Immunobiol., vol. 180, pp. 316-327 (1990)

Division of Immunology and Allergy (Hans Wilsdorf Laboratory), Department of Medicine, H6pital cantonal universitaire, Geneve, Switzerland

An Interleukin 1 Inhibitor Affects both Cell-Associated Interleukin I-Induced T Cell Proliferation and PGE2 / Collagenase Production by Human Dermal Fibroblasts and Synovial Cells PHILIPPE SECKINGER, MARIE-THERESE KAUFMANN, and JEAN-MICHEL DAYER Received September 22, 1989 . Accepted in Revised Form November 24, 1989

Abstract Lipopolysaccharide (LPS) induces cell-associated interleukin 1 (IL 1) production in the human promonocytic cell line U937. Demonstration of cell-associated IL 1 activity was based on the ability of LPS-treated U937 cells, subsequently fixed with paraformaldehyde, to stimulate thymocyte proliferation in the presence of phytohemagglutinin. Like soluble IL 1 (sIL 1), cell-associated IL 1 is capable of inducing PGE2 and/or collagenase production by dermal fibroblasts and human synovial cells in a dose-dependent manner. It is thus a mediator of the inflammatory response owing to a direct intercellular contact located at the membrane level, where bound molecules may trigger inflammation at a local site of action. We reported that the natural (~ 23 kDa) IL 1 inhibitor (IL 1 INH) from the urine of febrile patients inhibited all the sIL-l-induced biologic activities under investigation and that it acted by binding to the IL 1 receptor, thus blocking the interaction of the monokine with the receptor. Data demonstrate that the IL 1 INH also blocks cell-associated IL I-induced T cell proliferation and PGE2 production by both dermal fibroblasts and synovial cells as well as collagenase production by the latter cell type. Thus, as for the sIL 1, a feedback mechartism exists for cellassociated IL I-induced bioactivities.

Introduction In inflammatory diseases, infiltration of several types of circulating cells occurs, and thus, intercellular interactions become a potential phenomenon by which the inflammatory response is elicited. Therefore, several of our investigations have focussed on the question whether cell-associated molecules such as interleukin 1 (IL 1) mediate the inflammatory response via direct cell-cell interactions. IL 1 has pleiotropic biologic activities and plays a major role in immune and inflammatory response (1), as it induces collagenase and PGE 2 production by human dermal fibroblasts and synAbbreviations: LPS = lipopolysaccharide; slL-l = soluble interleukin 1; ILl INH = interleukin 1 inhibitor; LAF = lymphocyte-activating factor; MCF = mononuclear cell factor.

ILl Inhibitor and Cell-Associated ILl . 317

ovial cells (2-5), two factors involved in tissue destruction (5). The molecu:" lar sequencing has revealed that at least two non-allelic genes encode for IL 1, both synthesized as intracellular precursors with a molecular weight of 31 kDa. Both gene end products, IL la (pI ~5) and IL lB (pI ~7), share the same receptor and similar biologic activities, despite only 26 % homology in their amino-acid sequence (6-9). Originally described as a soluble protein of 17 kDa, IL 1 has recently been reported as a cell-associated form in different cells including macrophages, monocytes, B lymphocytes, endothelial cells and dermal fibroblasts (10-19). It was shown to induce, like soluble IL 1 (sIL 1), T cell proliferation and later on PGE z production by human chondrosarcoma cells (18). We have previously observed that urine from monocytic leukemia and febrile patients contained an inhibitor of IL 1 activity (IL 1 INH) with a molecular weight of 18-25 kDa as estimated by gel filtration (20). The IL 1 INH is effective against IL 1 a and B, as assessed in the lymphocyteactivating factor assay (IL l/LAF), the mononuclear cell factor assay (IL 1/ MCF) measuring PGE z and/or collagenase production by synovial cells and dermal fibroblasts, and the fibroblasts proliferation assay, whereas tumor necrosis factor-a (TNF-a) activity, shared with IL 1, was not affected by the present IL 1 INH (21). Subsequently, we were able to demonstrate that the IL 1 INH acts by binding to the IL 1 receptor, thus blocking the interaction of IL 1 with its receptor (22). The aim of the present investigation was to extend studies on the biologic activities of cell-associated IL 1 and to search for the presence of a feedback mechanism at this level.

Materials and Methods Reagents and Media

Dulbecco's modified Eagle's minimal essential medium (DMEM), RPMI 1640, phosphatebuffered saline (PBS), fetal calf serum (FCS), penicillin, streptomycin and glutamine were obtained from GIBCO (Paisley, Scotland). Phytohemagglutinin (PHA) was purchased from Wellcome Research Laboratories (Beckenham, Kent, U.K.), and lipopolysaccharide (LPS) (E. coli strains 055:B5) from Difco Laboratories (Detroit, MI, U.S.A.). A rabbit polyclonal antibody to recombinant human IL la (rh IL la) was provided by Biogen SA, Geneva, Switzerland. Partial Purification of IL 1 INH

Partial purification of IL 1 INH was carried out as described previously (23). Briefly, concentrated urine was ammonium sulfate-precipitated by 40-80 % of saturation, the pellet chromatographed on DEAE Sephadex, hydroxylapatite and finally gel-filtrated using an Ultrogel AcA54 column (2.6 x 100 cm; LKB, Bromma, Sweden). Fractions eluting between 18 and 25 kDa were used for the biologic studies. Cell Lines

The human monocytic tumor cell line U937, free of mycoplasma, was maintained in RPMI 1640 medium supplemented with 100 J-lg/ml streptomycin, 100 U/ml penicillin, 1 % glutamine

318 . P. SECKINGER, M.-T. KAUFMANN, and J.-M. DAYER and 10 % heat-inactivated FCS. Human synovial cells were isolated as described previously (2) from surgical synovectomy specimens, obtained from patients with rheumatoid arthritis or osteoarthritis. Human foreskin fibroblasts were similarly isolated from tissue specimens by protease treatment (4). The human synovial cells and dermal fibroblasts were maintained by passage in DMEM supplemented with 100 [!g/ml penicillin, 100 U/ml streptomycin, 1 % glutamine and 10 % non-heat-inactivated FCS.

Induction of Cell-Associated IL 1 Induction of cell-associated IL 1 was carried out as described previously (15, 17). Briefly, U937 cells (2.5 x 105/ml) were incubated for 24 h in RPMI 1640-complemented culture medium at 37°C with 10 [!g/ml of LPS, then washed by centrifugation in culture medium and treated with 1 % paraformaldehyde (w/v) at room temperature for 15 min. Excess of paraformaldehyde was removed by centrifugation, paraformaldehyde-fixed cells were washed three times in PBS and suspended at the required concentration in the appropriate culture medium. LPS-unstimulated paraformaldehyde-fixed U937 cells were utilized as control.

Cell-Associated IL llLAF Assay The IL 1 activity of paraformaldehyde-fixed cells was assayed by the C 3H/HeJ mouse thymocyte proliferation assay (21). Briefly, the thymocytes (1.5 x 106 /well, plated in 96~well V-bottom plates; Costar, Cambridge, MA, U.S.A.) were co-stimulated for 72 h with PHA (1 [!g/ml) in the presence of various concentrations of either stimulated or control unstimulated U937. Partially purified IL 1 INH from the AcA54 inhibitory fractions was tested for inhibitory properties at three different final dilutions on cells: 1:20, 1:40, and 1: 80. LPS-treated U937 and control U937 were preincubated with the rabbit polyclonal antiserum to rhIL 1a for 3 h at 37°C in the presence of FCS at 1:500 dilution. Thymocytes were added after this preincubation time in order that final dilution of the antiserum was two times higher than values indicated. In each case, the cells were cultured in triplicate for 3 days at 37°C in 5 % CO z/95 % air. Six h before end of culture, cells were pulsed with [3H]-TdR (0.5 [!Ci/well), harvested onto glass fiber filters and uptake of eH]-T dR was determined by using a scintillation counter.

Cell-Associated IL lIMCF Assay The IL 1 activity of paraformaldehyde-fixed cells was assayed on human synovial cells and dermal fibroblasts. Two days before bioassays, the synovial cells or fibroblasts were plated in 96-well flat bottom plates at a density of 2 x 104 cells/well. Cells were then stimulated with various concentrations of either stimulated or control U937 using DMEMsupplemented with 10 % FCS. PGE2 and collagenase production was measured after 72 hand 96 h, respectively, in culture supernatants. The IL 1 INH was tested with the appropriate controls for inhibitory properties at three different final concentrations (1 :20, 1:40, and 1:80). PGE2 production was measured by a double antibody radioimmunoassay by using an antiserum to PGEz, kindly provided by Dr. L. LEVINE (Brandeis University, Waltham, MA, U.S.A.) (2). Collagenase activity after 96h of culture was measured according to the method of JOHNSON-WINT using tritium-reconstituted collagen fibrils (2). One unit is defined as the release of 1 ltg/min of labeled collagen at 37°C.

Results

Regulation of Cell-Associated IL I-Induced T Cell ProlIferation by the ILIINH

IL 1 activity of cell-associated IL 1 of paraformaldehyde-fixed U937, activated with 10 [!g/ml LPS for 24 h and measured by the thymocyte

III Inhibitor and Cell-Associated III . 319

40T--------------------------------,

c7 0

30

"

i

i=

20

.S

~

':!...

10

0.05

0.10

0.25

0.50

0.75

1.00

ParaformaldOhvdO fixed U 937 added (x 10-5)

Figure 1. Murine thymocyte proliferation (1.5 x 106 cells/well) induced by various concentrations of LPS-stimulated paraformaldehyde-fixed U937 in the absence 0 or presence of various dilutions: 1:20 (.), 1:40 ~) , and 1:80 (JIEl) of IL 1 INH. Values represent means ± SEM (N=3) of [3H]-TdR incorporation in thymocytes in the presence of PHA (IL l/LAF).

proliferation assay, is illustrated in Figure 1. A dose-response of eH]-TdR incorporation was observed up to concentrations of 105 U937/welL Contrary to LPS-stimulated cells, the control U937 did not induce thymocyte proliferation, and in all the assays performed in this study, control unstimulated U937 failed to induce biologic activity including PGE 2 and/or collagenase production by dermal fibroblasts and human synovial cells (data not shown). Studying the activity of an IL 1 INH in the same IL lILAF assay, complete inhibition of co-stimulated thymocyte proliferation was observed in the presence of IL 1 INH at various dilutions. Addition of IL 1 INH to control U937 did not affect eH]-TdR incorporation, indicating that the inhibition that was observed after adding LPS-stimulated U937 to the thymocytes could not be due to a cytotoxic or non-specific effect but rather to the inhibition of cell-associated IL 1 activity. In all the assays performed throughout this study, LPS-unstimulated U937 in the presence of IL 1 INH at various dilutions failed to induce biologic activity, including PGE 2 and/or collagenase production by dermal fibroblasts and human synovial cells (data not shown). To additionally document the relationship

320 . P. SECKINGER, M.-T. KAUFMANN, and J.-M. DAYER

of cell-associated IL 1 activities and the present IL 1 INH, LPS-stimulated U937 and control cells were preincubated at 3rC with a polyclonal rabbit antiserum raised against rhIL 1a. This antibody neutralizes only biologic activities of rhIL 1a and not those induced by rhIL 1~ when tested in the IL 1/MCF assay at 1:200 final dilution. Thus, 250 pg/ml of either rhIL 1aor ~-stimulated dermal fibroblast PGE z production from 13.7 ± 4.1 to 148.2 ± 8.3 and 116.3 ± 11.8 ng/ml, respectively (mean values ± SEM). When medium, rhIL 1a or rhIL 1~ was preincubated with the anti-rhIL la, PGE z production amounted to 15.3 ± 2.9, 12.5 ± 0.9 and 163 .1 ± 12.5 ngl ml, respectively (mean values ± SEM). Thymocyte proliferation was inhibited by adding polyclonal anti-rhIL 1a serum, as shown in Figure 2. Diluting the antibody and increasing the number of U937 prevented in part the inhibition induced by the polyclonal antibody. Addition of antibody to unstimulated cells did not affect the PHA-induced stimulation (data not shown). The 72-h culture supernatants, harvested and tested in the IL 11 LAF assay of both stimulated and control U937, revealed no IL 1 activity. When supernatants from 105 LPS-stimulated U937 were tested in the' presence of PHA, thymidine incorporation of PHA alone (515 ± 135 cpm) 40

M

1

30

..

b

l

I

20

.S

¥

'2..

10

o

~

~

1m

0.10

0.25

0.50

I

0.75

1.00

Paraformaldehyde fixed U 937 added (x 10-5)

Figure 2. Murine thymocyte proliferation (1.5 x lOb cells/well) induced by various concentrations of LPS-stimulated paraformaldehyde-fixed U937 in the ab sence 0 or presence of a 1:1,000 dilution of rabbit anti-human recombinant IL 1a serum (~ . Values represent means ± SEM (N=3) of eHJ-TdR incorporation in thymocytes in the presence of PHA (IL l/LAF).

III Inhibitor and Cell-Associated III . 321

remained unaffected (560 ± 215) by the addition of 1:4 dilution of this supernatant. In contrast, addition of 100 pg/ml of rhIL-1 ~ increased thymidine incorporation to 2,785 ± 455 cpm. Regulation of Cell-Associated IL I-Induced PGE2 and/or Collagenase Production by an IL 1 INH

In order to extend investigations of its biologic activities, cell-associated IL 1 was tested on human synovial cells and dermal fibroblasts. Paraformaldehyde-fixed U937 previously activated with 10 f.tg/ml LPS induced a dose-dependent PGE z production on both dermal fibroblast (Fig. 3) and human synovial cells (Fig. 4). Human synovial cells are more sensitive to cell-associated IL 1, since as few as 5 x 104 U937/well sufficed to obtain a maximal effect on PGE z production, whereas the same concentration of U937 on dermal fibroblasts resulted in only ~25 % of maximal PGE z production. IL 1 INH activity was then studied in the present IL I/MCF assay, and inhibition of PGE z production by both human dermal fibroblasts and ~r------------------------------'

c-

o )(

15

~

.5. J!l

If

...~ ~

10

G>

~

];

.~

I

w'"

5

(!) Il.

0.50

1.00

2.50

5.00

Paraformaldehyde fixed U 937 added (x 10"5)

Figure 3. PGE z production by dermal fibroblasts induced by varipus concentrations of LPSstimulated paraformaldehyde-fixed U937 in the absence (0 or prbence of various dilutions: 1:20 (.), 1:40 ~) , and 1:80 (~) ofIL 1 INH. Values represent means ± SEM. (N=3) ofpGEz production after 3 days of culture (IL lIMCF).

322 . P . SECKINGER, M.-T.

KAUFMANN,

and J.-M.

DAYER

synovial cells was almost complete in the presence of IL 1 INH. This inhibitory property was overcome when adding higher concentrations of U937 to either cell type, but was less pronounced on dermal fibroblasts (Fig. 3) than on synovial cells (Fig. 4). Thus, the inhibitory activity of the IL 1 INH involves a competitive mechanism of action. Addition of IL 1 INH alone, without U937, to human synovial cells led to a slight but not significant increase of PGE z basal level (data not shown). Similar results were obtained in three different experiments carried out as mentioned above on human synovial cells and dermal fibroblasts. It is known that sIL 1 induces collagenase production; therefore, we wondered whether cell-associated IL 1 had similar biologic activities. We were unable to induce collagenase production by dermal fibroblasts, at least in the strain investigated. In contrast, human synovial cells produce trypsinactivatable collagenase, and this production is dependent on the amount of U937 added to cells, as seen in Figure 5. Maximal collagenase production occurred by 5 x 10 4 LPS-stimulated U937/well, corresponding to the concentration used to obtain a maximal PGE z production by the same target cells. IL 1 INH activity was then studied in the present IL lIMCF ~~-------------------------------,

-..

'1'

0

I

15

~

I

~

~ '"

10

0-

s

.t:

~

i N

W

a?

0.05

0.10

o.so

1.00

pararormaldehyde fixed U 937 added

5.00

Ix

10-5)

Figure 4. PGE2 production by human synovial cells induced by various concentrations of LPSstimulated parafonnaldehyde-fixed U937 in the absence 0 or presence of various dilutions : 1:20 (.), 1:40 ~) , and 1:80 (!!iii) of IL 1 INH. Values represent means ± SEM (N=3) of PGE2 production after 3 days of culture (IL I/MCF).

III Inhibitor and Cell-Associated III . 323

1.0

i i 1

0.8

~

~:>

0.6

.J:.

1;

~

0.4

Q.

~

II>

j

~

0.2

0.0 .............- -- -....

0.10

0.50

1.00

5.00

Paraformaldehyde fixed U 937 added (x 10-5)

Figure 5. Trypsin-activated collagenase production by human synovial cells induced by various concentrations of LPS-stimulated paraformaldehyde-fixed U937 in the absence 0 or presence of various dilutions : 120 : (.), 1 :40(ffa), and 1 :80 (~) of IL 1 INH. Trypsin-activated collagenase production was measured in the culture supernatant after 4 days of culture. Values represent means ± SEM (N=3) (IL l /MCF).

assay measuring collagenase production by human synovial cells. Addition of both LPS-stimulated U937 and IL 1 INH resulted in a partial inhibition of collagenase production by human synovial cells. Similar results were obtained on other target cells. As was the case with the PGE z production by the same target cells, addition of increasing concentrations of LPS-stimulated U937 overcame the inhibitory activity of the IL 1 INH. Discussion Cell surface-associated IL 1 has been reported to be elicited by different stimuli, e.g. LPS, TNF-a, or lymphotoxin (10-19). Not only monocytemacrophages have been shown to produce cell-associated IL 1 that until now has been described as antigenically similar to sIL 1, but so have B cells, human endothelial cells or fibroblasts. Studies on the biologic activities of cell-associated IL 1 have especially focussed on its ability to induce T cell

324 . P.

SECKINGER,

M.-T.

KAUFMANN,

andJ.-M.

DAYER

proliferation in the presence of lectins or antigens, and ~nly MERLUZZI et al. (18) have reported that ceIl-associated IL 1 is capable of stimulating PGE2 production by human chondrosarcoma ceIls. Despite extensive investigation, the existence of ceIl-associated IL 1 remains controversial, since cDNA analysis clearly showed the absence of hydrophobic regions, necessary for membrane insertion (24). In contrast, TNF~a has been reported as a ceIl surface transmembrane protein produced at the local level of inflammation (25). With the aim of further investigating the biologic activities of cell-associated IL 1, we were able to induce PGE2 andlor collagenase production by dermal fibroblasts and human synovial cells. In the context of inflammatory diseases, infiltration of stimulated monocyte macrophages might be responsible for tissue damage. In addition, it has recently been shown that IL 1 synthesis and secretion are two separate processes, since IL 1 synthesis occurs within the monocytes after low density LPS stimvlation, whereas high-density is necessary to induce both synthesis and release of IL 1, suggesting a distinct mechanism of regulation of synthesis and secretion (26). Therefore, infiltration of IL 1 stimulated macrophages may be instrumental to coIlagen breakdown by inducing proteolytic enzyme synthesis without releasing sIL 1. The discrepancy between the results of collagenase production obtained with dermal fibroblasts and synovial cells can be explained by the strain of dermal fibroblasts used which did not produce any collagenase, as far as could be detected with the sensitiy of our assay. However, when tested in an earlier passage the same strain of fibroblasts, cultured in the presence of either medium or rhIL-1 ~ at 250 and 500 pglml, produced 0, 0.13, and 0.35 U/ml, respectively, of trypsinactivatable collagenase. This does not exclude the possibility that other fibroblasts also respond to cell-associated IL 1 by producing detectable collagenase. It should be noted that the PGE2 production induced by cell-associated IL 1 was rather elevated and that we are unable to obtain such concentrations of PGE2 with sIL 1, even at high concentration. Thus, comparison of cell-associated IL 1 with sIL 1 is only based on biologic activity and cannot be considered as an indication of the relative concentration of cell-associated IL 1 versus sIL 1. Moreover, these two different molecules may also have different specific activities (16). We previously isolated an IL 1 INH in the urine of febrile patients, specific for IL 1, since it does not affect TNF-a (20, 21). We were able to show that this inhibitor interferes with the binding of 125I_IL la to its receptor when tested on the murine thymoma subline EL4-6.1 (22). All the sIL I-induced biological activities that we have studied up to now are blocked by the IL-1 INH of ~23 kD (27). We now report that just like in the case of sIL 1, a control mechanism exists for cell-associated IL 1-induced biologic activities. Recently ROSOFF et al. demonstrated that IL 1 stimulates diacylglycerol production on T lymphocytes that bear IL 1 receptors (such as EL 4 murine thymoma cell

ILlInhibitor and Cell-Associated III . 325 line) or not (such as Jurkat). The authors suggest that at least two separate transduction mechanisms exist by which IL 1 signals are transmitted to the nucleus: the first utilizes the It 1 receptor, and the second involves phosphatidylcholine production (28). The fact that the IL 1 INH, which acts by binding to the receptor, is capable of blocking cell-associated IL 1 activity on thymocytes, dermal fibroblasts and human synovial cells, shows that the IL 1 receptor pathway is involved rather than the phosphatidylcholine pathway. The use of such an inhibitor will lead to new insight into how IL 1 transmits its signal inside the cell. The biologic activities induced by cell-associated IL 1 were also inhibited in the IL 1/LAF assay by adding an antibody to rhlL 1a and not by antibody to rhlL 1~. We therefore confirm previous observations that in the monocytic cell line U937 cell-associated IL 1 is almost related antigenically to slL la, since most - although not all- of the thymocyte proliferation was inhibited by adding only anti-rhlL 1a antibody preincubated with the cells in the presence of FCS, ruling out a proteolytic activity of the antiserum (15,16). More recently, SINGER et al. reported that IL 1~ is not anchored to the plasma membrane but localized in the cytosolic substance (29). A major aspect of our data is the in vitro inhibition of the bioactivity of one mediator of the inflammatory process. The availability of such an inhibitor opens the perspective of controlling connective tissue destruction at the local level through the inhibition of cell-associated IL 1 (30). In contrast to PGEz production by human synovial cells, collagenase production could not be inhibited to the same extent by adding IL 1 INH to the cells, and various degrees in the inhibitory capacity of an identical final dilution on cells have already been reported for slL 1 (22). Comparison of slL 1 and cell-associated IL 1 on the different functional responses to IL 1, e.g., bohe resorption, proliferation of fibroblasts, acutephase protein synthesis, will clarify whether slL 1 and the cell-associated form have identical bioactivities. Considering that both forms share an inhibitor acting at the receptor level, they may also have the same receptor. Acknowledgments This work was supported partly by grant no. 3.400.0.86 from the Swiss National Science Foundation, by Glaxo 1MB, and by the Foundation «Centre de recherches medicales Carlos et Elsie de Reuter». The authors thank Ms. RACHEL CHICHEPORTICHE for technical assistance.

References 1. OPPENHEIM, J. J., E. J. KOVACS, K. MATSUSHIMA, and S. K. DURUM. 1986. There is more than one interleukin-1. Immunology Today. 7: 45. 2. DAYER, J.-M., J. BREARD, L. CHESS, and S. M. KRANE. 1979. Participation of monocytemacrophages and lymphocytes in the production of a factor that stimulates collagenase and prostaglandin release by rheumatoid synovial cells. J. Clin. Invest. 64: 1386.

326 . P. SECKINGER, M.-T. KAUFMANN, andJ.-M. DAYER 3. MIZEL, S. B., J.-M. DAYER, S. M. KRANE, and S. E. MERGENHAGEN. 1981. Stimulation of rheumatoid synovial cell collagenase and prostaglandin production by partially purified lymphocyte-activating factor (interleukin-l). Proc. Nat!' Acad. Sci. USA. 78: 2474. 4. DAYER, j.-M., C. ZAvADIL-GROB, C. UCLA, and B. MACH. 1984. Induction of human interleukin-l mRNA measured by collagenase and prostaglandin Erstimulating activity in rheumatoid synovial cells. Eur. j. Immuno!. 14: 898. 5. KRANE, S. M., S. R. GOLDRlNG, andJ.-M. DAYER. 1982. Interactions among lymphocytes, monocytes, and other synovial cells in the rheumatoid synovium. Lymphokines 7: 75. 6. TOCCI, M. J., N. I. HUTCHINSON, P. M. CAMERON, K. E. KIRK, D. J. NORMAN, J. CHIN, E. A. Rupp, G. A. LIMJUCO, V. M. BONILLA-ARGUDO, and J. A. SCHMIDT. 1987. Expression in E. coli of fully active recombinant human interleukin I-beta. J. Immuno!. 138: 1109. 7. MARCH, C. J., B. MOSLEY, A. LARSEN, D. P. CERRETTI, G; BRAEDT, V. PRICE, S. GILLIS, C. S. HENNEY, S. R. KRONHEIM, K. GRABSTEIN, P. J. CONLON, T. P. Hopp, and D. COSMAN. 1985. Cloning, sequence, and expression of two distinct human interleukin-l complementary DNAs. Nature 315: 641. 8. MATSUSHIMA, K., T. AKAHOSHI, M. YAMADA, Y. FURUTANI, and J. J. OPPENHEIM. 1986. Properties of a specific interleukin 1 (IL-l) receptor on human Epstein-Barr virustransformed B lymphocytes: identity of the receptor for IL-la and IL-l~. J. Immuno!. 136: 4496. 9. KILIAN, P. L., K. L. KAFFKA, A. S. STERN, D. WOEHLE, W. R. BENJAMIN, T. M. DECHIARA, U. GUBLER, J. j. FARRAR, S. B. MIZEL, and P. T. LOMEDICO. 1986. Interleukin-la and interleukin-l~ bind to the same receptor on T cells. J. Immuno!. 136: 4509. 10. KURT-JONES, E. A., D. I. BELLER, S. B. MIZEL, and E. R. UNANUE. 1985. Identification of a membrane-associated interleukin-l in macrophages. Proc. Nat!. Acad. Sci. USA. 82: 1204. 11. MATSUSHIMA, K., M. TAGUCHI, E. J. KOVACS, H. A. YOUNG, andJ. J. OPPENHEIM. 1986. Intracellular localization of human monocyte associated interleukin-l (IL-l) activity and release of biologically active IL-l from monocytes by trypsin and plasmin. J. Immuno!. 136: 2883. 12. LE, J., D. WEINSTEIN, U. GUBLER, and J. VILCEK. 1987. Induction of membrane-associated interleukin-l by tumor necrosis factor in human fibroblasts. J. Immuno!. 138: 2137. 13. LE, J., and J. VILCEK. 1987. Accessory function of human fibroblasts in mitogenstimulated interferon-y production by T lymphocytes. j. Immuno!. 139: 3330. 14. GERRARD, T. L., and D. Y. VOLKMAN. 1985. IL I-like activity in antigen-presenting human B cell lines. J. Immuno!. 135: 3217. 15. CONLON, P. j., K. H. GRABSTEIN, A. ALPERT, K. S. PRICKETT, T. P. Hopp, and S. GILLIS. 1987. Localization of human mononuclear cell interleukin 1. j. Immuno!. 139: 98. 16. KURT-JONES, E. A. W. FIERS, and J. S. POBER. 1987. Membrane interleukin 1 induction on human endothelial cells and dermal fibroblasts. J. Immuno!. 139: 2317. 17. HURME, M. 1987. Membrane-associated interleukin 1 is required for the activation of T cells in the anti-CD3 antibody-induced T cell response. J. Immuno!. 139: 1168. 18. MERLUZZI, V. J., R. B. FAANES, M.CZAJKOWSKI, K. LAST-BARNEY, P. C. HARRISON, J. KAHN, and R. ROTHLEIN. 1987. Membrane-associated interleukin 1 activity on human U937 tumor cells: stimulation of PGE2 production by human chondrosarcoma cells. J. Immunol. 139: 166. 19. BAKOUCHE, 0., D. C. BROWN, and L. B. LACHMAN. 1987. Subcellular localization of human monocyte interleukin-l evidence for an inactive precursor molecule and a possible mechanism for IL-l release. J. Immunol. 138: 4249. 20. BALAVOINE,j-F., B. DE ROCHEMONTEIX, K. WILLIAMSON, P. SECKINGER, A. CRUCHAUD, a~d J. M. DAYER.' 1986. Prostaglandin E2 and collagenase production by fibroblasts and sYnovial cells is regulated by urine-derived human interleukin 1 and inhibitor(s). J. Clin. Invest. 78: 1120. 21. SECKINGER, P., K. WILLIAMSON, J.-F. BALAVOINE, B. MACH, G. MAZZEI, A. SHAW, and J.-M. DAYER. 1987. A urine inhibitor of interleukin 1 activity affects both interleukin-la and interleukin-l~ but not tumor necrosis factor a. J. Immunol. 139: 1541.

III Inhibitor and Cell-Associated III . 327 22. SECKINGER, P., J. W. LOWENTHAL, K. WILLIAMSON, J.-M. DAYER, and H. R. MACDONALD. 1987. A urine inhibitor of interleukin 1 activity that blocks ligand binding. J. Immuno!. 139: 1546. 23. DAYER, J.-M., B. DE ROCHEMONTEIX, B. BURRUS, S. DEMCZUK, and C. A. DINARELLO. 1986. Human recombinant interleukin 1 stimulates collagenase and prostaglandin E2 production by human synovial cells. J. Clin. Invest. 77: 645. 24. MINNICH-CARRUTH, L. L., J. SUTTLES, and S. B. MIZEL. 1989. Evidence against the existence of a membrane form of murine IL-1u. J. Immuno!. 142: 526. 25. KRIEGLER, M., C. PEREZ, K. DEFAY, I. ALBERT, and S. D. Lu. 1988. A novel form of TNF-cachectin is a cell surface cytotoxic trans-membrane protein: ramifications for the complex physiology of TNF. Cell 53: 45. 26. BAKOUCHE, 0., W. C. KOFF, D. C. BROWN, and L. B. LACHMAN. 1987. Interleukin 1 release by human monocytes treated with liposome-encapsulated lipopolysaccharide. J. Immuno!. 139: 1120. 27. DAYER, J.-M., and P. SECKINGER. 1989. Natural inhibitors and antagonists of interleukin1. In: Interleukin 1 - Inflammation and Diseases. BOMFORD, R. H. R., and B. HENDERSON (Eds.). Amsterdam, Elsevier Science Publishers BV (in press). 28. ROSOFF, P. M., N. SAVAGE, and C. A. DINARELLO. 1988. Interleukin-1 stimulates diacylglycerol production in T lymphocytes by a novel mechanism. Cell 54: 73. 29. SINGER, I. I., S. SCOTT, G. L. HALL, G. LIMJUCO, J. CHIN, and J. A. SCHMIDT. 1988. Interleukin-1~ is localized in the cytoplasmic ground substance but is largely absent from the Golgi apparatus and plasma membranes of stimulated human monocytes. J .•Exp. Med. 167: 389. 30. Roux-LoMBARD, P., C. MODOUX, andJ.-M. DAYER. 1988. Inhibitors of IL-1 and TNFu activities in synovial fluids and cultured synovial fluid cell supernatants. Calc. Tissue Int. 42: (Supp!.): A47 (abstr.) Dr. JEAN-MICHEL DAYER, H6pital cantonal universitaire, Division d'immunologie et d'allergologie, CH-1211 Geneve 4, Switzerland