Immunotogicat Reviews 1991, No. 121 Published by Munksgaard, Copenhagen, Denmark No part may be reproduced by any process without written permission from the author(s)

Heat-Shock Protein 60: Implications for Pathogenesis of and Protection against Bacterial Infections STEFAN H . E . KAUFMANN, BERND SCHOEL, J. D. A. v. EMBDEN''", TETSUJA KOGA, ANGELA WAND-WURTTENBERGER, MARTIN E. MUNK & ULRICH STEINHOFF

INTRODUCTION Heat-shock proteins (hsp or stress proteins) are produced by all cells in response to various physiological and nonphysiological stimuli (for review see Kaufmann 1990). hsp are ubiquitous and among the most abundant proteins ofthe biosphere. In increasing their intracellular hsp content, cells can protect themselves from otherwise lethal assaults. However, hsp are not only produced under stress situations; many of them are already constitutively present in the cell and carry out important house-keeping functions, hsp of 70 k and 60 k molecular weight are chaperones, e.g.. they participate in the folding, unfolding, translocation and assembly of other proteins (reviewed in Langer & Neupert 1991). The hsp 60 was first described by Hendrix in Escherichia coli in 1979 and has been termed GroEL. Together with GroES, it facilitates protein translocation across membrane barriers and possibly also secretion, hsp 60 has been found to be a common antigen of many bacterial pathogens including species of Borrelia, Legionella, Chlamydia, Coxiella, Salmonella, Treponema, Rickettsia. Mycobacterium, and Pseudomonas (for review, see Shinnick 1991). It is an abundant heat-inducible protein that makes up 1 to 2% of all proteins in bacteria grown under normal conditions and increases 4- to 5-fold under heat shock. Under certain stress conditions almost 10% of all proteins of M. bovis are hsp 60 (DeBruyn et al. 1987). Recent studies by Buchmeier & Heflron (1990) have shown that Salmonella Department of Immunology, University of Ulm, Albert-Einstein-Allee 11, D-7900 Ulm, * National Institute for Public Health and Environmental Hygiene, Bilthoven, The Netherlands. Correspondence: Prof. Dr. S. H. E. Kaufmann, Department of Immunology, University of Ulm, Albert-Einstein-Allee 11, D-7900 Ulm, FRG. Tel.: 0731/176-3387; Telefax: 073i/1763372.

68

KAUFMANN ET AL.

typhimurium organisms, which had been phagocytized by murine macrophages, markedly increase their hsp level, including that of hsp 60. This mere abundance of hsp may be one reason for their immunodominance. Bacterial hsp 60 cognates are highly conserved and hence the host may frequently come into contact with this antigen through infection with various microorganisms. Thus, constant boosting ofthe immune response to hsp could be another reason for immunodominance (Kaufmann et al. 1990). Hsp 60 is not restricted to the microbial world and a cognate is also found in mammals (Jindal et al. 1989). A major task of the mammalian hsp 60 is the protein folding and assembly in the mitochondrial matrix (for summary see Langer & Neupert 1991). Although it was originally thought that the localization of hsp 60 was restricted to mitochondria, recent findings indicate their presence in the cytoplasm (R. A. Young pers. communication). Hsp are highly conserved and, in the case of hsp 60, approximately 60% sequence homology between the mycobacterial and human cognates has been observed (Jindal et al. 1989). Hence, T cells and antibodies with specificity for conserved sequences are potentially autoreactive. The availability, in large amounts, of recombinant hsp 60 of M. tuberculosis, M. bovis and M. leprae has stimulated many scientists to study the antigenic properties of this well-defined protein in infectious and autoimmune diseases. M. tuberculosis and M. bovis, the etiologic agents of tuberculosis, and M. leprae, the causative organism of leprosy, are intracellular bacteria capable of persisting and replicating inside mononuclear phagocytes (for summary see Kaufmann 1986, 1987). Although macrophages represent the major habitat, other host cells may be abused as well. A notable example are Schwann cells, which represent a preferred habitat for M. leprae. T lymphocytes are the major mediators of acquired resistance and, to a large extent, pathology seems to be caused by cellular immune mechanisms as well. Recent studies have suggested the involvement of alpha/beta and gamma/delta T cells in immunity to mycobacteria (Kaufmann 1988, Kaufmann & Kabelitz 1991). The activation of strong antimicrobial effector functions in macrophages through interleukins produced by T cells seems to represent the crucial step in acquired resistance. In addition, cytolytic T cells (CTL) may contribute to protection by releasing bacteria from intracellular niches (Kaufmann 1988). This mechanism, however, may also have detrimental consequences for the host. It has been known for a long time that autoreactive T cells and antibodies often arise during mycobacterial infections and that they may contribute to pathogenesis (Asherson 1968, Shoenfeld & Isenberg 1988). Increased autoantibody levels have been found in severe forms of mycobacterial diseases, in particular in patients suffering from lepromatous leprosy. Also the generation of autoreactive T lymphocytes following mycobacterial infection has been documented. Our own group found high numbers of autoreactive T cells in mice infected with

hsp 60 IN BACTERIAL INFECTIONS

69

viable M. bovis which were active both in vitro and in vivo (Muller et al. 1986a, b). For example, among a battery of T-cell hybridomas derived from M. bovisinfected mice, approximately V-i showed auto reactivity (Muller & Kaufmann 1985). Many hybridomas expressed dual reactivity, i.e., they responded to autologous cells alone and to those pulsed with mycobacterial antigens but to a different extent. This observation would be consistent with a conserved peptide serving as T-cell epitope. Our laboratory originally became engaged in hsp 60 while searching for vaccine candidates against tuberculosis and leprosy. During our studies, it soon became dear that hsp 60 represents a fascinating antigen which not only allows new insights into the mechanisms underlying acquired resistance to and immunopathology of bacterial infections but may also be pertinent to our understanding of surveillance and autoimmunity. hsp 60: A PROTECTIVE ANTIGEN? The strong antigenicity for T cells ofa mycobacterial 64 kD protein was already recognized before it was known to belong to the hsp family. When this protein had been cloned in E. eoli (Thole et al. 1987) it became possible to test its antigenicity on human CD4 T cells from a tuberculoid leprosy patient (Emmrich et al. 1986). It was found to stimulate two ofthe four CD4 T-cell clones analyzed. It was also observed that many T-cell clones from healthy PPD"^ donors recognized this antigen, an observation which was later validated by showing that T cells from nonnal healthy individuals respond to hsp 60 in primary in vitro proliferation assays (Munk et al. 1988). Subsequently, limiting dilution analyses performed with purified T cells revealed that, in certain individuals, up to 50% of the M. tubereulosis-reacti\e T cells are hsp 60-speciric (Kabelitz et al. 1990). Both alpha/beta and gamma/delta T cells respond to hsp 60 (Emmrich et al. 1986, Munk et al. 1988, Holoshitz et al. 1989, Haregewoin et al. 1989, KabeUtz et al. 1990). Thus, in patients as well as in healthy individuals hsp 60-specific T cells exist and seem to be realtively frequent, indicating that it is a dominant antigen of M. tuberculosis. Limiting dilution analyses performed in the murine system revealed that 10-20% of all T cells from M. tuberculosis-imm^xmztd mice whieh react to whole M. tuberculosis organisms are specific for hsp 60 (Kaufmann et al. 1987). We wondered, therefore, whether immunization with hsp 60 induces a strong cellular immune response against M. tuberculosis provided it is delivered in a potent carrier system. Mice were immunized with hsp 60 in the Ribi adjuvant consisting of trehalose dimycolate, monophosphoryl lipid A, and mycobacterial cell wall skeleton. Eight to 10 days after subcutaneous injection, draining lymph nodes were collected and T cells isolated. Afterwards, limiting dilution analyses were performed using hsp 60 or killed M. tuberculosis as antigens (Fig. 1). About 1/

70

KAUFMANN ET AL. Number of responder cetts 300 500 1200

2500

5000

Figure I. Frequency analysis of T cells from mice immunized with hsp 60 in a strong adjuvant. Mice were immunized with hsp 60 in Ribi adjuvant s.c. and 8 days later lymph node cells were collected. Purified T ceils were cultured in the presence or absence of the hsp 60 (2.5 //g per well) or killed M. tuberclosis (5 /ig per well). (D) hsp 60, frequency 1/ 2990; (o) M. tuberculosis, frequency 1/1820. Fractions xO.l. Data from Kaufmann et al (1987) with permission.

2000 T cells responded to M. tuberculosis, a frequency comparable with that seen after immunization with whole M. tuberculosis organisms in Freund's adjuvant (Kaufmann et al. 1987). These fmdings, therefore, indicate that immunization with hsp 60 in an appropriate adjuvant system activates high numbers of M. tuberculosis-responsWe T cells. To further improve its immunogenicity, hsp 60 was cloned into a viable carrier system, the Aro" Salmonella typhimurium. Aro" S. typhimurium strains are mutants which lack a gene essential for the biosynthesis of aromatic compounds and hence are unable to survive in the mammalian host. They are avirulent and have been used successfully for vaccination against mouse typhoid (Hoiseth & Stocker 1981). A plasmid carrying the left promotor Pj, of lambda upstream of the mycobacterial ribosomal binding site and the start codon of hsp 60 was cloned into Aro" S. typhimurium by J. Thole and J. v. Embden. This plasmid allows thenno-inducible transcription of the PL promotor; at 28°C, synthesis of hsp 60

hsp 60 IN BACTERIAL INFECTIONS

71

is negligible whereas, at 42°C, hsp 60 is produced as the major protein of S. typhimurium; at 37°C, an intermediate level of hsp 60 is produced. Mice were immunized with r-Aro" S. typhimurium expressing hsp 60 (Ml 175) or with r-Aro S. typhimurium lacking the hsp 60 gene (Ml 167). Several weeks after vaccination, the mice were challenged with PPD or hsp 60 and delayed-type hypersensitivity (DTH) reactions were measured 24 h later. Although salmonellae expressed hsp 60 in mice for a limited time period only, they induced significant DTH responses (Table I). Interestingly, this vaccination schedule did not induce DTH against PPD, probably because the preparation of PPD used did not contain sufficient amounts of hsp 60. Active vaccination against tuberculosis was also attempted with these constructs. Mice were immunized with the r-Aro"' S. typhimurium and afterwards infected with viable M. bovis; these studies, however, gave inconsistent results. The question as to whether hsp 60-specific T cells are protective, therefore, cannot be answered definitively as yet. On the one hand, the failure to demonstrate reproducible protection by vaccination with hsp 60 in a strong carrier argues against a role in acquired resistance. On the other hand, a baseline immune response to hsp 60 could have already been stimulated naturally by subchnical infections with various bacteria (although we used SPF mice) which could not be further increased by vaccination. One approach to decide between these two alternatives would be to perform vaccination studies with salmonella strains which stably express hsp 60 in vivo, using germ-free mice. The demonstration of primary T-cell responses to hsp 60 in healthy individuals indicates that preimmunizations indeed occur in humans without being recognized. Why do so many adults suffer from infectious diseases such as tuberculosis

TABLE I Delayed'type hypersensitivity induced by hsp 60 in an avirutent replicating vector Immunization

Challenge

MII75 Ml 167 Nil

hsp 60 (10 /ig) hsp 60 (10/ig) hsp 60 (10 fig)

M1175 M1167 Nil

hsp 60 (3 /ig) hsp 60 (3 fig) hsp 60 (3 fig)

M1175 Ml 167 Nil

PPD (5 /ig) PPD (5 /ig) PPD(5/iB)

DTH Response 12 4 3 7 •1:.. •

2

4 3 2

Mice (C57B1/6) were vaccinated with 5 x 10* viable organisms of M n 7 5 (r-Aro 5. typhimurium expressing hsp 60) Ml 167 (r-Aro" S. typhimurium lacking the hsp 60 gene) or were left untreated. After 5 weeks, animals were challenged with hsp 60 or PPD and DTH responses measured 24 hours later as described.

72

KAUFMANN ET AL.

ifthey already possess protective T cells prior to first encounter with the responsible agent? As many as 50-60 million people are afflicted by tuberculosis worldwide and about 5% of all mortalities are due to this disease. On the other hand, one may note that about 1.7 billion people, or I4 ofthe world's population, are infected with M. tuberculosis. Thus, in the case of tuberculosis the percentage of people infected but not diseased far exceeds the number of those infected and diseased. The pre-existence of crossreactive T lymphocytes, at least in part, may contribute to this difference. DO hsp CONTRIBUTE TO AUTOPROTECTION OF HOST CELLS? The last section discussed evidence that hsp 60 is a dominant T-cel! antigen of mycobacterial infections, hsp are not only produced by the predator but also by its prey becatise infected cells have to face insults from at least two sides. - First, intracellular parasitism markedly influences the affected cell even if the parasite is of low toxicity itself Mycobacteria, in particular M. leprae, are of low toxicity and this may be one reason for their longevity inside professional and nonprofessional phagocytes. Still, mycobacteria induce striking alterations in macrophages, such as the formation of multinucleated giant cells or foam cells. Furthermore, mycobacteria, despite their slow duplication rate, require certain nutrients for their survival which they have to obtain from their prey. - Second, sooner or later the immune system will recogtiize parasitized host cells and set into motion effector mechanisms aimed at the eradication of the foreign invader. This not only includes the activation of professional phagocytes, but also the destruction of host cells of low antimicrobial activity (for review, see Kaufmann 1988). It can therefore be envisaged that host cells parasitized by intracellular bacteria increase their own hsp synthesis as an autoprotective mechanism. Polla and coworkers have claimed preferential induction of hsp 60 in macrophages which have phagocytized M. bovis (for review, see Kantengwa et al. 1991). Besides macrophages, Schwann cells provide a major habitat for M. leprae (reviewed in Kaufmann 1986). Schwann cells are essential for the structural and functional integrity of the peripheral nervous system. Nerve damage as a direct consequence of Sehwann cell destruction occurs in all forms of leprosy and demyelinization and complete Schwann cell destruction is characteristic for the onset of lepromatous neuropathology (Antia 1982). Although the reasons for the preferential affinity of M. leprae to Schwann cells remain unclear, it has been claimed that this cell provides a niche that is protected from attack by the immune system. More recently, however, it has been shown that Schwann cells can present mycobacterial antigens in the context of MHC molecules and thus become recognizable by T lymphocytes (Steinhoff & Kaufmann 1988). Because they are hardly - if at all - reconstituted, damage to Schwann cells must be particularly deleterious

hsp 60 IN BACTERIAL INFECTIONS

73

for the host and extreme precautions may be required in order to protect these cells from harmful assaults. One way would be to increase endogenous hsp synthesis in response to sublethal stimuli such as those caused by low-level infection with M. leprae. We approached this question using the monoclonal antibody (mAb) IVDl kindly provided by Dr. J. DeBruyn which crossreacts with a shared epitope of mycobacterial and human hsp 60 (within the sequence AA 1-62 and AA 1-82 of M. tuberculosis and human hsp 60, respectively). Human Schwann cells were either infected with M. leprae or remained uninfected. Lysates were separated on SDS-PAGE under nonreducing conditions, immunoblotted on nitrocellulose and developed with mAb IVDl. Only a faint band was identified with uninfected Schwann cells, whereas a marked signal was observed after M. leprae infection (Fig. 2). Appropriate controls excluded this band from being mycobacterial hsp 60. In a complementary experiment, Schwann cells were cultured with "Smethionin following IFN-y stimulation, heat shock, or infection with M. leprae. Afterwards, cell lysates were immunoprecipitated with another crossreactive mAb. ML30, kindly provided by Dr. J. Ivanyi (Ivanyi et al. 1983). This mAb recognizes AA 275-297 and AA 301-322 of mycobacterial and human hsp 60, respectively (17 of 22 AA are identical in this stretch). Only very little 60 kD material was precipitated from control Schwann cells, whereas a significant quantity of a 60 kD molecule was demonstrable after IFN-y stimulation, heat shock, or M. leprae infection. These fmdjngs suggest that Schwann cells parasitized with M. leprae increase their intracellular levels of hsp 60. Although we assume that destruction of certain host cells with low antibacterial potential contributes to protective immunity by facilitating transition of bacteria into more potent effector ceils, generally host cell destruction remains a doubleedged sword which - almost by definition - also causes pathogenic alterations (Kaufmann 1988). Because of their extraordinary value for the host, the latter fate may be particularly critical for Schwann cells. For such a valuable cell, the bearing of an intracellular parasite may be preferable to attack by host cells. Evidence has been presented recently that natural killer (NK) cells are activated during leprosy and a role in protection has been claimed (Kaplan et al. 1989, Ab et al. 1990). NK cells are highly aggressive cells which have been implicated in defence against viral infections and tumor surveillance and scattered reports have appeared in the literature indicating that they may also be involved in defence against bacterial and parasitic infections (Blanehard et al. 1987, Carl & Dasch 1986, Klimpel et al. 1986). We analyzed the influence of M. leprae on Schwann cell destruction by such killer cells (Steinhoff et al. 1991). Peripheral blood leukocytes from normal healthy individuals were activated with dead mycobacteria and subsequently restimulated with IL-2. After some weeks, these cells expressed strong killer activity against the NK-susceptible target cells, K562, and against mononuclear phagocytes which had been pulsed with dead M. leprae.

74

KAUFMANN ET AL.

kDa

A

B

200 —

94 —

60— 46 — 30 — 21,5 —

Figure 2. Western blot analysis of hsp 60 in uninfected or M. teprae-infected human Schwann cells. Lysates were separated by SDS polyacrylamide gel electrophoresis, blotted onto nitrocellulose and stained with the anti-hsp 60 mAb, 1A4. A, M. bovis: B, uninfected Schwann cells; C, Schwann cells infected with viable M. leprae (100 organisms per cell) for 24 h.

Schwann cells were also killed effectively, whereas unpulsed cells remained virtually unaffected or were killed to a lesser degree (Fig. 3). Interestingly, Schwann cells which had been infected with viable M. leprae organisms were not attacked. To assess whether this lack of killing was due to a failure to recognize infected target cells or due to increased resistance, Schwann cells were infected with viable organisms and then pulsed with dead leprosy bacilli. Infection significantly reduced Schwann cell lysis (Fig. 4). Although, in this particular experiment, significant killing was already found with unpulsed Schwann cells, a further

75

hsp 60 IN BACTERIAL INFECTIONS

40-

40-

30-

R an.

20-

20

10-

tt-

0-"

030:1

E:T ratio

6:1

30:1

E:T ratio

5:1

Figure 3. EfTect of dead and viable M. leprae on lysis of Schwann cells by NK cells, (a) Schwann cells were pulsed with 3 /ig (A), 10 /jg (D), or 30 fig (O) dead M. leprae or were left untreated (o). (b) Schwann cells were infected with viable M. leprae at ratios of 30 (•), 10 ( • ) or 5 (A) hacteria per cell or were left untreated (o). Data reproduced from SteinhofT et al. (1991) with permission.

increase was caused by dead M. leprae which was reduced by about 50% when the cells had been infected with viable bacilli. Resistance to killing was also achieved by IFN-y stimulation and by heat shock. These findings are in agreement with data obtained by others who found increased resistance to killing by NK cells, TNF or specific CTL after treatment with heat, IFN-y, or chemicals (Sugawara et al. 1990, Renkonen et al. 1988, Gromkowski et al. 1989). Our data provide evidence that infection with viable leprosy bacilli causes

80 70 60 .2 50 -" 40 302010 0 60:1

E: T ratio

20:1

Figure 4. Reduced killing ot M. leprae-pulscd Schwann cells after infection with viable M. leprae. Schwann cells were pulsed with 3 /ig (A) or 1 fig (D) killed M. leprae or left untreated (o). Alternatively, Schwann cells were infecled with viable M. leprae at 5 bacteria per target cell alone (•) or 24 h before pulsing with 3 fig (A) or 1 ftg ( • ) dead M. leprae. Data reproduced from SteinhofT et al. (1991) with permission.

76

KAUFMANN ET AL.

increased resistance to killer cells. It therefore appears that the process of infection, but not the mere encounter of microbial components, provides a signal for resistance. Viable M. leprae are translocated from the endosomal into the cytoplasmic compartment (Mor 1983) and we assume that the existence of M. leprae in the cytoplasm was a necessary condition for resistance. Consistent with this notion, the presence of foreign proteins in the cytoplasm has been shown to cause hsp induction {Ananthan et al. 1986). Although our data do not provide a causal link between hsp synthesis and resistance, they provide evidence that not only heat shock, but also naturally occurring infection causes stress situations that can eventually protect the cell from another more dramatic assault. CROSSREACTIVITY OF ANTIBODIES AND T CELLS WITH SPECIFICITY FOR MYCOBACTERIAL hsp 60 hsp 60 cognates are ubiquitous and highly conserved. Hence, they harbor in their structure the potential to cause autoimmune responses. In mammals, hsp 60 is found in the mitochondrial matrix where it facilitates the translocation, folding and assembly of mitochondrial proteins (Langer & Neupert 1991). Class Irestricted T lymphocytes primarily recognize peptides derived from proteins that are present in the cytoplasmic compartment (Townsend & Bodmer 1989). Thus, presentation of newly synthesized hsp 60 to class I-restricted T cells should be possible, whereas processing of autologous hsp 60 through the class II pathway is considered less likely. In contrast, bacterial hsp 60 is primarily presented in the context of MHC class II gene products.

CROSSREACTIVE ANTIBODIES The intraceliular location of hsp 60 should prevent recognition ofthe host's own cognates by the humoral immune response, although antibodies against conserved epitopes have been generated by immunization with mycobacterial hsp 60 cognates (see above and Engers et al. 1986). Strikingly, we obtained evidence that certain host cells are recognized by cross-reactive antibodies. Macrophages were generated by in vitro cultivation of bone marrow cells as described (Flesch & Kaufmann 1987). After 9-11 d of culture, cells were stained with various mAb, including the mAb ML30, and with a rabbit antiserum raised against mycobacterial r-hsp 60 (Wand-Wiirttenberger et al. 1991). Cytofluorimetric analysis suggests that ML30 and the anti-hsp 60 rabbit antiserum defined a distinct cell population, whereas matched control antibodies and antisera gave only a small shift in peak intensity. Similar effects were seen with murine Schwann cells. On the other hand, we failed to detect a specific signal on peripheral blood leukocytes from normal healthy donors. Thus, our findings suggest that antibodies with specificity for mycobacterial hsp 60 react with surface molecules of certain,

,

hsp 60 IN BACTERIAL INFECTIONS

fl

but not all, host cells. Consistent with this, Pierce and coworkers have provided evidence for surface expression of hsp 70 by antigen-presenting B cells and macrophages (reviewed in Pierce et al. 1991), as have Fisch et al. (1990) for hsp 60 expression by Daudi cells. Jajour et al. (1990) found that a rabbit antiserum against human hsp 60 detects a 77 kD surface molecule on certain gamma/delta T cells. Finally, surface expression of hsp 100/hsp 90 cognates by chemically induced mouse sarcomas and of hsp 70 cognates by transformed rat fibroblasts has been described (for review see Srivastava & Maki 1991). Some degree of sequence homology exists between mycobacterial hsp 60 and Other mammalian proteins including mouse T-complex proteins known to be expressed on the cell surface (Gupta 1990). Thus, we do not want to exclude crossreactivity wilh an unrelated and hitherto unknown mammalian protein. Note that the IVDl mAb, which identifies a 60 kD band in lysates of human and murine cells including BMM, failed to surface-label BMM. Furthermore, in more recent studies employing supernatants, instead of ascites, of the ML30 hybridoma cell line, we failed to detect surface staining of BMM at antibody concentrations which stained a 60 kD band in cell lysates (our own observation and Kissling et al., this volume). Hence further studies will be required before definitely deciding whether host hsp 60 itself is expressed on the cell surface or not. A correlation between mycobacterial infection and autoimmune disease has been claimed (Shoenfeld & Isenberg 1988, Asherson 1968) and increased levels of serum antibodies against mycobacterial hsp 60 have been described in autoimmune diseases (Tsoulfa et al. 1989, Kaufmann 1990). The fine specificity of these antibodies, however, is still unknown, and we do not know whether they are directed against conserved or nonconserved regions. WHAT ABOUT T CELLS? The high conservation of hsp 60 cognates prompted us to ask two questions: (1) Do T cells with specificity for epitopes shared by mycobacterial and human hsp 60 cognates exist? (2) Do T cells with specificity for mycobacterial hsp 60 recognize stressed host cells? The first question was addressed in the human system using synthetic peptides corresponding to four highly conserved regions of hsp 60 (Table II). These peptides were chosen because they were at least 10 AA long and showed full or almost complete sequence homology in the mycobacteriat and human cognates. Note that these peptides do not represent preferred T-cell epitopes according to known algorithmic predictions. However, using overlapping peptides of mycobacterial hsp 60 we previously found that peripheral blood leukocytes of healthy individuals are stimulated by a large number of peptides independent of known predictions (Munk et al. 1990).

78

KAUFMANN ET AL.

TABLE II Alignment ofthe amino acid residues ofthe self-epitopes shared by human and mycobacterial hsp 60 hsp Human M. tuberculosis

Residue No.

Amino Add Sequenee*

109-120 84—95

AGDGTTTATVLA ***********

Human

269-280

KPLVI I AEDVDGE ALS TLV LN

M. tuberculosis

243-263

# » * L * * * * * * E * * * * * * * * V *

Human M. tuberculosis

298-307 272-286

VAV KAP G F G D **********

Human M. tuberculosis

430-442 403-414

A AV E E GI V L G GG ********y^*#*

i i

' The * indicates identicaJ amino acid; I indicates trypsin cleavage site within self epitopes.

Peripheral blood leukocytes from normal healthy individuals were stimulated in vitro with killed M. tuberculosis and afterwards their cytolytic activity was assessed on autologous adherent cells which had been pulsed with intact mycobacterial hsp 60 antigen or with a tryptic digest thereof (Munk et al. 1989). Targets pulsed with killed M. tuberculosis or with tryptic hsp 60 peptides were efficiently lysed whereas those pulsed with intact mycobacterial hsp 60 remained virtually unaffected. When peptides were added to target cells, responses were observed in 8/9 individuals against at least one of these peptides. Representative responses are shown in Table III. Only one ofthe four peptides contains a trypsin cleavage site and hence the corresponding region in hsp 60 must have been destroyed by tryptic digestion whereas the other three epitopes are contained in tryptic peptides. Applying L-cells transfected with human HLA genes, kindly provided by R. W. Karr, we found that the peptides were presented by histocompatible HLA-DR molecules. Recently, we found that a tryptic digest of the human hsp 60 is also recognized by M. tuberculosis-activated T cells from healthy individuals. Taken together, these fmdings demonstrate the existence of T cells with specificity to highly conserved regions of hsp 60 in the peripheral blood of healthy individuals. Preliminary experiments indicate that M. tuberculosis-acti\ate6 T lymphocytes also recognize less conserved regions of the human hsp 60 such as AA sequence 180-196 (8/17 identical AA) which seems to be a major epitope for gamma/delta T cells from the thymus of newborn mice (Bom et al. 1990). At the same time, our data are not consistent with the generation of shared epitopes through the class II processing pathway, because pulsing with intact mycobacterial hsp 60 antigen did not result in killing. Nevertheless, hsp 60 is able to stimulate T-cell proliferation from many healthy donors (Munk et al. 1988). Although T cells recognizing shared epitopes of hsp 60 in the context of MHC

"ll-

hsp 60 IN BACTERIAL INFECTIONS

79

TABLE III Specificity ofM. luhercnlosis-activaled CTL against shared self epitopes of hsp 60 Percent Specific Lysis at E:T Ratio Donor

Target Priming

1

M. tuberculosis hsp 60 (intact) Peptide 109-120 Peptide 298-307 Peptide 430-^1 None

2

3

4

M. tuberculosis hsp 60 (intact) Peptide 109-120 Peptide 269-289 Peptide 298-307 None M. tuberculosis hsp 60 (tryptic) Peptide 269-289 Peptide 298-307 Peptide 430-441 None M. tuberculosis hsp (intact) hsp 60 (tryptic) Peptide 109-120 Peptide 269-289 Peptide 298-307 Peptide 430-441 None

5:1 0

0 0 0 0 0 9 0 0 0

15;]

6 0 4 ]2

0 0 15 0

I

30:1 19 0 50

30 0 0 33 0

0

0 10

0 4 0 4 26 30 26 12

5

22

28

0

0 14 0

3 35 0 0 4

0 % 7 6 0

4 0 0 0

9

e2

3 0 2

19 17

0 5 0 0 6 2

0 0

4 6

90:1 37 0 42 36

0 0 38 .0 2 15 0 5 67 70 58 33 51 5 48 0

0 4 0 0 3 6

" Peripheral blood leukocytes from healthy individuals were activated with killed M. tuberculosis for 7 days and afterwards CTL activities were determined on autologous target celts primed with 5 /ig/wel! killed M. tuberculosis, intact 60-kDa hsp, or trypsinized 60-kDa hsp. Alternatively, targets were labelled with 5 /ig/well synthetic peptides. Data from Munk et al. (1989) with permission.

class II molecules exist in normal healthy individuals, it appears that they are not harmful, probably because these shared epitopes are not generated in sufficient amounts to activate an immune response. On the other hand, our preparation of killed M. tuberculosis should contain sufiicient quantities of shared epitopes since crossreactive T cells were activated. During certain stages of infection, large amounts of bacterial detritus accumulate locally and are exposed to proteases of host or microbial origin. Enzymatic protein degradation could give rise to crossreactive peptides which could then directly bind to MHC molecules, resulting in the activation of crossreactive T lymphocytes (Kaufmann et al. 1990). Such T

80

KAUFMANN ET AL.

cells are likely to be harmless as long as autologous self peptides are not presented. The folloviing set of experiments, which were performed in the murine system, suggests that this may be the case under certain conditions. CTL of CD8 phenotype were activated in vitro by high concentrations of a tryptic digest of mycobacterial hsp 60 (Koga et al. 1989, Steinhoff et al. 1990). In confinnation of studies hy Bevan and coworkers (Carbone et al. 1988), in this in vilro culture system CTL capable of recognizing targets pulsed with the homologous peptide preparation were activated. In the case of tryptic ovalbumin, such peptide killers fail to recognize ovalbumin peptides generated by endogenous class I processing. In contrast, when intact hsp 60 antigen was introduced into target cells hy osmotic shock, CTL raised against tryptic peptides recognized the epitope generated by endogenous processing (Fig. 5). These findings could he explained by two alternatives: first, in the case of hsp 60, the proteases involved in endogenous processing and trypsin could produce similar peptides; second, in vitro stimulation with tryptic hsp 60 peptides actually reflected secondary stimulation because mice might have been primed to hsp 60 through previous in vivo contact with environmental microbes (see above). The question then was whether autologous hsp 60 also undergoes class I processing and whether shared epitopes are presented. For these experiments, the two cell types which serve as major habitats for mycobacteria were used as targets: macrophages, which constitutively express

50

40

30

£ 2010

60:1

20;1

60:1

20:1

Effector: Target Ratio

Figure 5. CTL against tryptic hsp 60 peptides but not against tryptic ovalbumin peptides recognize the homologous protein after processing through the cytoplasmic pathway. CTL were activated with ovalbumin peptides (a) or hsp 60 peptides (b). hsp 60 or ovalbumin were introduced into EL4 cells by osmotic shock. Alternatively, EL4 cells were pulsed with hsp 60 peptides or with ovaibumin pepUdes. (A) hsp 60 peptides; ( • ) ovalbumin peptides; (A) hsp 60 introduced by osmotic shock; ( • ) ovalbumin introduced by osmotic shock; (o) osmotic shock control. Data from Steinhoffet al. (1990) with permission.

hsp 60 IN BACTERIAL INFECTIONS

81

class 1 MHC molecules, and Schwann cells, which are constitutively MHC class r but become class I^ after IFN-y stimulation (Steinhoff & Kaufmann 1988). CTL raised against tryptic peptides of mycobacterial hsp 60 not only recognized BMM pulsed with the homologous peptide digest, but also BMM that had been stimulated with IFN-y or infected with cytomegalovirus but had never contacted peptides of mycobacterial origin (Fig. 6). In contrast, unstressed BMM remained virtually unaffected. In the case of Schwann cells, the peptide pulse alone was not sufficient, probably because the required restriction element was lacking. However, IFN-y stimulation alone allowed for lysis which was only marginally increased by addition of tryptic hsp 60 peptides. Further experiments demonstrated that stressed host cells were not killed by T cells raised against tryptic

60 1 20:1

'

'•

71

60;1 20:1

7:1

60:1 20 1

7:1

Effector : Target Ratio

Figure 6. Lysis of peptide pulsed BMM or stressed BMM by CTL raised against mycobacterial hsp 60. CTL were activated against tryptic digest of mycohacterial hsp 60 (a, b, c) or ovalbumin (d) and then tested on unstimuiated BMM (•); BMM pulsed with hsp 60 (o); BMM activated with IF'N-y containing T cell factors (V); r-IFN-j'-activated BMM (A); cytomegalovirus-infected BMM (D); and BMM pulsed with ovalbumin peptides (O). Data reproduced from Koga et al. (1989) with pennission).

82

KAUFMANN ET AL.

ovalbumin peptides and that T cells raised against hsp 60 peptides failed to recognize targets pulsed with ovalbumin peptides. Furthermore, it was found that the CTL were class I-restricted, CD8-^ (Fig. 7) and expressed the alpha/beta TCR. Currently, we do not know which endogenous protein is the source of the crossreactive T-cell epitope(s). However, it is tempting to speculate that hsp 60 itself fulfills this task. Consistent with this assumption we find that heat shock of BMM also yields T-cell recognition. However, knowing that mycobacterial hsp 60 shows homology with other proteins, this question stili needs a definite answer (Kaufmann 1990).

© 30-

10-

tn 060:1 20:1

7:1

60:1 20:1

7:1

60:1 20:1

7:1

u

a.

25-, 201510-

5060:1

60:1

20:1

60:1

Effector : Target Ratio Figure 7. Phenotype and genetic restriction of hsp 60-activated CTL. CTL were activated in vitro with tryptic digest of hsp 60 and then tested on unstimulated BMM (a), BMM pulsed with hsp 60 peptides (b), or on IFN-y-stimulated BMM (c) in the presence of antibodies to CD4 or CD8. (O) no antibodies; (A) anti-CD4 antibodies; (D) anti-CD8 antibodies. T cells from C57BI/6 mice (d), BIO.MBR (e) or B10.A(2R) (0 miee were activated with tryptic hsp 60 and then tested on BMM from C57BI/6 mice, (o) unstimulated BMM; (G) BMM pulsed with hsp 60 pepUdes; (A) IFN-y-stimulated BMM. Data from Koga et al. (1989) with permission.

I

hsp 60 IN BACTERIAL INFECTIONS

83

The problem was further approached by employing ati appropriate anti-sense probe to specifically interfere with neosynthesis of hsp 60 in stressed host cells. A DNA sequence was selected which showed no homology with other known sequences according to the Genbank release 65 and EMBL release 24 within the UW GCG package. A thiophosphate-modified sense oligonucleotide probe corresponding to this sequence and a corresponding anti-sense probe were used. Incubation of acridin-labelled probes demonstrated marked uptake by BMM and Schwann cells during a 20-h incubation. Immunoprecipitation studies of "Smethionin-labelled cells indicated that the anti-sense probe indeed inhibited neosynthesis of endogenous hsp 60 whereas the sense probe was ineffective. We therefore assessed whether the probe specifically blocked recognition of stressed cells by hsp 60-reactive CTL. BMM were incubated with the anti-sense probe and then stimulated with IFN-y. Subsequently, lysis of BMM by CTL raised against a tryptic digest of mycobacterial hsp 60 was assessed. The anti-sense probe inhibited lysis of IFN-y-stimuIated BMM, but not that of peptide pulsed BMM. Additional experiments showed that the anti-sense oligonucleotides did not affect specific lysis in the tryptic ovalbumin system or in an allogeneic situation. Thus, our data favour autologous hsp 60 as the source ofthe crossreactive peptide presented by BMM in the context of MHC class I gene products. Yet we cannot exclude that hsp 60 is somehow involved in MHC class I assembly/ expression and hence that the anti-sense probe interfered with this more general step. k

SUMMARY AND CONCLUSIONS In this review we have focused on antigenic features of hsp 60 related to: its ubiquitous distribution in the biosphere; its extraordinary homology among various bacteria; its high conservation from prokaryotic to eukaryotie cells; and its abundant expression under stress situations occurring during infection. These unique features make hsp 60 an excellent candidate antigen relevant to protection and pathogenesis of bacterial infections and, perhaps in a broader sense, to surveillance and autoimmunity. We will briefly discuss these possibilities in the following. Acquired resistance If we assume that bacterial organisms contain some thousand different proteins which all represent potential antigens, the frequency of T cells with specificity for mycobacterial hsp 60 appears surprisingly high. Although, during the course of infection, high levels of hsp may be induced in bacteria, mere abundance appears to be an important though insufficient explanation. In addition, constant boosting by similar hsp 60 cognates from various microbes with which humans come into

84

KAUFMANN ET AL.

contact may contribute to dominance. This could easily explain the occurrence of hsp 60-specific T cells in healthy individuals with no clinical history of mycobacterial infections. Involvement of more sophisticated mechanisms, such as the affinity of hsp to other proteins, cannot be excluded (Flynn et al. 1989). Yet dominance does not necessarily mean protection and definite proof that hsp are protective antigens is lacking. Perhaps the immune response against epitopes shared by various mycobacterial pathogens represents a first line of defence preceding a more specific immune response. Such broadly reactive antigens would not qualify as prime candidates for vaccine design. Imtnunesurveillanee T cells with specificity for epitopes shared by bacterial and human hsp 60 are readily demonstrable and stressed host cells are recognized by hsp 60-specinc T cells. Such T lymphocytes are endowed with the capacity to identify host cells stressed by a variety of assaults such as inflammation, infection, trauma, or transformation. Although it has been claimed that hsp-reactive gamma/delta T cells are particularly destined for such surveillance functions (Born et al. 1990, Asamow et al. 1988), alpha/beta T cells could also participate. Pathogenesis The mechanisms causing pathogenesis should be similar to those underlying protection and surveillance. In the former case bacterial hsp would be responsible for both induction of immunity and expression of pathogenic reactions; in the latter case an immune response stimulated by conserved regions of bacterial hsp 60 would be converted against a host-derived cognate. In both cases, attraction of hsp 60-specinc T cells to sites suffering from a bacterial burden could cause pathogenic, probably DTH-like reactions. Direct evidence for such a situation comes from experimental infections of mice with Chlamydia sp. where an hsp 60 cognate has been identified as a stimulator of immunopathogenic reactions (Morrison et al. 1989). it remains to be established whether such a response contributes to the chronic ocular and genital tract infiammations caused by chlamydiae. It is also unclear whether bacteria-specific or conserved regions are the target of the pathogenic response. Another example which has been discussed here may be the destruction of stressed Schwann cells by autoreactive T cells after activation by hsp 60, which could contribute to nerve damage in leprosy (Steinhofret al. 1990). ; Autoimmunity

)

T cells with reactivity to self epitopes of hsp 60 have even been demonstrated in healthy individuals and antibodies to conserved epitopes of hsp 60 have been

hsp 60 IN BACTERIAL INFECTIONS

85

generated in mice by immunization with bacteria (Ivanyi et al. 1983, Engers et al. 1986, Munk et al. 1989, Lamb et al. 1989). Hence, such B cells and T cells exist and the questions arise as to how these autoreactive T cells evade thymic deletion and whether they cause pathologic damage. With regard to the first point, we need to distinguish between CD4 T cells studied in the human system and CD8 T cells analyzed in the murine system. Autoreactive CD4 T cells were only demonstrable by using synthetic peptides and not when intact hsp 60 was added to the cultures. Furthermore, we do not have evidence for class II processing/presentation of endogenous hsp 60. Therefore, we propose that endogenous class II processing fails to generate sufficient quantities of self epitopes required for CD4 T-cell deletion in the thymus and. later, for their activation in the periphery. In contrast, class I presentation of endogenously derived epitope(s) shared with mycobacterial hsp 60 was demonstrable using CD8 T lymphocytes after in vitro activation with peptides. In agreement with Bevan and coworkers (Carbone et al. 1988) CD8 T cells with low specificity may have been stimulated by hsp 60 peptides in vitro which, under physiological in vivo conditions, are governed properiy by regulatory mechanisms in the periphery. Furthermore, in vivo class I presentation of shared hsp 60 epitopes may be insufficient for CD8 T-ceil activation as long as appropriate CD4 T cells are not present, since recognition in the absence of signals from CD4 T cells leads to CD8 T-cell paralysis rather than activation (Mueller et al. 1989). Although, in the physiological situation, autoreactive B cells and T cells of both types can be effectively controlled, their pathogenic potential remains and could be evoked under certain conditions, e.g., during bacterial infections.

Autoimmune disease hsp 60-specific T cells have been implicated in two experimental models of organspecific autoimmune diseases: adjuvant-induced arthritis and diabetes of NOD mice (van Eden et al. 1988, Elias et al. 1990). Furthermore, an involvement of T cells and/or antibodies with specificity to hsp 60 in systemic lupus erythematosus, rheumatoid arthritis, and insulin-dependent diabetes mellitus has been suggested (Gaston et al. 1989, Rajagopalan et al. 1990, Holoshitz et al. 1989, Jones et al. 1990). However, evidence to the contrary also exists (Atkinson et al. 1990, Kampe et al. 1990). The major islet-cell autoantigen in diabetes, which has a molecular mass of 60 kD. is not hsp 60 but the enzyme glutamic acid decarboxylase (Baekkeskow et al. 1990). A recent study suggested recognition of endogenous hsp 60 by gamma/delta T cells in lupus patients, based on blocking studies using mAb against mycobacterial hsp 60 (Rajagopalan et al. 1990). One of the mAb (IIIE9) employed is specific for M. leprae hsp 60 and does not react with CHO or HeLA cells nor with any of the various mycobacterial species tested besides M. leprae (Anderson et al. 1988, Buchanan et al. 1987, Dudani & Gupta 1989).

86

KAUFMANN ET AL.

Thus, the precise relation of hsp 60-specific T cells and B cells to autoimmune disease requires further analysis. We anticipate that a secondary facilitating role of hsp in autoimmune diseases is more likely than a primary initiating function. Perhaps hsp are expressed by target cells following attack by tissue-specific T cells and/or antibodies. Such cells could then become the targets for hsp-specific T cells and/or antibodies previously activated by shared epitopes of microbial origin. In any case, the findings that application of hsp 60 markedly interferes with experimentally induced autoimmune disease points to a regulatory role for this antigen (von Eden et al. 1988, Ellis et al. 1990). As discussed here, hsp play a multifactorial role in immunity to mycobacteria. Not only do they act as double-sided antigens - also functionally they markedly interfere with the host response. This is perhaps best illustrated by our observations that, on the one hand, stress caused by mycobacterial infections renders target cells resistant to killer cells, and that, on the other hand, stress results in recognition of target cells by killer cells activated by mycobacterial hsp. ACKNOWLEDGMENTS

'

The authors wish to thank Mrs. R. Mahmoudi for typing the manuscript and to Drs. A. Gatrill and G. Follows for critically reading it. The work described was supported by the following grants to S. H. E. Kaufmann: UNDP/Worid Bank/ WHO Special Program for Research and Training in Tropical Diseases, WHO Program for Vaccine Development, German Leprosy Relief Association. ECIndia Science and Technology Cooperation Program, A. Krupp Award for young professors, SFB 322, Landesschwerpunkt 30. M. E. Munk was recipient ofa grant by CNPq Brazil; T. Koga received support by the A. v. Humboldt Foundation. We thank Drs. J. Ivanyi, J. DeBruyn, H. Waldmann, R. Kubo, S. Modrow, R. W. Karr and R. A. Young for helpful reagents. REFERENCES Ab, B. K., Kiessling, R., van Embden, J. D. A., Thole, J. E. R., Kumararatne, D. S., Pisa, P., Wondimu, A. & Ottenhoff, T H. M. (1990) Induction of antigen-specific CD4* HLA-DR-restricted cytotoxic T lymphocytes as well as nonspecific nonrestricted killer cells by the recombinant mycobacterial 65-kDa heat shock protein. Eur. J. Immunol. 20, 369. Ananthan, J., Goldberg, A. L. & Voellmy, R. (1986) Abnormal proteins serve as eukaryotic stress signals and trigger the activation of heat shock genes. Science 232, 322. Anderson, D. C , Barry, M. E. & Buchanan, T. M. (1988) Exact definition of species-specific and cross-reactive epitopes of the 65 kiiodalton protein of Myeobacterium leprae using synthetic peptides. J. Immunol. 141, 607. Antia, N. H. (1982) Leprosy a disease ofthe Schwann cell. Lepr India 54, 599. Asamow, D. M., Kuziel, W. A., Bonyhadi, M., Tigelaar, R. E., Tucker, P. W. & Allison, J. P. (1988) Limited diversity ofy/S antigen receptor genes of Thy-1 * dendritic epidermal cells. Cell 55, 837.

hsp 60 IN BACTERIAL INFECTIONS

87

Asherson, G. L. (1968) The role of microorganisms in autoimmune responses. Progr. Allergy 12, 192. Atkinson. M. A., Maclaren. N. K. & Scharp, D. W. (1990) Letter to the editor. Lancet 336, 1251. Baekeskov, S., Jan-Aanstoot. H. & Christgau, S. (1990) Identification of the 65K autoantigen in insulin-dependent diabetes as the G ABA-synthesizing enzyme glutamic acid decarboxylase. Nature 347, 151. Blachard, D. K.. Stewart II, W. E., Klein, T. W., Friedmann, H. & Djeu. J. Y. (1987) Cytolytic activity ofhuman peripheral blood leukocytes against Legionella pneumophila-mfecled monocytes; characterization of the efTector cell and augmentation by interleukin 2. J. Immunol. 139, 551. Bom, W.. Hall, L., Dallas. A., Boymei, J., Shionick, T.. Young, D.. Brennan, P. & O'Brien, R. (1990b) Recognition ofa peptide antigen by heat shock reactive yjS T lymphocytes. Science 249, 67. Bom. W.. Happ, M. P.. Dallas, A., Reardon, C , Kubo, R., Shinnick. T., Brennan, P & O'Brien, R. (1990a) Recognition of heat shock proteins and y/S T cell function. Immunol. Today 11, 40. Buchanan, T. M., Nomaguchi, H., Anderson, D. C , Young, R. A., Gillis, T. P.. Britton, W. J., Ivanyi, J., Kolk. A. H. J., Doss, O., Bloom, B. R. & Mehra, V. (1987) Characterization of anti body-reactive epitopes on the 65 kilodalton protein of Mycobacterium leprae. Infect. Immun. 55. 1000. Buchmeier. N. A. & Heffron, F. (1990) Salmonella proteins induced following phagocytosis by macrophages are controlled by multiple regulons. Science 248, 730. Carbone. F. R., Moore, M. W., Sheil, J. M. & Bevan, M. J. (1988) Induction of cytotoxic T lymphocytes by primary in vitro stimulation with peptides. J. Exp. Med. 167, 1767. Carl, M. & Dasch, G. A. (1986) Characterization of human cytotoxic lymphocytes directed against cells infected with typhus group rickettsiae: evidence for lymphokine activation of effectors. J. Immunol. 136, 2654. De Bruyn, J.. Bosmans, R.. Tumeer, M., Weckx, M., Myabenda, J., van Vooren, J. P., Falmagne, P.. Wiker, H. G. & Harboe. M. (1987) Purification, partial characterization, and identification of a skin reactive protein antigen of Mycobacterium bovis BCG. Infect. Immun. 55. 245. Dudani. A. K. & Gupta, R. S. (1989) Immunological characterization ofa homolog of tbe 65-kiIodalton mycobacterial antigen. Infect. Immun. 57, 2786. Elias, D., Markovits, D.. Reshef, T.. van der Zee, R. & Cohen, I. R. (1990) Induction and therapy of autoimmune diabetes in the non-obese diabetic (NOD/Lt) mouse by a 65kDa heat shock protein. Proc. Nati. Acad. Sci. USA 87. 1576. Emmrich, R, Thole, J., van Embden, J. & Kaufmann, S. H. E. (1986) A recombinant 64 kiioDalton protein of Mycobacterium bovis BCG specifically stimulates human T4 clones reactive to mycobacterial antigens. J. Exp. Med. 163, 1024. Engers, H. D., Houba. V., Bennedsen. J., Buchanan, T. M., Chaparas, S. D., Kadival, G., Closs, O., David. J. R., van Embden, J. D. A., Godal, T.. Mustafa, S. A., Ivanyi, J., Young, D. B.. Kaufmann. S. H. E., Khomenko. A. G.. Kolk, A. H. J.. Kubin, M., Louis. J. A.. Minden. P.. Shinnick, T. M.. Trnka. L. & Young, R. A. (1986) Letter to the Editor: Results of a World Health Organization-sponsored workshop to characterize antigens recognized by Mycobacterium-specific monoclonal antibodies. Infect. Immun. 51.718. Fisch, P., Malkovsky, M.. Kovats, S., Stunn, E., Braakman. E., Klein. B. S., Voss, D., Morrissey, L. W., DeMars, R., Welch, W. J., Bolhuis, R. L. H. & Sondel, P M. (1990) Recognition of human Vy9/V^2 T-cells of a GroEL homolog on Daudi Burkitt's lymphoma cells. Science 250, 1269.

88

KAUFMANN ET AL.

Flesch, L & Kaufmann, S. H. E. (1987) Mycobaetedal growth inhibition by interferon-)' activated bone marrow macrophages and differential susceptibility among strains of Mycobaeterium tuberculosis. J. Immunol. 138, 4408. Flynn, G. C , Chappell, T G. & Rothman, J. E, (1989) Peptide binding and release by proteins implicated as catalysts of protein assembly. Science 245, 385. Gaston, J. S. H., Lofe, P., Jenner, P. J., Colston, M. J. & Bacon, P. A. (1990) Recognition of a mycobaeteria-specific epitope in the 65 kilodalton heal-shoek protein by synovial fluid-derived T-cell clones. J. Exp. Med. 771, 831. Gromkowski, S. H., Yagi, J. & Janeway, C. A. Jr. (1989) Elevated temperature regulates tumor necrosis factor-mediated immune killing. Eur. J, Immunol. 19. 1709. Gupta, R. S. (1990) Sequenee and structural homology between a mouse T-complex protein TCP-1 and the "chaperonin" family of bacterial (GroEL, 60-65 kDa heat shock antigen) and eukaryotie proteins. Biochem. Intern. 20, 833. Haregewoin, A., Soman, G., Hom, C. R. & Finberg, R. W. (1989) Human yd* T cells respond to mycobaeterial heat-shock protein. Nature 340, 309. Hendrix, R. W. (1979) Purification and properties of GroEL a host protein in bacteriophage assembly. / Mol. Biol. 129, 375. Hoiseth, S. K. & Stocker, B. A. D. (1981) Aromatic-dependent Salmonella typhimurium are non-virulent and effective as live vaeeines. Nature 291, 238. Holoshitz, J., Koning, F., Coligan, J. E., De Bruyn, J. & Strober, S. (1989) IsoIaUon of CD4- CD8- myeobacteria-reactive T lymphocyte clones from rheumatoid arthritis synovial fluid. Nature 339, 226. Ivanyi, J., Sinha, S., Aston, R., Cussell, D., Keen, M. & Sengupta, U. (1983) Definition of species speeific and eross-reaetive antigenic determinants of Mycobacterium leprae using monoclonal antibodies. Clin. Exp. Immunol. 52, 528. Jarjour, W.. Mizzen, L. A., Weleh, W. J., Denning, S., Shaw, M., Mimura. T., Haynes, B. F. & Winfield, J. B. (1990) Constitutive expression of a groEL-related protein on the surface of human eells. J. Exp. Med. 172, 1857. Jindal, S., Dudani, A. K., Singh, B., Harley, C. B. & Gupta, R. S. (1989) Primary structure of a human mitochondrial protein homologous to the bacterial and plant chaperonins and to the 65-kiloDaiton myeobacteria! antigen. Mol. Cell. Biol. 9, 2279. Jones, D. B., Hunter, N. R. & Duff, G. W. (1990) Heat-shock protein 65 as a j5 cell antigen of insulin-dependent diabetes. Lancet 336, 583. Kabelitz, D., Bender, A., Schondelmaier, S., Schoel, B. & Kaufmann, S. H. E. (1990) A large fraction of human peripheral blood yd"" T cells is activated by Mycobacterium tuberculosis but not by its 65 kD heat shock protein. J. Exp. Med. I7I, 667. Kampe, O., Velloso, L., Andersson, A. & Karlsson, A. (1990) Letter to the editor Lancet 336, 1250. Kantengwa, S., Donati, Y. R. A., Clerget, M., Maridonneau-Parini, I., Sinclair, F., Mariethoz, E., Rees, A. D. M., Slosman, D. O. & Polla, B. S. (1991) Heat shock proteins: an autoprotective mechanism for inflammatory cells? Sem. Immunol, (in press). Kaplan, G., Kiessling, R., Teklemariam, S., Hancock, G.. Sheftel, G., Job, C. K., Converse, P, Ottenhoff, T. H. M., Becx-Bleumink, M., Dietz, M. & Cohn, Z. A. (1989) The reeonstitution of cell-mediated immunity in the cutaneous lesions of lepromatous leprosy by recombinant interleukin 2. J. Exp. Med. 169, 893. Kaufmann, S. H. E. (1986) Immunology of leprosy: new facts, future perspectives. Microbial Pathogen. 1, 107. Kaufmann, S. H. E. (1987) Towards new leprosy and tuberculosis vaccines. Micobiol. Sciences 4, 324. Kaufmann, S. H. E. (1988) CD8+ T lymphocytes in intraeellular microbial infections Immunol. Today 9, 168.

hsp 60 IN BACTERIAL INFECTIONS

89

Kaufmann, S. H. E. (1990) Heat shock proteins: a link between rheumatoid arthritis and infection? Curr. Opin. Rheumatot 2, 430. Kaufmann, S. H. E. & Kabelitz, D. (1991) Gamma/delta T lymphocytes and heat shock proteins. Curr. Top. Microbiol. Immunol. 167, 191. Kaufmann, S. H. E., Sehoel, B., Wand-Wiirttenberger. A., Steinhoff, U., Munk, M. E. & Koga, T. (1990) T cells, stress proteins and pathogenesis of mycobaeterial infections. Curr. Top. Microbiol. Immunol. 155, 125. Kaufmann, S. H. E., Vath, U., Thole, J. E. R., v. Embden, J. D. A. & Emmrieh, F. (1987) Enumeration of T eells reactive with Mycobacterium tuberculosis organisms and specific for the recombinant mycobacterial 64 kiloDalton protein. Eur. J. Immunol. 11, 351. Klimpel, G. R., Niese!. D. W. & Klimpel, K. D. (1986) Natural cytotoxic effector cell activity against Shigella /lexneri-infccled HeLa cells. J. Immunol. 136, 1081. Koga, T., Wand-Wurltenberger. A., DeBruyn, J., Munk, M. E., Schoel, B. & Kaufmann, S. H. E. (1989) T eells against a bacterial heat shock protein recognize stressed macrophages. Science 245, 1112. Lamb, J. R., Bal. V, Mendez-Samperio, P., Mehlert, A.. So, A., Rothbard, J., Jindal, S., Young, R. A. & Young, D. B. (1989) Stress proteins may provide a link between the immune response to infection and autoimmunity. Int. Immunol. 1, 191. Langer, T. & Neupert, W. (1991) Heat shoek proteins hsp 60 and hsp 70: their roles in folding, assembly and membrane translocation of proteins. Curr. Top. Microbiol. Immunol. 167, 3. Mehra, V, Sweetser, D. & Young, R. A. (1986) Efficient mapping of protein antigenic determinants. Proc. Natl. Acad. Sci. USA 83, 7013. Mor, N. (1983) Intracetlular location of Mycobacterium leprae in macrophages of normal and immune-deficient mice and effect of rifampin. Infect. lmmun. 42, 802. Morrison, R. P., Belland, R. J., Lying, K. & Caldwell, H. D. (1989) Chlamydial disease pathogenesis. The 57-kD chlamydial hypersensitivity antigen is a stress response protein./. Exp. Med 170, 1271. Mueller, D. L., Jenkins, M. K. & Schwartz, R. H. (1989) Clonal expansion versus functional clonal inactivation: a costimulatory signalling pathway determines the outcome of T cell antigen receptor occupancy. Annu. Rev. Immunol. 7, 445. Muller, I. & Kaufmann, S. H. E. (1985) Antigen-reactivity pattern o f T cell hybridomas from Mycobacterium bovis BCG-infected mice. Infect, lmmun. 49, 838. Muller, I., Mayer, B., Brinkmann, V. & Kaufmann, S. H. E. (1986a) Autoreactive T-cell clones from mice infected with Mycobacterium bovis, strain bacillus Calmette-Guerin (BCG): I. Phenotype, specificity and in-vitro function. Immunobiology 171, 366. Muller, L, Rolink, A.. Freudenberg, N. & Kaufmann, S. H. E. (1986b) Biological functions in vitro and in vivo of cloned autoreactive T cells. In: Leukocyte.% in Host Defense, Oppenheim, J. J. & Jacobs, D. M., eds. Alan R. Liss, Inc., New York, pp. 325-330. Munk, M. E., Sehoel, B. & Kaufmann, S. H. E. (1988) T cell responses of normal individuals towards recombinant protein antigens of Mycobacterium tuberculosis. Eur. J. Immunol. 18, 1835. Munk, M. E., Schoel, B., Modrow, S., Karr, R. W., Young, R. A. & Kaufmann, S. H. E. (1989) Cytolytic T lymphocytes from healthy individuals with specificity to self epitopes shared by the mycobacterial and human 65 kDa heat shoek protein. J. Immunol. 143, 2844. Munk, M. E., Shinnick, T. M. & Kaufmann, S. H. E. (1990) Epitopes ofthe Mycobacterium tuberculosis 65 kDa heat shock protein for human T cells comprise different secondary protein structures. Immunobiology 180, 272. Pierce, S. K., De Nagel, D. C. & van Buskirk, A. M. (199!) A role for heat shock proteins in antigen processing and presentation. Curr Top. Microbiol. Immunol. 167, 83. Rajagopalan, S., Zordan, T, Tsokos, G. C. & Datta, S. K. (1990) Pathogenic anti-DNA

90

KAUFMANN ET AL.

autoantibody-inducing T helper cell lines from patients with active lupus nephritis: Isolation of CD4"8" T helper cell lines that express the y/S T-cell antigen receptor Proc. Nati Acad. Sci. USA 87, 7020. Renkonen, R., Ristimaki, A. & Hayry, P. (1988) Interferon-gamma protects human endothelial cells from lymphokine-activated kifler cell-mediated lysis. Eur. J Immunol 18 1839. Shinnick, T. M. (1991) Heat shock proteins as antigens of bacterial and parasitic pathogens. Curr Top. Microbiol. Immunol. 167, 145. Shoenfeld, Y. & Isenberg, D. A. (1988) Mycobacteria and autoimmunity. Immunol. Todav 9, 178. Srivastava, P. K. & Maki, R. G. (1991) Stress-induced proteins in immune response to cancer. Curr. Top. Microbiol. Immunol. 167, 109. Steinhoff, U. & Kaufmann, S. H. E. (1988) Specific lysis by CD8-^ T cells of Schwann cells expressing Mycobacterium leprae antigens. Eur. J. Immunol. 18, 973. Steinhoff, U., Schoel, B. & Kaufmann, S. H. E. (1990) Lysis of interferon-y activated Schwann cells by crossreactive CD8 ajfi T cells with specificity to the mycobacterial 65 kDa heat shock protein. International Immunol. 2, 279. Steinhoff, U., Wand-Wurttenberger, A.. Bremerich, A. & Kaufmann, S. H. E. (1991) Mycobacterium leprae renders Schwann cells and mononuciear phagocytes susceptible or resistant against killer cells. Infect. Immun. 59, 684. Sugawara, S., Nowicki, M., Xie, S., Song, H. J. & Dennert, G. (1990) Effects of stress on lysability of tumor targets by cytotoxic T cells and tumor necrosis factor. J. Immunol 145, 1991. Thole, J. E. R., Keulen, W. J., Kolk, A. H. J., Groothuis, D. G., Berwald, L. C , Tiesjema, R. H. & van Embden, J. D. A. (1987) Characterization, sequence determination, and immunogenicity of a 64-kilodalton protein of Mycobacterium bovis BCG expressed in Escherichia coli K-I2. Infect. Immun. 55, 146. Townsend, A. & Bodmer, H. (1989) Antigen recognition by class I-restricted T lymphocytes. Ann. Rev. Immunol. 7, 601. Tsoulfa, G., Rook, G. A. W, van Embden, J. D. A., Young, D. B., Mehlert, A., Isenberg. D. A., Hay, F. C. & Lydyard, R M. (1989) Raised serum TgG and IgA antibodies in mycobacterial antigens in rheumatoid arthritis. Ann. Rheum. Dis. 48. 118. Van Eden, W., Thole, J. E., van der Zee, R., Noordzij, A., van Embden. J. D., Hensen, E. J. & Cohen, I. R. (1988) Cloning of the mycobacterial epitope recognized by T iymphocytes in adjuvant arthritis. Nature 331, 171. Wand-Wurttenberger, A., School, B., Ivanyi, J. & Kaufmann, S. H. E. (1991) Surface expression by mononuclear phagocytes of an epitope shared with mycobacterial heat shock protein 60. Eur J. Immunol, (in press).