Prodrug-Based Topical Vaginal Drug Delivery: a Physical Model Approach to the Simultaneous Membrane Transport and Bioconversion.

Abstract

A methodology for the investigation of drug transport and metabolism in the vaginal membrane was developed using mouse as animal model in two stages. The first stage involved the development of a basic physical model using tracer levels of drugs. During the second stage, predictions of actual drug efficacy were made by extending the model to high drug concentration levels. A simple laminate model for describing simultaneous drug transport and bioconversion at steady state was adopted in the first stage of this study. This physical model has the following properties: the vaginal membrane was assumed to be composed of two or three layers depending on the estrous phase, the drug diffusivities and the enzymatic distributions were assumed to be uniform within each layer, the enzymatic activities were approximated with first-order kinetics, and the drug molecules permeated the vaginal membrane via a lipoidal pathway or via an aqueous pore pathway which was enzyme-free. The model transport parameters, i.e. drug diffusivities, were first determined with permeation and desorption experiments. It was found that drug diffusivities in the mouse vaginal membrane changed significantly with the estrous cycle. This finding is believed to be the first of its kind. The enzymatic activities in each layer of the vaginal membrane were then obtained from "go-through" experiments by computer iteration according to the model. The esterase and adenosine deaminase activities were found to be much higher in the epithelium than in the lamina propria and the muscular layer. A technique for separating the epithelium from the lamina propria was developed for the purpose of independently determining the enzymatic activities in each layer of the membrane. The validity of the model was confirmed by showing the enzymatic activities obtained from this independent assay were in good agreement with those deduced from the computer iteration procedure. After the simple laminate model was proven to be suitable for the describing simultaneous transport and bioconversion of tracer levels of ara-A and its prodrugs in the vaginal membrane, it was modified to include the Michaelis-Menten kinetics for both enzymes and the prodrug inhibition of the adenosine deaminase for the situation of high drug concentrations. The appropriateness of these modifications were confirmed by conducting "go-through" experiments using saturated solutions and comparing the experimental fluxes of drug species with the theoretical values. Finally, drug concentration-distance profiles in the vaginal membrane were constructed for ara-A and its prodrugs. To simulate in vivo situations, the drug applied on the surface of the membrane was assumed to be in the form of a saturated aqueous solution and the profiles were plotted based on the modified model. Predictions of the antiviral efficacy of the prodrugs were made from these concentration profiles and literature information on the location of active herpes simples virus infection in the epithelium and the minimal inhibitory concentration (MIC). Ara-A and ara-A-5'-O-octanoate were predicted to be ineffective during all phases of the estrous cycle. Ara-A-5'-O-valerate (V-ara-A) may deliver ara-A above the MIC levels at the 75% depth position in the epithelium for the high permeability diestrus membranes, but gives only marginal or lower than MIC ara-A levels in membranes with intermediate or low permeabilities. Ara-A-5'-O-acetate, which has a comparable solubility but a much slower enzymatic hydrolysis rate than V-ara-A, can achieve a much higher ara-A level at the 75% depth position in the membrane with intermediate and low permeabilities.