Induction of micronuclei in mouse and rat by glycidamide, genotoxic metabolite of acrylamide

Article abstract of DOI:10.1016/S1383-5718(02)00281-4

Male CBA mice and male Sprague-Dawley rats were treated by i.p. injection of glycidamide (GA), the presumed genotoxic metabolite of acrylamide (AA). GA was obtained through a new way of synthesis. As an endpoint of chromosome damage, micronucleus (MN) induction in erythrocytes was measured. Hemoglobin (Hb) adducts were used as a measure of in vivo dose of GA. GA induced linear dose-dependent increases in adduct levels in both species. Rats exhibit, compared with mice, 30% higher Hb adduct levels per unit of administered amount of GA. The incremental MN frequencies per administered dose of GA in mice showed a linear-quadratic dose-dependent curve. In the rat no positive dose-response relationship was obtained, probably due to toxic effects to the bone marrow. The main result of this study is the finding that after treatment with synthetic GA the MN frequency per unit of the in vivo dose of GA in the mouse is very similar to that obtained in a previous study, where animals were treated with AA and GA as a metabolite. This equality in potency of GA, whether its in vivo dose is established by injection of synthetic GA or through metabolism of AA, supports the view that GA is the predominant genotoxic factor in AA exposure.

Full text of DOI:10.1016/S1383-5718(02)00281-4

Mutation Research 535 (2003) 15–24
Induction of micronuclei in mouse and rat by glycidamide,
genotoxic metabolite of acrylamide
Birgit Paulssona, Natalia Kotovab,c, Jan Grawéb,1, Alistair Hendersond,
Fredrik Granathe, Bernard Goldingd, Margareta Törnqvista,∗
a
Department of Environmental Chemistry, Stockholm University, S-10691 Stockholm, Sweden
b
Department of Genetic and Cellular Toxicology, Stockholm University, S-10691 Stockholm, Sweden
c
Department of Radiobiology and Environmental Medicine, International Sakharov Environmental University, Minsk, Belarus
d
Department of Chemistry, University of Newcastle, Bedson Building, Newcastle NE17RU, UK
e
Department of Medical Epidemiology, Karolinska Institutet, S-17177 Stockholm, Sweden
Received 23 May 2002; received in revised form 27 September 2002; accepted 14 October 2002
Abstract
Male CBA mice and male Sprague–Dawley rats were treated by i.p. injection of glycidamide (GA), the presumed genotoxic
metabolite of acrylamide (AA). GA was obtained through a new way of synthesis. As an endpoint of chromosome damage,
micronucleus (MN) induction in erythrocytes was measured. Hemoglobin (Hb) adducts were used as a measure of in vivo
dose of GA. GA induced linear dose-dependent increases in adduct levels in both species. Rats exhibit, compared with mice,
30% higher Hb adduct levels per unit of administered amount of GA. The incremental MN frequencies per administered
dose of GA in mice showed a linear-quadratic dose-dependent curve. In the rat no positive dose–response relationship was
obtained, probably due to toxic effects to the bone marrow. The main result of this study is the finding that after treatment with
synthetic GA the MN frequency per unit of the in vivo dose of GA in the mouse is very similar to that obtained in a previous
study, where animals were treated with AA and GA as a metabolite. This equality in potency of GA, whether its in vivo
dose is established by injection of synthetic GA or through metabolism of AA, supports the view that GA is the predominant
genotoxic factor in AA exposure.
© 2002 Elsevier Science B.V. All rights reserved.
Keywords: Glycidamide; Acrylamide; Micronuclei; Hemoglobin adducts; Carcinogenic
1. Introduction
clastogenic and is known to be carcinogenic in ani-
mal experiments and, therefore, classified as a prob-
able human carcinogen [1]. Glycidamide (GA) is a
metabolite of AA formed through metabolic transfor-
mation [1]. AA is metabolized to GA by cytochrome
P450 enzymes (Fig. 1), as shown by Sumner et al. [2],
by comparing metabolism of AA in wild-type mice,
mice pretreated with a P450 inhibitor and mice devoid
of cytochrome P450 2E1. AA has shown low reactiv-
ity towards DNA in vitro [3] and in AA-treated mice
and rats only DNA adducts from the metabolite GA
Acrylamide (AA) is a vinylic compound used as
monomer in the synthesis of polyAA products for, e.g.
grouting and soil stabilization. AA is neurotoxic and
∗
Corresponding author. Tel.: +46-8-163769;
fax: +46-8-152561.
E-mail address: margareta.tornqvist@mk.su.se (M. Törnqvist).
1
Present address: The Cell Analysis Core Facility, The Rudbeck
Laboratory, Uppsala University, S-75105 Uppsala, Sweden.
1383-5718/02/$ – see front matter © 2002 Elsevier Science B.V. All rights reserved.
PII: S1383-5718(02)00281-4

16
B. Paulsson et al. / Mutation Research 535 (2003) 15–24
eral blood and in rat bone marrow. In vivo GA doses
were measured as levels of GA adducts to N-terminal
valine in Hb. From a measured level of a specific, sta-
ble adduct to a macromolecule with known turnover
kinetics the in vivo dose can be calculated. The in
vivo dose implicates the net effects of all enzymatic,
chemical and physiological factors on rates of for-
mation and disappearance of a reactive chemical or
metabolite. The in vivo dose is, if defined as the
concentration in tissues integrated over time [10],
proportional to the frequencies of mutations [11].
Fig. 1. AA metabolized in vivo (by CYP 2E1 enzymes) forming
GA.
have been detected [4]. Therefore, GA is assumed to
be the genotoxic agent in AA exposure [4,5].
In cancer tests mice are more sensitive with re-
gard to tumor induction per unit of absorbed lifetime
dose of AA, compared to rats [1,6]. Furthermore, the
mouse is more effective than the rat in metabolizing
AA to GA, as shown by Sumner et al. [7] analyzing
urinary metabolites from rats and mice following AA
administration, and by Paulsson et al. [8] measuring
hemoglobin (Hb) adducts from AA and GA in mice
and rats after AA treatment. These observations were
taken by Paulsson et al. as a support of the applica-
bility of the multiplicative cancer risk model for AA
[5]. This model implies that the increase in cancer risk
of a genotoxic chemical is proportional to the in vivo
dose of the genotoxic compound, in this case GA, and
the background tumor frequency [9].
2. Materials and methods
2.1. Laboratory animals
Male CBA mice, 8 weeks old, weighing about
25 g, and male Sprague–Dawley rats, 8 weeks old and
weighing about 300 g, were used. The mice were sup-
plied from B&K Universal AB, Sollentuna, Sweden,
and the rats from Charles River Sverige AB, Uppsala,
Sweden. The animals were kept in the animal house at
Stockholm University, where they were allowed free
access to solid food and tap water. The study has been
reviewed and approved by the Ethical Committees on
Animal Experiments of Stockholm, Sweden.
This study is part of a project aiming at an improved
cancer risk assessment of AA. The study is focused on
the genotoxicity of GA per dose unit in mice and rat.
The level of Hb adducts of GA is used as a measure
of in vivo dose, while micronucleus (MN) frequen-
cies in erythrocytes are used as a measure of chromo-
some damage. In a previous study mice and rats were
treated with AA, MN frequencies were measured and
Hb adduct levels were determined from AA and the
metabolite GA. MN induction was, in that study, ob-
served only in mouse. This was suggested to be a re-
sult of the more effective metabolic conversion of AA
to GA in the mouse, compared with the rat, leading to
higher in vivo doses of GA per unit of administered
dose of AA in the mouse [8].
The objectives of the present study were to study
the genotoxic potency per in vivo dose of GA, for
further elucidation of the role of GA in the genotoxic
and carcinogenic effect of AA. The genotoxic effects
were studied using a flow cytometric in vivo MN
assay, measuring the frequencies of micronucleated
polychromatic erythrocytes (MPCE) in mouse periph-
2.2. Chemicals
AA (CAS no. 79-06-1) was obtained from Merck
(acrylamide for electrophoresis, 99%), oxone (potas-
sium monopersulfate) (CAS no. 37222-66-5) was
obtained from Sigma–Aldrich and pentafluorophenyl
isothiocyanate (CAS no. 35923-79-6) was obtained
from Fluka (purum, >96%, purified on a Sep-Pak sil-
ica cartridge [12]). All other chemicals and solvents
used were of analytical grade.
GA (CAS no. 5694-00-8) was synthesized through
reaction between AA and dimethyldioxirane (DMD)
(Fig. 2). Firstly, DMD was synthesized by separately
adding oxone (200 g) and water/acetone (500 ml of
420/320 ml mixture) to a vigorously stirred solu-
tion of NaHCO3 (100 g) in water/acetone (240 ml of
420/320 ml mixture) over a 3 h period, at a reduced
air pressure (730 mmHg) [13]. DMD in acetone was
collected over a 2.5 h period. After titration of the
distillate, AA (1 mol eq.) was added to DMD in

B. Paulsson et al. / Mutation Research 535 (2003) 15–24
17
For mice, peripheral blood PCE were studied for
MN induction, while for rats bone marrow PCE were
used. Peripheral blood is better suited for low-dose
studies due to a lower degree of variability, but in rats
the spleen selectively removes micronucleated PCE.
However, we have earlier shown, for both mice and
rats, that within species MN frequencies measured in
peripheral blood and bone marrow are quantitatively
comparable, providing that samples are taken at the re-
spective optimal times for each tissue [14,15]. In this
study, rat bone marrow was sampled 24 h after treat-
ment, and mouse peripheral blood 48 h after treatment,
based on previous studies where these times showed
maximum responses for the respective tissues [8]. The
animals were weighed before and after treatment.
Fig. 2. Racemic GA was synthesized through reaction between
AA and dimethyldioxirane in acetone.
acetone (1.5 mol eq.) for synthesis of GA. The solu-
tion was stirred under N2 at ambient temperature for
3 days. The acetone was removed and the product
was purified by distillation. The purity of the GA was
1
confirmed by NMR analysis: H NMR (d6-DMSO)
2.75 (1H, dd, J 6.3 and 2.4 Hz, CH), 2.89 (1H, dd, J
6.3 and 4.5 Hz, CH), 3.29 (1H, dd, J 4.3 and 2.6 Hz,
CH), 7.34 (1H, b, s, NH) 7.49 (1H, b, s, NH).
2.4. Hb adduct analysis by the N-alkyl
Edman method
2.3. Animal treatment
One of the major reactive sites in Hb of human
and rodents is the N-terminal valines, and many com-
pounds, for instance AA and GA [16] can be moni-
tored in vivo, through their adducts to the N-termini by
the N-alkyl Edman method [17,18]. After derivatiza-
tion with the Edman reagent, pentafluorophenyl isoth-
iocyanate (PFPITC), the adduct (the N-substituted
N-terminal valine) is detached from the globin as
an Edman derivative, the N-alkylvaline-pentafluoro-
phenylthiohydantoin (alkyl-Val-PFPTH). The deriva-
tive is then analyzed by gas chromatography–mass
spectrometry (GC–MS).
Red blood cells were separated from the plasma
and the globin precipitated [19]. Samples of globin
were prepared according to Törnqvist et al. [20]
with modifications according to Bergmark [21] and
analyzed by GC–MS/MS after acetonization of the
GA-Val-PFPTH derivative, according to Paulsson
et al. [8] and Paulsson et al. (submitted). The limit of
detection was about 10 pmol/g globin (40 mg globin
samples) and the coefficient of variation for duplicate
preparation (on different occasions) and measurement
was between 5 and 16%.
GA was dissolved in 0.9% saline under sterile con-
ditions prior to the treatment of the animals. All treat-
ments were done by single i.p. injection. The injec-
tion volume was for mice 10 ␮l/g, and for rats 8 ␮l/g.
A pilot study was performed where two animals per
species, treated with 0.7 mmol/kg body weight, and
one control animal per species, were analyzed by the
MN assay (see below). Both species showed, at this ad-
ministered dose, a decrease in the percentage of PCE,
an indication of toxic effects to the bone marrow. The
decrease was more pronounced in the mice (∼40%)
than in the rats (∼20%) (see also Section 3). There-
fore, it was decided that in the main study, 0.7 mmol/kg
body weight should be the highest administered dose
to mice. For the rats, however, the highest adminis-
tered dose was set to 1.4 mmol/kg body weight, with
expectations of obtaining higher MN frequencies.
In the main study three animals per administered
dose and four controls of each species were used.
The mice were treated with three doses (0.18, 0.35,
and 0.70 mmol/kg body weight). For the rats only
two different doses (0.70 and 1.4 mmol/kg body
weight) were used, in order to minimize the need
for animals and for GA. The last rat treated, sched-
uled for 1.4 mmol/kg body weight, was given only
0.90 mmol/kg body weight due to insufficient amount
of substance available.
2.5. Flow cytometric MN assay
The procedure for sampling, fixation, staining and
flow cytometric analysis of erythrocytes from mouse
peripheral blood and rat bone marrow, was the same as

18
B. Paulsson et al. / Mutation Research 535 (2003) 15–24
earlier reported [8,15,22]. Essentially, after anestheti-
zation about 100–200 ␮l of mouse peripheral blood
was drawn from orbital veins into heparinized tube
using a Pasteur pipette. Triplicate samples of 6 ␮l
blood were layered on 1 ml of 65% Percoll (Amer-
sham Pharmacia Biotech AB, Uppsala, Sweden) so-
lution in phosphate buffered saline (PBS) and cen-
trifuged for 20 min at 600×g. Rats were anesthetized
and sacrificed and, by a syringe, bone marrow from
both femurs was flushed into a conical tube (4.5 ml;
12/75 mm, Greiner, Kebo, Spånga, Sweden) with 2 ml
of RPMI-medium (Gibco, UK), centrifuged and resus-
pended in 120 ␮l PBS. Triplicate samples of 30 ␮l rat
bone marrow suspension were layered on 1 ml of 65%
Percoll solution in PBS, as above, and centrifuged.
Supernatants from mouse as well as rat samples were
carefully aspirated, leaving a pellet of erythrocytes and
some nucleated cells. The pellets obtained were sus-
pended in 30 ␮l Sörensen’s phosphate buffer, 0.05 M,
pH 6.8, with 1.5 mg/ml sodium dodecyl sulfate (SDS;
Sigma, St. Louis, MO, USA) included. After overnight
storage at 4 ◦C, the fixative was removed and staining
buffer was added under gentle vortexing. The stain-
ing buffer was prepared by adding 500 ␮l Hoechst
33342 stock (HO342; Molecular Probes, Pitchford,
OR, USA; 500 mM in distilled water) and 50 ␮l Thi-
azole Orange stock (TO; Molecular Probes, Eugene,
USA; 1 mg/ml in methanol) to 100 ml PBS. The sam-
ples were stained with 1 ml staining buffer during 1 h,
in the darkness, mixed every 15 min. Afterwards, the
flow cytometric analysis and calculation of PCE and
micronucleated PCE (MPCE) frequencies were per-
formed as described by Grawé et al. [22]. The coeffi-
cient of variation for the MN assay (pooled samples)
was 14%.
on the basis of the average adduct level of all ani-
mals in each dose group. The hypothesis of the same
linear-quadratic dose–response curve independent of
mode of exposure was tested by a likelihood ratio test.
3. Results
3.1. Synthesis of GA
Previously, GA has been prepared by reaction of AA
with alkaline hydrogen peroxide under strictly con-
trolled conditions of pH [24,25]. Our way of synthe-
sizing GA, via direct epoxidation of AA with DMD
avoiding the involvement of peroxides, has the advan-
tage that the labile GA can be isolated simply by evap-
oration of the reactive material and solvent. The GA
was obtained essentially pure (95%) and its 1H NMR
was fully consistent with the structure and with pre-
viously reported data [24,25].
3.2. Hb adduct levels
Analysis of GA adducts in Hb in blood sam-
ples from GA-treated rats and mice showed linearly
dose-dependent adduct level increments that were
significantly different in the two species (P = 0.003)
(Fig. 3). The increase in adduct level per unit dose
of GA (in mmol/kg body weight) were 82 nmol/g
globin (95% CI: 76–90) for rats and 61 nmol/g globin
(95% CI: 55–67) for mice. Thus, the adduct lev-
els in rats were 32% (95% CI: 12–65) higher per
administered dose of GA than the levels in mice.
Average adduct level in control animals was for mice
0.036 nmol/g globin (S.D. = 0.004) (n = 4) and for
rats 0.012 nmol/g globin (S.D. = 0.003) (n = 4).
2.6. Statistical methods
3.3. The MN test
Relations between adduct levels and administered
doses as well as between MN frequencies and in
vivo doses were assessed by linear or linear-quadratic
regression, using SAS® software [23]. Calculated
parameters are presented together with either 95%
confidence interval (CI), standard deviation (S.D.)
or P-value. The responses of the MN frequencies as
a function of the in vivo dose of GA obtained by
i.p. injection of synthetic GA or as a metabolite of
AA (data from an earlier study [8]), were calculated
The result from the MN test of peripheral blood
from mice showed a linear-quadratic increase in the
frequency of MPCE with administered dose of GA.
Combined results from the pilot study and the main
study are shown in Fig. 4.
The results from the MN test of rat bone marrow
showed a non-monotone dose–response. At the admin-
istered doses 0.7 and 0.9 mmol/kg body weight there
was a weak, but significant (P = 0.001 for combined

B. Paulsson et al. / Mutation Research 535 (2003) 15–24
19
Fig. 3. GA adduct levels in rat and mouse after GA treatment.
results from the pilot and main studies, P = 0.02 for
results from the main study), increase in the frequency
of MPCE while at the highest given dose, 1.4 mmol/kg
body weight, the MPCE frequencies were approxi-
mately similar to the background values. In Fig. 5
results from the main study and the pilot study are
shown.
In both mice and rats the frequency of PCE was
monitored as a measure of bone marrow toxicity im-
pairing cell proliferation. In both species the treated
animals showed, in comparison to the control ani-
mals, a decrease in the PCE frequency. In the mice the
decrease at the administered dose 0.7 mmol/kg body
weight was about 40%, the same as in the pilot study
Fig. 4. MN frequencies, measured as frequencies of MPCE, in mouse after GA treatment. Combined results from the pilot study and the
main study.

20
B. Paulsson et al. / Mutation Research 535 (2003) 15–24
Fig. 5. MN frequencies, measured as frequencies of MPCE, in rat after GA treatment. Combined results from the pilot study and the main
study.
(Fig. 6). In the rats, however, the decrease was more
pronounced in the main study than in the pilot study.
For rats given 0.7 mmol/kg body weight the decrease
in PCE frequency was about 40%, compared with 20%
in the pilot study. At higher administered doses, the
PCE frequencies in rats were as low as one third of
the background values (Fig. 7). It is known from our
earlier work that while MPCE frequencies are highly
reproducible between experiments, PCE frequencies
are more variable. The results show that in mice the
increase in MN is apparent even at doses producing
no or little bone marrow toxicity. In the rat, however,
the modest increase in MN is observed at doses with
pronounced bone marrow toxicity. This indicates that
Fig. 6. Frequencies of PCE, an indication of toxic effect, in mouse after GA treatment. Results from both the pilot study and the main study.

B. Paulsson et al. / Mutation Research 535 (2003) 15–24
21
Fig. 7. Frequencies of PCE, an indication of toxic effect, in rat after GA treatment. Results from both the pilot study and the main study.
the measured MN frequencies in rats may be underes-
timated due to reduced proliferation in the bone mar-
row, particularly at the highest administered doses of
GA.
4. Discussion
In both mice and rats GA induced linearly dose-
dependent increases in Hb-adduct levels (Fig. 3). The
GA adduct levels (reflecting in vivo doses) in rats were
significantly (32%) higher per administered dose than
in mice. This implies that the detoxification of GA is
faster in the mouse than in the rat. This conclusion
is based on the assumption that the reactivity of GA
toward N-terminal valine is approximately the same
in rats and mice (shown, e.g. for ethylene oxide by
Granath et al. [9].
The flow cytometric method used for the determi-
nation of MN frequencies is very stable over time,
as shown both in studies involving several repeat
samplings over a period of more than 3 months [26],
and when comparing identical exposures and ani-
mal strains in experiments made several years apart
[27,28]. Thus, experiments done several months apart
may be quantitatively compared. Here we have uti-
lized this reproducibility over time to combine the
results of pilot and main experiments with GA, and
in order to compare two ways of establishing GA
doses in vivo (AA treatment and GA treatment). The
combination of the GA data in the present study is
of advantage especially in the study of rats, where a
small number of animals were used.
3.4. Body weight
The animals were weighed before and after the
treatment in the main study and the result showed
that the rats were, in contrast to the mice, loosing
body weight during the treatment (Table 1). The
treated rats lost between 2 and 8% of their body
weight 24 h after treatment. This is seen as a symp-
tom of systemic effects caused by the treatment of
GA.
Table 1
Average difference in body weight for each dose group of mice
and rats after GA treatment
Administered dose
(mmol/kg)
n
Weight
difference—mice
Weight
difference—rats
g
%
g
%
Control
0.18
0.35
0.7
0.9
1.4
4
4
3
3
1
2
0.2
0.0
0.4
0.1
–
0.8
0.0
1.3
0.3
–
0.5
–
–
0.2
–
–
−2
−5
−8
−7.7
−14
−26
The incremental MN frequencies per adminis-
tered dose of GA in mice were compatible with
–
–

22
B. Paulsson et al. / Mutation Research 535 (2003) 15–24
Fig. 8. MN frequency per unit of in vivo GA dose in mouse after AA treatment (the previous study) as well as after GA treatment (the
present study). On the x-axis are the measured GA adduct levels (mean values from the different dose-groups) and on the y-axis the
frequency of MPCE (individual values).
linear-quadratic dose-dependence (Fig. 4). In the rats,
no positive dose–response relationship was obtained,
there was a weak increase in MN frequency at the
two lowest but no increase at the highest administered
GA dose (Fig. 5). Adduct levels (in vivo doses) of
GA were, however, higher in rats than in mice over
the whole range of administered doses.
At the highest administered doses the rats showed
symptoms of systemic toxicity, decreased erythrocyte
proliferation in the bone marrow together with loss
of body weight. The low MN frequencies at these
administered doses of GA in rats are probably due
to these toxic effects. The genotoxic effect of GA in
rat, measured by MN frequencies in the bone marrow
PCE, may have been overshadowed by the systemic
effects caused by GA. The MN assay, therefore, does
not seem to be suitable for the measurement of geno-
toxic effects caused by GA in the rat. For performing
a comparison of genotoxic potency of GA in mice and
rats another assay for measuring genotoxic endpoints
should be employed.
levels) similar to those in the present study. A com-
parison between the present and the preceding studies
with regard to MN frequency per unit of in vivo dose
of GA in mice, shows a similar response to the in
vivo GA dose whether it results from metabolism of
AA or from administered synthetic GA (Fig. 8). The
shapes of the two dose response curves are indicated
to be different, however. The curve corresponding to
the MN frequency per in vivo dose of GA after AA
treatment is linear while the curve corresponding to
the MN frequency per in vivo dose of GA after GA
treatment shows a linear-quadratic shape. The latter
was also the case for the curve corresponding to MN
increase per administered GA dose (Fig. 4).
The regression coefficients when fitting a simple
linear or a linear-quadratic dose response curve to
the observed MN frequencies as a function of GA
adduct levels in mice are shown in Table 2. As can be
seen the linear model yields almost identical slopes.
However, when fitting a linear-quadratic model, the
same linear-quadratic model does not apply to the
two modes of exposure (P = 0.02). In fact, the re-
sponse after AA treatment is almost purely linear,
while the response after GA treatment is almost purely
quadratic. This may reflect a true difference in the re-
sponse due to the way of exposure but one should not,
In the preceding study of mice and rats treated with
AA, Hb adducts from the parent compound AA and
the metabolite GA were measured together with MN
frequencies [8]. The AA treatment of mice in that
study led to in vivo GA doses (measured as GA adduct

B. Paulsson et al. / Mutation Research 535 (2003) 15–24
23
Table 2
Regression coefficients when fitting the same simple linear model or the same linear-quadratic model to the observed MN frequencies as
a function of measured GA adduct level in mice, as shown in Fig. 8
Parameters
AA treatment
GA treatment
Regression coefficient
P
Regression coefficient
P
Linear model
Background (intercept)
Linear (slope)
1.26
0.066
1.27
0.062
<0.0001
<0.0001
Linear-quadratic modela
Background (intercept)
Linear term
1.32
0.049
0.0005
1.59
−0.005
0.0016
0.004
0.31
0.69
<0.0001
Quadratic term
a The hypothesis of identical linear-quadratic dose–response independent of mode of exposure, can be rejected (P = 0.02).
at this stage, absolutely exclude unknown systematic
errors.
sential role in the induction of cell transformation
by AA. The authors collect a number of arguments
against a causative action of reaction with DNA of
the metabolite GA in the genotoxic effects of AA. In
this context measurements of doses as in the present
study, are most useful for the elucidation of genotoxic
mechanisms.
It should be noted that the agreement between the
two studies with different treatments includes that, in
the previous study, GA is formed through metabolism
of AA and that in the present study, the in vivo GA
dose is received through i.p. injection of synthetic GA.
This fact will have an impact on the dose rate, i.e.
the concentration, of GA in vivo, in the two studies.
A higher dose rate (as obtained by i.p. injection of
the genotoxic synthetic GA) might, for instance, influ-
ence the rate of DNA repair. Further studies would be
needed for a clarification of the discrepancy between
the two studies.
To conclude, the results from the present study in
combination with the preceding study show that the
increase in MN in AA-treated mice could be ascribed
to the formation of the genotoxic metabolite GA. It
seems that the in vivo doses of AA in the previous
study do not contribute to the observed genotoxic
effects from AA treatment. The results support the
hypothesis that it is the metabolite GA that is the pre-
dominant genotoxic agent following exposure to AA.
A similar comparison for rats could not be made
since, in the previous study AA-treated rats exhibited
no MN increase (most probably due to too low in vivo
GA doses); in the present study we could not estimate
any dose response curve for MN increase per in vivo
dose of GA (due to the probably underestimated MN
frequencies in the rats given the highest administered
dose of GA).
Acknowledgements
The authors gratefully acknowledge Prof. Em. Lars
Ehrenberg for valuable discussions during this work.
Thanks are also due to Dr. Emma Bergmark for provid-
ing reference compounds, to Mr. Ioannis Athanasiadis
for skillful technical assistance in mass spectrometric
analysis and to M.Sc. Christina Romare for performing
the NMR analysis. We acknowledge economic support
from the Swedish Council for Work Life and Social
Research and the Swedish Research Council for Envi-
ronmental Agricultural Science and Spatial Planning.
References
[1] IARC Monographs on the evaluation of carcinogen risk to
humans: some industrial chemicals, No. 60, International
Agency for Research on Cancer, Lyon, 1994.
[2] S.C.J. Sumner, T.R. Fennell, T.A. Moore, B. Chanas, F.
Gonzalez, B.I. Ghanayem, Role of cytochrome P450 2E1
in the metabolism of acrylamide and acrylonitrile in mice,
Chem. Res. Toxicol. 12 (1999) 1110–1116.
[3] J.J. Solomon, J. Fedyk, F. Mukai, A. Segal, Direct alkylation
of 2ꢀ-deoxynucleosides and DNA following in vitro reaction
with acrylamide, Cancer Res. 45 (1985) 3465–3470.
In a recent paper Park et al. [29] presented strong
indications of glutathione depletion playing an es-

24
B. Paulsson et al. / Mutation Research 535 (2003) 15–24
Thesis, Department of Radiobiology, Stockholm University,
[4] D. Segerbäck, C.-J. Calleman, J.L. Schroeder, L.G.
Costa, E.M. Faustman, Formation of N-7-(2-carbamoyl-2-hy-
droxyethyl)guanine in DNA of the mouse and the rat
following intraperitoneal administration of [14C]acrylamide,
Carcinogenesis 16 (1995) 1161–1165.
Stockholm, 1992.
[17] P. Rydberg, B. Lüning, C.A. Wachtmeister, L. Eriksson, M.
Törnqvist, Applicability of a modified Edman procedure for
measurement of protein adducts: mechanisms of formation
and degradation of phenylthiohydantoins, Chem. Res. Toxicol.
15 (2002) 570–581. 2002.
[18] M. Törnqvist, C. Fred, J. Haglund, H. Helleberg, B. Paulsson,
P. Rydberg, Protein adducts: quantitative and qualitative
aspects of their formation, analysis and applications, J.
Chromatogr. B 778 (2002) 279–308.
[19] J. Mowrer, M. Törnqvist, S. Jensen, L. Ehrenberg, Modified
Edman degradation applied to hemoglobin for monitoring
occupational exposure to alkylating agents, Toxicol. Environ.
Chem. 11 (1986) 215–231.
[20] M. Törnqvist, J. Mowrer, S. Jensen, L. Ehrenberg, Monitoring
of environmental cancer initiators through hemoglobin
adducts by a modified Edman degradation method, Anal. Bio-
chem. 154 (1986) 255–266.
[21] E. Bergmark, Hemoglobin adducts of acrylamide and acrylo-
nitrile in laboratory workers, smokers, and nonsmokers,
Chem. Res. Toxicol. 10 (1997) 78–84.
[22] J. Grawé, L. Abramsson-Zetterberg, G. Zetterberg, Low dose
effects of chemicals as assessed by the flow cytometric in
vivo micronucleus assay, Mutat. Res. 405 (1998) 199–208.
[23] SAS/STAT User’s Guide, Version 6, 4th ed., vol. 2, SAS
Institute Inc., Cary, NC, 1989.
[24] G.B. Payne, P.H. Williams, Reactions of hydrogen peroxide.
VI. Alkaline epoxidation of acrylonitrile, J. Org. Chem. 26
(1961) 651–695.
[25] S. Sugiyama, S. Ohigashi, R. Sawa, H. Hayashi, Selective
preparation of 2,3-epoxypropanamide and its facile conversion
to 2,3-dihydroxypropanamide with acidic resins, Bull. Chem.
Soc. Jpn. 62 (1989) 3202–3206.
[5] B. Paulsson, F. Granath, J. Grawé, L. Ehrenberg, M.
Törnqvist, The multiplicative model for cancer risk assess-
ment: applicability to acrylamide, Carcinogenesis 22 (2001)
817–819.
[6] M. Törnqvist, E. Bergmark, L. Ehrenberg, F. Granath,
Risk assessment of acrylamide, Report PM 7-98 National
Chemicals Inspectorate, Solna, Sweden, 1998 (in Swedish).
[7] S.C.J. Sumner, L. Selvaraj, S.K. Nauhaus, T.R. Fennell,
Urinary metabolites from F344 rats and B6C3F1 mice
coadministered acrylamide and acrylonitrile for 1 or 5 days,
Chem. Res. Toxicol. 10 (1997) 1152–1160.
[8] B. Paulsson, J. Grawé, M. Törnqvist, Hemoglobin adducts and
micronucleus frequencies in mouse and rat after acrylamide
or N-methylolacrylamide treatment, Mutat. Res. 516 (2002)
101–111.
[9] F. Granath, C. Vaca, L. Ehrenberg, M. Törnqvist, Cancer risk
estimation of genotoxic chemicals based on target dose and
a multiplicative model, Risk Anal. 19 (1999) 309–319.
[10] L. Ehrenberg, E. Moustacchi, S. Osterman-Golkar, G. Ekman,
Dosimetry of genotoxic agents and dose–response relation-
ships of their effects, Mutat. Res. 123 (1983) 121–182.
[11] H. Helleberg, H. Xu, L. Ehrenberg, K. Hemminki, U. Rannug,
M. Törnqvist, Studies of dose distribution, premutagenic
events and mutation frequencies for benzo[a]pyrene aiming
at low dose cancer risk estimation, Mutagenesis 16 (2001)
333–337.
[12] M. Törnqvist, Epoxide adducts to N-terminal valine of
hemoglobin, in: J. Everse, K.D. Vandegriff, R.W. Winslow
(Eds.), Methods Enzymol, 1994, pp. 650–657.
[13] C. Bleasdale, B.T. Golding, S.L. Kitson, Preparation of
ethenyl and oxiranol carbamates and their (2-2H2)-derivatives
using selenium chemistry; thermolysis of oxiranol carbamate
and reaction with adenosine, Bioorg. Med. Chem. Lett. 4
(1994) 291–296.
[14] L. Abramsson-Zetterberg, J. Grawé, G. Zetterberg, Erythro-
poiesis and the induction of micronuclei in mouse spleen
determined by flow cytometry, Mutat. Res. 394 (1997) 17–28.
[15] L. Abramsson-Zetterberg, J. Grawé, G. Zetterberg, The
micronucleus test in rat erythrocytes from bone marrow,
spleen and blood: the response to low doses of ionising
radiation, cyclophosphamide and vincristine determined by
flow cytometry, Mutat. Res. 423 (1999) 113–124.
[26] G. Zetterberg, J. Grawé, Flow cytometric analysis of micro-
nucleus induction in mouse erythrocytes by gamma-irradia-
tion at very low dose-rates, Int. J. Radiat. Biol. 64 (1993)
555–564.
[27] J. Grawé, G. Zetterberg, H. Amneus, Effects of extended
low-dose rate exposure to 137Cs detected by flow-cytometric
enumeration of micronucleated erythrocytes in mouse peri-
pheral blood, Int. J. Radiat. Biol. 63 (1993) 339–347.
[28] L. Abramsson-Zetterberg, J. Grawé, G. Zetterberg, Sponta-
neous and radiation-induced micronuclei in erythrocytes from
four species of wild rodents: a comparison with CBA mice,
Mutat. Res. 393 (1997) 55–71.
[29] J. Park, L.M. Kamendulis, M.A. Friedman, J.E. Klaunig,
Acrylamide-induced cellular transformation, Toxicol. Sci. 65
(2002) 177–183.
[16] E. Bergmark, Hemoglobin dosimetry and comparative toxi-
city of acrylamide and its metabolite glycidamide, Ph.D.