(687f) High-Throughput Isolation of Influenza Antibodies

Influenza A virus (IAV) primarily targets the respiratory system. Its segmented genome and reassortment ability, along with mutations due to error-prone polymerase activity, results in a rapidly mutating virus that is difficult to predict. Consequently, several flu pandemics have occurred over the past century and flu epidemics cycle every year. The current method of reformulating influenza vaccinations each year is both time- and resource-intensive. Thus, to protect humans from this pathogen, scientists are striving to develop a universal flu vaccine that can stimulate a protective immune response against all IAV strains. The continued discovery of broadly neutralizing influenza antibodies could enable the development of more effective therapeutics and vaccines that protect against most flu strains. However, current technologies lack the capability to achieve rapid, high-throughput screening of full antibody sequences for neutralization activity. Many of the current workflows that screen libraries of sizes 10⁴ rely on binding affinity of antibody fragments and purified antigens to select for improved antibodies. However, binding is not a direct indicator of neutralization activity. The fastest neutralization assays are currently performed in a well plate, which limits throughput to approx. 10² - 10³ over multiple days.

Emulsion droplets provide the ability to compartmentalize and miniaturize bulk experiments rapidly, allowing for the study of thousands or millions of antibodies in a short time period. Here, we show that a droplet-based platform can be used to positively select neutralizing antibodies. Our system uses a microfluidic chip to generate emulsion droplets that capture virus particles with a fluorescent marker and antibody-secreting cells that can be infected by the virus. If a cell is infected, it will express a fluorescent reporter, often GFP. However, if a cell is expressing a neutralizing antibody, it will be protected from infection and will not express the fluorescent reporter. The cells can then be harvested from droplets and sorted using a FACS machine into GFP+/- groups, wherein the cells expressing neutralizing antibodies are enriched in the GFP- group. The genetic material is then extracted and sequenced such that the antibody sequences of strong neutralizers are known.

In this study, we demonstrate that infectivity rates of cells in droplets can be adjusted by tuning variables including time, virus concentration, and environment. Using the highest concentration of virus and incubating cells in droplets leads to infectivity rates of approximately 40%, an improvement compared to other disease models used in the lab. We also demonstrate that a weakly neutralizing (against the PR8 strain) flu antibody, H2526, efficiently reduces the infectivity in droplets to support high-throughput screening. Additional efforts with panels of anti-PR8 antibodies and libraries of anti-PR8 antibodies will also be discussed.

We are currently applying this platform to screen larger libraries directly for their neutralization function, including for site-saturation mutagenesis screens of established antibodies, and to discover new anti-flu antibodies from diverse immune response libraries. The findings from these studies will support new scalable workflows to rapidly discover anti-influenza antibodies that enhance influenza treatments and preventions, including against potential pandemic influenza strains.