J. Clin. Invest. https://doi.org/10.1172/JCI91512 (2018)

The immune response to pathogens such as influenza virus involves the processing of pathogen proteins into peptides by antigen-presenting cells and the subsequent presentation of these antigenic peptides by MHC molecules to T cells. Administration of peptide-based vaccines, which are also presented via the MHC, can be difficult because of the instability of peptides. Miles et al. sought to generate an environmentally stable influenza vaccine by modifying an antigenic 9-mer peptide (GILGFVFTL) to contain all d-amino acids, but it was found to be nonimmunogenic. However, screening of a combinatorial peptide library for peptides that could mimic GILGFVFTL in eliciting a T-cell response identified an all-d peptide (gppqwnnpp) bearing no sequence resemblance to GILGFVFTL. This d-peptide had reduced potency due to decreased binding to MHC but was much more stable in human serum and under conditions that mimic gastric acid. This d-peptide had comparable MHC preference, could similarly amplify GILGFVFTL-specific human memory T cells in vitro, and could mobilize antigen-specific T-cell repertoires that closely mimic those elicited by GILGFVFTL. Finally, gppqwnnpp was immunogenic when injected into mice or when dosed orally and could protect the animals from an influenza challenge. These results highlight the benefits of d-amino acid engineering and provide a promising avenue toward more stable vaccines.