In vitro modulatory effects of flavonoids on human cytochrome P450 2C8 (CYP2C8)

Article abstract of DOI:10.1007/s00210-012-0731-5

The inhibitory effects of five flavonoidswith distinct chemical classes (flavones [luteolin], flavonols [quercetin and quercitrin], and flavanones [hesperetin and hespiridin]) on cDNA-expressed CYP2C8 were investigated. CYP2C8 was co-expressed with NADPH-

Full text of DOI:10.1007/s00210-012-0731-5

Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
DOI 10.1007/s00210-012-0731-5
ORIGINAL ARTICLE
In vitro modulatory effects of flavonoids on human
cytochrome P450 2C8 (CYP2C8)
Chia Yong Pang & Joon Wah Mak & Rusli Ismail &
Chin Eng Ong
Received: 7 August 2011 /Accepted: 23 January 2012 /Published online: 4 February 2012
#
Springer-Verlag 2012
Abstract The inhibitory effects of five flavonoids with distinct
chemical classes (flavones [luteolin], flavonols [quercetin and
quercitrin], and flavanones [hesperetin and hespiridin]) on
cDNA-expressed CYP2C8 were investigated. CYP2C8 was
co-expressed with NADPH-cytochrome P450 reductase in
Escherichia coli and used to characterise potency and
mechanism of these flavonoids on the isoform. Tolbutamide
4-methylhydroxylase, a high-performance liquid
chromatography-based assay, was selected as marker activity
for CYP2C8. Our results indicated that the flavonoids inhibited
CYP2C8 with different potency. The order of inhibitory activ-
ities was quercetin > luteolin > hesperetin > hesperidin >
quercitrin. All of these compounds however exhibited
mechanism-based inhibition. A number of structural factors
were found to be important for inhibition; these include the
molecular shape (volume to surface ratio), the number of
hydroxyl groups as well as glycosylation of the hydroxyl
group. Quercetin was the most potent inhibitor among the
flavonoids examined in this study, and our data suggest that it
should be examined for potential pharmacokinetic drug inter-
actions pertaining to CYP2C8 substrates in vivo.
.
.
.
Keywords Cytochrome P450 CYP2C8 Flavonoids
In vitro inhibition
Introduction
Flavonoids represent a group of phytochemicals which are
widely found in edible plants. Plant flavonoids have been of
interest to scientists for decades, owing to their observed
biological effects such as free-radical scavenging, modula-
tion of enzymatic activity, and inhibition of cellular prolif-
eration, as well as their potential utility as antibiotic, anti-
allergic, anti-diarrheal, anti-ulcer, and anti-inflammatory
agents (Havsteen 2002).
Flavonoids are also well known to interact with enzymes
involved in drug metabolism. Among the many drug-
metabolising enzyme systems, cytochrome P450 is consid-
ered to be important as CYP–flavonoid interactions have
been shown to give rise to important pharmacological and
toxicological implications. Flavonoids can induce the ex-
pression of several CYPs and modulate their metabolic
activity, thus enhancing metabolism of certain drugs. On
the other hand, inhibition of CYPs involved in carcinogen
activation and generation of scavenging reactive species can
be beneficial properties of various flavonoids. Modulatory
effects of flavonoids have been investigated in CYP1A1
(Zhai et al. 1998), CYP1A2 (Lee et al. 1998), CYP1B1
(Doostdar et al. 2000), CYP3A4 (Chan et al. 1998) and
more recently, CYP2C9 and CYP2A6 (Si et al. 2009;
Tiong et al. 2010). The flavonoid modulatory potency and
the structural features important for their effects on another
important CYP2C isoform, CYP2C8, have not been
C. E. Ong (*)
Jeffrey Cheah School of Medicine and Health Sciences,
Monash University Sunway Campus,
Jalan Lagoon Selatan,
46150, Bandar Sunway, Selangor, Malaysia
e-mail: ceong98@hotmail.com
:
C. Y. Pang J. W. Mak
School of Pharmacy and Health Sciences,
International Medical University,
126, Jalan Jalil Perkasa 19, Bukit Jalil,
57000, Kuala Lumpur, Malaysia
R. Ismail
Pharmacogenetics Research Group, Institute for Research
in Molecular Medicine, Universiti Sains Malaysia,
16150 Kubang Kerian,
Kelantan, Malaysia

496
Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
investigated. The role of the CYP2C8 has recently garnered
considerable interest in the fields of drug metabolism and
pharmacogenetics. The isoform is now known to play a
major role in the metabolism of several therapeutically
important drugs, including chemotherapeutic agents (i.e.
paclitaxel and all-trans-retinoic acid), anti-malarials (i.e.
amodiaquine), anti-diabetic agents (i.e. rosiglitazone and
repaglinide), antiarrhythmics (i.e. amiodarone), and HMG-
CoA reductase inhibitors (i.e. cerivastatin and fluvastatin)
(Totah and Rettie 2005). It also converts endogenous sub-
stances, such as arachidonic acid, to biologically active
epoxide metabolites (Totah and Rettie 2005). Relevant to
this area of study is the use of quercetin, a flavonoid com-
pound, as the in vitro CYP2C8 inhibitor probe in many drug
metabolism and pharmacokinetic studies in recent years
(Rahman et al. 1994; Li et al. 2002; Gao et al. 2010).
Although quercetin exhibited potent inhibitory effect on
CYP2C8 in these studies, its relative potency in comparison
to other structurally related flavonoid compounds have not
been explored. In view of this and the increasing role played
by CYP2C8 in drug metabolism, characterisation of the
interaction mechanism of flavonoids with this isoform
remains an area of research interest. In this study, CYP2C8
was co-expressed with NADPH-cytochrome P450 reductase
(OxR) in Escherichia coli, and the expressed protein was
examined for modulatory effects by five naturally occurring
flavonoids with different structural classes. The potency and
selectivity of CYP2C8 inhibition, as well as the inhibition
mechanism of these compounds were investigated.
Preparation of CYP2C8 monooxygenase systems
Co-expression of CYP2C8 and OxR in E. coli DH5α cells
and subsequent membrane isolation were carried out as
described previously (Singh et al. 2008). The membrane
fragments of E. coli were stored at −80°C in a 1:1 mixture
of pH 7.6 TES (100 mmol/L Tris : 0.5 mmol/L EDTA :
500 mmol/L sucrose buffer) and ice-cold distilled water before
utilised in the enzyme assay. Routinely, the CYP2C8 spectral
content obtained from culturing experiments was 0.37
0.05 nmol mg−1 (n03).
Tolbutamide 4-methylhydroxylase assay
Tolbutamide 4-methylhydroxylase assay was carried out
essentially according to the published protocol (Miners et
al. 1988). A standard 0.5 mL incubation mixture contained
0.4 mg DH5α membrane protein, NADPH generating sys-
tem (1 mmol/L NADP, 10 mmol/L G-6-P, 2 IU G-6-P
dehydrogenase and 5 mmol/L MgCl2) and incubated at
37°C for 120 min. All reactions were started with the addi-
tion of NADPH-generating system. Reactions were then
stopped by placing the samples in ice followed by addition
of 1.0 mL of 0.1 mol/L phosphoric acid. The sample was
then extracted with 8 mL of hexane: chloroform: iso-amyl
alcohol (1,000: 250: 1v/v) to remove excess tolbutamide.
After centrifuging the mixture for 5 min at 2,000 rpm, the
tube containing the mixture was stored in −80°C freezer for
10 min before the organic layer of the mixture was dis-
carded. Fourteen microlitres of phenytoin (1 g/L), the inter-
nal standard, and 8 mL of diethyl ether were later added to
the sample and extracted by vortexing for a few minutes.
The sample was then centrifuged for 5 min at 2,000 rpm
before being stored in the −80°C freezer for 15 min. The
resulting organic layer was decanted into a conical tube and
then evaporated to dryness using N2 gas. The residue was
reconstituted in 60 μL of mobile phase, of which 20 μL was
injected onto the HPLC system. The chromatography was
performed using an Inertsil® C18 column (15 cm×4.6 mm
inner diameter, 4 micron particle size; GL Sciences, USA)
and eluted with acetate buffer (10 mmol/L, pH 4.3) : aceto-
nitrile (78:22) at flow rate of 2.0 mL/min. The column was
connected to a Perkin Elmer HPLC Series 200 system
comprising a pump and a UV detector. Absorbance was
monitored at 230 nm. Under these conditions, the retention
times for hydroxytolbutamide and phenytoin were 16.2 and
30.6 min, respectively. Standard curves for hydroxytolbuta-
mide were constructed in the range of 4.5–25.0 μmol/L.
Unknown concentrations of hydroxytolbutamide were de-
termined by comparison of hydroxytolbutamide:phenytoin
peak height ratios with those of the standard curve. Assay
validation has been carried out previously in our laboratory.
Overall assay within-day precision was determined by
Materials and methods
Chemicals and reagents
Acetonitrile, phosphoric acid, hydrochloride acid and all other
HPLC grade solvents were purchased from Fisher Scientific
(Pittsburgh, PA, USA). Tris base and isopropyl β-D-1-thioga-
lactopyranoside were acquired from Promega (Madison, WI,
USA). Terrific broth media was purchased from Invitrogen
Corporation (Carlsbad, CA, USA). Nicotinamide adenine
dinucleotide phosphate (NADP), glucose-6-phosphate
dehydrogenase, glucose-6-phosphate (G-6-P), dimethyl
sulfoxide (DMSO), magnesium chloride, tolbutamide, 4-
hydroxytolbutamide and phenytoin were purchased from
Sigma-Aldrich (St. Louis, MI, USA). The five flavonoids used
in this study, quercetin dehydrate (purity ≥ 98%), hesperidin
(purity ≥ 97%), hesperetin (purity ≥ 95%), luteolin (purity ≥
98%) and quercitrin (purity ≥78%), were also purchased from
Sigma-Aldrich. The E. coli bacterial stocks harbouring plas-
mids, pCW-CYP2C8 and pACYC-OxR, used for the expres-
sion of CYP2C8 and OxR, were constructed and prepared in
our laboratory previously (Singh et al. 2008).

Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
497
measuring hydroxytolbutamide formation in separate incu-
bations of the same batch of the expressed enzyme.
Coefficients of variation for hydroxytolbutamide formation
were 7.1%, 5.0% and 3.5% at substrate concentration of 0.1,
0.25 and 1.0 mmol/L, respectively. The corresponding inter-
day precision values were 8.9%, 6.3% and 5.7%, respectively.
The mean recovery of hydroxytolbutamide, calculated by
comparing the peak height for extracted compound with that
of an equal amount injected directly into the chromatograph,
was 87.0 3.8% for six samples in the concentration range
4.5–25.0 μmol/L.
Kinetic analysis
Enzyme kinetic data were analysed by nonlinear least
squares regression analysis software EZ-FitTM (Perrella
Scientific, USA) and the kinetic parameters, Km and Vmax
,
were determined over the substrate concentration range
studied. The IC50 values for flavonoid inhibition were inter-
polated mathematically by SigmaPlot® software (version
2004, Systat Software Inc, USA). This was accomplished
by fitting the function V0100×IC50 / (IC50+C) (V0activity
expressed as percentage of control and C0concentration of
flavonoid) using the method of Marquardt in the nonlinear
regression program of SigmaPlot® software.
Enzyme inhibition experiments
In order to determine the modulatory effects of the flavo-
noids on CYP2C8, the enzyme assay was carried out in the
presence and absence of flavonoids. All flavonoids dis-
solved well in DMSO and stock solutions were subsequent-
ly prepared in this solvent. Individual flavonoid compounds
were added to incubations and the final concentration of
DMSO did not exceed 0.5% v/v. Control incubations
contained an equivalent volume of DMSO, although it was
shown that DMSO had a negligible effect on CYP2C8-
catalysed reaction at the concentration (0.5% v/v) employed.
Preliminary experiments were carried out first to determine
kinetic parameters Michaelis–Menten constant (Km) and
maximum velocity (Vmax) for CYP2C8. All inhibition
experiments were subsequently conducted by incubating a
single tolbutamide concentration around the determined Km
with a range of flavonoid concentrations to determine IC50
values (concentration of inhibitor causing 50% inhibition of
enzyme activity). Concentrations used in inhibition experi-
ments for each flavonoid were 20–200 μmol/L for luteolin,
10–200 μmol/L for quercetin, 10–500 μmol/L for quercitrin,
and 50–400 μmol/L for hesperetin and hesperidin. To assess
mechanism-based activities of the flavonoids, the enzyme
assay was conducted with a period of pre-incubation. For this,
the reaction mixtures containing 0.1 mol/L phosphate buffer
(pH 7.4), NADPH-generating system, expressed CYP2C8
and flavonoids were prepared on ice. After the reaction mix-
ture was pre-incubated at 37°C for 15 min, tolbutamide was
added to the mixture to initiate the reaction. The metabolism
of the substrate was subsequently analysed as described
above. The IC50 values with or without pre-incubation were
determined for each of the flavonoids studied. In addition, the
NADPH-dependency of inhibition was also examined by
incubating expressed CYP2C8 and the flavonoid in phosphate
buffer for 15 min before addition of tolbutamide and NADPH-
generating system to initiate the assay as described above. The
velocity rates were subsequently determined and compared to
control incubations where NADPH generating system was
added at the beginning of the pre-incubation.
Results
Kinetic parameters were determined for tolbutamide 4-
methylhydroxylation by incubating expressed CYP2C8 at
substrate concentration ranging from 200 to 2,000 μmol/L.
Hydroxytolbutamide formation by the expressed CYP2C8
was found to best fit the single Michaelis–Menten kinetic,
with apparent Km and Vmax values of 1.16 0.13 mmol/L
and 10.2 1.8 pmol min−1 mg protein−1 (or 27.6 4.9 pmol
min−1 nmol CYP−1; n03), respectively. These data were in
the similar range as reported for CYP2C8 in the literature
(Relling et al. 1990; Veronese et al. 1993). Based on these
calculated data, subsequent inhibition study with flavonoids
was conducted at tolbutamide concentration of 1.0 mmol/L
which corresponded to the apparent Km value calculated
above.
Five flavonoid compounds, luteolin, quercetin, querci-
trin, hesperetin and hespiridin, were chosen for the study
based on their structural classes (see Table 1). All these
compounds, dissolved in DMSO, were used separately in
the inhibition experiments. Inhibitory effect was examined
at 5–7 inhibitor concentrations at a fixed tolbutamide con-
centration described above. Quercetin, luteolin, hesperetin
and hesperidin exhibited inhibitory effects on CYP2C8
activity, giving IC50 values of 46.0, 82.0, 168.4 and
274.7 μmol/L, respectively (third column in Table 1). In
contrast to the above flavonoids, quercitrin did not show any
appreciable inhibition over the range of concentration inves-
tigated (10–500 μmol/L). Thus, quercitrin is the least potent
inhibitor among the five compounds examined and its IC50
value was designated as >500 μmol/L.
To further explore the inhibition mechanism of the four
flavonoids showing inhibitory effect (i.e. quercetin, luteolin,
hesperetin and hesperidin), kinetic analyses of CYP2C8-
catalysed tolbutamide hydroxylase were performed with or
without a period of pre-incubation with the flavonoids at se-
lected concentrations. As shown in Fig. 1, pre-incubation of the
four flavonoids at their respective selected concentrations

498
Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
Table 1 Inhibitory effects of flavonoids on CYP2C8 activity
n.d. Not determined
a Data are presented as mean SD from triplicate determinations
resulted in greater degree of inhibition for CYP2C8-catalysed
tolbutamide hydroxylation, as reflected by the lower velocity
rates observed (white-coloured bars vs dark grey bars in Fig. 1).
NADPH-dependency of the inhibition was also examined by
pre-incubating the flavonoids with CYP2C8 in the absence of
NADPH generating system during the 15-min pre-incubation

Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
499
incubation is illustrated in Fig. 2. Quercitrin was not examined
for the pre-incubation effect as it did not show appreciable
inhibition. Considered collectively, the above data have pro-
vided evidence for the mechanism-based inhibition exerted by
the investigated flavonoids on CYP2C8.
Discussion
Tolbutamide hydroxylase assay was used as the activity mark-
er for CYP2C8 as tolbutamide and 4-hydroxytolbutamide
were commercially available. This assay is a well established
HPLC method commonly used for CYP2C subfamily mem-
bers. Assay validation has been performed and the protocol
has been found to be highly reliable and reproducible (Miners
et al. 1988; Baldwin et al. 1999). Tolbutamide hydroxylase
activity reported in the present investigation (routinely 2.18–
4.13 pmol min−1 mg−1 or 5.89–11.16 pmol min−1 nmol
CYP−1 at 1.0 mmol/L tolbutamide) was consistent with activ-
ity levels reported in the literature which range from 0.44 to
2.48 pmol min−1 mg−1 (Relling et al. 1990; Veronese et al.
1993). Km value reported here was also close to that reported
by other studies (Veronese et al. 1993). Taken together, data
from kinetic characterization indicated that CYP2C8 and OxR
have been well expressed in DH5α cells and the two proteins
interacted optimally to generate a catalytically active system
for kinetic and inhibition studies.
To elucidate the structural features of flavonoids that are
responsible for modulating CYP2C8 activity, we examined
flavonoid derivatives that have distinct and different structural
features. Our results showed that flavonoids affected in vitro
CYP2C8-catalysed reaction differently. These differences
were related to the chemical structure of the flavonoids. The
five compounds investigated belong to three different struc-
tural classes. Luteolin, a flavone, differs from flavonol (quer-
cetin) in terms of absence of a hydroxyl group at C3 position.
Quercitrin, another flavonol, is the glycosylated form of quer-
cetin. Hesperetin and hesperidin, both flavanone, differ from
the others as they are buckled molecules, having the exocyclic
phenyl ring B that lies nearly perpendicular to the rest of the
molecule. This conformation is due to the absence of C2-C3
double bond in the ring C, unlike flavone and flavonol which
are fairly planar molecules. The presence of a C2-C3 double
bond restricts the orientation of phenyl ring B, hence the small
volume to surface (V/S) ratio in both flavone and flavonol. As
with quercitrin, hesperidin is the glycosylated form of hesper-
etin, having a rutinoside at C7 of the hesperetin.
Fig. 1 Velocity rates for tolbutamide hydroxylation in incubations of
tolbutamide (1.0 mmol/L) without (filled bars with white colour) or
with (bars with light and dark grey) pre-incubation with the flavonoid
compounds at the indicated concentrations. Pre-incubations were car-
ried out in the absent (light grey) or present (dark grey) of NADPH
generating system during the pre-incubation period. Data are shown as
mean SD from triplicate determinations
period. As shown in Fig. 1, this manoeuvre resulted in no
difference in velocity rates as compared to control incubations
without pre-incubation (white-coloured bars vs light grey bars
in Fig. 1). Based on the calculated velocities, 15-min pre-
incubations (with the flavonoids and in the presence of
NADPH generating system) have increased inhibitory potency
towards CYP2C8 to 2.5 times for both quercetin and luteolin,
3.3 times for hesperetin, and 2.1 times for hesperidin, respec-
tively. In view of the substantial reduction in the observed
CYP2C8 activity, IC50 values were re-determined for the four
flavonoids with the 15-min pre-incubation treatment at various
inhibitor concentrations. Consistent with the enhanced inhibi-
tory effect observed with a single concentration (Fig. 1), quer-
cetin, luteolin, hesperetin and hesperidin exhibited lower IC50
values of 26.7, 52.2, 127.6 and 209.7 μmol/L respectively
(fourth column in Table 1). A representative IC50 plot for
hesperetin showing the curves with and without pre-
Based on the IC50 values, the order of inhibitory potency
among the flavonoids examined was: quercetin > luteolin >
hesperetin > hesperidin > quercitrin. Examination of the struc-
ture–activity relationship of these compounds seems to suggest
the importance of certain structural factors in CYP2C8 inhibi-
tion. Planar molecules, quercetin and luteolin, having smaller
Fig. 2 Effect of different concentrations of hesperetin on the tolbuta-
mide hydroxylase activity catalysed by CYP2C8 in the absence (solid
line) and the presence (dashed line) of pre-incubation. Data are shown
as mean SD from triplicate determinations

500
Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
(V/S) ratio inhibited more potently than buckled molecules,
hesperetin and hesperidin. This may indicate that CYP2C8
binds preferentially planar molecules than buckled molecules
which have high (V/S) ratio. This result is in agreement to a
study involving CYP1A2 where it was shown that planar
molecules possessed higher inhibitory activity (Lee et al.
1998). Another factor that controlled CYP2C8 inhibition was
the number of free hydroxyl groups attached to flavonoid
nucleus. Among the flavonoids examined, quercetin, with five
hydroxyl groups, showed higher inhibitory activity than luteo-
lin which carries four hydroxyl groups. Luteolin, in turn,
exhibited two-fold greater potency than hesperetin that carries
only three free hydroxyl groups. This result is again in agree-
ment with that of Lee et al. (1998), which showed increasing
inhibitory potency toward CYP1A2-mediated caffeine N3-
demethylation with an increasing degree of hydroxylation in
flavonoids from both flavone and flavonol classes. Several
other earlier studies have also demonstrated that the numbers
and positions of hydroxyl groups on the A and B rings of
flavonoids were important in determining their effects on en-
zyme activities (Lasker et al. 1984; Guengerich and Kim 1990;
Chae et al. 1991). It has been suggested, based on the finding
from these studies that free hydroxyl groups, in particular, the
C5 and C7 hydroxyl groups, preferentially interact with Fe3+ of
the CYP active site due to their steric availability and adequate
acidity (Li et al. 1994). The results of the present investigation
also indicated that glycosylation of the free hydroxyl groups
decreased the ability to inhibit CYP2C8. This effect was clearly
observed when comparing the inhibitory effect of quercetin
with its glycoside, quercitrin as well as that of hesperetin and
hesperidin. Rhamnoglucoside, attached to C3 of quercitrin,
essentially abolished the potent inhibitory effect seen in quer-
cetin, even though quercetin was the most potent inhibitor
among the five flavonoid compounds investigated. Similarly,
hesperidin, with a rutinoside attached to its C7, exhibited a 1.6-
fold increase in the IC50 value when compared to hesperetin.
This seems to indicate that the large polar substituent at position
C3 or C7 of the flavonoid nucleus diminishes or reduces the
affinity of the compounds toward CYP2C8. It is likely that the
free hydroxyl groups at these positions are essential for high
affinity binding of the flavonoids to CYP2C8. The finding from
this study is also consistent with results obtained in other
studies (Lee et al. 1998) which showed decrease in inhibitory
effect of glycosylated flavonoids when comparison to their
respective aglycones was made.
at various flavonoid concentrations revealed substantially
lower IC50 values for quercetin, luteolin, hesperetin and hes-
peridin (Table 1). Mechanism-based inhibition involves me-
tabolism of drugs or other chemicals to products that, in turn,
inactivate CYP enzymes. Mechanisms involved may take
several forms, including direct interaction with the haem of
CYP leading to formation of haem-protein adduct, as well as
formation of metabolite-intermediate (MI) complex that binds
tightly to CYP haem (VandenBrink and Isoherranen 2010). In
the case of flavonoids, it is well known that CYP enzymes are
involved in their metabolism leading to hydroxylated and O-
demethylated metabolites (Hollman and Katan 1997). It is
probable that some reactive metabolites of flavonoids were
formed by CYP2C8 in this study, leading to mechanism-based
inactivation. This is also consistent with the fact that polyhy-
droxylated flavonoids were shown to bind to CYP proteins as
Fe3+ ligands. The free hydroxyl groups, in particular, those
attached to C5 and C7, have been suggested to form direct
ligand binding to the Fe3+ atom at the CYP active site (Beyeler
et al. 1988). It is possible that these interactions may lead to
formation of MI complex and/or haem-protein adduct, hence
inactivation of CYP2C8 seen in the present investigation.
Quercetin was the most potent inhibitor among the flavo-
noids investigated in the present project. A number of studies
have reported the use of quercetin as an inhibitor probe for
CYP2C8 in in vitro studies (Rahman et al. 1994; Ono et al.
1996; Li et al. 2002; Gao et al. 2010). Specificity of quercetin
toward CYP2C8 was, however, found to be concentration
dependent. When used at 100 μmol/L, quercetin inhibited
CYP isoform selective substrate probes non-selectively,
showing inhibition of 30–100% of all CYPs tested (including
CYPs 3A4, 3A5, 2E1, 2D6, 2C8, 2C9, 2B6, 2A6 and 1A2)
with isoforms CYPs 2A6, 2B6, 2C8 and 2C9 inhibited to a
greater extent (Ono et al. 1996). Another study
(Masimirembwa et al. 1999) has indicated that quercetin,
when used at a lower concentration (10 μmol/L), exhibited
significant selectivity towards paclitaxel 6α-hydroxylase ac-
tivity mediated by CYP2C8. At this concentration, CYP2C8
was inhibited by more than 80% while other isoforms (CYPs
1A1, 1A2, 2C9, 2D6, 3A4 and 3A5) were inhibited to a lesser
degree (about 20%). This showed that quercetin only works
well as CYP2C8 isoform-selective inhibitory probe in a nar-
row concentration range. When used in high concentration,
the selectivity seems to be lost, thus it is important to carefully
select the inhibitor concentration when any inhibitory probe is
to be used in in vitro inhibition study. Although the inhibitory
effect of quercetin across various CYP isoforms have been
studied, the comparative study of inhibitory effects of flavo-
noid compounds which are structurally closely related to
quercetin on CYP2C8 activity has not been reported. The
present project specifically looks into this particular aspect,
and the results showed that quercetin was the most potent
inhibitor among the flavonoids examined. As discussed
Further investigation on the mechanism of inhibition
revealed that the four flavonoids examined, quercetin, luteo-
lin, hesperetin and hesperidin, exhibited mechanism-based
inhibition on CYP2C8. The inhibitory effects of these inhib-
itors increased considerably after pre-incubation with
NADPH. Furthermore, control incubations without NADPH
did not result in reduced activity, implying NADPH-
dependency of the inhibition (Fig. 1). Subsequent incubations

Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
501
above, molecular shape, position and number of hydroxyl
groups as well as glycosylation have been found to be the
factors governing the differential inhibitory effect observed.
Studies have shown that humans ingest, on average,
0.78–1.06 g of flavonoid daily (Gosnay et al. 2002; Woods
et al. 2003). Inhibition of CYP activities, including CYP2C8
examined here, could be considered relevant in drug–food
interaction. Foods contain a great variety of flavonoids, and
the overall metabolism and bioavailability of these com-
pounds in man remain an area of research interest. Recent
studies have shown that, following ingestion of diets rich in
flavonoid, maximum concentration for total flavonoids
(both conjugate forms and aglycones) achieved may vary
from 3.1 to 22.5 μmol/L (Clifford 2004). However, it is
considered unlikely that plasma flavonoid concentration
will routinely exceed 10 μmol/L in total after normal dietary
intake. Hence, when this is considered with the IC50 values
of various flavonoids obtained in this study (26.7 μmol/L or
larger), the likely in vivo inhibitory effect on CYP2C8-
mediated tolbutamide hydroxylation would only be at most
marginal as the determined IC50s were higher than the
typical plasma level attained in our body. Nevertheless, this
notion of marginal inhibition may not hold true for other
CYP2C8 substrates such as amodiaquine and paclitaxel.
Data from other studies have demonstrated lower IC50 or
Ki values of quercetin and other flavonoids towards
CYP2C8 substrates. As discussed above, quercetin inhibited
CYP2C8-catalysed amodiaquine N-deethylation potently
with a Ki of 2.0 μmol/L (Li et al. 2002). This level is well
below the typical flavonoid level attained from the normal
diets (about 10 μmol/L) implying potential in vivo interac-
tion between quercetin and amodiaquine via CYP2C8 inhi-
bition. Furthermore, the study by Masimirembwa and
coworker which demonstrated potent inhibition of paclitaxel
6α-hydroxylation by 10 μmol/L quercetin (see above) may
also have in vivo implication for paclitaxel–quercetin inter-
action (Masimirembwa et al. 1999). This is also consistent
with two other studies which showed low Ki and IC50 values
for quercetin against CYP2C8-catalysed paclitaxel hydrox-
ylation with values of 1.3 μmol/L and 5.4 μmol/L, respec-
tively (Rahman et al. 1994; Gao et al. 2010). On the basis of
the discussion above, it is thus apparent that the potential for
clinical interaction between flavonoid and CYP2C8 would
depend on the substrate taken as well as the plasma flavo-
noid attained in individual patients affected. Further studies
are needed to obtain additional evidence that quercetin and
other flavonoids examined in the present study can be
effective in vivo inhibitors of CYP2C8.
potent inhibitor among the flavonoids examined and further
work is warranted to study the in vivo interactions between
CYP2C8 substrates and this flavonoid.
Acknowledgements This work was supported by a research grant
from the International Medical University (Grant no: BMedSc 101/04),
Kuala Lumpur, Malaysia.
Conflicts of interest The authors declare that they have no conflict
of interest.
References
Baldwin SJ, Clarke SE, Chenery RJ (1999) Characterisation of the
cytochrome P450 enzymes involved in the in vitro metabolism of
rosiglitazone. Br J Clin Pharmacol 48:416–423
Beyeler S, Testa B, Perrissoud D (1988) Flavonoids as inhibitors of rat
liver monooxygenase activities. Biochem Pharmacol 37:1971–
1979
Chae YH, Marcus CB, Ho DK, Cassady JM, Baird WM (1991) Effects
of synthetic and naturally occurring flavonoids on benzo[a]pyrene
metabolism by hepatic microsomes prepared from rats treated
with cytochrome P-450 inducers. Cancer Lett 60:15–24
Chan WK, Nguyen LT, Miller VP, Harris RZ (1998) Mechanism-based
inactivation of human cytochrome P450 3A4 by grapefruit juice
and red wine. Life Sci 62:PL135–PL142
Clifford MN (2004) Diet-derived phenols in plasma and tissues and
their implications for health. Planta Med 70:1103–1114
Doostdar H, Burke MD, Mayer RT (2000) Bioflavonoids: selective
substrates and inhibitors for cytochrome P450 CYP1A and
CYP1B1. Toxicology 144:31–38
Gao Y, Liu D, Wang H, Zhu J, Chen C (2010) Functional character-
ization of five CYP2C8 variants and prediction of CYP2C8
genotype-dependent effects on in vitro and in vivo drug–drug
interactions. Xenobiotica 40:467–475
Gosnay SL, Bishop JA, New SA, Catterick J, Clifford MN (2002)
Estimation of the mean intakes of fourteen classes of dietary
phenolics in a population of young British women aged 20–
30 years. Proc Nutr Soc 61:125A
Guengerich FP, Kim DH (1990) In vitro inhibition of dihydropyridine
oxidation and aflatoxin B1 activation in human liver microsomes
by naringenin and other flavonoids. Carcinogenesis 11:2275–
2279
Havsteen BH (2002) The biochemistry and medical significance of the
flavonoids. Pharmacol Ther 96:67–202
Hollman PC, Katan MB (1997) Absorption, metabolism and health
effects of dietary flavonoids in man. Biomed Pharmacother
51:305–310
Lasker JM, Huang MT, Conney AH (1984) In vitro and in vivo
activation of oxidative drug metabolism by flavonoids. J Pharmacol
Exp Ther 229:162–170
Lee H, Yeom H, Kim YG, Yoon CN, Jin C, Choi JS, Kim BR, Kim DH
(1998) Structure-related inhibition of human hepatic caffeine N3-
demethylation by naturally occurring flavonoids. Biochem Pharmacol
55:1369–1375
Li Y, Wang E, Patten CJ, Chen L, Yang CS (1994) Effects of flavo-
noids on cytochrome P450-dependent acetaminophen metabolism
in rats and human liver microsomes. Drug Metab Dispos 22:566–
571
In conclusion, the works carried out in this study have
confirmed the differential modulatory effects of some natu-
rally occurring flavonoids on CYP2C8, and that certain
structural factors have been found to be important to exert
inhibitory effect to the isoform. Quercetin was the most
Li XQ, Bjorkman A, Andersson TB, Ridderstrom M, Masimirembwa
CM (2002) Amodiaquine clearance and its metabolism to N-
desethylamodiaquine is mediated by CYP2C8: a new high affinity

502
Naunyn-Schmiedeberg's Arch Pharmacol (2012) 385:495–502
and turnover enzyme-specific probe substrate. J Pharmacol Exp
Ther 300:399–407
Si D, Wang Y, Zhou YH, Guo Y, Wang J, Zhou H, Li ZS, Fawcett JP
(2009) Mechanism of CYP2C9 inhibition by flavones and flavo-
nols. Drug Metab Dispos 37:629–634
Masimirembwa CM, Otter C, Berg M, Jonsson M, Leidvik B,
Jonsson E, Johansson T, Backman A, Edlund A, Andersson
TB (1999) Heterologous expression and kinetic characteriza-
tion of human cytochromes P-450: validation of a pharmaceu-
tical tool for drug metabolism research. Drug Metab Dispos
27:1117–1122
Miners JO, Smith KJ, Robson RA, McManus ME, Veronese ME,
Birkett DJ (1988) Tolbutamide hydroxylation by human liver
microsomes. Kinetic characterisation and relationship to other
cytochrome P-450 dependent xenobiotic oxidations. Biochem
Pharmacol 37:1137–1144
Ono S, Hatanaka T, Hotta H, Satoh T, Gonzalez FJ, Tsutsui M (1996)
Specificity of substrate and inhibitor probes for cytochrome
P450s: evaluation of in vitro metabolism using cDNA-expressed
human P450s and human liver microsomes. Xenobiotica 26:681–
693
Rahman A, Korzekwa KR, Grogan J, Gonzalez FJ, Harris JW (1994)
Selective biotransformation of taxol to 6α-hydroxytaxol by human
cytochrome P450 2 C8. Cancer Res 54:5543–5546
Singh R, Ting JG, Pan Y, Teh LK, Ismail R, Ong CE (2008) Functional
role of Ile264 in CYP2C8: mutations affect haem incorporation
and catalytic activity. Drug Metab Pharmacokinet 23:165–174
Tiong KH, Yiap BC, Tan EL, Ismail R, Ong CE (2010) In vitro
modulation of naturally occurring flavonoids on cytochrome
P450 2A6 (CYP2A6) activity. Xenobiotica 40:458–466
Totah RA, Rettie AE (2005) Cytochrome P450 2 C8: substrates,
inhibitors, pharmacogenetics, and clinical relevance. Clin Phar-
macol Ther 77:341–352
VandenBrink BM, Isoherranen N (2010) The role of metabolites in
predicting drug–drug interactions: focus on irreversible P450
inhibition. Curr Opin Drug Discov Devel 13:66–77
Veronese ME, Doecke CJ, Mackenzie PI, McManus ME, Miners JO,
Rees DL, Gasser R, Meyer UA, Birkett DJ (1993) Site-directed
mutation studies of human liver cytochrome P-450 isoenzymes in
the CYP2C subfamily. Biochem J 289:533–538
Woods E, Clifford MN, Gibbs M, Hampton S, Arendt J, Morgan L (2003)
Estimation of mean intakes of 14 classes of dietary phenols in a
population of male shift workers. Proc Nutr Soc 62:60A
Zhai S, Dai R, Friedman FK, Vestal RE (1998) Comparative inhibition
of human cytochromes P450 1A1 and 1A2 by flavonoids. Drug
Metab Dispos 26:989–992
Relling MV, Aoyama T, Gonzalez FJ, Meyer UA (1990) Tolbutamide
and mephenytoin hydroxylation by human cytochrome P450s
in the CYP2C subfamily. J Pharmacol Exp Ther 252:442–
447