Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The function of serotonin (5-hydroxytryptamine; 5-HT) as a modulator of neuronal activity and, in particular, of transmitter release in the mammalian central nervous system is well established. 5-HT receptors located on soma/dendrites and on terminals of 5-HT neurons can function as autoreceptors (Barnes and Sharp, 1999; Piñeyro and Blier, 1999). On the other hand, 5-HT_1B,D, 5-HT₂, or 5-HT₃ heteroreceptors located on nonserotonergic neuronal terminals can regulate release of diverse neurotransmitters, such as acetylcholine, noradrenaline, dopamine, -aminobutyric acid, glutamate or cholecystokinin (Barnes and Sharp, 1999; Sarhan and Fillion, 1999).

We previously found that 5-HT can potently inhibit the release of endogenous glutamate from adult rat cerebellar slices by acting at 5-HT₁ and 5-HT₂ receptors (Maura et al., 1988). Experiments with synaptosomes isolated from the cerebellum by a standard procedure indicated that the 5-HT₁ receptors, belonging to the 5-HT_1D subtype, are sited on glutamate-releasing terminals of parallel/climbing fibers (Raiteri et al., 1986; Maura and Raiteri, 1996); surprisingly, release-regulating 5-HT₂ receptors appeared to be lost during the preparation of synaptosomes.

Immunohistochemical, electrophysiological, and pharmacological evidence indicates that, in addition to parallel/climbing fibers, cerebellar mossy fibers use glutamate as their major transmitter (Beitz et al., 1986; Somogyi et al., 1986). Terminals of mossy fibers are relatively large; therefore, during homogenization of cerebellar tissue, two types of synaptosomes are produced: "standard" synaptosomes (the smaller synaptosomes that concentrate in the crude mitochondrial fraction) and giant synaptosomes (larger synaptosomes sediment in the nuclear fraction). According to Israel and Whittaker (1965), giant synaptosomes originate from the large terminals of mossy fibers. Because of the different sedimentation properties, giant synaptosomes are not present in standard synaptosomal preparations. However, they can be obtained from the nuclear fraction of the cerebellar homogenates. We previously prepared giant synaptosomes from rat cerebellum (Maura et al., 1991). Their morphology appeared well preserved and similar to that originally described by Israel and Whittaker (1965). In particular, giant synaptosomes released endogenous glutamate upon K⁺-depolarization in a calcium-dependent manner, indicating good viability. Furthermore, the evoked release of glutamate was inhibited by 5-HT or by (±)-DOI, suggesting that the 5-HT₂ receptors originally identified in cerebellar slices could be heteroreceptors located on mossy fiber endings lost during standard synaptosomal preparation. Accordingly, 5-HT₁ agonists displayed no activity in the giant synaptosome preparation indicating little or no contamination by standard glutamatergic synaptosomes carrying 5-HT_1D receptors (Maura et al., 1991).

Receptors of the 5-HT₂ type have been shown to exist in at least three structurally different subtypes termed 5-HT_2A, 5-HT_2B, and 5-HT_2C; moreover, ligands endowed with selectivity for these subtypes have been developed (for review, see Barnes and Sharp, 1999). It was therefore possible to approach a more detailed pharmacological characterization of the release-regulating 5-HT₂ receptor presumably present on rat cerebellar mossy fibers. This work was carried out mainly with preparations of giant synaptosomes. 5-HT and (±)-DOI, which display similar affinities for the 5-HT_2A, 5-HT_2B, or 5-HT_2C subtype (Hoyer et al., 1994; Baxter et al., 1995) were used as agonists; the results originally obtained with the two drugs (Maura et al., 1991) have been completed by constructing full concentration-response curves in this study. The known antidepressant trazodone, generally thought of as a 5-HT receptor antagonist (Haria et al., 1994; Takeuchi et al., 1997), was studied because during preliminary experiments, carried out previously in our laboratory, it had exhibited some activity at the 5-HT₂ receptors on giant cerebellar synaptosomes (see also Garrone et al., 2000). Various 5-HT₂ receptor antagonists were used to discriminate between receptor subtypes. In a set of experiments the role of endogenous 5-HT acting at the 5-HT₂ receptors present on mossy fiber terminals in regulating glutamate release was investigated using cerebellar slices.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Animals. Adult Sprague-Dawley male rats weighing 200-250 g were housed at constant temperature (22 ± 1°C) and relative humidity (50%) under a regular light/dark schedule (lights on 7:00 AM to 7:00 PM). Food and water were freely available. The animals were killed by decapitation. The cerebellum was rapidly removed and placed in ice-cold medium. Experimental procedures were approved by the Ethical Committee of the Pharmacology and Toxicology Section, Department of Experimental Medicine in accordance with the European legislation (European Communities Council Directive of 24 November 1986, 86/609/EEC).

Experiments with Giant Synaptosomes. Giant synaptosomes were prepared according to Israel and Whittaker (1965) with some modifications. The cerebellum was removed and homogenized in 40 volumes of 0.32 M sucrose buffered with phosphate at pH 7.4. The homogenate was centrifuged at 1000g for 5 min. The pellet (P₁; crude nuclear fraction) was resuspended in an equal volume of sucrose 0.32 M, pH 7.4, filtered through a double gauze layer, and then centrifuged at 1000g for 5 min. The pellet containing the giant synaptosomes was resuspended in a physiological salt solution of the following composition: 125 mM NaCl, 3 mM KCl, 1.2 mM MgSO₄, 1.2 mM CaCl₂, 1 mM NaH₂PO₄, 22 mM NaHCO₃, and 11 mM glucose (aerated with 95% O₂ and 5% CO₂ at 37°C), pH 7.2 to 7.4. Giant synaptosomes were incubated 15 min at 37°C in standard medium. After incubation, identical aliquots of the synaptosomal suspension were distributed at the bottom of a set of parallel superfusion chambers maintained at 37°C (Raiteri et al., 1974). Superfusion was started at a rate of 0.6 ml/min with standard medium aerated continuously with 5% CO₂ in O₂. Glutamate release was evoked by exposure to 15 mM KCl (replacing an equimolar concentration of NaCl) for 120 s, after 38 min of superfusion. Two 3-min samples (basal outflow) were collected before and after the 6-min sample containing the glutamate released by the depolarization pulse. Superfusate fractions were collected in plastic vials and rapidly frozen. Agonists were added concomitantly with high K⁺ and antagonists 8 min before depolarization. Glutamate released in the collected fractions was expressed as picomoles per milligram synaptosomal protein or as percentage of the glutamate tissue content at the beginning of the fraction. Protein determination was carried out according to Bradford (1976). The K⁺-evoked overflow was calculated by subtraction of glutamate content in the two 3-min fractions (pre- and poststimulation basal outflow) from the total content in the 6-min fraction corresponding to high K⁺ stimulation and was measured as percentage of increase with respect to the prestimulation basal outflow. Depolarization-evoked overflow in the presence of drugs was calculated as percentage of variation with respect to the control.

Experiments with Slices. The isolated cerebellum was chopped into 250-µm slices using a McIlwain tissue chopper. After 10 min of incubation at 37°C in standard medium, slices were transferred at the bottom of a set of parallel superfusion chambers (3 slices per chamber) at 37°C. Superfusion was started at a rate of 0.6 ml/min with standard medium aerated with 5% CO₂ in O₂. Two 120-s depolarizations (35 mM KCl replacing an equimolar NaCl concentration) were applied 38 and 78 min after start of superfusion (S₁ and S₂, respectively). Two 4-min samples (basal outflow) were collected before and after 8-min samples containing glutamate released by depolarization pulses; superfusate fractions were collected in plastic vials and rapidly frozen. Glutamate released in the collected fractions was expressed as picomoles per milligram protein or as percentage of the glutamate tissue content at beginning of the fraction. The K⁺-evoked overflow during S₁ or S₂ was calculated by subtraction of glutamate content (picomoles or percentage of the glutamate tissue content) in the pre- and poststimulation basal fractions from the total content in the 8-min fractions corresponding to K⁺ stimulation. The antagonists were added 8 min before S₂; their effects on K⁺-evoked overflow were measured as percentage of variation of the S₂/S₁ ratio with respect to the control value.

Glutamate Determination. The amount of endogenous glutamate released in the fractions collected (or remaining in synaptosomes or slices) was measured by high-performance liquid chromatography. Glutamate content of synaptosomes or slices was measured in the supernatant obtained after homogenization (Ultra Turrax, M. Cella, Milan, Italy; maximum speed, 20 s) in ice-cold distilled water and centrifugation at 20,000g for 10 min. The analytical method involved automatic precolumn derivatization (Waters 715 ultra wisp; Waters, Milford, MA) with o-phthalaldehyde followed by separation on C₁₈ reverse phase chromatography column (Chrompack International, Middleburg, The Netherlands; 10 cm × 4.6 mm, 3 µm) and fluorometric detection. Buffers and gradients program were as follows: solvent A, 0.1 M sodium acetate (pH 5.8)/methanol, 80:20; solvent B, 0.1 M sodium acetate (pH 5.8)/methanol, 20:80; solvent C, 0.1 M sodium acetate (pH 6.0)/methanol, 80:20; gradient program, 100% solvent C for 4 min; 90% solvent A and 10% solvent B in 1 min; isocratic step for 2 min; 78% solvent A and 22% solvent B in 2 min; isocratic step for 6.5 min; 66% solvent A and 34% solvent B in 1.1 min; isocratic step for 1.5 min; 42% solvent A and 58% solvent B in 1.1 min; isocratic step for 3.5 min; flow rate 0.9 ml/min (Waters 600 MS gradient system). Homoserine was used as internal standard. The detection limit was 100 fmol/µl.

Calculation. EC₅₀ values for agonists were determined from curves obtained using a four-parameter logistic function fitting routine (Sigma Plot software). pD₂ (log EC₅₀) values were taken as measure of the agonist potency. Means ± S.E. of n number of experiments are presented throughout. Significance of the difference was analyzed by analysis of variance followed by post hoc Scheffé multiple comparison or Student's t test. Level of significance was set at p < 0.05.

Drugs. The following drugs were purchased: (±)-1-(2,5-dimethoxy-4-iodophenyl)-2-aminopropane [(±)-DOI] from Sigma/RBI (Natick, MA); ketanserin from Tocris Cookson (Bristol, UK); 5-HT creatinine sulfate from Calbiochem (Los Angeles, CA). The following drugs were donated: R-(+)--(2,3-dimethoxyphenyl)-1-(2-(4-fluorophenyl)ethyl)-4-piperidine-methanol (MDL 100907) from Hoechst Marin Russel (Cincinnati, OH); 6-chloro-5-methyl-1-[6-(methylpyridin-3-yloxy)pyridin-3yl-carbamoyl] indoline (SB 242084) from SmithKline Beecham Pharmaceuticals (West Sussex, UK); trazodone from Istituto Ricerche Francesco Angelini (Pomezia, Roma, Italy)

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Release from Giant Synaptosomes. The overflow of endogenous glutamate (6-min samples) evoked during depolarization with 15 mM KCl (120 s) of superfused rat cerebellar giant synaptosomes amounted to 470 ± 52.4 pmol/mg of protein (n = 14); the basal outflow in the 3-min samples collected before and after K⁺ depolarization amounted to 134 ± 11.8 pmol/mg of protein (n = 14) and 119 ± 9.5 pmol/mg of protein (n = 14), respectively. Expressed as percentages of the total synaptosomal glutamate content, the basal release in the 3-min sample collected before and after K⁺ stimulation amounted to 0.60 ± 0.05/min (n = 4) and 0.51 ± 0.04/min (n = 4), respectively, whereas the K⁺-evoked overflow (6-min sample) was 9.35 ± 0.90% (n = 4). Previous experiments (Maura et al., 1991) had shown that the overflow of endogenous glutamate evoked by K⁺ depolarization of rat cerebellar giant synaptosomes in superfusion was almost entirely Ca²⁺-dependent.

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Inhibition by 5-HT (), (±)-DOI (), or trazodone () of the K+ (15 mM)-evoked glutamate overflow from rat giant synaptosomes. Data are expressed as percentage of the control value. Agonists were added concomitantly with high K+. Points are means ± S.E. of three to six independent experiments in triplicate (three superfusion chambers for each experimental condition).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Antagonism by MDL 100907, but not by SB 242084, of the 5-HT inhibition of the K+-evoked glutamate overflow from rat giant synaptosomes. Data are expressed as percent inhibition of the K+-evoked overflow of glutamate. 5-HT was added concomitantly with high K+; antagonists 8 min before. Data are means ± S.E. of three to five independent experiments in triplicate. *, significant difference (p < 0.05) versus 5-HT alone.

View larger version (23K): [in this window] [in a new window]   Fig. 3.   Antagonism by ketanserin or MDL 100907 but not by SB 242084 of the (±)-DOI inhibition of the K+-evoked glutamate overflow from rat giant synaptosomes. Data are expressed as percent inhibition of the K+-evoked overflow of glutamate. (±)-DOI was added concomitantly with high K+; antagonists 8 min before. Data are means ± S.E. of three to six independent experiments in triplicate. *, significant difference (p < 0.05) versus (±)-DOI alone.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Antagonism by ketanserin or MDL 100907 of the trazodone inhibition of the K+-evoked glutamate overflow from rat giant synaptosomes. Data are expressed as percent inhibition of the K+-evoked overflow of glutamate. The agonist was added concomitantly with high K+; antagonists 8 min before. Data are means ± S.E. of three to four independent experiments in triplicate. *, significant difference (p < 0.05) versus trazodone alone.

Release from Cerebellar Slices. The fractional basal outflow of glutamate in the 4-min fraction collected before S₁ (30.79 ± 2.62 pmol/mg of protein/min; mean ± S.E., n = 6) was 0.163 ± 0.0145%/min. The glutamate overflow evoked by K⁺ (35 mM) depolarization during S₁ (758.7 ± 63.6 pmol/mg of protein; mean ± S.E., n = 6) amounted to 4.02 ± 0.32%. The mean control S₂/S₁ ratio (0.99 ± 0.07; mean ± S.E., n = 6) was significantly increased to 1.35 ± 0.10 (+37%; mean ± S.E., n = 6) when MDL 100907 (1 µM) was added 8 min before S₂. The S₂/S₁ ratio in the presence of the 5-HT_2A antagonist did not significantly differ from the value obtained in parallel superfusion chambers in the presence of ketanserin (1 µM): 1.32 ± 0.08 (mean ± S.E., n = 6).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Our results are compatible with the idea that serotonergic receptors of the 5-HT_2A subtype mediating inhibition of evoked glutamate release are present on the terminals of mossy fibers in the rat cerebellum. A number of considerations lead to this conclusion. The synaptosomal preparations used have previously been found to resemble morphologically the giant endings of cerebellar mossy fibers observed in situ and purified from cerebellar homogenates (Israel and Whittaker, 1965; Maura et al., 1991). In addition, these preparations release glutamate upon depolarization in a Ca²⁺-dependent manner (Maura et al., 1991), and they are insensitive to agonists at 5-HT₁ receptors known to be localized on parallel fiber/climbing fiber axon terminals (Raiteri et al., 1986; Maura and Raiteri, 1996). Furthermore, the system of synaptosome superfusion used (a very thin layer of synaptosomes up-down superfused; for technical details, see Raiteri and Raiteri, 2000) can prevent indirect effects. Any compound released is rapidly removed by the superfusion fluid before it can act on the releasing terminal or on neighboring particles; therefore, the release modulations provoked by a drug added to the superfusion fluid can be assumed to be due to direct actions on the releasing particle. In our case, the inhibition of glutamate release caused by 5-HT/(±)-DOI is therefore probably due to direct action on 5-HT₂ receptors sited on mossy fiber glutamatergic terminals. Finally, the use of selective drugs tends to exclude the involvement of 5-HT₂ receptors of the 5-HT_2B/C subtype and favors the conclusion that the receptors in point are 5-HT_2A subtype. Ketanserin, a 5-HT_2A/5-HT_2C receptor antagonist (Baxter et al., 1995), blocked the (±)-DOI effect, in accordance with previous results obtained using 5-HT (Maura et al., 1991). The inability of the potent and selective 5-HT_2C receptor antagonist SB 242084 (Kennett et al., 1997) to block the effect of 5-HT tends to exclude the involvement of the 5-HT_2C receptor subtype. The finding that the selective 5-HT_2A receptor antagonist MDL 100907 (Baxter et al., 1995) could prevent the effect of 5-HT or (±)-DOI allows us to classify the serotonergic receptor located on the glutamatergic giant nerve terminals of the rat cerebellar cortex as 5-HT_2A subtype.

Trazodone could inhibit the K⁺-evoked glutamate release from giant synaptosomes, although with potency and efficacy lower than those of 5-HT or (±)-DOI. These findings, together with the antagonism by ketanserin or MDL 100907 of the effect of trazodone, suggest that the compound can behave as a partial agonist at the 5-HT₂ presynaptic receptors sited on mossy fibers. Based on the technical arguments discussed above it can be assumed that trazodone acts directly on glutamate-releasing mossy fiber terminals by activating the 5-HT_2A receptors located on these terminals. It should be added that the superfusion fluid should immediately remove trazodone metabolites possibly formed during superfusion before they can act on nerve terminals. Thus the release inhibition observed most likely represents a genuine effect of trazodone at presynaptic 5-HT_2A receptors. When this work was almost completed, a paper was published proposing that trazodone inhibition of glutamate release from cerebellar mossy fiber synaptosomes could be related to "its effects on a determinant common to  compounds" and simultaneous action "as a partial 5-HT receptor agonist" (Garrone et al., 2000). Our data are in line with the idea that trazodone is a partial agonist at 5-HT receptors regulating glutamate release from cerebellar mossy fiber synaptosomes. In addition, we classified pharmacologically the receptor involved as 5-HT_2A subtype. The possibility that 5-HT_2A and  receptors cooperate in regulating glutamate release cannot be ruled out and certainly a better understanding of such a possible receptor-receptor interaction would merit attention. No doubt all these new results add further complexity to the pharmacological profile of trazodone, an antidepressant drug formerly considered a 5-HT₂ receptor antagonist (see Haria et al., 1994).

In the cerebellar cortex no morphological evidence for presynaptic 5-HT_2A receptors has been provided; receptors of this subtype were found to be expressed in the somatodendritic region of Purkinje cells (Maeshima et al., 1998). Nevertheless, it is of interest that mRNA encoding for the 5-HT_2A receptor or 5-HT_2A receptor-like protein was detected in reticular and pontine nuclei (Palacios et al., 1993; Pompeiano et al., 1994), from which mossy fibers project to the cerebellar cortex (see Palacios et al., 1993). The cerebellum has usually been considered an area of the central nervous system extremely poor in 5-HT receptors, based on binding or morphological studies. However, types and subtypes of the 5-HT receptor have been identified and characterized in this area as release regulating presynaptic auto- and heteroreceptors in functional studies (Raiteri et al., 1986, 1991; Maura et al., 1991; Maura and Raiteri, 1996; see below).

Inhibition of neurotransmitter release by presynaptic receptors coupled to phosphoinositide hydrolysis, including the 5-HT_2A subtype (see Lucaites et al., 1996; Grotewiel and Sanders-Bush, 1999), is not surprising. Indeed, either potentiation or inhibition of glutamate release was observed following activation of phosphoinositide-coupled metabotropic glutamate receptors; in this case, a "switch" from potentiation to inhibition of release would depend on nerve terminal activity (Herrero et al., 1998). As to 5-HT_2A receptors, Ca²⁺-dependent opening of hyperpolarizing K⁺ channels was observed after 5-HT_2A receptor activation in C6 glioma cells (Bartrup and Newberry, 1994). Clearly, the precise mechanism by which 5-HT_2A receptors mediate inhibition of the evoked glutamate release in cerebellar mossy fiber terminals remains to be determined.

A possible physiological role for the presynaptic 5-HT_2A receptors characterized in this study is suggested by the effect of MDL 100907 in slices. The ability of the selective 5-HT_2A antagonist to increase glutamate release during depolarization of cerebellar slices indicates that, at least under some conditions of stimulation, endogenously released 5-HT can reach 5-HT_2A receptors able to modulate glutamate release onto granule cells.

Serotonergic projections originating from raphe nuclei or other nuclei in the reticular formation terminate with fine and diffuse varicosities in the three layers of the cerebellar cortex (Hökfelt and Fuxe, 1969; Chan Palay 1975; Bishop and Ho, 1985) and seem to exert a very sophisticated inhibitory control of glutamatergic transmission. Electrophysiological studies have shown that 5-HT can affect the firing of rat cerebellar neurons and modulate excitatory amino acid effects (see Strahlendorf and Hubbard, 1983; Lee et al., 1985). Based on our studies with isolated nerve terminals and cerebellar slices (Raiteri et al., 1986, 1991; Maura et al., 1988, 1991, 1995; Maura and Raiteri, 1996; Marcoli et al., 1997, 1998), multiple types, and subtypes of 5-HT receptors appear involved in the control of cerebellar glutamate transmission in the adult rat. The mechanisms identified include glutamate release regulation by 5-HT_1D and 5-HT_2A presynaptic receptors and inhibition by postsynaptic 5-HT_1A and 5-HT_2C receptors of the nitric oxide/cyclic GMP pathway coupled to ionotropic glutamate receptor activation (Raiteri et al., 1986, 1991; Maura et al., 1995; Maura and Raiteri, 1996; Marcoli et al., 1997, 1998).

A better knowledge of the serotonergic control of cerebellar glutamate transmission is of potential interest for a number of pathological conditions including cerebellar ataxia. In fact, ionotropic glutamate receptor overactivation and excitotoxic degeneration of Purkinje or granule cells have been proposed to be involved in cerebellar ataxia (Butterworth, 1993; O'Hearn and Molliver, 1997; Zuo et al., 1997). Significant improvements were observed in ataxic patients administered 5-HT_1A receptor agonists (Lou et al., 1995; Trouillas et al., 1996). Trazodone, potentially able to activate postsynaptic 5-HT_2C receptors inhibiting the glutamate/nitric oxide/cyclic GMP pathway (Marcoli et al., 1998) and presynaptic 5-HT_2A receptors controlling glutamate release onto granules (Garrone et al., 2000; present work), might display antiataxic effects.