Scientists Use CRISPR and AI to Discover New Ebola Drug Targets

Scientists at the Broad Institute of MIT and Harvard, in collaboration with researchers at Boston University’s National Emerging Infectious Diseases Laboratories (NEIDL), have employed a novel technique called optical pooled screening (OPS) to uncover human genes critical to Ebola virus infection. This breakthrough, published in Nature Microbiology, could pave the way for new therapeutic strategies by targeting host proteins the virus depends on, rather than the virus itself. Ebola, though rare, remains a severe and often fatal disease with limited treatment options. Traditional methods for identifying host factors that regulate viral infection are challenging, particularly for high-risk pathogens like Ebola, which require strict biosafety protocols. The new approach overcomes these barriers by combining CRISPR gene-editing with high-content imaging and artificial intelligence (AI) to analyze how silencing specific human genes impacts the virus’s ability to replicate. The team tested nearly 40 million human cells, each with a single gene knocked out using CRISPR. After infecting the cells with Ebola, they fixed and inactivated them to allow analysis outside high-containment labs. Using machine learning, they examined images of the cells to detect changes in viral protein and RNA levels, while an AI model developed by the Broad Institute’s Eric and Wendy Schmidt Center helped classify the infection stage of individual cells. This allowed researchers to identify hundreds of host proteins involved in different phases of Ebola’s life cycle, including those necessary for viral entry and replication. One key discovery was the role of the gene UQCRB, which is linked to mitochondria. Silencing UQCRB in cells reduced Ebola infection without harming the cells’ normal function. A small molecule inhibitor targeting this gene showed promise as a potential therapeutic. Another gene, STRAP, when disrupted, altered the balance of viral RNA and protein, suggesting it could be a target for intervention. Further experiments are underway to explore these genes’ roles and their viability as drug targets. The method also revealed that silencing certain genes disrupted replication of related filoviruses, such as Sudan and Marburg, which share similarities with Ebola but lack approved treatments. This implies the approach might lead to broad-spectrum therapies for multiple viruses. OPS, developed by the Broad Institute, merges high-content imaging—which captures detailed cellular changes—with pooled CRISPR screens, which efficiently test genetic elements. By analyzing vast datasets, the technique enables researchers to map viral dependencies with unprecedented precision. “AI gave us an unprecedented ability to do this at scale,” said co-senior author Robert Davey, director of BU’s NEIDL. The study’s co-first authors, Rebecca Carlson and J.J. Patten, emphasized that the method allows simultaneous measurement of multiple cellular features, offering insights into virus-host interactions that traditional screens cannot. This could accelerate the discovery of treatments for hard-to-treat infections. Funded by institutions including the Broad Institute, National Institutes of Health, and others, the research highlights the growing role of AI in biomedical science. By focusing on host factors, the team aims to reduce reliance on direct viral targeting, which can be less effective due to rapid viral mutations. The findings underscore the potential of OPS to transform how scientists study complex pathogens and develop therapies for emerging infectious diseases.