Therapeutic vaccination trials, in which patients with cancer were vaccinated with minimal CTL peptide in oil-in-water formulations, have met with limited success. Many of these studies were based on the promising data of mice studies, showing that vaccination with a short synthetic peptide in IFA results in protective CD8(+) T cell immunity. By use of the highly immunogenic OVA CTL peptide in IFA as a model peptide-based vaccine, we investigated why minimal CTL peptide vaccines in IFA performed so inadequately to allow full optimization of peptide vaccination. Injection of the minimal MHC class I-binding OVA(257-264) peptide in IFA transiently activated CD8(+) effector T cells, which eventually failed to undergo secondary expansion or to kill target cells, as a result of a sustained and systemic presentation of the CTL peptides gradually leaking out of the IFA depot without systemic danger signals. Complementation of this vaccine with the MHC class II-binding Th peptide (OVA(323-339)) restored both secondary expansion and in vivo effector functions of CD8(+) T cells. Simply extending the CTL peptide to a length of 30 aa also preserved these CD8(+) T cell functions, independent of T cell help, because the longer CTL peptide was predominantly presented in the locally inflamed draining lymph node. Importantly, these functional differences were reproduced in two additional model Ag systems. Our data clearly show why priming of CTL with minimal peptide epitopes in IFA is suboptimal, and demonstrate that the use of longer versions of these CTL peptide epitopes ensures the induction of sustained effector CD8(+) T cell reactivity in vivo.