Structural and kinetic basis for heightened immunogenicity of T cell vaccines

Abstract

Analogue peptides with enhanced binding affinity to major histocompatibility class (MHC) I molecules are currently being used in cancer patients to elicit stronger T cell responses. However, it remains unclear as to how alterations of anchor residues may affect T cell receptor (TCR) recognition. We correlate functional, thermodynamic, and structural parameters of TCR-peptide-MHC binding and demonstrate the effect of anchor residue modifications of the human histocompatibility leukocyte antigens (HLA)-A2 tumor epitope NY-ESO-1(157-165)-SLLMWITQC on TCR recognition. The crystal structure of the wild-type peptide complexed with a specific TCR shows that TCR binding centers on two prominent, sequential, peptide sidechains, methionine-tryptophan. Cysteine-to-valine substitution at peptide position 9, while optimizing peptide binding to the MHC, repositions the peptide main chain and generates subtly enhanced interactions between the analogue peptide and the TCR. Binding analyses confirm tighter binding of the analogue peptide to HLA-A2 and improved soluble TCR binding. Recognition of analogue peptide stimulates faster polarization of lytic granules to the immunological synapse, reduces dependence on CD8 binding, and induces greater numbers of cross-reactive cytotoxic T lymphocyte to SLLMWITQC. These results provide important insights into heightened immunogenicity of analogue peptides and highlight the importance of incorporating structural data into the process of rational optimization of superagonist peptides for clinical trials.

Figure 1.

Structure of 1G4 TCR bound to A2–NY–ESO-1 complexes. (A) Diagram highlighting the CDR loops of the 1G4 TCR complexed to the A2–ESO 9C. The CDR loops are colored as follows: V_αCDR1, blue; V_αCDR2, cyan; V_αCDR3, green; V_βCDR1, red; V_βCDR2, orange; and V_βCDR3, magenta. The peptide (yellow) projects a bulky methionine-tryptophan peg, which is embedded deep within the TCR CDR loops. (B) Simulated annealing omit electron density map yellow chicken wire showing the M₄W₅ residues bound by the 1G4 TCR. TCR contact residues are colored according to CDR color as in panel A. Hydrogen bonds between the CDR3_β and the M₄W₅ residues are shown in black. The map is contoured at 3σ. (C) Illustration of the complexed TCR colored according to panel A, and showing the close interaction of the hydrophobic peptide MW peg (yellow) with the TCR Y31α and Y100α side chains. (D) Schematic diagram showing the interactions between the M₄W₅ peg residues and the surrounding 1G4 TCR residues. TCR contact residues with a minimum interatomic distance of ≤3.8 Å are shown in green, whereas other residues surrounding the M₄W₅ peg are shown in blue. The individual distances between the TCR and the M₄W₅ peg are shown as dashed lines (distances ≤3.4 Å are shown in pink; distances between 3.4 Å and 3.8 Å are in light blue; hydrogen bonds are shown with distances in black). (E) View of the TCR antigen combining surface from the perspective of pMHC with CDR loop surfaces colored according to panel A. The figure shows the structure of the uncomplexed TCR, highlighting the Y30_α and Y100_α side chains. (F) α1α2 helix superimposition of the ESO 9C (green peptide and dark-green TCR) and ESO 9V (pink peptide and red TCR) TCR–peptide structures showing the subtle differences the anchor residue modification propagates throughout the complex interface. Notably, the COOH-terminal half of the ESO 9C peptide is 0.6 Å further out of the HLA–A2 groove than the 9V peptide and the tip of the CDR3_β loop is lower compared with the TCR–A2–ESO 9V complex.

Figure 2.

Binding measurements of peptide to A2 molecules and pMHC complex to TCR. (A) HLA A2–peptide complexes were precipitated from T2 lysate pulsed either with ESO 9C or ESO 9V peptide at concentrations shown. HLA-A2 molecules from T2 cells pulsed with either 3 μM flu matrix58–66 peptide or in the absence of peptide (open bar) were used as positive and negative controls. ESO 9C, ESO 9V, and flu matrix58–66 peptides were treated with 20 mM TCEP at room temperature for 1 h before being added to the T2 cell lysate at the final concentration of 200 μM. (B) T2 cells were stained with the A2–ESO 9C-specific Fab antibody 3M4E5. T2 cells were pulsed with either ESO 9C (left column) or ESO 9V (right column) at concentrations shown. Negative control staining with flu matrix peptide is shown in the top panel. Mean channel fluorescence is shown in each panel. (C) Affinity measurements for TCR binding to A2 complexed with ESO 9C and ESO 9V are shown in panels a and b and c and d, respectively. Panels a and c show measurements at equilibrium, whereas panels b and d show kinetic measurements. Panels a and c show 10 serial dilutions in duplicates of 1G4 NY–ESO-1–TCR (analyte), starting from a concentration of 240 μM. Panels b and d show six serial dilutions in duplicates of 1G4 NY–ESO-1–TCR, starting from a concentration of 24 μM.

Figure 3.

1G4 CTL–T2 conjugates formation is very efficient when targets are loaded with ESO 9V peptide. The panel shows a series of selected images taken from the live-cell videos in Videos 1 and 2 (available at http://www.jem.org/cgi/content/full/jem.20042323/DC1). CTLs loaded with lysotracker green were added to the adherent T2 target cells pulsed with 1 μM ESO 9C (A–E; and Video 1) or pulsed with 1 μM ESO 9V (F–L; and Video 2). The images selected from the Video 1 correspond to time 0 (A) indicating the first contact of a CTL with T2 pulsed with ESO 9C, 5 min (B), 12 min, and 20 s (C), 15 min and 20 s (D), 22 min and 40 s (E) and the ones taken from Video 2 are, respectively: time 0 (F), 1 min and 40 s (G), 5 min and 40 s (H), 6 min and 20 s (I), and 9 min and 40 s (L). The Normaski differential interference contrast image (blue) and 1 μM of green confocal image have been super-imposed at each time point. White, red, black, and light blue arrows indicate the CTLs described in the results. Bars, 10 μm.

Figure 4.

ESO 9V is very efficient in inducing polarization of CTL granules at the immunological synapse. (A) The confocal images show 1G4 CTLs conjugated with T2 pulsed with 1 μM ESO 9V. Lytic granules stained with anti-cathepsin D (green) and anti-p58 cis Golgi protein (red). Upon target cell recognition both Golgi and lytic granules move from the rear of the cell (a), around the nucleus (b), and polarize at the immunological synapse (c). (B and C) Quantitative analysis of granule polarization in 1G4 CTL clone recognizing targets pulsed with ESO 9C (B) or ESO 9V peptides (C). Cell conjugates with granules at the rear (white bar), granules moving laterally toward the synapse (gray bar) and granules at the synapse (black bar) as illustrated in A were counted using a fluorescence microscope.

Figure 5.

Role of CD8 in the recognition of the ESO 9V and ESO 9C peptides. CIR B cells transfected with either wild-type A2 or A2 DT227/8KA were pulsed with either ESO 9V or ESO 9C peptides. Supernatant was assayed for MIP-1β (top) or IFN-γ (bottom) by ELISA. Standard deviation from the mean of two duplicate assays is shown. Diamonds indicate target cells pulsed with ESO 9V peptide. Circles indicate target cells pulsed with ESO 9C peptide.

Figure 6.

Enhanced in vivo immunogenicity of ESO 9V vaccines. (A) A2K^b transgenic mice were immunized with plasmid DNA followed by recombinant vaccinia virus encoding either ESO 9C or ESO 9V peptides. Splenocytes from immunized mice were stained ex vivo 1 wk after vaccinia injection with A2K^b tetramers loaded with ESO 9C peptide (57). Results from four separate experiments were combined, each experiment includes five to eight mice (P = 0.0001). (B and C) Splenocytes from ESO 9V– immunized mice are capable of recognizing ESO 9C peptide. B shows that PBL from mice immunized with ESO 9V vaccines were ex vivo stained with A2K^b tetramers containing either ESO 9C (top row) or ESO 9V (bottom row) peptides. Percentage of tetramer⁺ of CD8⁺ cells is shown in each panel. C shows lysis of target cells (Jurkat cells transfected with A2K^b cDNA) pulsed with either ESO 9V or ESO 9C peptide. Target cells pulsed with the irrelevant peptide flu matrix58–66 (Flu) were used as negative controls. To ensure optimal presentation of the ESO 9C peptide, lysis of target cells pulsed with ESO 9C peptide was compared in the presence or absence of the reducing agent TCEP. In agreement with our previous results (19), ESO 9C pulsed target cells were killed more efficiently in the presence of the reducing agent TCEP (ESO 9C+TCEP) than in the absence of TCEP (ESO 9C). (D) Enhanced IFN-γ response by T cells from ESO 9V–primed mice. Splenocytes from either ESO 9V– or ESO 9C–immunized mice were tested for their ability to recognize target cells pulsed with ESO 9C peptide. Duplicate samples were used. Peptide concentrations and numbers of tetramer positive T cells are shown.