Development of serum-free media for the culture and transfer of human blastocysts

Development of serum-free media for the culture and transfer of human blastocysts David K.Gardner Colorado Center for Reproductive Medicine, 799 East Hampden Avenue, Suite 300, Englewood, CO 80110, USA The culture and transfer of the blastocyst stage embryo has several advantages for assisted reproduction in the human. However, due to inadequacies of present culture conditions in human in-vitro fertilization (IVF), embryos are routinely transferred to the uterus on either day 2 or day 3 of development around the 4- to 8-cell stage, with resultant implantation rates of only 10-25%. In other mammalian species the transfer of cleavage stage embryos, which normally reside in the oviduct, results in a significantly lower implantation rate compared with the transfer of blastocysts. Extended culture of human embryos in vitro will help to identify those embryos with little, if any, developmental potential. It is therefore plausible that the blastocyst has an intrinsically higher viability than the cleavage stage embryo. It has now been shown in human IVF that sequential serum-free media can support >50% blastocyst development, with an implantation rate per blastocysts of 50%, double that obtained for cleavage stage embryos. As the implantation rate of the blastocyst is higher than the cleavage stage embryo, fewer blastocysts are required for transfer. The development of completely defined embryo culture media may prove feasible by the replacement of protein with the glycosaminoglycan hyaluronate. Hyaluronate, which is protein-free, is more suitable than albumin in supporting implantation in the mouse, and can eliminate the biological variation inherent when using protein and the potential for contamination when using blood products such as albumin. Key words: albumin/amino acids/carbohydrates/hyaluronate/sequential culture/ viability Introduction There are numerous advantages to being able to culture the embryo to the blastocyst stage prior to transfer in human in-vitro fertilization (IVF). These include the synchronization of the embryonic stage with the uterus (Bavister, 1995), and the identification of those embryos with little developmental potential (Dawson et al, 1995), culminating in an increased implantation rate per embryo 21 o European Society for Human Reproduction and Embryology Human Reproduction Volume 13 Supplement 4 1998

Culture and transfer of human blastocysts (Gardner and Lane, 1997, 1998). Such an increase in embryo implantation rate will not only increase the overall efficiency of human IVF per se, but will significantly reduce the number of multiple gestations, as fewer embryos will be required at transfer to establish a successful pregnancy. Blastocyst culture will also facilitate trophectoderm biopsy and the application of embryo viability tests such as the non-invasive assessment of metabolism (Gardner and Leese, 1993; Gardner and Lane, 1997). The latter approach has been used successfully in animal models to predict pregnancy outcome. The question, therefore, is how can viable human blastocysts be obtained in culture? The emphasis here is on the development of viable blastocysts. Unfortunately, blastocyst development in culture does not necessarily mean that the embryo is viable, an observation demonstrated quite clearly in the mouse (Gardner et al, 1997a; Lane and Gardner, 1997a; Gardner and Lane, 1998). This point can initially appear rather paradoxical, and has contributed to the overall confusion over the potential role of blastocyst transfer in human IVF. A good example of this is the work of Bolton et al. (1991), in which human embryos were cultured to the blastocyst stage in Earle's balanced salts supplemented with pyruvate and maternal serum. Although a respectable 40% blastocyst development was achieved, the resultant implantation and pregnancy rate was only 7%, below that for the transfer of cleavage stage embryos. It was subsequently concluded from this study, and others like it, that the transfer of human blastocysts did not offer a solution to the low implantation rate observed when cleavage stage embryos are transferred, typically only 10-25%. Sequential culture media It is the premise of this discussion that, in order to support development of a competent zygote to the viable blastocyst stage, one needs to use more than one culture medium to take into account the significant changes in embryo physiology and metabolism which occur during the preimplantation period (Leese, 1991; Gardner, 1998a,b,c). The term competent zygote is used here in order to highlight the fact that a significant number of embryos conceived through IVF, ~25-35% are chromosomally abnormal (Kola et al., 1993; Van Blerkom, 1994) and therefore will have impaired developmental potential. Thus the maximum percentage of pronuclear embryos that one may expect to reach the blastocyst stage is around 65-75%. The nature of the changes in embryo physiology and metabolism which occur during the preimplantation period have been discussed extensively elsewhere (Biggers et al., 1989; Leese, 1991; Rieger, 1992; Gardner and Lane, 1993a; 1997; Gardner, 1998a,b). As a result of the changes in the developing embryo's requirements, the levels of carbohydrates and those amino acids which support excellent development of the cleavage stage mouse embryo, do not support optimum development and differentiation of the blastocyst (Gardner and Lane, 1993b; Lane and Gardner, 1997a). In contrast, the levels of carbohydrates and those amino acids which do support blastocyst development 219

D.K.Gardner and differentiation can actually impair the cleavage stage embryo (Lane and Gardner, 1994, 1997b; Gardner and Lane 1997; Gardner et al, 1997a), lending further support to the hypothesis that optimal development of the human embryo in culture will require more than a single medium formulation. In support of such in-vitro studies, the environment within the oviduct and uterus differ markedly, the human embryo therefore being exposed to a changing nutrient pool as it develops in vivo (Gardner et al, 1996). However, as well as having to meet the changing requirements of the embryo as it develops in culture, it is also of paramount importance that the amount of cellular stress in the embryo, induced by the culture conditions, is kept to a minimum. Examples of cellular stress experienced by the embryo include osmotic stress, shifts in intracellular ph (phj) and perturbations in energy metabolism. Manifestations of such cellular stress include retarded cleavage, cleavage arrest, cytoplasmic blebbing, aberrant metabolism, impaired energy production, inadequate genome activation and gene transcription. Understanding the nature of such cellular stress has been important in the development of new media that can minimize such culture-induced effects (some would say culture-induced artefacts). A well-documented example of culture-induced stress is that observed when glucose and phosphate are present in a simple culture medium. A simple embryo culture medium is one based on balanced salt solutions, such as Earle's salts (Edwards, 1981) and human tubal fluid (HTF) (Quinn et al, 1985), and which lack amino acids. In such media, glucose has been shown to impair the development of embryos from several mammalian species including the human (Conaghan et al, 1993). There has therefore been a trend to remove glucose from such culture media in order to alleviate this problem (Quinn, 1995; Pool et al, 1998). However, this may be perceived as alleviating an in-vitro induced artefact by the implementation of a second artefact, i.e. the removal of glucose from culture media even when it is present in both oviduct and uterine fluids. An alternative approach has been to determine the nature of the glucose toxicity (Seshagiri and Bavister, 1991; Gardner and Lane, 1993c; Lane and Gardner, 1997c; Gardner, 1998a), and then to design media which alleviate the problem (Gardner and Lane, 1993b; 1996; 1998). For example, when present in a simple medium such as HTF glucose can induce its own utilization through glycolysis at the expense of substrate oxidation in the cleavage stage embryo (Seshagiri and Bavister, 1991; Gardner and Lane, 1993c; Vella et al, 1997), a phenomenon somewhat analogous to the Crabtree effect in tumour cells (Koobs, 1972). However, when specific amino acids and EDTA are included in the culture medium, this induction of glycolysis is prevented. Amino acids and EDTA suppress glycolysis through different mechanisms and act in combination to further suppress glycolysis (Gardner and Lane, 1993c; Gardner 1998a). Importantly the role of both amino acids and EDTA in culture medium is not limited to this effect on embryo metabolism. Amino acids also act as both osmolytes and intracellular ph (phj) buffers, whilst EDTA helps to chelate potentially toxic divalent cations present as contaminants in the culture system. The question is therefore which approach to alleviating glucose toxicity in embryo culture is 220

Culture and transfer of human blastocysts most suitable for human IVF? Unfortunately we do not know the answer yet. Both approaches have their merits. Indeed it could be argued that the human embryo in vivo does not 'see' glucose until the end of the first cell cycle to the midpoint of the second cell cycle, when the surrounding cumulus cells have dispersed. This is because the cumulus cells readily convert glucose to both lactate and pyruvate (Gardner et al, 1996). However, it is also evident that the later stage embryo, especially from the 8-cell stage onward has an increasing requirement for glucose. Therefore, the presence of glucose in culture medium for blastocyst development is warranted. The above discussion about the toxicity of glucose and the methods employed to resolve this is a good example of how the minimization of cellular stress, irrespective of the method employed, facilitates improved embryo development in culture (Gardner and Schoolcraft, 1998). However, as amino acids and EDTA confer secondary benefits to the embryo, it is proposed that it is more physiological and effective to first identify the mechanism of the problem and solve it, rather than simply to remove the causal ingredients. In the case of glucose, the latter approach ignores any potential benefits glucose may have in embryo development, specifically its role in biosynthesis (Gardner, 1998a,b). Taking the above principles into account, the sequential human embryo culture media Gl and G2 were formulated (Gardner, 1994; Barnes et al, 1995; Gardner and Lane, 1997). Medium Gl was designed to support development of the pronuclear embryo to the 8-cell stage, while G2 was designed to support development of the 8-cell embryo to the expanded blastocyst stage. Using these media, blastocyst development of >60% has been reported (Gardner et al, 1997a; 1998a). In another programme, a modification of this system in which a commercially available medium (IVF 50; IVF Science Scandinavian) was used for the first 48 h of development prior to culture in medium G2, and gave rise to a >50% blastocyst development (Jones et al, 1998). In both cases, the resultant blastocysts had an implantation rate of twice that of the cleavage stage embryos in the respective clinical IVF programmes (Gardner et al, 1998a; Jones et al, 1998). Similarly, using different media, but sticking to the principles of sequential media, two other groups have reported excellent blastocyst development (>50%) and the establishment of pregnancies after blastocyst transfer (Behr et al, 1997; T.Pool, personal communication). So it is evident that the use of two or more sequential media to support the extended culture of the human embryo to the blastocyst stage can give rise to viable blastocysts. The efficacy of blastocyst transfer after culture in sequential media has recently been confirmed through a prospective randomised clinical trial (Gardner et al., 1998b). Development of completely defined media for human IVF One criticism raised continually in human IVF is the inclusion of blood products, such as serum or serum albumin in the culture system. The use of any blood product raises the serious issue of potential for contamination and infection of the 221

D.K.Gardner gametes and embryos. With the widespread concerns regarding the transmission of Creutzfeldt-Jakob disease through donated serum products, there is an increasing urgency to move to a protein-free embryo culture environment. The reasons for removing whole serum from embryo culture will be discussed separately. The use of serum albumin for human embryo culture stems from its routine use in the culture of mouse, sheep and cattle embryos. Interestingly, the inherent problems of using serum albumin in animal models have been well documented, specifically the significant batch to batch variation in embryotrophic activity (Kane, 1983; Batt et al, 1991; McKiernan and Bavister, 1992), thus making batch to batch comparisons very difficult. This problem has fuelled the search for alternative macromolecules to serum albumin. Synthetic polymers such as polyvinylalcohol (PVA) and polyvinylpyrrolidone (PVP) have both been used in assisted reproductive technologies (Bavister, 1995), however neither can be considered a physiological alternative to protein, and the teratological nature of these compounds has not been determined. Furthermore, the true role of serum albumin in embryo culture has yet to be fully elucidated, although it is plausible that albumin helps to bind and stabilize growth factors. On a practical level, the presence of macromolecules in embryo culture media serves to facilitate manipulation of gametes and embryos, which would otherwise adhere to the glass or plastic culture vessel. A physiological alternative to albumin is the glycosaminoglycan hyaluronate. Not only do the levels of this glycosaminoglycan increase in the uterus around the time of implantation in the mouse (Zorn et al, 1995), the human embryo expresses the receptor for it throughout preimplantation development (Campbell et al., 1995). Although hyaluronate is a glycosaminoglycan, unlike other glycosaminoglycans it is not covalently linked to a core protein, and can therefore be considered a polysaccharide. This therefore removes both the problems of variation and contamination, as hyaluronate can be synthesized and isolated in a pure form. In preliminary trials of hyaluronate in mouse embryo culture and transfer, it was found that not only could hyaluronate (0.5 mg/ml) readily replace serum albumin in culture, but more importantly it significantly increased the implantation rate of resultant blastocysts (Gardner et al., 1997b). Furthermore, this increase in implantation rate could be attributed to the presence of hyaluronate in the medium used at transfer alone. When mouse embryos were cultured in the appropriate sequential media, there was no evident benefit in vitro of having any macromolecule/protein present. Embryos formed blastocysts at the same rate in medium in the absence of any macromolecule. However, when such blastocysts were pre-equilibrated in medium with 0.5 mg/ ml hyaluronate for 5 min prior to transfer in medium with hyaluronate, the resultant implantation rate was equivalent to that obtained for embryos cultured for the entire preimplantation period in the presence of hyaluronate (Gardner et al., 1997b). The effect of hyaluronate in human embryo transfers warrants investigation. Removal of serum albumin therefore has several advantages, but what about the removal of whole serum? The need to remove whole serum from extended culture systems appears to be more pressing than the removal of albumin for the 222

Culture and transfer of human blastocysts following reasons; serum has been shown to induce a wide range of abnormalities in the blastocysts of sheep cultured in its presence from the zygote stage. These abnormalities include abnormal ultrastructure of the mitochondria (Thompson et al., 1995), abnormal energy metabolism (possibly the result of impaired mitochondrial function) (Gardner et al, 1994), premature blastocoel development, and (by far the most alarming) the development of abnormally large fetuses (Thompson et al, 1995). Although hundreds of babies have been born through IVF after blastocyst transfer using co-culture and/or serum and none of the resultant births have been reported to be abnormal, there is still cause for concern that certain sera may indeed induce excess cellular stress which may then be manifest during gestation or during the lifetime of the child. The mechanism by which serum induces abnormalities in the embryos/fetus of domestic animals is unresolved, though the over-expression of certain growth factor genes has been implicated. This therefore questions the merit of supplementing media for human embryo culture with growth factors. The mammalian blastocyst expresses the ligands and receptors for several growth factors, many of which are known to exhibit cross-reactivity. It is questionable then that an observed effect after adding an individual ligand to a culture medium is of any physiological significance. Until a direct link between any growth factor's action and subsequent fetal development can be demonstrated conclusively in animal models, it is advisable at this stage of our understanding of blastocyst development and differentiation that growth factors not be utilized in human embryo culture systems. Conclusions Using sequential serum-free culture media, not only is it possible to culture the human pronuclear embryo to the blastocyst at high frequencies, but the resultant blastocysts have a higher implantation rate than the cleavage stage embryo after transfer, culminating in an increased efficiency in human IVF. This approach will significantly reduce the need for the transfer of several embryos, thereby reducing the number of multiple gestations. Further refinements in human embryo culture media are likely to include the replacement of protein with such physiological macromolecules as hyaluronate. Such media will have inherently less biological variation and should significantly reduce the potential of infection from the culture system. References Barnes, F.L., Crombie, A., Gardner, D.K. et al. (1995) Blastocyst development and pregnancy after in vitro maturation of human primary oocytes, intracytoplasmic sperm injection and assisted hatching. Hum. Reprod., 10, 3243-3247. Batt, P.A., Gardner, D.K. and Cameron, A.W.N. (1991) Oxygen concentration and protein source affect the development of preimplantation goat embryos in vitro. Reprod. Fertil. Dev., 3, 601-607. 223

D.K.Gardner Bavister, B.D. (1995) Culture of preimplantation embryos: facts and artifacts. Hum. Reprod. Update, 1, 91-148. Behr, B., Moore, D., Gebhardt, J. and Hall, S. (1997) Human blastocyst development in vitro without co-culture using a sequential culture system. Proc. Am. Soc. Rep. Med. P081. Biggers, J.D., Gardner, D.K. and Leese, H.J. (1989) Control of carbohydrate metabolism in preimplantation mammalian embryos. In Rosenblum, I.Y. and Heyner, S. (eds), Growth Factors in Mammalian Development. CRC Press, Boca Raton, FL, USA, pp. 19-32. Bolton, V.N., Wren, M.E. and Parsons, J.H. (1991) Pregnancies after in vitro fertilization and transfer of human blastocysts. Fertil. Steril., 55, 830-832. Campbell, S., Swann, H.R., Aplin, J.D. et al. (1995) CD44 is expressed throughout pre-implantation human embryo development. Hum. Reprod., 10, 425-430. Conaghan, J., Handyside, A.H., Winston, R.M.L. and Leese, H.J. (1993) Effects of pyruvate and glucose on the development of human preimplantation embryos in vitro. J. Reprod. Fertil., 99, 87-95. Dawson, K.J., Conaghan, J., Ostera, G.R. et al. (1995) Delaying transfer to the third day postinsemination, to select non-arrested embryos, increases development to the fetal heart stage. Hum. Reprod, 10, 177-182. Edwards, R.G. (1981) Test-tube babies. Nature, 293, 253-256. Gardner, D.K. (1994) Mammalian embryo culture in the absence of serum or somatic cell support. Cell Biol. Int., 18, 1163-1179. Gardner, D.K. (1998a) Changes in requirements and utilization of nutrients during mammalian preimplantation embryo development and their significance in embryo culture. Theriogenology, 49, 83-102. Gardner, D.K. (1998b) Embryo development and culture techniques. In Clark, J. (ed.), Animal Breeding: Technology for the 21 s ' Century. Harwood Academic, Berkshire, UK, pp. 13^-6. Gardner, D.K. (1998c) Improving embryo culture and enhancing pregnancy rate. In Shoham, Z. Howies, C. and Jacobs, H. (eds), Female Infertility Therapy: Current Practice. Martin Dunitz, London, 288-299. Gardner, D.K. and Lane, M. (1993a) Embryo culture systems. In Trounson, A. and Gardner, D.K. (eds), Handbook of in Vitro Fertilization. CRC Press, Boca Raton, FL, USA, pp. 85-114. Gardner, D.K. and Lane, M. (1993b) Amino acids and ammonium regulate mouse embryo development in culture. Biol. Reprod., 48, 377-385. Gardner, D.K. and Lane, M. (1993c) The 2-cell block in CF1 mouse embryos is associated with an increase in glycolysis and a decrease in tricarboxylic acid (TCA) cycle activity: alleviation of the 2-cell block is associated with the restoration of in vivo metabolic pathway activities. Biol. Reprod., 49 (Suppl. 1), 52. Gardner, D.K. and Lane, M. (1996) Alleviation of the '2-cell block' and development to the blastocyst of CF1 mouse embryos: role of amino acids, EDTA and physical parameters. Hum. Reprod, 11, 2703-2712. Gardner, D.K. and Lane, M. (1997) Culture and selection of viable human blastocysts: a feasible proposition for human IVF? Hum. Reprod. Update, 3, 367-382. Gardner, D.K. and Lane, M. (1998) Culture of viable human blastocysts in defined sequential serum-free media. Hum. Reprod., 13 (Suppl. 3), 148-159. Gardner, D.K. and Leese, H.J. (1993) Assessment of embryo metabolism and viability. In Trounson, A. and Gardner, D.K. (eds), Handbook of in Vitro Fertilization. CRC Press, Boca Raton, FL, USA, pp. 195-211. Gardner, D.K. and Schoolcraft, W.B. (1998) Human embryo viability: what determines developmental potential and can it be assessed? J. Assist. Reprod. Genet., 15, 455-458. Gardner, D.K., Lane, M., Spitzer, A. and Batt, P.A. (1994) Enhanced rates of cleavage and development for sheep zygotes cultured to the blastocyst stage in vitro in the absence of serum and somatic cells: amino acids, vitamins and culturing embryos in groups stimulate development. Biol. Reprod, 50, 390-400. Gardner, D.K., Lane, M., Calderon, I. and Leeton, J. (1996) Environment of the preimplantation human embryo in vivo: metabolite analysis of oviduct and uterine fluids and metabolism of cumulus cells. Fertil. Steril., 65, 349-353. Gardner, D.K., Lane, M., Kouridakis, K., and Schoolcraft, W.B. (1997a) Complex physiologically 224

Culture and transfer of human blastocysts based serum-free culture media increase mammalian embryo development. In Gomel, V. and Leung, P.C.K. (eds), In Vitro Fertilization and Assisted Reproduction. Monduzzi Editore, Bologna, Italy, pp. 187-191. Gardner, D.K., Lane, M. and Rodriguez-Martinez, H. (1997b) Fetal development after transfer is increased by replacing protein with the glycoaminoglycan hyaluronate for embryo culture. Hum. Reprod., 12 (Abstr. Book 1), O-215. Gardner, D.K., Vella, P., Lane, M. et al. (1998a) Culture and transfer of human blastocysts increases implantation rates and reduces the need for multiple embryo transfers. Fertil. Steril, 69, 84-88. Gardner, D.K., Schoolcraft, W.B., Wagley, L. et al. (1998b) A prospective randomised trial of blastocyst culture and transfer in human in-vitro fertilization. Hum. Reprod., 13, 3434-3440. Jones, G.M., Trounson, A.O., Gardner, D.K. et al. (1998) Evolution of a culture protocol for successful blastocyst development and pregnancy. Hum. Reprod., 13, 169-177. Kane, M. (1983) Variability in different lots of commercial bovine serum albumin affects cell multiplication and hatching of rabbit blastocysts in culture. J. Reprod. Fertil., 69, 555-558. Kola, I., Sathanathan, A.H. and Gras, L. (1993) Chromosomal analysis of preimplantation mammalian embryos. In Trounson, A. and Gardner, D.K. (eds), Handbook of in Vitro Fertilization. CRC Press, Boca Raton, FL, USA, pp. 173-193. Koobs, D.H. (1972) Phosphate mediation of the Crabtree and Pasteur effects. Science, 178,127-133. Lane, M. and Gardner, D.K. (1994) Increase in postimplantation development of cultured mouse embryos by amino acids and induction of fetal retardation and exencephaly by ammonium ions. J. Reprod. Fertil, 102, 305-312. Lane, M. and Gardner, D.K. (1997a) Differential regulation of mouse embryo development and viability by amino acids. J. Reprod. Fertil, 109, 153-164. Lane, M. and Gardner, D.K. (1997b) Non-essential amino acids and glutamine decrease the time of the first three cleavage divisions and increase compaction of mouse zygotes in vitro. J. Assist. Reprod. Genet., 14, 398-403. Lane, M. and Gardner, D.K. (1997c) EDTA stimulates development of cleavage stage mouse embryos by inhibiting the glycolytic enzyme phosophoglycerate kinase. Biol. Reprod., 57 (Suppl. 1), 193. Leese, H.J. (1991) Metabolism of the preimplantation mammalian embryo. Oxf. Rev. Reprod. Biol, 13, 35-72. McKiernan, S. and Bavister, B.D. (1992) Different lots of BSA inhibit or stimulate in vitro development of hamster embryos. In Vitro Cell Dev. Biol, 28A, 154-156. Pool, T.B., Atiee, S.H. and Martin, J.E. (1998) Oocyte and embryo culture: Basic concepts and recent advances. Infertil Reprod. Med. Clin. N. Am., 9, 181-203. Quinn, P. (1995) Enhanced results in mouse and human embryo culture using a modified human tubal fluid medium lacking glucose and phosphate. /. Assist. Reprod. Genet., 12, 97-105. Quinn, P., Kerin J.F. and Warnes, G.M. (1985) Improved pregnancy rates in human in vitro fertilization with the use of a medium based on the composition of human tubal fluid. Fertil. Steril, 44, 493-498. Rieger, D. (1992) Relationship between energy metabolism and development of the early embryo. Theriogenology, 37, 75-93. Seshagiri, P.B. and Bavister, B.D.(1991) Glucose and phosphate inhibit respiration and oxidative metabolism in cultured hamster eight-cell embryos: evidence for the 'Crabtree effect'. Mol. Reprod. Dev., 30, 105-111. Thompson, J.G., Gardner, D.K., Pugh, P.A. et al. (1995) Lamb birth weight is affected by culture system utilized during in vitro pre-elongation development of ovine embryos. Biol. Reprod., 53, 1385-1391. Van Blerkom, J. (1994) Developmental failure in human reproduction associated with chromosomal abnormalities and cytoplasmic pathologies in meiotically mature oocytes. In Van Blerkom, J. (ed.), The Biological Basis of Early Human Reproductive Failure. Oxford University Press, New York, USA, pp. 283-326. Vella, P., Lane, M. and Gardner, D.K. (1997) Induction of glycolysis in the day-3 mouse embryo by glucose. Biol. Reprod., 57 (Suppl. 1), 26. Zorn, T.M.T., Pinhal, M.A.S., Nader, H.B. et al. (1995) Biosynthesis of glycosaminoglycans in the endometrium during the initial stages of pregnancy of the mouse. Cell. Mol. Biol., 41, 97-106. 225