MK-0354

Endogenous ligands for alicarboxylic acid receptors are important metabolic intermediates that play a significant role in regulating body energy and maintaining homeostasis. However, the molecular mechanism of alicarboxylate ligand-mediated counterpart receptors is currently unclear. We resolve the active state structure of HCA2-niacin, and the structural analysis explains the mechanism of niacin selectivity in the alicarboxylic acid receptors family. Homology modeling, molecular dynamics simulation and mutagenesis experiments reveal different ligand recognition modes and activation mechanisms of the alicarboxylic acid receptors, analyze the flexibility of the binding pocket and elucidate the important role of disulfide bonds on receptor activation and ligand binding. These more detailed molecular mechanisms further elucidate the relevant mechanisms of human metabolism and provide key clues for subsequent drug development of alicarboxylic acid receptors.

Mathematical models predicting in vivo pharmacodynamic effects from in vitro data can accelerate drug discovery, and reduce costs and animal use. However, data integration and modeling is non-trivial when more than one drug-target receptor is involved in the biological response. We modeled the inhibition of non-esterified fatty acid release by dual G-protein-coupled receptor 81/109A (GPR81/GPR109A) agonists in vivo in the rat, to estimate the in vivo EC₅₀ values for 12 different compounds. We subsequently predicted those potency estimates using EC₅₀ values obtained from concentration-response data in isolated primary adipocytes and cell systems overexpressing GPR81 or GPR109A in vitro. A simple linear regression model based on data from primary adipocytes predicted the in vivo EC₅₀ better than simple linear regression models based on in vitro data from either of the cell systems. Three models combining the data from the overexpressing cell systems were also evaluated: two piecewise linear models defining logical OR- and AND-circuits, and a multivariate linear regression model. All three models performed better than the simple linear regression model based on data from primary adipocytes. The OR-model was favored since it is likely that activation of either GPR81 or GPR109A is sufficient to deactivate the cAMP pathway, and thereby inhibit non-esterified fatty acid release. The OR-model was also able to predict the in vivo selectivity between the two receptors. Finally, the OR-model was used to predict the in vivo potency of 1651 new compounds. This work suggests that data from the overexpressing cell systems are sufficient to predict in vivo potency of GPR81/GPR109A agonists, an approach contributing to faster and leaner drug discovery.