You are here

Article Text

Article menu
Basic tumor immunology
Original research
Identification of neoepitope reactive T-cell receptors guided by HLA-A*03:01 and HLA-A*11:01 immunopeptidomics
  1. Catherine M Ade1,
  2. Matthew J Sporn1,
  3. Sudipto Das2,
  4. Zhiya Yu1,
  5. Ken-ichi Hanada1,
  6. Yue A Qi3,4,
  7. Tapan Maity3,5,
  8. Xu Zhang3,6,
  9. Udayan Guha3,7,
  10. Thorkell Andresson2 and
  11. James C Yang1
  1. 1Surgery Branch, National Cancer Institute, Bethesda, Maryland, USA
  2. 2Cancer Research Technology Program, Frederick National Laboratory for Cancer Research, Frederick, Maryland, USA
  3. 3Thoracic and GI Malignancies Branch, National Cancer Institute, Bethesda, Maryland, USA
  4. 4Center for Alzheimer's and Related Dementias (CARD), National Institute on Aging and National Institute of Neurological Disorders and Stroke, Bethesda, MD, USA
  5. 5Laboratory of Cell Biology, National Cancer Institute, Bethesda, MD, USA
  6. 6Division of Cancer Treatment and Diagnosis, National Cancer Institute, Rockville, MD, USA
  7. 7NextCure Inc, Beltsville, MD, USA
  1. Correspondence to Dr Catherine M Ade; catherine.ade{at}nih.gov

Abstract

Background Tumor-specific mutated proteins can create immunogenic non-self, mutation-containing ‘neoepitopes’ that are attractive targets for adoptive T-cell therapies. To avoid the complexity of defining patient-specific, private neoepitopes, there has been major interest in targeting common shared mutations in driver genes using off-the-shelf T-cell receptors (TCRs) engineered into autologous lymphocytes. However, identifying the precise naturally processed neoepitopes to pursue is a complex and challenging process. One method to definitively demonstrate whether an epitope is presented at the cell surface is to elute peptides bound to a specific major histocompatibility complex (MHC) allele and analyze them by mass spectrometry (MS). These MS data can then be prospectively applied to isolate TCRs specific to the neoepitope.

Methods We created mono-allelic cell lines expressing one class I HLA allele and one common mutated oncogene in order to eliminate HLA deconvolution requirements and increase the signal of recovered peptides. MHC-bound peptides on the surface of these cell lines were immunoprecipitated, purified, and analyzed using liquid chromatography-tandem mass spectrometry, producing a list of mutation-containing minimal epitopes. To validate the immunogenicity of these neoepitopes, HLA-transgenic mice were vaccinated using the minimal peptides identified by MS in order to generate neoepitope-reactive TCRs. Specificity of these candidate TCRs was confirmed by peptide titration and recognition of transduced targets.

Results We identified precise neoepitopes derived from mutated isoforms of KRAS, EGFR, BRAF, and PIK3CA presented by HLA-A*03:01 and/or HLA-A*11:01 across multiple biological replicates. From our MS data, we were able to successfully isolate murine TCRs that specifically recognize four HLA-A*11:01 restricted neoepitopes (KRAS G13D, PIK3CA E545K, EGFR L858R and BRAF V600E) and three HLA-A*03:01 restricted neoepitopes (KRAS G12V, EGFR L858R and BRAF V600E).

Conclusions Our data show that an MS approach can be used to demonstrate which shared oncogene-derived neoepitopes are processed and presented by common HLA alleles, and those MS data can rapidly be used to develop TCRs against these common tumor-specific antigens. Although further characterization of these neoepitope-specific murine TCRs is required, ultimately, they have the potential to be used clinically for adoptive cell therapy.

  • antigen presentation
  • antigens
  • immunotherapy
  • immunotherapy, adoptive
  • receptors, antigen

Data availability statement

Data are available upon 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-2023-007097

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

    • Isolating neoepitope reactive T-cell receptors (TCRs) for each patient is a laborious process that requires identifying the precise HLA allele restriction and minimal epitope. Targeting recurring mutations in common driver oncogenes can quickly extend T-cell adoptive therapy to many patients with common cancers using TCR-engineered autologous lymphocytes.

    WHAT THIS STUDY ADDS

    • Mass spectrometry was used to identify processed and presented oncogene-derived neoepitopes, which were then used to rapidly develop TCRs recognizing these common tumor-specific neoantigens.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • Murine TCRs have functioned effectively in vitro and in vivo when introduced into human lymphocytes. This comprehensive survey of neoepitopes from recurrent oncogenic mutations presented by the common HLA alleles HLA-A3 and HLA-A11 identifies T-cell targets and led to the rapid generation of TCRs of relevance to clinical trials.

    Introduction

    Tumor-specific mutations found across a wide variety of human cancers are becoming increasingly appreciated in the clinical setting. As investigators implement new highly effective immunotherapies, they have discovered that tumor-specific ‘neoantigens’ drive the immune responses underlying these therapies. One form of effective immunotherapy is adoptive cell therapy (ACT), wherein tumor-reactive T-cells, first procured from tumor infiltrating lymphocytes (TIL), are expanded in vitro and given to a properly prepared recipient. This has been reported to cause objective responses in over half of treated patients with metastatic melanoma with a quarter achieving complete and durable rejection of their cancer.1 More recently, similar successes have been reported in individual patients with common epithelial cancers.2 Extensive studies on the antigens recognized by effective TIL have again revealed that mutated neoantigens are playing a key role in these clinical outcomes.3 4

    Identifying mutation-reactive TIL from a patient is laborious and time-consuming,5 6 and sometimes does not generate robust T-cell responses. In one clinical study, most patients had detectable neoantigen reactive TIL, yet only 1.6% of tumor-associated mutations generated a T-cell response.7 Additionally, this strategy does not identify all reactivities in a patient,8 and unfortunately certain cancer histologies require additional measures to sufficiently enrich T-cell populations for effective neoantigen reactivity.9 Given the complexity and sometimes inefficient results of established TIL screening assays, there has been a push to identify T-cell responses against the most common shared activating oncogenic mutations in human cancers, which potentially could treat multiple patients with off-the-shelf therapies. Targeting these antigens has the additional advantage of avoiding antigen loss as a means of immune evasion, as these are essential to the malignancy.

    Multiple neoepitope-reactive T-cell receptors (TCRs) have been successfully identified that recognize different mutated oncogenes by traditional TIL screening methods3 10 11 but this remains a hit-or-miss process. Several different strategies exist to determine whether a particular major histocompatibility complex (MHC)/mutation combination can create a neoepitope before T-cell isolation occurs. Epitope-prediction algorithms can be used to quickly predict the binding of mutated peptides to a diverse array of HLA alleles. One disadvantage to this approach is that these algorithms only predict binding of the peptide, not processing and presentation of the neoepitope. Additionally, the accuracy of the algorithms depend on how much data is available for a specific HLA restriction12 13 and which amino acids reside in the anchor residues.14 Mass spectrometry (MS) has emerged as a powerful tool that can assess the processing and presentation of precisely defined minimal epitopes at the cell surface. Recently, several neoepitopes have been identified by MS,15–21 showing the feasibility of this approach. Analyzing engineered monoallelic cell lines expressing a single HLA allele eliminates the need for in silico HLA-deconvolution,22 which is a considerable limitation and impediment in classic immunopeptidomic experiments.23 24

    Once a neoepitope and HLA molecule are known, neoepitope reactive TCRs need to be identified. Either a patient’s peripheral blood lymphocytes (PBL) can be stimulated,25 or HLA transgenic mice can be vaccinated before sorting using the appropriate minimal determinant tetramer in order to isolate these TCRs. Unmodified murine TCRs are highly active when inserted into human PBL,26 and these receptors have the theoretical advantage of not pairing with endogenous human TCR alpha and beta chains. TCR ‘mis-pairing’ can potentially generate hazardous new reactivities and often negatively affects expression at the cell surface.27–30 Multiple clinical protocols have demonstrated that human PBL modified with murine TCRs can mediate objective tumor regressions.31–33

    Here, we present an MS survey of 10 commonly occurring tumor-specific mutations presented by two high-frequency alleles in the US population: HLA-A*03:01 and HLA-A*11:01. We isolated naturally processed neoepitopes using MS and assessed their immunogenicity by identifying neoepitope reactive TCRs in HLA-transgenic mice. Using our pipeline, we identified four HLA-A*11:01 restricted neoepitopes and five HLA-A*03:01 restricted neoepitopes naturally processed and detected by MS. We were able to isolate TCRs reactive to all four HLA-A*11:01 neoepitopes and three HLA-A*03:01 neoepitopes. These off-the-shelf TCRs can potentially be used to conveniently and quickly engineer autologous peripheral blood lymphocytes for ACT clinical trials.

    Materials and methods

    Cell line characterization

    Endogenous HLA class-I expression

    The 721.221 (American Type Culture Collection, RRID:CVCL_6263) and K562 (gifted from Paul Robbins, RRID:CVCL_0004) cell lines were confirmed to be HLA class-I negative by flow cytometric analysis (BD FACSCanto II) using pan-class-I antibody fluorescein isothiocyanate (FITC)-labeled W6/32 (BioLegend, RRID:AB_314873).

    Isolating murine TCRs reactive to tumor-specific mutations

    Testing candidate TCRs by donor PBL transduction

    Confirmation of neoantigen reactivity of isolated TCRs: One week after transduction, transduced T-cells (5×104) were co-cultured with the appropriate target cells (5×104) for 24 hours. Target cell lines were first prepared by cognate HLA transduction. These HLA-monoallelic cell lines were transduced an additional time with a mutation-containing construct, as described in the supplemental materials and methods (online supplemental table 1) or pulsed with serially diluted peptides. For peptide titrations, both the mutant and matched wild-type peptides were used. The supernatants were harvested for human interferon gamma (IFN-γ) ELISA following the manufacturer’s protocol (DuoSet, R&D Systems). Experiments were performed in triplicate.

    Supplemental material

    Supplemental material

    Supplemental material

    Other details and additional experimental procedures are provided in the online supplemental materials and methods.

    Results

    Establishing tumor lines and optimizing epitope identification

    The 721.221, a B-lymphoblastoid line, and K562, an erythroleukemia line, were selected for initial testing due to the lack of endogenous class I HLA expression.34 35 Both cell lines were retrovirally transduced with either HLA-A*03:01 or HLA-A*11:01, sorted, and evaluated to verify that both lines could maintain high levels of MHC expression in culture (online supplemental figure 1A). To test whether these cell lines could function as antigen presenting cells, KRAS G12V7–16 and KRAS G12V8–16 peptides were pulsed onto 721.221/A*11:01 and K562/A*11:01 cell lines as a positive control for a known neoepitope/HLA pair.36 Both lines pulsed with peptide were recognized by their respective murine TCR (online supplemental figure 1B,C). The 721.221 and K562/A*11:01 cell lines were subsequently transduced with the KRAS G12V minigene to assess endogenous processing and presentation of mutated epitopes from a retroviral minigene construct (online supplemental table 1). Both cell lines were recognized by the murine HLA-A*11:01 restricted anti-KRAS G12V TCR (online supplemental figure 1D). We repeated these preliminary experiments using our HLA-A*11:01 restricted KRAS G12D TCR,36 demonstrating that both cell lines are capable of processing and presenting multiple HLA-A*11-restricted neoepitopes (online supplemental figure 1E,F). Due to reports of differential epitope processing by the constitutive (standard) and immunoproteasome,37–39 we tested proteasomal subunit expression in each line. 721.221 expresses the immunoproteasome and constitutive proteasome while K562 expresses only the constitutive proteasome (online supplemental figure 1G). From these data we elected to move forward testing both cell lines to evaluate how differences in the proteasome affect processing and presentation of epitopes.

    We developed an MS pipeline that could reliably and reproducibly identify neoepitopes presented on our cell lines (figure 1). We compared multiple reagents, conditions, procedures and software to optimize the recovery and identification of known HLA-A*11:01 restricted KRAS G12V and G12D epitopes (online supplemental figure 2A). KRAS G12V8–16 (VVGAVGVGK), KRAS G12V7–16 (VVVGAVGVGK), and KRAS G12D7–16 (VVVGADGVGK) peptides were identified in all samples tested (online supplemental table 2 and figure 2B), consistent with the previously reported data for this epitope.18 21 36

    Figure 1
    Figure 1

    Immunoprecipitation and mass spectrometry strategy for isolating and identifying peptides. Schematic of the strategy used for isolating and identifying MHC class I peptides. (1) The 721.221 and K562 cells were stably transduced with either HLA-A*03:01 or HLA-A*11:01. Transduced cells were sorted based on their class I MHC expression, selecting for the cells expressing the most via the pan-class-I antibody W6/32. Sorted mono-allelic cell lines were stably transduced with an oncogenic ‘minigene’ construct coding for between one and three mutations (online supplemental table 1). (2) Individual cell lines were lysed and incubated with W6/32 cross-linked sepharose beads overnight. (3) Sepharose bead complexes were gently washed before HLA class I peptides were acid eluted from MHC molecules. Eluted peptides were loaded onto preconditioned C18 columns and washed to remove impurities. Purified peptides were eluted using an acetonitrile gradient, lyophilized, and resuspended for mass spectrometry. (4) After mass spectrometry runs, isolated peptides were identified using PEAKS proteomics software with a modified proteome containing mutations of interest. Figure created with BioRender.com. LC-MS/MS, liquid chromatography-tandem mass spectrometry; MHC, major histocompatibility complex.

    MS results of mutated neoepitopes processed and presented by HLA-A*03:01 and A*11:01

    We evaluated the processing and presentation of 10 common oncogene mutations that occur across multiple tumor types (table 1).40 41 Most mutations were cloned individually into separate retroviral ‘minigene’ vectors (online supplemental table 1). The three EGFR mutations were cloned into the same retroviral construct as sequential mutated minigenes (referred to as ‘Trivalent Tandem Minigene’, or TMG) or sequential mutated exons (referred to as ‘Trivalent Tandem Exon’, or TE). All constructs were retrovirally transduced into 721.221 and K562 cell lines and kept under selection for the duration of the experiments to ensure constant expression of the oncogene.

    View this table:
    Table 1

    Clinical significance of selected mutations

    Mutated epitopes presented by HLA-A*11:01

    Class I MHC molecules were immunoprecipitated and peptides were eluted and analyzed as performed previously with the HLA-A*11:01 KRAS positive controls (figure 1). We identified four HLA-A*11:01 restricted neoepitopes from KRAS G13D, EGFR L858R, BRAF V600E, and PIK3CA E545K (figure 2A; online supplemental table 2). KRAS G13D peptides were identified in seven out of eight samples analyzed on the mass spectrometer. Using similar numbers of cells (~1e8 cells) expressing comparable levels of HLA for each replicate preparation allowed us to use peptide ion intensity as a semi-quantitative indicator of eluted peptide abundance. We observed that peptide ion intensities of 1×106 or greater (shown as dark green boxes in figure 2) correlated with more reproducible peptide identification between multiple replicates, while peptides identified with an ion intensity below 1×106 were often not identified in every replicate, although when present, could be verified by fragmentation analysis (light green boxes). The KRAS G13D7–16 peptide (VVVGAGDVGK) was identified in four samples with an ion intensity over 1×106 (dark green boxes in figure 2A) and in two replicates, a shorter KRAS G13D9–16 peptide (VGAGDVGK) was identified (online supplemental table 2).

    Figure 2
    Figure 2

    Mass spectrometry results of HLA-A*11:01 or HLA-A*03:01 transduced cell lines with at least one positive identification. Three to four samples were analyzed for each monoallelic line. The peptide sequence derived from the most abundant neoepitope identified in each cell line is shown in this table. The mutated amino acid for each peptide is bolded and in red. Green boxes indicate that at least one mutation-containing peptide was identified in the sample. Dark green boxes correspond to peptides with intensity values assigned by the mass spectrometer of over 1×106 when compared with all identified peptides in the sample. Light green boxes correspond to peptides with intensity values assigned by the mass spectrometer of less than 1×106 when compared with all identified peptides in the sample. Red boxes indicate that no mutation-containing peptide was identified in the sample. White boxes correspond to samples where no mass spectrometry data are available. All mutation-containing peptide sequences and relative peptide intensities identified for each cell line can be found in online supplemental table 2. (A) Mass spectrometry results of HLA-A*11:01 transduced lines. Four neoepitopes were identified from KRAS G13D, EGFR L858R, BRAF V600E, and PIK3CA E545K using our mass spectrometry pipeline. (B) Mass spectrometry results of HLA-A*03:01 transduced lines. Five neoepitopes were identified in at least one sample using our mass spectrometry pipeline from KRAS G12V, KRAS G12D, KRAS G13D, BRAF V600E and EGFR L858R. Spectra for identified peptides can be found in online supplemental table 2. One sample is defined as one cell line (721.221 or K562 expressing HLA-A*11:01 or HLA-A*03:01 and either one minigene or one tandem construct). Each sample listed is a biological replicate, as samples were processed independently of one another.

    EGFR L858R peptides were identified in five out of seven samples transduced with the EGFR TMG construct. The EGFR L858R552–560 peptide (KITDFGRAK), which has been previously shown to be immunogenic42, was identified with peptide ion intensities over 1×106 in four of those samples (figure 2A; online supplemental table 2). Interestingly, all mutant EGFR 9mer peptides recovered from K562 samples had peptide ion intensities of over 1×106, while only one out of the four 721.221 samples analyzed returned the EGFR L858R552–560 peptide with an intensity of over 1×106. Additionally, this peptide was identified in four out of seven samples transduced with the EGFR TE construct (figure 2A; online supplemental table 2). As observed with the TMG samples, this peptide was identified from all K562 samples run, yet was found in only one 721.221 sample run on the mass spectrometer. These results indicate a possible processing discrepancy between the standard and immunoproteasome of this HLA-A*11:01 restricted epitope. Our MS pipeline was unable to detect peptides containing the EGFR E746-A750Δ or T790M mutation in all EGFR HLA-A*11:01 samples run.

    We identified the BRAF V600E591–601 peptide (KIGDFGLATEK) in two K562 replicates using our MS pipeline (figure 2A; online supplemental table 2). Peptides recovered were below the intensity value of 1×106, suggesting that it might not be an abundant epitope on the surface of the HLA-A*11:01 K562 cell line. However, the MS2 spectra confirms the identity of the peptide with good confidence.

    The final HLA-A*11:01 restricted epitope identified by our survey was the PIK3CA E545K535–545 peptide (STRDPLSEITK) (figure 2A; online supplemental table 2). This peptide was identified with ion intensities above 1×106 in all eight samples analyzed, indicating it is robustly processed and presented by both lines. Although three common PIK3CA mutations were analyzed (table 1), we were unable to identify peptides containing either the E542K or H1047R mutation (online supplemental figure 3A).

    Mutated epitopes presented by HLA-A*03:01

    We identified five HLA-A*03:01 restricted epitopes from our transduced cell lines (figure 2B; online supplemental table 2). It has recently been reported that KRAS G12V, G12D, and G13D can be processed and presented by HLA-A*03:01.15 16 18 21 We were able to confirm these epitopes through our survey in both K562 and 721.221 cell lines. In our study, both KRAS G12V7–16 (VVVGAVGVGK) and G12V8–16 (VVGAVGVGK) peptides were identified on HLA-A*03:01 with intensities above 1×106 in all samples analyzed (figure 2B; online supplemental table 2), demonstrating that both KRAS G12V epitopes are efficiently processed and presented by HLA-A*03:01 in both cell lines.

    The KRAS G12D7–16 peptide (VVVGADGVGK) was identified in six out of the eight samples run (figure 2B), consistent with the published literature.15 18 ,21 This peptide was recovered with peptide ion intensities above 1×106 in all K562 samples run, but only identified in two 721.221 samples tested. We also recovered the KRAS G12D5–16 peptide (KLVVVGADGVGK) in all four K562 samples (online supplemental table 2); this peptide has not been identified by previous MS surveys. We did not detect G12D5–16 in any of the 721.221 experiments, indicating a possible discrepancy in processing and presenting these two HLA-A*03:01-restricted KRAS G12D epitopes between K562 and 721.221.

    We identified KRAS G13D7–16 (VVVGAGDVGK) and KRAS G13D5–16 (KLVVVGAGDVGK) peptides in one out of seven samples analyzed (figure 2B). The peptide ion intensity was below our arbitrary ion intensity of 1×106 (online supplemental table 2), suggesting that the A*03:01 restricted KRAS G13D epitope is not abundant on the cell surface, which agrees with available targeted MS data.15 Together, these data indicate that HLA-A*03:01-restricted KRAS G13D epitopes are not processed and presented efficiently and may therefore be a poor target for T-cell immunotherapy.

    We identified the BRAF V600E591–601 peptide (KIGDFGLATEK) in all eight samples analyzed using our MS pipeline (figure 2B; online supplemental table 2). Notably, peptides were recovered with ion intensities over 1×106 in every sample, demonstrating this epitope is processed and presented well in both of our engineered cell lines. The EGFR L858R552–560 peptide (KITDFGRAK) was identified in 11 out of 15 samples transduced with the TMG construct or TE construct (figure 2B; online supplemental table 2). The EGFR 9mer peptide was identified with peptide ion intensities of over 1×106 in 10 out of the 11 samples, indicating this peptide is processed and presented well.

    Similar to our HLA-A*11:01 query, our MS strategy was unable to detect peptides containing the EGFR E746-A750Δ or T790M mutation in all EGFR HLA-A*03:01 samples run. Additionally, we were unable to identify any mutation containing PIK3CA peptides derived from the three minigene constructs tested (online supplemental figure 3B). This was surprising, as the PIK3CA E545K535–545 peptide was consistently and robustly detected by our HLA-A*11:01 query (figure 2A), and HLA-A*03:01 and HLA-A*11:01 are part of the same superfamily.43 Additionally, an HLA-A*03:01 restricted PIK3CA H1047L peptide has been recently described, demonstrating this specific PIK3CA region can be processed and presented on HLA-A*03:01.19 Taken together, these data emphasize that all HLA allele and oncogene combinations must be evaluated independently of one another to determine which are bona fide epitopes.

    Generating neoantigen-reactive TCRs in HLA transgenic mice

    Ultimately, the immunogenicity of an epitope is established when a TCR is isolated that specifically recognizes that epitope. From our proteomics work, we identified nine potential neoepitopes restricted by HLA-A*03:01 or HLA-A*11:01 that are potential candidates for TCR isolation.

    Neoepitope reactive TCRs were generated using a modified murine immunization model as previously described (figure 3).36 Transgenic mice expressing either the chimeric HLA-A3/Kb or HLA-A11/Kb were vaccinated as described (online supplemental materials and methods). Spleens and draining lymph nodes (DLN) were stimulated in vitro with the minimal epitope for approximately 10 days and were assayed for reactivity using an IFN-γ ELISA. Reactive cultures were single-cell sorted into a 96-well plate using HLA and minimal epitope-matched tetramer. The TCR alpha and beta chains were amplified and cloned into separate retroviral constructs for co-transduction into anti-CD3 stimulated donor PBLs. Candidate TCRs were tested for recognition of peptide-pulsed and minigene-transduced HLA matched target cells. For TCRs that recognized target lines well in initial experiments, bicistonic vectors coding for both the alpha and beta chains were synthesized for further testing and future clinical applications.

    Figure 3
    Figure 3

    Mouse immunization strategy for isolating and identifying neoepitope-reactive TCRs. Schematic for isolating and identifying neo-epitope reactive TCRs by mouse immunization. (1) Transgenic mice are vaccinated using an emulsion containing either CFA or IFA, the minimal epitope as determined by mass spectrometry, and HBVc helper peptide. Spleens and draining lymph nodes from vaccinated mice were harvested, processed, stimulated in vitro with the minimal epitope for over 1 week in 24-well plates. (2) Stimulated wells were screened individually for reactivities using peptide-pulsed and fully-transduced (HLA and minigene) cell lines. (3) Wells with positive reactivities were single-cell sorted using epitope-specific tetramers into 96-well plates. (4) For each sorted T cell, individual alpha and beta chains were amplified, sequenced, identified, and cloned. (5) Isolated TCRs were transduced into healthy donor PBL and tested against peptide-pulsed targets and the fully-transduced lines from which they were originally identified. Figure created with BioRender.com. CFA, complete Freund’s adjuvant; HBVc, Hepatitis B Virus Core; IFA, incomplete Freund's adjuvant; TCR, T-cell receptor; PBL, peripheral blood lymphocytes.

    Identification of HLA-A*11:01-restricted reactive TCRs

    Three TCRs were initially isolated from KRAS G13D7–16 vaccinated HLA-A11/Kb transgenic mice (online supplemental figure 4A and supplemental table 3). Only TCRs 3 and 7 demonstrated reactivity to KRAS G13D7–16, recognizing 1 ng/mL of peptide pulsed onto K562/A*11:01 cells (figure 4A, left panel). No TCR recognized the cognate wild-type KRAS peptide (figure 4A, middle panel). Lastly, TCR 3 and 7 recognized the K562/A*11:01 cell line transduced with the KRAS G13D minigene, but not K562/A*11:01 alone line (figure 4A, right panel), demonstrating that both TCRs can recognize endogenously processed and presented epitope.

    Figure 4
    Figure 4

    Target recognition of anti-mutated KRAS, PIK3CA, EGFR, and BRAF TCRs isolated from vaccinated HLA-A11/Kb mice. Alpha and beta chains of individual TCRs were isolated and cloned from HLA-A11/Kb transgenic mice vaccinated with the cognate peptide identified through our mass spectrometry pipeline. TCRs were then tested for specific recognition of neoepitopes. Initial targets were K562/A*11:01 cells pulsed with either titrated mutant peptide (left panels) or titrated wild-type peptide (middle panels) to determine specificity and sensitivity of TCRs. TCRs that showed specificity to the mutant peptide, but not wild-type peptide, were tested for recognition of processed and presented antigen on K562/A*11:01 transduced with a mutation-containing construct (right panels). (A) Result of KRAS G13D TCRs. (B) Result of PIK3CA E545K TCRs. (C) Result of EGFR L858R TCRs. (D) Result of BRAF V600E TCRs. All peptide concentrations are shown as ng/mL. Interferon release is measured as pg/mL. N=3. Mean±SD is shown. IFN-γ, interferon gamma; TCRs, T-cell receptors.

    Nine TCRs were initially isolated from PIK3CA E545K535–545 vaccinated HLA-A11/Kb transgenic mice (online supplemental figure 4B and online supplemental table 3). Seven TCRs demonstrated specific reactivity to PIK3CA E545K535–545 peptide-pulsed K562/A*11:01 cells, with TCRs 4, 11, 13, and 16 displaying peptide recognition at 0.1 ng/mL (figure 4B, left panel). No TCR recognized the PIK3CA wild-type peptide (figure 4B, middle panel) or K562 cells transduced with HLA-A*11:01 alone (figure 4B, right panel). All seven TCRs demonstrated specific reactivity to K562/A*11:01 stably transduced with the PIK3CA E545K minigene, showing these seven TCRs recognize a PIK3CA neoepitope endogenously processed and presented on HLA-A*11:01.

    Three TCRs (TCRs 1, 4, and 9) were isolated from EGFR L858R752–760 reactive splenocytes and DLN lymphocytes (online supplemental figure 4C and online supplemental table 3). All three TCRs demonstrated specific reactivity to EGFR L858R752–760 peptide-pulsed K562/A*11:01 cells down to a concentration of 1 ng/mL (figure 4C, left panel), with no reactivity to wild-type peptide (figure 4C, middle panel). All three TCRs recognized the K562/A*11:01 cell line transduced with the EGFR TMG and TE constructs, while none recognized the K562 line transduced with HLA-A*11:01 alone (figure 4C, right panel).

    Although the MS data were not robust for HLA-A*11:01 BRAF V600E591–601 (figure 2A; online supplemental table 2), we were able to initially isolate four TCRs from vaccinated HLA-A11/Kb mice (online supplemental figure 4D and online supplemental table 3). Two TCRs, TCR 3 and 4, demonstrated specific reactivity to peptide-pulsed K562/A*11:01 cells down to 1 ng/mL and did not show appreciable reactivity to wild-type peptide (figure 4D, left and middle panel). These two receptors also demonstrated specificity to the processed and presented epitope, as they recognize K562/A*11:01 cells transduced with the BRAF V600E minigene while showing no detectible recognition of K562/A*11:01 alone (figure 4D, right panel).

    Identification of HLA-A*03:01-restricted reactive TCRs

    HLA-A3/Kb transgenic mice were vaccinated with the KRAS G12V7–16 peptide, the most abundant KRAS-derived epitope identified through our MS survey (online supplemental table 2). Eleven TCRs were initially isolated from KRAS G12V7–16 reactive DLN lymphocytes (data not shown), however only TCR 807 specifically recognized KRAS G12V7–16 peptide-pulsed K562/A*03:01 cells at a concentration of 1 ng/mL (figure 5A, online supplemental figure 4E, online supplemental table 3). TCR 807 did not demonstrate any reactivity to wild-type peptide, and it recognized the HLA-A*03:01 restricted G12V7–16 epitope when processed and presented by K562/A*03:01 cells transduced with the KRAS G12V minigene (figure 5A, middle and right panel).

    Figure 5
    Figure 5

    Target recognition of anti-mutated KRAS, EGFR, and BRAF TCRs isolated from vaccinated HLA-A3/Kb mice. Alpha and beta chains of individual TCRs were isolated and cloned from HLA-A3/Kb transgenic mice vaccinated with the cognate peptide identified through our mass spectrometry pipeline. TCRs were then tested for specific recognition of neoepitopes. Initial targets were mono-allelic HLA-A*03:01 transduced cells pulsed with either titrated mutant peptide (left panels) or titrated wild-type peptide (middle panels) to determine specificity and sensitivity of TCRs. TCRs were further tested for recognition of processed and presented antigen on K562/A*03:01 transduced with a mutation-containing minigene construct (right panels). (A) Result of KRAS G12V TCRs. (B) Result of EGFR L858R TCRs. (C) Result of BRAF V600E TCRs. All peptide concentrations are shown as ng/mL. Interferon release is measured as pg/mL. N=3. Mean±SD is shown. IFN-γ, interferon gamma; TCRs, T-cell receptors.

    We attempted to isolate TCRs from HLA-A3/Kb transgenic mice that were vaccinated with either the KRAS G12D5–16 or the KRAS G12D7–16 peptide, as both were identified in our MS pipeline (online supplemental table 2). Although six TCRs were identified (online supplemental table 3), no TCR recognized either the KRAS G12D5–16 or the KRAS G12D7–16 epitope (data not shown). This demonstrates that, although an epitope may be processed and presented, it might not be immunogenic enough to elicit a T-cell response.

    Six TCRs were isolated from HLA-A3/Kb transgenic mice immunized with the EGFR L858R752–760 peptide (online supplemental figure 4F and online supplemental table 3). All six candidates demonstrated specific reactivity to EGFR L858R752–760 peptide-pulsed K562/A*03:01 cells at a concentration of 10 ng/mL, with no recognition of the respective EGFR wild-type peptide (figure 5B, left and middle panel). Additionally, all six TCRs demonstrated specificity to K562/A*03:01 cells transduced with the EGFR TMG or TE construct (figure 5B right panel).

    Three TCRs were isolated from BRAF V600E591–601 immunized HLA-A3/Kb transgenic mice (online supplemental figure 4G and online supplemental table 3). All three TCRs recognized BRAF V600E591–601 peptide-pulsed K562/A*03:01 cells at a concentration of 10 ng/mL (figure 5C, left panel). Importantly, only TCR 4 had no discernable recognition of wild-type peptide-pulsed target cells (figure 5C, middle panel). All three TCRs recognized processed and presented epitopes on the K562/A*03:01 cell line stably transduced with the BRAF V600E minigene (figure 5C, right panel). These data show the range of TCRs that can be isolated from our murine immunization protocol and highlights the importance of careful vetting to minimize off-target effects.

    Taken together, our entire data set shows that it is possible to detect minimal peptide sequences of neoepitopes through immunopeptidomics, which can then be used to generate neo-epitope reactive TCRs in HLA transgenic mice.

    Discussion

    Mounting an immune response against the mutated proteins of a patients’ cancer has proven to be the basis for many of the currently available effective immunotherapies for cancer treatment. Unfortunately, detailed studies of these neoantigens and the repertoire targeting them have shown that they are highly patient specific and require developing personalized Good Manufacturing Practice (GMP) reagents that must function in the face of an immunosuppressive tumor microenvironment. Although ACTs suffer these limitations, the immune response might be rapidly created by expanding the repertoire ex vivo to very large numbers, avoiding tumor induced inhibition. One way to simplify ACT is through targeting shared neoepitopes using off-the-shelf TCRs to genetically engineer PBL for administration. MHC restriction complicates this approach because a tumor must have the shared mutation plus the presenting MHC allele, creating hundreds, if not thousands, of potential mutation and HLA allele combinations to explore. MHC binding predictions can increase the complexity by suggesting multiple overlapping epitopes for each mutation, as well as returning many false positive and negative peptide identifications.

    We chose to narrow the search for candidate shared neoepitope-MHC targets by performing a comprehensive survey of eluted peptides from monoallelic cell lines using MS. This report focuses on the results of neoantigens presented by HLA-A*03:01 and HLA-A*11:01, members of the same HLA superfamily. Although we conducted the study using cell lines expressing the constitutive proteasome as well as the immunoproteasome, significant differential processing was not a common finding. This MS approach provided an estimate of relative epitope abundance to allow us to focus attention on the dominant epitope when multiple variants were present. Identifying the exact dominant epitope also facilitated vaccination, in vitro sensitization and tetramer construction, all of which aided in the generation and isolation of reactive TCRs. We were able to generate HLA-A*03:01 and HLA-A*11:01 restricted murine TCRs that recognize seven different mutated epitopes including KRAS G12V, KRAS G13D, PIK3CA E545K, EGFR L858R and BRAF V600E (figures 4 and 5). There are multiple peptides derived from the mutated oncogenes we tested in this study that are predicted to be strong binders to HLA-A*03:01 or HLA-A*11:01 (NetMHCPan4.1). However, we did not detect peptides from every epitope predicted by the algorithm. Notably, we were only able to detect PIK3CA E545K peptides presented on HLA-A*11:01, yet the same peptide is predicted to bind to HLA-A*03:01 as well. By using our MS approach, we can demonstrate with certainty which epitopes are processed and presented, and therefore, which epitopes should be pursued for TCR isolation.

    Our MS pipeline overcomes some key hurdles of traditional neoepitope screening assays.44 By using monoallelic cell lines, the laborious deconvolution process is removed as we know which HLA allele a mutated peptide is restricted to. Additionally, we used a discovery proteomics approach to identify neoepitopes in an unbiased manner. For example, this method was agnostic to peptide length, so we were able to identify epitopes that would not be considered by more focused approaches. Although there are many exceptions, most MHC class I peptides are generally accepted to be 9–10 amino acids long.45 46 This assumption vastly limits the identifiable epitopes in certain MS pipelines, as four of our discovered epitopes might not have been considered in targeted approaches. Our study identified the HLA-A*03:01 restricted KRAS G12D 12mer, which had not been identified in previous targeted MS approaches for this oncogene and HLA combination (online supplemental table 2).15 18 Additionally, we identified three epitopes with 11mer peptides as the minimal epitope (figure 2, online supplemental table 2). These data demonstrate that longer epitopes are not only processed and presented but can be immunogenic.

    Although we did not pursue them specifically, theoretically this system has the capability to detect post-translationally modified epitopes. Phosphorylation, deamidation, side chain modifications and peptide splicing are all known to occur on immune epitopes. The software employed has limited ability to detect these alterations, but performs best if it searches for specific predesignated modifications. One concern with, but perhaps the advantage of our MS system is the use of protein overexpression to evaluate epitopes. Because T cells have been reported to recognize between 1 and 5 peptide-MHC complexes,47 we felt the need to enhance the sensitivity of MS as much as possible. As we are using discovery proteomics, any peptide identified via the mass spectrometer needs to be abundant enough to have a signal to noise ratio sufficient to be selected by the mass spectrometer for analysis. It is challenging to perform these studies on cell lines with endogenous expression levels of both the MHC and oncogene. Physiological levels of epitopes might not be identified by MS even if they are present at levels a T cell could recognize. Still, this system could be grossly overestimating the abundance of mutated peptides expressed at the cell surface, potentially impacting the ability of our TCRs to recognize tumor lines with natural HLA and mutated oncogene expression.

    This study presents multiple TCRs that recognize cancer-specific neoepitopes that could be clinically relevant. When retrovirally engineered into donor PBL, we identified TCRs that recognized their target neoantigen in seven out of eight peptide-HLA combinations examined (figures 4 and 5). TCRs isolated through our murine immunization protocol demonstrated avidity for cognate peptide pulsed onto target cells at concentrations of 10 nM down to 0.1 nM, with little to no background reactivity to wild-type peptide. Additionally, these TCRs recognized HLA-matched cell lines transduced with the appropriate mutation without recognition of lines expressing the HLA moiety only. Isolating neoepitope reactive TCRs validates our MS approach and demonstrates that this is an efficacious strategy to identify novel TCRs to various HLA/peptide complexes.

    There are, however, some current limitations to TCRs we have identified in this study. First, these TCRs have only demonstrated reactivity to transduced cell lines, which typically express supraphysiologic levels of MHC and oncogene. Minigenes were used to express tumor specific mutations to maximize transduction efficiency and there may be quantitative differences in the processing and presentation of epitopes from minigenes versus full length mutated oncogenes. As selected TCRs are advanced towards clinical testing, larger panels of tumor lines, patient-derived xenografts, and patient-specific organoids are being assembled to determine the recognition frequency of tumors with appropriate HLA and antigen status.

    PBL engineered with murine TCRs can cause objective responses in patients. A gp100 reactive TCR was generated via murine immunization, which caused cancer regression in patients with metastatic melanoma and had an overall response rate of 19%. This murine TCR was shown to persist in circulation at similar levels as a fully human DMF5 TCR 1 month after treatment, suggesting that immune destruction of murine TCR sequences may not be a limiting factor in using these TCRs for human use.31 Additionally, our group tested an HLA-A*02 restricted murine TCR generated by vaccinating HLA-A*02 transgenic mice that recognized the tumor-germline antigen NY-ESO1. Eleven patients were treated, and seven had objective responses with no evidence of autoimmunity.48 However, there is a potential for anti-TCR immune reactions against murine TCRs, which could limit efficacy. We know from prior experiences that only a minority of patients generate humoral responses. In early trials using murine TCRs clinically, only 6 out of 26 patients surveyed developed anti-murine TCR antibodies. Importantly, there was no relationship between antibody detection and clinical response.49 This low humoral response rate may be related to the lymphodepleting chemotherapy given prior to ACT and a low CD4 T-cell count that can persist after treatment.

    Another concern are off-target toxicities, which have been observed in clinical trials using murine TCRs. In one study, a carcinoembryonic antigen (CEA)-reactive TCR was isolated from an HLA-A*02:01 transgenic mouse immunized with the CEA691–699 peptide. This TCR recognized its specific peptide as well as HLA-A*02+ CEA+ colon cancer cell lines in vitro.50 When used clinically, one in three patients had objective regression of metastatic liver lesions. More clearly, all had severe colitis due to low levels of CEA expressed in normal GI mucosa, which ultimately led to its abandonment in spite of clinical efficacy. Similar results were seen with a fully murine TCR targeting the MAGE-A3 antigen presented in the context of HLA-A*02:01. Five of nine patients achieved objective responses, but severe irreversible neurotoxicity was encountered in two patients. It was subsequently shown that this was due to off-target, off-tumor recognition of the highly homologous protein MAGE-A12 expressed in normal brain tissue.33 Consequently, in-depth analyses will be performed using cross-reactivity assays in order to identify possible off-target recognition of normal proteins.51 Nevertheless, immunogenicity may be TCR dependent, and TCRs might be generated from byproducts of introduced transgenes.52 Due to these concerns, parallel efforts are underway to use these MS data to pursue fully human TCRs from patients through in vitro sensitization.11

    In summary, we describe the isolation of TCRs to seven different HLA/peptide complexes using epitope identification by MS and vaccination of HLA transgenic mice. These TCRs demonstrate peptide avidities in the nM range, high specificity for the mutated versus wild-type epitope and the ability to recognize endogenously processed antigen using transduced cell lines. Targeting common shared mutated oncogenes, we were able to identify TCRs against seven neoepitope-HLA combinations, several of which have no other TCRs described in the literature. This provides the basis for clinical development which will require further study of each candidate for efficacy in murine models and potential toxicity. A library of similar TCRs would be a major step forward in facilitating the use of neoantigen-reactive T cells in the adoptive therapy of many patients with advanced cancer.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study was approved by the NIH Animal Care and User Committee (ACUC) guidelines, animal protocol number SB201.

    Acknowledgments

    We would like to thank C Van Wagoner for her guidance on immunoprecipitation techniques and troubleshooting. We would like to thank C Cultraro for designing and cloning the EGFR plasmids, Q Wang for the KRAS plasmids, and P Robbins for the K562 line used in these experiments. We would like to thank S Krishna for his help making the UV-exchange tetramers used for sorting the murine TCRs. Lastly, special thanks to S Farid and A Mixon for their guidance in the FACS Core.

    References

    1. ↵
      1. Goff SL,
      2. Dudley ME,
      3. Citrin DE, et al
      . Prospective evaluation comparing intensity of Lymphodepletion before adoptive transfer of tumor-infiltrating lymphocytes for patients with metastatic Melanoma. J Clin Oncol 2016;34:2389–97. doi:10.1200/JCO.2016.66.7220
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Zacharakis N,
      2. Chinnasamy H,
      3. Black M, et al
      . Immune recognition of somatic mutations leading to complete durable regression in metastatic breast cancer. Nat Med 2018;24:724–30. doi:10.1038/s41591-018-0040-8
      OpenUrlCrossRefPubMed
    3. ↵
      1. Tran E,
      2. Robbins PF,
      3. Lu Y-C, et al
      . T-cell transfer therapy targeting mutant KRAS in cancer. N Engl J Med 2016;375:2255–62. doi:10.1056/NEJMoa1609279
      OpenUrlCrossRefPubMed
    4. ↵
      1. Robbins PF,
      2. Lu Y-C,
      3. El-Gamil M, et al
      . Mining Exomic sequencing data to identify Mutated antigens recognized by Adoptively transferred tumor-reactive T cells. Nat Med 2013;19:747–52. doi:10.1038/nm.3161
      OpenUrlCrossRefPubMed
    5. ↵
      1. Rosenberg SA,
      2. Restifo NP,
      3. Yang JC, et al
      . Adoptive cell transfer: a clinical path to effective cancer Immunotherapy. Nat Rev Cancer 2008;8:299–308. doi:10.1038/nrc2355
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Rosenberg SA,
      2. Restifo NP
      . Adoptive cell transfer as personalized Immunotherapy for human cancer. Science 2015;348:62–8. doi:10.1126/science.aaa4967
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Parkhurst MR,
      2. Robbins PF,
      3. Tran E, et al
      . Unique neoantigens arise from somatic mutations in patients with gastrointestinal cancers. Cancer Discov 2019;9:1022–35. doi:10.1158/2159-8290.CD-18-1494
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. Yossef R,
      2. Tran E,
      3. Deniger DC, et al
      . Enhanced detection of Neoantigen-reactive T cells targeting unique and shared Oncogenes for personalized cancer Immunotherapy. JCI Insight 2018;3:e122467. doi:10.1172/jci.insight.122467
    9. ↵
      1. Tran E,
      2. Turcotte S,
      3. Gros A, et al
      . Cancer Immunotherapy based on Mutation-specific Cd4+ T cells in a patient with epithelial cancer. Science 2014;344:641–5. doi:10.1126/science.1251102
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Kim SP,
      2. Vale NR,
      3. Zacharakis N, et al
      . Adoptive cellular therapy with Autologous tumor-infiltrating lymphocytes and T-cell receptor-engineered T cells targeting common P53 neoantigens in human solid tumors. Cancer Immunol Res 2022;10:932–46. doi:10.1158/2326-6066.CIR-22-0040
      OpenUrlCrossRef
    11. ↵
      1. Levin N,
      2. Paria BC,
      3. Vale NR, et al
      . Identification and validation of T-cell receptors targeting RAS Hotspot mutations in human cancers for use in cell-based Immunotherapy. Clin Cancer Res 2021;27:5084–95. doi:10.1158/1078-0432.CCR-21-0849
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Reynisson B,
      2. Alvarez B,
      3. Paul S, et al
      . Netmhcpan-4.1 and Netmhciipan-4.0: improved predictions of MHC antigen presentation by concurrent motif Deconvolution and integration of MS MHC eluted ligand data. Nucleic Acids Res 2020;48:W449–54. doi:10.1093/nar/gkaa379
      OpenUrlPubMed
    13. ↵
      1. Nielsen M,
      2. Andreatta M
      . Netmhcpan-3.0; improved prediction of binding to MHC class I molecules integrating information from multiple receptor and peptide length Datasets. Genome Med 2016;8:33. doi:10.1186/s13073-016-0288-x
    14. ↵
      1. Sachs A,
      2. Moore E,
      3. Kosaloglu-Yalcin Z, et al
      . Impact of Cysteine residues on MHC binding predictions and recognition by tumor-reactive T cells. J Immunol 2020;205:539–49. doi:10.4049/jimmunol.1901173
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Wang Q,
      2. Douglass J,
      3. Hwang MS, et al
      . Direct detection and Quantification of neoantigens. Cancer Immunol Res 2019;7:1748–54. doi:10.1158/2326-6066.CIR-19-0107
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Douglass J, et al
      . Bispecific antibodies targeting mutant RAS neoantigens. Sci Immunol 2021;6:57.
    17. ↵
      1. Cleyle J,
      2. Hardy M-P,
      3. Minati R, et al
      . Immunopeptidomic analyses of colorectal cancers with and without Microsatellite instability. Mol Cell Proteomics 2022;21:100228. doi:10.1016/j.mcpro.2022.100228
    18. ↵
      1. Choi J,
      2. Goulding SP,
      3. Conn BP, et al
      . Systematic discovery and validation of T cell targets directed against Oncogenic KRAS mutations. Cell Rep Methods 2021;1:100084. doi:10.1016/j.crmeth.2021.100084
    19. ↵
      1. Chandran SS,
      2. Ma J,
      3. Klatt MG, et al
      . Immunogenicity and therapeutic targeting of a public Neoantigen derived from Mutated Pik3Ca. Nat Med 2022;28:946–57. doi:10.1038/s41591-022-01786-3
      OpenUrlCrossRef
    20. ↵
      1. Bassani-Sternberg M,
      2. Bräunlein E,
      3. Klar R, et al
      . Direct identification of clinically relevant Neoepitopes presented on native human Melanoma tissue by mass Spectrometry. Nat Commun 2016;7:13404. doi:10.1038/ncomms13404
    21. ↵
      1. Bear AS,
      2. Blanchard T,
      3. Cesare J, et al
      . Biochemical and functional characterization of mutant KRAS epitopes Validates this Oncoprotein for immunological targeting. Nature Communications 2021;12:.:4365. doi:10.1038/s41467-021-24562-2
    22. ↵
      1. Abelin JG,
      2. Keskin DB,
      3. Sarkizova S, et al
      . Mass Spectrometry profiling of HLA-associated Peptidomes in mono-Allelic cells enables more accurate EPITOPE prediction. Immunity 2017;46:315–26. doi:10.1016/j.immuni.2017.02.007
      OpenUrlCrossRefPubMed
    23. ↵
      1. Sidney J,
      2. Grey HM,
      3. Southwood S, et al
      . Definition of an HLA-A3-like Supermotif demonstrates the overlapping peptide-binding Repertoires of common HLA molecules. Hum Immunol 1996;45:79–93. doi:10.1016/0198-8859(95)00173-5
      OpenUrlCrossRefPubMedWeb of Science
    24. ↵
      1. Falk K,
      2. Rötzschke O,
      3. Takiguchi M, et al
      . Peptide motifs of HLA-A1,-A11,-A31, and-A33 molecules. Immunogenetics 1994;40:238–41. doi:10.1007/BF00167086
      OpenUrlCrossRefPubMedWeb of Science
    25. ↵
      1. Bharathan M,
      2. Trebska-McGowan K,
      3. Anna P, et al
      . Tetramer based approach for efficient identification and isolation of Neo-antigen specific Cd8 T cells from peripheral blood (PBL) of patients with metastatic cancers. J Immunother Cancer 2015;3:47. doi:10.1186/2051-1426-3-S2-P47
    26. ↵
      1. Cohen CJ,
      2. Zhao Y,
      3. Zheng Z, et al
      . Enhanced antitumor activity of murine-human hybrid T-cell receptor (TCR) in human lymphocytes is associated with improved pairing and TCR/Cd3 stability. Cancer Res 2006;66:8878–86. doi:10.1158/0008-5472.CAN-06-1450
      OpenUrlAbstract/FREE Full Text
    27. ↵
      1. Bialer G,
      2. Horovitz-Fried M,
      3. Ya’acobi S, et al
      . Selected murine residues endow human TCR with enhanced tumor recognition. J Immunol 2010;184:6232–41. doi:10.4049/jimmunol.0902047
      OpenUrlAbstract/FREE Full Text
    28. ↵
      1. Sommermeyer D,
      2. Neudorfer J,
      3. Weinhold M, et al
      . Designer T cells by T cell receptor replacement. Eur J Immunol 2006;36:3052–9. doi:10.1002/eji.200636539
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Schumacher TNM
      . T-cell-receptor gene therapy. Nat Rev Immunol 2002;2:512–9. doi:10.1038/nri841
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Kessels HW,
      2. Wolkers MC,
      3. van den Boom MD, et al
      . Immunotherapy through TCR gene transfer. Nat Immunol 2001;2:957–61. doi:10.1038/ni1001-957
      OpenUrlCrossRefPubMedWeb of Science
    31. ↵
      1. Johnson LA,
      2. Morgan RA,
      3. Dudley ME, et al
      . Gene therapy with human and Mouse T-cell receptors mediates cancer regression and targets normal tissues expressing cognate antigen. Blood 2009;114:535–46. doi:10.1182/blood-2009-03-211714
      OpenUrlAbstract/FREE Full Text
    32. ↵
      1. Parkhurst MR,
      2. Yang JC,
      3. Langan RC, et al
      . T cells targeting Carcinoembryonic antigen can mediate regression of metastatic colorectal cancer but induce severe transient colitis. Mol Ther 2011;19:620–6. doi:10.1038/mt.2010.272
      OpenUrlCrossRefPubMed
    33. ↵
      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
    34. ↵
      1. Lozzio CB,
      2. Lozzio BB
      . Human chronic Myelogenous leukemia cell-line with positive Philadelphia Chromosome. Blood 1975;45:321–34.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Shimizu Y,
      2. Geraghty DE,
      3. Koller BH, et al
      . Transfer and expression of three Cloned human non-HLA-A,B,C class I major Histocompatibility complex genes in mutant Lymphoblastoid cells. Proc Natl Acad Sci U S A 1988;85:227–31. doi:10.1073/pnas.85.1.227
      OpenUrlAbstract/FREE Full Text
    36. ↵
      1. Wang QJ,
      2. Yu Z,
      3. Griffith K, et al
      . Identification of T-cell receptors targeting KRAS-Mutated human tumors. Cancer Immunol Res 2016;4:204–14. doi:10.1158/2326-6066.CIR-15-0188
      OpenUrlAbstract/FREE Full Text
    37. ↵
      1. Morel S,
      2. Lévy F,
      3. Burlet-Schiltz O, et al
      . Processing of some antigens by the standard Proteasome but not by the Immunoproteasome results in poor presentation by Dendritic cells. Immunity 2000;12:107–17. doi:10.1016/s1074-7613(00)80163-6
      OpenUrlCrossRefPubMedWeb of Science
    38. ↵
      1. Valmori D,
      2. Gileadi U,
      3. Servis C, et al
      . Modulation of Proteasomal activity required for the generation of a cytotoxic T lymphocyte-defined peptide derived from the tumor antigen MAGE-3. J Exp Med 1999;189:895–906. doi:10.1084/jem.189.6.895
      OpenUrlAbstract/FREE Full Text
    39. ↵
      1. Chapiro J,
      2. Claverol S,
      3. Piette F, et al
      . Destructive cleavage of Antigenic peptides either by the Immunoproteasome or by the standard Proteasome results in differential antigen presentation. J Immunol 2006;176:1053–61. doi:10.4049/jimmunol.176.2.1053
      OpenUrlAbstract/FREE Full Text
    40. ↵
      1. Holt ME,
      2. Mittendorf KF,
      3. LeNoue-Newton M, et al
      . My cancer genome: Coevolution of precision oncology and a molecular oncology Knowledgebase. JCO Clin Cancer Inform 2021;5:995–1004. doi:10.1200/CCI.21.00084
      OpenUrl
    41. ↵
      1. Tate JG,
      2. Bamford S,
      3. Jubb HC, et al
      . COSMIC: the catalogue of somatic mutations in cancer. Nucleic Acids Res 2019;47:D941–7. doi:10.1093/nar/gky1015
      OpenUrlCrossRefPubMed
    42. ↵
      1. Li F,
      2. Deng L,
      3. Jackson KR, et al
      . Neoantigen vaccination induces clinical and immunologic responses in non-small cell lung cancer patients harboring EGFR mutations. Journal for Immunotherapy of Cancer 2021;9:e002531. doi:10.1136/jitc-2021-002531
    43. ↵
      1. Guan P,
      2. Doytchinova IA,
      3. Flower DR
      . HLA-A3 Supermotif defined by quantitative structure–activity relationship analysis. Protein Eng Des Sel 2003;16:11–8. doi:10.1093/proeng/gzg005
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Qi YA,
      2. Maity TK,
      3. Cultraro CM, et al
      . Proteogenomic analysis unveils the HLA class I-presented Immunopeptidome in Melanoma and EGFR-mutant lung adenocarcinoma. Molecular & Cellular Proteomics 2021;20:100136. doi:10.1016/j.mcpro.2021.100136
      OpenUrl
    45. ↵
      1. Trolle T,
      2. McMurtrey CP,
      3. Sidney J, et al
      . The length distribution of class I-restricted T cell epitopes is determined by both peptide supply and MHC allele-specific binding preference. J Immunol 2016;196:1480–7. doi:10.4049/jimmunol.1501721
      OpenUrlAbstract/FREE Full Text
    46. ↵
      1. Remesh SG,
      2. Andreatta M,
      3. Ying G, et al
      . Unconventional peptide presentation by major Histocompatibility complex (MHC) class I allele HLA-A*02:01: BREAKING CONFINEMENT. J Biol Chem 2017;292:5262–70. doi:10.1074/jbc.M117.776542
      OpenUrlAbstract/FREE Full Text
    47. ↵
      1. Irvine DJ,
      2. Purbhoo MA,
      3. Krogsgaard M, et al
      . Direct observation of ligand recognition by T cells. Nature 2002;419:845–9. doi:10.1038/nature01076
      OpenUrlCrossRefPubMedWeb of Science
    48. ↵
      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
    49. ↵
      1. Davis JL,
      2. Theoret MR,
      3. Zheng Z, et al
      . Development of human anti-murine T-cell receptor antibodies in both responding and Nonresponding patients enrolled in TCR gene therapy trials. Clin Cancer Res 2010;16:5852–61. doi:10.1158/1078-0432.CCR-10-1280
      OpenUrlAbstract/FREE Full Text
    50. ↵
      1. Parkhurst MR,
      2. Joo J,
      3. Riley JP, et al
      . Characterization of genetically modified T-cell receptors that recognize the CEA:691-699 peptide in the context of HLA-A2.1 on human colorectal cancer cells. Clin Cancer Res 2009;15:169–80. doi:10.1158/1078-0432.CCR-08-1638
      OpenUrlAbstract/FREE Full Text
    51. ↵
      1. Linette GP,
      2. Stadtmauer EA,
      3. Maus MV, et al
      . Cardiovascular toxicity and Titin cross-reactivity of affinity-enhanced T cells in myeloma and Melanoma. Blood 2013;122:863–71. doi:10.1182/blood-2013-03-490565
      OpenUrlAbstract/FREE Full Text
    52. ↵
      1. Riddell SR,
      2. Elliott M,
      3. Lewinsohn DA, et al
      . T–cell mediated rejection of gene–modified HIV–specific cytotoxic T lymphocytes in HIV–infected patients. Nat Med 1996;2:216–23. doi:10.1038/nm0296-216
      OpenUrlCrossRefPubMedWeb of Science

    Supplementary materials

    Footnotes

    • Contributors CMA: Data Curation, Formal Analysis, Investigation, Methodology, Project Administration, Validation, Visualization, Writing—original draft, Writing—reviewing and editing. MS: Data Curation, Formal Analysis, Investigation, Validation, Visualization, Writing—original draft, Writing—reviewing and editing. SD: Data Curation, Formal Analysis, Investigation, Validation. ZY: Data Curation, Investigation, Methodology, Resources. KH: Conceptualization, Methodology, Resources, Supervision, Writing—reviewing and editing. YAQ: Data Curation, Formal Analysis, Investigation, Methodology, Software. TM and XZ: Investigation. UG: Conceptualization, Funding Acquisition, Methodology, Resources. TA: Conceptualization, Data Curation, Formal Analysis, Funding Acquisition, Methodology, Resources, Software, Supervision, Writing—original draft, Writing—reviewing and editing. JCY: Guarantor, Conceptualization, Funding Acquisition, Methodology, Project Administration, Resources, Supervision, Writing—reviewing and editing.

    • Funding The research was funded by the Intramural Research Program.

    • Competing interests The authors do not have competing interests to declare. Two non-provisional international patents have been filed based on the work described in the study.

    • 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.