126 JOURNAL OF COSMETIC SCIENCE aqueous solution of SDS (1-200 mM) and (b) an aqueous solution of SDS (1-200 mM) + 10wt% glycerol. Specifically, an increase in the radius and/or in the number density of the aqueous pores corresponds to an increased perturbation in the skin barrier (1,6,7, 10,42). The radius and the number density of the skin aqueous pores resulting from the exposure to contacting solutions (a) and (b) above were determined using the hindered-transport model of the skin aqueous porous pathways, along with the in vitro mannitol transdermal permeability and the average skin electrical resistivity measure­ ments. For completeness, we also conducted similar measurements on p-FTS, which was exposed to: (c) the PBS control, and (d) 10 wt% glycerol aqueous contacting solutions. In Figure 6, we have plotted the log of the mannitol transdermal permeability, P (cm/h), against the log of the average skin electrical resistivity, R (kohm-cm2), over the same exposure time, exhibited by p-FTS samples exposed to solutions (a), the diamonds, and (b), the triangles, above. Each diamond/triangle represents a log P value of one p-FTS sample at steady state and the corresponding log R ( the log of the average skin electrical resistivity value). The slopes of the best-fit curves resulting from linear regressions, the dashed line for (a) and the solid line for (b), are not statistically different from the theoretically predicted slope value of -1, thereby indicating consistency with the hin­ dered-transport aqueous porous pathway model analysis for p-FTS samples exposed to contacting solutions (a) and (b) above (6,7). Also, note that the dashed line has a larger intercept value than that corresponding to the solid line, which reflects a larger average pore radius, r pow for p-FTS samples exposed to (a) than to (b). Having determined r pore • the pore number density was determined using equation 6, in which all the parameters, except eh, the pore number density, are known in advance (recall that dX = 15 µm) -2.15.---------------------------------------, -3.25 - -3.50 -:: -3.15 -4.00 -4.25 -4.50 -4.75.._ ___ ....__ ___ ___. ___________ __,_ _______________......______. 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 log R (kohm-cm2) Figure 6. Experimental correlation between the in vitro mannitol transdermal permeability, P (cm/h), and the in vitro skin electrical resistivity, R (kohm-cm2), exhibited by p-FTS samples exposed to an aqueous solution of SDS (1-200 mM), the diamonds, and to an aqueous solution of SDS (l-200 mM) + 10 wt% glycerol, the triangles. Each data point corresponds to a log P value of one p-FTS sample at steady state and the associated log R, the log of the average skin electrical resistivity value over the same time period. The slopes of the best-fit curves resulting from a linear regression are: (i) -0.9768 ± 0.06 for SDS (1-200 mM), with R2 = 0.9636, shown as the dashed line, and (ii) -1.0453 ± 0.06 for SDS (1-200 mM) + 10 wt% glycerol, with R2 = 0.9653, shown as the solid line. Note that these slope values are not statistically different from the theoretically predicted value of -1.

SDS MICELLES IN SKIN BARRIER PERTURBATION 127 (6,7). Using the model described above, we found that the average pore radius does not depend on the SC thickness, LiX, while the pore number density is directly proportional to LiX. The aqueous pore number density, e!T, values resulting from exposure of the p-FTS samples to contacting solutions a-d above were normalized by the e/'T value resulting from exposure of the p-FTS samples to the PBS control solution, solution (c), which served as the baseline, and have been denoted as (e/T)00rmal (see Appendix, where we illustrate how to obtain r pore and (e/T)00rmal for p-FTS samples exposed to (a)). Our deduced values of r po re and (e/'T)00rmal corresponding to solutions a-d above are reported in Table I. As can be seen, the average pore radius, r po w corresponding to (a) is 33 ± 5A, while that corresponding to (b) is 20 ± 5A, which is similar to the average pore radius corresponding to (c), 20 ± 3A. In addition, the normalized pore number density, (e!T)00rmaI, corresponding to (a), 7 ± 1, is about twice that corresponding to (b), 3 ± 1. Interestingly, we also see that a 10 wt% glycerol aqueous solution (contacting solution d) reduces r pore and (e/'T) 00rmal by about 50% relative to the PBS control. The results in Table I indicate that an SDS aqueous contacting solution containing micelles, in the presence of 10 wt% glycerol, induces a lower extent of skin barrier perturbation, as reflected in the lower average pore radius and normalized pore number density, when compared to an SDS aqueous contacting solution, in the absence of glycerol. In fact, in the absence of glycerol, an SDS micelle of 19.5 ± lA hydrodynamic radius experiences no steric hindrance in penetrating through aqueous pores in the SC that have an average pore radius of 33 ± 5A (see Table I). However, in the presence of 10 wt% glycerol, an SDS micelle of 18.5 ± lA hydrodynamic radius experiences sig­ nificant steric hindrance in penetrating through smaller aqueous pores in the SC that have an average pore radius of 20 ± 5A (see Table I). Moreover, the presence of 10 wt% added glycerol in the SDS aqueous contacting solution reduces the (e!T) 00 r mal value from 7 ± 1 to 3 ± 1, which is more than a 50% reduction in the normalized pore number density. Hence, adding 10 wt% glycerol to an aqueous SDS micellar contacting solution minimizes the micellar contribution to SDS skin penetration in vitro by minimizing both the average pore radius and the pore number density of the skin aqueous pores. The results of this study indicate that the data is consistent with hypothesis (iii): Glycerol reduces both the radius of the aqueous pores in the SC relative to that of the SDS micelles, as well as the aqueous pore number density, which if not reduced, would allow SDS micelles to contribute to SDS skin penetration in vitro. POSSIBLE STRUCTURAL MODES OF INTERACTION OF GLYCEROL AND SDS WITH THE SKIN BARRIER Our results indicate that the addition of 10 wt% glycerol to an aqueous contacting solution of SDS mitigates skin barrier perturbation in vitro by reducing the skin aqueous pore radius and the aqueous pore number density. We propose two scenarios to ratio­ nalize these results. According to the first scenario, it is well-accepted that because of its strong hygroscopic property and ability to modulate water fluxes in the SC, glycerol can diffuse into the SC and bind water within the SC (24,28,29). In fact, researchers have observed a significant positive correlation in vivo between the skin-moisturizing ability of glycerol, as determined through skin conductance measurements, and the correspond­ ing amount of glycerol found in the skin barrier (52). As a result, water binding by glycerol in the SC reduces the mobility of water within the SC. The limited mobility of