Machine Learning-Based Virtual Screening for the Identification of Novel CDK-9 Inhibitors.

Area of Science:

• Oncology
• Computational Chemistry
• Drug Discovery

Background:

• Cyclin-dependent kinase 9 (CDK9) is vital for cancer cell survival, regulating oncogene and anti-apoptotic protein expression.
• CDK9 overexpression in malignancies positions it as a key therapeutic target for anticancer strategies.

Purpose of the Study:

• To develop and apply a machine learning (ML) approach for identifying novel inhibitors of CDK9.
• To validate potential CDK9 inhibitors through enzymatic assays and cancer cell line testing.

Main Methods:

• Systematic data collection and preprocessing to build seventy predictive ML models.
• Virtual screening (VS) guided by the best-performing ML model to identify candidate inhibitors.
• Enzymatic assays and subsequent testing of lead compounds in diverse cancer cell lines.

Main Results:

• Identification of 14 promising compounds from virtual screening.
• Two compounds demonstrated low micromolar inhibitory activity.
• A novel compound exhibited therapeutic potential, both alone and in combination with Camptothecin (CPT), with varying efficacy across cell lines.

Conclusions:

• Machine learning effectively integrated into drug discovery pipelines for identifying novel anticancer agents.
• The identified novel compound represents a promising starting point for developing new CDK9-targeted therapies.
• The study highlights the potential of CDK9 inhibition in cancer treatment, with combination therapy showing promise.