A team at the Universitat Autonoma de Barcelona has designed a small antibody fragment, called scFv-h3D6, that targets the amyloid-beta peptide without the structural region that pulls immune cells into the brain. The Protein Design and Immunotherapy group, led by Chair Professor Sandra Villegas of the Department of Biochemistry and Molecular Biology, reports that the fragment works at the molecular, cellular and cognitive levels in mice.
The problem it aims to solve is well documented. The European Medicines Agency approved two anti-amyloid antibodies in 2025, lecanemab from Biogen and donanemab from Eli Lilly, the first treatments shown to modify the course of Alzheimer's rather than just ease symptoms. Both carry a characteristic risk of bleeding in the brain, visible on MRI, affecting between 10 and 27 percent of treated patients in clinical trials. The risk is concentrated in people who carry the APOE epsilon 4 allele, and European regulators consequently restrict these drugs to patients with one copy of the variant or none at all.
Villegas's group has argued for years that full-length antibodies may be inherently unsafe in the brain. The brain lacks the cells and molecules of the systemic immune system, and full antibodies recruit those immune cells across the blood-brain barrier, which can destabilise it and lead to bleeding. Getting antibodies into the brain is necessary for the drugs to clear amyloid, but the same access appears to drive the side effect. Cutting away the recruiting region is meant to keep the amyloid-clearing action while removing the trigger.
If the approach holds up, it could widen access to disease-modifying treatment for the substantial group of patients currently excluded on genetic grounds, particularly APOE epsilon 4 homozygotes. The evidence so far comes from mouse models, and the fragment has yet to be tested in people, so the safety advantage remains a hypothesis supported by animal work rather than a clinical result.