In the guinea pig it has been demonstrated that pregnant females injected daily with testosterone propionate (T.P.) did not show behavioral masculinization nor clitoral hypertrophy to the extent seen in comparably treated nonpregnant controls (Diamond and Young, 1963). Thus, protection against androgen-induced masculinization during pregnancy was demonstrated at two levels: behavioral and morphologic. A possible mechanism for this protection was suggested by showing that physiological amounts of progesterone, when injected simultaneously with testosterone, prevented the behavioral masculinization. The somatic changes were not inhibited.

The present investigation was undertaken to determine if endogenous progesterone or substances produced either by the ovary of pregnancy or the placentae could be implicated in this androgen antagonism. A preliminary report has appeared (Diamond, 1964).


Five main experimental groups were prepared from among pregnant and non- pregnant guinea pigs of a heterogeneous (Louisville) stock. The gestation period in this stock usually is between 65 and 72 days. The groups were:

1. Females ovariectomized on day 46 of pregnancy.

2. Females ovariectomized on day 36 of pregnancy.

3. Nonpregnant females hysterectomized on day 6 of the estrus cycle; ovariectomized 60 days later.

4. Nonpregnant females given a sham hysterectomy on day 6 of the estrus cycle; 60 days later given progesterone for 50 days.

5. Nonpregnant females hysterectomized on day 66.

A number of animals which aborted following ovariectomy were continued in the experiment as a separate group.

Castration was performed with ether anesthesia using a standard dorsal flank approach. Initially, due to an anticipated high rate of abortion, day 46 of pregnancy was selected for spaying. Successful completion of pregnancy for a majority of females castrated at this late date prompted a series of ovariectomies on the thirty-sixth day of pregnancy. Both dates allowed comparison of pre- and postoperative behavior and genital condition.

Hysterectomy was performed with sodium pentobarbital anesthesia using a midventral incision and involved total uterine excision from the cervix to the utero-tubal junction (similar to the method of Rowlands and Short, 1959). The sixth day of the cycle was selected for surgery since Rowlands (1961) and Howe (1965) have demonstrated that hysterectomy on this day maintains corpora lutea in an apparently functional state for a period much longer than normal gestation, i.e., in a pseudopregnant condition. A group of animals subjected to sham hysterectomies on day 6 served as controls. The sham hysterectomy consisted of incisions and ligations similar in all respects to those of the true hysterectomy save that the mesometria and arterial supply to the uterus were not interrupted and the uterus remained in situ. The group hysterectomized on day 66 (late hysterectomy group) enabled us to test whether virilization, once started, might be ameliorated. The administration of progesterone on day 66 to animals subjected to a sham hysterectomy on day 6 likewise enabled us to see whether established masculinization could be reversed. Following castration or experimentation all ovaries were serially sectioned and checked microscopically for the presence of active corpora lutea to confirm the presence or absence of pseudopregnancy.

Testosterone propionate (Upjohn; 50 mg/ml) was injected into all animals daily. Injections were made alternately into the right or left gluteal and hamstring muscles. Pregnant animals were injected starting the seventeenth day after mating (day 18); nonpregnant females were injected starting 17 days after ovulation. Dates of mating and ovulation were determined by daily vaginal checks for sperm and vaginal membrane rupture respectively. Only females with a minimum of three consecutive 16 ± 2 day cycles were used in the study. The day of mating or ovulation was considered as day 1. Five milligrams of the androgen were given on day 18 and 1 mg was given daily thereafter until the termination of the experiment on day 116. Each female was thus injected with androgen for a period comparable to the last 50 days of pregnancy and 50 days following parturition. Progesterone, given to the sham hysterectomized animals, was administered while the animals continued to receive their daily injections of T.P. Twenty-five milligram of progesterone (Upjohn; 50 mg/ml) were given on day 66 and 5 mg were given daily thereafter.

As an index of neural masculinization, behavior tests were given to each animal at 10-day intervals from day 26 to 116 with an additional test on day 71. These 11 tests allowed for observation prior to and subsequent to each of the experimental procedures. The quantitative procedures used have previously been described (Diamond and Young, 1963). Essentially, the incidence of mounting of a stimulus female by the test female provided the behavioral index of masculinization. This behavior is not normally seen in an anestrous or non-androgen-injected female. As an index of morphological virilization the genitalia of all animals were observed daily and graded in accordance with a 0 to 5 standardized scale of clitoral hypertrophy.

Throughout the investigation the animals were kept in an air-conditioned laboratory with the temperature maintained between 68° and 75°F. The diet included guinea pig chow and oats and water ad lib with supplementation of cabbage or hay.

Statistical evaluation was with chi square and Mann-Whitney U tests as outlined by Siegel (1956).


Pregnant females, even though ovariectomized, and pseudopregnant (hysterectomized) females did not respond to the neural masculinizing action of testosterone propionate to the extent demonstrated by the nonpregnant, aborting or sham hysterectomized animals which do respond. This difference is evident from the behavioral measurements of male-like mounting (Table 1, Figs. 1-4).


Mounting behavior in androgen-treated female guinea pigs during the pregnancy period
  N Mean
no. of
No. of
No. of
tests in
     ovariectomy day 36
4 0 0 0
     ovariectomy day 46
7 0.20 4 6
     day 6
13 1.23 5 13
Sham hysterectomy:
     day 6
9 4.31 9 21
     day 66
5 5.60 5 16


Fig. 1: Pseudopregnant - ovariectomized. All receiving testosterone propionate from day 18 until day 116.

  Fig 2: Pregnant - ovariectomized. All receiving testosterone propionate from day 18 until day 116.

Fig 3: Sham hysterectomy. All receiving testosterone propionate from day 18 until day 116.   Fig 4: Late hysterectomy. All receiving testosterone propionate from day 18 until day 116.


Of 11 pregnant and ovariectomized females, male-like mounting was demonstrated only by four animals in six of 55 tests. Of 13 pseudopregnant (hysterectomized) females, mounting was seen only in five animals in 13 of 65 tests. In significant contrast (P <0.001), nine non-pseudopregnant (sham-hysterectomized) females and five nonpregnant (late hysterectomized) females all mounted; the two groups mounted in 21 of 45 tests and in 16 of 25 tests.

Following parturition by the pregnant animals, or ovariectomy in the pseudopregnant group, antimasculinizing protection was no longer evident and females now showed definite male-like mounting (Figs. 1, 2). Additionally, pregnant animals which did not mount prior to abortion did so afterwards when injections were continued.

The administration of progesterone starting on day 66 to the androgen-treated non- pregnant females of the sham hysterectomy group led to a significant reduction in mounting (P <0.025, Fig. 3). In contrast, hysterectomies performed on day 66 on a group of nonpregnant androgen-treated females did not induce a comparable decrease in mounting behavior (Fig. 4). In fact, the mean rate of mounting in this group stayed high. Histological examination of ovaries from the hysterectomized females confirmed that functional corpora lutea were present if the surgery was performed on day 6 of the cycle but were not present if hysterectomy is performed on day 66 in already androgenized animals. On the contrary, ovaries from animals of this last mentioned group were quite cystic.

The tissues of the clitoris possessed a sensitivity to androgens distinct from that of the neural tissues. This organ rapidly and completely manifested androgen-induced hypertrophy in all groups not truly pregnant (Figs. 1, 3, 4). A maximal rating of five was seen within 10 to 20 days of androgen treatment in all members of the pseudopregnant, sham and late hysterectomy groups. The pregnant females, in contrast, usually had no or only moderate (rating of 3 or less) clitoral hypertrophy after similar androgen treatment (Fig. 2). Moreover, in the pregnant group, castration, regardless of whether done on day 36 or 46, was not followed by a significant increase in clitoral masculinization. After abortion or parturition, however, full clitoral hypertrophy appeared within 10 to 20 days.

Neither progesterone treatment nor pseudopregnancy reduced existing clitoral hypertrophy or prevented its development (Figs. 1, 3).


The finding that the hysterectomized female was not behaviorally masculinized by injected androgen until after ovariectomy is indicative of a protective mechanism associated with the ovary of a pseudopregnant female. The ovary of the pregnant female, in contrast, was found not to be indispensable for the antiandrogenic effect, although it may have contributed to it, since castration during pregnancy was not followed by masculinization. Data from the groups of pregnant females indicated that these females, even after ovariectomy on the thirty-sixth or forty-sixth day of gestation, possessed protection from androgen-induced neural masculinization as long as abortion or parturition did not occur. Thus it may be assumed that the placentae, fetuses or some associated feature of pregnancy other than the ovary was still providing antiandrogenic protection.

Rowlands and Short (1959) have shown that the corpora lutea of ovaries from comparably hysterectomized female guinea pigs contain high concentrations of progesterone for a period longer than that of normal gestation, and Heap and Deanesly (1966) have shown placentae are capable of maintaining, in ovariectomized animals, a serum level of progesterone of 2 to 5 times that of non pregnant animals. Since removal of the prime progesterone sources in these animals by ovariectomy, abortion or parturition was soon followed by the display of male-like behavior patterns, data from the present preparations may be considered to substantiate the initial expectations that endogenous progesterone is capable of inhibiting behavioral masculinization in androgen-treated female guinea pigs. Previously, administered progesterone was demonstrated to have this capability (Diamond and Young, 1963)

The clitoral tissues, on the other hand, did not seem to be protected either by endogenous or exogenous progesterone. All androgen-injected females except those pregnant exhibited full clitoral hypertrophy within three weeks of treatment. If the female was not pregnant, by day 36 she normally had developed full clitoral hypertrophy. Hall (1938) had similarly reported that clitoral tissues in the rat are not protected from testosterone-induced masculinization by estrone plus progesterone. In contrast, clitoral protection was not lost to the pregnant female even following castration and full hypertrophy was not seen in any pregnant animal during the present study. This indicates that for the somatic tissue, as for the neural tissues, in the pregnant female the ovaries are not indispensable for antiandrogenic effects and some other protective factor(s) or pregnancy are available.

Progesterone, in larger quantities, does not seem to be the protective entity involved. This may be deduced from two facts: first, progesterone levels in ovariectomized pregnant guinea pigs are generally lower than those in the intact pregnant female (Heap and Deanesly, 1966), and second, nonpregnant female guinea pigs injected daily with 20 mg of progesterone while given 2 mg of T.P. daily showed full clitoral hypertrophy within 20 days (Diamond, unpublished data). An additional pilot study indicated that simultaneous administration of estradiol benzoate or estradiol dipropionate with the progesterone also does not effect clitoral masculinization. It may be significant, however, that Neumann, Richter and Gunzel (1965) with prenatal administration of a synthetic progesterone, 1, 2α-methylene-6-chloro-Δ6-17α-OH-progesterone-acetate were able to prevent normal androgen induced differentiation of male genitalia in the rat.

Thus, as reported previously (Diamond and Young, 1963), the genital tissues were seen quantitatively and qualitatively to exhibit a different response pattern than the neural tissues to androgen-induced masculinization. While the pregnant, pseudo-pregnant and progesterone-injected groups were refractory to androgen-induced behavioral masculinization, only the pregnant group failed to develop full somatic masculinization. In those groups displaying both behavioral and genital changes the clitoral hypertrophy developed first and maximally while the behavioral maxima occurred later. When virilization was already established (in the sham hysterectomy group), progesterone was seen capable of reducing the behavioral effects but had no effect on reducing the genital hypertrophy. Pseudopregnancy did not prevent the induction of clitoral masculinization as did true pregnancy.

The nature of the antimasculinizing protective action of pregnancy in regard to the neural tissues and behavior thus seems to be progesterone dependent; the mechanism protecting the somatic tissues of the genitalia is still unknown and awaits elucidation.

With the clear delineation of this normal anti-androgenic feature of pregnancy, it may be speculated that this phenomenon has adaptive significance in protecting the nervous system of the mother and the female fetuses she carries from androgenic influences from male fetuses. Recently, Cséffalvay and Klose (1966) observed clitoral hypertrophy during pregnancy in 47.5% of 434 women bearing male fetuses. A comparable behavioral effect of androgens on the mother would be temporary but the effect on a female fetus, guinea pig or human, could be permanent (see discussion by Diamond, 1965 and Diamond, in press). The fact that the clitoral tissues are less protected from androgen than the neural tissues suggests that alteration of this sexual structure is of less importance in evolution than alteration of the sexual behavior.


Thanks are due Mr. Walter Strybel who, under the University of Louisville School of Medicine Summer Scholarship Program, assisted with various surgical and testing portions of this work, and to Mrs. Susan Wheeler, who assisted with the histological preparations.

The testosterone propionate and progesterone used in the study were generously supplied by Dr. S.S. Stubbs and the Upjohn Company, Kalamazoo, Michigan.

This investigation was supported by Public Health Service Research grants GM-10632 and HD-02326 from the National Institutes of Health.



Cséffalvay, Von T., and S. Klose, 1966. Durch fetale Androgene bedingte mütterliche Virilisation. Zeit. f. Geburt. Gynäkol., 165: 44-49.

Diamond, M., 1964. Pseudopregnancy, ovariectomy and induced masculinization. Anat. Rec., 148: 276.

Diamond, M., 1965. Discussion in “Sex and Behavior,” Frank A. Beach (ed.). p. 101. John Wiley & Sons, Inc., New York.

Diamond, M., In press. Genetic-endocrine interactions and Human Sexual Behavior. In: Progress in Human Behavior Genetics. Steven C. Vandenberg (ed).

Diamond, M., and W.C. Young, 1963. Differential responsiveness of pregnant and non-pregnant guinea pigs to the masculinizing action of testosterone propionate. Endocrinology, 72: 429-438.

Hall, K., 1938. The clitoris of the rat after ovariectomy and the injection of sex hormones. J. Path. Bact., 47: 19-26.

Heap, R.B., and R. Deanesly, 1966. Progesterone in systemic blood and placentae of intact and ovariectomized pregnant guinea-pigs. Endocrinology, 34: 417-423.

Howe, G.R., 1965. Influence of the uterus upon cyclic ovarian activity in the guinea pig. Endocrinology, 77: 412-414.

Neumann, F., K.-D. Richter and P. Günzel, 1965. Wirkungen von Antiandrogenen. Zent. f. Veterinärmedizen, 12: 171-188.

Rowlands, I.W., 1961. Effect of hysterectomy at different stages in the life cycle of the corpus luteum in the guinea-pig. J. Reprod. Fertil., 2: 341-350.

Rowlands, I.W., and R.V. Short, 1959. The progesterone content of the guinea-pig corpus luteum during the reproductive cycle and after hysterectomy. J. Endocrinol., 19: 81-86.

Siegel, S., 1956. Nonparametric statistics for the behavioral sciences. McGraw Hill, New York.

Back to top