Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Signaling through neuronal nAChRs is being recognized increasingly as playing an important role in neurotransmission in the central nervous system (Arneric et al., 1996a; McGehee et al., 1995; Bertrand and Changeux, 1995). Accumulating preclinical and clinical evidence suggests that alterations in either the levels and/or activity of this ligand gated ion channel receptor family may play a role in several CNS disorders including Alzheimer's disease and Parkinson's disease (Williams et al., 1994; Arneric et al., 1996a). Modulation of nAChRs can elicit several functional responses including enhancement of fast excitatory neurotransmission (McGehee et al., 1995), facilitation of neurotransmitter release (Sacaan et al., 1995), the neurogenic control of cerebral blood flow (Linville et al., 1993) and a diversity of behavioral responses including cognitive enhancement (Levin, 1992), neuroprotection (Donnelly-Roberts et al., 1996), anxiolytic (Brioni et al., 1993) and analgesic activity (Sullivan and Bannon, 1996).

Rapid advances in the molecular biology and pharmacology of nAChRs in the past decade have revealed that a diversity of nAChRs may mediate the wide spectrum of behavioral effects of nAChR ligands (McGehee and Role, 1995; Arneric et al., 1996a; Brioni et al., 1996). Eleven gene products (₂-₉; ₂-₄) have been identified to date in brain, sensory systems and autonomic ganglia (Deneris et al., 1991; Sargent, 1993; Elgoyhen et al., 1994). Functional responses can be elicited either in Xenopus oocytes injected with pair-wise combinations of  and  subunits (Deneris et al., 1991; Papke and Heineman, 1993), or in cell lines stably expressing the 42 and 34 subtypes (Gopalakrishnan et al., 1996; Papke and Heinemann, 1993; Whiting et al., 1991; Wong et al., 1995) confirming biochemical observations which suggest that many native nAChRs consist of / heterooligomers. Further, more complex combinations that include 5 as part of the complex (i.e., 354) may also occur (Role and Berg, 1996). The 7, 8 and 9 gene products, however, differ from other members of the nAChR superfamily in that they can form functional receptors when expressed as homo-oligomers in oocytes (Seguela et al., 1993; Gerzanich et al., 1994; Elgoyhen et al., 1994; Briggs et al., 1995) or cell lines (Gopalakrishnan et al., 1995).

Although it is not clear which subunit combinations form nAChRs in situ, the pharmacology of the putative nAChR subtypes and selectivity of known nAChR ligands is beginning to emerge from in vitro heterologous expression studies. For example, studies in transfected cell lines indicate that the high-affinity ()-nicotine binding site in brain corresponds to the 42 subunit combination (Whiting et al., 1991; Flores et al., 1992) whereas the distribution of the 7 subunit largely coincides with the distribution of high-affinity [¹²⁵I]-BgT binding sites in rat brain (Seguela, et al., 1993; Clarke et al., 1985). Furthermore, recent studies demonstrating ()-nicotine to be an antagonist at 9 nAChRs together with the potent inhibitory effects of cytisine on 2-containing subunit combinations indicate that the pharmacological profile of classical nAChR ligands is more complex than previously recognized (Elgoyhen et al., 1994; Papke and Heinemann 1993).

The historical precedent and pharmacology has led to the conventional nomenclature of calling functional combinations of these novel neuronal gene products related to the neuromuscular receptor, i.e., "nicotinic receptors." This nomenclature has carried over to describe compounds that interact with nAChRs as either nicotinic agonists or nicotinic antagonists. An alternative nomenclature that refers to nAChRs as "cholinergic channels" serves not only to highlight mechanistic distinctions versus muscarinic receptors, but also to acknowledge the nonequivalent interactions of nicotine at the various subtypes (see above). The term "cholinergic channel modulators" (ChCMs) then defines the broad class of agents that includes competitive activators, allosteric activators and allosteric facilitators (collectively, cholinergic channel activators), as well as cholinergic channel inhibitors, which may act through any of at least four likely mechanisms: competitive antagonism, noncompetitive (allosteric) inhibition, ion channel blockade or receptor inactivation (e.g., "desensitization"). The term cholinergic channel modulator further emphasizes that it is possible for a compound to possess one set of properties (e.g., activate) at one subtype of nAChR and a different set of properties (e.g., inhibit) at a different subtype or different properties at the same subtype depending on the conditions (e.g., either activate or desensitize depending on concentration of the cholinergic channel modulator).

ChCMs that possess appropriate combinations of modulatory properties at cholinergic channel subtypes may be able to selectively influence central neurotransmission without having the side-effect liabilities associated with ()-nicotine. Therapeutically, ChCMs lacking cardiovascular or CNS side effects associated with ()-nicotine may represent a potential strategy to ameliorate many of the CNS deficits accompanying Alzheimer's disease or related disorders. Accordingly, there has been a flurry of medicinal chemistry efforts in this area targeted at the development of novel and safe ChCMs (Holladay et al., 1995).

The concept that site- and/or subtype-selective modulation of nAChR function is possible has led to the identification and characterization of ABT-418 [(S)-3-methyl-5-(1-methyl-2-pyrrolidinyl) isoxazole] and SIB 1508Y [(S)-5-ethynylnicotine], analogs of ()-nicotine, and GTS-21 [3-(2,4)-dimethoxybenzylidine], an anabaseine analog, as novel ChCMs (Arneric et al., 1995; Lloyd et al., 1995; Meyer et al., 1994). Relative to ()-nicotine, ABT-418 displays selectivity toward 42 compared with the 7 nAChR. Further, ABT-418 is less potent and efficacious than ()-nicotine to activate human ganglionic nAChRs (Arneric et al., 1995). In contrast, GTS-21 is a weak 7 agonist which may display greater efficacy at the 7 nAChR than the 42 subtype (deFiebre et al., 1995; Briggs et al., 1997). Both compounds demonstrate cognitive enhancing activities in rodents and primates and are neuroprotective in in vitro models of excitoxicity (Arneric et al., 1995; Martin et al., 1994; Woodruff-Pak et al., 1994). Acute administration of ABT-418 has been shown to improve cognitive performance in patients with Alzheimer's disease (Newhouse et al., in press, 1997). However, as has been observed with many ChCMs, both ABT-418 and GTS-21 are poorly (<10%) bioavailable in primates (Arneric et al., 1995; Briggs et al., 1997).

ABT-089 (fig. 1), a ChCM derived from a series of pyridyl ether compounds (Lin et al., 1997), has high oral bioavailability, excellent safety and behavioral efficacy comparable with ABT-418 and GTS-21 (Arneric et al., 1996b; Decker et al., 1997, companion paper). In the present and accompanying paper, the initial pharmacological characterization of ABT-089 is described. Herein, ABT-089 is shown to potently and selectively interact with neuronal nAChRs. In in vitro preparations that differentially express nAChR subtypes, ABT-089 is shown to display a complex profile having agonist, partial agonist and inhibitory activities depending on the neuronal nAChR subtype with which it interacts. Additionally, after acute and subacute exposure, ABT-089 is shown to be neuroprotective against an excitotoxic insult. In the accompanying paper, ABT-089 is shown to display cognitive enhancing activity in rodents and primates (Decker et al., 1997, companion paper).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Structure of ABT-089.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials

ABT-089 [2-methyl-3-(2-(S)-pyrrolidinylmethoxy)pyridine dihydrochloride], and its enantiomer A-94224 [2-methyl-3-(2-(R)-pyrrolidinylmethoxy)pyridine dihydrochloride] were synthesized at Abbott Laboratories as described (Lin et al., 1997). DHE and MLA were purchased from Research Biochemicals International (Natick, MA). ()-Nicotine (hydrogen tartrate salt), morphine sulfate, atropine and mecamylamine hydrochloride were obtained from Sigma Chemical Co. (St. Louis, MO). All radioligands were obtained from DuPont-NEN (Boston, MA). -BgT was obtained from Biotoxins Inc. (Miami, FL).

Cells of the IMR-32 human neuroblastoma and TE 671 human medulloblastoma clonal cell lines (ATCC, Rockville, MD) were maintained in a log phase of growth according to established procedures (Lukas, 1993). Cell lines stably expressing the human 42 and 7 subtypes (K177 and K28, respectively) were maintained as described (Gopalakrishnan et al., 1995, 1996).

Animals were treated according to a protocol approved by Abbott's Institutional Animal Care and Use Committee.

Membrane Preparation

Rat cerebral cortical membranes were prepared from male Sprague-Dawley rats as described by Enna and Snyder (1977) with some modifications as described (Sullivan et al., 1994). Brains were rapidly removed after decapitation, homogenized in 15 volumes of 0.32 M sucrose and centrifuged at 1000 × g for 10 min at 4°C. The supernatants were removed and centrifuged at 20,000 × g for 20 min at 4°C. The resultant P₂ pellets were homogenized with a Polytron (10 s, setting of 6) in ice-cold water and spun at 8,000 × g for 20 min. The supernatant and loose buffy coat were carefully removed and centrifuged at 40,000 × g for 20 min. The membrane pellet was washed with ice-cold H₂O and recentrifuged at 40,000 × g before storage at 80°C. Before use, the frozen membrane pellets were slowly thawed, washed and resuspended in 30 volumes of assay buffer (composition, 120 mM NaCl, 5 mM KCl, 2 mM CaCl₂, 2 mM MgCl₂ and 50 mM Tris-Cl, pH 7.2, 4°C).

Confluent K177 cells stably expressing the human 42 subunit combination were rinsed with ice-cold binding buffer (composition, mM: Tris HCl, 50; NaCl, 120; KCl, 5; MgCl₂, 1 and CaCl₂, 2.5; pH 7.4 at 4°C), mechanically disaggregated and homogenized using a polytron for 10 s. The homogenate was centrifuged at 45,000 × g for 20 min at 4°C, and the pellet was resuspended in ice-cold buffer at a concentration of 40 to 50 µg protein. Confluent K28 cells stably expressing the human 7 subunit were rinsed with ice-cold binding buffer (composition: 118 mM NaCl, 4.8 mM KCl, 2.5 mM CaCl₂, 1.2 mM MgSO₄, 20 mM HEPES, pH 7.5), mechanically disaggregated and homogenized with a polytron for 10 s. The homogenate was centrifuged at 45,000 × g for 20 min at 4°C, and the pellet was resuspended in ice-cold buffer at a concentration of 40 to 50 µg protein.

[³H]()-Cytisine binding. Binding conditions were as described previously (Anderson et al., 1995; Gopalakrishnan et al., 1996). Samples containing 20 to 200 µg of protein, 0.7 nM [³H]()-cytisine (30 Ci/mmol) and the indicated concentrations of drug were incubated in a final volume of 500 µl for 75 min at 4°C in triplicate. Nonspecific binding was determined in the presence of 10 µM ()-nicotine.

[¹²⁵I]-BgT binding. [¹²⁵I]-BgT binding was determined in membranes prepared from rat brain and K28 cell membranes and from Torpedo californica electroplax.

Additional receptor selectivity binding studies. To further assess the selectivity of ABT-089 as an nAChR ligand, the compound was evaluated in a PROFILE receptor binding selectivity screen by NovaScreen (Oceanix, Hanover, MD) by standard receptor binding protocols that are documented in table 2. ABT-089 was tested in 45 binding assays for several peptides, channel proteins, peptide factors, reuptake sites, second messenger systems and neurotransmitters. ABT-089 was evaluated at three concentrations (1, 100 and 10,000 nM) in duplicate.

Data analysis. In competition experiments, the drug concentration producing 50% inhibition (IC₅₀) of radioligand binding and the Hill coefficient (n_H) were determined from plots of log (B₀  B)/B versus log (concentration of drug), where B₀ and B are specific binding in the absence and presence of competitor, respectively, with a four-parameter logistics program in RS/1 (Bolt, Beranek and Newman Inc. Cambridge, MA). Inhibition constant (K_i) values were determined with the Cheng-Prusoff equation (Cheng and Prusoff, 1972).

⁸⁶Rb⁺ Efflux from K177, IMR 32 and TE 671 Cells

Cells of the IMR-32 human neuroblastoma clonal cell line (ATCC, Rockville, MD) and the TE 671 human medulloblastoma cell line were maintained in a log phase of growth according to established procedures (Lukas, 1993). K177 cells stably expressing the human 42 nAChR subunit combination were maintained as described (Gopalakrishnan et al., 1996). Experimental cells were seeded at a density of 500,000 cells/ml into a 24-well tissue culture dish. Plated cells were allowed to proliferate for at least 48 h before loading with 2 µCi/ml of ⁸⁶Rb⁺ (35 Ci/mmol) overnight at 37°C. The ⁸⁶Rb⁺ efflux assays were performed according to previously published protocols (Sullivan et al., 1994; Gopalakrishnan et al., 1995) except serum-free DMEM was used during the ⁸⁶Rb⁺ loading, rinsing and agonist-induced efflux steps. Data were analyzed by a two-way ANOVA with StatView II (Abacus Concepts, Inc., Berkeley, CA). The criterion of statistical significance was P < .05.

⁸⁶Rb⁺ Efflux from Mouse Thalamic Synaptosomes

The ability of ABT-089 and ()-nicotine to activate ion channels was investigated by measuring efflux of ⁸⁶Rb⁺ from mouse thalamus with a slight modification of the method of Marks et al. (1993) as described previously (Arneric et al., 1994).

A P₂ fraction equivalent to two thalami was incubated for 30 min at 21°C in 35 ml of perfusion buffer containing 4 mCi of ⁸⁶Rb⁺ (35 Ci/mmol). At the end of the incubation period, tissue was harvested and separated from the incubation medium by filtration onto 6-mm diameter glass fiber filters (Type GC50, Microfiltration Systems, Dublin, CA) under gentle vacuum (0.2 atm) followed by three washes at room temperature with perfusion buffer. The filter containing the ⁸⁶Rb⁺ loaded synaptosomes was placed on a 13-mm glass fiber filter (Type GC50, Microfiltration Systems, Dublin CA) and perfused continuously at 21°C. After an initial average wash period of 8 min, fractions were collected every 30 s by use of a Retriver II fraction collector (ISCO, Inc., Lexana, KS). Exposure to ABT-089 and ()-nicotine usually occurred 3 min into a 10-min collection period. In any experiment, five concentrations of each ligand were tested and the tissue on each filter stimulated only once. ()-Nicotine (10 µM) was included in each experiment as control to normalize values between experiments. Radioactivity was measured with a Packard Auto-Gamma counter (Packard, Naperville, IL) and the magnitude of the ⁸⁶Rb⁺ response amplitude calculated by determining the increase in radioactivity above the base line after stimulation of the tissue. The average base line underlying the peak was calculated by averaging the radioactivity present in the tubes immediately before and after the peak. Peak size was determined by subtracting the average base-line value from each fraction in the peak. To correct for differences in total tissue content and base-line release, the response was normalized by dividing by the amount of ⁸⁶Rb⁺ present in the tissue at the time of stimulation. EC₅₀ values and the maximum response obtained for stimulation of ⁸⁶Rb⁺ efflux were calculated by use of Inplot^TM (Graphpad, San Diego, CA). Data were analyzed by a two-way ANOVA with use of StatView II (Abacus Concepts, Inc., Berkeley, CA). The criterion of statistical significance was P < .05.

Channel Currents in Xenopus Oocytes Expressing the Human 7 nAChR

The ability of ABT-089 to activate the human 7 nAChR expressed in Xenopus laevis oocytes was determined as described previously (Briggs et al., 1995). Recordings were made using a two-electrode voltage clamp at a holding potential of 60 mV. Experiments were done in modified Barth's solution (90 mM NaCl, 1 mM KCl, 0.66 mM NaNO₃, 0.74 mM CaCl₂, 0.82 mM MgCl₂, 2.4 mM NaHCO₃, 2.5 mM sodium pyruvate and 10 mM Na-HEPES buffer, pH 7.55) containing 10 mM BaCl₂ in place of CaCl₂ and MgCl₂ to prevent a secondary activation of Ca²⁺-dependent Cl current by Ca²⁺ influx through the 7 nicotinic channels. The duration of agonist exposure generally was 1.25 to 2.5 s, but was increased up to 20 s to elicit a plateau in the slower responses. The interval between agonist applications was 5 min. Response stability was assessed through multiple applications. Antagonists were superfused in the bathing solution for 3 to 8 min before testing the response and were applied with agonist so that the only change was the addition of agonist to the superfusion. Current responses were quantified by measuring the peak amplitude of the response relative to base-line holding current and were normalized to the average response to 100 µM ()-nicotine determined in the same oocyte to account for variability in receptor expression among oocytes.

Striatal [³H]Dopamine Release

nAChR-evoked release of [ring-2,5,6-³H]dopamine (24.4 Ci/mmol) was measured in superfused rat striatal slices. Striata were dissected from two male Sprague-Dawley rats per experiment and sliced 0.35 × 0.25 mm by a McIlwain Tissue Chopper (Brinkman Instrument Co., Westbury, NY). After two washes with Krebs-HEPES buffer (137 mM NaCl, 4.7 mM KCl, 1 mM MgSO₄, 2.5 mM CaCl₂, 1.25 mM NaH₂PO₄, 10 mM glucose, 15 mM HEPES-NaOH, pH 7.4, containing 10 µM pargyline and 10 µM ascorbic acid), slices were preincubated for 10 min at 37°C under 95%/5% O₂/CO₂. After replacing the buffer, slices were labeled with 100 nM [³H]dopamine for 25 min in Krebs-HEPES at 37°C. Aliquots of slices were placed in 18 superfusion chambers of a Brandel SP2000 superfusion apparatus (Brandel, Gaithersberg, MD). After 47 min of washout, slices were exposed to agonist for 4 min. Antagonists when present were introduced 4 min before and during agonist exposure. Collected fractions were counted in 5 ml of Ecolume. Tissue was recovered from superfusion chambers, solubilized with 1 ml of Solvable (DuPont-NEN) and counted in 15 ml of Ecolume.

Fractional release of [³H]dopamine was calculated from radioactivity above base line as a fraction of total radioactivity. Relative potencies were calculated by the release evoked by 100 nM ()-nicotine as a standard. EC₅₀ values were determined by nonlinear least squares regression analysis with Inplot.

Hippocampal [³H]ACh Release

nAChR-evoked release of [³H]ACh was measured in superfused rat hippocampal synaptosomes. The F4 synaptosomal fraction was washed twice in Krebs/bicarbonate buffer (NaCl, 118.5 mM; NaHCO₃, 24.9 mM; KCl, KH₂PO₄, 1.2 mM; CaCl₂, 2.5 mM; MgSO₄, 2.5 mM; glucose, 10 mM gassed with 95% O₂/5% CO₂, to give pH 7.4), and resuspended to a protein concentration of 1 mg/ml. The synaptosomes were loaded with [³H]choline (Amersham International, Buckinghamshire, England) by incubation for 30 min with 0.8 µM [³H] choline (diluted with unlabeled choline; specific activity, 2 Ci/mmol). Aliquots (150 µl) were loaded into perfusion chambers of a Brandell superfusion apparatus (Brandel, Gaithersgerg, MD), and perfused with Kreb's buffer at 37°C, flow rate 0.25 ml/min. Three-minute fractions of perfusate were collected and counted for radioactivity. Agonists were administered in Kreb's buffer as 20-s pulses, separated from the bulk flow of the buffer by 10-s air bubbles. Dose-response curves were determined by comparing up to six different agonist concentrations in parallel in a single experiment. A standard response was provided by challenging another parallel chamber with 5 µM ()-nicotine.

Evoked release of [³H]ACh from rat hippocampal synaptosomes was measured as the area under the peak above basal release. This was converted to picomoles of [³H]ACh released per mg protein by reference to the specific activity of the [³H]choline used to load the synaptosomes and the amount of tissue loaded into the perfusion chambers. Results were normalized for variations in [³H]choline uptake between experiments: mean uptake was 99.8 ± 8.3 pmol [³H]choline/mg protein/30 min (mean ± S.E.M. for 24 experiments) and subsequently was normalized with reference to ()-nicotine-evoked [³H]ACh release. EC₅₀ values were determined by nonlinear least squares regression analysis with Inplot (Graphpad, San Diego, CA).

Neurotoxicity Studies

Primary cortical cultures were prepared from Sprague-Dawley rats (Charles River; Wilmington, MA) at day 18 of gestation as described previously (Donnelly-Roberts et al., 1996). Dissected cortices were placed on ice in Hanks' Balanced Salt Solution (Gibco/BRL; Gaithersburg, MD) and meninges and blood vessels removed. The cortex was disaggregated by mechanical trituration through a fire-narrowed Pasteur pipette, then plated onto poly-L-lysine coated 96-well culture dishes at a density of about 50,000 cells per well in DMEM/10% FCS (reduced to 1% FCS 24 h after plating)/33 mM glucose/2 mM glutamine/50 U/ml pen:strep/B27 supplement. Cultures were maintained at 36°C in a humidified atmosphere of 10% CO₂. The IMR 32 cells were maintained in proliferative growth phase following routine protocols (Lukas, 1993). Before use, the cells (50-70% confluency) were differentiated with 100 ng/ml nerve growth factor in 1% FCS/DMEM for 3 days before use.

After 7 to14 days in vitro, cells were pretreated with test compound diluted in DMEM/N2 supplement (Gibco/BRL) for 2 h. This pretreatment solution was replaced by Hanks' Balanced Salt Solution (without magnesium, but containing 3 mM calcium chloride) containing L-GLU (300 µM) and coapplied with the test compound for an additional 15 min. This compound/glutamate solution was removed and replaced with fresh DMEM/N2 supplement for 24 h. Neuronal damage was assessed by measuring the levels of the cytosolic enzyme LDH released into the medium by the damaged cells using a Cytotox 96 assay kit (Promega; Madison, WI). Basal LDH release was typically between 8 and 10% of the LDH released after lysis of the cells with 0.8% Triton X-100, whereas GLU treatment typically resulted in a 3- to 4-fold increase in release over basal levels. The final levels of LDH release, both from control and treated cells, varied from plate to plate because of variability in cell density. Therefore, to facilitate plate-to-plate comparison all values were normalized to the 300 µM GLU-induced maximal LDH release (assigned 100%). Data were analyzed by a two-way ANOVA with StatView II (Abacus Concepts, Inc., Berkeley, CA). The criterion of statistical significance was P < .05.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Receptor Binding

[³H]()-Cytisine has been shown to bind with high affinity to the 42 subtype of nAChRs, a major subtype in brain (Flores et al., 1992). ABT-089 displaced [³H]()-cytisine binding to the human 42 subunit combination stably expressed in the K177 cell line (Gopalakrishnan et al., 1996) in a concentration-dependent manner with a K_i value of 16 ± 2 nM and a Hill coefficient value of 0.99 ± 0.04 (n = 3; table 1). The R-enantiomer of ABT-089, A-94224, displaced binding to this subtype with a K_i value of 35 ± 5 nM (n = 3). ABT-089 and A-94224 displaced [³H]()-cytisine binding to rat brain with K_i values of 17 ± 3 nM and 39 ± 6 nM, respectively. ()-Nicotine displaced [³H]()-cytisine binding to K177 cells and rat brain with K_i values of 1.0 ± 0.1 nM and 1.0 ± 0.1 nM, respectively.

View this table: [in this window] [in a new window]   TABLE 1 Cholinergic binding properties of ABT-089, A-94224 and ()-nicotine Receptor binding studies were performed as described under "Materials and Methods." Results are means ± S.E.M of three to five separate determinations. The Hill coefficients were not significantly different from unity in all cases (data not shown).

In contrast to its activity at the 42 nAChR subtype, ABT-089 was >500-fold less potent (K_i  10,000 nM, n = 3) in displacing [¹²⁵I]-BgT binding from the -BgT-sensitive nAChR subtype present in rat brain and the [¹²⁵I]-BgT binding site present on the human 7 nAChR subtype (K_i > 10,000 nM, n = 3) (table 1). A-94224, the R-enantiomer of ABT-089, also displayed very weak affinity for the rat brain and human 7 [¹²⁵I]-BgT binding site (K_i > 10,000 nM, n = 3). ()-Nicotine displaced [¹²⁵I]-BgT binding from rat brain and human 7 with K_i values of 6000 ± 876 nM and 2000 ± 178 nM, respectively. ABT-089 and ()-nicotine were also weak inhibitors (K_i > 1000 nM) of the binding of [¹²⁵I]-BgT to the 11 nAChR subtype found on Torpedo electroplax membranes (table 1).

ABT-089 was also examined in 45 other receptor binding and enzyme activity assays (table 2) and showed negligible affinity (K_i > 10 µM) for muscarinic, 5-HT₃ and the benzodiazepine receptors as well as other members of the ligand-gated ion channel superfamily, including -aminobutyric acid_A, benzodiazepine, N-methyl-D-aspartate, MK 801, quisqualate, kainate; L, N and T calcium, chloride and potassium channel proteins; members of G-protein-coupled receptor superfamily, adenosine, alpha-1, alpha-2, beta-1 and beta-2 adrenergic, 5-hydroxytryptamine, bradykinin, endothelin, neuropeptide Y, opioid, vasoactive intestinal peptide; choline, norepinephrine, serotonin and dopamine uptake sites; and did not inhibit the activity of acetylcholinesterase, protein kinase C or monoamine oxidase A and B (table 2).

Ion Flux Studies

Mouse thalamus. nAChR-mediated ⁸⁶Rb⁺ efflux from mouse thalamic synaptosomes has been proposed to reflect an activation of the 42 subtype (Marks et al., 1993). ABT-089 was less potent (EC₅₀ = 5 ± 2 µM; n = 3), and significantly (P < .05) less efficacious (34% of ()-nicotine) than ()-nicotine (EC₅₀ = 1 ± 0.4 µM) in evoking ⁸⁶Rb⁺ efflux from mouse thalamic synaptosomes, an effect that was sensitive to the noncompetitive nAChR antagonist, mecamylamine (10 µM) (fig. 2).

View larger version (17K): [in this window] [in a new window]   Fig. 2.   Effect of ABT-089 and ()-nicotine to stimulate [86Rb+] efflux from mouse thalamic synaptosomes. P2 fractions from mouse thalamus that had been loaded with [86Rb+] were exposed for 1 min to the concentrations of compound indicated. Values are the mean ± S.E.M., n = 3. ABT-089 and ()-nicotine had EC50 values of 5 ± 2 µM and 1 ± 0.4 µM, respectively.

Human 42. ABT-089 did not activate ion ⁸⁶Rb+ efflux from the human 42 subtype at concentrations up to 300 µM (fig. 3A). However, ABT-089 did inhibit ()-nicotine-induced cation efflux with an IC₅₀ value of 176 µM (fig. 3A, inset; n = 3), whereas the R-enantiomer, A-94224, displayed negligible activity as either an agonist or antagonist at concentrations up to 300 µM. In contrast, ()-nicotine (EC₅₀ = 4.6 ± 0.7 µM; n = 5) potently activated cation efflux from this cell line.

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effect of ABT-089, A-94224 and ()-nicotine to stimulate cation (86Rb+) efflux from K177 cells expressing the human 42 subunit combination (A) and IMR 32 cells (B). Cells that had been loaded with 86Rb+ were exposed for 5 min to the concentrations of compound indicated. Values are the mean ± S.E.M., n = 3-5. The insets show the ability of ABT-089 to inhibit ()-nicotine (100 µM)-induced cation efflux from K177 cells (A inset) and IMR 32 cells (B inset). In these experiments, cells were exposed to ABT-089 at the indicated concentrations for 15 min before the addition of ()-nicotine for 5 min and assessment of cation efflux.

Human IMR 32 cells. nAChR-mediated cation efflux from the human neuroblastoma IMR 32 is thought to reflect the activation of a ganglionic nAChR subtype possibly 34 (Lukas, 1993). ABT-089 was significantly (P < .05) less potent (EC₅₀ > 300 µM) and less efficacious (<15%) than ()-nicotine (EC₅₀ = 21 ± 4 µM, n = 3) in activating the efflux of ⁸⁶Rb⁺ through this nAChR subtype (fig. 3B), a finding that agrees with the greatly diminished ability of this agent to elicit depressor/pressor responses in anesthetized dogs (Arneric et al., 1996b). When evaluated for its ability to block nAChR-mediated ion flux in IMR 32 cells, ABT-089 was found to inhibit ()-nicotine-elicited cation efflux with an IC₅₀ of 100 µM (fig. 3B, inset).

Human TE 671 cells. ABT-089 was more potent (EC₅₀ = 30 µM) but less efficacious (60%) than ()-nicotine (EC₅₀ = 180 µM) to stimulate cation efflux from the human medulloblastoma cell line TE 671, effects that were completely blocked by mecamylamine (100 µM) and tubocurare (100 nM) (fig. 4). The R-enantiomer of ABT-089, A-94224, activated cation efflux in this cell line at concentrations in excess of 300 µM. At a concentration of 1 mM, A-94424 was 40% as efficacious as ()-nicotine. The nAChR subunits involved in mediating the effects of ABT-089 in this cell line are unclear at present (see discussion below).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effect of ABT-089, A-94224 and ()-nicotine to stimulate cation (86Rb+) efflux from human TE 671 cells. TE 671 cells that had been loaded with 86Rb+ were exposed for 5 min to the concentrations of compound indicated. ABT-089 and ()-nicotine had EC50 values of 30 ± 5 µM and 180 ± 20 µM, respectively. Values are the mean ± S.E.M., n = 3.

Channel Currents

Human 7. The activity of ABT-089 at the human 7 homo-oligomeric nAChR was determined electrophysiologically in Xensopus laevis oocytes injected with 7 nAChR RNA. In this preparation, ()-nicotine acts as an agonist with an EC₅₀ of 83 µM (Briggs et al., 1995). In contrast, ABT-089 acted as a very weak agonist (fig. 5). At a concentration of 1 mM, the efficacy of ABT-089 was only 1.5 ± 0.2% (n = 9) of that found for ()-nicotine (100 µM). This effect was concentration-dependent, because 100 µM ABT-089 elicited smaller responses (0.6 ± 0.1%, n = 4). The response to 1 mM ABT-089 was blocked reversibly by the 7 nAChR antagonist, MLA (10 nM, not shown). A-94224, the R-enantiomer of ABT-089, was a slightly weaker agonist; 100 µM and 1 mM A-94224 elicited responses of 0.5 ± 0.2% (n = 3) and 1.0 ± 0.2% (n = 3) of the maximal ()-nicotine response, respectively. The response to 1 mM A-94224 was 58 ± 6% as large as the response to 1 mM ABT-089 in the same oocytes (n = 3).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Human 7 nAChR response to ABT-089 compared with ()-nicotine. Recordings from one oocyte expressing human 7 nAChR exemplify the smaller, slower responses to 100 µM and 1 mM ABT-089 compared with the rapid spike-like response to 100 µM ()-nicotine. The inset shows the same traces on an expanded vertical scale.

Neurotransmitter Release Studies

[³H]ACh release. In Alzheimer's disease, the most consistent neurochemical abnormality is the decrease in cholinergic neurotransmission (Coyle et al., 1983). Concentration-response curves for nAChR-mediated enhancement of ACh release from rat hippocampal synaptosomes by ABT-089 and ()-nicotine are shown in figure 6. ABT-089 (EC₅₀ = 3 ± 0.4 µM) displayed a similar potency to ()-nicotine (EC₅₀ = 1.0 ± 0.23 µM) in this assay. Unlike other preparations, ABT-089 was as efficacious as ()-nicotine to stimulate [³H]ACh release.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effect of ABT-089 and ()-nicotine in stimulating the release of [3H]ACh from rat hippocampal synaptosomes. ABT-089 and ()-nicotine had EC50 values to evoke release of [3H]ACh of 3 ± 0.4 µM and 1 ± 0.23 µM, respectively. Values are the mean ± S.E.M., n = 3.

[³H]Dopamine release. In contrast to its effects on nAChR-mediated ACh release, ABT-089 was 30% less efficacious and 25-fold less potent (EC₅₀ = 1.1 ± 0.3 µM) than ()-nicotine (EC₅₀ = 0.04 ± 0.02 µM) in stimulating the release of [³H]dopamine from rat striatal slices (fig. 7), an effect blocked by the competitive nAChR antagonist, DHE (10 µM). The R-enantiomer of ABT-089, A-94224, was 10-fold less potent than ABT-089, but equally efficacious, to stimulate the release of [³H]dopamine.

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Effect of ABT-089, A-94224 and ()-nicotine in stimulating the release of [3H]dopamine from rat striatal slices. ABT-089, A-94224 and ()-nicotine had EC50 values to evoke release of [3H]dopamine of 1.1 ± 0.3 µM, 12 ± 3 µM and 0.05 ± 0.01 µM, respectively. Values are the mean ± S.E.M., n = 3-5.

Neuroprotective Effects of ABT-089

The ability of ABT-089 to protect against glutamate-induced neurotoxicity was investigated in primary cultures of rat cortical cells and in differentiated human IMR 32 cells. L-Glutamate (300 µM) administered to cultures of either rat cortical cells or IMR 32 cells for 15 min elicited a significant (P < .05) increase in the levels of LDH above basal levels (basal, 8-10%; glutamate-treated, 30-35%) assessed 24 h later. Pretreatment of either the IMR 32 or rat cortical cells with ABT-089 protected against the excitotoxic insult in a concentration-dependent manner (fig. 8, A and B). EC₅₀ values for ABT-089 were 3 ± 2 µM (IMR 32 cells) and 10 ± 3 µM (rat cortical cells). The neuroprotection elicited by ABT-089 was time-dependent; maximal effects were observed after a 2-h pretreatment with ABT-089 (data not shown). The R-enantiomer of ABT-089, A-94224, did not protect against the glutamate insult at concentrations up to 100 µM. The protective effects of ABT-089 appeared to be mediated via an interaction with the nAChRs because mecamylamine (100 µM) and -BgT (1 nM) added to the cells 15 min before ABT-089 addition attenuated the protective effects (fig. 9). Mecamylamine and -BgT were not neuroprotective by themselves under these conditions.

View larger version (49K): [in this window] [in a new window]   Fig. 8.   Neuroprotective effects of ABT-089 in differentiated IMR 32 cells (A) and rat cortical cultures (B). ABT-089, at the indicated concentrations, was added to the cells for 2 h before exposure to 300 µM GLU for 15 min. Release of LDH into the extracellular media was determined as described under "Materials and Methods." Values were normalized to maximum LDH released by GLU (100%) and represents the mean ± S.E.M. of six to eight determinants in three separate experiments. EC50 values for ABT-089 were 3 ± 2 µM (IMR 32 cells) and 10 ± 3 µM (rat cortical cells). * P < .05.

View larger version (41K): [in this window] [in a new window]   Fig. 9.   The neuroprotective effects of ABT-089 are mediated via nAChRs. Either media, mecamylamine (MEC; 100 µM) or -BgT (1 nM) were added to the rat cortical cells 15 min before the addition of ABT-089 (10 µM). Two hours later 300 µM glutamate (GLU) was added for 15 min. Release of LDH into the extracellular media was determined as described under "Materials and Methods." Values were normalized to maximum LDH released by GLU (100%) and represents the mean ± S.E.M. of six to eight determinants in three separate experiments. * P < .05.

In behavioral paradigms assessing cognition, the potency and efficacy of ABT-089 is enhanced after subacute treatment (Decker et al., 1997, companion paper). Accordingly, it was of interest to determine whether the potency of ABT-089 to elicit neuroprotection is also enhanced after subacute treatment. Primary rat cortical cells were exposed to ABT-089 (0.01-10 µM) for 7 days before treatment with glutamate (300 µM) for 15 min and assessment of LDH release 24 h later. As shown in figure 10, subacute exposure of rat cortical cells to ABT-089 at concentrations of 0.01 and 0.1 µM elicits significant neuroprotection against glutamate-induced toxicity. In parallel experiments, rat cortical cells were maintained in culture for 7 days and then exposed to ABT-089 at the same concentrations for 2 h. In this instance, no neuroprotection was observed after the acute (2 h) exposure to 0.01 and 0.1 µM ABT-089 (data not shown).

View larger version (38K): [in this window] [in a new window]   Fig. 10.   The neuroprotective effects of ABT-089 are maintained after subacute treatment. Primary rat cortical cells were exposed to ABT-089 at the indicated concentrations for 7 days before treatment with glutamate (300 µM) for 15 min and assessment of LDH release 24 h later. * P < .05.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present study describes the in vitro pharmacological properties of a novel ChCM, ABT-089. In radioligand binding studies, ABT-089 was shown to display selectivity toward the high-affinity ()-cytisine binding site present on the 42 nAChR subtype (K_i = 16 nM) relative to the [¹²⁵I]-BgT binding site present on the 7 (K_i >10,000 nM) and 11 (K_i > 1000 nM) nAChR subtypes. In cation flux and electrophysiological studies ABT-089 displayed a more complex profile than ()-nicotine having agonist, partial agonist and inhibitory activities depending on the nAChR subtype with which it interacts. ABT-089 differentially stimulated neurotransmitter release; the compound displayed a potency and efficacy similar to ()-nicotine to facilitate ACh release but was markedly less potent and less efficacious than ()-nicotine to stimulate DA release. ABT-089 also displayed neuroprotective properties in primary cultures of rodent cortical cells, as well as human neuroblastoma cells, which suggested the potential to prevent the process of neurodegeneration. The differential full agonist/partial agonist profile of ABT-089, as compared with ()-nicotine and ABT-418, illustrates the complexity of nAChR activation and the potential to target responses at subclasses of neuronal and peripheral receptors.

The cloning of rat, chick and human brain cDNAs for nAChRs indicates a diversity of receptor subtypes in mammalian brain (Sargent, 1993; Role and Berg, 1996). The widespread distribution, yet regionally selective combination of 2, 3, 4, 7 and 2 transcripts in the brain suggests that subtypes of neuronal nAChRs may be selectively localized, although the physiological functions that these serve are currently unknown. Nonetheless, the use of binding and functional assays to design compounds that selectively interact with these various nAChRs represents a logical approach to delineating the effects of cholinergic channel modulators. Two major classes of nAChRs have been identified in brain by radioligand binding studies; a high-affinity ()-nicotine binding site that likely corresponds to the 42 subtype and a high-affinity -BgT binding site on the 7 subtype and also on the 8 and 9 nAChR subtypes (Clarke et al., 1985; Elgoyhen et al., 1994; Gerzanich et al., 1994; Seguela et al., 1993). In the present study the interaction of ABT-089 with the high-affinity ()-nicotine and -BgT binding sites in rodent brain and on the human 42 and 7 subtypes stably expressed in mammalian cell lines was investigated. ABT-089 was significantly more potent at the 42 subtype than at the 7 nAChR. Comparison of the radioligand binding data obtained in this study with those recently reported for ABT-418 and GTS-21 (Arneric et al., 1995; Briggs et al., 1997), indicates that ABT-089 (K_i = 16 nM) is equipotent to GTS-21 but is approximately 5-fold less potent than ABT-418 (K_i = 3 nM) at the 42 subtype. At the -BgT-sensitive binding site on the human 7 subtype, ABT-089 is less potent (K_i > 10,000 nM) than either ABT-418 (K_i = 4000 nM) or GTS-21 (K_i = 2000 nM). The lack of stereoselectivity of ABT-089 versus its enantiomer, A-94224, observed in the radioligand binding studies differentiates this agent from nicotine and other classical nAChR ligands. However, as newer ligands for nAChRs appear in the literature, it is becoming apparent that enantioselectivity is not a prerequisite property of ChCMs. For example, the enantiomers of epibatidine display similar potency (~50 pM) in their interaction with the high-affinity-nicotine binding site present on the 42 nAChR (Sullivan and Bannon, 1996).

Functionally, the effects of ABT-089 and ()-nicotine were compared on neurotransmitter release, ion flux and current flow measured electrophysiologically. In contrast to ()-nicotine and ABT-418, which have primarily cholinergic channel activator activity, ABT-089 demonstrates a complex pattern of activity with agonist activities at some subtypes of nAChRs and inhibitory activities at others.

ABT-089 was evaluated in two assay preparations thought to reflect an interaction with the 42 nAChR subtype. The human 42 subunit combination stably expressed in a mammalian cell line (K177) has recently been shown to have pharmacological properties similar to those of the avian 42 subtype and that found in rodent brain (Whiting et al., 1991; Flores et al., 1992; Gopalakrishnan et al., 1996). ABT-089 was found to display negligible agonist activity at this subtype at concentrations up to 300 µM but did weakly inhibit ()-nicotine-evoked cation efflux (IC₅₀ = 176 µM). As noted above, ABT-089 potently interacts with the agonist binding site on the 42 subtype in radioligand binding experiments thought to reflect an interaction with the desensitized state of the receptor (Lippiello et al., 1987). Thus, the very weak functional activity of ABT-089 at this subtype may be suggestive of a preferential interaction with a desensitized nonconducting state of this receptor.

In contrast to its effects on the human 42 subtype, ABT-089 was a relatively potent (EC₅₀ = 5 µM) (34% of ()-nicotine) partial agonist to stimulate cation efflux from mouse thalamic synaptosomes. As noted above, nAChR-mediated stimulation of cation efflux from mouse thalamic synaptosomes has been proposed to reflect the activation of the 42 subtype (Marks et al., 1993). The differences in activity of ABT-089 in these two 42 preparations may reflect species differences in the effects of the ligand or differences in the methodology used to assess cation efflux. Unfortunately, amino acid sequence information is not yet available for the murine 4 subunit preventing analysis of the contribution, if any, of species differences. Alternatively, it maybe that in mouse thalamus ABT-089 is modulating cation efflux via interactions with nAChR subunits in addition to 42 because 3 and 5 are expressed in this region (Wada et al., 1989; Sargent, 1993).

The 7 nAChR will self-assemble to form functional homomeric nAChRs in Xenopus oocytes (Seguela et al., 1993; Briggs et al., 1995), and in a mammalian cell line stably expressing this subunit (Gopalakrishnan et al., 1995). In this preparation ABT-089 acted as a very weak agonist. Indeed, the responses to ABT-089 could appear "negligible" when plotted on the same scale as the control response to ()-nicotine. However, the effect of ABT-089 could be measured when the recorded data were plotted on an expanded scale (see fig. 5). These small responses were blocked by the selective antagonist MLA (10 nM), confirming their being caused by activation of the 7 nAChR.

nAChR-mediated activation of cation efflux in the IMR 32 neuroblastoma cell line may be mediated via a "ganglionic-like" nAChR subtype, most likely containing 3 and 4 nAChR subunits and possibly an 5 subunit (Lukas, 1993). Thus, ChCMs with reduced activity relative to ()-nicotine in this preparation might be expected to have an enhanced cardiovascular safety profile compared with the ()-nicotine. Indeed, it was recently reported that ABT-418 and (±)-epibatidine, ChCMs with reduced and enhanced potency, respectively, relative to ()-nicotine in this preparation display diminished and increased cardiovascular effects, respectively, relative to ()-nicotine at equivalent doses in anesthetized dog (Arneric et al., 1995; Sullivan and Bannon, 1996). In the present study, ABT-089 was found to have negligible activity at this putative human ganglionic subtype. This finding agrees with the greatly diminished ability of this agent to elicit adverse cardiovascular effects in anesthetized dog (Arneric et al., 1996b).

Cells of the human medulloblastoma clonal line TE 671 have been shown to express the human 1, 1,  and  subunits that form high-affinity -BgT binding sites with pharmacological properties similar to the neuromuscular junction subtype (Lukas, 1986). Recently, we have found with use of reverse transcriptase-polymerase chain reaction and radioligand binding techniques that this cell line also expresses human neuronal subunits (L.M. Monteggia, manuscript in preparation). This finding raises the possibility that the 1,  and  subunits may combine with various neuronal nAChR subunits to form subtypes with unique pharmacological properties. ABT-089 is more potent (EC₅₀ = 30 µM) but less efficacious (60%) than ()-nicotine (EC₅₀ = 180 µM) to stimulate cation efflux in this cell line (fig. 4), effects that are blocked by mecamylamine (100 µM). Thus, ABT-089 is more potent than ()-nicotine to activate this human nAChR in contrast to its effects at the human 42, 7 and 3x subtypes. Experiments elucidating the subunit combination responsible for the ABT-089 mediated cation efflux in this human cell line are ongoing and may provide additional insight into the mechanism of action of this compound.

nAChRs have a modulatory role on several neurotransmitter systems including cholinergic, noradrenergic, serotonergic, GABAergic and dopaminergic systems (Wonnacott et al., 1990; Sacaan et al., 1995; Wonnacott, 1997). A decrease in cholinergic transmission, arising from the degeneration of the basal forebrain cholinergic system, is the most consistent neurochemical abnormality in patients with Alzheimer's disease (Coyle et al., 1983). Agents that act to selectively enhance cholinergic transmission have been a major focus of pharmaceutical research for the past decade (Arneric et al., 1996a). In the present study, ABT-089 was found to be a selective modulator of ACh release compared with ()-nicotine. ABT-089 was found to be slightly less potent (EC₅₀ = 3 µM), but equally efficacious compared with ()-nicotine (EC₅₀ = 1 ± 0.23 µM) to stimulate [³H]ACh release from rat hippocampal synaptosomes (fig. 6). Although the subtype(s) mediating this effect are uncertain at present, this finding suggests that ABT-089 is a full agonist at some nAChR subtypes.

In contrast to its effects on nAChR-mediated ACh release, ABT-089 is 30% less efficacious and 25-fold less potent (EC₅₀ = 1.1 µM) than ()-nicotine (EC₅₀ = 0.04 µM) in stimulating the release of [³H]dopamine release from rat striatal slices (fig. 7). It is thought that the addiction liability and locomotor stimulant effects associated with chronic exposure to ()-nicotine are mediated via dopamine release. The decreased potency of ABT-089 to stimulate dopamine release, compared with ()-nicotine, thus suggests it may have less dependence potential.

Neurodegeneration caused by excitotoxic damage has been implicated in the etiology of both Alzheimer's disease and Parkinson's disease. Recent studies suggest that ChCMs can block N-methyl-D-aspartate-receptor-mediated glutamate toxicity in rat cortical cells and in differentiated neuroblastoma cell lines (Akaike et al., 1994; Marin et al., 1994; Donnelly-Roberts et al., 1996). Activation of either rat cortical or differentiated IMR 32 nAChRs by ()-nicotine and ABT-418 afforded in vitro neuroprotection against glutamate toxicity (Donnelly-Roberts et al., 1996). In the present study, ABT-089 was found to protect rat cortical neurons and differentiated IMR 32 cells from an excitotoxic insult (fig 8).

The effects of ABT-089 were time-dependent, with maximal neuroprotection being observed when ABT-089 was administered 2 h before the glutamate insult, which suggests that some early immediate gene response process may be involved in mediating the beneficial effects of the compound. Further, the potency of ABT-089 to elicit neuroprotection was dramatically increased after subacute (7-day) treatment. In a behavioral paradigm assessing cognition, the Morris water maze, the potency and efficacy of ABT-089 is enhanced after subacute treatment (Decker et al., 1997, companion paper). It is noteworthy that the minimally effective neuroprotective concentrations of ABT-089 (0.01-0.1 µM) af