Induction of an angiogenic phenotype in endometriotic stromal cell ...

2 downloads 0 Views 323KB Size Report
herald sites of endometriotic implants and morphometric interleukin-1 .... Nuclear, Boston, MA, USA), washed in 5% trichloroacetic acid, forthe VEGF cDNA ...
Molecular Human Reproduction vol.6 no.3 pp. 269–275, 2000

Induction of an angiogenic phenotype in endometriotic stromal cell cultures by interleukin-1β

Dan I.Lebovic1, Frauke Bentzien1, Victor A.Chao1, Evelyn N.Garrett1, Y.Gloria Meng2 and Robert N.Taylor1,3 1Center

for Reproductive Sciences, Department of Obstetrics, Gynecology and Reproductive Sciences, University of California, HSE 1689, Box 0556, San Francisco, CA 94143–0556, and 2Division of Immunology, Genentech Inc, South San Francisco, CA 94080–4990, USA 3To

whom correspondence should be addressed

Activated peritoneal macrophages are associated with endometriosis and may play a central role in its aetiology by releasing interleukin-1β (IL-1β) in response to refluxed endometrium. Pari passu with the establishment of endometriotic implants is the development of a vascular supply. In this study we investigated the angiogenic properties of two endometrial proteins, vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6), and assessed their production in response to IL-1β stimulation in human stromal cells isolated from normal endometrium (NE) and endometriotic lesions (EI). Proliferation of bovine brain capillary endothelial cells (BBCE) with a [3H]-thymidine incorporation assay was observed when VEGF (2.1 ⍨ 0.2-fold; P < 0.05) or VEGF and IL-6 (1.8 ⍨ 0.1-fold; P < 0.05) were added in vitro, relative to saline-treated control cultures. Northern blot analysis showed induction of VEGF mRNA (2.6-fold; P < 0.05) and IL-6 mRNA (6.3fold; P < 0.05) transcripts in EI cells, but not NE cells, exposed to IL-1β. A similar induction was seen with VEGF and IL-6 protein secretion in the responsive EI cells. Reverse transcription–polymerase chain reaction (RT–PCR) for the IL-1 receptor type I (IL-1 RI) indicated that the differential effects of IL-1β on NE and EI cells was associated with 2.4 ⍨ 0.1-fold more receptor mRNA in EI versus NE cells. We propose that the ability of IL-1β to activate an angiogenic phenotype in EI stromal cells but not in NE cells, is mediated by the IL-1 RI. Key words: angiogenesis factor/cytokines/endometriosis/interleukin-1/neovascularization

Introduction Retrograde menstruation is an almost universal phenomenon (Halme et al., 1984) and with each cycle, some endometrial tissue is sloughed into the peritoneal cavity. Endometrial cells that enter the peritoneum have two potential fates. They are either completely phagocytosed or they adhere to the peritoneal surface and develop into endometriotic implants (Halme et al., 1984; Bartosik and Jacobs, 1986; Olive and Henderson, 1987). Macrophages, the primary peritoneal mononuclear phagocytes, infiltrate ectopic endometrial implants and coexist within these lesions (Klein et al., 1994). Moreover, the concentration of activated macrophages and their products are increased in the peritoneal fluid of women with endometriosis (Halme et al., 1983; Haney et al., 1983). We (Ryan and Taylor, 1997) and others have proposed that macrophage-derived cytokines, such as interleukin (IL)-1β, participate in an integrated inflammatory cascade that facilitates implantation and growth of ectopic endometrial cells in some women. Hyperaemic peritoneal surfaces observed during laparoscopic investigation often herald sites of endometriotic implants and morphometric analyses have shown neovascularization around and within endometriotic implants (Nisolle et al., 1993). Drawing from the analogy of tumour metastasis, it was proposed that a developing endometriotic implant must acquire a new blood supply to achieve a volume of ⬎2–3 mm3 (Folkman, 1995). Thus, the factors responsible for the acquisition of an © European Society of Human Reproduction and Embryology

angiogenic phenotype are likely to be critical for the establishment, invasion and progression of endometriotic lesions (Oosterlynck et al., 1993; Taylor et al., 1997). Vascular endothelial growth factor (VEGF) is a mitogen for vascular endothelial cells and one of the most potent permeability factors (Keck, 1989; Leung et al., 1989; Charnock-Jones et al., 1993). We reported the production of VEGF in endometriotic lesions and observed a positive correlation between the severity of endometriosis and the concentration of VEGF in peritoneal fluid (McLaren et al., 1996b; Shifren et al., 1996). Interleukin-6 (IL-6) is a multifunctional cytokine produced by various cell types including endometrial stromal cells (Tseng et al., 1996; Van der Molen and Gu, 1996). In the mouse, IL-6 was found to be angiogenic in the uterus during embryonic implantation and in the ovary during folliculogenesis (Motro et al., 1990). In other studies, IL-6 induces endothelial cell motility (Rosen et al., 1991) and is directly mitogenic to endothelial cells (Giraudo et al., 1996). IL-1β is a pleiotropic cytokine involved in the nascent inflammatory immune response. This cytokine is the dominant interleukin-1 secreted by activated peritoneal macrophages (Chensue et al., 1989) and its concentrations are elevated in peritoneal fluid of women with endometriosis (Fakih et al., 1987; Mori et al., 1992). Since IL-1β can induce VEGF and IL-6 production in other cell types (Jones et al., 1993; Ben-Av et al., 1995; Li et al., 1995; Chung et al., 1997; Levitas et al., 269

D.I.Lebovic et al.

1997), we investigated whether IL-1β could induce VEGF and IL-6 expression in isolated primary endometriotic cells. If so, this would suggest a central role for co-ordinated, macrophagederived cytokine-induced angiogenesis in the pathogenesis of endometriosis.

Materials and methods Sources of tissues Tissue specimens were obtained from patients undergoing laparoscopy or laparotomy after providing written informed consent under a study protocol approved by the Committee on Human Research, University of California, San Francisco, USA. Healthy ovulatory women, who had not received hormones or gonadotrophin-releasing hormone (GnRH) agonist therapy for at least 3 months before surgery, were recruited. Women with endometriosis (mean ⫾ SD age, 35 ⫾ 5 years; n ⫽ 4) were staged intra-operatively according to a modification of the revised American Society for Reproductive Medicine (ASRM, 1985) classification. Control subjects were women with subserosal leiomyomata or without pelvic pathology requesting tubal ligation (age, 35 ⫾ 5 years; n ⫽ 4). Endometrial and endometrioma biopsies were collected under sterile conditions and transported to the laboratory on ice in α-minimal essential medium (MEM) with 10% fetal bovine serum (FBS). All samples and cycle stages were estimated histologically (Noyes et al., 1950) and considered to be endometriotic lesions when epithelium and stroma were seen. All normal endometrial biopsies were in phase and consistent with the patient’s menstrual dating. Typically, biopsies from endometriotic lesions showed flattened epithelium and compact stroma. Endothelial cell cultures We chose the well-established bovine brain capillary endothelial (BBCE) cells as a model for peritoneal capillary angiogenesis in endometriosis. Two independent assays were selected to evaluate different functional characteristics of the angiogenic activities of VEGF and IL-6. To assess endothelial cell proliferation, the classical [3H]-thymidine incorporation assay was used (Ferrara et al., 1991). The BBCE cells were grown in low glucose Dulbecco’s modified Eagle’s medium (DMEM), 10% calf serum, glutamine (2 mmol/l), and antibiotics (100 IU/ml penicillin/streptomycin and 2.5 µg/ml fungizone). The BBCE cells were plated at 3000 cells/cm2 plate in 0.5 ml incubation medium (low glucose DMEM with 10% calf serum, glutamine and antibiotics). 2 nmol/l VEGF (provided by N.Ferrara, Genentech, South San Francisco, CA, USA) alone or together with an IL-6 dose-range of 0.15 pmol/l to 1.0 nmol/l, were added to the dishes every other day for 4 days. On day 4, the cells were incubated for 4 h with 0.6 µCi [3H]-thymidine (20.1 Ci/mmol; New England Nuclear, Boston, MA, USA), washed in 5% trichloroacetic acid, solubilized with 0.25 N NaOH, and counted as previously described (Ferrara et al., 1991). Endothelial cell migration The invasive property of endothelial cells in response to VEGF and IL-6 was assessed using a modified collagen I invasion assay. BBCE cells were plated (40 000/cm2) on 300 µl type I collagen gels (Vitrogen 100; Collagen Corporation, Palo Alto, CA, USA). Gels were prepared by adding 8 volumes of type I collagen solution with 1 volume of 10⫻ phosphate-buffered saline (PBS) and 1 volume of 0.1 N NaOH, dispensed into tissue culture wells and allowed to gel at room temperature for 1 h followed by addition of BBCE cells in 250 µl incubation medium (low glucose DMEM with 10% calf serum,

270

glutamine and antibiotics) with or without PBS, 2 nmol/l VEGF or 1.5 pmol/l IL-6. The tissue culture wells were then placed in 37°C for 48 h after which the conditioned media were washed off with PBS and the gels were fixed in situ with 4% paraformaldehyde. The gels were then embedded in paraffin, sectioned at 5 µm thickness and stained with haematoxylin and eosin. Randomly selected equidistant fields of the BBCE cell monolayers were identified by conventional light microscopy. Human endometrial cell cultures Normal endometrial (NE, n ⫽ 4) tissue was obtained by Pipelle® (Cooper Surgical, Shelton, CT, USA) biopsies from women without pelvic pathology. Biopsies of ovarian endometriotic implants (EI, n ⫽ 3) were obtained from consenting patients, as we have reported previously (Ryan et al., 1994). Histological dating (Noyes et al., 1950) of the biopsy specimens was used to confirm that these were collected in the proliferative phase of the endometrial cycle. Primary endometrial cell cultures were prepared from biopsies as described previously (Ryan et al., 1994). Glandular epithelial cells were separated from stromal cells and debris by filtration through narrow gauge sieves. Stromal cells were subcultured to eliminate contamination by macrophages or other leukocytes. Extensive characterization of cell cultures prepared using this protocol confirmed that they were ⬎95% pure and retained functional markers of their endometrial origin in vitro (Ryan et al., 1994). Interleukin-1β stimulation Cultures of NE and EI stromal cells were plated in 10 cm culture dishes (Becton Dickinson, Lincoln Park, NJ, USA) and allowed to grow to confluence in 10% FBS-supplemented media. Prior to the addition of cytokine, the medium was changed to a low serum medium (MEM-α supplemented with 2.5% FBS, nucleosides and non-essential amino acids). Some cultures were treated with recombinant human IL-1β (10 ng/ml ⫽ 0.6 nmol/l; Sigma Chemical Co, St Louis, MO, USA). Conditioned media were removed and analysed after 4, 8, 12, and 24 h. Pilot experiments showed that 85% of maximal cytokine accumulation was reached after 12 h. The 2.5% FBS-supplemented MEM-α used for the experiments was tested for IL-1β, VEGF and IL-6 concentrations and all were below the limit of detection for the respective enzyme-linked immunosorbent assay (ELISA). Preparation of total RNA and Northern analysis Total RNA was extracted from cell cultures using the TRIzol reagent kit (Gibco BRL, Gaithersburg, MD, USA). Total RNA (10 µg) was subjected to electrophoresis and blotted by capillary transfer onto a nylon membrane (Schleicher and Schuell, Keene, NH, USA). The membrane was hybridized with a [32P]-labelled VEGF complementary DNA (cDNA) probe and a [32P]-labelled IL-6 cDNA probe synthesized by random primer extension (Clontech, Palo Alto, CA). The template for the VEGF cDNA probe is a 921 bp fragment of VEGF corresponding to the protein-coding region between nucleotides 335–1256 (Leung et al., 1989). The template for the IL-6 cDNA probe is a 159 bp fragment of IL-6 corresponding to the protein-coding region between nucleotides 551–709 (May et al., 1986). The integrity and relative amount of RNA loaded into each lane were confirmed using a [32P]-labelled glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA 240 bp probe as a constitutively expressed marker. Data were analysed as ratios of the density of the hybridization signals of VEGF or IL-6 to GAPDH mRNA, as determined by a phosphorimager (Storm-Moleculer Dynamics, Menlo Park, CA, USA). To determine the dependency of VEGF expression on initiation of RNA synthesis rather than post-transcriptional RNA stabilization, cells were pre-treated with the transcriptional inhibitor, actinomycin

Angiogenic phenotype in endometriotic cells

D (Sigma Chemical Co; 5 µg/ml), for 1 h prior to stimulation with IL-1β (0.6 nmol/l). VEGF enzyme-linked immunosorbent assay The ELISA plates were coated with 2.5 µg/ml monoclonal antibody to VEGF (mAb 3.5F8) in 50 mmol/l carbonate buffer, pH 9.6, at 4°C overnight and blocked with 0.5% BSA in PBS. All monoclonal antibodies were prepared and specificity characterized as previously described (Kim et al., 1992). Standards (0.03–2 ng/ml recombinant VEGF165) and 3-fold serially diluted samples in PBS containing 0.5% BSA, 0.05% polysorbate 20, 0.25% 3-[(3-cholamidopropyl) dimethylammonio]-1-propanesulphonate, 0.2% bovine γ-globulins (Sigma Chemical Co), 5 mmol/l ethylenediamine tetraacetate, and 0.35 mol/l NaCl were incubated on the plate for 2 h. Bound VEGF was detected using biotinylated monoclonal antibody to VEGF (mAb 4.6.1), followed by streptavidin peroxidase (Sigma) and 3,3⬘5,5⬘tetramethyl benzidine (Kirkegaard and Perry Laboratories, Gaithersburg, MD, USA) as the substrate. Absorbance was read at 450 nm on a Vmax plate reader (Molecular Devices, Menlo Park, CA, USA). The standard curve was fitted using a four parameter non-linear regression curve-fitting program (developed at Genentech). Data points that fell in the linear range of the standard curve were used for calculating the VEGF concentration in samples. The assay was linear for VEGF in conditioned medium and was sensitive to 0.03 ng/ml. Interleukin-6 enzyme-linked immunosorbent assay Specific sandwich ELISA for IL-6 was performed on the conditioned medium using a commercial kit (Quantikine; R&D Systems, Minneapolis, MN, USA). In our laboratory the assay was linear in conditioned medium samples and sensitive to 0.7 pg/ml, with intraand inter-assay coefficients of variation of 3.1 and 2.7% respectively. The assay is specific for human IL-6, has no known cross-reactivity with other cytokines and is not interfered with by the presence of soluble IL-6 receptors (R&D Systems, 1997 Immunoassay Catalogue). Aliquots of culture supernatants were each tested in duplicate at several dilutions and compared to reference standards of human recombinant IL-6 (R&D Systems). Reverse transcription–polymerase chain reaction (RT–PCR) To compare expression of IL-1 receptor type I (IL-1 RI) in NE and EI cells, RT–PCR was performed using primers derived from the human IL-1 RI sequence. The sense primer began at base 1373 (5⬘ position) and was 25 bases in length. The antisense primer began at base 1604 (3⬘ position) and was 22 bases in length. These amplifed a 253 bp PCR product. Complementary DNA (cDNA) was reverse transcribed from total RNA obtained from NE or EI cells. Ten-fold dilutions of cDNA representing 25 ng to 25 fg of total RNA were amplified. The cDNA was subjected to 30 cycles of PCR amplification consisting of 40 s at 95°C, 20 s at 60°C and 70 s at 75°C. The resulting PCR products were visualized on a 4% agarose gel stained with ethidium bromide. Data were analysed as ratios of IL-1 RI to GAPDH, as determined by computer-assisted densitometry (NIH Image 1.54, Springfield, VA, USA). Statistical analysis All experiments were repeated a minimum of three times and analysed by unpaired or paired t-tests as appropriate. Results are presented as mean ⫾ SEM. Significant differences were accepted when two-tailed analyses yielded P ⬍ 0.05 (Glantz, 1992).

Figure 1. Endothelial cell proliferation. The [3H]-thymidine incorporation assay for endothelial cell proliferation showed a significant mitogenic response to vascular endothelial growth factor (VEGF) (2 nmol/l, P ⬍ 0.05) IL-6 (1.5 pmol/l) had minimal and non-significant stimulatory effects over phosphate-buffered saline (PBS) control. IL-6 also showed no additive effects with VEGF. The number of samples analysed in each group is indicated as the number at the base of each histogram. Asterisks indicate treatments that significantly differ (P ⬍ 0.05) from PBS-treated controls by paired t-tests with Bonferroni corrections.

Results In-vitro assays of functional angiogenic activity Pilot dose-response experiments were used to establish the optimal concentrations of the various angiogenic factors, and then these were performed in triplicate in multiple experiments. As depicted in Figure 1, VEGF (2 nmol/l) stimulated [3H]thymidine incorporation 2.1 ⫾ 0.2-fold over saline treated controls (P ⬍ 0.05) and IL-6 (1.5 pmol/l) together with VEGF stimulated [3H]-thymidine incorporation 1.8 ⫾ 0.1-fold over saline treated controls (P ⬍ 0.05). VEGF induced migration of the BBCE cells (Figure 2B) relative to the saline control (Figure 2A). Alone, IL-6 had no significant mitogenic effect on BBCE cells but, as shown in Figure 2C, did influence the migration of the endothelial cells as seen by the invasion into the collagen matrix. Cytokine gene expression Confluent cultures of four independent EI and three independent NE stromal cell preparations were evaluated for the expression of angiogenic gene products. Total RNA prepared from cell lysates was separated on agarose gels. Northern hybridization was used to identify and quantify VEGF and IL-6 mRNA transcripts (Figure 3). As reported previously (Shifren et al., 1996), VEGF mRNA transcripts migrated as 2–3 differentially-spliced transcripts of 4.2–4.4kb. The IL-6 probe detected a single transcript of 1.3 kb, in agreement with other reports (May et al., 1986). This probe also showed low level cross-hybridization with 28S and 18S rRNA bands on the total RNA blots, but these could be easily distinguished from the bona fide IL-6 mRNA signal. As an internal control for RNA quantity and integrity, the blots were reprobed to quantify mRNA representing the constitutive GAPDH gene 271

D.I.Lebovic et al.

Figure 2. Modified collagen I invasion assay. In addition to its mitogenic effect on endothelial cells (EC), vascular endothelial growth factor (VEGF) caused invasive behaviour of ECs plated on collagen matrix (panel B) interleukin-6 (IL-6) also stimulated EC bud formation and invasion into a collagen plug (panel C) compared with phosphate-buffered saline (PBS) control (panel A) Haematoxylin and eosin staining; scale bar ⫽ 10 µm.

Figure 4. Phosphorimaging quantification of Northern analyses for vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6) mRNA. The left panel represents VEGF mRNA for normal endometrial stromal cells (NE) and endometriotic stromal cells (EI) without and with interleukin-1β (IL-1β) stimulation. The right panel represents IL-6 mRNA in NE and EI cells without and with IL-1β stimulation. While the IL-1β has no significant effect on NE cells, the EI cells showed significant increases in VEGF and IL-6 mRNA in response to IL-1β (P ⬍ 0.05) The number of samples analysed in each group is indicated at the base of each histogram. Asterisks indicate treatments that significantly differ (P ⬍ 0.05) from untreated controls by paired t-tests.

Figure 5. Actinomycin D effect on vascular endothelial growth factor (VEGF) mRNA. Northern analysis demonstrating inhibition of VEGF mRNA synthesis in endometriotic stromal cells (EI) cells treated with interleukin-1β (IL-1β) (0.6 nmol/l, lanes 3 and 4) in the presence (lanes 2 and 4) or absence (lanes 1 and 3) of actinomycin D (Act-D, 5 µg/ml)

Figure 3. Representative Northern blot analyses demonstrating the induction of vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6) mRNA in interleukin-1β (IL-1β)-treated (0.6 nmol/l) cultured human endometrial stromal cells (NE) or endometriotic stromal cells (EI) VEGF transcripts of 4.2 to 4.4-kb were detected. IL-6 transcripts of 1.3 kb also were detected. The position of 28S and 18S ribosomal RNA on the original gels are marked. The integrity and amount of total RNA loaded were confirmed by subsequent hybridization of the blot with a glyceraldehyde-3-phosphate dehydrogenase (GAPDH) probe (1.2 kb).

272

(1.2 kb), which was used to normalize the phosphorimaging data (Figure 4). Relative to untreated EI cells, incubation for 12 h in the presence of IL-1β (0.6 nmol/l) resulted in a 2.6fold increase in VEGF mRNA and 6.3-fold induction of IL-6 mRNA (P ⬍ 0.01; paired t-test). By contrast, exposure to IL-1β had no significant effect on NE stromal cell VEGF mRNA (0.8-fold) or IL-6 mRNA (1.9-fold) induction (P 艌 0.10). IL-1β induction of VEGF mRNA was blocked by pretreatment with actinomycin D (Figure 5), suggesting that IL-1β signals new VEGF mRNA synthesis rather than a posttranscriptional regulation. Cytokine protein expression The increased secretion of immunoreactive VEGF and IL-6 proteins from EI cell cultures exposed to IL-1β (0.6 nmol/l) was confirmed by ELISA. The ratios of IL-1β-treated to

Angiogenic phenotype in endometriotic cells

Table I. Vascular endothelial growth factor (VEGF) and interleukin-6 (IL-6) induction by IL-1β in stromal cells. Values are presented as mean ⫾ SEM of cells from three patients in each group

VEGF (pg/ml) IL-6 (pg/ml)

Without IL-1β Mean ⫾ SEM

With IL-1β Mean ⫾ SEMa

Increase (-fold) Mean ⫾ SEM

162 ⫾ 93 2990 ⫾ 672

671 ⫾ 470 12 700 ⫾ 4222

3.8 ⫾ 0.7* 4.6 ⫾ 1.2*

a0.6

nmol/l for 12 h. *P ⬍ 0.05.

Figure 6. Reverse transcription–polymerase chain reaction (RT–PCR) for interleukin-1 receptor type I (IL-1 RI) mRNA. Primers were taken from the human IL-1 RI sequence which amplified a 253 bp PCR product. Insert shows representative semiquantitative RT–PCR for IL-1 RI at 10-fold dilutions in normal endometrial stromal cells (NE) and endometriotic stromal cells (EI) cells. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control for genomic DNA contamination, RNA integrity and normalization of the IL-1 RI data. The amount of IL-1 RI mRNA was quantified using NIH Image software and the results are presented in the bar graph showing a significantly greater amount (2.4 ⫾ 0.1-fold) of IL-1 RI in EI versus NE cells (P ⬍ 0.05) by unpaired t-test.

untreated cells showed statistically significant increases for VEGF and IL-6 secretion (P ⬍ 0.05; Table I). Analysis of IL-1 receptor type I mRNA expression To understand the enhanced production of two independent angiogenic factors in EI cells in response to stimulation with IL-1β, we sought to quantify the predominant signalling receptor (IL-1 RI) for this ligand. Repeated attempts to analyse this transcript by Northern blotting were unsuccessful, owing to the known low quantity of IL-1 RI under the conditions of our assays (Simo´ n et al., 1993). Semi-quantitative analyses using RT–PCR were established using the amplification of GAPDH cDNA as an internal control (Figure 6). These results indicated that EI cells contain ~2.5-fold more IL-1 RI mRNA than NE cells analysed under identical conditions.

Discussion Previous studies from our laboratory (Ryan et al., 1995; Shifren et al., 1996) prompted us to postulate that the establishment and progression of human ectopic endometrial implants are dependent upon their neovascularization. While NE cells have

the capacity to synthesize and secrete angiogenic molecules (e.g. basic FGF, VEGF, IL-6, IL-8), enhanced production of angiogenic factors by cells derived from endometriotic lesions might further increase the likelihood of implant establishment and proliferation. We proposed that the immune system participates in this angiogenic response via macrophage-derived cytokines, such as IL-1β, which stimulate the stromal cell production of angiogenic molecules. In the current study, we demonstrated that the synthesis of at least two such factors, VEGF and IL-6, are preferentially induced in endometriotic stromal cells by IL-1β. Peritoneal macrophages from women with endometriosis are increased in number, concentration, and degree of morphological activation compared to normal women (Halme et al., 1983; Haney et al., 1983; Hill and Anderson, 1989). The peritoneal recruitment and activation of these macrophages appear to be mediated by the expression of specific monocyte chemoattractants and growth factors. Pelvic fluid concentrations of two of these, Regulated upon Activation, Normal T cell Expressed and Secreted (RANTES) (Khorram et al., 1993) and monocyte chemotactic protein (MCP)-1 (Akoum et al., 1995; Arici et al., 1997), are increased in women with endometriosis compared with normal, matched controls. Moreover, endometriotic implants and isolated cells derived from the lesions express RANTES mRNA and protein (Hornung et al., 1997). The immunostaining intensity of a third cytokine, granulocyte-macrophage colony stimulating factor-1 (GMCSF-1), was reported to be elevated in endometriotic implants during the secretory phase (Sharpe-Timms et al., 1994). Once macrophages have been mobilized into the peritoneum and infiltrate endometriotic lesions, they are postulated to produce cytokines, growth factors, and complement (Halme et al., 1988; Isaacson et al., 1990; Olive et al., 1991; McLaren et al., 1996a; Ryan and Taylor, 1997) that stimulate endometriotic implant growth by paracrine pathways. A pivotal macrophage-derived cytokine is IL-1β, which plays a key role in primary immune responses (Sunderkotter et al., 1994). Some studies showed elevated peritoneal fluid levels in stage I and II endometriosis cases (Fakih et al., 1987; Mori et al., 1992; Taketani et al., 1992), while others found no differences from control women (Awadalla et al., 1987) or failed to detect IL-1 at all (Koyama et al., 1993). Fakih et al. (Fakih et al., 1987) found that IL-1 was secreted from peritoneal macrophages collected from patients with endometriosis, but not by peritoneal macrophages isolated from normal controls. Our experiments were designed to test the hypothesis that IL-1β could activate an angiogenic phenotype in endometriotic stromal cells in vitro. While the overall number of patients in each group was small, appropriate unpaired or paired statistics were applied. We have shown that both epithelial and stromal cells are important sources of VEGF. However, our recent studies showing vectorial VEGF secretion by epithelial cells toward the lumen (Hornung et al., 1998) suggested that stromal cell VEGF production may be most important for angiogenesis due to their proximity to developing vascular networks. The results of our study show that incubating human EI stromal cells for 12 h in the presence of IL-1β significantly increased expression of VEGF and IL-6 mRNA. Since the 273

D.I.Lebovic et al.

induction of VEGF was blocked by actinomycin D we conclude that an increase in transcription rate rather than a posttranscriptional mechanism explains the VEGF mRNA induction by IL-1β in stromal cells. This is in contrast to the significant post-transcriptional effect on VEGF mRNA accumulation in hypoxic rat pheochromocytoma cells (Levy et al., 1996). High affinity IL-1 receptors have been detected on NE stromal cells (Simo´ n et al., 1993), although the concentrations in the latter are reportedly lower than those detected in NE epithelial cells (Tabibzadeh et al., 1990; Simo´ n et al., 1993). Our data confirm that EI stromal cells are targets for IL-1β, with the resultant stimulation of VEGF and IL-6 mRNA and protein. The enhanced EI cell production of VEGF and IL-6 in response to IL-1β suggests an augmented sensitivity of these stromal cells to the biological actions of IL-1β. Moreover, the results support an immunological link between activated macrophages and increased vascularity seen in endometriotic lesions. Our results confirmed a mitogenic and migratory role for VEGF in BBCE cells. While IL-6 did not induce proliferation it did stimulate bud formation and invasion of endothelial cells into a collagen plug. In summary, the present study supports a central role of activated macrophages and their secretory products in the progression of endometriotic implants and offers an immunological explanation for the neovascularization that surrounds these lesions. We propose that IL-1β is a prototypic example of an activated peritoneal macrophage secretory product in endometriosis. Under the conditions of our primary cell model, IL-1β appears to have the capability of efficiently activating an angiogenic phenotype in endometriotic stromal cells, but not in NE cells. Analyses of IL-1 RI mRNA suggest that the enhanced sensitivity of EI stromal cells, relative to NE cells, is conferred, in part, via increased IL-1β ligand binding and signalling. Functional studies of the receptor protein will be necessary to confirm this hypothesis. It remains to be determined whether the apparent up-regulation of the IL-1 RI gene is a result of the ectopic location of endometriotic implants in a cytokine-rich environment or if the establishment of lesions reflects the successful selection of highly cytokine-sensitive endometrial cells displaced into the peritoneal cavity by retrograde menstruation. Elucidation of this ‘chicken-or-egg’ issue is a challenge for future investigations of the cellular pathophysiology of endometriosis.

Acknowledgements The authors thank Drs. Eldon Schriock, Carolyn Givens and Isabelle Ryan from the Department of Obstetrics, Gynecology and Reproductive Sciences, UCSF for their generous clinical contributions to the study, Mari Matli, from the Department of Surgery, UCSF for her helpful technical suggestions and Shy Tassa and Lisa Caris, from Genentech, Inc. for performing the VEGF ELISA. This work was supported by the following grants and fellowships: NIH grants HD08517–01 (DIL), HD33283 (RNT) and HD37321, through the Specialized Cooperative Centers Program in Reproduction Research.

References Akoum, A., Lemay, A. and C., B. (1995) Cytokine-induced secretion of monocyte chemotactic protein-1 by human endometriotic cells in culture. Am. J. Obstet. Gynecol., 172, 594–600.

274

American Society of Reproductive Medicine (1985) Revised American Fertility Society classification of endometriosis: 1985. Fertil. Steril., 43, 1–2. Arici, A., Oral, E., Attar, E. et al. (1997) Monocyte chemotactic protein-1 concentration in peritoneal fluid of women with endometriosis and its modulation of expression in mesothelial cells. Fertil. Steril., 67, 1065–1072. Awadalla, S., Friedman, C., Haq, A. et al. (1987) Local peritoneal factors: their role in infertility associated with endometriosis. Am J. Obstet Gynecol, 157, 1207–1214. Bartosik, D. and Jacobs, S.L. (1986) Endometrial tissue in peritoneal fluid. Fertil. Steril., 46, 796–800. Ben-Av, P., Crofford, L., Wilder, L. et al. (1995) Induction of vascular endothelial growth factor in synovial fibroblasts by prostaglandin E and interleukin-1: a potential mechanism for inflammatory angiogenesis. FEBS Letts., 372, 83–87. Charnock-Jones, D., Sharkey, A., Rajput-Williams, J. et al. (1993) Identification and localization of alternately spliced mRNA’s for vascular endothelial growth factor in human uterus and estrogen regulation in endometrial carcinoma cell lines. Biol. Reprod., 48, 1120–1128. Chensue, S.W., Shmyr-Forsch, C., Otterness, I.G. et al. (1989) The beta form is the dominant interleukin 1 released by murine peritoneal macrophages. Biochem. Biophys Res. Comm., 160, 404–408. Chung, K.W., Ando, M. and Adashi, E.Y. (1997) Interleukin (IL)-1 dependent regulation of IL-6 in the immature rat ovary: A specific receptor-mediated eicosanoid-dependent effect. J. Soc. Gynecol. Invest., 4, 87A. Fakih, H., Baggett, B., Holtz, G. et al. (1987) Interleukin-1: a possible role in the infertility associated with endometriosis. Fertil. Steril., 47, 213–217. Ferrara, N., Winer, J. and Burton, T. (1991) Aortic smooth muscle cells express and secrete vascular endothelial growth factor. Growth Factors, 5, 141–148. Folkman, J. (1995) Angiogenesis in cancer, vascular, rheumatoid and other disease. Nature Med., 1, 27–31. Giraudo, E., Arese, M., Toniatti, C. et al. (1996) IL-6 is an in vitro and in vivo autocrine growth factor for middle T antigen-transformed endothelial cells. J. Immunol., 157, 2618–2623. Glantz, S. A. (1992) Primer of Biostatistics. McGraw-Hill, New York, USA. Halme, J., Becker, S., Hammond, M. et al. (1983) Increased activation of pelvic macrophages in infertile women with mild endometriosis. Am. J. Obstet. Gynecol., 145, 333–337. Halme, J., Hammond, M., Hulka, J. et al. (1984) Retrograde menstruation in healthy women and in patients with endometriosis. Obstet. Gynecol., 64, 151–154. Halme, J., White, C., Kauma, S. et al. (1988) Peritoneal macrophages from patients with endometriosis release growth factor activity in vitro. J. Clin. Endocrinol. Metab., 66, 1044–1049. Haney, A.F., Muscato, J.J. and Weinberg, J.B. (1983) Peritoneal fluid cell populations in infertility patients. Fertil. Steril., 35, 696–698. Hill, J. and Anderson, D. (1989) Lymphocyte activity in the presence of peritoneal fluid from fertile women and infertile women with and without endometriosis. Am. J. Obstet. Gynecol., 161, 861–864. Hornung, D., Ryan, I.P., Chao, V.A. et al. (1997) Immunolocalization and regulation of the chemokine RANTES in human endometrial and endometriosis tissues and cells. J. Clin. Endocrinol. Metab., 82, 1621–1628. Hornung, D., Lebovic, D.I., Shifren, J.L. et al. (1998) Vectorial secretion of vascular endothelial growth factor by polarized human endometrial epithelial cells. Fertil. Steril., 69, 909–915. Isaacson, K.B., Galman, M., Coutifaris, C. et al. (1990) Endometrial synthesis and secretion of complement component-3 by patients with and without endometriosis. Fertil. Steril., 53, 836–841. Jones, T.H., Kennedy, R.L., Justice, S.K. et al. (1993) Interleukin-1 stimulates the release of interleukin-6 from cultured human pituitary adenoma cells. Acta Endocrinol. (Copenh.), 128, 405–410. Keck, P.J. (1989) Vascular permeability factor, an endothelial cell mitogen related to PDGF. Science, 246, 1309–1312. Khorram, O., Taylor, R., Ryan, I. et al. (1993) Peritoneal fluid concentrations of the cytokine RANTES correlate with the severity of endometriosis. Am. J. Obstet. Gynecol., 169, 1545–1549. Kim, K.J., Li, B., Houck, K. et al. (1992) The vascular endothelial growth factor proteins: identification of biologically relevant regions by neutralizing monoclonal antibodies. Growth Factors, 6, 53–64. Klein, N.A., Pergola, G.M., Tekmal, R.R. et al. (1994) Cytokine regulation of cellular proliferation in endometriosis. Ann. N.Y. Acad. Sci., 734, 322–332. Koyama, N., Matsuura, K. and Okamura, H. (1993) Cytokines in the peritoneal fluid of patients with endometriosis. Int. J. Gynecol. Obstet., 43, 45–50.

Angiogenic phenotype in endometriotic cells Leung, D.W., Cachianes, G., Kuang, W.J. et al. (1989) Vascular endothelial growth factor is a secreted angiogenic mitogen. Science, 246, 1306–1309. Levitas, E., Chamoun, D., Udoff, L.C. et al. (1997) Periovulatory and interleukin (IL)-1-dependent upregulation of intraovarian vascular endothelial growth factor (VEGF): Potential role for VEGF in the promotion of periovulatory angiogenesis and vascular permeability. J. Soc. Gynecol. Invest., 4, 87A. Levy, A.P., Levy, N.S. and Goldberg, M.A. (1996) Post-transcriptional regulation of vascular endothelial growth factor by hypoxia. J. Biol. Chem., 271, 2746–2753. Li, J., Perrella, M.A., Tsai, J.-C. et al. (1995) Induction of vascular endothelial growth factor gene expression by interleukin-1 beta in rat aortic smooth muscle cells. J. Biol. Chem., 270, 308–312. May, L.T., Helfgott, D.C. and Sehgal, P.B. (1986) Anti-beta-interferon antibodies inhibit the increased expression of HLA-B7 mRNA in tumor necrosis factor-treated human fibroblasts: Structural studies of the beta-2 interferon involved. Proc. Natl Acad. Sci. USA, 83, 8957–8961. McLaren, J., Prentice, A., Charnock-Jones, D.S. et al. (1996a) Vascular endothelial growth factor is produced by peritoneal fluid macrophages in endometriosis and is regulated by ovarian steroids. J. Clin. Invest., 98, 482–489. McLaren, J., Prentice, A., Charnock-Jones, D.S. et al. (1996b) Vascular endothelial growth factor (VEGF) concentrations are elevated in peritoneal fluid of women with endometriosis. Hum. Reprod., 11, 220–223. Mori, H., Sawairi, M., Nakagawa, M. et al. (1992) Expression of interleukin1 (IL-1) beta messenger ribonucleic acid (mRNA) and IL-1 receptor antagonist mRNA in peritoneal macrophages from patients with endometriosis. Fertil. Steril., 57, 535–542. Motro, B., Itin, A., Sachs, L. et al. (1990) Pattern of interleukin 6 gene expression in vivo suggests a role for this cytokine in angiogenesis. Proc. Natl Acad. Sci. USA, 87, 3092–3096. Nisolle, M., Casanas-Roux, F., Anaf, V. et al. (1993) Morphometric study of the stromal vascularization in peritoneal endometriosis. Fertil. Steril., 59, 681–684. Noyes, R.W., Hertig, A.T. and Rock, J. (1950) Dating the endometrial biopsy. Fertil. Steril., 1, 3–25. Olive, D. and Henderson, D.Y. (1987) Endometriosis and mullerian anomalies. Obstet. Gynecol., 69, 412–415. Olive, D.L., Montoya, I., Riehl, R.M. et al. (1991) Macrophage-conditioned media enhance endometrial stromal cell proliferation in vitro. Am. J. Obstet. Gynecol., 164, 953–958. Oosterlynck, D., Meuleman, C., Sobis, H. et al. (1993) Angiogenic activity of peritoneal fluid from women with endometriosis. Fertil. Steril., 59, 778–782. Rosen, E.M., Liu, D., Setter, E. et al. (1991) Interleukin-6 stimulates motility of vascular endothelium. EXS, 59, 195–205. Ryan, I.P. and Taylor, R.N. (1997) Endometriosis and infertility: New concepts. Obstet. Gynecol. Surv., 52, 365–371. Ryan, I., Schriock, E. and Taylor, R. (1994) Isolation, characterization, and comparison of human endometrial and endometriosis cells in vitro. J. Clin. Endocrinol. Metab., 78, 642–649. Ryan, I., Tseng, J., Schriock, E. et al. (1995) Interleukin (IL)-8 concentrations are elevated in peritoneal fluid of women with endometriosis. Fertil. Steril., 63, 929–932. Sharpe-Timms, K.L., Bruno, P.L., Penney, L.L. et al. (1994) Immunohistochemical localization of granulocyte-macrophage colony-stimulating factor in matched endometriosis and endometrial tissues. Am. J. Obstet. Gynecol., 171, 740–745. Shifren, J.L., Tseng, J.F., Zaloudek, C.J. et al. (1996) Ovarian steroid regulation of vascular endothelial growth factor in the human endometrium: implications for angiogenesis during the menstrual cycle and in the pathogenesis of endometriosis. J. Clin. Endocrinol. Metab., 81, 3112–3118. Simo´ n, C., Piquette, G. N., Frances, A. et al. (1993) Localization of interleukin1 type I receptor and interleukin-1 beta in human endometrium throughout the menstrual cycle. J. Clin. Endocrinol. Metab., 77, 549–555. Sunderkotter, C., Steinbrink, K., Goebeler, M. et al. (1994) Macrophages and angiogenesis. J. Leukocyte Biol., 55, 410–422. Tabibzadeh, S., Kaffka, K.L., Satyaswaroop, P.G. et al. (1990) Interleukin-1 (IL-1) regulation of human endometrial function: Presence of Il-1 receptor correlates with IL-1-stimulated prostaglandin E2 production. J. Clin. Endocrinol. Metab., 70, 1000–1006. Taketani, Y., Kuo, T. and Mizuno, M. (1992) Comparison of cytokine levels and embryo toxicity in peritoneal fluid in infertile women with untreated or treated endometriosis. Am. J. Obstet. Gynecol., 167, 265–270.

Taylor, R.N., Ryan, I.P., Moore, E.S. et al. (1997) Angiogenesis and macrophage activation in endometriosis. Ann. N.Y. Acad. Sci., 828, 194–207. Tseng, J.F., Ryan, I.P., Milam, T.D. et al. (1996) Interleukin-6 secretion in vitro is up-regulated in ectopic and eutopic endometrial stromal cells from women with endometriosis. J. Clin. Endocrinol. Metab., 81, 1118–1122. Van der Molen, D.T. and Gu, Y. (1996) Human endometrial interleukin-6 (IL-6): in vivo messenger ribonucleic acid expression, in vitro protein production, and stimulation thereof by IL-1beta. Fertil. Steril., 66, 741–747. Received on October 8, 1999; accepted on December 10, 1999

275