Artificial intelligence-powered discovery of small molecules inhibiting CTLA-4 in cancer

Area of Science:

• Immunology
• Computational Chemistry
• Drug Discovery

Background:

• Checkpoint inhibitors have transformed cancer immunotherapy but have limitations.
• Monoclonal antibodies targeting CTLA-4 can cause severe immune-related adverse events.
• Existing small molecules target CD80, not CTLA-4, limiting direct therapeutic intervention.

Purpose of the Study:

• To identify novel small molecule inhibitors of the CTLA-4/CD80 interaction using AI-powered virtual screening.
• To validate and optimize these inhibitors for potential cancer therapy.

Main Methods:

• AI-driven virtual screening of approximately ten million compounds to identify CTLA-4 targeting agents.
• In vitro and in vivo validation using biochemical, biophysical, immunological, and animal assays.
• Optimization of lead compounds to achieve potent inhibitors of the CTLA-4/CD80 interaction.

Main Results:

• AI screening successfully identified validated hit molecules targeting CTLA-4.
• Optimized compounds inhibited the CTLA-4/CD80 interaction at micromolar concentrations without degrading CTLA-4.
• Several compounds demonstrated prophylactic and therapeutic efficacy in preclinical cancer models.

Conclusions:

• AI-based frameworks are effective for designing small molecules targeting immune checkpoints.
• The identified compounds show potential for developing novel cancer immunotherapies.
• This approach offers a promising strategy for overcoming limitations of current checkpoint inhibitor therapies.