Adhesion of lymphoid cells to fibroblasts in tissue culture

CELLULAR

122,33-47

IMMUNOLOGY

Adhesion DAVID

(1989)

of Lymphoid

ABRAHAM,

GEORGE

Cells to Fibroblasts BOW-GHARIOS,

HELEN

in Tissue Culture

MUIR, AND IRWIN OLSEN’

Kennedy Institute ofRheumatology, 6 Bute Gardens, London W6 70 W, United Kingdom Received October 14, 1988; accepted March 20,1989 In this study we have examined the cellular and molecular specificity of lymphocyte interaction with fibroblasts. Using mitogen-activated T-cells, we found that attachment to fibroblasts was highly sensitive to protease treatment, and to an antibody raised against the purified lymphocyte plasma membrane, but it was not mediated by the MEL- 14 surface antigen or phosphomannosy1 receptors. Lymphocyte interaction with fibroblasts was also unaffected by monoclonal antibodies against the LFA- 1, Mac- 1, and Class II MHC antigen complexes. In contrast, adhesion of both T- and B-lymphocytes was strongly inhibited by fucoidan, a polymer of sulphated fucose, whereas fucose, mannan, and mannose 6-phosphate had no effect. Both B- and T-lymphoid cell lines were able to recognise and adhere to fibroblasts, although the marked differences between the attachment of the different types of cell did not appear to be related to their immunological function. The attachment of most of the cell lines was prevented by the presence of fucoidan, whereas the inhibition of binding of each of the lymphoid lines in the presence of the anti-T-lymphocyte plasma membrane antibody varied widely. These findings suggest that lymphocyte attachment to libroblasts involves multiple cell surface receptors, and that these are expressed at different levels on specific T- and B-cells. 0 1989 Academic PPXS, Inc. INTRODUCTION

The interaction of lymphocytes with other types of cells is thought to be fundamental to the expression of many normal lymphocyte functions. Lymphocyte emigration from the vascular system and circulation through peripheral lymphoid organs involve recognition of certain endothelial cells within the various lymphoid tissues (1, 2). Similar adhesive reactions are also believed to be responsible for the preferential homing patterns of specific types of T- and B-lymphocytes to selective organs in vivo (3), and for the extravasation of effector lymphocytes into areas of inflammation, as in the rheumatoid synovium (4), in degenerative muscle tissue (5, 6), in some types of hepatic disease (7), and in certain dermal lesions (8,9). The essential role of lymphocyte attachment is also evident in other functions such as cytotoxicity, in which target cell lysis occurs only following conjugate formation ( 10). Lymphocyte interaction with fibroblasts in vivo has also been shown to occur, in the epidermis during wound healing (11) and when fibroblasts function as accessory cells in antigen presentation ( 12). Moreover, fibroblasts induced by lymphokines to express Class II major histocompatibility complex (MHC) antigens are functionally recognized by cloned cytotoxic T-lymphocytes, a reaction that may be involved in ’ To whom correspondence should be addressed. 33 0008-8749/89 $3.00 Copyright 0 1989 by Academic Press, Inc. All rights of reproduction in any form reserved.

34

ABRAHAM

ET

AL.

the rejection of allografis ( 13). Lymphocytes have also been shown to adhere to and explore the surface of fibroblasts in tissue culture (14, 15). This type of cell-to-cell contact is now known to be accompanied by the direct transfer of certain lysosomal enzymes from the lymphocytes (16). Although the molecular basis of lymphocyte adhesion to fibroblasts is still poorly understood, the results of recent studies have suggested that multiple receptors are present on different subpopulations of lymphocytes and that B-cells do not share the same receptors as T-lymphocytes ( 17, 18). The present study examines the interaction between lymphocytes and dermal fibroblasts and provides further evidence for the presence of multiple adhesion molecules on both T- and B-cells. Our results also show that differential levels of these surface receptors expressed at the plasma membrane confer major differences in the adhesive properties of the cell. MATERIALS

AND METHODS

Tissue culture plastics were obtained from Costar (Northumberland, UK), culture media were from Flow Laboratories (Irvine, UK), heat-inactivated fetal calf serum was from Sera Lab. Ltd. (Sussex, UK), and tissue culture plastics were from GIBCO (Middlesex, UK). Monoclonal antibodies were prepared from the respective hybridoma cell lines obtained from the American Type Culture Collection (Rockville, MD; ATCC) or were gifts, as noted below. Normal goat serum, FITC-conjugated anti-murine IgM, and both FITC- and RITC-conjugated anti-rat immunoglobulins were from ICN Biomedicals Ltd. (Bucks, UK). Lowicryl K4M resin and glutaraldehyde were obtained from Taab Laboratories Ltd. (Berks, UK). Goat anti-rabbit IgG conjugated to 10 nm gold (AuroProbe) was obtained from Janssen Pharmaceutical Ltd. (Oxon, UK). Tween 20, EIA grade, was from Bio-Rad (Her&, UK), and Protein A-Sepharose was supplied by Pharmacia Fine Chemicals (Bucks, UK). All other reagents were obtained from the Sigma Chemical Co. Ltd. (Dorset, UK) and were of the highest purity available.

Cell Culture Single-cell suspensions of lymphocytes were prepared from spleens of CBA strain mice as previously described ( 19). They were activated by incubating 1O6cells/ml for 72 hr at 37°C in 5% CO2 in air, in RPM1 1640 medium supplemented with 5% heatinactivated foetal calf serum (FCS), 100 U/ml penicillin, 100 pg/ml streptomycin, 20 mMHepes, and 50 pM2-mercaptoethanol. To generate T-lymphoblasts, concanavalin A (Con A) was added at the start of culture at 4 &ml. For activated B-cells, lipopolysaccharide (LPS) was added at 20 &ml. The following murine lymphoid cell lines were grown as for lymphocytes, but in RPM1 medium containing 10% FCS. DO- 11.10, a Balb/c X AKR functional T-helper-cell hybridoma (20) M 12 and A20, spontaneous Balb/c B-cell lymphomas (2 I), TA3, an M 12 X LPS-stimulated Balb/c X AJ spleen cell hybridoma (22), and HB32, an NSl X Ig-immunized Balb/c spleen hybridoma (23), were supplied by Dr. B. Chain, University College, London. Dr. M. Feldman, Charing Cross Sunley Research Centre, London, supplied EL4, a chemically induced C57/Bl T-cell lymphoma; K2S, a C57/B6 radiation-induced leukemia virus (Rad-LV/Nu 1) lymphoma with suppressor T-cell function; BW 5 147, a spontaneous AKR T-cell lymphoma; and NSl and

LYMPHOCYTE-LlBROBLAST

INTERACTION

35

SP2/0, both Balb/c plasmacytomas (24). BA4, a Balb/c antigen-specific cytotoxic Tcell clone (25), and A3, a CBA/Ca X NSl hybridoma (26), were generously provided by Dr. B. Askonas, National Institute of Medical Research, London. Professor R. N. Maini, Kennedy Institute and Charing Cross and Westminster Medical School, London, supplied Ag8 (P3/X63 Ag8), a Balb/c myeloma (27), and two human cell lines: Molt 4, a T-lymphoblastoid cell, and Wil-2, a B-lymphoblastoid cell. Normal human skin fibroblasts were grown to confluence as monolayers in 48-well tissue culture plates (approximately 1 X lo4 cells/well), in 1 .Oml of Eagle’s minimum essential medium containing 10% FCS, 100 U/ml penicillin, and 100 pg/ml streptomycin (MEM). They were used between passages 8 and 15. Numbers of lymphocytes and fibroblasts were determined by direct counting in a haemocytometer, in the latter case after detachment of the cells from the monolayer using 0.25% trypsin (19).

Lymphocyte Binding Assay Lymphocyte adhesion to fibroblasts was measured quantitatively using [3H]DNAprelabeled cells (17). These were prepared by the addition of [methyl-3H]thymidine ([3H]TdR; 0.5 &i/ml, 40 Ci/mmol) to the cultures for 24 hr prior to harvesting the cells, after which they were washed four times with fresh medium and resuspended at a density of 5 X lo6 cells/ml in MEM buffered with 20 mM Hepes, pH 7.4. An aliquot (200 ~1) of this suspension was added to the confluent fibroblast monolayers and incubated for 2 hr at 37°C unless indicated otherwise. The nonadherent cells were carefully removed by aspiration, and the fibroblast monolayers and attached lymphocytes were washed five times with phosphate-buffered saline (PBS) to remove remaining unbound cells. The adherent cells were solubilized with 500 ~1 of lysis buffer (5 mMTris-HCl, pH 7.4,200 mMNaCl,O.5% Triton X-100, and 0.2% SDS), and the radioactivity that remained associated with the fibroblast monolayer was measured by scintillation counting. The adhesion index (AI) is equivalent to the number of lymphocytes bound per fibroblast cell, calculated on the basis of the initial specific radioactivity of the lymphocytes which was measured separately at the beginning of each experiment. This value is shown as the arithmetic mean of five separate determinations. In each experiment the binding of the cells to the culture vessel, in the absence of the fibroblast monolayer, was subtracted from that obtained in their presence.

Treatment of Lymphocytes with Trypsin and Preparation of Released Peptides The procedure for trypsin treatment was similar to that reported by Woodruff et al. (28). Briefly, Con A-activated lymphocytes were prepared from murine spleen, as described above, and 1 X 10’ cells/ml were incubated with trypsin (20 pg/ml in PBS) or with trypsin and soybean trypsin inhibitor (200 pg/ml), for 5 min at 37°C. After incubation, the trypsin inhibitor was added to all the cell suspensions, which were then washed twice with fresh MEM and collected by centrifugation. The cells were resuspended to a concentration of 1 X 10’ cells/ml and aliquots were maintained for 3 hr at 4°C or recultured at 37°C either alone or in the presence of cycloheximide (CHX; 0.1 mM), an inhibitor of protein synthesis. After this time the cells were washed, resuspended to 5 X lo6 cells/ml, and immediately used in the binding assay.

36

ABRAHAM

ET AL.

The direct effect of CHX on lymphocyte attachment was also assessed by adding it directly to the adhesion assay. The supernatant from lo9 trypsin-treated Con A-activated T-lymphocytes was clarified by centrifugation at 15,000 rpm for 10 min at 4°C. Trypsin was removed by affinity chromatography using soybean trypsin inhibitor coupled to agarose. The peptides released from the lymphocyte cell surface by trypsin treatment did not bind to the column and were dialyzed exhaustively against 5 mMTris-HCl, pH 7.4. SDSPAGE analysis showed that this preparation was composed of a heterogeneous mixture of plasma membrane-derived peptides in the molecular weight range between 20 and 90 kDa. Approximately 250 pg of protein was obtained from lo9 cells. Preparation

and Analysis of Lymphocyte

Plasma Membranes

Plasma membranes were purified from lysates of lo9 Con A-activated T-lymphocytes by centrifugation on a discontinuous sucrose gradient as previously described (29). This preparation contained the marker enzymes, 5’nucleotidase and alkaline phosphodiesterase, at specific activities between 20 and 30 times higher than that of the original cell homogenate (30), and was shown by electron microscopy to consist of vesicles between 200 and 800 nm diameter, surrounded by characteristic trilaminar membranes (3 1). Approximately 2 mg of protein was obtained from lo9 cells. Preparation PM)

andAnalysis

ofAntibody

against Lymphocyte Plasma Membranes

(Anti-

Antibody was raised in rabbits against the purified plasma membranes of activated murine T-lymphocytes using standard procedures, and the IgG fraction was obtained by affinity chromatography on Protein A-Sepharose. Fab fragments were prepared by digestion of 20 mg of IgG protein with 12.5 U of papain-agarose in 100 mM sodium phosphate, pH 6.5, containing 2 mMEDTA and 10 mM cysteine, for 16 hr at 37°C. After centrifugation, the supematant was diluted with 5 vol of PBS and the Fc component removed using Protein A-Sepharose. SDS-PAGE analysis of the unbound material showed that there were no remaining IgG heavy chains that could be detected by staining with Coomassie blue. Immunofluorescence using the anti-PM was carried out by coupling the IgG directly to FITC at alkaline pH (32). Excess fluorochrome was removed by dialysis against PBS. Lymphocytes (2 X lo6 cells) were incubated at 4°C with the FITC-conjugated anti-PM (100 &ml) in 0.5% bovine serum albumin (BSA) in PBS for 30 min. The cells were washed three times with PBS, resuspended in 1% paraformaldehyde for 15 min, and washed again with PBS. After mounting in 90% glycerol, they were examined under a Zeiss photomicroscope III fitted with an epifluorescence condenser IIIRS and a 25X objective. For immunoelectron microscopy of the anti-PM, the lymphocytes were fixed in 0.5% glutaraldehyde and embedded in Lowicryl K4M, as previously described (3 1). Ultrathin sections (80- 100 nm) were cut using a Reichart ultramicrotome and incubated with anti-PM in PBS containing 20% normal goat serum, 0.1% gelatin, 1% BSA, and 1% Tween 20. Detection was carried out by the application of goat antirabbit IgG conjugated to 10 nm colloidal gold and visualized under a Phillips 300 electron microscope.

LYMPHOCYTE-FIBROBLAST

INTERACTION

37

The effects of anti-PM on lymphocyte adhesion were examined by adding the antibody (250 pg/ml) to the lymphocyte suspension, for 15 min prior to the binding assay. It was then also added back directly to the assay, at the same final concentration.

Identijcation of Lymphocyte Binding Using SpecificAntibodies Rat monoclonal antibodies (MAbs) were used for T-cell phenotyping by indirect immunofluorescence, with either fluorescein- or rhodamine-conjugated rabbit antirat immunoglobulin as the secondary antibody. The MAbs were NIMR-1, specific for the anti-Thy-l antigen of most T-lymphocytes (33); and YTS 169.4 and YTS 19 1.1 (kindly provided by Dr. B. Chain, University College, London), specific for the Lyt-2 (CDs) and L3T4 (CD4) antigens characteristic of suppressor/cytotoxic and helper/inducer T-lymphocytes, respectively (34). Polyclonal FITC-conjugated antimurine IgM was used to identify B-lymphocytes.

Efects of Monoclonal Antibodies on Lymphocyte Adhesion The following MAbs were used to examine the role of specific lymphocyte cell surface/adhesion proteins: anti-mu&e Class II (Mouse Ia), M5/114.15.2 (35); antimurine T3 complex, 145-2C 11 (36) a generous gift of Professor J. Owen, University of Birmingham; anti-LFA- 1, FD44 1.8 (37); anti-LFA- 1 (a-subunit), M 17/4.2, and anti-LFA- 1/Mac- 1 (&subunit), M 18/2.a.8 (38); anti-Mac- 1 (a-subunit), M l/ 70.15.11.5 (39); anti-gp80, 5D2-27, the major cell adhesion glycoprotein of certain strains of mice (40); and MEL-14 (4 1). In the case of MAb 5D2-27, binding was assessed using Con A-activated spleen cells from strain C57/B6, since the antigen which it recognizes, gp80, is a major cell surface glycoprotein of this strain but not of strain CBA. Like the anti-PM these MAbs were also added (250 pg/ml) to the lymphocytes for 15 min before, and their presence was maintained during, the binding assay.

Preparation and Use of Fucoidan Commercially available fucoidan was obtained and purified as described by Brandley et al. (42), except that gel filtration was carried out on a HPLC TSK G3000 SW column (7.5 mm X 60 cm) using water as the mobile phase, at a flow rate of 1 ml/min ( 15 bar; 225 psi). The high-molecular-weight carbohydrate fraction obtained was found to be free of contaminating protein. It was dialyzed extensively against distilled water, lyophylized, and stored at -20°C until use. The effects of fucoidan, L- and D-fUCOSe, mannose 6-phosphate, and mannan on lymphocyte binding were measured by adding these components, at the concentrations indicated, to the lymphocyte suspensions 15 min prior to the binding assay. They were then added back to the assay at the same concentration.

Adhesion of Lymphoid Cell Lines The binding of the lymphoid cells was carried out in the same way as that of the lymphocytes. However, because some of the cell lines (BW 5 147, SP2/0, NSl, and Ag8) are deficient in thymidine kinase, they were radiolabeled by the addition of [methyl-3H]thymidine 5’-monophosphate (1 &i/ml, 40 Ci/mmol) to the culture medium for 24 hr. The adhesion ratio (AR) is the ratio of the binding of the lymphoid

38

ABRAHAM

ET AL.

Treatment

ElX Adhesion

None

2030 105060708090 1001 0 X of control binding

1. Effect of trypsin on lymphocyte adhesion. Con A-activated T-cells were prepared from murine spleen as described under Materials and Methods, and their attachment to fibroblasts was measured (1) without trypsin treatment (control adhesion), (2) following trypsin treatment of the lymphocytes, (3) following trypsin treatment in the presence of trypsin inhibitor (TI), (4) after trypsin-treated lymphocytes were recultured at 37”C, (5) after trypsin-treated lymphocytes were recultured at 4°C (6) after trypsintreated lymphocytes were recultured at 37°C in the presence of cytoheximide (CHX), and (7) without trypsin treatment, but in the presence of CHX. In the control binding assay(1, above), 5.9 + 1.1 lymphocytes adhered per fibroblast cell. FIG.

cell lines relative to that of the Con A-activated splenic lymphocytes, defined as equivalent to 1.O. Experiments in which the effects of anti-PM and fucoidan on the attachment of lymphoid cells were examined were carried out in the same way as those for the lymphocytes. RESULTS

Trypsin Sensitivity of Lymphocyte Adhesion Treatment of human, mouse, and rat lymphocytes with low concentrations of trypsin has previously been shown to prevent their attachment to endothelial cells (28, 43). Figure 1 shows that trypsin pretreatment of Con A-activated T-lymphocytes from murine spleen also markedly inhibited their binding to fibroblasts (1.5 f 0.3 Tcells/fibroblast, band 2, compared with 5.9 & 1.1 untreated, control lymphocytes/ fibroblast, band 1). When the cells were simultaneously exposed to trypsin in the presence of excess trypsin inhibitor (band 3), binding to the fibroblast monolayer remained at the control level (6.0 f 1.3 T-cells/fibroblast, band 1). Lymphocytes treated with trypsin and then recultured in the absence of trypsin for 3 hr at 37°C (band 4), but not at 4°C (band 5), recovered nearly 70% of their adhesion capability. Partial recovery of lymphocyte adhesiveness also occurred when they were recultured for 3 hr at 37°C in the presence of cycloheximide, an inhibitor of protein synthesis (2.6 f 0.7 T-cells/fibroblast, band 6). The presence of this reagent during the binding assay, at levels which blocked protein synthesis (>95% inhibition within 5 mm), had no effect on lymphocyte attachment (band 7).

Eflects ofAntibody against the Lymphocyte Plasma Membrane The addition to the binding assay, of increasing amounts of peptides released by treatment of the activated T-lymphocytes with trypsin (data not shown) or of the

LYMPHOCYTE-PIBROBLAST

39

INTERACTION

A.

B.

X of control odherion 110

X of control adhesion 120

100 -

100

90 60

60 80 70 -

60

60 -

10

50 20

40 30

, , , , , , , , , , , 0.00 0.01 0.06 0.12 0.16 0.20 Plormo membrane(mg/ml)

0 0.0

0.2

0.1

0.6

0.6

1.0

Anti-PM (mg/ml)

PIG. 2. Effects of plasma membranes and anti-plasma membrane antibody (anti-PM) on adhesion of activated T-lymphocytes. (A) Increasing concentrations of the lymphocyte plasma membranes were added to the Con A-activated lymphocytes from murine spleen, for 15 min at 37-C, and the subsequent adhesion of the lymphocytes was measured, as described under Materials and Methods. (B) As in (A), using increasing concentrations of the anti-PM. The control adhesion is the binding of untreated activated T-lymphocytes (6.6 per fibroblast cell), which is taken as 100%.

purified T-lymphocyte plasma membranes themselves (Fig. 2A), progressively prevented the attachment of the intact activated T-cells to the fibroblasts. Maximum inhibition was obtained by adding 0.20 mg of plasma membrane protein, equivalent to the amount present in lo* cells. Polyclonal antibody raised against the trypsin-released peptides or against the purified plasma membranes of Con A-activated T-lymphocytes intensely stained more than 95% of the cells in cultures of the murine splenic lymphocytes incubated with Con A (Fig. 3A). Anti-PM was similarly reactive with LPS-stimulated B-lymphocytes, but the antibody did not cross-react with either human lymphocytes or human fibroblasts (data not shown). The localization of the fluorescence to the plasma membrane of the murine cells was confirmed by immunoelectron microscopy, which showed antibody reactivity only at the cell surface, on microvilli, and in some sections in cytoplasmic vesicles just below the plasma membrane (Fig. 3B). Like the plasma membranes, the addition of increasing amounts of the anti-PM also progressively inhibited T- and B-lymphocyte adhesion to the fibroblasts. Figure 2B shows that the binding of the activated T-cells was reduced approximately 50% by 0.2 mg of antiPM (from 6.6 to 3.1 lymphocytes per fibroblast), and more than 70% by 0.5 mg of the antibody. The attachment of the activated B-cells was similarly affected, being reduced, in the presence of 0.5 mg of anti-PM, to 35% of the control value of 17.3 lymphocytes per fibroblast. Fab fragments of anti-PM were as effective as the intact IgG in inhibiting both T- and B-lymphocyte adhesion (data not shown). Role of Specific Cell Surface Proteins in Lymphocyte Adhesion A number of monoclonal antibodies, previously shown to react with plasma membrane proteins involved in lymphocyte adhesion to other types of cell, were also ex-

40

ABRAHAM

ET AL.

FIG. 3. Immunomicroscopy of the lymphocyte cell surface. (A) Immunofluorescence light microscopy of intact Con A-activated murine T-lymphocytes using the rabbit anti-PM as the primary antibody and an anti-rabbit FITC-conjugated secondary antibody. Similar surface staining was also observed with LPSactivated murine B-cells (not shown). Bar, 15 pm. (B) Immunogold electron microscopy of Lowicrylembedded sections of the Con A-activated T-lymphocyte sections using the anti-PM as the primary antibody and a goat anti-rabbit IgG conjugated to 10 nm gold as the secondary antibody. Bar, 0.2 pm.

amined for their possible inhibitory effect on the attachment of lymphocytes to fibroblasts. Anti-Thy- 1, anti-Lyt-2 (CD8), and anti-L3T4 (CD4) all produced readily detectable lymphocyte immunofluorescence when they were applied to Con A-activated cells previously bound to fibroblast monolayers. However, the attachment of the Con A-activated T-lymphocytes was not mediated by the respective antigens Thy1, CD8, and CD4, since the addition of each of the MAbs to the binding assay itself did not decrease the number of lymphocytes which adhered in the absence of the

LYMPHOCYTE-FIBROBLAST

41

INTERACTION

TABLE 1 Effects of Antibodies against Lymphocyte Surface Antigens on T-Lymphocyte Adhesion to Fibroblasts Surface antigens/ adhesion molecules Thy- 1 Lyt-2 (CD8) L3T4 (CD4) Class II MHC (Mouse Ia) T3 complex LFA- 1 LFA- 1 (a-chain) LFA- 1/Mac-l @-chain) Mac- 1 (a-chain) gp80 gp80-90 Plasma membranes of Con A-lymphocytes

Antibodies added to binding assay

Inhibition of lymphocyte binding”

NIMR- 1 YTS 169.4 YTS 191.1 M51114.15.2 145-2Cll FD441.8 Ml7j4.2 Ml8/2.a.S Ml/70.15.11.5 5D2-27 MEL-14

0 0 0 0 3 12 5 15 15 0 9

Anti-PM

64

u Inhibition of lymphocyte binding is expressed as the percentage inhibition of attachment of the Con A-activated T-lymphocytes in the presence of the respective antibodies. This was calculated relative to the attachment of the lymphocytes in the absence of any added antibody (5.7 lymphocytes per fibroblast cell in this experiment), which is taken as 100%.

MAbs (5.7 activated T-cells per fibroblast) (Table 1). Nor did T-cell attachment involve certain components of the mouse Class II MHC and T3 complex, surface antigens which are known to modulate T-cell function (36). There was also no inhibition of lymphocyte binding which resulted from the addition of MAbs prepared against LFA- 1 and Mac- 1, both involved in several contact-dependent reactions of leukocytes (44, 45), and against gp80, a major adhesion glycoprotein of some strains of mouse (40). Similarly, the MAb MEL- 14, which has been widely investigated because of the important role of this antigen in the adhesion of lymphocytes to peripheral lymph node endothelia (41, 46), also had no effect on lymphocyte attachment to fibroblasts (Table 1). As before, the addition of the polyclonal anti-PM markedly inhibited the binding of the activated T-lymphocytes by 64%. It also significantly reduced the attachment of activated B-cells (AIs of 19.1 and 6.1 in the absence and presence of anti-PM, respectively). In contrast, although FITC-conjugated anti&M reacted with nearly all LPS-activated lymphocytes which had previously bound to fibroblast monolayers, the addition of this antibody to the binding assay itself did not prevent the adhesion of the B-lymphocytes (93% of control binding in the presence of anti-&M). Role of Carbohydrate

Receptors in Lymphocyte Adhesion to Fibroblasts

Carbohydrate-binding receptors have also been implicated in cell-cell recognition, and in lymphocyte adhesion to endothelial cells (47). The attachment of cells expressing these types of receptors is inhibited in the presence of polysaccharides such as fucoidan (42) and phosphomannans (43,48). In the present study we have purified the carbohydrate component of fucoidan, a highly sulphated polymer of fucose, and show that it produced a dose-dependent decrease in attachment, inhibiting the bind-

42

ABRAHAM

ET AL.

X of control adhesion 140 , 120 ,*--ht--C--, c’ 100 - i::: - r.-*..&

3

.-a. ‘4.

-..!.a

80 -

I Al

h2 ,*\ I *I: I

60 40 -

2

20 -

1

0 0.1

1

5

10 25

Carbohydrate

50 100 250 (pg/ml)

FIG. 4. Effect of carbohydrates on T- and B-lymphocyte adhesion. Increasing amounts of fucoidan (1, 2) and mannose 6-phosphate (3, 4) were added to suspensions of Con A-activated T-lymphocytes (1, 3) and LPS-activated B-lymphocytes (2,4). The subsequent adhesion of the lymphocytes was then measured, as described under Materials and Methods. The control adhesion is the attachment of each type of lymphocyte in the absence of the carbohydrates (5.3 T-cells and 19.7 B-cells bound per fibroblast), which is taken as 100%.

ing of Con A-activated T-lymphocytes to the fibroblasts by more than 90% at 50 pg/ ml (Fig. 4). Similar results were obtained when it was added to binding assays using activated B-lymphocytes, although the attachment of these cells was reduced by a maximum of only 70% even when fucoidan was added at 250 hg/ml. The addition of either D- or L-fucose to the binding assay, at concentrations as high as 1 mg/ml, had no significant effect on T- or B-lymphocyte adhesion (94 and 98% of control binding, respectively). In contrast to the effects of fucoidan, the addition of increasing amounts of mannose 6-phosphate, which has previously been reported to be a potent inhibitor of lymphocyte adhesion reactions involving phosphomannosyl receptors (43,49) and the MEL-14 antigen (48), failed to prevent T- or B-lymphocyte attachment to fibroblasts, as shown in Fig. 4. The addition of mannan to the binding assay, at concentrations as high as 1.O mg/ml, also did not inhibit lymphocyte attachment to fibroblasts (9 1 and 94% of control binding of T- and B-lymphocytes, respectively).

Efects ofAnti-PM and Fucoidan on Adhesion of Lymphoid Cells A number of studies have shown that differences in specific surface components among various lymphoid cell lines determine their individual level of adhesion to endothelial cells (41, 50). The results in Table 2 show that these cells also differ in their level of attachment to fibroblasts in tissue culture, and that there is no direct relationship between binding affinity and immunological function. Thus, while some T-cell lines, pre-B-cells, and mature B-cells (for example, DO- 11.10, M 12, and SP2/ 0, respectively) were far more adherent than Con A-activated T-lymphocytes (i.e., had adhesion ratios much greater than 1.O), the binding of some cells in each of these functionally distinct groups (for example, EL4, A20, and A3, respectively) was similar to or lower than that of the Con A-lymphoblasts. Moreover, while the very high adhesion index of LPS-activated B-lymphocytes from murine spleen was much higher

LYMPHOCYTE-FIBROBLAST

43

INTERACTION

TABLE 2 Effects of Anti-PM and Fucoidan on Lymphoid Cell Adhesion to Fibroblasts Adhesion Lymphoid cell T-cells DO-1 1.10 EL4 KS BW 5 147 BA4 Pre-B-cells Ml2 TA3 A20 B-cells SP2/0 A3 HB32 NSl A@

Inhibition of binding”

Index

Ratio

+Anti-PM

+Fucoidan

(helper) (lymphoma) (suppressor) (thymoma) (cytotoxic)

21.8 4.3 7.3 13.1 5.3

3.1 0.6 1.1 1.9 0.9

31 61 22 48 -

66 3s 91 58 -

(lymphoma) (lymphoma) (lymphoma)

19.7 18.0 4.5

2.8 2.7 0.6

59 62

56 96 81

(myeloma) (hybridoma) (hybridoma) (myeloma) (myeloma)

22.3 8.1 14.8 11.2 10.3

3.1 1.2 2.1 1.6 1.6

62 64 83

80 77 84 82 7s

n The inhibition of adhesion of the lymphoid cells is expressed as the percentage reduction in the respective binding of each of the individual cell lines in the absence of anti-PM or fucoidan (the adhesion index, as shown in the column on the left).

than that of a human B-cell line, Wil-2 (AIs of 20.5 and 5.5, respectively), a human T-lymphoblastoid cell, Molt 4, also adhered strongly to fibroblasts (AI of 14.8). The presence of anti-PM affected the adhesion of all the murine T- and B-lymphoid cell lines, but the extent of inhibition varied widely between them, and also did not appear to be related to immune function, as shown in Table 2. However, although the binding of T-, pre-B-, and B-cells (e.g., EL4, M12, and Ag8) was very strongly inhibited by the anti-PM, the presence of fucoidan reduced the adhesion of only two of these cell lines, M 12 and Ag8, but it had little effect on EL4 (35% inhibition of the level of EL4 binding in the absence of anti-PM). Conversely, although fucoidan almost totally prevented the attachment of BW 5 147 and severely inhibited the binding of DO- 11.10 (9 1 and 66% inhibition, respectively, of the control binding of these cells in the absence of anti-PM), both of these T-cells were only slightly affected by the presence of anti-PM (22 and 3 1% inhibition of control binding, respectively). This antibody had no effect on the binding of the human T- or B-cell lines (data not shown). DISCUSSION Studies on the adhesive reactions of T- and B-lymphocytes have suggested that a number of distinct recognition systems exist which regulate the attachment of these cells to endothelial cells (2, 3, 50), and to other vascular as well as nonlymphoid tissues (6, 7). Recent studies have demonstrated that a structurally related group of lymphocyte surface receptors is responsible for all three functionally distinct types of

44

ABRAHAM

ET AL.

lymphocyte-high endothelial venule (HEV) interaction found in intestinal (mucosal) (46), peripheral lymph node (41), and the synovial membrane (51). Such interactions also appear to control lymphocyte migration (52), directing certain types of lymphocytes to specific tissue in viva (3), and the near absolute specificity shown by certain lymphoid cell lines for different types of HEV (50). Lymphocytes are also able to recognise fibroblasts (14), and the data reported in the present study suggest that at least two distinct types of cell surface receptors are involved in lymphocyte-fibroblast interaction. Inhibition of adhesion by both the anti-PM and the mild trypsin treatment shows that certain protein structures on the lymphocyte cell surface are in some way critically involved in binding to fibroblasts. Protease-sensitive components have also been shown to mediate rat and human lymphocyte binding to peripheral lymph node HEV and endothelial cell cultures (28, 53), and to be involved in the adhesion of mouse lymphocytes to PPME beads by the phosphomannosyl receptor (43). The protease-sensitive molecules are reexpressed at the lymphocyte cell surface after trypsin treatment even in the absence of protein synthesis, showing that intracellular pools of these proteins are present within the lymphocyte and able to recycle back to the plasma membrane and restore lymphocyte adhesion to fibroblasts. A number of specific lymphocyte surface proteins (e.g., LFA-1 and Mac-l) have previously been shown to play important roles in accessory mechanisms mediating and strengthening cell-cell interactions (10, 54). These membrane molecules have been shown, by blocking studies with MAbs, to be involved in lymphocyte cytotoxicity and natural killer cell attachment to target cells (45). Anti-LFA-1 has also been reported to inhibit the adhesion of T-lymphocytes to endothelial cells in culture (55), a process which requires divalent cations, unlike lymphocyte adhesion to fibroblasts. In the present study, MAbs against LFA- 1 and Mac- 1 had no effect on binding, suggesting that these antigens are not involved in the interaction between lymphocytes and fibroblasts. A number of other MAbs against cell surface proteins known to be involved in certain contact-mediated lymphocyte functions, such as Class II MHC and the T3 complex, were found to be equally ineffective. In marked contrast, lymphocyte adhesion to fibroblasts was potently inhibited by fucoidan, indicating that specific lymphocyte surface receptors are blocked by the fucose 4-sulphate polysaccharide. Adhesion of lymphocytes to fucoidan-derivatized gels via a fucoidan receptor has been reported (42) and exhibits characteristics similar to those demonstrated here and previously (17) to be involved in lymphocyte binding to fibroblasts, such as the independence of divalent cations and temperature dependence. Both fucoidan and phosphomannans have previously been shown, using different types of adhesion assay, to block lymphocyte adhesion to HEVs in vitro (42, 43,47,48), suggesting that lymphocyte adhesion to fibroblasts might also be mediated by similar receptors. However, while the results presented here showed that the attachment of the (murine) lymphocytes to the (human) fibroblasts involved the fucoidan receptor, as discussed above, it was also found that these lymphocytes did not recognize fibroblasts by the phosphomannosyl receptor, since neither mannan nor mannose 6-phosphate prevented lymphocyte attachment. This is consistent with our finding, using this same binding assay, that a MAb against the MEL-14 antigen, which has recently been shown to be closely related to the phosphomannosyl receptor (48), also had no effect on lymphocyte-fibroblast interaction. Moreover, in studies using the Woodruff tissue section assay, mitogenic activation of lymphocytes was

LYMPHOCYTE-FIBROBLAST

INTERACTION

45

recently found to markedly down-regulate the MEL-14 antigen (56), whereas the results of experiments using fibroblast monolayers, as in the present study, showed that the adhesion of activated lymphocytes was nevertheless substantially greater than that of resting cells ( 17) which express much higher levels of MEL- 14. It is also notable that lymphoid cell lines having only low levels of this surface protein (48), BW 5 147 and SP2/0, still strongly adhered to the fibroblasts. Taken together these results suggest that while the fucoidan receptor appears to play an important part in lymphocyte-fibroblast interaction, it is unlikely that the phosphomannosyl receptor is involved in this type of adhesive reaction. All the lymphoid cell lines examined were able to adhere to the fibroblast monolayers, although the adhesion index varied by more than threefold between them and did not appear to be related to their immunological function. Although the binding to fibroblasts of the helper T-cell DO- 11.10 was substantially greater than that of the suppressor T-cell, KS (AIs of 2 1.8 and 7.3, respectively), similar to the preferential adhesion of human Leu-3+ (helper/inducer) lymphocytes to endothelial cells (53) and keratinocytes (57), the latter was found to be dependent on the expression of the Class II MHC (HLA-DR) complex. Since cultured human fibroblasts express Class I but not Class II MHC antigens (58) it is unlikely that these mediate the interaction between specific types of lymphocytes and fibroblasts. The inhibition of attachment of all the lymphoid cell lines by anti-PM suggests that some common surface proteins are involved in the binding to fibroblasts. Differences in attachment in the presence of this antibody probably reflect the relative level of expression of the adhesive molecules in each of the lymphoid cells. Thus the low inhibition of binding of the suppressor T-cell line K2S (22%), compared with the Bcell myeloma Ag8 (83%) could be explained if K2S expressed four times the levels of adhesive molecules. In contrast, the binding of both K2S and Ag8 was mediated primarily by the distinct fucoidan receptor, since the presence of fucoidan inhibited the adhesion of these cells by 9 1 and 75%, respectively. EL4, on the other hand, was inhibited only 35% by fucoidan but by 67% by the anti-PM. These results indicate that binding to fibroblasts is mediated by multiple surface receptors which coexist on different types of lymphocytes, as has previously been suggested for the adhesion of specific populations of rat lymphocyte to endothelial cells (42). Although the present study has examined the interactions of murine lymphocytes and lymphoid cells with fibroblasts of human origin, other types of assays have shown that a number of properties of this adhesive process are similar to those which mediate the attachment of human lymphocytes to human fibroblasts (59). Moreover, it is notable that several adhesion proteins of mouse and human origin appear to be very closely related both structurally and functionally (45, 60, 6 1). The close homology of these proteins and their high degree of genetic conservation through evolution (38,49) is also suggested by the ready formation of interspecies heteropolymers (45). The precise molecular basis of lymphocyte interaction with fibroblasts is not yet clear. Adhesive events of this type appear to be of importance in wound healing (11) and certain types of fibroses (62), in which the migration of lymphocytes into the epidermis may be prompted by their attachment to certain fibroblast matrix components. Interactions between certain T-cells and fibroblasts are probably also important in inflammatory reactions of skin such as scleroderma, an autoimmune rheumatic disease (63-65). In tissue culture, cell-to-cell contact with fibroblasts has been shown to enhance the mobility (66) and certain metabolic activities (67) of lympho-

46

ABRAHAM

ET AL.

cytes. Interaction with lymphocytes has also been shown to induce accessory cell function and antigen presentation by the fibroblasts (12). This sometimes leads to cytotoxicity, perhaps involving the contact-mediated transfer of lysosomal enzymes from the lymphocytes, which has been shown to occur both in vitro (67, 68) and possibly in vivo (69,70). REFERENCES 1. Harlan, J. M., B/o&65,513, 1985. Jalkanen, S., Med. Biol. 65,223, 1987. 3. Woodruff, J. J., Clarke, L. M., andchin, L. H., Annu. Rev. Zmmunol. 5,201, 1987. 4. Cavender, D., Haskard, D., Yu, C-L., Iguchi, T., Miossec, P., Oppenheimer-Marks, N., and Ziff, M., Fed. Proc. 46, 163, 1987. 5. Behan, W. M. H., Behan, P. O., Durward, W. F., and McQueen, A., J. Neural. Neurosurg. Psychiatry 50,1468,1987. 6. Engel, A. G., and Arahata, K., Ann. Neural. 16,209, 1984. 7. Kawanishi, H., Arch. Pathol. Lab. Med. 101,286, 1977. 8. Poulter, L. W., Seymour, G. J., Duke, O., Janossy, G., and Panayi, G., Cell. Zmmunol. 74,358, 1982. 9. Temowitz, T., and ThestrupPederson, K., J. Invest. Dermatol. 87,6 13, 1986. 10. Shaw, S., GintherLuce, G. E., Quinones, R., Gress, R. E., Springer, T., and Sanders, M. E., Nature (London) 323,262, 1987. 11. Kanzler, M. H., Gorsulowsky, D. C., and Swanson, N. A., J Dermatol. Surg. Oncol. 12, 1156, 1986. 12. Umetsu, D. T., Katzen, D., Jabara, H. H., and Geha, R. S., J. Zmmunol. 136,440,1986. 13. Collins, T., Krensky, A. M., Clayberger, C., Fiers, W., Gimbrone, M. A., Burakoff, S. J., and Pober, J. S., J. Zmmunol. 133, 1878, 1984. 14. Piela, T. H., and Kom, J. H., Cell. Zmmunol. 114, 149, 1988. 15. Chang, T. W., Celis, E., Eisen, H. N., and Solomon, F., Proc. Natl. Acad. Sci. USA 76,29 17, 1979. 16. Olsen, I., Muir, H., Smith, R., Fensom, A., and Watt, D. J., Nature (London) 306,75, 1983. 17. Abraham, D., Muir, H., and Olsen, I., Zmmunology65,385, 1988. 18. Prieto, J., Beatty, P. G., Clark, E. A., and Patarvoyo, M., Immunology 63,63 1, 1988. 19. Olsen, I., Dean, M. F., Muir, H., and Harris, G., J. Cell Sci. 55,2 11, 1982. 20. White, J., Haskins, K. M., Marrack, P., and Kappler, J., J. Zmmunol. 130, 1033, 1983. 2 1. Jin Kim, K., Kanellopoulos-Langevin, C., Merwin, R. M., Sachs, D. H., and Asofsky, R., J. Zmmunol. 2.

122,549,1979. 22.

23. 24. 25. 26. 27. 28. 29.

30. 3 1. 32. 33. 34.

Glimcher, L. H., Hamano, T., Asofsky, R., Sachs, D. H., Pierres, M., Samelson, L. E., Sharrow, S. O., and Paul, W. E., J. Zmmunol. 130,2287, 1983. Ozato, K., Mayer, N., and Sachs, D. H., J. Zmmunol. 124,533, 1980. Shulman, M., Wilde, C. D., and Kiihler, G., Nature (London) 276,269, 1978. Taylor, P. M., Wraith, D. C., and Asknonas, B. A., Immunology 54,607, 1985. Klaus, G. G. B., Pepys, M. B., Kitajima, K., and Askonas, B., Immunology 38,687, 1979. Keamey, J. F., Radbruch, A., Liesegang, B., and Rajewsky, K., J. Zmmunol. 123, 1548, 1979. Woodruff, J. J., Katz, I. M., Lucas, L. E., and Stamper, H. B., J. Zmmunol. 119,1603, 1977. Abraham, D., Bou-Gharios, G., Muir, H., Olsen, I., Smith, R., and Winchester, B., Eur. J. Cell Biol. 40, 167, 1986. Olsen, I., Oliver, T., Muir, H., Smith, R., and Partridge, T., J. CellSci. 85,231, 1986. Bou-Gharios, G., Moss, J., Olsen, I., and Partridge, T., Histochemistry89,69, 1988. Brandtzaeg, P., Stand. J. Zmmunol. 2,273, 1973. Chayen, A., and Parkhouse, R. M. E., J. Zmmunol. Methods 49, 17, 1982. Cobbold, S. P., Jayasuriya, A., Nash, A., Prospero, T. D., and Waldmann, H., Nature (London) 312, 548,1984.

Bhattacharya, A., Dorf, M. E., and Springer, T. A., J. Zmmunol. 127,2488,1981. Leo, O., Foo, M., Sachs, D. H., Samelson, L. E., and Bluestone, J. A., Proc. Natl. Acad. Sci. USA 84, 1374, 1987. 37. Sarmiento, M., Dialynas, D. P., Lancki, D. W., Wall, K. A., Lorber, M. I., Loken, M. R., and Fitch, F. W., Zmmunol. Rev. 68, 135, 1982. 38. Sanchez-Madrid, F., Simon, P., Thompson, S., and Springer, T. A., J. Exp. Med. 158,586, 1983. 35. 36.

LYMPHOCYTE-FIBROBLAST 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 5 1. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.

INTERACTION

47

Springer, T. A., Galfre, G., Secher, D. S., and Milstein, C., Eur. J. Immunol. 9,301, 1979. Jacobson, K., O’dell, D., Holifield, B., Murphy, T. L., and August, J. T., J. Cell Biol. 99, 16 13, 1984. Gallatin, W. M., Weissman, I. L., and Butcher, E. C., Nature (London) 304,30, 1983. Brandley, B. K., Ross, T. S., and Schnaar, R. L., J Cell Biol. 105,99 1, 1987. Yednock, T. A., Stoolman, L. M., and Rosen, S. D., J. CellBiol. 104,713, 1987. Springer, T. A., Teplow, D. B., and Dreyer, W. J., Nature (London) 314,540, 1985. Springer, T. A., Dustin, M. L., Kishimoto, T. K., and Marlin, S. D., Annu. Rev. Immunol. 5, 223, 1987. Jalkanen, S., Bargatze, R. F., de 10sToyos, J., and Butcher, E. C., J. Cell Biol. 105,983, 1987. Stoolman, L. M., and Rosen, S. D., J. Cell Biol. 96,722, 1983. Yednock, T. A., Butcher, E. C., Stoolman, L. M., and Rosen, S. D., J. Cell Biol. 104,725, 1987. Stoolman, L. M., Yednock, T. A., and Rosen, S. D., Blood 70, 1842, 1987. Jalkanen, S., Steere, A. C., Fox, R. I., and Butcher, E. C., Science 233,556, 1986. Oppenheimer-Marks, N., and Ziff, M., Arthritis Rheum. 29,789, 1986. Oppenheimer-Marks, N., and Ziff, M., Cell. Immunol. 114,307, 1988. Masuyama, J-I., Minato, N., and Kano, S., J. Clin. Invest. 77, 1596, 1986. Hamann, A., Jablonski-Westrich, D., Duijvestijn, A., Butcher, E. C., Baisch, H., Harder, R., and Thiele, H-G., J. Immunol. 140,693, 1988. Mentzer, S. J., Burakoff, S. J., and Faller, D. V., J. Cell. Physiol. 126,285, 1986. Hamann, A., Jablonski-Westrich, D., Scholz, K-U., Duijvestijn, A., Butcher, E. C., and Thiele, H-G., J. Immunol. 140,737, 1988. Nickoloff, B. J., Lewinsohn, D. M., Butcher, E. C., Krensky, A. M., and Clayberger, C., J. Invest. Dermatol. 90, 17, 1988. Pober, J. S., Collins, T., Gimbrone, M. A., Cotran, R. S., Gitlin, J. D., Fiers, W., Clayberger, C., Krensky, A. M., Burakoff, S. J., and Reiss, C. S., Nature (London) 305,726, 1983. Piela, T. H., and Korn, J. H., Cell. Immunol. 114, 149, 1988. Clayton, L. K., Sayre, P. H., Novotny, J., and Reinherz, E. L., Eur. J. Immunol. 17, 1367, 1987. Sewell, W. A., Brown, M. H., Owen, M. J., Fink, P. J., Kozak, C. A., and Crumpton, M. J., Eur. J. Immunol. 17,1015,1987. Wahl, S. M., Wahl, L. M., and McCarthy, J. B., J. Immunol. 121,942, 1978. Alcocer-Varela, J., Laffon, A., Alracon-Segovia, D., and Ibanez de Kasep, G., J. Rheumatol. 11, 1, 1984. Alcocer-Varela, J., Martinev-Cordero, E., and Alarcon-Segovia, D., Clin. Exp. Immunol. 59, 666, 1985. Morrow, J., and Isenberg, D. (Eds.), “Autoimmune Rheumatic Disease,” Chap. 7, p. 256. Blackwell, Oxford, 1987. Arencibia, I., Pedari, L., and Sundqvist, K-G., Exp. CellRes. 172, 124, 1987. Olsen, I., Abraham, D., Shelton, I., Bou-Gharios, G., Muir, H., and Winchester, B., Biochim. Biophys. Acta968,312, 1988. Olsen, I., Dean, M. F., Harris, G., and Muir, H., Nature (London) 291,244, I98 1. Gasper, P. W., Thrall, M. A., Wenger, D. A., Macy, D. W., Ham, L., Domsife, R. E., McBiles, K., Quackenbush, S. L., Kessel, M. L., Gillette, E. L., and Hoover, E. A., Nature (London) 312, 467, 1984. Taylor, R. M., Farrow, B. R. H., Stewart, G. J., Healy, P. J., and Tiver, K., Transplant. Proc. 19,27X), 1987.