Development of Pharmacophore Models for the Important Off-Target 5-HT_2B Receptor.

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Pharmacology

Background:

• Drug toxicity, particularly off-target effects, is a major hurdle in pharmaceutical development.
• 5-HT_2B receptor agonism is linked to safety issues like valvular heart disease, leading to market withdrawals.

Purpose of the Study:

• To develop early detection methods for potential 5-HT_2B receptor binding.
• To identify key amino acid residues in the 5-HT_2B active site.
• To create predictive models for filtering compounds with potential 5-HT_2B liabilities.

Main Methods:

• Molecular dynamics simulations to identify key amino acid residues.
• Development and performance evaluation of pharmacophore models on in-house data.
• Machine learning model applied to a diverse chemical library for activity prediction.

Main Results:

• Identification of critical amino acid residues in the 5-HT_2B active site.
• Validated pharmacophore models for predicting 5-HT_2B activity.
• A machine learning model successfully labeled a diverse compound subset for 5-HT_2B activity.

Conclusions:

• The developed models serve as effective filters for early identification of compounds with potential 5-HT_2B off-target liabilities.
• This approach can aid in preventing drug attrition due to 5-HT_2B-mediated toxicity.
• Facilitates the development of safer pharmaceuticals by minimizing risks associated with valvular heart disease.