Soluble urokinase-type plasminogen activator receptor is over ...

Molecular Human Reproduction vol.3 no.12 pp. 1101–1105, 1997

Soluble urokinase-type plasminogen activator receptor is overexpressed in uterine endometrium from women with endometriosis M.Sillem1, S.Prifti, B.Monga, P.Buvari, U.Shamia and B.Runnebaum Division of Gynaecological Endocrinology and Reproductive Medicine, Department of Obstetrics and Gynaecology, Ruprecht-Karls-Universita¨t, Voss-Straβe 9, D-69115 Heidelberg, Germany 1To

whom correspondence should be addressed

Extracellular matrix degradation by secreted proteases, e.g. plasmin, is essential for endometrial functions such as blastocyst implantation and menstruation. We investigated whether the expression of plasmin(ogen) activating or inhibiting factors in endometrial cells from women with endometriosis was different from women without the disease. Endometrial biopsies were obtained from 10 patients with and 16 women without endometriosis. Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM)/F12 supplemented with diethylstilboestrol (10–10 M) alone or combined with promegestone (5 H 10–8 or 5 H 10–6 M). Urokinase plasminogen activator (uPA), plasminogen activator inhibitor (PAI)-1 and -2, and soluble uPA receptor (suPAR) concentrations were assayed by enzyme-linked immunosorbent assay (ELISA) in the conditioned media. uPA and PAI-2 concentrations were not influenced by steroid treatment and did not differ between women with and without endometriosis, whereas PAI-1 was significantly up-regulated by promegestone in both groups. In contrast, suPA-R expression was not influenced by steroid treatment but was significantly higher in cells from endometriosis patients. This is the first report on suPA-R secretion in endometrial cells and the results indicate an altered activation of plasmin(ogen) in endometrium from women with endometriosis that could lead to a higher proteolytic potential of retrogradely menstruated endometrial fragments with consecutive development of endometriotic foci. Key words: endometriosis/extracellular matrix/invasion/in-vitro model/plasminogen

Introduction An altered interaction between the extracellular matrix of the peritoneal lining and retrogradely menstruated endometrial fragments appears to be one explanation for the fact that some women develop endometriosis and others do not (Evers, 1996). Recently, experimental conditions that lead to an altered adhesiveness (van der Linden et al., 1996) or invasiveness (Wild et al., 1994) of endometrial cells in vitro have been described. In a previous study, we were able to demonstrate that complete enzymatic digestion as well as proteinase inhibition leads to significant impairment of ectopic implantation in an animal model of endometriosis (Sillem et al., 1996). In this context, two families of secreted proteases, serine proteinases [essentially the plasmin(ogen) cascade] and matrix metalloproteinases are of particular interest. While secreted proteinases are important for extracellular matrix turnover, a second set of endometrial proteinases, i.e. cell surface peptidases, seem to play a role in lymphocyte homing and degradation of growth factors and cytokines (Imai et al., 1996). Plasminogen, an ubiquitous protein secreted by the liver, is activated to the highly potent protease plasmin by two types of activators, namely, tissue type plasminogen activator (tPA) and urokinase type plasminogen activator (uPA), which binds to a cell surface receptor (uPA-R). While the activation by tPA is important for fibrinolysis, uPA triggers localized pericellular proteolysis of the extracellular matrix. The system © European Society for Human Reproduction and Embryology

is tightly controlled by potent inhibitors, at the level of plasminogen activators by plasminogen activator inhibitor type 1 (PAI-1) and plasminogen activator inhibitor type 2 (PAI-2), and at the level of plasmin by α2 macroglobulin and α2 antiplasmin. The uPA pathway plays an important role in uterine physiology in general (Littlefield, 1991) and in the initiation of menstruation in particular (Tabibzadeh, 1996). Accordingly, a regulation by steroids and local paracrine factors has been reported (Cassle`n et al., 1992; Schatz and Lockwood, 1993). Urokinase PA can be inactivated by progesterone in endometrial stromal cell cultures through increased expression of PAI-1 and cell surface uPA-R (Cassle`n et al., 1995). One physiological function of the system is the activation of matrix metalloproteinases. While this family of secreted proteases has many synergistic functions regarding implantation and initiation of menstruation (Hulboy et al., 1997), its role in the pathogenesis of endometriosis is the subject of ongoing studies in our group. Apart from in physiological functions, the plasmin(ogen) activator/inhibitor system seems to be involved in the pericellular proteolytic processes occurring during tumour cell migration and metastasis (Mignatti and Rifkin, 1993). Plasminogen and single chain urokinase have been detected by immunohistochemistry in higher concentrations in endometriosis biopsies compared to corresponding eutopic endometrium (Ferna´ndezShaw et al., 1995). Recently, a soluble form of uPA-R has been described, 1101

M.Sillem et al.

which is encoded by the same gene as the surface receptor. The binding domain for uPA is similar, but the carboxyterminal end by which the surface receptor is anchored to the cell membrane is modified through alternative splicing, which suggests a retained binding activity (Pyke et al., 1993; Mizukami et al., 1995). In this series of experiments, we examined the ability of endometrial cells to secrete suPA-R in vitro, and we attempted to substantiate the hypothesis that uterine endometrium from women with endometriosis shows an altered regulation of plasminogen activation, thus facilitating ectopic implantation of endometrial fragments after retrograde menstruation in endometriosis.

Materials and methods Subjects Endometrial biopsies were obtained from normally cycling women with (n 5 10) or without (n 5 16) endometriosis during diagnostic laparoscopy using a flexible sampling device (Pipelle de Cornier, Prodimed, Neully-en-Thelle, France). In all, 20 patients were in the late proliferative phase and six (three with and three without endometriosis) were in secretory phase. None had received steroid or gonadotrophin-releasing hormone (GnRH) analogue treatment in the 3 months preceding the procedure. Normal cycles were assumed when a cycle length of 25–32 days was reported by the patient and the biopsies were dated according to the last menstrual period. Of the patients with endometriosis, three had mild disease, five had moderate disease and two had severe disease as judged from reviewing the surgery reports. Of the patients without endometriosis, four had post-inflammatory pathology, eight had male factor or unexplained infertility, three had functional cysts and one had a tubal ligation. Patients with suspected endometrial pathology were excluded. All subjects gave informed consent and the study was approved by the Ethical Committee of the University of Heidelberg Faculty of Medicine. Reagents All media, as well as antibiotic antimycotic solution (AAS), gentamycin, non-essential amino acids (NEAA), insulin–transferrin–selenite (ITS), Phenol Red, and trypsin–EDTA, were from Sigma Chemical Co. (Deisenhofen, Germany). Glutamax was obtained from Carl Roth GmbH (Karlsruhe, Germany), and fetal bovine serum (FBS) was from CCPro, Neustadt, Germany. Diethylstilboestrol (DES) was purchased from Sigma, and promegestone (R5020) from NEN Life Science Products (Cologne, Germany). For uPA, PAI-1, PAI-2, and suPA-R, commercially available enzyme-linked immunosorbent assay (ELISA) kits were used (American Diagnostica, Greenwich, CT, USA). Cell culture All endometrial samples were collected in Dulbecco’s modified Eagle’s medium (DMEM)/F12 supplemented with AAS and gentamycin at 4°C and processed within 24 h. After mincing the tissue with scalpels, trypsin digestion was performed until a single cell suspension was obtained. An aliquot was drawn and counted using the Trypan Blue (Sigma, Deisenhofen, Germany) exclusion method in a Neubauer haemocytometer. Cells were then seeded in 6-well plates at a standard density of 2.53105 per well and cultivated at 37°C and 5% CO2 in DMEM/F12, containing Phenol Red and supplemented with 10% FBS. Upon confluence, which

1102

was reached after 36–48 h, cells were switched to Phenol Redfree, serum-free medium, supplemented with ITS, NEAA, and glutamax. At this time, about one third of cells showed positive immunocytochemical staining with the monoclonal antibody Bw 495/36, indicating epithelial endometrial cells (Kruitwagen et al., 1991) and the remainder showed positive staining for vimentin, indicating stromal cells (Moll et al., 1982) (results not shown). The relationship between the epithelial and stromal cells did not change perceptibly during the treatment period and in preceding experiments, the viability of cells was found to be .90% in the Trypan Blue exclusion test after 6 days in serum-free medium. Hormone treatment Upon confluence, hormone treatment was initiated. As a negative control, one well was treated with vehicle alone (ethanol at 0.1% v/v). A second well was treated with DES 10–9 M, and two further wells with DES 10–9 M and R5020 at 5310–8 or 5310–6 M respectively. Media were changed 2, 4, and 6 days after initiation of treatment and stored at –20°C until the assays were performed. Enzyme-linked immunosorbent assays All ELISAs were performed according to the manufacturer’s instructions on conditioned media derived from all cultures and all samples were assayed in duplicate. For uPA, the lower detection limit of the assay was 10 pg/ml. For PAI-1 and PAI-2, the lower detection limits of the assays were 50 pg/ml. For suPA-R, the lower detection limit of the assay was 0.1 ng/ml of total uPA-R. This assay recognizes soluble, native, and recombinant uPA-R as well as uPA-R/uPA and uPA-R/uPA/PAI-1 complexes. All readings were within the linear range of the standard curves after dilutions were performed as required. The intra-assay coefficient of variation was always ,10% and assays were repeated if this criterion was not met. Statistical analysis Statistical analyses were performed using Statxact Turbo Version 2.11, a statistical package for exact non-parametric inference (Cytel Software Corporation, Cambridge, MA, USA). In order to take into account the effect of the cycle phase, the samples were stratified into two independent strata, proliferative and the secretory phases. For comparison between samples from women with and without endometriosis, a stratified Wilcoxon rank sum test was performed, while simultaneously adjusting for the covariate cycle phase. Values were considered to be significant when P ø 0.05.

Results In all experiments performed, we were unable to detect a significant difference between cultures treated with vehicle alone, compared with cultures treated with DES alone. Therefore, comparisons were made between treatment with DES alone and treatment with DES and R5020, in order to have only one variable parameter. The levels of significance did not differ between comparisons that were stratified for cycle phase and those that were not. However, stratified comparisons are quoted below. Urokinase-type plasminogen activator uPA concentration in the conditioned media of cells treated with DES alone for 4 days was 18.5 ng/ml (1.85–60 ng/ml) (median, min–max) in the group without endometriosis and 16 ng/ml (8.45–60 ng/ml) in samples from endometriosis

Soluble urokinase receptor in endometriosis

patients. This difference was not significant. R5020 treatment led to a decrease in uPA concentration that was less pronounced in samples from endometriosis patients. However, neither the decrease nor the difference between groups were significant. Plasminogen activator inhibitor type 1 In conditioned media from cells treated with DES alone for 4 days, a concentration of 20 ng/ml (2–128 ng/ml) was detected in the group of samples without endometriosis. In samples from endometriosis patients, the corresponding finding was 27 ng/ml (5–142 ng/ml). This difference was not significant. In both groups, R5020 treatment at both concentrations led to a significant increase in PAI-1 concentrations (P , 0.001). After DES 1 R5020 (5310–8 M) for 4 days, PAI-1 concentration was 165 ng/ml (46–1113 ng/ml) in the absence and 264 ng/ml (10–468 ng/ml) in the presence of endometriosis, but this difference between both groups, again, was not significant (Figure 1). All values given in this paragraph were not significantly different from corresponding values obtained at the other time points. Plasminogen activator inhibitor type 2 Concentrations of PAI-2 in conditioned media after administration of DES alone for 4 days were 0.5 ng/ml (0.1–4.7 ng/ml) in samples from patients without endometriosis and 0.6 ng/ml (0–5.5 ng/ml) in controls. This difference was not significant. Neither treatment with R5020 nor presence or absence of endometriosis had a significant effect on PAI-2 concentrations, and the same was true for the other time points. Soluble urokinase plasminogen activator receptor Cells from women without endometriosis secreted suPA-R at 1.1 ng/ml (0.3–6.7 ng/ml) after 4 days of DES alone, compared with 3.2 ng/ml (0.8–9.8 ng/ml) by cells from endometriosis patients. This difference was significant (P 5 0.036). R5020 treatment did not influence suPA-R expression significantly. The situation was similar at the other time points with levels of significance ranging from P 5 0.01 to P 5 0.05. Also, the samples from the two groups of women were always significantly different when corresponding R5020 treatments were compared (Figure 2).

Discussion Retrograde menstruation, which plausibly explains many aspects of endometriosis (Jenkins et al., 1986), seems to be a phenomenon that frequently occurs in women with patent tubes, irrespective of the presence of endometriosis (Blumenkrantz et al., 1981; Halme et al., 1984). Therefore, additional factors that increase tubal reflux and/or facilitate ectopic development of viable endometrial fragments that reach the peritoneal cavity can be postulated (Bartosik et al., 1986; Leyendecker et al., 1996). Our approach focuses on factors that physiologically mediate extracellular matrix turnover in uterine endometrium and that might be involved in pericellular proteolysis occurring during ectopic implantation of an endometrial fragment. One such factor is the system that regulates plasmin(ogen) activation.

Figure 1. Plasminogen activator inhibitor-1 (PAI-1) concentrations in endometrial cell culture supernatants from women without (C) and with endometriosis (E) in ng/ml, secreted by 2.53105 cells. Circles indicate readings from individual samples. Cultivated cells were treated with steroids for (a) 2, (b) 4 or (c) 6 days. R5020 treatment led to a significant increase in PAI-1 concentrations (P , 0.001)

The regulation of PAI-1 in endometrial stromal cells by progestin in vitro has been described before (Schatz and Lockwood, 1993). PAI-1, in turn, seems to be the major inhibitor of plasminogen activation in the endometrium (Schatz et al., 1995). Cassle`n et al. (1995) observed a synergistic negative effect of progestin treatment on PA activity, due to increased expression of both PAI-1 and cell surface bound uPA receptor. Complexing of uPA to PAI-1 and binding to uPA-R both seem to facilitate clearance of active plasmin. On the other hand, uPA binding to its receptor greatly increases its activity, e.g. in U-937 cells (Ellis et al., 1991) or in 1103

M.Sillem et al.

Figure 2. Soluble urokinase plasminogen activator receptor (suPA-R) concentrations in endometrial cell culture supernatants from women without (C) and with endometriosis (E) in ng/ml, secreted by 2.53105 cells. Circles indicate readings from individual samples. Cultivated cells were treated with steroids for (a) 2, (b) 4 or (c) 6 days (P , 0.05).

keratinocytes (Kramer et al., 1995). Of the paracrine factors, epidermal growth factor has been shown to be a regulator in endometrial cells in vitro: uPA, tPA and PAI-1 were upregulated and the effect on the latter two was increased by progesterone (Miyauchi et al., 1995). While the plasmin(ogen) activator/inhibitor system has not been investigated explicitly in the uterine endometrium of endometriosis patients before, our results confirm previous reports with respect to uPA, PAI-1 and PAI-2 (Gleeson et al., 1993) secretion and regulation by endometrial cells in vitro. We were unable to demonstrate a difference between cells from women with and without endometriosis in these parameters. Exvivo studies of the plasmin(ogen) activator/inhibitor system in 1104

endometriosis have yielded heterogeneous results. Plasminogen activity in peritoneal fluid was reported to be similar in peritoneal fluid from patients with and without this condition by several authors (Åstedt and Nordenskjo¨ld, 1984; Batzofin et al., 1985). Others reported on a decreased activity (Ohtsuka, 1980) or an increased proteinase inhibiting activity (Fazleabas et al., 1987). To our knowledge, the ability of endometrial cells to secrete the soluble form of uPA receptor has not been previously reported. Soluble uPA receptor can be detected at low levels in plasma from healthy volunteers and at high concentrations in plasma from patients with septicaemia or in pleural effusions and ascites from patients with inflammatory or malignant conditions (Mizukami et al., 1995). Whether soluble uPA-R in fact enhances uPA activity like surface-bound uPA-R or decreases it by binding uPA without consecutive activation, is as yet unclear (Pyke et al., 1993). In vitro, the invasiveness of glioblastoma cells mediated by uPA-R can be inhibited by a specific monoclonal antibody (Mohanam et al., 1993). In vivo, the absence of uPA-R from normal breast tissue as opposed to malignant tissue as well as a correlation between uPA-R expression and prognosis in breast cancer patients has been described (Bianchi et al., 1994; Duggan et al., 1995). However, the sequence homology of the molecule with elapid snake venom toxins (Casey et al., 1994) suggests further, as yet unidentified functions of soluble uPA receptor. Recent findings indicate that soluble urokinase receptor can increase the local availability of uPA by retarding its inhibition by PAI-1 and also its inactivation and clearance (Al Roof Higazi et al., 1996). Additionally, the integrin mediated adhesion of cells to vitronectin, an extracellular matrix component, was found to be inhibited by suPAR (Wei et al., 1996) which would facilitate shedding of cells or tissue fragments. We conclude that our findings have a pathophysiological significance in that an over-expression of suPA-R is likely to lead to the generation of ‘coarser’ endometrial fragments and to the increased availability of uPA for pericellular proteolysis. Both mechanisms would help explaining why endometrial tissue in the peritoneal cavity leads to endometriosis in some but not in the majority of women. In future studies, it may be interesting to study endometriotic implants or cell lines derived from endometriosis implants for soluble urokinase receptor expression (Bouquet de Jolinie`re et al., 1997).

Acknowledgements The authors wish to thank Michael D.Kramer, for critically discussing the experimental design and the manuscript. Financial support was generously provided by Schering Forschungsgesellschaft, Berlin, Germany. The expert technical help of Ms Julia Jauckus is gratefully acknowledged.

References Al Roof Higazi, A., Mazar, A., Wang, J. et al. (1996) Single-chain urokinasetype plasminogen activator bound to its receptor is relatively resistant to plasminogen activator inhibitor type 1. Blood, 87, 3545–3549. Åstedt, B. and Nordenskjo¨ld, F. (1984) Plasminogen activators in endometriosis. Acta Obstet. Gynecol. Scand., 63 (Suppl. 123), 23–24. Bartosik, D., Jacobs, S.L. and Kelly, L.J. (1986) Endometrial tissue in peritoneal fluid. Fertil. Steril., 46, 796–800.

Soluble urokinase receptor in endometriosis Batzofin, J.H., Holmes, S.D., Gibbons, W.E. et al. (1985) Peritoneal fluid plasminogen activator activity in endometriosis and pelvic adhesive disease. Fertil. Steril., 44, 277–279. Bianchi, E., Cohen, R.L., Thor, A.T. et al. (1994) The urokinase receptor is expressed in invasive breast cancer but not in normal breast tissue. Cancer Res., 54, 861–866. Blumenkrantz, M.J., Gallagher, N. and Bahore, R.A. (1981) Retrograde menstruation in women undergoing chronic peritoneal dialysis. Obstet. Gynecol., 57, 667–670. Bouquet de Jolinie`re, J., Validire, P., Canis, M. et al. (1997) Human endometriosis-derived permanent cell line (FbEM-1): establishment and characterization. Hum. Reprod. Update, 3, 117–123. Casey, J.R., Petranka, J.G., Kottra, J. et al. (1994) The structure of the urokinase-type plasminogen activator receptor gene. Blood, 84, 1151–1156. Cassle`n, B., Urano, S., Lecander, I. et al. (1992) Plasminogen activators in the human endometrium, cellular origin and hormonal regulation. Blood Coagul. Fibrinolysis, 3, 133–137. Cassle`n, B., Nordengren, J., Gustavsson, B. et al. (1995) Progesterone stimulates degradation of urokinase plasminogen activator (in endometrial stromal cells by increasing its inhibitor and surface expression of the u-PA receptor. J. Clin. Endocrinol. Metab., 80, 2776–2784. Duggan, C., Maguire, T., McDermott, E. et al. (1995) Urokinase plasminogen activator and urokinase plasminogen activator receptor in breast cancer. Int. J. Cancer, 61, 597–600. Ellis, V., Behrendt, N. and Danø, K. (1991) Plasminogen activation by receptor-bound urokinase. J. Biol. Chem., 266, 12752–12758. Evers, J.L.H. (1996) The defense against endometriosis. Fertil. Steril., 66, 351–353 Fazleabas, A.T., Khan-Dawood, F.S. and Dawood, M.Y. (1987) Protein, progesterone, and protease inhibitors in uterine and peritoneal fluids of women with endometriosis. Fertil. Steril., 47, 218–224. Ferna`ndez-Shaw, S., Marshall, J.M., Hicks, B. et al. (1995) Plasminogen activators in ectopic and uterine endometrium. Fertil. Steril., 63, 45–51. Gleeson, N., Gonsalves, R. and Bonnar, J. (1993) Plasminogen activator inhibitors in endometrial adenocarcinoma. Cancer, 72, 1670–1672. Halme, J., Hammond, M.G. and Hulka, J.F. (1984) Retrograde menstruation in healthy women and in patients with endometriosis. Obstet. Gynecol., 64, 141–154. Hulboy, D.L., Rudolph, L.A. and Matrisian, L.M. (1997) Matrix metalloproteinases as mediators of reproductive function. Mol. Hum. Reprod., 3, 27–45. Imai, K., Kanzaki, H. and Mori, T. (1996) Cell surface peptidases in human endometrium. Mol. Hum. Reprod., 2, 425–431. Jenkins, S., Olive, D.L. and Haney, A.F. (1986) Endometriosis: pathogenetic implications of the anatomic distribution. Obstet. Gynecol., 67, 335–338. Kramer, M.D., Schaefer, B. and Reinartz, J. (1995) Plasminogen activation by human keratinocytes: molecular pathways and cell-biological consequences. Biol. Chem. Hoppe-Seyler, 376, 131–141. Kruitwagen, R.F.P.M., Poels, L.G., Willemsen, W.M.P. et al. (1991) Endometrial epithelial cells in peritoneal fluid during the early follicular phase. Fertil. Steril., 55, 297–303. Leyendecker, G., Kunz, G., Wildt, L. et al. (1996) Uterine hyperperistalsis and dysperistalsis as dysfunctions of the mechanism of rapid sperm transport in patients with endometriosis and infertility. Hum. Reprod., 11, 1542–1551. Littlefield, B.A. (1991) Plasminogen activators in endometrial physiology and embryo implantation. A review. Ann. N.Y. Acad. Sci., 622, 166–175. Mignatti, P. and Rifkin, D.B. (1993) Biology and biochemistry of proteinases in tumor invasion. Physiol. Rev., 73, 161–195. Miyauchi, A., Momoeda, M., Nakabayashi, M. et al. (1995) Regulation of the plasminogen activator/plasmin system by epidermal growth factor in cultured endometrial cells. Hum. Reprod., 10, 3284–3288. Mizukami, I.F., Faulkner, N.E., Gyetko, M.R. et al. (1995) Enzyme-linked immunosorbent assay detection of a soluble form of urokinase plasminogen activator receptor in vivo. Blood, 86, 203–211. Mohanam, S., Sawaya, R., McCutcheon, I. et al. (1993) Modulation of in vitro invasion of human glioblastoma cells by urokinase-type plasminogen activator receptor antibody. Cancer Res., 53, 4143–4147. Moll, R., Franke, W.F., Schiller, D.L., et al. (1982) The catalog of human cytokeratins: patterns of expression in normal epithelia, tumors and cultured cells. Cell, 31, 11–24. Ohtsuka, N. (1980) Study on the pathogenesis of adhesions in endometriosis. Acta Obstet. Gynecol. Jap., 32, 1758–1766. Pyke, C., Eriksen, H., Solberg, B. et al. (1993) An alternatively spliced variant

of mRNA for the human receptor for urokinase plasminogen activator. FEBS Lett., 326, 69–74. Schatz, F. and Lockwood, C.J. (1993) Progestin regulation of plasminogen activator inhibitor type 1 in primary cultures of endometrial stromal and decidual cells. J. Clin. Endocrinol. Metab., 77, 621–625. Schatz, F., Aigner, S., Papp, C. et al. (1995) Plasminogen activator activity during decidualization of human endometrial stromal cells is regulated by plasminogen activator inhibitor 1. J. Clin. Endocrinol. Metab., 80, 2504–2510. Sillem, M., Hahn, U., Coddington, C.C. et al. (1996) Ectopic growth of endometrium depends on its structural integrity and proteolytic activity in the cynomolgus monkey (Macaca fascicularis) model of endometriosis. Fertil. Steril., 66, 468–473. Tabibzadeh, S. (1996) The signals and molecular pathways involved in human menstruation, a unique process of tissue destruction and remodelling. Mol. Hum. Reprod., 2, 77–92. Van der Linden P.J.Q., de Goeij A.F.P.M., Dunselman G.A.J. et al. (1996) Endometrial cell adhesion in an in vitro model using intact amniotic membranes. Fertil. Steril., 65, 76–80. Wei, Y., Lukashev, M., Simon, D. et al. (1996) Regulation of integrin function by the urokinase receptor. Science, 273, 1551–1555. Wild, R.A., Zhang, R. and Medders, D.(1994) Whole endometrial fragments form characteristics of in vivo endometriosis in a mesothelial co-culture system: an in vitro model for the study of the histogenesis of endometriosis. J. Soc. Gynecol. Invest., 1, 65–68. Received on July 21, 1997; accepted on October 9, 1997

1105