- Email: [email protected]
Expression of the α2β1 and α3β1 integrins at the surface of mesothelial cells: a potential attachment site of endometrial cells
FERTILITY AND STERILITY威 VOL. 74, NO. 3, SEPTEMBER 2000 Copyright ©2000 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A.
Expression of the ␣21 and ␣31 integrins at the surface of mesothelial cells: a potential attachment site of endometrial cells Craig A. Witz, M.D.,a Akiyuki Takahashi, M.D.,b Iris A. Montoya-Rodriguez, B.S.,a Sook Cho, M.D., Ph.D.,a and Robert S. Schenken, M.D.a The University of Texas Health Science Center at San Antonio, San Antonio, Texas
Received December 23, 1999; revised and accepted March 28, 2000. Presented at the Conjoint Annual Meeting of the American Society for Reproductive Medicine and the Canadian Fertility and Andrology Society, Toronto, Ontario, Canada, September 25–29, 1999. Reprint requests: Craig A. Witz, M.D., Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7836 (FAX: 210-567-4958; Email: [email protected]). a Department of Obstetrics and Gynecology. b Department of Cellular and Structural Biology. 0015-0282/00/$20.00 PII S0015-0282(00)00701-9
Objective: To localize ␣21 and ␣31 integrins in the cell membrane of peritoneal mesothelium in vivo and in vitro. Design: Descriptive study using confocal and two-photon laser-scanning microscopy. Setting: University-based laboratory. Patient(s): Women without endometriosis undergoing surgery for benign conditions. Intervention(s): None. Main Outcome Measure(s): Peritoneal biopsies (n ⫽ 9) and mesothelial monolayer cultures (n ⫽ 4) were incubated with antibodies to the ␣2 and ␣3 subunits and to the intact ␣21 and ␣31 integrins. Specimens were examined with laser-scanning microscopy. Result(s): The ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins were identified at the base of the mesothelial cells (i.e., toward the basement membrane). There was also expression of the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins at the cell surface (i.e., toward the peritoneal cavity). Conclusion(s): The resolution of the confocal and two-photon laser-scanning microscope enabled localization of integrins in mesothelial cells. The presence of ␣21 (collagen-laminin receptor) and ␣31 integrins (collagen-laminin-fibronectin receptor) at the base of mesothelial cells suggests a role for these molecules in adhesion to the basement membrane. The presence of these molecules at the cell surface suggests a potential locus for cell adhesion in such processes as endometriosis and cancer metastasis. (Fertil Steril威 2000;74: 579 – 84. ©2000 by American Society for Reproductive Medicine.) Key Words: Peritoneum, mesothelium, integrin, endometriosis, cancer metastasis, confocal, laser-scanning microscopy
We recently described a model of endometriosis with use of explants of human peritoneum and endometrium. This study showed that endometrial cells can attach to intact peritoneal mesothelium. After attachment, the endometrium seems to rapidly invade through the mesothelium (i.e., in ⬍16 –24 hours) (1). Similarly, Wild et al. (2) showed that fragments of endometrium attach to monolayers of mesothelium grown on a collagen matrix. After attachment of endometrium, the mesothelial monolayer is disrupted and later grows over the attached fragment. The mechanisms involved in endometrial cell attachment to the peritoneum remain poorly defined because little is known about
the expression of cell adhesion molecules at the mesothelial cell surface. Integrin expression by ovarian cancer cells has been implicated in the binding of cancer cell lines to mesothelial cells (3, 4). Integrins are heterodimeric, transmembrane proteins composed of an ␣ and  subunit that mediate cell-cell and cell-extracellular matrix interactions (5). We previously described the expression of ␣ integrin subunits in mesothelium of anterior peritoneum, uterine serosa, and in monolayer culture. This study showed strong expression of ␣2 (collagen-laminin receptor) and ␣3 (collagen-laminin-fibronectin receptor) integrin subunits and variable expression of ␣6 (laminin receptor) in vivo. There were no differ579
ences seen in the expression of these integrin subunits when the anterior peritoneum and uterine serosa were compared. However, in the monolayer cultures there was strong expression of the ␣5 subunit in addition to the ␣2 and ␣3 integrin subunits. In addition, in vitro there was minimal expression of the ␣v subunit and no expression of ␣6 (6). Similar findings of integrin subunit expression by cultured mesothelial cells were recently reported by Tietze et al. (7). In their study, cultured mesothelial cells strongly expressed the 1, 3, ␣2, ␣3, ␣5, and ␣v subunits. There was minimal reactivity for the ␣1 and ␣6 subunits. We also examined the expression of the mesothelial ␣2 and ␣3 integrin subunits in vivo and in monolayer culture with use of immunoelectron microscopy. These subunits were seen distributed throughout the mesothelial cell cytoplasm, were expressed in the plasma membrane, and were present on the cell surface (i.e., toward the peritoneal cavity) (6). Our previous investigation was limited in that it examined the expression of integrin subunits. Similar to our study, other investigators have reported cytoplasmic immunohistochemical localization of integrin subunits (8, 9). However, appearance on the plasma membrane, the site of integrin activity, depends on heterodimerization of the ␣ and  integrin subunits (8). The purpose of this study was to localize intact ␣21 (collagen-laminin receptor) and ␣31 (collagen-laminin-fibronectin receptor) integrins in the cell membrane of peritoneal mesothelium in vivo and in vitro. Because mesothelial cells are usually ⬍10 m thick (6), confocal (CLSM) and two-photon laser-scanning microscopy (TPLSM) were used to discriminate the basal from the apical plasma membrane.
MATERIALS AND METHODS Approval for this study was given by the Institutional Review Board of The University of Texas Health Science Center at San Antonio.
Peritoneum Tissue Collection
Peritoneum from the anterior abdominal wall (n ⫽ 9) was obtained from reproductive aged women without endometriosis who were undergoing surgery for benign conditions. The tissue was immediately placed in Tissue-Tek O.C.T. Compound (Miles Laboratories, Elkart, IN), frozen in the vapors of liquid nitrogen, and stored at ⫺70°C until sectioning. Serial 4-m sections of the peritoneum were placed on silane-coated slides and fixed for 5 minutes in 4°C acetone. Slides were then placed in Tris-buffered saline before hematoxylin and eosin staining (H & E) and immunohistochemistry. The H & E stains were used to properly orient tissue sections and ensure an intact mesothelium. 580
Witz et al.
Mesothelial Monolayer Cultures Mesothelial cells were enzymatically dispersed from anterior peritoneum biopsies (n ⫽ 4) with use of 0.1% type I collagenase (Worthington Biomedical, Freehold, NJ) and 0.1% DNAse I (Sigma Company, St. Louis, MO). Cells were plated in 75-cm2 polystyrene flasks (Corning Costar, Cambridge, MA). Cells were grown in Minimum Essential Medium Eagle (D-Valine modification; Sigma) supplemented with epidermal growth factor (20 ng/mL; Gibco BRL, Gaithersburg, MD) and 10% defined fetal calf serum (Hyclone, Logan, UT) (2). Cells were grown to confluence and then passaged. Morphologic assessment of the monolayer cultures was performed with use of H & E staining. In addition to assessment of morphology, purity of mesothelial monolayer cultures was determined by incubation with monoclonal antibodies to human cytokeratin (Oncogene Science, Uniondale, NY), vimentin (Oncogene Science), and von Willebrand factor (Dako, Carpinteria, CA) (6, 10). Slides were incubated with biotinylated horse anti-mouse secondary antibody followed by avidin-biotin peroxidase complex (Vectastain Elite, Burlingame, CA). The slides were then reacted with diaminobenzidine (DAB) and hydrogen peroxide, yielding a brown reaction. On the second or third passage, monolayers were grown on Aclar fluoropolymer film (ProPlastics Inc., Linden, NJ) cut into 1.5-cm squares placed at the bottom of 24-well plates (Corning Costar). The cells were grown to subconfluence on the films, and the films were embedded and frozen in the same manner as sections of anterior peritoneum. Previous studies showed that mesothelial cells attach and proliferate on this film and maintain a similar morphology to those grown on glass and plastic surfaces. Furthermore, this film can be embedded and processed for frozen section (6).
Immunohistochemistry Sections of anterior peritoneum and monolayer cultures were incubated with monoclonal antibodies to ␣2 integrin subunit (clone P1E6; Chemicon, Temecula, CA), intact ␣21 integrin (clone BHA2.1; Chemicon) (11, 12), ␣3 integrin subunit (clone P1B5; Chemicon), and intact ␣31 integrin (clone M-KID 2; Chemicon) (13). Slides were then incubated with fluorescein-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Some slides underwent tyramide signal amplification. After reaction with primary antibody, slides were incubated with biotinylated horse anti-mouse secondary antibody followed by streptavidin-horseradish peroxidase (Vectastain). Slides were then incubated with biotinyl tyramide followed by streptavidin-fluorescein isothiocyanate (NEN Life Science Products, Boston, MA).
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000
FIGURE 1 Peritoneum biopsy showing the distribution of the ␣21 integrin in mesothelium. M ⫽ mesothelial cells. Arrowheads illustrate expression on the basal plasma membrane. Arrows illustrate expression at the cell surface. (CLSM, original magnification, ⫻750; bar ⫽ 10 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
Controls Negative controls were performed by substituting nonimmune mouse immunoglobulin (Dako) for the primary antibody. The concentration of the nonimmune mouse immunoglobulin was determined by matching the concentration of the primary antibody. Negative controls were performed for both nonamplified and tyramide signal amplified studies.
RESULTS Mesothelial Monolayer Culture The cultured mesothelial cells had a squamoid polygonal shape, strongly expressed the intermediate filaments vimentin and cytokeratin, and did not stain for von Willebrand factor (6) (data not shown).
Mesothelial Cell Integrin Expression
Confocal and Two-Photon Laser-Scanning Microscopy CLSM and TPLSM was performed with use of an Olympus IX70 inverted microscope and the Olympus Fluoview System (Nagano, Japan). Slides were examined with a ⫻60 objective with numeric aperture of 1.20 under water immersion. For CLSM, fluorescein was excited with an argon laser at 488 nm (NEC, Nagano, Japan). For TPLSM, a Verdi/Mira femto-second pulsed Ti:sapphire laser (Coherent, Inc., Santa Clara, CA) was used at 800 nm. Fluorescein emission was detected at 510 –550 nm. Digital images of 512 ⫻ 512 pixels were recorded with use of a speed of 16 s/scan. Simultaneous bright-field images were recorded with fluorescein emission. Images were processed for printing with use of Adobe Photoshop (San Jose, CA) on a Dell personal computer (Austin, TX). FERTILITY & STERILITY威
In the peritoneal biopsies, the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins were identified at the base of the mesothelial cells (i.e., toward the basement membrane). In addition, there was expression of the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins at the cell surface (i.e., toward the peritoneal cavity). Figure 1 shows labeling of the intact ␣21 integrin in the plasma membrane at the base and apical surface of the mesothelial cell. A similar pattern was found for the intact ␣31 integrin. In addition to the cell membrane, a diffuse labeling of the ␣3 subunit was seen in the cytoplasm of the mesothelial cells in vivo and in vitro. Figure 2 illustrates the distribution of the ␣3 subunit in mesothelial cells in vivo. The patterns of expression of the integrin subunits and intact ␣21 and ␣31 integrins in monolayer culture were similar to peritoneal biopsies. There was expression of the ␣3 subunit and the ␣31 integrin at the base of the cell (i.e., 581
FIGURE 2 Peritoneum biopsy showing the distribution of the ␣3 integrin subunit in mesothelium. Expression of the ␣3 integrin is seen at the base and surface of the cell. Diffuse labeling is seen throughout the cytoplasm. (CLSM, original magnification, ⫻450; bar ⫽ 10 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
toward the Aclar film). In addition, there was expression at the apical cell surface. However, the expression of the ␣2 subunit and the ␣2 1 integrin at the base of the cell was decreased compared with mesothelium in the peritoneal biopsies. Figure 3 shows the labeling of the intact ␣21 integrin in mesothelial cell culture.
Negative Controls There was no significant staining of tissue in the peritoneal biopsies and mesothelial monolayers when nonimmune mouse immunoglobulin was substituted for primary antibody in the nonamplified and in the tyramide amplified slides (data not shown).
DISCUSSION Integrins seem to be the major receptor by which cells attach to components of the extracellular matrix such as collagen, laminin (LM), fibronectin (FN), and vitronectin. Some integrins also mediate important cell-cell adhesion (5). As a result of integrin binding, intracellular signals are transduced that make a major contribution to the regulation of cell phenotype (14, 15). In the present study, the use of monoclonal antibodies to the ␣21 (11, 12) and ␣31 heterodimers (13) and the resolution of the CLSM and TPLSM allowed for localization of intact integrins in mesothelial cells. We showed the pres582
Witz et al.
ence of ␣21 (collagen-laminin receptor) and ␣31 integrins (collagen-laminin-fibronectin receptor) at the base of mesothelial cells, suggesting a role for these molecules in adhesion to the basement membrane. In a previous investigation using CLSM and immunohistochemistry, we evaluated the composition of the submesothelial ECM. Laminin and collagen IV (C IV) were seen intermixed in the basement membrane underlying the mesothelium. Intermingled with the basement membrane was a condensation of FN and collagen I (C I) (unpublished). In addition, the present study showed the ␣21 and ␣31 integrins at the surface of the mesothelial cell (i.e., toward the peritoneal cavity). It is possible that mesothelial surface integrins could play a role in the initial attachment of cells to the peritoneum in such processes as endometriosis and cancer metastasis. The composition of the ECM surrounding endometrial glands and stroma includes type III, type IV, and type V collagens, LM, and FN, which are all ligands of the ␣21 and ␣31 integrins (16). Similarly, ovarian cystadenocarcinoma cells synthesize the ECM proteins C IV and LM when grown in serum-free medium (17). In vivo, we found similar localization of the ␣2 and ␣3 integrin subunits compared with the ␣21 and ␣31 heterodimers. A diffuse labeling of the ␣3 integrin subunit was also seen in the mesothelial cell cytoplasm. Our previous study using the same monoclonal antibodies to the ␣2 (clone
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000
FIGURE 3 Distribution of the ␣21 integrin subunit in a mesothelial cell grown on Aclar film. There is decreased expression at the base of the cell (arrowheads) and stronger expression at the cell surface (arrows). (TPLSM, original magnification, ⫻1,050; bar ⫽ 5 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
P1E6) and ␣3 (clone P1B5) integrin subunits and immunoelectron microscopy showed the presence of these subunits in the plasma membrane and the cytoplasm. The technique of immunoelectron microscopy was not able to show a quantitative difference between the concentration of these subunits in the plasma membrane and mesothelial cell cytoplasm. There was a similar density of labeling throughout the cell (6). In contrast, the present study using CLSM and TPLSM clearly showed the most intense signal in the cell membrane. The mesothelial expression of the ␣3 subunit and the ␣31 integrin was the same in monolayer culture and peritoneum biopsies. In contrast, the expression of ␣2 subunit and the ␣21 integrin was slightly different. There was a decreased intensity of labeling of the ␣2 subunit and the ␣21 integrin at the base of cultured cells. It is possible that the decreased expression could be a result of cellular binding to FN, which becomes interposed between the cell and the Aclar film. Because FN is present in serum in high concentration (approximately 280 g/mL), ample FN to serve this purpose was present in the cell culture medium that contained 10% fetal calf serum. Furthermore, mesothelial cells grown in monolayer culture have been shown to produce FN (18). Consistent with the hypothesis of mesothelial cell binding to FN in vitro is the differential expression of integrins in FERTILITY & STERILITY威
monolayer culture. In culture, we have found that mesothelial cells do not express the ␣6 integrin subunit, which is a receptor for LM, a component of normal basement membrane (6). Others have found only a weak immunoreactivity for the ␣6 subunit in cultured mesothelium (7). In contrast, cultured mesothelial cells express the ␣5 and ␣v integrin subunits that are associated with FN binding (6, 7). Previous investigators have reported the expression of integrins in peritoneal biopsies and in epithelial cells recovered from peritoneal fluid in women with and without endometriosis. These investigators reported that ␣2, ␣3, ␣4, ␣5, and ␣6 are expressed by peritoneum in women without endometriosis. Unfortunately, the investigators did not state whether the staining in the peritoneum biopsy samples was limited to the mesothelium (19). We believe that the present study is the first to clearly show the presence of integrins on the cell surface of mesothelial cells. These integrins are potential binding sites for ectopic endometrium and metastatic cancer cells. We found that the ␣21 and ␣31 integrins are present on the surface of mesothelium in vivo and in vitro. Mesothelial monolayer culture, therefore, seems to be an appropriate model to study the role of these integrins during the initial interactions of ectopic endometrium and cancer cells with the peritoneum. 583
References 1. Witz CA, Montoya-Rodriguez IA, Schenken RS. Whole peritoneal explants—a novel model of the early endometriosis lesion. Fertil Steril 1999;71:56 – 60. 2. Wild RA, Ren-jie Z, Medders D. Whole endometrial fragments form characteristics of in vivo endometriosis in a mesothelial cell co-culture system: an in vitro model for the study of the histogenesis of endometriosis. J Soc Gynecol Invest 1994;1:65– 8. 3. Strobel T, Cannistra SA. Beta1-integrins partly mediate binding of ovarian cancer cells to peritoneal mesothelium in vitro. Gynecol Oncol 1999;73:362–7. 4. Lessan K, Aguiar DJ, Oegema T, Siebenson L, Skubitz AP. CD44 and beta1 integrin mediate ovarian carcinoma cell adhesion to peritoneal mesothelial cells. Am J Pathol 1999;154:1525–37. 5. Hynes R. Integrins: versatility, modulation, and signalling in cell adhesion. Cell 1992;69:11–25. 6. Witz CA, Montoya-Rodriguez IA, Miller DM, Schneider BG, Schenken RS. Mesothelium expression of integrins in vivo and in vitro. J Soc Gynecol Invest 1998;5:87–93. 7. Tietze L, Borntraeger J, Klosterhalfen B, Amo-Takyi B, Handt S, Gunther K, et al. Expression and function of beta(1) and beta(3) integrins of human mesothelial cells in vitro. Exp Mol Pathol 1999;66: 131–9. 8. Bosman FT. Cell adhesives and modulators of cell function. Histochem J 1993;25:469 –77. 9. Tabibzadeh S. Evidence of T-cell activation and potential cytokine action in human endometrium. J Clin Endocrinol Metab 1990;71:645–9. 10. La Rocca P, Rheinwald J. Coexpression of simple epithelia keratin and vimentin by human mesothelium and mesothelioma in vivo and in culture. Cancer Res 1984;44:2991–5.
584
Witz et al.
11. Wang XQ, Frazier WA. The thrombospondin receptor CD47 (IAP) modulates and associates with alpha2 beta1 integrin in vascular smooth muscle cells. Mol Biol Cell 1998;9:865–74. 12. Hangan D, Uniyal S, Morris VL, MacDonald IC, von Ballestrem C, Chau T, et al. Integrin VLA-2 (alpha2beta1) function in postextravasation movement of human rhabdomyosarcoma RD cells in the liver. Cancer Res 1996;56:3142–9. 13. Bartolazzi A, Fraioli R, Tarone G, Natali PG. Generation and characterization of the murine monoclonal antibody M-KID 2 to VLA-3 integrin. Hybridoma 1991;10:707–20. 14. Humphries MJ, Mould AP, Tuckwell DS. Dynamic aspects of adhesion receptor function-integrins both twist and shout. Bioessays 1993;15: 391–7. 15. Giancotti FG, Ruoslahti E. Integrin signaling. Science 1999;285:1028 – 32. 16. Aplin JD, Charlton AK, Ayad S. An immunohistochemical study of human endometrial extracellular matrix during the menstrual cycle and first trimester of pregnancy. Cell Tissue Res 1988;253:231– 40. 17. Golombick T, Dajee DR, Bezwoda WR. Extracellular matrix interactions. 1. Production of extracellular matrix with attachment and growthsustaining functions by UWOV2 ovarian cancer cells growing in protein-free conditions. In Vitro Cell Dev Biol 1995;31:387–95. 18. Cannistra SA, Ottensmeier C, Niloff J, Orta B, DiCarlo J. Expression and function of beta 1 and alpha v beta 3 integrins in ovarian cancer. Gynecol Oncol 1995;58:216 –25. 19. van der Linden P, de Goeij A, Dunselman G, van der Linden E, Ramaekers F, Evers J. Expression of integrins and E-cadherin in cells from menstrual effluent, endometrium, peritoneal fluid, peritoneum and endometriosis. Fertil Steril 1994;61:85–90.
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000
Expression of the ␣21 and ␣31 integrins at the surface of mesothelial cells: a potential attachment site of endometrial cells Craig A. Witz, M.D.,a Akiyuki Takahashi, M.D.,b Iris A. Montoya-Rodriguez, B.S.,a Sook Cho, M.D., Ph.D.,a and Robert S. Schenken, M.D.a The University of Texas Health Science Center at San Antonio, San Antonio, Texas
Received December 23, 1999; revised and accepted March 28, 2000. Presented at the Conjoint Annual Meeting of the American Society for Reproductive Medicine and the Canadian Fertility and Andrology Society, Toronto, Ontario, Canada, September 25–29, 1999. Reprint requests: Craig A. Witz, M.D., Division of Reproductive Endocrinology, Department of Obstetrics and Gynecology, The University of Texas Health Science Center at San Antonio, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7836 (FAX: 210-567-4958; Email: [email protected]). a Department of Obstetrics and Gynecology. b Department of Cellular and Structural Biology. 0015-0282/00/$20.00 PII S0015-0282(00)00701-9
Objective: To localize ␣21 and ␣31 integrins in the cell membrane of peritoneal mesothelium in vivo and in vitro. Design: Descriptive study using confocal and two-photon laser-scanning microscopy. Setting: University-based laboratory. Patient(s): Women without endometriosis undergoing surgery for benign conditions. Intervention(s): None. Main Outcome Measure(s): Peritoneal biopsies (n ⫽ 9) and mesothelial monolayer cultures (n ⫽ 4) were incubated with antibodies to the ␣2 and ␣3 subunits and to the intact ␣21 and ␣31 integrins. Specimens were examined with laser-scanning microscopy. Result(s): The ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins were identified at the base of the mesothelial cells (i.e., toward the basement membrane). There was also expression of the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins at the cell surface (i.e., toward the peritoneal cavity). Conclusion(s): The resolution of the confocal and two-photon laser-scanning microscope enabled localization of integrins in mesothelial cells. The presence of ␣21 (collagen-laminin receptor) and ␣31 integrins (collagen-laminin-fibronectin receptor) at the base of mesothelial cells suggests a role for these molecules in adhesion to the basement membrane. The presence of these molecules at the cell surface suggests a potential locus for cell adhesion in such processes as endometriosis and cancer metastasis. (Fertil Steril威 2000;74: 579 – 84. ©2000 by American Society for Reproductive Medicine.) Key Words: Peritoneum, mesothelium, integrin, endometriosis, cancer metastasis, confocal, laser-scanning microscopy
We recently described a model of endometriosis with use of explants of human peritoneum and endometrium. This study showed that endometrial cells can attach to intact peritoneal mesothelium. After attachment, the endometrium seems to rapidly invade through the mesothelium (i.e., in ⬍16 –24 hours) (1). Similarly, Wild et al. (2) showed that fragments of endometrium attach to monolayers of mesothelium grown on a collagen matrix. After attachment of endometrium, the mesothelial monolayer is disrupted and later grows over the attached fragment. The mechanisms involved in endometrial cell attachment to the peritoneum remain poorly defined because little is known about
the expression of cell adhesion molecules at the mesothelial cell surface. Integrin expression by ovarian cancer cells has been implicated in the binding of cancer cell lines to mesothelial cells (3, 4). Integrins are heterodimeric, transmembrane proteins composed of an ␣ and  subunit that mediate cell-cell and cell-extracellular matrix interactions (5). We previously described the expression of ␣ integrin subunits in mesothelium of anterior peritoneum, uterine serosa, and in monolayer culture. This study showed strong expression of ␣2 (collagen-laminin receptor) and ␣3 (collagen-laminin-fibronectin receptor) integrin subunits and variable expression of ␣6 (laminin receptor) in vivo. There were no differ579
ences seen in the expression of these integrin subunits when the anterior peritoneum and uterine serosa were compared. However, in the monolayer cultures there was strong expression of the ␣5 subunit in addition to the ␣2 and ␣3 integrin subunits. In addition, in vitro there was minimal expression of the ␣v subunit and no expression of ␣6 (6). Similar findings of integrin subunit expression by cultured mesothelial cells were recently reported by Tietze et al. (7). In their study, cultured mesothelial cells strongly expressed the 1, 3, ␣2, ␣3, ␣5, and ␣v subunits. There was minimal reactivity for the ␣1 and ␣6 subunits. We also examined the expression of the mesothelial ␣2 and ␣3 integrin subunits in vivo and in monolayer culture with use of immunoelectron microscopy. These subunits were seen distributed throughout the mesothelial cell cytoplasm, were expressed in the plasma membrane, and were present on the cell surface (i.e., toward the peritoneal cavity) (6). Our previous investigation was limited in that it examined the expression of integrin subunits. Similar to our study, other investigators have reported cytoplasmic immunohistochemical localization of integrin subunits (8, 9). However, appearance on the plasma membrane, the site of integrin activity, depends on heterodimerization of the ␣ and  integrin subunits (8). The purpose of this study was to localize intact ␣21 (collagen-laminin receptor) and ␣31 (collagen-laminin-fibronectin receptor) integrins in the cell membrane of peritoneal mesothelium in vivo and in vitro. Because mesothelial cells are usually ⬍10 m thick (6), confocal (CLSM) and two-photon laser-scanning microscopy (TPLSM) were used to discriminate the basal from the apical plasma membrane.
MATERIALS AND METHODS Approval for this study was given by the Institutional Review Board of The University of Texas Health Science Center at San Antonio.
Peritoneum Tissue Collection
Peritoneum from the anterior abdominal wall (n ⫽ 9) was obtained from reproductive aged women without endometriosis who were undergoing surgery for benign conditions. The tissue was immediately placed in Tissue-Tek O.C.T. Compound (Miles Laboratories, Elkart, IN), frozen in the vapors of liquid nitrogen, and stored at ⫺70°C until sectioning. Serial 4-m sections of the peritoneum were placed on silane-coated slides and fixed for 5 minutes in 4°C acetone. Slides were then placed in Tris-buffered saline before hematoxylin and eosin staining (H & E) and immunohistochemistry. The H & E stains were used to properly orient tissue sections and ensure an intact mesothelium. 580
Witz et al.
Mesothelial Monolayer Cultures Mesothelial cells were enzymatically dispersed from anterior peritoneum biopsies (n ⫽ 4) with use of 0.1% type I collagenase (Worthington Biomedical, Freehold, NJ) and 0.1% DNAse I (Sigma Company, St. Louis, MO). Cells were plated in 75-cm2 polystyrene flasks (Corning Costar, Cambridge, MA). Cells were grown in Minimum Essential Medium Eagle (D-Valine modification; Sigma) supplemented with epidermal growth factor (20 ng/mL; Gibco BRL, Gaithersburg, MD) and 10% defined fetal calf serum (Hyclone, Logan, UT) (2). Cells were grown to confluence and then passaged. Morphologic assessment of the monolayer cultures was performed with use of H & E staining. In addition to assessment of morphology, purity of mesothelial monolayer cultures was determined by incubation with monoclonal antibodies to human cytokeratin (Oncogene Science, Uniondale, NY), vimentin (Oncogene Science), and von Willebrand factor (Dako, Carpinteria, CA) (6, 10). Slides were incubated with biotinylated horse anti-mouse secondary antibody followed by avidin-biotin peroxidase complex (Vectastain Elite, Burlingame, CA). The slides were then reacted with diaminobenzidine (DAB) and hydrogen peroxide, yielding a brown reaction. On the second or third passage, monolayers were grown on Aclar fluoropolymer film (ProPlastics Inc., Linden, NJ) cut into 1.5-cm squares placed at the bottom of 24-well plates (Corning Costar). The cells were grown to subconfluence on the films, and the films were embedded and frozen in the same manner as sections of anterior peritoneum. Previous studies showed that mesothelial cells attach and proliferate on this film and maintain a similar morphology to those grown on glass and plastic surfaces. Furthermore, this film can be embedded and processed for frozen section (6).
Immunohistochemistry Sections of anterior peritoneum and monolayer cultures were incubated with monoclonal antibodies to ␣2 integrin subunit (clone P1E6; Chemicon, Temecula, CA), intact ␣21 integrin (clone BHA2.1; Chemicon) (11, 12), ␣3 integrin subunit (clone P1B5; Chemicon), and intact ␣31 integrin (clone M-KID 2; Chemicon) (13). Slides were then incubated with fluorescein-conjugated goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA). Some slides underwent tyramide signal amplification. After reaction with primary antibody, slides were incubated with biotinylated horse anti-mouse secondary antibody followed by streptavidin-horseradish peroxidase (Vectastain). Slides were then incubated with biotinyl tyramide followed by streptavidin-fluorescein isothiocyanate (NEN Life Science Products, Boston, MA).
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000
FIGURE 1 Peritoneum biopsy showing the distribution of the ␣21 integrin in mesothelium. M ⫽ mesothelial cells. Arrowheads illustrate expression on the basal plasma membrane. Arrows illustrate expression at the cell surface. (CLSM, original magnification, ⫻750; bar ⫽ 10 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
Controls Negative controls were performed by substituting nonimmune mouse immunoglobulin (Dako) for the primary antibody. The concentration of the nonimmune mouse immunoglobulin was determined by matching the concentration of the primary antibody. Negative controls were performed for both nonamplified and tyramide signal amplified studies.
RESULTS Mesothelial Monolayer Culture The cultured mesothelial cells had a squamoid polygonal shape, strongly expressed the intermediate filaments vimentin and cytokeratin, and did not stain for von Willebrand factor (6) (data not shown).
Mesothelial Cell Integrin Expression
Confocal and Two-Photon Laser-Scanning Microscopy CLSM and TPLSM was performed with use of an Olympus IX70 inverted microscope and the Olympus Fluoview System (Nagano, Japan). Slides were examined with a ⫻60 objective with numeric aperture of 1.20 under water immersion. For CLSM, fluorescein was excited with an argon laser at 488 nm (NEC, Nagano, Japan). For TPLSM, a Verdi/Mira femto-second pulsed Ti:sapphire laser (Coherent, Inc., Santa Clara, CA) was used at 800 nm. Fluorescein emission was detected at 510 –550 nm. Digital images of 512 ⫻ 512 pixels were recorded with use of a speed of 16 s/scan. Simultaneous bright-field images were recorded with fluorescein emission. Images were processed for printing with use of Adobe Photoshop (San Jose, CA) on a Dell personal computer (Austin, TX). FERTILITY & STERILITY威
In the peritoneal biopsies, the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins were identified at the base of the mesothelial cells (i.e., toward the basement membrane). In addition, there was expression of the ␣2 and ␣3 subunits and the intact ␣21 and ␣31 integrins at the cell surface (i.e., toward the peritoneal cavity). Figure 1 shows labeling of the intact ␣21 integrin in the plasma membrane at the base and apical surface of the mesothelial cell. A similar pattern was found for the intact ␣31 integrin. In addition to the cell membrane, a diffuse labeling of the ␣3 subunit was seen in the cytoplasm of the mesothelial cells in vivo and in vitro. Figure 2 illustrates the distribution of the ␣3 subunit in mesothelial cells in vivo. The patterns of expression of the integrin subunits and intact ␣21 and ␣31 integrins in monolayer culture were similar to peritoneal biopsies. There was expression of the ␣3 subunit and the ␣31 integrin at the base of the cell (i.e., 581
FIGURE 2 Peritoneum biopsy showing the distribution of the ␣3 integrin subunit in mesothelium. Expression of the ␣3 integrin is seen at the base and surface of the cell. Diffuse labeling is seen throughout the cytoplasm. (CLSM, original magnification, ⫻450; bar ⫽ 10 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
toward the Aclar film). In addition, there was expression at the apical cell surface. However, the expression of the ␣2 subunit and the ␣2 1 integrin at the base of the cell was decreased compared with mesothelium in the peritoneal biopsies. Figure 3 shows the labeling of the intact ␣21 integrin in mesothelial cell culture.
Negative Controls There was no significant staining of tissue in the peritoneal biopsies and mesothelial monolayers when nonimmune mouse immunoglobulin was substituted for primary antibody in the nonamplified and in the tyramide amplified slides (data not shown).
DISCUSSION Integrins seem to be the major receptor by which cells attach to components of the extracellular matrix such as collagen, laminin (LM), fibronectin (FN), and vitronectin. Some integrins also mediate important cell-cell adhesion (5). As a result of integrin binding, intracellular signals are transduced that make a major contribution to the regulation of cell phenotype (14, 15). In the present study, the use of monoclonal antibodies to the ␣21 (11, 12) and ␣31 heterodimers (13) and the resolution of the CLSM and TPLSM allowed for localization of intact integrins in mesothelial cells. We showed the pres582
Witz et al.
ence of ␣21 (collagen-laminin receptor) and ␣31 integrins (collagen-laminin-fibronectin receptor) at the base of mesothelial cells, suggesting a role for these molecules in adhesion to the basement membrane. In a previous investigation using CLSM and immunohistochemistry, we evaluated the composition of the submesothelial ECM. Laminin and collagen IV (C IV) were seen intermixed in the basement membrane underlying the mesothelium. Intermingled with the basement membrane was a condensation of FN and collagen I (C I) (unpublished). In addition, the present study showed the ␣21 and ␣31 integrins at the surface of the mesothelial cell (i.e., toward the peritoneal cavity). It is possible that mesothelial surface integrins could play a role in the initial attachment of cells to the peritoneum in such processes as endometriosis and cancer metastasis. The composition of the ECM surrounding endometrial glands and stroma includes type III, type IV, and type V collagens, LM, and FN, which are all ligands of the ␣21 and ␣31 integrins (16). Similarly, ovarian cystadenocarcinoma cells synthesize the ECM proteins C IV and LM when grown in serum-free medium (17). In vivo, we found similar localization of the ␣2 and ␣3 integrin subunits compared with the ␣21 and ␣31 heterodimers. A diffuse labeling of the ␣3 integrin subunit was also seen in the mesothelial cell cytoplasm. Our previous study using the same monoclonal antibodies to the ␣2 (clone
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000
FIGURE 3 Distribution of the ␣21 integrin subunit in a mesothelial cell grown on Aclar film. There is decreased expression at the base of the cell (arrowheads) and stronger expression at the cell surface (arrows). (TPLSM, original magnification, ⫻1,050; bar ⫽ 5 m).
Witz. Mesothelium ␣21 and ␣31 integrin expression. Fertil Steril 2000.
P1E6) and ␣3 (clone P1B5) integrin subunits and immunoelectron microscopy showed the presence of these subunits in the plasma membrane and the cytoplasm. The technique of immunoelectron microscopy was not able to show a quantitative difference between the concentration of these subunits in the plasma membrane and mesothelial cell cytoplasm. There was a similar density of labeling throughout the cell (6). In contrast, the present study using CLSM and TPLSM clearly showed the most intense signal in the cell membrane. The mesothelial expression of the ␣3 subunit and the ␣31 integrin was the same in monolayer culture and peritoneum biopsies. In contrast, the expression of ␣2 subunit and the ␣21 integrin was slightly different. There was a decreased intensity of labeling of the ␣2 subunit and the ␣21 integrin at the base of cultured cells. It is possible that the decreased expression could be a result of cellular binding to FN, which becomes interposed between the cell and the Aclar film. Because FN is present in serum in high concentration (approximately 280 g/mL), ample FN to serve this purpose was present in the cell culture medium that contained 10% fetal calf serum. Furthermore, mesothelial cells grown in monolayer culture have been shown to produce FN (18). Consistent with the hypothesis of mesothelial cell binding to FN in vitro is the differential expression of integrins in FERTILITY & STERILITY威
monolayer culture. In culture, we have found that mesothelial cells do not express the ␣6 integrin subunit, which is a receptor for LM, a component of normal basement membrane (6). Others have found only a weak immunoreactivity for the ␣6 subunit in cultured mesothelium (7). In contrast, cultured mesothelial cells express the ␣5 and ␣v integrin subunits that are associated with FN binding (6, 7). Previous investigators have reported the expression of integrins in peritoneal biopsies and in epithelial cells recovered from peritoneal fluid in women with and without endometriosis. These investigators reported that ␣2, ␣3, ␣4, ␣5, and ␣6 are expressed by peritoneum in women without endometriosis. Unfortunately, the investigators did not state whether the staining in the peritoneum biopsy samples was limited to the mesothelium (19). We believe that the present study is the first to clearly show the presence of integrins on the cell surface of mesothelial cells. These integrins are potential binding sites for ectopic endometrium and metastatic cancer cells. We found that the ␣21 and ␣31 integrins are present on the surface of mesothelium in vivo and in vitro. Mesothelial monolayer culture, therefore, seems to be an appropriate model to study the role of these integrins during the initial interactions of ectopic endometrium and cancer cells with the peritoneum. 583
References 1. Witz CA, Montoya-Rodriguez IA, Schenken RS. Whole peritoneal explants—a novel model of the early endometriosis lesion. Fertil Steril 1999;71:56 – 60. 2. Wild RA, Ren-jie Z, Medders D. Whole endometrial fragments form characteristics of in vivo endometriosis in a mesothelial cell co-culture system: an in vitro model for the study of the histogenesis of endometriosis. J Soc Gynecol Invest 1994;1:65– 8. 3. Strobel T, Cannistra SA. Beta1-integrins partly mediate binding of ovarian cancer cells to peritoneal mesothelium in vitro. Gynecol Oncol 1999;73:362–7. 4. Lessan K, Aguiar DJ, Oegema T, Siebenson L, Skubitz AP. CD44 and beta1 integrin mediate ovarian carcinoma cell adhesion to peritoneal mesothelial cells. Am J Pathol 1999;154:1525–37. 5. Hynes R. Integrins: versatility, modulation, and signalling in cell adhesion. Cell 1992;69:11–25. 6. Witz CA, Montoya-Rodriguez IA, Miller DM, Schneider BG, Schenken RS. Mesothelium expression of integrins in vivo and in vitro. J Soc Gynecol Invest 1998;5:87–93. 7. Tietze L, Borntraeger J, Klosterhalfen B, Amo-Takyi B, Handt S, Gunther K, et al. Expression and function of beta(1) and beta(3) integrins of human mesothelial cells in vitro. Exp Mol Pathol 1999;66: 131–9. 8. Bosman FT. Cell adhesives and modulators of cell function. Histochem J 1993;25:469 –77. 9. Tabibzadeh S. Evidence of T-cell activation and potential cytokine action in human endometrium. J Clin Endocrinol Metab 1990;71:645–9. 10. La Rocca P, Rheinwald J. Coexpression of simple epithelia keratin and vimentin by human mesothelium and mesothelioma in vivo and in culture. Cancer Res 1984;44:2991–5.
584
Witz et al.
11. Wang XQ, Frazier WA. The thrombospondin receptor CD47 (IAP) modulates and associates with alpha2 beta1 integrin in vascular smooth muscle cells. Mol Biol Cell 1998;9:865–74. 12. Hangan D, Uniyal S, Morris VL, MacDonald IC, von Ballestrem C, Chau T, et al. Integrin VLA-2 (alpha2beta1) function in postextravasation movement of human rhabdomyosarcoma RD cells in the liver. Cancer Res 1996;56:3142–9. 13. Bartolazzi A, Fraioli R, Tarone G, Natali PG. Generation and characterization of the murine monoclonal antibody M-KID 2 to VLA-3 integrin. Hybridoma 1991;10:707–20. 14. Humphries MJ, Mould AP, Tuckwell DS. Dynamic aspects of adhesion receptor function-integrins both twist and shout. Bioessays 1993;15: 391–7. 15. Giancotti FG, Ruoslahti E. Integrin signaling. Science 1999;285:1028 – 32. 16. Aplin JD, Charlton AK, Ayad S. An immunohistochemical study of human endometrial extracellular matrix during the menstrual cycle and first trimester of pregnancy. Cell Tissue Res 1988;253:231– 40. 17. Golombick T, Dajee DR, Bezwoda WR. Extracellular matrix interactions. 1. Production of extracellular matrix with attachment and growthsustaining functions by UWOV2 ovarian cancer cells growing in protein-free conditions. In Vitro Cell Dev Biol 1995;31:387–95. 18. Cannistra SA, Ottensmeier C, Niloff J, Orta B, DiCarlo J. Expression and function of beta 1 and alpha v beta 3 integrins in ovarian cancer. Gynecol Oncol 1995;58:216 –25. 19. van der Linden P, de Goeij A, Dunselman G, van der Linden E, Ramaekers F, Evers J. Expression of integrins and E-cadherin in cells from menstrual effluent, endometrium, peritoneal fluid, peritoneum and endometriosis. Fertil Steril 1994;61:85–90.
Mesothelium ␣21 and ␣31 integrin expression
Vol. 74, No. 3, September 2000