Critically for clinical value, this vaccine design has also been demonstrated to induce durable epitope-specific CTL responses against tolerized antigens 27–29, and it is now in several clinical trials. The availability of third generation MHC class I-transgenic mice expressing the human HLA-A2 molecule (HHD mice) provides a powerful tool for the investigation

of both induction and performance of CD8+ T cells recognizing human HLA-A*0201-binding epitopes 30, 31. PLX-4720 ic50 In the present study, we investigated the ability of three PSMA-derived HLA-A*0201-binding epitopes, delivered as p.DOM-epitope vaccines, to prime CD8+ T cells in the HHD transgenic mice. We show that, in sharp contrast to full-length PSMA-encoding vaccines, all three p.DOM-PSMA epitope vaccines generated CD8+

T-cell responses. However, the key point is that the target peptides must be naturally presented by PSMA-expressing tumor cells. This has not been clear in the past since most strategies have used human CD8+ T cells expanded in vitro with candidate Cell Cycle inhibitor peptides. By this approach, PSMA27-specific T cells showed weak but definite killing of PSMA-expressing LNCap prostate tumor cells 32. The same study reported that PSMA663 and PSMA711-specific CTLs appeared unable to kill the target cells, suggesting that these peptides were not efficiently processed and presented. However, processing of PSMA663 and possibly PSMA711 was observed subsequently 33. The divergent evidence 3-mercaptopyruvate sulfurtransferase on the processing

status highlights the difficulties in using human CD8+ T cells expanded in vitro, making decisions about potential peptide targets for vaccination difficult. Testing in the “humanized” model now reveals that T cells specific for PSMA27 and PSMA663, but not PSMA711, could specifically kill PSMA-expressing tumor cells in vitro and in vivo, thereby providing evidence for efficient processing and presentation of these two epitopes. Data on p.DOM-PSMA27 provide validation of the clinical trial in patients with PCa, where induction of CD8+ T-cell responses in the majority of vaccinees is evident 34. Three DNA fusion vaccines encoding PSMA-derived peptide epitopes were constructed according to the previously described vaccine design 26. Each vaccine encoded the first domain of FrC from TT, DOM, genetically fused to a discrete human PSMA HLA-A*0201-binding epitope, to create the p.DOM-PSMA27, p.DOM-PSMA663, and p.DOM-PSMA711 vaccines. The DOM sequence encodes the p30 promiscuous helper T-cell epitope that provides linked T-cell help for the vaccine response. DNA vaccines encoding the full-length human PSMA protein which contains all three epitopes were also constructed for comparison, either alone (p.PSMA) or fused to DOM (p.PSMA-DOM) (Fig. 1A).