Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Serotonergic pathways in the CNS have received significant attention over the past two decades because of their involvement in the treatment of psychiatric illness. Despite this effort, the mechanisms by which these systems affect psychiatric conditions have remained elusive. One reason for this is the complexity of serotonergic systems, with at least 14 receptor subtypes identified in mammalian tissue (Hoyer and Martin, 1997). The 5-HT₃ receptor initially received little attention in behavioral studies, with research limited to its role in emesis. This receptor is present in high concentration in gut, where it affects gastric motility, and in the CNS, at highest concentrations in the area postrema, a postulated site of antiemetic action. Although most of the 5-HT receptor subtypes use second messengers to modulate postsynaptic signal transduction, the 5-HT₃ receptor is unique in its functional role as the only identified serotonergic ligand-gated ion channel receptor. Outside the brainstem, the 5-HT₃ receptor is present in relatively low abundance, mostly concentrated in amygdala, entorhinal cortex, and hippocampus (Kilpatrick et al., 1988; Barnes et al., 1990; Laporte et al., 1992; Steward et al., 1993).

During the past decade, there has been a developing interest in the functional role of 5-HT₃ limbic sites in relation to psychiatric illness. 5-HT₃ antagonists have been reported to modulate the release of dopamine and other neurotransmitters, including 5-HT itself (for a review, see Gozlan, 1997). Behavioral studies with 5-HT₃ antagonists have shown that these antagonists have the ability to alter behaviors in animal models of anxiety, dementia, and substance abuse and withdrawal, suggesting that these receptors may play an important role in cognition and motivation (Costall and Naylor, 1992). Clinical trials with these compounds in relation to their anxiolytic and cognitive action have been limited and inconclusive (Bentley and Barnes, 1995).

A primary factor hampering the investigation of 5-HT₃ receptors has been the dearth of selective compounds appropriate for radioligand studies (Gaster and King, 1997). Because these receptors are present in low abundance in the brain (e.g., <3 fmol/mg tissue) (Laporte et al., 1992, Steward et al., 1993), 5-HT₃ ligands must have high affinity, high selectivity, and low nonspecific binding and must be radiolabeled with high specific activity to accurately assess effects of treatments that might alter receptor characteristics. Until recently, no single ligand has met all of these requirements. Previous radiolabeled 5-HT₃ ligands have come from five structural classes (Gaster and King, 1997). Although members from each these classes have been radiolabeled with tritium, these radioligands have lacked adequate specific activity and/or selectivity to study changes in 5-HT₃ binding characteristics.

Recently, we reported the development of a series of zacopride derivatives as potential high-affinity 5-HT₃ antagonists radiolabeled with ¹²⁵I, an isotope with a specific activity (2170 Ci/mmol) more suitable for studies of receptors present in low abundance. Previous work had demonstrated that exchanging the chlorine atom of zacopride for an iodine atom resulted in an 8- to 10-fold loss of activity (Ponchant et al., 1991; de Paulis et al., 1997; Hewlett et al., 1997). The resulting compound, deschloro-5-iodo-(S)-zacopride, in its radiolabeled form, (S)-[¹²⁵I]iodozacopride, is the only ¹²⁵I-labeled 5-HT₃ ligand that is commercially available. Contrary to an early report (Gehlert et al., 1993), however, it has only modest affinity for the 5-HT₃ receptor (K_D = 1.3 nM; Ponchant et al., 1991; Hewlett et al., 1997). Two other analogs of zacopride, [¹²⁵I](S)-N-(1-azabicyclo[2.2.2]oct-3-yl)-5-iodo-2,3-dimethoxybenzamide (MIZAC) (K_D = 1.5 nM), and its iodoallyl homolog, [¹²⁵I]trans-(S)-N-(1-azabicyclo-[2.2.2]oct-3-yl)-5-chloro-2-(3-iodo-2-propenyloxy)-3-methoxybenzamide (LIZAC) (K_D = 0.3 nM), have been characterized. Both of these compounds appear to specifically bind to an uncharacterized "benzac" site (Hewlett et al., 1998a, 1998b) in addition to the 5-HT₃ receptor. Yet another substituted benzamide, (S)-N-(1-azabicyclo-[2.2.2]oct-3-yl)-4-amino-5-chloro-3-iodo-2-methoxybenzamide (TRIZAC), showed high 5-HT₃ affinity (0.05 nM) (de Paulis et al., 1996) but proved problematic to radiolabel. Based on the facts that a large substituent in the aromatic 3-position enhanced activity and that the 4-amino group of zacopride was not an absolute structural requirement for high activity (Wong et al., 1993; de Paulis et al., 1997), DAIZAC (Fig. 1) was prepared and conveniently radiolabeled (Mason et al., 1996). Here, we present the in vitro pharmacological characterization of [¹²⁵I]DAIZAC as a highly selective, high-affinity antagonist radioligand for 5-HT₃ sites using murine preparations of rat brain 5-HT₃ receptors and mouse 5-HT₃ receptors expressed in cloned NCB-20 cells (Lovinger and White, 1991).

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Chemicals. (S)-Zacopride [(S)-4-amino-5-chloro-2-methoxy-N-(1-azabicyclo[2.2.2]oct-3-yl)benzamide] and (S)-iodozacopride [(S)-4-amino-5-iodo-2-methoxy-N-(1-azabicyclo[2.2.2]oct-3-yl)benzamide] were synthesized as reported previously (Hewlett et al., 1997). Acetylcholine, adenosine, chlorpheniramine, cimetidine, dexamethasone, kainic acid, and spironolactone were obtained from ICN Diagnostics (Costa Mesa, CA). Other neurotransmitter ligands were obtained from Research Biochemicals (Natick, MA). Unlabeled DAIZAC was prepared as reported previously (Mason et al., 1996). Radiolabeled [¹²⁵I]DAIZAC was prepared from the corresponding 3-stannyl derivative, obtained from unlabeled DAIZAC using bis(tributyltin) in the presence of triphenylphosphine-tetrakispalladium in triethylamine as reported previously (Mason et al., 1996). Other biochemicals for binding assays were obtained from Sigma Chemical (St. Louis, MO) and were reagent grade.

Rat Brain Membrane Preparation. Male Sprague-Dawley rats (250-300 g; Harlan Inc., Indianapolis, IN) or frozen rat brains (Pel Freeze Biologicals, Rogers, AK) were used. Dissection of brain regions was carried out as described previously (Hewlett et al., 1998a). Brain regions or whole brain without cerebellum were homogenized (13 ml/g tissue) using a Brinkman Polytron model PT 3000 (15 s at 21,700 rpm) in 50 mM HEPES buffer (pH 7.4), containing 5 mM CaCl₂ and 2.4 mM MgCl₂ (Hewlett et al., 1998a). The homogenate was centrifuged at 10,000g at 4°C for 15 min, and the supernatant was discarded. The membrane pellet was resuspended in the same volume of buffer and centrifuged a second time at 10,000g at 4°C for 15 min, and the resulting pellet was resuspended in cold HEPES buffer at 60 mg/ml (whole brain) or 20 mg/ml (brain regions).

NCB-20 Cell Membrane Preparation. NCB-20 cells (Chinese hamster × neuroblastoma hybrid) were grown in Dulbecco's modified Eagle's high glucose medium (Lovinger and White, 1991). Near confluent cells (70-80%) were pooled by centrifugation and resuspended in 50 mM NaH₂PO₄ buffer at pH 7.5, containing 140 mM NaCl, 3 mM EDTA, 50 µg/ml bacitracin, and 0.1 nM PMFS. The cells were collected and disrupted by Polytron (12,500 rpm for 20 s), followed by low-speed centrifugation (600g for 10 min). The pellet containing nuclei was discarded. The supernatant, containing the membranes, was centrifuged at 40,000g for 20 min. The pellet was diluted to ~0.7 mg protein/ml in 10 mM HEPES (pH 7.5), containing 50 µg/ml bacitracin and 0.1 mM phenylmethylsulfonyl fluoride. Protein content was determined by the Folin-phenol method using bovine serum albumin as standard.

In Vitro 5-HT₃ Binding Assay. For saturation analysis, whole rat brain without cerebellum (final concentration, 37.5 mg wet weight/ml), or NCB-20 cell membranes (final concentration, 0.08 mg protein/ml), were incubated with [¹²⁵I]DAIZAC in concentrations ranging from 0.006 to 1.6 nM in 50 mM HEPES buffer (pH 7.4) containing 5 mM CaCl₂ and 2.4 mM MgCl₂ at 20°C for 30 to 60 min in a total volume of 0.40 ml. Nonspecific binding was determined by incubation in the presence of the highly selective, structurally dissimilar 5-HT₃ receptor antagonist, bemesetron (MDL-72222): 4 to 50 µM for rat brain tissue and 50 µM for NCB-20 cell membranes. Competition studies for both membrane preparations were performed using varying concentrations of 5-HT₃ agonists or antagonists to inhibit binding of 0.1 nM [¹²⁵I]DAIZAC. For kinetic analyses, the rates of specific [¹²⁵I]DAIZAC binding to 5-HT₃ receptors were determined at 20°C. For determining the association rate, [¹²⁵I]DAIZAC (0.10 nM final concentration) was added to 5 ml of whole rat brain membrane preparation. At specified time intervals, 0.5-ml aliquots were filtered. For determining the dissociation rate, the rat brain membrane preparation (5 ml) was preincubated with [¹²⁵I]DAIZAC (0.10 nM final concentration) for 30 min; 10 µM bemesetron was added; and at specified time intervals, 0.5-ml aliquots were filtered. Bound and free [¹²⁵I]DAIZAC were separated by rapid filtration through 32 fiberglass filters (Schleicher and Schuell, Keene, NH), presoaked in 0.3% polyethylenimine, using a Brandel M-24R cell harvester. The filters were washed three times for 10 s with 50 mM Na₂HPO₄ buffer (pH 7.4). Gamma spectrometry was performed using an ICN Biomedic Isomatic 4/600 HE instrument at 80% efficiency.

Selectivity for the 5-HT₃ Receptor. The selectivity of DAIZAC for the 5-HT₃ receptor was assessed by three methods. The first method involved comparisons of displacement of [¹²⁵I]DAIZAC by excess DAIZAC (10 µM) and bemesetron (4 µM) at different concentrations of [¹²⁵I]DAIZAC to determine whether the labeled ligand was bound to a site not accessible to bemesetron. As a second method, labeled [¹²⁵I]DAIZAC was incubated with high concentrations (3-30 µM) of unlabeled antagonists for different CNS neurotransmitter receptors and uptake carriers. In these experiments, [¹²⁵I]DAIZAC (3 nM) was used at a concentration 20 times higher than its K_D to increase the probability of labeling sites other than the 5-HT₃ receptor. The third method involved competition studies with radioligands selective for other binding sites using either full displacement curves of unlabeled DAIZAC or, in some cases, a single high concentration of unlabeled DAIZAC (10 µM) to displace the non-5-HT₃ radioligand binding. Full inhibition curves were carried out when 10 µM DAIZAC displaced >25% of the radioligand. Thirty-four of these competition assays were performed with NOVASCREEN (Hanover, MD). Affinities of DAIZAC for the human 5-HT₆ and 5-HT₇ receptors were determined by Dr. M. Hamblin (Veterans Affairs Medical Center, Seattle, WA) using previously reported methods (Shen et al., 1993; Kohen et al., 1996). Briefly, transfected cell lines (HeLa E6-7 and HeLa E7-2) were grown, harvested, homogenized, and incubated with 1.85 nM [³H]lysergic acid diethylamide (DuPont NEN, Billerica, MA) at 37°C for 90 min in 100 mM Tris·HCl, 5 mM MgSO₄, 1 mM EDTA, and 1 mM Na ascorbate, pH 7.4. Nonspecific binding was defined by coincubation with 5 µM methiothepin. In the case of the 5-HT₄ receptor, rat striatal membranes were incubated with the 5-HT₄ receptor selective radioligand [³H]GR-113808 (0.2 nM; Amersham Life Science, Arlington Heights, IL) at 37°C for 30 min as described by Grossman et al. (1993) using unlabeled DAIZAC or (S)-zacopride in concentrations ranging from 0.5 to 5000 nM. Nonspecific binding was defined by coincubation with 10 µM SDZ 205-557. After filtration as described, the filters were suspended in scintillation fluid and counted in a Beckman LS 1801 instrument operating at 45% efficiency.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Chemical structure of DAIZAC.

Data Analysis. For saturation analysis, the binding data were analyzed by Lundon Software (Chagrin Falls, OH). The k_on and k_off were calculated according to the method of Weiland and Molinoff (1981). An estimate of K_D was determined using the equation K_D = k_off/k_on. IC₅₀ values were calculated from log-logit analysis. Inhibition constants were calculated using the Cheng-Prusoff equation, K_i = IC₅₀/(1 + L/K_D), where L is the concentration of [¹²⁵I]DAIZAC and K_D is 0.15 nM, and compared with values found in the literature. Percent occupancy of a ligand at given site, derived from the law of mass action, was estimated as F/(F + K_D), where F is the free concentration of the ligand and K_D is the affinity of the ligand for the site.

Autoradiography. Rat brain was removed and rapidly frozen in 2-methylbutane at 20°C. Blotted, frozen rat brain was coronally sectioned on a cryostat microtome (Richard-Jung, Cryocut 1800) in slices of 30-µm thickness. The thawed, slide-mounted rat brain sections were preincubated in a buffer containing 50 mM HEPES buffer (pH 7.6), 5 mM CaCl₂, and 2.4 mM MgCl₂, at 20°C for 30 min, and then incubated in buffer with [¹²⁵I]DAIZAC (0.075 nM) at 20°C for 1 h. Nonspecific binding was determined in every fifth section by coincubation with 4 µM bemesetron. After incubation, sections were washed twice in ice-cold 50 mM Na₂HPO₄ buffer (pH 7.4; 1 min and then 2 min) and then washed with cold distilled water for 1 min and allowed to dry overnight at 20°C. Autoradiograms were prepared using Hyperfilm B-Max (Amersham, Arlington Heights, IL) with an exposure time of 24 h. Film was developed using Kodak D-19 developer (5 min), rinsed in 2% acetic acid (20 s), and fixed with Kodak rapid fixer (4 min). Autoradiograms were scanned using Pixeltools Nubus Board (Perceptic Corporation), installed on a Macintosh Quadra 900 computer, using COHU solid state camera illuminated by Chroma Pro 45 light source. Image analysis was performed using Image 1.49 software (National Institutes of Health, Bethesda, MD).

Whole-Cell Patch-Clamp Recording. Cell currents due to activation of 5-HT₃ receptors were recorded from NCB-20 cells as described previously (Lovinger and White, 1991). Whole-cell recordings were performed using glass microelectrodes (1-3 M) at room temperature with an Axopatch 200 (Axon Instruments, Foster City, CA) patch-clamp amplifier. Cells were perfused at 1.5 ml/min using an extracellular medium containing 150 mM NaCl, 5 mM KCl, 2.5 mM CaCl₂, and 10 mM HEPES (pH 7.4; osmolarity, 0.34 mol/kg). The patch electrode dialyzing the interior of the cell contained 140 mM CsCl, 5 mM EGTA, and 10 mM HEPES (pH 7.4 using CsOH; osmolarity, 0.31 mol/kg). Series resistance was generally <10 M. All of the experiments were performed under voltage-clamp with the membrane potential at 60 mV. The neurotransmitter and DAIZAC were dissolved in extracellular medium and delivered from an array of 12 large-bore (>0.15 mm inner diameter) HPLC tubes placed near the bottom of the culture dish. The tubes were connected to solution-containing reservoirs placed above the preparation, allowing for steady gravity-induced flow of solutions from the tubes. Solution flow was gated by plastic stopcocks connected to the tubing, which was continuous with each tube. Solution exchange could be completed within 150 ms by initiating flow and manually moving the tube array such that the desired tube was facing the cell at a distance of ~0.1 mm. In experiments to determine the effects of different DAIZAC concentrations on 5-HT-induced current, DAIZAC was applied for 60 s (0.01-1 nM) or 5 s (10-100 nM) before combined 5-HT/DAIZAC application. The different DAIZAC pretreatment times were necessary to ensure that maximal inhibition by DAIZAC was achieved at each concentration. The signal was filtered (2-5 kHz cutoff frequency, three-pole bessel filter), digitized at 1 kHz with a Labmaster DMA (Axon Instruments), and recorded on a 386 microcomputer (Dell System 310, San Antonio, TX) using the pClamp software (Axon Instruments). Current records were analyzed off-line using pClamp software. Peak current amplitude was measured as the difference between current before agonist application and at the time when inward current was maximal. Data are represented as mean ± S.E.M. values.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Affinity of [¹²⁵I]DAIZAC for the 5-HT₃ Receptor. Specific [¹²⁵I]DAIZAC binding reached equilibrium within 30 min at temperatures of 4-37°C and remained stable at 20°C for 24 h. Tissue from either fresh rat brain or frozen homogenate stored at 70°C for up to 90 days showed no deterioration in binding. Fig. 2, A and B, shows representative saturation curves and Scatchard analysis. Using 4 µM bemesetron to determine non-5-HT₃-specific binding, saturation analysis of [¹²⁵I]DAIZAC binding to whole rat brain homogenate showed a single site with K_D of 0.15 ± 0.01 nM and B_max of 0.64 ± 0.04 fmol/mg wet tissue (Fig. 2A). Nonspecific binding increased linearly with increasing ligand concentration and represented 17% of total counts bound in rat brain membranes when DAIZAC was present at a concentration equal to its K_D value. Specific binding of [¹²⁵I]DAIZAC to NCB-20 cell membranes, as defined by 50 µM bemesetron, showed a single binding site with K_D of 0.140 ± 0.025 nM and B_max of 340 ± 58 fmol/mg protein (Fig. 2B). When [¹²⁵I]DAIZAC was present at a concentration equal to its K_D, nonspecific binding represented <1% of total bound ligand.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Saturation and Scatchard analysis of specific [125I]DAIZAC binding in (A) whole rat brain homogenate and (B) cloned NCB-20 cell membranes gave equilibrium dissociation constants KD of 0.15 and 0.14 nM, respectively. Nonspecific binding was determined by coincubation with 4 and 50 µM bemesetron, respectively.

1

1

1

1

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Association and dissociation of 0.1 nM [125I]DAIZAC binding at 20°C in the rat brain. The apparent association rate constant was kobs of 0.167 ± 0.014 min1. Dissociation rate was determined in a separate experiment by the addition of 10 µM bemesetron, giving the monophasic dissociation rate constant koff of 0.080 ± 0.001 min1. The calculated association rate was kon of 0.87 ± 0.10 min1 nM1 and the derived KD was estimated at 0.093 nM.

Competition Studies with 5-HT₃ Receptor Ligands. A comparison of the ability of previously characterized 5-HT₃ receptor antagonists and agonists to displace [¹²⁵I]DAIZAC or (S)-[¹²⁵I]iodozacopride binding in rat brain and NCB-20 cell membranes is shown in Table 1. Displacement of 0.1 nM [¹²⁵I]DAIZAC binding in whole rat brain membranes by unlabeled DAIZAC gave inhibition constant K_i of 0.138 ± 0.012 nM. The corresponding value in NCB-20 cell membranes was K_i of 0.163 ± 0.011 nM. Hill slopes in either case were not significantly different from unity. Inhibition constants of 5-HT₃ receptor antagonists against [¹²⁵I]DAIZAC were consistent between the two preparations. The affinities of these ligands in displacing [¹²⁵I]DAIZAC and (S)-[¹²⁵I]iodozacopride binding in the rat brain were also similar. Finally, the abilities of 5-HT₃-selective antagonists and agonists to displace [¹²⁵I]DAIZAC binding in rat brain matched their previously reported affinities for the 5-HT₃ receptor with a correlation coefficient of 0.98 (p < .001). It is of note that although antagonists exhibited similar K_i values in the two preparations, with the exception of serotonin itself, full agonists were 4 to 10 times more potent in displacing [¹²⁵I]DAIZAC binding in rat brain than they were in the NCB-20 cell preparation. As seen with other radioligands for the 5-HT₃ receptor, pseudo-Hill slopes of some agonists were substantially >1, consistent with allosteric cooperativity within the drug/receptor complex (Kilpatrick et al., 1990). Hill slopes of antagonists were close to unity (Table 1). Representative curves for displacement of 0.1 nM [¹²⁵I]DAIZAC binding in whole rat brain membranes by 5-HT₃ receptor ligands are shown in Fig. 4.

View larger version (16K): [in this window] [in a new window]   Fig. 4.   Competition curves for 5-HT3 receptor ligands in displacing 0.1 nM [125I]DAIZAC binding in whole rat brain homogenates. Nonspecific binding was determined using 4 µM bemesetron. Hill slopes for antagonists, bemesetron, LY-278584, and (S)-zacopride were 1.0 ± 0.1, whereas those of agonists, such as 2-methyl-5-HT, were 1.6 ± 0.2.

Selectivity of [¹²⁵I]DAIZAC for the 5-HT₃ Receptor. The use of 4 or 50 µM bemesetron or 10 µM unlabeled DAIZAC gave similar results in defining nonspecific binding at the different concentrations of [¹²⁵I]DAIZAC in both NCB 20 cells and rat brain homogenates, consistent with selective binding only to the 5-HT₃ site. In further assessing the selectivity of [¹²⁵I]DAIZAC, antagonists for every other known 5-HT binding site at concentrations sufficient to occupy 95% of their respective receptors (3-30 µM) were unable to displace >3% of specifically bound [¹²⁵I]DAIZAC (Table 2). Similarly, antagonists of numerous other neurotransmitters and uptake carriers were unable to displace [¹²⁵I]DAIZAC (3 nM) to any significant degree (Table 3). The dopamine D₂/D₃ receptor antagonist epidepride (Kessler et al., 1991) displaced [¹²⁵I]DAIZAC binding in rat brain homogenate with a K_i of 433 nM. However, 10 µM DAIZAC was unable to displace 0.05 nM [¹²⁵I]epidepride binding, indicating that the [¹²⁵I]DAIZAC displacement by epidepride represents the affinity of epidepride for the 5-HT₃ receptor and not the affinity of DAIZAC for dopamine D₂ or D₃ receptors.

Distribution of [¹²⁵I]DAIZAC Binding in Rat Brain and Peripheral Tissue. Specific binding of 0.075 nM [¹²⁵I]DAIZAC to membranes prepared from dissected rat brain regions is shown in Fig. 5. Highest binding was seen in temporalimbic cortex and hippocampus. Low levels of binding were seen in striatum and cerebellum. Scatchard analyses of binding demonstrated relatively low forebrain receptor densities even in regions with the highest concentrations of receptors. This distribution is consistent with that of the 5-HT₃ receptor as reported by Laporte et al. (1992). Analysis of specific binding of [¹²⁵I]DAIZAC in entorhinal cortex gave K_D of 0.153 ± 0.004 nM and B_max of 1.75 ± 0.08 fmol/mg tissue (n = 14). A similar analysis in hippocampus gave K_D of 0.130 ± 0.008 nM and B_max of 1.29 ± 0.05 fmol/mg tissue (n = 13). Examination of the specific binding of 0.1 nM [¹²⁵I]DAIZAC in different peripheral tissues found the highest levels of binding in ileum. When the data were normalized for protein concentration in each tissue preparation, no significant specific binding was observed in kidney, liver, spleen, testes, or heart. Scatchard analysis of binding to ileal membranes gave K_D of 0.22 nM and B_max of 3.25 fmol/mg protein.

View larger version (41K): [in this window] [in a new window]   Fig. 5.   Regional specific in vitro binding of 0.075 nM [125I]DAIZAC in the rat cerebellum (CER), thalamus (THA), pons (PON), striatum (STR), medulla oblongata (MED), frontal cortex (FCX), hippocampus (HIP), and entorhinal cortex (ENT), expressed as fmol/mg tissue.

Autoradiography. Autoradiograms of rat brain show the highest specific binding in nucleus solitarius, with intermediate levels of binding in entorhinal cortex and hippocampus (Fig. 6). Analysis of the autoradiograms revealed distinct structures with highly localized [¹²⁵I]DAIZAC binding. In the brainstem, binding was seen in area postrema, nucleus of the spinal tract of the trigeminal nerve, and nucleus solitarius and dorsal motor nucleus of the vagus. In the limbic regions, binding was seen in the cingulate, entorhinal and olfactory cortex, amygdala, lateral dorsal and ventral hippocampus, dental gyrus, and interpeduncular nucleus. Neither striatum nor substantia nigra showed significant binding.

View larger version (94K): [in this window] [in a new window]   Fig. 6.   Autoradiographic localization of binding of 0.07 nM [125I]DAIZAC in 30-µm coronal slices of the rat brain. Images of autoradiograms correspond roughly to sections shown in plates 28, 49, and 61 of Paxinos and Watson (1986). Total binding (left) and binding in the presence of 4 µM bemesetron (right). The autoradiograms demonstrate specific binding in caudal medulla, hippocampus, entorhinal cortex, and amygdala, with minimal 5-HT3 binding in other regions.

Electrophysiology. The effect of DAIZAC on the individual fast inward cation current in individual NCB-20 cells elicited by 5-HT perfusion was examined by whole-cell patch-clamp recordings. Fig. 7 (top) shows the current produced from a single NCB 20 cell while the cell was subjected to 2 µM 5-HT application in the absence or presence of DAIZAC. The antagonist (DAIZAC) was applied in the presence of agonist (5-HT) for 10 s. It can clearly be seen that the inward current produced by 5-HT was reduced in amplitude by 10-s exposure to 0.1 nM DAIZAC and was completely eliminated by 10 nM DAIZAC. The average inhibition was 45.6 ± 8.8% at 0.1 nM DAIZAC (n = 5) and 96.5 ± 2.3% at 10 nM (n = 8). A similar high level of inhibition of current was observed in the presence of 100 nM zacopride (data not shown). Recovery of the full response to 5-HT could be achieved after a 3-min wash with drug-free solution after each 0.1 nM DAIZAC exposure. However, recovery was incomplete even after a 20- to 35-min wash after a brief exposure to 10 or 100 nM DAIZAC (data not shown).

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Effects of DAIZAC on current induced by 2 µM 5-HT in intact NCB-20 cells. Top, currents in the presence of 10 nM, 0.1 nM, and no DAIZAC. Bottom, inhibition curve showing an IC50 of 0.24 nM.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

To effectively identify 5-HT₃ sites that are present in low abundance in the CNS, a ligand must have properties that maximize its detectability. Such a ligand must be radiolabeled with a specific activity sufficiently high to detect the presence of the receptor in these tissues. Given that tritium has a specific activity of 29 Ci/mmol, methyl-labeled tritiated ligands can produce no more than 87 cpm/fmol at maximum specific activity, assuming a counting efficiency of 45%. Because measurable regional densities of 5-HT₃ sites vary from 0.3 to 3 fmol/mg tissue, tritiated ligands can produce no more than 26 to 260 cpm/mg tissue at saturating concentrations, a level too low to reliably detect changes in receptor binding characteristics. By contrast, compounds labeled with ¹²⁵I, an isotope with a specific activity of 2170 Ci/mmol, can provide a maximal signal of 1,150 to 11,500 cpm/mg tissue at saturating concentrations, assuming 80% counting efficiency. Scatchard analyses or displacement studies using radioligands at receptor occupancies on the order of 5% can still maintain specific binding levels sufficient to make reliable estimates of equilibrium and kinetic parameters if nonspecific binding does not interfere.

We previously demonstrated that [¹²⁵I]DAIZAC binds specifically to 5-HT₃ receptors in rat CNS homogenates with an affinity 3 to 20 times higher than any previously reported radioiodinated ligand (Mason et al., 1996) and an order of magnitude higher than the only commercially available radioiodinated ligand, (S)-[¹²⁵I]iodozacopride (Hewlett et al., 1997). The present results extend these findings, demonstrating saturable, specific binding to 5-HT₃ sites in NCB 20 cell membranes. Displacement of [¹²⁵I]DAIZAC by 5-HT₃ antagonists was similar in both preparations; however, with the exception of serotonin, full agonists had significantly lower K_i values at receptors from rat brain tissue homogenates than at mouse-derived 5-HT₃ binding sites in NCB-20 cell membranes. This latter difference may be due to species differences. Such intermurine differences in agonist binding have been reported previously (Bonhaus et al., 1993). Equilibrium constants (K_D = 0.15 and 0.14 nM) are in good agreement in the two preparations and are not different than the value obtained through kinetic studies (K_D = 0.09 nM) in rat brain membranes. The kinetic analyses revealed that the high affinity of DAIZAC is primarily due to slow dissociation from the receptor, relative to other 5-HT₃ ligands. The high affinity of DAIZAC at the 5-HT₃ receptors makes it one of the more potent 5-HT₃ ligands presently available.

High specific activity and high affinity alone are not sufficient to identify receptors present in low abundance. High nonspecific binding can obscure binding to 5-HT₃ sites both in vitro and in vivo. Earlier radioligands for 5-HT₃ sites such as [³H]GR65630 exhibited a considerable degree of nonspecific binding that made it difficult to identify these sites in tissues with low densities of 5-HT₃ receptors (Kilpatrick et al., 1990; Miller et al., 1992). Nonspecific binding of [¹²⁵I]DAIZAC, as determined with unlabeled DAIZAC (3-10 µM) or bemesetron (3-50 µM), was low in both rat CNS and NCB-20 membrane preparations and increased linearly with increasing ligand concentration. Virtually 99% of total binding was specific in the NCB-20 cell membranes (Fig. 2, A and B). When [¹²⁵I]DAIZAC was present in rat brain membranes at a concentration equal to its K_D value (0.15 nM), the signal-to-noise ratio for binding (specific binding/nonspecific binding) was 4.9. By contrast, the signal-to-noise ratio of (S)-[¹²⁵I]iodozacopride at its K_D concentration (1.3 nM) was 0.65, less than one seventh of the ratio for [¹²⁵I]DAIZAC in the same tissue (Hewlett et al., 1997). The high nonspecific binding of (S)-[¹²⁵I]iodozacopride is due in part to its lower affinity for the 5-HT₃ site, necessitating higher concentrations of the radioligand in the preparation. (S)-[¹²⁵I]Iodozacopride additionally has a greater tendency to associate nonspecifically with the membrane preparation than does DAIZAC (~1.6 times), even when used at the same concentration (Hewlett et al., 1997). The more favorable signal-to-noise ratio of [¹²⁵I]DAIZAC enhances the use of this ligand in assessing changes in 5-HT₃ binding characteristics in CNS tissues.

Ligands used to identify the 5-HT₃ receptors also must have high selectivity for this site. This is particularly important for substituted benzamides such as iodozacopride and DAIZAC in relation to the 5-HT₄ and dopamine D₂ and D₃ binding sites. Ligands in this class are known to bind to these sites (Bockaert et al., 1990; Schmidt and Peroutka, 1990; Kessler et al., 1991), which are present in some brain regions at densities nearly 20 times the highest forebrain densities of 5-HT₃ sites (Barnes et al., 1990; Kessler et al., 1991; Laporte et al., 1992; Waeber et al., 1994). Even a ligand with 100-fold selectivity for 5-HT₃ receptors over D₂ receptors, for example, would predominantly label D₂ binding sites in striatum, confounding the interpretation of binding and displacement studies in this tissue. Although numerous 5-HT₃ receptor antagonists have been described in the literature, few have had the selectivity necessary for regional binding studies. Ondansetron, tropisetron, and their derivatives have shown binding to  or 5-HT₄ sites, or both (Kilpatrick et al., 1990; van Wijngaarden et al., 1993; Schiavi et al., 1994). Substituted benzamides such as metoclopramide and (S)-zacopride also bind to 5-HT₄ sites (Schiavi et al., 1994), where they appear to act as partial agonists (Bockaert et al., 1990). Additionally, [¹²⁵I]iodozacopride appears to bind to a 5-HT-sensitive receptor in the rat ileum different than the 5-HT₃ receptor, tentatively identified as the 5-HT_1P receptor (Wade et al., 1991; Ge et al., 1995).

In contrast, DAIZAC exhibited significant selectivity for 5-HT₃ receptors over every other CNS receptor tested. This was demonstrated by several methods. First, the structurally dissimilar 5-HT₃ receptor antagonist bemesetron (K_i = 13 nM; Fozard, 1984) was used to define nonspecific binding. This ligand has high selectivity for the 5-HT₃ receptor, having 3 orders of magnitude in selectivity for the 5-HT₃ over any other known site (Meyerson and Park, 1991; Schiavi et al., 1994). Its highest non-5-HT₃ affinities are for the 5-HT₄ receptor (K_i = 1690 nM; Schiavi and Park, 1994) and the 5-HT transporter site (K_i = 1970 nM; Meyerson et al., 1991). Binding of [¹²⁵I]DAIZAC was inhibited to the same extent by 4 and 50 µM concentrations of both bemesetron and unlabeled DAIZAC in NCB-20 cell and rat brain membranes preparations. Even in whole rat brain at concentrations of [¹²⁵I]DAIZAC as high as 10 times its K_D value (1.5 nM), bemesetron (4 µM) produced displacement equal to that seen with 10 µM unlabeled DAIZAC, indicating that [¹²⁵I]DAIZAC is unable label sites other than those binding bemesetron, including the non-5-HT₃ "benzac" binding site previously found to bind other substituted benzamide 5-HT₃ receptor antagonists (Hewlett et al., 1998a, 1998b). Additionally, it should be noted that unlike (S)-[¹²⁵I]iodozacopride, [¹²⁵I]DAIZAC binding was displaced completely by bemesetron in the rat ileum, suggesting that [¹²⁵I]DAIZAC does not bind to 5-HT_1P receptors.

Specificity was also demonstrated using unlabeled ligands with known affinities at different CNS sites to displace the binding of [¹²⁵I]DAIZAC. Antagonists present at a sufficient concentration to occupy >95% of other known serotonergic CNS binding sites (including the serotonin uptake transporter and 5-HT₄ binding site) were ineffective in displacing [¹²⁵I]DAIZAC binding in rat brain membranes, even when the radioligand was present at concentrations 20 times its K_D value for the 5-HT₃ site. Because other serotonergic binding sites are present in the rat CNS at densities up to 40 times higher than highest forebrain densities of the 5-HT₃ receptor (Mellerup and Plenge, 1986), if DAIZAC had appreciable affinity for such a site, binding of [¹²⁵I]DAIZAC to that site would be reduced in the presence of unlabeled ligand. For example, if [¹²⁵I]DAIZAC bound to a second site with an affinity of 15 nM, 1/100th of its K_D value at the 5-HT₃ site, and that site was present in the preparation at a density equal to that of the 5-HT₃ receptor, occupancy theory would predict that 14% of [¹²⁵I]DAIZAC-specific counts would be bound to that site. Thus, the absence of an appreciable reduction in binding in the presence of unlabeled ligand implies that the affinity of [¹²⁵I]DAIZAC for the other serotonergic sites would have to be at least 2 orders of magnitude less than its affinity for the 5-HT₃ site. A similar argument can be made for nonserotonergic CNS receptors and uptake sites where a displacement of <8% displacement was observed. It is important to note that each of the competing ligands used has weak affinity for the 5-HT₃ receptor. Thus, any displacement of [¹²⁵I]DAIZAC by these ligands could represent displacement of binding to the 5-HT₃ site.

Finally, selectivity was assessed using unlabeled DAIZAC to displace binding of selective serotonergic and nonserotonergic radioligands in preparations enriched in their respective receptor sites. No receptor assessed in this manner showed an affinity for DAIZAC within 3 orders of magnitude of the affinity of DAIZAC for the 5-HT₃ receptor. The selectivity ratio of DAIZAC for the 5-HT₃ receptor over the muscarinic receptor (the non-5-HT₃ receptor at which it was most potent) was >2800, greater than that for any other 5-HT₃ ligand yet reported (Gozlan, 1997). DAIZAC showed a selectivity for 5-HT₃ over 5-HT₄ receptors >120 times that of (S)-zacopride and >200 times that of (S)-iodozacopride. The selectivity ratio for 5-HT₃ over every other serotonergic site tested was >4 orders of magnitude.

There are several reasons for the high receptor selectivity of DAIZAC. DAIZAC is a structural analog of zacopride, a conformationally restricted substituted benzamide with high affinity for 5-HT₃ sites (Barnes et al., 1990). Although substituted benzamides may have affinities for 5-HT₄ receptors and the dopaminergic D₂ receptor family, unlike other 5-HT₃ ligands in this class (e.g., zacopride and iodozacopride), DAIZAC does not have an amino group in the aromatic 4 position, a substituent that appears to enhance 5-HT₄ receptor affinity relative to 5-HT₃ receptor affinity (Eglen et al., 1994; Gaster and King, 1997). Additionally, molecular modeling of ondansetron and tropisetron has suggested that the 5-HT₃ receptor pharmacophore contains two electrostatic features: a hydrogen bond acceptor and the access for a planar lipophilic ring system (Rizzi et al., 1990; Gaster and King, 1997). By using the 3-quinuclidinyl group as the amine moiety, access of DAIZAC for the recognition site of the receptor is restricted to a perpendicular direction relative to the benzamide plane (Schmidt and Peroutka, 1990), effectively eliminating affinity of DAIZAC for the dopamine D₂ and D₃ receptors (de Paulis et al., 1997).

The use of [¹²⁵I]DAIZAC in assessing regional 5-HT₃ receptor distribution was demonstrated in rat membrane homogenates. [¹²⁵I]DAIZAC specific binding was differentially distributed, with the highest specific binding levels in posterior cortex and hippocampus, intermediate binding levels in medial and frontal cortex, and the lowest binding levels in striatum and cerebellum. Autoradiographic studies also demonstrated the usefulness of this ligand. The highest specific binding level was seen in caudal sections of medulla oblongata (nucleus solitarius), with intermediate levels of binding in entorhinal cortex, amygdala, and hippocampus. Although 5-HT₃ receptors are known to modulate dopamine release (Gozlan, 1997), only low levels of binding were seen in striatum, substantia nigra, and ventral tegmentum (Fig. 6, middle), suggesting that 5-HT₃ effects on dopaminergic transmission may be mediated through cortical and limbic receptors. The distribution of [¹²⁵I]DAIZAC binding conforms to results reported by others (Barnes et al., 1990; Laporte et al., 1992), obtained with [³H]zacopride and [¹²⁵I]iodozacopride. Autoradiographic images of [¹²⁵I]DAIZAC binding, however, were obtained at a concentration equal to half its K_D value, using overnight exposures, and were almost completely devoid of nonspecific artifact. These convenient conditions and the high signal-to-noise ratio associated with [¹²⁵I]DAIZAC binding make it an ideal radioligand for use in autoradiography.

This study also has demonstrated that DAIZAC is functionally active in inhibiting the cation inward current in NCB-20 cells. The application of 5-HT to individual cells elicits a fast inward depolarization current that has been shown to be mediated exclusively by 5-HT₃ receptors (Lovinger and White, 1991). DAIZAC produced lasting but reversible inhibition of 5-HT-induced currents in individual cells, reducing the current activated by 5-HT in a concentration-dependent manner, with an IC₅₀ of 0.24 nM. DAIZAC itself did not produce any detectable ion current at concentrations up to 100 nM, demonstrating a lack of any partial agonist activity.

Recovery of 5-HT-induced fast inward current after DAIZAC administration was slow, consistent with the 7.5-min half-life for DAIZAC occupancy seen in binding studies. A similar slow recovery from antagonist block has been observed during experiments examining the effects of nanomolar concentrations of zacopride and tropisetron (Neijt et al., 1988; Lovinger and White, 1991). Estimation of the K_D values of antagonists with long dissociation times using an acute electrophysiological preparation has been problematic. Neijt et al. (1988) found that IC₅₀ values for tropisetron and bemesetron in antagonizing 5-HT-induced currents in N1E-115 cells were roughly one-tenth their corresponding K_i values at the 5-HT₃ receptor. In our experiments, however, the IC₅₀ (0.24 nM) of DAIZAC was in good agreement with the K_D and K_i values derived from saturation and competition studies.

The 5-HT₃ receptor is potentially a significant, but understudied, effector site that could be important in mediating responses to psychotherapeutic drugs, such as the selective serotonin reuptake inhibitors. It is present in limbic structures and cortical association areas and may play an important role in anxiety, impulsivity, and substance abuse (Costall and Naylor, 1992; Gozlan, 1997). Additionally, these sites may be important in processes related to learning and consolidation of memory (Gozlan, 1997). The mechanism by which 5-HT₃ receptors might affect such processes has been difficult to elucidate because of the low abundance of these CNS sites and the lack of suitable ligands. The availability of [¹²⁵I]DAIZAC will make possible studies assessing alterations in regional 5-HT₃ binding characteristics in response to different pharmacological treatments and should enhance our knowledge of mechanisms and circuitry through which serotonin might act via this unusual 5-HT receptor subtype.