Fig. 1. Logarithm of the apparent permeability vs. lipophilicity, based on data from four literature studies (Caldwell et al., 1998, Yazdanian et al., 1998, Irvine et al., 1999 and Hilgers et al., 2003). The highest observed permeabilities appear to be aqueous boundary layer limited, but the trends are not readily apparent, based on the stirring rates used.

Fig. 2. The logarithm of permeability vs. pH for verapamil, based on PAMPA data. The unfilled circles correspond to unstirred effective permeability data, and the filled circles correspond to stirred data. The dashed curves refer to the membrane permeability (effective values corrected for the aqueous boundary layer). The dotted lines refer to the aqueous boundary layer permeability coefficients under the two sets of conditions.

Fig. 3. Logarithm permeability vs. pH profiles for (a) alprenolol, and (b) propranolol. In both plots, the unstirred PAMPA data are indicated by unfilled circles. In both plots, the filled circles are based on data taken from reference (Adson et al., 1995). In (a), the filled squares are based on data taken from reference (Yamashita et al., 2000), and in (b), the filled squares are based on data taken from reference (Pade and Stavchansky, 1997).

Fig. 4. The logarithm of permeability vs. pH for cimetidine, based on (a) PAMPA measurement, (b) Caco-2 measurement (Nagahara et al., 2004), and for alfentanil, based on (c) PAMPA data, and (d) Caco-2 data (Nagahara et al., 2004). The large thick-line unfilled circles in (b) are apparent permeability coefficients less the estimated paracellular permeability coefficients. This corrected data was used in the refinement. The uncorrected data are shown in thin-line small circles.

Fig. 5. Logarithm permeability vs. pH profiles of several β-blockers: (a) atenolol (Neuhoff et al., 2003), (b) metoprolol (Neuhoff et al., 2003), (c) propranolol (this work). The filled circles refer to PAMPA results. The small unfilled circles refer to the apparent permeability values, and the larger unfilled circles refer to the same data corrected for the estimated ion permeability.

Fig. 6. Caco-2 permeability of verapamil: (a) without BSA in the receiver compartment, and (b) with 4% wt./vol. in receiver compartment. The filled squares refer results from plates stirred at 700 rpm; the unfilled circles refer to results from plates stirred at 100 rpm. The dashed curves refer to transcellular permeability, free of the effects of the aqueous boundary layer, as well as ion transport. Dotted lines refer to the aqueous boundary layer permeability and solid curves are the best-fit lines to the apparent permeability measured. Below pH 6.0, the monolayer appears to be more permeable than expected, based on the pH partition hypothesis.

Fig. 7. Prediction of the intrinsic Caco-2 permeability, using intrinsic PAMPA permeability, calculated apparent partition coefficients (AB/log D, pH 7.4), and Abraham's H-bond acidity, α. The data used in the analysis are listed in Table 5.

Physicochemical properties of the bases^a

Caco-2 ion/paracellular permeability analysis (pH 7.4)^a

Caco-2 analysis^a

Caco-2 hydrodynamic analysis: log P_ABL = log K + α log ν

Prediction of Caco-2 intrinsic permeability using PAMPA data