Discovery of indoleamine 2,3-dioxygenase inhibitors using machine learning based virtual screening

Area of Science:

• Medicinal Chemistry
• Computational Chemistry
• Cancer Immunotherapy

Background:

• Indoleamine 2,3-dioxygenase (IDO) is a key immune checkpoint and a target for cancer immunotherapy.
• Existing IDO inhibitors lack clinical approval, necessitating the search for novel therapeutic agents.

Purpose of the Study:

• To identify novel indoleamine 2,3-dioxygenase (IDO) inhibitors using machine learning.
• To explore the potential of tanshinone derivatives as IDO inhibitors for cancer treatment.

Main Methods:

• Development of naive Bayesian (NB) and recursive partitioning (RP) machine learning models.
• Virtual screening of an in-house compound library using predictive models.
• Enzyme-based and cell-based assays to evaluate IDO inhibitory activity.
• Surface plasmon resonance (SPR) and molecular dynamic (MD) simulations for target interaction and binding mode analysis.

Main Results:

• Three novel IDO inhibitors belonging to the tanshinone family were identified with IC50 values ranging from 1.30 to 4.68 μM.
• The identified compounds demonstrated IDO inhibitory activity in both enzyme and cell-based assays.
• SPR and MD simulations confirmed the direct interaction of the inhibitors with the IDO target and elucidated their binding modes.

Conclusions:

• Machine learning models successfully identified novel, potent IDO inhibitors from the tanshinone scaffold.
• This study reveals a new therapeutic application for Danshen (Salvia miltiorrhiza) derivatives in cancer immunotherapy.
• The discovered compounds represent promising candidates for further development as IDO-targeted cancer therapeutics.