Picture credits: Achiko
A new approach to vaccines with a touch of machine learning could put an end to recalls and seasonal variations, say researchers at MIT. This “pan-variant” vaccine would ignore the virus itself but quickly control infections by cracking down on infected cells.
To be completely clear, this is still in animal testing and a long way from being rolled out. But with COVID becoming a resident virus in the human population, longer-lasting solutions than occasional boosters for particularly bad strains are in demand.
The problem is that, as amazing as mRNA vaccines are, they’re reactive, not proactive: you see a variant, you sample its spike protein or other distinguishing feature, and you slip it into the immune system so it knows that she is on the look for it. It’s a bit like letting a rescue dog sniff out a lost hiker’s belongings.
Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory wanted to find another, more sustainable way to protect the body against COVID attacks. A paper describing their findings was published today in the journal Frontiers in Immunology.
The team decided to pitch the idea of attacking the virus itself because its most distinctive feature, the spike protein, is constantly changing. Instead, they focused on certain molecular signals that reliably appear on cell surfaces. infected by the virus. If these could be spotted early and the immune system’s T-cells deployed rapidly, the infection would be stopped before it reached dangerous or even potentially infectious levels.
These surface signals, called human leukocyte antigens, present a variety of peptides to T cells, much like raising semaphore flags. If everything is in order, it’s the usual combination of familiar peptides and the T cell moves on. If something goes wrong, a fragment of the virus can be hoisted onto the flagpole and the T cells open fire.
So what does machine learning have to do with all of this? There are lots of data listing the different proteins and chains of amino acids found in COVID, and what they become once they have infiltrated a cell, and how cells indicate using HLAs that they are infected.
Machine learning algorithms are good for solving optimization problems like this, where lots of noisy data needs to be sorted for specific combinations of qualities. In this case, they designed algorithms to catalog the relevant peptides and select about 30 of them that are present or “conserved” in all versions of the virus, but are also associated with HLAs and are likely to be used as flags for that T cells can see them.
Transgenic mice that received our versions of HLA and this new vaccine showed a much larger immune response in the short term after infection, and none died from the virus.
“This study provides evidence in a living system, a real mouse, that vaccines we designed using machine learning can provide protection against the COVID virus,” said MIT PhD student Brandon Carter. one of the main authors of the article, in an MIT news article. .
An interesting possible benefit is that immunocompromised people can get significant protection from this approach when mRNA vaccines don’t work for them. People suffering from “long COVIDs” may also get some relief in the form of a fuller immune attack against their particularly resistant infection.
As the abstract of the study puts it:
The undetectable specific antibody response in mice immunized against MIT-T-COVID demonstrates that specific T cell responses alone can effectively attenuate the pathogenesis of SARS-CoV-2 infection. Our results suggest that further study is warranted for pan-variant T-cell vaccines, including for people who cannot produce neutralizing antibodies or to help mitigate Long COVID.
It’s a promising avenue of research and a great way to use advances in computing for global health. But it’s also important to recognize that the “pan-variant” option is still in its infancy. For one thing, it can work with or against existing vaccines – what if the best peptides for the immune-response vaccine were those targeted to be destroyed by mRNA priming? The two would work at cross purposes. And too strong an immune response also runs the risk of collateral damage, mistaken elimination of ambiguous signaling cells, etc.
But these are the right questions – the ones that are relevant because the basic function of the new vaccine appears to be working. We’ll know more as the team progresses through further testing of this promising approach.