JBRA Assist. Reprod. 2000;04(02):64-70
ATUALIZAÇÃO
doi: 10.5935/1518-0557.2000.4.2.04
1MD. Research Fellow of UNESP - Botucatu
2PHD FIBMS. Scientific Consultant
Abstract
This paper reviews the use of co-culture in assisted reproduction. It has been suggested that co-culture procedures may have a significant beneficial impact on IVF, as it is believed that "factors" released by the co-culture cells. It has been argued that with conventional IVF the absence of various in vivo influences compromises the viability of IVF embryos. We have therefore set out to look at the current status of co-culture by exploring the different methods and examining the rationale behind their use. We have not attempted a priori to answer the question of which is the best co-culture method nor is it clear in the current climate whether co-culture offers superior results to more modem sequential media formulations which have recently been reported to improve percentages of blastocysts for transfer as compared with traditional media.
Our review of the literature leaves us with the following conclusion; the majority of published papers on co-culture, suggest that they have demonstrated significantly improved blastocyst rates (>50%), though noone has yet demonstrated unequivocally, how co-culture exerts this beneficial effect. The concerns over xeno co-culture and the lack in our understanding of the real underlying factors that underpin the apparent improvements mean that for the time being, co-culture is still only an interesting scientific model with which to study interactions between somatic cell cultures and human embryos. Currently defined media systems are beginning to yield comparable results to those demonstrated by the majority of co-culture studies, and the field is developing rapidly. Under the circumstances, although we feel that co-culture has much to offer as a research tool, for the time being, further research is required to characterise the factors that are involved, with a view to the future likelihood that these will be added to defined media, which now appear likely to supersede the need for coculture in the very near future.
Key words: co-culture, growth factors, animal, human blastocyst culture, media
Resumo
O presente estudo é uma revisão do uso de co-cultura em reprodução assistida. Sugere-se que o procedimento de co-cultura tem um impacto benéfico e significante na FIV, acredita-se que alguns fatores soltam em co-cultura. Argumenta-se que com FIV convencional, a ausência de várias influências in vivo comprometem a viabilidade de embriões para FIY. Temos pesquisado o atual status da co-cultura pela pesquisa de diferentes métodos e examinando a razão de seu uso.
Não estamos tentando, à princípio, responder a questão de qual é o melhor método de co-cultura, e nem está claro se a co-cultura oferece resultado superiores as modernas formulações de meios sequenciais, nos quais tem sido demonstrado melhora nas porcentagens de blastocisto por transferência quando comparado com o meio tradicional.
Nessa revisão de literatura chegamos a conclusão que a maioria dos artigos publicados sobre co-cultura, sugerem um aumento significativa nas taxas de blastocisto (>50%), embora nenhuma ainda tivesse demonstrou que a co-cultura exerce um efeito benéfico. O conceito acima da xeno co-cultura e a falta de entendimento sobre os reais fatores que sustentam a idéia de aparente melhora que por algum tempo tivemos, co-cultura é ainda um modelo de interesse cientítico com o que se estuda as interações entre as células somáticas em culturas e os embriões humanos. Atualmente, os sistemas dos meios de cultura são definidos por estar começando com rendimentos comparando resultados para demonstrar que a maioria dos estudos de co-cultura e o campo está desenvolvendo rapidamente. De acordo com as circunstâncias, sentimos que a co-cultura tem muito para oferecer como ferramenta de pesquisa, além disso as pesquisas estão desejando caracterizar os fatores que são envolvidos com a visão de futura probabilidade que isso irá adicionar para definir o meio, que agora aparece felizmente suplantar a necessidade de co-cultura no futuro.
Introduction
Assisted Reproduction Technology (ART) brings new expectation for subfertile couples. Since the first experiences with animals (Thibault, 1949) and the first test tube baby in 1978, many discoveries have been made in an attempt to improve the chances of pregnancy in such couples. Numerous techniques in ART have been developed to improve results, but in the UK the average pregnancy rates (PR) vary between IS to 20% and approximately 15% take home baby rate (Human Fertilization and Embryology Authority, 1998). Some of the failings of ART have been blamed on suboptimal conditions in the culture media for human embryo development in vitro (Bongso & Ratnam, 1990). It is pertinent to note that the PR with ART in domestic animais is approximately 60% when the embryos are replaced in the uterus at blastocyst stage (Iritani, 1988). In fertile women fertilization occurs in the ampullary-isthmic region ofthe fallopian tube. The subsequent cleavage until the blastocyst stage takes approximately 4 to 5 days before the embryo descends to the uterus for implantation (Johnson & Everitt, 1990). In normal in vitro fertilization (IVF) procedures when the embryo is replaced at 4-6 cell stage into the uterus, it will therefore be there for 90h or so before implantation. Thus IVF results may be affected detrimentally by this asynchronous transfer. Qn the other hand one factor which may also contribute to poor results are the inadequate culture conditions which seem, to date, to be largely unable to support good embryo development in vitro to the blastocyst stage Bolton et al., 1989; Hardy et al., 1989; Bolton et al, 1990; Dokras et al., 1991; Dokras et al, 1993;) Some authors have tried to follow human embryo development and to keep them until the blastocyst stage for replacement (Bolton et al., 1989; Bolton et al, 1990; Dokras et al., 1991 ; Dokras et al, 1993; Hardy etal., 1989), since with early transfer, abnorrnalities of oocyte maturation, fertilization and early cleavage may not be manifested until later stages, following embryo transfer (ET) when it is no longer possible to visually monitor the progress of pre-implantation development (Bolton et.al., 1989; Hardy et al., 1989; Bolton et al , 1990; Dokras et al., 1991; Dokras et al, 1993; Gardner & Lane 1997). If the majority ofhuman IVF embryos succeeded in developing to blastocyst stage following embryo transfer; if we assumed that most blastocysts will hatch and then implant in the endometrium, then most pregnancy failures should be associated with positive ß-hCG results (socalled: biochemical pregnancies). Since, this is usually not the case, either our assumption is invalid or it seems likely that defective embryonic development prior to implantation may be one of the prime causes of failure in IVF-ET (Gardner & Lane, 1997). In many IVF clinics, probably in order to reduce the risk of fragmentation and degeneration of the human embryo they are replaced earlier (before 8-cells) (Gardner & Lane, 1997). As a matter of fact the large percentage of embryos showing a block at the 8 cell stage, when in culture, led Bavister (1988) to investigate the possible presence of artifacts in the culture medium or failure in genomic activation. It remains unclear though whether or not this might be produced by blocking/embryotoxic factors or indeed, whether failure of genome activation per se is a major factor in human IVF failure or not. Nevertheless, the possible existence of blocking/ embryotoxic factors gains credence when considering the findings of Gandolfi et al., (1989 a,b) who showed that the culture of sheep embryos on sheep oviductal cells enhanced the development and the viability of the embryos. Since then a number of authors have begun research into co-culture, using a variety of animais such as cow, goat, monkey and rat (for example: bovine kidney cells, monkey kidney cells (Vero Cell) and buffalo rat liver cells). The co-culture technique has also been used in humans including a large variety of feeder cells: oviductal cells, fallopian tube cells, endometrium cells and luteinized granulosa cells. Co-culture of human embryos with any of the above seems to increase the PR and also provides an opportunity to further evaluate the embryo (since most studies have looked at blastocyst transfer) before replacement (Bongso & Fong 1993 ; Bongso & Fong 1995; Guerin & Nicollet 1997).
Growth factors and embryonic blocks
Until recently, the culture of human embryos up to the expanded blastocyst stage has been suboptimal under existing in vitro conditions in most IVF centers. (Gardner & Lane, 1997; Bavister, 1988; Bavister, 1992; Bavister, 1995). Even under the strictest quality control measures and the best laboratory conditions only 25 - 30% of "excess" embryos cleave normally to the expanded blastocyst stage (Bolton et al., 1989; Dokras et al., 1991; Hardy et al., 1989; Fehilly et al., 1985). The objective of the co-culture system, therefore, is to provide an environment for the embryo that contains beneficial factors, which might more c10sely simulate those conditions in the human fallopian tube than existing culture media. There are two hypotheses to explain how the co-culture cells might improve the fertilization rate and enhance embryo development. First: specific and nonspecific embryotrophic factors might add to and improve culture conditions and second, the co-culture cells might remove undesirable factors from the culture medium. (Bongso et al., 1991a; Bongso et al 1991b). In several species of animais including the human, embryotoxic factors can block the embryo development ofdifferent stages in vitro (Gardner & Lane, 1997; Bavister, 1988; Bavister, 1992; Bavister, 1995 ; Bongso et al, 1988). This block appears to happen when activation of the genome (Brande et al., 1988) occurs and the co-culture system appears to overcome it in most mammals (Bongso & Fong, 1993; Bongso et al., 1995 ; Menezo et al., 1995), which, it is generally postulated, may involve non-specific factors such as glycoproteins (which may or may not be growth factors) being secreted by the feeder cells, thereby facilitating /embryonic development and bypassing any block factors mammals (Bongso & Fong, 1993 ; Bongso et al, 1995; Menezo et al., 1995). Some specific glycoproteins have been described in animais (Gandolfi et al., 1989a), and the molecular weight of proteins varies between 36 and 215 kDa and can be found in different stages of the female cycle, whilst moreover, Butzow, 1989 has demonstrated the presence of a 36 kDa glycoprotein (placental protein 5) in the secretory phase of the human menstruai cycle. However, as yet, no unequivocal proof has been put forward that any of these factors specifically play a role in the co-culture systems we have discussed to date. More definitive studies are needed. The mode of action ofthe putative growth factors need to be determined by studying the release of glycoprotein(s) by the co-culture cells into conditioned medium. Presumably these factors may simply diffuse into the embryo via the zona pellucida or act by direct contact between zona and ampullary cell allowing cross-transfer of glycoprotein(s) and other important metabolites. Bongso & Fong (1993) has suggested an explanation for the action of coculture on embryos in both positive and negative aspects as compared with traditional media. The negative aspects he attributes to non co-culture situations include:
Presence of hypoxanthine, which is known to cause the twocell, block in mouse embryo.The positive conditions which he subscribes to be the benefits of co-culture, such as the presence of embryotrophic factors:
Need for stabilization of the biochemical in vitro environment (for example, pH, 02, C02) and decreasing oxygen metabolite leveis.
Need for redesign ofthe medium by reducing glucose and increasing lactate leveis.
The need to use patients serum in simple glucose containing media.
Possible secretion of antioxidant taurine by the cells which in tum improves embryonic development.
The belief that low-molecular weight fractions in cellconditioned medium induce a positive effect.
Postulated glycoproteins released by the cells.
Postulated growth factors (such as transforrning growth factor TGF 131, insulin like growth factor IGF) released by the cells.
Co-Culture Systems
Much has been written about the evaluation of embryos and it is generally believed that embryos which divide evenly and a at a rate of one cell division every 20-24 hours are the embryos with the highest viability when replaced in the uterus. (Bolton et al., 1989; Dokras et. al ., 1991, 1993; Hardy et al., 1989; Gardner & Lane, 1997; Bavister, 1992, 1995; Hartshorne et al, 1991; Trounson et. al., 1982). However, many examples have been noted when retarded and unevenly cleaved embryos develop to term so there are no absolute criteria available at the present to determine which embryos are viable and which are not. (Bolton et al., 1989; Dokras et al., 1991, 1993; Hardy et al., 1989; Gardner & Lane, 1997; Bavister, 1992, 1995; Hartshorne et al, 1991; Trounson et al., 1982). In an attempt to determine whether in vitro culture methods might be improved upon, co-culture systems have been developed in which embryos are grown on a layer of somatic helper cells (Gandolfi et al., 1989a, b; Bongso & Fong 1993, Bongso, 1995 ; Menezo et al, 1995) , since it is believed that this approach may provide a stimulus to embryonic development. This approach may, in the long run, help to define the requirements of embryos during development. Trophoblastic vesicles have been found to be useful only in animais according to Camous et al. 1984.
Other studies have also demonstrated the beneficial effect of various cellular monolayers on the development of mammalian embryos, for instance uterine fi broblasts (Wiemer et al , 1989a,b), human tubular cells (Bongso et al, 1991 a, b; Bongso et al, 1990), and Vero Cells (Huang et al, 1997; Menezo et al, 1990) have all been used in an attempt to enhance early in vitro human embryo development.
Human Co-Culture System
A logical choice for co-culture system are cells from the human fallopian tube, although the ability of such cells, cultured as monolayers, to mimic their in vivo function per se, is somewhat doubtful. Verhage et al., 1979, have shown that complex changes occur in these ciliated and secretory epithelial cells during the menstrual cycle. In the early foll icular phase, hypertrophy and reciliation occurs, although in the human fallopian tube it is thought that the cilia only play a secondary role in gamete and embryo transport. Then as the cycle progresses, in the late follicular phase, the epithelial cells attain maximum height and ciliation and finally, at the end of the luteal phase, atrophy and deciliation occurs. Jansen et al. (1984), observed that around ovulation the isthmus secretions became abundant when estrogen levels were high at the midcycle, and this is possibly an important regulatory role in sperm transport through the isthmus. Oviductal fluid is therefore a crucial environment in which the movement of ovum and spermatozoa, fertilization, embryo transport and early development takes place. Glucose, lactate and pyruvate support sperm, oocyte and embryo survival. (Bongso et al., 1995 ; Leese, 1988). According to Gardner & Lane (1996) the lactate and glucose concentration in the oviduct changes with the day of the cycle and the human embryo is exposed to different metabolite concentrations as it moves along the tract.
Furthermore, cumulus cells readily consume glucose, producing lactate. Therefore, the early human embryo is exposed to low glucose and high lactate leveis in vivo. Thus, embryos in culture are likely to be in a suboptimal environment, which even in the presence of human ampullary epithelial cells is unlikely to mimic the in vivo environment. Nevertheless, Bongso et al. (1989), reports that the support of human embryo cleavage and growth with human ampullary epithelial cells in vitro resulted in a reduction in the cleavage abnormalities, producing good quality embryos with equal size blastomeres and minimal fragmentation when compared with conventional medium. Bongso et al. (l992a), have also evaluated embryonic behaviour in vitro, examining the pregnancy and implantation rates of embryos grown to blastocysts in co-culture with human ampullary cells in fifty women undergoing a single IVF treatment cycle. This study demonstrated that the expanded blastocyst rates in co-culture were significantly better than in the control patients (44% compared with 23%, p<O.OI). In this study a significantly higher percentage of good quality four cell stage embryos (regular blastomeres, absent to slight fragmentation) were also observed. In a similar study (Bongso et al., 1992b). II has also been shown that high success rates with IVF involving co-culture can be achieved (44% per cycle, implantation rate (IR) of 31.8%).Other possibilities for co-culture are granulosa cells, for instance, Dirnfeld et al. (1997) have shown an improvement in embryo quality in the short term development of early stage embryos and concluded that poor quality embryos may be rescued to cleave regularly in the presence of autologous human granulosa cells. Another system for study is human endometrial cells, obtained after hysterectomy (Bongso et al, 1991 b). Gland and stromal cells can be separated and established as cell lines for periods of over one menstrual cycle. Monolayers may be established in 3-7 days and in vitro cell growth has been found to consist of a mixed growth of epithelial fibroblast-like cell types (Bongso et al., 1991 b). Future studies could also be developed to characterise the exact in vitro behavior of the endometrium at the time of embryo replacement. The secretory cells from endometrial cell lines may affect the embryo, perhaps through the release of unknown/unspecified implantation factors or signals which it is speculated may eventually yield improved pregnancy rates (Birkenfeld & Navot, 1991).
Animal Co-Culture System
Co-culture of human embryos with various cellular monolayers has been described in the section beforehand, however, there is often a difficulty in obtainin·g samples from human cells, and furthermore there remain issues over possible bacterial or viral contamination. In animals, there remains similar potential for contamination, but they represent a more readily available source of cells for co-culture, although it should be borne in mind that the dangers of xeno co-culture remain an unknown quantity and significant problems may yet arise in the future. Animal feeder cells have been tried successfully by Rexroad and Powell (1988) who used ovine oviductal epithelial cells cultured in vitro and the subculture cells from established cell lines for co-culture with ovine embryos. Soon after Wiemer et al. (1989a, b) found markedly improved human embryo morphology when they were cocultured for one day on monolayers of fetal bovine uterine fibroblast. Increased blastomere diameter was the most evident morphological feature of embryos that developed in the presence of helper cells.
Menezo et al. (1990) used immortalised ceh lines of extra genital origin such as Vero Cell, which originate from kidney epithelium of African Green Monkeys. These cells were selected because kidney and genital tract have a common embryological origin (mesoderm). Moreover, Vero Cell are perhaps the most appropriate form of animal cells to use for coeculture with human embryos, since they are highly controlled for viruses and other contaminants because they are used for vaccine production. In this study the authors showed that > 50% of embryos of poor quality could reach the blastocyst stage when co-cultured. The Vero Cell co-culture system with B2 medium gives better results in terms of rate of blastocyst formation, with best grade of embryos. This system appears to have good development potential (61 % vs.3% of the control group). This observation suggests that in this case, the co-culture system improved human embryonic development, rescuing them from degeneration. Zetova et al., (1993) also studied the use of Vero Cell Monolayers (no complications reported) in standard culture medium, achieving a pregnancy rate (PR) of 33.33%, and an IR of 19.1 %. Guerin and Nicollet (1997) reported on a total of 1603 co-cultures performed by 11 groups over a 2 year period. The mean rate of clinical pregnancies per transfer after co-culture was 32.9% which was significantly higher than the mean value obtained by using conventional IVF procedures. They concluded that almost half of the cleavage embryos were able to develop into blastocyst during co-culture with Vero Cell monolayers, and the subsequent IR of these blastocysts were approximately twice as high as that of four-cell embryo. They suggest that previous successive failure with IVF was a main indicator for treatment with co-culture. Buffalo Rat Liver cells (BRL) have also been used for the co-culture of human embryos obtained from poor prognosis patients, and have been reported to have a positive effect on both the implantation and the clinical pregnancy rates (Hu et al, 1997; Hu et al., 1998) have also reported that co-culture with BRL cells and assisted hatching in a population of 200 first-time IVF patients was able to achieve a 58% clinical PR and a 49% live birth rate (26% IR). The rate of abnormalities detected in uterus and at birth was not different from that observed in the general population. AIthough Guerin and Nicollet, (1997) found that there was an apparent high frequency of trisomies, mainly trisomy 21 (0.8%), the majority of these cases corresponded to situations in which the mother was aged > 37 years. Finally, although comparison between human and animals co-culture are rare, it has been done by Feng et al., (1996) who studied different culture systems including human granulosa monolayer cells, co-culture with bovine oviductal epithelial cells (BOEC) arid co-culture with bovine uterine epithelial cells (BUEC). This study indicated that co-culture with BOEC and BUEC is more efficient than co-culture with human granulosa cells in enhancing the development of embryos to the eight-cell stage at 72h after egg retrieval. This result only serves to highlight the fact that in vitro behavior of cell types are unlikely to mimic exactly their in vivo function.
Blastecyst Culture Using Defined Media
So far, we have seen the potential of co-culture techniques, but there are disadvantages, such as the exposure to animal cells, factors released by such cells and the possibility of crossinfection (Gardner & Lane, 1997; Bavister, 1992, 1995). With defined media, there is no need for the relatively more complicated culture methods needed with co-culture (Gardner & Lane, 1997). Early attempts to improve culture methods Bolton et al., (1989); Bolton et al., (1991); Hardy et al., (1989); Quinn et al., (1997) yielded relatively poor results (e.g. low rates of blastocyst formation or poor implantation rates), hence the impetus for co-culture (Gandolfi et al., 1989a; Gandolfi et al 1989b; Bongso & Fong, 1993; Bongso & Fong, 1995; Menezo et al., 1990; Menezo et al 1995). However, recent developments in formulations of defined media have resulted in a shift in philosophy amongst those carrying out human IVF (Gardner & Lane, 1997; Barnes et al., 1995; Jones et al., 1998). Until recently, embryo' transfer at the 4-cell stage had been established global practice, in part, tacit acknowledgement of the deficiencies in culture media (Gardner & Lane, 1997; Desai & Goldfarb, 1998; Desai et al., 1997). Better understanding of the physiology of preimplantation mammalian embryos has resulted in improvement in the contents of media (Gardner & Lane, 1997; Bavister & Mckierman, 1993; Gardner et al., 1998; Scholtes & Zeilmaker, 1996). The advances in media formulations, means that the move towards blastocyst stage human embryo culture and transfer has taken a significant step forward. Now, as with co-culture, up to 50% or more of ali embryos derived through human IVF may develop on to the blastocyst stage. In particular, the most recent developments suggest that sequential media, which take into account the changing milieu ofthe human reproductive tract provide good blastocyst culture results (Gardner & Lane 1997). Taking into consideration the changing needs of a precompaction and a postcompaction embryo with regards for their needs for energy (Iactate, pyruvate, glucose etc) and the developing need for non-essential amino acids, the use for two defined media has provided data more than comparable with the best ofco-culture (Gardner & Lane, 1997). Altogether, these results in addition to criticism of co-culture by others, Van Blerkom, (1993); Sakkas, (1994) provide evidence that the use ofdefined culture media, as opposed to co-culture methods for obtaining blastocyst in human IVF may now have edged ahead.
For those who share concerns about co-culture, advari,ces in formulation technology of defined media, allows for a real glimpse of a future where, defined media may provide a simpler and safer method for blastocyst culture.
Discussion and conclusion
Numerous investigators have performed the application of coculture systems to in vitro developing human embryos. The aim of these researchers has been to see if co-culture is able to provide improved results. These studies suggest that, in certain hands, the use of co-culture procedures may have a significant beneficial impact on IVF (Gandolfi et al., 1989a, b; Bongso & Fong 1993, 1995; Guerin & Nicollet, 1997). Many authors also suggest that co-culture offers a significant alternative for patients with previously failed IVF replacement attempts (Bongso & Fong, 1993, 1995; Guerin & Nicollet, 1997; Menezo et al, 1995; Wiemer et al, 1989a,b). A number of different somatic cells have been used for co-culture, but there is no clear evidence that any particular cell type is more effective than another. For instance some authors have attempted to provide a rationale for the use of oviductal cell co-culture, by suggesting that oviductal epithelial cells in vitro may provide an environment which mimics the natural environment more closely (Gandolfi et al., 1989a, b; Bongso & Fong, 1993 e 1995). Conversely, they argue that the absence of oviductallike influences in vitro may compromise the viability of IVF embryos. Unfortunately, there is little or no evidence to suggest that oviductal cells in culture are able to behave as they do in vivo. Furthermore, it would seem that co-culture with any somatic cell and even monkey Vero cells are able to provide similar results (Menezo et al., 1995; Hu Y. et al., 1997 e 1998; Feng et al, 1996).
It hasbeen argued that the improvement(s) offered by co-culture centre on improved quality and/or development of embryos, arising as a consequence perhaps from the following two reasons: (i) empirical use of co-culture to improve embryo development and/or viability and (ii) the belief that the coculture technology itself has an embryotrophic effect through some secretory or buffering property of the somatic cells (Gandolfi et al., 1989a, b; Bongso & Fong 1993 e 1995; Bavister, 1992). A further issue arising from research done on co-culture, is that it might offer the possibility ofenhanced embryo viability and consequently the development of routine blastocyst transfer in IVF programmes (Bongso & Fong 1995 e 1993). When transferring human blastocysts, it is believed that there may be a synchrony between blastocyst transfer and endometrial receptivity (Hartshorne et al., 1991; Olivennes et al., 1994; Lelaidier et al., 1995).
Other advantages of such a procedure include the possible selection of embryos with a higher implantation potential Nakayama et al. (1995), supporting the concept that some (but not necessarily all) embryos with chromosomal and genetic abnormalities may be excluded by their failure to reach the blastocyst stage. Finally, another additional benefit of coculture, is supposed to be the availability of larger numbers of supernumerary embryos for freezing, since it is believed that fewer blastocysts are needed for embryo transfer because of their supposed enhanced viability. Since cryopreservation of human embryos fertilised in vitro, has become an established clinical procedure (Trounson & Mohr, 1983; Lassalle & Renard, 1985; Testart et al., 1986; Cohen et al., 1988) the replacement therefore of cryopreserved spare human embryos provides infertility patients with further chances of having a live baby. As a counter to the idea ofco-culture, it is important to consider that a number of improvements in culture techniques that have taken place over the last decade, allowing acceptable rates of development of human embryos to blastocyst stage (Gardner & Lane, 1997). Although the question of which is the best method remains unanswered, it is clear that several techniques are able to provide a good percentage of blastocyst formation after the use of defined medium culture. Moreover, Gardner et al. (1998) have demonstrated the successful use of two sequential serum-free media formulated for the culture ofhuman zygotes to the blastocyst stage. Designated G 1 and G2, these media were formulated around the leveis of carbohydrates present in the human oviduct and uterine fluids at the time when the embryo is present. They found that viable human blastocysts can be obtained in such sequentiál culture media in the absence of co-culture and serum and an improvement in PRs were observed when day 5 transfer occurred. This observation opens again the question of whether co-culture is really necessary to improve results or if other important considerations related to implantation rate Iike: age, quality of embryos replaced, quality and duration of culture, previous IVF cycle, pregnancies, uterine anomalies and general health problems, are more important with respect to treatment outcome (Bavister 1995 e 1992; Van Blerkom, 1993 ; Sakkas et al., 1994). Bavister, 1992 e 1995 has argued that a great deal of the enthusiasm for co-culture as a means to improve embryo development is probably misplaced. In Iight of recent reports Gardner & Lane, 1997; Jones et al., 1998; Bavister, 1993 his arguments are finding increasing support world-wide. Coculture, it seems, remains somewhat of an enigma. Clearly in some hands, it produces excellent results; certainly comparable with those obtained with defined media (Gardner & Lane, 1997; Bongso & Fong 1995 e 1993; Guerin & Nicollet, 1997; Menezo et al., 1995), however its role in the wider world may rest in its use as an experimental clinical tool, whose use will perhaps in time, allow us to characterise the factor(s) that might be beneficial to human embryos developing in culture, with a view to the future Iikelihood that these will subsequently be added to defined media, which now appear likely to supersede the need for co-culture.
Acknowledgements
WC would like to aeknowledge the following: the support by way of a travelling Bursary from the Botueatu Medical Sehool (for WC), the kind help and adviee from Professor J. G. Franco lr, Sue Smith and Kathryn Parkinson of the Portland Hospital for allowing aeeess to the exeellent facilities and lastly, but by no means least, ali the staff at the Portland Hospital Fertility Unit for their kindness and hospitality.
Butzow R. - The human fallopian tube contains placental protein 5. Hum Reprod; 4, 17-20. 1989.
Jansen R. P. - Endoerine response in the fallopian tube. Endoerinol. Rev. 5, 525-51, 1984.
Leese H. J. - The formation and funetion of oviduet fluido J. Reprod. Fertil. 82, 843-56, 1988.