There is increasing evidence that compounds with selectivity for gamma-aminobutyric acid(A) (GABA(A)) alpha2- and/or alpha3-subtypes may retain the desirable anxiolytic activity of nonselective benzodiazepines but possess an improved side effect profile. Herein we describe a novel series of GABA(A) alpha2/alpha3 subtype-selective agonists leading to the identification of the development candidate 17, a nonsedating anxiolytic in preclinical animal assays.

The in vivo occupancy of brain benzodiazepine binding sites by compounds A and B was measured using a [(3)H]Ro 15-1788 binding assay and related to plasma and brain drug concentrations. The plasma concentration associated with 50% occupancy was higher for compound A than compound B (73 and 3.7 nM, respectively), however, there was little difference in the brain concentrations required (73 and 63 nM). Both compounds showed a non-linear relationship between plasma and brain concentrations such that above brain concentrations of approximately 100 nM increasing plasma concentrations did not result in a concomitant increase in brain concentrations. This is consistent with brain concentrations being dependent on a saturable compartment which was postulated to be the benzodiazepine binding site-containing GABA(A) receptors. This hypothesis was tested in alpha1H101R mice, in which the alpha1 subunit of the GABA(A) receptor is rendered insensitive to benzodiazepine binding resulting in an approximate 50% reduction in the total benzodiazepine-containing GABA(A) receptor population. It was shown that the Occ(50) brain concentrations in the alpha1H101R animals was lower (17 nM) than in wild type mice (63 nM), as was the plateau concentration in the brain (105 and 195 nM, respectively). These data suggest measured concentrations of compounds A and B in brain tissue are dependent on receptor expression with a minimal contribution from unbound and non-specifically bound compound.