Recent developments in novel antidepressants targeting α4β2-nicotinic acetylcholine receptors.

Nicotinic acetylcholine receptors (nAChRs) have been investigated for developing drugs that can potentially treat various central nervous system disorders. Considerable evidence supports the hypothesis that modulation of the cholinergic system through activation and/or desensitization/inactivation of nAChR holds promise for the development of new antidepressants. The introductory portion of this Miniperspective discusses the basic pharmacology that underpins the involvement of α4β2-nAChRs in depression, along with the structural features that are essential to ligand recognition by the α4β2-nAChRs. The remainder of this Miniperspective analyzes reported nicotinic ligands in terms of drug design considerations and their potency and selectivity, with a particular focus on compounds exhibiting antidepressant-like effects in preclinical or clinical studies. This Miniperspective aims to provide an in-depth analysis of the potential for using nicotinic ligands in the treatment of depression, which may hold some promise in addressing an unmet clinical need by providing relief from depressive symptoms in refractory patients.

Introduction

Depression is a common and frequently severe psychological condition with a distinct change of mood, characterized by sadness, loss of interest and pleasure, feelings of guilt or low self-worth, disturbed sleep or appetite, and feelings of tiredness and poor concentration, affecting approximately 120 million people worldwide.[1] Numerous therapeutic agents exist for the treatment of depression that target monoamine transporters regulating the uptake of the neurotransmitters dopamine, serotonin, and norepinephrine.[2] However, a considerable proportion of patients respond poorly to these drugs,[3] as demonstrated by the NIMH-funded sequenced treatment alternatives to relieve depression (STAR*D) study conducted between 2001 and 2006, which highlighted the inadequacy of current medications for major depressive disorder (MDD).[4] Therefore, there is still an urgent need for potent pharmacotherapies associated with novel biological mechanisms of action. In this Miniperspective, we review the association between depression and nicotinic acetylcholine receptors (nAChRs), especially the α4β2-nAChR subtype, from the perspective of clinical and preclinical findings. We highlight the most recently developed α4β2-nAChR agonists and antagonists that exhibit antidepressant-like effects in vivo.

The cholinergic hypothesis of depression proposes that hyperactivity of the cholinergic system over that of the adrenergic system leads to depression (Figure 1).[5] Several lines of evidence from rodent and human studies support this hypothesis. Flinders sensitive rats, a line selectively bred for increased cholinergic sensitivity, were found to exhibit several depression-like behaviors,[6,7] and increased acetylcholine (ACh) signaling in the hippocampus was found to promote behaviors in mice related to anxiety and depression.[8] In humans, physostigmine, which potentiates cholinergic transmission by inhibiting acetylcholinesterase (AChE), the enzyme that breaks down ACh, produces depressive-like symptoms in individuals with and without a history of depression.[5] Administration of the nonselective nicotinic antagonist mecamylamine (1) (Scheme 1),[9−11] the partial agonist varenicline (2),[12] or the muscarinic antagonist scopolamine (3)[13,14] demonstrated putative antidepressant-like effects, especially in treatment-resistant patients receiving their regular psychotropic medications such as the selective serotonin reuptake inhibitor (SSRI) citalopram (4).[9] Magnetic resonance imaging studies have shown that the levels of choline, the rate-limiting precursor to endogenous ACh, were elevated in the brains of patients with depression as well as in the frontal cortex of adolescents with depression.[15,16] Additionally, a number of key antidepressants such as SSRIs (fluoxetine, sertraline (5), paroxetine, and citalopram 4), the norepinephrine reuptake inhibitor reboxetine (6), the norepinephrine dopamine reuptake inhibitor bupropion (7), and tricyclics (amitriptyline (8), imipramine (9), and nortriptyline)[17] have all been shown to possess antagonistic activities at nAChRs,[18−20] although, in most cases, drug concentrations achieved in the human brain would not be adequate to affect nAChR functions.[21] Nicotine (10) itself, and some nicotinic agonists or antagonists, can potentiate the antidepressant-like effects of the SSRIs and SNRIs in rodent models,[22,23] likely due to the common end point of reduced function due to receptor desensitization by agonists or antagonism. The development of nAChR ligands to attenuate cholinergic activity to treat depression could conceivably help to treat depressive symptoms in refractory patients.

Figure 1

Role of the cholinergic system in depression. The cholinergic hypothesis of depression postulates a hyperactivity of the cholinergic system over that of the adrenergic system in the brain. Choline (the rate-limiting precursor to endogeneous ACh) crosses the blood–brain barrier to enter the brain and is actively transported into the cholinergic presynaptic terminals by an active uptake mechanism. The neurotransmitter ACh is synthesized from choline and acetyl coenzyme A, catalyzed by the enzyme choline acetyl transferase. ACh is sequestered into secretory vesicles by vesicular ACh transporters. Once released from the presynaptic terminals, ACh can interact with a variety of presynaptic and postsynaptic receptors. Two classes of the cholinergic ACh receptors are muscarinic (G protein-coupled) and nicotinic (ionotropic). Once activated, nAChRs form transient open cationic channels that allow the ions Na+, K+, and Ca2+ to flow across the plasma membrane and induce cellular responses. Prolonged exposure to ACh or nicotinic agonist causes a gradual decrease in the rate of this ionic response, leading to a high affinity, longer-lasting functionally inactive state, referred to as desensitization. ACh has its signal terminated primarily by the enzyme AChE, unlike many other monoaminergic neurotransmitters where reuptake mechanisms predominate.

Scheme 1

Compound Structures

Nicotinic Acetylcholine Receptors and Depression

The two major types of cholinergic receptors are muscarinic ACh receptors (mAChRs) and nAChRs, both of which are widely distributed in the central and peripheral nervous systems.[24] G protein-coupled mAChRs are believed to be involved in mood regulation and AChE-induced depressive behavior.[5,13,14,25] Neuronal nAChRs belong to the ligand-gated ion channel superfamily of neurotransmitter receptors. Varying combinations of nAChR subunits (α1−α10, β1−β4, γ, δ, and ε; α2−α7 and β2−β4 are expressed in the brain) assemble into pentameric ion channels, allowing for diverse pharmacological properties.[26] Each nAChR subunit consists of a large amino-terminal extracellular domain (ECD), a transmembrane domain comprising four α-helices (M1–M4), and a variable cytoplasmic domain between M3 and M4. ACh binding sites are thought to form between the subunit interfaces of the ECD bound by the C-loop containing the face of an α-type subunit and the face of an adjacent subunit. When acutely activated by endogenous ACh or exogenous nicotinic ligands, nAChRs form transient open cationic channels that allow the ions Na+, K+, and Ca2+ to flow across the plasma membrane and induce cellular responses.[27] Prolonged exposure to ACh or nicotinic agonists causes a gradual decrease in the rate of this ionic response, leading to a longer-lasting functionally inactive state through a process referred to as desensitization. nAChRs have been found to contribute to mood control by regulating behavioral reinforcement in the mesolimbic dopamine system, corticotropin releasing factor and function in the hypothalamic–pituitary–adrenal (HPA) axis, circadian rhythms in the suprachiasmatic nucleus, and cytoplasticity in the hippocampus.[11] Given nAChR subtype diversity and their involvement in the modulation of various key neurotransmitter systems, including dopamine, serotonin, norepinephrine, glutamate, and γ-aminobutyric acid (GABA), nicotinic ligands have the potential to treat a multitude of neurological and psychiatric disorders, including depression.[9,28]

The α4β2 heteropentameric and α7 homopentameric subtypes are the two major nAChR subtypes expressed in the brain (Figure 2).[29] Both subtypes are implicated in the mediation of the pharmacological and behavioral effects of compound 10 and in nAChR-mediated modulation of monoamine release[30,31] and are likely involved in the antidepressant effects of nicotinic ligands. Studies investigating the role of nAChRs in depression have focused primarily on the α4β2* (asterisk indicates possible assembly with other subunits) and α7-nAChR subtypes. α4β2*-nAChRs are widely distributed in the neuroanatomic regions implicated in depression, including the thalamus, basal ganglia, striatum, hypothalamus, amygdala, ventral tegmental area (VTA), locus coeruleus, and dorsal raphe nucleus,[9] and α7-nAChRs are highly expressed in the hypothalamus, hippocampus, and cortex.[32,33] α4β2*-nAChRs are thought to regulate the release of monoamine neurotransmitters through action in these areas.[34,35] β2-knockout mice showed decreased immobility in the forced swim test (FST) compared to that of wild-type mice, indicating that the absence of β2-nAChRs-mediated signaling could manifest in an antidepressant-like phenotype in vivo.[36] The antidepressant-like effect of the nAChR antagonist compound 1 was diminished when the β2- or α7-subunits were knocked out.[37] Similarly, the antidepressant-like effects of the nAChR agonist sazetidine-A (11)[38] and the tricyclic antidepressant compound 8(36) were absent in mice lacking the β2-subunit. Additionally, the efficacies of compounds 4 and 6 in the mouse FST were enhanced by agonists at either α4β2*- or α7-nAChRs.[23] These findings suggest the involvement of β2- and α7-receptor subtypes in mediating the antidepressant-like effects of nicotinic ligands.

Figure 2

Selected nAChR subtypes. The high sensitivity (HS) α4β2-nAChR has a presumed α4/β2 subunit ratio of 2:3 and exhibits comparatively high sensitivity to nicotinic agonists, whereas the low sensitivity (LS) α4β2-nAChR, at which nicotinic agonists have lower observed potency, is composed presumably of α4 and β2 subunits in a 3:2 ratio.

Clinical studies provide additional evidence for a relationship between α4β2*-nAChRs and depression. Single photon emission computed tomography (SPECT) using the α4β2*-nAChR specific radioligand [123I]5-I-A-85380 ([123I]12) revealed that the β2*-nAChR availability across all brain regions in depressed patients was lower than that in healthy subjects.[39] Additionally, positron emission tomography using 2-[18F]fluoro-3-(2[S]-2-azetidinylmethoxy)-pyridine showed reduced levels of ligand binding to α4β2*-nAChR in Parkinson’s patients with depressive symptoms.[40] It was recently reported that the clinical action of compound 6 may be at least partially due to its inhibitory action on α4β2-nAChR.[41]

Designing a nicotinic ligand to provide maximal therapeutic efficacy and minimal side effects depends on a ligand’s ability to specifically target the desired combination of nAChR subtypes. Observations that nAChR α6 subunits are not widely distributed in the brain but are most prevalent in midbrain dopaminergic regions in the mammalian CNS suggest their potential involvement in mood control. As such, targeting α6*-nAChR may be indicated.[11,42−44] In the basal ganglia, the VTA and substantia nigra, in particular, α6- and possibly nAChR β3 subunits, are included in α4β2*-nAChRs that appear to have high affinity for nicotinic agonists.[34] α3β4*-nAChR subtypes are expressed at relatively low levels in the brain[45] except for the interpeduncular nucleus, fasciculus retroflexus, and median habenula.[46] However, activation or blockade of α3β4*-nAChR, which are also expressed in the peripheral nervous system, may result in side effects in vivo, including dysregulation of the autonomic nervous system.[47,48] Selective and potent partial agonists of α4β2*-nAChRs, especially those with low affinity for α3β4*-nAChRs, are considered to have higher efficacy and likely fewer side effects in rodent behavioral models, although the involvement of α3β4*-nAChRs in mood control cannot be completely ruled out.[49−51] We also point out that because of their roles as “accessory” subunits that are incapable of forming functional receptors alone or even in combination with nAChR α or β subunits (except, perhaps, as α7α5-nAChR), we have not delved deeply into the potential roles played by the α5 subunits. However, α5 subunits can integrate into α3β4*- and α4β2*-nAChRs and perhaps into α6*-nAChRs to further extend the diversity in receptor subtypes and isoforms, which may affect pharmacological profiles.[52,53] Continuing studies will extend our knowledge of the role of α5 subunits in nAChRs.

The essential pharmacophore presented by nicotinic ligands consists of a cationic center (e.g., a quaternized or protonated nitrogen) and a hydrogen-bond acceptor (e.g., the pyridine nitrogen atom in the case of compound 10). The cationic nitrogen binds to a tryptophan residue of the principal α-subunit, forming a cation−π interaction.[54−56] The hydrogen-bond acceptor has been shown to bind to the complementary β-subunit, likely mediated by a water molecule to the backbone NH of a Leu residue in the α4β2 receptor.[57,58] Additionally, various studies support the idea that agonist interactions with the principal α-subunit are important for binding affinity, while interactions with key amino acids present in the complementary β-subunit affect agonist efficacy.[59,60]

High-resolution crystal structures of integral membrane protein mammalian nAChRs are not yet available. Invertebrate pentameric acetylcholine binding protein (AChBP) subunits share 20–24% sequence identity with that of the homologous ECDs of nAChR subunits. However, the putative binding site and some aromatic residues found in nAChR subunits are reported to be conserved. Therefore, AChBPs have been characterized and used to provide insights into ligand recognition sites and other elements of nAChR subunit ECDs and ligand binding site interfaces. The majority of ligand-bound structures have been obtained with AChBP from Aplysia californica (Ac) or Lymnaea stagnalis (Ls) (Figure 3). The Ls-AChBP was postulated to be a more suitable surrogate of α4β2-nAChR for a series of 1-(pyridin-3-yl)-1,4-diazepane analogues because the Trp 53 in Ls-AChBP subunits is Tyr in Ac-AChBP subunits, which interacts less favorably with the quaternized or protonated nitrogen of nicotinic ligands. Recently, the nonmolluscan acetylcholine binding protein from the marine annelid Capitella teleta (Ct), namely, Ct-AChBP, was identified[61] and reported to more closely mimic α4β2-nAChR than does the Ac-AChBP based on comparisons of the ligand binding affinities.[62] The X-ray crystal structure of Ct-AChBP with lobeline or compound 2 bound along with mutagenic studies highlight the importance of key interactions that are responsible for receptor activation or desensitization and for location of key residues in loops D (W57) and E (V111, F119, and L121) in the complementary subunit opposed to the α subunit at the ligand binding subunit interface.

Figure 3

Top view of X-ray crystal structures of Ac-, Ls-, and Ct-AChBPs. The figure was generated using PDB files 2BR7, 1I9B, and 4B5D by PyMOL.

Because of the higher homology between Ct-AChBP and the human α4 and β2 subunits (as computed by ClustalX: α4/Ct-AChBP = 64.6% and β2/Ct-AChBP = 62.1%), especially for key residues involved in ligand recognition (Figure 4), a homology model was generated and refined based on a Ct-AChBP for the ligand binding domain of the human α4β2-nAChR. A ribbon structure for the ECD modeled α4β2-nAChR shows a 10-stranded β-sandwich capped by an N-terminal α-helix for each subunit (Figure 5A), for which the modeled β-sandwich can be nicely superimposed on that of Ct-AChBP (Figure 5B). This homology model may provide useful information for designing other selective α4β2-nAChR ligands.[63]

Figure 4

Amino acid sequence alignment of the Ct-AChBP with extracellular domains of nAChR α4 or β2 subunits. Green boxes highlight positions of key residues of the α4 subunit, and blue boxes outline key residues of the β2 subunit. Residue numeration refers to that for the human α4 subunit. (This is for the mature α4 subunit, not including cleavage of the leaders sequence including the translational start methionine residue.).

Figure 5

Homology model of the human α4β2-nAChR ECD including the ligand binding interface: (A) ribbon structure representation colored by subunit (yellow, α4; azure, β2) for the human α4β2-nAChR ECD and (B) superimposition of the modeled structure (in blue) with the experimental template from Ct-AChBP (in red).

nAChR Agonists and Antagonists Exhibiting Antidepressant-Like Effects

The diversity of nAChR subtypes provides an opportunity to generate subtype-specific ligands that could treat a variety of conditions, although the high sequence homology across individual subtypes of brain nAChRs poses a substantial challenge for the development of subtype-selective nicotinic drugs. Both academia and industry have contributed to a growing body of literature concerning the rational design of potential antidepressant ligands that more selectively target β2*-nAChR than does compound 10.

Rodent behavioral models have been used to assess antidepressant-like effects of these new chemical entities. The most widely used behavioral assay for antidepressant-like efficacy is the FST, in which a mouse or rat is placed in a beaker of water and the amount of time the animal spends passively floating is measured. Clinically therapeutic antidepressants typically reduce the time an animal will spend immobile. Similarly, the tail suspension test (TST) measures immobility time when the rodent is suspended by its tail. This test is relatively short in duration, and in the case where an antidepressant agent is given, the animal will struggle for longer periods of time compared to vehicle treated animals.[64] The novelty-suppressed feeding (NSF) test is a behavioral paradigm originally utilized to measure anxiolytic-like effects of drugs, but more recently the NSF assay has been proposed as a behavioral platform sensitive to the chronic but not acute administration of antidepressants. In NSF, food-deprived mice experience a conflict between feeding and the fear of exploring the novel environment of a brightly lit open area or an unfamiliar cage containing food. Chronic treatment with antidepressants reduces latency to eat in the novel environment. Additionally, an alternative version of the NSF has been advanced that is called the novelty-induced hypophagia (NIH) test, but it utilizes a palatable food to eliminate the need for food restriction. In general, these models predict the onset of action of antidepressants consistent with the therapeutic time course found in humans and therefore validate the NSF/NIH tests as useful behavioral paradigms to gauge the antidepressant efficacy of compounds.[65]

nAChR Antagonists

(1R,2S,4S)-N,2,3,3-Tetramethylbicyclo[2.2.1]heptan-2-amine, 13

The nicotinic ligand that has received the most attention as a potential antidepressant is compound 1, a racemic noncompetitive and nonselective antagonist of nAChRs (IC50,α3β4 = 91–610 nM, IC50,α4β2 = 0.6–2.5 μM, and IC50,α7 = 1.6–6.9 μM). Originally developed as an antihypertensive agent, anecdotal reports of mood modulation and a hypothesis that traditional antidepressants might be acting in part through noncompetitive antagonism of nAChRs[19] led to a preliminary controlled study that demonstrated therapeutic effects of compound 1 for mood disorders, including MDD, that were comorbid with Tourette’s disorder.[10,50] However, the subgroup sizes were very small, and the comorbidity with Tourette’s disorder precluded any conclusions from being drawn about the antidepressant efficacy of compound 1 monotherapy in patients who do not have Tourette’s disorder.

In studies using receptor expression in Xenopus oocytes, the 2S-(+)-enantiomer of compound 1, TC-5214 (13),[51] was found to dissociate more slowly than 2R-(−)-mecamylamine (14) from α4β2- and α3β4-nAChRs, as well as more slowly from α4β2- than α3β4-nAChRs. IC50 values for compound 13 at α3β4-, α4β2-, α7-, or α1β1γδ-nAChRs are similar to those of compound 14 (IC50,α3β4 = 0.2–0.6 and 0.05–0.4 μM, IC50,α4β2 = 0.5–3.2 and 0.5–1.7 μM, IC50,α7 = 1.2–4.6 and 2.2–5.8 μM, and IC50,α1β1γδ = 0.6–2.2 and 0.3–1.1 μM).[51] Further pharmacological studies suggested that compound 13 is a more efficacious antagonist of LS α4β2-nAChRs than compound 1 and can act as a positive allosteric modulator at HS α4β2-nAChRs.[66] This was hypothesized to be mechanistically important; however, the positive allosteric modulation at α4β2-nAChRs was inferred from a slight potentiation of the HS α4β2-selective agonist TC-2559 (15) in a mixed population of HS and LS α4β2-nAChRs. Potentiation of ACh was not demonstrated, and because the effect was not reproduced in a pure population of HS α4β2-nAChRs, activation of LS α4β2-nAChRs by the coapplication of compounds 15 and 13 was not excluded.

Compound 13 exhibited higher anxiolytic- and antidepressant-like effects in several animal models than racemic mecamylamine (forced swim and social interaction tests, light/dark assay). These behavioral activities were attributed to antagonist effects at α4β2*-nAChRs.[67] Moreover, compound 13 showed a superior preclinical safety profile compared to that of either the racemic compound or the 2R-(−)-enantiomer. Compound 13 was found to be well-tolerated in acute and chronic toxicity studies in different animal models (mice, rats, and dogs) and showed acceptable pharmacologic, pharmacokinetic, and metabolic profiles for therapeutic development.

Targacept advanced compound 13 to a phase 2 study as an augmentation in patients who were inadequate responders to the SSRI compound 4. Initial results were very promising, with patients showing an average 6 point improvement on the primary end point, the Hamilton depression rating scale (HAM-D), as well as improvement on other secondary measures.[68] A collaboration with AstraZeneca (Targacept and AstraZeneca’s RENAISSANCE program) followed, advancing compound 13 to phase 3 development as an adjunctive therapy in MDD patients who were inadequate responders to SSRI or SNRI monotherapy. Unfortunately, four phase 3 trials (two fixed and two flexible dose trials) failed to meet the primary end point of achieving a greater change in the Montgomery–Asberg depression rating scale total score for the experimental group receiving an adjunctive therapy of compound 13 combined with an SSRI or SNRI than for the placebo group receiving monotherapy with an SSRI or SNRI alone.[69−72]

There are several possible explanations for the failure of compound 13 in the phase 3 trials. One explanation is that the studies that utilized compound 13 as an adjunct therapy had an unusually high placebo response. Every dose group (compound 13 and placebo) showed at least a 40% improvement in the MADRS total score after 8 weeks of adjunct treatment. A contemporaneous phase 2b monotherapy study of compound 13, which had a lower placebo response, provided some indications of a dose-related antidepressant response, but the study was terminated early.[73] It has been suggested that pursuing compound 13 as a monotherapy may have led to a better outcome than as a combination therapy with an SSRI or SNRI.[73,74] Subjects in placebo groups in the combination study received SSRI or SNRI treatment, and manipulation of the serotonergic or noradrenergic system may have interfered with the antidepressant response of compound 13.[73] Additionally, other authors have suggested that a combination study with a tricyclic antidepressant, which would have antimuscarinic effects, rather than an SSRI or SNRI devoid of antimuscarinic activity, may have been more effective. An antidepressant with antimuscarinic effects could have pro-nicotinic effects through a neuro-adaptive upregulation of cholinergic tone, which could be normalized by compound 1. Alternatively, combining an antinicotinic with an antimuscarinic may have produced a complementary effect in reducing a hypercholinergic state than either treatment alone. Papke and Picciotto[74] also suggest that compound 1, in particular, the 2S-(+)-enantiomer 13, may not have been the best choice for the clinical trials. The rationale for using compound 13 was based on a study that suggested that compound 13 had both activating and inhibitory effects at α4β2-nAChRs, a better in vivo profile in animal studies, and a better safety profile than the 2R-(−)-enantiomer.[67,75] However, more recent studies found little in vitro evidence to differentiate the pharmacological properties of compound 1 and the 2R- and 2S-enantiomers and only a modest difference in potency in vivo in the tail flick test between the stereoisomers.[76] Furthermore, compound 1 is an antagonist that is more potent at inhibiting ganglionic α3β4-nAChRs than α4β2-nAChRs in the brain that are thought be the primary target of hypercholinergic activity associated with depression.[25,77] A recent estimate of the free compound 1 concentration in the brain indicates that it is likely insufficient to be pharmacologically relevant, blocking approximately 20% of α4β2*-nAChR function.[21] Further evidence that compound 13 did not provide sufficient inhibition on nAChRs comes from the failure of CP-601927 (16),[78] an α4β2 partial agonist that was tested in a phase 2 clinical trial as an augmentation in treatment resistant subjects with major depression.[79,80] Weber et al.[21] propose that compound 16, which was predicted to inhibit only 23% of α4β2-nAChRs,[81] would not translate into improved antidepressant efficacy.

Although it is possible that compound 13 could be an effective monotherapy, given its poor nAChR subtype selectivity and the failure in the phase 3 clinical trial, it seems unlikely that this ligand will have a future in the treatment of depressive disorders. The disappointing failure of compound 13 illustrates the need for a better understanding of the optimal pharmacological properties to maximize clinical antidepressant efficacy. More potent nAChR inhibitors that block >25% α4β2 nAChR would likely be more effective than compound 1 or compound 13. Inhibition may be achieved with more potent antagonists, partial agonists with very low intrinsic activity, or desensitizing agents.

2-(tert-Butylamino)-1-(3-chlorophenyl)propan-1-one, 7

Compound 7(82) is an α-aminoketone believed to elicit antidepressant effects by acting as a dopamine and norepinephrine reuptake inhibitor. It was first reported to show efficacy as a smoking-cessation aid in nondepressed patients in 1994 and has subsequently proven to be a noncompetitive antagonist at a variety of nAChR subtypes.[83,84] Compound 7 is most potent as a dopamine reuptake inhibitor, with an IC50 of 550 nM, while potencies at norepinephrine transporters and a number of nAChR subtypes fall in the low micromolar range. However, it has long been appreciated that clinical outcomes correlate poorly with plasma concentrations of compound 7(85) and that active metabolites reaching higher and longer-lasting concentrations in plasma likely play the leading role. The principal metabolite, 2S,3S-hydroxybupropion (17), has reduced potency at dopamine transporters and most nAChR subtypes, whereas its potency at norepinephrine transporters and α4β2-nAChRs is enhanced.[76] It appears that a determining factor of compound 7’s efficacy is its metabolism to sufficiently high concentrations of compound 17, as differences in its plasma concentrations between responders and nonresponders was the most statistically significant difference in a 2006 study.[86] Plasma concentrations reported for compound 17 were in the low micromolar range. Although compound 7 and another active metabolite, threohydrobupropion, were present at lower concentrations, they may have additive effects in combination with the principal metabolite that provide a significant contribution to a complex pharmacology mediated by norepinephrine transporters, dopamine transporters, and α4β2-nAChRs. Evaluation of compound 7 and its hydroxyl metabolites in the mouse FST showed that compound 17 was equally potent to that of compound 7, whereas the 2S,3R-isomer showed no effect on immobility.[83] The 2S,3S-isomer was also a more potent inhibitor of [3H]norepinephrine uptake and [3H]dopamine update as well as a more potent antagonist of α4β2-nAChR function than that of the 2S,3R-isomer, suggesting that the 2S,3S-isomer may be a better candidate than compound 7 for antidepressant treatment.

(2S,13bS)-2-Methoxy-2,3,5,6,8,9,10,13-octahydro-1H,12H-pyrano[4′,3′:3,4]pyrido[2,1-i]indol-12-one, 18

Dihydro-β-erythroidine (DHβE, 18)[87] is an alkaloid that is isolated from the seeds of Erythrina L., a genus of trees and shrubs that is found in the tropics and subtropics around the globe. In Central America, Erythrina had an important role in folk medicine during precolonial times. Compound 18 acts as a competitive antagonist at α4β2-nAChRs (IC50,α4β2 = 30 nM and IC50,α3β4 = 23 μM)[87] that blocks compound 10-induced dopamine release from rat striatal slices (IC50 = 30 nM).[88] Compound 18 shows antidepressant-like effects in the mouse FST (3.0 mg/kg) and TST (1.0 and 3.0 mg/kg) without affecting locomotor activity.[89] Additionally, compound 18 is able to potentiate the antidepressant-like effects of compound 9 (4 and 20 mg/kg) in the TST.[22]

((3S,6S,6aS,7R,7aR,8S,9R,10S,11aR,12S,12aS,13S)-1-Ethyl-11a,12-dihydroxy-6,8,10,13-tetramethoxydodecahydro-2H-3,6a,12-(epiethane[1,1,2]triyl)-7,9-methanonaphtho[2,3-b]azocin-3(4H)-yl)methyl 2-((S)-3-methyl-2,5-dioxopyrrolidin-1-yl)benzoate, 19

Methyllycaconitine (MLA, 19),[90] a complex diterpenoid alkaloid isolated from the seeds of Delphinium brownii, is a competitive and selective α7-nAChR antagonist with an IC50 value in the subnanomolar range, as tested in Xenopus oocytes and α7-transfected SH-SY5Y cells.[90,91] Compound 19 was found to partially inhibit anatoxin-evoked dopamine release from rat striatal slices[92] and decrease the time spent immobile in NMRI female mice in the mouse FST (10 mg/kg) and TST (10 mg/kg) without affecting locomotor activity.[89] Another study, however, found no effect of compound 19 (10 mg/kg) in the FST in BALB/c male mice.[38] The discrepancies between these studies, which may be related to mouse strain, gender, or procedural differences, warrant further investigation into the role of α7 nAChR inhibition and antidepressant function.

7-(Pyridin-3-yl)-1,7-diazaspiro[4.4]nonane, 20

Another α4β2-nAChR antagonist is the Targacept compound TC-2216 (20).[93] This compound was reported to show beneficial effects in preclinical studies, thereby promoting its further development for the treatment of depression and anxiety disorders.[93] A phase 1 clinical trial of racemic compound 20 was initiated in 2008, but no further trials were conducted.

nAChR Agonists

Modulation of nAChRs with partial agonists is pharmacologically complex, as it involves potentially simultaneous and interacting effects. Under acute conditions, partial agonists elevate baseline cholinergic tone while simultaneously lowering the ceiling of nAChR-mediated signaling. Depending on potency, dose, and exposure, partial agonists could have an additive effect with endogenous ACh, thereby lowering the threshold for ACh signaling while simultaneously lowering the efficacy of that signal. Under conditions of chronic administration, partial agonists desensitize nAChR, and it is believed that the resulting attenuation of cholinergic signaling is the principal pharmacotherapeutic end point.

Cytisine and Derivatives

(1R,5S)-1,2,3,4,5,6-Hexahydro-8H-1,5-methanopyrido[1,2-a][1,5]diazocin-8-one, 21

(−)-Cytisine (21),[94] a tricyclic quinolizidine alkaloid isolated from the seeds of Cytisus laburnum L., has been used to treat tobacco dependence in Eastern Europe since the 1960s. Compound 21 is an α4β2-nAChR partial agonist (Ki,α4β2 = 2 nM) and an α3β4 (Ki,α3β4 = 480 nM) and α7 (Ki,α7 = 5890 nM) full agonist that shows antidepressant-like activities in several rodent models,[95] presumably mediated by a reduction of neuronal activity in the basolateral amygdala.[96] Compound 21 (10 μM) was reported to exhibit 56% of the response relative to that of compound 10 and to inhibit 30% of the current evoked by compound 10 (10 μM) in Xenopus oocytes expressing human α4β2-nAChRs.[97] However, it was reported to have unfavorable side effects, which include nausea, vertigo, abdominal pain, respiratory stimulation, and muscle weakness.[98] Mineur et al. observed that compound 21 was not tolerated by mice at a dose of >1.5 mg/kg.[95] Additionally, the poor absorption and brain penetration[94,99−101] of compound 21 may also limit its application as a clinical antidepressant.

(1R,5S)-11-Bromo-1,2,3,4,5,6-hexahydro-8H-1,5-methanopyrido[1,2-a][1,5]diazocin-8-one, 23

SAR studies of cytisine analogues revealed that phenyl ring replacements of the pyridone ring of compound 21 reduced the binding affinities and functional efficacies.[97] Substitution of 21 at the C-3 (−CH=CH2,[102] −NO2,[103] or pyridinyl[95]), C-4 (−CH3),[102] or C-5 (-Br)[78] position could lead to the same or improved α4β2-nAChR binding affinity, whereas substitution at the piperidine nitrogen,[97,103,104] the C-3 position (aromatic rings),[102] or at the C-12 position (−CO2CH3 or −COCH2CH3)[102] reduced affinity. Introduction of substituents on the piperidine ring nitrogen,[97] C-4 (−CH3 or −CH2OH),[102] or the C-5 (−NO2)[103] position increases the selectivity for α4β2 receptor. Among these substituted cytisine analogues, 3-(2-pyridyl)cytisine (22)[95] and 5-bromocytisine (23)[95] were further tested in mouse antidepressant efficacy models. Both compounds displayed high affinity at α4β2*-nAChRs (Ki,α4β2* = 0.9 for 22 and 0.3 nM for 23), whereas compound 22 was found to show lower affinities at α3β4*-nAChRs and α7*-nAChRs (Ki,α3β4* = 119 nM and Ki,α7 = 1100 nM for 22 vs Ki,α3β4* = 3.8 nM and Ki,α7 = 28 nM for 23). In electrophysiology assays, compound 22 activates both HS and LS α4β2*-nAChRs (EC50,HS = 12 nM and EC50,LS = 31 nM) expressed in oocytes with low efficacies (less than 10% relative to the efficacy of ACh at both HS and LS receptors) while exhibiting little agonism at α3β4*.[95] In the same assays, compound 23 activated both HS and LS α4β2*-nAChRs with similar potencies (EC50,HS = 13 nM and EC50,LS = 15 nM) and efficacies (17% relative to ACh). A comparison of binding and functional data for compound 21 and its selected analogues are presented in Table 1. The antidepressant-like effects of compound 22 were demonstrated in the TST (0.6 mg/kg, but not at 0.3 or 0.9 mg/kg), FST (0.3–0.9 mg/kg), and chronic NSF tests (15 days at 0.3 mg/kg, but not at 0.6 mg/kg) in C57BL/6 mice. On the other hand, compound 23 (0.3–1.2 mg/kg) failed to show any significant effects in the same tests 30 min postintraperitoneal injection.[95] It has been suggested that the lack of efficacy of compound 23 was likely due to the low brain penetration, as 50 ng of the compound in 1 μL of artificial cerebrospinal fluid showed a significant antidepressant-like effect in the TST when administered centrally.

Table 1

Binding Affinities and Maximal Responses and Potencies of Compounds 2 and 21–23 with Respect to Activation of nAChRs Expressed in Oocytes[95,101]

		α4β2 activation				α4β2 inactivation
ID	ref	LS efficacy	HS efficacy	LS EC50 (nM)	HS EC50 (nM)	LS IC50 (nM)	HS IC50 (nM)
ACh	(95)	100%	100%	73	1.7
2	(101)		22%		1400	50 000	70
21	(101)		6.5%		2000	28 000	50
22	(95)	3%	8%	31	12
23	(95)	17%	17%	15	13

(6R,10S)-7,8,9,10-Tetrahydro-6H-6,10-methanoazepino[4,5-g]quinoxaline, 2

Inspired by morphine and its simplified analogues [3.3.1]- and [3.2.1]-bicyclic benzomorphans, Coe et al. identified the simplified cytisine analogue, benzazapine 24,[105] as a nicotinic antagonist with a Ki value of 20 nM. Attachment of a trifluoromethyl group to the C-4 position of benzazapine 24 gives 16 and CP-601932 (25)[78] (Figure 6). Compound 16 is a selective α4β2-nAChR partial agonist with weaker activity at the α3β4-nAChRs (Ki,α4β2 = 1.2 nM vs Ki,α3β4 = 102 nM),[78] whereas its enantiomer, 25, shows similar affinities at both receptor subtypes, with Ki values of 21 nM and lower affinities for the α6- and α7-nAChR subtypes (Ki > 300 nM).[106,107] Compound 16 decreased the time spent immobile in the FST at 0.75, 1, and 1.5 mg/kg, whereas no significant effect was found in the TST at the same doses.[78] Moreover, compound 16 was found to be relatively safe and well-tolerated in phase 1 single- and multiple-dose studies, but it failed to show a statistically significant change on the primary efficacy scale in favor of the drug when used in the augmentation of antidepressant therapy in major depression in a phase 2 clinical study.[108]

Figure 6

Selected nicotinic benzazapine analogues.

Nitration of benzazapine 24 in the presence of 2 equiv of nitronium triflate (CF3SO2O–NO2+) in dichloromethane afforded the 4,5-dinitrated benzazapine 26(105) with an unexpected regioselectivity (the meta-isomer was found to be less than 10% of the mixture). Reduction of 26 to the diamine, condensation with glyoxal, and deprotection followed by salt formation provided compound 2 (Figure 6).[105] This achiral quinoxaline 2 is the most publicized cytisine analogue, which is known as varenicline. Its tartrate salt has been launched and marketed by Pfizer for smoking cessation in the U.S. (Chantix), Canada, Europe, and other countries (Champix). Compound 2 is a potent partial agonist at α4β2*-nAChRs, with a Ki value of 0.4 nM and an agonist efficacy of 40–60% of that of compound 10.[109,110] In Xenopus laevis oocytes expressing α4β2-nAChRs, compound 2 was found to have an EC50 of 2.3 μM and an efficacy of 13.4% relative to that of ACh.[109] It potently blocks compound 10 binding to α4β2*-nAChRs and produces a sustained increase in dopamine release to 60% of the maximal effect of compound 10 (188% at 0.32 mg/kg sc).[110] In addition, 1 mg/kg po compound 2 reduced the dopamine-enhancing effects of a subsequent dose of 0.32 mg/kg sc compound 10 to the level of the effect of compound 2 alone.[110] In a SPECT imaging study, compound 2 (0.5 mg) was found to completely saturate α4β2*-nAChRs in human brain.[111] The dissociation half-life of compound 2 is about 5.4 h in a mouse ex vivo receptor occupancy study.[38] Compound 2 alone has antidepressant-like effects in the FST and was able to augment the effects of compound 5 when administered in combination.[112] It enhanced the mood and cognitive function associated with compound 10 withdrawal in patients in clinical trials for smoking cessation.[113] In an 8 week open-label study, compound 2 augmented the effects of traditional antidepressants in depressed smokers who derived minimal benefit from standard antidepressant treatment.[12] Recently, studies on a larger scale of community volunteers who wanted to quit smoking revealed that 12 week administration of compound 2 (0.5 mg/day for days 1–3 followed by 0.5 mg twice a day for days 4–7 and 1 mg twice a day thereafter) was coupled with a generalized suppression of depression compared with that of the placebo-controlled treatment.[114]

A-85380 and Analogues

(S)-3-(Azetidin-2-ylmethoxy)-5-iodopyridine, 12

The pyridine ether-based ligands, in which a CH2O linker is inserted between the pyrrolidine ring (or in case of its analogues, an azetidine ring) of compound 10 and its pyridine ring have attracted considerable interest as α4β2-nAChR agonists due to their high potency.[115−118] A-85380 (27)[119] is a 3-pyridyl ether that exhibits both high potency and selectivity for β2-containing nAChRs, primarily α4β2* in brain (Ki,α4β2* = 0.02–0.05 nM for rat brain receptors and 0.04 nM for human brain receptors) relative to that of human α7-nAChRs (Ki,α7 = 148 nM) or muscle-type α1β1γδ-nAChRs (Ki,α1β1γδ = 314 nM).[119,120] Functionally, compound 27 acts as a potent activator of both human α4β2-nAChRs (EC50 = 0.7 μM) and ganglionic α3β4*-nAChRs (EC50 = 0.8 μM), the effects of which are blocked by pretreatment with compound 1. In the FST, compound 27 was found to have antidepressant-like effects that could be blocked by pretreatment with the nAChR antagonists compounds 1 and 18, as well as by the nonselective serotonin receptor antagonist methiothepin, suggesting that compound 27 exerts its effects via neuronal nicotinic receptor activation of serotonergic pathways.[119,121] As exemplified by compound 12(120) (Ki,α4β2* = 0.01 nM), introduction of a halogen substituent at the C-2 (fluoro only), C-5, or C-6 position of compound 27 leads to ligands that retain subnanomolar affinity for the α4β2*-nAChRs, as measured by radioligand binding assays utilizing rat brain membrane preparations.[120] Similar to compound 27, 12 was also found to decrease immobility in the mouse FST.[38] Recently, β2*-nAChR availability was found to be decreased across brain regions in depressed patients compared to that of healthy subjects using SPECT with the β2*-selective radioligand tracer [123I]12.[39] A unifying explanation of these effects stems from the idea that desensitization and antagonism provide similar end points for agonists and antagonists. However, because agonists have an acute activation phase, behavioral assays may measure effects due to activation and/or inactivation of nAChRs. In this context, one would expect coadministration of an antagonist to reduce the behavioral signature of an agonist administered alone. More complicated explanations involving multiple nAChR subtypes also can be made, but they are highly speculative.

(S)-5-(Azetidin-2-ylmethoxy)-3-methylisoxazole, 28

Replacement of the pyridine core of these ether-based ligands with a methylisoxazole group led to compound 28, which binds to α4β2-nAChRs with higher affinity than to the α3β4-nAChRs (Ki,α4β2 = 4.6 nM vs Ki,α3β4 = 692 nM).[122] The functional potency of compound 28 (1.2 μM) at the α4β2-nAChR is similar to that of compound 2 (1.4 μM), whereas the efficacy of compound 28 is higher (110% vs 53%) in cells heterologously expressing a mixture of HS and LS α4β2-nAChRs (Table 2). Broad screening showed that this compound was highly selective for nAChRs and did not have significant binding affinity to the other 45 neurotransmitter receptors and transporters tested. In the mouse FST, compound 28 was found to decrease immobility at 1–5 mg/kg i.p. and 5 mg/kg po. This compound showed no significant hERG or CYP (1A2, 2B6, 2C9, 2C19, 2D6, and 3A4) inhibition at 10 μM. Between 53.4 and 73.7% of compound 28 was found to remain unchanged after 60 min incubation with mouse or human liver microsomes (1 and 10 μM).[122]

Table 2

Functional Potencies and Efficacies of Ligands 2, 10, 11, and 28–31: Agonism and Inactivation at Human α4β2-nAChRs[122−125]

	agonism			inactivation
compound	EC50 (nM)	efficacy (%)a	efficacy HS (%)b	IC50 (nM)	efficacy (%)
2	1400	53		110	85
11	5.8	55	100	4.8	37
28	1200	110	92	169	78
29	10	21	92	9.4	63
30	42	22	61	31	68
31	8.4, 2300c	21	45	58	85
10	290	88	110	430	93

The efficacies were measured in a mixture of HS and LS α4β2-nAChRs.

The efficacy values were extrapolated using compound 11 defined as a full agonist at the HS α4β2-nAChR.

Compound 31 activates both HS and LS α4β2-nAChRs with sufficient selectivity to distinguish activity at each subtype. Efficacy of 31 at LS α4β2-nAChRs is approximately 17%.

(S)-6-(5-(Azetidin-2-ylmethoxy)pyridin-3-yl)hex-5-yn-1-ol, 11

Compound 11, an analogue of compound 27 having an alkynyl substituent attached to the 5-position of the pyridine ring, was identified as a highly potent full agonist at HS α4β2-nAChR and a low efficacy agonist at LS nAChR (Table 2).[126] Compared with compound 27, the selectivity of compound 11 at α4β2- relative to ganglionic α3β4*- and homomeric α7-nAChRs is greatly improved.[127,128] Compound 11 potently binds to α4β2-nAChRs (Ki = 0.4 nM) but not to α3β4*- (Ki > 104 nM) or α7-nAChRs (Ki > 104 nM). Ex vivo receptor occupancy in the thalamus showed prolonged receptor occupancy. The dissociation half-life of compound 11 (8–24 h) was found to be significantly longer compared with that of compound 12 (3–5 h),[38] which is likely due to the presence of the side chain in compound 11.

Robust antidepressant-like effects of compound 11 were observed in the FST, TST, and NIH assays, with no effect on baseline locomotor activity when given chronically.[35,38,129] Compound 11 (1 mg/kg) significantly reduced immobility 2, 3, and 4 h but not 5 h after administration in the mouse FST. Repeated administration of compound 11 in the FST at 1 and 3 mg/kg doses for 2 weeks did not result in tolerance. Moreover, its activity in the FST was completely blocked by either compound 18 or compound 1,[38] suggesting that activation of nAChRs may responsible for the activity of compound 11 in this behavioral model of antidepressant efficacy. In addition, β2 subunit knockout mice did not show any antidepressant response to compound 11 in the FST, suggesting that interaction with β2*-nAChRs plays a key role in the behavioral activity of compound 11.[38] Stimulation of dopamine and noradrenaline release in rat striatal slices was observed with compound 11, an effect that was blocked by nicotinic antagonists 18 and compound 1.[126] In addition, the α6*-selective antagonist α-conotoxin MII (100 nM) reduced the maximum effect of compound 11-induced dopamine release in rat by 48%, suggesting that α6*-nAChRs may be involved in the antidepressant-like effects of compound 11.[126]

Compound 11 has been reported to have other behavioral effects including increased hypothermia.[130,131] It is unlikely that the hypothermic effects of compound 11 are related to its antidepressant-like effects observed in the FST. Although reduced, β2 knockout mice still showed a hypothermic response to compound 11,[131] whereas the antidepressant-like effects of compound 11 was completely abolished in β2 knockout mice.[38] The hypothermic effects of compound 11 (1 mg/kg) returned to baseline approximately 2 h postinjection,[132] whereas the antidepressant-like effect persisted up to 4 h.[38] Compound 11 has shown additional beneficial behavioral effects in animal models including analgesia,[133] improving attentional function after disruption with compound 3 or MK-801,[132,134] reducing anxiety following withdrawal from chronic compound 10,[135] decreasing body weight gain following chronic treatment,[136] and decreasing compound 10 self-administration and alcohol consumption.[137−139] Interestingly, unlike compound 10 or compound 2, compound 11 did not upregulate nAChR in the brain after chronic administration[140] and did not maintain receptor upregulation following chronic compound 10 administration,[136] providing further evidence that compound 11 may exert its behavioral effects through a unique mechanism.

(S)-N-Phenyl-5-(pyrrolidin-2-ylmethoxy)pyridin-3-amine, 31

Replacing the acetylene bond of compound 11 with a substituted cyclopropane or an isoxazole ring lead to compounds 29 and 30, respectively.[63,124,125,141] Introduction of different amino groups at the 5-position of the pyridine ring yielded a series of 3-alkoxy-5-aminopyridine derivatives exemplified by compound 31.[123] These compound 11 analogues (29–31) act as α4β2-nAChRs partial agonists (Table 2) with low nanomolar binding affinities (Ki = 0.1–1.2 nM) and excellent subtype selectivity at β2*- over β4*-nAChRs (Table 3). Compound 29 has been cocrystallized with Ct-AChBP, and its binding mode was found to be similar to that reported for compound 2; both compounds rely on the presence of two cationic centers spaced ∼5.8 Å from each other for their binding interactions, including cation−π interactions with W153 (loop B) and Y201 (loop C) as well as a set of H-bonds with the backbone atoms of W153 (loop B), Q116 (loop E), and I128 (loop E). (Figure 7.) The hydroxyethyl group of 29 appears to be flexible and is able to adopt different orientations, thus highlighting the ability to carry out structural modifications of this appendage without causing significant alterations in activity. The antidepressant-like properties of compounds 29–31 were demonstrated in the classical mouse FST (1–30 mg/kg). The lack of significant interactions with other neurotransmitter receptors and transporters widely distributed throughout the CNS as well as the favorable preliminary absorption, distribution, metabolism, excretion, and toxicity (ADME-Tox) profiles place these specific α4β2-nAChR ligands in a favorable position to be further studied as new antidepressant drug candidates.

Table 3

Binding Affinities of Compounds 10, 11, and 27–31 at Seven nAChR Subtypes[122−125]

	Ki (nM)a
compound	α2β2	α2β4	α3β2	α3β4	α4β2	α4β4	α4β2*b	selectivity α4β2/α3β4
11				>10 000	0.4		0.9	24 000
27					0.05		0.05
28	4.3	311	8.7	692	4.6	86.0	12.0	150
29	0.1	249	3.0	6520	0.1	82.6	0.5	65 200
30	1.0	935	15.4	>10 000	0.6	1790	3.0	>16 300
31	1.7	559	40.6	5640	1.2	16.9	1.4	4700
10	5.5	70	29	260	4.9	23	9.8	53

Ki values were determined by competition for [3H]epibatidine binding sites using radioligand binding.

α4β2*, prepared from rat forebrain.

Figure 7

X-ray crystal structure of the Ct-AChBP in complex with compound 2 or 29.[63]

Other Ligands

(S,E)-5-(5-Isopropoxypyridin-3-yl)-N-methylpent-4-en-2-amine, 32

TC-1734 (32) is a potent α4β2-nAChR partial agonist (Ki = 11 nM) and is selective over α7-nAChRs (Ki > 50 000 nM).[142] Compound 32 (10 μM) showed no significant inhibition when tested at 135 other receptor and enzyme systems. In functional studies, the agonist activity of compound 32 was measured using a 86Rb+ efflux assay in rat thalamic synaptosomes (α4β2 subtype, EC50 = 220 nM, efficacy = 57%) and the dopamine-release assay in striatal synaptosomes (α4β2/α6/α3β2, EC50 = 106 nM, efficacy = 55%). Compound 32 (up to 100 μM) did not show any detectable effects at either muscle-type (α1β1γδ) or ganglionic (α3β4) nAChRs. Compound 32 had favorable pharmacokinetic and metabolic profiles and was well-tolerated in acute and chronic oral toxicity studies in various animal species (mice, rats, and dogs), as well as in a human clinical studies using either single- or repeated-dose oral administration.[143] Compound 32 exhibited antidepressant-like effects in the mouse FST when administered intraperitoneally, with a significant reduction in immobility at the minimal dose of 1 μmol/kg, and locomotor activity was not affected by repeated administration.[142] In clinical trials, compound 32 has been studied in a variety of indications characterized by varying types and degrees of cognitive impairment, such as attention deficit/hyperactivity disorder, mild cognitive impairment, and age-associated memory impairment.[144,145] Currently, Targacept is recruiting patients for a phase 2b study with this compound in order to assess efficacy, safety, and tolerability in patients with mild to moderate dementia due to Alzheimer’s disease.[146,147]

(S)-3-Ethynyl-5-(1-methylpyrrolidin-2-yl)pyridine, 33

SIB-1508Y (33)[148] is a selective α4β2-nAChR partial agonist (EC50 = 1.8 μM; efficacy = 49% relative to that of compound 10), having no activity at homomeric α7- or muscle-type α1β1γδ-nAChRs and weak activity at ganglionic α3β4-nAChRs (EC50 = 23 μM; efficacy = 52% relative to that of compound 10).[148] In vitro, compound 33 stimulated dopamine release from slices of rat striatum prepared from various brain regions, and the effect was blocked by nAChR antagonists compound 1 or compound 18.[149] On the other hand, compound 33 was found to show relatively weak effects on NE or 5-HT release. Subchronic or chronic administration of compound 33 demonstrated robust antidepressant-like effects in the learned helplessness model in rat, which were significantly blocked by nAChR antagonist compound 1 in the subchronic study.[150] In addition, compound 33 was also found to improve cognitive and motor function in monkey models of Parkinson’s disease.[151,152]

1-(5-Chloropyridin-3-yl)-1,4-diazepane, 34

NS3956 (34)[23] was found to be a partial agonist at human α4β2-nAChRs (Ki = 0.36 nM) with good selectivity (Ki,α7 = 399 nM and Ki,α1 = 944 nM). In the oocyte expression system, this compound was very potent at HS α4β2-nAChRs (EC50 = 11 nM), with an efficacy of 47% of the response to a maximally efficacious concentration of ACh.[23] At LS α4β2-nAChRs, potency and efficacy were both reduced, with an EC50 of 604 nM and an efficacy of 38% relative to that of ACh. In reference to the subtype selectivity and off-target activity, actions at α7- or α3β4-nAChRs of compound 34 were weak or absent, and it did not show any significant interactions with the monoamine transporters, including 5-HT, DA, and NE. Moreover, compound 34 exhibited an ED50 value of 0.033 mg/kg in the [3H]epibatidine binding assay when tested in vivo, indicating that it was able to cross the BBB. Interestingly, compound 34 (0.3–3.0 mg/kg) was shown to be inactive in the mouse FST when tested alone, whereas when tested as a combination therapy, it significantly enhanced the antidepressant-like effects of the SSRI compound 4 and the SNRI compound 6 at 1.0 mg/kg.[23] These results suggest the possibility of a synergistic interaction between nicotinic agonism and the action of antidepressant medications. In vitro binding studies along with the detailed interactions with Ls-AChBP revealed that a bromine substitution at the 6-position increased the efficacy at HS α4β2-nAChR, and a relatively small substitution, such as chloro, bromo, or ethoxy, at the 5-position of the pyridine ring in a 1-(pyridin-3-yl)-1,4-diazepane scaffold improved selectivity for α4β2- over α7-nAChRs.[57]

(R)-4-Chloro-N-(quinuclidin-3-yl)benzamide, 35

PNU-282987 (35)[153] is a potent and selective α7-nAChR agonist (Ki = 26 nM) with good selectivity over α3β4- or α1β1γδ-nAChRs (≥60 μM). Compound 35 did not show significant interactions with monoamine, muscarine, glutamate, or GABA receptors when tested at concentrations of 1 μM, with the exception of the 5-HT3 receptor, where it showed a Ki value of 930 nM. Furthermore, compound 35 was able to evoke whole-cell currents from cultured rat hippocampal neurons and enhance GABAergic synaptic activity in hippocampal slices.[153] Compound 35 (10–20 mg/kg) show no antidepressant-like activity when tested alone, but it significantly enhanced the antidepressant-like effect of the SSRI compound 4 or the SNRI compound 6.[23]

Conclusions

Considerable evidence supports the hypothesis that hyperactivity of the cholinergic system over that of the adrenergic system leads to depression. Clinical and preclinical studies provide evidence that modulating nAChRs may lead to antidepressant effects when given alone and in combination with traditional antidepressants (summarized in Table 4). Compound 10 has a clear and long-recognized activity as an antidepressant, and a childhood history of depression elevates the risk for those individuals becoming users of tobacco products, reflecting self-medication with compound 10 as a “street” antidepressant.[154] It is difficult to determine the critical drug features that are responsible for compound 10’s antidepressant activity, given the pharmacodynamic and pharmacokinetic variables and the various effects of compound 10 on nAChRs in vitro. Our understanding is incomplete about where and which nAChR subtypes are involved in excitatory neurotransmission and/or in modulation of neurotransmission mediated by monoamines or other chemical messengers implicated in mood regulation. Pharmacokinetics bring additional uncertainties with regard to drug availability and efficacy in vivo and translation of ligand activities in vitro to those in vivo. For example, effects of several agonists in animal models are attenuated by coadministered antagonists, suggesting that nAChR activation contributes to some antidepressant-like drug effects. Antagonists alone also elicit antidepressant-like effects, suggesting that inhibition of nAChR function through processes such as desensitization also contribute to antidepressant efficacy, consistent with the cholinergic/adrenergic hypothesis. Moreover, compound 10 and many nicotinic ligands have limited selectivity across nAChR subtypes or subtype isoforms, meaning that they could have a matrix of effects across subtypes as well as on the activation–inactivation axis. There still is a need for reliable assays for several nAChR subtypes, in vitro, complicating the interpretation of results and strategies in identification of promising compounds. It is also important to be mindful that behavioral tests in animals measure changes in behavior that correlate with clinical efficacies of currently available antidepressants and do not necessarily predict effective antidepressant activity in humans, particularity in treatment-resistant patients.

Table 4

Clinical and Preclinical Evidence for the Viability of Targeting nAChRs in Depression

ID	receptor subtype selectivity	antidepressant/anxiolytic results	ref
2	α4β2 partial agonist, less potent α3β4 and α7 full agonist	Improved FST; augmented the antidepressant effects in depressed smokers	(12, 38, and 109−114)
11	α4β2 partial agonist	Improved FST, TST, and NIH	(35, 38, and 126−129)
12	α4β2 full agonist	Improved FST	(38, 39, and 120)
13	Nonselective noncompetitive antagonist	Improved FST, social interaction test, light/dark assay; failed phase 3 clinical study as add-on in treating resistant patientsdark assay	(19, 21, 50, 51, and 66−72)
16	α4β2 partial agonist	Improved FST Failed as augmentation of antidepressant therapy for major depression in phase 2 clinical study	(78 and 108)
18	α4β2 competitive antagonist	Improved FST and TST	(22 and 87−89)
19	α7 antagonist	Improved FST and TST	(89−92)
20	α4β2 antagonist	Antidepressant and anxiolytic like response in preclinical studies	(93)
21	α4β2 partial agonist, α3β4 and α7 full agonist	Antidepressant-like activities in several rodent models; safety issues, poor absorption and limited brain penetration	(94−101)
22	α4β2 partial agonist	Improved FST, TST, and chronic NSF	(95 and 97)
27	α4β2 partial agonist, α3β4 agonist	Improved FST	(119−121)
28	α4β2 partial agonist, less potent α3β4 agonist	Improved FST	(122)
29	α4β2 partial agonist	Improved FST	(63 and 125)
30	α4β2 partial agonist	Improved FST	(124 and 141)
31	α4β2 partial agonist	Improved FST	(123)
32	α4β2 partial agonist	Improved FST	(142−147)
33	α4β2 partial agonist	Improved learned helplessness	(148−152)
34	α4β2 partial agonist	No effect alone in FST; enhanced the antidepressant-like effect of SSRI (compound 4) and SNRI (compound 6)	(23 and 58)
35	α7 agonist	No effect alone at FST; enhanced the antidepressant-like effect of SSRI (compound 4) and SNRI (compound 6)	(23 and 153)

Recent phase 3 findings indicated that the nAChR antagonist compound 13, the 2S-(+)-enantiomer of the broad-spectrum, noncompetitive nAChR antagonist, compound 1, did not show efficacy as an adjunct therapy for patients that were nonresponders to traditional antidepressant treatment. This suggests that simple antagonism of what is likely to be a number of nAChR subtypes has an insufficient antidepressant effect. However, the lack of nAChR subtype specificity and the marginal free concentration in brain are also significant issues for compound 13 that may have contributed to the negative outcomes in the phase 3 trials. Those outcomes are not likely to be accurate predictors of effects of nicotinic full or partial agonists with greater nAChR subtype selectivity.

Designing a compound that has actions more like those of compound 10 (short-lived nAChR agonism mediated by binding to the orthosteric sites followed by receptor desensitization) may produce a more efficacious antidepressant. Factors outlined above need to be addressed toward optimization of ligand action of the desired type and at the appropriate nAChR subtypes. These challenges aside, preclinical evidence shows potent antidepressant-like efficacy of compound 11 and some of its analogues in the mouse FST. These ligands have high selectivity at α4β2*-nAChR over other nAChR subtypes and act as partial agonists at the (α4)2(β2)3-nAChR isoform. These findings suggest and support the hypothesis that ligands selectively targeting the (α4)2(β2)3-nAChR isoform may hold significant promise as efficacious antidepressants. These ligands would offer advantages over nicotinic agents that are relatively nonselective and produce adverse side effects through actions at peripheral nAChRs or at central non-α4β2*-nAChRs.

Aside from their selectivity for α4β2*-nAChR, compound 11 and its analogues have favorable pharmacokinetic and toxicological profiles. Advancement of the best of these ligands to and through clinical trials still requires substantial investment of effort and capital. Our work funded under a National Cooperative Drug Discovery and Development program has provided important impetus to such advancement, but private sector engagement and commitment to treatment of depression or other psychiatric disorders will be necessary. This poses an additional challenge given the perhaps more proximal benefit of attention on new drugs to treat diseases of aging and metabolism and acute/chronic pain than to treat psychiatric disorders that are poorly understood, perhaps because of their etiopathogenic heterogeneity. If nAChR ligands are to be prescribed as antidepressants, then it may happen through the inverse route that compound 7 took to be prescribed as a smoking-cessation aid. α4β2*-selective agonists and partial agonists that are attractive as potential antidepressants may also be efficacious as smoking-cessation aids. It is entirely possible that this class of compounds will first reach the clinic as an improvement over compound 2 and that antidepressant efficacy is ascertained thereafter. Regardless, studies with α4β2*-nAChR-selective ligands have already contributed to an improved understanding about nAChR subtypes and isoforms, and elucidation of roles played by the nicotinic system in the effects of these ligands certainly have illuminated the neurobiology underlying depression.