Nina Desai 1 and James Goldfarb

Transcription

1 Human Reproduction vol.13 no.6 pp , 1998 Co-cultured human embryos may be subjected to widely different microenvironments: pattern of growth factor/ cytokine release by Vero cells during the co-culture interval Nina Desai 1 and James Goldfarb Department of Reproductive Biology, University MacDonald Women s Hospital, Case Western Reserve University, Cleveland, Ohio 44106, USA 1 To whom correspondence should be addressed This study was designed to identify and quantify concentrations of growth factors/cytokines released by Vero cells during the co-culture interval. The factors screened for in this preliminary investigation, namely platelet-derived growth factor (PDGF), transforming growth factor β (TGFβ), interleukin-6 (IL-6), leukaemia inhibitory factor (LIF) and epidermal growth factor (EGF) have each been identified to impact on early embryo development or are secreted by embryos themselves, suggesting an autocrine regulatory role. Vero cell culture supernatants were collected at 2, 3, 4, 5 and 6 days after seeding. Samples were assessed by enzyme-linked immunoassay for growth factor/ cytokine secretion at each designated time interval. Conditioned medium from all days contained IL-6, PDGF and LIF. The concentration of IL-6 increased from 294 pg/well on day 2 to almost 1600 pg/well on day 6. PDGF also accumulated rapidly in co-culture wells, rising from pg/well early in the culture period to around 500 pg/ well by day 6. In the second half of this study, medium supernatants from patients enrolled in our co-culture programme were analysed. Retrospective evaluation of medium supernatants collected at the time of transfer from co-cultures from 11 randomly selected patients showed considerable patient-to-patient variation in concentrations of secreted growth factors and cytokines. These findings indicate that during the co-culture interval embryos are exposed to a dynamic environment, with increasing concentrations of growth factors and cytokines. The positive effects of co-culture on embryo quality and in-vitro blastulation need to be balanced against the variation that this technique can potentially introduce into the embryo culture system. Keywords: co-culture/embryo/growth factors/interleukin-6/vero Introduction In recent years, co-culture systems as a means to improve clinical pregnancy rates in human in-vitro fertilization (IVF) programmes have received a good deal of attention. This technique involves culture of human embryos on somatic cell monolayers for 3 6 days before transfer to the patient s uterus. No tissue or species specificity has been demonstrated. Human embryos cultured on oviductal epithelial cells (Bongso et al., 1990; Wiemer et al., 1993), uterine fibroblasts (Wiemer et al., 1989), uterine endometrial cells (Birkenfield and Navot, 1991), kidney epithelial cells (Ménézo et al., 1990; Guerin and Nicollet, 1997) and cumulus cells (Quinn and Margalit, 1996) show improvements in different embryo parameters. The exact mechanism whereby co-culture cells exert their embryotrophic effect is not yet clear. They may help to dilute the effect of any potential growth-inhibiting compounds in the embryo s microenvironment. In addition, during the course of their own proliferation, these somatic cells may be secreting factors which are beneficial to the embryo. Factors that might potentially influence subsequent embryo development include low molecular weight polypeptide growth factors and cytokines. The mitogenic effects of the growth factors platelet-derived growth factor (PDGF), epidermal growth factor (EGF) and transforming growth factors α and β (TGFα and TGFβ) on embryo development in vitro have been demonstrated with several animal species (Paria and Dey, 1990; Dardik and Schultz, 1991; Larson et al., 1992; Lim et al., 1993). In addition to their mitogenic effects, each of these growth factors can influence specific stages of preimplantation development. Leukaemia inhibitory factor (LIF) enhances the in-vitro blastulation of human (Dunglison et al., 1996) as well as animal preimplantation embryos (Robertson et al., 1990; Fry et al., 1992; Marquant-Le Guinne et al., 1993). Conceptuses of several species (Murray et al., 1990; Mathialagan and Roberts, 1994), including humans (Austgulen et al., 1995), secrete the cytokine interleukin-6 (IL-6), raising the possibility that this factor plays an autocrine embryotrophic role during early embryonic development. The present study was designed to identify and quantify concentrations of growth factors and cytokines secreted by coculture cells. Although co-culture systems have been used in clinical IVF programmes since at least 1989 (Wiemer et al., 1989; Bongso et al., 1990; Ménézo et al., 1990), little quantitative information has been available on the nature and quantity of growth factors/cytokines being released during the co-culture interval and their patterns of secretion. In this investigation we focus on the Vero African Green Monkey kidney cell coculture system described initially by Ménézo et al. (1990) for cultivation of human embryos to the blastocyst stage. The factors screened for in this preliminary investigation, namely PDGF, TGFβ, EGF, LIF and IL-6 have each been identified to impact on embryo development, the implantation process or are secreted by embryos themselves, suggesting an autocrine regulatory role European Society for Human Reproduction and Embryology

2 Co-cultured embryos and growth factor/cytokine release Materials and methods Experimental design Vero cell monolayers for this study were set up according to the protocol used in our clinical co-culture programme (see next section) but without human embryos. Cell growth during the culture interval was assessed by trypsinization of monolayers and determination of cell number per well. For each growth factor a series of 10 Nunc 4-well dishes were seeded. Cumulative secretion was measured by harvesting wells in duplicate at 2-, 3-, 4-, 5- or 6-day intervals. Culture supernatants were collected and stored at 20 C until assay. For measurement of daily secretion, culture supernatants were collected in duplicate at 24 h intervals and fresh pre-equilibrated B2 medium was added back to each well. Co-culture programme In our IVF Clinic, co-culture on Vero cell monolayers is performed routinely for patients with a history of two or more failed IVF cycles, previous cycles with poor embryo quality, or aged 39 years. The Vero cell line was purchased from the American Type Culture Collection (ATCC; Rockville, MD, USA). At the time of this study, Vero cell monolayers for patients enrolled in our clinical co-culture programme were prepared once a week. Nunc 4-well dishes containing Ménézo s B2 medium (Fertility Technologies, Natick, MA, USA) were seeded with Vero cells per well. No additional protein supplementation was utilized. Dishes were incubated at 37 C with 5% CO 2 in a humidified chamber. Prior to use, monolayers were extensively washed and fresh pre-equilibrated B2 medium was added to each well. After pronuclear formation, patient embryos were cultivated on Vero cell monolayers (1 4 per well) for 48 h. The best quality embryos were selected for transfer to the patient on day 3. Culture supernatants were collected and stored at 20 C for future assay. Embryos not transferred ( spares ) were left in co-culture wells with fresh medium for an additional 2 3 days and cryopreserved at the blastocyst stage, if of good quality. Assay for growth factors/cytokines Culture supernatants were assayed for PDGF, LIF, EGF, TGFβ and IL-6 using enzyme-linked immunoassay (ELISA) kits from R&D Systems (Minneapolis, MN, USA). Assays were performed following the manufacturer s instructions. The sensitivity of the ELISA assays utilized in this study were as follows: PDGF 8.4 pg/ml, LIF 2.0 pg/ml, IL pg/ml, EGF 0.2 pg/ml and TGFβ 2.0 pg/ml. Unconditioned medium was used for controls and all measurements were made in duplicate. Total growth factor content per well was calculated. Results A typical growth curve for Vero cell monolayers following seeding is illustrated in Figure 1. Cell number increased from cells to approximately cells per well after six days of culture. By day 3, Vero cell monolayers appeared to be confluent, judging by cell growth over the entire well surface. Cultures displayed typical epithelial cell morphology and the borders of adjacent cells were easily distinguished. Although contact-inhibited, Vero cell cultures can become extremely dense and it was noted that cells continued to replicate over the next 3 days, albeit at a much slower rate. Cells were so tightly packed together by day 6 of culture that individual cells were difficult to distinguish and morphologically the cells occupied less surface area, appearing more cuboidal. Figure 1. Pattern of Vero cell growth in Nunc 4-well dishes, following seeding at an initial density of cells/well (n 2). Cultures were set up in duplicate and the cell number per well was determined daily. Vero cell-conditioned medium from all days sampled contained LIF, PDGF and IL-6. EGF and unbound (active) TGFβ were not detected in any of the medium samples screened. Medium supernatants from non-co-cultured embryos at different cell stages for the above factors were consistently negative in the ELISA assays. For this reason, we have not considered embryo-secreted growth factors as major contributors to the total secretions being measured. The minute quantities of factors being secreted by the embryo itself were likely to be outside the limits of detection of the ELISA assays utilized in this study. Figure 2A depicts the pattern of LIF secretion by Vero cell monolayers over the co-culture interval. On days 2 and 3, LIF concentration ranged from 2 to 6 pg/well. By day 6, the mean LIF concentration had increased to 35 pg/well. IL-6 was present at extremely high levels in Vero cellconditioned medium (Figure 2B). The concentration of this cytokine increased from 294 pg/well on day 2 to almost 1600 pg/well by day 6. PDGF also accumulated rapidly in the coculture wells. Early in the culture period on days 2 and 3, PDGF concentration ranged from pg/well. A 25-fold increase in PDGF concentration, to ~500 pg/well, was measured by day 6 of culture. In a separate experiment, the daily secretion of PDGF by monolayers of different ages (i.e. the number of days after initial seeding) was measured (Figure 3). As might be expected, daily PDGF output was highest from older monolayers, where cell density was maximal. Sixday-old Vero cell monolayers secreted up to 115 pg of PDGF into the culture medium in 24 h, while two-day-old Vero cell monolayers secreted only 20 pg PDGF. These data suggested that human embryos placed in co-culture could potentially be exposed to vastly different concentrations of growth factors/ cytokines, depending on the cell density and growth characteristics of the co-culture cells. In the second half of this study, this possibility was addressed by retrospective analysis of medium supernatants from cocultures of 11 randomly selected patients who had been through our clinical co-culture programme. Concentrations of PDGF and IL-6 were measured in medium supernatants collected from co-culture wells on the day of embryo transfer (day 3). 1601

3 N.Desai and J.Goldfarb Figure 2. Growth factor/cytokine accumulation in Vero cellconditioned medium collected over a six-day interval. Co-culture well supernatants were harvested in duplicate on days 2, 3, 4, 5 and 6 and assayed for (A) leukaemia inhibitory factor (LIF) and (B) platelet-derived growth factor (PDGF) and interleukin-6 (IL-6). Figure 4. Assay of (A) platelet-derived growth factor (PDGF) and (B) interleukin-6 (IL-6) concentrations in culture medium from cocultures from 11 randomly selected patients enrolled in our clinical co-culture programme. Vero cell-conditioned medium samples were collected on day 3 at the time of embryo transfer. Table I. Variation in the ratio of interleukin-6 (IL-6) to platelet-derived growth factor (PDGF) secretion in patient co-culture supernatants on the day of embryo transfer (n 11) Patient Ratio IL-6/PDGF Pregnancy outcome Figure 3. Relationship between the age of the monolayer (days after initial seeding) and the amount of platelet-derived growth factor (PDGF) secreted in a 24-h interval. Culture supernatants were collected in duplicate on days 2 to 6. Fresh medium was added back to each well after each sampling. Figure 4A and B summarizes concentrations of PDGF and IL-6 measured in the day 3 supernatants of each co-culture patient. It is evident from this preliminary report that no two patients being co-cultured will likely be exposed to identical concentrations of secreted factors. PDGF concentration was 75 pg/well for three patients, between pg for six patients, and 800 pg for two patients. IL-6 concentration also varied among co-cultures, ranging from 500 to 1000 pg/ well. The ratio of IL-6 to PDGF in each supernatant was also Negative Negative Negative Biochemical Negative Negative Clinical Negative Negative Clinical Negative highly variable. (Table I). LIF was not assayed in these samples as the quantity of culture supernatant available for each patient was limited. In a separate set of ten patient co-culture medium samples, LIF concentrations ranged from 3 to 10 pg/well in seven samples and were undetectable by ELISA in three. Discussion The potential benefit of transferring human embryos at the blastocyst stage has sparked renewed interest in co-culture as

4 Co-cultured embryos and growth factor/cytokine release a means to achieve this end. Despite the abundance of literature on the embryotrophic effects of different somatic cells on embryo development, there are few quantitative data describing levels of various factors being released and their pattern of expression throughout the culture interval (Papaxanthos-Roche et al., 1994; Plachot et al., 1995). Investigators working with both mouse (Paria and Dey, 1990; O Neil, 1997) and human embryos (Moessner and Dodson, 1995) have suggested that autocrine secretion of growth factors may account for the observed improvement in development of embryos cultured in groups versus individually. If the minute quantities of factors secreted by embryos themselves can indeed influence in-vitro development, then the nature of the interaction between growth factors secreted by co-culture cells and the developing preimplantation embryo is worthy of consideration. In the present study, Vero cell culture supernatants were screened for several growth factors and the cytokine IL-6. The cumulative increases in IL-6, PDGF and LIF concentrations with increasing days of culture and cell density, although intuitive, have not been previously described. The dynamic nature of the co-culture environment and the degree of variability that the use of dividing cells can potentially introduce into the embryo culture system is often overlooked. Clinical in-vitro programmes performing co-culture, whether it be with primary tissue such as oviduct or an established cell line like monkey Vero cells, try to organize their tissue culture laboratories such that monolayers are always available for patients as required. To accomplish this, cell cultures are trypsinized once a week and co-culture dishes are seeded. These dishes are then utilized for co-culture of IVF patient embryos for the upcoming week. One consequence of this type of arrangement is that patient embryos cultured for similar time periods on co-culture cells may be subjected to quite different microenvironments depending on when during the week their eggs were retrieved, the age (i.e. days post seeding) of the monolayer when co-culture was initiated, and the growth characteristics of the cell line itself. For instance, in our laboratory, co-culture wells are seeded with Vero cells on Fridays. A patient requiring cells on Sunday (2 days post-seeding) would therefore have their embryos placed on a younger monolayer than a patient requiring cells on say, Thursday (5 days post-seeding). Based on the growth factor secretion curves depicted in Figures 2 and 3, it appears that embryos cultivated on the younger semi-confluent monolayer might initially be exposed to overall lower amounts of growth factors compared with embryos placed on an older, more densely packed monolayer. On this basis, it is postulated that medium supernatants collected from different patients co-cultures at the time of transfer might vary considerably in growth factor content. We were, in fact, able to very clearly illustrate this type of patient-to-patient variation in embryo culture milieu (Figure 4A and B). Retrospective evaluation of day 3 culture media from 11 randomly selected patients, for just two factors (PDGF and IL-6) revealed a wide range of concentrations. The relative amounts of each of these factors in individual culture supernatants was also inconsistent (Table I). In our three patients with positive pregnancy tests, the ratio of IL-6 to PDGF ranged from ~4:1 (patient 7) to ~20:1 (patient 10) and almost 50:1 (patient 4). Whether or not the variations observed in growth factor/cytokine concentrations in culture supernatants have any biological or clinical significance is impossible to interpret at this time. A much more extensive screening for secreted factors and examination of patient- and embryospecific parameters, along with clinical pregnancy outcomes would be necessary to address this question adequately. What is truly striking about the data presented in Figure 4 and Table I is that, at the time of medium collection, patient embryos had been cultured for only 48 h on the Vero cell monolayers and already the preimplantation embryos had experienced differing levels of contact with secreted factors. Considering that these are only two of many known and unknown factors released by co-culture cells, it becomes obvious that the co-culture environment is extremely dynamic and difficult to standardize. Surprisingly, it was also found that Vero cell monolayers arising from seed stocks purchased from different sources at different passage numbers could be quite dissimilar in levels of growth factor/cytokine secreted. Cell lines derived from Vero cell seed stocks purchased from the American Type Culture Collection (ATCC) at passage 121 secreted 2- to 10- fold higher amounts of LIF and IL-6 than cultures initiated from seed stocks obtained from a commercial vendor (BioWhittaker, Walkersville, MD, USA) at passage 161 or higher (unpublished data). Repeated subpassaging of a cell line prior to embryo co-culture has been noted to yield poorer results (Ménézo et al., 1992) and may be related to slower cell division and ascribed to slower cell division. Reduction in levels of secreted factors with slower cell division may be one mechanism by which repeated subpassaging might alter co-culture performance. The contribution of secreted growth factors/cytokines to the subsequent development of human embryos placed on coculture requires careful analysis. Data from several animal species indicate that the effects of growth factors and cytokines on embryos can be both cell stage and concentration dependent. Optimal and inhibitory levels, if any, of LIF, IL-6 and PDGF for human embryo culture have yet to be determined. Conflicting results have been obtained with LIF supplementation of human embryo culture media. Jurisicova et al. (1995), in testing recombinant human LIF (5 20 ng/ml) in serum-supplemented Hams F-10 medium, were unable to demonstrate any improvement in embryo development. In contrast, Dunglison et al. (1996) reported significantly enhanced rates of blastocyst formation by human embryos cultured in serum-free medium supplemented with LIF (1000 IU/ml or approximately pg/ml). The concentrations of LIF utilized in both of these studies were at least 1000-fold higher than those measured in our Vero cell co-culture system. The function of LIF in the in-vitro-developing blastocyst may be to maintain the proliferation of the stem cell pool, that is, the cells comprising the inner cell mass (Stewart, 1994; Nichols et al., 1996). Messenger RNA transcripts for the LIF receptor have been identified in both human and mouse blastocysts (Nichols et al., 1996; Van Eijk et al., 1996). LIF receptor transcripts were shown to be localized mainly in the inner cell 1603

5 N.Desai and J.Goldfarb mass, whereas LIF expression was in the trophectoderm (Nichols et al., 1996). These data suggest a paracrine interaction in the blastocyst, such that cells secreting LIF activate receptors on the inner cell mass. Secretion of LIF by co-culture cells may mediate a similar response. It could be further speculated that provision of supplemental LIF either through co-culture cells or direct addition to culture medium may only be beneficial if the developing blastocyst is itself unable to produce this factor, either due to suboptimal in-vitro culture conditions or poor embryo quality. The cytokine IL-6 was the most actively secreted factor detected in the Vero cell culture medium and its role in early embryo development is the least understood. Our laboratory first reported the secretion of large amounts of this cytokine by Vero cells in an earlier study (Desai and Goldfarb, 1996), in which the secretions of Vero cells and an endometrial cell line with embryotrophic properties were compared. Subsequent testing of this cytokine on murine IVF embryos, showed that both blastulation and hatching could be modulated in a dose-dependent fashion with IL-6 supplementation of culture medium (Desai et al., 1996). The effect of PDGF at early stages of human embryonic development has not yet been directly assessed, but human chorionic gonadotrophin secretion by in-vitro-cultured blastocysts can be stimulated by culture medium supplementation with this factor (2000 pg/ml) (Lopata and Oliva, 1993). Detection of PDGF by radioimmunoassay in culture supernatants from blastocyst stage human embryos (Svlander et al., 1991) and identification of mrna transcripts for this factor and its receptor in human 8-cell, morula and blastocyst stage embryos (Osterlund et al., 1996) are certainly indicative of an autocrine function for this factor. The milestone studies of Thibodeaux and colleagues (1993) and Larson et al. (1992) define a role for PDGF in mediating the transition from maternal to embryonal transcription in in-vitro-cultured bovine embryos. Interestingly, in this animal model, prolonged exposure to supplemental PDGF (1000 pg/ml) after genomic activation, negatively affected subsequent blastocyst formation. The possibility of a negative effect of high levels of coculture cell-secreted growth factors has to our knowledge never been raised. Our experience has been that blastulation of untransferred human embryos ( spares ) is poorer if coculture is initiated on very densely populated Vero cell monolayers (i.e. 6 days old). While the current findings demonstrate differences in overall growth factor levels, it is not possible at this time to distinguish between any negative effect due to high concentrations of accumulated growth factors and negative effects from a too-rapid accumulation of Vero cell metabolic by-products in the culture milieu. Direct analysis of individual growth factor effect on human embryos in a cell-free system will ultimately prove necessary to delineate their role in early development. In conclusion, the positive attributes of co-culture, namely increased embryo quality and blastocyst formation, need to be balanced against the potential variation that this technique can introduce to the in-vitro culture environment. This study has screened for Vero cell-secreted factors and described their pattern of secretion during the co-culture interval. It is evident 1604 from this preliminary report that no two patients undergoing co-culture will be likely to have their embryos exposed to identical levels of secreted factors. Clinical laboratories utilizing co-culture need to become more conscious of the changing profile of growth factor secretion by the selected cell line and the relationship to cell proliferation. It is clear that alternatives to co-culture are needed to allow for greater standardization of embryo culture media. The identification and quantification of growth factor and cytokine secretion by co-culture cells may be a first step towards understanding the basic growth requirements of the preimplantation human embryo and may ultimately allow the design of specific culture milieu for optimal in-vitro growth of human embryos to the blastocyst stage. References Austgulen, R., Arntzen, K., Vatten, L. et al. (1995) Detection of cytokines (interleukin-1, interleukin-6, transforming growth factor-beta) and soluble tumor necrosis factor receptors in embryo culture fluids during in vitro fertilization. Hum. Reprod., 10, Birkenfield, A., and Navot, D. (1991) Review: endometrial cultures and their application to new reproductive technologies: a look ahead. J. In Vitro Fert. Embryo Transfer, 8, Bongso, A., Ng, S. and Ratnam, S. (1990) Co-cultures: their relevance to assisted reproduction. Hum. Reprod., 5, Dardik, A. and Schultz, R. (1991) Blastocoel expansion in the preimplantation mouse embryo: stimulatory effect of TGF and EGF. Development, 113, Desai, N. and Goldfarb, J. (1996) Growth factor/cytokine secretion by a permanent human endometrial cell line with embryotrophic properties. J. Assist. Reprod. Gen., 13, Desai, N., Scarrow, M., Kinzer, D. and Goldfarb, J. (1996) Design of a new culture system using human extracellular matrix and interleukin-6. Am. Soc. Reprod. Med. (Suppl.), 166. Dunglison, G., Barlow, D. and Sargent, L. (1996) Leukemia inhibitory factor significantly enhances the blastocyst formation rates of human embryos cultured in serum-free medium. Hum. Reprod., 11, Fry, R., Batt, P., Fairclough, R. and Parr, R. (1992) Human leukemia inhibitory factor improves the viability of cultured ovine embryos. Biol. Reprod., 46, Guerin, J. and Nicollet, B. (1997) Interest of co-cultures for embryos obtained by in-vitro fertilization: a French collaborative study. Hum. Reprod., 12, Jurisicova, A., Ben-Chetrit, A., Varmuzu, S. and Casper, R. (1995) Recombinant human leukemia inhibitory factor does not enhance in vitro human blastocyst formation. Fertil. Steril., 64, Larson, R., Ignotz, G. and Currie, W. (1992) Platelet derived growth factor (PDGF) stimulates development of bovine embryos during the fourth cell cycle. Development, 115, Lim, J., Bongso, A. and Ratnam, S. (1993) Mitogenic and cytogenetic evaluation of transforming growth factor-beta on murine preimplantation embryonic development in vitro. Mol. Reprod. Dev., 36, Lopata, A. and Oliva, K. (1993) Chorionic gonadotrophin secretion by human blastocysts. Hum. Reprod., 8, Marquant-Le Guinne, B., Humbolt, P., Guillon, N. and Thibier, N. (1993) Murine LIF improves the development of IVF cultured bovine morula. J. Reprod. Fertil., 12, 61. Mathialagan, N., Bixby, J. and Roberts, R. (1994) Expression of interleukin- 6 in porcine, and bovine preimplantation conceptuses. Mol. Reprod. Dev., 32, Ménézo, Y., Guerin, J. and Czyba, J. (1990) Improvement of human early embryo development in vitro by co-culture on monolayer of Vero cells. Biol. Reprod., 42, Ménézo, Y., Hazout, A., Dumont, M. et al. (1992) Co-culture of embryos on Vero cells and transfer of blastocysts in humans. Hum. Reprod., 7, Moessner, J. and Dodson, W. (1995) The quality of human embryo growth is improved when embryos are cultured in groups rather than separately. Fertil. Steril., 64, Murray, R., Lee, F. and Chiu, C. (1990) The genes for leukemia inhibitory

6 Co-cultured embryos and growth factor/cytokine release factor and interleukin-6 are expressed in mouse blastocysts prior to the onset of hemopoiesis. Mol. Cell. Biol., 10, Nichols, J., Davidson, D., Tetsuya, T. et al. (1996) Complementary tissuespecific expression of LIF and LIF-receptor mrnas in early mouse embryogenesis. Mech. Dev., 57, O Neill C. (1997) Evidence for the requirement of autocrine growth factors for development of mouse preimplantation embryos in vitro. Biol. Reprod., 56, Osterlund, C., Wramsby, H. and Pousetto, A. (1996) Temporal expression of platelet-derived growth factor (PDGF-A) and its receptor in human preimplantation embryos. Mol. Hum. Reprod., 2, Papaxanthos-Roche, A., Taupin, J., Mayer, G. et al. (1994) Human interleukin for DA cells or leukemia inhibitory factor is released by Vero cells in human embryo co-culture. Fertil. Steril., 63, Paria, B. and Dey, S. (1990) Preimplantation embryo development in vitro: cooperative interactions among embryos and role of growth factors. Proc. Natl Acad. Sci. USA, 87, Plachot, M., Heyman, D. and Godard, A. (1995) Embryo coculture: the interaction embryo feeder. In IXth World Congress on In Vitro Fertilization and Assisted Reproduction, Monduzzi, Bologna, pp Quinn, P. and Margalit, R. (1996) Beneficial effects of coculture with cumulus cells on blastocyst formation in a prospective trial with supernumerary human embryos. J. Assist. Reprod. Gen., 13, Robertson, S., Lavranos, T. and Seamark, R. (1990) In vitro models of the maternal fetal interface. In Wegman, T., Nisbett-Brown, E. and Gill, T. (eds), The Molecular and Cellular Immunology of the Maternal-Fetal Interface. pp Oxford University Press, New York. Stewart, C. (1994) Leukemia inhibitory factor and the regulation of preimplantation development of the mammalian embryo. Mol. Reprod. Dev., 39, Svlander, P., Holmes, P., Olovsson, M. et al. (1991) Platelet-derived growth factor is detected in human blastocyst culture medium but not in human follicular fluid a preliminary report. Fertil. Steril., 56, Thibodeaux, J., Del Vecchio, R. and Hansel, W. (1993) Role of plateletderived growth factor in development of in vitro matured and in vitro fertilized bovine embryos. J. Reprod. Fertil., 98, Van Eijk, M.J.T., Mandelbaum, J., Salat-Baroux, J. et al. (1996) Expression of leukemia inhibitory factor receptor subunits LIFRβ and gp130 in human oocytes and preimplantation embryos. Mol. Hum. Reprod., 2, Wiemer, K., Cohen, J., Wiker, S. et al. (1989) Coculture of human zygotes on fetal bovine uterine fibroblasts: embryonic morphology and implantation. Fertil. Steril., 52, Wiemer, K., Hoffman, D., Maxson, W. et al. (1993) Embryonic morphology and rate of implantation of human embryos following co-culture on bovine oviductal epithelial cells. Hum. Reprod., 8, Received on May 30, 1997; accepted on March 20,