Use of an α3β4 nicotinic acetylcholine receptor subunit concatamer to characterize ganglionic receptor subtypes with specific subunit composition reveals species-specific pharmacologic properties

Abstract

Drug development for nicotinic acetylcholine receptors (nAChR) is challenged by subtype diversity arising from variations in subunit composition. On-target activity for neuronal heteromeric receptors is typically associated with CNS receptors that contain α4 and other subunits, while off-target activity could be associated with ganglionic-type receptors containing α3β4 binding sites and other subunits, including β4, β2, α5, or α3 as a structural subunit in the pentamer. Additional interest in α3 β4 α5-containing receptors arises from genome-wide association studies linking these genes, and a single nucleotide polymorphism (SNP) in α5 in particular, to lung cancer and heavy smoking. While α3 and β4 readily form receptors in expression system such as the Xenopus oocyte, since α5 is not required for function, simple co-expression approaches may under-represent α5-containing receptors. We used a concatamer of human α3 and β4 subunits to form ligand-binding domains, and show that we can force the insertions of alternative structural subunits into the functional pentamers. These α3β4 variants differ in sensitivity to ACh, nicotine, varenicline, and cytisine. Our data indicated lower efficacy for varenicline and cytisine than expected for β4-containing receptors, based on previous studies of rodent receptors. We confirm that these therapeutically important α4 receptor partial agonists may present different autonomic-based side-effect profiles in humans than will be seen in rodent models, with varenicline being more potent for human than rat receptors and cytisine less potent. Our initial characterizations failed to find functional effects of the α5 SNP. However, our data validate this approach for further investigations.

Introduction

Early single-channel studies of the nicotinic acetylcholine receptors of autonomic neurons revealed a rich diversity of channel subtypes (Papke, 1993). Although it is now appreciated that, at least in embryonic ganglia, rapidly desensitizing α7-containing nAChRs are of functional importance, blocking α7 receptors in adult animals generally does not impair ganglionic function, and it is likely that the early single-channel studies only detected an array of more slowly desensitizing heteromeric receptor subtypes. We now know, based on recent studies using knockout animals, that ganglionic receptors are primarily assembled as pentameric complexes containing varying arrangements of α3, β2, β4, and α5 subunits (David et al., 2010).

Although ganglionic blockers were the first drugs used clinically to target neuronal nAChR, most current drug development programs intended to target nAChR in the CNS, either for therapeutics or nicotine dependence, view ganglionic receptors containing α3 in various combinations with β2, β4, and α5 as potential sites for off-target side effects. The human α3-β4-α5 genes are in a cluster at chromosomal location 15q24, and recent genome-wide association studies indicated strong correlations between single nucleotide polymorphisms in the α3-β4-α5 gene cluster and risk for both cancer and nicotine dependence (Chen et al., 2009; Stevens et al., 2008). Nicotine addiction and dependence has been clearly linked to α4* and α6* receptors (Wu and Lukas, 2011), and α5 co-assembly into α4* receptors also promotes high sensitivity to nicotine, suggesting a link between nicotine use and α4β2α5 receptors (Kuryatov et al., 2011). However, recent studies have also demonstrated a link between α3β4α5-containing receptors in the medial habenula and nicotine-related behavior, promoting receptors with combinations of these subunits as an alternative target for the development of smoking cessation drugs (Fowler et al., 2011; Frahm et al., 2011; Gallego et al., 2011; Salas et al., 2009).

It has been shown both in vivo (Grady et al., 2009) and in heterologous expression systems (Boulter et al., 1990; Gerzanich et al., 1998) that α3 will form receptors in various combinations with β2, β4, and α5 subunits. However, α3 and β4 subunits readily form functional receptors without additional subunits, and functional effects of α5 co-expression are much more easily detectable in β2-containing than in β4-containing receptors (Gerzanich et al., 1998). Therefore, since most effectively targeted drug development relies on the use of receptors with known subunit composition, we adopted a strategy previously shown to be useful for controlling the subunit composition of α4* receptors (Zhou et al., 2003), by constructing a concatamer of β4 and α3 (β4−6−α3), suitable for co-expression with monomeric α3, β2, β4, or α5 subunits. The β4−6−α3 construct will provide ligand-binding domains with α3−β4 interfaces, so that co-expressed subunit monomers will, with high likelihood, take the fifth position as a structural subunit in the assembled pentamer.

We provide pharmacological validation of hypothesized subunit compositions and characterize the agonist and partial-agonist profiles of the α3β4 receptor subtypes for ACh, nicotine, and the smoking cessation agents, cytisine and varenicline. Cytisine and varenicline have been proposed to have therapeutic utility through potent partial agonist effects on CNS α4-containing receptors. However, it has been a concern that the reportedly high efficacy of these agents on ganglionic α3-containing receptors might be a source of autonomic side effects. We reevaluate those data and show significant differences from the previously reported data based on the use of rodent receptor subtypes and our current studies based on the use of human receptor clones. Additionally, we used the β4−6−α3 construct to study the D376N variant of α5, specifically associated with smoking and cancer risks.