Increased susceptibility to kainic acid–induced seizures in Engrailed-2 knockout mice

II

PSGa

b

c

Pd

Be

v

AiatthbbomttmafsbEamikzzmLlogtb

1

HU*sTEAIRt

Neuroscience 159 (2009) 842–849

0d

NCREASED SUSCEPTIBILITY TO KAINIC ACID–INDUCED SEIZURES

Kh

T(awpebirSsvsadfKbcaiataph

tbdmbcsnuBaMdassaiD

. P. TRIPATHI,a,b P. SGADÒ,c1 M. SCALI,a,b C. VIAGGI,c

. CASAROSA,d1 H. H. SIMON,e F. VAGLINI,c

. U. CORSINIc AND Y. BOZZIa1*

Institute of Neuroscience, C.N.R., Pisa, Italy

Laboratory of Neurobiology, Scuola Normale Superiore, Pisa, Italy

Department of Neuroscience, Section of Pharmacology, University ofisa, Italy

Department of Biology, Laboratory of Cellular and Developmentaliology, University of Pisa, Italy

Centre for Neuroscience (IZN), Department of Neuroanatomy, Uni-ersity of Heidelberg, Germany

bstract—The En2 gene, coding for the homeobox-contain-ng transcription factor Engrailed-2 (EN2), has been associ-ted to autism spectrum disorder (ASD). Due to neuroana-omical and behavioral abnormalities, which partly resemblehose observed in ASD patients, En2 knockout (En2�/�) miceave been proposed as a model for ASD. In the mouse em-ryo, En2 is involved in the specification of midbrain/hind-rain regions, being predominantly expressed in the devel-ping cerebellum and ventral midbrain, and its expression isaintained in these structures until adulthood. Here we show

hat in the adult mouse brain, En2 mRNA is expressed also inhe hippocampus and cerebral cortex. Hippocampal En2RNA content decreased after seizures induced by kainic

cid (KA). This suggests that En2 might also influence theunctioning of forebrain areas during adulthood and in re-ponse to seizures. Indeed, a reduced expression of parval-umin and somatostatin was detected in the hippocampus ofn2�/� mice as compared to wild-type (WT) mice, indicatingn altered GABAergic innervation of limbic circuits in En2�/�

ice. In keeping with these results, En2�/� mice displayed anncreased susceptibility to KA-induced seizures. KA (20 mg/g) determined more severe and prolonged generalized sei-ures in En2�/� mice, when compared to WT animals. Sei-ures were accompanied by a widespread c-fos and c-junRNA induction in the brain of En2�/� but not WT mice.ong-term histopathological changes (CA1 cell loss, upregu-

ation of neuropeptide Y) also occurred in the hippocampusf KA-treated En2�/� but not WT mice. These findings sug-est that En2�/� mice might be used as a novel tool to studyhe link between epilepsy and ASD. © 2009 IBRO. Publishedy Elsevier Ltd. All rights reserved.

Present address: Neurogenetics Laboratory, A. Meyer Children’sospital, Florence, Italy (P. Sgadò); Centre for Integrative Biology,niversity of Trento, Italy (S. Casarosa, Y. Bozzi).

Correspondence to: Y. Bozzi, Centre for Integrative Biology, Univer-ity of Trento, Via delle Regole 101, 38060 Mattarello, Trento, Italy.el: �39-0461-882961; fax: �39-0461-883937.-mail address: [email protected], [email protected] (Y. Bozzi).bbreviations: ASD, autism spectrum disorder; En2, engrailed-2;

EGs, immediate early genes; KA, kainic acid; NPY, neuropeptide Y;

mT-PCR, reverse transcription–polymerase chain reaction; WT, wild-

ype.

306-4522/09 © 2009 IBRO. Published by Elsevier Ltd. All rights reserved.oi:10.1016/j.neuroscience.2009.01.007

842

ey words: autism spectrum disorder, epilepsy, homeobox,ippocampus, glutamate.

he homeobox-containing transcription factor Engrailed-2EN2) is crucially involved in the regionalization, patterningnd neuronal differentiation of the midbrain and hindbrain,hich represent the regions of the CNS in which its ex-ression has been detected. Indeed, En2 is predominantlyxpressed in the developing cerebellum and ventral mid-rain, starting at the neural plate stage (E8.5) and continu-

ng throughout embryonic and postnatal development (foreviews see Joyner, 1996; Herrup et al., 2005; Hidalgo-ánchez et al., 2005; Gherbassi and Simon, 2006). Initialtudies performed on En2 knockout (En2�/�) mice re-ealed a phenotype consistent with this restricted expres-ion pattern. En2�/� mice displayed cerebellar hypoplasia,reduced number of Purkinje cells and a subtle but repro-ucible defect in the anteroposterior pattern of cerebellaroliation (Joyner et al., 1991; Millen et al., 1994, 1995;uemerle et al., 1997). More recently, deficits in socialehaviors were detected in En2�/� mice, including de-reased play, reduced social sniffing and allogrooming,nd reduced aggressiveness (Cheh et al., 2006). Deficits

n spatial learning and memory tasks (Morris water mazend object recognition test), as well as in specific motorasks (rotarod) were also reported in En2�/� mice (Cheh etl., 2006). These studies suggest that En2 might be ex-ressed also in more anterior brain areas during adult-ood.

Due to their complex neurodevelopmental, neuroana-omical and behavioral phenotype, En2�/� mice haveeen proposed as a novel model for autism spectrumisorder (ASD). Indeed, abnormalities observed in En2�/�

ice resemble—at least in part—some of those that haveeen reported in ASD patients, such as hypoplasia oferebellar vermal lobules (Courchesne et al., 1988). Manytudies also reported a significant reduction in Purkinje cellumber, cerebellar nuclei and inferior olive in ASD individ-als (Williams et al., 1980; Bauman and Kemper, 1985;auman, 1991; reviewed in DiCicco-Bloom et al., 2006),s also observed in En2�/� mice (Kuemerle et al., 1997).ore recently, altered anatomical structure of the amyg-ala has also been reported in En2�/� mice (Kuemerle etl., 2007). It is interesting to notice that in ASD patients,ignificant neuropathological alterations have been de-cribed in several telencephalic structures, including themygdala, hippocampus and other limbic areas (reviewed

n Palmen et al., 2004; Bauman and Kemper, 2005;iCicco-Bloom et al., 2006). Finally, the human En2 gene

tia

stmtwpmStbEd

A

TeWibUoa(eKeiewmimiacNtaa

Qr

TIcmaPmtrWRuPsmi

flw(cqEcmLscafl7ctowwsvtKptLpt

B

KdEes1fmafzsTe(fhainAis

I

Mrthepplfisl

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849 843

o ASD, and two single-nucleotide polymorphisms (SNPs)n the En2 gene have been associated to ASD (Gharani etl., 2004; Benayed et al., 2005; Brune et al., 2008).

Since a high prevalence of epilepsy has been de-cribed in ASD patients (Deonna and Roulet, 2006; Cani-ano, 2007), we studied seizure susceptibility in En2�/�

utant mice. Seizures were evoked by systemic adminis-ration of the glutamate receptor agonist kainic acid (KA), aidely used model to study the behavioral and neuro-athological consequences of seizures in rodents (Loth-an and Collins, 1981; Sperk, 1994; Schauwecker andteward, 1997). The effects of KA administration in wild-

ype (WT) and En2�/� mice were evaluated by means ofehavioral and histological analyses. Here we show thatn2�/� mice display increased susceptibility to KA-in-uced seizures and long-term histopathology.

EXPERIMENTAL PROCEDURES

nimals

he generation of En2�/� mice was previously described (Joynert al., 1991). The original En2 mutants (mixed 129 Sv�Swiss–ebster genetic background) were crossed at least three times

nto a C57BL/6 background. En2�/� heterozygous mice of thisackground were obtained from the central animal facility of theniversity of Heidelberg, and used as founder animals to establishur own colony at the University of Pisa. Because En2�/� micere viable and fertile (Joyner et al., 1991), two separate En2�/�

WT) and En2�/� colonies of the same genetic background werestablished, according to previous studies (Cheh et al., 2006;uemerle et al., 2007). This was obtained as follows: from het-rozygous mating (En2�/��En2�/�), WT and En2�/� mice were

dentified by PCR genotyping (Sgadò et al., 2006) and used tostablish the two different colonies. The WT and En2�/� coloniesere maintained by (En2�/��En2�/�) and (En2�/��En2�/�)ating, respectively. For this reason, WT and En2�/� mice used

n this study were not littermates; however, age-matched adult (5onths old; weight�25–35 g) male mice were used in all exper-

ments. Animals were housed in a 12-h light/dark cycle with foodnd water available ad libitum. Experiments were conducted inonformity with the European Communities Council Directive ofovember 24, 1986 (86/609/EEC). All experiments conformed to

he Italian Ministry of Health guidelines on the ethical use ofnimals. Care was taken to minimize the number of animals usednd their suffering.

uantitative reverse transcription–polymerase chaineaction (RT-PCR)

otal RNAs were extracted by Trizol© reagent (Invitrogen, Milan,taly) from the cerebral cortex, hippocampus, ventral midbrain anderebellum of four adult WT mice and pooled. For KA experi-ents, RNAs were extracted and pooled from hippocampi of fourdult WT mice, killed 3 h after i.p. administration of KA (KA; Oceanroduce International, Shelburne, NS, Canada) at the dose of 20g/kg KA-treated mice experienced generalized seizures within

he first 2 h after KA administration (see also below for seizureating scale). RNA extracted from the hippocampus of four adult

T mice treated with saline was used as a control. DNase-treatedNAs were purified and concentrated with Nucleospin RNA col-mns (Macherey-Nagel, Düren, Germany). cDNA for real-timeCR was synthesized from RNA (2 �g) using the reverse tran-criptase Core kit (Eurogentec, Seraing, Belgium) according to theanufacturer’s instructions. Quantitative PCR was performed us-

ng a Rotor-gene 2000 thermal cycler with real-time detection of m

uorescence (Corbett Research, Sydney, Australia). PCR reactionsere conducted in a volume of 25 �l using the Mesa Green qPCR kit

Eurogentec) according to manufacturer’s instructions. Mouse mito-hondrial ribosomal protein L41 (Mrpl41) was used as a standard foruantification. Primers (Sygma-Genosys, Milan, Italy) were as follows:n2 forward 5=-agagagggcgcagttctttg-3=; En2 reverse 5=-cgacacagacg-agacacac-3= (GenBank accession no. NM_010134.3; expected frag-ent size: 151 bp); L41 forward 5=-GGTTCTCCCTTTCTCCCTTG-3=;41 reverse 5=-GCACCCCGACTCTTAGTGAA-3= (GenBank acces-ion no. NM_001031808.2; expected fragment size: 179 bp). Each PCRycle consisted of denaturation for 10 s at 94 °C, annealing for 20 st 62 °C (60 °C for L41), and extension for 30 s at 72 °C. Theuorescence intensity of SYBR Green I was read and acquired at2 °C after completion of the extension step of each cycle. PCRonditions for individual primer sets were optimized by varyingemplate cDNA and magnesium ion concentration in order tobtain amplifications yielding a single product and melt curvesith a single uniform peak. Quantification of individual transcriptsas performed using the dComparative QuantitationT softwareupplied with Rotor-gene. En2 and L41 mRNA concentrations inentral midbrain, cerebral cortex and hippocampus were referredo those detected in the cerebellum (comparative quantitation). InA experiments, En2 and L41 mRNA concentrations in the hip-ocampus of KA-treated mice were referred to those detected inhe hippocampus of saline-treated animals. Ratios of En2 mRNA/41 mRNA comparative concentrations were then calculated andlotted as the average of three different technical replicates ob-ained from each RNA pool.

ehavioral observation of KA-induced seizures

A (Ocean Produce International, Shelburne, NS, Canada) wasissolved in saline and administered i.p. at 20 mg/kg body weight.ight WT and 12 En2�/� mice were used. In all experiments, thexperimenter was blind to the genotype of the animals. Seizureeverity was determined according to Racine’s scale (Racine,972): stage 0: normal behavior; stage 1: immobility; stage 2:orelimb and/or tail extension, rigid posture; stage 3: repetitiveovements, head bobbing; stage 4: forelimb clonus with rearingnd falling (limbic motor seizure); stage 5: continuous rearing andalling; stage 6: severe whole body convulsions (tonic–clonic sei-ures); stage 7: death. For each animal, the rating scale value wascored every 20 min for a maximum of 3 h after KA administration.he maximum rating scale values reached by each animal overach 20 min interval were used to calculate the rating scale value�SE) for each treatment group. Statistical analysis was per-ormed by two-way repeated measures ANOVA followed by postoc Holm-Sidak test. At the end of behavioral observation (3 hfter KA), a subset of mice (WT, n�3; En2�/�, n�5) was killed for

n situ hybridization. Another subset of mice (WT, n�5; En2�/�,�7) was killed at 7 days after KA for histopathological analyses.n additional group of WT and En2�/� mice received saline

njection and was used as controls for behavioral observations, initu hybridization and histopathology.

n situ hybridization

ice were killed at 3 h after KA injection, and brains were rapidlyemoved and frozen on dry ice. Coronal cryostat sections (20 �mhick) were fixed in 4% paraformaldehyde. Non-radioactive in situybridization was performed as previously described (Antonuccit al., 2008) using a digoxigenin-labeled c-fos and c-jun ribo-robes (Bozzi et al., 2000). Signal was detected by alkaline phos-hatase–conjugated anti-digoxigenin antibody followed by alka-

ine phosphatase staining. The specificity of the results was con-rmed by the use of sense riboprobes which gave no detectableignal (not shown). Brain areas were identified according to Frank-in and Paxinos (1997). To quantify the level of c-fos and c-jun

RNAs, digital images of three matching sections per animal,

ttswpba

I

Fts(bcfwaaGmptabB

atCaarNfiEe

taocCiwpicf

Ea

Naaagatdta

bFcagembopaapgtlt

EG

SCanal2atrstl2sl(

Et

RmtWmsElseto(if(et

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849844

aken at the level of the dorsal hippocampus, were analyzed usinghe Image J free software (http://rsb.info.nih.gov/ij/). For eachection, signal intensity was measured in 10 different circularindows (area�0.01 mm2) placed in layers 2–3 and 5–6 of thearietal/temporal cortex. Mean signal intensity was divided by theackground labeling calculated in the corpus callosum. Statisticalnalysis was performed by Student’s t-test.

mmunohistochemistry

or parvalbumin and somatostatin staining, three WT adult micehat never received KA were used. For neurodegeneration and NPYtaining studies on KA-treated mice, five WT and seven En2�/� micesee above) were killed 7 days after 20 mg/kg KA. Brains were fixedy immersion in 4% paraformaldehyde, cryoprotected in 30% su-rose/1� PBS and coronal sections (40 �m thick) were cut on areezing microtome. Serial sections from the dorsal hippocampusere incubated overnight with appropriate antibodies as follows:nti-parvalbumin monoclonal (Sygma, Milan, Italy; 1:5000 dilution);nti-somatostatin polyclonal (Peninsula-Bachem, Well am Rhein,ermany; 1:5000 dilution); anti-NeuN monoclonal (Chemicon, Te-ecula, CA, USA; 1:500 dilution) or anti–neuropeptide Y (NPY)olyclonal (Bachem, Well am Rhein, Germany; 1:5000 dilution) an-ibodies. Signals were revealed with appropriate biotinylated second-ry antibodies (Vector Laboratories, Burlingame, CA, USA) followedy avidin–biotin–peroxidase complex (ABC kit, Vector Laboratories,urlingame, CA, USA) and diaminobenzidine (DAB) reaction.

Neuronal damage was qualitatively assessed in CA1/CA3reas of NeuN-stained sections from KA-treated mice, accordingo the following scale (Bozzi et al., 2000; Bozzi and Borrelli, 2002;ilio et al., 2001): little damage, presence of scattered degener-ted cells; mild damage, small areas with degenerated cellsnd/or tissue sclerosis; severe damage, extended areas of neu-onal and fiber degeneration, accompanied by tissue sclerosis.eurodegeneration was also confirmed by Nissl staining, per-

ormed on sections adjacent to those used for NeuN histochem-stry. Nissl and NeuN staining confirmed cerebellar hypoplasia inn2�/� brains (not shown), according to previous studies (Joynert al., 1991; Millen et al., 1994, 1995; Kuemerle et al., 1997).

To quantify NPY staining in the mossy fiber pathway of KA-reated mice, three sections through the dorsal hippocampus werenalyzed per mouse. Images of mossy fiber pathway and of theverlying corpus callosum were digitized. Light intensity and mi-roscope settings were optimized initially and then held constant.are was taken to avoid saturation at either end of the pixel

ntensity range (0–255). For each section, NPY signal intensityas calculated in 10 boxes (40�40 �m) placed in mossy fiberathway using the Image J free software (http://rsb.info.nih.gov/

j/). Mean signal intensity was divided by the background labelingalculated in the corpus callosum. Statistical analysis was per-ormed by Mann–Whitney rank sum test.

RESULTS

n2 is expressed in the adult mouse hippocampusnd cerebral cortex

euroanatomical and behavioural studies performed ondult En2�/� mice suggest that En2 might be expressedlso in anterior brain structures during adulthood (Cheh etl., 2006; Kuemerle et al., 2007). We therefore investi-ated En2 mRNA expression in different brain areas of thedult mouse brain. To this purpose, we performed quanti-ative real-time RT-PCR experiments using the mitochon-rial ribosomal L41 protein mRNA as a standard for quan-ification. As expected, L41 amplification gave comparable

ellum, ventral midbrain, hippocampus, cerebral cortex;ig. 1A). En2 transcripts were detected at low but signifi-ant levels in the hippocampus and cerebral cortex; En2mplification curves from these two areas were indistin-uishable (Fig 1B). According to previous studies (Joynert al., 1991; Millen et al., 1994; Simon et al., 2001), En2RNA was detected at higher levels in the ventral mid-rain and cerebellum (Fig. 1B). Comparative quantificationf real-time RT-PCR experiments showed that in hip-ocampus and cerebral cortex, En2 mRNA was presentbout 100 times less than in cerebellum (Fig. 1C). In thedult hippocampus, En2 mRNA levels were regulated byathological hyperactivity. In animals that experiencedeneralized seizures following systemic administration ofhe glutamate agonist KA (20 mg/kg i.p.), En2 mRNAevels were decreased by 30%, as compared to saline-reated controls (Fig. 1D).

n2�/� mice have a reduced expression ofABAergic markers in the hippocampus

ince En2 is expressed in the adult hippocampus (Fig. 1B,), we sought to investigate the presence of subtle neuro-natomical defects in this structure of En2�/� mice. Immu-ohistochemistry experiments with anti-parvalbumin andnti-somatostatin antibodies were performed to detect se-

ected GABAergic interneuron populations (Matyas et al.,004; Jinno and Kosaka, 2006) in the hippocampus of WTnd En2�/� mice. In WT mice, parvalbumin revealed theypical staining around the cell bodies of pyramidal neu-ons of CA1 (not shown) and CA3 (Fig. 2) hippocampalubfields (see also Matyas et al., 2004), whereas soma-ostatin predominantly labeled stratum lacunosum molecu-are and hilar interneurons (Fig. 2; see also Matyas et al.,004). In En2�/� mice, staining for both parvalbumin andomatostatin was markedly reduced in CA3 pyramidal

ayer and stratum lacunosum moleculare, respectivelyFig. 2).

n2�/� mice display an increased susceptibilityo KA seizures

educed expression of parvalbumin and somatostatinight indicate an altered GABAergic innervation (and,

hus, excitability) of hippocampal circuitry in En2�/� mice.e therefore investigated seizure susceptibility in En2�/�

ice. En2�/� mice never showed hyper-excitability orpontaneous seizures during standard housing. However,n2�/� mice presented a peculiar response when chal-

enged with KA. The time course of the behavioral re-ponse of WT and En2�/� mice to KA (20 mg/kg) wasvaluated over a period of 3 h after administration. In WT,his dose of KA generally resulted in the sole appearancef pre-convulsive behaviors at all time-points analyzedFig. 3). Only four out of eight WT mice displayed brief,solated episodes of limbic motor seizures (rearing withorelimb clonus, stage 4), and never showed tonic–clonicstage 6) seizures. The same KA dose in En2�/� micelicited clear signs of focal epilepsy (head bobbing) withinhe first 20 min, rapidly culminating in stage 4 limbic motor

fw

pm(te

Fmmqvqs cortex;r Web vers

FEpmAl

Fm

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849 845

rom that observed in WT mice (Table 1). In sharp contrastith WT, the majority (7 out of 12) of En2�/� mice dis-

ig. 1. En2 mRNA is expressed in adult mouse hip and ctx and is reitochondrial ribosomal protein L41 (A) and En2 (B) mRNAs from theice. The graphs report the appearance of fluorescence in PCR amuantification of real-time RT-PCR experiments. In (C), values are ealues of three technical replicates) in different adult brain areas, normauantitation ratios (average values of three technical replicates), fromaline-treated controls. Abbreviations: cb, cerebellum; ctx, cerebraleferences to color in this figure legend, the reader is referred to the

ig. 2. Downregulation of GABAergic markers in the hippocampus ofn2�/� mice. (A) Representative parvalbumin staining in the CA3yramidal cell layer of WT and En2�/� mice. (B) Representative so-atostatin staining in the slm and hilus of WT and En2�/� mice.

acunosum moleculare. Scale bar�50 �m.tD

layed severe tonic–clonic seizures (Table 1). En2�/�

ice showed generalized stage 4–6 seizures for about 2 h40–160 min; Fig. 3). Statistical analysis performed bywo-way repeated measures ANOVA revealed a significantffect of genotype (F1,136�7.522, P�0.014). Multiple com-

by seizure activity. (A, B) Real-time RT-PCR amplification profiles ofline), vmb (green line), hip (black line) and ctx (red line) of WT adult

s a function of the number of PCR cycles. (C, D) Graphs report theas En2 mRNA/L41 mRNA comparative quantitation ratios (average

b. In (D), values are expressed as En2 mRNA/L41 mRNA comparativeof KA-treated adult mice (3 h post KA, 20 mg/kg i.p.), normalized tohip, hippocampus; vmb, ventral midbrain. For interpretation of theion of this article.

ig. 3. Increased susceptibility to KA-induced seizures in En2�/�

ice. Progression of behavioral changes after systemic KA adminis-�/�

gulatedcb (blue

plicons axpressedlized to cthe hip

ration (20 mg/kg i.p.) in WT and En2 over a 3 h observation period.ata are mean seizure scores �SE.

pcpEgw

Ei

TdaaiEaWpepap4bbEwmctcibtdca

L

Iism

tIbtemlaHer

F(ssvpctbpat

Fm

T

ALAL

ma

b

P*

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849846

arison procedure showed that En2�/� mice had signifi-antly higher behavioral scores than WT mice at all timeoints analyzed (Fig. 3; Holm-Sidak post hoc test, WT vs.n2�/� mice, P�0.05). Saline-injected animals of bothenotypes showed no sign of epileptic activity during thehole period of observation (not shown).

n2�/� mice display increased induction ofmmediate early genes (IEGs) after KA seizures

he occurrence of limbic seizures induced by KA in ro-ents correlates with specific patterns of activation of brainreas, as revealed by IEG induction studies (Willoughby etl., 1997; Bozzi et al., 2000). Thus, c-fos and c-jun mRNA

nduction was analyzed by in situ hybridization on WT andn2�/� brains, 3 h after KA administration, to map brainreas differentially activated by KA in the two genotypes. InT mice, c-fos mRNA was mainly detected in the hip-

ocampus and in other limbic areas (amygdala, pyriform/ntorhinal cortex). No signal was detected in the caudate-utamen and thalamus (Fig. 4A). This induction profile waslso observed for c-jun mRNA, with the exception of theyriform cortex (that showed no c-jun mRNA labeling) (Fig.A). In keeping with seizure generalization observed byehavioral analysis, a widespread and strong induction ofoth c-fos and c-jun mRNAs was detected throughoutn2�/� brains. In particular, c-fos and c-jun mRNA labelingas evident in caudate-putamen, pyriform cortex, thala-us, amygdala, hippocampus, entorhinal cortex and other

ortical areas (Fig. 4A). Quantification of in situ hybridiza-ion experiments performed at the level of the parietalortex confirmed that c-fos and c-jun mRNAs were signif-cantly increased in both layers 2–3 and 5–6 in En2�/�

rains, as compared to WT controls (P�0.001, Student’s-test) (Fig. 4B, C) saline-treated mice of both genotypesid not show any c-fos mRNA signal, while basal levels of-jun mRNA were detected in the hippocampus in thesenimals (not shown).

ong-term histopathology in KA-treated En2�/� mice

n order to assess whether increased susceptibility to KA-nduced seizures in En2�/� mice also resulted in increasedusceptibility to long-term damage, the histology of pyra-idal cell layers and mossy fiber pathway was evaluated in

able 1. Generalized seizures in WT and En2�/� mice

WT En2�/�

nimals with stage 4 seizures 4/8 10/12a

atency (minutes) to 1st stage 4 seizure 35.5�7.7 24.7�3.2b

nimals with stage 6 seizures 0/8 7/12*atency (minutes) to 1st stage 6 seizure n.d. 49.8�13.9

Seizure latency (mean�SE) is calculated from the time of KA ad-inistration.Not different between the two groups (z-test, P�0.05).Not different between WT and En2�/� mice (Mann–Whitney test,�0.05).Significantly different between the two groups (z-test, P�0.05).

cAbbreviation: n.d., not determined.

he hippocampus of WT and En2�/� mice, 7 days after KA.mmunostaining for the pan-neuronal marker NeuN onrain sections from WT mice did not reveal any damage inhe CA1 pyramidal cell layer (Fig 5), and scattered degen-rated cells were only occasionally observed in CA3 pyra-idal neurons (Table 2). Conversely, in En2�/� mice, cell

oss and tissue sclerosis were detected in four out of sevennimals in both CA1 and CA3 regions (Fig. 5, Table 2).istological changes in the mossy fiber pathway werevaluated using NPY immunohistochemistry. A robust up-egulation of NPY immunoreactivity was found in the

ig. 4. IEGs are differentially induced by KA in WT and En2�/� mice.A) c-fos and c-jun mRNA in situ hybridizations, 3 h after KA. Repre-entative sections at the level of the CPu and dorsal hippocampus arehown. Genotypes and relevant brain areas are as indicated. Abbre-iations: Amy, amygdala; CA1/CA3, pyramidal cell layers of the hip-ocampus; CPu, caudate-putamen; DG, dentate gyrus; ent, entorhinalortex; ht, hypothalamus; pyr, pyriform cortex; Sep, septum; thal,halamus. 2–3 And 5–6 indicate layers of the cerebral cortex Scalear�2 mm. (B, C) Quantification of c-fos (B) and c-jun (C) mRNAs inarietal/temporal cortex of KA-treated WT and En2�/� mice. Valuesre mean normalized signal intensities �SE. ** P�0.001 (Student’s

-test).

ig. 5. KA-induced neurodegeneration in the CA1 subfield of En2�/�

ice. NeuN immunostaining of coronal sections from the dorsal hip--/-

orpus callosum; pyr, pyramidal cell layer. Scale bar�150 �m.

mw(pins

BcalstEil

sdmmw2ma6c

itzgtiPslf(bAarvztssmm

hmKgmdafiCoiadtaaMd(gs

TK

B

C

C

K

F(SNlr

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849 847

ossy fibers of dentate gyrus in all En2�/� mice (Fig. 6B),hereas no such labeling was detected in WT animals

Fig. 6A). Quantification of NPY staining in the mossy fiberathway from WT and En2�/� mice confirmed these find-

ngs (Fig. 6C). Saline-treated mice of both genotypes didot show any sign of hippocampal histopathology (nothown).

DISCUSSION

y using behavioral, gene expression and neuroanatomi-al analyses, we provided evidence that En2�/� mice haven increased susceptibility to KA-induced seizures and

ong-term histopathology. En2�/� mice displayed moreevere and prolonged generalized seizures as comparedo WT mice. The occurrence of generalized seizures inn2�/� mice was accompanied by the widespread mRNA

nduction of the IEGs c-fos and c-jun, as well as CA1 celloss and NPY upregulation in mossy fibers.

Systemic KA administration has been widely used totudy the susceptibility to acute seizures and seizure-in-uced long-term histopathology in inbred and mutantouse strains. In the present study, control and En2�/�

ice of a mixed 129/Sv�C57BL/6 genetic backgroundere used. According to our previous studies (Bozzi et al.,000), WT mice displayed a very mild response to 20g/kg KA, never showing continuous generalized epilepticctivity (stage 5 seizures) or tonic–clonic seizures (stage) after KA. Conversely, En2�/� mice displayed high sus-eptibility to seizures induced by the same dose of KA,

able 2. Cell damage in hippocampal CA1/CA3 pyramidal layers ofA-treated WT and En2�/� mice

Degree of cell damage (no. of animals)

rain area/genotype None Little Mild Severe

A1WT (n�5) 5 0 0 0En2�/� (n�7) 3 2 1 1

A3WT (n�5) 3 1 1 0En2�/� (n�7) 3 2 0 2

Brain damage was evaluated in NeuN-stained sections, 7 days afterA, according to the scale described in Experimental Methods.

ig. 6. NPY upregulation in the mf pathway of KA-treated En2�/� miceB) mice, 7 days after KA. The almost complete loss of CA1 is also viscale bar�500 �m. (C) Quantification of NPY staining intensity in thePY signal-to-background ratio (intensity of NPY label divided by the b

ines in the box denote the 25th, 50th, and 75th percentile values. The error barsank sum test).

ndicating a seizure-promoting effect of En2 inactivation inhe mixed 129/Sv�C57BL/6 background. Increased sei-ure susceptibility in En2�/� mice was also confirmed byene expression studies. A precise correlation exists be-ween the occurrence of generalized seizures and thenduction pattern of the IEGs c-fos and c-jun following KA.re-convulsive behaviors (stages 1–3 of the Racine’scale) induce IEGs mainly in the hippocampus and other

imbic areas (hippocampus, amygdala, entorhinal and pyri-orm cortices) whereas continuous generalized seizuresstages 4–6) result in a widespread expression in severalrain areas (Willoughby et al., 1997; Bozzi et al., 2000).ccordingly, WT mice experienced no or very brief gener-lized seizures after KA, showing IEGs mRNA inductionestricted to the hippocampus and other limbic areas. Con-ersely, En2�/� mice showed long-lasting generalized sei-ures and widespread induction of IEGs mRNA throughouthe brain. It is important to point out that spontaneouseizures never occurred in naive En2�/� mice (data nothown), indicating that increased susceptibility of En2�/�

ice is evident only in response to a potent seizure-pro-oting stimulus, such as KA.

KA-induced seizures were also followed by long-termistopathological changes in the hippocampus of En2�/�

ice. In mice, the occurrence of brain damage followingA-evoked seizures strongly depends on the genetic back-round; certain mouse strains, such as inbred C57BL/6 orixed 129/Sv�C57BL/6, are resistant to KA-induced celleath (Schauwecker and Steward, 1997; Schauwecker etl., 2000; McLin and Steward, 2006). In keeping with thesendings, our WT mice never showed cell damage in theA1 pyramidal layer and only occasionally presented littler mild damage in the CA3. A marked degeneration was

nstead observed in all En2�/� mice analyzed in the CA1nd, to a lesser extent, in the CA3. This variability in theegree of brain damage following systemic KA administra-ion, that we observed in both WT and En2�/� mice, is ingreement with previous studies (Schauwecker and Stew-rd, 1997; Bozzi et al., 2000; McLin and Steward, 2006).ost importantly, En2�/� but not WT mice treated with KAisplayed NPY upregulation in the mossy fiber pathwaydentate gyrus to CA3). This pattern of NPY upregulation isenerally considered a reliable marker of long-term, post-eizure synaptic rearrangements, and is thought to be

epresentative NPY staining in the hippocampus of WT (A) and En2�/�

. Abbreviations: CA1, pyramidal cell layer; (H) hilus; mf, mossy fibers.and En2�/� mice. Each box chart summarizes the distribution of thed staining) for all hippocampal sections in each group. The horizontal

. (A, B) Rible in (B)mf of WTackgroun

denote the 5th and 95th percentile values. * P�0.05 (Mann–Whitney

ipd2po(vmc

ictOpgrRtwfiheit

iGrtitptwcmcdcmtBmvm

bpucmdopads

wtBeata

AasWPe(GUfidGa

A

B

B

B

B

B

B

B

B

C

C

C

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849848

ndicative of an acquired hyper-excitability of the hip-ocampus following seizures (Morimoto et al., 2004; Na-ler et al., 2007; Dudek and Sutula, 2007; Sperk et al.,007; Sutula and Dudek, 2007). Moreover, as observed foryramidal cell loss, rearrangements in mossy fibers do notccur following KA in mouse strains derived from C57BL/6Schauwecker et al., 2000). Our data suggest that inacti-ation of the En2 gene also results in long-term anatomicalodifications and, likely, hyper-excitability of hippocampal

ircuitry in response to KA seizures.It remains to be determined how the En2 mutation can

mpact hippocampal excitability. Indeed, the En2 gene isommonly known for its crucial role in pattern formation ofhe midbrain and hindbrain regions (see the introduction).ur data show that in the adult mouse brain, En2 is ex-ressed in the hippocampus and cerebral cortex, two re-ions crucially involved in seizure generation and sp-ead. These results confirm and expand non-quantitativeT-PCR expression data previously published on-line in

he mouse genome informatics (MGI) database (http://ww.informatics.jax.org). Moreover, we showed for therst time that the expression of the En2 gene in the adultippocampus is downregulated by seizure activity. Thisxpression profile indicates that En2 may be also involved

n the functioning of adult brain areas and in their responseo seizures.

The reduced staining of parvalbumin and somatostatinn the hippocampus of En2�/� mice suggests a reducedABAergic innervation onto CA3 and CA1 pyramidal neu-

ons, that might explain the hyper-excitability observed inhese mutants. The present study, along with recent find-ngs from other authors (Kuemerle et al., 2007), suggestshat En2 is implicated also in the development of telence-halic structures. Kuemerle et al. (2007) reported an an-erior shift of the amygdala in En2�/� mice. Specifically, itas found that in mutant mice the lateral, basolateral,entral and medial nuclei of the amygdala are located in aore anterior position in the cortex when compared to

ontrols. These anatomical alterations, together with thoseetected in the hippocampus, might contribute to alterircuitry and excitability of the limbic system in En2�/�

ice. EN2 secretion from posterior structures to anteriorarget areas could explain these effects (Joliot et al., 1998;runet et al., 2005; Sonnier et al., 2007). Further studiesight be aimed at determining whether EN2 released from

entral midbrain projections can influence excitability ofore anterior limbic structures.

Our results strengthen the notion that En2�/� mice cane used as a model to study at least some of the multipleathological aspects of ASD. In particular, we propose tose these mutants to further investigate the role of alteredircuitry/excitability of the limbic system in ASD and, ulti-ately, the relationship between epilepsy and ASD. In-

eed, morphological abnormalities (likely of developmentalrigin) have been detected in the limbic system of autisticatients, that might underlie increased excitability. Classicalnatomical studies showed increased cell density and re-uced cell size in limbic structures including hippocampus,

ere also observed in the same structures from autistic pa-ients (Kemper and Bauman, 1993; Palmen et al., 2004;auman and Kemper, 2005). It will be also interesting toxploit En2 mutant mice to understand how the inactivation ofgene involved in midbrain/hindbrain patterning can affect

he embryonic development (and adult functioning) of morenterior, telencephalic structures.

cknowledgments—We thank Giulio Cappagli, Adriano Tacchind the technical/administrative staff of the Istituto di Neuro-cienze del C.N.R. for excellent assistance, and Matteo Caleo,olfgang Wurst, Nilima Prakash, Sara Migliarini and Massimoasqualetti for helpful discussions and reagents. Y.B. is a recipi-nt of a research grant from the National Research CouncilCNR—“Ricerche Spontanee a Tema Libero”—RSTL Program)..U.C. is a recipient of a research grant from the Italian Ministry ofniversity and Research (Prin, 2005, Prot. No. 2005055445). Thenancial support from the “Fondo per gli Investimenti della Ricercai Base” (FIRB, CHEM-PROFARMA-NET Project) of the Italianovernment to the Istituto di Neuroscienze del C.N.R. is alsocknowledged.

REFERENCES

ntonucci F, Di Garbo A, Novelli E, Manno I, Sartucci F, Bozzi Y,Caleo M (2008) Botulinum neurotoxin E (Bont/E) reduces CA1neuron loss and granule cell dispersion, with no effects on chronicseizures, in a mouse model of temporal lobe epilepsy. Exp Neurol210:388–401.

auman M, Kemper TL (1985) Histoanatomic observations of thebrain in early infantile autism. Neurology 35:866–874.

auman ML (1991) Microscopic neuroanatomic abnormalities in au-tism. Pediatrics 87:791–796.

auman ML, Kemper TL (2005) Neuroanatomic observations of thebrain in autism: a review and future directions. Int J Dev Neurosci23:183–217.

enayed R, Gharani N, Rossman I, Mancuso V, Lazar G, Kamdar S,Bruse SE, Tischfield S, Smith BJ, Zimmerman RA, Dicicco-BloomE, Brzustowicz LM, Millonig JH (2005) Support for the homeoboxtranscription factor gene ENGRAILED 2 as an autism spectrumdisorder susceptibility locus. Am J Hum Genet 77:851–868.

ozzi Y, Vallone D, Borrelli E (2000) Neuroprotective role of dopamineagainst hippocampal cell death. J Neurosci 20:8643–8649.

ozzi Y, Borrelli E (2002) Dopamine D2 receptor signaling controlsneuronal cell death induced by muscarinic and glutamatergicdrugs. Mol Cell Neurosci 19:263–271.

rune CW, Korvatska E, Allen-Brady K, Cook EH, Dawson G, DevlinB, Estes A, Hennelly M, Hyman SL, McMahon WM, Munson J,Rodier PM, Schellenberg GD, Stodgell CJ, Coon H (2008) Heter-ogeneous association between engrailed-2 and autism in theCPEA network. Am J Med Genet B Neuropsychiatr Genet 147B:187–193.

runet I, Weinl C, Piper M, Trembleau A, Volovitch M, Harris W,Prochiantz A, Holt C (2005) The transcription factor engrailed-2guides retinal axons. Nature 438:94–98.

anitano R (2007) Epilepsy in autism spectrum disorders. Eur ChildAdolesc Psychiatry 16:61–66.

heh MA, Millonig JH, Roselli LM, Ming X, Jacobsen E, Kamdar S,Wagner GC (2006) En2 knockout mice display neurobehavioraland neurochemical alterations relevant to autism spectrum disor-der. Brain Res 1116:166–176.

ilio MR, Bolanos AR, Liu Z, Schmid R, Yang Y, Stafstrom CE, MikatiMA, Holmes GL (2001) Anticonvulsant action and long-term effectsof gabapentin in the immature brain. Neuropharmacology 40:139–

C

D

D

D

F

G

G

H

H

J

J

J

J

K

K

K

L

M

M

M

M

M

N

P

R

S

S

S

S

S

S

S

S

W

W

P. P. Tripathi et al. / Neuroscience 159 (2009) 842–849 849

ourchesne E, Yeung-Courchesne R, Press GA, Hesselink JR, Jerni-gan TL (1988) Hypoplasia of cerebellar vermal lobules VI and VIIin autism. N Engl J Med 318:1349–1354.

eonna T, Roulet E (2006) Autistic spectrum disorder: evaluating apossible contributing or causal role of epilepsy. Epilepsia 47 (Suppl2):79–82.

iCicco-Bloom E, Lord C, Zwaigenbaum L, Courchesne E, Dager SR,Schmitz C, Schultz RT, Crawley J, Young LJ (2006) The develop-mental neurobiology of autism spectrum disorder. J Neurosci 26:6897–6906.

udek FE, Sutula TP (2007) Epileptogenesis in the dentate gyrus: acritical perspective. Prog Brain Res 163:755–773.

ranklin KBJ, Paxinos G (1997) The mouse brain in stereotaxic coor-dinates. San Diego: Academic Press.

harani N, Benayed R, Mancuso V, Brzustowicz LM, Millonig JH (2004)Association of the homeobox transcription factor, ENGRAILED 2, 3,with autism spectrum disorder. Mol Psychiatry 9:474–484.

herbassi D, Simon HH (2006) The engrailed transcription factors andthe mesencephalic dopaminergic neurons. J Neural Transm Suppl70:47–55.

errup K, Murcia C, Gulden F, Kuemerle B, Bilovocky N (2005) Thegenetics of early cerebellar development: networks not pathways.Prog Brain Res 148:21–27.

idalgo-Sánchez M, Millet S, Bloch-Gallego E, Alvarado-Mallart RM(2005) Specification of the meso-isthmo-cerebellar region: theOtx2/Gbx2 boundary. Brain Res Brain Res Rev 49:134–149.

inno S, Kosaka T (2006) Cellular architecture of the mouse hip-pocampus: a quantitative aspect of chemically defined GABAergicneurons with stereology. Neurosci Res 56:229–245.

oliot A, Maizel A, Rosenberg D, Trembleau A, Dupas S, Volovitch M,Prochiantz A (1998) Identification of a signal sequence necessaryfor the unconventional secretion of engrailed homeoprotein. CurrBiol 8:856–863.

oyner AL (1996) Engrailed, Wnt and pax genes regulate midbrain–hindbrain development. Trends Genet 12:15–20.

oyner AL, Herrup K, Auerbach BA, Davis CA, Rossant J (1991)Subtle cerebellar phenotype in mice homozygous for a targeteddeletion of the en-2 homeobox. Science 251:1239–1243.

emper TL, Bauman ML (1993) The contribution of neuropathologicstudies to the understanding of autism. Neurol Clin 11:175–187.

uemerle B, Gulden F, Cherosky N, Williams E, Herrup K (2007) Themouse engrailed genes: a window into autism. Behav Brain Res176:121–132.

uemerle B, Zanjani H, Joyner A, Herrup K (1997) Pattern deformitiesand cell loss in engrailed-2 mutant mice suggest two separate pat-terning events during cerebellar development. J Neurosci 17:7881–7889.

othman EW, Collins RC (1981) Kainic acid induced limbic seizures:metabolic, behavioral, electroencephalographic and neuropatho-logical correlates. Brain Res 218:299–318.

atyas F, Freund T, Gulyas AI (2004) Immunocytochemically definedinterneuron populations in the hippocampus of mouse strains used

cLin JP, Steward O (2006) Comparison of seizure phenotype andneurodegeneration induced by systemic kainic acid in inbred, out-bred, and hybrid mouse strains. Eur J Neurosci 24:2191–2202.

illen KJ, Wurst W, Herrup K, Joyner AL (1994) Abnormal embryoniccerebellar development and patterning of postnatal foliation in twomouse engrailed-2 mutants. Development 120:695–706.

illen KJ, Hui CC, Joyner AL (1995) A role for en-2 and other murinehomologues of Drosophila segment polarity genes in regulatingpositional information in the developing cerebellum. Development121:3935–3945.

orimoto K, Fahnestock M, Racine RJ (2004) Kindling and statusepilepticus models of epilepsy: rewiring the brain. Prog Neurobiol73:1–60.

adler JV, Tu B, Timofeeva O, Jiao Y, Herzog H (2007) NeuropeptideY in the recurrent mossy fiber pathway. Peptides 28:357–364.

almen SJ, van Engeland H, Hof PR, Schmitz C (2004) Neuropatho-logical findings in autism. Brain 127:2572–2583.

acine RJ (1972) Modification of seizure activity by electrical stimulation.II. Motor seizure. Electroencephalogr Clin Neurophysiol 32:281–294.

chauwecker PE, Ramirez JJ, Steward O (2000) Genetic dissection ofthe signals that induce synaptic reorganization. Exp Neurol 161:139–152.

chauwecker PE, Steward O (1997) Genetic determinants of suscep-tibility to excitotoxic cell death: implications for gene targetingapproaches. Proc Natl Acad Sci U S A 94:4103–4108.

gadò P, Albéri L, Gherbassi D, Galasso SL, Ramakers GM, AlavianKN, Smidt MP, Dyck RH, Simon HH (2006) Slow progressivedegeneration of nigral dopaminergic neurons in postnatal en-grailed mutant mice. Proc Natl Acad Sci U S A 103:15242–15247.

imon HH, Saueressig H, Wurst W, Goulding MD, O’Leary DD (2001)Fate of midbrain dopaminergic neurons controlled by the engrailedgenes. J Neurosci 21:3126–3134.

onnier L, Le Pen G, Hartmann A, Bizot JC, Trovero F, Krebs MO,Prochiantz A (2007) Progressive loss of dopaminergic neurons inthe ventral midbrain of adult mice heterozygote for Engrailed1.J Neurosci 27:1063–1071.

perk G (1994) Kainic acid seizures in the rat. Prog Neurobiol 42:1–32.

perk G, Hamilton T, Colmers WF (2007) Neuropeptide Y in thedentate gyrus. Prog Brain Res 163:285–297.

utula TP, Dudek FE (2007) Unmasking recurrent excitation gener-ated by mossy fiber sprouting in the epileptic dentate gyrus: anemergent property of a complex system. Prog Brain Res 163:541–563.

illiams RS, Hauser SL, Purpura DP, DeLong GR, Swisher CN (1980)Autism and mental retardation: neuropathologic studies performed infour retarded persons with autistic behavior. Arch Neurol 37:749–753.

illoughby JO, Mackenzie L, Medvedev A, Hiscock JJ (1997) Fosinduction following systemic kainic acid: early expression in hip-pocampus and later widespread expression correlated with sei-

(Accepted 1 January 2009)(Available online 10 January 2009)