Modeling Variations in Antibody Response Magnitude and Longevity

Long-lasting antibody responses are pivotal for both protective immunity and autoimmunity. Yet, the intricate mechanisms that dictate the duration of these responses remain only partially elucidated. By employing an agent-based in silico model, we simulated the generation of short-lived and long-lived plasma cells during the immune response to an adenoviral COVID-19 vaccine, postulating that antigen-specific plasma cells have a certain probability of attaining an extended half-life. This hypothesis implies that the quantity of antigen-specific plasma cells generated in the initial immune response, coupled with their likelihood of becoming long-lasting, influence the magnitude of the antibody response months after immunization. Interestingly, our simulations unveiled two distinct clusters among individuals several months post-vaccination, delineating markedly divergent dynamics in antibody titers: one group exhibited sustained elevated antibody levels (sustainers), while another witnessed a decline (decayers). Notably, the absence of long-lived plasma cells in the decayers distinguished them from the sustainers. Leveraging machine learning clustering on antibody titers, we achieved an accuracy of 0.925 in identifying the decayers 28 weeks following the initial dose.In this in-silico system, the difference between sustainers and decayers stems from stochastic inter-individual differences in the immune repertoire and the efficacy of priming. We speculate that, in real life, aged and immunocompromised people may be prone to the decayer pattern and may benefit from receiving their vaccine booster after a shorter interval. We are comparing our model's predictions with clinical data on the antibody response to SARS-CoV-2 Nucleoprotein and Spike post-COVID-19 infection or vaccination to validate and refine our model. Specifically, we are harnessing machine learning methodologies on data sourced from published immunological studies to discern patterns in the dynamics of the antibody response.