TARGET PEPTIDE PRESENTED BY A MAJOR HISTOCOMPATIBILITY COMPLEX

CROSS REFERENCE TO RELATED APPLICATIONS / INCORPORATION BY REFERENCE

All documents cited or referenced herein ("herein cited documents"), and all documents cited or referenced in herein cited documents, together with any manufacturer's instructions, descriptions, product specifications, and product sheets for any products mentioned herein or in any document incorporated by reference herein, are hereby incorporated herein by reference, and may be employed in the practice of the invention. More specifically, all referenced documents are incorporated herein by reference to the same extent as if each individual document was specifically and individually indicated to be incorporated by reference.

FIELD OF THE INVENTION

The present invention relates to a method for predicting whether a binding peptide, which binds to a target peptide, preferably a target peptide that is presented in the context of major histocompatibility complex (MHC), and is for administration to a subject, will cross react with another peptide in the subject in vivo.

BACKGROUND OF THE INVENTION

MHC class I and class I I are immunoglobulin superfamily proteins specialised for antigen presentation, with a polymorphic peptide binding site which enables them to present a diverse array of short peptide fragments at the surface of the antigen presenting cell. Peptides presented by MHC are derived from proteins which have been proteosomally processed within the cell. In humans, MHC molecules are known as human leukocyte antigens (HLA).

A number of emerging immunotherapies rely on the administration to a subject to be treated of a binding peptide that binds a target peptide-MHC complex. The binding peptide may be an immune binding peptide such as, for example, an antibody or antigen binding fragment thereof or a T cell receptor or antigen binding fragment thereof. Such binding peptides bind to a binding sequence, comprising the amino acid sequence of the target MHC presented peptide. Often the binding sequence of the target peptide is known. There is a risk with such therapies that the binding peptide binds to peptides other than the target peptide (referred to herein as "off target peptides"), causing unwanted side effects. It is therefore desirable to identify whether such off target peptides exist. This allows binding peptides to be chosen and designed that do not bind to off target peptides and consequently have a far greater chance of not causing unwanted side effects. Reasons for unwanted side effects derived from off target specificities in adoptive T cell therapy are; mispairing of transduced TCR chains with endogenous chains, insertion mutagenesis associated with TCR transduction, or alloreactivity (Amos et al., Blood 2011, 118(3):499-509; Daniel-Meshulam et al., Front Immunol 2012, 3:186). To date, the way to prevent off target toxicity is to include apoptosis genes to destroy T cells if toxicity arises after administration to the patient (Restifo, et al., Nat Rev Immunol 2012, 12(4): 269-81.)

Thus a person skilled in the art would be motivated to provide mechanisms that deal with binding to off target peptides after administration of a binding peptide. It is therefore unlikely that a person skilled in the art would consider trying to identify off target peptides before the binding peptide is administered. In the highly unlikely event that the skilled person took this latter approach, one option would be to search protein sequence databases for peptides with similarity to the target MHC presented peptide. But, this often returns a large number of peptides, all of which would need to be tested and even then off-target peptides may not be identified. Lowering the stringency of the search parameters would further increase the number of potential epitopes that would have to be tested and again may still not reveal off target peptides. Alternatively, the skilled person may measure any immune response generated by the binding peptide in the presence of cells derived from normal tissue(s) (which preferably do not express the target peptide). However, this can be a difficult process, depending on the number and type of cells tested. Furthermore, primary cells cultured in vitro, may have a different protein expression profile compared to the same cell type in vivo. This may result in a false assessment of potential cross reactivity in vivo. Finally, the skilled person may use animal models to measure any immune response generated by the binding peptide. Because of the differences between human protein sequences and those of the animal, the absence of unwanted side effects in the animal may not translate to humans.

Thus, these approaches do not accurately indicate, especially when administered to a subject, whether the binding peptide will indeed give rise to unwanted side effects derived from off target specificities, especially in an individualised or personalised setting. I n short, any attempt heretofore to identify off target peptides before the binding peptide is administered had no reasonable expectation of success. I ndeed, as is explained in more detail below, the inventors have found that such approaches will not necessarily identify off target peptides that cause an unwanted side effect. Although these approaches were performed on the a3a T cells described in the examples below, off-target activation of a3a T cells only become apparent when the binding peptide was were administered to patients (manuscript in preparation).

Citation or identification of any document in this application is not an admission that such document is available as prior art to the present invention.

SUMMARY OF THE INVENTION

It is therefore desirable to provide an alternative method for predicting whether a binding peptide will or at least is likely to bind to an off target peptide, which addresses drawbacks of the prior art.

In an aspect, the present invention provides a method for predicting whether a binding peptide, which binds to a target peptide presented by a Major Histocompatibility Complex (MHC) and is for administration to a subject, has the potential to cross react with another peptide in the subject in vivo, the method comprising: identifying at least one binding motif in the target peptide to which the binding peptide binds; and

searching for peptides that are present in the subject that comprise the at least one binding motif and that are not the target peptide,

wherein the presence of one or more such peptides indicates that the binding peptide has the potential to cross react in vivo.

In another aspect, the invention comprehends a method of treating a human or animal mammalian subject in need thereof - advantageously comprising providing individualized or personalized treatment involving a binding peptide - comprising:

predicting whether the binding peptide, which binds to a target peptide presented by a Major Histocompatibility Complex (MHC) in the subject to thereby provide said treatment and is for administration to the subject, has the potential to cross react with another peptide in the subject in vivo, comprising:

identifying at least one binding motif in the target peptide to which the binding peptide binds; and

searching for peptides that are present in the subject that comprise the at least one binding motif and that are not the target peptide,

wherein the presence of one or more such peptides indicates that the binding peptide has the potential to cross react in vivo, and the absence of one or more of such peptides indicates that the binding peptide has the potential not to react in vivo.

The at least one binding motif may be identified by:

creating a series of mutants of the target peptide, each mutant having the amino acid residue at one position in the binding sequence thereof that is involved in binding to the binding peptide substituted for an alternative amino acid, such that over the series of mutants the amino acid residue in each position in the binding sequence is substituted for an alternative amino acid; and

testing each mutant in the series for its activity relative to the wild type target peptide, wherein an amino acid residue at a position within the binding sequence is identified as being part of the binding motif if the mutant in which the amino acid at that position is mutated to an alternative amino acid has a substantial loss of activity relative to the wild type target peptide.

Methods of the invention may further comprise, where an amino acid residue at a position in the binding sequence is not identified as being part of the binding motif, substituting this position with at least one additional amino acid and testing for activity relative to the wild type peptide,

wherein amino acid substitutions which result in a substantial loss of activity relative to the wild type target peptide are considered to be non-tolerated amino acids and not part of the binding motif and/or amino acid substitutions which do not result in a substantial loss of activity relative to the wild type target peptide are considered as part of the binding motif.

Methods of the invention may further comprise creating a series of mutants, each mutant having the amino acid residue at one position in the binding sequence substituted for an alternative amino acid, such that over the series of mutants the amino acid residue in each position in the binding sequence is substituted for all alternative amino acids, and testing each mutant in the series for activity relative to the wild type peptide,

wherein amino acid substitutions which result in a substantial loss of activity relative to the wild type target peptide are considered to be non-tolerated amino acids and not part of the binding motif and/or amino acids substitutions which do not result in a substantial loss of activity relative to the wild type target peptide are considered as part of the binding motif.

The activity that is tested may be the ability of the mutant to bind to the binding peptide and/or to elicit the biological response caused by binding to the binding peptide. The alternative amino acid may have a different side chain to that of the amino acid for which it is being substituted.

The alternative amino acid may be one that does not appear in the sequence that is involved in binding to the target peptide.

The alternative amino acid may be alanine or glycine.

The search may be carried out for peptides that are expressed in selected tissue(s) and/or accessible to the binding peptide.

Methods of the invention may further comprise testing binding to the target peptide of any peptide that is present in the subject that comprises the at least one binding motif. Methods of the invention may further comprise:

when there is the absence of one or more of such peptides and hence the indication that the binding peptide has the potential not to cross react in vivo, administering a treatment effective amount of the binding peptide to the subject, and/or

when there is the presence of one or more such peptides and hence an indication that the binding peptide has the potential to cross react in vivo, identifying the potential for each peptide to cause off target side effects in vivo, and where necessary, preparing an alternative binding peptide having the absence of one or more of such peptides and hence the indication that the alternative binding peptide has the potential not to cross react in vivo, and administering a treatment effective amount of the alternative binding peptide to the subject.

The treatment effective amount of the binding peptide is that amount typically given to the suitable mammalian patient. Thus, the invention comprehends testing known binding peptide treatments for whether such treatment binding peptides will bind to peptides other than the target peptide. The treatment effective amount of the alternative binding peptide is within the ambit of the skilled person from this disclosure and the knowledge in the art. For example, the treatment effective amount can be determined by comparing the binding properties and/or ability to elicit the desired treatment biological response of the binding peptide with the binding properties and/or ability to elicit the desired treatment biological response of the alternative binding peptide, and adjusting the dosage of the binding peptide based on the difference in binding and/or ability to elicit the biological response of the alternative binding peptide in comparison with the binding and/or ability to elicit the biological response of the binding peptide. I n advantageous embodiments, the alternative binding peptide has binding properties and/or biological response eliciting properties akin to that of the binding peptide, and hence its dosage or amount to be administered is analogous to that of the binding peptide.

In certain aspects, the invention therefore provides an improvement in a method of treating a human or animal mammalian subject in need thereof comprising administering a binding peptide. This improvement comprises:

predicting whether the binding peptide, which binds to a target peptide presented by a Major Histocompatibility Complex (MHC) in the subject to thereby provide said treatment and is for administration to the subject, has the potential to cross react with another peptide in the subject in vivo, comprising:

identifying at least one binding motif in the target peptide to which the binding peptide binds; and

searching for peptides that are present in the subject that comprise the at least one binding motif and that are not the target peptide,

wherein the presence of one or more such peptides indicates that the binding peptide has the potential to cross react in vivo, and the absence of one or more of such peptides indicates that the binding peptide has the potential to not to react in vivo; and, when there is the absence of one or more of such peptides and hence the indication that the binding peptide has the potential will not to react in vivo, and/or when binding to the off target peptide is not expected to cause unwanted side effects in vivo, administering a treatment effective amount of the binding peptide to the subject. This method can include, when there is the presence of one or more such peptides and hence an indication that the binding peptide has the potential to cross react in vivo, preparing an alternative binding peptide having the absence of one or more of such peptides and hence the indication that the alternative binding peptide has the potential not to cross react in vivo, and administering a treatment effective amount of the alternative binding peptide to the subject.

Moreover, the invention has utility in the preparation of pharmaceutical compositions comprising a binding peptide. In this respect, the invention provides a method for preparing a pharmaceutical composition comprising a binding peptide, or an improvement to methods for preparing a pharmaceutical composition comprising a binding peptide, comprising:

predicting whether the binding peptide, which binds to a target peptide presented by a Major Histocompatibility Complex (MHC) in the subject to thereby provide said treatment and is for administration to the subject, has the potential to cross react with another peptide in the subject in vivo, comprising:

identifying at least one binding motif in the target peptide to which the binding peptide binds; and

searching for peptides that are present in the subject that comprise the at least one binding motif and that are not the target peptide,

wherein the presence of one or more such peptides indicates that the binding peptide has the potential to cross react in vivo, and the absence of one or more of such peptides indicates that the binding peptide has the potential not to cross react in vivo. Accordingly, it is an object of the invention not to encompass within the invention any previously known product, process of making the product, or method of using the product such that Applicants reserve the right and hereby disclose a disclaimer of any previously known product, process, or method. It is further noted that the invention does not intend to encompass within the scope of the invention any product, process, or making of the product or method of using the product, which does not meet the written description and enablement requirements of the USPTO (35 U.S.C. §112, first paragraph) or the EPO (Article 83 of the EPC), such that Applicants reserve the right and hereby disclose a disclaimer of any previously described product, process of making the product, or method of using the product.

It is noted that in this disclosure and particularly in the claims and/or paragraphs, terms such as "comprises", "comprised", "comprising" and the like can have the meaning attributed to it in U.S. Patent law; e.g., they can mean "includes", "included", "including", and the like; and that terms such as "consisting essentially of" and "consists essentially of" have the meaning ascribed to them in U.S. Patent law, e.g., they allow for elements not explicitly recited, but exclude elements that are found in the prior art or that affect a basic or novel characteristic of the invention.

These and other embodiments are disclosed or are obvious from and encompassed by, the following Detailed Description.

BRIEF DESCRIPTION OF THE DRAWINGS

The following Detailed Description, given by way of example, but not intended to limit the invention solely to the specific embodiments described, may best be understood in conjunction with the accompanying drawings, wherein:

Figure 1 shows IFNv production by a3a T cells in response to peptide-pulsed cells presenting either MAGE A3 peptide (denoted WT), or alanine-substituted peptides. Figure 2 shows IFNv production by IMCmagel redirected T cells in response to peptide- pulsed cells presenting either MAGE A3 peptide (denoted WT), or alanine-substituted peptides. Figure 3 shows IFNv production by IMCmagel redirected T cells in response to peptide- pulsed cells presenting either MAGE A3 peptide (denoted WT), or glycine-substituted peptides. Figure 4 shows IFNv production by a3a T cells in response to cells pulsed with peptides identified in the motif search. Non-transduced T cells (ntd) were used as a negative control.

Figure 5 shows IFNv production by IMCmagel redirected T cells in response to cells pulsed with peptides identified in the motif search. Experiments carried out in the absence of IMCmagel were used as a negative control.

Figure 6 shows IMCmagel redirected T cell killing of cells pulsed with either MAGE A3 or Titin peptides. Cell killing was determined using the IncuCyte platform. A negative control was carried out using in the absence of IMCmagel (E + T).

Figure 7 shows the cytokine profile a3a T cells in the presence of iCells, as determined by the Luminex assay. EJM and colo205 cells were used as a positive and negative control respectively. Figure 8 shows phase contrast images, obtained using the IncuCyte platform, of Titin positive iCells being killed by a3a T cells. The images were taken after 24 hours incubation. Non- transduced T cells (ntd) were used as a negative control.

Figure 9 shows details of 15 peptides identified in a BLAST search carried out to find human peptides with a similar sequence to MAGE A3 peptide. Activation of a3a T cells or IMCmagel redirected T cells, as determined by IFNv ELISpot, is indicated.

Figure 10 shows activation of a3a T cells, as determined by IFNv ELISpot, in response to peptide-pulsed cells presenting either MAGE A3, human Titin, or mouse Titin. Figure 11 shows IFNv production by I mmTAC-NYESO redirected T cells in response to peptide-pulsed cells presenting either NY-ESO-1 peptide (denoted WT), or alanine- substituted peptides. Figure 12 shows IFNv production by I mmTAC-NYESO redirected T cells in response to native (WT) NY-ESO-1 peptide and each amino-acid-substituted peptide.

DETAILED DESCRIPTION

The inventors have found that, by identifying the binding motif in the target peptide to which the binding peptide binds, rather than the known binding sequence, off target peptides - that have the potential to, or will, cross react in vivo - can be identified far more accurately. If off target peptides are found, this allows binding peptides to be altered so that they do not bind to these off target peptides. The binding motif(s) may be identified by creating a series of mutants of the target peptide, each mutant having the amino acid residue at one position in the sequence thereof that is involved in binding to the binding peptide (the "binding sequence") substituted for an alternative amino acid, such that over the series of mutants the amino acid residue in each position in the binding sequence is substituted for an alternative amino acid. Each mutant in the series is then tested for its activity relative to the wild type target peptide. An amino acid residue at a position within the binding sequence may be identified as being part of the binding motif if the mutant in which the amino acid at that position is mutated to an alternative amino acid has a substantial loss of activity relative to the wild type target peptide, such as 50, 55, 60, 65, 70, 75, 80, 85, 90% or greater loss of activity. This may result in a binding motif comprising amino acid(s) at one or a plurality of positions (such as 2, 3, 4, 5, 5, 6, 7, 8, 9, 10 or more) within the binding sequence being identified.

Where an amino acid residue at a position in the binding sequence is not identified as being part of the binding motif (using the technique described above for example), this position may be further substituted with at least one additional amino acid and tested for activity relative to the wild type peptide. Amino acid substitutions which result in a substantial loss of activity relative to the wild type target peptide, such as 50, 55, 60, 65, 70, 75, 80, 85, 90% or greater loss of activity, are considered to be non-tolerated amino acids and/or not part of the binding motif. Conversely, amino acid substitutions which do not result in a substantial loss of activity (for example of at least 50%) relative to the wild type target peptide may be considered as part of the binding motif. This may result in a binding motif as defined above and additionally in which one or a plurality of positions are represented by more than one but not all amino acids. Alternatively or additionally, a series of mutants may be made in which the amino acid residue at each position of the binding sequence is mutated to all alternative amino acids (relative to the wild type amino acid). For a binding sequence of nine amino acids, this would mean 171 peptides. Each mutant in the series is tested for activity relative to the wild type peptide. Amino acid substitutions which result in a substantial loss of activity relative to the wild type target peptide, such as 50, 55, 60, 65, 70, 75, 80, 85, 90% or greater loss of activity, are considered to be non-tolerated amino acids and/or not part of the binding motif. Conversely, amino acid substitutions which do not result in a substantial loss of activity (for example of at least 50%) relative to the wild type target peptide may be considered as part of the binding motif. This may result in a binding motif in which each position is represented by at least one but not all amino acids.

The activity that is tested may be the ability of the mutant to bind to the binding peptide (this can be measured using Surface Plasmon Resonance for example) and/or to elicit the biological response caused by binding to the binding peptide. The biological response may be, for example, activation of immune system cells such as T cells, measured by cytokine production or destruction of the target cell; activation of an enzyme, measured by accumulation of product or disappearance of substrate; or activation of a signalling cascade (measured by monitoring protein phosphorylation, or changes in gene expression and protein production). The alternative amino acid substituted into the mutants may be alanine, glycine or indeed any amino acid, provided that it has a different side chain to that of the amino acid for which it is being substituted. Preferably, the alternative amino acid is one that does not appear in the binding sequence. Thus, any one of the following amino acids may be used: alanine, asparagine, aspartic acid, arginine, cysteine, glutamine, glycine, glutamic acid, histidine, isoleucine, lysine, leucine, phenylalanine, methionine, serine, proline, tryptophan, threonine, tyrosine, valine, as well as non-naturally occurring amino acids. Such techniques, known as "amino acid scanning", are known in the art and have been used to determine a binding motif. See Wells, Methods Enzymol 1991; 202: 390-411 for example. A similar approach is also known for detecting other specificities of a T cell receptor (for example Udyavar et al. J Immunol 2009; 182(7):4439-47). However, such amino acid scanning has not been used to identify off-target peptides which may cause undesirable side effects.

Once the binding motif has been identified, protein databases may be searched for proteins which contain the binding motif. Suitable protein databases include but are not limited to UniProtKB/Swiss-Prot (http://www.uniprot.org/), Protein Information Resource (PI R) (http://pir.georgetown.edu/pirwww/index.shtml), and/or Reference Sequence (RefSeq) (www.ncbi.nlm.nih.gov/RefSeq). Searching for a peptide motif may be carried out using any one of a number of tools, which may be found on bioinformatics resource sites such as ExPASY (http://www.expasy.org/). For example, the search tool ScanProsite identifies user-defined motifs in all protein sequences in the UniProtKB/Swiss-Prot Protein Knowledgebase (De Castro et al. Nucleic Acids Res. 2006 Jul 1; 34 (Web Server issue):W362-5).

It is preferred that peptides containing the exact binding motif, i.e. with 100% identity to the binding motif, are identified. However, proteins containing motifs that have less than 100% identity to the binding motif may be identified. For the purposes of searching, the binding motif may be modified to include ambiguity at certain positions, for example with amino acids which have similar properties (e.g. leucine/ isoleucine, etc), or where it is already known in the literature that a particular amino acid is tolerated at a certain position (such as HLA anchor residues).

The search may be carried out for peptides that are of human origin or of organisms which are commonly present in humans, such as viral or bacterial pathogens, or commensal bacteria. However, where the method of the present invention is applied to non-human animals, such as non-human mammals, the search may be carried out for peptides that are of the relevant non-human animal origin or of organisms which are commonly present in such non-human animals. Additionally or alternatively, the search may be carried out for peptides that are expressed in selected tissue(s) and/or accessible to the binding peptide. Such information may be obtained from the literature.

Peptides identified in the search as comprising the at least one binding motif may be confirmed as an off target peptide and cross reacting with binding peptide by determining the ability of the identified peptide to bind to the binding peptide (for example using Surface Plasmon Resonance), or assessing the biological response generated by binding of the binding peptide to the identified peptide. The biological response may be for example, activation of immune system cells such as T cells, measured by cytokine production or destruction of the target cell; activation of an enzyme, measured by accumulation of product or disappearance of substrate; or, activation of a signalling cascade (measured by monitoring protein phosphorylation, or changes in gene expression and protein production).

Once one or more off target peptides have been identified, the potential of the (or each) off target peptide to cause unwanted side effects when bound by the binding peptide may be determined. This may include searching literature sources to determine expression of the off target peptide in normal tissue. For example, where expression of the off target peptide in normal tissue is non-existent or limited (for example with cancer testis antigens), the binding peptide may be considered suitable for administration in vivo. In cases where expression of the off target peptide in normal tissue is widespread or is in critical tissues, such as heart cells, binding may optionally be additionally confirmed in vitro using cells which express the off target peptide. In some situations the binding peptide may not be administered in vivo because of the cross reaction that this can cause. The binding peptide may be redesigned so that there is no longer any cross reactivity to the off target peptide(s), while maintaining binding, preferably with high affinity, to the target peptide. For example, T cell receptors can be redesigned by mutagenesis using the methods described in WO 03/020763. Where redesigning does not prevent cross reactivity to the off target peptide, an alternative binding peptide may be sought, for example, an alternative T cell receptor, which binds the target peptide.

If no off target peptides are found, or if the binding to the off target peptides is not expected to cause unwanted side effects (for example by virtue of the off target peptide being expressed in limited or non-critical tissues), the binding peptide may be used in a method of preventing or treating a disease or condition which is ameliorated by administration of the binding peptide. Methods of treatment include but are not limited to immunotherapies; for example, administration to a patient of modified T cells (adoptive therapy), such as those transduced with affinity enhanced T cell receptors or chimeric antibody receptors;

administration of monoclonal antibodies or monoclonal antibody fragments, especially TCR- like antibodies; administration of novel bi-specific immunotherapeutic agents such as ImmTACs (Immune mobilising TCRs against cancer) (Liddy, et al. (2012) Nat Med 18: 980- 987) or BiTEs (Bispecific T cell engaging antibodies) (Baeuerle, et al. (2009). Curr Opin Mol Ther 11(1): 22-30).

Such treatments may be provided in a pharmaceutical composition together with one or more pharmaceutically acceptable carriers or excipients. Therapeutic TCRs, or cells, will usually be supplied as part of a sterile, pharmaceutical composition which will normally include a pharmaceutically acceptable carrier. This pharmaceutical composition may be in any suitable form, (depending upon the desired method of administering it to a patient). It may be provided in unit dosage form, will generally be provided in a sealed container and may be provided as part of a kit. Such a kit would normally (although not necessarily) include instructions for use. It may include a plurality of said unit dosage forms. The pharmaceutical composition may be adapted for administration by any appropriate route, such as a parenteral (including subcutaneous, intramuscular, or intravenous) route. Such compositions may be prepared by any method known in the art of pharmacy, for example by mixing the active ingredient with the carrier(s) or excipient(s) under sterile conditions.

Dosages of the substances of the present invention can vary between wide limits, depending upon the disease or disorder to be treated, the age and condition of the individual to be treated, etc; for example, a suitable dose range for an ImmTAC reagent may be between 25 ng/kg and 50 μg/kg. A physician will ultimately determine appropriate dosages to be used.

In the present invention, the binding peptide that binds to MHC presented peptide may be an immune binding peptide, which may be an immunotherapeutic peptide. Binding peptides can be derived from natural sources, or they may be partly or wholly synthetically produced. Examples of immune binding peptides include T cell receptors ("TCRs" - which term includes antigen binding fragments of T cell receptors). As is described in WO 99/60120, TCRs mediate the recognition of specific Major Histocompatibility Complex (MHC)-peptide complexes by T cells and, as such, are essential to the functioning of the cellular arm of the immune system. The TCR is a heterodimeric cell surface protein of the immunoglobulin superfamily which is associated with invariant proteins of the CD3 complex involved in mediating signal transduction. TCRs exist in αβ and γδ forms, which are structurally similar but T cells expressing them have quite distinct anatomical locations and probably functions. The extracellular portion of the receptor consists of two membrane-proximal constant domains, and two membrane-distal variable domains bearing polymorphic loops analogous to the complementarity determining regions (CDRs) of antibodies. It is these loops which form the binding site of the TCR molecule and determine peptide specificity.

The TCR may be in soluble form (e.g. having no transmembrane or cytoplasmic domains), for example a monoclonal TCR as described in WO03/020763, and/or in single chain form, as described in WO2004/033685. For stability, soluble TCRs preferably have an introduced disulphide bond between residues of the respective constant domains, as described, for example, in WO 03/020763. Single chain formats include αβ TCR polypeptides of the Va-L- νβ, νβ-L-Va, Va-Ca-L-νβ or Ν/α-ί-νβ-Οβ types, wherein Va and νβ are TCR a and β variable regions respectively, Ca and 0β are TCR a and β constant regions respectively, and L is a linker sequence. Alternatively or additionally, the TCR may be fused to an immune effector domain for use as a targeting agent for delivering therapeutic agents to an antigen presenting cell. Such therapeutic agents include for example antibodies or antibody fragments such as an anti-CD3 fragment, immunomodulators such as cytokines, enzymes such as perforin, or chemotherapeutic agents, such as cis-platin. TCRs may also be expressed on a cell, such as a T cell. Said T cells may be used in adoptive therapy.

Other binding peptides encompassed by the present invention include antibodies, such as TCR like antibodies, which have been engineered to bind to MHC presented peptides (for example see, Sergeeva, A., G. et al. (2011). Blood 117(16): 4262-72 and/or Dahan, R., and Y. Reiter. 2012. Expert Rev Mol Med. 14:e6. The term "antibody" as used herein refers to immunoglobulin molecules and immunologically active portions of immunoglobulin molecules, i.e., molecules that contain an antigen binding site that specifically binds an antigen, whether natural or partly or wholly synthetically produced. The term also covers any polypeptide or protein having a binding domain which is, or is homologous to, an antibody binding domain. Examples of antibodies are the immunoglobulin isotypes (e.g., IgG, IgE, IgM, IgD and IgA) and their isotypic subclasses; fragments which comprise an antigen binding domain such as Fab, scFv, Fv, dAb, Fd; and diabodies. Antibodies may be polyclonal or monoclonal. A monoclonal antibody may be referred to herein as "mab".

It is possible to take monoclonal and other antibodies and use techniques of recombinant DNA technology to produce other antibodies or chimeric molecules which retain the specificity of the original antibody. Such techniques may involve introducing DNA encoding the immunoglobulin variable region, or the complementary determining regions (CDRs), of an antibody to the constant regions, or constant regions plus framework regions, of a different immunoglobulin. See, for instance, EP-A-184187, GB 2188638A or EP-A-239400. A hybridoma or other cell producing an antibody may be subject to genetic mutation or other changes, which may or may not alter the binding specificity of antibodies produced. As antibodies can be modified in a number of ways, the term "antibody" should be construed as covering antibody fragments, derivatives, functional equivalents and homologues of antibodies, humanised antibodies, including any polypeptide comprising an immunoglobulin binding domain, whether natural or wholly or partially synthetic. Chimeric molecules comprising an immunoglobulin binding domain, or equivalent, fused to another polypeptide are therefore included. Cloning and expression of chimeric antibodies are described in EP-A-0120694 and EP-A-0125023. A humanised antibody may be a modified antibody having the variable regions of a non-human, e.g. murine, antibody and the constant region of a human antibody. Methods for making humanised antibodies are described in, for example, US Patent No. 5225539

It has been shown that fragments of a whole antibody can perform the function of binding antigens. Examples of binding fragments are (i) the Fab fragment consisting of VL, VH, CL and CHI domains; (ii) the Fd fragment consisting of the VH and CHI domains; (iii) the Fv fragment consisting of the VL and VH domains of a single antibody; (iv) the dAb fragment (Ward, E.S. et ol., Nature 341:544-546 (1989)) which consists of a VH domain; (v) isolated CDR regions; (vi) F(ab')2 fragments, a bivalent fragment comprising two linked Fab fragments (vii) single chain Fv molecules (scFv), wherein a VH domain and a VL domain are linked by a peptide linker which allows the two domains to associate to form an antigen binding site (Bird et ol., Science 242:423-426 (1988); Huston et ol., PNAS USA 85:5879-5883 (1988)); (viii) bispecific single chain Fv dimers (PCT/US92/09965) and (ix) "diabodies", multivalent or multispecific fragments constructed by gene fusion (WO94/13804; P.

Hollinger et ol., Proc. Notl. Acad. Sci. USA 90: 6444-6448 (1993)). Diabodies are multimers of polypeptides, each polypeptide comprising a first domain comprising a binding region of an immunoglobulin light chain and a second domain comprising a binding region of an immunoglobulin heavy chain, the two domains being linked (e.g. by a peptide linker) but unable to associated with each other to form an antigen binding site: antigen binding sites are formed by the association of the first domain of one polypeptide within the multimer with the second domain of another polypeptide within the multimer (WO94/13804). Where bispecific antibodies are to be used, these may be conventional bispecific antibodies, which can be manufactured in a variety of ways (Hollinger & Winter, Current Opinion Biotechnol. 4:446-449 (1993)), e.g. prepared chemically or from hybrid hybridomas, or may be any of the bispecific antibody fragments mentioned above. It may be preferable to use scFv dimers or diabodies rather than whole antibodies. Diabodies and scFv can be constructed without an Fc region, using only variable domains, potentially reducing the effects of anti-idiotypic reaction. Other forms of bispecific antibodies include the single chain "Janusins" described in Traunecker et al., EMBO Journal 10:3655-3659 (1991). Bispecific diabodies, as opposed to bispecific whole antibodies, may also be useful because they can be readily constructed and expressed in E. coli. Diabodies (and many other polypeptides such as antibody fragments) of appropriate binding specificities can be readily selected using phage display (WO94/13804) from libraries. If one arm of the diabody is to be kept constant, for instance, with a specificity directed against antigen X, then a library can be made where the other arm is varied and an antibody of appropriate specificity selected. An "antigen binding domain" is the part of an antibody which comprises the area which specifically binds to and is complementary to part or all of an antigen. Where an antigen is large, an antibody may only bind to a particular part of the antigen, which part is termed an epitope. An antigen binding domain may be provided by one or more antibody variable domains. An antigen binding domain may comprise an antibody light chain variable region (VL) and an antibody heavy chain variable region (VH). Also encompassed within the present invention are binding peptides that bind to MHC presented peptides and are based on engineered protein scaffolds. Protein scaffolds are derived from stable, soluble, natural protein structures which have been modified to provide a binding site for a target molecule of interest. Examples of engineered protein scaffolds include, but are not limited to, affibodies, which are based on the Z-domain of

staphylococcal protein A that provides a binding interface on two of its a-helices (Nygren, P. A. (2008). FEBS J 275(11): 2668-76); anticalins, derived from lipocalins, that incorporate binding sites for small ligands at the open end of a beta-barrel fold (Skerra, A. (2008) FEBS J 275(11): 2677-83), nanobodies, and DARPins. Engineered protein scaffolds are typically targeted to bind the same antigenic proteins as antibodies, and are potential therapeutic agents. They may act as inhibitors or antagonists, or as delivery vehicles to target molecules, such as toxins, to a specific tissue in vivo (Gebauer, M. and A. Skerra (2009). Curr Opin Chem Biol 13(3): 245-55). Short peptides may also be used to bind a target protein. Phylomers are natural structured peptides derived from bacterial genomes. Such peptides represent a diverse array of protein structural folds and can be used to inhibit/disrupt protein-protein interactions in vivo (Watt, P. M. (2006). Nat Biotechnol 24(2): 177-83)].

Although the present invention has been described with reference to predicting whether a binding peptide, which binds to a target peptide that is presented in the context of MHC, will cross react with another peptide in a subject in vivo, it is to be understood that the techniques described herein can be applied to any target peptide, regardless of whether it is presented in the context of MHC. Thus, binding peptide and target peptide may be any pair of molecules which have binding specificity for one another. Other examples of such pairs of molecules include hormone-hormone receptor, receptor-ligand, enzyme-substrate.

With regard to the methods of the invention, as an example, the invention can be practiced with regard to any known antibody treatments. Examples of FDA-approved therapeutic monoclonal antibodies are set forth in the following table.

Example FDA approved therapeutic monoclonal antibodies

Brand Approval Indication

Antibody Company Type Target

name date (Targeted disease)

Abcfximab ReoPro Eii Lilly 1994 chimeric inhibition of Cardiovascular disease glycoprotein Example FDA approved therapeutic monoclonal antibodies

Brand Approval Indication

Antibody Company Type Target

name date (Targeted disease) llb/llla

inhibition of Several auto-immune

A alimuiTiab Humira Abbot 2002 human

T F^a signaling disorders

Chronic lymphocytic

AJermuz ii-nab Campath Go:vy;ri<:- 2001 humanized CD52

leukemia lj.-2Ra. receptor

asi!iximab Simulect Novartis 1998 chimeric Transplant rejection

(CD ¾)

Bejimumab Benlysta GlaxoSmithKiine 2011 human cell activating erythematosus'^' ¹¹ ^""^ factor ^M1

Vascular

Colorectal cancer. Age endothelial

Bevaci Limab Avastin Genentech/ oche 2004 humanized

growth factor

. JLeD.gr3tjon.(of -iabei

(VivGF)

Anagiastic large ce!i

Brentuximab

Adcetris 2011 Chimeric CD30 iym& QIO (ALCL) and vedotin

Canakinumab Maris Novartis 2009 Human IL fi Cryopyrin-associated periodic syndrome Example FDA approved therapeutic monoclonal antibodies

Brand Approval Indication

Antibody Company Type Target

name date (Targeted disease)

(CAPS)

and neck cancer

Li!!y/Merck KGaA receptor

Certolizumab inhibition of

Cimzia UCB (company) 2008 humanized Crohn's disease pego; ' I Ed* signaling jj.-2 ot receptor

Dadizurnab Zenapax Genentech/Roche 1997 humanized Transplant rejection

(CD25)

Postmenopausal

Prolia , RAN K .igand

Denosurnab Amsen 2010 Human osteoporosis , Solid

Xgeva inhibitor

tumor ^' s bony metasteses

Eculizumab Soliris

Efaliziimab Raptiva 2002 humanized CD11 Psoriasis

Gemtuzumab Mylotarg W e h 2000 humanized CD33 leukemia (with

caiicheamicin) Example FDA approved therapeutic monoclonal antibodies

Brand Approval Indication

Antibody Company Type Target

name date (Targeted disease)

Johnson & Rheumatoid arthritis,

Goiimumab Simponi Johnson/Merck & 2009 Human Psoriatic arthritis, and

Inihibiior

Co, inc. Ankylosing spondylitis

Spectrum

Ibriturnomab

Zevalin Rb _/ Lanacejrdcais, 2002 murine CD20 (with ytiTlurn-SO or tiuxetan

Inc. indiurr¾;_lll)

Janssen Biotech, inhibition of Several autoimmune jnfi i¾j a Remicade 1998 chimeric

Inc./ erck & Co TN F-a signaling disorders

Ipillrnurnab (

Yervoy 2011 Human blocks ΠΊ.Α-4 Mci noi i:;

MDX-101 )

Muromona - Orthoclone T :eil ^ΏΆ

Janssen-Ciiag 1986 murine Transplant rejection

C OKT3 Receptor alpha-4 (a4) M u 11 i i e _. cj eras] and

Nataiizumab Tysabri Bio en idec/Eian 2006 humanized

i0i.eg. i. j Crohn ^' s dj sea se

Chronic lymphocytic

Ofal:tj _j r;njmab Arzerra 2009 Human CD20

leukemia immunoglobulin mainly ajierg^-related

QiTiaiiz roab Xolair Ger.entech/Novartis 2004 humanized

I (lgE) asthma Example FDA approved therapeutic monoclonal antibodies

Brand Approval Indication

Antibody Company Type Target

name date (Targeted disease) an epitope of

Palivizumab Synagis Med immune 1998 humanized the RSV F

Virus

protein epidermal

Panitumumab Vectibix Amgen 2006 human growth factor GjGrectaj. anci:

receptor

Vascular

endothelial

Ranibizum b Lucentis Genentech/Novsrtis. 2006 humanized

growth factor A

(VEG -A)

Rituxan, Biogeri

Rituximab 1997 chimeric CD20 Non-Hodgkin lymphoma

Mabthera Idec/Geneniech

Actemra

and 2010 Humanised Anti- IL--6R Rheumatoid arthritis or Ai!izumab 5

RoActemra

Tositurnomab Bexxar GlaxoSmithKline 2003 murine CD20 Non-Hodgkin lymphoma

Trastuzumab Herceptin Genentech 1998 humanized Erb82 Breast cancer In this regard, mention is also made of bispecific antibodies with which the invention can also be practiced. Bispecific antibodies are a particular class of therapeutic antibodies that have yielded promising results in clinical trials, and in April 2009, the bispecific antibody catumaxomab was approved in the European Union.

The methods of the invention can be practiced with such binding peptides as these monoclonal and bispecific antibody treatments, and hence the skilled person readily knows the formulation and dose and means to administer if there is no cross reactivity detected using the instant invention, and from the dosages of these known binding peptides, can readily determine, using the herein disclosure and the knowledge in the art, the formulation, and dose and means to administer alternative binding peptides.

Preferred features of each aspect of the invention are as for each of the other aspects mutatis mutandis. The prior art documents mentioned herein are incorporated to the fullest extent permitted by law. The present invention will be further illustrated in the following Examples which are given for illustration purposes only and are not intended to limit the invention in any way; but rather, the Applicants reserve the right to both generalise from the Examples when claiming, and provide from the Examples specific claims. Examples

The present invention will now be described with reference to the following non-limiting examples.

Example 1

The target MHC presented peptide used in this example is derived from human cancer testis antigen MAGE A3 and has the following amino acid sequence; EVDPIGHLY. MAGE A3 peptide is presented on antigen presenting cells in the context of HLA-A*01. The binding peptide used in this example comprises a modified T cell receptor (TCR) which has been engineered to possess enhanced affinity for MAGE A3 peptide. Methods to produce affinity enhanced TCRs are known in the art (for example, phage display WO 03/020763). The native MAGE A3 TCR was obtained from a MAGE A3 T cell clone, as described in WO2012/013913. Two versions of the modified MAGE A3 TCR are used. A moderately affinity enhanced version, expressed by transduced T cells (termed a3a T cells) as described in WO2012/013913, and a high affinity version, produced as a soluble protein fused to a T cell activating anti-CD3 fragment (termed IMCmagel), according to the method of WO2010/133828.

1.1 Identification of the binding motif

Variants of the native MAGE A3 peptide were obtained in which each amino acid position was sequentially replaced with alanine, as shown below (in each case the alanine substitution is underlined). Peptides were obtained from Peptide Protein Research Limited, UK.

AVDPIGHLY EADPIGHLY

E VAP I G H LY

E VD Al G H LY

E VD PAG H LY

EVDPI AH LY EVDPIGALY

EVDPIGHAY

EVDPIGH LA

The native and alanine-substituted peptides were pulsed on to antigen presenting cells, and interferon γ (IFNv) production, as measured using the ELISpot assay, used as a read-out for T cell activation. Essential positions were defined by a greater than 50% reduction in T cell activity relative to the native peptide.

1.1a) Activation of a3a T cells by alanine-substituted peptides

ELISpot assays were carried out according to the manufacturer's instructions (BD

Biosciences). HLA-A1+ hepatocyte cells were used as target cells and pulsed with 10 μΜ of each peptide. Target cells were counted and plated at 50,000 cells per well in 50 μΙ assay buffer (10% FCS, 88% RPMI 1640, 1% glutamine and 1% penicillin/streptomycin). Effector T cells used in this method were a 1:1 mix of CD4+ and CD8+ T cells (obtained by negative selection (using the CD4 and CD8 Negative Isolation Kits, Dynal) from peripheral blood lymphocytes (PBL) obtained from a healthy donor). Cells were stimulated with anti CD3/CD28 coated beads (T cell expander, Invitrogen), transduced with lentivirus carrying the gene encoding a3a T cell receptor, and expanded in assay media containing 50 U/ml IL-2 until between 10 and 13 days post transduction. Effector T cells were plated at 15,000 cells per well. Plates were incubated overnight at 37 °C/5% C0 ₂ and quantified, after

development, using an automated ELISpot reader (Immunospot Series 5 Analyzer, Cellular Technology Ltd.). Non-transduced PBLs from the same healthy donor were used as a negative control. All experiments were carried out in triplicate.

Figure 1 shows IFNv production by a3a transduced T cells in response to native (wt) MAGE A3 peptide and each alanine-substituted peptide. Five of the alanine-substituted peptides resulted in a greater than 50 % decrease in IFNv production compared to native MAGE A3 peptide. The corresponding native residue at each of these five positions comprises the binding motif. In this case the binding motif is defined as EXDPIXXXY, where X is any amino acid. 1.1b) Activation of IMCmagel redirected T cells by alanine-substituted peptides

ELISpot assays were carried out as described in section (1.1a), except effector cells were prepared from peripheral blood mononuclear cells (PBMCs) blood using standard procedures utilising Lymphoprep (Axis-Shields, cat# NYC-1114547) and Leucosep tubes (Greiner, cat# 227290), and plated at 25,000 cells per well. IMCmagel was added to a final concentration of 0.1 nM per well. Controls were carried out in the absence of IMCmagel (effectors + targets + peptide); and in the absence of peptide-pulsed target cells (effectors + IMCmagel).

Figure 2 shows IFNv production by IMCmagel redirected T cells in response to native (wt) MAGE A3 peptide and each alanine-substituted peptide. Five of the alanine-substituted peptides resulted in a greater than 50 % decrease in IFNv production compared to native MAGE A3. The corresponding native residue at each of these five positions comprises the binding motif. In this case the binding motif is defined as EXDPIXXXY, where X is any amino acid.

1.1c) Activation of IMCmagel redirected T cells by glycine-substituted peptides The same procedure was followed as detailed in section (1.1b) except the HLA-A*01+ hepatocyte target cells were pulsed with glycine-substituted peptides. MAGE A3 peptide contains one native glycine residue; in this case the native glycine was considered nonessential for the purposes of defining the motif. Figure 3 shows IFNv production by IMCmagel redirected T cells in response to native (wt) MAGE A3 peptide and each glycine-substituted peptide. Five of the glycine-substituted peptides resulted in a greater than 50 % decrease in IFNv production compared to native MAGE A3. The corresponding native residue at each of these five positions comprises the binding motif. In this case the binding motif is defined as EXDPIXXXY, where X is any amino acid. 1.2 Identification of potential off-target peptides

The ScanProsite tool (http://prosite,expasy,org/seanprosjte) was used to search all

UniProtKB/Swiss-Prot (release 2012_10 of 31-Oct-12: 538259 entries) database sequences, for proteins which contain the motif identified above (entered as E-X-D-P-l-X-X-X-Y). No filters were used. Pattern options were set to allow at most 1 X sequence characters to match a conserved position in the pattern and the match mode was set to 'greedy, overlaps, no includes'.

Five unique human proteins were identified: native MAGE A3, MAGE family members A6 and B18, the muscle protein Titin, and a protein known as Dos. The amino acid sequences of the motif-containing peptides are shown in the table below (residues which comprise the motif are underlined).

The search results also identified a number of motif-containing peptides from common human pathogens. Two peptides were selected from the list as examples for further testing; a nuclease protein present in three strains of Epstein Barr Virus (EBV), and a ribosomal maturation factor from Clostridium difficile, the amino acid sequences of these peptides are shown below. Protein (Accession number) Sequence

EBV protein (Q1HVE7)

E F D P 1 Y P S Y (P03217) (Q3KSR5)

C. difficile protein (Q18BH3) E K D P 1 K E N Y

1.3 Confirming cross reactivity of off target peptides Potential off target peptides, identified above, were tested for their ability to cross react with a3a T cells and IMCmagel redirected T cells using an IFNv ELISpot assay.

1.3a) Testing for activation of a3a T cells by HLA-A*01+ hepatocyte cells pulsed with potential off-target peptides

All peptides were produced synthetically by Peptide Protein Research Limited, UK.

Activation of a3a T cells was determined by IFNv ELISpot assay using the same procedure as described in section 1.1a. HLA-A*01+ hepatocyte cells were used as targets and pulsed with 10 μΜ of each peptide. T cells were prepared from peripheral blood lymphocytes (PBLs) obtained from a healthy donor, and transduced with the a3a TCR. Non-transduced donor T cells were used as a negative control. T cell activation in response to cells pulsed with each of the motif-containing peptides was compared to activation by native MAGE A3 peptide pulsed cells.

Figure 4 shows that cells pulsed with MAGE A3, MAGE A6, MAGE B18 and Titin peptides led to activation of a3a T cells. Peptide Dos as well as peptides from EBV and C. difficile did not induce T cell activation. 1.3b) Testing for activation of IMCmagel redirected T cells by HLA-A*01+ hepatocyte cells pulsed with potential off target peptides

Peptide pulsed cells were prepared as described in 1.3a. Activation of IMCmagel redirected T cells was determined by IFNv ELISpot assay using the same procedure as described in section 1.1b.

Figure 5 shows that cells pulsed with MAGE A3, MAGE A6, MAGE B18 and Titin peptides led to activation of IMCmagel redirected T cells. Peptide Dos as well as peptides from EBV and C. difficile did not induce T cell activation.

The experiments described in 1.3a and 1.3b confirm that there are three off target peptides which are bound by a3a T cells and IMCmagel redirected T cells; MAGE A6, MAGE B18, and Titin. Of the three only Titin is relevant in a clinical context. Normal tissue expression of MAGE family proteins is restricted to male germ-line cells, whereas Titin is expressed in cardiac and skeletal muscle (Uniprot Protein Knowledgebase

(htvp://vvvvw.ur:;p! L.or¾/t:niprot).

1.3c) Measuring affinity to T cell activating peptides

Affinity was determined by surface plasmon resonance using a BIAcore 3000 instrument and reported in terms of an equilibrium dissociation constant (K _D ). Soluble versions of the a3a and IMCmagel TCRs were prepared using the method described in Boulter, et al., Protein Eng, 2003. 16: 707-711. Biotinylated specific and control pMHC monomers were prepared as described in Garboczi, et al. Proc Natl Acad Sci U S A 1992. 89: 3429-3433 and

O'Callaghan, et al., Anal Biochem 1999. 266: 9-15, and immobilized on to a streptavidin- coupled CM-5 sensor chips. All measurements were performed at 25°C in PBS buffer (Sigma) supplemented with 0.005% Tween (Sigma) at a constant flow rate. To measure affinity, serial dilutions of the soluble TCRs were flowed over the immobilized pMHCs and the response values at equilibrium were determined for each concentration. Equilibrium dissociation constants (K _D ) were determined by plotting the specific equilibrium binding against protein concentration followed by a least squares fit to the Langmuir binding equation, assuming a 1:1 interaction.

The results summarised in the table below confirm Titin peptide binds to soluble versions the TCRs used in a3a T cells and IMCmagel. The affinity (K _D ) is shown in the table below:-

1.3d) Cytotoxicity of IMCmagel redirected T cells to HLA-A *01+ hepotocyte cells pulsed with Titin peptide

Killing assays were carried out using the IncuCyte FLR-Platform (Essen Biosciences). Flat bottomed 96 well plates were used for assay. The assay medium was RPMI (w/o phenol red + 10% FCS +1% Pen/strep + 1% Glu). HLA-A*01+ hepatocyte cells were plated at 10,000 cells per well and incubated overnight to allow them to adhere. Peptides were made up at 10 μΜ final concentration and added at 25 μΙ (60 μΜ) per well. Effector T cells (prepared as described in 1.1b) were thawed from liquid N2, counted and plated at 100,000/well.

IMCmagel was added to a final concentration of 1 nM. Control measurements were carried out in the absence of IMCmagel (effectors + targets). Images were taken every 2 h, over a three day period, and the number of apoptotic cells per mm ² was quantified using the CellPlayer 96-well Kinetic Caspase 3/7 reagent and the IncuCyte FLR-Platform. The reagent is cleaved by activated Caspase 3/7 upon target cell apoptosis resulting in the release of the dye and green fluorescent staining of nuclear DNA. Figure 6 shows IMCmagel redirected T cells kill cells pulsed with Titin peptide to a similar degree as native MAGE A3.

1.3e) Activation of a3a T cells by cells naturally presenting Titin peptide iCell cardiomyocytes were obtained from Cellular Dynamics International. iCell

cardiomyocytes are highly purified human cardiomyocytes derived from induced pluripotent stem (iPS) cells and are electrically active with typical mechanical characteristics of cardiac tissue. These cells are positive for Titin and negative for MAGE A3 as determined by standard RT-PCR methods. iCell cardiomyocytes were revived from liquid nitrogen and plated at 50,000 cells per well four days prior to the assay and treated as per the manufacturer's instructions. The iCell cardiomyocytes were virally transduced with HLA-A*01. EJM and colo205 (positive and negative controls respectively) were plated at 50,000 cells per well on the day of assay. iCell cardiomyocytes were washed once with R10 (RPMI + 10% FCS +1% Pen/strep + 1% Glu) and then incubated in R10 for the assay. a3a transduced T cells and corresponding non- transduced cells were taken from culture and plated at 50,000 cells per well. The cells were cultured for 24 h and then centrifuged at 800 xg for 4 min, 100 ul of supernatant was then removed and placed in a clean plate. The plates were stored at -70 °C until thawing and developing to measure multiple cytokine release (IFNv, GM-CSF and MIP-Ιβ) by Luminex (25- plex kit), in accordance with manufacturer's instructions.

Figure 7 shows a3a T cells release IFNv, GM-CSF and MIP-Ιβ in response to iCell

cardiomyocytes. 1.3f) Killing of cells naturally presenting Titin peptide by a3a T cells

Phase contrast images were obtained after 24 h on the the IncuCyte FLR-Platform (Essen Biosciences), using the same experimental set-up as that described above (1.3e). Figure 8 shows the resulting images obtained. In the presence of a3a T cells, iCells are destroyed.

Comparative Example 1

The following experiments show how the methods currently available in art do not, or would not, identify Titin as an off target peptide. These methods were not able to predict unwanted side effects when a3a T cells were administered to patients (manuscript in preparation). a) Similarity search

The RefSeq database was searched for human peptides with a similar sequence to MAGE A3, using protein BLAST (Basic Local Alignment Search Tool) version 2.2.27 (Altschul, et al, Nucleic Acids Res. 1997 25:3389-3402,

http://biast.ncbj.nim. nih.gov/Biast g!?CMD=Web&PAGE TYPE^SIastHome). Algorithm parameters were adjusted automatically for a short protein. (General parameters; expect threshold = 200000, word size = 2: Scoring parameters; matrix = PAM30, gap costs = existence 9 extension 1; no compositional adjustments.)

The BLAST search returned over 100 hits, including MAGE A3, MAGE A6, and MAGE B18, but not Titin. 15 sequences were selected for testing using peptide pulsed targets (HLA-A*01+ hepatocyte cells) and IFNv ELISpot assays as described in Example 1 (section 1.1a and 1.1b). The peptide sequences are detailed in Figure 9. None of these peptides was able to activate a3a T cells or IMCmagel redirected T cells. b) Primary cell screening

HLA-A*01+ human cells from various tissues were screened for their ability to induce activation of a3a T cells and IMCmagel redirected T cells. T cell activation was assessed by IFNv release using an ELISpot assay as described in Example 1 (section 1.1a and 1.1b).

Multiple lots were used for most primary cell types. Cells were grown in 2D culture.

• Astrocytes

• Cardiac Myocytes

• Non-pigmented ciliary epithelial cells

· Bronchial smooth muscle cells

• Dermal fibroblasts

• Dermal microvascular endothelial cells

• Hepatocytes

• Pulmonary Fibroblasts

· Renal Epithelial Cells

• Melanocytes

• Skeletal Muscle

• Pulmonary microvascular endothelial cells Results showed no obvious reactivity for IMCmagel and a3a transduced T cells. Since expression of MAGE family proteins is restricted to male germ line cells cross reactivity with MAGE A6 and MAGE B18 is not detected. However, despite Titin being expressed in cardiac and skeletal muscle, this experiment failed to detect cross reactivity with skeletal and muscle cells. Since titin expression is known to be upregulated in differentiated tissue (Van der Loop, et al. (1996). J Muscle Res Cell Motil. 17:23-36) detecting cross reactivity required more specialised cell cultures (such as the iCells shown above). c) Would a mouse model have identified off-target toxicity?

Sequence alignment of the full length human Titin protein with the mouse Titin protein was carried out using the align function on the Uniprot knowledgebase ( http :/ / ^' www. u n i rot.org/ajjgrt) . The equivalent mouse Titin peptide has the following sequence:-

Activation of a3a T cells was assessed using HLA-A*01+ hepatocyte cells pulsed with MAGE A3, human Titin or mouse Titin peptides. T cell activation was determined by I FNv ELISpot assay as described above. Synthetic peptides were obtained from Peptide Protein Research Limited, UK. Figure 10 shows IFNv release in response to HLA-A*01+ hepatocyte cells pulsed with MAGE A3 and human Titin; no IFNv was detected when cells were pulsed with the equivalent Titin peptide from mouse. Therefore, a mouse model would not have identified Titin as an off target peptide leading to unwanted side effects when administered to a patient. Example 2

The target MHC-presented peptide used in this example is derived from human cancer testis antigen NY-ESO-1 and has the following amino acid sequence; SLLMWITQC. NY-ESO-1 peptide is presented on antigen presenting cells in the context of HLA-A*02.

The binding peptide used in this example comprises a modified T cell receptor (TCR) fusion protein which has been engineered to possess enhanced affinity for NY-ESO-1 peptide, and is termed ImmTAC-NYESO in this example. Production of such fusion proteins is described in McCormack et a I, 2013, Cancer Immunol I mmunother, 62 (4):773-85 and Liddy et al, 2012, Nat Med, 8:980-987. 2.1 Identification of the binding motif by substitution with alanine

Variants of the native NY-ESO-1 peptide were obtained in which each amino acid position was sequentially replaced with alanine, as shown below (in each case the alanine

substitution is underlined). Peptides were obtained from Peptide Protein Research Limited, UK.

A L L M W I T Q C

S A L M W I T Q C

S L A M W I T Q C S L L A W I T Q C

S L L M A I T Q C

S L L M W A T Q C

S L L M W I A Q C

S L L M W I T A C S L L M W I T Q A

The native and alanine-substituted peptides were pulsed on to antigen presenting cells, and interferon γ (IFNv) production, as measured using the ELISpot assay, used as a read-out for T cell activation. Essential positions were defined by a greater than 50% reduction in T cell activity relative to the native peptide.

ELISpot assays were carried as described in Example 1 section 1.1b except that HLA-A2+ T2 cells were used as target cells and effector PBMCs were plated at 40,000 cell per well.

ImmTAC-NYESO was added to a final concentration of 0.1 nM per well.

Figure 11 shows IFNv production by ImmTAC-NYESO redirected T cells in response to native (wt) NY-ESO-1 peptide and each alanine-substituted peptide. Five of the alanine-substituted peptides resulted in a greater than 50 % decrease in IFNv production compared to native NY- ESO-1 peptide. The corresponding native residue at each of these five positions comprises the binding motif. In this case the binding motif is defined as XLXMWITQX, where X is any amino acid. 2.2 Identification of potential off-target peptides

The ScanProsite tool (http://prosite,expasy,org/seanprosjte) was used to search the

UniProtKB/Swiss-Prot database (release date 13 ^th Nov 2013) for proteins which contain the motif identified above (entered as X L X M W I T Q X). The search was limited to human sequences.

Two unique human proteins were identified, NY-ESO-1 (accession no: P78358) and LAGE-IA (accession no: 075638-2) an alternative cancer testis antigen possessing the same 9 amino acid sequence as the NY-ESO-1 peptide. TCRs that bind to the NY-ESO-1 peptide are known to recognise cancer cells expressing LAGE-A1 (McCormack et al, 2013, Cancer Immunol Immunother, 62 (4):773-85). 2.3 Identification of the binding motif by substitution with all alternative amino acids

Variants of the native NY-ESO peptide were obtained in which the amino acid residue at each position was sequentially replaced with all 19 alternative naturally-occurring amino acid, such that 171 peptides were prepared in total. The native and amino-acid substituted peptides were pulsed on to antigen presenting cells, and interferon γ (IFNy) production, as measured using the ELISpot assay, used as a read-out for T cell activation. Essential positions were defined by a greater than 50% reduction in T cell activity relative to the native peptide. ELISpot assays were carried as described in section 2.1a.

Figure 12 shows IFNy production by ImmTAC-NYESO redirected T cells in response to native (wt) NY-ESO-1 peptide and each amino-acid-substituted peptide. The tolerated residues at each position of the peptide are shown below. For each position the first tolerated residue shown corresponds to the native (WT) residue at that position.

2.4 Identification of potential off-target peptides

The ScanProsite tool was used as described in section 2.2, to search for proteins containing the tolerated residues at the indicated positions (entered as [SVTYMFAGHNQKPRWLI]-[LI]- [LVAMICQNH]-[MQVTN]-W-[ITLMS]-[TSD]-[QG]-[CLVMSTGAI])

Two unique proteins were found (exincluding NY-ESO and LAGE Al).

Peptides from Plexin-Dl and protein 5B1 are tested in vitro as described in Example 1.3, to confirm recognition by ImmTAC-NYESO. The potential of these off-targets peptides to cause unwanted side effects in vivo is determined by, for example, reference to literature sources detailing expression of Plexin-Dl and protein 5B1 in normal tissue. Where expression in normal tissue is widespread, or in critical tissues, binding of ImmTAC-NYESO may be additionally be confirmed in vitro using cells which express the off target peptide. Comparative Example 2 a) Similarity search

A similarity search was performed on the WT sequence of the NY-ESO-1 peptide using the BLAST tool as described in Comparative Example la. The top 100 hits did not return either of the two motif containing peptides identified above. b) Comparison with peptides from homologous proteins in mouse

Sequence alignment of the full length human Plexin Dl and 5B1 proteins with the homologous protein from mouse (accession no: NP_080652.2 and NP_083677.1,

respectively) was carried out using the align function on the Uniprot knowledgebase

(http://www.uniprot.org/aiign). Comparison of the respective peptide sequences is shown below:-

Neither of the mouse peptides contains the motif identified in section 2.3, therefore a mouse model may not identify potential cross reactivity.