The absence of differences between GalR knockout and wild
The absence of differences between GalR1 knockout and wild type mice in kainic acid-induced seizures suggests that GalR1 do not interfere with the seizure-inducing action of kainic acid. Postsynaptic kainate receptors are excitatory ionotropic glutamate receptors and their activation leads to seizures and neurotoxicity (Lerma et al., 2001). Presynaptic kainate receptors are metabotropic receptors, and their activation inhibits glutamatergic transmission (Chittajallu et al., 1996, Lerma et al., 2001, Kamiya and Ozawa, 1998). If GalR1 activation inhibits seizures through presynaptic inhibition of glutamate release (Mazarati et al., 2000, Zini et al., 1993b), rather than through downstream postsynaptic mechanisms, the absence of differences between GalR1 knockout and wild type mice in kainate-induced seizures are not surprising. On the other hand, kainic homoharringtonine induced more severe seizures in galanin peptide knockout mice, compared to their wild type littermates, while overexpression of galanin gene mitigated kainate-induced convulsions (Mazarati et al., 2000). Taken together with the results obtained from GalR1 knockout animals, and considering that both GalR subtypes (or at least GalR1) are intact in galanin peptide mutant animals (Hohmann et al., 2003), these data suggest that GalR2 inhibit kainic acid-induced seizures through postsynaptic, rather than presynaptic mechanism. Postsynaptic localization of GalR2 was also implied in the studies on neuroprotective effects of galanin (see below). Self-sustaining SE induced by perforant path stimulation strongly depends on the enhanced release of glutamate from entorhinal cortex- dentate gyrus projection and subsequent activation of N-methyl-d-aspartate receptors (Wasterlain et al., 2000). Attenuation of galanin-mediated block of glutamate release from perforant path would promote SE, as it was observed in GalR1 KO animals. The role of GalR2 in seizures was first addressed by Mazarati et al. (2004a), who used an antisense approach, particularly peptide nucleic acid (PNA) antisense, to downregulate GalR2 in the hippocampus in vivo. PNA is a DNA or RNA mimic, which binds to DNA or RNA in complementary antiparallel fashion, thus inhibiting transcription or translation (Nielsen and Egholm, 1999, Pooga et al., 1998, Rezaei et al., 2001). PNA targeted at mRNA encoding GalR2 at positions 18–38 was administered into the dentate gyrus of the rat over a 1-week period. This resulted in 50% decrease of GalR2 binding in the infused hippocampus, without affecting GalR1 binding (Mazarati et al., 2004a). Such semi-acute focal partial GalR2 knockdown, even though not perfect, allowed studying GalR2 in seizure regulation. GalR2 knockdown animals required PPS of the same duration as controls (30 min) to elicit SE. However, once established, seizures were significantly more severe in GalR2 knockdowns than in control rats (treated with missense PNA). These results supported the idea, that GalR2 inhibited SE during its maintenance phase, and did not affect the initiation phase, as it was suggested for GalR1. Still, it is quite difficult to draw definitive conclusions on GalR subtype-specific regulation of seizures, due to the differences in species and methodologies, which could affect one phase of SE, and not the other (e.g., it is difficult to study maintenance phase in GalR knockout mice). Fig. 4 presents a diagram which summarizes the results of our experiments (Mazarati et al., 1998, Mazarati et al., 2004a, Saar et al., 2002, as well as personal unpublished data), which involved focal intrahippocampal GalR1/GalR2, or GalR2 activation, GalR1 subtype downregulation, or combinations of such treatments. Downregulation of GalR2 using PNA antisense resulted in the increase of seizure severity. Activation of GalR1 and GalR2 using non-selective agonist galanin (1–29) had strong anticonvulsant effects. However, selective activation of GalR2 without affecting GalR1, using galanin (2–11), Ala-2-galanin (1–29), or D-Trp-2 galanin (1–29) in the amounts equimolar to the effective dose of galanin (1–29), had anticonvulsant effect which was as strong as that of GalR1/GalR2 activation. Activation of GalR1 in the rats with GalR2 knockdown, achieved by galanin (1–29) injection into the GalR2 PNA-pretreated hippocampus attenuated seizures, but this effect was not as strong as after selective GalR2 activation. It must be noted that GalR2 PNA injection did not completely abolish GalR2 binding; therefore, even partial downregulation of GalR2 attenuated anticonvulsant effects of galanin. Finally, downregulation of GalR1 using GalR1 PNA antisense, increased seizure severity, however this increase was not as profound as after GalR2 downregulation. Overall, both GalR1 and GalR2 seem to contribute to the anticonvulsant effects of galanin, however, GalR2 activation has more profound anticonvulsant effect; furthermore, the above-discussed studies suggests that while GalR1 inhibit seizures through presynaptic inhibition of excitatory transmission, GalR2 might have postsynaptic component in their anticonvulsant effect.