Immobilized Artificial Membrane HPLC Derived Parameters vs PAMPA-BBB Data in Estimating in Situ Measured Blood-Brain Barrier Permeation of Drugs.

The affinity indexes for phospholipids (log kW(IAM)) for 42 compounds were measured by high performance liquid chromatography (HPLC) on two different phospholipid-based stationary phases (immobilized artificial membrane, IAM), i.e., IAM.PC.MG and IAM.PC.DD2. The polar/electrostatic interaction forces between analytes and membrane phospholipids (Δlog kW(IAM)) were calculated as the differences between the experimental values of log kW(IAM) and those expected for isolipophilic neutral compounds having polar surface area (PSA) = 0. The values of passage through a porcine brain lipid extract (PBLE) artificial membrane for 36 out of the 42 compounds considered, measured by the so-called PAMPA-BBB technique, were taken from the literature (P0(PAMPA-BBB)). The values of blood-brain barrier (BBB) passage measured in situ, P0(in situ), for 38 out of the 42 compounds considered, taken from the literature, represented the permeability of the neutral forms on "efflux minimized" rodent models. The present work was aimed at verifying the soundness of Δlog kW(IAM) at describing the potential of passage through the BBB as compared to data achieved by the PAMPA-BBB technique. In a first instance, the values of log P0(PAMPA-BBB) (32 data points) were found significantly related to the n-octanol lipophilicity values of the neutral forms (log P(N)) (r(2) = 0.782) whereas no significant relationship (r(2) = 0.246) was found with lipophilicity values of the mixtures of ionized and neutral forms existing at the experimental pH 7.4 (log D(7.4)) as well as with either log kW(IAM) or Δlog kW(IAM) values. log P0(PAMPA-BBB) related moderately to log P0(in situ) values (r(2) = 0.604). The latter did not relate with either n-octanol lipophilicity indexes (log P(N) and log D(7.4)) or phospholipid affinity indexes (log kW(IAM)). In contrast, significant inverse linear relationships were observed between log P0(in situ) (38 data points) and Δlog kW(IAM) values for all the compounds but ibuprofen and chlorpromazine, which behaved as moderate outliers (r(2) = 0.656 and r(2) = 0.757 for values achieved on IAM.PC.MG and IAM.PC.DD2, respectively). Since log P0(in situ) refer to the "intrinsic permeability" of the analytes regardless their ionization degree, no correction for ionization of Δlog kW(IAM) values was needed. Furthermore, log P0(in situ) were found roughly linearly related to log BB values (i.e., the logarithm of the ratio brain concentration/blood concentration measured in vivo) for all the analytes but those predominantly present at the experimental pH 7.4 as anions. These results suggest that, at least for the data set considered, Δlog kW(IAM) parameters are more effective than log P0(PAMPA-BBB) at predicting log P0(in situ) values for all the analytes. Furthermore, ionization appears to affect differently, and much more markedly, BBB passage of acids (yielding anions) than that of the other ionizable compounds.