Epigenetic modifications in valproic acid-induced teratogenesis

Toxicology and Applied Pharmacology 248 (2010) 201–209

Contents lists available at ScienceDirect

Toxicology and Applied Pharmacology

j ourna l homepage: www.e lsev ie r.com/ locate /ytaap

Epigenetic modifications in valproic acid-induced teratogenesis☆

Emily W.Y. Tung a, Louise M. Winn a,b,⁎a Department of Pharmacology and Toxicology, Queen's University, Kingston, Ontario, Canada K7L 3N6b School of Environmental Studies, Queen's University, Kingston, Ontario, Canada K7L 3N6

Abbreviations: AcH3, acetylated histone H3; AcH4disintegrations per minute; GD, gestational day; H3K4histone H3 lysine 9; NTD, neural tube defect; VPA, valp☆ Preliminary results from these studies have been prAnnual Symposium of the Society of Toxicology of CanAnnual Meeting of the Society of Toxicology (U.S.A.), MaMeeting of the Teratology Society, June 2009.⁎ Corresponding author. Room 557 Botterell Hall,

Ontario, Canada K7L 3N6. Fax: +1 613 533 6412.E-mail address: winnl@queensu.ca (L.M. Winn).

0041-008X/$ – see front matter © 2010 Elsevier Inc. Adoi:10.1016/j.taap.2010.08.001

a b s t r a c t

Article history:Received 15 June 2010Revised 30 July 2010Accepted 3 August 2010Available online 10 August 2010

Keywords:Valproic acidHistone deacetylaseMethylationEmbryo

Exposure to the anticonvulsant drug valproic acid (VPA) in utero is associated with a 1–2% increase in neuraltube defects (NTDs), however the molecular mechanisms by which VPA induces teratogenesis are unknown.Previous studies demonstrated that VPA, a direct inhibitor of histone deacetylase, can induce histonehyperacetylation and other epigenetic changes such as histone methylation and DNA demethylation. Theobjective of this study was to determine if maternal exposure to VPA in mice has the ability to cause theseepigenetic alterations in the embryo and thus contribute to its mechanism of teratogenesis. Pregnant CD-1mice (GD 9.0) were administered a teratogenic dose of VPA (400 mg/kg, s.c.) and embryos extracted 1, 3, 6,and 24 h after injection. To assess embryonic histone acetylation and histone methylation, Western blottingwas performed on whole embryo homogenates, as well as immunohistochemical staining on embryonicsections. To measure DNA methylation changes, the cytosine extension assay was performed. Resultsdemonstrated that a significant increase in histone acetylation that peaked 3 h after VPA exposure wasaccompanied by an increase in histone methylation at histone H3 lysine 4 (H3K4) and a decrease in histonemethylation at histone H3 lysine 9 (H3K9). Immunohistochemical staining revealed increased histoneacetylation in the neuroepithelium, heart, and somites. A decrease in methylated histone H3K9 staining wasobserved in the neuroepithelium and somites, METHYLATED histone H3K4 staining was observed in theneuroepithelium. No significant differences in global or CpG island DNA methylation were observed inembryo homogenates. These results support the possibility that epigenetic modifications caused by VPAduring early mouse organogenesis results in congenital malformations.

, acetylated histone H4; dpm,, histone H3 lysine 4; H3K9,roic acid.eviously presented at the 40thada, December 2008; the 48thrch 2009; and the 49th Annual

Queen's University, Kingston,

ll rights reserved.

© 2010 Elsevier Inc. All rights reserved.

Introduction

Valproic acid (VPA) is a widely prescribed broad spectrumantiepileptic agent that is also used in the management ofmigraines and bipolar disorders (Bowden, 2009; Evers, 2008).Unfortunately, in utero exposure to VPA during the first trimesterof pregnancy results in an increased risk of major and minorcongenital malformations (Ornoy, 2009). Described by the term“fetal valproate syndrome,” maternal use of VPA has beenassociated with congenital heart defects, craniofacial abnormalities,skeletal and limb defects, and functional and cognitive effects inchildren (Clayton-Smith et al., 1995). The developing nervoussystem is particularly sensitive to VPA. The rate of neural tube

defects (NTDs) in infants exposed to VPA in utero is 10–20 timesthat of the general population (Ornoy, 2009). Although VPA hasbeen known to be teratogenic for almost 30 years, the precisemolecular mechanisms by which VPA exerts its teratogenicity arenot yet fully elucidated.

VPA is a direct inhibitor of class I and II histone deacetylases andprevious studies have suggested that VPA-induced histone hyperace-tylationmay be a possiblemechanism ofmediating teratogenesis (Phielet al., 2001). Exposure of Xenopus and zebrafish embryos to VPA duringthe early gastrula stage resulted in developmental anomalies such asgrowth retardation, defects in gut coiling, and eye, heart and tail defects(Gurvich et al., 2005). In the same study, analogs of VPA lacking histonedeacetylase inhibitory activity were shown to be less teratogenic.Studies have also shown that in utero exposure to VPA results in anincrease in the expression of acetylated histones in somites ofdeveloping mouse embryos, which correlates with increased skeletalabnormalities observed at term (Menegola et al., 2005).

Post-translational histone modifications play an important rolein regulating gene expression by controlling accessibility of DNA totranscription factors. Acetylation of lysine residues on histone tailsreduces their overall positive charge, thus weakening the interac-tion between DNA and histone proteins and consequently facil-itates the binding of transcription factors to DNA (Eberharter and

202 E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

Becker, 2002). In addition, transcription activation and histoneacetylation are correlated with the methylation of histone H3 atlysine 4 (H3K4), whereas methylation of histone H3 at lysine 9(H3K9) is associated with gene silencing (Berger, 2007). Interest-ingly, recent studies have demonstrated that histone deacetylaseinhibitors including VPA can trigger increases in H3K4 methylationin vitro (Nightingale et al., 2007). Similar increases in histoneacetylation and H3K4 methylation were observed at specificpromoter regions in morula-stage embryos after exposure to VPA(VerMilyea et al., 2009).

In addition to these findings, several studies have found thatVPA can induce the expression of genes that are regulated by DNAmethylation. For example, VPA was shown to induce the expres-sion of reelin, a gene that is methylated, in neuronal precursor cells(Mitchell et al.,), as well as prevent methionine-induced hyper-methylation of reelin and GAD67 promoters in a murine model ofschizophrenia (Dong et al., 2007). Furthermore, in vitro studiesdemonstrated that VPA was able to induce active DNA demethyla-tion independent of replication, suggesting the presence andinvolvement of DNA demethylases (Detich et al., 2003). Subse-quent studies by the same investigators indicated that VPA-induced hyperacetylation and DNA demethylation were genespecific (Milutinovic et al., 2007).

Together, the evidence presented above suggests that VPA canmodulate gene expression through epigenetic alterations such ashistone acetylation, histone methylation, and DNA methylation.The purpose of this study was to examine these epigeneticmodifications in the developing mouse embryo after maternalexposure to VPA during a critical developmental time period. Inaddition, exencephalic embryos were examined separately todetermine if embryos affected by a NTD display differences inthese epigenetic markers. Our data demonstrate that earlyepigenetic changes occur in the embryo, further supporting a rolefor histone acetylation and histone methylation in VPA-initiatedteratogenesis.

Materials and methods

Experimental animals. Virgin female CD-1 mice (Charles RiverLaboratories Inc., St. Constant, QC, Canada) were purchased at 4 to6 weeks of age and were maintained in a temperature-controlledroom with a 12-h light/dark cycle. Standard rodent chow (PurinaRodent Chow, Ralston Purina International, Strathroy, ON, Canada)and tap water were provided ad libitum. Mice were acclimated for aweek and then bred by housing three females with one maleovernight. The presence of a vaginal plug the following morningwas designated as gestational day (GD) 1, and these females wereseparated and housed together. All practices were in accordancewith the guidelines of the Canadian Council on Animal Care andexperimental procedures were approved by the Queen's UniversityAnimal Care Committee.

Animal treatment. On the morning of GD 9, dams were injectedsubcutaneously with 400 mg/kg of valproic acid (Sigma-AldrichCanada Ltd., Oakville, Canada) or the vehicle control (0.9% saline).This dose was selected as previous studies from our laboratorydemonstrated a significant induction of exencephaly in the CD-1mouse strain (Dawson et al., 2006). Dams were sacrificed 1, 3, 6,and 24 h after injection and embryos were collected for subsequentanalyses. Embryos collected at the 24-h time point were separatedaccording to cranial neural tube closure status.

Histone extraction and Western blotting. Litters of whole embryoswere pooled in order to obtain a sufficient amount of histones forWestern blotting. The number of litters pooled together was: 3litters for the 1-h time point, 3 litters for the 3-h time point, and 3

litters for the 6-h time point, with each pool containing 30–46embryos. In addition, 4 embryos from 1 litter were pooled for the24-h time point. Therefore, the n value is represented by a pool ofembryos, where at least 3 pools were used at each time point. Forthe 24-h time point, embryos were assessed for neural tube closurestatus, as the critical time period for neural tube closure occursbetween GD 9-GD 10 in mice (Dawson et al., 2006), and were thenseparated according to normal (closed neural tube) or exe-ncephalic (opened neural tube) groups. Embryos were homoge-nized on ice in triton extraction buffer [PBS containing 0.5% TritonX (v/v), 2 mM phenylmethylsulfonyl fluoride, 0.02% sodium azide(w/v)], washed, and resuspended in 0.2 N HCl overnight. Proteinconcentrations of the supernatants were quantified using theBioRad assay.

Histones (5 μg) were separated by SDS-PAGE on 15% polyacryl-amide gels followed by transfer to PVDF membranes. Membraneswere blocked in 5% bovine serum albumin, then incubated witheither anti-acetylated histone H3 (dilution 1:5000) (UpstateBiotechnology, Billerica, United States), anti-acetylated histone H4(dilution 1:5000) (Upstate Biotechnology, Billerica, United States),anti-di- and trimethyl histone H3K4 (dilution 1:500) (Abcam,Cambridge, United States) or anti-monomethyl histone H3K9antibodies overnight (dilution 1:1000) (Abcam, Cambridge, UnitedStates). Membranes were incubated with appropriate secondaryantibodies and visualized using an enhanced chemiluminescencekit (Perkin Elmer, Boston, United States). To control for differencesin protein loading, membranes were stripped and re-probed foreither total histone H3 or histone H4 (dilution 1:5000) (UpstateBiotechnology, Billerica, United States). ImageJ software (NIH,Bethesda, MD) was used to measure the relative optical densitiesof the bands.

Cytosine extension assay. The cytosine extension assay was per-formed on embryonic DNA as previously described (Pogribny et al.,1999). Briefly, genomic DNA was isolated from embryonic tissuesusing the Qiagen DNeasy™ kit. Embryos were pooled from severallitters in order to obtain DNA to perform the assay. The number oflitters pooled together was: 3 litters for the 1-h time point, 1 litter forthe 3-h time point, 1 litter for the 6-h time point, and 2 embryos fromone litter for the 24-h time point, with embryos being categorized byneural tube closure status in the VPA treated group. Again, each poolwas considered to be an n of 1, with the number of embryos rangingfrom 10 to 15 per litter. For positive and negative controls, 5-aza-2-deoxycytidine treated Jurkat genomic DNA and CpG methylatedJurkat genomic DNA (New England Biolabs, Pickering, Canada) wereused respectively to establish the assay in our laboratory (data notshown). Aliquots (0.5 μg) of DNA were digested with the methylationsensitive restriction endonucleases HpaII and BssHII (10 U perreaction). A third aliquot of undigested DNA served as the backgroundcontrol. A single nucleotide extension was performed in a 30 μLreaction containing 0.5 μg of DNA, 0.75 U of Taq DNA polymerase, 1×PCR buffer, 1 mM MgCl2, and 0.15 μL of [3H]-dCTP (57.4 Ci per mmol)and incubated at 56 °C for 1 h. Samples were applied toWhatman DE-81 ion exchange filters and washed 5 times with 0.5 M sodiumphosphate buffer. The filters were then dried and processed for liquidscintillation counting. The incorporation of [3H]dCTP is directlyproportional to the number of unmethylated CpG sites in the digestedsamples. Disintegrations per minute (dpm) incorporated in theundigested samples was subtracted from the dpm values obtainedfrom digested samples.

Immunohistochemistry. Embryoswere collected from at least 8 littersin each treatment group and fixed in 10% neutral buffered formalin.Embryos were paraffin embedded, sectioned at 5 μm, and mounted onglass slides. Sections were deparaffinised, cleared, and rehydrated. Thesectionswere then treatedwith0.1 Msodiumcitrate at95 °C for antigen

203E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

retrieval, washed in PBS, and blocked for endogenous peroxidaseactivity with 10% hydrogen peroxide in water. Slides were probed withthe same antibodies used for Western blotting and stained using theanti-rabbit Vectastain Elite ABC kit or M.O.M. immunodetection kit(Vector Laboratories) with 3,3’-diaminobenzidine as the substratefollowed by counterstaining with hematoxylin.

Statistical analysis. Results were analyzed using a one sample t-testto compare rates of exencephaly in embryos between vehicle or VPA-treated dams. Statistical significance for Western blotting or cytosineextension assay was assessed using a Student's t-test or one wayanalysis of variance, followed by the Neumann–Keuls multiplecomparison test for post-hoc analysis (GraphPad Prism 4.0, GraphPadSoftware, San Diego, CA). pb0.05 was considered significant.

Results

VPA-induced exencephaly

Consistent with previous studies (Dawson et al., 2006), maternalexposure to VPA resulted in a significant increase in the percentage ofembryos with exencephaly (35±8.4%) when compared to controlembryos (Fig. 1). All control embryos displayed proper neural tubeclosure on GD 10.5.

VPA-induced acetylation of histone H3

To determine a time course of changes in histone acetylation inembryos aftermaternal exposure to VPA,Western blotting analysis wasperformed. In utero exposure to VPA resulted in a significant increase inthe expression of acetylated histone H3 at 1 h after treatment (171±4.7% compared to control), 3 h after treatment (261±64% compared tocontrol), and 6 h after treatment (159±19% compared to control)(Fig. 2A). No significant changeswere observed 24 h after VPA exposurebetween control, VPA closed neural tube, and VPA opened neural tubegroups. Immunohistochemical staining performed on embryos 3 h afterVPA exposure showed increased staining for acetylated histone H3particularly in the neuroepithelium and mesenchyme, as well as someincreased staining in the heart and somites (Fig. 2B–G).

VPA-induced acetylation of histone H4

After maternal injection of VPA, expression of acetylated histoneH4 was significantly increased in embryos 1 h (248±42% comparedto control) and 3 h (429±49% compared to control) after treatment(Fig. 3A). Expression of acetylated histone H4 was not significantlychanged 6 and 24 h after VPA exposure, regardless of neural tubeclosure status. Increased staining for acetylated histone H4 wasobserved in embryo sections 3 h after VPA administration, particularlyin the neuroepithelium and heart (Fig. 3B–G).

control VPA 400 mg/kg0

5% e

xenc

epha

ly (±

SE

M)

10

15

20

25

30

35

40

45 *

Fig. 1. VPA-induced exencephaly. The percentage of embryos with exencephaly on GD10.5 was expressed per litter; n=10 for the saline treatment group, and n=8 for theVPA treated group. * denotes significance from controls (pb0.005).

VPA-induced di- and trimethylation of histone H3K4

A statistically significant increase in the expression of di- andtrimethylated histone H3K4was observed 3 h (392±117%) after drugexposure (Fig. 4A). No significant changes were observed at any othertime point. Immunohistochemical staining revealed increased stain-ing for di- and trimethylated histone H3K4 in the neuroepithelialregion of embryonic heads (Fig. 4B–G).

VPA-induced alterations in monomethylation of histone H3K9

A significant decrease in the expression of monomethyl histoneH3K9 was observed 3 h (37±12% compared to control) aftermaternal injection with VPA (Fig. 5A). No significant changes wereobserved at any other time point. A decrease in monomethyl histoneH3K9 staining was seen in the neuroepithelium as well as in thesomites in embryonic sections following immunohistochemicalstaining performed 3 h following VPA exposure (Fig. 5B–G).

Effect of VPA on global DNA methylation and CpG island methylation

To assess the effects of VPA on global DNA methylation, thecytosine extension assay with HpaII was performed. VPA did notinduce any significant changes in 3[H]dCTP incorporation at 1, 3, 6,and 24 h after VPA exposure (Fig. 6). To assess the effects of VPA onCpG island methylation, the cytosine extension assay with BssHII wasperformed. VPA did not induce any significant changes in 3[H]dCTPincorporation at 1, 3, 6, and 24 h after VPA exposure, regardless of thepresence of a NTD at the 24-h time point (Fig. 7).

Discussion

The use of VPA during pregnancy is associated with a 3-foldincrease in major congenital malformations when compared to thegeneral population. In addition, increased cognitive deficits inchildren who were exposed to VPA in utero have also been noted(Koren et al., 2006). Although the teratogenicity of VPA was firstnoted almost 30 years ago (Bjerkedal et al., 1982), the mechanisms bywhich this drug induces congenital anomalies are not yet fullyunderstood. In the present study, we examine the ability of VPA tomodulate epigenetic markers as a mechanism of teratogenesisthrough which consequent changes in gene expression may giverise to a birth defect. A teratogenic dose of VPA was administeredduring a key period of susceptibility (i.e. organogenesis) in order toexamine epigenetic changes during formation of the heart, somites,and neural tube, as VPA targets these organs. Specifically in regards toneural tube closure, precise expression of genes controlling the cellcycle and cell viability are required in order for the neural tube to closewhereby disturbances during this process can lead to a defect asobserved by transient exposure of xenobiotics through epigeneticmechanisms (Liu et al., 2009; Copp et al., 2010). Our resultsdemonstrate that VPA exposure during pregnancy causes a transientincrease in histone acetylation that correlate with changes in histonemethylation, suggesting that VPA can induce epigenetic alterations bydirectly inhibiting histone deacetylase.

In rodent models, several studies have demonstrated the inductionof exencephaly (Dawson et al., 2006), cardiac abnormalities (Wu et al.,2010), and skeletal malformations (Menegola et al., 2005) followingadministration of VPA at a dose of 400 mg/kg. Results from our presentstudy show similar rates of exencephalic embryoswith this dose of VPAuponmorphological examination onGD10.5, which are consistentwithpublished studies indicating that this single dose of VPA is sufficient toinduce NTDs. Also consistent with previous findings are our resultsshowing a significant increase in the acetylation of histone H4 1 h afterVPA exposure (Menegola et al., 2005). In addition, we are the first todemonstrate a significant increase in the expression of acetylated

Fig. 2. VPA-induced acetylation of histone H3. (A) Acetylation of histone H3 was assessed by Western blotting for AcH3 at 1, 3, 6, and 24 h after maternal VPA treatment.n=3 for the Western blotting. * significantly different from control (pb0.05). (B–G) Localization of AcH3 in embryonic sections. (B, D, F) show control embryonic head(200×) (↑ indicates neuroepithelium, Δ indicates cranial mesenchyme), heart (100×), and somites (200×) (↑ indicates somite tissue), respectively. (C,E,G) show VPAtreated embryos 3 h after VPA administration. C shows embryonic head (200×) (↑ indicates neuroepithelium, Δ indicates cranial mesenchyme), E shows embryonic heart(100×), and G shows somites (200×) (↑ indicates somite tissue).

204 E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

Fig. 3. VPA-induced acetylation of histone H4. (A) Western blotting was used to assess acetylation of histone H4 at 1, 3, 6, and 24 h after maternal VPA treatment. n=3 for theWestern blotting. * significantly different from control (pb0.05). (B–G) Localization of AcH4 in embryonic sections. (B, D, F) show control embryonic head (200×) (↑ indicatesneuroepithelium, Δ indicates cranial mesenchyme), heart (100×), and somites (200×), respectively. (C, E, G) show VPA treated embryos 3 h after VPA administration. C showsembryonic head (200×) (↑ indicates neuroepithelium, Δ indicates cranial mesenchyme), E shows embryonic heart (100×), and G shows somites (200×).

205E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

Fig. 4. VPA-induced di- and trimethylation of histone H3K4. (A) Methylation of histone H3K4 was assessed by Western blotting for di- and trimethylated H3K4 at 1, 3, 6, and 24 hafter maternal VPA treatment. n=3 for theWestern blotting. * significantly different from control (pb0.05). (B–G) Localization of di- and trimethylated H3K4 in embryonic sections.(B, D, F) show control embryonic head (200×) (↑ indicates neuroepithelium, Δ indicates cranial mesenchyme), heart (100×), and somites (200×), respectively. (C, E, G) showembryos 3 h after VPA administration, where C shows embryonic head (200×) (↑ indicates neuroepithelium,Δ indicates cranial mesenchyme), E shows embryonic heart (100×), andG shows somites (200×).

206 E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

Fig. 5. VPA-induced alterations in monomethylation of histone H3K9. (A) Methylation of histone H3K9 was assessed by Western blotting for monomethylated H3K9 at 1, 3, 6, and24 h after maternal VPA treatment. n=3 for the Western blotting * significantly different from control (pb0.05). (B–G) Immunohistochemical staining performed on embryonicsections with anti-monomethyl H3K9. (B, D, F) show control embryonic head (200×) (↑ indicates neuroepithelium, Δ indicates cranial mesenchyme), heart (100×), and somites(200×), respectively. (C, E, G) VPA treated embryos 3 h after VPA administration, where C shows embryonic head (200×) (↑ indicates neuroepithelium, Δ indicates cranialmesenchyme), E shows embryonic heart (100×), and G shows somites (200×).

207E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

Fig. 6. Effect of VPA on global DNA methylation. Global DNA methylation wasdetermined 1, 3, 6 and 24 h after maternal treatment with VPA by DNA digestion withHpaII followed by a single nucleotide extension with 3[H]dCTP; n=3.

208 E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

histoneH3, and that the acetylation of both histonesH3 andH4peak 3 hafter VPA administration. Since these changes do not persist and are notsignificantly different between normal and exencephalic embryos onGD 10, our results suggest that early changes in histone acetylation areresponsible for causing teratogenesis.

Alterations in histone acetylation during organogenesis haveimportant implications for VPA-induced teratogenesis. Acetylatedhistone tails allow transcription factors to access DNA by aconformational change to permissive or open chromatin (Eberhar-arter and Becker, 2002). In addition, acetylated histones arerecognized by other proteins that allow for recruitment of ATP-dependent chromatin remodelling complexes which consequentlymay lead to gene activation (Grant et al., 1999). VPA can induce awide range of gene expression changes in the mouse embryofollowing maternal injection of the drug, including alterations ingenes coding for structural and heat shock proteins, as well astranscription factors and regulators of translation (Kultima et al.,2004). Specifically in the head of the mouse embryo, genesregulating cell cycle arrest and apoptosis were disrupted followingVPA exposure, suggesting that exencephaly could be caused byincreased cell death (Okada et al., 2005). Interestingly, weobserved an increase in acetylated histone H3 and H4 staining inthe cranial mesenchyme and neuroepithelium in VPA-treatedembryos. As histone deacetylase activity is required duringembryonic development to maintain cell proliferation (Lagger etal., 2002), histone hyperacetylation in the cranial neural tube andsurrounding tissues may result in cell cycle arrest, cell death, and

Fig. 7. Effect of VPA on CpG islandmethylation. CpG islandmethylation was determined1, 3, 6 and 24 h after maternal treatment with VPA by DNA digestion with BssHIIfollowed by a single nucleotide extension with 3[H]dCTP; n=4.

ultimately a NTD. Regarding cardiac anomalies, VPA administrationresulted in a significant increase in the number of majormalformations seen in the cardiac tissues of fetuses, and inhibitionof histone deacetylase was observed in cultured cardiomyocytes(Wu et al., 2010). Our results support a role for histone deacetylaseinhibition in VPA-induced heart malformations, as increasedstaining for acetylated histone H3 and H4 was observed inembryonic hearts.

Acetylation of histone tails is one of several covalent modifica-tions involved in the regulation of transcription, where onemodification can trigger subsequent modifications, contributing tothe overall control of gene expression (Zhang et al., 2001).Methylation of lysine residues was traditionally considered to bea stable epigenetic marker, however the discovery of histonedemethylases in recent years suggest histone methylation can betransient (Shi et al., 2004; Whetstine et al., 2006). The con-sequences of histone methylation on transcription are dependenton the specific residue and degree of methylation (Berger, 2007).Methylation of H3K9 is associated with transcriptional silencing,whereas methylation of H3K4 is correlated with gene transcription(Berger, 2007). Specifically in mammals, trimethyl histone H3K9 isfound predominantly in pericentric heterochromatin, whereasmono- and dimethyl histone H3K9 are found to silence specificsites within euchromatin (Mehedint et al., 2010). Our data fromthis study are particularly interesting as changes in histonemethylation were detected in response to the administration of ahistone deacetylase inhibitor, supporting recent findings thathistone deacetylase inhibition can alter the epigenetic landscape(Gupta et al., 2010; Szyf, 2007). In our study, we measured asignificant decrease in the expression of monomethyl histone H3K9and a significant increase in dimethyl and trimethyl histone H3K9expression after VPA exposure, suggesting a temporary global shifttowards active gene transcription in the embryo. Histone methyl-ation is critical in regulating gene expression during pre- andpostimplantation embryonic development, and undergoes dynamicchanges during mouse neural tube closure (Biron et al., 2004).Other studies also support the requirement of specific histonemethylation markers during neurulation (Mehedint et al., 2010;Pourebrahim et al., 2007), however future studies will be needed toelucidate the significance of histone modifications with regards tospecific genes following VPA exposure.

Adding another layer of complexity to the epigenetic control ofgene regulation is DNA methylation, where methyl groupscovalently bound to the 5' position of cytosine project into themajor groove of DNA, thus inhibiting the binding of transcriptionfactors (Jones et al., 2001). Interestingly, despite observing histonehyperacetylation, decreased methylation of histone H3K9, andincreased methylation of histone H3K4, we did not observe anydifferences in global or CpG island methylation following VPAtreatment. It is possible that due to our use of whole embryohomogenates, alterations in DNA methylation status were dilutedand not detectable as measured by the cytosine extension assay,despite observing significant global changes in whole embryohomogenates in our histone acetylation and methylation assays.Regardless, the significance of DNA methylation as a mechanism ofVPA-induced teratogenesis remains debatable. Previous studieshave examined the effect of VPA administration on folic acid, amethyl group donor in one carbon metabolism that can ultimatelyaffect the methionine cycle and DNA methylation. Though somestudies have shown that VPA reduces serum folate levels (Wegneret al., 1992) and alters methionine metabolism (Ubeda et al.,2002), other studies have suggested a minimal role for folate inVPA-induced teratogenesis (Spiegelstein et al., 2003). In addition,while some studies have demonstrated that VPA has the ability toincrease histone acetylation and decrease DNA methylationconcomitantly (Perisic et al., 2009), other studies have measured

209E.W.Y. Tung, L.M. Winn / Toxicology and Applied Pharmacology 248 (2010) 201–209

increases in histone acetylation that did not occur alongside DNAdemethylation (Tabolacci et al., 2008). Therefore it is evident thatfurther studies are required to deduce the gene and tissue specificeffects of maternal VPA administration on embryonic DNAmethylation.

In summary, in the present study we demonstrate that earlyincreases in histone acetylation occur within the embryo followingmaternal VPA exposure. These changes in acetylation were correlatedwith an increase in histone methylation at histone H3K4 and adecrease in methylation at histone H3K9. Though global DNAdemethylation was not detected, future studies will be required toexamine local changes in methylation status. In summary, these novelfindings support the hypothesis that early, transient epigeneticchanges contribute to the mechanism of VPA-induced teratogenesis.

Acknowledgments

The authors would like to thank John DaCosta for assistance withtissue processing and embedding, as well as Dr. Igor Pogribny and Dr.Tetyana Bagnyukova for providing the protocol and consultationregarding the cytosine extension assay. This work was supported by aCanadian Institutes of Health Research (CIHR) grant (MOP 86593) toLMW.

References

Berger, S.L., 2007. The complex language of chromatin regulation during transcription.Nature 447, 407–412.

Biron, V.L., McManus, K.J., Hu, N., Hendzel, M.J., Underhill, D.A., 2004. Distinct dynamicsand distribution of histone methyl-lysine derivatives in mouse development. Dev.Biol. 276, 337–351.

Bjerkedal, T., Czeizel, A., Goujard, J., Kallen, B., Mastroiacova, P., Nevin, N., Oakley Jr., G.,Robert, E., 1982. Valproic acid and spina bifida. Lancet 13, 1096.

Bowden, C.L., 2009. Anticonvulsants in bipolar disorders: current research and practiceand future directions. Bipolar Disord. 11, 20–23.

Clayton-Smith, J., Donnai, D., 1995. Fetal valproate syndrome. J. Med. Genet. 32, 724–727.Copp, A.J., Greene, N.D.E., 2010. Genetics and development of neural tube defects. J.

Pathol. 220, 217–230.Dawson, J.E., Raymond, A.M., Winn, L.M., 2006. Folic acid and pantothenic acid

protection against valproic acid-induced neural tube defects in CD-1 mice. Toxicol.Appl. Pharmacol. 211, 124–132.

Detich, N., Bovenzi, V., Szyf, M., 2003. Valproate induces replication-independent activeDNA demethylation. J. Biol. Chem. 278, 27586–27592.

Dong, E., Guidotti, A., Grayson, D.R., Costa, E., 2007. Histone hyperacetylation inducesdemethylation of reelin and 67-kDa glutamic acid decarboxylase promoters. Proc.Natl. Acad. Sci. U.S.A. 104, 4676–4681.

Eberharter, A., Becker, P., 2002. Histone acetylation: a switch between repressive andpermissive chromatin. EMBO Rep. 3, 224–229.

Evers, S., 2008. Treatment of migraine with prophylactic drugs. Expert Opin.Pharmacother. 9, 2565–2573.

Grant, P.A., Eberharter, A., John, S., Cook, R.G., Turner, B.M., Workman, J.L., 1999.Expanded lysine acetylation specificity of Gcn5 in native complexes. J. Biol. Chem.274, 5895–5900.

Gupta, S., Kim, S.Y., Artis, S., Molfese, D.L., Schumacher, A., Sweatt, J.D., Paylor, R.E.,Lubin, F.D., 2010. Histone methylation regulates memory formation. J. Neurosci. 30,3589–3599.

Gurvich, N., Berman, M.G., Wittner, B.S., Gentleman, R.C., Klein, P.S., Green, J.B.A., 2005.Association of valproate-induced teratogenesis with histone deacetylase inhibitionin vivo. FASEB J. 19, 1166–1180.

Jones, P.A., Takai, D., 2001. The role of DNA methylation in mammalian epigenetics.Science 293, 1068–1070.

Koren, G., Nava-Ocampo, A.A., Moretti, M.E., Sussman, R., Nulman, I., 2006. Majormalformations with valproic acid. Can. Fam. Phys. 52, 441–442.

Kultima, K., Nyström, A.N., Scholz, B., Gustafson, A.L., Dencker, L., Stigson, M., 2004.Valproic acid teratogenicity: a toxicogenomics approach. Environ. Health Perspect.12, 1225–1235.

Lagger, G., O'Carroll, D., Rembold,M., Khier, H., Tischler, J.,Weitzer, G., Schuettengruber, B.,Hauser, C., Brunmeir, R., Jenuwein, T., Seiser, C., 2002. Essential function of histonedeacetylase 1 in proliferation control and CDK inhibitor repression. EMBO J. 21,2672–2681.

Liu, Y., Balaraman, Y., Wang, G., Nephew, K.P., Zhou, J., 2009. Alcohol exposure altersDNA methylation profiles in mouse embryos at early neurulation. Epigenetics 4,500–511.

Mehedint, M.G., Niculescu, M.D., Craciunescu, C.N., Zeisel, S.H., 2010. Choline deficiencyalters global histone methylation and epigenetic marking at the Re1 site of thecalbindin 1 gene. FASEB J. 24, 184–195.

Menegola, E., Di Renzo, F., Broccia, M.L., Prudenziati, M., Minucci, S., Massa, V.,Giavini, E., 2005. Inhibition of histone deacetylase activity on specific embryonictissues as a new mechanism for teratogenicity. Birth Defects Res. B Dev. Reprod.Toxicol. 75, 392–398.

Milutinovic, S., D'Alessio, A.C., Detich, N., Szyf, M., 2007. Valproate induces widespreadepigenetic reprogramming which involves demethylation of specific genes.Carcinogenesis 28, 560–571.

Mitchell, C.P., Chen, Y., Kundakovic, M., Costa, E., Grayson, D.R., Histone deacetylaseinhibitors decrease reelin promoter methylation in vitro. J. Neurochem. 93,483-492.

Nightingale, K.P., Gendreizig, S., White, D.A., Bradbury, C., Hollfelder, F., Turner, B.M.,2007. Cross-talk between Histone Modifications in Response to Histone DeacetylaseInhibitors. J. Biol. Chem. 282, 4408–4416.

Okada, A., Kushima, K., Aoki, Y., Bialer, M., Fujiwara, M., 2005. Identification of early-responsive genes correlated to valproic acid-induced neural tube defects in mice.Birth Defects Res. A Clin. Mol. Teratol. 4, 229–238.

Ornoy, A., 2009. Valproic acid in pregnancy: Howmuch are we endangering the embryoand fetus? Reprod. Toxicol. 28, 1–10.

Perisic, T., Zimmermann, N., Kirmeier, T., Asmus, M., Tuorto, F., Uhr, M., Holsboer, F.,Rein, T., Zschocke, J., 2009. Valproate and amitriptyline exert common and divergentinfluences on global and gene promoter-specific chromatin modifications in ratprimary astrocytes. Neuropsychopharmacology 35, 792–805.

Phiel, C.J., Zhang, F., Huang, E.Y., Guenther, M.G., Lazar, M.A., Klein, P.S., 2001. Histonedeacetylase is a direct target of valproic acid, a potent anticonvulsant, moodstabilizer, and teratogen. J. Biol. Chem. 276, 36734–36741.

Pogribny, I., Yi, P., James, S.J., 1999. A sensitive new method for rapid detection ofabnormal methylation patterns in global DNA and within CpG islands. Biochem.Biophys. Res. Commun. 262, 624–628.

Pourebrahim, R., Van Dam, K., Bauters, M., DeWever, I., Sciot, R., Cassiman, J.J., Tejpar, S.,2007. ZIC1 gene expression is controlled by DNA and histone methylation inmesenchymal proliferations. FEBS Lett. 581, 5122–5126.

Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J.R., Cole, P.A., Casero, R.A., Shi, Y.,2004. Histone demethylation mediated by the nuclear amine oxidase homologLSD1. Cell 119, 941–953.

Spiegelstein, O., Merriweather, M.Y., Wicker, N.J., Finnell, R.H., 2003. Valproate-inducedneural tube defects in folate-binding protein-2 (Folbp2) knockout mice. BirthDefects Res. A Clin. Mol. Teratol. 12, 974–978.

Szyf, M., 2007. The dynamic epigenome and its implications in toxicology. Toxicol. Sci.100, 7–23.

Tabolacci, E., De Pascalis, I., Accadia, M., Terracciano, A., Moscato, U., Chiurazzi, P.,Neri, G., 2008. Modest reactivation of the mutant FMR1 gene by valproic acid isaccompanied by histone modifications but not DNA demethylation. Pharmacogenet.Genomics 18, 738–741.

Ubeda, N., Alonso-Aperte, E., Varela-Moreiras, G., 2002. Acute valproate administrationimpairs methionine metabolism in rats. J. Nutr. 132, 2737–2742.

VerMilyea, M.D., O'Neill, L.P., Turner, B.M., 2009. Transcription-independent heritabilityof induced histone modifications in the mouse preimplantation embryo. PLoS ONE4, e6086.

Wegner, C., Nau, H., 1992. Alteration of embryonic folate metabolism by valproic acidduring organogenesis: implications for mechanism of teratogenesis. Neurology 42,17–24.

Whetstine, J.R., Nottke, A., Lan, F., Huarte, M., Smolikov, S., Chen, Z., Spooner, E., Li, E.,Zhang, G., Colaiacovo, M., Shi, Y., 2006. Reversal of histone lysine trimethylation bythe JMJD2 family of histone demethylases. Cell 125, 467–481.

Wu, G., Nan, C., Rollo, J., Huang, X., Tian, J., 2010. Sodium valproate-induced congenitalcardiac abnormalities in mice are associated with the inhibition of histonedeacetylase. J. Biomed. Sci. 17, 16.

Zhang, Y., Reinberg, D., 2001. Transcription regulation by histone methylation:interplay between different covalent modifications of the core histone tails.Genes Dev. 15, 2343–2360.