Functional characterization of 12 allelic variants of CYP2C8 by assessment of paclitaxel 6α-hydroxylation and amodiaquine N-deethylation

Article abstract of DOI:10.1016/j.dmpk.2015.07.003

Cytochrome P450 2C8 (CYP2C8) is one of the enzymes primarily responsible for the metabolism of many drugs, including paclitaxel and amodiaquine. CYP2C8 genetic variants contribute to interindividual variations in the therapeutic efficacy and toxicity of p

Full text of DOI:10.1016/j.dmpk.2015.07.003

Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
Contents lists available at ScienceDirect
Drug Metabolism and Pharmacokinetics
journal homepage: http://www.journals.elsevier.com/drug-metabolism-and-
pharmacokinetics
Regular article
Functional characterization of 12 allelic variants of CYP2C8 by assessment
of paclitaxel 6 -hydroxylation and amodiaquine N-deethylation
a
Chiharu Tsukada a, Takahiro Saito a, Masamitsu Maekawa b, Nariyasu Mano b, Akifumi Oda c,
,
*
Noriyasu Hirasawa a, Masahiro Hiratsuka a
a Laboratory of Pharmacotherapy of Life-Style Related Diseases, Graduate School of Pharmaceutical Sciences, Tohoku University, Sendai 980-8578, Japan
b Department of Pharmacy, Tohoku University Hospital, Sendai 980-8574, Japan
c Institute of Medical, Pharmaceutical and Health Sciences, Kanazawa University, Kanazawa 920-1192, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Cytochrome P450 2C8 (CYP2C8) is one of the enzymes primarily responsible for the metabolism of many
drugs, including paclitaxel and amodiaquine. CYP2C8 genetic variants contribute to interindividual
variations in the therapeutic efficacy and toxicity of paclitaxel. Although it is difficult to investigate the
enzymatic function of most CYP2C8 variants in vivo, this can be investigated in vitro using recombinant
CYP2C8 protein variants. The present study used paclitaxel to evaluate 6a-hydroxylase activity and
amodiaquine for the N-deethylase activity of wild-type and 11 CYP2C8 variants resulting in amino acid
substitutions in vitro. The wild-type and variant CYP2C8 proteins were heterologously expressed in COS-
Received 14 May 2015
Received in revised form
14 July 2015
Accepted 21 July 2015
Available online 31 July 2015
Keywords:
Cytochrome P450
CYP2C8
7 cells. Paclitaxel 6
measuring the concentrations of 6
a
-hydroxylation and amodiaquine N-deethylation activities were determined by
-hydroxypaclitaxel and N-desethylamodiaquine, respectively, and the
a
kinetic parameters were calculated. Compared to the wild-type enzyme (CYP2C8.1), CYP2C8.11 and
CYP2C8.14 showed little or no activity with either substrate. In addition, the intrinsic clearance values of
CYP2C8.8 and CYP2C8.13 for paclitaxel were 68% and 67% that of CYP2C8.1, respectively. In contrast, the
CLint values of CYP2C8.2 and CYP2C8.12 were 1.4 and 1.9 times higher than that of CYP2C8.1. These
comprehensive findings could inform for further genotype-phenotype studies on interindividual dif-
ferences in CYP2C8-mediated drug metabolism.
Genetic polymorphisms
Paclitaxel
Amodiaquine
Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Paclitaxel is useful for the treatment of breast, lung, and ovarian
cancers. However, inter-patient variability in its toxicity (e.g., pe-
Cytochrome P450 2C8 (CYP2C8) is a major metabolic enzyme
that contributes to the metabolism of at least 5% of clinical drugs,
including the anticancer drug, paclitaxel, and the antimalarial drug,
amodiaquine [1]. Interindividual variability in cytochrome P450
activity significantly influences the metabolism of these drugs,
resulting in alterations in efficacy and adverse effects. Recently,
polymorphisms in CYP2C8 have been reported to contribute to al-
ripheral neuropathy and neutropenia) and therapeutic efficacy
limit its use in chemotherapy [7e9]. Paclitaxel is primarily
metabolized by CYP2C8, with secondary metabolism by CYP3A4/5
(Fig. 1). In a previous cohort study, Hertz et al. reported that breast
cancer patients carrying the CYP2C8*3 variant achieved a more
complete clinical response to paclitaxel than did patients carrying
the wild-type allele [10]. The odds ratio for an individual with the
CYP2C8*3 allele to achieve complete clinical response was 3.92.
Subjects with CYP2C8*3 were also more susceptible to severe pe-
ripheral neuropathy caused by paclitaxel, showing an odds ratio
was 3.13. A study by Green et al. stratified thirty-eight ovarian
cancer patients (treated with paclitaxel and carboplatin) by their
ABCB1 G2677 T/A genotype, and found that those who were het-
erozygous for CYP2C8*3 showed reduced clearance of paclitaxel
[11]. Bergmann et al. reported that carriers of CYP2C8*3 had 11%
lower clearance than non-carriers [12]. In addition to CYP2C8*3, a
terations in enzymatic activity [2]. CYP2C8*3 (2130G
> A,
30411A > G; R139K, K399R), which is one of the major allelic var-
iants of CYP2C8, has an allele frequency of 0.13 among Caucasians
[3]. CYP2C8.3, the protein encoded by CYP2C8*3, has been reported
to show substrate-dependent alterations in activity [4e6].
* Corresponding author.
E-mail address: mhira@m.tohoku.ac.jp (M. Hiratsuka).
http://dx.doi.org/10.1016/j.dmpk.2015.07.003
1347-4367/Copyright © 2015, The Japanese Society for the Study of Xenobiotics. Published by Elsevier Ltd. All rights reserved.

C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
367
Fig. 1. Metabolic pathways for paclitaxel showing the primary route, catalyzed by cytochrome P450 2C8 (CYP2C8) and secondary routes, catalyzed by CYP3A4/5.
large study by Abraham et al. found that CYP2C8*4 was associated
with an increased risk for paclitaxel-related peripheral neuropathy
[13]. An association between CYP2C8 genetic polymorphisms and
changes in paclitaxel pharmacodynamics and pharmacokinetics
has been found in several previous studies [14e16]. However, it is
difficult to investigate the enzymatic function of most of the
CYP2C8 allelic variants in vivo. The association between these var-
iants and patient responses to paclitaxel, or the incidence of severe
adverse effects, is also difficult to determine because of their low
frequency in the population (below 0.01) [17]. Therefore, most of
their enzymatic functions in vivo remain unclear.
To assess the functional differences between the enzymatic ac-
tivities of CYP2C8 variants in vitro, previous studies have used a
range of heterologous expression systems, including bacteria, yeast,
insect cells, and mammalian cells [3,4,6,18e27]. Using these sys-
tems, several CYP2C8 variants were shown to exhibit increased or
decreased enzymatic activity for CYP2C8 substrates. However,
different kinetic values have been reported in the different
expression systems used. Therefore, the present study expressed
CYP2C8 variants under the same conditions in mammalian cells
and compared their catalytic activities.
from Toronto Research Chemicals (North York, Ontario, Canada).
Oxidized -nicotinamide-adenine dinucleotide phosphate
(NADPþ), glucose-6-phosphate (G6P), glucose-6-phosphate dehy-
drogenase (G6PDH), and -nicotinamide-adenine dinucleotide
phosphate (NADPH) were obtained from Oriental Yeast (Tokyo,
Japan). The polyclonal anti-human CYP2C8 antibody was pur-
chased from Nosan Corporation (Kanagawa, Japan). The polyclonal
anti-calnexin antibody was purchased from Enzo Life Sciences
(Farmingdale, NY, USA). Polyclonal goat anti-rabbit immunoglob-
ulin/horseradish peroxidase was purchased from DakoCytomation
(Glostrup, Denmark). COS-7 cells were obtained from Riken Cell
Bank (Ibaraki, Japan). All other chemicals and reagents were of the
highest quality or analytical quality commercially available.
b
b
2.2. CYP2C8 cDNA cloning and construction of expression vectors
CYP2C8 cDNA fragments, obtained from a human liver cDNA
library (TaKaRa, Shiga, Japan), were amplified by polymerase chain
reaction with the forward primer 50-CACCATGGAACCTTTTGT
GGTCCTG-30 and the reverse primer 50-TCAGACAGGGATGAAGCA-
GATCTGGTATG-30 using PfuUltra High-Fidelity DNA polymerase
(Agilent Technologies, Santa Clara, CA, USA). The underlined
sequence in the forward primer was introduced for directional
TOPO cloning. Amplified fragments were subcloned into the
pENTR/D-TOPO vector (Life Technologies, Carlsbad, CA, USA).
Plasmids containing CYP2C8*1 cDNA were used as a template to
generate 11 CYP2C8 allelic variant constructs (CYP2C8*2-CYP2C8*4,
CYP2C8*6, CYP2C8*8-CYP2C8*14) using a QuikChange Site-directed
Mutagenesis Kit (Stratagene, La Jolla, CA, USA), according to the
manufacturer's instructions. All prepared constructs were
confirmed by direct sequencing. Wild type and variant CYP2C8
cDNA sequences were subsequently subcloned into the mammalian
expression vector, pcDNA3.4 (Life Technologies), by TA cloning.
To date, 16 variant alleles of CYP2C8 have been designated
(http://www.cypalleles.ki.se/cyp2c8.htm).
A
previous study
showed that CYP2C8*5 and CYP2C8*7 variant alleles produced
incomplete proteins that were not enzymatically active [24,28].
This lack of activity was due to a frame shift (for CYP2C8*5) and a
nonsense mutation (for CYP2C8*7). In the present study, we focused
on CYP2C8*1-CYP2C8*14 (Table 1) and evaluated their function after
their expression in COS-7 cells. The use of two CYP2C8 substrates,
paclitaxel and amodiaquine, enabled us to examine whether any
functional differences were substrate-specific.
2. Materials and methods
2.1. Chemicals
2.3. Expression of CYP2C8 variant proteins in COS-7 cells
Paclitaxel was purchased from Wako Pure Chemical Industries
(Osaka, Japan). Docetaxel was purchased from SigmaeAldrich (St
Louis, MO, USA). 6a-Hydroxypaclitaxel was obtained from Calbio-
COS-7 cells were cultured in Dulbecco's modified Eagle's me-
dium (Nacalai Tesque, Kyoto, Japan) containing 10% fetal bovine
serum at 37 ꢀC in the presence of 5% CO2. Cells were transfected
chem (San Diego, CA, USA). Amodiaquine, N-desethylamodiaquine
hydrochloride, and N-desethylamodiaquine-d5 were purchased
with plasmids (7
reagent (Bio-Rad Laboratories, Hercules, CA, USA), according to the
mg) carrying CYP2C8 cDNA using TransFectin lipid

368
C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
Table 1
CYP2C8 allelic variants characterized in this study.
Allele
Protein
dbSNP numbera
Nucleotide changes
Amino acid changes
CYP2C8*1
CYP2C8*2
CYP2C8*3
CYP2C8*4
CYP2C8*6
CYP2C8*8
CYP2C8*9
CYP2C8*10
CYP2C8*11
CYP2C8*12
CYP2C8*13
CYP2C8*14
CYP2C8.1
CYP2C8.2
CYP2C8.3
CYP2C8.4
CYP2C8.6
CYP2C8.8
CYP2C8.9
CYP2C8.10
CYP2C8.11
CYP2C8.12
CYP2C8.13
CYP2C8.14
rs11572103
rs11572080; rs10509681
rs1058930
rs142886225
rs72558195
e
11054A > T
2130G > A; 30411A > G
11041C > G
4472G > A
4517C > T
10989A > G
26513G > T
23452G > T
32184_32186delTTG
10918T > G
I269F
R139K; K399R
I264M
G171S
R186G
K247R
K383N
E274X
461delV
I223M
A238P
e
rs78637576
rs3832694
e
rs188934928
10961G > C
CYP2C8, cytochrome P450 2C8; SNP, single nucleotide polymorphism.
a
RefSNP accession number in dbSNP (http://www.ncbi.nlm.nih.gov/snp/).
manufacturer's instructions. After incubation for 24 h at 37 ꢀC, cells
were scraped off, centrifuged at 1500 g for 5 min, and re-suspended
in homogenization buffer (10 mM TriseHCl [pH 7.4], 1 mM ethyl-
enediaminetetraacetic acid [EDTA], and 10% glycerol). Microsomal
fractions were prepared by differential centrifugation at 9000 g for
20 min, followed by centrifugation of the resulting supernatant at
105 000 g for 60 min. The microsomal pellet was re-suspended in
50 mM TriseHCl (pH 7.4) containing 1 mM EDTA, 20% glycerol,
150 mM KCl, and Protease Inhibitor Cocktail Set III (Calbiochem)
and stored at ꢁ80 ꢀC. The protein concentration was determined
using a BCA protein assay kit (Thermo Fisher Scientific, Waltham,
MA, USA).
20 min, 6
of up to 50
After protein removal by centrifugation (10 000 ꢂ g for 3 min),
L of the supernatant was subjected to high-performance liquid
chromatography (HPLC). The HPLC system consisted of a Waters
2695 Separations Module, a Waters 2487 dual absorbance de-
tector (Waters, Milford, MA, USA), and an XBridge C18 analytical
column (4.6 ꢂ 150 mm, 5 m particle size; Waters), maintained at
40 ꢀC. 6
-Hydroxypaclitaxel was eluted isocratically with water
and acetonitrile (60:40, v/v) at a flow rate of 1.0 mL/min. Detection
of 6 -hydroxypaclitaxel was performed by measuring absorbance
at 230 nm. The lower limit of 6 -hydroxypaclitaxel quantification
was 10 nM. Standard curves for 6 -hydroxypaclitaxel were con-
structed in the 0.01e4 M range using authentic metabolite. The
coefficient of variation was 6.3% for a 6 -hydroxypaclitaxel con-
centration of 0.8 M, which was observed at substrate concentra-
a
-hydroxypaclitaxel formation was linear in the presence
mg microsomal protein (data not shown).
50
m
l
m
a
a
a
a
2.4. Determination of protein expression levels by immunoblotting
m
a
COS-7 microsomal fractions (2
mg microsomal protein) were
m
separated by electrophoresis on 10% sodium dodecyl sulfate-
polyacrylamide gels. Western blotting was performed according
to standard procedures. Recombinant CYP2C8 Supersomes (BD
Biosciences, Woburn, MA, USA) were used as the standard (range,
0.02e0.25 pmol) in each gel to quantify the CYP2C8 protein level.
CYP2C8 protein was detected using a polyclonal anti-human
CYP2C8 antibody (diluted at 1:1000), and calnexin was detected
by a polyclonal anti-calnexin antibody (diluted at 1:5000). Sec-
ondary detection was carried out with horseradish peroxidase-
conjugated goat anti-rabbit IgG (diluted at 1:10 000). Immuno-
blots were visualized using SuperSignal West Femto Maximum
Sensitivity Substrate (Thermo Fisher Scientific). Chem-
iluminescence was quantified using a ChemiDoc XRSþ, with the
help of Image Lab Software (Bio-Rad Laboratories).
tions in the range of Michaelis constant (Km). The enzymatic activity
was normalized to the CYP2C8 expression level.
2.6. Amodiaquine N-deethylation assay
The amodiaquine N-deethylation activity of CYP2C8 was deter-
mined using previously reported methods, with minor modifica-
tions [29]. The reaction mixture contained the microsomal fraction
(15
and 50 mM potassium phosphate buffer (pH 7.4) in a total volume
of 150
L. Following pre-incubation at 37 ꢀC for 3 min, reactions
mg), amodiaquine (0.25, 0.5, 1.0, 2.5, 5.0, 10, 20, 40, or 80 mM),
m
were initiated by the addition of 1 mM NADPH. The mixtures were
incubated for an additional 20 min at 37 ꢀC. Reactions were
terminated by adding 150 mL acetonitrile containing 300 nM N-
desethylamodiaquine-d5 (the internal standard). Determination of
2.5. Paclitaxel 6
a
-hydroxylation assay
amodiaquine N-deethylation activity in the microsomal fraction
containing CYP2C8.1 (20
mg of microsomal protein) and 40 mM
Paclitaxel 6
ported previously, with several modifications [18]. The incubation
mixture consisted of the microsomal fraction (50 g), paclitaxel
(0.25, 0.5, 1.0, 2.5, 5.0, 10, 20, 40, or 80 M), and 50 mM potassium
phosphate buffer (pH 7.4) in a total volume of 150 L. Following
a
-hydroxylation by CYP2C8 was measured as re-
amodiaquine revealed that N-desethylamodiaquine formation was
linear for incubations of up to 20 min. When the reaction con-
taining 40 mM amodiaquine was incubated for 20 min, N-desethy-
m
m
lamodiaquine formation was linear in the presence of up to 15
microsomal protein (data not shown).
mg of
m
pre-incubation at 37 ꢀC for 3 min, reactions were initiated by
addition of the NADPH-generating system (1.3 mM NADPþ,
3.3 mM G6P, 3.3 mM magnesium chloride, and 0.4 units/mL
G6PDH). The mixture was incubated at 37 ꢀC for 20 min. Reactions
After the removal of protein by centrifugation (15 000 ꢂ g for
10 min), the supernatant was diluted five times with water, and
10 mL of the diluted solution was injected into a liquid chromatog-
raphy tandem mass spectrometry (LC-MS/MS) system. N-Desethy-
lamodiaquine was measured using the LC-MS/MS system in
the positive ion detection mode at the electrospray ionization
interface (TSQ Vantage Triple Stage Quadrupole Mass Spectrometer;
Thermo Fisher Scientific). Separation by ultra HPLC was conducted
using a Nexera Ultra High Performance Liquid Chromatography
system (SHIMADZU, Kyoto, Japan). Chromatographic separationwas
were terminated by adding 150
docetaxel (the internal standard). Measurement of paclitaxel 6
hydroxylation using 40 M paclitaxel and 50 g of the microsomal
fraction containing CYP2C8.1 showed that 6 -hydroxypaclitaxel
formation was linear for incubations of up to 20 min. Moreover,
when the reaction containing 40 M paclitaxel was incubated for
mL of methanol containing 2 mM
a
-
m
m
a
m

C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
369
performed using an XBridge C18 analytical column (2.1 ꢂ 150 mm,
3.5
m particle size; Waters), maintained at 50 ꢀC. N-Desethyla-
modiaquine was eluted isocratically with a mobile phase consisting
of 5 mM ammonium acetate and 0.1% (v/v) formic acid in water and
5 mM ammonium acetate and 0.1% (v/v) formic acid in acetonitrile
value for CYP2C8.8 was reduced to 25% of that for wild-type.
Compared to the Vmax value for CYP2C8.1, the Vmax values for
CYP2C8.9 and CYP2C8.12 were significantly increased (by 1.2- and
1.3-fold, respectively). The Vmax values of CYP2C8.3, CYP2C8.4,
CYP2C8.8, CYP2C8.10, CYP2C8.13, and CYP2C8.14 were significantly
decreased. The CLint values for CYP2C8.2 and CYP2C8.12 were
significantly higher, and those of CYP2C8.3, CYP2C8.8, CYP2C8.13,
and CYP2C8.14 were significantly lower than that of CYP2C8.1.
The kinetics of amodiaquine N-deethylation were determined
for nine CYP2C8 variants, as shown in Table 3. The Km, Vmax, and
m
(90:10, v/v) at a flow rate of 200 mL/min.
Quantitative MS/MS analyses were performed in the selective
reaction monitoring mode. The areas under the peak of the m/z
328 / 283 (collision energy, 17 V and S-Lens RF amplitude voltage,
67 V) for N-desethylamodiaquine and that of the m/z 333 / 283
(collision energy, 15 V and S-Lens RF amplitude voltage, 66 V) for N-
desethylamodiaquine-d5 were measured. The optimized parame-
ters for MS were as follows: spray voltage, 3.5 kV; sheath gas
pressure, 50 psi; vaporizer temperature, 400 ꢀC; capillary temper-
ature, 392 ꢀC; and collision pressure, 1.4 mTorr. The sheath gas was
nitrogen, and the collision gas was argon. The LC-MS/MS system
was controlled by Xcalibur software (Thermo Fisher Scientific),
which was also used to analyze the data. The lower limit of N-
desethylamodiaquine quantification was 5 nM. Standard curves for
N-desethylamodiaquine were constructed in the 5e2560 nM range
using authentic metabolite. The coefficient of variation was 2.4% for
an N-desethylamodiaquine concentration of 160 nM, which was
observed at substrate concentrations in the range of Km. The
enzymatic activity was normalized to the corresponding CYP2C8
expression level.
CLint for amodiaquine N-deethylation by CYP2C8.1 were 1.35
mM,
11.3 pmol minꢁ1 pmolꢁ1 CYP2C8, and 8.37 L minꢁ1 pmolꢁ1
m
CYP2C8, respectively. The kinetic parameters of CYP2C8.11 and
CYP2C8.14 could not be determined because the amount of the
product, N-desethylamodiaquine, produced by these variants was
below the limit of quantification at low substrate concentrations.
Three variants, CYP2C8.2, CYP2C8.4, and CYP2C8.6, showed Km
values that were 1.3-fold higher than that of CYP2C8.1. The Vmax for
CYP2C8.2 was also increased 1.4-fold over that of CYP2C8.1. The
Vmax values of CYP2C8.3, CYP2C8.6, CYP2C8.8, CYP2C8.9, CYP2C8.10,
CYP2C8.12, and CYP2C8.13 were significantly lower than that of
CYP2C8.1. The CLint values of six variants, CYP2C8.6, CYP2C8.8,
CYP2C8.9, CYP2C8.10, CYP2C8.12, and CYP2C8.13, were significantly
lower than the CLint for CYP2C8.1.
4. Discussion
2.7. Data analysis
CYP2C8 is a clinically important metabolic enzyme. It has been
recognized that CYP2C8 genetic polymorphisms are a possible
source of interindividual variability in the efficacy and adverse ef-
fects of drugs metabolized by this enzyme. The present study
revealed, for the first time, functional alterations in the activities of
11 CYP2C8 allelic variants expressed in COS-7 cells. The kinetic
The kinetic data were analyzed in the Enzyme Kinetics Module
of SigmaPlot 12.0 (Systat Software, Inc., Chicago, IL, USA), a curve-
fitting program based on nonlinear regression analysis, and the
Km, maximum velocity (Vmax), and intrinsic clearance (CLint ¼ Vmax
Km) values were determined. All values were expressed as the
mean standard deviation (SD) of experiments performed in
/
parameters for paclitaxel 6a-hydroxylation were determined for 10
triplicate. Statistical analyses of enzymatic activity and kinetic pa-
rameters used analysis of variance, Dunnett's test for CLint of
paclitaxel, and Dunnett's T3 test for the paclitaxel Km and Vmax and
for the amodiaquine CLint, Km, and Vmax (IBM SPSS Statistics, Inter-
national Business Machines, Armonk, NY, USA). Differences with P-
values less than 0.05 were considered statistically significant.
variants and those for amodiaquine N-deethylation were deter-
mined for 9 variants. The kinetics of wild-type and variant forms of
CYP2C8 determined in this study and in other studies are shown in
Supplemental Table 1.
Some studies have reported that CYP2C8.4, which includes an
I264M substitution, showed altered enzymatic activity
(Supplemental Table 1). Although previous in vitro studies reported
reduced paclitaxel clearance by CYP2C8.4, as compared with the
wild-type enzyme, our study did not identify any statistically sig-
nificant difference in CLint. This inconsistency may reflect the
different experimental systems employed (Supplemental Table 1).
It has been reported that the I264M substitution in CYP2C8.4
decreased the ratio of holoprotein to apoprotein and made
CYP2C8.4 less thermally stable than CYP2C8.1 [22]. Moreover,
2.8. 3D structural modeling of CYP2C8
CYP2C8 3D structural modeling was based on the CYP2C8 X-ray
structure reported by Schoch et al., Protein Data Bank code: 1PQ2
[30]. The CDOCKER protocol of Discovery Studio 2.5 (Accelrys, San
Diego, CA, USA) was used to create docked CYP2C8-paclitaxel and
CYP2C8-amodiaquine structures.
3. Results
Protein levels of the CYP2C8 variants expressed in COS-7 cells
were measured by immunoblotting. The polyclonal CYP2C8 anti-
bodies recognized all CYP2C8 variant proteins (Fig. 2). Expression
levels of the endoplasmic reticulum-resident protein, calnexin,
were almost constant in microsomes from the transfected cells.
CYP2C8 protein was not detected in the cells transfected with
empty vector (mock transfection).
The kinetic parameters of paclitaxel 6
determined for 10 CYP2C8 variants, as shown in Table 2. For
CYP2C8.11, 6 -hydroxypaclitaxel was not detected, even at the
highest concentration of paclitaxel tested (80 M). Km, Vmax, and
CLint values for paclitaxel 6 -hydroxylation by the wild-type
enzyme (CYP2C8.1) were 7.18
M, 2.18 pmol minꢁ1 pmolꢁ1
CYP2C8, and 0.31
L minꢁ1 pmolꢁ1 CYP2C8, respectively. The Km
a-hydroxylation were
Fig. 2. Western blots showing immunoreactive cytochrome P450 2C8 (CYP2C8)
variant proteins (upper panel) and calnexin (lower panel). Following electrophoresis
on 10% sodium dodecyl sulfate-polyacrylamide gels and protein transfer, CYP2C8
variant proteins and calnexin (loading control) were detected using polyclonal anti-
bodies. The lane numbers in the upper panel correspond to the CYP2C8 variant
expressed in the sample (see Table 1).
a
m
a
m
m

370
C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
Table 2
Kinetic parameters for paclitaxel 6
a
-hydroxylation by microsomes of COS-7 cells expressing wild-type (CYP2C8.1) and CYP2C8 allelic variants.
Variant
Km
(
mM)
Vmax (pmol minꢁ1 pmolꢁ1 CYP2C8)
CLint (m
L minꢁ1 pmolꢁ1 CYP2C8) (% of CYP2C8.1)
CYP2C8.1
CYP2C8.2
CYP2C8.3
CYP2C8.4
CYP2C8.6
CYP2C8.8
CYP2C8.9
CYP2C8.10
CYP2C8.12
CYP2C8.13
CYP2C8.14
7.18 0.62
8.37 1.87
6.02 0.53
5.15 0.85
6.54 0.55
1.78 0.40**
6.90 0.30
5.05 0.27
4.88 0.46
5.25 0.93
17.7 3.15***
2.18 0.06
3.60 0.41
0.30 0.02
0.44 0.05*** (143%)
0.21 0.02*** (68%)
0.25 0.03 (81%)
0.29 0.014 (94%)
0.20 0.03*** (65%)
0.38 0.01* (123%)
0.28 0.003 (91%)
0.59 0.04* (192%)
0.20 0.02*** (67%)
0.04 0.002*** (13%)
1.25 0.02***
1.26 0.04***
1.86 0.07
0.34 0.02***
2.59 0.03*
1.40 0.02***
2.85 0.13**
1.06 0.08***
0.71 0.09***
CYP2C8, cytochrome P450 2C8; Km, Michaelis constant; Vmax, maximum velocity; CLint, intrinsic clearance (Vmax/Km). These data represent the mean SD of three independent
catalytic assays. *P < 0.05, **P < 0.01, ***P < 0.005, as compared to CYP2C8.1. The kinetics of CYP2C8.11 could not be determined because its enzymatic activity was not detected
at a high substrate concentration (80 mM paclitaxel).
Table 3
Kinetic parameters for amodiaquine N-deethylation by microsomes of COS-7 cells expressing wild-type (CYP2C8.1) and CYP2C8 allelic variants.
Variants
Km
(m
M)
Vmax (pmol minꢁ1 pmolꢁ1 CYP2C8)
CLint (m
L minꢁ1 pmolꢁ1 CYP2C8) (% of CYP2C8.1)
CYP2C8.1
CYP2C8.2
CYP2C8.3
CYP2C8.4
CYP2C8.6
CYP2C8.8
CYP2C8.9
CYP2C8.10
CYP2C8.12
CYP2C8.13
1.35 0.036
1.76 0.037***
1.19 0.095
1.70 0.022***
1.78 0.053**
1.31 0.062
1.28 0.092
1.18 0.063*
1.13 0.093***
1.48 0.032
11.30 0.15
8.37 0.29
16.13 0.23***
8.41 0.11***
11.28 0.13
5.81 0.03***
3.45 0.10***
5.49 0.09***
3.88 0.04***
4.40 0.17***
2.69 0.01***
9.18 0.07 (110%)
7.11 0.48 (85%)
6.62 0.03 (79%)
3.26 0.09*** (39%)
2.65 0.05*** (32%)
4.31 0.25*** (52%)
3.29 0.15*** (39%)
3.92 0.49** (47%)
1.82 0.04*** (22%)
CYP2C8, cytochrome P450 2C8; Km, Michaelis constant; Vmax, maximum velocity; CLint, intrinsic clearance (Vmax/Km). These data represent the mean SD of three independent
catalytic assays. *P < 0.05, **P < 0.01, ***P < 0.005, as compared to CYP2C8.1. The kinetic parameters of CYP2C8.11 and CYP2C8.14 could not be determined because the amount
of metabolite produced was below the detection limit at the lower substrate concentrations.
CYP2C8.4 has been shown to be more susceptible to proteasomal
degradation than the wild-type enzyme, resulting in decreased
protein levels and correspondingly lower enzyme activity [23].
CYP2C8.4 may therefore be less stable than CYP2C8.1 and further
investigation of its stability and expression level is required.
The CYP2C8*2 allele contains the nucleotide change 11054A > T,
which produces the I269F substitution. This is one of the most
common variants of CYP2C8. Previous studies using other expres-
sion systems reported that the recombinant CYP2C8.2 enzyme had
lower CLint values for paclitaxel and amodiaquine metabolism, as
compared with CYP2C8.1 [3,4,18,19]. However, the data presented
in this study indicated that CYP2C8.2 showed a significantly higher
CLint for paclitaxel than did CYP2C8.1. Inconsistencies between the
present and previous studies in terms of paclitaxel metabolism in
transient expression systems may be due to differences in the
experimental systems employed (Escherichia coli, yeast, and insect
cells). Additionally, the I269F substitution is located in the H-helix
on the surface of the CYP2C8 protein; this is close to the I264M
substitution found in CYP2C8.4. Thus, CYP2C8.2 may also be less
stable, resulting in decreased activity in vivo; however, in vitro,
CYP2C8.2 activity was increased. Further investigation is needed to
determine how the I269F substitution in CYP2C8.2 affects the
pharmacokinetics of CYP2C8 substrates in vivo. Aquilante et al. re-
ported that the mean area under the concentration-time (0e48 h)
curve (AUC0-48) of the ratio of two pioglitazone metabolites (M-
III:MꢁIV) was 33% lower in CYP2C8*2 carriers than in wild-type
carriers; however, the maximum plasma concentration and AUC
of pioglitazone did not differ significantly between these in-
dividuals [31]. Hertz et al. reported an association between the
CYP2C8 genotype and paclitaxel-induced peripheral neuropathy
(PIPN) [16]. In this study, CYP2C8*2, CYP2C8*3, and CYP2C8*4 car-
riers were categorized as low paclitaxel metabolizers, but neither
CYP2C8*2 nor CYP2C8*4 carriage showed a significant association
with PIPN. In the present study, the data indicated that CYP2C8*2
may not reduce drug metabolism. Therefore, the impact of
CYP2C8*2 on the pharmacokinetics of drugs metabolized by CYP2C8
remains unclear. Further studies will be required to determine the
impact of this genotype on drug metabolism in clinical settings.
CYP2C8.8, which contains an R186G substitution, showed a
lower level of expressed protein than did CYP2C8.1 and substan-
tially decreased activities for both substrates; this was consistent
with the results of a previous study [24]. Additionally, it has been
reported that the CYP2C8.8 holoprotein level expressed in insect
cells was significantly lower than that of CYP2C8.1 [24]. R186 in
CYP2C8.1 is located in the E-F loop, forming an ionic bond with
E148 in the D-helix. The R186G substitution in CYP2C8.8 may cause
a loss of this ionic bonding and may lead to hydrogen bonding with
E147 instead (Fig. 3). This conformational alteration may reduce the
affinity for heme and alter the ratio of holoprotein forms, making
CYP2C8.8 more susceptible to proteasomal degradation.
CYP2C8*11 contains the nucleotide change 23452G > T, which
produces the E274X substitution. Consequently, CYP2C8.11 lacks
approximately 40% of the C-terminal region, including the heme
binding site. The present study found that CYP2C8.11 was inactive for
both substrates. Yeo et al. reported that the CYP2C8*11 genotype
causedadecreaseinrosiglitazonehydroxylationinvivo [32]. Subjects
with CYP2C8*1/*11 heterozygosity showed a 31% decrease in rosi-
glitazone hydroxylation, as compared with CYP2C8*1 homozygotes.
These findings indicated that patients with CYP2C8*11 would show
decreased paclitaxel 6a-hydroxylation, resulting in more severe
adverse effects, including peripheral neuropathy and neutropenia.
The present study revealed previously uncharacterized func-
tional alterations in CYP2C8.12 activity. This variant showed a
markedly increased Vmax for paclitaxel and a significantly decreased

C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
371
Vmax for amodiaquine, as compared to the wild-type enzyme. Thus,
CYP2C8.12 exhibited substrate specificity in its activity. The dele-
tion of three nucleotides, TTG, at residues 32184e32186 in
CYP2C8*12, results in the deletion of V461 (461delV). This allele is
found at a very low frequency in the Japanese population (0.1%)
[17]. V461 is located in the C-terminal region of CYP2C8.1. The next
residue, D462, may form a hydrogen bond with N466. In CYP2C8.12,
461delV caused a shift to D461, which may form a hydrogen bond
with N465. Furthermore, D462 may also form a hydrogen bond
with N465 (Fig. 4A). Thereby, conformational changes in the active
sites, including the important residue I476, which is I475 in
CYP2C8.12, may occur (Fig. 4B). The differential activities for
paclitaxel and amodiaquine may relate to structural and size dif-
ferences between paclitaxel (molecular weight, 854) and amodia-
quine (molecular weight, 356), and to the size of the active sites.
The Vmax of CYP2C8.13 was significantly decreased for both
substrates. This variant contains the I223M substitution in the F0-
helix. The M side chain is bulkier than that of I and the I223M
substitution may therefore result in steric hindrance and improper
protein folding. Afzelius et al. reported that the conformational
flexibility in the F-G loop in CYP2C5 and CYP2C9 could be of great
importance for substrate recognition, binding, and the volume of
the active site [33]. Therefore, CYP2C8.13 may reduce the flexibility
of the F-G loop, resulting in decreased Vmax and CLint
.
Fig. 3. A diagram of a portion of the crystal structure of cytochrome P450 2C8 (CYP2C8) showing the R186G substitution in CYP2C8.8. Paclitaxel is shown in pink, heme is shown in
orange. Ionic bonding differs between the wild-type enzyme and the CYP2C8.8 variant.
Fig. 4. (A) A diagram of a portion of the crystal structure of cytochrome P450 2C8 (CYP2C8), showing V461 and D461 (caused by the V deletion). V461 in the wild-type enzyme and
D461 in CYP2C8.12 are shown in yellow. Hydrogen bonding differs between the wild-type enzyme and the CYP2C8.12 variant. (B) A diagram of the overall crystal structure of
CYP2C8. Paclitaxel is shown in pink, heme is shown in orange, and the I476 residue in the wild-type enzyme and I475 residue in CYP2C8.12 are shown in blue. The size of the active
site differs between the wild-type enzyme and the CYP2C8.12 variant.

372
C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
[2] Aquilante CL, Niemi M, Gong L, Altman RB, Klein TE. PharmGKB summary:
CYP2C8.14 showed markedly decreased activity for both sub-
very important pharmacogene information for cytochrome P450, family 2,
subfamily C, polypeptide 8. Pharmacogenet Genomics 2013;23:721e8.
[3] Dai D, Zeldin DC, Blaisdell JA, Chanas B, Coulter SJ, Ghanayem BI, et al. Poly-
morphisms in human CYP2C8 decrease metabolism of the anticancer drug
paclitaxel and arachidonic acid. Pharmacogenetics 2001;11:597e607.
[4] Yu L, Shi D, Ma L, Zhou Q, Zeng S. Influence of CYP2C8 polymorphisms on the
hydroxylation metabolism of paclitaxel, repaglinide and ibuprofen enantio-
mers in vitro. Biopharm Drug Dispos 2013;34:278e87.
[5] Kaspera R, Naraharisetti SB, Evangelista EA, Marciante KD, Psaty BM, Totah RA.
Drug metabolism by CYP2C8.3 is determined by substrate dependent in-
teractions with cytochrome P450 reductase and cytochrome b5. Biochem
Pharmacol 2011;82:681e91.
strates. CYP2C8.14 has an A238P substitution in the G-helix. The
residue A238 is located in the substrate recognition site, which is
involved in substrate affinity [34]. Moreover, Hanioka et al. [26]
reported that yeast expression of CYP2C8.14 produced a signifi-
cantly lower holoprotein level than that observed for the wild-type
enzyme, CYP2C8.1. Therefore, the CYP2C8.14 holoprotein levels
may be decreased in comparison to those of CYP2C8.1, resulting in
reduced metabolism of CYP2C8 substrates.
A study by Bergmann et al. found that CYP2C8*3 carriers had a
significantly lower clearance of paclitaxel than non-carriers, which
was consistent with our in vitro data for CYP2C8*3. However, there
are some limitations to predicting in vivo activity or phenotype
from in vitro data. Expression levels and protein stability may be
considerably different in vivo. Some CYP2C8 variant proteins may
decrease heme incorporation into the CYP2C8 apoprotein, but this
possibility could not be assessed in the present study because we
used polyclonal anti-human CYP2C8 antibodies to determine
CYP2C8 protein levels. This could be investigated by studying the
difference in the carbon monoxide-reduced spectra and by
analyzing the protein bands using MS. Ohtsuki et al. have devel-
oped such an approach to the quantification of various CYP proteins
[35]. These quantitative analyses would also overcome the poten-
tial limitation of antibody affinity differences between the variant
proteins, and could determine the active form of CYP2C8 protein.
However, the levels of CYP2C8 protein expressed in COS-7 cells
were too low to analyze by the carbon monoxide-reduced spectra.
Furthermore, the relationship between in vivo and in vitro data
using other substrates are required to confirm these results, in
addition to the measurement of the holoprotein level. Therefore,
further studies will be required to determine the impact of these
CYP2C8 allelic variants on drug metabolism in clinical settings.
In conclusion, wild-type CYP2C8 and 11 allelic variants of this
enzyme were expressed in COS-7 cells and their enzymatic activ-
ities were characterized in vitro. This study confirmed the func-
tional alterations previously reported for some variants and also
characterized the activity of several other less common allelic
variants. These variants could contribute to significant interindi-
vidual variability in the pharmacokinetics and pharmacodynamics
of drugs that are CYP2C8 substrates, such as paclitaxel and amo-
diaquine. These comprehensive findings may provide useful in-
formation for further genotype-phenotype correlation studies
regarding interindividual differences in drug efficacy and toxicity
arising from differences in CYP2C8 activity.
[6] Kaspera R, Naraharisetti SB, Tamraz B, Sahele T, Cheesman MJ, Kwok PY, et al.
Cerivastatin in vitro metabolism by CYP2C8 variants found in patients expe-
riencing rhabdomyolysis. Pharmacogenet Genomics 2010;20:619e29.
[7] Steed H, Sawyer MB. Pharmacology, pharmacokinetics and pharmacoge-
nomics of paclitaxel. Pharmacogenomics 2007;8:803e15.
[8] Taniguchi R, Kumai T, Matsumoto N, Watanabe M, Kamio K, Suzuki S, et al.
Utilization of human liver microsomes to explain individual differences in
paclitaxel metabolism by CYP2C8 and CYP3A4. J Pharmacol Sci 2005;97:83e90.
[9] Sonnichsen DS, Liu Q, Schuetz EG, Schuetz JD, Pappo A, Relling MV. Variability
in human cytochrome P450 paclitaxel metabolism. J Pharmacol Exp Ther
1995;275:566e75.
[10] Hertz DL, Motsinger-Reif AA, Drobish A, Winham SJ, McLeod HL, Carey LA,
et al. CYP2C8*3 predicts benefit/risk profile in breast cancer patients receiving
neoadjuvant paclitaxel. Breast Cancer Res Treat 2012;134:401e10.
[11] Green H, Soderkvist P, Rosenberg P, Mirghani RA, Rymark P, Lundqvist EA,
et al. Pharmacogenetic studies of Paclitaxel in the treatment of ovarian cancer.
Basic Clin Pharmacol Toxicol 2009;104:130e7.
[12] Bergmann TK, Brasch-Andersen C, Green H, Mirza M, Pedersen RS, Nielsen F,
et al. Impact of CYP2C8*3 on paclitaxel clearance: a population pharmacoki-
netic and pharmacogenomic study in 93 patients with ovarian cancer. Phar-
macogenomics J 2011;11:113e20.
[13] Abraham JE, Guo Q, Dorling L, Tyrer J, Ingle S, Hardy R, et al. Replication of
genetic polymorphisms reported to be associated with Taxane-related sen-
sory neuropathy in early breast cancer patients treated with paclitaxel. Clin
Cancer Res 2014;20:2466e75.
[14] Leskela S, Jara C, Leandro-Garcia LJ, Martinez A, Garcia-Donas J, Hernando S,
et al. Polymorphisms in cytochromes P450 2C8 and 3A5 are associated with
paclitaxel neurotoxicity. Pharmacogenomics J 2011;11:121e9.
[15] Green H, Khan MS, Jakobsen-Falk I, Avall-Lundqvist E, Peterson C. Impact of
CYP3A5*3 and CYP2C8-HapC on paclitaxel/carboplatin-induced myelosup-
pression in patients with ovarian cancer. J Pharm Sci 2011;100:4205e9.
[16] Hertz DL, Roy S, Jack J, Motsinger-Reif AA, Drobish A, Clark LS, et al. Genetic
heterogeneity beyond CYP2C8*3 does not explain differential sensitivity to
paclitaxel-induced neuropathy. Breast Cancer Res Treat 2014;145:245e54.
[17] Saito Y, Katori N, Soyama A, Nakajima Y, Yoshitani T, Kim SR, et al. CYP2C8
haplotype structures and their influence on pharmacokinetics of paclitaxel in
a Japanese population. Pharmacogenet Genomics 2007;17:461e71.
[18] Gao Y, Liu D, Wang H, Zhu J, Chen C. Functional characterization of five
CYP2C8 variants and prediction of CYP2C8 genotype-dependent effects on
in vitro and in vivo drug-drug interactions. Xenobiotica 2010;40:467e75.
[19] Parikh S, Ouedraogo JB, Goldstein JA, Rosenthal PJ, Kroetz DL. Amodiaquine
metabolism is impaired by common polymorphisms in CYP2C8: implications
for malaria treatment in Africa. Clin Pharmacol Ther 2007;82:197e203.
[20] Soyama A, Saito Y, Hanioka N, Murayama N, Nakajima O, Katori N, et al. Non-
synonymous single nucleotide alterations found in the CYP2C8 gene result in
reduced in vitro paclitaxel metabolism. Biol Pharm Bull 2001;24:1427e30.
[21] Soyama A, Saito Y, Komamura K, Ueno K, Kamakura S, Ozawa S, et al. Five
novel single nucleotide polymorphisms in the CYP2C8 gene, one of which
induces a frame-shift. Drug Metab Pharmacokinet 2002;17:374e7.
[22] Jiang H, Zhong F, Sun L, Feng W, Huang ZX, Tan X. Structural and functional
insights into polymorphic enzymes of cytochrome P450 2C8. Amino Acids
2011;40:1195e204.
Acknowledgments
This study was supported in part by a grant from the Ministry of
Health, Labour and Welfare (MHLW) of Japan (‘Initiative to facilitate
development of innovative drug, medical devices, and cellular &
tissue-based product’), the Smoking Research Foundation, and the
Japan Research Foundation for Clinical Pharmacology. We thank the
Biomedical Research Core of Tohoku University Graduate School of
Medicine for technical support.
[23] Singh R, Ting JG, Pan Y, Teh LK, Ismail R, Ong CE. Functional role of Ile264 in
CYP2C8: mutations affect haem incorporation and catalytic activity. Drug
Metab Pharmacokinet 2008;23:165e74.
[24] Hichiya H, Tanaka-Kagawa T, Soyama A, Jinno H, Koyano S, Katori N, et al.
Functional characterization of five novel CYP2C8 variants, G171S, R186X,
R186G, K247R, and K383N, found in a Japanese population. Drug Metab
Dispos 2005;33:630e6.
[25] Hanioka N, Matsumoto K, Saito Y, Narimatsu S. Influence of CYP2C8*13 and
CYP2C8*14 alleles on amiodarone N-deethylation. Basic Clin Pharmacol Tox-
icol 2011;108:359e62.
Appendix A. Supplementary data
[26] Hanioka N, Matsumoto K, Saito Y, Narimatsu S. Functional characterization of
CYP2C8.13 and CYP2C8.14: catalytic activities toward paclitaxel. Basic Clin
Pharmacol Toxicol 2010;107:565e9.
[27] Soyama A, Hanioka N, Saito Y, Murayama N, Ando M, Ozawa S, et al. Amio-
darone N-deethylation by CYP2C8 and its variants, CYP2C8*3 and CYP2C8
P404A. Pharmacol Toxicol 2002;91:174e8.
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.dmpk.2015.07.003.
References
[28] Ishikawa C, Ozaki H, Nakajima T, Ishii T, Kanai S, Anjo S, et al.
A frameshift variant of CYP2C8 was identified in a patient who suffered
from rhabdomyolysis after administration of cerivastatin. J Hum Genet
2004;49:582e5.
[1] Lai XS, Yang LP, Li XT, Liu JP, Zhou ZW, Zhou SF. Human CYP2C8: structure,
substrate specificity, inhibitor selectivity, inducers and polymorphisms. Curr
Drug Metab 2009;10:1009e47.

C. Tsukada et al. / Drug Metabolism and Pharmacokinetics 30 (2015) 366e373
373
[29] Dravid PV, Frye RF. Determination of N-desethylamodiaquine by hydrophilic
interaction liquid chromatography with tandem mass spectrometry: appli-
cation to in vitro drug metabolism studies. J Chromatogr B Anal Technol
Biomed Life Sci 2008;863:129e34.
[30] Schoch GA, Yano JK, Wester MR, Griffin KJ, Stout CD, Johnson EF. Structure of
human microsomal cytochrome P450 2C8. Evidence for a peripheral fatty acid
binding site. J Biol Chem 2004;279:9497e503.
[31] Aquilante CL, Wempe MF, Spencer SH, Kosmiski LA, Predhomme JA,
Sidhom MS. Influence of CYP2C8*2 on the pharmacokinetics of pioglitazone
in Healthy African-American Volunteers. Pharmacotherapy 2013;33:
1000e7.
[32] Yeo CW, Lee SJ, Lee SS, Bae SK, Kim EY, Shon JH, et al. Discovery of a novel
allelic variant of CYP2C8, CYP2C8*11, in Asian populations and its clinical
effect on the rosiglitazone disposition in vivo. Drug Metab Dispos 2011;39:
711e6.
[33] Afzelius L, Raubacher F, Karlen A, Jorgensen FS, Andersson TB,
Masimirembwa CM, et al. Structural analysis of CYP2C9 and CYP2C5 and an
evaluation of commonly used molecular modeling techniques. Drug Metab
Dispos 2004;32:1218e29.
[34] Lewis DF. Homology modelling of human CYP2 family enzymes based on the
CYP2C5 crystal structure. Xenobiotica 2002;32:305e23.
[35] Ohtsuki S, Schaefer O, Kawakami H, Inoue T, Liehner S, Saito A, et al. Simul-
taneous absolute protein quantification of transporters, cytochromes P450,
and UDP-glucuronosyltransferases as a novel approach for the characteriza-
tion of individual human liver: comparison with mRNA levels and activities.
Drug Metab Dispos 2012;40:83e92.