Hung-Ching Liu, Ph.D., Zhi-Ying He, M.S., Carol A. Mele, B.A., Lucinda L. Veeck, M.L.T., D.Sc., Owen Davis, M.D., and Zev Rosenwaks, M.D.

Transcription

1 FERTILITY AND STERILITY VOL. 71, NO. 2, FEBRUARY 1999 Copyright 1999 American Society for Reproductive Medicine Published by Elsevier Science Inc. Printed on acid-free paper in U.S.A. Human endometrial stromal cells improve embryo quality by enhancing the expression of insulin-like growth factors and their receptors in cocultured human preimplantation embryos Hung-Ching Liu, Ph.D., Zhi-Ying He, M.S., Carol A. Mele, B.A., Lucinda L. Veeck, M.L.T., D.Sc., Owen Davis, M.D., and Zev Rosenwaks, M.D. The Center for Reproductive Medicine and Infertility, The New York Hospital-Cornell Medical Center, New York, New York Objective: To demonstrate the mechanism by which human endometrial stromal cells improve embryo quality in coculture. Design: Randomized study. Setting: Academic research center. Patient(s): Patients undergoing IVF-ET. Intervention(s): Donated human embryos were cultured randomly either alone (group A) or with human endometrial stromal cells (group B), and the embryonic expression of insulin-like growth factors (IGFs) and their receptors was detected by reverse transcriptase polymerase chain reaction after culture. Main Outcome Measure(s): The embryo frequency distribution of groups A and B before and after culture and the embryonic transcripts of the IGF family genes of the two study groups after culture were compared. Result(s): The embryo frequency distribution of the day 3 embryonic stages in groups A and B was not different. However, after culture, a statistically significant difference in blastocyst formation was observed between groups A and B. A significant increase in the expression of IGF-1, IGF-2, the IGF-1 receptor, and the insulin-receptor also was noted. Among the embryos that reached the blastocyst stage, the expression of IGF-1 and the IGF-1 receptor also was significantly different in the two study groups. Conclusion(s): Human endometrial stromal cells enhanced the expression of IGFs and their receptors in cocultured human embryos, which may be essential for improving embryo quality. (Fertil Steril 1999;71: by American Society for Reproductive Medicine.) Key Words: Coculture, gene expression, IGFs, IGF receptors, RT-PCR Received June 5, 1998; revised and accepted October 4, Presented at the 10th World Congress on In Vitro Fertilization and Assisted Reproduction, Vancouver, British Columbia, Canada, May 24 28, Reprint requests: Hung- Ching Liu, Ph.D., Cornell University Medical College, 515 East 71st Street, Room S-500, New York, New York (FAX: ; hcliu@mail.med.cornell.edu) /99/$20.00 PII S (98) Embryo quality is essential for the success of IVF-ET. Unfortunately, current in vitro human embryo culture conditions have proven to be suboptimal, which may be a contributing factor to the low implantation rates noted after IVF-ET (1 3). Extensive investigations have been conducted in various areas in an attempt to improve embryo quality. These include altering in vitro culture medium and supplemental growth factors (4, 5) and using modern coculturing technology (6 10). Coculture techniques that use different types of monolayer cells as helper cells have been known to enhance embryo viability in various mammalian species, including humans (6 10). We have shown that both vero cells (monkey kidney cells) (11) and human secretory endometrial stromal cells (ESCs) (12) are effective helper cells. Endometrial stromal cells are the physiologic cells for implantation and have been studied thoroughly. We demonstrated previously that ESCs enhance mouse embryo viability (12, 13). Embryo outgrowth rates were significantly enhanced by the presence of ESCs in the control medium as well as the medium with exogenous insulin-like growth factor 1 (IGF-1; 20 g/ml) (13). The outgrowth rate was increased by the 361

2 presence of ESCs from 55.8% 78.5% in the control medium and from 58% 94.5% in the medium with the exogenous IGF-1 (13). This finding led us to use coculture clinically for improving implantation rates after IVF-ET. However, concerns regarding labor, cost, and donor tissue contamination with the human immunodeficiency virus and other infectious agents must be addressed before this new technology is introduced in conventional IVF-ET. With the advantage of preserving ESCs by cryopreservation, we elected to perform autologous endometrial coculture to avoid donor tissue problems. The ESCs were isolated from tissue biopsy samples that were obtained from patients before IVF-ET treatment and stored in liquid nitrogen until used for coculture. With the use of autologous coculture in patients who had had at least one previous failed IVF attempt, we proved that ESCs enhance human embryo quality (14). When the cohort of preembryos were assigned randomly to growth in a conventional medium or a coculture, the coculture appeared to reduce fragmentation and the cleavage rate. Transfer of the embryos with the best morphologic quality resulted in a dramatic increase in the pregnancy rate compared with previous conventional culture cycles (14). Many other investigators also have reported that coculture is effective in a clinical setting. However, the mechanism by which human ESCs improve embryo quality remains relatively unknown. In coculture, helper cells may enhance embryo development by secreting embryotrophic factors or by removing toxins and other inhibitory substances. We previously demonstrated that the major benefit of ESCs was their secretion of proteins that promote early embryo development (12). Numerous cytokines and growth factors have been found in endometrial secretory proteins that may serve as growthpromoting factors for human embryos. The IGF family is comprised of two polypeptide ligands (IGF-1 and IGF-2), their complementary receptors, and a group of binding proteins that regulate IGF bioactivity (15). This IGF system has been studied recently in the field of reproductive biology. Transcripts of IGFs and their receptors were detected in ovarian cells, oviductal cells, and ESCs (16 18). Thus, these cells may produce IGFs to nourish preembryos in the early stages of development. In animal studies, exogenous IGFs were found to increase embryo cleavage, blastocyst formation, and outgrowth rates in vitro (12, 16, 19). With the use of Western blot analysis, we identified many IGF-related factors in condition medium preexposed to ESCs (13). We suspect that IGFs may be one of the factors in endometrial secretory products that contribute to embryo growth in coculture. Further, with the use of semiquantitative reverse transcriptase polymerase chain reaction (RT-PCR), we were able to detect simultaneously the transcripts of IGFs and their receptors on individual mouse and human preembryos (20). Therefore, we postulated that ESCs may produce IGFs to promote the growth of embryos by acting through their receptors on the preembryos. It is of interest that the secretion of these growth factors by ESCs increased with the presence of embryos in culture (13). In this study, we examined whether the presence of ESCs, modifies the expression of IGFs and their receptors on cocultured embryos. MATERIALS AND METHODS Human Embryos The human embryos used in our study were donated by patients undergoing IVF-ET at The Center for Reproductive Medicine and Infertility in New York. The study was approved by the Committee on Human Rights in Research of The New York Hospital-Cornell Medical Center Institutional Review Board in accordance with the Helsinki Declaration of 1989 on human experimentation. After retrieval, the best-quality embryos were selected for transfer or frozen embryo storage. Only embryos rejected for use in IVF-ET were donated by patients. These spare embryos were of poor quality. In general, they either had 6 blastomeres after 72 hours of in vitro culture or demonstrated 25% fragmentation at any given stage. Occasionally, a few frozen embryos were donated by patients who did not intend to become pregnant again. Thirty-seven day 3 human preembryos were used in the study. They were allocated randomly according to their day 3 embryonic stage either to culture alone (group A) or to coculture with human ESCs isolated from endometrial biopsy samples (group B) (21) in Ham s F-10 medium (GIBCO, Grand Island, NY) plus 10% human serum in vitro for an additional 3 days. In addition, two six-cell, frozen, donated embryos also were used in the study and were divided into two groups. The embryonic stage of the day 3 and day 6 embryos as well as their embryonic growth potential in vitro was recorded. The embryonic growth potential in vitro was defined as arrested if the embryo exhibited a reduction or no change in the number of blastomeres or if it became degenerated or vacuolated from day 3 to day 6 of culture, as exhibiting slow growth if the embryo exhibited an increase in the number of blastomeres, and as active if the embryo reached the morula or blastocyst stage on day 6 (22). Ribonucleic Acid Isolation, Reverse Transcription, Polymerase Chain Reaction To elucidate the mechanism by which human ESCs improve embryo quality, the expression of IGFs and their receptors in whole embryos cultured alone or cocultured with human ESCs was compared. Cultured embryos were washed in phosphate-buffered saline until free of culture medium, treated with proteinase to remove the zona pellucida, and then washed again in phosphate-buffered saline 362 Liu et al. Mechanism of coculture Vol. 71, No. 2, February 1999

3 before being subjected to semiquantitative RT-PCR for the detection of IGFs and their receptors (17). Briefly, RNAs of individual embryos were extracted by phenol-chloroform-isoamylalcohol (25:24:1) and further purified through deoxyribonuclease treatment, phenol-chloroform-isoamylalcohol extraction, and ethanol precipitation. The purified RNA was used for complementary DNA synthesis by Moloney murine leukemia virus reverse transcriptase and random hexanucleotide or an oligo (deoxythymidine) primer. Subsequently, the synthesized complementary DNA was amplified by PCR with Taq DNA polymerase and specific primers. The amplified products then were separated on 3% agarose gel and visualized by ethidium bromide staining for qualitative identification of specific transcripts. The intensity of the ethidium bromide fluorescence was measured with a photographic imaging system for semiquantitative analysis. The intensity of a specific stained product was related to the number of templates, shown by the linear increase in the product amplified with rabbit globin messenger RNA templates from pg per tube (Fig. 1). External Controls To ensure the consistency of the assays, external controls with fixed amounts of rabbit globin messenger RNA (1 pg per tube) were run simultaneously with our samples. To monitor the interassay and intra-assay coefficients of variation and the recovery rate, some controls were run with exact sequential steps for extraction and RT-PCR and some controls were run with RT-PCR directly (Fig. 2). With the use of our highly reliable method, the intra-assay and interassay coefficients of variation were 7% and 10%, respectively, and the recovery rate was 90% (23). Limitations of Semiquantitative Analysis Although our semiquantitative RT-PCR can be used to compare the relative amounts of specific transcripts in different samples, it cannot be used to compare the relative amounts of different transcripts in a single sample, owing to the fact that all transcripts were amplified in different amplification conditions and the amplicons were different in size (19). Therefore, we only compared the expression of the same gene between the two study groups. Statistical Analysis For embryo distribution, the number of embryos in each category according to their embryonic stage or growth potential in vitro was analyzed by 2 test. The log-linear module of the software package Statistica (Statsoft, Tulsa, OK) was used. The marginal frequency was calculated. The observed frequencies and the expected frequencies were compared with Pearson s 2 statistic. For gene expression, the median test and the Wald-Wolfowitz Runs test were used for comparison. The median test has the null hypothesis that all samples come from populations with identical medians. FIGURE 1 Correlation of amplified product with amount of template used. Known amounts of rabbit globin messenger RNA (mrna) templates were used for amplification by RT-PCR as described in the text to show the linear relation between the amount of template and the amplified product. Top: Identification of the products by sizing on a 3% agarose gel. The amounts of rabbit globin mrnas used for amplification in lanes 1, 2, 3, 4, 5, and 6 were 0.125, 0.25, 0.5, 1, 2, and 4 pg per tube, respectively. Lane M represents the markers. Bottom: Linear relation between the amplified products and rabbit globin mrna up to 2 pg per tube. The dotted line is the exploration of the linear line. bp base pairs. Results were calculated by finding the overall median of the combined groups and counting the number of values in each separate group above and below the overall median. Then, the expected values of the counts were compared with the observed values with the 2 statistic. The Wald-Wolfowitz Runs test looked at the null hypothesis that two independent samples were drawn from two populations that were the same in some respect. P.05 was defined as statistically significant. RESULTS When analyzed by Pearson s 2 test, the frequency distribution of day 3 embryos according to their embryonic stage was no different in groups A and B (P.84) (Fig. 3). FERTILITY & STERILITY 363

4 FIGURE 2 Method for simultaneous detection of multiple gene expression in individual embryos. The messenger RNAs (mrnas) from individual embryos were extracted by detecting the expression of multiple genes as described in the text. To ensure the consistency of the assays, external controls with fixed amounts of rabbit globin mrna (rgmrna (1 pg per tube) were run simultaneously with the samples. To monitor the inter- and recovery rates, some of the controls were run with exact sequential steps for extraction and RT-PCR and some controls were run with RT-PCR directly. IGF-I insulin-like growth factor 1; IGF-II insulin-like growth factor 2; IGF-IR insulin-like growth factor 1 receptor; IGF-IIR insulin-like growth factor 2 receptor; insulin-r insulin receptor. FIGURE 3 Frequency distribution of day 3 embryos according to their embryonic stage. Forty-nine donated day 3 human preembryos were assigned randomly to culture alone (group A, ) or to coculture with human endometrial stromal cells isolated from endometrial biopsy samples (Group B, ) in Ham s F-10 medium plus 10% human serum in vitro. However, after culture, a statistically significant difference in blastocyst formation was observed between the two study groups (18.2% and 51.8% in groups A and B, respectively; P.04). The frequency distribution of day 6 embryos according to their embryonic stage also was significantly different (P.032) (Fig. 4). Sixty-seven percent (18/27) of the embryos in the cocultured group were at stages greater than or equal to the morula stage, whereas 68% (15/22) of the embryos cultured alone remained at the 16-cell stage. Again, after culture, the frequency distribution of the embryos according to their embryo growth potential in vitro was significantly different (P.012) in the two study groups. A significantly higher percentage of embryos in the cocultured group (78.9%) had a higher embryonic growth potential in vitro compared with that of the embryos cultured alone (31.8%) (Fig. 5). Accompanying the increase in blastocyst formation and growth potential in vitro, there was a significant increase in the gene expression of IGFs and their receptors in the cocultured group. The expression of IGF-1, IGF-2, the IGF-1 receptor, and the insulin receptor was significantly greater in the cocultured embryos. The expression of the IGF-2 receptor also was greater, albeit not significantly different, in the cocultured embryos compared with the embryos cultured alone (Fig. 6). It is of interest that, among the embryos that reached the blastocyst stage, the mean expression of IGFs and their receptors was higher in the cocultured group than in the group cultured alone. In particular, the expression of IGF-1 and the IGF-1 receptor was significantly higher in the cocultured group than in the group cultured alone (Fig. 7). DISCUSSION Coculture has been used to overcome the deficiencies of conventional in vitro embryo culture conditions in IVF-ET. 364 Liu et al. Mechanism of coculture Vol. 71, No. 2, February 1999

5 FIGURE 4 Frequency distribution of day 6 embryos according to their embryonic stage. All embryos in group A ( ) and group B ( ) were cocultured for an additional 3 days in vitro. Their day 6 embryonic stages then were recorded and analyzed by Pearson s 2 test. Conventional in vitro embryo culture conditions are known to be suboptimal, as demonstrated by the high incidence of abnormal embryos, slow cleavage rate, and low blastocyst formation rate, as well as the significantly lower pregnancy rate compared with that of GIFT. Coculture in which various types of monolayer cells are used recently has been used to enhance embryo viability in mammalian species, including humans. Sheep oviduct (8), bovine fibroblast (6), African green monkey kidney (vero) FIGURE 5 Frequency distribution of embryos according to their growth potential in vitro. The embryonic growth potential in vitro, defined in the text, of group A ( ) and group B ( ) also was recorded and analyzed by Pearson s 2 test. (7, 11), human cumulus (9), human ampullar (10), and human endometrial (12, 13) cells have been shown to improve human preembryo quality by increasing the number of blastomeres, decreasing the percentage of fragmentation, increasing blastocyst formation, and increasing the clinical pregnancy rate (6 13). However, contradictory results also have been obtained (24). Factors such as the type of helper cells used, the type of medium used, the ph of the medium, the type of supplemental proteins used, cell density, the use of cells in active growth or stationary phases, and the end points of evaluation affect the outcome of cocultures. To confirm a positive effect of coculture, all these factors must be considered and monitored carefully. Human ESCs, the physiologic cells for embryo growth, are effective helper cells and are unique in their ability to produce growth factors for embryo development. They markedly enhance embryo viability and growth rates (12, 13). The ability of these cells to enhance embryo growth is increased further when the cells are pretreated with progesterone and relaxin. Embryos that are cocultured with hormone-treated cells are good-quality embryos with a high potential for advanced development (12, 13). The major benefit of human ESCs is that they enhance the secretion of embryotrophic factors and/or proteins (12). To avoid the risk of disease transmission, autologous ESCs were used in a group of select patients in whom repeated attempts (average of 3.4) at IVF-ET had failed. The results were encouraging (14). In this study, we further confirmed that human ESCs effectively enhance the rate of blastocyst formation (Fig. 3) FERTILITY & STERILITY 365

6 FIGURE 6 Gene expression of embryos after culture alone or coculture with endometrial stromal cells. The expression of multiple genes was detected in embryos after culture alone ( ) or coculture with endometrial stromal cells ( ). The data are expressed as means SD. The expression of insulin-like growth factor 1 (IGF-I), and insulin-like growth factor 2 (IGF-II), insulin-like growth factor 1 receptor (IGF-IR), and insulin receptor (Insulin-R) was significantly higher in the cocultured embryos. * P.05 when the specific gene expression of embryos cultured alone and embryos cocultured was compared. IGF-IIR insulin-like growth factor 2 receptor. and the in vitro growth potential. Blastocyst formation was extremely high in the cocultured embryos (51.8%) despite the fact that most of the embryos studied were rejected spare embryos that were of generally poor quality. The rate of blastocyst formation could be even higher if coculture were used routinely for embryo culture after retrieval. A high rate of blastocyst formation in coculture offers the opportunity for blastocyst transfer to improve pregnancy rates. Accompanying the increase in blastocyst formation, there was a significant increase in the embryonic expression of IGF-1, IGF-2, the IGF-1 receptor, and the insulin receptor in FIGURE 7 Gene expression of embryos that reached the blastocyst stage after culture. The expression of insulin-like growth factor 1 (IGF-I), insulin-like growth factor 2 (IGF-II), insulin-like growth factor 1 receptor (IGF-IR), insulin-like growth factor 2 receptor (IGF-IIR), and Insulin receptor (Insulin-R) in embryos that reached the blastocyst stage was determined and compared. The expression of IGF-I and IGF-IR was significantly higher in the cocultured embryos ( ) than in the embryos cultured alone ( ). 366 Liu et al. Mechanism of coculture Vol. 71, No. 2, February 1999

7 the coculture. Because IGFs and insulin have been demonstrated to stimulate embryo development (12, 19, 25), the enhanced expression of IGFs and their receptors may be essential for improving embryo quality. As shown in Figure 6, both cocultured embryos and embryos cultured alone exhibited enormous fluctuations in the expression of all genes. This may be due to the fact that both study groups were composed of embryos at various embryonic stages, as illustrated in Figure 4. It is surprising that when expression was controlled for embryos that reached the blastocyst stage, great variation in gene expression was still noted among these embryos. In general, the mean expression of all genes was higher in cocultured embryos than in embryos cultured alone. However, the expression of IGF-1 and the IGF-1 receptor was significantly higher in the cocultured group. This may suggest that IGF-1 and the IGF-1 receptor play an important role in early embryo development. However, our data did not show a statistically significant difference in the expression of IGF-2 and the IGF-2 receptor between the two study groups. Whether this is due to the limited sample size or to different bioactivity of IGF-1 and IGF-2 requires further study. Because the high expression of embryonic IGFs and their receptors often was associated with embryos of high potential growth in vitro (22), the variation in gene expression among embryos at similar developmental stages (i.e., blastocyst) may reflect differences in their quality. The blastocysts in the cocultured group usually had better morphology, with more cells in the inner cell mass revealed by differential staining (data not shown). We now have demonstrated that in coculture, ESCs act directly on embryos to augment the expression of both embryonic growth factors and their receptors. We previously had shown that the presence of embryos in coculture enhanced the secretion of embryotrophic growth factors by ESCs (13). Thus, we strongly believe that IGFs and their receptors may be involved in embryo-endometrium interaction. In coculture, embryos may send their embryonic signals to stimulate ESCs to secrete IGFs. These IGFs then may trigger their biological activity after binding to their receptors on the embryo. In conclusion, human ESCs in coculture not only secrete the essential growth factors for embryo development but also enhance embryonic expression of these factors and their receptors to function as autocrine and paracrine regulators for improving embryo quality. Our data strongly suggest that IGF-1 and the IGF-1 receptor play an important role in early embryo development and that their expression on the embryo may serve as a marker for embryo quality. References 1. Barnea ER, Check JH, Grudzinskas JG, Maruo T. Implantation and early pregnancy in humans. In: Liu HC, editor. Advanced technologies improve embryo implantation after IVF-ET. London: The Parthenon Publishing Group, Liu HC. Recent advances to improve human implantation. Assist Reprod Tech/Androl 1993;4: Cohen J, Alikani M, Liu HC, Rosenwaks Z. Rescue of human embryos by micromanipulation. Baillieres Clin Obstet Gynaecol 1994;8: 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 Steril 1985;44: Fitzgerald L, DiMattina M. An improved medium for long term culture of human embryos overcomes the in vitro developmental block and increases blastocyst formation. Fertil Steril 1992;57: Wiemer K, Cohen J, Wiker S, Malter H, Wright G, Godke R. Coculture of human zygotes on fetal bovine uterine fibroblast: morphology and implantation. Fertil Steril 1989;52: Menezo Y, Guerin J, Czyba J. Improvement of human embryo development in vitro by co-culture on monolayers of vero cells. Biol Reprod 1990;42: Gandolfi F, Brezini J, Richardson L, Brown C, Moor R. Characterization of proteins secreted by sheep oviduct epithelial cells and their function in embryonic development. Development 1987;106: Quinn P, Margalit R. Beneficial effects of coculture with cumulus cells on blastocyst formation in a prospective trial with supernumerary human embryos. J Assist Reprod Genet 1996;13: Bongso A, Ng SC, Fong CY, Anandakumar C, Marshall B, Edirisinghe R, et al. Improved pregnancy rate after transfer of embryos grown in human fallopian tubal cell coculture. Fertil Steril 1992;58: Lai Y, Stein D, Soong Y, Tang Y, Grifo J, Malter H, et al. Evaluation of vero cell co-culture system for mouse embryos in various media. Hum Reprod 1992;7: Liu HC, Mele CA, Noyes N, Rosenwaks Z. Endometrial secretory proteins enhance early embryo development. J Assist Reprod Genet 1994;11: Liu HC, Mele C, Catz D, Noyes N, Rosenwaks Z. Production of insulin-like growth factor binding protein (IGFBPs) by human endometrial stromal cells is stimulated by the presence of embryos. J Assist Reprod Genet 1995;12: Barmat L, Liu HC, Spandorfer S, Kowalik A, Mele C, Xu KP, et al. Autologous endometrial co-culture in patients with repeated failures of implantation after IVF-ET. J Assist Reprod Genet. In press. 15. Giudice LC, Milkowski DA, Lamson G, Rosenfeld RG, Irwin JC. Insulin-like growth factor binding proteins in human endometrium: steroid-dependent messenger ribonucleic acid expression and protein synthesis. J Clin Endocrinol Metabol 1991;72: Adashi EY, Resnick CE, Hurwitz A, Ricciarellie E, Hernandez ER, Roberts CT, et al. The intra-ovarian IGF system. Growth Regul 1992; 2: Carsson B, Hillensjo T, Nilsson A, Tornell J, Billig H. Expression of insulin-like growth factor-i (IGF-I) in the rat fallopian tube: possible autocrine and paracrine action of fallopian tube-derived IGF-I on the fallopian tube and on the preimplantation embryo. Endocrinology 1993; 133: Giudice LC. Growth factors and growth modulators in the human uterine endometrium: their potential relevance to reproductive medicine. Fertil Steril 1994;61: Xias P, Tekpetey FR, Armstrong DT. Effect of IGF-I on pig oocyte maturation, fertilization, and early embryonic development in vitro, and on granulosa and cumulus cell biosynthetic activity. Mol Reprod Dev 1994;38: Liu HC, He ZY, Tang YX, Mele CA, Veeck LL, Davis O, et al. Simultaneous detection of multiple gene expression in mouse and human individual preimplantation embryos. Fertil Steril 1997;67: Liu HC, Tseng L. Estradiol metabolism in isolated human endometrial epithelial gland and stromal cells. Endocrinology 1979;104: Liu HC, He ZY, Mele CA, Veeck L, Davis OK, Rosenwaks Z. Expression of IGFs and their receptors is a potential marker for embryo quality. Am J Reprod Immunol 1997;38: He ZY, Liu HC, Mele CA, Barmat L, Veeck LL, Davis O, et al. Expression of inhibin/activin subunits and their receptors and binding proteins in human preimplantation embryos. J Assist Reprod Genet 1999;16: Sakkas D, Jaquenoud N, Leppens G, Campana A. Comparison of results after in vitro fertilized human embryos are cultured in routine medium and in coculture on vero cells: a randomized study. Fertil Steril 1994;61: Harvey MB, Kaye PL. Insulin increases the cell number of the inner cell mass and stimulates morphological development of blastocysts in vitro. Development 1990;110: FERTILITY & STERILITY 367