Multi-objective evolutionary design of adenosine receptor ligands.

Area of Science:

• Medicinal Chemistry
• Computational Drug Design
• Pharmacology

Background:

• Adenosine receptors are crucial drug targets.
• Developing selective antagonists with good ADMET properties remains challenging.
• De novo drug design requires efficient computational methods.

Purpose of the Study:

• To develop and apply a novel multiobjective evolutionary algorithm (MOEA) for de novo design of adenosine receptor antagonists.
• To optimize compounds for high affinity, selectivity, and favorable ADMET properties.
• To identify novel chemical scaffolds with potential therapeutic applications.

Main Methods:

• Utilized a multiobjective evolutionary algorithm (Molecule Commander) for iterative structure generation, evaluation, and selection.
• Employed a pharmacophore model for the human A1 adenosine receptor (hA1AR) as an objective function.
• Developed support vector machine models for other adenosine receptor subtypes (hA2A, hA2B, hA3) to ensure selectivity.
• Evaluated compounds based on predicted affinity, selectivity, and ADMET properties.

Main Results:

• Generated a library of 3946 unique compounds, from which chemical scaffolds were derived.
• Six selected scaffolds were synthesized and tested, with scaffolds 2 and 3 showing low micromolar affinity for adenosine receptor subtypes.
• Systematic modifications on scaffold 3 led to improved affinity and selectivity for hA1AR; compound 3a showed 280 nM affinity with 10-fold selectivity, and compound 3g exhibited 1.6 μM affinity with negligible activity at other subtypes.

Conclusions:

• The developed MOEA is effective for de novo design of adenosine receptor antagonists.
• The identified scaffolds represent promising starting points for developing novel therapeutics targeting adenosine receptors.
• The study demonstrates the power of evolutionary algorithms in accelerating drug discovery with desired pharmacological profiles.