Distinct patterns of expression of interleukin-1 and by normal and cancerous human ovarian tissues

ARTICLE

INTRODUCTION

Ovarian carcinoma remains the most lethal gynecological neoplasm and the majority of malignant ovarian neoplasms are of epithelial origin [1]. The ovary is not considered an immunologically privileged site. Resident ovarian mononuclear phagocytes, lymphocytes, and polymorphonuclear granulocytes can be observed at various stages of the ovarian life cycle [2, 3]. The immune system defends the body against tumors, among other mechanisms, by regulating the secretion of cellular growth and differentiation factors including cytokines [4, 5]. IL-1, an inflammatory cytokine, is a mediator of cellular cooperation in immune/inflammatory reactions and acts as an autocrine/paracrine factor with pleiotropic activity. It activates various cell types to produce cytokines and/or to express adhesion molecules and also participates in tissue remodeling and scar formation [4, 5]. IL-1 is produced by cells of the immune system after stimulation and also by a variety of cell types including fibroblaste, endothelial cells, epithelial cells and other cells of immune system origin [4, 5]. IL-1 functions with its receptors and antagonists in a concerted fashion. The expression of this "IL-1 system" has been demonstrated in the ovary [6] and several studies indicate the involvement of IL-1 in the physiology of ovarian function [7-9]. Haskill et al. [10] has demonstrated the infiltration of leukocytes into ovarian tumor tissue and ascites fluid. Immune response leads to the recognition of ovarian tumor cells as being non-self and to the production of autologous antibodies that lyse ovarian neoplastic cell lines in vitro [11]. IL-6 and TNF-alpha were shown to be elevated in the serum, as well as in the ascites of ovarian neoplasm bearing patients. However IL-1 serum levels were not different from normal [12], but elevated levels of IL-1ß were demonstrated in the effusions of ovarian epithelial neoplasms [13]. Tumor cells have been shown to produce various autocrine/paracrine factors which apparently are not related to the tissue origin of the tumor [2, 14]. Factors of cellular immune origin such as IL-1, IL-6 and TNF are also involved in some oncogenic processes, in addition to inflammation, autoimmune diseases and infection [5, 15]. IL-1, IL-6 and TNF-alpha induce anorexia in tumor-bearing and infected animals, and IL-1 supplementation to tumor-bearing animals has been shown to potentiate the tumor-bearing anorexia and to reduce the growth rate of the tumor [5, 16]. A direct anti-tumor effect of IL-1 has been reported against malignant melanoma in vitro, breast cancer cells in vitro, and murine pancreatic cancer in vivo [4, 5]. In the present study we have localized the production and the expression levels of both IL-1alpha and IL-1ß in normal and cancerous ovarian tissues by immunohistochemical staining under in vitro conditions.

MATERIALS AND METHODS

Fresh ovarian tissue was collected under sterile conditions from the operation room of the Department of Obstetrics and Gynecology, Soroka Medical Center, Beer-Sheva, Israel. Histopathologically confirmed ovarian carcinoma tissue was obtained from 10 untreated women with ovarian carcinoma at various stages of disease. Eight of these specimen were used for in vitro experiments. Normal ovarian tissue was obtained from 12 women who underwent oophorectomy for non-malignant disease and seven of these specimen were examined by immunohistochemical staining. Fresh tissue was washed immediately in cold PBS in order to eliminate residual blood cells and was prepared for bioassays and immunoassays.

MATERIALS

Dulbecco's Modified Eagle Medium (DMEM), L-glutamine (2 mM), penicillin (100 U/ml), streptomycin (100 mg/ml) (combined antibiotics), fetal calf serum (FCS). All ingredients for media were purchased from Biological Industries (Beth-Haemek, Israel). All culture reagents contained less than 0.025 ng/ml of LPS (lipopolysaccharide). LPS (E. coli 055:B; Difco Laboratories, Detroit, MI, USA). Human recombinant cytokines, IL-1 and IL-2 were kindly provided by Dr. Ron N. Apte, Ben-Gurion University of the Negev, Beer-Sheva, Israel. Other materials used were Bio-Rad Protein Assay Kit (Bio-Rad Lab. GmBH, Munchen, Germany), polyclonal rabbit anti-human IL-1alpha and anti-human IL-1ß were purchased from Genzyme Corp. (Cambridge, MA), broad spectrum biotinylated second antibody and streptavidin-peroxidase conjugate from Zymed (San Francisco, CA), casein (Sigma), urea ANALAR (BDH), trypsin solution (Beth-Haemek, Israel), NaH₂PO₄.2H₂O ANALAR (BDH), K₂HPO₄ (Sigma), Tween 20 (Sigma), proteinase K (Sigma), diamino-benzidine tetrahydrochloride (DAB) (Sigma), and Eukitt (GmbH). ELISA kits specific for IL-1alpha were purchased from R&D Systems (Minneapolis, MN, USA) and ELISA kits for IL-1ß were purchased from T-Cell Diagnostic, Inc. (Cambridge, MA. USA). Disposable tissue-culture flasks and microplates were purchased from Sterilin (Feltham, England).

Establishment of primary normal and cancerous ovarian cell lines:

approximately 1-2 grams of fresh ovarian tissue were used for the establishment of primary cell lines. All procedures and cell manipulations was carried out under sterile conditions. Ovarian tissue was minced with a scalpel into pieces of less than 1 mm³ and dissociated by collagenase (0.05% w/v), DNAse I (0.002% w/v) and hyaluronidase (0.01% w/v) for 2-3 hours at 37° C with stirring until complete dissociation. The cell suspensions were filtered through nylon mesh screens, then centrifuged at 1,200 RPM for 10 min. The cells were suspended in growth medium (DMEM, 5% FCS, glutamine and antibiotics) and cultured in 25 ml bottles. After 7-10 days when monolayers were formed, the cultures were trypsinized and then seeded into new bottles (first passage). At each passage cells were cultured in new bottles to remove contaminating macrophages. Assays were performed after 3 or 4 passages and cells were examined by immunohistochemical staining with anti-keratin antibodies to confirm that the cells were epithelial cells not contaminated by fibroblasts.

Tissue culture activation:

ovarian tumor tissue free from residual blood cells, debris and necrotic areas, was cut into small pieces (1-3 mm) in cold PBS. Weighed samples were cultured in DMEM-medium containing 5% FCS, L-glutamine (2 mM) and combined antibiotics, and stimulated with various concentrations of LPS (0.1-100 µg/ml). After 72 h of incubation (or as indicated), conditioned media were collected and stored at ­ 20° C for cytokine evaluation.

Primary ovarian cell activation:

normal and cancerous ovarian cells (5 x 10⁴ cells/ml) were cultured overnight in 1 ml growth media in 24 well plates. Monolayers were washed with PBS and fresh media with or without various concentrations of LPS (1, 10 and 100 µg/ml) were added for 24-120 hours. After different times of incubation, supernatants were collected and adherent cells were washed with PBS and growth medium added. Plates were freezed and thawed three times and centrifuged at 1,200 rpm for 10 min. Conditioned media (lysates) were collected. Lysates and supernatants were stored at ­ 20° C for cytokine evaluation.

Immunohistochemical staining of ovarian tissue:

formalin-fixed, paraffin-embedded tissue blocks of the ovarian specimens from the 17 patients, were processed by the Department of Pathology, Soroka Medical Center, for immunohistochemical investigation. Ten of the specimens showed primary carcinoma of the ovary (eight of these were also used for in vitro experiments) and seven cases demonstrated no evidence of neoplasm (these tissues were also used for in vitro experiments). Four micron-thick sections were mounted on silane-coated slides, dried at 37° C for 48 h and stored at room temperature. Blocking of the nonspecific background was done with PBS containing 0.05% casein, a modification of the method described by Tacha et al. [17]. This solution was also used to dilute the primary antibodies. Preliminary experiment using sections from two cases were exposed to either trypsin, proteinase K, PBS, urea with boiling, or citrate buffer with boiling for 15 minutes to determine which of the procedures would give the best procedure for antigen unmasking. The best results were obtained with both primary antibodies, polyclonal rabbit anti-human IL-1alpha antibodies (1:20) and polyclonal rabbit anti-human IL-1ß antibodies (1:10) diluted in PBS/casein pH 7.5 after boiling in 6 M urea for 15 min. Boiling in citrate buffer 0.001 M pH 6.0 gave reasonable good results. Negative results were obtained with the other procedures. To dilute the other antibodies and to wash the sections, 0.05% Tween 20 was added to the PBS/casein solution. The biotinylated antibody and the streptavidin-peroxidase conjugate were applied according to the suppliers' directions. Endogenous peroxidase was blocked with 3% H₂O₂ in 80% methanol for 15 min before the streptavidin-peroxidase conjugate was applied. Development was performed with 0.06% DAB and Mayer's haematoxylin was used for counter staining. The sections were mounted in Eukitt.

Preabsorption with the relevant recombinant peptide showed a significant decrease in positive staining for each primary antibody. Negative controls were included for each specimen using PBS/casein instead of the primary antibodies.

Tissue homogenate:

samples of LPS-activated or non-activated normal ovarian tissue were homogenized in ice, with 2 ml of PBS containing 2% combined antibiotics for about 3 minutes. The homogenate was then centrifuged for 10 minutes at 10,000 rpm, and supernatants were stored at ­ 20° C until IL-1 determination.

Protein measurement:

activated and non-activated ovarian tissue was examined for quantification of protein in order to compare the cytokine levels to constant amount of ovarian tissues. The results were expressed as optical density (OD) or pg/mg protein or per gram of tissues. The results were similar for each formula used. The protein contents of the ovarian tissues homogenates were determined using the Bio-Rad Protein Biossay, which is a dye-binding assay based on the differential color change of a dye in response to various concentrations of protein as determined at OD 595 nm.

IL-1 bioassay:

to determine IL-1 bioactivity in ovarian tissue conditionned media, the 1A-5 helper T-cell line was used. The 1A-5 helper T-cell line produces IL-2 in the presence of IL-1 and PHA [18]. This IL-2 activity was detected by using a CTLD cell line which proliferates in the presence of IL-2 [19]. Ovarian tissue conditioned media (10% v/v) were added to 1A-5 cells (5 x 10⁴ cells/well, 100 µl) in 96 microwell plates. Recombinant IL-1 was used as positive control for the bioassay. Twenty-four hours after incubation, the plates were frozen and thawed three times and CTLD cells (5 x 10³ cells/well, 20 µl) were added to the plates. Recombinant IL-2 was used as a positive control for the bioassay: 48-72 h later, cell proliferation was determined by addition of 1 mg/ml of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma), followed 3 h later by isopropanol in 0.04 N hydrochloric acid, at room temperature. Optical density of cleaved MTT molecules (brown color) was measured by spectrophotometry. Results were expressed at the net absorbance values at 450 nm and 630 nm [20].

IL-2 bioassay.

CTLD cells (5 x 10³ cells/well, 100 µl) were added to conditioned media (10% v/v) of ovarian tissue in 96 microwell plates. CTLD cell proliferation was detected as described for the IL-1 bioassay.

Each experiment (for each patient) was repeated at least three times with similar patterns of results. The results shown are from a single IL-1 assay and represent the mean ± SEM of triplicates. In our previous studies we have shown (unpublished data) that LPS did not affect the proliferation capacity of either 1A-5 or CTLD cells.

RESULTS

IL-1 expression in normal and cancerous ovarian tissues.

Normal ovarian tissue expressed low levels of IL-1alpha, mainly in epithelial cells and in some cases in endothelial cells of blood vessels, but not in the ovarian stroma (Figure 1A and Table 1). In contrast, high levels of IL-1alpha were expressed by epithelial tumour cells in all 10 cases with cancer and in 8 of these cases high levels of IL-1alpha were also expressed in endothelial cells (Figure 1B, C and table 1). IL-1alpha was expressed not only in the cytoplasm of tumor cells by also, in some cases, on the cytoplasmic membrane (Figure 1B). Normal ovarian tissue stained weakly when polyclonal rabbit anti-human IL-1ß antibodies (primary antibodies) were applied (Figure 2A and Table 1), while strong staining was observed in all neoplastic ovaries examined, especially in neoplastic epithelial cells. Endothelial cells also expressed IL-1ß (Figure 2B and Table 1). Over 80% of cancerous ovarian cells expressed high levels of IL-1alpha and IL-1ß. The number of immune cells, in normal and cancerous ovarian tissue, was within normal limits. Fibroblasts and the rest of the stroma stained only very weakly for both IL-1 types.

IL-1 secretion by normal and cancerous ovarian tissues.

Conditioned media (CM) from unstimulated cancerous ovarian tissue contained IL-1-like activity. This activity was increased in CM of LPS-stimulated tissues in a dose-dependent manner (Figure 3). On the contrary, IL-1 could not be detected in CM of unstimulated, normal ovarian tissues. Stimulation of these tissues with high doses of LPS (10 and 100 µg/ml) was followed by secretion of IL-1. Levels of secreted IL-1 were significantly lower in normal than in cancerous ovarian tissue with all LPS concentrations. Kinetics of IL-1 secretion by unstimulated normal ovarian tissue did not show any IL-1-like activity, even after 96 h of incubation (Figure 4A) while stimulation with high doses of LPS (10 and 100 µg/ml) over a long incubation period (72-96 h) resulted in secretion of low levels of IL-1. On the other hand, cancerous ovarian tissue secreted IL-1 constitutively even after 48 h of incubation, and stimulation with LPS not only increased the secretion of IL-1 but also induced earlier secretion (24 h) (Figure 4B). As shown in Figure 3 and Figure 4, the levels of IL-1-like activity identified in normal ovarian tissue were lower than in cancerous tissue even after stimulation by LPS. All conditioned media (CM) were also assessed for IL-2 activity by IL-2 bioassay in order to exclude the presence of residual IL-2 in the CM of ovarian tissues. We could not detect any significant IL-2 activity in these CM.

In order to identify the types of IL-1 secreted by ovarian tissue, conditioned media of normal and cancerous ovarian tissue were tested for IL-1alpha and IL-1ß by using ELISA kits specific for these cytokines. As depicted in Figure 5A, unstimulated normal ovarian tissue secreted constitutively neither IL-1alpha nor IL-1ß. This confirms the results shown in Figure 3 and 4 where biological activity of IL-1 could not be detected. After stimulation with LPS, these tissue secreted similar amounts of both IL-1alpha and ß in a dose-dependent manner. Unstimulated cancerous ovarian tissue secreted both IL-1alpha and ß, and this activity increased after LPS stimulation in a dose-dependent manner (Figure 5B). The levels of IL-1ß were very much higher than those of IL-1alpha in these CM. The amounts of IL-1alpha and IL-1ß secreted by cancerous ovarian tissue were very much higher than those of normal tissues.

IL-1 was not detected in CM of unstimulated normal ovarian tissue (Figures 3, 4). In order to assess whether normal ovarian tissue express IL-1 but does not secrete it into the CM, we assessed IL-1 activity in the homogenate of this tissue. As shown in Figure 6, optimal IL-1-like activity was detected in the homogenate of normal ovarian tissue 72-96 h after incubation with or without LPS (10 µg/ml). IL-1 activity in tissue homogenates was increased after stimulation with LPS in a dose-dependent manner and in a time-course manner. The majority of this activity was neutralized using polyclonal rabbit anti-human IL-1alpha antibodies in a bioassay system. IL-1 activity was also examined in the lysates of cancerous ovarian tissue, and was significantly higher than that in normal ovarian tissue (data not shown). Evaluation of IL-1 levels was performed by bioassay in order to assess IL-1 both quantitatively and qualitatively.

Established primary normal and cancerous epithelial ovarian cultures were examined for their capacity to express and secrete IL-1. IL-1 activity was optimal after 72 hours of culture. As shown in Figure 7, normal ovarian cells did not secrete IL-1 constitutively or following induction with LPS. However, lysates of these cells exhibited very low levels of IL-1 activity and this activity increased under LPS stimulation. On the other hand, cancerous ovarian cells secreted IL-1 constitutively and IL-1 levels were increased under LPS stimulation. IL-1 activity in the lysates of carcinoma cells was higher than in the supernatants. The levels of IL-1 in the lysates of ovarian carcinoma cells was very much higher than in lysates of epithelial cells from normal ovaries. The type of IL-1 in lysates was mainly IL-1alpha and in the supernatants IL-1ß. These results demonstrate that normal ovarian cells express IL-1 but do not secrete it, while cancerous ovarian cells both express and secrete high levels of IL-1. LPS stimulation increased the capacity of normal ovarian cells to express IL-1, but did not induce its secretion. In cancerous ovarian cells, LPS increased the capacity for both IL-1 expression and secretion.

DISCUSSION

Our results demonstrate that cancerous ovarian tissue expresses high levels of both IL-1alpha and IL-1ß which are produced mainly by the neoplastic cells. In contrast, normal ovarian tissue expresses low levels of both IL-1alpha and IL-1ß, and have the potential to secrete biologically active IL-1 in vitro. A large difference in the expression and secretion levels for both IL-1alpha and IL-1ß was demonstrated between normal and cancerous ovarian tissues. Cancerous ovarian tissue expressed high levels of IL-1alpha and IL-1ß in vivo and secrete high levels of IL-1 in vitro in a constitutive manner. Increased IL-1 secretion was observed in ovarian tissue after in vitro stimulation with LPS. Most of the IL-1 secreted by cancerous ovarian tissue was of the IL-1ß type. On the other hand, normal ovarian tissue secreted IL-1 only after stimulation with high doses of LPS and after a prolonged incubation period. We also demonstrated that both, IL-1alpha and IL-1ß were produced in similar low amounts by normal ovarian tissue. Even though IL-1 is not constitutively secreted by normal ovarian tissue, we detected IL-1 in the homogenate of these tissue, indicating that normal ovarian tissue is capable of producing IL-1 but does not secrete it. We suggest that epithelial cells in cancerous ovarian tissue are the main cellular component responsible for IL-1 production under in vitro conditions. In contrast to currently accepted views, we have shown for the first time by immunohistochemical staining that neoplastic epithelial cells in cancerous ovarian tissue are an important source of IL-1 compared with other cells such as infiltrated immune cells or connective tissue. Furthermore, we have also demonstrated, for the first time to the best of our knowledge, that, under in vitro conditions, the production/secretion of IL-1alpha and IL-1ß are regulated differently in normal and in cancerous ovarian tissue and that LPS is not involved in the mechanism of IL-1 secretion by normal epithelial cells.

Pro-inflammatory cytokines may mediate tumor growth by serving as growth factors or angiogenesis-promoting factors. Cytokines of tumor cell origin may serve as autocrine growth factors or stimulate neighboring stromal or immune cells to produce such growth factors. On the other hand, cytokines generated by malignant cells may recruit immune cells to the tumor site and elicit an anti-tumor immune response which may decrease growth rate or even cause tumor regression. There is ample evidence for the involvement of cytokines in tumor growth or regression, and our demonstration that tumor cells themselves contribute significantly to IL-1 production, adds a new aspect to the regulatory function of this substance. IL-1, has pleiotropic effects on proliferation, differentiation and activation of diverse anti-tumor effector cells [4]. It has also been demonstrated that IL-1 exerts direct cytotoxic effects on tumor cells [4, 5]. IL-1 can stimulate the growth of human astrocytoma cell lines, malignant trophoblastic cells and also ovarian tumor cells [21-23]. Constitutive expression and secretion of IL-1 by ovarian carcinoma tissue may indicate the involvement of IL-1 in the pathophysiology of ovarian carcinoma. It is possible that tumor cells could use IL-1 as an autocrine factor for cellular proliferation or other intracellular functions, or as a membrane-associated factor for activation of microsurrounding immunity or chemotaxis of immune cells. Indeed, recently Wu et al. [21] have shown that IL-1 induces ovarian cell proliferation. Also, the binding of IL-1alpha to nuclear DNA, an effect of antisense IL-1alpha on endothelial cell growth and large amounts of constitutive intracellular IL-1ra in the same cells expressing IL-1alpha, have been demonstrated [24, 25]. IL-1 may affect angiogenesis directly or indirectly by inducing the production of angiogenic factors such as PDGF and TNF. IL-1 may induce the expression of adhesive molecules on endothelial cells of blood vessels at the tumor site as well as upregulating surface receptors of tumor cells [4]. On the other hand, IL-1 may function as an inhibitory factor in the immune response [26, 27]. There are conflicting results as far as the levels of IL-1ß in cancerous tissue and normal ovarian tissue are concerned. In line with our results, Punnonen et al. [28] observed that benign and malignant ovarian tumors produce IL-1ß. In contrast to our results, these authors did not find difference in the levels of IL-1ß produced by malignant or benign tissues. Our data indicate significant differences in the levels of both IL-1alpha and IL-1ß secreted by normal and cancerous tissues. Our immunohistochemical results and results involving established primary cell lines show that both, IL-1alpha and IL-1ß are mainly produced by tumor cells in cancerous ovarian tissue. This is in accordance with other studies that have demonstrated production of IL-1 by various tumor cell lines such as squamous cell carcinoma [29, 30].

CONCLUSION

IL-1 may play a physiological role in the ovary; the constitutive levels of IL-1 secreted by cancerous ovarian tissue may indicate its involvement in the pathophysiology of ovarian cancer. This cytokine may function as a paracrine/autocrine growth factor. The production of both types of IL-1 by malignant ovarian tissue does not only differ quantitatively, but apparently with regard to the mechanisms and signals required for their induction.

Acknowledgement.

Supported by a grant from the Israel Cancer Association.

REFERENCES

1. Servo S F, Scully R E, Sobin L H. 1973. International histologic classification of ovarian tumors. 9. Geneva : WHO.

2. Rao B R, Slotman B J. 1991. Endocrine factors in common epithelial ovarian cancer. (Review) Endocr. Rev. 12: 14.

3. Adashi E Y. 1990. The potential relevance of cytokines to ovarian physiology: the emerging role of resident ovarian cells of the white blood cel series. (Review) Endocr. Rev. 11: 454.

4. Dinarello C A. 1996. Biologic basis for interleukin-1 in diseases. Blood 87: 2095.

5. Akira S, Hirano T, Taga T, Kishimoto T. 1990. Biology of multifunctional cytokines: IL-6 and related molecules (IL-1 and TNF). (Review) FASEB J. 4: 2860.

6. Hurwitz A, Loukides J, Ricciarell E, Botero L, Katz E, McAllister J M, Garcia J E, Rohan R, Adashi E Y, Hernandez E R. 1992. Human intraovarian interleukin-1 (IL-1) system: highly compartmentalized and hormonally dependent regulation of the genes encoding IL-1, its receptor, and its receptor antagonist. J. Clin. Invest. 89: 1746.

7. Takakura K, Taii S, Fukuoka M, Yasuda K, Tagaya Y, Yodoi J, Mori T. 1989. Interleukin-2 receptor/p55(Tac)-inducing activity in porcine follicular fluids. Endocrinology 125: 618.

8. Gottschall P E, Uehara A, Hoffmann S T, Arimura A. 1987. Interleukin-1 inhibits follicle stimulating hormone-induced differentiation in rat granulosa cells in vitro. Biochem. Biophys. Res. Commun. 149: 502.

9. Fukuoka M, Yasuda K, Taii S, Takakura K, Mori T. 1989. Interleukin-1 stimulates growth and inhibits progesterone secretion in cultures of porcine granulosa cells. Endocrinology 124: 884.

10. Haskill S, Becker S, Fowler W, Walton L. 1982. Mononuclear-cell infiltration in ovarian cancer. I. Inflammatory-cell infiltrates from tumor and ascites material. Br. J. Cancer 45: 728.

11. Kutteh S, Baker V V, Doellgast G J. 1990. Autologous antibodies eluted from membrane fragments in human ovarian epithelial nepoplastic effusion. III. Cytotoxic potential and characteristics of antigen(s). Am. J. Obstet. Gynecol. 163: 1301.

12. Moradi M M, Carson L F, Weinberg B, Haney A F, Twiggs L B, Ramakrishnan S. 1993. Serum and ascitic fluid levels of interleukin-1, interleukin-6, and tumor necrosis factor-alpha in patients with ovarian epithelial cancer. Cancer. 72: 2433.

13. Kutteh W H, Kutteh C C. 1992. Quantification of tumor necrosis factor-alpha, interleukin-1ß, and interleukin-6 in the effusion of ovarian epithelial neoplasm. Am. J. Obstet. Gynecol. 167: 1864.

14. Arbeit J M. 1990. Molecules, cancer, and the surgeon. A review of molecular biology and its implications for surgical oncology. (Review) Ann. Surg. 212: 3.

15. Benharroch D, Prinsloo I, Apte R N, Yermiahu T, Geffen D B, Yanai-Inbar I, Gopas J. 1996. Interleukin-1 and tumor necrosis factor-alpha in the Reed-Sternberg cells of Hodgkin's disease. Correlation with clinical and morphological "inflammatory" features. Eur. Cytokine Netw. 7: 51.

16. Gelin J, Andersson C, Lundholm K. 1991. Effects of indomethacin, cytokines, and cyclosporin A on tumor growth and the subsequent development of cancer cachexia. Cancer Res. 51: 880.

17. Tacha D E, Mckinney L. 1992. Casein reduced non-specific background staining in immunolabeling techniques. J. Histotechnol. 15: 127.

18. Gillis S, Mizel S B. 1981. T-cell lymphoma model for the analysis of interleukin-1 mediated T-cell activation. Proc. Natl. Acad. Sci. 78: 1133.

19. Gillis S M, Ferm W O, Smith K A. 1978. T-cell growth factor: parameters of production and quantitative microassay for activity. J. Immunol. 20: 2027.

20. Mosman T. 1983. Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays. J. Immunol. Meth. 65: 55.

21. Wu S, Rodabaugh K, Martinez-Maza O, Watson J M, Silberstein D S, Boyer C M, Peters W P, Weinberg J B, Berek J S, Bast R C Jr. 1992. Stimulation of ovarian tumor cell proliferation with monocyte products including interleukin-1, interleukin-6 and tumor necrosis factor-alpha. Am. J. Obstet. Gynecol. 166: 997.

22. Lachman L B, Brown D C, Dinarello C A. 1987. Growth-promoting effect of recombinant interleukin-1 and tumor necrosis factor for human astrocytoma cell line. J. Immunol. 138: 2913.

23. Berkowitz R S, Hill J A, Kurtz C B, Anderson D J. 1988. Effects of products of activated leukocytes (lymphokines and monokines) on the growth of malignant trophoblast cells in vitro. Am. J. Obstet. Gynecol. 158: 199.

24. Haskill S, Martin G, van Le L, Morris J, Peace A, Bigler C F, Jaffe G J, Hammerberg C, Sporn S A, Fong S. 1991. cDNA cloning of an intracellular form of the human interleukin-1 receptor antagonist associated with epithelium. Proc. Natl. Acad. Sci. USA 88: 3681.

25. Maier J A, Voulalas P, Roeder D, Maciag T. 1990. Extension of the life-span of human endothelial cells by an interleukin-1alpha antisense oligomer. Science 249: 1570.

26. Holtmann H, Wallach D. 1987. Down regulation of the receptors for tumor necrosis factor by interleukin-1 and 4beta-phorbol-12 -myristate-13-acetate. J. Immunol. 139: 1161.

27. Ye K, Koch K C, Clark B D, Dinarello C A. 1992. Interleukin-1 down-regulates gene and surface expression of interleukin-1 receptor type I by destabilizing its mRNA whereas interleukin-2 increases its expression. Immunology 75: 427.

28. Punnonen J, Heinonen P K, Kuoppala T, Jansen C T, Punnonen R. 1991. Production of interleukin-1ß and tumor necrosis factor-alpha in patients with benign or malignant ovarian tumors. J. Cancer Res. Clin. Oncol. 117: 587.

29. Watson J M, Sensinataffar J L, Berek J S, Martinez-Maza O. 1990. Constitutive production of interleukin-6 by ovarian cancer cell lines and by primary ovarian tumor cultures. Cancer Res. 50: 6959.

30. Li B Y, Mohanraj D, Olson M C, Moradi M, Twiggs L, Carson L F, Ramakrishnan S. 1992. Human ovarian epithelial cancer cells cultures in vitro express both interleukin-1alpha and ß genes. Cancer Res. 52: 2248.