GABAA Receptor Subtype-Selective Modulators. I. α2/α3-Selective Agonists as Non-Sedating Anxiolytics

The prototypic benzodiazepines, such as diazepam, are not only anxiolytic but also produce sedation. These effects are mediated by GABA(A) receptors containing either an α1, α2, α3 or α5 subunit at which the positive modulatory effects (i.e., agonist efficacy) of benzodiazepines are mediated via a specific benzodiazepine recognition site. Recent molecular genetic and pharmacological data point to α1-containing GABA(A) receptors as the "sedative" and α2- and/or α3-containing receptors as the "anxiolytic" subtype(s). Therefore, at Merck Sharp & Dohme attempts were made to identify subtype-selective compounds that modulate α2/α3 but not α1 receptor function with the prediction that such compounds would be non-sedating anxiolytics. The initial strategy for discovering such "anxioselective" compounds focussed on producing compounds with much higher affinity at the α2/α3 compared to α1 subtypes. The starting point for this approach was the triazolophthalazine series developed from a combination of a screening hit and a literature compound [1]. However, the maximum α3 versus α1 binding selectivity that could be achieved in this series was 12-fold and this was not considered sufficient for an appropriate in vivo pharmacological differentiation compared to non-selective compounds. Nevertheless, within this series compounds demonstrating (albeit to a limited extent) higher agonist efficacy at the α3 versus α1 subtype were also identified. This suggested that it might be possible to synthesize a compound with higher efficacy at the α2 and/or α3 compared to α1 subtypes, ideally with no efficacy at the latter subtype (i.e., a compound with subtype-selective efficacy). By changing the structure from a triazolophthalazine to a triazolopyridazine core, a number of either pharmacological tool compounds (L-838417, MRK-067 and MRK-696) or clinical development candidates (MRK-409 and TPA023) were identified. Encouraged by the success of this approach and the observation that the benzimidazole NS-2710 had a modest degree of α3 versus α1 selectivity efficacy, a structurally-related class of imidazopyridines was also explored. The introduction of an additional nitrogen into the imidazopyridine core gave the imidazopyrimidine series which initially had issues with poor dog pharmacokinetics. However, this was resolved and resulted in the identification of the development candidates MRK-623 and MRK-898. A fluoroimidazopyridine was found to be a bioisostere of the imidazopyrimidine core and in this series the α3-selective tool compound TP003 was identified. The addition of a further nitrogen into the imidazopyrimidine core produced the imidazotriazine series, which yielded the clinical candidate TPA023B. Imidazopyrazinone and imidazotriazinone compounds offered no advantages over their respective imidazopyrimidine and imidazotriazine analogues. Additional pharmacological tool compounds were identified within the pyridine, pyrazolotriazine, pyridazine and pyrazolopyridone series highlighting the general feasibility of GABA(A) receptor subtype selective efficacy as a strategy for developing compounds with novel in vitro and in vivo profiles.