We report the feasibility of using syngeneic cell lines, derived from the PTEN knockout model of prostate cancer, as tumor vaccines to elicit prophylactic effects against tumor establishment and progression. We further confirmed that vaccination and tumor inhibition were associated with a cellular immune response resulting in the generation of effector cells capable of exerting highly potent antiproliferative effects against these prostate cancer cells. Finally, we also noted that effective growth suppression could be achieved upon the adoptive transfer of these effector cells to preestablished tumors.
The PTEN-P8 cell line was isolated directly from true prostate adenocarcinoma arising spontaneously in this model, as generated by mating mice with biallelic floxed PTEN loci with mice expressing Cre recombinase under the control of a prostate specific promoter.16
Notably PTEN-P8 cells are deleted in only 1 PTEN allele, they are only weakly tumorigenic in vivo and they no longer express the recombinase.17
Subsequently PTEN-CaP8 cells were derived from PTEN-P8 by the introduction of a retroviral vector constitutively expressing Cre, resulting in deletions in the 2 PTEN alleles and showing much more robust tumorigenicity in vivo.17
Therefore, in these experiments we used the PTEN-CaP8 prostate cancer cell line for immunization and tumor establishment in vivo, and its parental cell line PTEN-P8 for confirmatory in vitro studies to characterize the cellular immune response. Effector cells were generated by MLTR after vaccinating nontumor bearing litter mates of PTEN deleted mice, which were of an identical genetic background except for the lack of prostate specific Cre. Hence, vaccination with PTEN-CaP8 best mimics the situation in the mice in which PTEN deleted prostate cancer developed, whose tumors initially expressed Cre. However, in vitro assays demonstrated that effector cells generated through vaccination and resensitization with PTEN-CaP8 showed robust reactivity against PTEN-P8 cells, which no longer expressed detectable levels of Cre. This indicated that the cellular immune response was not exclusively directed against Cre as a foreign antigen, but rather likely recognized endogenous tumor antigens.
Hence, it is possible that the expression of Cre as a foreign protein may have served an adjuvant function to stimulate an initial immune response against PTEN-CaP8 cells during vaccination, which then further resulted in epitope spreading to natural endogenous tumor antigens. However, notably no conventional immune adjuvants were used in the vaccination protocol. It is certainly possible that initial irradiation of the cells used in vaccination might also have somehow resulted in the exposure of epitopes that facilitated the immune response but the subsequent in vitro MLTR restimulation procedures used PTEN-CaP8 cells that had been treated with mitomycin C, rather than irradiated. Thus, a significant contribution of irradiation as an adjuvant seems less likely. In this context using radiation in situ is likely to have limited efficacy as an immuno-activating strategy due to intrinsic local mechanisms of tumor immunoresistance, including the expression of immunosuppressive cytokines, such as transforming growth factor-β, and the presence of inhibitory dendritic cells expressing anergizing co-regulators, such as B7-H1, as well as immunotolerizing T-regulatory cells in the tumor and draining lymph nodes. In fact, the ex vivo pre-activation of cytolytic T cells away from the immuno-suppressive tumor environment and their removal from these inhibitory influences may indeed represent a critical factor in the effectiveness of adoptive transfer strategies for immunotherapy.
In these experiments we used bioluminescence imaging to detect the presence and quantity of luciferase marked PTEN-CaP8/RL cells in vitro and in vivo. Bioluminescence signal intensity showed good correlation with the cell number in vitro and tumor size in vivo, and it confirmed the usefulness of this methodology for monitoring tumor growth and the response to adoptive immunotherapy in living animals with time. In these studies unmarked parental PTEN-CaP8 cells were used for vaccination in non-tumor bearing litter mates. Hence, the effector cells that mediated effective tumor growth inhibition upon adoptive transfer were not reacting to luciferase as a foreign antigen. Further studies using this methodology should be greatly facilitated by the recent generation of a new version of the prostate specific PTEN knockout model, which is also transgenic for prostate specific luciferase and, hence, spontaneously arising orthotopic prostate tumors are already marked.22
The antiproliferative effects in vitro and tumor growth suppression in vivo achieved by effector cells derived from PTEN-CaP8 vaccinated hosts were found to be highly potent, particularly considering that the responses were not likely to be directed against xenogenic antigens, such as Cre or RL. Based on the MLTR culture conditions it was expected that in the presence of target antigens on the sensitizing tumor cells the effector cell preparation would be enriched for CTLs. However, it is certainly possible that the effector cell culture also may have contained subpopulations of natural killer or lymphokine activated killer cells, which contributed to the potent antiproliferative responses in vitro and in vivo. In this context it should also be noted that the MTS assay measures only the metabolic activity of viable adherent target cells, and so it is difficult to distinguish between decreased target cell proliferation, eg in response to immunocytokine signals released from effector cells, and actual target cell death due to effector cell mediated cytotoxicity.
Therefore, future studies will focus on the further characterization of effector cells generated by vaccination with the syngeneic prostate cancer cell lines derived from the PTEN knockout model as well as the identification of relevant endogenous tumor antigens that are expressed by these cancer cells and may be specifically recognized by CTLs. Additionally, this model is also likely to be highly informative and useful for testing various augmentation strategies, such as lymphodepleting chemotherapy, which has shown promise as a preconditioning regimen to augment adoptive immunotherapy in experimental models dating back 25 years23–25
as well as in recent clinical trials.26,27
In addition to adoptive transfer strategies, the reliable and reproducible nature of spontaneous prostate cancer arising in this model is well suited for assessing vaccination strategies administered before carcinogenesis or at an early (PIN) stage. To date cancer vaccine strategies have been considerably more effective in a prophylactic setting or against minimal residual disease. In fact, they have proved rather ineffective against established bulky disease. As noted, it is increasingly recognized that solid tumors create a highly immunosuppressive environment through a wide variety of mechanisms, and so preventive vaccination may prove to be the most effective approach.
In this context it is also interesting to note a recent report showing that targeted deletion of PTEN in T cells regulated the peripheral homeostasis of Tregs in vivo and allowed their expansion in response to interleukin-2. Because prostate specific expression of Cre recombinase results in biallelic deletion of PTEN only in prostate cells in our current model, we are not currently able to directly observe the effects of PTEN deletion on T-cell function. However, in the human disease setting it is certainly conceivable that PTEN loss may represent a more generalized genetic predisposition in each cell compartment, which may conspire to simultaneously promote carcinogenesis and tolerize the immune system to the developing malignancy. Therefore, in future studies it may be valuable to evaluate the consequences of PTEN loss in prostate epithelium and T cells in a dual compartment knockout model.