122 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS STRATUM DONOR CORNEUM VIABLE SKIN RECEPTOR I I I I I I I DRUG DRUG Reservoir Diffusion Barrier METABOLITES Sink Condition Figure 2. Bilayer diffusion/bioconversion model for percutaneous absorption. The drug concentration on the surface of the skin is assumed to be constant during the period of skin penetration (skin-controlled transdermal drug delivery system). The governing equations and boundary/initial conditions have been described in detail in reference (5). not been elucidated. The penetration profile of vitamin C was theoretically analyzed by the dynamic bilayer-skin diffusion/bioconversion model (5). The diffusivity and the partition coefficient in each skin layer (stratum corneum and viable skin) were deter- mined from the lag times and the steady-state fluxes on the basis of the steady-state bilayer skin model (10). By assuming a homogeneous distribution for tissue vitamin C (initial concentration), the penetration profile was simulated (Figure 3) and compared with the experimental one. The initial tissue concentration of vitamin C was found to be approximately 2.7 }xmol/ml in the viable skin of the present animal model. The rate constant of vitamin C bioconversion in the skin was also found to be very small (1.5 ( 10 -5 s-1) compared to the skin bioconversion of vitamin E discussed in the next sec- tion of this article. VITAMIN E PENETRATION The penetration profiles of vitamin E are plotted in Figure 4. The significant difference is easily observed in the penetration profiles between HPLC assay and radioactivity counting. The radiolabeled vitamin E penetrated promptly across the skin, while the nonlabeled compound appeared after a remarkably long lag time (about 48 hours). We found previously that vitamin E bioconverted in the viable skin from a provitamin appeared in the receptor compartment after a long lag time (24-36 hours) (8). We therefore suggested that vitamin E bioconverted in the viable skin might diffuse back into the stratum corneum very slowly. If this is the case, the penetration profile not only of the radiolabeled vitamin E but also of the nonlabeled compound should provide a long lag time. Contrary to our previous hypothesis, the significant difference appears in the penetration profiles between radiolabeled and nonlabeled compounds. The long

SKIN PENETRATION 123 200 o o 6 12 18 24 30 36 Time (h) Figure 3. Penetration profile of vitamin C across hairless mouse skin. Key: ([•) HPLC assay. (O) Radioac- tivity counting. - .... Calculated (initial concentration in viable skin C o = 2.7 •mol/ml, no biocon- verson) ....... Calculated (C o = 0, no bioconversion). -- Calculated (initial concentration in viable skin C o = 2.7 •xmol/ml, first order bioconversion kinetics k o = 1.5 x 10-5 s-•, enzyme decay rate constant A = 5.6 x 10 -6 s-•). The diffusivity and the partition coefficient both in the stratum corneum and in the viable skin were determined from the penetration profiles of radiolabeled vitamins across the intact and stripped skins. The detailed procedure has been described in reference (10). Diffusivity in stratum corneum = 7.6 X 10 -• cm2/s. Diffusivity in viable skin = 5.8 x 10 -8 cm2/s. Partition coefficient between stratum corneum and viable skin = 0.28. Concentration on the skin surface = 136 •xmol/ml. Thickness of stratum corneum = 10 •m. Thickness of viable skin = 370 •tm. lag time in the penetration profile assayed by HPLC therefore suggests the extensive metabolism of vitamin E in the skin. At this stage of research on percutaneous absorp- tion of vitamins, little is known with respect to the skin metabolism. However, Shira- tori suggested that the skin may be an important storage site for vitamin E and play a major role in distribution and metabolism of vitamin E (11). After about 48 hours, vitamin E appeared gradually in the receptor solution. This is due to the fact that skin enzyme becomes gradually deactivated under the in vitro condi- tion. The enzyme which is responsible for esterification of estradiol esters was found to degrade in the hairless mouse skin under the in vitro condition by following the expo- nential decay law (12). Assuming the exponential decay law for the activity of skin enzymes, k = koexp( - At) (1) where A is the decay rate constant, k o is the intrinsic rate constant for bioconversion, and t is the time, the experimental profiles of vitamin E penetration were described by the present model. The results are shown in Figure 4 where the calculated profile based on the constant activity of enzymes (A = 0) is also plotted for comparison. The calcu-