ABSTRACT
Fetal fluid was studied in mice of the Q, C57BL/McL, C3H/Bi/McL and JU/Fa strains. The amount increases to day 16 of gestation, and then decreases. Amniotic fluid and fetal plasma contain three transferrins, five α-fetoproteins and albumin. Protein concentration in both increases during gestation; glucose decreases in amniotic fluid although it rises in fetal plasma; urea levels remain constant. The regulation of volume and biochemical composition of amniotic fluid, and its relation to the yolk-sac membrane, are discussed.
INTRODUCTION
Fetal fluids make up a large proportion of the mammalian conceptus during gestation. They alter in a number of ways throughout pregnancy, but neither their origin nor their role in development is clear: in an early review, for example, Needham (1931) gave more than 20 hypotheses for their origin. Recent studies have shown that substances in the fetal fluids may originate from either the maternal circulation, the fetal circulation or the uterine lumen: the relative importance of each source varies according to the species (Brambell, 1970).
Johnson (1971) observed volume changes of mouse sac fluid (amniotic fluid and exocoelomic fluid) during pregnancy in four strains of mice, one (C57BL/Gr) having consistently larger amounts of fluid than the others. Crosses between C57BL and CBA mice indicated a maternal effect on volume of fluid. Gustine & Zimmerman (1972, 1973), on the other hand, demonstrated a fetal origin of α-fetoproteins (αFP) 1–5 and transferrins (Trf) 1 and 2 in mouse amniotic fluid in vitro, and showed that the changes in protein composition of amniotic fluid in vivo reflected those that had occurred in fetal plasma 24 h previously. Amniotic fluid Trf3 is also of fetal origin in the mouse (Renfree & McLaren,1974), unlike in man where transferrin is of maternal origin (Usategui-Gomez & Morgan, 1966). The volume of extra-embryonic fluids in the rat is largely regulated by the yolk sac (Payne & Deuchar, 1972), which forms the main nutritive organ of the embryo before the formation of a functional allantoic placenta (New & Brent, 1972). In the mouse the yolk sac is similarly well developed, and completely surrounds the amnion for the greater part of gestation.
In the present study we describe, for four strains of mice, the changes that occur during pregnancy in fetal fluid volume and in the levels of protein, glucose and urea in maternal plasma, fetal plasma and amniotic fluid.
METHODS
Collection of samples
Female mice of three inbred strains (C57BL/McL, C3H/Bi/McL and JU/Fa) and of one randomly bred strain (Q), were checked daily for vaginal plugs. The day of the plug was considered day 0 of gestation. Pregnant females were exsanguinated by decapitation on days 11–19 of gestation and the maternal plasma collected in heparinized centrifuge tubes. Uteri were removed and individual fetuses dissected from the uterus with embryonic membranes intact. To minimize evaporation, they were placed in 30 mm Petri dishes with lids and kept on ice throughout the weighing process. Before weighing, each embryo was blotted on filter paper moistened with distilled water, and weighed (‘intact wet weight’ ). Amniotic fluid was collected by insertion of a 26-gauge needle directly into the amniotic cavity, carefully avoiding puncture of vitelline or allantoic vessels. The membranes and associated chorio-allantoic placentae were then removed from the embryos, blotted again to remove all traces of fetal fluids and both they and the embryo were re-weighed separately. Fetal plasma was collected in 10µl heparinized capillary tubes from the severed end of the umbilicus (days 11–13) or from a small incision made in the external carotids (days 14–19). A constant amount of blood (as measured by the distance along the capillary tube) was collected from each fetus on any particular day of gestation. In all, 528 conceptuses from 72 females were treated in this way. The total weight of the fetal fluids was obtained by subtraction of the wet weight of embryo, placenta and membranes from the intact wet weight. The embryos and placentae (with membranes) were then lyophilized, after which they were placed in an oven at 150 °C for 24 h before reweighing to obtain dry weight. All plasma and fluid samples were kept frozen at − 70 °C until used for analysis.
In addition to the maternal and fetal samples, plasma was collected from three adult mice and their new-born (1 day post-partum) young. Stomach contents were collected by aspiration of the fluid with a fine pipette through the stomach wall from six litters of embryos on day 18 of pregnancy.
Biochemical analysis
(i) Electrophoresis
The polyacrylamide slab gel electrophoresis system of Reid and Bieleski (1968) was used. Gels (″ (1.586 mm) thick) were poured using the solutions described in Davis (1964). Twenty-five µl of amniotic fluid and 2.5µl of fetal and maternal plasma in 20 µl of bromophenol blue in 40% sucrose was applied to each well. Each slab was run at 30 mA, 100 volts for
h (until the bromophenol tracking dye reached the bottom of the slab), stained in 0.025% Coomassie Brilliant Blue and 0.1% Amido Black in 25% methanol and 7% acetic acid in water for 45 min. Destaining was by diffusion for 3 h in 25% methanol and 7% acetic acid in water. Gels were photographed immediately after destaining.
(ii) Quantitative analyses
Protein concentration was determined by the Lowry technique, using bovine serum albumin as the standard (Lowry, Rosebrough, Farr & Randall, 1951). Glucose analysis followed the glucose oxidase-peroxidase method of Werner, Rey & Wielinger (1970). Urea was measured by the colorimetric assay of Fawcett & Scott (1960) and Chaney & Marbach (1962).
Statistical methods
The statistical methods used were those of simple and partial regression and of cubic polynomial regression, based on unweighted litter means.
Means were not weighted according to the number of fetuses present, because variation in embryo weight tends to rise as the number of fetuses increases (Healy, McLaren & Michie, 1960).
RESULTS
The changes in the relative proportions of embryo, placenta and extra-embryonic fluid were similar for each of the four strains studied. In the Q strain (Fig. 1), placental weight makes up the smallest component of the conceptus, and after day 14 does not alter significantly. At day 11, fluid weight is greater than embryo wet weight, and steadily rises to a peak weight at day 16, where it is equal to 50% of the weight of the fetus. It then falls so that at day 18 the fluid weight is only 10% of the embryo wet weight.
Wet weight of the intact conceptus and its components: embryo, placenta with extra-embryonic membranes, and extra-embryonic fluid in Q strain mice between days 11 and 18 of gestation.
The most obvious difference between the inbred strains was in their rate of development, in particular that of the C3H strain (Fig. 2). At birth, Q embryos are slightly larger than C57BL, JU and C3H in that order. Placental weights are also lowest in C3H embryos, but do not differ from the other strains by so much. Fluid weights all show a peak but the time is different in the four strains, C3H reaching its highest values 1–2 days after the others. To eliminate this variation between strains, and also the slight variation of up to 12 h introduced by timing from the morning of the vaginal plug, all subsequent analyses relate fluid weight to embryo dry weight (‘embryo weight’ ).
Embryonic dry weight (a), placental dry weight (b) and extra-embryonic fluid weight (c) in each of the four strains of mice, Q (x), C57 (●), C3H (◼) and JU (▴) between days 11 and 19 of gestation.
The amount of fluid in each conceptus increases from an embryo weight of 0.003-0.08 g, after which levels drop. When fluid weight is plotted with reference to embryo weight, the apparent temporal differences between strains disappear (Fig. 3), and there is no significant difference between the cubic regressions of the four strains (F9.52 = 0.22).
Polynomial regression line of best fit for weight of extra-embryonic fluid on embryonic dry weight. Litter means of Q (x), C57 (●), C3H (◼) and JU (▴) plotted around the line.
Q mice tended to have larger litters than the inbred strains. Partial regression analyses fitting litter size to the cubic regressions of fluid volume on dry weight showed that litter size had no significant effect on the amount of fluid (b = −0.0014 ±0.0095, d.f. = 69).
Progressive changes were observed during gestation in the protein patterns of amniotic fluid and fetal plasma but, as expected, the maternal plasma pattern did not change (Fig. 4). There were no differences observed between the strains except in the transferrin region of the gel where C3H were Tfra/Tfra (Renfree & McLaren, 1974), while the other three strains were all Trfb/Trfb. Fetal plasma and amniotic fluid contained three main groups of proteins: albumins, α-feto-proteins (five bands) and transferrins (three bands). Towards the end of gestation Trf1 and Trf2 have almost disappeared, being replaced almost exclusively by Trf3 in both fetal plasma and amniotic fluid (Figs. 4 and 5). Similarly, the α-fetoprotein pattern changed from five bands in the early samples to aFP5 only in the latest samples. The changes in amniotic fluid followed those in fetal plasma after an interval of about 24 h.
Proteins in amniotic fluid (Amn), fetal plasma (Fp) and maternal plasma (Mp) from Q mice at days 13, 15 and 17 of gestation. Note the progressive changes from the three fetal transferrins to the one adult type, and the shift from five α-fetoproteins at day 13 to one (αFP5) in day 17 fetal plasma samples. Electropherograms of proteins in mouse fetal and maternal fluids. Trf, transferrins 1, 2 and 3; Hp, haemoglobin/haptoglobins; α-FP, α-fetoproteins 1–5; Alb, albumin.
Proteins in amniotic fluid (Amn), fetal plasma (Fp) and maternal plasma (Mp) from Q mice at days 13, 15 and 17 of gestation. Note the progressive changes from the three fetal transferrins to the one adult type, and the shift from five α-fetoproteins at day 13 to one (αFP5) in day 17 fetal plasma samples. Electropherograms of proteins in mouse fetal and maternal fluids. Trf, transferrins 1, 2 and 3; Hp, haemoglobin/haptoglobins; α-FP, α-fetoproteins 1–5; Alb, albumin.
Proteins in stomach contents (St) at day 18 compared to amniotic fluid (Amn) at days 12 and 14 and fetal plasma (Fp) at day 18. Note the similarity of stomach contents at day 18 to amniotic fluid at day 17 (Fig. 4).
Proteins in stomach contents (St) at day 18 compared to amniotic fluid (Amn) at days 12 and 14 and fetal plasma (Fp) at day 18. Note the similarity of stomach contents at day 18 to amniotic fluid at day 17 (Fig. 4).
Stomach contents had the characteristic high viscosity and appearance of term amniotic fluids, and a drop of this fluid could be expressed from the fetal mouth by squeezing the stomach. Acrylamide electrophoresis showed that the fetal stomach content collected at day 18 had an almost identical protein pattern to that of amniotic fluid at day 17: it resembled fetal plasma less in its relative proportions of the transferrins and a-fetoproteins (Fig. 5).
Regression analysis showed that, in relation to embryonic weight, the levels of protein, glucose and urea in maternal and fetal plasma, and in amniotic fluid, did not vary significantly among the four strains. For each fluid, the means and standard errors were subsequently calculated on pooled data from the four strains.
In maternal plasma, protein, glucose and urea concentrations showed no overall upward or downward trend during pregnancy (Fig. 6). The mean values for the three substances were 49·8 ± 14·7 mg/ml (n = 60), 127.7 ±28.7 mg% (n = 62), and 5·3 ±2·1 µM/ml (n = 63).
Urea (A), glucose (B) and protein (C) in amniotic fluid (◼), fetal plasma (▴) and maternal plasma (●). Means for each weight category ± standard errors (total determinations in amniotic fluid, n = 64, in fetal plasma, n = 45 and in maternal plasma, n = 60) plotted against embryonic dry weight. The last point on each graph for fetal plasma and maternal plasma are the samples from young 24 h after birth.
Urea (A), glucose (B) and protein (C) in amniotic fluid (◼), fetal plasma (▴) and maternal plasma (●). Means for each weight category ± standard errors (total determinations in amniotic fluid, n = 64, in fetal plasma, n = 45 and in maternal plasma, n = 60) plotted against embryonic dry weight. The last point on each graph for fetal plasma and maternal plasma are the samples from young 24 h after birth.
Protein concentration of amniotic fluid rose from 2·4 mg/ml to 4·5 mg/ml between days 11 and 18. Fetal plasma likewise showed an increase, from about 20% of the maternal level at day 11 to 50% by the end of gestation. Urea concentrations were similar in all three fluids throughout gestation.
Glucose, on the other hand, decreased in amniotic fluid from levels little below those in maternal plasma to almost unmeasurable levels at term. In fetuses weighing less than 50 mg fetal plasma was difficult to obtain completely uncontaminated from amniotic fluid, owing to the very small volumes of blood. This contamination is most evident in the glucose assays, where the two fluids differ greatly in concentration. However, in fetal plasma, the variations in glucose level were not statistically significant until the last days of gestation when the level rose steeply. Newborn plasma glucose levels were between those of term embryos and adults.
DISCUSSION
In all the strains studied the amount of extra-embryonic fluid increased up to about day 16 of gestation. However, it is only from day 14 that the embryo wet weight begins to exceed that of the fluid, so that for a great part of gestation the fluid represents a significant portion of the conceptus. The total fluid content of the conceptus is made up of yolk-sac fluid, exocoelomic fluid and amniotic fluid. On day 14 of gestation, the parietal yolk-sac membrane breaks down and the outer embryonic surface exposed to the uterine lumen is the visceral yolk sac (‘yolk-sac inversion’ ). The exocoel is of relatively small volume, so that in the final arrangement the amniotic membrane is closely covered by the yolk sac over most of its surface (Fig. 7). The contents of the amniotic cavity are therefore likely to be regulated by two membranes: the yolk sac and the amnion. In addition, the allantoic placenta and the allantoic vessels are closely applied to the amnion, but only over a very small area.
Relationships of the fetal membranes in the mouse at day 16 of gestation (diagrammatic). AC, amniotic cavity; AM, allantoic mesoderm; AMN, amnion; AV, allantoic vessels; D, decidua; EX, exocoelomic cavity; PYS, parietal yolk sac; T, trophoblast (cytotrophoblast and syncitiotrophoblast); UE, uterine epithelium; UL, uterine lumen; UW, uterine wall; VV, vitelline veins; VYS, visceral yolk sac; ecto., extra-embryonic ectoderm; som.tn., somatic mesoderm; spl.m., splanchnic mesoderm; y.s.endo, yolk-sac endoderm.
Relationships of the fetal membranes in the mouse at day 16 of gestation (diagrammatic). AC, amniotic cavity; AM, allantoic mesoderm; AMN, amnion; AV, allantoic vessels; D, decidua; EX, exocoelomic cavity; PYS, parietal yolk sac; T, trophoblast (cytotrophoblast and syncitiotrophoblast); UE, uterine epithelium; UL, uterine lumen; UW, uterine wall; VV, vitelline veins; VYS, visceral yolk sac; ecto., extra-embryonic ectoderm; som.tn., somatic mesoderm; spl.m., splanchnic mesoderm; y.s.endo, yolk-sac endoderm.
The changes in the fluid throughout gestation may be related to the anatomical arrangement. Up to the time of yolk-sac inversion, fluid accumulates within the fetal membranes. However, within a day or two of the breakdown of the parietal yolk sac and Reichert’ s membrane, the volume of the extra-embryonic fluid declines. This suggests that the yolk sac may play a role in regulating fluid volume.
Experiments on the functions of embryonic membranes in the rat in vitro shed considerable light on this problem. Anti-yolk sac antibody added in vitro to embryos with intact visceral yolk sacs caused gross retardation of growth and differentiation, whereas when it was added to the amniotic cavity or the exocoel, it had little or no effect (New & Brent, 1972). Using a series of expiants of day-10 rat embryos, Payne & Deuchar (1972) concluded that the amnion alone was not capable of regulating the fluid volume of the amniotic cavity. Where the amnion was the only membrane left intact, the amniotic cavity almost disappeared due to collapse, but the embryos had a lower death rate than those in which the amnion had been removed too, indicating that while it had little function in fluid regulation, it did have a protective function. When parietal and visceral yolk sacs, including Reichert’ s membrane, were left intact the embryos expanded, rotated, and incorporated tracer into protein and had higher total protein values. The visceral yolk sac alone was inferior to a complete yolk sac, but most functions were little changed. They concluded that the visceral yolk sac was of greatest importance in regulating the volume of embryonic fluids, but that its nutritive function was enhanced by the parietal yolk sac and Reichert’ s membrane.
In the present study, the decrease in fluid weight occurs after the breakdown of the outer membranes, perhaps indicating a gradual decline in the importance of the yolk sac relative to the allantoic placenta. The glucose levels in the amniotic fluid are highest in the early stages, and drop very low by the end of gestation, while urea is apparently at equilibrium with both fetal and maternal fluids throughout. Total protein concentration changes very little through gestation, rising only slightly over the time period measured in this study, although the changes in the relative proportions of the a-fetoproteins and in the transferrins are quite marked. The similarity of protein pattern of the stomach contents to amniotic fluid strongly suggests that mouse embryos are ‘drinking’ amniotic fluid in the last days of gestation.
This idea does not conflict with the observation of a rise in amniotic fluid protein concentration, nor with the increase in viscosity during the last two days of gestation. The absorptive cells of the visceral yolk-sac endoderm of the rat have ultrastructural features characteristic of protein-absorbing cells (Deren & Wilson, 1964; Deren, Padykula & Wilson, 1966; Padykula, Deren & Wilson, 1966). This placental membrane may continue to transfer macromolecules, and proteins in particular, even after the breakdown of the parietal yolk sac. Lambson (1966) suggests that proteins stored in yolk-sac endoderm cells may pass directly into the amniotic cavity late in pregnancy. Furthermore, since several, if not all of the proteins in the amniotic fluid are of fetal origin (Gustine & Zimmerman, 1972, 1973; Renfree & McLaren, 1974), the transfer will most likely be from the vitelline circulation or endoderm cells rather than the uterine lumen.
The yolk sac thus emerges from these studies as a membrane likely to play an important part in the growth of mouse embryos, especially in the earlier stages of gestation, while the function of the amnion is apparently more protective than nutritive. The amniotic fluid occupies a large volume relative to the embryo up to day 13, but its final volume may be limited not only by the transport functions of the yolk sac, but also by the tonus of the uterine muscle and pressure within the uterus. During days 12–16, the rate of weight increase of the total conceptus parallels that of the embryo, but thereafter the fluid decreases by 40%, so that the embryo rate exceeds that of the total conceptus. A reduction in fluid at the end of gestation would allow a continuation of fetal growth and maturation, without any corresponding increase in intra-uterine pressure.
ACKNOWLEDGEMENTS
We thank Dr G. A. Clayton, Dr W. G. Hill and Mrs J. Smith for assistance with the polynomial regression programmes. The Ford Foundation provided financial support.