CYP3A-mediated generation of aldehyde and hydrazine in atazanavir metabolism

Article abstract of DOI:10.1124/dmd.110.036327

Atazanavir (ATV) is an antiretroviral drug of the protease inhibitor class. Multiple adverse effects of ATV have been reported in clinical practice, such as jaundice, nausea, abdominal pain, and headache. The exact mechanisms of ATV-related adverse effect

Full text of DOI:10.1124/dmd.110.036327

Supplemental material to this article can be found at:
http://dmd.aspetjournals.org/content/suppl/2010/12/09/dmd.110.036327.DC1.html
0090-9556/11/3903-394–401$20.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2011 by The American Society for Pharmacology and Experimental Therapeutics
DMD 39:394–401, 2011
Vol. 39, No. 3
36327/3668841
Printed in U.S.A.
CYP3A-Mediated Generation of Aldehyde and Hydrazine
□
S
in Atazanavir Metabolism
Feng Li, Jie Lu, Laiyou Wang,1 and Xiaochao Ma
Department of Pharmacology, Toxicology and Therapeutics, University of Kansas Medical Center, Kansas City, Kansas
Received September 9, 2010; accepted December 9, 2010
ABSTRACT:
Atazanavir (ATV) is an antiretroviral drug of the protease inhibitor known and 13 novel ATV metabolites. Three potential reactive
class. Multiple adverse effects of ATV have been reported in clin- metabolites were detected and characterized for the first time: one
ical practice, such as jaundice, nausea, abdominal pain, and head- aromatic aldehyde, one ␣-hydroxyaldehyde, and one hydrazine.
ache. The exact mechanisms of ATV-related adverse effects are These potential reactive metabolites were primarily generated by
unknown. It is generally accepted that a predominant pathway of CYP3A. Our results provide a clue for studies on ATV-related ad-
drug-induced toxicity is through the generation of reactive metab- verse effects from the aspect of metabolic activation. Further stud-
olites. Our current study was designed to explore reactive metab-
olites of ATV. We used a metabolomic approach to profile ATV
metabolism in mice and human liver microsomes. We identified 5
ies are suggested to illustrate the impact of these potential reac-
tive metabolites on ATV-related adverse effects.
Introduction
which was reported in ϳ40% of patients who received 400 mg of
ATV once daily. ATV-induced hyperbilirubinemia rarely led to dis-
continuation of treatment; however, ϳ8% of patients developed
clinical jaundice (Goldsmith and Perry, 2003; Sulkowski, 2004). In
addition, elevation of alanine aminotransferase and aspartate ami-
notransferase activity was noted in ϳ14% of the patients receiving
ATV and was unrelated to bilirubin levels (Goldsmith and Perry,
2003). Monitoring of liver function is recommended for ATV-treated
patients, in particular for the patients with existing liver diseases
(Eholie´ et al., 2004). The exact mechanisms of ATV-related adverse
effects are unknown.
It is generally accepted that a predominant pathway of drug-in-
duced toxicity is via the generation of reactive metabolites (Baillie,
2006; Guengerich and MacDonald, 2007). The reactive metabolites,
such as aldehyde, epoxide, quinone methide, and hydroxylamine, can
cause various adverse side effects (O’Brien et al., 2005; Tang and Lu,
2010). For instance, felbamate, a broad-spectrum antiepileptic agent,
resulted in hepatotoxicity by way of its metabolite atropaldehyde
(Dieckhaus et al., 2002). Until now, limited information has been
available on ATV metabolism. In 2009, five ATV metabolites were
reported, which included one N-dealkylation product, two metabolites
resulting from carbamate hydrolysis, one hydroxylated product, and
one keto metabolite (ter Heine et al., 2009). Reactive metabolites of
ATV have not been identified.
Atazanavir (Reyataz) (ATV), is a protease inhibitor (PI) used for
the treatment of the human immunodeficiency virus (HIV) infection
(Swainston Harrison and Scott, 2005; Croom et al., 2009). ATV was
approved by the U.S. Food and Drug Administration in 2003 and is
used in combination with other antiretroviral agents, such as ritonavir
(RTV). The Food and Drug Administration also approved the use of
ATV without cotreatment with RTV in selective PI-naive patients.
The recommended dose for treatment-naive patients is 300 mg of
ATV with 100 mg of RTV once daily. ATV can be given in a dose of
400 mg once daily without RTV for selected PI-naive patients who
cannot tolerate RTV (Rivas et al., 2009). ATV has particular advan-
tages over other PIs because of its moderate resistance profile, min-
imal effect on lipid profiles, low capsule burden, and once-daily
dosing (Rivas et al., 2009).
Despite these advantages, ATV is associated with various adverse
drug reactions (Busti et al., 2004; Havlir and O’Marro, 2004). Nausea
was reported in ϳ35% of the patients receiving ATV, followed by
abdominal pain, headache, and diarrhea (Goldsmith and Perry, 2003).
The most common laboratory abnormality is hyperbilirubinemia,
This work was supported by the National Institutes of Health National Center
for Research Resources [Grant COBRE 5P20-RR021940].
1 Current affiliation: Department of Pharmacology, Guangdong Pharmaceuti-
cal University, Guangzhou, China.
In the current study, we used a metabolomic approach, which has
been proved to be a powerful tool in studying drug metabolism (Chen
et al., 2006; Li et al., 2010), to investigate the metabolism of ATV in
mice and human liver microsomes (HLM). We identified 5 known and
13 novel ATV metabolites. Three potential reactive metabolites were
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.110.036327.
□S The online version of this article (available at http://dmd.aspetjournals.org)
contains supplemental material.
ABBREVIATIONS: ATV, atazanavir; PI, protease inhibitor; HIV, human immunodeficiency virus; RTV, ritonavir; HLM, human liver microsomes;
P450, cytochrome P450; INH, isoniazid; KCZ, ketoconazole; rcf, relative centrifugal forces; UPLC, ultraperformance liquid chromatography;
TOFMS, time-of-flight mass spectrometry; PBS, phosphate-buffered saline; PCA, principal component analysis; OPLS-DA, orthogonal projection
to latent structures-discriminant analysis; MS/MS, tandem mass spectrometry; TB, tuberculosis.
394

REACTIVE METABOLITES OF ATAZANAVIR
395
matography on silica gel (hexane-ethyl acetate 6:1) to yield the oxime ether
detected and characterized: one aromatic aldehyde, one ␣-hydroxyal-
dehyde, and one hydrazine. CYP3A4 was identified as the primary
enzyme involved in the formation of the two aldehydes and one
hydrazine. These results provide a clue for studies on ATV-related
adverse effects from the aspect of metabolic activation.
(338 mg, 80%) as a colorless oil. 1H NMR (400 MHz, CDCl3): ␦ 8.70 (dd, 3J ϭ
4
8.1 Hz, J ϭ 1.5 Hz, 1H, pyridinyl-H), 8.10 (s, 1H, CH ϭ NOCH3), 8.01 (d,
3J ϭ 8.0 Hz, 2H, phenyl-H), 7.72–7.80 (2H, pyridinyl-H), 7.67 (d, 3J ϭ 8.0 Hz,
2H, phenyl-H), 7.25 (m, 1H, pyridinyl-H), 3.99 (s, 1H, CH ϭ NOCH3); 13C
NMR (100 MHz, CDCl3): ␦ 156.5 (CH ϭ NOCH3), 149.7 (pyridinyl-C), 148.2
(pyridinyl-C), 140.5 (phenyl-C), 136.8 (pyridinyl-C), 132.6 (phenyl-C), 127.7,
(phenyl-C), 127.1 (phenyl-C), 122.4 (pyridinyl-C), 120.5 (pyridinyl-C), 62.1
(CH ϭ NOCH3); high-resolution mass spectrometry (electrospray ionization,
positive): m/z [M ϩ H]ϩ calculated for C13H13N2O: 213.1028; found:
213.1024. 1H and 13C NMR spectra were recorded on a 400 MHz Varian
spectrometer. Chemical shifts are reported in parts per million, and coupling
constants, J, are reported in Hertz (Supplemental Fig. 1).
Materials and Methods
Chemicals and Reagents. ATV (methyl N-[(2S)-1-[2-[(2S,3S)-2-hydroxy-
3-[[(2S)-2-(methoxycarbonylamino)-3,3-dimethylbutanoyl]amino]-4-phenyl-
butyl]-2-[(4-pyridin-2-ylphenyl) methyl]hydrazinyl]-3,3-dimethyl-1-
oxobutan-2-yl]carbamate) was supplied by the National Institutes of Health
AIDS Research and Reference Reagent Program. The recombinant human
P450s and HLM were purchased from XenoTech, LLC (Lenexa, KS). Isoni-
azid (INH), methoxylamine, ketoconazole (KCZ), semicarbazide, and NADPH
were obtained from Sigma-Aldrich (St. Louis, MO). 4-(Pyridin-2-yl)-
benzaldehyde and 4-(pyridin-2-yl)benzoic acid were purchased from Syn-
Chem, Inc. (Des Plaines, IL). All the solvents for liquid chromatography and
mass spectrometry were of the highest grade commercially available.
Animals and Treatments. All mice (2–4 months old) were maintained
under a standard 12-h dark/light cycle with water and chow provided ad
libitum. Handling was in accordance with study protocols approved by the
University of Kansas Medical Center Institutional Animal Care and Use
Committee. The mice were treated (orally) with ATV (50 mg/kg) or 4-(pyri-
din-2-yl)-benzaldehyde (12 mg/kg) and housed separately in metabolic cages
for 18 h. Urine and feces were collected for metabolite analysis. In brief,
urinary samples were prepared by mixing 40 ␮l of urine with 160 ␮l of 50%
acetonitrile and were centrifuged at 20,000 relative centrifugal forces (rcf) for
10 min. Feces were homogenized in water (1 mg of feces in 10 ␮l of H2O).
Then 200 ␮l of acetonitrile was added to 200 ␮l of the resulting mixture,
followed by centrifugation at 20,000 rcf for 10 min. The supernatant was
transferred to a new Eppendorf vial for a second centrifugation (20,000 rcf for
10 min). Each supernatant was transferred to an autosampler vial, and 5.0 ␮l
was injected into a system (Waters, Milford, MA) combining ultraperformance
liquid chromatography (UPLC) and time-of-flight mass spectrometry (TOFMS)
for metabolite analysis.
ATV Metabolism In Vitro. Incubations were conducted in 1ϫ phosphate-
buffered saline (PBS) (pH 7.4) containing 50 ␮M ATV, 0.1 mg of HLM, or 2
pmol of each cDNA-expressed P450 enzyme (control, CYP1A2, 2A6, 2B6,
2C8, 2C9, 2C19, 2D6, 2E1, and CYP3A4) in a final volume of 190 ␮l. After
5 min of preincubation at 37°C, the reaction was initiated by the addition of 10
␮l of 20 mM NADPH (final concentration 1.0 mM) and continued for 30 min
with gentle shaking. Incubations in the absence of NADPH were used as
controls. Coincubations of KCZ (10 ␮M) were performed to determine the role
of CYP3A in ATV metabolism. Incubations were terminated by adding 200 ␮l
of acetonitrile and vortexing for 1 min and centrifuging at 20,000 rcf for 10
min. Each supernatant was transferred to an autosampler vial, and 5.0 ␮l was
injected into the UPLC-TOFMS system for metabolite analysis. All incuba-
tions were performed in duplicate.
Biomarkers of Metabolic Activation. Because most reactive metabolites
are not stable, it is difficult to detect them directly. Reactive intermediates can
form adducts with trapping agents, such as GSH, potassium cyanide, methoxy-
lamine, and semicarbazide, which predict potential binding with cellular pro-
teins and/or some other molecules. For example, methoxylamine can form a
Schiff base with aldehydes, a process mimicking reactions between aldehyde
metabolites and lysine residues on proteins (Evans et al., 2004). UPLC-
TOFMS can be used to detect adducts of reactive metabolites and trapping
agents. In the current study, methoxylamine, semicarbazide, and INH were
used as trapping agents.
Synthesis and Characterization of Methyl Oxime of 4-(Pyridin-2-yl)
benzaldehyde. Methoxylamine hydrochloride (172 mg, 2.1 mmol) and pyri-
dine (167 mg, 2.1 mmol) were added to a solution of 4-(pyridin-2-yl)benzal-
dehyde (366 mg, 2.0 mmol) in methanol (10 ml). The reaction mixture was
refluxed for 30 min in a water bath. After most of the methanol had been
removed in vacuo, water (5.0 ml) was added to the residue, and the mixture
was extracted with CH2Cl2 (two 15-ml extractions). The organic phases were
washed with H2O (two 10-ml washes) and dried with MgSO4. After removal
Trapping 4-(Pyridin-2-yl)benzaldehyde Using Methoxylamine. One ar-
omatic aldehyde, namely, 4-(pyridin-2-yl)benzaldehyde, was detected in our
study using methoxylamine as a trapping agent. The experiment was per-
formed in 1ϫ PBS containing 50 ␮M ATV, 0.1 mg of HLM, and 20.0 ␮l of
50 mM methoxylamine (dissolved in 1ϫ PBS, final concentration 5 mM) in a
final volume of 190 ␮l. After 5 min of preincubation at 37°C, the reaction was
initiated by the addition of 10 ␮l of 20 mM NADPH (final concentration 1.0
mM) and continued for 30 min with gentle shaking. The same incubations were
conducted without NADPH as the control. The workup is identical to the
procedure described under ATV Metabolism In Vitro.
Trapping Methyl (S)-1-((2S,3R)-3-hydroxy-4-oxo-1-phenylbutan-2-
ylamino)-3,3-dimethyl-1-oxobutan-2-ylcarbamate Using INH. One ␣-hy-
droxyaldehyde, namely, methyl (S)-1-((2S,3R)-3-hydroxy-4-oxo-1-phenylbu-
tan-2-ylamino)-3,3-dimethyl-1-oxobutan-2-ylcarbamate, was detected in our
study. INH was used to trap this ␣-hydroxyaldehyde. The experiment was
performed in 1ϫ PBS containing 50 ␮M ATV, 0.1 mg of HLM, and 20.0 ␮l
of 50 mM INH (dissolved in H2O, final concentration 5 mM) in a final volume
of 190 ␮l. After 5 min of preincubation at 37°C, the reaction was initiated by
the addition of 10 ␮l of 20 mM NADPH (final concentration 1.0 mM) and
continued for 1 h with gentle shaking. The same incubations without NADPH
were conducted as the control. The reactions were quenched by adding 200 ␮l of
ice-cold acetonitrile. The mixture was vortexed for 1 min and centrifuged at 20,000
rcf for 10 min. The supernatant was transferred to an autosampler vial, and 5.0 ␮l
was injected into the UPLC-TOFMS system for metabolite analysis. In addition to
INH, semicarbazide was also used to trap this ␣-hydroxyaldehyde.
UPLC-TOFMS Analysis. ATV and its metabolites were separated using a
100 ϫ 2.1-mm (Acquity 1.7 ␮m) UPLC BEH C-18 column (Waters). The flow
rate of the mobile phase was 0.3 ml/min with a gradient ranging from 2 to 98%
aqueous acetonitrile containing 0.1% formic acid in a 10-min run. TOFMS was
operated in both positive and negative modes with electrospray ionization. The
source temperature and desolvation temperature were set at 120 and 350°C,
respectively. Nitrogen was applied as the cone gas (10 l/h) and desolvation gas
(700 l/h). Argon was applied as the collision gas. TOFMS was calibrated with
sodium formate and monitored by the intermittent injection of lock mass
leucine enkephalin in real time. The capillary voltage and the cone voltage
were set at 3.5 kV and 35 V in positive ion mode. Metabolites were screened
by using MarkerLynx software (Waters) on the basis of accurate mass mea-
surement (mass errors less than 10 ppm). The structures of ATV and its
metabolites were elucidated by tandem mass spectrometry fragmentation with
collision energy ramp ranging from 10 to 40 eV.
Data Analysis. Mass chromatograms and mass spectra were acquired by
MassLynx software in centroid format from m/z 50 to 1000. Centroid and
integrated mass chromatographic data were processed by MarkerLynx soft-
ware to generate a multivariate data matrix. Principal component analysis
(PCA) and orthogonal projection to latent structures-discriminant analysis
(OPLS-DA) were conducted on Pareto-scaled data. The corresponding data
matrices were then exported into SIMCA-Pϩ12 (Umetrics, Kinnelon, NJ) for
multivariate data analysis.
Results
Profile of ATV Metabolism in Mice Using a Metabolomic Ap-
proach. ATV and its metabolites were found in the feces and urine
but mainly in the feces. The results of chemometric analysis on the
of the organic solvent under vacuum, the residue was subject to flash chro- ions produced by UPLC-TOFMS assay of control and ATV-treated

396
LI ET AL.
mouse urine and feces are shown in Fig. 1. The unsupervised PCA tal Fig. 2. The patterns of ATV metabolites in urine and feces are
similar, but most of them are much more abundant in feces (Supple-
mental Figs. 3 and 4). The metabolic map of ATV in mice is sum-
marized in Fig. 2. Overall, 18 ATV metabolites were identified,
including 5 previously reported metabolites (M1, M2, M6, M13, and
M14) (ter Heine et al., 2009) and 13 novel metabolites (Fig. 2).
Among these novel ATV metabolites, 3 potential reactive metabolites
were detected and characterized: one aromatic aldehyde (m1), one
␣-hydroxyaldehyde (M15Ј), and one hydrazine (M16).
analysis score plot of the feces (Fig. 1A) revealed two clusters
corresponding to the control and ATV-treated groups. The S-plots
(Fig. 1, B and C) generated from OPLS-DA display the ion contri-
bution to the group separation in the feces and urine, respectively. The
top ranking ions were identified as ATV and its metabolites, which
were marked in the S-plots (Fig. 1, B and C). The MS/MS spectra of
M2–M14 and their structural elucidations are provided in Supplemen-
Identification of an Aromatic Aldehyde (m1) in ATV Metabo-
lism. A dealkylated ATV metabolite (M1) was reported in a previous
study (ter Heine et al., 2009). In theory, an aldehyde should be
generated together with the formation of M1. We confirmed the
existence of the aldehyde (m1) in ATV metabolism (Figs. 3 and 4). In
the ATV-treated mouse urine samples, the dealkylated ATV metab-
olite (M1) was detected, but the expected aldehyde (m1) was not
found. However, further metabolites of m1 were detected, which
included an acid (m2), a glycine-conjugate product (m3), and an
N-acetylcysteine-conjugate product (m5) (Fig. 3). In addition, the
existence of m1 was confirmed in incubations with HLM using
methoxylamine as a trapping agent. The evidence for m1 formation is
summarized in Fig. 4. The chromatograms of M1, m2, m3, and m5 are
presented in Fig. 4A. M1 was eluted at 5.51 min and had a protonated
molecule [M ϩ H]ϩ at m/z ϭ 538 Da. Compared with the MS/MS of
ATV (Fig. 4B), M1 had the same fragments at m/z 335 and 144. The
fragment ion at m/z 367 and the absence of ion at m/z 168 suggested that
the 4-(pyridin-2-yl)-benzyl moiety was lost from ATV. The other frag-
ment ions are interpreted in the inlaid structural diagram (Fig. 4C).
4-(Pyridin-2-yl)benzoic Acid (m2). m2 is an oxidized metabolite
of m1. In both the ATV-treated and 4-(pyridin-2-yl)benzaldehyde
(m1)-treated mouse urine samples, m2, which had a protonated mol-
ecule [M ϩ H]ϩ at m/z ϭ 200 Da, was detected. The fragment ions
of m2, at m/z 182, 154, 127, and 78, are interpreted in the inlaid
diagram (Fig. 4D). In addition, m2 was confirmed by comparison of
retention time and accurate mass with an authentic standard sample.
2-(4-(Pyridin-2-yl)benzamido)acetic Acid (m3). m3 is a glycine-
conjugated product of m2. Metabolite m3 was eluted at 3.42 min,
having a protonated molecule [M ϩ H]ϩ at m/z ϭ 257 Da. The major
ions at m/z 182, 154, 127, and 78 are interpreted in Fig. 4E. The ion
at m/z 211 was formed by loss of the HCOOH moiety.
2-Acetamido-3-(4-(pyridin-2-yl)benzylthio)propanoic acid (m5).
m5 is a further metabolite of m1, which is conjugated with N-acetyl-
cysteine, a moiety of GSH. Metabolite m5, eluted at 3.76 min, had a
protonated molecule [M ϩ H]ϩ at m/z ϭ 331 Da. MS/MS of m5
produced the major ions at m/z 289, 243, 200, and 168. The structural
elucidation was interpreted in Fig. 4F. Metabolite m5 was observed in
both ATV and 4-(pyridin-2-yl)benzaldehyde (m1)-treated mouse
urine samples.
4-(Pyridin-2-yl)benzaldehyde (m1). In the incubation with HLM
and ATV, 4-(pyridin-2-yl)benzaldehyde (m1) was trapped using me-
thoxylamine. The formed oxime was eluted at 5.47 min (Fig. 4G),
having a mass of [M ϩ H]ϩ ϭ 213 m/z. MS/MS of the m1-oxime
generated the ions at 181 (loss of CH3OH) and 155 (loss of C2H4NO).
The ions were interpreted in the inlaid structural diagram (Fig. 4H).
The structure of m1-oxime was confirmed by comparing the retention
time and accurate mass with those of the synthetic authentic standard
sample, which was characterized by NMR.
Identification of an ␣-Hydroxyaldehyde (M15ꢀ) in ATV Metab-
olism. In ATV-treated mouse urine samples and in vitro study using
HLM, a novel dealkylated metabolite (M15) of ATV was identified.
Meanwhile, an ␣-hydroxyaldehyde (M15Ј) was trapped by INH (Fig. 5) and
semicarbazide (data not shown) in the incubation of ATV in HLM.
FIG. 1. Metabolomic analysis of control and ATV-treated mouse urine and feces.
Wild-type mice (n ϭ 4) were treated with 50 mg/kg ATV (orally). Urine and feces
were collected for analysis. A, separation of control and ATV-treated mouse
feces in a PCA score plot. The t[1] and t[2] values represent the score of each sample
in principal component 1 and 2, respectively. B, loading S-plot generated by
OPLS-DA analysis of metabolome in ATV-treated mouse feces. C, loading S-plot
generated by OPLS-DA analysis of metabolome in ATV-treated mouse urine. The
x-axis is a measure of the relative abundance of ions, and the y-axis is a measure of
the correlation of each ion to the model. These loading plots represent the relation-
ship between variables (ions) in relation to the first and second components present
in the PCA score plot. ATV and its metabolites were labeled. The number of ions
(metabolite identification) was accordant with that in Fig. 2.

REACTIVE METABOLITES OF ATAZANAVIR
397
FIG. 2. The metabolite map of ATV. All metabolites were determined by accurate mass and mass fragmentations. Overall, 18 ATV metabolites were identified, including
5 previously reported metabolites (M1, M2, M6, M13, and M14) (ter Heine et al., 2009) and 13 novel metabolites. Among these novel ATV metabolites, one aromatic
aldehyde (m1), one ␣-hydroxyaldehyde (M15Ј), and one hydrazine (M16) were detected and characterized.
The formations of M15 and M15Ј were NADPH-dependent (Fig. 5A). less than that of ATV. The fragment ions at m/z 339 (loss of CH3OH),
The chromatograms of M15 and M15Ј-hydrazone are depicted in Fig. 200 (loss of C12H10N), and 168 (C12H10N) were interpreted in the
5A. Metabolite M15 was eluted at 4.23 min, having a protonated inlaid structural diagram (Fig. 5B). The formed M15Ј-hydrazone of
molecule [M ϩ H]ϩ at m/z ϭ 371 Da, 321 Da (loss of C17H25N2O4) INH, eluted at 4.67 min, had a protonated molecule [M ϩ H]ϩ at
FIG. 3. A scheme of m1 formation and its further metabolism. m1 was generated from ATV spontaneously with M1, and it was trapped using methoxylamine (MeONH2).
In mice, m1 can be further metabolized to acid (m2) and conjugated with glycine (m3). m1 may interact with GSH (m4) and be further metabolized to an
N-acetylcysteine-conjugate product (m5).

398
LI ET AL.
FIG. 4. Identification of metabolite m1. Urine and feces from mice were collected for 18 h after ATV treatment (50 mg/kg p.o.). In addition, in vitro studies were performed
to trap m1 using methoxylamine. Structural elucidations were performed on the basis of accurate mass measurement (mass errors less than 10 ppm) and MS/MS
fragmentations. MS/MS fragmentation was conducted with collision energy ramping from 10 to 40 eV. Major daughter ions from fragmentation were interpreted in the
inlaid structural diagrams. A, chromatograms of metabolite M1, m2, m3, and m5 in urine. B, MS/MS of ATV. C, MS/MS of M1. D, MS/MS of m2. E, MS/MS of m3.
F, MS/MS of m5. G, chromatograms of m1-oxime in the incubation with HLM. H, MS/MS of m1-oxime.

REACTIVE METABOLITES OF ATAZANAVIR
399
major fragment ions at m/z 349 (loss of H2O), 179 (C10H15NO2), and
144 (C7H14NO2). The fragment ions were interpreted in the inlaid
structural diagram (Fig. 6B).
ATV Metabolism in HLM. Fourteen ATV metabolites were
identified in the incubation with HLM (Supplemental Table 1),
including two dealkylated ATV metabolites (M1 and M15), two
aldehydes (m1 and M15Ј), four monohydroxylated ATV metabo-
lites (M2, M3, M4, and M5), one monohydroxylated ϩ monode-
hydrogenated metabolite (M6), one dihydroxylated ATV metabo-
lite (M10), one dihydroxylated ϩ monodehydrogenated metabolite
(M12), two hydrolyzed metabolites (M13 and M14), and one
hydrazine (M16). Their relative abundance is displayed in Fig. 7A.
M2 is the most abundant metabolite of ATV in the incubation with
HLM, followed by M15 and M1. The metabolic pathways of M1
and M15 should be emphasized because 1) they are major meta-
bolic pathways in ATV metabolism, 2) two aldehydes (m1 and
M15Ј) are generated simultaneously with the formation of metab-
olites M1 and M15, and 3) the generation of the hydrazine (M16)
is also dependent on M1.
Role of P450s in ATV Metabolism. The incubation of ATV with
nine different human cDNA-expressed P450s (CYP1A2, 2A6, 2B6,
2C8, 2C9, 2C19, 2D6, 2E1, and 3A4) revealed that CYP3A4 is the
primary enzyme contributing to ATV metabolism (Supplemental Ta-
ble 1). CYP3A4 and CYP2D6 cocontributed to the metabolic pathway
of metabolite M1, but CYP3A4 was more important than CYP2D6
(Supplemental Table 1). CYP3A4 was the dominant enzyme respon-
sible for the formation of metabolite M15 (Fig. 7, B and C). The
inhibitory effect on the M1 and M15 metabolic pathways was further
verified by coincubation of KCZ with HLM and cDNA-expressed
CYP3A4. In the incubation with HLM, the formations of M1 and M15
were suppressed up to 90% by KCZ at 10 ␮M (Fig. 7B). In the
incubation with CYP3A4, KCZ at 10 ␮M inhibited the formations of
M1 to 95% and M15 to 90% (Fig. 7C).
Discussion
In our current study, we used a liquid chromatography-mass
spectrometry-based metabolomic approach to profile ATV metab-
olism. We identified 18 ATV metabolites, including 5 previously
reported metabolites (M1, M2, M6, M13, and M14) (ter Heine et
al., 2009) and 13 novel metabolites (Fig. 2; Supplemental Fig. 2).
Three potential reactive metabolites, one aromatic aldehyde (m1),
one ␣-hydroxyaldehyde (M15Ј), and one hydrazine (M16), were
detected and characterized for the first time. The levels of the two
aldehydes (m1 and M15Ј) were high because they were generated
from two primary metabolic pathways of ATV, M1, and M15
(Figs. 2 and 7A).
The aromatic aldehyde (m1) that we identified in ATV metabolism
(Figs. 3 and 4) was further oxidized to the corresponding acid m2 and
then conjugated with glycine to yield the metabolite m3. This pathway
may contribute to the detoxication of aldehyde m1. On the other hand,
FIG. 5. Identification of metabolite M15Ј. Duplicate incubations were conducted in
1ϫ PBS (pH 7.4) containing ATV (50 ␮M), HLM (0.5 g of protein/l), and INH (5.0
mM), with or without NADPH (1.0 mM). The metabolites were analyzed using
UPLC-TOFMS. A, chromatograms of M15 and M15Ј-hydrazone. B, MS/MS of an N-acetylcysteine-conjugate product (m5) was detected in the ATV-
M15. C, MS/MS of M15Ј-hydrazone.
treated mouse urines. The formation of metabolite m5 indicates that
the further metabolite of m1 can interact and form an adduct with
m/z ϭ 470 Da. MS/MS of the M15Ј-hydrazone produced the major GSH (m4) (Fig. 3). The exact mechanism of m5 formation was not
ions 282 (loss of C8H15N2O3) and 162 (C10H12NO). The structural determined in the current study. On the basis of a previous report (Ji
elucidation was interpreted in the inlaid structural diagram (Fig. 5C). et al., 2007), an intermediate sulfate is proposed. In brief, m1 can be
Identification of a Hydrazine (M16) in ATV Metabolism. In the reduced to an alcohol, and subsequently sulfated to form a sulfate that
incubation with ATV and HLM, a hydrazine (M16), resulting from serves as a leaving group. The resulting sulfate reacts with GSH to
the dealkylation and hydrolysis, was identified (Fig. 6). The forma- form m4, which is further metabolized to an N-acetylcysteine-conju-
tions of M16 were NADPH-dependent (Fig. 6A). M16, eluted at 4.18 gate product (m5). Clarification of the mechanism of this interaction
min, had a protonated molecule [M ϩ H]ϩ at m/z ϭ 367 Da, 171 with GSH and its implication in ATV-related adverse effects is
(C8H13NO3) Da less than that of M1. MS/MS of M16 produced the desired in future studies.

400
LI ET AL.
FIG. 6. Identification of the metabolite M16. Duplicate incubations were conducted in 1ϫ PBS (pH 7.4) containing ATV (50 ␮M) and HLM (0.5 g of protein/l), with or
without NADPH (1.0 mM). The metabolites were analyzed using UPLC-TOFMS. A, chromatogram of M16. B, MS/MS of M16.
The hydrazine (M16) that we identified is a further metabolite of rearrangement, resulting in the toxicity (Spahn-Langguth and Benet,
1992; Tang and Lu, 2010).
M1. It has been reported that hydrazine can cause hepatic lesions and
neurotoxicity (Waterfield et al., 1993; Nicholls et al., 2001). The
detection of metabolite M16 provided more insight for the study of
ATV toxicity. The ␣-hydroxyaldehyde (M15Ј) is predicted as the
most active metabolite of ATV, which cannot be directly detected in
mouse urine or the HLM incubation system. Previous studies have
suggested that ␣-hydroxyaldehyde can form a Schiff base with the
amino group on the protein. The unstable imine can be converted into
a stable 1-amino-2-keto protein adduct by an intramolecular Amadori
CYP3A4 was determined as the primary enzyme contributing to
metabolic pathways of M1 and M15. Thus, cotreatment with ATV and
a CYP3A4 inducer will increase the formation of reactive metabolites
M15Ј, m1, and M16, which may augment ATV toxicity. Rifampicin,
which is also a CYP3A4 inducer, is a first-line drug for tuberculosis
(TB) treatment. According to the guideline of the World Health
Organization, HIV/TB coinfected patients are required to be treated
with anti-TB drugs for at least 2 weeks before the initiation of
anti-HIV treatment. In a recent clinical trial, nausea, vomiting, and
elevation of alanine aminotransferase activity were reported in all
healthy volunteers who were pretreated with rifampicin followed by
treatment with ATV and RTV (Haas et al., 2009). It is possible that
rifampicin induces CYP3A4 expression, which accelerates ATV me-
tabolism to reactive metabolites, such as m1, M15Ј, and M16, result-
ing in liver injury.
In summary, we identified three potential reactive metabolites of
ATV, which included two aldehydes and one hydrazine. CYP3A4 was
determined to be the primary enzyme contributing to the formation of
these aldehydes and hydrazine. Further studies are suggested to illus-
trate the role of these potential reactive metabolites in ATV-related
adverse effects.
Acknowledgments
We thank the National Institutes of Health AIDS Research and Reference
Reagent Program for providing the HIV protease inhibitors; Jin Chou and Dr.
Zhonghua Peng (University of Missouri-Kansas City) for NMR service; and
Dr. Martha Montello for editing the manuscript.
Authorship Contributions
Participated in research design: Li and Ma.
Conducted experiments: Li, Lu, and Wang.
Contributed new reagents or analytic tools: Li.
Performed data analysis: Li and Ma.
Wrote or contributed to the writing of the manuscript: Li and Ma.
FIG. 7. ATV metabolism in HLM and the role of CYP3A in ATV metabolism.
Duplicate incubations were conducted in 1ϫ PBS (pH 7.4) containing ATV (50
␮M), NADPH (1.0 mM), HLM (0.5 g of protein/l), or CYP3A4 (10 nM). KCZ (10
␮M) was used in the inhibitory test. A, relative abundance of ATV metabolites
generated from HLM. B, effects of KCZ (10 ␮M) on the formations of M1 and M15
in the incubation with HLM. C, effects of KCZ (10 ␮M) on the formations of M1
and M15 in the incubation with cDNA-expressed CYP3A4. The data are expressed
as means.
References
Baillie TA (2006) Future of toxicology-metabolic activation and drug design: challenges and
opportunities in chemical toxicology. Chem Res Toxicol 19:889–893.
Busti AJ, Hall RG, and Margolis DM (2004) Atazanavir for the treatment of human immuno-
deficiency virus infection. Pharmacotherapy 24:1732–1747.
Chen C, Meng L, Ma X, Krausz KW, Pommier Y, Idle JR, and Gonzalez FJ (2006) Urinary
metabolite profiling reveals CYP1A2-mediated metabolism of NSC686288 (aminoflavone).
J Pharmacol Exp Ther 318:1330–1342.

REACTIVE METABOLITES OF ATAZANAVIR
401
Croom KF, Dhillon S, and Keam SJ (2009) Atazanavir: a review of its use in the management
of HIV-1 infection. Drugs 69:1107–1140.
Dieckhaus CM, Thompson CD, Roller SG, and Macdonald TL (2002) Mechanisms of idiosyn-
cratic drug reactions: the case of felbamate. Chem Biol Interact 142:99–117.
Eholie´ SP, Lacombe K, Serfaty L, Wendum D, and Girard PM (2004) Acute hepatic cytolysis in
an HIV-infected patient taking atazanavir. AIDS 18:1610–1611.
Evans DC, Watt AP, Nicoll-Griffith DA, and Baillie TA (2004) Drug-protein adducts: an
industry perspective on minimizing the potential for drug bioactivation in drug discovery and
development. Chem Res Toxicol 17:3–16.
Goldsmith DR and Perry CM (2003) Atazanavir. Drugs 63:1679–1693; discussion 1694–1675.
Guengerich FP and MacDonald JS (2007) Applying mechanisms of chemical toxicity to predict
drug safety. Chem Res Toxicol 20:344–369.
O’Brien PJ, Siraki AG, and Shangari N (2005) Aldehyde sources, metabolism, molecular toxicity
mechanisms, and possible effects on human health. Crit Rev Toxicol 35:609–662.
Rivas P, Morello J, Garrido C, Rodríguez-No´voa S, and Soriano V (2009) Role of atazanavir in
the treatment of HIV infection. Ther Clin Risk Manag 5:99–116.
Spahn-Langguth H and Benet LZ (1992) Acyl glucuronides revisited: is the glucuronidation
process a toxification as well as a detoxification mechanism? Drug Metab Rev 24:5–47.
Sulkowski MS (2004) Drug-induced liver injury associated with antiretroviral therapy that
includes HIV-1 protease inhibitors. Clin Infect Dis 38 (Suppl 2):S90–S97.
Swainston Harrison T and Scott LJ (2005) Atazanavir: a review of its use in the management of
HIV infection. Drugs 65:2309–2336.
Tang W and Lu AY (2010) Metabolic bioactivation and drug-related adverse effects: current
status and future directions from a pharmaceutical research perspective. Drug Metab Rev
42:225–249.
ter Heine R, Hillebrand MJ, Rosing H, van Gorp EC, Mulder JW, Beijnen JH, and Huitema AD
(2009) Identification and profiling of circulating metabolites of atazanavir, a HIV protease
inhibitor. Drug Metab Dispos 37:1826–1840.
Haas DW, Koletar SL, Laughlin L, Kendall MA, Suckow C, Gerber JG, Zolopa AR, Bertz R,
Child MJ, Hosey L, et al. (2009) Hepatotoxicity and gastrointestinal intolerance when healthy
volunteers taking rifampin add twice-daily atazanavir and ritonavir. J Acquir Immune Defic
Syndr 50:290–293.
Havlir DV and O’Marro SD (2004) Atazanavir: new option for treatment of HIV infection. Clin
Infect Dis 38:1599–1604.
Ji T, Ikehata K, Koen YM, Esch SW, Williams TD, and Hanzlik RP (2007) Covalent modifi-
cation of microsomal lipids by thiobenzamide metabolites in vivo. Chem Res Toxicol 20:701–
708.
Waterfield CJ, Turton JA, Scales MD, and Timbrell JA (1993) Investigations into the effects of
various hepatotoxic compounds on urinary and liver taurine levels in rats. Arch Toxicol
67:244–254.
Li F, Wang L, Guo GL, and Ma X (2010) Metabolism-mediated drug interactions associated with
ritonavir-boosted tipranavir in mice. Drug Metab Dispos 38:871–878.
Nicholls AW, Holmes E, Lindon JC, Shockcor JP, Farrant RD, Haselden JN, Damment SJ,
Waterfield CJ, and Nicholson JK (2001) Metabonomic investigations into hydrazine toxicity
in the rat. Chem Res Toxicol 14:975–987.
Address correspondence to: Dr. Xiaochao Ma, Department of Pharmacol-
ogy, Toxicology and Therapeutics, University of Kansas Medical Center, Kansas
City, KS 66160. E-mail: xma2@kumc.edu