j. Soc. Cosmet. Chem., 41, 275-281 (September/October 1990) A range-finding method for approximating sunscreen efficacy and substantivity using guinea pigs JAMES J. KREUZMANN and EDWIN V. BUEHLER, Hill Top Biolabs, Inc., P.O. Box 429501, Cincinnati, OH 45242. Received January 3, 1990. Synopsis A method for approximating sunscreen efficacy using shaved/depilated guinea pigs was explored. Sun- screen-treated and untreated sites are irradiated with UVA or UVB radiation. For UVA experiments the guinea pig's responsiveness is heightened by orally administered 8-methoxypsoralen. Three untreated sites receive slightly different radiation doses covering a range centered around the expected M.E.D. Irradiation levels for sunscreen-treated sites are those for untreated sites multiplied by the sunscreen's expected protec- tion estimate. When corresponding treated and untreated sites develop a comparable erythema, the sun- screen is considered to be functioning at the expected protection estimate. If erythema levels are different, the qualitative directional deviation from the expected protection estimate is indicated. UVB testing of nine sunscreen formulations (including 8% homosalate) having known SPFs from 3 to 36 gave results in general agreement with expected protection levels. Additionally, use of the UVA method or incorporating a test of sunscreen wash-off resistance is demonstrated. The method is useful for providing rapid early guidance for formulation adjustments prior to more expen- sive and time-consuming human clinical testing. INTRODUCTION The several adverse effects of ultraviolet (UV) radiation on human skin have been well documented, although the interactive effects of the varying wavelengths are not com- pletely understood. It is clear, however, that the shorter wavelengths of ultraviolet radiation (UVB = 290-320 nm) have the highest levels of energy and are responsible for the most apparent and visible effects of overexposure. In addition to the immediate sunburn response, there is considerable experimental data to implicate long-term inten- sive exposures as resulting in skin cancer (1,3) and/or photo-aging effects (4). It is also recognized that in conjunction with other environmental exposures, UV light can pro- duce both phototoxic and photoallergic reactions (1,2). Less well known is the suppres- sive effect of UV radiation on immune function (5). With the increasing use of com- mercially available sunscreens, mostly effective in the UVB range, there is a growing concern that more intensive exposures to UVA may produce these and perhaps other adverse effects (5- 7). It is anticipated that these concerns will be addressed by the manufacturers of consumer 275

276 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS products and that testing will be required prior to commercialization. It would be particularly appropriate to have an animal model for prescreening with UVA sunscreens because to observe the biologic effects in the human requires either an extended expo- sure period or a presensitization with a phototoxic/photocarcinogenic psoralen (8,9). The guinea pig has been shown to yield well-defined, reproducible erythema to ultravi- olet B irradiation (UVB, 290- 320 nm) (10,11) and to ultraviolet A irradiation (UVA, 320-400 nm) (12), especially when predosed with 8-methoxypsoralen (8MOP) (3). The other animal model most suggested is the mouse, but this model does not develop marked erythema, and the investigator must rely on the development of edema as an endpoint evaluation. This is perhaps due to the fact that the guinea pig mast cell is a histamine releaser, whereas the mouse mast cell predominantly releases serotonin (1) the latter is known to produce edema in the rodent. Additionally, the guinea pig has a broad surface area allowing for multiple-site testing and its hair can provide an effective "blocking" template on areas where exposure is unwanted. It has also been shown to exhibit a high degree of correlation to the human response when used under both static and wash-off conditions (10, 11, 14). For these reasons, we selected it for use in preclin- ical sun protection studies. With regard to testing sunscreens for UVA efficacy, there is considerable controversy with regard to the proper approach (8,16). One of the commonly used approaches is to use the phototoxin/photocarcinogen 8MOP as a means to increase the human subjects' susceptibility to long-wave ultraviolet radiation. This should be considered an unaccep- table risk for these volunteers, except for very unusual circumstances. The other tech- niques (e.g., immediate tanning, delayed tanning) require extensive irradiation times and are expensive to run. Psoralen, dosed orally, results in a severe phototoxic reaction when animals or humans are exposed to the shorter wavelengths of UVA (320-340 nm, peak @ 335 nm). Although there is some question (8) as to whether testing at the shorter wavelengths of UVA (17) is proper for purposes of determining an SPF, it seems logical and appropriate to use this approach, since the shorter wavelengths are the most biologically active. Therefore, a methodology has been developed as a screening proce- dure that will identify the potential of a material to protect against UVB or UVA radiation in accordance with its expected clinical "sun protection factor." Since this clinical "sun protection factor" most often represents the protection value analyzed, we have defined this factor to be the "target value" for use in our assays. It must be stressed that label claims on sunscreen products are based solely on the results of human testing (14). It is not our suggestion that this screening procedure replace the human assay, but that it serve as an aid to identify raw materials and formulations worthy of subsequent evaluation at the clinical level. MATERIALS AND METHODS SUNSCREEN PREPARATIONS Eight experimental or proprietary sunscreen formulations were submitted to our labora- tory for testing. The data were to be compared to sun protection factors (SPF) derived from clinical evaluations conducted according to the FDA monograph (14). In addition, 8% homosalate, recognized by the FDA as a standard for clinical trials, was included in this investigation. Some of these materials were additionally tested in human panels by