A cancer vaccine induces expansion of NY-ESO-1-specific regulatory T cells in patients with advanced melanoma

Abstract

Cancer vaccines are designed to expand tumor antigen-specific T cells with effector function. However, they may also inadvertently expand regulatory T cells (Treg), which could seriously hamper clinical efficacy. To address this possibility, we developed a novel assay to detect antigen-specific Treg based on down-regulation of surface CD3 following TCR engagement, and used this approach to screen for Treg specific to the NY-ESO-1 tumor antigen in melanoma patients treated with the NY-ESO-1/ISCOMATRIX™ cancer vaccine. All patients tested had Treg (CD25(bright) FoxP3(+) CD127(neg)) specific for at least one NY-ESO-1 epitope in the blood. Strikingly, comparison with pre-treatment samples revealed that many of these responses were induced or boosted by vaccination. The most frequently detected response was toward the HLA-DP4-restricted NY-ESO-1(157-170) epitope, which is also recognized by effector T cells. Notably, functional Treg specific for an HLA-DR-restricted epitope within the NY-ESO-1(115-132) peptide were also identified at high frequency in tumor tissue, suggesting that NY-ESO-1-specific Treg may suppress local anti-tumor immune responses. Together, our data provide compelling evidence for the ability of a cancer vaccine to expand tumor antigen-specific Treg in the setting of advanced cancer, a finding which should be given serious consideration in the design of future cancer vaccine clinical trials.

Figure 1. Antigen-specific Treg and Teff can be detected by measuring down-regulation of surface.

CD3. PBMC from Patient 113 were cultured with pooled NY-ESO-1 18mer peptides as described in Methods, followed by re-stimulation with the indicated individual peptides during overnight culture to allow CD3 down-regulation. Cells were stained with antibodies to CD4, CD25, FoxP3 and CD3 and analyzed by flow cytometry. (A): an example of the staining patterns observed, illustrating the gating of Treg and Teff (on gated viable CD4⁺ T cells) and the down-regulation of CD3 within each population. (B–C): A summary of the responses detected within the Treg (B) and Teff (C) populations. Dotted lines indicate the baseline level of CD3-low cells (nil peptide condition).

Figure 2. CD4⁺ CD25⁺ FoxP3⁺ cells identified after culture have phenotypic and functional characteristics of Treg.

Patient PBMC were cultured for 21d with NY-ESO-1 18mer peptide(s) known from preliminary studies to induce a Treg response and then: (A) analyzed for CD127 expression by flow cytometry, gating on Treg (CD4⁺ CD25⁺ FoxP3⁺) or Teff (CD4⁺ FoxP3⁻) as indicated. Results shown are representative of three experiments with similar results; or (B): Treg were purified by sorting CD4⁺ CD25⁺ CD127^low cells and tested for their ability to suppress the proliferation of CFSE-labeled CD8⁺ T cells pre-stimulated for 16hr with plate-bound anti-CD3. The graph shows % suppression relative to cultures conducted in the absence of Tregs, while flow cytometry histograms below illustrate CFSE profiles obtained for responder T cells alone (left) or at 1∶2 ratio with Treg (right). Data are representative of three independent experiments using samples from three different individuals. In (C), cultured PBMC were re-stimulated with the relevant peptide and expression of LAP on the surface of Tregs or Teffs was determined by flow cytometry within the CD3-low (peptide-responsive) population for three patients.

Figure 3. CD3 down-regulation on Tregs is dependent on the concentration of peptide used for re-stimulation.

PBMC from four patients were cultured for 21d with NY-ESO-1 18mer peptides and then re-stimulated overnight with individual peptides at the indicated concentrations. The proportion of Treg down-regulating CD3 in response to peptide was determined by flow cytometry.

Figure 4. Summary of NY-ESO-1-specific Treg responses detected in a cohort of 9 patients vaccinated with NY-ESO-1/ISCOMATRIX^TM vaccine.

(A): For every patient within the cohort, each validated Treg response is summarized with a box. Responses are considered validated if they were observed in at least two independent cultures, using two independently synthesized batches of peptide. The position of the box indicates where in the NY-ESO-1 peptide sequence the response was localized. In the event that responses were detected to two peptides adjacent in sequence, this is shown as a single response spanning the two peptides. Shaded boxes indicate that the magnitude of the response was increased at least 2-fold in post-vaccination samples compared to pre-vaccination samples when both samples were tested in parallel under identical conditions. Solid boxes indicate that the response was only detectable in samples collected after vaccination. Open boxes indicate that the magnitude of the response was similar pre- and post-vaccination. (B): An example of a response that was induced by vaccination (Patient 124) is shown. Treg were gated on the basis of CD25 and FoxP3 expression, and CD3 down-regulation was assessed following re-stimulation with either control peptide or the same peptide used for expansion (NY-ESO-1_85–102).

Figure 5. Treg and Teff respond to an identical HLA-DP4-restricted epitope in the region NY-ESO-1_157–170.

Patient PBMC were cultured for 21d with the 18mer peptide NY-ESO-1_157–174 and then re-stimulated with the indicated short HPLC-purified peptides by either adding directly to the culture as usual (A–D) or by pulsing onto BCL followed by washing (E–F). After overnight incubation, cells were stained and analyzed by flow cytometry, gating on Treg (CD4⁺ CD25⁺ FoxP3⁺; A, C and E) or Teff (CD4⁺ FoxP3⁻; B, D and F). Peptides used for re-stimulation were based on the published epitope NY-ESO-1_157–170, with either truncation (A–B) or extension (C–D) at each terminus. Graphs in A–D show results obtained for Patient 102; similar results were also obtained for Patient 103. Graphs in E–F show mean + SEM from Patients 102 and 113; an asterisk indicates a p value of <0.05 (t test).

Figure 6. Tumor tissue contains populations of NY-ESO-1-specific Treg and Teff.

Tumour tissue was obtained from Patient 126 and TIL lines generated as described in Methods. (A–B): Cells were treated overnight with peptide NY-ESO-1_115–132 or control peptide, and then stained and analyzed by flow cytometry, gating on Treg (CD4⁺ CD25⁺ FoxP3⁺) or Teff (CD4⁺ FoxP3⁻) as indicated. Representative flow cytometry dot plots (A) show the gating of Treg and Teff, and the CD3 down-regulation response observed in each population following re-stimulation, while (B) shows a summary of results obtained in five experiments, each using one of the three different TIL lines generated. (C): The effect of blocking antibodies to HLA-DR, HLA-DP or HLA-DQ on the response to peptide NY-ESO-1_115–132 within Treg (left) and Teff (right) populations. Similar results were obtained in a second experiment. (C): TIL were stimulated overnight with NY-ESO-1_115–132 peptide and then peptide specific (CD3^lo) and non-specific (CD3^hi) Treg (CD4⁺ CD127^lo CD25^hi) were purified by cell sorting and tested for their ability to suppress the proliferation of CFSE-labeled CD8⁺ T cells pre-stimulated for 4hr with anti-CD3 at the indicated Treg:responder ratios. As a comparison, Treg were also sorted from previously frozen PBMC obtained from a healthy donor. Similar results were observed in a second experiment, although higher Treg:responder ratios were required to see suppression.