Novel and shared neoantigen derived from histone 3 variant H3.3K27M mutation for glioma T cell therapy

Abstract

The median overall survival for children with diffuse intrinsic pontine glioma (DIPG) is less than one year. The majority of diffuse midline gliomas, including more than 70% of DIPGs, harbor an amino acid substitution from lysine (K) to methionine (M) at position 27 of histone 3 variant 3 (H3.3). From a CD8⁺ T cell clone established by stimulation of HLA-A2⁺ CD8⁺ T cells with synthetic peptide encompassing the H3.3K27M mutation, complementary DNA for T cell receptor (TCR) α- and β-chains were cloned into a retroviral vector. TCR-transduced HLA-A2⁺ T cells efficiently killed HLA-A2⁺H3.3K27M⁺ glioma cells in an antigen- and HLA-specific manner. Adoptive transfer of TCR-transduced T cells significantly suppressed the progression of glioma xenografts in mice. Alanine-scanning assays suggested the absence of known human proteins sharing the key amino acid residues required for recognition by the TCR, suggesting that the TCR could be safely used in patients. These data provide us with a strong basis for developing T cell-based therapy targeting this shared neoepitope.

Figure 1.

H3.3K27M peptide induces T cell responses in DIPG patients and led to isolation of H3.3K27M-specific CTL clones from HLA-A2⁺ donor PBMCs. (A) Patient-derived and healthy donor-derived PBMCs were stimulated with H3.3WT, H3.3K27M, and flu peptide or without peptide and co-cultured with T2 cells pulsed with H3.3WT, H3.3K27M, and flu peptide or without peptide. Numbers of IFN-γ spots per 5 × 10⁴ T cells generated in each stimulation condition. Counts shown are normalized to no-peptide stimulation control. Bars and error bars represent median and SD, respectively, for IFN-γ spots per 5 × 10⁴ T cells observed in each stimulation condition. n = 3 in each group. Experiment was conducted once. *, P < 0.05; **, P < 0.01 by Student’s t test comparing H3.3WT and H3.3K27M stimulation groups. ns, not significant. (B) HLA-A*02:01⁺ healthy donor-derived PBMCs were stimulated in vitro with H3.3K27M peptide and evaluated for reactivity against HLA-A*02:01-H3.3K27M-specific tetramer and anti-CD8 mAb using flow cytometry (gated on live lymphocytes). Tetramer⁺ gate was based on control T cells. Tetramer^high population (2.42%) represented the high-affinity CTL population. (C) CTL clones were generated by flow-sorting followed by limiting dilution cloning of tetramer^high CD8⁺ T cells. The clone 1H5 was chosen for further investigation based on its excellent tetramer reactivity. The dot plot demonstrates that the clone 1H5 retains the reactivity levels to the HLA-A*0201/H3.3K27M-specific tetramer. (D) Bar graph representing IFN-γ production from the T cell clone 1H5 measured by ELISA after co-culture with T2 cells pulsed with H3.3K27M peptide at titrating concentrations or with H3.3WT peptide (5 µg/ml). Data represent two independent experiments with similar results. Bars and error bars represent median and SD, respectively, for IFN-γ production levels. **, P < 0.01 using Student’s t tests comparing 5 µg/ml H3.3WT and H3.3K27M peptide concentrations.

Figure 2.

The H3.3K27M peptide is detectable by LC-MS/MS in the HLA class I immunopeptidome of glioma cells bearing the H3.3K27M mutation. HLA class I peptides were biochemically purified from U87H3.3K27M glioma cells and analyzed by LC-MS/MS with a synthetic heavy version of the H3.3K27M peptide as the reference. (A) U87H3.3K27M HLA class I immunopeptidome shows two coeluting isotope patterns corresponding to the target m/z and mass difference of the oxidized forms of the heavy and the endogenous H3.3K27M peptides. (B) Fragmentation spectrum of the heavy peak, showing identification of the oxidized heavy H3.3K27M peptide. (C) Zoom-in of the light isotope pattern shows m/z values and distances between peaks as expected from the endogenous H3.3K27M peptide. This experiment was conducted once.

Figure 3.

Cloning of cDNA for the H3.3K27M-specific TCR and construction of a retroviral vector for efficient transduction of human T cells. (A) Schema of the TCR retroviral vector design. Synthesized TCR cDNA fragments derived from the CD8⁺ T cell clone 1H5 were inserted into the NotI/XhoI site of Takara siTCR vector plasmid together with the Kozak sequence, spacer sequence (SP), and P2A sequence. (B) T2 cells loaded with or without H3.3K27M, H3.3WT, or irrelevant influenza matrix M1_58–66 flu peptide (10 µg/ml) were co-cultured with either control or TCR-transduced Jurkat76CD8⁺ or Jurkat76CD8⁻ cells in a 1:1 ratio and assessed for IL-2 production by ELISA. Data represent three independent experiments with similar results. Bars and error bars represent median and SD, respectively, for IL-2 levels (n = 3 in each group). *, P < 0.05 by Student’s t test compared with each of the other groups. (C) Human PBMCs were transduced with the retroviral TCR vector, and CD3⁺ T cells were evaluated for transduction efficiency in CD8⁺ and CD8⁻ T cell populations by the specific tetramer. Of the 51% of CD8⁺ T cells, 22.5% were tetramer⁺, and of 49% of CD4⁺ T cells, 13.4% were tetramer⁺. (D) TCR-transduced or mock-transduced primary human CD4⁺ or CD8⁺ T cells were co-cultured with T2 cells loaded with H3.3K27M, H3.3WT, or flu peptide in a 1:1 ratio overnight for IFN-γ ELISPOT. Bars and error bars represent median and SD, respectively, for IFN-γ spot counts/5 × 10⁴ T cells in triplicate. **, P < 0.01 by Student’s t test comparing TCR-transduced CD8⁺ versus TCR-transduced CD4⁺ T cells, and H3.3WT versus H3.3K27M peptide-stimulated groups for TCR-transduced CD8⁺ T cells. Data represent two independent experiments with similar results. (E) TCR-transduced or mock-transduced CD8⁺ T cells were co-cultured with T2 cells pulsed with H3.3WT, H3.3K27M, or flu peptide in a 1:1 ratio, followed by evaluation of CD69 expression. Bars and error bars represent median and SD, respectively, for percentage CD69⁺ cells among CD8⁺ T cells (n = 3 in each group). Data represent two independent experiments with similar results. *, P < 0.05 by Student’s t test compared with each of the other groups.

Figure 4.

Alanine scanning mutagenesis determines key amino acid residues in the H3.3K27M epitope required for TCR recognition. Single alanine mutations (10 in total) were introduced at every amino acid residue within the H3.3K27M epitope (10-mer). Hence, 10 synthetic peptides each containing the specific substitution with alanine (A1–A3 and A5–A10) or valine (A4) were evaluated. In addition, synthetic peptides designed for nonmutated H3.3WT peptide and citrullinated H3.3K27M (Cit H3.3; i.e., the first amino acid of the H3.3K27M epitope was replaced by citrulline) were evaluated. (A) Relative HLA-A*0201 binding affinity of each peptide to that of H3.3K27M was determined by cell-free binding assay. This experiment was performed once. (B) J.RT3-T3.5 cells were transduced with H3.3K27M-specific TCR and evaluated for IL-2 production in response to T2 cells loaded with each peptide (10 µg/ml). Each group was assayed in triplicate. Bars and error bars represent median and SD, respectively, for IL-2 production levels. Data represent two independent experiments with similar results. *, P < 0.05 was calculated using Student’s t test comparing each peptide with the mutant H3.3.

Figure 5.

Evaluation of TCR avidity to the HLA-A2-peptide complex. (A and B) SPR analysis on the binding of the TCR and H3.3K27M-HLA-A2. (A) H3.3K27M TCR at concentrations of 14, 7, 3.5, 1.75, 0.9, 0.45, 0.23, 0.12, and 0 µM was injected over immobilized H3.3K27M-HLA-A2 (500 resonance units [RU]). (B) Fitted curve for equilibrium binding that resulted in a K_D of 2.9 µM. Data represent two independent experiments with similar results. (C) T2 cells (5 × 10⁴/well) loaded with titrating concentrations of the H3.3K27M peptide were co-cultured with TCR-transduced CD4⁺ or CD8⁺ T cells derived from three donors (5 × 10⁴/well) and assessed by IFN-γ ELISPOT. The EC₅₀ of the peptide was calculated using nonlinear regression analysis. Each experiment was performed in triplicate, and data represent two independent experiments with similar results. (D) T2 cells were pulsed with 10 µg/ml H3.3WT, H3.3K27M, flu peptide, or no peptide, and HLA-A2^negH3.3K27M⁺HSJD-DIPG019 cells were pulsed with 10 µg/ml H3.3K27M peptide. These cells were co-cultured with CD4⁺ or CD8⁺ TCR-transduced T cells and assessed by IFN-γ ELISPOT. Bars and error bars represent median and SD, respectively, for IFN-γ spots. Data represent two independent experiments with similar results. Experiments were performed in triplicate. ***, P < 0.001 by Student’s t test comparing the T2/H3.3K27M group with all the other stimulation conditions for CD8⁺ T cell groups.

Figure 6.

TCR-transduced T cells lyse H3.3K27M⁺HLA-A2⁺ glioma cells in an HLA-A*0201- and H3.3K27M-dependent manner. (A–E) Cytotoxicity of TCR-transduced T cells was evaluated by LDH cytotoxicity assay. Exogenous, synthetic H3.3K27M peptide (10 µg/ml) was added as a positive control group for TCR reactivity for each cell line. HLA-A2 blocking antibody was added in some experiments to determine HLA-A2–dependent TCR reactivity. (A–C) TCR-transduced or mock-transduced T cells were co-cultured with H3.3K27M⁺HLA-A*0201⁺ HSJD-DIPG-017 cells (A), H3.3K27M⁺HLA-A*0201⁺HSJD-DIPG-021 cells (B), or control H3.3K27M⁺HLA-A*0201^neg HSJD-DIPG-019 cells (C) at an E/T ratio of 1, 5, and 10 for 24 h. *, P < 0.05; **, P < 0.01 based on Student’s t test comparing TCR-transduced T cells with mock-transduced T cells. Each group was assessed in triplicate. Data represent two independent experiments with similar results. (D) HLA-A*0201-negative HSJD-DIPG-019 cells were stably transduced with lentiviral vector encoding HLA-A*0201 and evaluated for HLA-A2 expression by flow cytometry along with other cell lines. TCR-transduced or mock-transduced T cells were co-cultured with DIPG017, DIPG019, or DIPG-019-HLA-A2⁺ cells at an E/T ratio of 5, and cytotoxicity was measured by LDH assay. Bars and error bars represent median and SD, respectively, for specific percentage cytotoxicity. *, P < 0.05 based on Student’s t test. Each group was assessed in triplicate, and data represent two independent experiments with similar results. (E) TCR-transduced or control T cells were co-cultured with HLA-A*0201⁺ U87H3.3K27M cells or U87H3.3WT cells at an E/T ratio of 5. Each group was assessed in triplicate. Data represent three independent experiments with similar results. *, P < 0.05 based on Student’s t test. (F) CFSE-labeled target cells were co-cultured with TCR-transduced or control T cells with or without exogenous peptide at an E/T ratio of 5. After 24-h incubation, cells were stained with 7-AAD. %CFSE⁺7-AAD⁺ cells indicated specific percent cytotoxicity. Each group was assessed in triplicate. Data represent two independent experiments with similar results. *, P < 0.05 based on Student’s t test.

Figure 7.

Adoptive transfer of TCR-transduced T cells but not mock-transduced T cells results in inhibition of intracranial H3.3K27M⁺ glioma in NSG mice. NSG mice bearing intracranial U87H3.3K27M luciferase⁺ gliomas received intravenous infusion with PBS, mock-transduced T cells or TCR-transduced T cells. (A) Tumor growth is presented as radiance (10⁷ p/s/cm²/r) using BLI. Arrows indicate days on which mice received treatment. Lines and error bars represent median and SD, respectively, for radiance (10⁷ p/s/cm²/r) using BLI (n = 8 per group). (B) Representative BLI images of mice on days 10 and 32 after tumor inoculation. The background BLI signals were defined based on the levels seen in non–tumor-bearing mice. (C) Preferential accumulation of TCR⁺ T cells in the tumor site. At the time of intravenous infusion, ∼50% and 30% of the infused CD8⁺ and CD4⁺ T cells, respectively, were dextramer⁺ for TCR. On day 2 after the second intravenous infusion, the percentage of dextramer⁺ cells among CD8⁺ T cells and CD4⁺ T cells was evaluated in the peripheral blood and brain of mice that received TCR-transduced T cells. Bars and error bars represent median and SD, respectively, for percentage of dextramer⁺ cells among total live CD8⁺ or CD4⁺ T cells. Data indicate percentage dextramer⁺ cells among total live CD8⁺ or CD4⁺ T cells (n = 5 per group). The experimental results shown in this figure are representative of two independent experiments with reproducible results. *, P < 0.05; **, P < 0.01 using Student’s t test.