Experimental manipulation of sexual differentiation in wallaby pouch young treated with exogenous steroids

Sexual differentiation in eutherian mammals is believed to be a two-stage process. First, there is a genetically controlled differentiation of the indifferent gonad into a testis or an ovary, which is followed by differentiation of the Wolffian and Müllerian duct systems in the male under the influence of testicular hormones (Jost, 1947, 1953; Wilson et al. 1981). Marsupial mammals follow a broadly similar pattern, but there are some intriguing differences (Renfree et al. 1987; Renfree & Short, 1988).

The first studies on the hormonal control of sexual differentiation in marsupials were carried out in the Virginia opossum (Didelphis virginiana). In this species, the young are bom just after the first morphological differentiation of the testis (McCrady, 1938), so most of sexual differentiation occurs whilst they are readily accessible in the pouch. This made it easy to paint the skin of the pouch young with various steroids and study their subsequent sexual differentiation (see review by Burns, 1961). The responses of the Wolffian and Müllerian ducts to androgen and oestrogen treatment were similar to those of rat and mouse fetuses (Raynaud, 1958). However, neither androgen nor oestrogen treatment affected development of the scrotum, pouch or the mammary glands. Burns also reported that male young treated with oestrogen from the day of birth developed ovotestes. More recently, Fadem & Tesoreiro (1986) reported that oestrogen treatment of newborn male young grey opossums (Monodelphis domestica) prevents testicular differentiation.

In the tammar wallaby (Macropus eugenii), we have recently shown that the gonads are morphologically undifferentiated at birth, and there is no evidence of testicular hormone production (O et al. 1988). Despite this, the mammary and scrotal anlagen and the gubemaculum already show marked sexual dimorphism, suggesting that these structures differentiate under genetic rather than hormonal control. The purpose of the experiments described here was therefore to treat karyotypically sexed newborn wallaby pouch young with heterotypic gonadal steroids to determine which aspects of sexual differentiation are hormonally controlled.

Pregnant female tammar wallabies (Macropus eugenii) from our breeding colony at Monash University were checked daily for the presence of newborn young. Any neonates were examined in the pouch with an otoscope to detect the presence of scrotal bulges, which are confined to males (O et al. 1988).

Presumptive female pouch young without scrotal bulges on the day of birth (day 0) were treated with testosterone propionate (50 mg ml⁻¹ in arachis oil, Schering, AG, Berlin) (n = 17) or with an equivalent volume of arachis oil alone (n = 7). Presumptive male pouch young with scrotal bulges were treated with oestradiol benzoate (5 mg ml ⁻¹ in arachis oil, Organon Labs, Morden, UK) (n = 11), or with an equivalent volume of arachis oil alone (n = 7) (Table 1). These hormones were administered daily by mouth (Jurgelski, 1971) using a fine polythene capillary (ID 0-4mm, OD 0-8 mm, Dural Plastics, Dural, NSW, Australia) from day 0 to day 24. Since the young are growing rapidly at this stage, the volume of hormone administered was increased every 5 days to maintain the dosage rates between 24 and 50mgkg⁻¹ day⁻¹ for testosterone, and 1·2 to 2·5 mg kg⁻¹ day⁻¹ for oestradiol (Table 2). On day 25 all surviving pouch young were killed by decapitation, immer-sion fixed in neutral-buffered 10% formalin for 7-14 days, decalcified for 24h in 11 % formalin and 30 % formic acid, paraffin embedded, serially sectioned at 5 μm, and stained with haematoxylin and eosin.

The sex of the young was confirmed by karyotype (O et al. 1988) at the time of autopsy. In all cases, the presumptive sex was in agreement with the genetic sex.

The cross-sectional areas of the Wolffian and Müllerian ducts were determined on every 20th section using a projection microscope and a MOP-1 digitizing tablet (Carl Zeiss, Munich, Germany). The volumes of these ducts and their lumina were estimated by the sum of these areas multiplied by the distance between the measured sections (100 μm). The volume of the duct wall was determined by subtracting the lumen volume from the total duct volume.

The position of the gonads was determined relative to the lumbar vertebrae. The position of the intervertebral disks was determined by the presence of the nucleus pulposus. Once the testes had moved into the inguinal canal, the position was termed ‘inguinal’. The internal inguinal ring was located in sections containing the fifth lumbar vertebra.

Initial paired tests were used to establish that there was no systematic difference between measurements on the left and right sides of the reproductive tract. Subsequent analyses were made using the sum of the left and right measurements and unbalanced analysis of variance with Duncan-Waller multiple range tests. The data were log-transformed before analysis, where necessary, to correct for heteroscedasticity.

During the course of treatment a number of young died (Table 1). These were either found dead in the pouch or were lost from the pouch; only a few of these were control (oil-treated) animals. These deaths may have resulted from a combination of trauma during the daily catching and handling of the mother, damage to the mouth of the young by the capillary, and toxicity of the compounds being administered. More than half the testosterone-treated female pouch young died within the first 10 days of treatment, and more than half the oestradiol-treated male young died between day 10 and 19 of treatment. Most of these oestrogen-treated young were found dead in the pouch and showed significant oedema and grossly distended urinary bladders, ureters and kidneys at autopsy, suggesting that death resulted from blocked urinary ducts.

The gross anatomy of the adult female reproductive system of marsupials is different to that of eutherian mammals, due to a difference in the migration of the embryonic urinary and genital ducts (Tyndale-Biscoe & Renfree, 1987). In marsupials, the ureters pass medial to the Wolffian and Müllerian ducts (Figs 1–4), rather than laterally to them, as in eutherians. As a result, in female marsupials, the Müllerian ducts do not fuse to form a single median vagina or uterus. Instead, the uteri and vaginae remain separate. The developing Müllerian ducts fuse medially at the junction of the developing uterus and vaginal sections to form an anterior vaginal expansion, from which the two lateral vaginae develop, looping around the ureters as they pass down to the urogenital sinus.

In control (oil-treated) females, the Müllerian ducts were well developed (Fig. 1, Table 3). Regional differentiation had occurred. Near the ovary, the duct was relatively narrow with cuboidal to low columnar epithelium, becoming progressively larger, with a higher columnar pseudostatified epithelium in the region which would later develop as the uterus. The duct epithelium was pseudostratified columnar, with deeply staining oval nuclei. The lumen was patent for the whole length of the duct. Below this was a medial expansion which presumably would develop into the anterior vaginal expansion. From this, the duct passed laterally around the ureter before running medially to join the urogenital sinus. In this latter portion, the lumen became a narrow, slit-like opening, and close to the urogenital sinus, in the region where it fused with the Wolffian duct, the epithelium of the Müllerian duct had a stratified appearance similar to that of the urogenital sinus. Along their whole length the Müllerian ducts had dark-staining basophilic cytoplasm and nuclei. By contrast, the Wolffian duct epithelium, over most of its length, had a single layer of pale-staining cuboidal cells. The Wolffian duct had regressed significantly, with both duct wall and lumen having a reduced volume compared to control males (Table 3). The regression was most pronounced at the proximal and distal ends, where in places the duct was not patent (Fig. 1).

In testosterone-treated females, the Müllerian ducts were enlarged compared to control females, although their histological appearance was similar (Fig. 2, Table 3). The Wolffian ducts were greatly enlarged, although they had a cuboidal epithelium similar to control females, and they were patent for their whole length. The most marked cytological differences were seen at the ovarian end, where the duct appeared to be distended and had a flattened to squamous epithelium.

Control males had well-developed Wolffian ducts (Fig. 3, Table 3) with cuboidal epithelium and patent lumina over most of their length. Because the testes had migrated to the inguinal canal, the Wolffian ducts followed a fairly convoluted course, and between the internal inguinal ring and the urogenital sinus the surrounding connective tissue was reduced to a slender spermatic cord. The Müllerian ducts had regressed to rudiments; only the short section at the urogenital sinus, and the segment closest to the ovary and vestigial mesonephros remained, and these sections were substantially smaller than the corresponding sections in the control females.

In oestradiol-treated males, the lumina of the Wolffian duct were similar in size to control males, but the volume of the duct epithelium was reduced and more akin to control females (Fig. 4, Table 3). The epithelial cells of the Wolffian duct appeared similar to control males, except for apparently poor development close to the gonad.

Oestradiol treatment prevented Müllerian duct regression. The Müllerian ducts of treated males were significantly larger, both in terms of volume of the lumen and the volume of the wall of the duct, than in control males, and similar in volume to the Müllerian ducts of testosterone-treated females (Fig. 4, Table 3). In some individuals the lower portion of the Müllerian duct was absent, the duct ending blindly in a medial expansion, whilst in others, this expansion connected by a narrower portion to the urogenital sinus. The duct epithelium was high columnar or pseudostratified over most of its length, but towards the urogenital sinus the epithelium became stratified, similar to the sinus epithelium.

The ovaries of females at day 25 were smaller than the testes of males (Table 3, Fig. 5), irrespective of treatment.

Ovaries of the control females had a well-defined cortex distinctly separated from the medulla (Fig. 5). The cortex contained numerous nests of germ cells. A few germ cells near the medulla were undergoing meiosis, but the majority were in the leptotene or prophase stage as was reported by Alcorn (1975). The surface epithelium was a single cuboidal or pseudostratified layer. The medulla contained mainly stromal tissue dissected by rete cords. Testosterone treatment had no obvious effect on ovarian histology (Fig. 5) or volume (Table 3).

The testes of control males (Fig. 5) had well-developed seminiferous tubules surrounded by a distinct basement membrane. Most germ cells were in interphase and were contained within the tubules. A few germ cells were scattered between the tubules and occasionally outside the tunica albuginea, beneath the surface epithelium. The tunica albuginea was a well-defined band of fibrous connective tissue immediately below the surface epithelium which comprised a single layer of squamous to cuboidal cells. Small nests of Leydig cells were present between the tubules, particularly at the peripheral region of the gonad.

Oestradiol treatment had no effect on testicular volume (Table 3), but caused a number of gross histological changes (Fig. 5). Most of the gonad contained seminiferous tubules but portions at the periphery of the gonad lacked tubules and contained either fibrous connective tissue, or non-tubular stromal tissue which was reminiscent of ovarian tissue. The seminiferous tubules which were present appeared smaller and more widely spaced than in control testes. Between the tubules there appeared to be more fibrous connective tissue than in control testes and very few Leydig cells were seen. The tunica albuginea became indistinct at these cortical nontubular regions, although it was well defined over the tubular regions of the testis. The surface epithelium was a thin layer of squamous or cuboidal cells, as in control testes. Over the non-tubular regions the surface epithelium was a thicker layer of cuboidal or pseudostratified cells. Within the tubules many of the germ cells had dark-staining pyknotic nuclei. Germ cells were also abundant in the cortical non-tubular areas. Most of these appeared to be in interphase.

The ovaries of females were intra-abdominal, at the level of the last two lumbar vertebrae 4–5 (Figs 1, 2, 6). Testosterone treatment did not induce any descent; two of the treated females had ovaries noticeably higher than in controls. The gubernaculum of all females, irrespective of treatment, merged with the abdominal wall without a distinguishable processus vaginalis.

The testes of the control males had descended into the inguinal region whereas the testes of the oestra-diol-treated males were high in the abdomen, in a position similar to (or higher than) that of control females (Figs 1, 3, 4, 6). In control males, the processus vaginalis was patent right into the centre of the scrotum, whilst, in all oestradiol-treated males, the processus vaginalis and gubernaculum terminated in the abdominal wall outside the scrotum (Fig. 8) similar to normal males on day 10 of pouch life (unpublished observations).

The urogenital sinus was similar in appearance in control males, control females and testosterone-treated females (Fig. 7). In the oestradiol-treated males, the epithelium of the urogenital sinus was grossly hypertrophied and apparently hyalinized and the epithelial cells showed frequent mitoses (Fig. 7).

In control males, small prostatic buds were present in the urogenital sinus wall just caudal to the junction of the Wolffian ducts (Fig. 7C). Similar prostatic anlagen were also seen in testosterone-treated but not in control females (Fig. 7A,B). It is unclear if prostatic anlagen were present in a similar location to the urogenital sinus of oestradiol-treated males (Fig. 7D) because of the convoluted nature of the epithelium in these animals.

Female young all had well-developed pouches and mammary development was unaffected by testosterone treatment (Fig. 8A,B). In neither control nor oestradiol-treated males was there any evidence of pouch or mammary development.

All males had a well-developed scrotum into which the gubernaculum and processus vaginalis passed (Fig. 8C,D). Oestradiol treatment had no effect on the size or form of these structures. There was no sign of scrotal formation in any androgenized female.

Both male and female young have a prominent phallus at birth. At day 25, the phallus was similar in size and appearance in males and females, and neither androgen nor oestrogen had a noticeable effect. However, an additional female that was treated with testosterone propionate up to day 30 had a noticeably enlarged phallus compared to untreated male or female young at this stage (data not shown).

The treatment period, from birth to day 25, includes many critical events of sexual differentiation. At birth, the gonads of males and females are morphologically similar (O et al. 1988). Within two days of birth, the testes of male young develop seminiferous tubules (Renfree et al. 1987) and secrete MIS (Hutson et al. 1988). By day 25, the Müllerian duct of the control males has regressed over most of its length and the Wolffian duct is well-developed, whilst in control females the Müllerian duct is large and shows regional differentiation into Fallopian, uterine and vaginal segments but the Wolffian duct is regressing (Renfree et al. 1987; Alcorn, 1975).

In eutherian mammals, the differentiation of the Wolffian and Müllerian ducts and their derivatives is controlled by gonadal hormones. The results of this study show that the development of these ducts and the urogenital sinus is under similar control in the tammar wallaby (Table 4).

Testosterone treatment stimulated the development of the Wolffian duct, and induced the formation of prostatic buds in the urogenital sinus. These results are similar to those obtained by Burns (1961) in the opossum, and are comparable to the findings in eutherian mammals (Wilson et al. 1981). Since prostatic buds were present in the control males at day 25, the testes are presumably secreting androgens at this stage, a conclusion supported by the greater development of the Wolffian ducts in control male young compared to control female young.

Burns (1961) demonstrated that the phallus of the opossum was markedly enlarged by androgen treatment from day 0 to day 20. Although no effect of androgen treatment was observed up to day 25 in the tammar, more prolonged treatment of a single animal appeared to increase phallus size. The phallus seems to be less sensitive than some other androgen-responsive tissues. Normally no difference in phallus size is seen between male and female young until after day 60 (unpublished observations), even though sufficient androgen is present by day 25 to support Wolffian duct and prostate development. Since Bums used higher doses applied cutaneously over the abdominal region, it is probable that he achieved much higher doses in the phalluses of his treated opossums than we achieved by oral administration in our tammar young.

Oestradiol treatment stimulated development of the urogenital sinus and also appeared to inhibit testicular androgen secretion as evidenced by the reduced volume of the Wolffian duct wall in treated males. It also appeared to inhibit Müllerian-inhibiting substance secretion or action since the Müllerian ducts were retained in the treated males. Both these inhibitory actions of oestrogen could be a consequence of the inhibition of testicular development, although we do not know whether this effect is centrally or peripherally mediated.

MIS has also been implicated as a hormone involved in testicular differentiation (Vigier et al. 1987) so inhibition of MIS production or its action might have consequences for testicular development. Previous studies in opossums have shown that oestradiol treatment of neonatal male marsupials can prevent normal testicular formation (Fadem & Tesoreiro, 1986) and may even lead to formation of an ovotestis (Burns, 1961). In this study, less-dramatic effects on testicular development were seen. However, the increased thickness of the cortical tissue outside the tunica albuginea and the large number of extratubular germ cells in the cortical regions is evidence of partial gonadal sex reversal. The limited conversion we achieved may be related to the hormone dosage used, or may reflect the critical timing that is needed.

All adult male marsupials lack mammary glands, nipples and pouches (Tyndale-Biscoe & Renfree, 1987). Some eutherian males also lack mammary glands; in male rat fetuses, androgens act on the developing mammary anlagen to cause regression (Kratochwil, 1977). No evidence of androgen-induced mammary, nipple or pouch regression was seen in the testosterone-treated females in this study, nor was there any evidence of mammary development in oestradiol-treated male pouch young (Renfree et al. 1987; O et al. 1988). Similarly, in the newborn opossum, treatment with high doses of androgens or oestrogens did not inhibit development of the pouch or mammary anlagen in the female, or scrotal development in the male, nor was there any induction of mammary development in males, or scrotal development in females (Burns, 1961). These observations support the conclusion of O et al. (1988) that the development of these structures is not controlled by gonadal hormones, but is under autonomous genetic control.

The control of testicular migration and descent has been the subject of controversy. Some authors maintain that the process is mediated entirely by androgens (Rajfer, 1982; Hadziselimovic, 1983), whilst others suggest that another testicular hormone, possibly MIS, is involved in the initial stages (Donahoe et al. 1977; Habenicht & Neumann, 1983; Hutson, 1985). In newborn tammars, the indifferent gonads of males and females are at equivalent abdominal locations (O et al. 1988), and in both sexes the gonads descend slightly relative to the lumbar vertebrae up to about day 20 (unpublished observations). From this stage onwards, the testes move to the inguinal ring and pass through it before day 25. In oestradiol-treated males the testes remain intraabdominal, at a level similar to the ovaries of control females. Similar observations on fetally oestrogenized mice have been interpreted as due to an inhibition of androgen production (Hadziselimovic, 1983). However, since testosterone treatment of females throughout this period did not cause any ovarian migration (indeed the ovaries of some testosterone-treated females were higher in the abdomen than in controls), our experiments support the view that androgens alone do not mediate this stage of transabdominal testicular migration, so MIS may well play a role. However, the gubernaculum of neonatal tammars is sexually dimorphic prior to morphological differentiation of the gonads, so this may also affect the progress of gonadal migration in this species.

These studies have shown that the control of the development of the Wolffian and Müllerian ducts in this marsupial is similar to that seen in eutherian mammals. However, the differentiation of the mammary gland, pouch, scrotum and gubernaculum, which are not derived from these duct systems, are not under hormonal control (Table 4) and in this respect marsupials appear to differ from their eutherian counterparts, from whom they have been isolated for about 100 million years.

We thank Dr T. P. Fletcher for assistance with the animal handling. This work was supported by grants from ARGS and the Monash Special Research Fund, and permit no. 8428 from Dept of Conservation, Forests and Lands, Victoria, Australia.