Adjuvants boost the immune response to vaccine antigens, serving as key components in safe and effective vaccines. However, selecting a suitable adjuvant for a new vaccine can be challenging. This is due to the wide variety of adjuvants and the many mechanisms of vaccines they are meant to enhance. Therefore, the adjuvant selection process heavily relies on empirical experiments, which are time-consuming and resource-intensive. In this study, we introduce a machine learning approach leveraging non-human primate RNA transcriptomic data to predict immunogenic antibody levels after vaccination. Furthermore, analysis of the trained deep learning models enabled the identification of immune response mechanisms that are stimulated by adjuvants. Integration of machine learning has the potential to expedite vaccine adjuvant selection by focusing on evaluating adjuvant candidates with the highest probability of success. This may ultimately facilitate the development of more effective vaccines.

Background/Objectives: Adjuvants, added to vaccines to enhance immune responses, are central to shaping the magnitude and durability of immunity, yet their precise mechanisms remain incompletely defined. This study evaluated how diverse adjuvant combinations influence HPV vaccine immunogenicity in non-human primates, with a particular focus on impacts on hematopoietic biology—megakaryocytes and platelets—and broader innate and adaptive pathways. Methods: Eight adjuvanted formulations, each incorporating distinct immunomodulatory components and delivery platforms, were compared against an alum-only control in non-human primates. Longitudinal antibody titers (HPV16-specific) were measured up to 54 weeks, and blood transcriptomes were profiled at Day 1 and Day 7 after both prime and boost doses to assess pathway-level enrichment and gene-expression patterns. Results: Several adjuvant combinations significantly increased antibody titers at 54 weeks compared with alum alone. Formulations containing cationic lipid or monophosphoryl lipid A (MPL) were associated with enhanced antibody responses. Early upregulation of immune-related genes across innate and adaptive pathways was also observed, with some combinations (e.g., inclusion of QS21 or ISCOMs) showing similar trends. Distinct group- and time-dependent transcriptional signatures were observed, with higher-responding formulations exhibiting stronger enrichment in pathogen-influenced signaling and cellular/humoral immune programs. Conclusions: Adjuvant selection and formulation strategy substantially modulate vaccine immunogenicity and early transcriptional programs, including innate, adaptive, and hematopoietic pathways. While individual adjuvants differentially regulate immune and platelet-associated genes, common pathway-level patterns emerge across formulations. These findings suggest candidate mechanisms for prolonged vaccine efficacy and provide actionable insights to guide rational adjuvant design for sustained immune protection.