In vivo interaction between serotonin and galanin type 1 and type 2 receptors in dorsal raphe: implication for limbic seizures

Abstract

Neuropeptide galanin suppresses seizure activity in the hippocampus by inhibiting glutamatergic neurotransmission. Galanin may also modulate limbic seizures through interaction with other neurotransmitters in neuronal populations that project to the hippocampus. We examined the role of galanin receptors types 1 and 2 in the dorsal raphe in regulation serotonergic transmission and limbic seizures. Infusion of a mixed agonist of galanin receptors 1 and 2, galanin (1-29), into the dorsal raphe augmented the severity of limbic seizures both in rats and in wild type mice, and concurrently reduced serotonin concentration in dorsal raphe and in the hippocampus, measured by immunofluorescence, or HPLC. In contrast, injection of galanin 2 receptor agonist galanin (2-11), mitigated the severity of seizures in both species, and increased serotonin concentration in both areas. Injection of both galanin fragments into the dorsal raphe of galanin receptor 1 knockout mice exerted anticonvulsant effects. Both proconvulsant activity of galanin (1-29) and seizure suppression by galanin (2-11) were abolished in serotonin – depleted animals. Our data indicate that in the dorsal raphe, galanin 1 and 2 receptors modulate serotonergic transmission in a negative and a positive fashion respectively, and that these effects translate into either facilitation, or inhibition of limbic seizures.

Anticonvulsant effects of the neuropeptide galanin have been well established, and best described in models of limbic epilepsy (Haberman et al., 2003; Kokaia et al., 2001; Lin et al., 2003; Mazarati et al. 1998, 2000, 2004a, Mazarati et al. b, Mazarati and Lu, 2005). Two types of galanin receptors, GalR1 and GalR2, contribute to the inhibition of epileptic activity (Mazarati et al., 2004a,b). Suppression of seizures by galanin is due to the opening of G protein activated-, or ATP sensitive potassium channels, and ultimately presynaptic inhibition of glutamatergic transmission (Mazarati and Lu, 2005; Zini et al., 1993). In addition to regulating glutamatergic neurotransmission in the hippocampus, galanin coexists with classical neurotransmitters in other neuronal populations. The latter include acetylcholine in medial septum/diagonal band (Melander et al., 1986b); norepinephrine (NA) in locus coeruleus (Melander et al., 1986a, Xu et al., 1998a); serotonin (5-HT) in dorsal raphe (DR) (Hökfelt et al., 1998, Xu et al., 1998b). Hence, to fully understand the role of galanin in limbic epilepsy, it is important to learn how the peptide affects neurotransmission in extrahippocampal areas, which regulate hippocampal excitability. Serotonergic input into the hippocampus, which originates from 5-HT containing neurons in DR (Whitaker-Azmitia and Peroutka, 1990) regulates mood, cognition, and is involved in the pathophysiology of depression, anxiety and epilepsy (Graeff et al., 1996; Richter-Levin and Segal, 1996; Theodore, 2003). The major effect of 5-HT in the epileptic hippocampus is anticonvulsant, and depends on 5-HT_1A postsynaptic receptors (Gariboldi et al., 1996; Mauler et al., 2001; Sarnyai et al., 2000).

In DR, galanin is present in 5-HT negative nerve endings of unknown origin, which innervate serotonergic cells, and in a small number of 5-HT positive neurons (Hökfelt et al., 1998; Lu et al., 2005a,b). Galanin, acting through GalR1, hyperpolarizes the membrane of serotonergic neurons and inhibits 5-HT release in the hippocampus (Kehr et al., 2002; Hökfelt et al., 1998; Xu et al., 1998b), thus regulating serotonergic transmission in tandem with 5-HT_1A autoreceptors (Lanfumei and Hamon, 2000).

Eighty percent of galanin receptors in DR account for GalR1, and the remaining- to GalR2 (Lu et al., 2005a,b); the expression of galanin receptor type three is negligible (Mennicken et al., 2004). Because of the prevalence of GalR1, the net effect of galanin in DR is conceivably GalR1- dependent. A role of GalR2 in regulating the activity of serotonergic cells is a subject for the discussion of the present manuscript. Since intracellular signaling mechanisms coupled to GalR1 and GalR2 are different (the latter mediate G_q-G₁₁ dependent inositol phospholipid hydrolysis, and intracellular calcium mobilization, Branchek et al., 2000), the two types of raphe galanin receptors might differentially regulate both hippocampal serotonergic innervation and seizure activity. In the present study we investigated GalR1 and GalR2 modulation of DR – hippocampal serotonergic projection, and its role in regulating limbic seizures. First, we examined how activation of DR galanin receptors affects pattern of limbic seizures induced by electrical stimulation of the entorhinal cortex – hippocampal pathway; next we tested the hypothesis that galanin receptor mediated modulation of seizures requires intact DR – hippocampus serotonergic pathway. Finally, we correlated regulation of seizures by GalR1 and GalR2 with their effects on serotonin concentration in the hippocampus.

Materials and Methods.

Animals.

The experiments were performed on 10–12 week old male Wistar rats (Harlan, Indianapolis, IN) and 10–14 week old male C57 Black mice. Wild type and GalR1 knockout mice were provided by Tamas Bartfai, The Scripps Research Institute, La Jolla, CA. The mice were originally generated at the Garvan Institute, Sydney, Australia as previously described (Jacoby et al., 2002), where they were backcrossed into the C57BL/6J line for five generations. The line was then re-derived at the Jackson Laboratory (Bar Harbor, ME) and backcrossed into C57BL/6J for two additional generations. GalR1 knockout mice and their wild type littermates were derived from the crossing of heterozygous breeding pairs (Mazarati et al., 2004b). Each group consisted of 4–6 animals, as indicated for each experiment in figure and table legends. Animal studies were approved by the Institutional Animal Care and Use Committee of Veteran Administration Greater Los Angeles Healthcare System, and complied with NIH policies on animal use.

Surgery.

Under Isoflurane anesthesia, animals were stereotaxically implanted with a bipolar stimulating electrode into the angular bundle of perforant path (rats: 0.5 mm rostral, 4.5 mm left from lambda, 4.5 mm down from brain surface; mice: 0.5 mm anterior and 2.5 mm left from lambda, 2 mm down from brain surface), and with a tripolar recording skull screw electrode. In addition, animals were implanted with the guide cannula into dorsal raphe. The cannula internal diameter was 0.24 mm, the length was 8 mm for rats and 5 mm for mice (manufacturer catalog number C315G). Coordinates for rats were 7.8 mm caudal, from bregma, 0 mm from midline, 6.2 mm down from brain surface (Paxinos and Watson, 1986). Coordinates for mice were: 4.16 caudal from bregma, 0 mm from midline, 3.3 mm down from brain surface (Paxinos and Franklin, 2001). Electrodes and cannulas were from Plastics One (Roanoke, VA).

Intracerebral injections of galanin receptor ligands.

Galanin (1-29), a mixed GalR1 and GalR2 agonist (Branchek et al., 2000), or galanin (2-11), a preferential GalR2 agonist (Elliot-Hunt et al., 2004; Hua et al., 2004; Kerekes et al., 2003; Liu et al., 2001) were dissolved in 0.9% NaCl and injected into the DR through the injection cannula, which had been connected to the Hamilton microsyringe (Hamilton, Reno, NE) placed into the sp310i syringe pump (World Precision Instruments, Sarasota, FL). The injection cannula was inserted into the lumen of the guide cannula and peptide solution was delivered in a volume of 1 μl, at a rate of 0.5 μl/min. Control animals were injected with the vehicle. The peptides were injected 1 hour prior to seizure induction, or 1.5 hours prior to euthanasia in immunofluorescence and HPLC studies.

5-HT depletion.

Rats received a single intraperitoneal (i.p) injection of 5-HT selective neurotoxin parachloroamphetamine (PCA), in a dose of 20 or 60 mg/kg (Garcia et al., 2003). To alleviate side effects of PCA (hyperthermia and motor excitation), the animals were pretreated i.p. with chloropromazine, 10 mg/kg, 1 hr prior to PCA injection. Rats were subjected to perforant path stimulation (PPS), or euthanized for biochemical assays 10 days after the injection.

Seizure induction and quantification.

Seven to ten days after surgery, animals were connected to the DS8000 stimulator (World Precision Instruments) for stimulation, and to MP100/EEG100B acquisition system (Biopac, Santa Barbara, CA) for electroencephalogram recording. Self-sustaining status epilepticus (SE) was induced as described earlier. Rats were subjected to 30 minutes of PPS (5 mA, 20 Hz, 10 s trains of 1 ms bipolar square wave pulses, delivered every minute, together with 2 Hz continuous stimulation with 5 mA, 1 ms unipolar square wave pulses) (Mazarati et al., 1998, 2004a).

In mice seizures were induced by 5 s trains of 0.1 ms, 5 mA, 33 Hz square wave bipolar stimuli delivered every minute, together with continuous 3 Hz stimulation of 0.1 ms 5mA square wave monopolar pulses, for a total of 60 minutes (Mazarati et al., 2000, 2004b). Electrographic activity was acquired using AcqKnowledge 3.7 software (Biopac) continuously for up to 24 hours. Electrographic seizure activity was analyzed off line by an unbiased investigator. Seizures were defined as polyspike events with the frequency of 3Hz or higher and with the duration of 3 s or longer (Mazarati et al. Mazarati et al., 2004a,b). Total SE duration, equal to the time between the end of the stimulation and the occurrence of the last seizure episode, and the cumulative seizure time, defined as SE duration less seizure-free periods, were calculated (Mazarati et al., 2004a,b). When seizures continued longer than 24 hours, this value was assigned to SE duration for statistical purposes.

Immunohistochemistry.

Animals were deeply anesthetized with Isoflurane and perfused with 4% paraformaldehyde containing 0.3% glutaraldehyde in phosphate buffer (pH=7.4). Brains were removed, embedded in Paraffin and cut in the coronal plane at 10 micron on Leica RM2255 microtome (Leica, Microsystems, Nussloch, Germany). After dewaxing in EZ DeWax Solution (Biogenex, San Ramon, CA), and quenching of endogenous peroxidase with 3% hydrogen peroxide (Sigma, St. Louis, MO), sections were incubated for 18–24 hours at 4°C with rabbit polyclonal 5-HT antibodies (Immunostar, Hudson, WI) diluted 1:2,000 for hippocampal labeling, or 1:10,000 for DR staining. For semi-quantitative evaluation of 5-HT turnover, sections of DR adjacent to those processed for 5-HT immunofluorescence, were incubated with rabbit polyclonal antibodies against 5-HT metabolite 5-hydroxyindolacetic acid (5-HIAA, 1:4,000, Immunostar). In addition, some DR sections were incubated with rabbit polyclonal antibodies against galanin (1:4,000, Bachem, Torrance, CA). Afterwards, sections were incubated for 1 hour with goat anti-rabbit secondary antibodies at a dilution 1:1000 (Perkin Elmer, Boston, MA) and processed for signal amplification using Tyramide Signal Amplification Plus Fluorescent System (Perkin Elmer) according to the manufacturer’s protocol.

Green fluorescent images of brain sections were acquired under 10 X (for fibers) or 20X (for cells) magnification using Olympus AX70 microscope and Sony DKC5000 digital camera (both from Tokyo, Japan). Relative fluorescence was normalized against the background for each section, and was calculated by an unbiased investigator using Adobe Photoshop 5.02. Semi-quantitative analysis of relative fluorescence was performed in DR sections corresponding approximately to 7.8 mm posterior from bregma, and in hippocampal sections corresponding approximately to 4.16 mm posterior from bregma (Paxinos and Watson, 1986), in the areas defined on the insets of Fig. 3. 5-HT turnover was expressed as a ratio of 5-HIAA positive to 5-HT positive cells in pairs of adjacent sections of DR.

Fig.3. Effects of parachloroamphetamine (PCA, 60 mg/kg) treatment on 5-HT immunofluorescence in the rat dorsal raphe (DR) and in the hippocampus.

a–f: fluorescence photomicrographs illustrating 5-HT immunoflurescence in DR and hippocampus of rats pretreated with PCA 10 days prior to euthanasia. a–c: naïve rats, d-f- a and d- DR. Insets show higher magnification. b and e- stratum radiatum of CA1, c and f- same of CA3. Note profound depletion of 5-HT signal after PCA treatment. Scale bar is 50 μm, for main panels; for the insets on a and d the same scale bar corresponds to 12.5 μm. g and h- quantification of 5-HT immunofluorescence. Each group contained 4 animals. Insets (images from Paxinos and Watson, 1986) depict areas (shaded), in which quantification was performed. On g, DR is further magnified. Mean±SEM. *−p<0.05 vs. Control. (two-way ANOVA followed by Bonferroni post hoc test).

Measurement of 5-HT and NA in the hippocampus.

The concentration of 5-HT and NA was measured in the whole hippocampal tissue using a described method (Häidkind et al., 2004) with modifications. Animals were deeply anesthetized by Isoflurane, decapitated, brains were removed and hippocampi were dissected on ice. Tissue samples were weighed and placed in the solution containing 0.09 mol/L perchloric acid, 0.04 mmol/L EDTA and 5 mmol/L Sodium Bisulfite (1 ml per100 mg of wet tissue), homogenized using ultrasonic dismembranator (Fisher Scientific, Pittsburgh, PA), centrifuged at 4°C and 16,000 G in Eppendorf 1514 R centrifuge; supernatant was aliquotted at 200 μl and stored at −70°C.

For HPLC, 50 μl aliquots were used. HPLC was performed using HPLC system, equipped with L-ECD-6A electrochemical detector (Shimadzu, Kyoto, Japan), using HR-80 column, 4.6×8×3 μM (ESA Biosciences, Chelmsford, MA). The separation was done in isocratic elution mode using CAT-A-Phase II mobile phase (ESA). The measurements were done at an electrode potential of +0.7 V. The concentration of 5-HT and NA was expressed as pmoles per mg of wet tissue. Data were analyzed using Shimadzu EZ Start software.

Chemicals.

Galanin (1-29) and galanin (2-11) were from Bachem (Torrance, CA); PCA, chloropromazine, 5-HT and NA (the last two for HPLC standards) were from Sigma (St. Louis, MO).

Statistical analysis.

Data were analyzed using Prizm 4 software (GraphPad, San Diego, CA) by two-way, or one-way ANOVA followed by Bonferroni post hoc test, provided normal distribution. P<0.05 was accepted for statistical significance.

Results

In the first set of experiments we examined the effects of a non-selective GalR1 and GalR2 agonist galanin (1-29), and of a preferential GalR2 agonist galanin (2-11) on the severity of SE induced in the rat by 30 minutes PPS. Intra-DR injection of galanin (1-29) resulted in a dose-dependent augmentation of PPS-induced seizures. Proconvulsant effect of the peptide reached statistical significance at 0.5 nmole, when cumulative seizure time was 531±10 minutes vs. 337±9 minutes in controls, and total SE duration was 886±9.5 minutes vs. 716±37 minutes in controls; proconvulsant effect of galanin (1-29) further increased at 2.5 and 10 nmole (Fig.1).

Fig.1. Effects of galanin receptor agonists injected into the rat dorsal raphe (DR), on status epilepticus (SE) induced by 30 minutes perforant path stimulation.

A non-selective GalR1/GalR2 agonist galanin (1-29), or a selective GalR2 agonist galanin (2-11) were injected into DR 1 hour prior to 30 minutes PPS. Galanin (1-29) facilitated seizures in doses 2.5 and 10 nmole, evident as the increase of both cumulative seizure time and SE duration. Galanin (2-11) inhibited seizures in amounts of 2.5 and 10 nmole, although at 10 nmole the anticonvulsant effect was less pronounced. Data are presented as Mean±SEM. *− p<0.05 vs. Control; #− p<0.05 vs. 2.5 nmole (two-way ANOVA followed by Bonferroni post hoc test).

In contrast to galanin (1-29), intra-DR injection of galanin (2-11) attenuated the severity of PPS-induced convulsions. The effect was maximal at 2.5 nmole (cumulative seizure time 12±1.5 minutes, SE duration 65±5.3 minutes). A dose of 10 nmole was less effective in mitigating seizures, than 2.5 nmole (cumulative seizure time 166±59 minutes, SE duration 315±48 minutes), although SE was still less severe, than in control rats (Fig.1). For further studies, we selected 2.5 nmole as a test dose for each of the peptides. Thus, simultaneous activation of GalR1 and GalR2 in DR, as mimicked by galanin (1-29), facilitated limbic seizures, while preferential activation of GalR2 achieved through injection of galanin (2-11) exerted anticonvulsant effect. Since selective GalR1 agonists are not currently available, to further address the role of the two types of galanin receptors in seizures, we turned to GalR1 knockout mice. In wild type animals, the peptides (each in the amount of 2.5 nmole) modified PPS-induced seizure response in a manner similar to that in the rat: galanin (1-29) facilitated seizures (SE duration was 305±5 minutes vs. 223±4 minutes in controls), while galanin (2-11) acted as an anticonvulsant (SE duration 115±8 minutes) (Fig.2). In line with our previous studies (Mazarati et al., 2004b), GalR1 knockout mice exhibited more severe seizures in response to 60 minutes PPS (372±23 minutes), as compared to wild type animals. While intra-DR galanin (2-11) inhibited seizures in GalR1 knockout mice, the effect of galanin (1-29) on seizures was opposite to that observed in wild type mice: instead of facilitating seizure response, the peptide exerted anticonvulsant effect, comparable to that of galanin (2-11) (Fig.2).

Fig.2. Effects of galanin receptor agonists injected into dorsal raphe of wild type (GalR1+/+) and Galanin Receptor Type 1 knockout (GalR1 −/−) mice on status epilepticus induced by 60 minutes perforant path stimulation.

In wild type animals, galanin (1-29) (0.25 nmole) increased seizure severity, while galanin (2-11) (0.25 nmole) attenuated seizures. However, both peptides exhibited anticonvulsant effect in GalR1 knockout mice. Each group contained 4 animals. Mean±SEM *−p<0.05 vs. Control; #− p<0.05 vs. Galanin (1-29). (one-way ANOVA followed by Bonferroni post hoc test).

Thus, in the system, which lacked GalR1, and in which all effects of galanin (1-29) were mediated through GalR2, the peptide inhibited, rather than facilitated seizures, that is acted in the same fashion as a preferential GalR2 agonist. Taken together, the results of the first two sets of experiments suggested that activation of GalR1 and GalR2 in DR had opposite effects on limbic seizures: GalR1 were proconvulsant, while GalR2 exerted anticonvulsant effect.

We speculated that the regulation of seizures by DR GalR1 and GalR2 receptors involved modulation of serotonergic projection from DR into the hippocampus. To prove this point, we next examined, whether the peptides preserved their effects in a system in which 5-HT had been depleted by a systemic injection of a serotonin-selective neurotoxin PCA (60 mg/kg). To validate the system, we first examined the effects of PCA treatment on 5-HT immunofluorescence in the rat DR and in the hippocampus, as well as on monoamine concentration in the latter. In DR of naïve animals (Fig.3,a) abundant 5-HT immunofluorescence was observed both in fibers, and in cells (inset). In the hippocampus of naïve rats 5-HT fluorescence was confined to fibers only, and was most abundant in stratum radiatum in CA1 (Fig.3,b) and CA3 (Fig.3,c). Ten days after PCA injection, profound depletion of 5-HT from fibers was observed both in DR (Fig.3,d, g) and in the hippocampus (Fig. 3, e, f, h). In addition, the number of 5-HT positive cells in DR decreased. The observed depletion of 5-HT in the hippocampus was confirmed by HPLC: in PCA-treated animals the concentration of 5-HT fell below detectable level (Table 1). It is important to mention that the effect of PCA treatment was 5-HT selective, since it did not affect the concentration of NA (Table 1).

Table 1.

The effects of galanin (1-29), galanin (2-11) and parachloroamphetamine on the concentration of 5-HT and norepinephrine (NA) in the rat hippocampus.

Transmitter, pmole/mg	Control (n=5)	PCA60 (n=4)	PCA20 (n=5)	Gal(1-29) (n=5)	Gal(2-11) (n=5)
5-HT	0.297±0.007	(N/d)	0.148±0.002*	0.123±0.004*	0.485±0.034*
NA	0.497±0.035	0.427±0.012	0.423±0.027	0.433±0.009	0.490±0.067

Rats received i.p. parachloroamphetamine injection in doses of 20 (PCA20), or 60 (PCA60) mg/kg ten days prior to euthanasia. Galanin (1-29) (Gal(1-29)), or galanin (2-11) (Gal(2-11)) were injected into dorsal raphe in the amount of 2.5 nmole 1.5 hours prior to euthanasia. Catecholamine concentration was measured by HPLC. Numbers in parentheses indicate number of animals in a group. Data are expressed as Mean±SEM.

^*−p<0.05 vs. control (Two-Way ANOVA followed by Bonferroni post hoc test). After PCA60, 5-HT was not detectable (N/d).

Having established the effectiveness of PCA, we examined seizure response in 5-HT depleted rats. Thirty minutes of PPS administered 10 days after PCA injection, led to the development of SE, which was more severe, than in naïve animals: cumulative seizure time was 922±18 minutes, and all animals still exhibited seizure activity 24 hrs after the end of stimulation (Fig.4,a).

Fig.4. Systemic depletion of 5-HT in rats abolished effects of galanin peptides on seizures.

Parachloroamphetamine (PCA, 60 mg/kg, i.p.) was injected 10 days prior to perforant path stimulation PPS). a- Neither galanin (1-29) (Gal(1-29), n=6), nor galanin (2-11) (Gal(2-11), n=5) affected seizure severity in PCA-pretreated rats (n=5). b Fifteen minutes PPS delivered to PCA-treated animals (n=4) elicited seizures comparable to those observed in naïve animals after 30 minutes of stimulation (compare to fig 1., Control). However, under conditions of less severe seizures, both peptides failed to affect seizure severity (n=4 in each group).

Furthermore, systemic depletion of 5-HT completely abolished both proconvulsant effects of galanin (1-29) and anticonvulsant effects of galanin (2-11), when the peptides were injected into DR (Fig.4,a). It could be argued that no proconvulsant activity of galanin (1-29) in PCA treated animals could be revealed because PCA alone enhanced seizure severity to the maximal level. Similarly, it was possible, that galanin (2-11) failed to attenuate seizures, because the severity of the latter exceeded anticonvulsant potency of the peptide. To address these concerns, we subjected a separate group of PCA-treated rats to a PPS, which was limited to 15 minutes. This protocol failed to induce self-sustaining seizures in naive rats (not shown), but led to the development of seizures comparable to those observed in naive animals upon 30 minutes of stimulation (Fig.4,b). However, under these conditions, both galanin (1-29), and galanin (2-11) still failed to affect SE (Fig.4,b).

The results obtained from PCA-treated animals suggested that intact serotonergic transmission was required for GalR1 and GalR2 regulation of seizures. Since the outcomes of the activation of the two types of galanin receptor were opposite, it was conceivable, that GalR1 and GalR2 regulate serotonergic transmission in opposite directions. To study the regulation of DR-hippocampal serotonergic transmission by GalR1 and GalR2 we examined 5-HT immunofluorescence in the DR and in the hippocampus, as well as hippocampal monoamine concentration in the animals, which had been injected with galanin (1-29), or galanin (2-11) into DR.

Injection of galanin (1-29) into DR of naïve rats resulted in significant reduction of 5-HT immunofluorescence in the DR, as well as in both hippocampal subfields (Fig. 5, c-f). In DR, the attenuation of 5-HT signal was due to the depletion of transmitter from fibers, whereas 5-HT neuronal count was not affected (Fig.5,f). The effect of galanin (1-29) on serotonergic innervation of the hippocampus was confirmed by HPLC, which revealed a 60% decrease of 5-HT concentration, as compared to hippocampal tissue obtained from naïve animals (Table 1).

Fig.5. Effects of galanin (1-29) injected into the rat dorsal raphe (DR) on 5-HT immunofluorescence in DR and in the hippocampus.

a and b- galanin immunofluorescence on the DR of naïve rat (a) and 1.5 hours after injection of galanin (1-29) (2.5 nmole). Intense fluorescence in b reflects exogenously administered galanin. c- 5-HT immunofluorescence in the DR 1.5 hours after injection of galanin (1-29) (compare to 5-HT staining in naïve animal on 3a); inset shows higher magnification. d- 5-HT immunofluorescence in CA1 and e- in CA3 1.5 hours after intra-DR injection of galanin (1-29) (compare to 5-HT staining in naïve animal Fig.3b,c). Scale bar is 50 μm, for main panels; for the inset on c the same scale bar corresponds to 12.5 μm.. f- Quantification of 5-HT. Each group contained 4 animals. Administration of galanin (1-29) led to significant reduction of 5-HT immunofluorescence in all areas of interest; the number of 5-HT-positive neurons in DR was not changed. Mean±SEM *−p<0.05 vs. Control. (two-way ANOVA followed by Bonferroni post hoc test).

We failed to find any changes in 5-HT immunofluorescence in the DR and in the hippocampus after administration of galanin (2-11) (not shown). We speculated that if galanin (2-11) stimulated 5-HT (an effect opposite to that of galanin (1-29)), the effect of the peptide might not be noticeable due to the saturation of fluorescent signal. Therefore, we studied the effects of galanin (2-11) in the animals, in which 5-HT had been partially depleted by PCA in a dose of 20 mg/kg. In this dose PCA resulted in a significant depletion of 5-HT concentration in the DR, CA1 and CA3 of the hippocampus, evident by both immunofluorescence (Fig.6 a-c,g), and by HPLC (Table 1), although the extent of 5-HT decrease was not as robust as after PCA, 60 mg/kg. Importantly, PCA, 20 mg/kg did not result in the loss of 5-HT positive neurons in DR, suggesting that the source of the transmitter was not affected (Fig. 6, g). In PCA (20 mg/kg) pretreated animals galanin (2-11) returned the level of 5-HT immunofluorescence both in DR and in the hippocampus to the values observed in naïve rats (Fig.6, d-g). Furthermore, in naive animals (without PCA treatment), we detected a significant (63%) increase of 5-HT concentration in the hippocampus after galanin (2-11) injection into DR (Table 1).

Fig.6. Effects of galanin (2-11) injected into dorsal raphe (DR) on 5-HT immunofluorescence in DR and in the hippocampus of rats pretreated with parachloroamphetamine, 20 mg/kg (PCA20).

PCA treatment led to the decrease of 5-HT immunofluorescence in DR (a), CA1 (b) and CA3 (c) (compare to staining in naïve animals on Fig.3,a-c). Galanin (2-11) reversed PCA-induced decrease of 5-HT labeling (d.- DR, e- CA1, f- CA3). Scale bar: 50 μm. g- Quantification of the 5-HT immunofluorescence. Each group contained 4 animals. Mean±SEM. *−p<0.05 vs. Control; #− p<0.05 vs. PCA20 (two-way ANOVA followed by Bonferroni post hoc test).

To address the possibility that galanin receptors modulate 5-HT concentration through regulating transmitter metabolism, we compared immunofluorescence correlate of 5-HT turnover in DR sections of rats injected with either galanin (1-29), or galanin (2-11), with their respective controls (saline and PCA, 20 mg/kg). 5-HIAA immunofluorescence was not detectable in fibers, but was rather confined to cells, with approximately 4-fold higher abundance, as compared to 5-HT (not shown). We found no detectable changes in the 5-HIAA positive to 5-HT positive cell ratio after peptide administration (saline: 4.28±0.42, galanin (1-29): 4.38±0.51, p>0.05; PCA, 20 mg/kg: 4.26±0.29, PCA, 20 mg/kg + galanin (2-11): 4.48±0.37, p>0.05).

Thus, galanin (1-29) administered into DR resulted in the depletion of 5-HT in the site of injection, as well as in the hippocampus, while preferential activation of GalR2 by galanin (2-11) produced an increase of 5-HT concentration in these areas. The observed changes resulted from alterations of 5-HT release, rather than from changes in the neurotransmitter metabolism.

Discussion.

The major findings of the current study may be summarized as follows. First, the activation of GalR1 in DR facilitated, while the activation of GalR2 inhibited limbic seizures. Second, modulating effects of GalR1 and GalR2 in DR while opposite, both employed DR-hippocampal serotonergic projection. Finally, proconvulsant action of GalR1 was mediated through the inhibition of DR - hippocampus serotonergic input, while inhibition of seizures by GalR2 occurred through the increase of serotonin release. Activation of both GalR1 and GalR2 as mimicked by intra-DR administration of galanin (1-29), a natural agonist for both receptor subtypes (Branchek et al., 2000), likely reflects a scenario of how endogenous galanin controls 5-HT release. Since GalR1 in DR are more abundant, the net effect of galanin in this area pertains to the activation of GalR1, rather than of GalR2. Indeed, the application of a preferential GalR2 agonist, or of a system lacking GalR1 (i.e. GalR1 knockout mice) revealed the effect of preferential GalR2 activation, which appeared to be opposite to that GalR1. Furthermore, the reduction of anticonvulsant efficacy of galanin (2-11) in the highest applied amount (10 nmole), as compared to the amount of 2.5 nmole, might be attributed to the recruitment of GalR1, since it has been known, that galanin (2-11) in high concentrations (in vitro concentration of 5 μM and higher) acts both as GalR1 and GalR2 agonist (Hua et al., 2004; Liu et al., 2001; Lu et al., 2005a).

There is a potentially important translational implication of the reported differences between GalR1 and GalR2 in DR. The emergence of non-peptide low-molecular weight galanin receptor agonists (Bartfai et al., 2004; Saar et al., 2002) may lead to the development of synthetic galanin receptor ligands for the treatment of neurological disorders. Since the most practical mode of their delivery is systemic administration, such compounds would exert their effects in all neuronal populations which express target galanin receptors with possibly different, or even opposite (as our study suggested) outcomes. This concept, which is generally applicable to all synthetic ligands of neuropeptide receptors, is particularly relevant for galanin: the hippocampus expresses lower level of galanin peptide and galanin receptors, as compared to many extrahippocampal areas (Branchek et al., 2000; Depczynski et al., 1998; Lu et al., 2005a; O’Donnel et al., 1999;). With regard to epilepsy, potential beneficial effects of GalR1 activation in the hippocampus might be mitigated, or even negated by an undesirable inhibition of serotonergic innervation; this particular scenario prompts the development of selective GalR2 agonists, since GalR2 activation has been anticonvulsant both in the hippocampus (Mazarati et al., 2004a) and in DR.

Our studies confirmed the important role of serotonergic transmission in regulating hippocampal excitability and seizure activity. In line with previous studies (Gariboldi et al., 1996; Mauler et al., 2001; Sarnyai et al., 2000) we found that systemic depletion of 5-HT led to profound increase of the propensity to, and of the severity of limbic seizures. Furthermore, we established that an intact DR-hippocampal serotonergic projection is required for seizure modulating effects of both GalR1 and GalR2 in DR, since neither galanin (1-29), nor galanin (2-11) affected seizures in 5-HT depleted animals. Further immunohistological and biochemical analysis revealed, that proconvulsant effect of GalR1 correlated with the reduction of 5-HT in the DR and in the hippocampus, while inhibition of seizures by GalR2 paralleled with the increase of 5-HT concentration in both brain areas. These data, along with the lack of the effects of the peptides on 5-HT turnover, suggested that the activation of GalR1 decreased, while the activation of GalR2 increased 5-HT release, conceivably through the regulation of firing of 5-HT neurons. While GalR1- mediated inhibition of 5-HT release has been established (Kehr et al., 2002; Hökfelt et al., 1998; Xu et al., 1998b), a possible role of GalR2 in increasing serotonergic DR transmission has recently emerged (Lu et al., 2005a). Several observations, suggested that the effects of galanin in DR are not quite straightforward. Indeed, from the standpoint of the antagonism between galanin and 5-HT, it has been difficult to explain why synthetic agonists of galanin receptors exhibit antidepressant-like effect (Bartfai et al., 2004; Lu et al., 2005a). It thus has been speculated that GalR2 in DR increase firing rates of serotonergic neurons and of 5-HT release, an effect congruent with the intracellular signaling cascades coupled to GalR2 (Branchek et al. 2000; Lu et al., 2005a, Wang et al., 1998). Interestingly, chronic administration of fluoxetine, a serotonin uptake inhibitor commonly used as an antidepressant, upregulated GalR2, but not GalR1 in DR (Lu et al., 2005a). If indeed GalR2 enhance serotonergic transmission, as Lu et al (2005a) has suggested and as we have shown in the present study, this action of fluoxetine might be one of the mechanisms of its antidepressant effect. 5-HT_1A autoreceptors in DR have been known as a key feedback mechanism regulating the release of 5-HT from serotonergic neurons (Albert et al., 2001; Lanfumei and Hamon, 2000; Mauler et al., 2001). The presence of two galanin receptor types in DR might represent another fine-tuning tool in regulating serotonergic transmission and related hippocampal activity.

Another promising clinical implication of GalR2 mediated positive regulation of serotonergic transmission, is that GalR2 agonists might be beneficial not only for treating seizure syndrome, but also for depression which frequently accompanies epilepsy (e.g. Harden and Goldstein, 2002). Indeed, concurrent antidepressant effect of anticonvulsant therapy has been regarded important in managing epileptic patients with co-morbid depression (Sankar, 2004).

In conclusion, galanin regulates seizures not exclusively through the interaction with glutamatergic neurotransmission in the hippocampus, but also by modulating the activity of other neurotransmitters in different brain areas. In the DR, GalR1 and GalR2 activation affects seizure activity in opposite ways. It is conceivable, that the role of the two galanin receptor types in other neuronal populations also affects seizures differentially. This calls for further studies of the interaction between galanin receptors and classical neurotransmitters in such areas as cholinergic basal forebrain and locus coeruleus, and for the development of subtype – selective galanin receptor ligands.