encapsulated galanin-producing cells attenuate focal epileptic seizures in the hippocampus

Encapsulated galanin-producing cells attenuate focal

epileptic seizures in the hippocampus*Litsa Nikitidou, †Malene Torp, †Lone Fjord-Larsen, †Philip Kusk, †Lars U.Wahlberg, and

*M�erab Kokaia

Epilepsia, 55(1):167–174, 2014doi: 10.1111/epi.12470

Litsa Nikitidoufinished her PhD inSweden and iscurrently apostdoctoralresearcher in Hungary.

SUMMARY

Purpose: Encapsulated cell biodelivery (ECB) is a relatively safe approach, since the

devices can be removed in the event of adverse effects. The main objectives of the

present study were to evaluate whether ECB could be a viable alternative of cell ther-

apy for epilepsy.We therefore developed a human cell line producing galanin, a neuro-

peptide that has been shown to exert inhibitory effects on seizures, most likely acting

via decreasing glutamate release from excitatory synapses. To explore whether ECB

of genetically modified galanin-producing human cell line could provide seizure-sup-

pressant effects, and test possible translational prospect for clinical application, we

implanted ECB devices bilaterally into the hippocampus of rats subjected to rapid

kindling, amodel for recurrent temporal lobe seizures.

Methods: Two clones from a genetically modified human cell line secreting different

levels of galanin were tested. Electroencephalography (EEG) recordings and stimula-

tions were performed by electrodes implanted into the hippocampus at the same sur-

gical session as ECB devices. One week after the surgery, rapid kindling stimulations

were initiated.

Key Findings: Enzyme-linked immunosorbent assay (ELISA) measurements prior to

device implantation showed a release of galanin on average of 8.3 ng/mL/24 h per

device for the low-releasing clone and 12.6 ng/mL/24 h per device for the high-releas-

ing clone. High-releasing galanin-producing ECB devices moderately decreased stimu-

lation-induced focal afterdischarge duration, whereas low-releasing ECB devices had

no significant effect.

Significance: Our study shows that galanin-releasing ECB devicesmoderately suppress

focal stimulation-induced recurrent seizures. Despite this moderate effect, the study

provides conceptual proof that ECB could be a viable alternative approach to cell ther-

apy in humans, with the advantage that the treatment could be terminated by remov-

ing these devices from the brain. Thereby, this strategy provides a higher level of

safety for future therapeutic applications, in which genetically modified human cell

lines that are optimized to produce and release antiepileptic compounds could be clini-

cally evaluated for their seizure-suppressant effects.

KEY WORDS: Galanin, Epilepsy, Kindling, Hippocampus, Encapsulated cells, ECB

device.

Nonpathogenic viral vector–based gene delivery intothe brain has been proven to be a safe procedure inphase 1–2 clinical trials, for example, for Parkinson’sdisease.1–3 However, once the transgene of interest isexpressed in the host brain cells, it is impossible toreverse the process and to terminate its action in theevent that adverse effects arise. Encapsulated cell biode-livery (ECB) devices filled with genetically modified

Accepted October 14, 2013; Early View publication November 18, 2013.*Experimental Epilepsy Group, Wallenberg Neuroscience Center, BMC

A-11, Lund University Hospital, Lund, Sweden; and †NsGene A/S, Ballerup,Denmark

Address correspondence to M�erab Kokaia, Experimental EpilepsyGroup, Wallenberg Neuroscience Center, BMC A-11, Lund UniversityHospital, Lund, Sweden. E-mail: [email protected]

Wiley Periodicals, Inc.© 2013 International League Against Epilepsy

167

FULL-LENGTHORIGINALRESEARCH

human cells to release gene products into the host tissuehave an advantage of being a reversible treatment: TheECB devices can be removed from the brain with a rela-tively simple procedure and thereby terminate the exertedeffect. The cells in the ECB devices can have long-termviability when implanted into the brain because the nutri-ents from the surrounding host tissue can penetrate thesemipermeable membrane of the ECB device, and at thesame time, the gene products can be released into thehost tissue. The advantages of ECB devices as comparedto direct genetic modification of the host cells by viralvector–based gene delivery, or direct cell transplantation,are that the encapsulated cells do not alter host cells orintegrate into the host brain. Furthermore, the semiper-meable membrane isolates the cells in the ECB devicesfrom immune reactions against them in the host brain.Therefore, there is no need for immunosuppressant drugs.These considerations make it highly warranted from thetranslational perspective to explore possible therapeuticeffects of compounds delivered into the brain by usingECB technology.

The neuropeptide galanin was first discovered in por-cine intestine,4 but later it has been found in various partsof the body, including the peripheral nervous system(PNS) and the central nervous system (CNS). Galanin hasdiverse physiologic functions in the normal brain, but ithas also been implicated in pathophysiologic conditions,for example depression,5,6 Alzheimer’s disease,7–10 andepilepsy.11–14

Galanin signaling occurs through G-protein–coupled gal-anin receptor 1 (GalR1), 2 (GalR2), and 3 (GalR3).15–17

GalR1 and GalR2 are expressed within the hippocampus.18

The mechanism of action of galanin through the galaninreceptors is thought to be mediated through blockade ofvoltage-gated Ca2+ channels and/or activation of ATP-dependent K+ channels.19–21

Several studies suggest that galanin is involved in sei-zure regulation and can modulate epileptic activity in thebrain. During the epileptic seizures, galanin is releasedand exerts a presynaptic inhibitory effect on the glutama-tergic transmission.11,12,20 In galanin-overexpressingtransgenic mice or rats, in which galanin is overexpres-sed by gene transduction, prolonged latent period to theconvulsions and decreased susceptibility to generalizedseizures have been observed in a kindling model of epi-lepsy.14,22 In addition a viral vector–based gene therapyapproach has demonstrated a powerful seizure-suppres-sant effect of transgene galanin in other animal modelsof epilepsy, such as chemically and electrically inducedstatus epilepticus.22,23

The main objective of the present study was toexplore the therapeutic potential of intrahippocam-pally engrafted galanin-releasing ECB devices in an ani-mal model of stimulation-induced recurrent epilepticseizures.

Experimental ProceduresEthics statement

All experimental procedures were approved by the localMalm€o/Lund Ethical Committee for Experimental Animals(Ethical permit number M187-09), and were performedaccording to the guidelines of the Swedish Animal WelfareAgency and in agreement with international guidelines.

In vitro preparation and filling of ECB devicesECB devices consisted of a semipermeable polyether sul-

fone (PES) hollow fiber membrane filled with a polyvinylalcohol (PVA) cylindrical matrix serving as support for theencapsulated cells. Devices were built and sterilized beforethey were filled with low-passage human retinal pigmentepithelial cell line (ARPE-19). The cells and the filled ECBdevices were cultured in human endothelial serum-freemedia (HE-SFM) (Invitrogen, Stockholm, Sweden) in anincubator (37°C, 5%CO2). Several cell lines were generatedby genetic modification to release galanin. Two cell clonesthat produced and released galanin were selected, one with ahigher release (HR) and one with a lower release (LR) ofgalanin. The ECB devices were 5 mm long for the verticalplacement in the hippocampus, and 7 mm long for theangular placement (see below). Both had an outer diameterof 725 lm and an inner diameter of 525 lm. Each ECBdevice was filled with 60,000 cells (optimized by pilotexperiments). The ECB devices with nonmodified ARPE-19 cells (not producing galanin) and empty ECB deviceswere used as controls. Because no differences in any sei-zure parameters could be detected between groups treatedwith empty devices and ECB devices filled with controlcells (data not shown), the results from these two controlgroups were merged for analysis. One week and 3 weeksafter encapsulation, galanin release from the ECB devicesinto the incubation medium was sampled after 24 h andmeasured by a galanin enzyme-linked immunosorbentassay (ELISA; Bachem, Bubendorf, Switzerland).

AnimalsMale Sprague-Dawley rats (Charles River, Germany)

were used; these weighed 200–230 g at the beginning of theexperiment. The animals were housed individually at a 12 hlight/dark cycle with ad libitum access to food and water.All animals were weighed once a week throughout theexperiment.

Implantation of ECB devices and electrodeAnimals were anesthetized with isoflurane and fixed into

a stereotaxic frame (David Kopf Instruments, Tujunga, CA,U.S.A.). Three weeks after cell encapsulation, ECB deviceswere implanted bilaterally in two different positions. Wechose bilateral ECB implantation to prevent spread of theseizures from the hippocampus contralateral to the stimula-tion. One group of animals had the ECB devices implanted


168

L. Nikitidou et al.

in a straight vertical position and the other group at a26-degree angle, to ensure better coverage of the entirehippocampal axis. The following number of animals andgroups were used: Vertical placement of ECB devices—Empty n = 5, ARPE n = 4, LR n = 8, HR n = 5; Angularplacement of ECB devices—Empty n = 12, ARPE n = 12,LR n = 10, HR n = 7. The coordinates used for these twoECB device placements were as follows, reference pointsfrom bregma, midline. and dura: anteroposterior (AP)�4.8,mediolateral (ML) � 4.1, ventral (V) �6.0; and AP �5.3,ML � 2.7, V �8.0, respectively. Encapsulated cell biode-livery devices were stored in an incubator at 37°C (5% CO2)until implantation.

At the same surgical session a bipolar stainless steel stim-ulation/recording electrode (Plastics One, Roanoke, VA,U.S.A.) was implanted into the hippocampus at the follow-ing coordinates: AP �4.6, ML �4.9, V �6.3 (verticallyplaced ECB devices) and AP �4.8, ML �5.2, V �6.3 (26°angular placed ECB devices) from bregma, midline, anddura, respectively. A reference electrode was placedbetween the skull and the temporal muscle. Proximal elec-trode sockets were inserted into a plastic pedestal (PlasticsOne) and fixed on the skull with dental cement (Kemdent,Wiltshire, United Kingdom). The animals were allowed torecover for 1 week before rapid kindling electrical stimula-tions were started.

Rapid kindlingOne week after electrode and ECB device implantation,

the individual current threshold for epileptiform afterdis-charge (AD) induction was measured. The stimulation cur-rent started at 10 lA and increased by steps of 10 lA(1 msec square wave pulse, 100 Hz) until a focal EEG ADof at least 5 s duration was elicited. Subsequently, inductionof epileptic activity was initiated according to the rapidkindling protocol, 40 recurrent stimulations given every5 min, consisting of trains of 10 s duration (1 msec bipolarsquare wave pulses at 10 Hz), with a current intensity of400 lA.

The behavioral seizures during stimulation were scoredaccording to the Racine scale:24 Stage 0, no behavioralchanges; stage 1, facial twitches; stage 2, chewing andhead nodding; stage 3, unilateral forelimb clonus; stage4, rearing, body jerks, bilateral forelimb clonus; stage 5,imbalance. Electroencephalography (EEG) was recordedon a MacLab system (AD Instruments, Bella Vista,Australia) 1 min before and 1 min after electrical stimu-lation.

Explantation of ECB devices and perfusionFour weeks after rapid kindling stimulation, the animals

were deeply anesthetized with pentobarbital and were per-fused transcardially with 0.9% NaCl. The skull was openedand the ECB devices were removed and put into the pre-heated (37°C) medium (HE-SFM) and were stored in an

incubator (37°C, 5% CO2). Galanin levels in the incubationsolution were measured after 24 h by a galanin ELISA (Ba-chem) to estimate galanin release from surviving encapsu-lated cells. Moreover, ECB devices were embedded inresin, and cut (5 lm) and stained for hematoxylin & eosinto evaluate cell survival. The brains were removed and fixedin 4% paraformaldehyde for 24 h and then overnight in30% sucrose in 0.1 M sodium phosphate-buffered saline(KPBS). Brains were cut on a microtome in 30-lm–thickslices and stored in a cryoprotective solution in the freezeruntil use.

Immunohistochemistry and other staining proceduresfor brain slices

To determine the extent of the damage and localization ofthe ECB devices, hematoxylin & eosin stainings were per-formed on slices from all brains. To evaluate the inflamma-tion, double immunohistochemical stainings of ionizedcalcium-binding adapter molecule 1 (Iba1) together withectodermal dysplasia 1 (ED1) was performed. Slices wererinsed with 0.02 M KPBS and preincubated with 5% normalgoat serum and 5% normal donkey serum in 0.25% Triton-KPBS for 1 h in room temperature (RT). The slices werethen incubated with the sera, rabbit anti-Iba1 (Wako, Neuss,Germany; 1:1,000) and mouse anti-ED1 (AbD Serotec,Puchheim, Germany; 1:200) overnight at room temperature(RT). Next day the slices were rinsed with 0.02 M KPBS,followed by incubation with secondary antibodies for Iba1(FITC-goat anti-rabbit; Jackson Immunoresearch, Suffolk,United Kingdom; 1:400) and ED1 (Cy3-donkey anti-mouse;Jackson Immunoresearch; 1:400) for 2 h in RT. After 2 h,the slices were once again rinsed the same way as previouslyand were mounted on coated slides and coverslipped with1,4-diazabicyclo[2.2.2]octane (DABCO) (Sigma-Aldrich,Stockholm, Sweden).

Cell counting was performed in the motor cortex (1 mm2

including all layers of the cortex) on three serial (180 lmapart) Iba1/ED1 stained slices from each animal with verti-cally placed ECB devices. Four animals were counted bilat-erally, but no difference in stained cell numbers wasdetected between the sides (data not shown); therefore, therest of the animals were counted unilaterally, only ipsilat-eral to the electrode side. Cell counting was performed andthe images for figures were acquired with Olympus BX61fluorescence microscope.

Statistical analysisStatistical analysis of data was performed using

Student’s unpaired t-test. Differences between groupswere considered statistically significant at p < 0.05.Data are presented as mean � standard error of themean (SEM). The investigator conducting the behavioralgrading of seizures in the animals, EEG, and histologicalanalysis was unaware of the group identity of individualanimals.


169

Galanin-Producing Cells Inhibit Seizures

ResultsImplanted ECB devices with the HR galanin cell cloneshorten duration of focal seizures

After filling ECB devices with galanin-producing cells,the release of galanin in the incubation culture media wasmeasured (Fig. 1). One week after filling, the galaninrelease from the LR clone and the HR clone was similar,whereas 3 weeks after filling the HR galanin clone releasedabout 30% more galanin than the LR clone (HR12.6 � 0.4 ng/mL/24 h, LR 8.3 � 0.3 ng/mL/24 h;p < 0.05). During the initial phase of kindling stimulations,the seizures are usually focal, whereas subsequent stimula-tion-induced seizures spread and manifest as convulsions ofincreasing severity. The AD threshold for seizure inductionwas unaffected with vertically implanted and with angularECB devices filled with galanin-releasing cells (Vertical:control 48.9 � 8.1 lA, LR 52.5 � 10.1 lA, HR50.0 � 6.3 lA; Angular: control 39.2 � 4.7 lA, LR43.0 � 7.6 lA, HR 44.3 � 7.4 lA; p > 0.05) (Figs 2Aand 3A). However, the AD duration of focal seizures wasmoderately decreased in animals with vertically implantedECB devices, which released high levels of galanin com-pared to control (stage 1: control 50.1 � 1.5 s; LR,46.0 � 1.4 s; HR, 43.3 � 1.5 s; stage 2: control83.3 � 6.0 s; LR, 90.9 � 5.7 s; HR, 61.9 � 5.5 s), butwas unaltered at the generalized stages (stages 3–5: control96.9 � 11.4 s; LR, 106.6 � 16.3 s; HR, 90.5 � 14.0 s;p > 0.05) (Fig. 2B). The LR galanin clone showed adecreased AD duration during stage 1, but not for any otherseizure stages. Similarly, the angular-placed ECB devicesdecreased AD duration for focal seizures in animals withonly the HR clone (stage 1: control 49.4 � 1.2 s; LR47.0 � 1.6 s; HR 43.7 � 2.0 s; stage 2: control95.9 � 4.8 s; LR 93.3 � 11.1 s; HR 59.4 � 5.4 s)

(Fig. 3B. Representative EEG traces are shown in Fig. 4).There was no effect of LR ECB devices on AD duration atany seizure stage. The AD duration of generalized seizuresremained unaltered (stages 3–5: control 115.0 � 9.6 s; LR116.1 � 12.8 s; HR 112.4 � 7.2 s; p > 0.05).

After retrieval of the ECB devices, galanin-release levelswere measured once again in the culture media. The galaninrelease was decreased for about 50% or greater (LR3.5 � 0.5 ng/mL/24 h; HR 5.5 � 1.2 ng/mL/24 h;p > 0.05) compared to the levels right before the implanta-tion (Fig. 5A). No galanin release was detected from theempty or parental ARPE-19 cell line–filled ECB devices.Cell survival in the retrieved ECB devices was also verifiedwith hematoxylin & eosin staining of ECB-device sections(Fig. 5B–D). Overall, almost all ECB devices demonstrat-ing galanin release also contained surviving cells. Onlyanimals implanted with ECB devices that still released gala-nin after retrieval and those that had viable cells (all excepttwo animals) were included in the analysis.

Hematoxylin & eosin stainings were performed on sec-tions from brains of all animals to confirm the position ofthe ECB devices in the hippocampus. The expected posi-tions from two experimental animals are exemplified inschematic drawings on Figure 6A and 6C. The examplesof actual sections showing ECB device position in the

A

B

Figure 2.

High-releasing ECB devices with vertical placement attenuate focal

epileptic seizures. (A) AD duration at threshold stimulations was

unchanged in animals implanted with galanin-producing ECB

devices. (B) Average AD duration for kindling stages 1 and 2 was

attenuated by high-releasing galanin-producing ECB devices. Values

are presented as mean � SEM, *p < 0.05 and **p < 0.01 com-

pared to control.

Epilepsia ILAE

Figure 1.

Galanin release from low-releasing and high-releasing ECB devices.

Galanin release was measured using ELISA 1 week and 3 weeks

after filling the ECB devices with ARPE-19 cell lines. Values are pre-

sented as mean � SEM, ***p < 0.001 compared between low-

releasing and high-releasing clones.

Epilepsia ILAE


170

L. Nikitidou et al.

hippocampus are shown on Fig. 6B and 6D. Scoring of theextent of the hippocampal damage exerted by the ECBdevice implantation did not reveal any correlation betweenthe extent of damage and the severity of seizures (data notshown).

Inflammatory response in the motor cortex caused byencapsulated cells

Next we asked whether implantation of the ECB deviceswith galanin-producing cells could cause an inflammatoryresponse, thereby contributing to the observed effects onseizures. However, the ECB devices, regardless of whetherthey contained galanin-producing cells or those not modi-fied, induced inflammatory reaction of the same minormagnitude as judged by estimating the numbers of activatedmicroglia (number of Iba1-positive cells; Fig. 7A). Thenumber of activated microglia was, however, slightly less inthe brain slices from animals implanted by empty ECBdevices (empty 204.4 � 1.2 cells; ARPE 220.7 � 4.2 cells;LR 219.1 � 4.1 cells; HR 218.4 � 2.2 cells) (Fig. 7A).When ED1 immunoreactive cells were counted in the cor-tex, no difference between the different groups could bedetected (empty 12.1 � 0.7 cells; ARPE 13.7 � 1.0 cells;LR 11.3 � 0.7 cells; HR 12.1 � 1.0 cells; p > 0.05)(Fig. 7B). Similarly, double-labeled for Iba1 and ED1immunoreactive cell numbers were not different in variousgroups (empty 6.1 � 0.4 cells; ARPE 6.8 � 0.8 cells; LR5.7 � 0.5 cells; HR 5.7 � 0.4 cells; p > 0.05) (Fig. 7C).The minor inflammatory responses, number of focal andgeneralized seizures, as well as progression of kindling inboth experiments was not different from what has beenreported previously in animals without capsule implanta-tion,25,26 suggesting that implantation of empty ECBdevices does not lead to any significant alteration in inflam-matory responses or kindling outcomes. Taken together, ourdata suggest an only slightly increased number of activatedmicroglia in the cortex from implanted cell–containingECB devices.

A

B

Figure 3.

High-releasing ECB devices with angled placement attenuate focal

epileptic seizures. (A) AD duration at threshold stimulations was

unchanged in animals implanted with galanin-producing ECB

devices with different content. (B) Average AD for kindling stages

1 and 2 was attenuated by high-releasing galanin-producing ECB

devices. Values are presented as mean � SEM, *p < 0.05 and

***p < 0.001 compared to control.

Epilepsia ILAE

A

B

C

D

Figure 4.

EEG traces recorded during focal and generalized seizures. Repre-

sentative EEG recordings during focal (stage 2) and generalized sei-

zures (stage 5) in animals implanted with non–galanin-releasingECB devices (A and C) or ECB devices containing the high-releas-

ing cell line clone (B andD).

Epilepsia ILAE

A B

C

D

Figure 5.

Galanin release after explantation of ECB devices. (A) Average gal-

anin levels measured by ELISA in ECB devices after explantation.

Hematoxylin & eosin staining of a section from an ECB device filled

with (B) ARPE-19 control cells, (C) low-releasing galanin cells, and

(D) high-releasing galanin cells. Values are presented as

mean � SEM, p > 0.05.

Epilepsia ILAE


171


DiscussionHerein we demonstrate that implanted ECB devices that

release galanin decrease AD duration of focal seizures in a

well-established model of epilepsy: rapid kindling. Thesedata suggest that ECB devices could potentially be an alter-native source for exogenous long-term delivery of neuro-peptides or other biologic compounds to the brain, inparticular to the hippocampus, to suppress focal epilepticseizures. The observed seizure suppression of galanin byECB devices was relatively moderate, shortening the focalseizures but having no significant effect on duration of gen-eralized seizures.

Overall, the observed effect of grafted galanin-releasingECB devices is in line with previous publications, wherebygalanin has been shown to exert an inhibitory effect on sei-zures.11,22,23 The novel finding is that the ECB technology,which is a relatively safe treatment strategy compared todirect gene or cell therapy approaches, is a valid alternative,and may be considered for translational developmenttoward clinical applications, for example, in patients withtemporal lobe epilepsy. The advantage of the ECB technol-ogy from the point of patient safety is several-fold. First, thegrafted cells are isolated from the host cells by a semiperme-able membrane, and therefore, it is possible to remove thewhole graft in the event of adverse effects, or replace it ifnecessary; second, immune cells from the host cannotaccess the grafted cells, and therefore the risk of graft rejec-tion is minimized; third, there is no genetic manipulation ofhost cells, thereby diminishing the risk of unwanted muta-tions and carcinogenesis; fourth, there is no direct contact orinteraction between the grafted and host cells, thus reducingthe risk for transgene down-regulation or some other directregulatory effects from the host. Despite these advantages,the ECB devices may not be a first choice when graft–host

A B

C D

Figure 6.

Predicted placement of ECB devices in the hippocampus. (A) Pre-

dicted placement of vertically implanted ECB devices. (B) An

example of actual vertical placement of the ECB device in the hip-

pocampus. (C) Predicted placement of the ECB devices implanted

with a 26-degree angle. (D) An example of actual placement of the

ECB device implanted with an angle.

Epilepsia ILAE

A B C

Figure 7.

Inflammatory response in the cortex after implantation of ECB devices. (A) Average number of Iba1 immunoreactive cells in the cortex

of animals implanted with different types of ECB devices. (B) Average number of ED1 immunoreactive cells, a marker for activated micro-

glia in the cortex of same animals as in (A). (C) Double-labeled Iba1 and ED1 immunoreactive cells in the same animals as in (A) and (B).

Values are presented as mean � SEM, **p < 0.01 and ***p < 0.001 compared to empty.

Epilepsia ILAE


172

L. Nikitidou et al.

direct interaction and bidirectional integration is desirable,since ECB does not provide this possibility. This approachexcludes that grafted cells would integrate and become partof the existing network, and thereby induce release of theproduct of interest in a more regulated manner, just when itis needed. Such possible regulatory mechanism would allowfor perhaps more physiologic interaction of the graft and thehost, preventing, for example, down-regulation of the recep-tors in the host cells due to permanent high levels of theligands, as is the case for the ECB. Yet another limitation ofthe ECB technology may be a requirement of relatively highlevels of the gene product released by the encapsulated cellsto reach a therapeutically effective dose in the host brain tis-sue. The moderate effect of galanin-releasing ECB deviceson seizures, observed in the present study, may be related toinsufficient galanin levels provided by the implanteddevices. The moderate effect of galanin-releasing ECBdevices could be related to several factors. One possibility islow levels of galanin released by implanted ECB devices inthis study. In support of this notion, ELISA measurementsof galanin release from explanted ECB devices (5 weeksafter initial implantation) were less than half of what wasmeasured prior to implantation. Such decrease in galaninrelease could be caused by compromised survival of theencapsulated cells, which was apparent at least in some ofthe explanted ECB devices. Another reason for the moder-ate effect could be a glial scar formed around the devices.This would restrict diffusion of galanin into the host brain.Yet another factor potentially affecting the outcome of thepresent experiments could be related to down-regulation ofgalanin receptors induced by permanent increase of galaninlevels around the ECB devices. These questions need to beaddressed in future studies in more detail. Supporting animportant role of ligand levels released from ECB devices,previous studies with glial cell line–derived neurotrophicfactor (GDNF)-ECB25 have demonstrated suppression ofseizures with lower release levels of GDNF but not withhigher levels of GDNF. Galanin-based gene therapyapproaches22,27 have shown better outcome in seizure sup-pression in various models of epilepsy.

One possible confounding factor for seizure-suppressanteffects observed in this study could be the inflammatoryresponse to the galanin-producing ECB devices usingARPE cell line–containing devices. Inflammation has beenshown to play a profound role in epileptogenesis and icto-genesis.28,29 Therefore, possible inflammatory reaction ofthe host, despite the fact that these cells are behind the semi-permeable membrane (see above), could modulate seizuresand contribute to the observed effects. Indeed, our data sug-gest that ECB devices that contain the human ARPE celllines, galanin-producing or not, induce mild inflammatoryreaction, and elevate Iba1-positive cell counts in the corticalregion of the grafted animals as compared to thoseimplanted with empty devices. It should be noted that theECB devices increased the number of Iba1-immunoreactive

cells, but did not alter the number of ED1-immunoreactivecells. This would indicate a relatively mild level of inflam-matory reaction caused by the ECB devices. However, thenumber of Iba1-positive microglia was similar in all groupswith ECB devices containing ARPE cell lines, regardless ofwhether they released low or high levels of galanin or nogalanin at all. This would suggest that galanin-releasingECB devices per se did not induce any additional inflam-matory reaction (as judged by Iba1 immunostaining).Moreover, these data support the idea that galanin releasedfrom ECB devices was responsible for the observed seizure-suppressant effect but not the inflammatory reaction of thehost.

To strengthen clinical potential of the ECB technology,some optimizations need to take place. Among others, theexpression levels of galanin (or any other seizure-suppres-sant biologic compounds) could be increased further byusing transposon-based gene constructs as demonstratedwith the expression of nerve growth factor (NGF).30 More-over, recent unpublished experiments show that cell viabil-ity and density can be increased by using different scaffoldsthat allow better manufacturability and cell adhesion thanthat of the PVA foam used in the present study. Lastly, theECB allows for the expression and utilization of combina-tions of seizure-suppressant peptides and proteins, poten-tially strengthening the effect by their additive and orsynergistic action.

ConclusionThe ECB technology has been tested previously in

other neurologic diseases, such as Alzheimer’s disease,and has demonstrated good safety, tolerability, and indi-cations for positive functional outcomes.31,32 Our datasuggest that ECB devices could be a feasible strategy fordelivering galanin or other seizure-suppressant agentslocally into the focus of epileptic seizures. The variousadvantages of the ECB devices compared to other geneor gene product delivery techniques should be evaluatedagainst the efficacy and functional outcomes. In ourstudy, the observed seizure-suppressant effects by gala-nin-releasing ECB devices were moderate, and thereforemay need further optimization before it can be consid-ered for clinical application. Higher levels of galaninrelease over longer time periods seem to be necessary toachieve better outcomes in seizure control.

AcknowledgmentsThe authors are grateful to laboratory technician Nora Pernaa at Lund

University for the help with hematoxylin & eosin stainings and to the tech-nical staff at NsGene, Janni Larsen, Philip Usher, and Juliano Olsen fortheir help with the experiments. The study was supported by EU commis-sion FP7 grant EPIXCHANGE, and FP5 grant EPICURE, SwedishResearch Council, Kock Foundation, Hj€arnfonden and Hardebo Founda-tion.


173


DisclosureNone of the authors have any conflict of interest to disclose. We confirm

that we have read the Journal’s position on issues involved in ethical publi-cation and affirm that this report is consistent with those guidelines.

References1. Mandel RJ, Burger C. Clinical trials in neurological disorders using

AAV vectors: promises and challenges. Curr Opin Mol Ther2004;6:482–490.

2. LeWitt PA, Rezai AR, Leehey MA, et al. AAV2-GAD gene therapyfor advanced Parkinson’s disease: a double-blind, sham-surgerycontrolled, randomised trial. Lancet Neurol 2011;10:309–319.

3. Kaplitt MG, Feigin A, Tang C, et al. Safety and tolerability of genetherapy with an adeno-associated virus (AAV) borne GAD gene forParkinson’s disease: an open label, phase I trial. Lancet2007;369:2097–2105.

4. Tatemoto K, Rokaeus A, Jornvall H, et al. Galanin – a novelbiologically active peptide from porcine intestine. FEBS Lett1983;164:124–128.

5. Kuteeva E, Wardi T, Lundstrom L, et al. Differential role of galaninreceptors in the regulation of depression-like behavior andmonoamine/stress-related genes at the cell body level.Neuropsychopharmacology 2008;33:2573–2585.

6. Lu X, Barr AM, Kinney JW, et al. A role for galanin in antidepressantactions with a focus on the dorsal raphe nucleus. Proc Natl Acad SciUSA 2005;102:874–879.

7. Steiner RA, Hohmann JG, Holmes A, et al. Galanin transgenic micedisplay cognitive and neurochemical deficits characteristic ofAlzheimer’s disease. Proc Natl Acad Sci USA 2001;98:4184–4189.

8. Jhamandas JH, Harris KH, MacTavish D, et al. Novel excitatoryactions of galanin on rat cholinergic basal forebrain neurons:implications for its role in Alzheimer’s disease. J Neurophysiol2002;87:696–704.

9. Ding X, MacTavish D, Kar S, et al. Galanin attenuates beta-amyloid(Abeta) toxicity in rat cholinergic basal forebrain neurons. NeurobiolDis 2006;21:413–420.

10. Counts SE, He B, Che S, et al. Galanin fiber hyperinnervationpreserves neuroprotective gene expression in cholinergic basalforebrain neurons in Alzheimer’s disease. J Alzheimers Dis2009;18:885–896.

11. Kokaia M, Holmberg K, Nanobashvili A, et al. Suppressed kindlingepileptogenesis in mice with ectopic overexpression of galanin. ProcNatl Acad Sci USA 2001;98:14006–14011.

12. Mazarati AM, Hohmann JG, Bacon A, et al. Modulation ofhippocampal excitability and seizures by galanin. J Neurosci2000;20:6276–6281.

13. Mazarati AM, Liu H, Soomets U, et al. Galanin modulation of seizuresand seizure modulation of hippocampal galanin in animal models ofstatus epilepticus. J Neurosci 1998;18:10070–10077.

14. Schlifke I, Kuteeva E, Hokfelt T, et al. Galanin expressed in theexcitatory fibers attenuates synaptic strength and generalized

seizures in the piriform cortex of mice. Exp Neurol2006;200:398–406.

15. Mitsukawa K, Lu X, Bartfai T. Galanin, galanin receptors and drugtargets.Cell Mol Life Sci 2008;65:1796–1805.

16. Branchek TA, Smith KE, Gerald C, et al. Galanin receptor subtypes.Trends Pharmacol Sci 2000;21:109–117.

17. Iismaa TP, Shine J. Galanin and galanin receptors. Results Probl CellDiffer 1999;26:257–291.

18. Lu X, Mazarati A, Sanna P, et al. Distribution and differentialregulation of galanin receptor subtypes in rat brain: effects of seizureactivity.Neuropeptides 2005;39:147–152.

19. Palazzi E, Felinska S, Zambelli M, et al. Galanin reduces carbacholstimulation of phosphoinositide turnover in rat ventral hippocampus bylowering Ca2+ influx through voltage-sensitive Ca2+ channels.J Neurochem 1991;56:739–747.

20. Zini S, Roisin MP, Langel U, et al. Galanin reduces release ofendogenous excitatory amino acids in the rat hippocampus. EurJ Pharmacol 1993;245:1–7.

21. Kask K, Berthold M, Bartfai T. Galanin receptors: involvement infeeding, pain, depression and Alzheimer’s disease. Life Sci1997;60:1523–1533.

22. Kanter-Schlifke I, Toft Sorensen A, Ledri M, et al. Galanin genetransfer curtails generalized seizures in kindled rats without alteringhippocampal synaptic plasticity. Neuroscience 2007;150:984–992.

23. Lin EJ, Richichi C, Young D, et al. Recombinant AAV-mediatedexpression of galanin in rat hippocampus suppresses seizuredevelopment. Eur J Neurosci 2003;18:2087–2092.

24. Racine RJ. Modification of seizure activity by electrical stimulation.II. Motor seizure. Electroencephalogr Clin Neurophysiol 1972;32:281–294.

25. Kanter-Schlifke I, Fjord-Larsen L, Kusk P, et al. GDNF released fromencapsulated cells suppresses seizure activity in the epileptichippocampus. Exp Neurol 2009;216:413–419.

26. Sorensen AT, Nikitidou L, Ledri M, et al. Hippocampal NPY genetransfer attenuates seizures without affecting epilepsy-inducedimpairment of LTP. Exp Neurol 2009;215:328–333.

27. McCown TJ. Adeno-associated virus-mediated expression andconstitutive secretion of galanin suppresses limbic seizure activity invivo.Mol Ther 2006;14:63–68.

28. Auvin S, Mazarati A, Shin D, et al. Inflammation enhancesepileptogenesis in the developing rat brain. Neurobiol Dis2010;40:303–310.

29. Galic MA, Riazi K, Heida JG, et al. Postnatal inflammation increasesseizure susceptibility in adult rats. J Neurosci 2008;28:6904–6913.

30. Fjord-Larsen L, Kusk P, Emerich DF, et al. Increased encapsulatedcell biodelivery of nerve growth factor in the brain by transposon-mediated gene transfer.Gene Ther 2012;19:1010–1017.

31. Tornoe J, Torp M, Jorgensen JR, et al. Encapsulated cell-basedbiodelivery of Meteorin is neuroprotective in the quinolinic acid ratmodel of neurodegenerative disease. Restor Neurol Neurosci2012;30:225–236.

32. Eriksdotter-Jonhagen M, Linderoth B, Lind G, et al. Encapsulated cellbiodelivery of nerve growth factor to the Basal forebrain in patientswith Alzheimer’s disease. Dement Geriatr Cogn Disord 2012;33:18–28.


174

L. Nikitidou et al.

encapsulated galanin-producing cells attenuate focal epileptic seizures in the hippocampus

Documents