Mutagenesis vol.12 no.3 pp.133-140, 1997
Indication for thresholds of chromosome non-disjunction versus
chromosome lagging induced by spindle inhibitors in vitro in
human lymphocytes
Azeddine Elhajouji1'3, Fabian Tibaldi2 and
Micheline Kirsch-Volders1
'Anthropogenetics Laboratory, Vrije Universiteit Brussel, Pleinlaan 2,
1050-Brussels, Belgium and Institute de Calculo, University of Buenos
Aires, Argentina
3
To whom correspondence should be addressed
Risk assessment from exposure to spindle inhibitors should
take into account the possibility of threshold concentrationresponse curves for aneuploidy induction. We analysed
concentration-dependent induction of chromosome nondisjunction by well known spindle poisons (colchicine,
carbendazim, mebendazole and nocodazole) and a reference
clastogen, methyl methanesulphonate (MMS) in vitro in
human lymphocytes; and integrated these findings with
earlier results of chromosome loss in micronuclei. Chromosome non-disjunction was estimated on cytokinesis-blocked
lymphocytes after simultaneous fluorescent in situ hybridization labelling with two chromosome-specific centromeric
probes (chromosomes 1 and 17). The frequencies of spontaneous non-disjunction showed important inter-individual
variations and were surprisingly high (7.04-15.39%).
Lower concentrations of aneugens did not induce a statistically significant increase of non-disjunction frequencies over
the respective control levels, whereas higher concentrations
clearly induced a concentration-dependent increase in the
non-disjunction frequencies with the four aneugens tested.
On the contrary, even at high concentrations, MMS induced
a slight increase in the frequency of non-disjunction but
without being statistically significant when compared with
the control frequencies. We estimated the inflection points,
the first statistically significant concentrations, the last nonstatistically significant concentrations and the number of
events from concentration-response curves of chromosome
non-disjunction and chromosome loss. A threshold-type
of concentration-response for non-disjunction is highly
probable for colchicine and nocodazole. For carbendazim
and mebendazole the inflection point fell above the first
statistically significant concentrations. But since it is obvious from dose-response curves where the inflection point/
threshold lies, it appears that the model might be picking
up some irregularities (possibly due to experimental variability in the dose—response curve at concentrations greater
than the threshold). For accurate estimation of the threshold, analysis of more concentrations or more cells might
be needed. Our data strongly indicate that in cultured
human lymphocytes chromosome non-disjunction is a
major mechanism of aneuploidy induction by spindle inhibitors and since non-disjunction occurs at lower concentration than chromosome loss, the aneuploidy threshold should
be estimated on the basis of non-disjunction rather than
on micronuclei frequencies (chromosome loss).
Introduction
Aneuploidy characterizes any cell or organism with a chromosome number which deviates from a multiple of the haploid
• UK Environmental Mutagen Society/Oxford University Press 1997
set of chromosomes. Several potential targets/mechanisms for
chemically-induced aneuploidy are known (Liang et ai, 1985;
Dellarco et ai, 1986). Theoretically, these include: the chromosomes themselves (particularly the centromeres and telomeres),
the kinetochore proteins, microtubule synthesis and assembly,
the formation of mitotic dividing spindle, the synthesis and
functioning of the polar bodies, the movement of the segregating chromosomes on the spindle, membrane modifications
and those meiosis-specific events such as formation of the
synaptonemal complexes and recombinational exchanges.
Interference with such targets can lead to either chromosome
loss, resulting from lagging chromosomes at the anaphase, or
chromosome non-disjunction, where both sister chromatids
migrate to the same daughter nucleus.
The incidence of chromosomal non-disjunction in man is
relatively high compared with that in rodents (Bond and
Chandley, 1983). Meiotic non-disjunction has been implicated
in a wide range of abnormalities, such as fetal wastage, prenatal
and infant mortality, congenital malformation and mental
retardation (Bond and Chandley, 1983; Hook, 1985; Hoffmann
et ai, 1986). Mitotic non-disjunction is closely associated with
malignant diseases and neoplastic transformations (Evans,
1985). Indeed, there is considerable evidence that chemical
exposure leads to the induction of specific aneuploidies and
that they have a role in tumour progression (Oshimura and
Barrett, 1986; Fearon and Vogelstein, 1990; Herens et ai,
1992; Haesen et ai, 1993; Van Goethem et ai, 1995).
When comparing the type of mutations, ranging from gene
mutations on a single base (point mutation) to structural or
numerical changes of whole chromosomes with a number of
targets possibly involved in their induction, it is clear that
probably only genome mutations (numerical chromosome
changes) result from multi-target interaction with the mutagen;
indeed chromosome non-disjunction and loss are the consequences of the binding of aneugens on spindle or centriolar
tubulins, scaffolding and nuclear proteins or centromere and
kinetochore regions.
From a theoretical point of view, it is considered that
mutagens which can induce a lesion after interaction with a
single target, e.g. covalent binding with DNA, show doseresponse curves without threshold; indeed potentially any
DNA lesion which is not repaired may lead to a mutation.
Alternatively, mutations which require the involvement of
more than one target should follow threshold dose curves
(Crebelli and Carere, 1993; Parry et ai, 1993). As a consequence, for those mutagens which could be proven to exert
their effects only above a given threshold concentration, risk
assessment should be guided by this information. Up to now,
however, only few data are available to sustain these concepts
(Parry et ai, 1994). Therefore, our aim was to study in vitro
in human lymphocytes the concentration curve responses of
model multi-target acting aneugens with the most sensitive
cytogenetic methodologies allowing the detection of mutagenic
effects at low concentrations and to define the most appropriate
mathematical models to identify the critical concentrations.
133
A.Elhajouji, F.Tibaldi and M.Kirsch-Volders
In earlier studies we analysed the chromosome loss in
micronuclei by means of fluorescent in situ hybridization
(FISH) with human general centromeric probe (Elhajouji et al,
1995) and showed that chromosome loss induced by classic
spindle poisons [colchicine (COL), carbendazim (MBC),
mebendazole (MEB) and nocodazole (NOC)] seems to act via
a threshold. Indeed, with the compounds tested there was no
increase in the frequencies of centromere positive micronuclei
(MNCen+) over a range of lower concentrations followed by
steep direct increase to highly statistically significant values
of MNCen+ over a range of higher concentrations. However,
a recent study by our laboratory (Kirsch-Volders et al., 1996)
compared chromosome loss and chromosome non-disjunction
induced by ionizing radiation and observed that only a small
fraction of spontaneous and radiation-induced aneuploidy is
detected using the micronucleus assay and that non-disjunction
seems to be a far more important mechanism leading to
spontaneous and radiation-induced aneuploidy.
To have a general profile for the induction of aneuploidy
by chemicals and finally to assess the possible existence of a
threshold for aneuploidy, it is therefore necessary to investigate
the induction of non-disjunction by the compounds previously
analysed for chromosome loss in the same wide range of
concentrations.
In this paper, chromosome non-disjunction was estimated on
cytokinesis-blocked binucleated lymphocytes in combination
with FISH using chromosome specific centromeric probes
(pUC1.77 for chromosome 1 and D17Z1 for chromosome 17).
This combination allows accurate evaluation of non-disjunction
since artefacts are excluded from the analysis as only binucleates with the correct number of hybridization signals were taken
into account. Based on mathematical models, concentrationdependent effect and number of targets necessary for induction
of chromosome loss and chromosome non-disjunction were
assessed.
Materials and methods
Chemicals
MBC (CAS: 10605-21-7) was purchased from Aldnch Chemie (Steinheim,
Germany); MEB (CAS: 31431-39-7) from Sigma Chemical Co., Brussels,
Belgium; NOC (CAS: 31430-18-9) was provided by Janssen Pharmaceutica,
Beerse, Belgium. These three chemicals were dissolved in dimethylsulphoxide
(DMSO; Merck, Darmstadt, Germany) for spectroscopy. Methyl methanesulphonate (MMS; CAS: 66-27-3) was purchased from Merck, and dissolved
in phosphate-buffered saline (PBS). COL (CAS: 64-86-8) was purchased from
Janssen Chemica and dissolved in PBS. Cytochalasin B (Sigma) was dissolved
in DMSO and kept at a stock solution of 4 mg/ml at -20°C.
Cultures
Human peripheral blood samples were obtained from healthy women (for
COL, MBC and MEB) and men (for NOC and MMS) volunteers aged =£35
years. One donor per chemical was used. To allow the comparison between
chromosome loss data (Elhajouji et al., 1995) and non-disjunction the same
donor for a given chemical was used. Blood was drawn by venipuncture and
heparinized with Calparine® (Sanofi, Labaz, France). Undiluted whole blood
(10 ml) was stripped of erythrocytes and granulocytes using a LeucoPREP™
cell separation tube (Becton Dickinson Labware, Lincoln Park, NJ, USA).
Mononuclear cells were then put in 1 ml multi-well culture plates (Becton
Dickinson) at a concentration of 0.5X10 6 cells/ml (1 ml final volume) and
incubated in a humidified CO 2 (5%) incubator at 37°C in Ham's F-10
medium supplemented with HEPES buffer (Gibco BRL, Bethesda, MD, USA)
containing 15% of fetal calf serum (Gibco). The lymphocytes were stimulated
with 2% phytohaemagglutinin (PHA 16; Wellcome Diagnostics, Dartford,
UK) and treated with cytochalasin B (6 Hg/ml) at 44 h After 72 h, cultures
were harvested. Cells were subjected to a cold hypotonic treatment (0.075 M
KC1), immediately centrifuged and fixed three times with fixative (methanol:
acetic acid, 3:1). The fixed cells were dropped onto slides using Pasteur pipettes,
air-dried and stored at -20°C.
134
Treatment
Chemicals were always added in a volume =£20 \l\ for isolated lymphocyte
cultures. DMSO, if required, was always added at the same final concentration
of 1%, including the DMSO used for cytochalasin B.
The test products were added to the cultures 24 h after PHA stimulation
and lasted for 48 h except with colchicine where the treatment lasted for 19 h.
Two cultures were made in parallel for each concentration.
FISH
FISH with probes for centromeric regions of chromosomes 1 (pUC1.77) and
17 (D17Z1) were used. The probes were labelled by nick translation according
to the instructions of the suppliers (Life Technologies BRL, Paisley, Scotland).
FISH was performed as described by Pinkel et al. (1986) with some
modifications. Slides were treated with RNase (Sigma) [0.1 mg/ml in 2X
sodium chloride/sodium citrate (SSC) for 1 h and a mild pepsin solution
(Sigma) (0.001% in 10 mM HC1) for 10 min at 37°C in a waterbath. The
slides were denatured in 70% formamide/2x SSC for 2 min and dehydrated
in an ethanol series (50:70:100%). The probes were denatured at 90°C and
placed on each slide. Following an overnight hybridization at 37°C in a moist
chamber, the slides were washed with 55% formamide in 2X SSC at 42°C.
Detection of the biotinylated-labelled probe for chromosome 17 was performed
by means of avidin-FITC (fluorescein avidin D; Vector Laboratories, Burlingame, CA, USA) and biotinylated goat anti-avidin antibodies (Vector), allowing
signal amplification. The digoxigenin-labelled probe (chromosome 1) was
detected using a mouse anti-digoxigenin antibody (Boehringer Mannheim,
Mannheim, Germany) followed by a Texas Red™-conjugated sheep antimouse antibody (Amersham, Little Chalfont, UK). After dehydrating in an
ethanol series (50:70:100%), the cells were counterstained with 4',6-diamido2-phenylindole (DAPI; Boehringer Mannheim) in a p-phenylenediamine
antifade solution (Johnson and de C.Nogueira Araujo, 1981).
FISH analysis
For chromosomal non-disjunction analysis, 1000 binucleated lymphocytes per
culture (two cultures per concentration) were examined for chromosome 1
and 17 segregation. To restrict the scoring to the first mitosis after the
treatment and exclude technical artefacts only binucleated cells having the
diploid number (four spots for a given chromosome) of hybridization signals
were analysed. The preparations were examined with a Leitz Dialux 20
fluorescence microscope equipped with a double band pass filter (Leitz G/R,
excitation at 490/575 nm, emission at 525/635 nm) to allow the simultaneous
observation of both labelled chromosome centromeres and a single band pass
filter (Leitz A, excitation at 340-380 nm, emission at 430 nm) to visualize
the DAPI counterstaining.
To obtain chromosomal non-disjunction frequencies for the total genome
(% Tot. ND), the initial frequencies of non-disjunction for chromosomes 1
and 17 were multiplied by 23/2, assuming that non-disjunction occurs randomly
and independently for each chromosome.
Statistical analysis
Statistical differences between controls and treated samples were determined
with the x 2 test.
For the estimation of the inflection points, the %MNCen+ and the %oND and
(1 + 17) frequencies were used for chromosome loss and for chromosome nondisjunction respectively. The mathematical models for fitting chromosome
loss and chromosome non-disjunction data were respectively:
yi
= P, X ( l + e h 2
= Po + Pi* + P2* +
)"' + p 4 and
+ Pit 4 + 8<
x = compound concentration (u.M).
yx = estimated frequency (%) of chromosome loss by the mathematical model
at a given concentration x.
y2 = estimated frequency (%) of chromosome non-disjunction by the mathematical model at a given concentration x.
Pn, Pi
P5 = parameters to be estimated for the best fit with the
experimental data.
These functions were chosen since they showed the best fitness to the data.
Different models were tested to fit the data (e.g. exponential models, logistic
and semi-logistic, polynomial with different degrees). The best fitness to the
data was based on the correlation coefficient (r) and the x2 t e s t values after
applying the least squares method. The parameters Po, Pi, ., P5 were estimated
by using least squares method and in all the cases the goodness-of-fit was
statistically significant.
The number of events (X) necessary to induce chromosome loss and
chromosome non-disjunction was determined in function of the P values as
calculated by the Wilcoxon signed rank test (Conover, 1980). The multi-hit
and the multi-target equations (Kellerer, 1989) represent the general notion
that random hits (binding of a chemical molecule) accumulate lesions up to
Chromosome non-disjunction and chromosome tagging induced by spindle inhibitors
a critical threshold. Assuming identical targets one can utilize the Poisson
formula for the probability of i events at an expected frequency u. Postulating
a critical threshold of x events one obtains the multi-hit equation:
*
n=l-e-". V
;T0
u'
x
.•!
u2
li
.where V
. . o ••!
= 1 + H+
2
u3
+
6
if*
+•••+
u is assumed to be proportional to the mutagen concentration; p is the
percentage (probability) of MN containing a whole chromosome or of
chromosome non-disjunction events in function of the number of hits as a
consequence of at least x events. Since we cannot make any assumption on
the distribution of the data, the Wilcoxon signed rank test (Conover, 1980)
was applied in each case to analyse the existence of differences between the
experimental data and theoretical data for fixed values of x. Small values of
P indicate evidence of statistically significant differences between experimental
and theoretical data for each concentration. The raw data obtained directly
from the scoring of centromere positive flow sorted micronuclei (chromosome
loss) and non-disjunction provided the frequencies of spontaneous and induced
aneuploidy but overlooked probable differences in the mechanisms leading to
the presence of aneugenic events in untreated lymphocytes (spontaneous
background level) and their induction by aneugens. Therefore no calculation
of number of targets were performed on this data set where the two
different mechanisms of aneuploidy induction could not be distinguished. For
chromosome non-disjunction data, the number %cND (1 + 17) frequency, after
deduction of spontaneous non-disjunction (frequencies in controls), was used.
When the values obtained on the flow sorted micronuclei were used to estimate
the actual frequencies of centromere positive micronuclei in Cytochalasin-B
blocked binucleated lymphocytes, all frequencies were related to 1000 binucleates and allowed deduction of spontaneous background frequencies from those
observed in treated cells:
0
2.5 5
10 15 20 25 37.5 50 75 100
Cnnr ',.M1 \ '•»'
Fig. 1. Frequencies of non-disjunction for chromosomes 1 and 17 after
treatment with increasing concentrations of COL.
A = [(%cMNCBX%MNCen+),realed - (%cMNCBX%MNCen+)cm,rot] X 100.
For the estimation of the number of events necessary for the induction of
chromosome fragments by MMS, the same model as for chromosome loss
was used. The frequencies of MNCen" of flow sorted MN are related to 1000
binucleates allowing the deduction of spontaneously induced MNCen":
A = [(%cMNCBX%MNCen-)maled - (%oMNCBX%MNCen-)control] X 100.
Results
Chromosome non-disjunction in the macronuclei of cytokinesisblocked lymphocytes
Figures 1-5 give respectively for COL, MBC, MEB, NOC
and MMS the frequencies of non-disjunction between the
two daughter nuclei after simultaneous FISH labelling with
two chromosome-specific centromeric probes. The targeted
chromosomes were chromosomes 1 and 17 respectively. The
results (Figures 1-5) indicate that the frequencies of spontaneous non-disjunction (%oNDCB) do not differ significantly
between both chromosomes; chromosome 17 was however
always slightly more frequently involved in spontaneous
non-disjunction than chromosome 1 for the different donors
analysed and treatment with spindle inhibitors often reversed
this tendency at higher concentrations.
To estimate chromosomal non-disjunction frequencies for
the total genome, assuming that non-disjunction occurs randomly, the initial frequencies of non-disjunction for chromosomes 1 and 17 were multiplied by 23/2 (Table I). The
frequencies of total spontaneous non-disjunction are quite high
and show important inter-individual variations ranging from
7.04 to 15.39% in cytokinesis-blocked lymphocytes (Figure 6).
Lower concentrations of aneugens did not induce a statistically significant increase of non-disjunction frequencies over
the respective control levels, whereas higher concentrations
clearly induced a concentration dependent increase in the nondisjunction frequencies with the four aneugens tested. On the
contrary, MMS even at high concentrations induced a slight
increase in the frequency of non-disjunction but without being
Q
0 0.15 0.26 0.52 1.04 1.56 2.61 3.92 5.23 7.84 10.46
Cone. feiM)
Fig. 2. Frequencies of non-disjunction for chromosomes 1 and 17 after
treatment with increasing concentrations of MBC.
statistically significant when compared with control frequencies
(Figure 7).
In an ideal situation, the no-observed-effect level (NOEL)
or threshold corresponds to a single concentration above which
the compound exerts quite directly its maximal effect; however
since experimentally the concentrations were tested in a
discontinuous range, mathematical approaches were needed to
estimate the critical biological threshold.
Mathematical analysis of concentration-effect curves
Since aneuploidy can result from chromosome loss as well as
from non-disjunction, to assess a no-observed-effect level for
aneuploidy both non-disjunction and chromosome loss data
135
A.Elhajouji, F.Tibaldi and M.Kirsch-Volders
Ch. 1 I
Ch. 171
70
1 IiII 1
——
50 -
—,
30
20
11
••
0
.017 .033 .067 .135 .203 .271 .305 .338 .677 1.016
Fig. 3. Frequencies of non-disjunction for chromosomes 1 and 17 after
treatment with increasing concentrations of MEB.
0
2
3
7
13
20
27
33 66
100 132
Fig. 4. Frequencies of non-disjunction for chromosomes 1 and 17 after
treatment with increasing concentrations of NOC.
should be taken into account for the final evaluation. Therefore,
fitting of concentration-effect curves with adequate mathematical models was performed on non-disjunction data and on
chromosome loss data as well (Table I).
In concentration-response curves, the inflection point (IP)
may be useful as a parameter for an approximate determination
of the threshold level. We assumed that when the inflection
point is situated between the first statistically significant (FSS)
and the last non-statistically significant (LNSS) concentrations,
the inflection point would correspond to the threshold. To
estimate the inflection points, different mathematical models
were fitted to the data and those which showed the best
goodness-of-fit with less number of parameters were selected.
136
0
29
49
78
98 117 147 176 196 245 294
Cone. |iM)
Cone. (|iM)
Fig. 5. Frequencies of non-disjunction for chromosomes 1 and 17 after
treatment with increasing concentrations of MMS.
Two different mathematical models were applied on two types
of frequency data: the frequencies of centromere positive
micronuclei on flow sorted micronuclei and the frequencies
of chromosome non-disjunction in cytokinesis-blocked
lymphocytes.
Results from the calculations of inflection points in comparison with FSS and LNSS concentrations (Table II) clearly
indicate that for COL, MBC and NOC, the inflection point is
situated between the first statistically significant and the
last non-statistically significant concentrations, suggesting a
threshold mechanism for chromosome loss. For MEB the
inflection point is higher than the first statistically significant
concentration, however a threshold mechanism, as suggested
(Elhajouji et ai, 1995), might not be excluded since a NOEL
was shown at lower concentrations. From data obtained with
specific chromosome probes to analyse chromosome nondisjunction in binucleated cytokinesis-blocked lymphocytes,
the inflection point was situated between the first statistically
significant and last non-statistically significant concentrations
for COL and NOC; however, for MBC and MEB the inflection
point was higher than the respective first statistically significant
concentration.
Estimation of number of targets necessary for chromosome
loss and chromosome non-disjunction
When the values obtained on the flow sorted micronuclei were
used to estimate the actual frequencies of centromere positive
micronuclei in cytochalasin B-blocked binucleated lymphocytes, all frequencies were related to 1000 binucleates and
allowed deduction of spontaneous background frequencies
from those observed in treated cells. On the basis of this data
set, the analysis of the number of targets became relevant.
Figure 8 and Table II show that for COL, MBC, MEB and
NOC, the number of hits needed to be affected to induce
chromosome loss was higher than five for MBC or six for
COL, MEB and NOC. In contrast for MMS, a known clastogen
which induced increases only in chromosome breaks as centro-
Chromosome non-disjunction and chromosome lagging induced by spindle inhibitors
Table I. Frequencies of micronucleated CB, chromosome loss and non-disjunction
Cone. (uM)
Colchldne
Frequency of MN Chromosome loss
%o MNCB a
%MNCen+ D
*'
• *•
** '
** '
3.50
6.50
6.00
4.73
6.71
7.00
3.50
9.50
12.88
25.17
32.02
** "
** '
** '
2.08
3.45
7.35
5.71
3.84
5.03
7.81
8.43
20.91
55.51
89.74
PBS
0.0025
0.0050
0.0100
0.0150
0.0200
0.0250
0.0375
0.0500
0.0750
0.1000
•«•
••*
•••
•••
34.65
35.00
32.00
30.69
37.00
35.00
40.00
77.22
85.53
87.93
90.22
•
•
•
•
•
28.70
32.00
29.70
34.00
34.65
34.00
68.07
74.89
84.84
92.27
96.45
•
*•
• 4
•
**
•
*4
•
38.00
38.24
40.59
39.00
39.60
41.00
61.26
61.19
62.96
78.51
88.89
•••
•••
•*•
36.00
32.67
35.00
38.61
38.24
41.00
40.77
34.83
84.94
91.27
93.63
Chromosome non-disjunction
%o ND (1+17) c
% Tot. ND d
•
••
•*•
•••
••*
•••
6.13
7.45
5.50
9.49
9.09
25.28
37.97
45.72
59.29
64.38
64.24
•
••
"
•••
•••
•••
"•
9.99
13.47
12.35
13.40
29.83
35.88
41.16
45.28
59.06
57.41
65.30
•
•
*• •
*• •
• *•
*• •
7.04
8.57
6.32
10.92
10.45
29.07
43.67
52.58
68.18
74.03
73.88
•
•
•
• ••
*• •
• «•
**•
11.49
15.49
14.20
15.40
34.31
41.26
47.33
52.07
67.92
66.02
75.10
•
•
•
•* •
** •
**•
*• *
14.82
14.42
13.11
11.26
28.88
34.49
40.92
54.90
58.41
68.29
68.97
•
•••
•••
•••
11.79
10.34
12.22
9.64
11.88
14.17
18.70
35.77
61.01
68.35
75.33
CaitoendazJm
DMSO
0.157
0.262
0.523
1.046
1.569
2.615
3.923
5.231
7.846
10.461
Mebendazole
DMSO
0.017
0.033
0.067
0.135
0.203
0.271
0.305
0.338
0.677
1.016
Nocodazole
DMSO
0.002
0.003
0.007
0.013
0.020
0.027
0.033
0.066
0.100
0.132
MMS
PBS
29.36
48.93
78.32
97.90
117.44
146.80
175.80
195.80
244.70
293.70
6.85
7.50
8.25
7.00
9.00
11.00
17.00
•*
21.50
•*
34.29
• • » 159.93
**
285.71
**
•*
• *<
••
••
*•
••
1.50
1.10
1.50
1.50
3.00
2.62
2.14
2.58
10.77
45.57
93.97
4.09
5.81
1.81
4.35
6.74
9.12
6.36
26.57
18.43
31.66
55.15
36.27
33.66
30.00
29.81
28.43
24.75
23.00
23.50
18.50
16.00
9.95
*»
12.89
12.53
11.40
9.79
25.11
29.99
35.58
47.74
50.79
59.38
59.97
•
"•
•••
•••
10.26
8.99
10.62
8.39
10.33
12.32
16.26
31.10
53.06
59.43
65.51
*
•••
*• '
*• '
13.38
12.00
12.59
11.57
11.85
16.08
17.30
15.67
17.49
18.48
18.49
15.39
13.80
14.47
13.30
13.63
18.50
19.90
18.02
20.11
21.25
21.27
"Frequencies of micronucleated cytokinesis-blocked lymphocytes (%oMNCB). 1000 cytokinesisblocked lymphocytes per culture were examined.
b
Centromere positive micronuclei (%MNCen + ) as a measure for chromosome loss. 200 flow sorted
micronuclei per concentration were examined for the presence of centromeric spots.
Trequencies of non-disjunction for chromosomes 1 and 17 [%cND (1 + 17)].
••Chromosomal non-disjunction for the total genome (%Tot. ND). To obtain these frequencies, the
frequencies on column c were multiplied by 23/2, assuming that non-disjunction occurs at random
and independently for each chromosome.
"P < 0.05 compared with the respective control in the frequency of MNCB, MNCen + or ND;
»•/> < 0.001; ***/> < 0.0001.
mere negative MN, the number of events needed to induce an
acentric chromosome fragment was three.
The number of events necessary for chromosome nondisjunction, was lower than for chromosome loss with all
aneugens tested (Figure 8 and Table II). The number of
events was two for MBC and three for both COL and MEB
respectively. The highest number of events was seen with
NOC where five targets should be affected by the aneugen to
137
A.Elhajouji, F.Tibaldi and M.Kirsch-Volders
1.2
1
• - • • A
0.8
.
.
.
<
>
-
-•-•&
• - -
•
-
—
COL
MBC
NOC
MEB
MMS
p= 0.05
0.6
a.
0.4
0.2
:~.A-
3
Fig. 6. Frequencies of spontaneous non-disjunction for chromosomes 1 and
17 analysed in five different donors.
1
0.8
—•
•••••
A
--O
-COL
• MBC
-NOC
-MEB
10
4
5
6
7
number of events (X)
a
•
•••& •• M M S
-a
k
A'
0.6
•' .'
1
1
•••••
0.2
'\ •
^D'
. ' '" ' o..
•
0.4
A
'
0.
A
0
I
• A
i
i
i
3
O
*
t
1
1
4
5
6
7
number of events (X)
Fig. 8. The number of events (X) necessary to induce chromosome loss
(upper graph) and chromosome non-disjunction (lower graph) in function of
the P values as calculated by the Wilcoxon signed rank test.
0.01
1
10
100
Cone. (MM)
Fig. 7. The frequencies of chromosomal non-disjunction for the total
genome (data from Table I, column d) identified by FISH with human
chromosomes 1 and 17 specific probes in binucleated human lymphocytes
after exposure to respective mutagens. 1000 binucleates per culture were
examined with FISH. *The lowest concentration tested that showed a
statistically significant increase (P < 0.05) in non-disjunction compared
with the respective control (see Table I).
Table II. Statistical analysis of the data on chromosome loss and
chromosome non-disjunction
Chromosome loss
No. events (X)
LNSS concentration
Infection point (mM)
FSS concentration
Chromosome non-disjunction
No. events (X)
LNSS concentration
Infection point (mM)
FSS concentration
COL
MBC
NOC
MEB
> 6
0.025
0.033
0.037
>5
1.57
2.47
2.61
> 6
0.033
0.053
0.066
> 6
0.203
0.286
0.271
>3
0.015
0.020
0.020
>2
0.523
2.847
1.046
>5
0.027
0.032
0.033
>3
0.067
0.228
0.135
LNSS, last non-statistically significant; FSS, first statistically significant.
lead to chromosomal non-disjunction. For MMS the number
of events estimated for the induction of chromosome nondisjunction was three.
138
Discussion
Selection of the dose-incidence model that is appropriate for
predicting the risks of cancer or any adverse effect from lowlevel exposure to a given carcinogen or mutagen is among the
most contentious issues in public health (Upton, 1989). The
approximate extrapolation of risk from high to low doses is
usually applied. It might be accurate for effects that suppose
a linear dose-effect relationship. However, the best evaluation
of risk from exposure to mutagens or carcinogens, where a
non-linear dose-effect relationship is shown, is the analysis of
wide dose ranges that allowing estimation of critical points
such as a NOEL.
This study aimed to analyse concentration-dependent induction of chromosome non-disjunction by well known spindle
inhibitors in vitro using a wide range of concentrations
and integrate these findings with those of chromosome loss
(Elhajouji et al, 1995) to have a general profile of concentration-dependent induction of aneuploidy.
The frequencies of spontaneous non-disjunction showed
important inter-individual differences and were surprisingly
high, specially when total frequencies were estimated (7.04—
15.39%); the data are however close to those obtained by our
laboratory (18.4%) and others (11.5%) in a previous study
(Kirsch-Volders et al, 1996). As already suggested, it might
indicate either that in vitro the non-disjunction rate is increased,
that a strong selection occurs against aneuploid cells in vivo
or that aneuploidy does not affect the different chromosomes
at random.
Chromosome non-disjunction and chromosome lagging induced by spindle inhibitors
-ANEUPLOIDY
-NON-DISJUNCTION
- C H R O M O S O M E LOSS I
ANEUPLOIDY
NON-DISJUNCTION
CHROMOSOME LOSS
100.0
80.0
60.0
40.0
20.0
0.0
-0.02
0.02
0.04
0.06
0.08
0.1
Cone. (&iM)
-ANEUPLOIDY
- NON-DISJUNCTION
- C H R O M O S O M E LOSS
MEBENDAZOLE
ANEUPLOIDY
NON-DISJUNCTION
CHROMOSOME LOSS
100.0
80.0
60.0
40.0
20.0
0.0
-0.2
0.4
0.6
0.8
Cone. (ixM)
0 02
0.04
0.06
0.08
0.1
Cone. UM)
Fig. 9. Frequencies of non-disjunction for the total genome, chromosome loss and total aneuploidy induced by the different chemicals tested. To obtain the
frequencies of non-disjunction for the total genome, frequencies of non-disjunction for chromosomes 1 and 17 were multiplied by 23/2, assuming that ND
occurs at random and independently for each chromosome. Frequencies of chromosome loss were obtained from data of FISH centromere probing on flow
sorted micronuclei related to 1000 binucleates. The frequencies of total aneuploidy are the sum of non-disjunction and chromosome loss.
In this work, however, only slight differences between nondisjunction for chromosome 1 in comparison with chromosome
17 were observed in both controls (spontaneous frequencies)
and treated cells (chemically induced). These observations
disagree with non-disjunction frequencies for the same chromosomes in the same cells in a previous work (Kirsch-Volders
et al, 1996). These discrepancies might be due to interindividual variations for spontaneous non-disjunction or to the
fact that ionizing radiation affects chromosomes differently
depending on their centromeric heterochromatin content.
With the analysis of concentration-effect for non-disjunction,
a threshold-type of concentration-response for the tested
aneugens is dependent on the appropriateness of the modelled
inflection points. For COL and NOC the inflection point is
situated between the last non-statistically significant and the
first statistically significant concentration; for MBC and MEB
the inflection point fell above the first statistically significant
concentration. But since it is obvious from dose-response
curves where the inflection point/threshold lies, it appears that
the model might be picking up some irregularities (possibly
due to experimental variability in the dose-response curve
at concentrations greater than the threshold). For accurate
estimation of the threshold, analysis of more concentrations
or more cells might be needed.
When total aneuploidy was considered, the contribution of
chromosome non-disjunction seemed to be more important
than chromosome loss (Table I and Figure 9). Moreover,
since non-disjunction occurs at lower concentrations than
chromosome loss, the aneuploidy threshold should therefore
be estimated on the basis of non-disjunction rather than on
micronuclei frequencies (chromosome loss).
With this assumption, models can be selected to estimate
139
A.Elhajouji, F.Tibaldi and M.Kirsch-Volders
adverse effects of aneugens and, as for genotoxicants, risk
assessment might be determined in the future for chemical
aneugens. Whether the calculated threshold values correspond
to the real biological effect level is difficult to assess at the
present time; the difference between both will of course play
an important role in the final risk assessment decision.
Another more indirect way to approach the existence of
potential thresholds is to calculate the number of targets
involved in the induction mechanism. Indeed, the more targets
that have to be inactivated before the effect level is reached,
the higher the probability of observing a threshold curve
response.
When analysing the number of targets needed to be affected
by the spindle inhibitors to lead to aneuploidy, chromosome
loss seemed to require a higher number of affected targets
(five or six) than non-disjunction (two, three or five). This is
in agreement with the theoretical mechanism. Indeed, since
tubulin depolymerization is the main aneugenic mechanism to
be considered for the compounds studied, absence of some
tubulin protofi laments at one kinetochore is probably sufficient
to induce non-disjunction as compared with chromosome loss
where the kinetochores at both sides of the joined sister
chromatids should be detached from an organized spindle. At
concentrations higher than the ones inducing chromosome
loss, the whole mitotic apparatus might be affected resulting
in polyploidy (Athwal and Sandhu, 1985). Why the number
of events calculated as necessary to induce non-disjunction is
higher for NOC (five events) than for the other compounds
tested (two or three) is unclear. The four spindle inhibitors are
considered to have the same binding site on P-tubulin (Liang
and Brinkely, 1985; Nogales et al, 1995) and the three
benzimidazoles differ only by the substituents to the benzimidazole nucleus: NOC (6-C4H3SCO), MEB (5-C6H5CO) and MBC
(5-H).
Concomitantly, MMS was studied as a reference clastogen.
As shown with ionising radiation (Kirsch-Volders et al, 1996),
MMS induced a concentration-dependent increase of nondisjunction but without being statistically significant. A concentration-dependent increase of centromere negative MN
(Elhajouji et al, 1995) was of course found in the in vitro
micronucleus test and the number of targets which needed to
be affected to give rise to a MN was estimated equal to three.
This number of targets is not incompatible with the idea that
a chemically-induced centromere negative MN results from
two single strand DNA breaks or one non-repaired single
strand break, replicated in S phase. This would indirectly
suggest that in particular conditions of exposure, concentrationresponse curves for clastogenicity might show some type of
biological threshold.
In conclusion, the data presented here strongly suggest that
non-disjunction is a major mechanism of aneuploidy induction
by spindle inhibitors. The number of targets involved for
non-disjunction seemed to be lower than for induction of
chromosome loss. Since non-disjunction also occurs at lower
concentrations than chromosome lagging and that the threshold
concentration response are less clear for the former, it is highly
advisable to study accurately non-disjunction frequencies
before drawing conclusions about NOEL for aneugens.
The in vitro micronucleus test, when studied in combination
with FISH for specific chromosome or/and centromere regions,
allows discrimination between clastogenic events, chromosome
loss and chromosome non-disjunction; it becomes clear that
this easy, quick, well reproducible in vitro test provides a
140
unique tool to accurately approach risk assessment of both
clastogens and aneugens.
References
Athwal,R.S. and Sandhu.S.S. (1985) Use of a human X mouse hybrid cell
line to detect aneuploidy induced by environmental chemicals. Mutat. Res.,
149,73-81.
Bond,D.J. and Chandley.A.C. (1983) Aneuploidy. Oxford Monographs on
Medical Genetics, no. 11, Oxford University Press, London, pp. 198.
Conover,W.J. (1980) Practical Nonparametric Statistics. 2nd edn. John Wiley
and Sons, New York.
Crebelli.R. and Carere,A. (1993) Aneuploidy assays in routine screening? In
Madle.S. and Muller.L. (eds), Current Issues in Genetic Toxicology, bgaSchriften 1/93, MMV, Medizin-Verlag Munchen, pp. 70-89.
Dellarco,V.L., Mavournin.K.H. and Waters.M.D. (1986) Aneuploidy data
review committee: summary compilation of chemical data base and
evaluation of test methodology. Mutat. Res., 167, 149-169.
Elhajouji.A., Van Hummelen.P. and Kirsch-Volders,M. (1995) Indications for
a threshold of chemically induced aneuploidy in vitro in human lymphocytes.
Environ. Mol. Mutagen., 26, 292-304.
Evans,H.J. (1985) Neoplasia and cytogenetic abnormalities. In Dellarco.V.L.,
Woytek.P.E. and Hollaender.A. (eds), Aneuploidy: Etiology and Mechanisms.
Basic Life Sciences Vol. 36. Plenum Press, New York, pp. 165-177.
Fearon,E.R. and Vogelstein.B- (1990) A genetic model for colorectal
tumorigenesis. Cell, 61, 759-767.
Haesen,S., Timmermans.M. and Kirsch-Volders.M. (1993) Induction of
micronuclei and karyotype aberrations during in vivo mouse skin
carcinogenesis. Carcinogenesis, 14, 2319-2327.
Herens.C, Gonzales.M.L.A. and Barbason.H. (1992) Cytogenetic changes in
hepatocarcinogenesis from rats treated with chronic exposure to
diethylnitrosamine. Cancer Genet. Cytogenet., 60, 45-52.
Hoffmann,G.R., Dellarco.V.L. and Voytek.P.E. (1986) A review of the
symposium on aneuploidy: etiology and mechanisms. Environ. Mutagen.,
8,643-651.
Hook.E.B. (1985) The impact of aneuploidy upon public health: mortality
and morbidity associated with human chromosome abnormalities. In
Dellarco.V.L., Woytek.P.E. and Hollaender.A. (eds), Aneuploidy: Etiology
and Mechanisms. Basic Life Sciences Vol. 36, Plenum Press, New York,
pp. 7-33.
Johnson.G.D. and de C.Nogueira Araujo.G.M. (1981) A simple method of
reducing the fading of immunofluorescence during microscopy. /. Immunol.
Methods, 43, 349-350
Kellerer.A.M., (1989) Studies of the dose-effect relation. Experientia, 45,
13-21.
Kirsch-Volders.M., Tallon.I., Tanzarella.C, Sgura.A., Hermione.T., Parry.E.M.
and ParryJ.M. (1996) Chromosome non-disjunction is a major mechanism
for in vitro aneuploidy induction by X-rays in primary human cells.
Mutagenesis, 11, 307-313.
Liang.J C. and Brinkley.B.R. (1985) Chemical probes and possible targets for
the induction of aneuploidy. In Dellarco.V.L., Woytek.P.E. and Hollaender.A.
(eds), Aneuploidy: Etiology and Mechanisms. Basic Life Sciences Vol. 36,
Plenum Press, New York, pp. 491-505.
Nogales,E., Wolf,S.G., Khan, .A., Luduefia.R.F. and Downing,K.H. (1995)
Structure of tubulin at 6.5 A and location of the taxol-binding site. Nature,
375, 424-427.
Oshimura.M. and Barrett.J.C. (1986) Chemically induced aneuploidy in
mammalian cells: mechanisms and biological significance in cancer. Environ.
Mutagen., 8, 129-159.
ParryJ.M., Feilder.R.J. and McDonald,A. (1994) Thresholds for aneuploidyinducing chemicals. Muatgenesis, 9, 503—504.
Parry,J.M., Parry.E.M., Ellard.S., Warr,T., O'Donovan.J. and Lafi,A. (1993)
The detection, definition and regulation of aneugenic chemicals. In Vig.B.K.
(ed.), Chromosome Segregation and Aneuploidy. BK NATO ASI Series,
Vol. H12. Springer-Verlag, Berlin, Heidelberg, pp. 391-415.
Pinkel,D.. Straume.T. and GrayJ.W. (1986) Cytogenetic analysis using
quantitative, high-sensitivity, fluorescence hybridization. Proc. Nail. Acad.
Sci. USA, 83, 2934-2938.
Upton.A.C. (1989) The question of thresholds for radiation and chemical
carcinogenesis. Cancer Invest., 7, 267—276.
Van Goethem,F., de StoppelaarJ., Hoebee.B. and Kirsch-Volders.M. (1995)
Identification of clastogenic and/or aneugenic events during the preneoplastic
stages of experimental rat hepatocarcinogenesis by fluorescence in situ
hybridization. Carcinogenesis, 16, 1824—1834.
Received on April 15, 1996; accepted on January 29, 1997