CELLULAR

IMMUNOLOGY

144,80-94 (1992)

Inhibition of in Vitro T Cell Activation by Cornea1 Endothelial Cells’ WESLEYF.OBRITSCH,HIDETOSHI KAWASHIMA,ANTHONY EVANGELISTA, JEFFREYM. KETCHAM,EDWARDJ.HOLLAND,ANDDALES.GREGERSON Department of Ophthalmology, University of Minnesota, Minneapolis,

hfinnesota 554.55

Received August 5. 1991; acceptedJune 13, 1992 Cells and tissuesof the anterior uvea and aqueoushumor expressactivities which inhibit immune responses.These activities include soluble factors such as TGF-fl and uncharacterized cell surface interactions. Relatively little is known regarding the immunologic activities of comeal endothelium, despite its potentially important role in contributing to the immune privilege of the anterior chamber and the high successrate of cornea1transplantation. In this report, in vitro studies of cultured rat cornea1endothelial (CE) cells were done using S-antigen-specificLEW rat T cell lines, or S-antigen-specific T cell hybridomas, to examine the immunologic capabilities of CE cells. Monolayers of LEW rat CE cells were unable to present antigen or a mitogen, Con A, to T cell lines or hybridomas asassessedby the lack ofa proliferative responseor IL-2 secretion.Furthermore, the CE cells exerted a potent inhibitory effect when added to in vitro proliferation assaysof T cell lines stimulated with antigen or Con A. When T cells were preactivated on conventional antigen presenting cells and then transferred to wells containing CE cells, their proliferation was not inhibited. Although CE cells inhibited activation of T cell lines and hybridomas, they did not inhibit the growth of T cell hybridomas or CTLL cells, nor did the CE cells adversely affect the viability of resting T cells cultured on CE monolayers. The inhibitory effect was reversible as preincubation of T cells on CE cells for up to 6 days followed by washesrestored T cell responsivenesswhen assayedon splenocytes. The inability to stimulate proliferative responseswas not affectedby preincubation ofthe CE cellswith lymphokines which increaseMHC antigen expression. The inhibition observed in these assayswas not MHC-restricted as CE cells from both LEW and BN rats were equally inhibitory. CE cells from rabbits and cats were also potent inhibitors of T cell activation, suggesting that the mechanism is evolutionarily conserved. The mechanism of inhibition by CE cells is unknown at this time. o 1992 Academic PWS, IW.

INTRODUCTION Ocular tissue-specific cells, both bone marrow- and non-bone marrow-derived, have been shown to exert potent immunoregulatory effects on immune responses.In particular, the eye seems well equipped to suppress DTH reactions as shown by the anterior chamber-associated immune deviation (ACAID)* phenomena (l), which involves eye/spleen interaction, and also by the presence of cells which directly inhibit ’ This work was supported by NIH Grant EY05417 (D.S.G.), and Research to Prevent Blindness, Inc. D.S.G. is a Research to Prevent Blindness Senior Scientific Investigator. ’ Abbreviations used:ACAID, anterior chamber-associatedimmune deviation; CE cells, comeal endothelial cells; APC, antigen-presenting cells; bSAg, bovine retinal S-antigen; TGF-/3, transforming growth factor-@; BM, bone marrow; TCR, T cell antigen receptor; CAS, concanavalin A-activated supernatant; rINF-y, recombinant interferon-y; Con A, concanavalin A; SPF, specific pathogen-free. 80 0008-8749/92 $5.00 Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

T CELL INHIBITION

BY CORNEAL

ENDOTHELIAL

CELLS

81

lymphocyte activation (2,3). Aqueous humor and cells which line the anterior chamber of the eye are known to inhibit immunologic responsesdue to the presenceof soluble factors such as TGF-@ (4-6), an unidentified low molecular weight inhibitor (6) and by uncharacterized inhibitory cellular interactions (2,7,8). Detection of the inhibitory activities has been demonstrated most often by showing an effect on in vitro assaysof T lymphocyte activation. Few studies have been reported regarding the immunologic activities of cornea1 endothelium (CE), whether inhibitory or supportive of antigenspecific lymphocyte activation. Although the epithelial layers of the peripheral cornea contain Langerhans cells which are presumed to possessantigen-presenting activity (7, 9) other cells with cytokine-inducible Ia, including CE cells, are also present in the cornea and are potential mediators of immunoregulatory effects. The functions of Ia+ antigen-presenting cells (APC) include costimulation, the provision of a second signal required to fully activate antigen-specific CD4+ T cells, leading to IL-2 production and proliferation (10). Costimulatory potential appears to be a property of bone marrow-derived cells ( 11, 12) and may distinguish between “professional” APCs, such as dendritic cells, and “nonprofessional” APCs, i.e., non-BMderived cells with inducible Ia such as glia or vascular endothelium. Interaction of T cells with antigen in the context of Ia on cells lacking costimulatory ability may lead to induction of anergy ( 13- 15) i.e., long-term, antigen-specific unresponsiveness,which may be a mechanism for maintaining unresponsiveness in vivo (16). Although normal, noninflamed CE expresseslittle, if any, MHC antigens ( 17) CE of inflamed corneas is known to be positive for both classI and classII MHC antigens ( 18). Since CE has been reported to expressMHC class I and II molecules in response to cytokines, especially IFN-7 (19-2 I), and since MHC molecules are required as restriction elements for T cell activation, we examined the ability of cultured LEW and BN rat CE cells to present a soluble antigen, bovine retinal S-antigen (bSAg), or peptides from it, to bSAg-specific T cells and T cell hybridomas of LEW origin. As reported here, the endothelial cells were found to be: (I) unable to present antigen to T cell lines or T cell hybridomas, and (2) strongly inhibitory of in vitro lymphocyte proliferation assays,both antigen- and mitogen (Con A)-driven, when conventional proliferation assayswere performed on a monolayer of CE cells. MATERIALS AND METHODS Animals. Specific pathogen-free (SPF) female LEW and BN rats, lOO- 150 g, were purchased from Charles River Laboratories (Wilmington, MA) for use in all experiments and housed under SPF conditions. Antigens and peptides. bSAg isolation and characterization have been described (22). Synthetic peptides (listed in Table 1) were synthesized by conventional t-hoc and F-mot solid-phase synthesesaccording to sequencespredicted from the bSAg cDNA sequence ( 14). “bSAg peptides” indicate the mixture of synthetic peptides bSAg- 15, bSAg-28, and bSAg-32 which represent the major sites on bSAg which stimulate in vitro proliferative responsesin bSAg-specific T cell lines (24). Immunizations. LEW rats were immunized by one 50-~1 SCinjection on a hind thigh with 50 to 100 1.18of peptide emulsified in CFA containing 2.5 mg/ml final concentration of Mycobacterium tuberculosis H-37Ra (Difco, Detroit, MI). T cell lines. Isolation of the long-term, classII-restricted, pathogenic T cell line, R9, has been described (23). The pathogenic RI 170, R1104, and R858 T cell lines, as well

82

OBRITSCH ET AL. TABLE 1 T Cell Lines and Hybridoma Clones and Their Antigen/Peptide Specificities Specificity

T cell lines R9 R1104, R1170 R858 R849 Hybridoma clones lAll.2 lC3.4 E3

Sequence

Reference

bSAg bSAg bSAg273-289 bSAg143-162

see Ref. (43) see Ref. (43) SLTKTLTLVPLLANNRE CGVDFETKAFATHSTDVEED

(23) Unpublished (24) (24)

bSAg213-289 bSAg273-289 bSAgl43-162

SLTKTLTLVPLLANNRE SLTKTLTLVPLLANNRE CGVDFETKAFATHSTDVEED

(25)

(25) Unpublished

as the nonpathogenic line R849, were raised in a similar manner (24). Briefly, lymph node and spleen cells from LEW rats immunized with bSAg or peptides emulsified in CFA were cultured in a 1:1 mixture of DMEM (GIBCO, Grand Island, NY) and modified Click’s medium (Altick Associates,Riverfalls, WI) supplemented with 1 mM glutamine, 1 mM pyruvate, I g/liter glucose, 50 PM /3-mercaptoethanol, 0.1 mA4 nonessential amino acids, 1% syngeneic heat-inactivated rat serum, 50 pg/ml streptomycin, 50 U/ml penicillin, 40 pug/mlgentamycin, and 1 PMimmunizing Ag. T cells were further selected by repeated stimulations of 1 X lo6 viable, density gradient purified cells from the previous culture with 30 X lo6 fresh, 2000 rad-irradiated APC (pooled splenocytes and thymocytes) in 8 ml medium with antigen. No exogenous IL-2 or other cytokines were added. Lymphocyte proliferation assays. Proliferation assayswere done as previously described (23). Modifications done to include CE cells are noted on the tables and figures. When CE cells were used either alone or in combination with spleen cells, the CE cells were first grown to monolayers in the wells of flat-bottom 96-well microculture plates and then treated with cytokines as indicated. The assayswere done in the same wells by adding T cells, or spleen cells and T cells, onto the CE cell monolayer. T cell hybridomas. T cell hybridomas (25) were derived by fusion of the R858 or R849 rat T cell lines with the TCR al-/p- BW5 147thymoma variant BW- 1100.129.237 (26) and selection on HAT medium (supplemented DMEM, 10% FCS (Sigma, St. Louis, MO), antibiotics, 4 g/liter glucose, 100 PA4hypoxanthine, 400 ti aminopterin, and 16 PALM thymidine). After HAT selection, positive wells were expanded, subcloned by limiting dilution, and maintained in a 1:1 mixture of NCTC135 and modified Click’s medium containing 10% FCS, antibiotics, and 4 g/liter glucose. Hybridomas derived from the R858 line (clones lAll.2 and lC3.4) are primarily responsive to the bSAg sequence in residues 273-289, but show definite cross-reactivity with the corresponding sequence in the rat protein, which differs by two amino acid residues within that epitope (25). The hybridoma clone, E3, derived from the R849 T cell line recognizesonly the bSAg sequenceat residues 143-162. No evidence of cross-reactivity with the rat sequenceat that site has been found (unpublished observation). Accordingly, the clones lAll.2 and lC3.4 recognize both self- and nonself-sequenceswhile clone E3 recognizesnonself only; the epitope recognized by E3 is functionally a foreign determinant.

T CELL INHIBITION

BY CORNEAL

ENDOTHELIAL

CELLS

x3

Assay of IL-2 production by T cell hybridomas. T cell hybridomas were incubated with and without Ag in 96-well microtiter plates at 5 X lo4 cells/well in the presence of a monolayer of nonirradiated CE cells and/or normal LEW splenocytes ( 106/well) as APC. After 24 hr, supernatants were removed and diluted in 96-well, flat-bottom microculture plates. IL-2-dependent CTLL indicator cells were added (20 X 104/well) to a final volume of 150 ~1 in a 1:1 mixture of complete RPM1 1640 and modified Click’s with 10% FCS. CTLL cells were cultured for 18 hr, pulsed with 1 &i/well [3H]TdR (Amersham, Arlington Heights, IL), and harvested onto glassfiber filters for counting. Cornea1endothelial cells. The comeal endothelial cells were isolated from rat, rabbit, and cat corneas by modification of a previously reported procedure (27). The freshly enucleated eyes were immersed in Betadine for 3 min and rinsed in sterile D-PBS. The globe was dissected2 mm behind the limbus. The iris was removed and the scleral rim was trimmed away, leaving the cornea. Descemet’s membrane was peeled off the cornea, chopped into small pieces, and incubated at 37°C in DMEM containing 20% FCS and 1.5%chondroitin sulfate. The FCS and chondroitin sulfate were included in the CE culture medium to promote establishment and expansion of the primary cultures. After l-2 weeks, these cells reach confluency and are then trypsinized for l-2 min to obtain a single cell suspension. The cells were washed and plated so that the cells collected from 12 rat corneas were used to seed72 wells of a 96-well plate. These cells again grew to confluency within 3-4 days, at which time they were used for the experiments. C:vtokines. “CA&” the unconcentrated, l-day culture supernatant from Con Astimulated LEW rat splenocytes, upregulates the expression of MHC antigens on the cultured CE cells. Recombinant rat IFN-7 (provided by Dr. R. Caspi, Bethesda, MD) was also used in some experiments to verify the activity of the CAS preparation and demonstrated equivalent results. All of the experiments reported here were done with the CAS preparation. Zmmunqjluorescence. CE cells were cultured on sulfonated plastic coverslips (28) for 2-4 days in the presence of 20% CAS or 500 U/ml rat rIFN-y or medium alone. Medium was rinsed off and replaced with a 1:100 dilution of mAb OX-l 8 (class Ispecific), OX-6 (class II-specific) or control mAb in D-PBS containing 10% fetal calf serum. The cover slips were kept on ice from this point. After 45-60 min, the cells were rinsed and incubated with fluorescein-labeled anti-mouse IgG ( 1:50) for 45-60 min. The cells were washed, and the coverslips were mounted on glass slides for observation. RESULTS CE cells lack direct antigen-presenting capability. Initial experiments were done to assessthe ability of CE cells to present bSAg to the bSAg-specific T cell line R9, which has been well characterized. The R9 line has been shown to be CD4’ and expresses a class II-restricted proliferative response to bSAg and several synthetic peptides of bSAg (23, 29). The APC activity of CE cells was tested by applying resting R9 cells and antigens to a monolayer of CE cells. Conventional assay conditions utilizing normal, irradiated splenic APC were done asa control. Although the R9 cells responded vigorously in the presence of antigen or mitogen and splenic APCs alone. no proliferation was observed when the CE cells were used as the sole “APC” (Fig. 1). Fur-

84

OBRITSCH ET AL.

0 Spleen cells CE Cells Cytokines

+ -

+ -

FIG. I. CE cells do not themselves have antigen-presenting capability, but inhibit the response of the bSAg-specific R9 T cell line to antigen and irradiated, splenic APCs. R9 line cells (3 X 104/well)were added to 0.5 X lo6 spleen cells/well (2,000 rad irradiated) or a monolayer of LEW rat CE cells (unirradiated) together with the indicated antigens or Con A for 2 days before adding 1 &i/well of [3H]thymidine. In other cases,the splenic APCs and R9 T cells were dispensedinto wells which already contained a monolayer of CE cells. In some casesthe CE cells had been pretreated for 2 days with either 5 or 20% CAS and washed three times prior to being used in the assays.

thermore, the CE cells were found to exert a potent inhibitory effect when normal, irradiated splenic APC were added to a CE cell monolayer and the resulting co-culture tested for the ability to support a proliferative response in the R9 line (Fig. 1). Since IFN-7 has been shown to upregulate MHC expression on CE cells ( 19-2 l), and classII expression is required for APC activity to CD4+ T cells (30, 3 l), the effect of the addition of a cytokine-containing supernatant to the CE cells on their APC activity was investigated. R9 cells and antigen were applied to a monolayer of CE cells that were either untreated or pretreated for 2 days to induce MHC expression with either 5 or 20% cytokine supernatant (Fig. 1). Pretreatment of the CE cells for 2 days with a cytokine-containing supernatant did not induce antigen-presenting ability and the R9 response on splenic APC remained strongly inhibited if cytokine-treated CE ceils were present. The ability of the CAS cytokine preparation and the recombinant rat IFN-y to upregulate MHC on rat CE cells was tested. In agreement with several previous reports (20,2 1), the 1IW-y was found to increaseclassI and classII expression. The CAS preparation, used for all of the data shown in this report, was also found to increase class I expression as demonstrated by staining with 0X-18, and class II expression (OX-6 staining), although the increase in class I expression following CAS treatment was much greater (Fig. 2). The lack of a T cell response in the presence of the CE cells was not found to be unique to the R9 T cell line which had been carried through at least 12 cycles of in vitro stimulation and propagation. The R1104 T cell line, which was also specific for bSAg and had been selected for only 3 cycles in vitro, was assayedon a monolayer of CE cells that had been pretreated with 0 or 20% cytokine supernatant, or with normal splenic APC. The RI 104 line was also unable to respond to antigen if CE cells were used as the APC, and the R1104 proliferative response using splenic APC was suppressedin the presenceof CE cells (data not shown). The proliferative responseof the

T CELL INHIBITION

BY CORNEAL ENDOTHELIAL

CELLS

85

FIG. 2. Treatment of CE cells with the rat CAS cytokine supematant increasedcell surfaceMHC expression. CE cells were incubated for 4 days in the presenceof conventional, cytokine-free medium (A and C) or 20% CAS (B and D) and stained with either OX- I8 (A and B) or OX-6 (C and D). Magnification. 150X.

R 1170 T cell line, which is also specific for bSAg and had been selected for 7 cycles in vitro, was suppressedin the presence of CE cells. The inhibition is not MHC-restricted. If the inhibition of the responsein the presence of splenic APC was due to the induction of anergy, or another MHC-dependent mechanism, then the effect should be MHC class II-restricted. This possibility was tested by examining the ability of LEW (syngeneic) and BN (allogeneic) CE cells to inhibit proliferation of the R9 cells (Fig. 3). The R9 cells were assayedwith normal splenic APC on a CE cell monolayer derived from either LEW or BN corneas. Under these conditions. both LEW and BN CE cells equally inhibited the proliferation of R9 cells to antigen and mitogen. CE cells of rabbits and cats also exhibit inhibition. Since the inhibitory property of the rat CE cells was not MHC-restricted, and iris/ciliary body cells have been reported to expressa non-species-specific inhibitory activity (2) CE cells of nonrodent animals

86

OBRITSCH ET AL. 3 0 r; E B m2 2 = E

250 200 150 100 50 0

LEW spl

LEW spl + LEW CE cells

LEW spl + BN CE cells

FIG.3. Inhibition of T cell line activation is not MHC-restricted. R9 line cells were incubated with normal irradiated splenic APC or splenic APC plus either syngeneic CE cells (LEW) or allogeneic CE cells (BN) as described in the legend to Fig. 1. The proliferative responsewas assessedas described for Fig. 1.

were also examined. CE cells from rabbits and cats were also found to inhibit the proliferative response of the rat R 1170 T cell line (Fig. 4). Inhibition is reversible. The inhibitory effect of exposure to the CE cells was found to be dependent on the continued presenceof CE cells in the T cell cultures. Removing the T cells from the CE cells and washing were found to substantially restore their responsivenesswhen assayedon splenic APC (Fig. 5). R9 cells were incubated for 2 days either on CE cells alone or on CE cells pretreated with 10%cytokine supernatant prior to the assay. The reversibility is unlike the anergy observed in murine systems where the unresponsiveness is not reversed by isolating and washing the T cells (12). Preactivated T cells proliferated well even when co-cultured with CE cells. T cells were preactivated for 11 days on fresh APCs with antigen and then transferred to empty wells, or wells containing a monolayer of rat CE cells, or to wells with fresh APCs. The results show that once T cells are activated, they are no longer susceptible to the inhibitory effects of co-culture with CE cells (Fig. 6). Inhibition is not due to the carry over of CE medium, and CE cells themselves exhibit inhibition independent of the medium. Since the medium used to culture CE

LEW spleen cells CE cells

+ -

LEW

L;W

Rabbit Ra+bbit

Cat

CL

FIG.4. CE cells of rabbits and cats also inhibit T cell proliferation. The R1170 T cells were incubated with normal irradiated splenic APCor splenic APCplus CE cells either from rabbits or from cats asdescribed in the legend to Fig. 1. The proliferative responsewas assessedas described for Fig. 1.

T CELL INHIBITION 10

X E

8 of = $ !-

BY CORNEAL ENDOTHELIAL no Ag

CELLS

87

I

75 50 25

E 0 CE cells

CE cells + cytokines (10%)

FIG. 5. The inhibition mediated by CE cells is reversible. CE cells were seededinto 96-well microplates and allowed to form a confluent monolayer. Wells were then left untreated, or incubated with 10%CAS for 2 days. After CAS treatment, all wells were washed three times and 3 X lo4 rested R9 cells/well were added. After 2 days, the R9 cells were recovered, washed, and assayedon irradiated splenic APC.

cells differs significantly from the medium usedto maintain and assaythe T lymphocyte lines, there is the possibility that CE cell medium components are carried over into the assaysand effect the inhibition. Inclusion of up to 25% CE culture supernatant had relatively little effect on the antigen or mitogen responsivenessof R9 cells assayed on irradiated splenic APC (Fig. 7A). This indicates that carry over of a few percent of CE medium containing 20% FCS and 1.5%chondroitin sulfate does little to contribute to the inhibition. Furthermore, it was found that, even when CE cells were cultured in lymphocyte medium for 4 days prior to the assay,they were still able to inhibit T cell proliferation (Fig. 7B), as well as CE cells cultured only in CE medium throughout the time prior to assay(Fig. 7C). The results thus far are consistent with the suppression of proliferation being mediated by either a soluble factor produced by the CE cells or

Transfer of line cell assays to:

control, fresh not irradiated transferred APC

empty wells

wells with CE cell monolayer

FIG. 6. The R 1170 line cells activated on fresh APCs prior to co-culture with CE cells proliferated well even in the presenceof CE cells. RI 170 cells were preactivated for 36 hr prior to assay,as described in the legend to Fig. I. These preactivated cells were then transferred without washing to wells with or without CE cells, or to wells with fresh APC (0.5 X lo6 spleen cells/well, 2000 rad irradiated). Control assay,with and without bSAg or Con A, was left undisturbed. After another 36 hr, 1 &i/well of [jH]thymidine was added and incorporation was counted after I6 hr.

88

OBRITSCH ET AL. Proliferation, Acpm x 1O9 3 no CE cell supernatant 25% CE cell supernatant

Splenocytes

lymphocyte medium

Splenocytes Splenocytes+CE cells cultured in CE cell medium

FIG. 7A. Carry over of CE cell culture medium from CE cell cultures shows little inhibitory activity. R9 line cells were assayedon irradiated splenic APC in the presence of normal medium or 25% supernatant from normal CE cell cultures. The CE cell supematant was collected from confluent monolayers of CE cells after 3 days incubation. CE cells cultured in either lymphocyte medium (B) or CE cell medium (C) inhibit T cell proliferation. Prior to assay,CE cells were cultured either in lymphocyte medium for 4 days or in CE cell medium only. Then, the proliferation assay was performed as described in the legend to Fig. 1.

by cell interaction. However, since up to 25% CE cell supernatant alone was unable to inhibit proliferation assaysto the degree observed when the cells were co-cultured, it would appear that if the inhibition were due to a soluble factor, only minimal amounts are produced, or it is highly unstable. Do the CE cells aj2ct the viability of the lymphocytes? It is possible that the CE cells, in not supporting a proliferative response, simply do not maintain the viability of the T cells, perhaps by not producing a factor neededby the T cells, or by consuming something needed by the T cells, or by producing a factor such as TGF-/3 which has been reported to be toxic for T cells (32). To examine these possibilities, the effect of prolonged incubation of T cells from the RI 104 line with CE was studied (Table 2). Culture of the T cells for up to 6 days on CE cells did not inhibit the antigen-specific proliferation of the T cells recovered from the cultures, and pretreatment of the CE cells with CAS had little effect. Furthermore, R1104 T cells incubated with the CE cells for up to 8 days had a dramatically greater viability and recovery than those incubated in control wells without a monolayer of CE cells (Table 3). Although the counts of R 1104 line cells recovered on Days 2 and 4 are lessthan the nominal number of cells put into culture on Day 0, it should be noted that recovery of the RI 104 cells

T CELL INHIBITION

BY CORNEAL ENDOTHELIAL

89

CELLS

TABLE 2 Effect of Longer Exposure of the T Cells to the CE Cells CE cells

% of cytokine

Time

SAg peptides

Bov SAg

Con A

None LEW LEW LEW LEW LEW

None 0 5 20 5 20

I day I day I day I day 6 days 6 days

193” 117 103 108 132 197

198 192 II5 120 I38 I58

94 79 I7 29 I2 62

Note. R9 T cells were cultured in the absenceor presenceof CE cells which had been preincubated with the CAS cytokine preparation as indicated prior to application of the R9 cells. After the indicated time, the R9 cells were recovered, washed,and assayedfor proliferation in conventional assaysusing irradiated splenic APC. ’ ACPM X IO-‘.

was incomplete due to their adherenceto the CE cells and also due to lossesin collecting and preparing the cells for counting. The maintenance of viability for the duration of the assay (8 days) is striking compared to the R1104 cells put into wells without CE cells. Two T cell hybridomas, lC3.4 and E3, and the CTLL indicator line used in the IL-2 assayswere also plated on CE cell monolayers or in control wells without CE cells to determine the effect of the CE cells on their viability and growth. No inhibition of growth was found (Table 3). To the contrary, the CTLL cells, after 2 days on CE cells without IL-2, were still viable and capable of proliferation when IL-2 was added on Day 2 and the cells counted on Day 4. These results also show that the lack of CTLL proliferation in the following assayswas due to the absenceof IL-2 rather than to a growth-inhibiting substance secretedby the CE cells. The hybridoma cells rapidly overgrew the wells whether or not CE cells were present.

TABLE 3 Growth and Survival of a T Cell Line, T Cell Hybridomas, and CTLL Cells on Monolayers of CE Cells Day 2

Day 4

Day 8

Cells

Day 0

CE”

CE/CAS”

No CE

CE

CE/CAS

No CE

CE

CE/CAS

No CE

RI104 IC3.4 E3 CTLL’

20 40 40 30

lob 113 81 31

I5 202 I74 73

2 112 98 35

I8 220 221 74

15 I83 238 96

1 265 240 171

24 -d -

20 -

0 -

a Monolayers of CE cells in 96-well plates were cultured with medium only or medium containing 20% CAS for 4 days and washed prior to addition of hybridomas, T cells, or CTLL cells in the indicated numbers. Traces of CAS cytokine may persist after the washes. ’ Number of viable cells as assayedby trypan blue exclusion, X IO-“. ’ CAS was added on Day 2 to sustain the growth of these IL-2-dependent cells. d These cultures overgrew the wells and died.

90

OBRITSCH

ET AL.

Activation of T cell hybridomas. T cell hybridomas are widely considered to have lessstringent activation requirements than do T cells from lines, clones, or fresh lymphocyte preparations, since their activation is largely dependent on receptor occupancy by Ia-peptide complexes. Since the CE cells have cytokine-inducible Ia, their ability to present antigen to a T cell hybridoma was tested (Fig. 8). The IA1 1.2 T cell hybridoma, which is specific for the synthetic peptide bSAg-28 from bSAg, was not activated by either the peptide or intact bSAg when CE cells were used as APC, but was activated by peptide and bSAg presented by conventional nonirradiated splenic APC. The slight elevation in incorporation found when CAS-treated CE cells were used is probably due to residual IL-2 from the CAS pretreatment. Similarly, assaysof two other hybridomas, lC3.4 and E3, on normal CE cells and CAS-treated CE cells yielded no detectable IL-2 secretion (Fig. 9 and data not shown). Eflect of CE cells on the activation of a T cell hybridoma by splenic APC. Since the addition of CE cells to proliferation assaysof the R9 and R 1104 T cell lines was highly inhibitory, the effect of the CE cells on the response of the E3 T cell hybridoma was examined. The responsivenessof the E3 cells was of particular importance in that this clone has no reactivity to the rat sequencewhich corresponds to the bovine sequence to which the E3 clone was made. The E3 cells were not activated to produce IL-2 by their antigen, peptide bSAg143-162, on CE cells alone. Pretreatment of the CE cells with 25% CAS for 4 days to induce Ia also had no effect. When CE cells were tested for their ability to inhibit E3 activation in the presence of splenic APC, the response was inhibited by approximately 50% relative to the Ag-specific response found in the presenceof irradiated splenic APC alone (Fig. 9). In this respect, the hybridomas were somewhat less sensitive to inhibition than were the T cell lines which were routinely inhibited by 80% or more. DISCUSSION In this report we have shown that CE cells do not themselvesact asantigen-presenting cells; conversely, they exert a potently inhibitory effecton antigen- and mitogen-induced

CE cells

CE cells + cytokines (20%)

splenic APC

FIG. 8. CE cells, with and without pretreatment with CAS, do not support antigen-specific activation of the lAl1.2 T cell hybridoma. Monolayers of CE cells were pretreated with either 20% CAS or control medium for 2 days and washed. Hybridoma cells, 5 X lo4 cells/well, were added together with antigen and incubated for 1 day. Supematants were collected and assayed for IL-2 on CTLL cells as described under Materials and Methods.

T CELL INHIBITION

p!

BY CORNEAL ENDOTHELIAL 0 El W I

60

0

;

40-

8 j

20-

CELLS

91

noAg bSAg143-16: bSAg270-28 ConA

5 oSpleen cells EC cells

+ -

Cytokine

-

+

+

+

+ +

+ +

+

FIG. 9. CE cells reduce, but do not fully abrogate, the ability of splenic APCs to support activation of the E3 T cell hybridoma. Monolayers of CE cells were pretreated with either 20% CAS or control medium for 2 days and washed. Irradiated splenic APC were then added to these wells and also to wells which did not contain CE cells. Immediately after, hybridoma cells, 5 X lo4 cells/well, were added together with antigens and controls and incubated for 1 day. The supernatants were collected and assayedfor IL-2 on CTLL cells as described under Materials and Methods.

and proliferation of T cells from antigen-specific T cell lines and peptideactivation specific T cell hybridomas. Streilein et al. (33, 34), in elucidating the mechanisms by which ACAID is expressed, have identified an F4/80+ cell, presumably a bone marrow-derived macrophage-like cell, present in the murine iris/ciliary body which contains ACAID-inducing properties. Although extraocular F4/80+ cells do not have this ability, they can acquire it by placement in the anterior chamber, indicating that an activity produced locally in the eye is required (34). Clearly, the CE cells and their suppressive activity cannot be equated with the F4/80+ cells, but they may contribute to the anterior chamber environment needed for the expression of ACAID. In some respects, the results are suggestive of the effects of TGF-/3, but differ in other aspects.For example, proliferation of the CTLL cells was not inhibited by CE supernatant, as TGF-0 has been reported to do (35), nor was their growth inhibited even when cultured directly with the CE cells. The CE cells also failed to inhibit proliferation of any of the three hybridomas and maintained the viability of resting T line cells for at least 8 days. Furthermore, it has been reported that TGF-/3 does not block IL-2 production (36), while the CE cells did inhibit IL-2 production by the E3 hybridoma. These results lead to the suggestion that if the inhibition is mediated by a soluble factor, such a factor would have to be produced in such small amounts that close proximity is also required, or that the factor was very labile. Experiments are in progress to establish the possible role of TGF-P in our observations. In our experimental system, it is conceivable that some carry over of the CE cell culture medium might occur in the proliferation assays,affecting T cell proliferation. But, the finding that up to 25% of CE cell supernatant showed only a minimal suppression, together with the result showing that CE cells still exhibited a strong inhibition even when cultured in lymphocyte medium for 4 days prior to assay show that the CE cells themselves are the source of the inhibitory activity. Roberge et al. suggesteda unique inhibitory mechanism of ocular tissue. They found that retinal Miiller cells (neural glia) can be induced by IFN-7 to express Ia (37) and have been found to produce IL- l-like activity (38). Caspi et al. (39) reported

92

OBRITSCH

ET AL.

that IFN-y-treated Miiller cells inhibited the antigen response of primed Th cells. However, trypsinization and glutaraldehyde fixation of Miiller cells pretreated with antigen and IFN--r removed the inhibitory properties of the Mtiller cells, paradoxically conferring antigen-presenting ability on them (33). An important difference in our results is that the CAS treatment of the CE cells had no effect on any of the assays, apart from elevating MHC expression. There are also Ia+ cells in surrounding uveal tissues (2) which have been shown to exhibit similar properties. For example, rat ciliary body cells were unable to present antigen effectively to a T cell line and strongly inhibited conventional proliferation assaysin a non-MHC-restricted manner (2). The overall inhibitory effect, which was mediated by both soluble and cell-bound mechanisms, was not found to have a TGF-fl component, but was partially reversible by indomethacin (2). Furthermore, the mechanism of this inhibition was found to be a non-species-specificproperty (2). We also found that CE cells from nonrodent animals, such as rabbits and cats, exhibited an inhibitory property, indicating that the inhibitory mechanism of CE cells is evolutionarily conserved. Many laboratories have studied the presence, distribution, and antigen-presenting capabilities of cells with inducible Ia from several tissues,especially non-bone marrowderived cells not thought to be “professional APCs,” including neural glia, vascular endothelium, and ciliary body epithelium. An important consideration of these experiments examining the roles of the various Ia+ cells in antigen presentation or regulatory mechanisms is that the in vitro activation requirements of cultured T cell clones and lines differ from those of fresh, resting T cells (40-42). Regardless of the mechanism by which the cornea1endothelial cells expresstheir inhibitory effects, our results show that the immunosuppressive environment of the anterior chamber could, at least in part, be due to the activity of the CE cells in addition to the iris/ciliary body cells which are already well known to participate in ACAID (34, 35). One possible explanation for the inhibition of T cell responsivenessis the induction of anergy. Current hypotheses regarding the costimulation requirement and its relationship to the induction of anergy postulate that anergy is induced in the absenceof a factor or cell-cell interaction, i.e., by costimulation-deficient cells (10, 15). Another interpretation is suggestedby experiments of Roberge et al. (38) who found that the inhibitory influence of Mtiller cells could be removed by trypsinization, making them capable of supporting a proliferative response.These data suggestthat another mechanism for inducing unresponsiveness, i.e., a receptor-ligand interaction which downregulates the responses,might be exhibited by some cells. The lack of a requirement for MHC compatibility and the ability to restore responsiveness by removing and washing the T cells rule out anergy or anergy-like phenomena as reported for murine T cells, even though the CE cells express class II antigens and could theoretically induce anergy. Another interesting feature of the inhibition stems from the result showing that preactivated T cells could proliferate well even in the presenceof CE cells. This indicates that the inhibitory effect is expressedat the activation stageof the T cell response. At a later stage,when the T cells are committed to proliferation, they became refractory, even on direct contact, to the inhibitory effect of the CE cells. This phenomena is being investigated. Combining the results shown in this report, we postulate that there would be three possible mechanisms that could work independently or in concert, i.e., (1) cell-cell interaction, (2) TGF-& and (3) some other soluble factor(s). Further analysis of this

T CELL INHIBITION

BY CORNEAL ENDOTHELIAL

CELLS

93

inhibition would help to understand local immunoregulatory mechanisms, which is essential for the manipulation of ocular inflammation. ACKNOWLEDGMENTS We thank Drs. Robert A. Prendergast and J. Wayne Streilein for their critiques of the manuscript. We also thank Lisa Kamp for technical assistance.

REFERENCES I. Streilein, J. W., FASEB J. 1, 199, 1989. 2. Helbig, H., Gurley, R. C., Palestine, A. G., Nussenblatt, R. B., and Caspi, R. R., Eur. J. Immunol. 20, 2457, 1990. 3. Caspi, R. R., and Roberge, F. G., J. Autoimmunity 2, 709, 1989. 4. Helbig, H., Kittredge, K. L., Coca-Prados, M., Davis, J., Palestine, A. G., and Nussenblatt, R. B., GraefeS Arch. Clin. Exp. Opthalmol. 229, 84, 1991. 5. Streilein, J. W., and Cousins, S. W., Curr. Eye Res. 9(Suppl.), 175, 1990. 6. Granstein, R. D., Staszewski, R., Knisely, T. L., Zeira, E., Nazareno, R., Latina, M., and Albert, D. M., J. Immunol. 144, 3021, 1990. 7. Williamson, J. S. P., Bradley, D., and Streilein, J. W., Immunology 67, 96, 1989. 8. Streilein, J. W., Cousins, S., and Williamson, J. S. P., Int. Ophthalmol. 14, 317, 1990. 9. Williamson, J. S. P., DiMarco, S., and Streilein, J. W., Invest. Ophthalmol. Visual Sci. 28, 1527, 1987. 10. Mueller, D. L., Jenkins, M. K., and Schwartz, R. H., Annu. Rev. Immunol. 7,445, 1989. Il. Gaspari, A. A., Jenkins, M. K., and Katz, S. I., J. Immunol. 141, 22 16, 1988. 12. Jenkins, M. K., and Schwartz, R. H., J. Exp. Med. 165,302, 1987. 13. Burkly, L., Lo, D., Kanagawa, O., Brinster, R. L., and Flavell, R. A., Nature (London) 342, 562, 1989. 14. Morahan, G., Allison, J., and Miller, J. F. A. P., Nature 339, 662, 1989. 15. Jenkins, M. K., Mueller, D., Schwartz, R. H., Carding, S., Bottomly, K., Stadecker, M. J., Urdahl, K. B., and Norton, S. D., In “Mechanisms of Lymphocyte Activation and Immunoregulation. III. Developmental Biology of Lymphocytes” (S. Gupta, W. Paul, M. Cooper, and E. Rothenberg, Eds.), pp. 167-178. Plenum Press,New York, 1991. 16. Schwartz, R. H., Science 248, 1349, 1990. 17. Whitsett, C. F., and Stulting, R. D., Invest. Ophthalmol. Visual Sci. 25, 5 19, 1984. 18. Pepose,J. S., Gardner, K. M., Nestor, M. S., Foos, R. Y., and Pettit, T. H., Ophthalmology 92, 1480, 1985. 19. Young, E., Stark, W. J., and Prendergast, R. A., Invest. Ophthalmol. Visual Sci. 26, 571, 1985.

20. Donnelly, J. J., Li, W., Rockey, J. H., and Prendergast, R. A., Invest. Ophthalmol. Visual Sci. 26, 575, 1985. 21. Foets, B. J. J., van den Oord, J. J., Billiau, A., Van Damme, J., and Missotten, L., Invest. Ophthalmol.

22. 23. 24. 25. 26.

VisualSci. 32, 341, 1991. Gregerson, D. S., Fling, S. P., and Wohlhueter, R. M., Exp. Eye Rex 43, 803, 1986. Gregerson, D. S., Obritsch, W. F., Fling, S. P., and Cameron, J. D., J. Immunol. 136, 2875, 1986. Fling, S. P., Donoso, L. A., and Gregerson, D. S., J. Immunol. 147,483, 1991. Fling, S. P., Gold, D. P., and Gregerson, D. S., Cell. Immunol. 142, 275, 1992. White, J., Blackman, M., Bill, J., Kappler, J., Marrack, P., Gold, D. P., and Born, W., J. Immunol. 143, 1822, 1989.

27. Yue, B. Y. J. I., Sugar, J., Gilboy, J. E., and Elvart, J. L., Invest. Ophthalmol. Visual Sci. 30,248, 1989. 28. Furst, A., and Mahowald, A. P., J. Immunol. Methods 70, 101, 1984. 29. Gregerson, D. S., Fling, S. P., Obritsch, W. F., Merryman, C. F., and Donoso, L. A., CeN. Immunol. 123,427, 1989. 30. Schwartz, R. H., Annu. Rev. Immunol. 3,237, 1985.

31. Unanue, E. R., Annu. Rev. Immunol. 2, 395, 1984. 32. Siepl, C., Bodmer, S., Frei, K., MacDonald, H. R., DeMartin, R., Hofer, E., and Fontana, A., Eur. J. Immunol. 18, 593, 1988. 33. Wilbanks, G., and Streilein, W., J. Immunol. 146, 2610, 1991. 34. Wilbanks, G., Mammolenti, M., and Streilein, J. W., J. Immunol. 146, 3018, 1991. 35. Stoeck, M., Mieshcer, S., MacDonald, H. R., and von Fleidner, V., J. CeN. Phys. 141, 65, 1989.

94

OBRITSCH ET AL.

36. Wahl, S. M., Hunt, D. A., Wang, H. L., Dougherty, S., McCartney-Francis, N., Wahl, L. M., Ellingsworth, L., Schmidt, J. A., Hall, G., Roberts, A. B., and Spom, M. B., J. Immunol. 140, 3026, 1988. 37. Roberge, F. G., Caspi, R. R., Chart, C. C., Kuwabara, T., and Nussenblatt, R. B., Curv. Eye Res. 4, 975,1985. 38. Roberge, F. G., Caspi, R. R., and Nussenblatt, R. B., J. Immunol. 140, 2193, 1988. 39. Caspi, R. R., Roberge, F. G., and Nussenblatt, R. B., Science 231, 1029, 1987. 40. Umetsu, D. T., Pober, J. S., Jabara, H. H., Fiers, W., Yunis, E. J., Burakoff, S. J., Reiss, C. S., and

Geha, R. S., J. Clin. Invest. 76, 254, 1985. 4 I. Inaba, K., and Steinman, R. M., J. Exp. Med. 60, 1717, 1984. 42. Chu, E. T., Lareau, M., Rosenwasser,L. J., Dinarello, C. A., and Geha, R., J. Immunol. 134, 1676, 1985. 43. Shinohara, T., Dietzschold, B., Craft, C. M., Wistow, G., Early, J. J., Donoso, L. A., Horwitz, J., and Tao, R., Proc. Natl. Acad. Sci. USA 84, 6975, 1987.