Researcher from the Computational Structural Biology group at the Institute for Research in Biomedicine (IRB, Università della Svizzera italiana), have shown that artificial intelligence (AI) can design a potent novel miniprotein for lipocalin-2 (LCN2), a protein linked to neuroinflammation. Their study, published in Frontiers in Immunology and titled “AI-assisted design of ligands for lipocalin-2” by Jacopo Sgrignani, Sara Buscarini, Patrizia Locatelli, Concetta Guerra, Alberto Furlan, Yingyi Chen, Giada Zoppi and Andrea Cavalli, paves the way for new diagnostics and therapies.
The IRB team engineered a small protein, MinP‑2, designed to target LCN2, a molecule produced in excess during infections and inflammation that can, for example, damage the blood-brain barrier (BBB). Elevated LCN2 levels are associated with brain inflammation and related disorders, as the protein interacts with other molecules in ways that weaken the brain’s defenses and disrupt immune responses. Using AI, the researchers generated thousand ligand candidate and tested the most promising in the lab. MinP‑2 proved the most effective, binding tightly to LCN2 and blocking its harmful interactions.
These findings make MinP2 a promising candidate for new strategies to detect or treat brain inflammation. Its small size, compared to traditional antibodies, may allow it to penetrate tissues such as the brain more effectively while still binding strongly to its target. Although further studies are required to confirm its performance in living systems, the work offers an exciting proof-of-concept for AI-designed therapeutics.
Article :
AI-assisted design of ligands for lipocalin-2.
Sgrignani, J., S. Buscarini, P. Locatelli, C. Guerra, A. Furlan, Y. Chen, G. Zoppi and A. Cavalli
Front Immunol. (2025) 16
