In Response: We thank von Bergwelt-Baildon et al. (1) for their comments on our study describing CD8 T-cell reactivity against cyclin B1 in breast cancer patients with tumors overexpressing p53. First of all, we want to clarify that our study only described spontaneous immune responses and not responses primed by repeated in vitro stimulation with professional antigen-presenting cells (APC). Because of the previously described relationship between mutation in p53 and overexpression of cyclin B1 (2), we looked for correlation between p53 accumulation and response to cyclin B1. Therefore, it was essential that the T-cell reactivity against cyclin B1 was not in vitro primed, but in vivo primed. Since the T-cell responses were not detectable ex vivo, we stimulated peripheral blood mononuclear cells (PBMC) once with peptide, which should only boost existing T-cell responses and not prime new ones (3, 4).
Second, our study showed the existence of T-cell responses against only two cyclin B1derived peptides [CB9 (AKYLMELTM) and CB-P4 (AGYLMELCC)]. Interestingly, these responses were predominantly found in patients harboring cancer cells with high levels of cytoplasmic p53. Statistically, T-cell reactivity against only one of these cyclin B1derived peptides (CB9) correlated with elevated p53 protein expression.
Third, the cyclin B1derived peptides used in our study were characterized by elution of peptides from cancer cells (5); thus, the presentation of the peptides on the cell surface of cancer cells is firmly established. To this end, we did not attempt to measure cyclin B1specific killing of cancer cells to verify that the epitopes were naturally processed and presented by cancer cells. Hence, for our study, in vitro generation of cyclin B1specific T-cell cultures was not essential. Nevertheless, we agree that repetitive stimulation with peptide-pulsed professional APCs is more efficient in expanding antigen-specific T cells than a single stimulation with peptide-pulsed PBMCs. However, this was not our intention.
A point mentioned by von Bergwelt-Baildon and coworkers was that cyclin B1specific T cells are readily expandable in healthy donors. In our study, the absence of detectable T-cell responses against cyclin B1derived peptides in healthy donors did not suggest that cyclin B1specific T cells were not expandable; however, to expand the cells was not the focus of our study. In general, we agree that it is important to be able to expand tumor antigenspecific T cells and that this is important for immune therapy, such as allogeneic stem cell transplantation and adoptive transfer.
As in the study by Kondo et al. (6), we have characterized the cyclin B1derived peptide 204(ILIDWLVQV). This peptide was identified by reverse immunology using computational peptide prediction algorithms. We analyzed spontaneous T-cell reactivity in blood from cancer patients and healthy donors after one in vitro stimulation with peptide-pulsed PBMCs, as in our previous study. Spontaneous T-cell responses were observed in 16 of 22 cancer patients and 10 of 11 healthy donors.1
Furthermore, we did not see a correlation between reactivity against the 204(ILIDWLVQV) peptide and p53 overexpression, even when excluding the healthy donors. Because we used blood from one third of the patients described in our previous study, the correlation between p53 overexpression and cyclin B1 reactivity is likely to be peptide dependent.
A question raised by von Bergwelt-Baildon et al. was if the difference between the peptides could be linked to the different epitope classes: peptide abundantly presented by MHC and, thus, efficiently elutable versus peptides identified by reverse immunology translating to differential thymic expression and deletion. When epitopes are identified by reverse immunology, there is a risk that T cells with the predicted specificity may have been deleted during negative selection in the thymus. Therefore, the existence of this T-cell specificity needs to be verified. Another risk is that the epitopes are not presented on the surface of cancer cells. This also needs to be verified, and here, repeated stimulation with peptide-pulsed professional APCs is crucial. Using this approach, antigen-specific T-cell cultures and clones are established and their capacity to kill cancer cells is analyzed. In our work with the 204(ILIDWLVQV) peptide, we established a CTL clone that was able to lyse breast cancer cell lines in an HLA-A2restricted and peptide-specific manner. This indicated that the peptide is presented on the surface of cancer cells. Thus, although healthy donors hosted T-cell responses against the peptide, our experiments suggested that the peptide has relevance in cancer.
Nonetheless, we do agree that thymic deletion of cognate TCRs or abundancy of the appropriate MHC-peptide complex on the cell surface could be denominators of whether the peptide elicits a spontaneous response or not.
Taken together, we agree that professional APCs provide a better stimulation of T cells than nonprofessional APCs, such as the PBMCs used in our study. However, depending on the purpose of the study, such stimulation may not be relevant. In general, the experimental setup should be carefully selected to address the questions in hand. Thus, in our previous study, we are confident that responses against the CB9 and CB-P4 peptides could have been detected in cultures from healthy donors if we had used a more potent and repeated stimulation. We would in that case not have been able to answer the question if responses coincide with overexpression of p53.
Concerning the 204(ILIDWLVQV) peptide examined both by the group of von Bergwelt-Baildon and our group, the data strongly suggests that a T-cell response against this peptide is elicited in healthy donors as well as cancer patientscontrasting the data related to the CB9 and CB-P4 peptides. The background for the difference is unknown.
Conclusively, we also believe that cyclin B1 is a potential promising target for anticancer immune therapy, both preventive and therapeutic. Interestingly, the different reactivity against cyclin B1derived peptides could present a model for further studies of the mechanisms that governs spontaneous immune reactivity against cancer.