Ligand- and Structure-Based Analysis of Deep Learning-Generated Potential α2a Adrenoceptor Agonists

The signaling and selectivity of α-adrenoceptor agonists for the human α2A, α2B and α2C-adrenoceptors and comparison with human α1 and β-adrenoceptors

α2‐adrenoceptors, (α2A, α2B and α2C‐subtypes), are Gi‐coupled receptors. Central activation of brain α2A and α2C‐adrenoceptors is the main site for α2‐agonist mediated clinical responses in hypertension, ADHD, muscle spasm and ITU management of sedation, reduction in opiate requirements, nausea and delirium. However, despite having the same Gi‐potency in functional assays, some α2‐agonists also stimulate Gs‐responses whilst others do not. This was investigated. Agonist responses to 49 different α‐agonists were studied (CRE‐gene transcription, cAMP, ERK1/2‐phosphorylation and binding affinity) in CHO cells stably expressing the human α2A, α2B or α2C‐adrenoceptor, enabling ligand intrinsic efficacy to be determined (binding K_D/Gi‐IC₅₀). Ligands with high intrinsic efficacy (e.g., brimonidine and moxonidine at α2A) stimulated biphasic (Gi‐Gs) concentration responses, however for ligands with low intrinsic efficacy (e.g., naphazoline), responses were monophasic (Gi‐only). ERK1/2‐phosphorylation responses appeared to be Gi‐mediated. For Gs‐mediated responses to be observed, both a system with high receptor reserve and high agonist intrinsic efficacy were required. From the Gi‐mediated efficacy ratio, the degree of Gs‐coupling could be predicted. The clinical relevance and precise receptor conformational changes that occur, given the structural diversity of compounds with high intrinsic efficacy, remains to be determined. Comparison with α1 and β1/β2‐adrenoceptors demonstrated subclass affinity selectivity for some compounds (e.g., α2:dexmedetomidine, α1:A61603) whilst e.g., oxymetazoline had high affinity for both α2A and α1A‐subtypes, compared to all others. Some compounds had subclass selectivity due to selective intrinsic efficacy (e.g., α2:brimonidine, α1:methoxamine/etilefrine). A detailed knowledge of these agonist characteristics is vital for improving computer‐based deep‐learning and drug design.