Int J Clin Lab Res 22:90-94, 1992
9 Springer-Verlag1992
Are heat shock proteins involved in autoimmunity? J. S. H. Gaston
Department of Rheumatology, University of Birmingham, Birmingham, UK
Summary. Heat shock proteins (HSPs) have been postu-
Characteristics o f autoantigens
lated to be critical antigens in both autoimmune disease and experimental models of autoimmunity. This postulate has been largely based on the remarkable conservation of aminoacid sequence between human and bacterial HSPs, so that it has been argued that immune responses initially directed against the HSP of an infectious agent, would have the potential to initiate or maintain autoimmune disease. This would apply especially to T celt recognition of HSPs, since the T cell focuses on short peptide epitopes within a protein antigen rather than on the antigen's secondary structure. This article critically evaluates the available experimental evidence relating to this hypothesis: although research has clearly highlighted the central role of HSPs in the cellular immune response to pathogenic organisms and has shown the potential for T cell responses directed against self HSPs, a role for self HSPs as major target antigens in autoimmune disease has yet to be firmly established.
The antigenic targets in autoimmune disease fall into two general categories. Firstly, there are antigens which are clearly involved in the disease process, and have tissue specificity, e.g. the acetylcholine receptor in myaesthenia gravis, or the thyroid-stimulating hormone receptor in thyrotoxicosis. The second category involves molecules or molecular complexes which are critical for the normal functioning of every cell. Examples include the protein/ RNA complexes required for mRNA processing and splicing, DNA topoisomerase I, and tRNA transferase enzymes, all of which are targets of particular autoantibodies characteristic of certain connective tissue diseases [37]. In these cases the autoantibodies are excellent markers of disease, but it is difficult to envisage how autoimmune responses to these ubiquitous and indispensible cellular components can result in very specific tissue-restricted disease processes.
Key words: Heat shock protein - Autoimmunity - TceUs
H S P s as putative autoantigens
- Arthritis - Bacterial antigens
Introduction
Heat shock proteins (HSPs) have recently been the subject of extensive discussion in relation to their possible involvement in autoimmune diseases such as rheumatoid arthritis (RA) [36] and insulin-dependent diabetes mellitus [9]. This article will cover some of the experimental evidence which has fuelled this discussion, concentrating on evidence obtained in studies of experimental and clinical arthritis, and will examine the extent to which HSP involvement in autoimmunity has indeed been demonstrated.
HSPs fall into the second category of putative autoantigens. It has often been the case in the past that autoantigens have been identified and discussed well in advance of their cellular functions being worked out, and in some cases patients' sera with high titres of antibodies specific for autoantigens have greatly aided the process of identifying their biochemical nature and function. Material from patients with autoimmune disease has not so far aided investigations into the functions of HSPs, but the physiology of HSPs is currently the focus of much research, and once again those interested in autoimmunity find themselves grappling with the fundamentals of cell biology. The present view is that HSPs are primarily involved in the correct folding, assembly and compartmentalization of intracellular proteins [15]. It is probable that stresses to the cell, such as heat shock, result in increased quantities of denatured or damaged proteins, and thus a greater requirement for HSPs. The increased synthesis of
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity? these proteins in response to stress was used therefore as a defining characteristic of HSPs, but it is now clear that many of the members of different HSP families are required at all times for normal cellular functioning, and indeed many HSPs are synthesized constitutively by cells. As with other autoantibodies to cellular constituents, a connection cannot be easily made between interference with the physiological function of HSPs and the occurrence of a disease like arthritis or diabetes. One feature common to the intracellular targets of autoantibodies is that they are usually conserved molecules, i.e. their amino acid sequence does not vary greatly between different species [22, 38]. No doubt the conservation reflects the critical functions which they perform, and which are similar in all cells. For HSPs the conservation is not just confined to eukaryotic organisms (as is the case for molecules involved in, for example, m R N A splicing) but extends to bacteria including Archaebacteria. Conservation of sequence leads to conservation of immunogenic epitopes, particularly those recognized by T cells, since T cell epitopes consist of short linear peptides, and do not depend on the secondary structure of the intact molecule. Thus marked conservation of sequence makes HSPs attractive potential targets for autoimmune recognition, perhaps as a complication of a conventional immune response to bacterial infection [27]. This review will consider the evidence as to whether this actually occurs in human autoimmune disease.
Humorai responses to HSPs in autoimmune disease
Although HSPs can be compared with other autoantigens, such as the splicing RNA/protein complexes, since both perform important intracellular functions, the immune responses they elicit seem quite different. A characteristic of many connective tissue diseases such as systemic lupus erythematosus (SLE) is the occurrence of very high titres of autoantibodies. In contrast, T cell recognition of the same autoantigens has not been emphasized, although it is true that studies o f T cell responses to autoantigens in these diseases are only in their infancy. In direct contrast to this, T cell recognition of HSPs has attracted attention right from the beginning, whilst strikingly high titres of antibody to HSPs are not seen in autoimmune diseases. Nevertheless there have been some observations on antibodies to HSPs in diseases which might have an autoimmune aetiology. Thus antibodies to a Drosophila antigen, which might be a HSP (the sera actually identify a chromosomal puff associated with the heat shock response), were found in 40% of sera from ankylosing spondylitis (AS) patients [25]. Immunoblotting has shown that these AS sera recognize several insect proteins and, although an epitope on one of these is also present on a human protein, it does not appear to be a HSP [26]. Thus an autoantibody response to HSPs has not been demonstrated unequivocally in AS. Antibodies to human HSP 70 and HSP 90 have been described in 30% - 5 0 % of SLE patients [29, 30], albeit in studies of small numbers, but perhaps this is inevitable in view of the multiplicity of autoantibodies seen in SLE. In
91 RA, patients have higher titres of antibody (especially IgG) to the 65-kDa mycobacterial HSP (HSP 65) than disease controls, but the elevations are not very striking, with 75% of patients having antibody levels in the normal range [40]. Furthermore, whether any one these antibodies cross-react with self HSPs is currently unclear. In summary, there is little evidence to suggest that HSPs elicit a major autoantibody response in human disease. T cell recognition of HSPs in autoimmune disease
Immunity to HSPs in infection Prior to discussing the possible role of HSPs as targets of cellular immunity in autoimmune disease, its worth emphasizing that recognition of HSPs by T cells is a major feature of the response to infections, particularly those involving intracellular organisms, whether bacteria or parasites [23, 35, 42]. Why the immune system should single out HSPs from amongst all the foreign antigens present in an infectious organism is unclear; presumably since HSPs are inevitably present in all organisms, and are highly conserved in terms of their structure and sequence, they may be reliable antigens for the immune system to target. Indeed, the conservation of HSPs might mean that encounter with one bacterial pathogen primes the immune system for responding to HSPs of a subsequent infecting species. However, as mentioned above, this preferential targeting of bacterial HSPs would seem to carry with it the opportunity for autoimmunity, because of the similarities between self and bacterial HSPs. A further feature of the response to intracellular organisms is the involvement of T cells which use the 76 receptor [18, 21]; again, a subset of these cells have a preference for responding to HSPs [17, 33], including self HSPs.
HSP-specific T cells and experimental arthritis Much of the evidence for the involvement of HSPs in autoimmunity comes from studies of experimental arthritis [6]. In particular mycobacterial HSP 65 was first implicated in the pathogenesis of adjuvant arthritis in rats by the finding that a T cell clone specific for this antigen was capable of inducing arthritis when transferred to naive rats [7]. More recent work has extended this observation to suggest that immune responses to HSP 65 are involved in several kinds of experimental arthritis, including arthritis induced by streptococcal cell walls, or by non-bacterial agents such as synthetic adjuvants and pristane [6, 11]. The ability of a HSP 65-specific T cell clone to induce arthritis was initially explained on the basis of cross-reactivity between HSP 65 and a component of cartilage proteoglycan [8]. Although this might still be the case, the identity of the cross-reacting component in proteoglycan has proved elusive, and the possibility that the cross-reaction might involve the homologous rat HSP has been considered. Mapping of the epitope within HSP 65 recognized by the arthritogenic T cell clone showed that it was
92 in a region (amino acids 180-188) where there is not great similarity between HSP 65 and the homologous rat HSP 60 [43]. However, despite this lack of similarity, subsequent experiments have produced the interesting finding that murine "),6T cell hybridomas which showed "autoreactivity" (spontaneous interleukin-2 production without antigen challenge), also recognized amino acids 180-188 of HSP 65 [31]. Although they recognized the corresponding peptide from mammalian HSP 60, which would account for their apparent autoreactivity, there was a striking preference for the mycobacterial peptide, since very high concentrations of mammalian HSP 60derived peptide only produced a sub-optimal response. Thus it is not certain that murine HSP 60 is the principal autoantigen recognized by these cells.
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity? 60-specific T cell clone showed cross-recognition of both recombinant human HSP 60 and recombinant mycobacterial HSP 65 [19]. The clone also responded to heattreated autologous mononuclear cells, presumably recognizing HSP 60 expressed by these cells. It is now of great importance to confirm the autoreactivity of this clone by determining the specific epitope common to human and bacterial HSPs to which it responds. Otherwise, the finding in ReA of an immune response to a bacteria-specific epitope in Yersinia or Salmonella HSP 60 would merely be another instance of the immune system's apparent preference for responding to HSPs in combating infection. However, the recognition o f the same target antigen by JA patients' T cells raises the possibility that occult infection might be playing a part in this disorder, most probably in one of the clinical subsets of JA, which is a highly heterogeneous group of arthritides.
HSP-specific T cells and human arthritis
In studies of human arthritis, T cell clones specific for both mycobacterial HSP 65 and HSP 60 of Gram-negative organisms have been isolated from joints; the epitopes in HSP 65 recognized by T cell clones from a patient with reactive arthritis (ReA) [12] and two patients with RA have been defined in detail [14, 32]. In each case there was no evidence of cross-recognition of cartilage proteoglycan or of type II collagen. For clones from the ReA patient, the epitope recognized was in a non-conserved region of the HSP 65 sequence, but those seen by the RA clones isolated in two laboratories were in highly conserved regions. Despite this, recognition of intact human HSP 60 has not been reported, so that it has not been fully established that the self HSPs serve as an autoantigen for these T cell clones. Nevertheless, when peptides from mycobacterial and human HSPs, were tested, the HSP 65-specific clones did respond to peptides containing the human sequence. However in our laboratory there has been a consistent difference of three logs in the concentration of mycobacterial and human peptides needed to stimulate the T cell clones: the bacterial peptides were recognized at nanogram concentrations, where microgram quantities of the human peptide were required (P. E Life, J. S. H. Gaston, in preparation). In juvenile arthritis (JA), there has recently been a report of T cell responses to recombinant human HSP 60 in experiments in which some care was taken to remove possible contaminating Escherichia coli antigens [16]. However, this recognition has not yet been demonstrated at a clonal level. In other studies clones specific for E. coli HSP 60 have been isolated from JA synovial fluid in patients in whom marked T cell recognition of Gramnegative bacteria is seen in synovial fluid (Life et al., submitted for publication). The epitopes seen by these E. coli-specific T cell clones have yet to be determined. The HSP 60 from Gram-negative bacteria is also a target of the synovial T cell response in ReA [19, 28], which is known to be triggered by specific enteric infection. In this disease it is not necessary to postulate autoimmunity, since it is now known that bacterial antigens reach the joint, where they could presumably activate T cell immunity. Nevertheless, one of the reports of a Yersinia HSP
Protective immune responses to H S P s
So far the possibility that HSP might be targets of an autoimmune T cell has been considered. Paradoxically however, induction of a response to mycobacterial HSP 65, in most animal models of arthritis, results not in disease but in protection of the animal from the subsequent induction of arthritis [1]. Interestingly, very similar conclusions have been reached in the non-obese diabetic mouse model of diabetes [10]. In this sense the HSP 65specifc "arthritogenic" rat T cell clone is the exception rather than the rule. Protection is seen not only in classical adjuvant arthritis, but also in streptococcal well waltinduced arthritis, arthritis due to synthetic adjuvant and pristane-induced arthritis [2, 39]. Protection can be obtained by immunization with intact HSP 65, with a peptide containing amino acids 180 - 188 of HSP 65 or infection with recombinant vaccinia virus expressing HSP 65 [20, 41]; most HSP 65-specific clones also confer protection when transferred to naive animals [5], although they have the same epitope specificity as the arthritogenic clone. It has been argued that prior exposure to HSP 65 or to HSP 65-specific T cells allows the immune system to mount a regulatory response, perhaps by recognizing idiotypes expressed by the HSP 65-specific T cells [4]. Such a response could down-regulate subsequent responses to HSP 65, or perhaps alter them qualitatively, e.g. by changing the pattern of lymphokines produced by the antigen-specific T cells. Neither of these possibilities has been conclusively demonstrated. However, the quality of the T cell response to HSP 65 is emphasized by work in pristane-induced arthritis, where both arthritic animals and those protected by prior immunization with HSP 65 showed marked T cell responses both to HSP 65 and to joint extracts [39]. Thus in the protected animals absence of arthritis does not equate with absence of antigen recognition or even absence of autoreactivity (as shown by the response to joint extracts). A physiological parallel of prior immunization with HSP 65 may be the effects of gut flora on susceptibility to arthritis. For both adjuvant and streptococcal cell wallinduced arthritis, disease susceptibility can be decreased
J. S. H Gaston: Are heat shock proteins involved in autoimmunity? by introducing particular organisms into the gut flora of previously germ-free animals [3, 24]. It has not been shown that HSPs are the critical antigens involved in the effect produced by gut flora, b u t this seems possible. The evidence that T cell responses to HSP 65 are normally protective in experimental animals may have an interesting parallel in RA. Early studies in RA suggested that T cell responses to H S P 65 were fairly c o m m o n , particularly in synovial fluid, b u t the interpretation of these findings has been complicated because it has now been clearly shown that r e c o m b i n a n t preparations of mycobacteria HSP 65 c o n t a i n immunogenic antigens from E. coli, including E. coli H S P 60. When this is allowed for, it now seems that T cells responses to HSP 65 are rather rare in RA, and p r o b a b l y less c o m m o n than in the normal p o p u l a t i o n [13, 34]. To highlight the relative rarity of such responses in RA, a recent study of HSP 65-specific synovial T clones from a single DR4 + RA patient who did respond to this antigen showed that the HSP 65 was only recognized when presented by H L A - D P rather than H L A - D R [14]. This is a very unusual pattern of HLA restriction for h u m a n T cell clones, and it is likely that the patient d i d not possess any T ceils which could respond to HSP 65 presented by DR4. This lack of DR4restricted T cells would occur if thymic deletion removed all the T cells whose receptors could recognize HSP 65 presented by DR4. The role of such thymic deletion is to prevent the emergence of T cells reactive with self. Thus D R 4 might generally be associated with a lack of a T ceil response to HSP 65, since it might not always be possible to generate a DP-restricted response instead. One could then argue that an i n a d e q u a t e response to HSP 65 in D R 4 + individuals might predispose to arthritis, if the protective effects of responses to HSP 65 in animal models of arthritis applied to arthritis in man.
Conclusion There is a considerable a m o u n t o f evidence to suggest that HSPs are " i n v o l v e d " in a u t o i m m u n i t y , but the details of this involvement are still far from clear. The evidence that HSPs are p r i m a r y target antigens of an autoi m m u n e response is limited, despite the enormous opportunities for such autoreactivity afforded by their marked conservation of sequence and their capacity to elicit T cell responses. The i m m u n e system's previous experience of HSPs seems to play a significant part in determining susceptibility to certain a u t o i m m u n e conditions.
References 1. Billingham MEL Carney S, Butler R, Colston ML A mycobacterial 65-kD heat shock protein induces antigen-specific suppression of adjuvant arthritis but is not itself arthritogenic. J Exp Med 171:339, 1990 2. Broek MF van den, Hogervorst EJM, Bruggen MCJ van, Eden W van, Zee R van der, Berg WB van den, Protection against streptococcal cell-wall induced arthritis by pre-treatment with the 65-kD mycobacterial heat shock protein. J Exp Med 170, 449. 1989
93 3. Broek MF van den, Bruggen MCJ van, Koopman JP, Hazenberg MP, Berg W van den, Gut flora induces and maintains resistance against streptococcal cell wall-induced arthritis. Clin Exp Immuno/(in press) 4. Cohen IR, Young DB, The immune system's view of invading micro-organisms, autoimmunity and the immunological homunculus. Immunol Today 12:105, 1991 5. Cohen IR, Holoshitz L Eden W van, Frenkel A, T lymphocyte clones illuminate pathogenesis and affect therapy of experimental arthritis. Arthritis Rheum 28: 841, 1985 6. Eden W van. Heat shock proteins as immunogenic bacterial antigens with the potential to induce and regulate autoimmune arthritis. Immunol Rev 121: 1, 1991 7. Eden W van, Holoshitz L Nero Z, Frenkel A, Klajman A, Cohen IR, Arthritis induced by a T lymphocyte clone that responds to Mycobacterium tuberculosis and to cartilage proteoglycans. Proc Nat! Acad Sci USA 82: 5117, 1985 8. Eden W van, Thole JER, Zee R van der, Noordzij A, Embden JDA van, Hensen EJ, Cohen IR, Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331: 171. 1988 9. Elias D, Markovitz D, Reshef T, Zee R van der, Cohen IR, Induction and therapy of autoimmune diabetes in the nonobese diabetic (NOD/Lt) mouse by a 65-kDa heat shock protein. Proc Natl Acad Sci USA 87: 1576, 1990 10. Elias D, Reshef T, Birk OS, Zee R van der, Walker MD, Vaccination against autoimmune mouse diabetes with a T celt epitope of the human 65-kDa heat shock protein. Proc Natl Acad Sci USA 87: 3088, 1990 11. Gaston JSH, Heat shock proteins and autoimmunity. Semin Immunol 3: 35, 1991 12. Gaston JSH, Life PF, Jenner PJ, Colston MJ, Bacon PA, Recognition of a mycobacteria-specific epitope in the 65 kD heatshock protein by synovial fluid-derived T cell clones. J Exp Med 171:831, 1990 13. Gaston JSH. Life PF, Jenner PJ, Tonks S, Colston MJ, Bacon PA, How important are responses to mycobacterial hsp65 in rheumatoid arthritis? Br J Rheumatol 30:68s, 1991 14. Gaston JSH, Life PK Zee R van der, Jenner PJ, Colston MJ. Tonks S, Bacon PA, Epitope specificity and MHC restriction of rheumatoid arthritis synovial T cell clones which recognize a mycobacterial 65 kD heat shock protein. Int Immunot 3:965. 1991 15. Gething M-J, Sambrook J, Protein folding in the cell. Nature 355:33, 1991 16. Graeff-Meeder ER de, Zee R van der, Rijkers GT, Schuurman HL Kuis W, Bijlsma JWJ, Zegers BJM, Eden W van, Recognition of human 60 kD heat shock protein by mononuclear cells from patients with juvenile chronic arthritis. Lancet 337: 1368, 1991 17. Haregewoin A, Singh B, Gupta RS, Finberg RW, A mycobacterial heat-shock protein-responsive 76 T cell clone also responds to the homologous human heat-shock protein: a possible link between infection and autoimmunity. J Infect Dis 163: 156, 1991 18. Havlir DV, Ellner JJ, Chervenak KA, Boom WH, Selective expansion of human y6 T cells by monocytes infected with live mycobacterium tuberculosis. J Clin Invest 87:729, 1991 19. Hermann E, Lohse AW, Zee R van der, Eden W van, Mayet WJ, Probsr P, Poralla T, Meyer zum Buschenfelde KH, Fleisher B, Synovial fluid derived Yersinia-reactive T cells responding to human 65 kD heat shock protein and heat stressed antigen presenting ceils. Eur J lmmunol 21:2139, 1991 20. Hogervorst EJM, Schouls L, Wagenaar JPA, Boog CJP, Spaan WJM, Embden JDA van, Eden W van, Modulation of experimental autoimmunity: treatment of adjuvant arthritis by immunization with a recombinant vaccinia virus. Infect lmmun 59:2029, 1991 21~ Janis EM, Kaufmann SHE, Schwartz RH, Pardoll DM, Activation of y6 T ceils in the primary immune response to Mycobacterium tuberculosis. Science 244:713, 1989
94 22. Jindal S, Dudani AK, Harley CB, Singh B, Gupta RS, Primary structure of a human mitochondrial protein homologous to bacterial and plant chaperonins and to the 65 kD mycobacterial antigen. Mol Cell Biol 9:2279, 1989 23. Kaufmann SHE, Schoel B, Embden JDA van, Koga T, WandWurttenberger A, Munk ME, Steinhoff U, Heat shock protein 60: implications for pathogenesis of an protection against bacterial infections. Immunol Rev 121 : 67, 1991 24. Kohashi O, Kohashi Y, Takahshi T, Ozawa A, Shigematsu N, Suppressive effect of Escherichia coli on adjuvant-induced arthritis in germfree rats. Arthritis Rheum 29: 547, 1986 25. Lakomek H-J, Will H, Zech M, Kruskemper HL, A new serological marker in ankylosing spondylitis. Arthritis Rheum 27:961, 1984 26. Lakomek H-J, Ptomann M, Specker C, Schwochau M, Ankylosing spondylitis: an autoimmune disease? Ann Rheum Dis 50:776, 1991 27. Lamb JR, Mendez-Samperio P, Mehlert A, Ivanyi J, So A, Rothbard J, Jindal S, Young RA, Young DB, Stress proteins may provide a link between the immune response to infection and immunity, lnt Immunol 1: 191, 1989 28. Life PF, Bassey EOE, Gaston JSH, T cell recognition of bacterial heat shock proteins in inflammatory arthritis. Immunol Rev 121:113, 1991 29. Minota S, Cameron B, Welch W, Winfield JB, Autoantibodies to the constitutive 73-kD member of the hsp70 family of heat shock proteins in systemic lupus erythematosus. J Exp Med 168: 1475, 1988 30. Minota S, Koyasu S, Yahara I, Winfield JB, Autoantibodies to the heat shock protein hsp90 in systemic lupus erythematosus. J Clin Invest 81:106, 1988 31. O'Brien RL, Happ MP, Dallas A, Cranfill R, Lang J, Fu Y-X, Kubo R, Born W, Recognition of a single hsp-60 epitope by an entire subset ofy6 T lymphocytes. Immunol Rev 121:155, 1991 32. Quayle A J, Black Wison K, Li SG, Oftung F, Kjeldson-Kragh L Shinnick T, Capra D, Forte O, Natvig JB, Cross reactivity of myeobacteriat heat shock protein 65 kD induced T cell clones
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity?
33. 34. 35. 36. 37. 38.
39. 40.
41. 42. 43.
from rheumatoid synovium with human hsp65 kD. Arthritis Rheum 34: $80, 1991 Rajasekar R, Sim G-K, Augustin A, Self heat shock and ~,6 T cell reactivity. Proc Natl Acad Sci USA 87: 1767, 1990 Res PCM, Telgt D, Laar JM van, Pool MO, Breedveld FC, 'Cries RRP de, High antigen reactivity in mononuclear cells from sites of chronic inflammation. Lancet 335: 1406, 1990 Shinnick TM, Heat shock proteins as antigens of bacterial and parasitic pathogens. Curt Top Microbiol Immunol 167:145, 1991 Strober S, Holoshitz J, Mechanisms of immune injury in rheumatoid arthritis: evidence for the involvement of T cells and a heat shock protein. Immunol Rev 118:233, 1990 Tan EM, Autoantibodies in pathology and cell biology. Cell 67: 841,199I Thole JER, Hindersson P, Bruyn J de, Cremers F, Zee J van der, Cock H de, Tommassen L Eden W van, Embden JDA van, Antigenic relatedness of a strongly immunogenic 65 kDa mycobacterial protein antigen with a similarly sized ubiquitous bacterial common antigen. Microbiol Path 4:71, 198 Thompson SJ, Rook GA, Brealey R J, Zee R van der, Elson C J, Autoimmune reactions to heat shock proteins in pristane-induced arthritis. Eur J Immunol 20:2479, 1990 Tsoulfa G, Rook GA, Embden JDA van, Young DB, Mehlert A, Isenberg DA, Hay FC, Lydyard PM, Raised serum IgG and IgA antibodies to mycobacterial antigens in rheumatoid arthritis. Ann Rheum Dis 48:118, 1988 Yang YD, Gasser L Feige U, Prevention of adjuvant arthritis in rats by a nonapeptide from the 65-kD mycobacterial heat shock protein. Clin Exp Immunol 81: 189, 1990 Young RA., Elliott T.I, Stress proteins, infection and immune surveillance. Cell 59: 5, 1989 Zee R van der, Eden W van, Meloen RH, Noordzij A, Embden JDA van, Efficient mapping of a T cell epitope by the simultaneous synthesis of multiple peptides. Eur J Immunol 19:43, 1989
9 Springer-Verlag1992
Are heat shock proteins involved in autoimmunity? J. S. H. Gaston
Department of Rheumatology, University of Birmingham, Birmingham, UK
Summary. Heat shock proteins (HSPs) have been postu-
Characteristics o f autoantigens
lated to be critical antigens in both autoimmune disease and experimental models of autoimmunity. This postulate has been largely based on the remarkable conservation of aminoacid sequence between human and bacterial HSPs, so that it has been argued that immune responses initially directed against the HSP of an infectious agent, would have the potential to initiate or maintain autoimmune disease. This would apply especially to T celt recognition of HSPs, since the T cell focuses on short peptide epitopes within a protein antigen rather than on the antigen's secondary structure. This article critically evaluates the available experimental evidence relating to this hypothesis: although research has clearly highlighted the central role of HSPs in the cellular immune response to pathogenic organisms and has shown the potential for T cell responses directed against self HSPs, a role for self HSPs as major target antigens in autoimmune disease has yet to be firmly established.
The antigenic targets in autoimmune disease fall into two general categories. Firstly, there are antigens which are clearly involved in the disease process, and have tissue specificity, e.g. the acetylcholine receptor in myaesthenia gravis, or the thyroid-stimulating hormone receptor in thyrotoxicosis. The second category involves molecules or molecular complexes which are critical for the normal functioning of every cell. Examples include the protein/ RNA complexes required for mRNA processing and splicing, DNA topoisomerase I, and tRNA transferase enzymes, all of which are targets of particular autoantibodies characteristic of certain connective tissue diseases [37]. In these cases the autoantibodies are excellent markers of disease, but it is difficult to envisage how autoimmune responses to these ubiquitous and indispensible cellular components can result in very specific tissue-restricted disease processes.
Key words: Heat shock protein - Autoimmunity - TceUs
H S P s as putative autoantigens
- Arthritis - Bacterial antigens
Introduction
Heat shock proteins (HSPs) have recently been the subject of extensive discussion in relation to their possible involvement in autoimmune diseases such as rheumatoid arthritis (RA) [36] and insulin-dependent diabetes mellitus [9]. This article will cover some of the experimental evidence which has fuelled this discussion, concentrating on evidence obtained in studies of experimental and clinical arthritis, and will examine the extent to which HSP involvement in autoimmunity has indeed been demonstrated.
HSPs fall into the second category of putative autoantigens. It has often been the case in the past that autoantigens have been identified and discussed well in advance of their cellular functions being worked out, and in some cases patients' sera with high titres of antibodies specific for autoantigens have greatly aided the process of identifying their biochemical nature and function. Material from patients with autoimmune disease has not so far aided investigations into the functions of HSPs, but the physiology of HSPs is currently the focus of much research, and once again those interested in autoimmunity find themselves grappling with the fundamentals of cell biology. The present view is that HSPs are primarily involved in the correct folding, assembly and compartmentalization of intracellular proteins [15]. It is probable that stresses to the cell, such as heat shock, result in increased quantities of denatured or damaged proteins, and thus a greater requirement for HSPs. The increased synthesis of
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity? these proteins in response to stress was used therefore as a defining characteristic of HSPs, but it is now clear that many of the members of different HSP families are required at all times for normal cellular functioning, and indeed many HSPs are synthesized constitutively by cells. As with other autoantibodies to cellular constituents, a connection cannot be easily made between interference with the physiological function of HSPs and the occurrence of a disease like arthritis or diabetes. One feature common to the intracellular targets of autoantibodies is that they are usually conserved molecules, i.e. their amino acid sequence does not vary greatly between different species [22, 38]. No doubt the conservation reflects the critical functions which they perform, and which are similar in all cells. For HSPs the conservation is not just confined to eukaryotic organisms (as is the case for molecules involved in, for example, m R N A splicing) but extends to bacteria including Archaebacteria. Conservation of sequence leads to conservation of immunogenic epitopes, particularly those recognized by T cells, since T cell epitopes consist of short linear peptides, and do not depend on the secondary structure of the intact molecule. Thus marked conservation of sequence makes HSPs attractive potential targets for autoimmune recognition, perhaps as a complication of a conventional immune response to bacterial infection [27]. This review will consider the evidence as to whether this actually occurs in human autoimmune disease.
Humorai responses to HSPs in autoimmune disease
Although HSPs can be compared with other autoantigens, such as the splicing RNA/protein complexes, since both perform important intracellular functions, the immune responses they elicit seem quite different. A characteristic of many connective tissue diseases such as systemic lupus erythematosus (SLE) is the occurrence of very high titres of autoantibodies. In contrast, T cell recognition of the same autoantigens has not been emphasized, although it is true that studies o f T cell responses to autoantigens in these diseases are only in their infancy. In direct contrast to this, T cell recognition of HSPs has attracted attention right from the beginning, whilst strikingly high titres of antibody to HSPs are not seen in autoimmune diseases. Nevertheless there have been some observations on antibodies to HSPs in diseases which might have an autoimmune aetiology. Thus antibodies to a Drosophila antigen, which might be a HSP (the sera actually identify a chromosomal puff associated with the heat shock response), were found in 40% of sera from ankylosing spondylitis (AS) patients [25]. Immunoblotting has shown that these AS sera recognize several insect proteins and, although an epitope on one of these is also present on a human protein, it does not appear to be a HSP [26]. Thus an autoantibody response to HSPs has not been demonstrated unequivocally in AS. Antibodies to human HSP 70 and HSP 90 have been described in 30% - 5 0 % of SLE patients [29, 30], albeit in studies of small numbers, but perhaps this is inevitable in view of the multiplicity of autoantibodies seen in SLE. In
91 RA, patients have higher titres of antibody (especially IgG) to the 65-kDa mycobacterial HSP (HSP 65) than disease controls, but the elevations are not very striking, with 75% of patients having antibody levels in the normal range [40]. Furthermore, whether any one these antibodies cross-react with self HSPs is currently unclear. In summary, there is little evidence to suggest that HSPs elicit a major autoantibody response in human disease. T cell recognition of HSPs in autoimmune disease
Immunity to HSPs in infection Prior to discussing the possible role of HSPs as targets of cellular immunity in autoimmune disease, its worth emphasizing that recognition of HSPs by T cells is a major feature of the response to infections, particularly those involving intracellular organisms, whether bacteria or parasites [23, 35, 42]. Why the immune system should single out HSPs from amongst all the foreign antigens present in an infectious organism is unclear; presumably since HSPs are inevitably present in all organisms, and are highly conserved in terms of their structure and sequence, they may be reliable antigens for the immune system to target. Indeed, the conservation of HSPs might mean that encounter with one bacterial pathogen primes the immune system for responding to HSPs of a subsequent infecting species. However, as mentioned above, this preferential targeting of bacterial HSPs would seem to carry with it the opportunity for autoimmunity, because of the similarities between self and bacterial HSPs. A further feature of the response to intracellular organisms is the involvement of T cells which use the 76 receptor [18, 21]; again, a subset of these cells have a preference for responding to HSPs [17, 33], including self HSPs.
HSP-specific T cells and experimental arthritis Much of the evidence for the involvement of HSPs in autoimmunity comes from studies of experimental arthritis [6]. In particular mycobacterial HSP 65 was first implicated in the pathogenesis of adjuvant arthritis in rats by the finding that a T cell clone specific for this antigen was capable of inducing arthritis when transferred to naive rats [7]. More recent work has extended this observation to suggest that immune responses to HSP 65 are involved in several kinds of experimental arthritis, including arthritis induced by streptococcal cell walls, or by non-bacterial agents such as synthetic adjuvants and pristane [6, 11]. The ability of a HSP 65-specific T cell clone to induce arthritis was initially explained on the basis of cross-reactivity between HSP 65 and a component of cartilage proteoglycan [8]. Although this might still be the case, the identity of the cross-reacting component in proteoglycan has proved elusive, and the possibility that the cross-reaction might involve the homologous rat HSP has been considered. Mapping of the epitope within HSP 65 recognized by the arthritogenic T cell clone showed that it was
92 in a region (amino acids 180-188) where there is not great similarity between HSP 65 and the homologous rat HSP 60 [43]. However, despite this lack of similarity, subsequent experiments have produced the interesting finding that murine "),6T cell hybridomas which showed "autoreactivity" (spontaneous interleukin-2 production without antigen challenge), also recognized amino acids 180-188 of HSP 65 [31]. Although they recognized the corresponding peptide from mammalian HSP 60, which would account for their apparent autoreactivity, there was a striking preference for the mycobacterial peptide, since very high concentrations of mammalian HSP 60derived peptide only produced a sub-optimal response. Thus it is not certain that murine HSP 60 is the principal autoantigen recognized by these cells.
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity? 60-specific T cell clone showed cross-recognition of both recombinant human HSP 60 and recombinant mycobacterial HSP 65 [19]. The clone also responded to heattreated autologous mononuclear cells, presumably recognizing HSP 60 expressed by these cells. It is now of great importance to confirm the autoreactivity of this clone by determining the specific epitope common to human and bacterial HSPs to which it responds. Otherwise, the finding in ReA of an immune response to a bacteria-specific epitope in Yersinia or Salmonella HSP 60 would merely be another instance of the immune system's apparent preference for responding to HSPs in combating infection. However, the recognition o f the same target antigen by JA patients' T cells raises the possibility that occult infection might be playing a part in this disorder, most probably in one of the clinical subsets of JA, which is a highly heterogeneous group of arthritides.
HSP-specific T cells and human arthritis
In studies of human arthritis, T cell clones specific for both mycobacterial HSP 65 and HSP 60 of Gram-negative organisms have been isolated from joints; the epitopes in HSP 65 recognized by T cell clones from a patient with reactive arthritis (ReA) [12] and two patients with RA have been defined in detail [14, 32]. In each case there was no evidence of cross-recognition of cartilage proteoglycan or of type II collagen. For clones from the ReA patient, the epitope recognized was in a non-conserved region of the HSP 65 sequence, but those seen by the RA clones isolated in two laboratories were in highly conserved regions. Despite this, recognition of intact human HSP 60 has not been reported, so that it has not been fully established that the self HSPs serve as an autoantigen for these T cell clones. Nevertheless, when peptides from mycobacterial and human HSPs, were tested, the HSP 65-specific clones did respond to peptides containing the human sequence. However in our laboratory there has been a consistent difference of three logs in the concentration of mycobacterial and human peptides needed to stimulate the T cell clones: the bacterial peptides were recognized at nanogram concentrations, where microgram quantities of the human peptide were required (P. E Life, J. S. H. Gaston, in preparation). In juvenile arthritis (JA), there has recently been a report of T cell responses to recombinant human HSP 60 in experiments in which some care was taken to remove possible contaminating Escherichia coli antigens [16]. However, this recognition has not yet been demonstrated at a clonal level. In other studies clones specific for E. coli HSP 60 have been isolated from JA synovial fluid in patients in whom marked T cell recognition of Gramnegative bacteria is seen in synovial fluid (Life et al., submitted for publication). The epitopes seen by these E. coli-specific T cell clones have yet to be determined. The HSP 60 from Gram-negative bacteria is also a target of the synovial T cell response in ReA [19, 28], which is known to be triggered by specific enteric infection. In this disease it is not necessary to postulate autoimmunity, since it is now known that bacterial antigens reach the joint, where they could presumably activate T cell immunity. Nevertheless, one of the reports of a Yersinia HSP
Protective immune responses to H S P s
So far the possibility that HSP might be targets of an autoimmune T cell has been considered. Paradoxically however, induction of a response to mycobacterial HSP 65, in most animal models of arthritis, results not in disease but in protection of the animal from the subsequent induction of arthritis [1]. Interestingly, very similar conclusions have been reached in the non-obese diabetic mouse model of diabetes [10]. In this sense the HSP 65specifc "arthritogenic" rat T cell clone is the exception rather than the rule. Protection is seen not only in classical adjuvant arthritis, but also in streptococcal well waltinduced arthritis, arthritis due to synthetic adjuvant and pristane-induced arthritis [2, 39]. Protection can be obtained by immunization with intact HSP 65, with a peptide containing amino acids 180 - 188 of HSP 65 or infection with recombinant vaccinia virus expressing HSP 65 [20, 41]; most HSP 65-specific clones also confer protection when transferred to naive animals [5], although they have the same epitope specificity as the arthritogenic clone. It has been argued that prior exposure to HSP 65 or to HSP 65-specific T cells allows the immune system to mount a regulatory response, perhaps by recognizing idiotypes expressed by the HSP 65-specific T cells [4]. Such a response could down-regulate subsequent responses to HSP 65, or perhaps alter them qualitatively, e.g. by changing the pattern of lymphokines produced by the antigen-specific T cells. Neither of these possibilities has been conclusively demonstrated. However, the quality of the T cell response to HSP 65 is emphasized by work in pristane-induced arthritis, where both arthritic animals and those protected by prior immunization with HSP 65 showed marked T cell responses both to HSP 65 and to joint extracts [39]. Thus in the protected animals absence of arthritis does not equate with absence of antigen recognition or even absence of autoreactivity (as shown by the response to joint extracts). A physiological parallel of prior immunization with HSP 65 may be the effects of gut flora on susceptibility to arthritis. For both adjuvant and streptococcal cell wallinduced arthritis, disease susceptibility can be decreased
J. S. H Gaston: Are heat shock proteins involved in autoimmunity? by introducing particular organisms into the gut flora of previously germ-free animals [3, 24]. It has not been shown that HSPs are the critical antigens involved in the effect produced by gut flora, b u t this seems possible. The evidence that T cell responses to HSP 65 are normally protective in experimental animals may have an interesting parallel in RA. Early studies in RA suggested that T cell responses to H S P 65 were fairly c o m m o n , particularly in synovial fluid, b u t the interpretation of these findings has been complicated because it has now been clearly shown that r e c o m b i n a n t preparations of mycobacteria HSP 65 c o n t a i n immunogenic antigens from E. coli, including E. coli H S P 60. When this is allowed for, it now seems that T cells responses to HSP 65 are rather rare in RA, and p r o b a b l y less c o m m o n than in the normal p o p u l a t i o n [13, 34]. To highlight the relative rarity of such responses in RA, a recent study of HSP 65-specific synovial T clones from a single DR4 + RA patient who did respond to this antigen showed that the HSP 65 was only recognized when presented by H L A - D P rather than H L A - D R [14]. This is a very unusual pattern of HLA restriction for h u m a n T cell clones, and it is likely that the patient d i d not possess any T ceils which could respond to HSP 65 presented by DR4. This lack of DR4restricted T cells would occur if thymic deletion removed all the T cells whose receptors could recognize HSP 65 presented by DR4. The role of such thymic deletion is to prevent the emergence of T cells reactive with self. Thus D R 4 might generally be associated with a lack of a T ceil response to HSP 65, since it might not always be possible to generate a DP-restricted response instead. One could then argue that an i n a d e q u a t e response to HSP 65 in D R 4 + individuals might predispose to arthritis, if the protective effects of responses to HSP 65 in animal models of arthritis applied to arthritis in man.
Conclusion There is a considerable a m o u n t o f evidence to suggest that HSPs are " i n v o l v e d " in a u t o i m m u n i t y , but the details of this involvement are still far from clear. The evidence that HSPs are p r i m a r y target antigens of an autoi m m u n e response is limited, despite the enormous opportunities for such autoreactivity afforded by their marked conservation of sequence and their capacity to elicit T cell responses. The i m m u n e system's previous experience of HSPs seems to play a significant part in determining susceptibility to certain a u t o i m m u n e conditions.
References 1. Billingham MEL Carney S, Butler R, Colston ML A mycobacterial 65-kD heat shock protein induces antigen-specific suppression of adjuvant arthritis but is not itself arthritogenic. J Exp Med 171:339, 1990 2. Broek MF van den, Hogervorst EJM, Bruggen MCJ van, Eden W van, Zee R van der, Berg WB van den, Protection against streptococcal cell-wall induced arthritis by pre-treatment with the 65-kD mycobacterial heat shock protein. J Exp Med 170, 449. 1989
93 3. Broek MF van den, Bruggen MCJ van, Koopman JP, Hazenberg MP, Berg W van den, Gut flora induces and maintains resistance against streptococcal cell wall-induced arthritis. Clin Exp Immuno/(in press) 4. Cohen IR, Young DB, The immune system's view of invading micro-organisms, autoimmunity and the immunological homunculus. Immunol Today 12:105, 1991 5. Cohen IR, Holoshitz L Eden W van, Frenkel A, T lymphocyte clones illuminate pathogenesis and affect therapy of experimental arthritis. Arthritis Rheum 28: 841, 1985 6. Eden W van. Heat shock proteins as immunogenic bacterial antigens with the potential to induce and regulate autoimmune arthritis. Immunol Rev 121: 1, 1991 7. Eden W van, Holoshitz L Nero Z, Frenkel A, Klajman A, Cohen IR, Arthritis induced by a T lymphocyte clone that responds to Mycobacterium tuberculosis and to cartilage proteoglycans. Proc Nat! Acad Sci USA 82: 5117, 1985 8. Eden W van, Thole JER, Zee R van der, Noordzij A, Embden JDA van, Hensen EJ, Cohen IR, Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331: 171. 1988 9. Elias D, Markovitz D, Reshef T, Zee R van der, Cohen IR, Induction and therapy of autoimmune diabetes in the nonobese diabetic (NOD/Lt) mouse by a 65-kDa heat shock protein. Proc Natl Acad Sci USA 87: 1576, 1990 10. Elias D, Reshef T, Birk OS, Zee R van der, Walker MD, Vaccination against autoimmune mouse diabetes with a T celt epitope of the human 65-kDa heat shock protein. Proc Natl Acad Sci USA 87: 3088, 1990 11. Gaston JSH, Heat shock proteins and autoimmunity. Semin Immunol 3: 35, 1991 12. Gaston JSH, Life PF, Jenner PJ, Colston MJ, Bacon PA, Recognition of a mycobacteria-specific epitope in the 65 kD heatshock protein by synovial fluid-derived T cell clones. J Exp Med 171:831, 1990 13. Gaston JSH. Life PF, Jenner PJ, Tonks S, Colston MJ, Bacon PA, How important are responses to mycobacterial hsp65 in rheumatoid arthritis? Br J Rheumatol 30:68s, 1991 14. Gaston JSH, Life PK Zee R van der, Jenner PJ, Colston MJ. Tonks S, Bacon PA, Epitope specificity and MHC restriction of rheumatoid arthritis synovial T cell clones which recognize a mycobacterial 65 kD heat shock protein. Int Immunot 3:965. 1991 15. Gething M-J, Sambrook J, Protein folding in the cell. Nature 355:33, 1991 16. Graeff-Meeder ER de, Zee R van der, Rijkers GT, Schuurman HL Kuis W, Bijlsma JWJ, Zegers BJM, Eden W van, Recognition of human 60 kD heat shock protein by mononuclear cells from patients with juvenile chronic arthritis. Lancet 337: 1368, 1991 17. Haregewoin A, Singh B, Gupta RS, Finberg RW, A mycobacterial heat-shock protein-responsive 76 T cell clone also responds to the homologous human heat-shock protein: a possible link between infection and autoimmunity. J Infect Dis 163: 156, 1991 18. Havlir DV, Ellner JJ, Chervenak KA, Boom WH, Selective expansion of human y6 T cells by monocytes infected with live mycobacterium tuberculosis. J Clin Invest 87:729, 1991 19. Hermann E, Lohse AW, Zee R van der, Eden W van, Mayet WJ, Probsr P, Poralla T, Meyer zum Buschenfelde KH, Fleisher B, Synovial fluid derived Yersinia-reactive T cells responding to human 65 kD heat shock protein and heat stressed antigen presenting ceils. Eur J lmmunol 21:2139, 1991 20. Hogervorst EJM, Schouls L, Wagenaar JPA, Boog CJP, Spaan WJM, Embden JDA van, Eden W van, Modulation of experimental autoimmunity: treatment of adjuvant arthritis by immunization with a recombinant vaccinia virus. Infect lmmun 59:2029, 1991 21~ Janis EM, Kaufmann SHE, Schwartz RH, Pardoll DM, Activation of y6 T ceils in the primary immune response to Mycobacterium tuberculosis. Science 244:713, 1989
94 22. Jindal S, Dudani AK, Harley CB, Singh B, Gupta RS, Primary structure of a human mitochondrial protein homologous to bacterial and plant chaperonins and to the 65 kD mycobacterial antigen. Mol Cell Biol 9:2279, 1989 23. Kaufmann SHE, Schoel B, Embden JDA van, Koga T, WandWurttenberger A, Munk ME, Steinhoff U, Heat shock protein 60: implications for pathogenesis of an protection against bacterial infections. Immunol Rev 121 : 67, 1991 24. Kohashi O, Kohashi Y, Takahshi T, Ozawa A, Shigematsu N, Suppressive effect of Escherichia coli on adjuvant-induced arthritis in germfree rats. Arthritis Rheum 29: 547, 1986 25. Lakomek H-J, Will H, Zech M, Kruskemper HL, A new serological marker in ankylosing spondylitis. Arthritis Rheum 27:961, 1984 26. Lakomek H-J, Ptomann M, Specker C, Schwochau M, Ankylosing spondylitis: an autoimmune disease? Ann Rheum Dis 50:776, 1991 27. Lamb JR, Mendez-Samperio P, Mehlert A, Ivanyi J, So A, Rothbard J, Jindal S, Young RA, Young DB, Stress proteins may provide a link between the immune response to infection and immunity, lnt Immunol 1: 191, 1989 28. Life PF, Bassey EOE, Gaston JSH, T cell recognition of bacterial heat shock proteins in inflammatory arthritis. Immunol Rev 121:113, 1991 29. Minota S, Cameron B, Welch W, Winfield JB, Autoantibodies to the constitutive 73-kD member of the hsp70 family of heat shock proteins in systemic lupus erythematosus. J Exp Med 168: 1475, 1988 30. Minota S, Koyasu S, Yahara I, Winfield JB, Autoantibodies to the heat shock protein hsp90 in systemic lupus erythematosus. J Clin Invest 81:106, 1988 31. O'Brien RL, Happ MP, Dallas A, Cranfill R, Lang J, Fu Y-X, Kubo R, Born W, Recognition of a single hsp-60 epitope by an entire subset ofy6 T lymphocytes. Immunol Rev 121:155, 1991 32. Quayle A J, Black Wison K, Li SG, Oftung F, Kjeldson-Kragh L Shinnick T, Capra D, Forte O, Natvig JB, Cross reactivity of myeobacteriat heat shock protein 65 kD induced T cell clones
J. S. H. Gaston: Are heat shock proteins involved in autoimmunity?
33. 34. 35. 36. 37. 38.
39. 40.
41. 42. 43.
from rheumatoid synovium with human hsp65 kD. Arthritis Rheum 34: $80, 1991 Rajasekar R, Sim G-K, Augustin A, Self heat shock and ~,6 T cell reactivity. Proc Natl Acad Sci USA 87: 1767, 1990 Res PCM, Telgt D, Laar JM van, Pool MO, Breedveld FC, 'Cries RRP de, High antigen reactivity in mononuclear cells from sites of chronic inflammation. Lancet 335: 1406, 1990 Shinnick TM, Heat shock proteins as antigens of bacterial and parasitic pathogens. Curt Top Microbiol Immunol 167:145, 1991 Strober S, Holoshitz J, Mechanisms of immune injury in rheumatoid arthritis: evidence for the involvement of T cells and a heat shock protein. Immunol Rev 118:233, 1990 Tan EM, Autoantibodies in pathology and cell biology. Cell 67: 841,199I Thole JER, Hindersson P, Bruyn J de, Cremers F, Zee J van der, Cock H de, Tommassen L Eden W van, Embden JDA van, Antigenic relatedness of a strongly immunogenic 65 kDa mycobacterial protein antigen with a similarly sized ubiquitous bacterial common antigen. Microbiol Path 4:71, 198 Thompson SJ, Rook GA, Brealey R J, Zee R van der, Elson C J, Autoimmune reactions to heat shock proteins in pristane-induced arthritis. Eur J Immunol 20:2479, 1990 Tsoulfa G, Rook GA, Embden JDA van, Young DB, Mehlert A, Isenberg DA, Hay FC, Lydyard PM, Raised serum IgG and IgA antibodies to mycobacterial antigens in rheumatoid arthritis. Ann Rheum Dis 48:118, 1988 Yang YD, Gasser L Feige U, Prevention of adjuvant arthritis in rats by a nonapeptide from the 65-kD mycobacterial heat shock protein. Clin Exp Immunol 81: 189, 1990 Young RA., Elliott T.I, Stress proteins, infection and immune surveillance. Cell 59: 5, 1989 Zee R van der, Eden W van, Meloen RH, Noordzij A, Embden JDA van, Efficient mapping of a T cell epitope by the simultaneous synthesis of multiple peptides. Eur J Immunol 19:43, 1989
