← The Health AI Brief
Loading summary
Transcript1 lines
- [00:01]
A
Every year we witness a familiar cycle with seasonal viruses like influenza and COVID 19. We develop vaccines based on the current version of the virus. But by the time those vaccines are manufactured and distributed, the virus has mutated. This represents a constant evolutionary lag. The traditional approach relies on matching the vaccine to a specific existing strain. The alternative explored today is proactive vaccine design, training our immune systems to recognize stable, unchanging parts of a virus before new variants even emerge. To break this cycle, researchers at the University of Cambridge have used artificial intelligence to design a synthetic vaccine target. The AI analysed the genetic codes of a wide range of different coronaviruses found in nature. By comparing these genetic blueprints, the system mapped out their revolutionary family tree. Through this analysis, the AI identified the core structural elements shared across the entire viral family. It then constructed a completely new synthetic vaccine component, a superantigen. This designed target contains the shared stable features of many different viruses, aiming to teach the human immune system to recognize the entire family, including future variants that don't yet exist. A major challenge in vaccine design is that viruses are covered in highly visible fast mutating regions. These act as immunological decoys. When the immune system encounters a standard vaccine, it naturally focuses on these dominant highly visible decoy targets. When the virus mutates those specific areas, then our vaccine induced immunity loses its effectiveness. To solve this, the research team used a design strategy called Glycan masking. The computational design tools identified where to insert sugar molecules on the surface of the vaccine target. These sugar molecules act as physical shields covering up the highly distracting fast mutating decoy regions. This biochemical shielding hides the highly variable areas. With the decoys hidden, the immune system is guided to focus on and memorize the deep, stable and universal regions of the virus that can't easily mutate. The AI designed vaccine has now undergone its first human testing in a phase 1 clinical trial published in the Journal of Infection. This trial evaluated a candidate named Pevac P.S. which delivered this computationally designed target in the form of a DNA vaccine. The study enrolled 39 healthy volunteers in the UK, all of whom who had previously received multiple doses of standard COVID 19 vaccines. The primary goal of this initial trial was to assess safety and tolerability. The trial was successful in demonstrating that the vaccine safe and well tolerated. There were no serious adverse effects and no significant safety concerns reported among the participants. Additionally, the study successfully used a needle free delivery device using a micro jet of fluid to deliver the vaccine through the skin. This needle free approach simplifies administration and eliminates the need for medical sharps, which is very beneficial for global distribution. So the safety data is very encouraging. Though the overall immune response generated in the trial was modest, while the DNA delivery system produced modest results in humans, parallel research that the authors published in a different journal, MPGA Vaccines, shows that the same AI designed antigens produced very robust broad spectrum protection when delivered using MRNA technology in animal models. Delivering these computationally optimised targets via an MRNA lipid nanoparticle successfully triggered stronger neutralizing antibodies against a wide spectrum of variants including highly mutated Omicron COVID 19 lineages. So the key takeaway is that the computational design of the vaccine target was successful. The AI successfully identified how to mask decoys and highlight stable viral structures. The modest results in the human trial reflect the limitations of the DNA delivery system and the complex background of pre existing human immunity to previous COVID 19 vaccines. But the underlying AI design remains a highly promising scientifically validated strategy. Now the next logical phase of this research is translating these computationally designed shielded antigens onto clinical grade MRNA delivery platforms. This combination will hopefully unlock the full potential of proactive universal viral protection. There's so much happening in this space of drug and vaccine design using AI and we'll do our best to keep updated as they emerge. So do hit like and subscribe if you want to keep abreast of those updates.