Discovering Dually Active Anti-cancer Compounds with a Hybrid AI-structure-based Approach

Area of Science:

• Oncology
• Computational Chemistry
• Drug Discovery

Background:

• Cancer cells evade death by maintaining telomere length and tolerating DNA damage.
• Targeting both telomere maintenance and DNA repair pathways offers a potent anticancer strategy.
• G-quadruplex (G4) structures and poly(ADP-ribose) polymerase (PARP)1 are key targets for cancer therapy.

Purpose of the Study:

• To computationally identify novel compounds with dual inhibitory activity against G4 structures and PARP1.
• To develop a hybrid virtual screening approach combining AI and structure-based methods.
• To validate lead compounds for their anticancer efficacy.

Main Methods:

• Utilized a hybrid virtual screening (VS) protocol integrating machine learning (PyRMD) and structure-based VS.
• AI-driven analysis identified potential PARP1 inhibitors from large chemical libraries.
• Structure-based VS prioritized inhibitors for G4 stabilization potential.

Main Results:

• Identified 50 promising candidate compounds through the two-step VS process.
• Experimental validation confirmed dual PARP1 inhibition and G4 stabilization for lead compounds.
• Four lead compounds demonstrated significant antiproliferative effects against cancer cell lines.

Conclusions:

• Combined AI and structure-based methods effectively discover multitarget anticancer agents.
• Dual-targeting compounds offer a promising strategy for overcoming cancer resistance.
• This approach accelerates the development of novel cancer therapeutics.