PHYTOTHERAPY RESEARCH

Phytother. Res. (2012)

Published online in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/ptr.4724

Evaluation of Fooddrug Interaction of Guava

Leaf Tea
Kimiyuki Kaneko,* Katsuya Suzuki, Emi Iwadate-Iwata, Ikuo Kato, Kazumi Uchida
and Masaharu Onoue
Safety Research Department, Yakult Central Institute for Microbiological Research

Guava leaf tea (GLT) contains guava leaf polyphenol (Gvpp), which regulates the absorption of dietary carbohydrate from the intestines. Borderline diabetics, who are at high risk of development of diabetes, take GLT
to suppress a rapid increase of blood sugar level after meals. However, patients with diabetes in whom diabetic
drugs or warfarin as a blood thinner are prescribed also take GLT with the expectation of glycemic control.
Therefore, we studied whether GLT had potential for inhibition or induction of cytochrome P450 (CYP) and
an inuence on the action of warfarin. Extract of guava leaf (GvEx) consists of carbohydrate and polyphenols,
which are Gvpp, quercetin, and ellagic acid. These polyphenols, but not GvEx, showed a certain level of inhibition of human-cDNA-expressed CYPs. In a comparison of GLT and grapefruit juice, GLT showed weaker
inhibition of CYP activities and of midazolam 1-hydroxylation than grapefruit juice. Furthermore, neither liver
weight nor CYP3A expression in the liver was changed in rats that received GvEx for 90 days compared with the control group. When rats were concomitantly treated with GLT and warfarin, the prolongation of clotting time of blood
by warfarin was not inuenced. These data suggest that GLT is unlikely to interact with drugs. Copyright 2012
John Wiley & Sons, Ltd.
Keywords: fooddrug interaction; guava leaf polyphenol; guava leaf tea; P450; warfarin.

INTRODUCTION
As a traditional folk remedy, decoction of leaves of a
subtropical plant, Psidiumguajava, has been used as a
cure for diabetes and diarrhea in Okinawa (Japan) and
in Taiwan. Guava leaf tea (GLT) is claimed to have
benecial health effects through regulation of absorption of dietary sugar in borderline diabetics.
Thus, GLT is a functional food and a new type of
beverage that is commercially available in Japan. Guava
leaf polyphenol (Gvpp) contained in guava leaf consists
of a tannin polymer with a molecular weight of about
500030 000. Because dietary polyphenols bind to
digestive enzymes and regulate their activities in the
gastrointestinal tract, some polyphenols are known to
contribute to health through poor absorption of dietary
sugar or lipid (Mai et al., 2007; Lo Piparo et al., 2008;
McDougall et al., 2008). As well as these polyphenols,
Gvpp can inhibit saccharolytic enzymes in the intestine
(Deguchi et al., 1998). In a clinical study, GLT was
effective for regulation of blood glucose level after
meals, when borderline diabetics consumed 200 mL
GLT containing 70 mg Gvpp (Deguchi et al., 2000).
Some polyphenols are reported to have potent inhibitory activity against CYP3A4 and CYP2C9 (Kimura
et al., 2010). Flavonoids are one type of polyphenol
and have been shown to inhibit human P450 enzymes
including CYP3A4 in in vitro and in vivo studies (Piver
et al., 2003; Gaudineau et al., 2004; Lin et al., 2005).
Quercetin, one of the avonoids, is a potent inhibitor
* Correspondence to: Kimiyuki Kaneko, 1796 Yaho, Kunitachi, Tokyo
186-8650, Japan.
E-mail: kimiyuki-kaneko@yakult.co.jp

Copyright 2012 John Wiley & Sons, Ltd.

of CYP2C8 (Walsky et al., 2005; Kajosaari et al., 2005)

and has been usually used as a typical inhibitor of
CYP2C8 in many in vitro studies. Meanwhile, quercetin
is reported to induce CYP3A4 mRNA expression in
primary cultures of human hepatocytes (Raucy, 2003).
In addition, some avonoids of Ginkgo biloba are known
to be strong inducers of CYP3A (Deng et al., 2008).
Among CYPs, CYP3A4 has been recognized as a key
enzyme in fooddrug interactions because it is responsible
for the metabolism of more than 50% of clinical pharmaceuticals. Many patients with diabetes experience complications. They are required to take various medicines
and need long-term treatment. There is a tendency for
patients with diabetes-related disorders to receive bloodthinning medication such as warfarin (WF). In addition,
many patients with diabetes tend to take a functional food
in the interest of health promotion even if the medicine is
prescribed by a doctor. Therefore, some medical experts
and patients with diabetes are concerned about the interactions between such medicines and functional foods.
To dispel their apprehensions, the present study
examined whether GLT would inhibit CYPs, induce
CYPs, and inuence the efcacy of WF. To assess the
inhibitory potency by GLT, we performed an in vitro
assay using human-cDNA-expressed CYPs. Especially
pertaining to the inhibition of CYP3A4, we examined
the inhibitory activity on midazolam 1-hydroxylation
using human liver microsomes. We examined the induction potency of GLT by western blot analysis for
CYP3A using liver specimens from rats fed a hot water
extract of guava leaves (GvEx) for 90 days. Furthermore, to clarify the interaction between GLT and WF,
we evaluated the inuence of GLT given together with
WF on the activated partial thromboplastin time
(APTT) and thrombo test (TBT).
Received 09 August 2011
Accepted 13 April 2012

K. KANEKO ET AL.

MATERIALS AND METHODS

Test samples
GvEx was extracted using hot water from the leaf of
Psidiumguajava, with a yield of 14%. In brief, 100 g of
dried leaves was added to 2 L of distilled water (DW),
and the mixture was decocted at 80  C for 30 min. After
ltering through four layers of gauze to remove leaves,
the extracted solution was freeze dried. GvEx contained
about 40% carbohydrate (analyzed in triuoroaceticacid-treated sample by gas chromatography) and
33%46% Gvpp (analyzed by HPLC). In addition,
Gvpp contained 10% quercetin (analyzed hydrolysate
by HPLC) and 10% ellagic acid (analyzed dissolved
material with dimethylacetamide by HPLC). Gvpp was
puried from GvEx using a Bond Elute C18 reversephase solid-phase extraction column (2 g, Varian Sample
Preparation Products, CA, USA). The extract was freeze
dried after dialysis, and Gvpp was obtained with a yield
of 9.6%. The purity of Gvpp was 98% analyzed by
HPLC. GLT was obtained as a commercial product BansoreichaW from Yakult Honsha Co., Ltd. (Tokyo, Japan).
GLT mainly consisted of GvEx and contained Gvpp at a
concentration of 0.35 mg/mL. Grapefruit juice Sunkist
GRAPEFRUIT 100% was also obtained commercially
from Morinaga Milk Industry Co., Ltd. (Tokyo, Japan).

CHEMICALS
Microsomes from baculovirus insect cells expressing
recombinant human P450 isoforms (cDNA-expressed
CYPs), each containing only one P450 enzyme, P450
reductase, cytochrome b5, and pooled human liver
microsomes, were purchased from BD Gentest
(Woburn, MA, USA). The synthetic substrates
(AMMC: 3-[2-(N,N-diethyl-N-methylammonium) ethyl]7-methoxy-4-methylcoumarin; BFC: 7-benzyloxy-triuoromethylcoumarin;
CEC:
3-cyano-7-ethoxycoumarin;
DBF: dibenzyluorecein; EFC: 7-ethoxy-4-triuoromethylcoumarin); MFC: 7-methoxy-4-triuoromethylcoumarin) and the uorescent metabolites (AHMC:
3-[2-(N,N-diethyl-N-methylamino)ethyl]-7-hydroxy-4-methyl
coumarin; CHC: 3-cyano-7-hydroxycoumarin; HFC:
7-hydroxy-4-triuoromethylcoumarin; and uorescein)
were also purchased from BD Gentest (Woburn,
MA, USA). Glucose 6-phosphate, glucose 6-phoshate
dehydrogenase, b-NADP, vitamin K1, and WF were purchased from Sigma-Aldrich Corporation (St. Louis, MO,
USA). Midazolam, phenacetin, quercetin, ellagic acid,
calcium phosphate monobasic (food grade), IGEPAL
CA-630, phenylmethylsulfonyl uoride (PMSF), and
aprotinin were purchased from Wako Pure Chemical
Industries, Ltd. (Osaka, Japan). A metabolite of
midazolam, 1-hydroxymidazolam, was purchased from
Toronto Research Chemicals Inc. (North York, Canada).
Anti-rat CYP3A1/3A2 antibody and alkaline-phosphataselinked anti-rabbit IgG were purchased from Daiichi Pure
Chemicals Co., Ltd. (Tokyo, Japan) and Invitrogen
(Carlsbad, CA, USA), respectively. A protein assay
kit and a 5-bromo-4-chloro-3-indoly phosphate / nitro
blue tetrazolium (BCIP/NBT) detection kit were purchased from Pierce Chemical Company (Rockford, IL,
Copyright 2012 John Wiley & Sons, Ltd.

USA) and Nacalai Tesque, Inc. (Kyoto, Japan), respectively. Diagnostic reagents for APTT and TBT were purchased from Sysmex Corporation (Kobe, Japan).
All chemicals and solvents were of the highest grade
available commercially.

ANIMALS
Male SpragueDawley (Crl:CD(SD)) rats were purchased from Charles River Japan, Inc. (Kanagawa,
Japan). All animals were housed in normal cages and
allowed free access to water and feed (Funabashi Farm,
Co., Ltd., Chiba, Japan). All experimental protocols
were approved by the Institutional Animal Care and
Use Committee of Yakult Central Institute.
P450 inhibition studies
Inhibition assay of cytochrome P450 using synthetic
substrates.
Enzyme activity was measured using black Falcon
96-well microplates. The experimental conditions for
each CYP isoform are summarized in Table 1. The test
samples, which were GvEx and its components, were
adjusted by eight serial threefold dilutions from the initial concentration (GvEx: 1 mg/mL, Gvpp: 0.46 mg/mL,
quercetin: 0.04 mg/mL, and ellagic acid: 0.04 mg/mL,
based on calculation from their composition ratio in
GvEx). Inhibitors (positive control for CYP inhibition)
and serial dilutions of test samples were dispensed at
4 mL per each well, and 96 mL of cofactor solution was
added. In the case of GLT and grapefruit juice, 50 mL
of test samples (eight serial threefold dilutions) was
added to 100 mL of cofactor solution. Instead of the test
sample, 4 or 50 mL DW was added as the control. The
nal concentrations of cofactors in the reaction mixture
were 1.3 mM NADP+, 3.3 mM MgCl2, 3.3 mM glucose-6phosphate, and 0.4 units/mL glucose-6-phosphate
dehydrogenase in potassium phosphate buffer (KPO4;
pH 7.4). After 10 min of preincubation at 37  C, the
reaction was initiated by the addition of 100 mL prewarmed enzyme/substrate mixture. After the incubation
times indicated in Table 1, the reaction was terminated
by addition of 75 mL of stop solution (0.1 M Tris base
in acetonitrile or 2 N NaOH for CYP2C8). Fluorescence
was determined using a microplate reader (FLUOstar
model, BMG Lab Technologies GmbH, Offenburg,
Germany) with the excitation and emission wavelengths
shown in Table 1. The concentrations of test samples
required for 50% inhibition of catalytic activities (IC50)
were calculated using probit transformation and
regression analysis. IC50 values are shown as the mean
of triplicate assays. When inhibition of more than 50%
was not observed in the tested range of a sample, the
test sample was judged not inhibited.
Inhibition assay of CYP3A indicated midazolam
1-hydroxylation
In the assay, the nal concentration of GLT or grapefruit
juice ranged from 0.002% to 5%. The reaction mixture consisted of 100 mM sodium potassium phosphate
Phytother. Res. (2012)

EVALUATION OF FOODDRUG INTERACTION OF GUAVA LEAF TEA

Table 1. Experimental conditions of inhibition assay for each P450 enzyme using synthetic substrates
Concentration
Cytochrome P450
enzymes
CYP3A4
CYP2D6
CYP2C8
CYP2C9
CYP2C19
CYP2B6
CYP1A2

P450 content (nM)

KPO4 (mM)

10
15
40
10
4
10
5

200
100
50
25
100
100
100

Substrate (mM)
BFC
AMMC
DBF
MFC
CEC
EFC
CEC

50.0
1.5
1.0
75.0
25.0
2.5
5.0

Incubation
time (min)

Metabolite
standards

Excitation/Emission
wavelength (nm)

30
30
40
45
30
30
15

HFC
AHMC
Fluorescein
HFC
CHC
HFC
CHC

409/520
390/460
485/544
409/520
409/460
409/520
409/460

BFC: 7-benzyloxy-trifluoromethyl coumarin; AMMC: 3-[2-(N,N-diethyl-N-methylammonium)ethyl]-7-methoxy-4-methylcoumarin; DBF:

dibenzylfluorecein; MFC: 7-methoxy-4-trifluoromethyl coumarin; CEC: 3-cyano-7-ethoxycoumarin; EFC: 7-ethoxy-4-trifluoromethylcoumarin; HFC: 7-hydroxy-4-trifluoromethycoumarin; AHMC: 3-[2-(N,N-diethyl-N-methylamino)ethyl]-7-hydroxy-4-methylcoumarin; CHC:
3-cyano-7-hydroxycoumarin.

buffer (pH 7.4), 50 mM EDTA disodium salt, cofactor

solution (0.5 mM NADP+, 5 mM MgCl2, 5 mM glucose
6-phosphate, and 1 unit/mL glucose 6-phosphate
dehydrogenase), and pooled human liver microsomes
(0.2 mg protein/mL) in a total volume of 0.5 mL.
Because the Km value of midazolam was in the range
of 2.55.57 mM (Ghosal et al., 1996; Oda et al., 1999),
10 mM was set as the concentration of midazolam. After
10 min of preincubation at 37  C, reactions were
initiated by the addition of midazolam followed by incubation at 37  C for 15 min, and 5 mL of ethyl acetate was
added to stop the reaction. Phenacetin (100 mL; 5 mM)
was added as an internal standard, and 120 mg of ammonium sulfate was added to remove protein. After mixing
for 10 min, the sample was centrifuged at 3000 rpm for
10 min. The organic layer was transferred to another
tube, and the solvent was evaporated. The residue was
dissolved in 75 mM sodium acetate (pH 6.5) and acetonitrile (3:2, v/v) (50 mL). Analysis of the 1-hydroxymidazolam metabolite was performed by HPLC using a
computerized HPLC system (Waters Alliance 2695,
Waters Corporation, Milford, MA, USA) equipped with
a Cadenza CD-18 column (4.6  250 mm, 3 mm, Imtakt,
Kyoto, Japan). The mobile phase consisted of 75 mM of
sodium acetate (pH 6.5) for solvent A and acetonitrile
for solvent B. The metabolite was separated using a
linear gradient of 50% to 70% solvent A (0 to 25 min)
at a ow rate of 0.7 mL/min. Quantication of the
metabolite was performed by comparing the HPLC
peak area monitored at 240 nm to that of the internal
standard. Using the means of triplicate determinations,
IC50 values were calculated using probit transformation
and regression analysis.

quantitative analysis using western blot assay, the liver

(right lobe) was cut into small pieces. Liver homogenates were prepared in homogenizing buffer (50 mM
HEPES-KOH (pH 7.5), 150 mM NaCl, 1% IGEPAL
CA-630 as nonionic surfactant, 1 mM PMSF, and
1 mg/mL aprotinin). After centrifugation of the liver
homogenate at 600 g for 10 min at 4  C, Laemmli buffer
(62.5 mM TrisHCl (pH 6.8), 2% sodium dodecyl
sulfate (SDS), 5% mercaptoethanol, and 10% glycerol)
at an equal volume to that of the excluded supernatant
liquid was added to the precipitate. The mixture was
denatured for 2 min at 100  C, and the protein content
was determined using a Pierce BCA Protein Assay
Kit. The samples (50 mg protein) were separated by
10% SDSpolyacrylamide gel electrophoresis, and the
resulting band was transferred to a polyvinylidene
diuoride membrane. The membrane was blocked with
5% nonfat dried milk in TBS-T buffer (25 mM TrisHCl
(pH 7.5), 150 mM NaCl, and 0.1% Tween 20) for 1 h at
room temperature (RT) and exposed to anti-rat CYP
3A1/3A2 antibody diluted with 0.5% nonfat dried milk
in TBS-T buffer for 1 h at RT. Then, the membrane
was incubated with alkaline-phosphatase-linked antirabbit IgG for 1 h at RT. After washing three times with
TBS-T, antigenantibody complexes were visualized by
development using a BCIP/NBT detection kit. After
capturing the band of CYP3A1/3A2 on a Fujilm
LAS-3000 CCD camera, the optical densities of bands
were quantitated using MultiGauge software (Fujilm
Corporation, Tokyo, Japan), and the relative density
was compared with the control band as 100%. For
histopathological examination, the liver was xed in
10% neutral buffered formalin, embedded in parafn,
sectioned, stained with hematoxylin and eosin (HE),
and examined microscopically.

Evaluation of enzyme induction in liver by GvEx

feeding in rats
Evaluation of interaction between WF and GLT in rats
The possibility of induction of drug metabolic enzymes
was assessed by western blotting and histological tissue
examination in the liver from male SD rats orally administered GvEx (2000 mg/kg/day) or DW for 90 days.
After an overnight fast following the nal administration, rats were exsanguinated under anesthesia. Half of
the liver of rats in each group was devoted to quantitative analysis and the other half to histopathological examination after organ weight measurement. In
Copyright 2012 John Wiley & Sons, Ltd.

An in vivo study was performed to determine whether

the ingestion of GLT inuenced the anticoagulative
effects of WF in blood. GLT (20 mL/kg) or DW was
administered orally to male SD rats (6 weeks old), and
30 min later, WF (0.1 mg/5 mL/kg) was administered
orally. These treatments were repeated for seven
consecutive days. It is well known that excessive vitamin
K negates the anticoagulative effect of WF (Booth et al.,
Phytother. Res. (2012)

K. KANEKO ET AL.

1997; Lubetsky et al., 1999). A vitamin-K-rich diet was

used as a suppressive control for the effect of WF on
blood coagulation. The vitamin-K-rich diet was composed of F-2 diet (Funabashi Farms, Co., Ltd., Chiba,
Japan), calcium phosphate monobasic, and vitamin K1
at a ratio of 100:0.24:0.125. Each group consisted of six
rats. The day after the nal administration, blood was
collected from the abdominal aorta of each rat under
anesthesia by injection of sodium pentobarbital. For
the measurement of APTT and TBT, the blood sample
was treated with 3.2% potassium citrate as an anticoagulant. After platelet poor plasma (PPP) separation,
APTT or TBT reagent was added to recalcied PPP or
PPP, respectively. The time required for clot formation
was measured using a Coagrex 800 (Sysmex Corporation, Kobe, Japan). For the platelet count, blood was
collected in a similar way using EDTA dipotassium as
an anticoagulant. Platelet count was determined on an
XT 2000iv (Sysmex Corporation, Kobe, Japan).

RESULTS
In vitro study of P450 inhibition
We tested the inhibitory effects of GvEx and its components using human-cDNA-expressed CYPs. Moreover,
we also tested a commercial product (GLT) using the
same method. IC50 values of GvEx, its components,
and GLT are shown in Table 2. The inhibitory effects
of GvEx were found to be weaker than those of each
control inhibitor by a factor 1/22 (CYP2B6)1/36 500
(CYP2D6). The components of GvEx demonstrated

stronger inhibition than did GvEx on each CYP. As

shown by Zou et al. (2002), we showed that the activity
of quercetin inhibited not only CYP2C8 but also
CYP2C19, CYP3A4, and CYP2C9. These inhibitory
effects of quercetin were stronger than those of Gvpp,
whereas Gvpp inhibited CYP1A2, CYP2B6, and
CYP2D6 more strongly than did quercetin. In the
inhibition assay using commercial products, GLT
showed low activity as follows: The IC50 values of
grapefruit juice were from 1/10 (CYP3A4) to 1/2
(CYP1A2) compared with those of GLT. To get further
conrmation that the ability of GLT to inhibit CYP3A4
is lower than that of grapefruit juice, the production of
1-hydroxymidazolam, which is a metabolite of
midazolam, was analyzed by HPLC after incubation
with pooled human liver microsomes. Although both
GLT and grapefruit juice inhibited the metabolite formation dose-dependently, it was clear that the inhibitory
activity of GLT was lower than that of grapefruit juice.
The IC50 value of GLT was sixfold higher than that of
grapefruit juice inasmuch as the values of GLT and
grapefruit juice were 4.41% and 0.69%, respectively
(Table 3). According to these results, it was considered
that the possibility of drug interaction with GLT through
CYP inhibition is lower than that for grapefruit juice.

In vivo study of P450 induction

To clarify whether GLT has potential ability to induce
P450 enzymes in the liver, rats received 2000 mg/kg/day
GvEx for 90 days, and western blot assay for CYP3A1/
3A2 expression was performed using their liver tissues.
Table 4 shows no change in liver weight after feeding

Table 2. IC50 values for inhibition by GvEx, its ingredients, and product on human-cDNA-expressed P450 enzymes
IC50 (mg/mL)

IC50 (%)
Ingredient

Isozyme
CYP3A4
CYP2D6
CYP2C8
CYP2C9
CYP2C19
CYP2B6
CYP1A2

Control inhibitor
Ketoconazole
Quinidine
Quercetin
Sulfaphenazole
Tranylcypromine
Tranylcypromine
Furafylline

0.02
0.01
0.59
0.10
0.35
1.44
0.52

Product

GvEx

Gvpp

Quercetin

Ellagic acid

Grapefruit juice

GLT

19.92
182.28
18.16
28.00
19.09
32.04
26.43

8.78
6.43
1.45
2.53
2.09
3.94
1.56

1.50
10.58
0.59
1.62
0.98
N.L.
2.24

4.80
N.I.
13.99
7.14
1.95
N.I.
5.14

0.21
0.58
2.61
0.18
0.15
0.58
0.50

1.98
2.60
6.67
1.43
1.21
1.78
1.04

Data are means of triplicate determinations.

GvEx: extract of guava leaves; Gvpp: guava leaf polyphenol; GLT: Bansoreicha; N.I.: not inhibited.

Table 3. Percentage of 1-hydroxymidazolam at each additive density and IC50 values of CYP-dependent 1-hydroxylase activity
Relative density of metabolite (1-hydroxymidazolam)
0

0.002

Samples
GLT
Grapefruit juice

0.007

0.02

0.06

0.19

0.56

1.67

5.0
IC50
(%)

(Final concentration of sample in reaction mixture (%))

100
100

98.1
111.8

102.1
102.3

105.8
97.9

105.7
99.3

104.5
87.6

91.1
65.2

71.4
25.4

14.0
4.2

4.41
0.69

Data are means of triplicate determinations.

GLT: guava leaf tea.
Copyright 2012 John Wiley & Sons, Ltd.

Phytother. Res. (2012)

EVALUATION OF FOODDRUG INTERACTION OF GUAVA LEAF TEA

Table 4. Liver weight of rats fed GvEx for 90 days

Groups

Final body weight (g)

Absolute weight (g)

Relative weight (g/100 g body weight)

563.1  59.6
535.4  36.1

13.84  2.53
13.03  0.96

2.45  0.23
2.44  0.14

DW
GvEx

Data are mean  SD (n = 10).

DW: distilled water; GvEx: extract of guava leaves.

GvEx for 90 days. Quantitative analysis of the liver by

western blotting showed no difference in liver CYP3A1/
3A2 expression between GvEx-treated rats and control rats (Fig. 1). From these results, no predictive
evidence suggesting induction of drug-metabolizing
enzyme by feeding GvEx was observed in liver weight
and CYP expression.

PLT
VK-WF
GLT-WF
GLT-DW
DW-WF
DW-DW

Interaction between WF and GLT

We tested the antagonistic activity of GLT against WF by

the change of blood coagulation time. Treatment of
rats with WF (0.1 mg/kg, 7 days) and/or GLT (20 mg/kg,
7 days) did not cause a signicant change in body weight
or food intake (data not shown). Moreover, the platelet
count in each group was not changed compared with the
control (DW-DW) group (Fig. 2). As shown in Fig. 3,
treatment with WF alone in the DW-WF group caused
prolongation of coagulation time of PPP, and antagonistic
activity of vitamin K to WF was observed in rats fed the
vitamin K1 diet (VK-WF group). There was no signicant difference in coagulation times (APTT and TBT)
between the DW-DW group and the GLT-DW group.
This means that consumption of GLT had no effect
on the coagulation of PPP. In the comparison of the
DW-WF group and the GLT-WF group, prolongation of
APTT and TBT by WF was equally likely to be observed.
This observation showed that consumption of GLT had
no effect on the anticoagulative effects of WF.
A

DW

GvEx

Relative density, %

120
100
80
60
40
20
0

DW

GvEx

Figure 1. No effect of GvEx on liver CYP3A1/3A2 expression. (A)

Western blot analysis: protein samples from GvEx or DW treatment
for 90 days. (B) Relative hepatic CYP3A1/3A2 expression of rats in
DW group (control) and GvEx group. CYP3A1/3A2 levels are
shown relative to mean control expression as 100%. Values represent mean  SD (n = 5).
Copyright 2012 John Wiley & Sons, Ltd.

200

400

600

800

1000 1200

Figure 2. No effect on platelet count in rats treated with GLT and/

or warfarin. Values are mean  SD (n = 6). PLT, platelet count; VK,
vitamin K1; WF, warfarin; GLT, guava leaf tea; DW, distilled water.

DISCUSSION
The present study demonstrated that the consumption
of GLT had little or no possibility of inhibition or
induction of drug-metabolizing enzymes (P450). As
shown in Table 2, GvEx inhibited various CYPs in the
order CYP2C8 > CYP2C19 > CYP3A4 > CYP1A2 >
CYP2C9 > CYP2B6 > CYP2D6. This order was similar
to that of quercetin, and the inhibitory activities of
GvEx on CYP2C8 and CYP2C19 were about 1/30 and
1/20 of those of quercetin, respectively. In view of the
composition ratio of GvEx, it was thought that the
inhibitory activities by GvEx might be mainly derived
from quercetin. On the other hand, IC50 values of Gvpp
in CYP2D6, CYP1A2, and CYP2B6 were lower than
those of quercetin or ellagic acid. Considering that
GvEx contained 33%46% Gvpp, such inhibitory activities of GvEx on these CYPs were weaker than the
expected IC50 values. Therefore, it was suggested that
there might be some substances that obstruct the afnity
between P450 and these polyphenols in the plant components of GvEx, in addition to Gvpp, quercetin, and
ellagic acid. We think that grapefruit juice could be an
indicator of CYP inhibition because furanocoumarins
in grapefruit can inhibit not only CYP3A4 but also other
CYPs (Tassaneeyakul et al., 2000; Wen et al., 2002). In
addition, it has become public knowledge that drinking
a glass of grapefruit juice augments the bioavailability
of drugs metabolized by CYP3A4 in humans (Dahan
and Altman, 2004; Fukatsu et al., 2006). We therefore
compared the inhibitory activity of GLT with that of
grapefruit juice. In our study, the IC50 value of GLT
was 9.4 times higher than that of grapefruit juice using
human-cDNA-expressed CYPs and 6.4 times higher
for 1-hydroxylation of midazolam. These combined
ndings suggest that there is no possibility that GLT
could increase a pharmacological effect or side effects
through inhibition of CYP3A4.
It is also known that fooddrug interactions occur
as a result of induction of hepatic-drug-metabolizing
Phytother. Res. (2012)

K. KANEKO ET AL.

APTT
VK-WF
GLT-WF
GLT-DW
DW-WF
DW-DW
0

10

20

30

(sec)

10

20

30

(sec)

TBT
VK-WF
GLT-WF
GLT-DW
DW-WF
DW-DW

Figure 3. Interaction of GLT with warfarin-induced anticoagulation effect. Values are mean  SD (n = 6).*p < 0.005, **p < 0.001: significant differences between two groups are indicated on the right. APTT, activated partial thromboplastin time; TBT; thrombo test; VK, vitamin
K1; WF, warfarin; GLT, guava leafs tea; DW, distilled water.

enzymes. For example, St. Johns wort is the most wellknown herb that acts as an inducer of CYP3A. Ginkgo
biloba, which has been used to alleviate symptoms associated with numerous cognitive disorders, is also known
to induce CYP2B and CYP3A (Tada et al., 2008).
Induction of CYP2D in the rat liver by St. Johns wort
(Dostalek et al., 2005) and in human hepatocytes by
Valerian or Ginkgo biloba (Hellum et al., 2007) has been
reported. We therefore administered GvEx to rats at a
dose of 2000 mg/kg for 90 days and examined whether
drug-metabolizing enzymes were induced. As a result, no
inuence of GvEx feeding on liver weight and induction
of hepatic-drug-metabolizing enzymes and CYP3A1/3A2
was observed. As for the other CYPs proliferation of
smooth endoplasmic reticulum, which has a characteristic
ground-glass appearance (Vzquez, 1990), was not
shown in histopathological observation of HE-stained
sections of the liver (data not shown). This in vivo
study suggested that the possibility of induction of drugmetabolizing enzymes by GLT ingestion is very small.
Concerns over the interaction between WF and
foods have been reported in case reports and reviews
(Holbrook et al., 2005; Nutescu et al., 2006; Paeng
et al., 2007; Leung et al., 2008; Hornsby et al., 2008). In
these reports, some symptoms induced by ineffectiveness of WF or unexpected prolongation of clotting time
including excessive bleeding have been encountered
mostly in patients who suffered venous thrombosis, cardiac infarction, or cerebral embolism. These ineffective
or excessive actions of WF may have critical inuences
on these patients. Therefore, the drug interaction
between WF and GLT was examined by determination
of the changes in coagulation time when rats were administered GLT and WF concomitantly. In the animal
Copyright 2012 John Wiley & Sons, Ltd.

experiment using WF, it was very important to determine the appropriate dosage of WF. Bleeding in the
body and a decrease in platelet count in blood are
unsuitable in a study to measure the blood coagulation
times. At the applied dosage of WF (0.1 mg/kg, 7 days)
in this study, the platelet count of rats in WF-treated
groups did not decrease (Fig. 3). Fig. 4 shows that treatment with WF signicantly prolonged the coagulation
time (APTT and TBT). APTT reects coagulation
factors starting from XII and followed by XI, IX, VIII,
X, V, II, and brinogen. TBT reects coagulation factors
II, VII, and X. Vitamin K plays an important role in the
blood-clotting reaction through activation of coagulation factors II, VII, IX, and X. These coagulation factors
are synthesized in the liver in the presence of vitamin K,
and WF inhibits the vitamin-K-metabolizing cycle. In
this in vivo study, it was demonstrated that GLT ingestion did not inuence the effect of WF, and the coagulation time with GLT ingestion was the same as that with
DW treatment. As a result, GLT does not seem to either
decrease or increase the anticoagulant activity of WF.
In conclusion, the present results showed no possibility
of interactions between GLT and medicines. Furthermore, a recent report that studied herbal tea also stated
that guava tea showed no interaction with medicine
(Matsuda et al., 2007). The safety of GLT in fooddrug
interactions may be supported by these ndings. Food
Drug interactions have been reported to occur in
various systems in the body. Foods contain complex
mixtures of plant components such as polyphenols
and avonoids. These phytochemicals have the potential
to affect the properties of medicines through dietary
macroconstituents (i.e., total protein, fat, and carbohydrate ratios and total energy intake) and by
Phytother. Res. (2012)

EVALUATION OF FOODDRUG INTERACTION OF GUAVA LEAF TEA

inhibiting or inducing the activity of drug-metabolizing

enzymes. Further studies needed required to elucidate
these points.

Conict of Interest
All authors are employees of Yakult Honsha Co., Ltd.

REFERENCES
Booth SL, Charnley JM, Sadowski JA, Saltzman E, Bovill EG,
Cushman M. 1997. Dietary vitamin K1 and stability of oral
anticoagulation: proposal of a diet with constant vitamin K1
content. Thromb Haemost 77: 504509.
Dahan A, Altman H. 2004. Fooddrug interaction: grapefruit juice
augments drug bioavailability-mechanism, extent and relevance. Eur J Clin Nutr 58: 19.
Deguchi Y, Osada K, Chonan O, et al. 2000. Effectiveness of consecutive ingestion and excess intake of Guava leaves tea in
human volunteers. J Jpn Counc Adv Food Ingred Res (in
Japanese) 3: 1928.
Deguchi Y, Osada K, Uchida K, et al. 1998. Effects of extract of
guava leaves on the development of diabetes in the db/db
mouse and on the postprandial blood glucose of human subjects. Nippon Nogeikagakukaishi 72: 923931.
Deng Y, Bi HC, Zhao LZ, et al. 2008. Induction of cytochrome
P450s by terpenetrilactones and flavonoids of the Ginkgo
biloba extract EGb 761 in rats. Xenobiotica 38: 465481.
Dostalek M, Pistovcakova J, Jurica J, et al. 2005. Effect of St
Johns wort (Hypericumperforatum) on cytochrome P-450
activity in perfused rat liver. Life Sci 5: 239244.
Fukatsu S, Fukudo M, Masuda S, et al. 2006. Delayed effect of
grapefruit juice on pharmacokinetics and pharmacodynamics
of tacrolimus in a living-donor liver transplant recipient. Drug
Metab Pharmacokinet 21: 122125.
Gaudineau C, Beckerman R, Welbourn S, Auclair K. 2004. Inhibition of human P450 enzymes by multiple constituents of the
Ginkgo biloba extract. Biochem Biophys Res Commun 318:
10721078.
Ghosal A, Satoh H, Thomas PE, Bush E, Moore D. 1996. Inhibition
and kinetics of cytochrome P4503A activity in microsomes
from rat, human, and cDNA-expressed human cytochrome
P450. Drug Metab Dispos 24: 940947.
Hellum BH, Hu Z, Nilsen OG. 2007. The induction of CYP1A2,
CYP2D6 and CYP3A4 by six trade herbal products in cultured
primary human hepatocytes. Basic Clin Pharmacol Toxicol
100: 2330.
Holbrook AM, Pereira JA, Labiris R, et al. 2005. Systematic overview of warfarin and its drug and food interactions. Arch
Intern Med 165: 10951106.
Hornsby LB, Hester EK, Donaldson AR. 2008. Potential interaction
between warfarin and high dietary protein intake. Pharmacotherapy 28: 536539.
Kajosaari LI, Laitila J, Neuvonen PJ, Backman JT. 2005. Metabolism
of repaglinide by CYP2C8 and CYP3A4 in vitro: effect of fibrates
and rifampicin. Basic Clin Pharmacol Toxicol 97: 249256.
Kimura Y, Ito H, Ohnishi R, Hatano T. 2010. Inhibitory effects of
polyphenols on human cytochrome P450 3A4 and 2C9 activity. Food Chem Toxicol 48: 429435.
Leung H, Hung A, Hui AC, Chan TY. 2008. Warfarin overdose
due to the possible effects of Lyciumbarbarum L. Food Chem
Toxicol 46: 18601862.
Lin HL, Kent UM, Hollenberg PF. 2005. The grapefruit juice effect
is not limited to cytochrome P450 (P450) 3A4: evidence for

Copyright 2012 John Wiley & Sons, Ltd.

bergamottin-dependent inactivation, heme destruction, and

covalent binding to protein in P450s 2B6 and 3A5. J Pharmacol Exp Ther 313: 154164.
Lo Piparo E, Scheib H, Frei N, Williamson G, Grigorov M, Chou CJ.
2008. Flavonoids for controlling starch digestion: structural
requirements for inhibiting human alpha-amylase. J Med Chem
51: 35553561.
Lubetsky A, Dekel-Stern E, Chetrit A, Lubin F, Halkin H. 1999.
Vitamin K intake and sensitivity to warfarin in patients consuming regular diets. Thromb Haemost 81: 396399.
Mai TT, Thu NN, Tien PG, Van Chuyen N. 2007. Alpha-glucosidase
inhibitory and antioxidant activities of Vietnamese edible
plants and their relationships with polyphenol contents. J Nutr
Sci Vitaminol 53: 267276.
Matsuda K, Nishmura Y, Kurata N, Iwase M, Yasuhara H. 2007.
Effects of continuous ingestion of herbal teas on intestinal
CYP3A in the rat. J Pharmacol Sci 103: 214221.
McDougall GJ, Kulkarni NN, Stewart D. 2008. Current developments on the inhibitory effects of berry polyphenols on digestive enzymes. Biofactors 34: 7380.
Nutescu EA, Shapiro NL, Ibrahim S, West P. 2006. Warfarin and its
interactions with foods, herbs and other dietary supplements.
Expert Opin Drug Saf 5: 433451.
Oda Y, Mizutani K, Hase I, Nakamoto T, Hamaoka N, Asada A.
1999. Fentanyl inhibits metabolism of midazolam: competitive
inhibition of CYP3A4 in vitro. Br J Anaesth 82: 900903.
Paeng CH, Sprague M, Jackevicius CA. 2007. Interaction between
warfarin and cranberry juice. Clin Ther 29: 17301735.
Piver B, Berthou F, Dreano Y, Lucas D. 2003. Differential inhibition
of human cytochrome P450 enzymes by epsilon-viniferin, the
dimer of resveratrol: comparison with resveratrol and polyphenols from alcoholized beverages. Life Sci 73: 11991213.
Raucy JL. 2003. Regulation of CYP3A4 expression in human hepatocytes by pharmaceuticals and natural products. Drug Metab
Dispos 31: 533539.
Tada Y, Kagota S, Kubota Y, et al. 2008. Long-term feeding of
Ginkgo biloba extract impairs peripheral circulation and
hepatic function in aged spontaneously hypertensive rats. Biol
Pharm Bull 31: 6872.
Tassaneeyakul W, Guo LQ, Fukuda K, Ohta T, Yamazoe Y. 2000.
Inhibition selectivity of grapefruit juice components on human
cytochromes P450. Arch Biochem Biophys 378: 356363.
Vzquez JJ. 1990. Ground-glass hepatocytes: light and electron
microscopy. Characterization of the different types. Histol
Histopathol 5: 379386.
Walsky RL, Gaman EA, Obach RS. 2005. Examination of 209
drugs for inhibition of cytochrome P450 2C8. J Clin Pharmacol
45: 6878.
Wen YH, Sahi J, Urda E, et al . 2002. Effects of bergamottin on
human and monkey drug-metabolizing enzymes in primary
cultured hepatocytes. Drug Metab Dispos 30: 977984.
Zou L, Harkey MR, Henderson GL. 2002. Effects of herbal components on cDNA-expressed cytochrome P450 enzyme catalytic
activity. Life Sci 71: 15791589.

Phytother. Res. (2012)