J Pharmacol Sci 124, 160 – 168 (2014) Journal of Pharmacological Sciences © The Japanese Pharmacological Society Full Paper Molecular Determinants Responsible for Sedative and Non-sedative Properties of Histamine H1–Receptor Antagonists Yoshihiro Uesawa1, Shigeru Hishinuma2,*, and Masaru Shoji2 1Department of Clinical Pharmaceutics, 2Department of Pharmacodynamics, Meiji Pharmaceutical University, Kiyose, Tokyo 204-8588, Japan Received September 11, 2013; Accepted November 26, 2013 Abstract. There is argument whether non-sedative properties of histamine H1–receptor antago- nists (antihistamines) are determined by their active extrusions from the brain via P-glycoprotein or their restricted penetration through the blood-brain barrier. We have reported that sedative and non-sedative antihistamines can be well discriminated by measuring changes in their binding to H1 receptors upon receptor internalization in intact cells, which depends on their membrane- penetrating ability. In this study, molecular determinants responsible for sedative and non-sedative properties of antihistamines were evaluated by quantitative structure-activity relationship (QSAR) analyses. Multiple regression analyses were applied to construct a QSAR model, taking inter- nalization-mediated changes in the binding of antihistamines as objective variables and their structural descriptors as explanatory variables. The multiple regression model was successfully constructed with two explanatory variables, i.e., lipophilicity of the compounds at physiological pH (logD) and mean information content on the distance degree equality (IDDE) (r2 = 0.753). The constructed model discriminated between sedative and non-sedative antihistamines with 94% accuracy for external validation. These results suggest that logD and IDDE concerning lipophilicity and molecular shapes of compounds, respectively, predominantly determine the membrane- penetrating ability of antihistamines for their side effects on the central nervous system. Keywords: histamine H1 receptor, antihistamine, sedation, membrane penetration, P-glycoprotein Introduction active extrusion from the brain via P-glycoprotein (1 – 3) or their restricted penetration through the blood–brain Histamine H1–receptor antagonists / inverse agonists barrier (4 – 7). (antihistamines) are well known to have side effects such Receptor internalization, movement of the receptor as sedation, hypnosis, and cognitive impairment, which from the cell surface to intracellular compartments, is are associated with the blockade of H1 receptors in the known to affect the binding properties of receptor ligands central nervous system (CNS). On the basis of these in intact cells, depending on their ability to penetrate the clinical side effects, antihistamines are generally divided biomembrane (8, 9). We therefore tested how receptor into two groups, i.e., sedative and non-sedative (or less internalization influenced the binding properties of a sedative) antihistamines. Non-sedative antihistamines variety of antihistamines under ice-cold conditions where have fewer side effects on the CNS as a result of less a P-glycoprotein–mediated extrusion pump might not blockade of H1 receptors in the CNS, although they work (10). Our finding is that there are clear differences might induce sedation at higher doses. It is yet still between the effect of H1-receptor internalization on the inconclusive, however, about whether non-sedative binding of sedative and non-sedative antihistamines to properties of antihistamines are determined by their intact cells, which provide strong evidence that simple diffusion through the plasma membrane predominantly *Corresponding author. [email protected] determines their sedative and non-sedative properties. Published online in J-STAGE on January 29, 2014 However, the variety of chemical structures and physico- doi: 10.1254/jphs.13169FP chemical properties of antihistamines makes it difficult 160 Sedative and Non-sedative Antihistamines 161 to accurately predict membrane-penetrating ability for knowledge, we succeeded for the first time in construct- their sedative and non-sedative properties. ing a QSAR model to explain and predict sedative and Quantitative analyses of the chemical structures of non-sedative properties of antihistamines, and the con- compounds can be useful to explain and predict their structed QSAR model may also contribute to optimizing effects on physiological functions (11): many studies on the development of novel antihistamines with respect quantitative structure–activity relationships (QSAR) to their side effects on the central nervous system. have been successful in the analyses and prediction of pharmacological effects (12), enzymatic activities (13), Materials and Methods affinities for receptor proteins (14), pharmacokinetic parameters (15), and drug metabolism (16). However, Training and external validation set of antihistamines there is no report concerning a QSAR model specifically assessed designated for the membrane-penetrating ability of Nineteen antihistamines, for which the internalization- antihistamines; therefore, we tried to establish a QSAR mediated changes in their binding to intact cells are model to explain and predict membrane-penetrating abil- already known (10), were assessed as a training set of ity of antihistamines for their sedative and non-sedative antihistamines (Fig. 1): sedative antihistamines were properties on the basis of our previous report (10). chlorpheniramine, clemastine, cyproheptadine, diphen- Here we show that sedative and non-sedative pro- hydramine, mepyramine, promethazine, azelastine, keto- perties of antihistamines can be predicted with extremely tifen oxatomide, ebastine, loratadine, and terfenadine (12 high accuracy by the QSAR model constructed on the compounds) and non-sedative antihistamines were basis of their membrane-penetrating ability alone. To our mequitazine, epinastine, bepotastine, carebastine, fexof- Fig. 1. Chemical structures of antihista- mines assessed as a training set. Chemical structures of 19 antihistamines are shown. The training set of antihistamines consisted of 12 sedative and 7 non-sedative (in italics) compounds. 162 Y Uesawa et al enadine, desloratadine, and olopatadine (7 compounds). The non-sedative behavior of ebastine, loratadine, and terfenadine is considered to be due to their corresponding active metabolites, carebastine, desloratadine, and fexof- enadine, respectively (10). Sixteen antihistamines, for which the internalization- mediated changes in their binding to intact cells are unknown, were assessed as an external validation set of antihistamines (Fig. 2): sedative antihistamines were alimemazine, azatadine, dimetindene, diphenylpyraline, homochlorcyclizine, hydroxyzine, imipramine, isothip- endyl, and triprolidine (9 compounds) and non-sedative antihistamines were acrivastine, astemizole, cetirizine, emedastine, levocabastine, mizolastine, and temelastine (7 compounds). Objective variables for assessment of sedative and non- Fig. 3. DAUC as an objective parameter to quantitatively assess in- sedative properties of antihistamines ternalization-mediated changes in displacement curves for antihista- 3 As an objective variable to assess the sedative and mines against [ H]mepyramine binding to H1 receptors. The figure non-sedative properties of antihistamines, the extent of shows displacement curves for a non-sedative H1-receptor antagonist, 3 changes in the binding of a training set of 19 anti- epinastine, against [ H]mepyramine binding to intact U373 MG astro- cytoma cells (taken from our previous paper, ref. 10, with permis- histamines by internalization of H1 receptors was ex- sion). DAUC was expressed as the difference in AUC between the pressed as the difference in area under the curve (AUC) displacement curves obtained with histamine-pretreated [i.e., inter- between the displacement curves obtained with hista- nalization-induced (closed circle) and histamine–non-pretreated con- mine-pretreated (i.e., internalization-induced) and hista- trol cells (open circle)]. mine–non-pretreated control cells (DAUC, Fig. 3). Fig. 2. Chemical structures of antihista- mines assessed as an external validation set. Chemical structures of 16 antihista- mines are shown. The external validation set of antihistamines consisted of 9 sedative and 7 non-sedative (in italics) compounds. Sedative and Non-sedative Antihistamines 163 Briefly, cells were pretreated with or without 0.1 mM from their optimized 3D structures by Dragon software histamine for 30 min at 37°C in HEPES buffer (120 mM ver. 5.5 (Talete srl, Milano, Italy). The Dragon descrip- NaCl, 5.4 mM KCl, 1.6 mM MgCl2, 1.8 mM CaCl2, tors of 1593 types were used in the present study. 11 mM D-glucose, and 25 mM HEPES, pH 7.4 at 37°C) Lipophilicity of the compounds at pH 7.5 (logD) was to induce the internalization of H1 receptors. Sub- calculated by a tautomer-considered KLOP method in sequently, the cells were washed with ice-cold HEPES arvinView. buffer and intact cell binding was performed at 4°C as described in our previous paper (9, 10). The displace- Construction and application of simple and multiple ment curves were fitted to either a one- or two-site regression models model as follows (KaleidaGraph; Synergy Software, The Dragon descriptors and logD values were used to Reading, PA, USA); construct simple and multiple regression models in the One-site model in the experiment without histamine training set. The best model was explored by genetic pretreatment: