You are on page 1of 11

Photochemistry and Photobiology, 2002, 75(5): 437–447

Review

Photogenotoxicity of Mammalian Cells: A Review of the Different Assays


for In Vitro Testing¶
Jean-Roch Meunier*1, Alain Sarasin2 and Laurent Marrot1
L’Oréal Advanced Research, Life Sciences Research, Aulnay-sous-Bois, France and
1

CNRS, UPR-2169, Laboratory of Genetic Instability and Cancer, Villejuif, France


2

Received 14 December 2001; accepted 22 January 2002

ABSTRACT the molecular level have provided new insight in the field
of DNA lesion characterization, DNA repair, gene expres-
During the past several years, phototoxicity has been
sion and cell cycle control.
studied at the molecular level, and these studies have pro-
vided new insights in the field of DNA lesion character- Apart from the direct effect of UV radiation, the expo-
ization, DNA repair and cell response to ultraviolet (UV)- sures to UV light in the presence of chemical compounds
induced stress. The development of new antibiotics and found in drugs or cosmetic ingredients can reveal an inter-
antiinflammatory drugs has highlighted the necessity to action between light absorbency and chemicals through en-
develop the assessment of phototoxicity in the safety eval- ergy transfer or chemical modifications. These interactions
uation of new chemical compounds. This paper aims at can produce deleterious, cytotoxic or genotoxic effects.
reviewing the known molecular mechanisms of the cel- Thus, some chemical compounds are well-known photosen-
lular response to UV-induced stress, the in vitro methods sitizers: they can be of endogenous (such as porphyrin and
that can be proposed and used to screen for toxicity of flavonoids) or of exogenous origins (fluoroquinolone class
sunlight and the photosensitization process resulting of antimicrobial agents and nonsteroidal antiinflammatory
from the activation of drugs by light. UV sources, bio- drugs) (2).
logical systems and endpoints of interest in that partic- This has highlighted the necessity to harmonize and eval-
ular objective are listed. Phototoxic effects span from the uate the strategy used to assess the phototoxic and, in par-
cytotoxic–apoptotic effect to the induction of primary ticular, the photogenotoxic potential of a chemical com-
DNA damage, DNA repair and a variety of stress genes pound. Up to now, photomutagenicity testing is only re-
acting on the cell cycle and the fate of the cell. Ultimately, quired for sunscreen development.
it can lead to the induction of hereditary DNA modifi- For this purpose, different symposia, meetings and work-
cation. A variety of assays are proposed to specifically shops have been organized since 1993, by the European
address all these particular consequences of UV-induced Center for the Validation of Alternative Methods in 1993 (3)
toxicity. and 1999 (4) in Europe, in 1996 by the Food and Drug
Administration (5) and in 1999 (International Workshop of
Genotoxicity Test Procedures, IWGTP, Washington) (6), to
INTRODUCTION address this question and evaluate the existing in vitro meth-
All living organisms, including humans, are exposed to the odology. No regulatory requirement has been specified, but
sun or to manufactured equipment that emit light. Ultraviolet a set of recommendations and guidances have been issued
(UV) light is a well-established mutagenic and carcinogenic (4–6). Until now, the photocytotoxicity test 3T3 neutral red
agent (1), and in the recent years, phototoxicity studies at uptake (NRU) is the only in vitro method that has been val-
idated so far by European regulatory affairs (7–9). Because
¶Posted on the website on 28 Jauary 2002. this test is devoted to the evaluation of cytotoxic effect, it
*To whom correspondence should be addressed at: L’Oréal Ad- represents a very preliminary step, and there is still a neces-
vanced Research, Life Sciences Research, Direction of Biological sity for the development of mechanistically based methods.
Assessment and Safety, Genotoxicity Group, 1, av. E. Schueller, With this background, the fundamental research done for
F-93601 Aulnay-sous-Bois cedex, France. Fax: 33-1-4868-9180;
e-mail: jrmeunier@recherche.loreal.com years by numerous teams to understand the molecular mech-
Abbreviations: BER, base excision repair; CPD, cyclobutane pyrim- anisms underlying the cellular response to UV-induced ge-
idine dimers; NER, nucleotide excision repair; PARP, poly-ADP- notoxic stress permitted one to set up various in vitro tests
ribosyl polymerase; PCR, polymerase chain reaction; 6-4 PP, py- for the assessment of (1) photoprotection strategies; and (2)
rimidine (6-4) pyrimidones; ROS, reactive oxygen species; SCGE,
single-cell gel electrophoresis; SSL, solar-simulated light; UDS, ingredient phototoxicity.
unscheduled DNA synthesis; UV, ultraviolet. Obviously, the best cell system for in vitro photobiology
q 2002 American Society for Photobiology 0031-8655/01 $5.0010.00 studies is the target cells of the skin, mostly normal human

437
438 Jean-Roch Meunier et al.

scientists involved in developing new assays or in testing


phototoxicity effects of various xenobiotics (Table 1).

PHOTOINDUCED CYTOTOXICITY AND


APOPTOSIS
Cytotoxicity
The photocytotoxic effect of UV light should be distin-
guished from its photogenotoxic effect. Photocytotoxicity is
the measurement of acute effect of UV stress, resulting in
life or death of a cell. Photocytotoxicity can be the result of
either necrosis or apoptosis induction. There are some spe-
cific methodologies to unequivocally demonstrate the occur-
rence of an apoptotic response, which are listed hereafter.
In contrast, the aims of photogenotoxicity studies are to
Figure 1. Biological endpoints for phototoxicity testing and mea- evaluate the ‘‘genetic health’’ of the surviving cells. It has
suring photogenotoxic treatment on mammalian cells.
to be understood that photogenotoxicity studies should be
performed in the frame of a very well-defined cytotoxic ef-
fect of the UV stress employed. Thus, an accurate evaluation
keratinocytes or eventually normal human melanocytes (10). of the photocytotoxic effect of the treatment (UV light alone
Normal skin irradiated ex vivo or reconstructed skin systems or photoactivation studies of chemical compounds) is an ab-
should be even better, but more complicated, to manipulate solute prerequisite for photogenotoxicity studies. To be bi-
on a large scale. They certainly enable topical treatment in ologically relevant, the genotoxicity endpoints evaluated
conditions closer to those in real life; moreover, they do through a particular methodology should only be observed
constitute an alternative to animal testing for chemicals that in cells treated by conditions leading to no or moderate tox-
would be applied on the skin. But the biological endpoints, icity (survival $ 50% of the untreated controls).
which can be evaluated on these reconstructed organs, are There are many different methods to assess cytotoxicity.
mostly cytotoxicity, cytokine release at the population level The neutral red uptake method is currently used as a pho-
and in situ immunochemical detection of various antigens totoxicity measurement and has been validated by European
(11). There is still a need for developing investigative meth- regulatory agencies as an in vitro test for phototoxicity (7,8).
ods on these models at the single-cell level. For technical This method allows an evaluation of the photoinduced cy-
easiness, most experiments have been, therefore, carried out totoxicity and thus should be regarded only as a primary
using fibroblasts or transformed cells. step.
A variety of light sources have been used in the literature Cytotoxicity can also be measured by the counting of cells
of photobiology. This diversity in the used UV sources and for several days after treatment, by assessment of the mito-
in the dosimetry measurements renders the comparison of chondrial activity (3-[4,5-dimethylthiazol-2-yl]-2,5-diphen-
results between laboratories difficult. Thus, one of the goals yltetrazolium bromide test), by the determination of the mi-
of the workshops and meetings cited earlier was also to har- totic index (as a measurement of the dividing capacity of
monize the UV sources and to issue a recommendation on the cell such as that used in the chromosomal aberration
this specific point. test), by cloning efficiencies and so on.
Most of the studies were done with UVC light, which is There is, up to now, no consensus on the choice of a
effectively screened by the ozone layer and does not reach preferred method for cytotoxicity measurement. Obviously,
the earth’s surface but is easy to use in laboratory environ- the criterion of a reduction of 50% of cell survival would
ment. Studies done with UVB or UVA tubes demonstrated not biologically signify the same toxicity in terms of UV
the specificity of the cellular response with regard to these dose and concentration of compounds, depending on the
particular wavelengths. In an attempt to mimic the solar UV method used. This explains why usually at least two differ-
spectrum, solar simulators are used, enabling the assessment ent cytotoxicity endpoints must be used in photogenotoxicity
of synergistic effects of UVB and UVA on cell systems. In studies (such as cell count and mitotic index for the photo-
all the recommendations and guidances issued so far, the chromosomal aberration test, for example).
solar simulator is the recommended UV source. Moreover, Apoptosis is an important component of the cytotoxic ef-
a detailed knowledge of the exact spectrum of the used fect of a treatment, and the occurrence of apoptosis is a
lamps is necessary to allow comparison of biological effects valuable information in understanding the mechanism of
and an extrapolation from in vitro testing assays with the in phototoxicity of a particular compound because it involves
vivo ‘‘real life.’’ Because these apparatus are usually very specific pathways and cellular responses.
expensive, the minimal requirement is for information on the
spectral output by determining the UVB–UVA ratio of the
Apoptosis
source used.
This review details some of the consequences of UV ir- Background and mechanisms. Cells respond to toxic stimuli
radiation of living cells in order to describe most biological in one of two ways: by activation of the cellular stress re-
assays used to screen for photogenotoxicity of light, drugs sponse or by undergoing cell death. Apoptosis is a form of
or a combination of both (Fig. 1). It should be useful for cellular suicide, with distinct morphological features, which
Photochemistry and Photobiology, 2002, 75(5) 439

Table 1. In vitro screening assays for photogenotoxicity. State of the various described photogenotoxicity assays in terms of whether they
are used in routine screening or still in research and development stage (R&D)

Adaptability for routine screeing*† Assays


still in
Biological endpoints and assays Unknown lesions Known lesions Regulatory position R&D state

I. Cytotoxicity
Cell toxicity R1/E1 R1/E1 3T3 NRU validated (7)
Apoptosis R12/E12 R12/E12
II. Primary DNA damage
Immunodetection R2/E12 R+/E+2 Mechanistic approaches: sensitivity 1
Comet assay R1/E1 R1/E1 to be determined with a set
Comet assay + repair enzymes R12/E12 R12/E12 of chemical compounds
UDS R2/E2 R2/E2 of the same class
III. Cell response
Stress genes R12/E12 R12/E12 1
Cell cycle R2/E1 R2/E1 1
IV. Hereditary consequences
Point mutagenesis R1/E2 R1/E2 1/2‡
Chromosomal damage R1/E12 R1/E12 1/2‡

*R 5 reliability and E 5 easiness.


†R1: reliable; R12: reliability depending on the method used and damage targeted (immunodetection); R2: questionable reliability; E1:
easy to perform; E12: easiness depending on the assay used and damage targeted (immunodetection); E2: not so easy or time consuming
or needs expert skills.
‡1/2: Depending on the assay.

occurs under both physiological and pathological circum- system seems to be involved in the efficiency of the UV
stances (12). These features include membrane blebbing, cy- phototherapy (19), but UVA alone does not induce apoptosis
toplasmic shrinkage, nuclear condensation, nuclear fragmen- at relevant physiological doses. UVB and UVC induce ap-
tation, destruction of apoptotic cells into apoptotic bodies optosis mostly by a p53-dependent pathway. Indeed, the lev-
and, finally, phagocytosis of these apoptotic bodies. Apopto- el of apoptosis is proportional to the blockage of transcrip-
sis is associated with activation of caspase proteases that tion by the presence of unrepaired DNA lesions in tran-
target over 100 cellular substrates. Recent developments scribed strands, as demonstrated by the high level of apo-
have identified the mitochondrion as central to their activa- ptosis induced in DNA repair–deficient cells even by a low
tion (13). During apoptosis, cytochrome c is released from dose of UV (20). Depending upon the cell types, the apo-
the mitochondria into the cytosol, where it interacts with the ptosis is more or less easy to be recorded. The choice of the
cytosolic protein Apaf-1, leading to sequential activation of cell system is very important in this test because the level
pro-caspases-9 and -3 (14–16). In fact, mitochondrial chang- of spontaneous or induced apoptosis is very variable accord-
es are one of the earliest processes in the apoptotic cascade. ing to the cell type. Fibroblasts are very difficult to apoptose,
One of the fundamental features of this type of cell death whereas blood cells or transformed cells exhibit a high level
is the recognition and phagocytosis of apoptotic cells by of apoptosis. Similarly, cells isolated from nucleotide exci-
macrophages or neighboring cells. However, the mechanism sion repair (NER)–deficient patients are about five times
of plasma membrane alterations that promote the recognition more sensitive to the UV-induced apoptosis at low UV dos-
of apoptotic cells under in vivo conditions is poorly studied. es. Tumoral HeLa cells are now being widely used in several
The loss of plasma membrane phospholipid asymmetry, be- tests to screen for apoptosis induction after various treat-
cause of the externalization of phosphatidylserine, on the cell ments.
surface is one of the most well-reported recognition signals Assays for apoptosis detection. Numerous assays have
for phagocytic uptake (17). This process allowed the devel- been proposed for the characterization and quantification of
opment of a specific assay for the quantification of apoptosis. the apoptosis process (Fig. 2). The easiest assay is the mor-
Photodermatologists are well aware of the presence of phological study of the cells labeled by various types of
sunburn cells in the epidermis exposed to sun. These cells fluorescent probes. The early apoptosis process can be rec-
express the classical morphology and behavior of apoptotic ognized by microscopy, showing distinct peripheral chro-
cells. It was quickly demonstrated that UVC and UVB in- mosome condensation against the intact nuclear membrane.
duce apoptosis, which is considered as the best way to elim- Then, chromatin fragmentation and cell blebbing, with a still
inate UV-damaged keratinocytes and therefore to protect the intact cytoplasmic membrane, occur. Fluorescent markers,
exposed skin from cancer (18). Although UVC irradiation such as the Hoechst 33342, allow one to follow living cells
has been shown to induce apoptosis via expression of the and to detect apoptotic cells by fluorescent microscopy. This
Fas ligand and ligation of its receptor, Fas, it is believed that assay can be standardized and automatized for a rapid
p53 stabilization plays a major role in this induction. The screening. The number of studied cells is, however, relative-
apoptosis of human T-helper cells mediated through the Fas ly limited, making this assay not very quantitative. The sec-
440 Jean-Roch Meunier et al.

idues that are externalized and can be quantified by flow


cytometry and thus can be used in large-scale screening as-
says, but its specificity also varies according to the cell
types.
Finally, protein modifications can also be followed during
the apoptotic pathway. Specific cleavages of the poly-ADP-
ribosyl polymerase (PARP) or of numerous caspases occur
during apoptosis. PARP cleavage has been described as a
prerequisite for the DNA fragmentation process in the apo-
ptotic cell (21) and can be followed up by the Western blot
analysis (22). These death proteases are part of a proteolytic
cascade activating proenzymes into active proteases. Serial
activation of these various caspases can be followed in cell
extracts to quantify the level of apoptosis, and several
ELISA-based immunoassays have been developed, enabling
a rapid and quantitative detection of caspase activation. Cas-
pase 3 activation is a particularly interesting endpoint be-
cause of its control role as caspase effector responsible for
the execution of the cell death program in the cell (23,24)
(Fig. 2). Several colorimetric immunoassays are now avail-
able to quantitate the activity of caspase 3 in cell lysate (25).
These methods are usually highly reliable and can be used
in large-scale screening assays (high throughput) for apopto-
sis determination.
Research conducted in the last years in the field of apo-
Figure 2. In vitro assays for quantifying the apoptosis level in pho- ptosis has lead to the identification of different molecular
totoxicity testing.
events, organized in a sequential way, committing the cell
in the apoptotic pathway. These molecular endpoints have
ond most popular determination of the apoptosis level is the been translated into methods enabling a rapid and quantita-
quantification of the number of cells that appeared in the tive measurement of the apoptosis in a cell population (Fig.
sub-G1 peak of the cell cycle profile. After labeling with 2). But in order to assess a reliable and specific apoptotic
ethidium bromide, suspended cells are analyzed through effect of the UV light or of the combination of UV light
flow cytometry. Apoptotic cells appear in the window be- with a photoreactive chemical compound, one should con-
tween cell debris and the classical G1 phase. This assay is firm the results by using at least two different endpoints, for
easy to be carried out, even in large-scale screening protocol, example, an early one (such as the annexin V externaliza-
and is considered one of the most specific. tion, or the caspase 3 activation) and a late one (such as the
DNA fragmentation can be recorded by many ways: gel sub-G1 determination).
electrophoresis, loss of labeled genomic DNA, in situ label-
ing of DNA nicks (by the TUNEL assay) and comet assay. QUANTIFICATION OF GENOTOXICITY
These assays can be used for quantifying apoptotic efficien-
Photolesions
cy, but they have to be correlated with other assays because
they mainly measure the production of single-stranded Bulky lesions. UVB and UVC irradiations generate, in cel-
breaks in cell DNA, which are observed not only during the lular DNA, bipyrimidic photoproducts, which are mainly
apoptotic response. Thus, the TUNEL assay could also react cis–syn cyclobutane pyrimidine dimers (CPD) and pyrimi-
to the induction of strand-breaks by the DNA repair ma- dine (6-4) pyrimidones (6-4 PP) at a ratio varying from 4:1
chinery. Moreover, if the well-known ‘‘DNA ladder’’ pic- to 10:1, according to various studies and at various wave-
ture of apoptosis can easily be observed in circulating cells lengths (26). The 6-4 PP have been found to undergo an
(lymphocytes and derived–cell lines), it is usually less ob- efficient photoconversion into Dewar valence isomers upon
vious in other cell types such as skin keratinocytes or dermal exposure to UVB light. Other types of DNA damage, such
fibroblasts. Some polymerase chain reaction (PCR)–based as strand breaks or single-base modifications, have been de-
methods have been developed in an attempt to increase the tected but at high nonphysiological doses (26). UVA irra-
sensitivity of detection of the DNA ladder formation (Clon- diation can hardly produce direct photolesion on the DNA,
tech, Palo Alto, CA), but the DNA ladder endpoint can be especially at the longest wavelengths, where photons are
considered to be highly cell-type–dependent and thus some- poorly or not absorbed. Even if the induction of CPD by
times quite insensitive. Modifications of cell membranes as UVA has been reported (27,28), this question is still a matter
a result of activation of phospholipid-scramblase or inhibi- of debate. Moreover, some photoactivable compounds can
tion of aminophospholipid-translocase (or both) cause the also produce bulky adducts. For example, psoralen and pso-
externalization of phosphatidylserine during apoptosis. This ralen derivatives can intercalate into DNA and upon UVA
process is thought to be a general marker of apoptosis. This irradiation give rise to monoadducts or cross-links between
event can be followed by the detection of annexin-V–fluo- two thymine residues (29). Finally, reactive compounds pro-
rescein isothiocyanate binding to the phosphatidylserine res- duced during lipid peroxidation could also reach nuclear
Photochemistry and Photobiology, 2002, 75(5) 441

DNA and form adducts (30). However, the significance of sions usually by a DNA glycosylase activity giving rise to
such damage for photomutagenesis in vivo needs to be fur- nicks on purified DNA. The DNA substrates can be a plas-
ther clarified. mid genome (usually treated in vitro) or cellular DNA ex-
Oxidized bases. UVA, and even visible light, can produce tracted from treated cells.
oxidative DNA damage via the production of reactive oxy- Specific enzymes. The phage T4 endonuclease V (or the
gen species (ROS). Oxidation of DNA by ROS produces a one from Micrococcus luteus) specifically recognizes the
large variety of lesions that have been extensively studied in CPD, whereas the Escherichia coli Fpg, NTh or Nfo proteins
(oligo)nucleotides (31). In the case of the oxidative stress cut DNA at oxidized purines, oxidized pyrimidines or abasic
induced by UV from sunlight, mainly 8-oxoguanine as well sites, respectively (for a review, see Sankar and Sankar
as oxidized pyrimidines such as thymine glycol or uracil- [40]). After treatment of damaged DNA with these enzymes,
derivatives have been described in cellular systems (32–35). the measurement of the DNA size between two nicks allows
Strand breaks. Endogenous photosensitizers (such as mel- the quantification of DNA lesions per kilobasepair of DNA.
anin, porphyrin, riboflavin) and exogenous drugs (antibiot- For example, using the T4-endonuclease V enzyme on UV-
ics, antiinflammatory drugs. . . ) produce ROS when exposed irradiated plasmid DNA, Kuluncsics et al. (28) determined
to UVA. Some of these ROS, such as hydroxy-radicals or the yield of CPD formation by UVC (1.5 3 1022 CPD/[kbp
singlet oxygen, induce DNA single-strand breaks and alkali- J m2]), UVB (1.5 3 1024 CPD/[kbp J m2]), UVA (1.4 3
labile sites (36). These DNA damages, albeit certainly effi- 1027 CPD/[kbp J m2]) and solar-simulated light (SSL) (1.5
ciently repaired, could be a source of genetic instability. On 3 1026 CPD/[kbp J m2]). With UVA and SSL irradiations,
the other hand, double-strand breaks, which are known to oxidized pyrimidines and purines as well as abasic sites were
trigger a high rate of recombination, are only induced by detected at similar levels as CPD. It is important to note that
strong photosensitizers and have never been clearly detected the absolute values concerning the production of DNA le-
with UV from sunlight alone. sions after various wavelength irradiations vary a lot from
one laboratory to another and among the various biological
DNA Repair systems used. Other than the method using the bacterial Fpg
protein, 8-oxoguanine can also be measured by high-perfor-
Nucleotide excision repair. Bulky DNA adducts, including
mance liquid chromatography (HPLC) of DNase-treated
UV-induced photoproducts, are recognized by the versatile
DNA, followed by electrochemical detection (41). Such a
nucleotide excision repair (NER) pathway able to detect
measurement can improve the detection of ROS effects in
DNA structure abnormalities and to repair them. During the
treated biological samples. This assay is still difficult to be
course of repair, a 26–30 bases–long oligonucleotide con-
performed on a large scale, and measurements of 8-oxo-
taining the lesion is removed, and the gap is then filled in
guanine are still controversial and subject to artifacts. A
by unscheduled DNA synthesis (UDS) carried out by DNA
more quantitative and specific determination of DNA lesion
polymerase d or e in the presence of PCNA and various
can be performed using HPLC followed by electrospray ion-
replication factors. Bulky DNA lesions inhibit RNA poly-
ization tandem mass spectrometry (LC-MS/MS) of purified
merase II giving rise to a signal that results in a preferential
irradiated DNA (42). This allows one to measure all classes
repair of lesions located on transcribed strands. This tran-
of UV-induced photoproducts after UVC or UVB irradia-
scription-coupled repair is vital for cell survival after UV
tion. However, this technique is used by sophisticated lab-
irradiation. Its impairment gives rise to serious DNA-repair–
oratories, and we do not yet know if this could be applied
deficient diseases associating cancer development and neu-
in the near future for screening assays.
rological disorders (37).
In terms of photoscreening, plasmid DNA or genomic
Base excision repair. Modified bases (such as oxidized
DNA can be used to quantify the protection level on known
pyrimidines or purines) or abasic sites are repaired by base
DNA lesions (such as CPD, for example) or to detect new
excision repair (BER). Basically, the damaged base is rec-
classes of DNA damage produced by given photoactivable
ognized and removed by a specific DNA glycosylase, giving
chemicals. This can be realized by using the Fpg, Nfo or
rise to an abasic site that is further recognized and eliminated
Nth enzymes, which recognize numerous classes of DNA
by several classes of Ap-endonucleases. The repair of 8-
damage, whereas the T4-endonuclease V can only be utilized
oxoguanine was particularly well studied, even in human,
for quantifying CPD.
where the specific repair enzyme hOGG was described (38).
The E. coli NER system can be reproduced in vitro using
The small gap is then filled in by either the b polymerase
the UvrA, UvrB and UvrC purified proteins, which are able
(small patch BER) or by the same polymerases as in the
to recognize bulky DNA lesions and to produce single-strand
NER (long patch BER). Very recently, a transcription-cou-
gaps in damaged DNA. This assay has been successfully
pled repair for oxidized bases such as thymine glycols and
used to study the respective repair kinetics of CPD and 6-4
8-oxoguanine has been shown (39).
DNA lesions after UV irradiation (43). However, the avail-
BER and NER pathways are error-free processes, in con-
able enzymes of commercial preparation are relatively dif-
trast to the translesion synthesis carried out by mutagenic
ficult to use in a quantitative way.
DNA polymerases.
Antibodies. A popular test involves the use of polyclonal
or monoclonal antibodies able to recognize different types
Quantification
of bulky lesions such as UV-induced photoproducts, chem-
To characterize and quantify the lesions produced by various ical carcinogen–induced lesions and some oxidized bases
UV lamps, one can use purified DNA repair enzymes, which (44). The use of antibodies against CPD or 6-4 PP (45,46)
specifically recognize DNA lesions or classes of DNA le- is possible in a screening assay on genomic DNA or for in
442 Jean-Roch Meunier et al.

situ immunohistochemistry to follow the presence of lesions (10). The comet assay enables the evaluation of photopro-
on irradiated native skin or reconstructed skin in vitro (47). tective effects of sunscreen or antioxidants (56,57), or the
This assay can be ‘‘roughly quantified,’’ although its cali- photogenotoxic potential of chemical compounds when ir-
bration in terms of number of lesions per unit of DNA is radiated with UV light, such as the antibacterial fluoroquin-
still difficult. It is, however, a relatively easy technique for olone Lomefloxacin (58). This can be performed on every
measuring a protection index by a given molecule on UV- cell type, and thus is a particularly interesting endpoint in a
irradiated skin. Antibodies against 8-oxoguanine, as a mea- strategy of in vitro photogenotoxicity testing.
sure of oxidized purines, do exist, but their specificity and Unscheduled DNA synthesis. Quantification of NER by
possible use on tissue slides are still not reliable enough. UDS, i.e. by incorporation of labeled DNA precursors during
Comet assay. Sensitive detection of DNA damage is the resynthesis, corresponds to an indirect measure of the pres-
goal of a variety of techniques used in environmental toxi- ence of DNA lesions in an NER-proficient cell. This ap-
cology. Because the effects of environmental genotoxics, proach has been used to study the repair of photolesions such
such as UV light, are often tissue- and cell-type specific, it as pyrimidine dimers (59). The level of UDS is proportional
is of importance to rely on techniques that can detect DNA to the UV dose until a plateau indicating a steady-state level
damage in the individual target cell. of repair is reached (which is around 15 J/m2 for UVC and
The comet assay, also called single-cell gel electropho- 500 J/m2 for UVB). This assay allows one to measure repair,
resis (SCGE), is a powerful method devoted to the assess- and therefore the amount of damage even for unknown
ment of primary DNA damage (but not their biological ef- bulky adducts. It does, however, not easily detect DNA le-
fects such as mutations or clastogenesis). Although other sions repaired by BER, such as oxidized bases because in
techniques, such as immunological detection of strand this case, the patch size is too small to give rise to detectable
breaks (48) or the alkaline elution, exist for detecting DNA labeling. UDS analysis can be used in automatic and large-
damage, the comet assay presents the advantage of being scale assays using labeled or fluorescent detection probes.
relatively easy, reliable and fast enough to perform. Positive results with UDS indicate the presence of DNA le-
Analysis of DNA damage by SCGE was first described in sions recognized and repaired by NER and therefore allow
1984 by Ostling and Johanson (49). They demonstrated that the quantification of initial genotoxicity. This does not, how-
DNA of cells embedded in a low-melting agarose microgel ever, permit one to determine if the repair step is complete
on microscope slides migrates in an electrophoretic field in and faithful. For a rapid and early test in a screening pro-
a pattern determined by the extent of DNA damage. Because cedure, UDS is one of the most reliable in vitro assays, but
the migrating DNA resembles comet tails, the assay has been in some cases (particularly with chemical compounds) it can
termed as comet assay. The technique first described with be relatively insensitive, especially for DNA damage re-
the use of a neutral electrophoresis buffer that allows only paired through BER mechanisms (small patch of nucleotide
the detection of double-strand breaks has then evolved (50) removed).
with the use of alkaline electrophoresis conditions that per- Reporter gene. Bulky DNA adducts as well as some ox-
mit the detection of single-strand DNA breaks, alkali-labile idized bases strongly inhibit RNA transcription because of a
sites as well as double-strand breaks. The comet assay pro- blockage of the RNA polymerase II elongation. Plasmid
tocols have been discussed and harmonized during the In- DNA expressing a reporter gene (chloramphenicol acetyl
ternational Workshop on Genotoxicity Test Procedures in transferase, luciferase or Green Fluorescent Protein) can be
1999 (for a review, see Tice et al. [51]). used as a genetic marker of DNA repair after transfection in
Thanks to image analysis, the comet assay allows quan- mammalian cells. Damaged plasmids will not express the
titative measurement of DNA strand-breaks induction, either reporter gene because of the inhibition of transcription, but
caused by the genotoxin itself or by the activity of DNA during the course of repair, the gene will become expressed
repair enzymes removing the damage. Several software again as a result of an efficient repair. This assay is relatively
packages are available in the market. They enable the cal- easy to be performed in a medium-scale range, and gene
culation of the tail moment, a measure of the extent of dam- expression, i.e. repair efficiency, can be quantified by auto-
age, taking into account the length of the comet tail and the matic procedures. However, this assay necessitates treating
amount of DNA found both in the tail and in the nucleus the plasmid in vitro as naked DNA, which exclude all bio-
(52). Statistical analysis, accurate cytotoxicity data and ki- logical reactions produced inside the cells. Damaged plas-
netic studies are necessary to interpret the biological impact mids need to be transfected in the host cell, this necessitating
of the observed comet results (51). the use of a control undamaged plasmid as a marker for
This assay can be combined with the use of specific an- DNA transfection efficiency. Most experiments on a large-
tibodies to detect the nature of the damage, its localization scale basis have been done with human lymphocytes, but in
and the number of specific DNA damage such as pyrimidine theory all mammalian cells can be used for this assay, de-
dimers (53) or with DNA repair enzymes such as endonu- pending upon their transfection efficiency (for a review, see
clease III or Fpg enzymes (54). Benhamou and Sarasin [60]).
The comet assay is widely used today both in in vivo and
in vitro studies for the evaluation of the genotoxic potential GLOBAL CELLULAR RESPONSE TO THE UV
of a variety of DNA-damaging agents such as chemical com- STRESS
pounds, ionizing radiation and UV radiation. In the field of
Gene expression
phototoxicology, the comet assay has been used to demon-
strate the genotoxic potential of UVC, UVB or UVA light UV irradiation of mammalian cells produces damage at the
(55) as well as SSL on normal human melanocytes in culture level of proteins, DNA and lipids, according to the wave-
Photochemistry and Photobiology, 2002, 75(5) 443

length utilized, but many of its deleterious effects have gene the same position in tumors isolated from different individ-
expression as endpoints. uals. This type of hot spot can be used as a marker for
Among the major responses to UV irradiation, the p53 screening the deleterious effects of UV exposure and UV
pathway is probably the most documented one (Fig. 1). The mutagenesis. Because the largest mutation database has been
p53 protein accumulates in a dose-dependent manner in constructed with the p53 gene, we used this gene to screen
mammalian cells after exposure to various genotoxic agents for hot spots demonstrating UV exposure. This assay is
(61). In the first hours after treatment, p53 is stabilized by a much easier with the UV-induced mutations because one can
posttranslational process (probably linked to phosphoryla- look for the presence of the double mutations CC to TT,
tion, acetylation or both) independent of its transcription lev- which renders the assay easier and more specific. A test us-
el (62). This then leads to the transactivation of p53-respon- ing the mutant allele–specific PCR has been developed for
sive genes such as those involved in cell cycle regulation, looking at the appearance of CC-to-TT mutations at the co-
apoptosis, DNA repair or replication. A relatively easy meth- don 247/248 hot spot of the p53 gene, with a sensitivity as
od to detect genotoxic stress consists in the analysis of p53 low as 1026. By using this assay, we could show that normal
stabilization and accumulation by Western blotting or im- skin biopsy from an exposed body site could easily be dif-
munohistochemistry techniques using one of the numerous ferentiated from unexposed skin from the same individual
commercially available antibodies. This can be performed (68). Usual sun exposure induced detectable CC-to-TT tan-
easily on cell extracts or in vitro cultured cells, as well as dem mutation in a normal skin biopsy. This assay, when
on skin biopsies (63). A comparable type of assay can be done in a very controlled manner, is relatively easy to be
performed by following the kinetics of some expressed p53- carried out, and it is possible to use it as a screening test for
responsive genes that are triggered upon p53 activation; P21, antimutagenic protocols from either in vivo or ex vivo skin
Gadd45, PCNA and Bax genes are classically used as pos- biopsies, or from reconstructed skin. It is obvious that the
itive markers of UV damage either at the mRNA level or mutation induction represents the first step in the carcino-
with specific antibodies directed toward the proteins. UVC genic process, and therefore monitoring the presence of mu-
and UVB induce this p53 pathway, whereas the impact of tations can be considered as a direct way to measure this
UVA is less clear and seems to depend on cell type. initial process.
A major consequence of p53 stabilization is a cell cycle The mutagenicity of the UVA wavelength has not been
arrest via the induction of the p21 pathway, to allow time studied as well as for UVB and UVC. Episomal shuttle vec-
for DNA repair before replication. The cell cycle arrest oc- tors stably replicating inside the cell nucleus have been used
curs mostly at the G1/S checkpoint after UV irradiation, al- in order that the total response of the cells such as production
though some G2 arrest has also been reported. The mea- of free radicals is obtained after UVA treatment. UVA-in-
surement of G1-arrested cells immediately after treatment by duced mutation spectra show an absence of the tandem CC-
a genotoxic agent can be considered as a qualitative analysis to-TT mutations, a majority of C-to-T transitions, but not
of the consequences of DNA damage induction, and the ki- always located at bipyrimidic sequences, and a high level of
netics of normal cell cycle recovery can be taken as a quan- mutations at A–T base pairs compared with the UVC- and
titative analysis of DNA repair. This kind of analysis can be UVB–induced mutations (69,70). Although this high level
performed automatically by monitoring the number of cells of mutations at A–T base pairs could be considered as some
in each cell cycle compartment. The choice of the utilized type of molecular fingerprint of UVA, the lack of obvious
cells is critical for this type of assay because one should hot spots, with the small numbers of studied cases, does not
choose cells that are very active in cycling. allow to propose a screening assay specific for UVA muta-
genesis yet.
Photoinduced mutagenesis On the other hand, UVA-induced 8-methoxypsoralen mu-
tagenesis has also been studied in the framework of psora-
The production of bipyrimidic lesions by UVC or UVB ir- len1UVA therapy (71).
radiation induces a spectrum of mutations characteristic of The point mutagenesis assay on the V79 cell, called the
these wavelengths. Results of experiments with cells from hgprt assay, has been adapted in a photo-hgprt assay and
different organisms, ranging from bacteria to human, clearly developed in some laboratories to perform photomutagen-
show that UVC and UVB radiation give rise mainly to base icity screening of compounds (72).
pair–substitutions, targeted at pyrimidine–pyrimidine se-
quences, and mostly to GC-to-AT transitions (64,65). A high Photoclastogenicity
level of tandem CC-to-TT transitions are found only in UV-
irradiated samples (around 10–20% in DNA repair–profi- As a result of DNA repair failure, a primary DNA damage
cient systems), whereas this number is less than 1023 in mu- induced by UV light may be converted into a hereditary
tation spectra produced by genotoxic agents different from damage after cellular DNA replication, which can be a point
UV. Interestingly enough, DNA repair–deficient cells (es- mutation, a deletion or a chromosomal damage. The impact
sentially xeroderma pigmentosum cells) exhibit a very high of UV-induced point mutagenesis is well documented, as
level of CC-to-TT tandem mutations (about 60–70%) for still described previously. These mutations represent a major
not fully understood reasons. This result is found in tumor event in nonmelanoma skin cancers but are still a matter of
suppressor genes isolated from xeroderma pigmentosum debate in their role in the development of the very aggressive
(XP) skin cancers, such as p53, PTCH and p16 (65–67). melanoma, at least for the p53 gene. Point mutagenesis
Mutation spectra of these genes are characterized by muta- should, however, not hinder the importance of other damage
tional hot spots where the same type of mutation is found at and in particular chromosomal damage. These can be divid-
444 Jean-Roch Meunier et al.

ed into clastogenic events (resulting in the breakage of the the human reconstructed skin models. These characteristics
chromosome) and aneugenic events (resulting in a complete enable the development of in vitro tests on human skin cells
loss of a chromosome). Such events are sources of genetic or on human reconstructed skin models, which can be con-
instability, which could ultimately lead to the development sidered as predictive for the photoprotective or phototoxic
of cancer. Data generated so far with known photosensitizing effect of chemical compounds in vivo. Thus, the develop-
compounds show a good correlation between in vitro mam- ment of in vitro methodology as an alternative to animal
malian photoclastogenic and photocarcinogenesis potential testing is particularly advanced and relevant in the field of
in animals (73). phototoxicology.
In general, chromosomal damage is assessed through the Photobiological research conducted in numerous labora-
chromosomal aberration test (74) or through the in vitro mi- tories worldwide has permitted researchers to unravel, at the
cronucleus test. Photoinduced chromosomal damage is then cellular and molecular levels, the biological mechanisms of
evaluated either through the photoaberration test (usually on photoinduced hazards. This fundamental research work en-
CHO cell line) or through the photomicronucleus test on ables the determination of biological endpoints, which can
various cell types (immortalized CHO cells, V79 and normal be used to set up in vitro tests. This review has aimed at
human skin cells). Because the photochromosomal aberra- describing these various endpoints, from cytotoxicity and ap-
tion test (although considered as a regulatory test) is partic- optosis to photomutagenesis and chromosomal damage,
ularly difficult to perform and sensitive to the effect of cy- through the assessment of primary DNA damage and cellular
totoxicity, the photomicronucleus test has been proposed as response to UV-induced DNA lesions. For all these partic-
a feasible alternative. The in vitro micronucleus test, in the ular biological endpoints, specific in vitro approaches can be
process of international validation, is by far more adapted used to set up a strategy for testing of photoinduced toxicity
for screening purpose: it is easier and faster to carry out. and genotoxicity.
Moreover, the use of the micronucleus test renders photo- This review does not aim to propose a definite and stan-
induced clastogenic studies on human normal skin cells fea- dard strategy for testing but attempts to demonstrate the
sible (studies which are very difficult with the chromosomal strength of the mechanistically based in vitro methods on
aberration test because of the two-fold increase in the num- mammalian cells for phototoxicology assessment. In this
ber of chromosomes to be analyzed, as compared with CHO context, one has to also consider other existing in vitro meth-
cells). Finally, the micronucleus test is recognized as the ods, which can be quicker and easier to perform than tests
only one being able to detect both clastogenic and aneugenic on mammalian cells in culture and could be used in a pre-
events (75). These two kinds of events can then be distin- screening step in a funnel-shaped strategy. Among these
guished using molecular gene probing technologies, such as steps, some use a bacterial system (Photo-Ames test), but
fluorescence in situ hybridization (76). A concordance of the extreme sensitivity of bacteria to UV light makes the use
80–95% has been found between these two tests. of UV doses irrelevant when compared with human expo-
Thus, the photomicronucleus test has been recommended sure. The yeast Saccharomyces cerevisiae can also be used
in the strategy for photochemical genotoxicity assessment in as a quick and very informative model for photocytotoxicity
the IWGTP workshop held in Washington in 1999 (6). UV and photogenototoxicity assessment (80,81).
light (UVB, UVA as well as SSL) has been shown to induce Finally, the use of the reconstructed skin models from the
chromosomal damage in the micronucleus test (77,78). A second generation, including melanocytes, Langherans cells
variety of photosensitizing compounds have been evaluated (82) or cells defective for DNA repair (83), will enable the
in the photomicronucleus test on V79 cells (79) or on cir- development of new in vitro models that are always more
culating lymphocytes. predictive and informative of the phototoxic potential of UV
The manual scoring of the slides under a microscope is, light and of photosensitizing xenobiotics or more useful to
today, the major limitation for an extensive use of the pho- quantify the photoprotective effects of sunscreens.
tomicronucleus test for screening purpose. Several attempts
have been made to ease the reading of the results, either REFERENCES
through image analysis system or through flow cytometry.
1. Urbach, F. (1997) Ultraviolet radiation and skin cancer of hu-
The development of a reliable automatized system for scor- mans. J. Photochem. Photobiol. B: Biol. 40, 3–7.
ing would enable the routine use of photomicronucleus test 2. Gocke, E., S. Albertini, A. A. Chetelat, S. Kirchner and W.
in screening programs for photoinduced genotoxicity of Muster (1998) The photomutagenicity of fluoroquinolones and
chemical compounds (78). other drugs. Toxicol. Lett. 102–103, 375–381.
3. Spielmann, H., W. W. Lovell, E. Hölzle, B. E. Johnson, T.
Maurer, M. A. Miranda, W. J. W. Pape, O. Sapora and D. Sla-
CONCLUDING REMARKS dovski (1994) In vitro phototoxicity testing, the report and rec-
ommendations of ECVAM workshop 2. ATLA 22, 314–348.
Among the toxicological sciences, phototoxicology appears 4. Spielmann, H., L. Müller, D. Averbeck, M. Balls, S. Brendler-
to be a quite new and emerging field, which is illustrated by Schwaab, J. V. Castell, R. Curren, O. de Silva, N. K. Gibbs, M.
Liebsch, W. W. Lovell, H. F. Merk, J. F. Nash, N. J. Neumann,
the relative lack of a definite regulatory framework for test- W. J. W. Pape, P. Ulrich and H.-W. Vohr (2000) The second
ing as compared with other toxicological endpoints such as ECVAM workshop on phototoxicity testing. ATLA 28, 777–814.
chemical-induced mutagenesis. But phototoxicology is, 5. FDA. U.S. Department of Health and Human services, Food and
above all, quite unique among the toxicological fields be- Drug Administration, Center for Drug Evaluation and Research,
Pharmacology and Toxicology (January 2000) Guidance for in-
cause (1) the biological hazard of its etiologic agent, UV dustry photosafety testing. Draft guidance available on http://
light, is well characterized; and (2) the target organ, the skin, www.fda.gov/cder.
is unique, quite well known and modeled in vitro through 6. Gocke, E., L. Müller, P. J. Guzzie, S. Brendler-Schwaab, S.
Photochemistry and Photobiology, 2002, 75(5) 445

Bulera, C. F. Chignell, L. M. Henderson, A. Jacobs, H. Murli, 22. Kidd, S., J.-R. Meunier, N. J. Traynor, L. Marrot, C. Agapakis-
R. D. Snyder and N. Tanaka (2000) Considerations on photo- Caussé and N. K. Gibbs (2000) The phototumorigenic fluoro-
chemical genotoxicity: report of the International Workshop on quinolone, lomefloxacin, photosensitises p53 accumulation and
Genotoxicity test procedures working group. Environ. Mol. Mu- transcriptionnal activity in human skin cells. J. Photochem. Pho-
tagen. 35, 173–184. tobiol. 58, 26–31.
7. Spielmann, H., M. Balls, J. Dupuis, W. J. W. Pape, G. Pechov- 23. Nunez, G., M. A. Benedict, N. Yuanmimg Hu and N. Inohara
itch, O. de Silva, H.-G. Holzhütter, R. Clothier, P. Desolle, F. (1998) Caspases: the proteases of the apoptotic pathway. On-
Gerberick, M. Liebsch, W. W. Lovell, T. Maurer, U. Pfannen- cogene 17, 3237–3245.
becker, J. M. Potthast, M. Csato, D. Sladowski, W. Steiling and 24. Wolf, B. B., M. Schuler, F. Echeverri and D. R. Green (1999)
P. Brantom (1998) The international EU/COLIPA in vitro pho- Caspase-3 is the primary activator of apoptotic DNA fragmen-
totoxicity validation study: results of phase II (blind trial). Part tation via DNA fragmentation factor-45/inhibitor of caspase-ac-
1 the in vitro 3T3 NRU phototoxicity test. Toxicol. In Vitro 12, tivated DNAse inactivation. J. Biol. Chem. 274(43), 30651–
305–327. 30656.
8. Spielmann, H., M. Balls, J. Dupuis, W. J. W. Pape, O. de Silva, 25. Saunders, P. A., J. A. Cooper, M. M. Roodell, D. A. Schroeder,
H.-G. Holzhütter, F. Gerberick, M. Liebsch, W. W. Lovell and C. J. Borchert, A. L. Isaacson, M. J. Schendel, K. G. Godfrey,
U. Pfannenbecker (1998) A study on UV filter chemicals from D. R. Cahill, A. M. Walz, R. T. Loegering, H. Gaylord, I. J.
Annex VII of European Union Directive 76/768/EEC, in the in Woyno, A. E. Kaluyzhny, R. A. Krzyzek, F. Mortari, M. Tsang
vitro 3T3 NRU phototoxicity test. ATLA 26, 679–708. and C. F. Roff (2000) Quantification of active caspase 3 in
9. Anon (1998) European Commission, Directorate General XXIV. apoptotic cells. Anal. Biochem. 284, 114–124.
The scientific committee on cosmetic products and non food 26. Sage, E. (1993) Distribution and repair of photolesions in DNA:
products intended for consumers—SCCNFP/0069/98 final. genetic consequences and the role of sequence context. Photo-
Opinion on ‘‘in vitro methods to assess phototoxicity in the chem. Photobiol. 57, 163–174.
safety evaluation of cosmetic ingredients or mixtures of ingre- 27. Ley, R. D. and A. Fourtanier (2000) UVAI-induced edema and
dients’’, adopted by the plenary session of the SCCNFP of 25 pyrimidine dimers in murine skin. Photochem. Photobiol. 72(4),
November 1998. (website: http://europa.eu.int./comm.food/fs/ 485–487.
sc/sccp/out49pen.html) 28. Kuluncsics, Z., D. Perdiz, E. Brulay, B. Muel and E. Sage
10. Marrot, L., J. P. Belaidi, J. R. Meunier, P. Perez and C. Aga- (1999) Wavelength dependance of ultraviolet-induced DNA
pakis-Caussé (1999) The human melanocyte as a particular tar- damage distribution: involvement of direct or indirect mecha-
get for UVA radiation and an endpoint for photoprotection as- nisms and possible artefacts. J. Photochem. Photobiol. B: Biol.
sessment. Photochem. Photobiol. 69(6), 686–693. 49, 71–80.
11. Bernerd, F. and D. Asselineau (1998) UVA exposure of human 29. Averbeck, D. (1994) Photoaddition by furocoumarins. In Pho-
skin reconstructed in vitro induces apoptosis of dermal fibro- tobiology in Medicine (Edited by G. Jori, R. H. Pottier, M. A.
blasts: subsequent connective tissue repair and implications in J. Rodgers and T. G. Truscott.), pp. 71–98. Plenum Press, New
photoaging. Cell Death Differ. 5, 792–802. York.
12. Webb, S. J., D. J. Harrison and A. H. Wyllie (1997) Apoptosis: 30. Burcham, P. C. (1998) Genotoxic lipid peroxidation products:
an overview of the process and its relevance in disease. Adv. their DNA damaging properties and role in formation of endog-
Pharmacol. 41, 1–34. enous DNA adducts. Mutagenesis 13(3), 287–305.
13. Kluck, R. M., E. Bossy-Wetzel, D. R. Green and D. D. New- 31. Cadet, J., M. Berger, T. Douki and J. L. Ravanat (1997) Oxi-
meyer (1997) The release of cytochrome c from mitochondria: dative damage to DNA: formation, measurement, and biological
a primary site for Bcl-2 regulation of apoptosis. Science 275, significance. Rev. Physiol. Biochem. Pharmacol. 131, 1–87.
1132–1136. 32. Pflaum, M., S. Boiteux and B. Epe (1994) Visible light gener-
14. Liu, X., N. C. Kim, J. Yang, R. Jemmerson and X. Wang (1996) ates oxidative DNA base modifications in high excess of strand
Induction of apoptotic program in cell-free extracts: requirement breaks in mamalian cells. Carcinogenesis 15(2), 297–300.
for dATP and cytochrome c. Cell 86, 147–157. 33. Kielbassa, C., L. Roza and B. Epe (1997) Wavelength depen-
15. Li, P., D. Nijhawan, I. Budihardjo, S. M. Srinivasula, M. Ah- dence of oxidative DNA damage by UV and visible light. Car-
mad, S. Alnemri and X. Wang (1997) Cytochrome c and dATP- cinogenesis 18(4), 811–816.
dependent formation of Apaf-1/caspase-9 complex initiates an 34. Kvam, E. and R. M. Tyrrell (1997) Induction of oxidative DNA
apoptotic protease cascade. Cell 91, 479–489. base damage in human skin cells by UV and near visible radi-
16. Zhou, H., W. J. Henzel, X. Liu, A. Lutschg and X. Wang (1997) ation. Carcinogenesis 18(12), 2379–2384.
Apaf-1, a human protein homologous to C. elegans CED-4, par- 35. Cooke, M. S., N. Mistry, A. Lapado, K. E. Herbert and J. Lunec
ticipates in cytochrome c-dependent activation of caspase-3. (2000) Immunochemical quantitation of UV induced oxidative
Cell 90, 405–413. and dimeric DNA damage to human keratinocytes. Free Radic.
17. Martin, S. J., C. P. Reutelingsperger, A. J. MacGahon, J. A. Res. 33, 369–381.
Rader, R. C. van Schie, D. M. LaFace and D. R. Green (1995) 36. Peak, M. J., J. G. Peak and B. A. Carnes (1987) Induction of
Early redistribution of plasma membrane phosphatidylserine is direct and indirect single-strand breaks in human cell DNA by
a general feature of apoptosis regardless of the initiating stim- far- and near- ultraviolet radiations: action spectrum and mech-
ulus: inhibition by overexpression of Bcl-2 and Abl. J. Exp. anisms. Photochem. Photobiol. 45(3), 381–387.
Med. 182, 1545–1556. 37. Hanawalt, P. C. (1994) Transcription coupled repair and human
18. Sarasin, A. (1999) The molecular pathways of ultraviolet-in- disease. Science 266, 1957–1958.
duced carcinogenesis. Mutat. Res. 428, 5–10. 38. Dhénaut, A., S. Boiteux and J. P. Radicella (2000) Character-
19. Morita, A., T. Wersel, H. Stege, C. Ahrens, K. Karmann, M. ization of the hOGGI promoter and its expression during the
Grewe, S. Grether-Beck, T. Ruzicka, A. Kapp, L. O. Klotz, H. cell cycle. Mutat. Res. 461, 109–118.
Sies and J. Krutmann (1997) Evidence that singlet oxygen-in- 39. Le Page, F., E. E. Kwoh, A. Avrutskaya, A. Gentil, S. A. Lea-
duced human T helper cell apoptosis is the basic mechanism of don, A. Sarasin and P. K. Cooper (2000) Transcription-coupled
ultraviolet-A radiation phototherapy. J. Exp. Med. 186(10), repair of 8-oxoguanine: requirement for XPG, TFIIH, and CSB
1763–1768. and implications for Cockayne syndrome. Cell 101, 159–171.
20. Queille, S., C. Drougard, A. Sarasin and L. Daya-Grosjean 40. Sancar, A. and G. B. Sancar (1988) DNA repair enzymes. Annu.
(2001) Effects of XPD mutations on UV-induced apoptosis in Rev. Biochem. 57, 29–67.
relation to skin cancer-proneness in repair deficient syndromes. 41. Douki, T., T. Delatour, F. Paganon and J. Cadet (1996) Mea-
J. Investig. Dermatol. 117, 1162–1170. surement of oxidative damage at pyrimidine bases in gamma-
21. Kaufmann, S. H., S. Desnoyers, Y. Ottaviano, N. E. Davidson irradiated DNA. Chem. Res. Toxicol. 9, 1145–1151.
and G. G. Poirier (1993) Specific proteolytic cleavage of 42. Douki, T., M. Court, S. Sauvaigo, F. Odin and J. Cadet (2000)
poly(ADP-ribose) polymerase: an early marker of chemothera- Formation of the main UV induced thymine dimeric lesions
py-induced apoptosis. Cancer Res. 53, 3976–3985. within isolated and cellular DNA as measured by high perfor-
446 Jean-Roch Meunier et al.

mance liquid chromatography-tandem mass spectrometry. J. 60. Benhamou, S. and A. Sarasin (2000) Variability in nucleotide
Biol. Chem. 275, 11678–11685. excision repair and cancer risk: a review. Mutat. Res. 462, 149–
43. Van Hoffen, A., J. Venema, R. Meschini, A. A. Van Zeeland 158.
and L. H. Mullenders (1995) Transcription-coupled repair re- 61. Lane, D. P. (1992) P53 guardian of the genome. Nature 358,
moves both cyclobutane pyrimidine dimers and (6-4) photo- 15–16.
products with equal efficiency and in a sequential way form 62. Maltzman, W. and L. Czyzyk (1984) UV irradiation stimulates
transcribed DNA in xeroderma pigmentosum group C fibro- level of p53 cellular tumor antigen in non transformed mouse
blasts. EMBO J. 14(2), 360–367. cells. Mol. Cell. Biol. 4, 1689–1694.
44. Eveno, E., F. Bourre, X. Quilliet, O. Chevallier-Lagente, L. 63. Hall, P. A., P. H. McKee, H. D. Menage, R. Dover and D. P.
Roza, A. P. Eker, W. J. Kleijer, O. Nikaido, M. Stefanini, J. H. Lane (1993) High levels of p53 protein in UV irradiated normal
J. Hoeijmakers, D. Bootsma, J. E. Cleaver, A. Sarasin and M. skin. Oncogene 8, 203–207.
Mezzina (1995) Different removal of ultraviolet photoproducts 64. Brash, D. E., J. A. Rudolph, J. A. Simon, A. Lin, G. J. Mc-
in genetically related xeroderma pigmentosum and trichothiod- Kenna, H. P. Baden, A. J. Halperin and J. Pontén (1991) A role
ystrophy diseases. Cancer Res. 55, 4325–4332. for sunlight in skin cancer: UV-induced p53 mutations in squa-
45. Clingen, P. H., C. F. Arlett, L. Roza, T. Mori, O. Nikaido and mous cell carcinoma. Proc. Natl. Acad. Sci. USA 88, 10124–
M. H. Green (1995) Induction of cyclobutane pyrimidine di- 10128.
mers, pyrimidine (6-4)pyrimidone photoproducts, and Dewar 65. Daya-Grosjean, L. N. Dumaz and A. Sarasin (1995) The spec-
valence isomers by natural sunlight in normal human mononu- ificity of p53 mutation spectra in sunlight induced human can-
clear cells. Cancer Res. 55, 2245–2248. cers. Photochem. Photobiol. 28, 115–124.
46. Chadwick, C. A., C. S. Potten, O. Nikaido, T. Matsunaga, C. 66. Bodak, N., S. Queille, M. F. Avril, B. Bouadjar, C. Drougard,
Proby and A. R. Young (1995) The detection of cyclobutane A. Sarasin and L. Daya-Grosjean (1999) High levels of patched
thymine dimers, (6-4) photolesions and the Dewar photoisomers gene mutations in basal cell carcinomas from patients with xe-
in sections of UV-irradiated human skin using specific antibod- roderma pigmentosum. Proc. Natl. Acad. Sci. USA 96, 5117–
ies, and the demonstration of depth penetration effects. J. Pho- 5122.
tochem. Photobiol. B: Biol. 28, 163–170. 67. Giglia, G., N. Dumaz, C. Drougard, M. F. Avril, L. Daya-Gros-
47. Bernerd, F., C. Vioux and D. Asselineau (2000) Evaluation of jean and A. Sarasin (1998) P53 mutations in skin and internal
the protective effect of sunscreens on in vitro reconstructed hu- tumors of xeroderma pigmentosum patients belonging to the
man skin exposed to UVB or UVA irradiation. Photochem. Pho- complementation group C. Cancer Res. 58, 4402–4409.
tobiol. 71(3), 314–320. 68. Ouhtit, A., M. Ueda, H. Nakazawa, M. Ichihashi, N. Dumaz, A.
48. Salles, B., P. Frit, C. Provot, J.-P. Jaeg and P. Calsou (1995) In Sarasin and H. Yamasaki (1997) Quantitative detection of ul-
vitro eukaryotic DNA excision repair assays: an overview. traviolet-specific p53 mutations in normal skin from Japanese
Biochimie 77, 796–802. patients. Cancer Epidemiol. Biomark. Prev. 6, 433–438.
49. Östling, O. and K. J. Johanson (1984) Microelectrophoretic 69. Drobetsky, E. A., J. Turcotte and A. Châteauneuf (1995) A role
study of radiation-induced DNA damage in individual cells. for ultraviolet A in solar mutagenesis. Proc. Natl. Acad. Sci.
Biochem. Biophys. Res. Commun. 123, 291–298. USA 92, 2350–2354.
50. Singh, N. P., M. T. McCoy, R. R. Tice and E. L. Schneider 70. Robert, C., B. Muel, A. Benoit, A. Dubertret, A. Sarasin and A.
(1988) A simple technique for quantitation of low levels of Stary (1996) Cell survival and shuttle vector mutagenesis in-
DNA damage in individual cells. Exp. Cell Res. 175, 184–191. duced by ultraviolet A and ultraviolet B radiation in a human
51. Tice, R. R., E. Agurell, D. Anderson, B. Burlinson, A. Hart- cell line. J. Investig. Dermatol. 106, 721–728.
mann, H. Kobayashi, Y. Miyamae, E. Rojas, J.-C. Ryu and Y. 71. Nataraj, A. J., P. Wolf, L. Cerroni and H. N. Ananthaswamy
F. Sasaki (2000) Single cell gel/comet assay: guidelines for in (1997) p53 Mutation in squamous cell carcinomas from psori-
vitro and in vivo genetic toxicology testing. Environ. Mol. Mu- asis patients treated with psoralen 1 UVA (PUVA). J. Investig.
tagen. 35, 206–221. Dermatol. 109, 238–243.
52. Olive, P. L., D. Wlodek, R. E. Durand and J. P. Banath (1992) 72. Pflaum, M., C. Kielbassa, M. Garmyn and B. Epe (1998) Oxi-
Factors influencing DNA migration from individual cells sub- dative DNA damage induced by visible light in mammalian
jected to gel electrophoresis. Exp. Cell Res. 198, 259–267. cells: extent, inhibition by antioxidant and genotoxic effects.
53. Sauvaigo, S., C. Serres, N. Signorini, N. Emonet, M.-J. Richard Mutat. Res. 408, 137–146.
and J. Cadet (1998) Use of the single-cell gel electrophoresis 73. Müller, L. and P. Kasper (1998) The relevance of phototmuta-
assay for the immunofluorescent detection of specific DNA genicity testing as a predictor of photocarcinogenicity. Int. J.
damage. Anal. Biochem. 259, 1–7. Toxicol. 17, 551–558.
54. Collins, A. R., M. Dusinka, C. M. Gedik and R. Stetina (1996) 74. O.E.C.D. Guideline for testing of chemicals genetic toxicology:
Oxidative damage to DNA: do we have a reliable biomarker? in vitro mammalian cytogenetic test, 21st July 1997, No. 473,
Environ. Health Perspect. 104(S3), 465–469. pp. 1–10.
55. Alapetite, C., T. Wachter, E. Sage and E. Moustacchi (1996) 75. Fenech, M. (1993) The cytokinesis-block micronucleus tech-
Use of the alkaline comet assay to detect DNA repair deficien- nique: a detailed description of the method and its application
cies in human fibroblasts exposed to UVC, UVB, UVA and g- to genotoxicity studies in human populations. Mutat. Res. 285,
rays. Int. J. Radiat. Biol. 69, 359–369. 35–44.
56. Marrot, L., J. P. Belaidi, C. Chaubo, J. R. Meunier, P. Perez 76. Kirsch-Volders, M., A. Elhajouhi, E. Cundari and P. Van Hum-
and C. Agapakis-Caussé (1998) An in vitro strategy to evaluate melen (1997) The in-vitro micronucleus test: a multi-endpoint
the phototoxicity of solar UV at the molecular and cellular level: assay to detect simultaneously mitotic delay, apoptosis, chro-
application to photoprotection assessment. Eur. J. Dermatol. 8, mosome breakage, chromosome loss and non-disjunction. Mu-
403–412. tat. Res. 392, 19–30.
57. Anderson, D., T.-W. Yu, B. J. Philipps and P. Schmezer (1994) 77. Weller, E. M., J. Hain, T. Jung, R. Kinder, M. Köfferlein, W.
The effect of various antioxidants and other modifying agents Burkart and M. Nüsse (1996) UVB induced cell cycle pertur-
on oxygen-radical-generated DNA damage in human lympho- bations, micronucleus induction, and modulation by caffeine in
cytes in the COMET assay. Mutat. Res. 307, 261–271. human keratinocytes. Int. J. Radiat. Biol. 69(3), 371–384.
58. Reavy, H. J., N. J. Traynor and N. K. Gibbs (1997) Photogen- 78. Emri, G., E. Wenczl, P. Van Erp, J. Jans, L. Roza, I. Horkay
otoxicity of skin phototumorigenic fluoroquinolone antibiotics and A. A. Shothorst (2000) Low dose of UVB or UVA induce
detected using the Comet assay. Photochem. Photobiol. 66, chromosomal aberrations in cultured human skin cells. J. Inves-
368–373. tig. Dermatol. 115, 435–440.
59. Cleaver, J. E. and G. H. Thomas (1981) Measurement of un- 79. Kersten, B., J. Zhang, S. Y. Brendler-Shwaab, P. Kasper and L.
scheduled synthesis by autoradiography. In DNA Repair, Vol. Müller (1999) The application of the micronucleus test in Chi-
1, Part B (Edited by E. C. Friedberg and P. C. Hanawalt), pp. nese hamster V79 cells to detect drug-induced photogenotox-
277–289. Marcel Dekker Inc., New York and Basel. icity. Mutat. Res. 445, 55–71.
Photochemistry and Photobiology, 2002, 75(5) 447

80. Averbeck, D. and E. Moustacchi (1979) Genetic effects of 3- 82. Régnier, M., M.-J. Staquet, D. Schmitt and R. Schmidt (1997)
carbethoxypsoralen, angecilin, psoralen and 8-methoxypsoralen Integration of Langerhans cells into a pigmented reconstructed
plus 365 nm irradiation in S. cerevisiae: induction of reversions, human epidermis. J. Investig. Dermatol. 109, 510–512.
mitotics crossing-over, gene conversion and cytoplasmic ‘‘pe- 83. Bernerd, F., D. Asselineau, C. Vioux, O. Chevallier-Lagente, B.
tite’’ mutations. Mutat. Res. 68, 133–148. Bouadjar, A. Sarasin and T. Magnaldo (2001) Clues to epider-
81. Marrot, L. and C. Agapakis-Caussé (2000) Differences in the mal cancer proneness revealed by reconstruction of DNA repair-
photogenotoxic potential of two fluoroquinolones as shown in deficient xeroderma pigmentosum skin in vitro. Proc. Natl.
diploid yeast strain (Saccharomyces cerevisiae) and supercoiled Acad. Sci. USA 98(14), 7817–7822.
plasmid DNA. Mutat. Res. 468, 1–9.

You might also like