In our efforts to improve treatment of patients with HNSCC we embarked on the induction of T cell responses against widely expressed tumor antigens. One such tumor antigen is survivin which is expressed in almost all tumor types, including HNSCC, and is essential for tumor cell survival. Our laboratories have a substantial track record with respect to the induction, isolation and maintenance of human T cells. In the past we have isolated T cells against a variety of tumor antigens, amongst which Human Papilloma Virus, Mart-1, human telomerase reverse transcriptase (hTERT), ErbB3-binding protein-1 (Ebp1), carcinoembryonic antigen (CEA) and Her-2/neu [41, 43–52].
Here we have shown the presence of survivin specific T cells in PBMC of an HNSCC patient by means of tetramer staining (Figure 1A; binding of survivin(5–14): TLPPAWQPF) and in the draining lymph node of a patient with advanced breast cancer by means of ELIspot (Figure 1B; recognition of native survivin95-104: ELTLGEFLKL). It was noted that survivin specific T cells derived from these patients could not be maintained in vitro for prolonged periods of times, unlike almost all other T cell clones we obtained in the past. One could argue that such survivin specific T cells may have become senescent due to continuous in vivo stimulation by tumor cells present in these patient, resulting in shortening of the telomeric ends of the chromosomes and subsequent clonal exhaustion. Due to low T cell numbers we could not determine the length of the telomeric ends of the chromosomes in these T cells.
This argument of telomere erosion, however, surely does not hold for survivin specific T cells induced from PBMC taken from healthy donor blood (results shown in Figures 4, 5 and 6). Prior to in vitro stimulation, these T cells were antigen inexperienced, naive, and therefore will have had long telomeres. The results shown in Figure 6 suggest that survivin specific T cells can only be maintained when the percentage remains below a certain level, which may vary per donor or activation status of the T cells. Another example of this is shown in Figure 5, where we found survivin specific T cells up to 10% of the cells in the life gate of the FACS. However, when these cells were enriched and activated via feeder mix, the percentage of survivin specific T cells dropped dramatically.
A number of research groups have documented the existence of survivin specific T cells, in mice as well as in man. Sorensen et al. published on the generation of an HLA-A2 restricted, survivin(96–104) specific T cell clone . This T cell clone, derived from a breast cancer patient, was able to kill peptide loaded T2-target cells in an HLA-A2 restricted fashion. Moreover the T cell clone was able to recognize and kill a series of HLA-A2 positive tumors cells in classical cytotoxicity assays. In contrast to this, Leisegang et al. reported on the lack of success in generating survivin specific T cell lines or clones employing RNA loaded dendritic cells from HLA-A2 positive donors.
Under normal physiologic conditions survivin is a short lived protein . Endogenous expression of survivin, detected by immunohistochemistry and/or western blot analysis, has been documented in human T cells derived from patients with Multiple Sclerosis or Crohn’s disease [55, 56]. Leisegang reported on high level expression of survivin mRNA in activated T cells. Here we have documented the presence of survivin protein in activated CD8+ T cells, and absence thereof in resting T cells (results shown in Figure 7). Since survivin is a protein which is rapidly ubiquitinylated and degraded , it would be reasonable to assume that survivin derived protein fragments are also processed and presented in MHC class I, thus making activated T cells an unintended target for survivin specific T cells.
In a commentary, Aqui and Vonderheide pointed out that endogenous expression of potential tumor antigens, like survivin and telomerase, in activated T cells might lead to fratricide killing of T cells . No signs of fratricide however were reported by Sorensen et al., nor did they report on survivin mRNA or protein expression in T cells. Possible explanations could be that their T cell clone carried a mutation in the epitope of survivin(96–104) thereby losing the ability to be presented, or it may have displayed overall malfunctions in HLA-A2 antigen expression, thereby escaping fratricide killing.
Recently, Leisegang et al. showed that in vitro cultured, T cell receptor transgenic, survivin(96–104) specific T cells underwent fratricide in an HLA-A2 dependent manner . They used dendritic cells and T cells from HLA-A2 negative donors. DC were electroporated with mRNA encoding full length survivin in combination with HLA-A2 for the induction of HLA-A2 restricted survivin specific T cells. Resulting, HLA-A2 negative, T cells were capable of killing HLA-A2, survivin double positive tumour cells. Introduction of HLA-A2 restricted, survivin specific T cell receptors into polyclonal HLA-A2 positive T cells led to fratricide killing of neighboring T cells .
In our experiments we noted that survivin specific T cells could not be maintained in culture for prolonged periods of time. Although we were unable to check for fratricide mediated killing of neighboring T cells, due to a lack of sufficient T cell numbers, our findings can very well be explained by the observations made by Leisegang et al. for the immunodominant T cell epitope survivin(96–104). Although speculative, we documented lack of T cell outgrowth for the survivin(5–14) epitope (TLPPAWQPF) shown in Figure 1 as well.
Fratricide amongst survivin specific CD8+ T cells has now been documented . It remains unclear whether fratricide will also occur in CD4+ T cells on the basis of survivin epitopes presented in MHC-class II [59, 60]. A number of different splice variants have been described for survivin [16–19]. Whether all of these are expressed in activated T cells remains to be elucidated. One could envision that T cell reactivity directed against survivin splice variants, expressed in tumor cells but not in T cells, may still be useful in cancer immunotherapy approaches [61, 62]. This notion may be exemplified by the HLA-A24 restricted survivin 2B epitope 80–88 generated by Takahasi and colleagues . The isolation of HNSCC tumor infiltrating T cells and the identification of epitopes being recognized may offer new therapeutic targets. Alternatively, clinical use of ex vivo expanded tumor infiltrating T cells with a range of (unknown) specificities may also provide a viable treatment option for HNSCC patients . Survivin as an immunotherapeutic target may still be useful in DC based active immunotherapy, since in vivo the percentages of survivin specific T cells will most likely remain below the ‘fratricide threshold’. However for adoptive transfer purposes the use of survivin specific T cells is self-destructive and therefore not feasible.