Designing Macrocyclic Kinase Inhibitors Using Macrocycle Scaffold Hopping with Reinforced Learning (Macro-Hop)

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Drug Discovery

Background:

• Macrocyclic compounds are increasingly important in drug discovery, with over 70 currently in clinical use.
• Developing effective methods for designing novel macrocycles remains a significant challenge in the field.

Purpose of the Study:

• To introduce Macro-Hop, a reinforced learning framework for efficient exploration of macrocycle chemical space.
• To generate novel macrocyclic scaffolds with desired physicochemical properties and 3D structural similarity to reference compounds.

Main Methods:

• Utilized a reinforced learning framework (Macro-Hop) to explore macrocycle chemical space.
• Applied Macro-Hop to design macrocycle inhibitors targeting PDGFRα D842V kinase.

Main Results:

• Macro-Hop successfully generated novel macrocyclic scaffolds.
• The designed compound L7 demonstrated high potency against PDGFRα D842V (IC50 = 23.8 nM biochemical, 2.1 nM cellular).
• L7 also inhibited key secondary mutants PDGFRα D842V/G680R (IC50 = 64.1 nM) and PDGFRα D842V/T674I (IC50 = 27.6 nM).

Conclusions:

• Macro-Hop enables rapid and comprehensive exploration of macrocycle chemical space.
• The framework facilitates the design of potent and selective macrocyclic inhibitors.
• The study highlights the rapid validation of AI-designed molecules in wet-lab experiments.