A microinjection technique has been devised for labelling individual blastomeres of preimplantation mouse embryos with a marker drop of inert silicone fluid placed in the cytoplasm either at the periphery of the egg or at the interface between two blastomeres (i.e. centrally). The following observations were made on blastocysts which developed from injected eggs:

  • (1) All drops derived from peripheral injections (at two-cell to morula stage) were found exclusively in the trophoblast.

  • (2) Drops injected centrally at two-four-cell stages have been found in both the trophoblast and inner cell mass.

  • (3) Peripheral labelling of one or both members of a pair of eggs (four-cell to morula) fused into a chimaeric blastocyst yielded markers in both trophoblast and inner cell mass.

No evidence was found of inherent bilateral symmetry or polarity in the cleaving egg. The results indicate that physico-chemical positional effects determine whether a cell will differentiate into trophoblast or inner cell mass.

The results are discussed in the light of current hypotheses relating to early embryonic differentiation in the mammal, and it is suggested that cleavage occurs without spatial disturbance of the cytoplasmic pattern of the egg so that its cortical region is converted directly to the outer cells of the morula and hence to the trophoblast.

The differentiation of cell types which occurs as the rodent embryo develops into a blastocyst, with trophoblast cells and inner cell mass (ICM), has been the subject of some controversy in recent years. An interpretation put forward largely by Dalcq (1957) and Mulnard (1961) suggested that there were regional differences in the cytoplasm of the egg so that during cleavage there was segregation of special types of cytoplasm into so-called ‘dorsal’ and ‘ventral’ blastomeres. The dorsal and ventral cytoplasms were believed to predispose cells receiving them to form ICM and trophoblast respectively. This suggestion, that a mosaic cytoplasm in the egg governed differentiation, was based on cytochemical studies, in particular of RNA, and clearly implied that the developmental fate of the egg was predetermined.

However, experiments on isolated blastomeres of the two-cell embryo of the rabbit and mouse have shown that these blastomeres may develop into blastocysts which, on transfer to a suitable host female, may give rise to normal, fertile individuals (Seidel, 1960; Tarkowski, 1959). Seidel explained his results in the rabbit by assuming the presence, in the uncleaved egg, of an organizing centre whose presence in whole or part was essential for the formation of a true blastocyst. Tarkowski had suggested that in the mouse the regulative capacity of the two-cell-stage blastomeres might be influenced by their cytoplasmic characteristics. Later, however, together with Wroblewska (1967), he put forward an hypothesis suggesting that the fate of blastomeres remains labile at least until the eight-cell stage, their fate being determined epigenetically in the morula simply by their position in the embryo, i.e. whether the cells are on the outside of the mass or on the inside and so enclosed in an inter-cellular environment. The outer cells would then develop into trophoblast and the inner cells into the inner cell mass.

In the experiments reported here, we hoped to study the lineal relationship between particular regions of the cleaving egg and the components of the blastocyst and to show whether this could be classed as an epigenetic or predetermined pattern of development. A microinjection technique for introducing small drops of an inert fluid into cells as markers was initially devised for studying the process of orientation which occurs (normally) in the uterus as the blastocysts differentiate with the commencement of implantation (Jenkinson & Wilson, 1970). Using the same microinjection technique it has been possible to label individual blastomeres of the mouse egg at different stages of development and determine subsequently the position of the drop at the blastocyst stage. The position of the drop within the blastomeres can be controlled to some extent at injection so we have labelled either the peripheral cytoplasm or that lying near the interface of two blastomeres in the centre of a two- or four-cell egg. Eggs from the two-cell to the morula stages were used, and after appropriate labelling were cultured singly or as conjoined pairs.

Eggs were collected from spontaneously ovulating females of a hybrid strain (C57BL×A2G) by flushing the Fallopian tubes or uteri at various intervals after fertilization depending on the stages required (2-cell, 4-cell, 8-cell, 16-cell and morula stages). Initial experiments were carried out using a medium devised by Mulnard (1967), containing lactate and pyruvate, for two-four-cell eggs, and Earles balanced salt solution +10 % calf serum for eight-cell and subsequent stages. Subsequently it was found that a new medium formulated by Whittingham (1971) gave good development from the two-cell to blastocyst. The zona pellu-cida was removed by placing the eggs in a 0·5 % solution of pronase in Earles BSS and gently agitating with a mouth-controlled pipette. This could be done immediately after injection for the eight-cell and later stages, but eggs injected at the two-four-cell stage were left in their zonae till the blastocyst stage since they appeared susceptible to bursting on treatment with the pronase solution. Removal of the zona greatly facilitated identification of injected drops.

A Leitz micromanipulator was used to place the marker droplets in the individual blastomeres as the eggs were observed under a Laborlux microscope at a magnification of×320. An inert, non-toxic silicone fluid (Edwards Surgical Supplies Ltd., D.C. 550), which is immiscible with the cytoplasm, was generally used as the marker, though corn-oil is also well tolerated by the cell and was used in some early experiments. The injection pipette was made by drawing out a very fine capillary tube on a Palmer microelectrode puller, bending it by means of a de Fonbrune microforge, and very gently breaking the tip off by contact with a glass anvil -the last step being observed under the microscope with the tip in the desired injection fluid. The injection was performed by introducing this very fine pipette (tip diameter of less than 0·5 μm) into a hanging drop containing the eggs on a coverslip placed over a chamber containing liquid paraffin (Fig. 1). The volume of fluid injected was about 30 μm3. Two methods for holding the egg during the injection procedure were used :

Fig. 1.

Diagram of injection chamber: A, plan view; B, in section.

Fig. 1.

Diagram of injection chamber: A, plan view; B, in section.

  • (a) All the two-four-cell stages were held by compression, i.e. the volume of fluid in the hanging drop was reduced till the egg was trapped between the coverslip and liquid paraffin. The injection pipette could then pierce the zona pellucida and enter the blastomere.

  • (b) The alternative method was to use a pipette with an annealed tip of a diameter slightly less than the egg which was then held by means of gentle suction pressure.

After piercing the zona pellucida and introducing the pipette into the peripheral area of a blastomere a drop of the silicone fluid was forced out using a de Fonbrune suction and force pump. As the pipette is withdrawn the droplet adheres to the tip until it comes up against the plasma membrane, when it is ‘nipped off’ and lies close against the blastomere membrane. In order to place a drop directly against the interface between two blastomeresofa two- or four-cell, the injection pipette has to be passed completely through one blastomere and into the opposite blastomere so that the drop remains against the central membrane on withdrawal (Fig. 2). It is difficult to be certain that droplets injected thus are accurately located at the centre of the egg, as, especially in these early cleavage stages, the area of interblastomeric contact is small compared with the free surface area where peripheral injections are made. Some of the membrane in the region of blastomere contact at the two- or four-cell stage may become’externalized’ in succeeding divisions of the egg. Hence peripheral droplets are more likely to remain peripheral (as the results show) whereas central droplets may secondarily become peripheral. In order for a four-cell egg to be injected centrally it is now realized that the injection pipette must be passed obliquely through the egg. At the four-cell stage the egg appears as a regular tetrahedron, and the most satisfactory injection is likely to be obtained by passing through the ‘apical’ blastomere into any one of the other three (in the same plane). The low return of blastocysts with drops in the ICM from centrally injected eggs would seem to indicate, accepting an epigenetic theory, that the method was not satisfactory; however, the fact that we have obtained 3/20 eggs with a drop in the ICM proves that such an epigenetic hypothesis is tenable.

Fig. 2.

Method of injection of drops at the two-four-cell stage. Each pair of eggs represents the situation during and after completion of the injection process. A, Peripheral injection of a two-cell egg. B, Central injection of a four-cell egg held in position by compression.

Fig. 2.

Method of injection of drops at the two-four-cell stage. Each pair of eggs represents the situation during and after completion of the injection process. A, Peripheral injection of a two-cell egg. B, Central injection of a four-cell egg held in position by compression.

After injection the eggs were left to recover for 1 h or more in an incubator at 37 °C and were then transferred to a small culture chamber, in some instances being subjected to pronase treatment before this transfer.

The experiments can most easily be divided into three main groups.

  • (1) Eggs at the 2-cell, 4-cell, 8-cell, 16-cell and early morula stages were injected with a varying number of silicone fluid droplets (1–6) in the peripheral area of the cytoplasm of the blastomere(s). These eggs were then incubated until the blastocyst stage (Fig. 5 A).

  • (2) Eggs at the two-cell and four-cell stage were injected with one silicone fluid droplet placed centrally near the interface between the blastomeres. These, also, were incubated until blastocyst formation (Figs. 4A, 6A).

  • (3) Eggs at the 4-cell, 8-cell, 16-cell and early morula stage were injected peripherally with a number of silicone fluid droplets. After removal of the zona pellucida pairs of eggs were conjoined, with labelled + unlabelled or two labelled of the same developmental stage. Conjoined eggs were then incubated until the blastocyst stage, either in the Earles BSS + 30 % calf serum or in the Whittingham medium (Fig. 7 A). It was found that a greater proportion of the eggs cultured in the Earles BSS + 30 % calf serum differentiated into normal blastocysts than those cultured in the Whittingham medium. This was probably due to the greater adhesiveness promoted by the increased proportion of serum, so that the two eggs remained in close contact until cell intermingling was effected.

Fig. 3.

Diagrammatic representation of results showing position of drops at injection and then their position in the blastocyst. Eggs in group 1 and group 2 were cultured singly; in group 3 pairs of eggs were fused and then cultured in vitro. (Shading of the two-cell in group 2 indicates a burst blastomere.)

Fig. 3.

Diagrammatic representation of results showing position of drops at injection and then their position in the blastocyst. Eggs in group 1 and group 2 were cultured singly; in group 3 pairs of eggs were fused and then cultured in vitro. (Shading of the two-cell in group 2 indicates a burst blastomere.)

Fig. 4.

A, Two-cell mouse egg injected centrally with 1 drop, immediately after injection. B, Blastocyst developed from A with 1 drop in the ICM (the attached mass is the remains of the unlabelled blastomere which burst at the two-cell stage).

Fig. 4.

A, Two-cell mouse egg injected centrally with 1 drop, immediately after injection. B, Blastocyst developed from A with 1 drop in the ICM (the attached mass is the remains of the unlabelled blastomere which burst at the two-cell stage).

Fig. 5.

A, Four-cell mouse egg injected peripherally with 2 drops, immediately after injection. B, Blastocyst developed from A with 2 drops in the trophoblast.

Fig. 5.

A, Four-cell mouse egg injected peripherally with 2 drops, immediately after injection. B, Blastocyst developed from A with 2 drops in the trophoblast.

Fig. 6.

A, Four-cell mouse egg injected centrally with 1 drop, immediately after injection. B, Blastocyst developed from A with 1 drop in the ICM.

Fig. 6.

A, Four-cell mouse egg injected centrally with 1 drop, immediately after injection. B, Blastocyst developed from A with 1 drop in the ICM.

Fig. 7.

A, Pair of conjoined mouse eggs at the morula stage, one of which is labelled peripherally with 3 drops. B, Blastocyst developed from A with 1 drop in the ICM and 1 drop in the trophoblast (T). Only 2 of the 3 injected drops are shown as the third is rejected and not visible in this plane of focus.

Fig. 7.

A, Pair of conjoined mouse eggs at the morula stage, one of which is labelled peripherally with 3 drops. B, Blastocyst developed from A with 1 drop in the ICM and 1 drop in the trophoblast (T). Only 2 of the 3 injected drops are shown as the third is rejected and not visible in this plane of focus.

These three experimental groups are illustrated diagrammatically in Fig. 3.

At least two untreated eggs were kept as controls for each experiment. The control and treated eggs were then cultured until the blastocyst stage in a slide chamber containing small drops of the medium under liquid paraffin gassed at 5 % CO2 in air. The slide chamber was placed inside a larger perspex chamber gassed with the same mixture and incubated at 37 °C. The blastocysts were then examined in a hanging drop in the injection chamber to determine the position of the silicone fluid drops, whether they were situated in the trophoblast or inner mass cells. A complete examination was made possible by manipulation of the blastocysts within the hanging drop.

A number of silicone drops were found rejected, either completely outside the blastocyst as a naked drop or in a cell(s) not incorporated into the blastocyst, or apparently rejected into the blastocyst cavity, often in a cell larger than those of the trophoblast or ICM (presumably because it had stopped dividing) (Fig. 8 A, B). Some of the injected droplets became split into two or more smaller droplets, but these usually remained close together whilst some drops could not be traced. These facts account for any apparent discrepancy between the number of drops injected and the total number of drops as shown in the Tables.

Fig. 8.

A, Blastocyst with 1 drop rejected ‘externally’. B, Blastocyst with 1 drop enclosed in a cell rejected into the cavity.

Fig. 8.

A, Blastocyst with 1 drop rejected ‘externally’. B, Blastocyst with 1 drop enclosed in a cell rejected into the cavity.

Most drops were identified as definitely in the trophoblast or ICM though some are recorded as ‘unclassified’ since they occurred at the junction of the trophoblast and inner mass and it was impossible to locate them precisely. Only eggs developing into morphologically normal blastocysts are considered. Occasionally natural lipid droplets appear in large numbers in the fully formed blastocyst and this makes certain identification of the injected droplets difficult. Such blastocysts, which may be suffering from metabolic disturbances, have not been included in the results given here.

The results are presented in Tables 1, 2 and 3.

Table 1.

Group 1

Group 1
Group 1
Table 2.

Group 2

Group 2
Group 2
Table 3.

Group 3

Group 3
Group 3

The following main points may be noted:

  • (1) In all cases where drops were injected in the peripheral area of the cytoplasm, the drop or drops were later found solely in the trophoblast cells (Fig. 5 B).

  • (2) Injection of drops near the interface between the blastomeres of two-cell and four-cell eggs led subsequently to the drops being found in both the trophoblast and inner mass cells (Fig. 6B). In two instances in this group of experiments the blastomere through which the injection pipette had passed (at the two-cell stage) burst, but the surviving blastomere containing the drop developed to a fully expanded blastocyst with the remains of the burst blastomere still attached (Fig. 4B). These two eggs gave different results: in one the drop was in the trophoblast and in the other in the ICM.

    A lower return of blastocysts was obtained with eggs which were injected internally, this may be due to the timing of the injection in relation to the blastomere division. As far as possible, injections were carried out soon after a division had taken place, so that the passage of the micropipette through the blastomere created the least disturbance to the dividing nucleus and the spindle fibres, though it remains a more hazardous treatment of the egg than is peripheral injection.

  • (3) Fusi on of pairs of four-eight-cell eggs or morulae, one or both of which had been injected with drops, resulted in the drops being found in the tropho-blast and inner mass cells (Fig. 7B). With fusion at the 16-cell stage, drops were identified in the trophoblast cells only, an anomalous result which we can only explain as being due to chance intermingling of the cells during fusion such that those with drops become external.

On the basis of cytochemical studies Dalcq (1957) and Mulnard (1965) and colleagues believed that the eggs of both rat and mouse possessed both a bilateral symmetry and polarity. Two recent cytochemical observations cast a little doubt on some of their early work. Rodé, Damjanov & Skreb (1968) have examined whole mounts of the early stages of the rat embryo for acid phosphatase and alkaline phosphatase and found both to be present in practically all parts of the blastocyst; these results differ from those of Dalcq (1957) who demonstrated that the ICM gave a positive reaction when tested for the presence of alkaline phosphatase whereas the trophoblast contained no stainable material. However, in the mouse, Mulnard (1965) detected a diffuse staining of the ICM whilst the trophoblast contained only a few perinuclear granules with reactions for acid phosphatase. Rodé et al. suggested that the inner mass cells might appear to have a greater enzyme activity because of superposition of cells to give an optical effect and, also, possibly because of enzyme diffusion during incubation and other fixation artifacts. The second observation, made by Cerisola & Izquierdo (1969), was on mouse embryos from one-cell to the blastocyst stage which were studied for the presence of RNA and other basophilic material. As development proceeded they detected an overall increase in the basophilic material extractable by ribonuclease, particularly in the morula. However, the preparations did not reveal any spatial distribution of the material that might be interpreted as a plane of polarity or bilateral symmetry (unless extremely labile), though at the blastocyst stage the inner cell mass is more basophilic than the trophoblast. (The same criticism of superposition of cells to give an optical effect of greater enzyme activity must apply here also.)

Mulnard (1965) separated blastomeres of 2-cell mouse eggs and cultured both partners. On his and Dalcq’s hypothesis one would predict a segregation of cytoplasmic constituents in the blastomeres such that the two partners would develop in a complementary fashion. However, this was not the experimental result, as in some instances both blastomeres of a pair developed into trophoblastic vesicles, with no indication of ICM.

In experiments with the rabbit and the mouse Seidel (1960) and Tarkowski (1959) (respectively) found that when single blastomeres of the two-cell (1/2) and four-cell (1/4) stage were cultured, some developed into blastocysts apparently normal in structure though reduced in size (having only about one-half or one-quarter of the number of cells in normal blastocysts) whilst other blastomeres developed into anomalous forms such as ‘trophoblastic vesicles’. Seidel explained his results with rabbit by assuming the existence in the uncleaved egg of an organizing centre. Trophoblastic vesicles would be produced by blastomeres lacking the cytoplasm from this organizing centre. The latter idea was supported by his finding that there was a greater incidence of trophoblastic vesicles among forms from 1 /4 blastomeres than from 1 /2 blastomeres. Clearly on both Seidel’s and Dalcq/Mulnard’s hypotheses much would depend on the relationship between the planes of cleavage and the positions of the organizing centre or special types of cytoplasm. Tarkowski (1959) concluded that in the mouse the regulative capacity of two-cell stage blastomeres might be influenced by their cytoplasmic characteristics but that the cytoplasmic segregation hypothesis was tenable only if the first cleavage plane bears no constant relation to the plane of bilateral symmetry, so allowing for the obvious variability in the development of isolated 1 /2 or 1 /4 blastomeres.

Later work by Tarkowski & Wroblewska (1967) involved the culture to blastocysts of disaggregated blastomeres of four-cell and eight-cell mouse eggs. Blastocysts were obtained from 1/4, 2/8 and 1/8 blastomeres, though not with equal facility. Fewer blastocysts, as opposed to trophoblastic vesicles, were formed from pairs of blastomeres from the eight-cell than from single blastomeres of four-cell stages. With single blastomeres of the eight-cell stage, the incidence of blastocysts was still lower. To explain their results Tarkowski & Wroblewska put forward the idea of an epigenetically controlled differentiation of the blastomeres into trophoblast or inner mass cells. This would be dependent upon the position of the cells at the morula stage, i.e. whether a cell was external in the morula and exposed to the medium (trophoblast) or situated inside the morula, in an intercellular environment (ICM). If, as with 1/8 blastomeres, there are relatively few cells at the time of cavitation, then the chances of any of these being totally enclosed by others, so that they become ICM, are not very great and in fact only 15 % of 1/8 blastomeres formed blastocysts with ICM. Thus the number and arrangement of cells both play an important part, this being primarily a physical rather than a cytochemical condition.

Our results from group 1, where all the peripherally situated droplets in any egg from the two-cell to the early morula stage were subsequently found in the trophoblast, supports this idea of epigenesis. When, as in group 2, two-cell and four-cell eggs which were injected with one drop at the interface between two blastomeres developed with some of the drops in the inner cell mass, this demonstrates that the cytoplasm central to the egg mass may differentiate into inner cell mass. Not all the drops, however, end up in this position. This may in part be due to the fact, as explained in the Materials and Methods section, that it is much more difficult to be sure about the ‘centre’ of the egg, and drops which had been placed ‘centrally’ at these early stages could well become externalized with successive cleavage divisions.

Experiments performed by other workers demonstrate that when fusion of two eggs occurs there is not a complete intermingling or ordered rearrangement of the two types of cells, i.e. a random migration of blastomeres occurs (Mintz, 1965). It seems from the results of Tarkowski & Wroblewska (1967) and from our group 1 above that all blastomeres have the inherent property to differentiate and form trophoblastic cells, and certainly up to the eight-cell stage none are determined to form just inner cell mass, as Dalcq and Mulnard have suggested. The fusion experiments of our group 3 show that until the morula stage the outer blastomeres are labile and may form inner mass cells, probably when placed at the junction of the fusion where they would come into an intercellular environment. Further experiments in this laboratory, in which pairs of eggs of synchronous and asynchronous stages were fused, have also demonstrated the lability of the outer blastomeres up to the late morula stage in their ability to differentiate into inner mass cells (M. Susan Stern, in preparation). This substantiates the evidence of Mintz (1965) and shows that up to the morula stage the development, at least of the externally situated blastomeres, can be regulated.

In two cases from group 2, where the blastomere through which the injection pipette had passed subsequently burst, the remaining blastomere containing the drop developed into a fully expanded blastocyst. In one case the drop was found in the trophoblast and in the other in the ICM. This is of note since the drop was originally placed into the ‘central’ cytoplasm of the two-cell egg, and yet on bursting of the second blastomere the drop is now in the ‘peripheral’ cytoplasm. It remains to be seen if the presence of the burst blastomere still attached to the cleaving egg affects its development.

The experiments described here substantiate the accumulating evidence that there is no bilateral symmetry or polarity of significance for development of the mouse egg and that blastocyst differentiation is epigenetic. During cleavage the mouse egg is cellularized apparently without any significant disturbance of its spatial cytoplasmic organization, a feature which is characteristic of the cleaving eggs of most of the lower vertebrates. It is (presumably) the predetermined pattern of cleavage which determines the fate of each intrinsically totipotent region. In this manner, quite simply, the outer regions of the egg become the outer cells of the blastocyst and it would be surprising if this were not true for all eutherian mammals.

We are grateful to the Wellcome Trust for money to purchase the micro-manipulator equipment, and to the A.R.C. and S.R.C. for financial support. Mrs V. Roberts has provided skilled technical assistance.

Cerisola
,
H.
&
Izquierdo
,
L.
(
1969
).
Distribution del RNA y embriogenesis en el raton
.
Arch. Biol. Med. Exp
.
6
,
10
16
.
Dalcq
,
A. M.
(
1957
).
Introduction to General Embryology
, pp.
103
128
.
London
:
Oxford University Press
.
Jenkinson
,
E. J.
&
Wilson
,
I. B.
(
1970
).
In vitro support system for the study of blastocyst differentiation in the mouse
.
Nature, Lond
.
228
,
776
778
.
Mintz
,
B.
(
1965
).
Experimental genetic mosaicism in the mouse
.
In Preimplantation Stages of Pregnancy, Ciba Foundation Symposium
(ed.
G. E. W.
Wolstenholme
and
M.
O’Connor
), pp.
194
216
.
London
:
J. and A. Churchill Ltd
.
Mulnard
,
J. G.
(
1961
).
Problèmes de structure et d’organisation morphogénétique de l’œuf des Mammifères
.
In Symposium on the Germ Cells and Earliest Stages of Development
, pp.
639
688
.
Milan
:
Fondazione A. Baselli, Instituto Lombardo
.
Mulnard
,
J. G.
(
1965
).
Studies of regulation of mouse ova in vitro
.
In Preimplantation Stages of Pregnancy, Ciba Foundation Symposium
(ed.
G. E. W.
Wolstenholme
and
M.
O’Connor
), pp.
138
144
.
London
:
J. and A. Churchill Ltd
.
Mulnard
,
J. G.
(
1967
).
Analyse microcinématographique du développement de l’œuf de souris du stage II au blastocyste
.
Archs Biol. (Liège)
78
,
107
138
.
Rodé
,
B.
,
Damjanov
,
I.
&
Skreb
,
N.
(
1968
).
Distribution of acid and alkaline phosphatase activity in early stages of rat embryos
.
Bull, scient. Cons. Acads RPF Yougosl. A
13
,
304
.
Seidel
,
F.
(
1960
).
Die Entwicklungsfahigkeiten isolierter Furchungszellen aus dem Ei des Kaninchens Oryctolagus cuniculus
.
Wilhelm Roux Arch. EntwMech. Org
.
152
,
43
130
.
Tarkowski
,
A. K.
(
1959
).
Experimental studies on regulation in the development of isolated blastomeres of mouse eggs
.
Acta theriol
.
3
,
191
267
.
Tarkowski
,
A. K.
&
Wroblewska
,
J.
(
1967
).
Development of blastomeres of mouse eggs isolated at the 4- and 8-cell stage
.
J. Embryol. exp. Morph
.
18
,
155
180
.
Whittingham
,
D. G.
(
1971
).
Culture of mouse ova
.
J. Reprod. Fert., Suppl
.
14
,
7
21
.