Effects of testosterone implants in pregnant ewes on their female offspring

The first suggestion that female foetuses might be masculinized by testicular hormones was provided by Lillie’s classical study of the bovine freemartin (Lillie, 1916,1917). Pfieffer (1936) demonstrated a similar effect by transplanting testicular tissue to neonatal female rats. Later studies revealed that a single neonatal injection of testosterone propionate would masculinize female rats. Since this time the androgen-sterilized rat has been the subject of many detailed investigations (Harris, 1964; Barraclough, 1968; Harris, 1970; Beach, 1974; Brown-Grant, 1974; Davidson, 1974). Androgenization of the female foetus or neonate has also been studied in the guinea-pig (Barraclough & Gorski, 1961; Goy, Bridson & Young, 1964), hamster (Paup, Coniglio & Clemens, 1972; Swanson & Brayshaw, 1973), mouse (Manning & McGill, 1974), dog (Beach & Kuehn, 1970; Beach, 1975), sheep (Short, 1974), cow (Jost, Chodkiewicz & Mauleon, 1963), monkey (Goy, 1970) and man (Money & Erhardt, 1972).

Androgenized sheep have a number of advantages for experimental studies.

They are large enough to permit serial blood sampling over long periods of time; this is essential if one is to investigate pituitary gonadotrophin secretion and gonadal feedback mechanisms. They are also ideal for observing the behavioural consequences of androgenization, since the social and sexual behaviour of sheep reveals a clear-cut dimorphism between the two sexes. The purpose of the present investigation has been to document as fully as possible the anatomical and behavioural changes in newborn female sheep exposed to testosterone at different stages of intrauterine life, as a prelude to subsequent studies on their endocrinology and reproductive behaviour in adulthood.

During the 1973 mating season the oestrous cycles of twenty-four Finnish Landrace × Dorset Horn ewes were synchronized with progestagen-impregnated intra vaginal sponges (Synchromate; G. D. Searle & Co., High Wycombe, Bucks, England) containing 30 mg fluorogestone acetate (SC 9880; 17α-acetoxy-9α-fluoro-11 β-hydroxypregn-4-ene-3:20 dione). Mating to Finn-Dorset rams took place at the second oestrus following sponge withdrawal. The ewes were then randomly divided into three groups of eight. Implants of 1 g (1·010 ± S.E. 0-003) testosterone (Organon Laboratories Ltd., Newhouse, Lanarkshire, Scotland) were placed subcutaneously in the neck of two groups of ewes, using aseptic procedures under local Xylocaine anaesthesia. One group was implanted on day 30 (range 29–30) of gestation and the implants were removed on day 80 (range 80–81); another group received implants on day 50 (range 48–51) which were removed on day 100 (range 98–101). The third group of ewes served as controls and were not implanted. One ewe in the control group died, and one animal in the D30–80 group was not pregnant. During April 1974 the remaining 22 ewes produced 54 lambs (24 male, 30 female) over a ten-day period. Large litters were reduced to two lambs and the surplus animals were reared artificially. All the females were retained as experimental subjects. Six males were kept as controls, and were bilaterally vasectomized at 8 months of age.

During the 1974 mating seasons 1 g testosterone was implanted into two further groups of six pregnant ewes on days 70 or 90 of gestation and the implants were removed on days 120 or 140 respectively. Oestrus had not been synchronized in these ewes but they all lambed in February 1975 over a four-week period, giving 21 lambs (9 male, 12 female). All the females were kept and the males were discarded. Table 1 summarizes these results.

Blood samples were taken from the jugular veins of all the pregnant sheep at the time of implant removal, and plasma testosterone levels were measured by the technique of Corker & Davidson (1976). Single samples were taken from the ewes lambing in 1974 and three half-hourly samples were taken from those lambing in 1975.

The size of the clitoris in the implanted ewes was subjectively assessed on a 0–5 scale at the time of implant removal and again just prior to parturition. After removal, the testosterone implants were dried and weighed to determine the amount of testosterone released over the implantation period, and hence it was possible to estimate the daily release rate.

The appearance of the external genitalia of all the lambs was recorded at birth, and the size of the clitoris assessed at 10 months of age. The penile diameters of the female offspring showing complete masculinization were compared with those of normal male controls at 10 months of age, using callipers to obtain three measurements immediately anterior to the scrotum.

The D30–80, D50–100, D70–120 and D90–140 and control lambs were run together after weaning (12 weeks of age) and their urination behaviour was observed under field conditions.

The internal genitalia of the D30–80, D50–100 and control lambs were examined at 12–18 months of age by laparotomy under nembutal anaesthesia. Two sheep from each of the D30-80 and D50–100 groups were killed during the course of study. Post-mortems were carried out immediately following death, and ovaries, accessory glands, uterus and vagina were fixed in 10 % formal saline, sectioned at 6 μm and stained with haematoxylin and eosin.

Following implantation the D30–80 and the D50–100 ewes showed increased aggressive behaviour. This was not shown by the control ewes or those implanted later in gestation. No objective methods were used to quantitate this behaviour.

The implanted ewes all lambed after a normal gestation of approximately 143 days and no dystokia was encountered. There was some intrauterine death, and although this was higher in the groups implanted later in gestation (Table 1), the group differences did not achieve significance (0·05 < P < 0·10; Table 1). There was no difference in the number of dead lambs born during 1974 or 1975. All the ewes lactated, although two D90·140 mothers did not produce enough milk to feed even a single lamb, so supplementary feeding was necessary.

Testosterone concentrations in the jugular venous plasma of the ewes at the time of implant removal are given in Table 2, together with the estimated daily release rate. The variability was greater in the single samples collected in 1974, and the three half-hourly samples taken in 1975 reduced the standard errors considerably (Table 2). The D30-80 ewes had significantly (P < 0·05) higher testosterone levels than the D70–120 and the D90–140 animals. Correlation of plasma testosterone levels with time of implant removal was significant (r = − 0·526; P < 0·05). The daily release rate of testosterone was similar in all groups. The plasma testosterone levels in the control ewes were below the limit of sensitivity of the assay (< 0·1 ng/ml).

Subjective assessment of clitoral size in pregnant ewes at implant removal are given in Table 3. Enlargement was greater in the D30–80 and D50–100 groups than in the D70–120 and D90–140 groups. The hypertrophy diminished rapidly following withdrawal of the implants and the clitoral size had returned to normal by the time of parturition.

The genetically female lambs could be identified at birth by the fact that regardless of the extent of masculinization of the external genitalia, the ovaries remained in their normal intra-abdominal position; thus the scrotum of the masculinized animals was always empty, whereas the normal males had palpable testes present in the scrotum from birth onwards.

Females exposed to testosterone between days 30–80 of gestation showed complete masculinization of the external genitalia; they were born with a penis and an empty scrotal sac (Fig. 1C). The processus urethrae appeared similar to that of the normal male, but the prepuce could not be drawn back from the galea capitis, and the preputial opening appeared abnormal (Fig. 1C and D). The penile diameters of the D30–80 females (0·77 ± 0·10 cm; M±S.E.M.) were significantly smaller (P < 0·01) than the male controls (1·20 ± 0·10 cm; M ± S.E.M.) when measured at 10 months of age.

The external genitalia of the androgenized females are compared with normal female controls at 10 months of age in Table 4. In the normal ewe the clitoris is short and lies in the fossa clitoridis, surrounded by the labia (Fig. 1 A). The D50–100 females showed marked elongation and fusion of the labia within which the enlarged clitoris could be palpated, although it could not be extruded. In addition to the small vaginal opening, these ewes displayed a second opening at the tip of the fused labia (Fig. 1 B). The D70–120 animals showed less marked external masculinization, and the external genitalia of the D90–140 group did not differ from the normal females (Table 4).

Laparotomy revealed that all the androgenized ewes possessed ovaries. The uteri and vaginae of the D30–80 ewes were markedly distended with fluid but this was not so in the other experimental groups. One D50–100 ewe had a congenitally absent left uterine horn, with a fallopian tube that was occluded and distended with fluid.

Two D30–80 animals were killed at 10 months of age, during their first mating season. A single corpus luteum was present in the ovary of one animal, and it appeared functional when examined histologically; the uterine and vaginal histology was also characteristic of the luteal phase of the cycle (McKenzie & Terrill, 1937; Robinson, 1959). The other ewe showed a heavily luteinized follicle in one ovary and no corpora lútea. The uterus was grossly distended; the endometrium was attenuated, with sparse glandular tissue. Normal follicles and oocytes were identified in the ovaries of both animals. In both sheep the caudal part of the vagina merged into the base of the penis, and at this point histologically recognizable bulbo-urethral glands were present. Small seminal vesicles were also located on the lateral vaginal walls of both sheep.

Two D50–100 ewes were killed during the summer anoestrous period of 1975. The ovaries of both animals were small and inactive, although follicles and oocytes were present. The uteri were histologically typical of anoestrus (McKenzie & Terrill, 1937); neither bulbo-urethral glands nor seminal vesicles were evident.

Sheep show a distinct sexual dimorphism in their pattern of urination (Short, 1974). Whereas the female normally arches her back and crouches with her hind legs whilst voiding urine in a continuous stream, the male stands normally whilst voiding urine in pulsatile jets. Table 5 shows the urination behaviour of the androgenized female sheep. All those animals with either partial or complete masculinization of their external genitalia (D30–80, D50–100, D70–120) showed a male urination pattern, whereas the D90-140 females showed a female pattern.

There were strong indications that the effects of testosterone on the pregnant ewes were less noticeable when implants were inserted during the later stages of gestation. For example, ewes implanted on day 30 or 50 developed marked aggressive behaviour, whereas those implanted on day 70 or 90 showed none; maternal clitoral hypertrophy was also most marked in the day 30 and 50 animals, less marked in the day 70 ones, and totally absent in the day 90 group. There was also a significant decline in the maternal plasma testosterone levels at later gestational ages, although the estimated daily release rate of hormone from the implant remained constant. This strongly suggests that the clearance rate of testosterone from the maternal circulation increases with advancing gestational age. It is known that the ovine foeto-placental unit is capable of metabolizing testosterone to oestrogen (Findlay & Seamark, 1971), and urinary oestrogen excretion increases dramatically in the ewe after the 70th day of gestation (Fevre, Piton & Rombauts, 1965), suggesting an increase in placental aromatizing activity. Another factor may be the protective effect of progesterone. Diamond & Young (1963) found that progesterone protected guinea-pigs against the masculinizing action of testosterone propionate, and it is well known that there is a marked rise in the peripheral plasma progesterone levels in ewes after the 80th day of gestation (Bassett, Oxborrow, Smith & Thorburn, 1969). These facts need to be taken into account when seeking to define apparent ‘critical periods’ in foetal development during which maternally administered testosterone can exert a masculinizing effect.

Another limitation of the present study is that the experiments were carried out over two different years, and so the results could be confounded by unknown seasonal effects. For example, it could be argued that the lower blood testosterone levels in the day 70–120 and day 90–140 ewes was a year difference and not due to the duration of gestation. Fortunately, the anatomical and behavioural results obtained in both years of the present study are entirely concordant with those obtained in a previous year by Short (1974), when ewes were implanted on days 20, 40, 60 or 80 of gestation and the implants left in until term (see Fig. 2). Therefore we feel that seasonal effects can be discounted.

Short (1974) showed that ewes implanted with testosterone on day 20 or 40 of gestation gave birth to female lambs with completely masculinized external genitalia, whereas ewes implanted on day 60 or 80 produced female lambs with essentially female-type external genitalia. In the present investigation the implants were only left in the ewe for 50 days, and these two approaches complement one another (see Fig. 2); the day 30–80 lambs in the present study were completely masculinized, whereas the day 50–100, 70–120 and 90–140 animals has essentially female type genitalia. We can now begin to suggest a period of development during which the genital tubercle of the female sheep foetus is most sensitive to testosterone. If exposure begins in either periods 1 or 2 and continues into period 3 then complete masculinization takes place. Exposure during period 4 causes no masculinization. Thus the ‘critical period’ for masculinization of the genital tubercle appears to be between days 40–50 (period 2; Fig. 2). However, androgen may continue to cause partial masculinization until day 80 of gestation (period 3; Fig. 2). Failure to induce masculinization after day 80 could be due either to the refractoriness of the target tissue, or to the failure of testosterone to reach the foetus in sufficient quantities as a result of the increased maternal clearance rate.

Green & Winters (1945) considered that the gonad of the sheep foetus first began to differentiate into a testis or an ovary between days 25 and 29 of pregnancy, whereas Mauleon (1961) found that the gonads could first be sexed histologically on day 35. Attal (1969) reported that testosterone could be detected in the testes of male sheep foetuses from day 30 of pregnancy and that the external genitalia begin to differentiate on day 45. These results are concordant with our findings that the external genitalia of the female sheep foetus can be masculinized between days 40 to 50 of pregnancy.

Short (1974) showed that ewes implanted from days 20, 40, 60 or 80 of gestation until term had difficulties at lambing and many of them failed to lactate. Such problems were not encountered in the present experiment where the implants were removed before parturition. Beach & Kuehn (1970) injected bitches for 10 days during mid-pregnancy with large doses of testosterone propionate and also encountered foetal death and lactational failure. The estimated daily dose of testosterone received by all the ewes in the present study was approximately 0-1 mg/kg/day, which is at least an order of magnitude less than that used to masculinize the females of other species (see Beach & Kuehn, 1970).

Inspection of the D30–80 offspring at 10 months of age revealed that the prepuce was adhering to the galea capitis, and the penile diameters were significantly smaller than in entire rams of a similar age. This is presumably due to a difference in the degree of androgenization between normal males and androgenized females. Ashdown (1957a, b) has shown that the growth of the bull’s penis and separation of the prepuce is dependent on post-natal testosterone secretion.

The pattern of urination behaviour in sheep is sexually dimorphic, and can be influenced by testosterone between days 20–80 of gestation (Short, 1974); in the present study this behavioural pattern was not masculinized after day 90 (Table 5). Only ewes showing complete or partial masculinization of their external genitalia voided urine in a pulsatile manner, and this could be a consequence of androgen-induced development of urethral musculature. However, the different urination postures presumably reflect a more fundamental influence on behaviour.

It appears that the nervous reflexes involved in urination are directly influenced by testosterone, and that this behavioural system differs from those involved in mating behaviour, since Short (1974) showed that an androgenized ewe with a male urination pattern was nevertheless capable of showing normal female oestrous behaviour. The urination posture is also sexually dimorphic in dogs, and the bitch will show a male pattern if treated with testosterone during foetal or early post-natal life (Berg, 1944; Martins & Valle, 1948; Beach, 1975). Beach suggests that the sexual dimorphism of urination postures is predominantly determined post-natally in the dog, but we have shown that the urination posture of the sheep can be masculinized during early (day 20-80) foetal life.

In conclusion, we have shown that maternally administered testosterone is capable of affecting various aspects of development of the female foetus in different ways up until day 90. The spectrum of effects could be due to the existence of ‘critical periods’ of development of the foetus, during which different systems pass through a stage of androgen sensitivity, but the results are confused by apparent changes in the maternal rate of testosterone metabolism with advancing gestation, and changing progesterone levels. It will be necessary to measure the amount of testosterone actually reaching the foetus before the existence of ‘critical periods’ can be unequivocally proven.

The authors gratefully acknowledge the help of the Animal Breeding Research Organisation’s farm staff for the care and management of the sheep. We also thank Mr W. G. Davidson, Mrs R. Cunningham, Mr L. McKenzie and Mr D. W. Davidson for technical assistance. One of us (I.J.C.) is in receipt of a William Georgetti Scholarship.

The concentrations of testosterone and dihydrotestosterone have recently been measured in ovine foetal gonads and plasma by Pomerantz & Nalbandov (Proc. Soc. exp. Biol. N.Y. 149, 413-419, 1975). The foetal testis contains more androgen than the foetal ovary during the first half of gestation and plasma levels in male foetuses are significantly higher than in females at this time. Testicular and plasma concentrations fall during the second half of gestation and this is compatible with our observations that the critical period during which testosterone causes masculinization is during the first half of gestation.