j. Cosmet. sci., 49, 69-88 (March/April 1998) On the use of fluorescein as a probe to monitor anionic surfactant finsability from skin K. P. ANANTHAPADMANABHAN, K. K. CHAN, X. LEI, and M. P. ARONSON, Unilever Research U.S., 45 River Road, Edgewater, NJ, 07020 (K.P.A., K.K.C., M.P.A.), and Columbia University, New York (X. L.). Accepted for publication February 19, 1998. Synopsis Recent studies (1) demonstrated that the intrinsic interactions between anionic surfactants and the stratum corneurh of skin could be monitored from the displacement of a competitive protein-binding probe. The method correlated with direct measurements of surfactant binding and with clinical mildness. However, the level of strongly bound surfactant that remained after rinsing did not agree with an earlier study by Wortzman eta/. (2) in which the "rinsability" of cleansing surfactants from skin was measured using fluorescein as a marker dye for the surfactants in bar slurries. This prompted the present careful reexami- nation of the fluorescein assay. It is shown that fluorescein does not track the binding of surfactants to skin and thus cannot measure any intrinsic interactions between a cleansing composition and the skin. Thus, the assay reported by Wortzman eta/. actually measures the way fluorescein applied to skin from a solution or slurry is rinsed from the skin under unrealistic conditions employing a limited amount of ambient temperature water with no mechanical agitation. Not only do differences in fluorescein retention disappear under rinsing conditions that are more characteristic of everyday use, the differences between products primarily arise from differences in pH and counterion type and the consequent differences in fluorescein solubility. Thus, the assay measures fluorescein rinsability and does not measure the surfactant finsability! Therefore, any conclusions reached about sur- factant rinsability are erroneous, arise from an artifact of the test method, and have little to do with surfactant rinsability or mildness. INTRODUCTION In a recent study, we showed that the displacement of a fluorescence probe, ANS (1,8-anilinonaphthalene sulfonate), could be used to assess the interactions of anionic surfactants with proteins of the stratum corneum of skin (1). Probe displacement was found to correlate well with direct measurements of binding using radiolabeled pure surfactants (3) and with the clinical mildness of the bars as measured by the flex wash (4). However, the results based on ANS displacement did not correlate with the results reported by Wortzman et al. (2), who studied the "rinsability" of skin cleansing bar slurries spiked with another fluorescence probe, fluorescein. For example, ANS displace- ment indicated that a neutral pH bar based on acyl isethionate, which is known to be 69

70 JOURNAL OF COSMETIC SCIENCE mild, left considerably less surfactant strongly bound to stratum corneum proteins (i.e., resistant to thorough rinsing) than did a bar composition based on triethanolamine soaps. In contrast, Wortzman eta/. (2) concluded in their test that the TEA soap composition had better rinsability than the isethionate composition. Clearly, the two procedures and probe molecules are measuring very different phenomena. Fluorescence and molecular probe techniques have become increasingly popular in the study of the self assembly of surfactants, their interactions with proteins and polymers, and their binding at interfaces (5-13). The choice of the probe is obviously dependent upon the phenomena being monitored. For example, cooperative binding at interfaces or aggregation in solutions of surfactants can be effectively monitored by using a hydro- phobic probe such as pyrene, which has an affinity for the surfactant aggregates (5-10). A fundamentally different approach to monitoring surfactant binding is to use probes that compete with the surfactant molecule for binding sites. As mentioned earlier, one such competitive binding probe is 1,8-anilinonaphthalene sulfonate (ANS), which has been extensively used to investigate protein-ligand interactions (14-16). Since ANS is the salt of a strong acid, its spectral properties and solubility are independent of pH. Also, since ANS is anionic in character, it competes effectively with anionic surfactants for adsorption sites. Fluorescein, in contrast to ANS, is a weakly ionizable carboxylic acid that exhibits anionic character at pH values above its pKa(= 6.5) (17). Because of its pH-dependent spectral characteristics, fluorescein is widely used as a pH monitoring probe (17,18). It is not, however, clear whether fluorescein will act as a co-adsorbing/ tracking probe or as a competing probe for anionic surfactants. The inherent assumption in the Wortzman eta/. study (2) is that fluorescein tracks surfactant binding and that, therefore, higher fluorescein retention on skin reflects higher retention of cleansing surfactants on skin. The rinsability assay (2) involved rubbing a solution/slurry of the cleanser spiked with fluorescein on the skin to generate a lather, rinsing the skin by gently pouring over it a minimal amount of ambient temperature water, and quantifying residual fluorescein either visually or by measuring absorbance at 280 nm of a methanol-water extract from the treated skin. Since fluorescein is a safe and widely used probe, and the "rinsability assay" is relatively simple to carry out, it was of interest to understand what this test actually measured. It will be shown that the spectral and solution properties of fluorescein differ markedly in different surfactant solutions because of variation in pH and type of counterions, and that the detection procedure employing methanol-water extracts is subject to consider- able interference. More importantly, fluorescein does not behave as a hydrophobic probe that seeks out surfactant domains whether they be miceliar or protein-bound but, rather, functions as a very weak competitive binding probe that is easily displaced by bound surfactant. Therefore, the assay measures the way in which fluorescein rinses from skin rather than any intrinsic interactions of the surfactants with skin. It will be shown that this behavior is dominated by the differences in the intrinsic solubility of fluorescein and, in particular, by the differences in pH and counterions in the aqueous phase. Since this behavior does not reflect intrinsic interactions of the cleansing system with skin, the assay has little relevance in assessing rinsability of cleansing bar ingredients and product mildness.