This work lays the foundation for developing a single Ad vector encoding …

• Abstract
• Background
• Results
• Discussion
• Conclusion
• Methods
• Miscellaneous
• References
• Figures

DEN viruses, of which there exist four distinct serotypes (DEN-1, -2, -3 and -4), belonging to the family Flaviviridae, are transmitted to humans by the mosquito Aedes aegypti. Infection with DEN viruses poses a significant public health threat to ~40% of the global population. DEN virus infections, of which ~100 million cases occur annually, produce symptoms ranging from mild fever to severe hemorrhagic, potentially fatal, fever [1-3]. The pathogenesis of severe disease is not very well understood. Though several factors such as virulence of the infecting virus, age and genetic predisposition of the patient are implicated [4,5], the most important factor is held to be sequential infection by different serotypes in areas of endemicity [1-3]. Antibodies elicited by a given DEN virus serotype, which confer lifelong homologous immunity to that serotype, can cross-react with, but not protect against the remaining serotypes. These in fact have the potential to exacerbate disease severity through antibody dependent enhancement (ADE), during secondary infection with a different serotype [6,7]. Additionally, pre-existing memory T cell populations with lower avidity for the secondary serotype may be responsible for sub-optimal viral clearance and enhanced pro-inflammatory responses [8]. Therefore, an ideal dengue vaccine should be 'tetravalent', capable of affording protection against all four DEN virus serotypes [9,10].

In an effort to develop tetravalent dengue vaccines, several investigators are currently exploring plasmid DNA-based experimental vaccines in mice [11-13] and monkeys [14]. Simultaneously, some groups are also investigating recombinant protein-based monovalent vaccine candidates [15,16]. Of the numerous DEN vaccine approaches being explored, live attenuated [17-23] and yellow fever vaccine vector YF17D-based chimeric [24-27] vaccines are in advanced stages of development [10,28]. Vector backbones based on attenuated DEN viruses are also being used to develop intertypic chimeric vaccines [29,30]. Uniformly, all these vaccine candidates are 'monovalent' in that they are specific to a single DEN virus serotype. All these strategies rely on physically mixing the four monovalent vaccines to generate empirical 'tetravalent' formulations [reviewed in [31,32]]. It is becoming increasingly apparent that these formulations have the potential to elicit a skewed immune response, predominantly to a single serotype [18,23,25]. This has been attributed to 'viral interference', a poorly understood phenomenon, associated with increased replication of one of the monovalent vaccine viruses in the tetravalent formulation [reviewed in [31,32]]. This has necessitated the testing of multiple, empirical tetravalent formulations [20,22,27,30]. Recently, Zhou and Deem have suggested that injecting the monovalent vaccines at separate sites may alleviate this problem [33].

The objective of this work is to seek proof-of-concept for the hypothesis that a single vaccine vector targeting multiple DEN virus serotypes has the potential to bypass the risk of viral interference inherent in the currently prevalent approach of mixing 'monovalent' vaccines. For this purpose, we have developed a rAd vector encoding a novel 'bivalent' antigen, based on EDIII of DEN-2 and DEN-4 viruses. The choice of EDIII was based on several reasons. First, EDIII has emerged as the most important domain from a vaccine perspective. It has been shown that the host cell surface receptor-binding motif of the E protein maps to EDIII [34]; further serotype-specific multiple neutralizing epitopes have been localized to this domain [35-37]. Importantly, EDIII has only a very low intrinsic potential for eliciting cross-reactive antibodies against heterologous DEN virus serotypes [38,39], implicated in severe pathogenesis. Second, DEN-2 virus EDIII protein (EDIII-2) can block DEN-2 virus infectivity [40,41]. Third, several groups [13,15,16,38,39] including ours [42,43] have shown that EDIII-encoding gene-based vaccine candidates and recombinant EDIII fusion proteins elicit neutralizing antibodies.

In this paper, we describe the construction and characterization of a bivalent rAd vector encoding a chimeric antigen created by linking EDIIIs of DEN virus serotypes 2 and 4 (EDIII-4/2 antigen), and show that it elicits comparable levels of virus-neutralizing antibodies and T cell responses specific to these two serotypes.