The effect of epidermal growth factor, insulin and transferrin on the growth-promoting properties of serum depleted by repeated culture of postimplantation rat embryos

Although it has been apparent for some time that homologous serum has a finite growth-supporting capacity for postimplantation rat embryos cultured in vitro, the important biochemical factors involved have not been effectively identified. Some dilution of the serum is possible whilst still maintaining normal growth (Cockroft & Coppola, 1977; Huxham & Beck, 1985) but the studies performed indicate that 50% serum is required for growth to be comparable to that observed in utero. When the rat serum is replaced by human serum, even with the addition of excess glucose, growth of rat embryos is not completely normal unless a small proportion (10 %) of rat serum is added (Gupta & Beck, 1983; Lear et al. 1983). This implies that there are species-specific factors in rat serum which the embryo requires.

The nutritional role of the rat visceral yolk sac is well established (Freeman et al. 1981). In addition, it is now clear that the yolk sac transports many macromolecules to the embryo intact (Huxham & Beck, 1984, 1985). Indeed, the fact that many such factors are of importance to embryonic growth is demonstrated by the evidence that fluid contained within the cultured visceral yolk sac has some growthpromoting properties for rat embryos (Dunton et al. 1986).

It has been suggested by Klein et al. (1978) that certain specific proteins are depleted from serum when rat embryos are cultured at high density. Also Sanyal (1980) has shown that rat embryos mediate changes in the components of the culture medium, particularly in terms of pH, ⁠, glucose, urea and creatinine. More recently Priscott (Priscott et al. 1983) has to some degree extended the studies on serum proteins and has tentatively identified the proteins that are depleted by rat embryos in culture as α₂-macroglobulin and transferrin, amongst others.

In this study, we have attempted to ‘exhaust’ the growth-supporting capacity of rat serum by using it for repeated culture. We have assessed the extent of embryonic growth and development with recycled serum and have attempted to elucidate whether the limited growth observed was due to depletion of specific factors or to toxic changes in the serum brought about by the repeated culture. In addition, the serum has been supplemented with factors that may be depleted during repeated culture in an attempt to identify some of the specific agents involved.

Wistar rats were mated overnight and pregnancy was timed from midnight preceding the morning on which vaginal plugs were observed. Conceptuses were explanted at 9 ·5 days (early head-fold stage, containing three somites), according to the standard method of rodent embryo culture described by New (1978). Briefly, explanted embryos were cultured in homologous sera (see below) at 37 °C in glass bottles in a roller incubator. At the start of culture the incubation bottles were gassed with 5 % O₂: 5 % CO₂: 90 % N₂, after 24 h with 20% O₂: 5 % CO₂: 75 % N₂ and after 44 h with 40 % O₂: 5 % CO₂: 55 % N₂.

Serum was prepared from both male and female rats by immediate centrifugation and stored at – 18 °C until used. Before culture the serum was heat-inactivated at 56 °C for 30min. From a pool of approximately 50ml serum, 10ml was reserved as control serum and was treated in exactly the same way as experimental serum except that it was cultured without embryos. In this way, the effects of incubation, freezing and thawing and filtration were eliminated. The remainder was used as culture medium for embryos from 9 ·5 –11 ·5 days of development at a density of 1 embryo ml ⁻¹. After culture for 50 h the serum was frozen and stored. It was then sterilized by passage through a Millex-GV 0-22 pm filter unit (Millipore ^R) and reused for culture of further embryos with or without the addition of 2 mg ml ⁻¹ glucose and 10 μl ml ⁻¹MEM vitamin concentrate (Flow Laboratories). A small sample of medium was removed at each stage for biochemical analysis. Serum was recycled in this way until the culture of explants resulted in no significant growth as assessed by the morphological scoring system of Brown & Fabro (1981) which assesses the extent of growth and differentiation of 13 different embryological features including development of neural tube, limb bud, heart, otic and optic systems and yolk sac. In addition, measurements were made of protein levels (Lowry et al. 1951), yolk sac diameter, crown-rump length and somite number. Such serum was considered to be ‘exhausted’.

‘Exhausted’ serum and whole serum were then dialysed using Visking size 2 tubing (Medicell International Ltd, London, UK) at 4 °C against glucose-free balanced salt solution (Cockcroft, 1979) for 2 days with two changes of dialysis medium. After sterilization dialysed sera were used to culture embryos with the addition of glucose (2 mg ml ⁻¹) and vitamin supplement (MEM Vitamin concentrate, 10 μl ml ⁻¹). Such sera, both with and without dialysis, were supplemented with various concentrations (10 –50%) of whole rat serum and the growth of embryos was assessed as before. In addition ‘exhausted’ dialysed serum supplemented with glucose and vitamins was also used with the addition of insulin (0 –25 ng ml ⁻¹), epidermal growth factor (0 –25 ng ml ⁻¹) or rat transferrin. The extent of growth and differentiation of embryos was assessed after a total of 50 h in culture as above. Epidermal growth factor (from mouse submaxillary gland) and insulin (porcine, crystalline, 0 ·5 % Zn) were obtained from Sigma Chemical Co. Ltd, Poole, Dorset, UK. Rat transferrin was prepared by the method of Schreiber et al. (1979).

The pH of samples of sera were measured with a Pye Unicam digital pH meter. Care was taken to keep samples in airtight containers before measurements were performed to avoid any changes in pH. The osmolality was measured using a Halbmikro osmometer.

Lactate levels were measured using a diagnostic kit from Sigma Chemical Co., glucose levels were measured by the glucose oxidase method (kit from Sigma Chemical Co.) and total protein was assessed by a modification of the Lowry method (Lowry et al. 1951).

The levels of insulin in fresh and recycled sera were assayed using a double antibody radiolabelling method (Wellcome Laboratories).

Table 1 shows that the growth of embryos in serum that has been recycled once is greatly retarded. If the serum was used again hardly any embryonic growth was obtained. Even when glucose (2 mg ml ⁻¹) and vitamins (10 μl ml ⁻¹ minimum essential medium concentrate) were added back to recycled serum, embryonic growth was severely impaired. Control embryos grown in serum subjected to the same regime of freezing and thawing grew normally.

When recycled serum was dialysed (to remove any low molecular weight toxic waste products) and glucose and vitamins added, only limited improvement in growth was observed, indicating that the loss of growth-supporting properties can only be partially reversed in this way. After similar dialysis of whole serum embryonic growth was hardly affected in terms of morphological score or yolk sac diameter, although there was some effect on the protein level of the embryos.

When recycled serum is supplemented with whole rat serum at concentrations between 10 and 50%, there is an improvement in the growth and differentiation of embryos as shown by protein content, morphological score and yolk sac diameter. However, even at 50 % supplementation growth is not restored to the levels observed with control serum (Fig. 1A-C). It has previously been shown that supplementation of Hank’s balanced salt solution with 50% rat serum permits normal embryonic growth (Huxham & Beck, 1985), and therefore it is likely that these results are indicative of the presence of toxic moieties in ‘exhausted’ serum. If recycled serum is dialysed prior to supplementation as above, it is found that the restoration of growth occurs at much lower whole rat serum concentrations, and indeed growth is completely normal for all parameters at 30% supplementation (Fig. 2A-C).

Table 2 shows that whilst repeated culture has no effect on the total protein content of serum, the levels of glucose fall dramatically with a concomitant rise in lactate levels. In order to investigate the effects of high lactate concentrations on the growth of embryos, lactate (25 mmol l ⁻¹) was added to control serum. At this level, lactate severely inhibited embryonic growth (Table 3).

Repeated culture of embryos in serum causes a slight drop in the pH of the medium which is not observed in the absence of embryos. The osmolality of the medium was found to rise with repeated culture, and to a lesser extent in control serum incubated, frozen and thawed in the same way (Table 4).

Tables 2 and 4 also show that dialysis restores lactate. pH and osmolality to levels comparable with those found in control serum.

Insulin is known to have growth-promoting properties for many cell types and fetal tissues (Underwood & D’Ercole, 1984; Gluckman. 1986). It was, therefore, of interest to measure the changes in insulin levels of serum with repeated culture. Table 5 shows that the insulin levels fall with repeated culture, suggesting that hormones and related molecules may be specifically depleted by the embryos under such culture conditions.

Addition of porcine insulin at concentrations between 5 and 25 ng ml ⁻¹ slightly improved the growth and differentiation of embryos grown in dialysed ‘exhausted’ serum supplemented with glucose and vitamins, although there was no significant improvement above 10 ng ml ⁻¹ (Fig. 3A-C). At concentrations above 50 ng ml ⁻¹ the addition of insulin had a detrimental effect (results not shown). When insulin was added to whole rat serum at concentrations less than 50 ng ml ⁻¹ it was without effect, whereas concentrations above this level proved to be embryotoxic.

When epidermal growth factor was added to dialysed recycled serum, an improvement in the growth and differentiation of cultured embryos was observed with increasing concentrations over the range 5 –25 ng ml ⁻¹ (Fig. 4A-C).

Addition of rat transferrin at 500 μg ml ⁻¹ caused an improvement in embryonic growth (Fig. 5). In addition, the embryos were observed to have numerous, apparently normal, red blood cells in the circulation, whereas in those cultured in unsupplemented sera, when the circulation was sufficiently well developed for this to be assessed, the blood cells were colourless. The other factors (with the exception of whole serum supplementation) added to recycled serum did not have an effect on the overt anaemia.

Embryonic growth is clearly poorer in recycled serum than in fresh serum (Table 1). This is in general agreement with the work of New et al. (1976). Klein et al. (1978) and Sanyal (1980). It is possible to suggest several reasons for this. One explanation is that when serum is used in this way the repeated freezing, thawing and filtration of the serum renders it incapable of supporting normal growth and differentiation of embryos. However, when serum is treated in exactly the same way but no embryos are cultured in it, it is still able to support normal embryonic development. There are two other explanations for this effect; one is that the growth of embryos in serum adds factors to serum, possibly waste products, which render it toxic to further embryos; the other is that the embryos deplete material from the serum, possibly proteins or hormones, which makes further embryonic growth impossible. This study goes some way to elucidating the mechanism by which serum is ‘exhausted’.

Our data indicate that as glucose in serum is used by embryos there is a build up of lactate (Table 2). Since other workers have shown that at this stage of gestation the embryonic metabolism is based primarily on glycolysis (Tanimura & Shepherd, 1970), it is likely that lactate is the major breakdown product of carbohydrate metabolism. Although we have shown that lactate itself very severely inhibits embryonic growth when present at a concentration of 25 mmol l ⁻¹ (Table 3), at a level of around 15 mmol 1 ⁻¹, as found at the end of a single embryo culture (Table 2), the effect is far less marked providing the glucose levels are restored to normal.

This is shown by results obtained following the second passage plus glucose and vitamins (Table 1). However, when lactate is removed by dialysis of serum following the third passage, then glucose is added, very poor growth of embryos is obtained. The other potentially harmful waste products such as urea, uric acid and creatinine (Sanyal, 1980) are of similar molecular weight to lactate and are therefore also likely to be removed by dialysis. It is possible, however, that some nondialysable toxic agents are formed during embryos culture and thus inhibit further embryonic growth.

When whole rat serum was used to supplement ‘exhausted’ serum, it was not possible to restore its growth-promoting properties even when it was supplemented at the 50% level (Fig. 1). When Hank’s balanced salt solution and whole rat serum are used in a 1:1 (v/v) ratio, normal growth ensues (Huxham & Beck, 1985). This is further evidence that one of the causes of the lack of growth of embryos in ‘exhausted’ serum is the presence of toxic metabolites.

The removal of low molecular weight toxic materials by mild dialysis has the effect of permitting normal embryonic growth when 30 % whole serum supplementation is employed. 30% levels of whole serum supplementation are insufficient to support normal embryonic development when added to balanced salt solution only. It is clear therefore that low molecular weight toxic metabolites are released into the serum as it is recycled for culture of embryos, but since the adverse effect can be removed by dialysis, the presence of other higher molecular weight toxic materials is excluded. However, since normal growth can only be obtained when such dialysed recycled sera are supplemented with whole rat serum, it is probable that some growth-promoting materials, which are normally present in serum at low levels, are depleted during repeated culture so that they become growth limiting. Many growth factors and hormones fall into this category.

Other workers (Klein et al. 1978; Priscott et al. 1983) have suggested that certain specific proteins, including transferrin, α₂ macroglobulin, α₁, lipoprotein and ceruloplasmin, and also some uncharacterized proteins of 125 and >200 ×10³ relative molecular mass are depleted from serum by embryo culture. Although we present no comprehensive evidence for this, it is certainly the case that insulin levels drop quite markedly with repeated culture (Table 5). It has recently been shown that extremely low insulin levels found in serum specifically depleted by column chromatography preclude normal embryonic growth in culture, the level for this effect to be apparent being less than 0 ·5 ng ml ⁻¹ (Travers & Beck, 1988).

Three specific factors have been chosen for investigation: insulin, epidermal growth factor and transferrin. Previous work (Calvert et al. 1986; Andrews et al. 1987; Cumberland et al. 1987) has suggested a role for these factors in the growth and development of rat embryos in vitro.

Insulin improved growth and development with a maximum effect at around 10 ng ml ⁻¹, which is close to the physiological level. These results are interesting in regard to the recent data indicating that normoglycaemic serum, specifically depleted of insulin using an affinity column, is not able to support normal growth and development, whilst this effect can be reversed by addition of insulin at 10 ng ml ⁻¹ (Travers & Beck, 1988).

The role of insulin and insulin-like growth factors (IGFs) in the regulation of fetal growth is fairly well established (Gluckman, 1986), but the role of these factors at earlier stages of gestation is less well understood. The receptor for insulin has been identified on both embryonic tissues and extraembryonic membranes for rat conceptuses at gestational ages 10 and 12 days (Unterman et al. 1986), and recent studies have identified receptors for IGF-I and IGF-II, as well as insulin, in mouse embryos at the stage of organogenesis (Smith et al. 1987). It has been known for some time that the major fetal somatomedin in both humans and rats is IGF-II, and also that receptors for both IGF-I and IGF-II are present in fetal tissues (Sara & Hall, 1984). With the advent of in situ hybridization histochemistry, it has been possible to demonstrate recently the presence of mRNA for IGF-II in both human (Han et al. 1987) and rat (Beck et al. 1987) embryonic tissues as well, although only low levels were present during the early stages of organogenesis, as studied here. In the rat study it was also shown that very little IGF-I is produced by the conceptus at early gestational ages. The earliest identification of insulin in rat embryonic pancreas by immunocytochemistry was at 13 ·5 days of gestation (Travers & Beck, 1988), which suggests that any insulin requirement by the embryo at earlier stages must be provided from maternal sources. When embryos are grown in vitro the source of exogenous factors is the serum culture medium.

When epidermal growth factor was added to serum depleted of factors by repeated use for the culture of rat embryos, it was found to improve the growth of further embryos in a dose-dependent manner. The concentrations at which a plateauing of this effect was observed (15 –25 ng ml ⁻¹) are in the physiological range. It has been proposed that effects observed at such concentrations of ligand are likely to be by interactions with a highly specific receptor (Gospodorowicz, 1981; Heath & Rees, 1985; Pratt, 1983). The presence of an EGF receptor on fetal mouse and rat tissues has been reported (Hortsch et al. 1983), although the same study failed to identify receptors at early stages, especially in the yolk sac. Earlier studies (Nexo et al. 1980) had shown the presence of both receptors for EGF and an endogenous growth factor, later identified as transforming growth factor α (TGFα). Mouse embryonic tissues have been demonstrated to contain TGFα (Twardzik, 1982) which acts via the EGF receptor. It is, however, possible that exogenous EGF is required before the establishment of EGF or TGFα synthesis within the embryo (Rizzino, 1987) and that, in the model used here, this is a factor which becomes depleted from serum by recycling.

Transferrin had two effects when added to recycled serum; first it improved the growth of the embryos, and second it averted the anaemia observed with unsupplemented serum. The latter effect is not surprising since transferrin has a major role in iron transport into cells, as well as being essential for cell viability (Morgan et al. 1978) and is therefore likely to have effects on haemoglobin synthesis. In addition, it has previously been reported that the level of transferrin in serum is decreased by embryonic culture (Priscott et al. 1983). There have been reports that transferrin is synthesized by the visceral yolk sac (Williams et al. 1986) but it is clear that a large proportion of the embryonic transferrin requirement in the postimplantation conceptus is of maternal origin and is transported across the yolk sac placenta (Huxham & Beck, 1985; Cumberland et al. 1987).

The improved growth and differentiation of embryos in recycled serum in the presence of exogenous rat transferrin may be entirely due to its irontransporting role, since many enzyme systems in cells require iron or other metals for their activity. However, it has been suggested that transferrin and its receptor may have a growth-promoting role for cultured cells independent of any iron-transport function. Studies performed using alternative sources of iron carrier for the cells or occupancy of the receptor by antibodies against the receptor have attempted to separate the iron-transport function from other potential intracellular signalling events which may occur during receptor occupancy and internalization (Ekblom & Thesleff, 1985; Brock et al. 1986; Landschulz et al. 1984; and Beck et al. 1987). It remains unclear whether or not there are two independent actions of transferrin in this context, and whether or not this applies to the effect on embryonic growth is difficult to ascertain.

Not surprisingly, in no case was it possible to restore growth and differentiation of rat embryos to the level observed when whole rat serum supplementation was used by the addition of a single growth promoting factor. However, several reviews discuss a role for such molecules in embryonic and fetal development (e.g. Adamson, 1983; Heldin & Westermark, 1984; Rizzino, 1987) and it is likely that the factors added in this study are normally provided by the maternal system. Whilst evidence is now accumulating regarding the synthetic capability of the embryonic tissues for growth factors (e.g. Williams et al. 1986; Han et al. 1987; Beck et al. 1987), in most cases synthesis is only at a significant level in the late organogenetic period and it is possible that the early postimplantation embryo, as studied here, is dependent on maternal factors for the regulation of its growth and differentiation.

Thanks are due to the Wellcome Trust and the Medical Research Council for grants in aid of this research. The skilled technical assistance of Timothy Jefferson is also gratefully acknowledged.