GABAA Receptor Subtype-Selective Modulators. II. α5-Selective Inverse Agonists for Cognition Enhancement

Benzodiazepine site agonists (such as diazepam) are well-known to impair cognition. Since benzodiazepines exert their effects via modulation of alpha 1-, alpha 2-, alpha 3- and alpha 5-containing GABA(A) receptors, the cognition-impairing effects of diazepam must be associated with one or several of these subtypes. Of these different subtypes, alpha 5-containing GABA(A) receptors represent an attractive option as the "cognition" subtype based upon the preferential localization of these receptors within the hippocampus and the well-established role of the hippocampus in learning and memory. As a result, it is hypothesized that an inverse agonist selective for the alpha 5 subtype should enhance cognition. For example, L-655708, a partial inverse agonist with 50-100-fold higher affinity for the alpha 5 relative to the alpha 1, alpha 2 and alpha 3 subtypes of GABA(A) receptors, enhanced cognitive performance in rats. Unfortunately, however, the pharmacokinetic properties of this compound prevented it being developed further. In order to try achieve binding selectivity in a series structurally distinct from the imidazobenzodiazepines, the group at Merck Sharp & Dohme commenced studies within the triazolopyridazine series. Although a degree of binding selectivity could be achieved (a maximum of 22-125-fold for alpha 5 versus alpha 1, alpha 2 or alpha 3, this approach was dropped in favour of a strategy to identify compounds with either a combination of selective affinity and selective efficacy or purely selective efficacy. With respect to the former, screening of the Merck chemical collection identified a novel, moderately alpha 5 binding selective thiophene series and further optimization of this series produced MRK-536, which demonstrated a modest alpha 5 binding selectivity (similar to 10-fold) as well as alpha 5 efficacy selectivity. However, the structure-activity relationship within this and the analogous tetralone series proved unpredictable and these series were not pursued further. The success of the selective efficacy approach on the alpha 2/alpha 3-selective agonist project led a similar paradigm being adopted for the alpha 5 project. The starting point for this strategy was the triazolopyridazine 3 which, like MRK-536, possessed a degree of both alpha 5 binding and efficacy selectivity. By changing the core from a triazolopyridazine to a triazolophthalazine structure, alpha 5 binding selectivity was lost but with subsequent optimization, compounds with the desired profile (low or antagonist efficacy at the alpha 1, alpha 2 and alpha 3 subtypes and marked inverse agonism at alpha 5-containing receptors) could be achieved, allowing the clinical candidate alpha 5IA as well as the structurally-related pharmacological tool compound alpha 5IA-II to be identified. By appending features of the prototypic alpha 2/alpha 3-selective triazolopyridazine L-838417 (t-butyl and 1,2,4 triazole groups) along with the isoxazole of alpha 5IA to a pyrazolotriazine core, an additional clinical candidate, MRK-016, was described. Finally, a degree of alpha 5 efficacy selectivity was achieved the pyridazine series but metabolic instability within this chemotype limited its further optimization. Overall, these studies demonstrate the feasibility of adopting a selective efficacy approach in the identification of alpha 5 selective GABA(A) receptor inverse agonists.