The ultrastructure of germ cells in foetal and neonatal male rats

The Gonads of the rat undergo sex differentiation on the 14th day post coitum (p.c.), the testis becoming clearly distinguishable by the presence of an incipient tunica albuginea. The male germ cells become incorporated into medullary cords (the precursors of seminiferous tubules). In contrast, the germ cells in ovaries are scattered in cortical nests.

Recent quantitative studies have shown that at 14 · 5 days p.c., the number of germ cells is somewhat greater in the male than the female (Beaumont & Mandl, 1963; cf. Beaumont & Mandl, 1962). In both sexes mitotic activity ceases at about 18 · 5 days p.c. Thereafter, the male germ cells remain at prolonged interphase; a proportion of them show histological changes frequently associated with degeneration. Quantitative estimates, on the other hand, indicate that none are eliminated from the testis. In the coeval female, the germ cells enter the prophase of meiosis, whereafter no further mitotic divisions are possible. Due to several ‘waves’ of degeneration, the total number of germ cells falls rapidly (Beaumont & Mandl, 1962, 1963; Franchi & Mandl, 1962).

At 4 to 5 days after birth, the germ cells of the male divide to give rise to the first generation of spermatogonia type-A. A cytological study suggests that the primordial germ cells change their appearance shortly before renewal of mitotic activity (‘transitional’ cells; see Beaumont & Mandl, 1963). Shortly before the reappearance of mitoses, a number of germ cells undergo lysis : the cytoplasm seemingly ‘streams out’ following the breakdown of the cell membrane, and the nucleus becomes highly distorted. The degenerating gonocytes are frequently situated in the centre of the seminiferous cords, while the ‘transitional’ cells tend to be more peripheral in distribution. The latter observation (see also Sapsford, 1962a; Huckins, 1963) suggests that proximity to the basement membrane is somehow associated with the survival of the germ cell.

Primordial germ cells in the male rat have been shown to be remarkably sensitive to ionizing radiations. As judged by the percentage of normal, ‘regenerating’ and sterile tubules 25 days after birth, sensitivity rises between 13 · 5 and 17 · 5 days p.c. and remains at a high level throughout the period of interphase; with the renewal of mitotic activity, it declines (Beaumont, 1960, 1962; Starkie, 1961; Hughes, 1962). These observations suggest that the relationship between mitotic activity and susceptibility to ionizing radiation is the reverse of that observed in most other cells (see Mandl, 1964).

The present study was undertaken in order to answer the following questions :

1. Do male germ cells undergo ultrastructural changes during the prolonged radiosensitive phase of mitotic inactivity? If so, can these be correlated with the apparent ‘degenerative’ changes observed in histological preparations ?

2. Does the ultrastructure of mitotically active germ cells, shortly after sex differentiation, differ between the two sexes ?

3. Is the renewal of mitotic activity, shortly after birth, associated with marked changes in the ultrastructure of germ cells ?

4. By what process do male germ cells undergo degeneration, and how are they removed from the testis (cf. female: Franchi & Mandl, 1962)?

Adult female rats of the Birmingham colony were housed with males, and the time of mating determined by means of daily vaginal smears. The does were killed at intervals ranging from 14 · 5 to 22 · 5 days p.c. Some of the testes were derived from foetuses whose litter-mates were used for previous studies of foetal gonads (Beaumont & Mandl, 1962, 1963; Franchi & Mandl, 1962). Neonatal male rats were killed at intervals between 1 and 7 days post partum (p.pf Some of these were again litter-mates of specimens used for cytological and quantitative studies (Beaumont & Mandl, 1963). A total of seventy testes was examined with the electron microscope. .

The pregnant females were killed by means of chloroform vapour. The foetuses were rapidly dissected out and decapitated. Neonatal rats were killed either by decapitation or by a sharp blow on the head.

The gonads (together with their adnexa and fragments of mesonephroi in the youngest specimens) were rapidly dissected out and placed at once in ice-cold fixative (1 per cent, osmic acid in Michaelis’ veronal/acetate buffer at pH 7 · 6, containing a balanced salt solution with Na, Ca and K ions). The fixative used was recommended by Dr J. D. Robertson in a personal communication. It was prepared from the following solutions. Solution 1: Michaelis’s Buffer solution, consisting of 19 · 428 g. sodium acetate; 29 · 428 g. sodium veronal dissolved in 500 ml. CO₂-free distilled water. Solution 2: N/5 hydrochloric acid. Solution 3: 100 ml. of solution containing 96-8 g. sodium chloride per litre; 2 · 17 ml. of solution containing 3 · 42 g. calcium chloride per litre; 1 · 73 ml. of solution containing 2 · 85 g. potassium chloride per litre. Final fixative; 5 ml. sodium acetate buffer (Solution 1); 5 ml. N/5 hydrochloric acid (Solution 2); 13 ml. distilled water; 2 ml. salt solution (Solution 3). Then add equal parts (i.e. 25 ml.) 2 per cent. OsO₄ in distilled water.

The gonads were fixed for 2 – 3 hr.; they were then rinsed in buffer and dehydrated in an ethanol series. At the 70 per cent, alcohol stage of dehydration, non-gonadal tissues were removed as far as possible by careful trimming under a dissecting microscope. Some mesonephric tissue remained, however, in specimens aged 16 · 5 days p.c. or less; these were trimmed further after embedding (see below). The tissues were further dehydrated (up to 100 per cent, alcohol); they were then stained in 0 · 5 per cent, phosphotungstic acid in 100 per cent, alcohol before being finally dehydrated in absolute acetone. The embedding medium employed was Vestopal W (see Ryter & Kellenberger, 1958). The embedded tissues were sectioned on a Porter-Blum microtome. Silver, pale gold and occasionally thicker sections were selected for examination under a Siemens Elmiskop I electron microscope.

Specimens aged 14 · 5 to 16 · 5 days p.c. could not be sexed accurately at autopsy. In order to determine the sex of the gonads, and to remove non-gonadal tissues, thick sections (1 – 2 μ)were cut after embedding in Vestopal, mounted on microscope slides, and stained with 0 · 5 per cent, aqueous toluidine blue (see Franchi, 1963). The sections were examined under a light microscope, using both normal and phase contrast illumination. The embedded blocks were trimmed progressively by means of a razor blade, cutting away as much non-gonadal tissue as possible. The gonads were sexed by the usual histological criteria (presence of tunica albuginea and formation of medullary cords in the male; random distribution of germinal and somatic elements in the female).

At 14 · 5 days p.c. the majority of male germ cells are situated within the incipient cords in the inner region of the gonad; a few have been observed immediately beneath the germinal epithelium. Somatic cells within the cords (‘supporting’ cells) show no specific orientation with respect to the germ cells.

In the female, however, oögonia are scattered in rather ill-defined nests, frequently associated with somatic cells; the latter may be closely applied to the germ cells.

The germ cells of both sexes are characterized by their large size, and by rounded nuclei with sharply defined membranes. In the female, some cells are irregular in shape, and extensions of the cytoplasm may pass between adjacent somatic cells; in general, however, the germ cells are of simple ovoid or spherical form.

At 15 · 5 days p.c., male germ cells have a somewhat more angular outline than before, while the oögonia remain rounded. Practically all the gonocytes at this and subsequent stages are restricted to the developing medullary cords. Oögonia, tightly packed in nests, are again frequently associated with crescent-shaped somatic cells.

As the seminiferous cords develop, their inner regions become filled with germ cells of rounded or angular outline. The majority of ‘supporting’ cells appear to be attached to the basement membrane, but extending long irregular processes of cytoplasm between the gonocytes (Plate 1, Fig. 1). In contrast, oögonia are frequently seen in the immediate vicinity of the germinal epithelium. Nests of somatic and germinal cells become more clearly distinguishable. The intermingling of the two cell-types is more marked in the female than in the male (Plate 1, Fig. 2). For an account of the further development of female germ cells, see Franchi & Mandl (1962).

From the age of 16 · 5 days p.c. onwards, the general distribution of gonocytes remains more or less constant until some 4 days p.p. The most prominent change within the germ cells consists in the accumulation of mitochondria at one pole of the cell (mitochondrial ‘polarization’; see below). The cells remain closely packed, and cytoplasmic bridges between adjacent gonocytes are common; the conjoined cell membranes at these points may or may not have osmiophilic and/or phosphotungstic acidophilic material associated with them (Plate 2, Fig. 3). Similar observations have been made on the mammalian ovary (Franchi & Mandl, 1962) and the avian testis (Nagano, 1961).

By 19 · 5 days p.c., mitochondria again become distributed more evenly throughout the cytoplasm of the gonocyte. Extensions of the ‘supporting’ cells have infiltrated between the germ cells which thus tend to be ‘pushed apart’. Possibly as a result of the growth of somatic cells, the shape of the gonocytes may become somewhat more irregular.

The first clear-cut signs of degenerative changes were observed at 2 days p.p.; some components of individual gonocytes become swollen and the cell membrane discontinuous (see below).

At 4 days p.p., many germ cells are characterized by cytoplasmic processes, relatively free from organelles, which extend towards the basement membrane of the seminiferous cords (Plate 2, Fig. 4). The ‘supporting’ cells, whose nucleus is almost invariably in the basal portion of the cell, still possess long cytoplasmic processes which, together with the germ cells, fill the inner region of the semini-ferous cords. The latter are therefore still solid.

Between 5 and 7 days p.p., the proportion of cells which have established contact with the basement membrane increases. A few gonocytes remain seemingly unattached to the basement membrane; an unknown proportion of these may, however, extend cytoplasmic processes towards it in a different plane. Both ‘attached’ and ‘unattached’ gonocytes fall into two categories: (1) large cells, essentially similar to gonocytes seen at earlier stages, but for the irregularity of the shape of their nucleus (Plate 3, Fig. 5); and (2) smaller cells, characterized by a reduced cytoplasmic/nuclear ratio (Plate 3, Fig. 6). A comparison with histological preparations suggests that the former correspond to the ‘transitional’ cells, and the latter to spermatogonia type-A (see Beaumont & Mandl, 1963). In addition, a number of ‘unattached’ gonocytes retain their earlier configuration; with increasing age, these become rarer and some show signs of degeneration (see p. 299).

Mitotic figures, whose incidence rises at this time, have been observed more frequently in ‘attached’ than in ‘unattached’ cells. The plane of division in relation to the basement membrane varies; in some cases, one daughter cell is initially separated from the basement membrane, while the other lies with its long axis parallel to the basement membrane. Mitoses in centrally-placed germ cells are frequently abnormal, the cell contents spilling out of gaps in the membrane (see below).

Shortly after sex differentiation, the nuclei of both gonocytes and oögonia are typically large and rounded; occasionally they may be pear- or kidney-shaped. The nuclear envelope is sharply defined, particularly the outer membrane. A narrow layer of electron dense material is closely applied to the inner membrane. Obliquely cut sections show well defined annular nuclear pores.

The nuclear sap of gonocytes contains a heterogeneous assortment of granules which are frequently absent from oögonia. From 17-5 days p.c. onwards, the nuclear sap in the gonocyte contains small groups of extremely dense particles; these appear to belong to an irregular reticular network formed by other, less dense particles. In addition, the sap contains one or more areas of particles of moderate size (200 – 250 Å) forming groups of 0 · 7 – 1 · 0 μ in diameter (Plate 3, Fig. 7). Even before the onset of meiosis in the female, the nuclear sap is considerably more homogeneous.

With increasing age, the reticular network in the nuclear sap in some—but not all—male germ cells is less pronounced in the immediate vicinity of the nucleolus; a ‘halo’ or clear zone thus appears (Plate 4, Fig. 8). This change may be associated with nucleolar ‘vacuolation’ (see below).

The nuclei of cells believed to be ‘transitional’ are highly irregular in shape (Plate 3, Fig. 5), often showing invagination at one point. The nuclear sap both of these cells, and of spermatogonia type-A, is finely granular (Plate 3, Figs. 5, 6). Cells showing peripheral condensation of nuclear material, observed only in the oldest age-group studied, may correspond to spermatogonia of the intermediate type (see Clermont & Perey, 1957).

At 14 · 5 daysp.c., the nucleoli (probably two or three per cell) are variable in form. They appear to consist, in both sexes, of loosely arranged knots of granular cords; one or more nucleoli in each cell may be attached to the inner nuclear membrane. Portions may show two different electron densities (Plate 4, Fig. 9). Within the following 2 days, the nucleoli begin to acquire a more regular outline; those of some oögonia clearly show zones of two different densities (see Franchi & Mandl, 1962). In male specimens aged 17 · 5 days p.c. to 4 days p.p., the nucleoli consist of a network of dense granular material with light areas within the meshes (Plate 4, Fig. 8). They grow without undergoing any marked ultrastructural changes. In some gonocytes with a perinucleolar ‘halo’, the nucleoli appear foamy and vacuolated.

The nucleoli of ‘transitional’ cells and spermatogonia type-A differ from those of typical gonocytes in that they are angular or rod-shaped, rather than spherical; moreover, they may be larger and contain material of two distinct electron densities (Plate 4, Fig. 10).

There appears to be no marked difference between the sexes in the size and distribution of ribosomes. At early developmental stages, these are abundant in both gonocytes and oögonia; their incidence is of the same order as that in somatic cells. Subsequently, ribosomes tend to be more regularly arranged in groups or rosettes, their incidence becoming less pronounced in germinal than in somatic cells.

The endoplasmic reticulum of mammalian germ cells is known to be very poorly developed, only isolated vesicular profiles being seen in thin sections. The reticulum develops more rapidly in the female than in the male. Between 14 · 5 and 16 · 5 days p.c., that of gonocytes is sparse, consisting of a few rounded vesicles and occasional flattened long tubules or cisternae (Plate 1, Fig. 1); the latter may run parallel to the nuclear envelope. Oögonia already contain more abundant endoplasmic reticulum at 14 · 5 days; it consists of isolated small vesicles, and occasional chains of vesicles or long cisternae.

As the male foetus develops, the incidence of elongated profiles in the germ cells increases; in addition, some discontinuous or complete rings of endoplasmic reticulum of variable size have been observed. At birth, some gonocytes show long tracts in the form of chains of vesicular or short tubular form, studded with ribosomes. Groups of micro vesicles or whorls may be associated occasionally with mitochondria. The endoplasmic reticulum may also form complex structures near the nuclear membrane, reminiscent of multilamellar bodies.

Early post-natal development is not associated with any striking changes in the endoplasmic reticulum except in cells believed to be undergoing degeneration (see below).

Even at the youngest stages studied, the mitochondria of germ cells differ strikingly from those of somatic cells. They are generally larger, round or ovoid in shape (some may be elongated) and contain clearly marked cristae. Spherical mitochondria sometimes show relatively sparse, radially arranged cristae.

While the shape and numbers of mitochondria in gonocytes and oogonia are essentially similar, their distribution diverges between 15 · 5 and 18 · 5 days p.c. They remain randomly scattered throughout the cytoplasm in the female. In the male, they begin to aggregate at one pole of the cell at 15 · 5 days; their ‘polarization’ is most clearly seen at 16 · 5 to 18 · 5 days (Plate 1, Fig. 1).

From 19 · 5 days p.c. onwards, a progressively higher proportion of the mitochondria, which become evenly spread again, undergo slight ultrastructural changes. Some contain dense granules, while others may show vacuolated intercristal spaces, ‘stubby’ cristae or some of villous form (Plate 5, Fig. 11). The majority of the cristae, however, are parallel or radially arranged, as in earlier specimens. Signs taken to indicate mitochondrial budding were observed in specimens approaching full term; some mitochondria are clearly hour-glass shaped (Plate 5, Fig. 11), while others in the neighbourhood are very small. Enlargement of mitochondria associated with irregularity of cristae and a pale internal matrix, accompanied by swelling of the endoplasmic reticulum, is probably a sign of degeneration (see p. 299 and Plate 5, Figs. 12, 13).

Cells establishing contact with the basement membrane, presumed to be ‘transitional’, tend to have a higher proportion of mitochrondria with regularly spaced parallel cristae (Plate 5, Fig. 14). The small germ cells, corresponding to spermatogonia type-A, contain fewer but relatively larger mitochondria whose cristae are typically parallel (Plate 3, Fig. 6). In contrast, large ‘unattached’ gonocytes, persisting in 5- and 6-day-old animals, frequently contain irregularly shaped mitochondria of variable size and internal structure.

Between 14 · 5 and 16 · 5 days p.c., the Golgi apparatus is similar in the two sexes. It consists of a mass of small rounded vesicles with fairly dense contents; similar material may be seen in occasional stacks of elongated vesicles (Plate 5, Fig. 15). Associated with this organelle are several larger vesicles, some 20 0 · 25 – 0 · 35/x in diameter, enclosing a small knot of dense material (Plate 5, Fig. 16). Similar structures occasionally occur in oögonia (see Franchi & Mandl, 1962). Multivesicular bodies are rare, while centrioles are seen frequently.

The complexity of the Golgi apparatus increases with advancing age. By 18 · 5 days p.c., some gonocytes contain a distinctly annular structure, consisting more of micro vesicles than of sac-like stacks. The greatest degree of complexity, again showing an annular arrangement, is seen in germ cells believed to be ‘transitional’ and spermatogonia type-A at 5 to 7 days after birth (Plate 6, Fig. 17).

Occasional multivesicular bodies, apparently not associated with the Golgi apparatus, have been observed in both gonocytes and oögonia at the earliest stages examined. They are also observed occasionally in germ cells throughout the age-range studied; but they appear more consistently in ‘transitional’ cells and spermatogonia type-A.

Between 15 · 5 and 17 · 5 daysp.c., some germ cells contain small electron dense bodies, similar to those in the Golgi apparatus, but only 0 · 1 – 0 · 11 /x in diameter. Dense, osmiophilic fat droplets first appear, though infrequently, in gonocytes of 17 · 5-day-old specimens; while some are round, the majority are contracted and star-shaped. Larger fat droplets appear in older specimens (Plate 6, Fig. 18). Their incidence is variable, and they tend to aggregate in small groups. Somatic cells sometimes contain similar fatty inclusions.

Two distinct classes of cytoplasmic inclusions, referred to for convenience as A- and B-bodies, were observed in developing gonocytes.

These first appear at 18 · 5 to 19 · 5 days p.c., in the form of granular aggregates without membranes. As one or two are generally seen in each sectioned gonocyte, the number per cell is liable to be high. The A-bodies are commonly seen in the vicinity of the nuclear membrane, and are sometimes intimately associated with vesicles of the endoplasmic reticulum. The granules which compose them are finer than ribosomes and less well defined (Plate 6, Fig. 19). In comparison, the collection of granules in the nuclear sap, to which the A-bodies are superficially similar, are coarser and less compact.

The incidence of A-bodies appears to rise with age up to 1 · 5 or 2 days p.p. Thereafter it declines, as does the frequency with which gonocytes themselves occur. Only rarely were they found in germ cells believed to be ‘transitional’ (Plate 5, Fig. 14) or in spermatogonia type-A. In some post-natal specimens, occasional A-bodies were found to contain dense patches or to have a pale centre with a somewhat foamy rim.

Three varieties are included under this heading (see Plate 4, Fig. 8; Plate 6, Figs. 18, 20, 21). Although the individual organelles differ somewhat in appearance, they occur intermingled in clusters within the cytoplasm. All are membrane-bound. Transitional forms are common. B-bodies show no spatial relationships with the Golgi apparatus or other cytoplasmic organelles. The first to be encountered were in specimens aged 18-5 days p.c.

TYPE B₁ Small oval or comma-shaped bodies consisting of a single limiting membrane and a condensed electron opaque core of material sometimes showing signs of internal organization (Plate 6, Figs. 18, 20). Their size is of the same order as that of vesicles of the endoplasmic reticulum. They usually occur in groups in germ cells. Organelles of similar appearance have occasionally been seen in ‘supporting’ cells.

TYPE B₂. Intermediate in size between B₁ and B₃, usually bounded by a single membrane. These bodies have a fairly electron dense core of tightly packed granular or vesicular material which, as in type B₁ is separated from the limiting membrane by a narrow comparatively clear zone (Plate 4, Fig. 8; Plate 6, Fig. 18). This relatively scarce organelle appears in groups, in close association with other B-bodies; occasionally flocculent material develops in its core.

TYPE B₃. Irregular, but generally rounded bodies, initially about half the size of mitochondria. They are bounded by single or double membranes, but occasionally a double structure over part of their circumference coexists with a single one over the remainder (Plate 4, Fig. 8; Plate 6, Figs. 18, 21). Rarely, and only in specimens of more advanced age, do they become multilamellate. B₃-bodies show variable contents : some enclose distorted internal membranes and small vesicular structures, frequently associated with deposits of dense flocculent material applied to the inner limiting membrane; others contain small granules of variable density. B₃-bodies are preceded in appearance by type Bj, but they seem to arise in close association with them in clusters.

The incidence of B-bodies rises as development proceeds. Between birth and 2 days p.p., many gonocytes contain one or more clusters consisting of all three types. Thereafter their incidence, particularly of types B₁ and B₂, declines fairly rapidly. By 3 to 4 days p.p., B₁-bodies are no longer present, while B₂ are rare. As noted above, the number of ‘unattached’ gonocytes declines also. In the meantime, both the appearance and the incidence of B₃-bodies change considerably. Between 21-5 days p.c. and term, the average diameter of B₃-bodies increases to that of typical mitochondria (0 · 6 p). While the majority are smaller than this, some are 0 · 8 – 1 · 5/x in diameter and a few may even reach 2 – 3 μ. These enlarged organelles contain membranous material, partially obscured by amorphous deposits and occasionally also by a granular component (Plate 5, Fig. 12; Plate 6, Fig. 22). Some B₃-bodies of average size show some resemblance to mitochondria; others, however, are somewhat reminiscent of structures in the uterine epithelium, described by Nilsson (1962) and purported to be lysosomes. The larger varieties of B₃-bodies resemble the ‘lysosome-like’ organelles found in the epithelium of the rat prostate by Brandes, Groth & Gyorkey (1962). Although all enlarged B₃-bodies are basically similar in appearance, the nature and complexity of their internal constituents vary considerably.

By 3 to 4 days p.p., such B₃-bodies as persist are largely, though not entirely, restricted to ‘unattached’ gonocytes. No characteristic B-bodies of any type were observed in normal germ cells in 6- and 7-day-old animals. In contrast, germ cells believed to be undergoing degeneration appear to contain abundant B-bodies, particularly large and irregular forms of B₃ whose limiting membranes are frequently incomplete. Cells of this type were observed between 1 · 5 and 3 days p.p. and were also characterized by swelling of the endoplasmic reticulum and/or abnormalities of the mitochondria (see below).

‘Transitional’ cells and spermatogonia type-A in the oldest specimens studied occasionally contained the following cytoplasmic inclusions: (a) large bodies containing membranes and some finely granular or flocculent material; these are more like mitochondria than B₃-bodies; (b) vesicles, resembling those of the endoplasmic reticulum, but containing granules; (c) inclusions similar to but not identical with B₂-bodies; and (d) multivesicular bodies.

Organelles showing some resemblance to B₁-bodies have occasionally been observed in the basal region of ‘supporting’ cells of 6- and 7-day-old animals.

In view of the widely accepted belief that many primordial germ cells degenerate during late foetal and early neonatal life (see Beaumont & Mandl, 1963), particular attention was paid to any ultrastructural changes which might indicate the onset of cell-death.

Occasional pyknotic nuclei and other degraded cellular components, surrounded by abnormal cytoplasm, were observed in early foetal testes; it was impossible, however, to decide whether these were of germinal or somatic origin. The debris appeared to be intimately surrounded by adjacent somatic cells, but there was no unequivocal evidence that it had been engulfed (cf. Franchi & Mandl, 1962).

Germ cells showing signs of degeneration were sporadically observed in a few specimens aged 21 · 5 days p.c. to 2 days p.p. In some cases, swelling of the endoplasmic reticulum was evident in both germinal and adjacent somatic cells. Montages of several electron micrographs indicated that these seminiferous cords were in the immediate vicinity of a trimmed edge of the block. The possibility that the ‘degenerative’ changes were due to mechanical damage cannot, therefore, be discounted.

In some other instances, where the proximity of apparently degenerating cells to the trimmed edge of the block was uncertain, swelling of the endoplasmic reticulum in gonocytes was accompanied by excessively large numbers of B-bodies (Plate 5, Fig. 13), or unusually large (2 – 3 μ) mitochondria (Plate 5, Fig. 12) or marked swelling of the nuclear envelope (Plate 7, Fig. 23). Other abnormal cells were characterized by a reduction in the number of normal mitochondria and an increase in both unusual forms and of ruptured B-bodies containing dense flocculent material (Plate 5, Fig. 13; Plate 7, Fig. 24). Individual germ cells showing these changes were particularly noticed in specimens aged 1 · 5 and 2 days p.p. In the absence of evidence to the contrary, it is highly probable that the few cells showing these striking changes were in effect at early stages of degeneration.

Clear-cut signs of more widespread degeneration first occur in 5-day-old specimens. The affected germ cells are situated in the centre of the seminiferous cord and show no contacts with the basement membrane. Their nuclear and cell membranes are incomplete, and their mitochondria, still recognizable as germinal, disperse over a wide area (Plate 8, Figs. 25, 26). The cell membranes of neighbouring somatic cells also fragment in the vicinity of the gonocyte, though closer to the basement membrane they are bounded in a normal manner. The debris in the centre of the cord consists of a mass of protoplasm in which somatic and occasional germinal nuclei, although distorted, can still be distinguished. The available space is filled by somatic and germinal mitochondria, interspersed with swollen vesicles of the endoplasmic reticulum, occasional multivesicular bodies, and small profiles consisting of islands of cytoplasm, with scattered vesicles and perhaps mitochondria enclosed by a single membrane. Clear signs of canalization of the seminiferous cords are as yet absent. It was not possible to detect any evidence indicating that the debris was being actively removed or absorbed.

The incidence of lysis appears to rise between the 5th and 6th days p.p. At the latter age, it was observed that apical regions of ‘supporting’ cells frequently contain rounded masses of membranous and flocculent material (Plate 8, Fig. 27), similar in appearance to engulfed debris in foetal ovaries (Franchi & Mandl, 1962). These masses, however, are invariably bounded by double membranes.

Gametogenesis is at all times a dynamic process whose interpretation from fixed material is notoriously hazardous. A systematic study of serial sections, together with estimates of cell populations, partially overcome the inherent difficulties of interpretation. Such procedures, however, are hardly applicable to electron microscopy. Even though the present work was preceded by a quantitative histological study of the developing testis in the same strain of rat (Beaumont & Mandl, 1963), it was frequently difficult to equate individual developmental stages (e.g. ‘transitional’ cells) with those seen under the light microscope. Whether or not a cell seen under the electron microscope is cut through its largest diameter is always uncertain. Thus cells largely characterized by differences in size are hard to classify under the electron microscope. Moreover, there may be some unavoidable bias in selecting cells for photographic record. These pitfalls were constantly borne in mind in the course of the present study; and whenever new features first appeared in a proportion of germ cells, every effort was made, by repeated observations, to confirm their general occurrence.

The present findings indicate that shortly after sex differentiation, only minor differences are detectable between male and female germ cells. The most striking divergence, namely their topographical distribution in the gonad and their relationship to adjacent somatic cells, had already been established from light microscopic studies. Electron microscopy has revealed that the development of the oögonial endoplasmic reticulum is more advanced than that in gonocytes; that the latter possess a more angular outline; and that the incidence of desmosome-like contacts is higher in the female. At 15-5 days p.c., when germ cells in both sexes are still mitotically active (Beaumont & Mandl, 1962, 1963), the mitochondria in the male begin to aggregate at one pole of the cell. This ‘polarization’ becomes maximal at 16 · 5 to 18 · 5 days p.c., when mitotic activity begins to decline. Its significance is obscure, but it is worth noting that Nicander, Abdel-Raouf & Crabo (1961) observed similar changes in the gonocytes of bull calves. With advancing development, the mitochondria once again become more randomly distributed. Their temporary ‘polarization’ cannot therefore be assumed to indicate the onset of some irreversible process of differentiation. Since gonocytes undergoing mitosis at 15 · 5 to 18 · 5 days p.c. apparently contain randomly distributed mitochondria, it is possible that mitochondrial ‘polarization’ is a first step in the long period of interphase through which all gonocytes are known to pass before the resumption of mitotic activity at 4 to 5 days p.p. (Beaumont & Mandl, 1963). In the female, mitotic activity is abruptly halted when the germ cells enter the prophase of meiosis (Beaumont & Mandl, 1962; Franchi & Mandl, 1962). It is possible, therefore, that the increase in the population of gonocytes is checked in some manner by mitochondrial ‘polarization’. At present there is no evidence, however, that gonocytes with mitochondrial ‘polarization’ are incapable of dividing mitotically before entering upon their long interphase.

Following cessation of mitotic activity and re-distribution of mitochondria, the male germ cells acquire specific, though heterogeneous groups of organelles (A- and B-bodies) in their cytoplasm. The incidence of these inclusions is related to foetal age and, after birth, to the presence of cytoplasmic extensions linking the germ cell to the basement membrane. The ‘attached’ germ cells are believed to correspond to ‘transitional’ cells (Beaumont & Mandl, 1963) until they give rise, by mitosis, to spermatogonia type-A. Individual gonocytes show signs of degeneration at 2 to 4 days p.p., at a time when only few cells begin to extend cytoplasmic processes to the periphery of the cord. Widespread degeneration does not occur until 5 to 6 days p.p. The majority of cells affected appear to be ‘unattached’ gonocytes at interphase; but occasional abnormal mitoses have been observed in centrally situated germ cells. The significance of the specific organelles, of cytoplasmic contacts with the basement membrane, and of the degenerative changes observed, will be discussed in turn.

The nature and function of the granular cytoplasmic inclusions, here called A-bodies, are entirely obscure. The inclusions were first observed in specimens at 18 · 5 to 19 · 5 days; their incidence decreases rapidly after the 2nd day p.p., when their presence is largely restricted to centrally situated germ cells seemingly lacking contact with the basement membrane.

Three other types of organelle (B₁ B₂ and B₃) have been identified. Although forms intermediate in structure between each of the three types are never difficult to detect, there is no direct evidence that B₁-bodies, which appear first, subsequently become transformed to give rise to B₃-bodies. It is nonetheless possible—though unproven—that B₂-bodies, whose size is intermediate between that of the other two, are formed at the expense of B₁-bodies, and that they subsequently develop into the large, more complex B₃-forms.

B-bodies are particularly characteristic of gonocytes in the late foetal and early neonatal period, though it was not possible to determine with certainty that every germ cell contained them. The rapid decline in their incidence at 3 to 4 days p.p. coincided with the reduction in the numbers of typical rounded gonocytes; at this time, the majority of germ cells show signs of migrating towards the periphery of the seminiferous cords. Certain B-bodies, particularly those of type B₃, are prominent in germ cells whose nuclear envelope, endo-plasmic reticulum and mitochondria show presumed degenerative changes. Excessive numbers of B₃-bodies of irregular size and form are sometimes found in such cells as early as 1 · 5 or 2 days p.p. It would appear, therefore, that the development of large B₃-bodies is in some way correlated with abnormalities in gonocytes of post-natal animals. During the later part of the developmental period studied, widespread breakdown of the central regions of some cords takes place (cf. Beaumont & Mandl, 1963). It is tempting to suggest, therefore, that B-bodies play an important part in the autolysis of some gonocytes, especially those failing to make contact with the basement membrane. If this were so, the organelles might be considered to be lysosomal in nature, and their presence, coinciding with high radiosensitivity (see Beaumont, 1960, 1962; Hughes, 1962), of considerable significance.

In their reviews of morphological and biochemical properties of lysosomes, de Duve (1959a, b) and Novikoff (1961) point out that lysosomes are highly variable in form, and that biochemical evidence does not confirm the existence of a homogeneous class of particles. Indeed, functions attributed to intracellular hydrolases (de Duve, 1959") are so varied as to preclude the possibility of a common form for the lysosome. As judged by their form and/or enzymic properties, and the physiological conditions under which they become manifest, lysosomes apparently occur in many different types of cell; but in all sites in which they have been identified, they appear to play a rôle in either cellular differentiation, catabolism or involution (e.g. Novikoff, Beaufay & de Duve, 1956; de Duve, 1963; Novikoff & Essner, 1960; Moe & Behnke, 1962; Rahman, 1962; Nilsson, 1962; Brandes et al., 1962; Behnke, 1963; Bonneville, 1963; Napolitano, 1963).

Brandes et al. (1962) studied the prostatic epithelium in normal and castrated animals by both histochemical and electron microscopic procedures. They report that acid phosphatase activity is correlated with the presence of dense, membrane-bound organelles resembling lysosomes, both features becoming more prominent after castration. The organelles described and illustrated by Brandes et al. are strikingly similar to many of the B₃-bodies observed in gonocytes. Both are of variable size and shape, sometimes occur in a complex multiple form or in clusters, they are bounded by a limiting membrane and contain variably developed internal membranous and flocculent electron dense material. Nilsson (1962) similarly describes three types of dense cytoplasmic organelles in the uterine epithelium, whose incidence can be varied by the administration of oestrogen. Those of one type detected are similar not only to lysosomes, as described by Novikoff & Essner (1960), but also to some of the B₃-bodies in our material.

Lysosomes are apparently capable of storing a host of different enzymes (see Novikoff, 1961). Of those that have been identified, acid phosphatase is often prominent (de Duve, 1959a) and the simplest to detect by histochemical or by cytochemical methods (Holt & Hicks, 1961; Sabatini, Bensch & Barrnett, 1963). Preliminary attempts to establish the presence of acid phosphatase activity in the cytoplasm of gonocytes have so far failed, though sections of liver, processed simultaneously, indicated the presence of lysosomes in the expected manner. Moreover, Leydig cells in the sections of foetal rat testes showed some activity; Jirásek’s (1962) histochemical study of interstitial cells in the human embryonic testis yielded similar results. The apparent absence of detectable quantities of acid phosphatase in B₃-bodies does not, however, invalidate the hypothesis that they are lysosomal in nature (see Novikoff, 1961).

It is of interest to note that a marked rise in lysosomal enzyme activities, in some instances known to be associated with enlargement of lysosomes, occurs not only in spontaneously regressing tissues (e.g. embryonic Müllerian duct; tail of tadpole) but also in those subjected to a variety of noxious procedures (e.g. ligation, anoxia; see Novikoff, 1961). The levels of a variety of lysosomal enzymes in thymus and spleen have also been shown to rise steeply following irradiation (e.g. Roth & Eichel, 1959; Roth, Bukovsky & Eichel, 1962; Roth & Hilton, 1963). It would be premature to postulate that the presence of lysosomelike bodies in gonocytes is a major cause for their high radiosensitivity, ionizing radiations rupturing their membranes and ‘liberating’ their enzymes. First, it is possible that irradiation induces surviving cells in the spleen and thymus to produce more enzymes, rather than lysosomal breakdown and subsequent release of pre-existing ‘bound’ enzymes. Second, Rahman (1962) suggests that radiosensitivity and abundance of lysosomes in thymus cells may be inversely correlated. Third, if ionizing radiations kill gonocytes by rupturing the membranes of lysosome-like bodies, it would be expected that shortly after irradiation, there would be signs of widespread lysis. Beaumont’s observations (personal communication) point to the contrary: gonocytes exposed to a dose sufficient to induce subsequent sterility persist for several days after irradiation, and though many enlarge somewhat, none show signs of disintegration. It is as yet uncertain how long after irradiation gonocytes, or their daughter-cells, involute. If the major loss occurs at the first post-irradiation mitosis, the contribution of latent chromosomal damage may well outweigh that of enzymic release during interphase. Finally, mammalian oocytes in primordial follicles, known to be highly radiosensitive though mitotically inactive, apparently lack lysosome-like bodies. In his careful study of unirradiated and irradiated oocytes in young mice, Parsons (1962) makes no mention of such organelles.

The disappearance of B-bodies from germ cells which, after birth, become attached to the basement membrane poses further problems. If the lysosomelike bodies contain potentially autolytic enzymes, it would appear that either the enzymes are ‘resorbed’ or that they remain within the cell in some ‘bound’ or ‘inhibited’ form. The persistence of B-bodies in ‘unattached’ germ cells, and their abundance in cells with swollen endoplasmic reticulum and abnormal mitochondria, suggest that they may be somehow involved in spontaneous degeneration. The establishment of cytoplasmic contacts between the germ cell and the basement membrane appears to be a pre-requisite for subsequent survival and mitotic activity. What ‘advantages’, if any, the cell derives from such contacts are as yet unknown. It is difficult to decide, in the light of the present observations, whether B-bodies disappear from germ cells shortly before or after the first contact with the basement membrane has been established. All that can be said at present is that the two phenomena are related in time; but whether there is any causal relationship between the two remains to be determined.

Migration of gonocytes towards the periphery of the seminiferous cord has been observed in other species (e.g. Courot, 1962; Sapsford, 1962a, b). That they should be capable of doing so is hardly surprising in view of Blandau, White & Rumery’s (1963) recent demonstration that amoeboid movement in female germ cells persists up to the onset of pachytene. Blandau et al. were unable to observe any amoeboid movement in gonocytes after their inclusion within medullary cords, but the oldest animals they examined were approaching full term, when gonocytes are known to be tightly packed within the cords. With the subsequent rapid proliferation of somatic tissue (see Beaumont & Mandl, 1963) and the relative decrease in the incidence of germ cells per cross section of seminiferous cord, all that separates the centrally situated germ cell from the basement membrane is somatic cytoplasm. This would presumably offer no greater resistance to the ‘streaming’ of germinal cytoplasm than the loose tissue surrounding developing oöcytes.

The process of spontaneous degeneration, which affects a relatively small proportion of the total population of male germ cells (Beaumont & Mandl, 1963; see also Courot, 1962) differs strikingly both in time and nature from that observed in the female (Beaumont & Mandl, 1962; Franchi & Mandl, 1962). Pyknosis is rare, and it was impossible to decide whether such pyknotic nuclei as were recorded in the youngest specimens were germinal or somatic in origin. At the time that many female germ cells are eliminated from the ovary (18-5 days p.c. to 2 days p.p\ the gonocytes are at interphase, and none or few degenerate. Only as late as 4 to 5 days p.p., when the rate of spontaneous atresia in the female has begun to fall, does degeneration set in on a detectable scale in the male. When it does so, the cell membrane ruptures, the cytoplasm streams out and the nucleus disintegrates. Dying or dead cells in the foetal and neonatal female appear to be removed rapidly (Beaumont & Mandl, 1962) by the phagocyte-like activity of neighbouring somatic cells which engulf the debris (Franchi & Mandl, 1962). The process by which lysing gonocytes are eliminated from the testis appears to differ. Though the peripheral cytoplasm of ‘supporting’ cells occasionally contains cellular debris—which may or may not be germinal in origin—no clear-cut signs of rapid phagocytic removal have been observed. The cytoplasm of the degenerating germ cell disperses over a wide area, and small fragments may be absorbed or ‘digested’ by the cytoplasm of all the neighbouring somatic cells confluent with the lysing gonocyte. The ingestion of degenerating germ cells by Sertoli cells, observed by Lacy (1964) in adult rats following irradiation or treatment with oestrogen, does not appear to occur in the neonatal rat. Although the developing ‘supporting’ cells (precursors of Sertoli cells) are seemingly capable of engulfing cellular debris, the process whereby the germ cells degenerate clearly differs between the unirradiated neonatal and irradiated adult rat.

1. On a examiné au microscope électronique les gonocytes mÅles de foetus et de nouveau-nés de rat (de 14 jours post coitum jusqu’à 7 jours post partum). Pour faire la comparaison, on a examiné de nouveau des gonocytes femelles de jeunes spécimens (14 jours à 16 jours p.cf

2. Peu après la différenciation sexuelle, les gonocytes mÅles et femelles ont en commun beaucoup de caractères. Néanmoins, l’ergastoplasme des o ö gonies est le plus développé, tandis que les contours des gonocytes deviennent nettement plus anguleux pendant que le développement progresse. Entre 15 jours et 18 jours p.c., chez le mÅle, les mitochondries s’agrègent à un pöle de la cellule; chez la femelle, elles restent réparties au hasard. La signification de cette ‘polarisation’ mitochondriale est obscure, mais il est possible que le processus soit en relation d’une manière quelconque avec l’arrêt de l’activité mitotique.

3. A partir de 18 jours p.c., quand les mitochondries se trouvent de nouveau également réparties, apparaissent dans le cytoplasme des groupes d’organites spécifiques (corps A et B). On a identifié trois types de corps B, dont l’un ressemble aux lysosomes. Le nombre des corps B augmente pendant la fin de la vie intra-utérine et atteint son maximum entre la naissance et le 2e jour p.p. Ensuite, ils deviennent plus rares, comme les gonocytes situés au centre du cordon séminifère. A ce moment, une forte proportion de gonocytes émettent des expansions cytoplasmiques établissant un contact avec la membrane basale; leur noyau et leur nucléole changent de forme, tandis que leurs mitochondries contiennent des crêtes disposées plus régulièrement que ne le sont celles des gonocytes ‘non attachés’. On croit que les cellules présentant ces signes ‘d’écoulement’ ou de migration correspondent aux cellules de transition identifiées au microscope ordinaire, et sont les précurseurs immédiats de la première génération de spermatogonies de type A. Après le 2e jour p.p., les corps B persistent presque uniquement dans ceux des gonocytes qui n’ont pas établi de contact avec la membrane basale. On discute la signification possible des corps B dans la dégénérescence spontanée et la forte radiosensibilité des gonocytes.

4. Quelques gonocytes isolés montrent des modifications de dégénérescence pendant les trois premiers jours après la naissance. Ils montrent fréquemment un ergastoplasme gonflé, des mitochondries anormales et une abondance de corps B. La dégénérescence étendue ne s’instaure pas avant 5 jours environ et elle atteint largement les gonocytes ‘non attachés’ en interphase; on a observé la mitose de quelques cellules en dégénérescence. Celle-ci prend l’allure d’une lyse: la membrane cellulaire se rompt, le cytoplasme s’écoule sur une large surface et le noyau se déforme. On n’a pas pleinement établi le processus par lequel les gonocytes en cytolyse sont éliminés du testicule.

5. L’ultrastructure des spermatogonies du type A est essentiellement semblable à celle de leurs précurseurs. La plupart des petits gonocytes correspondant aux spermatogonies de type A vues au microscope optique ont leur grand axe parallèle à la membrane basale.

The expenses incurred in this study were defrayed from grants, made to Professor Sir Solly Zuckerman, F.R.S., by the Population Council, Inc. and by the Medical Research Council. The Authors are grateful to Professor Sir Solly Zuckerman, F.R.S., for his encouragement and valuable criticisms.