Introduction
Interleukin-1 (IL-1) and tumour necrosis factor (TNF), mainly produced by macrophages, are among the principal cytokines occurring in acute and chronic inflammatory diseases. They act on a large number of target cells by inducing, for instance, the expression of cytokines, chemokines, cytokine receptors, prostanoids, adhesion molecules and matrix metalloproteinases (MMPs). Under the control of T cells and macrophages, synoviocytes contribute to the destructive inflammatory process of the articular structures in chronic arthritic diseases [
1,
2,
3,
4]. This type of cellular interaction is not only mediated by soluble factors but also by direct contact between stimulated T cells, inducing macrophages to produce large amounts of TNF and IL-1 [
5,
6]. These cytokines induce the production of MMP-1, MMP-3 and inflammatory molecules such as prostaglandin E
2 (PGE
2), leading to the degradation of the extracellular matrix and the resorption of the mineral phase of the bone [
1,
2,
3,
4]. In addition, IL-1 and TNF induce the production of IL-6 and chemokines (e.g. IL-8, macrophage chemoattractant protein-1 [MCP-1]); IL-6 is mainly involved in activating B cells and MCP-1 is involved in recruiting inflammatory cells [
7,
8].
Three lines of results point to TNF and IL-1 as principal players in rheumatoid arthritis and probably also in other chronic inflammatory diseases. Studies have shown that patients suffering from joint inflammation have higher serum and/or synovial levels of TNF and IL-1 [
3,
4,
7,
9]. Another group has reported that transgenic mice secreting large amounts of TNF develop arthritis and joint destruction [
10]. Furthermore, natural inhibitors of TNF and IL-1 that are soluble receptors or antagonists (i.e. TNFsR, IL-1sRII and IL-1Ra) [
11,
12,
13] and blocking antibodies to TNF and IL-1 are able to limit joint inflammation, pain and bone erosion in patients with rheumatoid arthritis [
3,
14,
15,
16]. Despite impressive clinical results with anti-TNF therapy, however, not all patients respond to such treatment and, in many, joint destruction continues to progress [
17,
18]. This could imply that other cytokines, in particular members of the TNF family, may play an additional role.
TNF-like weak inducer of apoptosis (TWEAK) is another member of the TNF superfamily which currently includes more than 15 proteins involved in inflammation, cell differentiation, cell death and immune organ formation [
19]. It is a secreted protein with cytotoxic and proinflammatory activities [
20,
21,
22,
23,
24,
25]. TWEAK was also shown to act as a growth factor by inducing the proliferation of endothelial cells [
26]. As is the case for the other members of the TNF family [
27,
28,
29,
30,
31], the active form of TWEAK is a trimer [
20]. In this study, recombinant, soluble TWEAK and anti-TWEAK mAbs were tested for their ability to induce or block proinflammatory responses in human dermal fibroblasts and synoviocytes.
Materials and methods
Dermal fibroblasts and synoviocytes
To isolate human dermal fibroblasts and synoviocytes, respectively, foreskin of normal donors and surgical specimens from patients with rheumatoid arthritis and advanced osteoarthritis were subjected to protease treatment as already described [
1,
2,
6]. Synoviocytes were used from the third to the ninth passage. The cells were cultured in Dulbecco's modified Eagle's medium (Life Technologies, Paisley, UK) supplemented with 10% FCS (Seromed/Biochrom KG, Berlin, Germany), 50 μg/ml streptomycin, 50 IU/ml penicillin and 2 mM L-glutamine, at 37°C in 5% CO
2. The cells were seeded at 2 × 10
4 cells/well into 96- well flat-bottom plates (Costar, Acton, MA, USA) and maintained for 48 hours before testing. Fresh medium (200 μl) either alone or together with cytokines and/or antibodies was then added for 72 hours, in the presence of 2 μg/ml of polymyxin B (Sigma, St Louis, MO, USA). The conditioned media were collected and stored at -20°C until further analysis.
For blockade experiments, antibodies were preincubated in Eppendorf tubes at a final concentration of 10-7 M with cytokines or medium alone for 30 min at 37°C, and 200 μl of the mixtures were then added to the cells.
For mRNA experiments, 2 × 106 cells were plated in culture plates (Falcon, Becton Dickinson, Heidelberg, Germany) in the absence or presence of cytokines and/or antibodies in a total volume of 5 ml of medium containing 2 μg/ml of polymyxin B for the indicated time.
Recombinant proteins and monoclonal antibodies
TNF and IL-1β were from Biogen (Cambridge, MA, USA). Recombinant soluble human TWEAK contains an NH2-terminal myc tag followed by amino acid residues Ala106–His249 of hTWEAK (GenBank accession # AF030099). hTWEAK was expressed by the yeast Pichia pastoris (Invitrogen, Groningen, The Netherlands) and purified by conventional methods. Monoclonal antibody to human TNF (D2E7, LU 200134) was from Knoll (BASF Pharma, Ludwigshafen, Germany). Hamster monoclonal antibodies to hTWEAK were prepared using conventional methods by injecting Armenian hamsters intraperitoneally with recombinant soluble hTWEAK in complete Freund adjuvant, followed by intraperitoneal booster in incomplete Freund adjuvant. Initial screening of mAb activity was carried out by ELISA: purified hTWEAK was coated onto 96-well plates and various hamster mAbs were tested for their ability to bind to this immobilized protein. The capture of hamster mAb was visualized using donkey anti-hamster IgG (Jackson ImmunoResearch, West Grove, PA, USA) labelled with horseradish peroxidase. Monoclonal antibodies were further tested to determine whether they were able to block hTWEAK-flag binding to HT29 cells.
Total RNA was isolated from cultured cells using TRIzol reagent (Life Technologies, Paisley, UK) according to supplier's instructions. RNA was electrophoresed in 1% formaldehyde agarose gel, transferred to Hybond N nylon membrane (Amersham, UK) and hybridized to
32P-labeled MMP-1 (Dr Pierce, St Louis, MO, USA), cyclooxygenase 2 (COX-2) (a gift from Dr PE Poubelle, Ste-Foy, Canada) or glyceraldehyde-3 phosphate dehydrogenase (GAPDH) cDNA probes as already described by Rezzonico
et al. [
32]. RNAse protection assay was performed with the human chemokine multi-probe template set hcK5 from Pharmingen (San Diego, CA, USA) as described by the manufacturer.
Determination of IL-8, IL-6, MMP-1, RANTES, IP-10 and PGE2
Human IL-8 and RANTES were measured by ELISA using the quantikine immunoassay (R&D, Minneapolis, MN, USA). Human IL-6 ELISA was performed with innotest hIL-6 from Innogenetics (Zwijhaarde, Belgium). PGE
2 was quantitized using the products and protocol from Neupert
et al. [
33]. The determination of interstitial collagenase (MMP-1) was carried out as described by Lacraz and colleagues [
34]. Human IP-10 was measured using Hbt ELISA (HyCult Biotechnology, Uden, The Netherlands).
Discussion
Cultures of human dermal fibroblasts and synoviocytes have proved relevant to
in vitro models for studying surrogate markers of the mechanisms of chronic inflammatory diseases resulting in inflammation and tissue destruction of the joint structures. Synoviocytes are exquisitely sensitive to mediators of inflammation, such as TNF and IL-1β, and react strongly to direct contact with inflammatory T cells [
1,
2,
3,
4,
5,
6,
35]. They are able to secrete numerous proinflammatory molecules, such as PGE
2, chemokines (i.e. IL-8 and MCP-1) and proteolytic enzymes involved in extracellular matrix destruction such as MMP-1. The home-ostasis of inflammatory cytokines and metalloproteinases, as well as their respective inhibitors (i.e. IL-1Ra, TNFsR, tissue inhibitor of metalloproteinases-1) controls the process of tissue destruction [
3,
14,
15]. In addition, the process of repair of collagen and proteoglycan is also inhibited by IL-1 and TNF [
32,
35].
We recently reported that a member of the TNF family, TWEAK, has inflammatory and cytotoxic activities like TNF and lymphotoxin (LT)α [
20]. Marsters
et al. [
21] have reported that the receptor for TWEAK is the death-domain-containing receptor Apo3/DR-3/TRAMP, whereas Schneider
et al. [
22] demonstrated that TWEAK-induced cell death occurs via TNF and TNF receptor I on the Kym-1 cell line. Nakayama
et al. [
24] and Kaptein
et al. [
36] showed that TWEAK binding and TWEAK-induced cell death occurs in a number of tumour cells deficient in Apo3/DR-3/TRAMP. There must exist a TWEAK receptor that does not contain a death domain because TWEAK also promotes the growth of endothelial cells [
26]. In addition, in astrocytes, which proved resistant to TWEAK-induced cell death, the binding of TWEAK results in the increased secretion of IL-8, IL-6 and intercellular adhesion molecule [
23].
The present study demonstrates another noncytotoxic effect of TWEAK in that hTWEAK stimulated dermal fibroblasts and synoviocytes to produce significant amounts of RANTES, IP-10, IL-8 and MMP-1 as well as some PGE2 and IL-6. Although hTWEAK has proinflammatory properties, these are much less pronounced than those of IL-1β and TNF (with the exception for RANTES and IP-10), even though it potentiates the proinflammatory activities of TNF or IL-1β. In contrast with TNF, induction by TWEAK of most of the inflammatory mediators requires a longer time of incubation. Consequently, it cannot be ruled out that the prolonged effect of TWEAK in stimulating these mediators may be due to the rapid induction of one or more intermediate mediators, which are unidentified to date.
The production of PGE2, MMP-1, IL-8, IL-6, RANTES and IP-10 induced by hTWEAK (but not the production induced by TNF and IL-1β) was completely blocked by the mAb to hTWEAK, BCB10, confirming the specificity of hTWEAK. Moreover, blocking anti-TNF antibodies were inactive in the secretion of IL-8, MMP-1, PGE2, RANTES and IP-10 induced by TWEAK, ruling out any involvement of TNF in TWEAK activity. In contrast, BEB3, another mAb to hTWEAK, increased the activity of hTWEAK significantly, possibly by promoting the trimerization and/or stability of TWEAK. Such a mechanism has been observed with a number of ligands of the TNF family. On synoviocytes, but not on dermal fibroblasts, BEB3 alone stimulated the secretion of IL-8, MMP-1 and PGE2, whereas the blocking mAb to TWEAK, BCB10, reduced the basal level of IL-8 and MMP-1, suggesting that synoviocytes secrete endogenous TWEAK. The latency of this activity – observed on synoviocytes whether stimulated with BEB3 alone or not – could indicate that a soluble TWEAK inhibitor may also be secreted by these cells. In dermal fibroblasts, BEB3-crosslinked hTWEAK was shown to increase the mRNA level of three chemokines of the CXC or CC families, IL-8, MCP-1 and RANTES. In synoviocytes, BEB3-crosslinked hTWEAK increased the mRNA level of IL-8, MCP-1, RANTES, IP-10 and MIP-1α. The delayed synthesis of these molecules, contrasting with TNF-induced stimulation, suggests that TWEAK could be relevant to the persistence of the inflammatory response by recruiting neutrophils and macrophages to the site of the inflammation.