at SciVerse ScienceDirect

Neuropharmacology 72 (2013) 88e95

Contents lists available

Neuropharmacology

journal homepage: www.elsevier .com/locate/neuropharm

Therapeutic doses of antidepressants are projected not to inhibithuman a4b2 nicotinic acetylcholine receptors

Mark L. Weber a,1, Corry M. Hofland b,1, Christopher L. Shaffer a, Gunnar Flik b,Thomas Cremers b, Raymond S. Hurst a,2, Hans Rollema c,*

aNeuroscience Research Unit, Worldwide Research & Development, Pfizer Inc., 700 Main Street, Cambridge, MA 02139, USAbBrains On-Line, Groningen, P.O. Box 4030, 9701 EA Groningen, The NetherlandscRollema Biomedical Consulting, 20 Holdridge Court, Mystic, CT 06355, USA

a r t i c l e i n f o

Article history:Received 21 December 2012Received in revised form20 March 2013Accepted 10 April 2013

Keywords:Antidepressantsa4b2 Nicotinic acetylcholine receptorHuman unbound brain drug concentrationsNortriptylineMecamylamineVarenicline

* Corresponding author. Tel./fax: þ1 860 536 0169.E-mail address: hans.rollema@gmail.com (H. Rolle

1 Joint first authors (contributed equally to the stu2 Present address: Department of Neuroscience, Am

Cambridge, MA 02142, USA.

0028-3908/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.neuropharm.2013.04.027

a b s t r a c t

Inhibition of central a4b2 nAChRs by antidepressants, proposed to contribute to their clinical efficacy,was assessed for monoamine reuptake inhibitors (amitriptyline, nortriptyline, fluoxetine, sertraline,paroxetine, citalopram) by comparing projected human unbound brain drug concentrations (Cu,b) attherapeutic doses with concentrations that inhibit human a4b2 nAChRs in vitro. Inhibitory concentra-tions (IC50) were determined by patch clamp and ranged from 0.8e3.2 mM, except for nortriptyline(IC50 ¼ 100 nM). Cu,b values were calculated from human unbound plasma drug concentrations (Cu,p) andrat-derived brain-to-plasma and extracellular fluid-to-plasma ratios for the unbound drug, which arenear unity, due to much higher brain tissue binding than plasma protein binding of these drugs.Accordingly in humans, antidepressant Cu,b are projected to essentially equal Cu,p, with average valuesfrom 3e87 nM, which are 30-to-250-fold below their IC50 concentrations.

Based on our model, monoaminergic antidepressants minimally inhibit central nAChRs and it isunlikely that a4b2 nAChR antagonism contributes to their antidepressant activity. Nortriptyline is anexception with a Cu,b that is 2-fold below its IC50, which is comparable to the nAChR antagonist(�)-mecamylamine, for which Cu,b is 4-fold below its IC50; both drugs will inhibit a substantial fraction ofa4b2 nAChRs. The Cu,b of the a4b2 nAChR partial agonist varenicline, which has antidepressant-likeactivity in a murine model, is higher than its IC50 and varenicline is projected to cause w70% inhibi-tion of a4b2 nAChRs. Taken together these data may help explain the negative outcome of recent an-tidepressant augmentation trials with mecamylamine and the partial agonist CP-601927.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Modulation of the cholinergic system by decreasing the func-tional activity of nicotinic acetylcholine receptors (nAChRs) hasbeen proposed as a potential novel antidepressant mechanismbased on several observations that are consistent with the cholin-ergiceadrenergic hypothesis, which postulated that hyperactivityof the cholinergic system contributes to depression (Janowsky et al.,1972). These findings (reviewed by Shytle et al., 2002; Mineur andPicciotto, 2010; Philip et al., 2010) include: the association betweendepression and smoking in the context of nAChR desensitization by

ma).dy).gen, Inc., 360 Binney Street,

All rights reserved.

nicotine, the mood-improving effect of the non-selective nAChRantagonist (�)-mecamylamine in patients with Tourette’s Disorderand in treatment-resistant depressed patients, and theantidepressant-like activity of several nAChR antagonists and par-tial agonists in rodent models. Additional support for the role ofnAChRs in depression is the fact that all currently used antide-pressants, which act through monoamine transporters, are alsonon-competitive nAChR antagonists (Fryer and Lukas, 1999; López-Valdés and García-Colunga, 2001; Shytle et al., 2002; Arias et al.,2006; Bianchi, 2008; Arias et al., 2010a,b). Although the inhibi-tory potencies of most antidepressants are reported to be modest,with inhibitory concentrations (IC50) in the low-to-mid mM range, ithas been proposed that nAChR antagonism can neverthelesscontribute to clinical efficacy, as therapeutic doses of antidepres-sants are expected to result in high drug concentrations in humanbrain (Shytle et al., 2002; Bianchi, 2008; Arias et al., 2010a). This isbased largely on the high rodent-derived total brain-to-plasma

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e95 89

ratio (Cb/Cp), for these drugs, which does not take into accountpotential species differences in Cb/Cp and the “free drug” hypothesis(Tillement et al., 1988). Additionally, estimates of human unboundbrain concentrations (Cu,b) of these molecules have generallyassumed that brain tissue binding is similar to plasma proteinbinding. This is however not the case for the vast majority ofneurotherapeutics, including these antidepressants, since theirbinding to brain tissue is often much greater than to plasma pro-teins (Summerfield et al., 2007; Liu et al., 2009; Shaffer, 2010; Diet al., 2011; Doran et al., 2012). Assuming an identical unboundbrain drug concentration-to-unbound plasma drug concentration(Cu,b/Cu,p) in humans as in rats, for compounds showing net passivediffusion at the bloodebrain barrier in rodents (Doran et al., 2012),the projected steady-state Cu,b of the studied antidepressants attherapeutic doses are expected to be much lower and possiblyinsufficient to significantly inhibit central nAChRs.

The present study examined this concept in detail by comparingprojected human Cu,b of currently used antidepressants with theirIC50 for inhibiting human a4b2 nAChRs in vitro. Study test com-pounds were two tricyclic antidepressants (TCAs; amitriptyline,nortriptyline), four selective serotonin reuptake inhibitors (SSRIs;fluoxetine, sertraline, paroxetine, citalopram) and, for comparison,two potent inhibitors of nAChRs: the non-selective nAChR channelblocker (�)-mecamylamine (Young et al., 2001; Shytle et al., 2002;Bacher et al., 2009) and the a4b2 nAChR partial agonist varenicline(Rollema et al., 2007; 2010). For each drug, its human Cu,b was pro-jected from a reported efficacious human unbound plasma concen-tration and its single-dose rat-derived brain-to-plasma ratio for theunbound drug. This ratio was determined independently by twodifferent methods, an neuropharmacokinetic assay (Doran et al.,2012) and a microdialysis assay (Cremers et al., 2012), and themean of the two values was used for Cu,b calculations. Since the a4b2nAChR has been proposed to be one of the nAChR subtypes involvedin depression (reviewed by Mineur and Picciotto, 2010; Philip et al.,2010), we determined the IC50 of each test compound at the hu-man a4b2 nAChR to predict the extent of inhibition achieved withinthe brain. The results of this study suggest that therapeutic doses ofantidepressants do not substantially inactivate human a4b2 nAChRs.

2. Materials and methods

2.1. Animals

All animal studies were performed in strict accordance with the Care and Use ofLaboratory Animals guidelines (Institute of Laboratory Animal Resources, 1996) andwith Dutch law, using protocols reviewed and approved by the Institutional AnimalCare and Use Committees of Pfizer Worldwide Research and Development (PfizerWRD, Groton, CT) and the University of Groningen (Groningen, The Netherlands).Rats were individually housed in plastic cages on a 12 h light-dark cycle in a tem-perature- and humidity-controlled environment with free access to food and water.Neuropharmacokinetic studies were performed at Pfizer WRD (Groton, CT) in maleSpragueeDawley rats (250e300 g; Charles River Laboratories, Inc. Wilmington,MA); microdialysis studies were performed at Brains On-Line laboratories (Gro-ningen, The Netherlands) in male Wistar and SpragueeDawley rats (280e350 g;Harlan, Horst, The Netherlands).

2.2. Drugs and chemicals

(�)-Nicotine bitartrate, fluoxetine HCl, paroxetine HCl, S-citalopram HCl,(�)-citalopram HCl, amitriptyline HCl and nortriptyline HCl were obtained fromSigmaeAldrich Corp. (St Louis, MO, USA). Sertraline HCl, varenicline tartrate and(�)-mecamylamine HCl were synthesized (>99% chemical purity) at WRD (Groton,CT, USA). All other chemicals and solvents were analytical or HPLC grade and pur-chased from Aldrich Fine Chemical Co. (Milwaukee, WI), Biosolve B.V. (Val-kenswaard, The Netherlands), Fisher Scientific (Pittsburgh, PA), Merck KGaA(Darmstadt, Germany), Thermo Fisher Scientific (Waltham, MA), Mallinckrodt Baker(Phillipsburg, NJ, USA), SAFC Bioscience (Lenexa, KA) or Life Technologies (GrandIsland, NY, USA). Control rat plasma was obtained from Bioreclamation Inc. (Hicks-ville, NY, USA) or internally derived from rats at Brains On-Line. All samples collected

for compound quantification were stored at �70 or �80 �C until processing andbioanalysis.

2.3. Pharmacokinetic studies

2.3.1. Plasma and brain homogenate nonspecific bindingLiterature values were used when available for species-specific plasma (fu,p) and

brain homogenate (fu,b) unbound fractions; rat (fu,p and fu,b) and human (fu,p) valuesfor amitriptyline, (�)-mecamylamine and varenicline were determined by equilib-rium dialysis as described previously (Kalvass and Maurer, 2002). SeeSupplementary material for details.

2.3.2. Rat neuropharmacokinetic studiesThe majority of the rodent neuropharmacokinetic data reported here were

generated previously within Pfizer WRD (Groton, CT) and published (Maurer et al.,2005; Doran et al., 2012). Studies with amitriptyline (20 mg/kg, SC) and(�)-mecamylamine (1 mg/kg, PO) were conducted using established methodology(Doran et al., 2012), details of which are in the Supplementary material. Briefly, afterdosing male SpragueeDawley rats (N ¼ 3/time point, six time points from 0 to 7 hpost-dose), blood samples were collected via cardiac puncture from anesthetizedrats, after which whole brain was extracted and weighed.

2.3.3. Rat microdialysis studiesUltra-slow flow microdialysis studies in male Wistar and SpragueeDawley rats

were conducted as described (Cremers et al., 2009, 2012) and details are given in theSupplementary material. Briefly, two days after implantation of a jugular vein catheterand a guide cannula aimed at the prefrontal cortex under isoflurane anesthesia, a 4mmMetaQuant probe (Brain Link, Groningen, The Netherlands)was inserted into the guidecanula. One to two days later, the probe was perfused with Ringer’s buffer containing0.2% bovine serum albumin (BSA) at slow flow rates of 0.12e0.15 mL/min, while ultra-purified (UP) water without BSA was delivered through the dilution inlet at 0.80 mL/min. After 1 h, a single-dose of test compoundwas administered SC (except vareniclineand (�)-mecamylamine, which were administered PO) and microdialysate wascollected for 4.5 h at 30 min intervals and stored at �80 �C until analysis. Prior to theexperiments, in vitro probe recoveries were determined the same way with a Meta-Quant probe placed in a stirred Ringer solution containing test compound, to checkthat in vitro recoveries were >80% at flow rates of 0.12e0.15 mL/min. Following drugadministration, eight blood samples (100 uL) were taken over a 5 h period, collectedinto heparinized saline and centrifuged to obtain plasma.

2.3.4. Drug quantificationFor bioanalytical sample preparation, plasma and microdialysate were used as

such, while brain tissues were homogenized in a 4-fold volume (w/v) of 60% iPrOHin H2O. Brain tissue samples were processed for the quantification of dosed drugusing acetonitrile (ACN)-mediated matrix precipitation methodology, followed by acharacterized liquid chromatography-tandemmass spectrometry (LC-MS/MS) assay.Individual standard curves were prepared in respective control matrices, anappropriate dynamic range was achieved for all analytical assays, and instrumentsettings and potentials were adjusted to optimize theMS signal for each analyte. Fullexperimental details of the analytical systems, sample processing and compound-specific bioanalytical parameters are summarized in the Supplementary material.

2.3.5. Pharmacokinetic calculationsPharmacokinetic parameters were computed by noncompartmental analyses

using Watson Bioanalytical LIMS Version 7.4 (Thermo Fischer Scientific, Waltham,MA, USA) or WinNonlin� Version 5.2 (Pharsight Corporation, Mountain View, CA,USA). A value of 0 was used when a measured value was below the lower limit ofquantification (LLOQ). Means and standard deviations were calculated when at leasthalf of the values were >LLOQ. For each compound, AUC-derived inter-compart-mental neuropharmacokinetic ratios were calculated using AUC0eTlast, where Tlastrepresents the last time point post-dose for which compound concentrations were�LLOQ in all matrices.

For neuropharmacokinetics studies (Doran et al., 2012), no brain perfusionoccurred prior to brain tissue extraction for compound quantification nor was acorrection (Fridén et al., 2011) made to the measured Cb for compound in theanalyzed brain tissue’s capillary blood. Respective compound Cu,b/Cu,p was deter-mined using composite AUC0eTlast values, compound-specific rat fu values and anassumed brain tissue density of 1 g/mL, according to:

�AUCb$fu;b

�.�AUCp$fu;p

�¼ Cu;b=Cu;p (1)

For microdialysis studies (Cremers et al., 2012), respective compound CECF/Cu,pwas determined using mean AUC0eTlast neuromatrix values and rat fu,p, according to:

AUCECF=�AUCp$fu;p

�¼ CECF=Cu;p (2)

Predicted human Cu,b (in nM) was calculated from reported steady-state Cp (innM) at a clinically antidepressant dose, using the respective compound’s human fu,p

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e9590

and the mean of the rat-derived Cu,b/Cu,p and CECF/Cu,p ratios ((Cu,b/Cu,p)rat) accordingto:

Cp$fu;p$�Cu;b=Cu;p

�rat ¼ Cu;b (3)

2.4. Electrophysiology studies

Patch clamp experiments were performed as described previously (Rollema et al.,2007; see Supplementary material for cell culture and electrophysiology details).Briefly, a4b2 nAChR-expressing HEK293 cells were continuously superfused with anexternal buffer containing (inmM): NaCl (143), KCl (5), CaCl2 (2),MgCl2 (1), HEPES (10),and glucose (10), pH of 7.4 and 305-310mOsM. Electrodeswere filledwith intracellularbuffer containing (in mM): CsCH3SO3 (126), CsCl (10), NaCl (4), MgCl2 (1), CaCl2 (0.5),EGTA (5), HEPES (10), MgATP (3), NaGTP (0.3), and phosphocreatine (4), pH 7.2 and290-295 mOsM. Patch clamp experiments were conducted in the whole-cell config-uration at room temperature using pClamp 10 software (Molecular Devices, Sunnyvale,CA, USA), cells were voltage-clamped at �60 mV and exposed for 3 s to nicotine(10 mM) every 60 s. Once nicotine-evoked responses were stable, cells were pretreatedwith test compound for 60 s and then superfused continuously with test compoundwhile maintaining the nicotine challenges until a new stabile amplitudewas achieved.Each test compound was tested at increasing concentrations and IC50 and Hill slopevalues were calculated by fitting a logistic function to the normalized current ampli-tudesmeasured as a function of drug concentration. Amean IC50� SEMwas calculatedfrom the IC50 determined for individual cells.

2.5. Fraction of a4b2 nAChRs that is inactivated by test compounds

Using the inhibition potency (IC50) and Hill slope (nH) that are derived from thelogistic fit equation, the percent of a4b2 nAChRs that is inhibited by the projectedhuman unbound brain concentration of each compound after a clinical dose (Cu,b),can be estimated according to:

% Inhibition ¼ 100��Cu;b

�nH=��Cu;b

�nH þ �IC50

�nH�n o(4)

3. Results

3.1. Pharmacokinetics: projection of human unbound brain drugconcentrations

Rat-derived neuropharmacokinetic parameters are summarized inTable 1. In cases where rat data were not available in the literature,mouse data were used. Fig. 1 exemplifies the time courses of Cu,b (toprow) or CECF (bottom row) and Cu,p for citalopram and (�)-mecamyl-amine from rat neuropharmacokinetic and slow flow microdialysisstudies, respectively. Ratios derived from neuropharmacokinetic

Table 1Pharmacokinetic properties of test compounds in rats and humans used for the estimaefficacious doses. fu,p, fu,b: fraction of drug unbound in plasma or brain homogenate, respratio of unbound drug concentrations in brain and plasma; CECF/Cu,p: ratio of unbound drtotal plasma drug concentrations after clinical doses; Cu,p: range of steady-state humansteady-state human unbound brain drug concentrations after clinical doses.

Drug Rat

fu,p fu,b Cu,b/Cu,p CECF/Cu,p Mean r

Amitriptyline 0.13 0.0071 0.97 e 0.97Citalopram 0.355 0.028 1.08a 1.43b 1.26Fluoxetine 0.031 0.0023 0.89c 1.25 1.07Mecamylamine 0.85 0.31 0.94 1.05 0.99Nortriptyline 0.031 0.005 1.63c 0.85 1.24Paroxetine 0.015 0.006 0.86c 1.29 1.07Sertraline 0.011 0.0007 1.27c 0.96 1.12Varenicline 0.843 0.291 1.12d 0.96d 1.04

a Liu et al. (2009) and Doran et al. (2012).b S-citalopram, ratio for (�)-citalopram ¼ 1.57a.c Mouse data from Maurer et al. (2005).d Rollema et al. (2010) (Cu,b/Cu,p based on Cb/Cp in d).e All human Cp and fu,p data from Thummel et al. (2011) unless indicated otherwise.f Borgå et al. (1969).g Young et al. (2001).h Bacher et al. (2009).

(Cu,b/Cu,p) and slow flow microdialysis (CECF/Cu,p) methodologies arelisted in Table 1. The mean of the two ratios was used for the calcu-lation of human Cu,b values, except for amitriptyline forwhich no CECF/Cu,p value was generated.

Human Cu,b was projected via Eq. (3) using steady-state totalplasma drug concentrations (Cp) at clinically therapeutic doses(Table 1). Since all test compounds’ rat Cu,b/Cu,p and CECF/Cu,p areessentially unity, the assumption of a corresponding Cu,b/Cu,p inhumans is of high-confidence since this relationship has beenshown to exist in nonhuman primates for molecules demonstratingnet passive diffusion across the bloodebrain barrier in rats (Doranet al., 2012). Thus, projected human Cu,b for each test compoundclosely approximates its Cu,p, with Cu,b ranging from amean of 3 nMafter 200 mg/d sertraline to 87 nM after 40 mg/d citalopram(Tables 1 and 3). While clinical doses of mecamylamine and vare-nicline are 1-to-2-orders of magnitude lower than typical antide-pressant doses, their mean projected Cu,b are in the same range(40e56 nM) because both drugs have unbound brain-to-plasmaratios of about 1 and are largely unbound in plasma.

3.2. Pharmacodynamics: inhibition of human a4b2 nAChRs

The functional activity of a4b2 nAChRs was assessed by applyinga 3 s pulse of nicotine to voltage-clamped cells once every minute.After establishing a baseline nicotine-evoked current, test com-pounds were superfused continuously while maintaining therepeated nicotine challenges. Typical examples of theconcentration-dependent effects of two antidepressants (amitrip-tyline and citalopram) on nicotine-evoked currents at a4b2 nAChRsexpressed in HEK293 cells are shown in Fig. 2A. From these data aconcentration-response curve for inhibition of a4b2 nAChRs wasconstructed for each compound (Fig. 2B). Antidepressant drug IC50values ranged from 0.8 mM for sertraline to 3.2 mM for citalopram,except for nortriptyline that is about an order of magnitude morepotent a4b2 nAChR antagonist (IC50 ¼ 0.10 mM) than the otherantidepressants (Table 2). These data are in reasonable agreementwith values that have been published for some of the test com-pounds, while literature data for inhibitory potencies at variousnAChR subtypes suggest that the IC50 values at a3b4, a3b2, a7 anda1 nAChRs are comparable or higher (Table 2). Under these con-ditions the non-selective nAChR antagonist (�)-mecamylamine

tion of unbound brain drug concentrations in humans at recommended clinicallyectively; Cu,p, Cu,b: unbound plasma or unbound brain drug concentration; Cu,b/Cu,p:ug concentrations in extracellular fluid and plasma; Cp: range of steady-state humanunbound plasma drug concentrations after clinical doses; Cu,b: projected range of

Human

atio Dose, mg/d Cp,e nM fu,pe Cu,p, nM Cu,b, nM

100 360e900 0.04f 14e36 14e3540 230e460 0.20 46e92 57e11660 325e1620 0.06 20e97 21e1047.5 50e140g 0.60h 30e84 30e80

125 230e570 0.08 18e46 22e5730 30e100 0.05 2e5 0.1e5

200 80e165 0.02 1.5e3 2e42 20e50 0.80 16e40 17e42

Fig. 1. Typical time courses for unbound compound concentrations in rat brain and plasma, measured by single-dose neuropharmacokinetics (Cu,b and Cu,p, top row, A) and ultra-slow flow microdialysis (CECF and Cu,p, bottom row, B), shown for the SSRI citalopram and the nAChR antagonist mecamylamine.

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e95 91

was found to inhibit human a4b2 nAChRs with an IC50 of 230 nM,and we previously reported an IC50 of 6 nM for the a4b2 nAChRpartial agonist varenicline (Rollema et al., 2007).

The fraction of a4b2 nAChRs that is inhibited by a typical clinicaldose of each compound was estimated from the IC50, Hill slope andprojected mean human Cu,b of the drug after clinical doses ac-cording to Eq. (4) (Table 3). Therapeutic doses (30e200mg/d) of thetested antidepressants are predicted to inhibit <2.5% of a4b2nAChRs, except for nortriptyline that is predicted to inhibit a sub-stantially larger fraction (26%) of a4b2 nAChRs at a dose of 125 mg/d. (�)-Mecamylamine (7.5 mg/d) and varenicline (2 mg/d) areestimated to inhibit about 21% and 72% of human a4b2 nAChRs,respectively.

4. Discussion

According to the cholinergiceadrenergic hypothesis (Janowskyet al., 1972), depression can be caused or exacerbated by hyper-cholinergic tone and, thus, reducing cholinergic signaling mightbe an effective and novel antidepressant mechanism (Shytle et al.,2002; Lippiello et al., 2008; Bianchi, 2008; Arias et al., 2010a,b;

Mineur and Picciotto, 2010; Philip et al., 2010). Several preclinicalstudies on nAChR partial agonists and antagonists (reviewed byMineur and Picciotto, 2010; Philip et al., 2010), as well as initialclinical data on the nAChR antagonist (�)-mecamylamine (Dunbaret al., 2007; George et al., 2008; Bacher et al., 2009) and its enan-tiomer S-mecamylamine (url: ‘Targacept, 2009’) as adjunct therapyin treatment-resistant depressed patients, are consistent with thishypothesis. The notion that blocking nAChRs might result in aclinical improvement in depressed patients appeared to be furthersupported by the finding that currently used antidepressants arealso non-competitive nAChR antagonists with inhibitory potencies(IC50) in the low mM range (see references Table 2). Since thera-peutic doses of antidepressants are expected to result in total braindrug concentrations exceeding their IC50 values, it has been sug-gested that antidepressants can antagonize nAChRs in human brainand that this may be an additional antidepressant mechanism be-sides monoamine reuptake inhibition (Fryer and Lukas, 1999;López-Valdés and García-Colunga 2001; Shytle et al., 2002; Ariaset al., 2006; Bianchi, 2008; Arias et al., 2010a,b). However, basedon the “free drug” pharmacological hypothesis (Tillement et al.,1988), the actual brain drug concentrations available to modulate

Amitriptyline Paroxetine

300

pA

3 sec

10 µM Nicotine + 10 µM Amitriptyline

10 µM Nicotine + 3 µM Amitriptyline

10 µM Nicotine + 1 µM Amitriptyline

10 µM Nicotine + 0.3 µM Amitriptyline

10 µM Nicotine

10 µM Nicotine + 10 µM Paroxetline

10 µM Nicotine + 3 µM Paroxetine

10 µM Nicotine + 1 µM Paroxetine

10 µM Nicotine + 0.3 µM Paroxetine

10 µM Nicotine

0

10

20

30

40

50

60

70

80

90

100

0.0001 0.001 0.01 0.1 1 10 100

Cur

rent

(% o

f nic

otin

e re

spon

se)

Drug concentration (µM)

Amitriptyline Citalopram Fluoxetine Mecamylamine Nortriptyline Paroxetine Sertraline Varenicline

50% inhibition

A

B

Fig. 2. Inhibition of nicotine-evoked currents at human a4b2 nAChRs expressed in HEK293 cells. A. Typical examples of cumulative concentration-dependent inhibition by theantidepressants amitriptyline and paroxetine. B. Concentration-dependent inhibition curves of all test compounds.

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e9592

central nAChR signaling have usually been overestimated, sinceonly unbound brain drug concentrations functionally interact withbrain-based target receptors. The results of this exposure-responsestudy confirm that classical TCA and SSRI antidepressants caninhibit human a4b2 nAChRs, but more importantly, demonstratethat their human Cu,b at therapeutic doses are projected to be 30-to-250-fold lower than their IC50. This suggests that free brainconcentrations of antidepressants typically inhibit only a smallfraction (<3%) of a4b2 nAChRs and that nAChR antagonism con-tributes little, if any, to the clinical efficacy of monoaminergic an-tidepressants. Additionally, our results and recently reportedpreclinical and clinical data on nAChR ligands provide insight intothe relationship between the extent of nAChR inactivation andclinical antidepressant activity.

Before discussing the effects of the test compounds at a4b2nAChRs inmore detail, it is important to point out the limitations ofthis study. Anobvious caveat is our assumption that thehumana4b2nAChR expressed in HEK cells is an in vitro model for the functionalactivity of a4b2 nAChRs in the brain of depressed patients. Althoughthere are no reports of mutant receptors in depressed patients thatare more sensitive to inhibition or desensitization, as described forinstance formutant a4b2 nAChRs in a formof epilepsy that aremoresensitive to inhibition by the channel blocker carbamazepine(Bertrand et al., 2002), it cannot be excluded that subsets ofdepressed patients may have variations of either a4 or b2 nAChRsubunits that alter the pharmacology of antidepressants. While nosuch specific variations are known, there is emerging evidencesuggesting that a variation in CHRNA4 may be associated with

specific forms of depression (Tsai et al., 2012). In addition, differ-ences in stoichiometry of the a4b2 nAChR and/or incorporation ofother subunits can in theory also increase the sensitivity of thosereceptors to inhibition by the test compounds. A second limitation isthat our study is focused on a4b2 nAChRs, while other nAChR sub-types, e.g. a3b4* nAChRs in the medial habenula, have been pro-posed to be involved in the regulation of mood. However, from theliterature (see e.g. Table 2), it is clear that IC50 values of the antide-pressants, and of mecamylamine and varenicline, at a3b4* nAChRsare comparable to their IC50 values at a4b2 nAChRs, suggesting thatthe extent of a3b4* nAChR inhibition by the test compounds will becomparable to the inhibition of a4b2 nAChRs.

4.1. Antidepressants

This study investigated currently used standard antidepres-sants: the TCAs amitriptyline and nortriptyline, and the SSRIsfluoxetine, sertraline, paroxetine and citalopram. For comparison,we included two potent nAChR inhibitors: the non-selectivechannel blocker (�)-mecamylamine (Young et al., 2001; Shytleet al., 2002; Bacher et al., 2009) and the a4b2-selective partialagonist varenicline (Rollema et al., 2007, 2009, 2010). Typical Cpafter clinical doses of the test compounds and their human fu,p weretaken from the literature to calculate human Cu,p. The thirdparameter needed to project human Cu,b from Cu,p, the drug-dependent relationship between these two compartments, wasassessed by two different methods in rats. Neuropharmacokineticstudies measure total drug concentrations in plasma and brain

Table 2Potencies of test compounds (IC50 in mM) to inhibit human a4b2 nAChRs (this study)and published inhibitory potencies at a4b2 nAChRs and other subtypes

Compound IC50 at a4b2 nAChRmean � SEM (n)

IC50 at nAChR subtypesliterature data

Amitriptyline 0.87 � 0.18 mM (6) a4b2: 2.2 mMa

Citalopram 3.03 � 1.20 mM (6)

Fluoxetine 1.01 � 0.14 mM (4) a4b2: 2.0e2.5 mMb,c

a3b4: 4.7e11.2 mMb,c,d

a1: 2.1 mMc

Mecamylamine 0.23 � 0.03 mM (6) a4b2: 0.6e3.0 mMa,e

a3b4: 0.09e1.18 mMe,f,g

a3b4a5: 0.44e0.62 mMg

a3b2: 1.2e3.6 mMe

a7: 1.6e6.9 mMe

Nortriptyline 0.10 � 0.02 mM (7)

Paroxetine 1.51 � 0.39 mM (7) a4b2: 8.6 mMb

a3b4: 2.6e4.9 mMb,c

a7: 8.6e22.9 mMb,c,d

a1: 5.6 mMc

Sertraline 0.76 � 0.17 mM (6) a3b4: 3.1 mMb

a7: 2.1e11.1 mMb,c,d

a1: 3.5 mMd

Varenicline 0.006 � 0.002 mM (5)h a3b4: 0.22 mMi

a7: 0.08 mMj

a Arias et al. (2010a,b).b Arias et al. (2010b).c Fryer and Lukas (1999).d Feuerbach et al. (2005).e Papke et al. (2001).f Giniatullin et al. (2000).g Tammimäki et al. (2012).h Rollema et al. (2007).i Chatterjee et al. (2011).j Rollema et al. (2010).

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e95 93

homogenate over time following a single-dose, and then use neu-romatrix fu and Eq. (1) to afford an AUC-derived Cu,b/Cu,p (Shaffer,2010; Doran et al., 2012), with Cu,b being a CECF surrogate. Forexperimental efficiency, mouse Cu,b/Cu,p were used for compoundsthat already had such data published (Maurer et al., 2005). A secondmethod, ultra-slow flow microdialysis in rats, measures intracere-bral microdialysis-determined CECF by collecting drug from ECFwith >80% recovery, avoiding the need for recovery correctionsthat can give erroneous results (Cremers et al., 2009, 2012).Together with Cu,p and Eq. (2), the microdialysis assay providesCECF/Cu,p. These two methods generated ratios <2-fold different for

Table 3Inhibitory potencies (IC50), Hill slopes (nH), mean projected steady-state humanunbound brain drug concentrations (mean Cu,b) and fractions of a4b2 nAChRs thatare predicted to be inhibited (finhib a4b2 nAChR) at clinical doses of test compounds.

Compound IC50

(nM)nH Mean Cu,b

(nM)finhib a4b2nAChR (%)

TCAsAmitriptyline 867 1.040 24 2.3Nortriptyline 99 1.111 40 26.7

SSRIsCitalopram 3025 1.350 87 0.8Fluoxetine 1008 1.423 63 1.9Paroxetine 1513 1.190 3 0.06Sertraline 764 1.623 4 0.02

nAChR antagonistMecamylamine 225 0.960 56 20.9

nAChR partial agonistVarenicline 6 0.57 32 71.8

all test compounds (0.9e1.6), supporting the validity and accuracyof rat-derived inter-compartmental ratios, consistent with previousanalyses (Liu et al., 2009; Doran et al., 2012; Cremers et al., 2012).Amitriptyline needs additional microdialysis experiments toconfirm equal readouts for both methods. Furthermore, eachcompound’s Cu,b/Cu,p and CECF/Cu,p were near equilibrium in rats,and such compounds that exhibit net passive diffusion at thebloodebrain barrier in rats afford high-confidence Cu,b projectionsin large animals using Eq. (3) (Shaffer, 2010; Doran et al., 2012).Thus, projected human Cu,b are effectively equal to clinical Cu,p andare in the same range (approximately 2e100 nM) as their bindingaffinities for monoamine transporters (Derijks et al., 2008),consistent with an antidepressant mechanism mediated via inhi-bition of monoamine reuptake.

Inhibitory potencies of various antidepressants at a4b2, a7,a3b4, a3b4a5 or a1 nAChR subtypes have been reported in theliterature (see references Table 2), but datawere not available for alltest compounds at one specific nAChR subtype measured by thesame assay. The a4b2-containing nAChRs are among the subtypesthat have been implicated in depression, based on, among others,the antidepressant-like activity of selective a4b2 nAChR ligands inmurine models; the findings that active compounds are no longeractive in mice lacking the b2 nAChR subunit; and, that a4 nAChRsubunit knock-in mice exhibit increased anxiety (reviewed byMineur and Picciotto, 2010; Philip et al., 2010). In addition, a recentSPECT study showed a lower availability of b2-containing nAChRsin depressed patients than in healthy subjects (Saricicek et al.,2012), while polymorphism in the CHRNA4 gene has been impli-cated in depression in elderlymales (Tsai et al., 2012). Therefore, wechose the human a4b2 nAChR as an appropriate subtype tocompare inhibitory potencies of antidepressants and found thatIC50’s ranged from 0.8e3.2 mM, with the exception of nortriptyline,which has an IC50 of 100 nM (see further below). From these datathe fraction of human a4b2 nAChRs that will be inhibited by clinicaldoses of an antidepressant may be estimated according to Eq. (4)(Table 3). These calculations reveal that typical clinically effectiveantidepressant doses are predicted to inhibit only a minor fraction(0.02e2.3%) of human a4b2 nAChRs. Since inhibitory potencies ofantidepressants at other nAChR subtypes are comparable or lower(Table 2), it may be reasonably assumed that antidepressants willinhibit similar or smaller fractions of a3b4, a3b4a5, a7 or a1nAChRs at their projected Cu,b. Given theminimal inhibition of a4b2and other nAChR subtypes, it is unlikely that antidepressants willsufficiently reduce nAChR signaling to contribute to their clinicalefficacy. This conclusion is based on the assumption that inhibitingsuch a small fraction of nAChRs will not result in clinical activity,but questions the extent of nAChR inhibition required for an anti-depressant effect. Although speculative, the data for nortriptyline,(�)-mecamylamine and two a4b2 nAChR partial agonists, vareni-cline and CP-601927, may provide some insight.

4.2. Nortriptyline, mecamylamine, CP-601927 and varenicline

The TCA nortriptyline is 1-to-2 orders of magnitudemore potentas an a4b2 nAChR antagonist than the other studied antidepres-sants, and, with comparable Cu,b at a 125 mg/d dose, is estimated toinactivate about 25% of a4b2 nAChRs. Hence, nortriptyline com-bines monoamine reuptake inhibition with significant nAChRantagonism and would thus represent an augmentation approachin one molecule, provided that �25% receptor inhibition doesindeed entail antidepressant activity. Although one study foundnortriptyline to be efficacious in treatment-resistant depressedpatients (Nierenberg et al., 2003), there is no clinical evidence thatnortriptyline is more efficacious either as augmentation therapy intreatment-resistant patients or as stand-alone therapy, than other

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e9594

antidepressants (Al-Harbi, 2012; Power et al., 2012), suggestingthat inhibiting 25% of a4b2 nAChRs is insufficient for antidepres-sant augmentation. The finding that clinical doses of nortriptylineare estimated to inhibit a comparable fraction of a4b2 nAChRs as7.5mg/d (�)-mecamylamine (see below)may be relevant in viewofits use as a second-line smoking cessation aid, reported toapproximately double the odds of cessation (Hughes et al., 1998).This activity could be related to its nAChR antagonism since otherTCAs and SSRIs have no benefits to smoking cessation, with theexception of bupropion that is a also a monoamine reuptake in-hibitor with nAChR antagonist properties (Hughes et al., 1998).

The non-selective nAChR antagonist (�)-mecamylamine, atsteady-state clinical concentrations after 7.5 mg/d, is predicted toinhibit approximately 21% of a4b2 nAChRs, whereas its enantiomerS-mecamylamine (TC-5214) will likely inhibit a comparable frac-tion since the pharmacokinetic and functional properties of(�)-mecamylamine and its S-enantiomer do not differ greatly(Young et al., 2001; Papke et al., 2001; Lippiello et al., 2008). Initialclinical studies found that (�)-mecamylamine and S-mecamyl-amine augmented the effect of antidepressants in treatment-resistant depressed patients (Dunbar et al., 2007; George et al.,2008; url: ‘Targacept, 2009’), which seemed to indicate thatinhibiting �20% of central a4b2 nAChRs would be sufficient forclinical efficacy. However, four subsequent Phase 3 trials with S-mecamylamine as augmentation therapy in treatment-resistantdepressed patients did not meet the primary endpoints (url:‘Remaining, 2012’), again suggesting that antagonizing 20e25% ofa4b2 nAChRs does not result in enhanced antidepressant efficacy.

CP-601927, an a4b2 nAChR partial agonist that hasantidepressant-like activity in animal models (Mineur et al., 2011),was investigated in a Phase 2 trial (4 mg/d) as an augmentationstrategy in treatment-resistant subjects with major depressivedisorder, but ultimately lacked efficacy as an adjunct therapy inthese patients (Fava et al., 2012; Ramey et al., 2012). A recent pre-clinical study, employing identical methodology as described here(Hurst et al., 2012), projected that CP-601927 at 4 mg/d will inhibitapproximately 23% of a4b2 nAChRs, comparable to the inhibitioncaused by 7.5 mg/d (�)-mecamylamine. These data further suggestthat inhibiting about 20e25% of a4b2 nAChRs does not translateinto improved antidepressant efficacy.

One caveat regarding these clinical trials is that since treatment-resistant patients represent a different population than most pa-tients with major depression, it is conceivable that even thoughaugmentation strategies did not have a positive outcome intreatment-resistant patients, monotherapy with mecamylamine,CP-601927, varenicline, or with a selective a3b4* nAChR inhibitor,might have efficacy in major depression.

Finally, the a4b2 nAChR partial agonist varenicline can potentlydesensitize nAChRs (Rollema et al., 2007, 2010) after recommendeddoses for smoking cessation (2 mg/d) and is predicted to inhibit>70% of human a4b2 nAChRs. Varenicline has antidepressant-likeactivity in animal models (Rollema et al., 2009) and a small clin-ical trial in smokers with persistent depressive symptoms observedsignificant improvement in mood after adding varenicline to theregular antidepressant treatment (Philip et al., 2009). However, thiswas an open-label study in a very small sample of outpatients and,appropriately, is not unequivocal evidence that at least 70% inhi-bition is required for antidepressant activity. Recently, a vareniclinestudy (2 mg/d) in smokers with major depressive disorder, but onsuccessful antidepressant treatment, which would likely precludefurther mood improvement by co-administration of varenicline,found no difference in depression scores between varenicline- andplacebo-treatments (url: ‘Pfizer, 2012’). To date, no trial data withvarenicline exist from large double-blind studies in treatment-resistant depressed patients.

4.3. Conclusions

Based on estimated unbound brain drug concentrations andmeasured in vitro inhibitory potencies at human a4b2 nAChRs,therapeutic doses of antidepressants are predicted to cause insuf-ficient inhibition of central a4b2 nAChRs to contribute to theirclinical efficacy. In addition, since the fraction of a4b2 nAChRs thatis predicted to be inactivated by the antagonist mecamylamine andthe partial agonist CP-601927 in treatment-resistant depressedsubjects appears not enough for a detectable clinical augmentationeffect, we postulate that much more than 25% of central a4b2nAChRs e or of other subtypes such as a3b4* nAChRs e has to beinhibited to achieve significant augmentation of antidepressantefficacy. This could be accomplished with higher doses, if tolerated,with more potent antagonists, with more potent partial agonistswith very low intrinsic activity (exemplified by varenicline), or withmore potent desensitizing agents, approaches that are activelybeing pursued (e.g. Mineur et al., 2009, 2011; Kozikowski et al.,2009; Zhang et al., 2012). Predicting the extent of human nAChRinhibition at the Cu,b of novel compounds is important for the se-lection of clinical drug candidates and will help to identify theminimal amount of nAChR inhibition required for an antidepres-sant augmentation response.

Conflict of interest statement

The authors declare no conflict of interest.

Acknowledgments

The authors like to thank Dr. R.A. Stratford and C. Vu (XavierUniversity of Louisiana, New Orleans, LA) for help with modeling ofthe slow flow microdialysis data.

Appendix A. Supplementary material

Supplementary material related to this article can be found athttp://dx.doi.org/10.1016/j.neuropharm.2013.04.027.

References

Al-Harbi, K.S., 2012. Treatment-resistant depression: therapeutic trends, challenges,and future directions. Patient Prefer. Adherence 6, 369e388.

Arias, H.R., Bhumireddy, P., Bouzat, C., 2006. Molecular mechanisms and bindingsite locations for noncompetitive antagonists of nicotinic acetylcholine re-ceptors. Int. J. Biochem. Cell. Biol. 38, 1254e1276.

Arias, H.R., Feuerbach, D., Targowska-Duda, K.M., Russell, M., Jozwiak, K., 2010a.Interaction of selective serotonin reuptake inhibitors with neuronal nicotinicacetylcholine receptors. Biochemistry 49, 5734e5742.

Arias, H.R., Rosenberg, A., Targowska-Duda, K.M., Feuerbach, D., Jozwiak, K.,Moaddel, R., Irving, W., Wainer, I.W., 2010b. Tricyclic antidepressants andmecamylamine bind to different sites in the human a4b2 nicotinic receptor ionchannel. Int. J. Biochem. Cell. Biol. 42, 1007e1018.

Bacher, I., Wu, B., Shytle, D.R., George, T.P., 2009. Mecamylamine e a nicotinicacetylcholine receptor antagonist with potential for the treatment of neuro-psychiatric disorders. Expert. Opin. Pharmacother. 10, 2709e2721.

Bertrand, D., Picard, F., Le Hellard, S., Weiland, S., Favre, I., Phillips, H., Bertrand, S.,Berkovic, S.F., Malafosse, A., Mulley, J., 2002. How mutations in the nAChRs cancause ADNFLE epilepsy. Epilepsia 43 (Suppl. 5), 112e122.

Bianchi, M.T., 2008. Non-serotonin anti-depressant actions: direct ion channelmodulation by SSRIs and the concept of single agent poly-pharmacy. Med.Hypotheses 70, 951e956.

Borgå, O., Azarnoff, D.L., Forshell, G.P., Sjöqvista, F., 1969. Plasma protein binding oftricyclic antidepressants in man. Biochem. Pharmacol. 18, 2135e2143.

Chatterjee, S., Steensland, P., Simms, J.A., Holgate, J., Coe, J.W., Hurst, R.S.,Shaffer, C.L., Low, J., Rollema, H., Bartlett, S.E., 2011. Partial agonists of the a3b4*neuronal nicotinic acetylcholine receptor reduce ethanol consumption andseeking in rats. Neuropsychopharmacology 36, 603e615.

Cremers, T.I., de Vries, M.G., Huinink, K.D., Van Loon, J.P., Van der Hart, M., Ebert, B.,Westerink, Ben H.C., De Lange, E.C.M., 2009. Quantitative microdialysis usingmodified ultraslow microdialysis: direct rapid and reliable determination of

M.L. Weber et al. / Neuropharmacology 72 (2013) 88e95 95

free brain concentrations with the MetaQuant technique. J. Neurosci. Methods178, 249e254.

Cremers, T.I., Flik, G., Hofland, C., Stratford, R.E., 2012. Microdialysis evaluation ofclozapine and N-desmethylclozapine pharmacokinetics in rat brain. Drug.Metab. Dispos. 40, 1909e1916.

Derijks, H.J., Heerdink, E.R., Janknegt, R., De Koning, F.H.P., Olivier, B., Loonen, A.J.M.,Egberts, A.C.G., 2008. Visualizing pharmacological activities of antidepressants:a novel approach. Open Pharmacol. J. 2, 54e62.

Di, L., Umland, J.P., Chang, G., Huang, Y., Lin, Z., Scott, D.O., Troutman, M.D.,Liston, T.E., 2011. Species independence in brain tissue binding using brainhomogenates. Drug. Metab. Dispos. 39, 1270e1277.

Doran, A.C., Osgood, S.M., Mancuso, J.Y., Shaffer, C.L., 2012. An evaluation of usingrat-derived single-dose neuropharmacokinetic parameters to project accu-rately large animal unbound brain drug concentrations. Drug Metab. Disp. 40,2162e2173.

Dunbar, G.C., Kuchibhatla, R., Kumar, R., 2007. Mecamylamine in the treatment ofdepressed patients who were inadequate responders to citalopram first-linetherapy; a double-blind placebo controlled add-on study. J. Psychopharmacol.21, A40.

Fava, M., Ramey, T., Bell, J., Li, X., Boyer, S., Davidson, W., Billing, B., Arroyo, S., 2012.Augmenting SSRIs with an a4ß2 nAChR partial agonist: lack of efficacy ininsufficient response Major Depressive Disorder. Poster 31, ISCTM 8th Ann. Sci.Mtg., Washington, DC. https://isctm.org/public_access/Feb_2012/22Feb2012_Posters_Abstracts.pdf.

Feuerbach, D., Lingenhohl, K., Dobbins, P., Mosbacher, J., Corbett, N., Nozulak, J.,Hoyer, D., 2005. Coupling of human nicotinic acetylcholine receptors alpha7 tocalcium channels in GH3 cells. Neuropharmacology 48, 215e227.

Fridén, M., Bergström, F., Wan, H., Rehngren, M., Ahlin, M., Hammarlund-Udenaes, M., Bredberg, U., 2011. Measurement of unbound drug exposure inbrain: modeling of pH partitioning explains diverging results between the brainslice and brain homogenate methods. Drug Metab. Disp. 39, 353e362.

Fryer, J.D., Lukas, R.J., 1999. Antidepressants noncompetitively inhibit nicotinicacetylcholine receptor function. J. Neurochem. 72, 1117e1124.

George, T.P., Sacco, K.A., Vessicchio, J.C., Weinberger, A.H., Shytle, R.D., 2008. Nico-tinic antagonist augmentation of selective serotonin reuptake inhibitor-refractory major depressive disorder: a preliminary study. J. Clin. Psycho-pharmacol. 28, 340e344.

Giniatullin, R.A., Sokolova, E.M., Di Angelantonio, S., Skorinkin, A., Talantova, M.V.,Nistri, A., 2000. Rapid relief of block by mecamylamine of neuronal nicotinicacetylcholine receptors of rat chromaffin cells in vitro: an electrophysiologicaland modeling study. Mol. Pharmacol. 58, 778e787.

Hughes, J.R., Stead, L.F., Lancaster, T., 1998. Antidepressants for smoking cessation.Cochrane Database Syst. Rev. 2007 (1). http://dx.doi.org/10.1002/14651858.CD000031.pub3. Art. No.: CD000031.

Hurst, R.S., Rollema, H., Shaffer, C.L., Coe, J.W., Bertrand, D., 2012. Pharmacologicalprofile of CP-601927 at neuronal nAChRs. Program No. 476.12/NN6.2012Neuroscience Meeting Planner. New Orleans, LA, Soc. Neurosci. Online.

Institute for Laboratory Animal Resources, 1966. Guide for the Care and Use ofLaboratory Animals, Commission on Life Sciences. NRC, Nat Acad Press,Washington D.C.

Janowsky, D.S., Davis, J.M., El-Yousef, M.K., Sekerke, H.J., 1972. Cholinergiceadren-ergic hypothesis of mania and depression. Lancet 300, 632e635.

Kalvass, J.C., Maurer, T.S., 2002. Influence of nonspecific brain and plasma bindingon CNS exposure: implications for rational drug discovery. Biopharm. Drug.Dispos. 23, 327e338.

Kozikowski, A.P., Eaton, J.B., Bajjuri, K.M., Chellappan, S.K., Chen, Y., Karadi, S., He, R.,Caldarone, B., Manzano, M., Yuen, P., Lukas, R.J., 2009. Chemistry and phar-macology of nicotinic ligands based on 6-[5-(azetidin-2-ylmethoxy)pyridin-3-yl]hex-5-yn-1-ol (AMOP-H-OH) for possible use in depression. Chem. Med.Chem. 4, 1279e1291.

Lippiello, P.M., Beaver, J.S., Gatto, G.J., James, J.W., Jordan, K.G., Traina, V.M., Xie, J.,Bencherif, M., 2008. TC-5214 (S-(þ)-mecamylamine): a neuronal nicotinic recep-tor modulator with antidepressant activity. CNS. Neurosci. Ther. 14, 266e277.

Liu, X., Van Natta, K., Yeo, H., Vilenski, O., Weller, P.E., Worboys, P.D.,Monshouwer, M., 2009. Unbound drug concentration in brain homogenate andcerebral spinal fluid at steady state as a surrogate for unbound concentration inbrain interstitial fluid. Drug. Metab. Dispos. 37, 787e793.

López-Valdés, H.E., García-Colunga, J., 2001. Antagonism of nicotinic acetylcholinereceptors by inhibitors of monoamine uptake. Mol. Psychiatry 6, 511e519.

Maurer, T.S., Debartolo, D.B., Tess, D.A., Scott, D.O., 2005. Relationship betweenexposure and nonspecific binding of thirty-three central nervous system drugsin mice. Drug. Metab. Dispos. 33, 175e181.

Mineur, Y.S., Eibl, C., Young, G., Kochevar, C., Papke, R.L., Gündisch, D., Picciotto, M.R.,2009. Cytisine-based nicotinic partial agonists as novel antidepressant com-pounds. J. Pharmacol. Exp. Ther. 329, 377e386.

Mineur, Y.S., Einstein, E.B., Seymour, P.A., Coe, J.W., O’Neill, B.T., Rollema, H.,Picciotto, M.R., 2011. a4b2 nicotinic acetylcholine receptor partial agonists withlow intrinsic efficacy have antidepressant-like properties. Behav. Pharmacol. 22,291e299.

Mineur, Y.S., Picciotto, M.R., 2010. Nicotine receptors and depression: revisiting andrevising the cholinergic hypothesis. Trends Pharmacol. Sci. 31, 580e586.

Nierenberg, A.A., Papakostas, G.I., Petersen, T., Kelly, K.E., Iacoviello, B.M.,Worthington, J.J., Tedlow, J., Alpert, J.E., Fava, M., 2003. Nortriptyline fortreatment-resistant depression. J. Clin. Psychiatry 64, 35e39.

Papke, R.L., Sanberg, P.R., Shytle, R.D., 2001. Analysis ofmecamylamine stereoisomerson human nicotinic receptor subtypes. J. Pharmacol. Exp. Ther. 297, 646e656.

Pfizer: Chantix Safe for Smokers with Depression, 2012. Url: http://www.medpagetoday.com/PrimaryCare/Smoking/35384 (last accessed 12.01.12).

Philip, N.S., Carpenter, L.L., Tyrka, A.R., Price, L.H., 2010. Nicotinic acetylcholine re-ceptors and depression: a review of the preclinical and clinical literature. Psy-chopharmacology (Berl) 212, 1e12.

Philip, N.S., Carpenter, L.L., Tyrka, A.R., Whiteley, L.B., Price, L.H., 2009. Vareniclineaugmentation in depressed smokers: an 8-week, open-label study. J. Clin.Psychiatry 70, 1026e1031.

Power, R.A., Muthén, B., Henigsberg, N., Mors, O., Placentino, A., Mendlewicz, J.,Maier, W., McGuffin, P., Lewis, C.M., Uher, R., 2012. Non-random dropout andthe relative efficacy of escitalopram and nortriptyline in treating majordepressive disorder. J. Psychiatric Res. 46, 1333e1338.

Ramey, T., Bell, J., Boyer, S., 2012. Pharmacological and clinical profile of CP- 601,927:evidence for activity at the nAChR site. Program No. 780.11/KK7.2012 Neuro-science Meeting Planner. New Orleans, LA, Soc. Neurosci. Online.

Remaining TC-5214 Phase III efficacy studies do not meet endpoint, regulatory filingwill not be pursued. Url: http://www.reuters.com/article/2011/11/08/astrazeneca-targacept-idUSL6E7M80SI20111108?type¼companyNews&feedType¼RSS&feedName¼companyNews&rpc¼43 (last accessed 12.01.12).

Rollema, H., Chambers, L.K., Coe, J.W., Glowa, J., Hurst, R.S., Lebel, L.A., Lu, Y.,Mansbach, R.S., Mather, R.J., Rovetti, C.C., Sands, S.B., Schaeffer, E., Schulz, D.W.,Tingley III, F.D., Williams, K.E., 2007. Pharmacological profile of the a4b2nicotinic acetylcholine receptor partial agonist varenicline, an effective smokingcessation aid. Neuropharmacology 52, 985e994.

Rollema, H., Guanowsky, V., Mineur, Y.S., Shrikhande, A., Coe, J.W., Seymour, P.A.,Picciotto, M.R., 2009. Varenicline has antidepressant-like activity in the forcedswim test and augments sertraline’s effect. Eur. J. Pharmacol. 605, 114e116.

Rollema, H., Shrikhande, A., Ward, K.M., Tingley, F.D., Coe, J.W., O’Neill, B.T., Tseng, E.,Wang, E.Q., Mather, R.J., Hurst, R.S., Williams, K.E., De Vries, M., Cremers, T.,Bertrand, S., Bertrand, D., 2010. Pre-clinical properties of the a4b2 nicotinicacetylcholine receptor partial agonists varenicline, cytisine and dianicline translateto clinical efficacy for nicotine dependence. Br. J. Pharmacol. 160, 334e345.

Saricicek, A., Esterlis, I., Maloney, K.H., Mineur, Y.S., Ruf, B.M., Muralidharan, A.,Chen, J.I., Cosgrove, K.P., Kerestes, R., Ghose, S., Tamminga, C.A., Pittman, B.,Bois, F., Tamagnan, G., Seibyl, J., Picciotto, M.R., Staley, J.K., Bhagwagar, Z., 2012.Persistent b2*-nicotinic acetylcholinergic receptor dysfunction in majordepressive disorder. Am. J. Psychiatry 169, 851e859.

Shaffer, C.L., 2010. Defining neuropharmacokinetic parameters in CNS drug dis-covery to determine cross-species pharmacologic exposureeresponse re-lationships. Ann. Rep. Med. Chem. 45, 55e70.

Shytle, R.D., Silver, A.A., Lukas, R.J., Newman, M.B., Sheehan, D.V., Sanberg, P.R.,2002. Nicotinic acetylcholine receptors as targets for antidepressants. Mol.Psychiatry 7, 525e535.

Summerfield, S.G., Read, K., Begley, D.J., Obradovic, T., Hidalgo, I.J., Coggon, S.,Lewis, A.V., Porter, R.A., Jeffrey, P., 2007. Central nervous system drug disposi-tion: the relationship between in situ brain permeability and brain free fraction.J. Pharmacol. Exp. Ther. 322, 205e213.

Tammimäki, A., Herder, P., Li, P., Esch, C., Laughlin, J.R., Akk, G., Stitzel, J.A., 2012.Impact of human D398N single nucleotide polymorphism on intracellular cal-cium response mediated by a3b4a5 nicotinic acetylcholine receptors. Neuro-pharmacology 63, 1002e1011.

Targacept presents data from highly successful Phase 2b trial of TC-5214 asaugmentation treatment for Major Depressive Disorder. Url: http://www.targacept.com/wt/page/pr_1255642681 (last accessed 12.01.12).

Thummel, K.E., Shen, D.D., Isoherranen, N., 2011. Design and optimizations ofdosage regimens: pharmacokinetic data. In: Brunton, L., Chabner, B.,Knollman, B. (Eds.), Goodman & Gilman’s the Pharmacological Basis of Thera-peutics, 12th ed. McGraw-Hill, New York, pp. 1891e1990.

Tillement, J.-P., Urien, S., Chaumet-Riffaud, P., Riant, P., Bree, F., Morin, D., et al., 1988.Blood binding and tissue uptake of drugs. Recent advances and perspectives.Fundam. Clin. Pharmacol. 2, 223e238.

Tsai, S.-J., Yeh, H.-L., Hong, C.-J., Liou, Y.-J., Yang, A.C., Liu, M.-E., Hwang, J.-P., 2012.Association of CHRNA4 polymorphism with depression and loneliness inelderly males. Genes Brain Behav. 11, 230e234.

Young, J.M., Shytle, R.D., Sanberg, P.R., George, T.P., 2001. Mecamylamine: newtherapeutic uses and toxicity/risk profile. Clin. Ther. 23, 532e565.

Zhang, H., Tückmantel, W., Eaton, J.B., Yuen, P.W., Yu, L.F., Bajjuri, K.M.,Fedolak, A., Wang, D., Ghavami, A., Caldarone, B., Paterson, N.E., Lowe, D.A.,Brunner, D., Lukas, R.J., Kozikowski, A.P., 2012. Chemistry and behavioralstudies identify chiral cyclopropanes as selective a4b2-nicotinic acetylcholinereceptor partial agonists exhibiting an antidepressant profile. J. Med. Chem.55, 717e724.