Autologous granulosa cell coculture demonstrates zygote suppression of granulosa cell steroidogenesis*t

Transcription

1 FERTILITY AND Copyright 1996 American Society for Reproductive Medicine Vol. 66, No.3, September 1996 Printed on acid-free paper in U. S. A. Autologous granulosa cell coculture demonstrates zygote suppression of granulosa cell steroidogenesis*t David B. Seifer, M.D.* Melanie R. Freeman, M.Sc. Alice C. Gardiner, M.L.T.II George A. Hill, M.D. Alan L. Schneyer, Ph.D.~ Barbara C. Vanderhyden, Ph.D.**tt Brown University School of Medicine, Women's and Infants Hospital of Rhode Island, Providence, Rhode Island; Centennial Medical Center, Nashville, Tennessee; Massachusetts General Hospital, Boston, Massachusetts; and Ottawa Regional Cancer Centre, University of Ottawa, Ottawa, Ontario, Canada Objective: To determine if embryos can modulate steroid hormone production by luteinized granulosa cells. Design: Granulosa cells obtained from follicular aspirates were cultured alone or in the presence of a two-pronuclear zygote. The production ofe2 and P by these cultures was evaluated by RIA. Setting: In Vitro Fertilization Unit in an academic research environment. Patients: Sixteen women undergoing IVF. Interventions: Standard IVF-ET treatment cycle using leuprolide acetate for pituitary desensitization before hmg or urofollitropin for ovarian stimulation. Main Outcome Measures: Estradiol and P concentration in culture media of luteinized granulosa cells alone or granulosa cells cocultured with a two-pronuclear embryo. Results: Both E2 and P production by luteinized granulosa cells was reduced when cultured in the presence of an embryo. Conclusions: Human embryos secrete a factor that regulates granulosa cell steroidogenesis. Fertil Steril 1996;66: Key Words: Granulosa cell steroidogenesis, autologous granulosa cell coculture, embryo development, estradiol, progesterone Suboptimal embryo culture conditions may be responsible in part for the low pregnancy rate after Received December 7,1995; revised and accepted May 7,1996. * Presented in part at the 51st Annual Meeting ofthe American Society for Reproductive Medicine. Seattle, Washington, October 7 to 12, t Supported in part by Physician Scientist Award (AG00566 [D.B.S.] and R01HD31894 [A.L.S.]) from the National Institutes of Health, National Institute on Aging, Bethesda, Maryland. :j: The Ohio State University Medical Center, Division of Reproductive Endocrinology and Infertility, Department of Obstetrics and Gynecology. Center for Assisted Reproduction and Reproductive Endocrinology, Centennial Medical Center. II Division of Reproductive Endocrinology, Department of Obstetrics & Gynecology, Brown University School of Medicine; and Women and Infants Hospital of Rhode Island. ~ Reproductive Endocrine Unit and National Center for Infertility Research, Massachusetts General Hospital. ** Ottawa Regional Cancer Centre; and Departments of Medicine, Physiology, and Obstetrics and Gynecology, University of Ottawa. assisted reproductive technologies. Autologous granulosa cell co culture has been shown to enhance human embryo development, implantation, and pregnancy rates for IVF (1, 2). There is considerable evidence to suggest that the oviduct and the endometrium produce polypeptide growth factors that can control preimplantation development of the mammalian embryo (3, 4), and it can be postulated that the mechanism(s) by which granulosa cells enhance embryo development in autologous coculture is through the production of similar growth factors. Human embryos have been shown to express receptors for several growth factors, including the receptors for colony-stimulating factor-i, stem cell factor, interleukin-6, leukemia inhibitory factor, and tumor t t Reprint requests: Barbara C. Vanderhyden, Ph.D., Ottawa Regional Cancer Centre, Cancer Research Laboratories, 3rd Floor, 501 Smyth Road, Ottawa, Ontario, Canada, K1H 8L6 (FAX: ). Vol. 66, No.3, September 1996 Seifer et al. Zygotic inhibition of granulosa steroidogenesis 425

2 necrosis factor (5), which may mediate the effects through paracrine signaling. In addition to the paracrine effects of the surrounding somatic cells on preimplantation embryo development, human embryos have been shown to coexpress some growth factors (tumor necrosis factor-a, stem cell factor, and interleukin-6) and their receptors, an arrangement suggesting the autocrine enhancement of embryo development (5-7). Alternatively, embryonic growth factors may act on the surrounding endometrium to influence its receptivity to implantation. Although embryonic effects on somatic cells of the reproductive tract have not been demonstrated, evidence is accumulating to support a role for the oocyte in the regulation of granulosa cell activity. Numerous studies have demonstrated that loss of association with the oocyte alters the ability of granulosa cells to undergo proliferation (8, 9), differentiation (9), cumulus cell expansion (9-12), hyaluronic acid secretion (10, 11, 13), and expression of the plasminogen activator gene (14). In addition, oocyte factors appear to play an important role in the regulation of granulosa cell steroidogenesis (15-17). As a model, autologous embryo-granulosa cell coculture offers the unique opportunity to study communication between human somatic and germ cells and allows us to investigate the possibility that the human embryo can influence the activity of somatic cells. Therefore, we examined the hypothesis that steroid hormone production by luteinized granulosa cells is influenced by the presence of embryos when grown in coculture. Patient Population MATERIALS AND METHODS Sixteen women undergoing IVF, aged 32.6 ± 1.1 years (mean ± SEM), with a mean day 3 serum FSH of6.0 ± 0.2 miu/ml (BioMerieux-Vitek, Hazelwood, MO; enzyme-linked fluoroimmunoassay [conversion factor to SI unit, 1.00]). Etiologies of infertility consisted of 62.5% tubal, 18.8% endometriosis, 12.5% anovulation, and 6.2% unexplained. The autologous granulosa cell coculture studies were approved by the Centennial Medical Center Institutional Review Board. Stimulation Protocol Women received leuprolide acetate (LA, Lupron; TAP Pharmaceutical, North Chicago, IL) for pituitary desensitization during the follicular phase of the cycle. Leuprolide acetate was given twice daily at a dose ranging from 0.05 to 0.1 ml for a minimum of 1 week, after which serum E2 was obtained to assess whether adequate suppression had been achieved. When the serum E21evel was <16 pg/ml (conversion factor to SI unit, 3.671), one or two ampules ofhmg (Pergonal; Serono, Randoph, MA) containing 75 IU of FSH and 75 IU of LH and/or one or two ampules of urofollitropin (Metrodin; Serono) containing 75 IU offsh and 1 IU oflh were given 1M in the morning and evening in addition to half the daily dose of LA. Follicular development was monitored with a Toshiba SAL 77B (Toshiba, Tokyo) equipped with a 5- MHz transvaginal probe. When the serum E2 level was 330 pg/ml and two or more follicles measured 2::16 mm in mean diameter, 10,000 IU ofhcg were injected 1M. Transvaginal oocyte retrieval was performed 35 hours after hcg injection. Granulosa Cell Preparation Granulosa cell cultures were prepared as described previously (2). Briefly, granulosa cells in follicular aspirates were allowed to sediment (without centrifugation) until loose pellets of cells were visible (15 to 20 minutes). Red blood cells remained in suspension and granulosa cells were pipetted from the bottom of the tubes and pooled. The pooled cells were rinsed three times by resuspension in 3 ml Ham's F-lO (GIBCO BRL Life Technologies, Grand Island, NY) containing 10% maternal serum followed by 10 to 15 minute sedimentations at unit gravity. A hemocytometer cell count and a viability assessment by Trypan blue exclusion (range 85% to 95% viable) were made. Granulosa cell monolayers were prepared from patients after follicular aspiration and were plated at 100,000 cells in 2 ml culture medium containing 20% maternal serum per dish 24 hours before coculture. Most single cells and small clusters had attached and plated after 16 to 20 hours, resulting in an actively growing monolayer covering 40% to 50% of the culture area. Residual blood cells and unattached cells were removed by gentle flushing of culture medium over the cell surface with a transfer pipette followed by two rinses with culture medium. Equilibrated culture medium (1 ml) containing 15% maternal serum was added to each dish and a single two-pronuclear zygote was moved onto the cell layers for coculture incubation. For each patient, two granulosa cell-zygote cocultures were established and compared with granulosa cells cultured alone. Media was removed and replaced with fresh media at 48 and 120 hours of incubation and samples were analyzed for P and E2. For each coculture, the embryo was present throughout the 120 hours. 426 Seifer et al. Zygotic inhibition of granulosa steroidogenesis Fertility and Sterility

3 Table 1 Comparison of Geometric Means for P and E2 Concentrations in Media From Embryo-Granulosa Cell Cocultures and From Granulosa Cells Cultured Alone* Embryo with granulosa cellst Granulosa cells onlyt P (ng/ml) 48 h 90.2 (55.5 to 146.6) (660.7 to 966.1) 120 h 94.2 (53.5 to 165.6) (571.5 to 997.7) E2 (pg/ml) 48 h 58.2 (31.6 to 107.4) (81.8 to 311.9) 120 h 57.7 (28.5 to 116.9) (46.7 to 257.0) * Values in parentheses are 95% confidence intervals. t Cocultures established with one zygote and 100,000 granulosa cells isolated from follicular aspirates from 16 patients. Two cocultures were established per patient, with the values for the two cultures averaged before statistical analysis. t Cultures of 100,000 granulosa cells isolated from follicular aspirates of the same 16 patients; one culture per patient. P values derived from comparison of means using one-way analysis of variance. Hormonal Assays P Estradiol and P (Diagnostic Systems Laboratory Inc., Webster, TX) concentrations in the media samples were determined in duplicate by RIA. Assay standard curves were validated using Ham's F-I0 with 10% fetal calf serum. Intra-assay and interas say coefficients of variance were < 10% for all assays. The sensitivities of the assays were 8.0 pg/ ml for E2 and 0.12 ng/ml (conversion factor to SI unit, 3.180) for P. As no steroid substrates were added to the cultures, the measurements reflected basal levels of steroid production. Statistical Analysis Each patient served as her own control, with one culture containing only granulosa cells and two cultures containing one zygote with granulosa cells. The E2 and P concentrations for the two cocultures were averaged and the mean was used in statistical analysis. All hormonal concentrations were log transformed before analysis because of log normal distribution of the data. Mean ± SEM was calculated for the remaining parameters. Means were compared using one-way analysis of variance. Statistical significance was assumed at P : RESULTS Concentrations of P and E2 in the media from granulosa cells cultured in the presence or absence of zygotes are summarized in Table 1. Coculture of zygotes with granulosa cells resulted in an 8-fold decrease in P and a 1.5- to 3-fold decrease in E2 compared with granulosa cells cultured alone. Within each treatment group, concentrations of steroids at 48 and 120 hours were not different. DISCUSSION The data presented here suggest that intercellular communication between human somatic and germ cells may be mediated humorally. Although the beneficial effects of somatic cells on human embryo development have been described previously (1,2), this is the first observation to suggest that human germ cells may influence granulosa cell activity. The importance of this observation that human embryos affect luteinized granulosa cell activity may not be apparent immediately. These two cell types are in close proximity in vivo only for a limited period during very early embryo development, and the effects of the embryos on P production seen here are contrary to the corpora luteal production of P that normally occurs during early embryo development. However, we have found that mouse oocytes produce a P-inhibiting factor at several stages of development, from midgrowth to full-grown stages as well as ovulated oocytes and zygotes (Vanderhyden BC, unpublished data). Thus, we would speculate that production of this factor by early human embryos is simply a continuation of the constitutive production of this factor during oocyte development. The possibility that human oocytes-embryos may secrete factors into the culture media, particularly factors whose bioactivity is evaluated easily, would provide the opportunity to develop novel, noninvasive methods by which the health of oocytes in culture can be assessed. A regulatory role for the oocyte in granulosa cell luteinization was initially proposed long ago by EI Fouly et al. (18), who showed that surgical removal of oocytes from antral follicles in vivo resulted in luteinization of granulosa cells and elevated serum P levels. More recently, the development of the micromanipulation technique of oocytectomy (9, 10), which removes the oocyte from an oocyte-cumulus cell complex, has enabled investigators to confirm these initial observations and demonstrate that mammalian oocytes secrete factors that regulate numerous granulosa cell activities, including steroidogenesis. Murine and porcine oocytectomized complexes have been shown to secrete higher concentrations of P compared with intact oocyte-cumulus complexes, but the P production is inhibited when the oocytectomized complexes are cocultured with cumulus-free oocytes or in the presence of oocyte-conditioned media (15-17). These experiments indicate that an oocyte-secreted factor is responsible for suppression of granulosa P production, an observation that is in accordance with the results we have obtained with human embryos. Normal follicular maturation and ovulation require a hormonal milieu in which P levels are kept Vol. 66, No.3, September 1996 Seifer et a!. Zygotic inhibition of granulosa steroidogenesis 427

4 low. Premature luteinization of the granulosa cells is detrimental to the development of the oocyte and can prevent ovulation (19, 20). The ability of human embryos to inhibit P production suggests that the germ cells may play an active role in regulating their environment. It can be speculated that defects that impair the ability of the oocyte to produce this regulatory factor may contribute to conditions that potentially cause infertility, such as luteinized unruptured follicle syndrome or unexplained anovulatory cycles. The presence of an embryo in the cultures of granulosa cells appeared to inhibit the production of both P and E2. This inhibitory effect on E2 production is similar to the inhibition seen with porcine oocytes and granulosa cells (17) but is unlike the stimulatory effect seen when mouse oocytes are cultured with granulosa cells (15). The inhibition of both P and E2 suggests that the embryo-secreted factor generally may act to suppress steroidogenesis, although the mechanism of this inhibition is unknown. Although this is the most likely explanation, particularly in light of the studies performed in other species, it should be noted that our results do not exclude the possibilities of steroid uptake and retention or metabolism by the embryos. ;:' We currently are investigating possible candidates for the embryo-produced factor that suppresses granulosa cell steroidogenesis. Production of some cytokines by the embryo has been shown to be regulated in a stage-specific manner; therefore, clues regarding the possible identity of the steroid-regulating factor may be obtained by comparison with the stage-specific expression of known growth factors, although this information currently is very limited. There is, however, some evidence to indicate that human embryos can secrete growth factors. Tumor necrosis factor-a has been detected in the medium of human embryos cultured at all stages of preimplantation development, except as blastocysts (21). Using total RNA extracted from human embryos, interleukin-6 was found to be transcribed preferentially in blastocysts (5). In contrast, stem cell factor transcripts were detected at the two-cell stage, and again at the six- to eight-cell and morula stages of development (5). Although the biochemical nature of the steroid-regulating factor in this study is unknown, the factor was produced by embryos cultured from the zygote to the eight-cell stage (0 to 48 hours culture) and to the expanded blastocyst stage (48 to 120 hours culture), suggesting that the factor is produced at all stages of pre implantation development. An alternative strategy to determine the identity of the embryo-secreted factor is to compare its activity with known regulators of human granulosa cell steroidogenesis. In cultured human granulosa cells, activin A is able to suppress basal and gonadotropininduced progestin accumulation (22-24). However, our preliminary experiments investigating members of the transforming growth factor (TGF) (3 family as possible candidates for the steroid-regulating factor indicate that TGF-(3, inhibin, and TGF-a concentrations were not different between granulosa cells cultured alone as compared with those cocultured with zygotes, and activin was undetectable using the mesoderm induction activity of activin as a sensitive bioassay (25). Therefore, these factors are unlikely to be the putative factor. Although this study clearly demonstrates that embryos can influence the activity of somatic cells of the ovarian follicle, possible effects on the somatic cells of the reproductive tract have yet to be investigated. Identification of such factors produced by the germ cells and investigation of the mechanisms by which germ cells influence granulosa cell activity may provide tools that can be used to quantitatively evaluate embryo health, improve embryo culture conditions, and/or enhance the receptivity of the uterus for implantation. Acknowledgment. Many thanks to Jeffrey V. May, Ph.D., for performing assays for TGF-.tJ in preliminary experiments. REFERENCES 1. Plachot M, Antoine JM, Alvarez S, Firmin C, Pfister A, Mandelbaum J, et al. Granulosa cells improve human embryo development in vitro. Hum Reprod 1993; 12: Freeman MR, Whitworth CM, Hill GA. Granulosa cell coculture enhances human embryo development and pregnancy rate following in vitro fertilization. Hum Reprod 1995;2: Pampfer S, Arecci RJ, Pollard IW. Role of CSF-1 and other lymphatic factors in mouse preimplantation development. Bioessays 1991; 13: Tartakovsy B, Ben Yair E. Cytokines modulate preimplantation development and pregnancy. Dev BioI 1991; 146: Sharkey AM, Dellow K, Blayney M, Macnamee M, Charnock Jones S, Smith SK. Stage-specific expression of cytokine and receptor messenger ribonucleic acids in human preimplantation embryos. BioI Reprod 1995;53: Zolti M, Ben-Rafael Z, Meirom K, Shemesh M, Bider D, Mashiach S, et al. Cytokine involvement in oocytes and early embryos. Fertil Steril 1991; 56: Witkin S, Liu H-C, Davis OK, Rosenwatz Z. TNF is present in maleval sera in embryo culture fluids during IVF. J Reprod Immunol 1991; 19: Vanderhyden BC, Telfer EE, Eppig JJ. Mouse oocytes promote proliferation of granulosa cells from preantral and antral follicles in vitro. BioI Reprod 1992;46: Vanderhyden BC, Caron PJ, Buccione R, Eppig JJ. Developmental pattern of the secretion of cumulus expansion-enabling factor by mouse oocytes and the role of oocytes in promoting granulosa cell differentiation. Dev BioI 1990; 140: Buccione R, Vanderhyden BC, Caron PJ, Eppig JJ. FSH-in- 428 Seifer et at. Zygotic inhibition of granulosa steroidogenesis Fertility and Sterility

5 duced expansion of the mouse cumulus oophorus in vitro is dependent upon a specific factor(s) secreted by the oocyte. Dev BioI 1990; 138: Salustri A, Yanagishita M, Hascall VC. Mouse oocytes regulate hyaluronic acid synthesis and mucification by FSH-stimulated cumulus cells. Dev BioI 1990; 138: Prochazka R, Nagyova E, Rimkevicova Z, Nagai T, Kikuchi K, Motlik J. Lack of effect of oocytectomy on expansion of the porcine cumulus. J Reprod Fertil1991;93: Salustri A, Ulisse S, Yanagishita M, Hascall VC. Hyaluronic acid synthesis by mural granulosa cells and cumulus cells in vitro is selectively stimulated by a factor produced by oocytes and by transforming growth factor-.b. J BioI Chern 1990;265: Canipari R, Epifano 0, Siracusa G, Salustri A. Mouse oocytes inhibit plasminogen activator production by ovarian cumulus and granulosa cells. Dev BioI 1995; 167: Vanderhyden BC, Cohen IN, Morley P. Mouse oocytes regulate granulosa cell steroidogenesis. Endocrinology 1993; 133: Vanderhyden BC, Tonary AM. Differential regulation of progesterone and E2 production by mouse cumulus and mural granulosa cells by a factor(s) secreted by the oocyte. BioI Reprod 1995;53: Coskun S, Uzumcu M, Lin YC, Friedman CI, Alak BM. Regulation of cumulus cell steroidogenesis by the porcine oocyte and preliminary characterization of oocyte-produced factor(s). BioI Reprod 1995;53: EI-Fouly MA, Cook B, Nekola M, Nalbandov AV. Role of the ovum in follicular luteinization. Endocrinology 1970;87: Katz E. The luteinized unruptured follicle and other ovulatory dysfunctions. Fertil Steril1988;50: Check JH, Chase HS, Nowroozi K, Dietterich CJ. Premature luteinization: treatment and incidence in natural cycles. Hum Reprod 1991;6: Lachappelle MH, Miron P, Hemmings R, Falcone T, Granger L, Bourque J, et al. Embryonic resistance to tumour necrosis factor-a mediated cytotoxicity: novel mechanism underlying maternal immunological tolerance to the fetal allograft. Hum Reprod 1993;8: Rabinovici J, Spencer SJ, Doldi N, Goldsmith PC, Schwall R, Jaffe RB. Activin-A as an intraovarian modulator: action, localization, and regulation of the intact dimer in human ovarian cells. J Clin Invest 1992;89: Li W, Yuen BH, Leung PCK. Inhibition of progestin accumulation by activin-a in human granulosa cells. J Clin Endocrinol Metab 1992; 75: DiSimone ND, Lanzone A, Petraglia F, Ronsisvalle E, Caruso A, Mancuso S. Effect of activin-a on progesterone synthesis in human luteal cells. Fertil Steril1994;62: Melton DA. Pattern formation during animal development. Science 1991;252: Vol. 66, No.3, September 1996 Seifer et al. Zygotic inhibition of granulosa steroidogenesis 429