- Email: [email protected]
Possible Role of Bacterial Heat-Shock Proteins in Autoimmune Diseases
Zbl. Bakt. 278, 377-382 (1993) © Gustav Fischer Verlag, StuttgartlNew York
Possible Role of Bacterial Heat-Shock Proteins in Autoimmune Diseases W: VAN EDEN Department of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, University of Utrecht, 3508 TD Utrecht, The Netherlands
Summary The immune system is continuously exposed to a variety of microbial antigens. It is unavoidable that such antigens may resemble auto antigens, being present in the host. It has appeared now that the immune system is accepting recognition of such "mimicry" antigens and even may have developed a tendency to focus its attention to such antigens. This paper discusses the findings of immune responses to heat-shock proteins, as a family of proteins exhibiting extraordinary sequence conservation, creating a high degree of similarity between bacterial antigens and host antigens. On the basis of existing evidence, it is argued that immune responses directed at hsps are part of regulatory mechanisms which enable the immune system to safely contain potentially self-reactive lymphocytes. Furthermore, it is argued that by means of artificial immunization against hsps or selected determinants of hsps, a development of autoimmune diseases may be inhibited.
Zusammenfassung Das Immunsystem ist standig einer Vielzahl von mikrobiellen Antigenen ausgesetzt. Es ist unvermeidbar, dag solche Antigene auch Autoantigenen ahnlich sein konnen, die als wirtseigene Komponenten vorkommen. Es scheint so zu sein, dag das Immunsystem solche "Mimicry"-Antigene erkennt und sogar die Tendenz entwickelt hat, eine besondere Aufmerksamkeit darauf zu richten. Die vorliegende Arbeit diskutiert die Ergebnisse der Immunantwort auf Hitzeschockproteine. Diese gelten als eine Familie von Proteinen, die eine besondere Sequenzkonservierung aufweisen und in einem sehr hohen Mage wirtseigenen Antigenen ahneln. Auf der Grundlage der existierenden Daten wird argumentiert, dag die Immunantwort auf Hitzeschockproteine ein Teil des regulatorischen Mechanismus ist, welcher das Immunsystem befahigt, potentielle selbstreagierende Lymphozyten sicher zu enthalten. Es wird weiter argumentiert, dag mit Hilfe einer kiinstlichen Immunisierung gegen Hitzeschockproteine oder bestimmte ihrer Determinanten die Entwicklung einer Autoimmunkrankheit verhindert werden kann.
Autoimmune diseases are thought to result from immune responses inappropriately targeted at self-antigens. The origin of such inappropriately targeted responses has been a central theme in immunological research for long. Originally it was thought that according to Burnet, the immune repertoire was the composite of those clones that had
378
W. van Eden
passed "clonal selection". It was assumed that during the clonal selection process, the clones with a potential for self-recognition were to be deleted, which would imply that the adult repertoire would be devoid of clones (T and B lymphocytes) with the potential of autoreactivity. However, more recent technical advances in cellul~r cloning and antigen definition have led to findings that showed potentially autoreactive clones to be present in the healthy repertoire. These experimental advances have shifted the scientific debate from autoimmunity on the basis of the inadvertent action of the so-called "forbidden" (falsely non-deleted) clones to regulatory control of potentially harmful cells. The real issue now is which mechanisms are involved in the safe containment of dangerously selfreactive cells within the healthy immunological repertoire. It is well possible that the continued exposure to microbial non-self antigens is one of the forces implicated in this mechanism of containment, as we will discuss now. Autoimmune diseases such as arthritis, can be induced in experimental animals by immunization to microbial antigens. Alternatively, it has been shown that exposure to environmental infectious organisms may raise resistance to induced autoimmune diseases. Recently, evidence has accumulated that a group of evolutionary conserved proteins, present similarly in bacterial and in mammalian host cells, called heat-shock or stress proteins, can be intimately involved in both induction of diseases and protective resistance to such diseases. Furthermore, the recently discovered group of bacterial "superantigens", which can interact directly with non-variable parts of T-cell receptor V gene products, can cause major shifts in the T-cell repertoire, which theoretically, may also lead to changed suseptibility to autoimmune diseases. Further understanding of the role of microbial antigens in autoimmunity is supposed to lead to the development of novel immunological strategies for the specific therapeutic intervention in these diseases. Despite their frequent associations with MHC alleles and their frequent clustering within families, autoimmune diseases are not inherited diseases. The same can be said for infection: although autoimmune diseases may occur in association with infectious problems, they are certainly not infectious diseases. The associations with both genes and infection are likely to be of an indirect nature: in its development, the immune system is determined at large by both the genetic make-up of the individual and the experience of the system in its continuous contact with the environment. It may be selfevident that during evolution the system not only has developed in order to combat the environment or outside world of potentially aggressive microbial invaders, but also has learned to exploit the information of the outside non-self in order to improve its capacity to maintain self-integrity. The latter may explain the observations that microbial priming may protect against autoimmune diseases, despite the fact that under different conditions, similar microbial priming may elicit disease. Especially in the case of arthritis, infection is known to be capable of precipitating disease. Lyme arthritis is a chronic progressive arthritis barely distinct from classic reactive arthritis (ReA), induced by infection with the spirochaete, Borrelia burgdorferi. ReA is by definition a reactive event consequent to bacterial infection. Interestingly, in this case there is a very prominent association with HLA-B27. The fact that the same strong association is found with ankylosing spondylitis (AS) implicates that also AS is likely to be the result of bacterial immunization. In experimental animals, arthritis can be induced by bacterial infection, but also by immunization with bacterial antigens such as streptococcal cell walls (SCW) or heat-killed mycobacteria. In the latter model of mycobacteriainduced adjuvant arthritis in the rat (AA), the analysis of arthritogenic T-cell clones has revealed that responses directed at the mycobacterial 65 kD heat-shock protein (hsp65)
Bacterial Heat-Shock Proteins in Autoimmune Diseases
379
may be crucial for the development of disease (13). Heat-shock proteins are uniquely conserved proteins, due to their essential functional significance for living cells. They have been termed also "chaperonins", because of their chaperon-like activity in protecting the integrity of intracellular proteins during stressful events (12). Besides this, heat-shock proteins have been described to be involved in synthesis and assembly of complex multimeric proteins and the translocation of proteins over intracellular membranes. Due to the evolutionary conservation, hsp65 contains multiple amino acid sequences that are identical between the mycobacterial and the mammalian counterpart. Theoretically this may confound the immune response in properly distinguishing between the bacterial antigen and the host counterpart, with unintended autoimmune reactivity as a result. Fine mapping of the epitope seen by the arthritogenic T cells in the rat, revealed, however, that this epitope is of a relatively non-conserved nature and furthermore that this epitope may have a significant sequence homology with a cartilage proteoglycan protein (13). Thus, so far, in the experimental AA model no direct evidence is available for pernicious responsiveness directed at conserved sequences present in the rat endogenous hsp65. The hsp65 of the Lewis rat, however, is currently being cloned and will be used for the monitoring of T-cell responses during AA development. In contrast to this, it has been demonstrated in humans that T cells do respond to conserved hsp65 sequences. Both T helper cells and cytoxic T lymphocytes have been generated with specificity for conserved sequences and such cells have been shown to recognize autologous cells expressing hsp65. Furthermore, in a case of Yersinia-associated reactive arthritis, T cells have been collected from the synovial fluid of the inflamed joint that responded not only to Yersinia bacteria, but also to the mycobacterial and human hsp65 (5). Thus, in human arthritis, T cells responding to conserved areas of endogenous hsp65 may be found at the site of inflammation. Similar reactivity against the human endogenous hsp65 has been demonstrated recently to be present in the joints of children suffering from juvenile chronic arthritis UCA) (4). Interestingly, synovial biopsies taken from the same patients revealed a raised expression of hsp65 in the synovial lining cells (1). Thus, the presence of both responsive cells and the relevant antigen have been shown to coincide with the expression of disease. In cases of advanced rheumatoid arthritis in adults, no such T cell responsiveness to the human hsp65 was seen, despite the earlier reports on responses to the mycobacterial hsp65 in RA (8). It is possible that in advanced erosive RA, the responses to endogenous hsp65 have subsided and that responses directed at cartilage-related epitopes have taken over, caused by cartilage destruction. Since lCA usually is a non-erosive spontaneously remitting disease, one may speculate that T cell reponses directed at hsps are part of an adequate regulatory response that ultimately supports disease remission. That responses directed at hsp65 may have protective qualities has been demonstrated in various models of arthritis, such as AA, SCW-arthritis, pristane (oil) induced arthritis (10) and collagen type II-induced arthritis. In all these models, including the non-bacteria associated models (pristance and collagen), responses to hsp65 evolve along with arthritis development. This indicates that during arthritis, irrespective of the trigger leading to the disease, responses to hsp65 are stimulated, possibly through the raised expression of the molecule in the synovial tissues. Preimmunization with mycobacterial hsp65 prior to arthritis induction has been found to induce resistance against disease in the same models. Altogether, we may conclude that preexposure to hsp65 is triggering protective mechanisms, which apparently are capable of controlling disease when the potentially arthritogenic stimulus is introduced. Very similar to the phenomena de-
380
W. van Eden
scribed for arthritis, in the case of diabetes in the NOD mouse, hsp65 immunization may lead to an early appearance of disease, while preimmunization with a selected hsp65 peptide was found to create resistance to destruction of pancreatic beta-cells and expression of diabetes (2). Moreover, in BB rats, the occurrence of spontaneous diabetes can be prevented by the early immunization to mycobacteria in the form of CFA (9). This seemingly prominent recognition of hsp65 in autoimmune disease states may be interpreted as being suggestive of a preoccupation of the immune system with hsps, both with respect to endogenously expressed hsp and with respect to the microbial counterparts. It is possible that the immune system sets a focus on the recognition of these proteins, which seems profitable both in terms of carefully controlling autoimmune reactivity and in terms of controlling the unattended in invasion by bacterial and other parasitic invaders (3). A well balanced and carefully controlled recognition of these proteins may then add to resistance to infection and at the same time to resistance to autoimmune disease. That the immune system is paying ample attention to hsps may be concluded from observations that hsps are dominant bacterial antigens. In the case of hsp65, it has been demonstrated that after mycobacterial priming in mice at the polyclonal level, up to 20% of the responding T cells may see this particular hsp (7). Thus, despite the extensive homology with self, hsps tend to be very immunogenic. Furthermore, responses directed at presumably risky epitopes, such as the epitope 180-188 of hsp65 which is recognized by the arthritogenic T cells in the Lewis rat, seem to be under rather strict regulatory control. This became especially evident in analyzing Fisher rats (6). Fisher rats are relatively resistant to AA induction, despite the fact that their MHC type, at least for class II molecules, is fully identical to that of the Lewis rat. Despite this MHC identity, Fisher rats are, in contrast to Lewis rats, non-responders for epitope 180-188. However, in rare cases of AA in Fisher rats, responses to 180-188 did occur. Thus, in Fisher rats non-responsiveness to 180-188 went together with arthritis resistance. It is possible that Fisher rats have a more efficient T cell network control of responding to this arthritis-associated epitope. This possibility seems to be supported by the observations made in germ-free Fisher rats. Fisher rats, bred germ-free from birth, are susceptible to AA and SCW arthritis as are Lewis rats. The resistance, however, develops as soon as colonization with bacterial flora or just E. coli has taken place, showing that exposure to environmental bacterial antigens in itself is providing the Fisher rats with the means of controlling responses to potentially dangerous epitopes. In line with this are findings in Lewis rats housed under conditions with a variable microbial load in the environment. It is well known that arthritis is more severe in animals kept in specific pathogen-free or otherwise relatively clean facilities, as compared to the severity of arthritis seen in animals exposed to an environment replete with potentially pathogenic microorganisms. Probably, the infectious pressure is keeping the immune system on the alert and provides the animal with immunological experience that helps to control autoimmunity. The same has been observed for spontaneous diabetes in the NOD mouse: while in the conventional setting, 50-80% of the (female) animals developed arthritis, the frequency reached 100% under germ-free or specific pathogen-free conditions (11). In line with this seem to be findings that resistance to the development of a relapse of AA in Lewis rats upon reimmunization with mycobacteria also depends on the environmental infectious pressure (I. R. Cohen, pers. comm.). Animals kept in a clean environment were found to be less resistant to such relapses. In these animals, evidence was obtained a less well developed network control of anti-hsp65 T cell (anti-MI)
Bacterial Heat-Shock Proteins in Autoimmune Diseases
381
immunity (3). That immunity to critical antigens is evolving with time in the conventional environment has been shown also in the pristane model of arthritis in CBAlIgb mice (10). In the naive normal mice, T-cell responses were seen to develop spontaneously early in life, being maximal at the age of 2-4 months. After that period, a significant drop was observed, with a very low level of responsiveness from the age of 8 months onwards. This paralleled pristane arthritis susceptibility in an inverse manner. At the age of 8 months, 40% of the animals appeared susceptible, while at the age of 3-4 months only 1-10% of the animals developed disease. Mice kept germ-free, however, were all susceptible at the early ages, while their anti-hsp65 T cell responses were found to be very low. All such observations seem to be in concert with the proposition that immunity to critical environmental antigens protects against the development of autoimmune diseases. This all may well open perspectives for the use of such antigens in artificially optimizing this particular state of immunization by means similar to vaccination as used to raise anti-infectious immunity. Much, however, will depend on our ability to avoid the boostering of undesired reactivity to risky epitopes. The experience obtained in the model system of arthritis has teached us so far that prevention of disease may be a relatively easy and realistic goal to reach. Reduction of the ongoing disease process, however, will be more difficult to achieve. For this, further understanding of the cellular networks or interactions that ensure the maintenance of self-tolerance seems to be necessary. Nevertheless, the molecular approach in bacterial organisms has provided us not only with antigens but also with antigenic fragments or epitopes thereof that playa role in the regulation of self-tolerance. Together with a further understanding of the cellular interactions within the immune system, it is possible that in the end we will have to conclude that absent or improper immunity to critical bacterial antigens such as heat-shock proteins, may also playa role in the "molecular pathogenesis of bacteria", in this case leading to autoimmune diseases.
References
1. Boog, C. J. P., E. R. de Graeff-Meeder, M. A. Lucassen, R. van der Zee, M. M. Voorhorst-Ogink, P. J. s. van Kooten, H. J. Geuze, and W. van Eden: Two monoclonal antibodies generated against human hsp60 show reactivity with synovial membranes of patients with juvenile chronic arthritis. J. Exp. Med., 175 (1992) 1805-1810 2. Cohen, I. R.: Autoimmunity to chaperonins in the pathogenesis of arthritis and diabetes. Ann. Rev. Immunol. 9 (1991) 567-589 3. Cohen, I. R. and D. B. Young: Autoimmunity, microbial immunity and the immunological homunculus. Immunol. Today 12 (1991) 105-110 4. De Graeff-Meeder, E. R., R. Van der Zee, G. T. Rijkers, H. J. Schuurman, W. Kuis, J. W,J. Bijlsma, B.J.M. Zegers, and W. Van Eden: Recognition of human 60kD heat shock protein by mononuclear cells from patients with juvenile chronic arthritis. Lancet 337 (1991) 1368-1372 5. Hermann, E., A. W. Lohse, R. Van der Zee, W. Van Eden, W. J. Mayet, P. Probst, T. Poral/a, K. H. Meyer zum Buschenfelde, and B. Fleischer: Synovial fluid derived T cells responding to human 65kD heat-shock protein and heat-stressed antigen presenting cells. Eur. J. Immunol. 21 (1991) 2139-2143 6. Hogervorst, E. J. M., C. J. P. Boog, J. P. A. Wagenaar, M. H. M. Wauben, R. van der Zee, and W. van Eden: T cell reactivity to an epitope of the mycobacterial 65kD heat shock protein (hsp65) corresponds with arthritis susceptibility in rats and is regulated by hsp65 specific cellular responses. Eur. J. Immunol. 21 (1991) 1289-1296
382
W. van Eden
7. Kaufmann, S. H. E.: Heat shock proteins and the immune response. Immunol. Today 11 (1990) 129-136 8. Res, P. C. M., C. G. Schaar, F. C. Breedveld, W. van Eden, J. D. A. van Embden, I. R. Cohen and R. R. P. de Vries: Synovial fluid T cell reactivity against the 65 kD heatshock protein of mycobacteria in early onset of chronic arthritis. Lancet ii (1988) 478-481 9. Sadelain, M. W. j., H. Y. Qin, W. Sumoski, N. Parfrey, B.Singh, and A. Rabinovitch: Prevention of diabetes in the BB rat by early immunotherapy using Freund's adjuvant. J. Autoimmunity 3 (1990) 671-680 " 10. Thompson, S. j., G. A. W. Rook, R. j. Brealey, R. van der Zee, and C. j. Elson: Autoimmune reactions to heat shock proteins in pristane induced arthritis. Eur. J. Immunol. 20 (1990) 2479-2483 11. Todd, j. A.: Genetic control of autoimmunity in type 1 diabetes. Immunol. Today 11 (1990) 122~ 129 12. Van Eden, W.: Heat shock proteins as immunogenic bacterial antigens with the potential of inducing and regulating autoimmune arthritis. Immunol. Rev. 121 (1991) 5-28 13. Van Eden, W., j. E. R. Thole, R. van der Zee, A. Noordzij, j. D. A. van Embden, E. J. Hensen, and I. R. Cohen: Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331 (1988) 171-173
Dr. W. van Eden, Dept. of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, University of Utrecht, Yalelaan 1, 3508 TD Utrecht, The Netherlands
Possible Role of Bacterial Heat-Shock Proteins in Autoimmune Diseases W: VAN EDEN Department of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, University of Utrecht, 3508 TD Utrecht, The Netherlands
Summary The immune system is continuously exposed to a variety of microbial antigens. It is unavoidable that such antigens may resemble auto antigens, being present in the host. It has appeared now that the immune system is accepting recognition of such "mimicry" antigens and even may have developed a tendency to focus its attention to such antigens. This paper discusses the findings of immune responses to heat-shock proteins, as a family of proteins exhibiting extraordinary sequence conservation, creating a high degree of similarity between bacterial antigens and host antigens. On the basis of existing evidence, it is argued that immune responses directed at hsps are part of regulatory mechanisms which enable the immune system to safely contain potentially self-reactive lymphocytes. Furthermore, it is argued that by means of artificial immunization against hsps or selected determinants of hsps, a development of autoimmune diseases may be inhibited.
Zusammenfassung Das Immunsystem ist standig einer Vielzahl von mikrobiellen Antigenen ausgesetzt. Es ist unvermeidbar, dag solche Antigene auch Autoantigenen ahnlich sein konnen, die als wirtseigene Komponenten vorkommen. Es scheint so zu sein, dag das Immunsystem solche "Mimicry"-Antigene erkennt und sogar die Tendenz entwickelt hat, eine besondere Aufmerksamkeit darauf zu richten. Die vorliegende Arbeit diskutiert die Ergebnisse der Immunantwort auf Hitzeschockproteine. Diese gelten als eine Familie von Proteinen, die eine besondere Sequenzkonservierung aufweisen und in einem sehr hohen Mage wirtseigenen Antigenen ahneln. Auf der Grundlage der existierenden Daten wird argumentiert, dag die Immunantwort auf Hitzeschockproteine ein Teil des regulatorischen Mechanismus ist, welcher das Immunsystem befahigt, potentielle selbstreagierende Lymphozyten sicher zu enthalten. Es wird weiter argumentiert, dag mit Hilfe einer kiinstlichen Immunisierung gegen Hitzeschockproteine oder bestimmte ihrer Determinanten die Entwicklung einer Autoimmunkrankheit verhindert werden kann.
Autoimmune diseases are thought to result from immune responses inappropriately targeted at self-antigens. The origin of such inappropriately targeted responses has been a central theme in immunological research for long. Originally it was thought that according to Burnet, the immune repertoire was the composite of those clones that had
378
W. van Eden
passed "clonal selection". It was assumed that during the clonal selection process, the clones with a potential for self-recognition were to be deleted, which would imply that the adult repertoire would be devoid of clones (T and B lymphocytes) with the potential of autoreactivity. However, more recent technical advances in cellul~r cloning and antigen definition have led to findings that showed potentially autoreactive clones to be present in the healthy repertoire. These experimental advances have shifted the scientific debate from autoimmunity on the basis of the inadvertent action of the so-called "forbidden" (falsely non-deleted) clones to regulatory control of potentially harmful cells. The real issue now is which mechanisms are involved in the safe containment of dangerously selfreactive cells within the healthy immunological repertoire. It is well possible that the continued exposure to microbial non-self antigens is one of the forces implicated in this mechanism of containment, as we will discuss now. Autoimmune diseases such as arthritis, can be induced in experimental animals by immunization to microbial antigens. Alternatively, it has been shown that exposure to environmental infectious organisms may raise resistance to induced autoimmune diseases. Recently, evidence has accumulated that a group of evolutionary conserved proteins, present similarly in bacterial and in mammalian host cells, called heat-shock or stress proteins, can be intimately involved in both induction of diseases and protective resistance to such diseases. Furthermore, the recently discovered group of bacterial "superantigens", which can interact directly with non-variable parts of T-cell receptor V gene products, can cause major shifts in the T-cell repertoire, which theoretically, may also lead to changed suseptibility to autoimmune diseases. Further understanding of the role of microbial antigens in autoimmunity is supposed to lead to the development of novel immunological strategies for the specific therapeutic intervention in these diseases. Despite their frequent associations with MHC alleles and their frequent clustering within families, autoimmune diseases are not inherited diseases. The same can be said for infection: although autoimmune diseases may occur in association with infectious problems, they are certainly not infectious diseases. The associations with both genes and infection are likely to be of an indirect nature: in its development, the immune system is determined at large by both the genetic make-up of the individual and the experience of the system in its continuous contact with the environment. It may be selfevident that during evolution the system not only has developed in order to combat the environment or outside world of potentially aggressive microbial invaders, but also has learned to exploit the information of the outside non-self in order to improve its capacity to maintain self-integrity. The latter may explain the observations that microbial priming may protect against autoimmune diseases, despite the fact that under different conditions, similar microbial priming may elicit disease. Especially in the case of arthritis, infection is known to be capable of precipitating disease. Lyme arthritis is a chronic progressive arthritis barely distinct from classic reactive arthritis (ReA), induced by infection with the spirochaete, Borrelia burgdorferi. ReA is by definition a reactive event consequent to bacterial infection. Interestingly, in this case there is a very prominent association with HLA-B27. The fact that the same strong association is found with ankylosing spondylitis (AS) implicates that also AS is likely to be the result of bacterial immunization. In experimental animals, arthritis can be induced by bacterial infection, but also by immunization with bacterial antigens such as streptococcal cell walls (SCW) or heat-killed mycobacteria. In the latter model of mycobacteriainduced adjuvant arthritis in the rat (AA), the analysis of arthritogenic T-cell clones has revealed that responses directed at the mycobacterial 65 kD heat-shock protein (hsp65)
Bacterial Heat-Shock Proteins in Autoimmune Diseases
379
may be crucial for the development of disease (13). Heat-shock proteins are uniquely conserved proteins, due to their essential functional significance for living cells. They have been termed also "chaperonins", because of their chaperon-like activity in protecting the integrity of intracellular proteins during stressful events (12). Besides this, heat-shock proteins have been described to be involved in synthesis and assembly of complex multimeric proteins and the translocation of proteins over intracellular membranes. Due to the evolutionary conservation, hsp65 contains multiple amino acid sequences that are identical between the mycobacterial and the mammalian counterpart. Theoretically this may confound the immune response in properly distinguishing between the bacterial antigen and the host counterpart, with unintended autoimmune reactivity as a result. Fine mapping of the epitope seen by the arthritogenic T cells in the rat, revealed, however, that this epitope is of a relatively non-conserved nature and furthermore that this epitope may have a significant sequence homology with a cartilage proteoglycan protein (13). Thus, so far, in the experimental AA model no direct evidence is available for pernicious responsiveness directed at conserved sequences present in the rat endogenous hsp65. The hsp65 of the Lewis rat, however, is currently being cloned and will be used for the monitoring of T-cell responses during AA development. In contrast to this, it has been demonstrated in humans that T cells do respond to conserved hsp65 sequences. Both T helper cells and cytoxic T lymphocytes have been generated with specificity for conserved sequences and such cells have been shown to recognize autologous cells expressing hsp65. Furthermore, in a case of Yersinia-associated reactive arthritis, T cells have been collected from the synovial fluid of the inflamed joint that responded not only to Yersinia bacteria, but also to the mycobacterial and human hsp65 (5). Thus, in human arthritis, T cells responding to conserved areas of endogenous hsp65 may be found at the site of inflammation. Similar reactivity against the human endogenous hsp65 has been demonstrated recently to be present in the joints of children suffering from juvenile chronic arthritis UCA) (4). Interestingly, synovial biopsies taken from the same patients revealed a raised expression of hsp65 in the synovial lining cells (1). Thus, the presence of both responsive cells and the relevant antigen have been shown to coincide with the expression of disease. In cases of advanced rheumatoid arthritis in adults, no such T cell responsiveness to the human hsp65 was seen, despite the earlier reports on responses to the mycobacterial hsp65 in RA (8). It is possible that in advanced erosive RA, the responses to endogenous hsp65 have subsided and that responses directed at cartilage-related epitopes have taken over, caused by cartilage destruction. Since lCA usually is a non-erosive spontaneously remitting disease, one may speculate that T cell reponses directed at hsps are part of an adequate regulatory response that ultimately supports disease remission. That responses directed at hsp65 may have protective qualities has been demonstrated in various models of arthritis, such as AA, SCW-arthritis, pristane (oil) induced arthritis (10) and collagen type II-induced arthritis. In all these models, including the non-bacteria associated models (pristance and collagen), responses to hsp65 evolve along with arthritis development. This indicates that during arthritis, irrespective of the trigger leading to the disease, responses to hsp65 are stimulated, possibly through the raised expression of the molecule in the synovial tissues. Preimmunization with mycobacterial hsp65 prior to arthritis induction has been found to induce resistance against disease in the same models. Altogether, we may conclude that preexposure to hsp65 is triggering protective mechanisms, which apparently are capable of controlling disease when the potentially arthritogenic stimulus is introduced. Very similar to the phenomena de-
380
W. van Eden
scribed for arthritis, in the case of diabetes in the NOD mouse, hsp65 immunization may lead to an early appearance of disease, while preimmunization with a selected hsp65 peptide was found to create resistance to destruction of pancreatic beta-cells and expression of diabetes (2). Moreover, in BB rats, the occurrence of spontaneous diabetes can be prevented by the early immunization to mycobacteria in the form of CFA (9). This seemingly prominent recognition of hsp65 in autoimmune disease states may be interpreted as being suggestive of a preoccupation of the immune system with hsps, both with respect to endogenously expressed hsp and with respect to the microbial counterparts. It is possible that the immune system sets a focus on the recognition of these proteins, which seems profitable both in terms of carefully controlling autoimmune reactivity and in terms of controlling the unattended in invasion by bacterial and other parasitic invaders (3). A well balanced and carefully controlled recognition of these proteins may then add to resistance to infection and at the same time to resistance to autoimmune disease. That the immune system is paying ample attention to hsps may be concluded from observations that hsps are dominant bacterial antigens. In the case of hsp65, it has been demonstrated that after mycobacterial priming in mice at the polyclonal level, up to 20% of the responding T cells may see this particular hsp (7). Thus, despite the extensive homology with self, hsps tend to be very immunogenic. Furthermore, responses directed at presumably risky epitopes, such as the epitope 180-188 of hsp65 which is recognized by the arthritogenic T cells in the Lewis rat, seem to be under rather strict regulatory control. This became especially evident in analyzing Fisher rats (6). Fisher rats are relatively resistant to AA induction, despite the fact that their MHC type, at least for class II molecules, is fully identical to that of the Lewis rat. Despite this MHC identity, Fisher rats are, in contrast to Lewis rats, non-responders for epitope 180-188. However, in rare cases of AA in Fisher rats, responses to 180-188 did occur. Thus, in Fisher rats non-responsiveness to 180-188 went together with arthritis resistance. It is possible that Fisher rats have a more efficient T cell network control of responding to this arthritis-associated epitope. This possibility seems to be supported by the observations made in germ-free Fisher rats. Fisher rats, bred germ-free from birth, are susceptible to AA and SCW arthritis as are Lewis rats. The resistance, however, develops as soon as colonization with bacterial flora or just E. coli has taken place, showing that exposure to environmental bacterial antigens in itself is providing the Fisher rats with the means of controlling responses to potentially dangerous epitopes. In line with this are findings in Lewis rats housed under conditions with a variable microbial load in the environment. It is well known that arthritis is more severe in animals kept in specific pathogen-free or otherwise relatively clean facilities, as compared to the severity of arthritis seen in animals exposed to an environment replete with potentially pathogenic microorganisms. Probably, the infectious pressure is keeping the immune system on the alert and provides the animal with immunological experience that helps to control autoimmunity. The same has been observed for spontaneous diabetes in the NOD mouse: while in the conventional setting, 50-80% of the (female) animals developed arthritis, the frequency reached 100% under germ-free or specific pathogen-free conditions (11). In line with this seem to be findings that resistance to the development of a relapse of AA in Lewis rats upon reimmunization with mycobacteria also depends on the environmental infectious pressure (I. R. Cohen, pers. comm.). Animals kept in a clean environment were found to be less resistant to such relapses. In these animals, evidence was obtained a less well developed network control of anti-hsp65 T cell (anti-MI)
Bacterial Heat-Shock Proteins in Autoimmune Diseases
381
immunity (3). That immunity to critical antigens is evolving with time in the conventional environment has been shown also in the pristane model of arthritis in CBAlIgb mice (10). In the naive normal mice, T-cell responses were seen to develop spontaneously early in life, being maximal at the age of 2-4 months. After that period, a significant drop was observed, with a very low level of responsiveness from the age of 8 months onwards. This paralleled pristane arthritis susceptibility in an inverse manner. At the age of 8 months, 40% of the animals appeared susceptible, while at the age of 3-4 months only 1-10% of the animals developed disease. Mice kept germ-free, however, were all susceptible at the early ages, while their anti-hsp65 T cell responses were found to be very low. All such observations seem to be in concert with the proposition that immunity to critical environmental antigens protects against the development of autoimmune diseases. This all may well open perspectives for the use of such antigens in artificially optimizing this particular state of immunization by means similar to vaccination as used to raise anti-infectious immunity. Much, however, will depend on our ability to avoid the boostering of undesired reactivity to risky epitopes. The experience obtained in the model system of arthritis has teached us so far that prevention of disease may be a relatively easy and realistic goal to reach. Reduction of the ongoing disease process, however, will be more difficult to achieve. For this, further understanding of the cellular networks or interactions that ensure the maintenance of self-tolerance seems to be necessary. Nevertheless, the molecular approach in bacterial organisms has provided us not only with antigens but also with antigenic fragments or epitopes thereof that playa role in the regulation of self-tolerance. Together with a further understanding of the cellular interactions within the immune system, it is possible that in the end we will have to conclude that absent or improper immunity to critical bacterial antigens such as heat-shock proteins, may also playa role in the "molecular pathogenesis of bacteria", in this case leading to autoimmune diseases.
References
1. Boog, C. J. P., E. R. de Graeff-Meeder, M. A. Lucassen, R. van der Zee, M. M. Voorhorst-Ogink, P. J. s. van Kooten, H. J. Geuze, and W. van Eden: Two monoclonal antibodies generated against human hsp60 show reactivity with synovial membranes of patients with juvenile chronic arthritis. J. Exp. Med., 175 (1992) 1805-1810 2. Cohen, I. R.: Autoimmunity to chaperonins in the pathogenesis of arthritis and diabetes. Ann. Rev. Immunol. 9 (1991) 567-589 3. Cohen, I. R. and D. B. Young: Autoimmunity, microbial immunity and the immunological homunculus. Immunol. Today 12 (1991) 105-110 4. De Graeff-Meeder, E. R., R. Van der Zee, G. T. Rijkers, H. J. Schuurman, W. Kuis, J. W,J. Bijlsma, B.J.M. Zegers, and W. Van Eden: Recognition of human 60kD heat shock protein by mononuclear cells from patients with juvenile chronic arthritis. Lancet 337 (1991) 1368-1372 5. Hermann, E., A. W. Lohse, R. Van der Zee, W. Van Eden, W. J. Mayet, P. Probst, T. Poral/a, K. H. Meyer zum Buschenfelde, and B. Fleischer: Synovial fluid derived T cells responding to human 65kD heat-shock protein and heat-stressed antigen presenting cells. Eur. J. Immunol. 21 (1991) 2139-2143 6. Hogervorst, E. J. M., C. J. P. Boog, J. P. A. Wagenaar, M. H. M. Wauben, R. van der Zee, and W. van Eden: T cell reactivity to an epitope of the mycobacterial 65kD heat shock protein (hsp65) corresponds with arthritis susceptibility in rats and is regulated by hsp65 specific cellular responses. Eur. J. Immunol. 21 (1991) 1289-1296
382
W. van Eden
7. Kaufmann, S. H. E.: Heat shock proteins and the immune response. Immunol. Today 11 (1990) 129-136 8. Res, P. C. M., C. G. Schaar, F. C. Breedveld, W. van Eden, J. D. A. van Embden, I. R. Cohen and R. R. P. de Vries: Synovial fluid T cell reactivity against the 65 kD heatshock protein of mycobacteria in early onset of chronic arthritis. Lancet ii (1988) 478-481 9. Sadelain, M. W. j., H. Y. Qin, W. Sumoski, N. Parfrey, B.Singh, and A. Rabinovitch: Prevention of diabetes in the BB rat by early immunotherapy using Freund's adjuvant. J. Autoimmunity 3 (1990) 671-680 " 10. Thompson, S. j., G. A. W. Rook, R. j. Brealey, R. van der Zee, and C. j. Elson: Autoimmune reactions to heat shock proteins in pristane induced arthritis. Eur. J. Immunol. 20 (1990) 2479-2483 11. Todd, j. A.: Genetic control of autoimmunity in type 1 diabetes. Immunol. Today 11 (1990) 122~ 129 12. Van Eden, W.: Heat shock proteins as immunogenic bacterial antigens with the potential of inducing and regulating autoimmune arthritis. Immunol. Rev. 121 (1991) 5-28 13. Van Eden, W., j. E. R. Thole, R. van der Zee, A. Noordzij, j. D. A. van Embden, E. J. Hensen, and I. R. Cohen: Cloning of the mycobacterial epitope recognized by T lymphocytes in adjuvant arthritis. Nature 331 (1988) 171-173
Dr. W. van Eden, Dept. of Infectious Diseases and Immunology, Faculty of Veterinary Medicine, University of Utrecht, Yalelaan 1, 3508 TD Utrecht, The Netherlands