You are here

Article Text

Article menu
Immune cell therapies and immune cell engineering
Original research
Discovery of U2AF1 neoantigens in myeloid neoplasms
  1. Melinda Ann Biernacki1,2,
  2. Jessica Lok1,
  3. Ralph Graeme Black1,
  4. Kimberly A Foster3,
  5. Carrie Cummings1,
  6. Kyle B Woodward1,
  7. Tim Monahan1,
  8. Vivian G Oehler1,2,
  9. Derek L Stirewalt1,2,
  10. David Wu4,
  11. Anthony Rongvaux1,
  12. Hans Joachim Deeg1,2 and
  13. Marie Bleakley1,5
  1. 1Translational Sciences and Therapeutics Division, Fred Hutchinson Cancer Center, Seattle, Washington, USA
  2. 2Department of Medicine, University of Washington, Seattle, Washington, USA
  3. 3Department of Immunology, University of Washington, Seattle, Washington, USA
  4. 4Department of Laboratory Medicine and Pathology, University of Washington, Seattle, Washington, USA
  5. 5Department of Pediatrics, University of Washington, Seattle, Washington, USA
  1. Correspondence to Dr Melinda Ann Biernacki; mbiernac{at}fredhutch.org

Abstract

Background Myelodysplastic syndromes (MDS) arise from somatic mutations acquired in hematopoietic stem and progenitor cells, causing cytopenias and predisposing to transformation into secondary acute myeloid leukemia (sAML). Recurrent mutations in spliceosome genes, including U2AF1, are attractive therapeutic targets as they are prevalent in MDS and sAML, arise early in neoplastic cells, and are generally absent from normal cells, including normal hematopoietic cells. MDS and sAML are susceptible to T cell-mediated killing, and thus engineered T-cell immunotherapies hold promise for their treatment. We hypothesized that targeting spliceosome mutation-derived neoantigens with transgenic T-cell receptor (TCR) T cells would selectively eradicate malignant cells in MDS and sAML.

Methods We identified candidate neoantigen epitopes from recurrent protein-coding mutations in the spliceosome genes SRSF2 and U2AF1 using a multistep in silico process. Candidate epitopes predicted to bind human leukocyte antigen (HLA) class I, be processed and presented from the parent protein, and not to be subject to tolerance then underwent in vitro immunogenicity screening. CD8+ T cells recognizing immunogenic neoantigen epitopes were evaluated in in vitro assays to assess functional avidity, confirm the predicted HLA restriction, the potential for recognition of similar peptides, and the ability to kill neoplastic cells in an antigen-specific manner. Neoantigen-specific TCR were sequenced, cloned into lentiviral vectors, and transduced into third-party T cells after knock-out of endogenous TCR, then tested in vitro for specificity and ability to kill neoplastic myeloid cells presenting the neoantigen. The efficacy of neoantigen-specific T cells was evaluated in vivo in a murine cell line-derived xenograft model.

Results We identified two neoantigens created from a recurrent mutation in U2AF1, isolated CD8+ T cells specific for the neoantigens, and demonstrated that transferring their TCR to third-party CD8+ T cells is feasible and confers specificity for the U2AF1 neoantigens. Finally, we showed that these neoantigen-specific TCR-T cells do not recognize normal hematopoietic cells but efficiently kill malignant myeloid cells bearing the specific U2AF1 mutation, including primary cells, in vitro and in vivo.

Conclusions These data serve as proof-of-concept for developing precision medicine approaches that use neoantigen-directed T-cell receptor-transduced T cells to treat MDS and sAML.

  • Antigens, Neoplasm
  • Immunotherapy, Adoptive
  • Hematologic Neoplasms

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-2023-007490

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

    • Myelodysplastic syndromes (MDS) and secondary acute myeloid leukemia (sAML) evolving from MDS are susceptible to T cell-mediated killing, but targeted therapies are limited. Although MDS and sAML have relatively few total mutations compared with solid tumors, recurrent protein-coding mutations in spliceosome genes are frequent early events in myeloid malignancy development. T-cell responses to neoantigens created from protein-coding mutations play a well-established role in clinical responses of solid tumors to immune checkpoint blockade, and T cells targeting neoantigens have been used with clinical successes in solid tumors.

    WHAT THIS STUDY ADDS

    • We discovered novel neoantigens created from the recurrent U2AF1Q157R mutation that were processed and presented from endogenous protein. We engineered T cells to express T-cell receptors specific for the U2AF1Q157R neoantigens and demonstrated that U2AF1Q157R neoantigen-specific T cells effectively and selectively kill malignant myeloid cells bearing the neoantigen but do not recognize cells lacking the neoantigen, including normal hematopoietic cells.

    HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

    • These findings indicate that U2AF1Q157R neoantigens are promising targets for T-cell receptor-transduced T-cell therapy and other precision medicine approaches for individuals with MDS and sAML harboring these mutations. Additionally, these results suggest that T-cell therapies targeting aberrant U2AF1 and other neoantigens created from MDS/sAML-specific mutations should effectively eradicate malignant myeloid cells without harm to normal hematopoietic cells and warrant further investigation.

    Background

    Myelodysplastic syndromes (MDS) are a heterogeneous group of clonal hematopoietic stem cell neoplasms that arise largely from acquisition of somatic mutations in hematopoietic stem/progenitor cells (HSPC)1 2 and lead to ineffective hematopoiesis and cytopenias. A subset of patients with MDS will progress to secondary acute myeloid leukemia (sAML), which is often refractory to standard therapies and associated with high morbidity and mortality.3 Allogeneic hematopoietic cell transplantation (HCT) can be curative for MDS and sAML but is not accessible to all patients due to the refractory nature of disease, comorbidities, advanced age, or a combination of these factors. There is an unmet need for safer, more effective therapies for MDS and sAML. The effectiveness of non-myeloablative HCT and of donor lymphocyte infusion for MDS and sAML supports the concept that these malignant cells are susceptible to T cell-mediated killing.4 T cells can be engineered to express transgenic receptors, redirecting them to recognize specific target antigens including those derived from intracellular proteins and presented in the context of surface human leukocyte antigen (HLA) molecules.5 6 Adoptive T-cell therapies employing T cells transduced with antigen-specific T-cell receptor (TCR-T) generally have little toxicity5 7 and are suitable even for HCT-ineligible patients with comorbidities.

    Protein-coding mutations in spliceosome genes are prevalent in MDS and sAML2 8–14 and other myeloid neoplasms.15 16 Spliceosome mutations are early events that occur in founding clones and contribute to the dysplastic phenotype.8 17–21 Although mutations in spliceosome genes have been observed in older individuals without overt hematologic disease,22–24 their presence in younger, apparently healthy individuals predicts future development of AML,25–27 suggesting that these mutations exist only in aberrant, neoplastic or pre-neoplastic hematopoietic cells.28 T cells recognizing neoantigens created from these mutations should therefore have a limited likelihood of causing “on-target, off-tumor” toxicity. Owing to their probable role in transformation, spliceosome mutations are typically clonal and are unlikely to be lost through deletion or transcriptional repression, and neoantigens resulting from such mutations should be optimal immunotherapy targets.29 Protein products of mutations can be processed into short peptides and presented on the cell surface in the context of HLA. These peptide-HLA complexes, targetable by T cells, are presented on malignant cells but are low or absent on normal hematopoietic cells.28 30–32 The clinical importance of T-cell responses against neoantigens in antitumor immunity has been shown in multiple cancer settings,33–38 including myeloid malignancies.39

    To identify spliceosome mutation-derived neoantigens, we employed a reverse-immunology strategy using in vitro stimulation of healthy HLA-typed donor CD8+ T cells to isolate high-avidity T cells specific for epitopes derived from recurrent spliceosome mutations.32 MDS and sAML-associated mutations in spliceosome components occur predominantly in three genes: SF3B1, SRSF2, and U2AF1. Of these, mutations in SF3B1 are associated with a more favorable prognosis and response to standard therapies, while mutations in two other spliceosome genes predict poor outcome.40–42 We therefore focused on the unfavorable-risk SRSF2 and U2AF1 mutations as proof of concept. We determined that a peptide from U2AF1Q157R is processed from endogenous protein and presented on two closely related HLA-A molecules. Lentiviral transfer of U2AF1Q157R neoantigen-specific TCR conferred epitope specificity to third-party CD8+ T cells. U2AF1Q157R neoantigen-specific TCR-T cells recognized and killed U2AF1Q157R-bearing malignant myeloid cell lines and primary cells in vitro and did not recognize non-malignant HSPC. Finally, U2AF1Q157R neoantigen-specific TCR-T cells killed neoplastic myeloid cells bearing U2AF1Q157R in vivo in a cell line-derived xenograft murine model. These data indicate that U2AF1Q157R neoantigens are promising targets for precision medicine strategies including TCR-T cell therapy for individuals with myeloid neoplasms harboring these mutations.

    Methods

    Human samples

    Mononuclear cells were isolated from non-mobilized apheresis product or from whole blood (peripheral blood mononuclear cells, PBMC) or bone marrow mononuclear cells (BMMC) by Ficoll-Hypaque (Perkin-Elmer) density gradient centrifugation. PBMC and BMMC were cryopreserved in RPMI 1640 supplemented with 20% fetal bovine serum and 10% dimethylsulfoxide (DMSO) in vapor-phase liquid nitrogen in aliquots until use. In one case, patient AML was expanded by serially transplanting BMMC into immunodeficient humanized MISTRG mice (below) and harvested from secondary recipients prior to use in experiments.

    HLA binding predictions, processing and presentation predictions, and peptides

    HLA binding predictions were made using the Immune Epitope Database (IEDB) analysis resource (http://tools.iedb.org/main/) Consensus tool,43 which combines predictions from netMHC (4.0),44–46 SMM,47 and CombLib.48 Predictions were also made using netMHCpan 4.1.49–51 All predictions were made on March 20, 2022, using the most recent versions of the algorithms at the time. The 22 HLA class I molecules evaluated for predicted binding were HLA-A*01:01, HLA-A*02:01, HLA-A*03:01, HLA-A*33:01, HLA-A*33:03, HLA-A*11:01, HLA-A*24:02, HLA-B*07:02, HLA-B*08:01, HLA-B*15:01, HLA-B*35:01, HLA-B*40:01, HLA-B*44:02, HLA-B*44:03, HLA-C*03:03, HLA-C*03:04, HLA-C*04:01, HLA-C*05:01, HLA-C*06:02, HLA-C*07:01, HLA-C*07:02, and HLA-C*12:03. Peptides were defined as candidate epitopes if they had predicted IC50<250 nM for any HLA by >1 algorithm. Cytotoxic T lymphocyte epitope predictions assessing processing and presentation were performed using netCTLpan 1.152 for the same 22 HLA class I molecules. Similar peptides were identified using the motif search option in the online ScanProsite tool (https://prosite.expasy.org/scanprosite/), setting taxonomy to Homo sapiens and otherwise using default parameters.

    Control and neoantigen epitope peptides as well as similar peptides were synthesized using standard Fmoc chemistry (GenScript), reconstituted in DMSO to 10 µg/mL, and stored at −20°C in aliquots until use.

    Immunogenicity screening and identification of spliceosome neoantigen-specific CD8+ T cells

    CD8+ T cells from HLA-typed volunteer PBMC were purified by immunomagnetic bead depletion of CD8– cells (CD8+ T Cell Isolation Kit, Miltenyi Biotec). Autologous dendritic cells (DC) generated from monocytes by a modified fast DC protocol as described32 were used as antigen-presenting cells. For each immunogenicity screen, a minimum of 10×106 CD8+ T cells were plated at 3–6×104 T cells per well in 96-well plates, along with autologous mature DC in a T cell:DC ratio of 30:1. A total of up to seven plates were used in each experiment. Prior to co-culture with CD8+ T cells, DC were incubated for 2 hours at 37°C with putative spliceosome neoantigen epitope and control peptides based on donor HLA type. Each peptide was used at a final concentration of 1 µg/mL in culture medium for the incubation. After incubation with peptides, DCs were irradiated and washed before co-culturing with T cells. Cultures were supplemented with interleukin (IL)-12 (10 ng/mL) at initiation and IL-15 (10 ng/mL) at day 7. Split-well51Cr-release cytotoxicity assays (CRA) were performed on day 12–13, using autologous lymphoblastoid cell lines (LCL) with or without peptide as target cells. A well was considered positive if it exhibited >20% lysis of peptide-pulsed LCL and lysis of peptide-pulsed LCL was >2-fold higher than LCL without peptide. Peptide-specific T cells were cloned by limiting dilution using OKT3, IL-2, and feeder cells, then screened by split-well CRA on day 11–13. A clone was considered positive if it exhibited >20% lysis of peptide-pulsed LCL and lysis of peptide-pulsed LCL was ≥5-fold higher than lysis of LCL without peptide. Positive clones were expanded using OKT3, IL-2, and feeder cells,53 and their specificity and functional avidity were evaluated in functional assays (CRA and flow cytometry-based cytotoxicity assays) as described in online supplemental methods.

    Supplemental material

    TCR sequencing, transfer into lentiviral vectors, and transduction of T cells

    U2AF1Q157R neoantigen-specific TCR beta and alpha chains were sequenced by next-generation sequencing (Adaptive Biotechnologies, Takara Bio USA) as detailed in online supplemental methods. TCR were constructed by pairing the sequences encoding the dominant TCR beta and alpha chains in each spliceosome neoantigen-specific T-cell clone and including cysteine modifications and codon optimization as previously described.54 TCR constructs were synthesized as synthetic nucleotide blocks (GeneArt, Life Technologies or HiFi gBlock, IDT) and cloned into the pRRLPPT.MPSV.WPRE lentiviral vector (LV) that included the RQR8 selection marker55 by restriction digestion and ligation.

    CD8+ T cells immunomagnetically purified from normal donor PBMC underwent knockout of endogenous TCRαβ chains using CRISPR/Cas9 technology and were then transduced as described56 and as detailed in online supplemental methods. Four days after transduction, T cells were stained with U2AF1Q157R/HLA-A*33:03 or U2AF1Q157R/HLA-A*33:01 peptide-HLA (pHLA) tetramer and anti-CD8 monoclonal antibody. Tetramer+ CD8+ T cells were sorted to >80% purity, expanded,53 then evaluated by flow cytometry and functional assays and used in murine experiments.

    Cell line-derived xenografts

    TF-1 cells co-transduced with GFP-luciferase and a minigene encoding 24 amino acids of U2AF1 encompassing either U2AF1Q157R or the wild-type equivalent sequence were generated and purified (online supplemental methods). 5×105 dual-transduced TF-1 cells were injected intravenously into sublethally irradiated (150 cGy) NSG-SGM3 mice after baseline bioluminescence imaging (IVIS Spectrum). Five days later, 1×107 U2AF1Q157R-specific TCR32 transduced T cells (with TF-1 U2AF1Q157R, experimental; with TF-1 U2AF1 wild-type, control 1), 1×107 control CD8+ TCR-T cells specific for an irrelevant neoantigen (with TF-1 U2AF1Q157R, control 2), or vehicle (with TF-1 U2AF1Q157R, control 3) was injected intravenously on a randomized basis to ensure equivalent levels of TF-1 engraftment across groups (n=7 per group). Mice were subsequently followed weekly by peripheral blood flow cytometry and bioluminescence imaging. Mice were euthanized when they developed probable disease-related symptoms (weight loss >20%, poor body condition, hind limb paralysis) or based on saturation on bioluminescence imaging suggesting imminently lethal disease burden. Terminal bone marrow samples were harvested for flow cytometry for all euthanized animals.

    Results

    Identification of putative spliceosome neoantigens

    We first investigated whether aberrant amino acid sequences resulting from recurrent mutations in SRSF2 and U2AF1 had potential to produce candidate MDS and sAML epitopes. Four features are needed for epitopes to elicit T-cell responses: strong peptide binding to HLA, cleavage of the peptide from the parent protein, loading of the peptide onto HLA, and absence of tolerance to the epitope. We performed a multistep in silico analysis to screen for candidate spliceosome neoantigen epitopes (figure 1A). We first used the HLA-binding prediction algorithms NetMHCpan 4.149–51 and the IEDB consensus algorithm43 which combines artificial neural network (ANN) and stabilized matrix method (SMM) predictions to identify spliceosome mutation-derived peptides with a high probability of binding strongly to any of 22 HLA class I molecules, including 6 HLA-A, 7 HLA-B, and 8 HLA-C alleles. We evaluated amino acid sequences resulting from mutations occurring at two hotspots in U2AF1 (S34, Q157) and at one hotspot in SRFSF2 (P95).9 Peptides were considered epitope candidates if they were predicted favorably by at least two metrics: binding affinity of less than 250 nM by one or more algorithms (IEDB ANN, IEDB SMM, and/or netMHCpan binding affinity prediction), and/or percent rank less than one for mass spectrometry-based predictions in netMHCpan. Six candidate epitopes were predicted from two mutations in SRSF2 and 19 candidate epitopes were predicted from three mutations in U2AF1 (figure 1A, online supplemental table 1). We next assessed which of these peptides were likely to be processed from the parent protein and loaded onto HLA using the algorithm netCTLpan, which incorporates predictions of proteasomal cleavage and peptide transport. Analysis of a small set of peptides known to be immunogenic, processed, and presented (online supplemental table 2) suggested that setting a threshold rank <1% captured the majority of processed and presented peptides (six of seven). Using these criteria, we identified six peptides that were predicted to be processed from the parent amino acid sequence and presented on four different HLA molecules, generating seven total predicted processed/presented peptides (figure 1B, online supplemental table 1). Finally, we evaluated the probability of the equivalent wild-type peptides binding to the predicted HLA molecules for all candidate neoantigen epitopes, reasoning that tolerance to a candidate neoantigen epitope would be more likely if the wild-type peptide were presented. For 11 of the candidate epitopes, the wild-type peptides were not predicted to bind to the relevant HLA (figure 1B, online supplemental table 1). Of these candidate epitopes with no predicted wild-type binding, two were also predicted to be processed and presented by netCTLpan, thus meeting all our criteria for likelihood of being natural T-cell epitopes (figure 1A). We focused on these two candidate neoantigens from U2AF1Q157R as most suitable for T-cell recognition. The difference in predicted binding affinity of the same U2AF1Q157R peptide for HLA-A*33:01 and HLA-A*33:03 was only threefold (18 and 51 nM, respectively; online supplemental table 1), making these equally compelling for further study.

    Figure 1
    Figure 1

    Two candidate neoantigen epitopes are predicted from recurrent myelodysplastic syndromes-associated and secondary acute myeloid leukemia-associated U2AF1 mutations. (A) Schematic representation of the workflow used for in silico identification of candidate neoantigens. wt, wild-type. (B) Predictions were made using IEDB Stabilized Matrix Method and Artificial Neural Network methods, netMHCpan 4.1, and netCTLpan and summarized graphically. Predicted HLA binders are shown in green. Predicted processed and presented peptides are shown in blue. Epitopes for which the corresponding wild-type peptide is not predicted to bind HLA are in bright green and denoted with an asterisk. Candidate epitopes selected for further study (defined as having predicted IC50 was <250 nM or rank <1 by at least two methods, processing/prediction rank <1%, and no predicted HLA binding of the equivalent wild-type peptide) are shown in bright blue with double asterisk. HLA, human leukocyte antigen; IEDB, Immune Epitope Database.

    U2AF1Q157R neoantigen epitope is immunogenic and processed and presented from endogenous protein

    Algorithms predicting HLA binding and processing and presentation identify peptides that are likely to be relevant epitopes, but not all predicted epitopes will be processed and presented and thus require validation.57 58 Moreover, isolation of T cells specific for candidate neoantigens enables confirmation that the antigen is adequately processed and presented on the surface of malignant cells to induce T cell-mediated killing and thus represents a bona fide cancer neoantigen. Therefore, we directly tested the immunogenicity of the most promising candidate U2AF1Q157R neoantigens by stimulating CD8+ T cells with candidate neoantigen peptide.

    Primary in vitro stimulation of CD8+ T cells from two unique healthy donors, one HLA-A*33:03-positive and one HLA-A*33:01-positive, using autologous monocyte-derived DCs yielded CD8+ T cells specific for the peptide DFREACCRR, which is derived from the U2AF1Q157R mutation (figure 2A). In 4-hour CRA, one DFREACCRR-specific clone from the HLA-A*33:03-positive donor (D1.C32) and one clone from the A*33:01-positive donor (D2.C177) both killed peptide-pulsed autologous LCL targets proportionally to the concentration of peptide (figure 2B,C, respectively). We confirmed the restrictions of the clones to HLA-A*33:01 and HLA-A*33:03, respectively, by testing against panels of HLA-typed LCL (figure 2D,E, respectively). For both clones, only DFREACCRR peptide-pulsed LCL bearing HLA-A*33:01 or HLA-A*33:03 were recognized, while peptide-pulsed HLA-A*33:01/03-negative LCL bearing other HLA shared by the original T-cell donor were not recognized. We then asked whether T cells specific for U2AF1Q157R/A*33:01 could recognize U2AF1Q157R/A*33:03 and vice versa, since cross-recognition of peptide on similar HLA could broaden the applicability of a TCR. The U2AF1Q157R/A*33:03-specific clone lysed HLA-A*33:01-positive LCL pulsed with DFREACCRR peptide (figure 2F), and the U2AF1Q157R/A*33:01-specific T-cell clone lysed DFREACCRR-pulsed HLA-A*33:03-positive LCL (figure 2G), indicating that both clones could cross-recognize peptide presented on either HLA-A*33:03 or HLA-A*33:01. We then tested whether clones could kill the naturally HLA-A*33:03 positive AML cell line TF-1 transduced with a minigene encoding U2AF1Q157R or the equivalent wild-type sequence from U2AF1 (online supplemental figure 1). The HLA-A*33:03-restricted clone (D1.C32) eliminated U2AF1Q157R minigene-transduced TF-1 cells, but not control TF-1 cells transduced with the wild-type equivalent U2AF1 minigene, by 48 hours in a flow cytometry-based killing assay (figure 2H, gating strategy online supplemental figure 2). Similarly, HLA-A*33:01-restricted clone (D2.C177) eradicated U2AF1Q157R minigene-transduced TF-1 cells by 24 hours in this assay. These data indicate that the DFREACCRR epitope is processed and presented from endogenous protein.

    Figure 2
    Figure 2

    A U2AF1Q157R HLA-A*33-restricted neoantigen epitope is immunogenic and primes epitope-specific CD8+ T-cell clones. (A) Schematic of U2AF1Q157R protein and neoantigen epitope peptide. (B and C) One clone from an HLA-A*33:03-positive donor (B D1.C32) and one clone from an HLA-A*33:01-positive donor (C D2.C177) were each identified after primary in vitro stimulation of CD8+ T cells with DFREACCRR peptide, then tested for functional avidity in peptide titration CRA using autologous LCL pulsed with varying peptide concentrations (three technical replicate experiments). (D and E) HLA restriction of clone D1.C32 from HLA-A*33:03+ donor (D) and clone D2.C177 from HLA-A*33:01+ donor (E) were confirmed by testing in CRA against a panel of HLA-typed LCL with single HLA overlap with the original T-cell donor. LCL were pulsed with DFREACCRR peptide at 1000 ng/mL prior to co-culture. (F) Clone D1.C32 lyses DFREACCRR peptide-pulsed (1000 ng/mL) HLA-A*33:01+ as well as A*33:03+ LCL. (G) Clone D2.C177 lyses HLA-A*33:03+ as well as A*33:01+ LCL pulsed with 1000 ng/mL DFREACCRR peptide. (H and I) Per cent survival of TF-1 minigene-transduced cells co-cultured with either D1.C32 (H red), D2.C177 (I blue) or irrelevant neoantigen-specific clone (gray) in flow cytometry-based cytotoxicity assay. For (H and I) top plots, U2AF1Q157R minigene; bottom plots, U2AF1 minigene. Mean and SEM from >3 technical replicates shown. CRA, Cr-release cytotoxicity assays; HLA, human leukocyte antigen; LCL, lymphoblastoid cell lines.

    U2AF1Q157R -specific T-cell clones do not recognize similar peptides

    To assess for potential off-target recognition of non-U2AF1Q157R peptides by the DFREACCRR-specific clones, we first performed alanine scanning to determine the critical residues for the two clones and then performed an in silico search for proteins containing the critical residue motif for each clone motif using the ScanProsite tool.59 60 The critical residue motif was DxxExCCxR for the HLA-A*33:03-restricted clone (figure 3A). Two human peptides sharing the critical residue motifs (BTG397–105 DpcEvCCrR; TTLL4517–525 DglEdCCsR) were identified using ScanProsite, but the clone did not recognize either one of these similar peptides nor the wild-type U2AF1 peptide (DfrEaCCrQ) when presented by peptide-pulsed autologous LCL (figure 3B). The HLA-A*33:01-restricted clone was more permissive (xxxExCCxR) than the HLA-A*33:03-restricted clone (figure 3C) and 30 human peptides sharing the critical residue motif were identified by ScanProsite (online supplemental table 3). Of these, seven were predicted to bind well to HLA-A*33:01 (predicted affinity <500 nM and % rank <2) and thus could potentially be cross-recognized by the HLA-A*33:01-restricted clone (online supplemental table 3). However, the clone did not recognize autologous LCL pulsed with any of the seven similar peptides nor the wild-type U2AF1 peptide in CRA (figure 3D). These results suggest that both the HLA-A*33:03 and HLA-A*33:01 restricted clones and their TCR have low potential for off-target recognition of similar peptides.

    Figure 3
    Figure 3

    U2AF1Q157R/A*33-specific clones do not recognize similar peptides. (A) Alanine scanning for U2AF1Q157R/A*33:03-specific clone D1.C32 was performed using autologous LCL pulsed with a panel of peptides (1000 ng/mL) with alanine residues substituted at each position, along with two peptides with either a glycine or valine substitution at position five, a natural alanine residue in the DFREACCRR peptide (three technical replicate experiments). These data were used to identify critical residues for HLA and TCR binding and define the core motifs DxxExCCRR. (B) Peptides derived from wild-type human proteins and sharing the core motifs for the HLA-A*33:03-restricted clones clone were identified using the ScanProsite tool. To evaluate for cross-recognition of these peptides by clones, autologous LCL were pulsed with each peptide (1000 ng/mL) and used as targets for clones in CRA (three technical replicate experiments). (C) Alanine scanning for U2AF1Q157R/A*33:01-specific clone D2.C177 identified critical residues for HLA and TCR binding identified, defining the core motif xxxExCCxR for this clone. (D) Human peptides sharing the core motif were identified as using ScanProsite and clone D2.C177 tested for their recognition in CRA using autologous LCL pulsed with each peptide (1000 ng/mL) as targets. For all experiments, mean and SEM are shown. CRA, Cr-release cytotoxicity assays; HLA, human leukocyte antigen; LCL, lymphoblastoid cell lines; TCR, T-cell receptor.

    U2AF1Q157R -specific TCR can be transferred and confer specificity and leukemia killing

    Immunotherapies using adoptive transfer of T cells genetically engineered to express antigen-specific TCR can overcome numerical or functional defects in natural T-cell immunity against malignant cells. Having discovered U2AF1Q157R neoantigens, we next assessed the feasibility of transferring U2AF1Q157R-specific TCR as a step towards clinical translation. We first sequenced the TCRα and β chains from the U2AF1Q157R/A*33:01-specific and U2AF1Q157R/A*33:03-specific T-cell clones (online supplemental table 4). To generate transgenic TCR, we then paired the TCRα and β chains from each clone in LVs. We successfully cloned TCR from each of the two U2AF1Q157R neoantigen-specific clones into LV and transduced them separately into healthy third-party donor CD8+ T cells after knocking out the endogenous TCR.

    Third-party primary human CD8+ T cells transduced with U2AF1Q157R/A*33:03-specific or U2AF1Q157R/A*33:01-specific TCR expressed the transduction marker RQR855 and TCR, shown by staining with pHLA tetramer (figure 4A,B, respectively). Transduced CD8+ T cells lysed target cells pulsed with DFREACCRR peptide with functional avidity that was equivalent to that of parental clones (figure 4C,D, respectively). Similar to the parental clones, CD8+ T cells transduced with either the U2AF1Q157R/A*33:01-specific or U2AF1Q157R/A*33:03-specific TCR also killed the naturally HLA-A*33:03-positive TF-1 cell line expressing U2AF1Q157R but did not kill a wild-type U2AF1 control, with complete elimination occurring by or before 48 hours, (figure 4E,F). U2AF1Q157R/A*33-specific TCR-transduced T cells eliminated primary neoplastic myeloid cells bearing both the U2AF1Q157R mutation and HLA-A*33:01, but did not kill control HLA-A*33+ primary neoplastic myeloid cells lacking the mutation within 48 hours of co-culture (figure 4G, online supplemental figure 3A). Of note, due to very limited cell numbers in the original primary bone marrow sample, the U2AF1Q157R-positive HLA-A*33:01-positive sAML was expanded in humanized immunodeficient mouse model known to support normal and neoplastic myelopoiesis.61 62 The immunophenotype of human hematopoietic cells retrieved and purified from xenografts was similar to that reported by clinical assays on the patient’s blasts (online supplemental figure 3B,C). Importantly, U2AF1Q157R/A*33-specific TCR-transduced T cells showed minimal to no recognition of either primitive (CD34-enriched filgrastim-mobilized peripheral blood stem cells) or differentiated (PBMC) normal hematopoietic cells from a healthy HLA-A*33:01+ individual in a CD107a degranulation assay (online supplemental figure 4). These data indicate that transfer of U2AF1Q157R/A*33-specific TCR to third-party T cells is feasible and enables U2AF1Q157R neoantigen-specific killing primary neoplastic cells without off-target recognition of normal hematopoietic cells in vitro.

    Figure 4
    Figure 4

    Transfer of U2AF1Q157R neoantigen-specific TCR confers specificity and function. (A) Representative flow plots demonstrating expression of D1.C32 U2AF1Q157R/A*33:03-specific TCR (TCR32) transduced (TD) into primary human CD8+ T cells after CRISPR/Cas9-mediated knock-out of endogenous TCR alpha and beta chains (TCRko). Staining for the RQR8 transduction marker (CD34), left, and U2AF1Q157R/A*33:03-pHLA tetramer, right. (B) Representative flow plots demonstrating expression of clone D2.C177 U2AF1Q157R/A*33:01-specific TCR (TCR177) TD into TCRko primary human CD8+ T cells. CD34 staining, left, and U2AF1Q157R/A*33:03-pHLA tetramer, right. For A and B top, untransduced (UTD); middle, TCR TD; bottom, parental clone. (C) Testing of TCR32 TD T cells (dark red, solid) with parental clone control (light red, dashed) and (D) testing of TCR177 TD T cells (dark purple, solid) and parental clone control (light purple, dashed), for specific lytic activity in peptide titration CRA. For both C and D technical triplicates shown. (E) Per cent survival of TF-1 minigene-transduced cells co-cultured with TCR32 TD T cells (dark red) or parental clone control (light red) in flow cytometry-based cytotoxicity assay. (F) Per cent survival of TF-1 minigene-transduced cells co-cultured with TCR177 TD T cells (dark purple) or parental clone control (light purple) in flow cytometry-based cytotoxicity assay. For both E and F >3 technical replicates; top plot, U2AF1Q157R minigene TD TF-1; bottom plots, wild-type (WT) U2AF1 minigene TD TF-1. (G) Per cent survival of U2AF1Q157R/ HLA-A*33:01+ primary AML (left) or wild-type U2AF1 (Q157R-)/HLA-A*33:03+ primary AML (right) co-cultured with either TCR32 TD T cells (red) or TCR177 TD T cells (purple) in flow cytometry-based cytotoxicity assay. Data from >3 technical replicates. Mean and SEM are shown. AML, acute myeloid leukemia; HLA, human leukocyte antigen; pHLA, peptide-HLA; TCR, T-cell receptor.

    Spliceosome neoantigen-specific T cells control myeloid neoplasia in vivo

    We next evaluated the efficacy of U2AF1Q157R neoantigen-specific TCR-T cells in vivo in a cell line-derived xenograft (CDX) model (figure 5A). Immunodeficient NSG-SGM3 mice were engrafted with luciferase-expressing HLA-A*33:03+ TF-1 cell lines transduced with either a minigene encoding the mutated region of U2AF1Q157R or the equivalent wild-type amino acid sequence. Five days after intravenous injection of 5×105 TF-1 cells co-transduced with U2AF1Q157R minigene and luciferase, mice in the experimental group were treated with a single intravenous dose of 1×107 CD8+ T cells transduced with U2AF1Q157R/A*33:03-specific TCR32 (n=7). Control animals engrafted with U2AF1Q157R TF-1 cells were treated either with 1×107 CD8+ T cells transduced with an irrelevant neoantigen-specific control TCR (n=7) or with vehicle (n=7). An additional group of animals were engrafted with 5×105 TF-1 cells co-transduced with wild-type U2AF1 minigene and treated with TCR32 transduced T cells (n=7). This CDX allowed us to model an aggressive and proliferative myeloid neoplasm against which to challenge our TCR-T cells; within 38 days of TF-1 injection, all vehicle-treated animals had become moribund.

    Figure 5
    Figure 5

    U2AF1Q157 neoantigen-specific TCR-T cells control myeloid neoplasms in vivo. (A) Experimental schematic. After sublethal irradiation, immunodeficient NSG-SGM3 mice were injected intravenous with naturally HLA-A*33:03-positive TF-1 cells transduced with a U2AF1Q157 minigene, allowed to engraft for 5 days, then injected with either U2AF1Q157R/A*33:03-specific TCR-T cells (TCR32), irrelevant neoantigen-specific control TCR-T cells, or vehicle. A fourth group was engrafted with TF-1 cells transduced with a wild-type U2AF1 minigene, then treated with U2AF1Q157R/A*33:03-specific TCR-T cells (n=7 per group). (B) Serial bioluminescence imaging from mice engrafted with U2AF1 minigene-transduced TF-1 cells before and at time points after T-cell injection. (C) Quantitation of bioluminescence flux (p/s) in animals in each group at serial time points. Pairwise comparisons (t-tests) were performed on the ranked data between groups. (D) Kaplan-Meier survival curve (Mantel-Cox log-rank test). (E) Summary of minigene expression in retrieved TF-1 cells in terminal bone marrow as indicated by total absolute numbers of CD20 staining in treatment groups. (Mann-Whitney test, **, p value<0.005; ***, p value<0.001). In B–E animals engrafted with U2AF1Q157R minigene-transduced TF-1 cells and treated with TCR32 T cells, solid red; engrafted with U2AF1Q157R TD TF-1 and treated with control TCR-T cells, gray; engrafted with U2AF1Q157R TD TF-1 and treated with vehicle, black; engrafted with wild-type U2AF1 minigene-transduced TF-1 cells and treated with U2AF1Q157R/A*33:03-specific TCR-T cells, dashed/open red. TCR-T, T-cell receptor-transduced T cell; WT, wild-type.

    In contrast, an anti-leukemic effect was seen in mice in the experimental group that were engrafted with U2AF1Q157R-expressing TF-1 cells and treated with U2AF1Q157R/A*33:03-specific TCR32 T cells. In all time points after week 1, disease burden as measured by bioluminescence was consistently lower in experimental animals (≥102 lower flux, p/s) compared with bioluminescence in all three control groups (figure 5B,C). Correspondingly, survival was significantly prolonged in animals in the experimental group compared with those in each of the three control groups (figure 5D). Animals in the experimental group were euthanized due to >20% wt loss. Animals in the control groups were either found dead (n=1), euthanized due to disease-related symptoms (hind-limb paralysis or hunching, n=10; due to weight loss, n=4), or euthanized due to saturation of bioluminescence consistent with imminently lethal disease burden (n=13). Taken together, these results indicate that T cells bearing TCR32 controlled U2AF1Q157R-expressing neoplastic cells in vivo.

    While U2AF1Q157R/A*33-specific TCR-T cells showed activity in vivo, they were apparently unable to fully clear disease. To understand why disease clearance was incomplete, we first asked whether TCR-T cells were capable of in vivo persistence. U2AF1Q157R/A*33-specific TCR32 and control TCR-T cells both decreased in the peripheral blood over time and could be detected at similar percentages in the terminal bone marrow of animals in the experimental and control groups other than vehicle controls (online supplemental figure 5A–C). We next examined TF-1 cells in the marrow at the time of death to determine whether disease persistence could have been due to antigen loss. Although TF-1 cells persisted in animals in the experimental group, the TF-1 cells had lost expression of the U2AF1Q157R-encoding minigene. There was a significant decrease in TF-1 cells expressing the transduction marker (based on CD20 staining) linked to the minigene in the bone marrow of animals from the experimental group compared with controls (figure 5E, online supplemental figure 6A–C). In three of seven experimental group animals, bone marrow TF-1 cells were undetectable. Loss of either U2AF1 minigene by TF-1 cells in terminal marrow of control animals was not observed. Additionally, one of the longest-surviving animals in the experimental group was noted to have an abdominal focus of luminescence on imaging, correlating with 1.5 cm mass identified in the mesentery (online supplemental figure 6D). Flow cytometry on cells dissociated from the mass revealed it to be composed predominantly of human CD45+ GFP+ TF-1 cells with absent CD20 staining (online supplemental figure 6E), indicating complete loss of expression of the minigene, and therefore loss of the antigen.

    TF-1 cells have a rapid doubling time that could have permitted even the small proportion of untransduced TF-1 cells left after stringent pre-injection sorting to grow out under selective pressure from TCR-T cells. These results suggest that incomplete disease clearance in animals with U2AF1Q157R-expressing TF-1 cells and treated with U2AF1Q157R/A*33:03-specific TCR-T cells was due to antigen loss in this cell line model system and not T-cell failure, and in fact indicates that U2AF1Q157R/A*33-specific TCR-T cell were highly effective in eradicating U2AF1Q157R-expressing cells. Overall, these results indicate that U2AF1Q157R/A*33-specific TCR-T cells provide effective antigen-specific control of an aggressive myeloid neoplasm in vivo.

    Discussion/conclusions

    Few targeted therapies exist for MDS and sAML, and potentially curative HCT is not available to all patients. T-cell therapies targeting neoantigens could be safe and effective, even in HCT-ineligible patients, since residual normal hematopoiesis should be preserved. We hypothesized that mutations in spliceosome genes create neoantigens that would serve as targets for T-cell immunotherapy for MDS and sAML, and in this study, we identify neoantigens created from a recurrent mutation in U2AF1. We used in silico predictions to identify candidate U2AF1 mutation-derived neoantigen epitopes followed by in vitro immunogenicity screening to determine which candidates were recognized by CD8+ T cells. We identified a promising U2AF1Q157R peptide epitope, and confirmed that the peptide was naturally processed from endogenous protein and was presented on the expected restricting HLA-A*33 molecules on target cells. TCR-T cells specific for U2AF1Q157R/A*33 lysed neoplastic myeloid cells in vitro without off-target recognition of normal hematopoietic cells including HSPC, and controlled disease in vivo. Our findings indicate that U2AF1Q157R/A*33 is a bona fide MDS/sAML neoantigen and a promising precision medicine target for TCR-T cell therapies to treat a subset of patients with MDS and sAML, as well as myeloproliferative neoplasms and other myeloid malignancies.

    Although neoantigens can theoretically be created from any protein-coding mutation, aberrant peptides will not be processed from the protein product of every protein-coding mutation, not every aberrant peptide, that is, processed will be presented by all or even some HLA molecules, and not every aberrant peptide-HLA complex will be immunogenic. Predicting which mutations are likely to lead to immunogenic, naturally processed, and HLA-presented epitopes streamlines the process of neoantigen discovery and immunotherapy. We refined our in silico method to incorporate predictions of wild-type and aberrant peptide binding to HLA as well as peptide processing and transport. The candidate neoantigens we identified with this method proved to be immunogenic as well as naturally processed and presented on neoplastic cells with the appropriate mutation and HLA type. Using this multistep approach could facilitate the discovery of neoantigens and other types of antigens for downstream immunotherapy development.

    To our knowledge, this study is the first to describe neoantigens created directly from the protein products of a recurrent mutation in U2AF1. The U2AF1Q157R neoantigens are distinct from those that have been identified for de novo AML resulting from a recurrent NPM1 mutation63 or a leukemia-initiating CBFB-MYH11 gene fusion,32 neither of which are found with any frequency in MDS. Other shared neoantigens with relevance to patients with MDS and sAML have been identified, including those created from recurrent mutations in KRAS, the peptide sequences of which overlap with other RAS proteins including NRAS, and from mutations in TP53. Mutations in Ras family and other signaling genes and TP53 occur as secondary events in subclones and are linked to sAML transformation and poor outcomes.64 65 Neoantigens resulting from KRAS and TP53 mutations that have been found in cancers broadly28 also should occur in some cases of MDS and sAML. However, T-cell therapies targeting KRAS and TP53 mutation-derived neoantigens in high-risk MDS and sAML would target later subclones rather than early neoplastic clones and thus could debulk disease but would not be expected to be curative. Nevertheless, in a patient with suitable mutations and HLA type, therapeutic T cells targeting different antigens could be used in series, for example, using TP53-directed or Ras-directed T cells to control an aggressive clone and a U2AF1-directed clone for relapse prevention, or potentially even in combination. Moreover, because of HLA restriction and the heterogeneity of MDS and sAML, any TCR-T cell therapy directed at a single target is semi-personalized by nature, so multiple antigen targets will be needed to have TCR-T cell therapies for most patients.

    We also describe the first example of neoantigen-specific TCR-T cells for MDS. Other groups have developed neoantigen-directed T-cell immunotherapy for MDS that uses ex vivo autologous T cells expanded with peptides based on patient-specific mutations66 and that is currently being investigated in a clinical trial (NCT03258359). This approach relies on the ability to expand functional neoantigen-specific T-cell responses from patient T cells, which may be challenging in the immune dysregulated microenvironment of MDS.67 Preliminary data suggest that while manufacturing of such ex vivo-expanded T-cell products is feasible and administration is safe, the products are heterogeneous and persistence is limited.66 68 In contrast, our approach of using TCR transfer would enable patients to receive a product with a defined specificity and potency. TCR constructs can be modified further to improve the safety and efficacy of the transduced T cells by including CD8 co-receptors,54 safety switches,54 55 immunomodulatory fusion proteins,69 cytokine genes,70–72 or even a combination of modifications. Thus, while TCR-T cell therapy is technically relatively complex compared with ex vivo-expanded T cells, T-cell engineering offers multiple advantages over neoantigen-directed T cell lines for patients with MDS and sAML. Our results provide key foundational data for developing neoantigen-specific TCR-T cell therapies for myeloid neoplasms.

    Our findings provide proof-of-principle that neoantigens are relevant targets in malignancies like MDS and sAML that have relatively few total mutations in comparison to solid tumors.8 73–77 The percentage of individuals that express HLA-A*33:01 or HLA-A*33:03 worldwide are 2% and 3.2%, respectively, with higher prevalence of up to 6% and 22%, respectively, in certain populations, and mutations leading to U2AF1Q157R occur in around 2% of MDS and sAML,8 9 17 74 78 as well as uncommonly in certain solid tumors.79 While the specific neoantigens identified have narrow applicability based on their expected prevalence, TCR targeting these semi-personal antigens are a crucial part of a larger precision medicine “toolbox” containing multiple MDS-specific and sAML-specific antigens with diverse HLA restrictions that, together, will be broadly applicable to patients. In particular, because the U2AF1Q157R/A*33 neoantigen and similar neoantigens will be presented solely by aberrant cells bearing the mutation, not normal cells without the mutation, neoantigen-directed therapies should eliminate neoplastic cells while sparing non-malignant cells including HSPC that lack the mutation, as our in vitro data suggests. T cells targeting U2AF1Q157R and other MDS neoantigens thus warrant investigation especially for use in situations where myeloablation must be avoided (ie, in HCT-ineligible patients) and for whom potentially curative therapies currently do not exist.

    Data availability statement

    Data are available upon reasonable request.

    Ethics statements

    Patient consent for publication

    Ethics approval

    Blood and bone marrow samples from healthy volunteer donors and patients with MDS and AML were obtained after written informed consent in accordance with the Declaration of Helsinki to participate in research protocols approved by the Institutional Review Board of the Fred Hutchinson Cancer Center (FHCC). Samples from healthy donors for immunogenicity screening assays were collected on FHCC protocol 985 and 2684, and healthy donor samples for third-party T cells for TCR transduction on protocol 985. CD34-enriched cells from GCSF-mobilized peripheral blood stem cell products were obtained on FHCC protocol 2684. Samples from patients were obtained through the FHCC/University of Washington Hematopoietic Diseases Repository (protocol #1690), FHCC MDS Repository (protocol #1713), FHCC protocol 956, and FHCC protocol 2684. Animal experiments were approved by Institutional Animal Care and Use Committee of the FHCC (protocol #51082 for cell-line derived xenografts and #50941 for M-CSFh/h IL-3/GM-CSFh/h SIRPAh/h TPOh/h RAG2−/− IL2RG−/− (MISTRG) patient-derived xenografts).

    Acknowledgments

    We thank the patients and healthy donors who generously provided the samples used in these studies. We thank Yvonne Hsu, Dr Ana Dios-Esponera, Tanya Cunningham, Jacqueline Diaz, Michael Coon, Dr Jon Linton, and Jenny Harrington-Lill for their technical assistance, and Heather Persinger, Barbara Hilzinger RN, and Taylor Jones for sample collection support. We thank Dr Ted Gooley for his assistance with statistical analyses and Drs Kate Markey and Michelle Brault for their insightful comments on the manuscript. We are grateful to the FHCC Flow Cytometry Core Facility staff, especially Dr Michelle Black, Rebecca Reeves, and Ben Janoschek, for their help in panel development and tireless technical support. We appreciate FHCC Comparative Medicine for their outstanding mouse husbandry. We acknowledge the FHCC/University of Washington Hematopoietic Diseases Repository and FHCC MDS Repository for access to critical patient samples. We also acknowledge Yale University, the University of Zürich, and Regeneron Pharmaceuticals where MISTRG mice were generated with financial support from the Bill and Melinda Gates Foundation. Figures were made in BioRender (BioRender.com).

    References

    1. ↵
      1. Ogawa S
      . Genetics of MDS. Blood 2019;133:1049–59. doi:10.1182/blood-2018-10-844621
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Sperling AS,
      2. Gibson CJ,
      3. Ebert BL
      . The Genetics of myelodysplastic syndrome: from clonal haematopoiesis to secondary leukaemia. Nat Rev Cancer 2017;17:5–19. doi:10.1038/nrc.2016.112
      OpenUrlCrossRefPubMed
    3. ↵
      1. Estey EH
      . Acute myeloid leukemia: 2021 update on risk-stratification and management. Am J Hematol 2020;95:1368–98. doi:10.1002/ajh.25975
      OpenUrlCrossRefPubMed
    4. ↵
      1. Kolb HJ
      . Graft-versus-leukemia effects of transplantation and donor lymphocytes. Blood 2008;112:4371–83. doi:10.1182/blood-2008-03-077974
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Biernacki MA,
      2. Brault M,
      3. Bleakley M
      . T-cell receptor-based immunotherapy for hematologic malignancies. Cancer J 2019;25:179–90. doi:10.1097/PPO.0000000000000378
      OpenUrlCrossRef
    6. ↵
      1. Biernacki MA,
      2. Sheth VS,
      3. Bleakley M
      . T cell optimization for graft-versus-leukemia responses. JCI Insight 2020;5:e134939. doi:10.1172/jci.insight.134939
    7. ↵
      1. Chapuis AG,
      2. Egan DN,
      3. Bar M, et al
      . T cell receptor gene therapy targeting WT1 prevents acute myeloid leukemia relapse post-transplant. Nat Med 2019;25:1064–72. doi:10.1038/s41591-019-0472-9
      OpenUrlCrossRefPubMed
    8. ↵
      1. Papaemmanuil E,
      2. Gerstung M,
      3. Malcovati L, et al
      . Clinical and biological implications of driver mutations in myelodysplastic syndromes. Blood 2013;122:3616–27. doi:10.1182/blood-2013-08-518886
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Yoshida K,
      2. Sanada M,
      3. Shiraishi Y, et al
      . Frequent pathway mutations of splicing machinery in myelodysplasia. Nature 2011;478:64–9. doi:10.1038/nature10496
      OpenUrlCrossRefPubMedWeb of Science
    10. ↵
      1. Reinig E,
      2. Yang F,
      3. Traer E, et al
      . Targeted next-generation sequencing in myelodysplastic syndrome and chronic myelomonocytic leukemia AIDS diagnosis in challenging cases and identifies frequent spliceosome mutations in transformed acute myeloid leukemia. Am J Clin Pathol 2016;145:497–506. doi:10.1093/ajcp/aqw016
      OpenUrlCrossRefPubMed
    11. ↵
      1. Hershberger CE,
      2. Moyer DC,
      3. Adema V, et al
      . Complex landscape of alternative splicing in myeloid neoplasms. Leukemia 2021;35:1108–20. doi:10.1038/s41375-021-01197-2
      OpenUrlCrossRef
    12. ↵
      1. Lachowiez CA,
      2. Loghavi S,
      3. Furudate K, et al
      . Impact of splicing mutations in acute myeloid leukemia treated with hypomethylating agents combined with venetoclax. Blood Adv 2021;5:2173–83. doi:10.1182/bloodadvances.2020004173
      OpenUrl
    13. ↵
      1. Graubert TA,
      2. Shen D,
      3. Ding L, et al
      . Recurrent mutations in the U2AF1 splicing factor in myelodysplastic syndromes. Nat Genet 2011;44:53–7. doi:10.1038/ng.1031
      OpenUrlCrossRefPubMed
    14. ↵
      1. Thol F,
      2. Kade S,
      3. Schlarmann C, et al
      . Frequency and prognostic impact of mutations in SRSF2, U2AF1, and ZRSR2 in patients with myelodysplastic syndromes. Blood 2012;119:3578–84. doi:10.1182/blood-2011-12-399337
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Patnaik MM,
      2. Lasho TL
      . Myelodysplastic syndrome/myeloproliferative neoplasm overlap syndromes: a focused review. Hematology Am Soc Hematol Educ Program 2020;2020:450–9. doi:10.1182/hematology.2020000130
      OpenUrl
    16. ↵
      1. Palomo L,
      2. Meggendorfer M,
      3. Hutter S, et al
      . Molecular landscape and clonal architecture of adult myelodysplastic/myeloproliferative neoplasms. Blood 2020;136:1851–62. doi:10.1182/blood.2019004229
      OpenUrl
    17. ↵
      1. Shirai CL,
      2. Ley JN,
      3. White BS, et al
      . Mutant U2AF1 expression alters hematopoiesis and pre-mRNA splicing in vivo. Cancer Cell 2015;27:631–43. doi:10.1016/j.ccell.2015.04.008
      OpenUrlCrossRefPubMed
    18. ↵
      1. Fei DL,
      2. Zhen T,
      3. Durham B, et al
      . Impaired Hematopoiesis and leukemia development in mice with a conditional knock-in allele of a mutant splicing factor gene U2Af1 Proc Natl Acad Sci U S A 2018;115:E10437–46. doi:10.1073/pnas.1812669115
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Nagata Y,
      2. Makishima H,
      3. Kerr CM, et al
      . Invariant patterns of clonal succession determine specific clinical features of myelodysplastic syndromes. Nat Commun 2019;10:5386. doi:10.1038/s41467-019-13001-y
    20. ↵
      1. Pellagatti A,
      2. Armstrong RN,
      3. Steeples V, et al
      . Impact of Spliceosome mutations on RNA splicing in myelodysplasia: dysregulated genes/pathways and clinical associations. Blood 2018;132:1225–40. doi:10.1182/blood-2018-04-843771
      OpenUrlAbstract/FREE Full Text
    21. ↵
      1. Yip BH,
      2. Steeples V,
      3. Repapi E, et al
      . The U2Af1S34F mutation induces lineage-specific splicing alterations in myelodysplastic syndromes. J Clin Invest 2017;127:96202. doi:10.1172/JCI96202
    22. ↵
      1. Jaiswal S,
      2. Fontanillas P,
      3. Flannick J, et al
      . Age-related clonal hematopoiesis associated with adverse outcomes. N Engl J Med 2014;371:2488–98. doi:10.1056/NEJMoa1408617
      OpenUrlCrossRefPubMed
    23. ↵
      1. Genovese G,
      2. Kähler AK,
      3. Handsaker RE, et al
      . Clonal hematopoiesis and blood-cancer risk inferred from blood DNA sequence. N Engl J Med 2014;371:2477–87. doi:10.1056/NEJMoa1409405
      OpenUrlCrossRefPubMed
    24. ↵
      1. Chen S,
      2. Benbarche S,
      3. Abdel-Wahab O
      . Splicing factor mutations in hematologic malignancies. Blood 2021;138:599–612. doi:10.1182/blood.2019004260
      OpenUrlCrossRef
    25. ↵
      1. Abelson S,
      2. Collord G,
      3. Ng SWK, et al
      . Prediction of acute myeloid leukaemia risk in healthy individuals. Nature 2018;559:400–4. doi:10.1038/s41586-018-0317-6
      OpenUrl
    26. ↵
      1. Desai P,
      2. Mencia-Trinchant N,
      3. Savenkov O, et al
      . Somatic mutations precede acute myeloid leukemia years before diagnosis. Nat Med 2018;24:1015–23. doi:10.1038/s41591-018-0081-z
      OpenUrlCrossRefPubMed
    27. ↵
      1. Pritzl SL,
      2. Gurney M,
      3. Badar T, et al
      . Clinical and molecular spectrum and prognostic outcomes of U2Af1 mutant clonal hematopoiesis- a prospective Mayo clinic cohort study. Leuk Res 2023;125:107007. doi:10.1016/j.leukres.2022.107007
      OpenUrl
    28. ↵
      1. Biernacki MA,
      2. Bleakley M
      . Neoantigens in hematologic malignancies. Front Immunol 2020;11:121. doi:10.3389/fimmu.2020.00121
    29. ↵
      1. McGranahan N,
      2. Swanton C
      . Neoantigen quality, not quantity. Sci Transl Med 2019;11:eaax7918. doi:10.1126/scitranslmed.aax7918
    30. ↵
      1. Yarchoan M,
      2. Johnson BA,
      3. Lutz ER, et al
      . Targeting neoantigens to augment antitumour immunity. Nat Rev Cancer 2017;17:569. doi:10.1038/nrc.2017.74
    31. ↵
      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
    32. ↵
      1. Biernacki MA,
      2. Foster KA,
      3. Woodward KB, et al
      . CBFB-MYH11 fusion neoantigen enables T cell recognition and killing of acute myeloid leukemia. J Clin Invest 2020;130:5127–41. doi:10.1172/JCI137723
      OpenUrlCrossRef
    33. ↵
      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
    34. ↵
      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
    35. ↵
      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
    36. ↵
      1. Keskin DB,
      2. Anandappa AJ,
      3. Sun J, et al
      . Neoantigen vaccine generates Intratumoral T cell responses in phase IB glioblastoma trial. Nature 2019;565:234–9. doi:10.1038/s41586-018-0792-9
      OpenUrlPubMed
    37. ↵
      1. Ott PA,
      2. Hu Z,
      3. Keskin DB, et al
      . An Immunogenic personal neoantigen vaccine for patients with Melanoma. Nature 2017;547:217–21. doi:10.1038/nature22991
      OpenUrlCrossRefPubMed
    38. ↵
      1. Sahin U,
      2. Derhovanessian E,
      3. Miller M, et al
      . Personalized RNA mutanome vaccines mobilize poly-specific therapeutic immunity against cancer. Nature 2017;547:222–6. doi:10.1038/nature23003
      OpenUrlCrossRefPubMed
    39. ↵
      1. Cimen Bozkus C,
      2. Roudko V,
      3. Finnigan JP, et al
      . Immune checkpoint blockade enhances shared neoantigen-induced T-cell immunity directed against mutated calreticulin in myeloproliferative neoplasms. Cancer Discov 2019;9:1192–207. doi:10.1158/2159-8290.CD-18-1356
      OpenUrlAbstract/FREE Full Text
    40. ↵
      1. Bersanelli M,
      2. Travaglino E,
      3. Meggendorfer M, et al
      . Classification and personalized prognostic assessment on the basis of clinical and genomic features in myelodysplastic syndromes. J Clin Oncol 2021;39:1223–33. doi:10.1200/JCO.20.01659
      OpenUrl
    41. ↵
      1. Hamilton BK,
      2. Visconte V,
      3. Jia X, et al
      . Impact of allogeneic hematopoietic cell transplant in patients with myeloid neoplasms carrying Spliceosomal mutations. Am J Hematol 2016;91:406–9. doi:10.1002/ajh.24306
      OpenUrl
    42. ↵
      1. Makishima H,
      2. Visconte V,
      3. Sakaguchi H, et al
      . Mutations in the spliceosome machinery, a novel and ubiquitous pathway in leukemogenesis. Blood 2012;119:3203–10. doi:10.1182/blood-2011-12-399774
      OpenUrlAbstract/FREE Full Text
    43. ↵
      1. Kim Y,
      2. Ponomarenko J,
      3. Zhu Z, et al
      . Immune EPITOPE database analysis resource. Nucleic Acids Research 2012;40:W525–30. doi:10.1093/nar/gks438
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Nielsen M,
      2. Lundegaard C,
      3. Worning P, et al
      . Reliable prediction of T-cell epitopes using neural networks with novel sequence representations. Protein Sci 2003;12:1007–17. doi:10.1110/ps.0239403
      OpenUrlCrossRefPubMedWeb of Science
    45. ↵
      1. Lundegaard C,
      2. Lamberth K,
      3. Harndahl M, et al
      . Netmhc-3.0: accurate web accessible predictions of human, mouse and monkey MHC class I affinities for peptides of length 8-11. Nucleic Acids Res 2008;36:W509–12. doi:10.1093/nar/gkn202
      OpenUrlCrossRefPubMedWeb of Science
    46. ↵
      1. Andreatta M,
      2. Nielsen M
      . Gapped sequence alignment using artificial neural networks: application to the MHC class I system. Bioinformatics 2016;32:511–7. doi:10.1093/bioinformatics/btv639
      OpenUrlCrossRefPubMed
    47. ↵
      1. Peters B,
      2. Sette A
      . Generating quantitative models describing the sequence specificity of biological processes with the stabilized matrix method. BMC Bioinformatics 2005;6:132. doi:10.1186/1471-2105-6-132
    48. ↵
      1. Sidney J,
      2. Assarsson E,
      3. Moore C, et al
      . Quantitative peptide binding motifs for 19 human and Mouse MHC class I molecules derived using Positional scanning combinatorial peptide libraries. Immunome Res 2008;4:2. doi:10.1186/1745-7580-4-2
    49. ↵
      1. Jurtz V,
      2. Paul S,
      3. Andreatta M, et al
      . Netmhcpan-4.0: improved peptide-MHC class I interaction predictions integrating eluted ligand and peptide binding affinity data. J Immunol 2017;199:3360–8. doi:10.4049/jimmunol.1700893
      OpenUrlAbstract/FREE Full Text
    50. ↵
      1. Hoof I,
      2. Peters B,
      3. Sidney J, et al
      . Netmhcpan, a method for MHC class I binding prediction beyond humans. Immunogenetics 2009;61:1–13. doi:10.1007/s00251-008-0341-z
      OpenUrlCrossRefPubMedWeb of Science
    51. ↵
      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
    52. ↵
      1. Stranzl T,
      2. Larsen MV,
      3. Lundegaard C, et al
      . Netctlpan: Pan-specific MHC class I pathway EPITOPE predictions. Immunogenetics 2010;62:357–68. doi:10.1007/s00251-010-0441-4
      OpenUrlCrossRefPubMed
    53. ↵
      1. Riddell SR,
      2. Greenberg PD
      . The use of anti-CD3 and anti-CD28 monoclonal antibodies to clone and expand human antigen-specific T cells. J Immunol Methods 1990;128:189–201. doi:10.1016/0022-1759(90)90210-m
      OpenUrlCrossRefPubMedWeb of Science
    54. ↵
      1. Dossa RG,
      2. Cunningham T,
      3. Sommermeyer D, et al
      . Development of T-cell immunotherapy for hematopoietic stem cell transplantation recipients at risk of leukemia relapse. Blood 2018;131:108–20. doi:10.1182/blood-2017-07-791608
      OpenUrlAbstract/FREE Full Text
    55. ↵
      1. Philip B,
      2. Kokalaki E,
      3. Mekkaoui L, et al
      . A highly compact EPITOPE-based marker/suicide gene for easier and safer T-cell therapy. Blood 2014;124:1277–87. doi:10.1182/blood-2014-01-545020
      OpenUrlAbstract/FREE Full Text
    56. ↵
      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
    57. ↵
      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
    58. ↵
      1. Yadav M,
      2. Jhunjhunwala S,
      3. Phung QT, et al
      . Predicting immunogenic tumour mutations by combining mass spectrometry and exome sequencing. Nature 2014;515:572–6. doi:10.1038/nature14001
      OpenUrlCrossRefPubMedWeb of Science
    59. ↵
      1. Cameron BJ,
      2. Gerry AB,
      3. Dukes J, et al
      . Identification of a Titin-derived HLA-A1-presented peptide as a cross-reactive target for engineered MAGE A3-directed T cells. Sci Transl Med 2013;5:197ra103. doi:10.1126/scitranslmed.3006034
    60. ↵
      1. Chheda ZS,
      2. Kohanbash G,
      3. Okada K, et al
      . Novel and shared neoantigen derived from histone 3 variant H3.3K27m Mutation for glioma T cell therapy. J Exp Med 2018;215:141–57. doi:10.1084/jem.20171046
      OpenUrlAbstract/FREE Full Text
    61. ↵
      1. Rongvaux A,
      2. Willinger T,
      3. Martinek J, et al
      . Development and function of human innate immune cells in a humanized mouse model. Nat Biotechnol 2014;32:364–72. doi:10.1038/nbt.2858
      OpenUrlCrossRefPubMed
    62. ↵
      1. Song Y,
      2. Rongvaux A,
      3. Taylor A, et al
      . A highly efficient and faithful MDS patient-derived xenotransplantation model for pre-clinical studies. Nat Commun 2019;10:366. doi:10.1038/s41467-018-08166-x
    63. ↵
      1. van der Lee DI,
      2. Reijmers RM,
      3. Honders MW, et al
      . Mutated nucleophosmin 1 as immunotherapy target in acute myeloid leukemia. J Clin Invest 2019;129:774–85. doi:10.1172/JCI97482
      OpenUrlCrossRefPubMed
    64. ↵
      1. Guess T,
      2. Potts CR,
      3. Bhat P, et al
      . Distinct patterns of clonal evolution drive myelodysplastic syndrome progression to secondary acute myeloid leukemia. Blood Cancer Discov 2022;3:316–29. doi:10.1158/2643-3230.BCD-21-0128
      OpenUrlCrossRef
    65. ↵
      1. Menssen AJ,
      2. Khanna A,
      3. Miller CA, et al
      . Convergent clonal evolution of signaling gene mutations is a hallmark of myelodysplastic syndrome progression. Blood Cancer Discov 2022;3:330–45. doi:10.1158/2643-3230.BCD-21-0155
      OpenUrlCrossRef
    66. ↵
      1. Ferrari V,
      2. Tarke A,
      3. Fields H, et al
      . In vitro induction of neoantigen-specific T cells in myelodysplastic syndrome, a disease with low mutational burden. Cytotherapy 2021;23:320–8. doi:10.1016/j.jcyt.2020.10.003N
      OpenUrl
    67. ↵
      1. Winter S,
      2. Shoaie S,
      3. Kordasti S, et al
      . “Integrating the "Immunome" in the stratification of myelodysplastic syndromes and future clinical trial design”. JCO 2020;38:1723–35. doi:10.1200/JCO.19.01823
      OpenUrl
    68. ↵
      1. Tanaka TN,
      2. Ferrari V,
      3. Tarke A, et al
      . Adoptive transfer of neoantigen-specific T-cell therapy is feasible in older patients with higher-risk myelodysplastic syndrome. Cytotherapy 2021;23:236–41. doi:10.1016/j.jcyt.2020.11.003
      OpenUrl
    69. ↵
      1. Chao M,
      2. Thomay K,
      3. Goehring G, et al
      . Mutational spectrum of fanconi anemia associated myeloid Neoplasms. Klin Padiatr 2017;229:329–34. doi:10.1055/s-0043-117046
      OpenUrl
    70. ↵
      1. Kerkar SP,
      2. Muranski P,
      3. Kaiser A, et al
      . Tumor-specific CD8+ T cells expressing Interleukin-12 eradicate established cancers in Lymphodepleted hosts. Cancer Res 2010;70:6725–34. doi:10.1158/0008-5472.CAN-10-0735
      OpenUrlAbstract/FREE Full Text
    71. ↵
      1. Adachi K,
      2. Kano Y,
      3. Nagai T, et al
      . IL-7 and CCL19 expression in CAR-T cells improves immune cell infiltration and CAR-T cell survival in the tumor. Nat Biotechnol 2018;36:346–51. doi:10.1038/nbt.4086
      OpenUrlCrossRefPubMed
    72. ↵
      1. Bell M,
      2. Gottschalk S
      . Engineered cytokine signaling to improve CAR T cell Effector function. Front Immunol 2021;12:684642. doi:10.3389/fimmu.2021.684642
    73. ↵
      1. Qin G,
      2. Ilya S,
      3. Kim T-K, et al
      . Co-occurring mutation clusters predict drug sensitivity in acute myeloid leukemia. Blood 2020;136:12–3. doi:10.1182/blood-2020-142727
      OpenUrlCrossRef
    74. ↵
      1. Haferlach T,
      2. Nagata Y,
      3. Grossmann V, et al
      . Landscape of genetic lesions in 944 patients with myelodysplastic syndromes. Leukemia 2014;28:241–7. doi:10.1038/leu.2013.336
      OpenUrlCrossRefPubMedWeb of Science
    75. ↵
      1. Walter MJ,
      2. Shen D,
      3. Ding L, et al
      . Clonal architecture of secondary acute myeloid leukemia. N Engl J Med 2012;366:1090–8. doi:10.1056/NEJMoa1106968
      OpenUrlCrossRefPubMedWeb of Science
    76. ↵
      1. Flach J,
      2. Dicker F,
      3. Schnittger S, et al
      . An accumulation of cytogenetic and molecular genetic events characterizes the progression from MDS to secondary AML: an analysis of 38 paired samples analyzed by cytogenetics, molecular mutation analysis and SNP Microarray profiling. Leukemia 2011;25:713–8. doi:10.1038/leu.2010.304
      OpenUrlCrossRefPubMed
    77. ↵
      1. Kuykendall A,
      2. Duployez N,
      3. Boissel N, et al
      . Acute myeloid leukemia: the good, the bad, and the ugly. Am Soc Clin Oncol Educ Book 2018;38:555–73. doi:10.1200/EDBK_199519
      OpenUrl
    78. ↵
      1. Vainchenker W,
      2. Kralovics R
      . Genetic basis and molecular pathophysiology of classical myeloproliferative neoplasms. Blood 2017;129:667–79. doi:10.1182/blood-2016-10-695940
      OpenUrlAbstract/FREE Full Text
    79. ↵
      1. Kandoth C,
      2. McLellan MD,
      3. Vandin F, et al
      . Mutational landscape and significance across 12 major cancer types. Nature 2013;502:333–9. doi:10.1038/nature12634
      OpenUrlCrossRefPubMedWeb of Science

    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

    • Contributors MAB designed and performed experiments, analyzed data, and wrote the manuscript. MB reviewed data and wrote the manuscript. JL, KAF, RGB, KBW, and CC performed experiments, analyzed the data, and contributed figures and/or text to the manuscript. AR contributed to the design and execution of murine experiments involving MISTRG mice. TM and DW identified crucial patient samples. VGO, DLS, and HJD contributed to concept development. MAB and MB serve jointly as guarantors and accept full responsibility for the finished work and the conduct of the study, had access to the data, and controlled the decision to publish.

    • Funding This work was supported by the MPN Research Foundation and by the Flow Cytometry, Comparative Medicine and Biostatistics Shared Resource of the Fred Hutch/University of Washington Cancer Consortium (US Department of Health and Human Services, National Institutes of Health, National Cancer Institute P30 CA015704). MAB is an Amy Strelzer Manasevit Research Program Scholar (administered by the Be The Match Foundation). MB is the recipient of the Gerdin Family Endowed Chair for Leukemia Research, Fred Hutch.

    • Competing interests MB is a Founder and Scientific Advisory Board member of HighPassBio, and a Scientific Advisory Board member of Orca Bio, and has also received compensation from Miltenyi Biotec for presentations at conferences and corporate symposia pertaining to research unrelated to the current manuscript. MB and MAB have filed a provisional patent application number 63/274,681 covering applications of T cell immunotherapy for U2AF1-mutated malignancies.

    • 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.