Discovery of novel selective PI3Kγ inhibitors through combining machine learning-based virtual screening with...

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Oncology

Background:

• Phosphoinositide 3-kinase gamma (PI3Kγ) is a drug target for various diseases.
• Developing selective PI3Kγ inhibitors is challenging due to structural conservation among PI3K isoforms.

Purpose of the Study:

• To develop a novel machine learning-based virtual screening method for discovering selective PI3Kγ inhibitors.
• To identify and characterize novel PI3Kγ inhibitors with potential therapeutic applications.

Main Methods:

• A machine learning virtual screening model was developed using multiple PI3Kγ protein structures.
• A large chemical database was screened, and top compounds underwent bio-evaluation.
• Theoretical studies were conducted to understand the selective inhibition mechanism of identified compounds.

Main Results:

• The virtual screening identified 49 hits, leading to the discovery of JN-KI3.
• JN-KI3 selectively inhibited PI3Kγ (IC50 = 3,873 nM) without affecting Class IA PI3Ks.
• JN-KI3 induced selective cytotoxicity and apoptosis in hematologic cancer cells by inhibiting PI3K signaling.

Conclusions:

• The virtual screening model is effective for discovering novel PI3Kγ inhibitors.
• JN-KI3 demonstrates potential as a PI3Kγ-targeted therapy for hematologic tumors due to its selective cytotoxicity and apoptosis induction.
• JN-KI3 possesses novel structural characteristics for selective PI3Kγ inhibition, interacting with less common key residues.