Machine Learning-Based Virtual Screening for the Identification of Cdk5 Inhibitors

Area of Science:

• Medicinal Chemistry
• Computational Biology
• Drug Discovery

Background:

• Cyclin-dependent kinase 5 (Cdk5) is crucial in the central nervous system but its dysregulation is linked to neurodegenerative diseases (e.g., Alzheimer's, Parkinson's) and cancer.
• Cdk5 is a promising drug target due to its role in disease progression.
• Developing novel small-molecule Cdk5 inhibitors is a significant area of medicinal chemistry research.

Purpose of the Study:

• To identify novel small-molecule inhibitors of Cdk5 using a computational approach.
• To validate an in silico workflow for drug discovery targeting Cdk5.
• To provide lead compounds for further structure-based optimization.

Main Methods:

• Machine learning-based virtual screening.
• Molecular docking and simulations.
• Binding free energy calculations.

Main Results:

• Identification of two novel Cdk5 inhibitors (CPD1 and CPD4) with a 50% experimental hit rate.
• Both compounds demonstrated promising enzyme inhibitory activity.
• CPD1 exhibited significant antiproliferative effects in ovarian and colon cancer cells.

Conclusions:

• The in silico workflow is reliable for identifying Cdk5 inhibitors.
• CPD1 and CPD4 are valuable starting points for developing new Cdk5-targeted drugs.
• Further structure-based optimization can lead to potent Cdk5 inhibitors for therapeutic applications.