Optimizing a multi cancer antigen plasmid-based vaccine using an in situ prediction model.

Abstract

Abstract Little is known about optimization of a class II restricted plasmid-based vaccine targeting multiple cancer antigens. We set out to define parameters of an in situ prediction model in vaccine design for optimal immunogenicity. Using previously defined immunogenic class II epitopes from five cancer antigens, we evaluated the effects of using linkers between antigens and varying the antigen order. The construct with linkers was superior to the construct without linkers. The mice receiving the vaccine with linkers developed a greater magnitude (p<0.0001), incidence and breadth IFN-g immune response. To determine the optimal antigen order, the sequences of all possible permutations were generated. Each sequence was evaluated in four publicly available algorithms to assess potential generation of immunity and physiochemical properties of the fusion protein. Of all parameters assessed, only predicted class II immunogenicity, in vivo half-life and percent of epitopes in helical secondary structure differed considerably among the constructs. We evaluated three constructs with similarly high class II immunogenicity; one with (A) long in vivo half-life and high percent in helices, one with (B) short half-life and high percent in helices or one with (C) short half-life and low percent in helices. Construct A generated a higher magnitude IFN-g response than Construct B or C (p<0.006 for all). While there was 100% incidence of an immune response to all constructs, the breadth of the response for Construct A was greater than the other two constructs. Class II immunogenicity, in vivo half-life and the percent of the protein in a helical secondary structure can be used to discern the most immunogenic multi-antigen plasmid vaccine construct.