A virtual screening campaign using machine learning identified molecules with potential for development as novel CDK9 inhibitors for the treatment of cancer, according to early research findings published in Biomolecules. Integration of artificial intelligence (AI) into the drug discovery phase allowed for an accelerated analysis and identification of 14 molecule candidates that were then further tested and narrowed down to the one most promising candidate.
“This scientific work has confirmed for us how the use of [AI] in research can be fundamental for the identification of new molecules to be used in the fight against cancer,” said study author Tiziano Tuccinardi, MD, Full Professor of Medicinal Chemistry in the Department of Pharmacy at the University of Pisa, Italy. “AI, in fact, makes it possible to drastically reduce the time and costs of the initial drug discovery phase, while at the same time increasing the likelihood of identifying truly effective compounds.”
Methods and Findings
CDK9 is a target of interest for anticancer therapy due to its overexpression in several malignancies. In cancer cells, it acts as a regulator of transcriptional elongation and DNA repair, allowing the cells to survive through the continued expression of oncogenes and antiapoptotic proteins.
Researchers systematically collected data through five algorithms, two classification settings, and seven molecular representations to prepare 70 predictive models for possible molecule candidates. A virtual screening campaign narrowed down the candidates to 14 compounds that could bind to the CDK9 protein and showed any inhibitory effect. The compounds were tested in cellular models of cervical carcinoma and breast cancer until two molecules were selected based on their low micromolar range inhibitory activity; these were then further tested in three cancer cell lines, each with a different genetic background.
Ultimately, one compound was chosen for its therapeutic potential, which showed cytotoxic effects both as a single agent and when used in combination with camptothecin across several tested cell lines.
“[AI] is proving to be one of the fundamental tools for basic and applied research, thanks to its enormous analytical and predictive capabilities,” concluded study author Antonio Giordano, MD, PhD, Founder and Director of the Sbarro Health Research Organization. “However, experimental validation by humans remains a crucial and irreplaceable phase of the entire scientific process.”
DISCLOSURE: For full disclosures of the study authors, visit mdpi.com.
