Albumin and transferrin synthesis in whole rat embryo cultures

At about 10·5 days gestation in the rat, the liver diverticulum emerges from the embryonic gut. The rat liver diverticulum in organ culture, will, after 3 days, release plasma proteins into the medium (Parsa & Flancbaum, 1975). This suggests that the ability to synthesize plasma proteins is a property which is acquired by the liver at a very early stage in its development. It is of interest to determine when plasma-protein-synthesizing cells emerge from the primitive gut and to learn whether the ability to synthesize the different plasma proteins such as albumin and transferrin is acquired collectively or sequentially. Furthermore, it is of interest to determine whether extrahepatic sites for plasma protein synthesis exist and to assess the relative importance of these sites during embryonic development.

In a previous study using minced material from rat embryos and membranes, Yeoh & Morgan (1974) reported that incorporation of [¹⁴C] leucine into transferrin could be detected in foetal membranes but not the foetus at 13 days gestation. Albumin synthesis was present in preparations from both tissues. These results suggested that the liver may acquire the capacity for albumin synthesis prior to transferrin synthesis. Furthermore they indicated that extrahepatic sites of albumin and transferrin synthesis may reside in the extraembryonic membranes. Indeed, it has been shown in the mouse that transferrin is synthesized in the visceral yolk sac at an equivalent developmental stage (Janzen, Andrews & Tamaoki, 1982; Adamson, 1982). Using isolated hepatocytes derived from rat foetuses of varying gestational age maintained in culture, Yeoh, Wassenberg, Edkins & Oliver (1979) demonstrated that hepatocytes from as early as 12 days gestation secreted measurable amounts of both albumin and transferrin into the medium after 1 day in culture. Taken together these results suggest that at this stage of development, the foetal liver and hence the whole foetus as well as the extraembryonic membranes are capable of synthesizing transferrin. Failure to detect its synthesis in whole embryonic preparations may be due to a rapid turnover of the protein in that situation. Since albumin is detected in preparations from the whole embryo as well as extrahepatic membranes it is possible that it is not broken down as rapidly as transferrin in the whole embryo. In this study, albumin and transferrin synthesis were examined for by the incorporation of [³H]leucine into immunoprecipitates obtained with specific antiserum. To ensure specificity of the product, and to detect possible breakdown products, the immunoprecipitates were analysed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS–PAGE).

The Wistar Albino strain of rats was used throughout this study and was obtained from the Animal Resource Centre, Murdoch, WA. Timed matings were determined by the presence of spermatozoa in vaginal smears (taken as day 0).

Embryos were explanted on day 12 of gestation (36 ±0·5 somites, Witschi stage 21–22; Witschi, 1962) and prepared for culture by the method of Cockroft (1973). The culture medium used contained 75% Modified Eagles minimal essential medium (Flow Laboratories, Annandale, NSW) and 25% rat serum prepared by the method of Steele & New (1974). This was supplemented with glutamine (2·4 MM final concentration) and penicillin/streptomycin (100i.u.ml⁻¹ and 100 μg ml⁻¹ respectively; Grand Island Biological Co., NY). Embryos were cultured in 60 ml all-glass reagent bottles using the roller bottle method (Priscott, Yeoh & Oliver, 1984). The cultures were gassed with a mixture of 60% O₂/5% CO₂/35% N₂ and maintained at 37 °C. Embryos so cultured for 24 h remained viable and continued to differentiate normally, having 51·3 ±0·6 somites on completion of experiments. Initial experiments were conducted on 12-day conceptuses cultured for 24 h in the presence of 5 μCiml⁻¹ [³H]leucine (Radiochemical Centre, Amersham). Results from these indicated only limited amounts of radiolabelled albumin were present in both embryo and yolk-sac extracts. As a follow up to these results conceptuses were pulse chased in 25 μCi ml⁻¹ [³H]leucine for 2 h then transferred to normal culture media for time intervals of 0,1,2,4,8 and 12 h. In one experiment, embryos and yolk sacs were divided and cultured separately for a total of 4h. Yolk sacs and embryos were immunoprecipitated separately and in groups of four in all experiments.

After culture, tissues were washed three times in balanced salt solution, transferred to 3 ml of buffer A (20mM-Tris-HCl, pH 7-6, 0-14M-NaCl, 5HIM-EDTA, 1% Triton X-100) and sonicated. Samples were then centrifuged at 3000g for 5 min. 800/d aliquots of the remaining supernatants were immunoprecipitated for albumin or transferrin according to the method of Gross et al. (1982). Elution was achieved by incubation with 4% SDS, 10 % 2-mercaptoethanol, 20% glycerol and 0·25 M-Tris-HCl pH 6·8 at 90°C for 5 min. The eluted proteins were immediately subjected to SDS–PAGE (Cashman & Pitot, 1971). Samples were run at a constant current of ImA/gel for the first hour and then increased to 2mA/gel. Electrophoresed gels were frozen and cut into 2 mm slices and the radioactivity counted in individual slices in 5 ml gel counting solvent.

To determine the reliability of the immunoprecipitation technique 19-day foetal rat livers were cultured in primary monolayers for 48 h and incubated in the presence of 5 μCiml⁻¹ [³H]leucine for a subsequent 16h. SDS gel profiles of the immunoprecipitates are shown in Fig. 1. Based on these findings this immunoprecipitation protocol was applied to the embryo cultures. Initial experiments were conducted on whole conceptuses cultured for 24 h in the presence of 5 μCi ml⁻¹ [³H]leucine. SDS-PAGE profiles (Fig. 2) indicated albumin and transferrin may be present but at very low levels. Throughout all experiments the position of the immunoprecipitate peak along the gel was reproducible within ±2 mm. Additional peaks in the gels were initially thought to be due to nonspecific binding of synthesized protein to protein A and hence three clearing steps whereby protein A was added and the extracts centrifuged, prior to the addition of specific antibody, were incorporated into the protocol. Rabbit anti-rat IgG and protein A–Sepharose 4B were also tested as a superior alternative immunoadsorbent. Protein A–Sepharose 4B proved the most efficient immunoadsorbent with the least degree of non-specific binding and was used throughout subsequent experiments. However it was still necessary to incorporate blanks for each sample. These contained non-immune rabbit serum, and the radioactivity found in these was subtracted from the immunoprecipitates containing anti-albumin and antitransferrin respectively. It was possible that the low molecular weight proteins which coprecipitated with albumin or transferrin represented degraded products of the respective proteins. To test this, two series of experiments were conducted. Firstly, pulse-chase experiments were conducted on the whole conceptus. Cultures were preincubated for 4h in normal culture medium, pulsed for 2h in 25 μCiml⁻¹ [³H] leucine and then transferred to normal culture media for various time intervals. SDS–PAGE profiles (Fig. 3) showed an increase in the amount of radioactive transferrin in the embryo extracts and a corresponding fall in amount in the yolk-sac extracts during the 12 h in normal media. The presence of the low molecular weight material that was coprecipitated with specific antiserum was not evident until after 8 h in normal culture medium where it was most noticeable in yolk-sac extracts. Immunoprecipitation for albumin in the same tissue extracts could not be distinguished from background radioactivity. Subsequent cultures were therefore immunoprecipitated for transferrin only.

To further substantiate the possibility of degradation products a second experiment was conducted using radioiodinated proteins. ¹²⁵I-transferrin was added to the culture media and the whole embryos cultured for 24 h. Culture media, yolk sac and embryos were immunoprecipitated for transferrin. Only the native form of transferrin was found in media samples both before and after culture. However, in yolk-sac extracts there were proteolytic breakdown products of varying sizes in addition to the native transferrin (Fig. 4). Similar results were obtained when ¹²⁵I-albumin was added to the culture media, except that no native albumin was seen, indicating a more rapid proteolysis for that protein. No immunoreactive radiolabelled transferrin or albumin was found in the embryo.

To gain further insight into the potential transfer of transferrin from the yolk-sac membrane, yolk sacs were dissected free of the embryo and each cultured separately under conditions similar to those aforementioned for 4h (Fig. 5). Embryos remained viable during this time as judged by the presence of beating hearts. The extent of development was not determined. Gel profiles showed a decrease in transferrin levels in the yolk sac, similar to that observed with the whole conceptus (Fig. 3). However, in the absence of the yolk sac, there was no accumulation of radioactive transferrin in the embryo. This suggests that the increasing amounts of transferrin seen in the embryo when the whole conceptus was cultured were the result of active transport from the yolk sac.

Our experiments provide evidence that in the rat the visceral yolk sac is the primary site of transferrin synthesis and that a proportion of the protein is subsequently transported to the embryo. In contrast, we could not find evidence for albumin synthesis in either the embryo or yolk sac at this stage of development. Previous studies in our laboratories have demonstrated albumin synthesis in 13-to 14-day embryos (Yeoh & Morgan, 1974) and 12-day embryonic hepatocytes after 24h in culture (Yeoh et al. 1979). Furthermore, studies using 11-day liver organ cultures showed that albumin was secreted into the culture medium in detectable quantities by the third day, i.e. approximately equivalent to 14 days in vivo (Parsa & Flancbaum, 1975). Taken together these results suggest that albumin synthesis commences sometime on day 13.

The pattern of appearance of transferrin, on the other hand, is quite different. At 12 days gestation there is no embryonic production of the protein. This extends earlier in gestation the observation made by Yeoh & Morgan (1974) on 13-day embryonic material. In common with this earlier study definitive evidence of transferrin synthesis by the visceral yolk sac was obtained. The actual site of synthesis within the yolk sac was not studied by us, but may be in the sites of haemopoiesis in the blood islands that are still active at this stage (Fantoni, Lunadei & Ullu, 1975).

Unlike any previous study, we have shown that transferrin synthesized in the yolk sac is exported in significant quantities to the embryo. We suggest that the most likely pathway is by secretion into the vitelline circulation and thence to the embryo. Transferrin has a well-documented role as the principal iron-binding protein of serum (Morgan, 1981) and presumably also has this role in the early embryonic circulation. It has also been recently shown to be essential for the differentiation of mouse kidney tubules in vitro (Ekblom, Thesleff, Miettinen & Saxen, 1981) and it is therefore tempting to speculate that it may be involved as a specific growth factor in organ differentiation (Levin et al. 1984), perhaps together with other circulating proteins (Priscott, Gough & Barnes, 1983).

We have been careful to validate our protein detection methods and have confirmed the checks on antibody specificity made previously in this laboratory (Yeoh & Morgan, 1974; Yeoh et al. 1979). In developing the techniques for use with the whole-embryo culture method we became aware of greater problems, with regard to non-specific radioactivity, than we have encountered with hepatocyte cultures. This necessitated SDS–PAGE of immunoprecipitated proteins and subtraction of radioactive values derived from samples immunoprecipitated with non-immune serum. We believe this non-specific radioactivity probably derives from the complexity of the whole conceptus in culture. For example, the [³H] leucine will be taken up and utilized in a variety of biosynthetic pathways and is likely, therefore, to be present in a range of metabolic products in addition to the proteins we examined for specifically. Some of these associated with the immune complexes in a highly reproducible fashion as evidenced by characteristic patterns of radioactivity in gels after immunoprecipitation with non-immune serum. This was particularly evident in embryo preparations from the pulse-chase experiments where the radioactive peaks, in addition to being in the same positions, progressively increased in magnitude with time after the pulse. This was not a feature in the yolk-sac preparations and leads us to postulate that these peaks of radioactivity represent metabolic products unique to the embryo. Another confounding factor in the whole embryo culture technique is the proteolytic potential of the yolk sac for a variety of serum proteins (Freeman, Beck & Lloyd, 1981; Freeman & Lloyd, 1983) including both albumin and transferrin (McArdle & Priscott, 1984). We have confirmed that exogenous radioiodinated albumin and transferrin is proteolysed to smaller molecular weight immunoreactive fragments. Furthermore, we have obtained evidence that albumin is more susceptible to proteolytic breakdown than is transferrin in the yolk sac. It is thus conceivable that a low level of synthesis of albumin in the yolk sac would go undetected due to a rapid breakdown of the newly synthesized protein. However, proteolysis usually follows uptake from the apical surface of the yolk sac by pinocytosis (Lloyd, Williams, Moore & Beck, 1976). We have no direct evidence that proteins synthesized in either the embryo or yolk sac find their way to such sites of proteolysis from the proximal surface of the yolk sac. It is of interest that little intact exogenous transferrin reaches the embryo (McArdle & Priscott, 1984) and yet a high proportion of transferrin synthesized by the yolk sac does cross to the embryo. This suggests two pathways are in operation and may represent an evolutionary mechanism to ensure the embryo receives an adequate supply of an essential protein.

The present study demonstrates the usefulness of whole-embryo culture as an adjunct to other in vitro studies into embryonic development. In principal, the same approach could be applied to any protein synthesized by the embryo for which a specific antiserum is available.

We thank Dr H. J. McArdle for the ¹²⁵I-transferrin and ¹²⁵I-albumin. This work was supported by grants from the Raine Medical Research Foundation, the Australian Research Grants Scheme and the National Health and Medical Research Council.