Identification of the human cytochromes P450 involved in the oxidative metabolism of 'Ecstasy'-related designer drugs

Article abstract of DOI:10.1016/S0006-2952(00)00284-7

The human cytochrome P450 (CYP) isozymes catalyzing the oxidative metabolism of the widely abused amphetamine derivatives MDMA (N-methyl-3,4-methylenedioxyamphetamine, 'Ecstasy'), MDE (N-ethyl-3,4-methylenedioxyamphetamine, 'Eve'), and MDA (3,4-methylenedioxyamphetamine) were identified. Using a simplified non-extractive reversed-phase HPLC assay with fluorescence detection, biphasic Michaelis-Menten kinetics were obtained for formation of all three dihydroxyamphetamines in liver microsomes from a CYP2D6 extensive metabolizer subject. In contrast, no low K(m) component was detectable in microsomes from a poor metabolizer subject. Additional specific probes for CYP2D6 further confirmed this isozyme as the exclusive low K(m) component for demethylenation. P450-selective inhibitors applied to CYP2D6-inhibited microsomes and activity measurements in a series of recombinant P450s suggested CYP1A2 as the major high K(m) component with contributions by CYP2B6 and CYP3A4. Moreover, the relative CYP1A2 content of a panel of 12 human livers was weakly but significantly correlated to the high K(m) demethylenase activity (Spearman rank correlation coefficient [r(s)] = 0.58; P < 0.05). Microsomal maximal velocities for N-dealkylation were at least 7-fold lower than for demethylenation and were characterized by apparently monophasic kinetics. The most important isozyme for this reaction appeared to be CYP2B6, the microsomal content of which was found to be strongly correlated to N-deethylation of MDE (r(s) = 0.90; P < 0.001). We conclude that, in addition to CP2D6 as the sole high-affinity demethylenase, several other P450 isozymes have the capacity to contribute to microsomal oxidative metabolism of methylenedioxyamphetamines. This may be of particular importance in individuals genetically lacking functional CYP2D6. Copyright (C) 2000 Elsevier Science Inc.

Full text of DOI:10.1016/S0006-2952(00)00284-7

Biochemical Pharmacology, Vol. 59, pp. 1563–1571, 2000.
© 2000 Elsevier Science Inc. All rights reserved.
ISSN 0006-2952/00/$–see front matter
PII S0006-2952(00)00284-7
Identification of the Human Cytochromes P450
Involved in the Oxidative Metabolism of “Ecstasy”-
Related Designer Drugs
Klaus-Peter Kreth,* Karl-Artur Kovar,* Matthias Schwab† and Ulrich M. Zanger†‡
¨
¨
*PHARMACEUTICAL INSTITUTE, UNIVERSITY OF TUBINGEN, 72076 TUBINGEN; AND †DR. MARGARETE FISCHER-
BOSCH-INSTITUTE OF CLINICAL PHARMACOLOGY, 70376 STUTTGART, GERMANY
ABSTRACT. The human cytochrome P450 (CYP) isozymes catalyzing the oxidative metabolism of the widely
abused amphetamine derivatives MDMA (N-methyl-3,4-methylenedioxyamphetamine, “Ecstasy”), MDE (N-
ethyl-3,4-methylenedioxyamphetamine, “Eve”), and MDA (3,4-methylenedioxyamphetamine) were identified.
Using a simplified non-extractive reversed-phase HPLC assay with fluorescence detection, biphasic Michaelis–
Menten kinetics were obtained for formation of all three dihydroxyamphetamines in liver microsomes from a
CYP2D6 extensive metabolizer subject. In contrast, no low Km component was detectable in microsomes from
a poor metabolizer subject. Additional specific probes for CYP2D6 further confirmed this isozyme as the
exclusive low Km component for demethylenation. P450-selective inhibitors applied to CYP2D6-inhibited
microsomes and activity measurements in a series of recombinant P450s suggested CYP1A2 as the major high
Km component with contributions by CYP2B6 and CYP3A4. Moreover, the relative CYP1A2 content of a panel
of 12 human livers was weakly but significantly correlated to the high Km demethylenase activity (Spearman
rank correlation coefficient [rs] ϭ 0.58; P Ͻ 0.05). Microsomal maximal velocities for N-dealkylation were at
least 7-fold lower than for demethylenation and were characterized by apparently monophasic kinetics. The most
important isozyme for this reaction appeared to be CYP2B6, the microsomal content of which was found to be
strongly correlated to N-deethylation of MDE (rs ϭ 0.90; P Ͻ 0.001). We conclude that, in addition to CP2D6
as the sole high-affinity demethylenase, several other P450 isozymes have the capacity to contribute to
microsomal oxidative metabolism of methylenedioxyamphetamines. This may be of particular importance in
individuals genetically lacking functional CYP2D6. BIOCHEM PHARMACOL 59;12:1563–1571, 2000. © 2000
Elsevier Science Inc.
KEY WORDS. CYP2D6; cytochrome P450; designer drugs; ecstasy; MDE; MDMA
MDMA§ (street name “Ecstasy”, “Adam”), MDE (“Eve”),
and MDA (“love pills”) are methylenedioxyamphetamine-
based designer drugs which are among the most commonly
abused illicit drugs in the Western world. Their specific
psychotropic effects, described as increased communica-
tiveness, empathy, and self-knowledge, distinguish them
from typical stimulants and hallucinogens and lead to their
designation as “entactogens”, a new entity of psychotropic
agents [1]. Pharmacologically, they affect both the dopa-
mine and serotonin systems primarily through indirect
monoaminergic mechanisms, although serotoninergic ef-
fects appear to be more prominent [2]. For example,
continued use of MDMA leads to chronic serotonin deple-
tion and ultimately to irreversible damage of serotoninergic
neurons in laboratory animals [3–5] as well as in human
long-term users, as recently demonstrated by positron emis-
sion tomography studies [6] and further in vivo investiga-
tions [7, 8]. Although abusers commonly believe in the
innocuousness of these drugs, acute somatic adverse effects
such as hyperthermia, cardiovascular complications, renal
and hepatic failure, occasionally even leading to death [9,
10], as well as unwanted neuropsychiatric reactions [11, 12]
are increasingly being reported. As unexpected effects of
drugs are often related to their metabolism, the question
arises as to what impact individual drug metabolism capac-
ity may have on the adverse and neurotoxic effects of these
drugs.
The principal metabolic pathways of methylenedioxyam-
phetamine drugs (Fig. 1) have been elucidated and over a
dozen metabolites identified in animals and in humans.
The major oxidative degradation pathway of methyl-
enedioxyamphetamines involves demethylenation, leading
to reactive catechol metabolites which are further con-
verted by methylation, glucuronidation, and sulfation be-
‡ Corresponding author: Dr. Ulrich M. Zanger, Dr. Margarete Fischer-
Bosch-Institute of Clinical Pharmacology, Auerbachstr. 112, D-70376
Stuttgart, Germany. Tel. ϩ49-(0)711-81 01 37 04; FAX ϩ49-(0)711-85
92 95; E-mail: uli.zanger@ikp-stuttgart.de
§ Abbreviations: anti-LKM1, anti-liver/kidney microsome antibody type
1; CYP, cytochrome P450; EM, CYP2D6 extensive metabolizer; FMO-3,
flavin-containing monooxygenase 3; MDA, 3,4-methylenedioxyamphet-
amine; MDE, N-ethyl-3,4-methylenedioxyamphetamine; MDMA,
N-methyl-3,4-methylenedioxyamphetamine; PM, CYP2D6 poor metabo-
lizer; and rs, Spearman rank correlation coefficient.
Received 28 June 1999; accepted 30 December 1999.

1564
K-P. Kreth et al.
FIG. 1. Major pathways of oxidative microsomal
metabolism of N-substituted methylenedioxy-
amphetamines.
fore they are excreted [13–15]. A parallel side-chain deg-
radation pathway is initiated by N-dealkylation, first lead-
ing to the corresponding primary amines, which are then
oxidatively deaminated to the final benzoic acid derivatives
[15]. A third route via aromatic hydroxylation has been
proposed to result in the production of neurotoxic trihy-
droxyamphetamines based on in vitro studies [16]. Kumagai
et al. [17] investigated the demethylenation of MDMA
kinetically in rat liver microsomes and provided evidence
for the involvement of multiple enzymes, including one or
more phenobarbital-inducible cytochrome P450 forms with
high Km and, in particular, the CYP2D subfamily as the low
Km component. Subsequently, it was shown that the
homologous human CYP2D6 demethylenates MDMA with
low Km in a recombinant yeast system and that this isozyme
contributes to human liver microsomal activity [18, 19].
However, the metabolic pathways have not been studied in
detail in human liver microsomes, and other enzymes
participating in the demethylenation and N-dealkylation
reactions remain unidentified.
The aim of this study was, therefore, to systematically
investigate the enzymology of the demethylenation and
N-dealkylation pathways of ecstasy-related drugs in human
liver microsomes. The three related drugs MDMA, MDE,
and MDA were studied comparatively. The major enzymes
contributing to the two pathways were identified, and the
important role CYP2D6 plays in the demethylenation of all
three drugs was further established. These data form a better
basis to assess risk factors of entactogen abuse, such as
genetic polymorphism and drug interactions.
cin were purchased from Sigma. All other reagents were of
the highest grade available from commercial sources.
Human Liver Microsomes and Recombinant Enzymes
Microsomes were prepared from a kidney donor liver bank
and phenotyped in vitro as CYP2D6 EM or PM using
cumene hydroperoxide-mediated bufuralol 1Ј-hydroxyla-
tion and Western blotting as described [21]. Recombinant
CYP2D6 was prepared in insect cells and reconstituted with
NADPH:cytochrome P450 reductase as described [22]. All
other CYPs were purchased from Gentest Corp. as “super-
somes” co-expressed with NADPH:cytochrome P450 re-
ductase and, where indicated, with cytochrome b5. Infor-
mation on protein concentrations and CYP450 contents
was supplied by the manufacturer.
Incubation Conditions
To define optimal conditions for incubation and HPLC
analysis, human liver microsomes (0.1–2.5 mg/mL) were
incubated with MDE (0.1–2000 ␮M) and NADPH in the
presence or absence of superoxide dismutase [14] for up to
1 hr. The final conditions, for which linearity of product
formation with respect to time and protein had been
shown, were as follows: 50 ␮g of microsomal protein was
incubated in 100 mM sodium phosphate buffer, pH 7.4,
containing MDE, MDMA, or MDA at the given concen-
tration, 10 units of superoxide dismutase, and 1 mM
NADPH in a final volume of 100 ␮L. Reactions were
started by addition of NADPH and terminated by adding
10 ␮L of 70% perchloric acid. Incubation times were 15
min for N-dealkylation and 2 min for demethylenation,
respectively, at 37°. Incubations with recombinant enzymes
were performed under identical conditions at indicated
substrate concentrations with 5 pmol of cytochrome P450
or 20 ␮g FMO-3. Reaction mixtures containing diethyldi-
thiocarbamate, furafylline, or orphenadrine as chemical
inhibitor were preincubated in the presence of NADPH for
5 min at 37°, and the reactions were started by the addition
MATERIALS AND METHODS
Chemicals
MDE, MDMA, MDA, and the metabolites N-ethyl-3,4-
dihydroxyamphetamine and 3,4-dihydroxyamphetamine
were synthesized as racemic mixtures as previously de-
scribed [15, 20]. Furafylline was a kind gift from U. Fuhr
(University of Cologne, Germany). Sulfaphenazole was
obtained from Ciba-Geigy and ketoconazole from Janssen.
Diethyldithiocarbamate, orphenadrine, and troleandomy-

Microsomal Metabolism of Ecstasy-Related Drugs
1565
of substrate. The other inhibitor substances were not
preincubated. Bufuralol 1Ј-hydroxylation was determined
as described [21].
Life Science) was used for detection. Band intensities on
films were densitometrically analyzed with an Elscript 400
optical scanner (Hirschmann).
Analysis of Metabolites
Mathematical Methods
An HPLC method with fluorescence detection recently
developed in our laboratory to quantify MDE and its
metabolites in human plasma [23] was modified to analyze
metabolites in in vitro incubations. To avoid an extraction
step, protein was precipitated by the addition of perchloric
acid. It was shown that all metabolites quantified in this
study were stable under these conditions for at least 72 hr.
The incubation mixtures were then centrifuged at 9300 g
for 5 min, and the HClO4 supernatants were directly
injected into the HPLC system consisting of a LiChroCart
Superspher 60 RP-select B column, 5 ␮m, 250 ϫ 4 mm
connected to a LaChrom fluorescence detector L-7480 (all
HPLC equipment obtained from Merck). The mobile phase
consisted of 20 mM potassium phosphate buffer, pH 3.0,
and acetonitrile at varying concentrations (8% v/v for
3,4-dihydroxyamphetamine, 10% v/v for N-methyl-3,4-
dihydroxyamphetamine, and 12% v/v for N-ethyl-3,4-
dihydroxyamphetamine and MDA separation). The wave-
lengths for fluorescence detection were set at 286 nm
(excitation) and 322 nm (emission) for detection of all
metabolites. Under these conditions, retention times were
6.0 min for the dihydroxy metabolites and 17.5 min for
MDA. Quantification was performed using the authentic
substances as external standards except for N-methyl-3,4-
dihydroxyamphetamine, which was quantified using N-eth-
yl-3,4-dihydroxyamphetamine as standard.
Kinetic analyses were performed by non-linear least square
regression analysis using GraphPad Prism 2.0 (GraphPad
Software Inc.). In view of the unknown distribution of the
data, the non-parametric Spearman rank correlation coef-
ficient rs was calculated to compare the data obtained with
the human liver samples. Multiple regression analysis was
performed using Instat (GraphPad Software Inc.).
RESULTS
Kinetic Studies in Human Liver Microsomes
Since earlier studies had not indicated pronounced ste-
reospecific metabolism of MDMA, we used racemic mix-
tures throughout this study [18]. Detailed kinetic analyses
were performed with liver microsomes previously charac-
terized in vitro to be derived from an EM and a PM subject.
This had been shown by the presence (EM) or absence
(PM) of CYP2D6 activity, determined as cumene hydroper-
oxide-mediated bufuralol 1Ј-hydroxylation, and the pres-
ence or absence of CYP2D6 protein, as recognized by the
specific monoclonal antibody 114 [21]. The demethylena-
tion of all three amphetamines followed clearly biphasic
Michaelis–Menten kinetics in EM liver microsomes, as
shown in Eadie–Hofstee plots in Fig. 2. Over a substrate
concentration range of 0.4–1200 ␮M, the kinetic data
could be best modeled assuming a high- and a low-affinity
demethylenase with Km values in the low and upper
micromolar range, respectively (Table 1). The high-affinity
Km values for the N-substituted MDMA and MDE were 4-
to 5-fold lower than that for MDA. Vmax of the low-affinity
demethylenase was 3- to 5-fold higher than that of the
high-affinity component of the same liver (Table 1). In
contrast, PM microsomes completely lacked the high-
affinity/low-capacity component, and the fitted kinetic
parameters were in the same range as the low-affinity/high-
capacity component of EM microsomes (Fig. 2A, Table 1).
The N-dealkylation of MDE and MDMA, resulting in
production of MDA, was apparently monophasic in micro-
somes of EM and PM subjects, and the data were best fitted
to a one-site binding model with Km values in the upper
micromolar range (Fig. 2B and Table 1). However, in one
case (N-dealkylation of MDMA), a two-site binding model
with Km1 ϭ 82 ␮M and Km2 ϭ 3.06 mM gave a slightly
better fit, indicating that more than one enzyme may
participate in this reaction. The Vmax values for N-dealky-
lation reactions were between 7- and over 40-fold lower
than those for the corresponding demethylenations (Table
1).
Antibodies and Western Blotting
Monoclonal antibody 114 was used for Western blot detec-
tion of CYP2D6, and previously characterized strongly
inhibitory anti-LKM1 were used for CYP2D6 immunoinhi-
bition studies [24]. At the concentration used in the
experiments (1:200 of an ammonium sulfate fraction pre-
pared from human serum), this preparation inhibited
greater than 95% of MDE (5 ␮M) demethylenation.
Antibodies to CYP1A2, CYP2C8, CYP2E1, and CYP3A4
were kindly provided by Dr. Philippe Beaune, Paris [25].
Under the conditions used, anti-CYP2C8 recognized both
CYP2C8 and CYP2C9, while anti-CYP3A4 recognized
both CYP3A4 and CYP3A5 in a single band. The anti-
CYP2B6 antibody “WB-2B6” was purchased from Gentest
Corp. Recombinant enzymes were co-analyzed on the same
gels to identify the correct bands, because some of the
antibodies recognized several bands. Secondary antibodies
were obtained as standard reagents from common commer-
cial sources. Western blotting was performed according to
standard procedures using SDS–PAGE (10% polyacryl-
amide gels) with 25 ␮g of microsomal protein per lane. The
ECL (enhanced chemiluminescence) system (Amersham

1566
K-P. Kreth et al.
MDMA, and MDA, respectively. In contrast, furafylline (5
␮M), sulfaphenazole (5 ␮M), diethyldithiocarbamate (20
␮M), ketoconazole (10 ␮M), and troleandomycin (300
␮M) had no or only minor effects on the demethylenation
at 10-␮M substrate concentration (data not shown). Anti-
LKM1 autoantibodies, which potently and specifically in-
hibit CYP2D6, also decreased the demethylenation by over
95% with 5 ␮M MDE as substrate, whereas no significant
inhibition was observed for the demethylenation in PM
microsomes or for the N-dealkylation reaction (data not
shown).
In microsomes of 12 human livers, formation of dihy-
droxy and N-dealkylated metabolites was quantified at low
(5 ␮M) and high (200 ␮M) substrate concentrations, and
activities were correlated to each other and to the relative
content of several P450 isozymes determined by Western
blotting (Table 2). At low substrate concentration, de-
methylenase activities were highly correlated to each other,
to the relative CYP2D6 content, as well as to the CYP2D6-
specific bufuralol 1Ј-hydroxylase activity (Table 2).
Investigation of Low-Affinity Demethylenase
To determine and compare CYP2D6-dependent and -inde-
pendent demethylenase activities present in liver micro-
somes, N-ethyl-3,4-dihydroxyamphetamine formation from
MDE was individually measured in the 12 microsomal
samples in the absence and presence of anti-LKM1 at
200-␮M substrate concentration (Fig. 3). The contribution
by enzymes other than CYP2D6 varied between about 30%
and almost 100% of the total microsomal activity at higher
substrate concentration. To minimize effects related to
variability, a pool of liver microsomes from three EM
subjects was prepared and preincubated with anti-LKM1
autoantibodies to block CYP2D6 activity. Furafylline, a
specific mechanism-based inhibitor of CYP1A2, was the
most effective inhibitor of MDE demethylenation in this
pool, followed by orphenadrine, proposed to be a selective
inhibitor of CYP2B6 (Fig. 4). The CYP3A-selective inhib-
FIG. 2. Eadie–Hofstee plots for the formation of (A) N-ethyl-
3,4-dihydroxyamphetamine from MDE and (B) MDA from
MDE in liver microsomes from a CYP2D6 extensive and a
CYP2D6 poor metabolizer. MDE (0.4–1200 ␮M) was incu-
bated with microsomal protein. N-Ethyl-3,4-dihydroxyamphet-
amine (A) or MDA (B) was quantitated by HPLC and fluores-
cence detection. Data points represent the means of duplicate
incubations. The calculated kinetic parameters are summarized
in Table 1.
Investigation of High-Affinity Demethylenase
As expected, the high-affinity demethylenase observed in
EM microsomes was highly sensitive to the potent CYP2D6
inhibitor, quinidine, and Ki values of 0.46, 0.20, and 0.12
␮M were determined for the demethylenation of MDE,
TABLE 1. Summary of apparent kinetic parameters for the demethylenation and N-dealkylation reactions in human liver
microsomes
Demethylenation
N-dealkylation
CYP2D6
extensive
metabolizer
MDA
11.6 Ϯ 5.4
167 Ϯ 37
1277 Ϯ 287
825 Ϯ 403
CYP2D6 poor metabolizer
594 Ϯ 54
883 Ϯ 37
MDMA
2.2 Ϯ 1.6
223 Ϯ 54
377 Ϯ 177
745 Ϯ 85
MDE
MDMA
MDE
—
2.6 Ϯ 1.4
241 Ϯ 41
602 Ϯ 185
1261 Ϯ 130
K
m1 (␮M)
—
—
Vmax1 (pmol/mg ⅐ min)
K
Vmax2 (pmol/mg ⅐ min)
—
m2 (␮M)
1111 Ϯ 169
138 Ϯ 12
791 Ϯ 66
34 Ϯ 1.6
1460 Ϯ 293
2159 Ϯ 274
476 Ϯ 93
2045 Ϯ 195
Km (␮M)
Vmax (pmol/mg ⅐ min)
733 Ϯ 25
104 Ϯ 1.7
356 Ϯ 53
63 Ϯ 3.5
A two-enzyme model was applied to the demethylenation, which resulted in clearly biphasic Eadie–Hofstee plots (Fig. 2). In this case, Km1 and Vmax1 refer to Michaelis–Menten
parameters of the high-affinity component, whereas Km2 and Vmax2 refer to the low-affinity component. Values are means and standard errors (SE) obtained by non-linear least
square regression analysis of duplicate data points obtained at substrate concentrations between 0.4 and 1200 ␮M.

Microsomal Metabolism of Ecstasy-Related Drugs
1567
TABLE 2. Comparison of various enzyme activities and relative contents for individual CYPs determined by Western blotting in a
series of 12 human liver microsome samples
Parameter 1
Parameter 2
rs
Demethylenation
MDE (5 ␮M)
MDMA (5 ␮M)
MDA (5 ␮M)
MDE (5 ␮M)
MDE (5 ␮M)
MDMA (5 ␮M)
MDE (5 ␮M)
MDMA (5 ␮M)
MDA (5 ␮M)
CYP2D6
CYP2D6
CYP2D6
MDMA demethylenation (5 ␮M)
MDA demethylenation (5 ␮M)
MDA demethylenation (5 ␮M)
Bufuralol-1Јhydroxylation
0.89*
0.83*
0.94*
0.94*
0.97*
0.85*
0.98*
0.91*
0.95*
Bufuralol-1Јhydroxylation
Bufuralol-1Јhydroxylation
Demethylenation, CYP2D6 inhibited by anti-LKM1
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
N-dealkylation
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
MDE (200 ␮M)
CYP1A2
0.58†
0.31
0.10
0.01
0.44
0.46
0.10
CYP2B6
CYP2C8ϩ9
CYP2D6
CYP2E1
CYP3A4ϩ5
MDE demethylenation (5 ␮M)
CYP1A2
CYP2B6
CYP2C8ϩ9
CYP2D6
CYP2E1
CYP3A4ϩ5
MDMA N-dealkylation (200 ␮M)
0.48
0.90*
0.15
0.37
0.27
0.50
0.97*
For correlations, the Spearman rank correlation coefficient rs was calculated.
* P Ͻ 0.001.
† P Ͻ 0.05.
itors ketoconazole and troleandomycin exhibited less but
still significant inhibition, which was much more pro-
nounced with MDA and MDMA (51% and 43% inhibition
with troleandomycin, respectively).
The principal catalytic ability of human P450s to de-
methylenate methylenedioxyamphetamines was investi-
gated using a series of recombinant CYPs co-expressed in
insect cell membranes with CYP reductase and in some
cases with cytochrome b5. Of 13 different recombinant
enzyme preparations, CYP1A2, CYP3A4, and CYP1A1
showed the highest specific activities for MDE demethyl-
enation at 200-␮M substrate concentration, followed by
CYP2D6 (Fig. 5). Other tested P450s had measurable but
much lower activities. Interestingly, CYP3A4 was almost
inactive in the absence of co-expressed cytochrome b5.
NADH (1 mM) added to microsomal reaction mixtures
did, however, not increase activity for demethylenation.
FMO-3, the major human hepatic flavin-containing mono-
oxygenase form, had almost no measurable activity.
The MDE (200 ␮M) demethylenase activities deter-
mined in the panel of anti-LKM1-inhibited microsomes
were moderately but significantly (P Ͻ 0.05) correlated to
the relative levels of CYP1A2 (Table 2; Fig. 6A). The
correlations to the contents of CYP3A4, CYP2B6, and
CYP2E1 were still weaker and did not reach significance.
FIG. 3. MDE demethylenase activities of 12 human liver micro-
some samples. Enzymatic activities were determined at 200 ␮M
MDE in the absence (complete bars) and presence (lower bar
segment) of LKM1 autoantibodies to block greater than 95% of
the activity of CYP2D6. Liver sample 11 was shown by Western
blot analysis to be deficient in CYP2D6, while all other samples
were shown to contain immunodetectable CYP2D6. Bars rep-
resent the means of duplicate incubations.
Investigation of N-dealkylation
Several of the inhibitors were found to inhibit MDE
deethylation (Fig. 7). The most pronounced effect was
again achieved by furafylline, followed by orphenadrine,

1568
K-P. Kreth et al.
FIG. 4. Effect of CYP isoform-selective inhibitors on the low-
affinity component of MDE demethylenase. Pooled liver micro-
somes from three extensive metabolizers were preincubated with
anti-LKM1 autoantibodies to block CYP2D6. Activities were
then determined with 200 ␮M MDE at the indicated inhibitor
concentrations and expressed as percent of control activity
(100%, no inhibitor present). The results of two independent
incubations are represented by diamonds, and the means are
represented by bars.
FIG. 6. (A) Comparison of low-affinity MDE demethylenase
activity and relative content of CYP1A2 in 12 human liver
microsome samples. Microsomes were preincubated with anti-
LKM1 antibodies (dilution 1:200) to inhibit CYP2D6. Rates
for formation of N-ethyl-3,4-dihydroxyamphetamine (pmol/
mg ؋ min) were determined at 200 ␮M DME and compared to
relative levels of CYP1A2 determined by Western blotting. The
Spearman rank correlation coefficient rs was calculated (Table
2). Data points are means of duplicate incubations. (B) Com-
parison of N-dealkylase activity and relative content of CYP2B6
in 12 human liver microsome samples. Rates for MDA forma-
tion were determined at 200 ␮M MDE and compared to relative
levels of CYP2B6 determined by Western blotting. The Spear-
man rank correlation coefficient rs was calculated (Table 2).
Data points are means of duplicate incubations.
which inhibited between 25 and 63% at 100 and 500 ␮M,
respectively (Fig. 7). Ketoconazole (5 ␮M) and troleando-
mycin (100 ␮M) caused less but still significant inhibition.
In the recombinant system, CYP1A1, CYP1A2, and
CYP2B6 displayed equally high turnover numbers for N-
dealkylation of MDE (Fig. 5). All other enzymes, including
FMO-3, had much lower activities. Except for CYP2B6, the
turnover numbers for N-dealkylation were severalfold lower
than for demethylenation and on average reflected the
microsomal activity ratios (Table 1). Comparison of N-
dealkylation activities with relative P450 contents in the
same 12 liver samples used above revealed the best and the
only significant correlation to the CYP2B6 content (rs ϭ
0.90; P Ͻ 0.001). CYP2B6 varied more than 15-fold
among the 12 livers and was almost undetectable in two
livers (Fig. 6B). The weaker correlations to CYP3A4 (rs ϭ
0.50; P ϭ 0.05) and CYP1A2 (rs ϭ 0.48; P ϭ 0.06)
showed a trend towards significance, in agreement with
partial contributions by these enzymes. However, when
multiple regression analysis was performed, the only signif-
icant contribution to the total liver activity for N-deethy-
lation was made by CYP2B6 (P Ͻ 0.01). There was a
strong correlation between the microsomal N-dealkylation
activities for MDE and MDMA, indicating that both are
processed by the same enzymes in human liver (Table 2).
FIG. 5. Rates of demethylenation (filled bars) and N-deethyla-
tion (open bars) of MDE by recombinant human CYP isozymes.
Five picomoles of recombinant human P450 isozyme co-ex-
pressed with NADPH:P450 reductase and, where indicated,
with cytochrome b5, was incubated with MDE (200 ␮M).
N-Ethyl-3,4-dihydroxyamphetamine or MDA was quantitated
by HPLC. The results of two independent incubations are
represented by diamonds, and the means are represented by bars.

Microsomal Metabolism of Ecstasy-Related Drugs
1569
microsomal protein pool. Furthermore, CYP1A2 was the
only CYP to which the activities of the 12 microsome
samples were significantly correlated. The rather weak
correlation obtained (rs ϭ 0.58; P Ͻ 0.05) reflects the
contribution of additional isozymes.
In the recombinant system, CYP1A2 had the highest
turnover number, although similar turnover numbers were
also observed with recombinant CYP1A1, CYP2D6, and
CYP3A4 co-expressed with cytochrome b5. CYP1A1 is
expressed at very low levels in liver [27], and we were in fact
not able to clearly detect it in our samples. However, since
it is readily inducible, it may be expressed at higher levels
in some individuals. The fact that methylenedioxyamphet-
amines were shown to be efficient substrates for CYP1A1
and CYP1A2 in this study raises the question as to whether
these drugs themselves may be able to induce CYP1A1
and/or CYP1A2. This possibility, which may have impor-
tant toxicological implications, is further supported by the
fact that these drugs share the methylenedioxyphenyl
structure with safrole and isosafrole, which are preferential
inducers of CYP1A2 in mice and rats [28]. CYP3A4, on the
other hand, is the predominant liver P450 isozyme, and its
expression levels in human liver are approximately 2- to
5-fold higher than those of CYP1A2 [29]. Based on the
inhibition data with ketoconazole and troleandomycin, its
contribution to the total low-affinity MDE demethylenase
activity can be estimated to be 25% at most, i.e. less than
half of the CYP1A2-dependent activity. Based on recom-
binant turnover numbers and relative expression levels, one
would, however, expect roughly opposite proportions. This
discrepancy is most likely explained by the strong depen-
dence of P450 activity on the supporting enzymes CYP
reductase and cytochrome b5, one that has been observed
for many substrates of CYP3A4, complicating the establish-
ment of quantitative relationships between recombinant
turnover numbers and microsomal activity. A relative
overexpression of cytochrome b5 in the recombinant system
versus liver may lead to artificially high turnover numbers.
In previous investigations, the N-dealkylated metabolite
MDA was first described as an in vivo metabolite of MDMA
in the rat, where it appeared to be a major metabolite [13],
whereas it was clearly shown to be a minor metabolite of
MDE in humans [15, 23]. In animal studies, MDA was
shown to be equally effective to MDMA regarding effects
on the serotoninergic system [30] and interestingly, it was
described as the major metabolite in rat brain, in particular
in the male SD rat [31]. MDA formation may thus be
important for drug effect and toxicity, in particular if it
occurs in situ in the brain. Kinetic experiments suggested
that there is no high-affinity enzyme for N-dealkylation of
MDE and MDMA in human liver, and the low Vmax values
agree with the relatively minor in vivo metabolite formation
[23]. Inhibition data suggested that CYP1A2, CYP3A4,
and CYP2B6 may all participate in N-deethylation of MDE.
A strong correlation between activity and Western blot
detection of enzymes was, however, only seen for CYP2B6
content (rs ϭ 0.90; P Ͻ 0.0001; Fig. 6B and Table 2). Two
FIG. 7. Effect of CYP isoform-selective inhibitors on microso-
mal N-deethylation of MDE. Liver microsomes were incubated
with 200 ␮M MDE and the indicated inhibitors. The resulting
activity was expressed as percent of control activity (100%, no
inhibitor present). The results of two independent incubations
are represented by diamonds, and the means by bars.
DISCUSSION
Kinetic analysis of the demethylenation reaction revealed
clearly biphasic kinetics in liver microsomes derived from a
genetic CYP2D6 extensive metabolizer individual with
similar low Km/low-capacity and high Km/high-capacity
components for all three drug derivatives. The apparent Km
values of the two N-substituted drugs in microsomes were
almost identical to the Km values of (ϩ)-MDMA (1.72
␮M) and (Ϫ)-MDMA (2.90 ␮M) previously determined in
CYP2D6-expressing yeast microsomes [18]. Taken together,
these first kinetic data on the metabolism of methyl-
enedioxyamphetamines in human liver microsomes to-
gether with the additional experiments of this study suggest
that CYP2D6 is the only low Km component for the
demethylenation of all three methylenedioxyamphet-
amines.
At higher substrate concentration, other enzymes with
higher Km contribute substantially to the total microsomal
activity. This was shown by measuring MDE demethylenase
activity at 200-␮M substrate concentration in the micro-
some panel after complete inhibition of CYP2D6 with
anti-LKM1 antibodies. Between 30 and 100% of the total
microsomal activities were not affected by this treatment
and must therefore be attributed to other enzymes. In fact,
CYP2D6 appears to be relatively unimportant for the total
activity at this substrate concentration, as indicated by the
lack of correlation with the CYP2D6 content (Table 2).
Taken together, the experimental results suggest that
CYP1A2 constitutes the most important high Km enzyme.
Furafylline, a mechanism-based inhibitor with rather reli-
able selectivity for CYP1A2 [26], inhibited about 70% of
the CYP2D6-independent demethylenation of MDE in a

1570
K-P. Kreth et al.
livers had almost no detectable CYP2B6, and the CYP2B6
content of the other livers varied more than 15-fold. This
high variability agrees with previous reports [32]. In the
recombinant system, CYP1A1/1A2 had similar activities to
those of CYP2B6. The lack of a significant correlation to
CYP1A1/1A2 in the livers could mean that CYP2B6 makes
a larger contribution to the liver P450 content than
CYP1A1/1A2. However, this remains to be investigated.
Individuals who lack functional CYP2D6 due to a poor
metabolizer genotype or individuals under therapy with one
of the potent CYP2D6 inhibitors [33] lack the high-affinity
component for the major pathway of MDMA, MDE, and
MDA metabolism. It thus appears a reasonable and likely
possibility that the CYP2D6 PM genotype constitutes a
genetic risk factor predisposing users of methylenedioxyam-
phetamines to severe adverse drug effects. Conversely, as
the major neurotoxic metabolites appear to be generated
via catechol metabolites, PM individuals may be protected
from neurotoxic consequences. However, the fact that
additional enzymes are able to perform the same metabolic
steps makes predictions more difficult. For example, the
fraction of drug metabolized by low-affinity enzymes would
be increased if the drug accumulated in specific compart-
ments. Indeed, it was shown that MDMA concentrations in
rat brain exceed the plasma concentration almost 10-fold
[31]. In addition, because high-affinity CYP2D6 is easily
saturated, the other enzymes may become more important
in determining the plasma concentration of methyl-
enedioxyamphetamine metabolites.
There are few in vivo kinetic data for MDMA or related
drugs in humans available. In a recent study, plasma
samples collected over 4 hr from volunteers who had
ingested a single dose of 140 mg of MDE in the course of a
clinical study [34] were analyzed for MDE and the metab-
olites N-ethyl-4-hydroxy-3-methoxyamphetamine and
MDA [23]. We obtained blood samples from most of these
individuals and performed CYP2D6 genotyping [35]. One
individual, referred to as ‘P6’ (see Figs. 4, 6, and 8 of [23])
was shown to be homozygous for the nonfunctional
CYP2D6*4 allele, whereas five other volunteers had geno-
types compatible with the EM phenotype. Among these six
volunteers, individual P6 had the lowest area under the
curve for the demethylenated metabolite N-ethyl-4-hy-
droxy-3-methoxyamphetamine, namely 665 ng/mL*hr,
compared to an average of 1506 Ϯ 390 ng/mL*hr for the
other five participants. The maximal MDE plasma concen-
trations reached during the study did not differ between P6
(1.35 ␮M) and the other participants (1.24 ␮M Ϯ 0.24
␮M).
itself constitute a substantial risk factor for the adverse
effects of MDMA and MDE. In fact, it was recently shown
for the first time that three patients with fatal MDMA
intoxication were all extensive metabolizers [36]. Our in
vitro data suggest that in addition to CYP2D6 deficiency,
low levels or inhibition by drug interaction of the low-
affinity enzymes CYP1A2, CYP2B6, and CYP3A4 may be
responsible for adverse drug effects. Indeed, a fatal drug
interaction was reported between MDMA and ritonavir, a
strong inhibitor of CYP3A4 and CYP2D6 [37]. It appears
likely, therefore, that a number of different factors contrib-
ute to the risk of adverse and toxic effects of this group of
drugs.
We thank Philippe Beaune, Paris, for the generous gift of antibodies.
The excellent technical assistance of Britta Klumpp is also gratefully
acknowledged. This work was supported in part by the Robert Bosch
Foundation, Stuttgart, Germany. Preliminary results of this study were
presented at the XIIIth International Congress of Pharmacology,
IUPHAR, 26–31 July 1998 Munich, Germany, and published as an
abstract (Kreth KP, Zanger UM, Schwab M and Kovar KA, 1998,
Naunyn Schmiedebergs Arch Pharmacol 358: R 420).
References
1. Nichols DE, Differences between the mechanism of action of
MDMA, MBDB and the classic hallucinogens. Identification
of a new therapeutic class: entactogens. J Psychoactive Drugs
18: 305–313, 1986.
2. Steele TD, McCann UD and Ricaurte GA, 3,4-Methyl-
enedioxymethamphetamine (MDMA, “Ecstasy”): Pharmacol-
ogy and toxicology in animals and humans. Addiction 89:
539–551, 1994.
3. Ricaurte G, Bryan G, Strauss L, Seiden L and Schuster D,
Hallucinogenic amphetamine selectively destroys brain sero-
tonin nerve terminals. Science 229: 986–988, 1985.
4. Schmidt CJ, Neurotoxicity of the psychedelic amphetamine,
methylenedioxymethamphetamine. J Pharmacol Exp Ther
240: 1–7, 1987.
5. Scheffel U, Szabo Z, Mathews WB, Finley PA, Dannals RF,
Ravert HT, Szabo K, Yuan J and Ricaurte GA, In vivo
detection of short- and long-term MDMA neurotoxicity–a
positron emission tomography study in the living baboon
brain. Synapse 29: 183–192, 1998.
6. McCann UD, Szabo Z, Scheffel U, Dannals RF and Ricaurte
GA, Positron emission tomographic evidence of toxic effect
of MDMA (“Ecstasy”) on brain serotonin neurons in human
beings. Lancet 352: 1433–1437, 1998.
7. Bolla KI, McCann UD and Ricaurte GA, Memory impair-
ment in abstinent MDMA (“Ecstasy”) users. Neurology 51:
1532–1537, 1998.
8. Morgan MJ, Recreational use of “ecstasy” (MDMA) is asso-
ciated with elevated impulsivity. Neuropsychopharmacology
19: 252–264, 1998.
9. Henry JA, Jeffreys KJ and Dawling S, Toxicity and deaths
from 3,4-methylenedioxymethamphetamine. Lancet 340:
384–387, 1992.
These first in vivo data which relate plasma levels of MDE
metabolites to the CYP2D6 genotype thus demonstrate the
significance of the CYP2D6 polymorphism for the metab-
olism of MDE and likely also MDMA. However, they also
show that considerable metabolism via demethylenation
does occur in the absence of functional CYP2D6 in a PM
individual, demonstrating that other enzymes must be
active in vivo. Thus, the CYP2D6 PM genotype may not by
10. Milroy CM, Clark JC, and Forrest AR, Pathology of deaths
associated with “ecstasy” and “eve” misuse. J Clin Pathol 49:
149–153, 1996.
11. McCann UD and Ricaurte GA, Lasting neuropsychiatric
sequelae of (ϩ)-methylenedioxymethamphetamine (‘ecsta-
sy’) in recreational users. J Clin Psychopharmacol 11: 302–305,
1991.

Microsomal Metabolism of Ecstasy-Related Drugs
1571
12. McCann UD, Slate SO and Ricaurte GA, Adverse reactions
with 3,4-methylenedioxymethamphetamine (MDMA; ‘ecsta-
sy’). Drug Saf 15: 107–115, 1996.
13. Lim HK and Foltz RL, In vivo and in vitro metabolism of
3,4-(methylenedioxy)methamphetamine in rat: Identifica-
tion of metabolites using an ion trap detector. Chem Res
Toxicol 1: 370–378, 1988.
14. Hiramatsu M, Kumagai Y, Unger SE and Cho AK, Metabo-
lism of methylenedioxymethamphetamine: Formation of di-
hydroxymethamphetamine and a quinone identified as its
glutathione adduct. J Pharmacol Exp Ther 254: 521–527,
1990.
15. Ensslin HK, Maurer HH, Gouzoulis E, Hermle L and Kovar
KA, Metabolism of racemic 3,4-methylenedioxyethylamphet-
amine in humans. Isolation, identification, quantification,
and synthesis of urinary metabolites. Drug Metab Dispos 24:
813–820, 1996.
16. Lim HK and Foltz RL, Ion trap mass spectrometric evidence
for the metabolism of 3,4-(methylenedioxy)-methamphet-
amine to the potent neurotoxins 2,4,5-trihydroxymetham-
phetamine and 2,4,5-trihydroxyamphetamine. Chem Res
Toxicol 4: 626–632, 1991.
17. Kumagai Y, Lin LY, Hiratsuka A, Narimatsu S, Suzuki T,
Yamada H, Oguri K, Yoshimura H and Cho AK, Participation
of cytochrome P450-2B and -2D isozymes in the demethyl-
enation of methylenedioxymethamphetamine enantiomers by
rats. Mol Pharmacol 45: 359–365, 1994.
18. Tucker GT, Lennard MS, Ellis SW, Woods HF, Cho AK, Lin
LY, Hiratsuka A, Schmitz DA and Chu TY, The demethyl-
enation of methylenedioxyamphetamine (“Ecstasy”) by de-
brisoquine hydroxylase (CYP2D6). Biochem Pharmacol 47:
1151–1156, 1994.
19. Lin LY, Di Stefano EW, Schmitz DA, Hsu L, Ellis SW,
Lennard MS, Tucker GT and Cho AK, Oxidation of meth-
amphetamine and methylenedioxymethamphetamine by
CYP2D6. Drug Metab Dispos 25: 1059–1064, 1997.
20. Braun U, Shulgin AT and Braun G, Centrally active N-
substituted analogues of 3,4-methylenedioxyphenylisopro-
pylamine (3,4-methylenedioxyamphetamine). J Pharm Sci 69:
192–195, 1980.
21. Zanger UM, Vilbois F, Hardwick JP and Meyer UA, Absence
of hepatic cytochrome P450bufI causes genetically deficient
debrisoquine oxidation in man. Biochemistry 27: 5447–5454,
1988.
22. Evert B, Eichelbaum M, Haubruck H and Zanger UM,
Functional properties of CYP2D6 1 (wild-type) and CYP2D6
7 (His324Pro) expressed by recombinant baculovirus in insect
cells. Naunyn Schmiedebergs Arch Pharmacol 355: 309–318,
1997.
23. Brunnenberg M, Lindenblatt H, Gouzoulis-Mayfrank E and
Kovar KA, Quantitation of N-ethyl-3,4-methylenedioxyam-
phetamine and its major metabolites in human plasma by
high-performance liquid chromatography and fluorescence
detection. J Chromatogr B Biomed Sci Appl 719: 79–85, 1998.
24. Zanger UM, Hauri HP, Loeper J, Homberg JC and Meyer UA,
Antibodies against human cytochrome P-450db1 in autoim-
mune hepatits type II. Proc Natl Acad Sci USA 85: 8256–
8260, 1988.
25. Belloc C, Baird S, Cosme J, Lecoeur S, Gautier JC, Challine
D, de Waziers I, Flinois JP and Beaune PH, Human cyto-
chromes P450 expressed in Escherichia coli: Production of
specific antibodies. Toxicology 106: 207–219, 1996.
26. Newton DJ, Wang RW and Lu AY, Evaluation of specificities
in the in vitro metabolism of therapeutic agents by human
liver microsomes. Drug Metab Dispos 23: 154–158, 1995.
27. Drahushuk AT, McGarrigle BP, Larsen KE, Stegeman JJ and
Olson JR, Detection of CYP1A1 protein in human liver and
induction by TCDD in precision-cut liver slices incubated in
dynamic organ culture. Carcinogenesis 19: 1361–1368, 1998.
28. Lewandowski M, Chui YC, Levi PE and Hodgson E, Differ-
ences in induction of hepatic cytochrome P450 isozymes by
mice in eight methylenedioxyphenyl compounds. J Biochem
Toxicol 5: 47–55, 1990.
29. Shimada T, Yamazaki H, Mimura M, Inui Y and Guengerich
FP, Interindividual variations in human liver cytochrome
P-450 enzymes involved in the oxidation of drugs, carcino-
gens and toxic chemicals: Studies with liver microsomes of 30
Japanese and 30 Caucasians. J Pharmacol Exp Ther 270:
414–423, 1994.
30. Johnson M, Letter AA, Merchant K, Hanson GR and Gibb
JW, Effects of 3,4-methylenedioxyamphetamine (MDA) and
3,4-methylenedioxymethamphetamine (MDMA) isomers on
central serotonergic, dopaminergic and nigral neurotensin
systems of the rat. J Pharmacol Exp Ther 244: 977–982, 1988.
31. Chu T, Kumagai Y, DiStefano EW and Cho AK, Disposition
of methylenedioxymethamphetamine and three metabolites
in the brains of different rat strains and their possible roles in
acute serotonin depletion. Biochem Pharmacol 51: 789–796,
1996.
32. Code EL, Crespi CL, Penman BW, Gonzalez FJ, Chang TK
and Waxman DJ, Human cytochrome P4502B6: Interindi-
vidual hepatic expression, substrate specifity, and role in
procarcinogen activation. Drug Metab Dispos 25: 985–993,
1997.
33. Tucker GT, Adam and Eve make love. Analysis of the
enantiomers of 3,4-methylenedioxy-N-ethylamphetamine
(MDE, “Eve”) and its metabolite 3,4-methylenedioxyamphet-
amine (MDA) in rat brain. Hum Exp Toxicol 15: 455–457,
1996.
34. Hermle L, Spitzer M, Borchardt D, Kovar KA and Gouzoulis
E, Psychological effects of MDE in normal subjects. Are
entactogens a new class of psychoactive agents? Neuropsycho-
pharmacology 8: 171–176, 1993.
35. Schwab M, Griese U, Hermle L, Gouzoulis E, Zanger UM,
and Mikus G, Is there an impact of CYP2D6 genotype on the
toxicity of ‘ecstasy’ and related designer drugs? Naunyn
Schmiedebergs Arch Pharmacol 357 (Suppl): R163, 1998.
36. Schwab M, Seyringer E, Brauer RB, Hellinger A, and Griese
EU, Fatal MDMA intoxication. Lancet 353: 593–594, 1999.
37. Henry JA and Hill IR, Fatal interaction between ritonavir
and MDMA. Lancet 352: 1751–1752, 1998.