UCSF maps protein links to target autism
UCSF researchers mapped 1,800 autism-linked protein interactions, with 87% novel, published Aug. 27 in Science, offering a guide for drug discovery.
UCSF scientists mapped over 1,800 protein interactions linked to autism risk genes, revealing 87% new connections. The study, published Aug. 27 in Science, suggests protein interaction maps could guide drug discovery for autism.
Mapping the Autism Protein Landscape
The research team at the University of California, San Francisco, used 100 autism risk genes to construct a comprehensive protein interaction map. The analysis uncovered more than 1,800 protein interactions, 87% of which had not been reported before. The authors describe this effort as a blueprint for moving from genetic data to actionable drug targets.
| Metric | Value |
|---|---|
| Autism risk genes | 100 |
| Protein interactions identified | 1,800 |
| New interactions (not previously reported) | 87% |
The investigators identified highly interconnected cell hubs that converged on shared protein pathways. They employed AlphaFold, a Google DeepMind artificial intelligence program that predicts 3D protein structures, to determine which connections represented direct interactions.
Key Findings and Functional Validation
One notable interaction involved the forkhead box P1 (FOXP1) mutation, a high-risk mutation in autism, and the FOXP4 protein. In a human organoid model, the team showed that the FOXP1-FOXP4 interaction caused cells to exhibit autism-like characteristics. When FOXP4 was deleted, the normal cell phenotype was rescued, indicating that targeting the interaction could be therapeutically relevant.
The study highlights that proteins are dynamic, undergoing post-translational modifications and forming transient interactions that influence cellular function. According to Nevan Krogan, Ph.D., director of UCSF’s Quantitative Biosciences Institute, “Drug discovery is not exploiting protein interaction data like it should.”
Implications for Drug Discovery
Krogan explained that simply looking at genes and mutations would not reveal the need for a molecular glue or an inhibitor of an interacting protein. The findings suggest that drug discovery could benefit from a protein-protein interaction lens, potentially opening up previously undruggable targets.
The research aligns with earlier efforts, such as Insilico Medicine’s 2021 partnership with Huadong Medicine to prioritize protein-protein interactions in cancer, and Isomorphic Labs’ use of AlphaFold to discover small molecules with major pharma partners.
Future Directions and Industry Collaboration
Krogan told Fierce that he is in “deep talks” with big pharma to develop a drug discovery platform based on these findings. He said, “Let’s do drug discovery with this protein-protein interaction lens or that protein complex lens.”
For more on protein interaction data, see our stats. Our fixtures catalog includes related protein datasets, and the squad section lists researchers involved in similar projects.
Krogan’s statement underscores the potential shift in early-stage drug discovery prompted by this study.
