Enhanced TNF-R expression in RA synovial joints

Eur. J. Immunol. 1992. 22: 1907-1912

Fionula M. Brennann, Deena L. Gibbonso, Tracey Mitchelln, Andrew P. Copeo, Ravinder N. Mainin and Marc FeldmannO Charing Cross Sunley Research Centreo, London, and The Kennedy Institute of RheumatologyA, Bute Gardens, London

1907

Enhanced expression of tumor necrosis factor receptor mRNA and protein in mononuclear cells isolated from rheumatoid arthritis synovial joints* We previously proposed the hypothesis that the pro-inflammatory cytokine tumor necrosis factor-a (TNF-a) plays a pivotal role in the pathogenesis of rheumatoid arthritis (RA) based on our observations that it is the dominant inducer of interleukin-1 (IL-1) and granulocyte-macrophage colony-stimulating factor (GM-CSF) production in RA synovial joint mononuclear (MNC) cells in culture. Since TNF-a acts via two membrane receptors, we have extended those studies to investigate the distribution of the pS5 and p75 TNF receptors (TNF-R) in RA tissue. Surface receptor expression was quantitated by flow cytometry using monoclonal antibodies specific to the pSS (HTR-9) and the p75 (UTR-1) TNF-R. Both receptors were significantly increased on MNC isolated from the synovial membrane of RA patients compared to normal or RA peripheral blood MNC. Interestingly, the p7S TNF-R was increased both on large monocytic/macrophage-type cells and CD3+ lymphocytes. Furthermore, there was a significant increase in the proportion of CD3+ cells in RA synovial fluid expressing the p75 TNF-R, compared to matched peripheral blood MNC. In contrast to RA synovial MNC, p7S or pSS TNF-R expression was not significantly increased in osteoarthritis synovial MNC. In addition, Northern blot analysis indicated abundant expression of both pSS and p7S mRNA in RA synovial joint MNC. This was in contrast to normal peripheral blood MNC cells which contained little or no constitutive TNF-R mRNA; following stimulation with phytohemagglutinin and IL-2, a rapid and transient expression of both receptor mRNA was induced.These results, therefore, indicate that in RA synovial joint tissue there is up-regulation of both pSS and p7S TNF-R mRNA and surface protein expression, and with the presence of TNF-a in RA tissues, these results provide support to our hypothesis that TNF-a is of critical importance in the pathogenesis of RA.

1 Introduction

TNF-a and LT bind with equally high affinity to two distinct TNF receptors (TNF-R) which have been recently cloned Tumor necrosis factor-a (TNF-a) and lymphotoxin (LT) and expressed by a number of groups including ourselves are multipotent cytokines which are involved in a number [9, 10, 11, 12, 13l.These receptors have a relative molecuof diverse biological processes [I]. There are many reasons lar mass of S S kDa (pSS) and 75 kDA (p75) and both are why TNF-a has been implicated in the pathogenesis of members of a new receptor family which includes nerve rheumatoid arthritis (RA). It is a potent inducer of growth factor receptor and the B cell antigen CD40 [9]. cartilage-degradative enzymes such as collagenase [2] and Consistent with TNF's wide-range of activities, TNF-R are in bone culture systems, induces bone resorption [3].TNF-a expressed by most somatic cells [ 141.The cellular distribubioactivity has been found in RA synovial fluid 141, is tion of both receptors has recently been investigated using produced spontaneously in culture by RA synovial joint monoclonal antibodies specific for the 55-kDa and the cells [5, 61, and by immunostaining techniques [7] is 75-kDa TNF-R [15]. Cell lines of histiocytic lineage and localized to macrophages in the synovial tissue lining layer peripheral blood monocytes constitutively express both the and at sites of tissue destruction at the cartilage/pannus pS5 and p7.5 receptor [ 161whereas, several adenocarcinoma junction. Most telling, however, has been the effect of lines express only the p5S receptor. In contrast, resting neutralizing TNF-a in the rheumatoid joint cell cultures lymphocytes express few, if any,TNF-R and upon activation using anti-TNF-a antibody. In these cultures the production express predominantly the p75-receptor [15, 17, 181. NK of other pro-inflammatory cytokines such as IL-1 and cells express both the pS5 and p75 molecules, and both granulocyte-macrophage (GM)-CSF were abrogated [6, 81. receptors are involved in lymphokine-activated killer cell activity [19]. pSS and p75 TNF-R are independently [I 101791 regulated on cells. Thus, T cell-derived cytokines including I F N y 1201 and IL-2 [21] increase TNF-R expression on I This work was supported by the Arthritis and Rheumatism cells. Similarly, anti-y or staphylococcus aureus Cowan Council, Medical Resedrch Council and the Wellcome Founda- (SAC)-stimulated B cells markedly increases their p75 (but tlon not p55) receptor 1171. In contrast, activators of protein kinase C [22] LPS [23] and TNF-a itself 1241 induce Correspondence: Fionula M. Brennan, Charing Cross Sunley down-regulation of TNF-R on cell surface. As the level of Research Centrc, Lurgan Avenue, Hammersmith, London, W6 expression of TNF-R on cells is evidently central to the 8LW. GB action of TNF-a, we have investigated the expression of Abbreviations: RA: Rheumatoid arthritis OA: Osteoarthritis both receptors in synovial joint mononuclear cells derived MNC: Mononuclear cells from RA patients and compared the expression to matched 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1992

0014-2980/92/0707- 1907$3.50+ .25/0

1908

F. M. Brennan. D. L. Gibbons, T. Mitchell et a1

peripheral blood mononuclear cells (PBMC). The results show that there is increased TNF-R mRNA and surface protein expression in synovial joint mononuclear cells.This observation substantiates the hypothesis that disregulation of TNF-u and its receptor is of major importance to the pathogenic processes evident in RA.

2 Materials and methods 2.1 Tissue donors Surface TNF-R expression was determined on matched samples of peripheral blood and synovial h i d (SF) (n = 8) obtained from RA patients defined as ‘classical’or ‘definitive’ by the revised ARA criteria [2S] attending the Rheumatology Unit at the Charing Cross Hospital and on synovial membrane (SM) samples from both RA (n = 6) and osteoarthritic (OA, n = 5) patients obtained while patients were undergoing total joint replacement. Control samples of peripheral blood were obtained from healthy laboratory volunteers (t7 = 8). Due to limitations in sample size, separate samples of RA synovial fluid ( n = 6). RA SM (17 = 5) and OA SM ( n = 1) from the same clinic were used for RNA studies. For time-course experiments on mRNA and protein expression PBMC were isolated from plateletpheresis reridues from healthy donors (provided by the South London Blood Transfusion Centre, Tooting, GB).

Eur. J. Immunol. 1992. 22: 1907-1912

bodies not conjugated to a fluorochrome was incubated with 1: 100 dilution of PE-goat anti-mouse conjugate (Southern Biotechnology, Birmingham, AL) for 30 min at 4°C. In some cases, cells were further incubated with Leu-4-FlTC after blocking with 20 % normal mouse serum (Sigma).The cells were washed twice again, fixed with 1% formaldehyde, and analyzed using a FACScan flow cytometer (Becton Dickinson).TNF-R were analyzed on viable cells by gating on forward- and side-scatter, and calculated at mean FL2 (HTR-9 or UTR-1) units after subtraction of mean FL2 units (control antibody). Statistical differences between groups was determined by the Mann Whitney U test. The percentage lymphocytes positive for p7.5 TNFR was calculated on gated lymphocytes (determined by size on forward- and side-scatter), after subtracting background binding (control antibody). The proportion of p7S TNF-R+ CD3+ cells was calculated in a similar manner using two-color fluorescence to determine percentage CD3 and p75 positive cells. 2.4 RNA extraction and Northern blotting

Total RNA was prepared from PBMC, SF and SM cells after washing cells in PBS using guanidinium isothiocyanate and 10 kg or 15 pg of total RNA containing 0.5 pg ethidium bromide was separated by electrophoresis in 1% agarose gel containing 6 %, formaldehyde. RNA was transferred onto nylon membranes (Genescreen, Dupont, NEN, Boston, MA) by capillary blotting. Filters were prehybridized in a solution containing 50% formamide, 5X SSPE 7.SX 2.2 Tissue samples Denhardt’s solution and 1.50 pg/ml denatured yeast tRNA at 42 “C for 1 h.The p.55 TNF-R probe was a 1.3-kb Eco R l Blood and synovial fluids were collected in bottles, and fragment of pS.5 cDNA [9]. The p75 probe was a 600-bp mononuclear cells (MNC) isolated by FicoWHypaque Bgl I1 fragment of a cDNA clone isolated from a T cell centrifugation (specific density 1.077 g/ml). Synovial mem- library [26] using sequence-specific oligonucleotides and brane tissue was digested in RPMI 1640 (Gibco, Paisley, verified by DNA sequencing. Probes were labeled by Scotland) containing 5 %O FCS (Gibco), 5 mg/ml collagen- random oligonucleotide priming [27] using u-[”P] dCTP ase type IV (Sigma) and 0.15 mg/ml DNase type I (Sigma) (Amersham, Int. Amersham, GB), denatured and hybridas previously described [5, 61. PBMC, SM and SF mononu- ized to the filters for 16 h at 42°C. Filters were washed clear cells were cultured overnight in complete medium under high stringency conditions in 2X SSPE, 0.1 % SDS at which consisted of RPMI 1640, supplemented with 10 ‘70 room temperature, and twice in 0.1 x SSPE, 0.1 % SDS at (v/v) fetal calf serum and 2 mM I -glutamhe (Gibco) and 6.5 “C and exposed to Fuji X ray film at -70°C with then analyzed for TNF-R as described below. Mononuclear intensifying screens. cells isolated from plateletpheresis residues were cultured in complete medium (see above) containing 20 ng/ml IL-2 (Ajinomoto, Japan) and 1 pg/ml phytohemagglutinin. Cells were isolated at intervals up to 96 h for mRNA and 3 Results protein analysis of pSS and p7S TNF-R. Normal PBMC were cultured with PHA/IL-2 for 96 h to induce TNF-R expression. Surface TNF-R was determined 2.3 Monoclonal antibodies and fluorescence analysis using specific monoclonal antibodies (as described in Sect. 2.3) and analyzed by flow cytometry (Fig. la). Small The monoclonal antibodies used to detect the p.55 (HTR-9) amounts of both receptors were present on normal PBMC and the p75 (UTII-1) TNF-R were kindly provided by Dr. cells at time zero; after activation levels of both receptors Manfred Brockhaus (Roche, Basel, Switzerland). OX-20 increased on the cell surface reaching a maximum after 48 h (mouse anti-rat IgG,) which served as a control antibody in culture and declining slowly thereafter. Induction of was kindly provided by Dr. Don Mason (Oxford, GB). TNF-R expression was confirmed by extraction of RNA FTTC-Leu-4 (anti-CD3, Becton Dickinson, Cowley, GB) from PBMC during the culture period followed by Norwas used to determine the proportion of CD3+ T lympho- thern blot hybridization using pS.5 and p75 cDNA probes. cytes. For each analysis, 5 x los cells (PBMC, SF or SM PHA/IL-2 activation rapidly induced a transient expression MNC) were incubated for 30 min at 4°C with optimal of mRNA for both TNF-R (p75 more abundant than p55) concentrations (10 pg/ml) of the monoclonal antibodies with similar kinetics. Fig. 1b illustrates induction of cordescribed above and washed in PBS supplemented with rectly sized p75 transcript (4.5kb; pS.5 2.6kb, not shown) 2 YO fetal calf serum and 0.01 % azide. Monoclonal anti- reaching a maximum at 8 h and then declining.

Enhanced TNF-R expression in RA synovial joints

Eur. J. Immunol. 1992. 22: 1007-1912

In contrast to resting PBMC, RNA isolated from RA synovial fluid cells (Fig. 2a) hybridized strongly on Northern blots to both p55 and p75 TNF-R cDNA probes. Actin RNA served as a control for RNA loading. RNA isolated from RA and OA SM was also tested to determine

10

0 8 24 48 72 96 Time in culture (hours)

28s p75 TNF-R 0 2 4 8 24 48 hours

Figure 1. (a) p7S and p55 TNF-R induction on PBMC with PHA/IL-2. PBMC were cultured in complete medium with PHA/IL-2 ;IS dcscribcd in Sect. 2.2. Cells were harvested at the times indicated m d analyzed for TNF-R expression using monoclonal antibodies HTR-9 (anti-p55, hatched bars) and UTR-1 (antikp75, solid bars) detected with PE-goat anti-mouse and analyzed by flow cytometry. Results are expressed as mean FL2 units (after subtraction of control antibody staining. (b) Northern analysis of p75 TNF-R induction by PHA/IL-2. PBMC were culturcd in 50-ml flasks at 1 X 10h cells/ml in RPMI11640 containing 10 % FCS 20 ng/ml IL-2 and 1 pg/ml PHA. RNAwas harvested at the times indicated and 15 ~g of total cellular RNA separated on a denaturing agarose gel, transferred to a nitrocellulose filter and hybridized with a p7.5 probc.The gel was photographed under UV light prior to hybridization to indicate RNA loading.

123456

123456

-P%

1 2 3 4 5 6

amount of TNF-R mRNA present. After electrophoresis the gel was photographed under ultraviolet illumination to determine RNA loading. The filters were then hybridized withTNF-R probes. Correctly sized transcripts (4.5 kb and 2.6 kb) for both receptors were present (Fig. 2b). p75 transcript was less abundant than p55 mRNA. To determine whether TNF-R protein was also increased, p55 and p75 TNF-R in PBMC was compared to MNC isolated from the blood, SF, and SM of patients with RA, and SM of patients with OA. p55 TNF surface receptor (Fig. 3a) was significantly increased on RA synovial membrane MNC (mean = 13.8 k 16, pC0.036) compared to normal PBMC (mean = 1.3 k 1.9). There was no significant increase of p55 surface TNF-R expression on RA PBMC, (1.0 k 1.1) or RA SF MNC (1.9 k 4.7). OA synovial membrane MNC had increased, but not to significance, levels of p55 TNF-R (mean 8.6 k 15). p75 TNF surface receptor (Fig. 3b) was also significantly increased on RA SM MNC (mean = 21.2 k 8.6, p