15 according to Eq. (A.6). The log Ppara, log Pfilter, and log PABL were added as fixed contributions, as log P0 see more and log Puptake were refined ( Appendix A.5) for the non-inhibitor and added-inhibitor (50 μM PSC833) sets. Both the intrinsic and the uptake permeability values appeared to be affected by efflux ( Table 3). The two sets were

then combined, with the repeated refinement yielding log P0 = −5.28 ± 0.04, log Puptake = −5.73 (kept fixed), and log Pefflux = −5.80 ± 0.04 for the non-inhibitor set and log Pefflux < −8 for the +50 μM PSC833 set. This suggested that efflux was essentially suppressed by the inhibitor. With the log Pefflux of −5.80, it was possible to rationalize the extent to which the individual-set refined log Puptake and HIF inhibitor log P0 in the two sets were different. Fig. 4c and d shows colchicine and digoxin with added efflux inhibitor (checkered circle) and no-inhibitor (black circles). The addition of inhibitors increases the apparent permeability by nearly the same amount in both drugs, consistent with the suppression of efflux

transporter. To assess the ability to predict in vivo BBB permeability of a compound from permeability data measured using the PBEC model, P0 (in vitro) derived from our PBEC model permeability data was plotted against P0in situ (in vivo) derived from in situ brain perfusion data in rodents ( Fig. 5). Published data from other in vitro porcine BBB models were also included in the linear regression analysis. The r2 value Terminal deoxynucleotidyl transferase of 0.61 shows a good correlation for the pooled data. The in vitro blood–brain barrier

(BBB) model from primary porcine brain endothelial cells (PBEC) which shows a restrictive paracellular pathway was used for permeability studies of small drug-like compounds of different chemistry: acid, bases, neutrals and zwitterions. Assay at multiple pH was conducted for the ionizable compounds propranolol, acetylsalicylic acid, naloxone and vinblastine to plot permeability vs. pH. The pCEL-X software (Section 2.5 and Appendix A) was used for detailed permeability data analysis, including aqueous boundary layer (ABL) correction. The ABL was found to restrict propranolol permeability, which was also limited by low pore density of the Transwell®-Clear polyester filter membrane. The intrinsic transcellular permeability P0 showed good correlation with in situ data, indicating the predictive power of the in vitro model. Stirring helps to diminish the ABL thickness, but it cannot reduce it entirely. This is because the aqueous medium adjacent to the membrane surface is less mobile due to hydrogen bonds formed at the interface (Loftsson and Brewster, 2008). Hence, even vigorous stirring is unable to remove the ABL totally. Furthermore, excessive stirring is undesirable, since it can compromize tight junction integrity (cf., Zhang et al., 2006: 600 RPM). Application of the pKaFLUX method for ABL correction using pCEL-X proved useful particularly for ionizable compounds.