Metabolism and pharmacokinetics of allitinib in cancer patients: The roles of cytochrome p450s and epoxide hydrolase in its biotransformations

Article abstract of DOI:10.1124/dmd.113.056341

Allitinib, a novel irreversible selective inhibitor of the epidermal growth factor receptor (EGFR) 1 and human epidermal receptor 2 (ErbB2), is currently in clinical trials in China for the treatment of solid tumors. It is a structural analog of lapatinib but has an acrylamide side chain. Sixteen metabolites of allitinib were detected by ultra-high-performance liquid chromatography/quadrupole time-offlight mass spectrometry. The pharmacologically active a,b-unsaturated carbonyl group was the major metabolic site. The metabolic pathways included O-dealkylation, amide hydrolysis, dihydrodiol formation, hydroxylation, and secondary phase 2 conjugation. The metabolite of amide hydrolysis (M6) and 27,28-dihydrodiol allitinib (M10) were the major pharmacologically active metabolites in the circulation. The steady-state exposures to M6 and M10 were 11% and 70% of that of allitinib, respectively. The biotransformation of allitinib was determined using microsomes and recombinant metabolic enzymes. In vitro phenotyping studies demonstrated that multiple cytochrome P450 (P450) isoforms, mainly CYP3A4/5 and CYP1A2, were involved in the metabolism of allitinib. Thiol conjugates (M14 and M16) and dihydrodiol metabolites (M5 and M10) were detected in humans, implying the formation of reactive intermediates. The formation of a glutathione conjugate of allitinib was independent of NADPH and P450 isoforms, but was catalyzed by glutathione-Stransferase. P450 enzymes and epoxide hydrolase were involved in M10 formation. Overall, our study showed that allitinib was metabolized by the O-dealkylation pathway similar to lapatinib, but that amide hydrolysis and the formation of dihydrodiol were the dominant metabolic pathways. The absorbed allitinib was extensively metabolized by multiple enzymes.

Full text of DOI:10.1124/dmd.113.056341

Supplemental Material can be found at:
http://dmd.aspetjournals.org/content/suppl/2014/03/05/dmd.113.056341.DC1.html
1521-009X/42/5/872–884$25.00
http://dx.doi.org/10.1124/dmd.113.056341
DRUG METABOLISM AND DISPOSITION
Drug Metab Dispos 42:872–884, May 2014
Copyright ª 2014 by The American Society for Pharmacology and Experimental Therapeutics
Metabolism and Pharmacokinetics of Allitinib in Cancer Patients: The
Roles of Cytochrome P450s and Epoxide Hydrolase in
its Biotransformations
Lishan Lin, Cen Xie, Zhiwei Gao, Xiaoyan Chen, and Dafang Zhong
Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
Received November 28, 2013; accepted March 5, 2014
ABSTRACT
Allitinib, a novel irreversible selective inhibitor of the epidermal allitinib was determined using microsomes and recombinant me-
growth factor receptor (EGFR) 1 and human epidermal receptor 2
(ErbB2), is currently in clinical trials in China for the treatment of
solid tumors. It is a structural analog of lapatinib but has an acryl-
amide side chain. Sixteen metabolites of allitinib were detected by
ultra-high-performance liquid chromatography/quadrupole time-of-
tabolic enzymes. In vitro phenotyping studies demonstrated that
multiple cytochrome P450 (P450) isoforms, mainly CYP3A4/5 and
CYP1A2, were involved in the metabolism of allitinib. Thiol conjugates
(M14 and M16) and dihydrodiol metabolites (M5 and M10) were
detected in humans, implying the formation of reactive intermediates.
flight mass spectrometry. The pharmacologically active a,b-unsaturated The formation of a glutathione conjugate of allitinib was independent
carbonyl group was the major metabolic site. The metabolic path- of NADPH and P450 isoforms, but was catalyzed by glutathione-S-
ways included O-dealkylation, amide hydrolysis, dihydrodiol for- transferase. P450 enzymes and epoxide hydrolase were involved
mation, hydroxylation, and secondary phase 2 conjugation. The in M10 formation. Overall, our study showed that allitinib was me-
metabolite of amide hydrolysis (M6) and 27,28-dihydrodiol allitinib tabolized by the O-dealkylation pathway similar to lapatinib, but that
(M10) were the major pharmacologically active metabolites in the amide hydrolysis and the formation of dihydrodiol were the dominant
circulation. The steady-state exposures to M6 and M10 were 11% metabolic pathways. The absorbed allitinib was extensively metab-
and 70% of that of allitinib, respectively. The biotransformation of olized by multiple enzymes.
Introduction
and activation of these receptors to prevent downstream signaling
events by binding to the ATP-binding site of protein kinases and
competing with the ATP substrate (Shewchuk et al., 2000). Although
the safety profile of lapatinib is acceptable when treating breast cancer
(Bence et al., 2005; Burris et al., 2005; Geyer et al., 2006),
hepatotoxicity has been reported in some patients treated with
lapatinib (Peroukides et al., 2011).
Allitinib (AST1306; N-(4-(4-(3-fluorobenzyloxy)-3-chlorophenyl-
amino)quinazolin-6-yl) acrylamide) is a novel analog of lapatinib. In
contrast to lapatinib, it is an irreversible inhibitor of EGFR and ErbB2.
Allitinib is currently being evaluated in clinical trials in China for the
treatment of solid tumors. Unlike other known irreversible EGFR and
ErbB2 inhibitors, such as HKI-272 (Rabindran et al., 2004),
BIBW2992 (Eskens et al., 2008; Takezawa et al., 2010; Yap et al.,
2010), and PF-00299804 (Gonzales et al., 2008), allitinib is more
effective in ErbB2-dependent models than in EGFR-dependent
models. Compared with lapatinib, allitinib is associated with greater
inhibition of EGFR and ErbB2. After incubation of allitinib at
different concentrations with kinases at 37°C for 1 hour, allitinib
exhibited IC50 values of 0.5 nM and 3.0 nM toward EGFR and ErbB2,
respectively. Furthermore, allitinib shows inhibitory activity against
The ErbB family of receptor tyrosine kinases consists of four
members: epidermal growth factor receptor (EGFR), ErbB2, ErbB3,
and ErbB4 (Klapper et al., 2000; Olayioye et al., 2000). Activation of
these receptor tyrosine kinases plays important roles in the prolifer-
ation, survival, adhesion, migration, and apoptosis of cells. ErbB2
predominantly exists in an active conformation for lack of any known
endogenous ligands, and it prevents the degradation of EGFR. Because
of these features, ErbB2 has attracted increasing attention in recent
years (Garrett et al., 2003; Xie et al., 2011). Abnormal expression of
these receptors is related to the development and migration of tumors
(Gonzales et al., 2008). Thus, the ErbB family of receptor tyrosine
kinases is an important therapeutic target for the design of anticancer
drugs.
Lapatinib (Tykerb, GlaxoSmithKline, Brentford, UK), a 4-anilino-
quinazolie derivative, is an orally administered small-molecule reversible
inhibitor of EGFR and ErbB2. Lapatinib blocks the phosphorylation
dx.doi.org/10.1124/dmd.113.056341.
s This article has supplemental material available at dmd.aspetjournals.org.
ABBREVIATIONS: a-NF, a-naphthoflavone; ABT, 1-aminobenzotriazole; AUC, area under the concentration-time curve; CE, collision energy;
DMSO, dimethylsulfoxide; EGFR, epidermal growth factor receptor; ER, efflux ratio; GSH, glutathione; GST, glutathione-S-transferase; HBSS,
Hanks’ balanced salt solution; HIM, human intestinal microsomes; HLC, human liver cytosol; HLM, human liver microsomes; HPLC, high-
performance liquid chromatography; HPM, human pulmonary microsomes; IDA, information-dependent acquisition; KET, ketoconazole; LC-MS/
MS, liquid chromatography2tandem mass spectrometry; M1, O-dealkylallitinib; M2, O-dealkyl-27,28-dihydrogenated allitinib; M5, O-dealky-27,28-
dihydrodiol allitinib; M6, metabolite of amide hydrolysis; M10, 27,28-dihydrodiol allitinib; NB-2, (E)-N-(4-((3-chloro-4-((3-fluorobenzyl)oxy)phenyl)
amino)quinazolin-6-yl)but-2-enamide; P450, cytochrome P450; Papp, apparent permeability coefficient; UHPLC/Q-TOF MS, ultra-high-performance
liquid chromatography/quadrupole time-of-flight mass spectrometry.
872

Metabolism of Allitinib in Humans
873
after the first dose; before and at 1, 2, 3, 4, 5, and 6 hours after the second dose;
and before and at 1, 2, 3, 4, 8, 12, 24, and 36 hours after the third dose.
Additional samples were obtained before and 2 hours after the dose on days 10
and 23. Plasma samples were separated and stored at 220°C until analysis.
Fecal samples were collected on days 10212. The total fecal weight was
recorded at the end of each collection interval. Five volumes of acetonitrile-
water (2:1, v/v) were added to each fecal sample. The mixture was blended by
a motor-driven homogenizer and then vibrated by ultrasound waves for 30
minutes. A 50-ml aliquot of the mixture was removed and centrifuged at 3,500 ꢀ g
for 5 minutes to produce fecal homogenate.
Another 10 subjects received a single oral dose of 1,000 mg of allitinib
tosylate. Urine samples were collected before and at intervals of 026, 6212,
12224, 24248, and 48272 hours after the start of oral administration. The
total urine volume was recorded at each collection interval. The fecal homog-
enate and urine samples were stored at 220°C until analysis.
mutant forms of EGFR. Animal studies have also demonstrated that
allitinib potently inhibits tumor growth in human tumor xenograft
mouse models (Xie et al., 2011).
The a,b-unsaturated carbonyl group of allitinib was pharmacolog-
ically designed to covalently bind to cysteine residue Cys797 in EGFR
and to Cys805 in ErbB2 via Michael addition (Xie et al., 2011).
However, because this a,b-unsaturated carbonyl group is electro-
philic, allitinib might covalently bind to proteins and other nu-
cleophilic biomolecules, with potentially adverse effects. To evaluate
the pharmacology and safety of new drugs, it is necessary to determine
their disposition and metabolism during the clinical development
program. In particular, the metabolites of new drugs might contribute
to their pharmacological or toxicity profiles in vivo (Baillie et al.,
2002). The pharmacokinetics, metabolism, and disposition of allitinib
have been investigated in rats (data not shown). Allitinib was absorbed
quickly, and the maximum concentration was achieved at 1.023.0
hours after administration. Its concentration decreased quickly, with
t1/2 of 3.424.4 hours. Systemic available allitinib was extensively
metabolized to thiol conjugates and dihydrodiol metabolites, which
indicated that allitinib tended to form reactive intermediates. The
bioavailability of allitinib was only 5.7% in rats, which might be
related to 1) poor solubility, 2) first-pass metabolism, 3) poor per-
meability, and 4) efflux transport.
Metabolite Profiling of Human Plasma, Urine, and Fecal Samples
Sample Preparation. Representative pooled plasma, urine, and fecal
samples were prepared for metabolite profiling and identification experiments.
Plasma samples taken at 0 hour on day 1 and at 2, 4, and 6 hours after the third
dose on day 24 were pooled using equal volumes (100 ml) from each individual
sample. Urine samples collected at 0 hour and from 0224 hours were pooled
across all subjects by combining volumes proportional to the total volume
excreted by each subject in each collection interval. Three volumes of
acetonitrile were added to each plasma and urine sample. After vortex mixing
and centrifugation at 11,000 ꢀ g for 5 minutes, the supernatant was transferred
into a clean plastic tube and evaporated to dryness under a stream of air at 40°C.
The residues of the plasma and urine samples were reconstituted in 100 ml of
90% aqueous acetonitrile (v/v). Then, 10 ml of the reconstituted solution was
injected into an ultra-high-performance liquid chromatograph coupled to
a quadrupole time-of-flight mass spectrometer (UHPLC/Q-TOF MS) system
for analysis. Fecal samples were pooled by combining volumes proportional to
the total volume of each fecal homogenate from all subjects. The pooled fecal
homogenates were directly injected into the UHPLC/Q-TOF MS system for
analysis.
The objectives of the current study were as follow: 1) to
characterize the metabolic pathways of allitinib in cancer patients at
the steady state following its repeated oral administration, 2) to
determine the pharmacokinetic profile and routes of excretion of
allitinib, and 3) to explore the mechanisms involved in the formation
of the major metabolites in vitro to understand the metabolism of the
acrylamide group.
Materials and Methods
Chemicals
Enzyme Hydrolysis
Reference standards of allitinib (99.7% purity), NB-2 ((E)-N-(4-((3-chloro-
4-((3-fluorobenzyl)oxy)phenyl)amino)quinazolin-6-yl)but-2-enamide, 96% pu-
rity), O-dealkyl-27,28-dihydrogenated allitinib (M2), and the metabolite of
amide hydrolysis (M6) were supplied by Allist Pharmaceuticals Inc. (Shanghai,
China). The metabolites O-dealkylallitinib (M1), O-dealky-27,28-dihydrodiol
allitinib (M5), 27,28-dihydrodiol allitinib (M10), and allitinib epoxide (M23)
were synthesized and purified in our laboratory. Pooled mixed-gender human
liver microsomes, pooled mixed-gender human liver cytosol, pooled mixed-
gender human intestinal microsomes, pooled mix-gender human pulmonary
microsomes, and recombinant cytochrome P450 (P450) enzymes (CYP1A2,
CYP1B1, CYP2A6, CYP2B6, CYP2C8, CYP2C9, CYP2C19, CYP2D6,
CYP2E1, CYP3A4, CYP3A5, and CYP4A11) were purchased from BD
Gentest (Woburn, MA). Caco-2 cells at passage 17 were purchased from the
American Type Culture Collection (Rockville, MD). b-Glucuronidase,
sulfatase, 1-aminobenzotriazole (ABT), a-naphthoflavone (a-NF), sulfaphena-
zole, ticlopidine, quinidine, clomethiazole, ketoconazole (KET), quercetin,
Hanks’ balanced salt solution (HBSS), and cell-culture reagents were acquired
from Sigma-Aldrich (St. Louis, MO). All solvents were of analytical or high-
performance liquid chromatography (HPLC) grade. Purified water was
generated with a Milli-Q Gradient Water Purification System (Millpore,
Molsheim, France).
Pooled human urine samples were incubated with sulfatase and b-glucuronidase
in 1 M citrate buffer (pH 5.0) in a water bath at 37°C for 16 hours. The mixtures
were treated as described above and analyzed.
Synthesis of M1
Allitinib was stirred with aluminum chloride in acetone at room temperature
for 10 hours. The mixture was then separated on a YMC-Pack ODS column
(250 ꢀ 10 mm I.D., 5 mM; YMC Company Ltd., Kyoto, Japan) to obtain pure M1.
Synthesis of M5 and M10
M1 and citric acid (0.2 molar equivalent) were dissolved in acetone/water (5:1,
v/v) in a 100-ml round-bottom flask. Potassium osmate (0.2 molar equivalent)
was then added, followed by 4-methyl-morpholin N-oxide (1.1 molar
equivalent). The mixture was stirred at room temperature for 48 hours and
quenched by the addition of sodium hyposulfite. Acetone was removed by
placing the flask on a rotary evaporator. The aqueous residue was extracted
with ethyl acetate and separated on a YMC-Pack ODS column (250 ꢀ 10 mm
I.D., 5 mM; YMC Company Ltd.) to yield pure M5. The M10 synthetic
standard was acquired with the same chemical reaction after replacing M1 with
allitinib.
Clinical Study Design
Synthesis of Allitinib Epoxide
The study was performed at Fudan University Shanghai Cancer Center
(Shanghai, China) and was approved by the Ethics Committee of Fudan
Allitinib tosylate (0.18 mmol) was dissolved in 3.75 ml formic acid, and
University Shanghai Cancer Center. All subjects gave their written informed 2 ml H2O2 was added. The contents were heated at 45°C for 3 hours. After
consent before enrollment. Three Chinese cancer patients (two males and one
female) were given 1,000 mg of allitinib tosylate three times a day orally for 21
consecutive days. Blood samples were collected in heparinized tubes at the
following times on days 1 and 24: before and at 0.5, 1, 1.5, 2, 3, 4, and 6 hours
adding two volumes of acetonitrile-water (2:1, v/v) and adjusting the pH value to
3 with 1% ammonia, the mixture was then separated on a Shim-Pack PREP-ODS
(H) KIT column (250 ꢀ 20 mm I.D., 5 mM; Shimadzu, Kyoto, Japan) to yield
pure allitinib epoxide.

874
Lin et al.
Fig. 1. Q-TOF MS of allitinib under high CE in the ESI+ mode (A) and tentative structures of the major fragment ions of allitinib (B). The structure of allitinib can be
divided into three segments according to the fragmentation pattern.
Pharmacokinetic Analysis
In Vitro Pharmacological Activity of the Metabolites of Allitinib
The pharmacokinetic characteristics of allitinib, M6, and M10 were analyzed
using the noncompartmental model with Winnonlin 5.3 software (Pharsight,
Mountain View, CA). The Cmax and the time to reach the maximum concen-
tration were directly determined from the experimental data. The linear-log
trapezoidal method was used to calculate the area under the plasma concentration-
time curve for 0224 hours (AUC0–24 hours). The accumulation ratio (day 24/day 1)
The inhibitory activities of M1, M6, and M10 against EGFR and ErbB2
were determined as previously described (Xie et al., 2011).
Microsomal Incubation
A stock solution of allitinib was prepared in dimethyl sulfoxide (DMSO). A
for AUC0224 was also calculated. The elimination rate constant (ke) was mixture containing 1-mM allitinib was mixed with human liver microsomes
determined by least squares regression of the terminal log-linear phase of the
concentration-time curve. The terminal half-life was estimated as ln2/ke.
(HLM) (1 mg/ml), human intestinal microsomes (HIM) (1 mg/ml), and human
pulmonary microsomes (HPM) (1 mg/ml), in addition to phosphate-buffered

Metabolism of Allitinib in Humans
875
Fig. 2. Metabolic profiles of pooled plasma samples at the peak (A) (2 hours after the third dose on day 24) and trough (B) (4 hours after the third dose on day 24) steady-
state levels of allitinib after the oral administration of 1,000 mg of allitinib tosylate three times a day for 21 days to cancer patients. Upper trace, UHPLC-UV chromatogram.
Lower trace, mass defect filter–processed chromatogram.
saline (PBS) (100 mM), at pH 7.4 without or with 1 mM glutathione (GSH) at
a final volume of 200 ml. The final DMSO concentration was 0.1%. After
preincubation at 37°C for 3 minutes, 1 mM NADPH was added to initiate the
reaction. After incubation for 60 minutes, two volumes of ice-cold acetonitrile
were added to terminate the reaction. Control samples without NADPH or
microsomes were also prepared. To evaluate the contribution of glutathione
transferase (GST) to the formation of the reactive metabolites, separate samples
were prepared with 1 mg/ml human liver cytosol (HLC), 1 mM allitinib, and
1 mM GSH. Each incubation was performed in duplicate. After the samples
were centrifuged at 11,000 ꢀ g, the supernatants were evaporated to dryness
under an air stream at 40°C. The residues were reconstituted in 100 ml water/
acetonitrile (90:10, v/v). Then, a 10-ml aliquot of the reconstituted solution was
injected into the UHPLC/Q-TOF MS system for analysis.
Inhibition of the Oxidative Metabolism of Allitinib by Selective P450
Inhibitors
To evaluate the relative contributions of different microsomal enzyme
systems to the metabolism of allitinib in HLM, 1 mM allitinib was incubated at
37°C for 60 minutes with 1 mg/ml HLM alone or in the presence of different
chemical inhibitors of P450, including the nonspecific P450 inhibitor
1-aminobenzotriazole (ABT) (1 mM), the specific CYP1A1/2 inhibitor a-NF
(2 mM), the CYP2C8 inhibitor quercetin (1 mM), the CYP2C9 inhibitor
sulfaphenazole (6 mM), the CYP2B6/2C19 inhibitor ticlopidine (0.4 mM), the
CYP2D6 inhibitor quinidine (2 mM), the CYP2E1 inhibitor clomethiazole (0.1
mM), or the CYP3A4/5 inhibitor KET (1 mM). After incubation, two volumes
of ice-cold acetonitrile containing NB-2 were added to terminate the reaction.
Each incubation was performed in duplicate.
The MS peak area ratios of allitinib and the detected metabolites to the
internal standard in each incubation system were recorded to determine the
contributions of the P450 enzymes to the metabolism of allitinib. The results
were compared with those for control samples lacking the inhibitors.
Identification of CYP Enzymes Involved in the Metabolism of Allitinib
A stock solution of allitinib was prepared in DMSO. Allitinib (1 mM) was
incubated in the presence of individually expressed recombinant P450 enzymes
(50 pmol/ml 1A2, 1B1, 2A6, 2B6, 2C8, 2C9, 2C19, 2D6, 2E1, 3A4, 3A5, or
4A11 suspended in 100-mM PBS, pH 7.4) and 2-mM NADPH at 37°C for 60
minutes. The final DMSO concentration during incubation was 0.1%. Two
volumes of ice-cold acetonitrile containing an internal standard (500 ng/ml NB-2)
were added to terminate the reaction. A sample incubated without NADPH was
used as the negative control.
Mechanism of M10 Formation
Valpromide (10, 100, or 1,000 mM), an inhibitor of epoxide hydrolase, was
preincubated with HLM and NADPH for 3 minutes. Allitinib was added to the
mixture to a final volume of 200 ml. A control sample without valpromide was
Fig. 3. Metabolic profiles of pooled urine samples collected at 0224 hours after the oral administration of 1,000 mg of allitinib tosylate daily in cancer patients (A) and
pooled feces samples collected at days 10 to 12 after the oral administration of 1,000 mg of allitinib tosylate three times a day for 21 days in cancer patients (B). Upper trace,
UHPLC-UV chromatogram. Lower trace, mass defect filter–processed chromatogram.

876
Lin et al.
TABLE 1
UHPLC/Q-TOF MS data for allitinib metabolites detected in human plasma, urine, and feces
m/z
Error
(ppm)
RT
(min)
No.
Description
Formula
Fragment Ions
Matrices
P, U, F
[M + H]+
M0
Allitinib
449.116
C24H18ClFN4O2
24.5
14.0
340.074, 339.069, 313.054, 311.074,
305.109, 287.071, 277.107, 213.081
305.108, 277.111
M1
M2
O-Dealkylation
O-Dealkylation + alkenes
reduction
341.082
343.094
C
C
17H13ClN4O2
17H15ClN4O2
3.2
25.7
6.1
6.1
P, U, F
P, F
251.095, 287.075, 279.127, 307.117
M3
M4
O-Dealkylation + hydroxylation
O-Dealkylation + hydration
357.072
359.087
C
C
17H13ClN4O3
17H15ClN4O3
29
211.1
6.1
4.9
329.089, 321.111, 229.107
329.079, 313.083, 287.068,
277.103, 251.099
U, F
F
M5
O-Dealkylation + alkenes to
dihydrodiol
375.084
C17H15ClN4O4
25.6
3.5
313.052, 287.073, 251.096
P, U, F
M6
M7
M8
Amide hydrolysis
395.108
411.103
437.034
C
C
C
21H16ClFN4O
21H16ClFN4O2
17H13ClN4O6S
0.6
0.5
4.5
11.5
11.4
5.0
287.075, 286.062, 251.095
287.075, 286.062, 251.094
357.079, 329.095, 321.095,
303.080, 229.046
P, U, F
P, U, F
U
Amide hydrolysis + hydroxylation
O-Dealkylation + hydroxylation +
sulfation
M9
O-Dealkylation + cysteine
conjugation
Alkenes to dihydrodiol
462.107
483.123
491.056
504.109
517.115
570.138
587.125
587.141
612.153
C20H20ClN5O4S
C24H20ClFN4O4
14.8
20.2
27.1
24.4
4.1
3.6
11.0
7.3
341.085, 313.057, 305.098,
287.071, 277.119
375.072, 374.077, 357.072, 313.048,
287.073, 251.095
U, F
P, U, F
P
M10
M11
M12
M13
M14
M15-1
M15-2
M16
Αmide hydrolysis + hydroxylation +
sulfation
O-Dealkylation +N2acetylcysteine
conjugation
O-Dealkylation + glucuronide
conjugation
Cysteine conjugation
C
21H16ClFN4O5S
411.120, 286.063
C22H22ClN5O5S
C23H21ClN4O8
C27H25ClFN5O4S
C27H24ClFN4O8
4.9
375.063, 341.080, 287.068, 251.098
341.08
F
3.7
P, U
P, U, F
P
0
9.9
483.101, 461.117, 449.118, 432.162,
426.117, 398.072, 313.051, 287.075
411.102, 286.061
Αmide hydrolysis + hydroxylation +
glucuronide conjugation
Αmide hydrolysis + hydroxylation +
glucuronide conjugation
215.6
10.8
7.5
5.8
C
27H24ClFN4O8
9.7
411.102, 303.061, 302.059, 267.090,
229.130
449.123, 341.080, 340.085, 313.050,
311.081, 305.105, 287.063, 251.097
P, U
U
N-Acetylcysteine conjugation
C29H27ClFN5O5S
11.0
RT, retention time; P, plasma; U, urine; F, feces.
also prepared. Each incubation was performed in duplicate. After incubation for
60 minutes, two volumes of ice-cold acetonitrile were added to terminate the
reaction. After protein precipitation, the supernatant was evaporated to dryness
37°C for 2 hours in a humidified atmosphere containing 5% CO2. Next, 200 ml
of transport buffer was collected from the apical or basolateral side and mixed
with the same volume of methanol. HBSS containing allitinib at the donor side
at 40°C under a gentle stream of air. The residue was reconstituted in 80.0 ml of was collected before and after the assay to calculate assay recovery. After
90% aqueous acetonitrile (v/v). The MS peak area ratios of M10 to epoxide in
each incubation system were recorded to determine the role of valpromide in
the formation of M10. The results were compared with those for control
samples lacking valpromide.
washing three times, cell monolayers were lysed with 200-ml acetonitrile. All
incubations were performed in triplicate. The concentration of allitinib was
analyzed by LC-MS/MS.
The apparent permeability coefficients (Papp, cm/s) were calculated with the
following equation:
Incubation of Allitinib Epoxide in HLC
DQ
1
Papp
¼
ꢀ
Stock solutions of allitinib epoxide were prepared in DMSO. The allitinib
epoxide (20 nM or 2 mM) was incubated with 1 mg/ml HLC, GSH (2 mM),
magnesium chloride (5 mM), EDTA (0.1 mM), and PBS (100 mM, pH 7.4).
The final DMSO concentration during incubation was 0.1%. The incubation
was conducted in duplicate at 37°C in a total volume of 200 ml. The reaction
was terminated after 60 minutes of incubation by adding two volumes of ice-
cold acetonitrile. Control samples without HLC were also prepared. After
protein precipitation, the supernatant was evaporated to dryness at 40°C under
a gentle stream of air. The residue was reconstituted in 100 ml of 90% aqueous
acetonitrile (v/v). An aliquot of 10 ml of the reconstituted solution was injected
into the UHPLC/Q-TOF MS for analysis.
Dt A ꢀ Cdonor
where nQ/nt is the rate of permeability (nmol/s), A is the surface area of the
insert (cm2), and Cdonor is the initial concentration on the donor side (nmol/ml).
The efflux ratio (ER) was determined as follows:
PappðB 2 AÞ
ER ¼
PappðA 2 BÞ
where Papp(A2B) and Papp(B2A) represent the apparent permeability coefficients
of the substrate from the apical to the basolateral side and from the basolateral
to the apical side, respectively.
Caco-2 Cell Permeability Study
Analytical Conditions
For the cell permeability experiments, Caco-2 cells were seeded in 24-well
inserts at a density of 105 cells/cm2 to generate Caco-2 monolayers. The cells
were cultured for 21 days, during which time the medium was replaced every 3
days. Cell layers with a transepithelial electrical resistance value . 300 V×cm2
were used in the permeability experiment. Before the assay, the monolayers
were gently washed twice with warm HBSS (pH 7.4, 37°C). HBSS containing
2-mM allitinib was added to the apical side (150 ml) or the basolateral side (750
ml) of the inserts to initiate the experiment. The cells were then incubated at
Metabolite Profiling by UHPLC-UV/Q-TOF-MS. Metabolic profiling of
allitinib in biologic samples was performed on a Waters Acquity UHPLC
system (Waters, Milford, MA) equipped with a binary solvent delivery pump,
column oven, UV detector, and autosampler. Chromatographic separation was
performed on an Acquity UPLC HSS T3 column (100 ꢀ 2.1 mm I.D., 1.8 mm;
Waters) at 40°C. The mobile phase was a mixture of 5-mM ammonium formate
in water containing 0.05% formic acid (A) and acetonitrile (B) at a flow rate of

Metabolism of Allitinib in Humans
877
Fig. 4. Proposed metabolic pathways of allitinib in humans. Allitinib is extensively metabolized in vivo by O-dealkylation, amide hydrolysis, 27,28-dihydrodiolation, and
hydroxylation of the quinoline ring and the 3-fluorobenzyloxy moiety, combinations of the above pathways, and secondary phase 2 conjugation.
0.4 ml/min. Elution started with a 1-minute isocratic run with 10% solvent B,
followed by a linear gradient of 10% to 55% solvent B in 14 minutes, 55% to
99% solvent B in 1 minute, maintained for 1 minute, and then reduced to
10% B to equilibrate the column. The eluate was monitored by UV detection at
300 nm.
was used for data acquisition. At low CE, the transfer and trap CEs were 2 and
3 eV, respectively. At high CE, the transfer CE was 10 eV, and the trap CE
ranged from 10 to 20 eV. This mode of data collection allowed us to detect
intact precursor ions and fragments.
Data processing was performed using a MetaboLynx subroutine of the
MassLynx software (Waters). Mass defect filtering was used to screen
For metabolite profiling, MS detection was conducted with a Synapt Q-TOF
high-resolution mass spectrometer (Waters) in the positive electrospray ion- metabolites using a 40-mDa filter between the filter and the target metabolites.
ization mode. Nitrogen and argon were used as the desolvation and collision
gases, respectively. The capillary and cone voltages were set at 3,000 and 40 V,
respectively. Data from 80 to 1,000 Da were acquired using a source tem-
perature of 120°C and a desolvation temperature of 350°C. Data were corrected
The fragment ion spectra were compared between the parent compound and the
metabolites to help identify the metabolites, including the structure and site(s)
of modification in the parent molecule.
To determine M23 in HLC incubation, MS detection was conducted on
during acquisition using a reference (LockSpray) sample consisting of 400 a triple TOF 5600+ MS/MS system (AB Sciex, Concord, Ontario, Canada) in
ng/ml leucine encephalin (m/z 556.277) infused at 20 ml/min. An MSE with two
the positive ESI mode. Mass range was set at m/z 100–1,000. The following
separate scan functions programmed with independent collision energies (CEs)
parameter settings were used: ion spray voltage: 5,500 V, declustering

878
Lin et al.
TABLE 3
Pharmacokinetic characteristics of allitinib and its major metabolites following the
oral administration of 1,000 mg of allitinib tosylate three times a day for 21 days in
cancer patients
All values are mean 6 S.D.
Time
Parameter
Allitinib
M6
M10
Day 1
AUC0–24 (h·ng/ml)
Cmax (ng/ml)
tmax (h)
529 6 185
56.8 6 17.1
3.56 6 2.65
6.30 6 3.08
933 6 407
111 6 35
2.83 6 2.15
6.61 6 1.76
1.76 6 0.33
65.7 6 39.6
5.97 6 3.70
5.17 6 2.47
10.0 6 4.2
92.0 6 36.0
6.21 6 2.92
2.50 6 1.00
13.5 6 2.9
1.70 6 0.94
372 6 65
35.8 6 5.0
4.50 6 2.72
7.07 6 3.93
699 6 415
59.9 6 37.7
3.16 6 1.60
7.28 6 0.54
1.78 6 0.91
t1/2 (h)
Day 24
AUC0–24 (h·ng/ml)
Cmax (ng/ml)
tmax (h)
t1/2 (h)
Accumulation ratio
Cmax, maximum concentration observed in day 1 or day 24 after administration of allitinib
tosylate; tmax, time to the maximum concentration.
potential: 80 V, ion source heater: 550°C, curtain gas: 40 psi, ion source gas
1: 55 psi, ion source gas 2: 55 psi. For TOF MS scans, collision energy was
10 eV; for product ion scans, CE was 55 eV, and collision energy spread was 10
in the MS/MS experiment. Information-dependent acquisition (IDA) was used to
trigger acquisition of MS/MS spectra for ions matching the IDA criteria. The real-
time Multiple Mass Defect Filter was used in IDA criteria.
Quantitation of Allitinib and its Metabolites in Plasma with LC-MS/MS.
The concentrations of allitinib, M6, and M10 were quantified simultaneously
using a previously described LC-MS/MS method (Lin et al., 2013).
Quantification of M1 in Urine with LC-MS/MS. The LC system consisted
of two LC-20ADXR pumps and a SIL-20ACXR autosampler (Shimadzu,
Kyoto, Japan). A TSQ Quantum Vantage triple-quadrupole MS (Thermo Fisher
Scientific, Waltham, MA) equipped with a positive electrospray ionization
source was used for MS detection. Selected reaction monitoring was used to
detect the analyte and internal standard. The multiple reaction monitoring
transitions m/z 341.0 → m/z (251.0 + 277.0 + 285.0 + 305.0) and m/z 587.0 →
167.0 were used to detect M1 and irinotecan, respectively.
The concentration of M1 in urine was determined using a validated LC-MS/
MS method. After incubating equal volumes of b-glucuronidase (2,000 units of
type HA-4; Sigma-Aldrich) and sulfatase (15 units of type VIII; Sigma-
Aldrich) in 1 M citrate buffer (pH 5.0) at 37°C for 16 hours, M1 and the
internal standard, irinotecan, were separated on a Gemini C18 column (50 ꢀ 2.0
mm I.D., 5 mm; Phenomenex, Torrance, CA). The mobile phase was 10 mM
ammonium acetate containing 0.1% formic acid mixed with methanol
containing 0.1% formic acid (60:40, v/v).
Quantification of Allitinib and its Metabolites in Feces by HPLC-UV.
An Agilent 1260 HPLC system (Agilent) was used to determine the
concentrations of allitinib and its major metabolites (M2, M5, M6, and M10)
in fecal samples. An ASB C18 column (150 ꢀ 4.6 mm I.D., 5 mm; Agela
Technologies, Wilmington, DE) was used to separate Allitinib, M2, M5, M6,
M10, and NB-2 (internal standard). The mobile phase consisted of a mixture
of 0.1% formic acid in methanol (A) and 0.1% formic acid in 10-mM
ammonium acetate (B). Gradient elution was initiated at 40% A, maintained for
5 minutes, and then increased to 70% A in 2 minutes, which was maintained
for 5 minutes. The elution was then increased to 80% A in 1 minute,
maintained for 5 minutes, and immediately decreased to 40% A to equilibrate
the column. The flow rate was set at 1 ml/min, and a wavelength of 380 nm
was used to monitor the analytes.
Results
MS Fragmentation Behavior of Allitinib. The metabolites of
allitinib were identified by comparing their chromatographic and MS
fragmentation behaviors with those of the parent drug and the
synthesized reference standards. We first examined the chromato-
graphic and MS fragmentation behaviors of the reference allitinib.
Allitinib was eluted at 14.1 minutes and a protonated molecule
[M+H]+ was detected at m/z 449.120 in the positive ion mode. The

Metabolism of Allitinib in Humans
879
TABLE 4
Results of in vitro pharmacological studies of allitinib and its metabolites
All values are mean 6 S.D.
allitinib, M6, and M10 are listed in Table 3. The Cmax values for
allitinib, M6, and M10 were achieved at 3.625.2 hours after the first
dose. The AUC0–24 values for M6 and M10 were 14% and 65% of the
corresponding values of the parent drug (based on molar concen-
trations), respectively. After the last dose, the Cmax values for allitinib,
M6, and M10 were achieved at approximately 2.523.2 hours. At
steady state, the terminal half-lives of allitinib, M6, and M10 were
observed at 4.529.9 hours, at 6.3215.8 hours, and at 4.6211.6 hours,
respectively. The AUC0–24 values for M6 and M10 were 11% and
70% of that for the parent drug, respectively. The accumulation ratios
of allitinib, M6, and M10 were 1.3922.02, 1.0622.78, and 0.742
2.43, respectively.
At steady state, the daily fecal excretion of allitinib and its major
metabolites M2, M5, and M6 accounted for 32.5% 6 17.4% of the
dose. In the fecal samples, unchanged allitinib accounted for 30.4% 6
20.3% of the dose, whereas the major metabolites M2, M5, and M6
accounted for 1.41%, 0.067%, and 0.635% of the dose, respectively.
After a single oral dose of 1,000 mg of allitinib tosylate, a combination
of M1 and its glucuronide conjugate in urine accounted for 0.042% of
the dose.
Analyte
EGFR
ErbB2
IC50 (nM)
Allitinib
M1
M6
0.263 6 0.044
0.773 6 0.023
1.600 6 0.960
0.755 6 0.573
1.10 6 0.40
25.0 6 5.0
282 6 84
M10
72.0 6 13.0
high-CE MS (Fig. 1A) yielded fragment ions at m/z 340.074, 339.069,
313.054 (100% abundance), 311.074, 305.109, 287.071, 277.107, and
213.081. The predominant ion at m/z 313.054 was formed via the loss
of terminal olefin and cleavage of the alkyl C-O bond. The radical ion
at m/z 340.074 formed from the hemolysis of the alkyl C-O bond,
which was followed by the loss of a specific 109 Da moiety (1-fluoro-
3-toluene free radical), while m/z 339.069 was proposed to form from
the cleavage of alkyl C-O bond. The ions at m/z 311.074, 305.109,
287.071, 277.107, and 213.081 formed from the cleavage of N-(4-((3-
chloro-4-hydroxyphenyl)amino)quinazolin-6-yl)acrylamide. According
to the fragmentation pattern, the structure of allitinib was divided into
segments A, B, and C (Fig. 1B). The structures of the metabolites were
characterized by determining the changes in the m/z values of these
three segments.
In Vitro Pharmacological Activity of the Major Metabolites of
Allitinib. The results of in vitro pharmacological activity studies of the
major metabolites are presented in Table 4. The results showed that
M1, M6, and M10 were potent inhibitors of EGFR but weak inhibitor
of ErbB2.
Microsomal Incubation. After incubation with microsomes for 60
minutes, approximately 58% of the parent compound was consumed
by HLM, compared with 10% by HIM. Allitinib was not metabolized
by HPM. Therefore, the liver was considered to be the major site of
allitinib biotransformation.
Five metabolites were generated by incubating allitinib with HLM
in the presence of NADPH. Based on the MS peak areas, the main
metabolite in vitro was M5, followed by M1, M10, and M6. In
samples lacking NADPH, only M6 was detected, which indicated that
the formation of M1, M5, and M10 was NADPH-dependent, and that
P450 enzymes played important roles in the formation of these
metabolites. The formation of M6 was NADPH-independent, and
amidohydrolase was assumed to be responsible for its formation. The
properties of the metabolites identified after allitinib was incubated
with HLM are listed in Table 5.
Identification of P450 Enzymes Involved in the Oxidative
Metabolism of Allitinib. In vitro studies were performed to identify
which P450 enzymes were involved in the metabolism of allitinib and
the formation of M1 and M10. As shown in Fig. 5, after 1-mM allitinib
was incubated with recombinant P450 enzymes in the presence of
NADPH, it was metabolized by 11 of the enzymes tested, but not
CYP2C9. Based on the mean hepatic expression of the P450
enzymes (Rodrigues, 1999), the predominant enzyme involved in
A total of 11, 12, and 10 metabolites of allitinib were detected in
human plasma, urine, and feces, respectively (Fig. 2 and Fig. 3). The
metabolites were designated as M0 (parent compound, allitinib) to
M16 based on their m/z values. According to the UV chromatograms,
the major components in plasma were M0 followed by M6 and M10,
whereas those in urine were M1 and its glucuronide conjugate, M13.
M0 was the predominant compound in feces, whereas M2 and M6
were relatively minor. Table 1 lists the characteristics of the possible
metabolites of allitinib, including their retention times, proposed
elemental compositions, and characteristic fragment ions. The pro-
posed metabolic pathways of allitinib in humans are shown in Fig. 4.
The information used to assign each metabolite is summarized in the
Supplemental Results. The detection of cysteine conjugates and
dihydrodiol metabolites in plasma suggested the formation of reactive
intermediates. The structures of M1, M2, M5, M6, and M10 were
confirmed by comparing their chromatographic retention times and the
mass spectra with those of reference standards. The 1H-nuclear
magnetic resonance data for the reference standards are listed in
Table 2.
Pharmacokinetics and Excretion of Allitinib. The concentrations
of allitinib, M6, and M10 in human plasma were quantified after the
oral administration of 1,000 mg of allitinib tosylate three times a day
for 21 consecutive days. The pharmacokinetic characteristics of
TABLE 5
UHPLC/Q-TOF MS data for allitinib and its metabolites detected in human liver microsomes
RT
(min)
Measured
[M+H]+
+ NADPH +
GSH
No.
Description
Formula
+ NADPH
M0
M1
M5
M6
M10
M14
M22
Allitinib
O-Dealkylation
O-Dealkylation + alkenes to dihydrodiol
Amide hydrolysis
Alkenes to dihydrodiol
Cysteine conjugation
14.1
6.1
3.6
11.6
11.1
10.0
9.7
C24H18ClFN4O2
449.116
341.082
375.084
395.108
483.123
570.138
756.206
+
+
+
+
+
+
+
+
+
+
+
+
C
C
C
C
C
C
17H13ClN4O2
17H15ClN4O4
21H16ClFN4O
24H20ClFN4O4
27H25ClFN5O4S
34H35ClFN7O8S
2
2
Glutathione conjugation
RT, retention time; M14, thiol cysteine conjugate; M22, GSH conjugate.

880
Lin et al.
Fig. 5. Incubation of allitinib (1 mM) in recombinant
human P450 enzymes in the presence of NADPH at 37°C
for 1 hour. (A) Effect of recombinant human P450
enzymes on the consumption of allitinib. (B) Effect of
recombinant human P450 enzymes on the formation of
M1 and M10.
the metabolism of allitinib was CYP3A4, followed by CYP3A5 and involved in the formation of M1, whereas CYP3A4/5 was the
CYP2C8. In the presence of NADPH, M1 was mainly produced by predominant enzyme involved in the production of M10, followed
recombinant CYP3A4 and CYP2C8, followed by CYP2C19, CYP2E1, by CYP1A2.
and CYP3A5, whereas M10 was mainly produced by CYP1A2, fol-
lowed by CYP2D6 and CYP3A5.
Epoxide Hydrolase is Responsible for the Formation of M10. It
has been suggested that M10 is formed by P450 enzymes via an
To further determine the relative contributions of the different epoxide intermediate, and the resulting epoxide is hydrolyzed to M10
P450s in allitinib biotransformation, ABT (nonspecific P450 inhib- by epoxide hydrolase. To evaluate this possibility, HLMs were
itor), a-NF (CYP1A2 inhibitor), sulfaphenazole (CYP2C9 inhibitor), incubated with valpromide, an epoxide hydrolase inhibitor, in the
ticlopidine (CYP2B6/CYP2C19 inhibitor), quercetin (CYP2C8 inhib- presence of NADPH. The coincubation of allitinib with valpromide
itor), quinidine (CYP2D6 inhibitor), clomethiazole (CYP2E1 inhib- resulted in the production of epoxide, which was eluted at 13.2
itor), and KET (CYP3A4/5 inhibitor) were individually added to HLM minutes. The protonated molecule of epoxide was detected at m/z
preparations. The results are presented in Fig. 6. The metabolism of 465.105, and so was 16 Da larger than that of allitinib, indicating that
allitinib was inhibited by all seven P450 inhibitors. The amount of an oxygen atom was introduced into allitinib. The ion at m/z 356.067
allitinib that remained following incubation with ABT was approx- was formed by the specific loss of 109 Da. The product ion at m/z
imately three times greater than that in the control group, suggesting 287.056 was identical to that of the parent drug, whereas the product
that P450 enzymes played important roles in the oxidative metabolism ions at m/z 355.059, 327.064, 321.103, and 293.105 were 16 Da
of allitinib. The formation of M1 and M10 was markedly reduced in greater than those of the parent drug (m/z 339.069, 311.072, 305.109,
the presence of ABT, indicating that P450 enzymes were important and 277.103, respectively). These results suggested that an oxygen
in the production of M1 and M10. The formation of M1 was atom was probably attached to segment C of allitinib to form an
inhibited by 27.8% and 25.5% in the presence of quinidine and KET, epoxide. The identity of allitinib epoxide was further confirmed using
respectively. The formation of M10 was inhibited by 18.6% and synthetic standards.
44.0% in the presence of a-NF and KET, respectively. These results
As shown in Fig. 7, the formation of M10 decreased and the
indicated that CYP2D6 and CYP3A4/5 were the major enzymes production of allitinib epoxide increased with increasing concentrations

Metabolism of Allitinib in Humans
881
Fig. 7. Formation of epoxide and M10 after allitinib was incubated with HLM in the
presence 10, 100, or 1,000 mM of valpromide, an epoxide hydrolase inhibitor.
was independent of NADPH and P450. The chromatographic peak of
M22 increased when allitinib was incubated with human liver cytosol,
indicating that the formation of M22 was catalyzed by GST. Cysteine
conjugate M14 was also detected when allitinib was incubated with
GSH, HLM, and NADPH, suggesting that the cysteine conjugates of
allitinib in humans might be partly derived from GSH conjugates via
the mercapturate pathway.
In Vitro Metabolism of Allitinib Epoxide by HLC. The
coincubation of allitinib epoxide (2 mM) with GSH in HLC allowed
the formation of M6, M10, and a GSH conjugate, M23, which was
eluted at 9.3 minutes. M23 exhibited [M+H]+ ions at m/z 772.196, and
was 307.083 Da larger than that of allitinib epoxide, suggesting the
addition of a GSH molecule. The loss of pyroglutamic acid moiety
resulted in the formation of a fragment ion at m/z 643.154. The
product ions at m/z 355.062, 327.072, and 321.101 were identical to
those of allitinib epoxide. These results suggested that the GSH
molecule most likely be attached to the epoxy part of allitinib epoxide.
After incubation of 20-nM allitinib epoxide with GSH in HLC, M6
and M10, but not M10, were detected (Fig. 8).
Caco-2 Cell Permeability Study. The permeability of Caco-2 cells
to allitinib was determined at a concentration of 2 mM. The Papp(A2B)
of allitinib was 0.77 ꢀ 1026 cm/s, indicating that the cells showed poor
permeability to allitinib. When the transport of allitinib in the apical-
to-basolateral direction was compared with that in the basolateral-to-
apical direction, no apparent efflux transport was observed based on an
ER of 1.40.
Fig. 6. Effects of incubating of allitinib (1 mM) in HLM in the presence of NADPH.
Different enzyme inhibitors, including ABT (a nonspecific CYP inhibitor), a-NF
(a CYP1A1/2 inhibitor), quercetin (a CYP2C8 inhibitor), sulfaphenazole (a CYP2C9
inhibitor), ticlopidine (a CYP2C19 inhibitor), quinidine (a CYP2D6 inhibitor), KET
(a CYP3A4/5 inhibitor), and clomethiazole (a CYP2E1 inhibitor), were used to
evaluate the relative contributions of different microsomal enzyme systems to the
metabolism of allitinib. (A) Effects of P450 inhibitors on the metabolism of allitinib.
(B) Effects of P450 inhibitors on the formation of M1 and M10.
of valpromide. These results confirmed that epoxide hydrolase dis-
played an important role in the formation of M10.
Mechanism Involved in Thiol Conjugate Formation. Cysteine
conjugates (M9 and M14) and N-acetyl cysteine conjugates (M12 and
M16) were detected in humans, indicating the formation of reactive
intermediates. To determine the mechanism involved in the formation
Discussion
This study characterized the metabolic profiles of allitinib in cancer
of these thiol conjugates, in vitro studies were performed using GSH patients and established its biotransformation mechanisms.
as a trapping agent. The coincubation of allitinib with GSH resulted in
Allitinib was quickly absorbed, with a time to reach maximum
the production of a GSH conjugate, M22, which was determined as concentration of approximately 3.0 hours, consistent with the results
a major metabolite in rats after intravenous injection. It was eluted at of the preclinical study. Significant interpatient variability was ob-
9.6 minutes. The protonated molecule of M22 was detected at m/z served after the oral administration of allitinib tosylate, indicating that
756.206, and was 307.090 Da larger than that of allitinib, indicating dose modifications may be necessary to meet individual patients’
the addition of a GSH molecule. Product ions were observed at m/z needs. The unmodified drug was the predominant substance detected
627.163, 449.096, 341.092, and 287.073. The ion at m/z 627.163 in plasma. The amide hydrolysis metabolite (M6) and 27,28-dihydrodiol
resulted from the loss of the pyroglutamic acid moiety (129 Da). The allitinib (M10) were the major pharmacologically active metabolites,
product ions at m/z 341.092 and 287.073 were identical to those of the accounting for approximately 11% and 70% of the estimated AUC0–24
parent drug. These results suggested that the GSH molecule probably for the parent drug at steady state, respectively. The systemic clearance
attached to segment C of allitinib by Michael addition.
of M6 and M10 was slower than that of allitinib.
M22 was also observed when allitinib was incubated without
A total of 10 and 12 metabolites were identified in feces and urine,
NADPH or HLM, implying that the formation of the GSH conjugate respectively. Unchanged allitinib was detected in trace amounts in

882
Lin et al.
Fig. 8. Incubation of 2 mM (upper trace) and 20 nM (lower trace) allitinib epoxide with an HLC preparation in the presence of GSH. Extracted ion chromatograms of GSH
conjugate of allitinib epoxide M23 (A) and (B) and dihydrodiol metabolite M10 (C) and (D).
urine, suggesting that the systemically available allitinib was determined in human fecal samples was generated from the un-
metabolized extensively. The proposed structures of the metabolites absorbed parent drug due to poor bioavailability rather than through
M1, M2, M5, M6, and M10 were supported by comparisons with biliary excretion after absorption.
synthetic standards. The metabolic scheme of allitinib in humans is
The primary routes of allitinib biotransformation involved
presented in Fig. 4. The major metabolites in urine were O-deal- O-dealkylation (M1, M2, M3, M4, M5, M8, M9, M12, and M13),
kylallitinib (M1) and its glucuronide conjugate (M13). After the urine amide hydrolysis (M6), dihydrodiol formation (M5 and M10), and the
was incubated with b-glucuronidase and sulfatase for 16 hours, the subsequent phase 2 conjugation of these metabolites. The identified
recovery of allitinib in terms of M1 was less than 0.1%, indicating that metabolic processes of lapatinib and allitinib are shown in Fig. 9.
renal elimination makes a negligible contribution to the excretion of Allitinib shared several metabolic pathways with its analog lapatinib
allitinib. After the oral administration of allitinib tosylate, the mean (Castellino et al., 2012), including the O-dealkylation and hydroxyl-
fecal excretion ratio was 32.5% 6 17.4% (range from 0.95% to ation of the quinazoline moiety. However, the major metabolic sites of
59.4%). About 93.5% of the total fecal excretion was attributed to the allitinib are located at the a,b-unsaturated carbonyl group, and include
unchanged drug. The major metabolite, M2, accounted for 4.4% of the amide hydrolysis and dihydrodiol formation. The formation of
total fecal extraction, and the other two metabolites (M5 and M6) cysteine conjugate in vivo indicated that allitinib might covalently
accounted for ,3.0%. The recovery of allitinib in fecal samples bind to proteins or biologic macromolecules. The protein covalent
seemed to be related to the fecal weight excreted by the subjects. binding ratios after incubation of allitinib with human plasma and
Several factors may contribute to the low recovery of allitinib in HLM at 37°C for 1 hour were both less than 10.1%, and protein
humans after its oral administration, including the difficulty in fully covalent binding of allitinib was time-dependent (data not shown).
sampling the nonhomogeneously processed fecal samples, the loss of Further investigation is needed to understand the impact of protein
samples during collection or processing, or the covalent binding or covalent binding on the potency of this drug. In vitro studies revealed
noncovalent tissue uptake of the drug or its metabolites (Roffey et al., that the reaction between allitinib and GSH was independent of
2007). However, the results suggested that allitinib is predominantly microsomal P450 enzymes and NADPH, but that GST could catalyze
eliminated via fecal excretion. To confirm this possibility, radiolabeled the conjugation of allitinib and GSH.
allitinib will be used in future mass balance experiments.
In vitro metabolism studies using recombinant human P450
To further investigate the source of unchanged drug in feces, in isozymes and inhibition studies using selective chemical inhibitors
vitro Caco-2 permeability study was performed. The results showed of P450 enzymes suggested that the oxidative metabolism of allitinib
that the permeability of these cells to allitinib was poor (Papp , 2.0 ꢀ is mediated by multiple P450 enzymes. Allitinib was primarily me-
1026 cm/s). Thus, it might be concluded that the parent drug tabolized by CYP3A4/5, followed by CYP1A2, CYP1B1, CYP2A6,

Metabolism of Allitinib in Humans
883
Fig. 9. Identified metabolic processes of lapatinib (A) and allitinib (B) in humans.
CYP2B6, CYP2C8, CYP2C19, CYP2D6, CYP2E1, and CYP4A11. epoxide intermediate was more readily metabolized to dihydrodiol
The formation of M1 was primarily mediated by CYP3A4/5 and than to the GSH conjugate.
CYP2D6, whereas CYP1A2 and CYP3A4/5 were mainly involved in
An in vitro investigation showed that lapatinib was metabolized to
the formation of M10. The incubation of allitinib with the CYP1A2 O-dealkylated lapatinib by CYP3A4 and CYP2C8, and GSH and
inhibitor a-NF in an HLM preparation markedly increased the cysteinyl-glycine conjugates were also observed (Teng et al., 2010).
production of M1, which was probably attributable to the stimulation Although allitinib was metabolized by O-dealkylation, like lapatinib,
of the CYP3A system by a-NF (Ueng et al., 1997; Maenpaa et al., we detected no conjugates formed by the reaction between GSH and
1998). Because multiple P450 enzymes were involved in the the quinone imine intermediate. This might be explained by the greater
metabolism of allitinib, in vivo exposure to allitinib is unlikely to be reactivity of the a,b-unsaturated carbonyl group of allitinib. Con-
markedly affected by P450 enzyme inhibitors. Therefore, the likeli- sequently, GSH reacted with the a,b-unsaturated carbonyl group more
hood of drug-drug interactions mediated by the metabolism of allitinib readily than with the quinone imine group.
is minimal.
Lapatinib is suggested to be a quasi-irreversible inhibitor of
The incubation of allitinib with HLM and valpromide, an inhibitor CYP3A4, which is mediated by metabolites derived from N-oxidation
of epoxide hydrolase, confirmed that allitinib is metabolized to (Takakusa et al., 2011). Another in vitro study showed that lapatinib is
epoxide by P450 enzymes, and the epoxide metabolite is further also a mechanism-based inactivator of CYP3A5, and this inactivation
metabolized to M10 by epoxide hydrolase. Epoxides can react with is mediated by quinone imine (Chan et al., 2012). These studies
cellular macromolecules, such as DNA and protein, to exert mutagenic provide good insight into the possible mechanism of hepatotoxicity of
and genotoxic effects. Alternatively, epoxides can be metabolized to lapatinib observed in clinical trials. To determine whether allitinib
GSH conjugates by GST or to dihydrodiol metabolites by epoxide inhibits P450s because its structure is analogous to that of lapatinib
hydrolase with reduced toxicity (Ehrenberg and Hussain, 1981; and its introduced acrylamide group, in vitro enzyme induction and
Hooberman et al., 1993; Faller et al., 2001; Lee et al., 2005; Gonzalez- inhibition studies were performed. The results demonstrated that
Perez et al., 2012). The incubation of allitinib with HLM allitinib did not inhibit CYP1A2, CYP2C9, CYP2C19, or CYP2D6,
supplemented with NADPH and GSH confirmed that the pharmaco- but was a weak inhibitor of CYP3A4 (IC50 ; 100 mM with
logically active metabolite M10 was the major metabolite, but the midazolam as the substrate). This level of inhibition is negligible at
GSH conjugate of epoxide M23 was not observed. M10 and M23 clinical doses. Furthermore, we observed no apparent enzyme
were both derived from allitinib epoxide. Our studies showed that induction. These studies suggest that, even though the a,b-unsaturated
allitinib epoxide could form a GSH conjugate at the concentration of carbonyl group was introduced and a quinone imide intermediate
2 mM. However, when a much lower substrate concentration (20 nM) might be formed via O-dealkylation, allitinib does not affect P450
was used, the GSH conjugate of epoxide could not be detected, enzyme activity. The probability that allitinib has a pharmacologic
whereas the formation of M10 was observed, indicating that the effect on coadministered drugs via the induction or inhibition of P450

884
Lin et al.
reveals an active conformation, poised to interact with other ErbB receptors. Mol Cell 11:
495–505.
Geyer CE, Forster J, Lindquist D, Chan S, Romieu CG, Pienkowski T, Jagiello-Gruszfeld A,
Crown J, Chan A, and Kaufman B, et al. (2006) Lapatinib plus capecitabine for HER2-positive
advanced breast cancer. N Engl J Med 355:2733–2743.
Gonzales AJ, Hook KE, Althaus IW, Ellis PA, Trachet E, Delaney AM, Harvey PJ, Ellis TA,
Amato DM, and Nelson JM, et al. (2008) Antitumor activity and pharmacokinetic properties of
PF-00299804, a second-generation irreversible pan-erbB receptor tyrosine kinase inhibitor. Mol
Cancer Ther 7:1880–1889.
González-Pérez M, Gómez-Bombarelli R, Arenas-Valgañón J, Pérez-Prior MT, García-Santos
MP, Calle E, and Casado J (2012) Connecting the chemical and biological reactivity of
epoxides. Chem Res Toxicol 25:2755–2762.
Hooberman BH, Chakraborty PK, and Sinsheimer JE (1993) Quantitative structure-activity
relationships for the mutagenicity of propylene oxides with Salmonella. Mutat Res 299:85–93.
Klapper LN, Kirschbaum MH, Sela M, and Yarden Y (2000) Biochemical and clinical impli-
cations of the ErbB/HER signaling network of growth factor receptors. Adv Cancer Res 77:
25–79.
is low, even though the recommended dose of allitinib is 1,000 mg
three times a day.
In conclusion, we found that O-dealkylation is an important
metabolic pathway of allitinib, as it is for lapatinib. However, we also
found that amide hydrolysis and dihydrodiol formation are major
metabolic pathways, because the reactivity of the a,b-unsaturated
carbonyl group is high. Multiple enzymes, including P450 and
epoxide hydrolase, are involved in the metabolism of allitinib, but
allitinib does not induce or inhibit tested P450 enzymes, suggesting
a low potential for drug-drug interaction mediated by tested P450
enzymes in cases of coadministration.
Lee MS, Faller TH, Kreuzer PE, Kessler W, Csanády GA, Pütz C, Ríos-Blanco MN, Pottenger
LH, Segerbäck D, and Osterman-Golkar S, et al. (2005) Propylene oxide in blood and soluble
nonprotein thiols in nasal mucosa and other tissues of male Fischer 344/N rats exposed to
propylene oxide vapors—relevance of glutathione depletion for propylene oxide-induced rat
nasal tumors. Toxicol Sci 83:177–189.
Lin L, Gao Z, Chen X, and Zhong D (2013) Development and validation of a sensitive LC-MS/
MS assay for the simultaneous quantification of allitinib and its two metabolites in human
plasma. J Pharm Biomed Anal 86:49–55.
Mäenpää J, Hall SD, Ring BJ, Strom SC, and Wrighton SA (1998) Human cytochrome P450 3A
(CYP3A) mediated midazolam metabolism: the effect of assay conditions and regioselective
stimulation by alpha-naphthoflavone, terfenadine and testosterone. Pharmacogenetics 8:
137–155.
Ackowledgments
The authors thank the staff at Allist Pharmaceuticals, Inc. (Shanghai, China)
for synthesizing the standard compounds; Fudan University Shanghai Cancer
Center (Shanghai, China) for conducting the clinical studies; Dr. Hua Xie of
Shanghai Institute of Materia Medica (SIMM) for assistance with in vitro
pharmacological assays of the major metabolites of allitinib; Dr. Liang Li
(SIMM) for help for the NMR analysis; and Yanjun Bai and Xiuli Li (SIMM)
for assistance with the permeability studies.
Olayioye MA, Neve RM, Lane HA, and Hynes NE (2000) The ErbB signaling network: receptor
heterodimerization in development and cancer. EMBO J 19:3159–3167.
Peroukides S, Makatsoris T, Koutras A, Tsamandas A, Onyenadum A, Labropoulou-Karatza C,
and Kalofonos H (2011) Lapatinib-induced hepatitis: a case report. World J Gastroenterol 17:
2349–2352.
Rabindran SK, Discafani CM, Rosfjord EC, Baxter M, Floyd MB, Golas J, Hallett WA, Johnson
BD, Nilakantan R, and Overbeek E, et al. (2004) Antitumor activity of HKI-272, an orally
active, irreversible inhibitor of the HER-2 tyrosine kinase. Cancer Res 64:3958–3965.
Rodrigues AD (1999) Integrated cytochrome P450 reaction phenotyping: attempting to bridge the
gap between cDNA-expressed cytochromes P450 and native human liver microsomes. Bio-
chem Pharmacol 57:465–480.
Authorship Contributions
Participated in research design: Lin, Zhong, Chen.
Conducted experiments: Lin, Xie, Gao.
Contributed new reagents or analytic tools: Lin, Zhong, Chen.
Performed data analysis: Lin, Zhong, Xie, Gao, Chen.
Contributed to the writing of the manuscript: Lin, Zhong.
Roffey SJ, Obach RS, Gedge JI, and Smith DA (2007) What is the objective of the mass balance
study? A retrospective analysis of data in animal and human excretion studies employing
radiolabeled drugs. Drug Metab Rev 39:17–43.
Shewchuk L, Hassell A, Wisely B, Rocque W, Holmes W, Veal J, and Kuyper LF (2000) Binding
mode of the 4-anilinoquinazoline class of protein kinase inhibitor: X-ray crystallographic
studies of 4-anilinoquinazolines bound to cyclin-dependent kinase 2 and p38 kinase. J Med
Chem 43:133–138.
Takakusa H, Wahlin MD, Zhao C, Hanson KL, New LS, Chan EC, and Nelson SD (2011)
Metabolic intermediate complex formation of human cytochrome P450 3A4 by lapatinib. Drug
Metab Dispos 39:1022–1030.
Takezawa K, Okamoto I, Tanizaki J, Kuwata K, Yamaguchi H, Fukuoka M, Nishio K,
and Nakagawa K (2010) Enhanced anticancer effect of the combination of BIBW2992 and
thymidylate synthase-targeted agents in non-small cell lung cancer with the T790M mutation of
epidermal growth factor receptor. Mol Cancer Ther 9:1647–1656.
References
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.
Bence AK, Anderson EB, Halepota MA, Doukas MA, DeSimone PA, Davis GA, Smith DA,
Koch KM, Stead AG, and Mangum S, et al. (2005) Phase I pharmacokinetic studies evaluating
single and multiple doses of oral GW572016, a dual EGFR-ErbB2 inhibitor, in healthy sub-
jects. Invest New Drugs 23:39–49.
Burris 3rd HA, Hurwitz HI, Dees EC, Dowlati A, Blackwell KL, O’Neil B, Marcom PK, Ellis MJ,
Overmoyer B, and Jones SF, et al. (2005) Phase I safety, pharmacokinetics, and clinical activity
study of lapatinib (GW572016), a reversible dual inhibitor of epidermal growth factor receptor
tyrosine kinases, in heavily pretreated patients with metastatic carcinomas. J Clin Oncol 23:
5305–5313.
Castellino S, O’Mara M, Koch K, Borts DJ, Bowers GD, and MacLauchlin C (2012) Human
metabolism of lapatinib, a dual kinase inhibitor: implications for hepatotoxicity. Drug Metab
Dispos 40:139–150.
Chan EC, New LS, Chua TB, Yap CW, Ho HK, and Nelson SD (2012) Interaction of lapatinib
with cytochrome P450 3A5. Drug Metab Dispos 40:1414–1422.
Teng WC, Oh JW, New LS, Wahlin MD, Nelson SD, Ho HK, and Chan EC (2010) Mechanism-
based inactivation of cytochrome P450 3A4 by lapatinib. Mol Pharmacol 78:693–703.
Ueng YF, Kuwabara T, Chun YJ, and Guengerich FP (1997) Cooperativity in oxidations cata-
lyzed by cytochrome P450 3A4. Biochemistry 36:370–381.
Xie H, Lin L, Tong L, Jiang Y, Zheng M, Chen Z, Jiang X, Zhang X, Ren X, and Qu W, et al.
(2011) AST1306, a novel irreversible inhibitor of the epidermal growth factor receptor 1 and 2,
exhibits antitumor activity both in vitro and in vivo. PLoS ONE 6:e21487.
Yap TA, Vidal L, Adam J, Stephens P, Spicer J, Shaw H, Ang J, Temple G, Bell S, and Shahidi
M, et al. (2010) Phase I trial of the irreversible EGFR and HER2 kinase inhibitor BIBW 2992
in patients with advanced solid tumors. J Clin Oncol 28:3965–3972.
Ehrenberg L and Hussain S (1981) Genetic toxicity of some important epoxides. Mutat Res 86:
1–113.
Eskens FA, Mom CH, Planting AS, Gietema JA, Amelsberg A, Huisman H, van Doorn L, Burger
H, Stopfer P, and Verweij J, et al. (2008) A phase I dose escalation study of BIBW 2992, an
irreversible dual inhibitor of epidermal growth factor receptor 1 (EGFR) and 2 (HER2) tyrosine
kinase in a 2-week on, 2-week off schedule in patients with advanced solid tumours. Br J
Cancer 98:80–85.
Faller TH, Csanády GA, Kreuzer PE, Baur CM, and Filser JG (2001) Kinetics of propylene oxide
metabolism in microsomes and cytosol of different organs from mouse, rat, and humans.
Toxicol Appl Pharmacol 172:62–74.
Address correspondence to: Dr. Dafang Zhong, Shanghai Institute of Materia
Medica, Chinese Academy of Sciences, 501 Haike Road, Shanghai 201203, P.R.
China. E-mail: dfzhong@simm.ac.cn
Garrett TP, McKern NM, Lou M, Elleman TC, Adams TE, Lovrecz GO, Kofler M, Jorissen RN,
Nice EC, and Burgess AW, et al. (2003) The crystal structure of a truncated ErbB2 ectodomain