In silico studies on recreational drugs: 3D quantitative structure activity relationship prediction of classified and de novo designer benzodiazepines

Abstract

Currently, increasing availability and popularity of designer benzodiazepines (DBZDs) constitutes a primary threat to public health. To assess this threat, the biological activity/potency of DBZDs was investigated using in silico studies. Specific Quantitative Structure Activity Relationship (QSAR) models were developed in Forge™ for the prediction of biological activity (IC50) on the γ‐aminobutyric acid A receptor (GABA‐AR) of previously identified classified and unclassified DBDZs. A set of new potential ligands resulting from scaffold hopping studies conducted with MOE® was also evaluated. Two generated QSAR models (i.e. 3D‐field QSAR and RVM) returned very good performance statistics (r2 = 0.98 [both] and q2 = 0.75 and 0.72, respectively). The DBZDs predicted to be the most active were flubrotizolam, clonazolam, pynazolam and flucotizolam, consistently with what reported in literature and/or drug discussion fora. The scaffold hopping studies strongly suggest that replacement of the pendant phenyl moiety with a five‐membered ring could increase biological activity and highlight the existence of a still unexplored chemical space for DBZDs. QSAR could be of use as a preliminary risk assessment model for (newly) identified DBZDs, as well as scaffold hopping for the creation of computational libraries that could be used by regulatory bodies as support tools for scheduling procedures.

Applications and Potential of In Silico Approaches for Psychedelic Chemistry

“…In recent studies, the application of field-based 3D-QSAR modeling has emerged as a valuable approach for understanding biological activity and designing new molecules in the field of psychedelic chemistry. Catalani et al [ 92 ] employed 3D QSAR models, specifically the 3D-field QSAR and RVM (relevance vector machine) models, to predict the biological activity of designer benzodiazepines (DBZDs) on the γ-aminobutyric acid A receptor (GABA-AR). The models exhibited excellent performance statistics, indicating their reliability in predicting the potency of DBZDs.…”

“…Furthermore, the authors conducted scaffold hopping studies, revealing the potential for improving the biological activity of DBZDs by replacing the pendant phenyl moiety with a five-membered ring. This finding suggests the existence of an unexplored chemical space for DBZDs, highlighting the significance of computational techniques in expanding the design possibilities for novel psychedelic compounds [ 92 ]. Additionally, Floresta and Abbate [ 93 ] presented a comparative analysis between machine learning approaches and field-based 3D-QSAR models for predicting the affinity of substances acting on the serotonin 2A receptor (5-HT2AR).…”

Designer Benzodiazepines: Effects, Toxicity, and Interactions

“…Interestingly, flubrotizolam, the molecule predicted to be the most potent, is a new DBZDs for which no data are available in the literature. 76…”