Use of biopartitioning micellar chromatography and RP-HPLC for the determination of blood–brain barrier penetration of α-adrenergic/imidazoline receptor ligands, and QSPR analysis

For this study, 31 compounds, including 16 imidazoline/α-adrenergic receptor (IRs/α-ARs) ligands and 15 central nervous system (CNS) drugs, were characterized in terms of the retention factors (<i>k</i>) obtained using biopartitioning micellar and classical reversed phase chromatography (log <i>k</i>_BMC and log <i>k</i>_<i>w</i>RP, respectively). Based on the retention factor (log <i>k</i>_<i>w</i>RP) and slope of the linear curve (<i>S</i>) the isocratic parameter (<i>φ</i>_<i>0</i>) was calculated. Obtained retention factors were correlated with experimental log <i>BB</i> values for the group of examined compounds. High correlations were obtained between logarithm of biopartitioning micellar chromatography (BMC) retention factor and effective permeability (<i>r</i>(log <i>k</i>_BMC/log <i>BB</i>): 0.77), while for RP-HPLC system the correlations were lower (<i>r</i>(log <i>k</i>_<i>w</i>RP/log <i>BB</i>): 0.58; <i>r</i>(<i>S</i>/log <i>BB</i>): –0.50; <i>r</i>(<i>φ</i>_<i>0</i>/<i>P</i>_<i>e</i>): 0.61). Based on the log <i>k</i>_BMC retention data and calculated molecular parameters of the examined compounds, quantitative structure–permeability relationship (QSPR) models were developed using partial least squares, stepwise multiple linear regression, support vector machine and artificial neural network methodologies. A high degree of structural diversity of the analysed IRs/α-ARs ligands and CNS drugs provides wide applicability domain of the QSPR models for estimation of blood–brain barrier penetration of the related compounds.