Studies on the Transfer of Fertilized Mouse Eggs to Uterine Foster-Mothers: II. The Effect of Transferring Large Numbers of Eggs

In the previous paper of this series (McLaren & Michie, 1956) we described an experiment in which fertilized mouse eggs were transferred to the left uterine horn of recipient foster-mothers which were themselves pregnant with eggs of their own. Over the range tested (0–18 eggs), as the number of fertilized eggs transferred to the recipient female was increased, the number of embryos developing from them increased proportionately. We were thus unable to find a limit to the number of mouse embryos which could implant in a single uterine horn. There was, however, a suggestion that post-implantational mortality was greater when the total number of embryos implanted in a single horn was high.

The present work extends the range of inoculum sizes to thirty eggs, in order to investigate further the relation between implantation number and embryonic mortality, and also to determine whether a ‘law of diminishing returns ‘would set in when larger numbers of eggs were transferred.

As donors we used immature albino females induced to ovulate by hormone treatment and mated to albino males. The albino females were drawn from two random-bred strains. Some came from ‘The Mousery’, Rayleigh, Essex, and some from the TO (Theiler’s Original) strain maintained at the National Institute for Medical Research, Mill Hill, London. The albino males were from the TO strain.

As redpients we used female F₁ hybrids between the C3H and C57BL inbred strains mated to albino males of the TO strain. Since the recipients were homozygous for full colour, embryos of donor (alien) and recipient (native) origin could be distinguished at autopsy by their eye colour. All recipient females were adult and had previously given birth to one fitter.

Donor females were used $312$ days after mating, and recipient females $212$ days after mating. We had previously found this to be the most satisfactory combination.

Fertilized eggs were transferred to one uterine horn only (the left), so that its contents at autopsy could be assessed against those of the uninjected, control, horn.

Fourteen days after the operation (⁠$1612$ days post coitum) the recipients were killed. Each implantation in both horns was classified as ‘live alien embryo’, ‘live native embryo* or ‘dead’. The last group was further divided, according to the stage of pregnancy at which death was estimated to have occurred, into late deaths (13-16 days), middle deaths (10-12 days) and early deaths (before 10 days). We have discontinued the use of the terms ‘resorbing embryos ‘and ‘resorption rate ‘since we know of no compelling evidence that dead embryos undergo resorption in the mouse.

Other details of our experimental technique, including details of the egg transfer operation, are given in our earlier paper (McLaren & Michie, 1956).

The recipient females were divided into five groups, as follows:

1. Untreated control females.

2. Females which received a ‘dummy transfer’ of saline but no eggs into the left uterine horn.

3. Females which received five to ten eggs (‘low’ group).

4. Females which received fifteen to twenty eggs (‘middle’ group).

5. Females which received twenty-five to thirty eggs (‘high’ group).

The mice were selected for the five groups by a randomizing process which was subject to a restriction designed to spread the collection of data from the five groups more or less evenly over the same period of time.

In analysing the pregnancies obtained from the five groups, only surgically satisfactory operations were included. It had previously been found that gross surgical trauma affected embryonic survival. Recipients in which only alien embryos were found at autopsy have also been omitted, since they were presumably pseudopregnant, rather than pregnant, at the time of operation.

Table i shows that the pregnancy rate among the treated females (43/50 = 86%) did not differ significantly from that of the untreated control females (25/33 = 76 %). This confirms our earlier finding that the operation of egg transfer does not affect the pregnancy rate. The pregnancy rate among control females is significantly higher than the 57 % found in our earlier work. This can be attributed in part to our use on this occasion of F₁ hybrid females known for their high reproductive output, and in part to the improved conditions under which the females were kept prior to mating (in groups of not more than eight mice per cage, as against groups of twenty to thirty).

The reproductive superiority of the recipient females used in the present work over those used previously also manifests itself in the significantly higher number of implantations per horn in the control females (4·42 + 0·28 as compared with 3·39 ± 0·21). There is no significant correlation between the numbers of implantations in the two horns (r= −0·12). Once again the mean numbers of implantations in the right (uninjected) horns of the four groups of treated mice do not differ significantly from the corresponding value in the control females (see Table 1).

The low average number of implantations (3·85) in the right horn of ‘low group ‘females, and the high average number (6·10) in the right horn of ‘high group’ females, must represent chance fluctuations, apart from the small increment received by the ‘high group’ from the transmigration of embryos injected into the opposite horn (see Table 3 and McLaren & Michie, 1956). As stated above, females were selected at random for the various groups.

We concluded from our earlier work that the injection of saline alone (‘dummy transfers’) reduced by about one-third the expected number of implantations in the injected horn. Whether because of the improved quality of our recipient females, or through an improvement in our surgical technique, this effect was not apparent in the present work.

The data summarized in Table 1 were collected over a period of 18 weeks. When the yield of five embryos from transferred eggs is tabulated chronologically for the ‘low’, ‘middle’ and ‘high’ groups (Table 2), there appears to be an improvement of yield with time. A regression analysis confirms the significance of the effect (P < 0·05). The calculated regression lines for the three groups do not differ significantly from one another either in slope or in position. Using a common slope, the yields per injected egg adjusted to week 9·44 (i.e. the mean week) are 25·0, 18·9 and 22·2% in the ‘low’, ‘middle’ and ‘high’ groups, respectively. These percentages do not differ significantly from one another, and although the range of egg numbers has been doubled as compared with the previous experiment, there is still no consistent tendency for the yield per injected egg to fall off as the number of eggs injected is increased.

The time-trend, according to the regression line fitted to the three groups, amounts to an increase of about 1·7% per week, rising from 0·07 embryos per egg injected at the beginning of the experiment to 0·36 embryos at the end. This is presumably due to an increase of manual dexterity with time.

Fig. 1 shows the pregnancies in the three groups receiving eggs, arranged according to the yield of alien embryos found in each at autopsy. The class of zero yield is a small one except in the ‘low ‘group, where it can be largely accounted for through the loss of individual eggs. With relatively small inoculum sizes we expect a certain proportion of zero yields simply on the basis of random fluctuation. From our earlier work we concluded that over and above the random loss of individual eggs there was a second and distinct phenomenon of loss of the whole inoculum of eggs as a unit. We estimated that about one-third of all transfers failed through this cause. In the present experiment this hazard has been largely overcome, and accounts for the failure of perhaps 10% of transfers.

Table 3 gives the average number of implantations, classified as ‘five alien’, ‘five native’ and ‘dead’, in the right and left uterine horns of females of the five groups. As the number of eggs injected increases, the number of alien embryos in the injected horn increases, and the number of native embryos in the same horn decreases. A regression analysis indicates that there is a decline of 0·09 native embryos for each additional egg injected (P<0·01). This coefficient is the same as that found in our earlier experiment, but the phenomenon appears more striking in the present experiment as a greater range of inoculum sizes has been used.

Table 3 shows that in the ‘high’ group an average of less than one native embryo was found in the injected horn, as compared with five in the opposite horn. It is clear from the Table that the deficit of live native embryos cannot be accounted for from dead embryos visible at $1612$ days post coitum. We have independent unpublished evidence that embryos dying at a very early stage of pregnancy (c. 7 days) are still visible at 16-day autopsy. It follows that the suppression of native embryos by aliens must occur before or during the early stages of implantation and is not a result of post-implantational competition. The alien embryos, being a day ahead of the natives in developmental stage, are likely to implant first, and may render the uterus relatively refractory to the implantation of the native late arrivals.

There was significant evidence in our earlier work that when the number of implantations in the injected horn rose above the normal level, the implantation of native embryos in the uninjected horn was reduced. But in the new data there is no tendency for the number of implantations in the uninjected horn to decrease as the number of eggs injected increases.

We can therefore say that if the inhibitory effect upon implantation observed in our earlier experiment was a real one, it can only operate where the total number of implantations has been increased to a level approaching the reproductive capacity of the female, and that with our present females we have not reached this level. The frequency distributions of the total number of implantations in the recipient females of the two experiments are given in Fig. 2. The truncated distribution for the first experiment suggests a possible upper limit, or ‘ceiling’, to the number of implantations. The symmetrical distribution for the second experiment gives no evidence of any such ‘ceiling’.

The data on mortality are given in Table 4. In the control females the prenatal mortality rate was 10 %. In contrast to our previous finding there was no significant correlation between the two horns in respect of mortality. This bears witness to the greater uniformity of the females used in the present experiment. Of the twenty-two unsuccessful implantations, eighteen were classified as early deaths, four as middle deaths and none as late deaths.

The mortality in the uninjected horns of treated females did not differ significantly in extent or age distribution from that in the control females. In the injected horns of all groups the mortality rate was about double the control level and averaged 18·2%. The greater frequency of late deaths in the injected horn is not significant (P=0·1 by Fisher’s ‘exact method’).

Table 5 gives the mortality in each injected horn of the females receiving eggs, arranged according to the number of implantations in the horn. In our earlier experiment there was a significant tendency towards an increased mortality rate in the more crowded horns. No such effect is apparent in the present experiment, in spite of the increase in the degree of crowding. This finding is again in accord with the superior reproductive capacity of the females used in the second experiment.

Although the average yield of embryos per egg transferred was no higher in the present than in the previous experiment (just over 20%) the improvement with time during the present experiment meant that by the end we were achieving a yield of about 35%. This is comparable with the 40-5% yield achieved by Gates (1956), but not with the 50% yield which in our previous paper we concluded could be attained with our technique, given better control of technical and natural hazards.

The fact that our performance showed such marked improvement in the course of the present experiment, in spite of our having previously carried out some hundreds of similar egg transfer operations, suggest that the technique is difficult to standardize and requires modification before it can be recommended for general use.

The main aim of the present experiment was to investigate the reproductive limits of the female mouse, which we appeared to be nearing in our earlier experiment. But the F₁ hybrid mice used as recipients on the second occasion were so reproductively competent that in spite of doubling the effective range of the number of eggs injected we failed to detect any limits. Table 6 and Fig. 3 summarize the results of the two experiments.

It was not to be expected under the conditions of our experiment that a limit would be set by the capacity of the single injected uterine horn to accommodate the larger numbers of embryos. Hollander & Strong (1950) showed that unilaterally ovariectomized females gestate the normal number of live embryos per pregnancy : since in such pregnancies one uterine horn is empty, the average number of embryos gestated in the other horn is doubled. Even the number of eggs transferred in our ‘high’ group was not quite sufficient to double the expected number of implantations in the injected horn.

On the other hand, Bowman & Roberts (1958) have recently obtained results which are difficult to reconcile either with Hollander & Strong’s or with our own. These authors counted the corpora lutea and implantations in sixty-six normal pregnancies, and found that the proportion of fertilized eggs represented by implantations in a given uterine horn was inversely related to the number of corpora lutea in the ovary on that side. A similar relation held between the number of corpora lutea and the number of five embryos found at $1812$ days post coitum, i.e. the more eggs shed, the greater the proportion lost.

The data which Bowman & Roberts have published place the occurrence of this phenomenon in their material beyond reasonable doubt. Yet if a similar process had occurred in Hollander & Strong’s experiment, a substantial reduction in mean litter size should have been observed in their unilaterally spayed females. In the same way we would expect our own experiment to have shown a decline in the percentage yield of alien embryos with increasing numbers of eggs injected. But in fact, as we have seen, the yields obtained in the ‘low’, ‘middle’ and ‘high’ groups were approximately the same. The tendency which we found for the number of native implantations to diminish with increasing numbers of eggs injected is not strictly relevant to the effect described by Bowman & Roberts, which relates entirely to interaction between embryos at the same developmental stage as each other.

Leaving the question of single-horn limits, we may turn to the capacity of the female mouse as a whole to implant large numbers of embryos. Here we might have expected to find a limit. Yet in our material it proved possible to increase by about 50% the average number of implantations per pregnant female (i.e. from about 9 in the control and ‘dummy’ groups to about $1312$ in the ‘middle’ and ‘high’ groups) without provoking either a raised embryonic death rate or a lowered percentage yield of live embryos from the transferred eggs. The distribution of the number of live embryos in the ‘middle ‘and ‘high ‘groups was thus extended far beyond the control range, as shown in Fig. 4. This is surprising, since (in contrast to the situation studied by Hollander & Strong) the number of implantations must certainly have been raised above the number of corpora lutea available for their maintenance.

We are grateful to the Agricultural Research Council for financial support.

We have recently (McLaren & Michie, 1959) approached the problem by a different method: instead of increasing the number of implantations by the addition of alien eggs, the native quota was approximately doubled by the administration of gonadotrophins. In this way a limit was found above which foetal mortality sharply increased. This increase was restricted to deaths occurring after day 9, and only those embryos which survived the early period contributed to the effective crowding of the uterine horn. The females used were nulliparous, and foetal mortality rose when the number of implantations in a single horn exceeded eight.