Interaction Between Inner Cell Mass and Trophectoderm of the Mouse Blastocyst

/. Embryol. exp. Morph. Vol. 51, pp. 109-120, 1979 109 Printed in Great Britain © Company of Biologists Limited 1979

Interaction between inner cell mass and trophectoderm of the mouse blastocyst

II. The fate of the polar trophectoderm

By A. J. COPP1 From the Department of Zoology, Oxford

SUMMARY Selective labelling of polar trophectoderm cells in early mouse blastocysts has allowed the fate of polar cells to be followed during in vitro and in vivo blastocyst development. Results show that there is cell movement from polar to mural regions as blastocysts grow. This indicates that trophectoderm cells directly opposite the inner cell mass are the oldest mural cells. However, after implantation polar cells invaginate into the blastocoelic cavity and contribute to the extra-embryonic ectoderm. It is suggested that the morphogenetic changes occurring in the mouse embryo at implantation result from the maintenance of a balance between (a) regional differences in rates of cellular proliferation, and (b) mechanical constraints on the direction in which growth can occur.

INTRODUCTION It was predicted in a previous paper, on the basis of cellular proliferation rates in the various blastocyst regions, that there is cell movement from polar to mural trophectoderm as blastocysts develop (Copp, 1978; see also Gardner & Papaioannou, 1975). This paper reports experiments in which the fate of polar trophectoderm cells was followed, both before and after implantation, using melanin granules as a marker. Trophectoderm cells readily phagocytose melanin granules, and retain them at least until implantation has occurred, apparently with no subsequent transfer of granules between cells (Gardner, 1975). Further- more, melanin granules are easily visualized in standard histological prepara- tions. Therefore, apart from the problem of dilution as labelled cells continue to divide, melanin granule labelling fulfills most of the requirements for a useful biological marker (Weston, 1967).

MATERIALS AND METHODS Blastocysts were flushed from the uteri of pregnant CFLP females between 3 days 14 h and 3 days 16 h after the estimated time of ovulation (see Copp, 1978), and were stored in PB1-medium (PB1, Whittingham & Wales, 1969) 1 Author's address: Department of Zoology, South Parks Road, Oxford 0X1 3PS, U.K. 8 EMB 51 110 A. J. COPP

100

Fig. 1. Partially polar-herniated blastocyst before labelling. Fig. 2. Completely polar-herniated blastocyst after labelling. plus 10 % heat-inactivated foetal calf serum (FCS). Any blastocysts which had collapsed were allowed to re-expand in PB1 + FCS at 37 °C. Each expanded blastocyst was held by suction and a slit was made in the zona pellucida over the polar trophectoderm using a pair of straight needles controlled by a Leitz micromanipulator assembly (Gardner, 1978). Operated blastocysts were cultured at 37 °C in PB1 + FCS under paraffin oil (Boots, U.K. Ltd) until blastocoelic expansion caused herniation of the inner cell mass (ICM) plus covering polar trophectoderm through the slit in the zona pellucida. When part of the polar region had emerged, herniation was arrested by removing blastocysts to PB1 + FCS at room temperature. A suspension of melanin granules was prepared by teasing apart retinae from pigmented mice of various strains in alpha modification of Eagle's medium supplemented with 30 fim adenosine, guanosine, cytidine and uridine, and 10 /an thymidine (a, Flow Labs) plus 10 % foetal calf serum. Partially herniated blastocysts (Fig. 1) were cultured in hanging drops of this suspension for 1 h at 37 °C, in an atmosphere of 5 % CO2 in air, after which most embryos showed complete polar-herniation (Fig. 2). Any blastocysts which had collapsed or showed mural herniation were rejected. All others were washed thoroughly in a + FCS to remove loosely adhering granules and their zonae were removed in acid Tyrode's solution, pH 2-5 (Nicolson, Yanagimachi & Yanagimachi, 1975). In order to study the fate of polar trophectoderm cells before implantation, labelled blastocysts were cultured in a + FCS, in bacteriological plastic dishes (Sterilin) under 5 % CO2 in air. After either 1 or 24 h of culture, blastocysts were fixed and prepared for analysis by serial reconstruction as described previously (Copp, 1978). Cell numbers were determined by counting nuclei in both the polar and ICM regions and in the mural subregions. Since cell outlines were not always easy to see, the number of labelled cells was estimated Fate of mouse polar trophectoderm 111 indirectly. The distribution of melanin granules was usually clumped, and it was considered that each 'clump' of melanin granules corresponded to a labelled cell. A 'clump' was defined as one or more granules occurring within a volume of cytoplasm which extended around each nucleus in any direction for a distance equal to the average length of a nucleus in that particular region. However, it was decided that 'clumps' should be confined within regional boundaries and therefore they were usually more extensive in certain directions than others (e.g. elongated, narrow 'clumps' in the polar region). In this way, 'clumps' were intended to resemble cells in both size and shape (Fig. 3). The 'clump':cell number ratio gave an estimated labelling index for each blastocyst region. In order to control for a possible time-dependent spreading of melanin granules in all directions throughout the blastocyst, mural-herniated blastocysts were also labelled and analysed in the same way (Figs. 4 and 5). It was not possible to control precisely whether the proximal or distal mural subregion herniated, so all mural-herniated blastocysts were pooled for analysis. The fate of polar trophectoderm cells after implantation was followed by transferring polar-labelled blastocysts to the uteri of pseudopregnant recipients 3 days 17 h after the estimated time of mating to sterile males. Labelled blastocysts were transferred to one uterine horn of each recipient female, and the contralateral horn received an equal number of unlabelled polar-herniated blastocysts. Recipients were killed 38 h later and uteri containing implantation sites were prepared for histological examination as described previously (Copp, 1978). Embryos had reached a very early egg-cylinder stage in which the polar trophectoderm was multilayered (Fig. 6). Extra-embryonic ectoderm was clearly forming in all embryos but none showed an accumulation of polar cells above the level of origin of the primary giant cells, indicating that the ectoplacental cone had not yet developed. Extra-embryonic ectodermal cells could not be reliably disinguished from more superficial polar cells, and so the whole polar region was analysed together. It could be distinguished from the embryonic ectoderm on the basis of: (a) its greater intensity of cytoplasmic and nuclear staining; (b) the shape and orientation of its nuclei, and (c) a space between the two regions. In addition, primary trophoblastic giant cells, visceral and parietal endoderm were recognizable. Serial sections were scored for the presence or absence of each tissue and for the presence of one or more labelled cells within them.

RESULTS P re-imp Ian tat ion developmen t Twenty-four polar-labelled and 19 mural-labelled blastocysts were analysed. The number of melanin granules per blastocyst ranged from 28 to 648, except for a single polar-labelled blastocyst which contained four granules. This was felt to be an abnormally low level of labelling and the blastocyst was excluded from the granule-distribution analysis. The average number of granules per 8-2 112 A. J. COPP

Fig. 3. Drawings of serial sections of polar-labelled blastocysts fixed after (A) 1 h, and (B) 24 h of culture. Nuclei are drawn in outline. Dashed outlines indicate nuclei which were only faintly visible. Dots represent melanin granules, rectangles enclose 'clumps'. Note that 'clumps' usually extend to adjacent sections. Fate of mouse polar trophectoderm 113

100 Him

Fig. 4. Partially mural-herniated blastocyst before labelling. Fig. 5. Completely mural-herniated blastocyst after labelling.

Polar region

Primary giant cells Embryonic ectoderm Visceral endoderm Parietal endoderm

Fig. 6. (A) Section and (B) drawing of an early egg-cylinder developed from a polar-labelled blastocyst transferred to the uterus of a pseudopregnant recipient. Arrows indicate melanin granules. embryo did not differ significantly between polar-labelled blastocysts cultured for 1 and 24 h and the same was true for mural-labelled blastocysts (Table 1). This indicates that there is no substantial loss of melanin granules from blastocysts during 24 h of culture. Since there appears to be no transfer of granules between cells (Gardner, 1975), any alteration in the relative numbers of granules in different blastocyst regions must therefore indicate a redistribution of granule- containing cells, or their progeny, as development proceeds. Table 2 shows the distribution of melanin granule ' clumps' in blastocysts after 1 and 24 h of culture. These results indicate that, for polar-labelled 114 A. J. COPP

Table 1. Numbers of melanin granules in polar- and mural-labelled blastocysts after 1 and 24 h of culture

Average number of Region Hours of Number of melanin granules labelled culture blastocysts per blastocyst P Polar 1, ..11. 112-..^v6, I, 1?4 > 0-()5 24 13 171-3 Mural x1 s9 ji.^.322- 3J \'1-25 > 01 24 10 216-6 * Student's ^-values from 'comparison of two means' tests. blastocysts: (1) there is an increase in the proportion of labelled cells in the proximal mural subregion after 24 h of development; (2) there is no comparable increase in distal and mural labelling; (3) there is no fall in the proportion of labelled polar cells, showing that all polar daughter cells inherit granules at least during the next 24 h of development. However, there is some dilution of label during this period since the average number of melanin granules per polar 'clump' fell from 8-6 after 1 h, to 5-9 after 24 h of culture; and (4) ICM labelling is negligible. In addition, the failure of polar labelling indices to reach 100 % after 1 h of culture may indicate that a ' clump' is equivalent, on average, to more than one labelled cell. This inaccuracy applies equally to all blastocyst regions and so does not affect the conclusion of this experiment. The low levels of labelling observed in 'unlabelled' trophectodermal regions after 1 h of culture presumably represents either the penetration of granules between embryo and zona pellucida at the time of labelling, or secondary attachment of loose granules soon afterwards. Material resembling melanin granules could also be deposited during the histological procedure. Table 2 shows for mural-labelled blastocysts: (1) there is no time-dependent spread of label from mural to polar regions; (2) as development proceeds there is a fall in the proportion of labelled cells in the proximal mural subregion; and (3) there is no change in the level of distal mural labelling after 24 h culture. However, an influx of cells into this subregion during blastocyst development cannot be ruled out since a number of blastocysts were labelled initially in the distal mural subregion (see Materials and Methods). The marginally significant difference between levels of labelling in the polar regions of blastocysts 1 and 24 h after mural-labelling can probably be attributed to a single blastocyst,. fixed after 1 h, which had a very high labelling index (66 %) in this region. These results demonstrate a movement of cells from polar to proximal mural trophectoderm during mouse blastocyst development. Furthermore, they are consistent with a movement of cells, during the same period, from proximal to distal mural subregions. Table 2. Distribution of melanin granule ''clumps'' in polar- and mural-labelled blastocysts after 1 and 24 h of culture

Mural A Proximal Distal Polar ICM A A > Region Hours of Number of No. 'clumps' No. 'clumps' No. 'clumps' No. 'clumps' fo labelled culture blastocysts No. cells % No. cells °/ No. cells /o No. cells /o

/o now. 117 3 Polar 1 11 ^ 12-24 10-69 75-49 100 294 159 155 299 49 124 7 24 12 13-57 81-58 1-51 I" i -« 36T 152 464 < 0001* > 005 >005 > 005 phectt 46 14 2 Mural 1 9 5411 10-37 0-85 85 135 236 5 8 vj 24 10 TP, 14 '54 41-71 3-42 1-90 447 2U 146 42l •*? < 0001 > 005 005 > P > 0025 > 005 3 * P-values from ^ 2 'comparison of two proportions' tests. 116 A. J. COPP

2 1 10 19 0 0 0

I 7777? E C E C E C E C E C Primary giant Polar Embryonic Visceral Parietal cells region ectoderm endoderm endoderm Fig. 7. The distribution of melanin granules in egg-cylinders developed from polar-labelled blastocysts (E) and from unlabelled polar-herniated blastocysts (C). Bars represent the average number of sections in which each tissue appeared. Hatched regions represent the proportion of these sections in which the tissue was labelled. Numbers above bars indicate the total number of melanin granules counted.

Post-implantation blastocyst development Five early egg-cylinders developed from polar-labelled blastocysts and five control embryos developed from unlabelled polar-herniated blastocysts were analysed. Figure 7 shows, for each tissue, the average number of sections in which the tissue appeared, and the proportion of these sections which contained one or more labelled cells. The actual number of labelled cells was very low, and for instance, represented only a small fraction of the actively proliferating polar trophectoderm. This illustrates the limitations of melanin granules as a marker for post-implantation trophoblastic development when rapid dilution of granules occurs. Nevertheless, the primary trophoblastic giant cells and multilayered polar trophectoderm were consistently labelled, and granule-containing cells occurred frequently deep within the newly forming extra-embryonic ectoderm (which was included in the 'polar trophectoderm' fraction) as well as more superficially in the mesometrial part of the polar region (Fig. 6). Control levels of labelling were very low, as expected, and probably represent material, resembling melanin granules, which occurs endogenously in the uterus or was deposited during the histological procedure. Fate of mouse polar trophectoderm 111 These results indicate that, after implantation, polar cells begin to invaginate into the blastocoelic cavity and contribute to the developing extra-embryonic ectoderm.

DISCUSSION The results presented in this paper demonstrate that, during mouse blastocyst development, there is a shift of cells from polar to proximal mural trophectoderm, without any significant contribution to the distal subregion of later blastocysts. In addition, cells appear to leave the proximal mural subregion, during the same period of development, and these presumably enter the distal mural subregion. This evidence provides support for the idea that trophectoderm cells directly opposite the 1CM are the oldest mural cells (Gardner & Papaioannou, 1975; Copp, 1978). If continued mural cell division depends directly on the length of time that cells have been out of contact with the ICM, this can explain why mural cells directly opposite the ICM are the first to cease division and begin giant cell transformation. The analysis of polar-labelled blastocysts transferred to recipient uteri shows that, after implantation, polar cells invaginate into the blastocoelic cavity, and contribute to the extra-embryonic ectoderm. A number of previous experiments have indicated a trophectodermal origin for this tissue (Gardner & Johnson, 1975; Rossant & Ofer, 1977) but the present experiment provides the first direct demonstration of this. The results presented here, and elsewhere (Gardner & Papaioannou, 1975; Copp, 1978) provide the basis for a model (Fig. 8) which attempts to explain the morphogenetic events leading to egg-cylinder formation. Before implantation, while the blastocyst is free floating in the uterine lumen, there is cell movement from polar to mural trophectoderm (Fig. 8a). At the time of attachment to the uterine epithelium, however, mechanical constraints are imposed on the blastocyst so that further polar to mural cell movement is prevented. Continued polar cell division leads to an accumulation of cells over the ICM. An increase in volume of the polar region is inevitable and will occur in the direction of least physical resistance. Careful histological analysis of this period of development indicates that, at the mesometrial end of the implantation crypt, the walls of the uterine lumen are tightly apposed (see Fig. 3 in Reinius, 1965) and so mesometrial growth would require the walls to be forced apart. On the other hand, anti-mesometrially the blastocoel presents no such mechanical barrier to growth. It is suggested, therefore, that growth occurs initially in an anti-mesometrial direction (Fig. 86), and this idea is supported by the observation that very early egg-cylinders have small ectoplacental cones but well developed extra-embryonic ectodermal regions (see Fig. 6, and also figure 12.7 in Snell & Stevens, 1966). The lateral walls of the uterine luminal crypt, with covering primary giant cells and parietal endoderm, are closely apposed to the sides of the growing egg- cylinder (see figure 4 in Reinius, 1965). Gaps between visceral and parietal 118 A. J. COPP

Fig. 8. Model to explain egg-cylinder morphogenesis in the mouse. See text for explanation. endoderm normally seen in wax sections are probably artifactual. Consequently, growth of the egg-cylinder will be predominantly anti-mesometrial and not lateral. Once the anti-mesometrial tip of the egg-cylinder comes into contact with the relatively slowly expanding Reichert's membrane, its further growth in this direction is restricted. The path of least resistance to polar growth, now, is in a mesometrial direction since the ectoplacental cone can develop by forcing apart the apposed walls of the uterine lumen (Fig. 8 c). This series of morphogenetic events is completed within about 12 h of the onset of implantation, with pro- amniotic cavity formation occurring mainly during the third phase (Fig. 8 c) of the sequence. This model visualises in vivo egg-cylinder morphogenesis as resulting inevitably from the maintenance of a balance between regional differences in cellular proliferation rates and mechanical constraints on the direction in which growth may occur. However, it has been reported that approximately normal egg- cylinder development may occur, in the absence of uterine mechanical constraints, after blastocyst outgrowth in vitro (Hsu, Baskar, Stevens & Rash, 1974; Pienkowski, Solter & Koprowski, 1974; Wiley & Pedersen, 1977). Mechanical constraints may, nevertheless, operate in such systems if there is an increase in the strength of adhesion between trophoblast giant cells and their substratum soon after the onset of blastocyst outgrowth. This could result in a redirection of polar cell movement so that polar cells accumulate beneath the ICM and egg-cylinder formation results. Studies of in vitro egg-cylinder morphogenesis have indicated that polar cells cease to move into the trophoblast giant cell outgrowth shortly before egg-cylinder formation, as predicted by this hypothesis (A. J. Copp, in preparation). Fate of mouse polar trophectoderm 119 In other developing systems it has been suggested that morphogenesis may involve differential rates of cellular proliferation and mechanical constraints on the directions of growth. For instance, developing chicken lens epithelial cells initially elongate and subsequently the placode epithelium invaginates to form the lens. It has been proposed that these events are the results of continued cell proliferation within the placode and a restriction on the lateral spreading of the increasing placode cell population due to firm contact between the lens epithelium and the underlying optic vesicle (Zwaan & Hendrix, 1973). Similar explanations have been suggested for the evaginating thyroid gland, in which lateral spreading of thyroid cells is prevented by local mechanical constraints (Hilfer, 1973), and for the developing pancreas where a limitation on increase in surface area of the rudiment occurs despite its continued cell number increase (Pictet, Clark, Williams & Rutter, 1972). In addition, regional differences in cellular proliferation rates have been noted in the branching salivary rudiment (Bernfield, Banerjee & Cohn, 1972) and elongation of tubular glands in the chick oviduct is blocked by inhibitors of cell division (Wrenn, 1971). Finally, it should be noted that even in systems where regional differences in mitotic activity do not occur (e.g. the primordial lung, Wessels, 1970) localized mechanical constraints on cell movements accompanied by continued general cell division could nevertheless lead to the morphogenetic events observed.

I would like to thank Professor R. L. Gardner, Dr V. E. Papaioannou and Miss R. Beddington for valuable discussion. This work was supported by a Christopher Welch Scholarship and by the Medical Research Council.

REFERENCES BERNFIELD, M. R., BANERJEE, S. D. & COHN, R. H. (1972). Dependence of salivary epithelial morphology and branching morphogenesis upon acid mucopolysaccharide-protein (proteoglycan) at the epithelial surface. /. Cell Biol. 52, 674-689. COPP, A. J. (1978). Interaction between inner cell mass and trophectoderm of the mouse blastocyst. I. A study of cellular proliferation. /. Embryol. exp. Morph. 48, 109-125. GARDNER, R. L. (1975). Analysis of determination and differentiation in the early mammalian embryo using intra- and inter-specific chimaeras. In The Developmental Biology of Reproduction: 33rd Symposium of the Society for Developmental Biology (ed. C. L. Markert), pp. 207-238. New York: Academic Press. GARDNER, R. L. (1978). Production of; chimaeras by injecting cells or tissues into the blastocyst. In Methods in Mammalian Reproduction, (ed. J. C. Daniel, Jr.) pp. 137-165. New York: Academic Press. GARDNER, R. L. & JOHNSON, M. H. (1975). Investigation of cellular interaction and deploy- ment in the early mammalian embryo using interspecific chimaeras between the rat and mouse. In Cell Patterning {Ciba Fdn Symp.), pp. 183-200. Amsterdam: Associated Scientific Publishers. GARDNER, R. L. & PAPAIOANNOU, V. E. (1975). Differentiation in the trophectoderm and inner cell mass. In The Early Development of Mammals: 2nd Symposium of the British Society for Developmental Biology (ed. M.Balis & A. E. Wild), pp. 107-132. London: Cambridge University Press. HILFER, S. R. (1973). Extracellular and intracellular correlates or organ initiation in the embryonic chick thyroid. Amer. Zool. 13, 1023-1038. 120 A. J. COPP Hsu, Y.-C, BASKAR, J., STEVENS, L. C. & RASH, J. E. (1974). Development in vitro of mouse embryos from the two-cell stage to the early somite stage. /. Embryo/, exp. Morph. 31, 235-245. NICOLSON, G. L., YANAGIMACHI, R. & YANAGIMACHT, H. (1975). Ultrastructural localisation of lectin-binding sites on the zonae pellucidae and plasma membranes of mammalian eggs. J. Cell Biol. 66, 263-274. PICTET, R., CLARK, W. R., WILLIAMS, R. H. & RUTTER, W. J. (1972). An ultrastructural analysis of the developing embryonic pancreas. Devi Biol. 29, 436-467. PJENKOWSKI, M., SOLTER, D. & KOPROWSKI, H. (1974). Early mouse embryos: growth and differentiation in vitro. Expl Cell Res. 85, 424-428. REINIUS, S. (1965). Morphology of the mouse embryo, from the time of implantation to mesoderm formation. Z. Zellforsch. mikrosk. Anat. 68, 711-723. ROSSANT, J. & OFER, L. (1977). Properties of extra-embryonic ectoderm isolated from post- implantation mouse embryos. /. Embryol. exp. Morph. 39, 183-194. SNELL, G. D. & STEVENS, L. C. (1966). Early embryology. In Biology of the Laboratory Mouse (ed. E. L. Green), pp. 205-245. New York: McGraw-Hill. WESSELS, N. K. (1970). Mammalian lung development: interactions in formation and morphogenesis of tracheal buds. /. exp. Zool. 175, 455-466. WESTON, J. A. (1967). Cell marking. In Methods in Developmental Biology (ed. F. H. Wilt & N. K. Wessels), pp. 723-736. New York: Crowell. WHITTINGHAM, D. G. & WALES, R. G. (1969). Storage of two-cell mouse embryos in vitro. Aust. J. biol. Sci. 22, 1065-1068. WILEY, L. M. & PEDERSEN, R. A. (1977). Morphology of mouse egg-cylinder development in vitro: a light and electron microscopic study. /. exp. Zool. 200, 389-402. WRENN, J. T. (1971). An analysis of tubular gland morphogenesis in chick oviduct. Devi Biol. 26, 400-415. ZWANN, J. & HENDRIX, R. W. (1973). Changes in cell and organ shape during early development of the ocular lens. Amer. Zool. 13, 1039-1049.

{Received 4 September 1978, revised 9 December 1978)