Adjuvants have diverse, but often subtle, impacts on vaccination-induced immune responses that are important for vaccine effectiveness. In-depth profiling of vaccine-induced cytokine, cellular, and antibody responses coupled with machine learning has the potential to uncover adjuvant-specific immune response features that might guide rational adjuvant selection. Researchers evaluated human immune responses produced by vaccinations adjuvanted with two comparable, clinically relevant adjuvants, AS01B and AS02A, and found significant differentiating features, or immunological signatures, that they imprint on vaccine-induced immunity. The samples for this side-by-side comparison came from malaria-free people who had received a recombinant malaria subunit vaccine (AMA-1) that targets the parasite’s pre-erythrocytic stage. Since both adjuvant formulations contain the same immunostimulatory components, QS21 and MPL, this work sheds light on the subtle effect that adjuvant formulation has on immunogenicity. Adjuvant-mediated immune markers were discovered in two steps: First, creating a comprehensive immunoprofile. Second, using machine learning to combine the immunoprofiling data and determine which combination of immunological characteristics was most clearly capable of distinguishing vaccine-induced responses by adjuvant.
The computational study indicated statistically significant variations in cellular and antibody responses between cohorts and found a combination of immunological characteristics capable of distinguishing individuals by adjuvant with 71% accuracy. Furthermore, the in-depth analysis revealed an unanticipated activation of CD8+ T cells by the recombinant subunit vaccine, which is unusual and extremely significant for future vaccine design.