Mutagenesis vol.18 no.2 pp.119–126, 2003
The principal phenolic and alcoholic components of wine protect
human lymphocytes against hydrogen peroxide- and ionizing
radiation-induced DNA damage in vitro
William Greenrod1,2 and Michael Fenech1,2,3
1CSIRO
Health Sciences and Nutrition, PO Box 10041, Adelaide BC, South
Australia 5000, Australia and 2Clinical and Experimental Pharmacology,
The University of Adelaide, Adelaide, Australia
We have tested the hypothesis that the alcoholic and
phenolic components of wine are protective against the
DNA-damaging and cytotoxic effects of hydrogen peroxide
and γ-radiation in vitro. The components of wine tested
were ethanol, glycerol, a mixture of the phenolic compounds
catechin and caffeic acid and tartaric acid, all at concentrations that were 2.5 or 10.0% of the concentration in a
typical Australian white wine (Riesling). These components
were tested individually or combined as a mixture and
compared to a white wine stripped of polyphenols, as well
as a Hanks balanced salt solution control, which was
the diluent for the wine components. The effect of the
components was tested in lymphocytes, using the cytokinesis-block micronucleus assay, after 30 min incubation
in plasma or whole blood for the hydrogen peroxide or
γ-radiation challenge, respectively. The results obtained
showed that ethanol, glycerol, the catechin–caffeic acid
mixture, the mixture of all components and the stripped
white wine significantly reduced the DNA-damaging effects
of hydrogen peroxide and γ-radiation (P ⍧ 0.043–0.001,
ANOVA). The strongest protective effect against DNA
damage by γ-irradiation was observed for the catechin–
caffeic acid mixture and the mixture of all components (30
and 32% reduction, respectively). These two treatments as
well as ethanol produced the strongest protective effects
against DNA damage by hydrogen peroxide (24, 25 and
18%, respectively). The protection provided by the mixture
did not account for the expected additive protective
effects of the individual components. Ethanol was the
only component that significantly increased baseline DNA
damage rate, however, this effect was negated in the
mixture. In conclusion, our results suggest that the main
phenolic and alcoholic components of wine can reduce the
DNA-damaging effects of two important oxidants, i.e.
hydrogen peroxide and ionizing radiation, in this physiologically relevant in vitro system.
Introduction
Evidence for the protective effect of moderate consumption
of alcoholic beverages on cardiovascular disease was provided
initially by the epidemiological study of St Leger et al. (1979).
Later, ‘The French Paradox’ was coined to describe the low
incidence of cardiovascular disease in specific populations
despite elevated levels of risk, including high fat diets, sedentary lifestyles and prevalent cigarette smoking (Renaud and De
Lorgeril, 1993), and was hypothesized to be explained by their
moderate consumption of wine. Moderate alcohol intake is
1To
also associated with a reduced risk of mortality from cancer,
however, risk of cancer and other diseases increases progressively with excessive alcohol intake (Boffeta and Garfinkel,
1990). A study from eastern France reported a reduced risk of
death of 22% from all cancers in those people drinking up to
21 g ethanol/day (the majority in the form of wine, 82% of
total ethanol consumption) when compared with non-drinkers
(Renaud et al., 1998). A plausible explanation for these
observations is that wine or some of its components may
prevent DNA damage events caused by reactive oxygen species
that are implicated in the initiation events of cancer and
possibly atherosclerosis (DeFlora et al., 1997; Andreassi et al.,
2000; Ferguson, 2001; Ghiselli et al., 1998; Casalini et al.,
1999). A common link between cardiovascular disease and
cancer is the observation that elevated micronucleus frequency
(a marker of DNA damage) is associated with increased
severity of coronary artery disease (Botto et al., 2002).
We have previously shown that the genotoxic and cytotoxic
effects of reactive oxygen species can be efficiently measured
using the cytokinesis-block micronucleus (CBMN) assay following challenge by hydrogen peroxide, superoxide, activated
neutrophils and/or ionizing radiation (Fenech and Morley,
1986; Fenech et al., 1997; Umegaki and Fenech, 2000). The
assay, in its comprehensive mode, also allows measurement
of nucleoplasmic bridges between the two nuclei (an index of
chromosome rearrangement), necrosis, apoptosis and cytostasis
(i.e. inhibition of the extent and frequency of nuclear division)
(Fenech, 2000).
Pilot ex vivo experiments in our laboratory demonstrated a
reduction in hydrogen peroxide-induced genetic damage levels
at 1–3 h post-consumption of 300 ml of red or white wine,
with strongest inhibition (⬎70%) at the 1 h post-consumption
time point (Fenech et al., 1997). In vitro experiments in that
study showed that lymphocytes incubated in either 12% ethanol
solution or phenolic-stripped white wine prior to hydrogen
peroxide challenge resulted in a statistically significant 57–
87% reduction (P ⫽ 0.0354 and P ⫽ 0.0053, respectively) in
micronucleated (MNed) cell frequency at a concentration of
10% v/v in culture medium. These results of the in vitro
and ex vivo studies suggested that both the non-phenolic
components as well as the phenolic compounds of wine may
exert antioxidant effects. Because the ex vivo experiments did
not demonstrate a difference between red and white wine we
hypothesized that phenolic and/or non-phenolic components
common to red and white wine may be important in prevention
of oxidative DNA damage. Candidate molecules abundant in
red and white wine include alcohol, glycerol, tartaric acid and
the phenolic compounds caffeic acid and catechin (C & C)
[Table I lists their concentration in a typical Australian white
wine (Riesling) from the Clare Valley, South Australia.] The
aim of this study was to determine the relative contribution of
these compounds to the protective effect of wine against DNA
damage by reactive oxygen species. To answer these questions,
whom correspondence should be addressed. Tel: ⫹61 8 8303 8880; Fax:⫹61 8 8303 8899; Email: [email protected]
© UK Environmental Mutagen Society/Oxford University Press 2003 all rights reserved
119
W.Greenrod and M.Fenech
Table I. Concentrations of wine components in plasma or blood prepared for the in vitro experiments, shown relative to those in white wine
Component concentrations
White wine
Ethanol
Glycerol
Tartaric acid
Catechina
Caffeic acida
Mixb
PVPP stripped white winec
Control (HBSS)d
13.4% v/v
6.8 mg/ml
3.0 mg/ml
11.0⫻10–3 mg/ml
78.0⫻10–3 mg/ml
Plasma or blood
10% (1:10 dilution)
2.5% (1:40 dilution)
1.34% v/v
0.68 mg/ml
0.3 mg/ml
1.1⫻10–3 mg/ml
7.8⫻10–3 mg/ml
All above components
10% v/v
10% v/v
0.36% v/v
0.17 mg/ml
7.5⫻10–2 mg/ml
2.75⫻10–4 mg/ml
1.95⫻10–3 mg/ml
All above components
2.5% v/v
aCatechin
and caffeic acid were added together to the blood or plasma.
acid, catechin and caffeic acid.
polypyrollidone (PVPP).
dThe same diluent control (i.e. HBSS at 10% v/v) was used in both experiments testing wine components at the 2.5 and 10% dilutions.
bMix contained all components, i.e. ethanol, glycerol, tartaric
cWhite wine stripped of phenolic compounds using polyvinyl
two models of genetic damage were used: (i) isolated lymphocyte cultures in autologous plasma challenged with hydrogen
peroxide; (ii) lymphocytes in whole blood cultures challenged
with ionizing radiation. Effects of the wine components on
baseline and induced DNA damage and cell death were
determined using the comprehensive CBMN assay.
Materials and methods
Volunteers
The study was advertised to male students by Email after approval by
the CSIRO Health Sciences and Nutrition Human Experimentation Ethics
Committee and The University of Adelaide Human Research Ethics
Committee. Volunteers were not paid to participate. Four healthy male
volunteers aged 21–23 years were recruited based upon their responses to a
general health, lifestyle and ethanol consumption questionnaire and after
giving informed consent. Those who supplemented their diet with antioxidant
vitamins and/or multi-vitamins and/or smoked and/or drank in excess of
moderate drinking guidelines were excluded from the study. Young adult
males were selected for this study because previous studies have shown that
this group has the lowest baseline micronucleus frequency in the adult
population (Fenech, 1993a,b).
Wine component preparation
The current study examined the individual components ethanol, glycerol,
tartaric acid as well as the mixture C & C at 2.5 and 10% of the concentration
found in white wine. The C & C mixture of phenolic compounds was used
as they represent the most abundant phenolic compounds in white wine. The
wine components were dissolved in Hank’s balanced salt solution (HBSS)
and diluted in culture medium as shown in Table I to achieve final concentrations in vitro equivalent to a 2.5 and 10% dilution of the concentration in
white wine. The 2.5% concentration corresponds approximately to the theoretical average concentration of wine components to be found in the blood of
a 60 kg person after consuming 300 ml of white wine, assuming that 50% of
the components were absorbed and not rapidly excreted from the blood into
other tissues and extracellular fluids. For comparison we used the following
controls, also at corresponding dilutions: HBSS was the negative control and
white wine stripped of phenolic compounds using polyvinyl polypyrollidone
(PVPP) (Sigma Chemical Co, St Louis, MO) was a control for the major nonphenolic fraction of white wine. PVPP-stripped white wine was prepared by
passing white wine (1997 Deakin Estate Chardonnay, Victoria, Australia)
through a column of PVPP (pre-washed with 12% ethanol) under negative
pressure (Somers and Ziemalis, 1985).
Hydrogen peroxide challenge experiments
The basic protocol is outlined in Figure 1. On two occasions an overnight
fasted venous blood sample of 16 ml was collected into lithium–heparincoated Vacuette® tubes (Greiner Labortechnik, Austria). The whole blood was
spun for 10 min at 1400 g and the plasma separated. The pellet was
resuspended in an equal volume of HBSS and fresh lymphocytes were isolated
using Ficoll gradients (Pharmacia Biotech, Sweden). Samples of 475 µl of
plasma were reconstituted with autologous lymphocytes to a concentration of
120
Fig. 1. Protocol for in vitro hydrogen peroxide challenge experiments in
plasma. HBSS, Hanks balanced salt solution; CBMN assay, cytokinesisblock micronucleus assay.
1⫻106 cells/ml. Wine components were added in 25 µl aliquots to their
respective cultures to achieve the final concentrations shown in Table I. All
cultures were incubated at 37°C and 5% CO2 for 30 min. Following this,
15 µl of hydrogen peroxide (Ajax Chemicals, Sydney, Australia) was added
to achieve a final concentration of 750 µM. Cultures were then incubated for
a further 30 min at 37°C and 5% CO2, after which time the cells were washed
in 1 ml of RPMI 1640 (Trace Biosciences, Australia) and resuspended in
500 µl of RPMI 1640 medium supplemented with 10% fetal calf serum (Trace
Biosciences, Australia). Similarly prepared matched unchallenged cultures
with HBSS only were used as controls. Phytohaemagglutinin (PHA) (Murex
Biotech Ltd, Dartford, UK) was added to a final concentration of 22 µg/ml
and the cultures placed in a humidified incubator at 37°C and 5% CO2. Fortyfour hours after mitogen stimulation, cytochalasin B (final concentration
6 µg/ml; Sigma Chemical Co.) was added to accumulate the dividing
lymphocytes at the binucleate phase and 28 h later the cells were harvested
to slides using cytocentrifugation (Shandon, Runcorn, UK). The slides were
air dried, fixed and stained using Diff-Quik (LabAids, Narrabeen, Australia).
All cultures were prepared in duplicate, utilizing single batch reagents to
minimize confounding variation.
γ-Radiation challenge experiments
The protocol is outlined in Figure 2. Briefly, 500 µl of heparinized whole
blood was added to a 70 ml flat-bottomed pot (Sarstedt, Australia), prior to
the addition of wine components. The components were added in 50 µl
aliquots to their respective blood sample. All cultures were incubated at 37°C
for 30 min prior to irradiation. Genetic damage was induced by exposure to
a 137Cs source (Cis Bio IBL 437 C Blood Product Irradiator, dose rate 5.7
Gy/min). All cultures received a dose of 1.5 Gy, a level known to induce an
Wine components protect against DNA damage
Table III. Dose–response effect of hydrogen peroxide on MNed cell
frequency in lymphocytes from whole blood cultures of participating
subjects
Hydrogen peroxide dose
(µM)
Fig. 2. Protocol for in vitro γ-radiation challenge experiments in whole
blood. Gy, radiation dose in Gray; HBSS, Hanks balanced salt solution
control; CBMN assay, cytokinesis-block micronucleus assay.
V1
V2
MNed BN cell frequency/1000 BN cells
0
2
3
256
10
14
512
18
23
1024
21
26
ANOVA
Necrotic cell frequency/1000
BN cells
0
13.6
13.6
256
76.4
150.6
512
249.4
310.2
1024
469.7
462.8
ANOVA
V3
Mean ⫾ SEM
4
8
18
26
3⫾1
11 ⫾ 2
20 ⫾ 2
24 ⫾ 2
P ⬍ 0.001
13.9
177.9
399.6
560.8
14 ⫾ 0
135 ⫾ 30
320 ⫾ 44
498 ⫾ 32
P ⬍ 0.001
V1–V3, volunteer code number.
Table II. Dose–response effect of γ-radiation on MNed cell and NPB
frequency in lymphocytes from whole blood cultures of participating
subjects
γ-Radiation dose (Gy)
V1
V2
MNed BN cell frequency/1000 BN cells
0
0
2
1
127
129
2
516
539
4
820
687
ANOVA
NPBs/1000 BN cells
0
0
0
1
110
57
2
136
130
4
340
229
ANOVA
V3
Mean ⫾ SEM
3
133
482
640
2⫾1
130 ⫾ 2
512 ⫾ 16
716 ⫾ 54
P ⫽ 0.005
0
54
117
217
0⫾0
74 ⫾ 18
128 ⫾ 6
262 ⫾ 39
P ⬍ 0.001
V1–V3, volunteer code number.
~100-fold increase in MNed cell frequency relative to baseline, from preliminary dose–response experiments (Table II). Unirradiated blood samples were
used as controls. The CBMN assay was performed according to the protocol
of Fenech (1993a) with the following modifications for whole blood culture:
4.5 ml of RPMI 1640 culture medium (Trace Biosciences, Australia) supplemented with 10% FCS (Trace Biosciences) was added to each 500 µl of blood
sample. Cultures were kept in a humidified incubator at 37°C and 5% CO2
until PHA (Murex Biotech) was added in a 45 µl aliquot to a final concentration
of 200 µg/ml. The cultures were incubated for 44 h prior to addition of
cytochalasin B (Sigma Chemical Co.) to a final concentration of 6 µg/ml and
were then incubated for a further 28 h. Lymphocytes were harvested by
overlaying onto 1.5 ml Ficoll-Paque (Pharmacia Biotech) in 10 ml conicalbottomed centrifuge tubes (Sarstedt) and centrifuged for 25 min at 400 g at
19°C. The isolated lymphocyte (buffy) layer was transferred to a second 10
ml tube and 3 ml of HBSS was added. After centrifugation at 180 g at 19°C
for 5 min, the supernatant was removed and the pellet resuspended in 3 ml
of HBSS and again the cells were centrifuged for 5 min at 180 g at 19°C.
After removal of the supernatant, the cells were resuspended in 300 µl of
hypotonic solution (20 mg trypsin in a solution of 5 ml of RPMI 1640, 1.5
ml of dimethyl sulphoxide and 13.5 ml of MilliQ water, prepared within 1 h
of addition) and incubated at 37°C for 5 min. Cells were transferred to slides
and stained as described above for the hydrogen peroxide experiments.
Scoring procedure for MNed cells, nucleoplasmic bridges, necrosis and
apoptosis
All slides were coded prior to scoring. A single individual (W.G.) scored all
slides using a Leitz Dialux 20EB light microscope at 1000⫻ magnification
using an oil immersion lens. For the hydrogen peroxide challenge experiments,
the frequency of MNed cells in 1000 binucleated cells of mono-, bi- and
multi-nucleated cells and of necrotic and apoptotic cells was assessed according
to the latest standard procedure (Fenech, 2000). In the γ-radiation challenge
experiments, MNed cell frequency and nucleoplasmic bridge (NPB) frequency
were assessed in 1000 BN cells as markers of chromosome breakage/loss and
rearrangements, respectively. Nuclear division index (NDI) was calculated
from the ratio of the frequency of mono-, bi- and multi-nucleated cells
estimated by scoring 500 cells. It was not possible to reliably score necrotic
and apoptotic cells from whole blood cultures in the γ-irradiation experiment
due to the elimination or altered morphology of these cells by the hypotonic
treatment.
Statistical analysis
The significance of the dose–response effect for hydrogen peroxide and γradiation and the variation in mean values of the end-points scored for each
component treatment relative to the control was evaluated using a General
Linear Model Repeated Measures ANOVA. Parametric tests were used for
analysis because data for the end-points measured were normally distributed
around the mean. Data for cultures treated with hydrogen peroxide and γradiation were analysed separately for observed frequency of genotoxicity/
cytotoxicity biomarkers and also for induced frequency of these biomarkers
after subtracting baseline values in untreated cultures. Post hoc pairwise ttest comparisons were conducted using a least significant difference adjustment
for multiple comparisons. Relationships between variables were assessed using
the Spearman rank correlation. Data are expressed as means ⫾ SEM. Statistical
calculations were performed using SPSS for Windows 10.0.1 statistical
software (© SPSS Inc., 1999) and graphing was performed using Microsoft®
Excel 2002.
Results
Dose–response characteristics for gamma rays and hydrogen
peroxide
A γ-radiation dose–response curve was constructed from 0 to
4 Gy (Table II). Significant dose–response effects were
observed in the means of MNed BN cells and NPB frequency
in BN cells (Pearson correlation coefficient ⫽ 0.986, P ⬍
0.001 and P ⫽ 0.951, P ⬍ 0.001, respectively). From these
data we determined that exposure to 1.5 Gy γ-radiation would
provide a readily measurable increment in DNA damage using
both the MNed cell and NPB biomarkers.
A significant dose–response effect with hydrogen peroxide
dose was observed for frequency of MNed BN cells and
necrosis (Table III). From these experiments we determined
that 750 µM hydrogen peroxide was the highest appropriate
challenge dose because at this concentration we could expect
a 6-fold increase in MNed cell frequency without severe
cytotoxic or cytostatic effects that could have made it
impractical to determine MNed cell frequency in BN cells.
The doses selected were also based on a power calculation
indicating that a 20% change in MNed cell frequency in BN
121
W.Greenrod and M.Fenech
Fig. 3. (a) In vitro effect of catechin and caffeic acid (C & C), ethanol
(ETH), glycerol (GLY), a mixture of wine components (MIX), phenolicstripped white wine (SWW) and tartaric acid (TART) wine components on
baseline MNed cell frequency in BN cells from isolated lymphocyte
cultures. At 10% (v/v) concentration, the observed variation in means is
statistically significant (P ⫽ 0.050, ANOVA). #, Significantly different from
ethanol (P ⬍ 0.05, paired t-test). (b) In vitro effect of wine components on
baseline MNed cell frequency in BN cells from whole blood cultures. The
observed variation in means is not statistically significant at either
concentration (P ⬎ 0.05, ANOVA).
cells would be detectable with statistical significance
(P ⬍ 0.05) if blood from four subjects was tested in duplicate.
Effect of wine components on baseline genetic damage and
cytotoxicity in isolated lymphocyte cultures
Mean baseline MNed cell frequency for the four volunteers
was found to be 3.0 ⫾ 0.8 MNed cells/1000 BN cells in the
HBSS control. A statistically significant difference in MNed
cell frequency in BN cells was found across treatments at 10%
(P ⫽ 0.05, ANOVA) but not at 2.5% concentration (P ⫽ 0.144,
ANOVA) (Figure 3a). MNed cell frequency was increased on
treatment with ethanol at 10% concentration and although this
difference did not reach significance against the HBSS control,
it was significant when compared with glycerol, the mixture,
phenolic-stripped white wine and tartaric acid (P ⬍ 0.02,
P ⬍ 0.05, P ⬍ 0.01 and P ⬍ 0.02, respectively). Necrosis
was significantly reduced by all wine components except
tartaric acid at both concentrations (P ⬍ 0.0001, ANOVA)
(Figure 4a), however, there were no significant effects on NDI
and apoptosis.
122
Fig. 4. In vitro effect of catechin and caffeic acid (C & C), ethanol (ETH),
glycerol (GLY), a mixture of wine components (MIX), phenolic-stripped
white wine (SWW) and tartaric acid (TART) wine components on (a)
baseline necrotic cell frequency (*P ⬍ 0.05 versus all other treatments;
**P ⬍ 0.05 versus all other treatments except TART) and (b) hydrogen
peroxide-induced (actual – baseline) necrotic cell frequency in isolated
lymphocyte cultures (*P ⬍ 0.05 versus all other treatments except TART;
**P ⬍ 0.05 versus all other treatments except ETH).
Effect of wine components in whole blood cultures on
baseline DNA damage and NDI
The mean background level of MNed cells after whole blood
culture was 3.3 ⫾ 0.6 MNed cells/1000 BN cells in HBSS
controls (a result that is similar to that obtained for isolated
lymphocyte cultures). No NPB were observed in BN cells.
There were no significant differences between treatments in
MNed cell frequency (Figure 3b) at either concentration
amongst the various wine components (P ⬎ 0.05). As in
isolated lymphocyte cultures, there was a trend for elevated
MNed cell frequency on exposure to ethanol. A significant
difference in NDI was observed across treatments at both 10
(P ⫽ 0.046, ANOVA) and 2.5% dilution (P ⫽ 0.004, ANOVA).
At 10%, all treatments except glycerol reduced NDI significantly compared with the HBSS control (P ⬍ 0.05), whilst at
the lower concentration all treatments except the mixture and
tartaric acid reduced NDI values (P ⬍ 0.05).
Effect of wine components in plasma on hydrogen peroxideinduced DNA damage and cytotoxicity
A significant difference (P ⫽ 0.015, ANOVA) in induced
(observed – baseline) MNed cell frequency on challenge with
hydrogen peroxide was shown for those samples containing
Wine components protect against DNA damage
Fig. 6. In vitro effect of catechin and caffeic acid (C & C), ethanol (ETH),
glycerol (GLY), a mixture of wine components (MIX), phenolic-stripped
white wine (SWW) and tartaric acid (TART) on γ-radiation-induced
nucleoplasmic bridge frequency in BN cells from whole blood cultures. No
significant difference was seen at 2.5% (v/v) (P ⬎ 0.05, ANOVA), however,
at 10% (v/v) the significant ANOVA value (P ⫽ 0.012) was the result of a
significant reduction in NPB formation with all components (*, all
P ⬎ 0.05, paired t-test). #, ETH and MIX were significantly different from
all other treatments (P ⬍ 0.05, paired t-test), but not each other
(P ⫽ 0.125).
Fig. 5. (a) In vitro effect of catechin and caffeic acid (C & C), ethanol
(ETH), glycerol (GLY), mixture of wine components (MIX), phenolicstripped white wine (SWW) and tartaric acid (TART) on hydrogen peroxideinduced (actual – baseline) MNed cell frequency in isolated lymphocyte
cultures. Significant differences in MNed cell frequency were shown at both
2.5 and 10% (v/v) (P ⫽ 0.043 and P ⫽ 0.015, respectively). *, Significantly
different from the control (P ⬍ 0.05, paired t-test); #, not significantly
different from each other (P ⬎ 0.05, paired t-test). (b) In vitro effect of
wine components on γ-radiation-induced (actual – baseline) MNed cell
frequency from whole blood cultures. At both 2.5 and 10% (v/v)
concentration, the observed variation in means is statistically significant
(P ⬍ 0.05, ANOVA). *, Significantly different from control (P ⬍ 0.05,
paired t-test); #, significantly different from all other treatments (P ⬍ 0.05),
but not each other (P ⫽ 0.197).
wine components at 10% (v/v) concentration (Figure 5a).
Samples supplemented with C & C, ethanol or the mixture
of all components showed the largest and most significant
reductions in MNed cell frequency (P ⬍ 0.05). A similar
significant effect was observed for components at 2.5% dilution
for induced MNed cell frequency (P ⫽ 0.043, ANOVA)
(Figure 5a). Significant differences in induced necrosis between
treatments were found at both concentrations (P ⬍ 0.0001,
ANOVA) (Figure 4b). Necrosis significantly increased relative
to the HBSS control with all components at both concentrations
in all cases (P ⬍ 0.05 in all cases) except tartaric acid
(P ⬎ 0.2), which was in fact significantly different from all
other treatments (P ⬍ 0.05). At the 10% (v/v) concentration,
C & C produced an increase in hydrogen peroxide-induced
necrosis that was significantly greater than that observed for
all other treatments except ethanol. NDI and apoptosis did not
change significantly across treatments on hydrogen peroxide
challenge at either concentration (P ⬎ 0.05, ANOVA).
Effect of wine components in whole blood on γ-radiationinduced DNA damage and NDI
There was a significant difference in γ-ray-induced (observed –
baseline) MNed cell frequency between treatments at 2.5
(P ⫽ 0.037, ANOVA) and 10% (P ⬍ 0.001, ANOVA) dilution
(Figure 5b). All components significantly reduced MNed cell
frequency (P ⬍ 0.05) compared with the HBSS control at
10% (v/v) concentration, however, post hoc analysis demonstrated that the most significant reductions in MNed cell
frequency were achieved with the mixture of all components,
ethanol or C & C (P ⬍ 0.001, P ⫽ 0.001 and P ⫽ 0.001,
respectively). C & C and the mixture were not significantly
different from each other (P ⬎ 0.19). However, C & C was
significantly more protective than ethanol (P ⬍ 0.01). At 2.5%
dilution, all treatments except tartaric acid significantly reduced
induced MNed cell frequency (P ⬍ 0.05). Likewise, all
components significantly reduced NPB formation relative to
the HBSS control (P ⬍ 0.02 in all cases) at the 10% dilution
(Figure 6). The most significant reduction in NPB frequency
was in cultures supplemented with the mixture of all components or C & C (P ⬍ 0.005 and P ⬍0.001, respectively); these
two treatments were significantly different from all other
components (P ⬍ 0.02), but not from each other (P ⬎ 0.125).
The NDI was significantly reduced at the 10% concentration
on treatment with ethanol, the mixture of components and
phenolic-stripped wine (P ⬍ 0.05).
Cross-correlation of measures (Table IV)
For cross-correlations between assay end-points, data from
baseline and challenge experiments (observed frequency data
only) were combined. In isolated lymphocyte cultures (i.e. control and hydrogen peroxide challenge), a strong positive
correlation was observed between MNed cell frequency and
necrotic cell frequency (r ⫽ 0.665, P ⬍ 0.001), and NDI was
significantly positively correlated with necrotic cell and MNed
cell frequency (r ⫽ 0.410 and r ⫽ 0.277, respectively,
P ⬍ 0.005). For whole blood cultures (i.e. control and
123
W.Greenrod and M.Fenech
Table IV(a). Cross-correlation of assay end-points in isolated lymphocyte
cultures
Hydrogen peroxide challenge
APOP
NDI
MNed
NEC
–0.089
0.352
0.410
0.000
0.149
0.117
0.665
0.000
–0.170
0.074
0.277
0.003
APOP
NDI
Spearman correlation
P value
Spearman correlation
P value
Spearman correlation
P value
APOP, apoptotic cell frequency; MNed, micronucleated cell frequency;
NDI, nuclear division index; NEC, necrotic cell frequency. Coefficients
calculated using Spearman’s rank correlation.
Table IV(b). Cross-correlation of assay end-points in whole blood cultures
Gamma radiation challenge
NDI
MNed
NPB
–0.305
0.001
0.895
0.000
–0.284
0.002
NDI
Spearman correlation
P value
Spearman correlation
P value
MNed, micronucleated cell frequency; NDI, nuclear division index; NPB,
nucleoplasmic bridge frequency. Coefficients were calculated using
Spearman’s rank correlation.
γ-radiation challenge), NPB frequency was significantly and
positively correlated with MNed cell frequency (r ⫽ 0.895,
P ⬍ 0.001). Both markers of DNA damage were negatively
correlated with NDI (r ⫽ –0.284 and r ⫽ –0.305, respectively,
P ⬍ 0.005), which was opposite to the observation for MNed
cells in isolated lymphocyte cultures.
Discussion
The epidemiological evidence linking moderate wine consumption with reduced cardiovascular disease risk is compelling
and evidence is also accumulating for potential protective
effects against cancer (Boffetta and Garfinkel, 1990; Renaud
et al., 1998). Because increased oxidative stress is a recognized
risk factor for both cardiovascular disease and cancer some
investigators have examined oxidative damage to DNA. A
human intervention study suggested that moderate wine consumption reduces the concentration of 8-hydroxydeoxyguanosine (Leighton et al., 1999). A cross-sectional study of premenopausal, non-smoking women showed a significant inverse
correlation between alcoholic drink consumption and oxidation
of lymphocyte DNA (Bianchini et al., 2001). These results
are also supported by experiments in rodents which showed
that ingestion of wine solids reduces DNA oxidation in the
colon (Giovanelli et al., 2000).
The main questions that needed to be resolved are essentially
the following: (i) which of the major wine components are
protective against oxidative stress at concentrations that are
physiologically relevant; (ii) what is the relative contribution
of the wine components that are protective against oxidative
damage to DNA; (iii) are the antioxidant effects of the various
wine components additive or synergistic? The results we report
in this study provide a first comprehensive assessment of the
relative antioxidant properties of tartaric acid, C & C, ethanol
and, finally, glycerol. We have used the CBMN assay because
we have shown that this method enables the measurement of
genotoxic and cytotoxic events to be measured efficiently. In
124
addition, we have also shown that the CBMN assay is amongst
the most sensitive methods for measuring DNA damage
induced by ionizing radiation, superoxide, hydrogen peroxide
and activated neutrophils (Fenech and Morley, 1986; Umegaki
and Fenech, 2000). Although the CBMN assay does not
provide information on specific DNA base oxidation, it has
been shown to be more sensitive than the HPLC assay for
8-hydroxydeoxyguanosine in detecting DNA damage induced
by X-rays and UV radiation (Kobus et al., 1993).
The results from our experiments suggest that the wine
components may have small but significant effects on baseline
DNA damage rate, with ethanol tending to increase
micronucleus formation when compared with the other wine
components. The other wine components or ethanol in combination with the other wine components tended to reduce
DNA damage rate, although the differences were not significant
when compared with the HBSS control. The genotoxic effect
of ethanol is not surprising given that there is ample evidence
in the literature that ethanol can induce chromosome breakage
and sister chromatid exchange, particularly in those with
polymorphisms that alter the activity of aldehyde dehydrogenase (Morimoto et al., 1994). We have not genotyped the
volunteers in our study and therefore we are unable to say
whether the genetic background was an important variable,
however, future studies should consider this potential confounding effect. The isolated lymphocyte base-line experiments
(Figure 4a) suggested a prevention of necrosis by all wine
components except tartaric acid, which is consistent with a
possible antioxidant action given that necrosis in this in vitro
system is efficiently induced by oxidants such as hydrogen
peroxide (Fenech et al., 1999). In whole blood cultures wine
components reduced the NDI. However, it is difficult to
interpret the significance of this observation given that correlation analysis indicates that a reduced NDI is weakly but
significantly correlated with increased MNed cell frequency.
The effects observed in vitro do not suggest that moderate
wine consumption is likely to elevate genotoxicity, cytotoxicity
or cell proliferation in vivo.
The results for the hydrogen peroxide challenge suggest that
all wine components except tartaric acid significantly reduced
the level of induced DNA damage, however, this decrease was
at the expense of increased necrosis, with no effect on
apoptosis. It is known that hydrogen peroxide challenge can
cause cells to switch from apoptosis to necrosis as the main
mode of cell death (Hampton and Orrenius, 1997; Fenech
et al., 1999). It is therefore possible that the observed reduced
MNed cell frequency may have occurred due to improved
elimination by apoptosis and necrosis of cells with DNA
damage. However, a quenching effect of wine components
against reactive oxygen species generated by hydrogen peroxide is also a plausible explanation for the reduced MNed cell
frequency. An alternative explanation is the possibility that
some of the wine components may have been mildly prooxidant under the conditions of the experiment and may have
exerted an adaptive response by inducing an up-regulation of
antioxidant mechanisms within the cell. Perhaps the most
interesting observation was that ethanol, glycerol and C & C
were equally effective in reducing DNA damage induced by
hydrogen peroxide, however, the mixture of all wine components was not more effective than the individual contribution
of ethanol, glycerol or C & C. This suggests that the effect of
the components in the mixture is not additive and definitely
not synergistic. These results seem to indicate that the protective
Wine components protect against DNA damage
effects of ethanol, glycerol and C & C may be operating
through a similar mechanism and that their protective effects
are already at their maximum at the concentrations examined.
The latter is also supported by the observation that there is
virtually no difference between the extent of protection at the
2.5% dilution relative to the 10% dilution.
The results for effects on γ-radiation-induced DNA damage
show that all the components of wine examined, including
tartaric acid, significantly reduced MNed cell frequency and
NPB frequency when present at the 10% dilution, while
all but tartaric acid significantly protected against radiationinduced DNA damage at the 2.5% dilution, although to a
lesser extent than that observed at the 10% dilution. Unlike
the hydrogen peroxide challenge experiments, it is evident that
the protective effect is not saturated at the 2.5% dilution.
However, like the hydrogen peroxide challenge, the protective
effect of the mixture was not greater than that of C & C and
clearly did not exhibit an additive effect. An additive protective
effect should have resulted in a reduction in MNed cell
frequency of 64%, however, the observed reduction for the
mixture was only 32%, which is similar to the effect of C &
C (30% reduction) alone. Unlike hydrogen peroxide challenge,
C & C was more effective than ethanol in reducing DNA
damage induced by ionizing radiation. Therefore, it is apparent
that wine and its components also protect against the genotoxic
effect of ionizing radiation, however, it is still uncertain
whether this protective effect is saturated at the 10% dilution.
The lack of an additive effect suggests that the various wine
components examined may be exerting their protective effect
via a similar mechanism such as quenching of singlet oxygen,
as has been reported recently for caffeic acid (Foley et al.,
1999), or hydroxyl radical scavenging, as has been reported
for alcohols (Worm et al., 1993). The observation that phenolicstripped white wine was less protective than the mix also
suggests that wine phenolics play an important role in the
observed protective effects of the mixture. Our results support
previous observations that ethanol and glycerol are protective
against the cytotoxic effects of ionizing radiation (Worm et al.,
1993) and that plant phenolic compounds prevent chromosome
damage induction by ionizing radiation (Shimoi et al., 1994;
Parshad et al., 1998; Virinda and Devi, 2001).
Taken together, the results of our experiments support the
hypothesis that the major components of white wine (at
physiologically relevant concentrations that are in proportion
to their concentration in white wine) are protective against
oxidative stress generated by hydrogen peroxide and ionizing
radiation. The novelty of our results is the observation that
both wine phenolics and ethanol/glycerol are protective to
significant extents against the DNA-damaging effects of hydrogen peroxide and ionizing radiation and that these effects do
not appear to be additive. The latter may have important
implications with regard to the ongoing debate about whether
the phenolics in wine provide a health advantage over alcoholic
drinks that have much lower concentrations of these compounds. Also notable for this discussion is the observation that
ethanol was the only wine component increasing baseline
MNed cell frequency and that this effect was neutralized in
the mixture of all components. The results we have obtained
suggest that at physiologically relevant concentrations: (i)
ethanol on its own is genotoxic compared with phenolics; (ii)
that under conditions of hydrogen peroxide stress ethanol and
phenolics are equally protective; (iii) under conditions of
ionizing radiation stress phenolics appear to be more protective
than ethanol and glycerol. Our results suggest the hypotheses
that dealcoholized wines may be equally protective as complete
wine but wines stripped of phenolics are likely to be less
protective than complete wine and dealcoholized wines. These
hypotheses need to be tested in properly controlled intervention
studies. If these hypotheses turn out to be correct, they
may explain why it has been difficult epidemiologically to
distinguish between the health benefits of different alcoholic
drinks drunk in moderation and why some studies suggest that
wine may afford more protection than other alcoholic drinks
with lower phenolic and higher ethanol contents (Klatsky and
Armstrong, 1993; Gronbaeck et al., 1995).
Our future studies are now aimed at performing controlled
intervention studies in vivo to determine whether alcohol,
phenolic-stripped wine and whole wine differ significantly in
their capacity to alter the antioxidant properties of blood ex
vivo with respect to protection against hydrogen peroxide- and
ionizing radiation-induced DNA damage.
Acknowledgements
We are thankful to Creina Stockley and Dr Phil Burcham for their helpful
discussions and suggestions during the course of these studies. We are grateful
to the Department of Clinical and Experimental Pharmacology of the University
of Adelaide for supporting the PhD candidature of Will Greenrod. This study
would not have been realised without the help of the volunteers, the staff at
the CSIRO Health Sciences and Nutrition Clinic and the occasional technical
assistance of Julie Turner, Felicia Bulman and Phil Thomas. This study was
supported by a PhD scholarship granted to W.G. by the Grape and Wine
Research and Development Corp. (Australia).
References
Andreassi,M.G., Botto,N., Colombo,M.G., Biagini,A. and Clerici,A. (2000)
Genetic instability and atherosclerosis: can somatic mutations account for
the development of cardiovascular disease? Environ. Mol. Mutagen., 35,
265–269.
Bianchini,F., Jaeckel,A., Vineis,P., Martinez-Garcia,C., Elmstahl,S., van
Kappel,A.L., Boeing,H., Ohshima,H., Riboli,E. and Kaaks,R. (2001) Inverse
correlation between alcohol consumption and lymphocyte levels of
8-hydroxydeoxyguanosine in humans. Carcinogenesis, 22, 885–890.
Botto,N., Rizza,A., Colombo,M.G., Mazzone,A.M., Manfredi,S., Masetti,S.,
Clerico,A., Biagini,A. and Andreassi,M.G. (2001) Evidence for DNA
damage in patients with coronary artery disease. Mutat. Res., 493, 23–30.
Boffetta,P. and Garfinkel,L. (1990) Alcohol drinking and mortality among men
enrolled in an American Cancer Society prospective study. Epidemiology, 1,
342–348.
Casalini,C., Lodovici,M., Briani,C., Paganelli,G., Remy,S., Cheynier,V. and
Dolara,P. (1999) Effect of complex polyphenols and tannins from red wine
(WCPT) on chemically induced oxidative DNA damage in the rat. Eur. J.
Nutr., 38, 190–195.
De Flora,S., Izzotti,A., Walsh,D., Degan,P., Petrilli,G.L. and Lewtas,J. (1997)
Molecular epidemiology of atherosclerosis. FASEB J., 11, 1021–1031.
Fenech,M. (1993a) The cytokinesis-block micronucleus technique: a detailed
description of the method and its application to genotoxicity studies in
human populations. Mutat. Res., 285, 35–44.
Fenech,M. (1993b) The cytokinesis-block micronucleus technique and its
application to genotoxicity studies in human populations. Environ. Health
Perspect., 101, 101–107.
Fenech,M. (2000) The in vitro micronucleus technique. Mutat. Res., 455,
81–95.
Fenech,M. and Morley,A.A. (1986) Cytokinesis-block micronucleus method
in human lymphocytes: effects of in vivo ageing and low dose X-irradiation.
Mutat. Res., 161, 193–198.
Fenech,M., Stockley,C. and Aitken,C. (1997) Moderate wine consumption
protects against hydrogen-peroxide-induced DNA damage. Mutagenesis,
12, 289–296.
Fenech,M., Crott,J., Turner,J. and Brown,S. (1999) Necrosis, apoptosis,
cytostasis and DNA damage in human lymphocytes measured
simultaneously within the cytokinesis-block micronucleus assay: description
of the method and results for hydrogen peroxide. Mutagenesis, 14, 605–612.
Ferguson,L.R. (2001) Role of plant polyphenols in genomic stability. Mutat
Res., 475, 89–111.
125
W.Greenrod and M.Fenech
Foley,S., Navaratnam,S., McGarvey,D.J., Land,E.J., Truscott,T.G. and RiceEvans,C.A. (1999) Singlet oxygen quenching and the redox properties of
hydroxycinnamic acids. Free Radic. Biol. Med., 26, 1202–1208.
Ghiselli,A., Nardini,M., Baldi,A. and Scaccini,C. (1998) Antioxidant activity
of different phenolic fractions separated from an Italian red wine. J. Agric.
Food Chem., 46, 361–367.
Giovannelli,L., Testa,G., De Filippo,C., Cheynier,V., Clifford,M.N. and
Dolara,P. (2000) Effect of complex polyphenols and tannins from red wine
on DNA oxidative damage of rat colon mucosa in vivo. Eur. J. Nutr., 39,
207–212.
Gronbaeck,M., Deis,A., Sorensen,T.J.A., Becker,U., Schnohr,P. and Jensen,G.
(1995) Mortality associated with moderate intakes of wine, beer or spirits.
Br. Med. J., 310, 1165–1169.
Hampton,M.B. and Orrenius,S. (1997) Dual regulation of caspase activity by
hydrogen peroxide: implications for apoptosis. FEBS Lett., 414, 552–556.
Klatsky,A.L. and Armstrong,M.A. (1993) Alcoholic beverage choice and risk
of coronary heart disease mortality: do red wine drinkers fare best? Am. J.
Cardiol., 71, 467–469.
Kobus,K., Turner,J., Priestley,B. and Dreosti,I. (1993) A comparison of the
8-hydroxydeoxyguanosine and micronucleus techniques for the assessment
of X-ray and UV induced genotoxicity. Res. Commun. Chem. Pathol.
Pharmacol., 80, 249–252.
Leighton,F., Cuevas,A., Guasch,V. et al. (1999) Plasma polyphenols and
antioxidants, oxidative DNA damage and endothelial function in a diet and
wine intervention study in humans. Drugs Exp. Clin. Res., 25, 133–141.
Morimoto,K., Takeshita,T., Miura,K., Mure,K. and Inoue,C. (1994) Does the
genetic deficiency in ALDH2 determine the alcohol-drinking behaviour and
the induction of chromosome alterations in peripheral lymphocytes by
alcohol. In Obe,G. and Natarajan,A.T. (eds), Chromosome Alterations—
Origin and Significance. Springer-Verlag, Berlin, Germany, pp. 293–306.
Parshad,R., Sanford,K.K., Price,F.M., Steele,V.E., Tarone,R.E., Kelloff,G.J.
and Boone,C.W. (1998) Protective action of plant polyphenols on radiationinduced chromatid breaks in cultured human cells. Anticancer Res., 18,
3263–3266.
Renaud,S. and de Lorgeril,M. (1992) Wine, alcohol, platelets and the French
paradox for coronary heart disease. Lancet, 339, 1523–1526.
Renaud,S. and de Logeril,M. (1993) The French paradox: dietary factors and
cigarette-smoking related health risks. Ann. N. Y. Acad. Sci., 686, 299–309.
Renaud,S., Guegen,R., Schenker,J. and d’Houtaud,A. (1998) Alcohol and
mortality in middle-aged men from eastern France. Epidemiology, 9,
184–188.
Shimoi,K., Masuda,S., Furugori,M., Esaki,S. and Kinae,N. (1994)
Radioprotective effect and antioxidant activities of plant-derived
polyphenolic flavonoids in gamma-ray irradiated mice. Carcinogenesis, 15,
2669–2672.
Somers,T.C. and Ziemelis,G. (1985) Spectral evaluation of total phenolic
components in V.vinifera: grapes and wine. J. Sci. Food Agric., 36,
1275–1284.
St Leger,A.S., Cochrane,A.L. and Moore,F. (1979) Factors associated with
cardiac mortality in developed countries with particular reference to the
consumption of wine. Lancet, 12, 1017–1020.
Umegaki,K. and Fenech,M. (2000) Cytokinesis-block micronucleus assay in
WIL2-NS cells: a sensitive system to detect chromosomal damage induced
by reactive oxygen species and activated human neutrophils. Mutagenesis,
15, 261–269.
Vrinda,B. and Devi,P.U. (2001) Radiation protection of human lymphocyte
chromosomes in vitro by orientin and vicentin. Mutat. Res., 498, 39–46.
Worm,K.H., Klimczak,U. and Schulte-Frohlinde,D. (1993) Radiosensitisation
and radioprotection of E.coli by alcohols. Int. J. Radiat. Biol., 64, 485–495.
Received on July 4, 2002; accepted on September 30, 2002
126