The differential effect of sub-micron level HA aggregates on influenza potency assays

Abstract Influenza vaccines remain the most effective public health measure for the prevention of influenza-related illnesses. The primary immunogen in inactivated influenza vaccines is hemagglutinin (HA), the receptor binding protein of influenza. The concentration of HA during vaccine production and testing is standardized according to the level of antigen as measured by Single Radial Immunodiffusion Assay (SRID). This allows vaccine potency to be controlled such that individuals receive a dose known to provoke a clinically protective immune response. As compared to alternatives, SRID has the advantage of quantifying immunologically relevant forms of HA, but it depends on timely generation of novel reagents for each new vaccine strain. In recent years, a number of alternative assays have been suggested based on either epitope recognition, receptor binding or protection from proteolysis but it is unclear how they relate to vaccine potency in clinical trials. In this report we describe the development of a lectin-based, ELISA-type assay for HA potency and find it provides similar potency estimates to SRID except in the case of a vaccine with aggregated HA and other viral proteins. In that case, SRID predicted the immunologically active HA present and ELISA techniques did not. This difference was due to tested antibodies failing to pull down or bind to the HA present unless particle aggregates were first dissociated. Furthermore, detergent treatment alone was insufficient to complete this dissociation. While others have previously demonstrated that immunocapture-based techniques can misestimate the potency of influenza vaccines depending on the individual antibodies used we demonstrate that in this case the failure was due to an inability of all antibodies to capture HA contained in the aggregated influenza vaccine.