Wj. Roberts et Kb. Sloan, Correlation of aqueous and lipid solubilities with flux for prodrugs of 5-fluorouracil, theophylline, and 6-mercaptopurine: A Potts-Guy approach, J PHARM SCI, 88(5), 1999, pp. 515-522
The Potts and Guy equation that has been used to predict permeability coeff
icients for molecules being delivered from aqueous vehicles has been transf
ormed to accommodate lipid vehicles that are less polar than skin, and pola
r vehicles that are less polar than water. Solubilities in pH 4.0 aqueous b
uffer (S-AQ), solubilities in isopropyl myristate (S-IPM), and molecular we
ights (MW) of prodrugs of 5-fluorouracil (5-FU), theophylline (Th), and 6-m
ercaptopurine (6-MP) have been regressed against their fluxes from suspensi
ons in IPM (J(M)). Seven series (n = 39) of alkylcarbonyloxymethyl (ACOM),
alkyloxycarbonyl (AOC), alkylcarbonyl (AC), and alkylaminocarbonyl (AAC) pr
odrugs were used to determine the best fit to the transformed Potts and Guy
equation (eq 6): log J(M) = x + y log S-IPM + (1 - y)log S-AQ - z MW. The
estimated values for x, y, and z were -0.193, +0.525, and +0.00364, respect
ively, with r(2) = 0.945 for n = 39. Inclusion of a miscellaneous series co
mprised of the parent drugs and a branched alkyl chain prodrug gave an equa
lly good fit only if 6-MP was excluded from the analysis. The best performe
r (largest J(M)) in each series was usually correctly identified. The value
s for x, y, and z were consistent with values obtained by Potts and Guy, bu
t the inclusion of the (1 - y)log S-AQ term in eq 6 and the value for y, sh
ows that water solubility is almost as important as lipid solubility in pre
dicting flux. There were no significant changes in predicted log J(M) or x(
i) for each series if their log J(M) or x(i) were calculated using y and z
coefficients obtained for solutions to eq 6 from which the data for the ser
ies had been excluded. This suggests that the data from all the series is h
omogeneous. Data from Kasting, Smith, and Cooper for S-IPM, S-PG, and MW of
unrelated molecules were regressed against their fluxes from propylene gly
col (PG) using eq 7: log J(M) x + y log S-IPM + (1 - y) log S-PG - z MW. Th
e estimated values for x, y, and z were -1.673, +0.599, and +0.00595, respe
ctively, with r(2) = 0.852 for n = 28. These values for x, y, and z are als
o consistent with those previously reported by Potts and Guy, and, together
with the results for fluxes from IPM, show the general utility of the tran
sformed Potts and Guy equation in predicting flux from vehicles other than
water and in showing the importance of solubility in a polar solvent as wel
l as a nonpolar solvent in predicting flux.