Comparison of Aristolochic acid I derived DNA adduct levels in human renal toxicity models

Article abstract of DOI:10.1016/j.tox.2019.03.013

Aristolochic acid (AA) dependent human nephropathy results either from environmental exposure to Aristolochiaceae plant subspecies or their use in traditional phytotherapy. The toxic components are structurally related nitrophenanthrene carboxylic acids, i.e. Aristolochic acid I (AAI) and II (AAII). AAI is considered to be the major cause of Aristolochic acid nephropathy, characterized by severe renal fibrosis and upper urothelial cancer. Following enzymatic activation in kidney and/or liver, AAI metabolites react with genomic DNA to form persistent DNA adducts with purines. To determine whether AAI can be activated in human renal cells to form DNA adducts, we exposed telomerase immortalized renal proximal tubular epithelial cells (RPTEC/TERT1), the human embryonic kidney (HEK293) cell line, as well as primary human kidney cells (pHKC) to AAI in vitro. We modified an isotope dilution ultra-performance liquid chromatography/tandem mass spectrometry (ID-UPLC-MS/MS) based method for the quantification of dA-AAI adducts in genomic DNA. In addition, time dependent accumulation of adducts in renal cortex and bladder tissue from AAI/II treated Eker rats were used to validate the detection method. AAI-induced toxicity in human renal cells was determined by dA-AAI adduct quantification, the impact on cell viability, and NQO1 expression and activity. Our findings demonstrated adduct formation in all cell lines, although only pHKC and RPTEC/TERT1 expressed NQO1. The highest adduct formation was detected in pHKC despite low NQO1 expression, while we observed much lower adduct levels in NQO1-negative HEK293 cells. Adduct formation and decreased cell viability correlated only weakly. Therefore, our data suggested that i.) enzymes other than NQO1 could be at least equally important for AA bioactivation in human renal proximal tubule cells, and ii.) the suggested correlation between adduct levels and viability appears to be questionable.

Full text of DOI:10.1016/j.tox.2019.03.013

Accepted Manuscript
Title: Comparison of Aristolochic acid I derived DNA adduct
levels in human renal toxicity models
Authors: Heinke Bastek, Tabea Zubel, Kerstin Stemmer,
Aswin Mangerich, Sascha Beneke, Daniel R. Dietrich
PII:
DOI:
Reference:
S0300-483X(19)30073-3
https://doi.org/10.1016/j.tox.2019.03.013
TOX 52193
To appear in:
Toxicology
Received date:
Revised date:
Accepted date:
23 January 2019
22 March 2019
28 March 2019
Please cite this article as: Bastek H, Zubel T, Stemmer K, Mangerich A, Beneke S,
Dietrich DR, Comparison of Aristolochic acid I derived DNA adduct levels in human
renal toxicity models, Toxicology (2019), https://doi.org/10.1016/j.tox.2019.03.013
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Comparison of Aristolochic acid I derived DNA adduct levels in human renal toxicity
models
Heinke Bastek1*, Tabea Zubel2*, Kerstin Stemmer3, Aswin Mangerich2, Sascha Beneke1, Daniel
R. Dietrich1$
1 Human and Environmental Toxicology, University of Konstanz, Konstanz, Germany
2Molecular Toxicology, University of Konstanz, Konstanz, Germany
3 Division of Pharmacology and Toxicology, Institute for Diabetes and Obesity, Helmholtz
Zentrum München, Neuherberg, Germany
$ Corresponding author
* Equal contribution
Address all correspondence and requests for reprints to:
Daniel R. Dietrich, Human and Environmental Toxicology, University of Konstanz,
Universitätsstraße 10, 78464 Konstanz, Germany.
Email: daniel.dietrich@uni-konstanz.de
Phone: +49-7531-88-3518
1

Abstract
Aristolochic acid (AA) dependent human nephropathy results either from environmental
exposure to Aristolochiaceae plant subspecies or their use in traditional phytotherapy. The toxic
components are structurally related nitrophenanthrene carboxylic acids, i.e. Aristolochic acid I
(AAI) and II (AAII). AAI is considered to be the major cause of Aristolochic acid nephropathy,
characterized by severe renal fibrosis and upper urothelial cancer. Following enzymatic
activation in kidney and/or liver, AAI metabolites react with genomic DNA to form persistent
DNA adducts with purines. To determine whether AAI can be activated in human renal cells to
form DNA adducts, we exposed telomerase immortalized renal proximal tubular epithelial cells
(RPTEC/TERT1), the human embryonic kidney (HEK293) cell line, as well as primary human
kidney cells (pHKC) to AAI in vitro. We modified an isotope dilution ultra-performance liquid
chromatography/tandem mass spectrometry (ID-UPLC-MS/MS) based method for the
quantification of dA-AAI adducts in genomic DNA. In addition, time dependent accumulation
of adducts in renal cortex and bladder tissue from AAI/II treated Eker rats were used to validate
the detection method. AAI-induced toxicity in human renal cells was determined by dA-AAI
adduct quantification, the impact on cell viability, and NQO1 expression and activity. Our
findings demonstrated adduct formation in all cell lines, although only pHKC and
RPTEC/TERT1 expressed NQO1. The highest adduct formation was detected in pHKC despite
low NQO1 expression, while we observed much lower adduct levels in NQO1-negative
HEK293 cells. Adduct formation and decreased cell viability correlated only weakly.
Therefore, our data suggested that i.) enzymes other than NQO1 could be at least equally
important for AA bioactivation in human renal proximal tubule cells, and ii.) the suggested
correlation between adduct levels and viability appears to be questionable.
2

Keywords:
Aristolochic acid, DNA adducts, UPLC-MS/MS, NQO1, human kidney cells
1. Introduction
Aristolochic acids (AA) are a group of structurally related nitrophenanthrene carboxylic acids
and have been detected in several Aristolochiaceae and Asarum plant subspecies, also known
as birthwort or pipevine (Debelle et al., 2008). The major AA derivatives contained in the plant
extract of Aristolochia species are Aristolochic acids I (AAI) and II (AAII). Human kidney
diseases associated with AA exposure are summarized under the term "Aristolochic acid
nephropathy” (AAN), caused either by ingestion of plants containing AA as part of traditional
phytotherapies (formerly known as “Chinese herbs nephropathy”), or by the environmental
contaminants in food (Balkan endemic nephropathy) (Jadot et al., 2017). The pathology of
AAN in humans is basically characterized by extensive cortical tubular atrophy, loss of tubules
and dense interstitial fibrosis, most prominent in the outer area of the cortex, while the glomeruli
are relatively spared (Jelakovic et al., 2011, Depierreux et al., 1994). Progressive renal fibrosis
results in loss of function and subsequently to end stage renal disease and kidney failure,
requiring dialysis or transplantation. AAN patients are also frequently diagnosed with urothelial
(transitional cell) carcinomas of the renal pelvis and ureter (Cosyns, 2003). These upper urinary
tract cancers (UUCs), which account for only 5% of all urinary tract cancers worldwide, are
present in ∼50% of AAN cases (Nortier et al., 2000).
In contrast, in rodents, employed as surrogate species for understanding the mechanisms
underlying AA induced toxicity, nephropathy and carcinogenesis do not readily reflect the
pathologies observed in humans. Indeed, the observed pathologies appear to differ dramatically
dependent on whether AA exposure occurred via subcutaneous, intraperitoneal application or
via oral gavage. Mice and rats exposed acutely and chronically to AAI via gavage developed
3

ulceration in the forestomach, extensive tubular necrosis and primarily forestomach and renal
cortical tumors in absence of renal fibrosis, respectively. Specific rodent strains developed renal
fibrosis when treated via subcutaneous or intraperitoneal injection with AAI, as reviewed by
Debelle et al (2008). Consequently, metabolic activation of AAI, either in the liver or directly
in the renal cortex, appears critical for the development of renal proximal tubule cytotoxicity,
inflammation, and ensuing epithelial to mesenchymal transition and fibrosis, whereby the
relationship between cytotoxicity and AAI-DNA adducts is not clear to date. Indeed, metabolic
activation of AAI leads to a highly reactive intermediate, a nitrenium/carbenium ion, which is
able to form DNA adducts including 7-(deoxyadenosine-N6-yl) aristolactam I (dA-AAI) and 7-
(deoxyguanosine-N2-yl) aristolactam I (dG-AAI) (Pfau et al., 1990). Currently, nitroreduction
e.g. by NAD(P)H dehydrogenase (quinone 1) (NQO1) is considered to be the main pathway
for AAI bioactivation (Fig. 1) and thus is assumed to be required for AAI-DNA adduct
formation in humans (Stiborova et al., 2003). However, other phase I enzymes, abundant in the
liver, e.g. the cytosolic nitroreductases (e.g. xanthine oxidases), cytochrome P450s (CYP)
enzymes (e.g. CYP1A1, 1A2, or NADPH:CYP reductase) or peroxidases may be involved in
AAI metabolism (Stiborova et al., 2005, Stiborova et al., 2009). Thus, one would expect that
hepatic activation of AAI leads to cytotoxic AAI-intermediates capable of forming DNA
adducts and therefore to observable hepatotoxicity and/or DNA-adducts in hepatocytes.
Contrary to expectations, patient derived material demonstrated no overt hepatotoxicity.
However, the liver and other organs of an AAI exposed patient (spleen, lung, adrenal, stomach,
small intestine, bladder, lymph nodes, pancreas, brain) exhibited similar, lower or higher DNA-
32
adduct level (expressed as relative adduct labeling using P-postlabelling and expressed as
mean dA-AAI adducts per 109 nucleotides) compared to the corresponding kidney of the same
patient (Arlt et al., 2004, Nortier et al., 2003). The mean relative dA-AAI adduct levels reported
in the renal cortex of male and female patients ranged between 30 - 3880 dA-AAI adducts per
109 nucleotides analyzed using UPLC-MS/MS (Yun et al., 2012). Although less accurate, dA-
4

AAI adducts were also frequently quantified in kidney samples using 32P-postlabeling and
ranged between 1 - 2340 (Arlt et al., 2004, Nortier et al. 2003, Yun et al., 2012). Due to their
long persistence, the specific DNA adducts in patients can serve as biomarkers of exposure to
AAI (Stiborová et al., 2017). However, most likely the presence of DNA adducts are of little
help for a better understanding of AA bioactivation and thus of the development of AAI
mediated renal proximal cytotoxicity and ensuing nephropathy, as these adducts per se can be
detected also in organs without pathological changes.
In order to address several of the questions raised above, we used male and female Eker rats
exposed daily to a mixture of AAI and AAII to establish and validate an isotope dilution (ID)-
UPLC-MS/MS state-of-the-art quantification method for dA-AAI adduct levels based on the
method previously published by Yun et al. (2012). AAI exposure mediated dA-AAI adduct
levels were determined in renal tissues of exposed rat and humans, but not compared to in vitro
settings. Therefore, we employed this dA-AAI adduct detection and quantification method to
determine AAI derived DNA adduct formation not only in renal cortex and bladder of exposed
Eker rats, but also in human renal in vitro cell systems. For this we employed human embryonic
kidney (HEK293) cells, representing a viral immortalized cell model, proliferating or
differentiated hTERT immortalized human renal proximal tubular epithelial cells
(RPTEC/TERT1) (Wieser et al., 2008), and differentiated primary human kidney cells (pHKC).
Finally, we compared dA-AAI adduct formation to AAI-induced reduction in cell viability and
examined the expression and activity of NQO1 relevant for AAI-bioactivation.
Material and methods
2.1 Materials
Aristolochic acid I (AAI) and the AA sodium salt mixture (41% AAI and 56% AAII) were
purchased from Sigma-Aldrich. [15N5]-labeled 2’-deoxyadenosine ([15N5]-dA) was purchased
5

from Cambridge Isotope Laboratories, Inc. 2’-deoxyadenosine, calf thymus DNA, DNase I
(Type IV, from bovine pancreas), alkaline phosphatase (Escherichia coli), nuclease P1
(Penicillium citrinum), phosphodiesterase I (Crotalus adamanteus venom), NADPH,
menadione and formic acid (LC-MS grade) were purchased from Sigma-Aldrich. Thiazolyl
blue (MTT), dimethyl sulfoxide (DMSO) (>99,98 %) and N,N-dimethylformamide (DMF)
(>99,8 %) were purchased from Roth. All solvents for mass spectrometry were of LC-MS grade
from Roth.
2.2 Rodent experiments
Six to ten weeks old heterozygous Tsc2 mutant Eker rats (Tsc2+/-, Long Evans) were purchased
from the MD Anderson Cancer Center, Smithville, Texas, USA and housed at the University
of Konstanz animal research facility under standard housing conditions. Prior to exposure, male
and female Eker rats were randomly assigned to dose groups and allowed to acclimatize to
laboratory conditions for 4 weeks. All animal experiments were approved by the State of
Baden-Württemberg, Germany (Regierungspräsidium Freiburg, AZ: G-03/65). Groups of three
male and female Eker rats were treated daily by oral gavage with a mixture of AAI (41%) and
AAII (56 %) for 1, 3, 7, and 14 days (10 mg / kg bodyweight). Time-matched vehicle controls
(n=3) received corresponding volumes of 0.1 M NaHCO3. Rats were weighed daily. At the end
of each treatment period, anesthetized rats were sacrificed by exsanguination subsequent to
retrograde perfusion with PBS. Organs were collected, sliced and sections were snap frozen
and stored at -80 °C until further usage.
2.3 Routine cell culture
All cells were cultured routinely in T75 flasks (Sarstedt) at 37 °C in a 5 % CO2 humidified
atmosphere and subcultured by trypsinization. Fresh medium was provided three times per
week. The human proximal tubule cell line RPTEC/TERT1 (Wieser et al., 2008) were
6

purchased from Evercyte GmbH, Vienna, Austria and cultured in hormonally defined medium
consisting of a 1 to 1 mixture of Dulbecco’s modified Eagle’s medium (Life Technologies,
Darmstadt, Germany) and Ham’s F-12 nutrient mix (Life Technologies) supplemented with
2 mM Glutamax (Life Technologies), 5 µg/ml insulin (Sigma), 5 µg/ml transferrin (Sigma),
5 ng/ml sodium selenite (Sigma), 10 ng/ml epithelial growth factor (Sigma) and 36 ng/ml
hydrocortisone (Sigma), 100 U/ml penicillin and 100 µg/ml streptomycin. For subculturing,
cells were seeded at 30 % density and harvested or treated at day 3-4 (80-90 % confluence)
for proliferating cells or at day 16 for differentiated cells. Proliferating (day 3-4) and
differentiated (day 16) RPTEC/TERT1 cells reflect radically different phenotypes (Aschauer
et al., 2013). For transwell experiments the cells were seeded at 100 % density on PET
transwell® inserts with 0.4 µm pore size (Corning) and harvested or treated after 10 days
post-initial seeding. Human embryonic kidney cells HEK293 were purchased from the Leibniz
Institute DSMZ – German Collection of Microorganisms and Cell Cultures and cultured in
Dulbecco’s modified Eagle’s medium (DMEM) containing low glucose (1g/l) supplemented
with 10 % fetal bovine serum (FBS) gold, 100 U/ml penicillin and 100 µg/ml streptomycin.
Cells were seeded at a density of 5 x 104 cells/cm2 on poly D-Lysine (Sigma-Aldrich) coated
plates and harvested or subjected to treatment 48 h later.
Primary human renal epithelial cells (pHKC) were isolated from human biopsy material
(Klinikum Konstanz, Konstanz, Germany) from a 65 year old woman as described by O’Brien
et al. (2001). Cells were cultured in DMEM/Ham’s F-12 (Biochrom) supplemented with 1 x
Insulin-Transferrin-Selenium (ITS-G) (Invitrogen), 10 ng/ml epidermal growth factor (Sigma),
18 ng/ml hydrocortisone (Sigma), 4 ng/ml L-thyroxin (Sigma), 100 U/ml penicillin and 100
µg/ml streptomycin. Cells were seeded at a density of 0.5-1 x 106 cells/cm2 and processed
further after 48 h incubation (confluent monolayer).
2.4 Cell treatment for DNA adduct detection
7

Cells were seeded, as mentioned above, in 6 well plates (for DNA extraction), in 96 well plates
(for MTT cell viability assay and NQO1 assay) or in 6 well transwell inserts. At day of exposure
medium was exchanged with fresh medium containing various concentrations of AAI in 0.2 M
NaHCO3 and cells were incubated for additional 48 h at standard conditions.
2.5 Extraction and enzymatic digestion of DNA
DNA from cell and tissue samples was extracted with the PureLinkTM Genomic DNA Mini Kit
(Thermo Fisher Scientific) according to the manufacturer’s protocol. For cellular samples, cells
from one well of a 6 well plate were trypsinized, washed and processed according to
manufacturer’s protocol. For the rat tissue, 40-50 mg tissue were lysed for 1-1.5 h in the
genomic digestion buffer before DNA extraction. DNA was finally eluted in 100 µl nuclease
free water and DNA concentration was assessed via nanodrop.
The enzymatic digestion conditions used for DNA hydrolysis to 2’deoxynucleosides were
adapted from Yun et al. (2012) and Goodenough et al. (2007). DNA (5-10 µg) was diluted to
100 ng/µl H2O and 2 fmol/µl internal standard ([15N5]-dA-AAI) were added prior digestion.
DNase I (2542 U/mL in 0.15 M NaCl; 1016.8 U/mg DNA) was added, and the mixture was
incubated for 1.5 h. Next, nuclease P (Penicillium citrinum; 100 U/mL in 1 mM ZnCl ; 16
1
2
U/mg DNA) was added, and the incubation was further incubated for 3 h at 37 °C. Finally,
alkaline phosphatase (E. coli; 24 U/mL in 1 mM MgCl ; 8 U/mg DNA) and phosphodiesterase
2
I (Crotalus adamanteus venom; 1.7 U/mL in 110 mM Tris-HCl at pH 8.9, containing 110 mM
NaCl, 15 mM MgCl ; 0.2856 U/mg DNA) were added last, and the incubation was continued
2
for additional 18 h at 37 °C. After digestion, samples were evaporated to dryness by vacuum
centrifugation for 2 h and rehydrated in H2O for 2 h. Equal volume of DMSO was added to a
final concentration of 0.125 ng DNA/µl in 50 % DMSO. Samples were stored at -20 °C and
filtered through a 0.2 µm Millex-LG PTFE syringe filter (Millipore) before adduct
quantification.
8

2.6 Synthesis and purification of dA-AL-I standards
The synthesis of dA-AAI and [15N5]-dA-AAI standards were performed by the reaction of AA-
I with dA or [15N5]-dA, using a modification of the protocol described by Yun et al. (2012).
AA-I (1 mg) in DMF (100 µl) were mixed with 20 mg Zn dust (Sigma), pre-activated with 1 %
HCl for 15 min. Then, dA or [15N5]-dA (2 mg) in 50 mM potassium phosphate buffer (1 ml,
pH 5.8) were added and incubated at 37 °C for 16 h in the dark while shaking. Thereafter, the
solution was placed at 4 °C for 30 min to pellet the Zn dust. The supernatant was removed and
the Zn pellet was washed with 2 x 250 µl H2O. The combined supernatants were applied to
Isolute SPE C18(ec) cartridges (500 mg) (Biotage), which were preconditioned with CH3OH
(2 ml), followed by H2O (2 ml). Subsequently, the cartridges were washed with H2O (2 x 2 ml).
The adducts (together with AA-I, AL-I and residual dA) were eluted with CH3OH (2 ml) and
evaporated to dryness by vacuum centrifugation, finally dissolved in DMSO (250 µl).
Purification of standards was performed on a 2695 Alliance Separation Module (Waters) with
a Vydac protein & peptide C18 column (300 Å, 5 µm, 4.6 x 250mm). In the first purification
step dA-AAI and AL-I (tR: 31.2 – 32.7 min) were purified with a gradient starting with 90 %
buffer A (H2O supplemented with 0.1% acetic acid) at a flow rate of 0.6 ml/min. Buffer B
(acetonitrile supplemented with 0.1% acetic acid) was increased to 90% in 30 min and held
constant for 5 min. A second purification step was performed using the same conditions with
methanol instead of acetonitrile to purify dA-AAI (tR: 20.9 – 22.7 min) from AL-I. Final
fractions containing dA-AAI and [15N5]-dA-AAI standards were evaporated to dryness using
vacuum centrifugation and dissolved in DMSO.
2.7 Validation of chemical standard for mass spectrometric analysis
UV/Vis spectra were acquired with an Ultrospec 2100 pro from Amersham Biosciences.
Extinction coefficients were ε305=16,140 l×mol-1×cm-1 and ε415=5,400 l×mol-1×cm-1 (Yun et al.,
9

2012). Product ion scans were acquired at a Xevo TQ-S tandem MS from Waters. High-
resolution MS of each adduct was performed at a Thermo LTQ Orbitrap Discovery in the
Proteomics Facility of the University of Konstanz. NMR spectra were determined with a Bruker
Avance III 600 at the NMR core facility of the University of Konstanz.
2.8 Isotope dilution-UPLC-MS/MS quantification
Analysis was performed on an ACQUITY UPLC H-class coupled to a Xevo TQ-S tandem MS
(Waters). An ACQUITY UPLC Peptide CSH C18 column (130 Å, 1.7 µm, 2.1 × 50 mm,
Waters) was used for chromatographic separation. DNA samples were analyzed with a flow
rate of 0.4 ml/min. Mobile phase A was Millipore water and B was acetonitrile (LC-MS grade,
Carl Roth), both supplemented with 0.1% formic acid (Sigma-Aldrich). Initial conditions were
90 % A. B was increased to 90 % in 8 min and held for 2 min.
The source settings were: Capillary: 0.7 kV, Cone: 11V, Source Offset: 50 V, Source
Temperature: 150 °C, Desolvation Temperature: 500 °C, Cone Gas Flow: 150 l/h, Desolvation
Gas Flow: 1000 l/h, Collision Gas Flow: 0.15 ml/min, Nebuliser Gas Flow: 7 bar. The analyzer
was set to high mass resolution. The analytes were measured with cone voltage of 6 V, collision
energy of 10 V and an auto dwell-time (104 ms) in MRM mode. The transitions [M+H]+ were
dA-AAI 543.3 > 427.2 and [15N]-dA-AAI 548.3 > 432.2. DNA amounts were determined by
UPLC-MS/MS measurements (estimated injection of 0.6 ng DNA) with the same parameters
as described for nucleic acid adducts. The following ion transitions [M+H]+ were analyzed: dG
m/z 268.1>152, dA m/z 252>136, dT m/z 243>127 and dC m/z 228>112.
2.9 Validation of the analytical method
Serial dilution of standard stock solutions was performed to determine the linearity and
sensitivity. LOD (S/N of 3) and LOQ (S/N of 10) was determined. Quality control samples
were prepared in a low, medium and high concentration (LQC, MQC and HQC) of 5, 50 and
10

500 fmol to determine inter-day and intra-day precision (n=5) according to FDA guidelines
(FDA, 1995). Three different sample sets were prepared according to EMA guidelines
(European Medicines Agency, 2011) to define recovery and matrix effects: Standards were
spiked in digested blank matrix (Set 1), standards were spiked in water (Set 2) or standards
were digested in presence of blank matrix (Set 3). Samples were evaporated to dryness using
vacuum centrifugation and re-dissolved in 50 % DMSO. The percentage of recovery was
calculated by the ratio of the peak areas of Set 3 and 1. The ratio of the peak area of Set 1 and
2 determined matrix effects. All sets (1-3) were normalized to internal standards. The influence
of storage was evaluated by short-term storage at RT and 37 °C for 6, 12 and 24 h. The influence
of long-term storage at -20 °C and -80 °C was determined after 2 and 4 weeks. Impact of freeze-
thaw cycles (1-3 times) was also determined. All storage samples were analyzed in triplicates
of all three QC concentrations and normalized to internal standards.
2.10 MTT cell viability assay
Cell viability was determined by MTT reduction assay. Treated cells were incubated with 1 mM
MTT in medium at 37 °C for 1 h. MTT was removed and cells were lysed in 95 % isopropanol
and 5 % formic acid. Formazan absorption was measured at 550ꢀnm using a Tecan M200Pro
microplate reader (Tecan). Values obtained for dead cells (0.1 % Triton X-100) were subtracted
from all other values. Viability data were expressed as percentage of non-treated (control) cells.
2.11 Western Blot analysis
One well of a 6 well plate was used for protein sample preparation for each cell line. Cells were
trypsinized, washed in PBS and lysed with 200-300 µl RIPA buffer (50 mM Tris-HCl, 150 mM
NaCl, 0.1 % SDS, 0.5 % sodium deoxycholate, 1 % Triton X-100) for 10 min on ice. After
passage through a 26G needle, protein concentrations were determined using Pierce BCA
Protein Assay (Thermo Fisher Scientific) and samples for SDS Page were generated by addition
11

of 1.5 X Urea sample buffer (93.75 mM Tris-HCl pH=6.8, 9 M Urea, 7.5 % (v/v) β-mercapto-
ethanol, 15 % (v/v) glycerol, 3 % (w/v) SDS, 0.01 % (w/v) bromphenol blue), boiled at 95 °C
for 5 min. and stored at -80 °C. Subsequently, 4.5 µg total protein/sample were separated by a
12 % SDS-PAGE and blotted onto a nitrocellulose membrane. Membranes were detected either
with Ponceau S and then blocked with 5 % (w/v) milk powder in TTBS (100 mM Tris-HCl,
150 mM NaCl, 0.1 % (v/v) Tween 20, pH 7.6) for 1 h at RT. Proteins of interest were detected
either with anti-NQO1 antibody (1:1000 in 5 % milk, Thermo Fisher Scientific, #MA1-16672)
or with anti-β-actin (1:4.000 in 5 % milk, Sigma, #A2228) over night at 4 °C. After washing,
membranes were incubated for 1 h at RT with peroxidase-conjugated secondary antibody anti-
mouse IgG-peroxidase (1:80.000 in 5 % milk, Sigma-Aldrich, #A9044). Signals were detected
by enhanced chemiluminescence (Lumigen® ECL ultra, Lumigen) via ImageQuantTM LAS
4000 (GE Healthcare). Stripping of the membrane was performed with 0.2 M NaOH for 10 min
followed by incubation for 10 min in Millipore H2O.
2.12 NQO1 activity assay
NQO1 activity assay was performed using a modified protocol of the “Prochaska” microtiter
plate bioassay described by Fahey et al. (2004). For determination of NQO1 induction after
AAI exposure, cells were treated with solvent control, AAI in 0.2 M NaHCO3 or 5 µM
sulforaphane in DMSO for 48 h. Cells were washed with PBS and lysed by incubation with
50 µl 0.8 % digitonin in 2 mM EDTA (pH 7.8) at 37 °C for 10 min followed by an additional
incubation at 25 °C for 10 min while shaking. For transwell inserts cells were detached by
trypsinization, washed and lysed in 750 µl 0.8 % digitonin in 2 mM EDTA (pH 7.8). Briefly,
10 µl cell lysates were transferred into a 96 well plate and mixed with 200 µl of the reaction
mixture containing 25 mM Tris-HCl (pH 7.4), 0.066 % bovine serum albumin, 0,01 % Tween-
20, 7.2 µM MTT, 1 mM NADPH (freshly added) and 50 nM menadione (freshly added).
Control reactions were performed by addition of 50 µl 0.3 mM dicoumarol in 0.5 % DMSO
12

and 5 mM potassium phosphate (pH 7.4). Blank reactions were performed w/o menadione in
the reaction buffer. Absorption was quantified at 610 nm every min for 30 min in total using a
Tecan M200Pro microplate reader. The specific NQO1 activity was quantified using the slope
of the linear range of absorption increase. Values were normalized to total protein amount,
quantified by PierceTM BCA Protein Assay Kit.
2.13 Statistical analysis
Unless otherwise indicated data are presented as meanꢀ±ꢀSEM. Sample size (n) indicates the
number of independent experiments performed. Statistical significance was determined using
GraphPad Prism 7, statistical tests as indicated in the figure legends.
2. Results
3.1 Performance of the UPLC-MS/MS detection method of dA-AAI adducts
AAI derived DNA adducts are considered relevant and reliable biomarkers for AAI exposure,
whereby dA-AAI (Fig. 1) represents the predominant adduct found in animal and patient
material. As isotope-dilution LC-MS/MS is considered the gold standard for quantification of
DNA adducts, dA-AAI and [15N5]-dA-AAI standards were synthesized with a modified
biomimetic reaction described earlier by Yun et al. (2012). Pre-purification of standards was
performed by solid phase extraction (SPE). Final purification was achieved via HPLC with an
approximate yield of ~1-1.5 % compared to the AAI starting material. dA-AAI and [15N5]-dA-
AAI standards were verified by their characteristic UV/Vis spectra (Fig. S-1), product ion scans
(Fig. 2 a and b), high resolution MS (Fig. S-2), 1H-NMR (Fig. S-3 and S-4) and 2D-NMR (Fig.
S-5). Product ion scans resulted in detection of the aglycone adducts ([BH2]+ of dA-AAI m/z
427 and for [BH2]+ of [15N5]-dA-AAI m/z 432) (Fig. 2 c), which were subsequently used for
quantification with ID-UPLC-MS/MS. The sensitivity was determined at a limit of detection
13

(LOD, S/N of 3) of 0.1 fmol and a limit of quantification (LOQ, S/N of 10) of 0.5 fmol (Table
1). Serial dilutions of unlabeled standard, spiked with internal isotopically labeled standard
were used to prepare a calibration curve (Fig. 2 d and Table 1) with a linear range of at least
three orders of magnitude from 0.5 fmol to 1000 fmol (R2 = 0.9926). Representative
chromatograms of treated and control samples are shown in Fig. 2 e, f. Peak areas for dA-AAI
and [15N5]-dA-AAI were analyzed at a retention time of 4.2 minutes.
According to the EMA guidelines (European Medicines Agency, 2011), recovery and matrix
effects were determined (Table 2). Recovery rates were determined by comparing standards,
which were digested together with blank matrix, to standards which were spiked to separately
digested blank matrix. Matrix effects were determined by standards spiked to digested blank
matrix compared to standards spiked into water. All values were found to be within the
acceptable limits as suggested by EMA (European Medicines Agency, 2011). Quality
controls (QC) were prepared over a low (LQC), medium (MQC), and high (HQC)
concentration range of 5, 50 and 500 fmol. The values of percent relative standard deviation
(% RSD), which is a measure for the precision of a method, were <ꢀ20 % for LQCs and <ꢀ15 %
for MQCs and HQCs for samples measured on the same day (intra-day precision) and on
different days (inter-day precision). These values were fully compatible with FDA guidelines
for the validation of analytical procedures (FDA, 1995). Potential analyte loss due to storage
or repeated freeze-thaw cycles was determined (Table 3). Short-term storage at ambient
temperatures (RT and 37 °C), as well as long-term storage at lower temperatures (-20 °C and -
80°C) had no influence on measurement of the analyte. Taken together a reliable, robust,
sensitive and accurate ID-UPLC-MS/MS method was established to quantify dA-AAI adducts
from cells exposed in vitro.
3.2 UPLC-MS/MS analysis of dA-AAI adduct formation in fresh frozen rat tissue
14

We analyzed dA-AAI adduct formation in kidney cortex (Fig. 3 a, b) and bladder (Fig. 3 c, d)
tissue from AAI/II exposed male and female Eker rats (10 mg / kg bodyweight per day). We
were able to detect dA-AAI adduct formation already 24 h after oral exposure and observed a
time-dependent accumulation of DNA adducts in the kidney cortex and in bladder. Based on
the adduct curves in renal cortex, it appeared that male Eker rats had greater AAI activation and
thus DNA adduct formation than their female counterparts. Additionally, male Eker rats
displayed a severe reduction in bodyweight (Fig. 3 e), whereas AAI/II treated female Eker rats
showed only a decrease in bodyweight gain compared to the control group (Fig. 3 f). In contrast
to the renal cortex, DNA adduct levels in the bladder were approximately ten times lower and
showed no sex differences.
3.3 Comparison of dA-AAI adduct formation in human primary and immortalized renal cells
Based on the verification of dA-AAI adduct detection in rat tissue, we used our ID-UPLC-
MS/MS protocol to quantify DNA adduct formation in two different human renal cell lines
(HEK293 and RPTEC/TERT1) and primary human kidney cells (pHKC). Cells were exposed
to different AAI concentrations (1-100 µM) or solvent control for 48 h. We observed a
concentration dependent increase in dA-AAI adduct levels in all cells (Fig. 4 a). However, DNA
adduct levels appeared to reach saturation in pHKC already at very low AAI exposure
concentrations, whereas this was not observable in differentiated RPTEC/TERT1 or in HEK293
cells. The comparison of pHKC and HEK293 cells provided for an approximately 100-fold
difference in DNA-adducts quantified. When considering the 3 x 109 base pairs per diploid
genome in a human cell, the DNA adducts detected can be expressed as DNA adducts per cell.
Thus, the “saturation” in DNA adducts as observed is apparently either the result of cell type
specific saturable enzymatic activation capabilities and/or a saturation of available dA and dG
sites that are “adductable”. Alternatively, different cell lines may tolerate different adduct levels
at which cell death occurs (see below).
15

To assess the impact of cell differentiation on dA-AAI adduct formation, proliferating (day 3)
and differentiated (day 16) RPTEC/TERT1 cells on 2D wells and RPTEC/TERT1 cells grown
on transwell-inserts (day 10) were exposed to the same AAI concentration range (Fig. 4 b).
Adduct levels increased rapidly in differentiated compared to proliferating RPTEC/TERT1,
while proliferating RPTEC/TERT1 and RPTEC/TERT1 cultured on transwell inserts showed a
similar concentration response in adduct levels.
3.4 Influence of AAI exposure on cell viability
The effect of AAI on cell viability, as assessed via MTT reduction assay after 48 h exposure to
various AAI concentrations (Fig. 4 c), demonstrated an AAI concentration dependent loss of
cell viability, but also large differences in susceptibility amongst the three different cell types.
pHKC were most sensitive to AAI, showing a loss of viability already at 1 µM and a near
maximal loss of viability 10 µM AAI, i.e. the same concentration where DNA adduct levels
reached near saturation levels (Fig. 4 a). Conversely, RPTEC/TERT1 cells were much more
resistant to AAI, where demonstrable reduction of viability (70 % compared to ctrl.) was
observable as of 100 µM AAI, despite that DNA adducts were 5-fold higher than in pHKC at
1 µM with similar reduction in viability (Fig 4 d). HEK293 cell viability was similar to the one
observed for differentiated RPTEC/TERT1 cells, albeit at higher AAI concentrations
(>100 µM) HEK293 cells appeared to be more sensitive than RPTEC/TERT1 cells (Fig. 4 c).
The comparison of cell viability with DNA adducts on a cell type basis (Fig. 4 d) demonstrated
that a 5x and 25x lower number of DNA adducts were required in HEK293 cells to reduce cell
viability than was the case for differentiated pHKC and RPTEC/TERT1 cells, respectively.
Obviously, issues e.g. expression of organic anion transporter 1 (OAT1), known to be important
for AAI uptake (Dickman et al., 2011, Xue et al., 2011), as well as level of expression of phase
II enzymes (e.g. glutathione transferase (GST)) and levels of cytosolic glutathione, but not of
phase I enzymes (e.g. NQO1) will play a major role.
16

3.5 Comparison of NQO1 expression and activity in HEK293, RPTEC/TERT1 and pHKC
As noted earlier NQO1 is considered central in the metabolic activation of AAI and thus the
formation of DNA adducts. Accordingly, we examined the expression and the relative activity
of NQO1 in the three different cell types employed. As demonstrated (Fig. 5 a), HEK293 cells
do not express NQO1, and only a small amount of NQO1 can be detected for the differentiated
pHKC cells. Differentiation and/or the culturing technology appeared not to play a major role
with regard to NQO1 expression in RPTEC/TERT1, as demonstrated when NQO1 expression
levels were normalized to the housekeeping gene β-actin (Fig. 5 b). However, RPTEC/TERT1
cells displayed a significantly higher level of expression than pHKC cells. As level of
expression bears no information as to whether or not NQO1 is enzymatically active, NQO1
activity (Fig. 5 c) was assessed using a modified version of the “Prochaska” microtiter assay
(Fahey et al., 2004) or normalized to NQO1 expression levels (Fig. 5 d). In sum, NQO1
expression and activity was much higher in RPETC/TERT1 cells compared to pHKC cells,
while expression levels and activity of NQO1 in HEK293 were not significantly distinguishable
from background.
3.6 Lack of AAI mediated NQO1 induction in HEK293, RPTEC/TERT1 and pHKC
Currently, it is assumed that NQO1 can be induced by AAI exposure (Bárta et al., 2014),
suggesting that exposure to AAI would induce its own bioactivation in situ. To address the
latter, we determined NQO1 activity subsequent to 48 h exposure to AAI (1-100 µM), solvent
control, or 5 µM sulforaphane. Sulforaphane, a broccoli-derived phytochemical, was
demonstrated to activate Nrf2 and thus increased expression of NQO1 (Houghton et al., 2016).
Contrary to previous reports, AAI exposure did not induce an increased expression of NQO1
in any of the cells tested. However and as expected, sulforaphane treatment resulted in a
significant increase in NQO1 activity in differentiated RPTEC/TERT1 cells (Fig. 6 a) and
17

pHKC (Fig. 6 b), but not in HEK293 (Fig. 6 c) cells, thus corroborating the absence of
constitutive and inducible NQO1 expression in HEK293 cells.
3. Discussion
Although 32P-postlabeling is a commonly used technique for investigating AA mediated
formation of DNA adduct formation in experimental animal models (Dong et al., 2006) and
patients with nephropathy (Yun et al., 2012, Bieler et al., 1997), 32P-postlabeling encompasses
several limitations, e.g. variable labeling efficiencies, equivocal identities of carcinogen
adducts, semi-quantitative assessment, low through-put etc. Similarly, the use of antibodies for
detection of dA-AL I adducts (Chang et al., 2017) via secondary antibodies labeled with fluoro-
or chromophores, have similar limitations, e.g. sensitivity, semi-quantitative assessment, low
through-put and, most importantly, yet still lacks direct structural validation of detection using
DNA adduct quantification with stable, isotopically labelled internal standards. Modern
techniques e.g. LC-MS/MS represents a more reliable method to quantify AA derived DNA
adducts (Yun et al., 2012) that is also applicable to difficult samples such as paraffin-embedded
tissue and implementation for high through-put (Yun et al., 2017). As to date AAI mediated
DNA-adducts were not determined in the in vitro settings, we adapted the previously published
protocol by Yun et al. (2012, 2017) to our settings. By employing a de novo synthetized internal
standard for adducts, i.e. ion scan, high resolution MS, NMR and UV/Vis spectra characterized
unlabeled and stable isotope labeled ([15N5]-) dA-AAI (Fig. S-1 – S-5), we established a highly
sensitive ID-UPLC-MS/MS method. The use of the internal standard allows detection of sample
loss during preparation and variances during measurements, resulting in reduced variability in
the data. This highly sensitive ID-UPLC-MS/MS method was validated with regard to linearity,
recovery, matrix effects, precision, and storage conditions according to FDA and EMA
guidelines and was demonstrated to allow quantification of AAI mediated DNA-adducts in rat
18

renal and bladder tissues as well as in in vitro exposed HEK293, proliferating and differentiated
RPTEC/TERT1 and pHKC cells. The comparison of our ID-UPLC-MS/MS method with the
previously published approach by Yun et al. (2012, 2017), reporting an LOQ of 0.3 dA-AAI
adducts per 108 DNA bases which corresponds to 0.0917 fmol, demonstrated similar sensitivity
(LOD 0.1 fmol; LOQ: 0.5 fmol).
To evaluate our established method for dA-AAI quantification we analyzed the dA-AAI adduct
levels in Eker rats, orally exposed (1, 3, 7 or 14 days) to a mixture of AAI and II (10 mg / kg
bodyweight). In line with the published persistence of adducts in patient (Stiborová et al.,
2017) due to their ability to evade global genome repair (Sidorenko et al., 2012), we
observed a time dependent accumulation of dA-AAI adducts in rat renal cortex and bladder.
Adduct levels in renal cortex were ten times higher than in bladder tissue. Previous studies
reported that renal cortex from the very same rats did not display necrosis or increased cell
proliferation, but upregulation of NQO1 and several p53 dependent pathway genes, suggestive
of a DNA damage response (Stemmer et al., 2007). Furthermore, our data confirmed dA-AAI
adduct levels reported by Yun et al. (2012) in the kidney (1020 dA-AAI adducts/108 bases) for
mice treated with 1 mg/kg (unknown duration) (Yun et al., 2012). Unfortunately, no reference
values are available for dA-AAI adduct levels detected by UPLC-MS/MS in the bladder.
As we confirmed the applicability of our dA-AAI quantification method, we subsequently
investigated AAI toxicity in different human renal cell models. We compared the data for dA-
AAI adduct levels, cell viability as well as NQO1 expression and activity, hypothesized to be
required for AAI bioactivation, in pHKC, HEK293 and RPTEC/TERT1 cells. Although dA-
AAI adduct levels were significantly higher in differentiated RPTEC/TERT1 cells compared to
HEK293 cells, we did not observe a difference in cell viability between the cell lines.
Expression of hTERT can induce resistance to DNA-damaging response by suppression of p53-
activation (Jin et al., 2010). However, the response of RPTEC/TERT1 to the DNA damaging
substance B[a]P were deemed to be consistent with what is known regarding this cell type in a
19

normal, healthy kidney, including significant gene expression changes and induced BPDE-
DNA adducts (Simon-Friedt et al., 2015). Moreover, we were able to detect dA-AAI adduct
levels in RPTEC/TERT1 cells over the whole AAI-concentration range. The adduct levels were
similar to levels found in vivo in rat tissue in our study and also in human patient material (Yun
et al., 2012), indicating that the RPTEC/TERT1 cell systems represents a suitable model to
investigate dA-AAI adduct formation in more detail. Interestingly, reduction of cell viability in
the cell lines was not correlated to dA-AA adduct levels, suggesting either a cell-line specific
bearable threshold in DNA-adducts, at which cells succumb to the damage, or DNA-
independent toxicity of AAI. Substantial differences in activity of the involved DNA-repair
pathway, i.e. nucleotide excision repair, can be ruled out as both RPTEC/TERT1 (Simon-Friedt
et al., 2015) as well as HEK293 (Atanassov et al., 2004, Toga et al., 2014) are NER competent.
Primary cells possess a similar gene expression profile to the originate cell type and present
with, at least in the early passages, a relevant expression of organic anion transporters (OAT).
OAT1 and OAT3 are major uptake transporters for AAI (Dickman et al., 2011,Xue et al., 2011)
in proximal tubular epithelial cells and thus provide higher intracellular AAI levels. As
expected, we determined the highest dA-AAI adduct levels and a most severely reduced cell
viability in pHKC. Although we could not confirm OAT expression in our RPTEC/TERT1 cell
model due to unavailability of an appropriate OAT antibody, Aschauer et al. (2014)
demonstrated that the expression of organic anion transporters (OAT), mainly OAT1 and
OAT3, are increased in differentiated RPTEC/TERT1 cells. Consistent with the data reported
by Aschauer et al. (2014), we observed that dA-AAI adduct levels were higher in differentiated
than in proliferating RPTEC/TERT1 cells.
Metabolic activation of AAI is required for the DNA adduct formation. It has been
demonstrated that NQO1 is able to bioactivate AAI in cell free systems (Stiborova et al., 2011)
and its role in vivo is also strongly suggested (Bárta et al., 2014). Indeed, Arlt et al. (Arlt et al.,
2004) speculated that “the detection of dA-AAI adducts in several organs of patients suggests
20

the presence of AA activating enzymes in these organs and tissues and that AA and/or its
metabolites are distributed via the bloodstream to other organs”. Although the potential
involvement of phase II metabolic enzymes as sulfotransferases (SULTs) appeared inconsistent
(Meinl et al., 2006, Stiborova et al., 2011), nitroreduction followed by sulfation via SULT1B1
was demonstrated to further increase the mutagenic and cytotoxic potential of AAI (Sidorenko
et al., 2014). In support of the latter, the most recent findings using primary human hepatocytes
and renal proximal tubule epithelial cells in an organ-on-a-chip microphysiological system
(Chang et al., 2017), reported aristolactam-N-sulfated ester (AL-I N-OSO3) as the principle
intermediate that would be formed in the hepatocytes and transported to the renal cortex to
induce overt cytotoxicity and presumably formation of dA-AAI adducts. Based on the previous
findings of Sidorenko et al. (2014) and Chang et al. (2017), demonstrating that hepatic phase II
metabolism of the less reactive N-hydroxyaristolactam (AL-I-NOH) can lead to the formation
of highly cytotoxic AL-I-N-OSO3, immediate hepatoxicity would be expected either in the
primary hepatocytes employed in the organ-on-a-chip application or in the patients affected.
However, in neither of the exposure scenarios hepatotoxicity was reported (Yun et al., 2012,
Arlt et al., 2004, Nortier et al., 2003) or shown in Chang et al. (2017). Thus, it appears
questionable whether highly reactive metabolites can be metabolized and excreted from the
liver without affecting the hepatic cells themselves, while entering the blood stream and
exerting the toxic response in the kidney only. Rather, these findings suggest that metabolic
activation of AAI primarily occurs in the renal cortex proper.
We investigated NQO1 expression in the human renal primary cells and cell lines and could
not detect NQO1 expression or activity in HEK293, which correlated with the very low dA-
AAI adduct formation. RPTEC/TERT1 cells expressed NQO1, showed high enzyme activity
and also high adduct levels. In contrast, NQO1 expression and activity was low in pHKC, but
dA-AAI adduct levels were highest of all human renal cells tested, even in comparison to
21

differentiated RPTEC/TERT1 at high AAI concentrations. Thus, while NQO1 activity followed
expression levels, adduct formation did not. Continuous passaging of pHKC suggested
increasing NQO1 activity (Fig S-6), which could be explained by an adaptive counteraction
against oxidative stress (Dinkova-Kostova and Talalay, 2000). The observed variation of
NQO1 in different passage numbers of pHKC demonstrated the highly variable phenotype
of this cell model. AAI was not able to upregulate NQO1 activity in any of the applied renal
cell systems, in contrast to the positive control sulforaphane, as expected (Brooks et al., 2001),
but not in HEK293 devoid of NQO1. Consequently, we questioned the role of NQO1 as the
only bioactivator of AAI in human renal cells. Indeed, it has been shown that also cytochrome
P450 enzymes are able to activate AAI (Levová et al., 2011). Stiborova et al. (2012)
demonstrated that human and rat CYPs detoxified AAI to 8-hydroxyaristolochic acid I (AAIa)
via O-demethylation of the methoxy group, potentially resulting in a decreased renal toxicity
(Xiao et al., 2008). However, the latter work focused mainly on CYP1a1/2, isoforms that are
not expressed in human kidney (Knights et al., 2013). Other isoforms, such as human and rat
CYPs of the 2C subfamily and human CYP3A4/5, 2D6, 2E1, and 1B1 also form AAIa but with
much lower efficiency (Stiborová et al., 2015). Accordingly, while the role of CYP isoforms in
AAI bioactivation is still under debate, CYPs may play a role especially in HEK293 cells
showing absence of NQO1 expression. CYP mediated activation/detoxification highly depends
on the aerobic conditions: anaerobic conditions promote AAI activation, while detoxifying O-
demethylation is catalyzed at aerobic conditions, summarized by Stiborová et al. (2015).
Further investigations should therefore focus on the impact of the oxygen level on AAI DNA
adduct formation and AAI induced toxicity. Indeed, cultivation at physiologically normal
oxygen conditions for hepatic as well as renal cells, i.e. approx. 5% O2 for the hepatocytes
(Carreau et al., 2011) and 5-7% O2 for the renal proximal tubule epithelial cells (Schiffer and
Friederich-Persson, 2017), would provide for a much more realistic insight as to the
22

predominant pathways of AA metabolism and thus AAI DNA adduct formation and AAI
induced toxicity.
In summary, we demonstrated a sensitive and reliable UPLC-MS/MS method to quantify dA-
AAI adduct in DNA from human renal cell systems and tissue samples. We showed that
RPTEC/TERT1 cells represented an appropriate model to investigate mechanism(s) underlying
AAI-mediated DNA adduct formation in humans. In contrast, HEK293 may serve as excellent
negative control to elucidate specific steps in metabolism of AAI in renal cells.
Acknowledgement
The authors thank Sabine Drewitz for technical support with the cell culture and Western Blot
analysis. We thank the NMR Core facility at the University of Konstanz, especially Anke
Friemel for the NMR data and Prof. Dr. Tanja Gaich for helpful assistance in the interpretation.
We also thank the Proteomics Core facility at the University of Konstanz for the high-resolution
MS data. Animal experiments were supported by the Federal Ministry of Education and
Research (BMBF: 0313024). Mass spectrometric measurements were performed on an
instrument funded by the German Research Foundation and the federal state of Baden-
Wuerttemberg (DFG: INST 38/537-1). TZ was supported by a fellowship of the DFG-funded
Graduate School ‘Konstanz Research School Chemical Biology (KoRS-CB, GSC 218).
23

References
Arlt, V.M., Alunni-Perret, V., Quatrehomme, G., Ohayon, P., Albano, L., Gaid, H., Michiels,
J.F., Meyrier, A., Cassuto, E., Wiessler, M., Schmeiser, H.H., Cosyns, J.P., 2004.
Aristolochic acid (AA)-DNA adduct as marker of AA exposure and risk factor for AA
nephropathy-associated cancer. Int J Cancer 111, 977–980.
https://doi.org/10.1002/ijc.20316
Aschauer, L., Carta, G., Vogelsang, N., Schlatter, E., Jennings, P., 2014. Expression of
xenobiotic transporters in the human renal proximal tubule cell line RPTEC/TERT1.
Toxicol. In Vitro. https://doi.org/10.1016/j.tiv.2014.12.003
Aschauer, L., Gruber, L.N., Pfaller, W., Limonciel, A., Athersuch, T.J., Cavill, R., Khan, A.,
Gstraunthaler, G., Grillari, J., Grillari, R., Hewitt, P., Leonard, M.O., Wilmes, A.,
Jennings, P., 2013. Delineation of the Key Aspects in the Regulation of Epithelial
Monolayer Formation. Mol. Cell. Biol. 33, 2535–2550.
https://doi.org/10.1128/MCB.01435-12
Atanassov, B., Velkova, A., Mladenov, E., Anachkova, B., Russev, G., 2004. Comparison of
the global genomic and transcription-coupled repair rates of different lesions in human
cells. Z. Naturforsch. C. 59, 445–453.
Bárta, F., Levová, K., Frei, E., Schmeiser, H.H., Arlt, V.M., Stiborová, M., 2014. The effect
of aristolochic acid I on expression of NAD (P) H: quinone oxidoreductase in mice and
rats—A comparative study. Mutat. Res. Toxicol. Environ. Mutagen. 768, 1–7.
https://doi.org/http://dx.doi.org/10.1016/j.mrgentox.2014.01.012
Bieler, C.A., Stiborova, M., Wiessler, M., Cosyns, J.P., van Ypersele de Strihou, C.,
Schmeiser, H.H., 1997. 32P-post-labelling analysis of DNA adducts formed by
aristolochic acid in tissues from patients with Chinese herbs nephropathy.
Carcinogenesis 18, 1063–1067.
24

Brooks, J.D., Paton, V.G., Vidanes, G., 2001. Potent induction of phase 2 enzymes in human
prostate cells by sulforaphane. Cancer Epidemiol. Biomarkers Prev. 10, 949–954.
Carreau, A., El Hafny-Rahbi, B., Matejuk, A., Grillon, C., Kieda, C., 2011. Why is the partial
oxygen pressure of human tissues a crucial parameter? Small molecules and hypoxia. J.
Cell. Mol. Med. 15, 1239–1253. https://doi.org/10.1111/j.1582-4934.2011.01258.x
Chang, S.-Y., Weber, E.J., Sidorenko, V.S., Chapron, A., Yeung, C.K., Gao, C., Mao, Q.,
Shen, D., Wang, J., Rosenquist, T.A., 2017. Human liver-kidney model elucidates the
mechanisms of aristolochic acid nephrotoxicity. JCI insight 2.
Cosyns, J.-P., 2003. Aristolochic acid and “Chinese herbs nephropathy”: a review of the
evidence to date. Drug Saf. 26, 33–48. https://doi.org/10.2165/00002018-200326010-
00004
Debelle, F.D., Vanherweghem, J.-L., Nortier, J.L., 2008. Aristolochic acid nephropathy: a
worldwide problem. Kidney Int. 74, 158–169.
https://doi.org/https://doi.org/10.1038/ki.2008.129
Depierreux, M., Van Damme, B., Vanden Houte, K., Vanherweghem, J.L., 1994. Pathologic
aspects of a newly described nephropathy related to the prolonged use of Chinese herbs.
Am J Kidney Dis 24, 172–180. https://doi.org/S0272638694001484 [pii]
Dickman, K.G., Sweet, D.H., Bonala, R., Ray, T., Wu, A., 2011. Physiological and Molecular
Characterization of Aristolochic Acid Transport by the Kidney. J. Pharmacol. Exp. Ther.
338, 588–597. https://doi.org/10.1124/jpet.111.180984
Dinkova-Kostova, A.T., Talalay, P., 2000. Persuasive evidence that quinone reductase type 1
(DT diaphorase) protects cells against the toxicity of electrophiles and reactive forms of
oxygen. Free Radic. Biol. Med. 29, 231–240.
https://doi.org/https://doi.org/10.1016/S0891-5849(00)00300-2
Dong, H., Suzuki, N., Torres, M.C., Bonala, R.R., Johnson, F., Grollman, A.P., Shibutani, S.,
2006. Quantitative determination of aristolochic acid-derived DNA adducts in rats using
25

32P-postlabeling/polyacrylamide gel electrophoresis analysis. Drug Metab Dispos 34,
1122–1127. https://doi.org/dmd.105.008706 [pii]10.1124/dmd.105.008706 [doi]
European Medicines Agency, 2011. Guideline on bioanalytical method validation. Comm.
Med. Prod. Hum. Use.
Fahey, J.W., Dinkova-Kostova, A.T., Stephenson, K.K., Talalay, P., 2004. The “Prochaska”
microtiter plate bioassay for inducers of NQO1, in: Methods in Enzymology. Elsevier,
pp. 243–258. https://doi.org/https://doi.org/10.1016/S0076-6879(04)82014-7
FDA, U.S., 1995. Guideline for industry: text on validation of analytical procedures: ICH
Q2A. Rockville, MD Mar 1, 995.
Goodenough, A.K., Schut, H.A.J., Turesky, R.J., 2007. Novel LC-ESI/MS/MSn Method for
the Characterization and Quantification of 2 ‘-Deoxyguanosine Adducts of the Dietary
Carcinogen 2-Amino-1-methyl-6-phenylimidazo [4, 5-b] pyridine by 2-D Linear
Quadrupole Ion Trap Mass Spectrometry. Chem. Res. Toxicol. 20, 263–276.
https://doi.org/10.1021/tx0601713
Houghton, C.A., Fassett, R.G., Coombes, J.S., 2016. Sulforaphane and Other Nutrigenomic
Nrf2 Activators: Can the Clinician’s Expectation Be Matched by the Reality? Oxid.
Med. Cell. Longev. 2016, 7857186. https://doi.org/10.1155/2016/7857186
Jadot, I., Declèves, A.-E., Nortier, J., Caron, N., 2017. An Integrated View of Aristolochic
Acid Nephropathy: Update of the Literature. Int. J. Mol. Sci. 18, 297.
https://doi.org/10.3390/ijms18020297
Jelakovi, Cacute, B., Karanovi, Cacute, S., Vukovi, Cacute, I., Lela, F.M., Edwards, K.L.,
Nikoli, Cacute, J., Tomi, Cacute, K., Slade, N., Brdar, B., Turesky, R.J., 2011.
Aristolactam-DNA adducts are a biomarker of environmental exposure to aristolochic
acid. Kidney Int. 81, 559–567. https://doi.org/10.1038/ki.2011.371
Jin, X., Beck, S., Sohn, Y.-W., Kim, J.-K., Kim, S.-H., Yin, J., Pian, X., Kim, S.-C., Choi, Y.-
J., Kim, H., 2010. Human telomerase catalytic subunit (hTERT) suppresses p53-
26

mediated anti-apoptotic response via induction of basic fibroblast growth factor. Exp.
Mol. Med. 42, 574–582. https://doi.org/10.3858/emm.2010.42.8.058
Knights, K.M., Rowland, A., Miners, J.O., 2013. Renal drug metabolism in humans: the
potential for drug-endobiotic interactions involving cytochrome P450 (CYP) and UDP-
glucuronosyltransferase (UGT). Br. J. Clin. Pharmacol. 76, 587–602.
https://doi.org/10.1111/bcp.12086
Levová, K., Moserová, M., Kotrbová, V., Šulc, M., Henderson, C.J., Wolf, C.R., Phillips,
D.H., Frei, E., Schmeiser, H.H., Mareš, J., Arlt, V.M., Stiborová, M., 2011. Role of
cytochromes P450 1A1/2 in detoxication and activation of carcinogenic aristolochic acid
I: studies with the hepatic NADPH: cytochrome P450 reductase null (HRN) mouse
model. Toxicol. Sci. 121, 43–56. https://doi.org/https://doi.org/10.1093/toxsci/kfr050
Meinl, W., Pabel, U., Osterloh‐Quiroz, M., Hengstler, J.G., Glatt, H., 2006. Human
sulphotransferases are involved in the activation of aristolochic acids and are expressed
in renal target tissue. Int. J. cancer 118, 1090–1097.
https://doi.org/https://doi.org/10.1002/ijc.21480
Nortier, J.L., Martinez, M.C., Schmeiser, H.H., Arlt, V.M., Bieler, C.A., Petein, M.,
Depierreux, M.F., De Pauw, L., Abramowicz, D., Vereerstraeten, P., Vanherweghem,
J.L., 2000. Urothelial carcinoma associated with the use of a Chinese herb (Aristolochia
fangchi). N Engl J Med 342, 1686–1692.
https://doi.org/10.1056/NEJM200006083422301 [doi]
Nortier, J.L., Schmeiser, H.H., Muniz Martinez, M.C., Arlt, V.M., Vervaet, C., Garbar, C.H.,
Daelemans, P., Vanherweghem, J.L., 2003. Invasive urothelial carcinoma after exposure
to Chinese herbal medicine containing aristolochic acid may occur without severe renal
failure. Nephrol Dial Transpl. 18, 426–428.
O’Brien, E., Heussner, A.H., Dietrich, D.R., 2001. Species-, sex-, and cell type-specific
effects of ochratoxin A and B. Toxicol. Sci. 63, 256–264.
27

Pfau, W., Schmeiser, H.H., Wiessler, M., 1990. Aristolochic acid binds covalently to the
exocyclic amino group of purine nucleotides in DNA. Carcinogenesis 11, 313–319.
https://doi.org/https://doi.org/10.1093/carcin/11.2.313
Schiffer, T.A., Friederich-Persson, M., 2017. Mitochondrial Reactive Oxygen Species and
Kidney Hypoxia in the Development of Diabetic Nephropathy. Front. Physiol. 8, 211.
https://doi.org/10.3389/fphys.2017.00211
Sidorenko, V.S., Attaluri, S., Zaitseva, I., Iden, C.R., Dickman, K.G., Johnson, F., Grollman,
A.P., 2014. Bioactivation of the human carcinogen aristolochic acid. Carcinogenesis 35,
1814–1822. https://doi.org/https://doi.org/10.1093/carcin/bgu095
Sidorenko, V.S., Yeo, J.E., Bonala, R.R., Johnson, F., Scharer, O.D., Grollman, A.P., 2012.
Lack of recognition by global-genome nucleotide excision repair accounts for the high
mutagenicity and persistence of aristolactam-DNA adducts. Nucleic Acids Res 40,
2494–2505. https://doi.org/10.1093/nar/gkr1095
Simon-Friedt, B.R., Wilson, M.J., Blake, D.A., Yu, H., Eriksson, Y., Wickliffe, J.K., 2015.
The RPTEC/TERT1 cell line as an improved tool for in vitro nephrotoxicity
assessments. Biol. Trace Elem. Res. 166, 66–71. https://doi.org/10.1007/s12011-015-
0339-y
Stemmer, K., Ellinger-Ziegelbauer, H., Ahr, H.J., Dietrich, D.R., 2007. Carcinogen-specific
gene expression profiles in short-term treated Eker and wild-type rats indicative of
pathways involved in renal tumorigenesis. Cancer Res 67, 4052–4068.
https://doi.org/67/9/4052 [pii]10.1158/0008-5472.CAN-06-3587
Stiborová, M., Arlt, V.M., Schmeiser, H.H., 2017. DNA Adducts Formed by Aristolochic
Acid Are Unique Biomarkers of Exposure and Explain the Initiation Phase of Upper
Urothelial Cancer. Int. J. Mol. Sci. 18, 2144. https://doi.org/10.3390/ijms18102144
Stiborová, M., Bárta, F., Levová, K., Hodek, P., Schmeiser, H.H., Arlt, V.M., Martínek, V.,
2015. A Mechanism of O-Demethylation of Aristolochic Acid I by Cytochromes P450
28

and Their Contributions to This Reaction in Human and Rat Livers: Experimental and
Theoretical Approaches. Int. J. Mol. Sci. 16, 27561–27575.
https://doi.org/10.3390/ijms161126047
Stiborova, M., Frei, E., Arlt, V.M., Schmeiser, H.H., 2009. The role of biotransformation
enzymes in the development of renal injury and urothelial cancer caused by aristolochic
acid: urgent questions and difficult answers. Biomed. Pap. Med. Fac. Palacky Univ.
Olomouc 153. https://doi.org/DOI: 10.5507/bp.2009.001
Stiborova, M., Frei, E., Hodek, P., Wiessler, M., Schmeiser, H.H., 2005. Human hepatic and
renal microsomes, cytochromes P450 1A1/2, NADPH:cytochrome P450 reductase and
prostaglandin H synthase mediate the formation of aristolochic acid-DNA adducts found
in patients with urothelial cancer. Int J Cancer 113, 189–197.
https://doi.org/10.1002/ijc.20564 [doi]
Stiborova, M., Frei, E., Sopko, B., Sopkova, K., Markova, V., Lankova, M., Kumstyrova, T.,
Wiessler, M., Schmeiser, H.H., 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.
https://doi.org/10.1093/carcin/bgg119bgg119 [pii]
Stiborova, M., Levova, K., Barta, F., Shi, Z., Frei, E., Schmeiser, H.H., Nebert, D.W.,
Phillips, D.H., Arlt, V.M., 2012. Bioactivation versus detoxication of the urothelial
carcinogen aristolochic acid I by human cytochrome P450 1A1 and 1A2. Toxicol. Sci.
125, 345–358. https://doi.org/10.1093/toxsci/kfr306
Stiborova, M., Mareis, J., Frei, E., Arlt, V.M., Martinek, V., Schmeiser, H.H., 2011. The
human carcinogen aristolochic acid i is activated to form DNA adducts by human
NAD(P)H:quinone oxidoreductase without the contribution of acetyltransferases or
sulfotransferases. Env. Mol Mutagen 52, 448–459. https://doi.org/10.1002/em.20642
Toga, T., Kuraoka, I., Watanabe, S., Nakano, E., Takeuchi, S., Nishigori, C., Sugasawa, K.,
29