Phenotypic Polypharmacology Drug Discovery for CNS Applications

The complexity of the central nervous system (CNS), characterized by an intricate interplay among multiple neurotransmitter systems, poses a substantial challenge for drug discovery, especially for neuropsychiatric disorders. Not surprisingly, most psychiatric drugs rely on polypharmacology, the simultaneous modulation of multiple targets by a single drug, often the result of phenotypic, and in many cases serendipitous, observations. In this light, approaching CNS drug discovery through a target-agnostic in vivo phenotypic profiling of compounds may provide the best strategy for the successful identification of novel CNS drugs. To prosecute effectively, such a strategy must rely on infrastructure that is high throughput, standardized, and unbiased, i.e., a system that holistically assesses the CNS effects of compounds without preconceptions on pharmacological targets. We believe that target agnostic strategies should drive the identification of new leads and the development of structure-activity relationships (SAR) in lead optimization campaigns. After briefly analyzing the history of psychopharmacology to learn what worked (and what did not) in the drug discovery process, we will present examples of target agnostic approaches to drug discovery based on SmartCube^®, a proprietary platform that integrates behavior phenotype profiling in vivo with innovative machine learning (ML) and artificial intelligence (AI). We will present examples of the high-content behavioral phenotypic profiles obtained by exploring privileged scaffolds, such as tryptamine and 1,2 benzisoxazole, and highlight how structural modifications result in prominent changes in potency, therapeutic window, and overall CNS profile of the derivative compounds. As a different example, we highlight the power of target agnostic approaches in the discovery of ulotaront, a putative antipsychotic currently in Phase III clinical trials with a first-in-class mechanism of action (MOA) that does not target dopamine receptors and has a placebo-like tolerability profile. In conclusion, this chapter illustrates our view that in vivo phenotypic profiling, what we called behavioral phenotypic drug discovery (BPDD) is better suited for CNS drug discovery than traditional approaches due to its target agnostic nature and its potential to identify treatments with novel MOAs.