Introduction
Various culture media have been used to support early stages embryo development in human but also in farm animals. However embryo requirements are far from completely defined and length of culture time is still associated with loss of viability, especially at the individual level. In fact, during this preimplantation period, there are two phases. The first phase is one completely under maternal control, driven by the maternal material: mRNAs and proteins stored during oogenesis and the final stages of oocyte maturation. Then, secondly, after maternal to zygotic transition, the embryo will take the lead after the first real ‘embryonic’ transcriptions. If the culture time is shortened, embryo transfer is impossible in farm animals, due to their expulsion. In human, embryo transfer is usually performed in the uterus, at least 2 days before the normal entry time. Various types of co-culture systems have been devised to overcome these problems: (1) co-culture with trophoblastic tissue (Heyman et al., Reference Heyman, Ménézo, Chesne, Camous and Garnier1987), the first system to overcome the early cell block in farm animals; (2) co-culture with cell monolayers of fibroblasts have really limited value for embryo culture; and (3) monolayers of oviduct cells or uterine cells are successful to allow culture from the fertilised embryo to the blastocyst stage (Gandolfi & Moor, Reference Gandolfi and Moor1987; Thibodeaux et al., Reference Thibodeaux, Ménézo, Roussel, Hansel, Goodeaux, Thompson and Godke1992). In human, co-culture with Vero cells (Ménézo et al., Reference Ménézo, Guérin and Czyba1990), and with autologous uterine cells obtained from biopsies has been widely used.
In this article, we will review the physiological and biochemical aspects of the tubal environment and their impact on in vitro co-culture.
Physiological and biochemical aspects
Oviduct fluid is generated by transudation into the lumen as well as by secretion of compounds synthesized by epithelial cells (Leese, Reference Leese1988) (Fig. 1).
Mouse oviduct cells (A, C, D) and embryo into oviduct (B).
Collection of normal and physiologic tubal fluid has been performed in various ways (reviewed by Aguilar & Reyley, Reference Aguilar and Reyley2005): surgical installation of cannula in the lumen of the tube, aspiratory puncture of the dilated oviduct 3 days after the organ was tied at its two extremities, and absorption with filter paper after oviduct dissection and opening.
We have observed that only two techniques permit the collection of physiologic tubal fluid: (1) cannulation; and (2) aspiratory puncture after double ligature of the organ. Indeed the absorption technique with filter paper provides a fluid whose composition is close to that of the serum (personal observation). The absorption technique may select the transudate part of the oviduct fluid to the detriment of the secretory part. The flushing techniques do not allow the quantification of compounds.
The oviduct fluid amount is dependent on the stage of estrous cycle. The produced amount is increased by estradiol-17β and decreased by progesterone (Leese, Reference Leese1988).
Oviduct fluid contains amino acids, ions, proteins, hormones, growth factors, carbohydrates and lipids. Oviduct fluid composition has been analysed in several species, including cow, ewe, sow, rabbit, mouse and human (Guérin et al., 1995).
Amino acids
Certain amino acids, particularly alanine, glycine, glutamate, hypotaurine and taurine, are present in strikingly high concentrations in the oviductal fluid of cow, ewe, sow, rabbit and mouse (Ménézo, Reference Ménézo1976; Guérin et al. 1995) (Fig. 2). We have observed that oviduct cells support modifications of the amino acids environment. When methionine was incubated with oviduct epithelial cells we observed the production and release of taurine/hypotaurine (Guérin et al., Reference Guérin, Guillaud and Ménézo1995b), glutathione, cystathionine and homocysteine (Ouhibi et al., Reference Ouhibi, Hamidi, Guillaud and Ménézo1990). The particular pattern of amino acid concentrations may be an important factor to be considered for the improvement of embryo culture media.
Amino acid concentrations into serum and oviduct fluid of ewe.

Ions
Concentrations of ions in tubal fluid are similar to those in serum for the majority of species. Nevertheless, high potassium levels in tubal fluid seem to be constant across species. Potassium concentrations are considerably higher in tubal fluid than in serum (Table 1). Consequently, potassium concentration in embryo culture media is an important factor to consider. In a mouse IVF system, more pregnancies were obtained by using a medium based on human oviduct fluid potassium composition (HOF) rather than Tyrode's medium containing a low concentration of potassium (Quinn et al., Reference Quinn, Warner, Kerin and Kirby1985).
Concentration of ions in the oviduct fluid of various species collected from ampulla or the whole tube
Modified from Aguillar et al., 2005.
Glucose and its metabolites
Glucose levels in bovine oviduct are extremely low in comparison with serum concentrations suggesting that this is important for successful sperm capacitation and embryonic development. Preimplantation embryos gain their energy by oxidative phosphorylation, i.e. the oxidation of pyruvate and amino acids. Only at the compaction stage do embryos switch to glycolysis.
High levels of glucose in culture medium have been considered as major effectors of in vitro embryo developmental arrests. High glucose concentrations in the culture medium are implicated in the generation of reactive oxygen species (ROS) (Iwata et al., Reference Iwata, Akamatsu, Minami and Yamada1998). Glucose inhibits hypoxanthine phosphoribosyl transferase (HPRT) activity. The result is the production of ROS via xanthine oxidase (Guérin et al., Reference Guérin, El Mouatassim and Ménézo2001). Metabolism of glucose may be impaired in rodent in relation to a metabolic block with a poor entry into Krebs’ cycle and a block at glucose-6-phosphate isomerase leading to useless accumulation of glycogen. High levels of glucose in culture media may have deleterious effects in diabetic mammals and probably in humans.
Pyruvate is an interesting compound in that it acts not only as an energy source, but can also detoxify ammonia in the embryo through transamination and export of alanine, formed as a result. It also plays a role in protection against oxidative stress (see Oviduct production against oxidative stress) by preventing peroxide-induced injury (Morales et al., Reference Morales, Tilquin, Rees, Massip, Dessy and Van Langendonckt1999). Addition of lactate is questionable as it is a cul de sac product. To be re-introduced into the metabolism it has to be converted (oxidized) in pyruvate with the generation of NADH. Lactate can be used as an energy substrate by embryos and may serve to maintain the proper redox balance and pH. One of the interests of co-culture with oviduct cells may be the reduction of concentrations of glucose and the production of lactate (Ouhibi et al., Reference Ouhibi, Hamidi, Guillaud and Ménézo1990). Other glucose-derived metabolites (such as acetate, citrate, malate etc.) are sometimes added in media with no real established impact.
Vitamins
Ascorbic acid is present in follicular and tubal fluid (Hansen et al., Reference Hansen, Srikandakumar and Downey1991; Paszkowski & Clarke, Reference Paszkowski and Clarke1999). Several animal experiments have claimed the interest of water-soluble vitamins in culture media, but still no unequivocal demonstration has been made. Vitamin C and vitamin E have protective effects on embryo development. Experiments concerning embryo culture medium supplementation with vitamin C in the presence or absence of vitamin E strongly suggest that vitamin C is active via the protection of vitamin E against oxidation reactions.
Folic acid is another parameter: reduced folic acid is a methyl donor (as methionine through S-adenosyl methionine). There is no evidence that supplementation in vitro is necessary. Lipid-soluble vitamins (A and E) are natural potent anti-oxidants, but their introduction into culture media is problematic due to their low hydrosolubility.
Growth factors
Early embryos have receptors for a variety of growth factors, and co-cultured cells may secrete growth factors that stimulate embryo development. Several growth factors have been identified in oviduct and have beneficial effects on embryo development. Platelet-activating factor (PAF) is secreted by human and hamster embryo, and PAF receptor (PAF-R) has been identified in the oviduct. Bovine embryos synthesize growth hormone (GH) and its receptor in a stage-specific manner and bovine oviduct epithelial cells express the GH receptor. IGF binding proteins (IGFBPs) are present in early embryos (Pushpakumara et al., 1991). Locally produced IGFs, regulated by IGFBPs may act directly on the embryo or via modulation of oviduct secretions and muscular activity to positively influence early embryonic development (Pushpakumara et al., 1991). Vascular endothelial growth factor (VEGF) and FGF are secreted by the bovine oviduct in different amounts depending on the stage of the estrous cycle. FGF-1 is mainly secreted at ovulation whereas FGF-2 is increased after ovulation. FGF-2 promotes development of bovine embryos beyond the 8-cell block (Larson et al., Reference Larson, Ignotz and Currie1992). VEGF accelerates development of embryos (Luo et al., Reference Luo, Kimura, Aoki and Hirako2002), PDGF has beneficial effects on embryo development in vitro (Larson et al., Reference Larson, Ignotz and Currie1992).
Nevertheless, no specific embryotrophic factors secreted by the genital tract has been identified.
Other compounds such as macromolecules (OVGP) and hyaluronic acid (HA) are present in tubal fluid and have beneficial effects on embryo development. Expanding cumulus cells produce large amounts of HA. HA has a positive effect on cell number and apoptosis in blastocysts (Ulbrich et al., Reference Ulbrich, Schoenfelder, Thoene and Einspanier2004).
Serum is not necessary for IVF culture (Ménézo et al., Reference Ménézo, Testart and Perdrone1984): serum addition affects mRNA content and is involved in the ‘large calf syndrome’. Serum albumin is not fully necessary for embryo culture, but has always been added into culture media as it helps with embryo handling. The beneficial role of albumin is not fully understood: it binds several compounds of various molecular weights, such as lipids, amino acids, peptides, catecholamines. It may be incorporated directly into the embryo and may be a carrier of nutrients.
Oviduct protection against oxidative stress
ROS are produced by gametes and embryo, particularly during capacitation, fertilization and the second cell cycle division in mouse (Noda et al., Reference Noda, Matsumoto, Umaoka, Tatsumi, Kishi and Mori1991). Superoxide anion O2−•) induces the production of other ROS such as hydrogen peroxide (H2O2) and the extremely reactive hydroxyl radicals (OH•) via the Haber–Weiss or de Fenton reactions (Guérin et al., Reference Guérin, El Mouatassim and Ménézo2001). Oxidative stress affects embryo viability via alteration of proteins, nucleic acids and membrane lipids (Guérin et al., Reference Guérin, El Mouatassim and Ménézo2001). Oxidative stress is particularly important during in vitro embryo production (Goto et al., Reference Goto, Noda, Mori and Nakano1993) and is implicated in the in vitro embryo block (Noda et al., Reference Noda, Matsumoto, Umaoka, Tatsumi, Kishi and Mori1991). Several factors may explain that: exposition to high levels of oxygen, traces of metallic ions and/or light (Nakayama et al., Reference Nakayama, Noda, Goto and Mori1994), oxygen concentration etc. Furthermore, serum, used for in vitro maturation of animal oocytes contains amine oxidases (Parchment et al., Reference Parchment, Lewellyn, Swartzzendruber and Pierce1990). Reduction of oxygen concentration from 20% to 5% limits ROS production by the mouse embryo (Goto et al., Reference Goto, Noda, Mori and Nakano1993).
Several enzymatic and non-enzymatic mechanisms present in oviduct are implicated in embryo protection against oxidative stress: hypotaurine, ascorbic acid, superoxide dismutase (SOD), catalase, gluthathione peroxidase (GPX) etc.
Hypotaurine and taurine have been detected in oviduct and follicular fluids of cow, ewe, sow, rabbit and mouse (Guérin et al., Reference Guérin, Guillaud and Ménézo1995b) (Fig. 3). Oviduct epithelial cells in vitro are able to synthesise taurine and hypotaurine from methionine (Guérin et al., Reference Guérin, Guillaud and Ménézo1995b). The enzyme cysteine sulphinate decarboxylase (EC 4.1.1.29.) is present and active in the oviduct of cow and goat (Guérin & Ménézo, Reference Guérin and Ménézo1995). Hypotaurine neutralise OH• radicals. The byproduct of this reaction is taurine, which is able to neutralise cytotoxic aldehydes produced by oxidative stress. Furthermore, hypotaurine addition to embryo culture medium has beneficial effects on bovine embryo development (Barnett & Bavister, Reference Barnett and Bavister1992; Guyader-Joly et al., Reference Guyader-Joly, Guérin, Renard, Guillaud, Ponchon and Ménézo1998).
Taurine and hypotaurine concentrations into serum and oviduct fluid.
Other antioxidant compounds are present in tubal fluid, such as vitamin C and pyruvate (see section on Vitamins). It must be remembered that vitamin C may also have a pro-oxidant action via the reduction of ferric ions to the ferrous state (catalysing the Haber–Weiss reaction).
In addition the oviduct provides antioxidant enzymes: the transcripts encoding for antioxidant enzymes SOD, catalase and GPX are present in oviduct epithelial cells (El Mouatassim et al., Reference El Mouatassim, Guérin and Ménézo1999). These redundant and complementary mechanisms are schematically presented in Fig. 4.
Schematic representation of some protective mechanisms of oocyte and embryo against reactive oxygen species. CSD, cysteine sulfinate decarboxylase; MnSOD, manganese superoxide dismutase; CuZnSOD, copper zinc superoxide dismutase; GPX, glutathione; GSH, reduced glutathione.
Furthermore, the low oxygen pressure in the oviduct (40–60 mmHg) (Fisher & Bavister, Reference Fisher and Bavister1993), may limit auto-oxidation reactions. It must be emphasised that the benefit of co-culture may be partly related to the decrease of O2 concentration in the medium by oviduct epithelial cells (Nagao et al., Reference Nagao, Saeki and Kainuma1994).
Impact on in vitro culture
Culture systems
Culture in microdrops under a layer of mineral oil is often recommended on the basis of mimicking the in vivo situation and following the ‘mouse model’ for embryo culture. Although the volume of liquid surrounding the embryo in vivo is minimal. or ‘virtual’, the embryo is in contact with an epithelial cell layer, with a film of liquid that is permanently renewed (Fig. 1). This is not the case for microdrop culture, and there are three features to be considered: (1) Metabolic waste from the embryo may accumulate at a higher concentration, due to the sub-optimal volume of liquid. (2) The microdrop surface area allows a maximum exchange with oil, so that liposoluble compounds in the medium can be absorbed at a maximal level by the oil, thus depriving the embryo of these compounds. (3) Similarly, any micropollution of the oil with compounds that are water soluble will pass into the culture medium with a higher efficacy in a microdrop system.
Culture medium
Along with other teams, we began to design and define new, more complex culture media in the early 1970s that were based on the composition of genital tract secretions (Ménézo, Reference Ménézo1976). In this way, culture media are supplemented with amino acids, especially glycine, glutamine, alanine, hypotaurine and taurine.
Early preimplantation development, i.e. until maternal to zygotic transition (MZT), is maternally driven. Maternal mRNA and protein content of the oocytes, at the time of fertilization, is of major importance as there is an ineluctable turn over of these reserves. During MZT there is a minimal level of transcription only: the mature oocyte must contain a storage pool of proteins and/or mRNA transcripts in order to maintain its viability. Embryos’ needs before and after ZGA are not alike and the environment, i.e. culture conditions, will have a direct impact on transcription and translation.
In vitro culture conditions and environmental factors are very different from the in vivo conditions (light, variations in pH and temperature, static medium etc.). pH is regulated by HCO3/CO2 buffer according to the equation pH = pKa + log (HCO3)/(CO2). The Henderson–Hasselbach equation allows the calculation of pH under a 5% CO2 atmosphere, in relation with the concentration of bicarbonate. The embryo has an alkaline (7.4) pH: it is not able to match with an acidic pH before ZGA. Gas phase can be 5% CO2 in air or 5% CO2, 5% O2 and 90% N2. Even if a reduced O2 atmosphere seems to be more physiological, based on genital tract environment, and more susceptible to decrease the free radical formation, no clear cut data support the superiority of reduced oxygen tension in human especially in terms of ongoing pregnancy rates, especially if the medium is well protected by antireactive oxygen species (see Oviduct production against oxidative stress). In this way glutathione is sometimes added to culture media, however it has to be mentioned that it cannot enter the embryo.
Contaminants of the gas phase, such as NO and CO, are as deleterious as volatile organic compounds. An osmolarity between 280 and 300 mM allows early embryo development. No attention is generally paid to the red-ox potential, based on oviduct and uterine secretions it should be calibrated to −0.1 mV. The problem here is due to the absence of technical methods allowing the accurate measurement of the redox potential of a culture medium. A too strong reducing red-ox may have harmful effects on proteins.
Serum addition destabilises the culture media through a contribution of enzymes and unknown metabolites and catabolites. It is useless and even deleterious as it may increase pathologies (especially linked to imprinting).
Addition of growth factors in embryo culture media is still a matter of controversy. Growth factors are present in the female genital tract and the corresponding receptors are present on the embryos sooner or later. Beneficial effects have been observed after supplementation of embryo culture medium with growth factors in various species: FGF-2, VEGF, LIF and PDGF in bovine (Larson et al., Reference Larson, Ignotz and Currie1992; Luo et al., Reference Luo, Kimura, Aoki and Hirako2002). GH is synthesized by bovine embryo and bovine oviduct epithelial cells express the GH receptor (Hull & Harvey, Reference Hull and Harvey2001). In humans, GMCSF seems to have also a stimulatory and regulatory impact. Insulin is usually added for the second phase of culture: whether or not it has the same effect as IGF1 is questionable. Locally produced IGFs, regulated by IGFBPs, may act directly on the embryo or via modulation of oviduct secretions to positively influence the success of early embryonic development (Pushpakumara et al., Reference Pushpakumara, Robinson, Demmers, Mann, Sinclair, Webb and Wathes2002).
Growth factors may act through a balance between stimulatory and inhibitory effects. Mouse embryo development is significantly better when embryos are cultured in groups rather than alone. These results suggest an autocrine/paracrine effect but this effect has been denied in human systems.
Trophic effect of co-culture on embryo development
Even if it is now well established that tubal secretion is hormonally regulated, some pharmacological compounds may alter it weakly. There are, in vivo, relations between the oviduct and the passing embryo. The embryo probably regulates its transport and the microenvironment of the tubal cells and the cellular metabolism of the oviduct is modified during passage of the egg (Freese et al., Reference Freese, Orman and Paulos1973).
However, if we consider the trophic effect towards the embryo i.e. allowing a correct development without loss of viability, we can assess that hormones do not interfere in vitro with this process and that the exchanges between the embryo and the cells are not mandatory. Prepuberal oviducts can sustain early embryonic development in the mouse and in the bovine (Ménézo et al., Reference Ménézo, Hamidi, Khatchadourian and Nardon1989). Co-culture with genital tract cells collected at different stage of the estrous cycle effectively support embryonic development even if the cell morphology differs (Thibodeaux et al., Reference Thibodeaux, Ménézo, Roussel, Hansel, Goodeaux, Thompson and Godke1992). The final demonstration was given with the use of Vero or BRL cells for co-culture in human and bovine (Ménézo et al., 1994).
Co-cultured cell metabolism leads to modifications of the culture medium, mimicking tubal secretions. In this way, co-cultured cells may control important physico-chemical parameters of the medium (ionic composition, redox potential, pH, osmotic pressure etc.). It is generally assumed that the feeder cells in the co-culture systems act, on the one hand, via removal of toxic compounds from the culture medium (heavy metal cations, metabolic inhibitors, oxygen, glucose etc.) and, on the other hand, via the production of compounds regulating embryo development (hypotaurine, glutathione, amino acids, lipids, growth factors etc.). Cell monolayers seem able to regulate metabolic pathways of the embryos.
The in vitro embryo development block appears to be related to embryonic genome activation. Thus the overcoming of the in vitro block by co-culture systems suggests that co-cultured cells are able to help in activation of the transcription mechanism.
Conclusion
It is important to understand some of the interactions that take place between oviduct and early embryo in order to develop culture systems and culture media adapted to embryo needs. Prior to and during culture, it is important to be aware of, and to control, the interactions of different compounds with each other and with the gas phase. Sequential media are currently effective in human embryo culture, with blastocyst yields similar to those observed for co-culture systems. In order to evaluate new techniques and/or new culture media, in vitro grown embryos must be transferred to avoid misleading observations based upon morphology alone.
