WO2003016905A2

WO2003016905A2 - Determination of t-cell/mayor histocompatibility complex (mhc) reactivity

Info

Publication number: WO2003016905A2
Application number: PCT/GB2002/003753
Authority: WO
Inventors: Bent Karsten Jakobsen
Original assignee: Avidex Ltd
Current assignee: Avidex Ltd
Priority date: 2001-08-16
Filing date: 2002-08-15
Publication date: 2003-02-27
Anticipated expiration: 2004-02-16
Also published as: WO2003016905A3; GB0120042D0; AU2002319547A1

Abstract

The present invention provides a method for determining whether a T cell reacts with a predetermined Major Histocompatibility Complex (MHC) type. In the method, a sample comprising said T cell is brought into contat with a plurality of molecules of said MHC type, each MHC molecule being complexed with a peptide antigen whose contribution to a T cell receptor binding the MHC-peptide antigen complex is minimised. Whether said plurality of MHC molecules causes activation of the T cell is then determined. The method can assess allo-specific T cell activity, and may be used in matching transplant and donor patients, as well as in monitoring alloreactive responses following a transplant operation.

Description

METHODS

The present invention relates to methods for assessing allo-specific T cell activity. Such methods find particular use in matching transplant and donor patients, as well as in monitoring alloreactive responses following a transplant operation.

Transplantation of organs and other tissues is gaining increasing significance as a treatment for a range of human diseases. The major problem with this type of therapy is the intrinsic tendency of the host's immune system to reject the foreign tissue. A further complication can arise through immune responses from the transplanted tissue towards the host ("graft versus host disease" or GNHD). In order to minimise the risks of transplant rejection and graft-versus-host disease, two precautionary measures are currently being recommended as standard practice. The first is to tissue-type the donor and host, and to use this information to ensure the best possible "match" in order to minimise immune crossreactivity. The second is to treat the host, often extensively, with immunosuppressive drugs.

Tissue-typing, i.e., classification of a patient's Human Leukocyte Antigen (HLA) type, has provided a major advance in ensuring a higher frequency of successful transplant treatments, particularly in kidney and bone marrow transplantation. In principle, the better the match between the HLA types of the donor and host, the less destructive the immune-crossreactivity should be. Consequently, the risks of rejection or graft- versus-host disease are lower. Equally, development of immune-suppressing medication has provided tools with which to exercise some degree of control over the strength of crossreactive immune responses.

Clinical experience in transplantation

As mentioned above, graft rejection owing to alloreactivity can be controlled by powerful but dangerous immunosuppressive drugs. Because the risk of rejection is greatest in the early post-transplant period, maximum immunosuppression is given at this time. Protocols vary, but most regimens incorporate high-dose steroids, high-dose calcineurin inhibitors (cyclosporin or tacrolimus) and an anti-proliferative agent (azathioprine or mycophenolate mofetil). Anti-T cell antibodies are also used in some situations. The most widely used preparations are Muromonab CD3 (OKT3) and anti- thymocyte globulin/anti-lymphocyte globulin (ATG/ALG).

Maintenance therapy usually consists of the same three classes of drugs used at lower doses: a calcineurin inhibitor, an anti-proliferative agent and glucocorticoids. A number of new drugs are undergoing clinical evaluation, but equally important is the determination of the safest and most effective combinations of existing drugs. The relatively wide choice of potent immunosuppressive drugs allows flexibility in minimising side effects for the individual patient, but the side effects of long-term immunosuppression still remain a challenging clinical problem.

The arguments for and against minimising HLA mismatching in clinical transplantation have been debated for the past 30 years. Although it is not necessary to use an HLA genetically identical donor to have a successful transplant, it is equally clear that the likelihood of complications increases with each increment in histoincompatibility. The implication is that histocompatibility testing must be of the highest possible precision to choose an optimal donor, and to predict the risk of adverse alloreactivity. Global statistics support the view that cumulative mismatches in HLA are associated with poorer graft survival. However, such studies also suggest that mismatches at different HLA loci vary in potency. The strength of the mismatch^" seems to increase from HLA-A (the weakest), through HLA-B to HLA-DR (the most potent) (Bradley, Immunol Lett (1991) 29: 55-9). Despite the accumulating data that HLA matching correlates with graft survival rate for a number of transplant types, donor tissues or organs are generally only allocated on the basis of HLA type in kidney and bone marrow transplantation. HLA typing of donor and host for transplant operations

Soon after the first solid organ transplants were performed, it became apparent that some degree of HLA typing was prudent. This typing was performed with a panel of antisera specific for different antigenic products of the system in an assay that became known as the complement-dependent cytotoxicity (CDC) test (Mittal, et al, Transplantation (1968) 6: 913-27), or serology. CDC is robust and can be performed in three hours. However, there are drawbacks: viable lymphocytes are required, the antibodies required are generally non-renewable and the technique has limited powers of resolution, particularly for Class II HLA. Despite the lack of standardisation and simplicity, the CDC crossmatch assay remains a prerequisite in kidney transplantation.

The problem of discrimination of Class π antigens, which are arguably the most important antigens for solid organ transplantation, led to the development of a Class Et typing system based on restriction fragment length polymorphism (RFLP) (Bidwell, et al, Transplantation (1988) 45: 640-6). Although this method could identify serologically defined HLA-DR and -DQ alleles, it had a major problem in that it took 10 days from sampling to obtaining the result.

The development of PCR allowed the evolution of improved molecular HLA-typing techniques and extensive sequencing of HLA alleles followed. Transplantation laboratories were able to use PCR to amplify the polymorphic regions of HLA genes that could subsequently be analysed for the polymorphism within the amplicon, thus establishing tissue-type. Methods can be tailored to identify broad specificities (low resolution), serological specificities (medium resolution) or to discriminate between more than 90% of the alleles in the loci analysed (high resolution). In solid organ transplantation, medium resolution typing is usually carried out, whereas in unrelated bone marrow transplantation, low or medium resolution typing is always followed up with high resolution typing. Techniques for analysing the polymorphisms in amplified DNA can be divided into two basic groups: probe hybridisation and direct amplicon analysis. Current probe- based techniques generally require lengthy post-PCR steps which mean that the fastest method cannot be completed in less than five hours. However, there are other techniques which are suitable for limited numbers of samples that can be carried out more quickly. These include PCR-RFLP, PCR using sequence-specific primers, heteroduplex analysis and other conformational assays (Bunce, et al, Transplantation (1997) 64: 1505-13) .

Effect of HLA mismatching on allograft survival

HLA-matched renal allografts have a superior graft survival compared with HLA-mismatched transplants. However, an increasing number of studies indicate that not every HLA mismatch has the same detrimental effect on survival. HLA-DR mismatches seem to have a strong influence on survival, especially in the first six months after transplantation; this is not surprising as the HLA-DR molecules are mainly expressed on dendritic cells in the transplanted organ. The donor dendritic cells are eventually replaced by recipient-derived cells.

HLA-B mismatches have a more profound effect on renal allograft survival than HLA-A mismatches. However, differential effects of mismatches on graft survival have been reported at the level of individual alleles witiiin a locus. Maruya et al (Maraya, et al, In Clinical Transplants 1993 (eds PI Terasaki and JM Cecka), (1994) p511. UCLA Tissue Typing Laboratory, Los Angeles) were able to define a number of specific HLA mismatches that gave a graft survival similar to HLA-identical combinations. These mismatches have been called "permissible mismatches". Other studies have shown evidence that specific donor-recipient mismatches may influence graft survival in a positive or negative way. For example, a group studied by Vereerstraeten et al (Nereerstraeten, et al, Transplantation (1995) 60: 253-8) showed that certain HLA-DR incompatibilities have a high risk and others have a low risk for graft rejection. Another study focused on the definition of mismatched donor- recipient combinations that had a higher chance of graft rejection, the so-called "taboo" combinations. An example of such a taboo mismatch was found to be a mismatch for HLA-B7 which will lead to a significantly worse graft survival in HLA-A1 recipients than in those who are not HLA-A1 positive.

These empirical studies emphasise that not all HLA mismatches give similar outcomes. Some mismatches will not reduce graft survival (permissible or acceptable mismatches) whereas others have a strongly negative effect on graft function or survival (the detrimental or taboo mismatches). The problem with this type of analysis is that the definition of permissible or taboo combinations is based on studies of patient populations. These studies show that graft survival in one group of mismatches is superior to graft survival in another group of mismatches. However, in both the "good" and the "bad" groups, patients are found who reject their graft. It is not yet possible to predict which individual patients will actually reject their grafts. .

Current HLA matching policies are based on the selection of recipients who have the same HLA specificities as the donor; this is called "structural matching". The advent of molecular typing techniques (see above) has led to an enormous increase in the identification of HLA polymorphisms and clearly shows that structural matching of unrelated donor and recipient organs and tissues is practically unachievable. It is also clear that HLA matching policies would be more effective if they involved "functional matching" where the HLA molecules of the donor should be non-stimulatory to the immune system of the recipient.

Reliable in vitro assays that mimic allograft rejection are essential to test whether the HLA profile of a particular donor is acceptable to the immune repertoire of the recipient. Limiting dilution assays have been used to monitor the frequency of alloreactive T lymphocytes (Deacock, et al, J Immunol Methods (1992) 147: 83-92; Sharrock, et al, Immunol Today (1990) 11: 281-6). In bone marrow transplantation, several publications suggest that the number of recipient-specific CTL precursors in the donor is related to the degree of graft-versus-host disease occurring after transplantation (Kaminski, et al, Transplantation (1989) 48: 608-13; Roosnek, et al, Transplantation (1993) 56: 691-6).

The avidity of the T cells recognising donor antigen also needs to be taken into consideration; the presence of CTL that do not require CD8 molecules to stabilise then- binding (i.e., high avidity) is significantly associated with graft rejection (Roelen, et. al, Transplantation (1995) 59: 1039-42). In contrast, low avidity CTL are found mainly amongst graft-infiltrating cells in the absence of rejection. Similar data have been obtained with respect to HLA class II allospecific CTL; during rejection these CTL are resistant to treatment with anti-CD4 antibodies, while in the absence of rejection CD4+ graft-infiltrating CTL can be blocked by CD4-specific antibodies (van Emmerik, et al, J Heart Lung Transplant (1997) 16: 240-9).

Other studies have shown that discrimination between primed and naϊve CTL is also important. Naϊve CTL require re-stimulation with donor antigen in vitro before they express their cytolytic function whereas primed CTL are already cytotoxic for donor cells without re-stimulation in vitro. A significant association has been found between the presence of primed CTL and graft rejection (Naessen, et al, Clin Exp Immunol (1992) 88: 213-9).

These in vitro studies help to define, in the individual patient, which structural mismatches will be accepted by the patient's immune system. The results of a small pilot study in patients who were transplanted across a historical positive crossmatch show that these assays may well be relevant in this context. Those patients who experienced acute rejection of the graft showed a primed CTL repertoire against the donor mismatch on the day of transplantation. Four patients with a naϊve T cell repertoire in vitro experienced no graft rejection. Matching requirements are different for the different types of organs transplanted. It is clear that bone marrow transplantation demands a high degree of matching when structural mismatches, as defined by molecular typing, are very relevant. Nevertheless, individual patients have been transplanted with a mismatch for HLA-A and HLA-B molecules in the presence of a negative CTL precursor assay, showing that these in vitro assays are suitable for the definition of a functional match, even in the case of a very sensitive clinical model such as bone marrow transplantation. It is likely that functional matches in cases of structural mismatches occur more often in kidney and heart transplantation compared with bone marrow.

Although the currently available methods for detecting and quantifying CTL could be used to investigate the immune reaction between individual donors and potential recipients, the methodology is time-consuming and not available in a format in which a number of potential donors could be screened. It is therefore desirable to provide a method whereby the reactivity of T cells from a potential recipient for a range of HLA types could be predetermined in a functional assay such that the readout from this assay would identify the best matched donor type and minimise the risk of rejection following transplantation.

Herpes viruses and transplant rejection

Herpes viruses frequently cause serious complications after allogenic bone marrow and organ transplants. In particular, Epstein Barr virus (EBV) is associated with post- transplant lymphoproliferative disease (PTLD) in immunocompromised patients. PTLD is a spectrum of mainly B cell diseases that range from polyclonal lymphoproliferative disorders which resolve when immunosuppression is halted to highly malignant lymphomas. It is believed that the underlying immunosuppression inhibits the virus-specific CTL response that would normally control viral replication and maintain latency (Khatri, et al, J Immunol (1999) 163: 500-6). The importance of EBN matching between recipient and donor is unclear, although it has been suggested that PTLD does not occur with any greater frequency in mismatched patients (Harwood, et al, Pediatr Transplant (1999) 3: 100-3). Early diagnosis of EBN- associated PTLD is important because many patients respond to a reduction in immunosuppression, and several groups are developing assays to detect EBN markers (e.g., DΝA) in serum. The development of an assay to monitor the risk of graft rejection through an alloreactive response and therefore reduce the need for immunosuppression may also help to reduce the incidence of EBN-associated PTLD.

There are also several lines of evidence that viral infections, particularly with the herpes virus cytomegalovirus (CMN), play a role in the pathogenesis of solid organ allograft rejection. A diagnostic feature of acute rejection is infiltration of the allograft parenchyma by lymphocytes, a process regulated by the induction of adhesion molecules on vascular endothelial cells and their ligands on leucocytes. Data derived from biopsies of CMN-infected transplant recipients, as well as from experimental models of transplantation, indicate that CMN infection can result in an upregulation of such adhesion molecules, thereby facilitating the inflammatory process. Infection with

CMN is also associated with an increased expression of MHC class II on multiple cell types. An upregulation of these molecules could also contribute to graft failure. A diagnostic tool for detecting the vigour of the immune response between a recipient and potential donors may predict the likelihood of rejection in CMN positive recipients.

Despite the advances made, transplant treatment is still a dangerous and far from ideal treatment, and a clear and urgent need exists for more sophisticated methodology with which to predict the risk of rejection. Equally important is the ability to monitor immune crossreactivity following transplant operations. HLA typing has provided a major breakthrough in assessing how to match donor and host so as to reduce the risk of immune cross-reactivity. However, perfect HLA matching is virtually impossible to achieve. Furthermore, HLA typing only provides an indirect prediction of the likely level of crossreactivity that can be expected between host and transplant tissue. The critical issue is how vigorous the immune responses to the particular non-matched HLA types will be. Immune responses to different HLA types could vary substantially, and therefore a high degree of variability in transplant rejection is observed even when HLA matching is applied.

T cells and their T cell receptors are normally restricted by self MHC molecules and will only respond when presented with specific peptides. In recent years, substantial effort has gone into developing methods that allow the identification or targeting of the subset of T cells that are specific for a particular peptide antigen. Contrary to this prior effort, the present inventors have provided a method that allows the identification or targeting of T cells which are specific for particular MHC (in humans HLA) and not for a particular peptide antigen. This approach has particular use in detecting or targeting T cells that display alloreactivity as is observed in the rejection of tissue and organ transplants.

According to a first aspect of the present invention, there is provided a method for determining whether a T cell reacts with a predetermined Major Histocompatibility Complex (MHC) type, the method comprising: bringing a sample comprising said T cell into contact with a plurality of molecules of said MHC type, each MHC molecule being complexed with a peptide antigen whose contribution to a T cell receptor binding the MHC-peptide antigen complex is minimised; and determining whether said plurality of MHC molecules causes activation of the T cell.

The method of the present invention can be used to detect an alloreactive T cell restricted by the particular MHC type. Such a T cell may then form the basis of a therapy against the alloreaction. However, the method of the present invention finds particular use in determining the reactivity of an individual towards each of a plurality of Major Histocompatibility Complex (MHC) types (i.e. the ability of the individual to stimulate a T cell reaction against a particular MHC). Thus, the method of the present invention has particular utility in matching transplant and donor patients, as well as monitoring alloreactive responses following a transplant operation.

The method of the present invention is preferably used to determine the reactivity of a human individual. Thus, the MHC types are preferably Human Leukocyte Antigen (HLA) types. Because in most - but not all - instances, the method of the present invention is for use in determining the reactivity of a human, for example in an organ transplant procedure, for convenience, reference will be made in the following to HLA. However, it is to be understood that such references are intended to include MHC types of non-human animals.

In certain embodiments, the method of the present invention provides a convenient method for directly assessing, from patient T cell sample, e.g. a small blood sample, the relative levels of immune cross-reactivity towards a range of different HLA types. The ability to obtain a direct assessment of functional donor and host immune reactivity towards a range of HLA types constitutes a fundamental improvement over the simple genetic HLA matching of the prior art. With such information available, it is possible to combine donor and host so as to minimise the actual immune cross- reactivity, rather than merely minimising mis-matched genotypes. In turn, this can lead to a reduced requirement for treatment with immunosuppressive drugs following the transplant operation, thereby reducing the often severe side-effects associated with this type of therapy. Furthermore, following the transplant operation, the assay can be performed again at regular intervals, or when required, in order to follow the development in allo-specific T cell activity directed against the transplanted tissue. Thus the assay can be used as an indicator for decisions on the level of immuno- suppressor therapy to be employed at this stage in order to avoid rejection. Furthermore, the method of the present invention finds use in matching a transplant patient and a xeno-transplant organ, and for monitoring the progress of such a transplant. Recent data suggests that T cell alloreactivity is less dependent on antigen peptide than is the case for conventional, antigen-primed, and self-HLA type restricted, immune responses. By implication, activation of alloreactive T cell responses is therefore more dependent on interactions between T cell receptors (TCRs) and the HLA heavy chain component of the antigen complex, than on TCR-peptide interactions.

T cells mature in the thymus where they undergo at least two selection mechanisms, generally referred to as positive and negative selection. The structures of most, or all, TCRs are believed to share certain general architectural features (Chothia, et al, Embo J (1988) 7: 3745-55) that provide a framework suitable for MHC/peptide binding by the variable complementarity determining regions (CDRs). Thus, most TCRs may have intrinsic affimty for MHC/peptide complexes (Chothia, et al, Embo J(1988) 7: 3745-55). In the thymus, only TCRs with a certain mimmal level of affinity for one of the MHC molecules to which they are presented (the "self MHC molecules) will be positively selected. T cells with high affimty for one of the self MHC molecules will be negatively selected (Amsen & Kruisbeek. (1998). Immunol Rev 165: 209-29. Sebzda, et al (1999). Annu Rev Immunol 17: 829-74). It is unclear to what extent the peptide ligands presented influence these selection procedures.

Mature circulating T cells are believed to be able to recognise a substantial number of different peptide antigens presented by the same MHC molecule. A recent review (Mason & Powrie, Curr Opin Immunol (1998) 10: 649-55) estimates that an individual TCR may be able to bind up to IO⁶ different peptides. However, as each MHC molecule has an estimated capacity to bind ~10¹⁰ different peptides, each T cell will still be highly selective, responding only to 0.01% of the possible peptide antigens presented by the particular MHC molecule. The promiscuity displayed by T cells appears to be even more pronounced when exposed to cells presenting HLA molecules expressed from different alleles than those on which they were selected in the thymus. T cells exposed to cells expressing foreign MHC alleles will display "allo-reactivity", a phenomenon that is usually observed when an organ or tissue is transplanted into a heterologous host. The T cell alloreactivity is usually vigorous, leading to destruction (rejection) of the foreign cells.

The mechanisms involved in alloreactive T cell responses are not fully understood. However, a number of features indicate that peptide antigens play a substantially less prominent role in activation than is the case for serf-MHC restricted T cell activation. First, alloreactivity is mediated by an extensive diversity of T cells, and it is assumed that both "naϊve" T cells and T cells that have been previously activated by self-MHC restricted antigens are involved in the allo-response (Bill, et al. (1989). JExp Med 169: 115-33; Garman, et al. (1986). [published erratum appears in Proc Natl Acad Sci

USA 1986 Sep;83(18):7028]. Proc Natl Acad Sci USA 83: 3987-91; Lauzurica, et al. (1992). J Immunol 148: 3624-30). The heterogeneity of allo-responses and the fact that no particular antigens are required for their activation may indicate that more T cells are allo-reactive than are self-MHC/antigen reactive.

Secondly, the fact that so many T cells display reactivity to non-self MHC molecules indicate that many of the allo-reactive T cells would have been eliminated by negative selection in the thymus, had they undergone selection in an individual expressing the allo-restrictive MHC molecule. In other words, allo-reactive T cells may have TCRs with high affimty for the allo-restrictive MHC molecule. A good example of this has been observed with a dominant T cell response found in HLA-B8 patients who have been exposed to Epstein-Barr vims (EBN) infection (Burrows, et al, JExp Med (1994) 179: 1155-61). The majority of patients who are HLA-B44 negative have T cell alloresponses to HLA-B44 that coincide with their HLA-B8/EBN responses. In most patients, this is mediated by an immuno-dominant, "public", T cell clone (Burrows, et al, JExp Med (1994) 179: 1155-61). However, this clone is not observed in patients that are HLA-B44 positive, indicating that, in these patients, the public T cell clone is eliminated by negative selection. Instead, HLA-B44 positive patients display a variety of T cell responses to HLA-B8 presenting EBN antigen (Burrows, et al, EurJ Immunol (1997) 27: 1726-36). One response characterised from an HLA-B 8 and HLA-B44 positive patient was found to be alloreactive to HLA-B14 and HLA-B35, neither of which were expressed by that individual (Burrows, et al, Eur J Immunol (1997) 27: 1726-36). Thus, it seems likely that many allo-reactive T cells are better equipped to respond to certain non-self MHC molecules than to self-MHC molecules, making them less dependent on peptide antigen for activation.

Thirdly, it has been demonstrated that peptide antigens mediating allo-reactivity are relatively easy to identify. A T cell clone specific for EBN antigen restricted by HLA-B8 was found to be alloreactive for HLA-B35. Screening of the human protein sequence database (SwissProt) was used in a study to search for putative peptide sequences that would act as antigens when presented by HLA-B35. A previously defined fine specificity of the T cell clone for peptides presented by HLA-B8 was used to search for peptides with suitable HLA-B35 anchor residues. Strikingly, of 37 peptides tested, two acted as strong antigens for the T cell clone when presented by HLA-B35 (Burrows, et al, Eur J Immunol (1997) 27: 1726-36).

The method of the present invention is intended to optimise the detection of T cells with the ability to recognise certain HLA types. The contribution of the peptide antigen to a T cell receptor binding the MHC-peptide antigen complex may be minimised by complexing the MHC molecules with one of (a) a peptide antigen which presents substantially no T cell recognition features, and (b) a peptide antigen in which one or more T cell recognition features are randomly present.

In one aspect, the peptide antigens in the HLA molecules used in the method of the present invention are designed not to contribute to the T cell binding the HLA. Thus, the different types of HLA/peptide complexes may be provided in which the "normal" antigenic peptides have been substituted for "null" peptides which provide the absolute minimum of features for TCR recognition. Such peptides preferably comprise alanine residues except at amino acid positions which are required to bind the peptide to the HLA heavy chain (so-called "anchor residues" ). The null peptides can therefore form a complex with the HLA, but otherwise the peptide has no amino acid side chains (other than hydrogen) which could contribute to T cell binding. Instead of, or in addition to, alanine residues, null peptides may have glycine and/or serine residues to provide the minimum of T cell recognition features.

Alternatively, in another aspect, the "normal" peptides can be substituted for peptides from designed peptide libraries in which certain positions are randomised so as to make it more likely that antigen complexes are included that stimulate a larger proportion of T cells. Because of the high degree of variation in these peptide antigens, T cell recognition will not be dominated by a subset of the T cells which are specific for one particular antigen. Instead, the observed T cell response will be dominated by the bias towards the particular HLA molecule presenting the redundant antigens. Where only one or a few positions are randomised in the peptide antigen, the T cell response will be dominated by those cells which can react to the HLA molecule supported by minimal features by the peptide. Where all positions in the peptide antigen are randomised, the T cell response will be dominated by those cells which react to the HLA molecule supported by a variety of features supplied by the peptides, again biasing the response in favour of those T cells that rely mainly on contacts to the HLA heavy chain for activation.

The principle of using null peptides or designed peptide libraries containing randomised residues is to enhance the importance of TCR-HLA heavy chain interactions for the assay readout, rather than make this depend on the detection of a specific peptide antigen by particular T cell responses. Thus, the invention is based on a principle whereby HLA/peptide complexes are not used for the detection of peptide- specific, that is particular antigen-primed and self-MHC restricted, T cells, but rather that they are used for assessing the strength of alloreactive T cell responses towards individual HLA type molecules. Conveniently, each of the plurality of HLA molecules may be provided in an individual reaction chamber, e.g. immobilised in an individual well of an assay tray, allowing a sample of T cells to be incubated and react individually with each HLA type complex. Methods for the production of and immobilisation of HLA molecules are described in more detail below. The sample of the individual's T cells may be provided by a sample of peripheral blood leucocytes (PBLs) such that the method of the present invention detects all those T cells within the sample which are alloreactive to the provided HLA molecules.

When used for determining the reactivity of an individual towards each of a plurality of Major Histocompatibility Complex (MHC) types, the more different HLA types that are used for the method of the invention, the more informative and versatile it will be. However, the most common HLA types are preferably used as, in practical terms, it is virtually impossible to provide all known HLA molecule types and subtypes.- A further practical consideration is whether methodology for production of a specific

HLA molecule as a soluble complexes exists or can be developed.

The haploid genome contains three loci coding for type A, type B, and type C respectively of HLA class I molecules. A large number of alleles have been identified for each locus and any individual will express three different HLAs per haploid content equalling up to six different molecules on each cell.

Previously, the HLA nomenclature was based on serology, and the designation HLA-A2 describes the HLA-A molecules sharing the A2-characteristic serologic determinants (f. ex. HLA-A0201 - HLA-A02012). With the increasing number of known gene sequences, an HLA nomenclature which is based on sequence information has been introduced. According to these mles, a number is assigned to the alleles of the loci (f.ex. HLA-A* A0201) as their sequences become known. The internet provides some helpful sites that incorporate the HLAs as their sequences appear, along with information regarding their distribution in the population. The IMGT/HLA Database flιttp://www.ebi.ac.uk/imgt/hla is part of the international ImMunoGeneTics (EVIGT) project and provides a specialist sequence databases for sequences of the human major histocompatibiUty complex (HLA). This includes all official sequences for the WHO HLA Nomenclature Committee For Factors of the HLA System. Development of this database has been undertaken by James Robinson, Julia G. Bodmer and Steven G.E. Marsh. Two sites providing programmes for predicting peptide motifs in any given protein, authentic or synthetic have been opened. Both sites are supervised by acknowledged experts. In Germany, the

SYFPEITHI-site is provided by Dr. Hans-Georg Rammensee's group (Hans-Georg Rammensee, Jutta Bachmann, Niels Emmerich, Stefan Stevanovic: SYFPEITHI: An Internet Database for MHC Ligands and Peptide Motifs (access via : http://www.uni-tuebingen.de/uni/kxi ). This site is based on published reviews (Rammensee, et al, Immunogenetics, 1995, 41:178-228; Rammensee, et al: MHC ligands and peptide motifs. Landes Bioscience 1997).

In the USA, Dr. Kenneth Parker of the National Institute of Allergy and Infectious Diseases (NIAID) at the National Institutes of Health (NTH) in Befhesda, Maryland supervises the "HLA Peptide Binding Predictions"-site at http://bimas.dcrt.nih.gov/molbio/hla_bind/. This Web site allows users to locate and rank 8-mer, 9-mer, or 10-mer peptides that contain peptide-binding motifs for HLA class I molecules. These rankings employ amino acid/position coefficient tables deduced from the literature by Dr. Kenneth Parker. The Web site was created by Ronald Taylor of the Bioinformatics and Molecular Analysis Section (BIMAS), Computational Bioscience and Engineering Laboratory (CBEL), Division of Computer Research & Technology (CIT), National Institutes of Health, in collaboration with Dr. Parker. Such sites allow peptide antigens to be designed which have the required "anchor" residues, but which otherwise are either "null" peptides or are randomised.

Table I lists known HLA class I molecules, and is compiled from the "BHVIAS-site". Included is a list of semi-optimised synthetic poly-alanine peptide sequences calculated on the basis of the "HLA Coefficient Tables" used by the peptide prediction programme (Peptides marked by ♦ were designed on the basis of information in the indicated literature since no HLA Coefficient tables were available for these HLA types) and the references cited by the BIMAS-site for each HLA molecule.

Table I

Heavy chains References Optimal peptide

HLA-A*0101 DiBrino, et al. J Immunol. 1994; 152 (2): 620-31. AAEPAAAAY Kubo, et al. J Immunol. 1994; 152 (8): 3913-24. Falk, et al. Immunogenetics. 1994; 40 (3): 238-41. Sette, et al. Mol Immunol. 1994; 31 (11): 813-22. Colovai, et al. Tissue Antigens. 1994; 44 (2): 65-72.

HLA-A*0201 Falk, et al. Nature. 1991; 351 (6324): 290-6. ALAAAAAAV Hunt, et al. Science. 1992; 255 (5049): 1261-3. Rotzschke, et al. Eur J Immunol. 1992; 22 (9): 2453-6. Parker, et al. J Immunol. 1992; 149 (11): 3580-7. Ruppert, et al. Cell. 1993; 74 (5): 929-37. Parker, et al. J Immunol. 1994; 152 (1): 163-75. Chen, et al. J Immunol. 1994; 152 (6): 2874-81. Sette, et al. Mol Immunol. 1994; 31 (11): 813-22. del Guercio, et al. J Immunol. 1995; 154 (2): 685-93. Barouch, et al. J Exp Med. 1995; 182 (6): 1847-56.

HLA-A*0202 del Guercio, et al. J Immunol. 1995; 154 (2): 685-93. ALAAAAAAV* Barouch, et al. J Exp Med. 1995; 182 (6): 1847-56.

HLA-A*0204 Stryhn, et al. Eur J Immunol. 1996; 26 (8): 1911-8. ALAAAAAAV*

HLA-A*0205 del Guercio, et al. J Immunol. 1995 154 (2): 685-93. AVAAAAAAL Barouch, et al. J Exp Med. 1995; 182 (6): 1847-56.

HLA-A*0206 del Guercio, et al. J Immunol. 1995 154 (2): 685-93. ALAAAAAAV*

HLA-A*0214 Barouch, et al. J Exp Med. 1995; 182 (6): 1847-56. AVAAAAAAL*

HLA-A*0301 DiBrino, et al. PNAS U S A. 1993; 90 (4): 1508-12. ALAAAAAAK Kubo, et al. J Immunol.; 152 (8): 3913-24. Sette, et al. Mol Immunol. 1994; 31 (11): 813-22. Barouch, et al. J Exp Med. 1995; 182 (6): 1847-56.

HLA-A* 11 Zhang, et al. PNAS U S A. 1993; 90 (6): 2217-21. AVAAAAAAK Kubo, et al. J Immunol. 1994; 152 (8): 3913-24. Falk, et al. Immunogenetics. 1994; 40 (3): 238-41. Sette, et al. Mol Immunol. 1994; 31 (11): 813-22.

HLA-A*24 Kubo, et al. J Immunol. 1994; 152 (8): 3913-24. AYAAAAAAL Sette, et al. Mol Immunol. 1994; 31 (11): 813-22. HLA-A*2402 Ibe, et al. Immunogenetics. 1996; 44 (4): 233-41. AYAAAAAAW*

HLA-A*29 Boisgerault, et al. PNAS U S A. 1996; 93 (8): 3466-70 AEAAAAAAY*

HLA-A*3101 Falk, et al. Immunogenetics. 1994; 40 (3): 238-41. AVAAAAAAR Colovai, et al. Tissue Antigens. 1994; 44 (2): 65-72.

HLA-A*3302 Falk, et al. Immunogenetics. 1994; 40 (3): 238-41. AVAAAAAAR

HLA-A*68.1 Guo, at al. Nature. 1992; 360 (6402): 364-6. AVAAAAAAR

HLA-A*6901 Baroush, at al. J Exp Med. 1995; 182 (6): 1847-56. AVAAAAAAL*

HLA-B*0701 Huczko, et al. JImmunol. 1993; 151 (5): 2572-87. APAAAAAAL* Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

HLA-B*0801 Sutton, et al. Eur J Immunol. 1993; 23 (2): 447-53. AARARAAAL Malcherek, et al. Int Immunol. 199; 5 (10): 1229-37. DiBrino, et al. J Immunol. 1994; 152 (2): 620-31. Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

HLA-B*1402 DiBrino, et al. J Biol Chem. 1994; 269 (51): 32426-34. ARAARAAAL

HLA-B*2702 Rotzschke, et al. Immunogenetics. 1994; 39 (1): 74-7. ARAAAAAAY

HLA-B*2705 Jardetzky, et al. Nature. 1991; 353 (6342): 326-9. ARAAAAAAL Rotzschke, et al. rmmunogenetics. 1994; 39 (1): 74-7. Parker, et al. Biochemistry. 1994; 33 (24): 7736-43. Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

HLA-B3501 Hill, et al. Nature. 1992; 360 (6403): 434-9. APAAAAAAY Falk, et al. Immunogenetics. 1993; 38 (2): 161-2. Takamiya, et al. Int Immunol. 1994; 6 (2): 255-61. Sidney, et al. J Immunol. 1995; 154 (1): 247-59. Takiguchi, et al. Int Immunol. 1994; 6 (9): 1345-52. Schonbach, et al. J Immunol. 1995; 154 (11): 5951-8.

HLA-B3503 Sidney, et al. J Immunol. 1995; 154 (1): 247-59. APAAAAAAL* Steinle, et al. Immunogenetics. 1996; 43 (1-2): 105-7.

HLA-B*3701 Falk, et al. Immunogenetics. 1993; 38 (2): 161-2. ADAAAAAAL

HLA-B*3801 Falk, et al. Immunogenetics. 1995; 41 (2-3): 162-4. AHAAAAAAL Colovai, et al Tissue Antigens. 1994; 44 (2): 65-72.

HLA-B*3901 Falk, et al. rmmunogenetics. 1995; 41 (2-3): 162-4. AHAAAAAAL

HLA-B3902 Falk, et al. Immunogenetics. 1995; 41 (2-3): 162-4. AKAAAAAAL

HLA-B*40 Harris, et al. J Immunol. 1993 Dec l;151(l l):5966-74. AEAAAAAAL

HLA-B*4402 Fleischhauer, et al. Tissue Antigens. 1994; 44 (5): 311-7. AEAAAAAAY*

HLA-B*4403 Fleischhauer, et al. Tissue Antigens. 1994; 44 (5): 311-7. AEAAAAAAY DiBrino, et al. Biochemistry. 1995; 34 (32): 10130-8.

HLA-B5101 Falk, et al. Int Immunol. 1995; 7 (2): 223-8. APAAAAAAI Kikuchi, et al. Immunogenetics. 1996; 43 (5): 268-76.

HLA-B*5102 Falk, et al. Int Immunol. 1995; 7 (2): 223-8. APAAAAAAV Kikuchi, et al. rmmunogenetics. 1996; 43 (5): 268-76.

HLA-B*5103 Falk, et al. Int Immunol. 1995; 7 (2): 223-8. AAAAAAAAV Kikuchi, et al. Immunogenetics. 1996; 43 (5): 268-76.

HLA-B*5201 Falk, et al. Int Immunol. 1995; 7 (2): 223-8. AQAAAAAAV

HLA-B*53 Hill, et al. Nature. 1992; 360 (6403): 434-9. APAAAAAAV*

HLA-B*5401 Sidney, et aϊ. J Immunol. 1995; 154 (1): 247-59. APAAAAAAI*

HLA-B*58 Falk, et al. Immunogenetics. 1995; 41 (2-3): 165-8. ASAAAAAAW Colovai, et al. Tissue Antigens. 1994; 44 (2): 65-72.

HLA-B*60 Falk, et al. Immunogenetics. 1995; 41 (2-3): 165-8. AEAAAAAAL

HLA-B*61 Falk, et al. Immunogenetics. 1995; 41 (2-3): 165-8. AEAAAAAAV Boisgerault, et al. PNAS U S A. 1996; 93 (8): 3466-70.

HLA-B*62 Falk, et al. Immunogenetics. 1995; 41 (2-3): 165-8. ALAAAAAAF HLA-B*7801 Falk, et al. Int Immunol. 1995; 7 (2): 223-8. APAAAAAAL

HLA-C *03 Falk, et al. PNAS U S A. 1993; 90 (24): 12005-9. AALAAAAAL

HLA-Cw*0401 Falk, et al. PNAS U S A. 1993; 90 (24): 12005-9. AFAAAAAAL Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

HLA~C *0602 Falk, et al. PNAS U S A. 1993; 90 (24): 12005-9. AAAAAAAAL Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

HLA-C *0702 Falk, et al. PNAS U S A. 1993; 90 (24): 12005-9. AYAAAAAAY Sidney, et al. J Immunol. 1995; 154 (1): 247-59.

For class II molecules, the LMGT/HLA database is more useful. The database contains information regarding the amino acid and DNA sequences, ethnic distribution, information about the cell line from which the sequence was isolated, links to the databases (Genbank/EMBL/SwissProt) harbouring the original information, and medline links to the relevant literature of the particular entry. A query in the keywords field of this database using the search parameter "class II" returned 472 entries most representing an individual allele of a classical or non- classical HLA class II molecule.

Methods for the production of soluble class I and II MHC/peptide complexes are known. Soluble class I MHC-peptide complexes were first obtained by cleaving the molecules of the surface of antigen presenting cells with papain (Bjorkman, et al, J Mol Biol (1985) 186: 205-10). Although this approach provided material for crystallisation, in recent years it has been replaced for class I molecules by individual expression of heavy and light chains in E. coli, followed by refolding in the presence of synthetic peptide (Gao, et al, Prot. Sci. (1998) 7: 1245-49; Gao, et al, Nature (1997) 387: 630-4; Garboczi, et al, Proc Natl Acad Sci USA (1992) 89: 3429-33 Issn: 0027- 8424; Garboczi, et al, J Mol Biol (1994) 239: 581-7 Issn: 0022-2836; Madden, et al, [published erratum appears in Cell (1994) Jan 28 ;76(2): following 410]. Cell (1993) 75: 693-708 Issn: 0092-8674; Reid, et al, JExp Med (1996) 184: 2279-86; Reid, et al, FEBSLett (1996) 383: 119-23; Smith, et al, Immunity (1996) 4: 215-28 Issn: 1074- 7613; Smith, et al, Immunity (1996) 4: 203-13 Issn: 1074-7613). This approach has several advantages over previous methods; higher yields are obtained at a lower cost, peptide identity can be accurately controlled, and the final product is more homogeneous. Furthermore, expression of modified heavy or light chains, for instance fused to a protein tag, can be easily achieved.

Methods are also available from published literature for the synthesis of class II MHC/peptide molecules (Crawford, et al, Immunity (1998) 8: 675-82; Kozono, et al, Nature (1994) 369: 151-4; Matsui, et al, Science (1991) 254:1788-91; Mottershead, et al, Immunity (1995) 2: 149-54; Smith, et al, J Exp Med (1998) 188: 1511-20). These may be modified to make them soluble and to include biotinylation tag sequences to enable immobilisation on a streptavidin-modified surface. Full length DRB 1*0401 has been expressed on the surface of Drosophila melanogaster Schneider 2 cells under control of a copper sulphate-inducible Drosophila metallothionein promoter (Hansen, et al, Tissue Antigens (1998) 51: 119-28). This approach is readily modified to produce soluble MHC class II molecules simply by expressing a truncated version of the protein which contains a biotinylation tag sequence in place of the transmembrane domain. This protein will be secreted in a soluble form instead of bound to the extracellular surface of the cell membrane.

The α- and β-chains of HLA-DR1 have been expressed in E.coli as inclusion bodies and purified separately under denaturing conditions prior to co-refolding in vitro (Frayser, et al, Protein Expr Purif (1999) 15: 105-14). The synthesised protein was soluble, bound peptide in the expected manner, and was stable.

Increased stability between the α and β chains of soluble class II MHC molecules has been achieved by expressing them as fusion proteins. Membrane domains of the HLA-DR2 molecule α- and β-chains (DRA, DRB 1*1501 genes) have been replaced with leucine zipper domains from c-jun and c-fos (Kalandadze, et al, JBiol Chem (1996) 271 : 20156-62). Expression was achieved in methylrrophic yeast (Pichia pastoris) using the α-mating factor secretion signal to direct expression to the secretory pathway. In contrast to antibody-antigen interactions, interactions between TCRs and MHC/peptide complexes are characterised by low affinities and fast kinetics (Willcox, et al, Immunity (1999) 10: 357-65; Wyer, et al, Immunity (1999) 10: 219-225). The short half-life of the individual binding event between peptide-class I MHC and the

TCR and CD8 receptors makes this interaction unsuitable for use in detection methods (an MHC monomer itself would not stimulate a T cell). In nature, the interactions between T cells and antigen presenting cells are stabilised by multiple simultaneous receptor-ligand contacts, increasing the avidity of cell-cell interactions. Accordingly, in the present invention, it is preferred if the MHC types are provided as multivalent complexes. Such complexes may be in the form of monomers of the MHC coated in close proximity on a surface, e.g. provided in a liposome or attached to a surface, either directly or via biotin to avidin which is attached to the surface or alternatively in the form of MHC multimers. A suitable surface is the multiscreen Immobilon-P (Millipore) 96 well filtration plate which contains a sterile, high protein binding hydrophobic Immobilon-P membrane.

The production of tetrameric molecules of peptide-MHC complexes is described in Airman, et al, Science (1996) 274:94-6. The higher avidity of the multimeric interaction provides a dramatically longer half-life for the molecules binding to a T cell than would be obtained with binding of a monomeric peptide-MHC complex.

The tetrameric peptide-MHC complex may be made with synthetic peptide, β2microglobulin (usually expressed in E.coli), and soluble MHC heavy chain (also expressed in E. coli). The transmembrane domain is truncated from the heavy chain and replaced with a protein tag constituting a recognition sequence for the bacterial enzyme BirA (Barker & Campbell, JMol Biol (1981) 146: 451-67; Barker & Campbell, J Mol Biol (1981) 146: 469-92; SchsAz, Biotechnology N Y (1993) 11: 1138-43). Bir A catalyses the biotinylation of a lysine residue in a somewhat redundant recognition sequence (Schatz, Biotechnology N Y (1993) 11: 1138-43) . However, the specificity is high enough to ensure that the vast majority of protein will be biotinylated only on the specific position on the tag. The biotinylated protein can then be covalently linked to streptavidin which has four binding sites resulting in a tetrameric molecule of peptide-MHC complexes (Airman, et al, Science (1996) 274:94-6).

Multimeric class II MHC/peptide complexes have also been described (Crawford, et al, Immunity (1998) 8: 675-82), and it is becoming well established that both class I and class II restricted T cells can be activated by soluble MHC complexes provided in multimeric formulations.

Recently, soluble MHC/peptide molecules formulated in multimeric forms have been used to detect and activate cells expressing TCRs specific for the peptide ligand (Airman, et al, Science (1996) 274:94-6.; Nessey, et al, Eur JImmunol (1997) 27:

879-885). Indeed, MHC/peptide tetramers are now being used for the detection and quantification of antigen-specific T cells, and the technology holds promise for a multitude of uses in diagnostics and for the development of novel antigen-specific therapies (McMichael & O'Callaghan, J Exp Med (199 ) 187: 1367-71).

In contrast to these prior methods which use of MHC/peptide molecules for the detection of T cells which recognise a specific antigen peptide, the method of the present invention is intended to detect T cells which recognise certain MHC types regardless of the peptide. Thus, the peptide antigens in the MHC molecules used in the method of the present invention are peptides which are designed not to contribute to the T cell binding the MHC. The peptides may be "null" peptides that provide a mimmum of features for TCR recognition. Alternatively, the MHC complexes may contain designed peptide sequences in which one or a few positions are randomised in order to provide minimal features for TCR recognition. Recently a new strategy for expressing class I MHC complexes presenting a particular peptide sequence has been described. The advantage of this strategy is that it avoids the need for synthetic peptide. Instead, the peptide to be presented by the class I MHC complex is expressed as a part of the β2microglobulin polypeptide to which it is fused via a flexible linker sequence. This strategy has been applied to expression on the surface of antigen presenting cells (Uger & Barber, JImmunol (1998) 160: 1598-605) as well as for the generation of soluble class I MHC complexes (White, et al, J Immunol (1999) 162: 2671-6).

Following the incubation of the HLA complexes and T cell sample, whether any of the HLA types causes activation of the individual's T cells is determined. Although any suitable method known to the skilled person could be used, this is preferably by performing a T cell ELISPOT assay. The principle of this assay is that activation of signal transduction induces γ-interferon production and secretion by T cells. By plating the T cells, and incubating with an appropriate antibody; the number of activated cells can be easily counted. Thus, the counts obtained from each plate provide a measure of T cell reactivity to the HLA type molecule used for incubation. By comparing the activated T cell counts obtained by incubation with the different alloreactive and self-specific HLA molecules, a relative measure of crossreactivity is obtained. WO98/23960 describes a method for detecting activation of T cells.

Enzyme-linked immunospot (ELISPOT) assays have been used to determine the frequencies of cytokine-secreting T lymphocytes in peripheral blood from patients with viral, cancerous, and infectious diseases (Hagiwara, et al, Cytokine (1995) 7: 815- 22; Herr, et al, J Immunol Methods (1996) 191: 131-42; Lalvani, et al, JExp Med

(1997) 186: 859-65; Lalvani, et al, Proc Natl Acad Sci USA (1998) 95: 270-5). The ELISPOT assay detects secreted cytokine molecules (IFN-γ, IL-2, IL-4, IL-6, LL-10, TNF-α) in the immediate vicinity of the cell from which they are derived, and release of cytokine(s) represents a measure of effector function. Each spot in the read-out represents a "footprint" of the original activated T lymphocyte in response to antigenic peptide. Therefore, the numbers of spots provide an indication of the frequency of T cells that respond to the antigen which is used in the assay.

Limiting-dilution analysis (LDA) has been widely used to determine the frequency of peptide antigen specific T lymphocytes within lymphocyte populations. The frequencies of antigen specific CD8 T lymphocytes among peripheral blood lymphocytes are usually very low, and LDA effector lymphocytes require repeated stimulation with antigen and addition of cytokines to enable proliferation of antigen specific CD8⁺ T lymphocytes (CTLs) (Lehner, et al, JExp Med (1995) 181 : 79-91). Thus, results obtained by limiting dilution cultures may be influenced by the selectivity of culture conditions. The use of an ELISPOT assay to detect antigen specific T lymphocytes is therefore preferable, since freshly isolated peripheral blood lymphocytes can be used directly.

A number of studies have described the use of ELISPOT assay for quantification of the frequency of antigen specific T cell reactivity in peripheral blood lymphocytes. An ELISPOT assay was used for the detection and quantification of CD8 T lymphocytes isolated from peripheral blood, recognising melanoma peptide antigens presented by HLA- A*0201. CD8 T lymphocytes were isolated from peripheral blood and stimulated for 40 h with HLA-A2*0201 positive T2 cells loaded with melanoma peptide. Tumour necrosis factor α (TNF-α) secreted by activated CD8⁺T lymphocytes in response to melanoma peptide was trapped on nitrocellulose membranes precoated with anti-TNF-α antibodies and was then immunochemically visualised as spots (Herr, et al, J Immunol Methods (1996) 191: 131-42). Another study described the use of the ELISPOT assay to detect CD8⁺ T lymphocytes specific for influenza virus antigenic peptides. Peripheral blood mononuclear cells (PBMC) were isolated from HLA-B*2705, HLA-A*0201, HLA-A1, HLA-A3, HLA-B8, HLA-B35 positive donors and influenza vims antigenic peptides were used to identify frequencies of CD8⁺T lymphocytes by IFN-γ release (Lalvani, et al, JExp Med (1997)

186: 859-65). Detection of peptide specific CD8⁺T lymphocytes from freshly isolated PBMCs is complicated by the fact that the target cells used for peptide presentation elicit responses from T lymphocytes of other specificities. Heterologous B cell lines (BCLs) induce strong responses from alloreactive T lymphocytes, whereas autologous

BCLs result in strong EBV-specific responses. Allo-specific and EBV-specific responses can be circumvented by using the autologous fresh PBMCs themselves to present the antigenic peptide. Peptide-specific CD8⁺ T lymphocytes displayed IFN-γ release within 6 hour of peptide antigen stimulation and the frequency of peptide- specific CD8⁺ T lymphocytes enumerated by the ELISPOT assay was higher than the corresponding CTL precursor frequency as detected by LDA. The ELISPOT assay appears to be a more sensitive assay for detecting peptide-specific CD8⁺ T lymphocytes in peripheral blood.

As mentioned, the soluble HLA complexes will be generated with either 'null' peptides or peptides containing a degree of randomness at selected amino acid, positions. Both of these types of complexes can be generated with synthesised peptide or with the peptide expressed as a fusion with β2microglobulin. In the former case, for class I HLA molecules, the heavy chain, the light chain (i.e. β2m) and the peptide are folded together. For class II HLA molecules, the complex is expressed 'empty' and loaded with peptide subsequently.

In a second aspect, the invention provides a cloning vector, which is preferably DNA, encoding an MHC subunit (preferably human) into which a nucleotide (e.g.synthetic DNA) encoding a peptide antigen sequence of interest can be inserted such that expression of the vector produces a fusion protein comprising the MHC subunit with the peptide antigen fused thereto via a linker sequence. When the MHC is class I, the subunit is preferably β2microglobulin, with the peptide antigen fused at the N- terminus thereof. This vector may have different selective markers for propagation in E.coli,' different restriction enzyme sites for cloning; a different length or composition of the linker sequence between the antigenic peptide sequence and β2microglobulin; or with altered codon usage for expression of β2microglobulin. Furthermore, the peptide sequence to be expressed as a fusion with β2microglobulin can be altered by the polymerase chain reaction (PCR) or by cloning in synthetic DNA oligomers corresponding to the desired peptide sequence. In one such vector, recognition motifs for the restriction enzymes Ndel and BamHI in the DNA sequence facilitate alteration of the peptide sequence which is expressed together with β2microglobulin. When the MHC is class π, the subunit may be the α- or the β-chain, with the peptide antigen fused at the amino terminus thereof. The invention also provides, in a third aspect, a cell transformed with such a vector.

The design of a DNA plasmid vector allowing convenient insertion of DNA oligonucleotides encoding a chosen peptide sequence such that this will be expressed as a fusion with the β2microglobulin polypeptide via a flexible linker sequence is shown in Figure 1. Examples of designs of the DNA cassette for the expression of a HLA-A2 null peptide fused to β2microglobulin are shown in Figure 2. Examples of designs of the DNA cassette for the expression of a HLA-B8 null peptide fused to β2microglobulin are shown in Figure 3. Examples of designs for DNA cassettes suitable for the expression of β2microglobulin-peptide fusion proteins in which the peptide part contains positions with randomised amino acid identities are described in Examples 4-11.

β2microglobulin can be expressed in a number of factory cell lines, e.g., E.coli, Pichia Pastoris and mammalian cell lines. Similarly, peptide- β2microglobulin fusion proteins may be expressed in any of these systems. E.coli may be the most advantageous because it is less likely that protease processing in this organism will remove the fused peptide from the β2microglobulin polypeptide. In fourth aspect, the invention provides a multivalent class I or class U MHC/peptide complex, preferably a multimer, in which the peptide is optimised for assessment of T cell responses that are predominantly mediated through TCR-MHC contacts rather than through TCR-peptide contacts. The peptide may be a 'null' peptide which presents substantially no T cell recognition features . For instance, all amino acids in the peptide, except those residues that bind the peptide in the MHC groove (the 'anchor' residues) may be alanine, glycine and/or serine (such as glycine-serine mixtures) residues. Alternatively, the peptide may be one in which T cell recognition features are randomly present. For example, the peptide may be from a library of peptides in which certain positions are a mixture of selected or random residues. The other positions in these peptide libraries would be occupied by neutral residues, typically alanine, except for the 'anchor' residues which would be chosen as optimal as possible for binding to the respective HLA type molecules. The randomised positions are for providing a small peptide contribution to TCR binding to the HLA complex, so that binding is being dominated by TCR-heavy chain contacts.

The antigen peptide may be covalently linked to one of the subunits in the MHC complex by constituting part of the same polypeptide chain, and preferably forms part of a fusion protein with one of the MHC subunits. For class I HLA complexes, the peptide may be covalently linked to β2-microglobulin via a flexible linker sequence. For class II HLA complexes, the peptide may be covalently linked to the amino terminus of the α or β chain thereof via a flexible linker sequence (Kozono, et al, Nature (1994) 369: 151-4).

The invention also provides, in a fifth aspect, a kit for determining whether a T cell reacts with a predetermined Major Histocompatibility Complex (MHC) type, the kit comprising: a plurality of molecules of said MHC type, each MHC molecule being complexed with a peptide antigen whose contribution to a T cell receptor binding the MHC-peptide antigen complex is minimised.

In a preferred embodiment, the kit is for determining the reactivity of an individual towards each of a plurality of Major Histocompatibility Complex (MHC) types, and comprises a plurality of molecules of said plurality MHC types.

The kit preferably further comprises means for determining whether any of said MHC types causes activation of the individual's T cells. Such means may be the reagents required for an enzyme-linked immunospot assay (ELISPOT). The kit may further comprise instructions allowing a user of the kit to practice the invention.

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.

EXAMPLES

Reference is made in the following examples to the accompanying drawings in which:

Figure 1 shows the DNA sequence of a plasmid, pBJ196 which can be used for expression of a polypeptide which is a fusion of three elements: a peptide corresponding to an antigenic peptide from influenza virus ('flu matrix peptide) which is known to be presented by HLA-A0201 , a flexible linker sequence mainly comprised of Glycine and Serine residues, and β2microglobulin. The amino acid sequence of the resulting fusion protein is also shown. Figure 2 shows the DNA and amino acid sequences of the N-terminal part of the peptide-linker-β2microglobulin fusion polypeptide produced by plasmid pEX013, a modification of pBJ196.

Figure 3 shows the DNA and protein coding sequences of the N-terminal part of the peptide-linker-β2microglobulin fusion polypeptide produced by plasmid pEX014, a modification of pBJ196.

Example 1 -A DNA plasmid vector for expression of peptide- βlmicroglobulin fusions in E.coli.

This example describes the design and construction of a DNA plasmid vector for expression of peptide-β2microglobulin fusions in E.coli. The vector was designed so that the peptide to be expressed can be easily changed by inserting a pair of synthetic DNA oligonucleotides encoding the new peptide sequence.

A PCR reaction was performed with cDNA generated from human cytotoxic T lymphocytes (CTL) as a template. The primers used had the following sequences:

5' GGG GGG GGA TCC GGT GGG GGA GGC GGA TCA GGA GGC TCA GGT GGG TCA GGA GGC ATC CAG CGT ACT CCA AAG ATT CAG 3' (forward primer)

5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3 ' (reverse primer).

The PCR reaction generated a DNA fragment comprising the sequence encoding human β2microglobulin as expressed on the cell surface, i.e. without the signal peptide. The fragment also contained sequence 5' of the β2microglobulin gene which encodes a linker fusion sequence. The fragment was cloned into the DNA expression vector pGMT7 (Studier, et al, Methods Enzymol (1990) 185: 60-89 Issn: 0076-6879) using the unique BamHI (GGA TCC and HindUI (GAAJTC) restriction sites which are underlined in the primer sequences above.

The resulting DNA plasmid was then modified by cloning a pair of synthetic DNA oligonucleotides into the unique Ndel and BamHI sites. The sequences of the DNA oligonucleotides were as follows:

5' T ATG GGT ATT TTA GGA TTT GTT TTT ACA TTA G 3' (coding strand)

5' GA TCC TAA TGT AAA AAC AAA TCC TAA AAT ACC CA 3' (non-coding strand).

The resulting DNA plasmid, pB Jl 96, can be used for expression of a polypeptide which is a fusion of three elements: a peptide corresponding to an antigenic peptide from influenza vims ('flu matrix peptide) which is known to be presented by HLA- A0201, a flexible linker sequence mainly comprised of Glycine and Serine residues, and β2microglobulin. The DNA sequence of plasmid pBJ196 is shown in Figure 1. The DNA sequence of pBJ196 was verified by dideoxy sequencing. The amino acid sequence of the fusion protein is indicated above the DNA sequence of the relevant region. Also indicated are the start site of transcription ('+1 ') and the restriction sites that can be used to substitute the peptide coding sequence for DNA cassettes encoding other antigen peptides. The vector pBJ196 contains a T7 promoter for high-level expression in a suitable E.coli strain such as BL21DE3pLysS or BL21DE3pLysE

(Novagen). Example 2 -A DNA plasmid vector for expression of an HLA-A2 nullpeptide- β2microglobulin fusion in E.coli.

This example describes the design and construction f a DNA plasmid vector for expression of a HLA-A2 null peptide-β2microglobulin fusion.

Nector pBJ196 was modified by restriction enzyme digestion with enzymes Νdel and BamHI and insertion of the synthetic oligonucleotide pair:

5' T ATG GCA CTG GCT GCG GCA GCC GCA GCG GTT G 3' (coding strand)

5' GA TCC AAC CGC TGC GGC TGC CGC AGC CAG TGC CA 3' (non-coding strand).

This generated plasmid pEXO 13. Figure 2 shows the DΝA and protein coding sequences of pEX013 covering the Ν-terminal part of the peptide-linker- β2microglobulin fusion polypeptide. This sequence was verified by DΝA dideoxy sequencing. The peptide fused to β2microglobulin in the protein expressed from pEX013, ALAAAAAAN, consists of Alanine residues in all positions except positions 2 and 9 which are optimised for anchoring the peptide to the HLA-A*0201 heavy chain.

Example 3 -A DNA plasmid vector for expression of an HLA-B8 nullpeptide- β2microglobulin fusion in E.coli.

This example describes the design and construction of a DΝA plasmid vector for expression of a HLA-B8 null peptide-β2microglobulin fusion. Nector pBJ196 was modified by restriction enzyme digestion with enzymes Νdel and BamHI and insertion of the synthetic oligonucleotide pair:

5' T ATG GCA GCC AAA GCT AAA GCA GCC GCT CTG G 3' (coding strand)

5' GA TCC CAG AGC GGC TGC TTT AGC TTT GGC TGC CA 3' (non-coding strand).

This generated plasmid pEX014. Figure 3 shows the DΝA and protein coding sequences of pEXO 14 covering the Ν-terminal part of the peptide-linker- β2microglobulin fusion polypeptide. This sequence was verified by DΝA dideoxy sequencing. The peptide fused to β2microglobulin in the protein expressed from pEX013, AAKAKAAAL, consists of Alanine residues in all positions except positions 3, 5 and 9 which are optimised for anchoring the peptide to the HLA-B*0801 heavy chain.

Example 4 -A DNA plasmid vector for expression of an HLA-A2 single position randomised peptide-- β2microglobulin fusion in E.coli.

Examples 4-11 describe the design and methods used for generating DΝA plasmids from which randomised peptide-β2microglobulin fusions, suitable for forming complexes with HLA-A2 and HLA-B8 heavy chains, can be expressed. The same principles can be applied to peptide sequences suitable for loading into any class I or class JJ HLA heavy chain molecule. Example 4 describes the constmction of vectors for the expression of peptide-β2microglobulin fusions, in which the peptide part contains HLA-A2 anchor residues and in which one position in the peptide could be occupied by any one of sixteen different amino acids. Example 5 describes a similar approach whereby fusions in which two positions in the peptide are randomised, producing a total peptide variation of 256. Example 6 describes the construction of vectors to produce fusions in which three positions in the peptide are randomised, producing a total peptide variation of 4096. Example 7 describes the constmction of vectors to produce fusions in which seven positions in the peptide are randomised, producing a total peptide variation of 2.7 x ~10⁸. Examples 8-11 describe the equivalent strategies for HLA-B8.

The approaches described can be easily adapted to produce fusion proteins in which further residues in the peptide part are randomised. It can also easily be adapted to produce fusion proteins in which positions other than those chosen in Examples 4-11 are randomised. However, the preferred options are to randomise one, two or three positions in the peptide. The reason for this is that, although a higher degree of variation is likely to include antigen complexes that stimulate a larger proportion of the T cells, a certain concentration of each particular antigen may also be required. Thus, an increase in the degree of variation in antigens has to be balanced against a loss of concentration of the individual antigen. As described, alloreactive T cells are likely to depend significantly less on peptide identity than is the case for conventional T cell responses, and so even a limited number of antigen complexes may be able to stimulate a significant proportion of the alloreactive T cells present in a sample.

A PCR reaction is performed with vector pEX013 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3'). The forward primer contains redundant bases in three positions and has the following sequence:

5' GGGGGGCATATGGCACGT GCTGCGXNNGCC GCAGCGGTTGGA TCC GGT GGG 3',

where X=C/A/G and N=C/A/G/T. The PCR products are cloned into vector ρGMT7 using the unique Ndel and Hindiπ restriction sites. The resulting vectors have a sequence redundancy factor of 48. The peptide-linker- β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of 16. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) - Ala - Leu - Ala - Ala - Ran - Ala - Ala - Ala - Nal

where the HLA-A2 anchor residues, at positions 2 and 9, are underlined. "Ran" is the position which is randomised and corresponds to position 5 in HLA-A2 peptide ligand. Due to the redundancy in the PCR primer, position 5 ("Ran") in the peptide could be occupied by any of the following amino acids: Leucine, Proline, Histidine, Glutamine, Arginine, Isoleucine, Methionine, Threonine, Asparagine, Lysine, Serine, Naline, Alanine, Aspartic Acid, Glutamic Acid, Glycine.

Resides that are excluded from being present at position 5 are: Phenylalanine,

Tyrosine, Cysteine, Tryptophan. However, Cysteine residues are rarely suitable in peptide ligands presented by HLA complexes. The insertion of Stop codons in the sequence, which would lead to plasmids that do not express a full length fusion protein, is also avoided.

Example 5 -A DNA plasmid vector for expression of an HLA-A2 double position randomised peptide- β2microglobulin fusion in E.coli.

Similar to the approach described in Example 4, a PCR reaction is performed with vector pEX013 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3'). The forward primer contains redundant bases in six positions and has the following sequence: 5' GGGGGG CATATGGCACGT GCT GCGXNNXNNGCAGCG GTT GGA TCC GGT GGG 3',

where X=C/A/G and N=C/A/G/T. The PCR products are cloned as described in Example 4.

The resulting vectors have a sequence redundancy factor of 2304. The peptide-linker- β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of 256. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) - Ala - Leu - Ala - Ala - Ran - Ran - Ala - Ala - Val

where the HLA-A2 anchor residues, at positions 2 and 9, are underlined. "Ran" indicates the positions which are randomised (positions 5 and 6 in HLA-A2 peptide ligand). Due to the redundancy in the-PCR primer, positions 5 and 6 in the peptide may be occupied by any of sixteen different amino acids as described in Example 5.

Example 6 -A DNA plasmid vector for expression of an HLA-A2 triple position randomised peptide- β2microglobulin fusion in E.coli.

Similar to the approach described in Examples 4 and 5, a PCR reaction is performed with vector pEX013 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3 '). The forward primer contains redundant bases in nine positions and has the following sequence:

5' GGGGGG CAT ATGGCACGT GCT XNNXNNXNNGCAGCGGTT GGA TCC GGT GGG3', where X=C/A/G and N=C/A/G/T. The PCR products are cloned as described in Example 4.

The resulting vectors have a sequence redundancy factor of 110592. The peptide- linker-β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of 4096. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) - Ala - Leu - Ala - Ran - Ran - Ran - Ala - Ala - Val

where the HLA-A2 anchor residues, at positions 2 and 9, are underlined. "Ran" indicates the positions which are randomised (positions 4, 5 and 6 in HLA-A2 peptide ligand). Due to the redundancy in the PCR primer, position 4, 5 and 6 in the peptide could each be occupied by any of sixteen different amino acids as described in

Example 4.

Example 7 - A DNA plasmid vector for expression of an HLA-A2 seven position randomised peptide— βϋmicroglobulin fusion in E.coli.

Similar to the approach described in Examples 4, 5 and 6, a PCR reaction is performed with vector pEX013 (see Example 2) as template and the same reverse primer as used in Example 1 (5'- GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG -3 '). The forward primer contain redundant bases in nine positions and has the following sequence:

5'- GGGGGGCATATGXNNCGTXNNXNNXNNXNNXNNXNNGTT GGA TCC GGT GGG-3', where X=C/A/G and N=C/A/G/T. The PCR products are cloned as described in Example 4.

The resulting vectors have a sequence redundancy factor of -5.8x10 . The peptide- linker-β2microglobulin fusion polypeptides expressed from the plasmid mixture

(plasmid library) have a redundancy of factor of ~2.7xl0 . All variation is m the peptide part of the fusion polypeptide which has the following sequence:

(Met) - Ran - Leu - Ran - Ran - Ran - Ran - Ran - Ran - Val

where the HLA-A2 anchor residues, at positions 2 and 9, are underlined. 'Ran' indicates the positions which are randomised and corresponds to positions 1, 3, 4, 5, 6, 7 and 8 in HLA-A2 peptide ligands. Due to the redundancy in the PCR primer, these positions in the peptide could each be occupied by sixteen different amino acids as described in Example 4.

Example 8 - A DNA plasmid vector for expression of an HLA-B8 single position randomised peptide- β2microglobulin fusion in E.coli.

A PCR reaction is performed with vector pEX014 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3'). The forward primer contains redundant bases in three positions and has the following sequence:

5' GGGGGGCATATGGCAGCCAAAGCTAAAGCAXNNGCTCTGGGA TCC GGT GGGGGAG 3',

where X=C/A/G and N=C/A/G/T. The PCR products are cloned into vector pGMT7 using the unique Ndel and Hindiπ restriction sites. The resulting vectors have a sequence redundancy factor of 48. The peptide-linker- β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of 16. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) - Ala - Ala - Lys - Ala - Lys - Ala - Ran - Ala - Leu

where the HLA-B8 anchor residues, at positions 3, 5 and 9, are underlined. "Ran" is the position (position 7) which is randomised as described in Example 4.

Example 9 - A DNA plasmid vector for expression of an HLA-B8 double position randomised peptide- β2microglobulin fusion in E.coli.

Similar to the approach described in Examples 4, 5, 6 and 7, a PCR reaction is performed with vector pEX014 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3'). The forward primer contains redundant bases in six positions and has the following sequence:

5' GGGGGGCATATGXNNGCCAAAGCTAAAGCAXNNGCTCTGGGA TCC GGT GGGGGAG 3',

The resulting peptide-linker-β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of 256. All variation is in the peptide part of the fusion polypeptide which has the following sequence: (Met) -Ran - Ala - Lys - Ala- Lys - Ala- Ran - Ala - Leu

where the HLA-B8 anchor residues, at positions 3, 5 and 9, are underlined. "Ran" indicates the positions which are randomised (positions 1 and 7 in HLA-B8 peptide ligand). Due to the redundancy in the PCR primer, position 1 and 7 in the peptide are each occupied by any one of sixteen different amino acids as described in Example 4.

Example 10 - A DNA plasmid vector for expression of an HLA-B8 triple position randomised eptide- β2microglobulin fusion in E.coli.

Similar to the approach described in Examples 4, 5, 6, 7 and 8, a PCR reaction is performed with vector pEX014 (see Example 2) as template and the same reverse primer as used in Example 1 (5' GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG 3 '). The forward primer contains redundant bases in nine positions and has the following sequence:

5' GGG GGG CATATGXNNGCCAAAXNNAAA GCAXNNGCT CTGGGA TCC GGT GGG GGA G3',

where X=C/A/G and N=C/A G/T. The PCR products are cloned as described in Example 4.

The resulting vectors have a sequence redundancy factor of 110592. The peptide- linker-β2microglobulin fusion polypeptides expressed from the plasmid mixture

(plasmid library) have a redundancy of factor of 4096. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) -Ran - Ala - Lys - Ran - Lys - Ala - Ran - Ala - Leu where the HLA-B8 anchor residues, at positions 3, 5 and 9, are underlined. "Ran" indicates the positions which are randomised (positions 1, 4 and 7 in HLA-B8 peptide ligand). Due to the redundancy in the PCR primer, positions 1, 4 and 7 in the peptide could each be occupied by any one of sixteen different amino acids as described in

Example 4.

Example 11 -A DNA plasmid vector for expression of an HLA-B8 six position randomised peptide— β2microglobulin fusion in E.coli.

Similar to the approach described in Examples 4 to 10, a PCR reaction is performed with vector pEX014 (see Example 3) as template and the same reverse primer as used in Example 1 (5'- GGG GGG GAA TTC AAG CTT ACA TGT CTC GAT CCC ACT TAA CTA TCT TG -3'). The forward primer contains redundant bases in eighteen positions and has the following sequence:

5'- GGG GGG CAT ATGXNNXNNAAAXNNAAAXNNXNNXNN CTG GGA TCC GGT GGG GGA G-3',

The resulting vectors have a sequence redundancy factor of ~1.2xl0¹⁰. The peptide- linker-β2microglobulin fusion polypeptides expressed from the plasmid mixture (plasmid library) have a redundancy of factor of ~1.7xl0⁷. All variation is in the peptide part of the fusion polypeptide which has the following sequence:

(Met) -Ran — Ran - Lys - Ran - Lys - Ran - Ran — Ran - Leu where the HLA-B8 anchor residues, at positions 3, 5 and 9, are underlined. 'Ran' indicates the positions which are randomised and corresponds to positions 1, 2, 4, 6, 7 and 8 in HLA-B8 peptide ligands. Due to the redundancy in the PCR primer, these positions in the peptide could each be occupied by sixteen different amino acids as . described in Example 4.

Example 12 - Synthetic peptides randomised at one, two or three positions and suitable as HLA-A2 antigens for recognition by alloreactive T cells

Examples 12 and 13 describe designs for synthesis of mixed peptides which can used for detecting T cell alloreactivity against HLA-A2 and HLA-B8 respectively.

Soluble HLA complexes can be produced with synthetic peptide ligands. Peptide synthesis allows the incorporation of any selection of amino acids at individual positions. Thus, highly selected peptide mixes are used for generating HLA complexes suitable for the detection of alloreactive T cell activity. Synthetic peptides can be made to order by a number of companies, for example, Research Genetics Inc., Huntsville, AL USA.

Mixes of the following amino acids are incorporated at selected positions in the peptide antigens:

Mix = Phenylalanine, Leucine, Serine Tyrosine, Tryptophane, proline, Histidine, Glutamine, Arginine, Isoleucine, Methionine, Threonine, Asparagine, Lysine, Naline, Alanine, Aspartic Acid, Glutamic Acid, Glycine.

A mixed peptide antigen, randomised at a single position, and for presentation by HLA-A2 has the following sequence:

Ala - Leu - Ala - Ala - Mix - Ala -Ala- Ala - Nal A mixed peptide antigen, randomised at two positions, and for presentation by HLA- A2 has the following sequence:

Ala - Leu - Ala - Ala - Mix - Mix - Ala - Ala - Nak

A mixed peptide antigen, randomised at three positions, and for presentation by HLA- A2 has the following sequence:

Ala - Leu - Ala - Ala - Mix - Mix - Mix - Ala - Nal.

A mixed peptide antigen, randomised at seven positions, and for presentation by HLA- A2 has the following sequence:

Mi - Leu - Mix - Mi - Mi - Mi - Mi -Mix - Nal

Example 13 - Synthetic peptides randomised at one, two or three positions and suitable as HLA-B8 antigens for recognition by alloreactive T cells

Synthetic peptides with randomised positions are synthesised for other HLA molecules as described in Example 12. For HLA-B8, a mixed peptide antigen, randomised at a single position, has the following sequence:

Ala - Ala - Lys - Ala - Lys - Ala - Mix - Ala - Leu

A mixed peptide antigen, randomised at two positions, and for presentation by HLA- B8 has the following sequence:

Mix - Ala - Lys - Ala - Lys - Ala - Mix - Ala - Leu, A mixed peptide antigen, randomised at three positions, and for presentation by HLA- B8 has the following sequence:

Mix - Ala - Lys - Mix - Lys - Ala - Mix - Ala - Leu.

A mixed peptide antigen, randomised at six positions , and for presentation by HLA- B8 has the following sequence:

Mix - Mix - Lys - Mix - Lys - Mix - Mix - Mix - Leu

Example 14 - Purification ofpeptide-β2 microglobulin fusion proteins

Peptide -β2m fusions were expressed from the DNA plasmid pBJ196 and mutated derivatives thereof such as pEX013 described in Examples 1-11. These encode fusion proteins comprising the 'flu matrix peptide (or mutated versions thereof), a flexible linker sequence mainly comprised of Glycine and Serine, and -β2m, in a transformation competent strain of E. coli (a number of which are commercially available from Novagen, Madison, WI, USA). The plasmid pBJ196 contains the ρeptide-β2m fusion under the control of the strongly inducible T7 promoter in the vector pGMT7 (Studier, et al. Methods Enzymol 185: 60-89 Issn: 0076-6879 (1990)). E. coli BL21 cells transformed with one of the peptide -β2m fusion vectors were plated on LB/agar/100 mg/ml ampicillin plates made according to a standard recipe. Transformants were then grown in TYP medium with Ampicillin (16 g/1 Bacto- Tryptone, 16 g/1 Yeast Extract, 5 g/1 NaCl, 2.5 g/1 K2HPO4, 100 mg/1 Ampicillin) to an OD600 ~ 0.4. For large-scale expression, 1 1 volumes of TYP media were prepared in 2 1 conical flasks and these were covered with four layers of aluminium foil prior to being autoclaved. Cell densities were measured using optical density at 600 nm wavelength (OD600) on a Beckman DU530 spectrophotometer. Sterile TYP media was used as a blank. I'

44

Inclusion bodies were purified as described (Gao, et al, Prot. Sci.7: 1245-49 (1998)). Cells were lysed in 'Lysis Buffer' (10 mM EDTA (from 0.5 M stock pH 8.0), 2 mM DTT (from 1 M stock in 10 mM sodium acetate pH 5.2, stored at -20°C), 10 mM Tris pH 8.1 (from 2 M stock pH 8.1), 150 mM NaCl (from 4 M stock), 200 mg/ml lysozyme (from 20 mg/ml stock stored at -20°C), 10% glycerol (from fluid), 2500 units of DNAase I and lOmM MgCl₂ using a 50 ml Dounce homogeniser, (DNase I and lysozyme were from Sigma). Sonication, in lysis buffer, to break open the cells was performed using a 12mM probe sonicator (Milsonix XL2020). The probe was tuned according to the manufacturers instructions. The resulting suspension was then diluted 1:1 in 'Triton Buffer' (0.5% (w/v) Triton X-100 (from fluid), 50 mM Tris pH 8.1 (from 2 M stock), 100 mM NaCl (from 4 M stock), 0.1% sodium azide (from solid), 10 mM EDTA (from 0.5 M stock pH 8.0), 2 mM DTT (from 1 M stock in 10 mM sodium acetate pH 5.2, stored at -20°C) and left overnight. The inclusion bodies were separated from cell debris by centrifugation in a Beckman J2-21 centrifuge equipped with a JA-20 rotor as described (Gao, et al, Prot. Sci.7: 1245-49 (1998)) and stored at -20°C.

Inclusion bodies were then thawed and resuspended in 'Resuspension Buffer' (50 mM Tris pH 8.1 (from 2 M stock), 100 mM NaCl (from 4 M stock), 10 mM EDTA (from 0.5 M stock pH 8.0), 2 mM DTT (from 1 M stock in 10 mM sodium acetate pH 5.2, stored at -20°C)), and denatured in 'Denaturing Buffer' (6M Guanidine and lOmM DTT buffered with Tris-HCl pH 8.1 all chemicals from Sigma).

Example 15 - Refolding of HLA /peptide - % microglobulin fusion protein complexes and their subsequent tetramerisation

The refolding of HLA with peptide -β2m fusion proteins was carried using modified versions of the method described in Garboczi et al, 1992, Proc Natl Acad Sci USA. 89 (8): 3429-3433 and Gao, et al, 1998, Prot. Sci.7: 1245-49. Soluble peptide-MHC tetramers were produced using a similar method to that described by Altman et al, 1996, Science, 274: 94-96. Recombinant MHC class I heavy chain and β2 microglobulin-peptide fusion were produced in E. coli cells transformed with the relevant expression plasmid. HLA complexes were folded using 30 mg of heavy chain protein and 25 mg of β2 microglobulin-peptide fusion protein. When required, the MHC complexes were biotinylated using 5 μg/ml purified BirA enzymes, 0.5 mM biotin and 5 mM ATP. The reaction was incubated at room temperature for 16 h. MHC-peptide complexes were recovered by FPLC purification and ion exchange chromatography. Tetramers were made by mixing biotinylated complexes with streptavidin-PΕ (Sigma Chemicals Co) at a molar ratio of 4: 1. The labelled tetramers were concentrated to 3-4mg/ml and stored in PBS at 4°C.

Example 16 — Refolding of HLAf synthetic peptide/ β2 microglobulin complexes and their subsequent tetramerisation

The refolding of HLA with synthetic peptide and β2m was carried using modified versions of the method described in Garboczi etal, 1992, Proc Natl Acad Sci USA. 89 (8): 3429-3433 and Gao, et al, 1998, Prot. Sci.7: 1245-49. Soluble peptide-MHC tetramers were produced using a similar method to that described by Altman et al, 1996, Science, 274: 94-96. Recombinant MHC class I heavy chain and β2 microglobulin were produced in E. coli cells transformed with the relevant expression plasmid. HLA complexes were folded using 30 mg of heavy chain protein, 25 mg of β2 microglobulin and 10 mg of peptide. When required, the MHC complexes were biotinylated using 5 μg/ml purified BirA enzymes, 0.5 mM biotin and 5 mM ATP. The reaction was incubated at room temperature for 16 h. MHC-peptide complexes were recovered by FPLC purification and ion exchange chromatography. Tetramers were made by mixing biotinylated complexes with streptavidin-PΕ (Sigma Chemicals Co) at a molar ratio of 4:1. The labelled tetramers were concentrated to 3-4mg/ml and stored in PBS at 4°C.

Example 17 — Detection of T cell activation using the ELISPOT assay

T cell activation is detected using the following reagent and methods.

Reagents:

ELISPOT kit for human IFN-γ from MABTECH (cat.no. 3420-2 ALP). Lymphoprep from Nycomed Pharma AS (cat. no. 1031966). Microfilter plates from Milipore (cat. no. MAIP s45 10). Alkaline phosphatase conjugate substrate from BIORAD (cat. no.

170-6432) (McCutcheon, et al, J Immunol Methods (1997) 210: 149-66; Mer ille, et al, Transplantation (1993) 55: 639-46; Ou, etal, Virology (1992) 191: 680-6;

Ronnelid & Klareskog, J Immunol Methods (1997) 200: 17-26; Rufkowski, et al, Rheumatollnt (1997) 17: 151-8; Scheibenbpgen, et al, Clin Cancer Res (1997) 3: 221-

6; Scheibenbogen, et al, Int J Cancer (1997) 71: 932-6; Schmittel, et al, JImmunol

Methods (1997) 210: 167-74).

Methods: Wells are coated with anti-IFN-γ monoclonal antibody (mAb) in coating buffer (0.1 M

Na₂CO₃, pH9.6) and incubated o/n at 4 °C or for a mimmum of 1 hr at 37 °C. Plates are washed x4 with RPMI (200 μl) and 200 μl R10 (R10: RPMI supplemented with L- glutamine, penicillin, and 10% heat-inactivated pooled human AB serum) is added to block non-specific binding. This is incubated at 37 °C for 3 hr followed by 200μl of R10. PBMCs (separated from heparanized blood on Lymphoprep, washed x3 and resuspended in R10) are added at an optimum of 3 x 10⁵/well. Assays are performed in duplicate with appropriate positive and negative controls. A peptide concentration of 2 μM is used for octamers/nonamers with incubation o/n at 37 °C and a flick of cells and washing x6 with PBS plus 0.05% Tween. 50 μl anti-IFN-γ-biotin diluted in PBS is added and incubation at room temperature for 3 hr is carried out followed by washing x6 with PBS plus 0.05% Tween. 50 μl streptavidin conjugated alkaline phosphatase (AP) diluted in PBS is added and incubation at room temperature for 1 hr followed by washing x6 with PBS plus 0.05% Tween. 100 μl of colour reagent is added followed by incubation for 20-30 minutes and washing xl with tap water. Water is flicked out and the plate is dried. Spots are counted using a low power microscope with an eyepiece grid.

Claims

1. A method for determining whether a T cell reacts with a predetermined Major Histocompatibility Complex (MHC) type, the method comprising: bringing a sample comprising said T cell into contact with a plurality of molecules of said MHC type, each MHC molecule being complexed with a peptide antigen whose contribution to a T cell receptor binding the MHC-peptide antigen complex is minimised; and determining whether said plurality of MHC molecules causes activation of the T cell.

2. A method as claimed in claim 1, wherein the or each MHC type is a Human Leucocyte Antigen (HLA) type.

3. A method as claimed in claim 1 or claim 2, wherein each MHC molecule is complexed with one of (a) a peptide antigen which presents substantially no T cell recognition features, and (b) a peptide antigen in which one or more T cell recognition features are randomly present.

4. The method as claimed in claim 3, wherein the peptide antigen comprises alanine, glycine and/or serine residues except at the or each amino acid position which is required to bind the peptide to the MHC complex ("anchor residue").

5. A method as claimed in claim 3, wherein one or more predetermined amino acid positions in the peptide antigen have a randomly selected amino acid.

6. A method as claimed in claim 5, wherein the peptide antigen comprises alanine, glycine and/or serine residues except at the or each predetermined amino acid position, and the or each amino acid position which is required to bind the peptide to the MHC complex ("anchor residue").

7. A method as claimed in any preceding claim, wherein said sample is brought into contact with a plurality of MHC types.

8. A method as claimed in claim 7, wherein each of the plurality of MHC types is provided in an individual reaction chamber.

9. A method as claimed in any preceding claim, wherein the sample is a sample of peripheral blood leucocytes (PBLs).

10. A method as claimed in any preceding claim, wherein the MHC molecules are provided as a multivalent complex.

11. A method as claimed in claim 10, wherein the multivalent complex comprises an MHC multimer, such as a tetramer.

12. A method as claimed in any preceding claim, wherein the peptide antigen is covalently linked to a subunit of the MHC via a flexible linker.

13. A method as claimed in claim 12, wherein the MHC is class I MHC and the subunit is β2microglobulin.

14. A method as claimed in claim 12, wherein the MHC is class II MHC and the subunit is the β chain.

15. A method as claimed in any preceding claim, wherein determination of whether said MHC type causes activation of the T cell is carried out using an enzyme- linked immunospot assay.

16. A cloning vector encoding an MHC subunit (preferably human) into which a nucleotide encoding a peptide antigen sequence of interest can be inserted such that expression of the vector produces a fusion protein comprising the MHC subunit with the peptide antigen fused thereto via a linker sequence.

17. A cell transformed with a vector as claimed in claim 16.

18. A multivalent class I or class II MHC/peptide complex, preferably a multimer, in which the peptide antigen is not recognised in a specific manner by T cells.

19. A complex as claimed in claim 18 modified by the features of any one of claims 1-6 and 12-14.

20. A kit for determining whether a T cell reacts with a predetermined Major Histocompatibility Complex (MHC) type, the kit comprising: a plurality of MHC molecules, each MHC molecule being complexed with a peptide antigen whose contribution to a T cell receptor binding the MHC-peptide antigen complex is minimised.

21. A kit as claimed in claim 20 which is for determining the reactivity of an individual towards each of a plurality of Major Histocompatibility Complex (MHC) types, and which comprises a plurality of different MHC types.

22. A kit as claimed in claim 20 or claim 21 modified by the features of any one of claims 2-6, 8 and 10-14.