Modeling the Magnitude and Persistence of the Antibody ResponseThrough the Contributions of Different Subsets of Antibody-SecretingCells

Background The durability of vaccine-induced antibody responses varies widely across individuals and platforms, yet the underlying cellular dynamics remain elusive. Methods We present a mechanistic modeling framework to dissect the contributions of distinct subpopulations of antibody-secreting cells (ASCs). Using an agent-based model of COVID-19 vaccination, we simulated ASC generation and antibody output, accounting for probabilistic cell fate decisions and examining how antigen dose, the probability of generating long-lived ASCs, and their half-life shape antibody kinetics. Insights from these simulations informed the development of the two-ASC model, which decomposes antibody trajectories into biologically interpretable components attributed to plasmablasts and long-lived ASCs, and defines the waning pattern using two variables: Magnitude (response size) and Persistence (long-lived ASC contribution). We further expanded this approach in the three-ASC model, which allows theoretical estimation of the contribution of plasmablasts, short-lived plasma cells, and long-lived plasma cells. Results Within COVID-19 mRNA vaccine cohorts, Magnitude and Persistence were inversely correlated, indicating that peak responses are a poor predictor of durable antibody levels. In fact, the composite metric M×P, representing durable antibody levels, converged around vaccine-specific values despite individual variability in Magnitude. Notably, after two doses of mRNA vaccines, the component attributed to long-lived plasma cells was often negligible in virus-naive individuals, but more frequently detectable in those with hybrid immunity. Implications These models provide a unifying conceptual framework for describing antibody waning, linking serological trajectories to the underlying populations of antibody-secreting cells.