Introduction

Top Abstract Introduction Methods Results Discussion References

The 5-HT₃ receptor is unique among monoamine receptor subtypes in constituting a cation-selective ion channel, that mediates rapid synaptic transmission in the mammalian CNS (Derkach et al., 1989). A prominent consequence of its activation is the modulation of release of other neurotransmitters, e.g., cholecystokinin (Paudice and Raiteri, 1991), dopamine (Blandina et al., 1989), GABA (Ropert and Guy, 1991), norepinephrine (Blandina et al., 1991). These actions may eventually provide plausible biochemical substrates for the various purported psychoactive properties of 5-HT₃ receptor antagonists (Passani and Corradetti, 1996; Zifa and Fillion, 1992). Of particular interest is a possible role in cognition, which is supported by the observations that 5-HT₃ receptor antagonists have been shown to enhance cognitive performance in several rodent and primate models. Improvement was detected not only under basal conditions in young and old animals, but also when performance was impaired by cholinergic deficits (Barnes et al., 1990). Serotonergic innervation of cholinergic septohippocampal neurons (Milner and Veznedaroglu, 1993) represents one morphological basis for functional interactions between these two transmitter systems. In addition, the cortex receives both cholinergic (Mesulam et al., 1983) and serotonergic projections (Lidov et al., 1980; Moore et al., 1978), permitting postsynaptic interactions on target cells as well as presynaptic modulation of transmitter release. The nature and extent of these interactions remain controversial. In vitro studies on synaptosomes from various regions of human neocortex (Maura et al., 1992) and minces of rat entorhinal cortex (Barnes et al., 1989) showed an inhibitory action of 5-HT on [³H]-ACh release mediated by the activation of 5-HT₃ receptors. An attempt to replicate the finding on rat cortex was, however, unsuccessful (Johnson et al., 1993). Moreover, activation of multiple 5-HT receptors may produce opposing actions on release, so that nonselective agonists may have minimal effects except in the presence of appropriate antagonists (Barnes et al.,, 1989). In vivo dialysis studies of serotonergic modulation of ACh release from rat frontal cortex have been restricted to the use of the 5-HT-releasing drugs fenfluramine (Hirano et al., 1995) and norfenfluramine (Consolo et al., 1996) which increased spontaneous endogenous ACh release in freely moving rats. Based on systemic administration of both the releasers and the antagonists, the actions of endogenous 5-HT on ACh were attributed to 5-HT_2A (Hirano et al., 1995) and 5-HT_1B receptors (Consolo et al., 1996), 5-HT₃ antagonists were not tested. Our study describes the role of the 5-HT₃ receptor in the modulation of K⁺-evoked release of ACh from cortex and hippocampus of freely moving rats. The transverse microdialysis technique was used.

	Methods

Top Abstract Introduction Methods Results Discussion References

Animal housing and surgery. Male Wistar rats (225-275 g body weight) were housed in groups of three in a temperature-controlled room (20-24°C), allowed free access to food and water and kept on a 12 hr light/dark cycle. The rats were anesthetized with chloral hydrate (400 mg/kg, i.p.) and placed in a stereotaxic frame (Stellar, Stoelting, Wood Dale, IL). Microdialysis tubes were inserted transversally in frontoparietal cortices or in dorsal hippocampi. The microdialysis tubes were made of AN 69 membrane (Dasco, Modena, Italy), 220 µm internal diameter and 310 µm external diameter, molecular weight cut-off  15,000 Da. The membrane, mounted on a guide wire of tungsten-stainless steel, was covered with Super-Epoxy glue except for an 8-mm section in frontoparietal cortices and 6-mm section in the dorsal hippocampi. A thin stainless steel cannula (22 gauge, 1 cm long) was glued to the distal end of the tubing. After a sagittal incision, the overlying skin and the temporal muscles were retracted, and holes for placement of the dialysis tube were drilled bilaterally at coordinates: AP = -0.5 mm and H = 2.0 mm for the cortex and AP = -3.3 mm and H = 3.3 mm for the hippocampus. All coordinates (Paxinos and Watson, 1982) were measured on the bone surface and referred to bregma, with bregma and lambda on a horizontal plane. One end of the guide wire of the microdialysis tubing was fixed to the micro-manipulator of the stereotaxic instrument, and gently pushed through the brain. The exposed portion of membrane was positioned precisely by means of two reference marks corresponding to the outer surface of the temporal bones. The inner guide wire was removed, and another steel cannula (22 gauge, 1 cm) was secured to the end of the dialysis tubing. The cannulae, bent upward, were secured to the parietal bone with acrylic dental cement and the skin sutured. All surgical manipulations were performed under deep anesthesia. Rats were then replaced in their home cages (one rat per cage) to recover from surgery.

Microdialysis experiments. Twenty-four hours after surgery, each rat was placed in a Plexiglas cage. The inlet of the microdialysis probe was connected to a microperfusion pump (Carnegie Medicine, mod. CMA/100, Solna, Sweden) and the outlet was inserted into a 200-µl test tube containing 5 µl of 0.5 mM HCl to prevent hydrolysis of ACh. The microdialysis tubing was perfused at the rate of 3 µl/min with Ringer solution (NaCl 147 mM, CaCl₂ 1.2 mM, KCl 4.0 mM, pH 7.0). To recover detectable dialysate concentrations of ACh, a cholinesterase inhibitor (physostigmine sulfate, 7 µM) was included in the perfusion solution. The molecular weight cut-off of the membrane allowed low molecular weight solutes to cross the dialysis membrane according to their concentration gradients. Hence, both the collection of endogenous molecules, and the administration of exogenous compounds were feasible.

HPLC method. ACh was determined by HPLC-electrochemical detection (Giovannini et al., 1994). The HPLC apparatus consisted of a pump (model 1350, BioRad, Richmond, CA), a presaturation column (Chromspher 5 C18, 100 × 3 mm, Chrompack, Middleburg, the Netherlands), an injector (model 7725, Rheodyne, Cotati, CA), a guard column (reverse phase), an analytical column (Chromspher 5 C18, 100 × 3 mm, Chrompack), an enzyme reactor (10 × 2.1 mm, Chrompack), an electrochemical detector (model LC4C, BioAnalytical System, West Lafayette, IN) and a Perkin Elmer (Foster City, CA) chart recorder. The analytical column was transformed into a cation exchange column by loading it with sodium lauryl sulfate (0.5 mg/ml). The enzyme reactor consisted of a 1-cm long column containing Lichrosorb-NH₂ activated with glutaraldehyde, and to which acetylcholinesterase (E.C. 3.1.1.7) and choline oxidase (E.C. 1.1.3.17) were covalently bound. The mobile phase had the following composition (mM): K- phosphate buffer 100, KCl 5, tetramethylammonium 1 and EDTA 0.3 (pH 8.0). ACh was separated on the cation exchange column. ACh was hydrolyzed by acetylcholinesterase to form acetate and choline in the postcolumn enzyme reactor, then choline was oxidized by choline oxidase to produce betaine and hydrogen peroxide. Hydrogen peroxide was detected by a platinum electrode with the potential set at + 0.5 V. The flow rate was 0.75 ml/min. Peaks were identified by comparison of their retention times with those of the standards.

Quantification of ACh. The levels of ACh in the perfusates were calculated by comparison of sample peak heights with external standard peak height and expressed as pmol/10 min. Calibration curves for ACh were constructed by plotting the heights of peaks against the concentrations. Regression lines were then calculated and determination of unknown samples was carried out by the method of inverse prediction. The sensitivity limit was 500 fmol and the signal/noise ratio was higher than 3. Evoked release at the S₁ and S₂ periods of stimulation was calculated as the total minus the spontaneous release. Spontaneous release was obtained by averaging ACh content in the four samples immediately before the first K⁺ stimulation (S₁). Drug effects were evaluated by calculating the ratio of the evoked release (S₂/S1) for the two stimulation periods.

Statistical analysis. All values are expressed as means ± S.E.M., and the number of experiments (n) is also indicated. Comparisons between two means were carried out by Student's t test, paired or unpaired as appropriate. When experiments involved more than two treatment groups the presence of significant treatment effects was first determined by a one-way ANOVA and subsequent comparisons of pairs of means were made by Scheffe's test. For all statistical tests, P < .05 was considered significant.

Drugs. The substances used in this study included physostigmine hemisulfate, serotonin HCl, tetrodotoxin, (Sigma-Aldrich S.r.l., Milano, Italy); 8-OH-DPAT, tropisetron, 1-phenylbiguanide, m-chlorophenylbiguanide HCl (R.B.I., Natick, MA). All other reagents and solvents were of HPLC grade or the highest grade available (Sigma).

	Results

Top Abstract Introduction Methods Results Discussion References

Spontaneous and 100 mM K⁺-evoked release of ACh from cortex and hippocampus of freely moving rats. After 60 min of perfusion in the presence of 7 µM physostigmine, spontaneous ACh release was stable at, 3.2 ± 0.1 pmol/10 min (n = 172) in frontoparietal cortex and 2.7 ± 0.4 pmol/10 min (n = 12) in hippocampus.

Effects of 5-HT receptor agonists on the release of cortical Ach. Neither 50 µM 5-HT, nor 10 µM 1-PBG, a 5-HT₃ receptor selective agonist (Fozard, 1990; Ireland and Tyers, 1987; Wallis and Nash, 1981), in the perfusing Ringer solution altered spontaneous ACh release during a 30-min perfusion. ACh release averaged 3.6 ± 0.6 pmol/10 min before and 3.7 ± 0.6 pmol/10 min during 5-HT perfusion (n = 3), and 3.9 ± 0.4 pmol/10 min before and 4.1 ± 0.5 pmol/10 min during PBG perfusion (n = 3). However, 5-HT added to the perfusing medium 10 min before and during S₂ stimulation, significantly decreased the 100 mM K⁺-evoked ACh release from the cortex in a dose-dependent manner (fig. 1). This action was mimicked by PBG (fig. 2) and m-Cl-PBG (fig. 3), another 5-HT₃ selective agonist (Kilpatrick et al., 1990a), but not by the 5-HT_1A selective agonist 8-OH-DPAT (100 µM) which had no significant effect on 100 mM K⁺-evoked ACh release. The mean S₂/S₁ ratio in the presence of 8-OH-DPAT was 1.19 ± 0.18 (n = 5). 8-OH-DPAT (100 µM), added to the perfusing Ringer solution, also failed to alter spontaneous ACh release during a 30-min perfusion. ACh release averaged 3.4 ± 0.3 pmol/10 min before and 3.3 ± 0.5 pmol/10 min during 8-OH-DPAT perfusion (n = 3).

View larger version (37K): [in this window] [in a new window]   Fig. 1.   Inhibition by 5-HT of 100 mM K+-stimulated ACh release from the cortex of freely moving rats, and antagonism by systemic administration of tropisetron. ACh was measured in fractions collected every 10 min beginning 60 min after the onset of the perfusion. At 40 (S1) and 140 min (S2) the perfusion medium was changed from 4.0 mM KCl to 100 mM KCl for 10 min; isotonicity was maintained by reducing NaCl concentration. 5-HT (1-50 µM) was added to the perfusing medium 10 min before S2 and maintained during the S2 stimulation. Tropisetron, 0.5 mg/kg, was injected s.c. 40 min before the S2 stimulation. Shown are means ± S.E. of (n) of experiments. *P < .05, **P < .01 vs. control; P < .05 vs. 5-HT (10 µM) (ANOVA and Scheffe's test).

View larger version (36K): [in this window] [in a new window]   Fig. 2.   Inhibition by PBG of 100 mM K+-stimulated ACh release from the cortex of freely moving rats, and antagonism by systemic administration of tropisetron. Methods are identical to those described for figure 1. PBG (0.1-10 µM) was added to the perfusing medium 10 min before S2 and maintained during the S2 stimulation. tropisetron, 0.5 mg/kg, was injected s.c. 40 min before the S2 stimulation. Shown are means ± S.E. of (n) of experiments. *P < .05, **P < .01 vs. control; P < .01 vs. PBG (10 µM) (ANOVA and Scheffe's test).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Inhibition by m-Cl-PBG of 100 mM K+-stimulated ACh release from the cortex of freely moving rats. Methods are identical to those described for figure 1. m-Cl-PBG (1-10 µM) was added to the perfusing medium 10 min before S2 and maintained during the S2 stimulation. Shown are means ± S.E. of (n) of experiments. *P < .05 vs. control (ANOVA and Scheffe's test).

Effects of PBG on the release of hippocampal Ach. In contrast to its actions in frontoparietal cortex, PBG (10 µM), failed to modify the 100 mM K⁺-evoked ACh release in hippocampus. The mean S₂/S₁ ratio in the presence of PBG (1.35 ± .09, n = 7) was the same as that in controls.

Effects of systemic and local administration of a 5-HT₃ receptor antagonist, tropisetron, on the release of cortical Ach. Tropisetron (ICS 205-930), a 5-HT₃ receptor antagonist (Richardson et al., 1985), injected s.c. at a dose, 0.5 mg/kg, that blocks selectively 5-HT₃-mediated responses (Richardson and Buchheit, 1988), altered neither spontaneous ACh release (3.6 ± 0.5 pmol/10 min before; 3.4 ± 0.6 pmol/10 min 30 min after tropisetron n = 3), nor 100 mM K⁺-evoked ACh release (S₂/S₁ = 0.94 ± 0.03, n = 4), but it completely antagonized the inhibition of 100 mM K⁺-evoked release of ACh produced by 10 µM 5-HT (fig. 1) and by 10 µM PBG (fig. 2).

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Antagonism by local administration of tropisetron of 5-HT- and PBG-induced inhibition of 100 mM K+-stimulated ACh release from the cortex of freely moving rats. Methods are identical to those described for figure 1. 5-HT (10 µM) or PBG (10 µM) were added to the perfusing medium 10 min before S2 and maintained during the S2 stimulation. Tropisetron (10-100 nM) was added simultaneously with 5-HT or PBG, respectively. Shown are means ± S.E. of (n) of experiments. **P < .01 vs. 5-HT, 10 µM; P < .05 vs. PBG, 10 µM (ANOVA and Scheffe's test).

Effects of tetrodotoxin on the serotonergic modulation of the release of cortical ACh. TTX a voltage-dependent Na⁺-channel blocker, 0.5 µM, was infused directly into the frontoparietal cortex through the dialysis fiber 20 min before S₂ and maintained during S₂ stimulation. Spontaneous ACh release was significantly decreased by about 50% in the presence of TTX (P < .0001, paired Student's t test; n = 17). By pooling all 13 experiments with TTX, ACh spontaneous release averaged 3.27 ± 0.24 pmol/10 min, and ACh spontaneous release in the presence of TTX, calculated for each experiment by averaging the mean of the two 10-min samples of perfusate containing TTX and collected immediately before S₂, was 1.95 ± 0.19 pmol/10 min (fig. 5, A and B). Conversely, TTX had no effect on 100 mM K⁺-evoked release of ACh (S₂/S₁: 1.26 ± .26; n = 10) (fig. 5A). However, in the presence of TTX, 10 µM PBG, a concentration that can markedly reduce K⁺-evoked release of ACh (fig. 2), failed to inhibit the release of ACh (S₂/S₁: 1.07 ± .09; n = 7) (fig. 5B). Because TTX decreased spontaneous release, S₂ release was measured as the difference between the ACh content in the perfusion during S₂ and the mean of that in the two 10-min samples containing TTX and collected immediately before S₂.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Spontaneous and 100 mM K+-evoked release of ACh from the cortex of freely moving rats in the presence of .5 µM TTX alone (A) and in combination with 10 µM PBG (B). At 40 (S1) and 140 (S2) min the perfusion medium was changed from 4 to 100 mM KCl for 10 min. Isotonicity was maintained by reducing NaCl concentration. Bars show the duration of perfusion with 0.5 µM TTX (A and B) and 10 µM PBG (B). Shown are means ± S.E.M. of 10 (A) and 7 (B) experiments.

	Discussion

Top Abstract Introduction Methods Results Discussion References

Using microdialysis to simultaneously administer 5-HT and monitor changes in ACh release, we found that perfusion of frontoparietal cortex with 5-HT did not alter spontaneous release of ACh, but reduced up to 60% the release of ACh evoked by cortical perfusion with medium containing 100 mM K⁺. Although the concentration of K⁺ used in this study may appear high, the low recovery of K⁺ through the microdialysis membrane (Westerink and de Vries, 1996), and the rapid dilution of K⁺ in the extracellular space necessitate high concentrations in the perfusion fluid. Other investigators have shown that 60 mM K⁺ has only a slight effect on ACh output during brain dialysis (Westerink et al., 1987), and brain perfusion with 100 mM K⁺ evokes an increase in ACh release similar to that obtained with incubation of cortical slices in 20 mM K⁺ (Clapham and Kilpatrick, 1992).

The inhibition appears to be mediated by cortical 5-HT₃ receptors. It could be elicited by perfusion with PBG and m-Cl-PBG., compounds whose serotonergic agonist effects are essentially limited to activation of 5-HT₃ receptors (Fozard, 1990; Kilpatrick et al., 1990b; Tadipatri et al., 1992), and completely antagonized by systemic administration of tropisetron, at a dose (0.5 mg/kg, s.c.) considered selective for blockade of 5-HT₃ receptors (Richardson and Buchheit, 1988), and by inclusion of tropisetron in the perfusion fluid at submicromolar concentrations. The sensitivity of the agonist actions to the dose and concentration of tropisetron used argues strongly against the involvement of 5-HT₄ receptors (Bockaert et al., 1989; Donatsch et al., 1984; Dumuis et al., 1988) or catecholamine transporters (Benuck and Reith, 1992). The finding of 5-HT₃-mediated inhibition of K⁺-evoked ACh release in the cortex is consonant with the consistent demonstration of cortical 5-HT₃ binding (Barnes et al., 1988; Kilpatrick et al., 1987; Wong et al., 1989).

5-HT₃ receptors are ligand gated monovalent cation channels whose activation results in a net inward (depolarizing) current (Derkach et al., 1989). Actions in the central nervous system consonant with this mechanism have been reported (Sugita et al., 1992). Neuronal depolarization by activation of 5-HT₃ receptors may result in increased release of neurotransmitters, such as dopamine from slices of striatum (Blandina et al., 1988), cholecystokinin from synaptosomes of nucleus accumbens (Paudice and Raiteri, 1991), GABA from hippocampal slices (Ropert and Guy, 1991), and 5-HT from microdialysis of hippocampus (Martin et al., 1992). Numerous proposals could account for the apparent paradox that depolarization accompanying 5-HT₃ receptor activation inhibits K⁺-evoked transmitter release, such as that for hypothalamic norepinephrine (Blandina et al., 1991; Goldfarb et al., 1993) or cortical ACh (Barnes et al., 1989; Bianchi et al., 1990; and the present results). The inhibition may be indirect, and it could be most simply explained by postulating activation of an inhibitory interneuron. It is more difficult to rationalize the report of 5-HT₃-mediated inhibition of [³H]ACh from synaptosomes prepared from human cortex (Maura et al., 1992). However, the latter preparation minimizes but does not exclude indirect components, hence, the inhibition could be indirect, mediated by the release of another transmitter which, in turn, inhibits ACh release.

In our experiments, ACh release evoked by 100 mM K⁺ showed no significant TTX-sensitive component, thus excluding involvement of neuronal loops. In agreement with previous reports (Consolo et al., 1996), spontaneous ACh release was decreased by TTX, indicating the presence of impulse activity either in cholinergic afferents or in axons impinging on them. Perfusion with TTX abolished completely PBG-induced inhibition of K⁺-evoked ACh release, strongly suggesting that the 5-HT₃ receptors are located neither presynaptically on cholinergic nerve terminals, nor on terminals making axo-axonic contact with the cholinergic terminal. It appears most likely that the receptors are somadendritic receptors on interneurons and that excitation of these interneurons produces Na⁺-dependent action potentials that release an intermediary modulatory substance. Consistent with this hypothesis is the presence of 5-HT₃ receptor mRNA throughout cerebral cortex in the mouse (Tecott et al., 1993), and, in the rat, antibodies to the N terminal extracellular domain of the receptor labelled nonpyramidal cell bodies in layer II as well as in other layers of the cortex (Priestly et al., 1997). The simplest hypothesis, that these interneurons directly innervate the cholinergic presynaptic terminals and reduce ACh release is rendered unlikely by the observation that neither 5-HT nor either of the biguanides alters spontaneous ACh release, much of which is TTX sensitive. One synaptic arrangement consonant with the lack of 5-HT₃ modulation of spontaneous release is that the activated interneuron inhibits the release of an excitatory presynaptic modulator of cholinergic terminals. If this excitatory pathway were not spontaneously active, 5-HT₃ activation would have no effect on spontaneous ACh release. In the presence of K⁺, this excitatory modulator would be released and enhance the depolarization-induced release of ACh. Activation of 5-HT₃ receptors would remove this enhancement and partially, but not completely, depress K⁺-evoked ACh release. It is interesting in this regard that 5-HT₃ receptors facilitated the transmission of an inhibitory transmitter, GABA, in rat hippocampus (Ropert and Guy, 1991). In analogy with 5-HT₃ receptor/cortical ACh interactions, histamine, through H₃ receptor activation, inhibited significantly 100 mM K⁺-evoked release of ACh from the cortex of freely moving rats, but failed to alter spontaneous ACh release (Blandina et al., 1996). Histamine-induced inhibition is indirect, mediated by release of GABA from cortical interneurons (Giorgetti et al., 1997).

The effect of the use of cholinesterase inhibitors in these experiments on the characteristics of 5-HT₃ modulation of K⁺-evoked release is not known. The use of 100 nM neostigmine has been shown to alter the ability of dopaminergic agents to modulate resting ACh release in the striatum (De Boer and Abercrombie, 1996). The much lower levels of cortical as compared to striatal ACh release measured by microdialysis (Herrera-Marschitz et al., 1990) required concentrations of physostigmine of more than 1 µM in the perfusion fluid in order to reduce the variability of ACh release. The presence of physostigmine was not associated with any 5-HT₃ modulation of resting release and is unlikely to explain why we found a 5-HT₃ component to 5-HT modulation of cortical ACh release whereas the two previous studies with in vivo microdialysis did not. In both these studies anticholinesterases were used. Consolo et al. (1996) used the same concentration of physostigmine used in our study, and Hirano et al. (1995) used a concentration of neostigmine identical to that shown to alter striatal ACh release (De Boer and Abercrombie, 1996).

In contrast to the response of the K⁺-evoked ACh release from the cortex, that from the hippocampus was not regulated by PBG. Both, a facilitatory effect (Consolo et al., 1994) and the lack of effectiveness (Parkins et al., 1994) of 5-HT₃ receptors activation on in vivo hippocampal cholinergic activity has been reported. Nevertheless, we cannot exclude that, under our conditions, PBG was ineffective, for the release triggered by 100 mM K⁺ was already maximal.

The magnitude of the 5-HT₃ modulation of cortical ACh release may be underestimated in our experiments because of receptor desensitization. Very rapid desensitization of the 5-HT₃ receptor has often been seen: the fast inward current was transient despite continued application of 5-HT or 2-methyl-5-HT, a 5-HT₃ agonist (Richardson et al., 1985), in mouse N1E-115 neuroblastoma cells (Neijt et al., 1988), cultured mouse hippocampal cells (Yakel and Jackson, 1988) and in cloned human receptors expressed in Xenopus oocytes (Belelli et al., 1995). Similarly in peripheral neurons, rapid desensitization has been reported in rabbit preganglionic cervical sympathetic nerves (Elliott and Wallis, 1988) and in guinea pig celiac ganglion cells (Wallis and Dun, 1988). In our experiments, desensitization occurring during the first few minutes of superfusion with agonists would simply lower the magnitude of the response. It is possible, however, that the response we are studying may not desensitize, or may do so far more slowly than the 5-HT₃ receptor in other tissues. Variability of desensitization of the same receptor in different tissues is not uncommon. The depolarization in dorsal root ganglion cells produced by 5-HT₃ receptor activation did not desensitize (Todorovic and Anderson, 1990). Marked differences in desensitization rates in different tissues have been reported for other receptors. In crayfish, there was marked desensitization to GABA in the closer muscles but not the opener muscles of the walking leg (Dudel and Hatt, 1976). The half-time for desensitization of the 5-HT₂ response was 70 min in rabbit aorta and 5 min in the guinea pig trachea (Ben-Harari et al., 1987). The histamine H₂ receptors in guinea pig brain (Green et al., 1977) and heart (Johnson et al., 1979) membranes did not desensitize, whereas in HL-60 cells exhibited marked desensitization with an 80 to 90% loss of response (Johnson and Sawutz, 1985).

Accumulating evidence suggests that 5-HT may modulate cholinergic functions and that these interactions influence cognition (see review by Cassel and Jeltsch, 1995). Dysfunction of the serotonergic system has been found in aging (Schlicker et al., 1989; Wenk et al., 1989) and in Alzheimer's disease (Bowen et al., 1983). If cognitive deficits are related to reduced availability of ACh in the synaptic cleft (Blandina et al., 1996; Quirion et al., 1995) and 5-HT₃ receptor activation reduces ACh release, 5-HT₃ antagonists might facilitate ACh release and thereby improve cognitive function. Reports that 5-HT₃ receptor antagonists improve learning and memory (see review by Passani and Corradetti, 1996) support this contention.