Development of a computational model to quantify antigenic similarity between Dengue viruses

Abstract

Abstract Dengue virus (DENV) causes dengue fever, the most prevalent arthropod-borne viral illness in humans. Globally, four DENV serotypes cause an estimated 100 million new cases of dengue fever and 250,000 cases of dengue hemorrhagic fever per year. DENV serotypes differ by 25 to 40 percent in the envelope (E) protein. Although primary DENV infection results in life-long immunity to the infecting serotype, subsequent infection with a different serotype increases risk of severe disease. It has been argued that this is due to weakly neutralizing, cross-reactive antibodies that mediate viral uptake by Fc receptor-bearing cells. The unique role of DENV antibody cross-reactivity in disease outcome necessitates characterization of viral antigenic similarity and resulting disease pathogenesis. Inspired by influenza models, we are developing a model that incorporates sequence data, neutralization titers and structural information to quantify antigenic similarity between circulating DENV strains. We first identify important E protein antigenic positions by considering sequence conservation and correlation with cross-reactivity. Amino acid similarity at antigenic positions is then combined with structural data to calculate antigenic variance. The objectives of this study are to (1) identify E protein sequence positions and structural properties (epitopes) associated with antigenic variation and (2) introduce a model to quickly predict antigenic distance between DENV strains that can be assessed by neutralization experiments. We expect many antigenic positions to be located in epitopes described in the Immune Epitope Database, although some distant positions may exert allosteric affects by altering E protein structure or dynamics.