1521-009X/42/12/2023–2032$25.00 http://dx.doi.org/10.1124/dmd.114.059675DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 42:2023–2032, December 2014Copyright ª 2014 by The American Society for Pharmacology and Experimental Therapeutics

Disposition and Metabolic Profiling of [14C]Cerlapirdine UsingAccelerator Mass Spectrometry

Susanna Tse, Louis Leung, Sangeeta Raje, Mark Seymour, Yoko Shishikura, and R. Scott Obach

Pfizer, Inc., Groton, Connecticut (S.T., L.L., R.S.O.), and Collegeville, Pennsylvania (S.R.); Xceleron, Inc., Germantown, Maryland(M.S., Y.S.) and University of Dundee, Scotland, UK (Y.S.)

Received June 26, 2014; accepted September 12, 2014

ABSTRACT

Cerlapirdine (SAM-531, PF-05212365) is a selective, potent, full an-tagonist of the 5-hydroxytryptamine 6 (5-HT6) receptor. Cerlapirdineand other 5-HT6 receptor antagonists have been in clinical devel-opment for the symptomatic treatment of Alzheimer’s disease. Ahuman absorption, distribution, metabolism, and excretion studywas conducted to gain further understanding of the metabolismand disposition of cerlapirdine. Because of the low amount ofradioactivity administered, total 14C content and metabolic profilesin plasma, urine, and feces were determined using acceleratormass spectrometry (AMS). After a single, oral 5-mg dose of[14C]cerlapirdine (177 nCi), recovery of total 14C was almostcomplete, with feces being the major route of elimination of theadministered dose, whereas urinary excretion played a lesser role.The extent of absorption was estimated to be at least 70%.

Metabolite profiling in pooled plasma samples showed that un-changed cerlapirdine was the major drug-related component incirculation, representing 51% of total 14C exposure in plasma. Onemetabolite (M1, desmethylcerlapirdine) was detected in plasma,and represented 9% of the total 14C exposure. In vitro cytochromeP450 reaction phenotyping studies showed that M1 was formedprimarily by CYP2C8 and CYP3A4. In pooled urine samples, threemajor drug-related peaks were detected, corresponding tocerlapirdine-N-oxide (M3), cerlapirdine, and desmethylcerlapirdine.In feces, cerlapirdine was the major 14C component excreted,followed by desmethylcerlapirdine. The results of this studydemonstrate that the use of the AMS technique enables compre-hensive quantitative elucidation of the disposition and metabolicprofiles of compounds administered at a low radioactive dose.

Introduction

Serotonin (5-hydroxytryptamine, 5-HT) is a major neurotransmitterin both the central and peripheral nervous systems. At least seven5-HT receptor families (5-HT1–7) have been identified (Hoyer et al.,1994). Among the 5-HT receptor families, the 5-HT6 receptor islocalized primarily in the central nervous system and in regions suchas the cerebral cortex, nucleus accumbens, caudate-putamen, striatum,and hippocampus (Roberts et al., 2002; Hirst et al., 2003; Marazzitiet al., 2012) that are associated with cognition and memory (Nyberget al., 1996; Reinvang et al., 1998; Burgess et al., 2001). Inexperimental models, 5-HT6 antagonists have demonstrated the abilityto increase brain levels of acetylcholine and glutamate (Dawson et al.,2000; Riemer et al., 2003; Marcos et al., 2006) and to improvecognitive functions in animal models (Lindner et al., 2003; Foleyet al., 2004; Lieben et al., 2005; Arnt et al., 2010; Mohler et al., 2012).Several 5-HT6 antagonists have been evaluated in clinical trials for thetreatment of cognitive deficits and dementia associated withAlzheimer’s disease (Upton et al., 2008; Johnson et al., 2008;Geldenhuys and Van der Schyf, 2009). 5-HT6 antagonists also havebeen proposed to have utility as an adjunctive treatment of cognitivedysfunction in schizophrenia (Roth et al., 2004; Schreiber et al., 2006)

and other neurological diseases (Liu and Robichaud, 2009; Rosse andSchaffhauser, 2010; Codony et al., 2011). Recent clinical evidencefurther illustrates the potential utility of 5-HT6 antagonists in thesymptomatic treatment of Alzheimer’s disease (Maher-Edwards et al.,2010, 2011).Cerlapirdine (SAM-531, PF-05212365 [N,N-dimethyl-3-{[3-

(naphthalen-1-ylsulfonyl)-2H-indazol-5-yl]oxy}propan-1-amine]) is a se-lective, potent, full antagonist at the 5-HT6 receptor (Liu and Robichaud,2010) with procognitive effects in nonclinical species (Comery et al.,2010). Clinical safety, tolerability, and preliminary efficacy studieshave been conducted (Baird-Bellaire et al., 2010; Brisard et al., 2010;Mitchell, 2011). A study using radiolabeled [14C]cerlapirdine wasconducted in healthy young male subjects to gain a comprehensiveunderstanding of the absorption, metabolism, and excretion propertiesof cerlapirdine. Also, in vitro metabolism and reaction phenotypingstudies were conducted to elucidate the metabolic pathways and theoxidative enzymes responsible for the formation of the major metab-olite that was observed in vivo.Accelerator mass spectrometry (AMS) is an extremely sensitive

method for the detection and determination of isotopic ratios(Salehpour et al., 2008). AMS has been used in pharmaceutic researchand development since 2000 (Garner, 2000), including applications inpharmacokinetics (Sarapa et al., 2005; Hah et al., 2009) and metab-olism studies (Lappin and Stevens, 2008). In some radiotracer studiesdx.doi.org/10.1124/dmd.114.059675.

ABBREVIATIONS: AMS, accelerator mass spectrometry; ADME, absorption, distribution, metabolism, and excretion; AUC, area under theconcentration-time curveHLM, human liver microsomes; HPLC, high-performance liquid chromatography; 5-HT, 5-hydroxytryptamine, serotonin;LC/MS-MS, liquid chromatography coupled with tandem mass spectrometry; P450, cytochrome P450, ; PF-05212365, SAM-531, cerlapirdine,N,N-dimethyl-3-{[3-(naphthalen-1-ylsulfonyl)-2H-indazol-5-yl]oxy}propan-1-amine; RAF, relative activity factor; WAY-211560, 2-amino-3-methyl-5(3-phenylphenyl)-5-pyridin-4-ylimidazol-4-one.

2023

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

in which only very low amounts of radioactivity may be administeredfor various reasons—such as radiolytic instability or long plasma half-lives or retention in tissues—the ultra-high sensitivity of AMS enablesthe detection of very low levels of radioisotopes that may not bedetectable by conventional techniques. In the case of cerlapirdine,before the conduct of the absorption, distribution, metabolism, andexcretion (ADME) study in humans using [14C]cerlapirdine, dosim-etry calculations were performed to evaluate the risk involved withexposure to ionizing radiation in human subjects based on in vivodistribution, pharmacokinetics, and mass balance studies in rats (dataon file). Based on the absorption and distribution of radioactivity tothe whole body and major tissues, which indicated prolonged retentionof radioactivity in pigmented tissues, a single, low, oral radioactivedose of ;200 nCi (actual dose 177 nCi) of [14C]cerlapirdine wasselected for administration in the human study. Because of the lowamount of radioactivity anticipated in plasma and in excreta, whichwere likely to be below the levels that can be reliably measured byliquid scintillation counting, the levels of total 14C in plasma, wholeblood, urine, and feces were determined using AMS. Subsequently,metabolic profiles of [14C]cerlapirdine in plasma, urine, and feceswere determined in pooled samples using high-performance liquidchromatography (HPLC) fractionation followed by AMS analysis offractions (HPLC+AMS).

Materials and Methods

[14C]Cerlapirdine (Fig. 1) was prepared under good manufacturing prac-tices conditions by ABC Laboratories (Columbia, MO). Cerlapirdine and[14C]cerlapirdine were blended before the preparation of the final (capsule)dosage form. Each capsule was hand-filled, weighed, and capped. Each dose waspacked in one capsule per subject, with a target dose of 5.44 mg of thehydrochloride salt or approximately 4.96 mg of active moiety (free base). Eachdose vial contained approximately 200 nCi (;7.4 kBq; the actual meanradioactive dose was determined to be 177 nCi) of [14C]cerlapirdine. Authenticstandards of cerlapirdine, desmethylcerlapirdine (M1, N-desmethyl metabolite[N-desmethylcerlapirdine, N-methyl-3-{[3-(naphthalen-1-ylsulfonyl)-2H-indazol-5-yl]oxy}propan-1-amine]), N-oxide metabolite (M3) of cerlapirdine, [2H6]cerlapirdine(internal standard for cerlapirdine in LC/MS-MS method), [2H6]desmethyl-cerlapirdine (internal standard for desmethylcerlapirdine in the liquid chro-matography coupled with tandem mass spectrometry [LC/MS-MS] method),and WAY-211560 (2-amino-3-methyl-5(3-phenylphenyl)-5-pyridin-4-ylimidazol-4-one, internal standard used for reaction phenotyping study) were obtained fromPfizer Global Research & Development, Pearl River, NY or Princeton, NJ.

The following materials were used for the LC/MS-MS analyses of theplasma, urine, fecal samples and the metabolite profiling and identification.Methanol, acetonitrile, methyl-t-butyl ether, and ethyl acetate were obtained

from Honeywell Burdick & Jackson (Muskegon, MI). Isopropyl alcohol andsodium borate 10-hydrate were obtained from J.T. Baker (Phillipsburg, NJ);formic acid Superpur was obtained from EMD Millipore (Billerica, MA).Materials used for the graphitization process, AMS, and HPLC+AMS analysis,and methanol (HPLC and analytic grades) was obtained from Thermo FisherScientific (Waltham, MA). Liquid paraffin, copper oxide wire (ACS grade),cobalt powder (100 mesh, 99.9%), zinc powder (100 mesh, 99%), and titanium(II) hydride (325 mesh, 98%) were obtained from Sigma-Aldrich (St. Louis,MO). ANU sugar (certified 14C:12C ratio = 1.5061 Times Modern) used as anAMS standard was obtained from the Quaternary Dating Research Centre,Australian National University, Canberra, Australia. Synthetic graphite 200-35mesh (99.99%) was obtained from Alfa Aesar (Ward Hill, MA) and containedaluminum powder (99.99%, obtained from Acros Organics, Morris Plains, NJ).Solid aluminum cathode (used as machine control) was obtained from NationalElectostatics Corporation (Middleton, WI). Graphitization tubes, samples tubes,and borosilicate glass tubes (prebaked at 500�C for 2 to 4 hours) were obtainedfrom York Glassware Services Ltd. (York, United Kingdom), and tin capsulesand Chromosorb W were obtained from Elemental Microanalysis Ltd.(Okehampton, Devon, United Kingdom). Urea standard was obtained fromThermoQuest Corporation (San Jose, CA).

The following materials were used for in vitro cytochrome P450 reactionphenotyping experiments. Sulfaphenazole, tranylcypromine, quinidine, keto-conazole, quercetin, a-naphthoflavone, and diethyldithiocarbamate werepurchased from Sigma-Aldrich. HPLC-grade water, methanol, and acetonitrilewere obtained from E.M. Science (Gibbstown, NJ). All other chemicals werereagent grade or better. Human liver microsomes (pool of 50 male donors) werepurchased from XenoTech (Kansas City, KS). Recombinant human cyto-chrome P450 enzymes CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2C19,CYP2D6, CYP2E1, and CYP3A4 expressed in Escherichia coli were preparedat Pfizer (Collegeville, PA).

Dosing of Subjects. This was an open-label, nonrandomized, single-dosestudy conducted at a single clinical site (PRA International EDS, Zuidlaren, theNetherlands), in accordance with the International Conference on Harmoniza-tion (ICH) guidelines on good clinical practice and with ethical principles as perthe Declaration of Helsinki. The clinical protocol and informed consent formwere reviewed and approved by the institutional review board and independentethics committee before subject enrollment. Subjects were fasted for 10 hoursbefore dose administration, and fasted for 4 hours after dosing. On study day 1,each of six (6) healthy Caucasian male volunteers (age: 19–48 years; mean age:27 years; weight: 66.3–89.0 kg; mean weight: 76.8 kg) was administereda single capsule containing a 5-mg (free-base) oral dose of [14C]cerlapirdine(177 nCi radioactive dose) at approximately 8:00 AM, and received 240 ml ofroom temperature water immediately after dosing. All six study subjects weredischarged on day 15 after the 14-day sample collection and study period.

Collection of Samples. Blood (venous) samples were collected before drugadministration (predose sample) and at 0.5, 1, 1.5, 2, 3, 4, 6, 8, 12, 16, 24, 48,72, 96, 120, 144, 168, 192, 216, 240, 264, 288, 312, and 336 hours after drugadministration. Potassium EDTA was used as an anticoagulant. Each samplewas gently mixed and stored on ice until centrifugation. Aliquots of each bloodsample were centrifuged at 2500 RPM for 10 minutes at 4�C to obtain plasma.The plasma was removed and stored frozen at 270�C/280�C in polypropylenetubes before and after analysis.

Urine samples were collected before drug administration (predose specimen)and at 0–4, 4–8, 8–12, 12–24, 24–48, 48–72, 72–96, 96–120, 120–144, 144–168, 168–192, 192–216, 216–240, 240–264, 264–288, 288–312, 312–336

Fig. 1. Structure of [14C]cerlapirdine (as hydrochloride salt) and the position of the14C label.

TABLE 1

Demographic and baseline characteristics of study subjects

Characteristics Mean (S.D.) Min–Max (Median) n = 6

Age (yr) 27.0 (10.7) 19.0–48.0 (24.5)Baseline height (cm) 183.17 (4.62) 178.0–189.0 (182.5)Baseline weight (kg) 76.8 (8.86) 66.3–89.0 (78.8)Body mass index 22.98 (3.36) 18.76–28.09 (22.65)Gender and ethnicity

Sex Male (100%)Race Caucasian (100%)

2024 Tse et al.

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

hours after dosing. Urine samples were collected on wet ice during eachcollection period, then frozen and stored at 270�C/280�C before and afteranalysis. Fecal samples were collected before drug administration (predosespecimen), then all bowel movements were collected after drug administrationup to 336 hours after dosing. Fecal samples were quantitatively transferred intoa container per subject per 24-hour interval (predose, 0–24, 24–48 hours, etc.)and the total weight for each sample at each interval was recorded. The fecalsamples were homogenized with an Ultra Turax mixer (Ika Works,Wilmington, NC) after adding one to two weight equivalents of water.Aliquots of fecal homogenates were analyzed for metabolite profiling (HPLC +AMS) and metabolite identification (LC/MS-MS). Separate subsamples of fecalhomogenates were freeze-dried and analyzed for 14C content (AMS) in themass balance determination. The total amount of feces collected and the time ofeach defecation relative to the time of drug administration were recorded. Fecalhomogenate and freeze-dried feces samples were stored in plastic containers at270�C/280�C before and after analysis.

Accelerator Mass Spectrometry (AMS) Analyses. Whole blood, plasma,urine, and fecal samples were assayed for total 14C content using AMS atXceleron (York, United Kingdom). Metabolite profiling of pooled plasma,urine, and fecal samples were conducted using HPLC with offline 14C detectionby AMS at Xceleron. For the AMS analysis, plasma (60 ml), whole blood(20 ml), urine (diluted 10-fold with water; 100 ml), feces (homogenized, ~40 mg;freeze-dried, ~4 mg), HPLC fractions (100 ml), or fraction pools (250 ml) wereplaced in tubes containing CuO2. For samples containing little carbon(i.e., urine and HPLC fractions), liquid paraffin (equivalent to ~2 mg carbon)was added as carbon carrier. Samples were sealed into quartz tubes undervacuum, heated for 2 hours at 900�C, oxidizing all carbon in the sample toCO2. CO2 was cryogenically transferred and sealed into a glass tube containingTiH2 and Zn, with cobalt powder as catalyst. Samples were heated for 4 hoursat 500�C, then for 6 hours at 550�C, reducing CO2 to solid carbon. Carbon waspressed into aluminum cathodes, which were placed in the ion source of a5 MV tandem pelletron Accelerator Mass Spectrometer (Model 15SDH-2;National Electrostatics Corporation) and ionized using a Cs+ ion beam, and the14C:12C and 13C:12C ratios were determined. The 14C content of each samplewas calculated, based on the 14C:12C ratio and the total carbon content. Carboncontent was measured using an NA2100 Brewanalyser (CE Instruments,Wigan, United Kingdom), or generic values were used. Where appropriate,predose samples from the same subject were graphitized and analyzed todetermine background 14C:12C ratios.

Mass Balance and Blood:Erythrocyte Distribution. Mass balance wasdetermined based on the recovery of total 14C in urine and feces. Distribution ofblood radioactivity to erythrocytes at selected time points (up to 96 hours afterthe dose) was calculated from 14C contents in blood (Cblood), plasma (Cplasma),and hematocrit (Hct) using the formula: (Cblood 2 Cplasma)(12 Hct)/Hct. Ratiosof plasma cerlapirdine and M1area under the concentration-time curve (AUC)to total plasma 14C AUC were also determined as follows:

Cerlapirdine  or M1 AUC  ratio ¼ AUCcerlapirdine   or AUCM1�AUC

�total  14C

�

LC/MS-MS Analyses of Cerlapirdine and Desmethylcerlapirdine (M1) inHuman Plasma. The concentrations of cerlapirdine and desmethylcerlapirdine(M1) in plasma and urine samples were determined at Advion BioServices(Ithaca, NY) using validated LC/MS-MS assays. Plasma or urine samples (100 ml)were extracted by a liquid-liquid extraction procedure using a 96-well formatto isolate cerlapirdine and desmethylcerlapirdine. [2H6]Cerlapirdine and[2H6]desmethylcerlapirdine were used as internal standards for cerlapirdineand desmethylcerlapirdine, respectively. Sample extracts were injected ona 2.0 � 50 mm, 80 Å, (4 mm particle size) Synergi Hydro-RP column(Phenomenex Inc., Torrance, CA). A mobile phase gradient starting at 80%mobile phase A (10:90:0.01 methanol/water/formic acid) and 20% mobilephase B (90:10:0.01 methanol/water/formic acid) was used. Cerlapirdine anddesmethylcerlapirdine concentrations in plasma extracts were determined bya Sciex API 4000 mass spectrometer (ABSciex, Framingham, MA) with turboion spray in the positive ionization mode. The plasma method has a lowerlimit of quantitation of 0.2 ng/ml and a calibration curve range of 0.2 to100 ng/ml for both analytes. Cerlapirdine and desmethylcerlapirdine con-centrations in urinary extracts were determined by a Sciex API 3000 massspectrometer (ABSciex) with turbo ion spray in the positive ionization mode.The urinary method has a lower limit of quantitation of 10 ng/ml and acalibration curve range of 10–500 ng/ml.

In Vitro Cytochrome P450 Reaction Phenotyping. Experiments wereconducted to determine the intrinsic clearance by substrate depletion ofcerlapirdine or formation of desmethylcerlapirdine in human liver microsomes(HLM). Incubations were conducted at 37�C with shaking on Thermomixer R(Eppendorf, Hamburg, Germany) in 96-well square plates containing (1 mM) inpotassium phosphate buffer (100 mM, pH 7.4), MgCl2 (10 mM), and HLM(1 mg/ml) and were initiated with the addition of a NADPH regenerating system(glucose-6-phosphate, 3.6 mM; NADP+, 1.3 mM; and glucose-6-phosphatedehydrogenase, 0.4 units/ml) in the absence and presence of selective inhibitors

Fig. 2. Cumulative recoveries (mean 6 S.D.) of 14C (urine, feces, and total) aftera single oral 5-mg (177 nCi) dose of [14C]cerlapirdine.

Fig. 3. Plasma concentrations (mean 6 S.D.) versus time profiles of 14C in bloodand plasma after a single oral 5-mg (177 nCi) dose of [14C]cerlapirdine.

Metabolism of Cerlapirdine in Humans 2025

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

of cytochrome P450 enzymes including ketoconazole (1 mM) for CYP3A,quinidine (10 mM) for CYP2D6, sulfaphenazole (10 mM) for CYP2C9,tranylcypromine (50 mM) for CYP2C19, a-naphthoflavone (5 mM) forCYP1A2, tranylcypromine (10 mM) for CYP2A6, quercetin (20 mM) forCYP2C8, or diethyldithiocarbamate (50 mM) for CYP2E1. Reactions werequenched by the addition of 0.5 ml acetonitrile containing the internal standard(WAY-211560). The plates were then centrifuged in a Thermo Forma(Marietta, OH) centrifuge at room temperature for 10 minutes at 3400 rpm.Supernatants (200 ml) were transferred to 1-ml polypropylene 96-well platesfor the determination of the parent drug or M1 concentrations by LC/MS-MS.Similar incubations were conducted in recombinant human cytochrome P450enzymes CYP1A2, CYP2A6, CYP2C8, CYP2C9, CYP2C19, CYP2D6,CYP2E1, or CYP3A4 (100 pmol/ml) expressed in E. coli. The kinetics ofdesmethylcerlapirdine formation were determined in recombinant CYP2C8and CYP3A4 based on results from an initial study, and in HLM at parentdrug concentrations of 1, 5, 10, 25, and 50 mM similarly for an incubationduration of 10 minutes.

HPLC analysis was performed on an Agilent model 1100 HPLC equippedwith a binary pump (Agilent Technologies, Santa Clara, CA) and using an HTSPAL autosampler (LEAP Technologies, Raleigh, NC) maintained at 10�C.Aliquots (10 ml) of the acetonitrile supernatants (containing cerlapirdine,desmethylcerlapirdine, and the internal standard) were injected onto the HPLC.The column was a 5-mm Thermo Aquasil C18 column, 2.1 mm ID � 50 mm(Bellefonte, PA). Mobile phase A was 0.1% formic acid in water, and mobilephase B was 0.1% formic acid in acetonitrile. Mass spectrometry analyses wereconducted using an Applied Biosystems PE Sciex API 4000 triple quadrupolemass spectrometer (Applied Biosystems, Foster City, CA), with a turbo sprayinterface using the positive electrospray ionization mode with electrosprayionization source using a multiple reaction monitoring technique.

Data analysis was performed using Analyst, v.1.4.1 (Applied Biosystems).Peak area ratios for the parent drug against the internal standard were used toexpress the amount of parent remaining compared with that at time 0 in bothHLM with or without chemical inhibitors and in recombinant cytochrome P450(P450) enzymes. Intrinsic clearance by substrate depletion was determined bylinear regression of the log % substrate remaining versus time plots usingMicrosoft Excel, v.7.0 (Microsoft, Redmond, WA).

Concentrations of desmethylcerlapirdine formed were determined byextrapolating the response calculated from the peak-area ratios (peak areas ofdesmethylcerlapirdine versus the internal standard) to that of the standardcurve. Vmax and Km values were determined by nonlinear regression in plotsgenerated from the simple Emax model (model 101) using WinNonlinProfessional, version 4.1 (Pharsight, Mountain View, CA). The intrinsicclearance (Vmax/Km) of desmethylcerlapirdine formation determined ina recombinant human P450 enzyme was converted to the intrinsic clearancein HLM using the relative activity factor (RAF). The RAF is the ratio of thespecific activity in human liver microsomes to the activity by the recombinantP450 enzyme (Nakajima et al., 2002). The specific activity is measured asintrinsic clearance (Vmax/Km) using a specific probe substrate in HLM and inthe recombinant P450 (rP450) enzyme for the P450 enzyme of interest:

RAF ¼ CLint  HLM=CLint   rP450

or

CLint  HLM ¼ RAF� CLint   rP450

The RAF values for CYP2C8 and CYP3A4 determined internally were0.327 and 0.0432 nmol P450/mg protein, respectively. The % contribution ofthe intrinsic clearance by a particular P450 enzyme (CLint,P450i HLM) to totalP450-mediated intrinsic clearance (CLint,P450sum HLM) in human liver maythen be calculated as follows:

%P450i   Contribution ¼ 100%  CLint;  P450  i  HLM�CLint;  P450  sum  HLM

Metabolite Profiling in Plasma, Urine, and Feces. Plasma, urine, and fecalhomogenate samples were pooled across all six subjects at selected time points.A single 0–72 hours plasma pooled sample was prepared from aliquots ofplasma samples at each time point according to the Hamilton method (Hamiltonet al., 1981). Two pooled urine samples (0–96 hours and 96–168 hours) wereprepared by combining volumes of individual time point samples proportional

TABLE2

Pharm

acokinetic

parametersfortotal14Cin

plasmaandwhole

blood,

cerlapirdine

anddesm

ethylcerlapirdinein

plasmaafteradministrationof

asingle

oral

dose

of5mg(177

nCi)of

[14C]cirlapirdineto

healthymalesubjects

Plasma

14C(n

=6)

Blood

14C(n

=6)

Cerlapirdine(n

=6)

Desmethylcerlapirdine(n

=6)

Mean(%

CV)

Min–Max

Median

Mean(%

CV)

Min–Max

Median

Mean(%

CV)

Min–Max

GeometricMean

Mean(%

CV)

Min–Max

GeometricMean

Cmax

(ng/ml)

104(30)

71.3–122

99.0

69.2

(22)

58.5–92.4

64.2

81.8

(34)

50.6–127

78.1

5.00

(50)

3.00

–9.27

4.58

t max

(h)

5.7(27)

4.0–

8.0

6.0

4.0(39)

3.0–

6.0

3.0

3.42

(44)

1.5–

6.0

3.15

12.7

(137)

4–48

7.78

AUCt(ng•h/ml)

5266

(26)

4443

–7266

5191

2607

(26)

1847

–3751

2595

2835

(29)

2022

–4163

2745

473(39)

299–

771

446

AUCinf(ng•h/ml)

5610

(24)

3608

–7568

5475

3216

(25)

2283

–4472

3149

2866

(28)

2048

–4187

2778

507(36)

341–

791

482

t 1/2(h)

90(11)

79.4–106

87.8

37.0

(14)

a29.5–45.5

36.4

59.6

(14)

51.4–72.2

59.1

84.6

(22)

57.8–114

83.0

CL/F

(l/h

perkg)

NC

NA

NA

NC

NA

NA

0.024(21)

0.02

–0.03

0.024

NA

NA

NA

Vz/F(l/kg)

NC

NA

NA

NC

NA

NA

2.07

(28)

1.48

–2.92

2.01

NA

NA

NA

AUCinfratio

bNA

NA

NA

NA

NA

NA

0.51

NA

NA

0.09

NA

NA

CV,coefficientof

variation;

NA,no

tapplicable;NC,no

tcalculated.

aThe

apparent

t 1/2in

bloodmay

notreflectthetrue

term

inal

phasebecausebloodsamples

werenotanalyzed

beyond

96hours.

bAUCinfratio

relativ

eto

total14Cin

plasma.

2026 Tse et al.

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

to the volume collected over each time period for each subject. Two pooledfeces samples (0–96 hours and 96–168 hours) were prepared for each subjectby combining a mass of each individual sample proportional to the total massexcreted over each time period by that subject. An equal proportion of eachindividual subject feces pool was then pooled to provide two cross-subjectpooled samples (0–96 hours and 96–168 hours) for analysis.

Plasma extracts for metabolite profiling were prepared by adding 1 ml ofpooled plasma sample to 3 ml of acetonitrile, vortex mixing, then centrifuging(4500 rpm; 10 minutes; ambient temperature). The supernatant was transferredand evaporated under a stream of nitrogen to approximately 0.6 ml, followed byaddition of 0.4 ml of 5% v/v acetonitrile in 0.2% v/v formic acid. Aliquots ofpooled urine samples were centrifuged, and metabolite profiles weredetermined in the supernatants.

Fecal homogenate samples for metabolite profiling were prepared by mixingapproximately 200 mg of each overall pooled homogenate sample with 1 ml ofwater. Acetonitrile (4 ml) was added to 0.8 ml of the mixture, vortex mixed,and centrifuged. The supernatant was transferred to a clean tube and evaporatedunder a stream of nitrogen to approximately 0.6 ml, followed by adding 0.4 mlof 5% v/v acetonitrile in 0.2% v/v formic acid.

Plasma, urine, and feces metabolite profiling samples were spiked withreference standards (cerlapirdine; desmethylcerlapirdine, M1; and cerlapirdine-N-oxide, M3) then fractionated by HPLC. Metabolite peaks were separated onan YMC ODS-AQ column (4.6 � 250 mm, 5 mm) at a temperature set at 30�C.Mobile phase A consisted of 0.2% v/v formic acid in water, and mobile phase Bconsisted of 0.2% v/v formic acid in acetonitrile. The flow rate was 1 ml/min,and peaks were separated using a stepwise gradient from 15 to 95% mobilephase B over 58 minutes (15 to 20% B from 0 to 20 minutes and held at 20% Bfor 10 minutes, linear gradient to 30% B over 20 minutes, then to 95% B over5 minutes and held at 95% B for 3 minutes before column re-equilibration at15% B). The eluent was collected as a series of fractions at 15-second intervalsacross each run. Fractions were analyzed either individually or pooled thenanalyzed for 14C content by AMS as described earlier.

Metabolite Identification.Metabolite peaks from plasma, urinary, and fecalprofiling samples were collected using a fraction collector, and were analyzedby positive ion HPLC-UV-MS using a Thermo Finnigan LTQ-Orbitrap(ThermoQuest Corporation) in a full scan mode (m/z 200–650) at a resolutionsetting of 30,000. Capillary temperature was set at 275�C, and the source poten-tial was 3600 V. Other potentials were optimized to get optimal ionization andfragmentation of the parent compound. UV absorption spectra were obtained

by an in-line Surveyor photodiode array detector. A Polaris (Agilent) C18column was used (4.6 � 250 mm) with a flow rate of 0.8 ml/min. Mobile phaseA was comprised of 0.1% formic acid, and mobile phase B was comprised ofacetonitrile. The gradient system used was as follows: initially, 10% B held for5 minutes followed by a linear gradient to 50% B at 50 minutes, washing thecolumn at 95% B for 10 minutes, followed by column re-equilibration to initialconditions for 5 minutes. Collision induced dissociation spectra were obtainedfor m/z 410, 396, and 426. Proposed structures were compared with availableauthentic standards of cerlapirdine, desmethylcerlapirdine, and cerlapirdine-N-oxide.

Parent and Metabolite Excretion in Urine and Feces. Amounts ofcerlapirdine and observed metabolites excreted in urine and feces werecalculated by multiplying urine volume or feces weight and cerlapirdine (ormetabolites) concentration in pooled urine or fecal homogenate (adjusted byhomogenate volume) sample at each time interval. The total urinary or fecalexcretion of cerlapirdine (or metabolites) was obtained by summation across alltime intervals. Percentage dose urinary or fecal excretion of cerlapirdine (ormetabolites) was calculated by dividing total amount excreted in urine or fecesby the administered dose and multiplying by 100.

Pharmacokinetic and Statistical Analyses. The pharmacokinetics of total14C, cerlapirdine, and desmethylcerlapirdine in plasma were analyzed usingWinNonlin (version 5.1.1; Pharsight). Pharmacokinetic parameters werecalculated by noncompartmental analysis using the linear-up/log-down trapezoi-dal method. Based on the 336-hour plasma total 14C, cerlapirdine, anddesmethylcerlapirdine concentration data, peak plasma concentration (Cmax),and time at peak plasma concentration (tmax) values were taken directly from theplasma concentration versus time curve for each subject. The terminal-phasedisposition rate constant (lz) was determined by the log-linear regression of atleast three time points (maximum of five time points) that were judged to be inthe terminal phase. The apparent terminal half-life (t1/2) value was estimated bydividing 0.693 by lz. Total area under the concentration-time curve to the lastmeasurable concentration at time t (AUCt) was determined by the log-lineartrapezoidal rule from time 0 to the time of last observed concentration at timet (Ct). Total area under the concentration-time curve from time zero to infinity(AUCinf) was determined by: AUCinf = AUCt + Ct/lz. CL/F was calculated asdose/total area under the concentration-time curve, and Vz/F was determined bythe following: Vz/F = (dose/AUC) � (1/ lz). CL/F and Vz/F were calculated onlyfor the parent compound cerlapirdine.

Fig. 4. Plasma concentrations (mean 6 S.D.) versus time profiles of cerlapirdineand desmethylcerlapirdine after a single oral 5-mg (177 nCi) dose of [14C]cerlapirdine.

Fig. 5. Metabolite profiles in 0–72 hour plasma pooled sample after a single oral5-mg (177 nCi) dose of [14C]cerlapirdine.

Metabolism of Cerlapirdine in Humans 2027

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

Statistical analysis was limited to calculation of summary statistics (e.g.,mean, standard deviation [S.D.], and range) of the pharmacokinetic parametersacross the study subjects. Because this study was conducted in a single studygroup with a single treatment, no statistical comparison was performed.

Results

Excretion and Mass Balance. The study population consisted ofsix healthy, male, Caucasian subjects aged 19 to 48 years, with a meanage of 27 years. The subject demographic and baseline characteristicsare summarized in Table 1. The excretion of 14C in urine, feces, andtotal excretion with time is shown in Fig. 2. After a single oral dose of[14C]cerlapirdine, excretion was essentially complete, with mean(6 S.D.) total 14C recovery of 98.3% (6 4.7%) (range: 91.0–103%)up to 336 hours after the dose. Feces was the major route of excretion,accounting for 70.3% (6 5.1%) (range: 63.8–76.9%) of the admin-istered dose, and urine was the minor excretory route, accounting for28.0% (65.2%) (range: 24.4–37.8%) of the administered dose. Therate of excretion was rapid, with the majority (;68%) of the 14C doserecovered within 96 hours.Pharmacokinetic Profiles of Total 14C, Cerlapirdine, and

Desmethylcerlapirdine (M1). The plasma and blood concentrationsversus time profiles of total 14C are presented in Fig. 3. Pharma-cokinetic parameters of total 14C contents in plasma and blood aresummarized in Table 2. After a single oral dose of 5 mg/177 nCi of[14C]cerlapirdine, 14C was determined in selected time points (up to96 hours) in blood and for all time points (up to 336 hours) in plasma(Fig. 3). Mean peak concentrations (Cmax) of

14C were achieved withmean tmax values of ;3 hours and ;6 hours in blood and plasma,respectively. Mean Cmax values of 14C were 69.2 ng cerlapirdineequivalents/ml in blood and 104 ng eq/ml in plasma. Mean t1/2 of

14Cwas 37.0 hours in blood and 90.0 hours in plasma. It should be notedthat the apparent t1/2 in blood was shorter relative to plasma. This isprobably due to the limited (up to 96 hours only) analysis of bloodsamples such that the terminal phase was not completely character-ized. AUCinf was 3216 ng eq•h/ml in blood and 5610 ng eq•h/mlin plasma. Over the 96-hour sampling period, mean blood/plasmaratios for 14C ranged between 0.59:1 (1 hour) and 0.89:1 (12 hours)after the dose for all subjects, suggesting that [14C]cerlapirdine orrelated metabolites did not preferentially partition in whole bloodcomponents.The pharmacokinetic parameters of cerlapirdine and desmethyl-

cerlapirdine are also summarized in Table 2. The plasma concen-trations versus time profiles are depicted in Fig. 4. After a single oral5-mg dose of [14C]cerlapirdine, cerlapirdine was absorbed into the

systemic circulation with median time to peak plasma concentrations(tmax) of ;3 hours (range: 1.5–6 hours). Mean Cmax and AUCinf forcerlapirdine were ;82 ng/ml and 2866 ng•h/ml, respectively. MeanVz/F and and mean CL/F were 2.1 l/kg and 0.024 l/h per kg, respec-tively, with a mean apparent t1/2 of ;60 hours. For desmethyl-cerlapirdine, median tmax value was ;13 hours (range: 4–48 hours).Mean Cmax and AUCinf values for desmethylcerlapirdine were 5.0 ng/mland 507 ng•h/ml, respectively. Desmethylcerlapirdine was eliminatedfrom the systemic circulation with a mean apparent t1/2 of ;85 hours.Cerlapirdine and desmethylcerlapirdine accounted for 51 and 9%,respectively, of the total 14C in plasma. The ratio of desmethyl-cerlapirdine to cerlapirdine concentrations in plasma, based on theirrespective AUCinf values, was ;0.18.In Vitro Cytochrome P450 Reaction Phenotyping. In HLM, the

substrate depletion or formation of desmethylcerlapirdine was inhib-ited minimally by the various P450 inhibitors evaluated with theexception of quercetin (48–55%) and ketoconazole (54–63%). Theseobservations suggested that CYP2C8 and CYP3A were primarily re-sponsible for the metabolism of cerlapirdine to desmethylcerlapirdine,apparently with similar contribution of these enzymes in HLM.Because quercetin also inhibits CYP3A (albeit to a lower extent), thecontribution of CYP2C8 to desmethylcerlapirdine formation may havebeen somewhat overestimated.Recombinant CYP2C8 and CYP3A4 were also shown to be pri-

marily responsible for the metabolism of cerlapirdine to desmethyl-cerlapirdine, with Vmax values of 3.4 and 9.4 pmol/pmol P450/minfor CYP2C8 and CYP3A4, respectively, and Km values of 3.3 and11.2 mM for CYP2C8 and CYP3A4, respectively. When corrected bytheir corresponding RAF values, the intrinsic clearance (Vmax/Km)values of desmethylcerlapirdine formation were 0.34 and 0.036 ml/mgper min for CYP2C8 and CYP3A4, respectively, reflecting a corre-sponding contribution of ;90% for CYP2C8 by ;10% by CYP3A4.By taking into consideration results from both the recombinant P450and chemical inhibition in HLM, it appeared that the contributionof CYP2C8 and CYP3A to desmethylcerlapirdine formation was48–90% and 10–63%, respectively.

TABLE 3

Amounts (expressed as % of dose administered) of cerlapirdine and metabolitesexcreted (0–336 hours) in urine and feces after administration of a single oral dose of

5 mg (177 nCi) of [14C]cerlapirdine to healthy male subjects

Excreta Time Period Cerlapirdine Desmethylcerlapirdine M3 Unassigneda Totalb

h

Urine 0–96 4.9 3.1 7.1 6.7 21.896–168 0.6 1.5 1.1 0.8 4.00–168 5.5 4.6 8.2 7.5 25.8

Feces 0–96 24.5 14.7 ND 6.6 45.896–168 3.4 5.4 ND 3.3 12.10–168 27.9 20.1 ND 9.9 57.9

Total 0–168 33.4 24.7 8.2 17.4 83.7

ND, not detected.aMinor peaks of 14C in urine and feces homogenate pooled samples; structures could not be

identified.bAdditional 2.2% excreted in urine pooled sample and 12.5% excreted in feces homogenate

pooled sample between 168 and 336 hours.

Fig. 6. Metabolite profiles in 0–96 hour urine pooled sample after a single oral 5-mg(177 nCi) dose of [14C]cerlapirdine.

2028 Tse et al.

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

The reason behind the apparent discrepancy between the chemicalinhibition method and the recombinant P450 method on the contri-bution assessed for CYP2C8 and CYP3A4 is not clear; therefore, arange for their respective contributions was reported. The interindi-vidual variability in the systemic exposure of cerlapirdine in thepresent study is relatively low (25–30% coefficient of variation),which suggests that CYP3A4 may not play a major role in theclearance of cerlapirdine. CYP2C8 genetic polymorphisms have beenreported where the reduced function alleles CYP2C8*2 (found only inAfrican Americans with an allele frequency of 0.18) and CYP2C8*3(occurred primarily in Caucasians with an allele frequency of 0.13)may potentially lead to variability in pharmacokinetics, clinicalresponse, and toxicity (Dai et al., 2001). However, the size of thepresent study (n = 6 subjects) is likely insufficient to detect suchoccurrences if they indeed exist.Metabolite Profiling and Identification in Plasma, Urine, and

Feces. The radiochromatogram from the 0–72 hours AUC plasmapool is depicted in Fig. 5. The major peak (;92% of total sampleradioactivity) coeluted with cerlapirdine. One other peak was ob-served, coeluting with desmethylcerlapirdine. No other metabolite wasidentified in the plasma pooled sample.The amounts of 14C associated with peaks observed in the

radiochromatograms obtained for pooled urine and feces samples aresummarized in Table 3, and the metabolite profiles are depicted inFigs. 6 and 7. In the pooled 0–96 hours urine sample (Fig. 6), themajor peak present (7.1% of the administered dose) coeluted withcerlapirdine-N-oxide (M3). A second peak coeluted with cerlapirdine(4.9% of the administered dose), and another with desmethyl-cerlapirdine (M1; 3.1% of the administered dose). A fourth peak,representing 2.6% of the administered dose, did not align with any ofthe available reference standards.In the 96–168 hours pooled urine sample (data not shown),

desmethylcerlapirdine, cerlapirdine-N-oxide, the unknown peak, andcerlapirdine were observed, representing 1.5, 1.1, 0.6, and 0.8% of theadministered dose, respectively. In the 0–96 hours pooled feces homogenate

radiochromatograms (Fig. 7), cerlapirdine and desmethylcerlapirdine peakswere observed, with the parent compound showing higher abundance(24.5% of the administered dose) compared with the metabolite (14.7%) inthe earlier time point sample and the metabolite showing higher abundance(5.4% compared with 3.4%) in the later time point (96–168 fecal pooledsample, data not shown). Cerlapirdine-N-oxide or any other metabolite peakbesides cerlapirdine and M1 was not observed in either the 0–96 hours orthe 96–168 hours feces homogenate sample pools.Metabolite structures in the collected metabolite peaks from plasma,

urine, and feces homogenate sample pools were further confirmed bycomparing the collision-induced mass spectra and with high resolutionmass spectra of the metabolites to the authentic reference standards.Representative exact mass spectra for cerlapirdine, desmethyl-cerlapirdine, and cerlapirdine-N-oxide observed in fractions collectedfrom the 0–96 hours urine sample pool, and the correspondingauthentic standards are shown in Fig. 8. Similar results were observedin fractions from plasma, fecal homogenates, and the 96–168 hoursurine sample pools. For cerlapirdine, the parent ion at 410.15413(2.1 ppm) possesses the proper exact mass protonated molecular ion.The main fragment ions at m/z 365.09621 (2.1 ppm) represents theloss of dimethylamine, whereas fragments at 219.02271 (C10H7N2O2S;2.0 ppm) and 173.07123 (C10H9N2O; 1.68 ppm) represent fragmen-tation around both the dimethylamino and sulfone moieties (Fig. 8A).The desmethylcerlapirdine (M1) peak, which elutes on the liquidchromatography column slightly ahead of the parent compound,shows a parent ion at m/z 396.13779 and is consistent with anempirical formula of C21H22N3O3S (0.38 ppm). The fragment ions at365.09582, 219.02249, and 173.07098 are the same as for the parentdrug (Fig. 8B).Cerlapirdine-N-oxide, M3, was observed only in urine. The mass

spectrum of metabolite in urine fraction collected at a retention timealigned with the N-oxide shows a protonated molecular ion ofm/z 426.14812, which is consistent with an empirical formula ofC22H24N3O4S (20.20 ppm). The later retention time than parent isconsistent with an N-oxide metabolite on the dimethylamino nitrogen;however, the daughter spectrum was not informative.The pooled urine radiochromatogram shows a small peak not

associated with the parent drug, desmethyl, or N-oxide metabolites ata retention time of approximately 37.5 minutes, which may suggestthe presence of another metabolite. This peak accounted for ;10 and15% of the total 14C recovered for the 0–96 hours and the 96–168hours sample pools, respectively. However, the mass spectral data didnot offer enough information to propose a structure for this material.

Discussion

Cerlapirdine is a 5-HT6 receptor antagonist that has been evaluatedas a potential treatment of cognitive deficits in mild to moderateAlzheimer’s disease (Brisard et al., 2010). As part of the clinicaldevelopment program, a human ADME study with [14C]cerlapirdinewas conducted to elucidate the metabolism and disposition ofcerlapirdine. Because of the long retention time of cerlapirdine inpigmented tissues, a low radioactive dose (#200 nCi) was admin-istered. Because of the low amount of radioactivity that is generallybelow the detection limit of conventional detection method such asliquid scintillation counting, we used accelerator mass spectrometry todetermine the 14C drug-related material in the plasma, urine, and fecalsamples obtained from this study. The overall results show thatcerlapirdine is well absorbed (at least 70% based on amounts of un-changed cerlapirdine in feces and amounts of metabolites excreted inurine and feces) after oral administration. Recovery of 14C was almostcomplete (mean total recovery was 98.3%) within 14 days; feces was

Fig. 7. Metabolite profiles in 0–96 hour pooled fecal homogenate pooled sampleafter a single oral 5-mg (177 nCi) dose of [14C]cerlapirdine.

Metabolism of Cerlapirdine in Humans 2029

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

the major route of excretion, accounting for ;70% of the total dose,and urinary excretion was a minor route, with mean excretion of;28% of the dose. The rate of excretion was rapid, with majority(;68%) of the dose recovered within 96 hours.Metabolite profiling results showed that unchanged cerlapirdine

was the major drug-related compound in circulation, accounting forapproximately 51% of the total 14C in plasma. One circulatingmetabolite, desmethylcerlapirdine (M1), was identified. The ratio ofdesmethylcerlapirdine relative to cerlapirdine was 0.18 (based onAUCinf values). Total desmethylcerlapirdine was approximately 9% ofthe total 14C in plasma. Evaluation of the partial AUC data between0 and 72 hours also shows that cerlapirdine and M1 AUC0–72 hours

values were approximately 80% of the 14C AUC0–72 hours in plasma(data not shown), indicating that the parent compound and M1accounted for a majority of the total drug-related materials, eventhough the overall AUC values show that the parent and M1metabolite accounted for ;60% of total 14C AUC (0–336 hours). This

discrepancy is primarily due to low levels of 14C in plasma samplesbeyond 72 hours (concentrations ,15 ng eq/ml). The differencebetween the total 14C AUC and the measured cerlapirdine and M1may be due to other minor metabolites (with abundance lower thandesmethylcerlapirdine since no definitive peak has been observed) inplasma. The recovery of total 14C from pooled plasma was ;72%,which was similar to the extraction recoveries of cerlapirdine anddesmethylcerlapirdine from plasma when evaluated by an LC/MS-MSassay (data on file).It is possible that some of the minor metabolites of cerlapirdine may

not be quantitatively extracted and therefore not identified in thisstudy. In vitro metabolism studies in HLM have shown that otheroxidative metabolites (besides M1 and the N-oxide, M3) are formed(data on file). Other oxidative metabolites have also been observedin metabolism studies in nonclinical species in vitro and in vivo(data on file). However, in earlier metabolite profiling studies aftersingle or multiple doses of cerlapirdine in humans, cerlapirdine and

Fig. 8. HPLC/UV/HRMS-MS analysis of cerlapirdine, M1, and M3 in fractions obtained from 0–96 hours urine sample. (A) UV chromatograms of fractions containingcelapirdine, desmethylcelapirdine (M1), and celapirdine-N-oxide (M3). (B) High-resolution mass spectra for celapirdine, M1, and M3 metabolites.

2030 Tse et al.

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

desmethylcerlaperidine were the major drug-related materials incirculation, whereas cerlapirdine N-oxide was observed in traceamount after multiple dosing only (data on file). It is possible that lowlevels of these other oxidative metabolites (observed in HLM in vitroand in nonclinical species) may be present, but the abundance of thesewould be relatively low compared with cerlapirdine and desmethyl-cerlapirdine because no definitive metabolite peak was observed ineither the single/multiple dose metabolite profiling studies or in thisstudy. If present, the abundance of these metabolites would berelatively low compared with cerlapirdine and desmethylcerlapirdine.It should also be noted that no other metabolites besides cerlapirdine,M1, and M3 have been observed in the pooled urine and fecal samplesfrom later time points (96–168 hours); therefore, the presence ofanother metabolite at later time points (beyond 72 hours) is not likely.The N-oxide metabolite (M3) was the major drug-related entity

excreted in urine, along with desmethylcerlapirdine and unchangedcerlapirdine. Unchanged cerlapirdine was the predominant compoundexcreted in the feces, followed by desmethylcerlapirdine. No othermetabolite was identified in the feces. CYP2C8 and CYP3A wereidentified as the major P450 isozymes responsible for the formation ofdesmethylcerlapirdine in HLM or using recombinant enzymes. Boththe desmethyl and the N-oxide metabolites were also identified from invitro systems using HLM or S9 fractions (data on file). The overallmetabolism and disposition pathways of cerlapirdine are summarizedand depicted in Fig. 9.The AMS technique has been used in this metabolism study with

satisfactory outcomes. However, the samples were pooled across timepoints and subjects because of the low amounts of radioactivity and14C, thus time-related changes in the metabolite profiles and variabilitybetween subjects could not be assessed in this study. It should benoted that the metabolism of cerlapirdine is uncomplicated, with onemajor (desmethyl) and one minor (N-oxide) metabolite identified inplasma and in excreta. Other minor metabolites, if present, would

likely occur at very low amounts, precluding sufficient characteriza-tion and identification.Since its introduction, AMS has increasingly been used in phar-

maceutic research and development (Lappin and Stevens, 2008;Garner, 2010; Lappin and Seymour, 2010). The ultrahigh sensitivityof AMS (by measure isotopic 14C rather than radioactivity) enables theuse of low amounts of radioactivity in radiotracer pharmacokinetic andmetabolism studies. This is particularly important when the compoundunder evaluation shows radiolytic instability, has to be administered asa very low dose, or the compound exhibits long systemic and/or tissuehalf-lives such that the total amount of radioactivity that can be safetyadministered is limited. In a conventional human ADME study usingliquid scintillation counting as a quantitation method, the typicalradioactivity doses are in the range of 10–100 mCi. However, in thecase of ixabepilone, which shows radiolytic instability at high specificactivity but is stable at 1–2 nCi/mg (Comezo�glu et al., 2009), AMSprovided a tool by which the mass balance and metabolism of thiscompound could be determined in humans. Other examples of usingAMS for human mass balance and metabolism studies includevismodegib (Graham et al., 2011), which has a long half-life (over 200hours), and pasireotide, which is a somatostatin analog administered ata low dose of 600 mg (Lin et al., 2013).With the low amount of radioactivity administered in studies using

AMS as the detection method, the radiotracer study may be exemptfrom regulatory review by health authorities because the radioactiveexposure from the administered dose is comparable to ambient expo-sure. As a result, pharmacokinetics/ADME study can be conductedduring early phase clinical development, and valuable insights can begained and may result in an expedited development program (Garneret al., 2002; Garner 2010). This is particularly important in the case ofidentifying potential human metabolites that may require additionalsafety testing in animals and additional monitoring in clinical studies(Lappin and Seymour, 2010). In the case of cerlapirdine, long retention

Fig. 9. Proposed metabolic and elimination pathwaysof cerlapirdine, and the amounts (expressed as % ofdose administered) of cerlapirdine, desmethylcerlapirdine,and cerlapirdine-N-oxide excreted in urine and fecesover 168 hours.

Metabolism of Cerlapirdine in Humans 2031

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from

in pigmented tissues was observed in rodent tissue distribution studies(data on file), so the radioactivity dose was limited to ,200 nCi.Despite the low amount of radioactivity, the recovery of total 14C wasvirtually complete, and the metabolic and elimination pathways forcerlapirdine also were delineated in this study.

Authorship ContributionsParticipated in research design: Tse, Leung, Raje, Seymour.Conducted experiments: Leung, Shishikura, Obach.Performed data analysis: Tse, Leung, Raje, Seymour, Shishikura, Obach.Wrote or contributed to the writing of the manuscript: Tse, Leung, Raje,

Seymour, Shishikura, Obach.

References

Arnt J, Bang-Andersen B, Grayson B, Bymaster FP, Cohen MP, DeLapp NW, Giethlen B,Kreilgaard M, McKinzie DL, and Neill JC, et al. (2010) Lu AE58054, a 5-HT6 antagonist,reverses cognitive impairment induced by subchronic phencyclidine in a novel object recog-nition test in rats. Int J Neuropsychopharmacol 13:1021–1033.

Baird-Bellaire SJ, Wan H, Vandenhende F, Plotka A, Poola N, Danjou P, and Chalon S (2010)The effects of a single dose administration of SAM-531, a 5-HT6 antagonist, on sleep elec-troencephalogram (EEG) and quantitative wake EEG (QEEG) in healthy subjects (Abstract).Alzheimers Dement 6(4, Suppl)S510.

Brisard C, Safirstein B, Booth K, Hua L, Brault Y, Raje S, and Laventer S (2010) Safety,tolerability, and preliminary efficacy of SAM-531, a 5HT6 antagonist, in subjects with mild-to-moderate Alzheimer’s disease: results from a phase 2A study (Abstract). Alzheimers Dement 6(4, Suppl)S311.

Burgess PW, Quayle A, and Frith CD (2001) Brain regions involved in prospective memory asdetermined by positron emission tomography. Neuropsychologia 39:545–555.

Codony X, Vela JM, and Ramírez MJ (2011) 5-HT(6) receptor and cognition. Curr OpinPharmacol 11:94–100.

Comery TA, Aschmies S, Haydar S, Hughes Z, Huselton C, Kowal D, Kramer A, McFarlane G,Monaphan M, and Smith D, et al. (2010) SAM-531, N,N-dimethyl-3-{[3-(1-naphthylsulfonyl)-1H-indazol-5-yl]oxy}propan-1-amine, a novel serotonin-6 receptor antagonist with preclinicalpro-cognitive efficacy (Abstract). Alzheimers Dement 6(4, Suppl)S548–S549.

Cömezo�glu SN, Ly VT, Zhang D, Humphreys WG, Bonacorsi SJ, Everett DW, Cohen MB, GanJ, Beumer JH, and Beijnen JH, et al. (2009) Biotransformation profiling of [(14)C]ixabepilonein human plasma, urine and feces samples using accelerator mass spectrometry (AMS). DrugMetab Pharmacokinet 24:511–522.

Dai D, Zeldin DC, Blaisdell JA, Chanas B, Coulter SJ, Ghanayem BI, and Goldstein JA (2001)Polymorphisms in human CYP2C8 decrease metabolism of the anticancer drug paclitaxel andarachidonic acid. Pharmacogenetics 11:597–607.

Dawson LA, Nguyen HQ, and Li P (2000) In vivo effects of the 5-HT(6) antagonist SB-271046on striatal and frontal cortex extracellular concentrations of noradrenaline, dopamine, 5-HT,glutamate and aspartate. Br J Pharmacol 130:23–26.

Foley AG, Murphy KJ, Hirst WD, Gallagher HC, Hagan JJ, Upton N, Walsh FS, and Regan CM(2004) The 5-HT(6) receptor antagonist SB-271046 reverses scopolamine-disrupted consoli-dation of a passive avoidance task and ameliorates spatial task deficits in aged rats. Neuro-psychopharmacology 29:93–100.

Garner RC (2000) Accelerator mass spectrometry in pharmaceutical research and development—a new ultrasensitive analytical method for isotope measurement. Curr Drug Metab 1:205–213.

Garner RC (2010) Practical experience of using human microdosing with AMS analysis to obtainearly human drug metabolism and PK data. Bioanalysis 2:429–440.

Garner RC, Goris I, Laenen AAE, Vanhoutte E, Meuldermans W, Gregory S, Garner JV, LeongD, Whattam M, and Calam A, et al. (2002) Evaluation of accelerator mass spectrometry ina human mass balance and pharmacokinetic study-experience with 14C-labeled (R)-6-[amino(4-chlorophenyl)(1-methyl-1H-imidazol-5-yl)methyl]-4-(3-chlorophenyl)-1- methyl-2(1H)-quinolinone(R115777), a farnesyl transferase inhibitor. Drug Metab Dispos 30:823–830.

Geldenhuys WJ and Van der Schyf CJ (2009) The serotonin 5-HT6 receptor: a viable drug targetfor treating cognitive deficits in Alzheimer’s disease. Expert Rev Neurother 9:1073–1085.

Graham RA, Lum BL, Morrison G, Chang I, Jorga K, Dean B, Shin YG, Yue Q, Mulder T,and Malhi V, et al. (2011) A single dose mass balance study of the Hedgehog pathway inhibitorvismodegib (GDC-0449) in humans using accelerator mass spectrometry. Drug Metab Dispos39:1460–1467.

Hah SS, Henderson PT, and Turteltaub KW (2009) Recent advances in biomedical applications ofaccelerator mass spectrometry. J Biomed Sci 16:54 DOI: 10.1186/1423-0127-16-54.

Hamilton RA, Garnett WR, and Kline BJ (1981) Determination of mean valproic acid serum levelby assay of a single pooled sample. Clin Pharmacol Ther 29:408–413.

Hirst WD, Abrahamsen B, Blaney FE, Calver AR, Aloj L, Price GW, and Medhurst AD (2003)Differences in the central nervous system distribution and pharmacology of the mouse5-hydroxytryptamine-6 receptor compared with rat and human receptors investigated byradioligand binding, site-directed mutagenesis, and molecular modeling. Mol Pharmacol 64:1295–1308.

Hoyer D, Clarke DE, Fozard JR, Hartig PR, Martin GR, Mylecharane EJ, Saxena PR,and Humphrey PPA (1994) International Union of Pharmacology classification of receptors for5-hydroxytryptamine (Serotonin). Pharmacol Rev 46:157–203.

Johnson CN, Ahmed M, and Miller ND (2008) 5-HT6 receptor antagonists: prospects for thetreatment of cognitive disorders including dementia. Curr Opin Drug Discov Devel 11:642–654.

Lappin G and Seymour M (2010) Addressing metabolite safety during first-in-man studies using¹⁴C-labeled drug and accelerator mass spectrometry. Bioanalysis 2:1315–1324.

Lappin G and Stevens L (2008) Biomedical accelerator mass spectrometry: recent applications inmetabolism and pharmacokinetics. Expert Opin Drug Metab Toxicol 4:1021–1033.

Lieben CKJ, Blokland A, Sik A, Sung E, van Nieuwenhuizen P, and Schreiber R (2005) Theselective 5-HT6 receptor antagonist Ro4368554 restores memory performance in cholinergicand serotonergic models of memory deficiency in the rat. Neuropsychopharmacology 30:2169–2179.

Lin TH, Hu K, Flarakos J, Sharr-McMahon M, Mangold JB, He H, and Wang Y (2013) As-sessment of the absorption, metabolism and excretion of [14C]pasireotide in healthy volunteersusing accelerator mass spectrometry. Cancer Chemother Pharmacol 72:181–188.

Lindner MD, Hodges DB, Jr, Hogan JB, Orie AF, Corsa JA, Barten DM, Polson C, Robertson BJ,Guss VL, and Gillman KW, et al. (2003) An assessment of the effects of serotonin 6 (5-HT6)receptor antagonists in rodent models of learning. J Pharmacol Exp Ther 307:682–691.

Liu KG and Robichaud AJ (2009) 5-HT6 antagonists as potential treatment for cognitive dys-function. Drug Dev Res 70:145–168 DOI: 10.1002/ddr.20293.

Liu KG and Robichaud AJ (2010) 5-HT6 medicinal chemistry. Int Rev Neurobiol 94:1–34.Maher-Edwards G, Dixon R, Hunter J, Gold M, Hopton G, Jacobs G, Hunter J, and Williams P(2011) SB-742457 and donepezil in Alzheimer disease: a randomized, placebo-controlledstudy. Int J Geriatr Psychiatry 26:536–544.

Maher-Edwards G, Zvartau-Hind M, Hunter AJ, Gold M, Hopton G, Jacobs G, Davy M,and Williams P (2010) Double-blind, controlled phase II study of a 5-HT6 receptor antagonist,SB-742457, in Alzheimer’s disease. Curr Alzheimer Res 7:374–385.

Marazziti D, Baroni S, Pirone A, Giannaccini G, Betti L, Schmid L, Vatteroni E, Palego L,Borsini F, and Bordi F, et al. (2012) Distribution of serotonin receptor of type 6 (5-HT6) inhuman brain post-mortem. A pharmacology, autoradiography and immunohistochemistrystudy. Neurochem Res 37:920–927.

Marcos B, Gil-Bea FJ, Hirst WD, García-Alloza M, and Ramírez MJ (2006) Lack of localizationof 5-HT6 receptors on cholinergic neurons: implication of multiple neurotransmitter systems in5-HT6 receptor-mediated acetylcholine release. Eur J Neurosci 24:1299–1306.

Mitchell ES (2011) 5-HT6 receptor ligands as antidementia drugs. Int Rev Neurobiol 96:163–187.Mohler EG, Baker PM, Gannon KS, Jones SS, Shacham S, Sweeney JA, and Ragozzino ME(2012) The effects of PRX-07034, a novel 5-HT6 antagonist, on cognitive flexibility andworking memory in rats. Psychopharmacology (Berl) 220:687–696.

Nakajima M, Tane K, Nakamura S, Shimada N, Yamazaki H, and Yokoi T (2002) Evaluation ofapproach to predict the contribution of multiple cytochrome P450s in drug metabolism usingrelative activity factor: effects of the differences in expression levels of NADPH-cytochromeP450 reductase and cytochrome b(5) in the expression system and the differences in the markeractivities. J Pharm Sci 91:952–963.

Nyberg L, McIntosh AR, Cabeza R, Habib R, Houle S, and Tulving E (1996) General andspecific brain regions involved in encoding and retrieval of events: what, where, and when.Proc Natl Acad Sci USA 93:11280–11285.

Riemer C, Borroni E, Levet-Trafit B, Martin JR, Poli S, Porter RH, and Bös M (2003) Influenceof the 5-HT6 receptor on acetylcholine release in the cortex: pharmacological characterizationof 4-(2-bromo-6-pyrrolidin-1-ylpyridine-4-sulfonyl)phenylamine, a potent and selective 5-HT6receptor antagonist. J Med Chem 46:1273–1276.

Reinvang I, Magnussen S, Greenlee MW, and Larsson PG (1998) Electrophysiological locali-zation of brain regions involved in perceptual memory. Exp Brain Res 123:481–484.

Roberts JC, Reavill C, East SZ, Harrison PJ, Patel S, Routledge C, and Leslie RA (2002) Thedistribution of 5-HT(6) receptors in rat brain: an autoradiographic binding study using theradiolabelled 5-HT(6) receptor antagonist [(

125)I]SB-258585. Brain Res 934:49–57.Rossé G and Schaffhauser H (2010) 5-HT6 receptor antagonists as potential therapeutics forcognitive impairment. Curr Top Med Chem 10:207–221.

Roth BL, Hanizavareh SM, and Blum AE (2004) Serotonin receptors represent highly favorablemolecular targets for cognitive enhancement in schizophrenia and other disorders. Psycho-pharmacology (Berl) 174:17–24.

Salehpour M, Possnert G, and Bryhni H (2008) Subattomole sensitivity in biological acceleratormass spectrometry. Anal Chem 80:3515–3521.

Sarapa N, Hsyu P-H, Lappin G, and Garner RC (2005) The application of accelerator massspectrometry to absolute bioavailability studies in humans: simultaneous administration of anintravenous microdose of 14C-nelfinavir mesylate solution and oral nelfinavir to healthy vol-unteers. J Clin Pharmacol 45:1198–1205.

Schreiber R, Sleight A, and Woolley M (2006) 5-HT receptors as targets for the treatment ofcognitive deficits in schizophrenia, in The Serotonin Receptors: From Molecular Pharma-cology to Human Therapeutics (Roth BL, ed) pp 495–515, Humana Press, Totowa, NJ.

Upton N, Chuang TT, Hunter AJ, and Virley DJ (2008) 5-HT6 receptor antagonists as novelcognitive enhancing agents for Alzheimer’s disease. Neurotherapeutics 5:458–469.

Address correspondence to: Susanna Tse, Pharmacokinetics, Dynamics andMetabolism-New Chemical Entities, Pfizer Inc., Eastern Point Road, Groton,CT 06340. E-mail: susanna.tse@pfizer.com

2032 Tse et al.

at ASPE

T Journals on A

pril 15, 2017dm

d.aspetjournals.orgD

ownloaded from