P-glycoprotein limits the brain penetration of nonsedating but not sedating H1-antagonists.

The present study evaluates the impact of P-glycoprotein (P-gp) on plasma-brain disposition and transepithelial transport of sedating versus nonsedating H1-antagonists using multidrug-resistant (mdr) gene 1a and 1b (mdr1a/b) knockout (KO) mice and human MDR1-transfected Madin-Darby canine kidney (MDCK) cells. Three nonsedating (cetirizine, loratadine, and desloratadine) and three sedating (diphenhydramine, hydroxyzine, and triprolidine) H1-antagonists were tested. Each compound was administered to KO and wild-type (WT) mice intravenously at 5 mg/kg. Plasma and brain drug concentrations were determined by liquid chromatography-mass spectrometry analysis. Mean pharmacokinetic parameters (CL, V(ss), and t(1/2)) were obtained using WinNonlin. In addition, certirizine, desloratadine, diphenhydramine, and triprolidine (2 microM) were tested as substrates for MDR1 using MDR1-MDCK cells. The bidirectional apparent permeability was determined by measuring the amount of compound at the receiving side at 5 h. The brain-to-plasma area under the curve (AUC) ratio was 4-, 2-, and >14-fold higher in KO compared with WT mice for cetirizine, loratadine, and desloratadine, respectively. In contrast, the brain-to-plasma AUC ratio between KO and WT was comparable for hydroxyzine, diphenhydramine, and triprolidine. Likewise, the efflux ratio between basolateral to apical and apical to basolateral was 4.6- and 6.6-fold higher in MDR1-MDCK than the parental MDCK for certirizine and desloratadine, respectively, whereas it was approximately 1 for diphenhydramine and triprolidine. Our results demonstrate that sedating H1-antagonists hydroxyzine, diphenhydramine, and triprolidine are not P-gp substrates. In contrast, nonsedating H1-antagonists cetirizine, loratadine, and desloratadine are P-gp substrates. Affinity for P-gp at BBB may explain the lack of central nervous system side effects of modern H1-antagonists.