Early-stage endometriosis: adhesion and growth of human menstrual endometrium in nude mice

Transcription

1 FERTILITY AND STERILITY VOL. 74, NO. 2, AUGUST 2000 Copyright 2000 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Early-stage endometriosis: adhesion and growth of human menstrual endometrium in nude mice Michelle Nisolle, M.D., Ph.D., Françoise Casanas-Roux, B.S., and Jacques Donnez, M.D., Ph.D. Infertility Department Research, Service de Gynécologie, Université Catholique de Louvain, Cliniques Universitaires Saint-Luc, Brussels, Belgium Objective: To evaluate the implantation of menstrual endometrium and the early stages of evolution of endometriotic lesions. Design: Experimental prospective study. Setting: An academic research environment. Animals: Ten nude mice. Intervention(s): A minilaparotomy was performed to place fresh human menstrual endometrial samples in the peritoneal cavity. Removal of the transplants was performed successively on days 1, 3, and 5 by laparotomy. Main Outcome Measure(s): Adhesion of endometrial fragments and early stages of endometrial lesions was morphologically and immunohistochemically studied. Result(s): As early as day 1, stromal cells were found to be attached to the mesothelium. A progressive reorganization of epithelial and stromal cells into endometrial glands was observed. On day 5, cystic endometriotic lesions were characterized by more extensive proliferative activity in glandular cells and a higher VEGF score in stromal cells than that observed in previously removed transplants. Conclusion(s): Menstrual human endometrium is able to implant on intact mesothelium and to reorganize itself into structured glands and stroma under the influence of unknown factors. We suggest that stromal and glandular cells have two distinct roles: stromal cells are involved in the attachment process and glandular cells in the growth of the endometriotic lesion. (Fertil Steril 2000;74: by American Society for Reproductive Medicine.) Key Words: Menstrual endometrium, attachment, growth, early-stage endometriosis, transplantation Received October 27, 1999; revised and accepted February 2, Supported by a grant from the Fonds de la Recherche Scientifique Médicale (FRSM). Reprint requests: Michelle Nisolle, M.D., Ph.D., Service de Gynécologie, Université Catholique de Louvain, Cliniques Universitaires Saint-Luc, Avenue Hippocrate, 10, 1200 Bruxelles, Belgique (FAX: ; /00/$20.00 PII S (00) The pathogenesis of endometriosis still remains unclear. Retrograde menstruation, peritoneal adhesion of shed endometrial tissue, and outgrowth of these endometrial cells are essential steps in the pathogenesis of endometriosis according to Sampson s implantation theory (1). The mechanism of adherence of endometrial cells to the peritoneal lining is also unknown. Indeed, it is not clear whether endometrial stroma or epithelium can attach to the peritoneum and whether the mesothelium acts as a barrier to the attachment of regurgitated endometrium. Endometrial fragments in the peritoneal fluid and mesothelial cells express adhesion molecules that are possibly involved in cell-cell and cell-extracellular matrix interactions (2 4). Other factors in the peritoneal fluid, such as matrix metalloproteinases, steroids, growth factors, and cytokines, may also be involved in this process. To know if the mesothelium must be damaged to permit the development of peritoneal endometriosis, in vitro models have been developed with use of either amnion membranes (5, 6) or biopsies of human peritoneum (4, 7). These studies proved the adherence process in endometrium, but the conclusions were controversial, the former suggesting that mesothelial damage is not required. We recently strongly suggested that cultured endometrial explants could adhere to intact peritoneum (8). Recently, Koks et al. (9) observed that menstrual endometrial tissue adheres to subepithelial 306

2 structures of amnion and peritoneum in cases of damage to or absence of the epithelial lining. In vivo models of endometriosis showing an implantation rate as high as 87% after several weeks have already been published, but the early stage of endometriosis development was not investigated in these in vivo studies (8, 10 13). The purpose of this study was to evaluate the implantation of menstrual endometrium and the early stages of evolution of the endometriotic lesion, as soon as 1, 3, and 5 days after human menstrual endometrium transplantation, by morphological and immunohistochemical studies of the implant biopsies taken on days 1, 3, and 5. MATERIALS AND METHODS The use of human tissue for this study was approved by the Institutional Review Board of the Université Catholique de Louvain. Endometrium Menstrual endometrium was obtained from three reproductive-aged women (27 44 years) without endometriosis who were undergoing surgery for benign conditions. Surgery was performed during menstruation (day 2 or 3). The tissue samples were handled aseptically, placed in ice-cold sterile phosphatebuffered saline solution, ph7.4, and transported to the laboratory, where the specimens were cut into pieces not exceeding 1 mm 3. One part of the biopsy was fixed in 4% buffered formaldehyde, dehydrated, and embedded in paraffin. Sections were stained either with Gomori s trichrome for histological confirmation of the menstrual phase (14) or with specific antibodies (Ki-67, VEGF) for immunohistochemical studies. TABLE 1 Morphological and immunohistochemical results of the implants removed on days 1, 3, and 5. Day 1 Day 3 Day 5 No. of mice No. of removed transplants Presence of endometriotic glands (%) 4/12 (33) 4/18 (22) 8/13 (61) Proliferative activity Ki-67 (%) (median) Glandular cells a Stromal cells VEGF H-score (mean SD) Glandular cells Stromal cells b a In glandular cells, the proliferation index was significantly higher than that observed on days 1 and 3 (P.02). b In stromal cells, the VEGF-score was significantly higher than that observed on days 1 and 3 (P.001). Transplantation Into Nude Mice The homozygous nude mouse mutant (nu/nu) has congenital thymus aplasia resulting in a deficient T-lymphocyte system (15). Ten nude 8-week-old (Swiss nu/nu) female mice were used for the present study. Three mice were housed per cage under laminar-flow hamster egg penetration assay (HEPA)- filtered hoods in rooms maintained at 28 C with a 12:12- hour light-dark cycle. All housing materials and food and water were autoclaved before use. The mice were fed ad libidum with laboratory chow and acidified water. Handling was done under laminar-flow hoods. The mice were anesthetized with an i.p. injection of 0.07 ml of Imalgène 500 (Rhǒne Merieux, Belgium) and 0.16 ml of Rompun 2%, diluted 100 times vol/vol (Bayer, Belgium). Minilaparotomy (1 cm) was performed on the ventral midline just caudal to the umbilicus to place fresh endometrial samples in the peritoneal cavity (day 0). In each mouse, four to five transplants were deposited i.p. in the lower part of the pelvic cavity or the abdominal cavity, without any suture nor damage to the peritoneal layer. Opening of the peritoneal cavity lasted 2 minutes to decrease the time of exposure to air. The first endometrial biopsy (day 4) was transplanted into mice 1 4, the second (day 3) into mice 5 7, and the third (day 4) into mice After 24 hours, all the mice were reoperated on to remove one or two implants. On day 3, the procedure was repeated and at that time, two mice were killed because all the transplants had been removed. On day 5, the remaining eight mice were killed, after all the transplants had been removed from the peritoneal cavity. During this short experimental period, no hormonal therapy was administered to the mice. After extirpation, tissues were fixed in 4% buffered formaldehyde and embedded in paraffin. The whole sample was cut into serial sections that were either stained with Gomori s trichrome (16) for histologic evaluation or with specific antibodies for immunohistochemical studies. The proliferation index and the vascularization of the grafts were evaluated (17, 18). Measurement of Proliferative Activity Ki 67 labeling was conducted as previously described (17) with use of the peroxidase-antiperoxidase (PAP) complex and according to the immunohistochemical assay described by Immunotech (clone MIB1; Immunotec, Marseille, France). Diaminobenzidine-hydrogen peroxide substrate (Dako, Glostrup, Denmark) was used as the chromogen. The sections were lightly counterstained with hematoxylin. The proliferative activity of the implants was measured by evaluating the number of Ki 67 positive-staining nuclei within the glandular epithelium and the stromal cells with use of the 40 objective of an Axioscope light microscope (Zeiss, FERTILITY & STERILITY 307

3 FIGURE 1 Human endometrial implant on day 1. Human stromal cells are already found to be attached ( ) to the peritoneum. Normal mesothelium is found beneath stromal cells ( ). (Stain, Gomori s trichrome; original magnification, 40). Oberkochen, Germany). The number of counted nuclei ranged from 65 to 3,480. Results were expressed as median value, and one-way analysis of variance (19) was used for statistical analysis. Immunohistochemical Study of VEGF VEGF labeling was performed with use of the immunoperoxidase technique and the polyclonal rabbit anti-vegf (Santa Cruz Biotechnology, Santa Cruz, CA) as the primary antibody, 1/100 (vol/vol) (18). Hematoxylin was used as a light counterstain. Positive and negative controls were included in the series. All slides were evaluated by two observers (F.C., M.N.) on a blind basis. The VEGF semiquantitative score was calculated as follows: H-score Pi, where i is the intensity: from 0 (negative cells) to 3 (high staining intensity); and P is the percentage of stained cells for each given i; P 1( 15% positive-staining cells), P 2 (15% 50% positive-staining cells), P 3 (50% 85% positive-staining cells), P 4( 85% positive-staining cells), and P 5 (100% positive-staining cells). Results were expressed as means SEM and one-way analysis of variance was used for statistical analysis. RESULTS Macroscopic Appearance At the time of grafting, the transplants were red (menstruating endometrium), but within 24 hours, the transplants had attached to the peritoneum or the periintestinal fatty tissue and become completely white in appearance. On days 3 and 5, the transplants were still found to have the same macroscopic aspect. Morphological Study Before and After Transplantation In total, 43 transplants were morphologically evaluated: 12 samples were removed on day 1, 18 on day 3, and 13 on day 5 (Table 1). The endometrial biopsy was morphologically evaluated to be in the menstrual phase according to the criteria of Noyes et al. (14). The stroma appeared degenerative, undergoing necrosis, intermingling with inflammatory cells and red blood cells. Only some fragments of glandular epithelium and/or ruptured, collapsed glands were present. On day 1, attachment was histologically observed between stromal cells of the transplant and the mesothelium in all cases. The transplanted glandular and stromal cells appeared viable, but most of the transplant consisted of stromal and inflammatory cells (Fig. 1). Necrosis was still present in all the endometrial implants with areas of nuclear pyknosis and rhexis and red blood cells. A rapid healing capacity was observed; in 33% of cases, regenerative areas were seen in stromal cells and small restructured fragments of glandular epithelium. On day 3, endometriotic lesions were observed in 4 of the 18 removed transplants, but contrary to what was observed 308 Nisolle et al. Early stage endometriosis Vol. 74, No. 2, August 2000

4 FIGURE 2 Human endometrial implant on day 3. (A), Early invasion of stromal cells in the submesothelial area (between the arrows). (Stain, Gomori s trichrome; original magnification, 25). (B), Typical cylindrical glandular cells are noted (black arrow) in some areas but, obviously, invasion is due to stromal cells. (Stain, Gomori s trichrome; original magnification, 25). on day 1, the lesions appeared histologically to be completely reorganized and structured as typical endometriotic lesions (Fig. 2A and B). Necrosis was still present, but areas of regeneration were clearly visible in the stroma, and a neovascular network was already present at the junction between the transplant and mesothelium. On day 5, typical endometriotic lesions were noted in 8 of the 13 removed transplants from the 8 remaining mice (Fig. 3A and B). Necrosis was still present in some places, but large areas of normal-looking stroma surrounding dilated glands with flattened or cylindrical epithelium were observed. A vascular network was well developed in the area of FERTILITY & STERILITY 309

5 FIGURE 3 Human endometrial implant on day 5. (A), Dilated and nondilated endometriotic glands surrounded by stromal cells are clearly present in the submesothelial fatty tissue. (A), Stain, Gomori s trichrome; original magnification, 10). (B), Numerous wellformed endometriotic glands are noted. (Stain, Gomori s trichrome; original magnification, 40). adhesion between the graft and the peritoneal layer. It was surprising that the morphological aspect after 5 days was similar to that observed after a 3-week transplantation, as described in a previous study (8), where nude mice were grafted with proliferative or secretory endometrium. Proliferative Activity The proliferation index was negative in both the glandular and stromal cells of two menstrual endometrial biopsies. However, in the third endometrial biopsy, Ki-67 positivestained nuclei were noted in the stromal cells and less 310 Nisolle et al. Early stage endometriosis Vol. 74, No. 2, August 2000

6 frequently in epithelial cells. The proliferation index was 4.0% in endometrial stromal cells and 2.3% in epithelial cells. After 24 hours, the proliferation index (median) was 0% in the glandular cells and 0.47% in the stromal cells of the implant. On day 3, the proliferation index was 9.68% in the glandular cells and 0.23% in the stromal cells. On day 5, significantly greater (P.02) glandular proliferative activity (22.68%) was observed compared with that observed in implants removed after 1 or 3 days. The proliferation index in stromal cells was similar (0.21%) to that observed in implants removed on days 1 or 3. VEGF Expression In menstrual endometrial biopsies, VEGF positive-stained cells were observed in glandular epithelium and stroma. In glandular cells, the mean value of the VEGF H-score was on day 1, on day 3, and on day 5. No statistical change was observed in these VEGF scores according to the day of removal. In the stroma, VEGF positive-stained cells were present in 60% of cases on day 1. On days 3 and 5, VEGF positivestained cells were present in all cases (weak staining on day 3 and more pronounced on day 5). The VEGF scores remained low on days 1 and 3 ( and , respectively) and did not differ significantly from values observed in menstrual endometrial biopsies (4.0 0). On day 5, a significantly higher (P.001) VEGF H-score ( ) was noted compared with that observed in menstrual biopsies and in grafts removed on days 1 and 3. On days 3 and 5, in the area of adhesion developed between the graft and the mouse peritoneal wall, some capillaries of the neovascular network exhibited some VEGFpositive immunostained endothelial cells. DISCUSSION Several theories relating to the pathogenesis of endometriosis have been proposed; the most widely accepted is Sampson s theory (1). He postulated that retrograde menstrual flow transported desquamated endometrial cells through the Fallopian tubes into the peritoneal cavity and that the remaining viable cells subsequently implanted and grew. The mechanism of adherence of regurgitated endometrial cells to the peritoneum is still unclear, as is the speed at which the endometrial lesion evolves. Nude mice represent a well-known model of experimental endometriosis. Indeed, several investigators described the transplantation of proliferative or secretory human endometrium with an implantation rate varying from 33% to 87% (9, 10 13). However, the role of T-lymphocyte response on the proliferation of endometriotic implants has, therefore, not been taken into account. In the present study, we evaluated the first stage of attachment of menstrual endometrium with use of this in vivo model to mimic, as closely as possible, the natural process of regurgitation of menstrual debris. To evaluate the adherence process in endometrium, other in vitro models have been developed, but their conclusions remain controversial. Monolayer cultures of human mesothelium have been used to evaluate the effects of cytokines on the attachment of endometrial stromal cells and whole endometrial fragments to mesothelium (20, 21). Some investigators have questioned the role of the mesothelial cells in the attachment of ectopic cells to the peritoneum (5). Using amnion membranes, they suggested that an intact mesothelium may prevent the endometrium from adhering to the peritoneum. Using electron microscopy, Groothuis et al. (6) supported the hypothesis that the mesothelial layer acts as a barrier to the attachment of ectopic endometrium and that the initial attachment of refluxed menstruated endometrial cells in the peritoneal cavity depends on breaches in the mesothelial layer. Recently, Koks et al. (9) showed that antegradely shed menstrual tissue easily adheres to the ECM and basement membrane of amnion and peritoneum but not to an intact epithelial layer. This conclusion was recently questioned by Witz et al. (4) who showed attachment of endometrium via stromal cells to intact mesothelium. They developed a novel model of human explants of peritoneum cultured with mechanically dispersed endometrium. In the present study, menstrual endometrium was transplanted into 10 nude mice to mimic human menstrual regurgitation and to histologically evaluate the first stages of development of endometriotic lesions. We observed that as early as 24 hours after transplantation, menstrual endometrium was attached in all cases to the mesothelium. Most of the graft consisted of stromal cells with large areas of necrosis. Our results confirm the previously published report by Aoki et al. (12) who used proliferative or secretory human endometrium and showed attachment of endometrial cells and their viability as early as 1 day after transplantation. In our study, at the first stage of implantation, endometrial stromal cells were the only cells found at the transplantperitoneum interface. We showed that an intact layer of mesothelium could be usually identified up to the edge of the endometrial implant and that endometrial stromal cells were attached to the peritoneum, representing the initial step in the endometriotic process (Fig. 1). Our morphological study confirms that the attachment process is rapid and that stromal cells play a crucial role in this attachment step. This finding is in accordance with a recent publication by Witz et al. (4), highlighting the role of stromal cells in the initial phase of the endometriotic process. In the study by Aoki et al. (12), subcutaneously transplanted proliferative or secretory endometrial specimens were viable 1 day after transplantation, although degenerative changes were observed in the central part of the transplant, FERTILITY & STERILITY 311

7 probably caused by ischemia. Their time-serial morphological studies showed that by 2 weeks after transplantation, glandular epithelial cells and stromal cells had rearranged themselves and reconstituted glands showed some cystic figures. On day 14, the morphological aspects were reproduced in a similar manner up to day 70. Contrary to what was previously believed, the results of our study suggest that shortly after the attachment of stromal cells, a rearrangement of both types of cells occurs, leading to the development of endometriotic lesions and cystic glands in the space of just 5 days. This finding is probably the result of the extensive proliferation observed in glandular cells as early as 3 days after transplantation, which strongly suggests that stromal and glandular cells have two distinct roles: stromal cells are involved in the attachment process and glandular cells in the growth of the endometriotic lesion. It is becoming increasingly evident, as shown in one of our studies of the endometrium, that cytokines act as messengers between stromal and epithelial cells (22). The significant increase in stromal VEGF scores clearly highlights the role of the stromal cells in angiogenesis concomitantly with glandular reorganization. In the present study, we also clearly showed that menstrual endometrium consisting of necrotic tissue is able to implant and reorganize itself into structured glands and stroma under the influence of unknown factors. The first step of attachment involves the stromal cells because, in all cases, no epithelial cells were observed at the interface between the implant and the mouse tissue. The reorganization step is still poorly understood. Are the stromal cells responsible for the remodeling of epithelial cells into typical glandular structures? Further studies are needed to elucidate this crucial step of reorganization of menstrual endometrium into typical endometriotic lesions. In conclusion, this study highlights the following: [1] that menstrual endometrium is most likely to adhere (100% of cases); [2] the primordial role of stromal cells in the attachment process; [3] that damaged mesothelium is not required for adhesion; and [4] that after attachment, stromal cells are responsible for inducing early angiogenesis and further development, including the reorganization of epithelial and stromal cells into endometrial glands. References 1. Sampson J. Peritoneal endometriosis is due to the menstrual dissemination of endometrial tissue into the peritoneal cavity. Am J Obstet Gynecol 1927;14: Van der Linden PJ, de Goeij AF, Dunselman GA, Van der Linden EP, Ramackers FC, Evers JL. Expression of integrins and E-cadherin in cells from menstrual effluent, endometrium, peritoneal fluid, peritoneum and endometriosis. Fertil Steril 1994;61: Beliard A, Donnez J, Nisolle M, Foidart JM. Localization of laminin, fibronectin, E-cadherin and integrins in endometrium and endometriosis. Fertil Steril 1997;67: Witz CA, Monotoya-Rodriguez IA, Schenken RS. Whole explants of peritoneum and endometrium: a novel model of the early endometriosis lesion. Fertil Steril 1999;71: Van der Linden PQJ, de Goeij AF, Dunselman GA, Erkens HW, Evers JL. Endometrial cell adhesion is an in vitro model using intact amniotic membrane. Fertil Steril 1996;65: Groothuis PG, Koks CA, de Goeij AF, Dunselman GA, Arend SJW, Evers JL. Adhesion of human endometrium to the epithelial lining and extracellular matrix of amnion in vitro: an electron microscopic study. Hum Reprod 1998;13: Groothuis PG, Koks CA, de Goeij AF, Dunselman GA, Arends JW, Evers JL. Adhesion of human endometrial fragments to peritoneum in vitro. Fertil Steril 1999;71: Nisolle M, Casanas-Roux F, Marbaix E, Jadoul P, Donnez J. Transplantation of cultured explants of human endometrium into nude mice. Hum Reprod 2000;15: Koks CA, Groothuis PG, Dunselman GA, de Goeij AF, Evers JL. Adhesion of shed menstrual tissue in an in-vitro model using amnion and peritoneum: a light and electron microscopic study. Hum Reprod 1999;14: Zamah NM, Dodson MG, Stephens LC, Buttram VC, Besch PK, Kaufman RH. Transplantation of normal and ectopic human endometrial tissue into athymic nude mice. Am J Obset Gynecol 1984;149: Berqgvist A, Jeppsson S, Kullander S, Ljunberg O. Human endometrium transplanted into nude mice. Histologic effects of various steroid hormones. Am J Pathol 1985;119: Aoki D, Katsuki Y, Shimizu A, Kakinuma E, Nozawa S. Successful heterotransplantation of human endometrium in SCID mice. Obstet Gynecol 1994;83: Bruner KL, Matrisian LM, Rodgers WH, Gorstein F, Osteen KG. Progesterone suppresion of matrix metalloproteinases prevents establishment of endometriotic-like human lesions in nude mice. J Clin Invest 1997;99: Noyes RW, Hertig AR, Rock J. Dating the endometrial biopsy. Fertil Steril 1950;1: Wortis HH. Immunologic responses of nude mice. Clin Exp Immunol 1971;8: Gabe M. Les colorations topographiques. In: Masson et Cie, eds. Techniques histologiques. Paris: Masson, 1968; Nisolle M, Casanas-Roux F, Donnez J. Immunohistochemical analysis of proliferative activity and steroid receptor expression in peritoneal and ovarian endometriosis. Fertil Steril 1997;68: Donnez J, Smoes P, Gillerot S, Casanas-Roux F, Nisolle M. Vascular endothelial growth factor (VEGF) in endometriosis. Hum Reprod 1998; 13: Siegel S. Non parametric statistics for the behavioral sciences. International student edition. New York: McGraw-Hill, Zhang R, Wild RA, Ojago JM. Effect of tumor necrosis factor- on adhesion of human endometrial stromal cells to peritoneal mesothelial cells: an in vitro system. Fertil Steril 1993;59: Wild RA, Zhang R, Medders D. Whole fragments from characteristics of in vivo endometriosis in a mesothelial cell co-culture system: an in vitro model of the histogenesis of endometriosis. J Soc Gynecol Invent 1994;1: Singer C, Marbaix E, Kokorine I, Leaoine P, Donnez J, Eeckhout Y, Courtoy PJ. Paracrine stimulation of intersitial collagenase (MMP-1) in the human endometrium by interleukin 1a and its dual block by ovarian steroids. Proc Natl Acad Sci USA 1997;94: Nisolle et al. Early stage endometriosis Vol. 74, No. 2, August 2000