Researchers have successfully tested an AI-designed vaccine in humans that trains the immune system to recognize a broad range of coronaviruses, marking a major shift from reactive vaccines to preventive ones. The Phase 1 trial of pEVAC-PS, conducted by the University of Cambridge and its spin-out company DIOSynVax, enrolled 39 healthy volunteers and demonstrated that the vaccine was well tolerated with a clean safety profile. This represents the first time a vaccine whose active antigen was designed entirely through computational modeling has been tested in humans.

Why Does the Immune System Need to Recognize Virus Families, Not Just Current Strains?

Traditional vaccines have always worked the same way: scientists identify a circulating virus strain, develop a vaccine against it, and roll it out to the public. But by the time the vaccine reaches people's arms, the virus has often mutated into new variants, making the vaccine less effective. Over the past two decades, three major coronavirus outbreaks have shown how quickly these viruses can jump from animals to humans and spread globally. With roughly a 50 percent chance of another COVID-scale pandemic occurring within the next 25 years, waiting to react after an outbreak begins is no longer a viable strategy.

The AI-designed vaccine takes a fundamentally different approach. Instead of targeting one specific virus strain, it uses machine learning to analyze global coronavirus sequence data and identify conserved structural features shared across the entire sarbecovirus family, including viruses that have not yet crossed into humans. This means the vaccine trains your immune system to recognize common patterns that appear in SARS-CoV-2, SARS-CoV-1, and related bat coronaviruses with zoonotic potential, creating what researchers call a "super-antigen."

"We've converted vaccine development from being reactive to being future proof. Our vaccines will continue to provide protection against viruses even as they mutate into new strains," said Jonathan Heeney, pathologist and virologist at the University of Cambridge and the scientific lead of the research.

Jonathan Heeney, Pathologist and Virologist at the University of Cambridge

What Happened in the First Human Trial?

The trial tested the vaccine's safety and immune response in 39 healthy volunteers aged 18 to 50, all of whom had previously received two or three doses of a licensed COVID-19 vaccine. Participants received escalating doses of pEVAC-PS delivered as a DNA construct via needle-free intradermal injection, with doses given at day zero and day 28. The needle-free delivery method offers practical advantages: DNA vaccines are more stable than mRNA vaccines, can be manufactured and adapted at scale, and avoid sharps waste, making them easier to deploy globally, especially in low-resource settings.

The safety results were encouraging. Across all four dose levels, the vaccine was well tolerated, with the majority of adverse events being mild or moderate. Notably, fewer adverse events were reported after the second dose than after the first, suggesting that repeated intradermal administration was safe and well accepted by participants.

How Well Did the Vaccine Train the Immune System?

The immunogenicity results were more complex to interpret. Binding antibody responses to the vaccine construct were detectable, particularly at the highest dose, but responses were modest and variable. Small increases in neutralizing antibodies were seen against the Delta and Omicron variants of SARS-CoV-2, while activity against the ancestral Wuhan strain and SARS-CoV-1 was limited. A major reason for this variability was the timing of the study: recruitment took place between December 2021 and September 2023, during successive waves of Omicron infections and booster vaccination campaigns, meaning participants entered the study with widely varying immune histories.

However, the team used a more sophisticated analysis called peptide microarray analysis to probe whether the vaccine was directing immunity toward conserved viral regions, the central design goal of the platform. This revealed antibody binding to conserved receptor-binding domain epitopes, including regions corresponding to known broadly reactive antibody sites. While such binding does not necessarily translate directly into strong laboratory neutralization tests, these epitopes have been associated with protection in living organisms through Fc-mediated immune mechanisms, a process where antibodies recruit other immune cells to attack the virus.

Steps to Understand How This Vaccine Approach Differs From Traditional Vaccines

• Reactive Model: Traditional vaccines are developed after a virus outbreak begins, targeting the specific strain currently circulating. Manufacturing and distribution delays mean vaccines often lag behind viral evolution, offering limited protection against future spillovers from related viruses.
• Proactive Design: AI-designed vaccines analyze genetic sequences from entire virus families to identify shared features, allowing a single vaccine to protect against multiple related viruses before they emerge in humans.
• Computational Advantage: Machine learning identifies conserved structural patterns that humans might miss, enabling the creation of "super-antigens" that train the immune system to recognize broad viral categories rather than individual strains.
• Delivery Innovation: DNA vaccines delivered via needle-free intradermal injection are more stable than mRNA vaccines, easier to manufacture at scale, and more practical for global deployment in resource-limited settings.

Collectively, the trial data suggest that while pEVAC-PS did not generate robust neutralizing responses in this small Phase 1 study, it provided evidence that the vaccine successfully focused the immune system on shared coronavirus features, supporting the feasibility of the antigen design strategy.

"Viruses like Influenza, Coronaviruses and the Ebola group are evolving continuously and by the time vaccines are rolled out, they may be poorly matched. The current reactive vaccine system struggles to keep pace. If we can develop and clinically advance this new class of vaccines before a virus outbreak begins, millions of lives could be saved, lockdowns avoided and the economy preserved," explained Saul Faust, clinical researcher in pediatric infectious diseases and immunology at the University of Southampton and the trial's chief investigator.

Saul Faust, Clinical Researcher in Pediatric Infectious Diseases and Immunology at the University of Southampton

The results mark a significant proof of concept for AI-guided vaccine design targeting entire virus families. By demonstrating safety and immune engagement with conserved sarbecovirus regions, the study lays the groundwork for refining this approach and advancing broader, more durable coronavirus vaccines. This shift from reactive to proactive vaccination represents a fundamental change in how the immune system can be trained to protect against threats that haven't yet emerged.