THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 8, Issue of March 15, pp. 5834-5839, 1993 Printed in W.S.A.

Expression of the Cytokine RANTES in Human Rheumatoid Synovial Fibroblasts DIFFERENTIAL REGULATION OF RANTES AND INTERLEUKIN-8 GENES BY INFLAMMATORY CYTOKINES*

(Received for publication, September 10, 1992)

Palaniswami Rathanaswami, Mohamed Hachicha, Michael SadickS, Thomas J. SchallS, and Shaun R. McCollg From the Centre de Recherche en Inflammation, Zmmunologie et Rhumatologie, Centre de Recherche du CHUL, Universite Laual, Qutbec, GI V 4G2, Canada, and SGenentech Znc., South San Francisco, California 94080

A chronic inflammatory disease may be character- ized by an accumulation of activated leukocytes at the site of inflammation. Since the chemokine RANTES may play an active role in recruiting leukocytes into inflammatory sites, we investigated the ability of cul- tured human synovial fibroblasts isolated from pa- tients suffering from rheumatoid arthritis to produce this chemokine and compared its regulation to that of the closely related chemokine gene, interleukin-8 (IL- 8). In unstimulated synovial fibroblasts, the expression of mRNA for both chemokines was undetectable, but was increased in both a time- and dose-dependent man- ner upon stimulation with the monokines tumor necro- sis factor a (TNFa) and interleukin-18 (IL-18). Prein- cubation of the cells with cycloheximide “superin- duced” the level of IL-8 mRNA stimulated by TNFa and IL-18 and RANTES mRNA stimulated by IL-18, but decreased the expression of RANTES mRNA in response to TNFa. In addition, differential regulation of these genes was noted when synovial fibroblasts were stimulated with a combination of cytokines. IL-4 down-regulated and IFNy enhanced the TNFa- and IL- 18-induced increase in RANTES mRNA, whereas the induction of IL-8 mRNA by TNFa or IL-18 was inhib- ited by IFNy and augmented by IL-4. Moreover, a combination of TNFa and IL-18 synergistically in- duced IL-8 mRNA expression, whereas under the same conditions, the level of expression of RANTES mRNA was less than that induced by TNFa alone. These ob- servations were also reflected at the level of chemokine secretion. These studies demonstrate that by express- ing the chemokines RANTES and IL-8, synovial fibro- blasts may participate in the ongoing inflammatory process in rheumatoid arthritis. In addition, the obser- vation that these chemokine genes are differentially regulated, depending upon the presence of different cytokines, indicates that the type of cellular infiltrate and the progress of the inflammatory disease is likely to depend on the relative levels of stimulatory and inhibitory cytokines.

* This study was supported by grants (to S. R. M.) from the Medical Research Council of Canada and the Arthritis Society of Canada and a scholarship from the “Fonds pour la Recherche en Santb du Qubbec” (to S. R. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

§ T o whom correspondence should be addressed Centre de Re- cherche en Inflammation, Immunologie et Rhumatologie, Centre de

G1V 4G2, Canada. Tel.: 418-654-2240; Fax: 418-654-2765. Recherche du CHUL, Universitb Laval, 2705 Blvd. Laurier, Qubbec,

Various cytokines have been implicated as important me- diators of inflammation and joint destruction in rheumatoid arthritis (RA)’ (1) and other inflammatory process. Recently, a group of small molecular weight proinflammatory cytokines, known variously as the platelet factor 4 (PF4), the intercrine, or most recently the “chemokine”2 superfamily, has been described (2,3). They are related by having primary structural similarities and a conserved 4-cysteine motif and are subdi- vided into two groups, namely C-X-C and C-C, depending on whether or not there is an intervening amino acid between the first two cysteines (2, 3). In general, the chemokines of the C-X-C class, of which IL-8 is a member, are reported to have proinflammatory functions through their actions mainly on neutrophils, whereas the chemokines in the C-C class of the superfamily appear to act on mononuclear cells with various degrees of specificity. For example, RANTES has been shown to be a chemotactic factor for monocytes and T lymphocytes of the memory phenotype in vitro (4). To date, RANTES expression has been primarily observed in cultured T cell lines that are antigen-specific and growth factor-de- pendent (5).

Culturing human cells in vitro has shown that IL-8 expres- sion is up-regulated in a variety of cells such as monocytes, tissue macrophages, endothelial cells, epithelial cells, and dermal fibroblasts in response to stimuli including lipopoly- saccharide, TNFa, and IL-lP (3). However, the lymphocyte- derived cytokines IL-4 and IFNy are reported to down-regu- late the lipopolysaccharide-, TNFa-, or IL-1-stimulated IL-8 expression in monocytes (3,6) and in human thymic epithelial cells (7), respectively. In contrast to the known distribution and regulation of IL-8 in many cell types, the cellular distri- bution or factors controlling the regulation of the RANTES gene have not yet been characterized with the exception that, by in situ hybridization, RANTES mRNA has been shown to be present in synovial lining cells of patients with RA (8).

Synovial fibroblasts are believed to play an important role in the pathogenesis of RA (1). These cells are known to produce a number of putative mediators of inflammation such as plasminogen activator, prostaglandin Ea, collagenase, stro- melysin (g-ll), and several leukocyte-activating cytokines including granulocyte/macrophage- and granulocyte-colony

’The abbreviations used are: RA, rheumatoid arthritis; IFNy, interferon y; IL-la, interleukin-la; IL-4, interleukin-4; IL-8, interleu- kin-8; RANTES, Regulated upon Activation, Normal T Expressed, and presumably Secreted; TNFa, tumor necrosis factor a; ELISA, enzyme-linked immunosorbent assay.

The term “chemokines” was decided upon at the Third Interna- tional Symposium on Chemotactic Cytokines, September 1992, Baden-by-Vienna, Austria.

5834

Expression of RANTES and IL-8 in Synovial Fibroblasts 5835

stimulating factors and IL-6 (12, 13) in response to the macrophage-derived cytokines TNFa and IL-1. In addition, recent studies in our laboratory3 and others (14) have also demonstrated that synovial fibroblasts produce the chemo- kines IL-8 and MCP-1 (15) in response to stimulation by either TNFa or IL-lp, or upon engagement of major histo- compatibility complex class I1 molecules with superantigen (16). It is therefore possible that these fibroblasts may play a major role in leukocyte recruitment to, and activation in, the rheumatoid synovium by producing chemokines. Moreover, the type of leukocyte infiltrate present in the RA synovium could depend on the relative levels of chemokines such as RANTES and IL-8, which in turn may depend on various stimuli present in the synovial environment. Hence, we have also studied potential similarities and differences in the reg- ulation of these two genes by a variety of cytokines that are reported to be present in the rheumatoid synovium.

MATERIALS AND METHODS

Reagents-TNFa (recombinant human TNFa, specific activity lo7 units/mg, in the actinomycin D-free L-929 cytotoxicity assay), was a generous gift from Knoll Pharmaceuticals (Whippany, NJ). TNFa stock was stored at -80 “C in phosphate-buffered saline containing 0.01% bovine serum albumin. Recombinant human IL-1P and IFNy were a generous gift from Genentech Inc. (South San Francisco, CA). Recombinant human IL-4 was a generous gift from the Genetics Institute (Cambridge, MA). Tests on these solutions using the Lim- ulus amoebocyte assay for the presence of endotoxin were negative. Cycloheximide and actinomycin D were purchased from Sigma. Col- lagenase, trypsin-EDTA, RPMI 1640, and Dulbecco’s modified Ea- gle’s medium were purchased from GIBCO (Burlington, Ontario, Canada). Hyclone fetal calf serum was purchased from Professional Diagnostics (Edmonton, Alberta, Canada). Hybond N membranes, [y-32P]ATP, and [a-32P]dCTP were obtained from Amersham Canada (Oakville, Ontario, Canada). RNA guard (RNase inhibitor) and ri- bosomal RNA standards (28 and 18 S ) were purchased from Phar- macia (Montrbal, Qubbec, Canada). All other reagents used in this study were of molecular biological grade.

Cell Culture-The synovial fibroblast cultures used in this study were derived from patients who were diagnosed as suffering from either definite or classical RA according to the American Rheumatism Association 1987 criteria (17) and who were undergoing total knee replacement due to joint deterioration. The cultures were prepared as described previously (18, 19). Primary cultures were released by treatment with 0.5% trypsin, 0.03 M EDTA at 37 “C for 4 min and subcultured into 75-cmZ flasks. Morphologically, the passaged cells resembled the previously reported “fibroblast-like” synovial cells (18, 19). The cells were used between passages 3 and 14 (20). Routine FACS analysis of cell surface markers revealed that after the third passage, the cultures were consistently major histocompatibility com- plex class II-negative, Mo-l-negative (the latter indicating absence of myeloid cells and the type A “macrophage-like” synoviocytes) and factor VIII-negative (endothelial cell marker).

Isolation of Cytoplasmic RNA and Northern Blot Analysis-cul- tured synovial fibroblasts were grown to confluence in 75-cm2 culture flasks and treated as described in the figure legends, After the appropriate stimulation time with the cytokines, the cells were har- vested using trypsin-EDTA, the cytoplasmic RNA was prepared, and Northern blots were performed as previously described (21). The probes used in this study were the 244-base pair PstIIEcoRI cDNA fragment representing the coding region of the IL-8 cDNA from nucleotides 49 to 293 (22) and a 410-base pair EcoRI/ApaI cDNA fragment spanning the coding region of RANTES (5). The cDNA probes were radiolabeled with [a-32P]dCTP using the random primer DNA labeling system (GIBCO/BRL, Burlington, Ontario, Canada). To confirm equal loading of RNA, the membranes were rehybridized with a synthetic oligonucleotide (5’-GTTGGTTTCTTTTCCTC-3’) for 28 S ribosomal RNA, constructed from the known consensus sequences for 28 S RNA (Diethard Tautz, Institut fur Zoologic, Universitat Miinchen). The oligonucleotide was radiolabeled with [y- 32P]ATP using a 5’ DNA terminus labeling system (GIBCO/BRL,

P. Rathanaswami, M. Hachicha, W. L. Wong, T. J. Schall, and S. R. McColl, submitted for publication.

Burlington, Ontario, Canada). The- Northern blots shown in this report are taken from one experiment, which is representative of at least four others performed with similar results on the synovial fibroblasts of different donors. Some donor-to-donor variability with respect to the relative levels of TNFa- and IL-18-induced RANTES and IL-8 gene expression was observed.

ELISA for IL-8 and RANTES Chernokines-Cultured human syn- ovial fibroblasts were grown to confluence in 75-cm2 culture flasks and treated as described in the figure legends. The supernatants were collected and analyzed for chemokine content. Levels of IL-8 were measured using an IL-8 monoclonal antibody sandwich ELISA em- ploying two anti-IL-8 monoclonal antibodies recognizing different, non-competing determinants. Briefly, microtiter plates were coated with the first anti-IL-8 antibody, and then the samples and the second anti-IL-8 antibody conjugated to horseradish peroxidase were added. After washing, substrate was added and the absorbance was measured at 490 nm in a Vmax plate reader (Molecular Devices Corporation, Menlo Park, CA). A standard curve was generated by plottingA,% nm versus log rhIL-8 concentration, using a four-param- eter logistic curve fitting program (developed at Genentech Inc.). The specificity of the assay for IL-8 was verified using 13 other soluble immune proteins, including interleukins 1-6, and was found to have no cross-reactivity with any of these proteins tested. RANTES levels were assessed by a similar assay, the details of which are to be published elsewhere.‘

RESULTS

The Effect of TNFa and IL-lp on Steady State Levels of RANTES and IL-8 mRNA in Synovial Fibroblasts-Rheu- matoid synovial fibroblasts were grown to confluence in 75- cm2 culture flasks and stimulated either for increasing periods of time with fixed concentrations of TNFa or IL-lp or with increasing concentrations of TNFa or IL-la for 24 h, and Northern blots were performed. No RANTES mRNA was detected, regardless of the time point (data not shown), unless the cells were stimulated with cytokines (Figs. 1 and 2). However, RANTES mRNA was first detected after a stimu- lation of 6 h with TNFa or 2 h with IL-lp and the maximum levels were reached at 24 h and sustained with up to 48 h (Figs. 1A and 2 A , respectively). At 72 h, the level was de- creased to one-half of the maximum level (results not shown). The filters were then stripped and reprobed with an IL-8 cDNA fragment. A similar time course of induction of IL-8 mRNA in response to the monokines was observed. We then examined the effect of concentration of TNFa and IL-1@ on RANTES gene expression following a 24-h incubation. The steady state level of RANTES mRNA was increased in a dose- dependent manner upon incubation with either TNFa or IL- 1p (Figs. 1B and 2B, respectively). A similar dose response to the monokines was observed for IL-8 mRNA expression (Figs. 1B and 2B).

The Effect of Cycloheximide and Actinomycin D on the Monokine-induced Up-regulation of RANTES and IL-8 Gene Expression-To determine whether the TNFa- or IL-lp-in- duced up-regulation of RANTES and IL-8 genes was protein synthesis-dependent, we conducted experiments in presence of the protein synthesis inhibitor cycloheximide. Pretreat- ment with cycloheximide had no effect on the levels of RANTES or IL-8 mRNA in unstimulated synovial fibroblasts (Fig. 3, A and B) . In contrast, pretreatment with cyclohexi- mide led to a “superinduction” of IL-lp-induced RANTES and IL-8 gene expression. In addition, superinduction of TNFa-stimulated IL-8 gene expression was also observed in presence of cycloheximide (Fig. 3B). In contrast, cyclohexi- mide pretreatment caused varying degrees of inhibition of the TNFa-induced RANTES mRNA expression (Fig. 3B). In this experiment, cycloheximide exerted only a small inhibitory effect. We also conducted experiments by preincubating the

M. Sadick, manuscript in preparation.

5836 Expression of RANTES and IL-8 in Synovial Fibroblasts

a A

RANTES -

0 0.1 1.0 10 100

'I'SFcr [Ilg/lnl]

FIG. 1. Effect of TNFa on steady state level of RANTES and IL-8 mRNA in human rheumatoid synovial fibroblasts. Cells were grown to confluence in 75-cm2 culture flasks and stimulated with 100 ng/ml of TNFn for increasing periods of time ( A ) or with increasing concentrations of TNFn for 24 h ( B ) . The cells were harvested, cytoplasmic RNA was prepared, and RANTES mRNA was then detected as described under "Materials and Methods." The filters were then stripped and reprobed successively with an IL-8 cDNA fragment and an oligonucleotide probe for 28 S ribosomal RNA. The data presented in these figures are representative of at least seven other experiments performed with similar results.

synovial fibroblasts with the transcription inhibitor actino- mycin D to determine whether the up-regulation of RANTES and IL-8 by TNFa or IL-lP was occurring at the transcrip- tional level. Actinomycin D completely inhibited the induc- tion of RANTES and IL-8 mRNA expression in response to both monokines (Fig. 3A, shown only for IL-lP).

Differential Regulation of RANTES and IL-8 Gene Expres- sion in Synovial Fibroblasts-Since the synovial environment is likely to contain a mixture of several cytokines, we inves- tigated the effect of combinations of various cytokines on the expression of RANTES and IL-8 genes. Cultured synovial fibroblasts were therefore stimulated either with TNFa or IL- 18 in combination with IL-4 or IFNy or with TNFa in combination with IL-lP. As shown in Figs. 4 and 5, the

RANTES - -mmm

0 0.1 1.0 3.0 I O 30

I L l p [llg/Illl]

FIG. 2. Induction of steady state level of RANTES and IL- 8 mRNA by IL- 1j3 in human rheumatoid synovial fibroblasts. Cells were grown to confluence in 75-cm2 culture flasks and stimulated with 3 ng/ml of IL-10 for increasing periods of time ( A ) or increasing concentrations of IL-lD for 24 h ( B ) . RANTES and IL-8 mRNA and 28 S ribosomal RNA were then detected as described in Fig. 1. The data presented in these figures are representative of a t least seven other experiments performed with similar results.

expression of RANTES and IL-8 mRNA was differentially regulated by different combinations of these cytokines. IFNy had no direct effect on the level of either RANTES or IL-8 mRNA (Fig. 4, lanes 2-4). However, IFNy synergistically enhanced the stimulatory effect of TNFa and IL-lP on the level of RANTES mRNA but inhibited that on IL-8 mRNA in a dose-dependent manner (Fig. 4, lanes 6-8 and 10-12). Similarly to IFNy, IL-4 exerted no direct effect on the level of either RANTES or IL-8 mRNA (Fig. 5A, lanes 2-4). However, in contrast to IFNy, IL-4 decreased the stimulating effect of TNFa on the RANTES mRNA level but synergisti- cally increased the effect of TNFa on the level of IL-8 mRNA in a dose-dependent manner (Fig. 5A, lanes 6-8). An identical result was obtained for RANTES and IL-8 gene regulation when the cells were stimulated with IL-1P in combination with IL-4 (data not shown). The effect of combined presence of TNFa and IL-1/3 on synovial fibroblasts was also examined

Expression of RANTES and IL-8 in Synovial Fibroblasts 5837 a

RANTFS -

I I A -

21s -

I L I @

CllX

ACD

6

R A N I F S -

I I A ? -

21s -

TN Fa

CllX

” - + + + - + - - + -

” + - - +

” + + - + - +

FIG. 3. The effect of cycloheximide and actinomycin D on IL-18- or TNFa-induced steady state level of RANTES and IL-8 mRNA in human synovial fibroblasts. Confluent cells were princubated with 10 pg/ml cycloheximide or with 5 pg/ml actinomycin D for 1 h as indicated. The cells were then incubated with either diluent (lanes 1-3 in A and lanes 1 and 2 in B ) or 3 ng/ml IL-lP (lanes 4-6 in A ) or 100 ng/ml T N F a (lanes 3 and 4 in B ) for 24 h, harvested, and analyzed for RANTES and IL-8 mRNA and 28 S ribosomal RNA as described under “Materials and Methods.” This figure is from one experiment, which is representative of six others performed with similar results. Different times of exposure of the autoradiographic images for each mRNA species in this figure were chosen to clearly demonstrate the level of superinduction by cyclo- heximide.

(Fig. 5B). Incubation of cells with IL-18 alone (Fig. 5B, lanes 2-4) or TNFa alone (Fig. 5B, lane 5 ) stimulated the expres- sion of both RANTES and IL-8 mRNA as described above. However, IL-18 in combination with TNFa dose-dependently decreased the TNFa-induced RANTES mRNA level but syn- ergistically enhanced the TNFa-induced IL-8 mRNA level (Fig. 5B, lanes 6-8).

Inflammatory Cytokines Differentially Regulate the Secre- tion of RANTES and IL-8 in Synovial Fibroblasts-The ob- servations reported above were also verified at the level of protein secretion. Cultured human synovial fibroblasts were

RANTFS -

I I A - -m.I,.I,-“

28s- mmcra”Wm*”w-

I F S ~ (llg/llll) o I 10 IO o I In IO o I 10 100

IS Fw

11,1p

” ” + + + + - - - -

” ” - - - - + + + +

FIG. 4. The effect of IFNr on the TNFa- and IL-18-induced steady state level of RANTES and IL-8 mRNA in human synovial fibroblasts. Confluent cells were incubated with increas- ing concentrations of IFN-y either without or with 100 ng/ml TNFa or 3 ng/ml IL-10 for 24 h as indicated. The cells were then harvested and analyzed for RANTES and IL-8 mRNA and 28 S ribosomal RNA as described under “Materials and Methods.” This figure is from one experiment, which is representative of five others performed with similar results.

incubated for 24 h with either diluent, TNFa, IL-18, or TNFa and IL-18. The cells were also incubated with IFNy or IL-4 in the presence or absence of TNFa or IL-18 (Fig. 6). At the end of incubation, the supernatants were collected and ana- lyzed for the presence of RANTES and IL-8 by ELISA. It was observed that the effect of the cytokines on RANTES and IL-8 secretion, either alone or in combination, followed an identical pattern to their effect on the expression of RANTES and IL-8 mRNA.

DISCUSSION

RANTES and IL-8 are members of a recently discovered superfamily of low molecular weight immunoregulatory me- diators, known as “chemokines.” This superfamily includes platelet factor 4 (PF4), monocyte chemotactic protein-1 (MCP-l), and the macrophage inflammatory proteins (MIP) la and l@ (2, 3, 23-25). Studies on the chemotactic activity of several of these proteins in vitro have indicated relatively rigid patterns of target cell selectivity. For example, IL-8 is chemotactic for neutrophils, as well as some T cell subsets (26, 27), but not for monocytes. In contrast, MCP-1 is highly specific for monocytes (28, 29), whereas RANTES is a che- motactic factor for monocytes and T lymphocytes of the memory phenotype (4). Since the accumulation of activated T cells and macrophages at an inflamed site such as the synovial tissue of patients with RA may lead to significant structural damage to the joints, we examined the regulation of RANTES gene expression and its production in synovial fibroblasts, in response to stimulation by cytokines that are likely to be present in the synovial environment.

RANTES was originally characterized as a T cell-specific gene, which was expressed in cultured CD8’ CTL and also in Ag- or mitogen-activated T cells (5). Recent in situ hybridi- zation studies have shown that RANTES mRNA is present in synovial lining cells (8). In the present study, we show that cultured human synovial fibroblasts express and secrete RANTES in response to activation by the monokines TNFa and IL-18. This observation, combined with the fact that these cells are also capable of producing other chemotactic cytokines such as IL-8 (Refs. 3, 14, and 16, footnote 3, and present study) and MCP-1 (15, 30) in response to various stimuli including TNFa or IL-lfi, may be of major relevance

5838 A

RANTLS -

11r8 -

28s -

B

RANl*KS -

11,s -

28s -

ILIg (ng/ml)

TN Lrr

Expression of RANTES and IL-8 in Synovial Fibroblasts

1 T n l-

a

0

0 0.1 I I O 0 0.1 1 I O

- ” - + + + + TNFa - + - + - + - - - + IL-1B - - + + - - + - - IFN7 - - - - + + . I ” - I L - ~ - - - - - - - + +

FIG. 6. Effect of cytokine combination on the secretion of RANTES and IL-8 chemokines in human rheumatoid synovial

and stimulated for 24 h with TNFn (100 ng/ml), IL-10 (3 ng/ml), fibroblasts. Cells were grown to confluence in 75-cm’ culture flasks

1FN-y (100 ng/ml), and IL-4 (10 units/ml) either alone or in combi- nation as indicated. The supernatants were collected, and RANTES and IL-8 protein were detected by specific ELISA as described under “Materials and Methods.”

0 0.1 I 3 0 0.1 I 3

+ + + + ” ”

FIG. 5. Effect of IL-4 and IL-16 on TNFa-induced steady state level of RANTES and IL-8 mRNA in human synovial fibroblasts. The cells were grown to confluence and incubated with increasing concentrations of IL-4 ( A ) or increasing concentrations of IL-lP ( R ) either in combination with or without 100 ng/ml TNFn for 24 h as shown. The cells were then harvested and analyzed for RANTES and IL-8 mRNA and 28 S ribosomal RNA as described under “Materials and Methods.” This figure is from one experiment, which is representative of four others performed with similar results. Different times of exposure of the autoradiographic images for each mRNA species in E were chosen to clearly demonstrate the effect of the combined treatment of monokines.

to the pathogenesis of RA, particularly in the early stages of the disease, where such chemokines could play a major role in recruiting and activating peripheral blood leukocytes.

Studies to date indicate that chemokine gene expression is controlled a t several levels, including the transcriptional and post-transcriptional levels. The results of our studies in syn- ovial fibroblasts using actinomycin D and cycloheximide sup- port previous observations made in various other cell types concerning the regulation of the IL-8 gene (31-33). It has been shown that IL-8 promoter activity is positively driven by nuclear factor KB and CREB-like trans-acting factors (33) and possibly in a negative manner by (an) as yet unidentified repressor(s) (33). Moreover, IL-8 mRNA contains several AUUUA sequences (34) and may therefore be controlled a t

the level of mRNA stabilization (35). Although the mecha- nism of IL-1P-induced RANTES gene expression appears to closely resemble that of the IL-8 gene, the RANTES gene may be regulated by mechanisms other than (de)stabilization, since unlike IL-8, the 3”untranslated region of RANTES mRNA does not possess AUUUA sequences (5). Our results also suggest that TNFa- but not IL-10-induced RANTES mRNA expression involves the synthesis of one or more proteins. We are presently conducting experiments to deter- mine which of these mechanisms (transcriptional or post- transcriptional) are predominantly involved in regulating the expression of these two chemokine genes.

The differential regulation of the two chemokine genes in synovial fibroblasts by inflammatory cytokines may deter- mine the type of cellular infiltrate present in the rheumatoid synovium. As we have shown in the present report for IL-8 production, TNFa and IL-1P synergistically up-regulate the production of colony stimulating factors (granulocyte and granulocyte/macrophage) (12) and prostaglandins (36) in syn- ovial fibroblasts and in other human cellular systems (37,38). However, a relative inhibitory effect of this monokine com- bination on the expression of other cytokine genes, as ob- served for RANTES gene expression, has not yet been re- ported. This may be an important observation with respect to the progress of RA. When a critical number of monocytes have accumulated and been activated in the synovium, a certain level of TNFn and IL-lP may accumulate in the synovial tissue. The combined level of these two monokines may alter the balance of RANTES to IL-8 production, result- ing in a relative enhancement of neutrophil recruitment and a relative decrease in mononuclear cell accumulation. This observation may partly explain the transient “acute flares” involving a massive neutrophil infiltration into the rheuma- toid joint often observed during the disease.

The monokine-stimulated expression of the RANTES and IL-8 genes and the secretion of two chemokines wcrc differ- entially regulated by IFNy and IL-4. TNFa- and IL-10-

Expression of RANTES and IL-8 in Synovial Fibroblasts 5839

induced secretion and m~~~ expression of I L - ~ were inhib- 13. Tan, P. L. J., Farmiloe, S., Yeoman, S., and Watson, J. D. (1990) J .

ited and those of RANTES were synergistically enhanced by 14. Koch, A. E., Kunkel, S. L., Burrows, J. C., Evanoff, H. L., Haines, G. K.,

IFN?', whereas IL-4 exerted the.opposite effect. These 15. Hachicha, M., Rathanaswami, P., Schall, T. J., and McColl, S. R. (1993) bined observations are in keeping with previous results show- Arthritis Rheum., in press ing that the T cell products, IL-4 and IFNy, have opposing 16. Mourad, W. M., Mehindate, K., Schall, T. J., and McColl, S. R. (1992) J.

effects on TNFa- or IL-lfi-induced cellular functions (39, 40) 17. Arnett, F. C., Edworthy, S. M., Bloch, D. A., McShane, D. J., Fries, J. F., Exp. Med. 176,613-616

including the production of cytokines such as IL-lB (41, 42) Coo er, N. S , Healey, L. A., Kaplan, S. R., Liang, M. H., Luthra, H. S., Me& er Jr.,' T. A., Mitchell, D. M., Neustadt, D. H., Pinals, R. S.,

and granulocyte-colony stimulating factor (43). Thus, the Schalfer, J. G., Sharp, J. T., Wilder, R. L., and Hunder, G. G. (1988) present study extends the observation Of Opposing effects Of 18. Dayer J."., Krane, S. M., Russell, G. G., and Robinson, D. R. (1976)

Arthritis Rheum. 31,315-324

Pro;. Natl. Acad. Sci. U. S. A. 73,945-949

human synovial fibroblasts to include the regulation Of IL-8 20. Lafyatis, R., Remmers, E. F., Roberts, A. B., Yocum, D. E., Sporn, M. B.,

caution should be taken when arguing a general anti-inflam- Laboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory, told

they differentially regulate chemokine gene expression. 23. Yoshimura T. Yuhki N. Moore S. K., Appella, E., Lerman, M. I., and In view Of the different spectra Of activity Of the various 24. Furutani, Y., Nomura, H., Notake, M., Oyamada, Y., Fukui, T., Yamada,

chemokines, the observations in the present study showing M., Larsen, C. G., Oppenheim, J. J., and Matsushima, K. (1989) Biochem.

Of the genes coding for RANTES and 25. Obaru, K., Fukuda, M., Maeda, S., and Shimada, K. (1986) J. Biochem. IL-8 may have important ramifications with respect to the (Tokyo) 99,885-894 control of the type of leukocyte infiltrate observed in the 26. Peveri, P., Walz, A., DeWald, B., and Baggiolini, M. (1988) J. Exp. Med.

rheumatoid joint or other inflammatory sites. Overall, we can 27. Larsen, C. G., Anderson, A. O., Amelia, E., Oppenheim, J. J., and Mat-

hypothesize that the combined presence and opposing effects 28. Yoshimura, T., and Leonard, E. J. (1990) J. Immunol. 144,2377-2383 of these monokines and lymphokines on chemokine gene 29. Zachariae, C. 0. C., Anderson, A. O., Thompson, H. L., Appella, E., expression might be a method by which leukocyte trafficking

Mantovani, A., Oppenheim, J. J., and Matsushima, K. (1990) J. Exp. Med. 171,2177-2182

to inflammatory sites is normally regulated. 30. Villiger, 797 P. M., Terkeltaub, R., and Lotz, M. (1992) J. Immunol. 149 , 722-

Rheumatol. 17,1608-1612

Pope, R. M., and Strieter, R. M. (1991) J. Imrnunol. 147,2187-2195

IFNy and 1L-4 On monokine-induced 'ytokine expression in 19, Hamilton, J, A,, and Slywka, J, (1981) J , Immunol. 126,851-855

and RANTES. Our suggest that 21. Sambrook, J., Frisch, E. F., and Maniatis, T. (1989) Moleculur CIonin A

matory "le for IFNy and 1L-4 in chronic as 22. Schmid, J., and Weissmann, C. (1987) J. Immunol. 139,250-256

and Wilder, R. L. (1989) J. Clin. Inuest. 8 3 , 1267-1276

Spring Harbor, NY

Leonard,'E. >. (1989') FhBS Lek. 244,487-493

Biophys. Res. Commun. 169,249-255

167,1547- 1559

sushima, K. (1989) Science 243 , 1464-1466

2. 1.

3.

4.

5.

6.

7. 8.

9.

10.

11.

12.

REFERENCES Arend, W. P., and Dayer, J.-M. (1990) Arthritis Rheum. 33,305-315 Schall, T. J. (1991) Cytokine 3 , 165-183 Oppenheim, J. J., Zachariae, C. 0. C., Mukaida, M., and Matsushima, K.

Schall, T. J., Bacon, K., Toy, K. J., and Goeddel, D. V. (1990) Nature 3 4 7 ,

Schall, T. J., Jongstra, J., Dyer, B. J., Jorgensen, J., Clayberger, C., Davis,

Standiford, T. J., Strieter, R. M., Chensue, S. W., Westwick, J., Kasahara,

Galy, A. H. M., and Spits, H. (1991) J. Immunol. 1 4 7 , 3823-3830 Schall, T. J., Lu, L. H., Gillett, N., and Amento, E. P. (1991) Arthritis

Leizer, T., Clarris, B., Ash, P., van Damme, J., Saklatvala, J., and Hamilton,

Dayer, J.-M., de Rochemonteix, B., Burrus, B., Demczuk, S., and Dinarello,

Case, J. P., Lafyatis, R., Kumkumian, G. K., Remmers, E. F., and Wilder,

Liezer, T., Cebon, J., Layton, J. E., and Hamilton, J. A. (1990) Blood 76,

(1991) Annu. Reu. Immunol. 9,617-648

669-671

M. M., and Krensky, A. M. (1988) J. Immunol. 141,1018-1025

K., and Kunkel, S. L. (1990) J. Immunol. 146,1435-1439

Rheum. 3 4 , S117

J. A. (1987) Arthritis Rheum. 30,562-566

C. A. (1986) J. Clin. Inuest. 77.645-648

R. L. (1990) J. Immunol. 146,3755-3761

1989-1996

31. Koialsky, J., and Denhardt, D. T. (1989) Mol. Cell. Biol. 9,1946-1957 32. Sica, A,, Matsushima, K., Van Damme, J., Wang, J. M., Polentarytti, N.,

PTjana, E., Colotta, F., and Mantovani, A. (1990) Immunology 6 9 , 548-

33.

34.

35. 36.

37. 38.

39.

40. 41.

42.

43.

Mukaida, N., Mahe, Y., and Matsushima, K. (1990) J. Biol. Chem. 266 ,

Matsushima, K., Morishita, K., Yoshimura, T., Lavu, S., Kobayashi, Y., 21128-21133

Lew, W., Appella, E., Kung, H. F., Leonard, E. J., and Oppenheim, J. J. (1988) J. Ex Med. 167,1883-1893

Shaw, G., andgamen, R. (1986) Cell 46,659-667 Elias, J. A,, Gustilo, K., Baeder, W., and Freundlich, B. (1987) J. Immunol.

Rug iero, V., and Baglioni, C. (1987) J. Immunol. 1 3 8 , 661-663 Stasienko, P., Dewhirst, F. E., Peros, W. J., Kent, R. L., and Ago, J. M.

(1987) J. Immunol. 138,1464-1468 Hart, P. H., Vitti, G. F., Burgess, D. R., Whitty G. A,, Piccoli D. S. and

Becker, S., and Daniel, E. G. (1990) Cell. Immunol. 129,351-362 Hamilton, J. A. (1989) Proc. Natl. Acad. Sci. -??. S. A. 86,3863-380:

Arend, W. P., Gordon, D. F., Wood, W. M., Janson, R. W., Joslin, F. G.,

Donnelly, R. P., Fenton, M. J., Kaufman, J. D., and Gerrard, T. L. (1991)

Hamilton, J. A., Piccoli, D. S., Cebon, J., Layton, J. E., Rathanaswami, P.,

DDJ

138,3812-3816

and Jameel, S. (1989) J. Immunol. 143,118-126

J. Immunol. 146,3431-3436

McColl, S. R., and Leizer, T. (1991) Blood 79 , 1413-1419