The notion that TCRs are often of low affinity for TAA and not sufficiently activated by TAA has led to the hypothesis that antitumor immunity can be generated by activating TAA-specific T cells using mimotopes of higher affinity. Using a combinatorial peptide library we identified a panel of peptide mimotopes for the T cell epitope AH1 from the transplantable CT26 tumor (Table ). We studied 6 mimotopes that, when complexed with H-2Ld, have increasing affinities for a TCR that recognizes the H-2Ld/AH1 complex. These mimotopes all showed similar binding to H-2Ld, suggesting that the increase in affinity of the TCR-pMHC interaction is a result of enhanced binding of the MHC-mimotope to the TCR. Using these mimotopes, we found that to induce an optimal antitumor response, the affinity of the TCR-pMHC interaction must be higher than the TAA, but cannot be too high. Thus, the intermediate-affinity peptides elicited optimal antitumor responses.
The sequence of these mimotopes differs in the first 3 of 9 amino acids. The second amino acid is an anchor residue for binding the H-2Ld
). The high-affinity mimotopes and the AH1 peptide have the consensus P at this position. Typically mimotopes are designed with consensus anchor residues to ensure optimal presentation. Although the intermediate-affinity mimotopes, 39 and 27, have N at this position, N has been identified in other H-2Ld
–restricted peptides (Tum QNHRALDL; ref. 52
). Using the SPR competition assay in Figure C, we did not detect differences in binding H-2Ld
molecules whether position 2 is a P or an N, but it is interesting to note that the peptides that were most protective did not have the consensus anchor residue at position 2.
Increasing the affinity of the TCR-pMHC interaction correlated with increased proliferation and IFN-γ production from the T cell clone in vitro (Figure ). However, the increased binding did not correlate with antitumor immunity in vivo; vaccination with mimotopes whose EC50
was 3 orders of magnitude higher than the TAA resulted in no increase in antitumor immunity (Figure ). Mimotopes that promote antitumor immunity had a mid-range affinity that correlated with an EC50
2 orders of magnitude higher than the wild-type TAA. Interestingly, a previously characterized mimotope for the AH1 peptide, A5 (12
), which protects mice from tumor challenge, also falls within the range of the effective intermediate-affinity peptides (data not shown). It should be noted that none of the mimotopes chosen from the library exist in the mouse genome so the lack of antitumor activity seen by the high-affinity mimotopes cannot be attributed directly to central tolerance.
One important question that arises from these experiments is how representative results from one T cell clone are relative to a responding population of T cells. As discussed in the Introduction, the TCR used was identified multiple times, and the clone it was derived from had low functional avidity for its cognate ligand, as expected for T cells specific for self antigens (Figure ). Thus, both the precursor frequency and functional avidity likely represent the responding population. However, future experiments to identify mimotopes using nonclonal T cell populations that respond to the TAA will address whether a clone will predict optimal mimotopes for a polyclonal T cell population.
Increasing evidence suggests that the quality, not quantity, of the responding T cells determines the efficacy of the response. The mimotopes in the lower-affinity range expanded fewer AH1-specific T cells, potentially contributing to their inability to protect against tumor challenge (Figure A). Vaccination with the intermediate-affinity mimotopes results in more AH1-specific T cells relative to the highest-affinity mimotopes. TILs from mice vaccinated with both intermediate- and high-affinity mimotopes were antigen experienced and displayed CD107a in response to AH1 peptide (Figure ).
A more striking difference between T cells expanded by the intermediate- and high-affinity mimotopes was the observed functional differences. TILs from mice vaccinated with the high-affinity mimotopes produced significantly less IFN-γ. A similar loss of antigen-specific IFN-γ production by TILs was recently reported in melanoma patients (49
). IFN-γ is a potent activator of macrophages and increases the expression of MHC class I on tumor cells, making it crucial for an effective response against solid tumors (1
). The inability of the TAA-specific T cells to produce IFN-γ in response to vaccination with high-affinity mimotopes may explain why large numbers of T cells do not protect against tumor growth. It is important to note that although IFN-γ production was impaired following stimulation with the AH1 or mimotope peptide, treatment with PMA and ionomycin induced cytokine production, arguing against a global defect in these cells. High-affinity TCR-pMHC interactions have been shown to induce T cell deletion in the periphery (30
). High-affinity mimotopes may lead to deletion of the highest-avidity T cells and/or an expansion of a lower-avidity population of T cells. These low-avidity T cells have a low sensitivity for peptide and require high concentrations of cognate peptide for both cytotoxic activity and cytokine production (20
). This rationale predicts that vaccination with mimotopes of different affinities results in the expansion of T cells of varying avidity for the AH1 peptide. Therefore, the intermediate-affinity mimotopes may elicit a different repertoire of T cells relative to those expanded by the high-affinity mimotopes, resulting in improved antitumor activity. Although preliminary analysis of the responding repertoires of T cells elicited by both intermediate- and high-affinity mimotopes revealed no obvious differences in TCR Vβ chain expression (data not shown), critical differences in the CDR3 sequence may exist.
Alternatively, strong stimulation through the TCR by the high-affinity peptides may lead to the expansion of functionally anergic T cells that do not respond to further stimulation through the TCR. This response prevents autoimmunity following activation of self-reactive T cells in the periphery and occurs in mice harboring large numbers of transgenic T cells reactive against self antigens (55
). Similar to what we observed, these autoantigen-specific T cells proliferate and upregulate CD44 and CD69, but fail to develop effector functions. Functional impairment of CD8 T cells is also observed following lymphocytic choriomeningitis virus infection, where chronic viral loads lead to an exhausted population of T cells that lose the ability to produce IFN-γ (58
). In addition, administration of high-affinity peptide analogs of a myelin basic protein peptide prevent the onset of experimental autoimmune encephalomyelitis, suggesting that vaccination with mimotopes above an optimal affinity may be detrimental in inducing an optimal T cell response (54
). Further analysis of these functionally unresponsive cells is necessary to understand the mechanisms of functional impairment.
There is an affinity threshold for peptide vaccines below which T cells will not proliferate, and this threshold can be overcome by vaccination with peptides that increase TCR-pMHC affinity (6
). We propose here that there is also an upper boundary for the affinity of peptide vaccines above which the expansion of functionally unresponsive T cells occurs. MHC-mimotope complexes with intermediate affinity for the TCR induced the expansion of TAA-specific T cells that produce IFN-γ, demonstrating that this state of unresponsiveness was not evoked with lower-affinity antigen stimulation. These data have important implications for the design of peptide vaccines, since much of the previous work with mimotopes has used functional stimulation of a T cell clone as a readout for vaccine effectiveness.
In summary, we show here that while peptide stimulation of the T cell clone in culture correlates well with physical binding parameters involving the TCR-pMHC complex, it does not correlate well with immunogenicity of the same peptide against the tumor. The peptides of highest affinity may not be appropriate in the design of effective antitumor vaccines; as immunogens, these high-affinity mimotopes may functionally inactivate the relevant responding T cells by expanding cells with defective cytokine profiles.