Adeno-associated Virus-Mediated Expression and Constitutive Secretion of Galanin Suppresses Limbic Seizure Activity in Vivo

Transcription

1 doi: /j.ymthe ARTICLE Adeno-associated Virus-Mediated Expression and Constitutive Secretion of Galanin Suppresses Limbic Seizure Activity in Vivo Thomas J. McCown* Gene Therapy Center and Department of Psychiatry, University of North Carolina at Chapel Hill School of Medicine, Chapel Hill, NC 27599, USA *To whom correspondence and reprint requests should be addressed at the UNC Gene Therapy Center, 7119 Thurston, CB 7352, Chapel Hill, NC 27599, USA. Available online 30 May 2006 Intractable temporal lobe epilepsy presents an ideal target for gene therapy, but therapeutic success depends upon the ability to suppress limbic seizure activity. Adeno-associated virus vectors (AAV) were constructed in which the fibronectin secretory signal sequence (FIB) preceded the coding sequence for galanin (AAV-FIB-GAL) or green fluorescent protein (AAV-FIB-GFP), constructs that express and constitutively secrete the gene product. Bilateral AAV-FIB-GAL infusion into the rat piriform cortex (2 Ml/side) significantly attenuated kainic acid-induced seizures (10 mg/kg, ip) such that 11/12 rats exhibited no limbic seizures, while the remaining rat exhibited only a brief, single class III seizure. This AAV-FIB-GAL infusion also prevented electrographic seizure activity. In contrast, bilateral AAV-FIB-GFP infusion did not alter either behavioral or electrographic seizure activity. Since prior seizure exposure could influence vector efficacy, another group of rats received daily electrical stimulation of the piriform cortex until three consecutive class V seizures were elicited. Subsequently, AAV-FIB-GAL or AAV-FIB-GFP (3 Ml/30 min) was infused into the area of the electrode. One week later the AAV-FIB-GAL rats exhibited a significant increase in the stimulation current necessary to evoke limbic seizure activity, while AAV-FIB-GFP did not alter the seizure threshold. Thus, AAV-mediated galanin expression and secretion significantly suppress limbic seizure activity in vivo. Key Words: adeno-associated virus, epilepsy, seizures, galanin, piriform cortex INTRODUCTION Focal epilepsies present an attractive target for gene therapy, and recently, studies focusing upon neuroactive peptides have yielded some promising results. The endogenous neuropeptide, galanin, generally exerts an inhibitory action within the central nervous system (CNS) by increasing potassium conductance or reducing calcium conductances [1], and the magnitude of this inhibition has proven sufficient to suppress hippocampal seizure activity upon local infusion [2]. Based upon these findings, Haberman et al. [3] first demonstrated the antiseizure efficacy of vector-mediated neuropeptide gene expression, using a novel platform designed to circumvent the potential liabilities of viral vector tropism [4].An adeno-associated virus (AAV) serotype 2 vector was constructed in which the coding sequence for the active galanin peptide was preceded by the secretory signal sequence of the laminar protein, fibronectin. Thus, the active peptide would be secreted from the transduced cell in an unregulated fashion, essentially bypassing neuropeptide regulatory pathways. These studies showed that the fibronectin secretion signal sequence did cause constitutive secretion of the gene product from transfected cells in vitro and, upon transduction in vivo, significantly attenuated focal seizure sensitivity and prevented seizure-induced hippocampal cell damage. In a subsequent publication, Lin et al. [5] found that AAVmediated expression of a human galanin cdna significantly decreased the number of kainic acid-induced seizure episodes, but this approach to vector-mediated galanin expression did not alter seizure latency or kainic acid-induced neuronal damage. Neuropeptide Y also exhibits anti-seizure activity in vivo [6,7], and recently Richichi et al. [8] showed that AAV-mediated expression of preproneuropeptide Y exerts a range of anti-seizure effects. Using a human preproneuropeptide Y cdna and a hybrid AAV serotype to increase the area of gene expression, hippocampal /$30.00

2 ARTICLE doi: / j.ymthe expression of the preproneuropeptide Y increased the electrical seizure threshold in the hippocampus and significantly retarded the rate of kindling epileptogenesis. Thus, these studies clearly established the antiepileptic potential for vector-derived preproneuropeptide Y expression in vivo. Although these studies illustrate the anti-seizure potential of these gene therapy approaches, critical questions remain unanswered [9]. The galanin studies did not establish the ability of vector-derived galanin to attenuate limbic seizure activity. Because the most likely application of antiepileptic gene therapy will involve intractable temporal lobe epilepsy, efficacy must be demonstrated in animal models of limbic seizure activity. Examples include kainic acid-induced limbic seizures [10] and electrical kindling of limbic seizures [11]. As importantly, though, in all of these studies, viral vector transduction and gene expression transpired in tissue that had not experienced prior seizure exposure. Again, the most likely clinical application of such gene therapies will involve CNS structures that have experienced chronic seizure activity, and under such conditions galanin expression and constitutive secretion might be significantly altered. Therefore, the following studies first address the ability of AAV-mediated galanin expression and secretion to attenuate kainic acidinduced limbic seizure activity and second probe the ability of this gene therapy approach to attenuate seizure activity after multiple exposures to class V limbic seizures. RESULTS AAV-Derived Galanin Prevents Kainic Acid Seizures One week after we infused AAV-FIB-GAL (FIB, fibronectin secretory signal sequence; GAL, galanin; 2 Al, bilaterally; viral particles/ml; N = 12) or AAV- FIB-GFP (GFP, green fluorescent protein; 2 Al, bilaterally; viral particles/ml; N = 8) into the piriform cortex, all of the rats received a 10 mg/kg ip dose of kainic acid. The appearance of wet dog shakes reflected the first behavioral evidence of kainic acid excitation, and no differences emerged in the latency to wet dog shakes between the two AAV vector-infused groups and the uninjected control group (N = 8)(Fig. 1). Also, the latency to subsequent class III and class IV limbic seizures did not differ between the untreated control group and the AAV-FIB-GFP-infused group (Fig. 1). Thus, expression and constitutive secretion of the nontherapeutic protein, GFP, did not alter seizure sensitivity. In marked contrast, those animals that received AAV-FIB- GAL infusions exhibited a highly significant increase in the time to class III seizure activity in that only 1 of 12 animals exhibited a single, brief episode of class III seizure behavior over the 240-min observation period. The 11 other animals did not exhibit any limbic seizure FIG. 1. The effects of AAV-FIB-GFP and AAV-FIB-GAL vectors on the expression of limbic seizure behaviors. Seven days after infusion of AAV-FIB-GFP (N = 8) into the piriform cortex, kainic acid administration (10 mg/kg, ip) elicited wet dog shakes, class III and class IV seizure activity, with the same latency as in untreated controls (N =8)(t test, P N 0.1). Seven days after infusion of AAV- FIB-GAL (N = 12) into the piriform cortex, kainic acid administration elicited wet dog shake behaviors with the same latency as in untreated control animals. In contrast, a single, brief class III seizure was exhibited by only 1 of the 12 rats over the 240-min post-kainic acid observation period, while the other 11 rats did not exhibit any limbic seizure behaviors. None of the animals exhibited any class IV seizure behaviors over the entire 240-min post-kainic acid observation period (*t test, P b 0.01) (mean F SEM). behavior over the 240-min observation period. Furthermore, none of the AAV-FIB-GAL-infused rats exhibited any class IV seizure behavior over the 240-min observation period. As previously discussed [3], in vivo expression and secretion of galanin do not prove amenable to immunohistochemical localization, but as before [3], transduction with this AAV vector resulted in the in vivo expression of the appropriate vector-derived mrna (Fig. 2). The neuroanatomical localization of the AAV vector infusions and the level of pre-kainic acid GFP gene expression are illustrated in Figs. 3A and 3B, respectively. It must be noted that the area of GFP gene expression proved relatively localized, encompassing an area that extended approximately 250 Am in the rostral caudal direction within the piriform cortex. However, in those animals that received AAV-FIB-GFP infusions, only 1 of the 8 animals exhibited any GFP expression 2 weeks after kainic acid administration (Fig. 3C). Certainly, the well-documented kainic acid-induced cell death could account for a substantial loss of GFP expression. EEG recordings further substantiated the suppression of limbic seizures by the AAV-FIB-GAL vector (Fig. 4). In control rats (N = 2), kainic acid treatment caused the expected, marked electrographic seizure activity in the piriform cortex (Fig. 4A). The infusion of AAV-FIB-GFP (N = 2) did not alter the appearance of kainic acidinduced electrographic seizure activity that accompa- 64

3 doi: /j.ymthe ARTICLE FIG. 2. The in vivo presence of FIB-GAL mrna 1 week after AAV-FIB-GAL infusion into the piriform cortex. AAV-FIB-GAL vectors were infused into the rat piriform cortex and 1 week later the mrna was extracted. Subsequent RT- PCR shows that the appropriate 155-bp FIB-GAL product is present in the injected piriform cortex (lane A), while no product was found in the contralateral, uninjected cortex (lane B). Omission of the RT step (lane C) indicated the absence of contaminating viral DNA. The two outside lanes contain a 50 bp DNA step ladder with the base pair size indicated on the right. FIB-GAL-infused rats exhibited a significant elevation in the stimulation current necessary to evoke limbic seizure activity (Fig. 5). However, once limbic seizure activity was initiated, AAV-FIB-GAL infusion did not change the severity or duration of the limbic seizure (data not shown). Also, unlike findings 2 weeks after peripheral kainic acid administration, clear GFP expression was found around the stimulating electrode (Fig. 5, right). nied generalized limbic seizure behaviors (Fig. 4B). In AAV-FIB-GAL-treated rats (N = 2), however, no electrographic seizure activity could be found in the piriform cortex after kainic acid administration (Fig. 4C). This absence of local afterdischarge activity paralleled the total lack of any limbic seizure behaviors. Interestingly, this local suppression of electrographic seizure activity did not translate to a global suppression of electrographic seizure activity. In two rats in which AAV-FIB- GAL infusion totally prevented kainic acid-induced behavioral seizure activity, EEG recordings from the hippocampus revealed sporadic electrographic seizure activity (Fig. 4D). AAV-Derived Galanin Attenuates Seizure Sensitivity after Prior Limbic Kindling The findings with kainic acid suggested that severe seizure activity might lead to a decline in vector-derived gene product in vivo. To test if prior seizure activity altered the actions of AAV-FIB-GAL vectors in vivo, rats received daily electrical kindling stimulation of the piriform cortex until three consecutive class 5 limbic seizures occurred. Following the subsequent test for the seizure initiation threshold, the rats received infusions of either AAV-FIB-GFP (N = 5) or AAV-FIB-GAL (N =9) in the area of the stimulating electrode. One week later, the AAV-FIB-GFP-infused rats exhibited no significant difference in the threshold current necessary to elicit limbic seizure activity (Fig. 5, left). In contrast, AAV- FIG. 3. Illustrations of AAV vector infusion sites and expression of GFP in the piriform cortex. (A) The range of AAV-FIB-GFP and AAV-FIB-GAL infusion sites in the piriform cortex is shown (diagrams adapted from Paxinos and Watson [19]). (B) Typical GFP expression in the piriform cortex prior to treatment with kainic acid. This level of GFP expression coursed approximately 250 Am in a rostral caudal direction. (C) The level of GFP expression 2 weeks after kainic acid administration in the only animal of eight total in which GFP expression was evident. The other seven animals exhibited no discernable GFP expression. Horizontal bars, 40 Am. 65

4 ARTICLE doi: / j.ymthe DISCUSSION These studies not only show that the expression and constitutive secretion of galanin can suppress limbic seizure activity in vivo, but they also implicate the piriform cortex as a critical gate modulating the behavioral expression of limbic seizure activity. Haberman et al. [3] previously demonstrated that inclusion of the FIB sequence in AAV vectors resulted in the expression and constitutive secretion of GFP or galanin. Although AAV- FIB-GAL vectors protected hippocampal hilar neurons from kainic acid-induced cell death, the unilateral infusions did not alter the course of electrographic or behavior seizures following peripheral kainic acid administration. A wide range of limbic structures support electrically kindled seizures, including the hippocampus, amygdala, entorhinal cortex, perirhinal cortex, and piriform cortex [12], and the piriform cortex plays a prominent role in the development of kindled seizures [13]. The present findings show that bilateral infusion of the AAV-FIB-GAL vectors into the piriform cortex dramatically suppresses both behavioral and local electrographic seizure activity following peripheral kainic acid treatment. Eleven of the 12 AAV-FIB-GAL-treated animals FIG. 5. The effects of AAV-FIB-GAL (N = 9) or AAV-FIB-GFP (N = 5) infusion on the threshold for seizure genesis in animals that had been electrically kindled to class V seizure activity. As seen on the left, AAV-FIB-GFP infusion did not alter the stimulation threshold for seizure genesis, but AAV-FIB-GAL infusion significantly elevated the amount of current necessary to elicit seizure activity (paired t test, P b 0.02) (mean F SEM). The right shows GFP expression around the electrode tip, demarcated by *. Horizontal bar, 40 Am. FIG. 4. EEG recordings from the piriform cortex or the hippocampus following kainic acid administration. (A) Afterdischarge activity in the piriform cortex of an untreated control animal during a class IV seizure. (B) Afterdischarge activity in the piriform cortex of an AAV-FIG-GFP-treated animal during a class IV seizure, further reinforcing the lack of effect exerted by GFP expression and constitutive secretion. (C) The total lack of any afterdischarge activity in the piriform cortex of an animal that received an AAV-FIB-GAL infusion 7 days prior to the kainic acid treatment (10 mg/kg, ip.. At no time during the 240- min observation period was any behavioral or electrographic seizure activity noted. (D) EEG recording from the hippocampus of an animal that received an AAV-FIB-GAL infusion 7 days prior to the kainic acid treatment. During this recording the animal was motionless, yet the recording shows clear afterdischarge activity. Periods of this electrographic seizure activity occurred over the entire 240-min observation period, but at no time were any limbic seizure behaviors noted. exhibited no limbic seizure behaviors, while the remaining animal exhibited only a single, very brief episode of forelimb clonus. Moreover, the AAV-FIB-GAL treatment prevented local electrographic seizure activity. Although the extent of AAV transduction was substantially less compared to studies with neuropeptide Y [8], clearly, enough vector-derived galanin was expressed and secreted to prevent kainic acid-induced seizure activity. In agreement with previous results [3], AAV-FIB-GFP treatment did not alter either the behavioral or the electrographic seizure activity, so merely expressing and constitutively secreting a nontherapeutic protein does not influence kainic acid-induced seizure activity. Although peripheral kainic acid administration impacts the entire brain, AAV-FIB-GAL blockade of generalized seizure behaviors apparently arises from local actions in the piriform cortex. Wet dog shakes represent the first behavioral evidence of kainic acid excitation, and these behaviors appeared with the same latency across all of the treatment groups. Because the origin of wet dog shakes appears to be the hippocampus [14], a local action of the AAV-FIB-GAL in the piriform cortex would not be expected to influence these kainic acidinduced behaviors. Furthermore, expression and secretion of galanin in the piriform cortex did not prevent electrographic seizure activity in the hippocampus, even though limbic seizure behaviors were totally absent. It has been established previously that galanin can reduce neuronal excitability through interactions with galanin receptors [15,16], and clearly the AAV-FIB-GAL vectors 66

5 doi: /j.ymthe ARTICLE support sufficient vector-derived galanin secretion to suppress local neuronal excitability significantly. The fact that this local attenuation of neuronal excitability prevents generalized limbic seizure behaviors implies that the piriform cortex serves as an important gate for generalized limbic seizures. Finally, a key element to the success of any antiepileptic gene therapy will be the ability to suppress seizure activity in brain structures that have experienced chronic seizure activity. Following electrical kindling of the piriform cortex, AAV-FIB-GAL infusion supported enough galanin expression and secretion to elevate the seizure stimulation threshold significantly, but again, expressing and secreting a nontherapeutic protein, GFP, did not influence seizure sensitivity. As in the kainic acid studies, these effects of vector-derived galanin appear localized to the site of transduction, because once the seizure was initiated, no changes were found in the subsequent progression to generalized seizures, an event involving distal neuronal networks. Although electrical kindling does not cause the acute neuropathology characteristic of kainic acid administration, fully kindled animals do exhibit a significant, permanent enhancement of seizure sensitivity [17]. Whether such a gene therapy approach will prove effective likely will depend upon the nature and extent of seizure-induced pathology. Certainly frank neuronal cell loss will influence the effectiveness of any antiepileptic gene therapy. Thus, careful selection of both the AAV serotype and the specific promoter will prove crucial given the inevitable neuronal pathology that results from intractable temporal lobe epilepsy. MATERIALS AND METHODS Experimental animals. All of the animals were pathogen-free male Sprague Dawley rats obtained from Charles Rivers. The animals were maintained in a 12-h light dark cycle and had free access to food and water. All care and procedures were in accordance with the Guide for the Care and Use of Laboratory Animals (DHHS Publication No. [NIH]85-23), and all procedures received prior approval by the University of North Carolina Institutional Animal Care and Usage Committee. AAV plasmid construction and recombinant virus production. The AAV plasmids were constructed as previously described [3]. Briefly, the fibronectin signal sequence (nucleotides ) was derived from the rat fibronectin mrna sequence (GenBank Accession No. X15906) and oligonucleotides corresponding to both strands were generated (Midland Certified Reagent Co.). AgeI overhangs were included such that the annealed oligonucleotides could be ligated in front of GFP in an AAV plasmid in which gene expression is driven by the hybrid chicken h-actin promoter. This resulted in plasmid pfib-gfp. The galanin sequence was amplified by RT-PCR from rat brain RNA using primers directed to the mature peptide sequence such that melting and reannealing of two separate PCR products resulted in 5V AgeI and 3V NotI overhangs [18]. The 5V primer also included an additional ATG to ensure the translational start ofgalaninintheabsence ofthefibsequenceandthe3v primer contained a stop codon to terminate translation properly. The sequence of the primers was as follows: (1) 5V-CCGGTAATGGGCTGGACCCTGAAC-3V, (2) 5V- TAATGGGCTGGACCCTGAAC-3V, (3)5V-GCTCATGTGAGGCCATGCTTG- 3V, (4)5V-GGCCGCTCATGTGAGGCCATGCTTG-3V. The PCR product was ligated into the AgeI NotIsites replacing thegfp gene, resulting in plasmid pgal. pfib-gal was generated by adding the FIB oligonucleotide in the same manner as in the GFP constructs. All clones were sequenced to ensure accuracy. Virus production and purification. Recombinant AAV serotype 2 was produced as previously described [3] within the UNC Vector Core. Briefly, HEK 293 cells were transfected with the transgene plasmid, pacg2, and XX6-80 via the calcium phosphate method. Cells were lysed by freeze thaw or deoxycholate, and cell lysates were centrifuged through a discontinuous iodixanol gradient. The virus was further purified over a Hi-Trap heparin column (Amersham Pharmacia Biotech), dialyzed into phosphate-buffered saline, and stored at 808C until use. Recombinant virus titers were calculated by dot blot [19]. AAV vector infusion and piriform cortex electrode/cannula implantation. For virus vector infusions and electrode implants, animals first were anesthetized with 50 mg/kg pentobarbital and placed into a stereotaxic frame. Using a 32-gauge stainless steel injector and a Sage infusion pump, animals received 2 Al of AAV-FIB-GAL (N = 12) or AAV- FIB-GFP (N = 8) bilaterally over 20 min into the piriform cortex (IAL 6.7 mm, lateral 6.0 mm, vertical 8.4 mm, according to the atlas of Paxinos and Watson [20]). The injector was left in place for 3 min postinfusion to allow diffusion from the injectors. Immediately after the vector infusions, a subset of rats received electrode implants to monitor local EEG activity after peripheral kainic acid administration. Two AAV-FIB-GAL- and 2 AAV-FIB-GFP-injected animals were implanted with a bipolar platinum iridium electrode (0.007 in. diameter, insulated except for the tip cross section, 300 Am vertical tip separation) using the same coordinates as the vector infusion into the piriform cortex. Additionally, 2 AAV-FIB-GALinfused rats were implanted with a bipolar platinum iridium electrode (0.007 in. diameter, insulated except for the tip cross section, 300 Am vertical tip separation and a separate ground) into the hilar region of the hippocampus (IAL 4.7 mm, lateral 2.5 mm, vertical 3.6 mm, according to the atlas of Paxinos and Watson [20]). Finally, 2 uninjected control animals received a platinum iridium electrode implant into the piriform cortex. For the electrode implants, a ground wire was attached to one of three screws placed in the skull and the electrode was secured to these screws with cranioplastic cement. In all cases, the incisions were sutured, and the animals were allowed 7 days to recover. For the piriform kindling study, animals (N = 14) received electrode/ cannula implants that allowed electrical stimulation, EEG recording, and virus microinjection in the area of the electrode. As before, animals first were anesthetized with 50 mg/kg pentobarbital and placed into a stereotaxic frame. Next a bipolar platinum iridium electrode/cannula (0.007 in. diameter, insulated except for the tip cross section, 300 Am vertical tip separation; 26-gauge stainless steel hypodermic tubing) assembly was implanted into the piriform cortex using the same coordinates as above. The cannula was fixed to the electrode wires by epoxy, where the ventral electrode tip terminated 4 mm dorsal to the cannula, such that an infusion could be made in the area of the stimulating electrode. This assembly was secured by cranioplastic cement to three screws that were placed in the skull. The incision was sutured and the rats were allowed 7 days to recover from the surgery. In vivo detection of AAV-FIB-GAL-derived mrna by RT-PCR. As previously described [3] the vector-derived galanin could not be visualized in vivo by immunohistochemistry, likely due to the fact that the secreted galanin would be rapidly degraded, especially during the perfusion procedure. Thus, as previously described [3], in vivo activity of the AAV- FIB-GAL vector was validated by demonstrating the presence of vectorderived mrna. Four rats received AAV-FIB-GAL vector infusions into the piriform cortex as described above. Then, 1 week later, the animals received an overdose of pentobarbital (100 mg/kg, ip) and were subsequently decapitated. The brain was removed, and the piriform cortex was dissected out. The tissue was stored in RNAlater (Ambion, Austin, TX, USA) at 208C. Subsequently, the RNA was extracted from the tissue (Promega SV-40 total RNA isolation kit; Madison, WI, USA) and reverse transcribed using AMV reverse transcriptase and oligo(dt) primers. The subsequent PCR used primers designed to span the FIB-GAL sequence, which can be derived only 67

6 ARTICLE doi: / j.ymthe from the AAV vector (FIB, 5V-TGCTAGCAGTCCTGTGCCTG-3V; GAL, 5V- TGTGAGGCCATGCTTGTCGC-3V), leading to a 155-bp product. The PCR product was separated on a 1.8% NuSieve agarose gel. Kainic acid treatment. Seven days after piriform cortical infusion of the AAV vectors, the animals received a 10 mg/kg ip dose of kainic acid (Cayman Chemical Co., Ann Arbor, MI, USA). Using the limbic seizure scale of Racine [17], the latency was recorded to class I, II, III, IV, and V limbic seizure behaviors. In the case of the animals with recording electrodes, 30 min after kainic acid administration, the animalts electrode was attached to a Grass P55 AC preamplifier (3 Hz low filter, 100 Hz high filter; Grass Instruments). The output of the P55 was channeled into a PowerLabs 8/30 (ADInstruments, Colorado Springs, CO, USA) and the digital output signal was visualized and stored using Chart 5 software (ADInstruments). One hour after the first class IV or class V seizure, the animals received a 30 mg/kg dose of pentobarbital to terminate the seizures. Two weeks later, animals received an overdose of pentobarbital (100 mg/kg pentobarbital, ip) and subsequently were perfused transcardially with ice-cold 100 mm sodium phosphate-buffered saline (PBS) (ph 7.4), followed by 4% paraformaldehyde in 100 mm phosphate buffer (ph 7.4). After overnight fixation in the paraformaldehyde phosphate buffer, Vibratome sections (40 Am thick) were taken and rinsed in PBS. The sections were mounted and GFP fluorescence was visualized on an Olympus IX 70 fluorescence microscope. In addition, the infusion and electrode placements were determined. Electrical kindling of the piriform cortex and AAV vector infusion. Following a 7-day recovery period, the animals received a single bipolar stimulation (World Precision Instruments, A310 Accupulser and A365 stimulus isolator; 400 AA, 0.5-ms pulse width, 60 Hz, 2-s duration) each day until three class V seizures were evoked. On the subsequent test, the threshold for seizure genesis was determined by initiating the stimulation at 60 AA and increasing the stimulation current by 40 AA until limbic seizure activity was elicited. The threshold was determined at this point in the testing, because a hallmark of limbic kindling is the gradual reduction in seizure threshold over the course of repeated stimulations. Following this threshold test, the animals were sedated with 25 mg/kg pentobarbital, ip, and a 32-gauge injector was inserted into the 26-gauge cannula guide, such that the injector tip terminated near the electrode tips. As described above, 3 Al of either AAV-FIB-GAL (N = 9) or AAV-FIB- GFP (N = 5) was infused over a period of 30 min, and the injector was left in place for 3 min after the infusion. One week later, the same threshold titration procedure was used to evaluate the effects of the vector infusions on the threshold current necessary to evoke limbic seizure activity. Following this experiment, the animals were perfused, as described above, and the brains sectioned to locate the electrode placement. The area of GFP expression was visualized using an Olympus IX 70 fluorescence microscope. Statistics. Differences in the latencies to limbic seizure activity were evaluated using a Student t test, as separate control groups were tested for the AAV-FIB-GFP and AAV-FIB-GAL groups. There was no difference between the two control groups. For the limbic kindling study, a paired t test was employed, because each subject served as its own control. ACKNOWLEDGMENTS We thank Dr. Jude Samulski and the UNC Viral Vector Core for AAV vector production and Julia Hamra for expert technical assistance. These studies were supported by NIH Grant NINDS NS to T.J.M. RECEIVED FOR PUBLICATION FEBRUARY 15, 2006; REVISED MARCH 17, 2006; ACCEPTED APRIL 10, REFERENCES 1. Pieribone, V. A., Xu, S.-Q. D., Zhang, X., and Hfkfelt, T. (1998). Electrophysiologic effects of galanin on neurons of the central nervous system. Ann. N. Y. Acad. Sci. 863: Mazarati, A. M., et al. (1998). Galanin modulation of seizures and seizure modulation of hippocampal galanin in animal models of status epilepticus. J. Neurosci. 18: Haberman, R. P., Samulski, R. J., and McCown, T. J. (2003). Attenuation of seizures and neuronal death by adeno-associated virus vector galanin expression and secretion. Nat. Med. 9: Haberman, R. P., et al. (2002). Therapeutic liabilities of in vivo viral vector tropism: adeno-associated virus (AAV) vectors, NMDAR 1 antisense and focal seizure sensitivity. Mol. Ther. 6: Lin, E. J., Richichi, C., Young, D., Bauer, K., Vezzani, A., and During, M. J. (2003). Recombinant AAV-mediated expression of galanin in rat hippocampus suppresses seizure development. Eur. J. Neurosci. 18: Woldbye, D. P. D., Larsen, P. J., Mikkelsen, J. D., Klemp, K., Madsen, T. M., and Bolwig, T. G. (1997). Powerful inhibition of kainic acid seizures by neuropeptide Y via Y5-like receptors. Nat. Med. 3: Mazarati, A. M., and Wasterlain, C. G. (2002). Anticonvulsant effects of four neuropeptides in the rat hippocampus during self-sustaining status epilepticus. Neurosci. Lett. 331: Richichi, R., et al. (2004). Anticonvulsant and antiepileptic effects mediated by adenoassociated virus vector neuropeptide Y expression in the rat hippocampus. J. Neurosci. 24: Holmes, G. L., and Ben-Ari, Y. (2003). Seizing hold of seizures. Nat. Med. 9: Ben-Ari, Y. (1985). Limbic seizure and brain damage produced by kainic acid: mechanisms and relevance to human temporal lobe epilepsy. Neuroscience 14: McNamara, J. O., Byrne, M. C., Dasheiff, R. M., and Fitz, J. G. (1980). The kindling model of epilepsy: a review. Prog. Neurobiol. 15: Goddard, G. V., McIntyre, D. C., and Leech, C. K. (1969). A permanent change in brain function resulting from daily electrical stimulation. Exp. Neurol. 25: Lfscher, W., and Ebert, U. (1996). The role of the piriform cortex in kindling. Prog. Neurobiol. 50: Frush, D. P., and McNamara, J. O. (1986). Evidence implicating dentate granule cells in wet dog shakes produced by kindling stimulations of entorhinal cortex. Exp. Neurol. 92: Ben-Ari, Y. (1990). Galanin and gilbenclamide modulate the anoxic release of glutamate in rat CA3 hippocampal neurons. Eur. J. Neurosci. 1: Xu, Z. Q., Zheng, K., and Hokfelt, T. (2005). Electrophysiological studies on galanin effects in brain progress during the last six years. Neuropeptides 39: Racine, R. J. (1972). Modification of seizure activity by electrical stimulation. II: Motor seizure. Electroencephalogr. Clin. Neurophysiol. 32: Zeng, G. (1998). Sticky-end PCR: new method for subcloning. Biotechniques 25: Haberman, R. P., McCown, T. J., and Samulski, R. J. (1998). Inducible long-term gene expression in brain with adeno-associated virus gene transfer. Gene Ther. 5: Paxinos, G., and Watson, G. (1998). The Rat Brain in Stereotaxic Coordinates. Academic Press, New York. 68