0 400 800 1200 1600 2000 2400

Figure 12 Predicted versus experimental values for corneal steroid permeability as a function of partition coefficient.

The permeability coefficient K^ is just the flux divided by Cw. It is apparent that the permeability coefficient is linear with P for small distribution coefficients and constant for large P. Thus, for small P the epithelium is the barrier, and for large P the stroma is the barrier. A fit for steroid permeability is shown in Figure 12, where the regression analysis gave De= 1.4 x 10~9 cm2/sec and Ds = 2.0 x 10~6 cm2/sec for Ze = 4 x 10~3 cm and 4 = 3.6 x 10~2 cm (228). These values for the diffusion coefficients are reasonable compared with those of aqueous gels and lipid membranes.

A simple estimate of the diffusion coefficients can be approximated from examining the effects of molecular size on transport through a continuum for which there is an energy cost of displacing solvent. Since the molecular weight dependence of the diffusion coefficients for polymers obeys a power law equation (229), a similar form was chosen for the corneal barriers. That is, the molecular weight (M) dependence of the diffusion coefficients was written as

Using regression analysis on a data set of about 50 different molecules, it was found that a = -4.4, Ô = -0.5, De(0) = 12 cm2/sec, and Ds(0) = 2.5 x 10"5 cm2/sec (19). A graphic representation of the effect of relative molecular mass (Mr) and distribution coefficient on corneal permeability is shown in Figure 13. One observes a rapid reduction in permeability coefficient with decreasing P and increasing Mr. The addition of pores to the model, a mathematical construct, is necessary to account for permeability of polar molecules such as mannitol and cromolyn. These would also be required for correlating effects of compounds, such as BAC, which may compromise the epithelial barrier by increasing the volume of the extracellular space.

Another perspective provided by this model is the effect of three physicochemical parameters: solubility, distribution coefficient, and molecular mass on transcorneal flux. All of these properties can be influenced by molecular design. The effects of these properties are illustrated in Figure 13 in which the logarithm of the flux is plotted as a

0 0

Post a comment