The sun protection factor (SPF) inadequately defines broad spectrum photoprotection: demonstration using skin reconstructed in vitro exposed to UVA, UVB or UV-solar simulated radiation

ARTICLE

Auteur(s) : Françoise BERNERD, Corinne VIOUX, François LEJEUNE, Daniel ASSELINEAU

L’Oréal Recherche, C. Zviak Centre, 90 rue du Général Roguet, 92583 Clichy, France

Article accepted on 11/02/03

Material and methods

UV sources

UV-SSR, UVA and UVB irradiations were performed using a 1000 Watts Xenon lamp equipped with a dichroic mirror (Oriel, Les Ulis, France) filtered by a UG5, 2 mm thick (Schott, Clichy, France). UVA radiation alone was obtained by inserting a WG335 (3 mm) Schott filter. The measured irradiance of the UV-SSR source obtained with a WG 320 (1.5 mm) Schott filter, complied to the European Cosmetic and Perfumery Association [11] UV-SSR criteria. The relative cumulative erythemal efficacy, which is the key parameter to assess the relevance of the spectrum, was equal to 56.7% for the 290-310 nm waveband (respectively 56.4% for the standard sun defined in the Colipa SPF test method), 86.9% for the 290-320 nm waveband (84.2%), 92.5% for the 290-330 nm waveband (90.3%), 94.7% for the 290-340 nm waveband (91.1%) and 96.3% for the 290-350 nm waveband (95.1%). The standard erythemal dose (SED) which characterizes the erythemal efficacy independently of the individual sensitivity has been recommended by the Commission Internationale de l’Eclairage [28]. The dose rate of the UV-SSR source was 46 SED/ hour, compared to the standard sun dose rate corresponding to 7.8 SED / hour. The UVB spectrum was obtained using a custom made filter obtained by deposition of thin layers on fused silica (MicroModule, Le Plessix Paté, France). Specifications on the transmission of the filter were to absorb more than 50% of the incident light at 320 nm, to let pass a much light as possible under 320 nm and to absorb as much UVA light as possible. The spectral irradiances were carefully measured with a spectroradiometer (Macam Photometrics, Livingston, UK) calibrated against traceable standard lamps (National Physics Laboratories, Teddington, UK) (Fig. 1A).

Sunscreen formulations

Two prototype formulations, A and B, were prepared in the same simple oil-in-water vehicle. Formulation A contained 7% of UVB absorber octocrylene (Uvinul N539, BASF, Germany) and 3% of UVA absorber butylmethoxydibenzoylmethane (Parsol 1989, Givaudan-Roure Vernier, Switzerland). Formulation B contained 3.75% of UVB filter octyl methoxycinnamate (OMC or Parsol MCX, Givaudan Roure Vernier, Switzerland), and 7.5% of zinc oxide (Z-Cote, Sunsmart, USA). Both formulations were characterised by their SPF, UVA-PF and absorbance curves. SPF values were determined in 20 human volunteers according to the COLIPA SPF test method. The values of the SPF were 7.4 ± 1.5 for formulation A and 7.5 ± 1.6 for formulation B. UVA protection factors were determined using the in vivo method based on persistent pigment darkening (PPD) [29]. The values of UVA-PF were 7.2 ± 1.8 for formulation A and 2.8 ± 0.8 for formulation B. The transmission spectra of the sunscreen products were obtained using a Diffey and Robson method [30], through a roughened quartz plate (Fig. 1B).

Reconstructed skin in vitro

Dermal equivalents were prepared as previously described [31] using a collagen-fibroblast mixture containing 10⁶ fibroblasts. After contraction, human keratinocytes were seeded on this support. The culture was maintained during 7 days in immersed conditions and raised at the air-liquid interface for another 7 day period to obtain a complete differentiation process. The same strains of cells were used for the whole study. Normal human keratinocytes were isolated from breast skin (age 24) obtained after mammary reduction and used at the first passage for skin reconstruction. Normal human dermal fibroblasts were isolated after spreading from mammary skin explants, and cultured in DMEM 10% fetal calf serum. Dermal fibroblasts were used at passage 7 for skin reconstruction. Each condition was performed in duplicate in an experiment. Each experiment was repeated at least three times.

Irradiation procedure

Reconstructed skins on grids were irradiated without medium, 10 minutes after topical application of  ~ 2 mg/cm² (checked by weighing) of the vehicle or sunscreens preparation [23]. Formulations were applied using a curved sterile Paster pipette. Samples were rinsed after irradiation using 3 times 7 ml of PBS before adding fresh medium under the grid. Samples were fixed for classical histology or frozen in liquid nitrogen for immunolabelling, 24 h or 48 h after UV exposure. During the post irradiation period, samples were maintained at 37°C and 5% CO2, fed by capillarity with the medium.

Immunostaining

Immunostaining was performed on 5 μm cryostat sections as described previously [22]. Mouse monoclonal antibodies (Mab) were against cyclobutane pyrimidine dimers (CPD) (clone H3) [32], human vimentin (Monosan, The Netherlands). Fluorescein isothiocyanate (FITC)-conjugate rabbit anti-mouse immuno-globulins (Dako, Denmark) were used as second antibodies.

Results

Determination of the protective effect of both sunscreen formulations after UVA irradiation

The Biologically Efficient Dose after UVA exposure (UVA-BED) has been previously described in reconstructed skin as 25 J/cm² [22]. The effects corresponded to dermal alteration characterised by the disappearance of fibroblasts in the superficial area of dermal equivalent 48 hours after exposure. The disappearance of dermal cells is due to an apoptotic process occurring during this period, and which could be visualized using the TUNEL technique as soon as 6 hours after exposure. Epidermal alterations located in the granular layers occurred when UVA doses were increased above the BED (Fig. 2)
Two mg/cm² of the sunscreen preparations and vehicle were topically applied. From the BED, the UVA dose was progressively increased by 5 J/cm² increments. Visualization of fibroblasts within the dermal equivalent 48 hours after UVA exposure was performed by classical histology (Fig. 2A-E) and vimentin immunostaining (Fig. 2F-J). From 50 J/cm² the protection afforded by sunscreen formulations could be distinguished. The skin treated with formulation A still showed a normal epidermal differentiation and the presence of numerous dermal fibroblasts. The skin treated with formulation B exhibited clear alterations of the granular layers and the disappearance of superficial fibroblasts.
To determine the upper limit of protection obtained with formulation A, UVA doses were increased to 250 J/cm² (data not shown). At 100 J/cm² UVA, a good morphology of both epidermis and dermal equivalent was obtained. At 200 J/cm² UVA, the number of fibroblasts in the lattice was decreased, showing typical UVA-biological damage. At 250 J/cm², drastic alterations were observed, both dermal and epidermal. Therefore the upper dose was determined at the dose able to induce the disappearance of fibroblasts i.e. 200 J/cm².

Determination of protection afforded by both sunscreen formulations after UV-SSR irradiation

We wondered if the differential protective efficiency of both sunscreens after UVA irradiation was still relevant after more realistic UV exposure. Actually, obtention of pure UVA radiation required the use of Schott filter WG 335/3mm. However, this led to a low level of short UVA wavelengths within 320 to 340 nm (Fig. 1A). The two sunscreens were then tested using UV-SSR. In addition, since both sunscreens display a similar SPF, the question of their ability to protect similarly against UV-SSR damage was important.
The first experiments were focussed on the determination of the UV-SSR biological efficient dose (Fig. 3) based on the UVB-BED previously determined as the dose able to induce SBC 24 hours after exposure [21] and the UVA-BED determined as described above. A dose response experiment was performed and samples were taken after 24 and 48 hours.
24 hours after UV-SSR exposure, samples displayed typical SBC within the epidermis at a dose of 5.4 J/cm² UV (0.44 J/cm² UVB + 4.96 J/cm² UVA) (Fig. 3 A-B). Increasing the UV dose led to an increase in the number of SBC. At that dose, typical DNA lesions such as thymine dimers were induced, as shown on samples taken immediately after irradiation. Fig. 3 D-E show positive nuclei in all epidermal keratinocytes. Analysis of dermal fibroblasts within the dermal equivalent showed that the dose of 5.4 J/cm² UV-SSR induced the disappearance of fibroblasts 48 hours after exposure (Fig. 3 A,C, F,G). Higher UV-SSR doses led to more drastic damage including alterations of epidermal morphology (not shown).
The UV-SSR-BED was thus determined at a dose of 5.4 J/cm² UV, which corresponds to 3.1 SED.
In order to evaluate the protection afforded by the different formulations, the UV-SSR dose was progressively increased from the UV-SSR-BED up to 98 J/cm² UV (90 J/cm² UVA + 8 J/cm² UVB). The samples were taken 24 or 48 hours after UV-SSR exposure, and analysed for SBC formation and fibroblast disappearance.
Vehicle treated samples showed numerous SBCs and a total absence of dermal fibroblasts at a dose of 19.7 J/cm² (19.1 J/cm² UVA + 0.6 J/cm² UVB) (Fig. 4 A-B). Samples topically applied with both sunscreen formulations displayed a normal morphology up to a dose of 27.6 J/cm² UV (26.7 J/cm² UVA + 0.86 J/cm² UVB). From that dose, sunscreen formulation B did not prevent the disappearance of dermal fibroblasts within the dermal equivalent (Fig. 4 C-F). Higher doses led us to emphasise the differential efficiency between sunscreens (Fig. 4 G-H). Anti-vimentin immunostaining also revealed the poor protection of dermal fibroblasts afforded by formulation B (Fig. 4 I-J). The damages were similar to those obtained with pure UVA radiation (see fig. 2). In addition, at the latter dose (76.3 J/cm² UV), no SBCs were observed in samples pre-treated with either one or the other sunscreen formulation. Higher UV-SSR doses were then tested to analyse SBCs. However, higher UV-SSR dose 87.1 J/cm² UV (80 J/cm² UVA + 7.1 J/cm² UVB) led to total destruction of the tissue when samples were treated with sunscreen B (not shown). As a result, the two sunscreens could not be compared with regard to this specific biological parameter.

Analysis of the protection using a UVB source

Since it was not possible to assess SBCs using UV-SSR exposure, samples were exposed to pure UVB radiation corresponding to the UVB part of UV-SSR spectrum (Fig. 1A). This might avoid the destruction of the tissue at high UV-SSR doses, resulting from a high level of UVA.

A dose response experiment was performed to determine the BED. 0.85 J/cm² UV dose (0.62 J/cm² UVB + 0.23 J/cm² UVA), corresponding to 4 SED, led to the formation of typical SBCs at 24 hours (Fig. 5 A-B) and numerous DNA lesions in all the epidermal layers immediately after exposure (Fig. 5 E-F).

Comparison of both sunscreens was conducted at UVB doses calculated from the proportion of UVB contained in UV-SSR dose resulting in the destruction of the tissue, i.e. 87.1 J/cm² UV (UVB = 7.1 J/cm² + 80 J/cm ² UVA). Therefore the UVB doses chosen were 7.7 J/cm²; 8.5 J/cm² UV and 10.25 J/cm² UV (5.6; 6.1 and 7.45 J/cm² UVB respectively).

These experiments showed that even at the highest UVB dose, the two sunscreen preparations prevent SBC and pyrimidine dimer formation equally (Fig. 5 C-D, G-H). No damage was observed compared to sham-irradiated samples. These data support and confirm the hypothesis that the absence of protection in samples treated with formulation B and exposed to UV-SSR essentially resulted from UVA-induced damage.

Discussion

Relevant specific biological markers of UVB or UVA exposure have been identified in skin reconstructed in vitro. They made possible the evaluation of the efficiency of single absorbers in this model. The present study examines the relevance of the SPF value in broad spectrum photoprotection using more complex formulations and three relevant UV sources. Actually, sunscreen products usually combine chemical and/or physical absorbers [33] and solar UV radiation includes both UVB and UVA radiation with specific spectra [34]. The SPF value, which reflects the protection against sunburn and erythema, is the only protection value requested for commercial suncreen products. Recent studies evidenced that the SPF does not systematically reflect the level of protection provided with regard to other biological or clinical damages [19, 35-37]. These studies also emphasize the role of UVA wavelengths in numerous biological effects. The two tested formulations in our study were designed to follow several criteria: i) rigorous similar SPF, ii) broad spectrum classification and iii) different profiles of filtration. This appeared important to perform a correct comparative analysis of two filtrating systems avoiding different vehicle or additional active components of commercial products. Our results showed that the two formulations are different in their ability to protect dermal fibroblasts after UVA exposure. These alterations have been related to the photoaging process characterised by dermal drastic modifications of extracellular matrix [5, 6, 38]. Dose response experiments defined the upper UVA dose for both sunscreen formulations. Based on calculation of the value of SPF (ratio of MED with sunscreen/MED without sunscreen), we roughly estimate UVA protection factors in this model. The fibroblast alterations were taken as the referred biological end-point. Formulation A gave the value of 8 (200/25), and formulation B only 2 (50/25). These values are similar or at least of the same order of PPD values in vivo (PPD value for formulations A and B are 7.2 and 2.8 respectively).
These data also reinforce the fact that SPF values do not reflect the protection against various biological end-points, even for sunscreens classified as “broad spectrum” [26]. To properly address that point, experiments were performed using more physiological exposure conditions, obtained with UV-SSR. Moreover, the determination of SPF value is recommended with this UV-spectrum [11, 39]. In addition, this study allowed the identification of the biological effects of UV-SSR on skin reconstructed in vitro. The formation of CPD and SBC could be mostly attributed to the UVB part of the spectrum [21, 40-42]. The dermal alterations seemed to be due to the UVA range and could be related to penetration of UVA radiation through the dermis [5, 6, 22, 43]. In the UV-SSR conditions the formulation having the higher PF-UVA provided better skin protection. Because high UV-SSR doses (including high level of UVA) led to the destruction of the tissue protected with formulation B, the evaluation of SBC formation was not possible. The assumption that the low protection afforded by formulation B was due to poor UVA absorption was evaluated after UVB exposure. This study is the first one using such a “pure” UVB source which is not monochromatic. UVB radiation is generally obtained from fluorescent tubes. These tubes deliver a huge amount of short UVB wavelengths not present in the solar spectrum [44]. They are therefore often considered as physiologically irrelevant for photobiology experiments [45]. The present means of obtaining UVB radiation thus accurately reflects a more realistic UVB spectrum compared to outdoor conditions. SBCs as well as CPDs were indeed induced. However the doses required were higher than those using UVB fluorescent tubes [21]. This is probably due to the lack of highly energetic short UVB photons in the UVB-xenon spectrum. The two sunscreens were equally efficient with regard to SBC and CPD prevention, even at the highest UVB dose tested. These results are in agreement with those of Young et al, [46] showing that two sunscreens having similar SPF but a different spectral profile are equally efficient in preventing the formation of DNA dipyrimidine photolesions in human skin. Such a result is also in agreement with the correlation between erythema, thymidine dimers and SBC formation [47, 48].
Taken together, our results clearly point out that the value of SPF does not reflect the efficiency of sunscreens over the entire solar UV spectrum, and against major biological damage induced by sun exposure. In addition, they also confirmed human skin reconstructed in vitro as a powerful model for photobiology experiments and photoprotection evaluation. n

Acknowledgements. We would like to thank F. Christiaens for his expertise and assistance in monitoring the UV sources and dosimetry. We thank S. Forestier and D. Candau for the supply and characterization of sunscreen formulations.

References

1. Ross P M, Carter D M. Cutaneous DNA damage and repair. In “ Physiology, Biochemistry and Molecular Biology of the Skin ”. (Goldsmith L.A. Ed.) 2nd ed. Oxford University Press, New-York, 1991: 410-51.

2. Sarasin A. The molecular pathways of ultraviolet-induced carcinogenesis. Mut Res 1999; 428: 5-10.

3. LeVee G J, Oberhelman L, Anderson T, Koren H, Cooper KD. UVA II exposure of human skin results in decreased immunization capacity, increased induction of tolerance and a unique pattern of epidermal antigen-presenting cell alteration. Photochem Photobiol 1997; 65: 622-9.

4. Damian D, Barnetson R, Halliday G. Low-dose UVA and UVB have different time courses for suppression of contact hypersensitivity to a recall antigen in humans. J Invest Dermatol 1999; 112: 939-44.

5. Lavker R M, Gerberick G F, Veres D A, Irwin C J, Kaidbey K H. Cumulative effects from repetitive exposures to suberythemal doses of UVB and UVA in human skin. J Am Acad Dermatol 1995; 32: 53-62.

6. Kligman L, Akin F J, Kligman A M. The contributions of UVA and UVB to connective tissue damage in hairless mice. J Invest Dermatol 1985; 84: 272-6.

7. Drobetsky E A, Turcotte J, Châteauneuf A. A role for ultraviolet A in solar mutagenesis. Proc Natl Acad Sci USA 1995; 92: 2350-4.

8. Robert C, Muel B, Benoit A, Dubertret L, Sarasin A, Stary A. Cell survival and shuttle vector mutagenesis induced by ultraviolet A and ultraviolet B radiation in a human cell line. J Invest Dermatol 1996; 106: 721-8.

9. Naylor M F, Farmer K C. The case for sunscreens. A review of their use in preventing actinic damage and neoplasia. Arch Dermatol 1997; 133: 1146-54.

10. Rosen CF. In sunscreen photobiology: molecular, cellular, and physiological aspects. F. Gasparro ED. Springer-Verlag and Landes Bioscience. Chap 4, 1997; 63-68.

11. COLIPA. Sun protection factor test method. Brussels: The European Cosmetic Toiletry and Perfumery Association. 21, 1994.

12. Lowe N J, Meyers D P, Wieder J M et al. Low doses of repetitive ultraviolet A induce morphologic changes in human skin. J Invest Dermatol 1995; 105: 739-43.

13. Seite S, Moyal D, Richard S et al. Mexoryl® SX: a broad spectrum absorption UVA filter protects human skin from the effects of repeated suberythemal doses of UVA. J. Photochem Photobiol B:Biolog 1998; 44: 69-76.

14. Kelly D A, Young A R, McGregor J M, Seed P T, Potten C S, Walker S L. Sensitivity to sunburn is associated with susceptibility to ultraviolet radiation-induced suppression of cutaneous cell-mediated immunity. J Exp Med 2000; 191: 561-6.

15. Roberts L K, Learnb D B. Sunscreen SPF values and immune protection levels are equivalent when tests are conducted by appropriate methods. In: Sunscreen photobiology: molecular, cellular and physiological aspects., F. Gasparro, Ed. Landes Bio-sciences, Austin, TX, 1997: 127-55.

16. Ullrich S E, Kim T-H, Ananthaswamy H N, Kripke M. Sunscreen effects on UV-induced immune suppression. J Invest Dermatol Symposium Proceedings 1999; 4: 65-9.

17. Hochberg M, Enk C. Partial protection against epidermal IL-10 transcription and Langerhans cell depletion by sunscreen after exposure of human skin to UVB. Photochem Photobiol 1999; 70: 766-72.

18. Kligman L, Zheng P. The protective effect of a broad-spectrum sunscreen against UVA radiation in hairless mice: a histologic and ultrastructural assessment. J Soc Cosmet Chem 1994; 45: 21-3.

19. Kligman L, Agin P P, Sayre R M. Broad-spectrum sunscreens with UVA I and UVA II absorbers provide increased protection against solar-simulating radiation-induced dermal damage in hairless mice. J Soc Cosmet Chem 1996; 47: 129-55.

20. Rosentein B, Phelps R, Weistock M et al. p53 Mutations in basal cell carcinomas arising in routine users of sunscreens. Photochem Photobiol 1999; 70: 798-806.

21. Bernerd F, Asselineau D. Successive alteration and recovery of epidermal differentiation and morphogenesis after specific UVB-damages in skin reconstructed in vitro. Dev Biol 1997; 183: 123-38.

22. Bernerd F, Asselineau D. UVA exposure of human skin reconstructed in vitro induces apoptosis of dermal fibroblasts: subsequent connective tissue repair and implications in photoaging. Cell Death and Differentiation 1998; 5: 792-802.

23. Bernerd F, Vioux C, Asselineau D. Evaluation of the protective effect of sunscreens on in vitro reconstructed human skin exposed to UVB or UVA irradiation. Photochem Photobiol 2000; 71: 314-20.

24. Reece B T, Deeds D, Rozen M. An in vitro method for screening sunscreen formulations for sun protection factor using a full-thickness skin model. J Soc Cosmet Chem 1992; 43: 307-12.

25. Augustin C, Collombel C, Damour O. Measurements of the protective effect of topically applied sunscreens using in vitro three-dimensional dermal and skin equivalents. Photochem Photobiol 1997; 66: 853-9.

26. Diffey B L. A method for broad spectrum classification of sunscreens. Int J Cosmet Sci 1994; 16:47-52.

27. Moyal D, Chardon A, Kollias N. Determination of UVA protection factors using the persistent pigment darkening (PPD) as the end point. (Part 1) Calibration method Photodermatol Photoimmunol Photomed 2000; 16: 245-9.

28. Commission Internationale de l’Eclairage (CIE) (1998). Erythema reference action spectrum and standard erythema dose. Publ. CIE S 007/E, Vienna, Austria.(http://www.cie.co.at/cie/, Central Bureau, Kegelstrasse 27, 1030 Vienna, Austria).

29. Chardon A, Moyal D, Hourseau C. Persistent pigment darkening response as a method for evaluation of ultraviolet A protection assays in: Sunscreen Development, Evaluation and Regulatory aspects. N.Lowe, N. Shaath; M. Pathak Eds. Marcel Dekker Inc. New-York, Basel, Hong Kong 1997: 559-82.

30. Diffey B L, Robson J. A new substrate to measure sunscreen protection factors thoughout the ultraviolet spectrum. J Soc Cosmet Chem 1989; 40:127-33.

31. Asselineau D, Bernard B A, Bailly C, Darmon M. Epidermal morphogenesis and induction of 67kD keratin polypeptide by culture at the liquid-air interface. Exp Cell Res 1985; 159: 536-9.

32. Roza L, de Gruijl F R, Bergen Henegouwen J P A et al. Detection of photorepair of UV-induced thymine dimers in human epidermis by immunofluorescence microscopy. J Invest Dermatol 1991; 96: 903-7.

33. Pathak M A. Sunscreens: progress and perspectives on photoprotection of human skin against UVB and UVA radiation. J Dermatol 1997; 23: 783-800.

34. Sayre R M, Cole C, Billhimer W, Stanfield J, Ley R. Spectral comparison of solar simulators and sunlight. Photodermatol Photoimmunol Photomed 1990; 7: 159-65.

35. Fourtanier A, Gueniche A, Compan D, Walker S, Young A R. Improved protection against solar-simulated radiation-induced immunosuppression by a sunscreen with enhanced ultraviolet A protection. J Invest Dermatol 2000; 114: 620-7.

36. Seite S, Moyal D, Verdier M P, Hourseau C, Fourtanier A. Accumulated p53 protein and UVA protection level of sunscreens. Photodermatol Photoimmunol Photomed 2000; 16: 3-9.

37. Marrot L, Belaidi J P, Meunier J R, Perez P, Agapakis-Causse C. The human melanocyte as a particular target for UVA radiation and an end-point for photoprotection assessment. Photochem Photobiol 1999; 69: 686-93.

38. Gilchrest B A. Dermatoheliosis (Sun-induced aging). In « Skin and Aging Processes ». Gilchrest B A Ed., CRC Press, 1989: 97-116.

39. U.S. Food and Drug Administration, Tentative final monograph for sunscreen drug products for over-the-counter human use. Fed Reg 1993; 53 (90): 28194-302.

40. Freeman S E, Hacham H, Gange R W, Maytum D J, Sutherland J C, Sutherland B M. Wavelength dependence of pyrimidine dimer formation in DNA of human skin irradiated in situ with ultraviolet light. Proc Natl Acad Sci USA 1989; 86: 5605-9.

41. Perdiz D, Grof P, Mezzina M, Nikaido O, Moustacchi E, Sage E. Distribution and repair of bipyrimidine photoproducts in solar UV-irradiated mammalian cells. J Biol Chem 2000; 275: 26732-42.

42. Woollons A, Kipp C, Young A R et al. The 0,8% ultraviolet B content of an ultraviolet A sunlamp induces 75% of cyclobutane pyrimidine dimers in human keratinocytes in vitro. Br J Dermatol 1999; 140: 1023-30.

43. Scharffetter K, Wlaschek M, Hogg A et al. UVA irradiation induces collagenase in human dermal fibroblasts in vitro and in vivo. Arch Dermatol Res 1991; 283: 506-11.

44. Brown D B, Peritz A E, Mitchell D L, Chiarello S, Uitto J, Gasparro F P. Common fluorescent sunlamps are an inappropriate substitute for sunlight. Photochem Photobiol 2000; 72: 340-4.

45. Gasparro F P, Brown D B. Photobiology 102: UV sources and dosimetry- the proper use and measurement of “photons as a reagent”. J Invest Dermatol 2000; 114: 613-5.

46. Young A R, Sheehan J M, Chadwick C A, Potten C S. Protection by Ultraviolet A and B sunscreens against in situ dipyrimidine photolesions in human epidermis is comparable to protection against sunburn. J Invest Dermatol 2000; 115: 37-41.

47. Young A R, Chadwick C A, Harrison G I, Nikaido O, Ramsden J, Potten C S. The similarity of action spectra for thymine dimers in human epidermis and erythema suggests that DNA is the chromophore for erythema. J Invest Dermatol 1998; 111: 982-8.

48. Young A R. The sunburn cell. Photodermatol 1987; 4: 127-34.