- Email: [email protected]
γδ T-cell–human glial cell interactions. I. In vitro induction of γδ T-cell expansion by human glial cells
Journal of Neuroimmunology 74 Ž1997. 135–142
gd T-cell–human glial cell interactions. I. In vitro induction of gd T-cell expansion by human glial cells Mark S. Freedman a
a,)
, Sameer D’Souza b, Jack P. Antel
b
UniÕersity of Ottawa, Department of Medicine, DiÕision of Neurology, Ottawa General Hospital, 501 Smyth Road, Ottawa, Ont., Canada K1H 8L6 McGill UniÕersity, Neuroimmunology Unit, Department of Neurology and Neurosurgery, Montreal Neurological Institute, 3801 UniÕersity Street, Montreal, Que., Canada H3A 2B4
b
Received 26 August 1996; revised 1 November 1996; accepted 4 November 1996
Abstract
gd T-cells are found in increased proportion in multiple sclerosis ŽMS. white matter plaque infiltrates compared with peripheral blood or spleen, raising the possibility that they are either specifically attracted to lesion sites or, once present, are stimulated to expand. We have previously shown that human oligodendrocytes ŽOGC. preferentially express heat shock proteins Ž hsp ., molecules to which gd T-cells have been known to react and that in vitro expanded gd T-cells can lyse OGC. We therefore investigated whether human glial cells, that differentially express hsp, could stimulate gd T-cell expansion from peripheral blood. We compared the glial cell-induced expansion to cell lines which also differentially express hsp and have been shown to selectively stimulate gd T-cell expansion Že.g. RPMI 8226, Daudi.. We found that both OGC and human fetal astrocytes ŽhFA. expressed hsp and stimulated the preferential expansion of gd T-cells to about the same extent as the hsp expressing cell lines RPMI 8226 or Daudi, in the presence of exogenous interleukin-2 ŽIL-2. but without any T-cell mitogen. Furthermore, the type of gd T-cells expanded were of the Vd 2 subtype known to be particularly reactive to hsp. Microglia, U937 cell lines or purified myelin membranes, which express little or no hsp, did not support gd T-cell growth. These results therefore suggest that OGC may contribute to the local expansion of gd T-cells within MS plaques. Potential harmful effects of gd T-cells on OGC may thereby contribute to the immunopathogenesis of MS demyelination. Keywords: gd T-cell; Glia; Heat shock protein; T-cell stimulation; Multiple sclerosis
1. Introduction T-cells are concentrated within the multifocal perivenular areas of inflammation in the central nervous system white matter, that characterize the human autoimmune demyelinating disease multiple sclerosis ŽMS.. It is believed that these T-cells either directly Žthrough cell contact. or indirectly Žthrough cytokine release. contribute to the accompanying demyelination of MS, though the exact mechanisms are not known. The predominant T-cell type in the lesions expresses an ab T-cell receptor ŽTCR., some of which may be specifically reactive with myelin antigens, while others are not ŽSteinman et al., 1995.. Another T-cell type, expressing a gd T-cell receptor, has been identified in MS lesions and appears to be present in considerably increased proportion compared with periph-
) Corresponding author. Tel.: q1-613-7378917; fax: q1-613-7378857; e-mail: [email protected]
eral blood or spleen ŽWucherpfennig et al., 1992.. This suggests that either gd T-cells are specifically recruited from the periphery and persist in lesion sites or that gd T-cells infiltrating the CNS, as part of a generalized inflammatory response, are stimulated to expand within the CNS. Since activated and expanded gd T-cells can in vitro be shown to lyse human oligodendrocytes ŽOGC. ŽFreedman et al., 1991., it is possible they play an important role in the pathogenesis of the MS lesion. gd T-cells are known to react in a predominantly MHC non-specific matter with certain heat-shock proteins Ž hsp ., particularly the 65 kDa proteins of Mycobacteria or groEL-related mammalian homologues such as hsp60 ŽFisch et al., 1990; Kaufmann, 1992; Kaur et al., 1993; Haas et al., 1993.. Interaction with hsp60 expressed on the surface of Daudi cells has been shown to be key to the subsequent Daudi induced gd T-cell expansion ŽKaur et al., 1993.. OGC express little or no MHC ŽGrenier et al., 1989. but do express hsp60, possibly more so than other human glial cells ŽFreedman et al., 1992; Satoh et al.,
0165-5728r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 5 7 2 8 Ž 9 6 . 0 0 2 1 7 - 2
136
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
1992.. In this paper we investigated whether OGC, compared to other human glial cells, will stimulate the outgrowth and expansion of gd T-cells from peripheral blood. In a related study ŽFreedman et al., 1997., we address the specificity of the gd T-cell cytolytic response to hsp60.
2. Materials and methods 2.1. Cell lines and glial cell cultures RPMI 8226, Daudi and U937 cells were obtained from ATCC ŽRockville., grown and maintained in RPMI with 10% FCS ŽGibco BRL, Burlington. and 10% gentamicin Žexcept RPMI 8226.. Cell lines were split and fed with fresh media 2–3 times per week. Dissociated glial cell cultures enriched for OGC ŽFreedman et al., 1991., microglia ŽWilliams et al., 1992. or human fetal astrocytes ŽhFA. ŽD’Souza et al., 1995. were prepared as previously described in detail. Briefly, human brain tissue was obtained from patients undergoing surgery for epilepsy and cells isolated by trypsin-dissociation and Percoll gradient centrifugation. Microglia and OGC were enriched from mixed cultures based on their tendency to adhere Žmicroglia. or not ŽOGC. to poly-D-lysine coated coverslips ŽWilliams et al., 1992.. OGC cultures used for
induction experiments showed purities of ) 90% by galactocerebroside staining, whereas those used for histochemistry were ) 90% by CNPase immunoreactivity ŽFreedman et al., 1992.; microglia were ) 95% CD11c positive ŽWilliams et al., 1992.. Human fetal astrocytes ŽhFA. Ž) 95% GFAPq . were derived from 12–14 week abortuses following MRC ŽCanada. approval guidelines, as previously described ŽD’Souza et al., 1995.. All cell lines and cultures were maintained in a humidified 378C 5% CO 2 incubator. 2.2. T-cell cultures Freshly isolated peripheral blood mononuclear cells ŽPBMC. were obtained from healthy volunteers by FicollHypaque gradient centrifugation. 1 = 10 6 PBMC were placed in 24 well plates ŽNunclon, Gibco BRL, Burlington. with either 3.5 = 10 4 irradiated Ž5,000 Rads. RPMI 8226, U937 cells, OGC, microglia or hFA in 2 ml of serum-free media ŽAim-V, Gibco BRL, Burlington. and 50 Urml recombinant human IL-2 Ža generous gift from IMMUNEX, Seattle.. Media was changed with fresh IL-2 on day 6 of a 10 day culture. In some cases, purified human myelin Ža gift of J. McLaurin, Montreal PQ. was coated to wells and in other experiments, 2 = 10 5 Daudi cells were used as stimulators in 10% FCS according to a
Fig. 1. Constitutive surface ŽA, C. or intracellular ŽB, D. staining of either OGC ŽA, B. or hFA ŽC, D. with monoclonal Ab to hsp60, followed by biotinylated goat anti-mouse Ab and FITC-conjugated streptavidin; inset is the isotype control. Magnification: of ŽA. and ŽB. is =250, of ŽC. and ŽD. is =350.
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
137
previously published protocol ŽFisch et al., 1990.. For human glia, cells were pre-plated in experimental wells in order to maximize adherence and viability. After 10 days, cells were pipetted from wells, dead cells removed by Ficoll-Hypaque separation and viable Žtrypan blue excluding. cells counted and stained for flow cytometric analysis. 2.3. Immunohistochemisty For surface staining of hsp60, adult OGC or hFA on coverslips were incubated for 1 h with mouse anti-human hsp60 antibody ŽAb. ŽAffinity Bioreagents, Neshanic Station. or mouse anti-human IgG2a isotype ŽR & D Systems, Minneapolis. and washed three times with phosphate buffered saline ŽPBS.. Cells were then counterstained by incubation for 1 h with biotinylated goat anti-mouse Ab ŽVector Laboratories, Burlingame. at room temperature ŽRT., followed by another 1 h incubation with fluorescein-conjugated stretavidin ŽBoehringer Mannheim Biochemica, Mannheim., before fixing them in 4% paraformaldehyde ŽPFA. for 20 min at 48C and mounting coverslips with Gelvatol. For intracellular staining, coverslips were first fixed in ice cold Žy208C. 95% ethanolr5% acetic acid for 20 min, then rehydrated for 30 min with PBS and stained as above. Unless otherwise stated, all immunostaining was performed at RT, all Ab reagents were used at a dilution of 1:100 and all coverslips were washed extensively with PBS between steps. 2.4. Flow cytometry Staining of cells for flow cytometry has been previously described ŽFreedman et al., 1991. using either single or double staining techniques. Briefly, cell lines in suspension were incubated with a 1:40 dilution of polyclonal rabbit anti-hsp58 Ž hsp60, courtesy of W. Welch, San Francisco. and counter-stained with FITC-conjugated goat anti-rabbit IgG Ž1:100. ŽGibco BRL, Burlington.. ŽStaining of cell lines with the monoclonal Ab to human hsp60 used above did not demonstrate the same degree of surface expression as that seen with the polyclonal serum, similar to the reports of others ŽKaur et al., 1993... T-cell cultures were double stained with monoclonal antibodies ŽmAb.: antiTCR-d 1-FITC Žrecognizing all gd T-cells via the C d epitope. ŽT-cell Sciences, Intermedico, Markham. or antiTCR-arb-1-FITC, Leu-4-PE Žanti-CD3. ŽBecton-Dickinson, Mountain View.. All cell lines were maintained in 1% paraformaldehyde in PBS prior to analysis on a FACScan ŽBecton-Dickinson., gating out only debris. %CD3q gd T-cells was calculated based on the TCR-d 1-FITCrLeu-4PE quadrant and verified by conversely staining with TCR-arb-1-FITCrLeu-4-PE. In some cases, the gd Tcells grown were analyzed for subtype Ži.e. Vd 1r2. using either TCS-1 Žrecognizing Vd 1. or Tg iA Žrecognizing Vg 9 which is mostly associated with Vd 2 ŽHaas et al., 1993. mAb ŽT-cell Sciences. conjugated to FITC and
Fig. 2. RPMI 8226 or U937 cells were immunostained for hsp60 expression using rabbit polyclonal anti-hsp58, anti-GFAP antibodies or the second antibody Žgoat anti-rabbit-FITC. alone ŽFITC control. and analyzed by flow cytometry. Expression of hsp60 by RPMI 8226 cells supercedes that of U937 cells Ž ) 20=., which partially overlaps the non-specific staining by the rabbit polyclonal anti-GFAP control.
138
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
double stained with anti-TCR grd-1-PE ŽBecton Dickinson..
3. Results 3.1. Expression of hsp60 by human glial cells.
2.5. Statistical analysis Where indicated, statistical analysis was performed using a Mann–Whitney non-parametric unpaired analysis, calculating a one-tailed p value, with a software package called Instat from GraphPad e .
We have previously demonstrated that adult human glial cells differentially express hsp60 using polyclonal rabbit antiserum ŽFreedman et al., 1992.. We repeated these studies now using a monoclonal antibody to human mitochondrial hsp60 that cross reacts with the hsp homo-
Fig. 3. Fresh PBMC were cultured in serum-free media with irradiated stimulator cells ŽRPMI 8226 or U937. cells Ž30:1 PBMC:stimulator cell ratio. for 10 days with 50 Urml rhIL-2. Daudi cell cultures utilized a 5:1 PBMC:Daudi cell ratio in 10% FCS. In the presence of either RPMI 8226 or Daudi cells, there was a marked expansion of gd T-cell proportions Žtop panel. and cell numbers Žlower panel. from initial PBMC cultures containing proportions of 2–12%, whereas U937 cells did not significantly increase either proportions or cell numbers. Media alone contained rhIL-2 but no stimulator cells. Depicted are the results of at least three separate experiments from different individuals" S.E.M.
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
logue of Mycobacteria ŽSharif et al., 1992. and in addition we characterized the expression of hsp60 on hFA. As we previously observed, OGC constitutively express hsp60 on their surface, as well as intracellularly ŽFig. 1A, B.. In contrast to adult human astrocytes that only demonstrated a minor amount of punctate surface staining, if at all ŽFreedman et al., 1992., most hFA expressed both surface and intracellular hsp60 ŽFig. 1C, D.. Qualitatively, hsp60 surface staining of hFA appeared to be less than that of OGC.
139
3.2. Expression of hsp60 by cell lines Daudi cells have been shown elsewhere to express hsp60 ŽFisch et al., 1990; Kaur et al., 1993., but here we demonstrate that another B cell line RPMI 8226, that we and others have used extensively to expand peripheral blood gd T-cells ŽFreedman et al., 1991; Selin et al., 1992., also stains strongly for hsp60 Ža full log difference in mean channel fluorescence., compared to the lower levels seen on U937 cells ŽFig. 2..
Fig. 4. Fresh PBMC were co-cultured with irradiated stimulator cells ŽRPMI 8226 or glia; 30:1 PBMC:stimulator cell ratio. in serum-free media for 10 days. In the presence of either OGC or hFA, gd T-cells expanded, with statistically significant Ž ) . increases in both their proportions Žtop panel. and cell numbers Žlower panel. compared to culture in media alone. Neither purified myelin membranes nor microglia stimulated significant gd T-cell outgrowth. Depicted are the results of at least three separate experiments from different individuals" S.E.M.
140
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
3.3. Induction of gd T-cell expansion by cell lines It has been shown that hsp60 expressing Daudi or RPMI 8226 cells induced the expansion of gd T-cells from peripheral blood without the need for T-cell mitogen or TCR cross-linking with only the addition of exogenous IL-2 ŽFisch et al., 1990; Freedman et al., 1991; Selin et al., 1992.. Both polyclonal and monoclonal Ab to hsp60 would abrogate this stimulation ŽFisch et al., 1990; Selin et al., 1992; Kaur et al., 1993.. We and others have observed that the stimulation induced by the RPMI 8226 cells is maximal when cultures are maintained in serum free media ŽAim V. ŽFreedman et al., 1991; Selin et al., 1992.. We therefore conducted all experiments in serum free media using the PBMC:RPMI 8226 30:1 ratio previously determined to be optimal ŽFreedman et al., 1991; Selin et al., 1992. as a standard with which to compare other cells. Under similar serum-free conditions, U937 cells did not stimulate gd T-cell growth ŽFig. 3., even though low levels of hsp60 surface staining was noted ŽFig. 2.. U937 cells did not support growth whether or not serum was contained in the media Žresults not shown.. The 30:1 ratio used in our study was too dilute for Daudi cells to support the gd T-cell growth, even when using 10% FCS Žresults not shown., however at the reported 5:1 ratio ŽFisch et al., 1990., gd T-cells readily expanded ŽFig. 3.. These data demonstrate that RPMI 8226, like Daudi cells, both express high levels of hsp60 and strongly stimulate gd T-cell expansion from PBMC, whereas the lower hsp60 expressing U937 cells do not. 3.4. Induction of gd T-cell expansion by human glial cells Using the same 30:1 ratio that was optimal for gd T-cell expansion with RPMI 8226 cells, we investigated whether or not OGC, microglia or hFA had any similar stimulator capability. Some growth was observed using microglia, but this was not significantly different compared to media alone Ž p s 0.197 for % and 0.066 for cell number, Mann–Whitney.. The OGC induced gd T-cell growth to about the same extent as RPMI 8226 cells, both representing a statistically significant increases over media alone Ž p - 0.004 Ž%., p - 0.006 Žcell number. for the OGC; p - 0.0005 Ž%., p - 0.0009 Žcell number. for the RPMI 8226 cells. ŽFig. 4.. The proportion of gd T-cells did not increase as much as the absolute gd T-cell number in the OGC induced cultures compared with the RPMI 8226 cell line, suggesting that both ab and gd T-cells expanded in the presence of the OGC, whereas the RPMI 8226 may actually be suppressing ab T-cell growth ŽJ. Wilkins, personal communication.. hFA, unlike adult astrocytes, do express hsp60 and also induced a modest but significant gd T-cell expansion compared to media Ž p s 0.0465 Ž%., p s 0.0011 Žcell number.. To test whether it was the OGC itself or a myelin antigen that stimulated gd T-cell growth, we coated some
Table 1 gd T-cells derived from at least six different cultures stimulated via different cell types were stained for gd subtype Žsee Section 2. and analyzed by flow cytometry. Stimulator cell
RPMI 8226 Daudi Oligodendrocytes Fetal astrocytes
Proportion a of gd T-cells Ž"S.E.M.. Vd 1
Vd 2
4"1.7 5.9"1.7 13.93"2.19 ND
98.1".37 97.1"1.15 89"1.9 ND
a
% of CD3q cells. NDs not done.
wells with 0.5 ml or a 1:10 mixture of purified human myelin but did not see any significant growth compared to media ŽFig. 4.. We were not able to evaluate experiments in which the same anti-hsp58 rabbit antibody previously described to block Daudi cell induction was used to block the expansion of gd T-cells ŽFisch et al., 1990., because the Ab was toxic to both stimulator cell and PBMC in serum-free conditions. We did not have access to the same hybridoma supernatant for the hsp60 previously used to block Daudiinduced gd T-cell growth ŽKaur et al., 1993.. Nevertheless, these results do indicate that the ability to induce the expansion of gd T-cells from PBMC correlates with hsp60 expression. 3.5. Phenotype of gd T-cells stimulated hsp60, whether a soluble antigen or expressed by a cell or bacterium appears to stimulate mainly gd T-cells bearing the Vd 2 as opposed to the Vd 1 TCR ŽHaas et al., 1993.. In keeping with this observation, the gd T-cells stimulated by RPMI 8226, Daudi or OGC were all predominantly of the Vd 2 subtype ŽTable 1..
4. Discussion The mechanism of myelin damage in MS is unknown, however it is felt that T-cells either directly or indirectly mediate this phenomenon by their presence in MS plaques. gd T-cells are concentrated in MS plaque infiltrates, whereas they are few in number in peripheral blood or in other organs such as the spleen ŽWucherpfennig et al., 1992.. We now demonstrate that adult OGC are also capable of stimulating the growth and expansion of gd T-cells, which were predominantly of the Vd 2 phenotype, consistent with the reported findings in MS plaque tissue ŽWucherpfennig et al., 1992; Hvas et al., 1993.. In some studies ŽSelmaj et al., 1991; Wucherpfennig et al., 1992., gd T-cell concentrates co-localized with OGC expressing hsp60. Indeed, co-localization of gd T-cell infiltrates and areas of increased hsp60 immunoreactivity have even been
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
noted in chronic lesions of EAE ŽGao et al., 1995.. It would appear that it is the adult OGC itself and not just the myelin that stimulates the gd T-cell expansion, as we observed no growth in the presence only of purified myelin ŽFig. 4.. Cell lines and human glia that express hsp60 appear to share in the ability to stimulate the selective expansion of gd T-cells from the small numbers seen in peripheral blood to significant proportions. Others have shown this to be dependent on hsp60 recognition by the blocking of expansion using Ab to hsp60 on Daudi cells ŽKaur et al., 1993.. Kaur et al. Ž1994. have also demonstrated that other Burkitt’s lyphoma cells Žsuch as Raji., not expressing hsp60, are neither capable of inducing gd T-cell growth nor acting as good targets for gd T-cell-mediated lysis. hFA, unlike their adult counterparts, expressed detectable levels of hsp60. Although the levels of surface hsp60 were qualitatively lower than OGC by histochemistry, the hFA were capable of inducing significant gd T-cell expansion ŽFig. 4.. Others have observed similar lower level expression of hsp60 on murine astrocytes compared to OGC ŽSatoh et al., 1992.. We could not determine whether quantitatively the amount of hsp60 expression correlated with gd T-cell expansion, since it was not possible to perform flow cytometry on the glial cells. We have previously shown that some, but probably not all brain derived microglia express hsp ŽFreedman et al., 1992. and this may have accounted for the perceived induction of gd T-cell expansion we observed, though it was not significantly different from media; likely the result of wide variability among preparations ŽFig. 4.. Taken together, our data would suggest that hsp60 expression on glial cells correlates with the ability to induce gd T-cell expansion from peripheral blood and may be one brain antigen that contributes to the local accumulation of gd T-cells in MS lesions. Others have demonstrated that gd T-cells derived from either the blood or spinal fluid of MS patients have skewing of their TCR repertoire indicating a response to a common brain antigen, possibly hsp ŽBattistini et al., 1995; Stinissen et al., 1995; Nick et al., 1995.. Knowledge of the natural role and function of gd T-cells in the immune response is still in its infancy. It appears that these cells are not only MHC unrestricted, but they also do not require antigen processing and presentation, affording them the ability to directly recognize antigens on pathogens or on damaged tissues that may be expressing hsp60 as part of a stress response ŽHaas et al., 1993.. Infectious processes that turn on gd T-cells as part of a ‘first line defense’ ŽYoung and Elliott, 1989. are well known to precede attacks of MS ŽSibley et al., 1985.. Once activated in the periphery, gd T-cells could cross the blood–brain-barrier where they might encounter cells expressing hsp60. Although the fever associated with systemic illness could act as a direct inducer of hsp in the brain, we have shown that IL-1 Žthe endogenous brain pyrogen. induces OGC hsp expression in vitro ŽD’Souza
141
et al., 1994.. In mice, IL-1 and IL-12 synergistically activate and probably lead to the accumulation of gd T-cells at sites of infection ŽSkeen and Ziegler, 1995.. Recently it has been shown that acute MS plaque tissue also harbours significant amounts of IL-12 ŽWindhagen et al., 1995.. Burns et al. Ž1995. have demonstrated that gd T-cells expand in response to activated myelin reactive T-cell clones. There are therefore many possible reasons for the localized expansion of gd T-cells in MS plaques. Our current work suggests that hsp60 expressing OGC could serve to either attract or retain gd T-cells in the brain, by stimulating their expansion. These cells may be potentially harmful, given their ability in vitro to lyse OGC ŽFreedman et al., 1991.. Whether or not this lytic capability also depends on the expression hsp60 is the subject of a related paper ŽFreedman et al., 1997..
Acknowledgements We are indebted to Ms. Marta Bachetti for her excellent technical assistance. This work was supported in part by a research grant from the Multiple Sclerosis Society of Canada to M.S.F.
References Battistini, L., Salvetti, M., Ristori, G., Falcone, M., Raine, C.S. and Brosnan, C.F. Ž1995. gd T-cell receptor analysis supports a role for HSP 70 selection of lymphocytes in multiple sclerosis lesions. Mol. Med. 1, 554–562. Burns, J., Bartholomew, B. and Littlefield, K. Ž1995. gd T-cells participate in the immune response against activated, myelin basic proteinspecific, human T-cells. J. Neuroimmunol. 58, 177–182. D’Souza, S.D., Antel, J.P. and Freedman, M.S. Ž1994. Cytokine induction of heat shock protein expression in human oligodendrocytes: An interleukin-1-mediated mechanism. J. Neuroimmunol. 50, 17–24. D’Souza, S., Alinauskas, K., McCrea, E., Goodyer, C. and Antel, J.P. Ž1995. Differential susceptibility of human CNS-derived cell populations to TNF-dependent and independent immune-mediated injury. J. Neurosci. 15, 7293–7300. Fisch, P., Malkovsky, M., Kovats, S., Sturm, E., Braakman, E., Klein, B.S., Voss, S.D., Morrissey, L.W., DeMars, R., Welch, W.J., Bolhuis, R.L.H. and Sondel, P.M. Ž1990. Recognition by human Vg 9rVd 2 T-cells of a GroEL homolog on Daudi Burkitts lymphoma cells. Science 250, 1269–1273. Freedman, M.S., Ruijs, T.C.G., Selin, L.K. and Antel, J. Ž1991. Peripheral g-d T-cells lyse fresh human brain-derived oligodendrocytes. Ann. Neurol. 30, 794–800. Freedman, M.S., Buu, N.N., Ruijs, T.C.G., Williams, K. and Antel, J.P. Ž1992. Differential expression of heat shock proteins by human glial cells. J. Neuroimmunol. 41, 231–238. Freedman, M.S., Bitar, R. and Antel, J.P. Ž1997. gd T-cell-human glial cell interactions. II. Relationship between heat shock protein expression and susceptibility to cytolysis. J. Neuroimmunol. 74, 143–148. Gao, Y.L., Brosnan, C.F. and Raine, C.S. Ž1995. Experimental allergic encephalomyelitis. Qualitative and semiquantitative differences in heat shock protein 60 expression in the central nervous system. J. Immunol. 154, 3548–3556. Grenier, Y., Ruijs, T.C.G., Robitaille, Y., Olivier, A. and Antel, J.P.
142
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
Ž1989. Immunohistochemical studies of adult human glial cells. J. Neuroimmunol. 21, 103–115. Haas, W., Pereira, P. and Tonegawa, S. Ž1993. Gammardelta T-cells. Ann. Rev. Immunol. 11, 637–685. Hvas, J., Oksenberg, J.R., Fernando, R., Steinman, L. and Bernard, C.A. Ž1993. gd T-cell receptor repertoire in brain lesions of patients with multiple sclerosis. J. Neuroimmunol. 46, 225–234. Kaufmann, S.H.E. Ž1992. The cellular response to heat shock proteins. Experientia 48, 640–643. Kaur, I., Voss, S.D., Gupta, R.S., Schell, K., Fisch, P. and Sondel, P.M. Ž1993. Human peripheral gd T-cells recognize hsp60 molecules on Daudi Burkitts lymphoma cells. J. Immunol. 150, 2046–2055. Kaur, I., De Jong, J., Schell, K., Hank, J., Fisch, P. and Sondel, P.M. Ž1994. Human peripheral gd T-cells are stimulated by Daudi Burkitts lymphoma and not by any other Burkitts lymphoma tested. Cell. Immunol. 156, 54–61. Nick, S., Pileri, P., Tongiani, S., Uematsu, Y., Kappos, L. and De Libero, G. Ž1995. T-cell receptor gd repertoire is skewed in cerebrospinal fluid of multiple sclerosis patients: Molecular and functional analyses of antigen-reactive gd clones. Eur. J. Immunol. 25, 355–363. Satoh, J., Nomaguchi, H. and Tabira, T. Ž1992. Constitutive expression of 65-kDa heat shock protein ŽHSP65.-like immunoreactivity in cultured mouse oligodendrocytes. Brain Res. 595, 281–290. Selin, L.K., Stewart, S., Shen, C., Mao, H.Q. and Wilkins, J.A. Ž1992. Reactivity of gd T-cells induced by the tumour cell line RPMI 8226: Functional heterogeneity of clonal populations and role of GroEL heat shock proteins. Scand. J. Immunol. 36, 107–117. Selmaj, K., Brosnan, C.F. and Raine, C.S. Ž1991. Colocalization of lymphocytes bearing gd T-cell receptor and heat shock protein hsp65q oligodendrocytes in multiple sclerosis. Proc. Natl. Acad. Sci. USA 88, 6452–6456.
Sharif, M., Worrall, J.G., Singh, B., Gupta, R.S., Lydyard, P.M., Lambert, C., McCulloch, J. and Rook, G.A. Ž1992. The development of monoclonal antibodies to human mitochondrial 60 kDa heat-shock protein and their use in studying the expression of the protein in rheumatoid arthritis. Arthritis Rheum. 35, 1427–1433. Sibley, W.A., Bamford, C.R. and Clark, K. Ž1985. Clinical viral infections and multiple sclerosis. Lancet 1Ž8441., 1313–1315. Skeen, M.J. and Ziegler, H.K. Ž1995. Activation of gd T-cells for production of IFN-g is mediated by bacteria via macrophage-derived cytokines IL-1 and IL-12. J. Immunol. 154, 5832–5841. Steinman, L., Waisman, A. and Altmann, A. Ž1995. Major T-cell responses in multiple sclerosis. Mol. Med. Today 1, 79–83. Stinissen, P., Vandevyer, C., Medaer, R., Vandegaer, L., Nies, J., Tuyls, L., Hafler, D.A., Raus, J. and Zhang, J. Ž1995. Increased frequency of gd T-cells in cerebrospinal fluid and peripheral blood of patients with multiple sclerosis: Reactivity, cytotoxicity and T-cell receptor V gene rearrangements. J. Immunol. 154, 4883–4894. Williams, K., Bar-Or, A., Ulvestad, E., Olivier, A., Antel, J.P. and Yong, V.W. Ž1992. Biology of adult human microglia in culture: Comparisons with peripheral blood monocytes and astrocytes. J. Neuropathol. Exp. Neurol. 51, 538–549. Windhagen, A., Newcombe, J., Dangond, F., Strand, C., Woodroofe, N., Cuzner, M.L. and Hafler, D.A. Ž1995. Expression of co-stimulatory molecules B7-1 ŽCD80., B7-2 ŽCD86. and interleukin 12 cytokine in multiple sclerosis lesions. J. Exp. Med. 182, 1985–1996. Wucherpfennig, K.W., Newcombe, J., Li, H., Keddy, C., Cuzner, M.L. and Hafler, D.A. Ž1992. gd T-cell receptor repertoire in acute multiple sclerosis lesions. Proc. Natl. Acad. Sci. USA 89, 4588–4592. Young, R.A. and Elliott, T.J. Ž1989. Stress proteins, infection and immune surveillance. Cell 59, 5–8.
gd T-cell–human glial cell interactions. I. In vitro induction of gd T-cell expansion by human glial cells Mark S. Freedman a
a,)
, Sameer D’Souza b, Jack P. Antel
b
UniÕersity of Ottawa, Department of Medicine, DiÕision of Neurology, Ottawa General Hospital, 501 Smyth Road, Ottawa, Ont., Canada K1H 8L6 McGill UniÕersity, Neuroimmunology Unit, Department of Neurology and Neurosurgery, Montreal Neurological Institute, 3801 UniÕersity Street, Montreal, Que., Canada H3A 2B4
b
Received 26 August 1996; revised 1 November 1996; accepted 4 November 1996
Abstract
gd T-cells are found in increased proportion in multiple sclerosis ŽMS. white matter plaque infiltrates compared with peripheral blood or spleen, raising the possibility that they are either specifically attracted to lesion sites or, once present, are stimulated to expand. We have previously shown that human oligodendrocytes ŽOGC. preferentially express heat shock proteins Ž hsp ., molecules to which gd T-cells have been known to react and that in vitro expanded gd T-cells can lyse OGC. We therefore investigated whether human glial cells, that differentially express hsp, could stimulate gd T-cell expansion from peripheral blood. We compared the glial cell-induced expansion to cell lines which also differentially express hsp and have been shown to selectively stimulate gd T-cell expansion Že.g. RPMI 8226, Daudi.. We found that both OGC and human fetal astrocytes ŽhFA. expressed hsp and stimulated the preferential expansion of gd T-cells to about the same extent as the hsp expressing cell lines RPMI 8226 or Daudi, in the presence of exogenous interleukin-2 ŽIL-2. but without any T-cell mitogen. Furthermore, the type of gd T-cells expanded were of the Vd 2 subtype known to be particularly reactive to hsp. Microglia, U937 cell lines or purified myelin membranes, which express little or no hsp, did not support gd T-cell growth. These results therefore suggest that OGC may contribute to the local expansion of gd T-cells within MS plaques. Potential harmful effects of gd T-cells on OGC may thereby contribute to the immunopathogenesis of MS demyelination. Keywords: gd T-cell; Glia; Heat shock protein; T-cell stimulation; Multiple sclerosis
1. Introduction T-cells are concentrated within the multifocal perivenular areas of inflammation in the central nervous system white matter, that characterize the human autoimmune demyelinating disease multiple sclerosis ŽMS.. It is believed that these T-cells either directly Žthrough cell contact. or indirectly Žthrough cytokine release. contribute to the accompanying demyelination of MS, though the exact mechanisms are not known. The predominant T-cell type in the lesions expresses an ab T-cell receptor ŽTCR., some of which may be specifically reactive with myelin antigens, while others are not ŽSteinman et al., 1995.. Another T-cell type, expressing a gd T-cell receptor, has been identified in MS lesions and appears to be present in considerably increased proportion compared with periph-
) Corresponding author. Tel.: q1-613-7378917; fax: q1-613-7378857; e-mail: [email protected]
eral blood or spleen ŽWucherpfennig et al., 1992.. This suggests that either gd T-cells are specifically recruited from the periphery and persist in lesion sites or that gd T-cells infiltrating the CNS, as part of a generalized inflammatory response, are stimulated to expand within the CNS. Since activated and expanded gd T-cells can in vitro be shown to lyse human oligodendrocytes ŽOGC. ŽFreedman et al., 1991., it is possible they play an important role in the pathogenesis of the MS lesion. gd T-cells are known to react in a predominantly MHC non-specific matter with certain heat-shock proteins Ž hsp ., particularly the 65 kDa proteins of Mycobacteria or groEL-related mammalian homologues such as hsp60 ŽFisch et al., 1990; Kaufmann, 1992; Kaur et al., 1993; Haas et al., 1993.. Interaction with hsp60 expressed on the surface of Daudi cells has been shown to be key to the subsequent Daudi induced gd T-cell expansion ŽKaur et al., 1993.. OGC express little or no MHC ŽGrenier et al., 1989. but do express hsp60, possibly more so than other human glial cells ŽFreedman et al., 1992; Satoh et al.,
0165-5728r97r$17.00 Copyright q 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 6 5 - 5 7 2 8 Ž 9 6 . 0 0 2 1 7 - 2
136
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
1992.. In this paper we investigated whether OGC, compared to other human glial cells, will stimulate the outgrowth and expansion of gd T-cells from peripheral blood. In a related study ŽFreedman et al., 1997., we address the specificity of the gd T-cell cytolytic response to hsp60.
2. Materials and methods 2.1. Cell lines and glial cell cultures RPMI 8226, Daudi and U937 cells were obtained from ATCC ŽRockville., grown and maintained in RPMI with 10% FCS ŽGibco BRL, Burlington. and 10% gentamicin Žexcept RPMI 8226.. Cell lines were split and fed with fresh media 2–3 times per week. Dissociated glial cell cultures enriched for OGC ŽFreedman et al., 1991., microglia ŽWilliams et al., 1992. or human fetal astrocytes ŽhFA. ŽD’Souza et al., 1995. were prepared as previously described in detail. Briefly, human brain tissue was obtained from patients undergoing surgery for epilepsy and cells isolated by trypsin-dissociation and Percoll gradient centrifugation. Microglia and OGC were enriched from mixed cultures based on their tendency to adhere Žmicroglia. or not ŽOGC. to poly-D-lysine coated coverslips ŽWilliams et al., 1992.. OGC cultures used for
induction experiments showed purities of ) 90% by galactocerebroside staining, whereas those used for histochemistry were ) 90% by CNPase immunoreactivity ŽFreedman et al., 1992.; microglia were ) 95% CD11c positive ŽWilliams et al., 1992.. Human fetal astrocytes ŽhFA. Ž) 95% GFAPq . were derived from 12–14 week abortuses following MRC ŽCanada. approval guidelines, as previously described ŽD’Souza et al., 1995.. All cell lines and cultures were maintained in a humidified 378C 5% CO 2 incubator. 2.2. T-cell cultures Freshly isolated peripheral blood mononuclear cells ŽPBMC. were obtained from healthy volunteers by FicollHypaque gradient centrifugation. 1 = 10 6 PBMC were placed in 24 well plates ŽNunclon, Gibco BRL, Burlington. with either 3.5 = 10 4 irradiated Ž5,000 Rads. RPMI 8226, U937 cells, OGC, microglia or hFA in 2 ml of serum-free media ŽAim-V, Gibco BRL, Burlington. and 50 Urml recombinant human IL-2 Ža generous gift from IMMUNEX, Seattle.. Media was changed with fresh IL-2 on day 6 of a 10 day culture. In some cases, purified human myelin Ža gift of J. McLaurin, Montreal PQ. was coated to wells and in other experiments, 2 = 10 5 Daudi cells were used as stimulators in 10% FCS according to a
Fig. 1. Constitutive surface ŽA, C. or intracellular ŽB, D. staining of either OGC ŽA, B. or hFA ŽC, D. with monoclonal Ab to hsp60, followed by biotinylated goat anti-mouse Ab and FITC-conjugated streptavidin; inset is the isotype control. Magnification: of ŽA. and ŽB. is =250, of ŽC. and ŽD. is =350.
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
137
previously published protocol ŽFisch et al., 1990.. For human glia, cells were pre-plated in experimental wells in order to maximize adherence and viability. After 10 days, cells were pipetted from wells, dead cells removed by Ficoll-Hypaque separation and viable Žtrypan blue excluding. cells counted and stained for flow cytometric analysis. 2.3. Immunohistochemisty For surface staining of hsp60, adult OGC or hFA on coverslips were incubated for 1 h with mouse anti-human hsp60 antibody ŽAb. ŽAffinity Bioreagents, Neshanic Station. or mouse anti-human IgG2a isotype ŽR & D Systems, Minneapolis. and washed three times with phosphate buffered saline ŽPBS.. Cells were then counterstained by incubation for 1 h with biotinylated goat anti-mouse Ab ŽVector Laboratories, Burlingame. at room temperature ŽRT., followed by another 1 h incubation with fluorescein-conjugated stretavidin ŽBoehringer Mannheim Biochemica, Mannheim., before fixing them in 4% paraformaldehyde ŽPFA. for 20 min at 48C and mounting coverslips with Gelvatol. For intracellular staining, coverslips were first fixed in ice cold Žy208C. 95% ethanolr5% acetic acid for 20 min, then rehydrated for 30 min with PBS and stained as above. Unless otherwise stated, all immunostaining was performed at RT, all Ab reagents were used at a dilution of 1:100 and all coverslips were washed extensively with PBS between steps. 2.4. Flow cytometry Staining of cells for flow cytometry has been previously described ŽFreedman et al., 1991. using either single or double staining techniques. Briefly, cell lines in suspension were incubated with a 1:40 dilution of polyclonal rabbit anti-hsp58 Ž hsp60, courtesy of W. Welch, San Francisco. and counter-stained with FITC-conjugated goat anti-rabbit IgG Ž1:100. ŽGibco BRL, Burlington.. ŽStaining of cell lines with the monoclonal Ab to human hsp60 used above did not demonstrate the same degree of surface expression as that seen with the polyclonal serum, similar to the reports of others ŽKaur et al., 1993... T-cell cultures were double stained with monoclonal antibodies ŽmAb.: antiTCR-d 1-FITC Žrecognizing all gd T-cells via the C d epitope. ŽT-cell Sciences, Intermedico, Markham. or antiTCR-arb-1-FITC, Leu-4-PE Žanti-CD3. ŽBecton-Dickinson, Mountain View.. All cell lines were maintained in 1% paraformaldehyde in PBS prior to analysis on a FACScan ŽBecton-Dickinson., gating out only debris. %CD3q gd T-cells was calculated based on the TCR-d 1-FITCrLeu-4PE quadrant and verified by conversely staining with TCR-arb-1-FITCrLeu-4-PE. In some cases, the gd Tcells grown were analyzed for subtype Ži.e. Vd 1r2. using either TCS-1 Žrecognizing Vd 1. or Tg iA Žrecognizing Vg 9 which is mostly associated with Vd 2 ŽHaas et al., 1993. mAb ŽT-cell Sciences. conjugated to FITC and
Fig. 2. RPMI 8226 or U937 cells were immunostained for hsp60 expression using rabbit polyclonal anti-hsp58, anti-GFAP antibodies or the second antibody Žgoat anti-rabbit-FITC. alone ŽFITC control. and analyzed by flow cytometry. Expression of hsp60 by RPMI 8226 cells supercedes that of U937 cells Ž ) 20=., which partially overlaps the non-specific staining by the rabbit polyclonal anti-GFAP control.
138
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
double stained with anti-TCR grd-1-PE ŽBecton Dickinson..
3. Results 3.1. Expression of hsp60 by human glial cells.
2.5. Statistical analysis Where indicated, statistical analysis was performed using a Mann–Whitney non-parametric unpaired analysis, calculating a one-tailed p value, with a software package called Instat from GraphPad e .
We have previously demonstrated that adult human glial cells differentially express hsp60 using polyclonal rabbit antiserum ŽFreedman et al., 1992.. We repeated these studies now using a monoclonal antibody to human mitochondrial hsp60 that cross reacts with the hsp homo-
Fig. 3. Fresh PBMC were cultured in serum-free media with irradiated stimulator cells ŽRPMI 8226 or U937. cells Ž30:1 PBMC:stimulator cell ratio. for 10 days with 50 Urml rhIL-2. Daudi cell cultures utilized a 5:1 PBMC:Daudi cell ratio in 10% FCS. In the presence of either RPMI 8226 or Daudi cells, there was a marked expansion of gd T-cell proportions Žtop panel. and cell numbers Žlower panel. from initial PBMC cultures containing proportions of 2–12%, whereas U937 cells did not significantly increase either proportions or cell numbers. Media alone contained rhIL-2 but no stimulator cells. Depicted are the results of at least three separate experiments from different individuals" S.E.M.
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
logue of Mycobacteria ŽSharif et al., 1992. and in addition we characterized the expression of hsp60 on hFA. As we previously observed, OGC constitutively express hsp60 on their surface, as well as intracellularly ŽFig. 1A, B.. In contrast to adult human astrocytes that only demonstrated a minor amount of punctate surface staining, if at all ŽFreedman et al., 1992., most hFA expressed both surface and intracellular hsp60 ŽFig. 1C, D.. Qualitatively, hsp60 surface staining of hFA appeared to be less than that of OGC.
139
3.2. Expression of hsp60 by cell lines Daudi cells have been shown elsewhere to express hsp60 ŽFisch et al., 1990; Kaur et al., 1993., but here we demonstrate that another B cell line RPMI 8226, that we and others have used extensively to expand peripheral blood gd T-cells ŽFreedman et al., 1991; Selin et al., 1992., also stains strongly for hsp60 Ža full log difference in mean channel fluorescence., compared to the lower levels seen on U937 cells ŽFig. 2..
Fig. 4. Fresh PBMC were co-cultured with irradiated stimulator cells ŽRPMI 8226 or glia; 30:1 PBMC:stimulator cell ratio. in serum-free media for 10 days. In the presence of either OGC or hFA, gd T-cells expanded, with statistically significant Ž ) . increases in both their proportions Žtop panel. and cell numbers Žlower panel. compared to culture in media alone. Neither purified myelin membranes nor microglia stimulated significant gd T-cell outgrowth. Depicted are the results of at least three separate experiments from different individuals" S.E.M.
140
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
3.3. Induction of gd T-cell expansion by cell lines It has been shown that hsp60 expressing Daudi or RPMI 8226 cells induced the expansion of gd T-cells from peripheral blood without the need for T-cell mitogen or TCR cross-linking with only the addition of exogenous IL-2 ŽFisch et al., 1990; Freedman et al., 1991; Selin et al., 1992.. Both polyclonal and monoclonal Ab to hsp60 would abrogate this stimulation ŽFisch et al., 1990; Selin et al., 1992; Kaur et al., 1993.. We and others have observed that the stimulation induced by the RPMI 8226 cells is maximal when cultures are maintained in serum free media ŽAim V. ŽFreedman et al., 1991; Selin et al., 1992.. We therefore conducted all experiments in serum free media using the PBMC:RPMI 8226 30:1 ratio previously determined to be optimal ŽFreedman et al., 1991; Selin et al., 1992. as a standard with which to compare other cells. Under similar serum-free conditions, U937 cells did not stimulate gd T-cell growth ŽFig. 3., even though low levels of hsp60 surface staining was noted ŽFig. 2.. U937 cells did not support growth whether or not serum was contained in the media Žresults not shown.. The 30:1 ratio used in our study was too dilute for Daudi cells to support the gd T-cell growth, even when using 10% FCS Žresults not shown., however at the reported 5:1 ratio ŽFisch et al., 1990., gd T-cells readily expanded ŽFig. 3.. These data demonstrate that RPMI 8226, like Daudi cells, both express high levels of hsp60 and strongly stimulate gd T-cell expansion from PBMC, whereas the lower hsp60 expressing U937 cells do not. 3.4. Induction of gd T-cell expansion by human glial cells Using the same 30:1 ratio that was optimal for gd T-cell expansion with RPMI 8226 cells, we investigated whether or not OGC, microglia or hFA had any similar stimulator capability. Some growth was observed using microglia, but this was not significantly different compared to media alone Ž p s 0.197 for % and 0.066 for cell number, Mann–Whitney.. The OGC induced gd T-cell growth to about the same extent as RPMI 8226 cells, both representing a statistically significant increases over media alone Ž p - 0.004 Ž%., p - 0.006 Žcell number. for the OGC; p - 0.0005 Ž%., p - 0.0009 Žcell number. for the RPMI 8226 cells. ŽFig. 4.. The proportion of gd T-cells did not increase as much as the absolute gd T-cell number in the OGC induced cultures compared with the RPMI 8226 cell line, suggesting that both ab and gd T-cells expanded in the presence of the OGC, whereas the RPMI 8226 may actually be suppressing ab T-cell growth ŽJ. Wilkins, personal communication.. hFA, unlike adult astrocytes, do express hsp60 and also induced a modest but significant gd T-cell expansion compared to media Ž p s 0.0465 Ž%., p s 0.0011 Žcell number.. To test whether it was the OGC itself or a myelin antigen that stimulated gd T-cell growth, we coated some
Table 1 gd T-cells derived from at least six different cultures stimulated via different cell types were stained for gd subtype Žsee Section 2. and analyzed by flow cytometry. Stimulator cell
RPMI 8226 Daudi Oligodendrocytes Fetal astrocytes
Proportion a of gd T-cells Ž"S.E.M.. Vd 1
Vd 2
4"1.7 5.9"1.7 13.93"2.19 ND
98.1".37 97.1"1.15 89"1.9 ND
a
% of CD3q cells. NDs not done.
wells with 0.5 ml or a 1:10 mixture of purified human myelin but did not see any significant growth compared to media ŽFig. 4.. We were not able to evaluate experiments in which the same anti-hsp58 rabbit antibody previously described to block Daudi cell induction was used to block the expansion of gd T-cells ŽFisch et al., 1990., because the Ab was toxic to both stimulator cell and PBMC in serum-free conditions. We did not have access to the same hybridoma supernatant for the hsp60 previously used to block Daudiinduced gd T-cell growth ŽKaur et al., 1993.. Nevertheless, these results do indicate that the ability to induce the expansion of gd T-cells from PBMC correlates with hsp60 expression. 3.5. Phenotype of gd T-cells stimulated hsp60, whether a soluble antigen or expressed by a cell or bacterium appears to stimulate mainly gd T-cells bearing the Vd 2 as opposed to the Vd 1 TCR ŽHaas et al., 1993.. In keeping with this observation, the gd T-cells stimulated by RPMI 8226, Daudi or OGC were all predominantly of the Vd 2 subtype ŽTable 1..
4. Discussion The mechanism of myelin damage in MS is unknown, however it is felt that T-cells either directly or indirectly mediate this phenomenon by their presence in MS plaques. gd T-cells are concentrated in MS plaque infiltrates, whereas they are few in number in peripheral blood or in other organs such as the spleen ŽWucherpfennig et al., 1992.. We now demonstrate that adult OGC are also capable of stimulating the growth and expansion of gd T-cells, which were predominantly of the Vd 2 phenotype, consistent with the reported findings in MS plaque tissue ŽWucherpfennig et al., 1992; Hvas et al., 1993.. In some studies ŽSelmaj et al., 1991; Wucherpfennig et al., 1992., gd T-cell concentrates co-localized with OGC expressing hsp60. Indeed, co-localization of gd T-cell infiltrates and areas of increased hsp60 immunoreactivity have even been
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
noted in chronic lesions of EAE ŽGao et al., 1995.. It would appear that it is the adult OGC itself and not just the myelin that stimulates the gd T-cell expansion, as we observed no growth in the presence only of purified myelin ŽFig. 4.. Cell lines and human glia that express hsp60 appear to share in the ability to stimulate the selective expansion of gd T-cells from the small numbers seen in peripheral blood to significant proportions. Others have shown this to be dependent on hsp60 recognition by the blocking of expansion using Ab to hsp60 on Daudi cells ŽKaur et al., 1993.. Kaur et al. Ž1994. have also demonstrated that other Burkitt’s lyphoma cells Žsuch as Raji., not expressing hsp60, are neither capable of inducing gd T-cell growth nor acting as good targets for gd T-cell-mediated lysis. hFA, unlike their adult counterparts, expressed detectable levels of hsp60. Although the levels of surface hsp60 were qualitatively lower than OGC by histochemistry, the hFA were capable of inducing significant gd T-cell expansion ŽFig. 4.. Others have observed similar lower level expression of hsp60 on murine astrocytes compared to OGC ŽSatoh et al., 1992.. We could not determine whether quantitatively the amount of hsp60 expression correlated with gd T-cell expansion, since it was not possible to perform flow cytometry on the glial cells. We have previously shown that some, but probably not all brain derived microglia express hsp ŽFreedman et al., 1992. and this may have accounted for the perceived induction of gd T-cell expansion we observed, though it was not significantly different from media; likely the result of wide variability among preparations ŽFig. 4.. Taken together, our data would suggest that hsp60 expression on glial cells correlates with the ability to induce gd T-cell expansion from peripheral blood and may be one brain antigen that contributes to the local accumulation of gd T-cells in MS lesions. Others have demonstrated that gd T-cells derived from either the blood or spinal fluid of MS patients have skewing of their TCR repertoire indicating a response to a common brain antigen, possibly hsp ŽBattistini et al., 1995; Stinissen et al., 1995; Nick et al., 1995.. Knowledge of the natural role and function of gd T-cells in the immune response is still in its infancy. It appears that these cells are not only MHC unrestricted, but they also do not require antigen processing and presentation, affording them the ability to directly recognize antigens on pathogens or on damaged tissues that may be expressing hsp60 as part of a stress response ŽHaas et al., 1993.. Infectious processes that turn on gd T-cells as part of a ‘first line defense’ ŽYoung and Elliott, 1989. are well known to precede attacks of MS ŽSibley et al., 1985.. Once activated in the periphery, gd T-cells could cross the blood–brain-barrier where they might encounter cells expressing hsp60. Although the fever associated with systemic illness could act as a direct inducer of hsp in the brain, we have shown that IL-1 Žthe endogenous brain pyrogen. induces OGC hsp expression in vitro ŽD’Souza
141
et al., 1994.. In mice, IL-1 and IL-12 synergistically activate and probably lead to the accumulation of gd T-cells at sites of infection ŽSkeen and Ziegler, 1995.. Recently it has been shown that acute MS plaque tissue also harbours significant amounts of IL-12 ŽWindhagen et al., 1995.. Burns et al. Ž1995. have demonstrated that gd T-cells expand in response to activated myelin reactive T-cell clones. There are therefore many possible reasons for the localized expansion of gd T-cells in MS plaques. Our current work suggests that hsp60 expressing OGC could serve to either attract or retain gd T-cells in the brain, by stimulating their expansion. These cells may be potentially harmful, given their ability in vitro to lyse OGC ŽFreedman et al., 1991.. Whether or not this lytic capability also depends on the expression hsp60 is the subject of a related paper ŽFreedman et al., 1997..
Acknowledgements We are indebted to Ms. Marta Bachetti for her excellent technical assistance. This work was supported in part by a research grant from the Multiple Sclerosis Society of Canada to M.S.F.
References Battistini, L., Salvetti, M., Ristori, G., Falcone, M., Raine, C.S. and Brosnan, C.F. Ž1995. gd T-cell receptor analysis supports a role for HSP 70 selection of lymphocytes in multiple sclerosis lesions. Mol. Med. 1, 554–562. Burns, J., Bartholomew, B. and Littlefield, K. Ž1995. gd T-cells participate in the immune response against activated, myelin basic proteinspecific, human T-cells. J. Neuroimmunol. 58, 177–182. D’Souza, S.D., Antel, J.P. and Freedman, M.S. Ž1994. Cytokine induction of heat shock protein expression in human oligodendrocytes: An interleukin-1-mediated mechanism. J. Neuroimmunol. 50, 17–24. D’Souza, S., Alinauskas, K., McCrea, E., Goodyer, C. and Antel, J.P. Ž1995. Differential susceptibility of human CNS-derived cell populations to TNF-dependent and independent immune-mediated injury. J. Neurosci. 15, 7293–7300. Fisch, P., Malkovsky, M., Kovats, S., Sturm, E., Braakman, E., Klein, B.S., Voss, S.D., Morrissey, L.W., DeMars, R., Welch, W.J., Bolhuis, R.L.H. and Sondel, P.M. Ž1990. Recognition by human Vg 9rVd 2 T-cells of a GroEL homolog on Daudi Burkitts lymphoma cells. Science 250, 1269–1273. Freedman, M.S., Ruijs, T.C.G., Selin, L.K. and Antel, J. Ž1991. Peripheral g-d T-cells lyse fresh human brain-derived oligodendrocytes. Ann. Neurol. 30, 794–800. Freedman, M.S., Buu, N.N., Ruijs, T.C.G., Williams, K. and Antel, J.P. Ž1992. Differential expression of heat shock proteins by human glial cells. J. Neuroimmunol. 41, 231–238. Freedman, M.S., Bitar, R. and Antel, J.P. Ž1997. gd T-cell-human glial cell interactions. II. Relationship between heat shock protein expression and susceptibility to cytolysis. J. Neuroimmunol. 74, 143–148. Gao, Y.L., Brosnan, C.F. and Raine, C.S. Ž1995. Experimental allergic encephalomyelitis. Qualitative and semiquantitative differences in heat shock protein 60 expression in the central nervous system. J. Immunol. 154, 3548–3556. Grenier, Y., Ruijs, T.C.G., Robitaille, Y., Olivier, A. and Antel, J.P.
142
M.S. Freedman et al.r Journal of Neuroimmunology 74 (1997) 135–142
Ž1989. Immunohistochemical studies of adult human glial cells. J. Neuroimmunol. 21, 103–115. Haas, W., Pereira, P. and Tonegawa, S. Ž1993. Gammardelta T-cells. Ann. Rev. Immunol. 11, 637–685. Hvas, J., Oksenberg, J.R., Fernando, R., Steinman, L. and Bernard, C.A. Ž1993. gd T-cell receptor repertoire in brain lesions of patients with multiple sclerosis. J. Neuroimmunol. 46, 225–234. Kaufmann, S.H.E. Ž1992. The cellular response to heat shock proteins. Experientia 48, 640–643. Kaur, I., Voss, S.D., Gupta, R.S., Schell, K., Fisch, P. and Sondel, P.M. Ž1993. Human peripheral gd T-cells recognize hsp60 molecules on Daudi Burkitts lymphoma cells. J. Immunol. 150, 2046–2055. Kaur, I., De Jong, J., Schell, K., Hank, J., Fisch, P. and Sondel, P.M. Ž1994. Human peripheral gd T-cells are stimulated by Daudi Burkitts lymphoma and not by any other Burkitts lymphoma tested. Cell. Immunol. 156, 54–61. Nick, S., Pileri, P., Tongiani, S., Uematsu, Y., Kappos, L. and De Libero, G. Ž1995. T-cell receptor gd repertoire is skewed in cerebrospinal fluid of multiple sclerosis patients: Molecular and functional analyses of antigen-reactive gd clones. Eur. J. Immunol. 25, 355–363. Satoh, J., Nomaguchi, H. and Tabira, T. Ž1992. Constitutive expression of 65-kDa heat shock protein ŽHSP65.-like immunoreactivity in cultured mouse oligodendrocytes. Brain Res. 595, 281–290. Selin, L.K., Stewart, S., Shen, C., Mao, H.Q. and Wilkins, J.A. Ž1992. Reactivity of gd T-cells induced by the tumour cell line RPMI 8226: Functional heterogeneity of clonal populations and role of GroEL heat shock proteins. Scand. J. Immunol. 36, 107–117. Selmaj, K., Brosnan, C.F. and Raine, C.S. Ž1991. Colocalization of lymphocytes bearing gd T-cell receptor and heat shock protein hsp65q oligodendrocytes in multiple sclerosis. Proc. Natl. Acad. Sci. USA 88, 6452–6456.
Sharif, M., Worrall, J.G., Singh, B., Gupta, R.S., Lydyard, P.M., Lambert, C., McCulloch, J. and Rook, G.A. Ž1992. The development of monoclonal antibodies to human mitochondrial 60 kDa heat-shock protein and their use in studying the expression of the protein in rheumatoid arthritis. Arthritis Rheum. 35, 1427–1433. Sibley, W.A., Bamford, C.R. and Clark, K. Ž1985. Clinical viral infections and multiple sclerosis. Lancet 1Ž8441., 1313–1315. Skeen, M.J. and Ziegler, H.K. Ž1995. Activation of gd T-cells for production of IFN-g is mediated by bacteria via macrophage-derived cytokines IL-1 and IL-12. J. Immunol. 154, 5832–5841. Steinman, L., Waisman, A. and Altmann, A. Ž1995. Major T-cell responses in multiple sclerosis. Mol. Med. Today 1, 79–83. Stinissen, P., Vandevyer, C., Medaer, R., Vandegaer, L., Nies, J., Tuyls, L., Hafler, D.A., Raus, J. and Zhang, J. Ž1995. Increased frequency of gd T-cells in cerebrospinal fluid and peripheral blood of patients with multiple sclerosis: Reactivity, cytotoxicity and T-cell receptor V gene rearrangements. J. Immunol. 154, 4883–4894. Williams, K., Bar-Or, A., Ulvestad, E., Olivier, A., Antel, J.P. and Yong, V.W. Ž1992. Biology of adult human microglia in culture: Comparisons with peripheral blood monocytes and astrocytes. J. Neuropathol. Exp. Neurol. 51, 538–549. Windhagen, A., Newcombe, J., Dangond, F., Strand, C., Woodroofe, N., Cuzner, M.L. and Hafler, D.A. Ž1995. Expression of co-stimulatory molecules B7-1 ŽCD80., B7-2 ŽCD86. and interleukin 12 cytokine in multiple sclerosis lesions. J. Exp. Med. 182, 1985–1996. Wucherpfennig, K.W., Newcombe, J., Li, H., Keddy, C., Cuzner, M.L. and Hafler, D.A. Ž1992. gd T-cell receptor repertoire in acute multiple sclerosis lesions. Proc. Natl. Acad. Sci. USA 89, 4588–4592. Young, R.A. and Elliott, T.J. Ž1989. Stress proteins, infection and immune surveillance. Cell 59, 5–8.