Structures of covalent adducts between DNA and ochratoxin a: A new factor in debate about genotoxicity and human risk assessment

Article abstract of DOI:10.1021/tx900295a

The potent renal carcinogenicity of ochratoxin A (OTA) in rats, principally in the male, raises questions about mechanism. Chromatographic evidence of DNA adducts after ³²P-postlabeling analysis contrasts with experimental attempts to demonstrate the absence of OTA in such adducts. Proffered schemes for alternative epigenetic mechanisms in OTA carcinogenicity remain unsatisfying, while structural data substantiating DNA-OTA adducts has also been lacking. We report refined ³²P-postlabeling methodology revealing one principal adduct isolated in small amounts from the kidneys of all five Fischer and five Dark Agouti rats to which OTA had been given on four consecutive days. We also describe structural data for the principal adduct from OTA/DNA interaction in vitro and its subsequent preparative isolation by the postlabeling methodology (as C-C8 OTA 3′dGMP), essentially creating an ochratoxin B - guanine adduct. Reasoning for the unsuitability of experimental protocols in published evidence claiming nongenotoxicity of OTA is given. In vivo exposure of renal DNA to cycles of adduction with OTA, necessarily protracted for carcinogenesis to occur, can reasonably explain an occasional focal neoplasm from which metastasizing carcinoma could develop.

Full text of DOI:10.1021/tx900295a

Chem. Res. Toxicol. 2010, 23, 89–98
89
Structures of Covalent Adducts between DNA and Ochratoxin A:
A New Factor in Debate about Genotoxicity and Human Risk
Assessment
Peter G. Mantle,*,† Virginie Faucet-Marquis,‡ Richard A. Manderville,§ Bianca Squillaci,|
and Annie Pfohl-Leszkowicz‡
Centre for EnVironmental Policy, Imperial College London, London SW7 2AZ, U.K., Ecole Nationale Supe´rieure
Agronomique, UMR CNRS/INPT/5503, Toulouse, France, Departments of Chemistry and Toxicology, UniVersity
of Guelph, Guelph N1G 2W1, Canada, The Drug Metabolism and Pharmacokinetics Department,
GlaxoSmithKline, Ware SG12 0DP, U.K.
ReceiVed August 24, 2009
The potent renal carcinogenicity of ochratoxin A (OTA) in rats, principally in the male, raises questions
about mechanism. Chromatographic evidence of DNA adducts after 32P-postlabeling analysis contrasts
with experimental attempts to demonstrate the absence of OTA in such adducts. Proffered schemes for
alternative epigenetic mechanisms in OTA carcinogenicity remain unsatisfying, while structural data
substantiating DNA-OTA adducts has also been lacking. We report refined 32P-postlabeling methodology
revealing one principal adduct isolated in small amounts from the kidneys of all five Fischer and five
Dark Agouti rats to which OTA had been given on four consecutive days. We also describe structural
data for the principal adduct from OTA/DNA interaction in vitro and its subsequent preparative isolation
by the postlabeling methodology (as C-C8 OTA 3′dGMP), essentially creating an ochratoxin B-guanine
adduct. Reasoning for the unsuitability of experimental protocols in published evidence claiming
nongenotoxicity of OTA is given. In vivo exposure of renal DNA to cycles of adduction with OTA,
necessarily protracted for carcinogenesis to occur, can reasonably explain an occasional focal neoplasm
from which metastasizing carcinoma could develop.
exposure to a nongenotoxic carcinogen rather than to a
carcinogen concerning which adducts of defined structure are
known. Food and drinks industries with commercial interests
in the significance of natural and largely unavoidable traces of
OTA in some agricultural commodities, and a need to comply
with statutory limits, will naturally prefer that OTA is shown
not to be a genotoxin.
In recent years, several publications consistently claiming the
nonexistence of direct OTA/DNA adducts have been noted
(summarized in ref 7); concurrently, several other publications
have claimed the opposite (summarized in refs 8 and 9).
However, there has been little critical appraisal of the experi-
mental material used in publications claiming that DNA/OTA
covalent adducts do not occur. Meanwhile, we have addressed
the difficult challenge of obtaining structural data relating to
rat kidney DNA adducts. Demonstration of the structure of a
principal DNA/OTA adduct, prepared in vitro from calf thymus
DNA and isolated preparatively via postlabeling methodology,
here extends proof of covalent binding of OTA on guanine
beyond the quite compelling cochromatographic implication
described in ref 10.
Introduction
Whether the nephrotoxic mycotoxin ochratoxin A (OTA)
exerts its potent nephrocarcinogenicity in male rats by direct
interaction with DNA has long been a matter of debate. First
positive claims came from the application of 32P-postlabeling
methodology to DNA isolated from mice given OTA (1).
Autoradiographic evidence of radiolabeled compounds separated
after contact transfer on two-dimensional chromatograms was
proposed as showing that OTA had made covalent bonds with
nucleotides in vivo. Subsequently, postlabeling methodology has
consistently detected adduct spots on chromatograms which
were attributed to exposure of rats, mice, and pigs to OTA in
laboratories in Strasbourg, Bordeaux, Lyon (IARC) and Tou-
louse. Little or no evidence for this conclusion was experienced
in other laboratories (e.g., see ref 2), thus fueling the current
reluctance to accept positive postlabeling evidence as an
indicator of genotoxicity of OTA (3). Academic debate about
whether or not OTA conforms to the current definition of a
genotoxin might have rested as a minor unresolved esoteric
controversy if it were not for concerns that experimental
carcinogenicity in male rats and mice pointed to the possibility
that OTA could be a human carcinogen (4-6). Clear demon-
stration that a chemical carcinogen is a genotoxin influences
legislative attitude, whereby a 10-fold reduced human risk
differential may be applied when assessing the extent of
Materials and Methods
Rats and OTA. Ten 8-week old males of the Fischer and Dark
Agouti strains (B & K Universal Ltd., Grimston, Hull, UK) were
∼220 and 180 g, respectively, and were maintained on standard
(14.4% protein) diet (SDS Services, Witham, Essex, UK).
Design of OTA dosing strategy was based on previous experience
(11) of short-term administration with only mild clinical effects
and best incidence of renal DNA adducts achieved through daily
oral gavage. OTA (77 mg) was dissolved in ethanol (1.2 mL) and
* To whom correspondence should be addressed. E-mail: p.mantle@
imperial.ac.uk.
† Imperial College London.
‡ Ecole Nationale Supe´rieure Agronomique.
§ University of Guelph.
| GlaxoSmithKline.
10.1021/tx900295a  2010 American Chemical Society
Published on Web 11/23/2009

90 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Mantle et al.
the solution made up to 20 mL with 3% w/v NaHCO3 in water.
Then 0.4 mL (1.5 mg OTA) was given from this solution to each
of seven rats, at the same time of day by oral gavage on three
consecutive days. The daily dose was therefore ∼8.3 and 6.8 mg/
kg b.w. for Dark Agouti and Fischer rats, respectively. Controls
were given 0.4 mL of vehicle (5% ethanol in 3% aqueous NaHCO3).
Twenty-four hours after the last dose, four treated rats from each
strain and all controls were euthanized, and blood was withdrawn
from the heart. The kidneys were excised, separated into quarters,
flash frozen in liquid N2, transferred into separate Eppendorf tubes,
and stored at -20 °C. Blood was centrifuged immediately and
plasma separated for quantitative OTA analysis at the Central
Science Laboratory, York, UK. The mean weights of the remaining
three OTA-treated Fischer and Dark Agouti rats were 310.7 (
12.1 g and 261.7 ( 10 g, respectively, 5 weeks after the first OTA
dose. After 3 months (370 ( 17.3 g and 278.3 ( 4.7 g,
respectively), the rats were euthanized and the blood collected as
before. A narrow central transverse section was removed from each
kidney for H & E-stained histology and the remaining tissues flash
frozen in liquid N2 and stored at -20 °C pending 32P-postlabeling
analysis.
Figure 1. Comparison of the kidney DNA profile of an OTA-treated
rat. (a) D1 2.3 M phosphate buffer, + transfer of 4 cm. (b) D1 3 M
phosphate buffer, + transfer of 2 cm.
Postlabeling Method. The equivalent of 7 µg of DNA was dried
in vacuo, dissolved in 10 µL of the mix containing 1 µL of
micrococcal nuclease (2 mg/mL corresponding to 500 U), spleen
phosphodiesterase (15 mU/µg DNA), 1 µL of sodium succinate
(200 mM), and 1 µL of calcium chloride (100 mM, pH 6) and
digested at 37 °C for 4 h. The digested DNA was then treated with
5 µL of the mix containing 1.5 µL of nuclease P1 (4 mg/mL), 1.6
µL of ZnCl2 (1 mM), and 1.6 µL of sodium acetate (0.5 M, pH 5)
at 37 °C for 45 min. The reaction was stopped by the addition of
3 µL of Tris base 500 mM. The DNA adducts were labeled as
follows: to the nuclease P1 digest, 5 µL of the reaction mixture
containing 2 µL of bicine buffer [bicine (800 µM), dithiothreitol
(400 mM), MgCl2 (400 mM), and spermidine (400 mM) adjusted
to pH 9.8 with NaOH], 10 U of polynucleotide kinase T4, and 100
µCi of [γ-32P]ATP (specific activity 6000 Ci/mmol) was added and
the mixture incubated at 37 °C for 45 min. Normal nucleotides,
pyrophosphate, and excess ATP were removed by chromatography
on PEI/cellulose TLC plates (D1) in 3 M NaH2PO4 buffer, pH 5.7,
overnight. The origin (2 cm) areas containing labeled adducted
nucleotides were cut out and transferred to another PEI/cellulose
TLC plate, which was run (D2) in 4.8 M lithium formate and 7.7
M urea, pH 3.5, for 3 h. A further (D3) migration was performed
after turning the plate 90° anticlockwise in 0.6 M NaH2PO4 and
5.95 M urea, pH 6.4, for 3 h. Finally, the chromatogram was washed
in the same direction in 1.7 M NaH2PO4, pH 6, for 2 h (D4). Adduct
profiles were analyzed qualitatively and semiquantitatively by
autoradiography of the plates, carried out at -80 °C for 48 h in
the presence of an intensifying screen, using a radio-analytical
system of image analysis (AMBIS, Lablogic). In parallel to the
adduct analysis, a series of controls were performed to verify the
labeling efficiency of adducts and the efficiency of dephosphory-
lation of the nuclease P1. Refinement of adduct separation by use
of a 3 M phosphate buffer, instead of the former 2.3 M buffer, in
the D1 migration, and subsequent transfer of only 2 cm of the origin,
is shown in Figure 1. Steps in the refined procedure to resolve two
of our synthetic model adducts are illustrated in Figure 2.
Synthesis of OTA-dG and OTA-3′-dGMP. A sample of the
OTA-dG nucleoside adduct was prepared from the photoreaction
of OTA in the presence of excess dG, as outlined previously
(12, 13). A reaction mixture (1 mL total volume) of 500 µM OTA
and 40 mol equiv of dG in a mixture of 25 vol. % DMSO and 75
vol. % 100 mM aqueous phosphate buffer (pH 7.4) was irradiated
at 350 nm for 15 min using a Rayonet Chamber Reactor, Model
RPR-200. The OTA-dG nucleoside adduct was isolated using an
Agilent 1200 series HPLC equipped with an autosampler, auto-
collector, and diode array detector, and a mobile phase consisting
of 40 mM ammonium formate (pH 6.1)/acetonitrile and a semi-
preparative Phenomenex Kromasil C8 column (10 mm × 250 mm)
at a flow rate of 5 mL/min. The isolated sample of OTA-dG had
MS and UV-vis characteristics identical to those previously
reported (12, 13).
All kidney samples were coded for blind analysis for adducts,
to avoid suggestion of any analytical bias, and transferred personally
by air in dry ice between London and Toulouse.
DNA Extraction and Postlabeling Separation of DNA
Adducts. Materials. The following enzymes [proteinase K (used
as received), RNase A, RNase T1 (boiled 10 min at 100 °C to
destroy DNases), and microccocal nuclease (dialyzed against
deionized water)] were purchased from Sigma (Saint Quentin
Fallavier, France); spleen phosphodiesterase (centrifuged before use)
was from Calbiochem (VWR, France), and nuclease P1 (NP1) and
T4 polynucleotide kinase were from Roche diagnostics (Meylan,
France). [γ 32P-ATP] (444 Tbq/mmol, 6000 Ci/mmol) was from
Amersham (Les Ullis, France). Rotiphenol (phenol saturated with
TRIS-HCl at pH 8) was from Rothsichel (Lauterbourg, France);
cellulose MN 301 was from Macherey Nagel (Du¨ren, Germany);
polyethyleneimine (PEI) was from Corcat (Virginia Chemicals,
Portsmouth, VA); Whatman No 1 paper (ref 6130932) was from
VWR (France), and PEI/cellulose TLC plates used for 32P-
postlabeling analyses were prepared in the laboratory at Toulouse,
France. All reagents (potassium chloride, sodium hydrogen carbon-
ate, sulfuric acid, phosphoric acid, hydrochloric acid, acetic acid,
and sodium dihydrogen phosphate) were of normal grade.
DNA Extraction. Tissues (500 mg) were homogenized in 0.8
mL of a solution containing NaCl (0.1 M), EDTA (20 mM), and
Tris-HCl, pH 8 (50 mM) (SET) in an ice bath. To the homogenate,
200 µL of a 10% solution of sodium dodecylsulfate was added,
and following incubation for 10 min at 65 °C, 800 µL of potassium
acetate (6 M, pH 5) was added. The reaction mixture was kept at
0 °C for 30 min. After centrifugation for 25 min at 0 °C (10000g),
the supernatant, which contained nucleic acids, was collected, and
nucleic acids were precipitated overnight at -20 °C by adding 2
volumes of cold ethanol. The DNA pellets were collected and
washed once with 1 mL of 90% ethanol and dissolved in 500 µL
of SET (15 min at 37 °C). The total extract was mixed with 10 µL
of a mixture of RNase A (20 mg/mL) and RNase T1 (10 000 U/ml)
and incubated for 1 h at 37 °C; this treatment was repeated twice.
Samples were then treated with 25 µL of proteinase K solution
(20 mg/mL SET) for 1 h at 37 °C. After digestion, 500 µL of
rotiphenol was added, and the mixture was moderately shaken for
20 min at room temperature and centrifuged for 15 min at 15 °C
(10000g). The aqueous phase was collected after two extractions.
After a final extraction with one volume of chloroform/isoamyl
alcohol (24:1), the aqueous phase was collected, and 50 µL of
sodium acetate (3 M, pH 6) was added. The DNA was precipitated
by the addition of two volumes of cold ethanol overnight at -20
°C, and the precipitate was collected by centrifugation at 10000g
for 30 min. The DNA pellet was washed four times with 90%
ethanol. DNA was dissolved in deionized water and tested for purity
by recording UV spectra between 220 and 320 nm.
The protocol for the photochemical synthesis of OTA-3′-dGMP
was the same as that outlined previously (10). A reaction mixture
(1 mL total volume) of 100 mM OTA and 50 mol equiv of 3′-
dGMP in 100 mM phosphate buffer (pH 7.4) was irradiated for 2

CoValent Adducts between DNA and Ochratoxin A
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 91
for 13 min using a Rayonet Chamber Reactor, Model RPR-200.
The sample was then divided into four 250 µL fractions, and 1 mL
of 100% ethyl alcohol was added to each fraction in microcentrifuge
tubes. After incubation at -20 °C for 10 h, samples were
centrifuged for 15 min at 12000 rpm and 6 °C in an AccuSpin
centrifuge. The DNA pellet was washed with 70% ethanol, then
dried under vacuum and sent to the Pfohl-Leszkowicz laboratory
for 32P-postlabeling analysis and preparative isolation of two
prominent adducts.
Preparative Isolation of Adducts from the Photoreaction
of OTA and DNA. The protocol for postlabeling was exactly the
same as that for rat DNA, except that unlabeled ATP was used
instead of that radiolabeled with 32P. For postlabeling of 10 µg of
DNA, 20 pmol of ATP was used. After postlabeling, a preparative
chromatogram was prepared in which typically (Figure 3) were
applied radioactive spots of our standard C-C8 adduct, products of
the calf thymus DNA-OTA interaction postlabeled with 32P ATP,
our standard O-C8, and a mixture of C-C8 and O-C8 adducts. In
the central region of the plate, a stripe of products of the calf thymus
DNA-OTA interaction postlabeled only with unlabeled ATP was
applied. Chromatography in the D1 direction was run with NaH2PO4
(2.3 M, pH 5.7) for 16 h. The plate was washed twice by soaking
in ultrapure water, dried, and autoradiographed. Cellulose in the
central area corresponding to spots of interest was scraped off, the
cellulose suspended in methanol, shaken, centrifuged, the solution,
and the extraction cycle repeated. Residual cellulose was then
similarly extracted with triethanolamine. Combined extracts were
evaporated to dryness in vacuo. The preparative eluates from spots
1 and 2 (Figure 3) were sent, blind as samples 1 and 2, respectively,
for LC-MS analysis.
Chemicals for HPLC/MS Analysis of Adducts. OTA from
Aspergillus ochraceus was purchased from Sigma-Aldrich Co.
(Dorset, UK), formic acid was purchased from Perbio Science
(Northumberland, UK), HPLC grade acetonitrile was purchased
from Fisher Scientific (Loughborough, UK), and HPLC grade water
was produced by an Elga Maxima Ultra Pure Water system
(Buckinghamshire, UK). All solvents and chemical standards were
of the highest grade commercially available. Authentic standards
of C-C8 OTA 3′-dGMP (Figure 4) and O-C8 OTA 3′-dGMP
(Figure 5), and OTA-DNA samples were synthesized in Guelph.
Buffers, Reagents, and Standards. A 10 µg/mL MS standard
of OTA in 50:50 acetonitrile/water was prepared by diluting a 10
µL of a 1 mg/mL solution of OTA in water with 990 µL of 50:50
acetonitrile/water. A 0.1% formic acid solution was prepared by
diluting 1 µL of concentrated formic acid with 999 µL of water.
Figure 2. Autoradiographic illustration of steps in chromatographic
resolution of two of our model synthetic adducts after 32P-postlabeling,
showing discrete separation resolved from stray radioactivity after a
second D4 step. Transition from D1-D2 was by contact transfer of the
radioactive region after the D1 step.
min at 350 nm. The isolated OTA-3′-dGMP adduct had MS and
UV-vis characteristics identical to those reported previously (10)
and was used as a postlabeling standard.
Photoreaction of OTA with DNA. A reaction mixture (1 mL
total volume) of 500 µM OTA and 100 µg of calf thymus DNA
(repurified) in a mixture of 10 vol. % DMSO and 90 vol. % 100
mM aqueous phosphate buffer (pH 7.0) was irradiated at 350 nm
Method Development of MS Analysis of Model Synthetic
Adducts in Preparation for the Analysis of Products of in
Vitro OTA-DNA Interaction. Authentic synthetic standards of
model OTA-guanine compounds, C-C8 OTA 3′-dGMP, and O-C8
OTA 3′-dGMP were used to develop a suitable HPLC/MSn method.
Figure 3. Annotated autoradiograph of a model chromatogram (D1 separation) illustrating the preparative isolation of samples 1 and 2 used for
structural analysis. Lane 1 is synthetic C-C8dG-OTA processed via standard 32P-postlabeling. Lanes 2 and 3 are products of calf thymus DNA/OTA
photoirradiation processed via standard 32P-postlabeling to guide preparative excision of unlabeled compounds from the central region. Lane 4 is
synthetic O-C8dG-OTA also processed via 32P-postlabeling. It was necessary to avoid 32P contamination of samples for LC-MS.

92 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Mantle et al.
Results
Pharmacokinetics in Rats Treated with OTA. Plasma OTA
concentration in the four Fischer rats treated with OTA was
45.7 ( 2.6 µg/mL and in the Dark Agoutis was 36.8 ( 6.4
µg/mL. At first sight, the lower value in the smaller rats given
the same amount of OTA appears incongruous. Purely on a body
weight basis, the plasma concentration in the Dark Agoutis
might have been ∼56 µg/mL. Subsequent discovery that Dark
Agoutis have much shorter plasma OTA half-life than that of
Fischer rats (14) can account for the apparent disparity in
accumulation in plasma after the three consecutive doses.
However, substantial potential for delivery of OTA via vascular
circulation to kidneys in both rat strains was demonstrated.
Histology of kidneys from animals euthanized 3 months after
giving OTA showed sparse karyomegaly diffused across the
outer stripe of the outer medulla, consistent in our experience
with a typical response in renal tubule epithelia to circulating
OTA.
Figure 4. Structure of C-C8 OTA 3′-dGMP.
Results of Postlabeling Analysis. Autoradiographs relating
to the kidney of each animal were copied electronically from
Toulouse to London for matching to the blind code. No spots
were present in samples from control (untreated) rats or from
treated rats euthanized 3 months after the administration of OTA.
This applied to all replicates of both strains of rats. In contrast,
for the kidneys of all OTA-treated rats analyzed after the three
consecutive doses, one or two radioactive spots were present
in otherwise very clean autoradiographs of chromatograms,
indicating the usual 32P-postlabeling evidence of adducted DNA
arising from consistent delivery of OTA to nephron epithelia
(Figure 6). Notably, there was close replication of the strain-
specific pattern of DNA adducts. In Fischer rats, there was
predominantly one adduct, while in Dark Agoutis the same
adduct was evident together with a minor component with
slightly greater chromatographic mobility. The abundance of
the principal adduct across the 10 animals was in the range of
20-70 adducts per 109 nucleotides.
MS Analysis. Preparative isolation of two adducts from in
vitro interaction between OTA and DNA, occurring in ap-
proximately equal amounts, provided material for LC-MS
analysis coded only as samples 1 and 2. An approximate limit
of detection of 1 ng of material on column had been determined
during method development.
No peaks were identified in sample 1 at mass m/z 713, the
expected mass of the C-C8 OTA 3′dGMP adduct. However, a
prominent peak was identified in the m/z 517 extracted ion
chromatogram at 11.7 min. The mass spectrum under the peak
is shown in Figure 7. The base peak ion at m/z 517 corresponded
to the loss of the phosphate conjugated deoxyribose sugar moiety
from the C-C8 OTA 3′dGMP, which is consistent with hy-
drolysis due to the addition of formic acid in the sample
pretreatment step.
Figure 5. Structure of O-C8 OTA 3′-dGMP.
For all analytes, the instrument was tuned to give the best possible
MS performance. HPLC/MSn analysis was performed on an Agilent
1100 HPLC system (Santa Clara, USA) consisting of a binary pump
and column thermostat, coupled to a Waters QTof F Premier mass
spectrometer (Manchester, UK) equipped with an electrospray
ionization (ESI) source with separate Lockmass interface and
operated in negative ion mode. Samples were introduced to the
HPLC via a CTC-Analytics PAL autosampler (Zwingen, Switzer-
land) consisting of a 100 µL loop. The autosampler, HPLC system,
and mass spectrometer were controlled using Waters software
MassLynx version 4.1 (Manchester, UK). Calibration over the range
m/z 50-1000 was carried out using a NaI/CsI calibration standard
purchased from Waters. MS tuning was performed using direct
infusion of a 10 µg/mL standard solution of OTA, at a flow rate of
10 µL/min, using the integrated instrument syringe pump. The
following mass spectrometer conditions were common to all LC/
MS experiments performed: capillary voltage, 1.5 kV; sampling
cone, 35 V; source temperature, 140 °C; desolvation temperature,
350 °C; cone gas flow, 50 L/h; desolvation gas flow, 400 L/h; and
LM and HM resolution, 15 arbitrary units. A 1 µg/mL solution of
leucine enkephalin (m/z 554.2615) in 50:50 acetonitrile/water was
pumped into the lockmass interface at a rate of 20 µL/min using a
dedication HPLC pump (Shimadzu UK Limited, Manchester, UK).
Lockmass data was collected with a scan time of 0.5 at 10 s intervals
and averaged over 10 scans. Mass spectral data was collected in
MS centroid mode with a scan time of 1 s and interscan time of
0.02 s, over a mass range of m/z 100-1000.
Extracted ion chromatograms of m/z 517 for samples 1 and
2, together with the analysis of the authentic standard of C-C8
OTA 3′dGMP cochromatographed at 11.7 min, confirming the
deconjugated aglycone moiety, were present in all three analyses
(Figure 8).
All samples were pretreated with concentrated formic acid (50
µL), vortex mixed, and heated on a heating block at 80 °C for 60
min. The samples were then reduced to dryness under N2 gas,
resuspended in 100 µL of 0.1% formic acid (aqueous), and vortex
mixed. Ten microliters of each sample was injected onto a C18
column (Thermo Hypersil Gold 5 µm, 4.6 mm i.d. × 250 mm,
Thermo Fisher Scientific, Waltham, USA); 0.1% formic acid
(solvent A) and acetonitrile (solvent B) were used as mobile phases.
The samples were separated using a gradient elution between 5%
and 95% B in 15 min. The column was kept at 95% B for 5 min
and then reconditioned at 5% B for 10 min.
The mass spectrum under the peak at 11.7 min in OTA-DNA
sample 2 is shown in Figure 9.
The peak at m/z 517 corresponds to the loss of the phosphate
conjugated deoxyribose sugar moiety from the C-C8 OTA
3′dGMP as seen for OTA-DNA sample 1. However, additional
adduct related ions are also observed in the spectrum. The ion
m/z 655 corresponds to the sodium adduct of the C-C8 OTA
3′dGMP which has lost the phosphate conjugate but still has

CoValent Adducts between DNA and Ochratoxin A
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 93
Figure 6. Representative autoradiographic displays from Dark Agouti and Fischer rats illustrating similar adduct patterns for each strain but consistent
dominance of an adduct attributed to C-C8 OTA 3′-dGMP.
Figure 7. Background subtracted mass spectrum for OTA-DNA
sample 1.
the deoxyribose sugar moiety intact (Figure 10a). The ion at
m/z 815 represents the sodium adduct of the 3′,5′-bisphosphate
OTA-dGMP, showing additional phosphate conjugation on the
5′ hydroxyl group of the deoxyribose sugar (Figure 10b).
The structure of the deconjugated aglycone observed in OTA-
DNA sample 2 was confirmed by MS/MS analysis of the
precursor ion m/z 517 (Figure 11). MS/MS data was generated
on the Waters QT of Premier in continuum mode over the mass
range m/z 100-600 at a collision energy of 25 eV, a scan time
of 1 s, and an interscan time of 0.02 s.
Figure 8. Extracted ion chromatograms of m/z 517 for (A) OTA-DNA
sample 1, (B) OTA-DNA sample 2, and (C) the authentic standard of
C-C8 OTA 3′dGMP after pretreatment with formic acid. The data for
the authentic standard is weak but clearly shows a peak at the correct
retention time.
The ions m/z 473 and m/z 429 represent the loss of the
carboxylic acid group as CO2, with a further loss of CO2 from
the lactone moiety, thus confirming the presence of the OTA
adduct in OTA-DNA sample 2. Matching of analytical findings
to the blind sample code showed sample 1 to be C-C8 OTA
3′dGMP and sample 2 to be its triphosphorylated homologue.
the Pfohl-Leszkowicz laboratory by 32P-postlabeling in renal
tissue after exposure to OTA, as far as possible with limited
research resources. Limited resource of renal tissue precluded
preparative isolation for LC-MS of sufficient amounts of either
of the two key adducts revealed only by the 2D chromatography
of 32P-postlabeling after contact transfer. At least 1 ng would
Discussion
The present findings have taken the search for definitive
evidence on the structures of adducts, revealed consistently in

94 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Mantle et al.
Figure 9. Background-subtracted mass spectrum for OTA-DNA of sample 2.
Figure 10. (a) Structure of desphosphate C-C8 OTA 3′dGMP (m/z 655)
and (b) structure of 3′,5′-bisphosphate OTA-dGMP (m/z 815).
be needed on the column and would require ∼100 rats. The
present finding confirms previous recognition of C-C8-dG-OTA
as an in vivo DNA-OTA adduct by cochromatography in a 2D
system after contact transfer with an authentic standard (10).
We also show (Figure 12) similarly consistent chromatographic
behavior through five sequential steps in two dimensions for
C-C8-dG-OTA as a product of DNA-OTA interaction both in
vitro and in vivo. However, research effort may better be focused
on understanding the genetic changes in rat renal tumors caused
by OTA and the mechanism for marked gender difference in
sensitivity and pharmacokinetics (14, 15). Approximately 1018
molecules of OTA necessarily pass via a male rat’s kidneys
during several months to cause any renal cancer (16).
The experimental design described in ref 17 is presumably
based on previous OTA toxicokinetics (18) in which a single
dose (500 µg/kg body weight) was given by oral gavage. For
male rats in that study, assuming that the quoted rat weight
(∼250 g) is more reliable than the stated 8 week age (cf. Charles
River Web site for Fischer rats), approximately 130 µg of OTA
had been given per os in 500 µL of corn oil. Maximum
concentration of OTA in blood, measured in plasma 48 h later,
had reached ∼1.8 µg/mL, corresponding to ∼0.1 µg/g measured
in kidney from which blood had been flushed by syringing with
physiological saline after excision. Thus, in a rat given 130 µg
of OTA, less than 0.2% was actually found resident in its
kidneys. If blood containing 1.8 µg OTA/mL of the plasma
component circulates through the kidney in which blood
Figure 11. MS/MS spectrum (collision energy 25 eV) of m/z 517,
showing CID fragmentation of the deconjugated aglycone of C-C8 OTA
3′dGMP. The spectrum shows the two principal fragment ions m/z 473
and m/z 429.
constitutes 20-25% of tissue (19), OTA concentration in the
organ would be ∼0.4 µg/g. Since removal of blood from the
kidney can only be achieved efficiently, particularly from the
peritubular capillaries, by perfusion in situ under normal blood
pressure (19), the concentration of OTA in nonvascular tissue
within the kidney would have been considerably less than the
0.2% of the given dose recorded in ref 18. The similar dosing
of three adult male rats in ref 17, with assumed weight of ∼250
g and each rat given ∼130 µg 14C-OTA (0.25 mCi/mmol), thus
delivered ∼0.08 µCi 14C to each animal. This translates to
∼170,000 dpm of 14C, of which only about 50 dpm might have
been intracellular in the 300-400 mg of kidney used for the

CoValent Adducts between DNA and Ochratoxin A
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 95
This is consistent with strong binding of OTA to serum albumin
and a plasma half-life of about 10 days in the Fischer rat (14).
However, the pattern of OTalpha excretion is striking and
informative, although no discussion of this finding was made
in ref 25. OTalpha in urine indicates metabolism of OTA in
nephron epithelia and, at higher OTA doses, this activity is
prominent. However, OTalpha excretion falls quickly nearly to
zero during the two-day nondosing period, but exactly the same
pattern is repeated during the second week. Thus, repeated pulses
of OTA dosed by oral gavage are rather quickly subject to
metabolism in nephron epithelia. Therefore, waiting for 3 days
after a single OTA dose in ref 17 can be seen as counterproduc-
tive in an experiment designed to find evidence of DNA/OTA
adducts in whole kidney.
The general lack of DNA adduct detection in several
publications (2, 17, 25, 26) could be explained by several
drawbacks in the methodologies used. Indeed, the experimental
conditions and the methodologies used in these papers are not
optimal and did not reproduce those described by the others.
Notably, there is no quality verification of the DNA used for
ACMS analysis or 32P-postlabeling in the key publication (17)
used to underpin much of the current claims that OTA is not a
genotoxin. For some in vivo experiments, a low dose (0.2-1
mg/kg b.w.) was used for a short exposure time, 24 h; and for
ACMS detection, the conditions were 0.5 mg/kg b.w. for 72 h.
It has long been stated clearly that most of the adducts at a low
(0.6 mg/kg b.w.) dose or a medium (1.2 mg/kg b.w.) dose
become undetectable by 72 h (1, 27). Thus, the conditions were
not optimal for adduct recognition. Moreover, the ACMS
measurement had been performed without piloting authentic
samples for comparison, and they used a Turbo Ionspray source
and heated the samples to 400 °C. Most of the DNA adducts
and OTA-GSH will probably decompose at such high temper-
atures. For generation of OTA-DNA adduct standards, the
photoreaction was carried out with OTA at 500 mM and dG at
20 mM, conditions that are at an exceedingly high concentration
and do not match the conditions reported in ref 12 for successful
isolation and characterization of the adduct standards. Also, it
appears that no purification of the photoreaction was carried
out and, at best, the solution would have been a mixture of C-C8
and O-C8 adducts.
Figure 12. Autoradiographic comparison of chromatographic mobility
after 32P-postlabeling of synthetic C-C8dG-OTA, the product of
photoradiation of calf thymus DNA with OTA, and kidney DNA of a
Dark Agouti rat given the oral OTA regimen. (A) D1 separation,
showing sources of contact transfer of spot 1 to new plates for
subsequent D2, D3, and D4 steps and (B) outcome of the completed
protocol showing the mobility of the standard adduct similar to that of
a principal radioactive component after both in vitro and in vivo
interaction with OTA.
extraction of DNA for ACMS analysis. Unfortunately, there was
no measurement of 14C content of kidney of the dosed animals.
Even with very efficient extraction and isolation (which seems
not to have been the case), the total content of radioactivity of
the DNA above background, about one 14C atom to disintegrate
per second, could hardly have significantly changed its natural
abundance even if all potentially disintegrating carbon atoms
had been involved in DNA/OTA adduct(s). This simple calcula-
tion should have been possible from the authors’ own data. It
is also unfortunate that no preliminary studies to optimize the
delivery of 14C-OTA to kidney had been performed (17),
considering the costly resource investment made elsewhere to
produce the 14C-OTA. Optimum concentration of OTA circulat-
ing in blood may only be achieved after several weeks of
continuous daily intake of the amount of 14C-OTA given in ref
17, as indicated by in vivo rat data (14, 16, 20) and corroborated
by predictive pharmacokinetic modeling (15). All these factors
compound the unsatisfactory design of the radiolabeled study,
which, together with failure to have the planned positive control
evidence of successful 32P-postlabeling of parallel material in
the Pfohl-Leszkowicz laboratory, fails to justify the claim of
nongenotoxicity. The claim was also premature because sub-
stitution, of commercial 14C phenylalanine for the corresponding
unlabeled moiety in the 14C-OTA prepared biosynthetically (21),
by well-established methodology (22-24) was not done.
Experiments described in ref 25 showed integrated findings
on concentration of OTA in blood and ochratoxin R (OTalpha)
in urine over a 2-week period of 5-day-a-week gavage dosing
in vegetable oil. This information provides insight into phar-
macokinetics applicable to the NTP study and the renal tumors
produced. Clearly, with increased dose, blood OTA concentra-
tion increased progressively irrespective of the two nondosing
days in the middle and end of the 2-week experimental period.
Conditions for optimizing DNA extraction are not correct in
ref 17. Mainly for the purification, they used a Qiagen column,
a technique shown in three interlaboratory studies on postla-
beling to induce a loss of DNA adduct (EU project 5 V0448
Comparison and validation of 32P-postlabeling method; EU
project BEQALM; EU project OTA Risk assessment QKL1-
2001-0614). The purity of the DNA was also not good enough,
probably being contaminated by RNA and protein, which would
have decreased postlabeling efficiency (28). Figure 13 shows
comparison of the UV spectrum of DNA prepared in the
Wurzburg laboratory contemporaneously with the studies of ref
17 and markedly contaminated with RNA with that from which
the adduct data in Figure 6 was obtained. Other spectra showed
contamination with protein. The chromatographic conditions of
separation were also not adequate for OTA-DNA adduction.
The chromatographic pH conditions for 32P-postlabeling analy-
sis, with D1 at pH 6.8 and D3 at pH 3.5, do not conform to
standard published procedures, where D1 is run at a pH value
below 6.0, and D3 is run at pH ∼6.4 (10). Indeed, in ref 17,
the separation conditions described were those for polycyclic
aromatic hydrocarbons (PAH). PAH conditions were too strong
for OTA and, even for the aristolochic acid positive control,
the pattern was not optimal. The authors speculate that the DNA

96 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Mantle et al.
kidney from other protocols (25, 37), not specifically designed
to optimize the recognition of very small amounts of adducted
nucleotide, is disappointing. Kidneys from rats given the higher
OTA dose (25) were obtained 3 days after the last OTA dose
and from animals with marked polyuria, which indicated some
renal dysfunction. It is unreasonable to assume efficient
delivery of OTA to proximal tubule epithelia if there is overt
evidence of renal dysfunction. We calculate that circulating
OTA concentration in blood plasma can account for all
measured OTA in kidneys so that no renal parenchyma
accumulation of OTA could be perceived.
Material sourced from a 90-day study (38) for the study of
ref 37 had plasma OTA concentration for the highest dose group,
designed to mimic that of the NTP study (4), only at a mean
value of 3 µg/mL (with rather large standard deviation). In any
case, such animals would probably not have had a sufficient
period of OTA exposure for renal tumorigenesis to have been
put in place (16), and there was no evidence that plasma OTA
concentration had reached a steady state. This is an important
consideration since covalent binding of OTA to DNA could
relate only rather specifically to renal tumorigenicity. Adducts
can occur without implication that they constitute an actual
carcinogenic risk if the toxin insult is below threshold (39). DNA
adducts were recognized by 32P-postlabeling in our studies on
kidneys of rats with plasma OTA concentration stabilized around
8 µg/mL (19), which was typical of that associated with long-
term renal tumorigenesis in male Fischer rats (36). The rats were
at the 12 month stage from a large cohort continuing for life
on dietary OTA and causing renal cancer (36), but at that time,
it was not realized that further exposure to OTA might be
unnecessary for tumorigenesis because no one had yet explored
latency for this toxin. This emphasizes that sensitive 32P-
postlabeling analysis should have been applied to the kidney
tissues selected for the study of ref 37 to give assurance that
spectrometric methodologies could be used with confidence to
assess whether adducts were present. Otherwise, acquisition of
negative evidence from protocols that have not been challenged
concerning their fitness for purpose is an extremely weak basis
for constructing a putative mechanism in carcinogenesis.
Figure 13. Comparison of UV-spectra of rat kidney DNA isolated by
EU partners in Wurzburg (above) and Toulouse (below) as part of a
2003 interlaboratory study, as illustrated in attachments 11 and 7,
respectively, submitted to the EU project coordinator for the official
report but not included. Spectra are matched here for x-axis scale.
Considering the DNA quality criterion that UVmax should not be <257
nm and that the UV absorbance ratio (260:280 nm) should be 1.8-2.0,
all Wurzburg samples were contaminated by RNA; the spectrum
illustrated here represents nearly pure RNA (UVmax 252; 260:280 nm
absorbance ratio 2.2).
adduct observed is due to the cytotoxicity of oxygen reactive
species leading for example to 8-oxoguanine. This modification
can, however, not be detected with the postlabeling method
using nuclease P1 enrichment, as it detects only bulky lipophilic
adducts and not hydroxylated hydrophilic nucleotides. In the
same way, etheno bases cannot be detected by this technique
(29). Recently, we explained how an oxidative pathway leads
to the covalent OTA-DNA adduct via biotransformation of OTA
into the quinone derivative OTHQ (10, 30, 31).
It has recently been pointed out that the work presented in
ref 37 was flawed because the limit of detection (LOD) (10
fmol dG-OTA, corresponding to 3.5 adducts per 109 nucleotides)
was obtained from DNA hydrolysates (500 µg DNA equivalent)
and not from the actual rat kidney sample for which the negative
results were presented (40). This argument stemmed from the
fact that a high amount of sample spiking (1250 fmol) with an
isotopically labeled dG-OTA adduct of 96-98% isotopic purity
was used, and the nonisotopically labeled dG-OTA adduct
impurity was observed in the DNA hydrolysates but not in the
rat kidney sample. Clearly, the LOD for the DNA hydrolysates
versus the rat kidney sample were not the same. DNA
purification from the rat tissue sample is more complicated,
requiring extra steps, and interfering species could dramatically
change the LOD. It was also pointed out that in the rat kidney
sample a peak for the dG-OTA adduct following 90 days of
incubation was actually observed in the ref 37 study using the
m/z 633 f 429 fragmentation transition in MS/MS detection.
This apparent peak was not attributed to the dG-OTA adduct
because an adduct peak was not observed using the transition
m/z 633 f 517. In response to the criticism, the authors argued
(41) that their method of quantification was sound, but they
avoided discussing whether they were justified in their assump-
tion that the LOD from the DNA hydrolysates could serve as
an accurate LOD for the rat kidney sample. They also point
If histopathological change is a general indicator of renal cells
that are significant active or passive sites of OTA uptake and
may be sites of persistent genetic change necessary for tum-
origenesis, the outer stripe of the outer medulla is ideally the
preferred tissue for finding adducts because prominent aneuploid
karyomegaly in response to some mycotoxins is generally
located in and around this region (4, 32-34). Even this region
is not absolutely defined; packing of nephron P3 segments does
not rigidly conform to a sharp line of demarcation in renal tissue,
particularly concerning those from peripheral glomeruli. Rec-
ognition of precise anatomical transition in segment structure
and function from P2 to P3 in a single nephron cannot easily
be made in histology sections. There is no doubt that OTA can
cause genetic damage both in vitro and in vivo. Demonstration
of this (25) in the kidney of rats dosed for a short period has
been complemented by data from cells in tissue culture (35);
similar demonstration was also made (Mosesso, unpublished
data) in kidneys at the 12-month stage of the lifetime experiment
described in ref 36.
More recent analysis by isotopic dilution to determine whether
OTA forms DNA adducts in vivo has also concluded that it
does not (37). However, as with the findings in ref 17, data
from sophisticated spectrometric methodology can only be as
reliable as is the quality of the material analyzed. The use of

CoValent Adducts between DNA and Ochratoxin A
Chem. Res. Toxicol., Vol. 23, No. 1, 2010 97
out that the absence of a signal at transition m/z 633 f 517 is
a strong argument against the presence of dG-OTA in the
sample, even though an adduct peak was observed using the
m/z 633 f 429 transition. However, the m/z 633 f 429
transition is more intense than the m/z 633 f 517 (20% greater),
and here a clear adduct peak above background was observed
that comigrates with the adduct standard. Given that the m/z
633 f 429 transition provides a more intense adduct signal and
showed positive evidence for dG-OTA in rat kidneys (for OTA-
treated rat at a dose of 210 µg/kg b.w. for 90 days), we can
argue that the findings from the ref 37 study are in agreement
with our positive evidence for dG-OTA presented here.
thanks Region Midi-Pyre´ne´es ‘food safety program’, Ligue
Nationale Franc¸aise Contre le Cancer (2007) and Association
pour la Recherche Contre le Cancer (2005; 2008) for funding.
We thank Mariana Tozlovanu and Maud Juzio for skilled
technical assistance in Toulouse and Guelph, respectively, and
the Drug Metabolism and Pharmacokinetics Department, Glaxo-
SmithKline for the use of their analytical facilities.
References
(1) Pfohl-Leszkowicz, A., Chakor, K., Creppy, E. E., Dirheimer, G. (1991)
DNA-Adduct Formation, in Mice Treated with Ochratoxin A, in
Mycotoxins, Endemic Nephropathy and Urinary Tract Tumours
(Castegnaro, M., Plestina, R., Dirheimer, G., Chernozemsky, I. N.,
and Bartsch, H., Eds.) IARC Scientific Publications No 115, pp 245-
253, IARC Scientific Publications, Lyon, France,.
(2) Gautier, J.-C., Richoz, J., Welti, D. H., Markovic, J., Gremaud, E.,
Guengerich, F. P., and Turesky, R.J. (2001) Metabolism of ochratoxin
A: Absence of formation of genotoxic derivatives by human and rat
enzymes. Chem. Res. Toxicol. 14, 34–45.
(3) EFSA. (2006) Opinion of the scientific panel on contaminants in the
food chain on a request from the Commission related to ochratoxin A
in food. EFSA J. 365, 1-56. www.efsa.eu.int.
(4) Boorman, G. A. (1989) Toxicology and Carcinogenesis Studies of
Ochratoxin A (CAS No. 303-47- 9) in F344/N Rats (Gavage Studies),
1989 National Toxicology Program, Technical Report 358.
(5) Bendele, A. M., Carlton, W. W., Krogh, P., and Lillehoj, E. B. (1985)
Ochratoxin A carcinogenesis in the (C57BL/6J x C3H)F1 mouse.
J. Natl. Cancer Inst. 75, 733–742.
However, a genotoxicity mechanism operating in animals in
which renal tumors will form and in which genetic damage has
been demonstrated will evoke a relatively simple concept. The
concept turns on a reasonable probability of occasional persistent
genetic change related to tumorigenesis residing in perhaps just
one nucleus in the ∼30,000 highly differentiated nephrons (42)
in just one of a rat’s kidneys. Such could certainly be expected
after a lifetime of continual bombardment of rat tubular epithelial
cells with OTA molecules, as in the principal lifetime carcino-
genicity studies to date (4, 6, 36). It is not unreasonable that
other recent studies have shown that lifetime exposure is not
mandatory (16) but that less than half a lifetime or even as little
as 6 months of continuous toxic insult during the first half of
an ∼2-year rat lifespan would also be sufficient to determine
that renal tumorigenesis could arise in some individuals.
Experimental OTA tumorigenesis seems to be highly focal in
rat kidneys, according to observations on about 50 tumorous
kidneys throughout recent studies in London. Bilateral OTA
tumorigenesis seems, in our experience in the rat, to be dose-
related, and this would be consistent with the statistical
probability of causing one focus of the necessary genetic damage
in each kidney of an animal.
(6) Castegnaro, M., Mohr, U., Pfohl-Leszkowicz, A., Esteve, J., Steinmann,
J., Tillmann, T., Michelon, J., and Bartsch, H. (1998) Sex- and strain-
specific induction of renal tumours by ochratoxin A in rats correlates
with DNA adduction. Int. J. Cancer 77, 70–75.
(7) Turesky, R. J. (2005) Perspective: Ochratoxin A is not a genotoxic
carcinogen. Chem. Res. Toxicol. 18, 1082–1090.
(8) Manderville, R. A. (2005) A case for the genotoxicity of ochratoxin
A by bioactivation and covalent DNA adduction. Chem. Res. Toxicol.
18, 1091–1097.
(9) Pfohl-Leszkowicz, A., and Manderville, R. A. (2007) Ochratoxin A:
An overview on toxicity and carcinogenicity in animals and humans.
Mol. Nutr. Food Res. 51, 61–99.
(10) Faucet, V., Pfohl-Leszkowicz, A., Dai, J., Castegnaro, M., and
Manderville, R. A. (2004) Evidence for covalent DNA adduction by
ochratoxin A following chronic exposure to rat and subacute exposure
to pig. Chem. Res. Toxicol. 17, 1289–1296.
(11) Miljkovic, A., Pfohl-Leszkowicz, A., Dobrota, M., and Mantle, P. G.
(2003) Comparative responses to mode of oral administration and dose
of ochratoxin A or nephrotoxic extract of Penicillium polonicum in
rats. Exp. Toxicol. Pathol. 54, 305–312.
(12) Dai, J., Wright, M. W., and Manderville, R. A. (2003) Ochratoxin A
forms a carbon-bonded C8-deoxyguanosine nucleoside adduct: Im-
plications for C8 reactivity by a phenolic radical. J. Am. Chem. Soc.
125, 3716–3717.
(13) Frenette, C., Paugh, R. J., Tozlovanu, M., Juzio, M., Pfohl-Leszkowicz,
A., and Manderville, R. A. (2008) Structure-activity relationships for
the fluorescence of ochratoxin A: Insight for detection of ochratoxin
A metabolites. Anal. Chim. Acta 617, 153–161.
(14) Mantle, P. G. (2008) Interpretation of the pharmacokinetics of
ochratoxin A in blood plasma of rats, during and after acute or chronic
ingestion. Food Chem. Toxicol. 46, 1808–1816.
(15) Vettorazzi, A., Gonzales-Penas, E., Troconiz, I. F., Arbillaga, L.,
Corcuera, L. A., Gil, A. G., and Lopez de Cerain, A. (2009) A different
kinetic profile of ochratoxin A in mature male rats. Food Chem.
Toxicol. 47, 1921–1927.
(16) Mantle, P. G. (2009) Minimum tolerable exposure period and
maximum threshold dietary intake of ochratoxin A for causing renal
cancer in Dark Agouti rats. Food Chem. Toxicol. 47, 2419–2424.
(17) Mally, A., Zepnik, H., Wanek, P., Eder, E., Dingley, K., Ihmels, H.,
Volkel, W., and Dekant, W. (2004) Ochratoxin A: Lack of formation
of covalent DNA adducts. Chem. Res. Toxicol. 17, 234–242.
(18) Zepnik, H., Volkel, W., and Dekant, W. (2003) Toxicokinetics of the
mycotoxin ochratoxin A in F344 rats after oral administration. Toxicol.
Appl. Pharmacol. 192, 36–44.
(19) Garcia-Sanz, A., Rodriguez-Barnero, A., Bentley, M. D., Ritman, E. L.,
and Romero, J. C. (1998) Three-dimensional microcomputed tomog-
raphy of renal vasculature in rats. Hypertension 31, 440–444.
(20) Castegnaro, M., Canadas, D., Vrabcheva, T., Petkova-Bocharova, T.,
Chernozemsky, I. N., and Pfohl-Leszkowicz, A. (2006) Balkan
endemic nephropathy: role of ochratoxin A through biomarkers. Mol.
Nutr. Food Res. 50, 519–529.
All experimental dose-response data for OTA’s renal car-
cinogenesis, combined and plotted in the format of ref 43, makes
a compelling case for OTA being a thresholded carcinogen in
the male rat (16). There should be no philosophical conflict that
the mycotoxin can also make covalent adducts with DNA and
thereby, after many months of exposure to toxin molecules
passing through renal epithelium, occasionally cause the par-
ticular combination of persistent DNA change that programs
an occasional renal cell toward tumorigenesis, discovered late
in life. Recently, it has been demonstrated (44) that genotoxic
thresholds can exist for some genotoxins. Thus, determining
the lowest dose that does not increase background mutation rate
is essential and represents an important new aspect of geno-
toxicology. We recently showed that continuous treatment of
Dark Agouti rats with feed contaminated with 50 ng OTA/kg
for 4 weeks led to the formation of DNA adduct (20, 28); this
dosage is 8-fold less than that having no adverse clinical effect
during lifetime exposure (16).
The current findings relate solely to defining the fundamental
genotoxicity of OTA as revealed by 32P-postlabeling and the
structure determination of the principal adduct to demonstrate
direct covalent binding of OTA to DNA. Further understanding
of the precise mutations involved in rat renal tumorigenesis in
response to OTA will assist evidence-based application of
animal model data to human risk assessment.
Acknowledgment. Part of this work was funded by the 5th
RTD Framework Programme of the European Union, Project
QLK1-2001-01614. R.A.M. thanks the Natural Sciences and
Engineering Research Council of Canada for funding. A.P.-L.

98 Chem. Res. Toxicol., Vol. 23, No. 1, 2010
Mantle et al.
(21) Nestler, S. (2008) Experimental Renal Ochratoxicosis: A Molecular
and Pathological Study, Ph.D. Thesis, University of London.
(33) Turner, D. R., Mantle, P. G., McHugh, K. M., Adatia, R., and Gray,
T. (1993) The Effect of Penicillium Moulds on the Kidney and Their
Relationship with Balkan Nephropathy, Abstract: Proceedings, 14th
European Pathology Congress, Innsbruck.
(34) Brown, A. L., Odell, E. W., and Mantle, P. G. (2007) DNA ploidy
distribution in renal tumours induced in male rats by dietary ochratoxin
A. Exp. Toxicol. Pathol. 59, 85–95.
(35) Mosesso, P., Cinelli, S., Pinero, J., Bellacima, R., and Pepe, G. (2008)
In vitro cytogenetic results supporting a DNA nonreactive mechanism
for ochratoxin A, potentially relevant for its carcinogenicity. Chem.
Res. Toxicol. 21, 1235–1243.
(36) Mantle, P., Kulinskaya, E., and Nestler, S. (2005) Renal tumourigenesis
in male rats in response to chronic dietary ochratoxin A. Food Addit.
Contam, Suppl. 1, 58–64.
(37) Delatour, T., Mally, A., Richoz, J., Ozden, S., Dekant, W., Ihmels,
H., Otto, D., Gasparutto, D., Marin-Kuan, M., Schilter, B., and Cavin,
C. (2008) Absence of 2′-deoxyguanosine-carbon 8-bound ochratoxin
A adduct in rat kidney DNA monitored by isotope dilution LC-MS/
MS. Mol Nutr. Food Res. 52, 472–482.
(38) Rached, E., Hard, G. C., Blumbach, K., Weber, K., Draheim, R., Lutz,
W. K., Ozden, S., Steger, U., Dekant, W., and Mally, A. (2007)
Ochratoxin A: 13-week oral toxicity and cell proliferation in male
F344/N rats. Toxicol. Sci. 97, 288–298.
(22) Hult, K. (1986) Ochratoxin A labelled with 14C or 3H in the
phenylalanine moiety. J. Labelled Comp. Rad. 23, 801–805.
(23) Breitholz-Emanuelsson, A., Fuchs, R., Hult, K., and Appelgren, L.-
E. (1991) Distribution of 14C-ochratoxin A and 14C-ochratoxin B in
Rats: A Comparison Based on Whole Body Autoradiography, in
Mycotoxins, Endemic Nephropathy and Urinary Tract Tumours
(Castegnaro, M., Plestina, R., Dirheimer, G., Chernozemsky, I. N.,
and Bartsch, H., Eds.) pp 201-203, International Agency for Research
on Cancer, Lyon, France.
(24) Breitholz-Emanuelsson, A., Fuchs, R., Hult, K., and Appelgren, L. E.
(1992) Synthesis of 14C-ochratoxin A and 14C-ochratoxin B and a
comparative study of their distribution in rats using whole body
radiography. Pharmacol. Toxicol. 70, 255–261.
(25) Mally, A., Pepe, G., Ravoori, S., Fiore, M., Gupta, R., Dekant, W.,
and Mosesso, P. (2005) Ochratoxin A causes DNA-damage and
cytogenetic effects, but no DNA-adducts in rats. Chem. Res. Toxicol.
18, 1253–1261.
(26) Schlatter, C., Studer-Rohr, J., and Rasonyi, T. (1996) Carcinogenicity
and kinetic aspects of ochratoxin A. Food Addit. Contam. 13, 43–44.
(27) Pfohl-Leszkowicz, A., Grosse, Y., Kane, A., Creppy, E. E., and
Dirheimer, G. (1993) Differential DNA adduct formation and disap-
pearance in three mice tissues after treatment by the mycotoxin,
ochratoxin A. Mutat. Res. 289, 265–273.
(39) Waddell, W. J., Fukushima, S., and Williams, G. M. (2006) Concor-
dance of thresholds for carcinogenicity of N-nitrosodiethylamine. Arch.
Toxicol. 80, 305–309.
(40) Pfohl-Leszkowicz, A., Gabryelski, W., and Manderville, R. A. (2009)
Formation of 2′-deoxyguanosine-carbon 8-bound ochratoxin A adduct
in kidney DNA. Mol. Nutr. Food Res. 53, 154–155.
(28) Pfohl-Leszkowicz, A., and Castegnaro, M. (2005) Further arguments
in favour of direct covalent binding of Ochratoxin A (OTA) after
metabolic biotransformation. Food Addit. Contam., Suppl. 1, 75–87.
(41) Delatour, T., Mally, A., Dekant, W., Marin-Kuan, M., Schilter, and
Cavin, C. (2009) Answer to Prof. Pfohl-Leszkowicz’s letter to the
editor. Mol. Nutr. Food Res. 53, 156–157.
(42) Solomon, S. (1977) Developmental changes in nephron number,
proximal tubular length and superficial nephron glomerular filtration
rate of rats. J. Physiol. 272, 573–589.
(43) Waddell, W.J. (2006) Critique of dose response in carcinogenesis.
Hum. Exp. Toxicol. 25, 413–436.
(44) Johnson, G. E., Doak, S. H., Griffiths, S. M., Quick, E. M., Skibinski,
D. O. F., Zair, Z. M., and Jenkins, G. J. (2009) Non-linear dose-
response of DNA reactive genotoxins: recommendations for data
analysis. Mutat. Res. 678, 95–100.
(29) Bartsch, H., and Nair, J. (2000) Ultrasensitive and specific detection
methods for exocyclic DNA adducts: markers for lipid peroxidation
and oxidative stress. Toxicology 153, 105–114.
(30) Manderville, R. A., and Pfohl-Leszkowicz, A. (2008) Bioactivation
and DNA adduction as a rationale for ochratoxin A carcinogenesis.
World Mycotoxin J. 1, 357–367.
(31) Tozlovanu, M., Faucet-Marquis, V., Pfohl-Leszkowicz, A., and
Manderville, R. A. (2006) Genotoxicity of the hydroquinone metabolite
of ochratoxin A: Structure-activity relationships for covalent DNA
adduction. Chem. Res. Toxicol. 19, 1241–1247.
(32) Mantle, P. G., McHugh, K. M., Adatia, R., Gray, T., and Turner, D. R.
(1991) Persistent karyomegaly caused by Penicillium nephrotoxins in
the rat. Proc. R. Soc. London, Ser. B 246, 251–259.
TX900295A