QSAR Modeling of 5-HT1B, 5-HT1D and 5-HT1F Serotonin Receptor Ligands and Model-Based Parallel Virtual Screening for Novel Anti-Migraine Therapeutics

poster · 10 years ago
by Alexander Tropsha, Tiago L. Moda, Xiang S. Wang, Adriano D. Andricopulo (University of North Carolina at Chapel Hill, Universidade de São Paulo)
Migraine is a recurring, episodic neurological disorder characterized as unilateral, throbbing headache.1 The only class of marketed drugs to treat migraine are triptans sharing serotonin scaffold (Figure 1). However, current medications such as sumatriptan and naratriptan are not effective in all patients. Recent findings indicate that compounds that can simultaneously target a triplet of serotonin 5-HT receptors, i.e., 5-HT1B, 5-HT1D, and 5HT1F, can be very effective and safe therapeutics to treat acute migraine. 2 We have collected three 5HT subtype-specific data sets of molecules (532 in total) with known affinities for respective receptor subtypes. We have employed the data to develop robust and externally predictive Quantitative Structure-Activity Relationship (QSAR) models for every 5HT subtype, using both k-Nearest Neighbor (kNN) and Support Vector Machines (SVM) methods. We have employed these models for virtual screening of diverse chemical libraries to identify compounds predicted to bind to all three 5HT subtypes.