Metabolism of caffeic acid by isolated rat hepatocytes and subcellular fractions

Article abstract of DOI:10.1016/S0378-4274(02)00105-4

Caffeic acid (CA) is found in a wide variety of foods such as vegetables, fruits, tea, coffee, and wine. However, enzymes involved in its metabolism have not been identified. In the following, caffeic (CA), chlorogenic (CGA), and dihydrocaffeic (DHCA) acids were incubated with hepatocytes and shown to undergo metabolism by cytochrome P450, catechol-O-methyltransferase (COMT), and β-oxidation enzymes. Ferulic (FA) or dihydroferulic (DHFA) acids, formed as the result of CA- or DHCA-O-methylation by COMT, were also O-demethylated by CYP1A1/2 but not CYP2E1. DHCA or DHFA also underwent side chain dehydrogenation to form CA and FA, respectively, which was prevented by thioglycolic acid, an inhibitor of the β-oxidation enzyme acyl CoA dehydrogenase. The rates of glutathione conjugate formation catalyzed by NADPH/microsomes (CYP2E1) in decreasing order DHCA>CA>CGA trend which was in reverse order to the rates of their O-methylation by COMT. The CA- and DHCA-o-quinones formed by NADPH/P450 likely inhibited COMT but can readily form glutathione conjugates. CA, DHCA and DHFA were inter-metabolized to each other and to FA by isolated rat hepatocytes whereas FA was metabolized only to CA but not to DHCA or DHFA. CA, DHCA, FA, DHFA and CGA showed a dose-dependent hepatocyte toxicity and the LD₅₀ (2 h), determined were in decreasing order of effectiveness DHCA>CA>DHFA>CGA>FA. In summary, evidence has been provided that O-methylation, GSH conjugation, hydrogenation and dehydrogenation are involved in the hepatic metabolism of CA and DHCA. The O-methylation pathway for CA and DHCA is a detoxification route whereas o-quinones formation catalyzed by P450 is the toxification route.

Full text of DOI:10.1016/S0378-4274(02)00105-4

Toxicology Letters 133 (2002) 141–151
www.elsevier.com/locate/toxlet
Metabolism of caffeic acid by isolated rat hepatocytes and
subcellular fractions
Majid Y. Moridani a,b, Hugh Scobie b, Peter J. O’Brien b,*
a Department of Laboratory Medicine and Pathobiology, Mount Sinai Hospital, Uni6ersity of Toronto, 600 Uni6ersity A6enue,
Toronto, Ont., Canada M5G 1X5
b Faculty of Pharmacy, Uni6ersity of Toronto, 19 Russell Street, Toronto, Ont., Canada M5S 2S2
Received 4 December 2001; received in revised form 2 April 2002; accepted 2 April 2002
Abstract
Caffeic acid (CA) is found in a wide variety of foods such as vegetables, fruits, tea, coffee, and wine. However,
enzymes involved in its metabolism have not been identified. In the following, caffeic (CA), chlorogenic (CGA), and
dihydrocaffeic (DHCA) acids were incubated with hepatocytes and shown to undergo metabolism by cytochrome
P450, catechol-O-methyltransferase (COMT), and b-oxidation enzymes. Ferulic (FA) or dihydroferulic (DHFA)
acids, formed as the result of CA- or DHCA-O-methylation by COMT, were also O-demethylated by CYP1A1/2 but
not CYP2E1. DHCA or DHFA also underwent side chain dehydrogenation to form CA and FA, respectively, which
was prevented by thioglycolic acid, an inhibitor of the b-oxidation enzyme acyl CoA dehydrogenase. The rates of
glutathione conjugate formation catalyzed by NADPH/microsomes (CYP2E1) in decreasing order DHCA\CA\
CGA trend which was in reverse order to the rates of their O-methylation by COMT. The CA- and DHCA-o-
quinones formed by NADPH/P450 likely inhibited COMT but can readily form glutathione conjugates. CA, DHCA
and DHFA were inter-metabolized to each other and to FA by isolated rat hepatocytes whereas FA was metabolized
only to CA but not to DHCA or DHFA. CA, DHCA, FA, DHFA and CGA showed a dose-dependent hepatocyte
toxicity and the LD50 (2 h), determined were in decreasing order of effectiveness DHCA\CA\DHFA\CGA\
FA. In summary, evidence has been provided that O-methylation, GSH conjugation, hydrogenation and dehydro-
genation are involved in the hepatic metabolism of CA and DHCA. The O-methylation pathway for CA and DHCA
is a detoxification route whereas o-quinone formation catalyzed by P450 is the toxification route. © 2002 Elsevier
Science Ireland Ltd. All rights reserved.
Keywords: Caffeic acid; Chlorogenic acid; Ferulic acid; Hepatocyte; Metabolism; P450; COMT
1. Introduction
Caffeic acid (CA) is a catecholic acid that is
found naturally in a wide variety of foods such as
vegetables, fruits, tea, coffee, and wine (Buren et
al., 1973; Challis and Bartlett, 1975) with high
concentrations in certain foods and beverages.
* Corresponding author. Tel.: +1-416-978-2716; fax: +1-
416-978-8511.
E-mail address: peter.obrien@utoronto.ca (P.J. O’Brien).
0378-4274/02/$ - see front matter © 2002 Elsevier Science Ireland Ltd. All rights reserved.
PII: S0378-4274(02)00105-4

142
M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
For instance, red wine may contain 5 mg CA per
litre (Abu-Amsha et al., 1996). CA has been re-
ported to have anti-inflammatory, antimutagenic,
and anticarcinogenic activities (Challis and
Bartlett, 1975; Koshihara et al., 1984; Tanaka et
al., 1993a,b). Furthermore, there has been a grow-
ing interest in the multiple biological and pharma-
Previously, we showed that HRP/H2O2, tyrosi-
nase/O2, and liver microsomes prepared from
uninduced rats metabolized CA, chlorogenic acid
(CGA) and dihydrocaffeic acid (DHCA) (cate-
cholic acids) in vitro to form o- and p-quinones,
which were readily conjugated with glutathione or
formed hydroxylated adducts (Moridani et al.,
2001). We also reported hydrogen peroxide or
cumene hydroperoxide, P450 peroxidase sub-
strates (Anari et al., 1997a,b), or tyrosinase en-
hanced CA induced cytotoxicity in isolated rat
hepatocytes whereas benzylimidazole, a P450 in-
hibitor (Quan et al., 1992), prevented CA induced
hepatocyte cytotoxicity.
Further elucidation of the metabolic enzymes
involved in the human metabolism of CA and FA
should shed light on the role of different Phase I
(O-demethylation, hydrogenation and dehydro-
genation) and Phase II (O-methylation) enzymes
in the metabolism of dietary catecholic acids (CA
and DHCA) and phenolic acids (FA and dihydro-
ferulic acid (DHFA)), which have been previously
overlooked.
cological
properties
of
catecholic
acids
(Laranjinha et al., 1994) (Fig. 1).
Despite our dietary intake of CA very little is
known about its metabolism. There are many
conflicting pathways reported for the metabolism
of its O-methylated product ferulic acid (FA) or
its m-coumaric acid metabolite (Booth et al.,
1957; Gumbinger et al., 1993; Scheline, 1991).
Understanding the metabolic pathway for CA
may provide insight into the mechanism(s)
through which it exerts its beneficial physiological
effects.
In summary, evidence has been provided that
O-methylation, GSH conjugation, hydrogenation
and dehydrogenation are involved in the hepatic
metabolism of CA and DHCA. We have found
that the O-methylation pathway for CA and
DHCA is a detoxification route whereas o-qui-
none formation catalyzed by P450 is the toxifica-
tion route.
2. Materials and methods
2.1. Materials
CA, DHCA, FA, 3,4-dihydroxybenozoic acid,
CGA,
ethyl acetate (HPLC grade), hydrochloric acid
(HCl), acetic acid, S-adenosyl- -methionine
L-methionine, dimethyl sulfoxide (DMSO),
L
(SAM), glucose-6-phosphate, glucose-6-phosphate
dehydrogenase, trichloroacetic acid, Tris(hydrox-
ymethyl)aminomethane
aminopenta-acetic acid
(Tris),
diethylenetri-
5,
(DETAPAC),
5%-dithio-bis(2-nitrobenzoic acid) (DTNB), ben-
zylimidazole, phenylimidazole, magnesium chlo-
ride, sodium phosphate, sodium periodate,
Fig. 1. Chemical structure of catecholic acids (CA, CGA and
DHCA) and phenolic acids (FA and DHFA).

M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
143
nicotinamide adenosine dinucleotide phosphate
(oxidized form) nicotinamide
(NADP+),
HCl (1 M). The reaction mixtures were then
extracted into 2 ml of ice-cold ethyl acetate. After
centrifugation, a 1 ml aliquot of the organic ex-
tract was assayed for radioactivity by a liquid
scintillation counter (Packard, Tri-carb 2100TR).
Ethyl acetate extraction was found to be 99.8%
effective for the exclusion of unreacted [methyl-
3H] SAM.
adenosine dinucleotide phosphate (reduced form)
(NADPH), thioglycolic acid, glutathione (GSH),
sodium pentobarbital, pyrazole, 3-methylcholan-
threne, a-naphthoflavone and catechol O-methyl-
transferase (porcine liver) (COMT) were
purchased from Sigma-Aldrich Canada Co.
(Oakville, Ont.). Adenosine deaminase was ob-
tained from Roche Diagnostics (Laval, Quebec)
2.4. Preparation of induced li6er microsomes
and S-[methyl-3H]-adenosyl-
L-methionine (spe-
cific activity 5–15 Ci/mmol) was obtained from
NEN Life Science Products, Inc. (Boston, Massa-
chusetts). Stock solutions of CA, DHCA, FA,
DHFA, CGA, 3,4-DA, phenyl imidazole, ben-
zylimidazole were prepared in DMSO. All other
solutions were prepared in Millipore-filtered water
or buffer.
Adult male Sprague–Dawley rats, 250–300 g,
were obtained from Charles River Canada Labo-
ratories (Montreal, Quebec), fed ad libitum, and
allowed to acclimatize for 1 week on clay chip
bedding. Rats were injected intraperitoneally with
3-methylcholanthrene (3-MC) (25 mg/kg per day)
diluted in corn oil for 3 consecutive days to
induce CYP1A1/2 (Murphy et al., 1997). How-
ever, pyrazole (200 mg/kg per day) was injected
intraperitoneally to rats for 2 consecutive days to
induce CYP2E1 (Krikun and Cederbaum, 1984).
The animals were anesthetized by sodium pento-
barbital (60 mg/kg of body weight) 24 h following
last injection. Livers were removed under sterile
potassium buffer:KCl solution (1.18%, w/v, 4 °C)
as previously described (Anari et al., 1997a). Hep-
atic microsomes were prepared as described by
Dallner (1978).
2.2. Preparation of dihydroferulic acid
[3-(4-hydroxy-3-methoxyphenyl)propionic acid]
DHFA was prepared as previously described
with minor modification (Zhao et al., 1997). A
solution of FA (1 g) in ethanol (50 ml) was
hydrogenated over 10% Pd/C (100 mg) under 40
psi of hydrogen in a Parr apparatus (room tem-
perature, overnight). The reaction mixture was
filtered through filter paper, and concentrated in
100% yield. 1H-NMR (CD3OD) l=2.7 (t, 2H,
PhCH2), 2.9 (t, 2H, CH2), 6.5–6.8 (m, 3H, ArH);
positive ion electrospray-MS m/z (relative inten-
sity) [MH+] 197 (100%).
2.5. Glutathione conjugate formation catalyzed by
rat li6er microsomes
Glucose-6-phosphate (7.5 mM) was added to a
mixture of the test compound (0.5 mM), NADP+
(0.5 mM), magnesium chloride (5 mM), micro-
somes (1 mg/ml), glucose-6-phosphate dehydroge-
nase (2.5 U/ml), in 1 ml Tris–HCl buffer (0.1 M,
pH 7.4 containing 1 mM DETAPAC). When
GSH was present the GSH concentration was 0.5
mM. The mixture was pre-incubated at 37 °C
from which 250 ml samples were taken at 30 and
60 min and added to 25 ml trichloroacetic acid
(30% w/v), vortexed, left for 5 min, and cen-
trifuged to discard the protein pellet. An aliquot
of 62.5 ml DTNB (2 mg/ml prepared in Tris–HCl
buffer pH 8.94) was added to 100 ml of the
supernatant. The absorbance was monitored at
2.3. Catechol O-methylation assay
The COMT-catalyzed O-methylation of CA
and DHCA was carried out as previously de-
scribed (Borchardt, 1981). The reaction mixture
consisted of sodium phosphate buffer (0.1 M, pH
7.6), magnesium chloride (1.2 mM), 1.0 mM or
200 mM SAM containing 275 mCi [methyl-3H]
SAM, adenosine deaminase (10 mg/ml), and
DHCA or CA (2.0 mM) in a final volume of 150
ml. The reaction was started by the addition of
COMT (80 U/ml). The incubations were carried
out at 37 °C for 1 and 2 h. The O-methylation
reaction was terminated by the addition of 100 ml

144
M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
412 nm for CA and DHCA and at 460 nm for
CGA. The assay was repeated in the presence of
phenylimidazole (300 mM), a P450 heme ligand
(Quan et al., 1992).
2.9. Metabolism of catecholic and phenolic acids
catalyzed by hepatocytes
The test compounds (1 mM CA, DHCA,
DHFA, or FA) were added to the hepatocytes
(5×106 cells per ml) (10 ml), prepared as de-
scribed above, and was constantly in rotation
with a supply of 95% O2 at 37 °C for 4 h. The 10
ml sample of isolated hepatocytes was centrifuged
for 5 min and the pellet was discarded. The
supernatant was acidified to pH 2 by a HCl (1 M)
solution. The supernatant was extracted with 2×
10 ml ice-cold ethyl acetate. The ethyl acetate
fraction was evaporated to dryness and re-dis-
solved in methanol (1 ml). A 5 ml aliquot was
injected into the HPLC for analysis (see Section
2.10).
The experiment was repeated when hepatocytes
were pre-incubated for 30 min with the P450
inhibitor phenylimidazole (300 mM) (Quan et al.,
1992) or mitochondrial b-oxidation inhibitor thio-
glycolic acid (300 mM) (Boutellier et al., 1999)
prior to the addition of the test compound (1
mM).
2.6. O-Demethylation catalyzed by microsomes
CYP1A1/2 or CYP2E1 induced microsomes (4
mg/ml) were added to a solution of DHFA or FA
(1 mM), NADPH (1 mM) in 4 ml sodium phos-
phate buffer (0.1 M, pH 7.4). Incubations were
carried out at 37 °C for 1–4 h. The reaction was
stopped by removing the protein with centrifuga-
tion for 5 min. The supernatant was adjusted to
pH 2 using a HCl solution (1 M). The superna-
tant was extracted with 20 ml ice-cold ethyl ace-
tate. After a short centrifugation, the organic
extract was evaporated to dryness and re-dis-
solved in methanol (1 ml). A 5 ml aliquot was
analyzed by an HPLC method (see Section 2.10).
The experiment was repeated in the presence of
a-naphthoflavone, a CYP1A1/2 inhibitor (Zhou
et al., 2000).
2.10. HPLC analyses
2.7. Preparation of isolated rat hepatocytes
Preparation of the isolated hepatocytes (1×106
and 5×106 cells per ml) was performed using the
collagenase perfusion technique as described by
Moldeus et al. (1978). Cells were pre-incubated in
10 ml Krebs–Henseleit bicarbonate buffer (12.5
mM, pH 7.4) in a 50 ml round bottom flask for 30
min prior to the experiment and was constantly in
rotation with a supply of 95% O2 at 37 °C.
An aliquot of extracted sample (5 ml) was ana-
lyzed by HPLC using a Prodigy 5 mm ODS (3)
column (4.6×250 mm) (Phenomenex Co). A mix-
ture of 95.6% H2O, 4.1% ethyl acetate, and 0.3%
acetate was utilized as mobile phase at a flow rate
of 1.0 ml/min. A Shimadzu (SPD-6AV) UV–Vis
spectrophotometeric detector, a Shimadzu (LC-
6A) liquid chromatograph pump, and a Shimadzu
(C-R3A) Chromatopac integrator were used for
the HPLC analysis of the samples. The ab-
sorbance was monitored at 280 nm. Metabolites
(CA, DHCA, DHFA, and FA) (1 mM) were
injected into the column as standards for positive
identification or by co-elution with the hepatocyte
samples. All injections were performed in dupli-
cate. Quantification of the metabolites was carried
out through comparison with a standard curve
created from the injection of standard solutions of
known concentrations. 3,4-Dihydroxybenzoic acid
(DA) was used as an internal standard, and the
regression line used to calculate the amount of
each metabolite present.
2.8. Hepatocyte cytotoxicity
The dietary polyphenolics CA, DHCA, FA,
DHFA and CGA at 1–20 mM concentration
were incubated with the isolated rat hepatocytes
(1×106 cells per ml) (10 ml) prepared as de-
scribed above. The concentration (LD50 mM) at
which the polyphenolic compound caused 50%
cell death after incubation with the hepatocytes
for 2 h, was determined by measuring trypan blue
(0.1% w/v) exclusion by the hepatocytes with a
light microscope (Moldeus et al., 1978).

M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
145
Table 1
The COMT catalyzed O-methylation of CA, CGA and DHCA
Catechols
Amount of methylation (nmol)a
60 min
Vmax pmol/mg protein per minb
120 min
3,4-DA
CA
DHCA
CGA
91.790.7*
86.495.3*
66.992.0
163.493.0
117.593.1
77.491.5
188.991.1
1530
1440
1120
1940
116.593.1*
The reaction mixture consisted of sodium phosphate buffer (0.1 M, pH 7.6), magnesium chloride (1.2 mM), 1.0 mM or 200 mM
SAM containing 275 mCi [methyl-3H] SAM, adenosine deaminase (10 mg/ml), and catecholic acids (2.0 mM) in a final volume of 150
ml. The reaction was started by the addition of COMT (80 U/ml). The methylated catechols were then extracted with ethyl acetate
and counted (see Section 2). Ethyl acetate was 99.8% effective in excluding [methyl-3H] SAM from extraction and leaving it behind
in the aqueous solution. *, Significantly higher than DHCA (PB0.05).
a The data is average of three independent samples.
b The Vmax value was calculated from 60 min data.
2.11. Mass spectrometry analysis
3.2. O-Demethylation of FA/DHFA catalyzed by
microsomes
The extracted fractions from the hepatocyte
samples were centrifuged and used for mass spec-
trometry (MS) analyses. A 5 ml aliquot injected
directly into the mass spectrometer (PE Sciex9
II) and was analyzed in positive mode.
HPLC analysis of FA and DHFA incubated
with CYP1A1/2 induced liver microsomes. The
peaks at 17, 20, 33 and 38 min in the chro-
matogram corresponded to DHCA, CA, DHFA,
and FA, respectively. The rate of O-demethyla-
tion of FA or DHFA to form CA and DHCA,
respectively, was found to be linear over a 4 h
period. Incubation of CYP1A1/2 induced micro-
somes with a-naphthoflavone (a CYP1A1/2 in-
hibitor) resulted in a 90% decrease in DHCA and
CA formation. MS analysis of the sample pro-
vided evidence for metabolism of DHFA“
DHCA and FA“CA with identification of the
parent signals at m/z [M+1]+ 181, 183, 195 and
197 for CA, DHCA, FA, and DHFA,
respectively.
2.12. Statistical analysis
Statistical comparisons were carried out using a
two-way analysis of variance (ANOVA) to deter-
mine the significant differences between treat-
ments with a probability level of PB0.05. Results
represent the mean9S.D. of at least three inde-
pendent samples.
3. Results
The HPLC analysis of incubation mixtures of
FA or DHFA incubated with non-induced micro-
somes or CYP2E1 induced microsomes led to the
identification of single peaks corresponding only
to the original compounds.
3.1. O-Methylation of catecholic acids
Using a purified preparation of liver COMT
and [methyl-3H] SAM the rank order for O-
methylation found was CGA\3,4-DA\CA\
DHCA (Table 1). The rate and extent of
O-methylation for CGA was as much as twice of
that for DHCA. Decreasing SAM concentration
did not affect the rate of O-methylation. The rate
of O-methylation was more rapid during the first
hour of incubation.
3.3. GSH conjugate formation of catecholic acids
catalyzed by CYP2E1 induced microsomes
The amount of GSH depleted as a result of
dihydroxycinnamic acid oxidation catalyzed by
rat liver CYP2E1 induced microsomes/NADPH

146
M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
was determined to be 0.3, 0.8 and 1.0 equivalents
of GSH per mole of CGA, CA and DHCA,
respectively. This suggests that approximately 30,
80 and 100% of the corresponding compounds in
the reaction mixture underwent GSH conjugation.
The CGA, CA, and DHCA induced GSH deple-
tion by the CYP2E1 induced microsomes/
NADPH metabolic system was inhibited 85–91%
by benzylimidazole, a cytochrome P450 inhibitor.
toxic to the hepatocytes due to their very high
LD50 values.
3.5. Metabolism of phenolic and catecholic acids
by hepatocytes
The metabolism of CA, DHCA, FA, and
DHFA (1 mM) by isolated rat hepatocytes was
determined using HPLC and MS. An HPLC anal-
ysis of the metabolites formed from CA demon-
strated peaks at 17, 20, 33, and 38 min, which
corresponded to DHCA, CA, DHFA, and FA.
The major metabolite found was FA whereas the
minor metabolites were DHCA and DHFA
(Table 2). MS of the samples had parent signals at
m/z [M+1]+ 181, 183, 195 and 197 which corre-
sponded to CA, DHCA, FA, and DHFA, respec-
tively. Incubation of P450 inactivated hepatocytes
with CA resulted in a substantial increase in
DHCA formation and a moderate increase in
3.4. Hepatocyte cytotoxicity
CA, DHCA, FA, DHFA and CGA showed a
dose-dependent toxicity and their LD50 (mM),
determined as the concentrations required to in-
duce 50% cell death after incubation for 2 h with
the hepatocytes, were in decreasing order DHCA
(6 mM)\CA (10 mM)\DHFA (22 mM)\
CGA (23 mM)\FA (25 mM). FA and DHFA
the respective O-methylated products of CA and
DHCA, could practically be considered as non
Table 2
Summary of the quantification of the metabolites formed in isolated hepatocytes
Treatmenta
Metabolite concentration in hepatocyte supernatant (mM)b
DHCA
CA
DHFA
FA
CA (1 mM)c
CA+thioglycolic acid (300 mM)
CA+phenylimidazole (300 mM)
DHCA (1 mM)c
DHCA+thioglycolic acid (300 mM)
DHCA+phenylimidazole (300 mM)
DHFA (1 mM)c
DHFA+thioglycolic acid (300 mM)
DHFA+phenylimidazole (300 mM)
FA (1 mM)c
2890.4
6.091.2*¡
220912*
310920
300930
270950
2298.2
2596.1
1.090.1*¡
1.090.1
ND*¡
17090.1
210922*
150940
160930
130910*¡
140930
9396.2
120925
1192.7*¡
6397.4
2592.0
7.590.8*¡
3291.8*
7195.6
240930
210920*¡
280930*
7693.6
7894.1
110930*
62911
180920*
5296.1
2.590.26
7.591.2*
2294.2*
8293.4*
100920*
440912
390980
110920*
0.590.22
ND*¡
FA+thioglycolic acid (300 mM)
FA+Phenylimidazole (300 mM)
7191.2
3293.7*¡
ND*
8.090.58*
ND, not detected in sample. Hepatocytes (5×106 cells per ml) were incubated in 10 ml Krebs–Henseleit bicarbonate buffer (12.5
mM, pH 7.4) for 30 min prior to the experiment. The test compound (CA, DHCA, DHFA, or FA) (1 mM) and -methionine (300
L
mM) were added to the hepatocytes under a supply of 5% CO2, 95% O2 at 37 °C for 4 h. The hepatocytes were centrifuged and the
pellet was discarded. The supernatant pH was adjusted to 2, extracted with 2×10 ml ice-cold ethyl acetate, evaporated and
re-dissolved in methanol (1 ml). A 5 ml aliquot was injected into the HPLC for analysis. The experiment was repeated when
hepatocytes were pre-incubated with the P450 inhibitor phenylimidazole (300 mM) and the mitochondrial b-oxidation inhibitor
thioglycolic acid (300 mM) for 30 min prior to the addition of the test compound (1 mM). *, Significantly different from controls
(PB0.05).
a Thioglycolic acid, phenylimidazole, CA, DHCA, DHFA and FA at the given dose were not toxic to the isolated rat hepatocytes.
b The data are expressed as the mean9S.D. of at least three separate determinations.
c Controls.

M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
147
Fig. 2. Proposed metabolic fate of CA in isolated rat hepatocytes.
The major metabolites formed as a result of
DHFA incubation with isolated rat hepatocytes
were CA and FA as identified by MS, whereas
DHCA was a minor metabolite (Table 2). When
FA was incubated with hepatocytes CA and trace
amounts of DHFA and DHCA were formed.
Inhibition of P450 resulted in a substantial de-
crease in CA and DHCA formation thereby confi-
rming the involvement of P450 in the
O-demethylation of DHFA and FA. Inhibiting
mitochondrial b-oxidation with thioglycolic acid
prevented the conversion of FA to DHFA and
DHCA whereas the level of CA was only mar-
ginally elevated likely because more FA was avail-
able for O-demethylation. However, DHFA
metabolism to DHCA and CA formation was not
affected by thioglycolic acid. The apparent 3-fold
increase in the FA metabolite could be due to the
formation of an unknown metabolite with the
same retention time as FA.
DHFA and FA formation. Inhibiting mitochon-
drial b-oxidation with thioglycolic acid largely
prevented the conversion of CA to DHCA
whereas DHFA and FA formation was decreased.
These findings indicate that CA metabolism to
other metabolites such as o-quinone was pre-
vented as a result of P450 inhibition and, there-
fore, more CA was available for dehydrogenation
and O-methylation metabolism pathways.
Incubation of isolated hepatocytes with DHCA
formed CA as a major metabolite and DHFA and
FA as the minor metabolites. Treatment of P450
inactivated hepatocytes with DHCA showed a
30% increase in the formation of DHFA and FA
products but not CA. This was likely because
DHCA could not be oxidized to the o-quinone as
a result of P450 inhibition so that more DHCA
was available for metabolism to DHFA and FA.
Inhibition of mitochondrial b-oxidation resulted
in a 20% decrease in CA formation from DHCA.
This was because more DHCA was available for
the O-methylation pathway and resulted in 15%
more DHFA formation (Fig. 2).
As shown in Table 2, the extent at which
DHCA was metabolized to CA (DHCA“CA)
was 6-fold greater than that of CA metabolism to

148
M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
DHCA (CA“DHCA). Furthermore, the extent
at which DHFA was metabolized to FA
(DHFA“FA) was 100-fold greater than that of
FA metabolism to DHFA (FA“DHFA). These
findings show DHCA and DHFA dehydrogena-
tion was substantially more than the CA and FA
hydrogenation.
methylation by COMT. The CA was found to be
methylated and the rank order of methylation
found was CGA\3,4-DA\CA\DHCA. It was
found that CGA with its unique quinic ester
moiety displayed a 2-fold higher rate of O-methy-
lation when compared with DHCA.
Interestingly, we were able to correlate the rate
of O-methylation inversely to the cytotoxicity
(LD50) (2 h) as well as the amount of GSH
conjugate formation by NADPH supported mi-
crosomes. Furthermore, we showed that DHCA
o-quinone (100 mM), formed by the addition of
equivalent amount sodium periodate to a solution
of DHCA, inhibited the DHCA O-methylation
catalyzed by COMT (data not shown). Therefore,
the oxidized products of catechols (CA and
DHCA) formed by NADPH/P450 are likely in-
hibitors of COMT. We also showed that CYP2E1
induced microsomes and NADPH catalyzed the
oxidation of DHCA and CA resulting in GSH
depletion (GSH conjugate formation) much more
effectively than uninduced microsomes (Moridani
et al., 2001).
Previously the Vmax for the O-methylation of
3,4-DA, CA and DHCA were published as 53, 62
and 248 nmol/mg protein per min, respectively,
but the Vmax for CGA was not measured (Crevel-
ing et al., 1972). The rates were measured in vitro
using COMT that had been purified from rat
livers. However, our data suggests that the rate of
O-methylation for CA and DHCA catalyzed by
COMT (porcine) in vitro is much more rapid if
adenosine deaminase is added to the reaction
mixture so as to remove S-adenosylhomocysteine,
the demethylated product of SAM. S-Adenosyl-
homocysteine has previously been shown to effec-
tively inhibit COMT at a concentration as low as
25 mM (Zhu et al., 1994). Adenosine deaminase
converts the COMT inhibitory product S-adeno-
sylhomocysteine to its deaminated product S-
adenosylhomocysteine which was previously
found not to inhibit COMT (Coward and Slisz,
1973).
4. Discussion
CA is structurally a relatively simple com-
pound, but contains several metabolically active
sites, a property that leads to the formation of a
large number of metabolites. Previously we
showed that HRP/H2O2, tyrosinase/O2, and unin-
duced rat liver microsomes/NADPH metabolized
CA, CGA and DHCA in vitro to form o- and
p-quinones, which formed GSH conjugates or
hydroxylated adducts using mass spectrometry
technique (Moridani et al., 2001).
Further elucidation of the metabolic enzymes
involved in the metabolism of CA, DHCA, FA
and DHFA should shed light on the role of
different Phase
I metabolizing enzymes (O-
demethylation, hydrogenation and dehydrogena-
tion) and Phase II metabolizing enzymes
(O-methylation and GSH conjugation). However,
we did not include the sulfate and glucuronide
conjugations in this communication because the
prime focus of our studies was the inter-
metabolism of CA, DHCA, DHFA and FA.
The O-methylation of these hydroxycinnamic
acids through the catalytic activity of COMT,
although not demonstrated before, has been con-
sidered to be a major route of metabolism
(Bourne and Rice-Evans, 1998a,b; Scheline, 1991).
Despite this, CA has been called an in vitro
inhibitor of COMT, although significant amounts
of FA were found in the urine following the oral
administration of CA to rats (Guldberg and
Marsden, 1975; Booth et al., 1957). Measurement
of the in vitro activity of COMT catalyzing the
O-methylation of CA proved to be difficult using
less sensitive analytical techniques such as UV-Vis
spectroscopy and HPLC. A radioassay was, there-
fore, employed in the current study to accurately
measure the rates of CA, CGA and DHCA O-
Since the liver is the site for the widest variety
of biotransforming enzymes, hepatocytes were se-
lected as a model for the metabolism of the plant
phenolics and catecholics. The results of previous
studies using whole animals were comparable with
those found here for the metabolism of CA by

M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
149
hepatocytes. The use of hepatocytes to study the
metabolism of CA has not been explored before
despite the fact that it provides an excellent, reli-
able model for hepatic metabolism. We have
shown that hepatocytes catalyze the O-demethyla-
tion of FA and DHFA which was prevented by
the addition of the P450 inhibitor phenylimida-
zole (Quan et al., 1992). We have shown for the
first time that 3-MC induced microsomes
(CYP1A1/2) but not pyrazole induced micro-
somes (CYP2E1) catalyzed the O-demethylation
of FA“CA and DHFA“DHCA. Previous re-
search speculated that the conversion of CA to
FA was an irreversible reaction despite the lack of
evidence (Scheline, 1991). However, our research
suggests that this is not the case and DHFA, FA
undergoes O-demethylation to form CA. Induc-
tion of CYP1A1/2 by environmental or other
dietary substances may, therefore, affect the hep-
atic metabolism of dietary hydroxycinnamic acids
(CA, DHCA, FA and DHFA).
Our results suggest that the hepatic metabolism
of the plant phenolics (FA and DHFA) may
involve several enzymes (Fig. 2). Catecholic cin-
namic acids (CA and DHCA) can be metabolized
by COMT to form their corresponding O-methy-
lated products, FA and DHFA, respectively. We
have speculated that DHCA and DHFA may
undergo an oxidation reaction with the b-oxida-
tion enzyme acyl CoA dehydrogenase to form
their unsaturated products CA and FA,
respectively.
A subject of special interest in the study of CA
metabolism has been the role of intestinal bacte-
ria, which have been suggested to be solely re-
sponsible for several metabolic reactions
(Scheline, 1991). Other investigators using germ-
free rats suggested that the hydrogenation of the
CA and FA double bond is the result of anaero-
bic bacterial metabolism (Peppercorn and Gold-
man, 1972; Scheline, 1991). Thioglycolic acid
proved to be an effective in vivo inhibitor of fatty
acid b-oxidation and has been studied for use in
food intake control (Boutellier et al., 1999). When
thioglycolic acid was co-administered with CA to
isolated hepatocytes the major metabolite formed
was FA and the formation of the other metabo-
lites were significantly decreased. Our results show
that isolated hepatocytes readily catalyze the de-
hydrogenation of DHCA and DHFA to CA and
FA, respectively. This dehydrogenation reaction is
catalyzed by acyl CoA dehydrogenase of the fatty
acid b-oxidation pathway as it was inhibited by
thioglycolic acid, an inhibitor of b-oxidation. Fur-
thermore, because DHCA and DHFA were
formed from CA and FA, respectively, we con-
cluded for the first time that hydrogenation of the
double bond could also occur by an unidentified
hydrogenase present in the hepatocytes even un-
der aerobic (95% O2/5% CO2 atmosphere)
conditions.
Our findings indicate that the extent of dehy-
drogenation for DHCA and DHFA is substan-
tially more than the hydrogenation for CA and
FA. Briefly, CA can undergo both hydrogenation
and O-methylation at the same time whereas FA
can mainly undergo O-demethylation but not hy-
drogenation/reduction of its double bond. DHCA
can undergo both dehydrogenation and O-methy-
lation on incubation with rat hepatocytes. DHFA
on the otherhand, undergoes dehydrogenation
and O-demethylation (Fig. 2). FA was less readily
metabolized than CA, DHCA, and DHFA. The
O-methylation pathway is a detoxification route
in the hepatocytes as is evident from their LD50
values. In addition, CA and DHCA can undergo
oxidation much more readily by NADPH/P450 to
form o-quinones which can then form conjugates
with GSH (Moridani et al., 2001) (Fig. 2).
It has been shown that only 7% of the CA was
metabolised when CA was perfused once through
the isolated rat liver. However, only FA (1%),
isoferulic acid (0.2%) and esculetin (0.3%) were
identified and it was suggested that glucuronides
and sulfates may also be formed and released into
the bile (Gumbinger et al., 1993). Experiments are
underway to determine the extent CA undergoes
glucuronidation or sulfation when incubated with
hepatocytes.
In summary, evidence has been provided that
O-methylation, GSH conjugation, hydrogenation
and dehydrogenation are involved in the hepatic
metabolism of caffeic and DHCA acids. The O-
methylation pathway for CA and DHCA is a
detoxification route whereas o-quinone formation
catalyzed by P450 is the toxification route.

150
M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
Creveling, C.R., Morris, N., Shimizu, H., Ong, H.H., Daly, J.,
Acknowledgements
1972. Catechol O-methyltransferase. IV. Factors affecting
m- and p-methylation of substituted catechols. Mol. Phar-
macol. 8, 398–409.
This work was financially supported by a grant
received from the Natural Sciences and Engineer-
ing Research Council of Canada. The authors
wish to thank Dr Ling-Jie Meng, Medical Sci-
ences Building, University of Toronto, Canada,
for the technical assistance received with MS
analyses and Ford Barker for his excellent skills in
the preparation of isolated rat hepatocytes.
Dallner, G., 1978. Isolation of microsomal subfractions by use
of density gradients. Methods Enzymol. 52, 71–83.
Guldberg, H.C., Marsden, C.A., 1975. Catechol-O-methyl
transferase: pharmacological aspects and physiological
role. Pharmacol. Rev. 27, 135–206.
Gumbinger, H.G., Vahlensieck, U., Winterhoff, H., 1993.
Metabolism of caffeic acid in the isolated perfused rat
liver. Planta Med. 59, 491–493.
Koshihara, Y., Neichi, T., Murota, S., Lao, A., Fujimoto, Y.,
Tatsuno, T., 1984. Caffeic acid is a selective inhibitor for
leukotriene biosynthesis. Biochim. Biophys. Acta 792, 92–
97.
References
Krikun, G., Cederbaum, A.I., 1984. Increased microsomal
oxidation of alcohols after pyrazole treatment and its
similarities to the induction by ethanol consumption.
Biochim. Biophys. Acta 801, 131–137.
Laranjinha, J.A., Almeida, L.M., Madeira, V.M., 1994. Reac-
tivity of dietary phenolic acids with peroxyl radicals: an-
tioxidant activity upon low density lipoprotein
peroxidation. Biochem. Pharmacol. 48, 487–494.
Moldeus, P., Hogberg, J., Orrenius, S., 1978. Isolation and use
of liver cells. Methods Enzymol. 52, 60–70.
Abu-Amsha, R., Croft, K.D., Puddey, I.B., Proudfoot, J.M.,
Beilin, L.J., 1996. Phenolic content of various beverages
determines the extent of inhibition of human serum and
low-density lipoprotein oxidation in vitro: identification
and mechanism of action of some cinnamic acid derivatives
from red wine. Clin. Sci. (Colch) 91, 449–458.
Anari, M.R., Josephy, P.D., Henry, T., O’Brien, P.J., 1997a.
Hydrogen peroxide supports human and rat cytochrome
P450
1A2-catalyzed
2-amino-3-methylimidazo[4.,5-
f ]quinoline bioactivation to mutagenic metabolites: signifi-
cance of cytochrome P450 peroxygenase. Chem. Res.
Toxicol. 10, 582–588.
Moridani, M.Y., Scobie, H., Jamshizadeh, A., Salehi, P.,
O’Brien, P.J., 2001. Caffeic acid, chlorogenic acid and
dihydrocaffeic acid metabolism: glutathione conjugate for-
mation. Drug Metab. Dispos. 29, 1432–1439.
Murphy, S.E., Nunes, M.G., Hatala, M.A., 1997. Effects of
phenobarbital and 3-methylcholanthrene induction on the
formation of three glucuronide metabolites of 4-(methylni-
trosamino)-1-(3-pyridiyl)-1-butanone. NNK Chem. Biol.
Interact. 103, 153–166.
Anari, M.R., Khan, S., Jatoe, S.D., O’Brien, P.J., 1997b.
Cytochrome P450 dependent xenobiotic activation by
physiological hydroperoxides in intact hepatocytes. Eur. J.
Drug Metab. Pharmacokinet. 22, 305–310.
Booth, A.N., Emerson, O.H., Jones, F.T., DeEds, F., 1957.
Urinary metabolites of caffeic and chlorogenic acids. J.
Biol. Chem. 229, 51–59.
Peppercorn, M.A., Goldman, P., 1972. Caffeic acid
metabolism by gnotobiotic rats and their intestinal bacte-
ria. Proc. Natl. Acad. Sci. USA 69, 1413–1415.
Borchardt, R.T., 1981. Catechol O-methyltransferase. Meth-
ods Enzymol. 77, 267–272.
Bourne, L.C., Rice-Evans, C., 1998a. Bioavailability of ferulic
acid. Biochem. Biophys. Res. Commun. 253, 222–227.
Bourne, L.C., Rice-Evans, C.A., 1998b. Urinary detection of
hydroxycinnamates and flavonoids in humans after high
dietary intake of fruit. Free Radic. Res. 28, 429–438.
Boutellier, S., Lutz, T.A., Volkert, M., Scharrer, E., 1999.
2-Mercaptoacetate, an inhibitor of fatty acid oxidation,
decreases the membrane potential in rat liver in vivo. Am.
J. Physiol. 277, R301–305.
Buren, J.V., Vos, L.D., Pilnik, W., 1973. Measurement of
chlorogenic acid and flavonol glycosides in apple juice by a
chromatographic–fluorometric method. J. Food Sci. 38,
656–658.
Challis, B.C., Bartlett, C.D., 1975. Possible cocarcinogenic
effects of coffee constituents. Nature 254, 532–533.
Coward, J.K., Slisz, E.P., 1973. Analogs of S-adenosylhomo-
cysteine as potential inhibitors of biological transmethyla-
tion. Specificity of the S-adenosylhomocysteine binding
site. J. Med. Chem. 16, 460–463.
Quan, Z., Khan, S., O’Brien, P.J., 1992. Role of cytochrome
P-450 IIE1 in N-nitroso-N-methylaniline induced hepato-
cyte cytotoxicity. Chem. Biol. Interact. 83, 221–233.
Scheline, R., 1991. CRC Handbook of Mammalian
Metabolism of Plant Compounds. CRC Press, Boca Ra-
ton, FL.
Tanaka, T., Kojima, T., Kawamori, T., Wang, A., Suzui, M.,
Okamoto, K., Mori, H., 1993a. Inhibition of 4-nitroquino-
line-1-oxide-induced rat tongue carcinogenesis by the natu-
rally occurring plant phenolics caffeic, ellagic, chlorogenic
and ferulic acids. Carcinogenesis 14, 1321–1325.
Tanaka, T., Kojima, T., Kawamori, T., Yoshimi, N., Mori,
H., 1993b. Chemoprevention of diethylnitrosamine-in-
duced hepatocarcinogenesis by a simple phenolic acid pro-
tocatechuic acid in rats. Cancer Res. 53, 2775–2779.
Zhao, H., Neamati, N., Mazumder, A., Sunder, S., Pommier,
Y., Burke, T.R. Jr, 1997. Arylamide inhibitors of HIV-1
integrase. J. Med. Chem. 40, 1186–1194.

M.Y. Moridani et al. / Toxicology Letters 133 (2002) 141–151
151
Zhou, S., Paxton, J.W., Tingle, M.D., Kestell, P., 2000. Iden-
tification of the human liver cytochrome liver cytochrome
P450 isoenzyme responsible for the 6-methylhydroxylation
of the novel anticancer drug 5,6-dimethylxanthnone-4-ace-
tic acid. Drug Metabol. Dispos. 28, 1449–1456.
Zhu, B.T., Ezell, E.L., Liehr, J.G., 1994. Catechol-O-methyl-
transferase-catalyzed rapid O-methylation of mutagenic
flavonoids. Metabolic inactivation as a possible reason for
their lack of carcinogenicity in vivo. J. Biol. Chem. 269,
292–299.