Interleukin 8 in rheumatoid arthritis
Eur. J. Immunol. 1990. 20: 2141-2144
2141
Short paper Fionula M. Brennan, Claus 0. C. ZachariaeO, David Chantry, Christian G. LarsenO, Martin Turner, Ravinder N. Mainia, Kouji MatsushimaO and Marc Feldmann Charing Cross Sunley Research Centre, London, National Cancer Instituteo, Biological Response Modifiers Program, NCI-Frederick Cancer Research Facility, Frederick and Kennedy Institute of Rheumatologya, London
Detection of interleukin 8 biological activity in synovial fluids from patients with rheumatoid arthritis and production of interleukin 8 mRNA by isolated synovial cells* The presence of neutrophils in the synovial joint of patients with rheumatoid arthritis (RA) is thought to be due to the activity of chemotactic factors released by activated cells in the joint.We have shown in this report, for the first time, the abundance of one such factor, interleukin 8 (IL 8), in the synovial fluid of patients both with RA and other n o n - U joint diseases, and the spontaneous production of IL8 mFWA by RA synovial cells in culture.There was no correlation between the levels of chemotactic activity and IL8 protein, suggesting that other factors with similar neutrophil chemotactic activity are also present in the synovial fluid exudate. In support of this concept neither the level of chemotactic activity nor IL 8 protein levels correlated with neutrophil or leukocyte infiltration, indicating that the mechanism of migration into the inflammatory environment of the joint is complex. Such migration is likely to be due to a number of chemotactic signals in addition to IL8, which may either synergize with, or inhibit, the action of IL 8.
1 Introduction The presence of a large number of infiltrating mononuclear cells (MNC) and polymorphonuclear leukocytes (PMN), the latter concentrating in the synovial fluid (SF), is a common feature in the synovial joint of patients with rheumatoid arthritis (RA).The infiltrating cells are thought to contribute to the pathology of the disease through the release of degradative enzymes, prostaglandins, reactive oxygen species [ 11 and pro-inflammatory cytokines such as IL 1 [2] and TNF-a [3]. The migration of such cells into the joint is unlikely to be a passive process and it is thought to be mediated through the activity of chemotactic factors released by the activated cells in the joint.
TNF-a is chemotactic in vitro for neutrophils [5]. This apparent paradox was resolved by the discovery of a novel factor, monocyte-derived neutrophil chemotactic factor (MDNCF; [6]; now termed IL8) inducible by IL1 which was directly chemotactic for neutrophils in vitro [6-81. IL8 belongs to a large family of low (8-10 kDa) molecular mass proteins which have now been cloned and include the IFN-y-induced protein (yIP10; [9]), the murine MQ inflammatory proteins MIP 1and MIP 2 [lo], the gro gene product also known as melanoma growth stimulatory activity (MGSA), P-thromboglobulin [ l l ] and platelet factor 4 [lo]. I L 8 is also known as NAP (neutrophil-activating peptide; [12], NAF (neutrophil-activating factor; [ 13]), GCP (granulocyte chemotactic peptide; [14]) and TCF (T lymphocyte chemotactic factor; [15]).
Neutrophils show chemotaxis in response to a number of agents, including small peptides, complement components The gene for IL 8 has been cloned and IL 8 has been shown and cytokines. Based on studies in animals in which t o be produced by a number of cell types (including neutrophils accumulated at the site of injection of IL 1[4] or monocytes, T cells, fibroblasts, endothelial cells and keraTNF-a [5], it was originally thought that I L 1 and TNF-a tinocytes) following stimulation by a variety of prowere responsible for mediating neutrophil chemotaxis. inflammatory cytokines including IL 1 and TNF-a [16]. However, it has became clear that neither pure rIL1 or Many of these cell types are present in an activated state in the RA synovial membrane. In addition we have shown that IL 1and TNF-a (mRNA and protein) are produced at high levels by cells isolated from the synovial membrane of individuals with RA [17, 181. Thus, we were interested in [I 83251 determining whether if I L 8 (mRNA and bioactivity) was present in RA synovial cells and secreted into the SF. The * This work was funded by the Arthritis and Rheumatism Council, presence of this molecule could be an important factor in Xenova and the Sunley Trust. attracting both neutrophils and Tcells, the most abundant Correspondence: Fionula M. Brennan, Charing Cross Sunley cells infiltrating into the fluid and membrane, respectively, Research Centre, 1 Lurgan Avenue, Hammersmith, London to the site of inflammation. The influx of these cells is a critical step in the pathogenesis of RA and hence deterW68LW, GB mining the nature of the signals recruiting these cells is of importance . Abbreviations: RA: Rheumatoid arthritis SF: Synovial fluid 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
+
0014-2980/90/0909-2141$3.50 .25/0
2142
Eur. J. Immunol. 1990. 20: 2141-2144
F. M. Brennan, C. 0. C. Zachariae, D. Chantry et al.
2 Materials and methods 2.1 Patients The clinical data from the patients used in this study is summarized in Table 1. SF were obtained by arthrocentesis of knee joints of patients with an acute flare on a background of chronic disease. They were obtained from ten patients with RA (mean age 55 f 20 years) in whom the disease had lasted for at least 2 years. Eight patients in this group were RF+ (seropositive) and satisfied the criteria for classical disease as defined by the American Rheumatism Association [19], the remaining two patients were seronegative. In addition SF was obtained from nine patients with other arthritides including psoriatic arthritis (n = 2), polyarthritis (n = l ) , monoarthritis (n = 3), gonococcal arthritis (n = l), gout (n = 1)and osteoarthritis (n = l).The mean age in this group was 50 f 13 years. The drug regimen in the RA group consisted of non-steroidal anti-inflammatory drugs (NSAID) and/or steroids (prednisolone) and other disease-modifying drugs (methotrexate, azathioprine or gold). The non-RA group were either untreated or had been treated solely with NSAID.
samples (RA5 and RA6) by layering onto Lymphoprep followed by centrifugation as described previously [17].The MNC were then washed twice and resuspended in complete medium (RPMIplus 10% FCS). MNC were also prepared from two RA synovial membranes (obtained during total hip replacement surgery) for RNA analysis using enzymatic digestion of the tissue as previously described [17]. Total cellular RNA was extracted as described below.
2.2 mRNA analysis
Total RNA was prepared from MNC isolated from SF (n =2), synovial membrane ( n = 2 ) and as a positive control, the monocytic cell line THP-1 stimulated with PMA and LPS. CESS cells, a B lymphoblastoid cell line, served as a negative control. Guanidinium isothiocyanate lysis was followed by cesium chloride gradient ultracentrifugation. RNA was quantitated by absorbance at 260 nm, and 20 pg RNA was electrophoresed through 1% agarose gels containing 6% formaldehyde, with the addition of 10 ng/ml ethidium bromide to the running buffer. Following electrophoresis gels were photographed under ultraSFsamples were initially spun twice at 1000 x g to pellet the violet illumination, to ensure integrity of the RNA, and to cells and debris. The fluid was removed and stored at ensure equal loading as assessed by intensity of staining of -70°C until assayed for IL8 bioactivity. Prior to mononu- 18S and 28 S rRNA. RNA was transferred onto nitrocelluclear cell separation a differential cell count was performed lose by capillary blotting. Filters were baked for 2 h at 80°C on the sample usingTurk's stain (0.03% crystal violet acetic and prehybridized in a solution containing 50% formamacidPBS) to determine total white cell number and the ide, 5 X SSC (1 X SSC = 150 mM NaCl, 15 mM trisodium proportion of those cells which are neutrophils. MNC were citrate, pH 7.0), 7.5 x Denhardt's solution (1 x Denseparated from the PMN for mRNA analysis of two SF hardt's = 0.05% Ficoll400, 0.05% polyvinylpyrolidine, 0.05% BSA), 50 mM phosphate buffer (pH 6.6), 0.5 mg/ml denatured salmon sperm DNA. The cDNA probe for IL8 was a 125-bp Pst I-Eco RI fragment corresponding to the Table 1. Clinical data of patientsa) majority of the coding region of IL8 [7] and was labeled by random oliconucleotide priming [20] to a high specific activity (108-109 cpndpg DNA) using [a32P]dCTP(AmerNo. Disease Age Drugs sham Int., Amersham, GB). Hybridization was for 16-24 h (Years) at 42 "C in prehybridization solution. Filters were washed 63 1 RA+ NSAID twice in 2 x SSC, 0.1% SDS at room temperature, and once 2 RA+ 84 NSAID + prednisolone in 0.2 x SSC, 0.19'0 SDS at 50°C then exposed to Fuji X-ray 3 RA+ 47 NSAID + gold film at -70 "C with intensifying screens. 4 RA+ 70 NSAID 5 6 7 8 9 10 11
12 13 14 15
16 17 18 19
RA+ RA+ RA+ RA+ RA-
39 66
RA-
26
Psoriactic Psoriactic Polyarthritis Monoarthritis Monoarthritis Monoarthritis Gon-arthritis Gout OA
58
58
80 25
37 32 57 58
30 60 66 55
Methotrexate NSAID + prednisolone Azathioprine + prednisolone NSAID NSAID + prednisolone NSAID NSAID NSAID NSAID No treament No treatment NSAID NSAID Voltarol N o treatment
a) SF samples from patients with FL4 (n = 10) or a spectrum of other n o n - U joint disease ( n = 9)were used in this study. Drug treatment at time of exudate sample is documented. FL4 patients were treated with NSAID, disease-modifying drugs such as methotrexate, azathioprine, gold or steroids such as prednisolone. The non-RA patients were either untreated or treated with NSAID.
2.3 Detection of neutrophil chemotactic activity in SF
Neutrophil chemotactic activity was detected in conditioned medium using an in v i m assay, performed as described in detail previously [5]. Briefly, 1 ml of SFdiluted 1/5 in 0.05 M Tris-HC1 buffer was applied to a 4 x 10 mm heparin-Sepharose column (heparin-Sepharose CL-6B, Pharmacia, Uppsala, Sweden) equilibrated with 0.05 M Tris buffer, pH 7.5. After washing extensively IL8 was eluted with 3 ml 0.5 M NaCl in Tris buffer, and dialyzed against PBS overnight using a membrane with a molecular weight cut off of 3500. Data are shown as a chemotactic index, determined as mean number of cells per high-power field (performed six times) in the presence of the sample, divided by the number of cells migrated in the presence of medium alone. Chemotactic index of 1indicates that there is no chemotactic activity as the number of cells migrating towards the sample equals the number of cells migrating towards the medium.
Eur. J. Immunol. 1990. 20: 2141-2144
2.4 RIA determination of ILB in SF A rabbit was immunized with rIL 8 ([21]; 100 pg, four times a week over a 4-week period). Rabbit antiserum (1/450) and heparin-Sepharose-eluted samples (1/10) were mixed to a final volume of 500 pl in PBS with 0.5% BSA. After incubating for 30 min at room temperature 1251-labeledIL 8 (1251-IL8, 15000 cpm) was added to each sample and incubated overnight at 4 "C. Iron-conjugated goat antirabbit IgG (500 pl; Advanced Magnetics Inc., Cambridge,MA) was added to each sample. After further incubation for 10 min the samples were pelleted using a magnetic separation unit (Advanced Magnetics Inc.) and washed twice in washing buffer. The amount of rabbit antiserum and 1251-IL8used was calculated to give a nonspecific binding of between 3%-5% and a specific binding of between 30% -50%. The titration curve for IL 8 was established between 10 pg/ml and 100 ng/ml. All determinations were done in duplicate.
3 Results Chemotactic activity was detected in all SF both in patients with classical RA and other non-RA joint arthritides (Table 2) and is expressed as chemotactic index (CI) the conventional means of assessing neutrophils chemotaxis. There was no significant difference in the mean levels of IL8 activity between the RA group (CI = 2.9 f 1.8) and non-RA group (CI = 2.4 f 1.3). In SF samples from both groups there was a large number of infiltrating leukocytes, and of these a significant proportion were neutrophils (assessed as described in Sect. 2.3 methods). Although the bioassay measures neutrophil chemotactic activity, there
h
h
Interleukin 8 in rheumatoid arthritis
2143
Table 2. Presence of IL8 protein and neutrophil chemotactic activity in synovial fluidsa) No.
No. of cells per ml
No. of neutrophil cells per rnl
CI
IL 8 (ndm1)
1.8 x 106 0.3 x 106 3.2 x I@ 3.0 X 106 1.5 x 106 NAb) 1.5 x 106 NA 0.4 x 106
2.0 f 0.3 1.3 f 0.3 3.4 0.4 1.0 f 0.1 2.1 t 0.2 1.9 f 0.2 3.0 f 0.6 5.8 f 0.4 2.7 ? 0.4 6.2 f 0.8
0.96 1.18 1.17 1.60 1.72 0.83
1.8 x 106 NA 0.1 x 106 0.03 x 106
1.6 f 0.4 1.0 t 0.4 2.5 t 0.5 3.2 t 1.0 2.0 0.2 1.6 f 0.3 2.2 f 0.8 1.9 f 0.4 5.5 f 1.0
1.35 1.30 0.98 3.55 1.20 3.10 2.00 8.80 6.20
RA 1 2 3 4 5 6
I 8 9 10
Non-RA 11 12 13 14 15 16 17 18 19
6 3 8 6 2
x106 x106 xl06 X106 Xl06
x106 3 X l B 2 X1B 1 x106 Coagulated 4
3 x106 8 xloh 0.5 x 106 0.2 x 106 Acellular 0.6 x 106 NA NA 3 x106
-
-
0.15 x 106 NA NA 0.9 x l@
*
*
4.05 3.20 0.92 3.50
SF samples from patients with RA (n = 10) or other non-RA joint disease (n = 9) were tested for the presence of IL8 bioactivity (see text), results are expressed as chemotactic index (number of migrated cells with sample at a MOO dilution divided by the number of migrated cells with assay medium alone) determined by counting six high-power (10 x 40) fields. SF were also tested for IL 8 protein using a specific RIA. Results are expressed as protein nglml. Total cell composition of fluid was determined as described in the text and is expressed as cells/ml SF (total leukocyte number) and the proportion of those cells that are neutrophils. NA = Data not available.
was no correlation between the CI and the number of neutrophils in each sample (r = 0 . 1 2 3 ; ~> O.l), or the total leukocyte number (r = 0 . 2 4 9 ; ~>O.l).The fluids were also tested for IL 8 protein using a specific RIA. There was no difference in the mean levels of I L 8 protein between the RA group (1.9 f 1.2 ng/ml) and non-RA group (2.3 f 2.6 ng/ml). There was no significant correlation (r = 0.367, p > 0.1) between the chemotactic activity assessed as chemotactic index and I L 8 protein (ng/ml).
Figure I. Total RNA was isolated from RA SF MNC (SF1 = RA5, SF2 = RA6) RA synovial membrane cells (SM1, SM2 from two RA patients different from those analyzed for chemotactic activity in SF) and the cell IinesTHP-1 (positive control), and CESS (negative control) as described in Sect. 2.2. After electrophoresis the gels were photographed under ultraviolet illumination (A and C) t o ensure integrity and equal loading of RNA as assessed by 18S and 28 S rRNAintensity, [(A) THP-1, SF1, SF2; (C) SM1, SM2, CESS]. After transfer to nitrocellulose, filters were probed for IL 8 mRNA as described in Sect. 2.2. (B) shows correctly sized I L 8 mRNA in stimulated THP-1 cells, and in freshly isolated SF cells. (D) shows IL8 mRNA in synovial membrane cells but not in the negative control (CESS cells).
Total cellular RNA was isolated from RA SF (SF1 = RA5, SF2 = RA6) and synovial membrane (SM1, SM2) to determine whether IL8 mRNA was being produced. The results are illustrated in Fig. 1. Following electrophoresis the gel was photographed under ultraviolet illumination [(A) THP-1, SF1, SF2; (C) SM1, SM2 CESS] to determine the amount of RNA present. The filters were then hybridized with the IL8 cDNA probe as described in Sect. 2.2.THF'-1 cells which had been activated with 10 pglml LPS and 50 ng/ml PMA for 3 h, served as a positive control, and CESS cells containing no IL 8 mRNA served as a negative control. It can be seen that correctly sized IL8 transcripts (1.8 kb) were present in cells obtained both from both SF (B) and synovial membrane (D).
2144
Eur. J. Immunol. 1990. 20: 2141-2144
E M. Brennan, C. 0. C. Zachariae, D. Chantry et al.
4 Discussion The accumulation of PMN and lymphocytes within the synovial joint of patients with RA has been well documented. The migration of these cells into the joint is believed to be due to the production of chemotactic factors; however, the identity of such factor(s) and their production within the joint has not previously been addressed.We have attempted to investigate this by examining the expression of IL8, a cytokine identified principally on the basis of its chemotactic activity for neutrophils. Our results demonstrate for the first time the presence of IL 8 bioactivity in the SFof patients with RA and the expression of IL 8 mRNA by freshly isolated synovial cells, suggesting a role for IL8 in the accumulation of leukocytes in this disease. Although Ph4N could be readily detected in most of the SF analyzed it is noteworthy that there was no direct correlation between the levels of I L 8 detected and the number of PMN in the SEThe reason for this is unclear; however, it is now known that there are a large number of low-molecular weight molecules which are structurally related to IL 8 and several have been shown to have chemotactic activity. For example the gro gene product, also known as MGSA [ l l ] has recently been shown to be produced by fibroblasts and endothelial cells following stimulation with IL 1 or TNF-a [22], and has neutrophil chemotactic activity equivalent to IL8 (C. C. Zachariae and K. Matshushima, unpublished observations). It is likely that mediators such as this may also be expressed as a part of the chronic inflammatory process in RA, and may also act as chemoattractants thus contributing to the chemotactic activity of the SF. IL 8 is produced by a range of cell types following stimulation with LPS and more importantly by I L 1 and TNF-a [8,16]. Both I L1 and TNF-a have been shown to be present in SF [2,3] and we have demonstrated that freshly isolated synovial cells produce these mediators [ 171. Synovial cells consist of a heterogeneous population [18], many of which have been shown to be capable of making IL8 following stimulation including T cells, monocytes and resident cells within the membrane such as endothelial cells and fibroblasts. I L 1 and TNF-a have been shown to be strong inducers of IL8 in vitro and other related factors.We have recently demonstrated that TNF-a is the major stimulus driving IL 1 production in RA [18] thus it is likely that TNF-a and/or I L 1 are involved in the production of IL 8 in the synovial joint. It has been recently shown using a receptor competition assay that many of the neutrophil chemotactic factors inducible by IL 1and TNF-a, including groMGSA and MIP I1 bind to the IL 8R (K. Matshushima, unpublished observations), with different specific activities. Thus the relative contribution of such factors chemotactic for neutrophils in the pathogenesis of RA is likely to be complex. Experiments are in progress to ascertain the importance of such factors in addition to IL 8 in this process
and the relative importance of I L 1 and TNF-a in their induction in v i v o . I Walker . for collecting the tissue We would like to thank Ms. . samples, and Ms. G . Harris and Ms. K . Hartley for processing the synovial tissue and synovial fluids.
Received February 5, 1990; in revised form May 21, 1990.
5 References 1 Dayer, J.-M. and Demczuk, S., Springer Semin. lmmunopathol. 1984. 7: 387. 2 Fontana, A., Hengartner, H., Weber, E., Fehr, K., Grob, l? J. and Cohen, G., Rheumatol. lnt. 1982. 2: 49. 3 DiGiovine, F. S., Nuki, G. and Duff, G. W., Ann. Rheum. Dis. 1988. 47: 768. 4 Granstein, R. D., Margolis, R., Mizell, S. B. and Sauder, D. N., J. Clin. Invest. 1986. 77: 1020. 5 Yoshimura, T., Matshushima, K., Oppenheim, J. J. and Leonard, E. J., J. Immunol. 1987. 139: 786. 6 Yoshimura,T., Matshushima, K.,Tanaka, S., Robinson, E . A., Appella, E., Oppenheim, J. J. and Leonard, E. J., Proc. Natl. Acad. Sci. USA 1987. 84: 9233. 7 Matshushima, K., Morishita, K., Yoshimura, T., Lavu, S., Kobayashi, W., Lew, E. , Appella, E., Kung, H., Leonard, E. and Oppenheim, J. J., J. Exp. Med. 1988. 167: 1883. 8 Gronhoj Larsen, C., Anderson, A . O., Oppenheim, J. 0. and Matshushima, K., Immunology 1989. 68: 31. 9 Luster, A. D. and Ravetch, J. V., J. Exp. Med. 1987. 166: 1084. 10 Wolpe, S. D. and Cerami, A., FASEB J. 1989. 4: 2565. 11 Richmond, A., Balentien, E., Thomas, H. G., Flaggs, G., Barton, D. E., Spiess, J., Bordoni, R., Francke, U. and Derynck, R., EMBO J. 1988. 7: 2025. 12 Schroder, J. M., Mrowietz,V., Morita, E. and Christophers, E., J. Immunol. 1988. 139: 3474. 13 Peveri, l?, Walta, A., Dewald, B. and Baggiolini, M., J. Exp. Med. 1988. 167: 1547. 14 Van Damme, J. ,Van Beeumen, J., Opdenakker, G. and Billiau, A., J. Exp. Med. 1988. 167: 1364. 15 Gronhoj Larsen, C., Anderson, A. O., Appella, E., Oppenheim, J. 0. and Matshushima, K., Science 1989. 243: 1464. 16 Strieter, R. M., Kunkel, S. L., Showell, H. J., Remick, D. G., Phan, S. H.,Ward, P A . and Marks, R. M., Science 1989.143: 1467. 17 Buchan, G., Barrett, K. ,Turner, M., Chantry, D., Maini, R.N. and Feldmann, M., Clin. Exp. lmmunol. 1988. 73: 449. 18 Brennan, F. M., Chantry, D., Jackson, A., Maini, R.N. and Feldmann, M., Lancet 1989. ii: 244. 19 Silman. A. J.. Br. J. Rheumatol. 1988. 27: 341. 20 Feinberg, A. and Vogelstein, B., Anal. Biochem. 1983.132: 6. 21 Furuta, R.,Yamagishi, J., Kotani, H., Sakamoto, F., Fukui,T., Matsui,Y., Sohmuraa,Y ,Yamada, M.,Yoshimura,T., Larsen, C. G., Oppenheim, J. J. and Matshushima, K., J. Biochem. 1989. 106: 436. 22 Wen, D., Rowland, A. and Derynck, R., EMBO .I. 1989. 8: 1761.
k
Eur. J. Immunol. 1990. 20: 2141-2144
2141
Short paper Fionula M. Brennan, Claus 0. C. ZachariaeO, David Chantry, Christian G. LarsenO, Martin Turner, Ravinder N. Mainia, Kouji MatsushimaO and Marc Feldmann Charing Cross Sunley Research Centre, London, National Cancer Instituteo, Biological Response Modifiers Program, NCI-Frederick Cancer Research Facility, Frederick and Kennedy Institute of Rheumatologya, London
Detection of interleukin 8 biological activity in synovial fluids from patients with rheumatoid arthritis and production of interleukin 8 mRNA by isolated synovial cells* The presence of neutrophils in the synovial joint of patients with rheumatoid arthritis (RA) is thought to be due to the activity of chemotactic factors released by activated cells in the joint.We have shown in this report, for the first time, the abundance of one such factor, interleukin 8 (IL 8), in the synovial fluid of patients both with RA and other n o n - U joint diseases, and the spontaneous production of IL8 mFWA by RA synovial cells in culture.There was no correlation between the levels of chemotactic activity and IL8 protein, suggesting that other factors with similar neutrophil chemotactic activity are also present in the synovial fluid exudate. In support of this concept neither the level of chemotactic activity nor IL 8 protein levels correlated with neutrophil or leukocyte infiltration, indicating that the mechanism of migration into the inflammatory environment of the joint is complex. Such migration is likely to be due to a number of chemotactic signals in addition to IL8, which may either synergize with, or inhibit, the action of IL 8.
1 Introduction The presence of a large number of infiltrating mononuclear cells (MNC) and polymorphonuclear leukocytes (PMN), the latter concentrating in the synovial fluid (SF), is a common feature in the synovial joint of patients with rheumatoid arthritis (RA).The infiltrating cells are thought to contribute to the pathology of the disease through the release of degradative enzymes, prostaglandins, reactive oxygen species [ 11 and pro-inflammatory cytokines such as IL 1 [2] and TNF-a [3]. The migration of such cells into the joint is unlikely to be a passive process and it is thought to be mediated through the activity of chemotactic factors released by the activated cells in the joint.
TNF-a is chemotactic in vitro for neutrophils [5]. This apparent paradox was resolved by the discovery of a novel factor, monocyte-derived neutrophil chemotactic factor (MDNCF; [6]; now termed IL8) inducible by IL1 which was directly chemotactic for neutrophils in vitro [6-81. IL8 belongs to a large family of low (8-10 kDa) molecular mass proteins which have now been cloned and include the IFN-y-induced protein (yIP10; [9]), the murine MQ inflammatory proteins MIP 1and MIP 2 [lo], the gro gene product also known as melanoma growth stimulatory activity (MGSA), P-thromboglobulin [ l l ] and platelet factor 4 [lo]. I L 8 is also known as NAP (neutrophil-activating peptide; [12], NAF (neutrophil-activating factor; [ 13]), GCP (granulocyte chemotactic peptide; [14]) and TCF (T lymphocyte chemotactic factor; [15]).
Neutrophils show chemotaxis in response to a number of agents, including small peptides, complement components The gene for IL 8 has been cloned and IL 8 has been shown and cytokines. Based on studies in animals in which t o be produced by a number of cell types (including neutrophils accumulated at the site of injection of IL 1[4] or monocytes, T cells, fibroblasts, endothelial cells and keraTNF-a [5], it was originally thought that I L 1 and TNF-a tinocytes) following stimulation by a variety of prowere responsible for mediating neutrophil chemotaxis. inflammatory cytokines including IL 1 and TNF-a [16]. However, it has became clear that neither pure rIL1 or Many of these cell types are present in an activated state in the RA synovial membrane. In addition we have shown that IL 1and TNF-a (mRNA and protein) are produced at high levels by cells isolated from the synovial membrane of individuals with RA [17, 181. Thus, we were interested in [I 83251 determining whether if I L 8 (mRNA and bioactivity) was present in RA synovial cells and secreted into the SF. The * This work was funded by the Arthritis and Rheumatism Council, presence of this molecule could be an important factor in Xenova and the Sunley Trust. attracting both neutrophils and Tcells, the most abundant Correspondence: Fionula M. Brennan, Charing Cross Sunley cells infiltrating into the fluid and membrane, respectively, Research Centre, 1 Lurgan Avenue, Hammersmith, London to the site of inflammation. The influx of these cells is a critical step in the pathogenesis of RA and hence deterW68LW, GB mining the nature of the signals recruiting these cells is of importance . Abbreviations: RA: Rheumatoid arthritis SF: Synovial fluid 0 VCH Verlagsgesellschaft mbH, D-6940 Weinheim, 1990
+
0014-2980/90/0909-2141$3.50 .25/0
2142
Eur. J. Immunol. 1990. 20: 2141-2144
F. M. Brennan, C. 0. C. Zachariae, D. Chantry et al.
2 Materials and methods 2.1 Patients The clinical data from the patients used in this study is summarized in Table 1. SF were obtained by arthrocentesis of knee joints of patients with an acute flare on a background of chronic disease. They were obtained from ten patients with RA (mean age 55 f 20 years) in whom the disease had lasted for at least 2 years. Eight patients in this group were RF+ (seropositive) and satisfied the criteria for classical disease as defined by the American Rheumatism Association [19], the remaining two patients were seronegative. In addition SF was obtained from nine patients with other arthritides including psoriatic arthritis (n = 2), polyarthritis (n = l ) , monoarthritis (n = 3), gonococcal arthritis (n = l), gout (n = 1)and osteoarthritis (n = l).The mean age in this group was 50 f 13 years. The drug regimen in the RA group consisted of non-steroidal anti-inflammatory drugs (NSAID) and/or steroids (prednisolone) and other disease-modifying drugs (methotrexate, azathioprine or gold). The non-RA group were either untreated or had been treated solely with NSAID.
samples (RA5 and RA6) by layering onto Lymphoprep followed by centrifugation as described previously [17].The MNC were then washed twice and resuspended in complete medium (RPMIplus 10% FCS). MNC were also prepared from two RA synovial membranes (obtained during total hip replacement surgery) for RNA analysis using enzymatic digestion of the tissue as previously described [17]. Total cellular RNA was extracted as described below.
2.2 mRNA analysis
Total RNA was prepared from MNC isolated from SF (n =2), synovial membrane ( n = 2 ) and as a positive control, the monocytic cell line THP-1 stimulated with PMA and LPS. CESS cells, a B lymphoblastoid cell line, served as a negative control. Guanidinium isothiocyanate lysis was followed by cesium chloride gradient ultracentrifugation. RNA was quantitated by absorbance at 260 nm, and 20 pg RNA was electrophoresed through 1% agarose gels containing 6% formaldehyde, with the addition of 10 ng/ml ethidium bromide to the running buffer. Following electrophoresis gels were photographed under ultraSFsamples were initially spun twice at 1000 x g to pellet the violet illumination, to ensure integrity of the RNA, and to cells and debris. The fluid was removed and stored at ensure equal loading as assessed by intensity of staining of -70°C until assayed for IL8 bioactivity. Prior to mononu- 18S and 28 S rRNA. RNA was transferred onto nitrocelluclear cell separation a differential cell count was performed lose by capillary blotting. Filters were baked for 2 h at 80°C on the sample usingTurk's stain (0.03% crystal violet acetic and prehybridized in a solution containing 50% formamacidPBS) to determine total white cell number and the ide, 5 X SSC (1 X SSC = 150 mM NaCl, 15 mM trisodium proportion of those cells which are neutrophils. MNC were citrate, pH 7.0), 7.5 x Denhardt's solution (1 x Denseparated from the PMN for mRNA analysis of two SF hardt's = 0.05% Ficoll400, 0.05% polyvinylpyrolidine, 0.05% BSA), 50 mM phosphate buffer (pH 6.6), 0.5 mg/ml denatured salmon sperm DNA. The cDNA probe for IL8 was a 125-bp Pst I-Eco RI fragment corresponding to the Table 1. Clinical data of patientsa) majority of the coding region of IL8 [7] and was labeled by random oliconucleotide priming [20] to a high specific activity (108-109 cpndpg DNA) using [a32P]dCTP(AmerNo. Disease Age Drugs sham Int., Amersham, GB). Hybridization was for 16-24 h (Years) at 42 "C in prehybridization solution. Filters were washed 63 1 RA+ NSAID twice in 2 x SSC, 0.1% SDS at room temperature, and once 2 RA+ 84 NSAID + prednisolone in 0.2 x SSC, 0.19'0 SDS at 50°C then exposed to Fuji X-ray 3 RA+ 47 NSAID + gold film at -70 "C with intensifying screens. 4 RA+ 70 NSAID 5 6 7 8 9 10 11
12 13 14 15
16 17 18 19
RA+ RA+ RA+ RA+ RA-
39 66
RA-
26
Psoriactic Psoriactic Polyarthritis Monoarthritis Monoarthritis Monoarthritis Gon-arthritis Gout OA
58
58
80 25
37 32 57 58
30 60 66 55
Methotrexate NSAID + prednisolone Azathioprine + prednisolone NSAID NSAID + prednisolone NSAID NSAID NSAID NSAID No treament No treatment NSAID NSAID Voltarol N o treatment
a) SF samples from patients with FL4 (n = 10) or a spectrum of other n o n - U joint disease ( n = 9)were used in this study. Drug treatment at time of exudate sample is documented. FL4 patients were treated with NSAID, disease-modifying drugs such as methotrexate, azathioprine, gold or steroids such as prednisolone. The non-RA patients were either untreated or treated with NSAID.
2.3 Detection of neutrophil chemotactic activity in SF
Neutrophil chemotactic activity was detected in conditioned medium using an in v i m assay, performed as described in detail previously [5]. Briefly, 1 ml of SFdiluted 1/5 in 0.05 M Tris-HC1 buffer was applied to a 4 x 10 mm heparin-Sepharose column (heparin-Sepharose CL-6B, Pharmacia, Uppsala, Sweden) equilibrated with 0.05 M Tris buffer, pH 7.5. After washing extensively IL8 was eluted with 3 ml 0.5 M NaCl in Tris buffer, and dialyzed against PBS overnight using a membrane with a molecular weight cut off of 3500. Data are shown as a chemotactic index, determined as mean number of cells per high-power field (performed six times) in the presence of the sample, divided by the number of cells migrated in the presence of medium alone. Chemotactic index of 1indicates that there is no chemotactic activity as the number of cells migrating towards the sample equals the number of cells migrating towards the medium.
Eur. J. Immunol. 1990. 20: 2141-2144
2.4 RIA determination of ILB in SF A rabbit was immunized with rIL 8 ([21]; 100 pg, four times a week over a 4-week period). Rabbit antiserum (1/450) and heparin-Sepharose-eluted samples (1/10) were mixed to a final volume of 500 pl in PBS with 0.5% BSA. After incubating for 30 min at room temperature 1251-labeledIL 8 (1251-IL8, 15000 cpm) was added to each sample and incubated overnight at 4 "C. Iron-conjugated goat antirabbit IgG (500 pl; Advanced Magnetics Inc., Cambridge,MA) was added to each sample. After further incubation for 10 min the samples were pelleted using a magnetic separation unit (Advanced Magnetics Inc.) and washed twice in washing buffer. The amount of rabbit antiserum and 1251-IL8used was calculated to give a nonspecific binding of between 3%-5% and a specific binding of between 30% -50%. The titration curve for IL 8 was established between 10 pg/ml and 100 ng/ml. All determinations were done in duplicate.
3 Results Chemotactic activity was detected in all SF both in patients with classical RA and other non-RA joint arthritides (Table 2) and is expressed as chemotactic index (CI) the conventional means of assessing neutrophils chemotaxis. There was no significant difference in the mean levels of IL8 activity between the RA group (CI = 2.9 f 1.8) and non-RA group (CI = 2.4 f 1.3). In SF samples from both groups there was a large number of infiltrating leukocytes, and of these a significant proportion were neutrophils (assessed as described in Sect. 2.3 methods). Although the bioassay measures neutrophil chemotactic activity, there
h
h
Interleukin 8 in rheumatoid arthritis
2143
Table 2. Presence of IL8 protein and neutrophil chemotactic activity in synovial fluidsa) No.
No. of cells per ml
No. of neutrophil cells per rnl
CI
IL 8 (ndm1)
1.8 x 106 0.3 x 106 3.2 x I@ 3.0 X 106 1.5 x 106 NAb) 1.5 x 106 NA 0.4 x 106
2.0 f 0.3 1.3 f 0.3 3.4 0.4 1.0 f 0.1 2.1 t 0.2 1.9 f 0.2 3.0 f 0.6 5.8 f 0.4 2.7 ? 0.4 6.2 f 0.8
0.96 1.18 1.17 1.60 1.72 0.83
1.8 x 106 NA 0.1 x 106 0.03 x 106
1.6 f 0.4 1.0 t 0.4 2.5 t 0.5 3.2 t 1.0 2.0 0.2 1.6 f 0.3 2.2 f 0.8 1.9 f 0.4 5.5 f 1.0
1.35 1.30 0.98 3.55 1.20 3.10 2.00 8.80 6.20
RA 1 2 3 4 5 6
I 8 9 10
Non-RA 11 12 13 14 15 16 17 18 19
6 3 8 6 2
x106 x106 xl06 X106 Xl06
x106 3 X l B 2 X1B 1 x106 Coagulated 4
3 x106 8 xloh 0.5 x 106 0.2 x 106 Acellular 0.6 x 106 NA NA 3 x106
-
-
0.15 x 106 NA NA 0.9 x l@
*
*
4.05 3.20 0.92 3.50
SF samples from patients with RA (n = 10) or other non-RA joint disease (n = 9) were tested for the presence of IL8 bioactivity (see text), results are expressed as chemotactic index (number of migrated cells with sample at a MOO dilution divided by the number of migrated cells with assay medium alone) determined by counting six high-power (10 x 40) fields. SF were also tested for IL 8 protein using a specific RIA. Results are expressed as protein nglml. Total cell composition of fluid was determined as described in the text and is expressed as cells/ml SF (total leukocyte number) and the proportion of those cells that are neutrophils. NA = Data not available.
was no correlation between the CI and the number of neutrophils in each sample (r = 0 . 1 2 3 ; ~> O.l), or the total leukocyte number (r = 0 . 2 4 9 ; ~>O.l).The fluids were also tested for IL 8 protein using a specific RIA. There was no difference in the mean levels of I L 8 protein between the RA group (1.9 f 1.2 ng/ml) and non-RA group (2.3 f 2.6 ng/ml). There was no significant correlation (r = 0.367, p > 0.1) between the chemotactic activity assessed as chemotactic index and I L 8 protein (ng/ml).
Figure I. Total RNA was isolated from RA SF MNC (SF1 = RA5, SF2 = RA6) RA synovial membrane cells (SM1, SM2 from two RA patients different from those analyzed for chemotactic activity in SF) and the cell IinesTHP-1 (positive control), and CESS (negative control) as described in Sect. 2.2. After electrophoresis the gels were photographed under ultraviolet illumination (A and C) t o ensure integrity and equal loading of RNA as assessed by 18S and 28 S rRNAintensity, [(A) THP-1, SF1, SF2; (C) SM1, SM2, CESS]. After transfer to nitrocellulose, filters were probed for IL 8 mRNA as described in Sect. 2.2. (B) shows correctly sized I L 8 mRNA in stimulated THP-1 cells, and in freshly isolated SF cells. (D) shows IL8 mRNA in synovial membrane cells but not in the negative control (CESS cells).
Total cellular RNA was isolated from RA SF (SF1 = RA5, SF2 = RA6) and synovial membrane (SM1, SM2) to determine whether IL8 mRNA was being produced. The results are illustrated in Fig. 1. Following electrophoresis the gel was photographed under ultraviolet illumination [(A) THP-1, SF1, SF2; (C) SM1, SM2 CESS] to determine the amount of RNA present. The filters were then hybridized with the IL8 cDNA probe as described in Sect. 2.2.THF'-1 cells which had been activated with 10 pglml LPS and 50 ng/ml PMA for 3 h, served as a positive control, and CESS cells containing no IL 8 mRNA served as a negative control. It can be seen that correctly sized IL8 transcripts (1.8 kb) were present in cells obtained both from both SF (B) and synovial membrane (D).
2144
Eur. J. Immunol. 1990. 20: 2141-2144
E M. Brennan, C. 0. C. Zachariae, D. Chantry et al.
4 Discussion The accumulation of PMN and lymphocytes within the synovial joint of patients with RA has been well documented. The migration of these cells into the joint is believed to be due to the production of chemotactic factors; however, the identity of such factor(s) and their production within the joint has not previously been addressed.We have attempted to investigate this by examining the expression of IL8, a cytokine identified principally on the basis of its chemotactic activity for neutrophils. Our results demonstrate for the first time the presence of IL 8 bioactivity in the SFof patients with RA and the expression of IL 8 mRNA by freshly isolated synovial cells, suggesting a role for IL8 in the accumulation of leukocytes in this disease. Although Ph4N could be readily detected in most of the SF analyzed it is noteworthy that there was no direct correlation between the levels of I L 8 detected and the number of PMN in the SEThe reason for this is unclear; however, it is now known that there are a large number of low-molecular weight molecules which are structurally related to IL 8 and several have been shown to have chemotactic activity. For example the gro gene product, also known as MGSA [ l l ] has recently been shown to be produced by fibroblasts and endothelial cells following stimulation with IL 1 or TNF-a [22], and has neutrophil chemotactic activity equivalent to IL8 (C. C. Zachariae and K. Matshushima, unpublished observations). It is likely that mediators such as this may also be expressed as a part of the chronic inflammatory process in RA, and may also act as chemoattractants thus contributing to the chemotactic activity of the SF. IL 8 is produced by a range of cell types following stimulation with LPS and more importantly by I L 1 and TNF-a [8,16]. Both I L1 and TNF-a have been shown to be present in SF [2,3] and we have demonstrated that freshly isolated synovial cells produce these mediators [ 171. Synovial cells consist of a heterogeneous population [18], many of which have been shown to be capable of making IL8 following stimulation including T cells, monocytes and resident cells within the membrane such as endothelial cells and fibroblasts. I L 1 and TNF-a have been shown to be strong inducers of IL8 in vitro and other related factors.We have recently demonstrated that TNF-a is the major stimulus driving IL 1 production in RA [18] thus it is likely that TNF-a and/or I L 1 are involved in the production of IL 8 in the synovial joint. It has been recently shown using a receptor competition assay that many of the neutrophil chemotactic factors inducible by IL 1and TNF-a, including groMGSA and MIP I1 bind to the IL 8R (K. Matshushima, unpublished observations), with different specific activities. Thus the relative contribution of such factors chemotactic for neutrophils in the pathogenesis of RA is likely to be complex. Experiments are in progress to ascertain the importance of such factors in addition to IL 8 in this process
and the relative importance of I L 1 and TNF-a in their induction in v i v o . I Walker . for collecting the tissue We would like to thank Ms. . samples, and Ms. G . Harris and Ms. K . Hartley for processing the synovial tissue and synovial fluids.
Received February 5, 1990; in revised form May 21, 1990.
5 References 1 Dayer, J.-M. and Demczuk, S., Springer Semin. lmmunopathol. 1984. 7: 387. 2 Fontana, A., Hengartner, H., Weber, E., Fehr, K., Grob, l? J. and Cohen, G., Rheumatol. lnt. 1982. 2: 49. 3 DiGiovine, F. S., Nuki, G. and Duff, G. W., Ann. Rheum. Dis. 1988. 47: 768. 4 Granstein, R. D., Margolis, R., Mizell, S. B. and Sauder, D. N., J. Clin. Invest. 1986. 77: 1020. 5 Yoshimura, T., Matshushima, K., Oppenheim, J. J. and Leonard, E. J., J. Immunol. 1987. 139: 786. 6 Yoshimura,T., Matshushima, K.,Tanaka, S., Robinson, E . A., Appella, E., Oppenheim, J. J. and Leonard, E. J., Proc. Natl. Acad. Sci. USA 1987. 84: 9233. 7 Matshushima, K., Morishita, K., Yoshimura, T., Lavu, S., Kobayashi, W., Lew, E. , Appella, E., Kung, H., Leonard, E. and Oppenheim, J. J., J. Exp. Med. 1988. 167: 1883. 8 Gronhoj Larsen, C., Anderson, A . O., Oppenheim, J. 0. and Matshushima, K., Immunology 1989. 68: 31. 9 Luster, A. D. and Ravetch, J. V., J. Exp. Med. 1987. 166: 1084. 10 Wolpe, S. D. and Cerami, A., FASEB J. 1989. 4: 2565. 11 Richmond, A., Balentien, E., Thomas, H. G., Flaggs, G., Barton, D. E., Spiess, J., Bordoni, R., Francke, U. and Derynck, R., EMBO J. 1988. 7: 2025. 12 Schroder, J. M., Mrowietz,V., Morita, E. and Christophers, E., J. Immunol. 1988. 139: 3474. 13 Peveri, l?, Walta, A., Dewald, B. and Baggiolini, M., J. Exp. Med. 1988. 167: 1547. 14 Van Damme, J. ,Van Beeumen, J., Opdenakker, G. and Billiau, A., J. Exp. Med. 1988. 167: 1364. 15 Gronhoj Larsen, C., Anderson, A. O., Appella, E., Oppenheim, J. 0. and Matshushima, K., Science 1989. 243: 1464. 16 Strieter, R. M., Kunkel, S. L., Showell, H. J., Remick, D. G., Phan, S. H.,Ward, P A . and Marks, R. M., Science 1989.143: 1467. 17 Buchan, G., Barrett, K. ,Turner, M., Chantry, D., Maini, R.N. and Feldmann, M., Clin. Exp. lmmunol. 1988. 73: 449. 18 Brennan, F. M., Chantry, D., Jackson, A., Maini, R.N. and Feldmann, M., Lancet 1989. ii: 244. 19 Silman. A. J.. Br. J. Rheumatol. 1988. 27: 341. 20 Feinberg, A. and Vogelstein, B., Anal. Biochem. 1983.132: 6. 21 Furuta, R.,Yamagishi, J., Kotani, H., Sakamoto, F., Fukui,T., Matsui,Y., Sohmuraa,Y ,Yamada, M.,Yoshimura,T., Larsen, C. G., Oppenheim, J. J. and Matshushima, K., J. Biochem. 1989. 106: 436. 22 Wen, D., Rowland, A. and Derynck, R., EMBO .I. 1989. 8: 1761.
k
