Ultraviolet B-Induced DNA Damage in Human Skin and Its Modulation by a Sunscreen1

[CANCER RESEARCH 58. 2961-2964, July 15, 1998]

Advances in Brief

Ultraviolet B-induced DNA Damage in Human Skin and Its Modulation by a Sunscreen1

Vladimir J. Bykov, Jan A. Marcusson, and Kari Hemminki2

Department of Biosciences al Novum, Karolinska Institute, Novum, 141 57 Huddinge [V. J. B., K. H.Â¡,and Department of Dermatologv, Huddinge Hospital, 141 86 Huddinge [J. A. M.I, Sweden

Abstract participate. The ages of the volunteers varied from 18 to 56 years (median age, 26 years old). The UVB component of solar radiation is a risk factor for skin cancer, UV Radiation. Waldmann F 85/100 W-UV21 tubes, mounted in a Wald- the most common cancer in the Western world. Yet little is known about man UV 3003 K screen (Waldman, Germany) with automatic computerized the induction of DNA damage in human skin by UVB and its modulation light dose measurements, were used for UVB irradiations. This is a commonly by sunscreens. Here, we apply a novel postlabeling high-performance used clinical lamp covering a UVB area of 280-315 nm, emitting â€”¿5%ofthe liquid chromatography technique to quantify UVB-induced photoprod- relative spectral irradiation output in wavelengths of <290 nm and hardly any ucts in skin biopsies with and without sunscreen. The results showed in wavelengths of <280 nm. â€”¿30-foldinterindividual variations in levels of DNA damage in unpro For experiments, we chose a commercially available sunscreen, pH5-Eu- tected skin of the 14 subjects, probably relating to skin cancer suscepti cerin (Beiersdorf AG, Germany), with a stated SPF of 10 and a stated ability bility. On average, sunscreen guards against DNA damage as expected by to filter out both UVB and UVA. The sunscreen was applied as internationally the erythema! response, but some individuals are poorly protected. agreed, at a thickness of 2 /Â¿I/cm2(22) on a lumbar part of back. For a testing of MED with and without a sunscreen protection, the following single doses of Introduction UVB were used; 50, 100. 200, and 400 J/m2 on unprotected and protected sites and, additionally, 800 and 2000 J/m2 on sunscreen-protected areas. MED was Sunscreens are beneficial to skin because they reduce UV-induced assessed after 24 h according to the following grading: 0, no erythema at all; DNA damage (1, 2), lower mutation frequency (3), and inhibit pho- ( + ), mild erythema with partially defined edges (two edges or less); +, toaging (4). Sunscreens may also efficiently reduce the frequency of uniform erythema with sharp borders and 3 or more clearly visible edges that actinic keratosis, a precancerous lesion (5, 6). Thus, sunscreens appear were considered MED; + + , clear red color with slight induration; + + + , to be useful in preventing skin cancer (7). The potency of a sunscreen heavily red color with palpable edema. The test areas were squares measuring is determined by a SPF,3 defined as the ratio of the energy required to 2 cm x 3 cm. The areas were irradiated at 150, 150, 370, 920, and 2000 J/m2 produce a MED (skin reddening or minimal sunburn) through the and biopsied just below the waistline in the back. The latter four sites were covered with a sunscreen before the irradiation. Biopsies from one unirradiated sunscreen compared with the energy required to produce the same area were included as a background. Five volunteers were biopsied 5 times, reaction in the absence of the sunscreen (8). The efficacy of a excluding the 2000 J/m2 irradiated site, and nine volunteers were biopsied 6 sunscreen in preventing erythema may encourage longer sun expo times. After an intradermal injection of 5 mg/ml xylocaine-adrenaline (Astra sure, which may compromise the beneficial effects (9). Although AB, SÃ¶dertÃ¤lje,Sweden), 4-mm punch biopsies were collected. The time there is no doubt that sunscreens are able to reduce many effects of between irradiation and a biopsy excision was <15 min. The specimens were solar radiation, it is unclear how well the protection against erythema put in a plastic tube and rapidly frozen in carbon dioxide ice. The material was guards against DNA damage. stored in a freezer at -20Â°C until analysis. The sample codes were blinded We have developed a method enabling, for the first time, direct until completion of the analyses. identification and quantitation of UV-induced photoproducts in hu DNA Extraction and Photoproduct Analysis. DNA was isolated as de man skin biopsies (10-12). The method measures both cyclobutane scribed (12). From each 4-mm punch biopsy, we could isolate 4-8 /Ag of DNA. For each 32P-postlabeling assay, 3 ng of DNA were digested with pyrimidine dimers and 6-4 photoproducts, the major causes of the enzymes that produced photoproduct-containing trinucleotides and nucleosides lethal (13), mutagenic (14-17), transforming (18), and carcinogenic (10,11,23). (19, 20) properties of UV irradiation. We apply this technique here to Photoproduct standards were synthesized from nonmodified trinucleotides quantify individual skin response to experimental UV irradiation and (23). The standards of measuring photoproducts were mixed with hydrolyzed modulation of the response by sunscreen and prolonged exposure. nonmodified DNA and labeled together with the human samples. The quan tification took into account the amount of a standard added and a peak area Subjects and Methods obtained from human samples. The labeling efficiency at the conditions used varied: -100% for TT=T and 80% for TT=C (cyclobutane pyrimidine dimers) and 70% for TTâ€”T and 30% for TTâ€”C (6-4'-[pyrimidine-2'-one] Volunteers. Fourteen healthy Caucasian volunteers (7 females and 7 males) with skin types I and II (21) were recruited from medical students and pyrimidine photoproducts) (Refs. 10, 11, and 23). The reproducibility of assay laboratory staff at Huddinge Hospital. The study plan was approved by the was calculated as percentage of a difference between two measurements local ethical committee, and the volunteers gave their informed consent to divided by their mean. The average value was 56% for the detected photoproducts. All statistical calculations were performed with Origin Microcal software. Received 4/2/98; accepted 6/1/98. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with Results 18 U.S.C. Section 1734 solely to indicate this fact. 1This study was supported by the Swedish Radiation Protection Institute, EU Envi The 14 volunteers had skin types 1 and II and showed a MED of ronmental and Climate Program, the Swedish Environmental Board, and the Swedish 50-100 J/m2 in 24 h. Two cyclobutane dimers and two 6-4 photo- Society for Medical Research. 2 To whom requests for reprints should be addressed. Phone: 46-8-608-92-43; Fax: products were measured in their skin biopsies following UVB irradi 46-8-608-15-01; E-mail: [email protected]. 3 The abbreviations used are: SPF, sunscreen protection factor; MED. minima] ery ation using the postlabeling-high-performance liquid chromatography thema dose. method. The photoproducts were measured as trinucleotides, in which 2961

I I-TT=C â€¢¿â€¢-TT=T ^ -TT-T -TT-C

Fig. 1. Columns, average data of photoproduct levels on unprotected and sunscreen protected sites after UVB irra diations at a dose of 150 J/m2; bars. SE (n = 14).

UVB UVB-F10 the 5' nucleotide was unmodified and the two 3' nucleotides were spanning a range of >30 fold for the cyclobutane adducts, which were linked by the cyclobutane (marked as =) or 6-4 bond (marked as â€”¿). present in all of the samples (for some individuals, no 6-4 photoprod The measured trinucleotides were TT=C, TT=T, TTâ€”C, and ucts were detectable). Remarkably, there was no correlation between TTâ€”T.Each represents one-fourth of the possible photoproduct in the photoproduct levels between the unprotected and protected skin DNA; e.g., TT=T is one-fourth of the possible photoproducts TT=T, sites (r = â€”¿0.07).Forexample, volunteers A, C, H, L, and O were CT=T, AT=T, and GT=T. However, we only give the results for very effectively protected, whereas volunteers G, M, and N received the particular trinucleotide measured. little benefit from the sunscreen. We observed no effect of age or sex The mean skin responses and SEs among the 14 volunteers to 150 J/m2 are shown in Fig. 1. Cyclobutane dimers (TT=C and TT=T) on photoproduct levels. were formed in unprotected skin at â€”¿2.5photoproducts/IO6nucleo The volunteers received increasing doses of UVB on the sun screen protected skin, up to 2000 J/m2 (Fig. 3). The induction of tides and were ~5 times more abundant than 6-4 photoproducts (TTâ€”Cand TTâ€”T).The sunscreen of SPF 10 reduced the level of photoproducts increased approximately linearly with increasing dose, but at only â€”¿1/10-1/20ofthe response expected in unpro the adducts to â€”¿1/20,somewhatmore than expected by the SPF 10 tected skin (compare to the 150 J/m2 dose; Fig. 3, left). The solid factor. The sunscreen reduced the levels of the different photoproducts to an equal extent. The reduction of each photoproduct by the sun line shows the expected photoproduct levels, assuming that DNA screen was highly significant statistically (P < 0.001, paired t test). damage were reduced to 1/10 by sunscreen. Thus, the sunscreen The individual skin responses to 150 J/m2 are shown in Fig. 2 with performed as expected at the group level, but as seen in the large and without the sunscreen of SPF 10. Only the levels of TT=C in error bars, the large interindividual variability persists at all dose DNA are given in Fig. 2 because there was a good correlation among levels. The erythemal response in the protected area was checked the four photoproducts within individuals (r = 0.7-0.99). There was in nine volunteers. The sunscreen had been effective in all of them an astonishingly large difference in the individual response to UVB, (observed SPF varied between 3 and 20).

TT=C

7.0x1O'6- - unproieciea si --Ã« 6.0x1 0"6 proDÂ«-^iB | | - sunscreen

-"o

-CÂ®4.0x1 0"6 Fig. 2. Columns, individual levels of TT=

-â€¢o2.0x1 0"6 co 1.0X10"6-0.0â€”

^BH I|Bâ€”¿iBâ€”iiB^iBâ€”¿iÂ»-iiB~iiÂ»â€”¿i ABCDEFGH I KLMNO

volunteers 2962

Fig. 3. Columns, average data on the UVB irradiations of sunscreen-protected sites; bars, SE. The initial response on unpro tected site was taken as 100%. A prediction line shows expected levels of the photoproducts on the basis of the SPF (SPF = 10).

UV-150 UV-150-F10 UV-370-F10 UV-920-F10 UV-2000-F10

samples

Discussion the effective DNA doses resulting from sun exposure may be highly individual. If such a sensitivity persists over a lifetime, the DNA Considering that skin cancer is the most common cancer type in damage inflicted by tanning in fair-skinned people may vary by two Western countries and that solar UVB is an important risk factor, the orders of magnitude, at least. Because DNA damage is independent of lack of human data on UVB-induced DNA damage and its modulation erythema, there is no natural avoidance against tanning. The relation by sunscreens is surprising. Two relevant previous papers evaluated a ship, if any, of such a suggested sensitivity to the appearance of total of 11 individuals with indirect techniques, noting a clear effect of multiple skin cancers in individuals without family histories may the sunscreen (1, 2). Here, we apply, for the first time, a direct provide interesting mechanistic leads. The reasons for the individual technique on 14 individuals, analyzed in blinded samples, and find response remain unclear but probably relate to the structure of skin. two unexpected results. For one, there is a remarkable 30-fold differ We had no data on how the photoproducts actually distributed in the ence in the levels of photoproducts in unprotected skin among the individuals after a uniform dose of 150 J/m2. For the other, the epidermal cells. The recoveries of DNA from the skin biopsies were rather uniform at 4-8 jug. Removal of photoproducts by nucleotide protection by the sunscreen appeared to be very individual and en excision repair could not affect the results because the biopsies were tirely independent of the magnitude of the DNA damage in unpro placed on ice immediately after irradiation. Apparently, human skin tected skin or erythema! response. All of the subjects were fair- skinned, of skin types I and II, with recorded MEDs of 50-100 J/m2. contains agents that can modulate the effect of UVB on DNA, either by protecting or enhancing. For example, an equal dose of UVB The only previous indication of individual sensitivity is our pilot study caused more photoproducts in cultured skin than in DNA in solution, on eight volunteers, who also exhibited a large interindividual varia as was shown by the present technique recently (11). tion in response (12). The interindividual variation in this study was A second implication of these results is that the individual varia far larger than any methodological variation, which was â€”¿56%in bility in the protective effects of the sunscreen against DNA damage repeated analyses of the same samples. Although the sunscreen of SPF raises concerns that protection against erythema by the sunscreen is 10 provided a uniform reduction in the level of photoproducts by unrelated to the extent of DNA damage. It should be emphasized that 90-95%, as expected based on the erythema protection, over a dose range of 150-2000 J/m2, the individual responses varied from 99 to the sunscreen was applied uniformly by one person, and the erythemal response recorded in the sunscreen protected area confirmed the 65%. The results reported here are based on one commercial sun efficacy of the screen against erythemal response. Such uniform screen only but we also have tested this sunscreen with SPF 25 and application of the sunscreen in routine use is probably rare. On the two different sunscreens with SPF 10 and 25 from the same company basis of mitigated erythema response, sunscreens may encourage and obtained similar results. prolonged sun exposure among fair-skinned persons. The conse One qualification regarding the results is the UV source used. This quences may be most unfortunate to those whose DNA is poorly was a commonly used clinical lamp covering a UVB area of 280-315 protected by the sunscreen. nm, emitting ~5% of the relative spectral radiation output in wave lengths of <290 nm, with hardly any being <280 nm. Although the References emission from this source at wavelengths of <300 nm was relatively 1. Freeman. S. E.. Ley. R. D.. and Ley. K. D. Sunscreen protection against UV-induced higher than that from solar-simulated sources, the production and pyrimidine dimers in DNA of human skin in Â«VÂ«.Photodermatology, 5: 243-247, distribution of photoproducts in human skin were quite comparable 1988. 2. van Praag, M. C. G.. Roza, L., Boom, B. W.. Out-Luijting. C., Bergen Henegouwen. (12). Interestingly, the protection from photoproducts by the sun J. B. A.. Vermeer, B. J., and Mommaas. A. M. Determination of the photoprotective screen of SPF 10 used here was, on average, close to a factor of 10 efficacy of a topical sunscreen against UVB-induced DNA damage in human epider over the wide dose-range applied here. mis. J. Photochcm. Photobiol. B Biol.. 19: 129-134. 1993. 3. Ananthaswamy. H. N.. Loughlin, S. M., Cox, P., Evans. R. L.. Ullrich. S. E.. and The results have compelling implications for UVB-induced skin Kripke. M. L. Sunlight and skin cancer: inhibition of p53 mutations in UV-irradiated carcinogenesis, assuming that DNA damage is the initial trigger. First, mouse skin by sunscreens. Nat. Med.. 3: 510-514. 1997. 2963

4. Young, A. R. Senescence and sunscreens. Br. J. Dermatol., 122 (Suppl): 111-114, 14. Lawrence, C. W., Gibbs, P. E. M., Borden, A., Horsfall. M. J., and Kilbey, B. J. 1990. Mutagenesis induced by single UV photoproducts in E. coli and yeast. MutÃ¢t.Res.. 5. Thompson, S. C., Jolley. D., and Marks, R. Reduction of solar kÃ©ratosesby regular 299: 157-163. 1993. sunscreen use. N. Engl. J. Med.. 329: 1147-1151, 1993. 15. Horsfall. M.. and Lawrence. C. W. Accuracy of replication past the T-C (6-4) adduci. 6. Naylor, M. F., Boyd, A.. Smith. D. W., Cameron, G. S., Hubbard, D., and Neldner. J. Mol. Biol., 235: 465-471, 1994. K. H. High sun protection factor sunscreens in the suppression of actinic neoplasia. 16. Gentil, A., Le Page, F., Margot, A., Lawrence. C. W.. Borden. A., and Sarasin. A. Arch. Dermatol.. Â¡31: 170-175, 1995. Mutagenicity of a unique thymine-lhymine dimer or thymine-thymine pyrimidine 7. Naylor, M. F., and Farmer, K. C. The case for sunscreens. A review of their pyrimidone (6-4) photoproduct in mammalian cells. Nucleic Acids Res.. 24: 1837- use in preventing actinic damage and neoplasia. Arch. Dermatol.. 133: 1146-1154. 1840, 1996. 1997. 17. Smith, C. A., Wang, M., Jiang, N., Che, L., Zhao, X., and Taylor, J-S. Mutation 8. Farmer, K. C., and Naylor, M. F. Sun exposure, sunscreens, and skin cancer preven spectra of M13 vectors containing site-specific cis-syn, trans-syn-l (6-4), and Dewar tion: a year-round concern. Ann. Pharmacother., 30: 662-673, 1996. pyrimidone photoproducts of thymidylyl-(3'-5')-thymidine in Escherichia coli under 9. McGregor, J. M., and Young, A. R. Sunscreens, suntans, and skin cancer. Br. Med. J., SOS conditions. Biochemistry, 35: 4146-4154, 1996. 312: 1621-1622. 1996. 18. Pierceall. W. E., and Ananthaswamy, H. N. Transformation of NIH3T3 cells by 10. Bykov, V. J.. and Hemminki. K. UV-induced photoproducts in human skin expiants transfection with UV-irradiated human c-Ha-rai-l proto-oncogene DNA. Oncogene. analysed by TLC and HPLC-radioactivity detection. Carcinogenesis (Lond.). 16: 6: 2085-2091, 1991. 3015-3019. 1995. 19. Ananthaswamy, H. N., and Pierceall. W. E. Molecular mechanisms of ultraviolet 11. Bykov, V. J., and Hemminki, K. Assay of different photoproducts after UVA, B and radiation Carcinogenesis. Photochem. Photobiol., 52: 1119-1136, 1990. C irradiation of DNA and human skin expiants. Carcinogenesis (Lond.). 17: 1949- 20. Kraemer, K. H. Sunlight and skin cancer: another link revealed. Proc. Nati. Acad. Sci. 1955. 1996. USA, 94: 11-14, 1997. 12. Bykov, V. J., Jansen, C. T., and Hemminki. K. High levels of dipyrimidine dimers are 21. Jimbow, K., Fitzpatrick, T. B., and Wick, M. M. Biochemistry and physiology of induced in human skin by solar simulating UV-radiation. Cancer Epidemiol. Bio- melanin pigmentation. In: L. A. Goldsmith and J. H. Sterner (eds.). Physiology. mark. Prev., 7: 199-202, 1998. Biochemistry and Molecular Biology of the Skin, pp. 873-909. New York: Oxford 13. Clingen, P. H., Arieti, C. F.. Cole. J., Waugh, A. P. W., Lowe, J. E., Harcourt, S. A., University Press, 1991. Hermanova. N.. Roza, L., Mori, T., Nikaido, O., and Green, M. H. L. Correlation of 22. Diffey, B. L. People do not apply enough sunscreen for protection. Br. Med. J., 313: UVC and UVB cytotoxicity with the induction of specific photoproducts in T- 942, 1996. lymphocytes and fibroblasts from normal human donors. Photochem. Photobiol-, 61 : 23. Bykov, V. J., Kumar. R., FÃ¶rsli.A., and Hemminki. K. Analysis of UV-induced 163-170, 1995. photoproducts by 32P-postlabelling. Carcinogenesis (Lond.). 16: 113-118, 1995.

2964

Vladimir J. Bykov, Jan A. Marcusson and Kari Hemminki

Cancer Res 1998;58:2961-2964.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/58/14/2961

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/58/14/2961. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.