Abstract
Dengue, a significant public health problem in several countries around the world, is caused by four different serotypes of mosquito-borne dengue viruses (DENV-1, -2, -3, and -4). Antibodies to any one DENV serotype which can protect against homotypic re-infection, do not offer heterotypic cross-protection. In fact, cross-reactive antibodies may augment heterotypic DENV infection through antibody-dependent enhancement (ADE). A recently launched live attenuated vaccine (LAV) for dengue, which consists of a mixture of four chimeric yellow-fever/dengue vaccine viruses, may be linked to the induction of disease-enhancing antibodies. This is likely related to viral interference among the replicating viral strains, resulting in an unbalanced immune response, as well as to the fact that the LAV encodes prM, a DENV protein documented to elicit ADE-mediating antibodies. This makes it imperative to explore the feasibility of alternate ADE risk-free vaccine candidates. Our quest for a non-replicating vaccine centered on the DENV envelope (E) protein which mediates virus entry into the host cell and serves as an important target of the immune response. Serotype-specific neutralizing epitopes and the host receptor recognition function map to E domain III (EDIII). Recently, we found that Pichia pastoris-expressed DENV E protein, of all four serotypes, self-assembled into virus-like particles (VLPs) in the absence of prM. Significantly, these VLPs displayed EDIII and elicited EDIII-focused DENV-neutralizing antibodies in mice. We now report the creation and characterization of a novel non-replicating recombinant particulate vaccine candidate, produced by co-expressing the E proteins of DENV-1 and DENV-2 in P. pastoris. The two E proteins co-assembled into bivalent mosaic VLPs (mVLPs) designated as mE1E2bv VLPs. The mVLP, which preserved the serotype-specific antigenic integrity of its two component proteins, elicited predominantly EDIII-focused homotypic virus-neutralizing antibodies in BALB/c mice, demonstrating its efficacy. In an in vivo ADE model, mE1E2bv VLP-induced antibodies lacked discernible ADE potential, compared to the cross-reactive monoclonal antibody 4G2, as evidenced by significant reduction in the levels of IL-6 and TNF-α, suggesting inherent safety. The results obtained with these bivalent mVLPs suggest the feasibility of incorporating the E proteins of DENV-3 and DENV-4 to create a tetravalent mVLP vaccine.