Antigen presentation by interferon-γ-treated thyroid follicular cells inhibits interleukin-2 (IL-2) and supports IL-4 production by B7-dependent human T cells

62 G. Lornbardi et al. Eur. J . Irnrnunol. 1997.27: 62-71

Giovanna Lombard?, Kate Arnold’, Julia Uren’, Federica Marelli-Berg’, Roseanna Hargreaves’, Nesrina Imami’, Anthony Weetman’ and Robert Lechler’

’ Department of Immunology, Royal Postgraduate Medical School, London, GB

* Department of Medicine, The University of Sheffield, Clinical Sciences Centre, Northern General Hospital, Sheffield, GB

Antigen presentation by interferon-y-treated thyroid follicular cells inhibits interleukin-2 (IL-2) and supports IL-4 production by B7-dependent human T cells

The consequence of recognition of antigen on antigen-presenting cells that are induced to express major histocompatibility complex (MHC) class I1 molecules following an inflammatory process is still not clear. In this study, we have investigated the outcome of antigen presentation by epithelial cells and we have used as a model thyroid follicular cells (TFC) that are known to express MHC class I1 molecules in autoimmune thyroid diseases and acquire the capacity to present autoantigens to T cells infiltrating the thyroid gland. The result show that MHC class 11-expressing TFC were unable to stimulate a primary T cell alloresponse, using CD4’ T cells from three HLA-mismatched responders. Phenotypic analysis showed that TFC, after incubation with interferon-y, do not express the co- stimulatory molecules B7-1 (CD80) and -2 (CD86). Addition of murine DAP.3 cells expressing human B7-1 (DAP.3-B7) to cultures containing peripheral blood CD4’ Tcells and DR1-expressing TFC led to a proliferative response, suggesting that the failure of TFC to stimulate a primary alloresponse was due to a lack of co-stimulation. Similarly, HLA-DR-restricted, influenza-specific T cell clones dependent on B7 for co-stimulation did not respond to peptide presented by TFC; again the lack of response could be overcome by co-culture of TFC with DAP.3-B7. Furthermore, recognition of antigen on TFC inhibited interleukin-2 (IL-2) production in the B7-dependent T cells. In contrast, in T helper type 0 (Tho) T cells, IL-4 release was not affected by TFC presentation. In addition, antigen presentation by TFC favored IL-4 production relative to IL-2 production by B7-indpendent Tho clones. These results suggest that antigen presentation by MHC class 11’ TFC may induce tolerance in autoreactive Thl cells but may simultaneously favors a Th2 response in uncommitted T cells, and thereby support autoantibody production.

1 Introduction

The induction of interleukin-2 (IL-2) production by CD4’ T cells involves at least two independent signaling path- ways. The first results from engagement of the T cell receptor (TcR) : CD3 : CD4 complex, and the second from the interaction of CD2UCTLA-4 with the B7 family of co- stimulatory molecules expressed by specialized antigen- presenting cells (APC) [l, 21. Furthermore, it has recently become clear that the recognition of ligand in the absence of co-stimulatory signals can silence the IL-2 gene, and render the T cell refractory to subsequent challenge as far as IL-2 secretion is concerned [3-51. Indeed antigen presentation by cells that can be induced to express class I1 molecules following activation, such as keratinocytes [6, 71, pancreatic p cells [S] and human Tcells [9-121 has been shown to inhibit IL-2 production by T cells.

[I 157511

Received May Y, 1996; in final revised form October 9, 1996; accepted October 15, 1996.

Correspondence: Giovanna Lombardi, Department of Irnrnuno- logy, RPMS, Hammersmith Hospital, Du Cane Road, London W12 O N N , GB Fax: 44 181 7438602; e-mail: [email protected]

Abbreviations: TFC: Thyroid follicular cells HA: Hemagglutinin

Key words: Thyroid follicular cell / Human T cell clone / B7 co- stimulation I Tolerance

In this context, the functional consequences of antigen presentation by tissue parenchymal cells that are induced to express MHC class I1 molecules in the course of an inflammatory response is a matter of debate. For example, thyroid follicular cells (TFC) class I1 expression is a well- established feature of autoimmune thyroid diseases, such as Graves’ disease and Hashimoto’s thyroiditis [13]. It has also been demonstrated in v i m by addition of IFN-y [14, 151. The detection of class 11-expressing TFC in thyroid autoimmune diseases led to the suggestion that this “aber- rant” expression of class I1 molecules results in the activation of autoreactive lymphocytes that either participate in the initiation or perpetuation of an autoimmune disease (161. However, conflicting data have been reported regard- ing the APC function of TFC.

Several studies have shown that TFC are capable of presenting thyroid antigens and causing proliferation of T cell clones and lines derived from thyroid tissues from patients with Graves’ disease, although such stimulation is weak [17-211. Graves’ TFC were also found to present foreign peptide, derived from influenza hemagglutinin, to T cell clones but were unable to process the intact protein for presentation [22], and TFC have also been reported to stimulate T cells in an autologous mixed lymphocyte reac- tion (MLR) [23]. A class 11-negative rat thyroid cell line, WRT, weakly stimulated CD4’ T cells, and the induction of class I1 expression on the line with IFN-y had no further effect o n T cell stimulation, making the nature of this presentation pathway unclear [24].

0014-2980/97/0101-62$10.00 + .25/0 0 VCH Verlagsgesellschaft mbH, D-69451 Weinheim, 1997

Eur. J. Immunol. 1997.27; 62-71 Antigen presentation by thyroid follicular cells favors a Th2 response 63

In contrast, others have shown that TFC are either ineffi- cient at antigen presentation, or have no ability to stimulate primary or secondary T cell responses [25-281. In the mouse system it has been shown that a cloned thyroid- derived epithelial cell line, M.5, could be induced to express MHC class I1 molecules but failed to induce a primary MLR [25]. The early treatment of the line M.5 with X-irradiation or the addition of phorbol esters restored T cell proliferation to TFC [28, 291. Similar results were obtained in the human system; the addition of phorbol esters, but not IL-1, restored the lack of T cell proliferation to Graves’ disease TFC [30]. Although IFN-y-treated TFC were able to induce a weak peripheral blood T cell response to PPD, the addition of suboptimal numbers of monocytes synergistically augmented antigen presentation by thyrocytes [31].

These divergent results could be explained by the contami- nation of TFC preparations with a small population of “professional” APC in those experiments in which stimulation has been observed. Dendritic cells in particular are enriched in the thyroid glands of patients with Graves’ disease and Hashimoto‘s thyroiditis [32]. In addition, many of these studies antedated the recognition that CD4’ T cells can be divided into subpopulations according to the pattern of cytokines that they produce. Furthermore all the reported studies made exclusive use of proliferation as the read out for T cell activation and did not address the possibility that antigen presentation by non-specialist APC could lead to a different cytokine production profile.

The aim of the experiments described here was to investigate the consequences of antigen presentation by human TFC expressing MHC class I1 molecules, following IFN-y treatment. From the results obtained, we can conclude that TFC are unable to initiate an immune response. In addition, recognition of peptide on TFC by B7- independent Tho T cell clones may favor IL-4 production. In T cell clones that were not able to proliferate to TFC, the overnight culture with TFC inhibited IL-2 production during the rechallenge with professional APC, but left the capacity of the clones to secrete IL-4 intact. The possible relevance of the results to autoantibody-mediated autoimmune diseases is discussed.

2 Materials and methods

2.1 Antigens

The synthetic influenza virus hemagglutinin (HA) peptides (HA307-319 and HA100-115) were synthesized by the Imperial Cancer Research Fund (ICRF) Peptide Unit and kindly provided by Dr. Hans Stauss.

2.2 Monoclonal antibodies

Monoclonal antibody (mAb) DAKO-CK1, LP34 was pur- chased from DAKO (High Waycombe, GB) and the anti- B7/BB1 (anti-B7-l) from Becton Dickinson (Cowley, GB). The following antibodies were used for staining as hybridoma supernatants: L243 (anti-HLA-DRa, Amer- ican Type Culture Collection, ATCC, Rockville, MD), TS2/9 (anti-human LFA-3, ATCC), 6.5B5 (anti-ICAM-1,

kindly provided by Dorian Haskard), and BU63 (anti-B7- 2, kindly provided by Peter Beverly).

The following purified antibodies were used for enrichment of CD4’ T cells, Leu19 (anti-CD56, Becton Dickin- son), and mouse anti-human Ig (Fab-specific, Sigma Poole, GB). The OKT8 (anti-human CD8, ATCC) and L243 antibodies were purified from culture supernatant on protein A-Sepharose beads by standard methods. Eluted antibody was dialyzed against three changes of PBS.

The COS transfectant secreting the fusion protein CTLA4- Ig, was kindly provided by Peter Lane (Basel Institute for Immunology, Basel, Switzerland) [33].

2.3 Cell lines

TFC were isolated from two individuals undergoing thyroi- dectomy; the human thyroid 99 (HT99) line was derived from a 39-year-old male and the 100 (HT100) line from a 39-year-old female, both with Graves’ disease. HT99 and HTlOO were selected on the basis of their HLA-DR type, DRBl*O101. TFC were isolated as previously described [3]. Briefly, the thyroid tissue was finely minced under ste- rile conditions and rinsed with CA2+- and Mg2+-free saline solution. The tissue was digested in a solution of collage- nase (1 mg/ml) (Sigma) and dispase (4 mg/ml) (Sigma) for 30-min periods at 37 “C. Thyroid follicles were collected from the 2nd-4th digests, washed in RPMI 1640 (Gibco, BRL, Life Technologies Ltd, Paisley, Scotland) with 10 % fetal calf serum (FCS), 2 mM L-glutamine (Gibco). 50 IU/ ml penicillin (Gibco), and 50 pg/ml streptomycin (Gibco) and cultured overnight at 37 “C in 10 ml fresh medium per 75 cm2 culture flask (Greiner Labortechnik Ltd., Dursley, GB). Sixteen hours later the TFC-adherent layer was washed extensively in RPMI and fresh medium was added. After 3 days the cells were harvested with 0.05 YO trypsin (Gibco) and stored in liquid nitrogen until required, at which point the TFC were thawed and cultured in RPMI 1640 with 10 % FCS, 2 mM L-glutamine, 50 IU/rnl penicillin, and 50 pg/ml streptomycin, either with or without IFN-y (500 U/ml), for 4 days to deplete further conven- tional APC before being used in assays. Confirmation that there were no contaminating cells in the TFC cultures was achieved by FACS analysis after staining with an anti- cytokeratin mAb (DAKO-CK1, LP34) as described below.

Epstein-Barr virus (EBV)-transformed lymphoblastoid B cell lines (B-LCL), from the 10th International Histocom- patibility Workshop, were cultured in RPMI 1640 tissue culture medium (ICN Flow, Thame, GB) supplemented with 10% FCS, 2 mM L-glutamine, 50 IU/ml penicillin, and 50 p,g/ml streptomycin in 25-cm2 flasks, and were regu- larly passaged.

The M1 human fibroblast cell line [34] was co-transfected with the DRA (kindly provided by Dan Denney, Stanford, CA) and DRlB (kindly provided by Eric Long, NIH, Bethesda, MD) cDNA clones as previously described [16, 351. Transfection with B7-1 was performed as previously described [36]. In brief, Ml-DRl-B7 and DAP.3-B7 trans- fectants were produced by co-transfection with a cDNA clone encoding human B7-1 in the pcExV-3 vector (a kind

64 G. Lombardi et at. Eur. J. Immunol. 1997.27: 62-71

gift of Mark Jenkins, University of Minnesota, Minneapo- lis, MN) with pHyg which contains a hygromycin B (Sigma) resistance gene at ratios of 100 : 1, 50 : 1 and 20 : 1. After transfection cells were selected in medium containing 200 pg/ml hygromycin B and clones with high levels of B7.1 expression were selected following flow cytometric analysis using a Coulter Excel analyser (Coulter Electron- ics, Luton, GB). The fibroblast lines were maintained in DMEM tissue culture medium (ICN Flow, Thame, GB) supplemented wit 10 YO FCS, 0.2 YO sodium bicarbonate, 2 mM L-glutamine, 50 IU/ml penicillin and 50 pg/ml streptomycin, and 200 pg/ml G418 (Sigma) to maintain expression of the transfected genes. Cells were grown in 25-cm2 flasks, and passaged, following trypsinisation, once weekly.

The IL-Zdependent murine T cell line CTLL-2 (European Collection of Animal Cell Cultures, Salisbury, GB) was cultured in RPMI 1640 medium, supplemented with 2 mM L-glutamine, 50 IU/ml penicillin, 50 pg/ml streptomycin, 10 U/ml of human recombinant IL-2 (rIL-2, Boehringer Mannheim, Germany) and 10% FCS. The cells were cultured in 25-cm' flasks and were subcultured every 3 days. Prior to use in a proliferation assay, the CTLL-2 cells were washed twice and cultured overnight in normal culture medium, but without added rIL-2.

The IL-Cdependent murine cell line CT. h4S [37], transfected with a cDNA encoding the human IL-4 receptor (CD124; generous gift of W. Paul, NIH, Bethesda, MD) was cultured in RPMI 1640 medium, supplemented with sodium bicarbonate (0.24 % final concentration), 2 mM L- glutamine, 50 IU/ml penicillin, 50 pglml streptomycin, 1 mM sodium pyruvate, 50 mM 2-ME, human rIL-4 (100 U/ml, Genzyme, West Malling, GB) and 10% FCS. The cells were cultured in 25-cm2 flasks and were subcultured every 3 days. Prior to use in a proliferation assay, the CT.h4S cells were washed twice and kept on ice.

2.4 Enrichment of CD4' T cells

Peripheral blood mononuclear cells (PBMC) were obtained by Ficoll-Hypaque (Pharmacia, Biotech, St. Alban's, GB) centrifugation of heparinized blood, washed twice and resuspended in RPMI 1640 medium supplemented with 10% FCS, 2 mM L-glutamine, 50 IU/ml penicillin, 50 pg/ml streptomycin. The cell preparation was then depleted of adherent cells by two 45-min rounds of adherence to plastic on tissue culture dishes at 37 "C. The nonadherent cells were subsequently eluted and incubated with a cocktail of purified mAb, OKT8, Leul9, L243 and mouse anti-human Ig (Fab-specific) at saturating concentrations for 15 rnin at 4°C. The cells were washed to remove excess antibodies and further enriched by magnetic immunodepletion using sheep anti-mouse Ig-coated magnetic beads (Dynal Ltd, Merseyside, GB) as specified by manufacturer's instructions. In brief, antibody-treated cells were resuspended in cold tissue culture medium at lo7 cells/ml. Magnetic beads were subsequently added to the cell suspension and incubated at 4°C for 30 rnin followed by removal of the bead/mAb-coated B cells, monocytes and NK cells with a magnet. This step was repeated once more and the cells were resuspended in culture medium ready for the proliferation assay.

2.5 T cell clones

Clones NFl, NF2, NF3 and NF4, specific for HA 307-319 and restricted by DRB1*0101, were derived from PBMC isolated from a DRB1*0101 individual as described previously [14, 161. Clones HC6 (specific for HA307-319 and restricted by DRB1*0101) and HC3 (specific for HA100-- 115 and restricted by DRB1*0101) were generated from a DRB1*0101, DRB1*0403 individual by stimulating PBMC with purified influenza HA (5 pg/ml). The clones were maintained in culture by weekly stimulation with autologous PBMC, HA peptides and rIL-2, in RPMI 1640 medium supplemented with 10% human serum, 2 mM L- glutamine, 50 IU/ml penicillin, and 50 pg/ml streptomycin. Accessory cell-free preparations of T cells were pre- pared as described before [14]. Briefly, T cells were reseeded into new 24-well tissue culture plates on days 3 and 6 after restimulation in order to remove adherent cells. For use in the experiments the cells were purified by isolation on Ficoll-Paque 7 days after restimulation and washed five times by slow speed centrifigation (210 X g, 5 min) before use.

2.6 Flow microfluorimetric analysis

For flow microfluorimetric analysis, 5 x 10' T cells or TFC were incubated with the indicated mAb at 4°C for 30 min. After wasching twice in phosphate-buffered saline with 2.5 YO FCS, the cells were incubated for a further 30 min at 4°C with 100 p1 of 1:50 dilution of fluorescinated sheep anti-mouse Ig (Amersham Int., Amersham, GB). After two additional washes, stained cells were analyzed using the EPICS Profile Flow Cytometer (Coulter Electronics, Luton, GB).

For measuring cytokeratin expression, TFC (10') were fixed with 1 YO paraformaldehyde in PBS for 30 rnin at 4"C, washed and permeabilized with 0.1 YO saponin in PBS for 30 rnin at 4°C. After two further washes in PBS with 2.5 % FCS, the cells were stained with the indicated mAb following the above procedure.

2.7 T cell proliferation assays

T cell clones ( lo4 cells/well) were cultured in the presence of B-LCL, treated with 120 Gy X-irradiation, or TFC, treated with 30 Gy X-irradiation, or mitomycin-C treated M1 lines, in flat-bottom microtiter plates, in a total volume of 200 pl. For antigen-specific responses, the APC were pre-pulsed for 2-4 h, and then washed to remove any soluble peptide. In some experiments DAP.3-B7 cells were added to the cultures. Wells were pulsed with 1 pCi of ["HIdThd (Amersham International), after 48 h and the cultures harvested onto glass fiber filters 18 h later. Prolif- eration was measured as [3H] dThd incorporation by liquid scintillation spectroscopy. The results are expressed as the mean of triplicate cultures. SEM were routinely < 10 % .

For primary MLR assays, purified human CD4' T cells ( lo5 cells/well) were cultured with different numbers of irradiated stimulator cells, as indicated in figure legends in round-bottom plates (ICN Flow) for 6 days. Wells were pulsed with 1 pCi [3H] dThd 20 h before the end of the cul-

Eur. J. Immunol. 1997.27: 62-71 Antigen presentation by thyroid follicular cells favors a Th2 response 65

ture. Proliferation was measured and expressed as described above.

2.8 Two-stage cultures for tolerance induction

T cell clones were plated out in 24-well plates in the presence of peptide-prepulsed irradiated TFC in a total volume of 500 p1. In addition T cell clones were also cultured alone or in the presence of HA peptides (10 pg/ml), as previously shown [14]. After overnight incubation, the cells were harvested, washed extensively, and used in proliferation assays as described above. To “rest” the T cells, they were left for 3 days in the presence of a suboptimal dose of rIL-2.

2.9 CTLL and CT.h4S assays

Culture supernatants were collected and transferred into two sets of 96-well round-bottom microtiter plates as triplicate cultures for measurement of IL-4 and IL-2. The plates were stored at - 20 “C until used. In each well either 3 x lo3 CTLL or 5 X lo3 CT.h4S cells were added. In each experiment a standard titration for rIL-2 and/or rIL-4 was included. The CT.h4S cells were highly sensitive and specific. Proliferation was observed at 0.1 U/ml of rIL-4; no response was measured to rIL-2 below 50 U/ml. After respectively 8 and 24 h of incubation, wells were pulsed with 1 pCi of [3H] dThd (Amersham International). After 18 h the cultures were harvested onto glass fiber filters and [3H] dThd incorporation was measured by liquid scintillation spectroscopy, as described above.

3 Results

3.1 Characterization of TFC

TFC were obtained from two patients with Graves’ disease. TFC were firstly stained for cytokeratin, a marker for epithelial cells, using the mAb DAKO-CK1, LP34, to test their purity. Fig. l a shows that all the cells stained positive with this antibody, confirming the epithelial lineage of the cells. TFC express MHC class I constitutively and can be induced to express MHC class I1 molecules following IFN- y treatment. TFC were cultured with IFN-y for 4 days and then tested for expression of some of the molecules involved in T cell stimulation. The results shown in Fig. 1 were derived from TFC HT100, one of the two TFC lines induced to express HLA-DRB 1*0101. Similar results were obtained with other TFC cultures. In Fig. l b is shown the comparison of class I1 staining between TFC cultured either in the presence or absence of IFN-y (1000 U/ml) for 4 days. The level of expression of MHC class I1 on TFC treated with IFN-y was approximately threefold lower than that on a B-LCL (Fig. If). In addition, we tested the effect of IFN-y treatment on the expression of accessory molecules that are known to contribute to T cell reactivity. As shown in Fig. l c addition of IFNy up-regulated ICAM-1; however, LFA-3 expression was unaffected (Fig. Id) [38]. The level of ICAM-1 expression on the induced TFC was comparable to that on B-LCL (Fig. lf). No detectable expression of the B7 family of molecules was seen using either anti-B7/BBl mAb (Becton Dickin- son), specific for B7-1 (Fig. le) , or the CTLA-4-Ig fusion protein, which recognizes all B7 isoforms [33] (data not shown). The expression of MHC class I1 and of the accessory molecules, ICAM-1, LFA-3 and B7-1, on the HLA- DRB l”0101-positive B-LCL, used in the experiments

Relative Fluorescence Intensity

Figure I. TFC express MHC-class I1 but not B7-1 molecules after IFN-y treatment. Panel a shows the staining of TFC with anti-cytokeratin (thick line) and with an irrelevant mAb (thin line), after permeabilization with saponin as described in Sect. 2.6. From panel b to panel e TFC were cultured in the absence (thick lines) or in the presence (thin lines) of IFN-y (lo’ U/ml) for 4 days. The TFC were then stained with either the second mAb alone (solid lines) or with the following mAb (dotted lines): anti-DR (panel b), anti- ICAM-1 (panel c), anti-LFA-3 (panel d), anti-B7-l (panel e). In panel f the staining of a B-LCL is shown for comparison with the following mAb second mAb alone (thin solid line) (fluorescence intensity so low that negative cells compressed against the vertical axis); anti-DR (thin dotted line); anti- LFA-3 (thick solid line); anti-B7-l (thick dotted line): and anti-ICAM-1 (thin dashed line).

66 G. Lombardi et al. Eur. J . Immunol. 1997.27: 62-71

Relative Fluorescence Intensity

Figure 2. TFC lack cxpression of B7-2. Panels a and b show the staining ofTFC with anti-B7-l and anti-B7-2 mAb respectively at 48 h (thick lines) and 96 h (thin lines) after IFN-y treatment. TheTFC were stained with the second mAb alone (solid lines) or with the relevant mAb (dotted lines). In panel c the staining of a B-LCL with the 2nd mAb alone (thin solid line), anti-B7-1 (thin dotted line) and anti-B7-2 (thick solid line).

described below, are shown in Fig. If. To investigate further the expression of B7-1 and B7-2 molecules, a time- course experiment was performed using anti-B7lBB 1 and BU63 mAb. In Fig. 2 are shown the results ofTFC staining after IFN-y treatment for 48 and 96 h. Identical results were obtained at 0,24 and 72 h (data not shown). No detectable expression of B7-1 (Fig. 2a) and B7-2 (Fig. 2b) was seen at any point studied. For comparison the expression of B7-1 and B7-2 molecules on a B-LCL are shown in Fig. 2c.

(HT99 and HT100) or to B-LCL, prepulsed with different does of peptides, was measured. The results are shown in Fig. 4 and suggest that the T cell populations can be divided into two groups based on their capacity to respond to peptide presented by pre-pulsed TFC. Clones NFI, NF2 and HC3 showed no response, even at high peptide concentrations. The other three T cell clones showed some

3.2 TFC are unable to stimulate a primary alloresponse

During an autoimmune thyroid disease, TFC express MHC class I1 molecules and acquire the capacity to present autoantigen to T cells infiltrating the thyroid gland. We have investigated the capacity of IFN-y-treated TFC to induce a primary T cell response by measuring the proliferation of human CD4' T cells from three DRl-negative responders to IFN-y-induced TFC expressing HLA- DRB1*0101. CD4' T cells were unable to proliferate to TFC in an MLR. On the contrary, they were perfectly able to respond to B-LCL and to PBMC both expressing the same HLA-DR1 type. The results from one of these experiments are shown in Fig. 3a. To test the possibility that the lack of response to TFC was due to the lack of B7 expression, CD4' T cells were co-cultured with TFC in the presence of a murine fibroblast cell line transfected with a cDNA clone encoding human B7-1 (DAP.3-B7). As shown in Fig. 3b the addition of DAP.3-B7 reconstituted the response of CD4' Tcells toTFC. The involvement of B7 in the trans-co-stimulation provided by DAP.3-B7 was further demonstrated by the fact that the response of CD4' T cells was inhibited by the addition of CTL4-Ig (Fig. 3b). These results support the idea that the failure of TFC to stimulate a primary alloresponse is due to a lack of co- stimulation.

3.3 TFC are unable to stimulate '(B7-dependent" T cell clones

The antigen presenting function of TFC was further analyzed by testing their capacity to stimulate the proliferation of established human T cell clones and a polyclonal T cell line restricted by DRB1*0101 and specific for a defined peptide of HA. T cells were purified, as described in Sect. 2.4, and their proliferation to the two TFC lines

'? s o X 2.5 10 20 E stimulator cell numher ( x l ( k 3 3 a

Figure 3. TFC are unable to stimulate a primary MLR. CD4' T cells were purified as described in Sect. 2.4. CD4'Tcells (10' cel- Is/well) were cultured with different numbers of HLA-DR1- expressing B-LCL (open squares), PBMC (open triangles) and TFC pretreated (closed circles) or not (open circles) with IFN-y (10" Ulml). The results are shown in Fig. 2 a. In panel b, CD4' T cells were cultured either with TFC (filled bars) or with DAP.3 cclls expressing B7 (lightly shaded bar). In addition, CD4' Tcclls were co-cultured with TFC and DAP.3-B7 cells (darkly shaded bar) in the presence of either anti-DR mAb (hatched bar) or CTLACIg (open bar). The plates were incubated for 5 days and ['HIdThd was added for the last 18 h. Proliferation is shown as mean cpm of triplicatc cultures, corrected for background prolif- cration of stimulator cells alone (Acpm). SEM were routinely < 10%.


30 -

20 -

10 -

0 -

50 - 40 - 30 - 20 - 1 0 -

NF 1

’()- HC3

20 -

I 0 - - . 1 1 10

0- 1 HC6

Figure 4. TFC can stimulate antigen-specific T cell clones. IFN-y-treated TFC. HTlOO (solid circles) and HT99 (solid squares), and B-LCL (open circles) were pulsed overnight with different doses of peptides. HA-specific and DRBl*OlOl-restricted Tcell clones (NFI, NF2. NF3, NF4 and HC6 specific for HA307-319 and HC3 specific for HAlOO-llS), were cultured at 104/well with either 2 X 10‘/well prepulsed TFC or B-LCL. The plates were incub-

A

. I 1 10

15 -

10 -

5 -

0 -

. I 1 10

proliferation to peptide presented by TFC, although at a substantially lower level than to peptide-prepulsed B-LCL used as APC.

To determine whether the response patterns of the clones correlated with their dependence on B7 expression by the APC, their response was measured to a human fibroblast cell line (Ml-DR1) with or without co-expression of human B7-1. The results demonstrated that the same T cell clones that failed to respond to TFC were also unable to respond to B7-negative M1-DR1 cells. The co-expression of B7-1 on MI-DR1 completely restored their proliferation to levels similar to those induced by a B-LCL. In contrast, the T cell clones that were capable of recognizing TFC were also able to proliferate in response to B7- negative M1-DR1, and the presence of B7 on this line (Ml-DRl-B7) did not significantly augment the T cell resonse (R. Hargreaves, manuscript in preparation). The results from two T cell clones, one unable to respond to TFC (HC3) and the other perfectly capable of responding (NF4) are shown in Table 1.

The second approach was to try to restore the proliferative response to TFC by providing trans-co-stimulation. For

this purpose, T cell clones were cultured with pre-pulsed TFC (HT100) in the presence of DAP.3-B7. As shown in Fig. 5 the presence of DAP.3-B7 reconstituted the response to TFC cells by the B7-dependent T cell clones NF1 (panel a) and HC3 (panel b). In addition, the presence of trans-co-stimulation in the culture of B7- independent T cells made no consistent difference (data not shown). These data suggest that for T cell clones that can be defined as “B7-dependent” the TFC lines cannot provide the necessary co-stimulation for T cell activation.

3.4 Presentation by TFC induced a predominant production of IL-4 in B7-independent T cell clones

To characterize further the T cell response to TFC by B7- independent T cell clones, we analyzed and compared the lymphokine profiles of T cell clones, NF4 and HC6, in response to peptide presented by eitherTFC or B-LCL. As shown in Fig. 6a and c when B-LCL were used as APC high levels of IL-2 were produced in the presence of similar amounts of 1L-4. In contrast, when peptide was presented by TFC, IL-4 was the predominant lymphokine produced, Fig. 6b and d. Similar results, were obtained

68 G. Lombardi et al.

40 - 30 - 20 -

10 - 0 -

Eur. J. Immunol. 1997.27: 62-71

g 60- 0

d

40-

20 -

0 a

Table 1. B7-dependence of T cell clones"'

Cell lines T cell clones (A cpm X

HC3 NF4 MI-DRI 519 5,177 MI-DRI-B7 5,534 3,631

a) Tcell clones (104/weIl) were co-cultured with 3 X lo4 mitomycin C-treated M1 cells. After 48 h cultures were pulsed with [3H]dThd and harvested 18 h later. Proliferation is shown as mean cpm of triplicate cultures, corrected for background proliferation of stimulator alone ( A cpm).

when T cells and TFC were cultured in the presence of an anti-IL2 receptor mAb, to block any consumption of IL-2 by T cells (data not shown). In addition, the production of IFN-y and IL-4 was also measured by ELISA. The same pattern of cytokine production as those reported in Fig. 6 were observed (data not shown).

3.5 Peptide presentation by TFC inhibited IL-2, but not IL-4, production by B7-dependent T cells

It has been shown before that specific recognition of peptide/MHC class I1 complexes in the absence of co- stimulation can lead to a state of non-responsiveness in IL- 2-secreting T cells. We tested this possibility using T cell clones that were unable to proliferate in response to peptide presented by TFC. For this purpose T cell clones, NF1 or HC3, were purified and then cultured overnight under various conditions including TFC prepulsed with the specific peptide. The precultured Tcells were then washed and tested for their capacity to respond to the specific peptide presented by professional APC, namely B-LCL. After 24 h the supernatants were collected and tested for the presence of IL-2. IL-2 production by clone HC3, cultured overnight with TFC (HT99) with the relevant peptide

1 h)

Figure 5. The lack of response toTFC is due to the absence of B7 expression. Tcell clones, NFI (panel a) and HC3 (panel b), both restricted by HLA-DRBl*OlOl and specific for HA307-319 and HA100-115, respectively, were cultured with peptide-prepulsed TFC (HTIOO) in the absence (solid bars) or in the presence of DAP.3-B7 at a ratio of 1 : 1 (hatched bars) or 1 : 2 (shaded bars). Proliferation was assessed and reported as described in the legend to Fig. 3.

(HA100-115) or the irrelevant peptide (HA307-319), and then challenged with peptide-pulsed B-LCL, is shown in Fig. 7. Presentation of the specific peptide by TFC inhi-

peptide concentration ( p @mi)

Figure 6. TFC presentation induced a predominant production of IL-4 in B7-independent T cell clones. The production of IL-2 and IL-4 by two T cell clones specific for HA-307-319 and restricted by HLA-DRBI*OlOl, NF4 (panels a and b) and HC6 (panels c and d) in response to either B-LCL (panels a and c) or to TFC (panels b and d) is shown. Tcells were cultured with either B-LCL or TFC pre-pulsed with different doses of peptide. After 24 h supernatants were collected and IL-2 (open squares) or IL-4 (open triangles) were tested using CTLL or CT.h4S, respectively, as described in Sect. 2.9. Proliferation was assessed and reported as described in the legend to Fig. 3.


5 -

4 -

3 -

2 -

1 -

0 - IG .01 . 1 1 10

6 -

4 -

2 -

0-

”1

0: 1 i lb

peptide concentration ( p dml)

Figure 7. Inhibition of IL-2 production by Tcells incubated overnight with peptide-prepulsed IFN-y-treated TFC. The production of IL- 2 from overnight preincubated HC3 (HA100-115-specific and DRBl*OlOl-restricted) to peptide-prepulsed B-LCL is shown in panel a. T cells (lO‘/ml) were cultured under different conditions: with 10 pg/ml of the relevant peptide (filled circles), with IFN-y-treated TFC (HT99) prepulsed with 10 pg/ml of either the relevant (filled squares), or the irrelevant peptide (open squares), and in medium alone (open circles) in 24-well plates. After overnight incubation the T cells were collected, washed and cultured with peptide-prepulsed B- LCL. After 24 h the supernatants were collected and the IL-2 production was tested using the CTLL, as described in Sect. 2.9. In panels b and c clone NF1 (HA307-319-specific and DR1-restricted) was cultured overnight with TFC (HT100) prepulsed with either the relevant (solid squares) or with the irrelevant (open squares) peptide. After the overnight incubation the cells were collected, washed and either cultured immediately (panel b) or after 3 days rest (panel c) with peptide-prepulsed B-LCL. The supernatants were harvested and tested as described. Proliferation was assessed and reported as described in the legend to Fig. 3.

bited the release of IL-2 by HC3 in response to B-LCL. In addition, T cells were cultured overnight either with medium alone or with peptide in the absence of any added APC. As shown previously the production of IL-2 by T cells in response to subsequent stimulation with B-LCL was inhibited as a consequence of T : T presentation [14, 161. In Fig. 7b the IL-2 production by clone NF1 in response to peptide-prepulsed B-LCL after the overnight culture step is shown. The release of IL-2 in response to B- LCL was reduced only when NF1 was cultured overnight with relevant peptide-prepulsed TFC (HT100). NFl and HC3 were also left, after the overnight incubation step, to “rest” for 3 days and then challenged with peptide- prepulsed B-LCL. As shown in Fig. 7c, inhibition of IL-2 production persisted after the rest period.

To address the possibility that antigen presentation by TFC had a differential effect on the production of IL-2 and IL- 4, both cytokines were measured in response to re-

challenge with peptide-pulsed B-LCL. As shown in Fig. 8, for clone NF1, although IL-2 production was abolished by previous culture with TFC pulsed with the appropriate peptide, the T cells retained the ability to synthesize IL-4. These data further support the idea that peptide presentation by TFC induces tolerance in T helper cells by inhibit- ing IL-2 production and favoring a T helper switch towards a Th2 phenotype.

4 Discussion

In this study IFN-y-treated, DR-expressing TFC were unable to stimulate a primary alloresponse, and failed to induce proliferation by three of six antigen-specific T cell clonesAines tested. Furthermore, antigen presentation by the TFC to these three T cell clones inhibited IL-2 production when they were rechallenged with professional APC.

peptide concentration ( p dml)

Figure 8. Production of IL-4 in the absence of IL-2 following peptide presentation by TFC. The production of IL-2 and IL-4 from overnight preincubated NFI (HA307-319 specific and DRBl*OlOl-restricted) in response to peptide-prepulsed B-LCL is shown in panels a and b, respectively. Tcells (106/ml) were cultured with IFN-y-treated TFC (HT100) prepulsed with 10 pg/ml of either the relevant (filled circles), or the irrelevant peptide (open circles) in 24-well plates. After overnight incubation the T cells were collected, washed and cultured with peptide-prepulsed B-LCL. After 24 h the supernatants were collected and the IL-2 or IL-4 production was tested using either CTLL or CT.h4S, as described in Sect. 2.9. Proliferation was assessed and reported as described in the legend to Fig. 3.

70 G. Lombardi et al. Eur. J. Immunol. 1997.27: 62-71

The results also suggest that recognition of peptide on TFC can bias T cells towards a Th2 phenotype.

The provision of trans-co-stimulation [39-411, by addition of a transfected mouse cell line expressing human B7-1, overcame the deficiency in accessory cell function of the TFC. The significance of these findings relates to their implications for the functional effects of MHC class I1 molecule expression on epithelial cells in vivo. It has been proposed that this contributes to the pathogenesis of autoimmune inflammation, either by initiating or ampliyfing T cell responses to self proteins [16] The results described here support a modified form of this hypothesis, in that they suggest that antigen presentation by tissue parenchymal cells may tend to silence the capacity of tissue- infiltrating T cells to make IL-2, while leaving their capacity to secrete IL-4 intact. This could in turn provide help for autoantibody-producing B cells such as those involved in a condition like Graves’ disease. This hypothesis is supported by the work of Mullins et al. [42]. They showed that T cell clones generated from patient with Graves’ disease, specific for the thyrotropin receptor (TSHR), following stimulation with soluble OKT3 and phorbol ester were producing predominantly IL-4 [42].

The silencing of IL-2 production by CD4’ T cells as the result of antigen recognition on cells that lack specialized antigen-presenting function has been reported in murine and human systems with a variety of cell types that have been induced to express MHC class I1 molecules, either by gene transfection or by induction with IFN-y. These have included keratinocytes [6,7], fl cells of pancreatic islets [8], myoblasts [43], and human T cells [9-121. In contrast to these results, it has been reported that human TFC were fully able to stimulate autoreactive Tcells isolated from the thyroid glands of patients with Graves’ diesease [17-201. This discrepancy may reflect differences in the purity of the responder T cells used in the proliferation assays or in the way the T cells were grown in vitro. The demonstration here that the addition of a third-party cell expressing B7-1 (DAP.3-B7) restored the T cell clones’ proliferative response to TFC highlights the possible effects of any contaminating APC in the responder cell population.

It was notable that three of sixTcell clones did proliferate, albeit weakly, to antigen presented by the TFC. Prolifera- tion by these T cells was shown not to require B7 expression by the APC. This raises the important question as to whether this B7-independent phenotype exists in vivo in T cells that have recently been activated. If so, they may well be restimulated by encountering epithelial APC. It may be more likely that this phenotype arises as a result of prolonged maintenance in in vitro culture. Related to this issue, Dayan et al. [44] observed that T cells specific for thyroid peroxidase, that were isolated from the thyroid of a patient with Graves’ disease, were resistant to the induction of “anergy” in vitro. This raised the interesting possibility that the T cells responsible for autoimmune reactions in vivo are characterized by possessing low stringency activation requirements. However, it is difficult to be sure that the in vitro behavior of the thyroid-derived clones mirrors their in vivo phenotype [20, 441.

The in vitro finding reported here may parallel observa-

tion, in which the recipient’s immune system may be exposed to MHC class 11-expressing epithelial cells, in the absence of any bone marrow-derived specialized APC of donor origin. It is well established that the prolonged residence of kidney and heart allografts, in the absence of immunosuppression, induces a state of profound, donor- specific non-responsiveness [45]. We have recently observed related findings in man, in that the frequency of donor-reactive, IL-2-secreting T cells can be seen to fall in patients who have harbored renal allografts for extended periods [46]. This suggests that the in vivo consequence of T cell recognition of antigen on epithelial cells is to silence IL-2 production.

There is abundant evidence that self-reactive T cells exist in the normal T cell repertoire [47-491. Given that this is the case, it is highly possible that self-reactive T cells will be activated as a by-product of priming against an invading microbe. The data reported here support the suggestion that the induction of MHC class I1 expression on tissue parenchymal cells serves to silence and/or switch towards a Th2 phenotype such potentially deleterious self-reactive T cells. Given that the parenchymal cells are continuously exposed to high local concentrations of tissue-specific antigens, antigen presentation by these cells is likely to be biased in favor of self antigens and may provide a mecha- nism for selectively inactivating pro-inflammatory autoreactive cells. The data presented here further suggest that such antigen presentation may favor the production of T cell help for an antibody-mediated response. The relevance of this to the pathogenesis of tissue-specific autoimmune diseases will require further in vivo investigation.

Giovanna Lombardi was supported by the National Kidney Research Fund and Kate Arnold by the Sir Jules Thorn Trust. Julia Uren and Roseanna Hargreaves were supported by die Medical Research Council, Federica Marelli- Berg by Italfarrnaco and Nes- rina Irnarni by Action Research. This work was in part supported by the Medical Research Council and the Wellcome Trust.

5 References

1 Linsley, P., Brady, W., Grosmaire, L., Aruffo, A., Damlc, N.

2 Schwartz, R., Cell 1992 71: 1065. 3 Jenkins, M., Pardoll, D., Mizuguchi, J.. Quill, H. and

4 Quill, H. and Schwartz, R., J . Immunol. 1987. 138: 3704. 5 Jenkins, M. and Schwartz, R., J. Exp. Med. 1987. 165: 302. 6 Gaspari, A., Jenkins, M. and Katz, S . . J . Immunol. 1988. 141:

2216. 7 Bal, V., Mcindoe, A., Denton, G, Hudson, D., Lombardi, G.,

Lamb, J. and Lechler, R., Eur. J . Immunol. 1990.20: 1893. 8 Markman, J., Lo, D.. Naji, A, , Palmiter, R., Brinster. R. and

Herber-Katz, E., Nature 1988. 336: 476. 9 Lamb, J.. Skidmore, B., Green, N., Chiller, J. and Feldrnan,

M., J . Exp. Med. 1983. 157: 1434. 10 Sidhu, S. , Deacock, S. , Bal, V., Batchelor, J. R.. Lombardi,

G. and Lechler, R. I . , J . Exp. Med. 1992. 176: 875. 11 La Salle, J., Tolentino, P., Freeman, G., Nadler, L. and Haf-

ler, D., J . Exp. Med. 1992. 176: 177. 12 Lombardi, G., Sidhu, S., Dodi,T., Batchelor, R. and Lechler,

R., Eur. J . Immunol. 1994. 24: 523. 13 Hanafusa, Y.. Puiol-Borrell, R.. Chiovato, L., Russel, R.,

and Ledbetter, J., J. Exp. Med. 1991. 173: 721.

Schwartz, R., Immunol. Rev. 1987. 95: 113.

I .

tions made in experimental models of-oigan transplanta- Doniach, D. and Bottazzo, G. F., Lancet 1983. i i : 1110.


14 Todd, I., Pujol-Borrell, R. , Bottazzo, G. E and Feldmann, M., Clin. Exp. Immunol. 1985. 61: 265.

15 Weetman, A,, Volkman, D., Burman, K., Gerrard, T. and Fauci, A., J . Clin. Endocrinol. Metab. 1985. 61: 817.

16 Bottazzo, G., Pujol-Borrell, R., Hanafusa, T. and Feldmann, M.. Lance 1983. ii: 1115.

17 Londei, M., Bottazzo, G. and Feldmann, M., Science 1985. 228: 85.

18 Weetman, A., Volkman, D. , Burman, K., Marfolick, J., Pet- rick, P., Weintraub, B. and Fauci, A., Clin. Immunol. Immu- noputhol. 1986. 3Y: 131.

19 Iwatani, U., Gerstein, H., Iitaka, M., Row, V. V. and Volpe, R., J. Clin. Endocrinol. Metub. 1986. 63: 695.

20 Dayan, C., Londei, M., Corcoran, A., Grubeck-Loebenstein, B., James, R. , Rapoport, B. and Feldman, M., Proc. Nutl. Acad. Sci. USA 1991. 88: 7415.

21 Trieb, K., Sztankay, Hermann, M., Gratzel, R., Szabo, J., Jin- dall, S. and Grubeck-Loebenstein, G., J . Clin. Endocrinol. Metab. 1993. 77: 528.

22 Londei, M., Lamb, J., Bottazzo, G. and Feldmann, M., Nature 1984.312: 639.

23 Davies, T., J . Clin. Endocrinol. Metub. 1985. 61: 418. 24 Kirnura, H. and Davies, T., Clin. Immun. Immunoputhol.

25 Stein, M. and Staedecker, M., J. Immunol. 1987. 139: 1786. 26 Ebner, S., Stein, M. E., Minami, M., Dorf, M. and Stae-

decker, M., Cell. Immunol. 1987. 104: 154. 27 Grubeck-Loebenstein, B., Londei, M., Greenall, C., Pirich,

K., Kassal, H., Waldhausl, W. and Feldmann, M., J . Clin. Invest. 1988. 81: 1608.

28 Minami, M., Ebner, S., Staedecker, M. and Dorf, M., J . Immunol. 1987. 138: 393.

29 Czirjak, L., Danki, K., Gaulton, G. and Staedecker, M., Eur. J . Immunol. 1990. 20: 2597.

30 Tandon, N. and Weetman, A., Clin. Endocrinol. 1992.37: 274. 31 Eguchi, K., Otsubo, T., Kawabe, Y., Shimomura, C., Ueiki,

Y.. Nakao, H., Tezuka, H., Matsunga, M., Fukuda, T., Ishi- kawa, N . , Ito, K. and Nagataki, S., Clin Exp. Irnmunol. 1988. 72: 84.

1991. 58: 195.

32 Kabel, P., Voorbij, J., de Haan, M., Van der Gaag, R. and Drexhage, H., J . Clin. Endocrinol. Metah. 1988. 66: 199.

33 Lane, P., Gerhard, W., Hubele, S. A. L. and McConnel. F., J. Immunol. 1993. 80: 56.

34 Royer-Pokora, B., Peterson, W. and Haseltine, W., Exp. Cell. Res. 1984. 1.51: 408.

35 Dodi, A., Brett, S., Nordeng, T., Sidhu, S., Batchelor, J., Lornbardi, G., Bakke, 0. and Lechler, R. , Eur. J. Immunol. 1994. 24: 1632.

36 Greenlaw, R., Robinson, P., Lombardi, G. , Sidhu. S. and Lechler, R., Int. Immunol. 1992. 4: 6473.

37 Hu-Li, J . , Ohara, J., Watson, C., Tsang, W. and Paul, W. E., J. Immunol. 1989. 142: 800.

38 Weetman, A., Freeman, M., Borysiewicz, L. and Makgoba, M., Eur. J . Immunol. 1990.20: 271.

39 Jenkins, M., Ashwell, J. and Schwartz, R., J . Irnmunol. 1988. 140: 3324.

40 Otten, G. and Germain, R. N., Science 1991. 2.51: 1228. 41 Ding, L. and Shevach, E. M., Eur. J. Immunol. 1994.24: 859. 42 Mullins, R. J., Cohen, S. B., Webb, L. M., Chernajovsky, Y.,

Dayan, C. M., Londei, M. and Feldmann, M., J. Clin. Invest. 1995. I : 30.

43 Warrens, A., Zhang, J . , Sidhu, S., Watt, D., Lombardi, G. , Sewry, C. and Lechler, R., Int. Immunol. 1993. 6: 847.

44 Dayan, C., Chu, N., Londei, M., Rapaport, B. and Fcld- mann, M., J. Immunol. 1993. 151: 1606.

45 Hall, B., Jelbart, M., Gurley, K. and Dorsch, S. E., J. Exp. Med. 1985. 162: 1683.

46 Mason, P. D. , Robinson, C. M. and Lechler, R. I., Kidney Int. (In press). .

47 Liblau, R., Tournier-Lasserve, E., Maciazek, J. , Dumas, G., Siffert, O., Hashim, G. and Bach, M., Eur. J . Immunol. 1991. 21: 1391.

48 Tandom, N., Freeman, M. and Weetrnan, A, , Clin. Exp. Immunol. 1992. 89: 468.

49 Sun, J., Harcourt, G., Wang, Z., Hawke, S., Olsson, T., Fred- rikson, S. and Link, H. , Eur. J . Immunol. 1992. 22: 1553.

Documents

Antigen presentation by interferon-γ-treated thyroid follicular cells inhibits interleukin-2 (IL-2) and supports IL-4 production by B7-dependent human T cells