302 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS The objective of this study was to investigate the relationship between interactions of a cleanser with corneum proteins and its clinical harshness. To that end, the interaction potentials of a synthetic detergent-based bar known to be relatively mild to skin were compared with those of two soap-based compositions. In human clinical studies using the flex wash (10), these three bars have been found to be significantly different in their irritation potential, with the isethionate-based bar significantly milder than the two soap-based compositions. The interactions of the cleansers with stratum corneum were compared by their ability to displace the fluorescence probe, 1-anilinonaphthalene-8-sulfonic acid (ANS), from corneum proteins. The fluorescence characteristics of ANS have been studied extensively over the years (11,12) as a selective and sensitive marker for protein interactions. Recently we have applied this technique to probe the interactions of pure surfactants with stratum corneum proteins and have analyzed the spectral changes observed in some detail (13). The probe is essentially non-fluorescent in water but becomes appreciably fluorescent when bound to proteins. Anions that are more hydrophobic than ANS can displace it from its protein binding sites, with consequent decrease in its emission intensity (14). These spectral properties have made ANS a valuable probe for studying ligand-protein binding. Using the ANS displacement technique, the location and the relative amounts ofanionic surfactants bound to corneum during a one-minute wash with the three cleansers were measured. The treatment condition was kept sufficiently short to approximate a realistic cleansing situation. Experiments were carried out with isolated human stratum corneum as well as with dermatomed porcine skin. We also measured the corneum interactions of the pure surfactants sodium lauroyl isethionate and triethanolammonium laurate, which are important constituents of the formulated compositions, and compared the results with those of sodium dodecyl sulfate, a well known harsh surfactant. The fluorescence results were correlated with direct measurements of surfactant binding to corneum, using radiolabeled materials. EXPERIMENTAL MATERIALS The nominal compositions of the three skin cleansing bars are given in Table I. Bar A has a synthetic detergent cocoyl isethionate as its main ingredient. Bar B is a high- glycerol triethanolammonium soap composition, while Bar C is a pure soap composition Table I Ingredients of Test Products Bar A: Bar B: Bar C: Sodium cocoyl isethionate, stearic acid, sodium tallowate, water, sodium isethionate, coconut acid, sodium stearate, sodium dodecyl benzene sulfonate, sodium cocoate, fragrance, sodium chloride, titanium dioxide, trisodium EDTA, trisodium etidronate, BHT Sodium soap, TEA soap, TEA, glycerin, TEA lauryl sulfate, acetylated lanolin, alcohol, water, cocodiethanolamide Sodium cocoate, sodium tallowate, water, fragrance, sodium chloride

SURFACTANT-SKIN INTERACTIONS 303 containing primarily sodium salts of tallow (a mixture of stearic and palmitic) and coco (primarily lauric) fatty acid. These bars have significantly different mildness character- istics. This was evident from the results of the flex wash (10), soap chamber test (15), and zein dissolution test (7) described in Table II. The flex wash and soap chamber tests are reproducible clinical tests that have been used to distinguish the relative irritancy potential of various soaps and synthetic detergent bars. The results of the zein dissolu- tion test (7) given in Table II are also consistent with the flex wash results and indicate the different mildness characteristics of the three bars. Note that the higher the zein dissolution is, the higher is the irritation potential of the surfactant towards skin (7). Electrophoresis purity-grade sodium lauryl sulfate (SLS) was purchased from Bio- Rad Laboratories (Richmond, CA). Sodium laurate was purchased from Eastman Chemical Company (Rochester, NY). Sodium lauroyl isethionate (SLI CH3(CH2)1oCOO(CH2)2SO •-Na +) and •4C labeled SLI were synthesized at Unilever 14 ß - ß Research. C labeled launc ac•d was obtained from Amersham International. The skin irritation tendencies of SLS, SLI, and TEA-Na laurate, as determined by the zein dissolution technique, follow the order: TEA-laurate SLS SLI (see Table III). Potassium hydrogen phthalate, potassium dihydrogen phosphate (Aldrich Chemical), and sodium borate (Fisher Scientific) were used for buffer preparations. The fluorescence probes 1-anilinonaphthalene-8-sulfonic acid (1,8-ANS) *high purity* (Figure 1A) and fluorescein (Figure lB) were obtained from Molecular Probes (Eugene, OR). Full-thickness human skin was obtained from the International Institute for the Ad- vancement of Medicine (Exton, PA). Stratum corneum was isolated by incubating strips of skin, dermis side down, into a solution of 0.5 % trypsin in phosphate-buffered saline (PBS) at pH 7.4 for 2-4 hours at 37øC. The dermis was separated from the epidermis using forceps, and the latter was incubated in fresh trypsin-PBS solution overnight at 4øC. The enzyme solution was changed several times on the following day before the stratum corneum was washed with agitation and three changes of PBS solution to remove trypsin and residual cells. The stratum corneum was then spread on spectra mesh to dry overnight under desiccation. Full-thickness porcine back skins from 3-4-week-old white albino piglets were obtained from Buckshire Corp. (Perkasie, PA). After the hair from the specimen was clipped and the underlying subcutaneous tissues were removed, the skin was dermatomed to a thickness of approximately 450-500 Table II End Point Erythema Scores in In Vivo Flex Wash (10), Zein Dissolution Results, and Irritation Scores in Soap Chamber Tests (15) for Various Bars Mean total Soap chamber erythema score test-irritation Bar (10) % Zein dissolution* scores (15) BarA 5.7 -+ 1.2 55% 0.5 Bar B 20.4** 79% 2.8 Bar C 27.7 + 2.0** 82% 2.8 * 1% Surfactant active in the test solution. ** Significantly different (p 0.05) from Bar A.