SPECIAL INVITED PAPER ON AUTOIMMUNITY D O M I N I Q U E J. C H A R R O N , G U E S T E D I T O R F O R T H E SERIES

Autocrine Stimulation of Interleuldn 1 in Human Adherent Synovial Lining Cells: Down Regulation by Interferon Gamma Nathalie Temime, Arlette Joliviere, Dan/61e Lando, Luc Teyton, and Dominique Charron

ABSTRACT: lntedeukin 1 (IL-1) exerts biological properties on various immune and nonimmune cell types and tissues and thus may play an important role during chronic inflammatory processes. Here we have examined the IL-1 biosynthesis in adherent synovial lining cell (ASLC) cultures obtained from patients with rheumatoid arthritis (RA). We report that ASLCs in culture showed heterogeneous endogenous levels of IL-I a and fl expression. Recombinant interleukin 1 (rlL-l) a or/8 induced increases of IL-1 a and/3 mRNA and proteins levels in ASLCs. Although IL-I synthesis is enhanced by rlL-1 treatment, no soluble IL-1 t* or/8 could be detected by specific enzyme-linkod immunosorbent assays. A pretreatment with recombinant IFN gamma (rlFN'D downABBREVIATIONS ASLC adherent synovial lining cells DMEM Dulbecco's Modified Eagle Medium ELISA enzyme-linked immunosorbent assay FCS fetal calf serum IFN'F interferon gamma IL-1 interleukin 1 LAF lymphocyte activating factor

regulated the effect of rlL-1 on IL-1 synthesis in ASLCs. Acdnomycin D supressed the endogeneous and r l b t induced IL-I mRNA expression lndometh~in, in the presence of rlbl ~ or ~, up-regulates the level of expression of !I.-1 /3 in ASLCs pretreated with rIFNy, but has the opposite effect in non-pretreated cells. The increase of IL-I gene expression by rlL-1 in human ASLCs from RA patients may contribute as an amplification of the disease progress. These studies may also explain the beneficial effects of IFN7 in experimental models of IL-l-induced bone and cartilage degradation and in patients with diseases involving ILl. Human Iratuundogy 31. 2 6 1 - 2 7 0 (1991)

MHC PBMC PGE2 RA rlFN7 riL-I

major histocompatibillty complex peripheral blood mononuclear cells prostaglandin E2 rheumatoid arthritis recombinant interferon gamma recombinant interleukin 1

INTRODUCTION Rheumatoid arthritis (RA) is an autoimmune inflammatory process leading to joint destruction. Adherent synovial lining cells (ASLC) from RA joints are character-

From the Laboratoire ~lmmunog/nltique Mole'culaire, lnstitut Biom#dicaldes Conldkrs, Paris, France(N.T.; A.J.; L.T.; D.C.), and Roussd Udaf, Romainville, France (D.L.). Address reprint requests to ProfessorDominique Charron, Laboratoire d'lmmunoglndtique lffollculaire, lnstitut Biomldical des Corddiers, 15 rue d¢ l'Ecde de Midecine, 75006, Paris, France. Accepted February 28, 199l.

HumanImmunologT31,261-270 (1991) © AmericanSocietyfor Histocompatibilityand lmmunogenetics.1991

ized by hyperprolfferation and expression o f major histocompatibility complex (MHC) class II antigens [11. This aberrant expression could depend on cytokines released by mononuclear cells infiltrating local lesions [2]. Interferon gamma (IFNy) is known to induce M H C class II antigens in many cell types, including synovial cells [3], and to inhibit some inttarnmatory reactions, especially those associated with prostaglandin E2 (PGE2) synthesis (as in RA) [4] or bone resorption [5, 6]. Because of these properties, IFN'y is under investigation in clinical trials in patients with R A [7]. 261 0198-8839/91/$3.50

262

Different cytokines might be involved in the initiation or propagation of the inflammatory process of RA. Interleukin 1 (II.-1) plays a key role in the immune response by activating T and B lymphocytes [8], and exerts biological effects on various nonimmune cell types. Interleukin 1 is present at elevated levels in joint effusions from RA [9] and other rheumatic diseases [10], and is suspected to be involved in their pathologic process. For example, it may contribute to proliferation and fibrosis of tissues involved in the pannus formation, as well as degradation of the joint by increasing the secretion of PGE2 and collagenase from synovial cells [11] and chondrocytes [12]. Synovial cells are not only targets for the action of IL-1 but can also produce IL-1 [13]. Two IL-1 genes have been cloned [14, 15] encoding for IL-1 c~and IL-1 proteins. Although IL-I c~and/3 have similar biological activities [16], IL-1/3 is predominantly expressed by activated human monocytes [17]. Several groups reported that IL-1 could undergo an autocrine amplification of its synthesis by monocytes [18], endothelial cells [19], vascular smooth muscle cells [20], and leukemic cells [21]. Furthermore, we have recently documented that recombinant IL-1 (rIL-1) c~and ~ induce IL-1/3 gene expression in dermal human fibroblasts [22]. In the light of the central role of synovial cells in the pathology of RA, the regulation of rIL-l-increased IL-1 biosynthesis in synovial lining cells was investigated. We showed that rIL-1 c, and 3 can increase IL-I a or gene expression, resulting in IL-1 activity in ASLC in vitro. Neither the IL-1 a nor the IL-I /3 stimulation resulted in a marked IL-1 release by ASLC. The effect of recombinant IFN~/(rlFN~) on rIL-1-increased IL-1 synthesis was examined. Synovial cells from RA patients lose their MHC class II antigens expression within 96 hr of culture, as previously described, and synthesis can be reinduced by treatment with rIFN3, [3, 23]. The behavior of IL-1 synthesis was studied in pretreated cells (MHC class II positive) compared with non-pretreated cells (MHC class II negative). Recombinant IFN7 antagonized the positive effect of rIL-1 on IL-1 synthesis. When rlFN7 pretreated cells were stimulated with rlL-1 in the presence ofindomethacin (to block the production of PGE2), the IL-1 gene expression level was slightly increased as compared with IL-1 stimulation alone.

rlFN7 Down Regulates IL-I Synthesis

specific activity of 2.2 107 U/rag. Interleukin ~ enzymelinked immunosorbent assay kits were purchased from Endogen Inc.; IL-1/3 ELISAs were carried out as previously described [24], using monoclonal antibodies to IL-1/3 (Roussel Uclaf).

Cell cultures. Synovial membranes were obtained from patients with classical RA as defined by the American Rheumatism Association criteria during knee or hip arthroplasty joint replacements. The patients had a nonactive state of the disease and have had diverse therapeutic treatments; however, none had intensive treatment at the time of surgery. Synovial membranes were dissected from synovial fragments before enzymatic dissociation by sequential treatment with collagenase type I (2.5 mg/ml, Sigma) for 2 hr and trypsin (0.25%, Byosis, S.A. France) for the second hour of the collagenase digestion. The cells were filtered through a 0.1mm size pore filter, washed three times in Dulbecco's Modified Eagle Medium (DMEM), and plated at 8 x 104 cells/cm 2 in DMEM supplemented with 10% fetal calf serum (FCS) containing 100 U/ml penicillin, 100 /zg/ml streptomycin, 2 mM glutarnine, 10 mM Hepes buffer, and 1 mM pyruvate (designated complete medium). After 24 hr incubation at 37°C in 5% CO2, nonadherent cells were removed and the adherent cell population was washed three times. Synovial cells in culture showed three different cell types identified by their morphological criteria: fibroblast-like, dendritic-like, and macrophage-like cells. Macrophage-like cells are eliminated after the first passage. We performed esterase staining on synovial cell preparations. At the first passage, we detected less than 1% of positive cells; at the second passage no positive cells could be detected. Synovial cells were duplicated after treatment with trypsin 0.1%, and the cultures were used between the second and the ninth passages. Expe:'imentalprocedure. Cells were grown to confluence in 175 cm2 culture flasks, harvested, washed with complete medium, and adjusted to 2 x 106 cells/ml. Cells were plated in six-well plates for RNA studies (8 x 106 cells/well) or in 24-well plates for protein assays (1 x 106 cells/well). One hour later, 1 vol of complete medium containing different combinations of indomethacin (2 gtg/ml) IL-1 c,, 3 was added.

RNA preparation and analysis. Cells incubated in sixATERIALS AND METHODS

Recombinant interleukin and other reagents. The specific activity of rIL-1 a (Genzyme Corp, USA) was 10s and 107 U/rag for rIL-1 /3 (Roussel Uclaf, RomainviUe, France). Recombinant IFN3, (Roussel Uclaf) showed a

well plates were harvested by pipetting up and down, centrifuged, and lysed in 5 M guanidium thiocyanate [25]. RNA was pelleted by centrifugation through cesium chloride cushions [26]. Total RNA (15/zg) was run in 0.75% agarose gels containing 2.2 M formaldehyde and transferred to Hybond N nylon membrane

N. Temime et al.

(Amersham, U.K.). Prehybridization was performed in 50% formamide 5 x SSC, 1 x Denhardt's, 0.1% SDS, 0.25 mg/ml of denatured salmon sperm DNA, and 10% Dextran sulfate for 1 hr at 42°C. Hybridization was performed for 18 hr in the same buffer containing 2 x 106 cpm/ml of 3lp radiolabeled probe [27]. cDNA probe for IL-1 ,~ and IL-1 [3 m R N A were provided by Roussel Uclaf. c D N A for rat glyceraldehyde-3-phosphate deshydrogenase was used as the control [28]. After hybridization, filters were washed at room temperature in 1 x SSC, 0.1% 8D8 (2 x 15 rain), then twice in 0.1 x 88C, 0.1% SDS for 30 rain at 55°C. Blots were autoradiographed for 1 day.

Interleukin 1 activity assay. Medium collected from cultures of ASLC (conditioned medium) was centrifuged 10 rain at 1000 g and stored at - 8 0 ° C until assay. Cell layers were collected by pipetting up and down, washed twice with cold phosphate-buffered saline, and centrifuged. The pellets were lysed in 0.5 ml of 10 mM tris HCI p H 7.5, centrifuged (10 rain at 1000 g), and the supernatants (cytosolic fractions) were assayed for lymphocyte activating factor (LAF) activity by the murine thymocyte costimulation assay [29]. At least three dilutions (1 : 8, 1 : 16, 1 : 32) of the test samples were incubated in 96-well plates with 1.5 x 106 thymocytes from C 3 H / H e J mice ( 6 - 8 weeks old) (CESEAL Lab., France) in a final volume of 0.2 ml of RPM11640 containing 5% FCS, phytohemagglutinin (1 /zg/ml) (Wellcome, England), antibiotics, glutamine (2 raM), and 2-mercaptoethanol (2 raM). The cells were cultured for 72 hr, and (3H) thymidine (1/zCi/well, 1 Ci/mmol) was added to the wells for the last 4 - 8 hr. Cells were harvested onto glass fiber filters. The filters were counted in scintillation fluid, and quench corrections were performed with an external standard. The results are reported as the mean --- SD of counts per rain determined on triplicate cultures.

Enzyme-linked immunosorbent assay. After incubation, cultures were recovered by pipetting up and down, and pelleted by centrifugation. Pellets (lysed in 0.15 ml of 10 mM tris HCI pH 7.5) and supernatants were assayed by ELISA for IL-1 c~ and IL-1 [3. The antibodies do not cross-react with each other nor do they detect other cytokines such as IL-2, IL-6, IFN% or tumor necrosis factor a. The ELISA assay for IL-1 a detects the active forms of IL-1 c, and the pro IL-1 e~ (31 and 22 kD), whereas the ELISA for IL-1 [3 allows the assay of all forms of IL-1 [3 (active and inactive precursors, membrane and soluble forms). Both assays can detect as few as 10-20 pg/ml of IL-1 ~ or IL-1 [3.

263

Analysis of PGEz levels. Prostaglandin Ez was assayed in the medium by radiohnmunoassay [30] by using specific ant/bodies (Instirut Pasteur Production, France). RESULTS

Effect of rlL-I ~ and [3 on IL-1 mRNA levels in ASLC in culture. Primary cultures of ASLC derived from RA patients constitutively transcribe mRNA for both species of IL-1 as assessed by northern blot analysis (Fig. 1), and the IL-1 cdIL-1 [3 m R N A ratio was constant in all donors. A 24-hr incubation of ASLC with rIL-1 [3 or ~ (10 U/ml) resulted in a marked increase of IL-1 18 steadystate level m R N A (Fig. 2A). This increase was found in both low and h/gh endogenous IL-1 gene expression cultures (Figs. 3 and 4). The rehybridizatlon of the blot with IL-1 e~ probe showed a similar amplificatior, of basal IL-I a gene expression (Fig. 2B). IL-1 mRNA level increased in ASLC as early as 1 hr of incubation with rlL-1 [3 or ~ (data not shown). The effect was maximal after 4 hr of incubation with rIL-1 ~ or 24 hr with rlL-1 [3 and then declined rapidly; rlL-1 [3 induced a greater amplification of IL-1 m R N A than rIL-1 c, (data not shown). Incubation of ASLC for 24 hr with the R N A synthesis inhibitor actinomycin D (2 #g/ml) in the absence or presence of rlL-1 ,~ or [3 suppressed the constitutive and the rIL-1-increased IL-1 [3 gene expression; the same effect was found on IL-1 c~ mRNA level (data not shown). These findings indicate that the FIGURE 1 Variation of IL-I ~ and IL-fl constitutive gene expression: confluent ASLC from four different donors were incubated for 24 hr under basal conditions. Total RNAs (15 ~g) were analyzed by Northern blot an~ysis. The blot was hybridized with ~2p-labeled cDNA probe for 1I.-1 fl (A) and rehybridized with IL-I exprobe (B). To assess that the RNA in each lane was intact and in equal amount, the blot was finally rehybridized with glyceraldehyde phosphate dehydrogenase probe (C).

ASs

B

--

GAPDH

264

I L - I / 3 __,

QmqD'"

.18S

B IL-I(~

rlL-1

FIGURE 2 Recombinant interleukin 1 increases IL-I gone expression in ASLC. Confluent cells were incubated for 24 hr in the absence (lane 1) or presence of rIL- 1 c~or/3 (10 U/ml). Total RNAs (10 #g) were analyzed as described in Figure 1. The blot was hybridized with IL-I B probe (A) and rehybridized with IL-1 a probe (B). Glyceraldehyde phosphate dehyd rogenase probe served as control (C). Lane 1: control; lane 2: rlI.-I a; lane 3: rlL-I ~.

@

~

..

185..~._GAPD

H

I

8'~ %~1

i

m =l

-

+

~

-

+

-

+~,

-

-

~"

G(,mu/ml)

rlL-|J3(,mu/mi)

FIGURE 3 (A, B) Total cellular IL-I synthesis induced by rlL-1 in ASLC in culture. Cells were incubated for 24 hr with different doses of rlL-1 a or rlL-1/3. Following the incubation, cell pellets were lyzed and immunoassayed for IL-1/3 (A) or IL-1/3 (B). (C) Time course of rlL-1 induced total cellular IL-1/3 synthesis in ASLC. Cells were incubated with rlL-1 a (10 U/ml) (O) or rlI.-1/3 (10 U/ml) (El) for different time periods; total cellular extract was immunoassayed for II.-1 synthesis. (D) Modulation of total cellular IL-I/3 synthesis by indomethacin. ASLC were treated as in (A) with rlL-1 a and/ or indomethacin (1 /~g/aJ). Total cellular extract was immunoassayed for II.-1 B. :15"

3°TA

"C

10"

~

O/~ control

I0

r IL-I

I

Ir

~--~r

25

50

0

,]

: : ~216

O.°4

(U/ml)

I 24

TIME

I

II 48

~

(hours)

IL-I a l p h a

IL-I

beta

3 ~2

B

__

control

I0

rIL-1

I

Ir

~rIL,-~

25

(U/ml) IL-I alpha beta

0

~0 rIL-~

25

alpha

Without With

50

(U/ml) InOo

InOo

(I

~g/ml)

N. Temime et al.

265

45

r

::t.-~.

(U/m],)

FIGURE 4 Lymphocyte activating factor activity induced by rILol ~ or ~ in ASLC. Cells were incubated 24 hr with rlL1 /3 (O) or rIL-I ~ (El) in complete medium. Cytosolic LAF activity was measured and was reported as mean cpm ± SD of (JH) thymidine incorporation in triplicate culture~, in the thymocyte costimulation assay. (3H) Thymidine incorporation in the presence of phytohemagglutinin alone was 815 ± 58 cpm.

rlL-l-induced increase in IL-1 m R N A levels in ASLC depends on de novo R N A synthesis.

Modulation of IL-1 gene expression by rlFN,/ and indomethacin. Adherent synovial lining cells were ~ubjected to a 4-day pretreatment with rIFN'y (100 U/ml) to induce M H C class II antigen expression lost in the first day of in vitro culture. The induction of Ia antigens was assessed by Western blot analysis as previously described [23]. Recombinant IFNy-pretreated cells (MHC class II positive) were compared with non-pretreated ASLC (MHC class II negative) for IL-1 gene expression. Pretreatment with rlFN~, lowers the steadystate level of IL-i m R N A compared with non-pretreated cells (Figs. 5A and B, lanes 1 and 5). Recombinant IFNy-pretreated ASLC incubated for 24 hr with rIL-1 a showed an increase of IL-1 m R N A level. However, the magnitude of the increase of IL-1 m R N A was diminished compared with that found in non-pretreated ASLC (Fig. 5A and B, lanes 1, 2, 5, and 6). Similar results were obtained with rlL-1 ~ (data not shown). Prostaglandin Ez is spontaneously synthesized and released by ASLC, and rIL-1 stimulates PGE2 release from many cell types, including ASLC [11]. Therefore, blockage experiments of PGE2 synthesis by indomethacin (an inhibitor of the cyclooxygenase pathway of arachidonate metabolism) were performed to determine whether PGE2 could influence the rIL-l-induced IL-1 genes expression in ASLC. Warner et al. [19] reported

that indomethacin did not significantly affect rlL-i B-induced IL-1 • m R N A level in human endothel/ai cells. We have recently shown in cultured human foreskin fibroblasrs that indomethacin reduced IL-1 3 geae expression [22]. Here, we found that a 24-hr incubation with 1 /~g/rnl indomethacin slightly decreases the steady-state level of IL-1 gene expression by MHC class II negative ASLC (non-pretreated rlFNy), whereas it has the opposite effect on M H C class II positive ASLC (pretreated by rIFNy) (Fig. 5, lanes 1, 3, 5, and 7). Indomethacin added concomitantly with rlL-1 ~x in MHC class II negative ASLC did not change the IL-1 gene expression level compared with IL-1 ¢x alone (Fig. 5, lanes 2 and 4), whereas in MHC class II positive ASLC it potentiated the effect of rlL-1 ~ to induce a slight but significant increase of IL-1 m R N A levels (Fig. 5, lanes 6 and 8). Similar results were obtained following a 24-hr treatment with rlL-1 B and iodomethacin FIGURE 5 Modulation of IL-1 gene expression by rlFNy and indomethacin. Confluent ASLC were divided into two groups: non-pretreated cells, maintained 4 days in complete medium (lanes 1-4 ), and rlFNy-pretreated cells: (1O0 U/ral, 4 days; lanes 5-8). After 4 days, pretreated- and nonpretreated ASLC were harvested, washed, and incubated for 24 hr in the presence of IL-I a (10 U/ml) and/or indomethacin (1 t~g/ml). RNA were analyzed as described in Figure 1. The blot was first hybridized with IL-I ~ probe (A), then with IL-1 a probe (B), and finally with glyceraldehyde phosphate dehydrogenase probe (data not shown). Lanes 1 and 5: control; lanes 2 and 6:IL-1 a; lanes 3 and 7: indomethacin; lanes 4 and 8: indomethacin + IL-1 a. A

|S$

ii.-|~_.. '1

2

:3

4

'1

2

3

4

rIFHs(~oou/~t)

S

6

7

8

5

6

7

s

+

+

÷

+

r|L-la(loU/mll

-

÷

-

4-

-

.t-

-

.l-

Ind0(~9/~l)

-

_

+

+

-

_

.

÷

266

rlFN7 Down Regulates IL-I Synthesis

TABLE 1 Modulation of PGE2 secretion by rIL-1/3 and Indomethacin PGEz (ng/ml) in supernatant

0 Pretreatment None None rlFNy (100 U/ml)

167 123

IL-1 g + IL-113 Indomethacin indomethacin 5077 1472

037. 0.16

0.05 0.05

Non-pretreated cells (MHC class II negative) and rIFN'y pretreated cells (MHC class II positive) were prepared as described in Fig. 6 and incubated for 24 hr with IL-I ~ ( 10 U/ml) and/or indomethacin (1/,tg/ml). Prostaglandin E2 accumulated was estimated by specific radioimmunoassay in the culture medium.

(data not shown). These data suggest that the inhibitory effect of a pretreatment by rIFN'F on rIL-l-induced IL-1 m R N A synthesis is lowered but not reversed by the presence of indomethacin at a concentration of 1 /.Lg/ml, which blocks PGE2 synthesis. Because PGE2 secretion induced by IL-1 synthesis exerts a negative regulatory feedbv.ck on rIL-1-induced IL-I synthesis in peripheral blood mononuclear cells (PBMC) [18] and vascular endothelial cells [19], the PGE2 secretion by ASLC stimulated with rIL-1 and/or indomethacin was investigated. Table 1 shows that MHC class II negative ASLC maintained 24 hr in culture synthesize a large amount of PGE2. When stimulated for 24 hr by rIL-1/3 (25 U/ml), this secretion was increased and totally suppressed by 1 tzg/ml indomethacin (Table 1). We have reported that the IL-1 gene expression increased by rIL-1/3 in M H C class II positive ASLC was lowered compared with that found in M H C class II negative ASLC. Table 1 shows a similar pattern with regard to PGE2 secretion. Similar results were obtained with rIL-1 a (data not shown).

Effect of rlL-1 on cellular lL-1 synthesis. The level of IL-1 /8 proteins synthesized by ASLC cultured in basal conditions was time dependent, with a maximum level after a 24-hr incubation. No IL-1 /3 could be detected within the culture supernatants. The basal level of total cellular IL-1 /8 in ASLC cultured for 24 hr was different from individual to individual, consistent with our Northern blot analysis data (not shown). A 24-hr treatment with rIL-1 ot or/3 showed a dose-dependent increase of all forms of IL-1 B, with a maximum response for 25 U/ml or rIL-1 (Fig. 3A). A similar dose-dependent increase in total cellular IL-1 c~ synthesis was detected (Fig. 3B). Conditioned media were assayed for the presence of soluble IL-1 c~ or/8 using specific ELISA, and no newly synthesized soluble IL-1/3 or IL-1 c~ could ever be detected (data not shown). A time course induction of

total cellular IL-1/3 by rIL-1 ot (10 U/ml) showed a peak around 12 hr, and around 24 hr for rIL-1/3 induction (10 U/ml) (Fig. 3C). We showed that indomethacin did not influence the rIL-l-increased IL-1 R N A level in M H C class II negative ASLC (Fig. 5). In contrast, ASLC treated 24 hr with indomethacin (1/zg/ml) and rIL-I a showed a decrease of total cellular IL-1/3 synthesis compared with rIL-1 ~x alone (Fig. 3D). Again, no soluble IL-1 /3 or cx were detected in the presence of indomethacin. Similar effects of indomethacin were obtained with regards to rIL-1/3-increased total cellular IL-1/8 (data not shown).

Increase of cvtosolic LAF activity by treatment with rlL-l. The thymocyte coproliferation assay was used to determine whether the IL-1 induced by rIL-1 treatment in ASLC was biologically active. Treatment of ASLC with rIL-1 /3 or c~ (25 or 50 U/ml) for 24 hr showed an increase of cytosolic LAF activity (Fig. 4). However, no LAF activity was found in supernatants, indicating no secretion of de novo IL-1 (data not shown). Because the absence of detectable IL-1 activity in the conditioned media could be due to the presence of an inhibitor of IL-1 activity, we assayed 1 ng of rIL-1 o~or/3 on thymocytes in the presence or absence of conditioned media. N o variation of LAF activity could be detected (data not shown).

Effect of rlFN~ pretreatment on total cellular IL-I synthesis. Recombinant IFN y exerts a time- and dose-dependent inhibitory effect on total cellular IL-1/8 basal production in ASLC in culture (Fig. 6A), with a maximum effect after 48 hr and 100 U/ml. ASLC pretreated 4 days with rIFNy (100 U/ml) showed a decreased steady-state level of total cellular IL-1 fl consistent with the decrease of IL-1 m R N A levels (Fig. 6B). Major histocompatibility complex class II positive ASLC (rlFN'y pretreated) incubated for 24 hr with different doses of rlL-1 0t or/3 showed an inhibition of the positive autoregulation of total cellular IL-1/3 (Fig. 6B) and total cellular IL-1 a (Fig. 6C) compared with M H C class II negative ASLC (non-pretreated with rIFN~,). N o IL-1/8 or ~ was detected in supernatants. The inhibitory effect of the rIFN'y pretreatment on rlL-l-increased IL-I/3 synthesis was studied with different cultures of ASLC andconsistently showed a significant decrease of IL-1 /3 synthesis (from 15% to 60% inhibition). The time course production of total cellular IL-1/3 increased by rIL-1 ~ (10 U/ml) showed a maximum at 12 hr as did M H C class II negative ASLC (data not shown). Major histocompatibility complex class I! positive ASLC treated 24 hr with rIL-1 ~xin the presence ofindomethacin showed an increase of total cellular IL-1/3 synthesis compared with rlL-1 alone (Fig. 6D). These results con-

N. Temime et al.

267

A lOOT

o

50

~.

25

30-

~

20-

.. ... o.° ° °°.~~

r,

F~ 2

24

48

TIME

r7 72

,o 10'

T _1

96

-'°"°°

o

~'--~Without rIFNg

I 0

rIL-i °

~

~

I-t 25

10

~--~Nith rIFNg 3 0 - "B

£

_. . . . . . . . 0

(hours)

(ng/ml)

. ..° °...0 ...........................

20" . 0 .......................... ® °. ° .."'"

r,

10.---- 4- . . . . . . . . . . . . . . . . -+---q

0

"/ o

-----i

..........................

10

25 rIL-I

o

10

25 rlL-1

50

i alDha

(U/ml]

50 (U/ml)

FIGURE 6 Modulation of total cellular 1L-1 synthesis with rIFN7 in ASLC in culture. (A) Confluent ASLC were incubated for different time periods in the presence or absence of rIFN7 (100 U/ml). Total cellular extract was immunoassayed for IL-I 8. (B, C) Non-pretreated ( - - ) and rIFN7-pretreated ASLC (. . . . ) were prepared according to Figure 6; ASLC were then incubated 24 hr with different doses of rIL-l a (O) or/3 (I-1). Total cellular extracts were immunoassayed for IL-1 ~8 (B) or IL-1 /3 (C). (D) Recombinant IFNT-pretreated cells were incubated 24 hr with rlL-1 ~ and/or indomethacin; total cellular extract were assayed for IL-I/3.

firm the previous da:a obtained in Northern blot analysis (Fig. 5). DISCUSSION Interleukin 1 may play a key role during pathologic processes involving the synovial membrane. Adherent

---0--- W i t h o u t inao ....O .... W i t h inao [I

pg/ml)

synovial lining cells cultured in vitro respond to IL-1 by proliferation [31] and increase of prostaglandin and collagenase synthesis [11]. They not only respond to IL-I but also spontaneously produce this cytokine [13]. Dentritic-like cells, abundant in RA synovium and synovial fluid, are potent producers of IL-1 in response to mitogens [32]. Goto et al. [ 13] showed that cloned dendritic-like cells could release IL-I activity, as did fibroblast-like cells and macrophage-like cells, but to a lesser extent [13]. The cellular composition of synovial cell cultures has been debated at length. W e consider that a more accurate representation of the physiology of the tissue is obtained by using whole ASLC cultures instead of separating individual cell constituents of the whole ASLC cultures. Dalton et al. [33] showed that IL-t /8 could induce IL-1 ~x and ~ gene expression in human synovial cells using slot-blot analysis. We extend these results using Northern blot analysis and rlL-1 ~. The response o f ASLC to rIL-I ~ was rapid and transient,

268

with a peak level of IL-1 mKNA occurring at 4 hr (rIL-1 /3 induced a peak level of IL-1 mRNA at 24 hr). Interleukin 1 synthesis induced by rII.,1 was inves:igated by using specific ELISA. The increase of IL-I mRNA resulted in the stockage of IL-1 c~ and/3 within the cell with no release of soluble IL-1 a or/3 detectable by ELISA. No biologically active IL-1 could be detected in the supernatants as determined by thymocyte costimulation assay. These results correlate with the findings of Dalton et al. [33]. No inhibitor of the IL-1 activity in the supernatants of conditioned cultures was detected; thus, the newly synthesized IL-1 may require an additional signal to be released. In any event, IL-1 may be synthesized as proIL-1, then processed as membrane form and thus be biologically significant in the pathogenesis of RA. Therefore, ASLC that are in contact with chondrocytes and loaded with membrane IL-1 may constitute a signal for chondrocytes to secrete neutral proteases that will degrade the extracellular matrix of the joints of patients with RA. To determine the exact nature of the newly synthesized IL-1, Western blot analysis of the proteins are under current investigation. Interferon gamma, a lymphokine present at sites of inflammation, has been reported to modulate IL-1 production: IFNy enhances IL-1 production in monocytes and endothelial cells stimulated with lipopolysaccharides [34, 35] but specifically down-regulates the positive autoregulation of IL-1 in PBMC when added concomitantly with rIL-1 [36]. We found an inhibitory effect of a pretreatment with rIFN'y on rIL-l-induced total cellular IL-1 production by ASLC. Furthermore, rIFN7 pretreatment exerts its effect at the transcriptional level by inhibiting the rIL-l-induced IL-1 mRNAs. Recombinant IFN7 has been shown to induce the expression of MHC class II antigens in ASLC [3]. It remains to be determined whether MHC class II molecules themselves play a role in this down-regulation. Down-regulation of IL-l-induced IL-1 synthesis by rIFN-y in ASLC may explain its beneficial effects in experimental models of IL-l-induced bone and cartilage degradation in patients suffering from RA. Recombinant IFN7 used in the treatment of RA may be considered to have contradictory effects on disease progress because rIFN7 down-regulates IL-1 synthesis but upregulates MHC class II antigen expression, which can favor ASLC to behave as antigen-presentingcells. Previous reports indicate that one major pathway of inhibition of IL-1 synthesis involves arachidonic acid metabolites, particularly PGE2 and PGI2 [37, 38]. In monocytes, IL-1 synthesis induced by bacteriotoxins stimulates PGE2 synthesis, which is suggested to provide a negative regulatory feedback at the posttranscriptional level of IL-1 synthesis via cyclic adenosine monophosphate [38]. In addition, blockage of pros-

rlFN7 Down Regulates IL-I Synthesis

taglandin synthesis in PBMC by indomethacin results in an enhanced expression of IL-1 induced by itself [18]. We report here that blockage of prostaglandin synthesis in MHC class II negative ASLC results in a reduction of total cellular IL-1/3 induced by rIL-1 with no effect at the transcriptional level. The effect of indomethacin should be further analyzed at the transcriptional level using dot-blot analysis, which allows better quantification of a specific mRNA. We have previously found that indomethacin induces a reduction of II,-1/3 gene expression in human dermal fibroblasts [22], assuming that PGE2 may not act only at the posttranscriptional step. Ghezzi and Dinarello [36] documented that the inhibitory effect of rIFN% used concomittantlywith ILl, was not reversed by indomethacin in human PBMC stimulated with rlL-1. In ASLC pretreated with rIFN7, we report that indomethacin enhances the positive regulation of IL-1 synthesis. This effect is exerted at the transcriptional level and is not sufficient to totally reverse the IFN7 inhibition. These data taken together suggest that (1) IFN7 interrupts an amplification loop of IL-1 synthesis, which may occur at the level of IL-1 receptor; (2) IL-1 expression is submitted to complex regulation pathways that might be different in MHC class II negative and positive ASLC; (3) the mechanisms controlling IL-1 expression in class II negative ASLC are different from those found in monocytes but identical to those operating in dermal fibroblasts. It remains to be determined whether the positive regulation of IL-1 synthesis by rlL-1 is truly an autocrine or an associated paracrine stimulation because ASLC in culture are not a homogeneous cell population. We are currently addressing this question by in situ hybridization. Inasmuch as both IL-1 and IFN~ are key cytokines in RA, the mechanisms described in this paper might constitute important regulation pathways in inflammatory reactions. For instance, because rIFN'y reduces the inflammation in patients with RA [7] and reduces the IL-l-induced bone resorption [5, 6], the inhibitory effect of rIFN~/on the autocrine stimulation of IL-1 in ASLC could play a beneficial role in patients with RA.

ACKNOWLEDGMENTS The authors thank Dr. J. P. Pujol and Dr. Condamine for kindly providing ASLC, M. Daneaux for PGE2 assays, and C. le Contel for the LAF costimulationassays. We thank J. Bertoglio for helpful discussionsand critical reviews of the manuscript and M. Brandel for secretarial assistance. This work was supported by grants to D. Charron from INSERM, LNFCC, and ARC. N. Temime is a recipient of an IFSBMARC fellowship.

N. Temime et al.

REFERENCES 1. Klareskog, L, Forsum U, Malmmas-Tjerhund U., Kabeiitz D., Wigren A: Appearance of anti HLA-DR reactive cells in normal and rheumatoid synovial tissue. Stand J lmmunol 14:183, 1981. 2. Malone DG, CatteU H, Tsokos M, Decker JL, Wilder R: Immunohistelogic heterogeneity of synovium from patients with clinically active rheumatoid arthritis. Arth Rheum 26:517, 1983. 3. Amento EP, Bhan AK, McCullagh KG, Krane SM: Influences of gamma interferon on synovial fibroblastlike cells: la induction and inhibition of collagen synthesis. J Clin Invest 76:837, 1985. 4. Browning JL: Interferons and rheumatoid arthritis: Insight into interferon biology? Immunol Today 8:372, 1987. 5. Gowen M, Mundy GR: Actions of recombinant interleukin 1, interleukin 2 and interferon gamma on bone resorption in vitro. J Immunol 136:2478, 1986. 6. Takahashi N, Mundy GR, Roodman GD: Recombinant human interferon gamma inhibits formation of human osteoclast-like cells. J Immunol 137:3544, 1986. 7. Seitz M, Maria G, Frank M: Use of recombinant human gamma interferon in patients with rheumatoid arthritis. Rheumatology 45:93, 1986. 8. Dinarello CA, Cannon JG, Mier jW, Bernheim HA, Lopreste G, Lynn DL, Love RN, Webb AC, Auron PE, Reuben RC, Rich A, Wolff SM, Putney SD: Multiple biological activities of human recombinant interleukin 1. J Clin Invest 77:1734, 1986. 9. McGuire-Goldring MB, Meats JE, Wood DD, Ihric EJ, Ebsworth NH, Russel ROG: In vitro activation of human chnndrocytes and synoviocytes by a human interleukin 1 like factor. Arth Rheum 27:654, 1984. 10. Wood DD, Ihric EJ, Dinarello CA, Cohen PL Isolation of an interleukin 1 like factor from human joint effusions. Arth Rheum 26:975, 1983. 11. Dayer JM, Rochemonteix B, Burrus B, Demczuk S, Dinarello CA: Human recombinant interleukin 1 stimulates collagenase and prostaglandin E2 production by human synovial cell. J Clin Invest 77:645, 1986. 12. Deshmukh-Phadke K, Nanda S, Lee K: Macrophages factor that induces neutral proteases secretion by no, real rabbit chondrocytes. EurJ Biochem 104:175, 1980. 13. Goto M, Sasano M, Yamanaka H, Miyasaka N, Kamatani N, Inoue K, Nishioka K, Miyamoto T: Spontaneous production of an interleukin l-like factor by cloned rheumatoid synovial cells in long term culture. J Clin Invest 80:786, 1987. 14. Auron PE, Webb AC, Rosenwasser LJ, Mucci SF, Rich A, Wolff SM, Dinarello CA: Nucleotide sequence of human monocyte interleukin 1 precursor cDNA. Proc Narl Acad Sci USA 81:7907, 1984. 15. Lomedico PT, Gubler U, Heilman CP, Dukovich M, Girl

269

JG, Pan YCF, Collier K, Semionov R, Chua AO, Mizel SB: Cloning and expression of routine interleukin I cDNA in E. coil Nature 312:458, 1984. 16. Dinarello CA: An update on human interleukin h From molecular biology to clinical relevance. J Clin Immunol 5:287, 1985. 17. March CJ, Mosley B, Lersen A, Cerreti DP, Br~dt G, Price V, Gillis S, Henney CS, Kronheim SR, Grabstein IL Conlon PJ, Hopp TP, Cosman D: Cloning, sequence and expression of two distinct human interleukin 1 complementary DNAs. Nature 315:641, 1985. 18. Dinarello CA, Takashi I, Warner SIC, Orencole 81:, LOn. nemann G, Cannon JG, Libby P: Interleukin 1 induces interleukin 1, I: Induction of circulating interleukin 1 in rabbits in vivo and in human mononucleat cells in vitro. J Immunol 139:1902, 1987. 19. Warner SIC, Auger KR, Libby P: Interleukin 1 induces interleukln 1, II: Recombinant human interleukin 1 induces interleukin 1 production by adult human vascular endothelial cells. J Inununo1139:19t 1, 1987. 20. Warner SIC, Auger KR, Libby P: Human interleukin 1 induces interleukin 1 gene expression in human vascular smooth muscle cells. J Exp Med 165:1316, 1987. 21. Sakm K, Hattori T, Matsuoka M, Asou N, Yamamoto S, Sagawa K, Takatsuki K: Autocrine stimulation of interleukin 1 /3 in acute myelngenous leukemia cells. J Exp Med 166:1597, 1987. 22. Manviel A, Temime N, Chacron D, Loyau G, Pujol JP: Interleukin 1 c, and ~ induce interleukin 1 ~0gene expression in human dermal fibroblasts. Biochem Biophys Res Commun 156:1209, 1988. 23. Teyton L, Lottean V, Turmel P, Arenzana-Selsdedos F, Virelizier JL, Pujol JP, Loyan G, Piater-Tonneau D, Auffray C, Chart-on D: HLA-DR, DQ and DP ant/gen expression in rheumatoid synovial cells: A biochemical and quantitative study. J Immuno1138:1730, 1987. 24. Fontaine S, Damais C, Lando D, Cousin MA: Monoclonal antibodies to human interleukin 1 ~ and their use in a sensitive two-site enzyme linked immunosorbent assay. Lympholdne Res 8:129, 1989. 25. Chirgwin JM, Prysbyla AE, McDonald RJ, Putter WJ: Isolation of biologically active ribor.wleic acid from sources enriched in ribonuclease. Biochem 18:5294, i979. 26. Glisln, Crkvenjakov VR, Byus C: Ribonucle/c acid isolation by cesium chloride centrifugation. Biochem 13:2633, 1974. 27. Feinberg AP, Vogelstein B: A technique for radiolabelb ing DNA restriction endonuclease fragments to high specific activity. Anal Biochem 132:6, 1983. 28. Fort P, Marry L, Piacheczyk M, El Sabrouty S, Dani C, Jeanteur P, Blanchard JM: 1985. Various rat adult tissues express only one major mRNA specie from the glyceraldehyde-3-phosphate-deshydrngenase multigenic family. Nucleic Acid Res 13:1431, 1985.

270

29. Rosenwasser LJ, Dinarello CA: Ability of human leukocytic pyrogen to enhance phytohemagglutinin induced murine thymocyte proliferation. Cell lmmunol 63:134, 1981. 30. Dray F, Charbonnel B, MacloufJ: Radioimmunnoassay of prostaglandin F2c~, E1 and E2 in human plasma. Eur J Clin Invest 5:311, 1975. 31. Butler DM, Piccoli DS, Hart PH, Hamilton JA. Stimulation of human synovial fibroblast DNA synthesis by recombinant cytokines. J Rheumatol 15:1463, 1988. 32. Duff GW, Forre O., Waalen K, Dickens E, Naki G: Rheumatoid arthritis synovial dentritic cells produce interleukin 1. BrJ Rheum 24 (Suppl 1):94, 1985. 33. Dalton BJ, Connor JR, Johnson WJ: Interleukin 1 induces interleukin 1 ~xand interleukin 1 3 gene expression in synovial fibroblasts and peripheral blood monocytes. Arthritis Rheum 32:279, 1989. 34. Arenzana.Seisdedos F, VirelizierJL, Fiers W: Interferons

rIFN7 Down Regulates IL-I Synthesis

as macrophage activating factors III. Preferential effects of interferon gamma on the interleukine 1 secretory potential of fresh or aged human monocytes. J lmmunol 134:2444, 1985. 35. Miossec P, ZiffM: Immune interferon enhances the production of interleukin 1 by human endothelial cells stimulated with lipopolysaccharide. J Immunol 137:2848, 1986. 36. Ghezzi P, Dinarello CA: IL-1 induces IL-I III. Specific inhibition of IL-I production by IFN gamma. J Immunol 140:4238, 1988. 37. Kunkel SL, Chensue SW, Phan SH: Prostaglandins as endogenous mediators of IL-1 production. J Immunol 136:186, 1986. 38. Knudsenj PJ, Dinarello CA, Strom TB: Prostaglandins inhibit post-transcriptional IL-1 synthesis by increasing intracellular cyclic adenosine monophosphate. J Immunol 137:3189, 1986.