Objectives: Respiratory syncytial virus (RSV) is a widespread contagious virus that causes respiratory infections. RSV causes infections at all ages and repeated RSV infection is common throughout life. The aim of this analysis was to develop semi-mechanistic models to describe i) cellular and ii) humoral immunity dynamics after two times of vaccinations of compound X, a vaccine candidate for RSV infection, to eventually support dose selection.
Methods: The immunogenicity data for this analysis, IFN-γ response for cellular immunity and anti-RSV/A neutralization activity for humoral immunity, were obtained from a Phase 2 dose-finding study of compound X. The cellular immunity model was a two-compartmental model of active or memory IFN-γ secreting CD4+ T cells and included an antigen compartment. It assumes that memory T cells exist at baseline due to previous infection. The humoral immunity model was a four-compartmental model of active or memory B cells, long-lived plasma cells and neutralizing antibody titer and included an antigen compartment. It also assumes the existence of memory B cells, long-lived plasma cells and neutralizing antibody at baseline, attributable to previous infection. Given the complexity of the structural models and the sparsity of available data, the Cluster Gauss-Newton method (CGNM) [1] was used to determine fixed parameter values for structurally and/or practically unidentifiable parameters and initial values of parameters to be estimated. Model evaluation was based on the parameter precision, objective function value, biological plausibility of parameter estimates, goodness-of-fit (GOF) plots, and visual predictive checks (VPC).
Results: Key parameters related to the dynamics of active and memory T cells for the cellular immunity model as well as the dynamics of memory B cells and neutralizing antibodies for the humoral immunity model were well estimated with high precision ( < 20% relative standard error). Final models described the data adequately with predicted immune responses consistent with observed immune responses in GOF and VPC across all antigen doses. Additionally, the models accurately captured limited immune responses observed after revaccination compared to immune response observed after the initial vaccination which is in line with other RSV vaccines in older adults [2, 3].
Conclusions: Cellular and humoral semi-mechanistic immune response models for an RSV vaccine candidate were developed. These models will be utilized as a platform to support and accelerate future model-informed vaccine dose selection and development.
Citations: [1] Aoki, Y., et al. Optimization and Engineering, 2022; 23:169-199. [2] Kotb, S, et al. Respiratory Investigation, 2023, 61.2: 261-269. [3] Baber, J, et al. The Journal of Infectious Diseases, 2022, 226.12: 2054-2063.
