Int. J. Cancer: 71, 1066–1076 (1997) r 1997 Wiley-Liss, Inc. Publication of the International Union Against Cancer Publication de l’Union Internationale Contre le Cancer AUTOCRINE INTERLEUKIN-1 RECEPTOR ANTAGONIST CAN SUPPORT MALIGNANT GROWTH OF GLIOBLASTOMA BY BLOCKING GROWTH-INHIBITING AUTOCRINE LOOP OF INTERLEUKIN-1 Elisabeth OELMANN1, Annette KRAEMER1, Hubert SERVE1, Birgit REUFI1, Dorothea OBERBERG1, Stephan PATT2, Hermann HERBST3, Harald STEIN3, Eckhard THIEL1 and Wolfgang E. BERDEL1* 1Department of Hematology/Oncology, Benjamin Franklin Hospital, Freie Universität Berlin, 12200 Berlin, Germany 2Department of Neuropathology, Benjamin Franklin Hospital, Freie Universität Berlin, 12200 Berlin, Germany 3Department of Pathology, Benjamin Franklin Hospital, Freie Universität Berlin, 12200 Berlin, Germany In situ hybridization (ISH) of human glioblastoma tissue sections revealed expression of interleukin-1 (IL-1)a and/or b and IL-1 receptor types I and II (IL-1R I and II) in the majority of cases evaluable. To understand the function of IL-1-family members in human glioblastomas, we have studied 6 glioblastoma cell lines. RT-PCR, ISH, ELISA and 125I-IL-1-binding assays revealed expression of IL-1 and high-affinity receptors for human (h)IL-1 in all but 1 cell line. Using a colony growth assay in semi-solid media for testing serial plating efficacy (PE, number of colonies per number of cells seeded in %), only the IL-1R-negative cell line was not influenced by recombinant human (rh)IL-1a or -b, whereas IL-1 down-regulated the self-renewal of clonogenic cells of the other glioblastomas. Tritiated thymidine uptake was down-regulated by rhIL-1 in all cell lines studied. Cell viability remained unchanged by rhIL-1. Wherever growth modulation by rhIL-1 was detected, it could be reversed by either soluble IL-1R I or II or by rhIL-1 receptor antagonist (ra). IL-1ra not only was able to reverse rhIL-1-induced growth modulation but alone could modulate glioblastoma growth in comparison with control in cell lines producing IL-1. Our results show the presence of public autocrine loops for IL-1 leading to growth inhibition in some glioblastomas. To understand these loops, we have studied expression and function of IL-1ra in glioblastomas. ISH of human glioblastoma tissue sections revealed expression of hIL-1ra in all 8 cases evaluable. In 4 of 6 cell lines, IL-1ra was found in the supernatant under constitutive conditions, the IL-1R-negative line being among the 2 non-producers. The other non-producing cell line, HTB 17, showed expression of hIL-1R II. Most interestingly, a neutralizing antibody against IL-1ra down-regulated growth of IL-1- and IL-1ra-producing glioblastoma cells to approx. 30% of the controls. Thus, public autocrine loops for IL-1 in human glioblastomas exist and result in growth inhibition. An autocrine production of IL-1antagonizing molecules such as IL-1ra by these tumors can counteract this IL-1 function and represent a basic escape mechanism supporting malignant growth in some glioblastomas. Int. J. Cancer 71:1066–1076, 1997. r 1997 Wiley-Liss, Inc. IL-1a and IL-1b are pleiotropic cytokines with a multitude of activities in a wide range of cell types and different tissues (Dinarello, 1991, 1994, 1996; Platanias and Vogelzang, 1990; Schmidt and Tocci, 1991). They have some amino-acid homology, bind to the same type I and II cell surface receptors and share biologic activities (Dinarello, 1991, 1994). Type I receptors are important for signaling (Sims et al., 1993); type II receptors may be released from cells and function as a decoy target for IL-1 (Colotta et al., 1993; Symons et al., 1995). In addition to these shIL-1Rs, naturally occurring antagonists for IL-1 have been described (Arend, 1993; Dinarello, 1991; Dinarello and Thompson, 1991; Larrick, 1989). The cDNA of the receptor antagonist IL-1ra has been cloned (Eisenberg et al., 1990); the encoded protein shows some amino-acid homology with IL-1a and IL-1b (Dinarello, 1991), binding to IL-1R without initiating IL-1 signal transduction (Dripps et al., 1991), and can block a multitude of IL-1 effects (Arend, 1993; Dinarello and Thompson, 1991). Thus, IL-1 activity seems to be the result of a fine-tuned network of agonist and antagonist molecules within the IL-1 family. Diverse effects by IL-1 on the growth of cells from solid tumors have been reported. The cytokine can exert growth-inhibitory activity in some tumor cells (Bertoglio et al., 1987; Danforth and Sgagias, 1993; Gaffney and Tsai, 1986; Kilian et al., 1991; Lachman et al., 1986; Onozaki et al., 1985) but also stimulates growth of several tumor cell lines (Hamburger et al., 1987; Ito et al., 1993; Lahm et al., 1992), including autocrine growth stimulation (Zeki et al., 1993). Although these observations remain contradictory, several clinical studies with IL-1 in cancer patients are under way (Crown et al., 1991; Redman et al., 1994; Schuchter et al., 1994; Smith et al., 1992; Triozzi et al., 1995). Detailed studies have been published on the reduction of growth of myeloid leukemic cells by IL-1ra (Estrov et al., 1991; Yin et al., 1992). However, only few reports exist on the potential role of IL-1ra in solid tumors. Expression of IL-1ra has been reported in endometrial cancer (Van Le et al., 1991) and bronchogenic carcinoma (Smith et al., 1993). In both studies there was higher expression of IL-1ra in tumor cells than in normal tissues, which was discussed as being important in tumor evasion of host defense (Smith et al., 1993). We have shown that modulation of colony formation by rhIL-1 of some tumor cell lines was completely blocked by rhIL-1ra (Oelmann et al., 1994). There are many studies on the role and functions of IL-1 in the brain (Rothwell, 1991; Merrill, 1987; Smith, 1992). Of interest for this investigation are the following findings. It has been shown clearly that brain macrophages (microglia) produce IL-1 (Giulian et al., 1986, 1988; Hetier et al., 1988; Malipiero et al., 1990). Production is highest at the time of birth (Giulian et al., 1988). Furthermore, there is expression of IL-1R located on neurons throughout the brain (Dinarello, 1991; Rothwell, 1991), while IL-1R on glial cells are expressed after brain injury (Rothwell, 1991). Thus, IL-1 appears to be important for brain development and response to brain injury (Rothwell, 1991). IL-1 is expressed by astrocytoma and glioma cell lines (Fontana et al., 1982; Lee et al., 1989) and IL-1 mRNA was found in primary brain tumors (Merlo et al., 1993). Although growth-promoting effects of exogenous IL-1 have been reported in an astrocytoma cell line (Bertoglio et al., 1987; Lachman et al., 1987), these early observations must be interpreted with caution. The authors have employed only short incubation times, and a detailed study Abbreviations: BSA, bovine serum albumin; DMEM, Dulbecco’s modified Eagle’s medium; DTT, dithiothreitol; ELISA, enzyme-linked immunosorbent assay; FCS, fetal calf serum; h, human; HTCA, human tumorcloning assay; ICE, IL-1b-converting enzyme; IL-1, interleukin-1; IL-1ra, IL-1 receptor antagonist; IL-1R, IL-1 receptor; ISH, in situ hybridization; MAb, monoclonal antibody; PDH, pyruvate dehydrogenase; PE, plating efficiency; rh, recombinant human; RT-PCR, reverse transcriptasepolymerase chain reaction; sh, soluble human. *Correspondence to: Department of Hematology/Oncology, Benjamin Franklin Hospital, Freie Universität Berlin, 30 Hindenburgdamm, 12200 Berlin, Germany. Fax: 49 30 8445 4479. Received 9 December 1996; accepted 29 January 1997 INTERLEUKIN-1 NETWORK IN GLIOBLASTOMA performed later on the same cell line reported only transient growth stimulation but terminal differentiation of the cells after prolonged incubation with exogenous IL-1 (Tanaka et al., 1994). Up to now, there has been no comprehensive study on the expression and functional role of agonist and antagonist members of the IL-1 family in human glioblastoma. Using human glioblastoma tissue sections and cell lines, we show here the presence of autocrine loops for IL-1 in glioblastomas which can lead to growth inhibition and describe autocrine loops for IL-1-antagonizing molecules such as IL-1ra in cell lines which, by counteracting IL-1 loops, represent an escape mechanism supporting malignant growth of some glioblastomas. MATERIAL AND METHODS Tissue Tissue specimens were obtained from glioblastoma patients of our hospital during surgery according to our ethical board guidelines. Paraffin-embedded sections of glioblastoma multiforme were studied by ISH. Cells HTB 14 and HTB 17 are human glioblastoma cell lines and were purchased from the ATCC (Rockville, MD). All other human glioblastoma cell lines were kindly provided by Dr. D. Stavrou (Hamburg, Germany). Cell-culture techniques were performed according to standard procedures, and cells were routinely checked for Mycoplasma. Cytokines RhIL-1a was purchased from Genzyme (origin Escherichia coli, 105 units/µg; Cambridge, MA). RhIL-1b was from Genzyme (origin E. coli, 5 3 105 units/µg). RhIL-1ra was from PeproTech (origin E. coli; Rocky Hill, NJ). shIL-1R types I and II (origin Chinese hamster ovary cells) were a kind gift of Dr. J.E. Sims (Immunex, Seattle, WA). Antibodies Antibodies against IL-1a and -b were purchased from PeproTech (origin rabbit; catalog numbers 500-P21A and 500-P21B). The mouse MAb against hIL-1R I (Genzyme, code 1592-01) is an IgG1 antibody (expressed in C 127 cells). The rat MAb against hIL-1R II (Genzyme, code 80-3503-01) is an IgG2b antibody. Polyclonal antibodies neutralizing hIL-1ra were from either WakChemie (Bad Homburg, Germany, number AB-280-NA) or Genzyme (code 80-2975-01). ISH After linearization of plasmids (pGEM-3Z; Promega, Madison, WI; or pCR II for IL-1ra; Invitrogen, San Diego, CA) containing specific sequences of the genes for hIL-1a and -b (R&D Systems, Minneapolis, MN), hIL-1R type I and type II (kindly provided by Immunex) and hIL-1ra (produced by PCR from the common sequence of hIL-1ra from fetal liver), 35S-labeled run-off anti-sense and sense (control) transcripts were generated using Sp6 and T7 RNA polymerases. ISH for the detection of RNA transcripts was performed as previously described (Herbst et al., 1992). In brief, dewaxed and rehydrated paraffin sections were exposed to 0.2 N HCl and 0.125 mg/ml pronase (Boehringer, Mannheim, Germany) followed by acetylation with 0.1 M triethanolamine, pH 8.0/0.25% (v/v), acetic anhydride and dehydration through graded ethanols. Slides were hybridized to 2 to 4 3 105 cpm of labeled probes overnight at 54°C. Washing and autoradiography were performed as described by Milani et al. (1989). All sections were processed in parallel using the same batches of reagents and probes. Incubation of sections with Micrococcus nuclease (Boehringer, Mannheim) prior to ISH resulted in extinction of the specific autoradiographic signal, establishing that RNA sequences were targets of the hybridization procedure (Williamson, 1988). 1067 RT-PCR Total RNA was isolated according to the instructions of the supplier of RNAzol (Paesel and Lorei, Frankfurt, Germany). The amount of RNA isolated was determined spectrophotometrically. Reverse transcription was carried out using 1 µg of total RNA, 0.8 µM oligo-p(dT)15 primer, 0.125 mM of each dNTP, 13.5 units of RNAse inhibitor (Promega) and 200 units of MMLV reverse transcriptase in 13 final concentration of the reverse transcriptase buffer (Life Technologies, Eggenstein, Germany) and supplemented with DTT to yield a final concentration of 10 mM. The final reaction volume was 20 µl. The mixture was incubated at 37°C for 60 min and subsequently for 5 min at 95°C to inactivate the reverse transcriptase. PCR was performed in a total volume of 50 µl containing 4 µl of the RT reaction mixture, 10 mM Tris-HCL (pH 8.3), 50 mM KCl, 1.5 mM MgCl2, 0.001% (w/v) gelatin, 0.125 mM dNTPs, 1 µM of each up- and down-stream primer and 1.4 units of Taq polymerase (Angewandte Gentechnologie Systeme, Heidelberg, Germany). Amplification was carried out in a Perkin Elmer (Norwalk, CT) 9600 Thermocycler using a modified ‘‘hot-start’’ technique. The following RNA transcripts were detected via amplification of the corresponding cDNAs: (i) the b-subunit of PDH, using the primer pair composed of the (1) primer 58-GGT ATG GAT GAG GAC CTG GA-38 and the (2) primer 58-CTT CCA CAG CCC TCG ACT AA-38, yielding an amplicon of 105 bp; (ii) hIL-1a, using specific primer pairs (Clontech, Palo Alto, CA) which amplify a 491-bp fragment; (iii) hIL-1b, using specific primer pairs (Clontech) which amplify an 802-bp fragment; (iv) hIL-1R type I, using specific primer pairs (Clontech) which amplify a 300-bp fragment; (v) hIL-1R type II, using specific primer pairs composed of the (1) primer 58-GAA GAG ACC ATT CCT GTG ATC-38 and the (2) primer 58-GAA AGT CTT GAT GAT GAG GCC-38, which amplify an expected fragment of 481 bp; (vi) intracellular form of hIL-1ra, using a primer pair (Haskill et al., 1991) composed of the (1) primer GM 397 and the (2) primer GM 368, yielding a 512-bp PCR product, common sequence of hIL-1ra: (1) primer 58-TTA ACA TCA CTG ACC TGA GCG AGA ACA G-38 and (2) primer 58-CCT GGA AGT AGA ATT TGG TGA CCA TGA C-38, producing a 201-bp amplicon; (vii) ICE, using specific primer pairs (Thornberry et al., 1992) composed of modified primer pairs: p10 (1) primer 58-GCT ATT AAG AAA GCC CAC ATA GA-38 and (2) primer 58-TTC AGT GGT GGG CAT CTG CG-38, amplifying a 222-bp fragment and p20 (1) primer 58-GAC AAC CCA GCT ATG CCC AC-38 and (2) primer 58-CGG CTT GAC TTG TCC ATT ATT G-38, producing a 128-bp amplicon. The cycle program for each primer pair was preceded by an initial denaturation at 95°C for 4 min, a specific annealing temperature according to the Tm of each primer, and followed by a final extension at 72°C for 10 min. The cycle program comprised 35 cycles of 95°C for 1 min, annealing for 1.5 min and 72°C for 1.5 min. Amplification of transcripts of the PDH b-subunit was efficient using this cycle program. Amplification of the PDH b-subunit was used to judge DNA contamination in the RNA samples of the different cell lines examined and, moreover, to justify the comparability of amplification results of the specific target regions in the different cell lines. Amplicon identities were verified by Southern hybridization with the cDNA probes as mentioned above (details not shown). ELISA Cells were washed in RPMI 1640 medium without serum and incubated for 48 hr at pH 7.2, 37°C, 5% CO2 and high humidity in RPMI 1640 medium containing 0.2% BSA without serum. Cell concentrations at the beginning of the 48-hr incubation period were 1 to 2 3 106 cells/ml. Subsequently, cell supernatants were harvested and assayed for hIL-1a, hIL-1b, hIL-1b precursor or hIL-1ra, respectively. We have used ELISA kits obtained from Endogen (hIL-1a, code EH-IL1A; hIL-1b, code EH-IL1B; Boston, MA), R&D Systems (hIL-1ra, code DRA 00) and Cistron Biotech. (hIL-1b precursor, code 03-1000; Pine Brook, NJ). 1068 OELMANN ET AL. FIGURE 1 – Expression of hIL-1a (a), hIL-1b (b), IL-1 receptor type I (c), hIL-1 receptor type II (d) and hIL-1ra (e) in tissue sections (f, example for sense control) of human glioblastomas (results, see g) as shown by ISH. Black grains represent expression of specific messages. INTERLEUKIN-1 NETWORK IN GLIOBLASTOMA FIGURE 2 – RT-PCR analysis of hIL-1a, hIL-1b and ICE transcripts (a) and of hIL-1R types I and II transcripts (b) in human glioblastoma cell lines. M, molecular markers. For bp sizes see ‘‘Material and Methods’’. Binding assay Dissociation constants (Kd ) of hIL-1-binding sites were determined by Scatchard analysis of (3-[125I]-iodotyrosyl)-labeled hIL-1a (Amersham, Aylesbury, UK; specific activity, 2,000 Ci/mmol or 74 TBq/mmol) binding (Oelmann et al., 1995) with the following modifications. Cells were plated in 6-well plates at 1.0 to 10.0 3 105 cells/well 1–5 days before the binding assay to allow adherence. Cells were counted on the day of the assay. After washing the cells once with serum-free DMEM containing 2.5 mg/ml BSA, 125I-labeled hIL-1a (10 pM–1 nM) was added in 0.5 ml DMEM/ BSA to duplicate wells and incubated for 60 min at 37°C. Non-specific binding was determined at each concentration step by adding 100 3 molar excess hIL-1a to duplicate wells. Cells were then washed 3 times with ice-cold DMEM/BSA and removed from the wells by lysis with 2 3 0.5 ml 1 N NaOH. Radioactivity was counted in a Berthold gamma counter. Experiments were repeated 3–4 times. HTCA For evaluation of anchorage-independent clonal colony growth of the cell lines, a newly developed HTCA using mixtures of methylcellulose and agar (Oelmann et al., 1994; Topp et al., 1993) was used with the following modification: cells were detached by 1069 trypsinization and washed with their own growth medium, resuspended with 1 ml RPMI 1640 medium (GIBCO, Glasgow, UK) plus 10% FCS and counted by Trypan blue staining to yield a final concentration of 3 3 104 cells/ml. Viability of the cells .80% was required before cells were taken for an experiment. Methylcellulose solution was produced by boiling 0.5 I distilled water with 21 g methylcellulose (code M 0512; Sigma, Deisenhofen, Germany). Five hundred milliliters of cold Iscove’s modified Dulbecco’s medium (double concentrated; GIBCO, code 041-90132) was added to the methylcellulose after cooling down to 37.0°C. The mixture was kept in 3.6 ml aliquots at 220°C. Agar was dissolved by boiling 3 g Difco agar (Agar Noble; Difco, Detroit, MI) in 100 ml distilled water for 30 min, and consecutively 10 ml of the boiling agar were added to 20 ml RPMI 1640 (37.0°C). The incubation mixture was made up of 3.6 ml methylcellulose solution, 2.7 ml FCS (Hyclone, Logan, UT; code A-1111-D), 0.06 ml mercaptoethanol (8.4 3 1023 M mercaptoethanol in distilled water; GIBCO; code 043-01350 D), 0.3 ml cell suspension, 0.8 ml Iscove’s medium (Gibco; code 041-01980M) and 1.6 ml agar/ RPMI 1640 mixture. This incubation mixture was vortexed thoroughly and kept in the dark at 37.0°C for 20 min. RhIL-1 or control vehicle was added to Lux dishes (suspension culture dishes, 35 3 10 mm, code 174926; Miles, Naperville, IL) as a solution of 0.1 ml PBS (Gibco; code 14190-094) plus 0.1% BSA and rhIL-1. An aliquot of 1 ml of the incubation mixture was then added to the dishes. This final incubation mixture contained the final cytokine concentrations, as indicated in the results; thus, tumor cells were exposed to the cytokine for the complete assay period. For all experiments in which rhIL-1ra, shIL-1R or antibodies were used together with rhIL-1, these agents were incubated for 1 hr at room temperature or 37.0°C with the cells or the cytokines as indicated before being added to the assay. The number of cells finally seeded per dish was 1 3 103. Colony formation was evaluated with an inverted microscope before (to exclude cell clumping) and after an incubation period of 10 days at pH 7.2, 37°C, in an atmosphere of 5% CO2 and high humidity. PE was defined as the number of colonies per number of cells seeded, expressed as a percentage. PE1 represents PE observed at the completion of 1 HTCA. Serial PE was assayed by removing single colonies from either control or experimental cultures with micropipettes, obtaining single-cell suspensions by mechanical means, washing these cells and directly replating them at identical numbers per dish into HTCA. PE2 was calculated on the basis of the numbers of colonies counted at the end of a second HTCA. HTCA was used since this assay has been shown to reliably detect growth modulation of tumor cells by cytokines and to be predictive for in vivo tumorigenicity and modulation of in vivo tumor growth by cytokines (Freedman and Shin, 1974; Gross et al., 1988; Topp et al., 1993). Tritiated thymidine uptake Before the assay was started, cells were starved by reducing the serum concentration to 1.0% for 24 hr in order to obtain a synchronization effect on the cell cycle. Thymidine uptake kinetics under different serum concentrations had been obtained before and the serum concentrations used for starving allowed for 3–50% of tumor cell thymidine uptake over 24 hr compared with uptake in 10% FCS. For the tritiated thymidine uptake assay, 100-µl aliquots of the cytokines or antibodies in test medium (usual growth media of the tested cell lines) were seeded into 6 wells per test group of 96-well flat-bottomed microtiter plates (Greiner, Nürtingen, Germany) at indicated concentrations. The wells already contained 100 µl of the cell suspension (0.5 to 2 3 104 cells/well, depending on the growth kinetics of the cell line used). Controls contained 100 µl of pure test medium instead of the cytokine or antibody. Wherever antagonist members of the IL-1 family, such as rhIL-1ra or shIL-1R or neutralizing antibodies, were used in addition to rhIL-1, pre-incubation conditions with either the cells (e.g., rhIL-1ra, antibody against IL-1R) or the recombinant cytokine (e.g., shIL-1R types I or II, antibodies against IL-1 or IL-1ra) were as stated for OELMANN ET AL. 1070 FIGURE 3 – Specific binding of 125I-labeled hIL-1a to human glioblastoma cell lines. (a) X-axis, concentrations of Scatchard analysis of the binding data is shown as inset. (b) Kd values, sites per cell and correlation co-efficients. HTCA. Plates were incubated at 37°C, pH 7.2, in an atmosphere of 5% CO2 and high humidity for up to 72 hr. Cultures were pulsed for the last 6 hr with 1.0 µCi of [3H] thymidine (specific activity 5.0 Ci/mmol; Amersham) per well. Samples were processed and counted in an LKB Betaplate system (LKB Pharmacia, Freiburg, Germany). Values are given as means 6 SD. Cell numbers and viability All cell lines were incubated in 6-well plates (5 3 104 cells/well) in 3 ml medium plus 10% FCS with the cytokine concentrations as indicated for 24–168 hr in the above conditions (tritiated thymidine uptake assay). Subsequently, cells were detached by trypsinization and stained with Trypan blue to evaluate cell number and viability. Statistics Results were statistically evaluated by the Kruskal-Wallis test or the Mann-Whitney test as indicated; p values , 0.05 were interpreted as indicating significant differences. RESULTS Expression of IL-1 and IL-1R in human glioblastoma tissue sections in situ In a first set of experiments utilizing ISH, we assayed 10 human glioblastoma sections for the presence of hIL-1 and hIL-1R expression. The majority of human glioblastoma tissue sections evaluable expressed mRNA for hIL-1a, hIL-1b, and hIL-1R types I and II (Fig. 1). The main reason for being not evaluable was high 125I-labeled hIL-1a. FIGURE 4 – (a) Influence of rhIL-1a on colony formation (PE1 ) of human glioblastoma cell lines in HTCA. Values are expressed as % of controls (tumor cells only) and represent means 6 SD of 6-fold assays. PE values for controls were 87-HG-31, 10.8%; HTB 17, 2.2%; HTB 14, 11.2%; 86-HG-39, 10.7%; 87-HG-28, 1.6%; 88-HG-14, 11.6%. Since 103 cells were seeded per dish, the numbers of colonies per dish can be calculated as PE 3 10. *, p . 0.05 (indicating nonsignificant differences) when compared with the tumor cell (only) controls (Mann-Whitney test). All other values were significantly ( p , 0.05) different from controls. (b) Influence of rhIL-1 on the secondary plating efficiency (PE2 ) of human glioblastoma cell lines in HTCA. PE1 values after the first HTCA were calculated from means of 6-fold assays showing significant ( p , 0.05, Mann-Whitney test) stimulation (87-HG31) or inhibition (88-HG-14) of colony formation by rhIL-1a/b in comparison with control, given as 100%. Single colonies from either the controls or the rhIL-1-incubated cultures of the first HTCA were removed by micropipettes, gently agitated to yield single-cell suspensions, washed and immediately replated without rhIL-1 into a second HTCA at equal cell numbers per dish for calculation of PE2 from means of 6-fold assays. PE2 values from control colonies of the first HTCA did not decrease significantly when compared to PE1, indicating no toxicity of the replating procedure itself. All PE2 values obtained from rhIL-1-incubated cultures of the first HTCA were significantly ( p , 0.05, Mann-Whitney test) lower than those from control colonies of the first HTCA. (c) Blocking of the rhIL-1 effect by either rhIL-1ra or shIL-1R I using the 88-HG-14 cell line in HTCA. Values represent means 6 SD of 6-fold assays. Mann-Whitney tests were performed comparing control values with rhIL-1 condition (*) or comparing rhIL-1 condition with rhIL-1 plus inhibitor condition (**). Asterisks represent p values , 0.05. FIGURE 4 OELMANN ET AL. 1072 TABLE I – CONCENTRATION OF IL-1 NETWORK MEMBERS IN THE SUPERNATANTS OF HUMAN GLIOBLASTOMA CELL LINES AFTER 48 HR OF SERUM-FREE INCUBATION Cytokine Cell lines 86-HG-39 88-HG-14 87-HG-28 87-HG-31 HTB 14 HTB 17 hIL-1a1 11.3 6 3.65 64.8 6 55.7 129.2 6 80.7 5.7 6 1.0 60.3 6 42.2 19.3 6 3.3 hIL-1b precursor2 60.66 40.0 50.0 20.0 480.0 10.0 hIL-1b3 8.6 6 7.15 42.7 6 21.9 54.9 6 35.8 8.6 6 8.6 73.7 6 69.9 4.1 6 1.0 hIL-1ra4 n.d.6 2,800.0 14,500.0 12.8 9,500.0 n.d. 1Detection limit of ELISA at 13 pg/ml hIL-1a.–2Detection limit of ELISA at 50 pg/ml hIL-1b precursor.–3Detection limit of ELISA at 4 pg/ml hIL-1b.–4Detection limit of ELISA at 6.5 pg/ml hIL-1ra.–5Means 6 standard error of 2 or 3 experiments with 4-fold or 5-fold assays each, pg/ml.–6Means of one 4-fold assay, pg/ml. n.d., not detectable. non-specific background staining in the sense controls, which led to exclusion of the case. IL-1 message–positive sections and IL-1R message–positive sections were from identical patients. Signals could be located clearly over the tumor cells, however, some stromal cells also showed expression. For both IL-1 a/b and IL-1R I/II, mRNA expression was not equally distributed in all tumor cells of 1 section but occurred in patches of positive cells and/or in single cells distributed throughout the tumor (Fig. 1). Expression of IL-1 and IL-1R and functionality of specific IL-1-binding sites (receptors) in 6 human glioblastoma cell lines Next, we used RT-PCR, ISH and ELISA techniques to examine 6 human glioblastoma cell lines for expression of IL-1 a/b and IL-1R I/II. RT-PCR revealed message for IL-1R I in all cell lines, with 86-HG-39 showing only a faint band (Fig. 2). Type II receptor message expression was minimal in 2 lines, including 86-HG-39, and present in all other cell lines (Fig. 2). Both IL-1a and IL-1b messages were detected in all but 2 cell lines, and all cell lines expressed message for ICE (Fig. 2). Bands detected by RT-PCR were checked using Southern blotting to identify sequence homology with the sequences of interest (details not shown). Amplification of the PDH b-subunit was used to exclude DNA contamination in the RNA samples and to justify the comparability of amplification results of the specific target regions in the different cell lines (details not shown). In addition, data derived from ISH agreed almost exactly with RT-PCR and showed that mRNA expression for both IL-1 a/b and IL-1R I/II was not equally distributed in all tumor cells of 1 line but occurred in patches of positive cells and/or in single cells distributed throughout the cell line (details not shown). ELISA revealed the presence of either IL-1a or IL-1b or both in the supernatants of all cell lines, with 87-HG-31 being at the detection limit (Table I). There was some variability of IL-1 a/b production when ELISAs were performed at different times and numbers of cell passages, with 87-HG-31 ranging from below to clearly above detection limits for IL-1a/b (e.g., IL-1b range not detectable to 45.6 pg/ml; detection limit, 4 pg/ml). To test our cell lines for the presence of functional hIL-1R, we performed 125I-hIL-1-binding assays. These experiments revealed the presence and function of high-affinity receptors for hIL-1, with Kd values of 95–178 pM in all but 1 cell line (Fig. 3). Numbers of IL-1R ranged approx. 400–5,500 per cell. Only 86-HG-39 showed no specific binding of hIL-1. These experiments revealed a good overall agreement between IL-1 and IL-1R expression on the mRNA and the protein levels. Studies on the functional role of IL-1 in glioblastoma cell lines Having established the existence of expression of both IL-1 and IL-1Rs in tissue sections and cell lines of human glioblastomas, we were interested to understand the functional role of the IL-1 network in glioblastomas. Thus, we performed experiments using different assays for cell growth under the influence of rhIL-1 family members. Using a colony growth assay in semi-solid media (HTCA), we compared the activity of rhIL-1a with rhIL-1b and found both to TABLE II – EFFECT OF rhIL-1a ON TRITIATED THYMIDINE UPTAKE OF HUMAN GLIOBLASTOMA CELL LINES IN VITRO rhIL-1a (ng/ml) Control 0.1 1.0 10.0 Cell lines 88-HG-14 100.0 (3,071.3 6 101.0 63.02 27.02 87-HG-31 605.5)1 100.0 (5,467.1 6 943.3) 97.0 49.02 49.02 Cells were grown at 1% FCS for 24 hr for synchronizing, detached with EDTA and incubated with vehicle (control) or rhIL-1a for 48 hr. Tritiated thymidine was added for the last 6 hr before freezing/thawing and processing/counting (for details, see ‘‘Material and Methods’’). 1Values are shown as means of 6-fold assays 6 SD (SD , 20%) expressed as percentage of controls (controls in cpm in parentheses).–2p , 0.01, when compared with control (Mann-Whitney test). be equally effective (details not shown; for examples, see Fig. 4b). For most of the following experiments, therefore, we used rhIL-1a. The IL-1R-negative cell line 86-HG-39 was the only cell line not influenced by rhIL-1, whereas the first plating efficiency (PE1 ) was significantly and dose-dependently down-regulated by rhIL-1 in -2 and up-regulated in the remaining 3 cell lines (Fig. 4a). Formation of a colony generally can reflect self-renewal of clonogenic cells or mitosis with consecutive growth arrest and/or differentiation. The same end point of an HTCA can indicate more than one distinct biological event with consequent fundamental differences for tumor biology. Thus, this read-out system must be interpreted with caution, particularly when an experimental condition yields higher numbers of colonies over controls. To address this question, we removed single colonies from rhIL-1-incubated and control cultures with micropipettes and directly and serially replated them as single-cell suspensions. Testing serial PE of 1 cell line (87-HG-31) up-regulated by rhIL-1 in PE1 and 1 downregulated (88-HG-14) cell line by this method, we observed uniform and drastic reduction of colony formation by rhIL-1 in both lines already in PE2 (Fig. 4b). This reduction reached almost 1 log step. We therefore conclude that IL-1 generally down-regulates the self-renewal of clonogenic glioblastoma cells. To show IL-1 specificity of these observations, we used rhIL-1ra and soluble human IL-1R types I and II for blocking experiments. All 3 agents were able to dose-dependently reverse modulation of colony formation by rhIL-1 (Fig. 4c). Tritiated thymidine uptake was down-regulated by rhIL-1 in the cell lines studied (Table II), and this effect was uniform for 87-HG-31 and 88-HG-14. We did not perform tritiated thymidine uptake assays with HTB 17 and HTB 14 since terminal differentiation by rhIL-1 has been described for these cell lines (Tanaka et al., 1994). To clearly link this inhibition of proliferation by rhIL-1 to ligand–receptor interaction, we used rhIL-1ra also in this assay system. This receptor antagonist was able to completely reverse the rhIL-1 effects when up to 100-fold molar excess concentrations over rhIL-1 were used (Table III). IL-1 antagonism by IL-1ra in other biological systems also requires more than a 10-fold molar excess of the receptor antagonist (our HTCA results above and INTERLEUKIN-1 NETWORK IN GLIOBLASTOMA Arend et al., 1990; Dinarello, 1991; Granowitz et al., 1991a, 1992). As in the HTCA system, rhIL-1ra was able to override rhIL-1 activity and stimulated proliferation as measured by tritiated thymidine uptake in a cell line producing high amounts of IL-1 (88-HG-14; Table III). We counted cells to measure the activity of rhIL-1 during incubation periods of up to 168 hr. In rapidly growing cell cultures, rhIL-1 induced a significantly ( p , 0.05, Mann-Whitney test) lower increase of cell numbers in comparison with controls (e.g., 87-HG-31: mean [6-fold assay] number of cells 3 104 per well 6 SD; control/rhIL-1 cultures after 24 hr, 7.8 6 1.8/4.9 6 1.8; after 72 hr, 44.8 6 5.0/37.3 6 2.6; after 168 hr, 68.9 6 6.4/ 51.0 1 9.4). Where counting the cells after Trypan blue dye staining in parallel cultures after incubation times of up to 168 hr, we observed that cell viability remained unchanged by rhIL-1 (details not shown). FACS analysis of a few cell lines also has not revealed cell death after short-term incubation with rhIL-1 (details not shown). Together, these experiments indicate that IL-1 acts merely cytostatically and antiproliferatively but does not induce cell death. Expression and function of IL-1ra in glioblastomas Since we have shown expression and production of IL-1 and IL-1R in glioblastomas and growth modulation by rhIL-1 in the same tumors, our results suggest the presence of public autocrine loops for IL-1 leading to growth inhibition in some glioblastomas. The existence of such growth-inhibiting loops in a rapidly growing malignant tumor is hard to understand. Thus, we have looked for possible escape mechanisms. In several experiments, rhIL-1ra not only was able to reverse rhIL-1-induced growth modulation but alone could modulate glioblastoma growth in comparison with controls in cell lines producing IL-1 (see above), making this member of the IL-1 family a possible candidate for promotion of malignant growth. Using ISH, we found expression of hIL-1ra message in all 8 evaluable (for reasons of being non-evaluable, see above) glioblastoma tissue sections (Fig. 1) and in the majority of our glioblastoma cell lines also expressing hIL-1 and hIL-1R (details not shown). Expression was not equally distributed in all tumor cells but occurred in patches of positive cells and/or in single cells distributed throughout the tumor (Fig. 1). Details of the distribution of positive cells within a tumor are under further investigation. ISH of tissue sections revealed signals for hIL-1ra after shorter exposition time (21 days) and of higher intensity in the positive cells than for IL-1 or the receptors. However, some members of the IL-1 network, such as the type II receptor, revealed a much higher percentage of positive cells (details not shown). Testing our cell lines with RT-PCR, we found expression of message of the intracellular or secreted forms of IL-1ra in all but 2 1073 cell lines, the IL-1R-negative line (86-HG-39) being among the 2 non-transcribers (details not shown). Using ELISA, IL-1ra was found also in the supernatant under constitutive conditions in 4 of 6 lines, the IL-1R-negative line again being among the 2 nonproducers (Table I). There was good correlation between PCR and ELISA results. HTB 17 was the only cell line not producing IL-1ra but secreting IL-1 and possessing IL-1R, including expression of type II receptors. To study the functional role of autocrine IL-1ra secretion, we performed HTCA and tritiated thymidine uptake experiments with neutralizing antibodies against hIL-1ra. This neutralizing antibody against IL-1ra significantly down-regulated growth of IL-1- and IL-1ra-producing glioblastoma cells (88-HG-14) in both assays (see Table IV for results of tritiated thymidine uptake). In contrast, this antibody showed no non-specific toxicity in cell lines not producing IL-1ra (not shown). We have interpreted these results as reconstituting efficacy of growth-inhibiting IL-1 loops by blocking autocrine IL-1ra activity. These results convincingly show autocrine loops for IL-1ra which counteract the negative growth regulation by autocrine IL-1 loops in some glioblastomas. Further evidence for this conclusion comes from experiments showing inefficiency of neutralizing antibodies against hIL-1a or hIL-1b in growth modulation of glioblastoma cells producing autocrine IL-1ra (Table IV, almost identical HTCA results not shown in detail). Thus, the autocrine production of IL-1-antagonizing molecules such as IL-1ra represents a basic escape mechanism supporting malignant growth in some glioblastomas. DISCUSSION Our results show that the majority of human glioblastomas express IL-1a, IL-1b or both, as well as IL-1R types I or II or both, in situ (Fig. 1). This also holds true for human glioblastoma cell lines (Fig. 2), which can produce and secrete either IL-1a or IL-1b or both (Table I) and possess functional high-affinity receptors for IL-1 (Fig. 3). Addition of exogenous rhIL-1 down-regulates growth of IL-1R-positive cell lines in a variety of assays (Fig. 4, Tables II, III), and this can be blocked by antagonist members of the IL-1 family, such as rhIL-1ra (Table III, Fig. 4). This receptor antagonist alone can modulate glioblastoma growth in comparison with controls in the cell lines producing IL-1 (Table III). Our results clearly demonstrate the presence of public autocrine loops for IL-1 in human glioblastomas leading to growth inhibition. Interestingly, autocrine production of antagonist members of the IL-1 network, such as IL-1ra, can block this IL-1 activity (Tables I, IV). Thus, by blocking growth inhibition of autocrine IL-1, this represents a basic escape mechanism allowing and supporting malignant growth of some glioblastomas (Tables III, IV). Caution, however, is necessary when using the term ‘‘autocrine’’ since the distribution of the single IL-1 family members and the receptors is not homogeneous and paracrine activity may be operative as well. TABLE III – BLOCKING OF THE rhIL-1a (10 ng/ml) EFFECT ON TRITIATED THYMIDINE UPTAKE INTO HUMAN GLIOBLASTOMA CELL LINES BY rhIL-1ra Condition Control rhIL-1a rhIL-1a 1 rhIL-1ra Cell lines 88-HG-14 87-HG-31 4,016.9 6 293.6 (100.0)1 15.02 137.02 26,478.3 6 1,452.0 (100.0) 78.02 103.0 Cells were grown at 1% (88-HG-14) or 10% (87-HG-31) FCS for 24 hr, detached with EDTA and incubated with vehicle (control), rhIL-1a (10 ng/ml) or rhIL-1a plus 100-fold molar excess (over IL-1a) concentration of rhIL-1ra for 24 hr. Tritiated thymidine was added for the last 6 hr before freezing/thawing and processing/counting (for details, see ‘‘Material and Methods’’). 1Values are shown as means of 6-fold assays 6 SD for the controls in cpm (percentage in parentheses) and as percentage of controls for all other conditions (SD , 20%).–2p , 0.01, when compared with control (Mann-Whitney test). TABLE IV – INFLUENCE OF NEUTRALIZING ANTIBODIES AGAINST hIL-1ra AND hIL-1 ON TRITIATED THYMIDINE UPTAKE OF A HUMAN GLIOBLASTOMA CELL LINE (88-HG-14) PRODUCING IL-1ra AND IL-1 Antibody Cell line 88-HG-14 Control Anti-IL-1ra Anti-IL-1a Anti-IL-1b 37,169 6 3,698 (100.0)1 30.32 82.0 86.0 Cells were grown at 1% FCS for 24 hr for synchronizing, detached with EDTA and incubated with vehicle (control) or with antibody at concentrations of up to 1 mg/ml each for a further 24 hr. Tritiated thymidine was added for the last 6 hr before freezing/thawing and processing/counting (for details, see ‘‘Material and Methods’’). 1Values are shown as means of 6-fold assays 6 SD for the controls in cpm (percentage in parentheses) and as percentage of controls for the antibody conditions (SD , 10%).–2p , 0.01, when compared with control (Mann-Whitney test). 1074 OELMANN ET AL. There have been several reports on growth stimulation and growth inhibition of glioblastomas caused by addition of IL-1 (Bertoglio et al., 1987; Lachman et al., 1987; Tanaka et al., 1994), the reports on growth inhibition and terminal differentiation clearly prevailing. Other reports have noted production of IL-1 by astrocytoma and glioma cell lines (Fontana et al., 1982; Lee et al., 1989) and expression of IL-1 mRNA in primary brain tumors (Merlo et al., 1993). IL-1ra mRNA expression has been observed in some human glioblastoma tissue sections (Tada et al., 1994). We present here a comprehensive study on the expression and functional role of agonist and antagonist members of the IL-1 family in human glioblastoma. IL-1ra amounts found to be produced by some of the cell lines are rather high in comparison with IL-1 (Table I). In this respect, it is of interest that IL-1 antagonism by IL-1ra in other biological systems also requires more than a 10-fold molar excess of the antagonist (Arend et al., 1990; Dinarello, 1991; Granowitz et al., 1991a, 1992). Additionally, peak plasma concentrations of IL-1ra in endotoxinemia were found to be approximately 100-fold higher than IL-1 (Fischer et al., 1992; Granowitz et al., 1991b). However, cell lines may acquire altered properties in vitro and may not be totally representative for the tumor biology in situ. ISH of tissue sections revealed IL-1ra signals after shorter exposition time and of higher intensity in the positive cells than for IL-1 or the receptors, but some members of the IL-1 network, such as the type II receptor, revealed a much higher percentage of positive cells. Higher expression of IL-1ra message in the cell lines over the tissue sections prompted the hypothesis of the cell lines having developed from IL-1ra-positive cells of a tumor. Additionally, 1 cell line expressing the IL-1 loop (HTB 17) did not produce detectable amounts of IL-1ra, but did express mRNA for hIL-1R II. Further studies are needed to show whether other escape mechanisms, such as the production of soluble IL-1Rs, may be operative in this line. Soluble extracellular domains of IL-1R II have been described to be produced and secreted and to act as antagonist members of the IL-1 network (Colotta et al., 1993; Symons et al., 1995). It is not possible from our study to comparatively judge the importance of the single antagonist members of the IL-1 network in situ. Further studies must test the hypothesis of antagonist members of the IL-1 network, such as IL-1ra, as possessing proto-oncogene/ oncogene function for these tumors but may also shed some light on the concept of tumorigenesis as representing aberrant wound healing (Marshall et al., 1992). Normal brain expresses IL-1 and IL-1R, and IL-1 has been discussed as being important for brain development and response to brain injury (Giulian and Lachman, 1985; Rothwell, 1991) including wound healing. Glioblastoma development in scar tissue has been reported (Gruss et al., 1993), though there is no sound statistical basis for a correlation. IL-1 has a role in wound healing (Giulian and Lachman, 1985), and genetic alterations leading to over-expression and production of IL-1ra could represent an important step toward malignant aberration. These observations and hypotheses do not exclude the existence of other mechanisms allowing or facilitating malignant growth of glioblastoma. Interestingly, we have observed 1 cell line which has lost IL-1 control and expresses neither the receptor nor the ligands (86-HG-39). Differentiating autocrine loops have been described for TGF-b in glial progenitor cells (McKinnon et al., 1993). With the exception of some morphological changes, we have no clear evidence for induction of differentiation or senescence by autocrine IL-1 in our experiments, and further studies in this area are necessary. This (McKinnon et al., 1993) also draws attention to the fact that multiple growth factors can influence glioblastoma growth, some of them acting synergistically (Merzak et al., 1995). In this respect, it is of interest that our glioblastoma cell lines constitutively express messages for cytokines known to act antiproliferatively such as TGF-b and TNF (data not shown). The meaning of this observation in a rapidly growing cell line must be further studied. Expression of IL-1ra has been reported in endometrial cancer (Van Le et al., 1991) and bronchogenic carcinoma (Smith et al., 1993). In both studies there was higher expression of IL-1ra in tumor cells than in normal tissues (Smith et al., 1993; Van Le et al., 1991). Thus, our studies in glioblastoma may serve as a model, and we have begun to investigate the role of IL-1 antagonists such as IL-1ra in other tumor histologies. However, the activity of IL-1 in malignant disease is diverse and at times contradictory. Thoughtful and detailed studies clearly indicate a role for this cytokine in stimulating proliferation in tumors of epithelial origin (Hamburger et al., 1987; Ito et al., 1993; Lahm et al., 1992; Woodworth et al., 1995; Zeki et al., 1993), and each tumor type may be regulated individually. Indirect effects of this cytokine, such as augmentation of metastasis (Bani et al., 1991) through altered integrin expression (Garofalo et al., 1995) and their blockade by IL-1ra (VidalVanaclocha et al., 1994), must be taken into consideration when the role of the IL-1 network in progression of malignant disease in vivo is assessed. There is one more aspect of our results. Human tumor-cloning assays are widely used to estimate the cytostatic and cytotoxic potential of new drugs (Von Hoff, 1990). HTCA has been used in our study since it has been shown to reliably detect growth modulation of tumor cells by cytokines and to be predictive for in vivo tumorigenicity and modulation of in vivo tumor growth by cytokines (Freedman and Shin, 1974; Gross et al., 1988; Topp et al., 1993). However, as with other biological read-out systems, caution is necessary when HTCA results are interpreted, and this holds particularly true when any kind of stimulation is observed. As for hemopoiesis, careful studies are necessary to distinguish between stimulation of self-renewal and stimulation of mitotic divisions before growth arrest or during differentiation. Such studies unequivocally have revealed negative growth regulation by IL-1 in our as well as in other investigations (Tanaka et al., 1994) with glioblastomas. Thus, equating up- and down-regulation of colony formation by cytokines such as IL-1 in varying percentages with ‘‘growth’’ without further in-depth investigation of the biological meaning of this read-out may be misleading (Hanauske et al., 1992; Koch et al., 1995). In conclusion, we have shown that autocrine IL-1ra is overexpressed by some human glioblastoma cell lines and can support malignant growth of these tumors by blocking growth-inhibiting autocrine loops of IL-1. Our finding must be further studied in this and other tumor entities to learn more about the role of the IL-1 network in malignant disease, particularly since therapeutic interventions, e.g., with ribozymes against IL-1 antagonists such as IL-1ra, can be envisaged. ACKNOWLEDGEMENTS The authors thank Dr. J.E. Sims (Immunex, Seattle, WA) for kindly supporting this work. REFERENCES AREND, W.P., Interleukin-1 receptor antagonist. Adv. Immunol., 54, 167– 227 (1993). AREND, W.P., WELGUS, H.G., THOMPSON, R.C. and EISENBERG, S.P., Biological properties of recombinant human monocyte-derived interleukin 1 receptor antagonist. J. clin. Invest., 85, 1694–1697 (1990). BANI, M.R., GAROFALO, A., SCANZIANI, E. and GIAVAZZI, R., Effect of interleukin-1 beta on metastasis formation in different tumor systems. J. nat. Cancer Inst., 83, 119–123 (1991). BERTOGLIO, J.H., RIMSKY, L., KLEINERMAN, E.S. and LACHMAN, L.B., B-cell line-derived interleukin 1 is cytotoxic for melanoma cells and promotes the proliferation of an astrocytoma cell line. Lymphokine Res., 6, 83–91 (1987). COLOTTA, F., RE, F., MUZIO, M., BERTINI, R., POLENTARUTTI, N., SIRONI, M., GIRI, J.G., DOWER, S.K., SIMS, J.E. and MANTOVANI, A., Interleukin-1 type II receptor: a decoy target for IL-1 that is regulated by IL-4. Science, 261, 472–475 (1993). CROWN, J. and 15 OTHERS, A phase I trial of recombinant human INTERLEUKIN-1 NETWORK IN GLIOBLASTOMA interleukin-1b alone and in combination with myelosuppressive doses of 5-fluorouracil in patients with gastrointestinal cancer. Blood, 78, 1420– 1427 (1991). DANFORTH, JR., D.N. and SGAGIAS, M.K., Interleukin-1a and interleukin-6 act additively to inhibit growth of MCF-7 breast cancer cells in vitro. Cancer Res., 53, 1538–1545 (1993). DINARELLO, C.A., Interleukin-1 and interleukin-1 antagonism. Blood, 77, 1627–1652 (1991). DINARELLO, C.A., The biological properties of interleukin-1. Europ. Cytokine Netw., 5, 517–531 (1994). DINARELLO, C.A., Biologic basis for Interleukin-1 in disease. Blood, 87, 2095–2147 (1996). DINARELLO, C.A. and THOMPSON, R.C., Blocking IL-1: interleukin 1 receptor antagonist in vivo and in vitro. Immunol. Today, 12, 404–410 (1991). DRIPPS, D.J., BRANDHUBER, B.J., THOMPSON, R.C. and EISENBERG, S.P., Interleukin-1 (IL-1) receptor antagonist binds to the 80-kDa IL-1 receptor but does not initiate IL-1 signal transduction. J. biol. Chem., 266, 10331–10336 (1991). EISENBERG, S.P., EVANS, R.J., AREND, W.P., VERDERBER, E., BREWER, M.T., HANNUM, C.H. and THOMPSON, R.C., Primary structure and functional expression from complementary DNA of a human interleukin-1 receptor antagonist. Nature (Lond.), 343, 341–346 (1990). ESTROV, Z., KURZROCK, R., WETZLER, M., KANTARJIAN, H., BLAKE, M., HARRIS, D., GUTTERMAN, J.U. and TALPAZ, M., Suppression of chronic myelogenous leukemia colony growth by interleukin-1 (IL-1) receptor antagonist and soluble IL-1 receptors: a novel application for inhibitors of IL-1 activity. Blood, 78, 1476–1484 (1991). FISCHER, E., VAN ZEE, K.J., MARANO, M.A., ROCK, C.S., KENNEY, J.S., POUTSIAKA, D.D., DINARELLO, C.A., LOWRY, S.F. and MOLDAWER, L.L., Interleukin-1 receptor antagonist circulates in experimental inflammation and in human disease. Blood, 79, 2196–2200 (1992). FONTANA, A., KRISTENSEN, F., DUBS, R., GEMSA, D. and WEBER, E., Production of prostaglandin E and an interleukin-1 like factor by cultured astrocytes and C6 glioma cells. J. Immunol., 129, 2413–2419 (1982). FREEDMAN, V.H. and SHIN, S., Cellular tumorigenicity in nude mice: correlation with cell growth in semi-solid medium. Cell, 3, 355–359 (1974). GAFFNEY, E.V. and TSAI, S.C., Lymphocyte-activating and growthinhibitory activities for several sources of native and recombinant interleukin 1. Cancer Res., 46, 3834–3837 (1986). GAROFALO, A., CHIRIVI, R.G.S., FOGLIENI, C., PIGOTT, R., MORTARINI, R., MARTIN-PADURA, I., ANICHINI, A., GEARING, A.J., SANCHEZ-MADRID, F., DEJANA, E. and GIAVAZZI, R., Involvement of the very late antigen 4 integrin on melanoma in interleukin 1-augmented experimental metastases. Cancer Res., 55, 414–419 (1995). GIULIAN, D., BAKER, T.J., SHIH, L.-C.N. and LACHMAN, L.B., Interleukin 1 of the central nervous system is produced by ameboid microglia. J. exp. Med., 164, 594–604 (1986). GIULIAN, D. and LACHMAN, L.B., Interleukin-1 stimulation of astroglial proliferation after brain injury. Science, 228, 497–499 (1985). GIULIAN, D., YOUNG, D.G., WOODWARD, J., BROWN, D.C. and LACHMAN, L.B., Interleukin-1 is an astroglial growth factor in the developing brain. J. Neurosci., 8, 709–714 (1988). GRANOWITZ, E.V., CLARK, B.D., MANCILLA, J. and DINARELLO, C.A., Interleukin-1 receptor antagonist competitively inhibits the binding of interleukin-1 to the type II interleukin-1 receptor. J. biol. Chem., 266, 14147–14150 (1991a). GRANOWITZ, E.V., CLARK, B.D., VANNIER, E., CALLAHAN, M.V. and DINARELLO, C.A., Effect of interleukin-1 (IL-1) blockade on cytokine synthesis: I. IL-1 receptor antagonist inhibits IL-1-induced cytokine synthesis and blocks the binding of IL-1 to its type II receptor on human monocytes. Blood, 79, 2356–2363 (1992). GRANOWITZ, E.V., SANTOS, A.A., POUTSIAKA, D.D., CANNON, J.G., WILMORE, D.W., WOLFF, S.M. and DINARELLO, C.A., Production of interleukin-1receptor antagonist during experimental endotoxaemia. Lancet, 338, 1423– 1424 (1991b). GROSS, J.L., BEHRENS, D.L., MULLINS, D.E., KORNBLITH, P.L. and DEXTER, D.L., Plasminogen activator and inhibitor activity in human glioma cells and modulation by sodium butyrate. Cancer Res., 48, 291–296 (1988). GRUSS, P., SPOHR, A., LEIBER, A., TASLER, J., MENZL, H. and HOFSTÄDTER, F., On the questions of traumatic origin in a glioblastoma. Zentralbl. Neurochir., 54, 186–189 (1993). HAMBURGER, A.W., LURIE, K.A. and CONDON, M.E., Stimulation of anchorage-independent growth of human tumor cells by interleukin 1. Cancer Res., 47, 5612–5615 (1987). HANAUSKE, A.R., DEGEN, D., MARSHALL, M.H., HILSENBECK, S.G., BANKS, 1075 P., STUCKEY, J., LEAHY, M. and VON HOFF, D.D., Effects of recombinant human interleukin-1a on clonogenic growth of primary human tumors in vitro. J. Immunother., 11, 155–158 (1992). HASKILL, S., MARTIN, G., VAN LEE, L., MORRIS, J., PEACE, A., BIGLER, C.F., JAFFE, G.J., HAMMERBERG, C., SPORN, S.A., FONG, S., AREND, W.P. and RALPH, P., cDNA cloning of an intracellular form of the human interleukin 1 receptor antagonist associated with epithelium. Proc. nat. Acad. Sci. (Wash.), 88, 3681–3685 (1991). HERBST, H., STEINBRECHER, T., NIEDOBITEK, G., YOUNG, L.S., BROOKS, L., MÜLLER-LANTZSCH, N. and STEIN, H., Distribution and phenotype of Epstein-Barr virus-harboring cells in Hodgkin’s disease. Blood, 80, 484– 491 (1992). HETIER, E., AYALA, J., DENÈFLE, P., BOUSSEAU, A., ROUGET, P., MALLAT, M. and PROCHIANTZ, A., Brain macrophages synthesize interleukin-1 and interleukin-1 mRNAs in vitro. J. Neurosci. Res., 21, 391–397 (1988). ITO, R., KITADAI, Y., KYO, E., YOKOZAKI, H., YASUI, W., YAMASHITA, U., NIKAI, H. and TAHARA, E., Interleukin 1a acts as an autocrine growth stimulator for human gastric carcinoma cells. Cancer Res., 53, 4102–4106 (1993). KILIAN, P.L., KAFFKA, K.L., BIONDI, D.A., LIPMAN, J.M., BENJAMIN, W.R., FELDMAN, D. and CAMPEN, C.A., Antiproliferative effect of interleukin-1 on human ovarian carcinoma cell line (NIH:OVCAR-3). Cancer Res., 51, 1823–1828 (1991). KOCH, I., DEPENBROCK, H., DANHAUSER-RIEDL, S., RASTETTER, J.W. and HANAUSKE, A.R., Interleukin 1 modulates growth of human renal carcinoma cells in vitro. Brit. J. Cancer, 71, 794–800 (1995). LACHMAN, L.B., BROWN, D.C. and DINARELLO, C.A., Growth-promoting effect of recombinant interleukin 1 and tumor necrosis factor for a human astrocytoma cell line. J. Immunol., 138, 2913–2916 (1987). LACHMAN, L.B., DINARELLO, C.A., LLANSA, N.D. and FIDLER, I.J., Natural and recombinant human interleukin 1-b is cytotoxic for human melanoma cells. J. Immunol., 136, 3098–3102 (1986). LAHM, H., PETRAL-MALEC, D., YILMAZ-CEYHAN, A., FISCHER, J.R., LORENZONI, M., GIVEL, J.C. and ODARTCHENKO, N., Growth stimulation of a human colorectal carcinoma cell line by interleukin-1 and -6 and antagonistic effects of transforming growth factor b1. Europ. J. Cancer, 28A, 1894–1899 (1992). LARRICK, J.W., Native interleukin 1 inhibitors. Immunol. Today, 10, 61–65 (1989). LEE, J.C., SIMON, P.L. and YOUNG, P.R., Constitutive and PMA-induced interleukin-1 production by the human astrocytoma cell line T24. Cell. Immunol., 118, 298–311 (1989). MALIPIERO, U.V., FREI, K. and FONTANA, A., Production of hemopoietic colony-stimulating factors by astrocytes. J. Immunol., 144, 3816–3821 (1990). MARSHALL, G.M., VANHAMME, L., WONG, W.Y., SU, H. and VOGT, P.K., Wounding acts as a tumor promotor in chickens inoculated with avian sarcoma virus 17. Virology, 188, 373–377 (1992). MCKINNON, R.D., PIRAS, G., IDA, J.A., JR. and DUBOIS-DALCQ, M., A role for TGF-b in oligodendrocyte differentiation. J. Cell Biol., 121, 1397–1407 (1993). MERLO, A., JURETIC, A., ZUBER, M., FILGUEIRA, L., LÜSCHER, U., CAETANO, V., ULRICH, J., GRATZL, O., HEBERER, M. and SPAGNOLI, G.C., Cytokine gene expression in primary brain tumours, metastases and meningiomas suggests specific transcription patterns. Europ. J. Cancer, 29A, 2118–2125 (1993). MERRILL, J.E., Macroglia: neural cells responsive to lymphokines and growth factors. Immunol. Today, 8, 146–150 (1987). MERZAK, A., KOOCHEKPOUR, S., DKHISSI, F., RAYNAL, S., LAWRENCE, D. and PILKINGTON, G.J., Synergism between growth factors in the control of glioma cell proliferation, migration and invasion in vitro. Int. J. Oncol., 6, 1079–1085 (1995). MILANI, S., HERBST, H., SCHUPPAN, D., HAHN, E.G. and STEIN, H., In situ hybridization for procollagen types I, III and IV mRNA in normal and fibrotic rat liver. Evidence for predominant expression in nonparenchymal liver cells. Hepatology, 10, 84–92 (1989). OELMANN, E., SRETER, L., SCHULLER, I., SERVE, H., KOENIGSMANN, M., WIEDENMANN, B., OBERBERG, D., REUFI, B., THIEL, E. and BERDEL, W.E., Nerve growth factor stimulates clonal growth of human lung cancer cell lines and a glioblastoma cell line expressing high-affinity nerve growth factor binding sites involving tyrosine kinase signalling. Cancer Res., 55, 2212–2219 (1995). OELMANN, E., TOPP, M.S., REUFI, B., PAPADIMITRIOU, C., KOENIGSMANN, M., OBERBERG, D., THIEL, E. and BERDEL, W.E., Interleukin-1 receptor antagonist inhibits growth modulation of human tumor cell lines by interleukin-1 in vitro. Int. J. Oncol., 4, 555–558 (1994). ONOZAKI, K., MATSUSHIMA, K., AGGARWAL, B.B. and OPPENHEIM, J.J., 1076 OELMANN ET AL. Human interleukin 1 is a cytocidal factor for several tumor cell lines. J. Immunol., 135, 3962–3968 (1985). PLATANIAS, L.C. and VOGELZANG, N.J., Interleukin-1: biology, pathophysiology, and clinical prospects. Amer. J. Med., 89, 621–629 (1990). REDMAN, B.G., ABUBAKR, Y., CHOU, T., ESPER, P. and FLAHERTY, L.E., Phase II trial of recombinant interleukin-1b in patients with metastatic renal cell carcinoma [Abstract]. Proc. ASCO, 13, 805 (1994). ROTHWELL, N.J., Functions and mechanisms of interleukin 1 in the brain. Trends Pharmacol. Sci., 12, 430–436 (1991). SCHMIDT, J.A. and TOCCI, M.J., Interleukin-1. In: M.B. Sporn and A.B. Roberts (eds.), Peptide growth factors and their receptors I, pp. 473–521, Springer-Verlag, New York (1991). SCHUCHTER, L., NEUBERG, D., ATKINS, M., TESTER, B., RECLO, A., SICKLES, C., MOORE, J., LIGHT, B., ANDERSON, T., CHACHOUA, A. and DUTCHER, J., A phase I study of interleukin-1 alpha (IL-1) and high dose cyclophosphamide (Cy) in patients with advanced cancer [Abstract]. Proc. ASCO, 13, 327 (1994). SIMS, J.E., GAYLE, M.A., SLACK, J.L., ALDERSON, M.R., BIRD, T.A., GIRI, J.G., COLOTTA, F., RE, F., MANTOVANI, A., SHANEBECK, K., GRABSTEIN, K.H. and DOWER, S.K., Interleukin 1 signaling occurs exclusively via the type I receptor. Proc. nat. Acad. Sci. (Wash.), 90, 6155–6159 (1993). SMITH, D.R., KUNKEL, S.L., STANDIFORD, T.J., CHENSUE, S.W., ROLFE, M.W., ORRINGER, M.B., WHYTE, R.I., BURDICK, M.D., DANFORTH, J.M., GILBERT, A.R. and STRIETER, R.M., The production of interleukin-1 receptor antagonist by human bronchogenic carcinoma. Amer. J. Pathol., 143, 794–803 (1993). SMITH, E.M., Hormonal activities of cytokines. Chem. Immunol., 52, 154–169 (1992). SMITH II, J.W. and 20 OTHERS, The toxic and hematologic effects of interleukin-1 alpha administered in a phase I trial to patients with advanced malignancies. J. clin. Oncol., 10, 1141–1152 (1992). SYMONS, J.A., YOUNG, P.R. and DUFF, G.W., Soluble type II interleukin 1 (IL-1) receptor binds and blocks processing of IL-1b precursor and loses affinity for IL-1 receptor antagonist. Proc. nat. Acad. Sci. (Wash.), 92, 1714–1718 (1995). TADA, M., DISERENS, A.C., DESBAILLETS, I., JAUFEERALLY, R., HAMOU, M.F. and DE TRIBOLET, N., Production of interleukin-1 receptor antagonist by human glioblastoma cells in vitro and in vivo. J. Neuroimmunol., 50, 187–194 (1994). TANAKA, S., NAGASHIMA, T., MANAKA, S., HORI, T. and YASUMOTO, S., Growth suppression and astrocytic differentiation of glioma cells by interleukin-1. J. Neurosurg., 81, 402–410 (1994). THORNBERRY, N.A. and 28 OTHERS, A novel heterodimeric cysteine protease is required for interleukin-1b processing in monocytes. Nature (Lond.), 356, 768–774 (1992). TOPP, M.S., KOENIGSMANN, M., MIRE-SLUIS, A., OBERBERG, D., EITELBACH, F., VON MARSCHALL, Z., NOTTER, M., REUFI, B., STEIN, H., THIEL, E. and BERDEL, W.E., Recombinant human interleukin-4 inhibits growth of some human lung tumor cell lines in vitro and in vivo. Blood, 82, 2837–2844 (1993). TRIOZZI, P.L., KIM, J.A., MARTIN, E.W., YOUNG, D.C., BENZIES, T. and VILLASMIL, P.M., Phase I trial of escalating doses of interleukin-1b in combination with a fixed dose of interleukin-2. J. clin. Oncol., 13, 482–489 (1995). VAN LE, L., HASKILL, S., JAFFE, G.J. and FOWLER, W.C., Expression of interleukin-1 and interleukin-1 receptor antagonists in endometrial cancer. Gynecol. Oncol., 42, 161–164 (1991). VIDAL-VANACLOCHA, F., AMÉZAGA, C., ASUMENDI, A., KAPLANSKI, G. and DINARELLO, C.A., Interleukin-1 receptor blockade reduces the number and size of murine B16 melanoma hepatic metastases. Cancer Res., 54, 2667–2672 (1994). VON HOFF, D.D., He’s not going to talk about in vitro predictive assays again, is he? J. nat. Cancer Inst., 82, 96–101 (1990). WILLIAMSON, D.J., Specificity of riboprobes for intracellular RNA in hybridization histochemistry. J. Histochem. Cytochem., 36, 811–813 (1988). WOODWORTH, C.D., MCMULLIN, E., IGLESIAS, M. and PLOWMAN, G.D., Interleukin 1a and tumor necrosis factor a stimulate autocrine amphiregulin expression and proliferation of human papillomavirus-immortalized and carcinoma-derived cervical epithelial cells. Proc. nat. Acad. Sci. (Wash.), 92, 2840–2844 (1995). YIN, M., GOPAL, V., BANAVALI, S., GARTSIDE, P. and PREISLER, H., Effects of an IL-1 receptor antagonist on acute myeloid leukemia cells. Leukemia, 6, 898–901 (1992). ZEKI, K., NAKANO, Y., INOKUCHI, N., WATANABE, K., MORIMOTO, I., YAMASHITA, U. and ETO, S., Autocrine stimulation of interleukin-1 in the growth of human thyroid carcinoma cell line NIM 1. J. clin. Endocrinol. Metabolism, 76, 127–133 (1993).