Calibrating T cell help by antigen cross-linking as a universal vaccine platform against rapidly evolving pathogenic viruses

Abstract

Abstract The effectiveness of influenza vaccine is limited by the heterogeneous immune response it elicits. Factors such as age, sex, and pre-existing antibody titers only partially explain this heterogeneity. Using a novel deconvolution analysis we find subtype-specific bias as a major source of this inter-individual variation. Since the 1940’s it has been proposed that this phenomenon is due to prior strain exposure, referred to as “Original Antigenic Sin”. But our analysis of flu vaccine responses in a monozygotic twin cohort showed that antigenic imprinting from past exposures can only partially explain strain preference (r 2=0.25, p=4.7×10 −5). Thus, we hypothesized that host genetics may play a major role in subtype-bias. Consistent with this is our finding that strain-preferences are evident after vaccination in an influenza naïve infant cohort indicating that prior exposure is not essential. We determined the source of this genetic component to the wide variation in influenza hemagglutinin peptide presentation by different HLA class-II alleles resulting in an imbalanced, subtype-specific CD4 +T follicular help after vaccination. To address this, we developed a tunable, plug-and-play vaccine platform to better calibrate T cell help. In a human tonsil organoid system, we demonstrated that engineered, cross-linked HA induces a significantly higher response against multiple influenza subtypes in comparison to the inactivated influenza vaccine. In summary, we have developed an effective strategy to overcome response bias in a population with diverse genetic backgrounds and pre-existing immunity. Furthermore, we successfully adapted this platform to generate a pan-coronavirus vaccine. Supported by grants from NIH to MMD (5U19AI090019, 5U19AI057229). Additional support was provided by ITI-YIA to VM.