An in vitro strategy to evaluate the phototoxicity

ARTICLE

Skin cancers are the most frequent human cancers, and it is now well established that the ultraviolet components of sunlight radiation play a major role in the induction of photocarcinogenesis [1]. Sunlight is a complete carcinogen, and several lines of evidence implicate DNA as a central target. For instance, individuals with the genetic disorder xeroderma pigmentosum (XP) who are deficient in DNA repair, are extremely sensitive and susceptible to developing skin cancer on sun exposed areas [2].

Moreover, cyclobutane pyrimidine dimers (CPD), the most frequent photolesions induced by UVB, are thought to be directly involved in photo-induced melanogenesis and immunosuppression [3, 4]. Thus, in order to avoid health hazards, efficient photoprotection at the DNA level appears to be an important goal.

Prevention of skin erythema formation is usually taken as the criterion for the effectiveness of sunscreens. However, because of the complexity of the processes involved in photocarcinogenesis, it is clear that more information is needed at cellular and molecular levels in order to develop optimal protection against photocarcinogenic risks related to sunlight exposure. It looks even more important, when considering that DNA damage and p53 expression, a specific response to genotoxic stress, can be already induced by suberythemal doses [5].

Hence, we propose a strategy combining various simple in vitro models to test whether a sunscreen is devoid of genotoxic side effects and can provide reliable photoprotection in conditions similar to actual use.

This approach is a first and very informative step aiming at selecting products, which then will be studied in more complex systems such as skin in vitro or in vivo.

In order, to mimic natural sunlight in terms of spectral power distribution, we used a solar simulator as the radiation source, and the light reaching the biological sample was precisely characterized by spectroradiometry. The test battery includes the following biological systems:

1) Supercoiled circular DNA allowing the detection of structural alterations such as pyrimidine dimers. 2) The yeast Saccharomyces cerevisiae, which is considered as a paradigm for DNA repair studies [6], and gives information about phototoxicity and genetic alterations such as recombination and point mutagenesis. 3) Comet assay which indicates the level of DNA damage and DNA repair in human skin cells [7, 8]. 4) p53 expression, which can be taken as a hallmark of cellular response to genotoxic stress induced by UV [9, 10], was measured in the nucleus of human keratinocytes. 5) In human melanocytes, specific target cells for UV, phototoxicity as well as the induction of pigmentation were evaluated. The latter phenomenon has been reported to be a consequence of DNA repair [4].

Using the above strategy, the sunscreen Mexoryl^® SX was shown to effectively protect from the cytotoxicity and genotoxicity of solar UV even when dissolved in the medium and thus in contact with the cells. These results are in good agreement with those obtained in vivo [11].

Materials and methods

Chemicals

Media for yeast experiments were prepared as described elsewhere [12] with chemicals from Difco or Merck. Phosphate buffered saline (PBS) was from Gibco-BRL. Media for human cells were from Clonetics Inc.

Agarose for the comet assay was the low melting Incert^® Agarose from FMC.

Excell gels SDS from Pharmacia were used for SDS Page.

Nitro-cellulose membranes (Hybond, Amersham) were used for protein transfer.

Other chemicals were from Sigma.

Endonuclease V from phage T4 was obtained from Dr. Mullenders, Leiden University, The Netherlands, or, from Micrococcus luteus, as a gift from Dr. Ley, the Lovelace Institute, Albuquerque, USA.

The sunscreen agent used in this study was the terephtalylidene dicamphor sulfonic acid (Mexoryl^® SX).

Biological systems

Supercoiled circular DNA was pBR322 from Boehringer. Following irradiation, the digestion by endonuclease V was performed as described elsewhere [13] and the percentages of supercoiled DNA (form I) and nicked DNA (form II) were quantified by densitometry (Gel Doc 1000, Biorad, Ivry-sur-Seine) after electrophoresis in 1% agarose gel and ethidium bromide staining.

The yeast strain (Saccharomyces cerevisiae) was the D7-rad3 mutant, deficient in nucleotide excision repair with the following genotype: a/alpha, rad3/rad3, ade 240/ade 2-119, trp 5-12/trp 5-27, ilv1-92/ilv1-92. This strain was a gift from Dr. Averbeck (Institut Curie, Paris).

Normal human keratinocytes (NHK) and normal human melanocytes were neonatal cells from Clonetics Inc., which were cultured as described [14, 15].

Light source and spectral measurements

The light source was a solar UV simulator from Oriel equipped with a 1000 W Xenon short arc lamp, a dichromic mirror and the Oriel "atmospheric attenuation" filter (comparable to a 1 mm WG 320 filter). The beam size was 152 x 152 mm. The incident and the transmitted UV spectra were analysed with a spectroradiometer (Instaspec III, Oriel). When the sunscreen was used, the spectrum of the light reaching the samples was evaluated as follows: the probe of the spectroradiometer was placed under a quartz Petri dish (designed for us by Hellma) that contained the sunscreen dissolved in the buffer in the same conditions as in the experiments (especially with regard to the depth of the liquid) and the spectrum was recorded. Since each wavelength has its own biological effect, we preferred to consider the exposure time for a given UV spectrum rather than a global dose. By integration of the area under the spectrum, we could however measure the irradiance of the three regions of solar simulated UV (SSUV): UVB (290-320 nm): 13.4 W/m², UVA₂ (320-340 nm): 25.3 W/m² and UVA₁ (340-400 nm): 90 W/m².

In the yeast experiments, according to the photosensitivity of the D7-rad3 strain, the whole fluence was reduced 3.5 times by the aperture set from Oriel.

Figure 1 shows the spectral power distribution of the light transmitted by 2 mm thick solutions of the sunscreen Mexoryl^® SX in PBS. The concentrations of Mexoryl^® SX chosen for the experiments were such that the results could be extrapolated, in terms of optical density, to what could occur when a preparation is applied on the skin: a 2 mm deep solution at 0.016% equals a 20 µm deep layer at 1.6%.

Irradiation procedure

Supercoiled circular DNA (pBR322) was irradiated in NaCl 10 mM at a concentration of 20 ng/µl. Each sample was placed in the well of a multiwell titration plate and exposed for 5, 10 or 15 min. In order to mimic the experimental procedure used for cells, the sunscreen was placed in a quartz Petri dish (2 mm thick solutions) over the plate.

Yeast cells (10⁷ cells/ml) in stationary phase were exposed to SSUV in PBS in a plastic Petri dish (60 mm in diameter). The depth of the cell suspension was approximately 2 mm. During exposure, a gentle shaking ensured the homogeneity of the irradiation without changing turbidity. When Mexoryl^® SX was tested, it was added to the cell suspension so that it was in direct contact with the cells. Since the D7-rad3 strain is photosensitive, we had to reduce the initial output of our solar simulator (we used the aperture set proposed by Oriel in this regard) otherwise the toxicity would have been too high even for short exposure times.

Human cells were exposed to SSUV in PBS in 60 mm culture dishes. When tested, the sunscreen Mexoryl^® SX, was suitably diluted in PBS. In any case, the volume of liquid (PBS or PBS + Mexoryl^® SX) was such that its depth over the attached cell layer was approximately 2 mm. After irradiation, the cells were washed once with PBS and put back into their initial medium at 37° C in a 5% CO₂ atmosphere for an appropriate incubation time depending on the type of experiment.

Cell survival

For yeast cells, survivors and colonies genetically altered in the adenine locus, were counted after growth on complete solid medium while convertants in the tryptophan locus and revertants in the isoleucine-valine locus were selected on suitable supplemented minimal media [12, 16].

For human cells, cell survival was assessed by the MTT assay (Boehringer Mannheim, Germany) according to the manufacturer's instructions and quantified by colorimetric measurements with a twin reader plus (Labsystems).

Single-cell electrophoresis (comet) assay

Keratinocytes were exposed for 8 min to SSUV. After one hour, cells were embedded in a 0.5% agarose layer and the comet assay was performed as described by Singh [17]. After alkaline gel electrophoresis (20 min at 25 volts) and neutralisation with Tris pH8, DNA was stained with ethidium bromide and the comets were examined and photographed in a fluorescent microscope.

Preparation of nuclear extracts and Western blot analysis of p53 protein

Keratinocytes were exposed for 8 min to SSUV. After 4 hrs, 24 hrs and 48 hrs post-treatment incubation, the cells were harvested, and proteins of the nucleus were extracted according to Lin & Benchimol [18]. Equal amounts of protein were separated by SDS Page electrophoresis and transferred onto a nitro-cellulose membrane. Western blot analysis was carried out according to ECL western blot system's instructions (Amersham, Amersham place, GB), using the anti-p53 antibody clone D01 (Oncogene Sciences, Cambridge, MA, USA). The films were analysed by densitometry using the Gel Doc 1000 device (Biorad, Ivry-sur-Seine, France).

Melanin content

Experiments were performed with the pigment cells in their third to seventh passage. Constitutive melanin was determined 5 days after irradiation as described by Lee et al. [19].

Melanocytes were harvested and rinsed with PBS. Melanin was solubilized in 0.2 M NaOH and measured spectrophotometrically at an absorbance of 475 nm against a standard curve of known concentrations of synthetic melanin (Sigma Chemicals).

Tyrosinase activity

Tyrosinase activity was assayed spectrophotometrically by following the oxidation of L-dopa to dopachrome at 475 nm [20]. Cells were washed with PBS and lysed with 400 µl of 1% triton-X/PBS. After sonication and vibration, 150 µl of 5 µM L-dopa was added to 50 µl of the cellular extract. Control wells contained 50 µl lysis buffer or boiled cell lysate. The absorbance values were compared with a standard curve obtained with mushroom tyrosinase (Sigma Chemicals). The standard curve was linear within the range of experimental values, and there was no increase in absorbance in the control wells.

Results

Use of supercoiled circular DNA for the detection of structural alterations

The induction of pyrimidine dimers in DNA by the UVB portion of sunlight is well documented, but a contribution by UVA in such a process has also been reported [21]. Moreover in skin tumors, photomutagenesis hot spots occur mainly at dipyrimidine sites, suggesting a major role for dimers [22, 23]. Therefore, the first step to demonstrate photoprotection can be to ensure that pyrimidine dimer formation decreases in the presence of a sunscreen. The enzyme endonuclease V nicks the DNA at the level of cyclobutane pyrimidine dimers (CPD). This allows the conversion of a damaged supercoiled plasmid into a relaxed one, both forms being easily separated by agarose gel electrophoresis. Figure 2 shows that this approach can be used in a quantitative way and that the sunscreen Mexoryl^® SX prevents CPD induction in a dose-dependent manner when 2 mm-deep solutions are placed over the DNA. When the sunscreen and DNA were mixed, the screening effect was no longer homogeneous, especially for DNA molecules at the top of the liquid. We observed that, in this case, the protection was approximately three times less than when the sunscreen was placed over the sample. However, in order to exclude the possibility of photosensitization, experiments on cultured cells are obviously required.

Mitotic recombination and mutagenesis induced by solar simulated UV in Saccharomyces cerevisiae

The yeast Saccharomyces cerevisiae is a simple, unicellular eukaryote and it is considered to be a relevant model for DNA repair and genotoxicity studies. In fact, there is a strong conservation of the components of nucleotide excision repair in eukaryotes, for instance, homologues of human xeroderma pigmentosum cells exist in yeast [6]. Here, we used a diploid rad-3 mutant strain which is defective in excision repair. The effects of SSUV were assessed for the following biological end points: clonogenic cell survival, mitotic intergenic recombination involving the adenine locus, mitotic intragenic recombination (gene conversion) involving the tryptophan locus and point mutations (reversions) in the isoleucine-valine locus.

The yeast cells were exposed for up to 20 min to SSUV in the presence or in the absence of Mexoryl^® SX in the buffer. With the aperture used, the initial irradiance was reduced and the spectrum was qualitatively and quantitatively similar to natural sunlight although slightly more intense in the UVB region. In the absence of the sunscreen, the remaining fraction decreased with irradiation time in a dose-dependent manner, but in the presence of the sunscreen it was maintained at the control level even after 20 min of exposure (Fig. 3). This experiment clearly demonstrates that the cells were protected from SSUV-induced phototoxicity in the presence of Mexoryl^® SX.

In order to focus mainly on genetic alterations and to reduce the level of phototoxicity, the chosen irradiation times did not exceed 10 min in the subsequent experiments. The D7-rad3 strain being unable to repair DNA efficiently, the frequency of genetically altered colonies induced was significantly greater than that in untreated samples, and this already at low doses where toxicity was virtually absent. In Figure 3, we show that intergenic recombination, intragenic recombination and point mutations increased with the irradiation time in a dose-dependent manner and that Mexoryl^® SX markedly reduced the rate of those genetic events. It is also noteworthy that colonies showing specific phenotypes for genetic alterations in the adenine locus (mainly red colonies) could grow on the solid selection media used for tryptophan convertants or isoleucine-valine mutants. We assume that these colonies were the result of both types of genetic events. Interestingly, their occurrence was totally abolished in the presence of the sunscreen (not shown).

Thus, Mexoryl^® SX protected from cytotoxicity and genotoxicity induced by SSUV, and this, even in eukaryotic cells deficient in nucleotide excision repair.

Comet assay:
a test to assess genotoxicity of solar-simulated UV on human skin cells

When a cell is exposed to sunlight, various lesions are formed in its genomic DNA. UVB induces mainly pyrimidine dimers and pyrimidine (6-4) pyrimidone photoproducts whereas UVA produces mostly strand breaks and oxidative damage in DNA by type II photosensitization. The excision of these lesions provokes a transient breakage of the DNA backbone which can be detected by alkaline gel electrophoresis. Moreover, some damage is alkali labile and leads to additional strand breaks at high pH. This is the case for the 6-4 dimer and more particularly for its photoisomer, the Dewar photoproduct [24], as well as for abasic sites generated by various glycosylases during base excision repair [25].

The comet assay (also named single cell gel electrophoresis) is a simple and visual technique for measuring DNA breakage in individual cells. It has been extensively used to characterize genotoxins and also to analyse DNA repair in response to phototoxic effects of UV components of the solar spectrum [7, 8]. Figure 4 shows that after exposure of normal human keratinocytes to SSUV for 8 min, typical comets could be observed as early as one hour post-irradiation when no significant toxicity was detected by the MTT assay (not shown). These comets reflect the induction of photodamage and repair. If the same experiment was performed in the presence of Mexoryl^® SX dissolved in the medium (2 mm-deep solution, 0.008% and 0.016% in PBS), comet induction was almost totally abrogated (a few comets are still observed for 0.008%) suggesting that genomic DNA had been protected in the presence of the sunscreen. It is noteworthy moreover that this photoprotective effect of Mexoryl^® SX was also found at the cellular proliferation level assessed 24 hrs post-irradiation. The unprotected sample showed a 60% decrease in proliferation where the protected samples showed a 13% decrease (Mexoryl^® SX 0.008%) or was unaffected (0.016% Mexoryl^® SX). We can conclude that both SSUV cytotoxicity and genotoxicity were abrogated when the sunscreen was dissolved in the medium during the exposure. These results strongly suggest the absence of photosensitization at least in our experimental conditions. Furthermore, preincubation of the cells with the sunscreen before irradiation apparently did not affect the repair process since comet formation still occurred when Mexoryl^® SX was removed before SSUV exposure.

Analysis of p53 expression, a specific cellular response to photogenotoxic stress

The tumor suppressor gene p53 plays a major role in the protection of cells from DNA damage. In response to genotoxic stress, the p53 nucleoprotein accumulates and specifically binds to DNA, activates transcription and either regulates cell cycle progression by arresting cells in late G1 [26] or triggers apoptosis. The upregulation of p53 protein in response to UV is well documented. Moreover, p53 is the most commonly mutated gene found in human cancers and especially, in non-melanoma skin cancer [27]. Thus, the determination of the p53 protein status in normal human skin cells irradiated with SSUV can be taken as a hallmark of the cellular response to UV-induced stress. Since keratinocytes are the type of skin cells implicated in non-melanoma skin cancer i.e. basal cell carcinoma and squamous cell carcinoma (BCC and SCC), it is of particular interest to evaluate their p53 protein levels after solar-simulated UV irradiation and to study the photoprotective effect of a given sunscreen in this system. Here, in vitro cultured human keratinocytes were irradiated for 8 min with SSUV and p53 protein levels were analysed by Western Blotting.

Figure 5 shows that p53 protein started to accumulate 4 hrs after SSUV exposure, peaked at 24 hrs and decreased at 48 hrs. When Mexoryl^® SX was present in the medium, cells were protected from SSUV-induced genotoxic stress in a concentration-dependent manner. No p53 accumulation could be detected for 0.016% Mexoryl^® SX (data not shown).

Since a lower concentration (0.004%) of sunscreen gave rise to intermediate p53 activation kinetics, the photoprotective effect of Mexoryl^® SX cannot be assigned to the knock out of the p53-dependent pathway of cellular defence against UV-induced stress. As verified earlier, preincubation with Mexoryl^® SX did not disrupt p53 activation.

Induction of pigmentation:
the melanocyte as a particular target for solar UV

An obvious cutaneous effect of sun exposure is increased pigmentation which results from stimulation of melanogenesis and the transfer of melanin to keratinocytes. This process can be considered as a defence mechanism against photo-oxidative stress and some studies have suggested that DNA damage could trigger pigmentation [4]. In mammalian melanocytes, melanogenesis starts with a two step oxidation of tyrosine producing the reactive dopa-quinone and then dihydroxy-indole. Polymerisation of this molecule generates black eumelanin. Tyrosine hydroxylation and dopa oxidation are controlled by the same enzyme, tyrosinase, which can be activated at a post-translational level in response to various stresses. Tyrosinase-associated tyrosinase related proteins TRPI and TRP2 modulate the amount and nature of melanin produced [28]. On the other hand, the role of sun exposure in the induction of cutaneous malignant melanoma has been extensively discussed during the last two decades and some studies suggest that UV, especially at acute doses, could be an important etiological agent for this type of cancer. Thus, the melanocyte is a particularly relevant model for studying phototoxicity.

Here, normal human melanocytes were cultivated in a specific medium where they were able to respond to alphaMSH [15]. In these conditions, it was possible to measure a significant increase in tyrosinase activity and melanin content after exposure to SSUV (Fig. 6). It is noteworthy that the stimulation of melanogenesis occurred in parallel with cytotoxicity as if a minimum amount of cellular alterations affecting the whole metabolism was necessary for triggering pigmentation. This was confirmed by the effect of the sunscreen Mexoryl^® SX which ensured cell survival and maintained melanogenesis at a basal level. These results are in good agreement with those published by Abdel-Malek et al. [29] for UVB, but the contribution of UVA, especially, at the membrane level, cannot be ruled out. It is likely that the melanocytes are transiently arrested in cell cycle progression in order to repair DNA damage, this last process being able to stimulate pigmentation [30]. Altogether, these data show that, under our experimental conditions, Mexoryl^® SX not only protected the melanocytes from phototoxic effects of solar UV, but also that this sunscreen did not sensitize these cells at the concentrations used because, otherwise, stimulation of tyrosinase activity would have been observed.

Discussion

DNA damage induction by sunlight, and repair are at the origin of many specific biological processes such as immunosuppression, melanogenesis and, in the worst case, tumor development. Since photolesions can be produced in the genetic material even at sub-erythemal doses, it is essential to evaluate the protection provided by a given sunscreen at the molecular level. For ethical and practical reasons, animals and human skin explants cannot be used routinely to screen a large number of molecules. A panel of relevant in vitro tests is thus extremely useful for a first screening. The strategy reported here is based on complementary in vitro approaches, from simple plasmid DNA to much more complex human keratinocytes and melanocytes in culture.

Above all, the nature of the light used in the experiments is of the utmost importance. First, the spectral power distribution must be as similar as possible to that of solar UV. Recent controversies have underlined the importance of this particular point [31]. Without such caution, no reasonable extrapolation of the results to the natural conditions can be done. In addition to its own ability to induce DNA damage, the UVA part of solar UV could modulate the effects of UVB, in particular for the nature of mutation hotspots [32] and the influence of UVA on the activity of repair enzymes has been reported [33]. These data suggest that working with the full solar UV spectrum is truly a critical point. Moreover, thanks to spectroradiometry, it is possible to precisely characterize the light transmitted to the sample through the sunscreen. This is very helpful for understanding the persistence of some biological effects that could occur despite the presence of photoprotection.

The first and most simple model used is supercoiled circular DNA, which allows the detection of photodamage, that can be converted into strand breaks. In this study, we have shown that the amount of the well known photolesion, pyrimidine dimer, can be easily quantified after endonuclease V digestion. Such an experiment could also be performed for the 6-4 pyrimidine-pyrimidinone dimer taking advantage of its alkali-lability. Since this test is also well adapted for oxidative damage produced by type II photosensitization [34] it allows the discrimination of molecules that produce active oxygen species when irradiated.

However, the behaviour of isolated DNA in a test tube does not of course give the exact image of what happens in the nucleus of an eukaryotic cell. The yeast Saccharomyces cerevisiae is a very convenient and useful model as a second step in this strategy. Despite some metabolic differences with mammalian cell systems, the yeast Saccharomyces cerevisiae is considered a paradigm for DNA repair mechanisms because the proteins involved in nucleotide or base excision repair pathways are highly homologous to those of humans. From an experimental point of view, the D7-rad3 strain provides at the same time complementary information on cell survival, intergenic and intragenic mitotic recombination and point mutagenesis. At present, a comparable amount of data cannot be obtained with mammalian cells in culture.

Experiments can also be performed in the presence of exogenous photosensitizers that become phototoxic in specific spectral domains and particularly, upon exposure to solar UVA. Following the example of others [12], we are actually using psoralens in this regard. Our results presented above, clearly demonstrate that the sunscreen Mexoryl^® SX, when in contact with yeast cells, prevents phototoxic effects in terms of clonogenic cell death and solar UV-induced genotoxic effects.

When skin is exposed to sunlight, epidermal cells are in the front line and their DNA can be considered as the primary target for the initiation of carcinogenesis. A simple and sensitive test for detecting DNA damage is thus required. Single cell gel electrophoresis, i.e. the comet assay, is very suitable in this respect because it concerns a variety of lesions convertable into strand breaks: this is the case for alkali-labile sites but also for all kinds of damage subject to excision repair. Here, we have shown that when human keratinocytes are exposed to solar UV, comets are easily observed one hour after the irradiation. Moreover, recent improvements in the method enable the detection of the fast repaired single stand breaks induced by UVA [8]. Even in these conditions, we observed that Mexoryl^® SX protects cellular DNA (not shown).

In our experiments, the level of protection was high enough to prevent almost totally the formation of comets. However, in situations of partial protection only, quantification of the comets based on measurement of their tail moments as already described [35] would be very useful to compare sunscreens with regard to their efficacy. Work is in progress in our laboratory on such a quantification.

Induction of DNA damage by solar UV triggers many cellular responses. Among these, p53 activation appears to be a critical event. This process is already well documented. It has been shown that p53 protein accumulation results in growth arrest via the induction of genes such as p21^WAF1/CIP1 involved in cell cycle control. During cell cycle arrest, DNA repair can occur. However, if there is too much damage, the cell undergoes apoptosis. The importance of such regulation is clear if one considers that mutations in the p53 gene are associated with many types of cancer and, particularly strongly with squamous cell carcinoma. Thus, it is logical to use the p53 protein as a marker for "genotoxic stress" in addition to the detection of DNA damage. While the comet assay demonstrates that lesions are induced, p53 accumulation suggests that the level of DNA damage is high enough to disturb cellular metabolism. As shown recently [36], p53 activation could start when transcribed genes need to be repaired. From this, it is conceivable that an "acceptable" level of lesions could remain unrepaired before cell division or could be transmitted to the next cell generation. Beyond this level, the cell enters a state of emergency in which p53 plays a major role leading to the apoptotic pathway. The comet assay and the measure of p53 activation are thus quite complementary in order to evaluate photoprotection in distinct genotoxic situations.

The melanocyte is a very attractive model in our strategy because of its sensitivity to various environmental insults and, particularly, to photo-oxidative stress. Their specific response, melanogenesis, is easily quantified in terms of tyrosinase activity and melanin content. However, for normal human melanocytes in monoculture, we used particular experimental conditions in order to reproducibly determine the induction of pigmentation by solar UV. In particular, it was necessary to maintain the response to alphaMSH stimulation, otherwise a down regulation of tyrosinase activity takes place after a few days of post-irradiation incubation [37]. In our experiments, an increase of dopa oxidase activity and of melanin content was measured up to 5 days after the exposure. This is in good agreement with the kinetics of the tanning process.

Among the different assay systems described above the melanocytic test system links the evaluation of phototoxicity and photoprotection to a more general cellular response including membrane signalling pathways as well as nuclear events. In fact, this latter approach is not unrelated to genotoxicity since melanogenesis has been reported to be stimulated by lesions in DNA [4, 30]. Also noteworthy is the fact that the melanocytes irradiated in contact with the sunscreen Mexoryl^® SX keep their basal level of tyrosinase activity. From this, we conclude that in these experimental conditions, the melanocytes keep their capacity to perform melanogenesis but are no longer submitted to the photo-oxidative stress of solar UV.

Concluding remarks

We have developed an in vitro strategy in order to obtain a first set of data covering the genotoxic effects of solar UV and some biological consequences. Obviously, complementary experiments on more complex models such as reconstructed skin, skin explants or skin in vivo are necessary to perfectly evaluate the benefits of a given sunscreen in terms of photoprotection.

In order to allow extrapolability of the test results, we define the following criteria as being essential:

1. The use of the whole solar UV spectrum (290-400 nm) and a spectral power distribution of the solar simulated radiation as realistic as possible, especially for UVB.

2. The use of normal human skin cells, and particularly normal human keratinocytes and melanocytes that are specific targets for solar UV and are known to be involved in photocarcinogenesis.

3. Cell culture conditions defined to preserve as much as possible the characteristics of each cell type: use of cells from neonates in order to lessen the influence of the donor's history, limitation of the number of passages and use of defined culture media to keep cell-specific responses (melanocytes and alphaMSH).

4. Test of biological endpoints known as markers of UV exposures in vivo: DNA repair, p53 expression, melanogenesis stimulation.

Considering the fact that a monolayer of cells in culture is a very simplified model of the in vivo situation, we have defined a strategy including complementary approaches with increased complexity in order to circumscribe phototoxicity, photodamage and photoprotection. According to this, a sunscreen is evaluated by a sequential battery of tests that are somewhat linked to each other allowing confirmation of the presence (or absence) of DNA damage and specific biological consequences and vice versa. Obviously, the body of results obtained is more relevant than that of a single test alone. For example, data obtained by the comet assay are altogether more convincing when concomitantly analysed with photomutagenesis in yeast or with p53 status.

Finally, all these in vitro assays contribute to ensure the efficacy and the safety of products as demonstrated here for the sunscreen Mexoryl^® SX.

Abbreviations: BCC: basal cell carcinoma; SCC: squamous cell carcinoma; SSUV: simulated solar ultraviolet; MTT: (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide); XP: xeroderma pigmentosum

CONCLUSION

Acknowledgements

Thanks are due to Dr Averbeck (Institut Curie, Paris) for providing yeast strains, for helpful discussions and critical reading of the manuscript, to M. Maréchal for her help in the yeast experiments and to J. Caradec, N. Lequesne C. Olivry and A. Reinhardt for their help in preparing the manuscript.

REFERENCES

1. English DR, Armstrong BK, Kricker A, Fleming C. Sunlight and cancer. Cancer Causes and Control 1997; 8: 271-83.

2. Kraemer KH, Levy DD, Parris CN. Xeroderma pigmentosum and related disorders: examining the linkage between defective DNA repair and cancer. J Invest Dermatol 1994; 103: 96-101.

3. Nishigori C, Yarosh DB, Ullrich SE, Vink SE, Bucana CD, Roza L, Kripke M. Evidence that DNA damage triggers interleukin-10 cytokine production UV-irradiated murine keratinocytes. Proc Natl Acad Sci USA 1996; 93: 10354-9.

4. Eller MS, Ostrom K, Gilchrest BA. DNA damage enhances melanogenesis. Proc Natl Acad Sci USA 1996; 93: 1087-92.

5. Healy E, Reynolds NJ, Smith MD, Campbell C, Farr PM, Rees JL. Dissociation of erythema and p53 protein expression in human skin following UVB irradiation and induction of p53 protein and mRNA following application of skin irritants. J Invest Dermatol 1994; 103: 493-9.

6. Friedberg EC. Eukaryotic DNA repair: glimpses through the yeast Saccharomyces cerevisiae. BioEssays 1991; 13: 295-302.

7. Arlett CE, Lowe JE, Harcourt SA, Waugh APW, Cole J, Roza L, Diffey BL, Mori T, Nikaido O, Green MHL. Hypersensitivity of human lymphocytes to UVB and solar irradiation. Cancer Res 1993; 53: 609-14.

8. Alapetite C, Moustacchi E, Wachter T, Sage E. Use of the alkaline comet assay to detect DNA repair deficiencies in human fibroblasts exposed to UVC, UVB, UVA and gamma-rays. Int J Radiat Biol 1996; 69: 359-69.

9. Hall PA, McKee PH, Du P, Menage H, Dover R, Lane DP. High levels of p53 protein in UV-irradiated normal human skin. Oncogene 1993; 8: 203-7.

10. Yamaizumi M, Sugano T. UV-induced nuclear accumulation of p53 is evoked through DNA damage of actively transcribed genes independent of the cell cycle. Oncogene 1994; 9: 2775-84.

11. Ley RD, Fourtanier A. Sunscreen protection against ultraviolet radiation-induced pyrimidine dimers in mouse epidermal DNA. Photochem Photobiol 1997; 65: 1007-11.

12. Averbeck D, Averbeck S, Dubertret L, Young AR, Morliere P. Genotoxicity of bergapten and bergamot oil in Saccharomyces Cerevisiae. J Photochem Photobiol B Biol 1990; 7: 209-29.

13. Shiota S, Nakayama H. UV endonuclease of microloccus luteus, a cyclobutane pyrimidine dimer-DNA glycosylase/abasic lyase: cloning and characterization of the gene. Proc Natl Acad Sci USA 1997; 94: 593-8.

14. Boyce ST, Ham RG. Calcium-regulated differentiation of normal human keratinocytes in chemically-defined clonal culture and serum-free serial culture. J Invest Dermatol 1983; 81: 33s-40s.

15. Suzuki I, Cone RD, Im S, Nordlund J, Abdel-malek ZA. Binding of melanotropic hormones to the melanocortin receptor MC1R on human melanocytes stimulates proliferation and melanogenesis. Endocrinology 1996; 137: 1627-33.

16. Zimmermann FK, Kern R, Rasenberger H. A yeast strain for simultaneous detection of induced mitotic crossing over, mitotic gene conversion and reverse mutation. Mutation Res 1975; 28: 381-8.

17. Singh NP. Microgel electrophoresis of DNA from individual cells. Principles and methodology. Technologies for detection of DNA damage and mutations, Pfeiffer G, ed. Plenum Press, New York, 1996; 3-24.

18. Lin Y, Benchimol S. Cytokines inhibit p53 mediated apoptosis but not p53-mediated G1 arrest. Mol Cell Biol 1995; 15: 6045-54.

19. Lee TH, Lee MS, Lu MY. Effects of alphaMSH on melanogenesis and tyrosinase of B16 melanoma. Endocrinology 1972; 91: 1180-8.

20. Takahashi H, Parsons PG. Rapid and reversible inhibition of tyrosinase activity by glucosidase inhibitors in human melanoma cells. J Invest Dermatol 1992; 98: 481-7.

21. Freeman SE, Gange RW, Sutherland JC, Matzinger EA, Sutherland BM. Production of pyrimidine dimers in DNA of human skin exposed in situ to UVA radiation. J Invest Dermatol 1987; 88: 430-3.

22. Brash DE, Rudolph JA, Simon JA, Lin A. A role for sunlight in skin cancer: UV-induced p53 mutations in squamous cell carcinoma. Proc Natl Acad Sci USA 1991; 88: 10124-8.

23. Daya-Grosjean L, Dumaz N, Sarasin A. The specificity of p53 mutation spectra in sunlight induced human cancer. J Photochem Photobiol 1995; 28: 115-24.

24. Mitchell DL, Nguyen TD, Cleaver JE. Nonrandom induction of pyrimidine-pyrimidone (6-4) photoproducts in ultraviolet-irradiated human chromatin. J Biol Chem 1990; 265: 5353-6.

25. Collins AR, Ai-Guo M, Duthie SJ. The kinetics of repair of oxidative DNA damage (strand breaks and oxydised pyrimidines) in human cells. Mutation Res 1995; 336: 69-77.

26. Lane DP. p53 guardian of the genome. Nature 1992; 358: 15-6.

27. Ziegler A, Jonason AS, Leffell DJ, Brash DE. Sunburn and p53 in the onset of skin cancer. Nature 1994; 372: 773-6.

28. Hearing VJ, Tsukamoto K. Enzymatic control of pigmentation in mammals. FASEB J 1991; 5: 2902-9.

29. Abdel-Malek ZA, Swope V, Smalara D, Babcock G, Dawes S, Nordlung J. Analysis of the UV-induced melanogenesis and growth arrest of human melanocytes. Pigment Cell Res 1994; 7: 326-32.

30. Gilchrest BA, Zhai S, Eller MS, Yarosh DB, Yaar M. Treatment of human melanocytes and S91 melanoma cells with the DNA repair enzyme T4 endonuclease V enhances melanogenesis after ultraviolet irradiation. J Invest Dermatol 1993; 101: 666-72.

31. Roberts LK, Beasley DG, Learn DB, Giddens LD, Beard J, Stanfield JW. Ultraviolet spectral energy differences affect the ability of sunscreen lotions to prevent ultraviolet-radiation-induced immunosupression. Photochem Photobiol 1996; 63: 874-84.

32. Sage E, Lamolet B, Brulay E, Moustacchi E, Chateauneuf A, Drobetsky E. Mutagenic specificity of solar UV light in nucleotide excision repair deficient rodent cells. Proc Natl Acad Sci USA 1996; 93: 176-80.

33. Parsons PG, Hayward IP. Inhibition of DNA repair synthesis by sunlight. Photochem Photobiol 1985; 42: 287-93.

34. Epe B, Pflaum M, Boiteux S. DNA damage induced by photosensitizers in cellular and cell-free systems. Mutation Res 1993; 299: 135-45.

35. Ashby J, Tinwell H, Lefevre PA, Browne MA. The single cell gel electrophoresis assay for induced DNA damage (comet assay): measurement of tail length and moment. Mutagenesis 1995; 10: 85-90.

36. Sarasin A, Dumaz N, Marionnet C, Giglia G, Ehrhart C, Stary A, Mezzina M, Daya-Grosjean L. A molecular signal for the UV-induced p53 protein stabilization. Mutation Res 1997; 379: S38 (abst.).

37. Im S, Moro O, Peng F, Medran ED, Cornelius J, Babcock G, Nordlund JJ, Abdel-Malek ZA. Activation of the cyclic AMP pathway by alpha-melanotropin mediates the response of human melanocytes to ultraviolet B radiation. Cancer Res 1998; 58: 47-54.