Notes on the Development of the Germ-Layers in Diprotodont Marsupials

Notes on the Development of the Germ-layers in Diprotodont Marsupials. By T. Kerr, M.A. Assistant in the Zoology Department, Queen's University of Belfast.

With 5 Text-figures.

THE material for this investigation was collected in Tasmania by Professor T. Thomson Plynn from animals taken in the wild during 1920 to 1929, and forms part of the marsupial and mono- treme material obtained by him by the aid of grants allowed by the Ealston Trustees and by the Grants Committee of the Royal Society. The author would like to express now his gratitude both for the opportunity of examining it and for the help and encouragement he has received during the course of the work. The material itself consists of nine series of sections of blastocysts of Bettongia cuniculus cut by Professor Flynn him- self in Tasmania, and of fourteen blastocysts of Bettongia cuniculus and four of Potorous tridactylus which have been cut here. A series of early stages of these rather scarce little Marsupials is extremely hard to obtain, for only a single egg is produced at a time and it is almost impossible to breed the animals in captivity (Plynn, 1930). The blastocysts were fixed in Bouin or in Hill's fluid (picro-nitric-aceto-osmic mix- ture), and those not cut at the time were preserved in alcohol; the whole series covers a range of development from an early unilaminar blastocyst to a fully formed primitive streak. The process of section cutting is apt to be exceptionally troublesome owing to the small size of the blastocysts and the thickness of the shell-membrane, but the following method proved very satisfactory. The blastocysts Avere impregnated with thin celloidin, fixed to small pieces of liver by means of the celloidin, and embedded in paraffin; the blocks were cut by the excellent large precision microtome of Bausch and Lomb, kindly presented to the Department by the Rockefeller Trustees. 306 T. KERB All sections were cut at 5/x, and stained with iron haematoxylin and eosin; as it was not possible to orientate the smaller blastocysts, doubtful points in the sections had to be cleared up by reconstructions. The development of the germ-layers in Diprotodont Marsu- pials has not been worked out, so it was decided to treat this part of the material separately from that dealing with the primitive streak. There are various accounts of the early development of Polyprotodonts. Selenka's (1887) work on Didelphys virginiana was based on a very limited amount of material, some of which was certainly abnormal. An abnormal eight- celled egg, consisting of two irregular rings of four cells each, misled him into considering the four somewhat smaller cells ectoderm and the larger endoderm. In his forty-two- and sixty- eight-celled blastocysts, or 'gastrulae' as he called this stage, there is a definite polarity, which he took to be the continuing differentiation of his smaller-celled 'ectoderm' and larger-celled 'endoderm'. Each of these blastocysts has included in if a large ' L'rentoderm' cell, and Selenka apparently believed that the endoderm was formed by the gradual inclusion into the cavity of all the cells of his ' endoderm' pole from the region of a 'blastopore', their places being taken by encroaching 'ectoderm'. Possibly these 'Urentoderm' cells are similar to those described by Hartman (1919) as being included from about the fifty-celled stage onwards, and claimed by him also as endoderm, but Selenka's general views are certainly incorrect. In the above paper Selenka goes on to describe some primitive streak stages from Didelphys, but he did not have any blastocysts showing the origin of the mesoderm. In 1892 he published a description of the development of Hypsipry- mnus cuniculus ( = Bettongia cuniculus), but in the earliest stage with which he deals, namely a blastoeyst two days old and 2 mm. in diameter, the primitive streak is already well formed. Hill (1910) gave a very full account of early development in some Polyprotodonts, describing in detail the formation of endoderm in Dasyurus viverrinus. In this animal smaller, darker-staining 'endoderm mother cells' develop here and there throughout the formative area of the unilaminar GERM-LAYERS IN DIPROTODONTS 307 blastocyst, and when the blastocyst is about 4-5 mm. in diameter these give rise to the true internal endoderm. .This may occur in two ways: in the first, the endoderm mother cell prolongs itself along the inner side of the embryonic ectoderm, then migrates bodily out of place and comes to lie on the inner surface of the ectoderm; in the second, the prolongation divides to form an endoderm cell before the mother cell migrates. In either case the cells are capable of amoeboid activity, by which their migration is probably achieved, and of developing pseudopodia, which by secondary anastomoses with each other form the beginnings of a continuous, though at first fenestrated, endoderm layer. Eepeated cell division closes up the layer and extends it completely round the inside of the blastocyst. From shorter series of specimens Hill considers the method of endoderm formation in Macropus and Perameles to be similar; it is interesting to note, however, that in these the endoderm first appears in 0-85 mm. and just over 1 mm. diameter blastocysts respectively. Minot (1911) in a brief description of a late bilaminar blastocyst of the opossum added nothing new. Hartman (1916) studied the egg of Didelphys virginiana up to the fully formed bilaminar blastocyst; in this paper he describes endoderm appearing in blastocysts of 0-6 mm. or less, containing somewhat over 100 cells, and his conclusions regard- ing its origin are substantially the same as those of Hill. He mentions the occurrence in earlier stages of large cells included in the cavity of the blastocyst, as does Hill, but is inclined to agree with the latter that they are of no morphological importance. In a subsequent paper (1919), however, for which he had available a much more complete series of stages, he modifies his conclusions very considerably. He believes, in fact, that the included cells are the first appearance of the endoderm, which therefore appears at the fifty- or sixty-celled stage (about 0-15 mm. in diameter). He describes this as occurring by the rounding off of certain cells in a part of the wall representing the future embryonic area, and by their consequent detachment from the wall and their division to form endoderm. Hence the phenomena of endoderm formation described by Hill as the complete process in Dasyurus only represent the last stage 308 T. KERR in Didelphys. Hill (1918) described various stages of Didelphys aurita, but none dealing with layer formation. The following is a list of the various stages dealt with in this paper. When the blastocysts were in their natural spherical condition the diameter is given, when they were at all out of shape maximum and minimum measurements are given. Bettongia stages: I, 0-180 mm.; II, 0-210 mm.; Ill, 0-218 mm.; IV, 0-225 mm.; V, 0-266 mm.; VI, 0-275 mm.; VII, 0-308 mm.; VIII, 0-364 mm.; IX, 0-370 mm.; X, 0-380 mm. ;

TEXT-FIG. 1. ALB., albumen; COAO., coagulum; SH.M., shell-membrane; Y.B., yolk-body. X 200. XI, 0-392 mm.; XII, 0-704 by 0-432 mm.; XIII, 0-728 by 0-304 mm.; XIV, 0-761 by 0-680 mm.; XV, 0-938 mm.; XVI, 1-324 by 0-993 mm.; XVII, 1-600 mm. Potorous stages: I, 0-208 mm.; II, 0-458 by 0-396 mm.; III, 2-622 by 2-001 mm. The earliest blastocyst in the Bettongia series is a young unilaminar (0-180 mm.) in which the nuclei are large and ill defined, and both albumen coat and central mass are still well developed. The central mass can be differentiated into coagulum and yolk-body; but, as often happens, the zona pellucida has broken down under fixation (Text-fig. 1). The next three stages are simple unilaminar blastocysts, showing an increase in the number of nuclei with a decrease in their size; in all the albumen coat and central mass are distinguishable, though increasingly attenuated. It is not possible to make out any definite polarity in these stages, although it is shortly to appear. GERM-LAYERS IN DIPROTODONTS 309 The examination of unsectioned blastocysts is unfortunately not possible, owing to their small size in Bettongia and their thick shells; among other things this makes it difficult to be sure of the shape of the endoderm cells when they arise, and impossible to say whether they have pseudopodia or not. In stage V (0-266 mm.) the albumen has almost disappeared,

TEXT-FIG. 2. ALB., albumen; FORM. AR., formative area; PREC, precipitate; Z.P., zona pelluoida. X 250. but the zona pellucida can be clearly seen. The formative area has made its appearance as a concentration of cells forming a cap at one pole, the cap covering rather less than a quarter of the inside of the shell membrane (Text-fig. 2). The two areas remain distinct throughout succeeding stages, an increasing difference in thickness developing, and the endoderm mother cells when they appear are always within the formative portion. In this stage, which contains about eighty cells, and in many later ones, are present the inclusions noticed in early blastocysts by Hill and Hartman, and deemed by the former to be of no morphological importance inDasyurus. It was of the greatest interest, therefore, to discover whether these inclusions, or some of 310 T. KERR them, could possibly be taken as the first endoderm mother cells, whether, in other words, endoderm formation in Bettongia resembles that of Dasyurus or that of Didelphys. These preparations all indicate that it closely resembles that of Dasyurus. The cells described by Hartman as being normally included are not present, nor any that might reasonably be considered to be their descendants. Very rarely is there included a cell which might be a displaced blastomere, but typically the inclusions are irregular, both in size and in place of origin;

TEXT-FIG. 3. END.C, endoderm cell; END.M.C, endoderm mother cell; I.C., included cell; X, edge of formative area, x 350. although they are more common in the formative region, per- haps owing to a greater rate of metabolic activity there, they may be present anywhere on the blastocyst wall; the develop- mental age is no guide to the number to be expected, and they may even be entirely absent. They occur in Bettongia mainly as spherical masses attached to the inner side of normal cells (Text-fig. 3), and while sometimes apparently quite healthy are sometimes certainly degenerating; their nuclei are usually absent or represented merely by a few particles of chromatin. It seems probable that the substance of these cells is added to the nutritive fluid Avhich fills the interior of the blastocyst, and that they have no connexion here with endoderm formation. In stages X and XI (0-380 mm. and 0-392 mm.), containing GERM-LAYERS IN DIPROTODONTS 811 about 120 and 150 cells respectively, can be seen distributed throughout the formative area a number of compact cells with more granular cytoplasm, the endoderm mother cells. Their shape and the tendency of their cytoplasm to stain more deeply than that of neighbouring cells make them fairly easy to pick out in sections, and their arrangement throughout the formative area is seen to be irregular. Some are simply forming part of the blastocyst wall, others are elongating inwards on to the inner surface of the ectoderm. In stage XI a few have become entirely separated from the shell-membrane, and lie in shallow depressions in the ectoderm (Text-fig. 3). There are considerable resemblances, therefore, to the corresponding stage in D a s - y u r u s. It is difficult to be certain of the finer details of this inward migration of the endoderm, but the mechanism appears to be as follows. The originally rather rounded mother endoderm cells tend to become elongated in one or more directions along the inner surface of adjoining ectoderm cells; at the same time the ectoderm cells encroach upon the endoderm cells from the sides, driving a thin layer of protoplasm between them and the shell-membrane. This layer thickens until the ectoderm has become homogeneous, with the endoderm cells lying on its inner surface; in this position the latter flatten themselves out into very attenuated cells, which join up with each other to form a fenestrate layer under the formative region. Active cell divisions close up the layer and continue it round the inside of the blastocyst until the whole becomes bilaminar. It is possible that pseudopodia are present, as in Dasyurus, though not appreciable in sections; but if so their primary function seems likely to be to establish the first connexions throughout the developing endoderm. Hill suggests that part of the endoderm may be formed by proliferation of the endoderm mother cells in situ, i.e. by inwardly prolonged halves dividing off before the mother cell vacates its position in the wall. This appeared likely, so a case was looked for in which such a division was occurring or appeared to have occurred, but without success. Slightly later stages, unfortunately, are lacking, and in stage XII (0-704 by 0-432 mm.) the endodermal layer is complete. NO. 306 Y 812 T. KERR The ectoderm is divided quite sharply into formative and non- formative regions, the former occupying between one-quarter and one-third of the whole; the endoderm covers its inner surface as a thin layer, very attenuated between nuclei, and as yet undifferentiated. In stage XIV (0-761 by 0-680 mm.) the formative ectoderm nuclei are sperical or nearly so, measuring from about 8-5/u, to about 11-5/J. by 7/x; the non-formative are flattened and plate-like in form, measuring in sections about 9 p. by 4:5ix. The endoderm nuclei are similar in size and shape

TEXT-FIG. 4. EMB.ECT., embryonic ectoderm; END., endoderm; X, edge of formative area, x 250.

to those of the non-formative ectoderm. As growth proceeds a thickening takes place throughout all the ectoderm, but as this is relatively greater in the formative region the distinction between the two areas becomes more marked. In the endoderm a similar, though much less distinct, differentiation occurs; it consists of a rather irregular thickening under the formative area, and associated with this is a tendency for the nuclei to become more spherical (Text-fig. 4). In stage XVI (1-324 by 0-993 mm.) mesoderm has just begun to appear. It originates as a small, elongated oval area of proliferation in the ectoderm, orientated from near one edge of the formative area towards the centre. The proliferation results at first simply in a piling up of nuclei in this region, their cytoplasm continuous with that of the ectoderm to the outside and in close contact with that of the endoderm to the inside GERM-LAYERS IN DIPROTODONTS 313 (Text-fig. 5). Unlike the endoderm mother cells there is no histo- logical difference appreciable between ectoderm and first mesoderm, and there are no appearances suggesting emigration of the latter as a series of separate cells. Mesoderm would appear to arise merely as a more active proliferation of the ectoderm in a definitely linear area corresponding to the beginning of the future primitive streak. There is nothing to mark this proliferation as distinct from the ordinary division of the ectoderm cells, it is simply this division intensified over a restricted field. In stage

TEXT-FIG. 5. EMB.ECT., embryonic ectoderm; EMB.END., embryonic endoderm; M.Z., zone of mesoderm formation, x 250.

XVII (1-600 mm.) the area of proliferation of mesoderm has increased in size, but is still continuous with the ectoderm and has the same general appearance as before; the formative area also is still circular. These and the following stages will be considered later in relation to the primitive streak. The shell thicknesses were measured, the smallest measure- ment obtainable being taken as the most accurate in each case. As noticed by Hartman (1919) there is clearly great individual variation, and all that can be said is that the membrane begins to thin rapidly with the formation of the complete didermic blastocyst. The following table gives a series of sizes throughout the range: / blastacysl! Shell thickness Size of blastocyst Shell thickness in mm. in microns in mm. in microns 0-180 8-8 0-704 by 0-432 31 0-266 4-8 0-938 3-2 0-275 3-6 1-324 by 0-993 2-2 0-308 5-7 1-600 1-9 0-364 5-4 314 T. KERR The Potorous blastocysts are somewhat smaller in relation to their stage of development than those of Bettongia. Stage I (0-208 mm.) is a unilaminar blastocyst in which no formative area can be distinguished. Stage II (0-453 by 0-396 mm.) has a formative area with endoderm complete underneath it; in general appearance it resembles the corresponding stage of Didelphys much more closely than that of Bettongia. Stage III (2-622 by 2-001 mm.) has mesoderm well developed. The shell thicknesses decrease with age, being 7-8, 4-3, and 2-2 microns respectively. Otherwise no new features are to be seen in these stages. SUMMARY. Some early blastocysts of Bettongia cuniculus are described, ranging from the young unilaminar stage to that showing the first appearance of the mesoderm. The formation of formative and non-formative areas is described, and the origin of endoderm and mesoderm. Endoderm arises as a number of endoderm mother cells distributed throughout the formative area which move out of their positions in the blastocyst wall on to the inner surface of the formative ectoderm; later they join up to form a continuous layer. The process, therefore, is very similar to that found in Dasyurus viver- r i n u s by Hill; the formation of endoderm from large included cells as found by Hartman in Didelphys virginiana does not occur here. Mesoderm arises as an oval area of active proliferation in the formative ectoderm, orientated from near one edge of the still circular formative area towards the centre.

BIBLIOGRAPHY. Selenka, E. (1887).—'Studien iiber Entwickelungsgeschichte der Tiere', Heft IV, "Das Opossum (Didelphys virginiana)". Wiesbaden. (1892).—Ibid., Heft V, "Beutelfuchs und Kanguruhratte (Phalan- gista et Hypsiprymnus)". Wiesbaden. Hill, J. P. (1910).—"The Early Development of the Marsupalia, with special reference to the Native Cat (Dasyurus viverrinus)", 'Quart. Journ. Mior. Sci.', vol. 56. Minot, C. R. (1911).—"Note on the Blastodermic Vesicle of the Opossum", 'Anat. Rec.', vol. 5. GERM-LAYERS IN DIPROTODONTS 315

Hartman, C. G. (1916).—"Studies in the Development of the Opossum, Didelphys virginiana L., Parts I and II", ' Journ. of Morph.', vol. 27. Hill, J. P. (1918).—"The Early Development of Didelphys aurita", 'Quart. Journ. Mior. Sci.', vol. 63. Hartman, C. G. (1919).—"Studies in the Development of the Opossum, Didelphys virginiana L., Parts III and IV", 'Journ. of Morph.', vol. 32. Flynn, T. Thomson (1930).—"The Uterine Cycle of Pregnancy and Pseudo- pregnancy as it is in the Diprotodont Marsupial Bettongia cuniculus", 'Proc. Linn. Soc. N.S.W.', vol. 55.