Researchers at Integrated Biosciences and MIT apply deep neural networks to screen over 800,000 compounds, identifying three potent senolytics with high oral bioavailability that selectively induce apoptosis in senescent ‘zombie’ cells. These candidates bind Bcl-2, clear senescent cells in aged mice, and offer promising anti-aging therapeutic potential.
Key points
- Deep neural networks trained on experimental datasets screened over 800,000 compounds to predict senolytic activity.
- Three lead molecules exhibited high selectivity for senescent cells, binding the anti-apoptotic protein Bcl-2.
- In 80-week-old mouse models, one candidate cleared senescent renal cells and reduced senescence-associated gene expression.
Why it matters: These AI-discovered senolytics could revolutionize anti-aging therapies by selectively clearing harmful senescent cells with improved drug-like properties.
Q&A
- What are senolytics?
- How do deep neural networks predict senolytic activity?
- What role does Bcl-2 play in senescent cell apoptosis?
- Why is oral bioavailability important for drug development?
