Utilization of stable isotope labeling to facilitate the identification of polar metabolites of KAF156, an antimalarial agent

Article abstract of DOI:10.1124/dmd.116.072108

Identification of polar metabolites of drug candidates during development is often challenging. Several prominent polar metabolites of 2-amino-1-(2-(4-fluorophenyl)-3-((4-fluorophenyl)amino)-8,8-dimethyl- 5,6-dihydroimidazo[1,2-a]pyrazin-7(8H)-yl)ethanone

Full text of DOI:10.1124/dmd.116.072108

Supplemental material to this article can be found at:
http://dmd.aspetjournals.org/content/suppl/2016/08/02/dmd.116.072108.DC1.html
1
D
521-009X/44/10/1697–1708$25.00
RUG METABOLISM AND DISPOSITION
http://dx.doi.org/10.1124/dmd.116.072108
Drug Metab Dispos 44:1697–1708, October 2016
Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics
Utilization of Stable Isotope Labeling to Facilitate the Identification of
s
Polar Metabolites of KAF156, an Antimalarial Agent
Su-Er W. Huskey, Ry R. Forseth, Hongmei Li, Zhigang Jian, Alexandre Catoire, Jin Zhang,
Tapan Ray, Handan He, Jimmy Flarakos, and James B. Mangold
Drug Metabolism and Pharmacokinetics, Novartis Institutes for BioMedical Research, East Hanover, New Jersey
Received June 17, 2016; accepted August 1, 2016
ABSTRACT
1
3
Identification of polar metabolites of drug candidates during devel- stage in development, stable isotope labeled [
6
C ]KAF156-1
opment is often challenging. Several prominent polar metabolites of
was available as the internal standard for the quantification of
KAF156. We were thus able to design an oral dose as a mixture of
2
-amino-1-(2-(4-fluorophenyl)-3-((4-fluorophenyl)amino)-8,8-dimethyl-
14
14
13
5
,6-dihydroimidazo[1,2-a]pyrazin-7(8H)-yl)ethanone ([ C]KAF156),
[
C]KAF156-1 (specific activity 3.65 mCi/mg) and [
C
6
]KAF156-1
] as 0.9 and the mass difference
as 6.0202. By using this mass filter strategy, four polar metabolites
characterized by liquid chromatography-tandem mass spectrome- were successfully identified in rat urine. Subsequently, using a similar
12
13
an antimalarial agent, were detected in rat urine from an absorption, with a mass ratio of [ C]/[
distribution, metabolism, and excretion study but could not be
C
6
14
13
try (LC-MS/MS) because of low ionization efficiency. In such
instances, a strategy often chosen by investigators is to use a
dual labeling approach, [ C]KAF156-2 and [
2
C ]KAF156-2 were
synthesized to allow the detection of any putative polar metabolites
radiolabeled compound with high specific activity, having an iso- that may have lost labeling during biotransformations using the
1
2
14
14
topic mass ratio (i.e., [ C]/[ C]) and mass difference that serve as
the basis for a mass filter using accurate mass spectrometry.
previous [ C]KAF156-1. Three polar metabolites were thereby iden-
tified and M43, a less polar metabolite, was proposed as the key
14
Unfortunately, [ C]KAF156-1 was uniformly labeled (n = 1–6) with intermediate metabolite leading to the formation of a total of seven
the mass ratio of ∼0.1. This ratio was insufficient to be useful as a
mass filter despite the high specific activity (120 mCi/mg). At this
polar metabolites. Overall this dual labeling approach proved practical
and valuable for the identification of polar metabolites by LC-MS/MS.
Introduction
KAF156 (Wu et al., 2011; Nagle et al., 2012; Fig. 1) is a first-in-class,
antimalarial compound designed to eradicate both blood-stage and liver-
stage malaria parasites (Meister et al., 2011; Kuhen et al., 2014; Diagana,
Malaria affects approximately 200 million people worldwide every
year, particularly among young children in developing countries. Recent
reports of drug-resistant strains to artemisinin, a first-line treatment of
uncomplicated Plasmodium falciparum malaria, in Southeast Asia cause
concerns for health authorities and the pharmaceutical industry (Haque
et al., 2013; Na-Bangchang and Karbwang, 2013; Witkowski et al.,
013). Increased efforts have been undertaken to understand the
mechanism of resistance toward the current artemisinin containing
combination therapy (Carter et al., 2015; Hott et al., 2015).
Several antimalarial compounds are currently in phase II development
and their structures are diverse, because they are either from existing or
new chemical scaffolds, targeting different mechanisms of action (Held
et al., 2015). Furthermore, other approaches including vaccination and
next generation antimalarial drug candidates have been evaluated for
efficacy and safety for malaria prophylaxis treatment (Amet et al., 2013;
Sagara et al., 2014; Diagana, 2015; Nahrendorf et al., 2015; Teneza-
Mora et al., 2015). Preventative treatment in children is expected to
potentially avert 11 million cases and 50,000 deaths every year.
2
015). In clinical trials, 2-amino-1-(2-(4-fluorophenyl)-3-((4-fluorophenyl)-
amino)-8,8-dimethyl-5,6-dihydroimidazo[1,2-a]pyrazin-7(8H)-yl)ethanone
KAF156) has been shown to be safe and well tolerated in healthy
(
volunteers (Leong et al., 2014). In the proof of concept trial, a single dose
of KAF156 was administered to healthy volunteers either prior to or after
exposure to P. falciparum infected mosquitoes. In both cases, KAF156
was equally effective and prevented malaria in every volunteer receiving
the drug. KAF156 has the potential to be used safely for malaria
prophylaxis for travelers to endemic countries, deployed soldiers, and
children living in high -risk regions.
2
We conducted rat absorption, distribution, metabolism, and excretion
(ADME) and across species in vitro metabolism studies in preparation
for the above clinical investigation during the development of KAF156.
Several metabolites were identified in rat plasma and feces by liquid
chromatography-tandem mass spectrometry (LC-MS/MS) analysis;
however, four prominent polar metabolites of KAF156, accounting
for ;24–37% of the intravenous and oral dose, were detected by
radioactivity in rat urine but could not be characterized by LC-MS/MS. It
was very challenging to identify these polar metabolites because of their
low ionization efficiency and suspected low molecular weights.
dx.doi.org/10.1124/dmd.116.072108.
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; DQF-COSY, double quantum filter correlation spectroscopy; FPOAA, 2-(4-fluorophenyl)-2-oxoacetic acid; HPLC, high-performance liquid
chromatography; KAF156, 2-amino-1-(2-(4-fluorophenyl)-3-((4-fluorophenyl)amino)-8,8-dimethyl-5,6-dihydroimidazo[1,2-a]pyrazin-7(8H)-yl)ethanone;
E
LC-MS/MS, liquid chromatography-tandem mass spectrometry; MS , data were acquired in parallel utilizing alternating low and elevated collision
energies; QC, quality control; t1/2, the terminal elimination half-life.
1697

1
698
Huskey et al.
1
4
13
Fig. 1. Structure of (A) [ C]KAF156-1, (B) [ C
6
]KAF156-1,
]KAF156-2, and (E)
6
]KAF156-3. [ C]KAF156-1 was uniformly la-
1
4
13
(
[
C) [ C]KAF156-2, (D) [ C
2
3 14
1
C
14
beled with C (n = 1-6), as indicated by *, whereas
13
13
[
6 6
C ]KAF156-1 and [ C ]KAF156-3 were labeled with
13
six C atoms in the fluoroaniline or fluorophenyl ring, as
indicated by **, respectively.
For the identification of metabolites in biologic matrices by accurate mass (St. Louis, MO). OPTI-FLUOR liquid scintillant was from Packard (Downers
Grove, IL). Control blank plasma samples from rat were purchased from
spectrometry, one of the strategies routinely chosen by investigators is the
use of radiolabeled compound with high specific activity, where the isotopic
Bioreclamation (Hicksville, NY).
12
14
mass ratio (e.g., [ C]: [ C] = 1: 0.3) is adequate to serve as a mass filter (Ma
14
Animals
et al., 2006; Zhang and Mitra, 2012). However, the available [ C]KAF156-
12
14
Male Wistar Hannover rats (;250–320 g, ;8 week, n = 24) were from Harlan
Laboratories (Somerville, NJ). Catheters were surgically implanted into the
carotid artery and/or jugular vein of rats by the vendor (only one catheter was
implanted into carotid artery for blood collection from rats receiving oral dose).
All rats were housed individually in metabolism cages (Culex Autosampler, BAS,
Indianapolis, IN) in a temperature and humidity controlled room with free access
to food and water (food was withheld until 4 hours postdose).
1
(Fig. 1) was uniformly labeled (n = 1–6) with the mean C/ C mass ratio
of ;0.1, which was too low to be suitable as a mass filter, although the
specific activity was reasonably high (120 mCi/mg). At this stage in
13
6
development, stable isotope labeled compound ([ C ]KAF156-1) was
available for use as an internal standard for the quantification of KAF156 in
toxicology studies under good laboratory practice. We took advantage of the
availability of this internal standard and designed an oral dose as a mixture of
1
4
13
[
6
C]KAF156-1 (specific activity 3.65 mCi/mg) and [ C ]KAF156-1 with Dose Administration
12
13
the mass ratio of [ C]/[ C ] as 0.9 and mass difference of 6.0202. In
principle, after dosing of [ C]KAF156 and [ C
6
All doses were administered based on the individual animal body weights on
12
13
]KAF156-1 to rats, any
14
14
6
the day of dosing. As shown in Fig. 1, [ C]KAF156-1 and [ C]KAF156-2 with
14
metabolites containing fluoroaniline portion would be generated as isotopic different C labeling positions and two stable isotope labeled forms of KAF156
pairs in vivo, sharing the signature mass ratio of 0.9 and mass difference of ([ C ]KAF156-1 and [ C ]KAF156-2) were synthesized and used for the dose
13
13
6
2
6
.0202, thereby rendering them easily identified using a defined mass filter preparations. A total of four subgroups of rats were dosed either intravenously or
1
4
orally. For mass balance studies, [ C]KAF156-1 (specific activity 66 mCi/mg
at 3 mg/kg for intravenous dose and 20 mCi/mg at 10 mg/kg for oral dose) and
[
by accurate mass spectrometry.
Subsequently, an alternative tracer ([ C]KAF156-2) was prepared to
allow detection of any putative polar metabolites that may have lost
radiolabeling and were not detected using [ C]KAF156-1 (Fig. 1). By
using a similar dual labeling strategy, [ C ]KAF156-2 was synthesized
1
4
14
C]KAF156-2 (specific activity 51 mCi/mg at 3 mg/kg for intravenous dose and
21 mCi/mg at 10 mg/kg for oral dose) were dissolved in 5% Solutol HS 15 for
dosing. For identification of polar metabolites, two oral doses were prepared as
1
4
1
3
2
14
13
follows: (A) [ C]KAF156-1 (specific activity 3.65 mCi/mg) and [ C
6
]KAF156-
] of 0.9 and (B)
2
C]KAF156-2 (specific activity 3.1 mCi/mg) and [ C ]KAF156-2 were
1
4
and combined with [ C]KAF156-2 in the oral dose to facilitate the
identification of polar metabolites.
12
13
1
[
were combined to achieve the mass ratio [ C]/[ C
6
14
13
12
13
In this article, we describe the strategies and approaches used in the combined to achieve the ratio [ C]/[ C ] of 0.9. Both doses were prepared in 5%
6
identification of a total of seven polar metabolites of KAF156. We also Solutol HS 15 for oral dosing at 10 mg/kg.
describe two unexpected hurdles we encountered during our metabolite
investigations and describe solutions to overcome the challenges.
Overall, the dual labeling approaches have proven to be valuable for
the identification of polar metabolites with low molecular weights and
low ionization efficiency.
Each rat received an intravenous bolus injection via the jugular vein cannula.
Oral dose was administered by gavage in rats.
Blood Collection
All rats were housed individually in Culex metabolism cages to enable
automated blood sampling on the day of the study. Blood samples (200 ml) were
collected from the carotid artery of rats at selected time intervals. Saline (200 ml)
was automatically injected after sample was collected to clear the cannula and
replace the volume of blood samples. The total blood volume collected did not
exceed 1% of the body weights of rats.
Materials and Methods
Chemicals
KAF156 (Wu et al., 2011; Nagle et al., 2012; Fig. 1) was synthesized by
1
4
14
Novartis Institute for Tropical Diseases [ C]KAF156-1, [ C]KAF156-2,
13
13
Urine and Feces Collection
[
6
C ]KAF156-1, and [ C
2
]KAF156-2 were prepared by Novartis Isotope
Laboratories (East Hanover, NJ) and radiochemical purity was .99%.
Urine and feces were collected daily from each of the animals for 7 days. For up
4-Fluoromandelic acid, fluoroaniline, hydroxyphenylacetamide, acetonitrile, to 3 days, the urine collection tubes were cooled with ice. After the final
methanol, ammonium formate, and formic acid were from Sigma-Aldrich
collection, each cage was rinsed with water followed by 50% methanol. The cage

Identification of Polar Metabolites of KAF156
1699
wash was assayed for radioactivity. Urine and feces samples were stored at Plasma for Metabolic Profiling of KAF156
220ꢀC until analysis.
All plasma samples were thawed at room temperature and aliquots (150 ml)
were pooled from each animal at each time point. Each pool was diluted with
Sample Preparation
2
50 ml water and extracted with 2 ml of acetonitrile:methanol:acetic acid (50:50:
Plasma was obtained by centrifugation of blood samples at 4ꢀC (2000 g) for 0.1, v/v/v). The samples were vortex-mixed followed by centrifugation. The
0 minutes. An aliquot of each blood sample was used for radioactivity analysis. supernatants were evaporated to near dryness under a stream of nitrogen. The
1
An aliquot of plasma sample was counted directly for radioactivity. The remaining residues were reconstituted with 100 ml of acentonitrile:deionized water (50:50;
plasma samples were stored at 220ꢀC until analysis.
v/v) and an aliquot was analyzed by LC-MS/MS.
14
14
Plasma for the Quantification of KAF156
Urine for Metabolic Profiling of [ C]KAF156-1 and [ C]KAF156-2
Blank plasma and study samples were thawed at room temperature. Calibration
standards, control blanks, and quality control (QC) samples were prepared on the
day of analysis by adding appropriate standard or QC (25 ml) spiking solution to
Urine pools (10–20% by volume) were prepared with samples collected from
rats during 0–24 or 0–72 hours postdose. The selected time interval
represented .91–95% of the total urinary excretion of radioactivity from rats by
either dosing route. The pooled samples were centrifuged at 3500 rpm for
475 ml of blank plasma. An aliquot (20 ml) of study samples, blanks, standards, or
QC samples was transferred to the designated well of a 1-ml round bottom 96-well
plate (Analytical Sales and Services, Pompton Plains, NJ), followed by the
addition of the internal standard ([ C ]KAF156; 25 ml, 500 ng/ml) and
6
acetonitrile (200 ml) for protein precipitation. Samples were mixed for 5 minutes
and centrifuged at 2500 rpm for 15 minutes at 25ꢀC. The filtrate was evaporated to
1
0 minutes, and aliquots of the resulting supernatant were analyzed by LC-MS/MS.
13
14
14
Feces for Metabolic Profiling of [ C]KAF156-1 and [ C]KAF156-2
Rat feces were collected in 24-hour intervals for 7 days. Fecal samples from
dryness at ;45ꢀC under a stream of nitrogen (TurboVap LV; Zymark Corp., each rat from each collection interval were homogenized with 2 ꢀ water
Taunton, MA). The residues were reconstituted with 300 ml of acetonitrile:water: separately. Fecal homogenates, representing ;5–10% of total weight, were
formic acid (10:90:0.1, v/v/v). An aliquot (10 ml) was analyzed for KAF156 and pooled from three rats and from 0- to 24-hour and 24- to 48-hour intervals. The
13
[
6
C ]KAF156-1 by LC-MS/MS.
selected time interval represented .93% of the total radioactivity excreted in rat
Fig. 2. Representative metabolic profiles in (A) pooled
fecal extracts and (B) pooled urine after oral dosing of
1
4
[
C]KAF156-1 to rats. Urine and feces samples were
collected daily from rats up to 7 days. After homogeni-
zation, fecal homogenates were pooled, extracted, and
analyzed by LC-MS/MS. The HPLC separation of KAF156
and metabolites was performed, as described in Materials and
Methods (method A).

1
700
Huskey et al.
feces by either intravenous or oral dosing route. The resulting fecal pools were
extracted three times with three volumes of acetonitrile:methanol:acetic acid
gradient was 2%B for 8 minutes; increased to 25%B from 8 to
18 minutes; increased to 60%B from 18 to 34 minutes; increased to
95%B from 34 to 38 minutes; hold at 95%B from 38 to 40 minutes. The
flow rate was 0.5 ml/min.
(
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 reconstituted with acetonitrile:deionized water (50/50, v/v) and
analyzed by LC-MS/MS. The excretion recoveries of radioactivity from rat feces
were ;68–78% (i.v.) and ;69–77% (oral), respectively.
The HPLC column eluent was split 1:4 (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 (Perkin Elmer, Downers Grove,
IL) coated with solid scintillant at 0.15 minutes/well using a fraction collector. The
plates were dried at 40ꢀC under a stream of nitrogen. The plates were counted for
Sample Analysis
1
0–15 minutes/well in a TopCount Model NXT radioactivity detector (Perkin-
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, urine samples, and cage wash
was 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 of 100 mM EDTA as an
Elmer Life Sciences, Waltham, MA). Online monitoring with b-RAM employed
a 250 ml liquid cell, and eluent was mixed with 1.4 ml/min IN-Flow 2:1 (IN/US
Systems, Tampa, FL) liquid scintillant. The resulting data were processed using
the Laura Data System (LabLogic, Inc., ver. 4.0, Brandon, FL). All quantification
was based on the radioactivity associated with the radiochromatographic peaks.
antifoaming agent and 200 ml of 30% hydrogen peroxide were added to LC-MS Instrumentation and Operating Conditions
decolorize the samples. The sample vials were loosely capped and returned to
The structural characterization of metabolites was carried out using the above
the water bath for 3 hours. Thereafter, 10 ml of Eq. 989 scintillation cocktail was
added and the samples were placed in the dark overnight to reduce chemilumi-
nescence before counting for radioactivity.
UPLC method coupled to a two-channel Z-spray (LockSpray) Waters Synapt G1
or G2 quadrupole time-of-flight mass spectrometer (Manchester, UK). The
Q-TOF Synapt was operated in V-mode with a typical resolving power of at least
Fecal samples from animals were homogenized with two or three volumes of
water. Duplicate samples (;100 mg) of the slurry were weighed into scintillation
vials and processed as described above for blood samples before radioactivity
determination. The remaining fecal homogenate was stored frozen at 220ꢀC until
LC-MS/MS analysis.
1
0,000. Qualitative analyses were carried out using electrospray ionization in the
positive or negative ion 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 5–10 ml/min.
E
Accurate mass LC/MS data were collected using MS approach where data
were acquired in parallel utilizing alternating low and elevated collision energies.
In low-energy MS mode, data were collected at consistent collision energy of 12 eV.
In elevated MS mode, collision energy was ramped from 15 to 30 eV during data
Quantification of KAF156 by LC-MS/MS
E
Samples were analyzed on an LC-MS/MS system consisting of a Shimadzu
HPLC System and Sciex API5000 mass spectrometer using Analyst software
version 1.4.2 (Foster City, CA). The mass spectrometer was operated in the
positive ion mode, using turbo spray ionization, with a source temperature of
collection cycle.
Metabolites were characterized by TOF MS full scan and product ion scans.
Product ion scans were obtained either from dedicated TOF MS/MS experiments
E
or the MS approach. With TOF MS/MS experiments, the trapping collision energy
5
4
00ꢀC. Chromatographic separation was carried out on a Zorbax SB-C8 50 ꢀ
13
was ramped from 15 to 20 eV, whereas the transferred collision energy was set at
6
.6 mm column at 40ꢀC. KAF156 and the internal standard ([ C ]KAF156-1,
E
6
eV. In the MS approach, the data were acquired in parallel using alternating low
Fig. 1) were eluted using a gradient method with a mobile phase consisting of (A)
and elevated collision energies. Argon was used as the collision gas for both product
1
1
0 mM ammonium acetate in water containing 0.03% trifluoracetic acid and (B)
0 mM ammonium acetate in methanol containing 0.03% trifluoracetic acid. The
E
ion scanning techniques. Acquiring data with the MS approach provided for
collection of intact precursor ions and fragment ion information.
gradient was 30%B for 0.01 minute increased to 95%B from 0.01 to 1 minute;
followed by a 1.2-minute hold at 95%B. The flow rate was 0.6 ml/min. The
multiple reaction monitoring transitions for KAF156 and [ C
m/z 412.4 to m/z 312.1 and m/z 418.5 to m/z 318.1, respectively.
13
]KAF156-1 were NMR Analyses of Polar Metabolites (M7, M10.1, and M12)
6
Detailed description for the synthesis of 2-(4-fluorophenyl)-2-oxoacetic acid
Calibration curves were generated by plotting the respective peak area ratios (y)
of KAF156 to the internal standard versus the concentrations (x) of the calibration
(
M7) and the isolation of polar metabolites from rat urine can be found in
2
standards using weighted 1/x quadratic least-squares regression. The quantifi-
TABLE 1
cation was performed using Analyst software and Watson LIMS version 7.2.0.01.
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.
Fragmentation patterns of KAF156 and metabolites under positive ion mode
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
2
3
dependent MS and MS mass spectra. Comparison of metabolite fragment ions with those of
KAF156 allowed the assignment of regions of biotransformation.
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
+
Elemental
Formula
a
Compound [MH ] m/z
Diagnostic Fragment Ions
(
pump and an online radioactivity monitor (b-RAM) or a fraction collector. Two
UPLC methods were developed for the separation of KAF156 and metabolites.
KAF156 412.1959
C
C
C
22
22
22
H
H
H
24
25
25
N
N
N
5
5
4
OF
2
412, 312, 245, 124, 101, 72
410, 310, 243,122, 101
431, 413, 373, 231
M19
M24
M27
M28
M31
410.2000
431.1909
O
O
2
F
F
3
2
Method A: The separation of metabolites was carried out on a Zorbax
SB-C18 column (150 ꢀ 3.0 mm, 3.5 mm) using a linear gradient with a
mobile phase consisting of (A) 5 mM ammonium formate containing
452.2107 C24H27N O F 452, 310, 243, 143, 101
411.1844
428.1898
5
4
5
3
3
2
C
C
22
H
24
N
O
O
F
F
411, 310, 243, 200, 102
428, 410, 371, 312, 307, 207, 168, 123,
11, 95, 70
22
H
24
N
2
1
0.1% formic acid and (B) acetonitrile containing 0.1% formic acid. The
M33
M35
383.1693
444.1847
C
C
21
H
H
21
N
N
4
5
OF
2
383, 355, 339, 312, 245, 124, 122
444, 426, 383, 369, 329, 312, 290, 286,
gradient was 2%B for 3 minutes; increased to 25%B from 3 to
22
24
O
3
F
2
3
9
2 minutes; increased to 60%B from 32 to 45 minutes; increased to
5%B from 45 to 52 minutes; hold at 95%B from 52 to 60 minutes. The
2
45, 193, 141, 123, 98, 70
M35.8
428.1890
C
22
H
24
N
5
O
2
F
2
428, 410, 371, 331, 314, 288, 245, 220,
177, 141, 124, 98, 70
454, 312, 245, 143, 101
429, 383, 355, 312, 272, 245, 124, 84
413, 312, 245, 124, 102
flow rate was 0.7 ml/min.
M36
M37
M37.7
454.2061
429.1720
413.1794
C
C
C
24
H
22
H
22
H
26
N
23
N
23
N
5
4
4
O
2
O
3
O
2
F
F
F
2
2
2
Method B: The separation of polar metabolites was carried out on a UPLC
HSS T3 column (150 ꢀ 3.0 mm, 1.8 mm) using a linear gradient with a
mobile phase consisting of (A) 5 mM ammonium formate containing
a
0.1% formic acid; and (B) acetonitrile containing 0.1% formic acid. The
The most abundant fragment ions are highlighted in bold.

Identification of Polar Metabolites of KAF156
1701
TABLE 2
Data Processing
Fragmentation patterns of KAF156 and metabolites under negative ion mode
Quantification of KAF156 and Metabolites by Radiometry. KAF156 and
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).
The percent of radioactivity (PRA) in a particular peak, Z, was calculated as
following:
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
2
3
dependent MS and MS mass spectra. Comparison of metabolite fragment ions with those of
KAF156 allowed the assignment of regions of biotransformation.
a
Compound [MH-] m/z Elemental Formula
Diagnostic Fragment Ions
KAF156 410.1792
C
C
C
C
C
C
C
C
C
22
H
22
N
5
OF
FS
S
2
410, 335, 310
DPMꢀ inꢀ peakꢀ Z
%
ꢀ PRAꢀ inꢀ Z ¼
ꢀ 100
totalꢀ DPMꢀ inꢀ allꢀ integratedꢀ peaks
M3
205.9927
230.0126
167.0145
168.0468
169.0309
248.0030
196.0416
260.0525
6
8
8
8
8
8
9
H
5
NO
NO
4
206, 188, 126, 106
230, 213, 150, 107, 97
167, 139, 95, 75
168, 126, 106
169, 125, 95
248, 168, 126, 97
196, 153, 95
M5
M7
H
8
4
7
6
7
5
H
H
H
H
H
FO
3
The concentration or amount of each component was calculated as %PRA (as a
fraction) multiplied by the total concentration (ngEq/ml) or percent of dose in the
excreta.
M10
M10.1
M11
M12
M43
NO
2
F
FO
3
NO
5
FS
3
7
FNO
NO
14
H
8
2
F
2
260, 232, 206, 166, 149, 136,
Pharmacokinetic Parameters
1
16, 110, 95
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
a
The most abundant fragment ions are highlighted in bold.
1
Supplemental Materials. H spectra were acquired for enriched M7, M10.1, and lower limit of quantification were treated as zero for pharmakokinetic parameter
M12 from urine samples as well as two reference compounds [4-fluoromandelic calculations. The linear trapezoidal rule was used for AUC calculation. Re-
acid and 2-(4-fluorophenyl)-2-oxoacetic acid; ;100 mg]. All samples were gression analysis of the terminal plasma elimination phase for the determination of
2
dissolved in acetonitrile-d
3
.
t
1/2 included at least three data points after Cmax. If the adjusted R value of the
NMR spectra were acquired on a Bruker UltraShield 600 MHz/54 mm regression analysis of the terminal phase was less than 0.75, no values were
spectrometer equipped with a 5 mm CP TXI 600S3 H-C/N-D-05 Z Cryoprobe. reported for t1/2, AUCinf, V /F, and CL/F.
z
For the double quantum filter correlation spectroscopy (DQF-COSY) spectrum,
1
the 90ꢀ pulse (P1) was calibrated at a H transmitter power of 0.2 dB, and the
following acquisition parameters were used: acquisition time = 0.75 second,
complex increments = 1024, and number of scans = 32. Spectra were processed in
MestReNova. DQF-COSY spectra were processed using 8k (f2) by 2k (f1) zero
filling. A sine squared window function was applied in both the t1 and t2
Results
Mass Balance and Pharmacokinetics of KAF156 in Rats
14
14
After intravenous or oral dosing of [ C]KAF156-1 or [ C]KAF156-2
dimensions. Detailed description for the NMR analyses of three polar metabolites (Fig. 1), mass balance was achieved in rats, with 93–118% of the
from rat urine can be found in Supplemental Materials.
administered radioactivity dose being recovered in excreta. Higher than
Fig. 3. Negative product ion spectra of M3 from urine after oral
1
4
13
dosing of [ C]KAF156-1 and [
C ]KAF156-1 to rats. The oral
6
12
14
dose consisted of [ C]KAF156, trace amount of [ C]KAF156-
13
1
and [
6
C ]KAF156-1. The mass ratio of 0.9 and mass
difference of 6.0202 Da was used as mass filters for the
12
13
detection of C and
12.0129 in (C)]. As expected, product ion spectra of M3
showed pairs of fragments with mass difference of 6.0202 Da
e.g., 126.0303 and 132.0504 in (A and B)].
6
C molecular ions [e.g., 205.9927 and
2
[

1
702
Huskey et al.
Fig. 4. Negative product ion spectra of M5 from urine after oral
1
4
13
dosing of [ C]KAF156-1 and [
C ]KAF156-1 to rats. The oral
6
12
14
dose consisted of [ C]KAF156, trace amount of [ C]KAF156-
and [ C ]KAF156-1. As expected, full and product ion
6
13
1
spectra of M5 showed isotopic pairs of molecular ions (C) and
corresponding fragment ions (A and B) with mass ratio of 0.9
and mass difference of 6.0202 Da.
1
00% of radioactivity recovered in excreta was considered within
After intravenous dosing, KAF156 had a moderate terminal half-life
experimental errors (e.g., pipetting, weighing, and counting, etc.). Two (t1/2; 6.6 hours), high plasma clearance (CL; 5.4 l/h/kg) and large volume
1
4
14
14
tracers ([ C]KAF156-1 and [ C]KAF156-2 with different
C
of distribution at steady state (Vss; 29.5 l/kg). After oral doing, KAF156
labeling positions) were used in mass balance studies. After in- absorption rate was moderate with Cmax of 139 ng/ml observed at
1
4
travenous or oral dosing of [ C]KAF156-1, excretion of radioactivity 2 hours (time to reach the maximum concentration after drug
was about equal (;40–56% in urine and ;61–63% in feces). administration). The extent of oral absorption was at least 55%, based on
1
4
However, after intravenous or oral dosing of [ C]KAF156-2, the ratio of the dose-normalized total radioactivity AUC for plasma
radioactivity was primarily excreted into feces (;23% in urine obtained from orally versus intravenously dosed rats (data not shown).
and ;68–75% in feces).
The estimated oral bioavailability was ;48%, indicating minimal
Fig. 5. Negative product ion spectra of M5 (A) urine from rats
1
4
13
dosed orally with [ C]KAF156-1 and [
urine from control rats, and (C) mixture of urine from (A) and
B). LC-MS/MS analysis was performed from (A) urine
6
C ]KAF156-1, (B)
(
14
samples from rats orally dosed with [ C]KAF156-1 and
C ]KAF156-1 and (B) urine collected from control rats
where molecular ion of m/z at 230 was detected. [M-H] of m/z
at 230 was detected with identical UPLC elution time when
above two samples were mixed together.
13
[
6
2

Identification of Polar Metabolites of KAF156
1703
Fig. 6. Representative metabolic profiles in (A) pooled fecal
14
extracts and (B) pooled urine after oral dosing of [ C]KAF156-
2
to rats. Urine and feces samples were collected daily from rats
up to 7 days. After homogenization, fecal homogenate were
extracted and analyzed by LC-MS/MS. The HPLC separation of
KAF156 and metabolites was performed, as described in Materials
and Methods.
first-pass effect in rats. Detailed summary tables of mass balance and M43) were identified in rat feces, each accounting for ;2–10% of the
pharmakokinetic parameters can be found in the Supplemental Tables dose by either dosing route.
S1 and S2.
1
4
Characterization of Metabolites of [ C]KAF156-1 in Rat Feces
1
4
14
Metabolite Profiling of [ C]KAF156-1 or [ C]KAF156-2 in Rat
Plasma
The structures of the metabolites were proposed based on their
elemental composition derived from accurate mass measurements (,3–
1
4
14
2
3
After intravenous dosing of [ C]KAF156-1 or [ C]KAF156-2, the 5 ppm), fragment ions in their data-dependent MS and MS mass
prominent radiolabeled components in plasma were KAF156 and three spectra. Elemental formula and diagnostic fragmentation ions of
oxidative metabolites (M31, M35.8, and M37), each accounting for KAF156 and metabolites are summarized in Tables 1 and 2.
+
;
12–35% of the total radioactivity AUC0-8 hours. Nomenclature of
KAF156 ([MH] = 412) produced fragments at m/z 312 (loss of ethyl
metabolites was based on their HPLC elution time in the study in which amino acetamide), 245 (loss of HCN from fluorophenyl imidazole
they were first identified. Detailed graphic presentation can be found in amine), 101 (ethyl amino acetamide), and 124 (fluoro-benzylideneamine)
the Supplemental Figs. S1 and S2.
when analyzed in positive ion mode. Detailed product ion spectra of
KAF156 can be found in Supplemental Fig. S3.
1
4
Metabolite Profiling of [ C]KAF156-1 in Rat Feces
1
4
Metabolite Profiling of [ C]KAF156-1 in Rat Urine
KAF156 was well absorbed and extensively metabolized in rats such
that unchanged KAF156 in feces accounted for ;8–21% of the dose by
either intravenous or oral dosing route (Fig. 2A). Several minor urine after an oral dose of [ C]KAF156-1 to rats (Fig. 2B). KAF156 and
metabolites (M19, M27, M31, M33, M35, M35.8, M37, M37.7, and six minor metabolites (M31, M35, M35.8, M37, M37.7, and M43) were
A representative radiochromatogram shows the metabolic profiles in
1
4

1
704
Huskey et al.
Fig. 7. Negative product ion spectra of M10.1 from urine
1
4
13
after oral dosing of [ C]KAF156-2 and [
C
2
]KAF156-2 to
12
rats. The oral dose consisted of [ C]KAF156, trace amount of
14
13
[
C]KAF156-2 and [
2
C ]KAF156-2. The mass response ratio
of 0.9 and mass difference of 2.0060 Da were used as mass
12
13
filters for the detection of C and
2
C molecular ions [e.g.,
169.0300 and 171.0381 at (A)]. As expected, product ion spectra of
M10.1 showed pairs of fragments with mass difference of 2.0060
Da [e.g., 125.0388 and 127.0745 at (B)].
identified in rat urine. Additionally, four relatively polar metabolites
However, M5 remained unidentified when the mass filters were set at
(
M3, M5, M10, and M11) were detected by radiochromatography, isotopic mass ratio of 0.9 and mass difference at 6.0202 Da. Only when
together accounting for ;35% of the intravenous dose and ;23% of the the isotopic mass ratio of 0.9 was removed from the mass filter, was M5
2
oral dose. Their structures were not identified because of their poor ([M-H] = 230) detected with mass difference at 6.0202 Da. M5 was
ionization efficiency under either positive or negative ion mode, their assigned to be a O-sulfonate conjugate of acetylated and hydroxylated
suspected low molecular weights, their relatively short UPLC elution aniline. Interestingly, M5 has lost fluorine versus other polar metabolites
times (;3–11 minutes), and endogenous urine matrix background. A (M3, M10, and M11). Presumably, oxidative defluorination at fluoroani-
new UPLC method was developed to extend the UPLC elution times of line has occurred to form M5. The product ion spectrum showed
the polar metabolites (;7–18 minutes), as described as method B in fragments at m/z 150 (loss of SO ) and 107 (loss of acetyl group from
3
1
3
2
13
Materials and Methods (graphic presentation can be found in Supple- m/z 150) (Fig. 4A). Similarly, [ C
6
]M5 ([M-H] = 236, C isotope ion)
mental Fig. S4). generated corresponding fragments at m/z 156 and 113 (Fig. 4B).
1
2
13
As shown in Fig. 4C, the ratio of [ C]/[ C ] changed from 0.9 to ;2,
6
1
4
Characterization of Polar Metabolites of [ C]KAF156-1 in Rat indicating that this metabolite had lost fluorine and coeluted with an
Urine
endogenous substance with the same molecular ion of m/z at 230.
Indeed, an endogenous substance with [M-H] at m/z 230 was detected
in urine from untreated rat (Fig. 5B), which coeluted with M5 in the
UPLC system used in the study (Fig. 5C).
2
To facilitate the identification of polar metabolites, a mixture of
C]KAF156-1 (specific activity 3.65 mCi/mg) and [ C ]KAF156-1 was
6
14
13
[
dosed orally to rats. By using the dual labeling approach, three polar
metabolites (M3, M10, and M11) were identified by LC-MS/MS analysis
using defined mass filters and analyzed under negative ion mode.
As shown in Fig. 3A, M3 ([M-H] = 206) was assigned as a
O-sulfonate conjugate of hydroxylated fluoroaniline. The product ion [ C]KAF156-1, another tracer, [ C]KAF156-2 (Fig. 1), was synthe-
spectrum showed the characteristic fragments m/z 126 (loss of SO ) and sized for an additional rat ADME study. Consistent with findings from
1
4
Metabolite Profiling of [ C]KAF156-2 in Rat Feces
2
To identify any putative metabolite that had lost radiolabeling using
1
4
14
3
1
3
2
14
1
06 (loss of HF from m/z 126). As expected, [ C ]M3 ([M-H] = 212, [ C]KAF156-1, KAF156 was well absorbed and extensively metabo-
C isotope ion) generated corresponding fragments at m/z 132 and lized in rats such that unchanged KAF156 in feces accounted for ;8–
6
1
3
1
12 (Fig. 3B). By using the same isotopic filtering approach, two polar 25% of the dose by either intravenous and oral dosing route (Fig. 6A).
metabolites (M10 and M11) and one less polar metabolite (M43) were Several metabolites (M19, M27, M33, M36, and M43) were identified in
identified and their fragment ions are summarized in Tables 1 and 2. rat feces, each accounting for ;5–13% of the intravenous or oral dose.
Detailed product ion spectra of M43 can be found in Supplemental Fig. S5. Several minor metabolites (M28, M31, M35, M35.8, M37, M37.7), each

Identification of Polar Metabolites of KAF156
1705
Fig. 8. Negative product ion spectra of M12 from urine after
1
4
13
oral dosing of [ C]KAF156-2 and [
C
2
]KAF156-2 to rats.
1
2
The oral dose consisted of [ C]KAF156, trace amount of
C]KAF156-2 and [
.0060 Da was lost, presumably M12 had lost one C during
14
13
[
2
C ]KAF156-2. The mass difference of
13
2
biotransformation. Two companion molecular ions (e.g., 196.0416
versus 197.0471) were detected and product ion spectra showed
pairs of fragment ions (152.0560 versus 153.1311).
accounting for ,3% of the intravenous or oral dose, were identified in ion) generated corresponding fragments at m/z 127 and 95. Apparently,
12
13
rat feces. Elemental formula and diagnostic fragmentation ions of the fragment ion at m/z 95 was the same for both [ C] and [ C
2
] mole-
KAF156 and metabolites are summarized in Tables 1 and 2. cules because the fragment ion of m/z at 95 excluded two stable isotope
labeled carbon atoms. By using the same approach, M7 was identified as
the corresponding fluorophenyl oxoacetic acid. The fragment ions of
these polar metabolites are summarized in Tables 1 and 2.
However, the metabolite M12 remained unidentified when the mass
filters were set at isotope mass ratio of 0.9 and mass difference at 2.002
Da. Presumably one of the two stable isotope labeled carbon atoms was
lost during biotransformation, and the mass difference was no longer
1
4
Metabolite Profiling of [ C]KAF156-2 in Rat Urine
A representative radiochromatogram shows the metabolic profiles in
urine after an oral dose of KAF156 to rats (Fig. 6B). KAF156 and several
minor metabolites (M19, M24, M31, M33, M35, M35.8, M36, M37,
and M43) were identified in rat urine. However, three relatively polar
metabolites (M7, M10.1, and M12), accounting for ;13–14% of the
dose by either intravenous or oral dosing route, remained unidentified
because of their low ionization efficiency under positive or negative ion
2
.002 Da (see below NMR analysis and Fig. 8).
mode and substantial urine matrix background. Four previously identified Structural Elucidation of Polar Metabolites by NMR
polar metabolites (M3, M5, M10, and M11) were detected by LC-MS/MS
and the absence of the C radiolabel in these chemical species precluded
their quantification via radiochromatography in this study.
4
-Fluoromandelic acid was purchased from Sigma Aldrich and 2-(4-
14
fluorophenyl)-2-oxoacetic acid was synthesized according to the published
procedure (Lee and Chen, 1991). The metabolites M7, M10.1, and M12
were isolated from rat urine, and their structures were explored using
NMR analyses via comparison with the two reference compounds of
14
Characterization of Polar Metabolites of [ C]KAF156-2 in Rat Urine
14
By using the same dual labeling approach, a mixture of [ C]KAF156- 4-fluoromandelic acid and 2-(4-fluorophenyl)-2-oxoacetic acid (FPOAA).
1
3
2
[
(specific activity 3.1 mCi/mg) and [ C ]KAF156-2 with the ratio of
DQF-COSY spectra acquired for the enriched M7-containing urine
2
1
2
13
C]/[ C
2
] as 0.9 was dosed orally to rats. By using the isotope mass samples were compared with authentic FPOAA. Additionally, DQF-
ratio of 0.9 and the mass difference of 2.006 Da as mass filters, two polar COSY spectra for the enriched M10.1-containing urine samples were
metabolites (M7 and M10.1) were identified using LC-MS/MS analysis compared with the 4-fluoromandelic acid standard. For both M7 and
under negative ion mode.
M10.1, the spectral data matched their corresponding authentic
2
As shown in Fig. 7, the metabolite M10.1 ([M-H] = 169) was standards as assessed by comparison of chemical shifts and coupling
assigned as 4-fluoromandelic acid. The product ion spectrum showed patterns. These NMR data supported the proposed structures based on
characteristic fragments m/z 125 (loss of CO
2
) and 95 (loss of CH
2
O
the initial mass spectrometric analysis of the polar metabolites M7 and
1
3
2
13
2
from m/z 125). As expected, [ C ]M10.1 ([M-H] = 171, C isotope M10.1.

1
706
Huskey et al.
Fig. 9. Proposed metabolic pathways of KAF156 in rats (A) less polar metabolites and (B) polar metabolites. Structures of metabolites were characterized by LC-MS/MS
analyses. Positive ion mode was used for the identification of less polar metabolites whereas negative ion mode was used for the identification of polar metabolites except
negative ion mode was used for the identification of a less polar metabolite M43. Proposed structures of M7, M10.1, and M12 were further confirmed by NMR analyses.
1
14
The H spectrum acquired for the M12-enriched NMR sample Proposed Metabolite Pathways of [ C]KAF156 in Rats
showed two peaks shifted downfield that integrate with whole-number
ratios relative to each other (7.93 and 8.66 ppm; relative integration, 2:1).
KAF156 was well absorbed and extensively metabolized in rat and
The chemical shift and relative integration of the signal at 8.66 ppm
Metabolic pathways of KAF156 in rats are summarized in Fig. 9.
unchanged KAF156 only accounted for ;8–25% of the dose by either
dosing route. Briefly, KAF156 underwent oxygenation (M31 and
suggested that this peak represents an amide proton. A DQF-COSY
spectrum acquired for the sample revealed that the putative amide proton
M35.8), oxidative defluorination (M19), oxidative deamination (M37),
couples to a methylene at 3.82 ppm. Additionally, the DQF-COSY
combination of oxidative deamination and reduction (M37.7), hydroxy-
coupling patterns observed for the aromatic protons at 7.93 ppm were
lation and ring opening (M24), acetylation (M36), and hydrolytic
indicative of an aryl group with a fluorine atom substituted at the para
cleavage, leading to the imidazole ring opening (M43). A combination
position. Further structural analysis via the acquisition of a rotational
of the above reactions and O-sulfonation led to the formation of several
frame nuclear Overhauser effect spectroscopy spectrum demonstrated
downstream less polar metabolites (M27, M28, M35, and M33) and
that the amide proton is proximal to the aromatic protons observed at
polar metabolites (M3, M5, M7, M10, M10.1, M11, and M12).
7
.93 ppm. The chemical shift and position of the methylene, distal to the
aromatic ring, was consistent with it being positioned between the amide
nitrogen and a carboxylic acid group. This conclusion was supported by
mass spectrometric analysis (Fig. 8) and comparison of observed NMR
chemical shifts to those reported for 4-fluoro-hippuric acid (Chavez
et al., 2010).
Detailed description of the FPOAA synthesis and NMR analyses
can be found in the Supplemental Materials and Supplemental Figs.
S6–S13.
Discussion
KAF156 is an antimalarial agent, which was designed to eradicate
both blood-stage and liver-stage malaria parasites, providing a poten-
tially new treatment of malaria prophylaxis (Kuhen et al., 2014;
Diagana, 2015). To facilitate the understanding of metabolism and
disposition of KAF156 in humans, ADME studies in rats and in vitro
across species metabolism studies were conducted.
1
4
After intravenous or oral dosing of [ C]KAF156-1, mass balance was
achieved in rats. The radioactivity was excreted equally via urinary and
In Vitro Metabolism of KAF156 in Rat and Human Hepatocytes
1
4
[
C]KAF156-1 (2.5 and 12.5 mM) was incubated in freshly prepared fecal pathways. KAF156 showed high clearance, large volume of
hepatocyte suspensions for 24 hours. The prominent metabolites were distribution and long terminal half-life in rats. After oral dosing,
M37.7, M19, M27, and M43 in rat hepatocytes; M36, M37.7, and M43 KAF156 was well absorbed and extensively metabolized in rats such
in human hepatocytes, respectively. However, no polar metabolites were that unchanged KAF156 accounted for only ;8–21% of the dose in
generated from hepatocytes, although M43 was a prominent metabolite feces by either dosing route. The prominent radioactive components in
in vitro (graphic presentation can be found in Supplemental Fig. S14).
rat plasma were KAF156 and three oxidative metabolites (M31, M35.8,

Identification of Polar Metabolites of KAF156
1707
Fig. 10. Proposed mechanism for the formation of M43 from KAF156
and M37). Several metabolites were identified in urine and feces by based on the detection of isotopic pairs of molecular ions with the
LC-MS/MS analysis. expected mass ratio and mass difference. Nevertheless, M5 was not
However, four prominent polar metabolites, accounting for ;23– identified until the criteria of mass ratio of 0.9 was removed from the
3
5% of the dose by either dosing route were not identified in rat urine mass filter. It turned out that M5 had lost fluorine and coeluted with an
2
because of their low ionization efficiency under either positive or endogenous substance with the identical molecular ion [M-H] of m/z at
negative ion mode and the endogenous urine matrix background. To 230, resulting in the change of mass response ratio of [ C]/[ C ] from
6
1
2
13
facilitate the structural characterization of polar metabolites in rat urine, 0.9 to ;2. Interestingly, M5 was the only polar metabolite that had lost
we developed a new UPLC method that prolonged the retention times of fluorine, presumably via oxidative defluorination of the fluoroaniline
these polar metabolites (Godejohann, 2007; Liu et al., 2010; Gray et al., ring, and shared the same elemental composition as an endogenous
2
011; Heaton et al., 2012). In addition, we purchased two commercially substance. This conclusion was derived based on the detection of the
2
available putative metabolites (fluoroaniline and hydroxyphenylaceta- endogenous substance [M-H] of m/z at 230 in urine from control rats
mide) and established their poor ionization efficiency under either without any treatment.
positive or negative ion mode.
Concurrently M43, an imidazole ring-opened metabolite, was
We decided to use the mass filter feature in the accurate mass identified using this dual labeling approach under negative ion mode
spectrometer to accelerate the identification of polar metabolites (Ma (Supplemental Fig. S5). Apparently, M43 was derived via hydrolytic
et al., 2006; Zhang and Mitra, 2012). We designed an oral dosing cleavage at the imidazole ring of KAF156, followed by amide hydrolysis
1
4
solution as a mixture of [ C]KAF156-1 (specific activity 3.65 mCi/mg) to form a putative intermediate (fluoroaniline), which was further
1
3
13
and [ C
6 6
]KAF156-1. At this stage of drug development, [ C ]KAF156-1 hydroxylated, oxidative defluorinated, acetylated, or O-sulfonated to
was available and used as the internal standard for the quantification of generate above four polar metabolites (M3, M5, M10, and M11).
KAF156 in toxicology studies, therefore our strategy does not require
extra research efforts and resources.
Consistent with the above in vivo findings, M43 was identified to be
the major metabolite when [ C]KAF156-1 was incubated with rat and
1
4
1
2
13
6
Theoretically, a mixture of [ C]KAF156 and [ C ]KAF156-1 with human hepatocytes. However, none of polar metabolites were detected
the 0.9 mass ratio and 6.0202 mass difference would provide the same from any in vitro incubations, although most less polar metabolites
information in terms of mass filter application in accurate mass identified in vivo were also generated in rat hepatocytes. The limitation
spectrometry. However, in our dual labeling approach, we strategically of the in vitro system observed for KAF156 is not unusual, because the
14
added trace amount of [ C]KAF156-1 to allow identification of any same experiences have been reported in the literature for many drug
putative metabolite that no longer satisfied either or both criteria (i.e., mass candidates. Therefore, the characterization of polar metabolites of
ratio or/and mass difference) of a defined mass filter. As described below, KAF156 could only be achieved from an in vivo study.
we indeed identified two metabolites (M5 and M12) that would have been
missed without the addition of radiolabeled tracer in the oral dose.
As expected, three polar metabolites (M3, M10, and M11) and one have lost the radiolabeling of [ C]KAF156-1 and stable isotope labeling
Accordingly, we were concerned that we may miss the identification
of putative metabolites derived from hydrolysis of M43, which may
1
4
1
3
6
less polar metabolite (M43) of KAF156 were easily identified in rat urine of [ C ]KAF156-1 (e.g., metabolites containing fluorophenyl ring). It

1
708
Huskey et al.
14
would be challenging to identify these putative polar metabolites under Jones for the purification and analytical certification of [ C]KAF156, and Dr.
Tapan Majumdar and Ms. Shari Wu for the quantification of KAF156 in rat plasma.
either positive or negative ion mode without any mass filter.
14
Subsequently, we initiated synthesis of another tracer ([ C]KAF156-2)
and strategically radiolabeled the carbon in the imidazole ring adjacent to
fluorophenyl ring to allow detection of any putative hydrolysis metabolites
Authorship Contributions
Participated in study design: Huskey, Jian, Ray, He, and Flarakos.
Conducted experiments: Forseth, Li, and Zhang.
Performed chemical synthesis: Forseth and Jian.
Performed data analysis: Huskey, Forseth, Li, Catoire, and Zhang.
Wrote or contributed to the writing of manuscript: Huskey, Forseth, and
Mangold.
14
of M43. After intravenous or oral dosing of [ C]KAF156-2, mass balance
was achieved in rats. However, another three polar metabolites in rat urine,
accounting for ;13–14% of the dose by either dosing route, were detected
but could not characterized by LC-MS/MS analysis due to low ionization
efficiency.
To enable the identification of polar metabolites using the same dual
labeling approach, we considered two strategies for the stable isotope
References
1
3
Amet S, Zimner-Rapuch S, Launay-Vacher V, Janus N, and Deray G (2013) Malaria prophylaxis in
patients with renal impairment: a review. Drug Saf 36:83–91.
6
labeling of KAF156: (A) [ C ]KAF156-3 with stable isotope labeling at
1
3
six carbon atoms of the fluorophenyl ring and (B) [ C ]KAF156-2 with
2
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.
Chavez F, Kennedy N, Rawalpally T, Williamson RT, and Cleary T (2010) Substituents effect on
the Erlenmeyer-Plöchl reaction: Understanding an observed process reaction time. Org Process
Res Dev 14:579–584.
the stable isotope labeling at two adjacent carbon atoms of the imidazole
portion of KAF156 (Fig. 1). Based on the fact that carbon atom adjacent
1
4
to fluorophenyl ring was radiolabeled in [ C]KAF156-2, we anticipated
1
3
14
[
C ]KAF156-2 and [ C]KAF156-2 together would provide us with
2
Diagana TT (2015) Supporting malaria elimination with 21st century antimalarial agent drug
more information regarding any putative metabolism occurring at
imidazolopiperazine moiety. We designed the oral dose as a mixture of
discovery. Drug Discov Today 20:1265–1270.
Godejohann M (2007) Hydrophilic interaction chromatography coupled to nuclear magnetic res-
onance spectroscopy and mass spectroscopy–a new approach for the separation and identifica-
tion of extremely polar analytes in bodyfluids. J Chromatogr A 1156:87–93.
Gray N, Musenga A, Cowan DA, Plumb R, and Smith NW (2011) A simple high pH liquid
chromatography-tandem mass spectrometry method for basic compounds: application to
ephedrines in doping control analysis. J Chromatogr A 1218:2098–2105.
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.
Heaton J, Gray N, Cowan DA, Plumb RS, Legido-Quigley C, and Smith NW (2012) Comparison
of reversed-phase and hydrophilic interaction liquid chromatography for the separation of
ephedrines. J Chromatogr A 1228:329–337.
14
13
[
C]KAF156-2 (specific activity 3.1 mCi/mg) and [ C ]KAF156-2 with
2
1
2
13
2
the mass ratio of [ C]/[ C ] of 0.9 and the mass difference of 2.006.
As expected, two polar metabolites (M7 and M10.1) containing
fluorophenyl ring were identified in rat urine using a defined mass filter.
However, M12 was detected by radioactivity but was not selected using
the above isotopic mass ratio and mass difference as mass filters.
Apparently, M12 lost one labeled carbon atom during biotransformation
and could no longer be detected using the above mass filters. Therefore,
we initiated the isolation of these polar metabolites from rat urine and
elucidated the structure of M12 by NMR and LC-MS/MS analyses. M12
was assigned as a glycine conjugate of fluorobenzoic acid, presumably
M7 underwent oxidative decarboxylation and followed by glycine
conjugation. Retrospectively, M12 could have been easily identified by
LC-MS/MS analysis using the dual labeling approach (strategy A). We
did not identify any other new metabolites containing portion of
imidazolopiperazine using strategy B.
Held J, Jeyaraj S, and Kreidenweiss A (2015) Antimalarial compounds in Phase II clinical de-
velopment. Expert Opin Investig Drugs 24:363–382.
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.
Kuhen KL, Chatterjee AK, Rottmann M, Gagaring K, Borboa R, Buenviaje J, Chen Z, Francek C,
Wu T, Nagle A, et al. (2014) KAF156 is an antimalarial clinical candidate with potential for use
in prophylaxis, treatment, and prevention of disease transmission. Antimicrob Agents Chemother
5
8:5060–5067.
Lee DG and Chen T (1991) The oxidation of alcohols by permanganate. A comparison with other
high-valent transition-metal oxidants. J Org Chem 56:5341–5345.
Leong FJ, Zhao R, Zeng S, Magnusson B, Diagana TT, and Pertel P (2014) A first-in-human
randomized, double-blind, placebo-controlled, single- and multiple-ascending oral dose study of
novel Imidazolopiperazine KAF156 to assess its safety, tolerability, and pharmacokinetics in
healthy adult volunteers. Antimicrob Agents Chemother 58:6437–6443.
Liu M, Ronk M, Ren D, Ostovic J, Cauchon N, Zhou ZS, and Cheetham J (2010) Structure elucidation
of highly polar basic degradants by on-line hydrogen/deuterium exchange hydrophilic interaction
chromatography coupled to tandem mass spectrometry. J Chromatogr A 1217:3598–3611.
Ma S, Chowdhury SK, and Alton KB (2006) Application of mass spectrometry for metabolite
identification. Curr Drug Metab 7:503–523.
Meister S, Plouffe DM, Kuhen KL, Bonamy GM, Wu T, Barnes SW, Bopp SE, Borboa R, Bright
AT, Che J, et al. (2011) Imaging of Plasmodium liver stages to drive next-generation antimalarial
drug discovery. Science 334:1372–1377.
Nagle A, Wu T, Kuhen K, Gagaring K, Borboa R, Francek C, Chen Z, Plouffe D, Lin X, Caldwell
C, et al. (2012) Imidazolopiperazines: lead optimization of the second-generation antimalarial
agents. J Med Chem 55:4244–4273.
The formation of imidazole ring opened metabolite M43 was further
investigated. As mentioned previously, M43 was identified to be the
1
4
major metabolite when [ C]KAF156-1 was incubated with rat and
human hepatocytes. In addition, M43 was detected at low level when
incubated in buffer without hepatocytes (pH 7.4) for 24 hours. These
findings suggest that the hydrolytic cleavage of imidazole ring was
catalyzed enzymatically and with some extent of chemical degradation
during hepatocyte incubations. However, M43 was not generated when
1
4
[
C]KAF156-1 was incubated with all 19 commercially available
Na-Bangchang K and Karbwang J (2013) Emerging artemisinin resistance in the border areas of
Thailand. Expert Rev Clin Pharmacol 6:307–322.
recombinant CYP enzymes in the presence of NADPH, indicating that
this is not CYP450 catalyzed reaction (data not shown). Therefore, a
two-step hydrolytic cleavage mechanism for the formation of M43 from
KAF156 is proposed (Fig. 10).
Nahrendorf W, Scholzen A, Sauerwein RW, and Langhorne J (2015) Cross-stage immunity for
malaria vaccine development. Vaccine 33:7513–7517.
Sagara I, Oduro AR, Mulenga M, Dieng Y, Ogutu B, Tiono AB, Mugyenyi P, Sie A, Wasunna M,
Kain KC, et al. (2014) Efficacy and safety of a combination of azithromycin and chloroquine for
the treatment of uncomplicated Plasmodium falciparum malaria in two multi-country randomised
clinical trials in African adults. Malaria J 13:458–467.
Overall, the dual labeling approach proved to be practical and
valuable in the elucidation of polar metabolites of KAF156 with low
molecular weights and low ionization efficiency by either positive or
negative mode. This approach has been shown to be superior to using
stable isotope labeling alone in cases where unexpected biotransforma-
tions occurred. However, the decisions regarding the positions of stable
isotope labeling and radiolabeling require careful consideration and
biotransformation insight.
Teneza-Mora N, Lumsden J, and Villasante E (2015) A malaria vaccine for travelers and military
personnel: Requirements and top candidates. Vaccine 33:7551–7558.
Witkowski B, Khim N, Chim P, Kim S, Ke S, Kloeung N, Chy S, Duong S, Leang R, Ringwald P,
et al. (2013) Reduced artemisinin susceptibility of Plasmodium falciparum ring stages in western
Cambodia. Antimicrob Agents Chemother 57:914–923.
Wu T, Nagle A, Kuhen K, Gagaring K, Borboa R, Francek C, Chen Z, Plouffe D, Goh A,
Lakshminarayana SB, et al. (2011) Imidazolopiperazines: hit to lead optimization of new anti-
malarial agents. J Med Chem 54:5116–5130.
Zhang Z and Mitra K (2012) Application of mass spectrometry for metabolite identification, in
ADME-Enabling Technologies in Drug Design and Development (Zhang D and Surapaneni S
eds) pp 317–330, John Wiley & Sons, Inc., Hoboken, NJ.
Acknowledgments
Address correspondence to: Su-Er Wu Huskey, Novartis Institutes for BioMedical
Research, One Health Plaza, East Hanover, NJ 07936-1080. E-mail: Su.huskey@
novartis.com
The authors thank Dr. Francis Tse for constructive discussions and continued
encouragement and support, Professor Phil Huskey (Rutgers University) for
constructive discussion of reaction mechanism, Dr. Amy Wu and Mr. Lawrence