GKP

IMMa

Bb

7

Antssimnpegclkmaesswbvedmr

Ks

Tt

*EAcmgcqGKppct1

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin gene transfer curtails generalized seizures in kindled rats withoutaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1016/j.neuroscience.2007.09.056

Neuroscience xx (2007) xxx

0d

ARTICLE IN PRESS

ALANIN GENE TRANSFER CURTAILS GENERALIZED SEIZURES ININDLED RATS WITHOUT ALTERING HIPPOCAMPAL SYNAPTIC

LASTICITY.

pvH1pttebt1(eG(g

ttodpstTmies2bgmeslpktmmsmancag

. KANTER-SCHLIFKE,1a A. TOFT SØRENSEN,1a

. LEDRI,a E. KUTEEVA,b T. HÖKFELTb AND

. KOKAIAa*

Experimental Epilepsy Group, Wallenberg Neuroscience Center,MC A-11, Lund University Hospital, 221 84 Lund, Sweden

Department of Neuroscience, Retzius väg 8, Karolinska Institutet, 1717 Stockholm, Sweden

bstract—Gene therapy–based overexpression of endoge-ous seizure-suppressing molecules represents a promisingreatment strategy for epilepsy. Viral vector–based overexpres-ion of the neuropeptide galanin has been shown to effectivelyuppress generalized seizures in various animal models of ep-

lepsy. However, it has not been explored whether such treat-ent can also prevent the epileptogenesis. Using a recombi-ant adeno-associated viral (rAAV) vector, we induced hip-ocampal galanin overexpression under the neuron specificnolase promoter in rats. Here we report that in animals withalanin overexpression, the duration of electrographic afterdis-harges was shortened and initiation of convulsions was de-

ayed at generalized seizure stages. However, the hippocampalindling development was unchanged. Short-term plasticity ofossy fiber–cornu ammonis (CA) 3 synapses was unaltered, as

ssessed by paired-pulse and frequency facilitation of fieldxcitatory postsynaptic potentials (fEPSPs) in hippocampallices, suggesting that despite high transgene galanin expres-ion, overall release probability of glutamate in these synapsesas unaffected. These data indicate that hippocampal rAAV-ased galanin overexpression is capable of mediating anticon-ulsant effects by lowering the seizure susceptibility once gen-ralized seizures are induced, but does not seem to affect kin-ling development or presynaptic short-term plasticity inossy fibers. © 2007 IBRO. Published by Elsevier Ltd. All rights

eserved.

ey words: galanin, epilepsy, gene therapy, rAAV, kindling,ynaptic plasticity.

he neuropeptide galanin, first isolated from porcine intes-ine (Tatemoto et al., 1983), is widely expressed in the

Corresponding author. Tel: �46-0-46-2220547; fax: �46-0-46-2220560.-mail address: merab.kokaia@med.lu.se (M. Kokaia).bbreviations: aCSF, artificial cerebrospinal fluid; AD, afterdis-harge; AP, anterior–posterior; CA, cornu ammonis; CMV, cyto-egalovirus; DCG-IV, (2S,2=R,3=R)-2-(2=,3=-dicarboxycyclopropyl)lycine; DG, dentate gyrus; EEG, electroencephalogram/electroen-ephalographic; fEPSP, field excitatory postsynaptic potential; FF, fre-uency facilitation; FIB, fibronectin; GalOE, galanin overexpressing;alR, galanin receptor; IR, immunoreactivity; ISI, interstimulus interval;A, kainic acid; ML, medio-lateral; NSE, neuron specific enolase; PC,iriform cortex; PDGF-B, platelet-derived growth factor B; PPF, paired-ulse facilitation; Pr, release probability; rAAV, recombinant adeno-asso-iated virus; SE, status epilepticus; SSC, standard saline citrate; V, ven-

wral.These authors contributed equally to this work.

306-4522/07$30.00�0.00 © 2007 IBRO. Published by Elsevier Ltd. All rights reseroi:10.1016/j.neuroscience.2007.09.056

1

eripheral nervous system and CNS, and is involved in aariety of brain functions such as food intake (Kalra andorvath, 1998; Leibowitz, 2005), cognition (Crawley,996), mood (Weiss et al., 1998; Barrera et al., 2005), andain sensation (Xu et al., 2000; Liu and Hokfelt, 2002). In

he rat hippocampus, a common area of seizure generation inemporal lobe epilepsy, galanin is present in afferent cholin-rgic and noradrenergic fibers originating from the septum–asal forebrain complex and the locus coeruleus, respec-

ively (Melander et al., 1986; Gabriel et al., 1995; Xu et al.,998). Among the three G-protein-coupled galanin receptorsGalRs) cloned (Branchek et al., 2000), subtypes 1 and 2 arexpressed in the hippocampus (Lu et al., 2005), withalR1 being primarily found in the ventral cornu ammonis

CA) areas, and GalR2 in the ventral and dorsal dentateyrus (DG) (O’Donnell et al., 1999; Burazin et al., 2000).

Following experimental seizures, galanin-immunoreac-ivity (IR) disappears in hippocampal fibers normally con-aining galanin (Mazarati et al., 1998), indicating depletionf stored galanin. At the same time, seizure activity in-uces galanin mRNA expression in the entire hippocam-us (Wilson et al., 2005), as well as de novo galaninynthesis in interneurons of the polymorphic cell layer ofhe DG, as revealed by galanin-IR (Mazarati et al., 1998).hese changes have led to the suggestion that galaninight play a role in epilepsy. Indeed, infusion of galanin

nto the DG of rats before or during self-sustained statuspilepticus (SE) was able to shorten or completely abolisheizures (Mazarati et al., 1998; Mazarati and Wasterlain,002). In addition, galanin knockout mice were shown toe more susceptible to SE, whereas mice overexpressingalanin under the dopamine beta-hydroxylase (DBH) pro-oter were more resistant to the induction of SE (Mazaratit al., 2000). In line with these findings, we have previouslyhown that transgenic mice ectopically overexpressing ga-

anin under the platelet-derived growth factor B (PDGF-B)romoter exhibited delayed epileptogenesis induced byindling stimulations (Kokaia et al., 2001). In addition, elec-rophysiological recordings in hippocampal slices from theseice revealed that ectopically overexpressed galanin wasost likely released during repetitive high-frequency afferent

timulation causing reduced glutamate release via GalR-ediated action (Kokaia et al., 2001). Other studies havelso shown that binding of galanin to GalRs can modulateeuronal excitability. Activation of GalR subtypes 1 and 3auses the opening of ATP-dependent K� channels (Zini etl., 1993; Kask et al., 1997) and/or blockade of voltage-ated Ca2� channels (Palazzi et al., 1991). These are

ell-known mechanisms of inhibiting presynaptic gluta-

ved.

msGiarG

rbataloatckslsldmckigoCopues(

istabsp(mPNa

A

TS2kmt

iiEw

V

V2ma(mvaThg3tv�c(ipmtpemvb0

E

Talil(ewt

E

KwcbsdesbhbaVeh

E

V

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx2

ARTICLE IN PRESS

ate release, which could provide an explanation for theeizure suppressing effects of galanin. While GalR1 andalR3 probably suppress seizures through presynaptic

nhibition of excitatory transmission, the GalR2 possiblylso inhibits seizures via a postsynaptic component (Maza-ati, 2004). However, the mechanisms of action of thealR2 remain unclear (Mazarati, 2004; Lu et al., 2005).

Targeting galanin as an anticonvulsant molecule willequire a feasible approach to deliver the peptide into therain. Gene therapy represents such a novel approach,llowing local delivery and expression of genes for pep-

ides in the epileptic focus, with possibly less side effectsnd more efficacy as compared with conventional antiepi-

eptic drugs. Recently, it has been shown that transductionf hippocampal neurons with a recombinant adeno-ssociated viral (rAAV) vector encoding for galanin under

he neuron specific enolase (NSE) promoter could de-rease the time spent in seizures induced by hippocampalainic acid (KA) injection (Lin et al., 2003). In anothertudy, using a similar vector, but containing a cytomega-ovirus (CMV) promoter and the fibronectin (FIB) releaseequence, overexpressed and constitutively released ga-anin increased the threshold of wild running seizures in-uced electrically in the inferior collicular cortex (Haber-an et al., 2003). In addition, transduction of the piriform

ortex (PC) with rAAV–CMV–FIB–galanin in previouslyindled animals increased the threshold current for seizurenduction (McCown, 2006). Moreover, rAAV–CMV–FIB–alanin injection in the PC suppressed the generalizationf seizures induced by systemic KA-administration (Mc-own, 2006). These data demonstrate that local galaninverexpression mediated by a rAAV vector is able to sup-ress already generalized seizures. However, it remainsnknown whether such viral vector–induced galanin over-xpression could also delay kindling development, as washown in transgenic galanin overexpressing (GalOE) miceKokaia et al., 2001).

In the present study, we used an rAAV–NSE vector tonduce galanin overexpression in the hippocampus of ratsubsequently undergoing hippocampal kindling, a model ofemporal lobe epileptogenesis. We also explored possiblelterations in synaptic transmission and plasticity causedy transgene galanin in hippocampal mossy fiber–CA3ynapses. We chose these synapses since (i) we havereviously observed a reduction of frequency facilitationFF) of field excitatory postsynaptic potentials (fEPSPs) inossy fibers of mice overexpressing galanin under theDGF-B promoter (Kokaia et al., 2001), and (ii) rAAV–SE–galanin vector injections in the present study causedprominent galanin overexpression in the mossy fibers.

EXPERIMENTAL PROCEDURES

nimals

otally 15 adult male Sprague–Dawley rats (B&K, Sollentuna,weden), weighing 250 g at the beginning of the experiment, and2 pups (Sprague–Dawley; B&K) from two litters were used forindling and electrophysiological recordings, respectively. Ani-als were housed at a 12-h light/dark cycle with ad libitum access

o food and water. All experiments were performed according to a

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

nternational guidelines on the use of animals as well as to Swed-sh Animal Welfare Agency guidelines and approved by the localthical Committee for Experimental Animals. All experimentsere designed to minimize the number and suffering of animals.

iral vector injection

iral vectors were produced as described elsewhere (Lin et al.,003). Adult animals, subsequently used for the kindling experi-ent, were anesthetized by i.p. injection of ketamine (80 mg/kg)nd xylazine (15 mg/kg), and placed into a Kopf stereotaxic frameDavid Kopf Instruments, Tujunga, CA, USA). Drill holes wereade in the skull at the designated coordinates (see below), and

iral vector suspensions were injected through a glass pipettettached to a 5 �l Hamilton syringe (Hamilton, Reno, NV, USA).o achieve maximal spread of the virus throughout the adultippocampus, viruses (with genomic titers of 1�1012 for rAAV-alanin and 1�1013 for rAAV-empty, respectively) were diluted:1 with heparin (2000 IU/ml) and injected bilaterally in two sites ofhe dorsal (anterior–posterior (AP) �3.6, medio-lateral (ML) �2.0,entral (V) �3.0 and �2.1) and ventral (AP �4.8, ML �5.2, V6.5 and �3.8) hippocampus (in mm). Reference points for all

oordinates were bregma, midline and dura, tooth bar at �3.3 mmPaxinos and Watson, 1996). A schematic presentation of thenjection sites is shown in Fig. 1A. On each side of the hippocam-us, 6 �l of viral vector dilution was infused at a rate of 0.2 �l perin. After injection, the pipette was left in place for additional 3 min

o prevent backflow through the injection track when retracting theipette. Newborn animals (postnatal days 1–3), later used forlectrophysiological recordings, were anesthetized by cooling andounted on a pup tray attached to the stereotaxic frame. Pure

iral vector was injected into the DG, at (in mm): AP �1.5 (fromregma), ML �1.9 (from midline) and V �1.8 (from dura). Totally.5 �l viral vector was infused at a rate of 0.2 �l per min.

lectrode implantation

wo weeks following viral vector injection, adult animals werenesthetized as described above. A bipolar stainless-steel stimu-

ating/recording electrode (PlasticsOne, Roanoke, VA, USA) wasmplanted stereotaxically into the ventral hippocampus at the fol-owing coordinates (in mm): AP �4.8 (from bregma), ML �5.2from midline), V �6.3 (from dura), tooth bar at �3.3. A referencelectrode was inserted into the temporal muscle, and electrodesere fixed to the skull with dental cement. Animals were allowed

o recover for 1 week before undergoing electrical stimulation.

lectrical kindling stimulation

indling stimulations (1 ms square wave pulses of 100 Hz for 1s)ere given once daily at an initially sub-convulsive thresholdurrent. The threshold was determined on the first days of kindlingy delivering stimulations at increasing current intensity in 10 �Ateps, starting with 10 �A, until focal epileptiform activity (after-ischarge, AD) of at least 5 s duration was detected by electro-ncephalographic (EEG) recording. Behavioral seizures werecored according to the scale of Racine (1972): stage 0, noehavioral changes; stage 1, facial twitches; stage 2, chewing andead nodding; stage 3, unilateral forelimb clonus; stage 4, rearing,ody jerks, bilateral forelimb clonus; stage 5, imbalance. EEGctivity was recorded on a MacLab system (ADInstruments, Bellaista, Australia) before, during, and at least 1 min after the end ofach AD. Animals were considered fully kindled after having ex-ibited five stage 5 seizures.

lectrophysiology

iral vector–injected rats were randomly selected 35–57 days

fter surgery, and following decapitation their brains were rapidly

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

pKgovpCcNaargsat

NP(frrbsIMdaotp

Fvfie same as

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx 3

ARTICLE IN PRESS

laced into ice-cold sucrose solution (in mM; 225 sucrose, 2.5Cl, 0.5 CaCl2, 7.0 MgCl2, 28 NaHCO3, 1.25 NaH2PO4, 7.0lucose, 1.0 ascorbate and 3.0 pyruvate; adjusted to pH 7.4;smolarity 300 mOsm; equilibrated with 95% O2/5% CO2). Trans-erse slices (400 �m) from the dorsal hippocampus were pre-ared on a Vibratome (Vibratome 3000, Ted Pella, Inc., Redding,A, USA) using the same solution. Slices were stored in a holdinghamber containing artificial cerebrospinal fluid (aCSF) (mM; 119aCl, 2.5 KCl, 1.3 MgSO4, 2.5 CaCl2, 26 NaHCO3, 1.0 NaH2PO4,nd 11 glucose; pH 7.4; 296 mOsm) oxygenated at room temper-ture, and after �1 h of resting slices were transferred to aecording chamber constantly perfused (2 ml per min) with oxy-enated aCSF containing picrotoxin (PTX; 100 �M; Tocris Cook-on Ltd., Avonmouth, UK). Mossy fiber responses were evoked bybipolar stimulation electrode placed in the stratum lucidum close

ig. 1. Transduction with rAAV–NSE–galanin vector induces transgeniral vector injection sites, denoted by stars (A). Galanin mRNA in situollowing injection of rAAV–NSE–galanin (B, C) or rAAV–NSE–empmmunohistochemistry in the dorsal (F, H) and ventral (G, I) hippocammpty (H, I). Scale bars�500 �m in F (same as H), and 1 mm in G (

o the hilus of the DG and recorded by a pipette filled with 3 M t

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

aCl (resistance: 0.5–1 M�) placed in CA3 stratum lucidum.aired-pulse stimulations were delivered at interstimulus intervals

ISIs) of 25, 50, 100 and 200 ms (at 0.33 Hz) and paired-pulseacilitation (PPF) was calculated as the mean of five consecutiveesponses by measuring the initial slope (1–2 ms) of the secondesponse as compared with the first response. FF was assessedy alternating the stimulation frequency from 0.1 Hz to 1 Hz (10timuli at 1 Hz), and was calculated as the evoked paired-pulses.n some experiments, the putative non-specific GalR antagonist35 (Bartfai et al., 1992) (1 �M; Sigma-Aldrich, Stockholm, Swe-en) was added to the perfusion medium. To distinguish betweenfferent associational-commissural fibers and mossy fibers, whichriginate from different presynaptic neurons, the group II metabo-ropic glutamate agonist, (2S,2=R,3=R)-2-(2=,3=-dicarboxycyclo-ropyl)glycine (DCG-IV; 3 �M; Tocris Cookson Ltd.) was added to

overexpression in the hippocampus. A schematic presentation of thetion in the dorsal (B, D) and ventral (C, E) portion of the hippocampus. The dorsal hippocampus is highlighted by a frame in B. Galaninnimals transduced with rAAV–NSE–galanin (F, G) and rAAV–NSE–I), respectively.

e galaninhybridizaty (D, E)pus in a

he perfusion medium at the end of all recordings. Since the group

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

Immcs

I

Tmdahsri1aoB(prnssIo

I

Aatgsh(drmbbD(opwbwFpcaScJn

S

SfuMadri

rt

Thts–fmgu(aewTwn(pnmmmwpoane1t

wvaitppdr

ssrrswN

rdc

Vs

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx4

ARTICLE IN PRESS

I metabotropic glutamate receptor is only expressed on theossy fibers (Kirschstein et al., 2004), responses blocked byore than 95% were considered as mossy fibers fEPSPs. Re-

ordings and analysis were conducted using HEKA amplifier andoftware (HEKA Elektronik, Lambrecht, Germany).

mmunohistochemistry

wenty-four hours after the last seizure, i.e. approximately 2onths following viral vector injection, kindled animals wereeeply anesthetized with pentobarbital and perfused through thescending aorta with 0.9% NaCl followed by 4% paraformalde-yde. Brains were post-fixed (overnight, 4 °C), transferred to 20%ucrose and stored overnight (4 °C). Brains were cut on a mic-otome (30 �m thick slices), and free-floating sections were rinsedn KPBS and incubated overnight (at room temperature) in a:2000 dilution of rabbit antiserum to rat galanin (Theodorssonnd Rugarn, 2000). Slices were rinsed, and incubated with sec-ndary biotinylated goat-anti rabbit antibody (Vector Laboratories,urlingame, CA, USA) followed by the Alexa-488 fluorophore

Molecular Probes, Leiden, Netherlands). Slices used for electro-hysiological recordings were post-fixed in 4% paraformaldehyde,insed, and quenched in 3% hydrogen peroxide and 10% metha-ol before incubation with galanin antiserum (see above). Finallylices were incubated in biotinylated secondary antibody, and thetaining was visualized by 3–3=-diaminobenzidine (DAB) reaction.n all examined brain slices, no damage caused by either surgeryr virus was observed.

n situ hybridization

nimals were decapitated 6 h after the last kindling stimulation,nd brains were frozen on dry ice and subsequently cut into 14 �m

hick slices on a cryotome. Oligonucleotides complementary to ratalanin nucleotides 152–199 (Vrontakis et al., 1987) were used totudy galanin mRNA expression in the hippocampus by in situybridization. Oligonucleotide probe was labeled with 33P (dATP)PerkinElmer, Waltham, MA, USA) at the 3=-end using terminaleoxynucleotidyltransferase (Amersham, Amersham, UK) and pu-ified using ProbeQuantTM G-50 microcolumns (Amersham Phar-acia Biotech, Piscataway, NJ, USA). Probe was added to hy-ridization solution containing 50% of deionized formamide (Am-ion Inc., Austin, TX, USA), 4� standard saline citrate (SSC), 1�enhardt’s solution and 50 mg/l denaturated salmon testis DNA

Sigma-Aldrich, St. Louis, MO, USA). Sections were air-driedvernight and hybridized with the oligonucleotide probe (0.5 nger slide) for 16–24 h at 42 °C. After hybridization the sectionsere rinsed for 2 h (4�30 min) in 1� SSC at 55–60 °C followedy 1 h at room temperature. Sections were then rinsed in distilledater, rapidly dehydrated in ethanol and air-dried for at least 2 h.or the control of specificity, an excess (100�) of the unlabeledrobe was added to the hybridization solution. This procedureompletely abolished the hybridization signal. The slides werenalyzed by phosphoimaging (Fujix BAS 3000 system, Fujicolorverige AB, Skärholmen, Sweden), and digital images were pro-essed in Photoshop 9.0.2 (Adobe System Incorporated, Sanose, CA, USA), and optimized for brightness, contrast and sharp-ess.

tatistical analysis

tatistical analysis of differences between the groups was per-ormed using Student’s unpaired t-test, and a paired t-test wassed to evaluate the effect of drug (before and after application of35). Differences were considered significant at P�0.05 and datare presented as mean�SEM. At all times, the investigator con-ucting the behavioral grading of animals, electrophysiologicalecordings or histological analysis was unaware of the group

dentity of individual animals. s

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

RESULTS

AAV mediates robust galanin overexpressionhroughout the hippocampus

o confirm that galanin was indeed overexpressed in theippocampus, approximately 2 months following viral vector

ransduction, subgroups of kindled animals were taken for initu hybridization (n�4 and 2 for rAAV–NSE–galanin andempty, respectively) and immunohistochemistry (n�4 and 5

or rAAV–NSE–galanin and –empty, respectively). In ani-als transduced with rAAV–NSE–galanin, an increase ofalanin mRNA was found in the entire hippocampus, partic-larly pronounced in the DG granule cell layer, in the rostralFig. 1B) as well as the caudal dorsal (Fig. 1C) hippocampus,s compared with animals transduced with rAAV–NSE–mpty (Fig. 1D, E). In the latter, in situ hybridization signalas not distinguishable from background levels (Fig. 1D, E).ransgene mRNA expression was selective to neurons,hich is in good agreement with the previously demonstratedeuronal tropism of rAAV serotype 2 vectors in generalShevtsova et al., 2005), and of the rAAV vector used here inarticular (Lin et al., 2003). Moreover, the predominantlyeuronal expression of the rAAV vector-transduced galaninRNA was also due to using the neuron-specific NSE pro-oter (Tenenbaum et al., 2004). In agreement with increasedRNA levels, in animals injected with rAAV–NSE–galanine found that galanin-IR was elevated throughout the hip-ocampus, with relatively higher levels in the mossy fibersriginating from dentate granule cells (Fig. 1F, G). In controlnimals injected with rAAV–NSE–empty, on the other hand,o galanin-IR could be detected (Fig. 1H, I). The absence ofndogenous galanin-IR fibers in the hippocampus (Xu et al.,998) of control animals was most likely due to sensitivity of

he immunostaining used here (without signal amplification).To evaluate whether transgene galanin expression

as present already at the time when kindling was started,iral vector injections were also performed in non-kindlednimals. Three weeks following these injections, rAAV-

nduced galanin-IR was assessed and was comparable tohat described in kindled animals (data not shown). Com-arable to our results, stable transgene galanin overex-ression in naïve rat hippocampus has previously beenescribed 2 months after transduction using the sameAAV–NSE–galanin vector (Lin et al., 2003).

Additionally, we assessed galanin-IR in transducedlices following electrophysiology, which also confirmedtrong and stable galanin expression in the area whereecordings were made (i.e. stratum lucidum of CA3) in allAAV–NSE–galanin-transduced slices (similar to thathown in Fig. 1A), whereas galanin-IR in control slicesithin the same area was low and resembled that in rAAV–SE–empty slices of kindled animals.

AAV-based galanin overexpression decreases theuration of ADs and increases the latency toonvulsions at kindled generalized seizure stages

iral vector–injected animals received daily electricaltimulations of initially subconvulsive intensity, and sub-

equently developed seizures with increasing severity.

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

H(dtNsNrmadl2gomrddlbcarg

Gm

Ntrlsswa3rralffratcaw

Fsi lation tos pty, resp

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx 5

ARTICLE IN PRESS

ere, we found that compared with the control animalsrAAV–NSE– empty; n�7), rAAV–NSE– galanin trans-uction of the hippocampus (n�8) did not alter thehreshold to seizure induction (40.0�5.3 �A for rAAV–SE– galanin and 41.4�8.0 for rAAV–NSE– empty, re-pectively) or the AD duration (26.3�2.6 s for rAAV–SE– galanin and 29.6�3.1 s for rAAV–NSE– empty,

espectively) on the first day of kindling. The develop-ent of kindling was not different between the groups,s manifested by mean seizure stages at consecutiveays of stimulation (Fig. 2A) and the number of stimu-

ations required to reach different seizure stages (Fig.B). Also the duration of convulsions was similar in bothroups, and was on average 26.3�2.6 s in animalsverexpressing galanin and 27.5�1.4 s in control ani-als, respectively. However, in animals injected with

AAV–NSE– galanin, the duration of epileptic EEG ADsuring generalized (i.e. stages 4 and 5) seizures wasecreased (Fig. 2C). Moreover, in the same group, the

atency time from the stimulation to onset of generalizedehavioral convulsions was prolonged (Fig. 2D), asompared with control animals. These data suggest thatlthough the kindling development was not affected,AAV–NSE– galanin did attenuate the expression of

ig. 2. Effect of rAAV–NSE–galanin injection on hippocampal kindlintimulations required to reach different seizure stages and the fifth sta.e. all stage 1, 2, 4 and 5 seizures. (D) Average latency from stimueizures. n�8 For rAAV–NSE–galanin and seven for rAAV–NSE–em

eneralized seizure activity. m

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

alanin overexpression does not alter plasticity atossy fiber–CA3 synapses

ext, we asked whether transgene galanin could alter exci-atory synaptic transmission in the hippocampus. SinceAAV–NSE-mediated galanin overexpression was particu-arly pronounced in mossy fibers, and since we have earlierhown that in transgenic mice overexpressing galanin in theame fibers, FF of fEPSPs is reduced (Kokaia et al., 2001),e measured short-term plasticity in mossy fiber–CA3 syn-pses in the dorsal hippocampus by fEPSP recordings (Fig.A). Paired stimulation of the mossy fibers in slices from bothAAV–NSE–empty- (n�18 slices from six animals) andAAV–NSE–galanin- (n�16 slices from five animals) injectednimals induced high levels of PPF of fEPSPs in stratum

ucidum of CA3 (Fig. 3B). Paired stimulations applied at dif-erent ISI generated PPF at various magnitudes, but no dif-erences were detected between rAAV–NSE–galanin- andAAV–NSE–empty-injected animals (Fig. 3B). In addition,pplication of the putative non-specific GalR antagonist, M35,

o slices did not have any effect on PPF in both treated andontrol slices (data not shown). In the same synapses, welso investigated FF of mossy fiber fEPSPs, which occurshen increasing the frequency of stimulation from low to

eizure severity during kindling development. (B) Average number ofure. (C) Average duration of ADs during the different seizure stages,

the onset of behavioral generalized convulsions during all stage 4/5ectively.

g. (A) Sge 5 seiz

oderate. A switch of stimulation frequency from 0.1 Hz to

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

1ffsiisbcatdfig

Om

emrtsfi

V

IagtrP(vw

Ffaatv m five anr differen

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx6

ARTICLE IN PRESS

Hz (10 pulses) resulted in about threefold increase inEPSP magnitude in both rAAV–NSE–galanin (n�13 slicesrom five animals) and rAAV–NSE–empty slices (n�17lices from six animals) (Fig. 4A). The evoked fEPSP exhib-ted similar magnitude in both groups (Fig. 4B). In addition, FFn both groups was unaffected by M35 application (data nothown). Perfusion with DCG-IV, which is able to specificallylock mossy fiber fEPSPs (see Experimental Procedures),onfirmed that the responses recorded in CA3 were gener-ted by activation of mossy fiber synapses, and not associa-ional-commissural synapses (Fig. 4A). In summary, theseata suggest that short-term plasticity of hippocampal mossyber synapses is unaffected by high levels of transgenealanin.

DISCUSSION

ur results show that rAAV–NSE–galanin vector-

ig. 3. PPF of fEPSPs generated by mossy fiber–CA3 synapses is unarom stratum lucidum of the CA3 area in hippocampal slices. The stimulrea illustrates the mossy fiber terminals originating from dentate grannimals injected with rAAV–NSE–galanin (B) or rAAV–NSE–empty (Che CA3 area where the recording electrode was placed. Scale bar�1arious ISI recorded in slices from rAAV–NSE–galanin (n�16 slices froats. (D) (Upper panel) Superimposed traces of recorded fEPSPs with

ediated galanin overexpression can partially inhibit gen- (

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

ralized kindled seizures without altering kindling develop-ent and seizure induction threshold. Electrophysiological

ecordings did not reveal any alterations in normal synapticransmission or short-term plasticity of mossy fiber–CA3ynapses despite high galanin expression in the mossybers.

iral vector–mediated galanin action in epilepsy

n previous studies it was shown that, when overexpressednd constitutively released using a rAAV–CMV–FIB–alanin vector in the inferior colliculus, galanin increasedhe threshold current for stimulation used to induce wild-unning seizures (Haberman et al., 2003). Similarly, in theC, the kindled seizure-induction threshold was elevated

McCown, 2006). However, when rAAV–CMV–galaninector (lacking the FIB release sequence, as in our study)as used, no effect on the seizure threshold was found

y overexpression of transgene galanin. (A) The fEPSPs were recordedtrode was placed in the same area, but closer toward the DG. Shadedxons. An example of typical galanin-IR in slices (400 �m thick) fromtively, used for electrophysiological recordings. Arrowheads indicate

n B (same as C). (D) (Lower panel) PPF of fEPSPs in percentage atimals) and rAAV–NSE–empty (n�18 slices from six animals) injectedt ISIs from a rAAV–NSE–galanin-injected rat.

ffected bation elecule cell a), respec00 �m i

Haberman et al., 2003). Since constitutive release of

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

ttdrichechokri

rfp(e1aebcsnvh

tviggp2pataqMposeaao2sbaasp

Fri z, expresr empty-

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx 7

ARTICLE IN PRESS

ransgene galanin occurs mainly when the vector containshe FIB release sequence (Haberman et al., 2003), theseata indicate that high extracellular levels of constitutivelyeleased galanin are required to increase the seizure-nduction threshold. In accordance with this notion, in-reased seizure threshold has been observed after intra-ippocampal infusion of a galanin agonist in rats (Mazaratit al., 2006). In contrast, when galanin is predominantlyontained within fibers and can only be released followingigh-frequency stimulation, such as in transgenic miceverexpressing galanin under the PDGF-B promoter (Ko-aia et al., 2001) or in our present experiments using aAAV–NSE–galanin vector in the rat hippocampus, thenitial seizure threshold is not changed.

In the present study, transgene galanin significantlyeduced the duration of ADs during generalized but notocal seizures. The AD duration typically increases with therogression of kindled seizures from focal to generalizedMorimoto et al., 2004). Accordingly, we observed an av-rage increase of the AD duration by 52�17% from stageto generalized stage 4/5 seizures in five of seven controlnimals. One could speculate that gradually increasingxtracellular accumulation of transgene galanin releasedy high frequency activity reaches the levels necessary tourtail seizure activity only during longer ADs and moreevere seizures. Such accumulation of extracellular gala-in released in activity-dependent manner does not pre-ent the spread and generalization of the seizures from the

ig. 4. FF of fEPSPs generated by mossy fiber–CA3 synapses is uncesponse to 0.1 Hz (1), at 1 Hz stimulation frequency after 10 consecutis indicated by the arrow. (B) FF of each 10 consecutive fEPSP at 1 HAAV–NSE–galanin- (n�13 slices from five animals) and rAAV–NSE–

ippocampus, although it seems to delay this process, e

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

hus increasing the latency to the onset of behavioral con-ulsions. Similar to our results, in seizures induced by

ntrahippocampal KA injection, the number and duration ofeneralized seizures was decreased by rAAV–NSE–alanin-mediated galanin overexpression in the hippocam-us, but the onset of seizures was not affected (Lin et al.,003). Interestingly, galanin has earlier been shown torominently suppress generalized seizure activity withoutffecting the initiation of seizures. Infusion of galanin intohe rat DG following perforant path stimulation completelynd irreversibly abolished seizure activity in the subse-uent self-sustained phase of SE (Mazarati et al., 1998;azarati and Wasterlain, 2002), but its infusion beforeerforant path stimulation did not prevent the occurrencef seizures (Mazarati et al., 1998). It has recently beenuggested that the effect of hippocampal galanin on differ-nt parameters of seizure activity is dependent on thectivation of the various receptor subtypes. The GalR1ppears to mediate increase in threshold and suppressionf initial seizure activity in both kindling (Mazarati et al.,006) and SE models (Mazarati and Lu, 2005). Suppres-ion of generalized seizures, on the other hand, is said toe mediated by GalR2 (Mazarati and Lu, 2005; Mazarati etl., 2006). Since transgene galanin in our experimentsffected predominantly generalized seizures, one couldpeculate that it acts via activation of GalR2. However,articipation of GalR1 in the observed effects cannot be

by transgene galanin. (A) Representative fEPSP traces are shown ini (2), and following DCG-IV application (3). The presynaptic fiber volleysed as percentage of average fEPSP slope at 0.1 Hz, obtained from

(n�17 slices from six animals) injected rats.

hangedve stimul

xcluded.

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

Gt

TipftIoalpspSe1saGtcncwieqa2t2emsagsbrttsredml

Iertewtm

ctstkttd

ASCtpfo

B

B

B

B

B

C

D

G

H

K

K

K

K

L

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx8

ARTICLE IN PRESS

alanin, short-term plasticity and synapticransmission

he effect of galanin on synaptic transmission has beennvestigated in various brain structures, including the hip-ocampus, where it mediates predominately inhibitory ef-ects by hyperpolarizing cells via changing the conduc-ance of potassium or calcium channels (Xu et al., 2005).n the present study, we aimed to evaluate whether effectsf transgene galanin on seizures are mediated by alter-tions in plasticity of mossy-fiber synapses in the stratum

ucidum of CA3. Excitatory synaptic transmission in hip-ocampal mossy fiber synapses is distinct from most otherynapses and is characterized by its low overall releaserobability (Pr) of glutamate (Nicoll and Schmitz, 2005).ynapses with low Pr usually exhibit high PPF and FF ofxcitatory postsynaptic responses (Dobrunz and Stevens,997), as was also the case for fEPSPs in the presenttudy. However, both PPF and FF were similar in galaninnd empty vector-treated animals, and application of thealR receptor antagonist M35 did not change the facilita-

ion ratios. These data suggest that in these experimentalonditions, activity-dependent release of transgene gala-in from mossy fiber terminals is not sufficient to inducehanges in short-term synaptic plasticity, opposite to whatas observed in GalOE mice (Kokaia et al., 2001). This

nterpretation is supported by previous publications wherexogenously applied galanin was able to modify high fre-uency stimulation–induced EPSCs generated by syn-pses on pyramidal neurons of the PC (Schlifke et al.,006), as well as PPF of EPSPs generated by synapses ofhe perforant path on dentate granule cells (Bartfai et al.,004) in mice. One could speculate that the presence ofctopic galanin in mossy fibers already at early develop-ental stages, as in GalOE mice, could activate the re-

pective release machinery, which is then maintained indults. In contrast, viral vector–induced overexpression ofalanin directly in adult animals could fail to do so to theame extent, thus limiting the amount of galanin releasedy high frequency activity in these animals. Inability toelease sufficient amounts of transgene galanin could behe reason for absence of its modulatory effect on short-erm plasticity in mossy fiber synapses. Finally, possiblepecies and strain differences accounting for the differentesults should also be taken into account. The antiepilepticfficacy of transgene galanin released constitutively asescribed for the rAAV–CMV–FIB–galanin vector (Haber-an et al., 2003; McCown, 2006), would shed some more

ight on this issue.

CONCLUSIONS

n summary, our data suggest that transgene galanin over-xpressed by rAAV–NSE–galanin vector is predominantlyeleased during severe, high frequency epileptic activity inhe hippocampus and attenuates seizure activity by short-ning AD duration and delaying initiation of convulsions,hile kindling development is not affected. Normal synap-

ic transmission and short-term presynaptic plasticity in

ossy fibers are not changed, most likely due to lack of

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

onstitutive or low frequency activity-dependent release ofransgene galanin from these synapses. The present studyupports the idea that rAAV–NSE vector-based generansfer for galanin, although less effective for inhibitingindling and thus perhaps epileptogenesis, has some po-ential to attenuate already established generalized epilep-ic syndromes, and could in the future be optimized andeveloped into a possible novel therapeutic approach.

cknowledgments—This work was supported by grants from thewedish Research Council, the Segerfalk Foundation, therafoord Foundation, The Lars Hierta Memorial Foundation and

he Kock’s Foundation. We thank Matthew J. During for kindlyroviding the viral vectors (Lin et al., 2003), and Monica Lundahlor excellent technical assistant. We also thank Prof. Elvar The-dorsson, Linköping University, for providing the galanin antibody.

REFERENCES

arrera G, Echevarria DJ, Poulin JF, Laforest S, Drolet G, Morilak DA(2005) One for all or one for one: does co-transmission unify theconcept of a brain galanin “system” or clarify any consistent role inanxiety?Neuropeptides 39:289–292.

artfai T, Fisone G, Langel U (1992) Galanin and galanin antagonists:molecular and biochemical perspectives. Trends Pharmacol Sci13:312–317.

artfai T, Lu X, Badie-Mahdavi H, Barr AM, Mazarati A, Hua XY,Yaksh T, Haberhauer G, Ceide SC, Trembleau L, Somogyi L,Krock L, Rebek J Jr (2004) Galmic, a nonpeptide galanin receptoragonist, affects behaviors in seizure, pain, and forced-swim tests.Proc Natl Acad Sci U S A 101:10470–10475.

ranchek TA, Smith KE, Gerald C, Walker MW (2000) Galanin recep-tor subtypes. Trends Pharmacol Sci 21:109–117.

urazin TC, Larm JA, Ryan MC, Gundlach AL (2000) Galanin-R1 and-R2 receptor mRNA expression during the development of rat brainsuggests differential subtype involvement in synaptic transmissionand plasticity. Eur J Neurosci 12:2901–2917.

rawley JN (1996) Minireview. Galanin-acetylcholine interactions:relevance to memory and Alzheimer’s disease. Life Sci 58:2185–2199.

obrunz LE, Stevens CF (1997) Heterogeneity of release probabil-ity, facilitation, and depletion at central synapses. Neuron 18:995–1008.

abriel SM, Knott PJ, Haroutunian V (1995) Alterations in cerebralcortical galanin concentrations following neurotransmitter-specificsubcortical lesions in the rat. J Neurosci 15:5526–5534.

aberman RP, Samulski RJ, McCown TJ (2003) Attenuation of sei-zures and neuronal death by adeno-associated virus vector gala-nin expression and secretion. Nat Med 9:1076–1080.

alra SP, Horvath TL (1998) Neuroendocrine interactions betweengalanin, opioids, and neuropeptide Y in the control of reproductionand appetite. Ann N Y Acad Sci 863:236–240.

ask K, Berthold M, Bartfai T (1997) Galanin receptors: involve-ment in feeding, pain, depression and Alzheimer’s disease. LifeSci 60:1523–1533.

irschstein T, von der Brelie C, Steinhauser M, Vincon A, Beck H,Dietrich D (2004) L-CCG-I activates group III metabotropic gluta-mate receptors in the hippocampal CA3 region. Neuropharmacol-ogy 47:157–162.

okaia M, Holmberg K, Nanobashvili A, Xu ZQ, Kokaia Z, LendahlU, Hilke S, Theodorsson E, Kahl U, Bartfai T, Lindvall O, HokfeltT (2001) Suppressed kindling epileptogenesis in mice with ec-topic overexpression of galanin. Proc Natl Acad Sci U S A98:14006 –14011.

eibowitz SF (2005) Regulation and effects of hypothalamic galanin:relation to dietary fat, alcohol ingestion, circulating lipids and en-

ergy homeostasis. Neuropeptides 39:327–332.

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056

L

L

L

M

M

M

M

M

M

M

M

M

N

O

P

P

R

S

S

T

T

T

V

W

W

X

X

X

Z

I. Kanter-Schlifke et al. / Neuroscience xx (2007) xxx 9

ARTICLE IN PRESS

in EJ, Richichi C, Young D, Baer K, Vezzani A, During MJ (2003)Recombinant AAV-mediated expression of galanin in rat hip-pocampus suppresses seizure development. Eur J Neurosci18:2087–2092.

iu HX, Hokfelt T (2002) The participation of galanin in pain processingat the spinal level. Trends Pharmacol Sci 23:468–474.

u X, Mazarati A, Sanna P, Shinmei S, Bartfai T (2005) Distributionand differential regulation of galanin receptor subtypes in rat brain:effects of seizure activity. Neuropeptides 39:147–152.

azarati A, Lundstrom L, Sollenberg U, Shin D, Langel U, Sankar R(2006) Regulation of kindling epileptogenesis by hippocampal ga-lanin type 1 and type 2 receptors: The effects of subtype-selectiveagonists and the role of G-protein-mediated signaling. J PharmacolExp Ther 318:700–708.

azarati AM (2004) Galanin and galanin receptors in epilepsy. Neu-ropeptides 38:331–343.

azarati AM, Hohmann JG, Bacon A, Liu H, Sankar R, Steiner RA,Wynick D, Wasterlain CG (2000) Modulation of hippocampal ex-citability and seizures by galanin. J Neurosci 20:6276–6281.

azarati AM, Liu H, Soomets U, Sankar R, Shin D, Katsumori H,Langel U, Wasterlain CG (1998) Galanin modulation of seizuresand seizure modulation of hippocampal galanin in animal modelsof status epilepticus. J Neurosci 18:10070–10077.

azarati AM, Lu X (2005) Regulation of limbic status epilepticus byhippocampal galanin type 1 and type 2 receptors. Neuropeptides39:277–280.

azarati AM, Wasterlain CG (2002) Anticonvulsant effects of fourneuropeptides in the rat hippocampus during self-sustaining statusepilepticus. Neurosci Lett 331:123–127.

cCown TJ (2006) Adeno-associated virus-mediated expression andconstitutive secretion of galanin suppresses limbic seizure activityin vivo. Mol Ther 14:63–68.

elander T, Staines WA, Rokaeus A (1986) Galanin-like immunore-activity in hippocampal afferents in the rat, with special reference tocholinergic and noradrenergic inputs. Neuroscience 19:223–240.

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

icoll RA, Schmitz D (2005) Synaptic plasticity at hippocampal mossyfibre synapses. Nat Rev Neurosci 6:863–876.

’Donnell D, Ahmad S, Wahlestedt C, Walker P (1999) Expression ofthe novel galanin receptor subtype GALR2 in the adult rat CNS:distinct distribution from GALR1. J Comp Neurol 409:469–481.

alazzi E, Felinska S, Zambelli M, Fisone G, Bartfai T, Consolo S(1991) Galanin reduces carbachol stimulation of phosphoinosi-tide turnover in rat ventral hippocampus by lowering Ca2� influxthrough voltage-sensitive Ca2� channels. J Neurochem 56:

739 –747.

Please cite this article in press as: Kanter-Schlifke I, et al., Galanin genaltering hippocampal synaptic plasticity., Neuroscience (2007), doi: 10.1

axinos G, Watson C (1996) The rat brain in stereotaxic coordinates.New York: Academic Press.

acine RJ (1972) Modification of seizure activity by electrical stimula-tion. II. Motor seizure. Electroencephalogr Clin Neurophysiol32:281–294.

chlifke I, Kuteeva E, Hokfelt T, Kokaia M (2006) Galanin expressedin the excitatory fibers attenuates synaptic strength and general-ized seizures in the piriform cortex of mice. Exp Neurol200:398–406.

hevtsova Z, Malik JM, Michel U, Bahr M, Kugler S (2005) Promotersand serotypes: targeting of adeno-associated virus vectors forgene transfer in the rat central nervous system in vitro and in vivo.Exp Physiol 90:53–59.

atemoto K, Rokaeus A, Jornvall H, McDonald TJ, Mutt V (1983)Galanin: a novel biologically active peptide from porcine intestine.FEBS Lett 164:124–128.

enenbaum L, Chtarto A, Lehtonen E, Velu T, Brotchi J, Levivier M(2004) Recombinant AAV-mediated gene delivery to the centralnervous system. J Gene Med 6(Suppl 1):S212–S222.

heodorsson E, Rugarn O (2000) Radioimmunoassay for ratgalanin: immunochemical and chromatographic characterizationof immunoreactivity in tissue extracts. Scand J Clin Lab Invest60:411– 418.

rontakis ME, Peden LM, Duckworth ML, Friesen HG (1987) Isolationand characterization of a complementary DNA (galanin) clone fromestrogen-induced pituitary tumor messenger RNA. J Biol Chem262:16755–16758.

eiss JM, Bonsall RW, Demetrikopoulos MK, Emery MS, West CH(1998) Galanin: a significant role in depression?Ann N Y Acad Sci863:364–382.

ilson DN, Chung H, Elliott RC, Bremer E, George D, Koh S (2005)Microarray analysis of postictal transcriptional regulation of neu-ropeptides. J Mol Neurosci 25:285–298.

u XJ, Hokfelt T, Bartfai T, Wiesenfeld-Hallin Z (2000) Galanin andspinal nociceptive mechanisms: recent advances and therapeuticimplications. Neuropeptides 34:137–147.

u ZQ, Shi TJ, Hokfelt T (1998) Galanin/GMAP- and NPY-like immu-noreactivities in locus coeruleus and noradrenergic nerve termi-nals in the hippocampal formation and cortex with notes on thegalanin-R1 and -R2 receptors. J Comp Neurol 392:227–251.

u ZQ, Zheng K, Hökfelt T (2005) Electrophysiological studies ongalanin effects in brain-progress during the last 6 years. Neuropep-tides 39:269–275.

ini S, Roisin MP, Langel U, Bartfai T, Ben-Ari Y (1993) Galaninreduces release of endogenous excitatory amino acids in the rat

hippocampus. Eur J Pharmacol 245:1–7.

(Accepted 24 September 2007)

e transfer curtails generalized seizures in kindled rats without016/j.neuroscience.2007.09.056