WO2025224052A1

WO2025224052A1 - Methods and products for the generation and identification of t cells and tcrs

Info

Publication number: WO2025224052A1
Application number: PCT/EP2025/060849
Authority: WO
Inventors: Meng Amy LI; E-Chiang Lee
Original assignee: T Therapeutics Ltd
Current assignee: T Therapeutics Ltd
Priority date: 2024-04-22
Filing date: 2025-04-22
Publication date: 2025-10-30
Anticipated expiration: 2026-10-22
Also published as: GB2640828A; GB202405639D0

Abstract

Methods of discovering sequences of T cell receptors that bind a target MHC-peptide complex, by immunising laboratory animals with antigen-presenting cells expressing an MHC-peptide complex in which the MHC is non-self to the animal Allotype and/or species mismatch between the target MHC and the MHC of the laboratory animal leads to generation of T cells with high a inity T cell receptors that exhibit speciﬁc and selective binding to the target MHC- peptide complex. Transgenic animals with human TCR loci may be used in methods of allopriming or xenopriming to discover human TCRs for development as therapeutic agents for use in humans.

Description

Methods and Products for the generation and identification of T cells and TCRs

Introduction

Disclosed herein is a method for the generation and identification of T cells and T cell receptors (TCRs), and TCR sequences and TCRs identified by said methods, and uses thereof.

Background

In diseases such as cancer and autoimmune diseases, many potential targets for disease modification by pharmaceutical intervention are self antigens. For example, cancer associated antigens are encoded by genes which are overexpressed or ectopically expressed in cancer cells compared with healthy cells. In many incidences of autoimmune disease, the trigger of autoimmune reactions is T cells recognising and being activated against self peptide major histocompatibility complexes (pMHCs).

T cells and TCRs are useful therapeutic agents for treating diseases such as cancer and autoimmune diseases in patients, because they bind epitopes derived from cell surface and intracellular proteins, thereby enabling the targeting of self antigens in the patient. Because of the accessibility to intracellular targets, TCR based therapeutics expand the target landscape in treating cancer and autoimmune diseases compared to antibody based therapeutics which can only target cell surface proteins. The TCR binds a peptide presented in the context of a major histocompatibility complex (MHC) molecule on the cell surface. Most cells present pMHC, continually displaying a sample of peptides derived from intracellular processing of proteins, enabling surveillance by T cells.

For therapeutic purposes, especially in the context of soluble drug molecules such as bispecific proteins incorporating a TCR binding arm, high affinity TCRs against such targets are required to achieve efficacy. However, TCRs derived from mature T cells typically have low affinity against pMHCs presented by cells of the body. This is in part due to thymic T cell education during development called negative selection. While positive selection in the thymus eliminates T cell clones which fail to recognise self pMHC, negative selection ensures the T cell clones with high affinity against self pMHC are eliminated.

High-affinity TCRs specific for self tumour-associated antigens (TAAs) and autoantigens for autoimmune diseases are therefore rarely present in the natural T-cell repertoire. TCRs suitable for development as high affinity human therapeutics therefore cannot usually be found directly from screening T cells in human patients where the target pMHC complex is a self molecule.

One way to overcome the issue of negative selection is to generate T cells to human antigens in a different species, such as mice, where the human antigen is not conserved. Transgenic mice have been generated, containing humanised MHC and TCR genes. When immunised with human target peptide, such mice are able to present the human peptide in complex with the human MHC on their cells, and raise a T cell immune response to the target pMHC complex. Sequences of human TCRs can be discovered in such mice, and developed as therapeutic molecules for use in humans. For example, Obenaus et al., (Nature Biotechnology Volume 33 number 4 April 2015, pages 402-407) generated antigen-negative humanized transgenic mice ABabDII with a diverse human TCR repertoire restricted to the human leukocyte antigen (HLA) A*02:01. These mice were immunized with human TAAs in the form of peptides, for which they are not tolerant, allowing induction of CD8+ T cells with TCRs that had an affinity in a range that induced T cell expansion. This was possible because the human peptide target is not conserved between human and mice, so the mouse T cell repertoire had not undergone negative selection against the human target protein. However, peptides which are highly similar in amino acid sequence to the target but derived from the mouse proteome other than the homologous target gene could exist to induce tolerance.

Therefore there is still a need for a method for identifying high affinity TCRs which is applicable to other target types, including self peptides and those that are conserved in transgenic animals. T cells, and T cell receptors, that are capable of high affinity binding to peptides of interest are particularly useful for incorporation into biopharmaceuticals. T cells for use in cell therapy often require TCRs that bind their target with higher affinity than is typically isolated from natural repertoires, and TCRs for use as or in soluble therapeutic molecules generally require even higher affinity than that required for cell therapy.

Statements of invention

The present invention provides methods of discovering TCR sequences from T cells of laboratory animals, by priming an immune response in the T cells using a target pMHC complex wherein the MHC is non-self to the animal. Mismatch between the target MHC and the MHC allotype of the animal leads to generation of TCRs that bind the target MHC with higher affinity, compared with previously known methods in which the MHC allotype of the animal was matched to the target. The invention is useful for discovering TCR sequences with high affinity binding to target pMHC complexes, suitable for development as therapeutics for use in patients in which the target pMHC complex is associated with a disease or condition to be treated.

The present invention relates to:

A method for generation of a TCR in a laboratory animal, the method comprising i Delivering an antigen presenting cell to the laboratory animal, wherein the antigen presenting cell expresses an MHC-peptide complex in the laboratory animal, and wherein the MHC component of the MHC-peptide complex is not already expressed in the laboratory animal ii generation of T cells to the MHC -peptide complex in the laboratory animal; and iii optionally, isolation and/or purification of T cells from the laboratory animal reactive with the MHC -peptide complex, such as a polyclonal mixture of T cells or a monoclonal T cell, or cell lines thereof. The invention also comprises a method comprising steps (i) and (ii) above, optionally step (iii), together with a step of ex vivo enrichment of T cells of interest, for example based on binding affinity.

Also disclosed herein is a method for generation of a TCR in ex vivo, the method comprising i Delivering an antigen presenting cell to T cells of the laboratory animal, wherein the antigen presenting cell expresses an MHC-peptide complex in the laboratory animal, and wherein the MHC component of the MHC-peptide complex is not already expressed in the laboratory animal ii generation of T cells to the MHC -peptide complex; and iii optionally, isolation and/or purification of T cells from the laboratory animal reactive with the MHC -peptide complex, such as a polyclonal mixture of T cells or a monoclonal T cell, or cell lines thereof.

The invention also relates to a TCR, or variable region or domain(s) thereof, obtained or obtainable from the method, for use in a patient, e.g. human patient in need thereof having a host MHC, e.g. a HLA which is the same as that used to immunise the rodent and expressed by the APC.

The invention also relates to a bispecific molecule or multispecific molecule having as one arm a TCR, or variable region or domain(s) thereof, obtained or obtainable from the method. The molecules may be for use in a patient in need thereof having a host MHC which is the same as that used to immunise the laboratory animal and expressed by the APC.

The invention also relates to a method of treating an individual in need thereof, the method comprising delivery to said individual of a cell expressing a TCR, e.g. a human TCR or the variable region thereof, or a domain(s) thereof which is obtained or obtainable from the method disclosed herein, or comprising delivery of a nucleic acid (e.g. RNA e.g. mRNA or DNA) encoding the T cell receptor, or T cell receptor variable region or domain(s) thereof, such as a human TCR or variable region thereof, or a TCR polypeptide or fragment thereof as disclosed herein, formulated with a pharmaceutically acceptable excipient, into a patient in need thereof.

The TCR or variable region thereof, or variable domain(s) thereof, that is delivered or expressed is, in one aspect, in a soluble form, that is, not being associated with the surface of a T cell.

The laboratory animal may express a transgenic MHC allele, wherein the allele is different from the target MHC but is of the same species as the target MHC. Use of the target peptide- MHC complex (the MHC-peptide complex present in the target individual or population to be treated) for priming of an immune response in T cells of an animal with different MHC allotype is referred to herein as allopriming.

Alternatively, the laboratory animal may express an MHC allele that is of a different species compared with the target MHC. Use of the target peptide-MHC complex (the MHC peptide complex present in the target individual or population to be treated) for priming of an immune response in T cells of an animal with different species MHC is referred to herein as xenopriming. For xenopriming, the animal preferably is not transgenic for any MHC, so may express only endogenous host MHC.

The allo or xeno priming approach is illustrated herein with in vivo allopriming using an HLA- A*02:01 mouse. The invention provides a number of potential approaches and advantages, exemplified by but not limited to:

• Non-HLA-A*02:01 TCR discovery using HLA-A*02:01 mouse;

• TCR discovery of peptide presented by HLA-A*02:01 using non-HLA-A*02:01 mouse, e.g. HLA-A*ll:01, HLA-A*03:01, HLA-A*24:02.

• TCR discovery of any peptide presented by any human HLA using xeno MHC mouse, the genotype of which is stated in Figure 1. It will be appreciated that this principle applies beyond the human /mouse xeno combination, and the invention relates to, for example, the generation of TCR to any peptide presented by an MHC molecule in a laboratory animal which is not naturally expressed in the genome of the laboratory animal, such as TCR discovery of any peptide presented by any human HLA using a mouse naturally expressing only a mouse MHC.

• Polyclonal CTL lines establishment from an immunised mouse.

Immunisation using antigen presenting cells that present MHC-peptide complex with MHC allotype mismatched to the MHC of the host laboratory animal triggers an allogenic T cell response. An example is an APC presenting a first HLA allotype, e.g., HLA-A*24:02, used to immunise an animal expressing a second HLA allotype, e.g., HLA-A*02:01. Since the host T cell repertoire is not negatively selected on non-self HLA allotypes during development, the affinity of antigen-specific TCRs derived from such allogenic immunisation may be higher than TCRs derived from syngeneic priming, as in the latter situation the host T cells have been negatively selected to remove high affinity TCRs which could otherwise potentially cause autoimmune responses. Xenopriming is a further extension of allopriming concept: instead of mismatched MHC allotypes of one species (e.g., different HLA allotypes of human) between immunogen and the host, the MHC mismatch is between species (e.g., APC presenting human class I HLA, immunising animal lacking human class I HLA) and this mismatch is used to elicit a strong xenogeneic immune response. As demonstrated in Figure IB as an example, the cell-based immunogen contains MHC of a human HLA allotype, whereas the host has the native mouse MHC, H2-K, H2-D and H2-L.

In a xenopriming context, or optionally in allopriming, it can be advantageous to enhance the immune response by matching the CD8 co-receptor to the species of the class I MHC presented by the APC, so that there is native engagement between the MHC-binding site of the CD8 and the CD8-binding site of the MHC. Thus, where the MHC of the APC is human, or comprises a human al, a2 and a3 domain, the CD8 encoded by the genome of the animal may comprise a human MHC class I binding site. The CD8a and/or CD8|3 may be chimaeric (comprising part human, part non-human animal (e.g., endogenous) sequence) or may be fully human. The laboratory animal genome may encode CD8a and CD8|3 each comprising a human binding domain for MHC class I. Optionally, the laboratory animal genome comprises a CD8a gene comprising exon 1, exon 2, exon 3 and at least a part of exon 4 of human CD8a. Optionally, the laboratory animal genome comprises a CD8|3 gene comprising exon 1, exon 2 and at least a part of exon 3 of human CD8|3.

The presence of human CD8a|3 co-receptor in the animal facilitates host T cell interaction with, and activation by, the APC presenting human class I MHC-peptide complex, in order to elicit strong immune responses. LeFace et al (J. Exp. Med 1995, 182: 1315-1325) demonstrated that hCD8 transgenic mice had an enhanced response to foreign HLA class I, compared with mice with endogenous mouse CD8. The human CD8a|3 co-receptor has multiple interaction sites on human HLA I, so humanisation of CD8a|3 ensures optimal coreceptor function. Additionally, TCRa|3 humanisation in the animal enables the identification of human TCRs.

Animals (e.g., mice) for use in the present invention may thus have a genotype of humanised TCRaP and humanised CD8a|3, and optionally either humanised MHC (e.g., HLA class I) or endogenous (e.g., mouse) MHC.

All human HLA class I allotypes share sequence similarity, but alignment of HLA sequences reveals 3 major clans: HLA-A*23/A*24 clan; HLA-A*03/A*ll clan; and HLA-A*02/A*68 clan. The HLA-A*23/A*24 clan is the most distant clan diverged from the common ancestor (Figure 9). Mouse MHC (H2-D1, H2-K1) are relatively more diverged in sequence from the human HLA allotypes.

For allopriming, MHC pairings may be either within a clan or between different clans. In one embodiment, for allopriming between different clans of human MHC, the APC presents MHC-peptide complex in which the MHC component has an allotype of the HLA-A*23/A*24 clan (i.e., any HLA-A*23 or HLA-A*24 allotype) and the genome of the laboratory animal lacks any allotype of the HLA-A*23/HLA-A*24 clan and expresses an HLA allotype of the HLA- A*03/A*ll clan (i.e., any A*01, A*03, A*ll, A*25, A*26, A*29, A*30, A*31, A*32, A*33, A*34, A*66 or A*74 allotype) or the HLA-A*02/A*68 clan (i.e., any A*02 or A*68 allotype). In another embodiment, the APC presents MHC-peptide complex in which the MHC component has an allotype of the HLA-A*03/A*ll clan and the genome of the laboratory animal lacks any allotype of the HLA-A*03/A*ll clan and expresses an HLA allotype of the HLA- A*23/A*24 clan or the HLA-A*02/A*68 clan. For example, the MHC component of the APC may be HLA-A*03 and the laboratory animal may be transgenic for HLA-A*02. In another embodiment, the APC presents MHC-peptide complex in which the MHC component has an allotype of the HLA-A*02/A*68 clan and the genome of the laboratory animal lacks any allotype of the HLA-A*02/A*68 clan and expresses an HLA allotype of the HLA-A*23/A*24 clan or the HLA-A*03/A*ll clan.

The demonstrated success of allopriming, shown by experiments reported herein with HLA- A*24:02 or HLA-A*03:01 immunisation of HLA-A*02:01 host mouse, and xenopriming, shown by experiments reported herein with HLA-A*02:01 priming of endogenous mouse MHC host, can be operated with other allogeneic and xenogeneic pairings, selected according to the target MHC-peptide complex of interest and the desired laboratory animal species and/or transgenic animal genotype. The genome of a laboratory animal may be engineered to express any naturally occurring or synthetic MHC. For example, the genome of the animal may encode non-endogenous MHC, such as MHC of a non-human primate species or a computationally generated ancestral MHC sequence (e.g., representing a common ancestor of human and chimpanzee), and such animal will generate a xenogenic immune response to a delivered APC presenting an MHC-peptide complex in which the MHC component is of any human HLA allotype.

Preferably, a cohort of laboratory animals, comprising multiple identical animals, is immunised with the APC. Delivery of the APC to multiple animals, e.g., at least 3, 6 or 9 animals, increases the probability of obtaining a strong immune response in at least one animal, and may increase the number of different T cells and the diversity of TCRs that are recovered. Methods of the invention may thus be performed with multiple animals, e.g., mice.

Optionally, the invention is performed using multiple animals or multiple cohorts of animals, wherein the genome of each animal or cohort of animals encodes a different MHC. For example, a plurality of animals may comprise genomes encoding different HLA transgenes, e.g., a transgene of the HLA-A*23/A*24 clan, a transgene of the HLA-A*03/A*ll clan and a transgene of the HLA-A*02/A*68 clan. Thus an APC comprising an HLA-peptide complex may be delivered to a plurality of animals representing multiple different allopriming pairings. For example, an APC comprising an HLA-peptide complex may be delivered to a plurality of animals, wherein the HLA component of the HLA-peptide complex is of a first HLA allotype, and wherein the plurality of animals comprises a first animal (or first cohort of multiple identical animals) whose genome encodes an HLA of a second allotype, and a second animal (or second cohort of multiple identical animals) whose genome encodes an HLA of a third allotype, wherein the first, second and third allotypes are different HLA allotypes, and are optionally each selected from a different HLA allotype clan, wherein the HLA allotype clans are the HLA-A*23/A*24 clan, the HLA-A*03/A*ll clan and the HLA-A*02/A*68 clan.

A plurality of animals may include animals for allogeneic, xenogeneic and syngeneic priming. Thus, for example, an APC comprising an HLA-peptide complex may be delivered to a plurality of animals, wherein the plurality of animals comprises a first animal (or first cohort of multiple identical animals) whose genome encodes an HLA of an allotype different to the allotype of the HLA component of the APC and/or a second animal (or second cohort of multiple identical animals) whose genome encodes no HLA, and wherein MHC in the genome is optionally wild type MHC endogenous to the animal, or is a non-human MHC transgene, and optionally a third animal (or third cohort of multiple identical animals) whose genome encodes an HLA of the same allotype as the HLA component of the APC.

An advantage of this approach is that animals with different HLA allotypes have different TCR repertoires shaped by positive and negative selection. By immunsing animals with different TCR repertoires, there is a further enhanced opportunity to discover high affinity TCRs from different TCR clonotypes, representing an excellent starting point for TCR drug therapeutics.

Brief Description of the Drawings

Figure 1: Concept illustration. A, in vivo immunisation of mice. B, genotype of the cell immunogen and mouse platform for allo- and xeno-priming.

Figure 2: Illustrates mouse APC cell line DC2.4 engineering for antigen presentation. A: mouse B2m staining demonstrating the complete loss of mouse B2m in knockout cells compared to parental control. B: CD80 and CD86 surface upregulation when engineered DC2.4 cells were stimulated with LPS and I FNy.

Figure 3: Illustrates antigen specific CD8+ T cell responses obtained from allo allopriming. A: FACs plots showing antigen specific and non-specific tetramer staining of live CD8+ T cell population. Quadrant Q3 shows the amount of antigen specific response when mice immunised with antigen presenting cells with target and HLA overexpression compared to HLA only overexpressing cells. B: The proportion of cells from each quadrant shown in Figure 3A were summarised across different time points for each animal.

Figure 4: Bioinformatics analysis of sorted CD8+ T cells based on barcoding tetramer and cellular phenotype at single cell level.

Figure 5A: Data showing recovered TCR binding capacity measured by ELISA.

Figure 5B: Data showing affinity of recovered TCRs determined by SPR with TCR-Fc as analyte and captured MHC-peptide complex as ligand.

Figure 5C: Data showing affinity of recovered TCRs determined by SPR with soluble MHC- peptide complex as analyte and captured TCR-Fc as ligand.

Figure 5D: Affinity measurement of TCRs generated by allopriming, determined by SPR.

Figure 6: Affinity measurement of TCRs generated via syngeneic priming determined by SPR. Note that the detection limit is 100 uM and any weak binders with affinity close to or beyond detection limit is shown as lOOuM. Figure 7: Antigen specific CD8+ T cell responses obtained from xenopriming. FACS analysis demonstrates the antigen specific response measured by proportion of antigen specific pMHC tetramer positive and irrelevant peptide pMHC tetramer negative CD8 T cell population. Plots show antigen specific and non-specific tetramer staining of live CD8+ T cell population in mice immunised with target-presenting APC (left) or negative control APC (right).

Figure 8: Affinity measurement of a selected TCR generated by xenopriming, determined by SPR.

Figure 9: Phylogenetic tree of MHC class I sequences showing H2-D1 and H2-K1 allotypes (mouse), HLA-A*23/A*24 clan (human), HLA-A*03/A*ll clan (human), and HLA-A*02/A*68 (human). HLA-A*23/A*24 clan comprises HLA-A*23 and HLA-A*24 allotypes. HLA-A*03/A*ll clan comprises HLA-A*01, A*03, A*ll, A*25, A*26, A*29, A*30, A*31, A*32, A*33, A*34, A*66 and A*74 allotypes. HLA-A*02/A*68 clan comprises HLA-A*02 and HLA-A*68 allotypes.

Figure 10: Western blot analysis demonstrating the loss of mouse TAPI and TAP2 protein in DC2.4 cells.

Figure 11: Affinity measurement of TCRs generated by allopriming, determined by SPR with representative sensorgrams from selected high affinity TCRs.

Figure 12: Antigen specific CD8 T cell responses obtained from allopriming against neoantigen pMHC target. FACS analysis demonstrating neoantigen pMHC specific CD8 T cells in mice using dual neoantigen (PE) and wild type peptide (APC) pMHC tetramer staining of live CD8+ T cell population in mice immunised with target-presenting cells (left) or negative control cells (right), recovered after prime (top) or after one boost (bottom).

Figure 13: The specificity of TCR discovered from allopriming with a neoantigen pMHC as the target, measured in Jurkat signalling reporter T cell assay.

Detailed Description

In the present invention, presentation of a peptide antigen by a non-self MHC allows for the identification in vivo or ex vivo of T cells of relatively higher affinity to the presented non-self MHC -peptide complex compared to syngeneic MHC priming (syngeneic priming), as these T cells are not subject to negative selection. The non-self MHC can be an allotype from the same species, where it is termed allogenic MHC priming (a I lo-p rim i ng) or from a different species, which is termed xenogenic MHC priming (xeno-priming).

In particular, where the peptide is a self-peptide, or is highly homogolous to a self-peptide, T cells that recognise the peptide MHC complex will not be subject to negative selection in vivo, because the peptide - MHC complex is non self and the mature T cell repertoire is not negatively selected for non self peptide MHC complex.

We have observed that using allo-priming and xeno-priming methods disclosed herein, TCRs can be identified with relatively higher affinity for peptide- MHC complexes than typical TCR affinity against self peptide MHC. The MHC-peptide complex that is delivered to generate T cells may be referred to herein as the target MHC or target MHC-peptide complex.

Disclosed herein is therefore an improved way to generate higher affinity TCRs in vivo by employing a mismatch of MHC type found in the host species (e.g. mouse) and the MHC allotype used to present the peptide on the antigen presenting cell (APC). TCRs that bind to a target peptide-MHC complex can be discovered by these methods and their sequences, or derivatives thereof, e.g. variable regions or domain(s) thereof with peptide binding can be developed as therapeutic molecules for administration to patients, where the therapeutic molecule will bind cells presenting the target peptide MHC complex in the patient.

In the prior art, transgenic mice have previously been used to address the issue of tolerance, but in a different way. Specifically, mice expressing a human TCR repertoire restricted to the human leukocyte antigen HLA-A*02:01 were given a "non-self" immunogen (a human peptide) with low amino acid sequence homology to the mouse counterpart, expected to be immunogenic in the mouse. This mouse was used to identify high affinity TCRs to the non- self-peptide in the mouse, because such TCRs would not have been deleted through negative selection process. The HLA-transgenic mouse of the prior art therefore already had the advantage of avoiding negative selection of TCRs that recognised the target peptide-HLA, because the peptide (the priming antigen) was already non-self. See for example Obenaus et al (above).

Distinct from the in vivo platforms, alternative in vitro/ex vivo approaches have been tested for TCR discovery. Schendel (Frontiers in Oncology Front. Oncol. 13:1216829. doi:

10.3389/fonc.2023.1216829) discloses adding a non-self-HLA allotype to autologous DC and using these APC for HLA allo-restricted peptide presentation and stimulation of autologous donor T cells. However, this approach is limited compared with antigen presentation in vivo, e.g. in a laboratory animal such as a mouse or rat, because the donor T cells have a limited TCR repertoire. TCR diversity decreases with age, both due to the drop of the percentage of the naive T cells and due to the decrease of diversity within the naive T-cell pool. In mature individuals, development of new T cells in the thymus slows down due to thymus involution and T-cell numbers are maintained through division of mature T cells outside of the central lymphoid organs. In addition, the increased representation of antigen experienced memory T cells means such repertoire has low capacity in responding to new antigen. Therefore, approaches using autologous human T cells from healthy donors are inherently limited in the reactive T cell repertoire available. Amir et al (Clin Cancer Res. (2011) 17(17):5615-5625) disclose that higher potency TCRs could be identified in vivo from GvHD patients who underwent stem cell transplant with mismatched HLAs to the donor vs matching donor. The increase in potency is thought to be due to the increased TCR affinity via allo-priming. However, such seting would not be possible for TCR therapeutics discovery.

The use of laboratory animals in the present invention, such as rodents with a high percentage of naive T cells (e.g. using juvenile rodents of 6-8 weeks) provides a much wider T cell repertoire on which to start a selection for high affinity TCR. Therefore, the present invention advantageously accesses a T cell repertoire that is of a much broader profile than can be provided by T cells from a mature adult. Furthermore, since MHC allotype inherently shapes the TCR repertoire, immunisation of humanised rodents such as mouse with different MHCs could potentially further increase the accessible TCR repertoire.

In addition, the natural process of T cell generation and selective expansion in vivo, or priming and expansion of T cells ex vivo, may remove the need for any subsequent TCR affinity maturation for applications such as TCR based cell therapy.

Disclosed herein is a method for generation of a TCR in a laboratory animal, the method comprising i Delivering an antigen presenting cell to the laboratory animal, wherein the antigen presenting cell expresses an MHC-peptide complex in the laboratory animal, and wherein the MHC component of the MHC-peptide is not expressed in the laboratory animal ii generation of T cells to the MHC-peptide complex in the laboratory animal; and iii optionally, isolation and or purification of T cells from the laboratory animal reactive with the MHC-peptide complex, such as a polyclonal mixture of T cells or a monoclonal T cell, or cell lines thereof.

The invention also relates to a polyclonal mixture of T cells obtained or obtainable herein, and use of said polyclonal mixture in screening for TCR and T cells having binding to MHC- peptide complexes. In one aspect the MHC-peptide complex is not expressed from the genome of the laboratory animal because it is not encoded by the laboratory animal genome. In one aspect the MHC-peptide complex is not expressible within the genome, for example because it has been inactivated by mutation of the MHC gene or regulatory elements. Thus, the MHC component of the MHC-peptide complex, which is presented by the antigen presenting cell in the laboratory animal, is not expressed by cells of the laboratory animal. The MHC-peptide complex is only expressed in the animal through delivery of the antigen presenting cell, which presents the MHC-peptide complex at its cell surface.

Laboratory animals

Suitable laboratory animals include mice, rats, guinea pigs and rabbits , hamster, gerbils, chinchillas, voles, woodchucks, dogs, cats, swine, sheep and nonhuman primates. See MSD veterinary Manual 2023, section entitled Mice and Rats as Laboratory Animals

By Jennifer Frohlich, VMD, DACLAM, Office of Laboratory Animal Care, University of California, Berkeley (Reviewed/Revised Feb 2021).

In one aspect the laboratory animal is a rodent, such as a mouse or such as a rat.

TCR and MHC expression

In one aspect the MHC molecule expressed on the APC is a human HLA molecule. In one aspect the genome of a laboratory animal encodes a human MHC class I gene encoding a human HLA. In one aspect the laboratory animal has a knockout of the host MHC class I gene, such that the laboratory animal genome expresses only a human HLA class I gene product from a human MHC class I gene.

In one aspect the genome of a laboratory animal encodes a human MHC class II gene encoding a human HLA class IL In one aspect the laboratory animal has a knockout of the host MHC class II gene, such that the laboratory animal genome expresses only a human HLA class II gene product from a human MHC class II gene.

Mice expressing human HLAs are known in the art, for example as described in Obenaus et al (above) and references therein, such as Li, L.-P. et al. Transgenic mice with a diverse human T cell antigen receptor repertoire. Nat. Med. 16, 1029-1034 (2010). Further examples of class I HLA transgenes in mice are described in Pascolo et al., 1997 (J Exp Med 185(12):2043-2051 1997), W02014/130671, Wang et al., 2016 (Cancer Immunology Research 4(3):204 2016), Moore et al., 2021 (Science Immunology 6(66): 2021), and WO2021/139799. Examples of class II transgenes in mice are described in W02014/130671, and Ito et al., 1996 (J Exp Med 183:2635-2644 1996), Yatsuda et al., 2013 (PloS ONE 8(12):e84908 2013), Chen et all, 2017 (J Exp Med 214(ll):3417-3433 2017), Poncette et al., 2019 (J Clin Invest 129(l):324-335 2019) and Moore et al., 2021 (Science Immunology 6(66): 2021).

In one aspect the genome of a laboratory animal encodes a fully human TCR. Mice containing human TCR are known in the art, for example as described in Obenaus et al (above).

In one aspect the genome of a laboratory animal encodes a chimaeric TCR, a part of which is a human sequence and a part of which is a sequence of the laboratory animal TCR. The TCR in one aspect comprises at least a human variable region.

The laboratory animal genome may comprise, for example:

(i) a replacement of an endogenous T cell receptor (TCR) variable a (Va) gene with unarranged human TCR Va segments and unrearranged human TCR Ja gene segments, wherein the unrearranged human TCR Va and Ja gene segments are operably linked to a TCR constant a (Ca) gene of the laboratory animal, and/or

(ii) a replacement of an endogenous T-cell receptor (TCR) variable |3 (VP) gene with unrearranged human TCR VP gene segments, unrearranged human TCR DP gene segments, and unrearranged human TCR jp gene segments, wherein the unrearranged human TCR VP, DP, and jp gene segments are operably linked to a TCR constant P(CP) gene of the laboratory animal.

Mice encoding such chimaeric TCR are described, inter alia, US9113616B2 and Moore et al., Sci. Immunol. 6, eabj4026 (2021) 17 December 2021.

These mice are all suitable for use in the present invention and the teachings are incorporated by reference. Another suitable mouse is disclosed in co-pending UK patent application GB2312191.6 and GB2405373.8. See WO2025/032222. The laboratory animal may be a genetically modified rodent (e.g., mouse) as described in WO2025/032222, or may be a rodent (e.g., mouse) comprising a TCRa locus, TCR|3 locus, CD8a locus, CD8|3 locus, MHC class I locus and/or |32m locus described in WO2025/032222.

In one aspect the laboratory animal genome encodes a human TCR or variable region of a human TCR.

In one aspect the laboratory animal encodes a human MHC, as described herein.

As disclosed herein the MHC-peptide complex in the APC is not expressed from the genome of the laboratory animal, and in one aspect is not expressible. This creates a mismatch between the MHC of the APC and the MHC of the laboratory animal in the method of the invention.

Where the MHC type is from a different species to the MHC type found in the laboratory animal, this is described herein as xeno-priming.

Where the MHC type is from the same species to the MHC type found in the laboratory animal, but is a different sequence of MHC molecule, this is described herein as allo-priming.

In one aspect the laboratory animal genome comprises naturally occurring MHC genes, and the MHC molecule expressed on the APC is from a human (i.e. a human HLA molecule) . For example, where the laboratory animal is a mouse, the mouse genome contains mouse MHC class I genes expressing mouse MHCs, and the APC delivered to the mice express human HLA class I molecules. This is an example of xenopriming. Optionally in xenopriming embodiments the laboratory animal has no MHC transgene, so only has endogenous wild type MHC.

In one aspect the laboratory animal has a deletion or other inactivation of the natural MHC gene or genes, such that there is no expression of host MHC molecules, and the laboratory animal has a knock in of a gene which expresses an MHC found in a different member of the same species or from a different species, such as a human. Where the knock in is of a human class I HLA, the APC delivered to the laboratory animal can be selected to express a different type of human HLA. This is an example of allopriming.

Laboratory animals comprising a human HLA knock in and primed with an allogenic human HLA via the APC are particularly useful in the generation of TCR that can be used in humans as therapeutics. Specifically, the TCR that have suitably high affinity can be used in human patients who express the same HLA type as is present on the APC (the same target MHC).

The method of the invention can be used to generate TCR to peptides expressed by both MHC class I and MHC class II molecules. It will be appreciated that the reference to allopriming or xenopriming refers to a difference between different HLA class I molecules, or between different HLA class II molecules, but is not used to refer to a difference between HLA class I and HLA class II molecules. In one aspect the laboratory animal genome encodes a human HLA class I molecule, such as a human HLA-A*02, HLA-All*:01,HLA-A*ll:01, HLA-A*03:01, or HLA-A*24:02 knock in.

In one aspect the laboratory animal is a mouse and genome encodes a human HLA- A02*:02:01 knock in.

In one aspect the laboratory animal is a mouse and the genome encodes HLA-A02*:01, and is immunised by an APC expressing a peptide together with HLA-A*ll:01, HLA-A*03:01 or HLA-A* 24:02.

A soluble TCR, or portion thereof, as disclosed herein, is not associated with a T cell or cell membrane.

The TCR variable region and variable region domain(s) referred to herein suitably retain the same binding affinity for the peptide-MHC complex as the full length TCR, or at least 50% or more of that affinity.

In one aspect the invention relates to a soluble TCR, or a variable region thereof, or domains thereof, obtained by priming in vivo or ex vivo, by any method disclosed herein.

In one aspect the invention relates to a bispecific molecule or multispecific comprising a TCR arm, for example comprising all of a part of a TCR molecule produced by a method disclosed herein, for example wherein the part may be a TCR variable region, and to a laboratory animal encoding such a bispecific or multispecific.

TCRs or TCR variable regions or domain(s) produced by methods disclosed herein may be mutated by one or more amino acids, such as 2, 3, 4 or 5 amino acids, when compared to the original sequence identified in the laboratory animal, for example to improved specificity and/or affinity and/or solubility.

In one aspect the laboratory animal genome encodes a CD8 molecule that is matched in terms of species origin to the MHC class I, at least for those molecular regions that are known to interact with one another. Therefore, where the target MHC- peptide complex is a fully human HLA, then the CD8 may be a human CD8, or a chimaeric CD8 in which the region that interacts with the MHC class I is a human region.

In one aspect the laboratory animal genome encodes a CD4 molecule that is matched in terms of species origin to the MCH class II, at least for those molecular regions that are known to bind. Therefore, where the target MHC-peptide complex is a fully human HLA, then the CD4 may be a human CD4, or a chimaeric CD4 in which the region that interacts with the MHC class II is a human region.

In one aspect the MHC of the target MHC-peptide complex may be a chimaeric MHC e.g. a human-mouse chimaeric MHC. The same approach to matching the MHC and CD8 or MHC and CD4 mentioned above can also apply in this chimaeric scenario, providing an effective MHC and CD8 or CD4 interaction, as appropriate for the class of MHC. Thus it is possible to use a chimeric HLA in an antigen presenting cell, where the alpha 3 domain is mouse, to accommodate mouse CD8 binding. Alphal and Alpha2 domains of the chimaeric HLA remain human to ensure human antigen presentation. This way the laboratory animal does not require CD8 humanisation.

Antigen Presenting Cells

Any APC may be used to immunise the laboratory animal. In one aspect the APC is a dendritic cell. Methods for immunisation using antigen presenting cells is known in the art, as disclosed in e.g. Junji Yatsuda et al PLOS ONE | www.plosone.org 1 December 2013 | Volume 8 | Issue 12 | e84908.

The APC may be transfected with a DNA vector or mRNA from which the intended MHC and/or additional co-stimulatory receptors is expressed. The APC may also be co-transfected with a DNA vector, mRNA or peptide loaded with target peptide, for surface presentation of the target peptide on class I MHC or Class II MHC.

The APC may be transfected with nucleic acid comprising a full length gene encoding the polypeptide from which the peptide of the MHC-peptide originates (i.e., the peptide is a fragment of the encoded polypeptide). Where the MHC is class I MHC, this technique may allow peptide-agnostic TCR discovery - TCRs will be generated against whichever peptide or peptides are presented by the APC following over-expression of the gene and antigen processing within the cell. This is especially advantageous when it is unknown ab initio which peptide fragment of the polypeptide is most strongly presented as an MHC-peptide complex and/or which peptide is most immunogenic. For species-faithful antigen processing, it is preferable in this situation for the APC to be of the same species as the MHC component of the MHC peptide complex it presents, for example a human APC such as K562, presenting HLA class I peptide complex. Alternatively, the APC may be transfected with nucleic acid comprising a combination of ubiquitin-peptide minigenes. The selection of such a peptide combination can be based on in silico predictions, e.g. target HLA binding strength, human TAP processing efficiency; or one can use a peptide library using a tiling approach without prior in silico selection. Since a ubiquitin-peptide minigene is processed very efficiently, such an approach could ensure high surface pMHC presentation for all peptide candidates without the peptide processing bias/limitation introduced by e.g. proteasome processing and TAP transporter peptide sequence bias. A plurality of TCRs are generated in vivo, which may include TCRs specific for different MHC-peptide complexes, containing different peptide fragments of the overexpressed gene or peptides expressed from different individual peptide minigenes. The specificity of an individual TCR can subsequently be determined by detecting which MHC-peptide complex it binds. This can be done for example by a method comprising: (i) providing a panel of cells each presenting an MHC-peptide complex of known sequence, wherein each cell presents a different peptide fragment of the polypeptide, (ii) contacting the panel of cells with a TCR, (iii) detecting binding of the TCR to a cell of the panel, and (iv) identifying the MHC-peptide complex recognised by the TCR as being a complex of MHC and the peptide presented by that cell. One approach is to use an immunogencity readout such as ELISPOT by co-culturing reactive T cells derived from the animal with pMHC-presenting cells and to detect binding by measuring IFN gamma release. Other possibilities use monomeric pMHC or tetrameric pMHC for each candidate peptide to identify TCR binding specificity by either staining/sort reactive T cells or biochemical binding assays of monomeric or tetrameric pMHCs with all TCR sequences derived from the reactive T cell pools converted to soluble forms.

The species of the APC may match the species of the immunisation animal. The APC may be a mouse cell for mouse immunisation. To improve specificity of the T cell immune response, the APC may not present MHC other than the target MHC. The APC may not present class I MHC other than the target MHC. For example, knock out of endogenous |32 microglobulin in mouse dendritic cells prevents surface presentation of endogenous mouse MHC class I, so that an overexpressed single chain |32m MHC fusion molecule is the sole class I MHC presented at the cell surface. This restricts the immune response in the immunised animal to peptides presented by the target MHC.

A cell line deficient in the endogenous antigen processing/antigen presentation may be used to further improve specificity of the T cell response. For example, a TAP-deficient APC, e.g. TAPI and/or TAP2 deficient APC, or TAP 1 and/or TAP2 deficient, plus HM13 deficient APC, may be used. Knock-out of TAP inhibits the cell's endogenous antigen presentation pathway, so that a target peptide expressed with an N terminal signal peptide (e.g., from an exogenously introduced expression vector) is the sole peptide presented on MHC class I. In addition, signal peptide peptidase (SPP, encoded by HM13 gene) loss is valuable for enhancing HLA-A*02:01 peptide presentation specificity via the signal peptide driven target peptide transport, as hydrophobic signal peptide byproduct could be processed by SPP in ER and presented by HLA-A*02:01 due to its preferences for aliphatic amino acid residues that are commonly found in hydrophobic region of signal peptides (Bruno P et al. Nature Biotechnology 2023 Jul;41(7):980-992). Optionally, a full length gene is over-expressed within the cell (e.g., by introduction of an encoding expression vector as discussed above), wherein the gene comprises an N terminal signal peptide (either a native signal peptide of the gene, or one that is heterogenously added to the 5' end of the coding sequence of the gene). In the cell line deficient in antigen processing/presentation, peptide fragments of this gene are the only peptides presented at the cell surface as MHC-peptide complexes.

Examples of suitable cell lines for use as APC include DC2.4 (mouse dendritic cell line) and K562 (human cell line that may be used as a "chassis" antigen presenting cell by introducing a monoallelic HLA of interest).

An APC may be an artificial antigen presenting cell, for example as described in Perica et al Nanomedicine. 2014 Jan; 10(1): 119-129., and J Immunol Res Ther. 2017; 2(1): 68-79. , Neal et al.

It will be appreciated that a reference to antigen presenting cells (APC) includes presentation formats that may not be biological cells, but which may be formats, for example particles, that are able to present peptide to T cells, and which are included here within the definition of APCs.

Further details of antigen presenting cells that may be used for immunisation in the present invention are provided in the accompanying Examples.

Peptides

In one aspect peptides that form a part of the MHC-peptide complex are human peptides.

Reference to a "human" peptide or a "human" antigen herein is to a peptide or antigen that has been identified to be present in a human or human cell, or is encoded by a human genome. A human peptide or human antigen might therefore be seen in another species but must be found in humans. A similar nomenclature is taken to peptides or antigens from other species.

In one aspect peptides that form a part of the MHC- peptide complex are peptides that differ from any peptide sequence naturally in the laboratory animal, such as by at least 1 amino acid.

In one aspect peptides that form a part of the MHC-peptide complex are peptides that differ from any peptide sequence naturally in the laboratory animal by only 1 amino acid, by only 2 amino acids, or by only 3 amino acids or more.

In one aspect peptides that form a part of the MHC-peptide complex are peptides that are the same as peptide sequence naturally found expressed in the laboratory animal.

It will be appreciated that where the delivered peptide, such as a human peptide, and a selfpeptide are the same, or very similar in sequence (such as varying by 1, 2, or 3 amino acids), negative selection may occur in vivo, which may be overcome by the allopriming or xenopriming approach in vivo.

However, even if the delivered peptide is different from the self-peptide, the ability for the TCR repertoire to be selected in vivo based upon binding and maturation to both the peptide and MHC of the MHC-peptide complex (neither of which are self) still provides an advantages over an approach where MHC types are matched in vivo, as TCR that bind both to the non- self-peptide and non self MHC can contribute to the population of TCRs for selection.

The present in vivo approach is therefore applicable to the use of any peptide, even if it is not known or not possible to predict whether an antigen to which it is desired to raise a T cell response is a self-antigen or non-self in the laboratory animal in question.

TCR properties

The present invention provides a population of T cells with high affinity binding to a peptide that can be assessed for therapeutic suitability, such as for off target binding.

In one aspect the population of T cells obtained through allo or xeno priming as a whole has a greater quantity of binding, as measured for example by FACS, e.g. by antigen specific tetramer binding in FACS (see examples disclosed herein), when compared with the population obtained by syngeneic priming.

In one aspect the median binding affinity of a random population of 50 or 100 or 200 TCRs obtained by allo or xeno priming as disclosed herein is greater (higher affinity, lower KD) than the median binding affinity of a random population of 50 or 100 or 200 TCRs obtained by syngeneic priming to the same target, such as at least 10 fold greater, such as at least 100 fold greater.

In one aspect the population above obtained by allo or xeno priming comprises a TCR that has a higher affinity (lower KD) than any TCR obtained by syngeneic priming, in respect of the same peptide target sequence, such as at least 10 fold greater affinity, such as at least 100 fold greater affinity.

In one aspect a TCR obtained by allo or xeno priming comprises a TCR that has a higher affinity than any TCR obtained by syngeneic priming, in respect of the same peptide target sequence.

In one aspect the median binding affinity of the top 5, 10 or 15 highest affinity TCRs identified by allo or xeno priming is greater than the median affinity of the top 5, 10 or 15 highest affinity TCR identified by syngeneic priming, respectively, in respect of the same peptide target sequence.

Comparisons between allo/xeno priming and syngeneic priming can be made using methods as disclosed herein, such as Example 3.

In one aspect, the TCR directly obtained from the methods described herein has an affinity (KD) of lpM (1 micromolar) or less, such as 500nM or less, 400nM or less, 300nM or less, 200nM or less, preferably lOOnM or less. Affinity may be measured by methods disclosed in Examplel. A suitable method comprises SPR in which the TCR-Fc molecules are captured on a solid support e.g. onto a Protein G chip as the ligand, and the analyte comprises the MHC peptide complex , run over the ligand at 25 °C.

Formulations and methods

T cells have high affinity binding may be useful as therapeutic agents in treatment of human disease. In particular, the identification of T cells with high affinity binding and specificity for the target peptide allows the TCR to be sequenced, expressed and formulated with pharmaceutically acceptable excipients.

The T cell receptor sequences are suitable for use in patients who have the same MHC type as the delivered MHC of the MHC-peptide complex used in the priming method.

Methods disclosed herein therefore can comprise the step(s) of determining a nucleic acid sequence of the human TCR variable region(s) (or whole TCR) expressed by a T cell in the laboratory animal which is reactive to the MHC-peptide complex and a) expressing the human T cell receptor, or human T cell receptor variable region or domain(s) in a cell; optionally further formulating the expressed human T cell receptor, or human T cell receptor variable region or domain(s) with a pharmaceutically acceptable excipient; or b) inserting the nucleic acid encoding the human T cell receptor, or human T cell receptor variable region, into a cell, ex vivo or in vitro, such as into a human or animal cell, optionally wherein the cell containing the inserted nucleic acid is formulated for delivery to a human or animal, respectively; or c) formulating a nucleic acid (e.g. RNA or DNA) encoding the human T cell receptor, or human T cell receptor variable region, with a suitably delivery vehicle, such as a lipid or liposome, for delivery to a patient in need thereof in vivo.

The T cell receptor variable region or domain(s) are optionally expressed or encoded as part of a T cell receptor or as part of a soluble TCR molecule.

The TCR sequence, or TCR variable region sequence, or TCR domain(s) sequence may also be incorporated into or expressed as a part of a larger molecule, e.g. as a part of a bispecific molecule (or multi specific molecule) with one arm being the TCR or part thereof and the other specific arm being, for example, a part of an antibody or another TCR arm having a different specificity. For example the other arm of a bispecific can be an anti-CD3 arm.

Modifications to the TCR sequence, or TCR variable region sequence, or TCR domain(s) sequence identified in the methods of the invention may be made before therapeutic use. For example, changes to the nucleic acid that result in 1 or more, e.g. 2 or more or 3 or more amino acid changes in the TCR sequence may be made.

Modification to change the amino acid sequence of the TCR identified in the methods of the invention may be made before or after the inclusion of the TCR sequence into a different format, such as a bispecific or multi specific molecule or in a soluble format.

Methods of treatment

The invention also relates to a method of treating an individual in need thereof, the method comprising

(i) delivery of a cell according to step (b) above into the patient in need thereof, or

(ii) delivery of a nucleic acid (e.g. RNA e.g. mRNA or DNA) encoding the human T cell receptor, or human T cell receptor variable region formulated as in step (c) above into a patient in need thereof; or

(iii) delivery of a TCR polypeptide or variable region thereof, or domain(s) thereof, optionally in the form of a soluble TCR, region or domain(s) thereof, a bispecific molecule having as one arm a TCR or a variable region thereof or a domain(s) thereof, e.g., a TCR immune cell engager, such as anti-CD3 bispecific.

The invention also relates to a T cell receptor or T cell receptor variable region obtained or obtainable from the method of any preceding claim, for use in a patient having a host MHC (e.g. a host HLA) which is the same as that used to immunise the rodent and expressed by the APC.

Examples of TCR therapeutics are known and described in, for example, Robinson et al The FEBS Journal 288 (2021) 6159-6173 and Klebanoff et al Nature Reviews Drug

Discovery volume 22, pages996-1017 (2023). These also disclose suitable formats for use of TCR sequences identified by the present invention, such as soluble formats and TCR like molecules.

Clauses

The following numbered clauses represent embodiments of the invention.

1 A method for generation of a TCR in a laboratory animal, the method comprising

1 Delivering an antigen presenting cell to the laboratory animal, wherein the antigen presenting cell expresses an MHC-peptide complex in the laboratory animal, and wherein the MHC component of the MHC-peptide is not expressed in the laboratory animal. ii generation of T cells to the MHC-peptide complex in the laboratory animal; and iii optionally, isolation and/or purification of T cells from the laboratory animal reactive with the MHC-peptide complex.

2 The method of clause 1 wherein the MHC component of the MHC-peptide complex is allogenic.

3 The method of clause 1 wherein the MHC component of the MHC-peptide complex is xenogenic.

4. The method of any preceding clause, wherein the laboratory animal genome comprises human CD8 or comprises chimaeric CD8 having a region of human sequence and a region of sequence from the laboratory animal, wherein the region that interacts with the MHC is a human region.

5. The method of clause 4, wherein the MHC component of the MHC-peptide in the antigen presenting cell is human HLA class I and the laboratory animal genome does not encode any human HLA.

6. The method of clause 5, wherein the laboratory animal genome comprises endogenous MHC class I.

7. The method of any of clauses 1 to 4, wherein the MHC expressed in the laboratory animal genome is a human HLA. 8. The method of clause 7 wherein the laboratory animal genome encodes a human HLA-A*02:01, or HLA-A*ll:01, or HLA-A*03:01, or HLA-A*24:02 knock in.

9. The method of any preceding clause wherein the laboratory animal genome encodes a fully human TCR.

10. The method of any preceding clause wherein the laboratory animal genome encodes a chimaeric TCR, having a region of human sequence and a region of sequence from the laboratory animal.

11. The method of any preceding clause wherein the peptide is from a human antigen the expression of which is associated with disease, such as cancer, for example wherein the antigen is a tumour associated antigen.

12. The method of any preceding clause wherein the peptide is from a human antigen which has the same sequence as an antigen expressed in the laboratory animal, or is a variant of an antigen expressed in the laboratory animal that differs by only 1 or 2 or three amino acids.

13. The method of any preceding clause wherein the laboratory animal is a rodent, such as a mouse or rat.

14. The method of any preceding clause wherein the MHC expressed on the antigen presenting cell is expressed from a human MHC class I gene.

15. The method of any of clauses 1 to 13 wherein the MHC expressed on the antigen presenting cell is a human-mouse chimaeric MHC.

16. The method of any one of clauses 1-13 wherein the MHC expressed on the antigen presenting cell is expressed from a human MHC class II gene.

17. The method of clause 14 wherein the MHC in the MHC-peptide complex is a human HLA-A* 24:02.

18. The method of clause 17 wherein the MHC in the MHC-peptide complex is a human HLA-A*24:02 and the MHC encoded by the laboratory animal is HLA-A*02:01.

19. The method of any preceding clause, wherein the laboratory animal is mouse and the antigen presenting cell is a mouse antigen presenting cell.

20. The method of any preceding clause, wherein the antigen presenting cell is deficient in endogenous antigen processing /presentation, optionally being deficient in TAP and/or signal peptidase. 21. A method according to any of clauses 1 to 18, wherein the APC is a synthetic antigen presenting cell, optionally where the APC is not a biological cell.

22. The method of any preceding clause wherein there is isolation and/or purification of T cells from the laboratory animal which are reactive with the MHC-peptide complex, and optionally wherein the T cells are a polyclonal mixture of T cells or a monoclonal T cell, or a cell lines thereof.

23. The method of any preceding clause comprising determining a nucleic acid sequence of the human TCR variable regions expressed by a T cell from the laboratory animal reactive to the MHC-peptide complex and i expressing the human T cell receptor, or human T cell receptor variable region or variable domain(s) in a cell; optionally wherein the receptor is soluble, optionally wherein the TCR receptor, receptor variable region or variable domain(s) is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; and optionally wherein the TCR sequence has been modified by 1 or more, e.g. 2 or more e.g. 3 or more amino acid changes from that identified in the laboratory animal, optionally further formulating the expressed human T cell receptor, or human T cell receptor variable region or variable domain(s) with a pharmaceutically acceptable excipient; or ii inserting the nucleic acid encoding the human T cell receptor, or human T cell receptor variable region, or variable domain(s) into a cell, ex vivo or in vitro, such as into a human or animal cell,; or optionally wherein the nucleic acid is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; optionally wherein the TCR nucleic acid sequence has been modified to result in 1 or more, e.g. 2 or more e.g. 3 or more amino acid changes from the sequence of the TCR identified in the laboratory animal; optionally wherein the expressed T cell receptor or part thereof is soluble; optionally wherein the cell containing the inserted nucleic acid is formulated for delivery to a human or animal iii formulating a nucleic acid (e.g. RNA or DNA) encoding the human T cell receptor, or human T cell receptor variable region, or variable domain(s), with a suitably delivery vehicle, such as a lipid or liposome, for delivery to a patient in need thereof in vivo. optionally wherein the nucleic acid of the TCR is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; optionally wherein the TCR nucleic acid sequence has been modified to result in 1 or more, e.g. 2 or more or 3 or more amino acid changes from the sequence of the TCR encoded by the nucleic acid TCR identified in the laboratory animal optionally wherein the expressed T cell receptor or part thereof is soluble.

24. A method of treating an individual in need thereof, the method comprising

(i) delivery of a cell according to step (ii) of clause 23 into the patient in need thereof, or

(ii) delivery of a nucleic acid (e.g. RNA e.g. mRNA or DNA) encoding the human T cell receptor, or human T cell receptor variable region formulated as in step (iii) above into a patient in need thereof or

(iii) delivery of a TCR polypeptide or variable region thereof, or domain(s) thereof, optionally in the form of a soluble TCR, soluble variable region or soluble domain(s) thereof, or a bispecific molecule having as one arm a TCR or a variable region thereof or a domain(s) thereof, e.g., a TCR immune cell engager, such as anti-CD3 bispecific molecule.

25. A T cell receptor or T cell receptor variable region or domains(s) thereof obtained or obtainable from the method of any preceding clause, for use in a patient having a host MHC which is the same as that used to immunise the rodent and expressed by the APC.

26. A method for generation of a TCR in a mouse, the method comprising i Delivering an antigen presenting cell to the mouse, wherein the antigen presenting cell expresses an HLA-peptide complex in the mouse, wherein the HLA component of the MHC -peptide is not expressed in the mouse, wherein the mouse genome expresses a fully human TCR, or chimaeric mouse-human TCR, but not a fully mouse TCR wherein the mouse expresses a human HLA from the mouse genome but not a mouse MHC molecule ii generation of T cells to the MHC -peptide complex in the laboratory animal; and iii optionally, isolation and/or purification of T cells from the laboratory animal reactive with the MHC-peptide complex; and iv further optionally, the identification of the nucleic acid sequence encoding the TCR, and further optionally, mutation of the nucleic acid encoding the TCR v optionally including expressing a TCR or part thereof, optionally in the form of a larger molecule or complex, such as a bispecific molecule; and vi optionally formulation with a pharmaceutically acceptable excipient or carrier.

The invention is further described by the following non limiting examples.

Example 1 TCR discovery with in vivo allopriming to generate high affinity TCRs using humanised mice

To demonstrate the discovery using allo-priming method, we immunised transgenic mice containing humanised HLA*A02:01, CD8a/b, TCRa and TCR|3 loci as well as B2M with antigen presenting cells carrying non-HLA-A02:01 peptide MHC (pMHC) targets.

Generation of peptide MHC presenting cells

To generate the antigen presenting cells, mouse DC2.4 cell line (Merck-Millipore) was engineered to present the peptide MHC of interest by eliminating the mouse MHC presentation by knocking out the mouse B2m gene and overexpressing single chain human B2M HLA and peptide of interest.

DC2.4 cells were maintained in RPMI 1640 medium (Gibco) supplemented with 10% foetal calf serum (FCS; ATCC), IX L-Glutamine, IX non-essential amino acids (NEAA), IX HEPES buffer and 55 mM b-mercaptoethanol (all Gibco) ('culture medium').

Knock-out of mouse |32-microglobulin (mb2m) was achieved using CRISPR/Cas9 by cotransfection of cells with two plasmids encoding Cas9, guide RNAs (gRNAs) targeting exon 1 or exon 3 of mb2m and a GFP reporter. Cells were seeded onto 10 cm cell culture dishes to achieve 70-90% confluency for transfection, which was performed using Lipofectamine LTX & PLUS reagent (Thermo Fisher) according to the manufacturer's instructions and the scaling volumes for 10 cm dishes. Briefly, equal amounts of DNA for both mb2m gRNA plasmids were diluted in 0.5 mL OptiMEM (Gibco) together with the PLUS Reagent and the diluted DNA added dropwise to 500 uL of OptiMEM containing the diluted Lipofectamine LTX Reagent. The DNA-lipid transfection mixture was then incubated for 5 min at RT. During the incubation, the medium was aspirated from the cells and replaced with 10 mL of fresh culture medium. The DNA-lipid mixture was added dropwise to the cells on the dish and returned to the incubator at 37°C/5%CO2.

After 24 hr, GFP-positive cells were sorted on the Influx cell sorter (BD, Babraham Institute) based on transient GFP expression and maintained in culture for ~1 week before analysis of mb2m expression by flow cytometry using an anti-mb2m antibody (clone: S19.8, Biolegend or BD Biosciences) to assess the efficiency of knock-out. Staining of cells for flow cytometry was performed using ~ 1 x 10⁶ cells per sample on 96-well deep-well plates in 100 uL/well. The cells were first washed in PBS and then stained using LIVE/DEAD Fixable Near-IR Dead Cell Stain (Thermo Fisher) diluted 1:1000 in PBS for 30 min at 4°C. After ~30 min, cells were washed once in FACS Buffer (1% BSA/PBS + 0.01% sodium azide) and incubated with 20 ug/mL (2X) of mouse Trustain FcX reagent (Biolegend) for 15 min at 4°C to block the Fc receptors, before staining with 10 mg/10⁶ cells of anti-m|32m antibody (or mlgG2b isotype control antibody; Biolegend) for 30 min at 4°C in a final volume of 100 uL/sample. The cells were then washed three times with FACS buffer before fixing in 4% PFA/PBS for 15-20 min at room temperature. Cells were resuspended in 200 uL/sample of PBS for acquisition on CytoFLEX flow cytometer (Beckman Coulter) and the data was analysed using FlowJo vlO (BD Bioscience). To derive pure populations of cells deficient of mouse |32m ('DC2.4 m|32m KO' cells), the cells were subsequently re-sorted by staining for m|32m expression and gating on the m|32m-negative population. The confirmation of complete loss of m|32m by surface staining is shown in Figure 2A.

The DC2.4 m|32m KO cells were then engineered to express HLA-A*24:02 and a 9mer peptide target. This was achieved by co -transfection using the protocol above with vectors encoding human |32-microglobulin (h|32m) N terminal fused to human HLA-A*24:02 via a (G4S)3 linker as a single chain and a hygromycin resistance marker and a second vector encoding a minigene consisting of a gene encoding a fusion protein of ubiquitin linked via the C-terminal diglycine motif to the target peptide with a puromycin resistance marker, or empty vector (EV) as a control. The expression cassettes were flanked by DNA transposon piggyBAC inverted terminal repeats (ITRs). The two plasmids were transfected together with an additional plasmid encoding a hyperactive mutant of piggyBac transposase at a ratio of 20:1. Transfected cells were selected in 250 ug/mL hygromycin B (Roche) and 3 ug/mL puromycin (Merck Millipore) for a further ~7 days. Surface expression of HLA-A*24:02 and h|32m on the HLA-A*24:02- h|32m_EV or HLA-A*24:02-h|32m_target peptide cell lines was validated by flow cytometry using antibodies specific for HLA-A*24:02 (MBL, Clone: 17A10) and h|32m (clone A17082A, Biolegend).

Mouse DC2.4 cells are immortalised dendritic cells which possess key ligands such as CD80 and CD86 which can enhance peptide MHC presentation to T cells. In intro stimulation of engineered DC2.4 cells using LPS and I FNy induces both CD80 and CD86 surface level, Figure 2B. Therefore, using engineered DC2.4 as antigen presenting cells for allo and xeno priming is potentially advantageous.

Immunisation

The HLA-A*24:02-h|32m_EV or HLA-A*24:02-h|32m_target peptide stable expressing DC2.4 cell lines were detached from tissue culture flasks using TrypLE Express (Gibco), which was neutralised with a 5x volume of culture medium. Cells were pelleted by centrifugation at 300 xg for 5 min, resuspended in culture medium and counted before washing twice with PBS. Cells were resuspended to 25 x 10⁶ cells/mL or 12.5 x 10⁶ cells/mL in PBS and mixed 1:1 with Sigma Adjuvant System (Sigma-Aldrich, S6322) for dosing with 200 uL per mouse for the prime and boost immunisations, respectively. 2.5xl0⁶ cells and 1.25xl0⁶ cells were used for prime and boost respectively. Cell based immunogen were inoculated into mice via intra peritoneal (i.p.) route. Prime and one boost schedule was used for these studies with 21 day gap between prime and boost. Tissues were collected at day 9 and day 16 post prime, and day 7 post boost.

Tissue preparation, CD8+ T cell enrichment and staining for antigen specific T cell sorting

To identify antigen responding CD8+ T cells, mesenteric lymph nodes, inguinal lymph nodes and whole spleen were collected from each mouse. Briefly lymph nodes and spleen were diced and filtered through a 40pM cell strainer to generate cell suspension in tissue collection medium, which consists of RPMI 1640 medium (Gibco) supplemented with 10% foetal calf serum (FCS; ATCC), IX L-Glutamine (Gibco) and IX HEPES buffer (Gibco). CD8+ T cells were further enriched by using EasySep Mouse CD8+ T cell isolation kit (STEMCELL, #19853) together with isolation buffer which consists of PBS with 2% FCS (ATCC), 20mM HEPES (Gibco) and ImM EDTA (Gibco). Typically this protocol achieves 80% purity of CD8+ T cells. Cells could be either frozen down or proceed directly to labelling and cell sorting.

For staining, either fresh or thawed CD8+ enriched T cells were placed on ice. Up to 2.5 million CD8+ T cell enriched cell suspension were stained with 0.05-0.1pg peptide MHC specific and non-specific tetramers in lOOul on ice for 15 min. Peptide MHC specific and nonspecific tetramers were custom made by MBL and conjugated with PE and APC fluorophores respectively. APC in this context is an abbreviation for the fluorophore allophycocyanin. The HLA used in tetramer probes are CD8 binding deficient mutants. Such mutant HLA tetramer can facilitate the identification of higher affinity TCRs (Bodinier M et al. Nat Med. (2000) 6:707-710.)

Tetramer staining was followed by the addition of a staining cocktail containing anti-human CD8b antibody-FITC (Miltenyi, 130-11-567), mTCRbeta-BV421 (Biolegend, 109230), and eFluor 789 Fixable Viability Dye (ThermoFisher, 65-0865-18). Cell surface protein (CSP) markers (antibodies) and tetramers can be oligo barcoded for quantification of those CSP in NGS. Reagents were conjugated and quality controlled with oligo by following the manufacture's protocol (Abeam, ab270703/05/09).

TotalSeq-C series-hashtag antibodies (Biolegend, 155861, 155863, or any additional unique hastag antibodies depending on sample numbers) were also added into individual mouse sample for hash tagging purpose. Cells were incubated for further 30mins on ice before washing with staining buffer (PBS with 2% FCS and 20mM HEPES) twice prior to sorting.

Stained cells were resuspended in 300-400 pl of Staining buffer and filtered through a 40pM cell strainer and kept on ice prior to acquisition using Fusion FACSAria™ cell sorter (BD Bioscience). Cells were gated using the following gating strategy: Lymphocyte>Single cell>Live> hCD8b+> Target tetramer positive & negative control tetramer negative. Cells were sorted into a 1.5ml microtube with ~300pl of ice cold staining buffer.

As shown in Figure 3, a significant proportion of CD8+ T cells showed antigen specific tetramer staining after immunisation with engineered mb2m KO DC2.4 cells overexpressing target peptide HLA (Q3 in the FACs plots) compared to single chain HLA overexpression alone (QI in the FACs plots) both after priming and boost 1, in contrast to non-specific tetramer staining pattern where there is no difference. Freshly collected target cells were immediately taken for mRNA extraction for single Cell RNA-Seq. 10X Chromium Single cell platform from 10X genomics was applied. Alternative platforms are also applicable, such as BD Rhapsody™ si ngle-Cell Analysis System, Parse Evercode™ TCR single cell technology from PARSE Biosciences.

For 10X Chromium mRNA single cell technology, briefly, 3,000-10,000 sorted cells for each lane of a 10X ChipK Chip was used to generate single cell encapsulation. 1^st strand cDNA was sythesized by reverse transcription following 10X Chromium Next GEM Single Cell 5' v2 (Dual Index) with Feature Barcode technology Rev F protocol. cDNAs for both VDJs and CSP portions were cleaned-up and amplified followed by quality control using 2100 Bio-analyser system (Agilent). After QC, cDNA portion was used for the generation of libraries. Mouse TCR VDJ library, Gene Expression (GEX) library and CSP library were generated according to lOx Genomics's protocol. Final libraries were quality controlled before sequenced using Nova-Seq 6000 platform carried out by NovoGene.

Mapping raw sequences

The raw sequencing reads were obtained in FASTQ format, where the structure of the reads conform to that as specified by 10X Genomics. The reads were mapped to a set of custom- made reference sequences using 10X CellRanger running in multi-modality mode. VDJ and GEX reference sequences were made based on the human TCR sequences engineered in the transgenic mice and CSP reference sequences based on the barcoding kit used (BioLegend TotalSeq C and Abeam lightning). The obtained mapped data included, for example: for VDJ library, single cell TCR chain/contig nucleotide sequences with annotation of V(D)J genes, position of framework regions (FWR) and complementary determining regions (CDR), and raw quantification of chain abundance as in the number of unique molecular identifiers (UMI); for GEX library, calling of cell-containing partitions (droplets) and single-cell raw quantification of gene expression (UMI); for CSP library, single-cell raw quantification of antigen-probe-/antibody-conjugated barcodes (UMI).

QC, cell type annotation and expression phenotype scoring

Cells were assigned to each mouse donors by demultiplexing using hashtag as part of the CSP library. Cell QC were performed to remove low quality cells, potential doublets defined by gene expression or hashtag profile, and cells without a productive alpha chain or beta chain or with more than one beta chains.

For early datasets (samples), the gene expression data were normalised and clustered in a K- nearest-neighbour (KNN) graph (see next section) using standard single-cell analysis workflow and clusters were annotated as broad cell types using well-known marker genes. T cells were further re-clustered and annotated into finer cell types/states again using marker genes well-known in the field. Separate multinomial logistic regression classifiers were trained on the broad cell types and higher-resolution T cell types/states. These classifiers were used to call broad as well as high-resolution cell types/states from the normalised gene expression data for the later datasets (samples). Expression phenotype scoring was generated for each cell by comparing the abundance of a set of signature genes associated with certain phenotype against a set of randomly selected genes with matching abundance distribution in the total sample.

Quantifying antigen specificity / affinity

The raw quantification of binding to target antigen tetramer probe, negative control tetramer probe and the abundance of cell surface TCR were normalised by either the total abundance of CSP barcodes, or the total abundance of both CSP barcodes and GEX molecules, or a method that normalises the background level of non-specific barcodes. Antigen specificity of a given cell was calculated as the ratio of normalised abundance between the bound targetantigen probe and the bound negative nonspecific probe, and the antigen affinity of a cell was calculated as the ratio of normalised abundance between the bound target-antigen probe and the cell surface TCR.

Calling clonotypes

TCR clonotypes were defined as sharing the same V and J gene and the same CDR3 amino acid sequence for both TCR alpha and TCR beta chain. Exact sub-clonotypes were defined as sharing the same V and J gene and the same CDR3 nucleotide sequence for both TCR alpha and TCR beta chain. For the purpose of clonotype selection, we used amino-acid-sequence- based clonotype definition as the functions and properties of a TCR.

Quantifying and visualising clonotype sequence similarity

Clonotype sequence similarity were calculated on the basis of CDR amino acid sequences. We used a method that performed pairwise sequence alignment and conservation-based distance scoring. The distance between a pair of TCR clonotypes was calculated as the weighted sum of alignment distances score between CDRls, CDR2s and CDR3s of the same chain, where CDR3s were given higher weight than CDRls and CDR2s. Pairs with larger distance were considered less similar, therefore, the inverse of such distance (or adjacency) was used as a measure of similarity. Treating each clonotype as a node and the similarity / adjacency as the weight of the edge linking the nodes, the relationship between the clonotypes was represented as a weighted undirected graph. Using a force-directed graph layout, clonotypes with high sequence similarity were placed close to each other forming tight connected components whereas clonotypes unrelated to each other were placed far apart.

TCR clonal and cellular phenotype characteristics, candidate TCR selection and expression in bivalent Fc fusion format

Immunisation of two animals per each time point resulted in diverse clonal types: 139 for day 9 priming, 175 for day 16 priming and 140 for boost 1. At all three time points, singletons, multiplets and highly expanded clones were detected. The cellular phenotypes of obtained clones were largely non-naive cellular phenotype defined by gene expression: activated, proliferating, IFNy-stimulated, and exhaust like phenotypes. This suggests that allopriming induces immune-reactive CD8 T cell responses which can be detected at cellular level (Figure 4B). When antigen specific and non-specific tetramer abundance were plotted for each clonal type (Figure 4A), we noticed that majority of the clones deviate from the diagonal with more probe abundance for antigen specific tetramer than non-specific control, which mirrored the actual sort data shown in Figure 3A. This confirms the successful recovery of highly specific antigen reactive T cell clones. There was a trend of smaller clonal size with higher tetramer association (Figure 4A).

Based on these clonal type and phenotypic observations, TCR candidates were selected based on two criteria. Firstly, antigen specificity score over 1.4 fold, i.e. fold change of antigen specific tetramer probe over negative control tetramer probe, was used as a threshold to select TCR candidate with specific antigen binding. Secondly, T cells with activated, proliferating, IFNy-stimulated, and exhaust like phenotypes were included as these cellular phenotypes are associated with immune-reactive T cell phenotypes. In total, 25 clonal types were taken forward for expression.

The TCRa and TCR|3 variable and constant sequences were synthesized by Twist Bioscience as separate fragments. TCRa variable and constant sequence was cloned into an expression construct where it was fused to human IgGl Fc sequence at the C-terminal, whereas TCR|3 variable and constant sequence was cloned into a separate expression vector. Dual expression of both TCRa and TCR|3 constructs result in soluble bivalent TCR-Fc fusion molecule, which was used to examine the binding capacity to antigen of interest.

To produce TCR-Fc soluble proteins, Expi293™ transient expression system (Thermofisher) was used according to manufacturer's instruction. Six days after transfection, supernatants were harvested and purified using MabSelect Sure LX resin (GE Healthcare). Proteins were eluted in IgG Elution buffer (Pierce) at pH 2.8 and pH was neutralised with IM Tris pH 8.0. Protein purity was determined by SDS-PAGE.

Binding confirmation and affinity measurement of soluble TCR-Fc molecules

An Enzyme -linked immunosorbent assay (ELISA) was performed to assess binding ability of soluble TCR-Fc molecules. Single chain target peptide, B2M and HLA-A*24:02 recombinant protein was used as the substrate. The single chain molecule was constructed by using endogenous leader sequence of HLA-A*24:02 directly linked to peptide target and followed by N-terminally fusion to B2M with G3AS(G4S)2 linker. The extracellular domain of HLA- A*24:02 was C-terminally linked to the above construct via a G4S linker and a poly histidine tag. Such single chain peptide, B2M and HLA recombinant protein was expressed using Expi293™ transient expression system as described before and purified using single step immobilized metal affinity chromatography (IMAC) chromatography.

ELISA assay was first used to confirm peptide MHC binders first. Briefly, polystyrene 96-well plates flat bottom (Greiner Bio One , 3361) were coated for 2hr at 37°C with 2ug/ml of recombinant single chain target peptide, B2M and HLA-A*24:02 recombinant protein. After incubation, plates were washed four times with wash buffer (PBS containing 0.05% Tween-20 and 0.1% BSA) and blocked with blocking buffer (PBS containing 2% BSA) for 1 h at 37°C.

Wells were incubated with 100 pL of purified soluble TCR-Fc at six serial three-fold dilutions starting from 40nM concentration, in triplicate, for 2 hrs at 37°C. Anti-human P2M antibody was used as positive control. After washing four times with wash buffer, wells were incubated with HRP-conjugated anti-human Fc antibody (Abeam, 1:2000 dilution) in blocking buffer for 1 h at 37°C. Wells were washed four times again before incubating with 100 pL TMB substrate (ThermoFisher, N301). The TMB reaction was quenched after 5 min using 1 M sulfuric acid. The OD at 450 nm was measured on a EnVision PerkinElmer Plate Reader. Results for the binding was shown in Figure 5A, where five candidates were confirmed to be binders.

Surface Plasmon Resonance (SPR) was used to determine the binding affinity (KD) to single chain target peptide, B2M and HLA-A*24:02 recombinant protein using Biacore8K+ (Cytiva). A sensor Chip NTA was used to capture purified histidine-tagged single chain target peptide, B2M and HLA-A*24 recombinant protein here referred as ligand. Ligand was diluted with HBS-P IX running buffer (diluted from 10X HBS-P+ Buffer, pH 7.4 (Cytiva) and captured onto a Sensor Chip NTA approximately 50ug/ml. The ligands were injected for 60 seconds at 10 pl/min in all the active channels of all 8 flow channels. The run was performed at 25° C. using neutral pH HBS-P lx+ as running buffer. Protein A -purified soluble TCR-Fc molecules were diluted in the running buffer at lOug/ml and used as analyte. The analyte was injected in multiple cycle kinetics (MCK) mode at 4 concentrations (0.04nM, , 4nM, 20nM, and 66nM) with 180 seconds association phase and 600 seconds dissociation phase, at flow rate 30 pl/sec in both active and reference channels. The method can alternatively be performed with 5 concentrations, at 0.04 nM, 0.19 nM, 4 nM, 20 nM, and 66 nM respectively. Three injections of 10 mM Glycine pH 1.5 for 60 sec. at 10 pl/min were used for the regeneration phase. The values for association rate constant (k_on), dissociation rate constant (k_Off) and dissociation constant (KD) were calculated from the binding data by BIAevaluation software (Cytiva). Data were reference and buffer subtracted and fitted into one step biomolecular reaction (Langmuir 1:1) model. Figure 5B shows SPR sensorgrams of eight soluble TCR-Fc molecules showing binding to single chain target peptide, B2M and HLA-A*24:02 recombinant protein. Interaction for all eight binders was observed upon captured of TCR-Fc molecules onto a Protein G chip as ligand and injection of chain target peptide, B2M and HLA-A*24:02 recombinant protein as analyte. In this SPR orientation, since TCR-Fc fusion molecule is bivalent, the affinity determined could be a combined effect of avidity and affinity. Therefore, we conducted affinity measurement further in the reverse orientation where TCR-Fc was the ligand on the chip and pMHC was the analyte to eliminate the avidity effect.

Briefly, the SPR measurement in reverse orientation was conducted using Biacore8K+ (Cytiva). Protein A -purified soluble TCR-Fc molecules (referred as ligands) were diluted with HBS-P IX running buffer (diluted from 10X HBS-P+ Buffer, pH 7.4 (Cytiva) and ware captured onto the Protein G chip (Cytiva) at approximately 5ug/ml as the ligands. TCR-Fc was injected for 60 seconds at 10 pl/min in all the active channels of all 8 flow channels. The run was performed at 25°C. using neutral pH HBS-P lx+ as running buffer. Recombinant human peptide B2M-HLA-A*24:02 single chain molecule was diluted in the running buffer at 20ug/ml and used as the analyte. A concentration of 50 ug/ml can also be used. It was injected in multiple cycle kinetics (MCK) mode at 4 concentrations (0.19 nM, 2.4nM 12nM, 60nM and 300nM) with 120 seconds association phase and 200 seconds dissociation phase, at flow rate 30 pl/sec in both active and reference channels. A series of 5 concentrations can also be used: 0.19 nM, 4 nM, 20 nM, 66 nM and 300 nM. Three injections of 10 mM Glycine pH 1.5 for 60 sec. at 10 pl/min were used for the regeneration phase. The values for association rate constant (kon), dissociation rate constant (koff) and dissociation constant (KD) were calculated from the binding data by BIAevaluation software (Cytiva). Data were reference and buffer subtracted and fitted into one step biomolecular reaction (Langmuir 1:1) model. Figure 5C shows example SPR sensorgrams of soluble single chain target peptide- B2M-HLA-A*24:02 recombinant protein analyte binding to immobilised TCR-Fc ligand. Figure 5D summarises the affinity (KD) of the TCRs generated by allopriming, as determined by SPR. These affinity data are representative of monomeric (1:1) binding affinity between the TCR and its MHC-peptide target complex.

Example 2 - xenogeneic MHC in vivo priming to generate high affinity TCRs

The allopriming concept demonstrated in Example 1 is extended to xenopriming, whereby hTCR transgenic mice containing mouse MHCs are immunised with mouse B2m knockout DC2.4 cells overexpressing single cell human B2M-HLA class I and target peptide. Materials and methods described in Example 1 can be used for xenopriming, with the exception that the immunised mice express endogenous mouse MHC and do not express human MHC.

For discovery of target specific TCRs using a xenopriming method, transgenic mice contained humanised CD8a/b, TCRa and TCR|3 loci (see WO2025/032222). MHC and |32m loci in the mice were wild type and no MHC transgenes were present.

Engineered DC2.4 cells carrying the target peptide of interest and HLA of interest, HLA- A*02:01 (Experimental cohorts) and matched control DC2.4 cells carrying the HLA-A*02:01 but without the target peptide (Control cohort) were prepared separately. For immunogen preparation, lOOul PBS suspension of DC2.4 cells mixed with lOOul of Sigma Adjuvant System (SAS, S6322-1VL, Sigma) was prepared per mouse. 2.5xl0⁶ cells and 1.25xl0⁶ cells were used per injection for prime and boosts respectively. Cell based immunogen was introduced into mice via IP route. A schedule of prime and two boosts was used with a 21 day gap between prime and boost. Tissues were collected at day 9 post-prime, and day 7 post-boost 1 and post-boost 2. Tissue preparation, recovery of antigen-specific TCRs and TCR sequencing were essentially as described in Example 1. Antigen specific cell sorting was performed using a PE conjugated-pMHC tetramer of the target peptide pMHC (HLA-A*02 complex) and an APC conjugated pMHC tetramer of HLA-A*02 and a sequence-irrelevant peptide. Figure 7 shows FACS analysis of xenogenic response from a responding mouse in the experimental cohort (right panel) compared to a mouse from the control cohort (left panel). A strong antigen specific response was detected in the responding mouse, 0.96% of total CD8 cells were PE positive and APC negative (target tetramer +ve), contrasting with the detection of only 0.034% PE positive and APC negative CD8 T cells in the control mouse.

Affinity of a selected TCR derived from the xenogeneic responder mouse was determined by SPR. Protein A-purified soluble TCR-Fc molecules were diluted with HBS-P IX running buffer (pH 7.4 (Cytiva)) and captured on a Protein G chip (Cytiva) at approximately lOug/ml as the ligand. Bivalent soluble TCR-Fc molecules were injected for 60 seconds at 10 pl/min in all the active channels of all 8 flow channels. The run was performed at 25°C using neutral pH HBS-P lx+ as running buffer. Recombinant human target peptide B2M-HLA-A*02:01 single chain molecule was diluted in the running buffer at lOOug/ml and used as the analyte. It was injected in multiple cycle kinetics (MCK) mode at concentrations of 2uM, 623nM, 208nM and 70nM with 120s association phase and 200s dissociation phase, at flow rate 30 pl/sec in both active and reference channels. Three injections of 10 mM glycine pH 1.5 for 60s at 10 pl/min were used for the regeneration phase. The values for association rate constant (kon), dissociation rate constant (koff) and dissociation constant (KD) were calculated from the binding data by BIAevaluation software (Cytiva). Data were reference and buffer subtracted and fitted into one step biomolecular reaction (Langmuir 1:1) model. Figure 8 shows the affinity (KD) of a TCR derived from xenopriming, measured as 0.3 pM, with the SPR sensorgram and on-rate and off-rate kinetic information specified.

In conclusion, the data indicate that a xenopriming protocol, where a transgenic mouse containing wild type mouse MHC and humanised CD8a/b, TCRa and TCR|3 loci is immunised with engineered APC carrying epitope of interest with HLA-A*02:01 allele, can trigger a strong CD8 T cell antigen specific response. This provides a novel way of discovering TCR candidates with therapeutic potential. Inclusion of human CD8 in the mice may enhance the T cell response by engaging the HLA presented on the antigen presenting cell.

Comparative Example 3 -syngeneic HLA priming

As an example of syngeneic HLA priming, we describe immunisation of HLA-A*02:01 transgenic mice with a peptide target presented by HLA-A*02:01. The transgenic mice in this Example also contain humanised HLA-A*02:01, CD8a/b, TCRa and TCR|3, and B2M loci with mouse endogenous MHC and B2M deleted.

Peptide immunisation was conducted by using 100 pg of peptide and 50 pg of CpG ODN in 50 pl PBS emulsified with 50 pl of IFA for priming and 50 pg of peptide prepared in the same manner for the adjuvant was used for boosts. Peptide immunogen was administrated via intramuscular route. Prime and two to three boosts schedule are typically used for these studies and with 21 days gap between prime and boost and 14 days gap between each boost. Tissues are typically collected day 7 post each boost. T cells were sorted using antigen specific and non-specific tetramer probes as described in Example 1 at either prime day 9, 16 or boost 1, 2 or 3. TCR analysis was conducted similar to Example 1.

Material and methods for affinity determination for the binders were similar to what was described in Example 1 in the orientation where TCR-Fc molecules was captured. 50pg/ml TCR-Fc was captured onto protein G chip and the concentration range for the recombinant single chain target peptide MHC used was between 5-500 pg/ml. The affinity of TCRs derived from such syngeneic priming was weak with approximately half of binders being close to or beyond the detection limit of SPR, which is around 100 pM and is shown as 100 pM in Figure 6. The median affinity of the other half of the "stronger" binders is 9 pM and the strongest binder had an affinity of 2.6 pM. To confirm that allo and/or xeno priming can generate higher affinity TCRs compared to syngeneic priming, non-HLA-A*02:01 transgeneic mice, such as HLA-A*03:01, HLA-A*ll:01 and HLA-A*24:02 transgenic mice can be used to elicit allopriming and mouse MHC containing transgenic mice can be used for xenopriming. The cell antigen are generated by co-transfecting single chain B2M-HLA-A*02:01 with target peptide minigene construct. Material and methods for cell based immunisation, cell sorting and computational analysis are as described in Example 1. Candidate TCRs are selected and affinity is determined by SPR as described in Example 1.

Example 4 - TAP and HM13 knock out improving antigen cell presentation for immunisation

In cellular phenotype data described in Example 1, immune reactive T cell phenotypes were detected in significant quantity in antigen specific tetramer negative CD8+ T cells. This may result from T cell activation by alternative antigens to the intended target. In the experiment set up, although the mouse MHC system was not present on the cell surface due to mouse B2m deletion, peptides of mouse origin can be processed and presented on the exogenously introduced human B2M HLA single chain constructs. Surface presentation of such peptide MHC will elicit immune response in allo priming context and therefore possibly reduce the diversity and the amplitude of T cell responses against the intended target.

To improve on the specificity of the design for antigen presentation, TAP1/2 knockout can be engineered in m|32m deficient DC2.4 cells. TAP1/2 deficiency results in lack of ER transport of proteosome processed peptides, which leads to significant reduction in endogenous peptide presentation. In order to present target peptide on the surface in TAP1/2 KO cells, a minigene construct involving signal peptide, such as mouse mammary tumour virus (MMTV) signal peptide, fusion to peptide target will drive the TAP1/2 independent transport into ER. Subsequent signal peptide cleavage in the ER by signal peptide peptidases result in target peptide assembly with MHC and surface presentation (Bruno P et al. Nat Biotech 2023 Jul;41(7):980-992.). In addition, signal peptide peptidase (SPP, encoded by HM13 gene) loss is important for enhancing specifically HLA-A*02:01 peptide presentation specificity via the signal peptide driven target peptide presentation, as hydrophobic signal peptide byproduct could be processed by SPP in ER and presented by HLA-A*02:01 due to its preferences for aliphatic amino acid residues that are commonly found in hydrophobic region of signal peptides (Bruno P et al. Nature Biotechnology 2023 Jul;41(7):980-992.).

Knock-out of mouse Tap 1, Tap 2 and HM13 are achieved by using CRISPR/Cas9 by cotransfection of DC2.4 cells with a pair of plasmids encoding Cas9, guide RNAs (gRNAs) targeting each of the three genes. Since CRISPR/Cas9 containing plasmids also contain GFP reporter, GFP positive cells are sorted 24-48 hours post transfection to enrich for transfected cells. Since TAP1/2 and HM13 KO result in surface reduction of peptide HLA, anti-mouse b2m specific antibody staining is used to sort human HLA low cells clonally for expansion. Subsequent genotyping are used to confirm TAP1/2 and HM13 deficiency. Once successful clones have been identified, a subsequent round of CRISPR/Cas9 transfection is conducted to knockout mb2M with materials and methods described in Example 1. Genotyping is conducted to validate the loss of mb2m at genomic DNA level.

For target peptide expression in Tapl/2 and HM13 KO cells, target peptide was C-terminal seamlessly fused to MMTV signal peptide or mammalian gene derived signal peptides and the construct was driven by CAG promoter for expression. This expression construct together with a puromycin drug resistant cassette was flanked by DNA transposon piggyBac terminal inverted repeats to allow piggyBac mediated stable integration when co-transfected with a piggyBac transposase expressing construct.

In further work, TAPI and TAP2 were knocked out in DC2.4 cells with a pre-existing knockout of the mouse B2m (see Example 1) using CRISPR/Cas9. Plasmids encoding Cas9 and guide RNAs (gRNAs) targeting each of the two genes were cotransfected into the cells. Since CRISPR/Cas9 containing plasmids also contain GFP reporter, GFP positive cells were sorted at 24 hours post-transfection as single cells in 96 well plates containing culture media, to identify clonal KO cell lines. After 1 week's recovery, recovered clones were screened by Western blot analysis for the loss of mouse TAPI and TAP2 protein. lxlO⁶ cells were washed in cold lx PBS and lysed in 150 pL of 4% SDS lysis buffer (4% SDS, 150mM NaCI and 50mM Tris, pH 7.5). The lysate was then sonicated (5-10 seconds, 2.5-3 amplitude), and incubated at 95°C for 5 minutes. Absolute protein concentration was measured by NanoDrop Eight (ThermoFisher). Cell lysate was normalised based on protein amount and were then mixed 1:1 with 2x Laemmli buffer (4x LDS Sample buffer with lOx Sample Reducing Agent, NuPAGE, Invitrogen), and incubated at 95°C for another 5 minutes. Samples were run on 4-12% BisTris gel (NuPAGE, Invitrogen) in lx MES SDS running buffer (Bolt, Invitrogen) alongside a prestained protein ladder (CSL-BBL, Cleaver), and transferred using the iBlot3 system and MIDI NC transfer stacks (Invitrogen). Membrane was blocked in PBS with added 5% BSA (Sigma) for lh at RT before antibodies were probed O/N at 4°C in PBS with added 1% BSA and 0.05% sodium azide at 1:1000 dilution (TAPI - Invitrogen #MA5-55564, TAP2 - Invitrogen #PA5- 37414, GAPDH - CST #2118). HRP secondary antibody (CST #7074) was prepared at 1:3000 dilution in PBS with added 1% skim milk powder (Oxoid), and incubated with the membrane for lh at RT. After bloting, membrane was incubated in ECL Prime solution (Amersham) for 5 minutes and imaged on iBright 1500 imaging system (Invitrogen), using a wide range of exposures. Several clones showed a depletion of TAP1/2 protein, as shown in Figure 10. Subsequent genotyping was conducted to confirm TAP1/2 deficiency and clone CI-2 was selected for the subsequent KO of HM13 using CRISPR/Cas9 in the same way as the TAP KO.

An alternative cell line for use as antigen presenting cell is the human K562 cell line, which can be engineered to express an MHC transgene of interest (Cereb & Yang J Immunol 156(l):18-26 1996; Briten et al., J Immunol Methods 259(l-2):95-110 2002). For allopriming experiments, K562 cells were engineered to express monoallelic HLA-A*03:01, and to knock out TAPI and TAP2 to reduce endogenous peptide presentation. Example 5 - Allopriming HLA-A*24-peptide complex in HLA-A*02 hTCR transgenic mice

To demonstrate further utility of the allopriming approach, a further study was conducted, immunising HLA-A*02:01 transgenic mice with DC2.4 cells presenting peptide on HLA- A*24:02. Mice contained humanised HLA-A*02:01, |32m, CD8a/b, TCRa and TCR|3 loci and knock out of endogenous mouse MHC class I alleles, as described in WO2025/032222.

Engineered DC2.4 cells carrying the target peptide of interest and B2M-HLA-A*24:02 single chain transmembrane construct (Experimental cohorts) and matched control DC2.4 cells carrying the single chain B2M-HLA-A*24:02 without the target peptide (Control cohort) were prepared. Per mouse, lOOul of suspension was mixed with lOOul of Sigma Adjuvant System (SAS, S6322-1VL, Sigma) for a total 200ul of immunogen preparation. 2.5xl0⁶ cells and 1.25xl0⁶ cells were used per injection for prime and boosts respectively. Cell based immunogen was inoculated into mice via IP route. Prime and one boost schedule was used, with a 21 day gap between prime and boost. Tissues were collected at day 9 and day 16 postprime, and day 7 post-boost. Tissue preparation, recovery of antigen-specific TCRs and TCR sequencing were essentially as described in Example 1.

Affinity of selected TCRs was determined by SPR, essentially as described in Example 1 in the orientation with TCR-Fc molecules captured as ligand. The affinity summary of a panel of TCRs derived from alloreactive T cells is shown in Figure 11 left panel and representative SPR sensograms of selected high affinity TCRs are shown on the right. Many TCRs with affinities measured at less than 1 pM were readily obtained from this allopriming experiment (and see also the work reported in Example 1 and Figure 5c), in contrast to the syngeneic priming experiment described in Example 3, and Figure 6, where no TCR with affinity measured at less than 1 pM was obtained.

Example 6 - Allopriming HLA-A*03:01-peptide complex in HLA-A*02:01 hTCR transgenic mice

To demonstrate the application of neoantigen pMHC TCR discovery, a further allopriming study was conducted, immunising HLA-A*02:01 transgenic mice with DC2.4 cells presenting peptide on HLA-A*03:01. The target peptide was a human cancer neoantigen, containing a single amino acid substitution compared with the wild type peptide. Mice contained humanised HLA*A02:01, CD8a/b, TCRa and TCR|3 loci as well as humanised |32m. Endogenous mouse MHC class I alleles were knocked out.

Engineered DC2.4 cells carrying the target peptide of interest and HLA of interest, HLA- A*03:01 (Experimental cohorts) and matched control DC2.4 cells carrying the HLA-A*03:01 but without the target peptide (Control cohort) were prepared as immunogens as two separate cohorts. Cell based immunisation protocol is the same as described in Example 5. MHC-peptide complex tetramers containing either the neoantigen peptide or the wild type peptide were used in cell sorting with the aim to isolate T cells specific to neoantigen pMHC. Mice from the experimental cohort yielded a neoantigen pMHC specific response, demonstrated by an increased proportion of neoantigen pMHC tetramer positive CD8 cells (FACS plots in Figure 12, right panels), compared with mice from the control cohort (FACS plots in Figure 12, left panels). Antigen-specific T cell responses were observed at both collection time points, day 9 post prime (top panels) and day 7 post boost 1 (bottom panels). Note that T cells which can recognise both wild type and neoantigen were observed (in addition to T cells that only recognised the neoantigen), as demonstrated by a PE and APC double positive population in mice from the experimental cohort (Figure 12 right panels).

A Jurkat reporter cell assay was used to confirm specificity of the selected TCRs. An engineered Jurkat (T-cell derived) cell line, containing a stably integrated GFP reporter gene under the control of the NFAT response element, fluoresces upon TCR-pMHC engagement as a measure of T cell activation. Endogenous TCR|3 was knocked out in these cells using a CRISPR/Cas9 system as described previously in order to ensure signalling based NFAT activation is solely derived from exogenously introduced TCR. TCRs were formatted with mouse constant and transmembrane regions in order to ensure appropriate pairing. Candidate TCRs were independently transfected into the Jurkat reporter cells and the transfectants were co-cultured with K562 cells expressing monoallelic HLA-A*03:01 functioning as antigen presenting cells. K562 HLA-A*03:01 cells were further engineered for TAP1/2 KO as described in Example 4 to reduce endogenous peptide presentation. Neoantigen or wild type peptide was pulsed at 10 ug/ml overnight. Flow cytometry analysis was carried out to measure GFP induction upon TCR activation.

No NFAT-GFP signal was observed on untransfected Jurkat cells (Figure 12A).

A GFP positive population (~42%) was detected in Jurkat cells expressing a candidate TCR, designated CNT1423, in the presence of K562 cells pulsed with the neoantigen peptide, indicating that TCR recognition of neoantigen pMHC triggered Jurkat T cell activation (Figure 13B, fourth panel). In contrast, no NFAT-GFP signal was detected in the presence of K562 cells pulsed with wild type peptide (Figure 13B, third panel). This data indicates that TCR CNT1423 is specific to neoantigen and cannot recognise wild type peptide pMHC. As an additional negative control, an HLA-A*02:01 restricted peptide was also pulsed, and no NFAT-GFP signal was observed, indicating the TCR was specific for HLA-A*03:01 Neoantigen pMHC (Figure 13B, second panel). An alternative TCR, CNT1424, which does not bind the neoantigen pMHC, was used as a negative control. No NFAT-GFP signal was observed (Figure 13C).

Collectively, the data strongly suggest that allopriming can facilitate the recovery of high affinity TCRs, and that the specificity of recovered TCRs can be exquisite, with recovered TCRs able to differentiate peptide sequences differing by a single amino acid. This provides a novel means for TCR discovery with potential therapeutic applications for targeting epitopes that have a sequence highly similar to non-target epitopes.

Claims

2 The method of claim 1 wherein the MHC component of the MHC-peptide complex is allogenic.

3 The method of claim 1 wherein the MHC component of the MHC-peptide complex is xenogenic.

4. The method of any preceding claim, wherein the laboratory animal genome comprises human CD8 or comprises chimaeric CD8 having a region of human sequence and a region of sequence from the laboratory animal, wherein the region that interacts with the MHC is a human region.

5. The method of claim 4, wherein the MHC component of the MHC-peptide in the antigen presenting cell is human HLA class I and the laboratory animal genome does not encode any human HLA.

6. The method of claim 5, wherein the laboratory animal genome comprises endogenous MHC class I.

7. The method of any of claims 1 to 4, wherein the MHC expressed in the laboratory animal genome is a human HLA.

8. The method of claim 7 wherein the laboratory animal genome encodes a human HLA- A*02:01, or HLA-A*ll:01, or HLA-A*03:01, or HLA-A*24:02 knock in.

9. The method of any preceding claim wherein the laboratory animal genome encodes a fully human TCR.

10. The method of any preceding claim wherein the laboratory animal genome encodes a chimaeric TCR, having a region of human sequence and a region of sequence from the laboratory animal.

11. The method of any preceding claim wherein the peptide is from a human antigen the expression of which is associated with disease, such as cancer, for example wherein the antigen is a tumour associated antigen.

12. The method of any preceding claim wherein the peptide is from a human antigen which has the same sequence as an antigen expressed in the laboratory animal, or is a variant of an antigen expressed in the laboratory animal that differs by only 1 or 2 or three amino acids.

13. The method of any preceding claim wherein the laboratory animal is a rodent, such as a mouse or rat.

14. The method of any preceding claim wherein the MHC expressed on the antigen presenting cell is expressed from a human MHC class I gene.

15. The method of any of claims 1 to 13 wherein the MHC expressed on the antigen presenting cell is a human-mouse chimaeric MHC.

16. The method of any one of claims 1-13 wherein the MHC expressed on the antigen presenting cell is expressed from a human MHC class II gene.

17. The method of claim 14 wherein the MHC in the MHC-peptide complex is a human HLA-A* 24:02.

18. The method of claim 17 wherein the MHC in the MHC-peptide complex is a human HLA-A*24:02 and the MHC encoded by the laboratory animal is HLA-A*02:01.

19. The method of any preceding claim, wherein the laboratory animal is mouse and the antigen presenting cell is a mouse antigen presenting cell.

20. The method of any preceding claim, wherein the antigen presenting cell is deficient in endogenous antigen processing /presentation, optionally being deficient in TAP and/or signal peptidase.

21. A method according to any of claims 1 to 18, wherein the APC is a synthetic antigen presenting cell, optionally where the APC is not a biological cell.

22. The method of any preceding claim wherein there is isolation and/or purification of T cells from the laboratory animal which are reactive with the MHC-peptide complex, and optionally wherein the T cells are a polyclonal mixture of T cells or a monoclonal T cell, or a cell lines thereof.

23. The method of any preceding claim comprising determining a nucleic acid sequence of the human TCR variable regions expressed by a T cell from the laboratory animal reactive to the MHC-peptide complex and i expressing the human T cell receptor, or human T cell receptor variable region or variable domain(s) in a cell; optionally wherein the receptor is soluble, optionally wherein the TCR receptor, receptor variable region or variable domain(s) is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; and optionally wherein the TCR sequence has been modified by 1 or more, e.g. 2 or more e.g. 3 or more amino acid changes from that identified in the laboratory animal, optionally further formulating the expressed human T cell receptor, or human T cell receptor variable region or variable domain(s) with a pharmaceutically acceptable excipient; or ii inserting the nucleic acid encoding the human T cell receptor, or human T cell receptor variable region, or variable domain(s) into a cell, ex vivo or in vitro, such as into a human or animal cell,; or optionally wherein the nucleic acid is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; optionally wherein the TCR nucleic acid sequence has been modified to result in 1 or more, e.g. 2 or more e.g. 3 or more amino acid changes from the sequence of the TCR identified in the laboratory animal; optionally wherein the expressed T cell receptor or part thereof is soluble; optionally wherein the cell containing the inserted nucleic acid is formulated for delivery to a human or animal iii formulating a nucleic acid (e.g. RNA or DNA) encoding the human T cell receptor, or human T cell receptor variable region, or variable domain(s), with a suitably delivery vehicle, such as a lipid or liposome, for delivery to a patient in need thereof in vivo. optionally wherein the nucleic acid of the TCR is expressed as a part of a larger multispecific molecule, such as a bispecific molecule; optionally wherein the TCR nucleic acid sequence has been modified to result in 1 or more, e.g. 2 or more or 3 or more amino acid changes from the sequence of the TCR encoded by the nucleic acid TCR identified in the laboratory animal optionally wherein the expressed T cell receptor or part thereof is soluble.

24. A method of treating an individual in need thereof, the method comprising

(i) delivery of a cell according to step (ii) of claim 23 into the patient in need thereof, or

25. A T cell receptor or T cell receptor variable region or domains(s) thereof obtained or obtainable from the method of any preceding claim, for use in a patient having a host MHC which is the same as that used to immunise the rodent and expressed by the APC.