• · 放疗毒性与对DNA损伤异常转录反...
• · 环境内分泌干扰物的遗传毒性研究进...
• · 微核试验在中国的应用、发展与展望
• · Use of the alkal...
• · 遗传毒性评价系列文章
• · Compromised DNA ...
• · 光化学遗传毒性：原理和试验方法
• · Application of E...
• · 体外微核试验：从过去到未来
• · DNA-蛋白质交联的诱导、修复和...

光化学遗传毒性：原理和试验方法

a Toxicology, Bayer AG, D-42096, Wuppertal, Germany
b RCC-CCR, D-64380, Ro?dorf, Germany
c University of Mainz, D-55099, Mainz, Germany
d F. Hoffmann-La Roche, CH-4070, Basel, Switzerland
e University of Mainz, D-55131, Mainz, Germany
f Novartis Pharma AG, CH-4002, Basel, Switzerland
g Beiersdorf AG, D-20245, Hamburg, Germany
h Merck KGaA, D-64271, Darmstadt, Germany

Abstract
    In recent years, assessing the photogenotoxic potential of a compound became an issue for certain drugs and cosmetical products. Therefore, existing methods performed according to international guidelines (e.g. OECD guidelines) were adapted to the use of concurrent UV-visible (UV-Vis) light irradiation for the assessment of photomutagenicity/photogenotoxicity. In this review, photobiological bases of the processes occurring in the cell after irradiation with UV- and/or visible (vis)-light as well as a compilation of testing methods is presented. Methods comprise cell free investigations on naked DNA and in vitro methods, such as the photo-Ames test, the photo-HPRT/photo-mouse lymphoma assay (MLA), the photo-micronucleus test (MNT), the photo-chromosomal aberration test (CA) and the photo-Comet assay. A compilation of the currently available international literature of compounds tested on photogenotoxicity is given for each method. The state of the art of photogenotoxicity testing as well as the rational for testing are outlined in relation to the recommendations reached in expert working groups at different international meetings and to regulatory guidance papers. Finally, photogenotoxicity testing as predictor of photocarcinogenicity and in the light of risk assessment is discussed.
    Author Keywords: Photogenotoxicity; Photo-Ames test; Photo-HPRT assay; Photo-mouse lymphoma assay; Photo-micronucleus test; Photo-chromosomal aberration test; Photo-Comet assay; Photocarcinogenicity
    Abbreviations: 6-4PP, 6-4-photoproduct; AP, apurinic; BER, base excision repair; CA, chromosomal aberration test; CPD, cyclobutane pyrimidine dimer; GGR, global genomic repair; MLA, mouse lymphoma assay; MNT, micronucleus test; NER, nucleotide excision repair; ROS, radical oxygen species; SSB, single-strand breaks; TCR, transcription-coupled repair; UV, ultraviolet light; UVA, UVA-irradiation (320–400 nm); UVB, UVB-irradiation (280–320 nm); UVC, UVC-irradiation (200–280 nm); vis light, visible light (400–745 nm)
 
1. Introduction

    The property of a compound to induce genotoxic effects when irradiated with UV- and/or visible (vis)-light is called photochemical genotoxicity, in the following designated "photogenotoxicity". This effect requires that the compound absorbs light. The UV part of the electromagnetic spectrum of sunlight can be divided into a short-wave (200–280 nm; UVC), middle-wave (280–320 nm; UVB) and long-wave range (320–400 nm; UVA). Since UVC-irradiation is eliminated by the ozone layer, basically only UVA and UVB can affect life on earth. Hence, photogenotoxicity testing and hazard identification focus on visible- and/or UVA- and/or UVB-irradiation. Of course, this does not contradict the scientific interest in investigations using UVC.
    Based on precedence with photoabsorbing pharmaceuticals and cosmetic products over the past decade it became evident that selected compounds of these groups should be tested for photogenotoxicity. In 1990, a testing guideline was published by the Scientific Committee for Cosmetology of the Commission of the European Communities. In this guideline, testing of sunscreens for photogenotoxicity was requested [1]. Sunscreens per se are designed to absorb light due to their use as UV protecting agent. However, it has to be assured that in consequence of the high-energy uptake of these compounds no side effects like the damage of DNA are induced. In the pharmaceutic area light induced adverse effects including photogenotoxicity were reported for the antibacterial acting fluoroquinolones in the early 1990s [2]. Meanwhile, the need for a general photosafety evaluation including photogenotoxicity testing of drugs which absorb light and reach the skin in relevant concentrations either by systemic or by cutaneous application is considered by the European agencies and the FDA [3 and 4].
    One aim of this review is to compile available data in the open literature on the different test systems for photogenotoxicity and to discuss the current status of testing strategies also with respect to photocarcinogenicity. The focus will be on in vitro genotoxicity test methods currently used as standard tests in regulatory testing, which were adapted for testing chemicals in the presence of UV- and/or vis-light. Therefore, tests with yeast or the sister chromatide exchange (SCE) assay are not considered. In 1999, an international workshop on genotoxicity test procedures (IWGTP) was held in Washington, DC, where an expert group on photogenotoxicity testing was setting up first rules on the technical performance of photochemical testing [5]. The areas, which were already covered by the results of this working group, will only be touched briefly as long as no additional viewpoints can be added.
A second aim is to describe the biological processes known so far, which are initiated in irradiated cells by UV alone or during the concurrent presence of a light-absorbing chemical in a comprehensive manner. Studies with cell lines as well as with naked DNA are taken as examples.

2. Photobiological basis

2.1. DNA modifications induced by UV and visible light in the absence of exogenous compounds

    Solar radiation causes genotoxic effects basically by two types of mechanism, i.e. directly by photo-excitation of DNA and indirectly by excitation of other molecules, which can be endogenous cellular chromophores such as porphyrins and flavins or xenobiotic compounds. As a rule, the direct mechanism is much more efficient. The extent of DNA photo-excitation and all resulting genotoxic effects parallel the DNA absorption, which has a maximum at 260 nm and decreases exponentially in the UVB-range (280–320 nm) and UVA-range (320–400 nm). Since the direct photo-excitation of DNA gives rise to photoproducts (DNA modifications) with very high efficiency (quantum yield), a significant generation of direct photoproducts takes place even in the UVA range, although the relative absorption of DNA and thus the yield of products per dose unit induced at 300 nm is ~100-fold and at 360 nm is ~100,000-fold lower than at 260 nm [6].
    An important consequence of the steep decline by several orders of magnitude of the DNA absorption in the range between 290 and 360 nm is that the extent of direct DNA damage and genotoxic effects induced in this range depends very much on the exact emission spectrum of the light source. This is the reason for the strong biological impact of small changes of the solar spectrum caused by an ozone layer depletion. Similarly, very weak emissions of commercial UVA lamps in the UVB range, which are hardly visible on a linear scale, are often responsible for most of the DNA damage induced by these lamps.
    The product spectrum generated by direct DNA excitation at ~260 nm consists mostly of cyclobutane pyrimidine dimers (CPDs) (formed at TpT, CpC, CpT and TpC sequences) and so-called 6-4-photoproducts (6-4PPs) formed mostly at TpC and CpC sequences. Other photoproducts such as thymine glycols, cytosine hydrates, 8-7,8-dihydro-8-oxo- guanine (8-oxoG), formamidopyrimidines, single-strand breaks (SSB) and DNA–protein cross-links are generated in much lower amounts and probably do not significantly contribute to most of the biological consequences associated with the direct DNA damage. Some chemical structures of the photoproducts are shown in Fig. 1.

    Due to the high efficiency of the damage by direct excitation of DNA, all indirect mechanisms (both endogenous ones and those mediated by xenobiotics) are, on a theoretical basis, most probable biologically relevant only if the DNA absorption is negligible, i.e. in the UVA and visible range of the spectrum. Experimental evidence that indirect mechanisms are relevant in cells irradiated in that spectral range in the absence of xenobiotics comes from the "action spectra" (wavelength dependence) for the induction of oxidative DNA modification, which deviate from the DNA absorption spectrum at wavelengths >330 nm [7]. The indirect DNA damage induced at those wavelengths appears to be mediated by cellular chromophores such as flavins and porphyrins, which after photo-excitation form a relatively stable and thus chemically reactive triplet state and which therefore can act as "photosensitizers" (see below). In cultured cells exposed to natural sunlight, the yield of the indirectly induced oxidative guanine modifications was observed to be approx. 10% of that of the pyrimidine dimers. A major feature of the endogenous indirect DNA damage by vis light is its saturation at relatively low doses, most probably due to photodegradation (photobleaching) of the responsible chromophore.
 
2.2. DNA damage induced by UV-visible light in the presence of exogenous compounds

    In most cases, the absorption of UV or vis by organic compounds generates excited states that are too short-lived to react chemically (singlet states). In some cases, however, more stable excited triplet states are formed by "intersystem crossing". These can give rise to DNA damage either directly or indirectly (Fig. 2). In the former case, a one-electron or hydrogen abstraction is very frequent, i.e. an oxidation of DNA (type-I reaction). The resulting DNA damage appears to consist mostly of oxidative guanine modification including 8-oxoG. Sites of base loss and DNA SSB are generated in much lower yields (approx. 10% of that of the base modifications). Another direct type of DNA damage caused by photo-excited species is covalent binding, as observed in the case of furocoumarins. Lastly, an excited molecule can also transfer the excitation energy to DNA and thus give rise to pyrimidine dimer formation as observed upon direct DNA excitation. However, it is not clear whether this mechanism is relevant in cells.

    Compounds that after photo-excitation react by type-I or type-II reaction are called photosensitizers. Typically, they act as catalysts, i.e. they are ultimately re-generated in the ground state. If this is not the case, "photobleaching" (destruction of the absorbing chromophore) is generally observed during irradiation.
Several organic molecules decompose spontaneously after photo-excitation. Examples are several halogenated aromatic compounds, which yield halogen radicals and organic radicals, both of which potentially can damage DNA, e.g. by adduct formation.
    It is also possible that DNA damage can arise as consequence of damage to other cellular targets such as mitochondria or lipids, in particular because several photosensitizers accumulate in these compartments and because singlet oxygen reacts very efficiently with unsaturated fatty acids. Putative mechanisms are DNA adduct formation by, -unsaturated compounds generated from lipids and an increase of the endogenous production of reactive oxygen species (ROS).
While DNA can be regarded as the primary cellular target for UVB radiation, it is only a target among many others for most photoreactive compounds with the exception of some intercalators. Therefore, cytotoxicity frequently prevents the detection of the photogenotoxic effects of these compounds, although the intrinsic mutagenic potential of the DNA modifications generated is probably as high as that of pyrimidine dimers.

2.3. Repair of DNA lesions induced by UV-visible light

    There is no doubt that DNA repair is the major cellular defense mechanism against the deleterious effects of UV light. Repair mechanisms are closely coupled to cell cycle checkpoints, transcriptional control and DNA replication. UV light induces both, bulky lesions as well as oxidative damage in DNA. The relative amount of oxidative damage increases with increasing wavelength and thus becomes especially biologically relevant in the UVA/UVB range due to the formation of free radicals inside the cell. Consequently, notably in the UVA range, both nucleotide excision repair (NER) and base excision repair (BER) are involved in defense against UV-induced DNA damage.
    Bulky DNA adducts such as 6-4PPs or CPDs are repaired by NER in which about 30 proteins are involved. Some NER defects are viable on cellular and organismic level, resulting in the well-known UV-hypersensitive phenotypes [9]. For the excision repair cross-complementing (ERCC) group disorders only mouse models are available; obviously human individuals do not survive the defect. The fact that hypersensitivity to sunlight is related to increased cancer incidence (with the exception of Cockayne‘s syndrome) strongly points to the importance of DNA repair in health protection.
    NER of light-induced lesions can be subdivided into lesion recognition, excision and gap filling. Two pathways are involved: global genomic repair (GGR) and transcription-coupled repair (TCR) (for recent review, see [10]). GGR is transcription-independent and removes lesions from the entire genome. 6-4PPs, which distort the DNA more than CPDs, are removed more rapidly than CPDs by GGR. Also, they are removed more efficiently than CPDs from the transcribed strand of expressed genes by TCR [11]. TCR removes RNA polymerase blocking lesions from the transcribed strands of active genes [12]. It is coupled to transcription by RNA polymerase II (RNAP II). Thus, RNAP II can be considered to be a sensor for bulky UV-induced DNA damage. Another important complex recruited to the site of DNA damage is the transcription factor TFIIH which possesses several enzymatic functions such as ATPase and DNA helicase [13]. DNA unwinding by TFIIH is necessary for creating an open complex around a lesion. TFIIH may also play a role in distinguishing between the damaged and the undamaged DNA strand when a lesion leads to blockage of TFIIH translocation [13]. After damage recognition and the formation of an open complex, excision of the lesion is carried out by dual incision at defined positions flanking the DNA damage. The 3‘ incision is performed by XPG, the 5‘ incision by the XPF–ERCC1-complex. After incision, the 5‘ to 3‘ exonuclease activity of XPG removes the DNA fragment containing the lesion. The arising gap is filled in by DNA polymerase and, and DNA ligase I.
    Since UVA and UVB light produce free radicals inducing oxidative DNA lesions such as 8-oxoG [14], BER is also involved in repair of UV lesions. This repair is quite important since 8-oxoG is a critical pro-mutagenic lesion because of mispairing with adenine. Removal from DNA of 8-oxoG is performed by 8-oxo-guanine-DNAglycosylase-1 (OGG1) [15]. OGG1 displays weak apurinic lyase activity. Therefore, removal of 8-oxoG by OGG1 results in both apurinic (AP) sites and SSBs. AP sites are incised by AP-endonuclease resulting in a 5‘-deoxyribosephosphate (dRP) and a free 3‘-OH. In human cells, removal of 8-oxoG is performed mainly via short patch repair; only 25% of the lesions are repaired by long patch BER [16]. In the short patch pathway, DNA polymerase  (Pol ) is essentially involved. Pol  exhibits lyase-activity and is thereby able to remove 5‘-terminal deoxyribose phosphate residues from incised AP sites followed by insertion of a single nucleotide [17]. The ligation step is performed by DNA ligase III–XRCC1 complex or by DNA ligase I [18]. During long patch BER, Pol  plays a role by initiating DNA synthesis which is completed by Polor Pol resynthesizing 2–6 nucleotides. Ligation is mediated by DNA ligase I (for review, see [10]).
 
2.4. Cellular responses and signal transduction due to photo-induced DNA damage

    This expanding field can only briefly be addressed here to indicate that inducible functions might be modifiers of UV effects on living cells. Besides the effect of UV on DNA itself, there is a plethora of DNA damage-independent effects ranging from cell cycle disturbances, induction of signal transduction pathways, receptor activation, induction of repair activity and induction of apoptotic pathways. UVA, UVB and UVC light induce the transcription of several genes the products of which may exert, at least in part, protective function. They include immediate-early response genes, which are mainly transcription factors, and late-response genes. Transcription factors such as c-Fos, c-Jun, JunB, JunD, Elk-1 and ATF-2 are activated by kinases of the growth factor signal transduction pathway. Activation of this pathway by UV occurs within minutes upon exposure and is triggered by activation of surface receptors such as EGFR [19]. UVA, UVB and UVC are capable of activating the EGF receptor by autophosphorylation which is likely to be due to inhibition of a phosphatase due to oxygen radical formation [20]. The activated EGF-R initiates a cascade of phosphorylation events involving sequentially the proteins Ras/Rho, Raf which targets ERKs or MEKK which targets Jun kinase (JNK) and p38 kinase. These kinases phosphorylate Fos, Jun, Elk-1 and ATF-2 with different specificity thus activating to different levels the dimerized transcription factor AP-1 containing either one of these proteins. The different populations of AP-1 target different sets of genes whose activation controls growth and survival upon irradiation. The importance of this so-called "UV response" is exemplified by the fact that cells knockout for c-Fos or c-Jun are hypersensitive to UV light [21 and 22]. Induction of AP-1 is also related to stimulation of collagenase and stromylesin. Both might be involved in ultraviolet light-induced cutaneous aging [23]. Inhibition of phosphatases seems to be a general process involved in the UV-activation of receptors [24]. Another transcription factor activated within a few hours after UV-irradiation is NF- B which is involved in regulation of a variety of stress response genes including anti-apoptotic functions [25]. UV light is also able to trigger programmed cell death (apoptosis), activating the apoptotic Fas/CD95/Apo-1 death receptor by aggregation independent from binding of the natural ligand [26]. This leads to activation of caspase-8 which in turn activates the effector caspase-3 and caspase-activated DNAse (CAD) degrading nuclear DNA. Data at hand suggest the participation of nuclear as well as cytoplasmic signaling in receptor activation [27]. A link to DNA damage is given by the fact of p53-dependent upregulation of Fas receptor [28]. Since cells defective in NER are highly sensitive to UV-light induced apoptosis, non-repaired DNA damage is clearly involved in triggering the apoptotic process. Because many of the CHO mutants utilized for these studies are p53 mutated, UV-induced and DNA damage-triggered apoptosis is clearly not p53-dependent but involves mitochondrial damage at least in fibroblasts [29]. Overall, the available data indicates that UV light triggers apoptosis cell type specific both via receptor activation and the mitochondrial damage pathway. Only little is known about the interaction of light activated phototoxic compounds and signaling elements, but it is clear that irradiation might cause severe side effects due to activation of pro- and anti-apoptotic functions which could be even superimposed by a phototoxic compound.
 
3. Phototoxicity

    Phototoxic compounds are capable of inducing adverse effects in skin or eyes upon exposure to light. The initial photoreactive compound may reach the target tissue by dermal penetration after topical application or via the circulatory system after systemic uptake. Phototoxicity due to therapeutic or occasional contact with photoreactive compounds thereby represents an undesired side-effect of various drugs and chemicals. Consequently, their use in everyday products such as cosmetics or foodstuff should be avoided. Some chemicals which have been shown to be phototoxic were recently summarized [30]. The common feature of phototoxic compounds is their ability to absorb light energy within the wavelength range covered by sunlight, thus generating an excited intermediate. The underlying mechanisms leading to phototoxic insults such as direct energy transfer or generation of reactive oxygen species or eventually generation of a stable photo-metabolite are basically the same as for the induction of photogenotoxic effects (see Section 2.2).
Historically, phototoxicity testing is carried out using laboratory animals or human volunteers where phototoxotic effects can be assessed in terms of clinical symptoms such as erythema, edema or inflammation [31]. It should be noted however, that acute phototoxic reactions can resemble severe sunburn with blistered skin. Therefore, and due to ethical and legal considerations, in vitro test methods were proposed using red blood cells [32] and other mammalian cells including lymphocytes, skin cells and, more recently, reconstructed skin models for photosafety testing.
    In this context, a modification of the neutral red uptake cytotoxicity assay [33] employing light exposure of murine BALB/c 3T3 fibroblast cultures was developed and demonstrated to be a valid model for the identification of phototoxic chemicals in vitro [34, 35 and 36]. This method, referred to as 3T3 Neutral Red Uptake (3T3 NRU) phototoxicity test is based on a comparative evaluation of the cytotoxicity of a test compound in the presence and absence of irradiation with simulated sunlight including UVB, UVA and vis light. Cytotoxicity is assessed as concentration-dependent reduction in cellular uptake of the supravital dye neutral red 24 h after the combined exposure to chemical and light. From these data, concentration–response curves are established. The corresponding concentrations which lead to a 50% inhibition in cellular viability (IC50 values) with and without UV exposure are interpolated. Finally, the ratio of both IC50 values, referred to as the photo irritation factor (PIF), is calculated as a relative descriptor of phototoxicity. The PIF for chlorpromazine, which is recommended as concurrent positive control, is in the range of 6–10. This compound forms a phototoxic intermediate by the absorption of light and was shown to be an efficient inducer of phototoxic as well as photogenotoxic effects.
    A corresponding test guideline for the 3T3 NRU phototoxicity test based on a protocol validated in 1998 by ECVAM and COLIPA was drafted and submitted to the OECD and to the European Commission [37]. In 2000, the EU accepted this test system as validated and transferred it to a Council Directive making it mandatory for testing chemicals throughout the EU [38]. This test guideline now offers the possibility to conduct phototoxicity tests following a standardized protocol which finds acceptance by regulatory authorities [39]. It should be noted, that this test is neither designed to allow an assessment of phototoxic potency nor is it able to predict other adverse biological effects of photoreactive chemicals such as photogenotoxicity or photoallergenicity. However, in most cases compounds showing photogenotoxic or photoallergic effects generally turn out to be positive in the 3T3 NRU phototoxicity test as well, since the underlying photochemical mechanisms are similar [37].
    Further studies are in progress using human skin models instead of cell monolayer cultures thereby allowing a topical administration of test materials [40, 41, 42 and 43]. Accordingly, a major advantage is the possibility of a direct application of final products with suspected photoreactive ingredients and testing water-insoluble materials without the addition of solvents.
    As, at the present, cellular viability is the endpoint of choice in phototoxicity testing in vitro, the intrinsic phototoxic effects of UV alone, i.e. without the involvement of exogenous compounds, have to be considered too. In this respect, UVA and UVB dramatically differ in their effects on cellular viability as they employ different endogenous chromophores.
UVA is readily absorbed by cellular chromophores such as NAD coenzymes, riboflavin or melanin acting as endogenous photosensitizers. Subsequently, reactive intermediates among which reactive oxygen species play an important role are generated and ultimately lead to an oxidative damage of cellular macromolecules such as lipids, proteins and DNA. Accordingly, exposure of cells to UVA immediately results in a dose-dependent decrease in viability, predominantly due to damaged lipid membranes and proteins.
In contrast, for UVB the main cellular target is DNA itself, which efficiently absorbs UVB at wavelengths below 320 nm. So far, no further endogenous chromophores have been identified that are as nearly as efficient in absorbing UVB as DNA. Therefore, it can be assumed that the phototoxic effects of UVB alone are mainly mediated by its direct DNA damaging action. Accordingly, the onset of a significant reduction in cellular viability is delayed for at least several hours if usual short-term cytotoxicity assays such as neutral red uptake or MTT test are employed. In these assays, initial DNA lesions induced by UVB such as pyrimidine dimers or 6-4PPs remain obscure until secondary processes leading to an impaired survival of the irradiated cells become more and more prominent over time. In particular, activated p53 acts as a potent cell death mediator gene which in conjunction with cell cycle arrest accounts for most of the observed time-delayed decrease in viable cell numbers after UVB exposure. As a consequence, rather high UVB doses could be applied in phototoxicity assays if cellular viability is assessed immediately after exposure and before apoptotic cell death takes place. However, phototoxicity test protocols generally demand for UV exposure with a relatively low amount of UVB, thus avoiding interference with UVB-induced apoptosis.
 
4. In vitro testing of photogenotoxicity

4.1. Irradiation conditions

    In the paper published by the IWGTP working group "photochemical genotoxicity" [5] proposed irradiation conditions and influencing factors are described in detail. Therefore, only the main points will be strengthened here.
    As irradiation source a solar simulator should be used. In recent photogenotoxicity studies, mostly a xenon arc lamp was used with appropriate filters cutting off the UVC part. The emitted UVA:UVB ratio varies around 16:1–20:1. Measurement of the actual irradiation dose with UVA and UVB dosimeters is obligatory. In addition, a UV dose–response curve of the cells or bacteria used in the absence of the compound should be available for each test system. As a first approach it seems reasonable to irradiate the cells directly after application of the test compound. In an additional experiment, it might be necessary to preincubate the cells with the compound in the dark before irradiation in order to avoid false negative results due to limited compound uptake into the cell.
    There is no UV dose that can be generally recommended for photogenotoxicity testing. It has to be adapted in each case to the cell type (or bacteria) and the test system used. However, the UV-irradiation alone should induce a slight but reproducible photogenotoxic and/or phototoxic effect. In the case of very phototoxic compounds it might be necessary to evaluate a second, lower UV dose in order to be able to assess higher compound concentrations. Otherwise, a photogenotoxic effect might be missed.
The different test systems reveal different sensitivities to the UVB part of the irradiation probably dependent on the genetic endpoint measured. For example the photo-chromosomal aberration test (CA) and the photo-micronucleus test in vitro (MNT) are very sensitive to the phototoxic and photogenotoxic effect of UVB-irradiation whereas for the photo-HPRT assay or the photo-Comet assay higher UVB doses can be used before inducing measurable genotoxic effects or reaching the phototoxicity limit. In the first two assay systems, the testing window for addition of a compound might be very narrow due to strong phototoxic and/or photogenotoxic properties of the UVB alone. Therefore, it might be necessary to reduce the UVB amount of the irradiation depending on the test system used.
    However, discussions are going on whether UVB-irradiation at all should be used in photogenotoxicity testing. Though it is part of the sunlight reaching the earth and compounds absorb in this range, no increase in photogenotoxicity might be seen for the following reasons. When UV-irradiation is applied in photogenotoxicity tests, DNA damage by direct photo-excitation of DNA will unavoidably be induced in parallel with the DNA damage mediated by the photo-excitation of the test compound and cause a background effect. For both, the test compound-mediated and the direct/background process, the yields DNAmod of DNA modification per irradiation dose unit will depend on the quantum yield  of the damaging reaction, the absorption coefficient of the absorbing chromophor (DNA or test compound) at the irradiation wavelength (or wavelength range) and the concentration c of the absorbing chromophor:

DNAmod~ DNA×DNA×cDNA (absorption by DNA)

DNAmod~ ps×ps×cps (absorption by photosensitizer)

    In principle, the best discrimination between the drug-mediated genotoxicity (to be determined) and the undesirable genotoxicity caused by direct DNA absorption can be achieved by selecting an excitation wavelength at which the ratio of the absorption coefficients of the drug and the DNA is at its maximum. Since the absorption of DNA has been determined up to relatively long wavelengths, the calculation can be easily carried out provided that the absorption spectrum of the test compound is known as well. For theoretical reasons, it appears extremely unlikely that an optimum discrimination is obtained in the UVB range of the spectrum:
(1)For compounds which have an absorption maximum at a longer wavelength than DNA (260 nm), the decline of the absorbance with increasing wavelength in the UVB range will be less steep than that of DNA so that an optimum discrimination is not in the UVB.
(2)For compounds with an absorption maximum similar to that of DNA, the yield of directly generated pyrimidine dimers will always be higher than the yield of compound-mediated DNA modifications, since (i) the concentration c of the test compound is at the most similar to that of DNA, (ii) of the DNA will be not much lower than that of the test compound, and (iii) most importantly,  is much (orders of magnitude) lower for the indirect damage, because after excitation of a photosensitizer the fraction of molecules reacting with DNA is relatively small.     Only in the case where the genotoxicity assay is extremely insensitive to pyrimidine dimers and at the same time the absorption of the photosensitizer is exclusively restricted to the UVB range there is a likelihood that this assay will detect the photogenotoxicity of the photosensitizer. Since in most genotoxicity assays the intrinsic genotoxicity of the pyrimidine dimers will not be much lower than that of the compound-induced types of DNA modification, it appears unlikely that using UVB-irradiation will contribute to the detection of the photogenotoxic effect of the compound.
    Thus, on a theoretically basis, it appears reasonable to perform irradiation of cells in the presence of the compound with UVA alone. However, since solar irradiation conditions should be simulated as near as possible for the routine testing of unknown photogenotoxins, the recommendation was made by an international group to irradiate with a mixture of both UVA and UVB [5]. This advice is still followed in industrial testing protocols. It must be kept in mind, however, that the UVB part will limit the sensitivity of the genotoxicity test system in many cases and––dependent on the method––the use of filters might be necessary (see above).

4.2. Cell-free investigations on naked DNA

    Principal mechanisms responsible for the action of photogenotoxic chemicals may be studied in cell-free systems by irradiating mixtures of the test compound and `naked‘ DNA in aqueous solutions. Such studies can bridge the areas of photochemistry and photogenotoxicity. A compilation of published data in this field is given in Table 2.
Table 2. Investigations on naked DNA in the presence of UV-irradiation/vis
    Gel analysis of plasmids or alkaline elution of high molecular weight DNA has been employed in many studies. Direct breakage of the DNA backbone constitutes a prominent type of damage induced by photogenotoxins. Single- and double-strand breaks can easily be discriminated using supercoiled plasmids. Alkali or piperidine labile sites can be transformed into strand breaks by the respective treatment. Other DNA damages (base modifications) can be transformed into breaks by application of DNA repair enzymes such as FPG protein, endonuclease III, T4-endonuclease V, exonuclease III, etc. The spectra of enzyme sensitive sites can give information on the type of DNA damage induced. Gel analysis can give information on site specificity of action.
    Termination of replication by DNA polymerases or of transcription by RNA polymerases has been studied as another `biological‘ endpoint where site specific action can be monitored.   Transforming activity (e.g. with B. subtilis DNA), plasmid survival and mutation induction in irradiated plasmid transferred into bacteria or into S. cerevisiae have also been studied.
Chemical analysis of irradiated DNA, e.g. for 8-oxoG, can yield direct information on the involvement of oxidative radicals without interference from cellular processes and lengthy purification methods.
The described `simplified‘ test systems allow the identification of potential reactive species and types and site specificity of DNA damage without the interference of cellular processes. This obviously is of high relevance to the understanding of potential toxicological mechanisms and hazards. However, it should be kept in mind that the exposure conditions are even more artificial than in the in vitro (i.e. cellular) photogenotoxicity assays. The chance of photo-excited molecules to transfer their energy is completely different from the situation in cellular environment. The likelihood that energy or charge is transferred to DNA is clearly enhanced because the contact of substance and DNA can be much more intimate. This is specifically important for short-lived reactive species. Furthermore, there are practically no `checks‘ (e.g. by toxicity) of neither the substance concentration nor the irradiation dose and thus the possibility of using completely irrelevant conditions is much higher. Nevertheless, in combination with in vitro photogenotoxicity studies the information gained from acellular studies on the potency of a chemical for altering DNA under UV-irradiation can be very worthwhile.
4.3. Photogenotoxicity in bacteria (photo-Ames test)
    Mutation tests in bacteria are quicker and easier to perform than those in mammalian cells. Therefore, bacterial systems, such as the Ames test, are used as the initial test system in standard mutagenicity test batteries and the vast majority of photochemical genotoxicity data also derive from bacterial test systems.
The test methods described in the publications are diverse but reflect a general major problem of all methods used, i.e. the intrinsic cytotoxicity and genotoxicity of UV itself. Thus, a few publications describe methods irradiating the test chemicals prior to exposure of the tester organisms. This method has the advantage that higher UV doses can be used. However, distinct disadvantages of this method are that (a) short-lived photoproducts might be lost and (b) artificially high UV doses might be used. Consequently, biologically irrelevant results might arise. However, some publications describe experimental conditions taking both, cytotoxicity and genotoxicity of UV itself into account. For example, [44, 45, 46 and 47] describe investigations with the aim to establish a standard protocol for bacterial photogenotoxicity assays [48].
    Chemicals showing photogenotoxic effects in bacterial systems derive from diverse chemical classes; either aliphatic or aromatic, homocyclic or heterocyclic, halogen- or amino-substituted or non-substituted. The results, summarized in Table 3, point out that the majority of photogenotoxic chemicals are detected in a photo-modification of the Ames reverse mutation test using the standard Salmonella typhimurium plus an E. coli WP2 tester strain. An exception to this are the acridine dyes, which were detected in an E. coli B/r phage T5 resistance assay but not tested in the photo-Ames test (there are no publications on the latter test, to date). Using the standard Salmonella and E. coli strains, photochemical genotoxicity was detected in all type of strains, i.e. those carrying or carrying not the R-factor plasmid, pkM 101, with or without a deletion through uvr A or uvr B and indicating frameshift or base substitution mutations. Thus, for better comparison with results from the standard, non-UV-mediated genotoxicity assays, the modified reverse mutation assays applying the Ames Salmonella and E. coli WP2 strains are favorable to the E. coli B/r phage T5 system for testing of photogenotoxicity in bacteria.

Table 3. Chemicals showing positive photochemical genotoxic effects in bacterial assays in the presence of UV-irradiation/vis

4.4. Photo-HPRT/photo-MLA assay

    Tests on point mutations in mammalian cells evolved in the seventies and were used as an integral part of the routine genotoxicity test battery for new compounds since then. The same test principle as for tests with bacteria is used: A mutagen-treated cell population is placed under selective pressure so that only mutant cells are able to survive (for review, see e.g. [49]). For two of these test systems, the HPRT-test and the mouse lymphoma assay (MLA), data are available on an adapted protocol using concurrent UV and/or vis irradiation during substance treatment. As a rule established cell lines like V79 and CHO cells were used as test organisms in the HPRT assay. In the two citations found for a photo-MLA L 5178Y cells were treated.
    In Table 4, the results obtained with a photoversion of a mammalian point mutation assay are summarized. However, there is no comparable database existing as e.g. for the photo-Ames test which comprises compounds of many different chemical classes. Most of the investigations using the photo-HPRT test were performed with psoralen and angelicin and their derivatives from 1982 to 1986. This was due to the introduction of PUVA therapy to Psoriasis patients. All of them revealed a positive result at concurrent UV-irradiation. However, in many cases it is not clearly described which wavelengths of the UV range were used for cell irradiation. More recent studies were initiated in order to assess the photogenotoxicity of different fluoroquinolones with slight modifications of the chemical structure. These antibiotics seem to induce photoadverse effects when the X8 position is substituted by a halogen [2], like e.g. positive photogenotoxicity in the photo-HPRT test [50]. This property is abolished when a methyl group is added instead at X8 [51]. Noodt et al. [52] evaluated several porphyrins in the photo-HPRT test. Porphyrins are used in light-activated therapy of cancer patients with cytostatics. His conclusion was that the lipophilic derivatives were clearly positive in the photo-HPRT test whereas the hydrophilic derivatives showed no such photogenotoxicity in this test system.

                                                                                  Table 4. Chemicals tested in the HPRT-assay or MLA in the presence of UV-irradiation/vis

4.5. Photo-chromosomal aberration assay in vitro

    Tests on clastogenicity are an essential part of the testing strategy for chemicals, cosmetics and pharmaceuticals. Within the last decade, some compounds are found to be photoactivated to clastogens by UV or vis light. There is an indication that the determination of photoclastogenicity might be a good measure for the assessment of a photocarcinogenic potential [5 and 53]. However, mechanistically studies in the case of a photoclastogenic substance seem to be necessary to support the risk assessment as shown for fluoroquinolones.
    There are some recommendations published by [48, 54 and 55] how to perform a photo-CA. In addition, the test development for a photo-CA is published by [44 and 46]. However, as shown in Table 5, the photo-CA was performed with different cells lines using different protocols (e.g. light source, irradiation spectrum).

                                                           Table 5. Chemicals tested in the chromosomal aberration assay in vitro in the presence of UV-irradiation/vis

    Chemicals and pharmaceuticals derived from diverse chemical classes are found to be photoclastogenic in vitro, e.g. furocuoumarins and fluoroquinolones (see Table 5). Most of the assays were performed with established cell lines like Chinese hamster fibroblasts (CHO, V79 and CHL). However, also primary human lymphocytes were used. Up to now, the results show that it seems to be irrelevant which cell type is used for the detection of photoclastogenicty. The mode of genotoxic action of the known photoclastogenic compounds appears to be based on the same principle (i.e. generation of reactive oxygen species) in all cell types investigated so far.
 
4.6. Photo-micronucleus assay in vitro

    The in vitro micronucleus test (MNT) is currently under evaluation as an alternative to the traditional cytogenetic analysis of chromosomal aberrations. As a test for photochemical genotoxicity, the photo-MNT in vitro has been mainly used by two working groups on Chinese hamster V79 cells [56, 57, 58 and 59] (see Table 6). Compounds that have been tested for chromosomal aberrations and for micronuclei in a combined UV and chemical treatment approach have yielded consistent results (for comparison, see previous section). Both groups investigated the photochemical induction of micronuclei under various conditions of treatment and light. The test compounds included compounds that absorb in the vis light range such as methylene blue and neutral red [58]. Snyder and Cooper [57], which were using cytochalasin B for their micronucleus protocol, reported in 1999 that the phototoxicity and photogenotoxicity of fluoroquinolones is considerably reduced, if the formation of a DNA/topoisomerase II complex is inhibited via, e.g. sodium azide. These results indicate that the interaction of the fluoroquinolones with the topoisomerase II does not only play a role in the genotoxicity of fluoroquinolones in the dark but also influences their photogenotoxic potential.

                                                                            Table 6. Chemicals tested in the micronucleus assay in vitro in the presence of UV-irradiation/vis 

    In consideration of the primarily clastogenic mechanisms of photochemical genotoxicity, the photo-MNT should provide a good and faster alternative to the cytogenetic analysis of chromosomal aberrations. A lot of protocol aspects need standardization, similar to the CA for photogenotoxicity. In addition, the ongoing effort in the standardization of the in vitro MNT itself [60] will have an influence on the way to test for photogenotoxicity with micronucleus formation as an endpoint.

4.7. Photo-Comet assay in vitro

    The Comet assay, firstly described by Ostling and Johanson [61], was further developed by Singh et al. [62]. Treated cells are embedded in agarose, lysed and subjected to electrophoresis. The DNA damage is detected, e.g. by measuring the distance of DNA migration out of the cell nucleus into the gel. Using the alkaline protocol of the test method, DNA damage can be detected which is expressed as DNA SSB. In recent years, this protocol was adapted to the concurrent use of UV-irradiation during cell treatment. In Table 7, the available literature is summarized. The photo-Comet assay was predominantly used to assess fluoroquinolones for their photogenotoxic potency. The results obtained in different laboratories for the fluoroquinolones Bay Y 3118, lomefloxacin, fleroxacin and ciprofloxacin are in good agreement even when using different cell types. In contrast, contradictory results were published for TiO2. The negative result of 8-MOP in the photo-Comet assay can be explained by its DNA–DNA cross-linking activity under UV-irradiation which prevents the DNA from migration in the agarose gel [63 and 64]. Recently, several oral antidiabetics and diuretics were investigated using the photo-Comet assay [65]. Some of them showed positive results in the photo-Comet assay. It remains unclear whether the underlying mechanism is solely the generation of reactive oxygen species. The addition of the radical scavenger ascorbic acid showed no reducing effect on the DNA damage level whereas tocopherol decreased the photogenotoxic damage only for some but not all compounds.

Table 7. Chemicals tested in the Comet assay in vitro (alkaline protocol) in the presence of UV-Vis

4.8. Comparison of the different in vitro test systems

    Almost all of the in vitro test systems used in a routine genotoxicity test battery for new chemical compounds were adapted to the additional treatment with UV-Vis. The methods for the evaluation of photogenotoxicity described in this article are the ones most commonly used. Whereas first attempts for a photo-CA test and a photo-HPRT assay were performed in the first half of the 1980s due to the upcoming of a psoriasis therapy with psoralenes, the photo-Ames test was adapted in order to fulfill the request for photogenotoxicity testing of sunscreens at the beginning of the 1990s [66]. Beside the UV-irradiation part, these three test methods are performed according to an existing OECD guideline and can be regarded as more or less validated. The concurrent use of irradiation of the cells in a cell culture flask or bacteria on the agar plate with UV- and/or vis-light did not substantially change the standard protocol without irradiation. The far most compounds were tested in the photo-Ames test. However, some Salmonella strains seem to be more sensitive to detect photogenotoxic effects than others. As outlined in the COLIPA report of a task force dealing with photomutagenicity testing of sunscreens [48], there is experimental precedence to use strains TA102, TA1537 and WP2 (E. coli) as tester strains when UV-irradiation is used. TA102 and WP2 as repair proficient strains are less sensitive to the toxic/genotoxic action of CPDs and are, furthermore, able to recognize cross-linking agents (e.g. psoralens). TA102 appears to be the most sensitive strain for oxidative agents––likely the largest class of phototoxins. Because of the high sensitivity of the bacterial strains to CPDs, it appears more appropriate to reduce the UVB component of irradiation in the photo-Ames test compared to the other test systems. Adequate results were also obtained with TA100 [67]. Recently, Gocke and Chételat compared published and unpublished data on 19 phototoxins and concluded that there is no clear precedent of a photogenotoxin which is exclusively positive in the bacterial assay. Furthermore, many of the photogenotoxins which were only weakly positive in the photo-Ames test showed potent action in the photoclastogenicity assays [68].
    The standard photomutagen 8-methoxypsoralen (8-MOP) becomes clearly photogenotoxic in all three test systems and can therefore be used as positive control. Alternatively, chlorpromazine is a feasible positive control for photogenotoxic effects. The relative sensitivity of the photo-test systems towards other photomutagens according to the available database seems to reflect the relative sensitivity of the test systems without UV-irradiation. A recommendation, which test system has to be preferred as the first one in a testing battery cannot be done, since the database is too limited. Only very few compounds were tested in all test systems. The photo-CA test with mammalian cells seems to be a good and sensitive starting point for testing new compounds/new compound classes where nothing is known on their photogenotoxic properties.
Alternatively to the photo-CA, a photo-MNT in vitro can be employed. In recent years, an in vitro version of the MNT was developed which is still in the process of validation. So far, five publications are available on a photoversion of the MNT in vitro [56, 57, 58 and 59]. Again, 8-MOP as well as chlorpromazine are feasible positive controls in this test system. From the results obtained for the MNT in vitro with and without irradiation, it seems that this new test system might be a good alternative to the CA test possessing comparable sensitivity and specificity, provided that evidence for a sufficient cell proliferation was obtained in case of a negative result for a test compound.
    The Comet assay in vitro is a method broadly used in genotoxicity testing in the past 10 years. The measured genetic endpoint, DNA strand breaks, is a common consequence after different DNA damage events. Therefore, the Comet assay is regarded as a sensitive test system. Though the database is somewhat limited for the photo-version of the Comet assay in vitro, clearly positive results are available for chlorpromazine which can be therefore regarded as an adequate positive control in this test system. Due to its DNA–DNA cross-linking activity under UV-irradiation 8-MOP leads to a negative result in the photo-Comet assay and cannot be used as positive control substance. Good agreement of results was obtained for the same fluoroquinolone, when tested in different cell types and laboratories. However, when comparing with other photo-test systems, a contradictionary result was reached for nalidixic acid, which was negative in the photo-Comet assay and positive in the photo-MNT in vitro. Compounds revealing positive results in the photo-Comet assay in vitro and negative results in the photo-MNT in vitro are the tetracyclines doxycycline and oxytetracycline as well as griseofulvin, which is used as chemotherapeuticum. The reason for these differences is unknown. Since only such a few compounds are tested in both test systems it cannot be decided whether these are false positive results of the photo-Comet assay or false negative results of the photo-MNT. However, it might be reasonable to perform the photo-Comet assay in vitro as supplementary test in a photo-test battery. A more general comparison of sensitivity and specificity of the different test and endpoints examined after UV-Vis irradiation methods compiled in the present paper is not possible due to the limited database.
Just recently, a comparison of different in vitro methods for measuring effects induced by UV-irradiation including photogenotoxicity was published by Meunier et al. [69]. However, the main focus of this article lies more on mechanistically based methods.
 
5. In vivo testing for photogenotoxicity

    In vivo testing for photochemical genotoxicity is problematic since the skin cannot be easily utilized in standard approaches. Transgenic mutagenicity models may be useful in this area in that they allow determination of mutations in skin cells. Gorelick [70] was able to demonstrate UVR induced mutations in skin cells of lacI transgenic mice. In keratinocytes which were isolated from mice treated with the fluoroquinolone clinafloxacin it was possible to demonstrate clearly increased level of DNA strand breakage via a photo-Comet assay in vivo [71]. So far, no more data are available on in vivo studies on photogenotoxicity.

6. Predictivity for photochemical carcinogenicity

6.1. Rodent data

    The main target for photochemical carcinogenesis is the skin. Human skin differs considerably from that of the usually employed laboratory animals such as mouse, rat, dog, guinea pig, and rabbit, among which only mice and rats are used for photocarcinogenicity assessment. Due to a close similarity with human skin, minipigs would potentially serve as a good model for human skin effects, but there are no tumor models available for this species. The standard animal approach to test for photocarcinogenesis, photochemical carcinogenicity or photo-co-carcinogenicity is a test using genetically hairless mice (Skh) [72]. The database for photo co-carcinogenicity studies using hairless mice is considerably small [73]. It includes compounds that absorb UV or vis light and compounds that are supposed to act via other mechanisms such as immune suppression [73].
    Regarding the detection of photochemical genotoxins in this tumor model, only a few test results are available for evaluation. Testing for photochemical carcinogenesis in hairless mice of several furocoumarins (i.e. psoralens) and fluoroquinolones resulted in a higher incidence and a shorter latent period for skin tumors compared to UVR alone [74, 75 and 76]. Klecak et al. [76] tested the fluoroquinolones lomefloxacin, fleroxacin, ofloxacin, ciprofloxacin, and the quinolone nalidixic acid, in a photocarcinogenicity study with female Grl:Skh-1hr BR mice exposed to UVA (320–400 nm). In the UVA group, tumors were induced in 10% of the animals after a period of 76 weeks. Oral treatment with fluoroquinolones at subphototoxic doses and subsequent UVA radiation enhanced tumor prevalence and drastically shortened the median latent period in comparison to UVA alone. Lomefloxacin, fleroxacin and 8-MOP were the most potent photochemical carcinogens. Lomefloxacin plus UVR treatment led to tumors in 100% of the animals as early as 24 weeks after onset of treatment.
    The psoralens and fluoroquinolones that have been tested for photochemical carcinogenesis and have produced positive results are photochemical genotoxins. They are transformed to reactive metabolites when they absorb UV radiation and/or yield active oxygen species including singlet oxygen [77, 78 and 79]. Hence, the correlation of experimental data between photochemical carcinogens and photochemical genotoxins is convincing, but in vitro systems may detect some compounds as photochemical genotoxins which do not elicit a clear-cut positive result in the hairless mouse tumor model [73]. On the one hand, in vitro tests for photochemical genotoxicity are probably more sensitive compared to the rodent tumor model. On the other hand, irradiation conditions such as the use of a UVB dose with considerable tumorigenic potential for hairless mice have an impact on the animal test result, i.e. weak photochemical genotoxins may not be detectable under co-treatment with an efficiently tumorigenic UVB dose. Using established terms in carcinogenicity, UVA and fluoroquinolones or furocoumarins (8-MOP) are best described as co-carcinogenic.
    Hairless mice develop very rapidly basal cell carcinoma as well as squamous cell carcinoma upon UVB-irradiation but melanoma formation cannot be studied. Since melanoma formation is considered to be the most important risk aspect for humans, this standard animal model has considerable limitations in its predictivity for humans. Transgenic mouse melanoma models have been generated [80] and DMBA and UVB produced melanoma [81] but the animals have not been used for photo-cocarcinogenicity testing to date.
 
6.2. Human data

    UVB is a known human skin carcinogen and its direct genotoxic effects in vitro are considered to be an important factor in this process [82]. Pyrimidine dimers and 6–4PPs, which are the main UVB-related effects on DNA, result specifically in CC TT double base changes and C T substitutions. Such changes are found often in human squamous cell carcinoma, e.g. in the p53 tumor suppressor gene (e.g. [83 and 84]). Yet, it has to be acknowledged that a causal relationship of mutations in critical genes and human skin cancer has not been established unequivocally [85]. In addition to the effects of UVB on the genetic apparatus, its effects on the immune system and in particular immune suppression have to be taken into account [82 and 86].
    To date, the combination of 8-MOP and UVA, known as PUVA therapy, used in the treatment of psoriasis is the only well documented human photochemical genotoxic-carcinogen. Large doses of 8-MOP and UVA over several years yielded an increased risk for development of basal cell carcinoma, squamous cell carcinoma and possibly even to melanoma [87 and 88]. There are convincing data regarding the molecular signature of psoralen mutations. A significant percentage of human squamous cell carcinoma carry psoralen-related photoaddition-mutations in the p53 gene [89]. A majority of p53 mutations occurred at 5‘-TpG sites. 4,5‘,8-trimethylpsoralen plus 365 nm or 405 nm UVA leaves a highly specific signature in the HPRT gene in human lymphocytes. The HPRT mutations are highly targeted to psoralen photoaddition sites, i.e. at 5‘-TpA, 5‘-ApT and 5‘-TpG sites [90]. Skin cancer in tar, methotrexate and PUVA treated psoriasis patients, was associated with a low level of DNA repair compared to those patients who were treated, but did not develop skin cancer [91]. Hence, a causal relationship between psoralen-activation by UVA, its mutations and human skin cancer has been convincingly established.
    Coal tar preparations are discussed as human skin cancer inducers [92 and 93]. Although it is known that polycyclic aromatic hydrocarbons are present in coal tar, it is not known to which extent UV activation of such compounds is a contributing factor for human skin carcinogenesis (for review, see [94]). Several cases of basal cell carcinoma have been reported on amiodarone, a phototoxic antiarrhythmic [95 and 96]. No further mechanistical evaluations are available.
    In conclusion, there are only few data on human photochemical carcinogens available that can be used to evaluate the potential relevance of in vitro/in vivo photochemical genotoxicity testing and rodent photo-cocarcinogenicity testing.
 
7. Rationales for testing and risk assessment

7.1. Which compounds should be tested for photogenotoxicity?

    In contrast to routine genotoxicity testing not every compound should be tested for photochemical genotoxicity. Due to the special case that an UV activation of the compound is required to lead to photochemical genotoxicity, testing may be considered only, if the following conditions are met:
1. the compound absorbs in the UV- and/or vis-light range, and
2. the compound was already proven to reveal phototoxicity, and
3. the compound is intended for use on the skin (topical application), or
4. the compound (or a metabolite) is systemically distributed reaching significant concentrations in skin cells.
5. UV-Vis light-absorbing but non-phototoxic compounds might also be tested in the case that they belong to a chemical class already proven to have phototoxic or photogenotoxic representatives.
     A similar advice for the selection of compounds for photogenotoxicity testing was given by the IWGTP working group "photochemical genotoxicity" [54] and the second ECVAM workshop on "phototoxicity testing" held some months later [37]. However, the group pointed to the fact that there is currently no experimental basis to define specific levels for either the compound concentration in the skin or the molar absorbance below which testing for photochemical genotoxicity would not make sense. Therefore, also in the EMEA draft on photosafety testing [3] no such "threshold" is given.
    Only, if the compounds can absorb light it takes up the energy necessary to alter the DNA or to induce radical formation. However, it seems to be a very difficult task to set a limit of absorption underneath photogenotoxicity testing is not necessary anymore. Up to now, no such limit was defined by any scientific group or agency. Therefore, all available data of the test compound should be carefully reviewed considering all points listed above. If there is any hint to a possible photocytotoxic and/or photogenotoxic potential of the compound (or the corresponding compound class) it should be tested for photogenotoxicity even when showing low absorbance in the UV and vis range.
    Light absorption can only occur in cells, which are reached by UV-irradiation. In nearly all cases, these are skin cells or ocular membranes. In other words, compound and UV-irradiation have to be present in the same cell preferentially at the same time. Based on the current knowledge, setting a limit of a "relevant" concentration of the test compound, e.g. in the skin after systemic distribution is not possible. Again, decisions for testing should be taken by considering all available data and in the light of "responsible care". With respect to topical application, in addition data on percutaneous absorption should be taken into account.
    The main mechanism inducing photogenotoxicity seems to be the generation of ROS. However, there is some evidence that not in all cases the generation of ROS are the main source for the photogenotoxicity of a compound [65, 79 and 97]. Other ways like, e.g. the formation of photoproducts may also occur. Long-lived stable photoproducts may cross membranes and can be available in cells distant from the place of activation.

7.2. Risk assessment

    At this time, the scarcity of human examples and the limited dataset of rodent photochemical carcinogens and non-carcinogens makes the evaluation of the human relevance of short term photochemical genotoxicity tests and rodent photo-cocarcinogenicity tests problematic. Despite these limitations, genotoxicity may be addressed using tests that hold the most promise in assessing potential photochemical mutagenic/clastogenic effects, akin to standard genotoxicity/carcinogenicity testing [98]. From the above data and the parallel experience in standard genotoxicity testing, testing for photochemical genotoxicity should be considered in the phototoxicological assessment of chemicals. Importantly, the main purpose of such testing is to make an assessment of the likelihood of a compound to turn into a photochemical carcinogen upon activation with UV or visible radiation [99]. For such compounds, testing for photochemical genotoxicity preferably in mammalian cells in vitro may be an easy hazard identification approach for photochemical carcinogens [54 and 100]. However, further factors such as immunosuppression, irritation and dedifferentiation and most importantly, the risk that stems from exposure to UVB, are to be considered for risk assessment in photochemical carcinogenesis.
    With respect to the environmental or lifestyle-related risk of UVB exposure, the question has to be posed whether there is any additional and measurable risk to humans from the exposure to photochemical genotoxins as pharmaceuticals or cosmetic ingredients. Data on the relevant therapeutical condition par excellence in this context, PUVA (8-MOP + UVA), show that quite long exposure periods are needed to record an increase of skin cancer. The oral therapeutical use of fluoroquinolones leads to a wide systemic distribution with significant levels of the compound in skin cells that are exposed to UV-irradiation. Assessment of photochemical genotoxicity shows that fluoroquinolones induce chromosomal damage in the presence of UV-irradiation at concentrations as low as about 1  g/ml. The cmax for photocarcinogenic fluoroquinolones in the skin is within or even above this level. Consequently, there is a high probability that these fluoroquinolones induce chromosomal damage in the presence of UV-irradiation in human skin at therapeutic exposure levels. Since not all patients develop phototoxic skin reactions, cellular damage from some fluoroquinolones+UV irradiation may be induced even if no noticeable skin reaction occurs. However, it may be concluded that there is only a small increase in risk of photochemical carcinogenesis since patients are advised to strictly avoid exposure to UV-irradiation during the treatment that lasts maximum 14 days in standard cases.
    In conclusion, the relative skin cancer risk delivered from a treatment with or exposure to a photogenotoxic agent is a balance between UV as an existing risk factor, the possibility to avoid the activating wavelengths (UV and vis light) and the pharmacokinetics of the compound used.
 
References

1. SCC notes of guidance for testing of cosmetic ingredients for their safety evaluation, Commission of the European Communities Scientific Committee for Cosmetology, 1990.
2. J.M. Domagala, Structure–activity and structure–side-effect relationships for the quinolone antibacterials. J. Antimicrob. Chemother. 33 (1994), pp. 685–706.
3. EMEA note for guidance on photosafety testing, EMEA, CPMP/SWP/398/01, London, 27 June 2002.
4. Guidance for Industry: Photosafety Testing, US Department of Health and Human Services––Food and Drug Administration, Center for Drug Evaluation and Research (CDER), January 2000, Pharmacology and Toxicology, Rockville, MD, USA.
5. E. Gocke, L. Müller, P.J. Guzzie, S. Brendler-Schwaab, S. Bulera, C.F. Chignell, L.M. Henderson, A. Jacobs, H. Murli, R.D. Snyder and N. Tanaka, Considerations on photochemical genotoxicity: report of the IWGTP working group. Environ. Mol. Mutagen. 35 (2000), pp. 173–184
6. T. Douki and J. Cadet, Individual determination of the yield of the main UV-induced dimeric pyrimidine photoproducts in DNA suggest a high mutagenicity of CC photolesions. Biochemistry 40 (2001), pp. 2495–250
7. C. Kielbassa, L. Roza and B. Epe, Wavelength dependence of oxidative DNA damage induced by UV and visible light. Carcinogenesis 18 (1997), pp. 811–816.
8. J. Cadet, M. Berger, T. Douki, B. Morin, S. Raoul, L. Ravanat and S. Spinelli, Effects of UV and visible light radiation on DNA––final base damage. Biol. Chem. 379 (1997), pp. 1275–1286.
9. E.C. Friedberg, G.C. Walker, W. Siede, DNA Repair and Mutagenesis, ASM Press, Washington, 1995.
10. M. Christmann, M.T. Tomicic, W.P. Roos, B. Kaina, Mechanisms of human DNA repair: an update, Toxicology, in press.
11. D.L. Mitchell and R.S. Nairn, The biology of the 6-4-photoproduct. Photochem. Photobiol. 49 (1989), pp. 805–819.
12. V.A. Bohr, C.A. Smith, D.S. Okumoto and P.C. Hanawalt, DNA repair in an active gene: removal of pyrimidine dimers from the DHFR gene of CHO cells is much more efficient than in the genome overall. Cell 40 (1985), pp. 359–369.
13. L. Schaeffer, V. Moncollin, R. Roy, A. Staub, M. Mezzina, A. Sarasin, G. Weeda, J.H. Hoeijmakers and J.M. Egly, The ERCC2/DNA repair protein is associated with the class II BTF2/TFIIH transcription factor. EMBO J. 13 (1994), pp. 2388–2392.
14. H. Masaki and H. Sakurai, Increased generation of hydrogen peroxid possibly from mitochondrial respiratory chain after UVB-irradiation of murine fibroblasts. J. Dermatol. Sci. 14 (1997), pp. 207–216.
15. J.P. Radicella, C. Dherin, C. Desmaze, M.S. Fox and S. Boiteux, Cloning and characterization of hOGG1, a human homolog of the OGG1 gene of Saccharomyces cerevisiae. PNAS 94 (1997), pp. 8010–8015.
16. G. Dianov, C. Bischoff, J. Piotrowski and V.A. Bohr, Repair pathways for Processing of 8-oxoguanine in DNA by mammalian cell extracts. J. Biol. Chem. 273 (1998), pp. 33811–33816.
17. R. Prasad, W.A. Beard, P.R. Strauss and S.H. Wilson, Human DNA polymerase  deoxyribose phosphate lyase. J. Biol. Chem. 273 (1998), pp. 15263–15270.
18. R. Prasad, R.K. Singhal, D.K. Srivastava, J.T. Molina, A.E. Tomkinson and S.H. Wilson, Specific interaction of DNA polymerase beta and DNA ligase I in a multiprotein base excision repair complex from bovine testis. J. Biol. Chem. 271 (1996), pp. 16000–16007.
19. C. Sachsenmaier, A. Radler-Pohl, R. Zinck, A. Nordheim, P. Herrlich and H.J. Rahmsdorf, Involvement of growth factor receptors in the mammalian UVC response. Cell 78 (1994), pp. 963–972.
20. S. Gross, A. Knebel, T. Tenev, A. Neininger, M. Gaestel, P. Herrlich and F.D. Bohmer, Inactivation of protein-tyrosine phosphatases as mechanism of UV-induced signal transduction. J. Biol. Chem. 274 (1999), pp. 26378–26386.
21. S. Haas and B. Kaina, c-Fos is involved in the cellular defence against the genotoxic effect of UV radiation. Carcinogenesis 16 (1995), pp. 985–991.
22. R. Wisdom, R.S. Johnson and C. Moore, c-Jun regulates cell cycle progression and apoptosis by distinct mechanisms. EMBO J. 18 (1999), pp. 188–197.
23. M. Soriani, V. Hejmadi and R.M. Tyrrell, Modulation of c-jun and c-fos transcription by UVA and UVB radiations in human dermal fibroblasts and epidermoid KB cells. Photochem. Photobiol. 71 (2000), pp. 551–558.
24. A. Knebel, H.J. Rahmsdorf, A. Ullrich and P. Herrlich, Dephosphorylation of receptor tyrosine kinases as target of regulation by radiation, oxidants or alkylating agents. EMBO J. 15 (1996), pp. 5314–5325.
25. P.A. Baeuerle and T. Henkel, Function and activation of NF-kappa B in the immune system. Annu. Rev. Immunol. 12 (1994), pp. 141–179.
26. A. Rehemtulla, C. Hamilton, A. Chinnaiyan and V. Dixit, Ultraviolet radiation-induced apoptosis is mediated by activation of CD95 (Fas/APO-1). J. Biol. Chem. 272 (1997), pp. 25783–25786.
27. D. Kulms, B. Poeppelmann, D. Yarosh, A.T. Luger, J. Krutmann and T. Schwarz, Nuclear and cell membrane effects contribute independently to the induction of apoptosis in human cells exposed to UVB radiation. Proc. Natl. Acad. Sci. U.S.A. 96 (1999), pp. 7974–7979.
28. D. Munsch, R. Watanabe-Fukunaga, J.C. Bourdon, S. Nagata, E. May, E. Yonish-Rouach and P. Reisdorf, Human and mouse Fas (APO-1/CD95) death receptor genes each contain a p53-responsive element that is activated by p53 mutants unable to induce apoptosis. J. Biol. Chem. 275 (2000), pp. 3867–3872.
29. T. Dunkern, B. Kaina, Cell proliferation and DNA breaks are involved in ultraviolet light-induced apoptosis in nucleotide excision repair-deficient Chinese hamster cells, Mol. Biol. Cell 13 (2002) 348–361.
30. H. Spielmann, M. Balls, B. D?ring, H.G. Holzhütter, S. Kalweit, G. Klecak, H. L‘Eplattenier, M. Liebsch, W.W. Lovell, T. Maurer, F. Moldenhauer, L. Moore, W. Pape, U. Pfannbecker, J. Potthast, O. De Silva, W. Steiling and A. Willshaw, EEC/COLIPA project on in vitro phototoxicity testing: first results obtained with a Balb/c 3T3 cell phototoxicity assay. Toxicol. In Vitro 8 (1994), pp. 793–796.
31. L.A. Lambert, W.G. Warner, A. Kornhauser, Animal models for phototoxicity testing, in: H.I. Maibach (Ed.), Dermatotoxicology, Taylor & Francis, Washington, DC, 1996, pp. 515–530.
32. W.J.W. Pape, T. Maurer, U. Pfannenbecker and W. Steiling, The red blood cell phototoxicity test (photohemolysis and haemoglobin oxidation): EU/COLIPA validation programme on phototoxicity. ATLA 29 (2001), pp. 145–162.
33. E. Borenfreund and J.A. Puerner, Toxicity determination in vitro by morphological alterations and neutral red absorption. Toxicol. Lett. 24 (1985), pp. 119–124.
34. H. Spielmann, W.W. Lovell, E. H?lzle, B.E. Johnson, T. Maurer, M.A. Miranda, W.J.W. Pape, O. Sapora and D. Sladowski, In vitro phototoxicity testing: the report and recommendations of ECVAM Workshop 2. ATLA 22 (1994), pp. 314–348.
35. H. Spielmann, M. Balls, J. Dupuis, W.J.W. Pape, O. De Silva, H.G. Holzhütter, F. Gerberick, M. Liebsch, W.W. Lovell and U. Pfannenbecker, A study on UV filter chemicals from Annex VII of the European Union Directive 76/768/EEC, in the in vitro 3T3 NRU phototoxicity test. ATLA 26 (1998), pp. 679–708.
36. H. Spielmann, M. Balls, J. Dupuis, W.J.W. Pape, G. Pechovitch, O. De Silva, H.G. Holzhütter, R. Clothier, P. Desolle, F. Gerberick, M. Liebsch, W.W. Lovell, T. Maurer, U. Pfannenbecker, J.M. Potthast, M. Csato, D. Sladowski, W. Steiling and P. Brantom, The international EU/COLIPA in vitro phototoxicity validation study: results of phase II (blind trial). Part 1. The 3T3 NRU phototoxicity test. Toxicol. In Vitro 12 (1998), pp. 305–327.
37. H. Spielmann, L. Müller, D. Averbeck, M. Balls, S. Brendler-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, W.J.W. Pape, P. Ulrich and H.W. Vohr, The second ECVAM workshop on phototoxicity testing. The report and recommendations of ECVAM Workshop 42. ATLA 28 (2000), pp. 777–814.
38. Anon Commission Directive 2000/33/EC of 25 April 2000, adapting to technical progress for the 27th time Council Directive 67/548/EEC on the approximation of laws, regulations and administrative provisions relating to the classification, packaging and labelling of dangerous substances, ANNEX II: B.41, In vitro 3T3 NRU phototoxicity test, Off. J. Eur. Commun. L136 (2000) 90–107.
39. M. Liebsch and H. Spielmann, Currently available in vitro methods used in the regulatory toxicology. Toxicol. Lett. 127 (2002), pp. 127–134.
40. C. Cohen, K.G. Dossou, A. Rougier and R. Roguet, Episkin: an in vitro model for the evaluation of phototoxicity and sunscreen photoprotective properties. Toxicol. In Vitro 8 (1994), pp. 669–671.
41. M. Liebsch, B. D?ring, T.A. Donelly, P. Logemann, L.A. Rheins and H. Spielmann, Application of the human dermal model Skin2 ZK 1350 to phototoxicity and skin corrosivity testing. Toxicol. In Vitro 9 (1995), pp. 557–562.
42. A.M. Api, In vitro assessment of phototoxicity. Toxicol. In Vitro 10 (1997), p. 339.
43. P.A. Jones, W.W. Lovell, A.V. King and L.K. Earl, In vitro testing for phototoxic potential using the epiderm 3-D reconstructed human skin model. Toxicol. Methods 11 (2001), pp. 1–19.
44. S.W. Dean, M. Lane, R.H. Dunmore, S.P. Ruddock, C.N. Martin, D.J. Kirkland and N. Loprieno, Development of assays for the detection of photomutagenicity of chemicals during exposure to UV light. I. Assay development. Mutagenesis 6 (1991), pp. 335–341.
45. S.W. Dean, R.H. Dunmore, S.P. Ruddock, J.C. Dean, C.N. Martin and D.J. Kirkland, Development of assays for the detection of photomutagenicity of chemicals during exposure to UV light. II. Results of testing three sunscreen ingredients. Mutagenesis 7 (1992), pp. 179–182.
46. A. Chételat, S. Albertini, J.H. Dresp, R. Strobel and E. Gocke, Photomutagenesis test development. I. 8-Methoxypsoralen, chlorpromazine and sunscreen compounds in bacterial and yeast assays. Mutat. Res. 292 (1993), pp. 241–250.
47. L. Henderson, J. Fedyk, C. Bourner, S. Windebank, S. Fletcher and W. Lovell, Photomutagenicity assays in bacteria: factors affecting assay design and assessment of photomutagenic potential of para-aminobenzoic acid. Mutagenesis 9 (1994), pp. 459–465.
48. Photomutagenicity Task Force, Final report, The European Cosmetic Toiletry and Perfumery Association (COLIPA), Brussels, Belgium, August 1995.
49. E.R. Nestmann, R.L. Brillinger, J.P. Gilman, C.J. Rudd and S.H. Swierenga, Recommended protocols based on a survey of current practice in genotoxicity testing laboratories. II. Mutation in Chinese hamster ovary, V79 Chinese hamster lung and L5178Y mouse lymphoma cells. Mutat. Res. 246 (1991), pp. 255–284.
50. S.Y. Brendler-Schwaab, B.A. Herbold, A photomutagenic quinolone, in: Proceeding of the VII International Congress of Toxicology, Seattle, Washington, USA, 1995.
51. A.M. Jeffrey, L. Shao, S.Y. Brendler-Schwaab, G. Schlüter and G.M. Williams, Photochemical mutagenicity of phototoxic and photochemically carcinogenic fluoroquinolones in com-parison with the photostable moxifloxacin. Arch. Toxicol. 74 (2000), pp. 555–559.
52. B.B. Noodt, E. Kvam, H.B. Steen and J. Moan, Primary DNA damage HPRT mutations and cell inactivation photo-induced with various sensitizers in V79 cells. Photochem. Photobiol. 58 (1993), pp. 541–547.
53. L. Müller and P. Kasper, The relevance of photomutagenicity testing as a predictor of photocarcinogenicity. Int. J. Toxicol. 17 (1998), pp. 551–558.
54. N.W. Gibson, J.A. Hartley, R.J. LaFrance and K. Vaughan, Differential cytotoxicity and DNA-damaging effects produced in human cells of the Mer+ and Mer- phenotypes by a series of 1-aryl-3-alkyltriazenes. Cancer Res. 46 (1986), pp. 4999–5003.
55. H. Murli, M. Aardema, T. Lawlor and C. Spicer, Photoclastogenicity––an improved protocol, its validation, and investigation of the photogenotoxicity of DMBA. Environ. Mol. Mutagen. 40 (2002), pp. 41–49.
56. B. Kersten, J. Zhang, S.Y. Brendler-Schwaab, P. Kasper and L. Muller, The application of the micronucleus test in Chinese hamster V79 cells to detect drug-induced photogenotoxicity. Mutat. Res. 445 (1999), pp. 55–71.
57. R.D. Snyder and C.S. Cooper, Photogenotoxicity of fluoroquinolones in Chinese hamster V79 cells: dependency on active topoisomerase II. Photochem. Photobiol. 69 (1999), pp. 288–293.
58. B. Kersten, P. Kasper, S.Y. Brendler-Schwaab and L. Müller, Effects of visible light absorbing chemicals in the photo-micronucleus test in Chinese hamster V79 cells. Environ. Mutagen. Res. 23 (2001), pp. 97–102.
59. B. Kersten, P. Kasper, S.Y. Brendler-Schwaab and L. Müller, Use of the photo-micronucleus assay in Chines hamster V79 cells to study photochemical genotoxicity. Mutat. Res. 519 (2002), pp. 49–66.
60. M. Kirsch-Volders, T. Sofuni, M. Aardema, S. Albertini, D. Eastmond, M. Fenech, M.J. Ishidate, E. Lorge, H. Norppa, J. Surralles, W. von der Hude and A. Wakata, Report from the in vitro micronucleus assay working group. Environ. Mol. Mutagen. 35 (2000), pp. 167–172.
61. O. Ostling and K.J. Johanson, Microelectrophoretic study of radiation-induced DNA damages in individual mammalian cells. Biochem. Biophys. Res. Commun. 123 (1984), pp. 291–298.
62. N.P. Singh, M.T. McCoy, R.R. Tice and E.L. Schneider, A simple technique for quantitation of low levels of DNA damage in individual cells. Exp. Cell. Res. 175 (1988), pp. 184–191.
63. N. Babudri, B. Pani, S. Venturini and C. Monti-Bragadin, Mutagenic activity of four furocoumarins: angelicin, 4,5‘-dimethyl-angelicin, psoralen and 8-methyl-psoralen on V79 Chinese hamster cells. J. Environ. Pathol. Toxicol. 7 (1986), pp. 123–130.
64. S.Y. Brendler-Schwaab, B. Kersten, L. Müller, Detection of photochemical genotoxicity in V79 cells using the photo-Comet assay, 2003.
65. E. Selvaag, A.B. Petersen, R. Gniadecki, T. Thorn and H.C. Wulf, Phototoxicity to diuretics and antidiabetics in the cultured keratinocyte cell line HaCaT: evaluation by clonogenic assay and single cell gel electrophoresis (Comet assay). Photodermatol. Photoimmunol. Photomed. 18 (2002), pp. 90–95.
66. N. Loprieno, In vitro assay systems for testing photomutagenic chemicals. Mutagenesis 6 (1991), pp. 331–333.
67. D. Utesch and J. Splittgerber, Bacterial photomutagenicity testing: distinction between direct, enzyme-mediated and light-induced events. Mutat. Res. 361 (1996), pp. 41–48.
68. E. Gocke and A. Chetelat, Photochemical genotoxicity testing: experience with the Ames test and the in vitro chromosomal aberration assay. Environ. Mutagen. Res. 23 (2001), pp. 47–51.
69. J.-R. Meunier, A. Sarasin and L. Marrot, Photogenotoxicity of mammalian cells: a review of the different assays for in vitro testing. Photochem. Photobiol. 75 (2002), pp. 437–447.
70. N.J. Gorelick, Validation issues for the use of transgenic mouse mutation assays in risk assessment. Prog. Clin. Biol. Res. 395 (1996), pp. 81–108.
71. J. Theiss, T. Festerling, M. Dokmanovich, J. Samoy, M. Bush, S. Bulera and V. Ciaravino, Photogenotoxicity studies of clinafloxacin, a quinolone antiinfective agent. Environ. Mol. Mutagen. 29 Suppl. 28 (1997), p. 50.
72. P.D. Forbes, Relevance of animal models of photocarcinogenesis to humans. Photochem. Photobiol. 63 (1996), pp. 357–362.
73. A. Jacobs, J. Avalos, P. Brown and J. Wilkin, Does photosensitivity predict photocarcinogenicity?. Int. J. Toxicol. 18 (1999), pp. 191–198.
74. A.R. Young, I.A. Magnus, A.C. Davis and N.P. Smith, A comparison of the photomutagenic potential of 8-MOP and 5-MOP in hairless albino mice exposed to solar simulated radiation. Br. J. Dermatol. 108 (1983), pp. 507–518.
75. B.E. Johnson, N.K. Gibbs and J. Ferguson, Quinolone antibiotic with potential to photosensitize skin tumorigenesis. J. Photochem. Photobiol. B: Biol. 37 (1997), pp. 171–173.
76. G. Klecak, F. Urbach and H. Urwyler, Fluoroquinolone antibacterials enhance UVA-induced skin tumors. J. Photochem. Photobiol. B: Biol. 37 (1997), pp. 174–181.
77. N.J. de Mol, G.M.J. Beijersbergen van Henegouwen and B. van Beele, Singulet oxygen formation by sensitization of furocoumarins complexed with, or bound covalently to DNA. Photochem. Photobiol. 34 (1981), pp. 661–666
78. M. Hirose, K. Wakabayashi, S. Grivas, S. De Flora, N. Arakawa, M. Nagao and T. Sugimura, Formation of a nitro derivative of 2-amino-3,4-dimethylimidazo[4,5-f]quinoline by photo-irradiation. Carcinogenesis 11 (1990), pp. 869–871.
79. L.J. Martinez, G. Li and C.F. Chignell, Photogeneration of fluoride by the fluoroquinolone antimicrobial agents lomefloxacin and fleroxacin. Photochem. Photobiol. 65 (1997), pp. 599–602
80. F. Beermann, A. Hunziker and A. Foletti, Transgenic mouse models for tumors of melanocytes and retinal pigment epithelium. Pigment Cell Res. 12 (1999), pp. 71–80.
81. M.B. Powell, P.R. Gause, P. Hyman, J. Gregus, M. Lluria-Prevatt, R. Nagle and G.T. Bowden, Induction of melanoma in Tpras transgenic mice. Carcinogenesis 20 (1999), pp. 1747–1753.
82. IARC Solar and Ultraviolet Radiation, Evaluation of Carcinogenic Risks to Humans, IARC Monographs 55, 1992.
83. D.E. Brash, J.A. Rudolph, J.A. Simon, A. Lin, G.J. McKenna, H.P. Baden, A.J. Halperin and J. Pon-ten, A role for sunlight in skin cancer: UV-induce p53 mutations in squamous cell carcinoma. Proc. Natl. Acad. Sci. U.S.A. 88 (1991), pp. 10124–10128.
84. M.A. Nelson, J.G. Einspar, D.S. Alberts and C.A. Balfour, Analysis of p53 gene in human precancerous actinic keratosis lesions and squamous cell cancers. Cancer Lett. 85 (1994), pp. 23–29.
85. B.A. Bridges, UV-induced mutations and skin cancer: how important is the link?. Mutat. Res. 422 (1998), pp. 23–30.
86. M.L. Kripke, Immunologic unresponsiveness induced by UV radiation. Immunol. Rev. 80 (1984), pp. 87–102.
87. R.S. Stern and N. Laird, The carcinogenic risk of treatment for severe Psoriasis. Cancer 73 (1994), pp. 2759–2764.
88. R.S. Stern, K.T. Nichols and L.H. V?kev?, Malignant melanoma in patients treated for psoriasis with methoxsalen (psoralen) and ultraviolet A radiation (PUVA). N. Eng. J. Med. 336 (1997), pp. 1041–1045.
89. A.J. Nataraj, P. Wolf, L. Cerroni and H.N. Ananthaswamy, p53 mutation in squamous cell carcinomas from patients treated with psoralen + UVA (PUVA). J. Invest. Dermatol. 109 (1997), pp. 238–243.
90. A. Laquerbe, C. Guilouf, E. Moustacchi and D. Papadopoulo, The mutagenic processing of psoralen photolesions leaves a highly specific signature at an endogenous human locus. J. Mol. Biol. 254 (1995), pp. 38–49.
91. M. Dybdahl, G. Frentz, U. Vogel, H. Wallin and B. Nexoe, Low DNA repair is a risk factor in skin carcinogenesis: a study of basal cell carcinoma in psoriasis patients. Mutat. Res. 433 (1999), pp. 15–22.
92. F.J. van Schooten and R. Godschalk, Coal tar therapy. Is it carcinogenic?. Drug Safety 15 (1996), pp. 374–377.
93. I.A. Pion, K.L. Koenig and H.W. Lim, Is dermatologic usage of coal tar carcinogenic? A review of the literature. Dermatol. Surg. 21 (1995), pp. 227–231.
94. W.P. Arnold Tar. Clin. Dermatol. 15 (1997), pp. 739–744.
95. B.E. Monk, Amiodarone-induced photosensitivity and basal cell carcinoma. Clin. Exp. Dermatol. 15 (1990), pp. 319–320.
96. B.E. Monk, Basal cell carcinoma following amiodarone therapy. Br. J. Dermatol. 133 (1995), pp. 141–154.
97. C.E. Perrone, K.C. Takahashi and G.M. Williams, Inhibition of human topoisomerase IIalpha by fluoroquinolones and ultraviolet A irradiation. Toxicol. Sci. 69 (2002), pp. 16–22.
98. L. Müller, A. Kasper, B. Kersten and J. Zhang, Photochemical genotoxicity and photochemical carcinogenesis––two sides of a coin?. Toxicol. Lett. 102–103 (1998), pp. 383–387.
99. S.J. Bulera, J.C. Theiss, T.A. Festerling and F.A. de la Iglesia, In vitro photogenotoxic activity of clinafloxacin: a paradigm predicting photocarcinogenicity. Toxicol. Appl. Pharmacol. 156 (1999), pp. 222–230.
100. K.S. Loveday, Interrelationship of photocarcinogenicity. J. Photochem. Photobiol. 63 (1996), pp. 369–372.
101. D.L. Mitchell, J. Jen and J.E. Cleaver, Relative induction of cyclobutane dimers and cytosine photohydrates in DNA irradiated in vitro and in vivo with ultraviolet-C and ultraviolet-B light. Photochem. Photobiol. 54 (1991), pp. 741–746.
102. B. Demple and S. Linn, 5,6-Saturated thymine lesions in DNA: production by ultraviolet light or hydrogen peroxide. Nucleic Acids Res. 10 (1982), pp. 3781–3789.
103. P.W. Doetsch, T.H. Zastawny, A.M. Martin and M. Dizdaroglu, Monomeric base damage products from adenine, guanine and thymine induced by exposure of DNA to ultraviolet radiation. Biochemistry 34 (1995), pp. 737–742.
104. M.J. Peak, J.G. Peak and B.A. Carnes, Induction of direct and indirect single-strand breaks in human cell DNA by far- and near-ultraviolet radiations: action spectrum and mechanisms. Photochem. Photobiol. 45 (1987), pp. 381–387.
105. J.A. Lippke, L.K. Gordon, D.E. Brash and W.A. Haseltine, Distribution of UV light-induced damage in a defined sequence of human DNA: detection of alkaline-sensitive lesions at pyrimidine nucleoside–cytidine sequences. Proc. Natl. Acad. Sci. U.S.A. 78 (1981), pp. 3388–3392.
106. B. Epe, J. Hegler and D. Wild, Singlet oxygen as an ultimately reactive species in Salmonella typhimurium DNA damage induced by methylene blue/visible light. Carcinogenesis 10 (1989), pp. 2019–2024.
107. B. Epe, M. Pflaum and S. Boiteux, DNA damage induced by photosensitizers in cellular and cell-free systems. Mutat. Res. 299 (1993), pp. 135–145.
108. O. Will, E. Gocke, I. Eckert, I. Schulz, M. Pflaum, H.C. Mahler and B. Epe, Oxidative DNA damage and mutations induced by a polar photosensitizer, Ro 19-8022. Mutat. Res. 435 (1999), pp. 89–101.
109. M. Padula, S. Averbeck, S. Boiteux and D. Averbeck, Enzymatic recognition and biological effects of photodynamic damage induced in DNA by 1,6-dioxapyrene plus UVA. J. Photochem. Photobiol. B 41 (1997), pp. 60–66. |
110. M. Padula and S. Boiteux, Photodynamic DNA damage induced by phycocyanin and its repair in Saccharomyces cerevisiae. Braz. J. Med. Biol. Res. 32 (1999), pp. 1063–1071.
111. W. Adam, P. Groer, K. Mielke, C.R. Saha-Moller, R. Hutterer, W. Kiefer, V. Nagel, F.W. Schneider, D. Ballmaier, Y. Schleger and B. Epe, Photochemical and photobiological studies with acridine and phenanthridine hydroperoxides in cell-free DNA. Photochem. Photobiol. 66 (1997), pp. 26–33.
112. W. Adam, K. Mielke, C.R. Saha-Moller, M. Moller, H. Stopper, R. Hutterer, F.W. Schneider, D. Ballmaier, B. Epe, F.F. Gasparro, X. Chen and J. Kagan, Photochemical and photobiological studies of a furonaphthopyranone as a benzo-spaced psoralen analog in cell-free and cellular DNA. Photochem. Photobiol. 66 (1997), pp. 46–54.
113. J. Decuyper, J. Piette, M. Lopez, M.P. Merville and A. van De Vorst, Induction of breaks in deoxyribonucleic acid by photo-excited promazine derivatives. Biochem. Pharmacol. 33 (1984), pp. 4025–4031.
114. M.E. Churchill, A.M. Schmitz, J.G. Peak and M.J. Peak, Photosensitized damage to supercoiled plasmid DNA induced by 334-nm radiation in the presence of 2-thiouracil consists of alkali- and piperidine-labile sites as well as frank strand breaks. Photochem. Photobiol. 52 (1990), pp. 1017–1023.
115. A.A. Oroskar, F.P. Gasparro and M.J. Peak, Relaxation of supercoiled DNA by aminomethyl-trimethylpsoralen and UV photons: action spectrum. Photochem. Photobiol. 57 (1993), pp. 648–654.
116. A.A. Oroskar, C. Lambert and M.J. Peak, Effects of hydroxyl radical scavengers on relaxation of supercoiled DNA by aminomethyl-trimethylpsoralen and monochromatic UVA photons. Free Radic. Biol. Med. 20 (1996), pp. 751–756.
117. Z. Liu, Y. Lu, M. Lebwohl and H. Wei, PUVA (8-methoxypsoralen plus ultraviolet A) induces the formation of 8-hydroxy-2‘-deoxyguanosine and DNA fragmentation in calf thymus DNA and human epidermoid carcinoma cells. Free Radic. Biol. Med. 27 (1999), pp. 127–133.
118. M. Gulston and J. Knowland, Illumination of human keratinocytes in the presence of the sunscreen ingredient Padimate-O and through an SPF-15 sunscreen reduces direct photodamage to DNA but increases strand breaks. Mutat. Res. 444 (1999), pp. 49–60.
119. R. Dunford, A. Salinaro, L. Cai, N. Serpone, S. Horikoshi, H. Hidaka and J. Knowland, Chemical oxidation and DNA damage catalyzed by inorganic sunscreen ingredients. FEBS Lett. 418 (1997), pp. 87–90.
120. L. Martinez and C.F. Chignell, Photocleavage of DNA by the fluoroquinolone antibacterials. J. Photochem. Photobiol. B 45 (1998), pp. 51–59.
121. T.E. Spratt, S.S. Schultz, D.E. Levy, D. Chen, G. Schluter and G.M. Williams, Different mechanisms for the photo-induced production of oxidative DNA damage by fluoroquinolones differing in photostability. Chem. Res. Toxicol. 12 (1999), pp. 809–815.
122. J. Piette, C.M. Calberg-Bacq, M. Lopez and A. van de Vorst, Terminations of DNA synthesis on `proflavine and light‘-treated phi X174 single-stranded DNA. Biochim. Biophys. Acta 781 (1984), pp. 257–264.
123. M.P. Merville, J. Piette, M. Lopez, J. Decuyper and A. van de Vorst, Termination sites of the in vitro DNA synthesis on single-stranded DNA photosensitized by promazines. J. Biol. Chem. 259 (1984), pp. 15069–15077.
124. J. Decuyper, J. Piette, M.P. Merville-Louis and A. van de Vorst, Photosensitization of SV40 DNA mediated by promazine derivatives and 4‘-hydroxymethyl-4,5‘,8-trimethylpsoralen. Inhibition of the in vitro transcription. Biochem. Pharmacol. 36 (1987), pp. 1069–1076.
125. S. Sortino, S. Giuffrida and J.C. Scaiano, Phototoxicity of naphazoline. Evidence that hydrated electrons, nitrogen-centered radicals and OH radicals, trigger DNA damage: a combined photocleavage and laser flash photolysis study. Chem. Res. Toxicol. 12 (1999), pp. 971–978.
126. V. Shafirovich, A. Dourandin, N.P. Luneva, C. Singh, F. Kirigin and N.E. Geacintov, Multiphoton near-infrared femtosecond laser pulse-induced DNA damage with and without the photosensitizer proflavine. Photochem. Photobiol. 69 (1999), pp. 265–274.
127. M. Raha, L. Lacroix and P.M. Glazer, Mutagenesis mediated by triple helix-forming oligonucleotides conjugated to psoralen: effects of linker arm length and sequence context. Photochem. Photobiol. 67 (1998), pp. 289–294.
128. M.J. Ashwood-Smith, Frameshift mutations in bacteria produced in the dark by several furocoumarins absence of activity of 4,5‘,8-trimethylpsoralen. Mutat. Res. 58 (1978), pp. 23–27.
129. M.J. Ashwood-Smith, G.A. Poulton, M. Barker and M. Mildenberger, 5-Methoxypsoralen, an ingredient in several suntan preparations, has lethal, mutagenic and clastogenic properties. Nature 285 (1980), pp. 407–409.
130. B.L. Pool, R. Klein and R.P. Deutsch-Wenzel, Genotoxicity of 5-methoxypsoralen and near ultraviolet light in repair-deficient strains of Escherichia coli WP2. Food Chem. Toxicol. 20 (1982), pp. 177–181.
131. W.H. Koch, Psoralen photomutagenic specificity in Salmonella typhimurium. Mutat. Res. 160 (1986), pp. 195–205.
132. L. Bianchi, R. Melli, R. Pizzala, L.A. Stivala, L. Rehak, S. Quarta and V. Vannini, Effects of beta-carotene and alpha-tocopherol on photogenotoxicity induced by 8-methoxypsoralen: the role of oxygen. Mutat. Res. 369 (1996), pp. 183–194.
133. N.J. de Mol, G.M. Beijersbergen van Henegouwen, G.R. Mohn, B.W. Glickman and P.M. van Kleef, On the involvement of singlet oxygen in mutation induction by 8-methoxypsoralen and UVA-irradiation in Escherichia coli K-12. Mutat. Res. 82 (1981), pp. 23–30.
134. M.L. Riccio, G. Coratza, L. Bovalini and P. Martelli, Investigation of the mutagenic activity in Salmonella typhimurium of the furochromone khellin, proposed as a therapeutic agent for skin diseases. Mutat. Res. 279 (1992), pp. 103–108.
135. W.H. Koch, E.N. Henrikson, E. Kupchella and T.A. Cebula, Salmonella typhimurium strain TA100 differentiates several classes of carcinogens and mutagens by base substitution specificity. Carcinogenesis 15 (1994), pp. 79–88.
136. J.G. Jose, Photomutagenesis by chlorinated phenothiazine tranquilizers. Proc. Natl. Acad. Sci. U.S.A. 76 (1979), pp. 469–472.
137. R.B. Webb, B.S. Hass and H.E. Kubitschek, Photodynamic effects of dyes on bacteria. II. Genetic effects of broad-spectrum visible light in the presence of acridine dyes and methylene blue in chemostat cultures of Escherichia coli. Mutat. Res. 59 (1979), pp. 1–13.
138. A.A. Chetelat, S. Albertini and E. Gocke, The photomutagenicity of fluoroquinolones in tests for gene mutation, chromosomal aberration, gene conversion and DNA breakage (Comet assay). Mutagenesis 11 (1996), pp. 497–504.
139. H. Israel-Kalinsky, J. Tuch, J. Roitelman and A.A. Stark, Photoactivated aflatoxins are mutagens to Salmonella typhimurium and bind covalently to DNA in vitro. Carcinogenesis 3 (1982), pp. 423–429.
140. G.F. Strniste, J.W. Nickols and R.T. Okinaka, Photochemical oxidation of 2-aminofluorene: correlation between the induction of direct-acting mutagenicity and the formation of nitro and nitroso aromatics. Mutat. Res. 151 (1985), pp. 15–24.
141. S. De Flora, M. Bagnasco, A. Izzotti, F. D‘Agostini, M. Pala and F. Valerio, Mutagenicity of polycyclic aromatic hydrocarbon fractions extracted from urban air particulates. Mutat. Res. 224 (1989), pp. 305–318.
142. S. Arimoto-Kobayashi and H. Hayatsu, Formation of direct-acting mutagens from mixtures of N-nitrosomorpholine and carboxylates by UVA-irradiation. Environ. Mol. Mutagen. 31 (1998), pp. 163–168.
143. D. Averbeck, K. Polasa, J.P. Buisson, R. Bensasson, M. Rougee, J. Cadet, J.L. Ravanat, F. Perin, P. Vigny and P. Demerseman, Photobiological activities of 1,6-dioxapyrene in pro- and eukaryotic cells. Mutat. Res. 287 (1993), pp. 165–179.
144. E. Gocke, S. Albertini, A.A. Chetelat, S. Kirchner and W. Muster, The photomutagenicity of fluoroquinolones and other drugs. Toxicol. Lett. 102–103 (1998), pp. 375–381.
145. E. Gocke, A.A. Chetelat, M. Csato, D.J. McGarvey, R. Jakob-Roetne, S. Kirchner, W. Muster, M. Potthast and U. Widmer, Phototoxicity and photogenotoxicity of nine pyridone der. Mutat. Res. 535 (2003), pp. 43–54.
146. E. Gocke, Photochemical mutagenesis: examples and toxicological relevance. J. Environ. Pathol. Toxicol. Oncol. 20 (2001), pp. 285–292.
147. C. Selby, J. Calkins and H. Enoch, Detection of photomutagens in natural and synthetic fuels. Mutat. Res. 124 (1983), pp. 53–60.
148. D. Averbeck, D. Papadopoulo and I. Quinto, Mutagenic effects of psoralens in yeast and V79 Chinese hamster cells. Natl. Cancer Inst. Monogr. 66 (1984), pp. 127–136.
149. N. Babudri, B. Pani, S. Venturini, M. Tamaro, C. Monti-Bragadin and F. Bordin, Mutation induction and killing of V79 Chinese hamster cells by 8-methoxypsoralen plus near-ultraviolet light: relative effects of monoadducts and cross-links. Mutat. Res. 91 (1981), pp. 391–394.
150. B. Pani, N. Babudri, S. Venturini, M. Tamaro, F. Bordin and C. Monti-Bragadin, Mutation induction and killing of prokaryotic and eukaryotic cells by 8-methoxypsoralen 4,5‘-dimethylangelicin, 5-methylangelicin, 4‘hydroxymethyl-4,5‘-dimehylangelicin. Teratogen. Carcinogen. Mutagen. 1 (1981), pp. 407–415.
151. D. Papadopoulo, F. Sagliocco and D. Averbeck, Mutagenic effects of 3-carbethoxypsoralen and 8-methoxypsoralen plus 365-nm irradiation in mammalian cells. Mutat. Res. 124 (1983), pp. 287–297.
152. D. Papadopoulo, D. Averbeck and E. Moustacchi, Mutagenic effects photo-induced in mammalian cells in vitro by two monofunctional pyridopsoralens. Photochem. Photobiol. 44 (1986), pp. 31–39.
153. D. Papadopoulo and D. Averbeck, Genotoxic effects and DNA photoadducts induced in Chinese hamster V79 cells by 5-methoxypsoralen and 8-methoxypsoralen. Mutat. Res. 151 (1985), pp. 281–291.
154. R.N. Swart, M.A. Beckers and A.A. Schothorst, Phototoxicity and mutagenicity of 4,5‘-dimethylangelicin and long-wave ultraviolet irradiation in Chinese hamster cells and human skin fibroblasts. Mutat. Res. 124 (1983), pp. 271–279.
155. M. Tamaro, S. Gastaldi, F. Carlassare, N. Babudri and B. Pani, Genotoxic activity of some water-soluble derivatives of 5-methoxypsoralen and 8-methoxypsoralen. Carcinogenesis 7 (1986), pp. 605–609.
156. K.S. Loveday and B.A. Donahue, Induction of sister chromatid exchanges and gene mutations in Chinese hamster ovary cells by psoralens. Natl. Cancer Inst. Monogr. 66 (1984), pp. 149–155.
157. S. Venturini, N. Babudri, B. Pani, M. Tamaro, C. Antonello, S. Marciani Magno and C. Monti-Bragadin, Different substituents affect the genotoxic activity of psoralen derivatives: a study on mammalian cells and on bacteria in vitro. Med. Biol. Environ. 10 (1982), pp. 211–217.
158. F. Kelly-Garvert and M.S. Legator, Photoactivation of chlorpromazine: cytogenetic and mutagenic effects. Mutat. Res. 21 (1973), pp. 101–105. |
159. S.Y. Brendler-Schwaab, E. von Keutz, G. Schlüter, B.A. Herbold, A new approach to screen for photomutagenicity with the HPRT-assay, in: Photocarcinogenesis: Mechanisms, Models and Human Health Implications, Washington, DC, USA, 1995.
160. R.M. Rerko, M.E. Clay, A.R. Antunez, N.L. Oleinick and H.H. Evans, Photofrin II photosensitization is mutagenic at the tk locus in mouse L5178Y cells. Photochem. Photobiol. 55 (1992), pp. 75–80.
161. Y. Nakagawa, S. Wakuri, K. Sakamoto and N. Tanaka, The photogenotoxicity of titanium dioxide particles. Mutat. Res. 394 (1997), pp. 125–132.
162. A.H. Uggla, The induction of chromosomal aberrations and SCE by visible light in combination with dyes. I. The effect of hypoxia during light exposure in unsynchronized Chinese hamster ovary cells, sensitized with acridines. Mutat. Res. 201 (1988), pp. 229–239.
163. A.H. Uggla, The induction of chromosomal aberrations and SCEs by visible light in combination with dyes. II. Cell cycle dependence, and the effect of hydroxyl radical scavengers during light exposure in cultures of Chinese hamster ovary cells sensitized with acridine orange. Mutat. Res. 231 (1990), pp. 233–242.
164. S. Itoh, S. Nakayama and H. Shimada, In vitro photochemical clastogenicity of quinolone antibacterial agents studied by a chromosomal aberration test with light irradiation. Mutat. Res. 517 (2002), pp. 113–121.
165. A.T. Natarajan, E.A. Verdegaal-Immerzeel, M.J. Ashwood-Smith and G.A. Poulton, Chromosomal damage induced by furocoumarins and UVA in hamster and human cells including cells from patients with ataxia telangiectasia and xeroderma pigmentosum. Mutat. Res. 84 (1981), pp. 113–124.
166. M.J. Ashwood-Smith, E.L. Grant, J.A. Heddle and G.B. Friedmann, Chromosome damage in Chinese hamster cells sensitized to near-ultraviolet light by psoralen and angelicin. Mutat. Res. 43 (1977), pp. 377–385.
167. G. Abel, Chromosome damage induced in human lymphocytes by 5-methoxypsoralen and 8-methoxypsoralen plus UVA. Mutat. Res. 190 (1987), pp. 63–68.
168. Y.A. Alaoui, R. Arutyunyan and I. Emerit, Chromosome damage induced in human lymphocytes by 5-methoxypsoralen and 8-methoxypsoralen plus UVA. Mutat. Res. 190 (1987), pp. 63–68.
169. A. Chetelat, J.H. Dresp and E. Gocke, Photomutagenesis test development. II. 8-Methoxypsoralen, chlorpromazine and sunscreen compounds in chromosomal aberration assays using CHO cells. Mutat. Res. 292 (1993), pp. 251–258. |
170. K. Halkiotis, D. Yova and G. Pantelias, In vitro evaluation of the genotoxic and clastogenic potential of photodynamic therapy. Mutagenesis 14 (1999), pp. 193–198.
171. H.J. Reavy, N.J. Traynor and N.K. Gibbs, Photogenotoxicity of skin phototumorigenic fluoroquinolone antibiotics detected using the Comet assay. Photochem. Photobiol. 66 (1997), pp. 368–373.
172. L. Marrot, J.P. Belaidi, C. Chaubo, J.R. Meunier, P. Perez and C. Agapakis-Causse, Fluoroquinolones as chemical tools to define a strategy for photogenotoxicity in vitro assessment. Toxicol. In Vitro 15 (2001), pp. 131–142.
 
Corresponding author. Tel.: +49-202-368918; fax: +49-202-364345.
*1 Gesellschaft für Umweltmutationsforschung (German speaking section of the European Environmental Mutagen Society (EEMS)).
1 Present address: Aventis Pharma GmbH, D-65795 Hattersheim, Germany.