key: cord-0822225-r6789z4b
authors: Moots, R J
title: A fistful of T cells.
date: 1998-06-03
journal: Br J Rheumatol
DOI: 10.1093/rheumatology/37.6.602
sha: 90dae44a46c26a6e198bae2a7943455af100be9c
doc_id: 822225
cord_uid: r6789z4b

Evidence incriminating T cells in rheumatoid arthritis (RA) is strong but circumstantial--like a smoking gun at the scene of a crime. To investigate this, T lymphocytes were studied in health and disease. The effect of mutations in the groove of HLA-A2 on peptide presentation to T cells was studied to investigate normal T cell function. This allowed a detailed description of the interactions between individual MHC residues and antigens. Subsequently, T cells in the autoimmune disease, multiple sclerosis, were studied, to investigate the mechanisms for breakdown in peripheral tolerance. T-cell clones that recognized both autoantigens and viral proteins were isolated, suggesting that infection may trigger disease. Autoantigens would need to be defined to use this strategy in RA. T-cell responses to type II collagen, a candidate auto-antigen, were therefore studied in RA and an epitope successfully defined. The search for microbial 'mimics' triggering RA, and novel forms of immunotherapy are now possible--with potential rehabilitation of T cells.

T stranger arrived and silently got off his horse.

in the crime of causing RA, and is holding the smoking gun. I would like to prosecute, but any decent jury Outside the saloon lay the sheriff, wounded. In the middle of the gathering crowd stood a man holding a needs more than circumstantial evidence to convict. The case that I will prepare for my prosecution must smoking gun. He claimed to be innocent. Nobody had witnessed the crime. Whodunit? How can this be address two main questions. (1) The target: what particular antigens in autoimmune disease did the solved?

There are some similarities here with autoimmune T cell see that it did not like? (2) The motive: why did the T cell turn bad and bypass immunological tolerance disease. As we get closer to understanding the pathology, the answer seems to slip farther away and new mechanisms to become activated? To prepare my case, I have utilized new technologies questions pop up to confuse the situation. For example, there is a general assumption that rheumatoid arthritis that are constantly evolving in immunology. In particular, I have concentrated on studying the relationship (RA) is an autoimmune condition. However, the basic pathological mechanisms underlying this remain between structure and function in the interaction of T-cell antigen receptor ( TCR), antigen and MHC unclear. Indeed, some authorities still venture to suggest that it is not an autoimmune disease at all [1] ! In molecule [10] . This lies at the heart of both normal and abnormal immune responses. Solving it may pro-fairness, the evidence for cellular autoimmunity, whilst certainly provocative, is not conclusive. Many years vide the key to understanding autoimmune diseases such as RA. It may also allow rehabilitation of the ago, a profound lymphocyte infiltration was observed on simple histological examinations of inflamed syn-T cell, by the development of novel, effective therapy where the immune response is modulated specifically ovial tissue, even resembling the follicles normally seen in lymph nodes, suggesting that T cells are activated. and safely-to prevent or reduce disease. In this paper, I will summarize my work on the For many years, clinical studies, designed to reduce either lymphocyte numbers or function, have occurred presentation of antigens to T cells. Firstly, the rules governing peptide interactions with class I MHC and in parallel. A whole range of interventions, including thoracic duct drainage [2] , total lymphoid irradiation TCR will be investigated, culminating in the detailed functional analysis of peptide presentation to T cells. [3] , leucopheresis [4] , monoclonal anti-CD4 antibody [5] and cyclosporin A [6 ] , have been tried, each with Secondly, such principles are applied directly to the study of two MHC class II-related autoimmune dis-some, albeit variable, success. More recently, genetic techniques have revealed an association between dis-eases: multiple sclerosis (MS ) and RA. Here potential triggers for MS are characterized, and target self-ease severity and progression in RA and a small sequence of amino acids on class II major histocompat-peptides for RA studied, allowing the development of specific, novel therapies. By the end of this paper, the ibility complex (MHC ) molecules, the 'shared epitope' [7] [8] [9] . Since the only known function of these molec-case will be complete. It is then for the jury to decide upon the verdict on the T cell. ules is to present antigen to T lymphocytes, this suggests that these cells have a role in the pathogenesis ANTIGEN PRESENTATION BY CLASS I MHC of rheumatoid disease.

Let us say that the T lymphocyte is the prime suspect Introduction Whilst the existence of 'tissue-type' or 'immune response' genes had been known for many years, their Accepted 10 February 1998. physiological role-other than that of confounding by influenza A-specific, HLA-A2-restricted, T-cell lines. transplant surgeons-was unclear. Detailed studies with inbred animal strains localized these genes to Recognition by CTL was assessed by their ability to lyse target cells that had been surface labelled with an area termed the MHC. The seminal work of 51Cr. Cell lysis, occurring as a consequence of CTL Zinkernagel and Doherty in the 1970s [11] revealed recognition, resulted in a loss of membrane-bound that T cells recognized antigen only when the pre-51Cr, which could be measured with a c radiation senting cells expressed the same MHC genes as did the detector. The nine amino acid (9mer) peptide, FMP T cell. This was soon refined to a description of MHC 58-66, was presented by far the best [16 ] . molecules being cell surface molecules that were able to bind antigens, in the form of short peptides [12, Mutations in a cluster of residues in the a2 domain a 13]. Since then, the search was on to discover exactly helix of HLA-A2 abolish presentation of peptide how antigens fitted into the MHC molecule. A major

The system of HLA-A2-restricted peptide presentabreakthrough occurred in 1987 when Bjorkman et al. tion was now sufficiently refined to study the role of [14, 15] resolved the crystal structure of HLA-A2, one individual residues of the MHC and peptide in antigen of the commonest MHC molecules. They described a presentation. I was also able to use it to describe highly polymorphic groove or cleft, running along the interactions between the accessory molecule CD8 and membrane-distal surface of the MHC. This had two class I MHC [17] (not discussed further here). Previous main features. Firstly, it was of the appropriate size work on individuals with naturally occurring mutations and shape to accommodate peptide antigen, and also in class I MHC molecules indicated that the a2 domain it was accessible for interaction with the TCR (Fig. 1) . may be important for peptide binding and hence CTL It was hoped that a detailed characterization of the recognition. In collaboration with Professor Jeff molecular interactions underlying antigen presentation Frelinger and Dr Masanori Matsui ( University of would lead to more efficient means of stimulating the North Carolina, Chapel Hill, NC, USA), I determined immune system (for vaccination), and novel ways to the effect of introducing a series of mutations spanning switch off the immune system (in autoimmune disease).

the a2 domain a helix of HLA-A2 on presentation of X-ray crystallography, however powerful a tool, peptide to T cells [18] . This constituted the whole of could not provide information about the dynamics of one wall and one-half of the floor of the peptidepeptide/MHC interaction and initially was not of binding groove (Fig. 1 ). sufficient resolution to determine the exact orientation

The strategy for this large undertaking was to use of peptide within the groove of the MHC molecule.

the novel technique of 'saturation mutagenesis' as a Such information could best be gained by a functional tool to introduce a series of random amino acid analysis of peptide/MHC/TCR interactions. I set out to study this with Professor A. J. McMichael at Oxford University. At that time, the structure of class II MHC [responsible for presenting peptide antigen to T helper ( TH ) lymphocytes] was not known; however, the high sequence similarity suggested that it would be similar to class I MHC.

The optimal class I MHC-restricted peptide contains nine amino acids A full study of peptide presentation requires definition of the optimal peptide (with respect to both sequence and length) for a particular MHC molecule. Perhaps surprisingly, this key detail was not known. The system I selected to study was the presentation of influenza A virus by HLA-A2 to human cytotoxic T lymphocytes (CTL). This was chosen because HLA-A2 was a common MHC molecule, it was structurally defined, and T-cell responses to influenza A virus protein, restricted by HLA-A2, had already been observed. Although this particular MHC molecule was not directly associated with rheumatological disease, it had a significant sequence similarity with HLA-B27, for which no formal structural data were known. synthesized and tested for their ability to be recognized mutations along the whole of the a2 domain a-helix. MHC. In class II MHC, the third allelic hypervariable region corresponds closely to positions 152-156 of In this method, short oligonucleotides were synthesized, the reservoirs of each base deliberately contamin-class I MHC. This would suggest that the third allelic hypervariable region on class II MHC molecules might ated with a small equimolar mix of the other three bases. This was designed to result statistically in one also be intimately associated with peptide presentation. Indeed, it may be responsible for binding the 'arthrito-coding change per oligo, which were then cloned into Escherichia coli, expanded and sequenced. Those con-genic' peptide involved in the pathogenesis of RA. taining appropriate mutations were selected for study by transfecting into Hmy2.C1R cells that did not MHC heavy chain interaction with b 2 microglobulin is necessary for MHC class I function normally express class I MHC. The result was a large panel of clones that expressed HLA-A2 with point Analysis of the effect of further mutations in the a2 domain of HLA-A2 revealed that some residues lying mutations spanning a major part of the peptide-binding groove. The advantage of studying the effect of random in the floor of the groove also affected antigen presentation [21] . The importance of these particular positions substitutions at each position was that it reduced selection bias in choosing mutations to study.

was unexpected, because their side chains pointed not upwards, for interaction with peptide, but rather These transfected cells were then used to present FMP 58-66 peptide to influenza A matrix peptide-down-away from both peptide and TCR. This area of the floor of the groove lay above the small b 2 specific, HLA-A2-restricted T cells, as before. The results revealed that a striking cluster of residues from microglobulin chain that was non-covalently associated with the heavy chain ( Fig. 1 ). At that time, the func-positions 152 to 156 in the a2 domain a helix of HLA-A2 were critical both for peptide presentation tional importance, if any, of b 2 microglobulin was not known. This observation indicated that an interaction [18] and for an anti-HLA-A2-allospecific response [19, 20] . Further selective mutations with a variety of between the heavy and light chains was indeed necessary for class I MHC function. This has subsequently amino acid substitutions were introduced at these sites.

Even conservative mutation, with only subtle been confirmed, using mutant cell lines with defects in antigen presentation [22] . changes in the amino acid side chain, still resulted in the loss of recognition by T cells, indicating that this was a very important area for peptide presentation.

Pockets lining the groove of class I MHC bind peptide antigen and are important for peptide presentation The amino acid side chains of the critical residues were orientated either in, towards the groove (to bind A number of depressions or pockets in the peptidebinding groove were observed in the crystal structure peptide), or up, towards the TCR ( Fig. 2) . It is interesting to compare this site on the a2 domain a helix of of HLA-A2 and other class I MHC molecules [14, 15] . These appeared to be of sufficient size, shape and HLA-A2 with the a helix on the b chain of class II charge to accommodate amino acid side chains of the peptide antigen; however, whether they had any important role in binding peptides was not known. The effect of selected mutations introduced into many of the pockets of HLA-A2 by site-directed mutagenesis was therefore studied in collaboration with Professor Jack Strominger and Dr France Latron (Dana Farber Cancer Institute, Harvard University, USA). Most of these substitutions led to a loss of peptide presentation, suggesting that they were indeed responsible for binding peptide-even those pockets that looked too 'shallow' on X-ray crystallographic examination [21, 23, 24] . These residues were also found to have the paradoxical effect of selectively interfering with either the presentation of peptide fragment, or whole virus, but not both. This suggests that they are important in the dynamics of peptide binding to MHC during breakdown of larger fragments to bound peptide [25]. deduce the position of peptide antigen within the groove. The potential benefits of such a description Charge changes at either end of the peptide were then correlated with charge change in the HLA-A2 are great. These would include constructing criteria for the prediction of other class I MHC allele-associated molecule. In particular, a peptide with a change of charge from neutral to a positive residue at position peptides, improving vaccine design, and potentially generating 'blocking' or 'inhibitory' peptides that 66 could now be presented by an HLA-A2 molecule with a negative charge change at position 116 (on the bound MHC, but could not be recognized by TCRresulting perhaps in the selective switching off of left-hand side of the floor of the groove). This would indicate that these two residues lie in close proximity. autoimmune responses (Fig. 3, see below) .

Within the extensive panel of transfected cells, some Further analyses of other such charge change complementations were able to pin down the orientation and expressed HLA-A2 molecules with altered electrostatic charges at either end of the groove, and resultant loss position of peptide in the groove in detail [26 ] (Fig. 4) . Molecular modelling studies from these data reveal of peptide presentation. We hypothesized that the introduction of a complementary opposite charge that the FMP 58-66 peptide lies in the groove of HLA-A2 in an elongated form from the amino to change in the amino acid of the peptide that bound to that part of the MHC where the mutation occurred carboxyl ends, with a short 'hump' in the middle [21] . The relative importance of most of the amino acids may restore presentation. This would indicate that these two residues must interact and lie in close proxim-could be determined by this charge change complementarity technique. Later, high-resolution X-ray tech-ity. A panel of synthetic peptides based on the index FMP 58-66 peptide was therefore synthesized, with niques using purified recombinant class I molecules have confirmed that peptides lie in the groove of class I individual substitutions of positive and negatively charged amino acids at each position. They were not MHC molecules in such a manner [27] . presented to T cells expressing normal HLA-A2, but CLASS II MHC AND MOLECULAR MIMICRY some were presented by some of the abnormal HLA-A2

IN AUTOIMMUNE DISEASE molecules.

Multiple sclerosis Multiple sclerosis: a 'model' autoimmune disease for RA. Multiple sclerosis really is a T-cell-mediated, autoimmune disease. It is associated with class II MHC (HLA-DR2), dense T-cell infiltrates are found in inflammatory plaques and, in striking contrast to RA, autoantigens recognized by T cells have been discovered and characterized (reviewed by Moots and Wucherpfennig [28] ). The search for candidate antigens started by focusing on the target organ: the central nervous system. Various components of nerve sheaths, notably myelin basic protein (MBP), have been observed to contain T-cell epitopes in animals (and act as a target for T cells in experimental allergic encephalomyelitis (EAE ) [28] ). MBP-reactive T cells have also been observed in the blood of patients with MS where they are often clonally expanded and persistent [29] . In these key respects, considerably more is known about the immunology of MS than of RA. Since both diseases are T cell mediated and are associated with class II MHC molecules, I set out to investigate the potential mechanism whereby autoreactive T cells become activated and break immunological tolerance in MS. If this approach were successful, it may be possible to apply it to RA.

Molecular mimicry in MS. Potentially autoreactive T cells may escape from deletion in the thymus during maturation and are often found in normal people. The resultant 'resting' autoreactive T cells are considered a F. 3.-Peptide immunotherapy. One potential mechanism whereby normal part of the immune repertoire and do not altered synthetic peptides may be used to modify (or prevent) disease normally induce disease because they are held in check is illustrated here. The interaction between peptide autoantigen, by peripheral 'tolerance' mechanisms [28] . The key MHC and TCR must first be characterized in detail (above). An event in the induction of T-cell-mediated autoimmunanalogue peptide is then synthesized (below). This would have ity, and thus the development of MS or RA, would selective alterations, designed to retain binding to the MHC, but therefore be activation of these potentially autoreactive interact with the TCR in a manner that will either block TCR T cells. Two major hypotheses to explain this are: recognition, or even 'switch off ' the TCR so that it is unable to respond to other peptides. T-cell activation by superantigen and molecular M antigens, myocardial and glomerular tissue [30] . A comparable event in T cells would be somewhat more complicated. Here there would have to be a dual specificity of TCR for both self and foreign peptide, each presented by disease-specific MHC molecules such as HLA-DR2 for MS and HLA-DR4 in RA ( Fig. 5) . It has proved to be a considerable task to investigate this [28] . Many have searched in vain, hampered by the far greater complexity of antigen recognition for T cells, compared to antibody. As for humoral mimicry, computers have been used to search for linear sequences of similarity between self and foreign proteins. This approach was not successful, because unless the self and foreign sequences were identical (up to a 1 in 206 chance), a significant recognition would not have been anticipated. Over the last few years, we have learnt that TCR recognition is actually more degenerate, and cross-reactivity at the T-cell level may indeed be common (reviewed by Moots and Wucherpfennig [28] ). Indeed, the same MHC molecule (murine I-Au) has been observed to present two dissimilar peptides to the same specific T-cell clones, leading to the term 'space mimicry' [31] .

This approach was refined further by Wucherpfennig and Strominger [32] . T-cell clones reactive to human MBP were generated from patients with MS and used to characterize the structural requirements for recognition of the immunodominant MBP(85-99) peptide in terms of MHC binding and TCR recognition. A series of amino acid substitutions were then selectively introduced at each critical position and the effects on T-cell recognition analysed. From this, the sets of amino acids permitted at each of the critical residues were defined. These structural data, together with the knowledge that amino acid side chains required for binding peptides derived from different antigens.

The diverse nature of the viral peptides able to stimulate MBP-specific T-cell clones from these data mimicry. Both implicate microbial pathogens and are supported by some (albeit limited) epidemiological would suggest that more than one single pathogen is able to trigger autoimmunity. This may explain why it evidence. There is currently very little evidence to support the role of superantigen as a trigger of MS has been so difficult to link conclusively the pathogenesis of individual autoimmune diseases such as RA to or RA.

Molecular mimicry depends upon a structural homo-particular pathogens, in the face of considerable evidence for association of disease with infection in general logy between self and foreign antigens. This would result in a cross-reactive T-cell response if there were [28] . Rather, it is more likely that a group of common microbial pathogens are involved in the pathogenesis sufficient similarities between the two peptide epitopes for activation, yet enough differences to overcome of autoimmune processes. In this case, an immunological approach has considerably more potential than tolerance. In rheumatic fever, there is a cross-reaction at the level of antibody between group A streptococcal epidemiological methods to determine the triggers.

F. 5.-Molecular mimicry. This potential mechanism explains how tolerance to self-proteins may be broken in autoimmune disease. Here, T cells make an appropriate immune response to microbial antigens produced during infection. One T-cell clone, however, with an additional specificity for a self-antigen may become activated. If, during normal circulation through the body, this activated T cell encounters the appropriate self-antigen, it would proliferate further, mediating inflammation and hence disease.

The exciting discovery of the potential for molecular human herpes group viruses. MBP-specific T-cell lines were first set up from mimicry in MS was an important advance. However, if peptides synthesized in the laboratory from published lymphocytes isolated from the peripheral blood of patients with MS (courtesy of Charles Poser, Beth sequences of viruses cross-react with the self-protein MBP for presentation to TCR by the disease-associated Israel Hospital, Harvard University, Boston, USA). A total of 19 patients were studied. MBP-specific T-cell molecule HLA-DR2, can this occur in the more physiological circumstances when cells are infected with whole lines were grown from 11 of them, corresponding with other reports. However, in four of the patients with virus? Cells have complicated machinery for presenting antigens to T cells. Proteins are first broken down to MBP-reactive T-cell lines there was also some recognition of human coronavirus (CV ) 229E, a common long peptide fragments in discrete cellular compartments by specific proteases. These are loaded onto cause of upper respiratory tract infections (Fig. 6) . This was investigated further, by studying the other class II MHC molecules by helper proteins. Finally, after trimming of the peptide, it is exported to the cell human coronavirus serotype OC43 and the murine MHV A-59. Only 229E was recognized by MBP-surface in complex with the MHC, where it is available for recognition by a specific TCR [34] . Would this specific T-cell lines. Lymphocytes from the patient with the best cross-reactive response were studied in more complicated, specific cellular mechanism result in the exposure of the same sequences, that seem to work detail. MBP-specific T-cell clones were made by limiting dilution. Three independent clones with good react-well, when computer-predicted peptides are made in the test tube? ivity against MBP were tested against CV and the other viruses. In each case, the clones were able not I addressed this question in the laboratory of Strominger and Wucherpfennig, by assembling a panel only to recognize the MBP peptide (MBP 85-102), but also CV229E and herpes simplex type 1. Other of human viruses, and testing them for T-cell crossreactivity with MBP. Viruses were chosen in preference coronaviruses were not recognized, nor were the other viruses, including herpes simplex type 2. In this case, to bacteria because it was technically considerably easier to work with them in tissue culture. They were the antigen receptor of the MBP-specific T-cell clones appeared to have a triple specificity! Preliminary time selected on account of their availability, capacity to infect the APC used in proliferation assays, genome scale experiments indicated that one of the early HSV 1 proteins (expressed within the first 4 h after infection) size (hence potential to code for protein) and, in some tive T cells in the diseased compartment, the joint, would be somewhat more convincing.

I determined to investigate the prevalence of T-cell responses to C II in RA, applying the lessons learnt from analysis of the structure/function relationships of MHC. One problem from previous studies using collagen was that T cells were actually reactive to a contaminant, pepsin, rather than collagen. In addition, bovine collagen was usually used, because of its better availability. As for other T-cell work, the purity of the protein antigen was critical. I obtained highly purified C II from Dr Vic Duance of the Connective Tissue Research Unit, Cardiff University. Gel electrophoresis analysis suggested that this protein would be of sufficient purity for use in generating specific T-cell responses. Hospitals, Boston, MA, USA. Cells were grown in culture and their capacity to respond to purified human C II was estimated in simple [3H ]thymidine incorporacontains the mimicry epitope. Work is now in progress tion proliferation assays. T cells recognizing C II were to define this in detail and determine whether this maintained in culture and restimulated with further C II phenomenon holds true for independent clones from until stable T-cell lines emerged. other patients. It is exciting to consider that the triggers A degree of C II specificity was observed from T cells for MS may be elucidated by this method.

derived from synovial fluid T in 11 out of 16 patients with RA. No reactivity was observed in synovial or T-cell responses in RA peripheral blood T cells taken from any of the three Despite hard searching, no definitive autoantigen patients with diagnoses other than RA. Type II collahas been found in RA, suggesting to some that T cells gen reactivity was concentrated in the joint ( Fig. 7 ) are not important in this disease. The same had been with responses in blood (taken in parallel ) only said for MS a few years before the discovery of T-cell observed in two of those patients. No T-cell responses reactivity to MBP. I believed that the recent technical to C II were observed in the blood in the absence of a improvements in methods for generating T-cell lines synovial fluid response. were now sufficient to try to search for autoantigens in RA. Various candidate autoantigens have been proposed, including heat shock proteins, cartilage link protein and other proteoglycans (reviewed by Moots and Wucherpfennig [28] ). Perhaps the most compelling of these is type II collagen (C II ). This is expressed almost exclusively in synovial joints and is able to induce an inflammatory erosive arthritis in rodents, which shares many features with RA. This animal disease, collagen-induced arthritis (CIA), is mediated, at least in part, by C II -reactive T cells [35] . Indeed, a large amount of effort has been invested in studying the potential for inducing oral tolerance, with some preliminary success [36 ] . Previous studies had suggested that T cells reactive to C II may exist in patients with RA, and indeed may persist [37] , but little was known about the detail, if true.

The occurrence of autoreactive T cells alone would 'normal' people. However, a concentration of autoreac-A human C II epitope is defined outlook for inbred strains of mice with the murine model of MS, EAE, is now exceedingly good! With The observation of a T-cell response to C II in patients with RA can now be characterized in more detail. The peptide epitopes and TCR specificity known, it is possible both to prevent disease and induce remission key to this is defining the peptide epitope. The approach that I employed for this was to use informa-in established disease by the use of analogue peptides that are able either to block or actively 'switch off ' tion on the dynamics of peptide class II MHC interaction to predict peptides from within C II that would activated autoaggressive T cells [46 ] , and do this specifically and safely ( Fig. 3) . There are similar reports be capable of binding disease-specific MHC molecules such as HLA-DR4. It is known that slightly longer in the CIA model for RA [47] . Whilst this is obviously harder in humans-thankfully we are not inbred-it peptides bind to class II compared to class I MHC, of the order of 15 amino acids. The dynamics of this now becomes a possibility as the underlying mechanisms behind autoimmune disease become clearer. interaction also appear to differ from class I MHC and there would appear to be more degeneracy in binding to particular class II molecules [38, 39] . Crystal CONCLUSIONS structure analyses of a variety of class II MHC molec-T lymphocytes are at the same time good (protect ules indicate that relatively longer peptides are bound, us from disease), bad (induce autoimmune disease) overhanging the ends of the groove. and ugly (give us many clues, but show increasing Unlike class I MHC, the main interaction is with levels of complexity the nearer we get to understanding the peptide backbone rather than amino acid side them). Were Clint Eastwood a rheumatologist or chain, reflecting the lower specificity [39] . However, immunologist, he would have solved this long ago, peptide-binding pockets in class II MHC are present, and moved on to other things. As scientists, we require for binding peptide amino acid side chains. Moreover, a higher level of proof before we are convinced. This in vitro binding studies, where peptides with varying is now starting to happen. The rapid developments in sequences are studied for their ability to bind specific immunological techniques over recent years are there class II molecules such as HLA-DR4, have suggested for us to grasp, and apply to clinical problems such as that certain motifs or anchor residues may be found RA and MS. Understanding, as we do now, the in peptides that bind to individual class II molecules molecular immunology of antigen presentation in auto-[40-42]. A computer search for potential HLA-DR4immune disease, may have as great a consequence for restricted C II epitopes was performed on a human the future of rheumatology as did the invention of the protein database using criteria for HLA-DR4-binding

Winchester rifle for the Wild West. motifs from published data. Seven epitopes were pre-I conclude that in this awful crime of crippling dicted to bind to HLA-DR4. These were then synthedestruction of the joint in RA, there is good reason to sized and used to sensitize autologous antigencharge the T lymphocyte. Using molecular biological presenting cells for presentation to C II -specific T-cell techniques, I and others now have strong evidence to lines.

implicate the lymphocyte's unique antigen recognition One peptide, comprising residues 258-272, was in the crime. The target is possibly an immunodominrecognized by T-cell lines from patients with RA, and ant epitope within type II collagen. What was the not by those from control patients with diagnoses of motivation? I believe that it was not revenge (superantireactive arthritis, psoriatic arthritis or osteoarthritis. gen), but rather mistaken identity (molecular mimicry, Recently, other work in transgenic mice engineered to as for MS ). So what are the lessons for the future? express human HLA-DR4 and DR-1 molecules has Can the lymphocyte be rehabilitated? Perhaps better supported the idea that the immunodominant C II pepeducation, in the form of specific immunotherapy, tide lies in this region [43, 44] . In addition to this, the rather than our current methods of punishment with same peptide appears to mediate the bulk of the antirelatively crude immunosuppression, will allow the C II response in the H-2q mouse [45] . The characterizalymphocyte to retain some honour and lead a more tion of such an immunodominant C II peptide, restricted profitable life. by disease-associated MHC molecules in humans with disease, is of fundamental importance to the study of A potential triggering mechanisms of RA and lays the foundation for molecular mimicry work to occur, as I am especially grateful for help and advice from Professors P. A. Bacon (Birmingham University), for MS.

A. J. McMichael (Oxford University), J. L. Strominger POTENTIAL FOR IMMUNOTHERAPY and K. W. Wucherpfennig (Harvard University, USA), P. Emery (Leeds University) and J. S. H. Gaston Understanding the molecular immunology of antigen presentation is a major achievement. By elucidating (Cambridge University). I would also like to thank Dr R. W. Costello, Dr R. N. Thompson (North Liverpool the fine specificity of events between the tri-molecular complex of peptide, MHC and TCR, we are able to University Hospital ) and Mr J. H. Moots for their constructive criticism of this manuscript. My work at apply the lessons learnt to autoimmunity. The discovery of molecular mimicry in MS and a self-peptide in Oxford was supported by the Medical Research Council UK ( Training Fellowship) and at Harvard by RA may have significant clinical benefits. Indeed, immunotherapy already is a treatment of today. The the Wellcome Trust (Advanced Training Fellowship).

Limited regions R of the a2 domain a helix control anti-A2 allorecognition

The role of T cells in the immunopathogenesis An analysis using a panel of A2-mutants. of rheumatoid arthritis

Curd relationship between HLA-A2-restricted and anti-A2

Thoracic duct drainage in rheumatoid allospecific T-cell recognition

Total lymphoid irradiation

The fine specificity of antigen recognition by

The effects of HLA-A0201-restricted cytotoxic T lymphocytes. PhD repeated leukapheresis in patients with severe refractory Thesis, Institute of Molecular Medicine, Oxford rheumatoid arthritis

Barber clonal antibodies

Assembly of MHC class I molecules analyzed

Cyclosporine in rheumatoid arthritis: an overin vitro

Genetic typing for conserved residues in the cleft of HLA-A2 in presentaof patients with inflammatory arthritis at presentation tion of a nonapeptide to T cells

Significance of the six peptide binding pockets of

The shared HLA-A2.1 in influenza A matrix peptide-specific CTL epitope hypothesis. An approach to understanding the reactivity

Bacon cellular peptides in influenza A matrix peptide-specific PA, Bell JI. The association of HLA DR beta alleles CTL recognition

The precise localisation of a peptide in the cleft of 10. Moots RJ. In the groove: the fine detail of antigen HLA-A0201

Crystal structure of the major histocom

MHC-restricted cytotoxic patibility complex class I H-2Kb molecule containing a T cells: studies on the biological role of polymorphic single viral peptide: implications for peptide binding and major transplantation antigens determining T-cell restric-T-cell receptor recognition. Proc Natl Acad Sci USA tion-specificity, function, and responsiveness

The role of T cells in

Immunological recognize fragments of the influenza nucleoprotein

The epitopes of influenza expansion and persistence of human T cells specific for nucleoprotein recognized by cytotoxic T lymphocytes an immunodominant myelin basic protein peptide. can be defined with short synthetic peptides

Group A streptococcal M 14

Structure of the human class

HLA-A2

Degenerate recognition of a dissimilar antigenic peptide

by myelin basic protein-reactive T cells. Implications for Strominger JL, Wiley DC. The foreign antigen binding thymic education and autoimmunity

Identification of the Structural requirements for binding of an immunodominnonamer peptide from influenza A matrix protein and ant myelin basic protein peptide to DR2 isotypes and the role of pockets of HLA-A2 in its recognition by for its recognition by human T cell clones

A cross-species functional in T cell-mediated autoimmunity: viral peptides activate interaction between the murine major histocompatibility human T cell clones specific for myelin basic protein. complex class I alpha 3 domain and human CD8 revealed

42. Sinigaglia F, Hammer J. Defining rules for the peptide-MHC class II interaction. Curr Opin Immunol Effects of oral administration of type

Persistence of Specificity of an HLA-DRB1*0401-restricted T cell collagen type II-specific T-cell clones in the synovial response to type II collagen

membrane of a patient with rheumatoid arthritis

Three-DR1 transgene confers susceptibility to collagen-induced dimensional structure of the human class II histocompatarthritis elicited with human type II collagen

Holmdahl of the human class II MHC protein HLA-DR1 com-R. Identification of an immunodominant type-II collagen plexed with an influenza virus peptide. Nature peptide recognized by T cells in H-2q mice: self tolerance

Treatment of ibility complex class II-peptide interaction based on experimental encephalomyelitis with a peptide analogue peptide side chain scanning

Kang specificity of HLA-DR4 molecules: correlation with AH. Design of a synthetic peptide capable of treating rheumatoid arthritis association