A hemagglutinin and neuraminidase biased immunological memory shapes the dynamics of antibody responses to Influenza A virus

Abstract As influenza A virus (IAV) infection establishes a more diverse immunological memory to different viral proteins compared to vaccination, we hypothesize that this could skew antibody responses to immunizing antigens after infection. To explore this, we generated mouse models that possess either an IAV hemagglutinin (HA)- or neuraminidase (NA)-biased immunological memory by simultaneously or, sequentially-inoculating them with cocktails of isogenic viruses bearing four antigenically-distinct HA (H3v) or NA (N2v), that spans the IAV H3N2 human circulation history. Cocktail priming elicited comparable antibody responses to all immunizing antigens, whilst responses in sequentially-inoculated mice were antigen-dependent. We then challenged the mice with two H3N2 strains of opposing virulence and antigenic distance and examined the post-infection antibody landscapes. In both models, antibody increases after the two challenges occurred to antigenically-related, and not against antigenically-distant, HAs or NAs for which baseline titers were low. H3-imprinted mice seroconverted to challenge NA, but not HA, and vice versa. In primed mice, the antibody response was directed to the viral target with low pre-existing immunity, despite high titers of protective antibodies and no signs of symptomatic infections, while immune-naïve mice seroconverted to both HA and NA. Our findings show that the dynamics of antibody responses to HA and NA after infection can be altered by pre-existing immunity, i.e the “antibody ceiling effect” does not imply no-response, simply a "misdirected" response, to an antigenically-related antigen or an alternate viral protein. These results have implications on our seroepidemiological studies as well as our understanding of immune correlates of protection in populations with different influenza immune histories.