Methods and compositions for shifting the immunopeptidome of cancer cells

Abstract

Methods and compositions for inducing an immunopeptidome shift in a tumor cell are provided. Such methods and compositions are useful for eliciting an immune response against the tumor cell.

Description

METHODS AND COMPOSITIONS FOR SHIFTING THE IMMUNOPEPTIDOME OF CANCER CELLSCROSS-REFERENCE TO RELATED APPLICATIONS

[0001] This application claims benefit of U.S. Provisional Application No. 63 / 724,620 filed on November 25, 2024, U.S. Provisional Application No. 63 / 850,062 filed on July 24, 2025, and U.S. Provisional Application No. 63 / 902,799 filed October 21, 2025, which are incorporated herein by reference.SEQUENCE LISTING

[0002] The present specification makes reference to a Sequence Listing (submitted electronically as an .xml file name “DFCI_3550W01WO_SL.xml” on November 24, 2025. The .xml file was generated on 24 November 2025 and is 2.54 KB in size. The entire contents of the sequence listing are herein incorporated by reference.STATEMENT OF FEDERALLY SPONSORED RESEARCH

[0003] This invention was made with government support under CA265928, and Al 143565 awarded by the National Institutes of Health. The Government has certain rights in the invention.FIELD

[0004] Provided herein are methods and compositions for inducing an immunopeptidome shift in a cancer cell to elicit a polyclonal cytotoxic T lymphocyte (CTL) response against the cancer cell.BACKGROUND

[0005] Mutations that occur early as a normal cell transforms into a cancer cell are referred to as truncal (or clonal) mutations, i. e. , these mutations will be carried by virtually all subsequent generations of tumor cells that derive from the transformed founder cell. For example, p53 functions to protect cells from DNA damage and promote apoptosis, so TP53 mutations are often early truncal events and are a common genetic driver in cancer. All subclones of such a tumor share the TP 53 mutation.

[0006] Cancer cells can accumulate additional mutations during cell division. Further, a population of cancer cells can be heterogenous. That is, these mutations can vary between individual cells in a population, meaning that different cells may acquire different combinations of mutations over time. Mutations that confer a fitness advantage on cancer cells are also referred to as “driver” mutations, whereas other mutations that do not confer such an advantage are referred to as “passenger” mutations. If a particular mutation, or combination of mutations, offers a selective advantage, a subclone comprising the mutation(s) will emerge as dominant.

[0007] Various therapeutic approaches have been developed that focus on targeting tumorspecific antigens, including neoantigens. Therapeutic targeting of truncal neoantigens is advantageous compared to targeting passenger mutations as their presence in virtually all cancer cells can minimize the possibility that the tumor will escape the targeted therapy. Yet, targeting truncal neoantigens is not without challenge.

[0008] For example, tumor cells can evade the immune system by masking the neoepitope portion of a neoantigen. Alternatively, a tumor harboring a mutation in a truncal neoantigen may suppress surface display of the corresponding neoepitope. More specifically, a tumor cell may exhibit a locus-specific genetic defect that prevents a particular HLA allele from presenting peptides, such that the immunogenic truncal neoepitope is not displayed to the immune system. Alternatively, a tumor cell may increase aminopeptidase trimming to avoid cytotoxic T cell targeting.

[0009] As a result of these various mechanisms of preventing neoantigen presentation on the cell surface, tumor cells may be relatively epitope-barren. A recent clinical trial (Rappaport et al., Nat Med. 2024 Apr; 30(4): 1013-1022) utilized 20 driver neoantigens (including p53) to test therapeutic approaches that combine vaccination with immune checkpoint blockade for advanced cancers. The results of this trial have been disappointing but consistent with the notion that critical neoepitopes are not being effectively presented.

[0010] Various approaches of shifting the immunopeptidome have been investigated to enhance the immunogenicity of tumor cells. One approach has focused on the upregulation of MHC-I, i. e. , the molecule responsible for presenting peptides to the immune system. For example, Zhang et al. (Int J Mol Sci. 2024 Mar 25; 25(7):3660) showed that treatment with birinapant, an inhibitor of IAP (Inhibitor of Apoptosis) proteins, upregulates MHC-I in tumor cells. Their findings indicate that birinapant treatment influences antigen presentation in specific pathways including those related to RNA metabolism. However, this approach is not expected to produce a therapeutically significant shift in the immunopeptidome.

[0011] Other approaches of shifting the immunopeptidome have been directed to altering the peptides that are bound and presented by MHC-I. For example, endoplasmic reticulum aminopeptidase 1 (ERAP1) inhibitors can prevent trimming of peptides, in particular the excessive trimming employed by tumor cells, to prevent neoantigen presentation (Koumantou et al. , Cancer Immunol hnmunother. 2019 Aug; 68(8): 1245-1261). However, as a significant portion of the immunopeptidome is ERAP 1 -independent, pharmacological inhibition of ERAP1 may be insufficient to induce a potent immune response against a tumor cell.

[0012] Yet another approach involves pharmacologic modulation of splicing to generate immunogenic neoepitopes. Such neoepitopes can be presented on MHC-I to trigger an anti-tumor T cell response (Lu et al., Cell. 2021 Jul 22;184(15):4032-4047). As with the other approaches, this approach does not significantly increase the immunogenicity of tumor cells.

[0013] Thus, a need remains to identify approaches that can substantially shift the immunopeptidome in tumor cells, thus increasing the immunogenicity thereof.SUMMARY

[0014] The present disclosure is based on the discovery that a tumor-specific polyclonal cytotoxic T lymphocyte (CTL) response can be induced by shifting the immunopeptidome of tumor cells to present self, but non-tolerized, peptides via a major histocompatibility complex (MHC) to CD8+T cells by MHC-I (and / or CD4+T cells by MHC -II). By shifting the immunopeptidome in this manner, the tumor cells become highly immunogenic, resulting in T cell activation and subsequent destruction of the tumor. To limit destruction to the tumor, the shift is selectively induced in tumor cells using a targeting means (e. g. , a target moiety capable of specifically binding a tumor-specific or tumor-associated antigen such as an antibody) for targeted delivery of a drug that binds to and alters the peptide binding preference of the MHC expressed by the tumor cells.

[0015] Accordingly, in a first aspect, a method of inducing an immunopeptidome shift in a tumor cell of a subject in need thereof is provided. The method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

[0016] In another aspect, a method of inducing a polyclonal CTL response against a tumor cell of a subj ect in need thereof is provided. The method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises atargeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

[0017] In another aspect, a method of treating a tumor in a subject in need thereof is provided. The method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means (e.g., targeting moiety such as an antibody) that directs the drug conjugate to the cells that form the tumor, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the cells of the tumor.

[0018] In a further aspect, a method of treating a tumor in a human subject in need thereof is provided that comprises: (a) obtaining information about the subject’s human leukocyte antigen (HLA) alleles; and (b) steps for altering the peptide binding preference of a major histocompatibility complex (MHC) which is encoded by one of said HLA alleles and expressed by cells of the tumor, thereby inducing an immunopeptidome shift in the cells and eliciting a polyclonal cytotoxic T cell (CTL) response against the tumor.

[0019] In another aspect, a drug conjugate comprising a targeting means (e.g., a targeting moiety such as an antibody) is provided wherein the targeting means directs the drug conjugate to a tumor cell, and the drug is capable of binding to and altering the peptide binding preference of a major histocompatibility complex (MHC) expressed by a tumor cell.

[0020] Other features, objects, and advantages are apparent in the detailed description and drawings that follow. It should be understood, however, that the detailed description and drawings are given by way of illustration only, not limitation. Various changes and modifications will become apparent to those skilled in the art.BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Further description by way of example is provided with reference to the following drawings.

[0022] FIG. 1A shows a schematic of a cell displaying a non-self-peptide 101 (a viral peptide derived from a viral protein 102) and a self-peptide 103a (derived from a self-protein 104) by MHC -I 105 on its surface 106. Self-peptides with high binding affinity for MHC -I are shown as white-filled squares (103a), and self-peptides with low / no binding affinity are shown as grey- filled squares (103b). Non-self-peptides (viral peptides, 101) are shown as black-filled squares. Peptide loading occurs in the endoplasmic reticulum (ER) 107. Following protein cleavage in theproteasome 108, the peptides 101, 103a and 103b reach the ER 107 via a TAP transporter 110 in the ER membrane 120. Only the viral peptide 101 is recognized by the T cell receptor (TCR) 111 of a CD8+T cell 112.

[0023] FIG. IB shows a schematic of a cell displaying a non-tolerized “new” self-peptide 103b* by MHC-I 105 in the presence of a drug 114 that alters the peptide binding preference of MHC-I. Self-peptides with high binding affinity for MHC-I are shown as white-filled squares (103a), and self-peptides with low / no binding affinity are shown grey-filled squares (103b). Following protein cleavage in the proteasome 108, the peptides 103a and 103b reach the ER 107 via a TAP transporter 110 in the ER membrane 120. The drug 114 is shown as a star shape. The drug 114 inserts into the MHC-I 105 binding cleft (not shown in FIG. IB) in the ER 107 and alters the peptide binding preference of the MHC-I 105. Consequently, a self-peptide 103b that is not normally presented by the cell due to its low binding affinity is loaded into MHC-I 105 in the presence ofthe drug 114 and presented on the cell surface 106 as a “new” self-peptide 103b*. This “new” self-peptide 103b* is recognized by the T cell receptor (TCR) 111 of a CD8+T cell 112.

[0024] FIG. 2A shows a schematic of three MHC-I 205 complexes (each encoded by the HLA-B*57:01 allele). As shown, high binding affinity self-peptides (203a, black filled) are loaded into the binding cleft 215 of MHC-I 205, whereas self-peptides with low / no binding affinity (203b, grey-filled) are not loaded. No cytotoxic T cell (CTL) response is induced against these peptides.

[0025] FIG. 2B shows a schematic of the same three MHC-I 205 complexes shown in FIG. 2A in the presence of abacavir 214a. Abacavir 214a is illustrated as a mirror-image L shape marked with a star symbol. As shown, abacavir 214a binds in the peptide binding cleft of HLA- B*57:01, thereby altering the binding preference of MHC-I 205. As a result, low / no affinity self- peptides 203b are loaded into the binding cleft 215 and presented by MHC-I 205. These “new” non-tolerized self-peptides are recognized by CD8+T cells and induce a polyclonal CTL response.

[0026] FIG. 3 shows protein sequence alignment of amino acid sequences encoded by exons 2 and 3 of key A2 supertype alleles. B pocket and F pocket residues are shown in light grey and dark grey, respectively. Discrepancies within the pocket residues of different HLA-A*02 alleles are depicted by black border.

[0027] FIG. 4 shows the structure of the peptide binding cleft 315 of MHC-I encoded by HLA-B* 15:02 with a bound peptide 303 (black-filled structure) composed of amino acid residues P1-P9 in the presence of carbamazepine 314b (grey-filled structure). The a-helices forming the peptide binding cleft are indicated as 316a (al) and 316b (a2). Carbamazepine 314b interacts withthe amino acid (318, leucine) at position 156 in 316b, thereby altering the binding preference of the MHC-I molecule.

[0028] FIG. 5 shows a schematic of a tumor cell displaying a non-tolerized “new” selfpeptide 403b* by MHC-I 405 in the presence of a drug 414 that binds to and alters the peptide binding preference of MHC-I 405. Self-peptides with high binding affinity for MHC-I 403a are shown as white-filled squares, and self-peptides with low / no binding affinity 403b are shown as grey-filled squares. The drug 414 is shown as a star shape. The drug 414 is delivered to a tumor cell using an antibody-drug conjugate (ADC) 430. The antibody 431 of the ADC is shown as a Y shape. It binds to a tumor-specific or tumor-associated antigen 440 expressed on the surface of the tumor cell 406. The ADC 430 is then internalized and the drug is delivered to the endoplasmic reticulum (ER) 407. The drug 414 binds the MHC-I binding cleft (not shown in FIG. 5) in the ER 407 and alters the peptide binding preference of the MHC-I 405. Consequently, a self-peptide 403b that is not normally presented by the cell due to its low binding affinity is loaded into MHC-I 405 in the presence of the drug 414 and presented on the cell surface 406 as a “new” self-peptide 403b* . The “new” self-peptide 403b* is recognized by the T cell receptor (TCR) 411 of a CD8+T cell 412.

[0029] FIG. 6 illustrates shared peptides between the abacavir-treated and untreated controlH2228-B*57:01 cells. The figure maps the DDA analysis of obtained samples onto the DIA analysis (scan cycle).

[0030] FIG. 7 shows a representative self-peptide presented by H2228-B*57:01 cells in the presence of abacavir, but not in its absence. FIG. 7A shows a data-dependent acquisition (DDA) run of parent and fragment ions of abacavir-treated H2228-B*57:01 cells and a corresponding reference fragmentation pattern for a peptide having the amino acid sequence KTVEIVHIDI (SEQ ID NO: 1). The mass-to-charge ratio of the peptide is indicated. FIG. 7B shows Poisson detection plots for the DDA run of untreated H2228-B*57:01 cells. The expected scan position of the KTVEIVHIDI peptide is annotated by the block arrow and “predicted elution”. The extracted ion amplitude for the mass-to-charge ratio is shown by the black plot. The inverted Poisson probability is shown in grey below.

[0031] FIG. 8 illustrates activation of CD8+T cells after co-culture with abacavir (ABC) treated autologous dendritic cells (DCs). Co-culture DCs treated with a solvent (DMSO) served as a negative control. CD8+T cells incubated with anti-CD3 / CD28 beads were used as a positive control. The figure shows the percentage of 4-lBB+CD8+T cells following co-culture or bead treatment. Expression of 4- IBB on CD8+T cells were measured at 24-hour intervals for 4 days.

[0032] FIG. 9 illustrates the specificity of the anti-tumor effector function of CD8+T cells conditioned by co-culture with abacavir (ABC)-treated autologous DCs. Effector function against H2228-B*57:01 cells was assessed after a 6-hour co-culture with ABC-treated H2228-B*57:01 cells. As a negative control, H2228-B*57:01 cells were mock-treated with solvent (DMSO). FIG. 9A shows a flow cytometry plot of the intracellular expression of granzyme B and perforin in conditioned CD8+T cells following co-culture with H2228-B*57:01 mock- or ABC-treated cells. CD8+T cells expressing granzyme B and perforin are marked by a box, with the corresponding percentage of the overall CD8+T cell population provided on the figure. FIG. 9B shows the concentration of secreted granzyme B in the supernatants from the co-cultures with mock- or ABC- treated H2228-B*57:01 cells. CD8+T cells incubated with anti-CD3 / CD28 beads were used as a positive control. A no T-cell control was also included. FIG. 9C shows the percentage of relative maximum killing of mock- or ABC-treated H2228-B*57:01 cells by the conditioned CD8+T cells after 6 hours of co-culture.

[0033] FIG. 10 shows abacavir in the F pocket of HLA-B*57:01 in the crystal structure 3VRJ which was obtained using X-ray crystallography. The self-peptide LTTKLTNTNI also present in this structure is not shown. Overlaid on the abacavir structure shown in a stick representation are the structures predicted by two artificial intelligence (Al) models, namely, AlphaFold3 and Boltz-2. The root mean square deviation (RMSD) of the atomic positions in the modeled structures relative to 3VRJ are 0.714 A (Boltz-2) and 1.129 A (AlphaFold 3), respectively.

[0034] FIG. 11 shows abacavir’s predicted binding using Boltz-2. FIG. 11A illustrates that Boltz-2 predicts abacavir to bind in the F pocket of the MHC-I encoded by the HLA-B*57:01 allele. FIG. 1 IB illustrates that Boltz-2 predicts abacavir to be unable to bind in the F pocket (or the A and B pockets) of the MHC-I encoded by the HLA-B*57:03 allele, as it does not readily fit into any of the three pockets. The two MHCs differ by only two amino acids.

[0035] FIG. 12 shows the predicted interactions between (A) compound (1), (B) compound (4), (C) compound (5), and (D) compound (10) and particular amino acid residues within the A and B pockets. The dashed lines indicate interactions between specific amino acid residues and moi eties of the compounds.

[0036] FIG. 13 shows the predicted interactions between (A) compound (2), (B) compound (3), (C) compound (6), (D) compound (7), (E) compound (8), and (F) compound (9) and particular amino acid residues within the F pocket. The dashed lines indicate interactions between specific amino acid residues and moieties of the compounds.

[0037] FIG. 14 shows the predicted binding of compounds (1) and (2) using Boltz-2. FIG. 14A illustrates that compound (1) is predicted to bind in the A and B pockets of the MHC-I encoded by the HLA-A* 02:01 allele, and FIG. 14B illustrates that compound (2) is predicted to be bind in the F pocket of the MHC-I encoded by the HLA-A*02:01 allele.DETAILED DESCRIPTION

[0038] Unless otherwise defined herein, technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs and as commonly used in the art to which this application belongs. Exemplary methods and materials are described below, although methods and materials similar or equivalent to those described herein can also be used in the practice or testing of the present disclosure. In case of conflict, the present specification, including definitions, will control. To facilitate ready understanding, certain terms used herein are first defined below.

[0039] Unless otherwise required by context, singular terms shall include pluralities, and plural terms shall include the singular. Thus, the singular forms “a”, “an”, and “the” include plural referents unless the context clearly dictates otherwise. For example, “a tumor cell” is understood to represent one or more tumor cell(s) or a population of tumor cells. As such, the terms “a” (or “an”), “one or more”, and “at least one” can be used interchangeably herein.

[0040] “And / or” is to be taken as specific disclosure of each of the two specified features or components with or without the other. Thus, the term “and / or” as used in a phrase such as “A and / or B” herein is intended to include “A and B”, “A or B”, “A” (alone), and “B” (alone). Eikewise, the term "and / or" as used in a phrase such as “A, B, and / or C” is used interchangeably with “A and / or B and / or C” and is intended to encompass each of the following aspects: A, B, and C; A, B, or C; A or C; A or B; B or C; A and C; A and B; B and C; A (alone); B (alone); and C (alone).

[0041] The words “have” and “comprise”, or variations such as “has”, “having”, “comprises”, or “comprising” will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers. It is further understood that wherever aspects are described herein with the language “comprising” or “having” or grammatical equivalents thereof, otherwise analogous aspects described in terms of “consisting of’ and / or “consisting essentially of’ are also provided. In other words, if a composition comprising A, B and C is recited, a composition consisting essentially of A, B and C is also contemplated as is a composition consisting of A, B and C.

[0042] The term “about” refers to an interval of accuracy that a person skilled in the art will understand to still ensure the technical effect of the feature in question. The term indicates a deviation from the indicated numerical value of ± 10%, or ±5 %, or ± 1 % of the indicated numerical value.

[0043] The term “substantially” refers to the qualitative condition of exhibiting total or near-total extent or degree of a characteristic or property of interest. One of ordinary skill in the biological arts will understand that biological and chemical phenomena rarely, if ever, go to completion and / or proceed to completeness or achieve or avoid an absolute result. The term “substantially” is therefore used herein to capture the potential lack of completeness inherent in many biological and chemical phenomena.

[0044] The term “in vitro " refers to events that occur in an artificial environment, e.g., in a test tube or reaction vessel, or in a cell in cell culture, etc., rather than within a multi-cellular organism.

[0045] The term “in vivo " refers to events that occur within a multi-cellular organism, such as a human or a non-human animal.

[0046] The term “subject” relates to a human patient as the intended recipient of the compositions described herein. The subject may be an adult, i.e., the subject is 18 years or older. The subject may be a child, i.e., the subject is less than 18 years of age. The subject may have a solid tumor. The subject may have a hematological malignancy.

[0047] The phrase “administering” (or “administer” or “administration”) refers to the act of an attending physician or caregiver, prescribing a composition disclosed herein for administration and thereby causing the application of the composition to a subject, through infusion, injection, or any other means, whether self-administered or administered by a clinician or other qualified care giver.

[0048] The term “therapeutically effective amount” refers to the amount that is sufficient to achieve a meaningful benefit to the subject (e.g., treating, modulating, curing, inhibiting, preventing, and / or delaying cancer or its symptoms). For example, a therapeutically effective amount may be an amount sufficient to achieve a desired therapeutic effect. Generally, the amount of a therapeutic agent (e.g., a drug conjugate described herein) administered to a subject in need thereof will depend upon the characteristics of the subject. Such characteristics include the condition, disease severity, general health, age, sex, and body weight of the subject. One of ordinary skill in the art will be readily able to determine appropriate dosages depending on theseand other related factors. In addition, both objective and subjective assays may optionally be employed to identify optimal dosage ranges.

[0049] The term “immunopeptidome” refers to the set of peptides (self-peptides and non- self-peptides, e.g., viral peptides) presented by a major histocompatibility complex (MHC) on the surface of a cell to enable T cell surveillance. Accordingly, the terms “shifting the immunopeptidome” or “inducing an immunopeptidome shift” refer to changing the set or repertoire of peptides that are usually presented by the MHC. For example, an immunopeptidome shift may change at least 5% (e.g., 10% or 15%) of the repertoire of peptides that are presented by the MHC. In some instances, an immunopeptidome shift may change at least 20% (e.g., at least 30% or at least 40%), of the repertoire of peptides that are presented by the MHC.

[0050] The term “self-peptide” refers to a peptide that is presented on both peripheral cells and in the thymus. Central tolerance (i.e., the ability to distinguish between self and non-self) is acquired during T cell maturation in the thymus. During this process, self-reactive T cells are eliminated. The destruction of self-reactive T cells in the thymus typically prevents the occurrence of an aberrant immune response (e.g., an autoimmune response) in the periphery. However, in some instances, such an aberrant immune response can nevertheless occur if a T cell escapes destruction in the thymus and recognizes a self-peptide in the periphery. Self-peptides may be presented by an MHC encoded by a particular allele in the periphery that were not displayed by the same allele-encoded MHC in the thymus during T cell development. T cells are not tolerized to such self-peptides, and some T cells may recognize these self-, but non-tolerized, peptides when presented by the MHC in a cell in the periphery and become activated, resulting in a T cell- mediated immune response.

[0051] If the presentation of self, but non-tolerized, peptides is the consequence of an altered binding preference (e.g., binding affinity or specificity) of MHC, such peptides may be referred to as “new” self-peptides as they would not normally have been presented by the MHC.

[0052] The term “neoantigen” refers to a tumor-specific protein generated by a tumor as a result of one or more tumor-specific alterations, e.g. , genomic mutations (such as point mutations, deletion, or translocations), dysregulated RNA splicing, disordered post-translational modification, or integrated viral open reading frames. A neoantigen may typically be recognized as non-self to trigger an immune response that is not subject to central and peripheral tolerance.

[0053] The terms “cytotoxic T lymphocyte” and “CTL” refer to a T lymphocyte (or T cell) that expresses the cell surface CD8 or CD4 glycoprotein (i.e., a CD8+or CD4+T cell) and has cytotoxic effector function. The cytotoxic effector function may be mediated, e.g., bygranzyme / perforin-induced apoptosis or the Fas pathway. Cytotoxic T cells are primarily derived from the CD8+subset, but an activated CD4+T cell can also have cytotoxic effector function.

[0054] The terms “polyclonal CTL response” or “polyclonal T cell response” refers to the activation of multiple different T cell clones. Each clone will have different peptide specificity, i.e., each clone will have distinct immunorecognition for a particular peptide presented by a corresponding MHC. A polyclonal T cell response can be mediated either by multiple different CD8+T cell clones or by multiple different CD4+T cell clones. In some instances, a polyclonal T cell response can be mediated by both multiple different CD8+T cell clones and multiple different CD4+T cell clones.

[0055] The term “obtaining information about a subject’s HLA allele(s)” (or grammatical equivalents thereof) refers to providing the identity of the HLA alleles of a subject. The identity of the HLA alleles may have been determined by an HLA typing assay known in the art. The information may be provided from clinical notes or an external provider.

[0056] The phrase “steps for altering the peptide binding preference of an MHC” refers to the general acts described herein to achieve the function of altering the peptide binding preference of at least one MHC (e.g., an MHC -I), which is encoded by an HLA allele of, and expressed by, cells of a tumor (e.g., an MHC-I), and thereby inducing an immunopeptidome shift in these cells and eliciting a polyclonal CTL response against the tumor. It is contemplated that any drug which can non-covalently bind within the binding cleft of one or more MHC molecules may be capable of shifting the immunopeptidome sufficiently to induce the polyclonal CTL response. In particular, any drug which can non-covalently bind within the F pocket or the A and / or B pockets of the binding cleft of one or more MHC molecules is considered to have this capability. Drugs that bind within the F pocket or the A and / or B pockets of the MHC-I are believed to be particularly effective in inducing the desired polyclonal CTL response. Equivalent acts therefore include contacting a cell of a tumor with a drug that induces an immunopeptidome shift in the cell, in particular any drug that non-covalently binds the F pocket or the A and B pocket of an MHC-I.

[0057] The term “immune checkpoints” means a well-recognized group of molecules on the surface of various cells including CD4+and CD8+T cells, tumor cells, and immunoregulatory cells. In the context of cancer, these molecules serve as “brakes” to down-modulate or inhibit an anti-tumor immune response. Immune checkpoint molecules include but are not limited to Programmed Death 1 (PD-1), Cytotoxic T Lymphocyte Antigen 4 (CTLA-4), PD-L1 (also known as B7-H1), and LAG3, each of which can inhibit immune cell function. The action of immune checkpoints may be reversed or attenuated by immunotherapeutic agents which can act as immunecheckpoint inhibitors including, but are not limited to, anti-PD-1 antibodies (e.g., pembrolizumab, nivolumab, and pidilizumab), anti-CTLA-4 antibodies (e.g., ipilimumab), anti-PD-Ll antibodies (e.g., atezolizumab, durvalumab, avelumab, MDX-1105, or AMP -224), and anti-LAG3 antibodies (e.g., relatlimab).

[0058] Generally, techniques of cell and tissue culture, molecular biology, virology, immunology, microbiology, genetics, analytical chemistry, synthetic organic chemistry, medicinal and pharmaceutical chemistry, and protein and nucleic acid chemistry and hybridization used herein is in accordance with those well-known and commonly used in the art. Enzymatic reactions and purification techniques are performed according to manufacturer’s specifications, as commonly accomplished in the art or as described herein. Further, nomenclature used in connection with these technology areas herein is as commonly used in the art as can be seen by reference to, e.g., http: / / www.informatics.jax.org / mgihome / nomen / gene.shtml and https: / / www.genenames.org / .

[0059] All publications and other reference materials referenced herein are hereby incorporated by reference in their entirety. Although a number of documents are cited herein, this citation does not constitute an admission that any one of these documents forms part of the common general knowledge in the art.Major histocompatibility complex (MHC)

[0060] In humans, major histocompatibility complexes (MHCs) are encoded by the human leukocyte antigen (HLA) genes. The MHCs are expressed on the surface of cells and present peptides of cellular proteins to the immune system, which the immune system then recognizes as “self’ or “non-self ’. The HLA genes are categorized into three groups: Class I, II, and III.

[0061] HLA class I comprises HLA-A, HLA-B, and HLA-C, which encode MHC -I proteins. HLA class II comprises HLA-DP, HLA-DQ and HLA-DR, which encode MHC-II proteins. The class II genes can be further defined by whether they encode the a-chain or the [3- chain and their locus e.g., HLA-DPA1, HLA-DPB1, HLA-DQA1, HLA-DQB1, HLA-DRA and HLA-DRB1.

[0062] HLA class I proteins (MHC-I) comprise a single polymorphic chain containing three domains (alpha 1, 2, and 3) which associate with [32 microglobulin (J32M) at the cell surface. HLA class II proteins (MHC-II) comprise two polymorphic chains, each containing two domains (alpha 1 and 2, and beta 1 and 2). HLA class I proteins (MHC-I) are expressed on virtually allnucleated cells. HLA class II proteins (MHC-II) are primarily found on activated lymphocytes and professional antigen presenting cells (APCs) such as macrophages, dendritic cells and B cells.

[0063] Each individual inherits one paternal and one maternal haplotype of HLA I or II, each haplotype containing three class I (A, B, and C) and three class II (DP, DQ, and DR) loci, i.e., a total of six alleles. Further, HLA genes are highly polymorphic within the human population. Accordingly, an extensive range of self and non-self-peptides can be presented by MHCs to the immune system. There are currently over 22,000 HLA alleles reported to date, and some alleles are more common in the human population than others. Table 1 illustrates the top 40 HLA-I alleles which cover approximately 90% of all HLA alleles in the human population globally.Table 1. Top 40 global HLA-I allelesMHC peptide presentation

[0064] MHC-I and MHC -II proteins play important and distinct roles in the presentation of peptides to circulating CD8+and CD4+T cells, respectively. Due to MHC restriction, a T cell can bind to a peptide presented by a particular HLA allele but cannot recognize the same peptide when bound by a different HLA allele.

[0065] To avoid recognition of self-peptides, a T cell expressing a T cell receptor (TCR) that recognizes an HLA-self-peptide complex is typically destroyed during T cell development - a process called “central tolerance” (also known as “negative selection”). This process eliminates developing T lymphocytes that are autoreactive. Accordingly, circulating T cells predominantly comprise TCRs specific for non-self-peptides.

[0066] FIG. 1A illustrates schematically that proteins (both self- 104 and non-self 102, e.g. , viral) undergo proteasomal cleavage into peptides 101, 103a and 103b in the cytoplasm of a cell. Some of these peptides are translocated to the endoplasmic reticulum (ER) 107 via a transporter associated with antigen processing (TAP) 110 to allow for peptide loading in the ER 107 and presentation by MHC-I 105. A non-self-peptidelOl (e.g., virus-derived peptide) can be recognized by a TCR 111 on the surface of a CD8+T cell 112 since the non-self-peptide 101 is not found in the thymus and hence no central tolerance was developed to this peptide. Recognition triggers a CTL response against the cell presenting the non-self-peptide 101 on its MHC-I 105. This CTL response is mediated by CD8+T cells 112.

[0067] FIG. IB illustrates schematically that a drug 114 can bind to and alter the peptide binding preference (e.g., peptide binding affinity) of an HLA-encoded MHC-I 105. Consequently, during the loading in the ER 107, self-peptides with low / no binding affinity 103b can be bound by MHC-I 105 and are presented on the cell surface 106 as a “new” self-peptide 103b*. Since these “new” self-peptides 103b* are not usually presented by the MHC-I 105 (e.g. , due their low binding affinity in the absence of the drug) and thus were not displayed during thymic tolerance induction, they can be recognized by a CD8+T cell 112, inducing a CTL response similar to non-self (e.g., viral) peptides 101 in FIG. 1A.

[0068] The TCR mediates recognition with sensitivity and specificity. Other instances in which perturbation of self-peptide presentation can trigger a CTL response are known. For example, urushiol (contained in plants such as poison ivy) causes a CD8+T cell-mediated reaction. This reaction is directed against cells presenting self-peptides derivatized by the oxidized urushiol that are recognized by CD8+T cells as non-self, or foreign. Similarly, the post-translational modification of self-peptides with conversion of arginine to citrulline can give rise to T cell activation in rheumatoid arthritis, an autoimmune condition.

[0069] APCs additionally express MHC-II for the presentation of non-self-peptides from extracellular sources (e.g., proteins from phagocytosed bacteria) to CD4+T cells. In some instances, a CTL response can also be induced by CD4+T cells. For example, an HLA-DR1- restricted human melanoma neoantigen derived from a mutated form of triosephosphate isomerase (TPI) is recognized by CD4+tumor-infiltrating lymphocytes and triggers a CTL response against melanoma cells (Deng and Mariuzza, Trends Biochem Sci. 2007 Nov; 32(11): 500-8).

[0070] The methods and compositions disclosed herein artificially exploit these mechanisms to alter the immunopeptidome of tumor cells and thus induce a targeted CTL response thereagainst.Peptide binding cleft

[0071] As the MHC-I (HLA class I) peptide binding cleft is closed at the N and C termini, peptides capable of binding to MHC-I are restricted to an optimal length of about 8-10 amino acids, e.g., 9 amino acids. For MHC-II (HLA class II), each end of the peptide binding cleft is in an open configuration. MHC-II thus can accommodate peptides of greater length than the peptide binding cleft of MHC-I.

[0072] MHC-I molecules comprise six distinct binding pockets, named pockets A-F. Together, the six binding pockets form the binding cleft (Nguyen et al., Biochem Soc Trans. 2021 Nov 1;49(5):2319-2331). Each of the binding pockets allow for side chains of a peptide’s amino acids to bind inside the cleft. Thus, the amino acid sequence of a pocket dictates which side chains can bind and the conformation of these six pockets determines the peptide binding preference of the MHC-I molecule. Binding pockets B and F are responsible for housing the primary anchor residues of a peptide, i.e., the residues which form the main interaction between MHC-I and the peptide. As such, the binding specificities of pockets B and F are considered determinant of the peptide binding preference of an MHC-I molecule. A summary of the six binding pockets is provided in Table 2.Table 2. MHC-I (HLA-I) binding pockets

[0073] In particular, the compositions described herein include drugs (e.g, small molecules) that can alter the binding preference of MHC-I and, in some instances, MHC-II. It is believed that such drugs interact with amino acid acids near or within the binding pockets of MHC- I (and, in some instances, MHC-II) that form the cleft for peptide presentation. These interactions can change the peptide binding preference (e.g. , the binding affinity or specificity) of the cleft, and thus, the peptides that are capable of binding in the pockets of the cleft. As a result, self-peptidesthat are not normally presented by MHC-I (and, in some instances, MHC-II) and thus have escaped the afore-mentioned process of central tolerance, are presented.Drugs capable of altering the peptide binding preference of an MHC

[0074] Drugs that can alter the peptide binding preference of MHC molecules are well- known in the art (e.g., Sousa-Pinto etal., Int Arch Allergy Immunol. 2016; 170(3): 163-79; and Fan et al., J Immunol Res. 2017:2017:3186328). The peptide binding preference of an MHC molecule is determined by the HLA allele encoding it. By altering the peptide binding preference, such drugs cause the MHC molecule to bind peptides that are usually not compatible, or are less compatible, with the binding pocket of the MHC molecule, i.e., has no- to low binding affinity.

[0075] A drug capable of altering the peptide binding preference of an MHC may also be referred to as an “immunopeptidome shifter”. For example, an immunopeptidome shifter may bind to the peptide binding cleft of an MHC-I encoded by a class I HLA allele. Such an immunopeptidome shifter may referred to as a “class I HLA” immunopeptidome shifter. An immunopeptidome shifter may bind to the peptide binding cleft of an MHC encoded by an HLA- A allele. Such an immunopeptidome shifter may referred to as an “HLA-A” immunopeptidome shifter. Correspondingly, an immunopeptidome shifter may bind to the peptide binding cleft of an MHC encoded by an HLA-B allele. Such an immunopeptidome shifter may referred to as an “HLA-B” immunopeptidome shifter. Moreover, an immunopeptidome shifter may bind to the peptide binding cleft of an MHC encoded by an HLA-C allele. Such an immunopeptidome shifter may referred to as an “HLA-C” immunopeptidome shifter. In some instances, an immunopeptidome shifter may bind non-covalently within the F pocket, the A pocket, or the B pocket (or the A and B pockets) of the peptide binding cleft of an MHC-I encoded by a class I HLA-allele. Such an immunopeptidome shifter may be referred to as an “F pocket binder”, “A pocket binder”, or a “B pocket binder (or an “A and B pocket binder”), respectively.

[0076] As illustrated by Table 3, many immunopeptidome shifters are known, as are the corresponding allele(s) that encode the MHC molecule that they bind to and of which they alter the peptide binding preference of. For the avoidance of doubt, the immunopeptidome shifters in Table 3 were identified based on the aberrant immune response that they induce in individuals expressing the specific HLA-I or II allele. The methods described below allow the identification of further immunopeptidome shifters based on structural modelling and in vitro validation.

[0077] An exemplary immunopeptidome shifter is the nucleoside reverse transcriptase inhibitor, abacavir, that is approved for use against human immunodeficiency virus 1 (HIV-1).Abacavir has been documented to elicit an autoimmune attack against the cells of individuals that carry the HLA-B*57:01 allele. Abacavir has also been shown to bind non-covalently within the binding pocket F of the MHC-I encoded by the HLA-B*57:01 allele and shift the preference for the carboxy-terminal peptide anchor residues from tryptophan or phenylalanine to isoleucine or leucine. This alters the binding preference (or binding affinity) of that MHC-I-encoding allele and leads to a shift in the peptides presented by the MHC-I. In particular, it has been found that abacavir binding causes the presentation of multiple self-peptides in the MHC-I cleft that are not usually presented. These “new” self-peptides are recognized by CD8+T cells as foreign. Upon recognition of these “new” self-peptides, CD8+T cells become activated and initiate a polyclonal CTL response. Accordingly, abacavir is an HLA-B immunopeptidome shifter, or more specifically, an HLA-B*57:01 immunopeptidome shifter and can alternatively be characterized as an HLA- B*57:01 F pocket binder.

[0078] FIG. 2A and FIG. 2B provides a schematic depicting the presentation of peptides in the absence or presence of abacavir 214a, respectively. FIG. 2A and FIG. 2B show three MHC-I molecules 205 encoded by the HLA-B *57: 01 allele. FIG. 2A illustrates the binding preference of the MHC-I molecules 205 in the absence of abacavir 214a. Only self-peptides that fit well into the binding cleft 215 of the respective MHC-I molecules 205 are loaded and displayed. These are shown as the black-filled peptides 203a. Self-peptides with no or low binding affinity (shown as grey-filled peptides 203b) to the MHC-I molecules 205 encoded by HLA-B*57:01 remain unbound and are not loaded and thus are not displayed at the cell surface. As shown in FIG. 2B, abacavir 214a can insert itself into the peptide binding cleft ofHLA-B*57:01-encoded MHC-I 205 and alter the binding preference of this molecule. As a consequence, self-peptides which previously were not bound by HLA-B*57:01-encoded MHC-I 201 can now be loaded and displayed on the cell surface. These “new” self-peptides are recognized by CD8+T cells, resulting in the induction of a polyclonal CTL response.

[0079] Another example is the anti-epileptic drug carbamazepine. It has been reported that the use of carbamazepine in individuals with certain HLA class I and class II alleles is associated with delayed drug allergy reactions. For example, HLA-B* 15:02 is associated with carbamazepine-induced Stevens-Johnson syndrome (SJS). Similar to abacavir, detailed investigations have found that carbamazepine non-covalently binds to the MHC-I encoded by the HLA-B* 15:02 allele at secondary anchor sites and causes alterations in the binding preference of MHC-I, resulting in the presentation of multiple non-tolerized self-peptides and the induction of a polyclonal CTL response. Accordingly, carbamazepine can be classified as an HLA-B immunopeptidome shifter, more specifically an HLA-B* 15:02 immunopeptidome shifter. Asshown in Table 3, carbamazepine has also been associated with altering the peptide binding preference of HLA-A and HLA-C alleles. Accordingly, carbamazepine can be considered as a class I HLA immunopeptidome shifter. Similarly, carbamazepine has been associated with altering the peptide binding preference of class II HLAs and therefore can be considered as a class II HLA immunopeptidome shifter.

[0080] Other examples of drugs that can alter the peptide binding preference of MHC molecules include the non-nucleoside reverse transcriptase inhibitor nevirapine and the xanthine oxidase inhibitor allopurinol. Therapeutic use of these drugs is associated with hypersensitivity reactions in individuals with certain HLA class I alleles including HLA-C*04:01 and the HLA- B*58:01, respectively. Accordingly, nevirapine and allopurinol are HLA-C and HLA-B immunopeptidome shifters, respectively, more specifically, HLA-C*04:01 and the HLA-B*58:01 immunopeptidome shifters.

[0081] Table 3 provides exemplary associations between drug hypersensitivity reactions and HLA alleles. The drugs listed in this table are all believed to be capable of inducing an immunopeptidome shift, i.e., can be considered to be immunopeptidome shifters.Table 3. Exemplary drugs capable of inducing an immunopeptidome shift

[0082] As can be seen from Table 3, known drugs alter the peptide binding preference of a wide variety of HLA class I (and HLA class II) alleles, and therefore can readily be exploited for use in the methods and compositions described herein. Each exemplary drug in Table 3 is capable of inducing an immunopeptidome shift in at least one of the top 40 HLA-I alleles provided in Table 1. For example, carbamazepine is capable of inducing an immunopeptidome shift in 9 of the top 40 alleles provided in Table 1. The drugs are generally small molecules (e.g., with molecular weights of less than 500, or even less than 350 Da) with one or more cyclic moiety.

[0083] Included herein are means for altering the peptide binding preference of an MHC. The phrase “means for altering the peptide binding preference of an MHC” refers to a chemical compound or drug that can bind the MHC of a target cell, particularly within the peptide binding cleft, and the functional expression “altering” refers to a change in the peptide binding specificity or affinity of the MHC (e.g., because peptides no longer fit into the MHC’s peptide binding cleft in the presence of the chemical compound or drug) such that the resulting repertoire of self-peptides presented by a cell are shifted to present peptides (including “new” self-peptides) that are not normally presented by the MHC. Since such peptides were not presented to T cells in the thymus during central tolerance, a polyclonal CTL response is generated against the target cell. For example, the alteration may change at least 5% (e.g., 10% or 15%) of the repertoire of peptides that are presented by MHC on the tumor cell surface. In some instances, the alteration may change at least 20% (e.g., at least 30% or at least 40%) of the repertoire of peptides that are presented by MHC on the tumor cell surface. Equivalent drugs may vary by their chemical structure but perform the same function of altering the peptide binding preference of an MHC, thereby achieving the same result of changing the repertoire of self-peptides presented by a cell and inducing a polyclonal CTL response against the cell. For example, structural analogues of abacavir are known in the art that are capable of binding to and altering the peptide binding preference of an MHC (Thomson et al. Allergy. 2020 Mar;75(3): 636-647). Corresponding exemplary drugs are provided herein, including in Table 3. Equivalent means alters the peptide binding reference of at least one MHC expressed by a cell.

[0084] Moreover, the structural knowledge of key amino acid residues within the binding pockets forming the cleft of MHC-I can be used for rational drug design to provide further small molecule drugs (including derivatives or structural variants of the compounds listed in Table 3) that will similarly alter the binding capability of MHC-I. As described in Table 2, a self-peptide binds anchor residues within the B and F pockets of MHC-I. As shown for abacavir, drugs that interact with at least one of these anchor residues can alter the peptide binding preference of the MHC-I. Accordingly, thousands of “new” non-tolerized self-peptides can instead preferentiallybind to the MHC-I that could not previously do so. The repertoire of peptides that will be presented by the MHC-I is thus changed, and a polyclonal CD8+T cell response will be generated against the cell. The polyclonal nature of the T cell response induced by the presentation of multiple distinct “new” self-peptides makes it difficult for tumor cells to evade the immune system.

[0085] Accordingly, a drug for use with the methods and compositions described herein may interact with a single MHC-I binding pocket. For example, a drug may interact with the F pocket or the B pocket of the peptide binding cleft of MHC-I (e.g. , at least one, or each, of HLA- A, HLA-B, and HLA-C). The A and B pockets are topologically adjacent to each other, such that when a drug binds at the B pocket, it may also interact with or occupy the A pocket. Therefore, a drug may interact with more than one binding pocket, e.g. , the A and B pockets, or two or more of binding pockets F, C, and E of the peptide binding cleft of MHC-I (e.g, at least one, or each, of HLA-A, HLA-B, and HLA-C).

[0086] For example, the drug may interact with one or more amino acid residues at positions 7, 9, 24, 34, 45, 63, 66, 67, 70, and 99 corresponding to binding pocket B of MHC-I (e.g., at least one, or each, of HLA-A, HLA-B, and HLA-C), as provided in Table 2. In the MHC-I encoded by the HLA-A* 02:01 allele, these amino acid residues correspond to Tyrosine 7, Phenylalanine 9, Alanine 24, Valine 34, Methionine 45, Glutamate 63, Lysine 66, Valine 67, Histidine 70 and Tyrosine 99. The drug may interact with one or more of Tyrosine 7, Phenylalanine 9, Alanine 24, Valine 34, Methionine 45, Glutamate 63, Lysine 66, Valine 67, Histidine 70 or Tyrosine 99. In particular, the drug may interact with one or more of Glutamate 63 or Lysine 66 of the B pocket. In some instances, a drug interacting with the B pocket may also interact with Tyrosine 159 within the A pocket.

[0087] The same amino acids residues are present in the MHC-I encoded by other HLA- A*02 alleles (for example, HLA-A*02:02, HLA-A*02:03, HLA-A*02:05, HLA-A*02:06, HLA- A*02:07 and HLA-A*02:011) with the exception of the polymorphic amino acid residues at positions 9 and 99 (see FIG. 3). In the MHC-I encoded by either the HLA-A* 02: 05 or HLA- A*02:06 allele, the amino acid at position 9 is Tyrosine 9, and in the MHC-I encoded by HLA- A*02:07 allele, the amino acid at position 99 is Cysteine 99. Neither of these amino acid positions are predicted to be integral to the binding of a drug molecule. Accordingly, a drug that binds to the B pocket of the MHC-I encoded by the HLA-A* 02: 01 allele is likely to bind to other MHC-I peptides encoded by other HLA-A*02 alleles. For example, provided compounds (1), (4), (5) and (10) are predicted to bind to HLA-A* 02: 02, HLA-A* 02: 03, HLA-A* 02: 05, HLA-A* 02: 06, HLA- A*02:07 and HLA-A*02:011.

[0088] Drugs binding the B pocket may also bind to the topologically adjacent A pocket. In particular, drugs binding in the A pocket and / or B pocket may interact with amino acid residues 7, 63 and 66 of MHC-I (see Table 2).

[0089] Alternatively, the drug may interact with one or more amino acid residues at positions 77, 80, 81, 84, 95, 116, 123, 143, 146, and 147 corresponding to binding pocket F of MHC-I (e.g., at least one, or each, of HLA-A, HLA-B, and HLA-C MHC), as provided in Table 2. In the MHC-I encoded by the HLA-A* 02: 01 allele, these amino acid residues correspond to Aspartate 77, Threonine 80, Leucine 81, Tyrosine 84, Valine 95, Tyrosine 116, Tyrosine 123, Threonine 143, Lysine 146 and Tryptophan 147. The drug may interact with one or more of Aspartate 77, Threonine 80, Leucine 81, Tyrosine 84, Valine 95, Tyrosine 116, Tyrosine 123, Threonine 143, Lysine 146 or Tryptophan 147. In particular, the drug may interact with one or more of Aspartate 77, Tyrosine 64, Threonine 143 and Tryptophan 147 of the F pocket. In particular, the drug may interact with Aspartate 77.

[0090] The same amino acids residues are present in the MHC-I encoded by other HLA- A*02 alleles (for example, HLA-A*02:02, HLA-A*02:03, HLA-A*02:05, HLA-A*02:06, HLA- A*02:07 and HLA-A*02:011) with the exception of the polymorphic amino acid residue at position 95 (see FIG. 3). In the MHC-I encoded by the HLA-A*02:02 and HLA-A*02:05, the amino acid at position 95 is Leucine 95. The amino acid at position 95 is not predicted to be integral to the binding of a drug molecule. Accordingly, a drug that binds to the F pocket of the MHC-I encoded by the HLA-A* 02: 01 allele is likely to bind to other MHC-I peptides encoded by other HLA-A alleles. For example, compounds (2), (3), and (6) to (9) are predicted to bind to HLA- A*02:02, HLA-A* 02: 03, HLA-A* 02: 05, HLA-A* 02: 06, HLA-A* 02: 07 and HLA-A*02:011.

[0091] As illustrated schematically in FIG. 2B, abacavir 214a can bind in the peptide binding cleft 215 of MHC-I 205 encoded by the allele HLA-B*57:01. The insertion of abacavir 214a in or near one of the binding pockets in the cleft causes a shift in the binding preference of the MHC-I molecule 205 encoded by HLA-B*57:01. This alteration in the binding preference allows loading of “new” self-peptides 203b that, in the absence of abacavir 214a, had no or low binding affinity for the MHC-I 205 encoded by the HLA-B*57:01 allele.

[0092] Abacavir binding is believed to modify the F pocket architecture of MHC-I encoded by HLA-B*57:01 allele, specifically through interactions with amino acid residue 116 (Illing etal. Nature. 2012 Jun 28;486(7404):554-8). Accordingly, the drug may interact with the amino acid residue at position 116 in binding pocket F of MHC-I (e.g., at least one, or each, of HLA-A, HLA- B, and HLA-C).

[0093] Binding of carbamazepine 314b to an MHC-I encoded by HLA-B* 15:02 is shown in FIG. 4. FIG. 4 shows the structure of the peptide binding cleft 315 of MHC-I encoded HLA- B* 15:02 with a bound peptide composed of amino acid residues P1-P9 (black structure; 303) in the presence of carbamazepine 314b (grey structure). The a-helices forming the peptide binding cleft 315 are indicated as al 316a and a2 316b. Carbamazepine 314b is shown within the binding cleft 315wherein it interacts with the amino acid at position 156 318 in a2 316b, thereby altering the binding preference of the MHC-I molecule. Accordingly, the drug may interact with the amino acid residue at position 156 in the peptide binding cleft of MHC-I (e.g., at least one, or each, of HLA-A, HLA-B, and HLA-C). Specifically, HLA-B* 15:02 has a leucine residue at position 156 compared to a tryptophan residue, as found in HLA-B* 15:01. As discussed in filing et al. (ibid), the tryptophan at position 156 in HLA-B* 15: 01 would prevent binding of carbamazepine due to sterically clashing with the docked conformation of the drug.

[0094] The drug is selected for its capability to induce an immunopeptidome shift. The original target for which the drug was designed (e.g. , the reverse transcriptase of HIV-1 in the case of abacavir) and associated clinical indications are irrelevant. For greater safety, it may, however, be desirable to modify drugs that are capable of altering the peptide binding preference of an MHC such that they have reduced affinity for the target for which they were originally intended. Such modifications are not relevant to the function of altering the peptide binding preference of an MHC and therefore do not remove the drug from the range of equivalents.

[0095] Further drugs, or derivatives or structural variants of existing drug, that alter the peptide binding preference of an MHC can be identified using a suitable screening method. Accordingly, in another aspect, a method of identifying a compound that induces an immunopeptidome shift in a mammalian cell expressing an MHC is provided. The method comprises incubating a first cell expressing the MHC in the presence of the compound, incubating a second cell expressing the MHC in the absence of the compound, and determining whether the presence of the compound alters the repertoire of peptides presented by the first cell. The first and second cells are the same type of mammalian cell. A compound is identified as inducing an immunopeptidome shift (i.e., an immunopeptidome shifter) if the compound alters at least 5% (e.g. , 10% or 15%), or at least 20% (e.g., at least 30% or at least 40%) of the repertoire of peptides presented by the first cell relative to the second cell.

[0096] Various mass spectroscopy-based methods including liquid-chromatography data- dependent acquisition mass spectrometry (LC-DDA-MS), liquid-chromatography data- independent acquisition mass spectrometry (LC-DIA-MS) and hybrid methods combining ionmobility with LC-MS methods, such as Parallel Accumulation- Serial Fragmentation Mass Spectrometry (PASEF-MS) are known to the skilled person and available to determine the immunopeptidome of an MHC -expressing cell. LC-DDA-MS methods as described for conventional immunopeptidomics may be particularly useful for determining the immunopeptidome of an MHC -expressing cells.

[0097] Provided herein are small molecule compounds that are predicted to bind to the A and B pockets, or the F pocket of the MHC -I encoded by the HLA-A*02:01 allele. These compounds are predicted to act as immunopeptidome shifters.F pocket binders

[0098] Provided herein are compounds that are predicted to bind to the F pocket within MHC-I. Compounds interacting with the F pocket were found to favor compact, conformationally restricted scaffolds that maximize hydrophobic contacts and entropic gains.

[0099] In particular, Compounds (2), (3) and (6) to (9) were identified as possible F pocket binders. These compounds comprise amide or urea-like groups capable of acting as both a hydrogen bond donor and a hydrogen bond acceptor, in addition to a cyclic moiety that is capable of binding within the F pocket. The predicted orientation of binding in the F pocket is presented in Scheme 1 with the F pocket depicted as a dashed circle.Compound (6) Compound (7)Compound (8) Compound (9)Scheme 1 - Predicted orientation of binding of compounds (2), (3) and (6) to (9) in the F pocket of the MHC-I. The dashed lines represent the F pocket.

[0100] Scheme 1 shows that the cyclic moiety binds deep within the F pocket, while the amide or urea-like group forms hydrogen bonds at the opening of the pocket. It also illustrates that further chemical moieties comprised in these compounds are positioned away from the pocket, such that they can be functionalized. For example, these further chemical moieties may be amenable to modification with a conjugatable moiety, such that that the drug may be conjugated to a targeting means (e.g., an antibody).

[0101] Accordingly, drugs that bind to the F pocket generally comprise a polar group. In particular, drugs that bind to the F pocket may comprise an amide or urea-like group that binds around the opening of the F pocket, where the group can act as both a hydrogen bond donor and a hydrogen bond acceptor. The amide or urea-like group may interact with the amino acid residues at positions 77 and 147 at the opening of the pocket. In particular, the amide or urea-like group may interact with Aspartate 77 and Tryptophan 147.

[0102] Furthermore, the drug that binds to the F pocket may also comprise a cyclic group that binds deep within the F pocket. The cyclic group may be an aromatic or a non-aromatic group. When the cyclic group is an aromatic substituent, it may be a substituted or unsubstituted phenyl, a substituted or unsubstituted heteroaromatic group, or a substituted or unsubstituted bicyclic group, wherein at least one of the rings is aromatic. Where the cyclic group is substituted, the substituent is generally a small hydrophobic group, or the substituent is a carbonyl. The presence of a substituted or unsubstituted heteroaromatic group, such as pyridone, or fused bicyclic motifs, such as purine-derivatives, in these molecules, suggests that the F pocket tolerates planar structurescapable of hydrogen bonding, as well as hydrophobic stacking. The aromatic substituents are likely to enhance the affinity to the F pocket through 71-71 interactions that are available in the Tyrosinerich F pocket. When the further cyclic group is a non-aromatic substituent, it is a small cycloalkyl.

[0103] The co-existence of both aliphatic and aromatic chemotypes suggests that the F pocket is chemically versatile. The F pocket can accommodate both aromatic and non-aromatic groups, provided that a strategically placed amide or related polar group is retained near the opening of the pocket to anchor interactions.

[0104] Further still, the drugs may optionally comprise a further chemical moiety that sits outside the F pocket. The further chemical moiety, when present, may modulate the properties of the drug. The further chemical moiety, when present, may contribute to the binding of the drug by interacting with residues close to the F pocket within the MHC-I and / or ensure the desired hydrophobicity (e.g., LogP of < 3.5). The further chemical moiety, when present, may have a MW of < 300 Daltons to ensure that the drug (including the cyclic moiety) has a MW of < 350 Daltons. The further chemical moiety, when present, may comprise a cyclic moiety. The cyclic moiety may be a substituted or unsubstituted aromatic group, wherein the substituent is a further cyclic group such as a carbocycle or heterocycle. The cyclic moiety may be a substituted or unsubstituted non- aromatic group, wherein the substituent is a further cyclic group such as an aromatic group.A and B pocket binders

[0105] Provided herein are compounds that are predicted to bind to the A and B pockets of MHC-I. In particular, Compounds (1), (4), (5) and (10) were identified as possible A and B pocket binders. These compounds comprise an amide group capable of forming hydrogen bonds with the A pocket, in addition to an aliphatic moiety capable of binding within the B pocket. The predicted orientation of binding is presented in Scheme 2 with the A and B pockets depicted as a dashed circle.Compound (5) Compound (10)Scheme 2 - Predicted orientation of binding of compounds (1), (4), (5) and (10) in the A and B pocket of the MHC-I. The dashed lines represent the deep A and B pockets.

[0106] Scheme 2 shows that the amide framework and the hydrophobic moiety both bind within the A and B pockets. The amide framework binds within the A pocket through hydrogen bonding, while the hydrophobic group interacts with the B pocket. It also illustrates that further chemical moieties comprised in these compounds are positioned away from the pocket, such that they can be further functionalized. For example, these further chemical moieties may be amenable to modification with a conjugatable moiety, such that that the drug may be conj ugated to a targeting means (e.g., an antibody).

[0107] Compound (4) illustrates that there is no requirement for a further chemical moiety. Compound (4) already comprises a conjugatable moiety in the form of a thiol group and thus can be directly functionalized.

[0108] Accordingly, drugs that bind to the A pocket may comprise a pair of adjacent amide groups that provide strong polar interactions in the A pocket. The pair of adjacent amide groups establishes a directional hydrogen-bonding array that mimics a short peptide fragment. The drug may comprise a stereocenter adjacent to the amide framework, indicating that the A pocket is stereochemically constrained. The pair of adjacent amide groups may interact with the amino acid residues at positions 7, 63, 66, and 99 in the A pocket. In particular, it may interact with Tyrosine 7, Glutamate 63, Lysine 66, and Tyrosine 99. In addition, it may interact with Tyrosine 159 (if present).

[0109] Drugs that bind to the B pocket may comprise a compact hydrophobic substituent which fits in the B pocket’s hydrophobic cavity. The hydrophobic substituent may be an aliphatic group. In particular, the aliphatic group may be C3-C6 cycloalkyl or branched chain alkyl group. The hydrophobic group may be a substituted or unsubstituted aromatic group.

[0110] It was found that a drug interacting with the B pocket may also interact with the A pocket which is topologically adjacent to the B pocket. Thus, drugs may comprise a combination of structural features that enable them to bind to the A pocket and the B pocket.[oni] The identification of Compound (4) suggests that the A and B pockets can accommodate aromatic and heteroaromatic groups that are available for 71-71 stacking. Where the aromatic group is substituted, the substituent is a halogen, such that halogen bonds may form.

[0112] Drugs that bind to the A and B pockets may comprise a further chemical moiety that sits outside the A and B pockets. The further chemical moiety, when present, may modulate the properties of the drug. The further chemical moiety, when present, may contribute to the binding of the drug to the A and / or B pockets within the MHC-I. The further chemical moiety may ensure the desired hydrophobicity (e.g., LogP of < 3.5). The further chemical moiety, when present, may have a MW of < 300 Daltons to ensure that the drug (including the cyclic moiety and conjugatable moiety described below) has a MW of < 350 Daltons. The further chemical moiety, when present, may comprise an aromatic group. The aromatic group may be substituted or unsubstituted. The substituent may be one or more of a C3-C13 cycloalkyl or a C5-C13 heterocycle. The aromatic group may be a multicyclic aromatic group, wherein one or more of the rings are aromatic.Coniusatcible moiet

[0113] The drug may comprise a conjugatable moiety. The conjugatable moiety provides a means to conjugate the drug to a targeting means (e.g., an antibody), optionally via a means for linking (e.g., a linker, typically a cleavable linker). The conjugatable moiety may be selected from the group consisting of an alcohol (e.g., a phenol or aliphatic alcohol), a carboxylic acid, a carbonyl (e.g., an aldehyde or ketone), an amine, an amide and a thiol.

[0114] In some instances, the drug may be chemically modified to encompass a conjugatable moiety. For example, a compound may be identified as being capable of binding to and altering the peptide binding preference of an MHC. As demonstrated herein, such a compound may comprise a polar amide or urea-like framework, with either a cyclic moiety or hydrophobic moiety, depending on the binding pocket of the MHC of interest, and optionally a further chemical moiety, but may lack a conjugatable moiety.

[0115] Therefore, to form a drug that may be used in the drug conjugates or liposomes described herein, a compound (C) may be modified to comprise a conjugatable moiety (R), being a reactive group that is optionally selected from the group consisting of an alcohol (e.g., a phenolor aliphatic alcohol), a carboxylic acid, a carbonyl (e.g., an aldehyde or ketone) an amine, an amide and a thiol:

[0116] Such a compound (C) comprising a conjugatable moiety (R) is considered a derivative or structural variant of the compound (e.g., a compound listed in Table 3, or one of compound (1) to compound (10) provided herein).

[0117] For instance, the provided compounds can be readily modified for targeted delivery to a tumor cell. For example, although compounds (1) and (2) do not contain a moiety that can be conjugated to a targeting means, the skilled artisan can modify these compounds to add a conjugatable moiety such as an alcohol (e.g., a phenol or aliphatic alcohol), a carboxylic acid, a carbonyl (e.g., an aldehyde or ketone), an amine, an amide or a thiol. As shown herein, the morpholine ring of compound (1) is predicted to be distal to the A and B pockets and therefore may be amenable to modification with a conjugatable moiety. Similarly, the phenyl ring of compound (2) is predicted to be distal to the A and B pockets and so may be amenable to modification with a conjugatable moiety. Accordingly, where (C) is compound (1) or compound (2), the drug may be:(1) (2)

[0118] Computational modelling (employing, e.g., Boltz-2) may be used to predict if a drug -like small-molecule compound (e.g., a compound comprising a conjugatable moiety in addition to an amide or urea framework, or a cyclic or hydrophobic moiety, depending on the binding pocket) retains binding properties equivalent to a compound that is identified as being capable of binding to and altering the peptide binding preference of an MHC.Suitable properties of the drug

[0119] The drug is generally a small molecule that fulfils Lipinski’s “Rule of Five” (e.g., 5 or fewer hydrogen bond donors, 10 or fewer hydrogen bond acceptors, a molecular weight less than 500 Da, and / or a log P that does not exceed 5). The drug may have a molecular weight of < 350 Daltons (Da). The drug may have a LogP < 3.5. The number of hydrogen bond donors in the drug may be < 3. The number of hydrogen bond acceptors in the drug may be < 3.

[0120] The drug may have a disassociation constant (KD) in the micromolar range for the relevant MHC binding pocket. Alternatively, the drug may have a KD in the nanomolar range for the relevant MHC binding pocket. For example, the KD of the drug may be in the range of 0.5 pM to 30 nM. The KD for the drug can be predicted computationally, e.g., using Boltz-2.

[0121] Although abacavir was used as a model drug herein for inducing an immunopeptidome shift, the compositions and methods described herein may not comprise abacavir as the drug that is capable of altering the peptide binding preference of an MHC.Drug conjugates

[0122] Drugs that are capable of altering the peptide binding preference of an MHC (such as those listed in Table 3) have been associated with adverse events and systemic toxicities such as hypersensitivity reactions, which are believed to be the result of a polyclonal CTL response following a drug-induced immunopeptidome shift. Serious hypersensitivity reactions have been observed in individuals carrying specific HLA alleles. Accordingly, such drugs are contraindicated in individuals with the specific alleles encoding MHC-I and / or MHC-II molecules with which the drugs interact with to induce an immunopeptidome shift. In contrast, the present invention leverages this CTL response to reduce or eliminate tumor cells.

[0123] Abacavir has been shown to reduce tumor growth of an HLA-B* 57: 01 -expressing murine cancer cell line in an HLA-B*57:01 transgenic mouse (Susukida et al. Biol Pharm Bull. 2025;48(10): 1566-1571). Repetitively administering an immunopeptidome shifter such as abacavir on its own to human subjects suffering from cancer does not appear to be a feasible antitumor therapy, given that abacavir has been documented to elicit an autoimmune attack against the cells of individuals that carry the HLA-B*57:01 allele.

[0124] By using a targeting means that directs the drug to a tumor cell (e.g., a targeting moiety such as an antibody), the off-target toxicity (e.g., in form of a systemic hypersensitivity reaction) is mitigated. Accordingly, a drug conjugate or liposome can target the drug to tumor cellsand thus can induce a local, tumor-specific polyclonal CTL response. The drug is capable of binding to and altering the peptide binding preference of the MHC expressed by the tumor cells.Targeting means

[0125] Targeting means that are capable of directing drugs to a tumor cell are known in the art. Such targeting means include antibodies and other proteins or peptides (e.g., proteins or peptides that act as ligands for cell surface receptors), other antigen-binding molecules besides antibodies such as aptamers, as well as small molecules (e.g., small-molecules that act as ligands for cell surface receptors). For example, a targeting means may bind a tumor-specific or tumor- associated antigen (e.g., a cell surface receptor or other cell surface-expressed protein). Typically, such tumor-specific or tumor-associated antigens are expressed on the surface of the tumor cell. Upon binding of the targeting means to the antigen on the cell surface, the drug conjugate or liposome is internalized and enters the endosomal pathway of the tumor cell.

[0126] For example, included herein are means for binding tumor-specific or tumor- associated antigens, wherein such means is an antibody or an antigen-binding portion thereof (“antigen-binding means” or “antibody means for binding”). Structures corresponding to the antigen-binding means for binding tumor-specific and tumor-associated antigens, include the antibodies disclosed herein, including those listed in Table 4, and structural variants thereof. Equivalents include antibodies that differ in their amino acid sequence, particularly in the complementary binding region (CDR) of the antibody but perform the same function of binding the tumor-specific or tumor-associated antigen, thereby achieving the same result of targeted delivery of the drug to tumor cells.Antibody-drug conjugates (ADC)

[0127] Antibodies that are capable of specifically binding a tumor-specific or tumor- associated antigen expressed on the surface of a tumor are particularly suitable targeting moieties for described drug conjugates. Accordingly, in a further aspect, an antibody-drug conjugate (ADC) is provided, wherein the antibody is capable of specifically binding a tumor-specific or tumor- associated antigen, and the drug is capable of binding to and altering the peptide binding preference of a major histocompatibility complex (MHC) expressed by a tumor cell.

[0128] The antibody of the ADC is chosen to selectively bind to a tumor-specific or tumor- associated antigen on the surface of a tumor cell to achieve tumor selectivity and minimize adverse effects associated with the drug. As shown schematically in FIG. 5, following binding of the antibody 431 to the surface antigen 440, the ADC 430 is internalized and enters the endosomalpathway. As the endosome matures through fusion with lysosomes, the drug 414 is released from the antibody 431 and can diffuse into the endoplasmic reticulum 407 where it can induce a localized immunopeptidome shift to trigger a CTL response against the tumor cell.

[0129] Several ADCs for use in treating cancer are commercially available. These ADC technologies can be readily adapted for use in the methods and compositions disclosed herein by replacing the cytotoxic drug in the ADC with a drug that can alter the peptide binding preference of an MHC (such as those listed in Table 3). Currently, there are at least 11 FDA-approved ADCs for use in the treatment of cancer. These are listed in Table 4. Many others are undergoing clinical investigation.Table 4. FDA-approved ADCs

[0130] Typically, an ADC comprises an antibody (e.g., a full-length antibody or an antigen-binding fragment such as a scFv) connected via linker to a drug (sometimes referred to as the “payload”).Antibody

[0131] Antibodies can be produced, for example, by hybridoma technologies, recombinant technologies, phage display technologies, synthetic technologies (e.g., CDR or specificitydetermining residue, SDR, grafting), or combinations of such or other technologies known in the art. The use of hybridoma technologies typically involves immunization of an animal (e.g., a mouse) with the extracellular portion of the tumor-specific or tumor-associated antigen. A variety of well-known methods and tools can be used for producing and purifying the antibody, including vectors, for example, plasmids, viruses, or other vehicles for polynucleotide insertion or expression, and hosts, for example, microbial, yeast, insect, and mammalian organisms (see, e.g., Process Scale Purification of Antibodies, Uwe Gottschalk, ed., 2d ed. 2017).

[0132] Typically, the antibody of an ADC described herein is a monoclonal antibody. The antibody of an ADC described herein may be a bispecific antibody. The antibody may be a singledomain antibody (e.g., a nanobody). The antibody may be a single domain antibody comprising an antibody fragment (e.g., a single-chain Fv).

[0133] The antibody may comprise an Fc region. An antibody described herein may be provided with an Fc region for various reasons. For example, an Fc region may be included to extend the half-life of an antibody in circulation.

[0134] Suitable Fc regions for use with the antibody include Fc regions of an IgG subclass antibody (referred herein as “IgG Fc region”). The IgG Fc region may be human. The IgG Fc region may be humanized.

[0135] The IgG Fc region may be of the IgGl subclass (e.g., human IgGl). Or the IgG Fc region may be of the IgG2 subclass (e.g., human IgG2). Or the IgG Fc region may be of the IgG3 subclass (e.g. , human IgG3). Or the IgG Fc region may be of the IgG4 subclass (e.g. , human IgG4).

[0136] The Fc region may be modified, e.g., to reduce or eliminate one or more effector functions. Effector functions may not be desirable when the antibody is primarily used for targeting the drug to a tumor cell. Moreover, Fc receptor (FcR) binding may result in the delivery of the ADC to Fc receptor-expressing non-tumor cells. Induction of an immunopeptidome shift and resulting CTL response against these cells is typically not desired. Modifications that reduce or eliminate one or more effector functions typically also reduce or eliminate binding to FcR. Accordingly, the antibody of the ADC may comprise an Fc region that is modified to reduce or eliminate one or more effector functions.

[0137] Mutations in the Fc region that result in reduced or eliminated effector function include LALA mutations (typically referred to as L234A / L235A, optionally further comprising K322A). Accordingly, the Fc region may comprise LALA mutations. Effector functions such as antibody-dependent cell-mediated cytotoxicity (ADCC) are induced through binding of an antibody to one or more FcRs such as FcRI, FcRII, and / or FcRIIIa. PG mutations (typically referred to as P329G) can reduce or eliminate binding to FcR. Accordingly, the Fc region may comprise a PG mutation. The LALA and PG mutations may be combined, i.e., the Fc region may comprise the LALA mutations and the PG mutation. Accordingly, the antibody of the ADC may include an Fc region that comprises LALA mutations and / or PG mutations.

[0138] The antibody may have a disassociation constant (KD) in the nanomolar range. Alternatively, the antibody may have a KD in the picomolar range. For example, the antibody may have a KD of about 100 nM or less, e.g., from about 0.001 nM to about 100 nM, or from about 0.01nM to about 10 nM. The antibody may have a KD of about 10 nM or less, e.g., from about 0.01 nM to about 10 nM. The KD can be determined using a cell -based assay, an Enzyme Linked Immunosorbent Assay (ELISA), or Biolayer interferometry (BLI).

[0139] Included herein are means for binding a tumor-specific or tumor-associated molecule, wherein such means is selected from an antibody, a ligand, or a peptide.

[0140] Other parameters for consideration of a suitable antibody for use in a drug conjugate (i.e., an ADC) include aggregation, thermostability, polyreactivity and sequence liabilities. These parameters are unrelated to the binding function but are preferred for, for example, manufacturability. Thermostability may be examined using full IgG antibodies. It is desirable that an antibody displays a Fab stability of >68°C during thermostability testing. Polyreactivity can be assessed by incubating an antibody with KLH, LPS and / or human insulin. It is desirable that an antibody described herein has minimal polyreactivity, e.g., does not react with keyhole limpet hemocyanin (KLH), lipopolysaccharide (LPS) and / or human insulin.

[0141] It is desirable that an antibody for use in a drug conjugate (i.e., an ADC) is free of sequence liabilities including those that may be involved in the formation of aggregates. For example, two or more consecutive positively charged residues (e.g., RR or RKR) can contribute to the “stickiness” of antibodies, rendering them less desirable for therapeutic applications. The formation of aggregates can be assessed by incubating an antibody in a suitable storage buffer (e.g. , PBS), for 2-4 weeks (e.g., 4 weeks) at 4°C. Aggregation can be considered minimal when, e.g., 10% or less of the antibody is present as high molecular weight aggregates at the end of the incubation period as, assessed, e.g., by mass photometry. Other sequence liabilities include amino acid residues that may be subjected to post-translational modifications (PTMs). PTMs can affect affinity, stability, potency, and / or homogeneity of an antibody. PTMs can include deamidation, isomerization, oxidation, N-linked glycosylation, free thiol, pyro-Glutamate, and C-terminal lysine.Tumor-specific and tumor-associated antigens

[0142] A targeting means may specifically bind to a tumor-specific or tumor-associated antigen. The targeting means may bind any known tumor-specific or tumor-associated known antigen.

[0143] As a consequence of the accumulation of mutations in a tumor, the tumor cell can express tumor-specific antigens. A tumor-specific antigen is found on tumor cells only and is therefore considered advantageous for therapeutic targeting. Such antigens can originate as a resultof, e.g., genomic mutations including gene fusions, aberrant splicing, post-translational modifications (PTMs), and insertion of a viral open reading frame in a tumor cell.

[0144] Alternatively, a tumor cell may abnormally express a non-mutated protein (e.g., a cell surface receptor or other cell surface-expressed antigen). A tumor-associated antigen may be expressed at least 10-fold higher on a tumor cell compared to a normal cell, e.g., at least 10-fold, at least 20-fold or at least 30-fold.

[0145] The tumor-associated antigen may be associated with various cancers, e.g., EGFR. The tumor-associated antigen may be associated with a specific type, e.g., PSA. The tumor- associated antigen may be a protein or expressed on the surface of a tumor cell. Alternatively, the tumor-associated antigen may be a glycan expressed on the surface of a tumor cell (e.g., Tn and sTn).

[0146] Exemplary tumor-associated antigens that the targeting means (e.g., a targeting moiety such as an antibody of an ADC described herein) may bind to include PSA (prostate specific antigen), HER2 (human epidermal growth factor receptor 2), HER3 (human epidermal growth factor receptor 3), MAGE (melanoma-associated antigen), CD33, CD30, CD22, CD79b, CD19, Claudin 18.2, Nectin-4, B cell maturation antigen (BCMA), TROP2, tissue factor (TF), folate receptor-a (FR-a), CEA, CEACAM5, GPA33, GUCY2C, immature laminin receptor, IL-13Ra2, B7-H3, TAG-72, EpCAM, EphA3, mesothelin, SAP-1, survivin, PRAME, gplOO, TRP1, TRP2, MC1R, p-catenin, MUC1, EGFR (e g., EGFRvIII), GD2, GPC3, CD171, ErbB, PSMA (prostate specific membrane antigen), TERT, and Tn (CD 175), sTn (CD 175s). For example, overexpression of HER2 can be observed in some breast cancers, and elevated PSA and / or PSMA expression is associated with prostate cancer. Similarly, overexpression of GD2 can be observed in neuroectodermal-originating cancers, e.g., neuroblastoma, melanoma, and retinoblastoma, as well as breast, bladder and glioblastoma.

[0147] Several tumor-associated antigens have been used to target drugs to the cells of a cancer. Table 5 provides an overview of various cancers and the tumor-associated antigens that have been used to target them, e.g., to deliver a targeted cancer therapy.Table 5. Cancers and turn or- associated antigensDrug to antibody ratio

[0148] In an ADC, the drug-antibody ratio (DAR) typically refers to the average number of drug molecules conjugated to the antibody. It is well understood that the drug to antibody ratio (DAR) is an important parameter of an ADC. The optimal DAR for a given ADC may vary, e.g., dependent on the type of antibody, drug, and / or conjugation site that is / are used.

[0149] Typically, the DAR is between about 1 to about 20. The DAR may be between about 1 to about 16. The DAR may be between 1 to 10 (e.g., about 4-8). The DAR is about 1, about 2, about 4, about 6, about 8, or about 10. A DAR of 1, 2, or 4 may be suitable to achieve the desired therapeutic effect, e.g., inducing an immunopeptidome shift in a tumor cell or inducing a polyclonal CTL response against a tumor cell.

[0150] For example, it has been found that some T cells are capable of recognizing and destroying cells expressing only a few copies of peptide-MHC complexes (Akitsu et al., Sci Adv. 2024 Aug 14; 10(33)). Accordingly, a DAR as low as 1 or 2 may be enough in some circumstances to induce an effective CTL response against tumor cells. A low DAR may further help to reduce any off-target effects of the ADC.

[0151] Most individuals are heterozygotes at each HLA locus. Accordingly, there are up to six distinct alleles that may be targeted by six individual drugs, e.g. , individually, concomitantly, or sequentially. The ADC may comprise one drug. Alternatively, the ADC may comprise more than one drug, e.g., two different drugs. The two or more different drugs may alter the peptide binding preference of MHCs encoded by distinct alleles. By targeting two or more different MHC’ s encoded by different alleles, a (potential) loss of HLA heterozygosity in tumor cells may be prevented or overcome.Linker

[0152] A drug conjugate typically comprises a linker that joins the drug to the targeting means. As the endosome matures through fusion with lysosomes, the drug is released (e.g., by cleavage of the linker in the lysosome) and can diffuse into the endoplasmic reticulum to induce a localized immunopeptidome shift and trigger a CTL response against the tumor cell.

[0153] Accordingly, the drug capable of inducing an immunopeptidome shift in a cell may be coupled (“conjugated”) via a linker to a targeting means to form the drug conjugate. For example, the targeting moiety may be an antibody that specifically binds a tumor-specific or tumor- associated antigen expressed on the surface of the tumor cell in order to form an ADC.

[0154] Both cleavable and non-cleavable linkers have been proven to be safe in preclinical and clinical trials. For example, the enzyme-sensitive cleavable linker in brentuximab vedotin (ADCETRIS) is used to deliver antimicrotubule agent monomethyl auristatin E (MMAE), a synthetic antineoplastic agent, to CD30-positive tumor cells.

[0155] A cathepsin-cleavable tetrapeptidyl-aminomethoxy linker in trastuzumab deruxtecan (ENHERTU) is used to link the topoisomerase I inhibitor deruxtecan to trastuzumab.

[0156] A non-cleavable linker is employed in trastuzumab emtansine (KADCYLA), another approved ADC, to conjugate the microtubule-formation inhibitor mertansine (DM-1), a derivative of maytansine to the antibody trastuzumab.

[0157] Linkers may include disulfides, hydrazones, thioethers, or peptides. Peptide-based linkers can be rendered cleavable by the incorporation of a site that is cleaved by an enzyme present in the endosome or lysosome (e.g., cathepsin). A well-established cleavable linker is valine- citrulline-p-aminobenzyl carbamate. Another well-established cleavable linker is 6- maleimidohexanoyl-[Gly-Gly-Phe-Gly (SEQ ID NO: 2)]-OH (Mc-[GGFG (SEQ ID NO: 2)]-OH; CAS number 2413428-36-9). Linking the drug to the targeting means (e.g., a targeting moiety such as an antibody) can prevent drug activity outside of target cells, thereby reducing off-target effects.

[0158] Accordingly, the drug may be conjugated to the targeting means (e.g., targeting moiety such as an antibody) via a cleavable linker. A cleavable linker may be an enzyme cleavable linker. The enzyme cleavable linker may be a peptide-based linker. The enzyme cleavable linker may comprise a protease cleavage site. The protease may be a lysosomal enzyme (e.g., cathepsin, P-glucuronidase, sulfatase, phosphatase, or legumain). A cleavable tetrapeptidyl-aminomethoxy linker may be used in the drug conjugates (e.g., ADCs) described herein. For example, the linker may be an Mc-[GGFG (SEQ ID NO: 2)]-OH peptide-based linker.

[0159] A cleavable linker herein may be a chemically labile linker (e.g., acid-cleavable or reducible). Such a linker may be cleaved when encountering an acid (or acidic pH) or reducing environment (e.g., in an endosome or lysosome).

[0160] Alternatively, the drug may be conjugated to the targeting means (e.g., a targeting moiety such as an antibody) via a non-cleavable linker.

[0161] The linker may be a bond.

[0162] The linker may be a Ci to C20 alkylene chain, optionally a C5 to C15 alkylene chain. The alkylene chain may be interrupted one or more -O-, -S-, -N(R')-, -C=C-, -C(O)-, -C(O)O- -C(O)NR'-, group, wherein R' is H, or Me. The alkylene chain interrupted one or more -O- may be a polyethylene glycol (PEG) chain, optionally comprising 1-10 PEG units.

[0163] The linker may be a peptide chain.

[0164] The term “means for linking” describes the corresponding linkers described herein that link the desired targeting means (e.g., a targeting moiety such as an antibody) to the drug for targeted delivery to tumor cells. The “means for linking” may be cleavable or non-cleavable.Structure of the drug conjugate

[0165] A drug conjugate can be represented by Formula I:M - Ln -Dn, whereinM is a targeting means, e.g., a means for binding a tumor-specific or tumor-associated antigen expressed on the surface of a tumor cell (e.g., an antibody),L is a means for linking (e.g., a cleavable or non-cleavable linker),D is a means for altering the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell (e.g., a drug moiety such as small-molecule drug), and n is between about 1 to about 20. For example, n may be between about 1 to about 16, e.g. , between 1 to 10 or about 4-8. In some instances, n is about 1, about 2, about 4, about 6, about 8, or about 10.

[0166] The drug conjugate represented by Formula I may have a structure M - [L -D]n, such that each D group is connected to the M group via an L group.

[0167] Suitable targeting means such as means for binding a tumor-specific or tumor- associated molecule (e.g., antigen) expressed on the surface of a tumor cell, as well as suitablemeans for linking the targeting means and the means for altering the peptide binding preference of an MHC, are described herein, e.g., in Table 4. Suitable means for altering the peptide binding preference of an MHC are described in Table 3.

[0168] In some instances, the means for binding to and altering the peptide binding preference of an MHC may further comprise an ER-targeting means.Liposomal delivery vehicles

[0169] As an alternative to an ADC, a drug capable of altering the peptide binding preference of an MHC may be delivered by a liposome. The use of liposomal particles (e.g., liposomes or lipid nanoparticles) for the delivery of drugs (especially small -molecule drugs) is well known in the art. The liposomal delivery vehicle typically encapsulates the drug.

[0170] Liposomal delivery vehicles are commonly prepared from a mixture of lipids. This mixture commonly includes a PEGylated lipid that is capable of prolonging the half-life of the liposomal delivery vehicle in circulation. The presence of a PEGylated lipid also prevents clearance by the mononuclear phagocyte system. The polyethylene glycol (PEG) moiety of the PEGylated lipid may be about 1-3 kDa (e.g., about 2 kDa). It is typically covalently attached to a lipid with an alkyl chain length of 16 to 18 carbons to render the PEGylated lipid diffusible, thereby eventually enabling the uptake by tumor cells. To minimize uptake by the liver and spleen and elimination by the kidneys, liposomes may be provided that have a neutral or negative charge. A liposomal delivery vehicle may be prepared at a size suitable for extravasation. For example, a suitable liposomal delivery vehicle may be about 60-180 nm (e.g., about 80-160 nm) in size.

[0171] Such liposomal delivery vehicles and methods for producing them are well known in the art. Any suitable liposomal delivery vehicle is contemplated for use in the compositions and methods described herein.

[0172] In comparison to a drug conjugate, a liposomal delivery vehicle may be capable of delivering a larger amount of a drug capable of altering the peptide binding preference of an MHC to a tumor cell. For example, a single ADC typically is capable of delivering about 2-8 drug molecules, whereas a single liposomal delivery vehicle may be capable of carrying 1,000-10,000 drug molecules (depending on the size of the drug molecule).Passive targeting

[0173] Passive targeting using a liposomal delivery vehicle (e.g., liposome) is an established drug delivery strategy that exploits the enhanced permeability and retention (EPR)effect of such vehicles. Tumor targeting is achieved through the accumulation of drug-comprising liposomes in the tumor microenvironment (TME). Accordingly, this delivery strategy does not need to rely upon a tumor-specific targeting means such as an antibody to accumulate at the site of a tumor. The liposomal delivery vehicle is internalized by tumor cells and releases the drug into the cytosol from where it can diffuse to the ER and bind to MHC molecules in order to induce the desired immunopeptidome shift and the subsequent anti-tumor effect resulting from a polyclonal CTL response.

[0174] Liposomal delivery vehicles that reach a tumor via passive targeting have been employed as cancer therapeutics. Exemplary FDA-approved liposomal drug products for the treatment of cancer are provided in Table 6 (Gatto et al. Life (Basel). 2024 May 24;14(6):672).Table 6. FDA-approved liposomal drug products for cancer

[0175] To avoid the induction of a systemic polyclonal CTL response (e.g., an autoimmune response against non-tumor cells), the liposomal delivery vehicle may be administered directly to a tumor. For example, the liposomal delivery vehicle may be delivered intratumorally. Alternatively, the liposomal delivery vehicle may be delivered to a tumor via a blood vessel that directly supplies the tumor.Active targeting

[0176] More typically, a liposomal delivery vehicle is actively targeted to cells of a tumor. Liposomal delivery vehicles used for active targeting of tumor cells comprise a targeting means. A suitable targeting means may be an antibody or a ligand. Suitable targeting means are disclosed herein and include, e.g., the antibodies listed in Table 4. For example, the targeting means (e.g., the antibody or ligand) may specifically bind to a tumor-specific or tumor-associated antigen as provided herein (see, e.g., Table 5).

[0177] A liposome comprising an antigen-binding molecule such as an antibody may be referred to as an “immunoliposome”. For instance, an anti-tumor response was observed against a melanoma murine cell model using an immunoliposome loaded with doxorubicin and functionalized with monovalent-variable fragments (Fab’) of anti-PD-Ll (Merino et al. J Nanobiotechnology. 2021 Apr 13; 19(1): 102).

[0178] Ligand-conjugated liposomes have been designed, and their utility for active targeting of tumor cells has been demonstrated. For example, an estrogen-targeted oxaliplatin- loaded PEGylated liposome was used for targeted delivery to estrogen-receptor (ER)-expressing tumor cells and yielded anti-tumor efficacy (Sun et al. Int J Nanomedicine. 2021 Dec 23;16:8279- 8303).Moiety targeting the endoplasmic reticulum

[0179] The loading of self-peptides into MHC molecules occurs in the endoplasmic reticulum (ER). Accordingly, it may be advantageous to increase the amount of drug that is available at the ER. For example, the compositions described herein may optionally be modified to promote targeting of the drug to the ER such that the drug is present at high concentrations during peptide loading.

[0180] Accordingly, in some instances, the compositions described herein may include a small-molecule moiety that targets the drug to the ER. Such ER-targeting means may be linked covalently to the drug, e.g., via a means for linking (e.g., a linker as described in the preceding section).

[0181] Means and moieties that can target a small molecule (e.g., a drug) to the ER are known in the art and include tosyl-containing moieties such as N-tosylethylenediamine and tosylglycine, as well as sulfonamide-containing moieties such as N-(2-aminoethyl)-5- (dimethylamino)naphthalene - 1 -sulfonamide .

[0182] For example, sulfonamide moiety (a sulfonyl ligand) has been widely used in ER- targeting probes and drugs due to its excellent selectivity and affinity for the ER. Sulfonylurea and its derivatives can bind to the sulfonylurea receptors of adenosine triphosphate (ATP)-dependent potassium channel, which is highly expressed on the endoplasmic membrane. Two typical sulfonamide ligands, p-toluenesulfonamide and N-(2-aminoethyl)-5-(dimethylamino)naphthalene- 1 -sulfonamide, are often incorporated into ER-targeting nanosystems as selective binding units for the ER.

[0183] Sulfonamide ligands, such as p-toluenesulfonamide, are capable of binding to KDEL receptors present in endosomes. A drug described herein that comprises an ER retention moiety, such a sulfonamide ligand, can be bound by a KDEL receptor and induce retrograde transport to the ER. Providing a means of targeting the ER allows the drug to accumulate in the ER, even if it is not sufficiently hydrophobic to efficiently cross internal membranes. Alternatively, chemical modifications may be made to the drug to improve its hydrophobicity.

[0184] The drug conjugates described herein may comprise an ER-specific linker. For example, the linker may be specifically cleaved by ER-localized enzymes, e.g., ER-specific proteases or peptidases such as ERAP1. The ER has an oxidative environment for the formation of disulfide bonds for protein maturation and stability. Accordingly, the linker may be specifically cleaved in ER-associated conditions, e.g., within the oxidizing environment of the ER.Pharmaceutical compositions

[0185] A composition (e.g., a pharmaceutical composition) comprising a drug conjugate (e.g. , an ADC) or a liposome comprising the drug may comprise one or more of a pharmaceutically acceptable carrier, diluent, or excipient.

[0186] The pharmaceutically acceptable carrier, diluent, or excipient may be conventional (e.g., as described in Remington, The Science and Practice of Pharmacy, 22ndEdition, Loyd V., ed., Pharmaceutical Press, 2012). In general, the nature of the carrier depends on the mode of administration. For instance, parenteral formulations typically comprise injectable fluids that include pharmaceutically and physiologically acceptable fluids, such as water for injection, physiological saline, balanced salt solutions or the like as a vehicle. Pharmaceutical compositions can additionally include minor amounts of non-toxic auxiliary substances for stability (e.g., one or more buffering agent). Water of unknown provenance is generally not considered a pharmaceutically acceptable vehicle because it is not suitable for administration to a subject.

[0187] The carrier may be sterile and / or suspended or otherwise contained in a unit dosage form including one or more measured doses of the composition suitable for administration to a subject of an effective amount of a drug conjugate (e.g., an ADC) or a liposome comprising the drug described herein. Medications for use in therapy may also be included. The unit dosage form may be in a sealed vial that contains sterile contents or a syringe for injection into a subject, lyophilized for subsequent solubilization and administration.

[0188] A therapeutically effective amount of the drug conjugate (e.g., ADC) or the liposome comprising the drug may be diluted in a buffer suitable for infusion into a subject. Theinjectable suspension may be provided in one or more infusion bag(s) for intravenous administration to the subject. The infusion bag may be made of ethylene vinyl acetate (EVA). In some embodiments, the pharmaceutical composition is provided in a solid or controlled release dosage form.

[0189] A composition (e.g., a pharmaceutical composition) comprising a therapeutically effective amount of a drug conjugate (e.g., an ADC) or a liposome comprising the drug described herein may be formulated for parenteral administration, e.g, intramuscular, intravenous, subcutaneous, intraperitoneal, or intradermal administration. A composition comprising a therapeutically effective amount of a drug conjugate (e.g., an ADC) or a liposome comprising the drug described herein may be formulated for intranasal or inhalation administration.Methods of treatment

[0190] A method of inducing an immunopeptidome shift in a tumor cell of a subject in need thereof is provided, that comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

[0191] A method of inducing an immunopeptidome shift in a tumor cell of a subject in need thereof is provided, that comprises administering to the subject a therapeutically effective amount of a liposome comprising a drug that binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell, optionally wherein the liposome comprises a targeting means (e.g., an antibody or a ligand) that directs the liposome to the tumor cell.

[0192] Furthermore, a method of inducing a polyclonal cytotoxic T lymphocyte (CTL) response against a tumor cell of a subject in need thereof is provided that comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

[0193] Also provided is a method of inducing a polyclonal cytotoxic T lymphocyte (CTL) response against a tumor cell of a subject in need thereof is provided that comprises administering to the subject a therapeutically effective amount of a liposome comprising a drug that binds to andalters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell, optionally wherein the liposome comprises a targeting means (e.g., a targeting moiety such as an antibody or ligand) that directs the liposome to the tumor cell.

[0194] By altering the MHC’s peptide binding preference, the tumor cell presents non- tolerized self-peptides in the MHC and thereby induces a polyclonal CTL response against the tumor cell of the subject.

[0195] Accordingly, also provided is a method of treating a tumor in a subject in need thereof that comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the cells that form the tumor, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the cells of the tumor.

[0196] Also provided herein is a method of treating a tumor in a subject in need thereof that comprises administering to the subject a therapeutically effective amount of a liposome comprising a drug that binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell, optionally wherein the liposome comprises a targeting means (e.g., a targeting moiety such as an antibody or ligand) that directs the liposome to the cells that form the tumor.

[0197] In addition, a drug conjugate is provided for use in a method of treating a tumor in a subject in need thereof wherein the drug conjugate comprises a targeting means (e.g., a targeting moiety such as an antibody) that directs the drug conjugate to the cells that form the tumor, and the drug is capable of binding to and altering the peptide binding preference of a major histocompatibility complex (MHC) expressed by cells of the tumor.

[0198] Furthermore, a liposome is provided for use in a method of treating a tumor in a subject in need thereof wherein the liposome comprises a drug that binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell, and optionally a targeting means (e.g., a targeting moiety such as an antibody or ligand) that directs the liposome to the cells that form the tumor.

[0199] In some instances, the drug conjugate may be an antibody drug conjugate (ADC), wherein the antibody specifically binds a tumor-specific or tumor-associated antigen expressed on the surface of the tumor cell. In other instances, the liposome may comprise an antibody, wherein the antibody specifically binds a tumor-specific or tumor-associated antigen expressed on the surface of the tumor cell.Dosing and administration

[0200] The drug conjugate (e.g., ADC) or the liposome comprising the drug may be administered as a single dose or as multiple doses. The dose of the drug conjugate (e.g., ADC) or the liposome comprising the drug may be determined by the condition of the subject and the body weight of the subject to receive the drug conjugate (e.g., ADC) or the liposome. The dose of the drug conjugate (e.g., ADC) or the liposome comprising the drug may also be determined by the existence, nature and extent of any adverse side effects that might accompany the administration of the drug conjugate (e.g., ADC) or liposome. Typically, the attending physician will decide the dose of the drug conjugate (e.g., ADC) or liposome with which to treat each individual patient, taking into consideration a variety of factors, such as age, body weight, general health, diet, sex, route of administration, and the severity of the condition being treated.

[0201] In some instances, it may be useful to prime the subject’s naive T cells with the drug alone (i.e., not in the form of a drug conjugate or liposome) prior to treatment with the drug conjugate or liposome. Without wishing to be bound by theory, it is contemplated that priming a subject’s naive T cells may result in an enhanced polyclonal CTL response against tumor cells of the subject when the drug conjugate (e.g., ADC) or liposome is administered. It is known that subjects with an HLA-B*57:01 allele respond to abacavir treatment with a hypersensitivity reaction only after having received about 15 or more previous doses of the drug (e.g., 16 doses, 17 doses, 18 doses, 19 doses, or 20 doses). This indicates that naive T cells are primed by abacavir treatment and only once a critical amount of primed T cells exist does abacavir (at about the 16thdose) trigger a hypersensitivity reaction.

[0202] Accordingly, in some instances, a subject may receive one or more (e.g., two or more) doses of a therapeutically effective amount of a drug that can bind to and alter the peptide binding preference of an MHC expressed by the tumor cells of the subject. Subsequently, the subject is then administered a therapeutically effective amount of a conjugate or liposome that comprises the drug and a targeting means that directs the drug conjugate or liposome to the tumor cells of the subject. Thus, in some aspects, a method of inducing a polyclonal CTL response against a tumor cell of a subject in need thereof is provided that comprises (i) administering a therapeutically effective amount of a drug that can bind to and alter the peptide binding preference of a MHC expressed by the tumor cell of the subject, and (ii) administering a therapeutically effective amount of a conjugate that comprises the drug and a targeting means (e.g., an antibody) that directs the drug conjugate (e.g., ADC) to the tumor cells of the subject. In other aspects, a method of inducing a polyclonal CTL response against a tumor cell of a subject in need thereof isprovided that comprises (i) administering a therapeutically effective amount of a drug that can bind to and alter the peptide binding preference of a MHC expressed by the tumor cell of the subject, and (ii) administering a therapeutically effective amount of a liposome that comprises the drug and optionally a targeting means (e.g., an antibody or ligand) that directs the liposome to the tumor cells of the subject.

[0203] The therapeutically effective amount of the drug may be administered as multiple doses over a period of time sufficient to prime naive T cells. For example, the therapeutically effective amount of the drug may be administered as at least two (e.g., three to fifteen, or five to ten) doses over a period of at least two (e.g., three to fifteen, or five to ten) days. This is then followed by a single dose or multiple doses of the drug conjugate (e.g., ADC) or the liposome.

[0204] Clinical efficacy of the therapeutic treatment with a drug conjugate (e.g., ADC) or liposome comprising the drug described herein can be determined by measuring the clinical benefit rate (CBR). The clinical benefit rate is measured by determining the sum of the percentage of patients who are in complete remission (CR), the number of patients who are in partial remission (PR) and the number of patients having stable disease (SD) at a time point at least 6 months out from the end of therapy. The CBR for a particular therapeutic regimen is at least 25%, 30%, 35%, 40%, 45%, 50%, 55%, 60%, 65%, 70%, 75%, 80%, 85%, or more.

[0205] An effective amount may also be assessed by an improvement in important cancer endpoints, e.g., Overall Survival (OS), Disease-Free Survival (DFS), Objective Response Rate (ORR), Complete Response Rate (CRR), or Progression Free Survival (PFS). See Dept, of Health and Human Services, Food and Drug Admin, Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologies: Guidance for Industry (2018); E.A. Eisenhauer et al., New Response Evaluation Criteria in Solid Tumors: Revised RECIST Guideline (Version 1.1), 45 Eur. J. Cancer 228 (2009).

[0206] The degree of improvement in these endpoints that is clinically meaningful will vary on the tumor or cancer to be treated and other considerations by the tending physician. The OS or PFS is an anticipated relative improvement of at least 20% or at least 25% and an absolute increase of at least 2.5 months (e.g. , 6 months) in PFS and / or OS compared with that achieved with standard-of-care treatments. See Kumar et al., An Appraisal of Clinically Meaningful Outcomes Guidelines for Oncology Clinical Trials, JAMA Oncol. 2016 September 01; 2(9): 1238-1240; Ellis et al., American Society of Clinical Oncology perspective: Raising the bar for clinical trials by defining clinically meaningful outcomes, J Clin Oncol. 2014 April 20; 32(12): 1277-80.Testing for MHC expression

[0207] A drug conjugate (e.g., ADC) or a liposome comprising a drug as described herein can turn a subject’s “cold” tumor into a “hot” tumor by enabling recognition of tumor cells by the immune system of the subject. Of course, a prerequisite for this approach is that the cells of the tumor express an MHC such that the drug comprised in the drug conjugate (e.g. , ADC) or liposome can shift the immunopeptidome, i. e. , alter the peptide binding preference to present peptides that can be recognized by T cells.

[0208] The skilled person will appreciate that tumor cells are capable of evading the immune system through the loss of MHC (Dhatchinamoorthy et al. Front Immunol. 2021 Mar 9: 12:636568). For example, an MHC-low phenotype has been observed in many of the most frequent human cancers including non-small cell lung cancer (NSCLC), breast cancer, prostate cancer, colorectal cancer, head and neck squamous cell carcinoma (HNSC), hepatocellular carcinoma, and melanoma. The loss of MHC-I expression can vary for different types of these cancers and may range from 0-93%. Loss of expression of a single HLA class I allele has also been reported.

[0209] Many cancers downregulate MHC-I peptide presentation broadly. Empty MHC-I molecules are unstable without chaperone-binding and retained in the endoplasmic reticulum. Thus, loss of the MHC-I heavy chain, [32-microglobulin, immunoproteasome subunits, TAP, Tapasin, or ERAP1 can all result in a loss of MHC-I expression on the tumor cell surface. For example, loss of TAP expression ranging from about 10-80% has been documented in colorectal, renal cell, and cervical cancers, as well as melanomas.

[0210] Therefore, before a subject is treated with a drug conjugate (e.g., ADC) or a liposome comprising the drug using a method described herein, it may be determined whether a subject’s tumor expresses MHC to ascertain whether the subject is likely to respond to such treatment. For example, a sample of tumor cells obtained from the subject may be contacted with one or more antibodies that bind to the monomorphic determinants on the heavy chains of MHC-I molecules (HLA -A, HLA-B, and HLA-C) or to [32 -microglobulin to ascertain that tumor cells express MHC-I. Any known method to determine protein expression on a cell may be used.

[0211] For instance, one or more antibodies that specifically bind to HLA-A, HLA-B, HLA-C, and [32-microglobulin, respectively, may be used in a method for determining whether a subject’s tumor cells express MHC-I and therefore the subject is likely to benefit from treatment with an ADC or a liposome as described herein. Such a method may comprise contacting tumor cells obtained from the subject with one or more of such antibodies, wherein binding of at leastone of the one or more antibodies to the tumor cell indicates that the subject’s tumor cells express at least one MHC-I molecule whose binding preference may be altered by one or more drugs described herein. Any method for measuring protein expression in the subject’s tumor cell may be used, e.g., immunohistochemistry (IHC), Western blot, immunofluorescence (IF).

[0212] In some instances, a subject that is to be treated by a method disclosed herein may receive an inhibitor of Enhancer of Zeste Homolog 2 (EZH2), a histone-lysine N-methyltransferase enzyme, or an agonist of Stimulator of Interferon Genes (STING) prior to treatment, e.g., to overwrite epigenetic repression of MHC-I expression. The EZH2 inhibition and treatment with a STING agonist have both been found to increase MHC-I expression. Exemplary EZH2 inhibitors include, but are not limited to tazemetostat and valemetostat tosylate. Treatment with an S- adenosine-L-homocysteine (SAH) hydrolase inhibitor or S-adenosylmethionine (SAM) competitive inhibitor may likewise result in EZH2 inhibition. Exemplary STING agonists include, but are not limited to ulevostinag, GSK3745417, IMSA101, CDK-002, MK-2118, BMS-986301, BI 1387446, SB 11285, E7766, SNX281, TAK-676, and SYNB1891. Treatment with a cyclic dinucleotide or analogue of cyclic-GMP-AMP (cGAMP) may likewise result in activation of STING.

[0213] The combination of EZH2 inhibition and STING agonism has been shown to enhance T cell recognition of tumor cells in vivo (Mahadevan et al., Cancer Discov. 2021 Aug; 11(8): 1952-1969). Accordingly, the subject may receive a combination of an EZH2 inhibitor and a STING agonist, e.g., an EZH2 inhibitor followed by a STING agonist.HLA identification

[0214] As can be seen from Table 3, drugs that alter the peptide binding preference of MHC can often interact with more than one MHC-I type (and often more than one MHC -II type). To determine whether a drug targets one or more MHC-I types, it may be desirable to type the subject’s HLA alleles prior to therapy. Since tumor cells may lose expression of a single HLA class I allele, HLA typing may be performed in a sample of tumor cells obtained from the subject (e.g. , alongside an assay that confirms that the tumor cells express MHC).

[0215] Accordingly, the subject in need thereof has an HLA-I allele that can be bound by the drug to alter the peptide binding preference of the allele-encoded MHC-I. The subject’s HLA- I allele(s) may be determined from a sample obtained from the subject. For example, the subject’s HLA-I allele may be determined from a blood or saliva sample or a tumor biopsy obtained from the subject. In some instances, the methods provided herein may comprise a step of obtaininginformation about the subject’s HLA-I allele(s) to select a subject for treatment with a drug capable of binding to and altering the peptide binding preference of the subject’s MHC-I.

[0216] Standard assays for HLA typing are well-known and are typically available in all major transplant centers. Any suitable method for determining a subject’s allele can be used. For example, the subject’s HLA-I allele(s) can be determined, e.g., through a DNA-based analysis of a sample obtained from the subject (e.g., polymerase chain reaction (PCR) sequencing or sequencespecific oligonucleotide probes (SSOP). A PCR-based method may employ one or more primer pairs with sequence specificity for one or more HLA-I alleles. Alternatively, next-generation sequencing (NGS) may be used. Commercial kits are available for HLA typing, e.g. , SeCore HLA Sequence-Based Typing kits and Biotest HLA SSP kits. Information relating to the subject’s HLA- I allele may also be obtained from clinical notes or an external provider. The methods disclosed herein may further comprise a step of obtaining information about the subject’s HLA-I allele(s).Tumors

[0217] The tumor in the subject may be a solid tumor. The tumor in the subject may be a carcinoma, e.g., a squamous cell carcinoma, an adenocarcinoma, a transitional cell carcinoma or a basal cell carcinoma. The tumor in the subject may be a sarcoma, e.g., a bone sarcoma or a soft- tissue sarcoma. The tumor in the subject may be, but is not limited to, lung cancer, colorectal cancer (e.g., rectal cancer, anal cancer), bladder cancer, skin cancer (e.g., non-melanoma, melanoma, or Merkel cell carcinoma), prostate cancer, breast cancer, pancreatic cancer, gynecological cancer (e.g., cervical cancer, ovarian cancer, uterine cancer, endometrial cancer, vaginal cancer or vulval cancer), testicular cancer, penile cancer, a germ cell tumor, gastric (stomach) cancer (e.g., gastroesophageal junction cancer), esophageal cancer, head and neck cancer (e.g. , tonsil cancer, salivary gland cancer, laryngeal cancer), a brain tumor, thyroid cancer, thymus gland cancer, kidney cancer (e.g., Wilms tumor, upper urinary tract urothelial cancer, or adrenal gland tumor), liver cancer, a spinal cord tumor, nasopharyngeal cancer, mouth and oropharyngeal cancer (e.g., tongue cancer), eye cancer (e.g., retinoblastoma), ear cancer, neuroblastoma, a mesothelioma, cholangiocarcinoma, gallbladder cancer, or a neuroendocrine tumor (e.g., small bowel neuroendocrine tumor, large bowel and rectal neuroendocrine tumor, stomach neuroendocrine tumor, neuroendocrine tumor of the pancreas, lung neuroendocrine tumor).

[0218] The tumor in the subject may be a hematological malignancy. The tumor in the subject may be a leukemia, lymphoma or myeloma, e.g., acute lymphoblastic leukemia (ALL), acute myeloid leukemia (AML), non-Hodgkin lymphoma (NHL), Burkitt’s lymphoma, mantle cell lymphoma (MCL), diffuse large B cell lymphoma (DLBCL), chronic lymphocytic leukemia(CLL), chronic myeloid leukemia (CML), chromic myelomonocytic leukemia (CMML), follicular lymphoma, marginal zone lymphoma, small lymphocytic lymphoma (SLL), lymphoplasmacytic lymphoma (LPL), primary effusion lymphoma (PEL), primary mediastinal B cell lymphoma (PMBL), essential thrombocythemia, Hodgkin lymphoma, myelodysplastic syndromes (MDS), myelofibrosis, hairy cell leukemia, multiple myeloma, myeloproliferative neoplasms or polycythemia vera (PV).

[0219] The tumor in the subject may be classified as high-grade or low-grade. The tumor in the subject may be indolent. The tumor in the subject may be metastatic.Subject

[0220] The subject may be an adult, i.e., the subject is 18 years of age or older. Alternatively, the subject may be a child i.e., the subject is less than 18 years of age. The subject may be an infant, i.e., the subject is less than 36 months.

[0221] The subject may have relapsed. The subject’s tumor may be refractory to other cancer therapies. The subject may be treatment naive. Alternatively, the subject may have received one or more prior therapies. The one or more prior therapies may include one or more of chemotherapy, immunotherapy, and radiotherapy.Combination therap

[0222] The subject may be administered two or more different drugs comprised in one or more drug conjugates or liposomes. For example, each drug comprised in a drug conjugate or a liposome can be directed to induce an immunopeptidome shift in an MHC encoded by a different HLA allele. The two or more drugs comprised in two or more drug conjugates or liposomes may be administered concomitantly or sequentially. In the event that loss of heterozygosity (LOH) may occur in the tumor, targeting multiple alleles can avoid this. The two or more different drug conjugates may be the same type of drug conjugate, e.g., each may be ADCs. The two or more different drugs comprised in two or more drug conjugates or liposomes may have different targeting means (e.g., antibodies that bind different tumor-associated antigens).

[0223] Alternatively or additionally, the methods described herein may be used in combination with one or more additional therapies. Such additional therapies may be administered concomitantly or sequentially with a drug conjugate (e.g., an ADC) or liposome comprising the drug described herein, e.g., before or after administering the drug conjugate (e.g., an ADC) or liposome. Additional therapies may comprise radiotherapy, chemotherapy, surgery and / or immunotherapy.

[0224] In some instances, the subject may be administered a drug conjugate (e.g., an ADC) in combination with a liposome, each as described herein. In such instances, the drug comprised in the drug conjugate and the drug comprised in the liposomes are typically different from each other.

[0225] Additional therapies may include immune co-stimulatory therapies and / or immune checkpoint inhibitor therapies. Immune co-stimulatory therapies and immune checkpoint inhibitor therapies rely on pre-existing tumor antigen-specific T cells. The lack of such antigen-specific T cells may have contributed to the failure of such therapies in many cancer patients. Therefore, the use of an ADC or a liposome as described herein to induce an immunopeptidome shift in a tumor cell causes the presentation of new self-peptides that can be recognized by circulating CD8+T cells to elicit a tumor-specific immune response.

[0226] The methods described herein may further comprise treating the subject with an immune checkpoint inhibitor. For example, the immune checkpoint inhibitor may be administered to the subject concurrently with, or after administering the drug conjugate (e.g., an ADC) or liposome. The immune checkpoint targeting therapy can be selected from the group consisting of an antagonist anti-PD-1 antibody, an antagonist anti-PD-Ll antibody, an antagonist anti-PD-L2 antibody, an antagonist anti-CTLA-4 antibody, an antagonist anti-TIM-3 antibody, an antagonist anti -LAG-3 antibody, an antagonist anti-CEACAMI antibody and an IDO inhibitor, i. e. , an agent that inhibits the enzymatic activity of IDO (indoleamine-(2,3)-dioxygenase) and / or TDO (tryptophan 2,3 -dioxygenase).

[0227] For example, the immune checkpoint targeting therapy may be an anti-PD-1 antibody, e.g., pembrolizumab, nivolumab, pidilizumab, MEDI0680, PDR001, REGN2810, PF- 06801591, BGB-A317, TSR-042, or SHR-1210. Alternatively, the immune checkpoint targeting therapy may be an anti-PD-Ll antibody, e.g., atezolizumab, durvalumab, avelumab (MSB0010718C), MDX-1105, or AMP-224. The immune checkpoint targeting therapy may be an anti-CTLA-4 antibody, e.g., ipilimumab. Or, the immune checkpoint targeting therapy may be an anti-LAG-3 antibody, e.g., relatlimab. In some instances, the immune checkpoint targeting therapy may be an IDO inhibitor, e.g., epacadostat, F001287, indoximod, or NLG919.

[0228] Immune co-stimulatory therapy may include treatment with agonistic antibodies that specifically bind an immune co-stimulatory receptor, e.g., CD28, CD27, 0X40, 4-1BB, and GITR.

[0229] Regulatory T (Treg) cells have an immune response modulation function to avoid an aberrant immune response. In certain cancer patients, the amount and activity of Tregs may behigher than in healthy individuals, and may be triggered at the earlier acute phase. Accordingly, a subject receiving or to receive a drug conjugate (e.g., an ADC) or a liposome comprising the drug described herein may further receive administration of a Treg modulating therapy to inhibit or decrease the amount and activity of Tregs.

[0230] Treg modulating therapies are known in the art, and include without limitation antibodies (e.g, full antibodies, and antigen-binding fragments thereof) that specifically bind to CTLA-4, GITR, CCR4, PD-1, LAG3, CD25, or CD15s. The Treg modulating therapy can be administered prior to, contemporaneously with (e.g., during the same doctor visit), or subsequent to the administration of the drug conjugate (e.g., an ADC) or liposome. If the Treg modulating therapy is administered subsequent to the administration of the cell therapy, the patient's response to the cell therapy can be examined to determine the necessity and dose of the Treg modulating therapy.

[0231] Additional therapies may also (or alternatively) include treatment with an EZH2 inhibitor and / or a STING agonist. The EZH2 inhibitor and / or the STING agonist may be administered prior to, or contemporaneously with administration of the drug conjugate (e.g., an ADC). Exemplary EZH2 inhibitors include, but are not limited to, tazemetostat and valemetostat tosylate. Treatment with an S-adenosine-L-homocysteine (SAH) hydrolase inhibitor or S- adenosylmethionine (SAM) competitive inhibitor may likewise result in EZH2 inhibition. Exemplary STING agonists include, but are not limited to ulevostinag, GSK3745417, IMSA101, CDK-002, MK-2118, BMS-986301, BI 1387446, SB 11285, E7766, SNX281, TAK-676, and SYNB 1891. Treatment with a cyclic dinucleotide or analogue of cyclic-GMP-AMP (cGAMP) may likewise result in activation of STING. In some instances, the subject may receive a combination of an EZH2 inhibitor and a STING agonist, e.g. , an EZH2 inhibitor followed by a STING agonist, for instance, prior to administration of the drug conjugate (e.g., ADC) or liposome.EXAMPLES

[0232] The following examples are included for illustrative purposes only and are not intended to limit the scope of the invention. Those of ordinary skill in the art may be aware of materials and methods similar or equivalent to those described below and any of these can be used to practice or test the provided methods and compositions.Example 1. Tumor cells present “new” self-peptides as a result of a drug-induced immunopeptidome shift

[0233] This example illustrates that an immunopeptidome shift can be induced in a tumor cell following exposure to a drug capable of altering the peptide binding preference of a major histocompatibility complex (MHC) molecule.

[0234] Based on the finding that abacavir can shift the binding preference of MHC-I specifically encoded by HLA-B*57:01, it was hypothesized that HLA-B* 57:01 -expressing cells, but not cells expressing a different HLA allele, would present “new” self-peptides after treatment with abacavir.

[0235] The non-small cell lung cancer (NSCLC) cell line H2228 was transduced with either HLA-B*57:01 or HLA-B*07:02 (as a negative control) using the pLEX307 backbone plasmid lenti virus system. A plasmid-only control was also included as a negative control. Selection of transduced cells (i.e., HLA-B*57:01 or HLA-B*07:02 expressing cells) was performed using puromycin treatment. HLA-B*57:01 or HLA-B*07:02 expression was confirmed using quantitative real-time polymerase chain reaction (RT-PCR) with primers specific for HLA- B*57:01 or HLA-B *07: 02. These primers were designed to specifically recognize the transduced HLA alleles while not recognizing any H2228 -endogenous HLA alleles.

[0236] The HLA-allele expressing H2228 cells were seeded into 10 cm tissue culture- coated petri dishes at a density of IxlO6cells and incubated overnight. Parental H2228 cells were included as an additional negative control. H2228 parental, H2228-B*57:01, and H2228-B*07:02 were independently treated with 30 pM abacavir (Selleck Chem; S3165). The dose of abacavir was determined according to Chessman et al. (Immunity, 2008 Jun; 28(6): 822-32).

[0237] Pollowing a 24-hour incubation, peptide MHC (pMHC) immunoprecipitation (IP) was performed using an anti-HLA-ABC antibody (clone W6 / 32). Data-Dependent Acquisition (DDA) mass spectrometry was used to identify peptides extracted from affinity-isolated peptide- HLA complexes, known as the immunopeptidome. DDA has previously been used to demonstrate changes in the immunopeptidome of the mono-allelic C1R.B*57:O1 cell line that depend on abacavir treatment (Illing et al., supra).

[0238] Treatment of H2228-B*57:01 cells with abacavir shifted the immunopeptidome as B*57:01-restricted “new” self-peptides were identified that were absent in the untreated and negative control samples. From 106abacavir-treated H2228-B*57:01 cells, 1364 peptides were identified at a false discovery rate (FDR) of 5%. Of these peptides, 318 peptides were predicted tobind B*57:01 using NetMHC 4.0 based on % rank < 2, and 260 peptides were predicted not to bind any H2228-B*57:01 alleles (A*02:01, B*07:02, B*38:01, B*57:01, C*07:02 or C* 12:03). This prediction does not imply that the 260 predicted non-binder peptides are all associated with abacavir modification of B*57:01. Some DDA-identified peptides can arise from proteolysis during isolation. They could be distinguished from the abacavir-associated peptides, as they were also present in the control samples.

[0239] Some abacavir-associated peptides (i.e., peptides that were present only in the abacavir-treated H2228-B*57:01 cells) were predicted by NetMHC to bind B*57:01. For example, the peptide KTVEIVHIDI (SEQ ID NO: 1) was predicted to be aweak binderto B*57:01, but was absent in the untreated H2228-B*57:01 cells or the H2228 parental control.

[0240] Irrespective of the binding predicted by NetMHC, all fragmentation patterns from the DDA analysis of abacavir-treated samples were used to perform high-sensitivity Data- independent Acquisition (DIA) on untreated cells. This analysis revealed 147 peptides out of 1364 in the abacavir-treated cells that were not identified in the untreated H2228-B*57:01 cells, indicating that approximately 10% of the peptidome was modified by abacavir treatment. As expected, an abacavir-induced immunopeptidome shift was not observed in abacavir-treated H2228-B*07:02 cells (which lack expression of B*57:01; data not shown).

[0241] Mapping the DDA analysis (retention times in minutes) onto the DIA analysis (retention times in scan cycles) revealed a large number of shared peptides between the abacavir- treated and control samples that he on an elution line with low scatter (see FIG. 6), indicating the precision by which scan positions in the untreated samples can be predicted from the retentions times in the treated sample.

[0242] FIG. 7 shows a representative self-peptide presented by H2228-B*57:01 cells in the presence of abacavir, but not in its absence. As shown in FIG. 7A, the DDA run of parent and fragment ions of the abacavir-treated cells revealed a peak for a peptide with an expected mass-to- charge ratio (m / z) of 389.5642. This peak was assigned to the reference fragmentation pattern of a peptide with the amino acid sequence KTVEIVHIDI (SEQ ID NO: 1) with high confidence. Poisson detection is a highly sensitive DIA method for detecting a known fragmentation pattern in a complex ion fragment background. The absence of Poisson detection at the expected elution position with the untreated H2228-B*57:01 cells, as illustrated by the inverted Poisson probability (FIG. 7B), confirmed that an ion matching this mass and fragmentation pattern was not detected in the untreated H2228-B*57:01 cells.

[0243] This example illustrates that treatment with a drug capable of altering the peptide binding preference of MHC can induce an HLA-allele specific immunopeptidome shift in tumor cells.Example 2. Altering the peptide binding preference of MHC-I with abacavir induces a polyclonal T cell response specific to abacavir-treated tumor cells

[0244] This example illustrates that a drug capable of altering the peptide binding preference of an MHC molecule can elicit a polyclonal T cell response that is specifically targeting the tumor cells that received the drug.

[0245] To determine whether a targeted T cell response could be induced that resulted in the killing of tumor cells that had received an immunopeptidome shifter, human CD8+T cells were first stimulated with autologous antigen-presenting cells in the presence or absence of abacavir as a model compound. Briefly, peripheral blood mononuclear cells (PBMC) from HLA-B*57:01 healthy donors were used to generate mature dendritic cells (DCs). For this purpose, PBMCs were cultured for 5 days in a culture medium (RPMI supplemented with 2 mM L-glutamine, 10% human serum, and 1% penicillin / streptomycin) comprising human IL-4 (200 lU / mL) and GM-CSF (400 lU / mL). The resulting immature DCs were stimulated with LPS (100 ng / mL) and IFN-y (100 lU / mL) for 14 hours to induce maturation. The mature DCs were co-cultured with autologous CD8+T cells isolated from PBMCs in the presence of abacavir (10 pg / mL) or solvent (0.002% DMSO; negative control) for up to 14 days. Starting from day 3 of the co-culture, human IL-2 (100 lU / mL) was added every 3 days.

[0246] To determine abacavir-mediated CD8+T cell activation, the expression of 4-1BB was assessed. 4- IBB is a co-stimulatory receptor that is associated with T cell activation. CD8+T cells incubated with anti-CD3 / CD28 beads were used as a positive control. As shown in FIG. 8, the percentage of 4-lBB-expressing CD8+T cells increased to 5% of the total CD8+T cell population following a 4-day culture with abacavir-treated DCs, indicating that a subset of polyclonal CD8+T cells was activated by the DCs which had undergone an immunopeptidome shift and were presenting “new” self-peptides in their MHC-I molecules. In contrast, only about 1% of the CD8+T cells cultured with DMSO-treated DCs expressed 4- IBB. As expected, anti- CD3 / CD28 beads induced 4-1BB expression in the entire T cell population.

[0247] It was then assessed whether the CD8+T cells conditioned by co-culture with the abacavir-treated autologous DCs could recognize and kill HLA-B* 57: 01 -expressing tumor cells. Following the co-culture, CD8+T cells were harvested and exposed to H2228-B*57:01 cells asfollows. H2228-B*57:01 cells were seeded in 96-well plates and incubated in the presence or absence of abacavir as described in Example 1. Following exposure to abacavir-treated H2228- B*57:01 cells for 6 hours, increased intracellular expression of granzyme B and perforin was observed in a small subset of the CD8+T cells (about 2.5% of the total CD8+T cell population), as measured by flow cytometry (FIG. 9A). This indicates that the CD8+T cells could recognize the abacavir-treated H2228-B*57:01 cells. Intracellular expression of granzyme B and perforin was not observed in DMSO-treated H2228-B*57:01 cells which has not undergone an immunopeptidome shift, demonstrating that effector function of CD8+T cells was specific to tumor cells presenting “new” self-peptides as a result of the treatment with the immunopeptidome shifter.

[0248] Supernatants from the 96-well plates were collected for granzyme B ELISA to further confirm that abacavir treatment can successfully expand CD8+T cells reactive to “new” self-peptides which can recognize HLA-B*57:01+cancer cells. As a positive control, the supernatant from H2228-B*57:01 cells treated with anti-CD3 / CD28-activated CD8+T cells was included. In addition, a no T-cell control was also included. Consistent with the intracellular expression data, granzyme B was detected in the supernatant obtained from the co-culture of CD8+T cells and abacavir-treated H2228-B*57:01 cells. In contrast, no secreted granzyme B was detected in the supernatant obtained from the co-culture of CD8+T cells and DMSO-treated H2228-B*57:01 cells.

[0249] In parallel, T cell-mediated cytotoxicity of tumor cells was investigated. Briefly, CD8+T cells were sorted and labelled with Cell Proliferation Dye eFluor 450 before being contacted with target tumor cells (mCherry-labelled H2228 HLA-B*57:01 cells). The target tumor cells were irradiated with 35 Gy prior to a 24-hour treatment with 10 pg / ml abacavir or 0.002% DMSO. Following treatment, the target tumor cells were washed and cultured with human CD8+T cells, at an effector-to -target (E:T) ratio of 10: 1. Cytotoxic T lymphocyte (CTL) activity was measured every 4 hours using a Cytation 5 instrument. As shown in FIG. 9C, a specific anti-tumor response was observed only against the abacavir-treated H2228 HLA-B*57:01 cells. This response can be attributed to the induction of a tumor immunopeptidome shift by abacavir. Absolute killing level approached -50% at 48 hours without any detectable alloreactivity (FIG. 9C). This level of cytotoxicity is highly significant in view of the small fraction of CD8+T cells recognizing “new” self-peptides as a result of the abacavir-induced immunopeptidome shift, as shown in FIG. 8 and FIG. 9A.

[0250] This example illustrates the expansion of a subset of “neoepitope”-reactive human T cells following treatment with a drug capable of altering the peptide binding preference of MHC.Such T cells were capable of eliciting a cytotoxic anti-tumor response specifically against human tumor cells that had been treated with the immunopeptidome shifter.Example 3. Design and construction of an ADC comprising a drug capable of altering the peptide binding preference of an MHC

[0251] This example illustrates the successful preparation of a stable antibody-drug conjugate (ADC) comprising a tumor-targeting antibody and a drug capable of altering the peptide binding preference of an MHC.

[0252] Abacavir is well-known for being capable of altering the peptide binding preference of the MHC encoded by HLA-B* 57 : 01. As a proof of principle, abacavir was selected for inclusion in an ADC. The monoclonal antibody, datopotamab, was selected as a targeting means. Datopotamab is an anti-TROP2 monoclonal antibody. An ADC comprising this antibody, datopotamab deruxtecan (Dato-DXd, DS-1062a), has previously been manufactured (Okajima et al. Mol Cancer Ther. 2021 Aug 19;20( 12):2329-2340).

[0253] Abacavir was conjugated to the linker maleimidocaproyl-(Mc)-[GGFG (SEQ ID NO: 2)]-(OH). Subsequently, Mc-[GGFG (SEQ ID NO: 2)]-(OH) abacavir was conjugated to datopotamab at various drug to antibody ratios (DARs). Three different DAR formulations were tested: (1) a DAR of -8.00, (2) a DAR of -4.00, and (3) a DAR of -2.00. Briefly, the conjugation process comprised incubating datopotamab in TCEP for 2 hours at 25°C. Antibody reduction was performed at a molar ratio (TCEP: Ab) of 1.4, 2.5, and 10 for the conjugates to have a DAR of 2, 4, and 8, respectively. In parallel a linker-abacavir solution was prepared in DMSO. The datopotamab solution and linker-abacavir solution were mixed at a molar ratio (LA: Ab) of 4, 8, and 12, for the conjugates to have a DAR of 2, 4, and 8, respectively, in 10% DMSO followed by a 2-hour incubation at 25°C. Unconjugated linker-abacavir was removed using a Charcoal and Amicon® Ultracel-30 regenerated cellulose membrane with a 30 kDa molecular weight cut-off. The ADC was then extracted and filtered using a 0.22 pm filter. The ADC sample was then analyzed by Size-Exclusion High-Performance Liquid Chromatography (SEC-HPLC), hydrophobic interaction chromatography-high performance liquid chromatography (HIC-HPLC), liquid chromatography-mass spectrometry (LC-MS) Endotoxin and Free Drug.

[0254] Table 7 summarizes the results of the ADC conjugation at each of the three DARs tested. For each of the three ADC preparations, the amount of free linker-abacavir was below the level of quantification.Table 7. Results of ADC conjugation

[0255] Stability of the ADC at the three different DARs was then tested. The ADCs were subjected to 5 freeze-thaw cycles, stored at 4°C for 7 days, and incubated at 40°C for 3 days, and stored at 4°C for another 7 days. The prepared ADCs were analyzed by SEC-HPLC for aggregation, LC-MS for DAR after. The results are summarized in Table 8. Overall, the prepared ADCs exhibited good stability.Table 8. Results of ADC conjugation

[0256] The datopotamab-abacavir conjugate was incubated with the H2228-B*57:01 cells of Example 1 to assess cellular uptake. FACS analysis indicated that an excess of 50% of this antibody-drug conjugate (ADC) was internalized within the first 24 hours after its addition to the cell culture medium. Given datopotamab’s Ka of 0.74 nM, it was determined that a concentration of 0.45 pg / mL of the ADC should theoretically lead to a 70% saturation of TROP2 expressed on the surface of the tumor cells.

[0257] This example illustrates the successful preparation of an ADC comprising a tumortargeting antibody and a drug capable of altering the peptide binding preference of an MHC. The prepared ADCs were found to have suitable stability for therapeutic use. This example further illustrates that a conjugate comprising a drug capable of altering the peptide binding preference of an MHC and an antibody that specifically binds a surface-expressed tumor-associated antigen is taken up by tumor cells.Example 4. Modeling abacavir binding to HLA-B*57:01 using Al models

[0258] This example illustrates that artificial intelligence (Al) models can accurately predict the binding of abacavir to the F pocket of HLA-B*57:01.

[0259] Abacavir is known to bind non-covalently within the F pocket of the MHC -I encoded by the HLA-B*57:01 allele. The crystal structure 3VRJ shows abacavir bound to HLA- B*57:01 and the self-peptide LTTKLTNTNI (filing et al., Nature. 2012 Jun 28;486(7404):554- 8). FIG. 10 overlays abacavir in the F pocket ofHLA-B*57:01 as shown in 3 VRJ with the abacavir structures predicted by two Al methods, namely, AlphaFold3 and Boltz-2. The root mean squaredeviation (RMSD) of the atomic positions in the modeled structures relative to 3VRJ were 0.714 A (Boltz-2) and 1.129 A (AlphaFold 3), respectively.

[0260] This example illustrates that Al models such as Boltz-2 and AlphaFold 3 can be used to predict that abacavir can bind to the F pocket of HLA-B*57:01. The results in this example indicated that such models can be used to identify small-molecule drugs that may be capable of altering the peptide binding preference of an MHC.Example 5. Boltz-2 predicts abacavir binding to HLA-B*57:01 but not HLA-B*57:03

[0261] This example illustrates that an Al model such as Boltz-2 can accurately predict whether a small-molecule drug can bind to a pocket within the peptide binding cleft of an MHC.

[0262] The Boltz-2 (Passaro et al., bioRxiv. 2025; 2025.06.14.659707) is an Artificial Intelligence (Al) model that can be used to predict small molecule-protein binding affinities. Crucially, it achieves strong correlation with experimental readouts on many benchmarks. To explore whether this model could be used identify small-molecule drugs capable of altering the peptide binding preference of an MHC, Boltz-2 was first tested to predict the binding of abacavir to two different MHC-I structures. Abacavir is known to bind non-covalently within binding pocket F of the MHC-I encoded by the HLA-B*57:01 allele. As a control, Boltz-2 was used to predict binding to the MHC-I encoded by the HLA-B*57:03 allele. Both protein complexes differ by only two amino acids.

[0263] Boltz-2 accurately refolded the MHC-I encoded by the HLA-B*57:01 allele with abacavir, with the correct structure and docking location. Boltz-2 also predicted that abacavir would bind to HLA-B*57:01 but not to the closely related HLA-B*57:03. This is illustrated in FIG. 11A and FIG. 1 IB. FIG. 11A illustrates that Boltz-2 predicts abacavir binding in the F pocket of HLA-B*57:01. FIG. 11B illustrates that Boltz-2 predicts abacavir to be unable to bind in the F pocket (or plausibly in the A and B pocket) of HLA-B*57:03.

[0264] Table 9 summarizes the binding metrics and interface positional errors for abacavir binding to HLA-B*57:01 and HLA-B*57:03 predicted by Boltz-2.Table 9. Predicted B*57:03 / abacavir and B*57:01 / abacavir binding metrics

[0265] Table 9 shows a large positional error in the binding site for HLA-B*57:03, but a low positional error for HLA-B*57:01. Furthermore, the binding probability for HLA-B*57:03 was within the range of random molecules tested in the Boltz-2 validation (e.g., less than 0.3; see Figure 15 of Passaro et al., bioRxiv. 2025; 2025.06.14.659707), whereas HLA-B*57:01 had a binding probability of a borderline binder. The affinity predictions were in the range of 1-2 pM, which is consistent with the relatively high concentrations of abacavir required to elicit an effect on HLA-B*57:01 (30 pM in Example 2).

[0266] This example illustrates that Boltz-2 can correctly predict abacavir binding to an MHC, thus validating the methodology for in silico screening to identify small-molecule drugs that may be capable of altering the peptide binding preference of an MHC.Example 6. In silico screen for compounds that bind to the F pocket or the A and B pockets of HLA-A*02:01

[0267] This example illustrates the use of an Al model such as Boltz-2 to identify smallmolecule drug-like compounds that may be capable of altering the peptide binding preference of HLA-A*02:01.

[0268] Following validation of the Boltz-2 docking methodology described in Example 5, the same computational pipeline was used to perform an in silico screen to identify compoundsthat bind to the F or the A and B pockets of the MHC-I encoded by the HLA-A*02:01 allele. This HLA-I allele has a 15% prevalence in the human population.

[0269] An in silico docking screen was carried out using Ubuntu 24.04 LTS. A library of small molecules was exported from the publicly available ZINC20 database (Irwin et al., J. Chem Inf Model, 2020; 60(12): 6065-6073), amounting to ~ 625 million compounds. The compounds in this library were filtered based on their molecular weight (MW) and hydrophobicity. Molecules with a MW of < 350 Daltons and LogP of < 3.5 were retained for further analysis. The MW and LogP cut-offs were based on the “Rule of Three”, which is a set of guidelines for fragments used in Fragment-Based Drug Discovery (Congreve et al., Drug Discov Today. 2003; 8(19):876-7) and was determined to be optimal for gaining access to the F or the A and B pockets deep within the peptide binding cleft of HLA-A* 02:01.

[0270] A combination of flexible ligand docking (QuickVina2; Tang et al., IEEE / ACM Trans Comput Biol Bioinform. 2024; 21(6): 2382-2393) and Boltz-2 was utilized to further filter the remaining compounds. For QuickVina2-GPU-2, the HLA-A*02:01 was kept rigid while rotations were allowed for the compounds.

[0271] HLA-A*02:01 was modeled using the crystal structure of PDB ID 5HHN(Valkenburg et al., Proc Natl Acad Sci USA. 2016; 113(16):4440-5). Using PyMOL, crystallographic waters, the MHC-I peptide sequence, [32M and the a3 domain of the heavy chain (residues 181-274) were removed to allow for unimpeded docking of the compounds. Hydrogens were added to the structure using Reduce (Word et al., J Mol Biol. 1999; 285(4): 1735-47) with a pH of 7.4 before the final conversion to PDBQT format with AutoDock Tools (ADT) (Morris et al., J Comput Chem 2009; 30(16):2785-9E).

[0272] The center of the search box was set to be in the F-pocket of the MHC-I encoded by the HLA-A* 02: 01 allele and the size of the docking box was 40Ax40Ax40A to allow for the entire MHC platform to be searched including both the A, B and F pockets. QuickVina 2 utilizes a heuristic approach, and nine distinct potential binding conformations for each compound were generated. All compounds scoring -8.5 kcal / mol to -10.4 kcal / mol were selected for re-screening. This equates to a predicted dissociation constant in the range of 0.5 pM to 30 nM.

[0273] In order to confirm the hits identified by QuickVina 2, Boltz-2 was utilized for a rescreen of the selected ligands. Co-folding the HLA-A* 02: 01 -encoded MHC with a small molecule compound required the amino acid sequence of the MHC-I peptide, together with matching MSAs and SMILES representations of the small molecule. The sequence for the MHC- I encoded by the HLA-A* 02: 01 allele was obtained from the IPD-IMGT / HLA Database (Barkeret al. , Nucleic Acids Res. 2023; 51:D 1053-60) with residues 25-300 from the listed sequence being selected as input for both the multiple sequence alignment (MSA) and Boltz-2 input. MSA was performed using MMseqs2 packaged with Colabfold (Mirdita et al., Nature Methods. 2022; 19:679-682). SMILES representations of the selected hits were extracted from the ZINC20 library, with each hit having its own configuration file but utilizing the same MSA and HLA-A*02:01 sequence inputs.

[0274] The selected ligands were filtered following two different sets of cut-offs. Compounds meeting either of the cut-offs were combined to yield 31 candidates. One set of cutoffs followed recommendations based on the binding probability output from Boltz-2. A cutoff of greater than 0.5 was applied, with further refinement to retain only compounds predicted to bind at from 2 pM to 100 nM. This cut-off excludes compounds with predicted binding that is the same or worse than abacavir. The second set of cut-offs incorporated both binding metrics and predicted ligand-protein interface error (IPDE), selecting structures with a binding probability greater than 0.35, predicted affinity better than 3 pM, and an interface error of less than 0.5A. This cut-off excludes compounds with predicted binding that is worse than abacavir as well as molecules where the pocket is poorly defined due to a high positional error, which is consistent with non-specific binding and poor development potential.

[0275] The 31 candidate compounds were visually inspected in PyMOL. The candidate compounds were sorted by affinity and filtered based on availability for order. F pocket or A and B pocket-binding compounds (1) to (10) were selected for further experimental validation:

[0276] Compounds (1), (4), (5) and (10) were predicted to be localized in the A and B pockets. Predicted interactions between compounds (1), (4), (5), and (10) and particular amino acid residues in or adjacent to the pockets are shown in FIG. 12A, FIG. 12B, FIG. 12C, and FIG. 12D, respectively. Compounds (2), (3), (6), (7), (8) and (9) were predicted to bind in the F pocket. Predicted interactions between compounds (2), (3), (6), (7), (8), and (9) and particular amino acid residues in or adjacent to the pocket are shown in FIG. 13A, FIG. 13B, FIG. 13C, FIG. 13D, and FIG. 13E. The results from the Boltz-2 analysis of compounds (1) to (10) are summarized in Table 10, and compound (l)’s and (2)’s predicted binding orientations within the respective pockets are shown in FIG. 14A and FIG. 14B, respectively.Table 10. A*02:01 binding metrics for compounds (1) to (10) predicted by Boltz-2

[0277] Analysis of the predicted binding orientations of compounds (1), (4), (5) and (10) identified specific moieties that are consistently present in the compounds that bind to the A and B pockets. Based on the binding of compounds (1), (5) and (10), it was predicted that compounds engaging the A and B pockets display a pair of adjacent amide groups that establish a directional hydrogen-bonding array mimicking a short peptide fragment - a feature that likely underpins strong polar interactions in the A pocket, and a compact hydrophobic substituent which fits in the B pocket’s hydrophobic cavity. For example, the model predicted that compound (1) binds via the N-(cyclopentylmethyl)acetamide substituent. It was predicted that the acetamide is stabilized by hydrogen-bonding in the A pocket, mimicking a peptide backbone, and the hydrophobic cyclopentane moiety binds deeply within the hydrophobic B pocket. This is illustrated in FIG. 14A. Compounds (5) and (10) also comprise a pair of adjacent amide groups that forms polar interactions with the A pocket. Furthermore, compounds (5) and (10) comprise a hydrophobic branched alkylgroup (isopropyl group in compound (5) and isobutyl group in compound (10)) that is predicted to bind deep within the hydrophobic B pocket. Compounds (1), (5) and (10) also possessed stereocenters adjacent to the amide framework, indicating that the A pocket is stereochemically constrained.

[0278] While compounds (1), (5) and (10) adopt an aliphatic peptidomimetic profile, compound (4) illustrates that the A and B pockets can accommodate aromatic and heteroaromatic units along with fluorine and sulfur substituents providing opportunities for 71-71 stacking and halogen bonding. Compound (4) is a small compound that is predicted to fit entirely within the A and B pockets of the HLA-A* 02: 01 -encoded MHC. Although compound (4) does not contain the same pair of adjacent amide groups, it does have N-H groups, amide, and thiol available for hydrogen bonding and thus forms polar interactions with the A pocket. It also has a hydrophobic group that binds within the B pocket.

[0279] Analysis of the predicted binding orientations of compounds (2), (3), (6) to (9) identified specific moieties that are consistently present in the compounds that bind in the F pocket. It was predicted that compounds engaging in the F pocket displayed an amide or urea-like group that was predicted to bind around the opening of the pocket, where it can act as both a hydrogen bond donor and hydrogen bond acceptor. The compounds favor compact, conformationally restricted scaffolds that maximize hydrophobic contacts and entropic gains. For example, compound (2) is predicted to bind deeply within the hydrophobic F-pocket via the chloroindane moiety. This is illustrated by FIG 12B. Compounds (6) and (9) comprise small cyclic amides (i.e. pyridone groups), and compounds (7) and (8) comprise hydrophobic cyclopentane groups, which are predicted to bind within the F pocket. These findings illustrate that the F pocket accommodates hydrogen bonding as well as hydrophobic stacking.

[0280] The newly identified compounds that bind to the A and B pockets or the F pockets of the HLA-A* 02:01 -encoded MHC have predicted affinitie s in the low micromolar to nanomolar range. They can be modified with a conjugatable moiety to form a drug, that when coupled to a targeting means such as an antibody, optionally via a linking means, can form a drug conjugate suitable for use in the anti -cancer therapy of subj ects whose cancer cells express the MHC encoded by the HLA-A* 02:01 allele.

[0281] This example illustrates the identification of ten compounds that are predicted to bind to HLA-A* 02: 01 -encoded MHC in such a manner that they can alter the peptide binding preference of this MHC (as well as MHCs with similar binding pockets encoded by closely related HLA-A alleles) and thus induce an immunopeptidome shift. This example also illustrates that suchHLA-A immunopeptidome shifters may be readily identified using a validated, rapid, open-source Al model.

[0282] It should be understood that the details provided herein are given by way of illustration only, not limitation. Other features, objects, and advantages are apparent from the above detailed description, drawings and examples. Various changes and modifications will be apparent to those skilled in the art.

[0283] All patents, patent publications, and non-patent publications referenced herein are indicative of the level of skill of those skilled in the art to which this invention pertains. All these publications are herein incorporated by reference to the same extent as if each individual publication were specifically and individually indicated as being incorporated by reference.

Claims

CLAIMS1. A method of inducing an immunopeptidome shift in a tumor cell of a subject in need thereof, wherein the method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

2. A method of inducing a polyclonal cytotoxic T cell (CTL) response against a tumor cell of a subject in need thereof, wherein the method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means that directs the drug conjugate to the tumor cell, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

3. The method of claim 2, wherein the polyclonal CTL response comprises CD8+CTLs.

4. A method of treating a tumor in a subject in need thereof, wherein the method comprises administering a therapeutically effective amount of a drug conjugate to the subject, wherein the drug conjugate comprises a targeting means that directs the drug conjugate to the tumor cells forming the tumor, and the drug binds to and alters the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cells.

5. The method of any one of claims 1-4, wherein the targeting means specifically binds an antigen on the surface of the tumor cells.

6. The method of claim 5, wherein the antigen is a tumor-associated antigen.

7. The method of claim 6, wherein the tumor-associated antigen is expressed at least 10-fold higher on the tumor than on normal tissue.

8. The method of any one of claims 5-7, wherein the targeting means is an antibody9. The method of claim 8, wherein the antibody comprises an Fc region that is modified to reduce or eliminate one or more effector functions.

10. The method of claim 9, wherein the Fc region of the antibody comprises LALA mutations and / or PG mutations.

11. The method of any one of claims 1-10, wherein the method further comprises obtaining information about the subject’s HLA alleles.

12. The method of any one of the preceding claims, wherein the targeting means is conjugated to the drug via a cleavable linker.

13. The method of claim 12, wherein the cleavable linker is an enzyme cleavable linker comprising a protease cleavage site.

14. The method of claim 13, wherein the enzyme is an intracellular protease (e.g., cathepsin).

15. The method of claim 12 or 14, wherein the enzyme cleavable linker comprises tetrapeptidyl-aminomethoxy.

16. The method of any one of the preceding claims, wherein the drug is capable of altering a peptide binding cleft of the subject’s HLA-A, HLA-B or HLA-C MHC.

17. The method of claim 16, wherein the drug alters the binding of self-peptides to the B and / or F pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

18. The method of claims 16 or 17, wherein the drug is capable of interacting with one or more amino acids at positions 7, 9, 24, 34, 45, 63, 66, 67, 70 and 99 of the B pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

19. The method of any one of claims 16-18, wherein the drug is capable of interacting with one or more amino acids at positions 77, 80, 81, 84, 95, 116, 123, 143, 146 and 147 of the F pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

20. The method of any one of claims 1-16, wherein the drug is a compound selected from Table 3, or a derivative or structural variant thereof, or a compound selected from the group consisting of(10); or a derivative or structural variant thereof.

21. The method of claim 20, wherein the drug is abacavir, allopurinol, or carbamazepine, or a derivative or structural variant thereof.

22. A drug conjugate comprising a targeting moiety, wherein the drug conjugate comprises a targeting means that directs the drug conjugate to a tumor cell, and the drug is capable of binding to and altering the peptide binding preference of a major histocompatibility complex (MHC) expressed by the tumor cell.

23. The drug conjugate of claim 22, wherein the targeting means specifically binds an antigen on the surface of the tumor cell.

24. The drug conjugate of claim 23, wherein the antigen is a tumor-associated antigen.

25. The drug conjugate of claim 24, wherein the tumor-associated antigen is expressed at least10-fold higher on the tumor than on normal tissue.

26. The drug conjugate of any one of claims 22-25, wherein the targeting means is an antibody.

27. The drug conjugate of claim 26, wherein the antibody comprises an Fc region that is modified to reduce or eliminate one or more effector functions.

28. The drug conjugate of claim 27, wherein the Fc region of the antibody comprises LALA mutations and / or PG mutations.

29. The drug conjugate of any one of claims 22-28, wherein the drug conjugate is conjugated to the drug via a cleavable linker.

30. The drug conjugate of claim 29, wherein the cleavable linker is an enzyme cleavable linker comprising a protease cleavage site.

31. The drug conjugate of claim 30, wherein the enzyme is an intracellular protease (e.g., cathepsin).

32. The drug conjugate of claim 30 or 31, wherein the enzyme cleavable linker comprises tetrapeptidyl-aminomethoxy.

33. The drug conjugate of any one of claims 22-32, wherein the drug is capable of altering a peptide binding cleft of the subject’s HLA-A, HLA-B or HLA-C MHC.

34. The drug conjugate of claim 33, wherein the drug alters the binding of self-peptides to the B and / or F pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

35. The drug conjugate of claims 33 or 34, wherein the drug is capable of interacting with one or more amino acids at positions 7, 9, 24, 34, 45, 63, 66, 67, 70 and 99 of the B pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

36. The drug conjugate of any one of claims 33-35, wherein the drug is capable of interacting with one or more amino acids at positions 77, 80, 81, 84, 95, 116, 123, 143, 146 and 147 of the F pocket of the subject’s HLA-A, HLA-B or HLA-C MHC.

37. The drug conjugate of any one of claims 22-36, wherein the drug is a compound selected from Table 3, or a derivative or structural variant thereof, or a compound selected from the group consisting of(10) or a derivative or structural variant thereof.

38. The drug conjugate of claim 37, wherein the drug is or has molecular similarity to any one of compound (1) to compound (10), abacavir, allopurinol, or carbamazepine, or a derivative thereof.

39. A method of treating a tumor in a human subject in need thereof comprising:(a) obtaining information about the subject’s human leukocyte antigen (HLA) alleles; and(b) steps for altering the peptide binding preference of a major histocompatibility complex (MHC) which is encoded by one of said HLA alleles and expressed by cells of the tumor, thereby inducing an immunopeptidome shift in the cells and eliciting a polyclonal cytotoxic T cell (CTL) response against the tumor.

40. The method of claim 39, wherein the MHC of which the peptide binding preference is altered is encoded by an allele selected from HLA-A, HLA-B, or HLA-C41. The method of claim 39 or 40, wherein step (a) comprises identifying an HLA allele encoding the MHC in a sample obtained from the subject.

42. The method of claim 39, wherein the method comprises reversibly altering the peptide binding cleft of the MHC.

43. The method of claim 42, wherein the steps reversibly alter the binding of self-peptides expressed by cells of the tumor to the A and / or B pocket(s) of the MHC, or the F pocket of the MHC.

44. The method of claim 43, wherein one or more amino acids at positions 7, 9, 24, 34, 45, 63, 66, 67, 70 and 99 of the A and / or B pocket(s) of the MHC are prevented from contacting self- peptides.

45. The method of claim 42, wherein one or more amino acids at positions 77, 80, 81, 84, 95, 116, 123, 143, 146 and 147 of the F pocket of the MHC are prevented from contacting self- peptides.

46. The method of claim 39, wherein the steps comprise administering to the subject a drug that binds to and alters the peptide binding preference of the MHC.

47. The method of claim 46, wherein drug is conjugated to a targeting means that directs the drug to the cells of the tumor.