Halting a deadly virus like Ebola may seem impossible without targeting the virus directly. Yet, researchers at the Broad Institute of MIT and Harvard, in partnership with Boston University’s NEIDL, have taken a novel approach by disabling the virus’s critical support systems inside human cells. Their innovative work is opening unexpected new pathways for therapeutic intervention.
Optical Pooled Screening: A Game-Changer
Ebola outbreaks are rare but catastrophic, and current treatment options are limited. Traditional methods for identifying drug targets struggle with dangerous viruses, but optical pooled screening (OPS) changes the landscape by merging high-content imaging, CRISPR gene editing, and machine learning. OPS lets scientists silence each gene in the human genome across 40 million cells and observe how these changes affect Ebola’s ability to infect and replicate.
Combining CRISPR, Imaging, and Deep Learning
After systematically knocking out individual human genes, researchers infected the cells with Ebola and used advanced imaging to track infection outcomes. Tools such as CellProfiler and deep learning models enabled analysis of how Ebola progresses through infection stages at the single-cell level. By sequencing CRISPR guide RNAs, the team pinpointed which genes were disrupted and how this impacted viral replication.
- Hundreds of viral regulators identified: Many human proteins, when silenced, blocked Ebola’s ability to multiply.
- Key viral processes uncovered: Genes crucial for viral entry and replication, as well as those affecting inclusion bodies (Ebola’s “viral factories”), surfaced as potential drug targets.
- A new mitochondrial connection: Disabling the gene UQCRB revealed that mitochondria unexpectedly play a role in Ebola infection. Targeting UQCRB with a small molecule inhibitor significantly reduced infection rates without harming healthy cells.
- Balance of viral RNA and protein: Knocking out genes like STRAP shifted this balance, offering new leads for drug discovery.
Implications for Other Deadly Viruses
The OPS platform isn’t limited to Ebola. Follow-up screens showed that silencing some of the same host genes also disrupted the Sudan and Marburg viruses, dangerous relatives of Ebola with no current treatment. This paves the way for broad-spectrum therapies that target host factors shared by multiple viruses.
OPS and Machine Learning: Unlocking Cellular Secrets
OPS’s strength lies in simultaneously measuring many features in single cells, providing a nuanced map of virus-host interactions. Machine learning allowed researchers to analyze infection stages and cell responses at a scale never before possible, accelerating both basic science and drug discovery.
Expert Insights and What’s Next
Paul Blainey, MIT professor and co-senior author, emphasized the power of OPS to explore viral dependencies throughout infection. Robert Davey of BU highlighted AI-driven analysis as a historic deep dive into Ebola’s manipulation of human cells. While co-first author Rebecca Carlson noted that OPS’s ability to capture multiple cell features is unmatched. The team is now pursuing the most promising gene targets for future antiviral therapies.
Takeaway: Paving the Way for Next-Generation Antivirals
This research marks a significant advancement in antiviral discovery, shifting the focus from the virus itself to the human proteins it relies on. By combining CRISPR, advanced imaging, and AI, OPS enables rapid identification of new drug targets and expands our understanding of viral infection strategies. As this platform evolves, it could unlock treatments for not only Ebola but a range of deadly pathogens.
CRISPR and AI Are Revolutionizing Ebola Drug Discovery