In October 2007, the conference UV-Radiation-Induced Disease—Roles of UVA and UVB* was held in Stockholm, Sweden. Scientists, physicians, and industry representatives from Europe, the United States, and Australia gathered to present and discuss current knowledge of the biological effects of UV radiation and cellular defense mechanisms. The conference was organized by Rune Toftgård and Dan Segerbäck (Karolinska Institutet, Novum, Huddinge, Sweden) and Johan Hansson (Karolinska Institutet, Stockholm, Sweden). Presentations encompassed three main themes: cellular effects of UV radiation, experimental models of UV-induced disease, and human studies of UV-induced disease. Special reference was made to the relative importance of short (UVB)- and long (UVA)-wavelength UV radiation and the implications for the development of effective preventive measures. This report focuses on the new, unpublished data presented at the meeting.
In his introduction to the conference, Rune Toftgård emphasized the public health importance of solar UV–induced disease and the problem of implementing preventive measures.
In the first session, on cellular effects of UV radiation, Alain Sarasin (Institut Gustave Roussy, Villejuif, France) described work on the polymerases involved in repairing UV-induced lesions in human cells. Examination of the UV mutational spectra of a shuttle vector replicated in Burkitt's lymphoma (BL2) cells engineered specifically to knock out polymerases 
, 
and 
suggested a role for polymerase in the antimutagenic bypass of specific UV-induced lesions.
The intricacies of the roles of heme, Nrf2 (an upregulator), and Bach1 (a downregulator) in regulating UVA-induced heme oxygenase 1 (HO-1) levels in human skin fibroblasts were outlined by Rex Tyrrell (University of Bath, United Kingdom). Inhibition of heme synthesis by succinyl acetate was found to suppress UVA-induced accumulation of Nrf2 in the nucleus, suggesting that this accumulation results from activation mediated by heme release and stabilization of the Nrf2 protein. The refractoriness to reinduction of HO-1 by UVA observed 36–48 hours after the initial induction may be a result of the removal of free heme by HO-1. Cell-labeling experiments suggest that UVA radiation itself activates expression of Bach1 and its nuclear export. After a second UVA dose, Bach1 accumulates very rapidly in the nucleus, probably preventing reinduction of HO-1.
Further work on the cellular effects of UVA was reported by Jean Krutmann (Heinrich-Heine University, Düsseldorf, Germany), who used human skin equivalent models ('dermal equivalents') to examine mtDNA mutagenesis. Dermal equivalents from patients with Kearns–Sayre syndrome, who have an mtDNA deletion, cause increased contraction of collagen gels, indicating the increased production of reactive oxygen species, which is hypothesized to result in mtDNA mutagenesis and aging.
Given evidence from previous work that pyrimidine dimers are the major mutagenic lesions in both UVB and UVA exposure and that it is the cellular response that differs, Thomas Rünger (Boston University School of Medicine, Massachusetts) looked at the activation of p53 in human fibroblasts and keratinocytes. A prominent and longer-lasting activation was found following UVB treatment compared with UVA treatment in both cell types. Flow cytometric analysis demonstrated early S-phase arrest after UVA, but early recovery and late S-phase arrest after UVB treatment of human fibroblasts. These results might help to explain why UVA induces more mutations per initial dimer load.
Gerd Pfeifer (Beckman Research Institute of the City of Hope, Duarte, California) compared the repair of UVA-, UVB-, and simulated sunlight (SSL)-induced lesions in embryonic fibroblasts of the BB mouse, which carries cII and lacI transgenes. The oxidized purines induced by UVA and SSL were repaired within 30 minutes, whereas the 6–4 photoproducts and cyclobutane pyrimidine dimers predominantly induced by UVB and SSL, but not UVA, persisted for 6 and 24 hours, respectively. Footprinting experiments demonstrated that UVA was weakly mutagenic in the cII gene, whereas UVB and SSL were very mutagenic, with an increase in the frequency of C to T transitions at dipyrimidine sites. Rapid repair of UVA-induced oxidized guanines and persistence of primarily UVB-induced cyclobutane pyrimidine dimers could thus explain the characteristic mutations induced by sunlight.
Anna Asplund (Uppsala University, Sweden) described recent results in transcription profiling of microdissected cells from basal cell carcinoma and normal epidermal basal cells. A total of 202 upregulated and 161 downregulated genes were reported. These include known genes such as Frizzled homolog 8, cyclooxygenase, S100 proteins, and a number of 'new' genes of potential interest. The results are being validated by immunohistochemistry in conjunction with the Human Proteome Resource Project.
A urinary assay for biomonitoring of UV-induced DNA damage in humans, in which UV dimers are detected by 32P postlabeling, was used by Dan Segerbäck (Karolinska Institutet, Sweden). High levels of dimers were reported in the urine of lifeguards at a beach in southern Sweden. Lower levels were found in parents and children going swimming over a weekend. Background levels of dimers have also been reported in the urine of adults not engaged in outdoor activities, in the summer but not in winter; fivefold differences were found between individuals.
In the second session of the meeting, on experimental models of melanoma, Edward De Fabo (George Washington University, Washington, DC) described work with the HGF/SF mouse model, which after neonatal UV irradiation develops melanoma that closely resembles the human disease, with metastases. In this model, only UVB is effective; however, preliminary data with C57BL/6 HGF/SF transgenic black or albino mice suggest that UVA can initiate melanoma in the former. Purification of melanocytes by FACS sorting in a newly derived mouse model system (in collaboration with G. Merlino and R. Zaidi, National Cancer Institute, National Institutes of Health) that has inducible melanocyte-specific green fluorescent protein (GFP) expression and microarray determination of gene expression demonstrates specific persistent gene expression in UVB-treated animals that is absent in UVA-treated animals. The GFP mouse is being crossed with HGF/SF mice for experiments to define further UV-induced gene expression.
The role of melanocyte proliferation and migration after neonatal UVR exposure is being unraveled by Graeme Walker (Queensland Institute of Medical Research, Herston, Australia) in melanoma-prone transgenic mice carrying melanocyte-specific activated RAS mutations, HRAS or NRAS. If crossed with mice carrying either a pRb pathway or a p53 pathway defect, the tumors occur earlier and are more aggressive. After neonatal UVB, but not UVA irradiation, all mice show migration of melanocytes to the epidermal basal layer. Ongoing work includes attempts to elucidate the mechanisms of this melanocyte response and to determine whether it plays a role in the initiation of melanoma in the Ras models by priming melanocytes for transformation.
Graham Timmins (University of New Mexico, Albuquerque) described work supporting the hypothesis that melanin has two photosensitizing roles in the melanosome after UV irradiation: direct radical damage to DNA and reactive melanin radical damage to the melanosomal membrane, resulting in leakage of tyrosinase-derived quinones and subsequent production of bulky quinone–DNA adducts. B16 mouse melanoma cells are protected from SSL-induced damage by tyrosinase inhibitors such as methyl gentisate, whereas the addition of tyrosine increases loss of cell viability in this system. Tyrosine-derived DNA adducts were detected in SSL-treated B16 cells using 13C-tyrosine labeling. These adducts might be involved in the generation of T to A transversions in oncogenes such as BRAF. Timmins also discussed the experimental animal models, especially Xiphophorus and Monodelphis domestica, and their potential for melanin-mediated photobiology because the resistance to UV induction of melanoma in human albinos demonstrates its importance.
Meenhard Herlyn (Wistar Institute, Philadelphia, Pennsylvania) described the isolation of multipotent stem cell–like cells from the dermis of human foreskin, their cultivation in vitro as spheres, and their incorporation into synthetic skin. These cells express neural crest and embryonal stem cell markers and can form melanocytes after induction with appropriate growth factors. Comparison of the genomic profile of these dermal spheres with melanoma cell lines and fresh tumor specimens is expected to shed light on the possible role of stem cells in the initiation of melanoma.
Human studies of UV-induced disease were considered in the final session of the meeting. Julia Newton Bishop (University of Leeds, United Kingdom) reported an investigation of the interaction of melanoma susceptibility genes with sun exposure in conjunction with the International Melanoma Genetics Consortium, GenoMel. In a newly completed study, counts of nevi, an important risk factor for melanoma, were investigated in relation to holidays spent in sunny locations before the age of 17 in healthy women from Yorkshire, United Kingdom. Ongoing meta-analyses of recreational and occupational sun exposure and risk of melanoma indicate that intermittent sun exposure is causal.
A behavioral model for estimating population exposure to solar UV radiation was described by Brian Diffey (University of Newcastle, United Kingdom). Input data are ambient erythemal UV, the fraction of solar UV received on the face, and the frequency distribution of time spent outdoors. For northern Europe, a 1,000-fold variation in daily personal dose was predicted, with annual facial exposure of approximately 150 SED and holiday exposure representing approximately one-third of the annual dose. In contrast, in Florida, the daily personal dose was predicted to vary 200-fold, with annual facial exposure of approximately 400 SED and holiday exposure representing approximately one-quarter of the annual dose.
Results after 14 years of follow-up in the Norwegian-Swedish Women's Lifestyle and Health cohort study were reported by Marit B. Veierød (University of Oslo, Norway). Twofold increased risks of melanoma were observed for sunburn in childhood, the teenage years, and early adulthood for two or more sunburns per year relative to none in each age period. Use of a solarium at least once a month during adulthood appears to increase the risk of melanoma. The importance of taking into account the implementation of regulations concerning indoor tanning devices, as well as consequent changes in UVB/UVA ratios, was emphasized.
Boris Bastian (University of California, San Francisco) described four melanoma subtypes defined by the degree of sun damage (elastosis) in adjacent skin and whether the damage is mucosal or nonmucosal. The subtypes differed in genomic changes found and frequency of BRAF mutations. Recent genotype–phenotype correlation studies of 302 primary melanomas showed that BRAF mutations are associated with distinct histopathologic features. Classification trees can be constructed, allowing good prediction of mutation.
Work on gene-expression profiling of melanomas differing in BRAF and NRAS mutations was also described by Johan Hansson (Karolinska Institutet and Karolinska University Hospital, Sweden). A total of 79 genes were significantly overexpressed in BRAF compared with NRAS mutant tumors. Clustering of upregulated genes occurred on chromosome 7.
Antony Young (King's College London, United Kingdom) presented data on the effects of daily suberythemal SSR exposure, which more closely represents the real-life situation for most individuals. Protection against DNA photodamage and induction of DNA repair after a later high-dose SSR challenge were found in skin types that tan well. Young reported accumulation of p53 and DNA photodamage after daily suberythemal exposure that could be prevented by sunscreen application, but no induction of apoptosis. Loss of CD1a was found after daily suberythemal exposure and was also prevented by application of a sunscreen with a low sun protection factor. However, erythema was not a useful indicator of immunosuppression.
In the final presentation, Marianne Berwick (University of New Mexico, Albuquerque) described the international collaborative Genes and Environment in Melanoma (GEM) study. This study of more than 3,000 melanoma patients, which will be completed in 2008, is compiling data on properties of original biopsies, genotype, and sun exposure to determine survival in relation to solar exposure. These analyses were motivated by the results from an earlier study that demonstrated that sun exposure prior to the diagnosis of melanoma appears to have a protective effect in terms of survival. Thus, the replication of this result in the GEM study would have important implications for the directions of future research regarding the mechanism underlying such an effect.
In the discussion that followed the presentations, the importance of studying UVA and UVB in combination as well as separately was highlighted because of the uncertainty concerning the role of oxidative damage in UVA-induced mutagenesis and the possibility that the cellular response might be crucial. Results reported at the meeting indicated the need for more information on pigmentation and how it is regulated. The meeting ended with a lively debate on how to inform the public of the relative risks of UV exposure from sunbed use and sunny vacations.
Notes
* UV-Radiation-Induced Disease—Roles of UVA and UVB was held at the Nobel Forum, Karolinska Institutet, Stockholm, Sweden, 18–20 October 2007. The conference was sponsored by the Swedish Radiation Protection Authority and the Swedish Cancer Society.
