HIV vaccine development by computer assisted design: the GAIA vaccine

The design of epitope-driven vaccines that address the global variability of HIV has been significantly hampered by concerns about conservation of the vaccine epitopes across clades of HIV. We developed two computer-driven methods for improving epitope-driven HIV vaccines: the Epi-Assembler, which derives representative or "immunogenic consensus sequence" (ICS) epitopes from multiple viral variants, and VaccineCAD, which reduces junctional immunogenicity when epitopes are aligned in a string-of-beads format for insertion in a DNA expression vector. In this study, we report on 20 ICS HIV-1 peptides. The core 9-mer contained in these consensus peptides was conserved in 105-2250 individual HIV-1 strains. Nineteen of the 20 ICS epitopes (95%) evaluated in this study were confirmed in ELISpot assays using peripheral blood monocytes obtained from 13 healthy HIV-1 infected subjects. Twenty-five ICS peptides (all 20 of the peptides evaluated in this study and 5 additional ICS epitopes) were then aligned in a pseudoprotein string using "VaccineCAD", an epitope alignment tool that eliminates immunogenicity created by the junctions between the epitopes. Reordering the construct reduced the immunogenicity of the junctions between epitopes as measured by EpiMatrix, an epitope mapping algorithm. The reordered construct was also a more effective immunogen in vivo when tested in HLA-DR transgenic mice. These data confirm the utility of bioinformatics tools to design novel vaccines containing "immunogenic consensus sequence" T cell epitopes for a globally relevant vaccine against HIV. (c) 2005 Elsevier Ltd. All rights reserved.