Building a vaccine for future versions of the virus

But five years later, SARS-COV-2 is still circulating, and in the process it has evolved into a new variant that requires the latest vaccines to protect against them.

However, it takes time to design, manufacture and distribute new vaccines. This raises an important question. How can scientists create vaccines with versions of viruses that have not yet been released?

One solution comes from a predictive AI model called EVE-Vax, built by a team of scientists at Harvard Medical School, the Massachusetts Consortium on HMS-led Pathogen Preparation (MASSCPR), and other institutions.

New model explained on May 8th Immunity, Evolutionary, biological, and structural information about the virus is used to predict and design surface proteins that are likely to occur when pathogens mutate. Researchers have designed a viral protein that adapts EVE-VAX to SARS-COV-2 and induces a similar immune response as a real protein that evolved during the Covid-19 pandemic.

The study, supported in part by federal funding, suggests that this model can provide valuable information on future evolution of the virus and can be used to create a panel of “designer” proteins to assess future protections of vaccines. Researchers hope that the model will ultimately help scientists develop better vaccines and combat the outbreak and pandemics of viruses caused by rapidly mutating viruses.

First author Noor Youssef, the scientific lead in predictive modeling for HMS' Markslab's vaccine design group, said:

Eve's evolution

More than a decade ago, Debora Marks, co-author of Debora Marks, a professor of systems biology at HMS' Blavatnik Institute, began investigating whether millions of years of evolutionary genetic information could be used to predict protein structure and function.

In a 2021 paper, the researchers described the AI ​​model they created based on this idea. A model called EVE (Evolutionary Model of Variant Effects) uses large-scale interspecies evolutionary data to predict whether proteins function or not. When applied to humans, Eve was able to interpret genetic mutants as benign or as causes of illness.

As the Covid-19 pandemic unfolds, Marks and her team adapted the model to predict virus behavior.

They built Evescape, which applied the power of Eve's protein prediction to viral proteins. In a 2023 paper, scientists showed that if Ebscapes were present at the start of the Covid-19 pandemic, they predicted the most frequent SARS-COV-2 mutations and discovered the biggest variant of concern that is most likely to cause spikes in human infections.

The success of Evescape has led researchers to wonder whether their models can predict the future evolution of rapidly evolving viruses such as SARS-Cov-2.

Vaccines for such viruses are updated annually, so scientists need to make a year-old guess about how viral proteins evolve. This could lead to discrepancies between the predicted version of the virus used to design the vaccine and the actual version that circulates. The gap between prediction and reality can make the vaccine less effective than if the versions were more closely matched.

To solve this problem, researchers developed EVE-Vax. Evebax is a model that predicts and designs viral proteins that can be used to inform vaccine development in advance.

“We wanted to use our methods to see if we could create a brand new protein that is functional and has the same immune response as seen in real viruses,” Youssef said.

Predict future operations of SARS-Cov-2

In their latest research, scientists have used EVE-VAX to design 83 latest versions of the “spike” protein for SARS-COV-2, the main surface protein used by viruses to infect human cells. Newer versions of the spike protein had different combinations of up to 10 mutations.

To test the effectiveness of AI-designed proteins, the researchers teamed up with their colleagues and co-authors with Jeremy Luban, professor of molecular medicine at UMass Chan Medical School. Jacob Lemieux, assistant professor of medicine at Massachusetts General Hospital; Michael Sheiman, HMS Associate Professor at HMS Beth Israel Deacones Medical Center. Scientists conducted experiments in lab dishes using a safe, non-replica version of SARS-COV-2, designed for research. This experiment confirmed that the “designer” spike protein infects human cells infects human cells and induces immune responses that are roughly consistent with the real-life immune response to the virus at five different times during the Covid-19 pandemic.

“The basic insight here is that evolution will tell you what a virus and its proteins can do and what will happen in the future,” Marks said.

Finally, researchers have shown that hundreds of new spike proteins can be easily and cheaply engineered, which can be easily incorporated into the development of SARS-Cov-2 vaccines.

“Traditional vaccine designs use all sorts of different methods, but no one has used this approach before,” Marks said. “EVE-Vax opens up new areas of potential applications and design for vaccines.”

For example, researchers showed that EVE-Vax could predict a significant immune escape from the Covid-19 booster vaccine targeting omicron mutants.

“Using Evebax allows you to predict the immune response, rather than what the virus mutation is, which is more useful in real life situations,” Marks said.

The team is now spreading Evebax to other viruses, including avian flu, a growing problem in the US and around the world.

One important advantage of this model is that it works with limited information and can be opened for unstudied viruses such as Lassa and Nipah, as well as newly emerging viruses.

Ultimately, the team hopes Evebax will provide scientists with important information about the various ways the virus is likely to evolve, allowing them to design vaccines that will protect them from future versions of the pathogen.

Author, Funding, Disclosure

Paper Greff, Fadi Guntu, Kele Brock, Kelly Jaimé, Nicole Tantoni, Anne Dauphin, Amie Selman, Jorkoveckisk, Jorkoveskisk, Lalbot Atanconiel, Solomon Alexand Jas, Rose Elassemus, Debra Fuller, Fimer Pamela Bjorkman.

Funding for the work was provided by the coalition of Preparation Innovation for the Epidemic (CEPI), MASSCPR, and the National Institutes of Health (R37 AI147868).

Marks is an advisor to Dyno Therapeutics, Octant, Jura Bio, Tectonic Therapeutics and co-founder of earthquake treatment.