INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The 5-HT₃ receptor is unique among the serotonin receptors because it is the only ligand-gated ion channel in this family of neurotransmitter receptors. Since the first identification of 5-HT₃ receptors in cultured mouse neuroblastoma cells (Neijt et al. 1986) and the cloning of the first of two subunits (Maricq et al. 1991), many efforts have been made to characterize the receptor and channel properties in cultured cell lines and heterologous expression systems (for review see Jackson and Yakel 1995). However, information on the functional properties of 5-HT₃ receptors native to the CNS is still limited. The 5-HT₃ receptor is expressed in a large number of brain regions, including hippocampus, cortex, amygdala, striatum, and several brain stem nuclei (Barnes and Sharp 1999; Jackson and Yakel 1995). In many of these regions, the presence of 5-HT₃ receptors in presynaptic nerve endings has been well established (Nayak et al. 1999; Ronde and Nichols 1998). However, the functional role of presynaptic 5-HT₃ receptors in mediating or modulating neurotransmitter release remains elusive (van Hooft and Vijverberg 2000).

5-HT₃ receptor-mediated synaptic transmission has been observed in rat amygdala (Sugita et al. 1992), ferret visual cortex (Roerig et al. 1997), and rat sensorimotor cortex (Ferezou et al. 2002), suggesting a role for postsynaptic 5-HT₃ receptors in mediating fast serotonergic transmission. In cortex and hippocampus, the 5-HT₃ receptor is expressed in 35-66% of the population of cholecystokinin (CCK)-containing interneurons (Morales and Bloom 1997). One of the prominent functional features of the 5-HT₃ receptor-induced ion current in hippocampal interneurons is the region of negative slope conductance in the I-V curve (Kawa 1994; McMahon and Kauer 1997). This negative slope conductance is due to voltage-dependent block by Ca²⁺ ions, analogous to the voltage-dependent block by Mg²⁺ ions of the N-methyl-D-aspartate (NMDA) receptor (Nowak et al. 1984).

Interestingly, the voltage-dependent block by Ca²⁺ ions has not been observed with 5-HT₃ receptors in clonal cell lines or heterologously expressed 5-HT₃ receptors, except for a report on the expression of homomeric 5-HT₃ receptor in Xenopus oocytes (Maricq et al. 1991). It has been shown before that both 5-HT₃ receptors in N1E-115 neuroblastoma cells and cloned 5-HT₃ receptors can be inhibited by physiological concentrations of Ca²⁺ (Gill et al. 1995; Peters et al. 1988). However, this block is not voltage dependent and presumably involves an interaction of Ca²⁺ with the agonist recognition site (Niemeyer and Lummis 2001).

In this study, the voltage-dependent block of 5-HT₃ receptors by Ca²⁺ ions in hippocampal interneurons was experimentally determined and analyzed using a classical three barrier-two site (3B2S) model based on Eyring rate theory. The results suggest that the 5-HT₃ receptor ion channel contains two binding sites for Ca²⁺ ions, and that Ca²⁺ is slightly permeable at hyperpolarized membrane potentials.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Electrophysiology

Male Wistar rats, 14-16 days old (P14-P16), were decapitated, and the brain was quickly removed. Experiments were conducted according to the ethics committee guidelines of the University of Amsterdam. Parasagittal slices (250 µm) of the hippocampus were cut on a vibroslicer (752M, Campden Instruments, Loughborough, UK). Slices were allowed to recover for 30 min at 31°C in artificial cerebrospinal fluid (ACSF) containing (in mM) 120 NaCl, 3.5 KCl, 2.5 CaCl₂, 1.3 MgSO₄, 1.25 NaH₂PO₄, 25 NaHCO₃, and 25 glucose, continuously bubbled with 95% O₂-5% CO₂ (pH = 7.4). Interneurons in stratum radiatum of the hippocampal CA1 area were visualized using infrared differential interference contrast videomicroscopy on a Zeiss FS2 microscope with a VX44 CCD camera (PCO, Kelheim, Germany). Patch pipettes were pulled from boroscilicate glass and had a resistance of 2-4 M when filled with internal solution containing (in mM) 140 CsCl, 0.5 CaCl₂, 5 EGTA, 10 HEPES, and 2 Mg-ATP (pH = 7.3 with CsOH). For some experiments, KCl in extracellular solution and CsCl in intracellular solution were replaced with equimolar NaCl. Whole cell recordings were made using an EPC9 patch-clamp amplifier and PULSE software (HEKA Electronic, Lambrecht, Germany). Signals were filtered at 1-5 kHz and sampled at 2-10 kHz. Series resistance ranged from 5-20 M and was compensated for 80%. During recording, slices were kept submerged and were continuously superfused with ACSF, containing 0.5 µM TTX, at room temperature (20-22°C). Cells were voltage clamped at 60 mV (corrected for liquid junction potential), unless noted otherwise. A second pipette, connected to a picospritzer (General Valve, Fairfield, NJ) and containing 100 µM 5-HT in ACSF, was positioned in the vicinity of the cell soma. 5-HT was applied for 500 ms at 35-100 kPa. The 5-HT solution also contained Fast Green (approximately 1 mg/ml) to visually inspect the area of application. Previous experiments have shown that Fast Green inhibits miniature synaptic events (van Hooft 2002). Control experiments showed that Fast Green does not affect the amplitude or kinetics of 5-HT-induced ion currents (data not shown). 5-HT was applied at intervals of 3 min to allow for complete recovery from desensitization. Drugs and ACSF containing different concentrations of CaCl₂ were applied by bath perfusion.

Data analysis

Current-voltage relations of the 5-HT-induced ion current were recorded by a voltage ramp protocol. The cell was held at a holding potential of +20 mV for 10 s, allowing the inactivation of voltage-gated ion channels. Subsequently, a voltage ramp of 500 ms or 1 s from +20 to 140 mV was applied. This protocol was repeated in the presence of 5-HT, at the time of the ion current peak where little or no desensitization occurred. Current traces recorded in the absence of 5-HT were subtracted from those recorded in the presence of 5-HT. The subtracted current was normalized to the amplitude of the current at 60 mV.

Dose-response curves of Ca²⁺ block at a given membrane potential were fitted using a logistic equation

(1)

where I_max is the maximal normalized current at a given membrane potential, I_min is the fractional residual current in the presence of a saturating [Ca²⁺], K_d is the concentration of Ca²⁺ producing half-maximum block, and n is the Hill coefficient.

A 3B2S model based on Eyring rate theory (Hille 1992) was used to describe the I-V curves to investigate the relative permeation of Ca²⁺ ions and monovalent cations. The model assumes that the channel contains two binding sites for ions, reflected by local energy minima flanked by symmetrical barriers, and that the sites may be occupied simultaneously. Given two binding sites and two permeable ion species, nine different occupation states of the channel are possible (Begenisich and Cahalan 1980; Hille 1992). Transitions of the ions between different states of occupation are expressed as rate constants k, which are voltage-dependent

(2)

where k is Boltzmann's constant, h is Planck's constant, G is the free Gibbs energy of the transition to or from the binding site,  is the fraction of the electrical field experienced by the ion during the transition, z is the valence, and F, R, and T have their usual thermodynamic meaning. The term kT/h represents the intrinsic frequency of thermal vibration (6.15 × 10¹² s¹ at 22°C), indicating the maximum value of the rate constants. The rate constants for the transition of ions from the intra- or extracellular space to the binding site are of the same form, but multiplied by the intra- or extracellular ion concentration, respectively.

For each permeable ion, six rate constants can be derived in this way, giving rise to eight unknown parameters: the height of the three barriers, the depth of the two binding sites, and the three electrical distances defining the relative location of the binding sites along the electrical field across the channel. The energies of the barriers and sites for the different permeable ions were allowed to differ, but the electrical distances were assumed to be identical. Given the nine possible modes of occupation of the channel by two ions, we can calculate the steady-state probability that the channel is in a particular state by the matrix method as described in detail by Begenisich and Cahalan (1980). Given the rate constants of transition and the steady-state probabilities of channel occupation, the steady-state flux of each ion can subsequently be calculated as the net rate of the ion crossing any one of the barriers. The model was fitted to the I-V curves assuming that the ion current is due to the total charge transfer carried by the permeable ions. The fractional Ca²⁺ current was calculated as

(3)

and an estimate of the permeability ratio P_Ca/P_Na was subsequently obtained according to Schneggenburger et al. (1993) and Schneggenburger (1996)

(4)

For all calculations, concentrations were converted to ion activities, assuming activity coefficients of 0.5 for Ca²⁺ and 0.7 for Na⁺ (Schneggenburger 1996). In the text and figures, concentrations are mentioned rather than activities. The intracellular [Ca²⁺] was estimated to be 10 nM. Off-line analysis was performed using IGOR Pro (Wavemetrics, Lake Oswego, OR), and fitting of the I-V curves was done using fit routines written in Modula. All results are expressed as mean ± SE of n independent experiments. Comparisons were made using Student's t-test unless indicated otherwise. P < 0.05 was used to indicate a significant difference.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Extracellular Ca²⁺ blocks the 5-HT₃ receptor-mediated ion current in s. radiatum interneurons

Application of 100 µM 5-HT to whole cell voltage-clamped s. radiatum interneurons resulted in a transient inward current of 154 ± 63 pA (n = 34) in approximately 75% of the interneurons tested. The inward current was completely blocked by 100 nM of the selective 5-HT₃ receptor antagonist MDL72222 (Fig. 1A). The I-V curve of the 5-HT₃ receptor-mediated ion current shows a region of negative slope conductance from -100 to -60 mV (Fig. 1B). The shape of the I-V curve was similar when determined by evoking ion currents at different holding potentials (points in Fig. 1B). In the presence of 100 nM MDL72222, no ion current could be detected over a voltage range of 140 to +20 mV (Fig. 1B). Application of the 5-HT₁/5-HT₂ receptor antagonist methysergide (10 µM) did not affect the 5-HT-induced ion current (data not shown).

View larger version (13K): [in this window] [in a new window]   Fig. 1. I-V curve of the 5-HT3 receptor-mediated ion current in hippocampal stratum radiatum interneurons. A: ion current evoked by application of 100 µM 5-HT for 500 ms (asterisks) in a whole cell voltage-clamped s. radiatum interneuron (Vh = 60 mV). After bath application of 100 nM of the selective 5-HT3 receptor antagonist MDL72222, the current is completely blocked (arrow). B: I-V curve of the 5-HT-induced ion current. Points represent normalized current amplitudes recorded at different holding potentials. The trace represents the I-V curve as determined by a voltage ramp. In the presence of 100 nM MDL72222, the current is blocked over the whole voltage range (arrow). All ion current amplitudes are normalized to that recorded at -60 mV. Points and I-V trace are the mean of 6 cells.

In concordance with previous reports on 5-HT₃ receptor-mediated currents in hippocampal dentate gyrus interneurons (Kawa 1994) and CA1 s. radiatum interneurons (McMahon and Kauer 1997), Fig. 2A shows that the region of negative slope conductance of the I-V curve is dependent on extracellular [Ca²⁺]. Lowering the extracellular [Ca²⁺] shifts the region of negative slope conductance toward more negative potentials. At 100 µM Ca²⁺, the negative slope conductance region is absent down to a membrane potential of -120 mV (Fig. 2A). In contrast, Mg²⁺ does not affect the I-V curve of the 5-HT₃ receptor-mediated ion current. The I-V curves recorded in 1.3 mM Mg²⁺ and 100 µM Mg²⁺ (both in the presence of 2.5 mM Ca²⁺) are indistinguishable (Fig. 2B).

View larger version (26K): [in this window] [in a new window]   Fig. 2. Voltage-dependent Ca2+ block of the 5-HT3 receptor-mediated ion current. A: I-V curves of the 5-HT3 receptor-mediated ion current in the presence of the extracellular [Ca2+] indicated (not corrected for ion activities). Traces were recorded from the same cell. All current amplitudes are normalized to that recorded at -60 mV. B: superimposed I-V curves of the 5-HT3 receptor-mediated ion current recorded in the presence of 1.3 mM Mg2+ or 100 µM Mg2+, both in the presence of 2.5 mM Ca2+. Traces were recorded from the same cell. Similar results were obtained in 3 other cells.

A classical approach to analyze voltage-dependent block of ion channels derives from the description of block of voltage-dependent sodium channels by H⁺ (Hille 1992; Woodhull 1973). In this model, it is assumed that the blocking ion reaches a single binding site inside the ion channel, but cannot permeate the ion channel, resulting in a complete block of the ion current at saturating concentration of blocker and large negative membrane potentials (Woodhull 1973). The effect of Ca²⁺ was quantified by determining the block of the ion current at a given membrane potential as fraction of the maximum ion current at that membrane potential (which was taken from the ion current recorded at 100 µM Ca²⁺ up to a membrane potential of -120 mV, see Fig. 2A). The resulting dose-response curves (Fig. 3A) were fitted with a logistic equation (Eq. 1). The value of I_min (the fractional residual current in the presence of a saturating [Ca²⁺]) amounted to 0.11 ± 0.02 (n = 6), consistent with the residual current at hyperpolarized membrane potentials observed in the I-V curves (Fig. 2A). This indicates that the block by Ca²⁺ ions can be surmounted by voltage. Figure 3B shows that the Hill coefficients obtained from the dose-response curves (Fig. 3A) are larger than 1, suggesting that there is more than one binding site for Ca²⁺. Taken together, the data indicate that a Woodhull model of a non permeant blocker at a single site is not sufficient to adequately describe the voltage-dependent Ca²⁺ block of 5-HT₃ receptor channels.

View larger version (11K): [in this window] [in a new window]   Fig. 3. Analysis of voltage-dependent Ca2+ block. A: dose-response relations of Ca2+ block at different membrane potentials. The amount of block at a given membrane potential was determined as fraction of the estimated ion current in the absence of block. For each membrane potential, the curve was fitted with a logistic dose-response equation. For clarity, only 1 dose-response curve is shown, which corresponds to a membrane potential of -100 mV (). B: plot of the Hill coefficients, as obtained from the dose-response relations in A, as a function of membrane potential. Data represent mean ± SE from 6 cells.

Analysis of the voltage-dependent Ca²⁺ block according to a 3B2S model

A 3B2S model, based on Eyring rate theory, was subsequently used to describe the mechanism of Ca²⁺ block. This model is based on the assumption that two binding sites exist in the ion channel, and that Ca²⁺ and other permeant ions (Na⁺, K⁺, and under these experimental conditions, also Cs⁺) compete for these binding sites. The sites were assumed to be on the same electrical location for all ions. In addition, it was assumed that the relative permeabilities of the monovalent cations were identical (Davies et al. 1999; Lambert et al. 1989; Yang 1990). Figure 4A shows the fit of the I-V curve using the 3B2S model with the assumptions as outlined above. The estimates of the parameters were robust between cells (data not shown) and insensitive to the initial values of the parameters. Nevertheless, there are major discrepancies between the I-V curve and the fit. The boundary conditions in the model impose a reversal potential of 0 mV, as expected with equimolar monovalent cation concentrations on the inside and outside of the channel, and the assumption of equal permeability of monovalent cations. Experimentally, the reversal potential of the 5-HT₃ receptor-mediated ion current in rat hippocampal interneurons was found to be 11.4 ± 4.8 mV (n = 6). Lowering the barrier heights for Ca²⁺ in the model to increase Ca²⁺ permeability (which would result in a concomitant shift of the reversal potential to a more positive value) did not improve the fit of the I-V curve (data not shown). However, when I-V curves were recorded under ionic conditions where extra- and intracellular K⁺ and Cs⁺ were all replaced by Na⁺, the reversal potential was shifted to 0.5 ± 2.3 mV (n = 4), and the fit of the I-V curve with the 3B2S model under these ionic conditions improved considerably (Fig. 4B). Figure 4C shows the location of the barriers and binding sites for Ca²⁺ and Na⁺, expressed as fraction of the electrical field across the ion channel. The binding sites are located at 0.18 ± 0.05 and 0.94 ± 0.01 (n = 4) as fraction of the electrical field across the channel, with the barriers located symmetrically in between. Reeves et al. (2001) have reported the position of the channel-lining amino acid residues expressed as fraction of the electrical field. Based on these results, the physical position of the binding sites were inferred. Figure 4D shows that the first binding site is located approximately at position 13' (residue V291), whereas the second binding site is located at the far cytoplasmic side of the ion channel (more cytoplasmic than residue E277).

View larger version (30K): [in this window] [in a new window]   Fig. 4. Location of the Ca2+ binding sites in the 5-HT3 receptor channel. A: fit of an I-V curve with the three barrier-two site (3B2S) model under standard ionic conditions and assuming equal permeability of monovalent cations. B: fit of an I-V curve with the 3B2S model under modified ionic conditions (extra- and intracellular K+ and Cs+ substituted by Na+). C: distribution of the barriers and binding sites (G in units of RT) expressed as fraction of the electrical field across the ion channel (). Data represent mean ± SE from 4 cells. D: putative physical location of the binding sites of Ca2+ in the ion channel. The amino acid residues facing the lumen of the ion channel and the short-hand notation of their positions are after Reeves et al. (2001).

Predictions on the Ca²⁺ permeability of the 5-HT₃ receptor

Because the 3B2S model assumes two separate ion fluxes (Ca²⁺ and Na⁺), the charge, current and consequently the fraction of the total ion current carried by Ca²⁺ can be calculated. Figure 5A shows the Ca²⁺ current as function of voltage. At membrane potentials more positive than -120 mV, the conductance of the Ca²⁺ current is voltage dependent. At hyperpolarized membrane potentials more negative than -120 mV, the fractional Ca²⁺ current approaches 1%. However, at more physiological membrane potentials, the fractional Ca²⁺ current drops readily below 0.1% (Fig. 5B). Using the relationship between the fractional Ca²⁺ current and the permeability ratio P_Ca/P_Na (Schneggenburger 1996; Schneggenburger et al. 1993) (see Eq. 4), the permeability ratio P_Ca/P_Na was estimated to be 0.10 at -60 mV (Fig. 5B, inset).

View larger version (15K): [in this window] [in a new window]   Fig. 5. Predications of the 3B2S model concerning Ca2+ permeability. A: I-V curve of the Ca2+ current component, as predicted by the 3B2S model. The amplitude is normalized to that of the Na+ current component at -60 mV. B: I-V curve of the fractional Ca2+ current, expressed as percentage of the total current. Inset: PCa/PNa for the physiological relevant membrane potential range.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Our data confirm previously reported experiments on the voltage-dependent nature of Ca²⁺ block of 5-HT₃ receptor channels in native CNS neurons (Kawa 1994; McMahon and Kauer 1997). We extend these observations with in situ patch-clamp recordings in brain slices and with a quantitative prediction of the location of the binding sites and the relative Ca²⁺ permeability of the channel. For the latter two aspects, we employ Eyring rate theory and found that the minimum model that gives a sufficient quantitative description has 3B2S for calcium and for a monovalent cation. The rationale for using the 3B2S model is based on the observation that at large negative membrane potentials the Ca²⁺ block is surmounted by voltage (Fig. 2A), and that the Ca²⁺ block shows cooperativity (Fig. 3B). In a first attempt, we fitted the I-V curves with a Woodhull model, with either a finite or an infinite inner barrier, but this could not adequately describe the Ca²⁺ block (data not shown). One remaining discrepancy between our experimental data and the 3B2S model was caused by the fact that we lumped all experimental cations (Na⁺, K⁺, and Cs⁺) into one monovalent model ion. The clearest indication for this was a deviation between calculated and observed reversal potential. Experimentally, we confirmed this hypothesis by using only Na⁺ ions to carry the current, but this condition is to far outside the physiological range for standard use in slice experiments. Alternatively, a more elaborate model could include more monovalent permeable ions (K⁺ and/or Cs⁺), with associated barriers and binding sites. However, we think that such a refinement, which takes into account the relative permeability of K⁺ (and/or Cs⁺), but also adds many new parameters, will hardly enhance the explanatory power of the model as far as voltage-dependent Ca²⁺ block is concerned. The relative Ca²⁺ permeability, as calculated with Eq. 4, assumes independent monovalent/divalent ion permeation, which may not be the case for Na⁺ (K⁺) and Ca²⁺ ions through 5-HT₃ receptor channels. The exact determination of the relative permeabilities of Ca²⁺ and K⁺ compared with Na⁺ is beyond the experimental limits of the slice preparation used in this study. However, the 3B2S model provides a sufficient framework to describe the actions of Ca²⁺ in the 5-HT₃ receptor channel native to hippocampal interneurons.

Site of action of Ca²⁺

From the positions of the binding sites in the electrical field across the channel, and the published data on the location of channel-lining amino acid residues (Reeves et al. 2001), the location of the two binding sites were inferred to be approximately at the 13' and the -4' position (Fig. 4D). This is in excellent agreement with the data obtained from the nicotinic acetylcholine receptor (nAChR) 7 subunit, which is both functionally and structurally closely related to the 5-HT_3A receptor subunit. The amino acid residues lining the channel pore from the 13' position toward the cytoplasmic side are identical between nAChR 7 and 5-HT_3A subunit (Corringer et al. 2000). Site-directed mutagenesis studies of the nAChR 7 subunit revealed that mutation of residues L254 and L255 (approximately at the 16' position) and E237 (approximately at the -1' position) abolishes Ca²⁺ permeability (Bertrand et al. 1993; Corringer et al. 2000). This suggests that the basic properties of binding of Ca²⁺ ions in the channel is a conserved feature among these channels. The -1' position of the nAChR, located at the narrow side of the channel pore, has been implicated to determine ion selectivity (Corringer et al. 2000). Despite the fact that the nAChR 7 and the 5-HT_3A share the same amino acid sequence in this region, the nAChR 7 is highly permeable to Ca²⁺ (Bertrand et al. 1993), whereas the 5-HT₃ receptor in hippocampal interneurons is much less Ca²⁺ permeable (Fig. 5B). In addition, the nAChR 7 does not exhibit voltage-dependent block by Ca²⁺. Therefore it seems likely that, apart from the structural elements on the 5-HT_3A subunit, additional factors are involved in the voltage-dependent Ca²⁺ block of 5-HT₃ receptors in hippocampal interneurons.

Molecular determinant of Ca²⁺ block

Recombinant 5-HT₃ receptors expressed inXenopus oocytes display voltage-dependent Ca²⁺ block (Maricq et al. 1991), analogous to the voltage-dependent Ca²⁺ block observed with 5-HT₃ receptors native to hippocampal dentate gyrus interneurons (Kawa 1994) and CA1 s. radiatum interneurons (McMahon and Kauer 1997). However, this observation has not been corroborated in subsequent studies on the expression of homomeric 5-HT₃ channels in heterologous expression systems. The functional properties, especially Ca²⁺ permeability, of recombinant 5-HT₃ receptors, can be altered by co-expression with additional subunits such as the 4 nAChR subunit and the 5-HT_3B subunit (Davies et al. 1999; van Hooft et al. 1998). However, it has recently been reported that the 5-HT_3B subunit is not present in hippocampal interneurons (Ferezou et al. 2002; Morales and Wang 2002; Sudweeks et al. 2002). Besides, none of the heteromeric 5-HT₃ receptors examined in heterologous expression systems show voltage-dependent Ca²⁺ block. The most parsimonious explanation for this apparent discrepancy would be to propose the existence of yet another additional subunit conferring this property to 5-HT₃ receptors in hippocampal interneurons. As an alternative, posttranslational modification of specific sites involved in Ca²⁺ binding or interaction with additional intracellular factors may play a role. Concerning the latter option, it is of interest to note that intracellular polyamines have been shown to influence the Ca²⁺ permeability of nAChR (Haghighi and Cooper 2000). It remains to be determined whether a complex interaction between Ca²⁺, intracellular polyamines, and specific residues of the channel can account for the voltage-dependent Ca²⁺ block.

Little information is available on the functional properties of 5-HT₃ receptors expressed in other brain regions than hippocampus. It has been shown that the I-V curve of synaptically evoked 5-HT₃ receptor-mediated ion currents in layer 5 pyramidal neurons of ferret visual cortex also display a region of negative slope conductance (Roerig et al. 1997). However, the maximum ion current is around -20 mV compared with around -60 mV in hippocampal interneurons. In another study, the I-V curve of 5-HT₃ receptor-mediated ion currents in cortical layer 1 interneurons is linear; however, the I-V curve was recorded up to a membrane potential of only -65 mV (Zhou and Hablitz 1999). It would be of interest to know whether these apparent functional differences compared with hippocampal 5-HT₃ receptors truly reflect functional diversity of native 5-HT₃ receptors.

Physiological considerations

Voltage-dependent Ca²⁺ block appears to be a general property of hippocampal 5-HT₃ receptors (Kawa 1994; McMahon and Kauer 1997), and it is tempting to suggest that this phenomenon may represent a coincidence-detector function underlying a long-term potentiation (LTP)-like mechanism, analogous to the voltage-dependent Mg²⁺ block of NMDA receptors. There are a few reports suggesting that 5-HT₃ receptors are involved in LTP. It has been reported that systemic injections of the selective 5-HT₃ receptor antagonist ondansetron facilitate induction of LTP, increase the frequency of the theta electroencephalogram rhythm, and enhance retention of memory in hippocampus-dependent tasks (Staubli and Xu 1995). In addition, blockade of 5-HT₃ receptors in vivo results in a reduction of firing activity of hippocampal interneurons and an increase of firing of hippocampal pyramidal neurons (Reznic and Staubli 1997). It should be noted that according to these observations, it can be hypothesized that 5-HT₃ receptor activation would lead to a reduction of LTP in hippocampal pyramidal cells because of enhanced activity of interneurons excited by 5-HT₃ receptor activation. It remains to be determined whether a 5-HT₃ receptor-mediated LTP-like mechanism can be resolved in hippocampal interneurons.