Cutaneous metabolism, tissue binding, lipophilicity and electroosmosis : relevance for iontophoretic enhancement

In this work passive and iontophoretic transport of a dipeptide, Tyrosine-Phenylalanine (TyrPhe), a protected amino acid, Tyrosine-β-Naphthylamide (Tyrβ-NA), and a glucose derivative, Benzyl-2-Acetamido-2-deoxy-α-D-glucopyranoside (BAd-α-Glc), was investigated across heat-separated human epidermis in vitro. The modified Nernst-Planck Equation was used to theoretically describe permeation of these compounds. Permeation experiments were carried out with custom-made glass diffusion cells connected to a four-electrode system developed in our laboratories to generate constant voltage. Ag/AgCl electrodes with the anode in the donor compartment were used to apply the potential difference across the skin and reference Ag/AgCl electrodes, reaching either side of the membrane with Luggin capillaries, were used to precisely control and maintain constant the potential drop across the membrane. Permeation was studied at pH 3 and 4.5, for which the TyrPhe and Tyrβ-NA are positively charged.
The contribution of charge, lipophilicity, skin retention, electroosmosis and the influence of skin metabolism on the permeation rate of the above mentioned substances was examined.
pKa of TyrPhe and Tyrβ-NA was determined by potentiometric titration to be 3.50 and 4.78, respectively. The average net ionic valence of TyrPhe was 0.76 and 0.091 at pH 3 and 4.5, respectively and that of Tyrβ-NA was 0.98 and 0.66 at pH 3 and 4.5, respectively. n-octanol/aqueous buffer partition coefficients were determined using shaking flask method to be for BAd-α-Glc 0.13 and 0.11 at pH 3 and 4.5,respectively and for TyrPhe 0.80 and 0.61 at pH 3 and 4.5, respectively and for Tyrβ-NA 4.81 and 7.28 at pH 3 and 4.5, respectively. Enzymatic stability of TyrPhe was determined by reflection boundary set-up. A degradation of 12% at pH 3 and 45% at pH 4.5 was measured over 69 hours at 37°C. For BAd-α-Glc the generation of less than 0.5% of benzyl alcohol, the product of glycosidase action, was found. Tyrβ-NA was stable because of the side-chain protecting group.
A typical permeation experiment involved three experimental stages. After a first passive stage over 44 hours, 250 mV were applied for 3 hours followed by a second passive stage of 22 hours.
Membrane resistance during the three hours of iontophoresis remained constant or showed only small decline. These results guaranteed the integrity of the membrane under the present experimental conditions.
The influence of TyrPhe skin metabolism on its own permeation rate and skin retention as well as on the permeation rate and skin retention of BAd-α-Glc was examined. No reproducible permeation of TyrPhe at pH 3 and no permeation at all at pH 4.5 was measured when the dipeptide was used alone or in combination with BAd-α-Glc. Additionally, higher fluxes of the degradation products Tyrosine (Tyr) and Phenylalanine (Phe) were observed and higher levels of Tyr and Phe were recovered from the epidermis compared to blank runs. Blank runs reflected endogenous levels of Tyr and Phe. Degradation was inhibited by the use of o-phenanthroline at pH 3 but not at pH 4.5 and at both pH values at low temperature (4°C). Inhibition was verified by reproducible TyrPhe permeation and blank level Tyr and Phe fluxes and tissue recovery. These results confirm degradation of TyrPhe by cutaneous metabolism in heat-separated human epidermis, that was stronger at pH 4.5 compared to pH 3. BAd-α-Glc showed at both pHs smaller iontophoretic fluxes compared to the passive ones, indicating an electroosmotic flux from cathode to anode. In combination with TyrPhe, iontophoretic fluxes of BAd-α-Glc increased significantly when TyrPhe was metabolized in the tissue, while no such decrease was observed when TyrPhe metabolism was inhibited. This increase of BAd-α-Glc iontophoretic flux was accompanied by a considerable decrease of BAd-α-Glc amount retained in the epidermis. The generated Tyr and Phe appear, therefore, to decrease BAd-α-Glc amount retained in the epidermis and enhance iontophoretic flow. Thus, an interaction between the concurrent permeants at the level of tissue retention induced by metabolism can influence iontophoretic permeation of these poorly permeable compounds. So far only Tyr and Phe were able to show this effect. On the other hand, Tyr and Phe did not affect iontophoretic permeation of TyrPhe and Tyrβ-NA. This demonstrates that the effect of Tyr and Phe is not universal but depends on the permeant.
The effect of lipophilicity and electroosmosis was assessed simultaneously for all three permeants, which spanned a 70-fold range of lipophilicity, in the absence of extraneous Tyr and Phe using a theoretical model for the enhancement factor. This was based on the modified Nernst-Planck Equation, extended by the ratio of permeability coefficients for lipid and aqueous pathway and including an estimate of the ionic valence of permeants in the aqueous domain of the tissue. At pH 4.5 a very weak electroosmotic flow from cathode to anode was observed, indicating an isoelectric point of the skin slightly above 4.5. At pH 3 electroosmotic flow was approximately 10-fold stronger than at pH 4.5. The ratio of the permeability coefficients for lipid and aqueous pathways resulted for all compounds at both pHs in values much smaller than 1, indicating that the aqueous pathway dominated transepidermal permeation. The simultaneous analysis of the results of all substances shows that the applied model evaluation affords a quantitative assessment of the effect of ionic valence, lipophilicity and electroosmosis and makes possible to predict the outcome of iontophoresis.