Human 65-kDa hsp-reactive Tcells

Eur. J. Immunol. 1991. 21: 2139-2143

Elisabeth Hermanno, Ansgar W. LohseO, Ruurd Van der Zeeno, Willem Van Eden., Werner J. Mayet, Peter Probst, Thomas Poralla, Karl-Hermann Meyer zum Buschenfelde and Bernhard Fleischer I. Medizinische Klinik und Poliklinik, Johannes-GutenbergUniversitat, Mainz, Laboratory of Bacteriology., National Institute of Public Health and Environmental Protection, Bilthoven and Department of Infectious Diseases and Immunologyv,Veterinary Faculty, State University of Utrecht, Utrecht

Synovial fluid-derived Yersinia-reactive T cells responding to human 65-kDa heat-shock protein and heat-stressed antigen-presenting cells* Humoral and cellular immune reactions to heat-shock proteins have been implicated in the pathogenesis of arthritis. Heat-shock proteins occur in bacteria as well as all eukaryotes and have been highly conserved during evolution. Cross-reactivity between bacterial and human heat-shock proteins induced at the site of inflammation may underlie the pathogenesis of some forms of arthritis. In order to test this hypothesis, we raised and cloned a Yersinia-specificT cell line from the synovial fluid lymphocytes of a patient with Yersinia-induced reactive arthritis. From this line we obtained a CD4+ Tcell clone that proliferated in response to Yersinia antigens and both to the mycobacterial and the human 65-kDa heat-shock protein. This T cell clone also proliferated in response to autologous heat-stressed antigen-presenting cells as well as to synovial fluid mononuclear cells from the inflamed joint, thus showing true autoreactivity against endogenously synthetized self-antigen. These results demonstrate the induction of an autoimmune Tcell response by a natural bacterial infection and support the important role of heat-shock proteins in the pathogenesis of immune-mediated arthritis.

1 Introduction Heat-shock proteins (hsp) are induced in bacteria and eukaryotes under various conditions of stress [ l , 21. hsp are grouped according to molecular mass. The 60-kDa hsp family, which has been highly conserved during evolution, has been found t o be associated with autoimmune reactions. Animal models of autoimmune diseases have provided the initial evidence for an important role of hsp in the pathogenesis of autoimmune diseases. Studies of adjuvant arthritis have shown a Tcell response against the 65-kDa mycobacterial hsp to be responsible for the disease and the epitope in this model has been mapped onto amino acids 180-186 of this 65-kDa hsp [3,4]. Pretreatment of animals with the mycobacterial65-kDa hsp was effective not only in preventing adjuvant arthritis [3], but could also protect animals from streptococcal cell wall arthritis [5] or pristaneinduced arthritis [6], suggesting a role for 65-kDa hsp epitopes in different models of arthritis. Cross-reactivity between 65-kDa hsp epitopes and cartilage proteoglycans might account for these effects [3]. These animal experiments have led to the discovery of T cell reactivity against the mycobacterial65-kDa hsp in the synovial fluid (SF) of

[I 93631

*

A

2139

This work was done in the framework of the European Concerted Action on the Immunopathogenesis and Immunotherapy of Chronic Arthritis. Supported from the Kurt-Eberhard Bode Stiftung im Stifterverband fiir die Deutsche Wissenschaft. Supported from the Deutsche Forschungsgemeinschaft (DFGLo 368/2-2 and SFB311. Supported from the Institut Meneux.

Correspondence: Elisabeth Hermann, I. Medizinische Klinik, Johannes-Gutenberg-Universitat, D-6500 Mainz, FRG Abbreviations: hsp: Heat-shock protein S F Synovial fluid PBMC: Peripheral blood mononuclear cells SI: Stimulation index 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991

patients with active arthritis [7, 81. This T cell reactivity is not confined to any particular form of arthritis [9]. Reactivity against the human 65-kDa hsp, however, has thus far not been demonstrated in arthritis. Reactive arthritis is an immune-mediated process triggered by an infection. Gastrointestinal or genital infections usually precede the condition by a few days or weeks. Enterobacteria such as Salmonella, Shigella or Yersinia can often be identified as the causative organisms. Tcell clones reactive with the triggering microorganisms have recently been isolated from the SF of patients with Yersinia arthritis [lo] and Salmonella arthritis [ l l ] . In this study we have isolated from the SF of a patient with Yersinia-induced arthritis a Yersinia-reactive Tcell line and clone that was cross-reactive with human 65-kDa hsp and heat-stressed autologous cells.

2 Materials and methods 2.1 Patient and preparation of SF and PBMC A 19-year-old male patient presented with acute Reiter’s syndrome with oligoarthritis of both knees, the right ankle and the right wrist, conjunctivitis, and urethritis. He had a history of malaise, fever and diarrhea 2 weeks before the onset of arthritis. Stool cultures were negative, but seroconversion with a positive agglutination test for Yersinia enterocolitica 09 OH (1 :640) took place 1week after the first arthritic symptoms. Urethra smears and serum antibodies were negative for Chlamydia or mycoplasma. HLA typing revealed an HLA-B27- and DR2,DR4-positive phenotype. The patient required therapeutic arthrocentesis of the left knee. The SF was sterile and contained 12000 white cells/mm3. SF and peripheral blood (PB) samples were collected in heparinized tubes. Mononuclear cells were separated by Ficoll-Hypaque (Pharmacia, Freiburg, FRG) gradient centrifugation, washed three times and resuspended in RPMI 1640 (Gibco, Karlsruhe, FRG) con-

+

0014-2980/91/0909-2139$3.50 .25/0

2140

E. Hermann, A. W. Lohse, R.Van der Zee et al.

taining 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin and 10% heat-inactivated human AB serum. Monocyte- and B cell-enriched PB mononuclear cells (PBMC-E-) were obtained by removing T cells after rosetting with neuraminidase-treated SRBC. PBMC-E- were used as APC in the proliferation assays. 2.2 Antigens

Eur. J. Immunol. 1991.21: 2139-2143 reactive T cell line JP1 was performed in Terasaki microtest plates (Nunc, Roskilde, Denmark) by LD using Yersinia enterocolitica 09 (50 pg/ml) and rIL2 (20 U/ml)-supplemented medium and autologous irradiated feeder cells (1.2 x 104cells/well). Mitogens were not used in the cloning protocol. Growing Tcell clones were transferred to larger plates and then tested for reactivity to Yersinia in a proliferation assay. Yersinia-reactive clones were expanded further.

Yersinia enterocolitica 09 and Salmonella typhimurium, both isolated from patients with reactive arthritis, E . coli 2.5 T cell proliferation assay for T cell clones and Candida albicans were killed by heat, checked for sterility, washed and suspended in PBS. For the prolifera- Proliferative responses of theT cell clones were determined tion assays the bacterial antigens were diluted in the culture in triplicates in round-bottom 96-well plates using 104 medium mentioned above to a final concentration of cloned cells as responders and 104 autologous PBMC-E- as 50 pg/ml, a concentration previously determined to be AF'C. Cells were cultured for 48 h and pulsed with 0.25 pCi optimal [lo, 111. Tetanus toxoid (kindly provided by Beh- [3H]dThdfor the last 24 h of culture.The response of clone ringwerke, Marburg, FRG) was used at 10 pg/ml. hsp from JP1.2 to the 65-kDa hsp was tested for inhibition by a panel Mycobacterium bovis BCG (expressed from plasmid of mAb against HLA class I1 determinants. The mAb used pRIB 1300, see [3]) and the recombinant P-galactosidase were DA6.231 (kindly provided by Dr. V. Van Heyningen, human 65-kDa fusion protein, as well as P-galactosidase Edinburgh, Scotland) directed against common class I1 expressed from the cloning vector, pEX2, were added at determinants, the anti-HLA-DR antibody L243, the anti1 pg/ml and 5 pg/ml.The human 65-kDa hsp fusion protein HLA-DP antibody B7/21 and the anti-HLA-DQ antibody was expressed in E. coli from plasmid pRH710 (Van der Tii22. Zee, R. et al., to be published elsewhere). This pEX2derived plasmid was constructed from a phage hgtll clone The proliferative response of clone JP1.2 and of control that we obtained from Dr. R. A. Young, Boston, MA. clones from the same JP1 line to heat-stressed mononuclear pEX2-derived proteins were isolated as the insoluble cells were tested with culture medium alone and with material after lysis of recombinant E. coli strains. In additional Yersinia antigen. Irradiated autologous or alloaddition various deletion mutants of the mycobacterial geneic PBMC and autologous SFMC at 50000 cells/well 65-kDa protein were used in order to narrow down the were preincubated for 2 h at 37 "C or at 42 "C in 96-well epitope recognized by anti-65-kDa Tcells. These 0-galacto- microtiter plates. The responder cloned T cells were subsesidase fusion proteins were expressed in E . coli from quently added to the cultures with or without Yersinia mutants (plasmids pRIB 1404, pRIB 1426, pRIB 1444) of antigen at 10000 cells/well. Cultures were incubated at the gene encoding the 65-kDa protein fused to lacZ as 37°C for 48 h and measured as above. described [3]. PHA (Gibco) was used at 1 pg/ml.

3 Results 2.3 Lymphocyte prolieration assays PBMC and SF mononuclear cells (SFMC) were placed in round-bottom microtiter plates (Nunc, Roskilde, Denmark) at 5 x lo4cells/well with or without antigen in a final volume of 100 p1 of culture medium. All tests were performed in triplicates. Cultures were incubated for 6 days at 37 "C in a humidified atmosphere of 5% COz. On day 5 0.25 WCi = 9.25 kBq of [3H]dThd (Amersham Int., Amersham, GB) was added to each well. On day 6 the cells were harvested and [3H]dThd incorporation measured using a liquid scintillation counter. 2.4 Generation and cloning of an SF-derived Yersinia-reactiveT cell line, JP1

The JP1 Yersinia-reactive Tcell line was established by incubating SFMC (106/ml) with the Yersinia enterocolitica 09 antigen preparation at 50 pg/ml in culture medium in 12-well plates (Costar, Cambridge, MA). After 4 days 20 U/ml rIL2 (Boehringer Mannheim, Mannheim, FRG) were added. At 10-day intervals theTcell line was restimulated with Yersinia antigen and 1.5 x lo6 irradiated (4000 rad) autologous PBMC three times and propagated with IL 2-supplemented medium. Cloning of this Yersinia-

The proliferative response of PBL and SF lymphocytes against bacterial antigens and recombinant mycobacterial and human 65-kDa hsp are shown in Fig. 1. A weak but distinct response of PBL to Yersinia antigens could be seen. Only a very weak response against the hsp was measured in the PB. SF lymphocytes on the other hand showed a very marked response not only to the Yersinia antigens but also both to the mycobacterial 65-kDa hsp and to the pgalactosidase human 65-kDa fusion protein. Other bacterial antigen preparations (Salmonella, Escherichia coli and Tetanus) also induced proliferative responses, but no significant proliferation was induced by the vector medium (P-galactosidase expressed from pEX2) of the recombinant human 65-kDa protein in the concentrations used. A Yersinia-reactive Tcell line was established from the SF lymphocytes by stimulation with heat-killed Yersinia enterocolitica 09 organisms. After two cycles of antigen stimulation and expansion with IL2 the Tcell line was cloned by LD technique. Of 84 clones obtained 47 showed a significant response to Yersinia antigen.The 10 clones that showed the best responses were selected and characterized further. Seven out of the ten clones reacted only with Yersinia antigens, the other three also showed differing levels of response to other enterobacteria. One clone, JP1.2,

Eur. J. Immunol. 1991. 21: 2139-2143 PHA

Human 65-kDa hsp-reactive Tcells

b

TT Candtda alb E coil

S typhimur human 65KD hsp

w

mycob 65KD hsp

SFMC PBMC

Y enterocol 09

2141

due to low amounts of the E. coli hsp contained in the pEX2 preparation. None of the other clones showed a similar reactivity pattern (Table 1).The proliferative response of clone JP1.2 to various deletion mutants of the mycobacterial 65-kDa are shown in Table 2. From these results it appears that the cross-reactive epitope of the human and mycobacterial65kDa hsp recognized by JP1.2 is localized on the carboxy-terminal part of the protein between amino acid residues 370 and 540.

medium

Figure 1. Response of PB and SF mononuclear cells of the patient with Yersinia-induced reactive arthritis to bacterial antigens and the 65-kDa hsp.

JP1.2 was found to be a CD4+CD8- T cell clone with an dP TcR. JP1.2 as well as the other clones characterized were Vg5 and Vg8 negative. The proportion of Vp5+ and Vp8+ cells in the PB and the SF was within the normal range. JP1.2 recognized the 65-kDa in a dose-dependent and class 11-restricted manner as shown in Fig. 2. Inhibition studies with mAb suggested the response to be DRrestricted (Fig. 2).

showed a strong response to Yersinia antigen, other enterobacteria, Mycobacterium tuberculosis, the recombinant mycobacterial 65-kDa hsp and the recombinant human 65-kDa hsp (Table 1). JP1.2 did not respond to the vector medium (P-galactosidase expressed from pEX2). It also showed no reaction to Candida antigen and Tetanus toxoid (Table 1). M. tuberculosis was able to induce a similary strong response as the mycobacterial 65-kDa hsp (results not shown), which confirmed that the response was not to contaminants of the vector medium. Only very high concentrations of the pEX2-P-galactosidase preparation (20 pg/ml) were able to induce a low proliferative response (stimulation index of 5 ;Table 2). This response is probably

In order to test the response of the JP1.2 clone to endogenously synthesized hsp PBL of the patient were incubated at 37°C or 42°C for 2 h. JP1.2 or control clone cells were then added with and without antigen. JP1.2 cells showed a marked proliferative response to heat-stressed APC (Fig. 3). None of the other clones responded to heat-stressed APC (Fig. 3). HLA-incompatible heatstressed APC were not able to stimulate JP1.2 cells (Fig. 3). This is in accordance with the above findings showing the 65-kDa hsp response to be HLA-DR restricted. SF cells taken from the inflamed joint were by themselves able to stimulate JP1.2, and this effect could be augmented by prior heat stress (Fig. 3).

0

20000

10000

30000

40000

CPm

Table 1. Proliferative response (SI) of clone JP1.2 and control clone JP1.3 to various antigens

Clone JP1.3 cpm SI

Clone JP1.2

Antigen

CPm

SI

Medium control Yersinia enterocolitica Salmonella typhimwium E. coli Tetanus toxoid

133 11499

Candida albicans

160 535 125 10071 1097 8949 2033

pEX2-p-galactosidase Mycobact. 65-kDa hsp Human 65-kDa hsp

4665

1871

98 5wml l&ml 5&ml l&ml 5Wml 1&d

110 6976 493 245 423 270 101 NT 72 NT 105 NT

86.4 35.1 14.1 Cl

1.2 4.0 1.o

75.7 8.3 67.2

15.3

63 4.5 2.2 3.8 2.5 1 NT 50% identical and contain many conservative substitutions. However, the epitope on the mycobacterial6.5-kDa protein known to be critical in the induction of adjuvant arthritis in Lewis rats is not present on the human 65-kDa hsp [3, 131. Reactivity against the mycobacterial 65-kDa hsp has been described in various forms of arthritis [7-91. Gaston et al. recently established Tcell clones from the SF of a patient with reactive arthritis that recognized mycobacterial 65-kDa hsp, but the epitope recognized is not

We have in this study been able to show that Yersiniureactive SF lymphocytes from a patient with a Yersiniuinduced arthritis are cross-reactive with recombinant human 65-kDa hsp. We have, moreover, been able to show reactivity of the 65-kDa hsp-specific T cell clone JP1.2 against autologous heat-stressed APC. As the 65-kDa hsp of Yersiniu is not yet available it has not been possible to formally demonstrate that JP1.2 is indeed reactive with the Yersiniu hsp. However, the reactivity pattern of JP1.2 with its distinct response to various enterobacteria known to contain related hsp and the recombinant mycobacterial as well as the human 65-kDa hsp strongly suggests that it was triggered by Yersinia hsp. JP1.2 recognizes an epitope different from the mycobacteria-specific epitope on amino acids 180-188 recognized by T cells in adjuvant arthritis [3]. On the contrary JP1.2 is specific for a highly conserved epitope on human, mycobacterial and enterobacteria hsp. Low concentrations of the E. coli hsp are probably present in the recombinant protein preparations, which are able to induce a weak proliferative response of JP1.2 cells when high concentrations of the pEX-fl-galactosidase are used as antigen. However, these possible E. coli contaminants cannot be responsible for the vigorous Tcell responses to the recombinant human and mycobacterial65-kDa hsp. In addition, specificity of the response was confirmed by stimulation with pure mycobacteria. hsp are endogenous proteins whose synthesis is markedly enhanced by various forms of stress. The inflamed joint in arthritis provides conditions that will induce stress proteins [14, 151. The Tcells reacting to the 65-kDa hsp might therefore also react against autologous heat-stressed APC, as has been shown to be experimentally inducible in mice [16], and such autoreactivity may be critical in the inflamed joint. It has also been demonstrated in mice that y/6 T cells which have been primed in vivo can be further stimulated in vitro by heat-shock-induced self-stress proteins [17]. The TcR a/p+CD4+ clone JP1.2 showed a marked response both against the recombinant human 65-kDa hsp and against heat-stressed autologous APC. It is thus a truly autoreactive clone. SFcells taken from the inflamed joint of the patient were able to stimulate the JP1.2 clone without prior heat stress, although heat stress could enhance this stimulation (Fig. 3). Presumably these SFcells when taken from the patient were already expressing 65-kDa hsp due to the conditions in the inflamed joint. An additional explanation for the phenomenon that untreated synovial cells were recognized by JP1.2 could possibly be the presentation of Yersiniu hsp by these cells, since there is evidence that bacterial antigens implicated in the pathogenesis of reactive arthritis are indeed present in the affected joints [181.The increased response to heat-stressed synovial cells might then be due to the additional presentation of endogenous 65-kDa hsp epitopes on the synovial cells. JP1.2 may provide the missing link between infection and autoimmunity. It is a Tcell clone that was induced by a natural infection and was found to be reactive both with the infectious agent and a self component of the host. Studies of Tcell reactivity with hsp have so far suggested that the

Eur. J. Imrnunol. 1991. 21: 2139-2143

target epitopes recognized are those not present on the human host hsp [9]. O n the other hand, with the help of synthetic peptides it has been possible to induce cytotoxic Tcells with specificity for peptides present both on the mycobacterial and human 65-kDa hsp in normal individuals, suggesting that 65-kDa hsp-specific autoreactive T cell may be part of the normal T cell repertoire [191. It is conceivable that the occurrence of an autoreactive 65-kDa hsp-specificT cell clone such as JP1.2 plays a role not only in the pathogenesis of arthritis, but also in its control. By initiating an anti-idiotypic response it may trigger its own down-regulation as has been described in T cell vaccination with autoreactive Tcells in animal models of experimental autoimmune diseases [20, 211. In summary we have been able for the first time to isolate a Tcell clone specific for human 65-kDa hsp from the inflamed joint of a patient with reactive arthritis. In addition, we have also for the first time been able to demonstrate reactivity of human anti-65-kDa hsp-specific T cells against autologous heat-stressed cells. These results support the importance of hsp in arthritis.The role of hsp in the pathogenesis of autoimmune disease as suggested by animal models may be based on their high degree of conservation and prominent immunogenicity leading to the induction of cross-reactive T cell clones such as JP1.2 responding both to the microbial and the self hsp. We wish to thank Dr. R. A . Young for providing us with the hgtIl clone containing the human 65-kDa hsp gene. We are grateful to Dr. 1. R. Cohen and Dr. J. D. A . Van Ernbden for helpful discussions. Technical assistance by Ms. S. Meyer and Ms. J. M . Versantvoort is gratefully acknowledged. Received February 10, 1991; in revised form June 10, 1991.

5 References 1 Young, D. B., Ivanyi, J,. Cox, J. H. and Lamb, J. R., lrnrnunol. Today 1987. 8: 215. 2 Cohen, I. R., Annu. Rev. Irnrnunol. 1991. 9: 567.

Human 65-kDa hsp-reactive T cells

2143

3 Van Eden, W., Thole, J. E. R. ,Van der Zee, R., Noordzij, A., Van Embden, J. D. A., Hensen, E. J. and Cohen, I. R., Nature 1988. 331: 171. 4 Van der Zee, R.,Van Eden, W., Meloen, R. H., Noordzij, A . and Van Embden, J. D. A., Eur. J. Irnrnunol. 1989. 19: 43. 5 Van der Broek, M. F.,Van Bruggen, M. C. J., Hogervorst, E. J. M.,Van Eden, W.,Van der Zee, R. and Van der Berg, W. B., J. Exp. Med. 1989. 170: 449. 6 Thompson, S. J., Rokk, G. A., Bredley, R. J.,Van der Zee, R. and Elson, C. J., Eur. J. Imrnunol. 1990. 20: 2479. 7 Res, l? C. M., Schaar, C. G., Breedveld, F. C. ,Van Eden,W.,Van Embden, J. D. A., Cohen, I. R., DeVries, R. R. l?, Lancet 1988. ii: 478. 8 Gaston, J. S. H., Life, P. F., Bailey, L. C. and Bacon, P. A., J. Imrnunol. 1988. 143: 2494. 9 Gaston, J. S. H., Life, l? E , Jenner, P. J., Colston, M. J. and Bacon, l? A., J. Exp. Med. 1990. 171: 831. 10 Hermann, E., Fleischer, B., Mayet,W.-J., Poralla,T. andMeyer zum Biischenfelde, K.-H., Clin. Exp. Irnrnunol. 1989. 75: 365. 11 Hermann, E., Mayet,W.-J., Poralla, T., Meyer zum Biischenfelde, K.-H. and Fleischer, B., Scand. J. Rheurnatol. 1990. 19: 350. 12 Young, R. A., Annu. Rev. Irnrnunol. 1990. 8: 401. 13 Jindal, S., Dudani, A. K., Singh, B., Harley, C. B. and Gupta, R. S., Mol. Cell. Biol. 1989. 9: 2279. 14 Kubo, T., Towle, C. A., Mankin, H. J. and Treadwell, B. V., Arthritis Rheum. 1985. 28: 1140. 15 Karlsson-Parra, A., Soderstrom, K., Ferm, M., Ivanyi, J., Kiessling, R. and Klareskog, L., Scand. J. Irnrnunol. 1990.31: 283. 16 Munk,M. E., Schoel,B.,Modrow, S., Karr, R.W.,Young, R. A. and Kaufmann, S. H. E., J. Irnrnunol. 1989. 143: 2844. 17 Rajasekar, R., Sim, G.-K. and Augustin A., Proc. Natl. Acad. Sci. USA 1990. 87: 1767. 18 Granfors, K., Jalkanen, S.,Von Essen, R., Lahesmaa-Rantala, R., Isomaki, O., Pekkola-Heino, K., Merilahti-Palo, R., Saario, R., Isomaki, H. and Toivanen, A., N. Engl. J. Med. 1989. 320: 216. 19 Koga,T.,Wand-Wiirttenberger, A., DeBruyn, J., Munk, M. E., Schoel, B., Kaufmann, S. H. E., Science 1989. 245: 1112. 20 Lohse, A. W., Cohen, I. R., in Demaine, A. G., Banga, J. P., McGregor, A. M. (Eds.), Molecular Biology of Autoimmune Disease, Springer Verlag, New York 1990, p. 333. 21 Lider, O., Reshef,T., Beraud, E., Ben-Nun, A. R. and Cohen, I. R., Science 1988. 239: 181.