A study led by the University of Washington researchers showscases a new way to generate potential drugs using computational design. The approach holds promise for accelerating the long, cumbersome and expensive process of drug development.
Many current drugs are protein-based and bind specific targets – the drug Sotrovimab, which binds the COVID-19 virus, is one example. The new software tools similarly design proteins to stick to viruses or molecules in the body involved in diseases like diabetes and cancer.
While software is increasingly used in therapeutic protein design, the new study offers a key advance. The new tools generated designs using only a minimal amount of data as input: the known, three-dimensional structure of the target. The resulting proteins are also small and sleek, and easily synthesized.
Other software tools can also yield such blueprint designs, but they generally require more detailed input, such as data on how the target interacts with cellular molecules. And other tools may not spit out the optimal range of designs, said the authors of the study, led by researchers at the UW’s Institute for Protein Design.
According to the IPD, its new software can “scan a target molecule, identify potential binding sites, generate proteins targeting those sites, and then screen from millions of candidate binding proteins to identify those most likely to function.”
Using the new tools, the researchers designed proteins against 12 key molecular targets, including the insulin receptor, the COVID-19 virus, and several cancer-causing molecules. Proteins synthesized to match their designs bound their targets tightly and did not easily degrade, suggesting they might be good drugs for diseases like diabetes or cancer and COVID-19.
“When it comes to creating new drugs, there are easy targets and there are hard targets,” said co-first author Longxing Cao in a press release. “In this paper, we show that even the hard targets are amenable to this approach.”
The study was published Thursday in the journal Nature and was previously posted on the preprint server bioRxiv. The research is “protein design at its best,” said Alexander Hauser, a computational biologist and assistant professor at Copenhagen University. a tweet. The approach could result in faster, more efficient processes for drug discovery.
A similar approach underpins IPD spinout Neuleukin Therapeutics as well as “a couple very early stage commercialization groups still here at the IPD,” according to spokesperson Ian Haydon.
Future studies will combine the new methods with another tool developed by IPD researchers: RoseTTAfold, which predicts the three-dimensional structure of proteins using artificial intelligence and won the “Breakthrough of the Year” award last December from Science magazine.
Meanwhile, the researchers plan more extensive laboratory testing of the 12 proteins from their new study.
“We look forward to seeing how these molecules might be used in a clinical context, and more importantly, how this new method of designing protein drugs might lead to even more promising compounds in the future,” said IPD co-first author Brian Coventry. in the release.