NY-ESO-1 Vaccination in Combination with Decitabine Induces Antigen-Specific T-lymphocyte Responses in Patients with Myelodysplastic Syndrome

Abstract

Purpose: Treatment options are limited for patients with high-risk myelodysplastic syndrome (MDS). The azanucleosides, azacitidine and decitabine, are first-line therapy for MDS that induce promoter demethylation and gene expression of the highly immunogenic tumor antigen NY-ESO-1. We demonstrated that patients with acute myeloid leukemia (AML) receiving decitabine exhibit induction of NY-ESO-1 expression in circulating blasts. We hypothesized that vaccinating against NY-ESO-1 in patients with MDS receiving decitabine would capitalize upon induced NY-ESO-1 expression in malignant myeloid cells to provoke an NY-ESO-1-specific MDS-directed cytotoxic T-cell immune response.Experimental Design: In a phase I study, 9 patients with MDS received an HLA-unrestricted NY-ESO-1 vaccine (CDX-1401 + poly-ICLC) in a nonoverlapping schedule every four weeks with standard-dose decitabine.Results: Analysis of samples serially obtained from the 7 patients who reached the end of the study demonstrated induction of NY-ESO-1 expression in 7 of 7 patients and NY-ESO-1-specific CD4+ and CD8+ T-lymphocyte responses in 6 of 7 and 4 of 7 of the vaccinated patients, respectively. Myeloid cells expressing NY-ESO-1, isolated from a patient at different time points during decitabine therapy, were capable of activating a cytotoxic response from autologous NY-ESO-1-specific T lymphocytes. Vaccine responses were associated with a detectable population of CD141Hi conventional dendritic cells, which are critical for the uptake of NY-ESO-1 vaccine and have a recognized role in antitumor immune responses.Conclusions: These data indicate that vaccination against induced NY-ESO-1 expression can produce an antigen-specific immune response in a relatively nonimmunogenic myeloid cancer and highlight the potential for induced antigen-directed immunotherapy in a group of patients with limited options. Clin Cancer Res; 24(5); 1019-29. ©2017 AACRSee related commentary by Fuchs, p. 991.

Conflict of interest statement

Disclosure of Potential Conflicts of Interest

This study was supported by Celldex Therapeutics by the provision of CDX-1401/poly-ICLC. EAG has received honoraria and research support from Astex Pharmaceuticals and honoraria from Alexion Pharmaceuticals, Celgene Inc. and Pfizer Inc. CSH has received research support from Merck & Co., Inc. and SELLAS Life Sciences, Ltd. ESW has received honoraria from Incyte Pharmaceuticals, Astex Pharmaceuticals and Pfizer Inc. The remaining authors have no relevant financial conflicts to report.

©2017 American Association for Cancer Research.

Figures

Figure 1. The combination of NY-ESO-1 vaccine and decitabine promotes NY-ESO-1hypomethylation and expression

A. Schematic diagram of the clinical trial. CD11b+ cells and plasma samples were isolated from serial peripheral blood samples of patients during treatment. DLT = dose limiting toxicity window. B. Average percentage of methylated LINE-1 DNA in CD11b+ cells (solid line, circles) and plasma (dotted line, squares) harvested pre-treatment and at serial time points during treatment (n = 7). C. Average percentage of methylated NY-ESO-1 promoters in CD11b+ cells (solid line, circles) and plasma (dotted line, squares) harvested pre-treatment and at serial time points during treatment (n = 7). (D) NY-ESO-1 mRNA levels in patient samples harvested pre-treatment and at serial time points during treatment (n = 7). mRNA levels were determined using absolute quantification and normalized to β2-microglobulin (β2m) mRNA levels. For all panels, data are presented as the mean value and error bars represent the standard error of the mean. C = decitabine cycle number; D = day of each cycle. Each individual cycle has a range of 1 to 28 with decitabine treatment occurring on days 1 – 5. Statistical comparison of pre-treatment methylation in CD11b+ cells versus Cycle 1 methylation (n = 9) was performed using Wilcoxon signed rank test.

Figure 2. Myeloid blood cells from an NY-ESO-1 vaccinated MDS patient activate an NY-ESO-1-specific cytotoxic response in autologous T-lymphocytes

A. Flow cytometry analysis of T-lymphocyte response in clonal HLA-B35+ NY-ESO-1 specific CD8+ T-lymphocytes. T-lymphocyte clones were co-cultured with unselected peripheral blood mononuclear cells (PBMCs) collected from Patient 9 pre-treatment and at decitabine cycle 1, day 15 (C1D15) and decitabine cycle 2, day 15 (C2D15). NY-ESO-1 specific cells were detected using an NY-ESO-1 specific tetramer. T-lymphocyte responses were measured by intracellular cytokine staining for IFN-γ (y-axis for all plots) and cell-surface expression of CD107 (x-axis). B. Flow cytometry analysis of T-lymphocyte response in HLA-B35+ NY-ESO-1-tetramer positive CD8+ lymphocytes co-cultured with autologous CD11b+ myeloid cells. NY-ESO-1-tetramer positive T lymphocytes were enriched from samples collected from Patient 9 at the EOS. Autologous CD11b+ cells were collected from Patient 9 at pre-treatment, at C1D15, at CD215, and at the end of study (EOS). For all panels, gates were drawn based on un-stimulated T-lymphocytes and PMA/ionomycin stimulation acted as a positive control. Percentages of IFN-γ+/CD107+ cells are depicted. C. Average percentage of IFN-γ+/CD107+ NY-ESO-1-tetramer positive (Tet+; white bar) and tetramer negative (Tet; grey bar) CD8+ T-lymphocytes positive following co-culture with autologous CD11b+ blood cells at each time point. Statistical comparison of response using pre-treatment samples with other time-point was performed using Wilcoxon signed rank test (n = 7 replicates over two independent experiments; * = p < 0.05). Error bars represent standard error of the mean.

Figure 3. Frequency of CD141Hi and CD1c+ cDCs in vaccinated MDS/AML patients

Flow cytometry analysis was performed on peripheral blood samples isolated from the MDS/AML patients (enrolled on study) pre-treatment and from healthy age matched donors. A. Average frequency of CD141Hi, CLEC9A+ cDCs within the CD45+ population. N = 8 for healthy donors (circles) and MDS patients (squares). Data are presented as values for individual patients. The horizontal bar represents the mean value and error bars represent standard error of the mean. P values were determined using the Mann Whitney U test. B. Frequencies of CD141Hi, CLEC9A+ cDCs in CD45+ peripheral blood cells in individual MDS patients on study pre-treatment. C. Average frequency of CD1c+ cDCs within the CD45+ population. D. Average median fluorescent intensity (MFI) of anti-DEC-205 staining of CD141Hi (left) and CD1c+ (right) cDCs in healthy controls and MDS/AML patients on study. For all samples, anti-DEC-205 MFI was normalized to isotype control and Log2 transformed. Data are presented as values for individual patients. The horizontal bar represents the mean value and error bars represent standard error of the mean.

Figure 4. Frequency of CD141Hi and CD1c+ cDCs in vaccinated MDS/AML patients receiving decitabine

Flow cytometry analysis was performed on matched BM samples isolated from MDS/AML patients pre-treatment and at the EOS. A. Flow cytometry analysis of pre-treatment and EOS BM samples from Patients 5, 7, 8, and 9 (left to right respectively). Gates were drawn based on healthy BM controls. Percentages depict frequencies of CD141Hi and CD1c+ cDCs within the parental live/CD45+/Lin/HLA-DR+/CD11c+ population. B. Frequency of CD141Hi cDCs within the CD45+ population for pre-treatment (white bar) and EOS (gray bar) samples isolated from each patient. For comparison of the CD141Hi cDC populations in Pre versus EOS samples, double positive CD141Hi/CD1c+ cDCs were included. C. Frequency of CD1c+ cDCs within the CD45+ population for pre-treatment (white bar) and EOS (gray bar) samples isolated from each patient.