Cytotoxicity and Mutagenicity of UVB Assessed Using Cultured Rat Fibroblast

A retroviral vector carrying both positive (neo) and negative (herpes simplex virus thymidine kinase or HSV-tk) selection markers was constructed as a substrate for mutational assay in mammalian cells. Using a population of rat fibroblast cells carrying a single copy per cell of retroviral DNA randomly integrated in their chromosomes, we examined the cytotoxic and mutagenic activities of ultraviolet light (UV) at four wavelengths (254, 290, 300, and 320 nm). The action spectra for these activities are similar to some of the previously reported spectra for photochemical DNA modifications, erythema, cell killing, and mouse skin carcinogenesis, except at 290 and 320 nm. At 290 nm, no significant mutagenicity was observed. At 320 nm, both cytotoxic and mutagenic activities were 10 times higher than the values expected from the absorption spectrum for DNA and the action spectrum for bacterial inactivation and mutagenesis. Structural comparison of some of the HSV-tk mutants obtained after irradiation with 300 and 320 nm UV revealed partially different patterns of mutation specificity, suggesting the involvement of multiple molecular mechanisms in the genotoxicity associated with this range of UV. J Epidemiol, 1999 ; 9 : S72-S77.

A retroviral vector carrying both positive (neo) and negative (herpes simplex virus thymidine kinase or HSV-tk) selection markers was constructed as a substrate for mutational assay in mammalian cells.Using a population of rat fibroblast cells carrying a single copy per cell of retroviral DNA randomly integrated in their chromosomes, we examined the cytotoxic and mutagenic activities of ultraviolet light (UV) at four wavelengths (254, 290, 300, and 320 nm).The action spectra for these activities are similar to some of the previously reported spectra for photochemical DNA modifications, erythema, cell killing, and mouse skin carcinogenesis, except at 290 and 320 nm.At 290 nm, no significant mutagenicity was observed.
At 320 nm, both cytotoxic and mutagenic activities were 10 times higher than the values expected from the absorption spectrum for DNA and the action spectrum for bacterial inactivation and mutagenesis.Structural comparison of some of the HSV-tk mutants obtained after irradiation with 300 and 320 nm UV revealed partially different patterns of mutation specificity, suggesting the involvement of multiple molecular mechanisms in the genotoxicity associated with this range of UV.J Epidemiol, 1999 ; 9 : S72-S77.

UVB, cytotoxicity, mutagenicity HISTORICAL BACKGROUND
Various assay systems to assess the risk of environmental carcinogens have been developed [for a review see ref. 1)].In one type of assays, cultured cells are used to detect inactivation of certain marker genes after treatment with mutagens, and the nature of mutations are molecularly analyzed.Markers used for such purpose include both endogenous cellular genes and genes introduced exogenously.
The X-chromosome-linked hypoxanthine-guanine phosphoribosyltransferase (HGPRT) gene is frequently used as an endogenous marker, since there is only one copy of active gene per diploid cell and its inactivation is readily detected by reverse selection with an antimetabolite 6-thioguanine (6-TG).Such endogenous genes, however, are usually not suitable for molecular analyses because of their large sizes: e.g.HGPRT gene spans 44 kb 2).To circumvent this limitation, cells transfected with a "shuttle vector" carrying bacterial markers have been used [for a review see ref. 3)].Following mutagenesis, inactivation of these markers are detected by transforming individual plasmids into bacterial cells and testing their phenotypes.
Such an approach makes it possible to systematically analyze the specificity of mutations induced by certain mutagens mostly in primate cells.One drawback to such approach, however, is the difficulty in discriminating mutations induced by mutagenesis from those induced during other experimental steps, such as transfection and plasmid recovery.
To solve these problems, Ashman and coworkers 4,5) used retroviral shuttle vector carrying bacterial xanthine-guanine phosphoribosyltransferase (Eco-gpt) gene 4), along with another dominant selection marker (neo), and the origins of replication from bacterial plasmid as well as simian virus 40.When this retrovirus is infected into an HGPRT-deficient mouse fibroblast cell line (A9), the cells become resistant to HAT as well as G418 and sensitive to 6-TG.When the Eco-gpt gene is inactivated, the cells become resistant to 6-TG.The mutated genes can be easily rescued into bacteria after DNA amplification and excision in vivo induced by COS cell-fusion6).With retroviral vectors, one can adjust the condition of infection so that no more than one copy per cell of a marker gene is integrated into a non-specific site of host chromosomes.The major advantages of this system include i) high efficiency of gene transfer, ii) stable nature of the integrated proviruses, iii) direct detection of mutations by a phenotypic change of the host cells which lowers the probability of detecting non-specific mutations, and iv) small enough target size for molecular analyses.
The readiness in recovering viral DNA from the cells into bacteria was initially expected to be another important advantage of this system.This point, however, seems to have become less important these days, because the COS-cell fusion was found often to cause large deletion in the DNA 3).Furthermore, recent advance in the technology of polymerase chain reaction (PCR) has made it possible to directly amplify and analyze small segments of DNA in vitro.Nevertheless, this retroviral vector system was successfully used to study mutational specificity of various mutagens in the mouse cells [for a review see refx].

DESIGN OF LTK-15 RETROVIRAL VECTOR
Herpes simplex virus thymidine kinase (HSV-tk) is capable of converting anti-herpes prodrugs, such as acyclovir (ACV) and ganciclovir (GCV) to a toxic metabolites, and therefore makes the host cells sensitive to this drug.HSV-tk provides a unique opportunity in which one can confer the sensitivity, rather than resistance, to a certain drug via gene transfer techniques.If the introduced HSV-tk gene is inactivated or lost, the cells become resistant to these drugs 8), and this allows one to detect the gene defects in a positive manner.This marker has been widely and successfully used as a negative selection marker in gene targeting experiments.
We adapted this principle for mutational assay by constructing a retroviral vector carrying two markers, HSV-tk and neo 9).When a pool of rat fibroblast cells (CREF) previously infected with this virus (named LTK-15) and selected with G418 were exposed to certain mutagen, such as UV, a dose-dependent increase in the frequency of ACV-resistant colonies were observed.It was also possible with this vector to discriminate between small mutations and large deletions simply by testing the intactness of the linked second marker, neo, in the ACVresistant cell clones.This retroviral vector has an important advantage, because it does not require host cells pre-tagged with specific genetic markers for reverse selection, which makes it feasible to use this system in a wide variety of cells.One method previously used to study UV-induced mutations is to irradiate a purified "shuttle vector" DNA which is then passaged through mammalian cells and finally recovered into bacteria.This method allows the efficient identification of a large collection of UV-induced mutants33, 38,39).Although the retrovirus-based systems are not as efficient as such shuttle vector systems, they have their own advantages: e. g. minimum false positives that may occur during the passage and recovery of shuttle vectors and an ability to detect large deletions which would disable shuttle vectors.Furthermore, these retroviral systems provide a unique opportunity for us to detect mutations in a transcriptionally active gene integrated at various sites of mammalian chromosomes.and the cytotoxicity measured on bacteria 16) with a few exceptions as follows.

UVB-INDUCED CELL KILLING AND MUTATIONS DETECTED WITH THE LTK-15/CREFSYSTEM
The relative cytotoxicity and mutagenicity per quantum at 320 rim were about 10 fold higher than those measured on bacteria.Most previous studies, in which action spectra of UV on mammalian cells or whole animals have been determined, employed wavelengths up to 315 nm.In the limited number of studies using wider ranges of UV, shoulders of action spectra in the wavelength range longer than 320 rim have been commonly observed.Such cases include spectra for erythema 10) and cytotoxicity on human lymphoblastoid cells 23), P3 human epithelial cells 25) and rabbit lens epithelial cells 26).Our action spectra for cell killing and the mutagenicity measured with HSV-tk gene 4o) fit with these data.The obvious deviation from the DNA absorption spectrum suggests the possibility of a contribution by some mechanism(s) other than direct photoabsorption by the DNA.
At 290 rim, mutagenicity was very low, and a linear doseresponse curve could not be obtained 40).Jones et al.25)reported similar phenomenon, albeit at much longer wavelengths: that is, mutagenicity as measured by thioguanine-resistance in P3 human epithelial cells was undetectable at 334 rim but detectable at a longer wavelength, 365 rim.Mutagenicity detected in assays using living cells is expected to be influenced by many factors, including the ability of the agent to damage cells and DNA and the ability of the cells to protect DNA from damage or to repair it.Wavelengths which yield particularly low mutagenicity in these assays may reflect photochemical properties of key molecules working in such processes.

STRUCTURES OF THE MUTANTS 40)
In our assay system, the nature of mutations can be assessed at two levels: by testing the sensitivity of colonies to G418 and by direct sequencing.By the first approach, it was found that 300 rim UV induces relatively large deletions at higher frequencies.Although the generality and mechanism of this phenomenon is presently unknown, it may have some relevance to the fording by Miguel and Tyrrell 41) that 315 rim UV causes a significant yield of single-strand breaks in DNA.Determination of the exact deletion sites in our mutants may yield further information on the mechanism.Nevertheless, this phenomenon seems to be wavelength-dependent.It may be of interest to examine whether such deletion-inducing activity is even stronger at 290 nm, because such high frequency deletions may offer another explanation why 290 nm UV was found to be cytotoxic and yet non-mutagenic.
Sequence analyses of some of the mutant HSV-tk genes generated by 300 and 320 nor UV revealed that a majority of the mutations (34 out of 45) were the C to T transition and tandem substitutions mostly at di-or oligo-cytidines sites.Although tandem double mutations are thought to be a hallmark of UVinduced mutation 42), the frequencies of such mutations we observed after UVB irradiation were significantly higher than those detected previously with UVC by other workers.Reid and Loeb 38) reported that reactive oxygen species have the ability to induce tandem double CC>TT mutations.Our observations may therefore suggest the contribution of oxygen free radicals or related molecular species in these mutational events.Transition type single base substitution at the 3'-side of dipyrimidines is also common in UV mutagenesis 44).Our data indicated that the 3'-side of di-cytidine sequence appears to be the most prevalent site for substitutions.This is consistent with the contribution of (6-4) photoproduct; this photoproduct forms most often at CC and TC sequences 17) and the 3' base cannot pair correctly in this conformation 45,46).Some fraction (6 out of 21) of the C -> T transition mutations are associated with CpG sequence and may also be explain by (methyl-) cytosine deamination 47).
Besides these hallmark mutations, some complex mutations which cannot be explained by simple mechanisms were detected at 320 nor.Examples of such mutations are shown in Fig. 2. Some of these complex mutations (e.g.Fig. 2 a) and d)) may be explained by a mechanism analogous to the error-prone DNA synthesis in bacteria 48) ; others may represent multiple events which may or may not have occurred independently.Whether such complex mutations are specific to the UV of longer wavelengths remains to be established.Such information, if confirmed, may be important, because as Setlow et al. 49.50) have recently suggested based on experiments with hybrid fish, solar UVA (>320 nm) probably accounts for a majority of human melanoma cases.It is also interesting to note that Drobetsky et al.51) found a marked increase of T > G transversion in the aprt locus after exposure to UVA of relatively broad emission spectrum (>350 nm), as we found a similar tendency even at 320 nm and this mutation cannot be simply explained by known mechanisms.

CONCLUSIONS
In this study, we found that biological activities of UV at different wavelength differ not only in potency but also in quality.The molecular bases for such differences may be an interesting subject for future studies.Validity and usefulness of the present assay system have been confirmed through this study.Similar studies using other types of cells (e. g. skin keratinocytes which may be more relevant for UV risk assessment) and other types of mutagens may provide important