Surface expression of a cross-reactive hsp epitope
Eur. J. Immunol. 1991. 21: 1089-1092
1089
Short paper Angela Wand-Wiirttenbergero, Bernd Schoelo, Juraj IvanyiA and Stefan H. E. KaufmannO Department of Immunologyo, University of Ulm, Ulm and MRC Tuberculosis and Infections UnitA, Hammersmith Hospital, London
Surface expression by mononuclear phagocytes of an epitope shared with mycobacterial heat shock protein 60* Bone marrow-derived macrophages were stained with a variety of monoclonal antibodies (mAb) with specificity for the mycobacterial heat shock protein (hsp) 60 or with specific rabbit antiserum raised against mycobacterial hsp 60.The mAb h4L30 as well as the specific antiserum brightly stained bone marrowderived macrophages whereas all the other mAb as well as normal rabbit antiserum gave no or negligible reactivity. Surface expression of the shared epitope is observed by day 3 of in vitro cultivation and markedly increased by interferon-y stimulation on day 9. Although hsp 60 is thought to be restricted to the mitochondria1 compartment, antibody responses to this molecule have been implicated in autoimmunity. Our finding that bone marrow-derived macrophages express a cross-reactive epitope recognized by an mAb with specificity for amino acids 275 to 295 of the mycobacterial hsp60 is consistent with such a role.
1 Introduction Members of the heat shock protein (hsp) 60 family are essential for the folding, unfolding and translocation of intracellular proteins [ l , 21. Although their synthesis is increased under certain stress conditions, they are constitutively produced. hsp 60 are ubiquitous and highly conserved and cognates of bacterial and mammalian origin show about 60% homology [3,4]. Thus, hsp are potential targets for autoimmune responses and could provide a link between microbial infection and autoimmunity [3]. On the other hand, in mammalian cells the localization of hsp 60 is considered to be restricted to the mitochondrial compartment and it is thought that hsp, in general, are not expressed on the cell surface [ 5 , 6 ] . Here we show that an mAb, ML30, specific for amino acids 275 to 295 of mycobacterial hsp 60 as well as a specific rabbit anti-hsp 60 antiserum recognize a shared epitope on the surface of BM-derived MQ, (BMMQ) indicating that antibodies against bacterial hsp cross-react with surface molecules expressed by host cells.
2 Materials and methods 2.1 Cells BMMQ from C57BL/6 mice were used as a homogeneous population of mouse MQ,. Unless otherwise stated, BM
[I 91091
* This work received financial support from SFB 322, UNDPlWorld B a n W H O Special Program for Research and Training in Tropical Diseases, German Leprosy Association and A. Krupp award for young professors (S.H.E.K.).
Correspondence: Stefan H. E. Kaufmann, Department of Immunology, University of Ulm, Albert-Einstein-Allee 11, D-7900 Ulm, FRG Abbreviation: hsp: Heat shock protein 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991
cells taken from mouse femura were cultivated for 9-1 1 days in hydrophobic plastic bags as described earlier [7, 81. In some experiment cells were stimulated with rIFN-y.
2.2 Antisera and mAb Rabbit anti-hsp 60 antiserum was generated by immunizing a rabbit with 1 mg purified rhsp60 from M. bovis BCG in Ribi adjuvant i.m. hsp 60 for immunization was produced, isolated and purified from E. coli clone M1546 [9] essentially as described earlier [ 101.The hyperimmune polyclonal rabbit IgG fraction (rb-a-60) was prepared by affinity purification on a protein A-Superose (Pharmacia, Freiburg, FRG) column and titrated for saturating concentration on BMMQ cells by FCM. As a negative control we used a normal rabbit IgG preparation which had been absorbed on an hsp 60 affinity column (CNBr-Sepharose coupled with purified rhsp 60 of M. bovis BCG following the instruction of Pharmacia) to remove contaminating naturally occurring anti-hsp 60 antibodies. The following mAb were used for immunostaining: Mac-1 (rat IgG [ 111; as a positive control); Leu-19 (Becton Dickinson, Mountain, View, CA; as subclass-matched negative control to assess background binding of IgGl to FcR); mAb IVD1, IA1, XIXHS, XVIIIGl and IIGl (mouse anti-M. bovis BCGhsp 60 mAb of IgG2b(IA1) or IgGl (all other mAb) subclass (a gift of J. DeBruyn, Bruxelles); IIC8 (mouse IgGZt, anti-M. leprue hsp 60 mAb; gift of T. Gillis, [12]); ML30 (mouse) IgG1; [13]); and L7 (mouse IgGl mAb recognizing a species cross-reactive determinant on hsp 70 of M. feprue ([14]; gift of A. Basten, Sydney).
2.3 FCM BMMQ were harvested on day 11 and washed twice. FcR were blocked by irrelevant IgG for 15 min in RF'MI 1640 (Gibco, Grand Island, NY) supplemented with 1% mouse serum (when staining with rabbit IgG) or 1% rabbit serum 0014-2980/91/0404-1089$3S O + .25/0
Eur. J. Immunol. 1991. 21: 1089-1092
A. Wand-Wiirttenberger. B. Schoel, J. Ivanyi and S. H. E. Kaufmann
100
(when staining with mouse mAb). All steps were performed in 0.1% NaN3 on ice to prevent internalization of antibodies. Antibodies were diluted in the appropriate blocking buffer and incubated for 30min. For washing, ice-cold PBS containing 0.1% azide was used. Second antibodies were FITC-labeled affinity-purified F(ab’)z fragments of goat anti-mouse IgG and goat anti-rabbit IgG (Medac, Hamburg, FRG). Labeled cells were fixed in 2% paraformaldehydePBS and protected from light until FCM analysis.
,601 I80
2o ”
Murine BMMQ, were stained with various mAb with specificity for mycobacterial hsp 60 or hsp 70 and analyzed thereafter by FCM. Only the ML30 mAb which recognizes amino acids 275 to 295 of the mycobacterial hsp 60 showed positive staining of BMMQ, (Fig. 1). A specific rabbit Ig preparation raised against mycobacterial hsp60 also stained these cells. Fig. 2 shows a representative compari-
14)
120 IW
ar-
0
XI
~7
nm
.
0
3 Results
0
1
IWI
IAI
XIM
XVIIEI
mi
1161
iim
ldlw
MI
iuio
*ulo
Figure I . FCM analysis of mouse BMMQ stained with a panel of antibodies against mycobacterial hsp 60 and hsp70. Mac-1 and Leu-I9 served as positive or negative control, respectively. “Auto” shows the autofluorescent background of unstained cells.
TL 1
FL 1
. 1
. 2
. 3
. 4
.
.
.
5 6 7 Culture (days)
.
.
8
9
.
.
1 0 1 1
Figure 3. Analysis of cross-reactive epitopes during the 11 days of in vitro differentiation. Open circles: Mac- 1;closed circles: control IgG; open squares: ML30; closed squares: rb-a-hO.
son between the cross-reactive mAb ML30 and mAb L7, and between the rb-a-60 and control rb-IgG , respectively. Both the ML30 mAb and the rb-a-60 identify a distinct cell population whereas the normal rb-IgG and the L7 mAb produce a small shift in peak intensity only. This weak background staining by normal rb-IgG could be reduced by passage through an hsp 60-Sepharose column indicating that normal rb-IgG already contained antibodies against cross-reactive epitopes. In an attempt to follow surface expression in the course of BMMQ, development, the cells were stained daily during their in vitro cultivation for 11 days (Fig. 3). Between day 3 and day4, a steep increase in the expression of the cross-reactive epitope recognized by ML30 and rb-a-60 was observed which remained relatively stable until day 9. From day 10 to day 11 a further twofold increase in surface expression was detected with the ML30 mAb, whereas staining with the rb-a-60 was slightly reduced.Thus, the two antibody preparations may recognize different epitopes.
FL 1
FL I
Figure 2. Expression of an hsp 60 cross-reactive epitope. Positive staining of BMMQ, by anti-hsp 60 antibodies (left panel) compared to their isotype- and subclass-matched controls (right panel). Dotted line: autofluorescence; shaded line: Mac-I (positive control); solid line: ML30 and L7 mAb, respectively (upper panel); rb-a-60 and rb control IgG. respectively (lower panel).
Eur. J. Immunol. 1991. 21: 1089-1092 80
c
60
0
c
E40 Y
m
a
20
0
0
200 I000 IFN-gamma (U/ml)
Figure 4 . Induction of cross-reactive surface epitopes by rIFN-y. Staining with ML30 (W) or Leu-19 (Ed), respectively.
We wondered whether stimulation by rIFN-y would increase the expression of the shared epitope recognized by ML30. On day 9 of cultivation BMMQ were stimulated with rIFN-y for a further 2 days and thereafter analyzed by FCM. Stimulating the cells with lo00 U/ml IFN-y significantly increased surface expression of the cross-reactive epitope recognized by ML30 with a 2- to 8-fold increase in different experiments (Fig. 4). This increase is not due to FcR binding as shown by the control mAb L7. We have shown previously that rIFN-y stimulation of BMMQ renders them susceptible to killing by CD8 Tcells [15].Thus, it appears that IFN-y induces or increases expression of cross-reactive epitopes recognized by hsp 60-specificTcells and B cells.
Surface expression of a cross-reactive hsp epitope
1091
provide strong evidence that host cells can serve as targets of antibodies directed against hsp. The human rhsp60 is recognized by the ML30 mAb (unpublished) and Western blot analysis of HeLa cells reveals staining of a single 60-kDa band by this antibody [5, 221. Preliminary results from immunoprecipitation experiments indicate that the surface molecule expressing the cross-reactive epitope has an apparent molecular mass of 60 kDa and hence resembles the bacterial hsp 60 in this respect (unpublished). These findings would be consistent with surface expression of the mammalian hsp 60 cognate. Evidence for surface expression of other hsp by leukocytes has been presented recently [23,24]. On the other hand, it cannot be excluded that the shared epitope is contained in an unrelated structure as, for example, striking sequence homology between mycobacterial hsp 60 and unrelated proteins in mammalian cells including a mouse T complex protein have been described [25]. To obtain definite identification of the molecule, sequence determination of the immunoprecipitate will be performed in the future. In any case, our findings show that antibodies against bacterial hsp 60 recognize certain host cells and hence suggest a direct role for such antibodies in autoimmunity. We thank J. DeBruyn (Bruxelles) for providing a wide panel of anti-hsp 60 mAb; Z Gillis (Carville,L A ) for the IIC8 and A. Basten (Sydney) for the L7 hybridoma. We are grateful to Dr. J. D. A . Van Emhden (Bilthoven) for providing the recombinant E. coli clone MI546 expressing hsp 60 of M . bovis BCG and to Dr. I! Conradt for performing FACS analyses. riFN-y was a kind gift from Dr. Adolf (Boehringer, Vienna). Received December 17, 1990.
4 Discussion
5 References Because of their high conservation, cognates of the hsp60 family are potential candidates for an autoimmune response [3]. Indeed, evidence has been presented that Tcells with specificty for hsp 60 are involved in experimental models of rheumatoid arthritis and diabetes [16, 171. Furthermore, it has been shown that cytolytic T lymphocytes activated by tryptic fragments of bacterial hsp 60 recognize stressed BMMQ in an MHC class I-restricted fashion [15]. hsp60 is localized in the mitochondrial matrix where it is involved in the folding, unfolding and translocation of mitochondrial proteins [1,2, 18, 191. Class Irestricted T cells recognize peptides derived from endogenous antigens of cytoplasmic [20] as well as mitochondrial origin [21]. Hence, it can be envisaged that peptides derived from hsp60 can be translocated to the surface via class I-dependent processing. On the contrary, antibodies recognize their antigens directly. The intracellular location of hsp should, therefore, render cellular hsp unaccessible to the humoral immune response. Increased antibody levels to hsp, including mycobacterial hsp 60, have been found in several autoimmune diseases [3]. However, in these studies, it has not yet been elucidated whether the immune response is directed against epitopes shared by microorganism and host or against epitopes unique to the former. Our data demonstrate surface exmession of an eDitoDe shared by microbial hsp60 and 6ost cells. Henck, thky
Rothman, J. E., Cell 1989. 59: 591. Hartl, F.-U. and Neupert, W., Science 1990. 247: 930. Kaufmann, S. H. E., Immunol. Today 1990. 11: 129. Jindal, S., Dudani, A. K., Singh, B., Harley, C. B. and Gupta, R. S., Mol. Cell. Biol. 1989. Y: 2279. 5 Dudani, A. K. and Gupta, R. S., Infect. Immun. 9: 2786. 6 Mizzen, L. A., Chang, C., Garrels, J. J. and We1ch.W. J.. J. Biol. Chem. 1989.264: 20 664. 7 Flesch, I. and Ferber, E., Immunobiology 1986. 171: 14. 8 Flesch, I. and Kaufmann, S. H. E., J. Immunol. 1987. 138: 4408. 9 Thole, J. E. R., Dauwerse, H. G., Das, !F K., Groothuis, D. G., Schouls, L. M. and Van Embden, J. D. A., Infect. Immun. 1985. 50: 800. 10 Kaufmann, S. H. E.,Vath, U.,Thole, J. E. R..Van Embden. J. D. A. and Emmrich, F., Eur. J. Immunol. 1987. 17: 351. 11 Springer,T., Galfrk, G., Secher, D. S. and Milstein, C., Eur. J. Immunol.1978. 8: 539. 12 Gillis, T. I? and Buchanan, T. M., Infect. Immun. 1982. 37: 172. 13 Ivanyi, J., Sinha, S., Aston, R., Cussel, D., Deen, M. and Sengupta, U., J. Biol. Chem. 1983. 252: 7257. 14 Garsia, G. J., Mellqvist, L., Booth, R. J., Radford, A. J.. Bntton,W. J., Astbury, L., Trent, R. J. and Basten, A.. Infect. Immun. 1989.57: 204. A., DeBruyn, J..Munk, M. E., 15 Koga,T.,Wand-Wiirttenberger, Schoel, B. and Kaufmann, S. H. E.. Science 1989. 245: 1112. 16 Van Eden,W.,Thole, J. E . R.,Van der Zee, R., Woordzij, A., Van Embden, J. D. A., Heusen, E. J. and Cohen, J. R., Nature 1988. 331: 171. 1 2 3 4
1092
A. Wand-Wiirttenberger, B. Schoel, J. Ivanyi and S. H.
17 Elias, D., Markovits. D., Reshef, T., Van der Zeel, R. and Cohen, J. R., Proc. Natl. Acad. Sci. USA 1990. 87: 1576. 18 Ostermann, J., Norwich, A. L., Neupert,W. and Hartl, E-U., Nature 1989. 341: 125. 19 Cheng, M. Y., Hartl, E-U., Martin, J., Pollock, R. A.. Kalousek, F., Neupert,W., Hallberg, E. M., Hallberg. R. L. and Horwich, A. L., Nature 1989. 337: 620. 20 Townsend, A. and Bodmer. H., Annu. Rev. Immunol. 1989. 7: 601.
E. Kaufmann
Eur. J. Immunol. 1991. 21: 1089-1092
21 Loveland, B., Wang, C.-R., Yonekawa, H., Hermel, E. and Fischer-Lindahl, K., Cell 1990. 60: 971. 22 Evans, D. J., Norton, P.and Ivanyi, J., Acta Parhol. Microbiol. lmmunol. Scand. 1990. 98: 437. 23 Van Buskirk, A., Crump, B. L. and Pierce, S. K., J. Exp. Med. 1989. 170: 179. 24 Jarjour.W., Tsai,V. ,Woods,V., Welch,W., Pierce, W.. Shaw, M.. Mehta, H., Dillmann, W., Zvaifler, N. and Winfield, J., Arthritis Rheum. 1989. 32: 44. 25 Gupta. R. S., Biochem. Intern. 1990. 20: 833.
Letter to the Editor Last year we published a study about type I1 collagenspecific T cells in healthy donors (Lacour et al., Eur. J. Immunof. 1990. 20: 931) Several Tcell lines and clones were shown to specifically respond in a dose-dependent and MHC-restricted manner to human and chicken type I1 collagen as well as to human type I but not to human type IVcollagen. Some months ago,we became aware of a report by M. Anderson and R. Holmdahl (Eur.J. Immunof. 1990. 20: 1061) about murine T cell clones which were supposed to recognize murine type I1 collagen, but actually recognized pepsin as a contamination from the preparation of type I1 collagen. Due to this report, we retested all our T cell lines considered to be collagen-specific using pork pepsin as an antigen. Unfortunately, most T cell lines described in our study recognized pepsin in addition to collagen. Thus, we have to admit that these Tcell lines are not collagen specific but recognize a pepsin contamination. A strong response was observed at extremely low pepsin concentrations of 10 ng/ml, demonstrating that a contamination of collagens by
Eur. J. Immunol. 1991. 21: 1089-1092
1089
Short paper Angela Wand-Wiirttenbergero, Bernd Schoelo, Juraj IvanyiA and Stefan H. E. KaufmannO Department of Immunologyo, University of Ulm, Ulm and MRC Tuberculosis and Infections UnitA, Hammersmith Hospital, London
Surface expression by mononuclear phagocytes of an epitope shared with mycobacterial heat shock protein 60* Bone marrow-derived macrophages were stained with a variety of monoclonal antibodies (mAb) with specificity for the mycobacterial heat shock protein (hsp) 60 or with specific rabbit antiserum raised against mycobacterial hsp 60.The mAb h4L30 as well as the specific antiserum brightly stained bone marrowderived macrophages whereas all the other mAb as well as normal rabbit antiserum gave no or negligible reactivity. Surface expression of the shared epitope is observed by day 3 of in vitro cultivation and markedly increased by interferon-y stimulation on day 9. Although hsp 60 is thought to be restricted to the mitochondria1 compartment, antibody responses to this molecule have been implicated in autoimmunity. Our finding that bone marrow-derived macrophages express a cross-reactive epitope recognized by an mAb with specificity for amino acids 275 to 295 of the mycobacterial hsp60 is consistent with such a role.
1 Introduction Members of the heat shock protein (hsp) 60 family are essential for the folding, unfolding and translocation of intracellular proteins [ l , 21. Although their synthesis is increased under certain stress conditions, they are constitutively produced. hsp 60 are ubiquitous and highly conserved and cognates of bacterial and mammalian origin show about 60% homology [3,4]. Thus, hsp are potential targets for autoimmune responses and could provide a link between microbial infection and autoimmunity [3]. On the other hand, in mammalian cells the localization of hsp 60 is considered to be restricted to the mitochondrial compartment and it is thought that hsp, in general, are not expressed on the cell surface [ 5 , 6 ] . Here we show that an mAb, ML30, specific for amino acids 275 to 295 of mycobacterial hsp 60 as well as a specific rabbit anti-hsp 60 antiserum recognize a shared epitope on the surface of BM-derived MQ, (BMMQ) indicating that antibodies against bacterial hsp cross-react with surface molecules expressed by host cells.
2 Materials and methods 2.1 Cells BMMQ from C57BL/6 mice were used as a homogeneous population of mouse MQ,. Unless otherwise stated, BM
[I 91091
* This work received financial support from SFB 322, UNDPlWorld B a n W H O Special Program for Research and Training in Tropical Diseases, German Leprosy Association and A. Krupp award for young professors (S.H.E.K.).
Correspondence: Stefan H. E. Kaufmann, Department of Immunology, University of Ulm, Albert-Einstein-Allee 11, D-7900 Ulm, FRG Abbreviation: hsp: Heat shock protein 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1991
cells taken from mouse femura were cultivated for 9-1 1 days in hydrophobic plastic bags as described earlier [7, 81. In some experiment cells were stimulated with rIFN-y.
2.2 Antisera and mAb Rabbit anti-hsp 60 antiserum was generated by immunizing a rabbit with 1 mg purified rhsp60 from M. bovis BCG in Ribi adjuvant i.m. hsp 60 for immunization was produced, isolated and purified from E. coli clone M1546 [9] essentially as described earlier [ 101.The hyperimmune polyclonal rabbit IgG fraction (rb-a-60) was prepared by affinity purification on a protein A-Superose (Pharmacia, Freiburg, FRG) column and titrated for saturating concentration on BMMQ cells by FCM. As a negative control we used a normal rabbit IgG preparation which had been absorbed on an hsp 60 affinity column (CNBr-Sepharose coupled with purified rhsp 60 of M. bovis BCG following the instruction of Pharmacia) to remove contaminating naturally occurring anti-hsp 60 antibodies. The following mAb were used for immunostaining: Mac-1 (rat IgG [ 111; as a positive control); Leu-19 (Becton Dickinson, Mountain, View, CA; as subclass-matched negative control to assess background binding of IgGl to FcR); mAb IVD1, IA1, XIXHS, XVIIIGl and IIGl (mouse anti-M. bovis BCGhsp 60 mAb of IgG2b(IA1) or IgGl (all other mAb) subclass (a gift of J. DeBruyn, Bruxelles); IIC8 (mouse IgGZt, anti-M. leprue hsp 60 mAb; gift of T. Gillis, [12]); ML30 (mouse) IgG1; [13]); and L7 (mouse IgGl mAb recognizing a species cross-reactive determinant on hsp 70 of M. feprue ([14]; gift of A. Basten, Sydney).
2.3 FCM BMMQ were harvested on day 11 and washed twice. FcR were blocked by irrelevant IgG for 15 min in RF'MI 1640 (Gibco, Grand Island, NY) supplemented with 1% mouse serum (when staining with rabbit IgG) or 1% rabbit serum 0014-2980/91/0404-1089$3S O + .25/0
Eur. J. Immunol. 1991. 21: 1089-1092
A. Wand-Wiirttenberger. B. Schoel, J. Ivanyi and S. H. E. Kaufmann
100
(when staining with mouse mAb). All steps were performed in 0.1% NaN3 on ice to prevent internalization of antibodies. Antibodies were diluted in the appropriate blocking buffer and incubated for 30min. For washing, ice-cold PBS containing 0.1% azide was used. Second antibodies were FITC-labeled affinity-purified F(ab’)z fragments of goat anti-mouse IgG and goat anti-rabbit IgG (Medac, Hamburg, FRG). Labeled cells were fixed in 2% paraformaldehydePBS and protected from light until FCM analysis.
,601 I80
2o ”
Murine BMMQ, were stained with various mAb with specificity for mycobacterial hsp 60 or hsp 70 and analyzed thereafter by FCM. Only the ML30 mAb which recognizes amino acids 275 to 295 of the mycobacterial hsp 60 showed positive staining of BMMQ, (Fig. 1). A specific rabbit Ig preparation raised against mycobacterial hsp60 also stained these cells. Fig. 2 shows a representative compari-
14)
120 IW
ar-
0
XI
~7
nm
.
0
3 Results
0
1
IWI
IAI
XIM
XVIIEI
mi
1161
iim
ldlw
MI
iuio
*ulo
Figure I . FCM analysis of mouse BMMQ stained with a panel of antibodies against mycobacterial hsp 60 and hsp70. Mac-1 and Leu-I9 served as positive or negative control, respectively. “Auto” shows the autofluorescent background of unstained cells.
TL 1
FL 1
. 1
. 2
. 3
. 4
.
.
.
5 6 7 Culture (days)
.
.
8
9
.
.
1 0 1 1
Figure 3. Analysis of cross-reactive epitopes during the 11 days of in vitro differentiation. Open circles: Mac- 1;closed circles: control IgG; open squares: ML30; closed squares: rb-a-hO.
son between the cross-reactive mAb ML30 and mAb L7, and between the rb-a-60 and control rb-IgG , respectively. Both the ML30 mAb and the rb-a-60 identify a distinct cell population whereas the normal rb-IgG and the L7 mAb produce a small shift in peak intensity only. This weak background staining by normal rb-IgG could be reduced by passage through an hsp 60-Sepharose column indicating that normal rb-IgG already contained antibodies against cross-reactive epitopes. In an attempt to follow surface expression in the course of BMMQ, development, the cells were stained daily during their in vitro cultivation for 11 days (Fig. 3). Between day 3 and day4, a steep increase in the expression of the cross-reactive epitope recognized by ML30 and rb-a-60 was observed which remained relatively stable until day 9. From day 10 to day 11 a further twofold increase in surface expression was detected with the ML30 mAb, whereas staining with the rb-a-60 was slightly reduced.Thus, the two antibody preparations may recognize different epitopes.
FL 1
FL I
Figure 2. Expression of an hsp 60 cross-reactive epitope. Positive staining of BMMQ, by anti-hsp 60 antibodies (left panel) compared to their isotype- and subclass-matched controls (right panel). Dotted line: autofluorescence; shaded line: Mac-I (positive control); solid line: ML30 and L7 mAb, respectively (upper panel); rb-a-60 and rb control IgG. respectively (lower panel).
Eur. J. Immunol. 1991. 21: 1089-1092 80
c
60
0
c
E40 Y
m
a
20
0
0
200 I000 IFN-gamma (U/ml)
Figure 4 . Induction of cross-reactive surface epitopes by rIFN-y. Staining with ML30 (W) or Leu-19 (Ed), respectively.
We wondered whether stimulation by rIFN-y would increase the expression of the shared epitope recognized by ML30. On day 9 of cultivation BMMQ were stimulated with rIFN-y for a further 2 days and thereafter analyzed by FCM. Stimulating the cells with lo00 U/ml IFN-y significantly increased surface expression of the cross-reactive epitope recognized by ML30 with a 2- to 8-fold increase in different experiments (Fig. 4). This increase is not due to FcR binding as shown by the control mAb L7. We have shown previously that rIFN-y stimulation of BMMQ renders them susceptible to killing by CD8 Tcells [15].Thus, it appears that IFN-y induces or increases expression of cross-reactive epitopes recognized by hsp 60-specificTcells and B cells.
Surface expression of a cross-reactive hsp epitope
1091
provide strong evidence that host cells can serve as targets of antibodies directed against hsp. The human rhsp60 is recognized by the ML30 mAb (unpublished) and Western blot analysis of HeLa cells reveals staining of a single 60-kDa band by this antibody [5, 221. Preliminary results from immunoprecipitation experiments indicate that the surface molecule expressing the cross-reactive epitope has an apparent molecular mass of 60 kDa and hence resembles the bacterial hsp 60 in this respect (unpublished). These findings would be consistent with surface expression of the mammalian hsp 60 cognate. Evidence for surface expression of other hsp by leukocytes has been presented recently [23,24]. On the other hand, it cannot be excluded that the shared epitope is contained in an unrelated structure as, for example, striking sequence homology between mycobacterial hsp 60 and unrelated proteins in mammalian cells including a mouse T complex protein have been described [25]. To obtain definite identification of the molecule, sequence determination of the immunoprecipitate will be performed in the future. In any case, our findings show that antibodies against bacterial hsp 60 recognize certain host cells and hence suggest a direct role for such antibodies in autoimmunity. We thank J. DeBruyn (Bruxelles) for providing a wide panel of anti-hsp 60 mAb; Z Gillis (Carville,L A ) for the IIC8 and A. Basten (Sydney) for the L7 hybridoma. We are grateful to Dr. J. D. A . Van Emhden (Bilthoven) for providing the recombinant E. coli clone MI546 expressing hsp 60 of M . bovis BCG and to Dr. I! Conradt for performing FACS analyses. riFN-y was a kind gift from Dr. Adolf (Boehringer, Vienna). Received December 17, 1990.
4 Discussion
5 References Because of their high conservation, cognates of the hsp60 family are potential candidates for an autoimmune response [3]. Indeed, evidence has been presented that Tcells with specificty for hsp 60 are involved in experimental models of rheumatoid arthritis and diabetes [16, 171. Furthermore, it has been shown that cytolytic T lymphocytes activated by tryptic fragments of bacterial hsp 60 recognize stressed BMMQ in an MHC class I-restricted fashion [15]. hsp60 is localized in the mitochondrial matrix where it is involved in the folding, unfolding and translocation of mitochondrial proteins [1,2, 18, 191. Class Irestricted T cells recognize peptides derived from endogenous antigens of cytoplasmic [20] as well as mitochondrial origin [21]. Hence, it can be envisaged that peptides derived from hsp60 can be translocated to the surface via class I-dependent processing. On the contrary, antibodies recognize their antigens directly. The intracellular location of hsp should, therefore, render cellular hsp unaccessible to the humoral immune response. Increased antibody levels to hsp, including mycobacterial hsp 60, have been found in several autoimmune diseases [3]. However, in these studies, it has not yet been elucidated whether the immune response is directed against epitopes shared by microorganism and host or against epitopes unique to the former. Our data demonstrate surface exmession of an eDitoDe shared by microbial hsp60 and 6ost cells. Henck, thky
Rothman, J. E., Cell 1989. 59: 591. Hartl, F.-U. and Neupert, W., Science 1990. 247: 930. Kaufmann, S. H. E., Immunol. Today 1990. 11: 129. Jindal, S., Dudani, A. K., Singh, B., Harley, C. B. and Gupta, R. S., Mol. Cell. Biol. 1989. Y: 2279. 5 Dudani, A. K. and Gupta, R. S., Infect. Immun. 9: 2786. 6 Mizzen, L. A., Chang, C., Garrels, J. J. and We1ch.W. J.. J. Biol. Chem. 1989.264: 20 664. 7 Flesch, I. and Ferber, E., Immunobiology 1986. 171: 14. 8 Flesch, I. and Kaufmann, S. H. E., J. Immunol. 1987. 138: 4408. 9 Thole, J. E. R., Dauwerse, H. G., Das, !F K., Groothuis, D. G., Schouls, L. M. and Van Embden, J. D. A., Infect. Immun. 1985. 50: 800. 10 Kaufmann, S. H. E.,Vath, U.,Thole, J. E. R..Van Embden. J. D. A. and Emmrich, F., Eur. J. Immunol. 1987. 17: 351. 11 Springer,T., Galfrk, G., Secher, D. S. and Milstein, C., Eur. J. Immunol.1978. 8: 539. 12 Gillis, T. I? and Buchanan, T. M., Infect. Immun. 1982. 37: 172. 13 Ivanyi, J., Sinha, S., Aston, R., Cussel, D., Deen, M. and Sengupta, U., J. Biol. Chem. 1983. 252: 7257. 14 Garsia, G. J., Mellqvist, L., Booth, R. J., Radford, A. J.. Bntton,W. J., Astbury, L., Trent, R. J. and Basten, A.. Infect. Immun. 1989.57: 204. A., DeBruyn, J..Munk, M. E., 15 Koga,T.,Wand-Wiirttenberger, Schoel, B. and Kaufmann, S. H. E.. Science 1989. 245: 1112. 16 Van Eden,W.,Thole, J. E . R.,Van der Zee, R., Woordzij, A., Van Embden, J. D. A., Heusen, E. J. and Cohen, J. R., Nature 1988. 331: 171. 1 2 3 4
1092
A. Wand-Wiirttenberger, B. Schoel, J. Ivanyi and S. H.
17 Elias, D., Markovits. D., Reshef, T., Van der Zeel, R. and Cohen, J. R., Proc. Natl. Acad. Sci. USA 1990. 87: 1576. 18 Ostermann, J., Norwich, A. L., Neupert,W. and Hartl, E-U., Nature 1989. 341: 125. 19 Cheng, M. Y., Hartl, E-U., Martin, J., Pollock, R. A.. Kalousek, F., Neupert,W., Hallberg, E. M., Hallberg. R. L. and Horwich, A. L., Nature 1989. 337: 620. 20 Townsend, A. and Bodmer. H., Annu. Rev. Immunol. 1989. 7: 601.
E. Kaufmann
Eur. J. Immunol. 1991. 21: 1089-1092
21 Loveland, B., Wang, C.-R., Yonekawa, H., Hermel, E. and Fischer-Lindahl, K., Cell 1990. 60: 971. 22 Evans, D. J., Norton, P.and Ivanyi, J., Acta Parhol. Microbiol. lmmunol. Scand. 1990. 98: 437. 23 Van Buskirk, A., Crump, B. L. and Pierce, S. K., J. Exp. Med. 1989. 170: 179. 24 Jarjour.W., Tsai,V. ,Woods,V., Welch,W., Pierce, W.. Shaw, M.. Mehta, H., Dillmann, W., Zvaifler, N. and Winfield, J., Arthritis Rheum. 1989. 32: 44. 25 Gupta. R. S., Biochem. Intern. 1990. 20: 833.
Letter to the Editor Last year we published a study about type I1 collagenspecific T cells in healthy donors (Lacour et al., Eur. J. Immunof. 1990. 20: 931) Several Tcell lines and clones were shown to specifically respond in a dose-dependent and MHC-restricted manner to human and chicken type I1 collagen as well as to human type I but not to human type IVcollagen. Some months ago,we became aware of a report by M. Anderson and R. Holmdahl (Eur.J. Immunof. 1990. 20: 1061) about murine T cell clones which were supposed to recognize murine type I1 collagen, but actually recognized pepsin as a contamination from the preparation of type I1 collagen. Due to this report, we retested all our T cell lines considered to be collagen-specific using pork pepsin as an antigen. Unfortunately, most T cell lines described in our study recognized pepsin in addition to collagen. Thus, we have to admit that these Tcell lines are not collagen specific but recognize a pepsin contamination. A strong response was observed at extremely low pepsin concentrations of 10 ng/ml, demonstrating that a contamination of collagens by
