A new imaging approach monitors antibody responses to vaccines more quickly than current techniques, which could accelerate vaccine design
19 January 2022
By analysing high-resolution images, a computer can quickly predict the sequence of amino acids in antibodies. This could cut months off the time it takes to monitor antibody responses during vaccine development.
“It’s a shortcut on a process which typically takes months to achieve with the current set of tools and technologies,” says Andrew Ward at Scripps Research in California.
Antigens, such as the spike protein of the SARS-CoV2 virus, are key components of vaccines. They cause the immune system to produce a range of antibodies against the antigen, but some of these are more useful for the immune response than others. For example, a more useful antibody may block viral entry into a cell while another may not affect this process.
“Antigens are large surfaces so antibodies can target them in many different ways, but usually a small subset does most of the work,” says Ward.
Looking at the ratio of useful “on-target” to less useful “off-target” antibodies resulting from vaccination helps us to optimise the vaccine to tip the response towards more protective antibody production, says Ward.
However, this work takes time. It typically involves sequencing the DNA of individual antibody-producing B cells, generating antibodies from the sequences and then imaging the structure of these antibodies to predict where they bind to the antigen.
Ward and his colleagues have now developed a quicker method. They imaged frozen antibody structures using a technique called cryogenic electron microscopy and designed a computer algorithm that quickly predicts the amino acid sequences of the antibodies based on their structure.
To test their approach, the researchers vaccinated rhesus macaque monkeys using an antigen from HIV, which caused the monkeys to produce antibodies. They then drew blood from two monkeys and mixed each of the samples with the HIV antigen overnight.
The next day, they imaged the antibody-bound antigens in each sample in bulk, which produced detailed maps of the different antibody structures in the samples and how they bound to the antigen.
“This approach allows you to look at all of the antibody responses at once rather than one at a time [as with traditional approaches],” says Ward.
The researchers then focused on one antibody from each monkey and used a computer algorithm to compare the antibody structures with a known library of antibody sequences present in the monkeys, so they could find the existing sequences that best matched the structures.
They made antibodies from the predicted sequences and confirmed that the antibody structures fit those from the original images. The synthesised antibodies also bound in the same way to the antigen as the originally imaged antibodies.
“It’s a transformational tool for vaccine design, and for therapies that rely on antibodies,” says Ward.
Journal reference: Science Advances, DOI: 10.1126/sciadv.abk2039
Sign up to our free Health Check newsletter for a round-up of all the health and fitness news you need to know, every Saturday
More on these topics: