Identification of three novel ring expansion metabolites of KAE609, a new spiroindolone agent for the treatment of malaria, in rats, dogs, and humans

Article abstract of DOI:10.1124/dmd.115.069112

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] is a potent, fast-acting, schizonticidal agent being developed for the treatment of malaria. After oral dosing of KAE609 to rats and

Full text of DOI:10.1124/dmd.115.069112

Supplemental material to this article can be found at:
http://dmd.aspetjournals.org/content/suppl/2016/02/26/dmd.115.069112.DC1.html
1521-009X/44/5/653–664$25.00
DRUG METABOLISM AND DISPOSITION
http://dx.doi.org/10.1124/dmd.115.069112
Drug Metab Dispos 44:653–664, May 2016
Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics
Identification of Three Novel Ring Expansion Metabolites of KAE609,
a New Spiroindolone Agent for the Treatment of Malaria, in Rats,
Dogs, and Humans s
Su-Er W. Huskey, Chun-qi Zhu, Melissa M. Lin,1 Ry R. Forseth, Helen Gu, Oliver Simon,
Fabian K. Eggimann, Matthias Kittelmann, Alexandre Luneau, Alexandra Vargas, Hongmei Li,
Lai Wang, Heidi J. Einolf, Jin Zhang, Sarah Favara, Handan He, and James B. Mangold
Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, East Hanover, New Jersey (S.-E.W.H., C.Z.,
R.R.F., H.G., H.L., L.W., H.J.E., J.Z., S.F., H.H., J.B.M.); Technical Research and Development, Novartis Pharma, East Hanover, NJ
(M.M.L.); Novartis Institute for Tropical Diseases, Singapore (O.S.); and Global Discovery Chemistry, Bioreactions Group, Novartis
Institutes for BioMedical Research, Basel, Switzerland (F.K.E., M.K., A.L., A.V.)
Received December 21, 2015; accepted February 19, 2016
ABSTRACT
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] is a po-
tent, fast-acting, schizonticidal agent being developed for the
or rearrangement during biotransformation. Subsequent structural
elucidation of M37 revealed that KAE609, a spiroindolone, undergoes
an unusual C-C bond cleavage, followed by a 1,2-acyl shift to form a
treatment of malaria. After oral dosing of KAE609 to rats and ring expansion metabolite M37. The in vitro metabolism of KAE609 in
dogs, the major radioactive component in plasma was KAE609. hepatocytes was investigated to understand this novel biotransfor-
An oxidative metabolite, M18, was the prominent metabolite in rat mation. The metabolism of KAE609 was qualitatively similar across
and dog plasma. KAE609 was well absorbed and extensively the species studied; thus, further investigation was conducted using
metabolized such that low levels of parent compound (£11% of human recombinant cytochrome P450 enzymes. The ring expansion
the dose) were detected in feces. The elimination of KAE609 and reaction was found to be primarily catalyzed by cytochrome P450
metabolites was primarily mediated via biliary pathways (‡93% of (CYP) 3A4 yielding M37. M37 was subsequently oxidized to M18 by
the dose) in the feces of rats and dogs. M37 and M23 were the CYP3A4 and hydroxylated to M23 primarily by CYP1A2. Interestingly,
major metabolites in rat and dog feces, respectively. Among the M37 was colorless, whereas M18 and M23 showed orange yellow
prominent metabolites of KAE609, the isobaric chemical species, color. The source of the color of M18 and M23 was attributed to their
M37, was observed, suggesting the involvement of an isomerization extended conjugated system of double bonds in the structures.
Introduction
Karbwang, 2013; Witkowski et al., 2013). Growing global efforts are
focusing on finding new cures to combat malaria drug resistance.
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; Yeung
et al., 2010), a spiroindolone, represents a new class of potent, fast-acting,
schizonticidal agents for the treatment of malaria (Rottmann et al., 2010;
Meister et al., 2011; Spillman et al., 2013). More recently, KAE609 was
shown to be safe and well tolerated in healthy subjects up to a single dose of
300 mg and multiple doses of 150 mg daily for 3 days (Leong et al., 2014).
Evaluation in 21 adult patients with uncomplicated malaria due to either
Plasmodium vivax (n = 10) or Plasmodium falciparum (n = 11) showed
that once-daily dosing of KAE609 at 30 mg for 3 days resulted in a
rapid median parasite clearance time of approximately 12 hours for both
P. vivax and P. falciparum (White et al., 2014).
Malaria is one of the leading causes of death and disease worldwide,
especially among young children in developing countries. Drug
resistance to widely used artemisinin-based therapies (Carter et al.,
2015; Hott et al., 2015) has become an imminent issue since drug-
resistance strains were identified in Bangladesh, Cambodia, Thailand,
Vietnam, Myanmar and Laos (Haque et al., 2013; Na-Bangchang and
All authors are employees of Novartis. The authors have no other relevant
affiliations or financial involvement with any other organization or entity with a
financial interest in or financial conflict with the subject matter or materials
discussed in the manuscript apart from those disclosed. No writing assistance
was utilized in the production of this manuscript.
1Current affiliation: Merck Research Laboratories, Rahway, NJ
dx.doi.org/10.1124/dmd.115.069112.
In preparation for the clinical investigations described above, efforts
were undertaken to understand in vitro metabolism of KAE609 across
s This article has supplemental material available at dmd.aspetjournals.org.
ABBREVIATIONS: ADME, absorption, distribution, metabolism, and excretion; AUC, area under concentration-time curve; b-RAM, online radioactivity
monitor; COSY, correlation spectroscopy; CYP, cytochrome P450; DPM, disintegration per minute; DQF-COSY, double quantum filter correlation
spectroscopy; H/D, hydrogen/deuterium; HMBC, heteronuclear multiple bond correlation; HPLC, high-performance liquid chromatography; KAE579, (19
R,39S)-5-chloro-69-fluoro-39-methyl-29,39,49,99-tetrahydrospiro[indoline-3,19-pyridol[3,4-b]indol]-2-one; 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; MS, mass
spectrometry; MSE, MS technique involving dynamic collision between high and low energy switching; m/z, mass-to-charge ratio; NMR, nuclear
magnetic resonance; %PRA, percentage of radioactivity; PK, pharmacokinetics; QC, quality control; t1/2, terminal elimination half-life.
653

654
Huskey et al.
(Downers Grove, IL). Control blank plasma samples from rats and dogs were
purchased from Bioreclamation (Hicksville, NY).
species as well as absorption, distribution metabolism, and excretion
(ADME) of KAE609 in preclinical species. Following oral dosing,
KAE609 was the major component in plasma. An oxidative metabolite,
M18, was the major metabolite in rat and dog plasma, while M37 and
M23 were the major metabolites in rat and dog feces, respectively. To
note, nomenclature of metabolites was based on their high-performance
liquid chromatography (HPLC) elution time in the study in which they
were first identified. Based on liquid chromatography–tandem mass
spectrometry (LC-MS/MS) analysis, M37 shared the same molecular
weight with KAE609, suggesting that M37 was a product of isomeriza-
tion or rearrangement of KAE609 during biotransformation. Therefore,
the structural elucidation of M37 was undertaken to facilitate the under-
standing of this unusual biotransformation. During purification of M37
from rat feces, two brightly colored yellow fractions were identified
corresponding to M18 and M23. The strategies and approaches we used
in the elucidation of structures of M18 and M23 and the metabolic
pathways of KAE609 are described herein.
In parallel to the in vivo studies above, in vitro across-species
metabolism of KAE609 revealed that M23 was a prominent metab-
olite in human hepatocytes. Based on the preclinical ADME studies,
M23 was not detected in rat or dog plasma, although M23 was the
major metabolite in dog excreta. In view of Metabolites in Safety
Testing and International Conference on Harmonisation guidance
(Baillie et al., 2002; Baillie, 2008; Smith and Obach, 2009; U.S. Food
and Drug Administration 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), we describe herein the strategy we used regarding the
potential exposure coverage of M23 in humans.
In this article, we describe the strategies and approaches used in
the structural elucidation of three key metabolites of KAE609,
investigation of the cytochrome P450 (CYP) enzymes involved in a
novel ring expansion reaction of KAE609 to generate metabolite
M37 and subsequent formation of oxidative highly colored metab-
olites from M37. To the best of our knowledge, this is the first report
that revealed the ring expansion of a spiroindolone and generation of
two colored metabolites (M18 and M23) from a colorless precursor
metabolite (M37) by a single CYP-catalyzed oxidation reaction. All
these three metabolites (M37, M18, and M23) showed no pharma-
cological activity after the spiroindolone core of KAE609 was lost
by biotransformation.
Chemical Synthesis of Metabolites
Chemical synthesis of M37 (compound 1), M18 (compound 2), and M23
(compound 3) was carried out for structural confirmation and additional biologic
profiling. A detailed description of the preparation and characterization of the
metabolites can be found in the Supplemental Materials and Methods.
Biocatalytic Synthesis of Compound 3 (M23)
Preparation of Biocatalyst. The biocatalyst, Escherichia coli expressing
recombinant human CYP1A2, was prepared as described in Kittelmann et al.
(2012).
Biotransformation, Preparative Conditions (Compound 3, M23). In a
400-ml polyethylene flask, 40 ml cell suspension (optical density at 600 nm =
100) in PSE buffer (potassium phosphate 50 mM, sucrose 250 mM, and
EDTA 0.25 mM, pH 7.5) was mixed with 10 ml 0.9% NaCl solution and 5 ml
nutrient solution (40 g/l glucose, 40 g/l lactose, and 60 g/l aqueous sodium
citrate). The biotransformation was initiated by adding the substrate solution,
containing 2 mg compound 2 (M18) dissolved in 2.5 ml acetonitrile. The
reaction was incubated with open cap at 30ꢀC and 250 rpm for 16 hours in an
orbital shaker.
Purification of Biotransformation Product (Compound 3). After incubation,
the reaction was mixed with 5 g sodium chloride and extracted three times
with 100 ml ethyl acetate. The combined organic layers were mixed and
evaporated to dryness under reduced pressure. The residual raw product was
dissolved in 3 ml acetonitrile and injected into a 250 ꢀ 10 mm Nucloeodur
100-10 C18 ec column (Macherey-Nagel, Düren, Germany). The conditions
for preparative HPLC were as follows: solvent A, 0.05% trifluoroacetic acid
in water; solvent B, acetonitrile; gradient, 0–5 minutes 10% B, 5–48 minutes
10%–95% B, 48–53 minutes 100% B; flow rate, 4.5 ml/min; room temperature;
UV detection, 320 nm; and fraction size, 2 ml. The product eluted between 40%
and 50% B. The product containing fractions were combined and dried by
Speedvac lyophilization overnight. The product (1 mg) was obtained with
. 95% purity (HPLC/full diode array detector) and analyzed by nuclear
magnetic resonance (NMR) spectroscopy. The isolated yield was 46%.
In Vitro Incubations
To identify the CYP enzyme(s) involved in the metabolism of KAE609 or
M18 (compound 2) in humans, test compound (44 mM[14C]KAE609 or 50 mM
compound 2) was incubated with 19 commercially available recombinant human
CYP enzymes (listed in the chemicals section), or control microsomes in
100 mM potassium phosphate buffer (pH 7.4) containing 5 mM MgCl2. The
reactions were thermoequilibrated for 3 minutes at 37ꢀC and initiated by the
addition of 1 mM b-NADPH. The reactions were incubated for 30 minutes at
37ꢀC and were quenched by the addition of an equal volume of cold
acetonitrile. After vortex mixing and centrifugation, an aliquot of each sample
was analyzed by LC-MS/MS.
Materials and Methods
Chemicals
KAE609, KAE579 [(19 R,39S)-5-chloro-69-fluoro-39-methyl-29,39,49,99-
tetrahydrospiro[indoline-3,19-pyridol[3,4-b]indol]-2-one], compound 1 (M37),
compound 2 (M18), and compound 3 (M23) were synthesized by Novartis
Institute for Tropical Diseases and compound 3 (M23) was also prepared by the
Novartis Global Discovery Chemistry Bioreactions Group. [14C]KAE609 was
Animals
Male Wistar Hannover rats (approximately 227–310 g, n = 12) were
prepared by Novartis Isotope laboratories and radiochemical purity was . 99%. purchased from Harlan Laboratories (South Easton, MA). Catheters were
Microsomal preparations from baculovirus-infected insect cells expressing surgically implanted into the carotid artery and/or jugular vein of rats by the
recombinant human CYP enzymes coexpressed with CYP oxidoreductase
vendor (only one catheter was implanted into carotid artery for blood
[CYP1A1, CYP1A2, CYP1B1, CYP2C18, CYP2D6, CYP3A5, or CYP4A11], collection from rats receiving the oral dose). All rats were housed individ-
recombinant human CYP enzymes coexpressed with CYP oxidoreductase and
cytochrome b5 [CYP2A6, CYP2B6, CYP2C8, CYP2C9 (Arg144,Ile359),
CYP2C19, CYP2E1, CYP2J2, CYP3A4, CYP4F2, CYP4F3A, CYP4F3B or
ually in Culex metabolism cages (Culex Autosampler; BAS, Indianapolis,
IN) in a temperature- and humidity-controlled room (22 6 2ꢀC; 50% 6
15%) with free access to food and water (food was withheld until 4 hours
CYP4F12], as well as control microsomes were purchased from Corning postdose).
Discovery Labware (Corning, NY). Hepatocyte culture plates were obtained
from Corning Incorporated (Corning, NY). Krebs-Henseleit maintenance
Male beagles (approximately 10-14 kg, n = 5) were purchased from Marshall
farms (North Rose, NY) and kept in animal facility at Novartis for pharmaco-
medium, b-NADPH, dimethylsulfoxide, potassium phosphate (mono- and kinetics (PK) and ADME studies. The dogs were housed individually in a
dibasic), MgCl2, sodium bicarbonate, fructose, glycine, acetonitrile, metha-
temperature- and humidity-controlled room. Dogs were acclimated to the
nol, ammonium formate, and formic acid were purchased from Sigma-Aldrich metabolism cages 1 night prior to dosing. Dogs had free access to water and
(St. Louis, MO). OPTI-FLUOR liquid scintillant was purchased from Packard were given food once daily (food was withheld until 2 hours postdose).

Ring Expansion Metabolites of KAE609 (a Spiroindolone)
655
counted directly for radioactivity. The remaining plasma samples were stored
at 220ꢀC until analysis.
Plasma for the Quantification of KAE609
Blank plasma and study samples were thawed at room temperature.
Standards and quality control (QC) samples were prepared on the day of
analysis by adding appropriate standard or QC (25 ml) spiking solution to
475 ml blank plasma. Samples were loaded using a protein precipitation plate,
which was attached above a 1-ml 96-well assay plate. An aliquot (25 ml) of
study samples, blanks, standards, or QC samples was added to the appropriate
well followed by the addition of acetonitrile (100 ml) and the internal standard
(KAE579; 25 ml, 500 ng/ml) to each well. Samples were mixed for 5 minutes
and centrifuged at 2500 rpm for 15 minutes at 25ꢀC. The filtrate was
evaporated to dryness at approximately 45ꢀC under a stream of nitrogen
(TurboVap LV; Zymark Corp., Taunton, MA). The residues were reconstituted
with 300 ml acetonitrile/water/formic acid [10:90:0.1 (v/v/v)]. An aliquot (10 ml)
was analyzed for KAE609 and KAE579 by LC-MS/MS.
Plasma and Fecal Homogenates for Metabolic Profiling of KAE609
All plasma samples were thawed at room temperature and a 150-ml aliquot
was pooled from each animal at each time point. Each plasma pool was diluted
with 250 ml water and extracted with 2 ml acetonitrile/methanol/acetic acid [50:
50:0.1 (v/v/v)]. Fecal samples (5% and 1% by weight from rat and dog feces,
respectively) were pooled from fecal homogenate prepared from rats during
0–96 hours postdose and from dogs 0–72 hours postdose. The resulting pooled
fecal samples were extracted three times with 3 volumes of acetonitrile/methanol/
acetic acid [50:50:0.1 (v/v/v)].
After centrifugation at 3500 ꢀ rpm for 10 minutes, the supernatants were
combined and concentrated under a stream of nitrogen. Residues were recon-
stituted with acetonitrile/deionized water [50/50 (v/v)]. The recoveries of radio-
activity were 63.4% (i.v.) and 64.1% (oral) from rat feces and 60.7% (i.v.) and
66.7% (oral) from dog feces, respectively.
Fig. 1. Structure of KAE609 and KAE579. KAE609 is uniformly labeled in phenyl
ring (indicated by asterisk). KAE579, a derivative of KAE609, was used as an
internal standard for the quantification of KAE609.
Dose Administration
Sample Analysis
All doses were based on the individual body weights of animals on the day
of dosing. [14C]KAE609 (specific activity 40 mCi/mg at 5 mg/kg for rats; and
50 mCi/mg at 1 mg/kg for dogs) was dissolved in ethanol, polyethylene
glycol (PEG300) and pluronic F68 (5% in water) [10:20:70 (v/v/v)] for
intravenous dosing. [14C]KAE609 (specific activity 20 mCi/mg at 10 mg/kg)
was prepared in solution containing 1 N HCl (0.28%), solutol HS 15 (5%),
and 50 mM citrate buffer, pH 3 (94.7%) for oral doing to rats. [14C]KAE609
(specific activity 16.7 mCi/mg at 3 mg/kg) was prepared in suspension
containing 0.5% (w/v) methyl cellulose and 2% solutol HS 15 in water for
oral dosing to dogs.
Determination of Radioactivity. The radioactivity of all samples was
determined by liquid scintillation counting. For quench correction, an external
standard ratio method was used. Quench correction curves were established by
means of sealed standards. An aliquot of plasma and urine samples were counted
directly for radioactivity.
Solvable (500 ml) was added to each blood sample and incubated in a shaking
water bath at 50ꢀC for 2 hours. After incubation, 50 ml 100 mM EDTA as an
antifoaming agent and 200 ml 30% hydrogen peroxide were added to decolorize
the samples. The sample vials were loosely capped and returned to the water bath
for 3 hours. Thereafter, 10 ml Eq. 989 scintillation cocktail was added and the
samples were placed in the dark overnight to reduce chemiluminescence prior to
counting for radioactivity.
Each rat received an intravenous bolus injection via the jugular vein cannula.
Each dog received an intravenous slow bolus injection via the cephalic vein. The
oral dose was administered by gavage in both rats and dogs.
Fecal samples from animals were homogenized with 2 or 3 volumes of water.
Duplicate samples (approximately 100 mg) of the slurry were weighed into
scintillation vials and processed as described above for blood samples prior to
radioactivity determination. The remaining fecal homogenate was stored frozen
at 220ꢀC until LC-MS/MS analysis.
Blood Collection
Blood samples (200 ml) were collected from the carotid artery of rats at
selected time intervals in Culex tubes. Saline (200 ml) was automatically injected
after sample was collected to clear the cannula and replace the volume of blood
samples. Blood samples were collected via catheters placed in the cephalic vein
of each dog at selected time intervals. The total blood volume collected did not
exceed 1% of the body weights of rats and dogs.
Quantification of KAE609 by LC-MS/MS. Samples were analyzed on an
LC-MS/MS system consisting of a Shimadzu HPLC system (Shimadzu,
Kyoto, Japan) and Sciex API4000 mass spectrometer using Analyst software
version 1.4.2 (AB Sciex, Foster City, CA). The mass spectrometer was
operated in the positive ion mode, using turbospray ionization, with a source
temperature of 425ꢀC. Chromatographic separation was carried out on a Mac-
Mod Analytical ACE 5 C8 50 ꢀ 2.1 mm column. KAE609 and the internal
standard (KAE579, Fig. 1) were eluted using a gradient method with a mobile
phase consisting of the following: A, water containing 0.1% formic acid;
and B, acetonitrile containing 0.1% formic acid. The gradient was 35% B for
0.5 minutes, increased to 90% B from 0.5 to 2 minutes, followed by a 1-minute
hold at 90%B. The multiple reaction monitoring transitions for KAE609 and
KAE579 had a mass-to-charge ratio (m/z) of 390.2 to m/z 347.1 and m/z 356.2
Urine and Feces Collection
Urine and feces were collected daily from each of the animals for 7 days. Up to
3 days, the urine collection tubes were cooled with ice. After the final collection, each
cage was rinsed with water followed by 50% methanol. The cage wash was assayed
for radioactivity. Urine and feces samples were stored at 220ꢀC until analysis.
Sample Preparation
The plasma was obtained by centrifugation of blood samples at 4ꢀC (2000 ꢀ g) to m/z 313.1, respectively.
for 10 minutes. An aliquot of each blood sample was used for radioactivity
analysis (see the sample analysis section). An aliquot of plasma sample was
Calibration curves were generated by plotting the respective peak area ratios
(y) of KAE609 to the internal standard versus the concentrations (x) of the

656
Huskey et al.
calibration standards using weighted 1/x2 quadratic least-squares regression. The Data Processing
quantification was performed using Analyst software and Watson LIMS version
Quantification of KAE609 and Metabolites by Radiometry. KAE609 and
7.2.0.01 (Thermo Scientific, Philadelphia, PA). Concentrations in QC and study
samples were calculated from the resulting peak area ratios and interpolation
from the regression equations of the calibration curves. The lower limit of
quantification was 1.0 ng/ml.
metabolites were quantified in the extracts by radiochromatography. Peaks were
selected visually from the radiochromatogram, and their corresponding areas were
determined via peak integration (LAURA laboratory software from LabLogic).
The percentage of radioactivity (%PRA) in a particular peak, Z, was
calculated as follows:
Metabolic Profiling by LC-MS/MS. Metabolic profiling was performed
on a Waters Acquity UPLC System (Waters Corp., Milford, MA), equipped
with an autosampler and a quaternary pump, and an online radioactivity
monitor (b-RAM) or a fraction collector. Chromatographic separation was
carried out on a Mac-Mod Analytical Zorbax SB-C18 150 ꢀ 3.0 mm, 3.5-mm
column. KAE609 and the metabolites were eluted using a linear gradient
method with a mobile phase consisting of the following: A, 5 mM ammonium
acetate containing 0.1% formic acid; and B, acetonitrile/methanol [1:1 (v/v)]
containing 0.1% formic acid. The gradient was 5% B for 5 minutes, increased
to 50% B from 5 to 12 minutes, increased to 60% B from 12 to 30 minutes,
increased to 75% B from 30 to 40 minutes, increased to 85% B from 40 to
45 minutes, and increased to 98% B from 45 to 50 minutes, followed by a
3-minute hold at 98% B.
The HPLC column eluent was split 1:4 [mass spectrometry (MS)/fraction
collector or MS/b-RAM]. The eluent from the fraction collector was
mixed with methanol (0.5 ml/min) and collected directly into 96-Deepwell
LumaPlates coated with solid scintillant at 0.15 minutes per well using a
fraction collector. The plates were dried at 40ꢀC under a stream of nitrogen.
The plates were counted for 10–15 minutes per well in a TopCount Model
NXT radioactivity detector. Online monitoring with b-RAM employed a
250-ml liquid cell and eluent was mixed with 1 ml/min IN-Flow 2:1 (IN/US
Systems, Tampa, FL) liquid scintillant. The resulting data were processed
using the Laura Data System (version 4.0; LabLogic Systems Inc., Brandon,
FL). All quantification was based on the radioactivity associated with the
chromatographic peaks.
DPMꢀ inꢀ peakꢀ Z
%PRAꢀ inꢀ Z ¼
ꢀ 100
totalꢀ DPMꢀ inꢀ allꢀ integratedꢀ peaks
Where DPM is disintegration per minute. The concentration or amount of each
component was calculated as %PRA (as a fraction) multiplied by the total con-
centration (nanogram equivalent per milliliter) or percent of dose in the excreta.
Pharmacokinetic Parameters. The pharmacokinetic parameters were
calculated using actual recorded sampling times and noncompartmental
method(s) with Phoenix (WinNonlin version 6.2; Pharsight, Certara L.P.,
Princeton, NJ). Concentrations below the lower limit of quantification were
treated as zero for PK parameter calculations. The linear trapezoidal rule was
used for area under concentration-time curve (AUC) calculation. Regression
analysis of the terminal plasma elimination phase for the determination of
terminal elimination half-life (t1/2) included at least three data points after
Cmax. If the adjusted R2 value of the regression analysis of the terminal phase
was less than 0.75, no values were reported for t1/2, AUCinf, Vz/F, and
clearance/F.
Results
Mass Balance and PK of KAE609 in Rats and Dogs
After intravenous or oral dosing of [14C]KAE609, mass balance was
achieved in both rats and dogs, with 93%–100% of the administered
radioactivity dose being recovered in feces and , 2% of the dose
being recovered in urine. KAE609 had a long t1/2 (8.2 hours in rats
Online Hydrogen/Deuterium Exchange LC/MS-MS Method. Hydrogen/
deuterium (H/D) exchange LC/MS experiments were conducted. In these exper-
iments, water was substituted with deuterium oxide (D2O) in the HPLC mobile and 10.9 hours in dogs), low plasma clearance (0.252 l/h per
phase (A: 5 mM ammonium formate in D2O containing 0.1% formic acid). Samples
were analyzed under the identical LC-MS/MS method, as described below.
kilogram in rats and 0.201 l/h per kilogram in dogs), and moderate
volume of distribution at steady state (3.18 l/kg in rats and 2.17 l/kg
LC-MS Instrumentation and Operating Conditions. The structural in dogs). After oral doing, KAE609 was slowly absorbed in rats, with
characterization of metabolites was carried out using the above HPLC profiling
method coupled to a two-channel Z-spray (LockSpray) Waters Synapt G2S
quadrupole time-of-flight mass spectrometer. The Q-TOF Synapt was operated
in V-mode with a typical resolving power of at least 10,000. Qualitative analyses
were carried out using electrospray ionization in the positive ionization mode
using a lock spray source. Leucine enkephalin was used as the mass reference
standard for exact mass measurements and was delivered via the second spray
channel at a flow rate of 20 ml/min. Accurate mass LC/MS data were collected in
an alternating low-energy (MS) and elevated-energy (MSE) mode of acquisition.
In low-energy MS mode, data were collected at consistent collision energy of
a Cmax of 1210 ng/ml observed at 5.3 hours (tmax); however, it was
relatively rapidly absorbed in dogs, with a Cmax of 868 ng/ml observed
at 1.7 hours (tmax). The extent of oral absorption was complete in
rats and it was estimated to be 72.5% in dogs. The estimated oral
bioavailability was 84.6% in rats and 67.7% in dogs, indicating a
minimal first-pass effect in both rats and dogs. Detailed summary tables
of mass balance and PK parameters can be found in the Supplemental
Tables 1 and 2.
2 eV. In elevated MSE mode, collision energy was ramped from 15 to 30 eV Metabolite Profiling of [14C]KAE609 in Rat and Dog Plasma
during data collection cycle.
After intravenous or oral dosing, the major radiolabeled compo-
nent in rat and dog plasma was KAE609, accounting for approx-
imately 75%–84% of the total radioactivity AUC by either dosing
NMR Analysis of M37, M18, and M23. Detailed description of pro-
cedures for the isolation of metabolites from rat fecal extract can be found
in the Supplemental Materials and Methods. 1H spectra were acquired for
route. The prominent metabolite was M18, accounting for approx-
M37, M18, and M23 (dissolved in acetonitrile-d3) as well as reference
imately 10%–19% of total radioactivity AUC by either dosing
compounds 1, 2, and 3 (approximately 100 mg, dissolved in acetonitrile-d3).
route. Several minor metabolites were also detected, each account-
ing for # 3% of total radioactivity AUC by either dosing route.
Selective nuclear Overhauser effect spectroscopy data were acquired for M37
using a mixing time of 0.4 seconds, 32k data point collection over a spectral
Detailed graphic presentations can be found in the Supplemental
Figs. 1 and 2.
width of 12,019 Hz, and 2048 scans. A line broadening factor of 1 Hz was
applied to the free induction decay data before Fourier transformation. The
double quantum filter correlation spectroscopy (DQF-COSY) spectrum was
acquired for M23. NMR spectra were acquired on a Bruker DRX600 MHz
spectrometer equipped with a 3-mm dual CryoProbe at 300 K. For the DQF-COSY
spectrum, the 90ꢀ pulse (P1) was calibrated at a 1H transmitter power of 0.2 dB, and
the following acquisition parameters were used: acquisition time, 0.7 seconds;
complex increments, 1024; and number of scans, 64. Spectra were processed in
MestReNova. DQF-COSY spectra were processed using 8k (f2) by 2k (f1) zero
filling. Window functions were applied to the raw free induction decay (f2, sine
bell; f1, square sine function).
Metabolite Profiling of [14C]KAE609 in Rat and Dog Feces
KAE609 was well absorbed and extensively metabolized in rats and
dogs, such that unchanged KAE609 accounted for approximately 10%–
11% of the dose in rat feces and KAE609 was not detected in dog feces
by either dosing route (Fig. 2). M37 was the major component in rat
feces, accounting for approximately 23%–26% of the intravenous or
oral dose, whereas M23 was the major metabolite, accounting for

Ring Expansion Metabolites of KAE609 (a Spiroindolone)
657
Fig. 2. Representative metabolic profiles in
pooled rat feces (A) and pooled dog feces (B)
after oral dosing of [14C]KAE609. Feces samples
were collected daily from animals up to 7 days.
After homogenization, fecal homogenates were
extracted and analyzed by LC-MS/MS. The HPLC
separation of KAE609 and metabolites was
performed, as described in the Materials and
Methods. PO, oral administration.
approximately 25%–30% of either the intravenous or oral dose. In rat
Only 1%–2% of the intravenous and oral doses were recovered in
feces, several metabolites (M14, M18, M23, M24, M26, M36, M44, urine from either rats or dogs during the period of 0–168 hours;
M47, and M48) were identified, each accounting for approximately therefore, urine was not analyzed for metabolite profile.
2%–8% of the dose. Two metabolites (M20 and M20.1) were not
Structural Characterization of [14C]KAE609 and M37 by LC-MS/
MS and NMR
separated under the HPLC condition used in this study, together
accounting for approximately 12%–13% of the dose by either dosing
route. In dog feces, several prominent metabolites (M20, M21.2, M34,
The structures of the metabolites were proposed based on their
and M39), each accounting for approximately 10%–20% of the dose, elemental composition derived from accurate mass measurements
and several minor metabolites (M26, M37, M44, M46, and M46), each (,3–5 ppm), fragment ions in their data-dependent MS2 and MS3
accounting for approximately 4%–6% of the dose, were identified.
mass spectra, and exchangeable hydrogen atoms in the H/D exchange

658
Huskey et al.
metabolite spectrum. Acquisition of the one-dimensional nuclear
TABLE 1
Fragmentation patterns of KAE609 and metabolites
Overhauser effect spectroscopy spectrum demonstrated that the
proton at 5.84 ppm has a through-space correlation to the methyl
doublet at 1.28 ppm and the aromatic singlet at 7.50 ppm, indicating
that it is close to the chlorobenzene moiety. In addition, the aromatic
singlet at 7.50 ppm showed a nuclear Overhouser effect correlation to
the NH proton (H14) at 9.21 ppm. These observations suggest that the
spiro ring of KAE609 has expanded and rearranged. Based on the
NMR, exchangeable hydrogen atoms, and MS/MS results, a structure
was proposed for M37 as shown in Fig. 3. Because of the limited
quantity of samples, further structure confirmation via heteronuclear
multiple bond correlation (HMBC) experiments could not be per-
formed. The structure of M37 was further confirmed by chemical
synthesis.
Structural characterization of metabolites was carried out by LC-MS/MS analysis with
accurate mass measurements. The proposed structures of the metabolites were based on their
elemental composition derived from accurate mass measurements and fragment ions in their data-
dependent MS2 and MS3 mass spectra. Comparison of metabolite fragment ions with those of
KAE609 allowed the assignment of regions of biotransformation.
Compound [MH+] m/z Elemental Formula
Diagnostic Fragment Ions
KAE609 390.0573 C19H15Cl2FN3O 390, 347,a 319,a 311, 312, 284, 276
M14
M18
M20
M20.1
M21.2
M23
M24
M26
M34
M36
M37
M39
M40
M44
M46
M47
M48
404.0362
388.0416
404.0375
388.0425
404.0375
404.0363
404.0362
406.0515
406.0528 C19H15Cl2FN3O2
404.0042 C19H13Cl2FN3O2
390.0579
404.037
388.0419
402.0223
402.0209
401.9908
386.0256
C
C
C
C
C
C
C
C
19H13Cl2FN3O2
19H13Cl2FN3O
19H13Cl2FN3O2
19H13Cl2FN3O
19H13Cl2FN3O2
19H13Cl2FN3O2
19H13Cl2FN3O2
19H15Cl2FN3O2
404, 386,a 360, 351, 325, 278
388, 359,a 324
404,a 375, 340
388, 360,a 324
404,a 375,a 368,a 340,a 312
404, 375, 340, 235
404, 375, 340, 235
406,a 373, 327, 226a
406, 373,a 355,a 309,a 210a
404,a 375
Structural Characterization of M18 by LC-MS/MS and NMR
C
C
C
C
C
C
C
19H15Cl2FN3O
19H13Cl2FN3O2
19H13Cl2FN3O
19H11Cl2FN3O2
19H11Cl2FN3O2
19H11Cl2FN3O2
19H11Cl2FN3O
390, 373, 210,a 175
404, 355,a 337,a 319
388,a 359, 324, 210
402,a 339, 325, 197,a 172
402,a 310, 276, 172
402,a 367, 310
M18 ([MH]+ = 388) was an oxidative metabolite of KAE609 with a
net decrease of 2 Da in mass. The product ion spectrum showed
fragment ions at m/z 359 (loss of ethyl group) and 324 (loss of Cl radical
from m/z 359).
1H spectra of the metabolite M18 and KAE609 are shown in Fig. 4.
In the aliphatic region of the M18 spectrum, a multiplet was observed at
5.63 ppm corresponding to a methine proton H-8. The methyl group
(1.21 ppm) and the methylene protons (3.20 and 3.38 ppm) were
identified from COSY spectrum.
386,a 351, 323, 309
aIndicates the most abundant fragment ions.
experiment. The elemental formula and diagnostic fragmentation ions
of KAE609 and metabolites are summarized in Table 1. Detailed
product ion spectra of KAE609, M37, M18, and M23 can be found in
Supplemental Figs. 3–6.
1H spectrum of M18 showed five aromatic protons with coupling
patterns similar to the parent compound. However, the chemical shifts
of M18’s aromatic protons differed from KAE609, with H-1 showing
the most significant difference. The methine singlet observed in M37 is
absent. The MS results showed that M18 is 2 Da less than that of M37
and has one less exchangeable proton compare with M37. The LC-MS
and NMR results suggest that M18 appears to be derived from
previously identified M37. The structure of M18 was also confirmed
by chemical synthesis.
KAE609 ([MH]+ = 390) contained two chlorine atoms with an
isotope ratio of approximately 100:64 (35Cl:37Cl). The fragment ion at
m/z 347 was formed from loss of the vinyl amine moiety. Further loss of
the Cl radical from m/z 347 gave fragments at m/z 312, and loss of CO
from m/z 312 gave fragments at m/z 284.
Comparison of metabolite fragment ions with those of KAE609
allowed the assignment of regions of biotransformation. The product
ion spectrum of M37 ([MH]+ = 390) showed fragment ions at m/z 210
(loss of chloro-dihydro-quinazolin-one). There were no obvious similar
Structural Characterization of M23 by LC-MS/MS and NMR
M23 ([MH]+ = 404) was a hydroxylated metabolite of M18 with
fragmentation patterns between KAE609 and M37, which made the a net increase of 14 Da in mass versus KAE609. The product ion
assignment of M37 with only LC-MS/MS data not feasible. spectra of M23 showed fragment ions at m/z 375 and 340, which were
In the H/D exchange experiment, the full-scan mass spectra of 16 Da higher in mass than fragment ions from M18, supporting the
KAE609 and M37 showed a shift of the protonated molecular ion from assignment.
m/z 390 to MD+ at m/z 394 in KAE609 and from m/z 390 to MD+ at m/z
In the H/D exchange experiment, the full-scan mass spectrum of M18
393 in M37, suggesting that there are three exchangeable hydrogen and M23 showed a shift of the protonated molecular ion from m/z 388 to
atoms in KAE609 and only two exchangeable hydrogen atoms in M37 MD+ at m/z 390 for M18 and from m/z 404 to MD+ at m/z 407 for M23,
(data not shown).
consistent with the assignment that M23 was a hydroxylated metabolite
A comparison of 1H spectra acquired for the rat feces-derived sample of M18 (data not shown).
of M37 with KAE609 showed that M37 has five aromatic protons with
High-resolution MS data of compound 3, generated biocatalyti-
similar coupling patterns as were observed for KAE609 (Fig. 3). The cally, suggested a molecular formula of C19H12Cl2FN3O2 or one
chemical shifts of the M37 aromatic protons are different from KAE609, oxygen atom more than KAE609 and two hydrogen atoms less.
with the aromatic singlet of M37 at 7.29 ppm showing the most sig- Compared with the 1H-NMR spectrum of KAE609, the data of
nificant chemical shift difference relative to the parent molecule. These compound 3 showed only four aromatics protons (Table 2). The
NMR spectra suggest that no metabolism occurred on the aromatic rings presence of two aromatic protons (H-10 and H-13) which showed
of KAE609 during biotransformation to M37.
coupling to 19F suggested that the 6-chloro-5-fluoro-indole ring was
In the aliphatic region of the M37 spectrum, a multiplet was observed intact. However, the 5-chloro-indolone coupling pattern was no
at 4.95 ppm arising from the methine of M37 at H-8. The methyl group longer present. Instead, two singlets were observed at 6.20 and
(1.28 ppm) and the methylene protons (2.59, 3.09 ppm) were identified 8.25 ppm. There was no correlation in COSY and total correlation
from correlation spectroscopy (COSY). Both showed COSY correlations spectroscopy between these two protons. A rotational frame nuclear
to the H-8 proton. The chemical shift of the methine proton suggested that Overhauser effect spectroscopy correlation was observed between the
it was still connected to nitrogen as in KAE609’s structure.
NH-14 of the 6-chloro-5-fluoro-indole and the singlet at 8.25 ppm
The most significant difference between M37 and the parent drug (H-1). The detailed rotational frame nuclear Overhauser effect
spectra is an additional singlet observed at 5.84 ppm (1H) in the spectroscopy correlation spectrum of compound 3 can be found in

Ring Expansion Metabolites of KAE609 (a Spiroindolone)
659
Fig. 3. 1H NMR spectra for KAE609 and
M37 and NOESY spectrum for M37 isolat-
ed from rat fecal extracts after dosing of
[
14C]KAE609. 1H Spectra of M37 enrich-
ed from rat feces and authentic KAE609
were compared (top two panels). The one-
dimensional gradient NOESY spectrum was
acquired after the proton (at 5.84 ppm,
designated with an asterisk) was irradiated
to reveal the space correlations between the
aromatic singlet and methyl protons (bottom
panel). All samples were dissolved in
acetonitrile-d3. NOESY, nuclear Overhauser
effect spectroscopy.
Supplemental Fig. 7. The proton at 6.20 ppm (H-4) showed an HMBC KAE609 spectrum due to the anisotropy cone effect of the neigh-
correlation with the carbons C-2 and C-14c, confirming the position boring C=O.
on C-4. The chemical shift of C-3 at 174.1 ppm, extracted from the
HMBC experiment (H-1 to C-3) indicated that a hydroxylation had
occurred on position 3. The aliphatic CH resonance at 5.33 ppm
(H-8) showed a HMBC correlation to C-14b at 143.0 ppm. This chemical
NMR Analysis of Enriched M23 from Fecal Samples Compared
with Authentic Compound 3 from Bioreactions
Analysis of the 1H spectra revealed that the enriched M23-containing
shift indicated that C-14b was now sp2 hybridized in contrast with rat fecal sample included a chemical species with protons of the same
KAE609, where it was sp3. This suggested that the ring of the spiro chemical shifts and coupling patterns as that of compound 3: H-1, H-4,
had opened between C-14b and C-6 and rearranged. The chemical H-10, and H-13. In addition, 1H signals corresponding to H-4 for both
shift of H-8 indicated the formation of quinazolinone moiety. This the compound 3 and M23 samples appeared as broad singlets, sug-
was further supported by the down field shift compared with the gesting that the putative M23 H-4 proton is adjacent to a heteroatom

660
Huskey et al.
Fig. 4. 1H NMR spectra for KAE609 and M18 isolated
from rat hepatocytes. 1H spectra of M18 enriched from rat
hepatocyte incubations and authentic KAE609 were com-
pared. Both samples were dissolved in acetonitrile-d3.
bearing an exchangeable proton. The difference of proton chemical to obtain a partially deconvoluted picture of the 1H-1H spin systems in
shift of H-4 [6.20 ppm from compound 3 (Table 2) and 5.95 ppm from this region, a DQF-COSY spectrum was acquired. The DQF-COSY
M23 from fecal pool] may be attributed to a difference in sample spectrum shows a spin system with chemical shifts and coupling
concentration or pH due to the proximity of the hydroxyl at C-3.
patterns matching the CH3-CH-CH2 moiety in compound 3. Compar-
The aliphatic region of the M23-containing sample 1H spectrum was ison of NMR spectra of M23 derived from a rat fecal pool and
crowded with signals corresponding to matrix components. Therefore, compound 3 from biosynthesis can be found in Supplemental Fig. 8.

Ring Expansion Metabolites of KAE609 (a Spiroindolone)
661
TABLE 2
duct–cannulated rat, with the total accounting for approximately
1H- and 13C assignments and important HMBC correlations of compound 3 isolated 19% of the oral dose (data not shown).
from biosynthesis with recombinant human CYP1A2 in E. coli
Coupling constants are given in Hertz.
Discussion
Compound 3
Atom Numbera
KAE609 is an antimalarial agent that is progressing through clinical
development, providing a potentially new cure to combat malaria drug
resistance concerns (Rottmann et al., 2010; Meister et al., 2011;
Spillman et al., 2013). To facilitate the understanding of metabolism
and disposition of KAE609 in humans, ADME studies in nonclinical
species and across species in vitro metabolism studies were conducted.
After intravenous or oral dosing of [14C]KAE609, mass balance was
achieved in both rats and dogs. The radioactivity was primarily excreted
via biliary or fecal pathways (.93% of dose recovered in feces). After
oral dosing, KAE609 was well absorbed and extensively metabolized
in rats and dogs. KAE609 showed low clearance, moderate volume
distribution, and a long t1/2 in rats and dogs. The most abundant
radioactive component in rat and dog plasma was KAE609, accounting
for approximately75%–84% of the total radioactivity in plasma. M18
was identified to be the major circulating metabolite in rats and dogs,
accounting for approximately 10%–19% of the total radioactivity in
plasma. M37 was the major metabolite in rat feces, whereas M23 was
the major metabolite in dog feces. Despite M23 being the major
metabolite in dog feces, M23 was not detected in the circulation in
either rat or dog plasma after intravenous or oral dosing of KAE609.
As mentioned previously, M23 was identified as a human metabolite
based on in vitro incubations of KAE609 with hepatocytes. M23 was
not detected in rat and dog plasma from ADME studies, suggesting that
M23 may pose exposure coverage concerns if M23 is in the circulation
of humans. Therefore, to determine whether M23 is present in human
1H NMR
13C Shiftb
HMBC Correlations
1
2
3
8.25
s
128.3
130.4
174.1
H-1, H-4
H-1
3-OH
4
4a
6
8
8-Me
9R
9S
9a
9b
10
11
12
13
13a
14
14a
14b
14c
9.76
6.20
br s
br s
101.1
144.8
n.o.c
48.7
16.5
25.1
H-1
5.33
1.13
3.27
3.13
qd, 6.6, 6.1
d, 6.6
dd, 16.6, 6.4
d, 16.8
CH3-C8, H-9R, H-9S
H-9R, H-9S
CH3-C8
122.4
125.2
106.6
154.1
121.1
114.0
137.8
H-9R, H-9S
H-13
7.59
d, 9.2
H-13
H-10
7.77
d, 6.4
br s
H-10
10.44
127.0
143.0
102.5
H-9R, H-9S
H-1
H-4
br, broad; d, doublet; q, quartet; s, singlet.
aFor atom numbering, see Figure 5.
bThe 13C shifts were extracted from heteronuclear single quantum coherence and HMBC
correlation peaks.
cNot observed in the HMBC experiment.
In Vitro Metabolism of KAE609and M18 by Human Recombinant plasma, we conducted metabolite identification by LC-MS/MS using
CYP Enzymes
the plasma from the first-in-human study. M23 was confirmed to be
present in human plasma (data not shown). Consequently, we decided
to conduct a human ADME study to better define the exposure of M23
relative to total radioactivity exposure in plasma (Huskey et al.,
submitted manuscript). Our findings showed that M23 accounted for
approximately 12% of total radioactivity exposure in human plasma.
Based on the recommendation of Metabolites in Safety Testing and
International Conference on Harmonisation guidance (see U.S. Food
and Drug Administration 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), safety testing of M23 in nonclinical species is planned
accordingly.
To examine the roles of specific CYP enzymes involved in the
metabolism of [14C]KAE609, several recombinant CYP enzymes
were tested. The recombinant human CYP enzymes found capable
of oxidative metabolism of KAE609 were CYP2C9, CYP2C19,
CYP3A4, and CYP3A5 (Fig. 5A). M37 and M18 were among the
metabolites identified from the incubations. However, M23 was not
detected in any of the incubations with CYP enzymes.
As mentioned previously, the structure of M18 was proposed based
on NMR analysis and later confirmed by chemical synthesis (com-
pound 2). Thus, compound 2 (M18) was used as a substrate in in vitro
incubation to determine the CYP enzymes involved in the oxidative
metabolism of M18. M14 was the major metabolite generated in
incubations with all recombinant CYP enzymes tested. Trace levels of
M23 were detected in incubations with CYP1A1, CYP1A2, and
CYP1B1 (Fig. 5B).
To complement the aforementioned metabolite in safety testing,
it was critical to elucidate the structures of the metabolites involved in
the pathway leading to M23. KAE609, a spiroindolone derivative,
undergoes an unusual C-C bond cleavage, followed by a 1,2-acyl shift
to form a ring expansion product M37. This isomerization and re-
Proposed Metabolite Pathways of [14C]KAE609 in Rats and Dogs
Metabolic pathways of KAE609 in rats and dogs are summarized in arrangement reaction was catalyzed by CYP3A4 and the mechanism of
Fig. 6. KAE609 underwent C-C bond cleavage, followed by a 1,2-acyl this novel biotransformation reaction is proposed in Fig. 7. KAE609
shift and ring expansion to form M37, which underwent oxidation to undergoes oxidation by a single electron transfer to form a radical
form the colored metabolite M18, which was further hydroxylated to cation, followed by C-C bond cleavage to form an isocyanate radical
form another colored metabolite M23. M37 was the major metabolite cation intermediate. The isocyanate radical cation is reduced by a single
in rat feces, whereas M23 was the major metabolite in dog feces. electron transfer, followed by protonation. Nucleophilic attack by the
KAE609 also underwent oxidation to M40 and M20.1 with 2 mass amine on the carbonyl carbon of the isocyanate leads to ring closure to
units lower than that of KAE609 and one metabolite (M48) with form M37. A similar mechanism via a one-electron oxidized in-
4 mass units lower than that of KAE609. KAE609 was directly termediate has been proposed for the nonoxidative decarboxylation of
hydroxylated to form two metabolites (M26 and M34). Metabolites N-alkyl-N-phenylglycines by horseradish peroxidase in the absence of
(M40, M20.1, and M48) were further hydroxylated to several H2O2 and O2 (Totah and Hanzlik, 2002, 2004).
metabolites (M14, M20, M21.2, M24, M36, M39, M44, and M46).
As mentioned previously, structures of M37 and M18 were proposed
Several glucuronides were also identified in bile from the bile based on a combination of LC-MS/MS, exchangeable hydrogen atoms,

662
Huskey et al.
Fig. 5. In vitro metabolism of KAE609 and
compound 2 (M18) with human recombinant
CYP enzymes. [
14C] KAE609 (44 mM) or
compound 2 (M18; 50 mM) was incubated with
a panel of human recombinant CYP enzymes
(100 pmol CYP·ml21) or control microsomes in
the presence of NADPH. The reactions were
incubated at 37ꢀC for 30 minutes and quenched
with an equal volume of cold acetonitrile and
the products were analyzed by LC-MS/MS and/
or radioactivity monitoring. The assignment of
M23 in the extracted ion chromatogram was
based on MS/MS analysis. (A) Incubations with
CYP2C9 and CYP2C19 showed same results as
that with CYP3A5. (B) Incubations of com-
pound 2 with CYP1A1 and CYP1B1 showed
the same results as that with CYP1A2.
and NMR analyses. However, structural assignment for M23 was not
It is interesting that M23 was the major metabolite in dog feces and a
achievable due to an insufficient quantity of sample isolated from rat minor metabolite in rat feces; however, M23 was not in either rat or dog
feces. Therefore, our strategy was to use in vitro incubations to facilitate plasma. In contrast, M23 was the major metabolite in human plasma
the M23 structural assignment. Specifically, we incubated M18 with a and feces (Huskey et al., submitted manuscript). Apparently, M23 was
panel of human recombinant CYP enzymes and identified CYP1A1, generated enzymatically in the liver from KAE609 via intermediates
CYP1A2, and CYP1B1 as responsible for the formation of M23 from M37 and M18 in all three species. However, the distribution of these
M18. We then pursued this CYP-catalyzed reaction using recombinant three metabolites was different in rats and dogs versus humans. Thus,
human CYP1A2 (whole-cell biotransformation with genetically mod- we hypothesized that species differences in transporters dictate the
ified E. coli) to generate approximately 1 mg of M23 for structural distribution of M23 to blood (humans) or excretion into biliary or fecal
elucidation and evaluation of pharmacological activity.
pathways (rats, dogs, and humans).
The metabolic pathways of KAE609 are proposed (Fig. 6) based on
The prediction of circulating metabolites in humans and why certain
metabolite profiling and characterization efforts conducted using LC- metabolites circulate are topics of interest in drug metabolism research,
MS/MS and NMR analyses of metabolites. KAE609, a spiroindolone as discussed in several publications (Anderson et al., 2009; Dalvie et al.,
derivative, undergoes an unusual C-C bond cleavage, followed by a 2009; Smith and Dalvie, 2012). When species differences are observed
1,2-acyl shift to form a ring expansion product M37. This novel in the disposition of metabolites, as was the case for metabolite M23 of
biotransformation reaction was primarily catalyzed by CYP3A4. M37 KAE609, it may lead to the necessary safety evaluation of metabolites
was subsequently oxidized to M18 by CYP3A4 and hydroxylated to depending on their exposure in humans (see U.S. Food and Drug
M23 by CYP1A1, CYP1A2, and CYP1B1. All three metabolites had Administration 2008, 2010, and 2012, guidance available at http://
conjugated systems of double bonds, which imparted their character- www.fda.gov/cder/guidance/index.htm and http://www.fda.gov/Drugs/
istic color. As expected, M37 was colorless having two isolated GuidanceComplianceRegulatoryInformation/Guidances/default.htm).
conjugation systems. With only one additional double bond, M18 Thus, it is crucial to proactively address such concerns during early
and M23 gained an elongated conjugation system, resulting in M18 development to avoid potential delays in new drug application approval.
with an orange yellow color and M23 with a yellow color. The proposed
structures of the three metabolites were chemically synthesized and
thereby confirmed their structural assignments. They were found to be
pharmacologically inactive since the spiroindolone core, required for
encouragement and support, Dr. Robert Hanzlik (University of Kansas) for
Acknowledgments
The authors thank Dr. Francis Tse for constructive discussion and continued
pharmacological activity, was lost during biotransformation. Moreover,
safety evaluation of M23 showed it to be negative in the AMES test and
in the human lymphocyte chromosomal aberration test.
insightful and constructive discussion for the mechanism of the novel bio-
transformation reaction, Dr. Bryan Yeung for encouragement and insightful
suggestions in the chemical synthesis of metabolites, and Dr. Andreas

Ring Expansion Metabolites of KAE609 (a Spiroindolone)
663
Fig. 6. Proposed metabolic pathways of KAE609 in rats and dogs. Structures of M37, M18, and M23 were characterized by LC-MS/MS and NMR analyses and further
confirmed by chemical synthesis. However, the assignments of remaining metabolites were only based on LC-MS/MS analysis. Several glucuronides (M16.5, M21, M22.6,
M23.2, and M30) were identified in bile from bile duct–cannulated rats.
Fredenhagen for expert opinions and practical suggestions. The authors
also thank Dr. Zhigang Jian for synthesis of [14C]KAE609, Dr. Amy Wu
and Mr. Lawrence Jones for purification and analytical certification of
[
14C]KAE609, Dr. Ziping Yang and Ms. Cindy Chen for quantification of
KAE609 in rat and dog plasma, and Lolita Rodriguez for metabolic
profiling of KAE609 in biological matrices from rats.
Fig. 7. Proposed mechanism for the formation of M37. The structure of M37 was characterized by LC-MS/MS and NMR analyses and further confirmed by chemical synthesis.

664
Huskey et al.
Leong FJ, Li R, Jain JP, Lefèvre G, Magnusson B, Diagana TT, and Pertel P (2014) A first-in-
Authorship Contributions
human randomized, double-blind, placebo-controlled, single- and multiple-ascending oral dose
study of novel antimalarial spiroindolone KAE609 (Cipargamin) to assess its safety, tolera-
bility, 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, Bright
AT, and Che J, et al. (2011) Imaging of Plasmodium liver stages to drive next-generation
antimalarial drug discovery. Science 334:1372–1377.
Na-Bangchang K and Karbwang J (2013) Emerging artemisinin resistance in the border areas of
Thailand. Expert Rev Clin Pharmacol 6:307–322.
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 treatment
of malaria. Science 329:1175–1180.
Participated in study design: Huskey, Zhang, Favara, He.
Conducted experiments: Zhu, Forseth, Gu, Simon, Eggimann, Vargas, Li,
Wang, Zhang, Favara.
Performed data analysis: Huskey, Zhu, Lin, Forseth, Gu, Eggimann,
Kittelmann, Luneau, Einolf.
Wrote or contributed to the writing of manuscript: Huskey, Forseth, Simon,
Eggimann, Kittelmann, Luneau, Mangold.
Smith DA and Dalvie D (2012) Why do metabolites circulate? Xenobiotica 42:107–126.
Smith DA and Obach RS (2009) Metabolites in Safety Testing (MIST): considerations of
mechanisms of toxicity with dose, abundance, and duration of treatment. Chem Res Toxicol 22:
267–279.
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 cation
ATPase PfATP4 and is a target of the spiroindolone antimalarials. Cell Host Microbe 13:
227–237.
References
Anderson S, Luffer-Atlas D, and Knadler MP (2009) Predicting circulating human metabolites:
how good are we? Chem Res Toxicol 22:243–256.
Baillie TA (2008) Metabolism and toxicity of drugs. Two decades of progress in industrial drug
metabolism. Chem Res Toxicol 21:129–137.
Baillie TA, Cayen MN, Fouda H, Gerson RJ, Green JD, Grossman SJ, Klunk LJ, LeBlanc B,
Perkins DG, and Shipley LA (2002) Drug metabolites in safety testing. Toxicol Appl Phar-
macol 182:188–196.
Totah RA and Hanzlik RP (2002) Non-oxidative decarboxylation of glycine derivatives by a
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 in
Plasmodium falciparum isolates from Haiti. Am J Trop Med Hyg 92:552–554.
Dalvie D, Obach RS, Kang P, Prakash C, Loi CM, Hurst S, Nedderman A, Goulet L, Smith E,
and Bu HZ, et al. (2009) Assessment of three human in vitro systems in the generation of major
human excretory and circulating metabolites. Chem Res Toxicol 22:357–368.
Haque U, Glass GE, Haque W, Islam N, Roy S, Karim J, and Noedl H (2013) Antimalarial drug
resistance in Bangladesh, 1996-2012. Trans R Soc Trop Med Hyg 107:745–752.
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 development
in infected erythrocytes. Antimicrob Agents Chemother 59:3156–3167.
Huskey SE, Zhu C-Q, Fredenhagen A, Kühnöl J, Lumeaux A, Jian Z, Yang Z, Mia Z, Yang F,
Jain JP, Sunkara G, Mangold JB and Stein DS (2016) KAE609 (Cipargamin), a new
spiroindolone agent for the treatment of malaria: Evaluation of the Absorption, Distribution,
Metabolism and Excretion of a single oral 300 mg dose of [14C]KAE609 in healthy male
subjects Drug Metab Dispos (submitted for peer review)
peroxidase. J Am Chem Soc 124:10000–10001.
Totah RA and Hanzlik RP (2004) Oxidative and nonoxidative decarboxylation of N-Alkyl-N-
phenylglycines by horseradish peroxidase. Mechanistic switching by varying hydrogen per-
oxide, oxygen, and solvent deuterium. Biochemistry 43:7907–7914.
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 vivax
malaria. N Engl J Med 371:403–410.
Witkowski B, Khim N, Chim P, Kim S, Ke S, Kloeung N, Chy S, Duong S, Leang R,
and Ringwald P, et al. (2013) Reduced artemisinin susceptibility of Plasmodium falciparum
ring stages in western Cambodia. Antimicrob Agents Chemother 57:914–923.
Yeung BK, 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): a
new class of potent and orally efficacious compounds for the treatment of malaria. J Med Chem
53:5155–5164.
Kittelmann M, Kuhn A, Riepp A, Kühnöl J, Fredenhagen A, Oberer L, Ghisalba O, and Luetz S
(2012) Recombinant human cytochrome P450 enzymes expressed in Escherichia coli as whole
cell biocatalysts: preparative synthesis of oxidized metabolites of an mGlu5 receptor antago-
nist, in Practical Methods for Biocatalysis and Biotransformations 2 (Whittall J and Sutton PW
eds.) pp 138–146, John Wiley & Sons, Ltd., Chichester, UK.
Address correspondence to: Su-Er Wu Huskey, Drug Metabolism and Phar-
macokinetics, Novartis Institutes for BioMedical Research, One Health Plaza, East
Hanover, NJ 07936-1080. E-mail: Su.huskey@novartis.com