Synthesis and evaluation of nevirapine analogs to study the metabolic activation of nevirapine

Article abstract of DOI:10.1016/j.dmpk.2020.01.006

Nevirapine (NVP) is widely used as a non-nucleoside reverse transcriptase inhibitor of HIV-1, however, it is associated with severe skin and liver injury. The mechanisms of these adverse reactions are not yet clear, but the metabolic activation of NVP is thought to be related to the injury process. Until now, several metabolic activation pathways of NVP have been reported. In this study, in order to identify the reactive metabolite of NVP mainly responsible for CYP inhibition and liver injury, we synthesized five NVP analogs designed to avoid the proposed bioactivation pathway and evaluated their metabolic stabilities, CYP3A4 time-dependent inhibitory activities, and cytotoxicity. As a result, only a pyrimidine analog of NVP, which could avoid the formation of a reactive epoxide intermediate, did not inhibit CYP3A4. Outside of this compound, the other synthesized compounds, which could avoid the generation of a reactive quinone-methide intermediate, inhibited CYP3A4 equal to or stronger than NVP. The pyrimidine analog of NVP did not induce cytotoxicity in HepG2 and transchromosomic HepG2 cells, expressing major four CYP enzymes and CYP oxidoreductase. These results indicated that the epoxide intermediate of NVP might play an important role in NVP-induced liver injury.

Full text of DOI:10.1016/j.dmpk.2020.01.006

Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
Contents lists available at ScienceDirect
Drug Metabolism and Pharmacokinetics
journal homepage: http://www.journals.elsevier.com/drug-metabolism-and-
pharmacokinetics
Regular Article
Synthesis and evaluation of nevirapine analogs to study the metabolic
activation of nevirapine
,
Yasuhiro Tateishi a, Tomoyuki Ohe a *, Daisuke Yasuda a, Kyoko Takahashi a,
,
, **
Shigeo Nakamura b, Yasuhiro Kazuki c d, Tadahiko Mashino a
a Faculty of Pharmacy, Keio University, 1-5-30 Shibakoen, Minato-ku, Tokyo, Japan
b Department of Chemistry, Nippon Medical School, 1-7-1 Kyonan-cho, Musashino, Tokyo, Japan
c Department of Biomedical Science, Institute of Regenerative Medicine and Biofunction, Graduate School of Medical Science, Tottori University, 86 Nishi-
cho, Yonago, Tottori, Japan
d Chromosome Engineering Research Center (CERC), Tottori University, 86 Nishi-cho, Yonago, Tottori, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Nevirapine (NVP) is widely used as a non-nucleoside reverse transcriptase inhibitor of HIV-1, however, it
is associated with severe skin and liver injury. The mechanisms of these adverse reactions are not yet
clear, but the metabolic activation of NVP is thought to be related to the injury process. Until now, several
metabolic activation pathways of NVP have been reported. In this study, in order to identify the reactive
metabolite of NVP mainly responsible for CYP inhibition and liver injury, we synthesized five NVP an-
alogs designed to avoid the proposed bioactivation pathway and evaluated their metabolic stabilities,
CYP3A4 time-dependent inhibitory activities, and cytotoxicity. As a result, only a pyrimidine analog of
NVP, which could avoid the formation of a reactive epoxide intermediate, did not inhibit CYP3A4. Outside
of this compound, the other synthesized compounds, which could avoid the generation of a reactive
quinone-methide intermediate, inhibited CYP3A4 equal to or stronger than NVP. The pyrimidine analog
of NVP did not induce cytotoxicity in HepG2 and transchromosomic HepG2 cells, expressing major four
CYP enzymes and CYP oxidoreductase. These results indicated that the epoxide intermediate of NVP
might play an important role in NVP-induced liver injury.
Received 19 September 2019
Received in revised form
19 December 2019
Accepted 29 January 2020
Available online 28 February 2020
Keywords:
Nevirapine
Metabolic activation
Cytochrome P450
Hepatotoxicity
HepG2
© 2020 The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
1. Introduction
these adverse drug reactions are thought to be associated with the
metabolic activation of NVP described below.
Nevirapine (NVP, Fig. 1) is the first nonnucleoside reverse
transcriptase inhibitor widely used for the combination treatment
of HIV-1 infections, particularly in developing countries because of
its ease of use and low cost. Despite its effectiveness, NVP is asso-
ciated with severe liver and skin injuries, which occur with an
incidence of 1.0 and 0.3%, respectively [1]. Although it was reported
that NVP itself inhibits human hepatocyte growth [2] or induces
apoptosis and mitochondrial dysregulation in HepG2 cells [3,4],
In humans, NVP is metabolized by cytochrome P450 (CYP) and is
mainly excreted in the urine. An in vivo study showed that the
metabolites of NVP that are detected in human urine are the
glucuronide conjugates of 2-, 3-, 8-, and 12-hydroxyNVP [5]. It was
also reported that CYP3A4 and CYP2B6 are responsible for the
formation of Phase I NVP metabolites; the former is involved in the
formation of 2- and 12-hydroxyNVP and the latter catalyzes 3- and
8-hydroxyNVP [6]. 12-HydroxyNVP (12-OH-NVP) is further
metabolized by sulfotransferase 1A1 (SULT1A1) to yield 12-
sulfoxyNVP [7] or is oxidized to form 4-carboxyNVP [8].
Evidence for the in vivo bioactivation of NVP was also reported
where NVP could adduct with human serum albumin or hemo-
globin [8,9], or two NVP mercapturates substituted at the C-3 and
C-12 positions were identified, suggesting the possibility of the
metabolic activation of NVP in humans [10]. Although various
bioactivation pathways of NVP have been investigated to date, it is
generally recognized that 12-OH-NVP is a key metabolite related to
Abbreviations: NVP, nevirapine; CYP, cytochrome P450; 12-OH-NVP, 12-
hydroxyNVP; SULT1A1, sulfotransferase 1A1; TC-HepG2, transchromosomic
HepG2; G6P, glucose-6-phosphate; G6PDH, G6P dehydrogenase; HLM, human liver
microsomes; LC-MS, liquid chromatography with mass spectrometry; BSO,
buthionine-S,R-sulfoximine; AFT, aflatoxin B1; SIM, selected ion monitoring.
* Corresponding author.
L-
** Corresponding author.
E-mail address: ohe-tm@pha.keio.ac.jp (T. Ohe).
https://doi.org/10.1016/j.dmpk.2020.01.006
1347-4367/© 2020 The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.

Y. Tateishi et al. / Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
239
Fig. 1. Structure of NVP and its proposed bioactivation pathways.
NVP-induced toxicity. Chen et al. proposed that the reactive
quinone-methide formed from 12-OH-NVP is responsible for skin
rashes [7]. Other studies showed that 12-sulfoxyNVP formed from
12-OH-NVP is also related to NVP-induced skin rashes [11,12].
Furthermore, 12-OH-NVP causes upregulation of various genes in
the skin that induce immune reactions leading to skin rashes [13].
Recently, Fang et al. demonstrated that 12-sulfoxyNVP formed ad-
ducts with deoxynucleosides and amino acids [14].
of NVP and its analogs as an index of metabolic activation. In
addition, the hepatotoxicity of all the produced compounds was
assessed using transchromosomic HepG2 (TC-HepG2) cells
expressing major CYP enzymes (CYP2C9, CYP2C19, CYP2D6, and
CYP3A4) and CYP oxidoreductase [17]. Our data suggests that NVP-
2,3-epoxide, together with the reactive quinone-methide metabo-
lites, may play an important role in NVP-induced liver toxicity.
Bioactivation of NVP also appears to be the cause of liver in-
juries. Unlike the skin injury, several reactive metabolites were
reported for the liver injury. It was suggested that the quinone-
methide metabolite might contribute to NVP-induced liver in-
juries [15]. On the other hand, not only quinone-methide or 12-
sulfoxyNVP produced from 12-OH-NVP but other reactive metab-
olites such as quinone-imine, 2,3-NVP-quinone, and NVP-2,3-
epoxide have also been reported (Fig. 1) [7,15,16]. These metabo-
lites may covalently bind to proteins and ultimately cause liver
injuries.
2. Materials and methods
2.1. Materials
NADPþ, glucose-6-phosphate (G6P) and G6P dehydrogenase
(G6PDH) were purchased from Roche Diagnostics (Basel,
Switzerland). NVP was obtained from Tokyo Chemical Industry
(Tokyo, Japan). All other reagents used for microsomal and cyto-
toxicity experiments were of analytical grade.
In this study, for the purpose of determining the metabolic
activation pathway involved during hepatotoxicity, we synthesized
five NVP analogs (1e5, Fig. 2) designed to restrict the formation of
quinone-methide, quinone-imine, quinone and epoxide in-
termediates in the NVP A-ring (methylpyridine ring) which is
known to be metabolically activated as described above. Following
this, we evaluated the CYP3A4 time-dependent inhibitory activity
2.2. Synthesis of NVP analogs
We chemically synthesized NVP derivatives 1, 2 and novel NVP
analogs 3e5 as shown in Fig. 3. Briefly, for the synthesis of 1 or 5,
condensation of 3-amino-2-chloropyridine or 4-amino-3-chloro-5-
methylpyridazine prepared by methylation and the following ami-
nation of 4,5-dichloropyridazine-3(2H)-one, with 2-chloronicotinoyl
Fig. 2. Chemical structures of the synthesized NVP analogs.

240
Y. Tateishi et al. / Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
Fig. 3. Synthesis of NVP analogs. Reagents and Conditions; (a) BnBr, K2CO3, DMF, rt, 20 h; (b) dimethyl malonate, tBuOK, THF, rt (0.3 h)-reflux (1.5e3.5 h); (c) conc. HCl, reflux, 2 h,
then KF, H2O, DMF, 170 ꢀC, 3 h; (d) AlCl3, toluene, 50 ꢀC, 20 min; (e) POCl3, reflux, 1e2 h; (f) NaN3, DMF, 50 ꢀC, 2.5 h; (g) NaBH4, THF/CH3OH, rt, 0.5 h; (h) 2-chloronicotinoyl chloride,
NaH or pyridine, THF or DMF, 0 ꢀCert, 1e3 h; (i) cyclopropylamine, K2CO3, toluene/DMF or DMSO, 120 ꢀC, 4.5e22 h; (j) NaH, toluene/DMF, reflux, 2e23 h; (k) cyclopropylamine,
triethylamine or K2CO3, THF or toluene, rte120 ꢀC, 0.2e15.5 h; (l) conc. HCl, reflux, 1e1.5 h; (m) H2, Pd/C, CH3OH, rt, 12e19.5 h; (n) NaH, DMF, 60 ꢀC, 2 h, then CH3I, DMF, rt, 2 h.
chloride gave nicotinamides. These amides were reacted with
cyclopropylamine, followed by ring closure to produce 1 or 5. In the
case of compounds 3 and 4, condensation of diaminopyridine or
diaminopyrimidine prepared stepwise from 2,4-dichloro-3-
nitropyridine or 4,6-dichloro-5-nitropyrimidine as a starting mate-
rial, with 2-chloronicotinoyl chloride gave nicotinamides, and the
following ring closure reaction produced compounds 3 or 4. Com-
pound 2 was synthesized by direct N-methylation of NVP using
methyl iodide. More detailed procedures and analytical data for the
synthesized compounds are shown in the Supplemental Data.
gradient elution profile: 30% solvent B for 1 min, 30e100% B in
5 min, 100% B for 4 min, then 100 to 30% B in 0.1 min followed by
30% B for another 3.9 min (15 min in total). The eluent was intro-
duced directly into the mass spectrometer via electrospray ioni-
zation using the positive ion mode.
2.4. CYP inactivation
According to reported procedures [18], midazolam 10-hydroxy-
lase activity was determined to quantify the time-dependent loss of
CYP3A4 activity of HLM in the presence of NVP and its analogs. In
2.3. Metabolic stability in liver microsomes
brief, primary incubations included 100 mM of the following test
compounds, 1 mM NADPþ, 10 mM G6P, 2 unit/mL G6PDH, 1 mg/mL
HLM, 3 mM MgCl2, and 0.1 M K-Pi buffer (pH 7.4). The mixtures
Pooled human (200 donors) liver microsomes (HLM) were
purchased from XenoTech (Lenexa, KS). All incubations were con-
ducted at 37 ꢀC in a water bath. Stock solutions of the test com-
pounds were prepared in methanol. The incubation mixtures
were incubated for 30 min. Then, 50
mL aliquots of the primary
incubation mixtures were transferred to a secondary incubation
with a final volume of 0.5 mL, which included 40 mM midazolam,
containing HLM (0.4 mg protein/mL), the test compounds (1
m
M),
1 mM NADPþ, 10 mM G6P, 2 unit/mL G6PDH, 3 mM MgCl2, and
0.1 M K-Pi buffer (pH 7.4). The midazolam mixture was incubated
for 10 min and stopped by the addition of 1 mL of an ice-cold mixed
MgCl2 (3 mM), G6P (10 mM), and G6PDH (2 unit/mL) in 0.5 mL of
0.1 M K-Pi buffer (pH 7.4) were preincubated for 3 min. The reaction
was initiated by adding NADPþ (1 mM) and then incubated for
30 min. The reaction was stopped by the addition of 1 mL of an ice-
acetonitrile/methanol (2/1) solution containing 1
mixture was then centrifuged at 10,000ꢁg for 10 min. The super-
natant (10
L) was injected and 10-hydroxymidazolam was quan-
mM cortisone. The
cold mixed acetonitrile/methanol (2/1) solution containing 1
cortisone. The mixture was centrifuged at 10,000ꢁg for 10 min. The
supernatant (20 L) was injected and analyzed by liquid chroma-
mM
m
tified by the LC-MS analysis. The CYP3A4 remaining activities (%) of
test compounds were calculated as the percentage of the initial
activity obtained without preincubation. The CYP3A4 remaining
activity of control (treated with solvent only) was also calculated.
Then, time-dependent inhibition (%) of CYP3A4 was expressed as
the difference between the remaining activities of control and each
compound. The equations used for the calculation were shown as
follows:
m
tography with mass spectrometry (LC-MS, 6120; Agilent Technol-
ogies, Palo Alto, CA). The HPLC mobile phase A was water with 0.1%
formic acid, and mobile phase B was acetonitrile with 0.1% formic
acid. Chromatographic separations were performed using an
InertSustain C18 column (4.6 ꢁ 100 mm, 3.0
mm; GL Sciences,
Tokyo, Japan) at a flow rate of 0.5 mL/min under the following

Y. Tateishi et al. / Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
241
Table 1
Metabolic stability and CYP3A4 time-dependent inhibitory activity of NVP and its analogs.
Compound
Metabolic Stability (% remaining)
CYP3A4 Time Dependent Inhibition (%)
NVP
1
2
3
4
93.6 7.5
86.2 4.2
94.5 3.5
96.1 4.8
72.2 1.2
23.6 0.9
13.8* 4.9
3.9 8.8
13.3* 9.3
32.6*, y 1.8
0.4y 3.7
5
43.0*, y 4.6
For testing metabolic stability, the test compounds were incubated with HLM for 30 min. The remaining amount of test compounds was expressed
as percentages relative to the incubation (ꢂ) sample. For the CYP time-dependent assay, CYP3A4 remaining activity (%) in HLM was calculated as
the percentage of the initial activity obtained without preincubation. The data show differences between the remaining activities of control
(treated with solvent only) and each compound. Each value represents the mean of three samples S.D. *p < 0.01; significantly different from
negative control, yp < 0.01; significantly different from NVP, Student's t-test.
CYP3A4 remaining activity (%) ¼ (activity with preincubation/ac-
3.2. Time-dependent CYP3A4 inhibition of NVP and its analogs
tivity without preincubation) x 100
Time-dependent inhibition of CYP3A4 was calculated as the loss
of midazolam 10-hydroxylase activity compared with the control
which was preincubated with only solvent. NVP inhibited CYP3A4
significantly, and the N-methyl derivative 2 also showed the same
level of inhibitory activity as NVP. Although both the pyridine-
analogue 3 and the pyridazine-analog 5 inhibited CYP3A4 more
strongly, the pyrimidine-analog 4 showed weaker inhibitory ac-
tivity when compared with NVP. The desmethyl derivative 1
showed lower inhibitory activity than NVP, but the difference was
not statistically significant.
Time-dependent CYP3A4 inhibition (%) ¼ CYP3A4 remaining ac-
tivity of control e CYP3A4 remaining activity of compounds
2.5. Cell cultures
Human hepatocarcinoma HepG2 cells were purchased from the
American Type Culture Collection (Manassas, VA). TC-HepG2 cells
expressing major CYP enzymes involved in drug metabolism
(CYP2C9, CYP2C19, CYP2D6, and CYP3A4) and CYP oxidoreductase
were previously established [17]. The cells were maintained in
Dulbecco's modified Eagle's medium (NAKALAI TESQUE, Inc., Kyoto,
Japan) containing 10% fetal bovine serum (GIBCO/Life Technologies,
Grand Island, NY) and 1% penicillin-streptomycin (Sigma-Aldrich,
St. Louis, MO) in a 5% CO2 incubator under a humidified atmosphere
at 37 ꢀC.
3.3. Cytotoxicity of NVP and its analogs in HepG2/TC-HepG2 cells
To evaluate the metabolic activation of NVP and its analogs, the
cell toxicity of HepG2/TC-HepG2 cells was assessed by a WST-8
assay. We also evaluated the effects of pretreatment with buthio-
nine sulfoximine (BSO), an inhibitor of glutathione (GSH) biosyn-
thesis, since GSH traps reactive metabolites and detoxifies them. As
shown in Fig. 4A, no remarkable differences in cell viabilities be-
tween HepG2 and TC-HepG2 cells after the addition of NVP was
observed without pretreatment using BSO, while the cell viabilities
were significantly lower in TC-HepG2 cells treated with AFT than in
the parental HepG2 cells under the same condition. On the other
hand, under GSH-depleted conditions with the pretreatment of
BSO, the cytotoxicity of NVP in TC-HepG2 cells was significantly
stronger than in the HepG2 cells. Based on these results, the cyto-
toxicity of the synthesized compounds was investigated using GSH-
depleted TC-HepG2 cells, as the metabolic activation of compounds
can be more sensitively evaluated under these conditions. For a
comparison purpose, we also examined the toxicity without the
pretreatment of BSO. As shown in Fig. 4B, compounds 3 and 4 were
less toxic when compared with NVP, while compounds 2 and 5 had
significantly stronger toxicity in the BSO (þ) condition. Under the
BSO (ꢂ) condition, there are no remarkable differences in cell
viability among compounds.
2.6. Cytotoxicity assay
Cytotoxicity was assessed by a WST-8 assay using a Cell
Counting Kit-8 (CCK-8, Dojindo Laboratories, Kumamoto, Japan).
The cells were seeded (1 ꢁ 104 cells/well) on a 96-well plate (Iwaki/
Asahi Techno Glass Corporation, Tokyo, Japan). At the time of
seeding,
PBS (200
L
-buthionine-S,R-sulfoximine (BSO, Sigma) solutions in
M) were added. For the BSO (ꢂ) group, the cells were
m
treated with PBS. After 24 h of incubation, the test compounds or
aflatoxin B1 (AFT) as the positive control, dissolved in DMSO were
added. DMSO only was added for negative control (final DMSO
concentration was 0.5% v/v). The cells were incubated with the test
compounds for 48 h, and then the CCK-8 solution was added. After
another 4 h of incubation, the absorbance at 450 nm was measured
(Reference 600 nm) using Infinite M1000 PRO Microplate Readers
€
(TECAN, Mannedorf, Switzerland), and the cell viability was
determined.
3. Results
4. Discussion
3.1. Metabolic stability of NVP and its analogs
As shown in Fig. 1, several bioactivation pathways of NVP were
proposed to explain the mechanism of NVP-induced liver and/or
skin injury. The purpose of the present study is to identify the
reactive metabolite of NVP that is mainly responsible for the time-
dependent inhibition of CYP and eventual hepatic toxicity by using
chemically synthesized NVP analogs. Several reports suggested that
the reactive quinone-methide intermediate might contribute to the
toxicity, therefore we first synthesized desmethyl derivative 1 and
N-methyl derivative 2, which were previously reported by Hargrave
NVP and analogs 1 through 5 were incubated with pooled hu-
man liver microsomes. All compounds were detected using
selected ion monitoring (SIM) mode, and the metabolic stability
was defined as the percentage of parent compounds remaining
during the incubation. As shown in Table 1, 1, 2 and 3 were meta-
bolically stable, while the stability of both the pyrimidine-analog 4
and the pyridazine-analog 5 were lower than NVP.

242
Y. Tateishi et al. / Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
modified the NVP A-ring to other heterocycles. Interestingly, the
time-dependent CYP3A4 inhibition of the pyrimidine-analog 4 was
significantly reduced, while those of the pyridine-analog 3 and the
pyridazine-analog 5 were markedly increased in comparison with
NVP. Although these compounds may form quinone-methide me-
tabolites from a structural viewpoint, not all of these three com-
pounds showed CYP inhibition. This also suggests the presence of
an alternative reactive metabolite responsible for NVP-induced CYP
inhibition outside of the quinone-methide intermediate. Consid-
ering that the N-methyl derivative 2 inhibited CYP3A4 at the same
level as NVP, the quinone-methide intermediate is not thought to
be the only contributor to NVP-induced CYP inhibition, since this
type of reactive metabolite is difficult to be generated from com-
pound 2. On the other hand, reactive epoxide intermediates could
be formed from all the tested compounds except for the
pyrimidine-analog 4. Based on this structural insight along with the
results of the microsomal stabilities, the reactive epoxide in-
termediates of NVP might also be responsible for NVP-induced CYP
inhibition. Indeed, Dekker et al. [20] reported that NVP-3-GSH
derived from NVP-2,3-epoxide, not NVP-12-GSH derived from the
quinone-methide intermediate, was the only GSH-conjugate
detected when NVP was incubated with HLM and GSH.
Next, we examined the cytotoxicity of the synthesized com-
pounds in a liver-derived cell line, HepG2 cells. HepG2 cells express
little amount of CYP enzymes and are not appropriate to evaluate
the toxicity of metabolites. Thus, we used TC-HepG2 cells
expressing the four major CYP enzymes (CYP2C9, CYP2C19,
CYP2D6, and CYP3A4). Unexpectedly, the cytotoxicity of NVP in TC-
HepG2 cells was comparable to that in HepG2 cells (Fig. 4A, BSO
(ꢂ)). This result raised the possibility that excess GSH levels in cells
lessen the toxicity of NVP by trapping reactive NVP metabolites.
Therefore, to definitively detect the toxicity being caused by reac-
tive metabolites, we conducted cytotoxicity assays using BSO which
inhibits GSH biosynthesis. By using a pretreatment of BSO, NVP-
induced cytotoxicity was significantly increased in the TC-HepG2
cells when compared to the HepG2 cells and the BSO (ꢂ) group
(Fig. 4A). These data indicates that the metabolism of NVP by CYP
enzymes enhances NVP-induced cytotoxicity. Based on this result,
we assessed the toxicity of the synthesized NVP analogs under
depleted GSH conditions. For a comparison purpose, we also
examined the toxicity without the pretreatment of BSO. As shown
in Fig. 4B, compounds 3 and 4 significantly decreased toxicity when
compared with NVP, and compounds 2 and 5 showed stronger
toxicity than NVP in the BSO (þ) condition, while such differences
were not clearly observed in BSO (ꢂ) condition. This result also
suggests that the remarkable toxicity of 2 and 5 may be due to their
reactive metabolites detoxified by GSH in the same way as NVP and
the BSO (þ) condition is useful in order to compare the cytotoxicity
of reactive metabolites formed from the synthesized compounds in
the present study. The toxicity of desmethyl derivative 1 was
comparable to NVP. Since the cytotoxicity of compounds 1 and 2
were similar to or stronger than that of NVP, it is suggested that a
reactive epoxide intermediate, rather than a quinone-methide
metabolite, may play an important role in NVP-induced cytotox-
icity. Indeed, compounds 3 and 4, which have difficulty generating
reactive 2,3-epoxides, showed remarkably lower cytotoxicity than
NVP. On the other hand, compound 5, which was also thought to
avoid the generation of the 2,3-epoxide, did not reduce the toxicity.
As shown in Table 1, only compound 5 was metabolically unstable,
and this might contribute to its cytotoxicity. The cytotoxicity of NVP
analogs did not correlate completely with CYP inhibition such as
compound 3 which showed more potent time-dependent CYP3A4
inhibition but lower cytotoxicity than NVP. These results might
indicate that the key reactive metabolite is different between NVP-
induced time-dependent CYP inhibition and hepatotoxicity.
Fig. 4. Cytotoxicity to HepG2/TC-HepG2 cells of NVP and its analogs. A: Cytotoxicity
in HepG2 or TC-HepG2 cells after 48 h incubation with NVP with or without the
pretreatment with BSO (200
group. The control group was incubated with DMSO. AFT was used as a positive control
at 3.3
M. The bars indicate relative cell viability compared with control of BSO (ꢂ) in
each cell line. Each value represents the mean of six samples S.D. *p < 0.05,
***p 0.01,
0.05, yyp
mM). Cells were pretreated with PBS (ꢂ) for BSO (ꢂ)
m
<
0.001; significantly different from each control, yp
<
<
yyyp < 0.001; significantly different from corresponding viability of BSO (ꢂ), #p < 0.05,
##p < 0.01, ###p < 0.001; significantly different from corresponding viability of
HepG2 cells, Student's t-test. B: Cytotoxicity in TC-HepG2 cells after 48 h incubation
with NVP and its analogs (200
(200
mM) with or without the pretreatment with BSO
M). Cells were pretreated with PBS (ꢂ) for BSO (ꢂ) group. The control group was
m
incubated with DMSO. Bars indicate relative cell viability compared with control of
BSO (ꢂ). Each value represents the mean of three samples S.D. *p < 0.05, **p < 0.01,
***p < 0.001; significantly different from each control, $p < 0.05, $$p < 0.01; signifi-
cantly different from NVP, yyp < 0.01, yyyp < 0.001; significantly different from cor-
responding viability of BSO (ꢂ), Student's t-test.
et al. [19] which could avoid the formation of quinone-methide
metabolites.
Table 1 shows the metabolic stabilities of 1 and 2 and their
CYP3A4 time-dependent inhibitory activities. Although both com-
pounds 1 and 2 were metabolically stable to the same degree as
NVP, the CYP inhibitory activities were different between the two
compounds; compound 1 tended to inhibit CYP3A4 weaker than
NVP (no significant difference) while compound 2 inhibited com-
parable to NVP. These results indicate that not only the quinone-
methide intermediates but also other reactive metabolites might
play an important role in NVP-induced CYP inhibition. Encouraged
with this data, we also synthesized novel NVP analogs, which

Y. Tateishi et al. / Drug Metabolism and Pharmacokinetics 35 (2020) 238e243
243
Among the synthesized analogs, compound 4 which was
References
designed to mitigate the formation of reactive epoxide metabolites
tended to reduce the time-dependent CYP3A4 inhibition and TC-
HepG2 cytotoxicity compared with NVP. Previously, Srivastava
et al. showed that NVP-3-mercapturate, the in vivo metabolite
produced via NVP-3-GSH, was detected as a major metabolite in
urine of NVP-treated patients [10], while NVP-12-mercapturate was
also detected from human urine. In addition, Dekker et al. [20]
reported that NVP-3-GSH derived from NVP-2,3-epoxide was the
only GSH-conjugate detected in HLM fortified with GSH as
described above. Considering these reports and the present study
together, it is possible that the reactive epoxide metabolite of NVP
may play a significant role in the liver toxicity in human.
In the current study, we examined time-dependent CYP inhi-
bition and cytotoxicity in liver-derived cells to evaluate the meta-
bolic activation of compounds. However, it is widely accepted that
the emergence of drug-induced liver injury is thought to be a more
complex process and associated with an immune-mediated
pathway. Further studies are needed to fully understand the sig-
nificance of the reactive epoxide metabolite in the liver toxicity
induced by NVP.
[1] Pollard RB, Robinson P, Dransfield K. Safety profile of nevirapine, a non-
nucleoside reverse transcriptase inhibitor for the treatment of human im-
munodeficiency virus infection. Clin Therapeut 1998;20:1071e92. https://
doi.org/10.1016/S0149-2918(98)80105-7.
[2] Fang JL, Beland FA. Differential responses of human hepatocytes to the non-
nucleoside HIV-1 reverse transcriptase inhibitor nevirapine. J Toxicol Sci
2013;38:741e52. https://doi.org/10.2131/jts.38.741.
[3] Wongtrakul J, Paemanee A, Wintachai P, Thepparit C, Roytrakul S, Thongtan T,
et al. Nevirapine induces apoptosis in liver (HepG2) cells. Asian Pac J Trop Med
2016;9:547e53. https://doi.org/10.1016/j.apjtm.2016.04.015.
[4] Paemanee A, Sornjai W, Kittisenachai S, Sirinonthanawech N, Roytrakul S,
Wongtrakul J, et al. Nevirapine induced mitochondrial dysfunction in HepG2
cells. Sci Rep 2017;7:9194. https://doi.org/10.1038/s41598-017-09321-y.
[5] Riska P, Lamson M, Macgregor T, Sabo J, Hattox S, Pav J, et al. Disposition and
biotransformation of the antiretroviral drug NVP in humans. Drug Metab
Dispos 1999;27:895e901.
[6] Erickson DA, Mather G, Trager WF, Levy RH, Keirns JJ. Characterization of the
in vitro biotransformation of the HIV-1 reverse transcriptase inhibitor nevi-
rapine by human hepatic cytochromes P-450. Drug Metab Dispos 1999;27:
1488e95.
[7] Chen J, Mannargudi BM, Xu L, Uetrecht JP. Demonstration of the metabolic
pathway responsible for nevirapine-induced skin rash. Chem Res Toxicol
2008;21:1862e70. https://doi.org/10.1021/tx800177k.
[8] Caixas U, Antunes AMM, Marinho AT, Godinho ALA, Grilo NM, Marques MM,
et al. Evidence for nevirapine bioactivation in man: searching for the first step
in the mechanism of nevirapine toxicity. Toxicology 2012;301:33e9. https://
doi.org/10.1016/j.tox.2012.06.013.
[9] Meng X, Howarth A, Earnshaw CJ, Jenkins RE, French NS, Back DJ, et al.
Detection of drug bioactivation in vivo: mechanism of nevirapine-albumin
conjugate formation in patients. Chem Res Toxicol 2013;26:575e83. https://
doi.org/10.1021/tx4000107.
In conclusion, we have synthesized five NVP analogs in order to
identify the reactive metabolite of NVP mainly responsible for NVP-
induced liver injury. The present study indicates that an alternative
pathway, in which a reactive epoxide intermediate is generated,
might play an important role in the metabolic activation of NVP.
[10] Srivastava A, Lian LY, Maggs JL, Chaponda M, Pirmohamed M, Williams DP,
et al. Quantifying the metabolic activation of nevirapine in patients by inte-
grated applications of NMR and mass spectrometries. Drug Metab Dispos
2010;38:122e32. https://doi.org/10.1124/dmd.109.028688.
Authorship contributions
[11] Sharma AM, Klarskov K, Uetrecht JP. Nevirapine bioactivation and covalent
binding in the skin. Chem Res Toxicol 2013;26:410e21. https://doi.org/
10.1021/tx3004938.
[12] Sharma AM, Novalen M, Tanino T, Uetrecht JP. 12-OH-nevirapine sulfate,
formed in the skin, is responsible for nevirapine-induced skin rash. Chem Res
Toxicol 2013;26:817e27. https://doi.org/10.1021/tx400098z.
Participated in research design: Tateishi, Ohe, Takahashi,
Yasuda, Nakamura and Mashino.
Conducted experiments: Tateishi.
Contributed reagents or analytical tools: Kazuki.
Performed data analysis: Tateishi.
[13] Zhang X, Sharma AM, Uetrecht JP. Identification of danger signals in
nevirapine-induced skin rash. Chem Res Toxicol 2013;26:1378e83. https://
doi.org/10.1021/tx400232s.
Wrote or contributed to the writing of the manuscript: Tateishi,
Ohe and Mashino.
ꢀ
[14] Fang JL, Loukotkova L, Chitranshi P, Gamboa da Costa G, Beland FA. Effects of
human sulfotransferases on the cytotoxicity of 12-hydroxynevirapine. Bio-
All authors provided final approval of the manuscript.
chem
Pharmacol
2018;155:455e67.
https://doi.org/10.1016/
j.bcp.2018.07.016.
[15] Sharma AM, Li Y, Novalen M, Hayes MA, Uetrecht JP. Bioactivation of nevi-
rapine to a reactive quinone methide: implications for liver injury. Chem Res
Toxicol 2012;25:1708e19. https://doi.org/10.1021/tx300172s.
Declaration of competing interest
[16] Harjivan SG, Pinheiro PF, Martins IL, Godinho AL, Wanke R, Santos PP, et al.
Quinoid derivatives of the nevirapine metabolites 2-hydroxy- and 3-hydroxy-
nevirapine: activation pathway to amino acid adducts. Toxicol Res (Camb)
2015;4:1565e77. https://doi.org/10.1039/c5tx00176e.
[17] Satoh D, Iwado S, Abe S, Kazuki K, Wakuri S, Oshimura M, et al. Establishment
of a novel hepatocyte model that expresses four cytochrome P450 genes
stably via mammalian-derived artificial chromosome for pharmacokinetics
and toxicity studies. PloS One 2017;12. https://doi.org/10.1371/
journal.pone.0187072.
The authors declare no conflict of interest.
Acknowledgement
This research was supported by Platform Project for Supporting
Drug Discovery and Life Science Research (Basis for Supporting
Innovative Drug Discovery and Life Science Research (BINDS)) from
AMED under Grant Number JP19am0101089 (M.T.) and JP
19am0101124 (Y.K.). This work was also supported by JSPS
KAKENHI (Grant Number 16K08379) and a special grant generously
provided by the Hoansha Foundation.
[18] Wen B, Chen Y, Fitch WL. Metabolic activation of nevirapine in human liver
Microsomes : dehydrogenation and inactivation of cytochrome P450 3A4.
Drug
Metab
Dispos
2009;37:1557e62.
https://doi.org/10.1124/
dmd.108.024851.
[19] Hargrave KD, Proudfoot JR, Grozinger KG, Cullen E, Kapadia SR, Patel UR, et al.
Novel non-nucleoside inhibitors of HIV-1 reverse transcriptase. 1. Tricyclic
pyridobenzo-and dipyridodiazepinones.
https://doi.org/10.1021/jm00111a045.
J Med Chem 1991;34:2231e41.
[20] Dekker SJ, Zhang Y, Vos JC, Vermeulen NPE, Commandeur JNM. Different
reactive metabolites of nevirapine require distinct glutathione S -transferase
isoforms for bioinactivation. Chem Res Toxicol 2016;29:2136e44. https://
doi.org/10.1021/acs.chemrestox.6b00250.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.dmpk.2020.01.006.