Abstract 182P
Background
Vaccines represent attractive components of the tumor immunotherapy armamentarium. Although whole tumor-based vaccines and private neoantigens-based vaccines; each have pros and cons, their relative efficacy has never been comprehensively investigated. Furthermore, neoantigens prioritization to enrich for tumor-rejection antigens remains challenging.
Methods
In this study, we used a murine lung cancer model (LLC1) to investigate the effect of immunization with DCs pulsed with oxidized whole tumor lysate (WTL_DC vax) or neoantigen peptides (Neo_DC vax) on tumor growth. Distinct algorithms were used to predict and prioritize LLC1 candidate neoepitopes. Finally, we retrospectively meta-analyzed neoepitopes selection in a cohort of vaccine-treated ovarian and lung patients to improve correlates of immunogenicity and efficacy.
Results
WTL-DC vax was more effective in suppressing LLC1 tumor growth than Neo_DC vax (when neoepitopes were selected with any of the commonly-used in silico predictors based on peptide-MHC affinity). In contrast, vaccine efficacy significantly improved when immunogens were selected on the basis of the effective in vitro binding of candidate neoepitopes to cognate MHC alleles that was experimentally determined. Vaccine efficacy also significantly improved when immunogens had pre-validated or predicted (using PRIME1.0) immunogenicity. Finally, as opposed to their in silico scores, the effective in vitro binding efficacy of candidate immunogens better correlated with vaccine efficacy (clinical responses) in cohorts of lung cancer patients and with immunogenicity in cohorts of ovarian cancer patients.
Conclusions
Vaccine based on private neoepitopes predicted using common in silico predictors of peptide-MHC affinity are outperformed by WTL-based vaccines. Parameters such as in vitro peptide-MHC binding and peptide immunogenicity significantly improve the prioritization of tumor-rejecting neoepitopes and should be taken into consideration to increase vaccines’ clinical efficacy.
Legal entity responsible for the study
CHUV.
Funding
LICR CHUV UNIL.
Disclosure
All authors have declared no conflicts of interest.