Document Outline

enc-Lavier et al., 1996; Shih et al., 2000

). The classes of ﬂavonoids include chalcones,

ﬂavones, ﬂavonols, ﬂavanones, ﬂavanols, anthocyanins
and isoﬂavones (

Table 1

). The ﬂavonoid natural products

exert a wide range of biochemical and pharmacological
properties, with one of the most investigated eﬀects being
their cancer preventive activities. The cancer protective
eﬀects of ﬂavonoids have been attributed to a wide variety
of mechanisms, including free radical scavenging, modify-
ing enzymes that activate or detoxify carcinogens, and
inhibiting the induction of the transcription factor activa-
tor protein-1 (AP-1) activity by tumor promoters (

Caniv-

). Flavonoids also

have inhibitory eﬀect on the activities of many enzymes,
including b-glucuronidase (

Kim et al., 1994

), lipoxygenase

(

Laughton et al., 1991; Schewe et al., 2002

), cyclooxygen-

ase (

Laughton et al., 1991

), inducible nitric oxide synthase

(

Raso et al., 2001

), monooxygenase (

Siess et al., 1995

), thy-

roid peroxidase (

Doerge and Chang, 2002

), xanthine oxi-

dase (

Sheu et al., 1998

), mitochondrial succinoxidase and

NADH-oxidase (

), phosphodiesterase

), phospholipase A

and

protein kinase (

Cushman et al., 1991

The focus of this paper will be the eﬀects of ﬂavonoids

on cytochrome P450 (CYP) and phase II enzymes which
are key enzymes involved in the metabolism of xenobiotics.

Table 2

presents a summary of the ﬂavonoid-induced alter-

ations in enzyme expression and/or activity described in
this review.

2. Eﬀect of bioﬂavonoids on cytochrome P450 (CYP450)

Cytochrome P450 enzymes (phase I monooxygenase

enzymes) are widely known for their role in the metabolism
of drugs and other foreign compounds. Thus, modulation
of this enzyme system can inﬂuence the metabolism of
xenobiotics, producing eﬀects of pharmacological and tox-
icological importance. A number of naturally occurring
ﬂavonoids have been shown to modulate the CYP450 sys-
tem, including the induction of speciﬁc CYP isozymes, and
the activation or inhibition of these enzymes (

Wood et al.,

1986

Wattenberg et al. (1968)

began investigating the

eﬀects of ﬂavonoids on the CYP system over 36 years
ago, and a number of subsequent studies have shown diﬀer-
ent modulatory eﬀects of ﬂavonoids on CYP activity
in vitro and in vivo. Many carcinogens are metabolized
by CYP enzymes to either biologically inactive metabolites
or to chemically reactive electrophilic metabolites that
covalently

bind

DNA

producing

carcinogenicity

(

Fig. 1

) (

Conney, 2003

). The reactive metabolites may

undergo additional metabolism by phase I or II enzymes
to inactive products. Therefore, induction of either phase
I or phase II enzymes can result in increased detoxiﬁcation
of carcinogens (

Conney, 2003

). Since many chemical car-

cinogens are metabolized by CYP enzymes to both inac-
tive, as well as to carcinogenic metabolites, the eﬀects of
inducers of these enzymes on the carcinogenicity of a chem-
ical will depend on the inducerÕs eﬀects on the diﬀerent met-
abolic pathways. In animal studies, it has been reported
that inducers of CYPs usually decrease the carcinogenicity
of chemical carcinogens in vivo; this suggests that induc-

188

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

Table 1
The chemical structures and major food sources of the 6 major ﬂavonoid subgroups

Structure

Example

Major food sources

Basic structure of ﬂavonoids

Chalcone

Hop chalcones (xanthohumol and
dehydrocyclo-xanthohumol hydrate)

Hops, beer

Flavone

Acacetin

Parsley, thyme, celery, sweet red peppers,
honey, propolis

Apigenin
Baicalein
Chrysin
Diosmetin
Luteolin
Tangeretin

Flavonol

Galangin

Onions, kale, broccoli, apples, cherries,
berries, tea, red wine

Kaempferol
Morin
Myricetin
Quercetin

Flavanone

Eriodictyol

Citrus

Hesperetin
Homoeriodictyol
Naringenin

Flavanol

Catechin

Cocoa, green tea, chocolate, red wine,
hawthorn, bilberry, motherwort,
and other herbs

Epicatechin
Proanthocyanidins

Anthocyanin

Cyanidins

Cherries, grapes, berries, red cabbage

Pigmented
compounds

Isoﬂavone

Biochanin A

Red clover, alfalfa, peas, soy
and other legumes

Genistein
Diadzein

Equol
Formononetin

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

189

Table 2
Summary of ﬂavonoid-induced enzyme induction or inhibition (

"induction, #inhibition, * in vivo study)

Phase I

Flavonoids

Eﬀects

References

CYP1A1

Quercetin

"mRNA expression and activity

Ciolino et al. (1999)

Kaempferol

#TCDD or B[a]P-induced transcription

Kang et al. (1999)

Galangin

#Activity

Ciolino and Yeh (1999)

"mRNA expression

Ciolino and Yeh (1999)

#TCDD or DMBA-induced transcription

Ciolino and Yeh (1999)

Green tea extracts

"Activity

Liu et al. (2003)

Theaﬂavins

#Activity in the intestine

Catterall et al. (2003)

Xanthohumol

#Activity

Baicalein

#DMBA-induced transcription

Chan et al. (2002)

#Activity

Chan et al. (2002)

Dimethoxyﬂavone

#Activity

Wen et al. (2005)

#B[a]P-induced mRNA and protein expression, and Activity

Wen et al. (2005)

Chrysin

"Activity

Wen et al. (2005)

Diosmin

"mRNA expression

Ciolino et al. (1998)

Diosmetin

"mRNA expression

Ciolino et al. (1998)

#DMBA-induced transcription

Ciolino et al. (1998)

#Activity

Acacetin

#Activity

Flavone

#Activity

Zhai et al. (1998)

Hesperetin

#Activity

Biochanin A

#DMBA-induced transcription and activity

Genistein

#DMBA-induced transcription and activity

-Hydroxychalcone

#mRNA expression and activity

Wang et al. (2005)

CYP1A2

Quercetin

#Activity

Tsyrlov et al. (1994)

Biapigenin

#Activity

Galangin

#Activity

Zhai et al. (1998)

Green tea extracts

#TPA-induced transcription

Shih et al. (2000)

"Activity

Liu et al. (2003)

8-Prenylnaringenin

#Activity

Isoxanthohumol

#Activity

Diosmetin

#Activity

Acacetin

#Activity

Flavone

#Activity

Zhai et al. (1998)

Tangeretin

#Activity

Obermeier et al. (1995)

Naringin

#Activity

Fuhr et al. (1993)

Genistein

#Activity

#Activity

Chen et al. (2004)

Equol

#Activity

Diadzein

#Activity

Peng et al. (2003)

CYP1B1

Xanthohumol

#Activity

Baicalein

#DMBA-induced transcription

Chan et al. (2002)

Diosmetin

#Activity

Acacetin

#Activity

Hesperetin

#Activity

Biochanin A

#DMBA-induced transcription and activity

Genistein

#DMBA-induced transcription and activity

Diadzein

#Activity

Roberts et al. (2004)

Formononetin

#Activity

Roberts et al. (2004)

-hydroxychalcone

#mRNA expression and activity

Wang et al. (2005)

CYP2E1

Genistein

#Activity

Equol

#Activity

Theaﬂavins

#Protein level

Catterall et al. (2003)

Silybin (=Silibinin, Silybinin)

#Activity

Zuber et al. (2002)

CYP3A4

Naringenin

#Activity

Fuhr (1998)

Hyperforin

#Activity

"mRNA expression

Zhou et al. (2003)

"Activity

Bray et al. (2002)

Biapigenin

#Activity

Quercetin

#Activity

Venkataramanan et al. (2000)

Silymarin

#Activity

Silybin

#Activity (at high conc.)

Beckmann-Knopp et al. (2000)

190

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

Table 2 (continued)

Phase I

Flavonoids

Eﬀects

References

"Activity (at low conc.)

Genistein

#Activity

Foster et al. (2003)

Diadzein

#Activity

Foster et al. (2003)

Glabridin

#Activity

Zhou et al. (2004)

Myricetin

#Activity

Ho et al. (2001)

CYP19

Biochanin A

#Activity

#Activity

#Activity

Formononetin

#Activity

Equol

#Activity

Pelissero et al. (1996)

Chrysin

#Activity

#Activity

#Activity

Gambelunghe et al. (2003)

#Activity

Naringenin

#Activity

#Activity

#Activity

Flavone

#Activity

Apigenin

#Activity

Bicalein

#Activity

Galangin

#Activity

Galijatovic et al. (2001)

Phase II
UGT1A1

Chrysin

"mRNA expression

Walle et al. (2000)

"activity

Galangin

"mRNA expression

Walle and Walle (2002)

Isorhamnetin

"mRNA expression

Walle and Walle (2002)

Tangeretin

#Activity

Naringenin

#Activity

Flavone

#Activity

Quercetin

#Activity

Canivenc-Lavier et al. (1996)

UGT

Flavone

"Activity

Brouard et al. (1988)

"Activity

"Activity

van der Logt et al. (2003)

Quercetin

"Activity

van der Logt et al. (2003)

"Activity

Green tea

"Activity

Bu-Abbas et al. (1995)

"Activity

Sohn et al. (1994)

Biochanin A

"Activity

Formononetin

"Activity

Genistein

"Activity

Diadzein

"Activity

Naringenin

"Activity

Galangin

"Activity

Kaempferol

"Activity

GST

Genistein

"or #mRNA expression

Ansell et al. (2004)

(cell type speciﬁc)
"Activity

Diadzein

"Activity

Sharma and Sultana (2004)

Genistein + Diadzein

Prevent TPA-downregulated Activity

Flavone

"Activity

Nijhoﬀ et al. (1995)

Green tea extract

"Activity

Maliakal et al. (2001)

Morin

"Activity

Tanaka et al. (1999)

"Activity

Kawabata et al. (1999)

Silymarin

"Activity

Kohno et al. (2002)

GSTA2

Flavone

"mRNA expression

Kang et al. (2003)

-amino-3

methoxyﬂavone

"mRNA expression

Kang et al. (2003)

GSTP1-1

Quercetin

#Activity

van Zanden et al. (2003)

Tephropurpurin

"Activity

Chang et al. (1997)

Xanthohumol

"Activity

Miranda et al. (2000)

(continued on next page)

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

191

tion of Phase I and II detoxiﬁcation pathways may occur to
a greater extent than induction of CYPs involved in the
formation of carcinogenic metabolites (

Fig. 1

) (

Conney,

2003

2.1. CYP1

The CYP1 family consists of 1A1, 1A2, and 1B1 mem-

bers that are capable of activating procarcinogens.
CYP1A1 and 1B1 are both involved in the biotransforma-
tion of polycyclic aromatic hydrocarbons (PAHs), a class
of ubiquitous environmental chemicals, to carcinogenic
metabolites (

Guengerich and Shimada, 1991

). This process

is believed to contribute to pulmonary carcinogenesis,
because increased lung CYP1A1 expression and activity
are associated with a high risk of lung cancer (

McLemore

et al., 1990

). High CYP1A1 activity is also associated with

colorectal cancer (

Sivaraman et al., 1994

). CYP1A2 mainly

metabolizes important drugs such as phenacetin, theophyl-
line, caﬀeine, imipramine, and propranolol (

Brosen, 1995

and also activates some procarcinogens to carcinogens.

CYP1A2 plays a role in human tobacco-related cancers
(

Smith et al., 1996

). CYP1A1 is poorly expressed in human

liver although its synthesis can be markedly induced in
many extrahepatic tissues, notably the lungs (

Rendic and

Di Carlo, 1997

). In contrast, CYP1A2 is expressed princi-

pally in the liver (

Rendic and Di Carlo, 1997

). CYP1B1

is an extrahepatic estradiol 4-hydroxylase that activates
procarcinogens and elevated levels have been associated
with estrogen carcinogenesis (

Jefcoate et al., 2000

). Normal

human breast and human breast tumor tissues are known
to express CYP1B1, producing carcinogenic 4-hydroxy
estrogen. Inhibition of CYP1B1 aﬀects the production of
mutagenic estrogen 3,4-catechols (

Roberts et al., 2004

Xenobiotic responsive elements (XRE) are cis-acting

enhancer elements located in the promoter regions of xeno-
biotic responsive genes, which include genes encoding for
CYP1A1 and 1B1. The expression of these xenobiotic
responsive genes can be regulated through pathway involv-
ing aryl hydrocarbon receptor (AhR), which is a cytosolic
protein that can be activated by PAH. The activated
AhR then translocates to the nucleus, dimerizes with

Table 2 (continued)

Phase II

Flavonoids

Eﬀects

References

Morin

"Activity

Tanaka et al. (1999)

"Activity

Kawabata et al. (1999)

Pinostrobin

"Activity

Fahey and Stephenson (2002)

Kaempferol

"Activity

Galangin

"Activity

Quercetin

"Activity

Silymarin

"Activity

Kohno et al. (2002)

Genistein

"Activity

Yannai et al. (1998)

"Activity and mRNA expression

Wang et al. (1998)

"Activity

Diadzein

"Activity

Yannai et al. (1998)

"Activity

Flavone

"Activity

Siess et al., 1996

SULT1A1

Fisetin

#Activity

(P-PST or TS-PST)

Galangin

#Activity

Quercetin

#Activity

#Activity

#Activity

Catechin

#Activity

Equol

#Activity

Flavone

#Activity

Myricetin

#Activity

Kaempferol

#Activity

Chrysin

#Activity

Apigenin

#Activity

Genistein

#Activity

#Activity

#Activity

Diadzein

#Activity

#Activity

SULT1A3

Baicalein

#Activity

SULT1E1

Equol

#Activity

Hodek et al., 2002; Kohn et al.,

Quercetin

#Activity

Otake et al. (2000)

#Activity

Ohkimoto et al. (2004)

Genistein

#Activity

Ohkimoto et al. (2004)

Diadzein

#Activity

Ohkimoto et al. (2004)

#Activity

Wong and Keung (1997)

192

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

AhR nuclear translocator (ARNT), and interacts with
XRE (

Kronenberg et al., 2000

). Induction of CYP1A by

ﬂavonoids proceeds by various mechanisms, including the
direct stimulation of gene expression via speciﬁc receptor(s)
and/or CYP protein, or mRNA stabilization (

Lin and Lu,

1998; Shih et al., 2000

). Some ﬂavonoids induce CYPs

through binding to AhR, a ligand-activated transcription
factor, by acting in a similar way as 2,3,7,8-tetrachlo-
rodibenzo-p-dioxin (TCDD) (

Kohn et al., 2001

). Gener-

ally, substrates for AhR are planar aromatic compounds
with few bulky substituent groups. That might partly
explain the activity of ﬂavonoids, which have similar pla-
nar structures as AhR (

2001; MacDonald et al., 2001

). Other ﬂavonoids have been

shown to directly inhibit CYP1A1 activty (

Doostdar et al.,

2000; Tsyrlov et al., 1994; Wen et al., 2005

), commonly

demonstrated to be a competitive-type of inhibition, and
to aﬀect CYP1A1 transcription (

Ciolino et al., 1999; Kang

et al., 1999

The most abundant ﬂavonoids, ﬂavonols quercetin and

kaempferol, are both dietary ligands of the AhR, but they
exert diﬀerent eﬀects on CYP1A1 expression (

Ciolino

et al., 1999

). Treatment of MCF-7 cells with quercetin

resulted in a concentration- (

A) and time (

dependent increase in the amount of CYP1A1 mRNA.
The activity of enzyme CYP1A1 was measured by a 7-eth-
oxyresoruﬁn-O-dealkylation (EROD) assay. The increase
in EROD activity followed the increase in CYP1A1 mRNA.
Quercetin also causes a concentration- (

C) and time-

(

D) dependent increase in CYP1A1 enzyme activity

in MCF-7 cells (

). Kaempferol, by itself,

does not aﬀect CYP1A1 expression (

A), but it can

interact with the AhR, and act as an antagonist of TCDD-
induced CYP1A1 transcription (

Despite the structural similarity between quercetin and
kaempferol, their diﬀerential eﬀects might be due to the
absence of an additional hydroxy group on the B-ring of
kaempferol (

Fig. 3

), preventing it from achieving an optimal

ﬁt into the binding site on AhR to produce transcriptional
activation. The binding of kaempferol may block the bind-
ing of AhR ligands, and thus inhibit the activity of other
ligands such as TCDD (

). The importance

of the ortho-orientation of the hydroxyl group on the B ring
was tested by comparing the eﬀects of quercetin with a 3

substitution pattern and morin (2

-substitution pattern)

(

Tsyrlov et al., 1994

). 7-Methoxyresoruﬁn-O-dealkylation

(MROD) activities of mouse c-DNA-expressed CYP1A1
and 1A2, and human CYP1A2 were more eﬀectively inhib-
ited with quercetin than with morin (

Tsyrlov et al., 1994

Orientation of the methyl group can also inﬂuence inhibition
(

Tsyrlov et al., 1994

). Quercetin, a component in St. JohnÕs

wort inhibits CYP1A2 (

). St. JohnÕs wort

(Hypericum perforatum) extracts are commercially available
preparations used in the treatment of depression. Quercetin
inhibits phenacetin O-deethylase activity (mediated by
CYP1A2) in a recombinant CYP1A2 enzyme preparation
with an IC

value of 7.5 lM. Another ﬂavonoid in St.

JohnÕs wort, I3,II8-biapigenin was also shown to be a

Fig. 1. Flavonoids that block or suppress multistage carcinogenesis. Carcinogenesis is initiated with the transformation of the normal cell into a mutant
cell. These cells undergo tumor promotion into benign tumor cells, which progress to malignant cells. Flavonoids can interfere with diﬀerent steps of this
process. Some ﬂavonoids (for example, kaempferol, diosmetin, theaﬂavin, and biochanin A) can inhibit the metabolic activation of the procarcinogens to
their ultimate electrophilic species by phase I enzymes (predominantly CYPs), or their subsequent interaction with DNA. Therefore these agents block
tumor initiation (blocking agents). Alternatively, dietary ﬂavonoids (for example, naringenin, quercetin, biochanin A, and prenylchalcones) can stimulate
the detoxiﬁcation of carcinogens by inducing phase II enzymes, leading to their elimination from the body. Flavonoids such as genistein and EGCG
suppress the later steps (promotion and progression) of multistage carcinogenesis (suppressing agents) by aﬀecting cell cycle, angiogenesis, invasion, and
apoptosis. Adapted from (

Chen and Kong, 2004

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

193

potent, competitive inhibitor of CYP1A2 activity (IC

3.7 lM) (

Galangin, a ﬂavonol found in honey, is a potent inhibi-

tor of CYP1A1 activity, as measured by inhibition of
EROD activity, in intact cells and in microsomes isolated
from dimethylbenz[a]anthracene (DMBA) treated cells
(

Ciolino and Yeh, 1999

). The eﬀect is dose-dependent. Gal-

angin is a non-competitive inhibitor of CYP1A1 activity
(

Ciolino and Yeh, 1999

). Galangin increased the level of

CYP1A1 mRNA, indicating that it may be an agonist of
the AhR, but it inhibited the induction of CYP1A1 mRNA
by DMBA or by 2,3,5,7- TCDD (

Ciolino and Yeh, 1999

Galangin also inhibited the DMBA- or TCDD-induced
transcription of a reporter vector containing the CYP1A1
promoter. Thus, galangin is a potent inhibitor of DMBA
metabolism and an agonist/antagonist of the AhR (

Ciolino

and Yeh, 1999

). Similarly, quercetin causes an increase in

the level of CYP1A1 mRNA (

), whereas

it signiﬁcantly inhibits benzo(a)pyrene (B[a]P)-induced
CYP1A1 mRNA and protein expression in human Hep
G2 cells (

Kang et al., 1999

). Galangin is a very potent

CYP1A2 inhibitor, too (

Zhai et al., 1998

). It showed the

mixed-type inhibition, indicating that this compound can
compete for substrate binding at the active site and also
may bind to a region that does not participate directly in
substrate binding (

Zhai et al., 1998

Activator protein-1 (AP-1) may be involved in the regu-

lation of human CYP1A2 which contains two AP-1 bind-
ing sites (

Shih et al., 2000

Shih et al. (2000)

developed a

cell line T2Luc, which is a HepG2-derived cell line stably
integrated with a region of the human CYP1A2 5

-ﬂanking

gene containing two AP-1 binding sites linked to the thymi-

Fig. 2. Eﬀect of quercetin or kaempferol on expression of mRNA CYP1A1 and CYP1A1 enzyme activity. (A) Concentration response of CYP1A1
mRNA to quercetin (hatched bars) and kaempferol (open bars). MCF-7 cells were treated with quercetin the indicated concentration of quercetin. The
amount of CYP1A1 was normalized to the GPDH level. nd, not determined. (B) Time course of CYP1A1 mRNA increase caused by quercetin. MCF-7
cells were treated with 0.5 lM quercetin for the times indicated. The amount of CYP1A1 mRNA was normalized to GPDH levels. (C) The activity of
CYP1A1 in intact MCF-7 cells was determined by EROD assay. Cells were treated with the indicated concentrations of quercetin (j) or kaempferol ()
for 48 h. (D) Cells were treated with 5 lM quercetin for the times indicated. Reproduced from (

Mukhtar et al., 1992; Wang

) with permission from the Biochemical

Society.

194

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

dine kinase promoter-driven ﬁreﬂy luciferase reporter gene.
Green tea extracts (GTEs) inhibited the phorbol 12-o-
tetradecanoate 13-acetate (TPA)-induced AP-1 transcrip-
tional activation of a human CYP1A2 enhancer element,
while quercetin enhanced this activity. Green tea (Camellia
sinensis) and its extracts are rich source of catechins, a class
of ﬂavonoids (

Shih et al., 2000

). Catechins are the main

compounds in green tea; they consist of (

)-epicatechin,

(

)-epicatechin-3-gallate

(ECG),

(

)-epigallocatechin,

and

(

)-epigallocatechin-3-gallate

(EGCG)

(

Graham,

1992

). The mechanism of inhibition may be due to changes

in the composition of the AP-1 complex upon treatment of
cells with GTEs. An important and unexpected ﬁnding is
that GTEs themselves increased binding of nuclear pro-
teins to the AP-1 site to the same extent as TPA, yet did
not activate transcription of the luciferase reporter gene.
Unlike TPA, which induces AP-1 binding as early as 6 h
after its addition, GTE-induced AP-1 binding did not
occur until after 6 h, suggesting that other upstream signal-
ing pathways may also be targeted by this dietary agent
(data not shown) (

Shih et al., 2000

). In some in vitro exper-

iments, tea polyphenols had an inhibitory eﬀect on micro-
somal CYP enzyme system (

et al., 1988

). In contrast, long-term consumption of green

tea increases CYP1A1 and 1A2 activities in rats (

Liu

). The contents of CYPs (measured by the

method of Omura and Sato (

Heﬀernan and Winston,

)) in the livers of male rats in high dose (tea polyphe-

nols 833 mg/kg/day, intragastrically, for 6 months) groups
were signiﬁcantly increased (

Liu et al., 2003

Theaﬂavins in black tea are catechins generated by oxi-

dation of ﬂavanols during the fermentation of fresh tea leaf
(

Graham, 1992

). It has been shown to have antiprolifera-

tive and anticarcinogenic activities (

Yang et al., 1997

Male rats received intragastric administration of theaﬂav-
ins (20 mg/kg) for four weeks, and the liver and intestine
mucosa samples were obtained. Theaﬂavin treatment
markedly suppressed the CYP1A1 EROD activity in the
intestine (

Catterall et al., 2003

). However, there was no

eﬀect on any CYP activity in the liver (

Catterall et al.,

2003

); these results contrast with those from another study

where theaﬂavins suppressed CYP1A1 activity in hepatic
cell cultures (

Feng et al., 2002

). This discrepancy is pro-

bably due to the poor bioavailability of theaﬂavins as a
result of poor absorption and/or extensive pre-systemic
metabolism.

reported the in vitro inhibition

of cDNA-expressed human CYP1A1, CYP1B1, and
CYP1A2 by ﬂavonoids from hops (Humulus lupulus) (

Hen-

derson et al., 2000

). At 10 lM, the prenylated chalcone,

xanthohumol, almost completely inhibited the EROD
activity of CYP1A1. At the same concentration, other
hop ﬂavonoids decreased the EROD activity by 27.0–
90.8%. At 10 lM, xanthohumol completely eliminated
CYP1B1 EROD activity, whereas the other hop ﬂavonoids
showed varying degrees of inhibitory action ranging from
1.8% to 99.3% (

Henderson et al., 2000

). The most eﬀective

inhibitors of CYP1A2 acetanilide 4-hydroxylase activity
were the two prenylated ﬂavonoids, 8-prenylnaringenin
and isoxanthohumol, which produced >90% inhibition
when added at concentrations of 10 lM. CYP1A2 meta-
bolism of the carcinogen aﬂatoxin B1 was also inhibited
by isoxanthohumol and 8-prenylnaringenin as shown by
decreased appearance of dihydrodiols and aﬂatoxin M1
as analysed by HPLC (

Henderson et al., 2000

Baicalein, a ﬂavone extracted from the root of the Scu-

tellaria species, is a strong competitive inhibitor of EROD
activity induced by DMBA in MCF-7 cells (

Chan et al.,

). Baicalein can reduce the CYP1A1/1B1 mRNA

expression induced by DMBA, and the eﬀect on mRNA
abundance of CYP1A1 was greater than that of CYP1B1.
An XRE-luciferase gene reporter assay also indicated that
AhR transactivation was suppressed (

Chan et al., 2002

5,7-Dimethoxyﬂavone (DMF) is a major constituent of

the leaves of a Malaysian Piper species. DMF reduced
CYP1A1 EROD activity of the HepG2 cells virtually down
to zero (

). It also inhibited benzo(a)pyrene

(B[a]P)-induced DNA binding (

). B[a]P, a

major PAH procarcinogen, induces CYP1A1 protein and
its catalytic activity as well as CYP1A1 mRNA; DMF
clearly inhibits this transcriptional activation by decreasing
CYP1A1 catalytic activity as well as protein and mRNA
expression. DMF can also directly inhibit CYP1A1 protein
catalytic activity, as determined with recombinant protein
(

). It is remarkable that two so seemingly

Fig. 3. Structures of quercetin, kaempferol and 2,3,7,8-tetrachlo-
rodibenzo-p-dioxin (TCDD).

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

195

similar compounds as DMF and chrysin can have such dif-
ferent eﬀects, i.e. the 5,7-dimethoxy compound is a potent
inhibitor of the CYP1A1 protein whereas the 5,7-dihy-
droxy compound is a potent inducer (

Thus, methylation of ﬂavonoids seems to be an important
feature determining enzyme inhibitory properties.

The ﬂavone diosmetin is ingested (as its glucoside, dios-

min) in the commercially available drug ÔDaﬂon 500 mgÕ
(90% diosmin and 10% hesperetin). Both diosmin and dios-
metin are agonists of the AhR, causing a dose-dependent
increase in expression of CYP1A1 mRNA in MCF-7
human breast epithelial cancer cells (

Ciolino et al., 1998

);

however, diosmetin, but not diosmin, inhibits CYP1A1
activity in a non-competitive manner in microsomes iso-
lated from 7,12-dimethylbenz-[a]anthracene (DMBA) trea-
ted cells, as assayed by EROD activity (

Ciolino et al.,

Doostdar and co-workers (

Doostdar et al., 2000

) dem-

onstrated that six common ﬂavonoids present in citrus
juices at 70–200 mg/L are potent inhibitors of CYP1A1,
1A2 and 1B1 EROD activity in vitro (

Doostdar et al.,

). The ﬂavones acacetin and diosmetin were more

potent inhibitors than the ﬂavanones eriodictyol, hespere-
tin, homoeriodictyol and naringenin. This is probably
due to the presence of the reduced 2,3-bond in the C ring
of the structure in the ﬂavones (

Table 1

) (

Dai et al.,

et al., 1994; Zhai et al., 1998

). Potent inhibition of CYP1A1 and 1A2 enzyme activ-

ity by ﬂavone has also been reported in other studies (

Lee

). Flavone inhibits CYP1A1

and CYP1A2, with a 2-fold greater potency toward the lat-
ter (

Zhai et al., 1998

Tangeretin, a polymethoxylated ﬂavone present in large

amounts in citrus fruits, is a potent inhibitor of CYP1A2
EROD activity, with IC

of 0.8 lM in rat liver micro-

somes and 16 lM in human microsomes (

Obermeier

et al., 1995

). The inhibition was competitive with a inhibi-

tion constant (K

) value of 68 nM (

Obermeier et al., 1995

The ﬂavanone hesperetin is a selective substrate of

human CYP1A1 and CYP1B1 in the lymphoblastoid cell
line AHH-1, and it is a competitive inhibitor of CYP1B1
(

Fig. 4

) (

Doostdar et al., 2000

). Hesperidin, the glycoside

of hesperetin, is the major ﬂavonoid in orange juice. In vivo
studies have demonstrated that grapefruit juice consump-
tion increased the plasma half-lives of drugs such as caf-
feine. This eﬀect was attributed to inhibition of CYP1A2
by naringin, the major ﬂavonone in grapefruit juice (

Fuhr

et al., 1993

Biochanin A, the red clover (Trifolium pretense) isoﬂav-

one (

Chan et al., 2003

), and the soybean isoﬂavone geni-

stein (

Chan and Leung, 2003

) are eﬀective inhibitors of

DMBA-induced DNA damage in MCF-7 cells by inhibit-
ing CYP1A1 and CYP1B1. Both isoﬂavones could reduce
xenobiotic-induced CYP1A1 and 1B1 mRNA expression
through interference with XRE-dependent transactivation.
Enzyme kinetic studies also indicated that biochanin A
inhibits CYP1A1 and 1B1 with K

values of 4.00 and

0.59 lM, respectively (

Chan et al., 2003

), and genistein is

an eﬀective inhibitor of recombinant human CYP1A1
and CYP1B1 with K

values of 15.35 and 0.68 lM, respec-

tively (

Chan and Leung, 2003

). Extrahepatic CYP1B1

catalyzes the O-demethylation of biochanin A and formo-
nonetin to produce genistein and daidzein, respectively,
which inhibit CYP1B1 (

Roberts et al., 2004

). Inhibition

of CYP1B1 EROD activity by genistein was primarily
non-competitive (K

of 1.9 lM), and daidzein exhibited

mixed, but predominantly non-competitive inhibition of
CYP1B1 EROD activity (K

of 3.7 lM) (

Roberts et al.,

). Therefore, biochanin A and/or formononetin may

exert anticarcinogenic eﬀects directly by acting as compet-
itive substrates for CYP1B1 or indirectly through their
metabolites daidzein and genistein, which inhibit CYP1B1
(

Roberts et al., 2004

Genistein and equol inhibit CYP1A2 in liver micro-

somes from b-naphthoﬂavone-induced mice with IC

val-

ues of 5.6 mM and 1.7 mM, respectively (

Helsby et al.,

). Using human CYP1A2 from a speciﬁc expression

system, non-competitive inhibition was seen with both iso-
ﬂavones. This inhibition oﬀers a possible explanation for
the chemopreventive eﬀects of genistein in animals, but
inhibition of CYP1A2 is not likely to be achieved from
the concentration of genistein present in the human diet
(

Helsby et al., 1998

Dietary soy isoﬂavone containing 155 mg/g of genistein,

127 mg/g of daidzein, and other minor isoﬂavones had no
eﬀect on the hepatic mRNA abundance of CYP1A1 and
1A2 in rats, determined by real-time quantitative RT-
PCR (

Kishida et al., 2004

). This indicates that dietary iso-

ﬂavones do not induce CYPs in either the transcriptional
step or through post-transcriptional mRNA stabilization
(

Kishida et al., 2004

). Another in vivo study has reported

that genistein and equol only aﬀected the protein content
or activity of CYP1A1 and 1A2 following the administra-

Fig. 4. Lineweaver-Burk plot of EROD activity of lymphoblastoid
expressed human CYP1B1 (3.7 pmol) without hesperetin (d) and in
reactions containing 0.01 lM (j) or 0.02 lM hesperetin (m). Reproduced
from (

Doostdar et al., 2000

), with permission from Elsevier Science Inc.

196

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

tion of daily intraperitoneal injections to mice, when doses
of 40 mg/kg or higher of these isoﬂavones were adminis-
tered (

Helsby et al., 1997

Daidzein, a principal isoﬂavone in soybean, can inhibit

CYP1A2 activity and alter the pharmacokinetics of the-
ophylline in healthy volunteers (

Peng et al., 2003

). Theoph-

ylline is a bronchodilator with a narrow therapeutic index
(5–20 mg/L), and it is primarily eliminated by hepatic
metabolism mediated by CYP1A2. In the study by

Peng

et al. (2003)

, a single dose of 100 mg theophylline was taken

on day 3. Thereafter, one group received 200 mg daidzein
twice daily for 10 days, and the other group received pla-
cebo. On day 12, the test group received 200 mg daidzein
with 100 mg theophylline. Comparing the kinetic parame-
ters of theophylline determined on day 1 (without co-med-
ication) with those determined on day 12 (10-day daidzein),
the AUC, C

max

, and t

1/2

were signiﬁcantly increased (

Peng

Novak and Woodcroft, 2000

Genistein as well as combination of three (genistein,

daidzein, and glycitein) to ﬁve (plus biochanin A and for-
mononetin) isoﬂavones inhibited tamoxifen a-hydroxyl-
ation in female rat liver microsomes in vitro via
inhibition of CYP1A2 (

Chen et al., 2004

). The inhibition

of a-hydroxylation by genistein was mixed-type with a K

value of 10.6 lM (

Chen et al., 2004

). a-Hydroxytamoxifen

and its sulfate conjugate are thought to be responsible for
DNA adduct formation (

Umemoto et al., 2001

). Thus, gen-

istein and its isoﬂavone analogs have the potential to
decrease the side eﬀects of tamoxifen through metabolic
interactions that inhibit the formation of a-hydroxylation
(

Chen et al., 2004

Chalcones belong to the ﬂavonoid family and, naturally

occurring chalcones are predominantly metabolized by
hydroxylation (

De Vincenzo et al., 2000

). Among them,

-hydroxyl substituted chalcones have attracted much

attention, because they have biologically active properties,
such as prevention of platelet aggregation (

Lin et al., 1997

)

and LPS-induced septic shock (

Batt et al., 1993

) in animal

models. A XRE-luciferase reporter assay indicated that 2

hydroxychalcone was most eﬀective among ﬁve hydroxy-
chalcones in reducing CYP1A1 and 1B1 expression
through the disruption of XRE-transactivation (

Wang

et al., 2005

). The inhibition on CYP1A1 was competitive

and that of CYP1B1 was non-competitive. A decrease in
DMBA-DNA covalent binding was demonstrated in cul-
tures co-treated with 2

-hydroxychalcone and DMBA

(

Wang et al., 2005

). On the other hand, 2-hydroxychalcone

showed diﬀerent eﬀects; it inhibited CYP1A1 and 1B1
activities in recombinant microsomal preparation but
potentiated their expression in MCF-7 cells (

Wang et al.,

2005

). These ﬁndings suggested that the position and num-

ber of hydroxyl groups in hydroxychalcone might aﬀect the
CYP1 enzyme inhibition and gene expression.

A series of ﬂavonoids isolated from a resin of the tree

Dracaena cinnabari Balf. inhibited CYP1A activity in
hepatic microsomes isolated from TCDD treated mice
(

Machala et al., 2001

). The resin has been used in folk

medicine as an antiseptic and antiulcerative remedy. The
inhibitory potencies decreased in the following order:
galangin

> 7-hydroxyﬂavone > quercetin,

chrysin,

hydroxychalcone > 4,6-dihydroxychalcone > homoisoﬂav-
anes (

Machala et al., 2001

2.2. CYP2E

CYP2E1, an ethanol-inducible enzyme, is important in

the ﬁeld of toxicology and carcinogenesis, and it also has
a role in drug metabolism (

Guengerich et al., 1991

). For

example,

following

overdose

acetaminophen

CYP2E1 converts acetaminophen to toxic quinones, which
is responsible for the initiation of centrilobular liver toxic-
ity (

Lindros et al., 1990

). The activity of CYP2E1 is known

to be mainly regulated by post-transcrpitional protein sta-
bilization, but the contribution of the transcriptional step is
also signiﬁcant (

Genistein and equol, isoﬂavones in soy products, inhib-

ited p-nitrophenol (CYP2E1 substrate) metabolism in liver
microsomes from acetone-induced mice with IC

values of

approximately 10 mM and 560 lM, respecively (

Helsby

et al., 1998

). The 5-hydroxyl group and the 2,3-double

bond as well as hydroxyl groups in the ﬂavonoid B ring
may be essential for inhibition of aryl hydroxylation by
CYP2E1 (

Helsby et al., 1998

Theaﬂavins, catechins in black tea, decreased the protein

level of CYP2E1 in rat intestinal microsomes after oral
intake for four weeks (

Catterall et al., 2003

). Theaﬂavins

have been shown to antagonize the carcinogenicity of
nitrosamines in mice (

Shukla and Taneja, 2002

) and to

have antimutagenic activity (

Apostolides et al., 1997

Silybin (also known as silibinin or silybinin) is the main

component of silymarin, an extract from milk thistle. Sily-
bin was able to inhibit p-nitrophenol hydroxylation via
CYP2E1 in human liver microsomes (

Zuber et al., 2002

It displayed dose-dependent inhibition of enzyme activity
with IC

values in the micromolar range. However,

plasma concentrations of the individual ﬂavonolignans fol-
lowing dietary uptake do not exceed 0.5 lM (

Zuber et al.,

1997; Evans, 2000; Zhang and Benet, 2001

), so inhibition of CYP2E1 by ﬂavonolignans in the

diet may be unlikely.

2.3. CYP3A4

CYP3A4 is the largest subfamily of CYP enzymes

expressed in the human liver and gastrointestinal tract. It
is involved in the metabolism of 50% of therapeutic agents
as well as in the activation of toxic and carcinogenic sub-
stances. It has been reported that grapefruit juice alters
the pharmacokinetics of various drugs, including cyclo-
sporine (

Yee et al., 1995

), midazolam (

Kupferschmidt

et al., 1995

), dihydropyridine-type calcium channel block-

ers (

Bailey et al., 1994

), lovastatin (

Kantola et al., 1998

simvastatin (

Lilja et al., 1998

), ethinylestradiol (

Weber

et al., 1996

), and triazolam (

Hukkinen et al., 1995

). The

major mechanism for grapefruit juice-drug interactions is

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

197

thought to be due to the inhibition of intestinal CYP3A4
by ﬂavonoids in grapefruit juice (

Ameer and Weintraub,

), although the

inhibition of intestinal p-glycoprotein by ﬂavonoids has
also been observed. Particular interest has focused on the
inhibitory eﬀects of naringin (naringenin-7-rhamnogluco-
side), the major ﬂavonoid in grapefruit juice (>200 mg/
L), on the activity of intestinal CYP3A4 (

Ameer and Wein-

traub, 1997; Bailey et al., 2000; Evans, 2000

). Naringin is

partially metabolized by gastrointestinal bacteria to form
the ﬂavanone, naringenin (

Bailey et al., 2000

). Naringenin

exerts an inhibitory eﬀect on intestinal CYP3A4 within
30 min and impairs the metabolism of the calcium channel
blockers, felodipine, nitrendipine, nisoldipine, and verapa-
mil, when co-administered with grapefruit juice (

Fuhr,

Veronese et al. (2003)

demonstrated that consump-

tion of large amounts of grapefruit juice (double strength
three times daily for 3 days) inhibits both intestinal and
hepatic CYP3A4 activity, as quantiﬁed by the erythromy-
cin breath test and oral midazolam pharmacokinetics in
healthy male volunteers (

Veronese et al., 2003

). Consump-

tion of typical amounts (single strength) of grapefruit juice
caused less inhibition of intestinal CYP3A4 (

Veronese

Another clinically important inhibitor of CYP3A4 is St.

JohnÕs wort (

et al., 2001; Jurgenliemk and Nahrstedt, 2002; Nahrstedt
and Butterweck, 1997

). St. JohnÕs wort (Hypericum per-

foratum) is one of the most commonly used herbal medi-
cines in the United States. Its major constituents include
ﬂavonols, ﬂavonol glycosides and biﬂavones (

Barnes

using cDNA-

expressed enzymes showed that hyperforin was a potent
competitive inhibitor of CYP3A4 activities. The ﬂavonoid
compound I3, II8-biapigenin in St. JohnÕs wort was also
shown to be a potent, competitive inhibitor of CYP3A4
activity (

). Quercetin in St. JohnÕs wort was

found to inhibit CYP3A4 as well (

Obach, 2000; Zou

Moore et al., 2000; Wentworth et al., 2000; Zhou et al.,
2003

). However, enzyme induction with St. JohnÕs

wort has also been reported. In vitro studies using primary
cultures of human hepatocytes have demonstrated that St.
JohnÕs wort extract is a potent inducer of CYP3A4, and the
responsible component is hyperforin (

Goodwin et al., 2001;

). Hyperforin is a potent ligand (K

of 27 nM) for the

pregnane X receptor (

Moore et al., 2000

), which is an

orphan nuclear receptor regulating expression of CYP3A4,
as well as other enzymes and transporters (

Durr et al.,

2000; Moore et al., 2000; Wentworth et al., 2000

Animal studies using probe drugs have provided addi-

tional evidence that St. JohnÕs wort is a potent modulator
of various CYP enzymes, and the induction of CYPs
depends on the dosing regimen. Mice studies have indi-
cated that short-term treatment (four consecutive days)
of St. JohnÕs wort extract (435 mg/kg/day) or hyperforin
(10 mg/kg/day) did not alter erythromycin N-demethylase
(CYP3A) activity (

Bray et al., 2002

). In contrast, adminis-

tration of St. JohnÕs wort extract (140 or 280 mg/kg/day) to

mice for three weeks resulted in a two-fold increase in the
CYP3A activity (

Bray et al., 2002

). The protein level of

CYP3A was also increased six-fold (

Bray et al., 2002

The short-term treatment may not activate the pregnane
X receptor. Similarly, human studies indicated that long-
term ingestion (14-days) of St. JohnÕs wort administration
signiﬁcantly induced the activity of CYP3A4 as measured
by changes in alprazolam pharmacokinetics (

Markowitz

et al., 2003; Wang et al., 2001

), but short-term administra-

tion had no eﬀect on CYP3A4 activity (

Markowitz et al.,

Beckmann-Knopp et al. (2000)

Silymarin, milk thistle extract, is a naturally occurring

mixture of ﬂavonolignans (silibinin A, silibinin B, silichri-
stin, silidianin, taxifolin) (

Simanek et al., 2001

). It has been

used to treat liver diseases for hundreds of years. It is
known to protect cardiomyocytes against doxorubicin-
induced oxidative stress, due mainly to its radical scaveng-
ing and iron chelating potency (

Chlopcikova et al., 2004

Silymarin signiﬁcantly decreases CYP3A4 activity (

Fig. 5

)

in primary cultures of human hepatocytes (

Venkatarama-

nan et al., 2000

). The reason for the reduced activity of

CYP3A4 in silymarin-treated cells is not clear at this time.

used two model substrates,

denitronifedipine and erythromycin, and showed that sily-
bin, the major constituent of silymarin, inhibits oxidation
of denitronifedipine in a non-competitive manner, whereas
the eﬀect on erythromycin demethylation involves activa-
tion at low and inhibition at high silybin concentrations.
CYP3A4 shows very complex behavior, in that it may have
multiple conformations with distinct substrate speciﬁcities
(

Koley et al., 1996

). Furthermore, cooperativity in oxida-

tions catalyzed by the enzyme have been described and
attributed to the existence of more than one substrate bind-
ing site on the CYP3A4 molecule (

Hosea et al., 2000; Ueng

Treatments

RIF

Sily 0.1 mM Sily 0.25 mM

Form

on of 6-beta tes

tos

terone

(nmol/min/mg protein)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

Fig. 5. Eﬀect of silymarin on CYP3A4 activity in primary cultures of
human hepatocytes. The formation rate of 6b (OH) testosterone was
measured in primary cultures of human hepatocytes that were not treated
or treated for 48 h with rifampin, or treated for 48 h with silymarin (0.1
and 0.25 mM). *, indicates signiﬁcantly diﬀerent from untreated cells.
Reproduced from (

Venkataramanan et al., 2000

), with permission from

the American Society for Pharmacology and Therapteutics.

198

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

et al., 1997

). CYP3A4 activity, assessed by oxidation of

nifedipine, was also decreased by silybin in human liver
microsomes (

Zuber et al., 2002

). The inhibition was dose-

dependent with IC

values in the micromolar range.

CYP activities were unaﬀected by silybin at physiologically
relevant lower concentrations (

Zuber et al., 2002

). Addi-

tionally, milk thistle had no signiﬁcant eﬀect on CYP3A4
in vivo. Healthy female and male subjects have been
administered milk thistle (175 mg, twice daily, standardized
to 80% silymarins) for 28 days with no signiﬁcant changes
on phenotype ratios (1-hydroxymidazolam/midazolam) for
CYP3A4 (

Gurley et al., 2004

). After 21 days of milk thistle

extract administration (153 mg silymarin, 3 times daily), no
clinically signiﬁcant changes in the pharmacokinetics of
indinavir (a CYP3A4 substrate) were noted in humans
(

Piscitelli et al., 2002

). Another study of healthy volunteers

also found that 20 days of silymarin ingestion (160 mg 3
times daily) had no eﬀect on the pharmacokinetics of indi-
navir (

DiCenzo et al., 2003

). Administration of silymarin

(70 mg 3 times daily) for 28 days to healthy volunteers
had no eﬀect on the pharmacokinetics of aminopyrine or
phenylbutazone, two non-speciﬁc CYP probes (

Leber and

Knauﬀ, 1976

Green tea (Camellia sinensis) extract does not alter

CYP3A4 activity in healthy volunteers (

Donovan et al.,

). The probe drug alprazolam (2 mg, CYP3A4 activ-

ity)

was

administered

orally

healthy

volunteers

(n = 11) at baseline, and again after treatment with four
decaﬀeinated green tea (DGT) capsules/day for 14 days.
Each DGT capsule contained 211 ± 25 mg of green tea cat-
echins and <1 mg of caﬀeine. The plasma concentration of
the green tea ﬂavonoid, (

)-epigallocatechin gallate

(EGCG), reached 1.3 ± 1.8 lM, 2 h after DGT treatment.
There were no signiﬁcant diﬀerences in alprazolam phar-
macokinetics at baseline and after DGT treatment. In this
study, EGCG was bioavailable from this supplement
(

Donovan et al., 2004

). Therefore the lack of eﬀect of

DGT on CYP3A4 activity cannot be due to a lack of bio-
availability of the extract used in this study (

Donovan

et al., 2004

Genistein and diadzein, isoﬂavones present in soybeans,

were found to inhibit 3A4-mediated metabolism, but their
glycosides were inactive in human microsome preparations
(

Foster et al., 2003

). Genistein exhibited a similar inhibi-

tory activity against the human 3A isoform CYP3A7 (

Fos-

ter et al., 2003

). The results of this study were consistent

with the observations of Evans, indicating that daidzein
and genistein can inhibit oxidative metabolism (

Evans,

Kent et al., 2002; Zhou et al., 2004

Glabridin is a major ﬂavonoid in licorice (Glycyrrhiza

glabra). Kent et al. reported that the isoﬂavan glabridin
inactivated CYP3A4 in a time-, concentration-, and
NADPH-dependent manner, indicative of mechanism-
based inactivation (

Metabolism of glabridin by CYP3A4 resulted in the
destruction of the heme moiety, and the loss of the
CYP3A4-reduced carbon oxide spectrum and detectable

heme were correlated with glabridin concentrations (

Kent

). Glabridin may generate reactive intermediates

that result in heme fragmentation. Two hydroxyl groups on
the 2

and 4

position of the ﬂavonoid B ring of glabridin

seems to be required for CYP3A4 inactivation (

Kent

Miller et al., 1990; Zhou et al., 1996

The number of hydroxyl groups, as well as the position

of hydroxylation, also plays an important role in the inhib-
itory eﬀects of ﬂavones. Flavones having more hydroxyl
substitutions showed stronger inhibition of CYP3A4 activ-
ity than those with fewer hydroxyl groups (

Ho et al., 2001

In this study, the order of inhibitory potency was myrice-
tin > quercetin,

morin > kaempferol > apigenin > ﬂavone

(with 6, 5, 4, 3, 0 hydroxyl groups respectively). This was
conﬁrmed by the existence of a positive signiﬁcant correla-
tion (r = 0.89, p < 0.0005) between number of hydroxyl
groups and extent of CYP3A4 inhibition of these six com-
pounds. This ﬁnding is in agreement with a previous study
showing that the ability of ﬂavonoid compounds either to
inhibit or stimulate benzo(a)pyrene hydroxylation in
human liver microsomes was related to the presence or
absence of hydroxyl groups respectively (

Ho et al., 2001

2.4. CYP19

CYP19, also known as aromatase (11b-hydroxysteroid

dehydrogenase), is another member of the cytochrome
P450 enzyme superfamily. This enzyme represents a crucial
enzyme of estrogen biosynthesis; CYP19 converts andro-
stenedione and testosterone to estrone and estradiol,
respectively (

Zhu and Conney, 1998

). Increased expression

of aromatase has been observed in breast cancer tissue
(

). Since estrogens

cause cell proliferation and certain estrogen metabolites
are carcinogens, local expression of CYP19 has been corre-
lated with tumor initiation, promotion and progression.
Local regulation of aromatase by both endogenous factors
as well as exogenous medicinal agents inﬂuences the levels
of estrogen available for breast cancer growth (

Brueggeme-

ier et al., 2001

). Flavonoids and isoﬂavones are structurally

similar to the endogenous steroid hormone, estradiol, and
possess either estrogenic or antiestrogenic activities (

Brueg-

gemeier et al., 2001

), and they are known to be competitive

inhibitors of aromatase with respect to androgen sub-
strates. Flavonoids with antiestrogenic eﬀects have been
reported to be cancer preventive agents in breast cancer
(

Kumar et al., 2004

Flavones have previously been reported to be competi-

tive inhibitors of aromatase with respect to the androgen
substrate, with K

values at low micromolar concentrations

(

Adlercreutz et al., 1993; Campbell and Kurzer, 1993; Ibra-

him and Abul-Hajj, 1990; Kellis and Vickery, 1984; Pelis-
sero et al., 1996; Wang et al., 1994

). The binding

characteristics and the structural requirements necessary
for the inhibition of human aromatase by ﬂavonoids were
obtained using computer modeling and conﬁrmed by site-
directed mutagenesis (

Kao et al., 1998

). Flavonoids bind

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

199

to the active site of aromatase in an orientation in which
rings A and C mimic rings D and C of the androgen sub-
strate, respectively. Generally, ﬂavones and ﬂavanones
have higher aromatase inhibitory activity than isoﬂavones
and isoﬂavanones, which exhibit signiﬁcant binding aﬃni-
ties for the estrogen receptor (

Kao et al., 1998

). Based on

the study of a series of B ring substituted ﬂavanones with
a 7-methoxy group on the A ring, the structure-activity
relationships suggested that hydroxylation at position 3

and/or 4

are the optimal pattern of B ring substitution that

enhanced the anti-aromatase activity (

Pouget et al., 2002

Kao et al. showed that ﬂavones (chrysin, baicalein, and

galangin), ﬂavanone (naringenin) and isoﬂavones (biocha-
nin A) inhibit the activity of human aromatase in chinese
hamster ovary (CHO) cells, thus decreasing estrogen bio-
synthesis and circulating estrogen levels (

Kao et al.,

). Similarly, ﬂavone, chrysin, apigenin, naringenin,

and biochanin A inhibited human placental aromatase
(

Le Bail et al., 2000

). Isoﬂavone, equol was potent inhibitor

of the ovarian aromatase activity in rainbow trout and also
showed inhibitory eﬀect on human placental aromatase
(

Pelissero et al., 1996

). Daily treatment for 21 days with

natural supplements such as propolis and honey, contain-
ing chrysin, could block the conversion of androgens into
estrogens by inhibiting aromatase, with a consequent
increase of testosterone, eventually measurable in urine
samples (

Gambelunghe et al., 2003

Almstrup et al. have investigated several ﬂavonoids

using a pS2 mRNA essay for aromatase inhibition in
MCF-7 cells (

Almstrup et al., 2002

). Since aromatase con-

verts testosterone to 17a-estradiol, and pS2 mRNA is reg-
ulated by estrogen, aromatase activity can be measured by
diﬀerences in the expression level of the pS2 mRNA after
exposure to testosterone and the test compounds. Biocha-
nin A, formononetin, naringenin, and chrysin are aroma-
tase inhibitors at low concentrations (<1 lM) (

Almstrup

Jeyapaul and Jaiswal, 2000;

3. Eﬀect of bioﬂavonoids on phase II enzymes

Activation of phase II detoxifying enzymes, such as

UDP-glucuronyl transferase (UGT), glutathione S-trans-
ferase (GST), and NAD(P)H:quinone oxidoreductase
(QR) by ﬂavonoids results in the detoxiﬁcation of carcino-
gens and represents one mechanism of their anticarcino-
genic eﬀects (

Fig. 1

). The importance of induction of

Phase II metabolism in cancer prevention has been demon-
strated in studies of nrf-2 knockout mice; nrf-2 is a tran-
scription factor necessary for Phase II enzyme induction
(

Ramos-Gomez et al., 2001

). Nrf2 is normally localized

in the cytosol, where it is associated through protein-pro-
tein interactions with the chaperone Keap1 (

Itoh et al.,

1999

). The presence of an inducer disrupts the Keap1-

Nrf2 interactions, allowing Nrf2 to translocate to the
nucleus and bind to the antioxidant/electrophile response
element (ARE), in conjunction with small Maf proteins,
after activation (

Hayes and McMahon, 2001

). Nrf2 is

important for both the constitutive and inducible expres-
sion of several Phase II proteins, including QR, GST, c-
glutamylcysteine synthetase and UGT. In contrast to
normal mice, nrf2 knock-out mice are more susceptible
to benzo[a]pyrene carcinogenesis and are not protected
by Phase II inducers (

McMahon et al., 2001; Ramos-Gomez et al., 2001

Several papers have suggested that many of phase II

inducers have, or acquire by metabolism, a electrophilic
Michael reaction acceptor functionality (

Dinkova-Kost-

ova, 2002; Talalay et al., 1988

). Moreover, it was found

that the potency of inducers correlates with reactivity in
the Michael reaction (

Talalay et al., 1988

). The inducer

potency of Michael reaction acceptors is profoundly
increased by the presence of ortho- (but not other) hydro-
xyl substituent(s) on the aromatic ring(s) (

Dinkova-Kost-

ova et al., 2001

). Many ﬂavonoids contain Michael

reaction center(s) in their molecules, thus this characteristic
may be related to their eﬀects on phase II enzymes. Identi-
ﬁed dietary ﬂavonoids acting as phase II inducers include:
kaempferol (

Uda et al., 1997

), a ﬂavonoid present in high

amounts in kale; a ﬂavonoid fraction found in blueber-
ries/cranberries (

Bomser et al., 1996

); the ﬂavolignan silib-

inin (also known as silybin or silybinin) obtained from milk
thistle (

Zhao and Agarwal, 1999

3.1. UDP-glucuronyltransferase

Glucuronidation, catalyzed by the UDP-glucuronyl-

transferase family of enzymes, is a major metabolic path-
way of endogenous steroids, bile acids, drugs, and
carcinogens. UGTs have been divided into two families,
termed UGT1 and UGT2 (

Mackenzie et al., 1997

UGT1 enzymes mainly catalyse glucuronidation of exoge-
nous agents (drugs, pesticides, benzo[a]pyrene, etc.),
whereas UGT2 enzymes glucuronidate endogenous agents
(steroid hormones and bile acids). However exceptions
exist: for instance, human UGT1A1 can metabolize the
toxic heme breakdown product bilirubin, as well as cate-
chol estrogens, and ﬂavonoids (

King et al., 2001

). In

humans, glucuronidation capacity is prominently present
in the liver, but UGT activity toward bile acids, phenols,
and bilirubin is present in human intestinal, kidney, and
colon tissue (

Tukey and Strassburg, 2000

UGT1A1, 1A3, 1A4, 1A6, and 1A9 mRNA expression

is present in human livers. Of importance, UGT1A7 and
1A10 were discovered and cloned from gastric tissue and
UGT1A10 in biliary tissue, indicating these RNAs are
exclusively extrahepatic UGT1A gene products (

Tukey

and Strassburg, 2000

). UGT1A10 appears to be expressed

in all tissues of the gastrointestinal tract except liver. This is
signiﬁcant because UGT1A10 has one of the widest range
of substrate speciﬁcities of any of the UGTs, from small
phenolics to steroids, an indication that it may play a vital
role in most extrahepatic tissues for the glucuronidation of
endogenous and xenobiotic substrates (

Tukey and Strass-

burg, 2000

). The expression of UGT2 genes also follows

200

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

a tissue-speciﬁc pattern. The human olfactory UGT2A1,
which has been shown to be one of the more versatile of
the UGTs by recognizing all of the major classes of sub-
strates, is restricted in expression to olfactory tissue. One
could predict that UGT2A1 has evolved for the need to
serve as a ﬁrst line of metabolic defense for many sub-
stances that enter the body through the nasal mucosa. In
human liver, cDNAs have been identiﬁed for UGT2B4,
2B7, 2B11, and 2B15 (

Tukey and Strassburg, 2000

UGT2B7 transcripts are also found in intestine, esophagus,
brain, kidney, and pancreas. A signiﬁcant observation is
that UGT2B transcripts are abundantly expressed in ste-
roid-sensitive target tissues such as prostate and mammary
gland; UGT2B10, 2B11, 2B15, and 2B17 gene transcripts
have been identiﬁed in human prostate, and UGT2B11 is
also expressed in mammary gland tissue. The presence of
UGT2B17 may have a signiﬁcant impact on cancer of the
prostate gland by glucuronidating androgens and thus pro-
tecting this tissue from the carcinogenic actions of these
steroids (

Tukey and Strassburg, 2000

As reported by Walle et al., in human hepatoma HepG2

cells and the human intestinal cell line Caco-2, there was
a high level of induction of UGT1A1 by treatment with
the ﬂavone chrysin (

Walle et al., 2000

). Similarly, the

UGT1A1-mediated glucuronidation of quercetin was
greatly increased (

Galijatovic et al., 2001

). This induction

response was quite speciﬁc because UGT1A6, UGT1A9,
and UGT2B7 were not aﬀected by chrysin treatment.
Two of the ﬂavonoids that induced CYP1A1, galangin
and isorhamnetin, had no eﬀect on the UGT1A1 activity,
suggesting that the inducing eﬀect of UGT1A1 is not
related to the AhR (

Fig. 6

) (

Walle and Walle, 2002

The polymethoxyﬂavone tangeretin was the most potent

inhibitor of UGT1A1 catalyzed estradiol-3-glucuronida-
tion in human liver microsomes (IC

= 1 lM at 5 lM

estradiol concentration) (

Williams et al., 2002

). Naringenin

inhibited estradiol 3-glucuronidation to a similar extent at
all naringenin concentrations tested (5–100 lM) and there-
fore did not act as a competitive-type inhibitor (

Williams

van der Logt et al., 2003

). Flavone and quercetin were weak inhibitors

of estradiol 3-glucuronidation at the concentrations tested
(5–100 lM) (

Williams et al., 2002

). For chrysin, ﬂavanone,

nobiletin, and silymarin, the greatest inhibitory eﬀect on
estradiol 3-glucuronidation was at substrate concentrations
above 25 lM.

Flavone has been shown to induce hepatic 4-nitrophenol

(NP) UGT activity in rats at a concentration of 0.25%
(

Brouard et al., 1988

), 0.3% (

Canivenc-Lavier et al.,

1996

), and 1% (

) (w/w) in the diet

for 2 weeks. Van der Logt et al. quantiﬁed UGT enzyme
activity in hepatic and intestinal (proximal, mid and distal
small intestine and colon) tissue from male rats using NP
and 4-methlyumbelliferone as substrate (

van der Logt

Bu-Abbas et al., 1995; Sohn et al., 1994

). In their study, 1% (w/w) quercetin in the basal

diet for 2 weeks caused signiﬁcant increase of UGT enzyme
activity in liver and proximal and distal small intestine of
the rats. When determined using 2-aminophenol as the sub-

strate, long-term ingestion of green tea increases UGT
activity in rats (

and this induction is considered to contribute to the anti-
carcinogenic eﬀect of green tea. In contrast, three other
studies reported that quercetin has no eﬀect on hepatic
NP UGT activity (

Brouard et al., 1988; Canivenc-Lavier

et al., 1996; Siess et al., 1996

). This diﬀerence may be

caused by diﬀerent doses, food deprivation in animals or
other diﬀerences in experimental design. Since starvation
is known to rapidly decrease phase II enzyme activity (

Sieg-

ers et al., 1989

), an initial increase in UGT enzyme activity

may have disappeared during this food deprivation period.

found that UGT was induced by

selected ﬂavonoids (biochanin A, diadzein, formonone-
tin, genistein, prunetin, apigenin, galangin, kaempferol,
naringenin, and quercetin) at a concentration of 5 lM in
the prostate cancer cell line LNCaP. LNCaP cells were
exposed to each of the ﬂavonoids for 6 days. Those ﬂavo-
noids stimulated the activity of testosterone-UGT, which
conjugates testosterone to testosterone glucuronide. Bio-
chanin A was the most potent inducer of UGTs (10-fold
increase at 5 lM), with increased activity over the

Fig. 6. Induction of UGT1A1 (A) and CYP1A (B) in Hep G2 cells by
ﬂavonoids and classical AhR inducers. Api, apigenin; Chry, chrysin; Gal,
galanin; Iso, isorhamnetin; *, signiﬁcantly higher than control (DMSO),
p < 0.001; **, signiﬁcantly higher than control, p < 0.05. Reproduced from
(

Walle and Walle, 2002

), with permission from the American Society for

Pharmacology and Therapeutics.

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

201

concentration range of 0.5–50 lM (

Sun et al., 1998

). The

production and release of prostate speciﬁc antigen (PSA),
the prostatic tumor marker (

Catalona et al., 1991

), is tes-

tosterone dependent (

Montgomery et al., 1992

), and bioch-

anin A can signiﬁcantly decrease PSA concentrations,
likely due to increased glucuronidation of testosterone
(

Sun et al., 1998

). These studies suggest that the modula-

tion of hormone metabolism by ﬂavonoids may be impor-
tant in the prevention and treatment of prostate cancer
(

Sun et al., 1998

3.2. Glutathione-S-transferase (GST) and quinone
reductase (QR)

Flavonoids also contribute to the modulation of other

phase II detoxifying enzymes, such as GST and QR.
Human cytosolic GSTs are a family of dimeric biotransfor-
mation enzymes comprised of the four main classes; a, l, p,
and h (

Hayes and Pulford, 1995

). They catalyze the binding

of a large variety of electrophiles to the sulfydryl group of
glutathione, are involved in the detoxiﬁcation of (oxygen)
radicals, and have a main function in the binding and
transport of a wide variety of harmful compounds. GSTs
have a considerably important role in the detoxiﬁcation
of carcinogens (

Hayes and Pulford, 1995

). GSTs are pres-

ent in many species and tissues and also in relatively large
amounts in the epithelial tissues of the human gastrointes-
tinal tract (

Peters et al., 1991

). A signiﬁcant negative corre-

lation was demonstrated between GST enzyme activity and
tumor incidence in the mucosa along the human gastroin-
testinal tract, suggesting the importance of GSTs in cancer
prevention (

Peters et al., 1993

). QR prevents quinine redox

cycling and lowers levels of electrophilic quinines (

Kelly

et al., 2000

). Hence, the induction of GST and QR by

ﬂavonoids is possibly associated with cancer chemopreven-
tive eﬀects.

Genistein regulates estrogen receptor (ER)/ARE-depen-

dent gene expression in vitro (

Ansell et al., 2004

). ER reg-

ulation of a mouse GST Ya reporter gene was determined
in two cell lines in the presence of 1 lM of genistein. In
COS I cells expressing ERa and ERb, genistein repressed
GST Ya ARE-dependent gene expression (

Ansell et al.,

); however, treatment of C4-12-5 cells (ER-negative

breast cancer cell line derived from the MCF-7 cell line)
with genistein resulted in modest GST gene induction fol-
lowing transfection with ERa and ERb. This suggests that
the eﬀects of genistein on GST through ER/ARE signaling
are cell type speciﬁc (

Ansell et al., 2004

Flavone and 2

-amino-3

methoxyﬂavone induce the

nuclear translocation of a transcriptional factor—CCAAT/
enhancer-binding protein b (C/EBPb) and induce GSTA2
(a form) gene expression (

Fig. 7

) (

Kang et al., 2003

). Die-

tary administration of ﬂavone increases GST activities (a
and l isoforms) and the levels of glutathione in many tis-
sues of male Wistar rats (

Nijhoﬀ et al., 1995

). Long-term

ingestion of green tea extracts (GTEs) increases cytosolic
GST activity in female rats (

Maliakal et al., 2001

). On

the other hand, quercetin has been shown to eﬀectively
inhibit human placental GST (GSTP1-1), a subclass of
the GST family, in a time- and concentration-dependent
manner in vitro. GSTP1-1 activity is completely inhibited
following a 1 h-incubation with 100 lM quercetin or 2 h-
incubation with 25 lM quercetin. The inactivation mecha-
nism may involve the covalent modiﬁcation of cysteine
47 in GSTP1-1 by quercetin quinone or its quinone met-
hides (

van Zanden et al., 2003

Chang et al. (1997)

reported the isolation of a potent

QR inducer from the pantropical coastal shrub Tephrosia
purpurea, the chalcone, (+)-tephropurpurin. Hop ﬂavo-
noids (prenylchalcones and prenylﬂavanones) can induce
GST enzymes and QR in the mouse hepatoma Hepa1c1c7
cell line (

Miranda et al., 2000

). In contrast, the ﬂavanone,

naringenin, with no prenyl group, was ineﬀective in induc-
ing QR (

Miranda et al., 2000

). The hop chalcones, xan-

thohumol and dehydrocycloxanthohumol hydrate, also
induce QR activity in the AhR-defective mutant cell line,
Hepa1c1c7bp(r)c1 (

Miranda et al., 2000

). Dietary admin-

istration of morin to F344 rats led to signiﬁcant increases
in the activities of QR and GST in the liver, large bowel
and tongue and was protective against azoxymethane-
induced adenocarcinoma of the large intestine (

Tanaka

et al., 1999

), as well as against 4-nitroquinone 1-oxide-

induced tongue carcinogenesis (

Kawabata et al., 1999

The ﬂavonoid pinostrobin, present in honey and Thai gin-
ger (Boesenbergia pandurata) represents a potent inducer
of mammalian QR activity (

Fahey and Stephenson,

reported that a 2,3-double bond

Fig. 7. Immunoblot analysis of rGSTA2 in H4IIE cells treated with
ﬂavone. A representative immunoblot shows the level of rGSTA2 protein
in cells treated with PD98059 (3–30 lM) or ﬂavone (1–10 lM) for 24 h.
Each lane was loaded with 10 lg of cytosolic proteins. The relative
rGSTA2 level was assessed by scanning densitometry. Data represent the
mean ± SD with three separate experiments (signiﬁcant as compared with
control,

P < 0.05;

P < 0.01; control level = 1). Reproduced from (

Kang

), with permission of Oxford University Press.

202

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

in the C ring of ﬂavonoids is critical for the induction of
QR. Among the ﬂavonoids, the ﬂavonols are the most
eﬀective inducers of QR activity in Hepa1c1c7 cells
(kaempferol, galangin > quercetin > myricetin, apigenin
(a ﬂavone)) and ﬂavanols and ﬂavans are ineﬀective. The
2,3-double bond in the C ring may play a role as a
Michael reaction acceptor (

Dinkova-Kostova, 2002

). In

contrast to ﬂavone, ﬂavanone, which has a saturated
2,3-double bond in the C ring, cannot induce QR activity
(

Dinkova-Kostova, 2002

Silymarin can signiﬁcantly elevate GST and QR activi-

ties in the liver and colon of the rat. These eﬀects may be
related to the in vivo suppressive eﬀects of silymarin on
the occurrence of aberrant crypt foci, a putative precursor
lesions for colonic adenocarcinoma, and azoxymethane-
induced carcinoma (

Kohno et al., 2002

Genistein and daidzein were found to increase QR activ-

ities in Hepa1c1c7 cells (

Yannai et al., 1998

). The inhibition

of benzo[a]pyrene metabolite-DNA binding by genistein
may result from QR induction (

Lee et al., 1999

). Induction

of QR activity by genistein has been also observed in the
human colon cancer cell line Colo205 (

Wang et al.,

Glatt, 2000; Lewis et al., 2000; Yamazoe et al., 1999

). QR induction was further conﬁrmed by using

reverse transcription-polymerase chain reaction (RT-
PCR) techniques to measure mRNA expression. A signiﬁ-
cant correlation between the expression of QR mRNA and
the corresponding QR activity was observed (r = 0.76,
P < 0.001) (

Wang et al., 1998

). The single topical applica-

tion of 12-O-tetradecanoyl phorbol-13-acetate (TPA), a
well-known tumor promoter, down-regulated the level of
GST activity (47%) in mouse skin model (

Sharma and Sul-

tana, 2004

). Topical applications of soy isoﬂavones (mix-

ture of 33 mg of genistein and 67 mg of daidzein), 30 min
prior to the application of TPA prevented the decrease in
GST activity, in a dose dependent manner. Dietary geni-
stein and diadzein are also able to elevate the activities of
GST in the kidney and QR in the colon of female rats
in vivo (

Appelt and Reicks, 1999

Siess et al. (1996)

showed that dietary administration of

ﬂavone to male Wistar rats elevated the activities of phase
2 enzymes (GST and UGT) as well as phase I enzymes
(CYP 1A1/2 and 2B1/2) (

Siess et al., 1996

). Importantly,

the activities of CYP1A1/2 increased as early as 6 h after
the ﬁrst dose, reaching maximal induction after 4 days,
while the earliest elevation of the activities of CYP2B1/2
and phase II enzymes was observed 24 h after ﬂavone feed-
ing. This diﬀerence in the time course of induction may
suggest that the ultimate inducers of the CYP2B1/2 and
phase II enzymes are probably ﬂavone metabolites and
not the parent compounds themselves, or that the induc-
tion occurs through diﬀerent mechanisms for these diﬀerent
enzymes.

3.3. Sulfotransferases

The cytosolic sulfotransferases (SULTs) catalyze the sul-

fate conjugation of many hormones, neurotransmitters,

and xenobiotic compounds (

Coughtrie et al., 1998

). They

are involved in the Phase II detoxiﬁcation of xenobiotics,
as well as in the inactivation of endogenous compounds
such as steroid and thyroid hormones, catecholamines
and bile acids (

Coughtrie et al., 1998

). In contrast to this

protective function, sulfation is also a key step in the bio-
activation of a host of pro-mutagens and pro-carcinogens
(

). Sul-

fation activates carcinogens such as hydroxymethyl polycy-
clic aromatic hydrocarbons, allylic alcohols, benzylic
alcohols, and N-hydroxyarylamines, since their sulfate
esters are electrophiles that covalently bind to nucleic acids
and other macromolecules (

Glatt, 2000; Meerman et al.,

1994

Flavonoids have been suggested as potential chemopre-

ventive agents in sulfation-induced carcinogenesis (

Cought-

rie and Johnston, 2001; Eaton et al., 1996; Ghazali and
Waring, 1999; Tamura and Matsui, 2000; Walle et al.,
1995

). A number of ﬂavonoids exert inhibitory eﬀects on

sulfotransferase activity. A previous study has demon-
strated that ﬂavonoids (ﬁsetin, galangin, quercetin, myrice-
tin, kaempferol, chrysin, apigenin and genistein) represent
potent inhibitors of the P-form phenolsulfotransferase (P-
PST or SULT1A1)-mediated sulfation of acetaminophen
and minoxidol by human liver cytosol (

Eaton et al.,

1996

). Quercetin, ﬁsetin, and galangin demonstrated simi-

lar potencies for the inhibition of the P-form PST while
myricetin, chrysin, apigenin, and kaempferol were 3–10
times less potent.

The ﬂavonoids daidzein, genistein, quercetin, (+)-cate-

chin, equol and ﬂavone are non-competitive inhibitors of
human platelet P-PST with low K

and K

values (

Ghazali

and Waring, 1999

). The non-competitive nature of these

ﬂavonoid inhibitors of P-form PST is consistent with the
observation that they are poor substrates for this enzyme
(

Table 3

) (

Ghazali and Waring, 1999

). Consistent with pre-

vious ﬁndings, Harris et al. reported that quercetin was the
most potent inhibitor (IC

of 60 nM) (

Genistein and daidzein inhibited SULT1A1 with IC

val-

ues of 500 and 600 nM, respectively (

Table 3
K

, V

max

and K

values for inhibition of phenol sulfotransferase (PST) by

ﬂavonoids

Flavonoid

, lM

max

, lM

nmol/min/mg protein

Quercetin

0.051 ± 0.004

0.051 ± 0.007

0.10 ± 0.03

Genistein

0.047 ± 0.002

0.040 ± 0.001

0.21 ± 0.03

Daidzein

0.050 ± 0.007

0.049 ± 0.004

0.34 ± 0.08

Equol

0.058 ± 0.008

0.030 ± 0.007

0.49 ± 0.11

(+)-Catechin

0.053 ± 0.001

0.058 ± 0.009

0.76 ± 0.03

Flavone

0.061 ± 0.004

0.031 ± 0.006

0.94 ± 0.17

Control P-form-PST activity (without inhibitor) was 0.86 ± 0.09 nmol/
min/mg protein which also had a K

value of 0.050 ± 0.007 lM and V

max

of 0.080 ± 0.03 nmol/min/mg protein. Data are expressed as means ± SD.
Reproduced from (

Ghazali and Waring, 1999

), with permission from

Elsevier Science Inc.

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210

203

The most potent inhibitor of SULT1A3 in human platelets
was baicalein (IC

of 500 nM) (

Quercetin inhibits the sulfation of resveratrol in human

liver and duodenum in vitro (

De Santi et al., 2000

). Resve-

ratrol is a polyphenolic compound present in wine at con-
centration between 1 and 10 lM (

Soleas et al., 1997

). It has

beneﬁcial eﬀects against cancer and protective eﬀects on the
cardiovascular system. The IC

for quercetin was 12 pM

(liver) and 15 pM (duodenum) for the inhibition of resvera-
trol sulfation (

De Santi et al., 2000

). The potent inhibition

of resveratrol sulfation by quercetin suggests that com-
pounds present in the diet may inhibit the sulphation of
resveratrol, thus improving its bioavailability.

Marchetti et al. (2001)

also showed that quercetin is a

potent inhibitor of human liver and duodenum sulfotrans-
ferase in vitro. Drugs are often administered orally, and the
intestine and liver are therefore important sites of drug
metabolism. In this study, cytosolic fractions of human
liver samples and biopsies of the duodenum were used.
The IC

values of quercetin for the sulfation of four clin-

ically used drugs (dopamine, (

)-salbutamol, minoxidil

and paracetamol) are greater in duodenum than those in
liver. Such a diﬀerence may reﬂect the diﬀerent composi-
tion of sulfotransferase forms in the liver and duodenum,
since SULT1A1 is preferentially expressed in the liver
and SULT1A3 in the intestine (

Marchetti et al., 2001

In the case of estrogen sulfotransferase (SULT1E1),

equol was the most potent inhibitor. Equol has hydroxyl
groups that can potentially superimpose with the 3-hydro-
xyl group of 17b-estradiol (E

) (

). The

-hydroxyl group appears to play an important role in

the inhibitory eﬀect because formononetin, which lacks
this group, was the least potent of all the compounds
tested (

). Quercetin competitively inhi-

bits the sulfation of E

in normal human mammary epi-

thelial cells (

Otake et al., 2000

). Another study reported

that quercetin, genistein and daidzein inhibit the sulfation
of 17b-estradiol by zebraﬁsh SULT (

Ohkimoto et al.,

). Kinetic analyses showed that the mechanism of

action by these ﬂavonoids was competitive inhibition
(

Ohkimoto et al., 2004

The eﬀect of the ﬂavonoids may be mediated not just by

their inhibition of the bioactivating activity of the SULTs
on carcinogens, but by the eﬀects of the sulfoconjugates
of the ﬂavonoids produced by the activity of these enzymes
(

Pai et al., 2001

). It has been shown that daidzein sulfocon-

jugates are potent inhibitors of sterol sulfatase and
decrease the production of the biologically active estro-
genic steroids (shown to stimulate many breast tumors)
in mammary tissue from their inactive sulfoconjugates
(

Wong and Keung, 1997

4. Diﬃculties in the prediction of in vivo metabolic
eﬀects in humans

It is clear that eﬀects of ﬂavonoids in vivo may not

always be predicted on the basis of in vitro results alone.

For example, in the study of

Ueng et al. (1999)

, naringenin

but not naringin is an inhibitor of benzo(a)pyrene hydrox-
ylase (AHH) activity in vitro, whereas naringin inhibited
AHH activity and the expression of CYP1A2 in vivo.
Moreover, ﬂavonoids generally have low oral bioavailabil-
ity and can be degraded by gut bacteria. Therefore, concen-
trations in vivo may not reﬂect the concentrations tested
under in vitro conditions (

Kuhnau, 1976

). Isoﬂavone gly-

cosides are not absorbed intact across the enterocyte of
healthy adults, and their bioavailability requires initial
hydrolysis of the sugar moiety by intestinal b-glucosidases
for uptake to the peripheral circulation (

Setchell et al.,

DiCenzo et al., 2003; Gurley et al., 2004; Moore

). Apigenin, naringenin, and chrysin, which are strong

inhibitors of aromatase in vitro, did not inhibit androstene-
dione-induced uterine growth, indicating a lack of aroma-
tase-inhibiting eﬀect in vivo (

Saarinen et al., 2001

), a

diﬀerence that may be due to their relatively poor absorp-
tion and/or bioavailability. Tangeretin did not alter the
CYP3A4 activity in human volunteers, although it is a
potent stimulator of CYP3A4 activity in human liver
microsomes and microsomes containing cDNA-expressed
CYP3A4 (

Backman et al., 2000

). Quercetin has been shown

to inhibit the catalytic activity of P-PST using cell-free
enzyme preparations in vitro with an IC

value as low as

0.1 lM. In the intact human hepatoma cell line HepG2,
the potency of quercetin as an inhibitor of P-PST decreased
about 25-fold, yielding an IC

value of 2.5 lM (

Eaton

et al., 1996

). This diﬀerence is possibly due to the high

serum protein binding and poor plasma membrane perme-
ability of quercetin, and its metabolism to inactive metab-
olites. Silymarin also has inhibitory eﬀect on CYP3A4
in vitro, but not in vivo. This lack of in vitro–in vivo cor-
relation may be due to poor bioavailability, large inter-
individual variations in silibinin (also known as silybin or
silybinin) absorption, lower CYP binding aﬃnities of silib-
inin conjugates, product variability in silymarin content, or
poor dissolution characteristics of milk thistle dosage
forms (

et al., 2000

Flavonoid metabolites may have a higher or lower bio-

logical activity than the parent drug, and result in a change
of the overall cancer protective response.

Nielsen et al.

(2000)

has identiﬁed and quantiﬁed 10 diﬀerent metabo-

lites, each bearing an intact ﬂavan skeleton, in rat urine
and feces, after repeated administration of tangeretin
(abundant in citrus peal), as primarily demethylation
and/or hydroxylation products. The diﬀerences in metabo-
lite formation and disposition in vitro and in vivo may
account for some of the diﬀerences observed in in vivo
and in vitro studies.

Variable dietary exposure to a range of ﬂavonoid com-

pounds may contribute to some of the interindividual var-
iation in the pharmacokinetics and pharmacological
responses observed for drugs such as phenacetin, caﬀeine,
and theophylline, which are substrates for CYP1A2 (

Ren-

dic and Di Carlo, 1997

) as well as that observed for drugs

that are substrates for other P450 isozymes (

Zhai et al.,

204

Y.J. Moon et al. / Toxicology in Vitro 20 (2006) 187–210