveals anti-apoptoticand transcriptionally repressive activities of Daxx. J. Cell Sci.116 (Pt 2), 345–352.

Noda, H., et al., 2002. Increased proliferative activity caused byloss of p21(WAF1/CIP1) expression and its clinical significancein patients with early-stage gastric carcinoma. Cancer 94 (7),2107–2112.

Penry, J.K., Dean, J.C., 1989. The scope and use of valproate inepilepsy. J. Clin. Psychiatry 50, 17–22.

Plant, N., et al., 2002. Differential gene expression in rats followingsubacute exposure to the anticonvulsant sodium valproate.Toxicol. Appl. Pharmacol. 183 (2), 127–134.

Ray, S.D., et al., 1993. Ca2+ antagonists inhibit DNAfragmentation and toxic cell death induced by acetaminophen.FASEB J. 7 (5), 453–463.

Serrano, B.B., et al., 1999. Valproate population pharmacokineticsin children. J. Clin. Pharmacol. Ther. 24 (1), 73–80.

Tang, W., et al., 1996. Conjugation of glutathione with atoxic metabolite of valproic acid, (E)-2-propyl-2,4-pentadienoicacid, catalyzed by rat hepatic glutathione-S-transferases. DrugMetab. Dispos. 24 (4), 436–446.

Wang, S., et al., 1998. Murine caspase-11, an ICE-interactingprotease, is essential for the activation of ICE. Cell 92 (4),501–509.

Wang, J.-F., et al., 1999. Differential display PCR reveals noveltargets for the mood-stabilizing drug valproate including themolecular chaperone GRP78. Mol. Pharmacol. 55, 521–527.

Wu, S., et al., 2002. Ultraviolet radiation-induced apoptosis ismediated by Daxx. Neoplasia 4 (6), 486–492.

Zeise, M.L., et al., 1991. Valproate suppressesN-methyl-d-asparateevoked, transient depolarizations in the rat neocortex in vitro.Brain Res. 544, 345–348.

Zhou, T., et al., 2002. Immunobiology of tumor necrosis factorreceptor superfamily. Immunol. Res. 26 (1–3), 323–336.

sodium valproate induces apoptosis in the rat hepatoma cell line, fao

Documents