1521-009X/44/5/672–682$25.00 http://dx.doi.org/10.1124/dmd.115.069187DRUG METABOLISM AND DISPOSITION Drug Metab Dispos 44:672–682, May 2016Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

KAE609 (Cipargamin), a New Spiroindolone Agent for the Treatmentof Malaria: Evaluation of the Absorption, Distribution, Metabolism,and Excretion of a Single Oral 300-mg Dose of [14C]KAE609 in

Healthy Male Subjects s

Su-Er W. Huskey, Chun-qi Zhu, Andreas Fredenhagen, Jürgen Kühnöl, Alexandre Luneau,Zhigang Jian, Ziping Yang, Zhuang Miao, Fan Yang, Jay P. Jain, Gangadhar Sunkara,

James B. Mangold, and Daniel S. Stein

Drug Metabolism and Pharmacokinetics (S.W.H., C.Z., Z.J., Z.Y., Z.M., J.P.J., G.S., J.B.M) and Translational Medicine(F.Y., D.S.S.), Novartis Institutes for Biomedical Research, East Hanover, New Jersey; and Global Discovery Chemistry,

Basel, Switzerland (A.F., J.K., A.L.)

Received December 23, 2015; accepted February 19, 2016

ABSTRACT

KAE609 [(19R,39S)-5,79-dichloro-69-fluoro-39-methyl-29,39,49,99-tet-rahydrospiro[indoline-3,19-pyridol[3,4-b]indol]-2-one] is a potent,fast-acting, schizonticidal agent in clinical development for thetreatment of malaria. This study investigated the absorption, distri-bution, metabolism, and excretion of KAE609 after oral administra-tion of [14C]KAE609 in healthy subjects. After oral administration tohuman subjects, KAE609 was the major radioactive component(approximately 76%of the total radioactivity in plasma);M23was themajor circulating oxidative metabolite (approximately 12% of thetotal radioactivity in plasma). Several minor oxidative metabolites(M14, M16, M18, and M23.5B) were also identified, each accountingfor approximately 3%–8% of the total radioactivity in plasma.KAE609 was well absorbed and extensively metabolized, such thatKAE609 accounted for approximately 32% of the dose in feces. Theelimination of KAE609 and metabolites was primarily mediated via

biliary pathways. M23 was the major metabolite in feces. Subjectsreported semen discoloration after dosing in prior studies; there-fore, semen samples were collected once from each subject tofurther evaluate this clinical observation. Radioactivity excreted insemen was negligible, but the major component in semen wasM23, supporting the rationale that this yellow-colored metabolitewas the main source of semen discoloration. In this study, a newmetabolite, M16, was identified in all biologic matrices albeit at lowlevels. All 19 recombinant human cytochrome P450 enzymes werecapable of catalyzing the hydroxylation of M23 to form M16 eventhough the extent of turnover was very low. Thus, electrochemistrywas used to generate a sufficient quantity of M16 for structuralelucidation. Metabolic pathways of KAE609 in humans are sum-marized herein and M23 is the major metabolite in plasma andexcreta.

Introduction

Human malaria can be caused by five protozoan parasites of theplasmodia genera but the majority of all cases are attributable to eitherPlasmodium falciparum or Plasmodium vivax. After a bite from an

infected mosquito, the parasite infects hepatocytes and subsequentlyerythrocytes in the human host. It is the blood parasitemia (the schizontform) and erythrocyte rupture that produces the acute clinical symptoms(spiking fever, rigors, malaise, headache, muscle aches) and vascularcompromise and organ involvement, which results in complicationsthat can be fatal. P. falciparum is more commonly associated withsevere forms of malaria. P. vivax, unlike P. falciparum, maintains ahepatic reserve that can result in subsequent clinical relapses aftersuccessful treatment of the parasitemia.Drug resistance has occurred with prior therapies developed for

malaria and it has been detected to a limited extent for the artemisininclass (Carter et al., 2015; Hott et al., 2015). Despite the current emphasison sole use of combination therapies, it is predicted that, given the

All authors are employees of Novartis. The authors have no other relevantaffiliations or financial involvement with any other organization or entity with afinancial interest in or financial conflict with the subject matter or materialsdiscussed in the manuscript apart from those disclosed. No writing assistancewas utilized in the production of this manuscript.

dx.doi.org/10.1124/dmd.115.069187.s This article has supplemental material available at dmd.aspetjournals.org.

ABBREVIATIONS: %PRA, percentage of radioactivity; ADME, absorption, distribution, metabolism, and excretion; AUC, area underconcentration-time curve; CL/F, apparent systemic (or total body) clearance from plasma (or blood) after extravascular administration; DPM,disintegration per minute; HPLC, high-performance liquid chromatography; ICRP, International Commission on Radiologic Protection; KAE609,(19R,39S)-5,79-dichloro-69-fluoro-39-methyl-29,39,49,99-tetrahydrospiro[indoline-3,19-pyridol[3,4-b]indol]-2-one; LC-MS/MS, liquid chromatography–tandem mass spectrometry; LLOQ, lower limit of quantification; MS, mass spectrometry; m/z, mass-to-charge ratio; NMR, nuclear magneticresonance; P450, cytochrome P450; PK, pharmacokinetics; QC, quality control; SFC, supercritical fluid chromatography; t1/2, terminal elimination half-life; UPLC, ultra-performance liquid chromatography; Vz/F, apparent volume of distribution during the terminal elimination phase after extravascularadministration.

672

http://dmd.aspetjournals.org/content/suppl/2016/02/26/dmd.115.069187.DC1Supplemental material to this article can be found at:

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

current detection in clinical settings of artemisinin resistance, wide-spread artemisinin drug resistance will occur, as has been the historicalexperience for prior antimalarials. Therefore, there is an urgent medicalneed for the development of new antimalarial nonartemisinin-baseddrugs.KAE609 [(19R,39S)-5,79-dichloro-69-fluoro-39-methyl-29,39,49,99-

tetrahydrospiro[indoline-3,19-pyridol[3,4-b]indol]-2-one] (Fig. 1), aspiroindolone, represents a new class of potent, fast-acting, schizo-nticidal agents (Rottmann et al., 2010; Yeung et al., 2010; Meister et al.,2011; Spillman et al., 2013). KAE609 was safe and well tolerated inhealthy subjects up to the highest tested single dose of 300 mg and thehighest tested multiple dose of 150 mg daily for 3 days (Leong et al.,2014). Evaluation in 21 evaluable adult patients with uncomplicatedmalaria due to either P. vivax (n = 10) or P. falciparum (n = 11) showedthat once-daily dosing of KAE609 at 30 mg for 3 days was welltolerated and resulted in a rapid median parasite clearance time ofapproximately 12 hours for both P. vivax and P. falciparum (Whiteet al., 2014). The median fever clearance time was 8 hours and 12 hoursfor P. vivax and P. falciparum, respectively. The preliminary parasiteclearance time and fever clearance time results place KAE609 in thesame range as the standard-of-care artemisinin class of antimalarialssuch as the combination therapy artemether/lumefantrine (Coartem).To facilitate the understanding of metabolism KAE609, we conduted

absorption, distribution, metabolism, and excretion (ADME) studies innonclinical species (Huskey et al., 2016). Three prominent metabolites(M37, M18, and M23) were identified in biologic matrices (Fig. 1).Despite M23 being the major metabolite in dog feces, M23 was notdetected in either rat or dog plasma. In contrast, M23 was identified ininitial human studies to be a metabolite in plasma. In view of potentialmetabolite safety coverage concerns (see U.S. Food and DrugAdministration 2008, 2010, and 2012 guidance, available at http://www.fda.gov/cder/guidance/index.htm and http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm),our strategy was to conduct an investigation of human ADME duringearly development to define the metabolic pathways of KAE609,elucidate potential preclinical safety exposure concerns, and determinethe potential effect of intrinsic and extrinsic factors on KAE609.A single oral dose of 300 mg [14C]KAE609 (65 mCi) in healthy malesubjects was used as a tool to detect, track, and quantify drug andmetabolites in biologic samples.Semen discoloration was noted in prior clinical studies, especially at

higher KAE609 doses; therefore, our strategy was to collect a semensample once from each subject postdose to further evaluate this clinicalobservation from data on the distribution of the parent and metabolites,which would allow assessment of potential risks on sperm.In this article, we describe the definitive quantification of a

prominent metabolite (M23) in human plasma to evaluate metaboliteexposure coverage, the rationale for the source of the yellow color insemen observed in clinics, and the innovative approaches in thestructural elucidation of a new human metabolite (M16) identified inplasma, urine, feces, and semen, albeit at low levels. To our knowledge,this is one of the first examples using an electrochemical approach togenerate a sufficient quantity for structural elucidation of a metabolite.

Materials and Methods

Chemicals

Metabolites M23 and M16 were synthesized and coded as compounds 3 and4, respectively. The investigational drug, [14C]KAE609 (300 mg; total 2.41MBq, 65 mCi, specific activity 0.216 mCi/mg in six capsules of 50 mg), wasprepared by Novartis and supplied to the investigative site as open-labeledindividual subject packs. Chemical reagents (ammonium acetate, formic acid)

and solvents (methanol, isopropranol, and acetonitrile) were obtained fromSigma-Aldrich (St. Louis, MO). OPTI-FLUOR liquid scintillant was purchasedfrom Packard Instrument Co. (Downers Grove, IL). Control blank plasmasamples from three individual male human subjects were purchased fromBioreclamation (Hicksville, NY).

Electrochemical Synthesis of Compound 4 (M16)

For electrochemical oxidation, a one-compartment electrochemical synthesiscell with a volume of 80 ml was used attached to a ROXY potentiostat controlledby Dialogue software (version 2.02.199; Antec Leyden B.V., Zoeterwoude, TheNetherlands). This cell contained a 32-cm2 boron doped diamond (MagicDiamond) working electrode, a platinum wire auxiliary electrode, and a Pd/H2

(HyREF; Antec Leyden B.V.) reference electrode. The buffer solution contained0.9 mM compound 3 dissolved in 1:1 (v/v) acetonitrile/water and 0.5% formicacid as the electrolyte to provide sufficient conductivity for the electrochemicalprocess. A square wave pulse of +1.1 V (2 seconds) and +0.7 V (1 second) wasapplied over 600 minutes at room temperature. The reaction mixture wasconcentrated by removing the acetonitrile in vacuo, adsorbing on a solid-phaseextraction cartridge (Oasis HLB 20 cc; Waters Corp., Milford, MA), andextracting with 1:1 methanol/acetonitrile, followed by lyophilization.

Supercritical fluid chromatography (SFC) purification was achieved on aWaters Prep 100q SFC System equipped with aWaters 2998 PDA detector and aWaters 3100 mass spectrometer utilizing mass spectrometry (MS)–triggeredfractionation. The following equipment and conditions were used: Reprospher100 C18-WCX column, 5 mm, 30 � 250 mm (Dr. Maisch, Ammerbuch-Entringen, Germany); flow rate, 100 ml/min; eluent A, supercritical CO2; eluent B,methanol; gradient, 0 minutes 21% B and 10 minutes 26% B; columntemperature, 40�C; and injection volume, 1 ml in 1:1 methanol/dichloromethane.After solvent removal, 3.3 mg of pure material 4 was obtained, corresponding to ayield of 10%. The characterization of compound 4 by high-resolution MS andnuclear magnetic resonance (NMR) is described below.

Radiation Safety for Study Subjects

The expected radiation exposure of a subject was estimated according toInternational Commission on Radiologic Protection (ICRP; http://www.elsevier.com/wps/find/bookdescription.cws_home/713998/description#description)guidelines. The dosimetry calculations were based on prediction of human organexposure from conservative assumptions, as well as on extrapolation of availabledata on human pharmacokinetics (PK) using nonradiolabeled KAE609 and ratmass balance data using [14C]KAE609. The additional whole-body radiationburden in this study, due to the oral administration of 2.41 MBq (65 mCi) [14C]KAE609, was calculated to be maximally 0.88 mSv. This is below therecommended beyond background dose limit for the public of 1 mSv per yearand is in compliance with the ICRP guidance. The average natural radiationexposure is estimated to be approximately 2.4 mSv per year and approximately2.0 mSv in The Netherlands, where this clinical study was conducted. Forbiomedical investigations in small groups of healthy subjects, an effective doseof 0.1–1.0 mSv is considered to be a minor risk (Category IIa according to ICRP60 and ICRP 62) and therefore was acceptable.

Study Design and Conduct. This was an open-label ADME study with asingle oral dose of [14C]KAE609 (300 mg, 65 mCi). The study protocol wasreviewed and approved by both the radiation safety and ethics committees at theinvestigative site. All subjects provided written informed consent. Subjects hadto be healthy men aged 35–55 years, as indicated by medical history,examination, and standard safety laboratory assessments. They could not havereceived radiation of. 0.2 mSv for diagnostic, therapeutic, or research purposeswithin the prior 12 months. No other medications were allowed during the study.

Six healthy white male volunteers with a median age of 51 years were enrolledand all completed the study evaluations as planned. Dose administration was inthe fasted state. Subjects who met the eligibility criteria at screening wereadmitted for baseline (day21) evaluations. Subjects reported to the study centerat least 18 hours prior to dosing to undergo baseline safety assessment andevaluations to confirm continued eligibility. Subjects remained domiciled fromreporting to the study center on day 21 through the observation and sample(blood, plasma, urine, semen, and feces) collection period of 10–14 dayspostdose. Subjects were not discharged from the clinical research center untiltotal radioactivity recovery was$85% and radioactivity was#5% of theCmax in

ADME of KAE609, a Potent Antimalarial Agent, in Humans 673

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

blood, with a maximum in-house observation period of 4 additional days (days11–14) irrespective of percent recovery.

On the evening of admission, subjects received a standardized dinnerfollowed by a small snack in the evening. Thereafter, the subjects fastedovernight for a minimum of 10 hours. On the morning of dosing (day 1), subjectsunderwent predose safety evaluations and provided blood, plasma, urine, andfecal samples for pharmacokinetic and metabolism assessments as specified inthe assessment schedule. Subjects then received a single oral dose of 300 mg[14C]KAE609, administered as six capsules (50 mg each capsule) with water.Semen collection was done at 6, 12, or 24 hours postdose, with two subjectsrandomized to each time point (i.e., each subject provided one sample postdose).

Safety and pharmacokinetic assessments were performed throughout thestudy at specified time points. Safety assessments included physical examina-tions, ECG results, vital signs, standard clinical laboratory evaluations(hematology, blood chemistry, and urinalysis), and adverse event and seriousadverse event monitoring.

Blood, plasma, urine, feces, and semen were analyzed for total radioactivityand for determination of KAE609 and overall metabolic profiles. Subjects werereleased from the study center after the collection of the last pharmacokinetic andmetabolism samples (blood, plasma, urine, feces, and semen) and after all safetyassessments had been completed (study completion assessments). Urinecollection was stopped if daily recovery in urine was , 1% of the dose on2 consecutive days.

Sample Preparation

Blood, plasma, urine, feces, and semen samples were collected at selectedtime points after a single oral dose of 300 mg [14C] KAE609. All samples werereceived from PRA International (Utrecht, The Netherlands) and stored at280�Cuntil analysis (within 3 months).

Plasma for Quantification of KAE609 and M23. Blood samples weresubjected to centrifugation to obtain plasma samples. Plasma samples at selectedtime points were extracted by protein precipitation using a 96-well plate. A100-ml aliquot of human plasma [blank, zero, standards, quality control (QC)samples, and unknowns] and a 25-ml aliquot of internal standard solution, whichcontained 1000 ng/ml [13C6]KAE609 in 50:50 (v/v) acetonitrile/water solution,were added to each well in a 96-well plate. Then, the plate was vortex mixedbriefly. A 300-ml aliquot of acetonitrile was added to each well and the plate wasvortex mixed for 5 minutes. The assay plate was centrifuged at 1000 � g for10 minutes at 10�C. The supernatant (250 ml) was transferred and evaporatedto dryness under nitrogen at 40�C. The residues were reconstituted in 200 ml 20:80 (v/v) acetonitrile/water by vortex mixing for 10 minutes. A 5-ml aliquot wasinjected onto a selected column for quantification by liquid chromatography–tandem mass spectrometry (LC-MS/MS).

Plasma for Metabolic Profiling of KAE609. Human plasma samples werepooled according to the area under concentration-time curve (AUC) poolingmethod (Hamilton et al., 1981), using samples from 0.5, 1, 2, 3, 4, 6, 8, 24, 36,

48, 72, 96, 120, and 144 hours. The pooled samples (10 ml) were extracted with a40-ml mixture of acetonitrile and methanol [1:1 (v/v)] containing 0.1% aceticacid. Samples were vortex mixed and sonicated for 10 minutes prior tocentrifugation. The supernatants were evaporated to near dryness under a gentlestream of nitrogen on a TurboVap LV (Zymark Corp., Taunton, MA). Theresulting residues were reconstituted with 10 ml 5 mM ammonium formatecontaining 0.1% formic acid in water prior to LC-MS/MS analysis.

Urine and Feces for Metabolic Profiling of KAE609. Urine and fecal poolswere prepared (1% of weight), with samples covering the time interval 0–72 hours postdose. The urine sample was centrifuged at 3500 rpm and 10�C for10 minutes. The supernatant was concentrated to dryness under nitrogen at 35�C.The residue was reconstituted in 10 ml 5 mM ammonium formate with 0.1%formic acid in water prior to LC-MS/MS analysis. Each pooled fecal sample wasextracted three times with 5 volumes of acetonitrile/methanol/acetic acid [50:50:0.1 (v/v/v)]. Each extract was centrifuged at 3500 rpm for 10 minutes and thesupernatants were combined. The combined supernatants were concentratedunder nitrogen. Residues were reconstituted in acetonitrile/deionized water[50/50 (v/v)] prior to LC-MS/MS analysis.

Semen for Metabolic Profiling of KAE609. One semen sample wasobtained from each subject at the selected time points (6, 12, or 24 hours postdose).The sample volume was variable (0.5–2.0 ml) from six subjects. Proteins wereprecipitated by the addition of 5–10ml acetonitrile/methanol (1:1) containing 0.1%acetic acid to semen samples. The samples were centrifuged at 3500 rpm and 10�Cfor 10 minutes. The supernatants were concentrated to dryness under nitrogen at35�C. The residues were reconstituted in 5 ml 5 mM ammonium formatecontaining 0.1% formic acid in water prior to LC-MS/MS analysis.

Fig. 1. Structure of KAE609 and key metab-olites. KAE609 is uniformly labeled in thephenyl ring (indicated by asterisk). The num-bering of M23 and M16 is for the interpretationof NMR data.

Fig. 2. Mean cumulative recovery of total radioactivity in urine and feces fromhuman subjects. Urine and feces samples were collected up to 10–14 days from sixhuman subjects. Aliquots of urine samples were counted directly for radioactivity.After homogenization, aliquots of fecal homogenate were combusted and countedfor radioactivity. The mean cumulative recoveries of radioactivity in urine, feces,and combination are plotted against time.

674 Huskey et al.

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

For calculation of extraction recovery, an aliquot of the reconstituted samplewas assayed by liquid scintillation counting (Packard Instrument Co.).Recoveries of radioactivity were approximately 93%–100% in plasma, 93% inurine extract, 75.5% in fecal extract, and 95% in semen extract, respectively.

Sample Analysis

Determination of Radioactivity. Concentrations of total radioactivity inblood, plasma, urine, fecal homogenate, and semen were measured by liquidscintillation counting. An aliquot of plasma and urine samples was counteddirectly for radioactivity.

An aliquot of blood samples (0.3 ml) was added to 1 ml tissue solubilizer(Solvable; Perkin Elmer, Waltham, MA) and the mixture was incubated at 60�Cin a water bath for 60 minutes. After cooling in the water bath, Titriplex (0.1 M,100 ml) was added and the sample was decolorized by adding three times avolume of 75 ml 30% hydrogen peroxide (VWR International, Radnor, PA).After incubation for 15 minutes at room temperature, the mixture was heatedagain for 30 minutes at 60�C in a water bath. After cooling in a water bath, 18 mlof the scintillation cocktail Ultima Gold (Perkin Elmer) was added. Aftervortexing mixing for at least 5 seconds, the vial was placed in the liquidscintillation analyzer for at least 35 hours before counting.

Fecal samples were weighted and water (one to two weight equivalents) wasadded, and the mixture was homogenized with an Ultra Turrax mixer for at least2 minutes. Four accurately weighted aliquots of approximately 500 mg of thefecal homogenate were dried in a stove at 50�C for at least 3 hours. After theaddition of combustion aid (Perkin Elmer) to the dry homogenates, the sampleswere combusted in a model 307 sample oxidizer (Perkin Elmer). CarboSorb-E(7 ml) was used as absorber agent for carbon dioxide. At the end of thecombustion cycle, the absorber was mixed with 13 ml of the scintillantPermaFluor E. The samples were placed in the liquid scintillation analyzer atleast 60 minutes before counting.

Quantification of KAE609 by LC-MS/MS. Chromatographic separationwas achieved using a Mac-Mod Analytical ACE C8 (50 � 2.1 mm, 5 mm)column (heated at 30�C). The mobile phase consisted of two solvents (A, watercontaining 0.1% formic acid; and B, methanol containing 0.1% formic acid). Thefollowing linear gradient was used: 35% B from 0 to 0.5 minutes, 35% B to 90%B from 0.5 to 2 minutes, hold at 90% B until 3 minutes, 90% B to 35% B from 3to 3.2 minutes, and hold at 35% B until 4 minutes. The flow rate was 0.3 ml/min.

An API4000 mass spectrometer (AB Sciex, Foster City, CA) was used for thequantification of KAE609. Analyses were carried out using electrosprayionization in positive ion mode. The mass spectrometer was operated in multiplereaction monitoring mode. The mass transitions for KAE609 and internalstandard [13C6]KAE609 were monitored from a precursor ion mass-to-chargeratio (m/z) of 390.2 to product ion m/z 347.1 and precursor ion m/z 396.2 toproduct ion m/z 353.1, respectively. The calibration curves ranged from 1 to5000 ng/ml and QC samples were prepared at 2, 5, 400, 2000, and 4000 ng/ml.The assay accuracy and precision met the acceptance criteria set as615%, or620%

TABLE 1

Summary of the PK of KAE609 in human plasma

Blood samples were collected at selected time intervals from each subject. Plasma sampleswere prepared by centrifugation. Total radioactivity in blood and plasma was determined byscintillation counting. After extraction, plasma samples were analyzed by validated LC-MS/MSanalysis. PK parameters were calculated using Phoenix (WinNonlin, version 6.2).

Subject ID tmax CmaxAUC0–24

hAUClast AUCinf t1/2 CL/F Vz/F

h ng/ml mg×h/ml

h l/h liters

5101 8.00 2310 37.8 96.4 96.6 34.2 3.11 1535102 3.00 2290 30.9 96.5 97.0 39.6 3.09 1775103 6.00 1520 22.8 55.3 55.4 27.8 5.42 2175104 3.00 2330 28.0 60.0 60.0 30.5 5.00 2205105 3.00 623 11.3 38.6 38.8 36.4 7.73 4065106 4.00 1590 23.3 53.5 53.9 32.1 5.57 258Mean NA 1780 25.7 66.7 67.0 33.4 4.99 238SD NA 676 8.95 24.1 24.2 4.23 1.74 89.8%CV mean NA 38.0 34.8 36.1 36.1 12.7 35.0 37.7Geomean NA 1627 24.1 63.2 63.4 33.2 4.73 227%CV geomean NA 54.4 43.5 37.2 37.2 12.7 37.2 34.9Median 3.50 1940 25.7 57.7 57.7 33.2 5.21 219Minimum 3.00 623 11.3 38.6 38.8 27.8 3.09 153Maximum 8.00 2330 37.8 96.5 97.0 39.6 7.73 406

NA, not applicable (only the median and range are presented for tmax).

Fig. 3. Representative metabolic profiles inplasma (A), urine (B), feces (C), and semen(D) after a single oral dose of [14C]KAE609.Blood, urine, feces, and semen samples werecollected at selected time intervals from sixhuman subjects (subject 5103 is presented as anexample). After centrifugation, plasma sampleswere pooled according to the AUC poolingmethod from 0 to 120 hours or 0 to 144 hourspostdose. After extraction, the pooled plasmasamples, fecal homogenate, and urine andsemen samples were analyzed by LC-MS/MS.

ADME of KAE609, a Potent Antimalarial Agent, in Humans 675

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

for the lower limit of quantification (LLOQ), percentage of bias, and percentage ofthe coefficient of variation.

Quantification of M23 by LC-MS/MS. Another method was developed forthe quantification of M23 due to the presence of several isomeric metabolites inplasma. Chromatographic separation was achieved using a Mac-Mod AnalyticalACE C18 (50 � 2.1 mm, 5 mm) column (heated at 40�C). The mobile phaseconsisted of two solvents (A, 10 mM ammonium formate containing 0.2%formic acid; and B, acetonitrile containing 0.2% formic acid). The isocraticcondition (40% B) was used at a flow rate of 0.3 ml/min.

An AB Sciex API4000 mass spectrometer was used for the quantification ofM23. Analyses were carried out using electrospray ionization in positive ionmode. Themass spectrometer was operated inmultiple reactionmonitoringmode.The mass transitions for M23 (compound 3) and [13C6]KAE609 were monitoredfrom precursor ionm/z 404.2 to product ionm/z 375.3 and from precursor ionm/z396.2 to product ion m/z 353.2, respectively. The calibration curves ranged from1 to 500 ng/ml and QC samples were prepared at 5, 50, and 500 ng/ml. The assayaccuracy and precision met the acceptance criteria set as 6 15%, or 6 20% forLLOQ, percentage of bias, and percentage of the coefficient of variation.

Metabolic Profiling by LC-MS/MS

High-Performance Liquid Chromatography Method. The plasma, urine,and semen samples were subjected to two-dimensional high-performance liquid

chromatography (HPLC) for sample loading and analysis with offline radioac-tivity detection. The fecal samples were subjected to one-dimensional HPLCanalysis with online or offline radioactivity detection. The HPLC column (YMC-pack, ODS-AQ/S-3mm/12 nm 50 � 4.6 mm ID) was used for loading and themobile phase was 3% acetonitrile in water at a flow rate of 0.6 ml/min. Afterloading, the eluent was switched to a separation column (Zorbax SB-C18column, 150� 3.0 mm, 3.5 mm, at 35�C; Agilent Technologies) at a flow rate of0.5 ml/min. Mobile phases consisted of solvent A [5 mM ammonium formatecontaining 0.1% formic acid] and solvent B [1:1 (v/v) acetonitrile/methanolcontaining 0.1% formic acid]. The linear gradient started at 5% B for 5 minutes,increased to 50% B from 5 to 12 minutes, to 60% B from 12 to 30 minutes, to75% B from 30 to 40 minutes, to 85% B from 40 to 45 minutes, and to 98% Bfrom 45 to 50 minutes. A detailed description of the three HPLC/ultra-performance liquid chromatography (UPLC) methods for comparison ofcompound 4 and M16 from human fecal extract can be found in the Supplemen-tal Materials and Methods.

LC-MS Instrumentation and Operating Conditions. After HPLC separa-tion, metabolites were identified by MS (LTQ-Orbitrap Elite; Thermo FisherScientific, Waltham, MA). The metabolites were interfaced with the WatersAcquity UPLC system (Milford, MA) and an online or offline radioactivitymonitor. HPLC eluent was split 9:1; 450 ml/min was collected into a b-onlineradioactivity monitor or LumaPlates for radioactivity determination, and 50ml/minwas injected into a mass spectrometer for structure elucidation. Qualitative

TABLE 2

Exposure of KAE609 and its metabolites in human plasma after an oral dose of 300 mg [14C] KAE609

Plasma samples were pooled according to the AUC pooling method. The pooled plasma samples were extracted and analyzed by LC-MS/MS for metabolic profiling.

Subject IDExposure of KAE609 and Metabolites in Human Plasma

KAE609 M14 M16 M18 M23 M23 M23.5B

ngEq×h/ml

5101 92,191 11,703 1646 7679 18,843 18,843 30185102a 115,775 2454 ND 14,162 5488 5488 ND5103 61,945 4052 415 6931 12,515 12,515 ND5104 72,620 3042 ND 6369 13,330 13,330 ND5105 40,920 ND ND 4016 3796 3796 ND5106 66,537 4162 ND 7513 16,176 16,176 NDMean 74,998 4236 344 7778 11,691 11,691 503SD 25,951 3958 659 3397 5924 5924 1232Meanb 75.8 3.79 0.28 7.82 11.9 11.9 0.37SD 6.81 2.96 0.50 1.56 4.95 4.95 0.91

ND, not detectable (zero was used for calculation of mean values).aPlasma samples were pooled from all subjects (except 1502) from 0 to 120 hours postdose, and plasma samples from subject 1502 were pooled from 0 to 144 hours.bMean values represented as the percentage of total radioactivity in human plasma.

TABLE 3

Pharmacokinetic parameters of the major metabolite M23 in human plasma

Blood samples were collected at selected time intervals from each subject. Plasma samples were prepared by centrifugation. Total radioactivity in plasma was determined by scintillation counting.After extraction, plasma samples were quantified for KAE609 and M23 by LC-MS/MS analysis. PK parameters were calculated using Phoenix (WinNonlin, version 6.2).

Subject ID tmax Cmax AUC0–24 h AUClast AUCinf t1/2 CL/F Vz/F

h ng/ml mg×h/ml h l/h liters

5101 24.0 390 6.37 27.0 27.1 36.0 11.1 5755102 24.0 90.1 1.64 8.34 8.42 43.0 35.6 22085103 8.00 282 5.51 22.0 22.0 28.2 13.6 5545104 24.0 336 6.49 21.9 21.9 32.0 13.7 6315105 24.0 75.9 1.56 6.68 6.76 38.2 44.4 24475106 24.0 415 7.44 23.5 23.7 34.3 12.7 626Mean NA 265 4.83 18.2 18.3 35.3 21.9 1174SD NA 148 2.58 8.52 8.54 5.09 14.4 898%CV mean NA 56.0 53.4 46.8 46.6 14.4 65.7 76.5Geomean NA 217 4.04 16.0 16.1 35.0 18.6 938%CV geomean NA 89 83.00 66.0 65.5 14.6 65.4 80.3Median 24.0 309 5.94 22.0 22.0 35.1 13.6 629Minimum 8.00 75.9 1.56 6.68 6.76 28.2 11.1 554Maximum 24.0 415 7.44 27.0 27.1 43.0 44.4 2447

NA, not applicable (only the median and range are presented for tmax).

676 Huskey et al.

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

and quantitative analyses were carried out using electrospray ionization inpositive ion mode. The ion spray voltage, capillary voltage, and tube lens offsetwere adjusted to achieve maximum sensitivity using the parent KAE609 inpositive ion mode. The heated capillary temperature was 300�C, the sourcevoltage was 4.5 kV, and the capillary voltage was 45 V. Analyzer Fouriertransform mass spectrometry was set up with resolution 12,000 with massaccuracy maintained less than 3 ppm during the analysis. The MS/MS spectrawere obtained by high-energy collision-induced dissociation of [M + H]+. Theacquired data were analyzed using Xcalibur software (version 2.0; ThermoScientific Inc.).

A detailed description of HPLC/UPLCmethods for comparison of compound4 and M16 from human fecal homogenate and NMR analysis of compound4 (M16) can be found in the Supplemental Materials and Methods.

Data Processing

Quantification of KAE609 and Metabolites by Radiometry. Radioactivepeaks were selected visually from the radiochromatogram. The radioactivity inthe region encompassing the beginning and ending of the peak was summed. Allfurther calculations were based on radioactivity.

The percentage of radioactivity (%PRA) in a particular peak, Z, wascalculated as follows:

%PRA  in  Z ¼ DPM  in  peak  Ztotal DPM  in  all  integrated  peaks

� 100

Where DPM is disintegration per minute. The concentration or amount of eachcomponent was calculated as %PRA (as a fraction) multiplied by the totalconcentration (nanogram equivalent per milliliter) or the percentage of dose inthe excreta (the data were not adjusted for extraction recovery).

Pharmacokinetic Parameters. The pharmacokinetic parameters werecalculated using actual recorded sampling times and noncompartmentalmethod(s) with Phoenix software (WinNonlin, version 6.2; Pharsight, CertaraL.P., Princeton, NJ). Concentrations below the LLOQ were treated as zero forPK parameter calculations. The linear trapezoidal rule was used for AUCcalculation. Regression analysis of the terminal plasma elimination phase for thedetermination of terminal elimination half-life (t1/2) included at least three datapoints after Cmax. If the adjusted R2 value of the regression analysis of theterminal phase was less than 0.75, no values were reported for t1/2, AUCinf,apparent volume of distribution during the terminal elimination phase after

extravascular administration (Vz/F), and apparent systemic (or total body)clearance from plasma (or blood) after extravascular administration (CL/F).

Results

Six healthy white male volunteers with a median age of 51 years wereenrolled and all completed the study evaluations as planned. Nosignificant adverse events or safety concerns were reported. Everysubject reported semen discoloration, even though only two semensamples (subjects 5101 and 5104) were considered by the laboratorypersonnel to have a yellow color.

Mass Balance of KAE609 in Humans

After oral administration of [14C]KAE609 at 300 mg (65 mCi), massbalance was achieved in all six subjects (88.8% 6 3.02% of the dose),with 3.97% and 84.9% of the administered radioactivity dose beingrecovered in urine and feces, respectively (Fig. 2).

PK of KAE609 in Humans

PK parameters of KAE609 in each subject are summarized inTable 1. The plasma concentration of KAE609 declined in a mono-exponential manner. The mean Cmax was approximately 1780 ng/ml,with a median tmax of 3.5 hours (range, 3–8 hours). The CL/F of KAE609was approximately 4.99 l/h. The t1/2 of KAE609 was approximately 33.4hours. The mean Vz/F for KAE609 was 238 liters. The Vz/F valueindicates that KAE609 may not be extensively distributed into tissues.

Metabolite Profiling of [14C]KAE609 in Plasma

According to the AUC pooling method (Hamilton et al., 1981),aliquots of plasma samples at selected time intervals were combined fromeach subject for metabolite profiling. Representative metabolic profilesfrom subject 5103 in all matrices are depicted in Fig. 3. Quantitativeinformation on the metabolites from each subject appears in Table 2. Inall subjects, the major circulating component in plasma was KAE609,accounting for approximately 76% of the total radioactivity AUC0–120 h.Two prominent metabolites were identified to be M18 and M23,accounting for approximately 8% and 12% of total radioactivity in theAUC0–120 h pool, respectively. Several minor oxidative metabolites(M14, M16, and M23.5B) were also identified, each accounting for ,4% of total radioactivity AUC0–120 h. M23.5B and M16 were at tracelevels in plasma from only one and two subjects, respectively.

TABLE 4

Recovery of KAE609 and its metabolites in human excreta (expressed as percentageof dose) after an oral dose of 300 mg [14C] KAE609

The pooled urine and fecal samples were extracted and analyzed by LC-MS/MS for metabolicprofiling. The HPLC separation of KAE609 and metabolites was performed using a Zorbax SB-C18 column, as described in the Materials and Methods.

KAE609/MetaboliteFeces Urine

Mean SD Averageb

M14 0.385 0.943 1.45M16 2.65 0.653 0.45M18 0.392 0.960 NDM20.8 ND ND 0.15M21.8 ND ND NDM23 36.2 12.5 1.4M23.5A/M23.5Ba 2.41 1.19 NDM24 ND ND 0.4M29.8 2.29 1.45 NDM32 2.53 0.780 0.05M33 1.97 1.26 NDM34 1.97 2.17 NDM35.8 ND ND 0.10M48 2.39 0.546 0.05KAE609 31.7 17.1 NDPercentage of dose 85.0 2.47 4.05

ND, not detectable.aM23.5A/M23.5B: two metabolites were coeluted and the recovery represented the mixture of

two metabolites.bThe pooled urine samples were extracted from only two subjects (numbers 5103 and 5104).

TABLE 5

Concentration of KAE609 and its metabolites in human semen after an oral dose of300 mg [14C] KAE609

The semen samples were extracted and analyzed by LC-MS/MS for metabolic profiling. TheHPLC separation of KAE609 and metabolites was performed using a Zorbax SB-C18 column, asdescribed in the Materials and Methods.

Metabolite/SubjectConcentration of KAE609 and Metabolites in Human Semena

5101 5102 5103 5104 5105 5106 Mean SD

ngEq×h/ml

M16 634 287 807 433 231 518 485 216M18 283 203 228 208 162 263 224 44M20.8 ND ND 548 198 ND 189 312 205M21.8 823 123 279 143 ND 173 308 294M23 4699 1640 4023 4050 653 2332 2900 1599M32 ND 75 269 97 219 304 193 102M48 ND 61 331 115 231 287 205 114KAE609 1149 382 1138 475 167 369 613 423Total 7588 2772 7623 5717 1663 4435 4966 2470

ND, not detectable.aSemen samples were collected from subjects 5103 and 5106 at 6 hours postdose, subjects

5101 and 5105 at 12 hours postdose, and subjects 5102 and 5104 at 24 hours postdose.

ADME of KAE609, a Potent Antimalarial Agent, in Humans 677

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

PK of M23, the Major Circulating Metabolite, in Humans

M23 was previously identified as a human circulating metabolite ofKAE609 from a first-in-human study. The structure of M23 (Fig. 1),generated by biotransformation using compound 2 (M18) as a substratewith human recombinant CYP1A2, was elucidated by NMR (Huskeyet al., 2016). Subsequently, M23 was chemically synthesized and usedas the reference standard with the internal standard ([13C6]KAE609) forthe quantification of M23 by LC-MS/MS analysis. Thus, the concen-trations of M23 were determined by both radiometric and LC-MS/MSassays in this study.PK parameters of M23 in each subject are summarized in Table 3.

Concentrations of M23 were quantifiable as early as 1 hour postdose inall subjects and reached the maximal mean concentration of 265 ng/mlat a median tmax approximately 24 hours postdose. The t1/2 of M23 wasapproximately 35.3 hours. Exposure of M23 was similar amongsubjects 5101, 5103, 5104, and 5106; however, much lower exposureof M23 (approximately one-third of the other subjects) was observed insubjects 5102 and 5105.Exposure of KAE609 and M23 was compared using radiometric or

LC-MS/MS methods. The mean exposure of KAE609 (AUC0–120 h

approximately 75 mgEq×h/ml) by the radiometric method is com-parable to that obtained using LC-MS/MS quantification (AUCinf

approximately 67.0 mg×h/ml). Similarly, the mean exposure of M23(AUC0–120 h approximately 11.7 mgEq×h/ml) by the radiometricmethod was comparable to that by LC-MS/MS quantification(AUC0–120 h approximately 16.1 mg×h/ml). The mean exposure ratio ofM23 (AUCinf 18.3 mg×h/ml) to KAE609 (AUCinf 67.0 mg×h/ml) wasestimated to be approximately 29%.

Metabolite Profiling of [14C]KAE609 in Urine and Feces

Recovery of radioactivity from urine was less than 5% of the dose;thus, metabolic profiling was conducted from only two subjects (5103and 5104). Urine and fecal pools were prepared and quantitativeinformation on the metabolites from each subject appears in Table 4.Representative metabolic profiles in urine and feces from subject 5103are depicted in Fig. 3, B and C, respectively. A detailed summary ofKAE609 and metabolites in excreta from each subject can be found inSupplemental Tables 2 and 3.[14C]KAE609 was not detected in the urine. The prominent radio-

labeled components were M23 and M14, with an average value of

Fig. 4. Overall metabolism of KAE609. The exposureof KAE609 and metabolites in plasma is presented asthe percentage of AUC. The recovery of KAE609 andmetabolites in excreta (urine, feces, and semen) ispresented as the percentage of the dose.

678 Huskey et al.

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

approximately 1.4% of dose. All other metabolites were present at lowlevels, contributing, 0.5% of dose. M35.8 was a glucuronide based onLC-MS/MS analysis, presumably derived from M48.KAE609 and M23 were two major components in human feces,

accounting for a mean of approximately 32% and 36% of the dose,respectively. The amount of KAE609 and M23 excreted in feces variedto some extent among six subjects. Several minor metabolites were alsoidentified, each accounting for , 2%–3% of the dose. Three minormetabolites (M16, M23.5A, and M23.5B) were not observed fromprevious nonclinical ADME studies. The proposed structures of metab-olites are presented in Fig. 5.

Metabolite Profiling of [14C]KAE609 in Semen

Semen samples were collected at 6, 12, and 24 hours postdose (n = 2per time point and only one collection per subject). A representativemetabolic profile of a semen sample from subject 5103 is depicted inFig. 3D. Quantitative information on the metabolites from each subjectappears in Table 5. The amount of total radioactivity excreted in semenfrom each subject was minimal, ranging from 832 to 15,177 ngEq(,0.005% of the radioactivity dose). The most abundant radiolabeledcomponent in semen was M23, accounting for approximately 56% oftotal radioactivity in semen. Two prominent radiolabeled componentswere KAE609 andM16, accounting for approximately 12% and 10% oftotal radioactivity in semen. A few minor metabolites were alsoidentified, each accounting for approximately 3%–6% of total radio-activity in semen.

Overall Metabolism of [14C]KAE609 in Plasma and Excreta

Figure 4 presents the mean exposure of KAE609 and metabolites inplasma and mean recovery of KAE609 and metabolites in feces, urine,and semen. KAE609 was the major circulating component, accountingfor approximately 76% of the total AUC. M23 and M18 were twoprominent metabolites, accounting for approximately 12% and 8% ofthe total AUC. However, both KAE609 (approximately 32% of dose)and M23 (approximately 36% of the dose) were major components inhuman feces, the primary route of elimination. KAE609 was notdetected in urine and M14 and M23 were two minor metabolitesexcreted in urine. Only a trace amount of radioactivity was excreted insemen (,0.005%of the dose).M23was themajor component, followedby KAE609 and M16 in semen.

Structural Characterization of [14C]KAE609 and Metabolites byLC-MS/MS

Structural characterization of metabolites was based on theirelemental composition derived from accurate mass measurements andfragment ions in their data-dependent MS2 and MS3 mass spectra.Comparison of metabolite fragment ions with those of KAE609allowed the assignment of regions of biotransformation. The accuratemass, elemental formula, and diagnostic fragment ions of each metab-olite are summarized in Table 6 and proposed structures of themetabolites are presented in Fig. 5. Structural elucidation of M16,M18, M23, and M37 was based on LC-MS/MS and NMR analysis andwas further confirmed by chemical synthesis (Huskey et al., 2016).Detailed description of structural elucidation of compound 4 by

NMR can be found in the Supplemental Results and SupplementalTable 1. Compound 4 shared the identical molecular ion and theidentical product ion spectra with M16 from human feces. In addition,M16 and compound 4 shared identical elution times in three HPLC/UPLC systems tested. A detailed description of HPLC coelution can befound in the Supplemental Results. Therefore, we concluded thatcompound 4 is M16.

Metabolite Pathways of [14C]KAE609 in Humans

An overall metabolic pathway for KAE609 in humans is summarizedin Fig. 5. KAE609 was well absorbed and extensively metabolized inhumans, such that unchanged KAE609 in feces represented approxi-mately 32% of the dose. KAE609 underwent C-C bond cleavage and a1,2-acyl shift to form ring expansion metabolite M37. This novelbiotransformation was catalyzed by CYP3A4. M37 was oxidized toM18 by CYP3A4 and was further hydroxylated to M23 by CYP1A2(Huskey et al., 2016). M23 was further oxidized to M16 by all 19commercially available cytochrome P450 (P450) enzymes, albeit at lowlevels (data not shown). KAE609 also underwent oxidation to M40with 2 mass units lower than that of KAE609 and one metabolite (M48)with 4 mass units lower than that of KAE609. KAE609 and theseoxidative metabolites were further hydroxylated and/or oxygenated tonumerous minor metabolites (M29.8, M34, M14, M21.8, M23.5A,M24, M32, and M35). M48 underwent glucuronidation to M35.8 andM23.5B presumably was a carbamate metabolite of KAE609.

Discussion

KAE609 is an antimalarial agent in development (Rottmann et al.,2010; Meister et al., 2011; Spillman et al., 2013). An oxidativemetabolite, M23, was identified in human plasma; however, M23was not detected in rat and dog plasma (Huskey et al., 2016). M23 maypose concerns for metabolite exposure coverage (see U.S. Food andDrug Administration 2008, 2010, and 2012 guidance, available athttp://www.fda.gov/cder/guidance/index.htm and http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/de-fault.htm). Our strategy was to conduct an investigation of the humanADME early in development to define the metabolite pathways ofKAE609, to elucidate potential preclinical safety exposure concerns,and to determine the potential effect of intrinsic and extrinsic factors onKAE609.A single oral dose of 300 mg [14C]KAE609 was well tolerated by all

six subjects and no significant adverse events were reported. KAE609

TABLE 6

Fragmentation patterns of KAE609 and metabolites

Structural characterization of metabolites was carried out by LC-MS/MS analysis. Theproposed structures of the metabolites were based on their elemental composition derived fromaccurate mass measurements and fragment ions in their data-dependent MS2 and MS3 massspectra. Comparison of metabolite fragment ions with those of KAE609 allowed the assignmentof regions of biotransformation, as presented in Figure 5.

Compound [MH+] m/z Elemental Formula Diagnostic Fragment Ions

KAE609 390.0573 C19H15Cl2FN3O 390, 347,b 319,b 312, 284, 276M14 404.0362 C19H13Cl2FN3O2 404, 386,b 360, 351, 325, 278M16 420.0322 C19H13Cl2FN3O3 420, 402,b 376, 367, 341M18 388.0416 C19H13Cl2FN3O 388, 359,b 324M20.8 410.046 C18H15Cl2FN3O3 410, 392, 349, 226,b 184M21.8 420.0302 C19H13Cl2FN3O3 420, 402, 379,b 362,b 337, 252M23 404.0363 C19H13Cl2FN3O2 404, 375,b 340, 235M23.5A 420.0324 C19H13Cl2FN3O3 420, 391,b 356, 328M23.5B 434.048 C20H15Cl2FN3O3 434, 419,b 405M24 404.0362 C19H13Cl2FN3O2 404, 375,b 340, 235M29.8 406.0524 C19H15Cl2FN3O2 406, 210b

M32 404.0375 C19H13Cl2FN3O2 404,b 386M34 406.0528 C19H15Cl2FN3O2 406, 391, 210,b 182, 175M35 402.0216 C19H11Cl2FN3O2 402, 374, 367,b 346, 338, 323b

M35.8 562.0595 C25H19Cl2FN3O7 562, 386b

M37a 392.0574 14CC18H15Cl2FN3O 392, 327, 279, 210,b 175b

M40 388.0419 C19H13Cl2FN3O 388,b 359, 210M48 386.0256 C19H11Cl2FN3O 386,b 351

aThe molecular ion of M37 was previously detected in human hepatocyte after incubated with[14C]KAE609.

bIndicates the most abundant fragment ions.

ADME of KAE609, a Potent Antimalarial Agent, in Humans 679

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

was absorbed with a median tmax of approximately 3.5 hours and a Cmax

of approximately 1780 ng/ml. Mean clearance (CL/F) was approxi-mately 5 l/h, suggesting that KAE609 is a low clearance drug. Thepharmacokinetic parameters of KAE609 (Table 1) were in goodagreement with the data obtained from other clinical studies, with theexception that the t1/2 of KAE609 reported in this study was approx-imately 33 hours, which is longer than that estimated in other clinicalstudies (approximately 24 hours). This observation was attributed tothe sensitive quantitative assay and long collection time intervals (up to312 hours postdose) in our study. Similar to humans, the PK ofKAE609 also had a long terminal half-life (8.2–10.9 hours) and lowplasma clearance (0.201–0.252 l/h per kilogram) and good oral absorp-tion (72.5%–100%) in rats and dogs (Huskey et al., 2016).Radioactivity (KAE609 and metabolites) disappeared slowly from

the blood and plasma, with a mean t1/2 of approximately 41 hoursand 47 hours, respectively, indicating a slower elimination of totalradioactivity compared with the parent (approximately 33 hours) (datanot shown). The higher radioactivity in plasma relative to whole bloodsuggests that drug-related radioactivity has no special affinity anddistribution to red blood cells (data not shown). Plasma protein bindingof KAE609 is known from previous studies to be high ($99.7% inhumans).The most abundant radioactive component in plasma was KAE609,

accounting for approximately 76% of the total plasma radioactivity(pooled from 0 to 120 hours or from 0 to 144 hours). M23 was the majorcirculating metabolite, accounting for approximately 12% of the totalplasma radioactivity. Several minor oxidative metabolites (M14, M16,M18, and M23.5B) were also identified, each accounting for approx-imately 3%–8% of the total plasma radioactivity. M23 was further

quantified by LC-MS/MS using reference compound 3 and internalstandard ([13C6]KAE609). The t1/2 of M23 was comparable to that ofKAE609 (approximately 35 hours). Exposure (AUCinf) of M23 wasestimated to be approximately 29% of KAE609.Exposure of M23 was greater than 10% of total radioactivity in

plasma; thus, safety evaluation of M23 in nonclinical species isplanned according to guidance recommendations (see U.S. Food andDrug Administration 2008, 2010, and 2012 guidance, available athttp://www.fda.gov/cder/guidance/index.htm and http://www.fda.gov/Drugs/GuidanceComplianceRegulatoryInformation/Guidances/default.htm). M23 exposure was greater than 25% of KAE609exposure; thus, evaluation of in vitro drug–drug interaction potentialof M23 is in progress. Our investigation of the human ADME early indevelopment proved to have significant effect on subsequent pre-clinical safety evaluations for clinical development.After oral administration of [14C]KAE609, the recovery of total

radioactivity was greater than 85% of the dose in excreta of six subjectswithin 14 days postdose; these findings indicated that, in all likelihood,significant prolonged tissue retention of radioactivity was not occur-ring. Radioactivity was excreted slowly, consistent with the long half-life of total radioactivity in blood and plasma. Fecal excretion(approximately 85% of the dose) was the primary route of eliminationof KAE609 and its metabolites. KAE609 and M23 were two majorcomponents recovered in feces, accounting for approximately 32% and36% of the dose, respectively. Thus, KAE609 was well absorbed(.68%) and extensively metabolized in humans.Although it is not a common practice to collect semen in human

ADME studies, detection of drugs and their metabolites in semen hasbeen reported previously, particularly in patients with HIV (Ette et al.,

Fig. 5. Proposed metabolic pathways ofKAE609 in humans. M37 was not detectedin this study; however, it was detected inhepatocyte incubations across species and in ratplasma, bile, and feces. Structures of M37,M18, M23, and M16 were characterized byLC-MS/MS and NMR analyses and furtherconfirmed by chemical synthesis. The amountof each metabolite recovered in excreta, asexpressed as the percentage of the dose, ispresented; the trace amount represents , 3% ofthe oral dose.

680 Huskey et al.

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

1988; Taylor and Pereira, 2001; Avery et al., 2013). Semen wasevaluated since subjects in prior studies had reported semen discolor-ation after taking KAE609 at higher doses. Further investigationclarified the potential cause of the discoloration and distribution ofKAE609 and its metabolites. Semen samples were collected once fromeach subject at 6, 12, or 24 hours postdose (n = 2 per time point). Totalradioactivity excreted in semen was negligible, accounting for ,0.005% of the dose. M23 was the major component in semen,accounting for approximately 56% of the total radioactivity, followedby KAE609 and M16 at approximately 12% and 10% of the totalradioactivity.The source of the color of M18 and M23 was attributed to their

extended conjugated systems of double bonds in the structures (Huskeyet al., 2016). In this human ADME study, M23 andM18were identifiedto be the prominent radiolabeled components in semen, which explainedthe previous observation of colored semen reported in clinics. Furtherevaluation of semen samples did not indicate any abnormalities in spermcounts and sperm mobility.M16was detected in semen, accounting for 10% of total radioactivity

excreted; therefore, structural elucidation of M16 was further pursued.Based on product ion spectra, M16 was the oxidative product of M23.We incubated compound 3 (M23) with a panel of 19 human recom-binant P450 enzymes. All P450 enzymes were capable of generatingM16, albeit at very low levels. Because of the low yield by P450enzymes, it is not feasible to generate a sufficient quantity of M16 usingbioreactions for structural elucidation. Our attempt to isolate M16 fromhuman urine and feces also produced an insufficient quantity. Thus, analternative approach using electrochemical synthesis of M16 wasexplored.Electrochemistry has been proposed as a quick and simple tool to

prepare human metabolites (Faber et al., 2014). It is mostly used toprepare hydroxylated metabolites in the benzylic position or in aromaticrings, provided that it is activated by electron donating groups, such asphenols, ethers, or anilines (Jurva et al., 2003). Despite the high numberof investigated compounds, very few studies have prepared sufficientamounts for NMR analysis (Stalder and Roth, 2013; Fredenhagen et al.,2014). In our hands, electrochemistry is especially useful for com-pounds that are sparingly water soluble and that therefore are notreactive in aqueous enzyme reactions (Schroer et al., 2010). Operationunder constant voltage, rather than constant current, frequently giveshigher yields because overoxidation is reduced, but it leads to longerreaction times. Because the proposed structure of M16 suggestedbenzylic oxidation, electrochemical production of reference materialwas undertaken. A boron doped diamond electrode under acidicconditions gave the desired product in 10% yield. Surprisingly, theoxidation proceeded stereoselectively, because the hydroxyl radicalattacked from the opposite site of the neighboring methyl group.Stereoselective electrochemical allylic oxidation of cholesteryl acetatehas been described (Okamoto et al., 2004).The metabolic pathway of KAE609 was proposed (Fig. 5). KAE609,

a spiroindolone derivative, undergoes an unusual C-C bond cleavage,followed by an acyl 1,2 shift to form a ring expansion structure M37.This novel biotransformation reaction was catalyzed by CYP3A4(Huskey et al., 2016). M37 was subsequently oxidized to M18 byCYP3A, hydroxylated toM23 by CYP1A2, and further hydroxylated toM16 by all 19 P450 enzymes tested. KAE609 also underwent oxidationto M40 with 2 mass units lower than that of KAE609 and onemetabolite (M48) with 4 mass units lower than that of KAE609.KAE609 was directly hydroxylated to form two metabolites (M29.8and M34). Two metabolites (M40 and M48) were further hydroxylatedto several metabolites (M14, M21.8, M23.5A, M24, M32, and M35).One glucuronide (M35.8) was also identified in urine. Of note, M37

was the primary metabolite of KAE609 in humans even though M37was not detected in any matrices in this study.Metabolism of KAE609 was qualitatively similar but quantitatively

different across species. The major metabolite via CYP3A was M37 inrats and M23 in dogs and humans, which is formed from M18 byCYP1A2 (M18 was formed from M37 via CYP3A). M23 was detectedin rat and dog feces but not in plasma. We hypothesize that thequantitative differences across species are probably secondary to theknown differences in relative amounts of the different P450 enzymes ineach species and potentially to a difference in transporters.Two observations support a transporter difference hypothesis. In

dogs, a continuous intravenous infusion of M23 to achieve a plasmaM23 exposure similar to that found in humans after oral dosing ofKAE609 still yielded a dog semen concentration that was approxi-mately 300-fold below that found in humans. M23 was the major fecalmetabolite in both humans and dogs; therefore, M23 formed fromhepatic metabolism was distributed to blood and bile in humans, butonly to bile in dogs.Overall, a single 300-mg dose of KAE609 was generally safe and

well tolerated in these healthy male subjects. There were no serious orsignificant adverse events, and no clinically significant laboratoryabnormalities were reported in this study. KAE609 is mainly excretedinto feces with no urinary excretion. In contrast with dog and rat ADMEstudies, M23 is the major metabolite found in plasma, feces, and semenin humans.

Acknowledgments

The authors thank Dr. Francis Tse for constructive discussion and continuedsupport; Dr. Alana Upthagrove for insightful suggestions; and Drs. AlbrechtGlaenzel, Vishal Koradia, Stefanie Mayer, Nhat Quang, and Nguyen Trung forcontributions in the preparation of the final dosage. The authors also thank Dr.Amy Wu, Lawrence Jones, Ann Draghi, Cindy Chen, Dr. Zhenzhong Su, andRana Vani for analytical and operational support. Finally, the authors thankAlexander Marziale for SFC purification and Dr. Pascal Van Tilburg (studydirector from PRA) for conducting the clinical study and measurements ofradioactivity in plasma and excreta of KAE609.

Authorship ContributionsParticipated in study design: Huskey, F. Yang, Jain, Stein, Sunkara.Conducted experiments: Zhu, Kühnöl, Jian, Miao.Performed data analysis: Huskey, Zhu, Luneau, Z. Yang, Stein.Wrote or contributed to the writing of manuscript: Huskey, Fredenhagen,

Mangold, Stein.

References

Avery LB, VanAusdall JL, Hendrix CW, and Bumpus NN (2013) Compartmentalization andantiviral effect of efavirenz metabolites in blood plasma, seminal plasma, and cerebrospinalfluid. Drug Metab Dispos 41:422–429.

Carter TE, Boulter A, Existe A, Romain JR, St Victor JY, Mulligan CJ, and Okech BA (2015)Artemisinin resistance-associated polymorphisms at the K13-propeller locus are absent inPlasmodium falciparum isolates from Haiti. Am J Trop Med Hyg 92:552–554.

Ette EI, Ogonor JI, and Essien EE (1988) Passage of chloroquine into semen. Br J ClinPharmacol 26:179–182.

Faber H, Vogel M, and Karst U (2014) Electrochemistry/mass spectrometry as a tool inmetabolism studies-a review. Anal Chim Acta 834:9–21.

Fredenhagen A, Kittelmann M, Peukert S, Eggimann FK, Kuehnoel J (2014) inventors, NovartisAG, assignee. Preparation of sonidegib metabolites. World Intellectual Property Organizationpatent WO 2015092720 A1. 2015 Jun 25.

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.

Hott A, Casandra D, Sparks KN, Morton LC, Castanares G-G, Rutter A, and Kyle DE (2015)Artemisinin-resistant Plasmodium falciparum parasites exhibit altered patterns of developmentin infected erythrocytes. Antimicrob Agents Chemother 59:3156–3167.

Huskey SEW, Zhu CQ, Lin MM, Forseth RR, Gu H, Simon O, Eggimann FK, Kittelmann M,Luneau A, and Vargas A, et al. (2016) Identification of three novel ring expansion metabolitesof KAE609, a new spiroindolone agent for the treatment of malaria, in rats, dogs, and humans.Drug Metab Dispos, in press.

Jurva U, Wikström HV, Weidolf L, and Bruins AP (2003) Comparison between electrochemistry/mass spectrometry and cytochrome P450 catalyzed oxidation reactions. Rapid Commun MassSpectrom 17:800–810.

ADME of KAE609, a Potent Antimalarial Agent, in Humans 681

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from

Leong FJ, Li R, Jain JP, Lefèvre G, Magnusson B, Diagana TT, and Pertel P (2014) A first-in-human randomized, double-blind, placebo-controlled, single- and multiple-ascending oral dosestudy of novel antimalarial Spiroindolone KAE609 (Cipargamin) to assess its safety, tolerability,and pharmacokinetics in healthy adult volunteers. Antimicrob Agents Chemother 58:6209–6214.

Meister S, Plouffe DM, Kuhen KL, Bonamy GM, Wu T, Barnes SW, Bopp SE, Borboa R, BrightAT, and Che J, et al. (2011) Imaging of Plasmodium liver stages to drive next-generationantimalarial drug discovery. Science 334:1372–1377.

Okamoto I, Funaki W, Nakaya K, Kotani E, and Takeya T (2004) A new electrochemical systemfor stereoselective allylic hydroxylation of cholesteryl acetate with dioxygen induced by ironpicolinate complexes. Chem Pharm Bull (Tokyo) 52:756–759.

Rottmann M, McNamara C, Yeung BK, Lee MCS, Zou B, Russell B, Seitz P, Plouffe DM,Dharia NV, and Tan J, et al. (2010) Spiroindolones, a potent compound class for the treatmentof malaria. Science 329:1175–1180.

Schroer K, Kittelmann M, and Lütz S (2010) Recombinant human cytochrome P450 mono-oxygenases for drug metabolite synthesis. Biotechnol Bioeng 106:699–706.

Spillman NJ, Allen RJW, McNamara CW, Yeung BKS, Winzeler EA, Diagana TT, and Kirk K(2013) Na(+) regulation in the malaria parasite Plasmodium falciparum involves the cationATPase PfATP4 and is a target of the spiroindolone antimalarials. Cell Host Microbe 13:227–237.

Stalder R and Roth GP (2013) Preparative microfluidic electrosynthesis of drug metabolites. ACSMed Chem Lett 4:1119–1123.

Taylor S and Pereira AS (2001) Antiretroviral drug concentrations in semen of HIV-1 infectedmen. Sex Transm Infect 77:4–11.

White NJ, Pukrittayakamee S, Phyo AP, Rueangweerayut R, Nosten F, Jittamala P, Jeeyapant A,Jain JP, Lefèvre G, and Li R, et al. (2014) Spiroindolone KAE609 for falciparum and vivaxmalaria. N Engl J Med 371:403–410.

Yeung BKS, Zou B, Rottmann M, Lakshminarayana SB, Ang SH, Leong SY, Tan J, Wong J,Keller-Maerki S, and Fischli C, et al. (2010) Spirotetrahydro b-carbolines (spiroindolones): anew class of potent and orally efficacious compounds for the treatment of malaria. J Med Chem53:5155–5164.

Address correspondence to: Su-Er W. Huskey, Drug Metabolism andPharmacokinetics, Novartis Institutes for BioMedical Research, One Health Plaza,East Hanover, NJ 07936-1080. E-mail: su.huskey@novartis.com

682 Huskey et al.

at ASPE

T Journals on February 6, 2020

dmd.aspetjournals.org

Dow

nloaded from