Coculture of mouse embryos with cells isolated from the human ovarian follicle, oviduct, and uterine endometrium*t

Transcription

1 FERTILITY AND STERILITY Copyright or.> 1993 The American Fertility Society Vol. 59. No.1. January 1993 Printed on acid-free paper in U.S.A. Coculture of mouse embryos with cells isolated from the human ovarian follicle, oviduct, and uterine endometrium*t Melanie R. Freeman, M.Sc.:j: M. Cristina Bastias, M.D. George A. Hill, M.D. Kevin G. Osteen, Ph.D. Center for Fertility and Reproductive Research, Department of Obstetrics and Gynecology, Vanderbilt University School of Medicine, Nashville, Tennessee Objective: To examine the specificity of somatic cell support by comparing embryonic development during long-term in vitro coculture with feeder cells derived from the human ovarian follicle, oviduct, and endometrium. Design: Comparative study of murine embryo development and degeneration during 6 days of in vitro coculture. Results: All feeder-cell cultures were beneficial to embryonic development and viability. Few differences were observed between feeder cell types (epithelial or fibroblastic) or cell origin (ovarian follicle, oviductal, or endometrial). Embryos developed to the eight-cell stage in 24 hours whether in coculture (83.6% to 100%) or in media alone (85.2%); however, further development in media alone decreased compared with coculture (15.6% versus 63.4% to 87.7%, plating) and embryo degeneration increased (67.9% versus 5.5% to 19.4%) after 6 days. Conclusions: [1) Coculture of embryos with human reproductive tract cells is beneficial to embryonic development and viability. [2) Human somatic cell support of murine embryos during longterm in vitro coculture is not tissue specific nor dependent on cell type. Fertil Steril 1993;59: Key Words: Coculture, endometrial cells, oviductal cells, cumulus cells, granulosa-lutein cells, murine embryos Oocytes and embryos developing in vivo are nourished by the cells and fluid of the ovarian follicle, the fallopian tube, and the uterine endometrium. Received January 13, 1992; revised and accepted September 28,1992. * Supported by Biomedical Research Support, grant #RR from Vanderbilt University, Nashville, Tennessee, and grant #HD from The National Institutes of Health, Bethesda, Maryland. t Presented in part at the 47th Annual Meeting of The American Fertility Society, Orlando, Florida, October 19 to 24, :j: Reprint requests and present address: Melanie R. Freeman, M.Sc., Center for Assisted Reproduction and Reproductive Endocrinology, 2222 State Street, Suite D, Nashville, Tennessee Present address: Center for Assisted Reproduction and Reproductive Endocrinology, Nashville, Tennessee. Several nutrient culture media have been formulated to mimic the composition of these fluids (1, 2) and, provided that these media are supplemented with purified protein or protein-rich biological fluid such as serum (3), are able to meet the basic nutrient requirements for short-term in vitro growth for embryos of some species. However, these media fall short of providing an optimal embryo growth environment and may be a contributing factor to the disappointingly low pregnancy rates for in vitro fertilization-embryo transfer (IVF-ET) (4). Recent advances in embryo culture have been documented in studies of mammalian embryo coculture with somatic cells of reproductive and nonreproductive tissue origin (3, 5-8). These studies have consistently demonstrated improvement in embryo growth during coculture in comparison with 138 Freeman et ai. Human cell/mouse embryo coculture

2 1 standard culture methods. However, no study, to date, has examined the specificity of somatic cell support by comparing the embryotrophic effects of several distinct feeder-cell populations originating from the human female reproductive tract. This study was designed to examine two-cell murine embryo growth and degeneration during 6 days of coculture with six human reproductive tract cell populations. MATERIALS AND METHODS Feeder-Cell Density and Culture Medium This study was conducted with prior approval of, and in accordance with, the Vanderbilt University Committee for the Protection of Human Subjects and the Vanderbilt University Institutional Animal Care and Use Committee. To determine the optimal cell density for coculture experiments, uterine stromal cells were plated at 2 X 10 4 and at 6 X 10 4 cells/60-mm dish in a phenol red-free mixture (1:1) of Dulbecco's modified Eagle's medium (DMEM) and Ham's F-12 (Sigma Chemical Co., St. Louis, MO) + 10% fetal calf serum (FCS; GIBCO, Grand Island, NY). Media were changed to serum-free DMEM/Ham's F-12 after 24 hours. The ph of the media and the percentage confluence of the cultures were determined on the 8th day. To determine the optimal culture medium for coculture experiments, three commercially prepared culture media, Ham's F-I0 (GIBCO), Earle's balanced salt solution (EBSS; GIBCO), and DMEM/ Ham's F -12 were examined for their ability to maintain feeder cells in culture and for their ability to support murine embryo growth during coculture without serum supplements. Uterine stromal cells were plated at 6 X 10 4 /60-mm culture dish in media containing 10% FCS and were changed to serumfree media after 24 hours. Cell growth was monitored for 8 days. Two-cell mouse embryos were added on the 3rd day of culture and the number of hatching blastocysts was determined 4 days later. Initiation of Cell Cultures Cells were isolated from normally cycling women of reproductive age (31 to 43 years of age) during the stage in the menstrual cycle that would most likely coincide with the presence of egg(s) and embryo(s) in vivo, that is, harvesting during the follicular phase for granulosa-lutein cells (GC) and cumulus cells and during the secretory phase for oviductal and endometrial cells. A comparison of cells from many different patients was not within the scope of this study. Endometrial Cells Uterine cells were isolated from endometrial biopsies of three normally cycling women during the secretory phase of the menstrual cycle. The stromal (fibroblast) and epithelial fractions were separated as previously described (9) and plated in DMEM/ Ham's F % FCS and grown to confluency. Purity of isolated cell preparations was assessed by immunohistochemical localization of cytokeratins and vimentin (9), and <5% contamination of separated cells was routine. Oviductal Cells Excised oviducts from a 31-year-old patient were obtained during the secretory phase of the menstrual cycle. Oviductal cells were purified using a modification of the procedure for isolation of endometrial cell types (9). Epithelial and fibroblast cells were plated separately and grown to confluency. The epithelial-enriched fraction contained predominantly ciliated and nonciliated cuboidal cells, whereas the fibroblast-enriched fraction contained <10% cuboidal ciliated and nonciliated cells. No further characterization of cell purity was determined. Ovarian Follicle Cells Granulosa-lutein cells were obtained from a 37- year-old patient and cumulus cells were recovered from five patients (32 to 43 years of age) after the aspiration of oocytes for IVF-ET. All patients had undergone leuprolide acetate-human menopausal gonadotropin-stimulated cycles. The GCs were prepared according to the procedure previously described (10) and were grown to confluency. For cumulus cells, oocytes were cultured in Ham's F-lO containing 15% maternal serum and were inseminated with 125,000 sperm/ml. Oocytes were removed from the dish 18 hours later and the dissociated cumulus cells were incubated an additional 48 hours. The cells were rinsed and were cultured for 1 to 2 weeks before coculture with mouse embryos. Preparation of Feeder Layers for Coculture For coculture experiments endometrial and oviductal cells and GCs were plated at 6 X 10 4 cells/ 60-mm dish. All cells were rinsed to remove serum Vol. 59, No.1, January 1993 Freeman et al. Human cell/mouse embryo coculture 139

3 before coculture with mouse embryos in serum-free DMEM/Ham's F-12. Cumulus cells were not replated and the cell number in each dish was not determined. Preparation of Mouse Embryos Female mice (B6C3Fl; Harlan, Indianapolis, IN) were injected with 5 IV pregnant mare serum (Sigma Chemical Co.) and 5 IV human chorionic gonadotropin (hcg; Sigma Chemical Co.) 50 hours later and mated with CD-l males (Charles River, Raliegh, NC). Two-cell embryos were flushed from the oviducts 42 hours after hcg injection and distributed evenly between test dishes. Mouse Embryo Development Embryo development and degeneration were observed with a dissecting microscope every 24 hours during the 6-day culture period. The numbers of twocell, three- to seven-cell, eight-cell, morula, and four stages of blastocyst (expanding, hatching, attachment to the dish, and plating or trophoblastic outgrowth) were determined. Embryo development was assessed from the proportion of eight-cell embryos (after 24 hours of culture), morula (after 48 hours), expanding blastocysts (after 72 hours), hatching blastocysts (after 96 hours), attachment to the dish (after 120 hours), and plating (after 144 hours). The number of degenerating embryos was also determined. Early cleavage stage embryos were classified as degenerate if >25% of each embryo contained cytoplasmic fragments or if blastomeres appeared dark and granular. Morula and blastocyst stage embryos were considered degenerate if they collapsed or appeared dark and granular. Statistical Analysis Embryos from each animal were equally distributed between the culture dishes in each experiment, there being four to eight replicates per experiment. Differences in embryonic development and degeneration between cultures were compared with Fisher's exact test. RESULTS Feeder-Cell Density and Culture Medium The optimal initial cell density was determined to be 6 X 10 4 cells per dish. Cultures plated at 6 X 10 4 cells per dish reached 90% confluency after 8 days, whereas cultures plated at 2 X 10 4 cells per dish were 75% confluent after 8 days. The ph of the media did not vary significantly between the lower (ph 7.39) and higher (ph 7.36) cell densities. When cell growth in serum-free DMEM/Ham's F-12, EBSS, and Ham's F-I0 was observed, growth in EBSS and Ham's F-I0 was sluggish, and cells began to slough-off after 4 days, compared with 90% confluency in DMEM/Ham's F-12 after 8 days. Blastocyst hatching was supported equally well by DMEM/Ham's F-12 (76/106, 71.7%), Ham's F-I0 (80/107, 74.8%), and EBSS (84/102, 82.4%). Because DMEM/Ham's F-12 was the only media able to support the feeder-cell monolayer for the 8-day culture period, it was chosen as the media base for the experiments in this study. Embryonic Development in Cocultures and Control Embryonic development (Table 1) was enhanced by all feeder-cell cultures from 48 to 144 hours of culture compared with the media-alone controls; however, few significant differences were observed between cocultures. For the first 24 hours, embryo development in cumulus-cell co culture exceeded the other cocultures and the control. Likewise, at 120 hours, embryo development in GC coculture exceeded the other cocultures and the control. No significant differences in embryo development between cocultures were seen at any other observation time. Embryo Degeneration in Cocultures and Control Co culture with feeder-cell monolayers was associated with reduced embryo degeneration (Table 2) from 48 to 144 hours compared with culture in media alone. No significant differences in embryo degeneration were observed between cocultures. DISCUSSION In this study, long-term culture of mouse embryos was monitored to examine somatic cell specificity during coculture and to quantitate effects of cultural conditions that might relate to short-term culture during human IVF-ET. The presence of human somatic cells of reproductive-tract origin was associated with increased development and decreased degeneration of two-cell murine embryos during long-term in vitro culture. This effect was observed regardless of the feeder-cell origin within the female reproductive tract (ovarian follicle, oviduct, or endometrium) or the cell type (fibroblastic or epithelial). 140 Freeman et al. Human cell/mouse embryo coculture

4 Table 1 Comparison of Embryo Development in Control and Cocultures* Developmental stage after hours in culture 8-Cell Morula Treatment (n) 24 h 48 h Control (128) 109 (85.2) 104 (81.3)t Ovarian follicle cells Cumulus (134) 134 (100):j: 131 (97.8) GC (73) 61 (83.6) 70 (95.9) Oviductal cells Fibroblast (71) 65 (91.6) 67 (9404) Epithelial (71) 60 (84.5) 71 (100) Endometrial cells Stromal (83) 71 (85.5) 82 (98.8) Epithelial (65) 59 (90.8) 63 (96.9) Values are the sum of values from 4 to 8 replicates; percent in parentheses. t Controls are significantly lower than cocultures, P < 0.008, 48 hours; P < ,72 to 144 hours; Fisher's exact test. Expanded Hatching Attached Plating 72 h 96 h 120 h 144 h 75 (58.6)t 45 (35.2)t 25 (19.5)t 20 (15.6)t 123 (91.8) 114 (85.1) 91 (67.9) 94 (70.2) 64 (87.7) 67 (91.8) 64 (87.7) 64 (87.7) 66 (93.0) 57 (80.3) 50 (7004) 53 (74.7) 61 (85.9) 55 (77.5) 44 (62.0) 45 (6304) 77 (92.8) 63 (75.9) 57 (68.7) 66 (79.5) 55 (84.6) 51 (78.5) 42 (64.6) 48 (73.9) :j: Significantly higher than all other 24-hour cultures, P < 0.003; Fisher's exact test. GC cultures significantly higher than all other 120-hour cultures, P < 0.01; Fisher's exact test. Although the overall objective of this study was to optimize in vitro embryo culture, certain conditions beneficial to embryo growth can be identified within constraints of otherwise suboptimal culture conditions. In this respect, DMEM/Ham's F-12 is a suboptimal embryo cultures medium but is widely used for the initiation of primary cell cultures and effectively supports cell subcultures and for this reason was chosen as the medium base for this study. The mechanism of embryotrophic action(s) of somatic cells is not completely understood. Effects may result, in part, from the removal of inhibitory or toxic substances by the feeder cells from the culture media (3) either inherent in the formula or produced during embryo metabolism. Responses may also re- flect production and release of metabolic intermediates or growth factors by the feeder cells during embryo development (3, 5). The production of proteins, peptides, or lipids by feeder cell mono layers (3,5,7) and cell-to-embryo contact (3, 11) have also been cited as potential factors in somatic cell support of embryo development. A range of peptide growth factors have been identified in secretions of reproductive tract tissues and, therefore, may contribute to responses established in the present study. The factors include insulin-like growth factor(s), epidermal growth factor, transforming growth factora, and various isotypes of transforming growth factor-13 (12, 13). Properties of specific cellular products that may affect embryo development have not yet been defined. Table 2 Comparison of Embryo Degeneration in Control and Cocultures Hours in culturet Treatment (n) Control (168) 19 (11.3):j: 61 (36.3):j: Ovarian follicle cells Cumulus (134) 0(0) 1 (0.8) GC (73) 0(0) 0(0) Oviductal cells Fibroblast (71) 0(0) o (0) Epithelial (71) 0(0) 0(0) Endometrial cells Stromal (83) 0(0) o (0) Epithelial (65) 0(0) 0(0) * Values are the sum of values from 4 to 8 replicates; percent in parentheses. t No degeneration was observed after 24 hours of culture (47.0):j: 102 (60.7):j: 114 (67.9):j: 3 (2.2) 21 (15.7) 26 (1904) 3 (4.1) 4 (5.5) 4 (5.5) 1 (104) 8 (11.3) 11 (15.5) 0(0) 7 (9.9) 8 (11.3) 1 (1.2) 6 (7.2) 11 (13.5) 1 (1.5) 6 (9.2) 10 (1504) :j: All controls are significantly higher than cocultures, P < 0.006, 48 hours; P < , 72 to 144 hours; Fisher's exact test. Vol. 59, No.1, January 1993 Freeman et al. Human cell/mouse embryo coculture 141

5 In summary, mouse embryo development and viability were maintained effectively during coculture with cells from the human female reproductive tract. Enhanced development was observed regardless of the origin of feeder cells or the cell type. Although the mechanisms are not yet understood, coculture of embryos with somatic cell supports was beneficial to embryo development and to viability of embryos during extended culture in the system described. Similar modification to a culture system used with human embryos during IVF-ET might also be advantageous. However, with respect to differences in the chronology of early embryonic development between species (14), care must be taken when extrapolating experiences from one species to another. REFERENCES 1. Quinn P, Kerin JF, Warnes GM. Improved pregnancy rate in human in vitro fertilization with the use of a medium based on the composition of human tubal fluid. Fertil Steril1985;44: Tervit HR, Whittingham DG, Rowson LEA. Successful cui ture in vitro of sheep and cattle ova. J Reprod FertiI1972;30: Eyestone WH, First NL. Co-culture of early cattle embryos to the blastocyst stage with oviducal tissue or in conditioned medium. J Reprod Fertil 1989;85: Medical Research International, Society for Assisted Reproductive Technology (SART), The American Fertility Society. In vitro fertilization embryo transfer (lvf -ET) in the United States: 1990 results from the IVF-ET Registry. Fertil Steril 1992;57: Wiemer K, Amborski G, Denniston R, Godke R. Use of a hormone-treated fetal uterine fibroblast monolayer for in vitro development for bovine embryos. Theriogenology 1987;27: Wiemer KE, Cohen J, Amborski GF, Wright G, Wiker S, Munyakazi L, Godke RA. In-vitro development and implantation of human embryos following culture on fetal bovine uterine fibroblast cells. Hum Reprod 1989;4: Gandolfi F, Moor R. Stimulation of early embryonic development in the sheep by co-culture with oviduct epithelial cells. J Reprod FertiI1987;81: Ouhibi N, Hamidi J, Guillaud J, Menezo Y. Co-culture of 1- cell mouse embryos on different cell supports. Hum Reprod 1990;5: Osteen KG, Hill GA, Hargrove JT, Gorstein F. Development of a method to isolate and culture highly purified populations of stromal and epithelial cells from human endometrial biopsy specimens. Fertil Steril 1989;52: Hill GA, Herbert CM III, Wentz AC, Osteen KG. Use of individual human follicles to compare oocyte in vitro fertilization to granulosa cell in vitro luteinization. Fertil Steril 1987;48: Allen R, Wright R. In vitro development of porcine embryos in co-culture with endometrial cell monolayers or culture supernatants. J Anim Sci 1984;59: Morrish DW, Bhardwaj D, Dabbagh LK, Marusyk H, Siy o. Epidermal growth factor induces differentiation and secretion of human chorionic gonadotropin and placental lactogen in human placenta. J Clin Endocrinol Metab 1987;65: Skinner MK, Keski-Oja J, Osteen KG, Moses HL. Ovarian thecal cells produce transforming growth factor-beta which can regulate granulosa cell growth. Endocrinology 1987;121: Tesarik J. Developmental control of human preimplantation embryos: a comparative approach. J In Vitro Fert Embryo Transf 1988;5: Freeman et al. Human cell/mouse embryo coculture