Mntagenesis vol.12 no.4 pp.259-264, 1997
Damage proneness induced by genomic DNA demethylation in
mammalian cells cultivated in vitro
Paolo Perticone1-3, Giuseppe Gensabella1 and
Renata Cozzi2
1
Centre di Genetica Evoluzionistica del CNR, c/o Dipartimento di Genetica
e Biologia Molecolare, Universiti 'La Sapienza', 00185 Roma and
2
Dipartimento di Biologia Universita di Roma m , Italy
3
To whom correspondence should be addressed
Variations in the genomic DNA methylation level have
been shown to be an epigenetic inheritable modification
affecting, among other targets, the sister chromatid
exchange (SCE) rate in mammalian cells in vitro. The
inheritable increase in SCE rate in affected cell populations
appears as a puzzling phenomenon in view of the well
established relation between SCE and both mutagenesis
and carcinogenesis. In the present work we demonstrate
that, in a treated cell population, demethylation could
be responsible for the inheritable induction of damage
proneness affecting both damage induction and repair.
Normal and ethionine or azacytidine treated Chinese
hamster ovary cells, subclone Kl (CHO-K1), were challenged with UV light (UV) or mitomycin-C (MMC) at
different times from the demethylating treatment The SCE
rate was measured with two main objects in view: (i)
the induction of synergism or additivity in combined
treatments, where mutagen (UV or MMC) pulse is supplied
from 0 to 48 h after the end of the demethylating treatment,*
and (ii) the pattern of damage extinction, for the duration
of up to six ceil cycles after the end of the combined
(demethylating agent + mutagen) treatment Results
indicate both a synergism in SCE induction by rautagens
in demethylated cells even if supplied up to four cell cycles
after the end of the demethylation treatment and a delay
in recovery of induced damage, compared with normally
methylated cells. These data are discussed in the light of
the supposed mechanism of SCE increase and of the
possible biological significance in terms of mutagenesis and
carcinogenesis.
Introduction
The sister chromatid exchange (SCE) test, although routinely
employed to detect the potential mutagenic activity of chemical
and physical agents, leaves behind significant uncertainty as
to the precise mechanism of their formation, their relation to
mutagenic processes and their precise biological significance
(Sandberg, 1982). Nevertheless, the SCE test appears to be
very useful in the study of DNA damage for at least three
reasons: (i) variations in the SCE background level are compatible with cell growth and division (Sandberg, 1982; Perticone
etai, 1993); (ii) deriving from DNA breakage and strand
exchange, SCEs are believed to reflect the interaction of
damaging agents with DNA (Sandberg, 1982); and (iii) because
of the close association between initial steps in cell transformation and induction of DNA alterations, they appear to be a
significant indicator of neoplastic transformation processes, as
© UK Environmental Mutagen Society/Oxford University Press 1997
shown by the capacity of several carcinogens to affect the
SCE rate (Popescu etai, 1981).
Although a number of studies have focused (with conflicting
results) on ascertaining whether DNA lesions responsible for
the exchanges are repaired during the SCE formation (Kato,
1973; Shaffer, 1977; Ishii and Bender, 1980; Tice and
Schwartzman, 1982; Lindenhahan and Shubert, 1983) and
despite the observation that DNA lesions are unable to induce
SCE at the same locus in successive cell generations
(Schwartzman etai., 1985), very few reports have dealt with
the persistence of mutagen- and carcinogen-induced increase
in the SCE baseline in vitro through successive cell cycles
after treatments (Muscarella and Bloom, 1982; Popescu etai,
1985; Takeshita and Conner, 1985). This lack appears even
more surprising in the light of the possibility that specific
tumorigenic activity may be related to the duration of SCE
persistence (Takeshita and Conner, 1985).
Previous data obtained in our laboratory showed an induction
of DNA modifications (increase in both SCE and ribosomal
gene activity levels) and their persistence up to 10-16 cell
cycles after the end of a single pulse of demethylating agents
in mammalian cells in vitro (Perticone etai., 1987, 1990;
Giancotti etai, 1995). These data have been explained by
suggesting a close relationship between inherited demethylation
(by maintenance methylases) and persistence of an increased
SCE frequency (Perticone etai, 1987). Conversely, a rapid
reduction of the increased SCE level has been reported for all
mutagen agents tested in vitro (Muscarella and Bloom, 1982;
Conner and Cheng, 1983; Popescu etai, 1985; Takeshita and
Conner, 1985).
In the present work, ethionine (ETH) and azacytidine (AZA)treated Chinese hamster ovary (CHO-K1) cells were challenged at different times after the end of the demethylating
treatment with two known mutagens, namely 254 nm UV light
(UV) and mitomycin-C (MMC). Our aim was to compare the
in vitro patterns of SCE induction and their extinction, by
means of two classical mutagen agents, in normally methylated
versus demethylated mammalian cells. Two different challenging protocols were applied in order to determine whether the
effects of AZA and ETH on mutagen induced SCEs should be
ascribed to the induced demethylation (synergism or additivity
experiments), and to evaluate the variations in the damage
extinction trend (persistence experiments).
The results indicate a synergism in SCE induction and a
delay in the extinction of damage in demethylated, mutagentreated mammalian cells compared with normally methylated
ones. These findings are discussed in terms of their possible
biological significance in in connection with mutagenesis and
carcinogenesis.
Materials and methods
Experimental protocols
Synergism or additivity experiments. Cells were seeded 2 h before the first
treatment. Cells were seeded at a density of 1 x 105 cells/5 ml medium flasks
for Petri dishes in the case of UV irradiation) CFalconj. Demethylating agents
259
P.PertJcone, G.Gensabella and R.Cozzi
were supplied for 12 h (ETH, 5 mM; AZA, 10 nM) At the end of the
treatment, the demethylating agent was removed and cells were challenged
with a mutagen (MMC, 0.1 Hg/ml; UV, 10 J/m2/s) either immediately [time
interval (TT) = 0] or at different times ranging from 12 to 48 h after treatment.
At the end of the mutagen treatment, cells were allowed to undergo two cell
cycles (48 h) in the presence of bromodeoxyundine (BrdUrd) to obtain sister
chromatid differentiation.
In the case of single treatments (mutagens or demethylating agents), cells
were allowed to undergo two cell cycles in the presence of BrdUrd,
before fixation.
In the case of combined treatments, the mutagens were supplied immediately
or at 12, 24, 36 and 48 h after the end of the demethylating treatment, and
cells were fixed two cell cycles after the end of the mutagen treatment.
Colchicine (5X 10"7 M; Sigma, USA) was added 2 h before fixation.
Damage extinction experiments. Seeding, agent's doses, treatment durations
and colchicine, as above. After the removal of demethylating agent, the cells
were immediately challenged with the mutagen treatment At the end of this
second treatment, cells were allowed to perform two (24 h), four (48 h) and
six (72 h) cell cycles, with the last two cycles before fixation in the presence
of BrdUrd, to obtain sister chromatid differentiation.
Cell cultures
Chinese hamster ovary cells, subclone Kl (CHO-KI, n. C CL61, batch F9322) supplied from the American Type Culture Collection, were routinely
cultured in Ham's F-12 medium (Flow, USA) supplemented with 10% fetal
calf serum (Gibco, USA), 1% L-glutanune, 2% penicillin (5000 IU/ml)
and streptomycin (5000 g/ml). Under these growth conditions, the average
generation time was 12 h, as measured by BrdUrd incorporation (Ivett and
Tice, 1992)
Chemicals
The chemicals were freshly prepared immediately before use. Both L-ethionine
(Sigma) and 5-azacytidine (Sigma) were dissolved in Ham's F-12 medium
(stock solutions were 150 and 1 mM respectively) After complete dissolution
of the chemicals using a magnetic stirrer for 1 h at 37°C, the solutions were
filtered through sterile millex-gs 0.22 |im filters (Milhpore), and dispensed in
the final concentrations shown in Figures for 12 h (one cell cycle). MMC
(Kiowa, Japan) was dissolved in bi-distilled sterile water immediately before
treatments and supplied at a final concentration of 0.1 ng/ml to the cultures
in complete medium without serum for 2 h. After the pulse, all chemicals
were removed by gentle aspiration of medium. Cells were then washed twice
and refed with complete medium BrdUrd (Fluka, USA) was supplied to the
cultures from stock solutions stored at -20°C, at the final concentration of
5X10- 6 M.
UV-irradiation
Cells were UV-irradiated with a Spectroline lamp (model ENF-280 C/F),
50 Hz, 17 A (Spectronics Corporation, Westbury, NY, USA) al 254 nm wave
length, from a distance of 15 cm at a fluence rate of 2.50 J/rrr/s. Before
irradiation (10 J/m2) the medium was removed from the Petri dishes; cells
were then incubated in fresh complete medium.
SCE analysis
Harvesting and preparation of slides were performed according to the airdrying method. The usual Giemsa plus Hoechst technique (Perticone et al..
1987) was employed for the differential staining of sister chromatids. A
total of 40 metaphases from coded slides were scored for each point and
each treatment.
Statistical analysis
Means and SE were determined For each harvest time, the control and treated
cultures were compared by the i-test. The analysis of variance (ANOVA) test
was employed to compare the SCE rate obtained at differenl harvest times
(two to six cycles) in each treated culture as well as in the controls. The
Fisher '/•* test for a two-way ANOVA was applied to analyse additivity
or synergism. This test shows the statistical significance of the absence
of additivity.
Results
Background
Asynchronously growing CHO-KI cells were used. After 24 h
culture in the presence of BrdUrd the percentage of mitoses
with a complete sister chromatid differentiation (M2 cells)
was ~90-93% of the total number of mitotic cells. This
correlates to an average generation time of 12 h and good
260
Table I. Synergism or additivity in the SCE level induced by combined
treatments of demethylating (AZA, ETH) plus mutagen (UV, MMC) agents
in CHO-KI cells
Treatment(s)
Control
UV
MMC
AZA
ETH
AZA +
AZA +
AZA +
AZA +
AZA +
ETH +
ETH +
ETH +
ETH +
ETH +
AZA +
AZA +
AZA +
AZA +
AZA +
ETH +
ETH +
ETH +
ETH +
ETH +
UV
UV
UV
UV
UV
UV
UV
UV
UV
UV
MMC
MMC
MMC
MMC
MMC
MMC
MMC
MMC
MMC
MMC
Time interval*
SCE/cell
± SEM
0
12
24
36
48
0
12
24
36
48
0
12
24
36
48
0
12
24
36
48
9.90
39.20
39.95
22.50
14 80
_
93.32
80.64
64.47
64.71
_
58.02
49.10
50.13
42.08
53.95
55 00
47.50
61.60
56.95
46.50
47 70
41 40
41.05
48.90
0.53
2 27
1.17
1.45
0 83
3 69
2.36
4.26
3.75
_
346
2 26
2.51
3.35
2.49
1 53
2.30
2 32
2.00
1 50
201
1 06
1.32
200
nd
< 0.01
< 001
< 0.01
< 0.01
nd
< 0.01
< 0.01
< 0.01
> 0.01
> 0.01
< 0.01
> 0.01
< 0.01
< 0.01
< 001
< 0.01
< 0.01c
< 0.01°
< 0.01
AZA, azacytidine (10 uM); ETH, ethinine (5 mM), UV, UV light (254 nm)
10 /Jm2- MMC, mitomycin-C (0.1 (ig/ml); cf, culture failed; nd, not done.
Data are the mean of three or four separate experiments. For further details,
please see Materials and methods.
"Time interval, hours between demethylating agent removal and mutagen
supply
b
Calculated using Fisher's two-way analysis of variance.
°Values appear to be below the additivity level
synchronization, probably due to the trypsin shock (Perticone
et al., 1987). When appropriate, cultures were thus trypsinized
every 24 h, so that 24, 48 and 72 h cultures had performed
two, four and six cell cycles after the treatments respectively
(damage extinction experiments). A slight cell cycle delay was
induced only by AZA and MMC treatments. This delay
appeared to be restricted to the first and second cell cycle after
treatment, and it was no longer detectable at subsequent
fixation times, as indicated by the frequency of M2 cells (data
not shown).
In all experiments, the SCE control values did not change
significantly, ranging from 8.15 ± 0.54 to 9.9 ± 0.53 (Tables
I and II). The ANOVA statistical analysis performed showed
that all control cultures can be considered as being in the same
population, as expected.
Mutagen-induced SCEs
MMC induced a 4—5-fold increase of the SCE baseline value.
This level was reduced by 40%, four cell cycles after the
agent's removal, and returned to the control value after six
cell cycles (Tables I and II, Figure 1).
UV-light induced a 4-5-fold increase of the SCE baseline
value. This level was reduced by 70%, four cell cycles after
the end of the treatment, and returned to the control value
after four cell cycles (Tables I and II).
Demethylating agents-induced SCEs
AZA induced more than a doubling of the SCE rate, which
ranged from 19.15 to 22.5. whereas ETH induced a 659c
Damage proneness and genomic DNA demetbylatlon
Table n. Persistence of SCE induced by combined treatments (demethylating agent + mutagen) in CHO-K1 cells
Treatraent(s)
Level of SCE/ceU ± SEM after no. cycles from last treatment
Four cycles
Two cycles
Control
UV
MMC
AZA
ETH
AZA + UV
ETH + UV
AZA + MMC
ETH + MMC
Six cycles
SCE/ceU
±SEM
SCE/ceU
± SEM
8.2
39.91
37.12
19.15
14.43
66.76
45.25
59.85
57.17
0.74
1.56
2.89
1.35
0.75
2.36
2.66
3.19
2.53
9.85
11.05
23.12
21.5
14.08
27.15
23.71
49.35
33.37
0.85
0.86
0.94
1.04
0.55
1.65
1.54
2.73
1.53
< 0.01
> 0.01
< 0.01
<0.01
/»
SCE/ceU
±SEM
P2
<0.01
< 0.01
< 0.01
< 0.01
8.15
9.05
8.45
19.45
14.27
21.06
19.18
32.95
22.89
0.54
0.61
0.74
1.19
0.74
1.24
2.29
0.81
1.13
>0.01
< 0.01
<0.01
<0.01
Data are the mean of three separate experiments. For further details, please see Materials and methods. "Calculated using Fisher's two-way analysis of
variance.
increase of the SCE rate, which ranged from 14.4 to 14.8. Both
increases were significant (P < 0.001). Over the experimental
duration (up to six cell cycles after the end of the demethylation
pulse), this significant increase in the SCE level was maintained
(Tables I and II).
Synergism or additivity experiments
Cells were challenged by both demethylating agents in combination with both mutagen agents (see Table I). The SCE rate
was detected two cell cycles (24 h) after the end of the
treatment and compared with the control values. The time
elapsing between the end of the demethylating treatment and
the pulse of mutagen is indicated as TI, and ranges from 0 to
48 h. In both UV treatments, the TI = 0 was not performed,
because of the excessive toxicity.
At all TI, with the sole exception of ETH + UV at a TI of
48 h, AZA + MMC at a TI of 0 and 24 h, and ETH + MMC
at a TI of 24 and 36 h, all four combined treatments showed
a significant synergism induced by the challenge protocol, as
indicated by the Fisher 'F test. A noticeable high and
uniformly significant synergism was obtained by the combined
treatment AZA + UV.
Damage extinction experiments
Cells were challenged by both mutagen agents immediately
after removal of demethylating agents (see Table II). The
treatment protocol included a TI = 0 between the two
treatments, whereas the SCE rate was measured two (24 h),
four (48 h) and six (72 h) cell cycles after the end of the second
(last) treatment. The SCE rate of the combined treatments was
compared with that of the single ones.
For all sample times, with the exception of AZA + UV at
six cell cycles (72 h), the induced level of SCE persisted,
confirming that the extinction pattern had been profoundly
altered, as indicated by the Fisher 'F test.
Discussion
DNA damage is induced by a vast variety of agents (both
chemical and physical), that directly or indirectly produce
modifications to the deoxyribonucleic acid molecule. Once
induced, DNA damage is removed by the DNA repair cell
machinery (for a review, see Friedberg, 1985). As a result of
this process, damaged cells (in the vast majority of cases) face
two possibilities: death or restoration of normal conditions. A
third, more important fate, can ensure when cells become in
some way modified, maintaining their ability to grow and
divide. In this case, affected cells may set off along the
complex pathway leading to cancer.
Epigenetic modifications are defined as changes not affecting
DNA sequences. The most frequently studied among them are
modifications of cytosine. An enormous literature has been
produced on biological effects directly or indirectly related to
changes in cytosine DNA methylation (for a recent review,
see Jost and Saluz, 1993).
DNA damaging agents can modify methylation patterns
without inducing mutations (Holliday, 1979; Liebermann et al,
1983); cancer agents may also induce epigenetic variations
connected with tumour progression (Holliday and Jeggo, 1985;
Spruck et al., 1992). Moreover, variations in DNA methylation
are considered as inheritable (Holliday, 1993).
Apart from data obtained in our laboratory (Perticone
etal., 1987, 1990; Giancotti etal, 1995), the existence of a
relationship between demethylation and increase in the SCE
rate has been addressed in a number of reports (Chambers
and Taylor, 1982; Tada-Aki, 1983; Dcushima, 1984; Lavia
etal., 1985).
The work described here was designed to investigate the
possibility that the appearance of an inherited increase in SCE
baseline could be an indicator of some biologically valuable
inherited modification.
The first set of experiments (synergism or addiu'vity) was
designed to detect whether demethylated cells became more
sensitive to damage induction by known mutagens. As working
definition, in fact, synergism may be said to occur when
two treatments are observed to be mutually and positively
interacting towards a given end-point. As shown in combined
treatments (Table I), synergism is almost always observed and
maintained throughout the entire studied lapse of time, with
the sole exception of the combined treatments ETH + UV at
the 48 h TI, AZA + MMC at 0 and 24 h TI and ETH +
-MMC at 24 and 48 h TI. While ETH + UV data can be
accounted for in terms of the disappearance of the synergism
at the longest TI, AZA + MMC data appear not to be selfconsistent because of the presence of three TI positive and
two TI negative with respect to the synergism. The same
problem of consistency arises in ETH + MMC data, where
in two intermediate (24 and 36 h) times a sort of protective
effect occurs. On this subject, it must be underlined that the
synergism or additivity experimental protocols have been
deliberately built in a very restrictive fashion in such a way
SCE values from single treatment controls have been obtained
261
RPerticone, G.GensabeUa and R.Cozii
•—Control
Ail+UV
•
Azj+MMC
—•—Control
•
An
Azi
• -UUC
• -UV
T >-.
o ^_
— Control
Eth+MMC
_ _
r-
Eth+UV
Eth
'
- HUC
'
-EttHMMC
•i-
_
Fig. 1. Persistence of SCE induced by combined treatments (demethylating agent + mutagen) in CHO-KI cells, x axis, number of cell cycles; y axis, SCE/
cell Dose of agents as in Table I; no. cell cycles, number of cycles after the end of the last treatment (mutagen) and before fixing of the cells; data are the
mean of three separate experiments; see Table II for statistical significance. For further details see Materials and methods.
from TI = 0 populations (where SCE values are higher) not
considering extinction of damage due to the increase of the
Tl between given treatments and fixation of cells. Considering
the data overall, the presence of synergism not only in
concomitant or closely spaced treatments (as usually performed), but also when the interval between the twochallenges
lasts up to four cell cycles, supports the view of an inherited
variation affecting damage sensitivity and not of a trivial effect
due to a transitory mechanistic interaction between two agents
acting on the same target. Furthermore, it is highly likely that
this variation is epigenetic. In fact, in the case of AZA, if it
is conceivable that the involved modification is due to its
replacing cytosine in the DNA (Santi et ai, 1983), this action
does not explain the maintenance of synergism data up to 48 h
TI, where the final dilution of AZA substitution reaches
6.25% of the total number of cells in the treated population
(considering a dilution of 50%, for each cell cycle, 12 h).
More likely, in the case of ETH, the agent is known to act by
competitively inhibiting methyltransferase enzymes(s) without
binding to DNA (Cox and Irving, 1977; Bohem and Drahowsky,
1979,1981), thus behaving as a true epigenetic damaging agent.
The second set of experiments (persistence experiments)
were performed in order to further evaluate if induced
demethylation could affect the trend in the extinction of the
induced damage. The experimental protocol differed from the
previous set of experiments in such a way that the mutagen
treatments were always performed immediately after the
262
demethylating pulse. Our interest was no longer focused on
the discrimination between the induction and persistence of
synergism or additivity, but on the measure of the decrease in
the induced damage. In fact, during the time period considered,
SCE induced by mutagens returned to the control values
whereas SCE induced by demethylation maintained a significant difference from controls. In all combined treatments, over
the time interval considered, with the exception of AZA +
UV at the sixth cycle after the end of the last challenge, cell
repair machinery appeared to be unable not only to restore
control SCE values but even to bring cells closer to the
increased SCE background values induced by demethylation
alone.
Viewing the present results, together with literature data, two
main problems can tentatively be considered in greater detail:
Genomic demethylation and SCE 5-methykytasine
Variations in 5-methylcytosine patterns are regarded as epigenetic variations that once induced become inherited. In fact,
the capacity of de novo methylation to recognize and restore
the correct pattern appears to be at least incomplete (Kastan
et a\., 1982). Because of the evidence that in cancer cells the
methylation pattern is abnormal (for a review, see Spruck
et ai, 1992), it could be of some importance to find rapidly
and easily detectable cytogenetic end-points whose variation
is related to changes in methylation. Irrespective of the
mechanism of induction, the main characteristic of this end-
Damage proneness and genomic DNA demethylatlon
point, if it is to be regarded as relevant and eventually useful
for test purposes, must obviously be the possibility of being
inheritably varied without affecting cell viability and proliferation. SCEs (but also nucleolar organiser regions and probably
other parameters) are to be considered as cytogenetic endpoints whose modification does not affect either viability or
proliferation and are thus not expected to be selectively
eliminated in a given cell population. We have already suggested and extensively described a possible mechanism underlying the background increase in SCE by demethylation
(Perticone et al, 1987). This mechanism ascribes to the parental
strand methylation the role of strand discriminator in the case
of repair at the replication fork level. Any decrease in the
percentage 5mC would then increase the probability of errors,
ultimately leading to an increase in SCE. Present data support
and extend this view, indicating that this epigenetic modification results in a sort of damage proneness (through an error
proneness in the strand ligation step in otherwise normally
working repair machinery), viewed both as synergism
(demethylating agent/mutagen) and as modification in the
pattern of damage extinction.
Meaning of a SCE test in persistence protocol
As is well known, the SCE test is widely used as an indicator
of repair processes, even if the rapid spontaneous disappearance
in few cell cycles after their induction poses the question of
how the underlying damage is biologically significant at cell
population level. In fact although SCE formation may represent
a form of somatic recombination, it is clear that if SCEs
derive from exchanges at homologous loci, the result must be
genetically neutral. Even if recombinational events due to
unequal SCE have been detected by a small number of workers,
these same workers have found unequal SCE frequency to be
very rare (Tilley and Birshtein, 1985; Holden etal, 1987;
Morgan and Fero, 1987; Hellgren et al, 1990).
The finding that a particular class of agents (demethylating
agents) is able not only to induce an inheritable increase in
the SCE baseline, but also to modify the SCE level in the
same direction (inheritable increase) after the challenge with
mutagens normally able to induce only a temporary SCE
increase, suggests that this experimental protocol could be
applied to the testing of carcinogenic damage in vitro.
It is our working hypothesis that, in a given cell population,
the baseline SCE level could be considered as an indicator of the
'baseline' repair activity counteracting spontaneous damage.
Following this view, the rapid drop in the increased SCE level
to the baseline value after a given treatment is indicative, at a
population level, of a restoration of normal conditions.
Demethylating agents are a class of widely recognized cancer
agents; the possibility that other cancer agents (irrespective of
their mechanism of action) are able to irreversibly modify
both SCE baseline and SCE extinction profiles in mutagentreatment protocols in a given cell population is, in our opinion,
a good test of our hypothesis. Non-mutagenic carcinogens are
probably the most useful class of cancer agents to test.
Acknowledgements
This work was partially supported by Progetto Bilaterale CNR: Centra di
Genetics Evoluzionistica del CNR, Rome, Italy and Dr E.Patkin from
Department of Molecular Genetics, Institute for Experimental Medicine of
the Russian Academy of Medical Sciences, St. Petersburg, Russia. Prof. A.San
Martini's valuable statistical help is gratefully acknowledged. This paper is
dedicated to Amalia Giuffrida by P.P.
References
Bohem.T.LJ. and Drahousky.D. (1979) Effect of carcinogen ethionine on
enzymatic methylation of DNA sequences with various degrees of
repetiti\eness. Eur. J. Cancer, 16, 1167-1173.
BohemXLJ. and Drahousk\,D. (1981) Elevated transcriptional complexity
and decrease in enzymatic DNA mcthylotion in cells treated uith Lethionine. Cancer Res.. 41, 4101-4106.
ChambersJ.C. and TaylorJ.H. (1982) Induction of sister chromarid exchanges
by 5-fluorodeoxycytidine: correlation with DNA methylation. Chromosoma,
85,603-609.
ConnerJU.K. and ChengJVI. (1983) Persistence of ethylcarbamate-induced
DNA damage in vivo as indicated by sister chromatid exchanges analysis.
Cancer Res., 43, 965-971.
Cox.R., and Irving.C.C. (1977) Inhibition of DNA methylation by S-adenosylethionine with the production of methyl-deficent DNA in degenerating rat
liver. Cancer Res., 37, 222-225; Eur.]. Cancer, 16, 1167-1173.
Friedberg£.C. (ed.) (1985) DNA Repair. W.H.Freeman and Co., New York.
Giancotti.P., GrappelliX., Poggesi.I., AbatecolaAI., de CapoaA., Cozzi.R.
and Perticone J". (1995) Persistence of increased levels of ribosomal gene
activity in CHO-K1 cells treated in vitro with demethylating agents. Mutat.
Res., 348, 187-192.
Hellgren.D., Sahlen.S. and Lambert,B. (1990) Unequal SCE is a rare event
in homologous recombination between duplicated neo gene fragments in
CHO cells. Mutat. Res., 243, 75-80.
HoldenJJ., HoughJvl.R., Reimer,D.L. and White.B.N. (1987) Evidence for
unequal crossing-over as the mechanism for amplification of some
homogeneously staining regions. Cancer Genet. Cytogenet., 29, 139—149.
Holliday.R. (1979) A new theory of carcinogenesis. Br. J. Cancer, 40,513-522.
Holliday.R. (1993) Epigenetic inheritance based on DNA methylation. In
JostJ.P. and Saluz,H.P. (eds), DNA Methylation: Molecular Biology and
Biological Significance. Birkhauser Verlag, Basel, Boston, Berlin.
Holhday.R. and Jeggo.P.A. (1985) Mechanisms of changing gene expression
and their possible relationship to carcinogenesis. Cancer Sun:, 4, 557-581.
Ikushima,T. (1984) SCE and DNA methylation. Basic Life Science. 29 (Part
A), 161-172.
Ishii.Y. and Bender.M.A. (1980) Effects of inhibitors of DNA synthesis on
spontaneous and UV light induced sister chromatid exchanges in Chinese
hamster cells. Mutat. Res., 79, 19-32.
IvettJ.L. and Tice.R.R. (1982) Average generation time: a new method of
analysis and quantitation of cellular proliferation kinetics. Environ.
Mutagen., 4, 358.
JosU.P. and Saluz.H.P. (eds) (1993) DNA Methylation: Molecular Biology
and Biological Significance. BirkhSuser Verlag, Basel, Boston, Berlin.
Kastan,M.B., Gowans.B.J. and Lieverman,M.W. (1982) Methylation of
deoxycytidine incorporated by excision repair synthesis of DNA. Cell, 30,
509-516.
Kato.H. (1973) Induction of sister chromatid exchanges by u.v. light and its
inhibition by caffeine. Exp. Cell Res., 82, 383-390.
LaviaJ., Ferraro, M., Micheli.A. and Olivieri.G. (1985) Effect of 5-azacytidine
(5-azaC) on the induction of chromatid aberrations (CA) and sister chromatid
exchanges (SCE). Mutat. Res., 149, 463-467.
LiebermannJvI.W., Beach.L.R. and Palmiter.R.D. (1983) Ultraviolet radiationinduced metallothionein-I gene activation is associated with extensive DNA
methylation. Cell, 35, 207-214.
Lindenhahan.M. and ShuberU. (1983) On the origin of hydroxyurea induced
chromatid aberrations in G2 chromosomes with BrdUrd in only one of the
sister chromatids. Mutat. Res , 108, 301-316.
Morgan.W.F. and Fero.M.L. (1987) Induced sister chromatid exchange
frequency is not increased in homogeneously staining regions that contain
amplified genes. Cancer Genet. Cytogenet., 26, 245-251.
MuscarellaJXE. and Bloom.S.E. (1982) The longevity of chemically induced
sister chromatid exchanges in Chinese hamster ovary cells. Environ.
Mutagen., 4, 647-655.
Perticone.P, Cozzi.R. and Gustavino.B. (1987) Sister chromatid exchanges
induced by demethylating agents persist through several cell cycles in
mammalian cells. Carcinoxenesis, 8, 1059-1063.
Perticone.P., Palitti.F., Cozzi.R., D'Erme.M and Bona.R. (1990) Persistence
of azacy tidine-induced SCEs and genomic methylation in CHO cells in vitro.
Mutat. Res., 245, 211-215.
Perticone.P., Linguardo,M., Cozzi.R., Corbo.R.M. and Polani.S. (1993)
Persistence of sister chromatid exchange by 9-p-D-arabinofuranosyladenine
in Chinese hamster ovary cells cultivated in vitro. Mutagenesis, 8, 445-448.
Popescu.N.C Amsbaugh.S.C. and DiPaoloJ.A. (1981) Relationship of
carcinogen induced sister chromatid exchange and neoplastic cell
transformation. Int. J. Cancer, 28, 71-77.
Pofcsai.N.C, Amsbaugh,S.C. and DiPaoloJ.A. (1985) Persistence of sister
263
P.Perticone, G.Gensabella and R.Cozzi
chromatid exchanges and in vitro morphologica] transformation of Syrian
hamster fetal cells by chemical and physical carcinogens. Carcmogenesis,
6, 1627-1630.
Sandberg,A.A. (ed.) (1982) Sister Chromatid Exchange. Alan R.Liss Inc,
New York.
Santi.D.W., GarrettX E and Barr.P.J. (1983) On the mechanism of inhibition
of DNA-cytosine methyltransferase by cytosme analogs. Cell, 33, 9-10.
SchwartzmanJ.B , Goyanes.VJ., Campos.A., Lage.A.M., Veiras.C,
Silva.MC. and Ramos.S (1985) Persistence of DNA lesions and the
cytological cancellation of sister chromatid exchanges. Chromosoma, 92,
7-10.
Shaffer.D.A. (1977) Replication bypass model of sister chromatid exchanges
and implications for Bloom's syndrome and Fanconi's anemia. Hum. Genet.
39, 177-190.
Shaffer.D.A (1977) Replication bypass model of sister chromatid exchanges
and implications for Bloom's syndrome and Fancom's anemia Hum. Genet.,
39, 177-190.
Spruck.C.H ,111, Rideout.W.M ,111 and JonesAP (1992) DNA methylation
and cancer. In JosU-P. and Saluz.H P. (eds), DNA Methylation. Molecular
Biology and Biological Significance. Birkhauser Verlag, Basel, Boston,
Berlin.
Tada-Aki,H. (1983) Induction of chromosome decondensation, sister chromatid
exchanges and endoreduplication by 5-azacytidine, an inhibitor of DNA
methylation. Mutat Res., Ill, 47-52
Takeshita.T. and Conner.M.K. (1985) Persistence of cyclophosphamideinduced damage in bone marrow as indicated by sister chromatid exchanges
analysis Carcmogenesis, 6, 1097-1102.
Tice.R.R. and Schwanzman.J B. (1982) Sister chromatid exchange- a measure
of DNA damage persistence In Sandberg.A A (ed.) SCE. Alan R Liss,
New York, pp. 33-45.
Tilley.S.A and Birshtein.B.K. (1985) Unequal sister chromaud exchange A
mechanism affecting Ig gene arrangement and expression. J. Exp. Med.
162, 675-694
Received on December 13, 1996; accepted on March II, 1997
264