The relative contribution of human cytochrome P450 isoforms to the four caffeine oxidation pathways: An in vitro comparative study with cDNA-expressed P450s including CYP2C isoforms

Article abstract of DOI:10.1016/j.bcp.2008.05.025

The aim of the present study was to estimate the relative contribution of cytochrome P450 isoforms (P450s), including P450s of the CYP2C subfamily, to the metabolism of caffeine in human liver. The experiments were carried out in vitro using cDNA-expresse

Full text of DOI:10.1016/j.bcp.2008.05.025

b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/biochempharm
The relative contribution of human cytochrome P450
isoforms to the four caffeine oxidation pathways:
An in vitro comparative study with cDNA-expressed P450s
including CYP2C isoforms
*
Marta Kot, Władysława A. Daniel
Polish Academy of Sciences, Institute of Pharmacology, Sme˛tna 12, PL 31-343 Krako´w, Poland
a r t i c l e i n f o
a b s t r a c t
Article history:
The aim of the present study was to estimate the relative contribution of cytochrome P450
isoforms (P450s), including P450s of the CYP2C subfamily, to the metabolism of caffeine in
human liver. The experiments were carried out in vitro using cDNA-expressed P450s, liver
microsomes and specific P450 inhibitors. The obtained results show that (1) apart from the 3-
N-demethylation of caffeine – a CYP1A2 marker reaction and the main oxidation pathway of
caffeine in man – 1-N-demethylation is also specifically catalyzed by CYP1A2 (not reported
previously); (2) 7-N-demethylation is catalyzed non-specifically, mainly by CYP1A2 and, to a
smaller extent, by CYP2C8/9 and CYP3A4 (and not by CYP2E1, as suggested previously); (3) C-
8-hydroxylation preferentially involves CYP1A2 and CYP3A4 and, to a smaller degree,
CYP2C8/9 and CYP2E1 (and not only CYP3A, as suggested previously) at a concentration
of 100 mM corresponding to the maximum therapeutic concentration in humans. At a higher
caffeine concentration, the contribution of CYP1A2 to this reaction decreases in favour of
CYP2C8/9. The obtained data show for the first time the contribution of CYP2C isoforms to
the metabolism of caffeine in human liver and suggest that apart from 3-N-demethylation,
1-N-demethylation may also be used for testing CYP1A2 activity. Moreover, they indicate
that the C-8-hydroxylation is not exclusively catalyzed by CYP3A4.
Received 1 April 2008
Accepted 19 May 2008
Keywords:
Caffeine metabolism
Human cytochrome P450
cDNA-expressed isoforms
Liver microsomes
Cytochrome P450 inhibitors
# 2008 Elsevier Inc. All rights reserved.
1.
Introduction
specifically catalyzed by CYP1A2; moreover, they suggest that
other oxidation pathways of caffeine may be mediated – at
Caffeine (1,3,7-trimethylxanthine), the most widely used
psychoactive substance [1,2], recently found to possess
neuroprotective properties [3–5], is an established marker
substrate for testing CYP1A2 activity using 3-N-demethylation
in humans [6,7] and C-8-hydroxylation in rats [8].
least partly – by P450 isoforms different from CYP1A2. Using
liver microsomes and CYP1A specific inhibitors or antibodies,
it has been shown that human CYP1A2 plays a pivotal role in
caffeine metabolism, especially in catalyzing N-demethyla-
tion reactions [9–12]. Further studies with selected cDNA-
expressed CYP isoforms or CYP1A2 and CYP2E1 cell lines have
indicated that caffeine 3-N-demethylation is most efficiently
Numerous studies show that 3-N-demethylation to para-
xanthine (the main oxidation pathway) in humans is
* Corresponding author. Tel.: +48 12 6623266; fax: +48 12 6374500.
E-mail address: nfdaniel@cyf-kr.edu.pl (W.A. Daniel).
Abbreviations: CYP, cytochrome P450; HPLC, high performance liquid chromatography; DDC, diethyldithiocarbamate; FUR, furafylline;
SULF, sulfaphenazole; KET, ketoconazole; P450, cytochrome P450.
0006-2952/$ – see front matter # 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2008.05.025

544
b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
catalyzed by CYP1A2, while the CYP3A subfamily is the main
isoenzyme catalyzing C-8-hydroxylation to 1,3,7-trimethylu-
ric acid. In addition, CYP2E1 may also contribute to 1-N- and 7-
N-demethylation [13–18]. However, CYP2C isoforms (CYP2C8/
9/18/19) which constitute over 20% of total human liver P450
have not been studied in this respect so far. For this reason
precise qualitative and relative quantitative estimation of the
contribution of individual P450 isoforms to the specific
metabolic pathways of caffeine was not possible.
2.2.
studies
Human liver microsomes–kinetic and inhibition
The liver microsomes from individual patients HK23, HK37,
HG56, HG89, HG93 and the pooled liver microsomes from
patients HH02, HH03, HH04, HH10, HH12, HH16, HH24, HH25,
HH32, HH33, HH41, HH49, HH55, HH64, HH68, HH72, HH74,
HH89 were obtained from Gentest Co. (Woburn, MA, USA)
(Table 1).
Since our earlier studies with rats revealed that P450s of the
CYP2C subfamily were important for caffeine metabolism
(in particular for 7-N-hydroxylation) and that the contribution
of these isoforms to caffeine metabolism was concentration-
dependent [8,19], we attempted to carry out a complementary
investigation comprising all human drug-metabolizing P450s,
including CYP2C isoforms. The present comparative study
carried out on a full set of human cDNA-expressed P450s
allowed us to estimate for the first time the relative
contribution of individual P450s to the four oxidative meta-
bolic reactions of caffeine. The obtained results show that
CYP2C8/9 isoforms substantially contribute to the 7-N-
demethylation and C-8-hydroxylation of caffeine (Fig. 1).
Studies into caffeine metabolism in human liver micro-
somes were carried out at the linear dependence of product
formation on time, protein and substrate concentration. To
determine enzyme kinetics (Eadie–Hofstee plots), the caffeine
concentrations used ranged from 0.05 to 1.6 mM. For inhibi-
tion studies, 100 mM caffeine was incubated with pooled liver
microsomes in the absence or presence of one of the selective
inhibitors added in vitro
: 10 mM furafylline (a CYP1A2
inhibitor), 10 mM sulfaphenazole (a CYP2C9 inhibitor), 2 mM
ketoconazole (a CYP3A inhibitor). Incubation was carried out
in a system containing liver microsomes (ca. 0.5 mg of protein/
ml), a phosphate buffer (0.15 M, pH 7.4) and NADPH (1 mM).
The final incubation volume was 1 ml. After a 2-min (with
furafylline) or 3-min (with sulfaphenazole or ketoconazole)
preincubation at 37 8C, the reaction was initiated by adding
caffeine. After a 60-min incubation, the reaction was termi-
nated by adding 700 ml of a 2% ZnSO4 and 50 ml of 2 M HCl.
Caffeine and its metabolites were analyzed by the high-
performance liquid chromatography method (HPLC) described
below.
2.
Materials and methods
2.1.
Drugs and chemicals
Caffeine and its metabolites—theobromine, paraxanthine,
theophylline and 1,3,7-trimethyluric acid, as well as furafyl-
line, sulfaphenazole, ketoconazole and NADPH were pur-
chased from Sigma (St. Louis, USA). All the organic solvents
with HPLC purity were supplied by Merck (Darmstadt,
Germany).
2.3.
cDNA-expressed human P450s
Microsomes from baculovirus-infected insect cells expres-
sing CYP1A2, CYP2A6, CYP2B6, CYP2C8/9/18/19, CYP2D6,
Fig. 1 – The main metabolic pathways of caffeine and the contribution of P450 isoforms (based on the present study).

b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
545
Table 1 – Clinical characteristics of liver specimen donors
PatientAge (years) Gender
Cause of death
Medical history
Drugs
HH02
HH03
HH04
HH10
HH12
54
38
66
84
66
Male
Ascending aorta aneurysm
Head injury
Not available
Not available
Not available
Not available
Male
Allergy based asthma
Female
Male
Intracranial bleeding
Intracranial hemorrhage
Cardiac arrest
Hypertension
Hypertension, asthma, arthritis
Coronary artery disease,
peripheral vascular disease,
hypertension, bilateral endarectomy,
ischemic cardiomyopathy,
diabetes mellitus
Heparin, dopamine, pitressin
Dopamine, neosynephrine, T4
Female
HH16
54
Male
Anoxia
Myasthenia-gravis, appendectomy
Levoquin, clindamycin, mannitol,
dopamine, esmolul, nipride, T4
Not available
HK23
HH24
HH25
24
53
66
Male
Head injury
Not available
Female
Female
Cerebral vascular accident
Intracerebral hemorrhage
St. John’s wort
Ancef, heparin
Hypertension, stomach ulcer,
rheumatoid arthritis
Diabetes, hypertension,
hypothyroid, angina
Not available
HH32
77
Female
Intracranial hemorrhage
Glucotrol, synthroid, digoxin,
levostatin, ticlid, glipizide,
lopressor
HH33
HK37
HH41
65
56
58
Female
Female
Male
Intracranial bleeding
Brain hemorrhage
Not available
Not available
Not available
Not available
Pulmonary hypertension
Chronic heart failure, emphysema,
cardiac and respiratory problems
Hysterectomy, occasional depression
Hypertension
Not available
HH49
HH55
HG56
HH64
44
52
57
77
Female
Male
Intracranial bleeding
Cerebral vascular accident
Aneurysm
Welburtin
Nipride, dopamine, atropine, ancef
Not available
Female
Male
Not available
Cerebral vascular accident
Hypertension, mild diabetes,
gallbladder removed
Heparin, levophed, insulin,
propofol, lopressor, zantac,
vasopresson
HH68
HH72
HH74
HH89
HG89
HG93
47
54
55
33
71
45
Male
Head trauma
Not available
Not available
Not available
Not available
Not available
Not available
Dilantin, vadopressin, neosynepione
Not available
Male
Intracranial bleeding
Intracranial bleeding
Stroke
Male
Not available
Male
Not available
Female
Female
Intracranial hemorrhage
Closed head trauma
Not available
Dopamine, pitressin,
levofed, synthroid
CYP2E1 and CYP3A4/5 co-expressed with NADPH P450
oxidoreductase (Supersomes) were obtained from Gentest
Co. (Woburn, MA, USA). CYP2E1 was also co-expressed with
cytochrome b5. Microsomal protein expressing NADPH P450
oxidoreductase and cytochrome b5 was used as a control.
Studies into caffeine metabolism in the Supersomes were
carried out at the linear dependence of product formation
on time and the amount of P450 isoforms and substrate
concentration. Caffeine metabolism was studied under
experimental conditions similar to those described for liver
microsomes using 50, 100, 200, 400, 800 and 1600 mM
caffeine, except for the fact that the final concentration
and incubation time of P450s was 100 pmol/ml and 60 min
for CYP1A2, CYP2B6, CYP2C8/9/18/19, CYP2D6, CYP2E1 and
CYP3A4, but 200 pmol/ml and 15 min for CYP2A6 and
CYP3A5.
(8:1, v/v). The residue yielded after evaporation of the
microsomal extract was dissolved in 100 ml of the mobile
phase described below. An aliquot of 20 ml was injected into
the HPLC system. The Merck–Hitachi chromatograph,
‘‘LaChrom’’ (Darmstadt, Germany), equipped with a L-7100
pump, an UV detector and
a D-7000 System Manager
was used. The analytical column (Supelcosil LC-18,
15 cm  4.6 mm, 5 mm) was from Supelco (Bellefonte, USA).
The mobile phase consisted of 0.01 M acetate buffer (pH 3.5)
and methanol (91:9, v/v). The flow rate was 1 ml/min
(0–26.5 min), followed by 3 ml/min (26.6–35 min). The column
temperature was maintained at 30 8C. The absorbance of
caffeine and its metabolites was measured at a wavelength of
254 nm. The compounds were eluted in the following order:
theobromine (9.7 min), paraxanthine (15.8 min), theophylline
(16.9 min), 1,3,7-trimethyluric acid (23.4 min), caffeine
(30.5 min).
2.4.
Determination of caffeine and its metabolites
Caffeine and its four primary metabolites were assessed
using the HPLC method based on the procedure of Rasmussen
et al. [20], as described previously [21]. Briefly, after incuba-
tion, samples were centrifuged and the water phase contain-
ing caffeine and its metabolites was extracted with 6 ml of an
organic mixture consisting of ethyl acetate and 2-propanol
3.
Results
3.1.
Caffeine metabolism in human liver microsomes
It was shown that the 3-N-demethylation of caffeine was a
major metabolic reaction in human liver microsomes (ꢀ70%)

546
b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
Fig. 2 – The Eadie–Hofstee plots for caffeine 1-N-demethylation (A), 3-N-demethylation (B), 7-N-demethylation (C), C-8-
hydroxylation (D) in human liver microsomes. Human liver microsomes of donor HG56 were incubated in a 0.15 M
phosphate buffer (pH 7.4) with caffeine (0.05–1.6 mM) and NADPH (1 mM) for 60 min.
compared to 1-N- and 7-N-demethylation (7–8%) and C-8-
hydroxylation (ꢀ15%), as measured at the substrate concen-
tration of 100 mM. Fig. 2A–D shows the Eadie–Hofstee plots for
caffeine oxidation processes in liver microsomes of donor
HG56. The plot of 3-N-demethylation was linear, which
suggests a one-enzyme catalysis of the reaction. All the other
plots were non-linear (the curvature of the 1-N-demethylation
plot being very gentle, though), which indicates a multiple-
enzyme catalysis. Similar results (with regard to the form of
the curvature and quantitative results) were observed with
microsomes of donors HK23, HK37, HG89 and HG93.
CYP1A2 > CYP2D6 > CYP2C8 > CYP2C19 > CYP3A5 > CYP2C9
> CYP2B6 = CYP2C18 > CYP3A4 > CYP2A6 for 1-N-demethyla-
tion;
CYP1A2 > CYP3A4 > CYP2C9 > CYP2C8 > CYP2D6 > CYP2E1 >
CYP2C19 > CYP2C18 > CYP2B6 > CYP2A6 > CYP3A5 for 3-N-
demethylation;
CYP1A2 > CYP2D6 > CYP2C9 > CYP2E1 > CYP2A6 > CYP3A5 >
CYP2C19 > CYP2C8 > CYP2B6 > CYP2C18 > CYP3A4 for 7-N-
demethylation;
CYP1A2 > CYP2B6 > CYP2E1 > CYP2D6 > CYP2C18 > CYP3A4
> CYP3A5 > CYP2C19 > CYP2A6 > CYP2C8 > CYP2C9 for C-8-
hydroxylation. The ability of human cDNA-expressed P450s to
metabolize caffeine at the 100 mM concentration is shown in
Fig. 3A–D.
3.2.
Study with human cDNA-expressed P450s
Kinetic parameters showing caffeine metabolism in human
cDNA-expressed P450s, obtained by a non-linear analysis
(Program Sigma Plot 8.0; Enzyme Kinetics), are presented in
Table 2. These kinetic parameters indicate distinct inter-
isoform differences and are consistent with the Eadie–Hofstee
plots derived from liver microsomes (Fig. 2). CYP1A2 showed
the biggest intrinsic clearance (Vmax/Km) towards caffeine
metabolism, its highest value being reached for 3-N-demethy-
lation.
3.3.
Quantitative estimation of the contribution of P450
enzymes to the specific metabolic pathways of caffeine
We roughly estimated the contribution of the P450 isoforms
studied to caffeine oxidation pathways on the basis of the
rate of those reactions in Supersomes and the contribution
of each isoform to the total P450 content in the liver
(Table 3). The calculations done at the therapeutic con-
centration of caffeine (100 mM) indicated that CYP1A2 was
the main isoform responsible for caffeine metabolism.
According to the calculated intrinsic clearance values, the
preference of P450 enzymes for the catalysis of caffeine
metabolism was as follows (pmol of product/pmol of P450 iso-
form/min):
CYP1A2 was
a chief enzyme catalyzing 1-N- and 3-N-
demethylation (75 and 85%, respectively) and substantially

b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
547
Table 2 – Kinetic parameters of caffeine metabolism in cDNA-expressed human CYPs (Supersomes)
P450
1-N-demethylation
(theobromine)
3-N-demethylation
(paraxanthine)
7-N-demethylation
(theophylline)
C-8-hydroxylation
(1,3,7-trimethyluric acid)
Km
Vmax
Vmax/Km
Km
Vmax
Vmax/Km
Km
Vmax
Vmax/Km
Km
Vmax
Vmax/Km
1A2
2A6
2B6
2C8
2C9
2C18
2C19
2D6
2E1
2.56
40.64
2.31
0.92
2.17
4.93
11.93
74.74
n.d.
0.341
0.068
0.019
0.014
0.021
0.042
0.154
1.242
n.d.
0.133
0.002
0.008
0.015
0.010
0.008
0.013
0.017
n.d.
2.96
1.14
0.64
0.20
0.13
0.60
1.42
0.755
1.37
0.09
2.07
3.34
1.128
0.034
0.042
0.080
0.108
0.043
0.048
0.066
0.052
0.156
0.016
1.87
0.60
4.13
3.56
1.03
0.377
0.042
0.194
0.172
0.087
13.180
0.327
3.75
0.201
0.070
0.047
0.048
0.084
0.039
0.054
0.152
0.080
0.031
0.060
0.39
1.72
0.55
3.37
3.88
0.97
2.37
1.93
0.90
0.91
2.15
0.187
0.172
0.195
0.319
0.318
0.170
0.294
0.441
0.299
0.157
0.340
0.479
0.100
0.354
0.095
0.082
0.175
0.124
0.228
0.332
0.172
0.158
0.039
0.027
0.016
0.014
0.026
0.069
0.050
0.072
0.014
0.033
337.3
6.05
24.70
1.77
0.141
0.098
0.079
3A4
3A5
16.04
0.88
0.107
0.011
0.007
0.012
3.14
1.32
n.d.—not detected. Km [mM], Vmax [pmol/pmol P450 isoform/min].
contributing to 7-N-demethylation (38.7%) and C-8-hydro-
xylation (28.7%). Moreover, 7-N-demethylation was also
visibly catalyzed by CYP2C8 (12.8%), CYP2C9 (12.2%) and
CYP3A4 (13.6%), while C-8-hydroxylation was substantially
mediated by CYP3A4 (30%) and, to a lesser extent by CYP2C8
(10%), CYP2C9 (8%) and CYP2E1 (11%). At a higher concen-
tration of the substrate (1 mM), the contribution of CYP1A2
to caffeine C-8-hydroxylation decreased (to 16%) mostly in
favour of CYP2C8 and CYP2C9 (to 17 and 11.6%, respectively)
(Table 3 and Fig. 4).
3.4.
Inhibition of caffeine metabolism by specific P450
inhibitors in rat liver microsomes
Furafylline (a CYP1A2 inhibitor) exerted a strong inhibitory
effect on the rate of the four oxidative caffeine reactions, being
most effective for 1-N- and 3-N-demethylation (a drop to 5 and
2%, respectively, of the control value), and the least active
towards C-8-hydroxylation (a decrease to 45% of the control
value) (Fig. 5). Sulfaphenazole (a CYP2C9 inhibitor) signifi-
cantly decreased the rate of 7-N-demethylation and C-8-
Fig. 3 – The biotransformation of caffeine via 1-N-demethylation (A), 3-N-demethylation (B), 7-N-demethylation (C), C-8-
hydroxylation (D) by human cDNA-expressed P450s (Supersomes). Caffeine (100 mM) was incubated with the Supersomes
(100–200 pmol P450/ml) and NADPH (1 mM) for 60 min. Each bar represents the mean value of two independent analyses.

548
b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
Table 3 – Estimation of the contribution of P450 isoforms to the particular metabolic pathways of caffeine on the basis of
the rates of these reaction in Supersomes and average P450 contents in the liver
P450
Relative contribution
of the isoform to
the total P450 contents
in liver microsomes
(Fraction)
Relative contribution of the isoforms to caffeine metabolism in liver
microsomes (percentage)#
Caffeine 1-N-
demethylation
Caffeine 3-N-
demethylation
Caffeine 7-N-
demethylation
Caffeine C-8-
hydroxylation
100 mM
1000 mM
100 mM
1000 mM
100 mM
1000 mM
100 mM
1000 mM
CYP1A2
CYP2A6
CYP2B6
CYP2C8
CYP2C9
CYP2C18
CYP2C19
CYP2D6
CYP2E1
CYP3A4
CYP3A5
0.127a
0.040a
0.002a
0.150b
0.140b
0.005c
0.061c
0.015a
0.066a
0.288a
0.002d
75.12
66.88
85.16
92.00
38.74
35.28
28.75
2.91
15.98
2.17
0.30
0.02
8.31
5.50
0.08
3.66
1.20
n.d.
5.73
0.07
0.55
0.06
5.14
5.85
0.13
3.31
2.01
7.46
8.55
0.06
0.67
0.04
3.69
3.57
0.10
2.62
0.49
1.47
2.18
0.02
0.32
0.03
1.18
1.01
0.06
1.14
0.30
1.70
2.25
0.01
2.76
0.17
1.56
0.19
0.35
0.22
12.80
12.21
0.33
13.46
11.93
0.23
10.06
8.20
0.51
17.44
11.63
0.45
8.41
5.80
5.00
7.02
4.24
5.10
2.71
2.25
6.56
8.74
11.13
30.16
0.20
9.26
33.27
0.31
13.60
0.18
17.60
0.10
n.d.—not detected. Data according to aShimada et al. [22], bEdwards et al. [23], cJung et al. [24] and dRodrigues [25]. #Relative contribution of
P450s to the particular metabolic pathways of caffeine was calculated as percentage of the sum of predicted velocities in liver microsomes. The
predicted velocity in liver microsomes was calculated by multiplying the velocity in Supersomes (see Fig. 3) by the relative contribution of
isoform to the total P450 content in liver microsomes. For details of calculations, see Wo´jcikowski et al. [26]. Essential differences between 100
and 1000 mM concentration of caffeine are shown in bold.
hydroxylation (to 75 and 80%, respectively, of the control
value), not affecting other reactions. Ketoconazole (a CYP3A4
inhibitor) reduced the rate of 1-N- and 7-N-demethylation and
C-8-hydroxylation (to 85, 70 and 45%, respectively, of the
control value).
4.
Discussion
The data presented above constitute a complete report on the
qualitative and quantitative contribution of all the individual
drug-metabolizing P450 isoforms to the four oxidation path-
ways of caffeine metabolism in humans. The obtained results
Fig. 5 – The effect of P450-specific inhibitors on the rate of
caffeine 1-N-demethylation, 3-N-demethylation, 7-N-
demethylation and C-8-hydroxylation in pooled human
liver microsomes. The microsomes were incubated with
100 mM caffeine in the absence (control) or presence of
P450-specific inhibitors: 10 mM furafylline (FUR), 10 mM
sulfaphenazole (SULF), 2 mM ketoconazole (KET). The
absolute control values were 0.0023 W 0.00001 nmol of
theobromine/mg of protein/min, 0.0240 W 0.001 nmol of
paraxanthine/mg of protein/min, 0.0048 W 0.0002 nmol of
theophylline/mg of protein/min, 0.0026 W 0.0003 nmol of
1,3,7-trimethyluric acid/mg of protein/min. The mean
values W S.E.M. (n = 6) are presented. Statistical
Fig. 4 – The concentration-dependent contribution of P450
isoforms to caffeine 8-hydroxylation in human liver
(based on the rate of caffeine metabolism in the
Supersomes and the mean of P450s in the liver). For
further explanations see Fig. 3.
significance was assessed using Student’s test, and was
indicated with ***, p = 0.001; **, p = 0.01; *, p = 0.05. For
further explanations, see Fig. 2.

b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
549
confirm the principal role of the CYP1A2 isoform in the
metabolism of caffeine. They show that the main oxidation
pathway of caffeine in humans – 3-N-demethylation (70%) – is
specifically catalyzed by CYP1A2 (85%) at a concentration of
100 mM, which corresponds to the maximum therapeutic
concentration in humans [27,28] and supports the earlier
results of Berthou et al. [11,12] and Tassaneeyakul et al. [17].
Moreover, our study presents a few new findings indicating
that also 1-N-demethylation is governed to a great extent by
CYP1A2 (75%), while 7-N-demethylation and C-8-hydroxyla-
tion are catalyzed less specifically–by CYP1A2 and other CYP
isoforms, mainly CYP3A4 and CYP2C8/9. The above conclusion
is based on our consistent results of the Eadie–Hofstee
analysis, demonstration of the ability of cDNA-expressed
P450s to metabolize caffeine and inhibition studies.
CYP2C19. In general, the values of the above-discussed
contribution of individual P450 isoforms to the total P450
content [22–24] and to caffeine metabolism (Table 3) are
similar to those calculated on the basis of the meta-analysis
of data on P450s abundance in Caucasians [29], the latter
values concerning CYP2C8 being lower and those referring to
CYP2C9 and CYP2E1 higher, though. Importantly, however, in
the case of the meta-analysis [29] based on a large number
data from various sources, it should be kept in mind that P450
contribution values for individual isoforms were obtained
from different groups of subjects (as regards their number
and the examined individuals themselves), which accounts
for the qualitative difference in estimating particular
enzymes.
The above results obtained using human cDNA-expressed
P450s are in line with the data on inhibition of the four
metabolic pathways of caffeine by the CYP1A2 inhibitor
furafylline, which had the strongest effect on 1-N- and 3-N-
demethylation, and the weakest on C-8-hydroxylation (Fig. 5).
Furthermore, the above findings agree with the inhibition of
7-N-demethylation and C-8-hydroxylation by sulfaphenazole
(a CYP2C9 inhibitor) and ketoconazole (a CYP3A4 inhibitor) in
human liver microsomes.
Kinetic parameters showing caffeine metabolism in
human cDNA-expressed P450s indicate distinct inter-isoform
differences, which is consistent with the Eadie–Hofstee plots
derived from liver microsomes, the latter being indicative of
the one-enzyme catalysis of 3-N-demethylation (and the fairly
similar catalysis of 1-N-demethylation) on one hand, and the
visibly multiple-enzyme catalysis of 7-N-demethylation and
C-8-hydroxylation on the other.
The results obtained using cDNA-expressed P450s showed
that CYP1A2 displayed the highest intrinsic clearance towards
caffeine metabolism, its highest value being reached for 3-N-
demethylation (Table 2). The intrinsic clearance of CYP1A2
was also the highest among the P450s tested in the case
of other oxidation pathways, and it descended in the
following order: 3-N-demethylation > C-8-hydroxylation >
7-N-demethylation > 1-N-demethylation. We roughly esti-
mated the contribution of the P450 isoforms studied to
caffeine oxidation pathways on the basis of the rate of those
reactions in Supersomes and the contribution of each isoform
to the total P450 content in the liver (Table 3). The calculations
done at a caffeine concentration of 100 mM, which corre-
sponds to the maximum therapeutic concentration in
humans indicate that CYP1A2 is the main isoform responsible
for caffeine metabolism. CYP1A2 is the chief enzyme
catalyzing 1-N- and 3-N-demethylation (75 and 85%, respec-
tively) and substantially contributing to 7-N-demethylation
(38.7%) and C-8-hydroxylation (28.7%). Moreover, 7-N-
demethylation and C-8-hydroxylation are also considerably
mediated by CYP3A4 (13.6 and 30%, respectively) and CYP2C8/
9 (25 and 18%, respectively) and, to a lesser extent by CYP2E1
in the case of C-8-hydroxylation (11%). At a higher concen-
tration of the substrate, the contribution of CYP1A2 to
caffeine C-8-hydroxylation decreases, mostly in favour of
CYP2C8/9 (from 18 to 29%). It is noteworthy that the present
results, based on cDNA-expressed P450s and the average
values of P450s in the liver, are theoretical, since actual
results should depend on the inter-individual variability of
P450s and the diverse contribution of individual P450s to the
total content of P450 protein in the liver. We based our
calculations mainly on the mean values of P450s collected
from 60 donors (30 Caucasians and 30 Japanese), provided by
Shimada et al. [22]. Moreover, we included the results
obtained by Edwards et al. [23] who quantitatively determined
CYP2C8 and CYP2C9 in 30 individuals, and the data obtained
by Jung et al. [24] who estimated separately CYP2C18 and
Hence the results on the relative contribution of P450
isoforms to the metabolism of caffeine, obtained in the
present study, confirm the high CYP1A2-selective dependence
of caffeine 3-N-demethylation [11,17,18], but do not support
the formerly suggested involvement of CYP2E1 in 1-N- and 7-
N-demethylation [13,17,18], or the predominant role of
CYP3A4 in caffeine C-8-hydroxylation [17,18] in human liver.
The above-discussed differences between our present results
and those of other authors may stem from the different in vitro
model used (liver microsomes and selected P450 inhibitors;
selected cDNA-expressed P450 isoforms or CYP1A2 and
CYP2E1 cell lines) [14,15,17], the high caffeine concentrations
(above 0.1 mM) used in ‘in vitro experiments’ as reviewed by Ha
et al. [18], the use of unspecific P450 inhibitors [17], such as a-
naphtoflavone or diethyldithiocarbamate—DDC [30,31], and
the fact that the CYP2C isoforms were not taken into account.
The latter fact allowed us to calculate the relative contribution
of individual isoforms to the metabolism of caffeine in human
liver.
In conclusion, the obtained data show that (1) apart from
the 3-N-demethylation of caffeine–a CYP1A2 marker reaction
and the main oxidation pathway of caffeine in man–1-N-
demethylation is also specifically catalyzed by CYP1A2 (not
described previously); (2) 7-N-demethylation is non-specifi-
cally catalyzed, mainly by CYP1A2 and, to a lesser extent by
CYP2C8/9 and CYP3A4 (and not by CYP2E1, as suggested
previously); (3) C-8-hydroxylation preferentially involves
CYP1A2 and CYP3A4 and to a lesser degree CYP2C8/9 and
CYP2E1 (but not only CYP3A, as suggested previously) at a
concentration of 100 mM, which corresponds to the maximum
therapeutic concentration in humans. At a higher caffeine
concentration, the contribution of CYP1A2 to this reaction
decreases in favour of CYP2C8/9.
These data show for the first time the contribution of
CYP2C isoforms to the metabolism of caffeine in human liver
and suggest that, apart from 3-N-demethylation, 1-N-
demethylation may also be used for testing CYP1A2 activity.

550
b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
cDNA-expressed human CYP1A2 and CYP2E1.
Pharmacogenetics 1992;2:73–7.
Moreover, they indicate that the C-8-hydroxylation is not
exclusively catalyzed by CYP3A4.
[14] Eugster HP, Probst M, Wurgler FE, Sengstag C, Caffeine.
estradiol and progesterone interact with human CYP1A1
and CYP1A2. Evidence from cDNA-directed expression in
Saccharomyces cerevisiae. Drug Metab Dispos 1993;21:
43–9.
Acknowledgments
[15] Fuhr U, Doehmer J, Battula N, Wolfel C, Kudla C, Keita Y,
et al. Biotransformation of caffeine and theophylline in
mammalian cell lines genetically engineered for expression
of single cytochrome P450 isoforms. Biochem Pharmacol
1992;43:225–35.
This study was supported by Grant no. 4 PO5F 024 26 from the
State Committee for Scientific Research (KBN), Warszawa, and
was also part of the statutory activity of the Institute of
Pharmacology, the Polish Academy of Sciences in Krako´w.
[16] Tassaneeyakul W, Mohamed Z, Birkett DJ, McManus ME,
Veronese ME, Tukey RH, et al. Caffeine as a probe for
human cytochromes P450: validation using cDNA-
expression, immunoinhibition and microsomal kinetic
and inhibitor techniques. Pharmacogenetics 1992;2:
173–83.
[17] Tassaneeyakul W, Birkett DJ, McManus ME, Tassaneeyakul
W, Veronese ME, Andersson T, et al. Caffeine metabolism
by human hepatic cytochromes P450: contributions of
1A2, 2E1 and 3A isoforms. Biochem Pharmacol
1994;47:1767–76.
[18] Ha HR, Chen J, Krahenbuhl S, Follath F. Biotransformation
of caffeine by cDNA-expressed human cytochromes P-450.
Eur J Clin Pharmacol 1996;49:309–15.
[19] Kot M, Daniel WA. Effect of cytochrome P450 (CYP) inducers
on caffeine metabolism in the rat. Pharmacol Rep
2007;59:296–305.
[20] Rasmussen BB, Nielsen KK, Brøsen K. Determination of
theophylline and its metabolites in human urine and
plasma by high-performance liquid chromatography. J
Chromatogr B Biomed Appl 1996;676:169–74.
[21] Daniel WA, Kot M, Wo´jcikowski J. Effects of classic and
newer antidepressants on the oxidation pathways of
caffeine in rat liver. In vitro study. Pol J Pharmacol
2003;55:1045–53.
[22] Shimada T, Yamazaki H, Mimura M, Inui Y, Guengerich FP.
Interindividual variations in human liver cytochrome P-450
enzymes involved in the oxidation of drugs, carcinogens
and toxic chemicals: studies with liver microsomes of 30
Japanese and 30 Caucasians. J Pharmacol Exp Ther
1994;270:414–23.
[23] Edwards RJ, Adams DA, Watts PS, Davies DS, Boobis AR.
Development of a comprehensive panel of antibodies
against the major Xenobiotic metabolizing forms of
cytochrome P450 in humans. Biochem Pharmacol
1998;56:377–87.
[24] Jung F, Richardson TH, Raucy JL, Johnson EF. Diazepam
metabolism by cDNA-expressed human 2C P450s:
identification of P4502C18 and P4502C19 as low K(M)
diazepam N-demethylases. Drug Metab Dispos
1997;25:133–9.
r e f e r e n c e s
[1] Donovan JL, De Vane CL. A primer on caffeine
pharmacology and its drug interactions in clinical
psychopharmacology. Psychopharmacol Bull 2001;35:
30–48.
[2] Fisone G, Borgvist A, Usiello A. Caffeine as a psychomotor
stimulant: mechanism of action. Cell Mol Life Sci
2004;61:857–72.
[3] Chen JF, Xu K, Petzer JP, Staal R, Xu YH, Beilstein M, et al.
Neuroprotection by caffeine and A(2A) adenosine receptor
inactivation in a model of Parkinson’s disease. J Neurosci
2001;21:1–6.
[4] Xu K, Xu YH, Chen JF, Schwarzschild MA. Caffeine’s
neuroprotection against 1-methyl-4-phenyl-1,2,3,6-
tetrahydropyridine toxicity shows no tolerance to chronic
caffeine administration in mice. Neurosci Lett 2002;322:
13–6.
[5] Xu K, Xu Y, Brown-Jermyn D, Chen JF, Ascherio A, Dluzen
DE, et al. Estrogen prevents neuroprotection by caffeine in
the mouse 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine
model of Parkinson’s disease. J Neurosci 2006;26:535–41.
[6] Rostami-Hodjegan A, Nurminen S, Jackson PR, Tucker GT.
Caffeine urinary metabolite ratios as markers of enzyme
activity: a theoretical assessment. Pharmacogenetics
1996;6:121–49.
[7] Petersen MS, Halling J, Damkier P, Nielsen F, Grandjean P,
Weihe P, et al. Caffeine 3-N-demethylation (CYP1A2) in a
population with an increased exposure to polychlorinated
biphenyls. Eur J Clin Pharmacol 2006;62:1041–8.
[8] Kot M, Daniel WA. Relative contribution of rat cytochrome
P450 isoforms to the metabolism of caffeine: The pathway
and concentration dependence. Biochem Pharmacol
2008;75:1538–49.
[9] Grant DM, Campbell ME, Tang BK, Kalow W.
Biotransformation of caffeine by microsomes from human
liver. Kinetics and inhibition studies. Biochem Pharmacol
1987;36:1251–60.
[10] Butler MA, Iwasaki M, Guengerich FP, Kadlubar FF. Human
cytochrome P-450PA (P-450IA2), the phenacetin O-
deethylase, is primarily responsible for the hepatic 3-
demethylation of caffeine and N-oxidation of carcinogenic
arylamines. Proc Natl Acad Sci USA 1989;86:7696–700.
[11] Berthou F, Flinois JP, Ratanasavanh D, Beaune P, Riche C,
Guillouzo A. Evidence for the involvement of several
cytochromes P-450 in the first steps of caffeine metabolism
by human liver microsomes. Drug Metab Dispos
1991;19:561–7.
[12] Berthou F, Guillois B, Riche C, Dreano Y, Jacqz-Aigrain E,
Beaune P. Interspecies variations in caffeine metabolism
related to cytochrome P4501A enzymes. Xenobiotica
1992;22:671–80.
[13] Gu L, Gonzalez FJ, Kalow W, Tang BK. Biotransformation of
caffeine, paraxanthine, theobromine and theophylline by
[25] Rodrigues AD. Integrated cytochrome P450 reaction
phenotyping: attempting to bridge the gap between
cDNA-expressed cytochromes P450 and native human
liver microsomes. Biochem Pharmacol 1999;57:
465–80.
[26] Wo´jcikowski J, Pichard-Garcia L, Maurel P, Daniel WA.
Contribution of human cytochrome P-450 isoforms to the
metabolism of the simplest phenothiazine neuroleptic
promazine. Br J Pharmacol 2003;138:1465–74.
[27] Kaplan GB, Greenblatt DJ, Ehrenberg BL, Goddard JE,
Cotreau MM, Harmatz JS, et al. Dose-dependent
pharmacokinetics and psychomotor effects of caffeine in
humans. J Clin Pharmacol 1997;37:693–703.
[28] Carrillo JA, Benitez J. Clinically significant pharmacokinetic
interactions between dietary caffeine and medications.
Clin Pharmacokinet 2000;39:127–53.

b i o c h e m i c a l p h a r m a c o l o g y 7 6 ( 2 0 0 8 ) 5 4 3 – 5 5 1
551
[29] Rostami-Hodjegan A, Tucker GT. Simulation and prediction
inhibition of human cytochromes P450. Arch Biochem
of in vivo drug metabolism in human populations from in
vitro data. Nat Rev Drug Discov 2007;6:140–8.
[30] Chang TK, Gonzalez FJ, Waxman DJ. Evaluation of
triacetyloleandomycin, alpha-naphthoflavone and
diethyldithiocarbamate as selective chemical probes for
Biophys 1994;311:437–42.
[31] Eagling VA, Tjia JF, Back DJ. Differential selectivity of
cytochrome P450 inhibitors against probe substrates in
human and rat liver microsomes. Br J Clin Pharmacol
1998;45:107–14.