INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Serotonin (5-hydroxytryptamine, 5-HT) is an important modulatory neurotransmitter in the gastropod mollusk Aplysia californica, playing an important role in such behaviors as feeding (Rosen et al. 1989), locomotion (Mackey and Carew 1983), and the defensive withdrawal reflexes (Glanzman et al. 1989). 5-HT is released by behavioral stimuli that initiate nonassociative and associative learning (Mackey et al. 1989; Marinesco and Carew 2001) and triggers the alteration of properties of mechanosensory neurons (SNs) that provide afferent input to the circuits for the defensive gill, siphon, and tail withdrawal reflexes (Byrne and Kandel 1996). 5-HT stimulation of adenylyl cyclase (AC) acts via cAMP-dependent protein kinase to increase SN excitability and spike duration (Baxter and Byrne 1990; Goldsmith and Abrams 1992; Hochner and Kandel 1992; Klein et al. 1986) and to produce short-, intermediate-, and long-term facilitation of the synaptic connections between these SNs and postsynaptic neurons (Ghirardi et al. 1992, 1995; Schacher et al. 1988; Scholz and Byrne 1988). 5-HT activation of protein kinase C (PKC) and mitogen-activated protein kinase also contributes to facilitation of SN synapses (Braha et al. 1993; Byrne and Kandel 1996; Manseau et al. 2001; Martin et al. 1997; Sacktor et al. 1988; Sugita et al. 1994).

In studying the neuromodulatory roles of multiple 5-HT-activated second-messenger cascades, it would be advantageous to have selective pharmacological antagonists for the 5-HT receptors that activate AC in Aplysia CNS. In contrast, the 5-HT receptor antagonist cyproheptadine is widely used in Aplysia but affects multiple 5-HT receptor subtypes (Goldsmith and Abrams 1992; Sossin et al. 1994). To date, five G-protein-coupled 5-HT receptors have been cloned from Aplysia. Two of these, Ap5-HT_B1 and Ap5-HT_B2, are coupled to phospholipase C (PLC) (Li et al. 1995). Two other 5-HT receptors, 5-HT_ap1 and 5-HT_ap2, inhibit AC (Angers et al. 1998; Barbas et al. 2002). A fifth Aplysia 5-HT receptor, for which a partial cDNA clone has been obtained, is strongly expressed in the gill and weakly expressed in the CNS (Williams et al. 1997). 5-HT receptors that activate AC have not yet been cloned in Aplysia or any other gastropod mollusc (see Tierney 2001). We therefore pharmacologically characterized, in biochemical assays, the 5-HT receptor(s) that activate AC in Aplysia CNS. To assess the efficacy of these antagonists under physiological conditions, we also examined the 5-HT-induced cAMP-dependent modulation of the electrophysiological properties of SNs.

Recently, selective, high-affinity ligands have been developed for most known subtypes of mammalian 5-HT receptors (Bonhaus et al. 1997; Roth et al. 1994; Sleight et al. 1998; Wardle et al. 1994). Six subtypes of G-protein-coupled 5-HT receptors have been characterized in mammals (Hoyer and Martin 1997; Hoyer et al. 1994). Five of these receptor subtypes are coupled to AC: the 5-HT₄, 5-HT₆, and 5-HT₇ receptors, which activate AC, and the 5-HT₁ and 5-HT₅ receptors, which inhibit AC. The 5-HT₂ receptor subtypes activate PLC. A seventh receptor subtype, 5-HT₃, forms a nonselective cationic channel. Using 5-HT stimulation of AC in Aplysia CNS membranes as an assay, we tested nonselective high-affinity antagonists as well as several antagonists that are highly selective for specific mammalian 5-HT receptor subtypes. The pharmacology of the 5-HT receptor in Aplysia CNS that activates AC resembled most closely the pharmacology of the 5-HT₆ receptor subtype. Of the 14 compounds tested, methiothepin, a dibenzapine, was the most effective in inhibiting 5-HT stimulation of AC. Unfortunately, methiothepin inhibits multiple 5-HT receptors. However, in conjunction with spiperone, an antagonist selective for the PLC-coupled 5-HT receptors, methiothepin should be useful in studying cAMP-mediated, 5-HT-dependent neuromodulation in Aplysia.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

A. californica, (Aplysia Resource Facility, Miami, FL, and Alacrity, Redondo Beach, CA), weighing 100-250 g, were anesthetized by injection of isotonic MgCl₂, and their abdominal, cerebral, and pleural-pedal ganglia were removed. In biochemical experiments on whole CNS, the abdominal, cerebral, and pleural-pedal ganglia were trimmed to reduce nonneural tissue (i.e., sheath). In biochemical and electrophysiological experiments on SNs, pleural ganglia were desheathed to expose the SNs in the ventrocaudal (VC) cluster. In electrophysiological experiments, prior to desheathing, ganglia were treated with 0.5% glutaraldehyde for 50 s to prevent contraction of muscle cells in the remaining sheath.

Drugs

The following drugs were used: 5-carboxyamidotryptamine maleate (5-CT, Tocris Cookson), dopamine (DA) hydrochloride, 5-HT creatine sulfate, nifedipine (Sigma, St. Louis, MO), tetraethylammonium (TEA) chloride (JT Baker, Phillipsburg, NJ), clozapine, cyproheptadine hydrochloride, fluphenazine dihydrochloride, NAN-190 hydrobromide, metergoline, methiothepin mesylate, risperidone, ritanserin, Ro-04-6790, RS-102221 hydrochloride, spiperone hydrochloride (RBI, MA), GR-113808 (a gift of Glaxo-Wellcome), SB-204070 (a gift of SmithKline Beecham), and olanzapine (a gift of Eli Lilly). Stock solutions of 10 mM 5-HT were made in 0.1 M acetic acid or, when the final 5-HT concentration exceeded 100 µM, in deionized water. 5-CT, cyproheptadine, DA, fluphenazine, GR-113808, and methiothepin were made as 10 mM stock solutions in deionized water. The solubility of methiothepin is pH sensitive, decreasing in more alkaline solutions (Liao et al. 1999; Nelson et al. 1979). We found the limits of methiothepin solubility in physiological saline were 175 µM at pH 7.3, 100 µM at pH 7.6, and 32 µM at pH 8.3 (for the mesylate salt, which is more soluble than the maleate salt). Spiperone was dissolved at 400 µM in 37°C deionized water. Stocks of clozapine (80 mM), metergoline (40 mM), NAN-190 (5 mM), ritanserin (40 mM), and SB-204070 (40 mM) were dissolved in dimethylsulfoxide (DMSO). Stocks of risperidone (40 and 80 mM) and RS-102221 (40 mM) were dissolved in ethanol. Final concentrations of DMSO and ethanol were no more than 0.025%, except with NAN-190, in which case DMSO was 0.2%. For risperidone and RS-102221, control assays (without antagonist) contained 0.025% ethanol. For NAN-190, control assays contained 0.2% DMSO. Neither 0.2% DMSO nor 0.025% ethanol had any effect on AC activity or 5-HT stimulation. Initial assays with SB-204070, which binds to some plastics (Wardle et al. 1994), were performed in glass tubes; no differences were observed compared with assays in the polypropylene tubes used in most experiments.

Preparation of tissue for AC assays

For whole CNS membranes, trimmed ganglia were homogenized in a glass-glass homogenizer in homogenization buffer: 50 mM K-HEPES (pH 7.6), 75 mM KCl, 3 mM EGTA, 1 mM dithiothreitol (DTT), and protease inhibitors (10 µg/ml aprotinin, 1 mM benzamidine, 10 µg/ml leupeptin, and 25 µM p-nitrophenyl-p'-guanidinobenzoate hydrochloride; 0.8 ml buffer/CNS). Any residual sheath was removed, and the material was rehomogenized in a glass-Teflon homogenizer. For SN membranes, the VC cluster was dissected from 18 desheathed pleural ganglia and homogenized in a glass-Teflon homogenizer in 1.5 ml homogenization buffer. All processing of membranes was at 0°C; centrifugations were at 4°C. The CNS or SN homogenate was centrifuged at 1,000 g for 2 min to remove any crude particulate material. The supernatant was then centrifuged at 16,000 g for 20 min. The pellet was resuspended in homogenization buffer and recentrifuged. The final pellet was homogenized in resuspension buffer [50 mM K-HEPES, pH 7.6, 75 mM KCl, 1 mM DTT and protease inhibitors (described in the preceding text)] (500 µl/CNS and 380 µl/18 SN clusters) and then assayed immediately.

AC assays

Assays of AC activity in membrane preparations were carried out in 80 µl for most experiments on CNS and in 60 µl for dose-response experiments and SN experiments. Assay times were 5 min for CNS or 8 min for SNs. Assay temperature was 30°C; this higher-than-physiological temperature increases product synthesis, making assays more reliable. AC assay solution included 10 µM [³²P]-ATP (25 µCi/ml in whole CNS assays and 400 µCi/ml in SN assays), 50 µM [³H]-cAMP (~3 × 10⁵ cpm/ml), 10 µM GTP, 2.5 U/ml creatine phosphokinase, 5 mM creatine phosphate, 0.5 mM IBMX, 3 mM MgCl₂, 75 mM KCl, 250 µM EGTA, 50 mM K-HEPES (pH 7.6), 1 mM DTT, and protease inhibitors (described in the preceding text). This buffer was designed to produce a total ionic strength of ~100 mM for standard biochemical assays, which is approximately sixfold lower than in Aplysia tissues. Assays were terminated by addition of unlabeled ATP and cAMP, plus sodium lauryl sulfate (Salomon 1979). Cyclic AMP was separated from precursor ATP as described by Salomon (1979). The [³H]-cAMP enabled normalization for recovery after chromatography. Perfused membrane AC assays were conducted as previously described (Jarrard et al. 1993); assay buffer was the same as in steady-state assays, except with 15 µCi/ml [³²P]-ATP. Radioimmunoassays (RIAs) for cAMP were performed using a cAMP RIA kit (Biomedical Technologies, Stoughton, MA) according to the manufacturer's instructions.

Preparation of tissue for radioligand binding assays

The membranes were prepared as in the AC assays except that homogenization of CNSs was in 2 mM Tris-Cl (pH 7.3) with 0.7 M sucrose and protease inhibitors (described in the preceding text), and membranes were thoroughly washed in a series of four 30 min centrifugations at 47,000 g. Pellets were resuspended in 2 mM Tris-Cl, pH 7.3, without sucrose. The pellet after the first centrifugation was sonicated for 60 s (on ice) to remove any residual 5-HT. The final pellet was resuspended in 120 µl/CNS and aliquots stored in liquid nitrogen.

Radioligand binding assays

Binding experiments with 0.2 nM d-[¹²⁵I]-lysergic acid diethylamide (LSD) (2200 Ci/mmol, Dupont NEN) were performed with 25-50 µg of protein/sample either in 55 mM Tris-Cl, pH 7.3, with 5 mM MgSO₄ (low salt), or in (in mM) 460 NaCl, 10 KCl, 10 CaCl₂, 55 MgCl₂, and 10 Na-HEPES, pH 7.3, (physiological salt), in a final volume of 60 µl, at 37°C for 60 min. The binding reaction contained 10 µM paragyline (RBI, MA) and 1.8 mM ascorbic acid (JT Baker, Phillipsburg, NJ). LSD acts as a partial agonist at both 5-HT and DA receptors in Aplysia CNS (Drummond et al. 1980; Southall et al. 1997). To eliminate LSD binding to DA receptors, experiments were carried out in the presence of 300 µM cold DA. Nonspecific binding was defined as counts remaining in the presence of 10 µM cold LSD. Binding reactions were terminated by addition of 3 ml of wash buffer (50 mM Tris-HCl, pH 7.3) and rapid filtration under vacuum over Whatman GF/B glass fiber filters, pretreated for 30 min with 0.5% polyethylenimine (Sigma, St. Louis, MO) and 0.5% nonfat dry milk. Filters were rinsed with 3 × 10 ml of wash buffer and counted.

Electrophysiology

Desheathed pleural ganglia were secured with minuten pins on wax in a recording chamber. Preparations were studied at room temperature. In experiments on excitability, ganglia were superfused (at room temperature) with normal culture medium (in mM): 460 NaCl, 10 KCl, 11 CaCl₂, 55 MgCl₂, and 10 Na-HEPES, pH 7.6, supplemented with nutrients [7 mM glucose, MEM essential and nonessential amino acids (0.2 × normal concentration, GIBCO Invitrogen, Carlsbad, CA), and MEM vitamin solution (0.7 × normal concentration, GIBCO Invitrogen)]. In experiments on spike broadening, ganglia were superfused with high Mg²⁺/high Ca²⁺ culture medium (6 × normal Ca²⁺, 1.6 × normal Mg²⁺) (Goldsmith and Abrams 1991) to reduce spontaneous activity from modulatory interneurons: (in mM) 328 NaCl, 10 KCl, 66 CaCl₂, 88 MgCl₂, and 10 Na-HEPES, pH 7.6, supplemented with the same nutrients. The high-divalent culture medium does not alter transmitter release from the siphon SNs as compared with normal culture medium (Jiang and Abrams 1998) nor of the spike duration in TEA (B. A. Goldsmith and T. W. Abrams, unpublished results), suggesting that Ca²⁺ influx to SNs during action potentials is not altered. Spike broadening measurements were conducted in the presence of 100 mM TEA and 20 µM nifedipine (diluted 1,000-fold from a fresh nifedipine stock in DMSO).

SNs in the VC cluster were penetrated with 10-20 M glass microelectrodes filled with 2 M K-acetate/400 mM KCl. Data for spike duration and excitability were obtained in parallel from two SNs per ganglion and averaged. In spike broadening experiments, action potentials were stimulated with 2-ms depolarizing current pulses at a 15-s interstimulus interval (ISI). Spike duration was measured from the peak to the time at which it had decayed to 33% of the maximum amplitude. SNs in each ganglion were exposed to 5 µM 5-HT and methiothepin in the following sequence: 5-HT for 4 min, followed by 5-HT plus 20 µM methiothepin for 4 min, and finally 5-HT plus 100 µM methiothepin for 5 min. Spike duration before 5-HT, in 5-HT, and in antagonist plus 5-HT were each determined by averaging three consecutive spike widths. Excitability was measured by stimulating SNs with 500-ms depolarizing pulses at a 15-s ISI with two current intensities (1.25 and 2.5 × the threshold current for a 500-ms duration pulse); the two stimulus intensities were alternated every three to four stimuli. SNs were first exposed to 1 µM 5-HT followed by 5-HT plus either 100 µM methiothepin or 200 µM cyproheptadine. Electrophysiological data were acquired digitally with a Modular Instruments interface and were analyzed using Spike software (Hilal Associates, Englewood Cliffs, NJ).

Calculation of effects of antagonists on AC activity

Normalized stimulation of AC activity, expressed as percentage above basal activity was calculated, in the absence of antagonist, as

(1)

or in the presence of antagonist, as

(2)

In these equations, basal activity is defined as activity in the absence of exogenous transmitter. In Eq. 2, stimulation with antagonist is normalized to the same control basal activity (without antagonist) as in Eq. 1; thus any decrease in basal activity produced by an antagonist (without a change in fold-stimulation) is reflected as a reduction in stimulation. Transmitter stimulation of AC in the presence of antagonist as a percent of control stimulation was calculated as

(3)

(i.e., the ratio of Eq. 2 to Eq. 1). The inhibition of transmitter stimulation by antagonist is equal to this value subtracted from 100%.

Data analysis

Dose-response and dose-inhibition data were first normalized to basal activity and then normalized to maximal response within each assay. The assumption of simple competition giving a slope of one was confirmed with a Schild plot (Arunlakshana and Schild 1959). Statistical tests were performed using SPSS software (SPSS, Chicago, IL). Multivariate ANOVA, using a repeated-measures design for comparisons within preparations, was followed by post hoc pairwise comparisons with Bonferroni adjustment for multiple comparisons. Pearson correlation analysis for 5-HT receptors were performed using published inhibition data for mammalian 5-HT receptors.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Comparison of the effectiveness of diverse 5-HT receptor antagonists at the AC-coupled 5-HT receptor in Aplysia CNS

We began this study by examining how 5-HT stimulation of AC in Aplysia CNS was affected by antagonists selective for subtypes of mammalian 5-HT receptors. Although Aplysia hemolymph has a very high ionic strength, these AC assays were conducted in buffer with 100 mM ionic strength; the use of conventional biochemistry buffer enabled the pharmacological sensitivity of the Aplysia receptor to be compared with the pharmacology of mammalian 5-HT receptors. Stimulation of AC in CNS membranes was 117 ± 7% with 1 µM 5-HT and 217 ± 13% with 25 µM 5-HT (n = 33; means ± SE; stimulation expressed as percent above AC activity in the absence of exogenous transmitter). One micromolar 5-HT produces approximately half-maximal activation of AC, whereas 25 µM 5-HT produces near maximal activation. All selective antagonists tested were inactive: the selective 5-HT_1A receptor antagonist NAN-190 (Glennon et al. 1988), the selective 5-HT_2C receptor antagonist RS-102221 (Bonhaus et al. 1997), the selective 5-HT₄ receptor antagonists, GR-113808 and SB-204070 (Grossman et al. 1993; Wardle et al. 1994), and the selective 5-HT₆ antagonist Ro-04-6790 (Sleight et al. 1998) (Fig. 1). Olanzapine, a high-affinity antagonist for 5-HT₂ and 5-HT₆ receptors (Fuller and Snoddy 1992; Roth et al. 1994), was also inactive. All compounds were tested at 10 µM, which is at least three orders of magnitude above the K_i's for the corresponding most sensitive mammalian 5-HT receptor subtype.

View larger version (49K): [in this window] [in a new window]   Fig. 1. Effect of antagonists on serotonin (5-HT) stimulation of adenylyl cyclase (AC) in Aplysia CNS. Stimulation of AC in CNS membranes by 1 (A) or 25 µM 5-HT (B) in the presence of a 10 µM concentration of each antagonist is expressed as percent of control stimulation by the same concentration of 5-HT in the absence of antagonist. In this and subsequent figures, the sample size (n) represents the number of experiments, each of which was conducted on a separate CNS membrane preparation from 3 to 5 animals; within an experiment, each condition was assayed in quadruplicate. In this and other figures, data represent means ± SE (unless stated otherwise). Effects of antagonists on 5-HT stimulation were highly significant [repeated-measures ANOVA, testing stimulation × antagonist interaction, F(13,124) = 27.4 P < 0.001]. Post hoc pairwise comparisons, with Bonferroni adjustment for multiple comparisons, were used to identify significant effects of individual antagonists (*P < 0.05, **P < 0.01, ***P  0.001).

We next tested a number of less specific antagonists. Spiperone, which blocks the PLC-coupled 5-HT receptor from Aplysia CNS, Ap5-HT_B2, had no affect on 5-HT stimulation of AC. Clozapine, cyproheptadine, ritanserin (Ocorr and Byrne 1986), and risperidone were somewhat effective in inhibiting 5-HT stimulation of AC (Fig. 1). This antagonism was partially or fully surmountable with 25 µM 5-HT. Metergoline, fluphenazine, and methiothepin were the most effective antagonists. Metergoline produced 84 ± 7 and 66 ± 10% (n = 4) inhibition of AC stimulation by 1 and 25 µM 5-HT, respectively (Fig. 1). Fluphenazine produced 91 ± 1% inhibition of AC stimulation by 1 µM 5-HT and 58 ± 4% inhibition of AC stimulation by 25 µM 5-HT (n = 5). Methiothepin was the most active antagonist tested, producing 99.9 ± 1.0 and 95 ± 1% inhibition of AC stimulation by 1 and 25 µM 5-HT, respectively (n = 19; see figure legends for these and most other statistical results.)

Effects of antagonists on AC activity in the absence of exogenous 5-HT

Five of the active antagonists, risperidone, ritanserin, clozapine, fluphenazine, and methiothepin caused a decrease in AC activity in the absence of exogenous 5-HT [reducing AC activity on average by 39 ± 5% (n = 3), 12 ± 2% (n = 5), 39%± 2% (n = 3), 26 ± 7% (n = 5), and 40 ± 4% (n = 19), respectively (Fig. 2A). [Overall effects of antagonists on activity were highly significant; F(13,62) = 15.9, P < 0.001; individual probabilities from post hoc pairwise comparisons, with Bonferroni adjustment for multiple comparisons, P < 0.001 for each, except for ritanserin for which P < 0.05]. Olanzapine, which produced a nonsignificant (22 ± 5%) reduction in AC stimulation by 5-HT (Fig. 1A), caused a significant reduction in AC activity in the absence of 5-HT (by 27 ± 5%, P < 0.001, n = 4). Cyproheptadine, at 10 µM, did not have a significant effect on AC activity in the absence of 5-HT; at 200 µM, cyproheptadine reduced AC activity in the absence of exogenous 5-HT by 19 ± 2% (n = 5, P < 0.01, paired t-test). The inactive compounds GR-113808, NAN-190, Ro-04-6790, RS-102221, SB-204070, and spiperone had no effect on activity in the absence of exogenous transmitter. In contrast to the other antagonists, metergoline was a partial agonist; 10 µM metergoline increased AC activity by 42 ± 10% (n = 4) in the absence of 5-HT (Fig. 2A).

View larger version (34K): [in this window] [in a new window]   Fig. 2. Effects of methiothepin and metergoline on AC activity in the absence of exogenous 5-HT. A: AC activity without transmitter and with 1 and 25 µM 5-HT is plotted as a percent of basal activity (without either antagonists or 5-HT); activity is shown for control (without antagonists) and for methiothepin and metergoline. Methiothepin by itself significantly decreased, and metergoline by itself significantly increased, AC activity (P < 0.001 in each case). Antagonist concentration was 10 µM. B: inhibition of AC activity by methiothepin in the absence of 5-HT was highly variable as compared with inhibition of 5-HT-stimulated AC activity. In each experiment, inhibition of AC activity in the absence of 5-HT is expressed as percent of control activity in the absence of exogenous transmitter, and inhibition of 5-HT-stimulated AC activity is expressed as percent of control 5-HT-stimulated activity. [Note that inhibition of activity in the absence of 5-HT or 5-HT stimulated activity is expressed as a positive effect, (in contrast to Figs. 1 and 2A), such that 100% represents complete inhibition].

To analyze the mechanism of the inhibition of basal AC activity, we focused on methiothepin because it had the largest effect. The decrease in basal AC activity could, in principle, be due to a nonspecific effect either on AC itself or on the stimulatory G protein, G_s. To determine whether methiothepin directly inhibited either AC or G_s, independent of any effect on the 5-HT receptor, we examined whether methiothepin altered AC stimulation by small cardioactive peptide B (SCP_B). 5-HT and SCP_B activate AC through independent receptors (Abrams et al. 1984; Ocorr and Byrne 1986). Whereas methiothepin completely blocked 5-HT stimulation of AC, it did not decrease SCP_B stimulation of AC (Fig. 3). This suggests that the effect of methiothepin on basal AC activity results from an interaction with the 5-HT receptor, rather than an interaction with either G_s or AC, or any other nonspecific effects. Why would methiothepin affect AC activity in the absence of exogenous 5-HT? One possibility is that small amounts of residual endogenous 5-HT remain trapped in the membrane preparation and that the inhibition of "basal" activity represents inhibition of AC activity stimulated by residual 5-HT. Low levels of 5-HT trapped in endosomes in the membrane preparation (Schwartz et al. 1979) might gradually leak out, activating receptors. Consistent with the possibility that contaminating 5-HT produced modest AC stimulation that was inhibited by methiothepin, there was no inhibition of basal AC activity in assays on membranes from SNs or desheathed pleural ganglia, which were more dilute and therefore more extensively washed during preparation; thus, any residual 5-HT should be less concentrated (see following text, Fig. 8). Also consistent with this suggestion of contaminating endogenous 5-HT, inhibition of basal AC activity by methiothepin was highly variable among CNS membrane preparations, with inhibition ranging from 8 to 67%; in contrast, inhibition of AC activity stimulated by exogenous 5-HT was very consistent (Fig. 2B).

View larger version (43K): [in this window] [in a new window]   Fig. 3. Effect of methiothepin on AC stimulation by 5-HT and the neuropeptide small cardioactive peptide B (SCPB) in CNS membranes. AC stimulation is expressed as absolute activity (pmoles · mg1 · min1); this enables absolute stimulation of AC activity by SCPB to be compared with and without methiothepin despite the inhibition of activity in the absence of exogenous transmitter. Data are mean ± SE of 3 experiments, conducted in quadruplicate, each on separate CNS preparations. Note that whereas methiothepin was highly effective in inhibiting 5-HT stimulation of AC activity, it produced no inhibition of AC stimulation by SCPB. [F(2,6) = 444, P < 0.001, repeated-measures ANOVA testing antagonist × transmitter interaction; post hoc pairwise comparisons revealed methiothepin effects on stimulation by both 1 and 25 µM 5-HT were highly significant (P < 0.001), whereas the methiothepin effect on stimulation by 10 µM SCP was not significant.]

To directly test whether methiothepin was inhibiting AC activity stimulated by contaminating endogenous 5-HT, we performed a perfused-membrane assay on CNS membranes (Jarrard et al. 1993; Yovell et al. 1987). In this assay, homogenized CNS membranes are trapped on a filter and continuously perfused with large volumes of AC assay solution; because the filter chamber volume changes every 1.5 s, any endogenous 5-HT should be rapidly removed. With the perfused membrane assay, there was no detectable inhibition of activity in the absence of exogenous 5-HT by methiothepin (Fig. 4B). In contrast, in test-tubes assays on this same CNS preparation, there was 40 ± 2% (mean ± SD, n = 5) inhibition of AC activity (Fig. 4A). This lack of effect of methiothepin on basal AC activity in extensively washed membranes argues that contaminating 5-HT accounts for the inhibition of activity in the absence of exogenous 5-HT in the test-tube assays. These perfused membrane results also rule out the possibility that the inhibition of AC activity in the absence of exogenous 5-HT is due to activation of a G_i-coupled receptor by methiothepin.

View larger version (44K): [in this window] [in a new window]   Fig. 4. Extensive washing of CNS membranes eliminates inhibition of AC activity in the absence of 5-HT by methiothepin. AC activity in a single membrane preparation was assayed with 2 techniques: a conventional steady-state assay in test tubes and a perfused-membrane assay. A: methiothepin at 10 µM produced 40 ± 2% reduction in AC activity in steady-state assays on membranes conducted in test tubes. Data are the means ± SD of 5 replicate assays. AC activity is expressed as pmoles cAMP synthesized · mg protein1 · min1. B: methiothepin did not affect AC activity in a continuously perfused membrane preparation. AC activity is plotted before and during exposure to 10 µM methiothepin. Data are the means ± SD of 3 replicate assays from the same membrane preparation as in A. Membranes were trapped on a filter, washed with 30 ml of buffer to remove cytosol, and continuously perfused with [32P]-ATP-containing assay solution; 6-s fractions (250 µl) were collected and the synthesized cAMP was chromatographically separated from the precursor ATP. AC activity is expressed as pmoles cAMP synthesized · mg protein1 · min1, adjusted for steady deterioration of AC activity in membranes on the filter measured prior to the onset of methiothepin exposure (~a 2.3% decrease in activity/6-s interval; see Fig. 6 of Jarrard et al., 1993, for an example). For the assays in A and B, homogenized membranes homogenate from 6 CNSs was separated into 2 aliquots. One aliquot was centrifuged, and the membrane pellet washed to remove cytosol (as described in METHODS). The second aliquot, equivalent to the membranes from 3 CNSs, was divided into 3 portions, each of which was injected onto a filter and used for a single perfused-membrane assay.

If we assume that the AC activity in the presence of methiothepin represents the "true basal" activity in Aplysia CNS membranes, then total AC stimulation by 5-HT would actually be greater than the observed stimulation. When the minimum AC activity in each experiment determined in the presence of methiothepin is used as a measure of true basal activity (in Eq. 1), the total stimulation is 202 ± 15% with 1 µM 5-HT and 366 ± 24% with 25 µM 5-HT (n = 19). The hypothesized contaminating endogenous 5-HT would at least partly explain why 5-HT stimulation is greater in perfused membrane assays than in steady-state test-tube assays (e.g., Figs. 4 and 5 in Jarrard et al. 1993).

Both methiothepin and cyproheptadine block all of the 5-HT receptors in CNS that activate AC

Cyproheptadine has been used in several electrophysiological studies in Aplysia as an antagonist intended to be selective for specific 5-HT receptor subtypes in CNS (Emptage and Carew 1993; Mercer et al. 1991; Sun and Schacher 1996). However, published biochemical studies indicate there is no specificity. Sossin et al. (1994) found that cyproheptadine inhibited 5-HT-stimulated translocation of PKC, suggesting that it inhibits the PLC-coupled 5-HT receptor(s). Goldsmith and Abrams (1992) found that cyproheptadine inhibited 5-HT stimulation of AC in CNS and SN membranes. We confirmed that cyproheptadine inhibits the AC-coupled 5-HT receptor (Fig. 1). In dose-inhibition experiments, cyproheptadine inhibited 5-HT stimulation of AC with an IC₅₀ of 16 µM. Methiothepin was 31-fold more effective than cyproheptadine, inhibiting AC activity with an IC₅₀ of 510 nM (n = 3; Fig. 5A). We also tested 200 µM cyproheptadine, which is the concentration that was used in the earlier electrophysiological studies; at 200 µM, cyproheptadine completely blocked AC stimulation by 1 µM 5-HT and inhibited by 85 ± 1% AC stimulation by 25 µM 5-HT (Fig. 5B).

View larger version (33K): [in this window] [in a new window]   Fig. 5. Effects of methiothepin and cyproheptadine on 5-HT stimulation of AC in Aplysia CNS. A: dose-dependence of effects of methiothepin and cyproheptadine. Stimulation of AC activity by 5 µM 5-HT in the presence of antagonist is expressed as percent of 5-HT stimulation in the absence of antagonist. Data for methiothepin () and cyproheptadine () were obtained in separate experiments; in each experiment all concentrations of one antagonist were tested with a single membrane preparation. Each data point is the mean ± SE for 3 experiments, each of which was conducted in quadruplicate. (Some of the error bars are smaller than the symbols.) Dose-inhibition data were fit with a logistic equation of the form where [B] is the concentration of antagonist, Rmax is the stimulation of AC in the absence of antagonist, Rmin is the stimulation of AC in the presence of the maximally effective concentration of antagonist, IC50 is the concentration of antagonist giving half-maximal inhibition, and n is the Hill coefficient. B: effect of 200 µM cyproheptadine on 5-HT stimulation of AC. 5-HT stimulation of AC activity was calculated according to Eqs. 1 and 2 (see METHODS, n = 5).

We performed a Schild analysis to determine the affinity of methiothepin for the AC-coupled 5-HT receptor. In dose-response experiments, methiothepin behaved as a competitive antagonist, with a K_b of 18 nM (Fig. 6). This value for the K_b for methiothepin agrees well with the value obtained by determining the shift in the dose-response relationship for 5-HT produced by 10 µM methiothepin, (K_b =23 ± 6 nM, mean of 15 experiments). Although we cannot be certain this represents a single AC-coupled receptor, we call this class of receptor 5-HT_apAC.

View larger version (24K): [in this window] [in a new window]   Fig. 6. Competitive antagonism of 5-HT stimulation of AC in CNS membranes. Dose dependence of 5-HT stimulation of AC was determined in the presence of 0 nM (), 20 nM (), 80 nM (), and 320 nM (]) methiothepin. Each data point is the mean ± SD for one experiment, assayed in quadruplicate. The curves are fits to the data of equations of the form where [A] is the concentration of agonist, Rmax is the maximal stimulation of AC by 5-HT, Rmin is AC activity in the absence of agonist, EC50 is the concentration of agonist giving half-maximal response, and n is the Hill coefficient. Inset: Schild plot of these data using the EC50 values from the four logistic equations. Dose ratio (DR) is the ratio of the EC50 of the agonist measured in the presence of antagonist to the EC50 of the agonist alone. Data were fit with the equation where B is the concentration of antagonist; Kb = 18 nM and the slope = 0.88.

Comparisons of the pharmacological sensitivities of the 5-HT_apAC receptor and mammalian 5-HT receptors

We compared the antagonist profile of the 5-HT_apAC class of receptors with those of mammalian 5-HT receptors by calculating Pearson correlations. Competitive dose-response curves were fitted to inhibition data from AC assays for 5-HT_apAC, and dissociation constants were estimated for each antagonist. We used published K_i and K_d values for mammalian receptors (Table 1). Because all of the selective antagonists tested (GR-113808, NAN-190, olanzapine, Ro-04-6790, RS-102221, and SB-204070) had minimal affinity for the 5-HT_apAC class of receptor, they would bias the correlation analysis against those five receptors for which highly selective antagonists existed; therefore we excluded the data for these selective antagonists. Instead we used data for those eight nonselective antagonists for which binding data are available for most receptor subtypes: clozapine, cyproheptadine, fluphenazine, metergoline, methiothepin, risperidone, ritanserin, and spiperone. The Pearson correlation values between the 5-HT_apAC receptor and the 5-HT₁, 5-HT_2A, 5-HT_2C, 5-HT₆, and 5-HT₇ receptors, were 0.29, 0.02, 0.49, 0.92, and 0.53, respectively. Thus the strongest correlation was observed between the 5-HT_apAC receptor and the mammalian 5-HT₆ receptor. The correlations with the 5-HT₄ and 5-HT₅ receptors were not analyzed because for these receptors, information is not available about the affinity of many of these eight antagonists. However, the sensitivity of the 5-HT_apAC receptor to the two most potent antagonists, methiothepin and metergoline, is substantially different from the sensitivities of the 5-HT₄ and 5-HT₅ receptors. Both methiothepin and metergoline are inactive at the 5-HT₄ receptor. At the 5-HT_5A and 5-HT_5B receptors, methiothepin is active (with K_b's of 100 and 16 nM, respectively); however, metergoline is inactive at these receptors. To further assess whether the class of 5-HT_apAC receptors resembles the 5-HT₆ receptor, we examined the relative sensitivity to the agonist 5-CT. At the 5-HT_2A, 5-HT_2C, 5-HT₅, and 5-HT₇ receptors, 5-CT is a more active agonist than 5-HT, whereas at the 5-HT₆ receptor, 5-HT is more active than 5-CT (Hirst et al. 1997; Hoyer et al. 1994). At the 5-HT₄ receptor, 5-CT is inactive. 5-CT was less potent than 5-HT in stimulating AC in Aplysia CNS; EC₅₀s were 1.1 and 13 µM for 5-HT and 5-CT, respectively (Fig. 7). Thus in its sensitivity to both agonists and antagonists, the 5-HT_apAC class of receptor most closely resembles the mammalian 5-HT₆ receptor.

View larger version (17K): [in this window] [in a new window]   Fig. 7. Dose dependence of AC stimulation by 5-HT and 5-carboxyamidotryptamine maleate (5-CT). Stimulation of AC activity by 5-HT () or 5-CT () is expressed as percent of maximal AC stimulation by 5-HT. Data are means ± SE for 3 experiments, each of which was conducted in quadruplicate; within each experiment, both agonists were tested in a CNS membrane preparation from 4 animals. (Some of the error bars are smaller than the symbols.) Dose-response data were fit with a logistic equation of the form where [A] is the concentration of agonist, Rmax is the maximal stimulation of AC by 5-HT, Rmin is AC activity in the absence of agonist, EC50 is the concentration of agonist giving half-maximal response, and n is the Hill coefficient.

Methiothepin inhibits AC-coupled DA receptors

Methiothepin is a nonselective antagonist that affects a wide variety of 5-HT and DA receptors (Hoyer et al. 1994). We tested whether it affects AC-coupled DA receptors in Aplysia CNS. Because DA produces relatively weak AC stimulation, these experiments were carried out with a saturating DA concentration of 200 µM. Methiothepin effectively blocked DA stimulation of AC in CNS membranes; DA stimulation was decreased from 24 ± 2% without methiothepin to 2 ± 1% with methiothepin (stimulation expressed as a percentage above basal activity, without either methiothepin or DA; n = 3). Within these same experiments, fluphenazine caused a more modest and significantly smaller decrease in DA stimulation of AC; stimulation by DA in the presence of fluphenazine was 17 ± 1%. [Both antagonists inhibited DA stimulation significantly; F(1,4) = 34.8, P = 0.004, repeated-measures ANOVA, testing stimulation × antagonist interaction. For post hoc pairwise comparisons, P < 0.001 for methiothepin vs. control and P = 0.01 for fluphenazine versus control; the effects of the 2 antagonists were significantly different, P < 0.001 for methiothepin vs. fluphenazine.]

Inhibition by methiothepin of the AC-coupled 5-HT receptor in Aplysia SNs

We tested whether methiothepin also blocked 5-HT stimulation of AC in membranes from SN somata in the pleural ganglion VC cluster. Stimulation of AC by 5 µM 5-HT (as a percent of basal activity) was 1.0 ± 0.3% in the presence of 20 µM methiothepin versus 160 ± 24% in the absence of antagonist (Fig. 8A). A previous study of 5-HT effects on co-cultured SNs and postsynaptic motoneurons suggested that in SNs, there may be methiothepin-insensitive 5-HT receptors that activate AC (Sun and Schacher 1996). It seemed possible that SNs could have a second type of AC-coupled 5-HT receptor localized to their presynaptic processes in the neuropil. We therefore tested the effect of methiothepin on desheathed pleural ganglia, which contain presynaptic neuropilar processes of the VC cluster SNs. In pleural ganglion membranes, 10 µM methiothepin completely inhibited AC stimulation by 25 µM 5-HT (Fig. 8B). Thus in preparations enriched for SNs, no detectable 5-HT stimulation of AC was mediated by a methiothepin-insensitive receptor.

View larger version (30K): [in this window] [in a new window]   Fig. 8. Effect of methiothepin on 5-HT stimulation of AC in sensory neuron (SN) and pleural ganglion membranes. A: effect of methiothepin on AC in SN membranes. Membranes from pleural ganglion SN clusters were assayed with and without 5 µM 5-HT, in the presence and absence of 20 µM methiothepin. Data are the means ± SE of results from 3 experiments on separate membrane preparations, each from the pleural ganglion SN clusters of 9 animals. Methiothepin significantly inhibited AC stimulation by 5-HT (P < 0.05, 2-tailed t-test). B: effect of methiothepin on AC in pleural ganglion membranes. Membranes from 4 desheathed pleural ganglia, each from a separate animal, were assayed with and without 25 µM 5-HT, in the presence and absence of 10 µM methiothepin. Data are the means ± SD of 5 replicate assays. In both A and B, AC activity is expressed as percent of activity without exogenous 5-HT or methiothepin (basal activity). Note that in contrast to experiments on CNS membranes, in experiments on SN clusters or on desheathed pleural ganglia, methiothepin had no effect on AC activity in the absence of exogenous 5-HT.

Inhibition of the AC-coupled 5-HT receptor in SNs measured under physiological conditions

It has been observed that in intact SNs, methiothepin did not effectively block modulatory effects of 5-HT that are believed to be mediated by cAMP (Sun and Schacher 1996). It seemed possible that the efficacy of methiothepin might be reduced in physiological saline. To assess methiothepin inhibition of 5-HT stimulation of AC in intact SNs, we recorded the broadening of the SN action potential produced by 5-HT in the presence of 100 mM TEA and 20 µM nifedipine (Goldsmith and Abrams 1992; Jarrard et al. 1993). Together these two compounds block the 5-HT-modulated currents except for the two S-K⁺ currents, I_KS,slow and I_{KS,steady state} (Baxter and Byrne 1989, 1990; Edmonds et al. 1990; Goldsmith and Abrams 1992; Sugita et al. 1994). These two remaining 5-HT-sensitive currents are modulated via protein kinase A (PKA); therefore in TEA/nifedipine, spike broadening by 5-HT should be mediated exclusively by cAMP. We measured the effect of methiothepin using a protocol in which 5 µM 5-HT was applied initially in the absence of antagonist and then in the presence of first 20 µM methiothepin and finally 100 µM methiothepin. This approach with sequential comparisons within each SN enables more accurate quantification of the effects of antagonists (Goldsmith and Abrams 1992). In principle, prolonged exposure to 5-HT could result in desensitization, causing spike broadening to decrease during the late phase of the 5-HT exposure, at the time that the antagonist is applied. However, Jarrard et al. (1993) found that in TEA/nifedipine, after 10 min of exposure to 50 µM 5-HT, SNs exhibited no desensitization of the spike broadening response (see also Abrams et al. 1984).

Spike broadening by 5-HT was decreased 55 ± 10% by 20 µM methiothepin and 82 ± 8% by 100 µM methiothepin (P = 0.002 and P = 0.001, respectively). At 100 µM, methiothepin was also significantly more effective than at 20 µM (Fig. 9). We were unable to test higher concentrations to see whether spike broadening could be completely blocked because at concentrations >100 µM, methiothepin precipitates out of physiological saline at pH 7.6 (see METHODS for pH dependence of the solubility of methiothepin).

View larger version (47K): [in this window] [in a new window]   Fig. 9. Effect of methiothepin on broadening of the action potential produced by 5-HT in SNs. Action potentials in pleural ganglion SNs were recorded in high-divalent saline with 100 mM TEA and 20 µM nifedipine. Ganglia were superfused with saline in the following sequence: saline without 5-HT or antagonist, saline with 5 µM 5-HT (4 min), saline with 5 µM 5-HT and 20 µM methiothepin (4 min), and finally saline with 5 µM 5-HT and 100 µM methiothepin (5 min). A: examples of action potentials from one experiment. B: group data from spike broadening experiments. Spike duration in each condition is expressed as percent of control duration in the initial TEA/nifedipine saline. Data are means ± SE of results from SNs from 7 animals. Within each pleural ganglion, action potentials were recorded from 2 SNs in parallel at a 15-s interspike interval (ISI); the last 3 spike durations for both SNs prior to a solution change were averaged. Although 20 µM methiothepin produced a significant decrease in the duration of the 5-HT-broadened action potential, the effect of 100 µM methiothepin was significantly greater [F(2,12) = 38.0, P < 0.001, repeated-measures ANOVA; post hoc pairwise comparisons with Bonferroni adjustment revealed that spike durations in each 5-HT-containing solution differed significantly from the spike durations in the other 5-HT solutions; P = 0.002 for 5-HT +20 µM methiothepin vs. 5-HT, P = 0.001 for 5-HT +100 µM methiothepin vs. 5-HT, and P = 0.027 for 5-HT +20 µM methiothepin vs. 5-HT +100 µM methiothepin].

Methiothepin is lipophilic; in our experiments its effects did not reverse rapidly. Lukyanetz and Kostyuk (1996) and Kostyuk et al. (1992) similarly observed incomplete recovery of 5-HT sensitivity in Helix neurons after washout of methiothepin. In contrast, in studies of dissociated neurons in culture, Sun and Schacher (1996) found that 1 h after washout of methiothepin, the 5-HT response recovered.

Effects of physiological saline on antagonist binding

The partial block of 5-HT-induced spike broadening by methiothepin contrasted with the complete inhibition observed in AC assays on membranes from CNS or SN clusters. Because the cellular electrophysiological studies were conducted in Aplysia physiological saline, which has a high ionic strength (>675 mM), and the biochemical assays were conducted in low-ionic-strength buffer, we hypothesized that high ionic strength contributed to the reduced efficacy of methiothepin in the physiological experiments. We were unable to study directly the effect of physiological salt concentrations on methiothepin inhibition of AC stimulation in homogenized membranes because 460 mM NaCl resulted in a large reduction (approximately sixfold) in AC activity as compared with 100 mM ionic-strength buffer. In principle, this inhibition of basal AC activity by high-ionic-strength physiological saline could be due to an effect on the extracellular surface of the membrane; however, the EC₅₀ for the 5-HT-dependent increase in excitability in intact SNs in physiological saline is in the same range (~1 µM) (Stark et al. 1996) as the EC₅₀ for 5-HT stimulation of AC in membrane homogenates in 100 mM ionic-strength buffer (Fig. 7). Therefore the inhibition by NaCl of basal AC activity is most likely due to an effect on the cytoplasmic surface of the membrane. In several experiments, we examined the effects of methiothepin on 5-HT stimulation of AC in intact SNs using an RIA for cAMP. In contrast to the complete inhibition of AC stimulation by 10 µM methiothepin observed in 100 mM ionic-strength buffer (e.g., Fig. 8A), in physiological saline in the presence of 20 µM methiothepin, 5 µM 5-HT still increased intracellular cAMP in SNs (Fig. 10A). 5-HT stimulation of cAMP levels was 29.6 ± 14.7 fmole/cluster (n = 10) for methiothepin versus 47.5 ± 22.3 fmole/cluster for control saline (in both cases the 5-HT stimulation was significant, P < 0.05, 1-tailed t-test for paired comparisons with contralateral control clusters). With 20 µM methiothepin, the 5-HT stimulation was reduced by 38% as compared with control saline; this reduction in stimulation of cAMP levels, although not significant (unpaired comparisons), was reminiscent of the partial reduction in spike broadening produced by 20 µM methiothepin with the same concentration of 5-HT. Comparison of the partial inhibition by methiothepin of 5-HT stimulation in cAMP RIAs in physiological saline with the complete inhibition observed in AC assays in low-ionic-strength buffer suggested that the increased ionic strength of physiological saline can decrease the affinity of an antagonist. It should be pointed out that in these experiments, we did not distinguish whether the effect of physiological saline was due to ionic strength or to the high concentration of one specific ion.

View larger version (46K): [in this window] [in a new window]   Fig. 10. Effects of physiological saline on ligand binding to the 5-HT receptor. A: effect of methiothepin on stimulation of cAMP levels by 5-HT in intact SNs. RIA measurements of cAMP in pleural ganglion SN clusters were conducted in normal physiological saline with or without methiothepin. Each bilateral pair of SN clusters was preincubated either without methiothepin or with 20 µM methiothepin; one cluster was exposed to 5 µM 5-HT and the contralateral control cluster was incubated without 5-HT. The histogram shows the mean difference (± SE) in absolute quantities of cAMP between contralateral 5-HT-treated and control clusters. Note that in contrast to AC assays on SN membranes conducted in low-ionic-strength assay buffer, in intact SNs in physiological saline, there was not effective inhibition by methiothepin; 5-HT produced significant stimulation of cyclic AMP levels both with methiothepin (n = 10) and without methiothepin (n = 8) (P < 0.05, 1-tailed t-test for paired comparisons with contralateral control clusters). Inhibition by methiothepin was not significant (unpaired t-test). B: effect of ionic strength on d-[125I]-lysergic acid diethylamide (LSD) binding. CNS membranes were incubated with 200 pM [125I-LSD], in the presence of 300 µM dopamine (DA) to block DA receptors, either in low-ionic-strength buffer or in physiological saline. Nonspecific binding, determined by incubating membranes with 10 µM cold LSD in each buffer, was subtracted from total binding. Data are means ± SE of 4 experiments; each experiment consisted of 6 replicate assays per condition.

We now wanted to measure salt effects on the binding of methiothepin to 5-HT receptors. Radiolabeled methiothepin was not commercially available. The ligand [¹²⁵I]-LSD is commonly used to study 5-HT receptors in the presence of cold DA to block LSD binding to DA receptors (Drummond et al. 1980; Southall et al. 1997); we therefore attempted to use [¹²⁵I]-LSD as a ligand to quantify the effect of ionic strength on methiothepin binding to 5-HT receptors. However, physiological saline had the same effect on the affinity of LSD for 5-HT receptors that we predicted occurs with methiothepin: the specific binding of LSD was substantially reduced (Fig. 10B). It was therefore not possible to use this ligand to determine how salt influences the affinity of methiothepin. Nevertheless, these [¹²⁵I]-LSD results directly demonstrate that high-ionic-strength saline can dramatically reduce the affinity of a ligand for 5-HT receptors; a similar reduction in affinity in physiological saline may also occur with methiothepin, as our RIA measurements suggest.

Effects of methiothepin and cyproheptadine on the 5-HT-induced increase in SN excitability

Aplysia SNs normally exhibit dramatic spike frequency adaptation; when stimulated with prolonged depolarizing current pulses, these neurons typically stop firing within the first 100 ms. 5-HT produces an increase in excitability due substantially to a reduction in this spike frequency adaptation (Klein et al. 1986). This 5-HT modulation of excitability is mediated by effects on two K⁺ currents: a reduction in the slowly activating current I_KS,slow decreases accommodation (Goldsmith and Abrams 1992; Klein et al. 1982, 1986) and a reduction in the tonically activated, time-independent current I_{KS,steady state} decreases current threshold (Goldsmith and Abrams 1992; Siegelbaum et al. 1982). Although this modulation of SN excitability in the short term is mediated by cAMP (Goldsmith and Abrams 1992; Hochner and Kandel 1992),¹ cyproheptadine has been found not to affect or only partially affect the 5-HT-induced increase in excitability in pleural ganglion SNs (Mercer et al. 1991; Sun and Schacher 1996). This was puzzling because cyproheptadine blocks the AC-coupled 5-HT receptor in biochemical studies (Figs. 1 and 5) (Goldsmith and Abrams 1992). Furthermore, spike broadening in TEA/nifedipine is mediated by modulation of the same two 5-HT-sensitive S-K⁺ currents that are involved in the excitability increase, and this modulatory effect in intact SNs is completely blocked by cyproheptadine (Goldsmith and Abrams 1992).

We considered the possibility that these inconsistent results may be explained by the ionic-strength-dependent decrease in the efficacy of these antagonists. Would cyproheptadine block the increase in excitability produced by a lower concentration of 5-HT than previously tested? We reexamined the effect of cyproheptadine on the excitability increase by making within-cell sequential comparisons, exposing pleural ganglion SNs to 1 µM 5-HT, followed by 5-HT plus antagonist. This sequential comparison procedure enables more precise quantification of an antagonist's partial inhibitory effects. When SNs were exposed to 1 µM 5-HT in the absence of antagonist, the increase in excitability was maintained for 20 min, indicating that there was no desensitization of the AC-coupled 5-HT receptor (Fig. 11D). Cyproheptadine effectively reversed the 5-HT-induced increase in SN excitability. We observed that the extent of this reversal of the increase in excitability was dependent both on the duration of the exposure to cyproheptadine and on the amplitude of the test current (Fig. 11, A and B). Early during the exposure to cyproheptadine (within 2-4 min), the excitability tested with a current 1.25 × threshold was substantially reduced (P = 0.049), whereas the excitability tested with a current 2.5 × threshold was not significantly affected (P = 0.332, post hoc pairwise comparisons with Bonferroni adjustment for multiple comparisons). Later, when the cyproheptadine effect had reached a maximum (after 4-8 min), the increased excitability at both current intensities was completely reversed (P = 0.033 and P = 0.003 for the low and high current intensities, respectively; Fig. 11, A and B, see legend for ANOVA). Thus depending both on the duration of cyproheptadine exposure and on the test current, the block by cyproheptadine of the excitability increase by 5-HT was either partial or complete. The slower block of the 5-HT-induced excitability increase with the larger test current suggests that cyproheptadine is less effective in physiological saline than in the AC assays.

View larger version (46K): [in this window] [in a new window]   Fig. 11. Both methiothepin and cyproheptadine inhibit the 5-HT-induced increase in excitability in SNs. Excitability in SNs in pleural ganglia was tested by injecting neurons with 500-ms depolarizing current pulses at two intensities: 1.25 and 2.5 × threshold. Stimuli were given at a 15-s ISI in a protocol in which stimulus intensity was alternated every 3 to 4 stimuli. A: cyproheptadine (200 µM) completely reversed the increase in excitability produced by 1 µM 5-HT at both test currents. Note that at the early time point, when perfusion with cyproheptadine first reduced excitability (2 to 4 min after the onset of the exposure), the excitability increase with the lower test current was almost completely eliminated, whereas the excitability increase with the higher test current was not significantly reduced. When the effect of cyproheptadine had reached a maximum, 4-8 min later, the excitability increase with both test currents was reversed. Data are means ± SE from SNs from 5 animals. The overall effect of cyproheptadine to reverse the excitability increase produced by 5-HT was highly significant [F(2,7) = 15.2, P = 0.003, repeated-measures ANOVA testing effect of cyproheptadine exposure; see RESULTS for individual pairwise comparisons]. B: example of the time course of development of the cyproheptadine effect. Data are the average of responses of 2 SNs in a single pleural ganglion. Note that the excitability increase with the lower test current was reversed almost immediately at the start of the cyproheptadine exposure, whereas the excitability increase with the higher test current was blocked gradually and was nearly complete after 4 min of cyproheptadine exposure. This more rapid reversal of the excitability increase at the lower current intensity was observed consistently. C: methiothepin (100 µM) completely reversed the increase in excitability with a test current 1.25 × threshold and significantly reduced the increase in excitability with a test current 2.5 × threshold. Data are means ± SE of results from SNs from 3 animals. [F(1,4) = 42.8, P = 0.03, repeated-measures ANOVA; P = 0.005 for 1.25 × threshold current; P = 0.02 for 2.5 × threshold current, post hoc pairwise comparisons]. D: increased excitability is maintained during a prolonged exposure to 1 µM 5-HT. Example of excitability increase with 5-HT from one experiment; data are the average of responses of 2 SNs in a single pleural ganglion. During >20 min of exposure to 1 µM 5-HT, the excitability increase did not decline, indicating that at this concentration of 5-HT there is no desensitization of the stimulation of AC. [In D, with the onset of the 5-HT exposure, excitability increased earlier at the higher current intensity. This was not a consistent effect but was observed in approximately half of our experiments. In the other half of the experiments, excitability at the two current intensities rose approximately in parallel (e.g., B); in these cases, small differences in time course may have been missed because stimulus intensities were typically alternated every 45 s.]

We next examined whether methiothepin would act like cyproheptadine to inhibit the excitability increase produced by 1 µM 5-HT in SNs. Methiothepin at 100 µM completely blocked the increase in excitability produced by 1 µM 5-HT at a low stimulus intensity (1.25 × threshold), and significantly reduced the excitability increase at a higher stimulus intensity (2.5 × threshold) by 48 ± 4% (Fig. 11C). This excitability change remaining in the presence of 100 µM methiothepin may be explained by the decrease in the affinity of the antagonist in high-ionic-strength saline.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

5-HT antagonists as functional probes

The goals of this study were to pharmacologically characterize the 5-HT receptor or receptors that activate AC in Aplysia CNS. It would be useful to have potent, high-affinity antagonists that selectively block the 5-HT receptors that are coupled to either AC or PLC in Aplysia CNS to study the contribution of each of these signal transduction pathways to neural plasticity. Unfortunately, methiothepin, the most effective antagonist of 5-HT_apAC that we identified, blocks multiple Aplysia 5-HT receptors, including the PLC-coupled 5-HT receptors (Angers et al. 1998; Li et al. 1995); we also observed it blocked AC-coupled DA receptors. In contrast, spiperone blocks the known PLC-coupled 5-HT receptors, Ap5-HT_B1 and Ap5-HT_B2, as effectively as methiothepin (Li et al. 1995) but was inactive at the 5-HT_apAC receptor (Fig. 1). Spiperone is also inactive at the 5-HT_ap1 receptor, which inhibits AC (Angers et al. 1998). Using methiothepin and spiperone in parallel experiments, one can distinguish AC-mediated responses from PLC-mediated responses (B. Dumitriu, J. E. Cohen, and T. W. Abrams, unpublished results). If one can exclude the involvement of DA (e.g., if DA does not mimic the physiological response), methiothepin may be a useful tool in the analysis of the roles of the AC-coupled 5-HT receptor in behavioral plasticity. For example, recently, using methiothepin, Liao et al. (1999) were able to demonstrate that various components of long-term plasticity in Aplysia SNs were differentially dependent on 5-HT.

Effects of high-ionic-strength physiological saline

We observed a decrease in the inhibition by methiothepin in cellular electrophysiological assays conducted on intact SNs compared with the inhibition by methiothepin in biochemical assays. Our results suggest that the high ionic strength of the saline used for cellular studies (or the high concentration of either Na⁺, Mg²⁺, or Cl) weakens the binding of methiothepin to 5-HT receptors. Consistent with this possibility, 5-HT stimulation of cAMP levels in intact SNs in high-ionic-strength saline was not effectively blocked by a concentration of methiothepin (20 µM) that produced maximal inhibition of AC activity in low-ionic-strength buffer. A similar difference has been observed in the inhibition by methiothepin of 5-HT stimulation of AC in buccal muscle membranes and in intact Aplysia buccal muscle assayed in low salt buffer and physiological saline, respectively (L. E. Fox, P. E. Lloyd, J. E. Cohen, and T. W. Abrams, unpublished results). Also consistent with this hypothesized change in antagonist affinity, we found in Aplysia CNS membranes that physiological saline produced a threefold decrease in the binding of a radiolabeled antagonist (Fig. 10B). This concept that the high ionic strength of the physiological saline of marine animals can significantly alter the inhibition produced by an antagonist is supported by radioligand binding results demonstrating that ionic-strength differentially affects the affinity of various ligands (Hou et al. 1996).

Recently investigators have expressed recombinant Aplysia 5-HT receptors in mammalian cell lines and measured ligand binding and second-messenger stimulation under low-ionic-strength conditions (e.g., Angers et al. 1998; Li et al. 1995). Our results suggest that pharmacological assays of recombinant receptors performed at low ionic strength may not accurately reflect the efficacy of antagonists in the high-ionic-strength physiological saline of marine invertebrates.

Comparison of the 5-HT_apAC receptor with other Aplysia 5-HT receptors

Several Aplysia 5-HT receptor subtypes have been characterized using electrophysiological analysis, photoaffinity labeling, and molecular techniques. Gerschenfeld and Paupardin-Tritsch (1974) distinguished six different 5-HT responses in Aplysia neurons based on distinct pharmacologies that presumably are mediated by different receptor subtypes. Photoaffinity labeling of 5-HT receptors in Aplysia CNS suggested the existence of at least five serotonin receptor subtypes (Saitoh and Shih 1987). To date, five Aplysia receptors have been cloned. Li et al. (1995) cloned two PLC-coupled receptors: Ap5-HT_B1, which is expressed in the ovotestis and spermatheca, and Ap5-HT_B2, which is expressed in the CNS. Unlike the 5-HT_apAC class of receptor, these two PLC-coupled receptors were highly sensitive to spiperone. Angers et al. (1998) and Barbas et al. (2002) cloned two 5-HT receptors expressed in CNS, which inhibit AC: the 5-HT_ap1 and 5-HT_ap2 receptors. These two receptors display a pharmacological profile distinct from that of the 5-HT_apAC receptor. A fifth Aplysia 5-HT receptor has been partially cloned, which is expressed at high levels in peripheral tissues and at a low level in CNS (Williams et al. 1997). Methiothepin is an effective antagonist against all four recombinant Aplysia 5-HT receptors: Ap5-HT_B1, Ap5-HT_B2, 5-HT_ap1, and 5-HT_ap2 as well as against 5-HT_apAC. Similarly, methiothepin is an effective antagonist at all mammalian G-protein-coupled 5-HT receptor subtypes, except the 5-HT₄ receptor (Hoyer et al. 1994).

Which 5-HT receptors are blocked by cyproheptadine?

Several studies have used cyproheptadine to selectively interfere with specific forms of 5-HT-dependent plasticity in SNs (Emptage and Carew 1993; Mercer et al. 1991; Sun and Schacher 1996). Mercer et al. (1991) found that 200 µM cyproheptadine blocked spike broadening induced by 5 µM 5-HT without affecting the accompanying increase in excitability. Emptage and Carew (1993) used cyproheptadine to block 5-HT-induced short-term synaptic facilitation without interfering with long-term facilitation. Based on these dissociations, it was proposed that broadening of the SN action potential and short-term facilitation are triggered via cyproheptadine-sensitive receptors, at least some of which activate PLC, whereas enhanced excitability and long-term facilitation are triggered via cyproheptadine-insensitive receptors that activate AC. In contrast, the biochemical evidence indicates that cyproheptadine does not differentiate among 5-HT receptors in Aplysia CNS. In pleural-pedal-ganglia, cyproheptadine (at 200 µM) inhibits translocation of PKC by 5-HT, presumably by blocking PLC-coupled receptors (Sossin et al. 1994).² In SN membranes in low-ionic-strength buffer, AC stimulation by 5-HT is also blocked by 200 µM cyproheptadine (Goldsmith and Abrams 1992).

If there were a cyproheptadine-insensitive, AC-coupled 5-HT receptor that primarily mediates the increase in excitability, then one would expect to see residual AC stimulation by 5-HT in the presence of cyproheptadine. Although with the 5 µM 5-HT concentration used in the earlier electrophysiological experiments (Emptage and Carew 1993; Mercer et al. 1991), cyproheptadine never inhibited >92% of AC stimulation in biochemical assays (Fig. 5A), with 1 µM 5-HT, AC stimulation was completely blocked by 200 µM cyproheptadine (Fig. 5B). It could be argued that a second AC-coupled 5-HT receptor present in intact SNs does not contribute detectably to 5-HT stimulation of AC in homogenized membranes. It is unlikely that a 5-HT receptor that activates AC went undetected in our AC assays for several reasons. In a wide range of receptor systems, it is possible to study agonist stimulation of AC after homogenizing membranes; therefore it is not likely that the hypothesized cyproheptadine-insensitive receptor became nonfunctional. It is possible that a less abundant, regionally restricted, AC-coupled 5-HT receptor contributes importantly to SN physiology but was missed in these biochemical experiments. However, electrophysiological experiments also indicate that there is not a cyproheptadine-insensitive (or methiothepin-insensitive) 5-HT receptor in SNs that activates AC. Goldsmith and Abrams (1992) found that 200 µM cyproheptadine inhibited by >98% the spike broadening in TEA/nifedipine produced by 1 µM 5-HT; because TEA and nifedipine block the other 5-HT-modulated currents, spike broadening in the presence of these two compounds is almost entirely due to cAMP-dependent reduction in the two S-K⁺ currents.

If cyproheptadine is a broad spectrum 5-HT antagonist, how can it selectively block a subset of the responses to 5-HT? It is possible that 200 µM cyproheptadine only partially blocked the 5-HT-stimulated increase in cAMP levels in electrophysiological experiments, in contrast to the complete inhibition by cyproheptadine of 5-HT stimulation of AC in biochemical experiments. Such incomplete inhibition of AC stimulation could have resulted from a reduction in the affinity of cyproheptadine for the 5-HT_apAC receptor in high-ionic-strength physiological saline as we observed occurs with methiothepin and LSD. Thus a residual increase in cAMP levels that persists with cyproheptadine may have mediated the 5-HT-induced increase in excitability previously seen in the presence of this antagonist (Mercer et al. 1991). Consistent with the possibility that cyproheptadine at 200 µM is blocking the AC-coupled 5-HT receptor, though only partially, we observed that cyproheptadine was effective in blocking the excitability increase initiated by a fivefold lower concentration of 5-HT than tested in the studies of Mercer et al. (1991) (Fig. 11A). Also consistent with the possibility that 200 µM cyproheptadine only partially blocks the AC-coupled 5-HT receptor in physiological saline, we observed that the efficacy of cyproheptadine in blocking 5-HT modulation of SN excitability depended on the precise test conditions (Fig. 11, A and B). Differential inhibition of increased excitability depending on the test current was also observed with methiothepin (Fig. 11C). The observation that the efficacy of an antagonist varies substantially within a single neuron, depending on the intensity of the test current or the duration of exposure to the antagonist, suggests that the antagonist concentration is within the steep region of the dose-response curve; in this concentration range, small shifts in parameters can produce large changes in the physiological effect of the antagonist. To optimally detect inhibitory effects of an antagonist, it is important that the agonist (e.g., 5-HT) be used at concentrations that are substantially below saturating for the response being studied.

Increased excitability is particularly prone to being insensitive to antagonists that only partially block a population of receptors. The increase in SN excitability has a strikingly nonlinear dependence on the underlying decrease in K⁺ current, due, at least in part, to the discrete threshold for spike initiation. When measuring excitability, a partial inhibition of the 5-HT-induced reduction in K⁺ current may not be detected if the remaining K⁺ current is not sufficient to prevent the initiation of action potentials. Moreover, the slowly activating S-K⁺ current (I_KS,slow) is modulated by relatively low cAMP concentrations (Goldsmith and Abrams 1992); perhaps as a consequence, increased excitability occurs at relatively low concentrations of 5-HT (Stark et al. 1996) and may require the activation of only a modest percentage of receptors. This would make the increase in excitability difficult to block, particularly when the 5-HT concentration is substantially above the EC₅₀ for activation of AC. (The hypothesis that phosphorylation of S-K⁺ channels occurs at a high rate and is difficult to block is illustrated in Fig. 12A.)

Sun and Schacher (1996) found in cultured SNs that the inhibition by cyproheptadine of the 5-HT-stimulated increase in excitability varied substantially depending on the specific culture conditions. These authors interpreted this variation in the efficacy of cyproheptadine as indicating that under some culture conditions, the SNs express a cyproheptadine-insensitive (and methiothepin-insensitive) receptor that activates AC. Because our biochemical results indicate that the AC-coupled 5-HT receptors in CNS were all sensitive to both cyproheptadine and methiothepin, we interpret the variable block by these antagonists in electrophysiological studies as indicating that the antagonist concentration used was insufficient to completely block cAMP increases and that the modulatory cascade is expressed more powerfully at certain times as synapses are forming in culture (see cartoon in Fig. 12B). Given that in our experiments, the efficacy of both methiothepin and cyproheptadine varied with the precise experimental parameters, it is not surprising that the blockade of the excitability increase by these antagonists can change substantially as synapses develop in culture.

View larger version (47K): [in this window] [in a new window]   Fig. 12. Schematic representation of the hypothesis that multiple modulatory effects of protein kinase A (PKA) could be differentially sensitive to a 5-HT receptor antagonist, either because of differences in the phosphorylation rate of PKA substrates or because of developmental and regional differences in abundance of a protein in the cAMP cascade. Each cartoon illustrates a SN cell body (above) and a presynaptic terminal (below). A: high rate of phosphorylation of S-K+ channels by PKA makes these cAMP substrates less sensitive to blockade by a nonsaturating concentration of 5-HT receptor antagonist. A1: activation of receptor-Gs-AC complex by 5-HT stimulates synthesis of cAMP in a SN. Binding of cAMP to the regulatory units of PKA results in dissociation of the catalytic subunits, which then phosphorylate target proteins. In the synaptic terminal, the catalytic subunits phosphorylate an exocytosis-related protein, either in the vesicle membrane or nearby in the active zone, shifting vesicles into a high release probability state (indicated by the hatched pattern). In the cell body, phosphorylation of S-K+ channels by PKA converts the channels to a closed, nonactivatable state. These K+ channels are also located in the axonal membrane (Clatworthy and Walters 1993; Kaplan et al. 1994; Klein et al. 1986); they are shown exclusively in the soma because this is the site where excitability is measured. Some catalytic units are shown modulating multiple S-K+ channels, including recently phosphorylated channels (signified by dashed arrows) because the phosphorylation of the S-K+ channels, which influence excitability, appears to be extremely rapid (signified by thick arrows). This conclusion is based on the recent observation that a transient increase in cAMP produces maximal modulation of S-K+ channels within 2 s (Cohen and Abrams, unpublished). In contrast, phosphorylation of the exocytosis proteins is shown at lower rate (signified by thin arrows); thus a single catalytic unit can modulate fewer vesicles during the same period. A2: an antagonist (such as cyproheptadine) inhibits the majority of 5-HT receptors. Most of the S-K+ channels are still effectively modulated by the small number of activated PKA catalytic units. In contrast, phosphorylation of the exocytosis proteins at the active zones, which we suggest occurs at a slower rate, is almost completely blocked by the partial inhibition of 5-HT receptors. B: at a developmental stage at which synapses are forming, a protein involved in the cAMP cascade is upregulated, in this example the 5-HT receptor; this hypothetical upregulation reduces the sensitivity to a 5-HT receptor antagonist. B1: in the absence of 5-HT receptor antagonist, both PKA substrates, the S-K+ channel and the vesicle-associated protein are fully phosphorylated by 5-HT. B2: with inhibition of the majority of 5-HT receptors by antagonist (e.g., methiothepin or cyproheptadine), there is still effective enhancement of vesicle release because sufficient numbers of 5-HT receptors remain activated to produce substantial activation of PKA. (This reduced sensitivity to antagonist could similarly occur with an upregulation of AC or PKA.) We hypothesize that later, as synapses mature, 5-HT receptors are downregulated (as in A). Although there is no evidence that demonstrates a change in receptor density during synapse formation, similar up- and downregulation of cell-signaling proteins commonly occurs during development. Yovell et al. (1992) demonstrated that an increase in the concentration of the signaling protein calmodulin results in a substantial increase in the sensitivity to the second-messenger Ca2+. Such a change in receptor number could account for the observed developmental decrease in the efficacy of an antagonist (Sun and Schacher 1996).

It is important to emphasize that our finding that cyproheptadine does not discriminate among 5-HT receptors or among the second-messenger cascades that 5-HT initiates does not weaken the two central conclusions of the early studies that used cyproheptadine: 1) long-term facilitation can be initiated without short-term facilitation and 2) during short-term facilitation, presynaptic spike broadening and increased transmitter release are independent of changes in excitability (Mercer et al. 1991). It is also important to point out that our analysis does not address the issue of whether there are multiple AC-coupled 5-HT receptors in Aplysia CNS. We believe, however, that the predominant receptors that activate AC are sensitive to both cyproheptadine and methiothepin at concentrations of 100 µM.

In summary, our results indicate that it is possible to observe selective blockade of a subset of effects produced by a single modulatory cascade when there is incomplete inhibition of that cascade. Indeed, even a single modulatory effect may be differently inhibited depending on test conditions. Thus dissociation of different effects due to partial inhibition of a cascade may not demonstrate mediation by distinct signaling cascades. In general, caution is warranted in using receptor antagonists to dissociate second-messenger cascades when monitoring downstream physiological changes. Dose-inhibition curves in combination with the use of nonsaturating concentrations of agonists reduce the likelihood of being misled by selective inhibition of a subset of responses.

Comparison with classes of mammalian 5-HT receptor

None of the antagonists with high selectivity for mammalian 5-HT receptors effectively blocked the 5-HT_apAC receptor. Specific antagonists targeted against mammalian 5-HT receptor subtypes may be inactive against homologous molluscan 5-HT receptors because of the large evolutionary distance between these two phyla (Peroutka and Howell 1994); thus small numbers of amino acid substitutions near the binding site might substantially decrease the affinity of these highly specific antagonists and yet have more modest effects on the affinity of broad spectrum antagonists. We therefore compared the sensitivity of the 5-HT_apAC class of receptor to the less specific antagonists with the sensitivities of various mammalian 5-HT receptor subtypes to these same antagonists. Our Pearson correlation analysis suggested that the pharmacological profile of the 5-HT_apAC receptor was most similar to the profile of the 5-HT₆ receptor, with a correlation of >0.9, compared with correlations of ~0.5 for the next two most similar 5-HT receptor subtypes. Moreover, both the 5-HT_apAC and 5-HT₆ receptors differ from the 5-HT₄, 5-HT₅, and 5-HT₇ receptors in that, at the 5-HT_apAC and 5-HT₆ receptors, 5-HT has both higher potency and efficacy than the agonist 5-CT. Therefore we predict that 5-HT_apAC receptor is most homologous to the 5-HT₆ receptor. Interestingly, the 5-HT₆ class of receptors evolved relatively early, ~750 million years ago, well before the divergence of vertebrates from invertebrates (Peroutka and Howell 1994). The selective 5-HT₆ antagonist that we tested, Ro-04-6790, was completely inactive at the 5-HT_apAC receptor; however, this is consistent with the perspective that highly specific antagonists against mammalian receptors are likely to be ineffective against phylogenetically distant, though homologous, invertebrate receptors. Cloning of the 5-HT_apAC receptor or receptors will enable confirmation of the predicted homology between the 5-HT_apAC receptor and the mammalian 5-HT₆ receptor.