j. Soc. Cosmet. Chem., 43, 37-48 (January/February 1992) Simulation studies of skin permeation JOEL L. ZATZ, Department of Pharmaceutics, Rutgers College of Pharmacy, Piscataway, NJ 08855-0789. Received January 5, 199•. Presented at the International Cosmetic Science Meeting, October 1991. Synopsis Simulation studies were conducted using the multicompartmented membrane model to determine the effect on permeation of binding to the stratum corneum and washing of the skin surface. Under infinite dose conditions, binding lengthens the lag period but does not change the steady-state flux. Amounts pene- trating at any arbitrary time are reduced. Under finite dose conditions, binding causes significant reductions in the amount reaching the sink at a given time. There is a 50-fold decrease in the amount penetrated at 12 hours, when binding changes from 70% to 90% bound. Washing of the skin surface reduces the amount penetrated, but significant quantities may still get through the skin. At very short wash times, skin and blood concentrations are nearly proportional to contact time. Peak blood concentrations occur hours after washing, reflecting continuing permeation from the stratum corneum. Following a 30-minute application, the amount excreted during four days is proportional to the amount found in the stratum corneum at 30 minutes, provided that the transport coefficient through the stratum corneum and the elimination rate constant are not too small. If partition coefficient is the only variable, both total amount excreted and peak blood concentration are proportional to the amount found in the stratum corneum at 30 minutes. INTRODUCTION Simulation of percutaneous absorption can be used to predict penetration behavior and explore permeation mechanisms. Several models, diffusional and compartmental, have been used to describe the permeation of exogenous compounds into and through the skin. Approaches for estimating percutaneous absorption from physical-chemical data have been reviewed (1). Hadgraft modeled permeation through skin as a diffusion process and included the uptake by capillaries (2). He showed that a high oil/water partition coefficient favored skin retention and therefore increased reservoir function. Specific binding to skin com- ponents was not considered. Chandrasekaran et al. described the dual-sorption model of skin uptake, which divides permeant molecules into those that are "dissolved" in the stratum corneum, and thus free to diffuse, and those that are bound to membrane components (3). Binding was assumed to follow a Langmuir-type relationship. This model, applied to percutaneous absorption from an "infinite dose," accounted for the fact that diffusion coefficients based on the lag time differed in value from those calculated from the equation for 37

38 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS steady-state permeation. Experimental evidence suggested that the drug was bound to non-lipid components, perhaps proteins. Decontamination of skin exposed to toxins to prevent absorption has been the subject of several reports (4). The amount absorbed depends on the time elapsed after exposure as well as the washing treatment utilized. In some cases, solvents intended to remove contaminants serve also to compromise the skin barrier and are therefore counterpro- ductive. There are a number of experimental difficulties in conducting washing exper- iments. The importance of the simulation approach is that it eliminates experimental artifacts from consideration and describes the effect of cleaning the skin surface on skin permeation under ideal conditions. A previous report described a simulation model (multicompartmented membrane model) in which the stratum corneum (SC) is sectioned into component lamina and the permeation process is treated as the sum of a series of first-order transfers (5). Briefly, the model consists of a donor, five sequential stratum corneum compartments, a single aqueous tissue compartment, and a sink. The model was used to illustrate the effect of changes in donor volume on skin distribution and uptake. The multicompartmented membrane model has been extended to treat the questions of stratum corneum binding and skin washing (to remove an application from the skin surface after a fixed period of time). Intuitively, we would expect binding to the stratum corneum to affect absorption rate. However, the degree to which this occurs is not easily evaluated because it is difficult to design an experiment in which binding is the only variable. Simulated data provide a workable means for estimating the extent to which binding affects penetration. For the binding studies, a series of "dead-end" compartments was added to the basic model to represent bound permeant. The idea is a variation of the dual-sorption model previously described (3). However, it is more versatile in that it is not limited to infinite dose simulation. For the washing studies, the sink was replaced by a systemic central compartment (representing the body) from which excretion occurs by a single first-order process. With this variation in the model, it is possible to simulate blood concentration and urinary excretion if appropriate values of the elimination rate constant and volume of distribu- tion are selected. Other, more complex pharmacokinetic designs may also be employed. MODEL PARAMETERS Major model variables include the intercompartment transfer constant, K, which is analogous to a stratum corneum diffusion coefficient, and the ratio of rate constants describing transfer between donor and the first stratum corneum compartment, which represents a partition coefficient (PC). Permeant concentration and donor volume are also adjustable. In the binding studies, it was assumed that the fraction bound was constant. This corresponds to the low-concentration region of a sorption isotherm. It is also possible to model the entire isotherm by specifying the saturation binding value and an affinity constant. Several transfer coefficients in the models were kept constant. The transfer coefficients