1164

EXPRESSION AND FUNCTION OF SURFACE ANTIGENS ON SCLERODERMA FIBROBLASTS D. ABRAHAM, S. LUPOLI, A. McWHIRTER, C. PLATER-ZYBERK, T. H. PIELA, J. H. KORN, IRWIN OLSEN, and C. BLACK Dermal fibroblasts from patients with systemic sclerosis (SSc) bound a much greater number of T lymphocytes than did normal dermal fibroblasts. Monoclonal antibodies (MAb) against classes I and I1 antigens of the major histocompatibility complex (MHC) and their receptors, CDS and CD4, had no effect on T cell interaction with SSc and normal cells, while MAb against lymphocyte function-associated antigen type 3 (LFA-3) and CD2 both strongly inhibited lymphocyte attachment. MAb against intercellular adhesion molecule type 1(ICAM-1) and LFA-1 also prevented binding of T lymphocytes, but had a more marked effect on adhesion to SSc fibroblasts than to normal fibroblasts; they also completely abolished the increased binding to From the Cell Enzymology Unit and the Division of Clinical Immunology, The Mathilda and Terence Kennedy Institute of Rheumatology, Hammersmith, London, the Department of Rheumatology, Royal Free Hospital Medical School, London, England, and the Veterans Administration Medical Center and University of Connecticut School of Medicine, Newington. Supported in part by the Arthritis and Rheumatism Council of Great Britain, by NIH grant AR-32343, by the US Department of Veterans Affairs, and by the Connecticut Chapter of the Arthritis Foundation. D. Abraham, PhD: Cell Enzymology Unit, Kennedy Institute of Rheumatology; S. Lupoli, MD: Department of Rheumatology, Royal Free Hospital Medical School; A. McWhirter, BSc: Department of Rheumatology, Royal Free Hospital Medical School; C. Plater-Zyberk, PhD: Division of Clinical Immunology, Kennedy Institute of Rheumatology; T. H . Piela, PhD: VA Medical Center and University of Connecticut School of Medicine; J. H. Korn, MD: VA Medical Center, University of Connecticut School of Medicine; Irwin Olsen, PhD: Cell Enzymology Unit, Kennedy Institute of Rheumatology; C. Black, MD, FRCP: Department of Rheumatology, Royal Free Hospital Medical School. Address reprint requests to Irwin Olsen, PhD, Cell Enzymology Unit, The Mathilda and Terence Kennedy Institute of Rheumatology, 6 Bute Gardens, Hammersmith, London W6 7DW, UK. Submitted for publication November 6, 1990; accepted in revised form April 2, 1991. Arthritis and Rheumatism, Vol. 34, No. 9 (September 1991)

fibroblasts treated with interleukin-la, tumor necrosis factor a,and interferon-?. No difference was found in the proportion of normal and SSc fibroblasts that expressed MHC classes I and I1 and LFA-3, but more SSc cells expressed ICAM-1, and at a higher level, than did normal fibroblasts. These results show that cultured SSc cells have elevated binding to T lymphocytes, which possibly results from expansion of a subset of fibroblasts that produces high levels of ICAM-1.

Scleroderma (systemic sclerosis; SSc) is a multisystemic connective tissue disease characterized by microvascular obliteration and increased deposition of collagen resulting in fibrotic lesions (1-3). Although the pathogenesis of SSc is still uncertain, this complicated disorder involves inflammation, autoimmune features, vascular damage, activation of fibroblasts, and complex interactions between cells and components of the extracellular matrix (4-6). Studies of SSc have focused mainly on the altered metabolic activities of several types of cell that have been implicated in the disease process and the expression of specific surface receptors (7). Fibroblasts from affected areas of the skin of patients with SSc have been shown to produce greater amounts of types I and I11 collagen and other matrix proteins than normal fibroblasts (8-11) and to have higher levels of messenger RNA (mRNA) of these components (12,13). Recent in situ hybridization experiments showed that the increased amounts of collagen mRNA were expressed by only a subpopulation of the fibroblasts that were often localized adjacent to dermal blood vessels within the affected skin (14,15). These were frequently surrounded by mononuclear cells, a

SSc FIBROBLAST SURFACE ANTIGENS well-recognized feature of the inflammatory process which is particularly prominent in the early stages of SSc. The pathogenesis of SSc may therefore be related to direct cellular interactions between a unique population of clonally expanded fibroblasts and certain T lymphocytes (16-19), and could also involve specific effects exerted by the production of inflammatory mediators (20-22). In the present investigation, we examined the adhesive reactions of normal and SSc dermal fibroblasts with T cells, the influence of cytokines on these reactions, and the role of specific cell surface antigens in this process.

MATERIALS AND METHODS Fibroblast cultures. Punch biopsies were taken from the clinically affected skin of the forearm of 12 patients with SSc (6 men and 6 women) and from the normal skin of the forearm of 12 healthy individuals (9 men and 3 women) used as controls. The control subjects were 2 6 4 9 years of age (average 41 years); the SSc patients were 42-68 years of age (average 51 years). All SSc patients had severe, diffuse disease of 1-3 years duration (average 2.3 years) at the time of biopsy. Their modified Rodnan skin scores were between 7 and 48 (average 26). Eight patients were undergoing treatment with D-penicillamine. To obtain fibroblasts from the biopsy samples, the tissue was cut into 1-mm3 pieces, placed in sterile plastic dishes, and cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal calf serum (FCS), 2 mM L-glutamine, 1 mM sodium pyruvate, 100 unitdm1 penicillin, 100 pg/ml streptomycin, 2.5 pg/ml amphotericin B, and 50 pg/ml gentamycin. After 2-3 weeks of incubation in a humidified atmosphere of 5% CO, in air, the outgrowths of fibroblasts had reached confluence, and were passaged by brief trypsin treatment (3 minutes at 37°C) and recultured in DMEM. All experiments used fibroblasts between passages 2 and 4, depending on the number of cells obtained initially from the tissue sample. In each experiment, all the cell lines were examined at the same time and under the same conditions of culture (e.g., cell density, passage, days after plating). Cytokine treatment. Fibroblasts were seeded onto 96-well plates and grown to confluence. Recombinant cytokines interleukin-la (IL-la), interferon-y (IFNy), and tumor necrosis factor a (TNFa), kindly provided by Professor Marc Feldmann (Charing Cross Sunley Research Centre, London, UK), were used at concentrations of 200 units/ml, 1,000 units/ml, and 500 units/ml, respectively. Cytokines were added to the plates and incubated for 16 hours at 3TC, after which the monolayers were washed and the cells used immediately in the experiments, as indicated. Antibodies. Monoclonal antibodies (MAb) used in this study were W6/32 (anti-major histocompatibility complex [anti-MHC] class I), L243 (anti-MHC class II), TS1/22 (anti-lymphocyte function-associated antigen type 1 [antiLFA-11 a chain), TS2/9 (anti-LFA-3), and TS2/18.1.1 (anti-

1165

CD2). These were obtained as hybridoma cell lines from the American Type Culture Collection (Rockville, MD). RRl/1 (anti-intercellular adhesion molecule type 1 [anti-ICAM-11) was kindly provided by Dr. T. Springer (Boston, MA). BR2 is reactive against a fibroblast surface antigen apparently identical to aminopeptidase N (CD13) (Piela TH, Korn JH: unpublished observations). A polyclonal goat anti-human type I collagen antibody (Southern Biotechnology, Birmingham, AL) was also used. Preparation of lymphocytes. Peripheral blood was obtained from normal individuals by venous puncture, and the mononuclear cells were separated on a Ficoll gradient. Resting T cells were obtained by depleting the suspension of B cells and monocytes using nylon-wool columns and plastic adhesion techniques (23). To obtain activated T lymphocytes, the initial suspensions of lymphocytes were cultured for 5-7 days in RPMI medium containing penicillin, streptomycin, glutamine, 5% FCS, 5 x lOP5M 2-mercaptoethanol, and 4 pg/ml of concanavalin A. During this period, the cell number increased only slightly, from 1 x 10" to between 1.2 X 10" and 1.5 X 10" per milliliter, but the proportion of lymphoblasts accounted for 75-85% of the population, as determined by microscopic examination. Lymphocyte binding: morphology. Normal and SSc fibroblasts were grown on glass coverslips placed in 35-mm culture dishes. Prior to reaching confluence (i.e., during active proliferation in the exponential phase of cell growth), the culture medium was removed and replaced with 1.O ml of a suspension containing 5 x lo6 activated T lymphocytes. The cultures were incubated for 60 minutes at 37"C, the nonadherent cells removed by extensive washing with phosphate buffered saline (PBS), and the attached cells and monolayer fibroblasts were fixed by placing the coverslips into methanol for 10 minutes at 20°C. They were then air dried and stained with Giemsa. Lymphocyte binding: quantitative assay. The adhesion of lymphocytes to normal and SSc fibroblasts was measured using a radioisotopic assay. The T cells were labeled with 5'Cr by incubating 5 to 10 x lo6 cells/ml with 500 pCi/ml of Na,5'Cr04 (Amersham International, Buckinghamshire, UK), for 90 minutes at 37"C, in complete RPMI. Cells were then washed 5 times with fresh, warm medium, and 100-pl aliquots containing 5 x lo5 cells were allowed to adhere to 6 replicate wells of confluent fibroblast monolayers, each well containing approximately 2 x lo4 cells, for 60 minutes at 37°C. 'The nonadherent lymphocytes were aspirated; the monolayers were carefully washed and then solubilized by adding 100 pl of 50 mM NaOH. Radioactivity associated with the monolayers, due to the presence of adherent radiolabeled T cells, was measured by gamma counting. The results are presented as the arithmetic mean (+SEM) of the replicate wells, which were used to calculate the actual number of adherent lymphocytes, based on the specific radioactivity of the lymphocyte suspension initially applied to the monolayers. In some experiments, MAb were added to the labeled T cells for 15 minutes prior to assay, remaining throughout the period of binding. In other experiments, the fibroblast monolayers were first pretreated with the MAb, and in some cases also with cytokines, which were then removed and replaced with fresh MAb and lymphocytes. MAb were used at a concen-

ABRAHAM ET AL

1166

tration between 10 pg/ml and 50 pg/ml. The results of these experiments are presented as the average of the percentage of change in lymphocyte binding compared with control cultures containing no antibody or cytokine, using at least 6 different SSc and normal fibroblast strains. Phenotypic characterization of fibroblast cell cultures. The expression of surface antigens on normal and SSc fibroblasts was examined by flow cytometry (FCM) in the following way. The fibroblasts were detached from the monolayer by treating with 20 mM EDTA, pH 7.4, for 20 minutes at 4°C. Approximately 1 x lo6 cells were incubated with 1-10 pg/ml of the primary antibody in 2% FCS in PBS, for 60 minutes at 4°C. After washing with PBS-FCS, the cells were incubated with a species-specific fluorescein isothiocyanate-labeled secondary antibody (10 pg/ml), for 30 minutes at 4°C. The cells were again washed with PBS, fixed with 1% paraformaldehyde in PBS, and the level of fluorescence was analyzed on a FACScan (Becton Dickinson, Twickenham, UK), collecting data from lo4 cells. Fibroblasts were gated on the basis of their size (forward light scatter) and granularity (side light scatter). Cells treated with only the secondary antibody were used to set the fluorescence boundary markers for positive cells, i.e., to obtain the specific fluorescence resulting from the application of the primary antibodies. The percentage of positive cells is presented as the arithmetic mean (-+SEM) of determinations using fibroblasts from 6 different SSc patients and 6 different normal individuals. The AFI is the average net fluorescence intensity per cell, and is also presented as the arithmetic mean (5SEM) of the AFI of at least 6 different SSc and normal fibroblast cultures analyzed at the same time for each primary antibody. Measurement of collagen secretion. The secretion of type I collagen by cultured fibroblasts was measured using an inhibition enzyme-linked immunosorbent assay (ELISA) (24). Ascorbate was first added to the cultures and incubated for 48 hours, removed, and then replaced with DMEM for a 24-hour incubation prior to assay. Aliquots of the media were removed, incubated with the anti-human type I collagen antibody, and then transferred to 96-well plates that had been coated with type I collagen. The amount of collagen secreted by the fibroblasts was determined by assessing the extent to which the fibroblast culture media inhibited antibody binding to the collagen-coated plates. A standard curve was obtained by determining the extent to which known amounts of type I collagen inhibited the binding of anticollagen antibody to the plates.

A

RESULTS Binding of T lymphocytes to fibroblasts in vitro. The adhesion of T lymphocytes was examined to determine whether there are any significant functional differences in leukocyte interactions with normal and SSc fibroblasts. When suspensions of T cells were incubated with subconfluent monolayers of normal and SSc fibroblasts, and the nonadherent cells removed, many more lymphocytes remained attached to the SSc fibroblasts than to the normal cells (Figure 1).

B Figure 1. Interaction of T cells with fibroblasts in vitro. Subconfluent monolayers of normal dermal fibroblasts (A) and systemic sclerosis (SSc) dermal fibroblasts (B)were grown on coverslips and incubated with activated T cells for 60 minutes, as described in Materials and Methods. Adherent cells were fixed, then stained with Giemsa. There are many more round, dark-stained lymphocytes attached to the SSc cells than to the normal fibroblasts.

S S c FIBROBLAST SURFACE ANTIGENS To measure the precise numbers of lymphocytes that adhered to the fibroblasts, resting and activated T cells previously radiolabeled with "Cr were used. Although the adhesion of the activated T lymphocytes to normal and SSc fibroblasts was much higher than that of the resting T cells, the binding of both types of lymphocyte to the sclerotic fibroblasts was approximately twice their level of binding to the normal cells (Figure 2). Thus, whereas an average of 1.06 (k0.06 SEM) and 2.7 (k0.15 SEM) resting and activated T cells, respectively, were bound per normal fibroblast, 2.25 ( k O . 13 SEM) and 5.7 (k0.22 SEM) lymphocytes adhered to each SSc cell (P< 0.005, by 2-tailed t-test). Specific adhesion pathways of SSc fibroblasts. In view of the enhanced level of adhesive interaction between SSc fibroblasts and T lymphocytes, we examined the role of a number of specific cell surface antigens which have previously been shown to be involved in this process. This was carried out using antibody-blocking experiments, which measured the extent of inhibition of lymphocyte binding to monolayers of SSc and normal fibroblasts when MAb against specific adhesion molecules were added to the binding assay. The results showed that antigens of the MHC classes I and I1 complexes, and their respective lymphocyte receptors CD8 and CD4, do not participate in T cell adhesion to either normal or SSc cells (Figure 3). In marked contrast, the addition of a MAb

-q 0

u

81

1167

mAb

3-6 Inhibition of Binding 0

20

40

MHC I

I

CD2 ....,.

CD3

1

Ip

5.70

& 0.22)

6

2.25 ( 0.13) i

Normal SSc

2.70

& 0.151

Normal

SSc

Fibroblast cells Figure 2. Measurement of T lymphocyte binding to dermal fibroblasts from normal subjects and from patients with systemic sclerosis (SSc). Resting and activated T lymphocytes were radiolabeled with 51Cr, and their adhesion to fibroblasts (6 monolayer cultures each) was measured as described in Materials and Methods, the average number of which was calculated from the initial specific radioactivity of the lymphocytes and direct counting of the number of fibroblasts per well (expressed as the arithmetic mean 2 SEM).

Figure 3. Adhesion molecules involved in lymphocytefibroblast interactions. Activated T lymphocytes were radiolabeled with "Cr and their binding to 6 monolayer cultures of normal (open bars) and scleroderma (solid bars) dermal fibroblasts was measured in the presence of various monoclonal antibodies (mAb), as described in Materials and Methods. Nonspecific antibody, produced by the Ag8 cell line, was used as the control. Values are the average of the percentage of inhibition of binding of the lymphocytes to the 6 different cultures of cells in the absence of any added antibody (mean 2.5 and 7.5 lymphocytes per normal and scleroderma fibroblast, respectively). The SEM in these experiments were