You are here

Article Text

Article menu
Basic tumor immunology
Original research
Novel H-2Db-restricted CD8 epitope derived from mouse MAGE-type antigen P1A mediates antitumor immunity in C57BL/6 mice
  1. James McAuliffe1,
  2. Silvia Panetti1,
  3. Emily Steffke1,
  4. Amanda Wicki1,
  5. Vinnycius Pereira-Almeida1,
  6. Laurine Noblecourt1,
  7. Yushu Hu1,2,
  8. Shi Yu William Guo1,
  9. Julie Lesenfants3,4,
  10. Ramiro Andrei Ramirez-Valdez1,
  11. Vineethkrishna Chandrasekar1,
  12. Maryam Ahmad1,
  13. Vincent Stroobant3,4,5,
  14. Nathalie Vigneron3,4,5,
  15. Benoit J Van den Eynde1,3,4,5 and
  16. Carol Sze Ki Leung1,6
  1. 1Ludwig Institute for Cancer Research, Nuffield Department of Medicine, University of Oxford, Oxford, UK
  2. 2Immunology Section, Department of Medicine, University of Verona, Verona, Italy
  3. 3de Duve Institute, UCLouvain, Brussels, Belgium
  4. 4Ludwig Institute for Cancer Research, de Duve Institute, Brussels, Belgium
  5. 5WELBIO Department, WEL Research Institute, Brussels, Belgium
  6. 6Centre for Immuno-Oncology, Nuffield Department of Medicine, University of Oxford, Oxford, UK
  1. Correspondence to Dr Carol Sze Ki Leung; carol.leung{at}immonc.ox.ac.uk; Professor Benoit J Van den Eynde; benoit.vandeneynde{at}ludwig.ox.ac.uk

Abstract

Background Melanoma antigen gene (MAGE)-type antigens are promising targets for cancer immunotherapy as they are expressed in cancer cells but not in normal tissues, except for male germline cells. The mouse P1A antigen shares this MAGE-type expression pattern and has been used as a target antigen in preclinical tumor models aiming to induce antitumor CD8+ T-cell responses. However, so far only one MHC I-restricted P1A epitope has been identified. It is presented by H-2Ld in mice of the H-2d genetic background such as DBA/2 and BALB/c. Given the availability of multiple genetically altered strains of mice in the C57BL/6 background, it would be useful to define P1A T-cell epitopes presented by the H-2b haplotype, to facilitate more refined mechanistic studies.

Methods We employed a heterologous prime-boost vaccination strategy based on a chimpanzee adenovirus (ChAdOx1) and a modified vaccinia Ankara (MVA) encoding P1A, to induce P1A-specific T-cell responses in C57BL/6 mice. Vaccine-induced responses were measured by intracellular cytokine staining and multiparameter flow cytometry. We mapped the immunogenic CD8 epitope and cloned the cognate T-cell receptor (TCR), which we used for adoptive cell therapy.

Results ChAdOx1/MVA-P1A vaccination induces a strong P1A-specific CD8+ T-cell response in C57BL/6 mice. This response is directed against a single 9-amino acid peptide with sequence FAVVTTSFL, corresponding to P1A amino acids 43–51. It is presented by H-2Db. P1A vaccination, especially with ChAdOx1 administered intramuscularly and MVA delivered intravenously, protected mice against P1A-expressing EL4 (EL4.P1A) tumor cell challenge. We identified and cloned four TCRs that are specific for the H-2Db-restricted P1A43-51 peptide. T cells transduced with these TCRs recognized EL4.P1A but not MC38.P1A and B16F10.P1A tumor cells, likely due to differences in the proteasome subtypes present in these cells. Adoptive transfer of these T cells in mice bearing EL4.P1A tumors reduced tumor growth and increased survival.

Conclusions We identified the first CD8+ T-cell epitope of the MAGE-type P1A tumor antigen presented in the H-2b background. This opens new perspectives for mechanistic studies dissecting MAGE-type specific antitumor immunity, making use of the wealth of genetically altered mouse strains available in the C57BL/6 background. This should facilitate the advancement of specific cancer immunotherapies.

  • Vaccine
  • T cell
  • T cell Receptor - TCR
  • Major histocompatibility complex - MHC
  • Adoptive cell therapy - ACT

Data availability statement

Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as online supplemental information. Additional information required is available from the corresponding author on reasonable request.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/jitc-2024-008998

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    WHAT IS ALREADY KNOWN ON THIS TOPIC

    • The mouse P1A tumor antigen is one of the first identified tumor antigens and shares strong similarities with the human MAGE-type antigen. To date, only one P1A CD8 epitope presented by H-2Ld has been identified, limiting its study to mice with the H-2d haplotype present in the DBA/2 and BALB/c strains.

    WHAT THIS STUDY ADDS

    • We identified a novel H-2Db-restricted P1A CD8 epitope by investigating the strong P1A-specific CD8+ T-cell response induced following ChAdOx1/modified vaccinia Ankara-P1A vaccination in C57BL/6 mice. T cells transduced with T-cell receptors specific to this epitope showed antitumor activity against the syngeneic P1A-expressing EL4 tumor cell line. However, they did not recognize MC38.P1A and B16F10.P1A cells, likely due to differences in the proteasome subtypes.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • These findings provide new tools for studying MAGE-specific CD8+ T cell responses, facilitating the development of new and improved cancer immunotherapies.

    Introduction

    CD8+ cytotoxic T lymphocytes (CTLs) play a major role in antitumor immunity through their ability to kill cancer cells following the recognition of tumor antigenic peptides presented by MHC I molecules. Melanoma antigen gene (MAGE)-type antigens are one of the major classes of tumor antigens and are promising therapeutic targets for cancer immunotherapy strategies.1 2 MAGE-type antigens are encoded by a family of X chromosome-linked genes known as cancer-germline (CG) genes, which are largely silenced by DNA methylation in somatic tissues. In normal physiology, CG genes are expressed only in cells of the male germline and developing trophoblasts, both immune-privileged sites lacking expression of classical MHC I molecules. They remain silenced in other healthy cell types. However, due to the aberrant epigenetic events characterizing cancer progression, such as promoter region hypomethylation, they are often re-expressed by tumor cells, which do largely express MHC I molecules. This expression is detected in a substantial number of patients across various cancer types.3 MAGE-type antigens are highly immunogenic, with spontaneous CTL responses against several epitopes observed in cancer patients.4 5 Immunotherapeutic strategies targeting MAGE-type antigens, based on therapeutic cancer vaccines and adoptive T-cell transfer (ACT), have shown notable promise and are currently being tested in clinical trials.6–8 Given the critical importance of MAGE-type antigens as targets for cancer immunotherapy, further research into T-cell responses against MAGE-type antigens is urgently needed to improve the efficacy of existing immunotherapies and to facilitate the development of new modalities.

    The P1A tumor antigen is the best-known murine equivalent of the human MAGE-type antigens. P1A was one of the first identified tumor antigens and was used in pivotal studies that provided the first evidence that antigen-specific CTL responses could mediate tumor rejection and clearance.9 10 Utilization of the model antigen P1A has contributed to our understanding of antitumor immune responses and serves as a key experimental tool.11 To date, the only identified CD8+ T-cell epitope for the P1A antigen is a nonameric peptide with sequence LPYLGWLVF(35-43), which is presented by MHC I molecule H-2Ld in mice of the H-2d haplotype present in DBA/2 and BALB/c mice.12 Studying P1A35-43-specific CD8+ T-cell responses in P1A-expressing murine tumor models has been key in the preclinical development and evaluation of recent novel immunotherapies.13–15

    To our knowledge, P1A-specific CD8+ T-cell responses have so far been investigated only in mice of the H-2d haplotype. Given that the majority of transgenic and knockout mouse strains useful for mechanistic studies are only available in the C57BL/6 strain, which has the H-2b haplotype, there is a pressing need to extend the study of P1A-specific T-cell responses into this genetic background. We, therefore, set out to identify P1A epitopes in the C57BL/6 mouse strain to open up additional avenues to explore MAGE-type-specific immunity.

    We used a recently described vaccination strategy based on heterologous prime-boost with chimpanzee adenovirus (ChAdOx1) and modified vaccinia Ankara (MVA) viral vectors.13 We observed that ChAdOx1/MVA-P1A vaccination induced a strong P1A-specific CD8+ T-cell response in C57BL/6 mice. This response is directed against a single nonameric peptide, FAVVTTSFL, which corresponds to P1A amino acids 43–51 and is presented by H-2Db. We sequenced and cloned four T-cell receptors (TCRs) recognizing this peptide, and confirmed their antitumor activity against the syngeneic P1A-expressing EL4 tumor cell line in vitro and in vivo. These results provide new tools for preclinical exploration of MAGE-specific CTL responses.

    Materials and methods

    Mice

    Female C57BL/6 mice aged 6–8 weeks old used in this study were purchased from Envigo, UK. Experimental procedures were carried out in accordance with the terms of the UK Animals (Scientific Procedures) Act Project License (PPL) (PB050649E).

    Viral vectors

    The viral vectors encoding for the P1A antigen were constructed, isolated and purified as previously described.13 Briefly, the coding sequence of P1A was inserted into the E1 locus of ChAdOx1 under a CMV immediate early promoter. A sequence coding for the 26 amino acid transmembrane domain of the MHC II invariant chain was linked to the N terminus of the transgene. The MVA vectors encoding P1A were constructed with the F11 promoter driving transgene expression. The purity and identity of the viral vectors were confirmed by PCR.

    Vaccinations

    Vaccinations were performed at varying doses of either 1×107–5×108 infectious units (IU) of ChAdOx1 virus and 1×106–1×107 plaque forming units (PFU) of MVA virus, given via intramuscular injection in 50 µL total volume or via intravenous injection in 100 µL total volume.

    Cell lines

    EL4.P1A, MC38.P1A, and B16F10.P1A cell lines were generated by lentiviral transduction of parental cell lines with virus expressing the P1A protein coding sequence (cloned into pCCL lentiviral transfer vector) and established from clones screened for P1A-expression following single-cell sorting. The T429 melanoma tumor line was established previously from an induced Amela TiRP tumor,15 and T429 cells expressing standard proteasome (SP), single intermediate proteasome (SIP), double intermediate proteasome (DIP) and immunoproteasome (IP)16 were generated. Detailed methods for the generation of the T429 proteasome subtype cell lines can be found in online supplemental materials.

    Supplemental material

    Western blot

    Cellular expression of the inducible proteasome subunits β1i, β2i and β5i was assessed by Western blot. One million cells were lysed, 10–20 µg of protein lysate were separated by NuPage Bis-Tris 4%–12% gel (Invitrogen) and subsequently transferred to PVDF membrane. The membrane was blocked using 5% non-fat milk or 5% BSA in PBS-T, and then probed with the following specific antibodies: β5i PSMB8 (D1K7X, Cell Signaling), β1i PSMB9 (E7J1L, Cell Signaling), β2i PSMB10 (E6R7O, Cell Signaling), purified anti-P1A (AB 51.18), β-Actin (13E5). After incubation with secondary antibody, the membrane was visualized using a ChemiDoc imaging system.

    Tumor implantations and measurements, checkpoint inhibitor treatment

    To induce tumor growth, 2×105 EL4.P1A tumor cells were injected subcutaneously in the right flank of the mice. Tumor growth was measured 2–3 times per week and mice were sacrificed when tumor size reached a mean diameter of 15 mm. Tumor volume (V) was calculated according to the formula: V=((length (mm)×width2 (mm))×0.52). Anti-CTLA-4 (9H10, BioXcell), anti-PD-1 (RMP1-14, BioXcell) or isotype controls monoclonal antibody (mAb) were injected with a dose of 100 µg/mouse via I.P. injection in 100 µL total volume every third day, for five consecutive times. For therapeutic efficacy studies, tumor-bearing mice were randomized according to tumor size before treatment.

    Surface staining, intracellular cytokine staining and flow cytometry

    Peripheral blood mononuclear cells (PBMCs) and splenocytes were harvested as described previously.17 For surface staining, cells were incubated for 10 min at 4°C with 5 µg/mL anti-CD16/CD32 (2.4G2, BD Biosciences) to block Fc receptors, washed, and then stained for 30 min at 4°C with PE-conjugated H-2Db/P1A43-51- FAVVTTSFL multimer (manufactured and provided by Ludwig Institute for Cancer Research, Brussels, Belgium). Cells were then washed and stained for 20 min at 4°C with viability dye (LIVE/DEAD Aqua, Invitrogen) and fluorescently conjugated mAbs against surface molecules according to different staining panels; anti-CD3-APC or anti-CD3-BV421 (17A2), anti-CD8-FITC or anti-CD8-BUV395 (53–6.7), anti-CD4-BV421 or anti-CD4-Alexa-Fluor700 (GK1.5), anti-CD44-BUV805 (IM7), anti-CD62L-BV650 (MEL-14), anti-PD1-PECy7 (29F.1A12). Intracellular cytokine staining (ICS) was performed as previously described,13 with anti-IFN-γ-APC (XMG1.2), anti-IL-2-PE (JES6-5H4), and anti-TNF-α-BV650 (MP6-XT22), and then acquired on an LSRFortessa X-20 flow cytometer (BD Biosciences). For TCF1 staining, cells were surface-stained, fixed and permeabilized (eBioscience Foxp3/Transcription Factor Staining Buffer Set) and stained overnight with anti-TCF1-Alexa Fluor 647 (C63D9, Cell Signaling). All other mAbs were purchased from BioLegend. Data were analyzed with FlowJo software V.10 (Tree Star). Analysis of multifunctional CD8+ T-cell responses was performed via a Boolean analysis of IFN-γ+, TNF-α+, IL-2+ events in the CD8+ gate using FlowJo. Pestle (NIH) and SPICE (Vaccine Research Center, NIH) software were used to generate graphical representations of proportions of T cells expressing 1, 2 or all three cytokines.

    Identification of TCRs by single-cell RNA-sequencing

    Splenocytes from ChAdOx1/MVA-P1A vaccinated C57BL/6 mice were labeled with PE-conjugated H-2Db/P1A43-51-FAVVTTSFL multimer as just described. Biological replicates (n=10) for each experimental condition were pooled together at the tissue processing stage. Around 5000 live CD3+CD8+P1A43-51-FAVVTTSFL+ cells were sorted by fluorescence-activated cell sorting using a BD FACSAria III (BD Biosciences). Sorted cells were loaded into a Chromium single-cell sorting system (10×Genomics). Single Cell 5' v2 Dual Index V(D)J RNA libraries were prepared using 10×Genomics Chromium platform and reagents according to the manufacturer’s instructions by the Oxford Genomics Centre, University of Oxford. scRNA-seq data were preprocessed using the 10×Genomics CellRanger (Oxford Genomics Centre) and distinct T-cell clonotypes and full TCR sequences were analyzed using Loupe V(D)J Browser software (10×Genomics).

    Cloning of TCRs

    Selected P1A43-51 TCR cDNA sequences were produced as linear double-stranded DNA fragments (GeneArt Strings synthesis, Invitrogen). The sequences of the TCRα chain and TCRβ chain were linked via a T2A peptide, with a Ser-Gly (SG) spacer between TCRα and the 2A peptide. TCR cDNA sequences were inserted into a mouse stem cell virus (MSCV)-based retroviral vector containing an IRES-GFP cassette (TCR OTI-2A.pMIG-II) by restriction enzyme cloning, following removal of the OT-I TCR sequence in the open reading frame.

    Transgenic expression of TCR sequences and adoptive T-cell transfer

    HEK293T cells were transfected with pMIG-II TCR plasmids and a plasmid encoding gag-pol-env genes to produce ecotropic pseudotyped MSCV retrovirus encoding either P1A or OT-I TCR sequences. CD8+ T cells were isolated using MACS separation beads and columns (Miltenyi Biotec), activated with Dynabeads Mouse T-Activator CD3/CD28 (ThermoFisher) for 48 hours and then transduced with TCR-encoding retroviruses via spinfection. Transduced CD8+ T cells were then stimulated with rounds of either FAVVTTSFL or SIINFEKL peptide-pulsed irradiated EL4 cells and cultured in the presence of IL-2 (20 U/mL), IL-7 (5 ng/mL) and IL-15 (5 ng/mL) to expand Tg TCR expressing cells to a ≥90% pure P1A- or OT-I-specific population of T cells. Following expansion of a ≥90% pure P1A-specific or OT-I-specific population of T cells, live cells were isolated on a Ficoll gradient, then resuspended in sterile PBS. Ten million cells in 200 µL volume were injected intravenously in tumor-bearing mice.

    Stimulation assay of TCR-engineered T cells

    TCR A/B/C/D-transduced live T cells were counted and plated in 200 µL of RPMI, 10% FBS supplemented with 2-mercaptoethanol (50 µM, Gibco), DNase I (20 µg/mL, Roche) and anti-CD28 (2 µg/mL, Tonbo Biosciences) in presence of P1A(43−51) peptide (4 µg) or DMSO control, or EL4.P1A (with or without 30 min of P1A(43−51) peptide (4 µg) pulsation) or EL4 wild-type control, at a ratio of 1:1 T cells:tumor cells. Cells were incubated at 37°C for 5 hours, with brefeldin A (10 µg/mL, BioLegend) added for the last 4 hours and accumulation of intracellular cytokines was quantified by flow cytometry ICS. Stimulation assays using B16F10 and MC38 tumor lines included an IFN-γ pretreatment of the target cells (200 U/mL, BioLegend) for 48 hours, to improve MHC class I surface expression.

    Statistical analysis

    Statistical analyses were carried out using Prism V.10.0 (GraphPad). For immunogenicity studies, to determine the significance of comparing multiple groups a Kruskal-Wallis test with a Dunn’s post hoc analysis was performed. For comparisons between only two groups a Mann-Whitney U-test was performed. Statistically significant differences in tumor growth between different groups were determined by two-way analysis of variance, followed by Tukey’s post hoc test for multiple comparisons. Survival curves were created using the Kaplan-Meier method and statistical significance between groups was determined by log-rank (Mantel-Cox) test. All p values <0.05 were considered statistically significant.

    Results

    ChAdOx1/MVA-P1A vaccination induces a strong P1A-specific CD8+ T-cell response in C57BL/6 mice

    To investigate T-cell responses to P1A in mice with MHC haplotype H-2b, we vaccinated C57BL/6 mice with ChAdOx1 and MVA viral vectors encoding the P1A antigen, as prime and boost, respectively. This vaccination strategy was chosen for its strong stimulation of CD8+ T-cell responses,13 which are particularly relevant for cancer immunotherapy.

    We first injected viral vectors via the intramuscular route at a range of doses and time intervals (figure 1A). We assessed vaccine immunogenicity by measuring IFN-γ production by flow cytometry following ex vivo stimulation of PBMCs with a pool of overlapping 15-mer peptides spanning the P1A protein. ChAdOx1/MVA-P1A vaccination induced strong peripheral P1A-specific CD8+ T-cell responses (figure 1B). We detected significantly higher frequencies of IFN-γ-producing CD8+ T cells in the blood of mice vaccinated with the ChAdOx1-P1A prime (5×108 IU), followed by the MVA-P1A boost (107 PFU) compared with either sham vaccinated control mice or mice vaccinated with ChAdOx1/MVA expressing the irrelevant antigen DPY. Furthermore, variation of postboost response magnitude was observed dependent on vaccination schedule. The optimal regimen was a 4-week interval between prime and boost, with a mean IFN-γ production of 6.2% of the total CD8+ T-cell population, as detected by ICS. Therefore, we defined the optimal P1A vaccination regimen in the C57BL/6 strain as the ChAdOx1-P1A prime at 5×108 IU, followed by MVA-P1A boost at 107 PFU 4 weeks apart, which will be used here, unless otherwise specified. Interestingly, this optimal vaccination schedule differs from that reported for the DBA/2 strain of mice, that is, ChAdOx1-P1A at 108 IU and MVA-P1A at 107 PFU 4 weeks apart,13 which produced a more modest response of 1.9% of IFN-γ-producing CD8+ T cells in the C57BL/6 strain. Like in DBA/2 mice, vaccination stimulated a solely CD8+ T-cell response to the P1A antigen, with no detectable IFN-γ production observed within the CD4+ population (figure 1C). In addition to IFN-γ, P1A peptide stimulation induced significant TNF-α and IL-2 production in the vaccinated cohort, while negligible amounts of cytokines were detected in the control group (figure 1D). The P1A-specific CD8+ T cells triggered by this ChAdOx1/MVA vaccination strategy were polyfunctional and produced multiple cytokines in response to P1A peptide stimulation. Following boost vaccination, about 75% of the vaccine-induced P1A-specific CD8+ T cells produced two or more cytokines (figure 1E).

    Figure 1
    Figure 1

    ChAdOx1/MVA-P1A vaccination induces a strong P1A-specific CD8+ T cell response in C57BL/6 mice. C57BL/6 mice were vaccinated intramuscularly with ChAdOx1 prime and MVA boost expressing P1A or DPY (irrelevant antigen) or PBS at different doses and time intervals. (A) Experimental scheme. (B) Ex vivo IFN-γ production by CD8+ T cells in the peripheral blood on stimulation with P1A peptide pools detected by intracellular cytokine staining (ICS) and flow cytometry in the different treatment groups. ChAdOx1-P1A at 5×108 IU prime and MVA-P1A at 107 PFU boost were given 4 weeks apart and blood sampling was performed 2 weeks postprime and 2 weeks postboost. (C) Representative dot plot of IFN-γ production within the CD8 and CD4 populations of an individual mouse; (D) IFN-γ+, TNF-α+ and IL-2+ percentage of CD8+ T cells in all mice; (E) polyfunctionality of responding CD8+ T cells are represented as pie charts of mean relative proportions. Data are shown as mean±SEM, each symbol represents an individual mouse, 5–10 mice per group with key groups pooled from 2 to 3 experiments. Statistical analysis: Kruskal-Wallis with Dunn’s multiple comparisons (B), two-way ANOVA (D); *p<0.05, **p<0.01, ****p<0.0001. ANOVA, analysis of variance; IU, infectious unit; MVA, modified vaccinia Ankara; PFU, plaque forming unit.

    We concluded that C57BL/6 mice can mount a strong CD8+ T-cell immune response to tumor rejection antigen P1A in response to ChAdOx1/MVA-P1A vaccination.

    Mapping of the H-2b-restricted P1A epitope

    Next, we aimed to identify the immunogenic CTL epitopes of P1A through peptide epitope mapping experiments. We stimulated splenocytes harvested from ChAdOx1/MVA-P1A vaccinated C57BL/6 mice ex vivo with five subpools of overlapping 15-mer peptides, each covering a subsection of the P1A protein sequence, and measured IFN-γ production within the CD8 population by ICS (figure 2A). The IFN-γ response stimulated by subpool one was comparable to that induced by the full peptide pool, while sub-pools 2–5 did not induce any response. We next stimulated the splenocytes with all the individual 15-mers from subpool 1, and observed IFN-γ responses to two 15mers, P1A37−51 and P1A41−55, indicating the location of the epitope within the overlapping section 41–51 (figure 2B). The peptide-binding prediction tool Net MHC 4.0 predicted a peptide binding with high affinity to H-2Db in position 43–51, with sequence FAVVTTSFL (figure 2C).

    Figure 2
    Figure 2

    Epitope mapping of the P1A protein in C57BL/6 mice. Spleens from vaccinated C57BL/6 mice (ChAdOx1-P1A 5×108 IU prime and MVA-P1A 107 PFU boost) were harvested, processed to single-cell suspension and stimulated with different peptide fractions of the P1A protein or vehicle control (DMSO). Frequency of IFN-γ+ cells in CD8+ T cells detected by ICS on: (A) stimulation with overlapping 15-mer P1A peptides organized in five sub-pools or all peptides covering the entire protein; (B) stimulation with all individual 15-mers contained in Sub-pool 1. (C) Net-MHC 4.0 peptide binding prediction of the top five ranked P1A peptides in H-2b background within the amino acid interval 37–55. (D) Proportion of IFN-γ+ cells in CD8 detected by ICS on stimulation with previously determined immunogenic intervals or their adjacent peptides, and the first-ranked predicted nonameric epitope FAVVTTSLF P1A43-51. (E) IFN-γ-producing cells in response to two 8-mer generated as truncated versions of P1A43-51-FAVVTTSFL: FAVVTTSF or AVVTTSFL. (F) Graphical representation of the P1A protein and the two immunogenic epitopes in H2d mice (blue) or H2b mice (orange), with one shared amino acid (purple). Data are shown as mean+SEM, each symbol represents an individual mouse, n=3. Statistical analysis: Friedman test with Dunn’s multiple comparisons, *p≤0.05. ICS, intracellular cytokine staining; IU, infectious unit.

    This peptide induced significant IFN-γ production by CD8+ T cells, while response was absent in the adjacent portions of the protein (intervals 33–47 and 45–59) (figure 2D). To confirm FAVVTTSFL as the minimal epitope, we further tested the response to two 8-mer peptides, FAVVTTSF(43−50) and AVVTTSFL(44−51), lacking the amino acid present at either ends of FAVVTTSFL(43−51) (figure 2E). Both 8-mers failed to stimulate CD8+ splenocytes from vaccinated mice to a comparable level to the 9-mer epitope, confirming FAVVTTSFL(43−51) as the minimal epitope.

    These data show that the CD8 response against P1A in C57BL/6 mice is mounted against a single immunodominant epitope, with a minimal sequence corresponding to the 9-mer peptide FAVVTTSFL43−51, predicted to be presented by H2-Db. This novel P1A epitope differs from the previously identified P1A epitope LPYLGWLVF35-43 epitope, presented by H-2Ld. Notably, both P1A epitopes are contiguous in the P1A sequence and share the phenylalanine at position 43 (figure 2F).

    Identification of P1A43–51-specific TCR sequences

    Identifying TCR sequences from tumor-reactive T cells enables a more in-depth study of tumor-specific T-cell responses. We, therefore, aimed to identify the sequences of cognate TCRs recognizing P1A43-51 (FAVVTTSFL) presented by H-2Db. First, we produced a phycoerythrin (PE)-conjugated H2-Db tetramer loaded with the P1A43-51 peptide. This tetramer could specifically bind to the P1A-specific CD8+ T cells from splenocytes of vaccinated mice (online supplemental figure 1A). This enabled FACS sorting of the P1A43−51-specific CD8+ T cells for downstream single-cell V(D)J RNA sequencing (figure 3A,B). Clonotype analysis revealed expansion of defined T-cell clones, which we ranked in order of presence (% total barcode) (figure 3C). We selected four of the most highly represented clonotypes, arbitrarily named TCR A, B, C and D, and cloned them into retroviral expression vectors18 (TCR sequences shown in online supplemental table 1). We transduced T lymphoblast cell line BW5147 (online supplemental figure 1B) and murine primary CD8+ T cells isolated from spleen (figure 3D). Inclusion of GFP in the retroviral construct allowed for quantification of transduction efficiency. P1A43-51-specific T cells from transduced CD8+ splenocytes could be expanded with consecutive rounds of antigenic stimulation to obtain a population of >95% GFP+ cells (online supplemental figure 1C) for downstream functional assays.

    Figure 3
    Figure 3

    Identification of P1A(43-51)-specific TCR sequences. (A) Experimental scheme. Spleens from ChAdOx1/MVA-P1A vaccinated mice were harvested and processed to single-cell suspension. P1A-specific CD8+ T cells were labeled with an H-2Db P1A43-51 tetramer and isolated by FACS. cDNA libraries were prepared from P1A-specific CD8+ single cells; RNA from 1252 cells was sequenced. (B) Representative dot plot of H-2Db P1A43-51 tetramer+ CD8+ T cells sorted by FACS. (C) V(D)J gene sequences for TCR-α and TCR-β chain were determined and stratified by their frequency of total sequenced barcodes. (D) Four of the most represented clonotypes (A–D) were cloned into pMIG-II retroviral vectors to transduce CD8+ splenocytes; TCR transduction efficiency was detected by GFP expression compared with mock untransduced cells. (E) Representative flow cytometry plots of IFN-γ expression by TCR-transduced CD8+ T cells cocultured with either P1A- EL4 wild-type (WT), EL4.P1A cells or peptide-pulsed EL4.P1A cells, or stimulated with soluble P1A43-51 peptide or DMSO as positive and negative controls, respectively. (F) Magnitude of IFN-γ expression by TCR-transduced CD8+ T cells, pooled experiments (n=3, duplicates) shown relative to P1A43-51 peptide positive control. (G) P1A-specific T-cell responses on stimulation with endogenous or peptide-pulsed P1A+solid tumors (MC38.P1A and B16F10.P1A) compared with the parental P1A-WT lines, shown as frequency of IFN-γ+ CD8+ cells relative to P1A peptide control. (H) Representative western blot to assess the presence of proteasome inducible subunits (β1i, β2i and β5i) within the C57BL/6 tumor lines EL4, MC38 and B16F10, compared with T429 cells engineered to contain only a single proteasome subtype, as indicated (SP (β1-β2 -β5); SIP (β1β2β5i): DIP (β1iβ2β5i); IP (β1iβ2iβ5i). Statistical analysis: RM one-way ANOVA with Tukey’s multiple comparisons, *p≤0.05. ANOVA, analysis of variance; DIP, double intermediate proteasome; IP, immunoproteasome; SIP, single-intermediate proteasome; SP, standard proteasome.

    The TCR-engineered CD8+ T cells expressed functional TCRs, inducing IFN-γ production on recognition of the cognate antigen in a short 5-hour stimulation assay (figure 3E). In particular, the cells produced IFN-γ when co-cultured with EL4.P1A cells, a T lymphoblast cell line lentivirally transduced to express P1A (online supplemental figure 1D), compared with the parental line EL4 WT (a P1A-negative cell line), showing the capability of these TCRs to recognize the P1A epitope processed from the endogenous protein (figure 3E,F). No statistically significant difference in performance was observed across different TCR clonotypes, and magnitude of response remained comparable for all TCRs (online supplemental figure 1E-F). In addition to EL4.P1A, we tested recognition of other H-2b background tumor cell lines transduced to express P1A, such as MC38 and B16F10. However, these adherent P1A-expressing cell lines failed to significantly stimulate the P1A TCR-engineered CD8+ T cells unless when bypassing the antigen processing step by pulsing them with the P1A43-51 peptide (figure 3G). We considered this might result from differences in the proteasome subtypes present in these different cell lines. The standard proteasome (SP) core comprises two copies each of three catalytic subunit types, named β1, β2 and β5, each endowed with a specific peptide-cleavage activity. In some cells and tissues, and particularly in the context of inflammation, these catalytic subunits can be replaced by their IFNγ-inducible counterpart, β1i, β2i and β5i.16 The resulting proteasomes, named immunoproteasome (IP; β1i-β2i-β5i), double intermediate proteasome (DIP; β1i-β2-β5i) and single intermediate proteasome (SIP; β1-β2-β5i) have different peptide-cleavage specificities, with the IP and DIP generally favoring the production of peptides more suitable for MHC I presentation.16 We used Western blot to analyze the expression of these proteasome subtypes in EL4, MC38 and B16F10. As controls, we used melanoma T429 cells15 that we engineered for homogeneous expression of either SP, SIP, DIP or IP (figure 3H). Interestingly, EL4 expressed β1i and β5i subunits, with low level of β2i, corresponding to the DIP or IP proteasome subtypes, while MC38 and B16F10 rather contained the SIP and SP subtypes, respectively (figure 3H).

    Prophylactic ChAdOx1/MVA-P1A vaccination provides control of EL4.P1A tumor growth in C57BL/6 mice

    Next, we asked whether the P1A-specific immune response generated by ChAdOx1/MVA-P1A vaccination in C57BL/6 mice could improve EL4.P1A tumor control in vivo in a prophylactic setting (figure 4A). Recent publications showed that intravenous administration of viral vectors as part of a heterologous prime-boost vaccination schedule could improve immunogenicity19 and tumor regression as compared with intramuscular injection.20 We investigated whether these findings were relevant in the context of our ChAdOx1-P1A and MVA-P1A vaccination schedule. We vaccinated C57BL/6 mice 4 weeks apart with three different combinations of prime-boost administration routes: ChAdOx1 prime and MVA boost both intramuscular, prime intramuscular and boost intravenous or both intravenous. We then collected the peripheral blood post-prime and post-boost and assessed the magnitude of the CD8 immune response by ICS (figure 4B). Indeed, we observed the highest overall response on intravenous-intravenous vaccination, with an average of 12.3% of IFNγ-producing CD8+ T cells in the blood, as opposed to 5.3% and 7.9% of the intramuscular-intramuscular and intramuscular-intravenous combinations, respectively. No P1A-specific immunity was recorded in the groups receiving PBS or irrelevant antigen vaccination with ChAdOx1-DPY (intramuscular) and MVA-DPY (intravenous).

    Figure 4
    Figure 4

    Prophylactic P1A vaccination improves tumor control in C57BL/6 mice. C57BL/6 mice were vaccinated with 5×108 IU ChAdOx1-P1A and 107 PFU MVA-P1A 4 weeks apart via either intramuscular or intravenous routes, or with ChAdOx1-DPY (intramuscular) and MVA-DPY (intravenous) irrelevant antigen control or PBS sham. Mice were then implanted with EL4.P1A 0.2×106 cells subcutaneously in the right flank 2 weeks post boost. (A) Experimental scheme. (B) Vaccine immunogenicity was assessed through IFN-γ+ CD8+ T cells detected by ICS on ex vivo P1A peptide stimulation of the peripheral PBMCs, 14–17 days after prime (circle) and 10–13 days after boost (diamond); data representative of 2–4 experiments. (C) Tumor volume was measured, shown as mean±SEM; (D) tumor volume at day 17 for individual mice; (E) overall survival following tumor implantation; (F) individual tumor growth curves for each mouse in the treatment group. (G) Correlation analysis of day 17 tumor volume and post-boost CD8+ IFN-γ+ frequency for each vaccinated mouse; lines summarize correlation by vaccination route. Pooled data from 2 to 4 experiments of n=6–12 mice per group (C–G). Bars represent mean (B, D). Statistical analysis: Kruskal-Wallis with Dunn’s multiple comparisons (B, D) Two-way ANOVA with Tukey’s post hoc multiple comparisons (C), log-rank Mantel-Cox test (E) and Spearman rank correlation (G). *p<0.05, **p<0.01, ****p<0.0001. ANOVA, analysis of variance; ICS, intracellular cytokine staining; MVA, modified vaccinia Ankara; PBMCs, peripheral blood mononuclear cell.

    We then implanted the vaccinated mice subcutaneously with EL4.P1A cells and monitored tumor growth. All vaccinated groups showed significantly delayed tumor growth compared with controls, regardless of route of vaccine administration (p<0.0001 for intramuscular-intramuscular and intramuscular-intravenous, p=0.0016 for intravenous-intravenous) (figure 4C). However, when looking at day 17 postimplantation in particular, the most significant benefit was observed in the intramuscular-intravenous combination, with a mean tumor volume of 398 mm3, compared with a threefold higher mean of 1146 mm3 in the PBS group, and a twofold higher mean in the intramuscular-intramuscular and intravenous-intravenous cohorts (752 mm3 and 799 mm3, respectively) (figure 4D). Corroborating previous findings,19 the majority of the tetramer-positive CD8+ T cells induced by vaccination were effector memory TEM cells (CD62L−CD44+) (online supplemental figure 2A). Interestingly, T cells from mice receiving the vaccination via different routes showed some phenotypic changes. For example, intramuscular-intravenous and intravenous-intravenous vaccinations produced a decreased proportion of central memory TCM cells (CD62L+CD44+) among P1A-specific CD8+ T cells, as compared with the intramuscular-intramuscular vaccination group (online supplemental figure 2B).

    Tumor control translated into overall improved survival, with 80% of mice in the intramuscular-intramuscular and 87.5% in both the intramuscular-intravenous and intravenous-intravenous cohorts alive at day 20 postimplantation, contrasting with 32.3% and 25% in PBS and DPY control groups, respectively (figure 4E). The combination of ChAdOx1-P1A intramuscular and MVA-P1A intravenous resulted in the longest median survival (day 30), while intramuscular-intramuscular (day 24) and intravenous-intravenous (day 23) displayed a more modest improvement compared with the PBS or DPY controls (day 17 and day 20, respectively) (figure 4E,F). Finally, we looked at the association between tumor volume at day 17 and the frequency of P1A-specific CD8+ T cells in the blood. The strongest negative correlation was found in the intramuscular-intravenous cohort (Spearman rank, r=−0.5265), suggesting that this route of administration induced CD8+ T cells more efficacious in tumor control (figure 4G). These observations suggest that, while the magnitude of the induced T-cell response is pivotal for effective antitumor immune surveillance, the phenotype and quality of the induced T cells also play a key role in their resilience and success in tumor control.

    Adoptive T-cell transfer of P1A-specific TCRs induces therapeutic control of EL4.P1A tumors in C57BL/6 mice

    Despite the aggressiveness of the EL4.P1A tumor model, we next evaluated the therapeutic efficacy of ChAdOx1/MVA-P1A vaccination in EL4-P1A tumor-bearing mice, implementing a weekly alternated vaccination schedule (figure 5A–C). This continuous treatment schedule modestly but significantly improved tumor control and survival in these mice. On the other hand, a single prime-boost approach, even when combined with anti-PD-1 or anti-CTLA-4 checkpoint inhibitors, failed to provide appreciable effects (online supplemental figure 3A–B). We also evaluated the therapeutic efficacy of ChAdOx1/MVA-P1A in mice bearing established B16F10-P1A tumors, but a single prime-boost vaccination also did not significantly delay tumor growth nor improve survival (online supplemental figure 3C–E).

    Figure 5
    Figure 5

    Adoptive cell transfer of P1A43-51 TCR-transduced CD8+ T cells exhibits good therapeutic control of EL4.P1A tumor growth in C57BL/6 mice. (A) Experimental scheme: C57BL/6 mice were implanted with 0.2×106 EL4.P1A cells subcutaneously, before receiving intramuscular vaccination with alternating ChAdOx1-P1A (5×107 IU) and MVA-P1A (5×106 PFU) at 1-week intervals, for the duration of the study; (B) mean tumor volume and (C) survival were recorded. Data pooled from two experiments, 5–6 mice/group. (D) C57BL/6 mice were implanted with 0.2×106 EL4.P1A cells subcutaneously before receiving 107 TCR-transduced CD8+ cells (TCR A, TCR B or OT-I irrelevant TCR control or PBS) administered intravenously on day 7 following tumor implantation. (E) Mean tumor growth, (F) survival and (G) individual tumor growth curves are shown. (H) Proportion of IFN-γ+ CD8+ T cells detected by ICS following ex vivo P1A-specific stimulation of the PBMCs 10 days post-ACT. Bars represent the mean. (I) Correlation between tumor size (day 15) and CD8+ IFN-γ+ (day 17). Data pooled from two experiments, symbols represent individual mice, n=8–12 mice per group. Statistical analysis: Ordinary two-way ANOVA (B); two-way ANOVA with Dunnett’s multiple comparisons (E); log-rank Mantel-Cox test (C, F); Kruskal-Wallis with Dunn’s multiple comparisons (H) and Spearman rank correlation (I). *p<0.05, **p<0.01, ****p<0.0001. ACT, adoptive cell transfer; ANOVA, analysis of variance; IM, intramuscular; MVA, modified vaccinia Ankara; PBMC, peripheral blood mononuclear cell.

    In the last decade, MAGE-type antigens have been targeted in the clinical setting via ACT of autologous or engineered MAGE-specific T cells.21 22 For this reason, we explored a TCR-transduced T-cell therapy approach in our C57BL/6 model of murine MAGE-type antigens, using two of the P1A43-51-specific TCRs identified above. TCR A-transduced and TCR B-transduced CD8+ splenocytes from C57BL/6 mice underwent 2 rounds of P1A-specific stimulation to expand to a population of >90% P1A43-51-specific TCR-expressing T cells for injection into mice (online supplemental figure 4A). We analyzed the phenotype of the cells by flow cytometry prior to injection and confirmed their functionality (online supplemental figure 4B–D).

    We implanted C57BL/6 mice with EL4.P1A tumors subcutaneously and, once tumors were engrafted and palpable (day 7), we injected P1A43-51-specific TCR-expressing T cells or OT-I irrelevant TCR controls intravenously (figure 5D). Tumor growth was significantly reduced in mice treated with the TCR A-transduced or TCR B-transduced T cells as compared with control groups (p=0.0087 and p=0.0197, respectively) (figure 5E). This increased tumor control also translated into improved survival (p=0.0441) (figure 5F). TCR A appeared to be the most effective, achieving complete tumor clearance in 3/12 mice (figure 5G). The GFP signal of the adoptively transferred T cells remained detectable in the peripheral blood 10 days after injection (online supplemental figure 4E) and correlated with the presence of P1A-specific CD8+ T cells able to produce IFN-γ (figure 5H, online supplemental figure 4F), suggesting good persistence and functionality of the adoptively transferred T cells. Notably, tumor volume at day 15 correlated negatively with the frequency of IFN-γ+ T cells (p<0.0001, Spearman rank, r=−0.4887), as well as with the frequency of GFP+ T cells in the blood (figure 5I, online supplemental figure 4G).

    Overall, these adoptive T-cell transfer experiments consolidate the relevance of the novel H-2Db-restricted P1A epitope and the cognate TCRs to elicit antitumor activity against P1A-expressing tumors in C57BL/6 mice.

    Discussion

    In this study, we identified and characterized a new CD8+ T-cell epitope from the murine MAGE-type tumor antigen P1A presented in the H-2b haplotype. It is an H-2Db-restricted nonamer peptide with sequence FAVVTTSFL, corresponding to amino acids 43–51 of the P1A protein. This opens new experimental avenues to investigate antitumor immune responses targeting MAGE-type antigens. Great progress has been made in recent years to develop cancer immunotherapies designed to generate CTL responses against MHC I-restricted tumor antigens. Of these, MAGE-type antigens are a key class, and multiple clinical trials are currently underway assessing ACT of TCR-transduced CTLs or cancer vaccines targeting MAGE-type antigens.2 7 23 24 P1A remains the closest known murine equivalent to MAGE in terms of expression profile, and for many years, the only identified P1A CTL epitope has been the H-2Ld-restricted nonamer P1A35-43, through which much of our understanding of MAGE-specific CTL responses has been derived.9 12 The newly characterized H-2Db-restricted P1A43-51 epitope can now serve as a parallel tool in additional mouse strains of the H-2b haplotype, such as C57BL/6. Given that a huge tranche of the genetically modified mouse strains generated for basic research and mechanistic studies of tumor immunology is derived from H-2b mice (to highlight just some, Baft3 KO mice25 and TCF1 KO mice26), the ability to study P1A-specific CTL responses in the H-2b background represents a significant development. Like the P1A35-43 counterpart in H-2d mice, we found through epitope mapping studies that the P1A43-51 epitope is entirely immunodominant in the P1A CTL response in C57BL/6 mice.

    Using P1A as a model MAGE-type antigen in H-2b mice could serve as a significant improvement over other commonly used model antigen systems. To this day, a major workhorse in the experimental tumor immunology field is the chicken egg ovalbumin (OVA) antigen. Multitudes of murine tumor studies have been undertaken using common syngeneic cancer cell lines such as B16F10,27 EL428 and MC3829 genetically engineered to express OVA. In concert with TCR-transgenic OT-I or OT-II mice bearing a clonality of either CD8+ or CD4+ T cells specific for the H-Kb restricted OVA257-264 or the IAb-restricted OVA323-339 epitope, this represents a complete experimental system for the study of antitumor T-cell immunology. Studying OVA-specific T-cell responses has provided invaluable insights and findings into the workings of antitumor immunity. However, these insights can prove limited when looking from a translational point of view, because OVA, as a xenogeneic chicken antigen, is much more immunogenic than classical tumor antigens. MAGE-type antigens, for instance, are encoded by non-mutated self-genes, and as such, the MAGE-specific T-cell repertoire is subject to shaping by the processes of central and peripheral tolerance and the removal of high-avidity T-cell clones.30 Therefore, the degree to which OVA-specific T-cell responses emulate antitumor responses against natural tumor antigens is relatively low. Targeting the P1A43-51 epitope in the H-2b background, therefore, provides a more natural and translatable alternative in which to model T-cell responses against MAGE-type antigens, and thereby a new tool to test and develop novel immunotherapeutic strategies in preclinical settings.

    To set up a model system to study antitumor immunity targeting the P1A43-51 epitope in vivo, we generated cells of the H-2b-derived EL4 lymphoma line expressing P1A by transgenic methods, as we could not identify an H-2b tumor line that expressed P1A naturally. EL4.P1A cells were recognized by P1A43-51 specific CTL clones in vitro, indicating that this epitope was processed and presented through the MHC I pathway. We then performed in vivo studies and showed that a prime-boost vaccination strategy based on ChAdOx1 and MVA viral vectors encoding P1A, induced a robust CD8+ T-cell response focused on the sole epitope P1A43-51. This prophylactic vaccination delayed the growth of EL4.P1A tumors injected subsequently into vaccinated mice. Interestingly, administration of the ChAdOx1-P1A prime via the intramuscular route and the MVA-P1A boost via the intravenous route was particularly effective at controlling tumor growth. This builds on recent work showing that intravenous boosting following intramuscular prime with viral vectors is superior to boosting via the intramuscular route for generating tumor-antigen-specific T-cell responses and promoting tumor regression.19 20 In the therapeutic setting, ChAdOx1/MVA-P1A vaccination induced modest control of established EL4.P1A tumors and improved survival when alternate vaccines were given weekly. However, one dose of ChAdOx1/MVA-P1A vaccination failed to control the tumor. This might result from the insufficient time to develop a complete immune response before these aggressive tumors reach an uncontrollable stage, and some P1A-expressing tumors can evade CD8+ T-cell responses with different mechanisms.31 32 Further research is needed to investigate combinatorial immunotherapies aimed at enhancing antitumor therapeutic efficacy.

    Using scRNA-sequencing methods, we identified TCR sequences of P1A43-51-specific CD8+ T cells from vaccinated C57BL/6 mice, enabling the cloning of these TCRs, their expression via transgenic means and the generation of TCR-transgenic mice, which we are currently developing. In this study, we used an expression system based on a murine retrovirus to express the P1A43–51-specific TCRs in a range of cell lines and primary CD8+ T cells isolated from C57BL/6 mice.18 We observed that CD8+ T cells transduced with the P1A43–51 TCR recognized the P1A antigen presented by H-2b target cells and mediated therapeutic activity when adoptively transferred into mice bearing P1A+ tumors. The retroviral vector that we used relies on random integration of the transgene at non-physiological genomic sites. ACT efficacy could be further improved through CRISPR-Cas9 gene editing methods that both knockout endogenous TCR expression and insert TCR transgenes at physiological loci under endogenous promoters, thereby generating more functional tumor antigen-specific T cells.33 34 This represents one avenue for further research in this model. Overall, this provides a novel experimental tool for evaluating transgenic TCR expression of MAGE-antigen-specific TCRs for in vivo tumor immunology studies. ACT with TCR-transduced T cells directed against MAGE-type antigens are currently under active clinical evaluation, and therefore, experimental research to further understand MAGE-specific CTL responses in vivo and to improve such therapies is warranted.8

    We did not observe P1A43-51 CD8+ T-cell recognition of B16F10 or MC38 cancer lines of H-2b background engineered to express P1A in vitro, nor a significant antitumor effect against B16F10.P1A in vivo. This may be due to differences in the processing of the P1A protein by the proteasomal machinery in these particular tumor cell lines. The repertoire of peptide epitopes presented by MHC I molecules on a cell is dictated in part by the proteasome subtypes present in this cell.16 Our data show that EL4 cells express the inducible immuno-subunits of the DIP and IP proteasome subtypes, in contrast to MC38 and B16F10 expressing SIP and SP subtypes respectively. This differing proteasomal composition and possible effects on P1A43−51 epitope processing may underly differences in recognition by P1A43−51 CD8+ T cells. Further investigation is necessary to understand the impact of these proteasomal subtypes on P1A-antigen processing and T-cell recognition. This observation further reflects some of the challenges faced in the cancer immunotherapy field to induce effective tumor-specific T-cell responses against solid tumors, with different proteasome subtypes in these tumor types influencing the generation of immunogenic epitopes. Corroborating reports by other authors, we could not identify an H-2b murine cancer cell line with endogenous P1A expression.35 Lack of endogenous expression could be viewed as a limitation; however, this also mimics the natural landscape of MAGE-expression, which can be heterogeneous in human tumors and is not present in all of them.1 Nonetheless, MAGE-type antigen expression can be manipulated by pharmacological means, such as targeting the DNA methylation patterns through which their activation is regulated. Expression of MAGE-type antigens, including P1A, can be induced in tumors by treatment with DNA hypomethylating agents such as 5-aza-2′-deoxycytidine.35 This presents the tantalizing possibility of combining such drugs with MAGE-directed immunotherapies as a cancer therapeutic strategy, one which is being evaluated in the clinic.36

    Taken together, our findings expand the available toolkit for preclinical and experimental research to further understand the immunology of MAGE-specific tumor responses. Given the wide range of tumor types expressing MAGE-type antigens, we expect this additional model to enable the development of new and improved MAGE-specific immunotherapies, as well as aiding exploration of successful treatment combinations, thereby benefitting patients.

    Data availability statement

    Data are available on reasonable request. All data relevant to the study are included in the article or uploaded as online supplemental information. Additional information required is available from the corresponding author on reasonable request.

    Ethics statements

    Patient consent for publication

    References

    1. ↵
      1. Coulie PG,
      2. Van den Eynde BJ,
      3. van der Bruggen P, et al
      . Tumour antigens recognized by T lymphocytes: at the core of cancer immunotherapy. Nat Rev Cancer 2014;14:135–46. doi:10.1038/nrc3670
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Raza A,
      2. Merhi M,
      3. Inchakalody VP, et al
      . Unleashing the immune response to NY-ESO-1 cancer testis antigen as a potential target for cancer immunotherapy. J Transl Med 2020;18:140. doi:10.1186/s12967-020-02306-y
    3. ↵
      1. Florke Gee RR,
      2. Chen H,
      3. Lee AK, et al
      . Emerging roles of the MAGE protein family in stress response pathways. J Biol Chem 2020;295:16121–55. doi:10.1074/jbc.REV120.008029
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Yuan J,
      2. Gnjatic S,
      3. Li H, et al
      . CTLA-4 blockade enhances polyfunctional NY-ESO-1 specific T cell responses in metastatic melanoma patients with clinical benefit. Proc Natl Acad Sci U S A 2008;105:20410–5. doi:10.1073/pnas.0810114105
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Yuan J,
      2. Adamow M,
      3. Ginsberg BA, et al
      . Integrated NY-ESO-1 antibody and CD8+ T-cell responses correlate with clinical benefit in advanced melanoma patients treated with ipilimumab. Proc Natl Acad Sci U S A 2011;108:16723–8. doi:10.1073/pnas.1110814108
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Robbins PF,
      2. Morgan RA,
      3. Feldman SA, et al
      . Tumor regression in patients with metastatic synovial cell sarcoma and melanoma using genetically engineered lymphocytes reactive with NY-ESO-1. J Clin Oncol 2011;29:917–24. doi:10.1200/JCO.2010.32.2537
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Robbins PF,
      2. Kassim SH,
      3. Tran TLN, et al
      . A pilot trial using lymphocytes genetically engineered with an NY-ESO-1-reactive T-cell receptor: long-term follow-up and correlates with response. Clin Cancer Res 2015;21:1019–27. doi:10.1158/1078-0432.CCR-14-2708
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Nowicki TS,
      2. Berent-Maoz B,
      3. Cheung-Lau G, et al
      . A Pilot Trial of the Combination of Transgenic NY-ESO-1-reactive Adoptive Cellular Therapy with Dendritic Cell Vaccination with or without Ipilimumab. Clin Cancer Res 2019;25:2096–108. doi:10.1158/1078-0432.CCR-18-3496
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Brändle D,
      2. Bilsborough J,
      3. Rülicke T, et al
      . The shared tumor-specific antigen encoded by mouse gene P1A is a target not only for cytolytic T lymphocytes but also for tumor rejection. Eur J Immunol 1998;28:4010–9. doi:10.1002/(SICI)1521-4141(199812)28:12<4010::AID-IMMU4010>3.0.CO;2-5
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Van den Eynde B,
      2. Lethé B,
      3. Van Pel A, et al
      . The gene coding for a major tumor rejection antigen of tumor P815 is identical to the normal gene of syngeneic DBA/2 mice. J Exp Med 1991;173:1373–84. doi:10.1084/jem.173.6.1373
      OpenUrlAbstract/FREE Full Text
    11. ↵
      1. Näslund TI,
      2. Uyttenhove C,
      3. Nordström EKL, et al
      . Comparative prime-boost vaccinations using Semliki Forest virus, adenovirus, and ALVAC vectors demonstrate differences in the generation of a protective central memory CTL response against the P815 tumor. J Immunol 2007;178:6761–9. doi:10.4049/jimmunol.178.11.6761
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Lethé B,
      2. van den Eynde B,
      3. van Pel A, et al
      . Mouse tumor rejection antigens P815A and P815B: two epitopes carried by a single peptide. Eur J Immunol 1992;22:2283–8. doi:10.1002/eji.1830220916
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. McAuliffe J,
      2. Chan HF,
      3. Noblecourt L, et al
      . Heterologous prime-boost vaccination targeting MAGE-type antigens promotes tumor T-cell infiltration and improves checkpoint blockade therapy. J Immunother Cancer 2021;9. doi:10.1136/jitc-2021-003218
    14. ↵
      1. Lopes A,
      2. Vanvarenberg K,
      3. Kos Š, et al
      . Combination of immune checkpoint blockade with DNA cancer vaccine induces potent antitumor immunity against P815 mastocytoma. Sci Rep 2018;8. doi:10.1038/s41598-018-33933-7
    15. ↵
      1. Zhu J,
      2. Powis de Tenbossche CG,
      3. Cané S, et al
      . Resistance to cancer immunotherapy mediated by apoptosis of tumor-infiltrating lymphocytes. Nat Commun 2017;8:1404. doi:10.1038/s41467-017-00784-1
      OpenUrlCrossRefPubMed
    16. ↵
      1. Abi Habib J,
      2. Lesenfants J,
      3. Vigneron N, et al
      . Functional Differences between Proteasome Subtypes. Cells 2022;11:421. doi:10.3390/cells11030421
    17. ↵
      1. Leung CS,
      2. Maurer MA,
      3. Meixlsperger S, et al
      . Robust T-cell stimulation by Epstein-Barr virus-transformed B cells after antigen targeting to DEC-205. Blood 2013;121:1584–94. doi:10.1182/blood-2012-08-450775
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Holst J,
      2. Szymczak-Workman AL,
      3. Vignali KM, et al
      . Generation of T-cell receptor retrogenic mice. Nat Protoc 2006;1:406–17. doi:10.1038/nprot.2006.61
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Vardeu A,
      2. Davis C,
      3. McDonald I, et al
      . Intravenous administration of viral vectors expressing prostate cancer antigens enhances the magnitude and functionality of CD8+ T cell responses. J Immunother Cancer 2022;10. doi:10.1136/jitc-2022-005398
    20. ↵
      1. Ramirez-Valdez RA,
      2. Baharom F,
      3. Khalilnezhad A, et al
      . Intravenous heterologous prime-boost vaccination activates innate and adaptive immunity to promote tumor regression. Cell Rep 2023;42:112599. doi:10.1016/j.celrep.2023.112599
      OpenUrl
    21. ↵
      1. Morgan RA,
      2. Chinnasamy N,
      3. Abate-Daga D, et al
      . Cancer regression and neurological toxicity following anti-MAGE-A3 TCR gene therapy. J Immunother 2013;36:133–51. doi:10.1097/CJI.0b013e3182829903
      OpenUrlCrossRefPubMedWeb of Science
    22. ↵
      1. Hong DS,
      2. Van Tine BA,
      3. Biswas S, et al
      . Autologous T cell therapy for MAGE-A4+ solid cancers in HLA-A*02+ patients: a phase 1 trial. Nat Med 2023;29:104–14. doi:10.1038/s41591-022-02128-z
      OpenUrlCrossRef
    23. ↵
      1. Blackhall F,
      2. Moliner L,
      3. Samson A, et al
      . 1502TiP A phase I/IIa trial of ChAdOx1 and MVA vaccines against MAGE-A3 and NY-ESO-1. Ann Oncol 2023;34:S846. doi:10.1016/j.annonc.2023.09.2533
    24. ↵
      1. Ishihara M,
      2. Nishida Y,
      3. Kitano S, et al
      . A phase 1 trial of NY-ESO-1-specific TCR-engineered T-cell therapy combined with a lymph node-targeting nanoparticulate peptide vaccine for the treatment of advanced soft tissue sarcoma. Int J Cancer 2023;152:2554–66. doi:10.1002/ijc.34453
      OpenUrl
    25. ↵
      1. Hildner K,
      2. Edelson BT,
      3. Purtha WE, et al
      . Batf3 deficiency reveals a critical role for CD8alpha+ dendritic cells in cytotoxic T cell immunity. Science 2008;322:1097–100. doi:10.1126/science.1164206
      OpenUrlAbstract/FREE Full Text
    26. ↵
      1. Utzschneider DT,
      2. Charmoy M,
      3. Chennupati V, et al
      . T Cell Factor 1-Expressing Memory-like CD8(+) T Cells Sustain the Immune Response to Chronic Viral Infections. Immunity 2016;45:415–27. doi:10.1016/j.immuni.2016.07.021
      OpenUrlCrossRefPubMed
    27. ↵
      1. Albershardt TC,
      2. Leleux J,
      3. Parsons AJ, et al
      . Intratumoral immune activation with TLR4 agonist synergizes with effector T cells to eradicate established murine tumors. NPJ Vaccines 2020;5:50. doi:10.1038/s41541-020-0201-x
    28. ↵
      1. Nagasaki J,
      2. Inozume T,
      3. Sax N, et al
      . PD-1 blockade therapy promotes infiltration of tumor-attacking exhausted T cell clonotypes. Cell Rep 2022;38:110331. doi:10.1016/j.celrep.2022.110331
      OpenUrlCrossRef
    29. ↵
      1. Kurtulus S,
      2. Madi A,
      3. Escobar G, et al
      . Checkpoint Blockade Immunotherapy Induces Dynamic Changes in PD-1−CD8+ Tumor-Infiltrating T Cells. Immunity 2019;50:181–94. doi:10.1016/j.immuni.2018.11.014
      OpenUrlCrossRefPubMed
    30. ↵
      1. Aleksic M,
      2. Liddy N,
      3. Molloy PE, et al
      . Different affinity windows for virus and cancer-specific T-cell receptors: implications for therapeutic strategies. Eur J Immunol 2012;42:3174–9. doi:10.1002/eji.201242606
      OpenUrlCrossRefPubMed
    31. ↵
      1. Sarma S,
      2. Bai X-F,
      3. Liu J, et al
      . On the role of unmutated antigens in tumor rejection in mice with unperturbed T-cell repertoires. Cancer Res 2003;63:6051–5.
      OpenUrlPubMed
    32. ↵
      1. Bai X-F,
      2. Liu J-Q,
      3. Joshi PS, et al
      . Different lineages of P1A-expressing cancer cells use divergent modes of immune evasion for T-cell adoptive therapy. Cancer Res 2006;66:8241–9. doi:10.1158/0008-5472.CAN-06-0279
      OpenUrlAbstract/FREE Full Text
    33. ↵
      1. Legut M,
      2. Dolton G,
      3. Mian AA, et al
      . CRISPR-mediated TCR replacement generates superior anticancer transgenic T cells. Blood 2018;131:311–22. doi:10.1182/blood-2017-05-787598
      OpenUrlAbstract/FREE Full Text
    34. ↵
      1. Oh SA,
      2. Senger K,
      3. Madireddi S, et al
      . High-efficiency nonviral CRISPR/Cas9-mediated gene editing of human T cells using plasmid donor DNA. J Exp Med 2022;219:e20211530. doi:10.1084/jem.20211530
    35. ↵
      1. Guo ZS,
      2. Hong JA,
      3. Irvine KR, et al
      . De novo induction of a cancer/testis antigen by 5-aza-2’-deoxycytidine augments adoptive immunotherapy in a murine tumor model. Cancer Res 2006;66:1105–13. doi:10.1158/0008-5472.CAN-05-3020
      OpenUrlAbstract/FREE Full Text
    36. ↵
      1. Odunsi K,
      2. Matsuzaki J,
      3. James SR, et al
      . Epigenetic potentiation of NY-ESO-1 vaccine therapy in human ovarian cancer. Cancer Immunol Res 2014;2:37–49. doi:10.1158/2326-6066.CIR-13-0126
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    • Supplementary Data

      This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Footnotes

    • X @PanettiSilvia, @amandawii, @momosapien1906, @CarolSKLeung

    • JM and SP contributed equally.

    • BJVdE and CSKL contributed equally.

    • Contributors BJVdE and CSKL conceived the experiments. JM, SP and CSKL designed the experiments. JM, SP, ES, AW, VP-A, LN, YH, SYWG, JL, RAR-V, VC and MA performed the experiments. VS contributed reagents and material. JM, SP, ES, NV and CSKL analysed the data. JM, SP and CSKL wrote the manuscript, and all authors reviewed and edited the manuscript. CSKL is responsible for the overall content as guarantor.

    • Funding This work was supported by the Ludwig Institute for Cancer Research and the CRUK Oxford Centre Development Fund (CRUKDF 0216-BVDE) to BJVdE. CSKL was supported by a fellowship from Swiss National Science Foundation (P300P3_155374). AW was supported by the Berrow Foundation.

    • Competing interests BJVdE and CSKL are inventors on a patent that covers viral vectors and methods for the prevention and treatment of cancer. All other authors declare no conflict of interest.

    • Provenance and peer review Not commissioned; externally peer reviewed.

    • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.