Cytochrome P450 1A2 detoxicates aristolochic acid in the mouse

Article abstract of DOI:10.1124/dmd.110.032201

Aristolochic acids (AAs) are plant-derived nephrotoxins and carcinogens responsible for chronic renal failure and associated urothelial cell cancers in several clinical syndromes known collectively as aristolochic acid nephropathy (AAN). Mice provide a useful model for study of AAN because the renal histopathology of AA-treated mice is strikingly similar to that of humans. AA is also a potent carcinogen in mice with a tissue spectrum somewhat different from that in humans. The toxic dose of AA in mice is higher than that in humans; this difference in susceptibility has been postulated to reflect differing rates of detoxication between the species. Recent studies in mice have shown that the hepatic cytochrome P450 system detoxicates AA, and inducers of the arylhydrocarbon response protect mice from the nephrotoxic effects of AA. The purpose of this study was to determine the role of specific cytochrome P450 (P450) enzymes in AA metabolism in vivo. Of 18 human P450 enzymes we surveyed only two, CYP1A1 and CYP1A2, which were effective in demethylating 8-methoxy-6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-5-carboxylic acid (AAI) to the nontoxic derivative 8-hydroxy-6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-5- carboxylic acid (AAIa). Kinetic analysis revealed similar efficiencies of formation of AAIa by human and rat CYP1A2. We also report here that CYP1A2-deficient mice display increased sensitivity to the nephrotoxic effects of AAI. Furthermore, Cyp1a2 knockout mice accumulate AAI-derived DNA adducts in the kidney at a higher rate than control mice. Differences in bioavailability or hepatic metabolism of AAI, expression of CYP1A2, or efficiency of a competing nitroreduction pathway in vivo may explain the apparent differences between human and rodent sensitivity to AAI. Copyright

Full text of DOI:10.1124/dmd.110.032201

0090-9556/10/3805-761–768$20.00
DRUG METABOLISM AND DISPOSITION
Copyright © 2010 by The American Society for Pharmacology and Experimental Therapeutics
DMD 38:761–768, 2010
Vol. 38, No. 5
32201/3582079
Printed in U.S.A.
Cytochrome P450 1A2 Detoxicates Aristolochic Acid in the Mouse
Thomas A. Rosenquist, Heidi J. Einolf, Kathleen G. Dickman, Lai Wang, Amanda Smith,
and Arthur P. Grollman
Laboratory of Chemical Biology, Department of Pharmacological Sciences, Stony Brook University, Stony Brook, New York
(T.A.R., K.G.D., A.P.G.); and Novartis Institutes of Biomedical Research, East Hanover, New Jersey (H.J.E., L.W., A.S.)
Received January 13, 2010; accepted February 17, 2010
ABSTRACT:
Aristolochic acids (AAs) are plant-derived nephrotoxins and car- specific cytochrome P450 (P450) enzymes in AA metabolism in
cinogens responsible for chronic renal failure and associated vivo. Of 18 human P450 enzymes we surveyed only two, CYP1A1
urothelial cell cancers in several clinical syndromes known collec- and CYP1A2, which were effective in demethylating 8-methoxy-6-
tively as aristolochic acid nephropathy (AAN). Mice provide a use- nitro-phenanthro-(3,4-d)-1,3-dioxolo-5-carboxylic acid (AAI) to the
ful model for study of AAN because the renal histopathology of nontoxic derivative 8-hydroxy-6-nitro-phenanthro-(3,4-d)-1,3-di-
AA-treated mice is strikingly similar to that of humans. AA is also a oxolo-5-carboxylic acid (AAIa). Kinetic analysis revealed similar
potent carcinogen in mice with a tissue spectrum somewhat dif- efficiencies of formation of AAIa by human and rat CYP1A2. We
ferent from that in humans. The toxic dose of AA in mice is higher
also report here that CYP1A2-deficient mice display increased
than that in humans; this difference in susceptibility has been sensitivity to the nephrotoxic effects of AAI. Furthermore, Cyp1a2
postulated to reflect differing rates of detoxication between the knockout mice accumulate AAI-derived DNA adducts in the kidney
species. Recent studies in mice have shown that the hepatic cy- at a higher rate than control mice. Differences in bioavailability or
tochrome P450 system detoxicates AA, and inducers of the aryl- hepatic metabolism of AAI, expression of CYP1A2, or efficiency of
hydrocarbon response protect mice from the nephrotoxic effects a competing nitroreduction pathway in vivo may explain the ap-
of AA. The purpose of this study was to determine the role of
parent differences between human and rodent sensitivity to AAI.
Aristolochic acids (AAs) are nitrophenanthrene carboxylic acids tolactam I (ALI) and aristolactam II (ALII). An intermediate in this
found in various Aristolochia species used as traditional herbal med-
icines throughout the world (Jameson et al., 2008). Chronic exposure
to AA is responsible for aristolochic acid nephropathy (AAN) and for
Balkan endemic nephropathy [for review, see Grollman et al. (2007)
and Debelle et al. (2008)]. Hallmarks of each disease are proximal
tubule atrophy and tubulointerstitial fibrosis, leading to end-stage
renal disease. At least half of these patients also develop cancers of the
upper urinary tract.
Commercial preparations of AA are composed of mixtures of
8-methoxy-6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-5-carboxylic acid
(AAI) and 6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-5-carboxylic acid
(AAII), compounds that differ only by the presence of an O-methoxy
group in AAI. AAI is far more cytotoxic than AAII both in simian
kidney cells in culture (Balachandran et al., 2005) and in treated
animals (Shibutani et al., 2007).
pathway, N-hydroxyaristolactam, is believed to form a cyclic N-
acylnitrenium ion that forms adducts with exocyclic amines of purines
in DNA. Indeed, the promutagenic 7-(deoxyadenosine-N6-yl)aristo-
lactams (ALI-dA and ALII-dA) have been observed in the DNA of all
species examined after treatment with AA. A:T3T:A transversions,
the most frequently observed mutation in the TP53 gene in urothelial
tumors of patients with AAN and Balkan endemic nephropathy (Lord
et al., 2004; Grollman et al., 2007), have been proposed to be a
“fingerprint” mutation for aristolochic acid exposure (Grollman et al.,
2007). Furthermore, aristolactam metabolites are observed in the urine
of various species treated with AA (Krumbiegel et al., 1987). Thus,
this pathway is postulated to be universal in mammals.
The second detoxication pathway involves demethylation of AAI to
form the nontoxic 8-hydroxy-6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-
5-carboxylic acid (AAIa). AAIa and its metabolites have been ob-
served in the urine and feces of rabbits, dogs, mice, and rats treated
with AAI, but not in urine from humans exposed to AA (Krumbiegel
et al., 1987).
In animals and cell culture, AA is metabolized by several mecha-
nisms. One pathway involves reduction of the nitro group and con-
comitant condensation with the carboxylic acid moiety to form aris-
A number of defined enzyme systems can activate AA to form
DNA adducts in vitro. These include the microsomal enzymes NADPH:
cytochrome P450 reductase (Stiborova´ et al., 2001c, 2005a), prosta-
glandin H synthase (Stiborova´ et al., 2001a, 2005a), and CYP1A1/2
under anaerobic conditions (Schmeiser et al., 1986; Stiborova´ et al.,
This work was supported by the National Institutes of Health National Institute
of Environmental Health Sciences [Grant ES04068].
Article, publication date, and citation information can be found at
http://dmd.aspetjournals.org.
doi:10.1124/dmd.110.032201.
ABBREVIATIONS: AA, aristolochic acid; AAN, aristolochic acid nephropathy; AAI, aristolochic acid I, 8-methoxy-6-nitro-phenanthro-(3,4-d)-1,3-
dioxolo-5-carboxylic acid; AAII, aristolochic acid II, 6-nitro-phenanthro-(3,4-d)-1,3-dioxolo-5-carboxylic acid; ALI, aristolactam I; ALII, aristolactam
II; ALI-dA, 7-(deoxyadenosine-N6-yl)aristolactam I; ALI-dG, 7-(deoxyguanosine-N2-yl)aristolactam I; P450, cytochrome P450; UDPGA, UDP-
glucuronic acid; HPLC, high-performance liquid chromatography; h, human; r, rat; 3-MC, 3-metylcholanthrene; TBST, Tris-buffered saline/Tween 20.
761

762
ROSENQUIST ET AL.
preincubated with alamethicin (60 ␮g of alamethicin/mg protein, final con-
centration) for 15 min on ice before addition of MgCl2 and [3H]AAI. All
reactions were preincubated at 37°C for 3 min before cofactor initiation (1 mM
NADPH and/or 4 mM UDPGA). Human liver microsomes (1 mg of protein/
ml) were incubated with [3H]AAI (95 ␮M), and reactions were initiated with
NADPH and/or UDPGA and were incubated for 30 min at 37°C (final reaction
volume of 0.2 ml). Control incubations contained all reaction components
without cofactors. Reactions were quenched by the addition of an equal
volume of cold acetonitrile, and the precipitated protein was removed by
centrifugation at 39,000g for 10 min at ϳ4°C.
2001b, 2005b). Cytoplasmic enzymes implicated in AA activation
include NAD(P)H:quinone oxidoreductase (Stiborova´ et al., 2002,
2003, 2005a) and sulfotransferase A1 (Meinl et al., 2006). Under
aerobic conditions, hepatic microsomes from rats and humans de-
methylate AAI to form AAIa (Schmeiser et al., 1986; Sistkova et al.,
2008).
More recently, mice deficient in hepatic cytochrome P450 activity
were shown to have increased sensitivity to the nephrotoxic effects of
AA (Xiao et al., 2008). Conversely, pretreatment of mice with
3-methylcholanthrene (Xue et al., 2008) or ␤-naphthoflavone (Xiao et
al., 2009), agonists of the arylhydrocarbon receptor that induce CYP1
enzymes and other xenometabolizing activities, protects mice from AA.
The purpose of this study was to delineate the role of specific P450
enzymes in vivo in AA detoxication. We report that CYP1A1 and
CYP1A2 are the most active of 18 human P450s tested in demethy-
lating AAI. Kinetic analyses revealed that rat and human CYP1A2
enzymes were similarly efficient in catalyzing the formation of AAIa.
However, species differences were found in the efficiency of CYP1A1
versus CYP1A2 to catalyze the demethylation reaction.
We show also that Cyp1a2-null mice are relatively more sensitive
to AAI-elicited nephrotoxicity. This increased sensitivity can be re-
versed by pretreatment with 3-methylcholanthrene. In addition,
CYP1A2-null mice accumulate ALI-DNA adducts at a higher rate
than control mice. Taken together, these results indicate that, in
rodents, AAI demethylation, mediated by CYP1A2, is a primary
pathway of AAI detoxication. If the demethylation pathway is com-
promised, DNA adducts continued to accumulate, indicating that AAI
nitroreduction is increased.
HPLC Analysis of [3H]AAI and Metabolites. Samples (25 ␮l) were
analyzed by reverse-phase HPLC on an XTerra MS C18 column (250 ϫ 4.6
mm, 5 ␮m; Waters, Milford, MA) at ambient temperature. Gradient elution
was achieved using solvent A (0.1 M ammonium acetate, pH 7.5, v/v) and
solvent B (acetonitrile) at a flow rate of 1.2 ml/min. The HPLC eluate was
collected with a fraction collector (FC204; Gilson, Inc., Middleton, WI) at 0.2
min/fraction into Deepwell LumaPlate-96 plates (PerkinElmer Life and Ana-
lytical Sciences, Waltham, MA). The fractions were dried with a stream of
nitrogen, and radioactivity was counted with a TopCount NXT Microplate
Scintillation and Luminescence Counter (PerkinElmer Life and Analytical
Sciences) at a counting time of 10 min/well. Chromatograms from the Top-
Count counter were evaluated using WinFLOW HPLC application software
(version 1.4a; IN/US Systems, Tampa, FL) and plotted using SigmaPlot
software (SigmaPlot 2002 for Windows, version 8.0; Jandel Corporation,
Chicago, IL).
Metabolism of [3H]AAI by Specific Human Recombinant P450 En-
zymes. [3H]AAI (95 mM) was incubated with the recombinant human P450
enzymes, CYP1A1, CYP1A2, CYP1B1, CYP2A6, CYP2B6, CYP2C8,
CYP2C9, CYP2C18, CYP2C19, CYP2D6, CYP2E1, CYP2J2, CYP3A4,
CYP3A5, CYP4A11, CYP4F2, CYP4F3A, and CYP4F3B (100 pmol of P450/
ml), or control P450 microsomes in the presence of NADPH for 30 min at
37°C (final reaction volume of 0.4 ml). The buffer components, sample
processing, and HPLC analysis were as described above.
Materials and Methods
Kinetic Analysis of [3H]AAI Metabolism by Recombinant Human or
Rat CYP1A1 and CYP1A2. Steady-state kinetic parameters associated with
recombinant human (h) CYP1A2 and rat (r) CYP1A2 were determined to
establish the efficiency of [3H]AAI metabolism by these enzymes. hCYP1A1
(10 pmol of P450/ml ⅐ mlϪ1, 0.11 mg of microsomal protein/ml), hCYP1A2
(10 pmol of P450/ml, 0.067 mg of microsomal protein/ml, final concentration),
rCYP1A1 (25 pmol of P450/ml, 0.0625 mg of microsomal protein/ml, final
concentration), or rCYP1A2 (10 pmol of P450/ml, 0.05 mg of microsomal
protein/ml, final concentration) was incubated with various concentrations of
[3H]AAI (in duplicate) in the presence of NADPH for 10 min (final reaction
volume of 0.2 ml). The concentration of P450 enzyme and reaction time were
predetermined to be optimal to ensure ϳϽ20% turnover of AAI during the
incubation. Control samples at each concentration of [3H]AAI did not contain
NADPH. The buffer components, sample processing, and HPLC analysis were
as described above. [3H]AAI metabolism activity was plotted against substrate
concentration and the kinetic parameters, Km and Vmax, were determined by
nonlinear regression.
Treatment of Mice with Aristolochic Acid. Eight-week-old male mice
were given intraperitoneal injections of 2 mg/kg AAI dissolved in phosphate-
buffered saline without divalent cations (Sigma-Aldrich). Several groups of
mice were pretreated the previous day with a single intraperitoneal injection of
3-methylcholanthrene (60 mg/kg) in corn oil. This dose of 3-MC has previ-
ously been shown to induce resistance to AA nephrotoxicity in mice (Xue et
al., 2008). Control animals received vehicle-only injections. For urine collec-
tions, mice were housed overnight in metabolic cages. Mice were euthanized
by CO2 asphyxiation, and tissues were collected for microsome and DNA
preparation.
Materials and Reagents. Reagents. AAI was purified from a mixture of
AAI and AAII (40:60) purchased from Thermo Fisher Scientific (Waltham,
MA) as described previously (Shibutani et al., 2007). 3-Methylcholanthrene
was purchased from Sigma-Aldrich (St. Louis, MO). Human and rat recom-
binant CYP1A1 and CYP1A2 enzymes and pooled human liver microsomes
were purchased from BD Biosciences (San Jose, CA). Micrococcal nuclease
and potato apyrase were purchased from Sigma-Aldrich, spleen phosphodies-
terase was from Worthington Biochemical Corp. (Lakewood, NJ), and 3Ј-
phosphatase-free T4 polynucleotide kinase and nuclease P1 were from Roche
Applied Science (Indianapolis, IN). [␥-32P]ATP (specific activity, Ͼ6000
Ci/mmol) was obtained from GE Healthcare (Little Chalfont, Buckingham-
shire, UK). [3H]AAI was kindly provided by Tapan Ray, Novartis Institutes of
Biomedical Research (East Hanover, NJ). The radiochemical purity of
[3H]AAI was Ͼ98% and chemical purity was Ն93%. Anti-Cyp1A1 rabbit IgG
and anti-Cyp1A2 goat IgG and corresponding peroxidase-conjugated second-
ary antibodies were purchased from Santa Cruz Biotechnology, Inc. (Santa
Cruz, CA). Protein electrophoresis gels, membranes, buffers, and electrochem-
ical detection kits were obtained from Thermo Fisher Scientific. All other
chemicals and reagents were purchased from commercial sources.
Urinalysis kits. Mouse Albuwell microalbuminuria enzyme-linked immu-
nosorbent assay kits were purchased from Exocell (Philadelphia, PA). Creat-
inine QuantiChrom assay kits were purchased from BioAssay Systems (Hay-
ward, CA). All urinalysis kits were used following the manufacturer’s
instructions.
Mice. Animal protocols were reviewed and approved by the Stony Brook
Institutional Animal Care and Use Committee. Breeding pairs of Cyp1A2
knockout mice were obtained from Dr. F. Gonzalez (National Cancer Institute,
National Institutes of Health, Bethesda, MD). A colony was maintained by
breeding in the Stony Brook University animal facility. Control 129S1/SvImJ
Microsomal Preparation and Protein Quantification. Hepatic micro-
somes were prepared by homogenization of freshly thawed mouse liver in a
mice were purchased from The Jackson Laboratory (Bar Harbor, Maine). All Potter S homogenizer (B. Braun Biotech Inc., Allentown, PA) containing
experiments used 8-week-old male mice.
microsomal homogenization buffer (0.25 M sucrose, 0.154 M KCl, 0.05 M
Metabolism of [3H]AAI by Human Liver Microsomes. For all the in vitro Tris-HCl, 0.001 M EDTA, and 0.25 mM phenylmethylsulfonyl fluoride, pH
metabolism incubations in this report, the buffer components were 100 mM
7.4) at a concentration of 3 ml of buffer/g of tissue. The homogenate was
potassium phosphate buffer (pH 7.4) and 5 mM MgCl2 (final concentrations). centrifuged at 10,000g for 20 min at 4°C. The supernatant (S9) was transferred
For the glucuronidation reactions (containing UDPGA), microsomes were to fresh ultracentrifuge tubes and centrifuged at 105,000g for 1 h at 4°C. The

CYP1A2 DETOXICATES ARISTOLOCHIC ACID
763
top lipid layer and cytosol were aspirated, and the microsomal pellet was
resuspended in storage buffer (0.1 M potassium phosphate buffer, pH 7.4, 20%
glycerol, v/v) at an estimated final concentration of 1 ml of storage buffer per
g of tissue (starting amount) using a Dounce homogenizer. Aliquots of ap-
proximately 50 ␮l each were stored at Ϫ80°C. The amount of microsomal
protein was determined by the Bradford protein assay method.
32P-Postlabeling/Polyacrylamide Gel Electrophoresis Analysis. DNA di-
gestion mixtures were incubated at 37°C for 40 min with 10 mCi of
[␥-32P]ATP and 3Ј-phosphatase-free T4 polynucleotide kinase (10 U), fol-
lowed by incubation with apyrase (50 mU) for 30 min, as described previously
(Dong et al., 2006). The 32P-labeled products were separated by electrophore-
sis for 4 to 5 h on a nondenaturing 30% polyacrylamide gel (35 ϫ 42 ϫ 0.04
cm) with 1500 to 1800 V/20 to 40 mA. The position of 32P-labeled adducts was
established by ␤-PhosphorImager analysis (GE Healthcare). To quantify the
32P-labeled products, integrated values were measured using a ␤-PhosphorIm-
ager and compared with the standards. Because ALI-DNA adducts were not
available, a known amount (0.0152 pmol) of ALII-dA- or ALII-dG-modified
oligodeoxynucleotides was used as a standard (obtained from Francis Johnson,
Department of Chemistry, Stony Brook University) (Attaluri et al., 2010).
Immunoblotting Microsomal Proteins. Fifty micrograms of each micro-
somal protein preparation was subjected to electrophoresis through a 4 to 20%
gradient SDS-polyacrylamide gel (Thermo Fisher Scientific) and then electro-
transferred to a nitrocellulose membrane (Thermo Fisher Scientific). Mem-
branes were preblocked with 5% milk protein in TBST (150 mM sodium
chloride, 50 mM Tris, pH 7.5, and 0.2% Tween 20 detergent) and then were
incubated at room temperature with primary antibody, at 0.2 ␮g/ml, in block-
ing buffer for 2 h followed by washing in TBST. Secondary antibody was
diluted (1:20,000) in blocking buffer and incubated with the membrane for 1 h
followed by washing in TBST. Antibody was detected with electrochemical
reagents and film exposure following the manufacturer’s protocols (Thermo
Fisher Scientific).
Metabolism of [3H]AAI in Mouse Liver Microsomes. The metabolism of
[3H]AAI was examined in mouse liver microsomes prepared from control or
3-MC-treated wild-type or Cyp1a2(Ϫ/Ϫ) mice. Mouse liver microsomes (1 mg
of protein/ml) were preincubated with alamethicin, as described above, before
incubation with [3H]AAI (95 ␮M, final concentration). The reactions were
initiated with UDPGA and NADPH, and the samples were incubated for 30
min at 37°C. Control incubations contained all reaction components without
cofactors. The buffer components, sample processing, and HPLC analysis were
as described above.
Extraction and Digestion of Renal Cortical DNA. DNA was extracted
from cortical slices from freshly isolated kidneys using a DNeasy Blood and
Tissue kit (QIAGEN, Valencia, CA) according to the manufacturer’s protocol.
The concentration of DNA was determined by UV spectroscopy. One micro-
gram of DNA was digested enzymatically at 37°C for 16 h in 100 ml of 17 mM
sodium succinate buffer (pH 6.0) containing 8 mM CaCl2, micrococcal nucle-
ase (30 U), and spleen phosphodiesterase (0.15 U) (Dong et al., 2006). The
reaction mixture was incubated for 1 h with nuclease P1 (1 U), whereupon 200
␮l of water was added. The reaction samples then were extracted twice with
200 ␮l of butanol; the butanol fractions were combined, back-extracted with 50
␮l of distilled water, and evaporated to dryness.
A
COOH
AAIa-glucuronide
AAIa-sulfate
NO2
UGT/ST
OH
AAIa
H
UGT/ST
*
COOH
NO2
AAIaL-glucuronide
AAIaL-sulfate
ALIa
OH
OCH3
CYP
AAI
FIG. 1. A, pathways for AAI metabolism in
mammals. AAI is demethylated to form AAIa in
H
OH
UGT/ST
a reaction catalyzed by cytochrome P450 (CYP).
Subsequent conjugation to glucuronic acid or
sulfate, catalyzed by UDP glucuronyl trans-
ferases (UGT) or sulfotransferases (ST), en-
hances excretion. Cellular nitroreductases (NR)
catalyze formation of the biologically inactive
aristolactam (ALI). A reactive intermediate in
this pathway, possibly N-OH-ALI forms cova-
lent adducts with DNA and proteins. B, struc-
tures of AAI-derived DNA adducts: ALI-dA and
ALI-dG. ء, position of 3H in the radiolabeled
AAI used in this study.
AAIL-glucuronide
AAIL-sulfate
NR
OCH3
OCH3
N-OH-ALI
ALI
DNA and Protein adducts
B
ALI-dA
ALI-dG
CH3O
deoxyribose
CH3O
deoxyribose

764
ROSENQUIST ET AL.
HLM with NADPH
HLM no cofactors
A
HLM with NADPH+UDPGA
3000
2500
2000
1500
1000
500
3000
3000
2500
2000
1500
1000
500
AAI
AAI
AAI
2500
2000
1500
1000
500
AAIa
AAI-glucuronide
AAIa
0
0
0
0
5
10 15 20 25 30 35 40 45 50
Time (min)
0
5
10 15 20 25 30 35 40 45 50
Time (min)
0
5
10 15 20 25 30 35 40 45 50
Time (min)
CYP1A1
CYP1A2
B
P450 Control
3000
3000
2500
2000
1500
1000
500
3000
2500
2000
1500
1000
500
AAI
AAI
AAI
2500
2000
1500
1000
500
AAIa
AAIa
0
0
0
0
5
10 15 20 25 30 35 40 45 50
Time (min)
0
5
10 15 20 25 30 35 40 45 50
Time (min)
0
5
10 15 20 25 30 35 40 45 50
Time (min)
FIG. 2. Metabolism of AAI by human P450s. A, [3H]AAI was incubated with human liver microsomes (HLM). Reaction products were separated by HPLC, and the
radioactivity in each fraction was determined. Peaks corresponding to AAIa and AAI glucuronide are indicated. B, recombinant human P450 enzymes were tested for their
ability to metabolize AAI. Shown is the production of AAIa from AAI by human CYP1A1 and CYP1A2.
TABLE 1
Results
Kinetic parameters of human and rat CYP1A1 and CYP1A2 for AAI metabolism
Figure 1A shows the current model of AAI metabolism in mam-
Km
Vmax
Vmax/Km
mals. Two pathways are envisioned; intermediates in either pathway
may be substrates for the other. AAI can be directly demethylated to
form AAIa, as shown in the upper arm of the figure. AAIa is relatively
nontoxic in kidney cells in culture (Balachandran et al., 2005) and in
mice (Shibutani et al., 2010). AAI undergoes nitro-reduction to ALI,
as shown in the lower arm of the pathway (Fig. 1A). Activated
intermediates in this pathway form covalent adducts with DNA (Fig.
␮M
nmol ⅐ hϪ1 ⅐ nmol P450Ϫ1 ml ⅐ hϪ1 ⅐ nmol P450Ϫ1
Human CYP1A1 19.7 Ϯ 5.9
Human CYP1A2 53.6 Ϯ 15.0
2330 Ϯ 236
1692 Ϯ 256
227 Ϯ 111
523 Ϯ 62
118
31.6
0.344
43.2
Rat CYP1A1
Rat CYP1A2
659 Ϯ 247
12.0 Ϯ 5.0
1B) and proteins. The end products are conjugated to glucuronic acid lower than that for rat CYP1A2, whereas human CYP1A1 was ϳ4-
or sulfate by phase II enzymes, enhancing solubility and promoting fold more efficient than human CYP1A2 in producing this metabolite.
excretion.
To determine whether CYP1A2 is required in vivo for metabolism
As shown in Fig. 2A, [3H]AAI is metabolized by human liver of AAI in mice we treated Cyp1a2(Ϫ/Ϫ) and Cyp1a2(ϩ/ϩ) control
microsomes in the presence of NADPH by demethylation to form mice with a single, moderate dose (2.0 mg/kg) of AAI. Urine was
AAIa. In the presence of UDPGA, a glucuronidation product was collected during the week after treatment and analyzed for albumin
found with an elution time of ϳ27 min under these chromatographic and creatinine content. The toxin induces proximal tubule cell toxicity
conditions. Identification of AAIa and the glucuronidation product in manifested by reduced reuptake of low-molecular-weight proteins in
the human liver microsome samples was confirmed by liquid chro- the proximal nephron and the presence of injury markers in the urine
matography-tandem mass spectrometry (data not shown). We exam- (Huang et al., 2009). Four days after AAI treatment, there is an
ined 18 individual human cytochrome P450 enzymes for their ability increase in the amount of urinary albumin, as shown in Fig. 3. As
to metabolize AAI. AAIa was formed primarily by CYP1A2 and proximal tubule function is restored, albuminuria decreases and even-
CYP1A1 (Fig. 2B). With the exception of marginal CYP2C9 activity, tually disappears. The urine of AAI-treated Cyp1a2(Ϫ/Ϫ) mice con-
no other human P450 tested was active in this assay. Because of tained higher levels of albumin and for a longer period of time than
negligible expression of CYP1A1 in uninduced human liver, it is that of Cyp1a2(ϩ/ϩ) control mice, indicating a higher degree of
likely that CYP1A2 predominates in the formation of this metabolite nephrotoxicity for AAI in CYP1A2-deficient mice.
in human liver. Kinetic parameters associated with metabolism of
3-MC induces several biotransformation activities, including
[3H]AAI, catalyzed by human or rat recombinant CYP1A1 and CYP1A, and pretreatment of mice with 3-MC is reported to increase
CYP1A2 enzymes, are summarized in Table 1. Rat and human resistance to AAI-elicited nephrotoxicity (Xue et al., 2008). We
CYP1A2 enzymes were equally efficient in demethylating AAI pretreated Cyp1a2(Ϫ/Ϫ) or control mice with 60 mg/kg 3-MC or
(Vmax/Km values of 43.2 and 31.6 ml ⅐ hϪ1 ⅐ nmol of P450Ϫ1
,
vehicle and 24 h later treated them with AAI as in the previous
respectively). However, species differences were found for CYP1A1. experiment. Both 3-MC-pretreated control and Cyp1a2(Ϫ/Ϫ) mice
Thus, formation of AAIa catalyzed by rat CYP1A1 was 126-fold displayed increased resistance to AAI relative to noninduced mice

CYP1A2 DETOXICATES ARISTOLOCHIC ACID
765
FIG. 3. Nephrotoxic effects of AAI in mice. Mice lacking CYP1A2
[Cyp1a2(Ϫ/Ϫ)] or control mice from the isogenic 129S1/SvImJ
strain [Cyp1a2(ϩ/ϩ)] were given intraperitoneal injections on day
0 with AAI (2 mg/kg); urine was collected at the times indicated,
and albumin and creatinine levels were measured. The mice were
given injections of 3-MC (60 mg/kg) or corn oil (Ϫ3MC) 24 h
before treatment with AAI. All data represent averages of two mice
(error bar ϭ S.E.M.). Urinary albumin levels in untreated mice of
both strains are Ͻ0.1 mg/mg creatinine.
(Fig. 3). AAI-dependent albuminuria in the induced Cyp1a2(Ϫ/Ϫ)
mice was equivalent to that of uninduced control mice.
In vitro studies have established that AAI may be metabolized by
nitroreduction to ALI and/or demethylation to AAIa [for review, see
Stiborova´ et al. (2008)]. Aristolactams are formed by several cellular
enzymes, including NADPH:cytochrome P450 reductase, prostaglan-
din H synthase, and NAD(P)H:quinone oxidoreductase. Microsomal
AA metabolism, particularly reactions catalyzed by CYP1A enzymes
under anaerobic conditions also lead to the formation of genotoxic
intermediates (Schmeiser et al., 1997; Stiborova´ et al., 2001b,
2005a,b). Under aerobic conditions, formation of aristolactam-DNA
adducts in reactions catalyzed by hepatic microsomes was greatly
reduced (Schmeiser et al., 1997) and AAIa was formed (Sistkova et
al., 2008). We confirm that under aerobic conditions microsomal
CYP1A enzymes demethylate AAI to form AAIa and show that they
are kinetically competent to do so in vivo. AAIa is nontoxic in cells
(Balachandran et al., 2005; Shibutani et al., 2010) and even in the
presence of reducing agents is incapable of forming covalent adducts
with DNA (Shibutani et al., 2010).
In this study, we report that among 18 human cytochrome P450
enzymes tested, only the CYP1A1 and CYP1A2 enzymes have robust
AAI demethylating activity. The only other P450 with detectable
activity against AAI was CYP2C9. Under homeostatic conditions
CYP1A2 is the major CYP1A enzyme expressed in livers of humans
and mice (Gonzalez et al., 1984; Ikeya et al., 1989). Consistent with
these observations, we detected little AAI demethylation activity in
hepatic microsomes prepared from CYP1A2-null mice. In a pilot
genetic study, a weak association between the human CYP3A5*1
allele and Balkan endemic nephropathy, a disease resulting from
chronic dietary exposure to AA (Grollman et al., 2007), was noted
(Atanasova et al., 2005). However we did not detect AAI demethyl-
ation activity by CYP3A5.
To assess the AAI-metabolizing activities in Cyp1a2(Ϫ/Ϫ) mice,
we isolated hepatic microsomes from knockout and control mice both
with and without pretreatment with 3-MC. These microsomes were
used in the AAI assay described above, and results are shown in
Fig. 4, A and B. Microsomes from control mice produce primarily
AAIa, the level of which is increased by pretreatment with 3-MC.
Microsomes from Cyp1a2(Ϫ/Ϫ) mice produced little AAIa, providing
confirmatory evidence that CYP1A2 is the primary hepatic enzyme
responsible for demethylating AAI. Microsomes from 3-MC-treated
Cyp1a2(Ϫ/Ϫ) mice produce increased AAIa, indicating the induction
of another P450 activity [most likely CYP1A1 (vide infra)]. Several
unidentified metabolites (indicated by peaks numbered 1, 2, and 3 in
Fig. 4B) also are produced.
3-MC is known to induce the expression of CYP1A1 and CYP1A2
and, in Fig. 4C, we show that in the Cyp1a2(Ϫ/Ϫ) mice CYP1A1
protein was indeed induced in the liver. In Cyp1a2(ϩ/ϩ) mice, we
observed induction of CYP1A2 but not of CYP1A1 under these
conditions. The increase in CYP1A2 is consistent with the elevated
AAIa levels observed and increased resistance of these mice to AAI
relative to that of uninduced mice.
We hypothesize that animals unable to detoxicate AAI through
demethylation would be more susceptible to formation of aristolactam
I-DNA adducts (ALI-dG and ALI-dA) that are side-products of ni-
troreduction of AAI, the second pathway contributing to AAI metab-
olism. To assess this reaction, we isolated kidney cortex DNA from
mice euthanized 4 and 20 h after treatment with AAI and measured
ALI-DNA adducts by the 32P-postlabeling technique. As shown in
Fig. 5, A and B, Cyp1a2(Ϫ/Ϫ) mice, deficient for AAI-demethylating
activity, accumulate higher levels of ALI-DNA adducts.
Mice lacking hepatic P450 activity are extremely sensitive to AA,
indicating that AA metabolism occurs primarily in the liver (Xiao et
al., 2008). Consistent with this report, AAIa was not detected in renal
microsomes (data not shown). Induction of CYP1A activity with
3-MC (Xue et al., 2008) or ␤-naphthoflavone (Xiao et al., 2009)
increases resistance to AAI-elicited nephrotoxicity. In this study,
3-MC treatment increased hepatic microsomal AAI demethylation
activity in both control and CYP1A2-null mice. However, in
Cyp1a2(ϩ/ϩ) mice 3-MC treatment increased the hepatic expression
of CYP1A2 and also increased AAIa formation. In Cyp1a2(Ϫ/Ϫ)
mice, we observed induction of CYP1A1 expression. In addition to
Discussion
The ability to demethylate AAI may determine the toxic potential
of AA in mammalian species. In this article, we established that
CYP1A2 is the major enzyme responsible for demethylation of AAI
in mice. We report the most extensive survey of P450 enzymes for
AAI-metabolizing activities to date and the first kinetic analysis of the
capabilities of the CYP1A enzymes with respect to AAI demethyl-
ation. Human and rodent CYP1A2 are kinetically similar with respect
to this important reaction. If expressed, human hepatic CYP1A1 is
also kinetically competent to contribute to detoxication of AAI.

766
ROSENQUIST ET AL.
Uninduced
A
1000
1000
AAI
Cyp1a2-/-
AAI
Cyp1a2+/+
800
600
400
200
0
800
600
400
200
0
AAI
AAI
a
a
1
1
0
5
10 15 20 25 30 35 40 45
0
5
10 15 20 25 30 35 40 45
Time (min)
FIG. 4. Metabolism of AAI by hepatic micro-
somes from Cyp1a2(Ϫ/Ϫ) mice. Microsomes
isolated from livers of Cyp1a2(ϩ/ϩ) or
Cyp1a2(Ϫ/Ϫ) mice were incubated with
[3H]AAI as described under Materials and
Methods. Products were analyzed by HPLC
with off-line low-level radioactivity detection.
A, products produced by microsomes from un-
induced mice. B, products produced by micro-
somes from mice pretreated with 3-MC 24 h
before treatment with AAI. Peaks correspond-
ing to AAI and AAIa are indicated. Peaks 1, 2,
and 3 are unidentified products. Each result was
reproduced in two mice. C, immunoblots of
hepatic microsomal protein from two mice ob-
tained from each treatment described above and
reacted with antibodies to CYP1A2 (␣-
CYP1A2) or CYP1A1 (␣-CYP1A1). Arrow-
head indicates Cyp1A1.
Time (min)
Induced with 3MC
B
1000
800
600
400
200
0
1000
Cyp1a2-/-
AAI
AAI
Cyp1a2+/+
800
600
400
200
0
AAI
a
AAI
a
1
1
2
3
0
5
10 15 20 25 30 35 40 45
Time (min)
0
5
10 15 20 25 30 35 40 45
Time (min)
Cyp1A2 -/-
Cyp1a2 +/+
C
3-MC
- - + +
- - + +
α-CYP1A2
α-CYP1A1
increased AAIa production, hepatic microsomes from 3-MC-induced (Olivier et al., 2002). To date, AA carcinogenicity has been investi-
Cyp1a2(Ϫ/Ϫ) mice also produced several, as yet unidentified, me- gated only in the NMRI strain of mice (Mengs, 1988). In these mice,
tabolites of AAI. We speculate that these metabolites include AAIa AA induces primarily cancer of the forestomach and a lower fre-
phase II conjugation products. Thus, in the absence of CYP1A2, quency of other tumors including renal adenomas. Our results indicate
different activities are induced by 3MC, suggesting the presence of that the levels of ALI-DNA adducts increase significantly in the
additional enzymes that metabolize AAI.
renal-urothelial system of Cyp1a2(Ϫ/Ϫ) mice. It will be interesting to
We tested Cyp1a2 knockout mice for their sensitivity to AAI. determine whether CYP1A2 deficiency increases the rate of urothelial
Microalbuminuria, an indicator of renal proximal tubule dysfunction cell carcinogenesis in mice.
was elevated after AAI treatment in CYP1A2-deficient mice. Thus,
Both AAI and AAII induce aristolactam DNA adducts (Schmeiser
CYP1A2 does play an important role in detoxication of AAI in mice. et al., 1988; Shibutani et al., 2007), but AAII is relatively nontoxic to
According to the current model of AAI metabolism, reduction renal cells in vitro (Balachandran et al., 2005) and in vivo (Shibutani
in AAI demethylation activity is expected to increase production of et al., 2007). Thus, AA-induced DNA damage is unlikely to be
aristolactam metabolites and associated toxicity resulting from responsible for the profound toxicity observed in renal proximal
ALI-DNA and protein adducts. As predicted, we observed greater tubule cells. It is yet to be established whether the cytotoxic com-
accumulation of ALI-DNA adducts in the renal cortex of pound involved is AAI itself or a metabolite. Reduction of AAI
Cyp1a2(Ϫ/Ϫ) mice relative to wild-type mice. This result is consis- demethylation in Cyp1a2(Ϫ/Ϫ) mice can have the effect of increasing
tent with the proposed model of AAI metabolism.
In humans, at least 50% of patients with AA-associated nephrop-
either.
Finally, it has been noted that urine from AA-exposed humans
athies develop urothelial carcinomas of the upper urinary tract, char- contains metabolites of aristolactam but not those derived from AAIa
acterized by a predominance of A:T3T:A transversions in the TRP53 (Krumbiegel et al., 1987). This implies that the nitroreduction path-
gene (Grollman et al., 2007). This effect is in sharp contrast to way is more efficient than demethylation in metabolizing AAI in
sporadic urothelial cancers in which A:T3T:A transversions are rare humans. However, our results indicate that human CYP1A2 is kinet-

CYP1A2 DETOXICATES ARISTOLOCHIC ACID
767
FIG. 5. Increase in ALI-DNA adducts in mice
lacking CYP1A2. Genomic DNA was isolated
from the renal cortex 4 and 20 h after treatment
with AAI (2 mg/kg). 32P-postlabeling analysis
was performed as described under Materials
and Methods. A, 32P ALI-DNA adducts in renal
cortex of three Cyp1a2(Ϫ/Ϫ) mice (lanes 2–4)
and three Cyp1a2(ϩ/ϩ) mice (lanes 5–7) ana-
lyzed 20 h after treatment with AAI. Standards
(S) in lane 1 were ALII-dG and ALII-dA. ALII
nucleosides are 30 Da less than ALI nucleosides
and migrate slightly faster than ALI adducts
(indicated by ALI-dG and ALI-dA) under the
conditions used here. B, radioactivity in each
band from the experiment shown in A and a
similar experiment with a 4-h time point was
determined by PhosphoImager analysis, and the
concentrations of ALI-dG and ALI-dA were
established by comparison to the standards. The
concentrations of the promutagenic ALI-dA
DNA adduct from either Cyp1a2(ϩ/ϩ) or
Cyp1a2(Ϫ/Ϫ) mice is shown. The y-axis indi-
cates ALI-dA adducts per 100,000 nucleotides.
n ϭ 3. ءء, p Ͻ 0.01, Student’s one-tailed t test.
as a method for biomarker discovery in the urine of aristolochic-acid-treated mice. Electro-
phoresis 30:1168–1174.
Ikeya K, Jaiswal AK, Owens RA, Jones JE, Nebert DW, and Kimura S (1989) Human CYP1A2:
sequence, gene structure, comparison with the mouse and rat orthologous gene, and differences
in liver 1A2 mRNA expression. Mol Endocrinol 3:1399–1408.
Jameson CW, Lunn R, Jeter S, Garner S, Atwood S, Carter G, Levy D and JPC (2008)
Background document for aristolochic acid-related exposures, in Report on Carcinogens, 12th
ed, pp 1–228, National Toxicology Program, Research Triangle Park, NC.
Krumbiegel G, Hallensleben J, Mennicke WH, Rittmann N, and Roth HJ (1987) Studies on the
metabolism of aristolochic acids I and II. Xenobiotica 17:981–991.
Lord GM, Hollstein M, Arlt VM, Roufosse C, Pusey CD, Cook T, and Schmeiser HH (2004)
DNA adducts and p53 mutations in a patient with aristolochic acid-associated nephropathy.
Am J Kidney Dis 43:e11–e17.
ically similar to rodent CYP1A2 with respect to AAI demethylation
activity. Thus, some other aspect of AAI metabolism, perhaps hepatic
bioavailability, CYP1A2 expression levels, or increased activity of
nitroreductases, underlies the differential susceptibility to AAI be-
tween humans and rodents.
Acknowledgments. We are indebted to Penelope Strockbine for
expert technical assistance and to Francis Johnson, Stony Brook
University, for a gift of AAII-modified oligonucleotides.
Meinl W, Pabel U, Osterloh-Quiroz M, Hengstler JG, and Glatt H (2006) Human sulphotrans-
ferases are involved in the activation of aristolochic acids and are expressed in renal target
tissue. Int J Cancer 118:1090–1097.
Mengs U (1988) Tumour induction in mice following exposure to aristolochic acid. Arch Toxicol
61:504–505.
Olivier M, Eeles R, Hollstein M, Khan MA, Harris CC, and Hainaut P (2002) The IARC TP53
database: new online mutation analysis and recommendations to users. Hum Mutat 19:607–
614.
Schmeiser HH, Frei E, Wiessler M, and Stiborova M (1997) Comparison of DNA adduct
formation by aristolochic acids in various in vitro activation systems by 32P-post-labelling:
evidence for reductive activation by peroxidases. Carcinogenesis 18:1055–1062.
Schmeiser HH, Pool BL, and Wiessler M (1986) Identification and mutagenicity of metabolites
of aristolochic acid formed by rat liver. Carcinogenesis 7:59–63.
References
Atanasova SY, von Ahsen N, Toncheva DI, Dimitrov TG, Oellerich M, and Armstrong VW
(2005) Genetic polymorphisms of cytochrome P450 among patients with Balkan endemic
nephropathy (BEN). Clin Biochem 38:223–228.
Attaluri S, Bonala RR, Yang IY, Lukin MA, Wen Y, Grollman AP, Moriya M, Iden CR and
Johnson F (2010) DNA adducts of aristolochic acid II: total synthesis and site-specific
mutagenesis studies in mammalian cells. Nucleic Acids Res 38:339–352.
Balachandran P, Wei F, Lin RC, Khan IA, and Pasco DS (2005) Structure activity relationships
of aristolochic acid analogues: toxicity in cultured renal epithelial cells. Kidney Int 67:1797–
1805.
Debelle FD, Vanherweghem JL, and Nortier JL (2008) Aristolochic acid nephropathy: a world-
wide problem. Kidney Int 74:158–169.
Schmeiser HH, Schoepe KB, and Wiessler M (1988) DNA adduct formation of aristolochic acid
I and II in vitro and in vivo. Carcinogenesis 9:297–303.
Dong H, Suzuki N, Torres MC, Bonala RR, Johnson F, Grollman AP, and Shibutani S (2006)
Quantitative determination of aristolochic acid-derived DNA adducts in rats using 32P-
postlabeling/polyacrylamide gel electrophoresis analysis. Drug Metab Dispos 34:1122–1127.
Gonzalez FJ, Tukey RH, and Nebert DW (1984) Structural gene products of the Ah locus.
Transcriptional regulation of cytochrome P1–450 and P3–450 mRNA levels by 3-methyl-
cholanthrene. Mol Pharmacol 26:117–121.
Grollman AP, Shibutani S, Moriya M, Miller F, Wu L, Moll U, Suzuki N, Fernandes A,
Rosenquist T, Medverec Z, et al. (2007) Aristolochic acid and the etiology of endemic
(Balkan) nephropathy. Proc Natl Acad Sci USA 104:12129–12134.
Shibutani S, Bonala RR, Rosenquist T, Rieger R, Suzuki N, Johnson F, Miller F, and Grollman
AP (2010) Detoxification of aristolochic acid I by O-demethylation; less nephrotoxicity and
genotoxicity of aristolochic acid Ia in rodents Int J Cancer doi: 10.1002/ijc.25141.
Shibutani S, Dong H, Suzuki N, Ueda S, Miller F, and Grollman AP (2007) Selective toxicity of
aristolochic acids I and II. Drug Metab Dispos 35:1217–1222.
Sistkova J, Hudecek J, Hodek P, Frei E, Schmeiser HH, and Stiborova M (2008) Human
cytochromes P450 1A1 and 1A2 participate in detoxication of carcinogenic aristolochic acid.
Neuro Endocrinol Lett 29:733–737.
Huang F, Clifton J, Yang X, Rosenquist T, Hixson D, Kovac S, and Josic D (2009) SELDI-TOF

768
ROSENQUIST ET AL.
Stiborova´ M, Frei E, Arlt VM, and Schmeiser HH (2008) Metabolic activation of carcinogenic
aristolochic acid, a risk factor for Balkan endemic nephropathy. Mutat Res 658:55–67.
Stiborova´ M, Frei E, Breuer A, Wiessler M, and Schmeiser HH (2001a) Evidence for reductive
activation of carcinogenic aristolochic acids by prostaglandin H synthase-32P-postlabeling
analysis of DNA adduct formation. Mutat Res 493:149–160.
Stiborova´ M, Frei E, Hodek P, Wiessler M, and Schmeiser HH (2005a) Human hepatic and renal
microsomes, cytochromes P450 1A1/2, NADPH:cytochrome P450 reductase and prostaglan-
din H synthase mediate the formation of aristolochic acid-DNA adducts found in patients with
urothelial cancer. Int J Cancer 113:189–197.
Stiborova´ M, Frei E, Sopko B, Sopkova´ K, Markova´ V, Lankova´ M, Kumsty´rova´ T, Wiessler M,
and Schmeiser HH (2003) Human cytosolic enzymes involved in the metabolic activation of
carcinogenic aristolochic acid: evidence for reductive activation by human NAD(P)H:quinone
oxidoreductase. Carcinogenesis 24:1695–1703.
Stiborova´ M, Frei E, Sopko B, Wiessler M, and Schmeiser HH (2002) Carcinogenic aristolochic
acids upon activation by DT-diaphorase form adducts found in DNA of patients with Chinese
herbs nephropathy. Carcinogenesis 23:617–625.
adducts found in DNA of patients with Chinese herbs nephropathy. Gen Physiol Biophys
20:375–392.
Stiborova´ M, Sopko B, Hodek P, Frei E, Schmeiser HH, and Hudecek J (2005b) The binding of
aristolochic acid I to the active site of human cytochromes P450 1A1 and 1A2 explains their
potential to reductively activate this human carcinogen. Cancer Lett 229:193–204.
Xiao Y, Ge M, Xue X, Wang C, Wang H, Wu X, Li L, Liu L, Qi X, Zhang Y, et al. (2008)
Hepatic cytochrome P450s metabolize aristolochic acid and reduce its kidney toxicity. Kidney
Int 73:1231–1239.
Xiao Y, Xue X, Wu YF, Xin GZ, Qian Y, Xie TP, Gong LK, and Ren J (2009) ␤-Naphthoflavone
protects mice from aristolochic acid-I-induced acute kidney injury in a CYP1A dependent
mechanism. Acta Pharmacol Sin 30:1559–1565.
Xue X, Xiao Y, Zhu H, Wang H, Liu Y, Xie T, and Ren J (2008) Induction of P450 1A by
3-methylcholanthrene protects mice from aristolochic acid-I-induced acute renal injury. Neph-
rol Dial Transplant 23:3074–3081.
Stiborova´ M, Frei E, Wiessler M, and Schmeiser HH (2001b) Human enzymes involved in the
metabolic activation of carcinogenic aristolochic acids: evidence for reductive activation by
cytochromes P450 1A1 and 1A2. Chem Res Toxicol 14:1128–1137.
Address correspondence to: Dr. Thomas A. Rosenquist, Department of
Pharmacological Sciences, State University of New York, One Nicolls Road,
Stony Brook, NY 11794-8651. E-mail: rosenquist@pharm.stonybrook.edu
Stiborova´ M, Ha´jek M, Frei E, and Schmeiser HH (2001c) Carcinogenic and nephrotoxic
alkaloids aristolochic acids upon activation by NADPH:cytochrome P450 reductase form