Pharmacophore Development And Virtual Screening Of Corticotropin-Releasing Factor 1 Receptor Antagonists

Corticotropin-releasing factor (CRF) is the prime regulator of the hypothalamus-pituitary-adrenal (HPA) axis, coordinating endocrine, autonomic, behavioral, and immune responses to stress. Numerous reports suggest that alterations in CRF function contribute to the pathogenesis of depression. Moreover, selective nonpeptide CRF type 1 (CRF1) receptor antagonists have demonstrated antidepressant efficacy in extensive preclinical studies as well as limited clinical studies. The CRF1 receptor has been considered a renewed target for potential treatment.In our study, both quantitative and qualitative chemical function based pharmacophore models of CRF1 receptor antagonists were generated using two algorithms (HypoGen and HipHopRefine) of Catalyst software. The training set of HypoGen included twenty-five CRF1 antagonists exhibiting Ki values spanning from 0.5 nM to over 10000nM. The best HypoGen model consisted of five features: four hydrophobic aliphatic and one ring aromatic features. The best HipHopRefine model was derived from eight potent and four poor CRF1 antagonists. It composed of six excluded volumes, four hydrophobic aliphatic and one ring aromatic features. Both models were validated on a wide set of test molecules. They were able to identify potent antagonists and estimate their activities.Based on our pharmacophore models, computational screening of commercial Maybridge-3D and in-house databases identified several hits. Further refinement of one of these molecules, compound DYF-022-9 lead to a series of potential CRF1 antagonists, with the most potent compound having an estimated activity of the same or even better level of known antagonistsThe results of our study provide insight into the chemical features essential for biological activity. It should be useful in lead modification and virtual screening, contributing to the identification of CRF1 antagonists with novel structures.