Arthritis & Rheumatism Official Journal of the American College of Rheumatology

CURRENT COMMENT

CYTOKINES AND CYTOKINE INHIBITORS OR ANTAGONISTS IN RHEUMATOID ARTHRITIS WILLIAM P. AREND and JEAN-MICHEL DAYER relative concentrations of various cytokines and inhibitors in the pericellular environment in any diseased tissue. One objective of this review is to summarize the current understanding of the role of mediator molecules in RA. This article will focus primarily on interlcukin-1 (IL-1) and 1L-l inhibitors, but other molecules, including tumor necrosis factor a (TNFa), a T N F a inhibitor, transforming growth factor p (TGFP), 1L-6, y-interferon (7-IFN), granulocytemacrophage colony-stimulating factor (GM-CSF), platelet-derived growth factor (PDGF), and fibroblast growth factor (FGF) will also be discussed. Although other cytokines may be relevant, they will not be discussed in this brief review.

Cytokines have been implicated as important mediators of inflammation and joint destruction in rheumatoid arthritis (RA) (1). The results of many in vitro and in vivo studies suggest that cytokines may mediate cell-cell interactions that result in the release of tissue-damaging enzymes. In fact, many of the disease-modifying drugs used to treat RA may alter the production o r action of these cytokines. Research in many laboratories is directed at developing new therapeutic agents that would more effectively block the synthesis, release, or effects of particular cytokines. Although cytokines have important functions in normal cellular processes, their purported roles in pathophysiology must result from excessive production or inadequate inhibition. The results of recent studies indicate that the biologic effects of a single cytokine may vary considerably with the target cell or tissue. Furthermore, some cytokines may antagonize or oppose the effects of other cytokines. Lastly, natural inhibitor molecules appear to exist that may selectively block the effects of a particular cytokine. The net biologic effects would then depend upon the --

Pathophysiology of rheumatoid synovitis A hypothetical model of cytokine involvement in rheumatoid synovitis is depicted in Figure 1. In this model, antigen-presenting cells bearing the class I1 major histocompatibility complex molecules (la or HLA-DR) stimulate T lymphocytes to release y I F N and other cytokines. The possible role of antigen in this process remains unclear and will not be discussed in this review. Unknown factors from T lymphocytes and other cells are thought to stimulate macrophages to synthesize and secrete IL-I, TNFa, and IL-6. Gamma-interferon alone does not induce IL-1 and T N F a production, but its primary role may be to augment the effects of other agents. However, the precise mechanisms whereby macrophages are stimu-

~

From the Division of Kheumatology, Department of Medicine, University of Colorado Health Sciences Center. Denver. and the Division of Immunology and Allergy, HBpital Cantonal Univcrsitaire, Geneva, Switzerland. William P. Arend, MD: Professor and Head, Rheumatology Division, University of Colorado School of Medicine; Jean-Michel Dayer, MD: Associate Professor, HBpital Cantonal Universitaire. Address reprint rcqucsts to William P. Arend, MD, Division of Rheumatology, Box B 115, Univcrsity of Colorado Health Scicnccs Center, 4200 East Ninth Avcnuc, Denver, CO 80262. Submitted for publication June 19, 1989: accepted in revised form October 26. 19x9. Arthritis and Rheumatism, Vol. 33, No. 3 (March 1990)

305

AREND AND DAYER

306

------ .

and antagonist molecules in these cell-cell interactions.

IL-1

Effects of IL-1 on synovial fibroblasts and chondrocytes I

GM-CSF

Fibroblasts

-

Articular Cartilage

Chondrocytes

Production of collagenase and other neutral proteases

Figure 1. Stimulatory or agonist effects of cytokines on cell-cell interactions in rheumatoid synovitis. T cells, macrophages, and fibroblasts are all present in the rheumatoid pannus. Enzymes, such as collagenase and other neutral protcases, are secreted by synovial fibroblasts and chondrocytes. and result in destruction of cartilage, bone, and penarticular structures. IL-I = interlcukin-1; y1FN = yinterferon; GM-CSF = granulocyte-macrophage colonystimulating factor; TNFa = tumor necrosis factor a.

lated remain unclear, since there is little evidence to suggest the presence of lymphokines in the rheumatoid synovium (see below). Alternative hypotheses would be that synovial macrophages exist in a state of preactivation, possibly secrete cytokines without continuous induction, or are capable of paracrine and autocrine stimulation (2). For example, the results of recent studies suggest that TNFa, possibly secreted by macrophages, may be responsible for induction of IL-1 production in cultured rheumatoid synovial cells (3). Both IL-1 and TNFa may induce the production of collagenase and other neutral proteases in synovial fibroblasts and in chondrocytes located in the adjacent articular cartilage. These enzymes degrade proteoglycans and collagen, resulting in cartilage destruction. GM-CSF is hypothesized to amplify this process of cell-cell interaction by enhancing HLA-DR expression on macrophages and other antigenpresenting cells (4). The following discussion will review the data supporting this model of tissue destruction in RA, and summarize recent data suggesting that the explanation may be more complex. A repeating theme that will be developed is the probable importance of both agonist

IL-I is a 17-kd cytokine that is primarily a product of monocytes and macrophages (5). Other cells, including endothelial cells, keratinocytes, mesangial cells, astrocytes, B lymphocytes, and activated T lymphocytes, may also produce IL-1. Two forms of IL-1 (a and p) have been described, and they share -26% amino acid sequence homology. IL-1p is the predominant form synthesized by human monocytes and represents -90% of the IL-1 found in the supernatant of stimulated cells. 1 L - l a may function primarily as a membrane-bound molecule. Both IL-la and IL-lp bind to the same receptor on target cells and exhibit the same spectrum of biologic activities. However, I L - l a and IL-1/3 may bind with different affinities to the same receptor, and the 1L-1 receptor may vary between different cell types. The systemic effects of IL-1 are exerted on the central nervous system, bone marrow, blood vessel walls, and other tissues, and include the induction of metabolic changes ( 5 ) (Table 1). However, the local effects of IL-I on both immune and inflammatory cells may be more important in RA (6). These responses include augmentation of T and B lymphocyte function, chemotaxis of neutrophils, lymphocytes, and monocytes, proliferation of fibroblasts, and production of prostaglandin E2 (PGE2) and collagenase by synovial fibroblasts (7) and chondrocytes (8). In addition, IL-I may further enhance chemotaxis indirectly, through the induction of other chemoattractant cytokines (see Table 1. Biologic effects of interleukin-1 that may occur in rheumatoid arthritis

Site

Effects

Systemic

Fever Decreased appetite Increased granulocyte-macrophage colonystimulating factor production Synthesis of acute-phase proteins

Local

Chemotaxis of polymorphonuclear cells, lymphocytes, and monocytes Adherence of white blood cells to endothelial cells Fibroblast proliferation Prostaglandin E,, collagenase, and neutral protease production by fibroblasts and chondrocytes lncreased production of collagen and an inhibitor of neutral proteases Stimulation of T and R lymphocytes

CYTOKINES AND RA

below). IL-I may be involved in early events in rheumatoid synovitis by assisting in the migration of other cells and stimulating a variety of responses in endothelial cells. IL-I induction of PGEz and collagenase production may be responsible for some of the bone resorption and cartilage destruction that is observed in later stages of thc disease. In contrast, IL-1 may also contribute to joint scarring and fibrosis by stimulating fibroblast proliferation either directly (9) or indirectly, through the induction of PDGF (10). IL-1 also induces fibroblast production of fibronectin ( 1 I), type I collagen (9,111, proteoglycans (12), and an inhibitor of collagenase and other neutral proteases (9). In addition, IL-I alters collagen production by chondrocytes, reduces production of type I1 collagen, and enhances synthesis of type I and type I11 collagen (13). Since type I1 collagen is the predominant form in articular cartilage, these effects of IL-I may result in a further weakening of the rheumatoid joint. Furthermore, collagen types 11, 111, and IX may stimulate IL-1 production (l4,15). Lastly, IL-1 has been shown to have opposing effects on proteoglycans in cartilage, leading to both the inhibition of synthesis and the enhancement of degradation (12,16). These effects of IL-1 on synovial fibroblasts and chondrocytes represent an example of agonist and antagonist influences occurring simultaneously. Collagenase and its inhibitor may be produced at the same time by the same cells; the relative amounts of induction of each may vary with the cell stimulus. In addition, both collagcnase and collagen may be synthesized simultaneously, although the collagen type that is produced may lead to fibrosis, and not to adequate tissue repair. Thus, the degree and distribution of destruction in the rheumatoid joint may depend, in part, upon the relative production of collagenase, enzyme inhibitors, and collagen by chondrocytes and synovial fibroblasts. However, it has not been determined whether all of these in vitro effects of IL-1 on cell function also occur in vivo. In the following sections, studies on the role of IL-1 in animal models of experimental synovitis, the role of IL-I in RA, and the possible role of recently described IL-I inhibitors will be reviewed.

IL-1 and IL-1 inhibitors in experimental synovitis in vivo Observations in experimental animal models of arthritis support the hypothesis that IL-1 may be an

307

important mediator of tissue destruction. Intraarticular injection of IL-1 in rabbits induces a transient infiltration of neutrophils into the joint space, followed by mononuclear cell infiltration, although few cells are observed in the synovium (17). Loss of proteoglycans from the midzone of the articular cartilage also follows IL-1 injection into the joint, a response that appears to be separate from the cellular infiltration. Repeated injection of IL-I into the ankles of normal rats results in a chronic synovitis that is characterized by the presence of mononuclear cells and fibrosis, but is without cartilage and bone destruction (18). However, in joints that were previously injected with a streptococcal cell wall peptidoglycan-polysaccharide complex, subsequent injection of IL-1 markedly accentuates the inflammatory response, with production of pannus and cartilage destruction (18). In addition, it was shown that collagen-induced inflammatory arthritis in mice is accelerated by subcutaneous infusion of 1L-1 (19). These studies in animal models of arthritis all suggest that IL-1 is capable of mediating tissue destruction. It has also been observed that injection of IL-1 into the knee joints of rats with antigen-induced arthritis leads to a reduction in inflammation and joint destruction (20). This ameliorative effect of IL- 1 was observed with either 1L-1 pretreatment or administration of IL-I after the induction of arthritis. Further studies are necessary to establish whether this phenomenon might be due, in part, to desensitization to IL-I or possibly to the induction of IL-1 inhibitors (21) (see below). A study on antigen-induced arthritis in rabbits has suggested the presence of substances that may inhibit the effects of IL- 1. It was observed that biologically active IL-1 could be detected in synovial fluids only from arthritic joints at an early disease stage (22). However, IL-1 inhibitor activity was present in control synovial fluids and in increased amounts in arthritic synovial fluids. Markedly increased IL- I production by synovial explants was observed after the induction of arthritis, when no IL-1 activity could be detected in the synovial fluids (22). This finding implies that IL-I activity levels in joint fluids may not be an accurate reflection of what is occurring in the synovial tissues, primarily because of the presence of IL-I inhibitors in these fluids. Thus, in animal models of synovitis, IL-1 may accelerate tissue destruction, induce primarily reparative responses, or be accompanied by the production

AREND AND DAYER

Table 2. Evidence suggesting the role of interleukin-l (IL-I) in rheumatoid arthritis

Location

Features ~~

Synovialkid

~

-Variable levels of IL-I bioactivG Elevated levels of IL-la and II,-lp protcins Cells may not spontaneously produce IL-1

Synovial tissue

Spontaneous in vitro production of 1L-la and IL-lP High levels of I L - l a and IL-lp messenger KNA

Blood

Presence of IL-1p in levels that correlate with disease activity

of IL-1 inhibitors. These findings reemphasize the dual nature of the biologic effects of IL-1.

IL-1 in rheumatoid arthritis Much evidence exists to suggest that IL-l may play a role in rheumatoid arthritis (Table 2). Studies that have used bioassays in attempts to measure 1L-1 activity in human synovial fluids have yielded variable results, probably because of the ubiquitous presence of IL-1 and IL-2 inhibitors in these fluids (for review, see ref. 23). Enzyme-linked immunosorbent assays or radioimmunoassays for IL- 1a and IL- Ip proteins have demonstrated the presence of increased levels of thesc cytokines in the synovial fluids of both patients with active RA and patients without RA (23,24). However, these assays may be influenced by the viscosity of synovial fluids and by the presence of rheumatoid factors. Studies on IL-I production by RA synovial fluid or tissue macrophages have also yielded conflicting results. Although synovial fluid cells may not produce high levels of 1L-1 (25), rheumatoid synovial tissue obtained at arthroscopy synthesizes large amounts of both IL-la and IL-lp in vitro (26). This increased 1L-1 production was not observed in nonrheumatic synovia1 tissue and was correlated with arthroscopic and radiologic evidence of destructive arthritis in the RA patients. Thus, the presence of IL-1 in the synovial fluids of both RA patients and non-RA patients may be derived primarily from cells in the inflamed synovium. This conclusion is supported by the detection of high levels of IL-la and IL-lp messenger RNA (mRNA) in the rheumatoid synovial membrane (27,28). The results of additional studies offer further evidence that IL-1 may be an important mediator of tissue destruction in RA. Levels of circulating IL-lp in RA patients have been correlated with clinical disease

activity (29). In addition, circulating monocytes from RA patients with recently active disease spontaneously secrete 1L-l in vitro (30), a finding that probably reflects monocytopoiesis as immature cells secrete more IL-1. Lastly, although suggestive evidence exists, further studies are necessary to establish whether disease-modifying drugs used in the treatment of RA may inhibit the production or biologic effects of IL-1.

IL-1 inhibitors The results of recent studies have indicated the existence of natural inhibitors of IL-I that could potentially modulate the local effects of IL-1 in the rheumatoid joint. A variety of IL-1 inhibitors have been identified in cell supernatants and in human urine (for review, see refs. 31-33). However, the term “IL-1 inhibitor” is a general one, and these substances could potentially be acting at many different lcvels in a specific or nonspecific manner. These levels of activity include reducing IL-1 synthesis, binding to IL-1 and inhibition of its action, degradation of 1L-I, blocking IL-1 receptors, and interfering with IL-I effects at a postreceptor level. An additional level of regulation of cytokine effects could be the modulation of receptor expression. However, many of the activities called “IL-I inhibitors” have not been purified, and their mechanisms of action have yet to be determined. The subsequent paragraphs will summarize recent studies on a specific IL-I inhibitor that binds to the IL-I receptor on immune and inflammatory cells, and prevents 1L-1 from interacting with target cells.

Table 3. Characteristics of the 22-kd interleukin-l ( I L - I ) inhibitor

Sources

Human monocytes cultured on adherent immune complexes or IgG Urine of patients with fever or myelonionocytic leukemia

Properties

MW 22 kd, pl 4.54.7 Is not transforming growth factor p or another known cytokine No immunologic cross-reactivity with IL-1a or

IL-IP Mechanism

Binds to the 1L-l rcceptor on target cells

Effects

Prevents IL-l augmentation of thymocyte proliferation Prevents IL-l induction of prostaglandin Ez and collagenase production by synovial fibroblasts and chondrocytes Does not block tumor necrosis factor a stimulation of target cells

CYTOKINES AND RA

Table 4.

309

Other cytokines and inhibitors that may be involved in rheumatoid synovitis Factor

Properties

Tumor necrosis factor a ('I'NFa)

Produced primarily by macrophages and possesses biologic properties similar to interleukin- I (IL- I )

T N F a inhibitor

MW 31-33 kd. pl 5.5-6.1 Binds to soluble 1"Fa and prevents target cell stimulation

Interleu kin-6

Induced by 1L-1 and TNFa; stimulates hepatic synthesis of acute-phase proteins and Ig production by B lymphocytes

Granulocyte-macrophage colony-stimulating factor

Enhances HLA-DR expression and IL-1 production by monocytes

Platelet-derived growth factor

Stimulates fibroblast proliferation and anchorage-independent growth

Transforming growth factor /3

Stimulates matrix production, inhibits matrix degradation, and may induce immunosuppression

It has been observed that human monocytes cultured on a substrate of adherent immune complexes (34) or adherent IgG (32) secrete a 22-kd protein that inhibits the effects of IL-I on immune and inflammatory cells (Table 3). This protein binds specifically to the IL-1 receptor and blocks IL-1 induction of PGE, and collagenase production in human synovial fibroblasts and rabbit chondrocytes. This 1L-l inhibitor does not block IL-2 stimulation of IL-2-dependent cell lines, is not TGFP, and exhibits no immunologic cross-reactivity with IL-la or IL-1P (32). An IL-1 inhibitor possessing identical characteristics has been purified from the urine of both febrile patients and patients with myelomonocytic leukemia (35-37) (Table 3). This protein does not block T N F a stimulation (36), and it inhibits both soluble and membrane-bound IL- I (31). It has been observed that as monocytes mature into macrophages during in vitro culture, IL-1 production progressively decreases, and production of the 22-kd IL-1 inhibitor is enhanced (38). This observation suggests that tissue macrophages may be an important source of the IL-1 inhibitor. A 25-kd IL-1 inhibitor has been found in rheumatoid synovial fluids, although part of this activity may be due to T G F p (39). Cells in the synovial fluid have been shown to produce an IL-1 inhibitor in vitro (40,41). Furthermore, IL-1 inhibitory activity has been found in the serum and urine of patients with juvenile rheumatoid arthritis (JRA), and maximum levels of this activity have been correlated with the presence of

fever (42). However, the biochemical characteristics of the IL-l inhibitory activities present in synovial fluids, in supernatants of synovial cells, and in the serum or urine of JRA patients, and the correlation of these activities with the 22-kd 1L-1 inhibitor have not yet been determined. The relevancc of this 22-kd IL-I inhibitor produced by human monocytes and macrophages in RA remains to be established. The relative amounts of IL-1 or of the 22-kd IL-I inhibitor produced by human monocytes varies with the stimulus, although it is not known whether the same cell synthesizes both proteins (Arend WP: unpublished observations). However, this finding is another example of the simultaneous presence of agonist and antagonist effects. The production of net IL-I inhibitory activity by rheumatoid synovial macrophages may represent a natural control mechanism. It is possible that continued active synovitis is due to inadequate synthesis of antagonist proteins, and not to excess production of agonists.

TNFa and TNFa inhibitors in rheumatoid arthritis Other molecules that are synthesized by monocytes also may be involved in rheumatoid synovitis (Table 4). T N F a is a 17-kd cytokine that is produced primarily by monocytes and macrophages (43). IL-I and TNFa are usually synthesized and secreted simul-

3 10 taneously , although production of these 2 proteins appears to be independently regulated and controlled by different mechanisms. As monocytes mature into macrophages, the ability to produce IL-I appears to decrease dramatically, but TNFa production may be relatively unaffected. TNFa binds to a separate receptor on target cells, but shares many biologic activities with IL-1. This cytokine was originally described as an agent that was cytotoxic to tumor cells in vitro. The main pathologic role of TNFa appears to be that of a mediator of septic shock (43). However, TNFa has been shown to stimulate PGE, and collagenase production in vitro by synovial fibroblasts and chondrocytes, as well as bone resorption and fibroblast proliferation (44). In addition, both IL-1 and TNFa act in an autocrine manner on macrophages to stimulate their own transcription, as well as the production of each other. TNFa may be involved in pathophysiologic events in experimental synovitis, but its role appears to be secondary to that of IL-I . Intraarticular injection of TNFa into rabbit knees induces a lower inflammatory response than does IL-Ij3. Injection of TNFa alone also does not lead to a loss of cartilage proteoglycans, as is observed with IL-I. However, injection of both TNFa and IL-I produces a greater inflammatory response than is observed with either cytokine alone (45). Thus, TNFa may augment the tissuedamaging potential of IL- 1 in experimental synovitis. TNFa is also present in rheumatoid synovial fluids and appears to be synthesized by synovial tissue (46). It has been observed that TNFa protein can be detected in more than 50% of rheumatoid synovial fluids, particularly in those from patients with seropositive or active disease (47,48). Furthermore, immunoperoxidase staining has revealed detectable TNFa proteins in 6 of 10 rheumatoid synovial tissue specimens (49). The localization of TNFa is primarily in synovial lining cells and in interstitial macrophages (46,49). Similar to IL-1, the local effects of TNFa in the rheumatoid joint may be counteracted by regulatory or inhibitory proteins. A specific inhibitor of TNFa has been found in the urine of febrile patients (50) and in the supernatants of cells cultured from rheumatoid synovial fluids (41). This TNFa inhibitor is a 31-33-kd protein, with a PI range of 5.5-6.1, and thus, differs from the 1L-1 inhibitor described above. The TNFa inhibitor has been purified to homogeneity (5 I), and amino acid sequence analysis of its NH, terminal region has failed to reveal homology with any known

AREND AND DAYER

Macrophage

Increared prokction of colaOen and fibronectin. Decreaaed recretbn of coLgenare and neutral proteaaee. Enhanced ryntheair of Inhibitors of neutral protearer.

Figure 2. Inhibitory or antagonist factors that may be involved in cell-cell interactions in rheumatoid synovitis. The effects of interleukin-1 (IL-1) and tumor necrosis factor a (TNFa) on synovial fibroblasts and chondrocytes may be antagonized by specific inhibitors. The IL-1inhibitor blocks receptor binding of IL-I, while the T N F a inhibitor interacts with T N F a in solution and prevents receptor binding. Furthermore, the direct effects of transforming growth factor P (TGFP) on these cells may contribute to fibrosis and tissue repair.

protein (52,53). Analogous to the IL-I inhibitor, the TNFa inhibitor shows specificity for TNFa; it has no effect on IL-la or IL-Ij3 stimulation of cells, and shows only a slight inhibitory effect on TNFj3. This inhibitor prevents TNFa-induced PGE, production and expression of cell-associated IL-I in human dermal fibroblasts. Unlike the IL-I inhibitor, the TNFa inhibitor does not bind to the TNFa receptor, but rather to the TNFa molecule itself (51-53). The results of recent studies indicate that the TNFa inhibitor may represent a soluble version of a cell-surface receptor (54). The role of this TNFa inhibitor in ameliorating rheumatoid synovitis, like that of the IL-I inhibitor, remains to be determined (Figure 2).

The role of IL-6 in rheumatoid arthritis IL-6 is a 26-kd cytokine that is produced by monocytes, T lymphocytes, and fibroblasts (55). Both IL-I and TNFa induce the synthesis and secretion of IL-6. Furthermore, the biologic activities of IL-6 are similar to those of IL-I and TNFa. IL-6 differs from IL-I and TNFa in its failure to stimulate PGE, and collagenase production in chondrocytes and synovial fibroblasts (for discussion, see ref. 56). In fact, IL-6 appears to block IL- 1-induced PGE, production by these cells in vitro (Dayer J-M: unpublished observa-

CYTOKINES AND RA

tions). However, IL-6 is a more potent inducer of hepatic synthesis of acute-phase proteins and of Ig production by B lymphocytes than is either IL-l or TNFa. High levels of IL-6 are present in inflammatory synovial fluids, particularly from RA patients (56-59). These IL-6 levels exceed those of 1L-1 or TNFa in the same fluids, and are correlated with systemic parameters of disease activity, such as the erythrocyte sedimentation rate and the presence of acute-phase proteins. Furthermore, cells cultured from synovial tissue have been shown to produce 1L-6 in vitro (56,60). Immunohistologic staining of rheumatoid synovium reveals that the 1L-6 protein is present primarily in fibroblasts. Thus, the role of IL-6 in the rheumatoid synovium may be as an amplifier of certain IL-1 and TNFa effects, but it may reduce other biologic sequelae. However, IL-6 may also be involved in induction of synthesis of acute-phase proteins and in local stimulation of rheumatoid factor production.

y-IFN and GM-CSF Both y I F N and GM-CSF can stimulate HLADR and HLA-DQ expression by synovial tissue macrophages and fibroblasts. This effect could amplify the inflammatory cycle by enhancing lymphocyte stimulation and induction of a local immune response (61). However, little y I F N is found in rheumatoid synovial fluid or is produced by synovial tissue explants (62). Further evidence suggests a paucity of other T cell lymphokines in the rheumatoid joint, with an absence of IL-2 and IL-3 mRNA (63). However, the production of colony-stimulating factor and mast cell growth factor activities by rheumatoid synovial tissue in vitro emphasizes the important role of additional mediators from inflammatory cells (63). Colony-stimulating factors are a heterogeneous group of cytokines that act on bone marrow progenitor cells to induce growth of hematopoietic colonies. However, colony-stimulating factors can also affect more mature cells. GM-CSF is a predominant growth factor that may be synthesized by macrophages, fibroblasts, endothelial cells, and activated lymphocytes. GM-CSF production may be stimulated by IL-I, TNF. and IL-6. Furthermore, GM-CSF stimulates 1L-1 production, activates neutrophils, and induces HLA-DR expression on monocytes (for review, see ref. 64). The results of recent studies indicate the presence of significant levels of GM-CSF in rheumatoid synovial fluids (64,65). In addition, rheumatoid syn-

31 1

ovial explants have been shown to produce GM-CSF in vitro (64). An interesting observation in these studies is that GM-CSF was often undetectable in synovial fluids, unless the samples were diluted. This finding suggests the possible presence of GM-CSF inhibitors (64). Further studies are necessary to establish precisely which cells in the rheumatoid synovium produce GM-CSF, and the importance of this cytokine in the local inflammatory cycle.

PDGF and FGF Growth factors with characteristics of PDGF and FGF are present in synovial fluids from noninflammatory forms of arthritis (66), and PDGF has been reported in rheumatoid synovial fluids (67). In addition, explants of rheumatoid synovial tissue have been shown to produce PDGF in vitro (67). Both PDGF and FGF may be produced by macrophages, and they can stimulate DNA synthesis and proliferation of human synovial fibroblasts, without inducing PGE, production (68). Also, PDGF enhances anchorage-independent growth of rheumatoid synovial cells in soft agarose cultures (69). PDGF and IL-1 exhibit complex interactions on cultured rheumatoid synovial cells. IL-I inhibits PDGF-induced synoviocyte proliferation through the stimulation of PGEz production (70). In contrast, PDGF enhances PGE, production but inhibits collagenase transcription, both after IL-I induction (70). These results suggest that growth factors, particularly PDGF, may be responsible for the aggressive growth and proliferation of fibroblasts in the human rheumatoid joint.

TGFP Othcr cytokines may function primarily to suppress acute inflammation and to accelerate scarring or fibrosis. TGFP is a 25-kd cytokine that is produced by a variety of cells including platelets, bone cells, macrophages, lymphocytes, and synovial fibroblasts (71). This cytokine has multiple biologic effects, but its primary functions are to stimulate matrix synthesis and inhibit matrix degradation (Figure 2). TGFP has been shown to increase transcription of collagen and fibronectin genes in multiple cells, decrease secretion of proteases, and enhance production of protease inhibitors (for review, see ref. 71). TGFp also is chemotactic for monocytes and induces IL-1 production in these cells. Lastly, TGFP reduces yIFNinduced HLA-DR expression in human cells (72), and

312

inhibits growth of rheumatoid synovial cells in soft agarose cultures (69). High levels of TGFP are found in synovial effusions from patients with RA or gout, while fluids from osteoarthritis patients demonstrate lower levels (73). TGFP is produced by rheumatoid synovial tissue in vitro, and TGFP mRNA is present in both synovial fibroblasts and macrophages in vivo (67,74). In addition, exogenous TGFP stimulates collagen transcription and inhibits collagenase mRNA levels in cultured synoviocytes (74). TGFp is secreted in a latent form and may be activated by proteases present in the rheumatoid synovium. In addition to enhancing the fibrotic phase of rheumatoid synovitis, TGFP is immunosuppressive and may be responsible for the deficient cell-mediated immune response demonstrated by rheumatoid synovial lymphocytes in vitro (75). Thus, TGFP possesses the potential to exert both agonist and antagonist effects in the rheumatoid synovium.

Summary This review has summarized some of the evidence suggesting that cytokines may play an important role in mediating pathophysiologic events in RA. However, these proteins are capable of mediating both stimulatory (agonist) and inhibitory (antagonist) effects in the rheumatoid synovium. GM-CSF, IL-1, TNFa, and PDGF are all produced in the rheumatoid synovium and may function to induce inflammation, enzyme release, fibroblast proliferation, and tissue destruction. Local release of IL-6 may alter the effects of IL-1 and TNFa, as well as induce Ig production and hepatic synthesis of acute-phase proteins. However, specific inhibitors of 1L-1 and TNFa exist, which, if also released into the synovium, may antagonize the proinflammatory effects of these cytokines. In addition, IL-1 may have antiinflammatory effects, such as the induction of the synthesis of collagen and enzyme inhibitors by chondrocytes and synovial fibroblasts. Stimulation of these latter cells by TGFP also may result in decreased matrix degradation and increased formation of scar tissue. The developing scenario is one of cell-cell interactions that are influenced in positive and negative manners by the local release of various mediators. A further understanding of cytokines and cytokine inhibitors in the rheumatoid synovium may lead to the development of more specific and effective therapeutic agents .

AREND AND DAYER

Addendum. The 22-kd IL-I inhibitor found in the supernatants of human monocytes cultured on adherent IgG has recently been purified and sequenced (Hannum CH et al: Interleukin-I receptor antagonist activity of a human interleukin-1 inhibitor. Nature [in press]). This molecule is a unique protein, not previously described, that possesses -26% amino acid sequence homology to IL-Ip. A complementary DNA for this IL-1 inhibitor has been expressed in Escherichia coli with successful production of biologically active recombinant protein (Eisenberg SP et al: Primary structure and functional expression from complementary DNA of a human interleukin-1 receptor antagonist. Nature [in press]).

REFERENCES 1. Dayer J-M, Demczuk S: Cytokines and other mediators

in rheumatoid arthritis. Springer Semin Immunopathol 7:387-413, 1984 2. Zvaifler NJ: New perspectives on the pathogenesis of rheumatoid arthritis. Am J Med 85A: 12-17, 1988 3. Brennan FM, Chantry D, Jackson A, Maini R, Feldmann M: Inhibitory effect of T N F a antibodies on synovial cell interleukin- 1 production in rheumatoid arthritis. Lancet 11:244-247, 1989 4. Alvaro-Gracia JM, Zvaifler NJ, Firestein GS: Cytokines in chronic inflammatory arthritis. IV. Granulocyte/ macrophage colony-stimulating factor-mediated induction of class I1 MHC antigen on human monocytes: a possible role in rheumatoid arthritis. J Exp Med 170: 865-875, 1989 5. Dinarello CA: Interleukin-I and its biologically related cytokines. Adv Immunol 44: 153-205, 1989 6. Miossec P: The role of interleukin 1 in the pathogenesis of rheumatoid arthritis. Clin Exp Rheumatol 5:305-308, 1987 7. Dayer J-M, de Rochemonteix B, Burrus B, Demczuk S , Dinarello CA: Human recombinant interleukin 1 stimulates collagenase and prostaglandin E, production by human synovial cells. J Clin Invest 77:645-648, 1986 8. EvCquoz V, Bettens F, Kristensen F, Trechsel U, Stadler BM, Dayer J-M, de Weck AL, Fleisch H: Interleukin 2-Independent stimulation of rabbit chondrocyte collagenase and prostaglandin E, production by an interleukin I-like factor. Eur J Immunol 14:490495, 1984 9. Postlethwaite AE, Raghow K, Stricklin GP, Poppleton H, Seyer JM, Kang AH: Modulation of fibroblast functions by interleukin 1: increased steady-state accumulation of type I procollagen messenger RNAs and stimulation of other functions but not chemotaxis by human recombinant interleukin l a and p. J Cell Biol 106: 31 1-318, 1988 10. Raines EW, Dower SK, Ross R: Interleukin-1 mitogenic activity for fibroblasts and smooth muscle cells is due to PDGF-AA. Science 243:393-396, 1989

CYTOKINES AND RA 1 1 . Krane SM, Dayer J-M, Simon L E , Byrne MS: Mononuclear cell-conditioned medium containing mononuclear cell factor (MCF), homologous with interleukin 1, stimulates collagen and fibronectin synthesis by adherent rheumatoid synovial cells: effects of prostaglandin E, and indomethacin. Coll Relat Res 5:99-117, 1985 12. Yaron I, Meyer FA, Dayer J-M, Bleiberg I, Yaron M: Some recombinant human cytokines stimulate glycosaminoglycan synthesis in human synovial fibroblast cultures and inhibit it in human articular cartilage cultures. Arthritis Rheum 32:173-180, 1989 13. Goldring MB, Birkhead J , Sandell LJ, Kimura T, Krane SM: Interleukin I suppresses expression of cartilagespecific types I1 and IX collagens and increases type I and I11 collagens in human chondrocytes. J Clin Invest 82:202&2037, 1988 14. Dayer J-M, Trentham D E , Krane SM: Collagens act as ligands to stimulate human monocytes to produce mononuclear cell factor (MCF) and prostaglandins (PGE,). Coll Relat Res 2:523-540, 1982 15. Dayer J-M, Rikard-Blum S, Kaufmann M-T, Herbage D: Type IX collagen is a potent inducer of PGE, and interleukin I production by human monocyte macrophages. F E B S Lett 198:208-212, 1986 16. Arner EC, Pratta MA: Independent effects of interleukin-I on proteoglycan breakdown, proteoglycan synthesis, and prostaglandin Ez release from cartilage in organ culture. Arthritis Rheum 32:288-297, 1989 17. Pettipher E R , Higgs GA, Henderson B: Interleukin 1 induces leukocyte infiltration and cartilage proteoglycan degradation in the synovial joint. Proc Natl Acad Sci USA 83:8749-8753, 1986 18. Stimpson S A , Dalldorf FG, Otterness IG, Schwab JH: Exacerbation of arthritis by IL-I in rat joints previously injured by peptidoglycan-polysaccharide. J Immunol 140:2964-2969, 1988 19. Hom JT, Bendele AM, Carlson DG: In vivo administration with IL-1 accelerates the development of collageninduced arthritis in mice. J Immunol 141:834-841, 1988 20. Jacobs C, Young D, Tyler S, Callis G , Gillis S, Conlon PJ: In vivo treatment with IL-I reduces the severity and duration of antigen-induced arthritis in rats. J Immunol 141 :2967-2974, 1988 21. Wallach D, Holtmann H , Engelmann H , Nophar Y: Sensitization and desensitization to lethal effects of tumor necrosis factor and IL-1. J Immunol 140:2994 2999, 1988 22. Henderson B, Rowe FM, Bird CR, Gearing AJH: Production of interleukin 1 in the joint during the development of antigen-induced arthritis in the rabbit. Clin Exp Immunol 74:371-376, 1988 23. Hopkins SJ, Humphreys M , Jayson MIV: Cytokines in synovial fluid. I. The presence of biologically active and immunoreactive IL-I . Clin Exp Immunol 72:422427, 1988

313

24. Yamagata N. Kobayashi K, Kasama T , Fukushima T, Tabata M , Yoneya I , Shikama Y, Kaga S, Hashimoto M, Yoshida K , Sekine F, Negishi M, Ide H , Mori Y, Takahashi l': Multiple cytokine activities and loss of interleukin 2 inhibitor in synovial fluids of patients with rheumatoid arthritis. J Rheumatol 15:1623-1627, 1988 25. Bhardwaj N , Lau L L , Rivelis M, Steinman RM: Interleukin- 1 production by mononuclear cells from rheumatoid synovial effusions. Cell Immunol 114:405423, 1988 26. Miyasaka N , Sato K , Goto M , Sasano M, Natsuyama M. Inoue K , Nishioka K: Augmented interleukin-I production and HLA-DR expression in the synovium of rheumatoid arthritis patients: possible involvement in joint destruction. Arthritis Rheum 3 1 :480486, 1988 27. Buchan G, Barrett K , Turner M , Chantry D, Maini RN, Feldmann M: Interleukin-I and tumour necrosis factor mRNA expression in rheumatoid arthritis: prolonged production of IL-la. Clin Exp Immunol 73:449455, 1988 28. Firestein GS. Alvaro-Gracia JM, Zvaifler NJ: Quantitative analysis of cytokine gene expression in rheumatoid arthritis (abstract). Clin Res 37:586A, 1989 29. Eastgate JA, Symons JA, Wood N C , Grinlinton F M , di Giovinc FS, Duff GW: Correlation of plasma interleukin 1 levels with disease activity in rheumatoid arthritis. Lancet II:70&709, 1988 30. Shore A, Jaglal S, Keystone EC: Enhanced interleukin 1 generation by monocytes in vitro is temporally linked to an early event in the onset or exacerbation of rheumatoid arthritis. Clin E x p Immunol 65:293-302, 1986 31. Dayer J-M, Seckinger P: Natural inhibitors and antagonists of interleukin 1 , Interleukin 1, Inflammation and Disease. Edited by RHR Bomford, B Henderson. Amsterdam, Elsevier, 1989 32. Arend WP, J o s h F G , Thompson RC, Hannum CH: An interleukin 1 inhibitor from human monocytes: production and characterization of biological properties. J Imrnunol 143:1851-1858, 1989 33. Larrick JW: Native interleukin 1 inhibitors. Immunol Today 10:61-66, 1989 34. Arend WP, Joslin FG, Massoni RJ: Effects of immune complexes on production by human monoeytes of interleukin 1 or an interleukin 1 inhibitor. J Immunol 134: 3868-3875, 1985 35. Balavoine J-F, de Rochemonteix B, Williamson K , Seckinger P, Cruchaud A , Dayer J-M: Prostaglandin E, and collagenase production by fibroblasts and synovial cells is regulated by urine-derived human interleukin 1 and inhibitor(s). J Clin Invest 78:1120-1124, 1986 36. Seckinger P, Williamson K , Balavoine J-F, Mach B, Mazzei G , Shaw A , Dayer J-M: A urine inhibitor of interleukin 1 activity affects both interleukin l a and 18 but not tumor necrosis factor a. J Immunol 139: 1541-1545, 1987

3 14

37. Seckinger P, Lowenthal JW, Williamson K, Dayer J-M, MacDonald HR: A urine inhibitor of interleukin 1 activity that blocks ligand binding. J Immunol 139:1546-1549, 1987 38. Roux-Lombard P, Modoux C, Dayer J-M: Production of interleukin 1 (IL-I) and a specific 1L-1 inhibitor during human monocyte-macrophage differentiation: influence of GM-CSF. Cytokine 1:45-51, 1989 39. Lotz M, Carson DA: Transforming growth factor p and cellular immune responses in synovial fluids (abstract). Arthritis Rheum 32 (suppl4):S42, 1989 40. Lotz M, Tsoukas CD, Robinson CA, Dinarello CA, Carson DA, Vaughan JH: Basis for defective responses of rheumatoid arthritis synovial fluid lymphocytes to anti-CD3 (T3) antibodies. J Clin Invest 78:713-721, 1986 41. Roux-Lombard P, Modoux C, Dayer J-M: Inhibitors of IL-1 and T N F a activities in synovial fluids and cultured synovial fluid cell supernatants (abstract). Calcif Tissue Int 42:S(A47), 1988 42. Prieur A-M, Kaufmann M-T, Griscelli C, Dayer J-M: Specific interleukin-1 inhibitor in serum and urine of children with systemic juvenile chronic arthritis. Lancet 11:1240- 1242, 1987 43. Sherry B, Cerami A: Cachectin/tumor necrosis factor exerts endocrine, paracrine, and autocrine control of inflammatory responses. J Cell Biol 107: 1269-1277, 1988 44. Dayer J-M, Beutler B, Cerami A: CachecWtumor necrosis factor stimulates collagenase and prostaglandin E, production by human synovial cells and dermal fibroblasts. J Exp Med 162:2163-2168, 1985 45. Henderson B, Pettipher ER: Arthritogenic actions of recombinant IL-1 and tumour necrosis factor a in the rabbit: evidence for synergistic interactions between cytokines in vivo. Clin Exp Immunol 75:306-310, 1989 46. Yocum DE, Esparza L, Dubry S, Benjamin JB, Volz R, Scuderi P: Characteristics of tumor necrosis factor production in rheumatoid arthritis. Cell Immunol 122: 131-145, 1989 47. Hopkins SJ, Meager A: Cytokines in synovial fluid. 11. The presence of tumour necrosis factor and interferon. Clin Exp Immunol 73:88-92, 1988 48. Saxne T, Palladino MA Jr, HeinegLd D, Tala1 N, Wollheim FA: Detection of tumor necrosis factor a but not tumor necrosis factor p in rheumatoid arthritis synovial fluid and serum. Arthritis Rheum 31: 1041-1045, 1988 49. Husby G, Williams RC Jr: Synovial localization of tumor necrosis factor in patients with rheumatoid arthritis. J Autoimmun 1:363-371, 1988 50. Seckinger P, Isaaz S, Dayer J-M: A human inhibitor of tumor necrosis factor a.J Exp Med 167: 1511-1516, 1988 51. Seckinger P, Isaaz S, Dayer J-M: Purification and biologic characterization of a specific tumor necrosis factor a inhibitor. J Biol Chem 264: 11966-1 1973, 1989

AREND AND DAYER

52. Engelmann H, Aderka D, Rubinstein M, Rotman D, Wallach D: A tumor necrosis factor-binding protein purified to homogeneity from human urine protects cells from tumor necrosis factor toxicity. J Biol Chem 264: 11974-1 1980, 1989 53. Olsson I, Lantz M, Nilsson E, Peetre C, Thysell H, Grubb A, Adolf G: Isolation and characterization of a tumor necrosis factor binding protein from urine. Eur J Haematol 42:270-275, 1989 54. Novick D, Engelmann H , Wallach D, Rubenstein M: Soluble cytokine receptors are present in normal urine. J Exp Med 170:1409-1414, 1989 55. Wong GG, Clark SC: Multiple actions of interleukin 6 within a cytokine network. Immunol Today 9:137-139, 1988 56. Guerne P-A, Zuraw BL, Vaughan JH, Carson DA, Lotz M: Synovium as a source of interleukin 6 in vitro: contribution to local and systemic manifestations of arthritis. J Clin Invest 83585-592, 1989 57. Houssiau FA, Devogelaer J-P, van Damme J, Nagant de Deuxchaisnes C, van Snick J: Interleukin-6 in synovial fluid and serum of patients with rheumatoid arthritis and other inflammatory arthritides. Arthritis Rheum 31 :78C 788, 1988 58. Hirano T, Matsuda T, Turner M, Miyasaka N, Buchan G, Tang B, Sat0 K, Shimizu M, Maini R, Feldmann M, Kishimoto T: Excessive production of interleukin 6/B cell stimulatory factor-2 in rheumatoid arthritis. Eur J Immunol 18:1797-1801, 1988 59. Bhardwaj N, Santhanam U, Lau LL, Tatter SB, Ghrayeb J, Rivelis M, Steinman RM, Sehgal PB, May LT: IL-6/IFN-P2 in synovial effusions of patients with rheumatoid arthritis and other arthritides: identification of several isoforms and studies of cellular sources. J Immunol 143:2153-2 159, 1989 60. Miyasaka N , Sato K, Hashimoto J , Kohsaka H , Yamamot0 K, Goto M, Inoue K, Matsuda T, Hirdno T, Kishimoto T, Nishioka K: Constitutive production of interleukin 6/B cell stimulatory factor-2 from inflammatory synovium. Clin Immunol Immunopathol 52: 238-247, 1989 61. Burmester GR, Jahn B, Rohwer P, Zacher J, Winchester RJ, Kalden JR: Differential expression of Ia antigens by rheumatoid synovial lining cells. J Clin Invest 80: 595-604, 1987 62. Firestein GS, Zvaifler NJ: Peripheral blood and synovial fluid monocyte activation in inflammatory arthritis. 11. Low levels of synovial fluid and synovial tissue interferon suggest that y-interferon is not the primary macrophage activating factor. Arthritis Rheum 30:864-871, 1987 63. Firestein GS, Xu W-D, Townsend K, Broide D, AlvaroGracia J, Glasebrook A, Zvaifler NJ: Cytokines in chronic inflammatory arthritis. I. Failure to detect T cell lymphokines (interleukin 2 and interleukin 3) and pres-

CYTOKINES AND RA ence of macrophage colony-stimulating factor (CSF-I) and a novel mast cell growth factor in rheumatoid synovitis. J Exp Med 168:1573-1586, 1988 64. Xu W-D, Firestein GS, Taetle R, Kaushansky K, Zvaifler NJ: Cytokines in chronic inflammatory arthritis. 11. Granulocyte-macrophage colony-stimulating factor in rheumatoid synovial effusions. J Clin Invest 83:876-882, 1989 65. Williamson DJ, Begley CG, Vadas MA, Metcalf D: The detection and initial characterization of colonystimulating factors in synovial fluid. Clin Exp Immunol 72:67-73, 1988 66. Hamerman D, Taylor S, Kirschenbaum I, Klagsbrun M, Raines EW, Ross R, Thomas KA: Growth factors with heparin binding affinity in human synovial fluid. Proc SOCExp Biol Med 186:384-389, 1987 67. Lafyatis R, Thompson N, Remmers E, Flanders K, Roberts A, Sporn M, Wilder RL: Demonstration of local production of PDGF and TGF-p by synovial tissue from patients with rheumatoid arthritis (abstract). Arthritis Rheum 31 (suppl 4):S62, 1988 68. Butler DM, Leizer T , Hamilton JA: Stimulation of human synovial fibroblast DNA synthesis by plateletderived growth factor and fibroblast growth factor. J Immunol 142:3098-3103, 1989 69. Lafyatis R, Remmers EF, Roberts AB, Yocum DE, Sporn MB, Wilder RL: Anchorage-independent growth of synoviocytes from arthritic and normal joints: stimulation by exogenous platelet-derived growth factor and

3 15

70.

71.

72.

73.

74.

75.

inhibition by transforming growth factor-beta and retinoids. J Clin Invest 83:1267-1276, 1989 Kumkumian GK, Lafyatis R, Remmers EF, Case JP, Kim S-J, Wilder RL: Platelet-derived growth factor and IL- 1 interactions in rheumatoid arthritis: regulation of synoviocyte proliferation, prostaglandin production, and collagenase transcription. J Immunol 143:833-837, 1989 Sporn MB, Roberts AB, Wakefield LM, de Crombrugghe B: Some recent advances in the chemistry and biology of transforming growth factor-beta. J Cell Biol 105:103% 1045, 1987 Czarniecki CW, Chiu HH, Wong GHW, McCabe SM, Palladino MA: Transforming growth factor-p modulates the expression of class I1 histocompatibility antigens on human cells. J Immunol 140:4217-4223, 1988 Fava R, Olsen N , Keski-Oja J, Moses H , Pincus T: Active and latent forms of transforming growth factor p activity in synovial effusions. J Exp Med 169:291-296, 1989 Lafyatis R, Thompson NL, Remmers EF, Flanders KC, Roche NS, Kim S-J, Case JP, Sporn MB, Roberts AB, Wilder RL: Transforming growth factor-p production by synovial tissues from rheumatoid patients and streptococcal cell wall arthritic rats: studies on secretion by synovial fibroblast-like cells and immunohistologic localization. J Immunol 143:1142-1 148, 1989 Wahl SM, Hunt DA, Bansal G, McCartney-Francis N , Ellingsworth L, Allen JB: Bacterial cell wall-induced immunosuppression: role of transforming growth factor p. J Exp Med 168:1403-1417, 1988