• 17 November 2026
    • DOI:10.20944/preprints202611.1636.v1
Abstract
Six years after the first emergency use authorizations, mRNA-lipid nanoparticle (mRNA-LNP) vaccines have gone from a platform without a licensed product to one that supplies a large share of the world's COVID-19 and respiratory syncytial virus (RSV) vaccine doses and has completed phase 3 trials for seasonal influenza. Their performance depends on mechanisms that are being resolved with unusual speed at the level of individual cell types, receptors and molecular interactions, and several textbook assumptions have had to be revised in the process. This review follows the vaccine from the formulation vial to the germinal centre. We first describe how the RNA is engineered for efficient translation and for partial escape from cytosolic and endosomal RNA sensors, and why N1-methylpseudouridine is neither immunologically silent nor translationally neutral. We then examine how the ionizable lipid, the helper phospholipid, cholesterol and the PEG-lipid determine particle behaviour, cell uptake and endosomal release, and how the lipid and RNA components act as separable adjuvant signals. Evidence from mouse models and from human lymph node sampling is compared, with attention to the substantial species differences in innate cytokine control that limit extrapolation from mice to people. The clinical section summarizes efficacy and immunogenicity data for COVID-19, RSV and influenza vaccines, and the preclinical and early clinical status of candidates against cytomegalovirus, mpox, HIV, malaria, Nipah virus and Clostridioides difficile. Safety data are discussed in absolute terms rather than as signals alone. We close with the questions that currently block rational improvement: the molecular identity of the lipid sensor, the difficulty of separating immunogenicity from reactogenicity, the absence of a quantitative description of endosomal escape, and the design of antigens for pathogens that do not tolerate a single fixed immunogen.