Four Decades of Prophylactic EBV Vaccine Research: A Systematic Review and Historical Perspective

BACKGROUND: Epstein-Barr virus (EBV) is the causal agent of infectious mononucleosis and has been associated with various cancers and autoimmune diseases. Despite decades of research efforts to combat this major global health burden, there is no approved prophylactic vaccine against EBV. To facilitate the rational design and assessment of an effective vaccine, we systematically reviewed pre-clinical and clinical prophylactic EBV vaccine studies to determine the antigens, delivery platforms, and animal models used in these studies. METHODS: We searched Cochrane Library, ClinicalTrials.gov, Embase, PubMed, Scopus, Web of Science, WHO’s Global Index Medicus, and Google Scholar from inception to June 20, 2020, for EBV prophylactic vaccine studies focused on humoral immunity. RESULTS: The search yielded 5,614 unique studies. 36 pre-clinical and 4 clinical studies were included in the analysis after screening against the exclusion criteria. In pre-clinical studies, gp350 was the most commonly used immunogen (33 studies), vaccines were most commonly delivered as monomeric proteins (12 studies), and mice were the most used animal model to test immunogenicity (15 studies). According to an adaptation of the CAMARADES checklist, 4 pre-clinical studies were rated as very high, 5 as high, 13 as moderate quality, 11 as poor, and 3 as very poor. In clinical studies, gp350 was the sole vaccine antigen, delivered in a vaccinia platform (1 study) or as a monomeric protein (3 studies). The present study was registered in PROSPERO (CRD42020198440). CONCLUSIONS: Four major obstacles have prevented the development of an effective prophylactic EBV vaccine: undefined correlates of immune protection, lack of knowledge regarding the ideal EBV antigen(s) for vaccination, lack of an appropriate animal model to test vaccine efficacy, and lack of knowledge regarding the ideal vaccine delivery platform. Our analysis supports a multivalent antigenic approach including two or more of the five main glycoproteins involved in viral entry (gp350, gB, gH/gL, gp42) and a multimeric approach to present these antigens. We anticipate that the application of two underused challenge models, rhesus macaques susceptible to rhesus lymphocryptovirus (an EBV homolog) and common marmosets, will permit the establishment of in vivo correlates of immune protection and attainment of more generalizable data. SYSTEMATIC REVIEW REGISTRATION: https://www.crd.york.ac.uk/prospero/display_record.php?RecordID=198440, identifier PROSPERO I.D. CRD4202019844.