Mutagenesis vol. 28 no. 3 pp. 371–374, 2013
Advance Access publication 27 February 2013
doi:10.1093/mutage/ges082
RESPONSE
RE: Insensitivity of the in vitro cytokinesis-block micronucleus assay with human
lymphocytes for the detection of DNA damage present at the start of the cell culture
(Mutagenesis, 27, 743–747, 2012)
Günter Speit*,†
Univeristät Ulm, Institut für Humangenetik, D-89069 Ulm, Germany
*To whom correspondence should be addressed. Tel: +49-731-50065440;
Email: [email protected]
On behalf of co-authors.
†
In their letter to the Editor, Michael Fenech and Micheline
Kirsch-Volders express their concerns about the design of
a study we published in Mutagenesis (1) and challenge our
conclusions. In our opinion, most of their criticisms are not
justified and here we take the opportunity to detail why.
1. Fenech and Kirsch-Volders criticise the title of our publication. However, our results actually demonstrate the insensitivity of the CBMN assay for the detection of DNA damage
present at the start of the lymphocyte culture. The insensitivity of this protocol in comparison with an in vitro protocol that is recommended by the OECD guideline 487 (2)
is shown for six different chemical mutagens including the
cross-linking agent mitomycin C. A difference in sensitivity
can also be shown for ionising radiation (Figure 1). Ionising
radiation, radiomimetic chemicals and cross-linking agents
do induce micronuclei (MN) when lymphocytes are exposed
before the start of the culture (because they induce DNA
double-strand breaks or other poorly repairable lesions),
whereas most of the so-called ‘S-phase dependent’ mutagens
(i.e. those which induce excision-repairable lesions) do not.
We discuss the insensitivity of the CBMN assay as typically
used for human biomonitoring, for mutagens causing excision-repairable damage, which is the majority of genotoxic
agents to which humans are exposed. It is not evident how
Fenech and Kirsch-Volders can conclude that there is ‘overwhelming evidence’ that this assay, as usually conducted,
is ‘highly sensitive’. Our results, which are appropriately
discussed, demonstrate the relative insensitivity of the biomonitoring protocol in comparison with the in vitro protocol
recommended by the OECD guideline 487 for the sensitive
detection of genotoxic compounds (2). However, based on
the MN data we presented (1), even this ‘sensitive’ protocol needs high levels of DNA damage to induce a significant
increase in the frequency of micronucleated binucleated cells
(BNC).
2. Fenech and Kirsch-Volders question the originality of our
observations and stress the necessity of the ARA-C protocol
to detect agents that predominantly induce excision-repairable lesions. However, as clearly stated in our publication, it
is the first attempt to directly compare two protocols of the in
vitro cytokinesis-block micronucleus assay (CBMN assay).
The results presented are new and have important implications for the interpretation of human biomonitoring studies.
The comment regarding the formation of MN is rather
confusing and needs clarification. There should be no doubt
that the CBMN assay detects MN formed as a consequence
of chromatid-type aberrations and chromosome-type aberrations (i.e. acentric fragments) produced within one cell
cycle after exposure. This is the basis for the detection of all
types of clastogens in the in vitro CBMN assay according
to OECD guideline 487 (2). If acentric chromosome fragments were necessary for the formation of MN, the majority
of clastogens would not be detected. Interestingly, the schematic diagram presented by Fenech and Neville (3) shows a
chromatid break as the cause of MN formation in the presence of ARA-C.
It is correct that in vitro studies have shown that ARA-C
enhances the formation of MN by agents inducing excisionrepairable lesions and these studies are discussed in our
publication (1). However, these studies also demonstrate the
insensitivity of the CBMN assay in the absence of ARAC. With one exception, human biomonitoring studies were
all performed without ARA-C and positive results after in
vivo exposure to agents inducing excision-repairable lesions
cannot easily be explained. The only biomonitoring study
that used this approach (in vitro ARA-C treatment for the
first 16 h of culture) failed to demonstrate any significant
effect on MN frequencies in BNC using blood from individuals potentially exposed to genotoxic pollutants and/or
tobacco smoke (4). If Fenech and Kirsch-Volders state that
‘the proper protocol to detect in vitro excision-repairable
DNA lesions and agents that predominantly induce them is
the ARA-C protocol’, does this mean that all biomonitoring
studies investigating such effects without ARA-C are inappropriately performed?
3. We agree that a key question is the comparability of exposure in vitro versus in vivo. However, the plausible assumption is that the types of DNA damage induced in vitro and in
vivo are the same and it has never been shown that damage
levels induced in vivo are higher than those induced in vitro
under controlled experimental conditions. Furthermore, the
DNA repair mechanisms involved in the removal (excision) of lesions induced in vitro should be equally sensitive
towards DNA damage induced in vivo and present in lymphocytes at the start of the culture.
We also agree that many studies reported ‘associations’
between exposure to chemical mutagens and increased MN
frequencies in human biomonitoring. But these associations
do not prove a causal relationship between DNA damage
induced in vivo and the frequency of MN in BNC.
Exposures to environmental and occupational chemicals
should not be equated with exposures resulting from chemotherapy. Chemotherapy includes exposure to high doses
of strong mutagens (including cross-linking agents) and
systemic cytotoxic effects that frequently lead to reduced
© The Author 2013. Published by Oxford University Press on behalf of the UK Environmental Mutagen Society.
All rights reserved. For permissions, please e-mail: [email protected].
371
G. Speit
500
2.2
2.1
450
2.0
400
1.9
MN [‰]
350
1.8
300
1.7
250
1.6
200
1.5
1.4
150
1.3
100
1.2
50
1.1
0
1.0
Co
2Gy - 0h
2Gy - 45h
Fig. 1. Frequency of micronucleated BNC and NDI in the CBMN assay after exposure of blood cultures to gamma-irradiation (2 Gy) at the start of the culture
(0 h) or 45 h later (45 h). Mean of three independently repeated experiments. (**P < 0.01, ***P < 0.001; Student’s t test.)
proliferation of cultured lymphocytes (5). Positive findings
in chemotherapy patients do not explain positive findings
after environmental and occupational exposure. There is
no doubt that chemotherapy may induce MN in the CBMN
assay and this has never been questioned. However, radiomimetic chemicals and cross-linking agents are rare among
environmental mutagens. The majority of mutagens in our
environment and at the workplace produce excision-repairable lesions, which are not readily detected by the standard
CBMN assay in human biomonitoring.
4. We agree that the CBMN assay is used to address different questions. Its use as an indicator of exposure is still
one of the main applications. Our discussion only refers to
the frequent use of the CBMN assay in human biomonitoring in an attempt to detect genotoxic effects in cultured
lymphocytes after occupational and environmental exposure to genotoxic chemicals. Only these studies and the
plausibility and reliability of their results are the subject of
our concern. The use of the CBMN assay after exposure to
ionizing radiation or as an indicator of genomic instability
and potential cancer risks (‘early effects’) is not considered
and should be discussed separately. In fact, Bonassi et al.
(6) provided ‘preliminary evidence that MN in peripheral
blood lymphocytes is predictive of cancer risk’. However,
in this study, occupational exposure to mutagens or smoking status did not significantly modify the relationship
between MN frequency and cancer risk.
It seems to be clear that MN are induced as a consequence
of unrepaired or misrepaired DNA damage and different
MN frequencies may occur because of differences in individual DNA repair capacities and/or exposure. However, to
measure significantly increased MN frequencies in human
biomonitoring of exposed populations, sufficient damage
has to be induced in vivo and has to persist in vitro to lead
to the formation of MN. Our results (1) question whether
372
these requirements are fulfilled after occupational and environmental exposure to chemical mutagens.
As discussed previously (7), the scoring of MN frequencies in both binucleated and mononucleated lymphocytes
in the standard CBMN assay may provide some useful
additional information but is not suited to differentiate
clearly between MN induced in vivo or in vitro. Most of
the published human biomonitoring studies report MN frequencies in BNC only and these effects are discussed in
our publication.
5. We agree that there are differences between exposures in
vivo and in vitro. However, we do not have sufficient reliable
data to assess these differences and their consequences for
the formation of MN. Circulating lymphocytes in G0 phase
are in fact the most relevant target cells for the CBMN assay.
If proliferating lymphocytes are actually exposed to genotoxins in vivo during S phase, they will go through mitosis
in vivo and produce MN if sufficiently damaged. These MN
may be present in lymphocytes at the start of the culture,
but will most likely not contribute to the MN frequency in
BNC formed in the presence of cytochalasin B 44 h later.
Our study with cancer patients exposed to chemotherapy
indicated that even under these extreme exposure conditions
only a few MN are produced in vivo (5). However, this is
still the only study addressing this question and more data
are needed to better understand what is actually happening with lymphocytes in vivo and how these effects can be
appropriately considered in the CBMN assay.
With regard to the hypothesis that cumulative DNA
damage acquired by lymphocytes in vivo during chronic
exposure conditions explains the positive CBMN biomonitoring studies, it may in fact be useful to perform
comparative experiments in vitro using acute versus
chronic exposure conditions. However, without experimental data to support this hypothesis, we do not believe
Response to Letter to the Editor re CBMA assay
80
70
0.01
60
2.07
50
4.14
40
6.20
tail m om ent
30
8.26
20
10
10.32
20µM MMS
Co
13.07
40µM MMS
0
80
70
0.02
60
1.42
50
2.81
40
4.21
tail m om ent
30
5.60
20
7.00
10
0
50µM SO
25µM SO
Co
8.39
Fig. 2. Examples for the distribution of damage across cells in the comet assay with blood cultures exposed to MMS (A) or SO (B) at the start of the cultures.
that it is reasonable to expect that chronic exposure conditions in vivo can lead to higher levels of DNA damage
in lymphocytes than what can be achieved by short-term
treatment in vitro of lymphocytes, where much higher
concentrations of genotoxic compounds can be used.
Consequently, without such data, the in vitro results are
an appropriate indicator of the sensitivity of the in vivo
assay.
373
G. Speit
6. Fenech and Kirsch-Volders are correct that a direct
comparison between genotoxic effects in the population
of white blood cells (investigated by the comet assay)
and stimulated T lymphocytes (investigated by the
CBMN assay) may be of limited value because of the
different cell populations investigated. However, there
is no reason to assume that T lymphocytes were not
adequately damaged in our study by the directly acting
mutagens tested. For example, the distribution of damage
across cells does not give any indication for a bimodal
distribution of damaged cells (i.e. an indication for an
undamaged subpopulation). Figure 2 shows two examples
for such a distribution (original data from our published
comet assay experiments) (1). In blood exposed to methyl
methanesulphonate (MMS) (Figure 2A), the percentage
of cells with tail moments >0.7 increases from 26% in
the control to 52% at 20 µM and then 96% at 40 µM. In
the case of styrene oxide (SO) (Figure 2B), the response
is even more definitive and all cells measured exhibited
increased DNA damage at both concentrations tested.
Therefore, the comet assay results appropriately reflect
the initial DNA damage induced in blood cells including
lymphocytes. We do not doubt that the transformation of
primary DNA damage into MN is influenced by various
factors including individual repair capacities. These
factors may also be highly relevant for spontaneously
occurring MN and make them an indicator for ‘early
effects’—but this is not the subject of our study.
7. Fenech and Kirsch-Volders assert that our comet assay data
may have been confounded by apoptotic and necrotic cell
death. However, it is generally known that the comet assay
is a reliable indicator of DNA damage and even high damage levels (large comets) mainly represent repairable damage and not apoptotic cells (8). We do not know whether
it has ever been shown that DNA-damaging chemicals (as
used in our study) induce apoptosis in the CBMN assay with
lymphocytes without reducing the nuclear division index
(NDI). However, the fact that the substances investigated
clearly increased MN frequencies in the standard protocol
of the in vitro CBMN assay (i.e. following the OECD protocol) provides enough confidence that apoptosis/necrosis
does not have a relevant influence on their MN-inducing
374
potential. We just used the comet assay to demonstrate the
induction of initial DNA damage by the mutagens tested.
In summary, we understand that Fenech and Kirsch-Volders are
concerned about the CBMN assay and want to defend the assay
against unjustified criticism. However, the data presented in our
publication and our interpretation of the results should lead to
the conclusion that there is need for further discussion about
the plausibility and reliability of many human biomonitoring
studies performed with the standard CBMN assay. It should be
a common goal to critically re-evaluate the performance of the
assay and the protocols used. Investigators, regardless of their
depth of experience, should critically assess the methods they
are using. In this respect, the concerns raised by Fenech and
Kirsch-Volders represent a useful starting place for a reasoned
discussion of the usefulness of the CBMN assay when used for
human biomonitoring studies to detect genotoxicity induced in
peripheral blood lymphocytes.‍
References
1.Speit, G., Linsenmeyer, R., Schütz, P. and Kuehner, S. (2012) Insensitivity
of the in vitro cytokinesis-block micronucleus assay with human lymphocytes for the detection of DNA damage present at the start of the cell culture. Mutagenesis, 27, 743–747.
2.OECD. (2010) Guideline for the Testing of Chemicals No. 487: In Vitro
Mammalian Cell Micronucleus Test (MNvit). OECD, Paris, France.
3.Fenech, M. and Neville, S. (1992) Conversion of excision-repairable
DNA lesions to micronuclei within one cell cycle in human lymphocytes.
Environ. Mol. Mutagen., 19, 27–36.
4.Leopardi, P., Zijno, A., Marcon, F. et al. (2003) Analysis of micronuclei
in peripheral blood lymphocytes of traffic wardens: effects of exposure,
metabolic genotypes, and inhibition of excision repair in vitro by ARA-C.
Environ Mol. Mutagen., 41, 126–130.
5.Arsoy, N. S., Neuss, S., Wessendorf, S., Bommer, M., Viardot, A., Schütz,
P. and Speit, G. (2009) Micronuclei in peripheral blood from patients
after cytostatic therapy mainly arise ex vivo from persistent damage.
Mutagenesis, 24, 351–357.
6.Bonassi, S., Znaor, A., Ceppi, M. et al. (2007) An increased micronucleus
frequency in peripheral blood lymphocytes predicts the risk of cancer in
humans. Carcinogenesis, 28, 625–631.
7.Speit, G., Zeller, J. and Neuss, S. (2011) The in vivo or ex vivo origin of
micronuclei measured in human biomonitoring studies. Mutagenesis, 26,
107–110.
8.Collins, A. R., Oscoz, A. A., Brunborg, G., Gaivão, I., Giovannelli, L.,
Kruszewski, M., Smith, C. C. and Stetina, R. (2008) The comet assay: topical issues. Mutagenesis, 23, 143–151.