Broad-Spectrum Sunscreens Offer Protection Against Urocanic Acid Photoisomerization by Articial Ultraviolet Radiation in Human Skin1

*Center for Electron Microscopy, Departments of Dermatology and Immunohaematology and Bloodbank, Leiden University Medical Center, Leiden, The Netherlands; Merck KGaA, Cosmetics, Health, Nutrition Division, Business Unit Cosmetics, Research and Development, Darmstadt, Germany

Renate G. van der Molen,* Coby Out-Luiting, Hansju rgen Driller, Frans H.J. Claas, Henk K. Koerten,* and A. Mieke Mommaas*

Cis-urocanic acid (UCA) has been indicated as an important mediator of ultraviolet (UV)-induced immunosuppression. In this study we describe a rapid, noninvasive method for the determination of the protective capacity of various sunscreens against the UV-induced isomerization of trans-UCA into its cis form. For this purpose we applied sunscreens prior to in vivo exposure of human volunteers with single or repeated broadband UVB irradiations of 100 mJ per cm2. We found signicant but different levels of protection against UCA photoisomerization by all sunscreens that correlated with the sun protection factor. A comparison of various sunscreens with a sun protection factor of 10, showed that the best protection was offered by the sunscreens (containing organic UV lters or TiO2) with broad absorption spectra. The ability to inhibit cis-UCA formation was not inuenced by the penetration characteristics of sunscreens, as determined by application of the

sunscreen on quartz glass that was placed on the skin, preventing penetration of sunscreen in the skin. In addition ex vivo UV exposure of human skin was employed to permit other tests of immunomodulation, in this case the mixed epidermal cell lymphocyte reaction. The advantage of this ex vivo method is that there is no need to take biopsies from volunteers. Ex vivo irradiation of human skin with a single dose of 200 mJ per cm2 resulted in similar protection by the sunscreens against cis-UCA formation as in the in vivo system. Furthermore, the mixed epidermal cell lymphocyte reaction data correlated with the cis-UCA ndings. We conclude that UCA isomerization is an excellent method to determine sunscreen efcacy and that broad-spectrum sunscreens offer good immunoprotection. Keywords: immunoprotection/MECLR/titanium dioxide. J Invest Dermatol 115:421426, 2000

t is well known that ultraviolet (UV) radiation, and especially UVB, can cause damage to biologic systems. Besides being erythemogenic and carcinogenic, UVB exposure might result in suppression of some T cell mediated immune responses to antigens encountered within a short period after exposure (reviewed in Kripke, 1984). Several possible mechanisms have been offered to explain UV-induced immunosuppression, one of which involves initiation by specic photoreceptors located in the skin that link UVB with the immune system. Two chromophores have been proposed, DNA and urocanic acid (UCA). Trans-UCA is produced in the stratum corneum in the process of keratinization (Baden and Pathak, 1967). Upon UV
Manuscript received September 14, 1999; revised May 18, 2000; accepted for publication May 30, 2000. Reprint requests to: Dr. R.G. van der Molen, Center for Electron Microscopy, Leiden University Medical Center, PO-Box 9503, 2300 RA Leiden, The Netherlands. Email: R.G.van_der_Molen@lumc.nl Abbreviations: MECLR, mixed epidermal cell lymphocyte reaction; MED, minimal erythema dose; SPF, sun protection factor; UCA, urocanic acid. 1An abstract of part of this study was presented at the 26th Annual Meeting of the American Society for Photobiology in Snowbird, Utah, USA 1998, and published in Photochem Photobiol 67:71S, 1998.
0022-202X/00/$15.00

exposure it isomerizes to its cis form in a dose-dependent manner until a photostationary state is reached, with approximately equal quantities of the two isomers (Anglin et al, 1961; Baden and Pathak, 1967). Ample studies have demonstrated that this molecule plays an important role in UV-induced immunosuppression, although the mechanism of the immunosuppressive activity of cis-UCA is not clear (De Fabo and Noonan, 1983; reviewed in Mohammad et al, 1999). These include modication of antigen-presenting cell function, suppression of both contact hypersensitivity and delayed type hypersensitivity responses, and delay of allograft rejection. Sunscreens are widely used to protect against the damaging effects of UV radiation; however, there are many conicting reports regarding the ability of commercially available chemical sunscreens to protect against UV-induced immunosuppression (see reviews Young and Walker, 1995; Elmets and Anderson, 1996; and recent original article Finlay-Jones and Hart, 1998). There are several explanations for this discrepancy, one of which concerns the penetration characteristics of organic sunscreens. Penetration of sunscreens in the epidermis could lead to insufcient protection against the interaction of UV with UCA (Wolf and Kripke, 1997). In this respect, physical sunscreens (inorganic), containing micronized pigments, would be a good alternative to chemical sunscreens. The metal particles present in these sunscreens are

Copyright # 2000 by The Society for Investigative Dermatology, Inc.

421

422

VAN DER MOLEN ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

not able to penetrate the skin because of their size (Van der Molen et al, 1997). Therefore they remain on the skin surface, where they can offer broadband protection for both the UVA and UVB regions and prevent interaction of UV with UCA. Usually the relative effectiveness of sunscreens is evaluated on the basis of their ability to prevent erythema in human skin, indicated by their sun protection factor (SPF). Erythema is a simple, noninvasive, rapid endpoint for measuring sunscreen protection; however, it is clear from several studies that immunosuppression appears at a lower UV dose than erythema (Walker and Young, 1997; Halliday et al, 1998; Sheehan et al, 1998). Because of this it is desirable to investigate other endpoints for immunoprotection. Photoisomerization of UCA might be a suitable candidate to predict sufcient immunoprotection, because it can be detected at UV doses lower than erythemal exposure (Kammeyer et al, 1995). In this study we have compared the protective capacity of organic sunscreens (containing UV lters) and inorganic sunscreens (containing spherical TiO2 particles) against photoisomerization of UCA. First, we performed in vivo experiments with healthy human volunteers using different UV irradiation protocols. Next, we compared data from these in vivo studies with data obtained from ex vivo studies, using skin from plastic surgery. One of the reasons for this is to evaluate the possibility of avoiding in vivo studies. If biopsies from volunteers are not required, then other parameters for immunoprotection can easily be included. In this study the mixed epidermal cell lymphocyte reaction (MECLR), a method to investigate UV-induced modulation of the alloactivation capacity of Langerhans cells, was used to demonstrate this.
MATERIALS AND METHODS Sunscreens Seven sunscreens (organic and physical) with different in vivo determined SPF, developed for these investigations by Merck KGaA (Darmstadt, Germany), were studied. The SPF was determined with Philips TL-12 lamps and human volunteers with skin types IIV. The composition of the sunscreens is summarized in Table I. Three organic sunscreens contained high concentrations of UVB lters (B10, B16, and B19) and one contained high concentrations of both UVB and UVA lters (AB10). Furthermore, three inorganic sunscreens with increasing concentrations of TiO2 particles (T5, T8, and T10) were included in this study. The vehicle, an oil-in-water formulation, was the same for all preparations. Sunscreens or vehicle were evenly applied (using the recommended standard of 2 mg per cm2) on the skin, 1015 min before irradiation. UV radiation source A Waldmann UV 800 K light booth (Waldmann, Schwenningen, Germany) equipped with Philips TL-12 lamps was used as the UVB source. The major peak of these lamps is at 313 nm and 58% of the total energy is within the UVB range (280320 nm). Of the remaining energy, 41.6% is within the UVA range (320400 nm) and 0.4% within the UVC range (250280 nm). The total irradiance at 30 cm distance was 9.7 W per m2, as measured by a calibrated IL700 spectroradiometer

(International Light, Newburyport, MA) with a cosine-corrected SEE 400 detector and a WBS 320 lter. Experimental design in vivo study Healthy male and female volunteers (skin types IIV), aged 2059, were divided into several groups after informed consent. None of the volunteers had experienced sun exposure of their backs for at least 4 wk preceding the study. For each volunteer the minimal erythema dose (MED) of the back skin was determined for the lowest dose of UVB at which clearly demarcated erythema was observed at 24 h. Furthermore, four volunteers participated in a doseresponse experiment. They received UV-irradiation from 40 to 140 mJ per cm2 on 20 cm2 areas on the back skin without sunscreen or vehicle. In the sunscreen protection study, the rst group of volunteers (n = 14) received a single UVB dose of 100 mJ per cm2 and a second group (n = 18) received this same daily dose for four consecutive days. Prior to each irradiation the various sunscreens or vehicle were applied on 10 cm2 back skin areas. Furthermore, one area was irradiated without sunscreen and two areas were not irradiated and served as nonirradiated controls. Additionally, a few volunteers (n = 6) in the rst group received the sunscreens on both skin and quartz glass. Quartz glass allows UV radiation to pass but prevents penetration of sunscreen in the skin. After each irradiation all areas were cleaned with a tissue to remove remnants of sunscreen or vehicle. Samples for determination of UCA isomers were taken immediately after the last irradiation n et al (1991) and adapted according to the method described by Janse by Kammeyer et al (1997). Briey, two lter papers moistened with 20 ml 0.1 M potassium hydroxide were xed to each test area with adhesive tape. After 4 h, the lters were removed, added to 472 ml 0.1 M potassium hydroxide, and shaken for 30 s. Subsequently, the pH was adjusted to pH 5 with 0.5 M phosphoric acid, followed by stirring. Finally, extracts were ltered through 0.2 mm membrane lters and stored at 20C until they were analyzed for UCA with high-pressure liquid chromotography. Experimental design ex vivo study Breast skin obtained from plastic surgery was cut into pieces of 15 cm2. These pieces were pinned down on polystyrene foam before irradiation. First a doseresponse curve was performed by irradiating pieces of skin with increasing UV doses of 40 to 280 mJ per cm2 (n = 4). For the sunscreen protection studies (n = 7) the pieces were irradiated with a dose of 200 mJ per cm2 with or without prior application of the various sunscreens or vehicle. Immediately after irradiation the pieces were cleaned as described for the in vivo experiments. Part of the skin was cut into pieces of 1 by 1 cm and UCA was extracted by soaking in 1 ml 0.1 M potassium hydroxide for 4 h at 32C (skin temperature in vivo). Then the skin was removed and the pH was adjusted to pH 5 with 0.5 M phosphoric acid. Extracts were subsequently ltered through 0.2 mm membrane lters and stored at 20C until they were analyzed with high-pressure liquid chromotography. The remainders of the pieces of skin were prepared for the MECLR. High-pressure liquid chromotography Cis- and trans-UCA were separated on a 4.6 3 250 mm reversed-phase C18 column, particle size 5 mm (Alltech, Breda, The Netherlands) with a ow of 1 ml per min. Samples of 20 ml were injected by an autosampler with a loop of 50 ml and

Table I. Composition of the sunscreen formulations

Composition (%) Sunscreen Vehicled (SK-03-72) T5 (SK-03-62) T8 (SK-03-61) T10 (SK-03-60) AB10 (SK-03-91) B10 (SK-03-75) B16 (SK-03-74) B19 (SK-03-73)
aIn bEusolex

SPFa 1 5 8 10 10 10 16 19

Eusolex 232b UVB 1 1 2 4

Eusolex 6300b UVB 1 1 2 2

Eusolex 9020b UVA

Eusolex T2000c UVB + UVA 2 4.5 9.5

2 0.25 0.75 1

vivo determined SPF using Philips TL-12 lamps and volunteers with skin types IIV. 232, phenylbezimidazole sulfonic acid; Eusolex 6300, 4-methylbezylidene camphor; Eusolex 9020, butyl methoxydibenzoylmethane. cEusolex T2000, TiO particles coated with aluminum and simethicone. 2 dOil-in-water emulsion.

VOL. 115, NO. 3 SEPTEMBER 2000

SUNSCREEN PROTECTION AGAINST UCA PHOTOISOMERIZATION

423

UCA isomers were detected at 268 nm. Isocratic elution was performed with a mixture of methanol/ethanol/2-propanol (90:9:1), 0.2 M Na2HPO4 and 0.2 M NaH2PO4 in a ratio of 0.5:80:19.5 (vol/vol/vol) (pH 7.3). Standard solutions of pure trans-UCA (Sigma, St. Louis, MO) or chromatographically puried cis-UCA (kindly provided by Dr. Mary Norval of the University of Edinburgh, Scotland) were used to calculate the percentage of cis-UCA. Mixed epidermal cell lymphocyte reaction MECLR experiments (n = 7) were performed as described elsewhere (Hurks et al, 1997a). Briey, cells were isolated from epidermal sheets by trypsinization. Epidermal cells were suspended (1 3 106 cells per ml) in culture medium, i.e., RPMI 1640 supplemented with 1 mM L-glutamine, 100 i.u. penicillin per ml, 100 mg streptomycin per ml, and 15% heat inactivated human serum. Peripheral blood mononuclear cells, used as responder cells, were obtained from heparinized blood by Ficoll-Isopaque gradient centrifugation. The epidermal stimulator cells (20 Gy cesium irradiated) were cultured with allogeneic responder peripheral blood mononuclear cells in a humidied 5% CO2 atmosphere at 37C for 6 d. Eighteen hours before harvesting, each culture was pulsed with 0.5 mCi of tritiated thymidine [3H]TdR and the amount of incorporated [3H]TdR (in counts per minute, cpm) was determined by liquid scintillation spectroscopy. All cultures were carried out at least in triplicate. Hardly any proliferation (< 500 cpm) was found in cultures containing responder peripheral blood mononuclear cells only. The nal results are expressed as percentage thymidine incorporation in relation to the control (cells obtained from the nonirradiated arm). Statistics One-way analysis of variance (ANOVA) and the paired t test were used for analyzing protective capacity of sunscreens. p values of < 0.05 were considered signicant.

RESULTS The percentage of UCA rises with increasing UV doses The percentage of cis-UCA in non-UV-exposed in vivo skin was very low (0.8%). After UV exposure this percentage rose with the dose until a photostationary state was reached of 52% cisUCA at a dose of 100 mJ per cm2 (Fig 1). This dose equals 1.21.5 MED (1 MED was about 75 mJ per cm2). We chose to use this dose to investigate the photoprotective capacity of sunscreen in vivo. The dose response with ex vivo skin showed a similar trend (Fig 1); however, in nonexposed skin the percentage of cis-UCA was somewhat higher than in vivo skin (5.2%) and a higher UV dose of 200 mJ per cm2 was necessary to reach the photostationary state of 44% cis-UCA. Therefore the dose of 200 mJ per cm2 was used to examine the protection of sunscreens against isomerization of UCA and the modulation of the MECLR ex vivo. Sunscreens can protect against photoisomerization of UCA After in vivo exposure to single or repeated UV

irradiations of 100 mJ per cm2, we found an average percentage of cis-UCA of 53%, with no signicant difference between the two UV protocols (Fig 2). As expected, the percentage cis-UCA found with the vehicle was similar to UV alone (Fig 2). All investigated sunscreens could afford signicant protection (p < 0.01) against the isomerization of UCA following single or repeated UV exposures of 100 mJ per cm2 (Fig 2). Application of the sunscreens on the skin without UV irradiation resulted in similar percentages of cisUCA as found in control nonirradiated skin (data not shown). The efcacy of protection was SPF correlated as increased SPF values showed improved protection. The lowest protection was found with sunscreen T5: after a single UV exposure of T5 covered skin 24% cis-UCA was formed. Repeated exposure resulted in 35% cisUCA (Fig 2). The best protection in these experiments was found with both sunscreens B19 and T10 (no signicant difference), respectively, 4.9% and 5.4% cis-UCA after a single exposure and 12% and 14% after repeated exposures. Comparison of the two sunscreens with the same SPF showed that T10, a TiO2 containing sunscreen, offered better protection (p < 0.01) than B10, a sunscreen containing primary organic UVB lters with both UV protocols (Fig 2). In Fig 3 we have plotted the SPF of the sunscreens against the percentages of cis-UCA (data obtained from Fig 2). For both the organic (B) and the inorganic (T) sunscreens two regression lines are drawn for single and repeated UV exposures. Extrapolation of these lines results in a prediction of the SPF value by which complete protection can be found after either a single or repeated UV exposures. For the organic sunscreens (B) this is SPF 20 after a single exposure and SPF 23 after repeated exposures. Inorganic sunscreens (T) provide good protection with an SPF of 11 after a single dose and SPF of 13 after repeated exposures. The results of the ex vivo experiments were very similar to the in vivo data (Fig 4, compare with Fig 2). We found an average percentage of 4.3% cis-UCA in control unirradiated skin. After a dose of 200 mJ per cm2 the percentage rose to 44%. All sunscreens again afforded signicant protection (p < 0.01) against UCA isomerization, which was correlated with their SPF. Also in this experiment it is obvious that sunscreen T10 offered signicantly better protection (p < 0.01) than sunscreen B10, even though they have a similar SPF.

Figure 1. Doseresponse curves of in vivo and ex vivo UCA photoisomerization in human skin. The mean percentage of cisUCA (6SEM) rose after in vivo (solid line) UV irradiation with increasing UV doses until a photostationary plateau was reached of 52% cis-UCA with a dose of 100 mJ per cm2 (n = 4). After ex vivo (dotted line) irradiation of human skin a plateau of 44% cis-UCA was reached after a dose of 200 mJ per cm2 (n = 4).

Figure 2. Sunscreens can signicantly reduce cis-UCA formation after in vivo exposure to single or repeated UV irradiations of 100 mJ per cm2. The mean percentages of cis-UCA (6SEM) are shown for single exposure (h, n = 14) and irradiation for four consecutive days (j, n = 18). With both UV protocols all sunscreens showed signicant protection against the UV-induced increase of cis-UCA, although at different levels (p < 0.01). Signicantly, more cis-UCA was found in the sunscreen protected groups after repeated UV exposures compared with values found after a single dose (p < 0.05). Comparing the protective capacity of the sunscreens with the same SPF (B10 and T10) after a single dose showed signicantly better protection by sunscreen T10 (p < 0.01). Also after repeated UV exposures T10 showed signicantly better protection than B10 (p < 0.01). Cis-UCA percentages found with the vehicle were comparable with the percentages found after UV exposure without sunscreen.

424

VAN DER MOLEN ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Figure 3. In vivo sunscreen protection against UCA photoisomerization is correlated with the SPF. For both the organic sunscreens B and physical sunscreens T the SPF was plotted against the percentages of cisUCA (data from Fig 2) found at each SPF. Two regression lines for each sunscreen type are drawn through the observations made after single and repeated exposures (equations are shown). After a single exposure using the sunscreen type B the regression line crosses the x-axis at 6SPF 20 and after repeated UV exposure this occurs with 6SPF 23. For sunscreen type T the regression line crosses the x-axis at 6SPF 11 after a single UV exposure and at 6SPF 13 after repeated UV exposures.

Figure 4. After ex vivo exposure of skin to 200 mJ per cm2 UV all sunscreens showed signicant protection against UCA photoisomerization. The mean percentages of cis-UCA (6SEM) are shown (n = 7, for each bar). All sunscreens afforded signicant protection against the UV-induced increase of cis-UCA (p < 0.01). Comparing the sunscreens with the same SPF (B10 and T10) showed signicantly better protection by sunscreen T10 after UV exposure (p < 0.01). Cis-UCA percentages found with the vehicles were comparable with the percentages found after UV exposure without sunscreen.

The UV absorption spectrum of a sunscreen is more important than the penetration characteristics with respect to protection Sunscreen B10 and T10 offered differential protection against UCA isomerization. The main differences in characteristics between these two sunscreens are penetration properties and absorption spectra. Therefore, the inuence of the differences of these two parameters on the protective capacity against photoisomerization of UCA was studied in vivo. The rst parameter was investigated by applying the sunscreens on quartz glass, allowing UV to pass but preventing penetration of the sunscreens. Application of sunscreens on quartz glass did not result in signicant differences in protective capacity compared with application of the same sunscreen directly on the skin (Fig 5). The second parameter was examined by adding another organic sunscreen with SPF 10 (AB10) containing a broadabsorption spectrum comparable with sunscreen T10. Figure 6 shows the absorption spectra of the sunscreens AB10, B10, and T10. We found that sunscreen AB10 showed protection that was not signicantly different from T10, although it offered signicantly better protection (p < 0.01) than sunscreen B10 (Fig 5). Sunscreens show differential protection against modulation of MECLR responses ex vivo A single ex vivo UV exposure of 200 mJ per cm2 caused a signicant suppression (p < 0.01) of the MECLR responses to 17% as compared with control, nonirradiated skin (Fig 7). Both sunscreens T10 and AB10 could completely prevent this suppression. Percentages of 96% and 106%, respectively, were found, which were not signicantly different from control responses (100%). Application of sunscreen B10 on the other hand, resulted in a signicant suppression of 19% (p < 0.01). The MECLR responses of vehicle-treated skin were suppressed to a similar extent as UV-treated skin. DISCUSSION In this study we describe a rapid, noninvasive method for the determination of the protective capacity of sunscreens against the UV-induced UCA isomerization, an important endpoint for photoimmunosuppression. First we have evaluated the effectiveness of various sunscreens with different composition, such as organic and physical, narrow and broad-absorption spectra and different SPF, in in vivo human skin. After a single irradiation of 100 mJ per

Figure 5. A broad absorption spectrum is more important for good protection against UV-induced isomerization of UCA than the penetration characteristics of a sunscreen. The mean percentages of cis-UCA (6SEM) are shown (n = 7, for each bar). Values were increased from 0.8% in nonexposed skin to 51% (h) after in vivo exposure with an UV dose of 100 mJ per cm2 without quartz glass and to 50% with quartz glass (j). Again there was a signicant difference in the percentages of cisUCA when comparing the sunscreens B10 and T10 (p < 0.01). Application of T10 or B10 either directly on the skin or on quartz glass did not result in a signicant difference between the two application methods. Organic sunscreen AB10 with a broad absorption spectrum showed comparable (not signicantly different) protection as physical suncreen T10 with the same SPF of 10, although it was signicantly different from B10 (p < 0.01).

cm2, a dose by which a plateau of cis-UCA was reached in each individual that participated in the study, all sunscreens showed signicant reduction of cis-UCA formation but by various degrees. Using sunscreens with increasing SPF made it possible to predict at which SPF we would nd complete protection against isomerization at a particular UV dose. We found a difference of approximately two SPF between single and repeated UV exposures of 100 mJ per cm2, with the lowest SPF for the single dose. This probably reects accumulation of cis-UCA after repeated UV exposures (see also Krien and Moyal, 1994). A plateau is reached after several days as was shown by the nding that prolongation of the UV protocol for up to 3 wk (trice weekly) did not lead to different results when compared with UV exposure for four consecutive days (data not shown). In spite of the fact that there is some accumulation of cis-UCA after repeated UV exposures, all sunscreens still showed signicant protection compared with nonprotected skin. Besides the difference in SPF after a single or repeated UV dose of 100 mJ per cm2, we showed that for the

VOL. 115, NO. 3 SEPTEMBER 2000

SUNSCREEN PROTECTION AGAINST UCA PHOTOISOMERIZATION

425

Figure 6. UV absorption spectra of the three sunscreens with SPF 10 and vehicle. Spectra were obtained by spectrophotometric measurements of the formulations applied as a thin layer (1 mg per cm2) on quartz glass. Sunscreen AB10 () and T10 () show absorption/reection in both the UVB and UVA range. Sunscreen B10 ( ) mainly absorbs UVB wavelengths. The vehicle () shows minimal to no absorption.

organic sunscreens, with absorption mainly in the UVB range (B), an SPF was necessary that was approximately nine-values higher than for the broad-band physical sunscreens (T). To explain the higher protective capacity of the physical sunscreens (T) compared with the organic sunscreens (B), we investigated the inuence of the two main differences of these sunscreens, namely penetration characteristics and absorption spectra. Penetration characteristics of organic sunscreens in particular could lead to accumulation of sunscreens below the stratum corneum, the top layers of the skin, where they might not effectively block the isomerization of UCA (Wolf and Kripke, 1997). Furthermore, by penetration, organic sunscreens could promote the formation of cis-UCA by transferring absorbed energy to trans-UCA (Morrison et al, 1980). We have shown here, however, that it is not the penetration characteristics but the absorption spectrum that is more important with respect to protection. There was no difference in protective capacity of a given sunscreen when penetration was prevented by applying the sunscreens on quartz glass. On the other hand, the broad-spectrum physical sunscreen (T10) offered similar protection as a broadspectrum organic sunscreen (AB10) with the same SPF. The superior protection of broad-spectrum sunscreens over a narrowband UVB absorber is in agreement with other publications concerning photoimmunoprotection (Bestak et al, 1995; Damian et al, 1997) using broad-absorption spectra sunscreens. The in vivo human action spectra of UCA isomerization conrm our results concerning the importance of UVA protection.1 In another publication by Webber et al (1997), it was shown that UVA can isomerize UCA, albeit with lesser efciency. In this study a detailed doseresponse analysis of the isomerization of UCA was performed in mouse skin using UVA-I, UVA-II, and UVA-I + UVA-II wavelength ranges. Furthermore, Kammeyer et al (1995) reported that 326 nm is the most effective wavelength for in vivo UCA photoisomerization in human beings. Although the TL-12 UV lamp emits predominately UVB irradiation, a large part of its spectrum is in the 320330 UV region (spectral output of the TL12 is shown by El-Ghorr and Norval, 1999). In addition to the in vivo experiments, we have extended our study with ex vivo skin to nd out whether the results are similar. This would enable us to determine the protective capacity of sunscreens against other parameters for immunosuppression without the need to take
____________________________________________________________________________________________________

Figure 7. Sunscreens show differential protection against UVinduced modulation of the MECLR (n = 7). Data are expressed as percentages of control, nonirradiated skin (rst bar) 6SEM (*p < 0.01). The 100% control responses were 45 100 (64400) cpm. Hardly any proliferation (< 500 cpm) was found in cultures containing only responder peripheral blood mononuclear cells. After 200 mJ per cm2 MECLR responses of nonprotected skin were signicantly reduced to 17% as compared with control, nonirradiated skin. Sunscreens AB10 and T10 both completely prevented this reduction. MECLR responses with sunscreen B10 were signicantly suppressed to 81%. Responses with the vehicle were reduced, comparable with the UV irradiated skin without sunscreen.

1Gibbs NK, McLoone PR, Simics E: An action spectrum for urocanic acid isomerization in human skin in vivo. Photochem Photobiol 65:103S, 1997 (abstr.)

biopsies. An example of a frequently used endpoint is the MECLR (Hurks et al, 1997b). Therefore, we investigated both suppression of the MECLR and cis-UCA formation simultaneously. We found that a higher single dose of 200 mJ per cm2 was necessary to reach the photostationary state of UCA isomerization compared with the in vivo experiments. This higher dose could be caused by a red-shift in UCA photoisomerization as noted by others in mouse (Jones et al, 1996) and human skin (Kammeyer et al, 1995). In humans the effective wavelength in vitro was 10 to 15 nm lower than in vivo. Our data suggests this is also true when the skin is irradiated ex vivo. Additionally, there is a body site difference between the in vivo and ex vivo experiments that could explain the differences in the UCA dose response results (De Fine Olvarius et al, 1997). For the in vivo studies back skin was irradiated whereas for the ex vivo experiments breast skin from plastic surgery was used. Nevertheless, with a single dose of 200 mJ per cm2, we clearly demonstrated that the ex vivo results concerning the protective capacity of sunscreens are very similar to the in vivo ndings. In addition, using this ex vivo model, we were able to show that both the broad-spectrum sunscreens (AB10 and T10) with an SPF of 10 could prevent UVinduced modulation of the MECLR, whereas the narrow-band sunscreen (B10) with the same SPF could only partially prevent the suppression of the MECLR. Thus we can conclude that in our experiments UCA and MECLR data correlate very well. These results are in contrast to Reeve et al (1994) who showed that photoimmunoprotection by sunscreens was unrelated to cis-UCA formation, although they studied UV-induced systemic contact hypersensitivity responses in the hairless mice, which is a totally different system. On the other hand, a study performed by FinlayJones and Hart (1998) using BALB/C mice did show a correlation of cis-UCA formation and sunscreen immunoprotection of contact hypersensitivity responses. The mechanism of cis-UCA induced immunosuppression is still not clear. In previous in vitro experiments we observed only minimal to no suppression after adding pure cis-UCA to the MECLR (Hurks et al, 1997a), indicating that cis-UCA probably initiates a cascade of events that can take place in vivo and ex vivo but not in vitro. Furthermore, it indicates that the mode of action of cis-UCA is relatively quick because the cell suspension is made of the ex vivo skin immediately after irradiation. Mediators produced from the nervous system may represent one possible mechanism. Recent evidence indicates two such interactions between cis-UCA and neuropeptides, namely cis-UCA has strong homology with g-amino butyric acid (GABA) and stronger afnity for the GABA receptor than trans-UCA (Laihia et al, 1998)

426

VAN DER MOLEN ET AL

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

and it was found by Garssen et al (1998) that calcitonin-gene related peptide plays a crucial role in UV-induced immunomodulation and possibly also cis-UCA induced immunosuppression. It is not clear yet how much cis-UCA is necessary to induce immunosuppression but, given the ndings that the broad-spectra sunscreens did not completely prevent cis-UCA formation although they completely prevented the suppression of the MECLR, it seems likely that a certain threshold needs to be crossed. Further research is needed to elucidate the exact mechanism of cis-UCA-induced immunosuppression. In summary, we have shown that broad-spectra sunscreens (organic and inorganic) are superior over a narrow-spectrum sunscreen in protection against UCA photoisomerization. Furthermore, these ndings corresponded with the MECLR results. On the basis of the correlation of the protective capacity of sunscreens against cis-UCA formation and the suppression of the MECLR, in vivo UCA photoisomerization is an excellent tool to screen the immunoprotective efcacy of sunscreens. This noninvasive method allows very rapid screening of a large number of sunscreens on the same volunteer with different UV protocols, avoiding interindividual differences.
The authors wish to thank Arthur Kammeyer and Arij Weerheim for technical assistance of the UCA analysis and Dr. Mary Norval (University of Edinburgh, Scotland) for critically reading the manuscript and helpful discussions. This work was supported by grant 2817391 of the Praeventiefonds, The Netherlands.

REFERENCES
Anglin JH, Bever AT, Everett MA, Lamb JH: Ultraviolet-light induced alterations in urocanic acid in vivo. Biochim Biophys Acta 53:408409, 1961 Baden HP, Pathak MA: The metabolism and function of urocanic acid in skin. J Invest Dermatol 48:1117, 1967 Bestak R, Barnetson RSC, Nearn MR, Halliday GM: Sunscreen protection of contact hypersensitivity responses from chronic solar-simulated ultraviolet irradiation corrrelates with the absorption spectrum of the sunscreen. J Invest Dermatol 105:345351, 1995 Damian DL, Halliday GM, Barnetson RSC: Broad-spectrum sunscreens provide greater protection against ultraviolet-radiation-induced suppression of contact hypersensitivity to a recall antigen in humans. J Invest Dermatol 109:146151, 1997 De Fabo EC, Noonan FP: Mechanisms of immune suppression by ultraviolet irradiation in vivo. I. Evidence for the existence of a unique photoreceptor in skin and its role in photoimmunology. J Exp Med 157:8498, 1983 De Fine Olvarius F, Wulf HC, Therkildsen P, Poulsen T, Crosby J, Norval M: Urocanic acid isomers: relation to body site, pigmentation, stratum corneum thickness and photosensitivity. Arch Dermatol Res 289:501505, 1997 El-Ghorr AA, Norval M: The UV waveband dependencies in mice differ for the suppression of contact hypersensitivity, delayed-type hypersensitivity and cisurocanic acid formation. J Invest Dermatol 112:757762, 1999 Elmets CA, Anderson CY: Photocarcinogenesis. Mechanisms, models and human health implications. Subsection: Sunscreens and photocarcinogenesis: an objective assessment. Photochem Photobiol 63:435440, 1996

Finlay-Jones JJ, Hart PH: Photoprotection: sunscreens and the immunomodulatory effects of UV irradiation. Mutat Res Fundam Mol Mechanism Mutagen 422:155 159, 1998 Garssen J, Buckley TL, Van Loveren H: A role for neuropeptides in UVB-induced systemic immunosuppression. Photochem Photobiol 68:205210, 1998 Halliday GM, Yuen KS, Bestak R, Barnetson RSC: Sunscreens and vitamin E provide some protection to the skin immune system from solar-simulated UV radiation. Austral J Dermatol 59:7175, 1998 Hurks HMH, Out-Luiting C, Van der Molen RG, Vermeer BJ, Claas FHJ, Mommaas AM: Differential suppression of the human mixed epidermal cell lymphocyte reaction (MECLR) and mixed lymphocyte reaction (MLR) by cisurocanic acid. Photochem Photobiol 65:616621, 1997a Hurks HMH, Van der Molen RG, Out-Luiting C, Vermeer BJ, Claas FHJ, Mommaas AM: Differential effects of sunscreens on UVB-induced immunomodulation in humans. J Invest Dermatol 109:699703, 1997b yra n C, Lammintausta K, Pasanen P, Neuvonen K, Varjonen E, Kalimo K, A Janse s P: A non-invasive chamber sampling technique for HPLC analysis of human epidermal urocanic acid isomers. Acta Derm Venereol 71:143180, 1991 Jones CD, Barton AK, Crosby J, Norval M, Gibbs NK: Investigating the red shift between in vitro and in vivo urocanic acid photoisomerization action spectra. Photochem Photobiol 63:302305, 1996 Kammeyer A, Pavel S, Asghar SS, Bos JD, Teunissen MBM: Prolonged increase of cis-urocanic acid levels in human skin and urine after single total-body ultraviolet exposures. Photochem Photobiol 65:593598, 1997 Kammeyer A, Teunissen MBM, Pavel S, De Rie MA, Bos JD: Photoisomerization spectrum of urocanic acid in human skin in vitro: effects of simultaed solar and articial ultraviolet radiation. Br J Dermatol 132:884891, 1995 Krien PM, Moyal D: Sunscreens with broad-spectrum absorption decrease the trans to cis photoisomerization of urocanic acid in human stratum corneum after multiple UV light exposures. Photochem Photobiol 60:280287, 1994 Kripke ML: Immunologic unresponsiveness induced by ultraviolet radiation. Immunol Rev 80:87102, 1984 Laihia JK, Attila M, Neuvonen K, Pasanen P, Tuomisto L, Jansen CT: Urocanic acid binds to GABA but not to histamine (H-1, H-2, or H-3) receptors. J Invest Dermatol 111:705706, 1998 Mohammad T, Morrison H, HogenEsch H: Urocanic acid photochemistry and photobiology. Photochem Photobiol 69:115135, 1999 Morrison H, Avnir D, Bernasconi C, Fagan G: Z/E photoisomerization of urocanic acid. Photochem Photobiol 32:711714, 1980 Reeve VE, Boehm-Wilcox C, Bosnic M, Reilly WG: Differential photoimmunoprotection by sunscreen ingredients is unrelated to epidermal cis urocanic acid formation in hairless mice. J Invest Dermatol 103:801806, 1994 Sheehan JM, Potten CS, Young AR: Tanning in human skin types II and III offers modest photoprotection against erythema. Photochem Photobiol 68:588592, 1998 Van der Molen RG, Spies F, van't Noordende JM, Boelsma E, Mommaas AM, Koerten HK: Tape stripping of human stratum corneum yields cell layers that originate from various depths because of furrows in the skin. Arch Dermatol Res 289:514518, 1997 Walker SL, Young AR: Sunscreens offer the same UVB protection factors for inammation and immunosuppression in the mouse. J Invest Dermatol 108:133 138, 1997 Webber LJ, Whang E, De Fabo EC: The effects of UVA-I (340400 nm), UVA-II (320340 nm) and UVA-I+II on the photoisomerization of urocanic acid in vivo. Photochem Photobiol 66:484492, 1997 Wolf P, Kripke ML: Immune aspects of sunscreens. In: Gasparro FP (ed.). Sunscreen Photobiology: Molecular, Cellular and Physiological Aspects. New York: Springer Verlag, 1997, pp 99118 Young AR, Walker SL: Photoprotection from UVR-induced immunosuppression. In: Krutman J, Elmets CA (ed.). Photoimmunology. Cambridge: Blackwell Science, 1995, pp 285297