Behavioral demasculinization of female quail is induced by estrogens: Studies with the new aromatase inhibitor, R76713

HORMONES

AND

BEHAVIOR

26,

179-203 (1992)

Behavioral Demasculinization of Female Quail Is Induced by Estrogens: Studies with the New Aromatase Inhibitor, R76713 J. BALTHAZART,

A. DE CLERCK, AND A. FOIDART

Laboratory of General and Comparative Biochemistry, University of Likge (Bat. LI), 17 place Delcour, B-4020 Li$ge, Belgium The injection before Day 12 of incubation of estradiol benzoate (EB) into Japanese quail eggs produces a complete behavioral demasculinization of adult males that will hatch from these eggs. These males never show copulatory behavior even after administration of high levels of exogenous testosterone (T). It is usually assumed that such a demasculinization normally takes place in female embryos under the influence of endogenous estrogens but few experimental data are available to confirm the validity of this model. A series of four experiments was performed during which R76713, a triazole derivative that specifically inhibits aromatase (estrogen synthetase) activity, was injected into quail eggs at different stages of incubation to prevent the production of endogenous estrogens. The consequences of these embryonic treatments on the T-activated sexual behavior in adults were then quantified. When injected before Day 12 of incubation, R76713 completely blocked the behavioral demasculinization of females without affecting the behavior of the males. After a treatment with T, almost all R76713-treated females showed as adults a masculine copulatory behavior that was undistinguishable from the behavior of intact males. This effect was fully reversed by the injection in egg of EB demonstrating that the effects of R76713 were specifically due to the suppression of endogenous estrogens. Injection of R76713 during the late phase of the incubation (Day 12 or Day 15) only maintained weak copulatory behavior in females which confirmed that the behavioral demasculinization in quail takes place mainly though not exclusively during the early stages of ontogeny, In a last experiment, we combined an early R76713 treatment with an injection of EB either on Day 9 or on Day 14 of incubation. This showed that the sensitivity to differentiating effects of estrogens varies with age in a sexually differentiated manner. The EB injection on Day 9 demasculinized both male and female embryos. If this injection was delayed until Day 14, it was no longer effective in males but still caused a partial demasculinization of females. This demonstrates that even if females are not yet behaviorally demasculinized on Day 9 of incubation (suppression of aromatase activity at that age will maintain the behavior), their sensitivity to estrogens is already different from that of males. 0 1992 Academic Press. Inc.

In birds, like in mammals, reproductive behavior is sexually differentiated. The nature of this sex dimorphism is however different in the two 179 0018-506X/92

$4.00

Copyright 0 1992 by Academic Press. Inc. All rights of reproduction in any form reserved.

180

BALTHAZART,

DE CLERCK,

AND FOIDART

vertebrate groups. In mammals, it is mainly the capacity to display female sexual behaviors that is differentiated (defeminization of males) but the masculine behavior is also somehow more intense in males than in females exposed to a similar hormonal environment (females are demasculinized) (Gay and McEwen, 1980). In birds, it is by contrast the male sexual behavior which demonstrates a robust sex dimorphism while little or no dimorphism is observed in the female-typical behaviors. Treatment with testosterone (T) elicits masculine copulatory behavior in male quail or chicken but has almost no effect in females (Adkins, 1978; Adkins-Regan, 1983; Balthazart, 1989; Wilson and Glick, 1970). Similarly T induces singing in male but not in female zebra finches (Gurney and Konishi, 1980). It is clearly established that these behavioral dimorphisms are not the direct consequence of genetic differences between the sexes but, rather, result from the early exposure to a different hormonal milieu. In quail and chicken, it is largely admitted that the absence of male-type sexual behavior in adult females results from their early exposure to estrogens (Adkins, 1978; Adkins-Regan, Pickett, and Koutnik, 1982; Adkins-Regan, 1983). Females are demasculinized by embryonic estrogens in galliforms. A more complex situation would be present in song birds, in which the perinatal estrogens would demasculinize copulatory behavior (AdkinsRegan and Ascenzi, 1987; Adkins-Regan, 1990) but also masculinize singing (Gurney and Konishi, 1980; Adkins-Regan and Ascenzi, 1987). How a same hormone could enhance some masculine behavioral characteristics and, at the same time, decrease others is unclear at present. It has been argued that these differentiation processes could take place at different ages but the comparison of the behavioral changes obtained by hormonal manipulations of neonates (Adkins-Regan and Ascenzi, 1987; Gurney and Konishi, 1980; Adkins-Regan and Ascenzi, 1990; Pohl-Ape1 and Sossinka, 1984) with the plasma levels of steroids measured during ontogeny (Hutchison, Wingfield, and Hutchison, 1984; Adkins-Regan, Abdelnabi, Mobarak, and Ottinger, 1990) does not clearly support such an interpretation (Adkins-Regan, 1990). It is important to notice that the models of avian sexual differentiation which are outlined here (demasculinization of copulatory behavior in females and masculinization of singing in males by embryonic/neonatal estrogens) are based only on indirect evidence. In quail, it has been shown very convincingly that injections of estrogens in male embryos demasculinize their copulatory behavior (Adkins, 1975,1979; Schumacher, Hendrick, and Balthazart, 1989) but this effect might well be of a pharmacological nature. There is no or little direct proof that the same mechanism is responsible for the demasculinization of females under physiological conditions. One study only has addressed this question. Adkins injected female embryos with the antiestrogen CI-628 and showed that this em-

SEXUAL

DIFFERENTIATION

OF QUAIL

181

bryonic treatment increased the proportion of adult females mounting a stimulus animal after treatment with T (Adkins, 1976). Presumably, the demasculinization of behavior by endogenous (ovarian?) steroids had been blocked by the antiestrogen. However, these treated females were partly demasculinized since only three out of nine experimental females showed the full copulatory sequence including the cloaca1 contact movements. in addition, the intensity of the behavior exhibited by these CI-62%treated females (frequency and latency to first behavior) was also clearly different from that observed in normal males. Therefore, these data were in agreement with the notion that estrogens demasculinize female quail during ontogeny but could not be taken as conclusive evidence that this was the only process implicated in the sexual differentiation of reproductive behavior in quail. In addition, attempts in our laboratory to reproduce this effect using either CI-628 or another antiestrogen, tamoxifen, w.ere unsuccessful: low doses of these compounds never produced a significant behavioral effect while the higher doses killed embryos in the egg (Schumacher and Balthazart, unpublished data). The interpretation of the antiestrogens’ effects in avian embryos is further complicated by results obtained in zebra finches. In this species, treatment of young females with three different antiestrogens resulted in a hypermasculinization of brain nuclei controlling vocalizations (Mathews, Brenowitz, and Arnold, 1988; Mathews and Arnold, 1990) while it had been shown previously that these same nuclei were masculinized by exogenous estrogens (Gurney and Konishi, 1980; Gurney, 1981). These results therefore suggested one of two conclusions: either the three compounds tested (tamoxifen, LY1170018, and CI-628) were good antiestrogens and the model of sexual differentiation of the zebra finch brain was incorrect or the model was essentially correct but the antiestrogens did not act as such in this species (Mathews and Arnold, 1990). Whatever the conclusion in zebra finches, these data clearly throw doubts on the current model of sexual differentiation in quail that is primarily based on one single experiment using one of these suspicious compounds (U-628). Other models might be contemplated (Schumacher and Balthazart, 1985). The demascuiinization of male quail by exogenous estrogens appears to be restricted to a sensitive period of the embryonic life. It has been shown that EB injections into male embryos demasculinize copulatory behavior only if they are performed before the 12th day of incubation (Adkins, 1979; Adkins-Regan, 19$3; Schumacher et al., 1989). Delayed injections were without effect. This has been used .as evidence suggesting that, in femaIes, the behavioral demasculinization is also completed before incubation Day 12 under the influence of endagenous, presumably ovarian, estrogens. Yet experiments in females suggest that their behavioral demasculinization is not completed even at hatching (Day 17 of incubation). Neonataliy ovariectomized females show a weak masculine sexual

182

BALTHAZART,

DE CLERCK,

AND FOIDART

behavior in response to T treatment in adulthood (Hutchison, 1978; Schumacher and Balthazart, 1984). This behavior is seen only in a fraction of the birds and exclusively in optimal testing conditions (home cage test), but since it is completely absent in females ovariectomized later in life, this indicates that the demasculinization of females is in progress but nor fully completed at the time of hatching. Ovarian estrogens, apparently, complete the behavioral demasculinization of these females during the first 4 weeks post-hatch (Balthazart and Schumacher, 1984). There is therefore an apparent discrepancy between results showing that male quail are demasculinized by exogenous estrogens only if the treatment is applied before Day 12 of incubation and results showing that the demasculinization of females is not completed at hatching. This suggests that the time course of the differentiation process in females might be different from the expectation based on the results of hormonal manipulations of males. However, the study of this question has been difficult so far since there was no easy way to interfere with the action of estrogen in female embryos. A new very potent aromatase inhibitor (Wouters et al., 1989; De Coster et al., 1990) with great specificity was recently tested in adult quail. The triazole derivative, R76713 (6-[(4-chlorophenyl)(lH-12,4-triazol-1-yl) methyl]-1-methyl-lH-benzotriazole), was able to block the T-induced copulatory behavior and preoptic aromatase activity in castrated male quail (Balthazart, Evrard, and Surlemont, 1990) thereby demonstrating its great efficacy in this avian model. Aromatase inhibitors represent an alternative to antiestrogens for the physiological study of estrogen-dependent events: instead of blocking estrogen action at the receptor level, their synthesis is suppressed in the gonads and in other potential sources of steroid. This approach was applied here to the study of the sexual differentiation of quail copulatory behavior. If masculine sexual behavior is lost in females during embryonic life under the influence of estrogens (as suggested by the effect of exogenous estrogens in male embryos), it should be possible to maintain this behavior by blocking estrogen synthesis with a potent aromatase inhibitor such as R76713. We demonstrate here in a series of four experiments that this is indeed the case which strongly supports the model of differentiation according to which the behavior of quail differentiates by a demasculinization of females under the influence of estrogens. This new aromatase inhibitor was then used to determine the sensitive period during which estrogens act on the female brain to produce the sexually dimorphic circuitry that mediates masculine behavior. MATERIAL AND METHODS Animals and Hormonal Treatments Experiments were performed on male and female Japanese quail (Corurnix coturnix japonica). Eggs were purchased from local breeders (Fr.

SEXUAL

DIFFERENTIATION

OF QUAIL

183

Lefevre, Boneffe, Belgium or C. Dujardin, Liernu, Belgium), set in the incubator (38°C and 50-60% of relative humidity) in the morning of Day 0 and returned twice every day throughout the incubation period. Chicks hatched on Day 17. During these experiments, eggs were injected with estradiol benzoate (EB; Sigma E-9000; 25 pg/egg) and/or with the aromatase inhibitor, R76713 (6-[(4-chlorophenyl)(W-1,2,4-triazol-lyl)methyl]-1-methyl-1H-benzotriazole; Janssen Pharmaceutics, Beerse, Belgium) at the dose of 10 or 50 pg/egg. EB was always dissolved in 50 yl of sesame oil (Henry Lamotte, Bremen). R76713 was dissolved at a dose of 500 pg/ml in saline (0.9%) containing 20% of polyethyleneglycol (PEG400; UCB7908, Leuven, Belgium). Control birds were always injected with the corresponding volumes of solvent (50 ~1 oil or 20-100 ~1 saline/PEG). Specific protocols for each experiment are given below. Injections were administered with a 25-gauge needle inserted approximately 5 mm into the albumen of the egg (small end, opposite to the air chamber). Before each injection, needles were sterilized in a flame and needle holes in the shells were sealed with melted paraffin. Solutions were always sterilized prior to injection by heating them in boiling water (EB) or passing them through a 0.45~pm filter (R76713). One day after hatching, all remaining eggs were opened to figure out the reasons why they did not produce a young bird. By visual inspection, it was then possible to discriminate infertile eggs from eggs containing embryos that died before or after the injections. In the following, infertile eggs have been excluded from the analysis and all data on numbers of injected eggs and on hatching success only concern fertile eggs. Chicks were raised in heterosexual groups at 38-40°C for the first 2 weeks of life and at 28-30°C for the next 2 weeks. Throughout the experiments, birds were kept under long days (16L:8D, lights on at 6.00 a.m.). Water and food were available ad libimn. In all experiments, birds were castrated at the age of 3-4 weeks under total anesthesia (Hypnodil, Janssen Pharmaceutics, Belgium; 15 mg/kg). Both testes were removed through a unilateral incision on the left side. Only the left ovary of females was taken away. The right one is not developed and does not regenerate even after removal of the left gonad (Gibson, Follett, and Gledhill, 1975). At the age of 6-7 weeks, birds were isolated in individual cages and implanted with 40 mm (2 x 20 mm) Silastic implants filled with testosterone (Sigma T-1500) (Dow Corning Silastic Tubing 602-265; 1.57 mm id., 2.41 mm o.d.). Implants were always cleaned and preincubated for at least 12 hr in a 0.9% saline solution at 41°C to initiate steroid diffusion through the tube wall. Birds were sacrificed at the completion of the behavioral tests (see below), 2-3 weeks after the beginning of T treatment. At that time, completeness of castration and presence of the T implants were checked.

184

BALTHAZART,

DE

CLERCK,

AND

FOIDART

Behavioral Tests The copulatory behavior of the experimental subjects was observed in two different test situations (Schumacher and Balthazart, 1984). Birds were tested with a receptive female either in their home cage (19 x 30 x 17 cm height) or in a test arena (30 x 80 x 60 cm height). In both situations, a stimulus female was introduced for 5 min and the frequencies and latencies for the following behavior patterns were recorded: strut, neck-grab (NG), mount attempt (MA; counted only when a bird which is showing a neck-grab raises one leg and puts it over the back of the test female), mount (M) and cloaca1 contact movements (CCM) (see Adkins and Adler, 1972; Hutchison, 1978 for a detailed description). Birds were tested once or twice in each condition (see specific protocols). We demonstrated previously that weakly active birds are more likely to show sexual behavior in the home cage situation by comparison with the test arena (Schumacher and Balthazart, 1984; Schumacher et al., 1989) which justifies the use of both procedures in the present experiments. Morphology Birds were weighed to the nearest gram on several occasions: after hatching, just before the start of T treatment, and before sacrifice. At sacrifice, their cloaca1 gland, an androgen-dependent (Sachs, 1967) sexually differentiated (Adkins, 1975; Adkins and Adler, 1972; Balthazart, Schumacher, and Ottinger, 1983) structure, was also measured with a caliper [greatest length x greatest width = cloaca1 gland area (CGA)]. Statistics Morphological data and behavioral frequencies were analyzed by analysis of variance (ANOVA) followed when appropriate by the Fisher protected least significant difference (PLSD) test to compare individual means. Percentages of birds showing a given behavior and differences between hatching rate according to experimental treatments were analyzed by nonparametric statistics (x2 and Fisher’s exact probability test). Effects were considered significant for P S .05. Probabilities reported in this paper are two-tailed. RESULTS Experiment

1

In a first experiment, the effects of two different doses of R76713 injected in quail embryos on Day 8 of incubation were tested. Methods A total of 131 fertile eggs were incubated and distributed in three experimental groups that were injected with 10 pg (group RlO; N = 40)

SEXUAL

DIFFERENTIATION

OF QUAIL

185

or 50 pg (group R50; N = 43) R76713/egg or with the solvent (20% PEG in saline) as control (group C; N = 48). One half of the control group was injected with 20 ~1, the other half received 100 ~1 of the solvent. Injections were performed in the afternoon of the 8th incubation day (eggs set in the incubator = Day 0). All birds were gonadectomized at 4 weeks post-hatch and implanted with T silastic capsules at 6 weeks. They were tested twice in the arena and twice in their home cage during the 3rd week following the placement of the implants. Given the usual hatching rate in quail (60-80%) and the mortality normally observed during early postnatal life, the number of subjects observed in the behavior tests was 13 C, 6 RlO, and 8 R50 males, and 15 C, 17 RlO, and 10 R50 females. Results The hatching rate was significantly affected by the experimental treatments (RlO: 28/40, R50: 21/43, C: 38/48; x2 = 9.686, P = .0079). This was essentially due to a reduction of hatching rate in the R50 group. The group treated with the lowest dose of aromatase inhibitor hatched with the same success as the controls (x” = 0.978, P = .327). The body weight of the chicks at hatching was similar in the three groups. It tended to be slightly higher in birds that had been injected with the aromatase inhibitor (RlO: 8.6 + 0.8, R50: 8.7 ? 0.9, C: 8.3 & 0.7 g, mean -+ SD) but the difference was far from being significant (Fzvg6 = 1.502, P = .lll). In the following presentation, the data corresponding to the two subgroups of control birds (injected with 20 or 100 ~1 of solvent) are always pooled for the sake of clarity. This is justified by the fact that little or no effect of the amount of solvent could be observed by twoway ANOVA (sex of the birds and dose of solvent) in separate analyses of the body weight, cloaca1 gland area, and frequency of MA or CCM in the arena for control birds alone (P > .05 in each case). The frequency of MA and CCM in the home cage situation was however lower in the birds that had received the largest volume of solvent (MA: F,,24 = 4.545. P = -043, CCM: Fl,24 = 7.461, P = ,016 in the two-way ANOVA on control birds with the sex and dose of solvent as factors) but this did not justify a separate analysis of the data for these birds especially since the behavioral frequencies observed in the arena are the basis of the presentation (see below). Body weights measured at the end of the experiment were analyzed by a two-way ANOVA with the sex of the birds and their embryonic treatment as factors. This confirmed the larger body weight of females by comparison with males [females: 219.1 ? 26.9, males 207.1 +- 16.4 g (mean t SD); Fl,63 = 5.109, P = .027] but there was no effect of the

186

BALTHAZART,

DE CLERCK,

AND FOIDART

Experiment 1 q

PC

R10

Experiment 2 lil

RM

*

Males

*

Females

Males

Experiment 3 WC

q

R76

0

EB

Experiment 4 Ei

n

RWEB

100

400

300

300

200

200

100

Males

Females

SEX

Females

100

R7Mil.W q

0 R7fdEB.W

Males

R7610iI.Dl3 q R76iEB.Dl.3

Females

SEX

1. Effects of embryonic treatments with the aromatase inhibitor R76713 or with estradiol benzoate (EB) on the sexual differentiation of the cloaca1 gland. Data presented are the means r SEM of the cloaca1 gland areas measured at the end of the experiments when gonadectomized birds had been treated for approximately 2-3 weeks with Silastic implants filled with testosterone. Data were analyzed by ANOVA followed when appropriate by Fisher PLSD tests comparing groups two by two (see text for detail). Results of these tests are reported at the top of the corresponding bars as follows: *, P < .05 for the comparison with the control birds of the same sex; #, P < .05 for the comparison with males submitted to the same treatment; A, P < .05 for the comparison with females treated with R76713 on Day 6 (D6) in Experiment 2, with birds of the same sex treated with R76713 alone (R76) in Experiment 3, and with birds treated with EB on Day 9 (R76/EB.D9) in Experiment 4. The same notations are also used in Figs. 2-5 where more information can be found. FIG.

embryonic treatments (F,,, = 2.352, P = .103) and no interaction between these two factors (F2,63 = 1.235, P = .297). In contrast, the cloaca1 gland area at the end of the experiment was affected by the embryonic treatments (F2,63= 6.081, P = .004) and the sex of the birds (Fl,63 = 4.938, P = .029). The treatment effect was extremely different in males and females (interaction: F2,63= 10.371, P < .OOOl). CGA are normally smaller in females than in males even after treatment with a same dose of T (Adkins, 1975; Adkins and Adler, 1972; Balthazart et al., 1983). This difference was confirmed here but was completed abolished by the embryonic treatment with R76713 which blocked the demasculinization of the gland in females while it had no effect in males (see Fig. 1 for the detail of statistical comparisons by the Fisher PLSD test).

SEXUAL

DIFFERENTIATION

Mount Attempt

187

OF QUAIL

Cloacal Contact Movement 100 50

0

Males

Females

Males

Females

SEX

”

Males

Females

Males

Females

SEX

2. Effects of the aromatase inhibitor R76713 injected on Day 8 of incubation on the sexual differentiation of masculine sexual behavior in quaif. Each egg was injected with 10 (RlO) or 50 (R50) ,ug R76713 or with the vehicle as control (C). Data presented are the total percentage of birds that performed at least one mount attempt or one cloaca1 contact movement during the behavioral tests performed in the arena (top graphs) and the mean frequency (*SEM) per test of these behaviors (bottom graphs). The percentages of active birds were analyzed by the Fisher exact probability test, and the frequency data were analyzed by ANOVA followed when appropriate by Fisher PLSD tests comparing groups two by two. Results of group comparisons are reported at the top of the corresponding bars as foHows: * ) P < .OS for the comparison with the control birds of the same sex; #, P i .O5 for the comparison with males submitted to the same treatment. FIG.

The same striking effects were observed during the behavioral tests whose results are summarized in Fig. 2. Intact females almost never show copulatory behavior even after a treatment with T that fully activates the behavior in castrated males. This was confirmed here: females in the control group (N = 15) never showed MA or CCM when tested in the arena. Treatment with R76713 on Day 8 of incubation completely suppressed this behavioral sex dimorphism: more than 90% of the injected females (16/17 in the RlQ group, lo/IO in the R50 group) showed the full copulatory sequence including MA and CCM when presented in the arena to a sexually receptive egg-laying female. This effect was obviously highly significant (see Fig. 2 for the detailed results of the Fisher Exact probability tests). In addition, the behavioral frequencies observed in the R76713-treated females were similar to those observed in intact males (see Fig. 2, bottom part). The aromatase inhibitor had, by contrast, no significant effect on the sexual behavior of the males (percentage of active birds or behavior frequencies).

188

BALTHAZART,

DE CLERCK,

AND FOIDART

The analysis of the behavioral frequencies by two-way ANOVA confirmed the significant effects of the sex, embryonic treatment, and/or interaction between these factors (MA: sex, Fl,63 = 7.899, P = .007, treatment: F 2.63 = 9.204, P = .OOl, interaction: F2,63 = 11.633, P = .OOl; CCM: sex, 4,63 = 3.533, P = .065, treatment: FzTe3= 3.847, P = .026, interaction: F2,63 = 10.467, P = .OOl). As mentioned above, sexual behavior is more frequently observed when birds are tested in their home cage by comparison with the arena (Schumacher and Balthazart, 1984). During the present experiment, all subjects were also tested twice in this other situation. This confirmed the increased incidence of sexual behavior in birds that were poorly active in the arena (8 C females out of 15 showed MA and 2 showed CCM in their cage compared to 0 for MA and CCM in the arena). The home cage situation could not affect the behavior in the other groups that were already fully active in the arena. Despite this difference in the behavior of the control females, the statistical analysis of the behavioral frequencies observed in the home cage essentially led to the same conclusion as obtained for the tests in the arena. The frequency of both MA and CCM was influenced by the sex of the bird, their treatment and/or the interaction of the two factors (MA: sex, FI,63 = 11.642, P = .oOl, treatment: F2,63 = 4.737, P = .012, interaction: F2.63 = 5.434, P = .007; CCM: sex, Fl,63 = 10.183, P = .002, treatment: F2.63 = 1.630, P = .204, interaction: F2,63 = 3.558, P = .034; data not shown). Experiment 2

Since the first experiment had revealed that R76713 injected on Day 8 of egg incubation prevented female demasculinization, a second experiment was carried out to determine whether this effect was limited in time. This was done by testing the behavior of birds that had been injected at different embryonic ages with the aromatase inhibitor. Methods A total of 246 fertile eggs were randomly distributed into six experimental groups and set in the incubator. They were then injected with R76713 (10 ,ug in 20 ~1 of PEG400-saline) on Day 6 (N = 49), Day 9 (N = 45), Day 12 (N = 44), or Day 15 (N = 46) of incubation. The remaining eggs were injected with 20 ,ul of the solvent on Day 6 (N = 35) or 9 (N = 27) of incubation. All birds were gonadectomized on Days 24-25 post-hatch and implanted with T (40-mm Silastic capsules) when 7 weeks old. They were then tested for masculine copulatory behavior twice in the arena and twice in their home cage during an S-day period starting 12 days after the beginning of the T treatment. Final numbers of subjects at the end of the experiment were 10 males and 10 females in the Day 6-R76713 (D6) group, 9 males and 4 females in the Day 9-

SEXUAL

DIFFERENTIATION

OF QUAIL

189

R76713 (D9) group, 13 males and 5 females in the Day 12-R76713 (D12j group, 14 males and 11 females in the Day 15-R76713 (D15) group, 4 males and 3 females in the control Day 6 (Ctrl-D6) group, and only 2 males in the control Day 9 (Ctrl-D9) group.

Results The hatching rate differed significantly among the 6 groups (D6: 29/49, D9: 20/45, D12: 23/44, D15: 29/46, Ctrl-D6: 11/35, Ctrl-D9: 5/27, xz = 31.106, P = .OOOS). The difference between groups was mainly due to a lower hatching success in the control eggs injected on Day 9 and to some extent on Day 6. This cannot therefore be related to the treatment with the aromatase inhibitor. When these two groups were removed from the analysis, hatching success was homogeneous between the 4 remaining groups (x’ = 1.112, P = .774). The simple exclusion of the controls injected on Day 9 also made the hatching rate similar in the remaining groups (x” = 9.389, P = .052). Given the small number of birds available in the Ctrl-D9 group (2 males and no females at the adult age), these are ignored in the subsequent presentation and birds injected with solvent on Day 6 are simply called controls (Ctrl). Body weight at hatching differed significantly from one group to the other (F,,,,, = 7.442, P < .OOl). Fisher PLSD tests showed that this resulted essentially from a lower weight in the D12 and D15 groups (they are significantly different from the other groups for P = .05). The magnitude of these differences is however small (D6: 9.89 k 1.10, D9: 9.74 t 1.09, D12: 8.82 2 0.77, D15: 8.69 i- 1.03, Ctrl: 9.90 5 1.04 g; mean + SD) and cannot receive an obvious interpretation. The two-way ANOVAs performed of the adult body weight (before and after T treatment) confirmed the significantly larger size of the females compared to males (effect of sex: before T: F,,73 = 9.252, P = .033, after T: F1,73 = 12.057, P < .OOl) but failed to detect an effect of the embryonic manipulations (before T: Fd,73= 1.46, P = .223, after T: F4,73= 0.118, P = .976) or an interaction of these treatments with the sex of the birds (before T: FJ,73= 0.682, P = .606, after T: FJ.73 = 0.673, P = .613). Adult females were 7 to 9% heavier than males. The cloaca1 gland areas measured at the end of the experiment were significantly affected by the sex of the birds (FI,r3 = 20.200, P < .OOl), their embryonic treatment (Fe,73= 6.438, P < .OOl), and the interaction between these factors (F,,, = 3.493, P = .Oll). As shown in Fig. 1, the CGA was similar in all groups of males but was significantly reduced in the females that had not been injected with the aromatase iiihibitor before Day 9 of incubation. The behavior of these birds was affected in the same way. Data concerning the tests performed in the arena are summarized in Fig. 3. As in the previous experiment, females that had been injected early

BALTHAZART, Mount

Males

DE CLERCK,

AND FOIDART Cloaca1 Contact

Attempt

Females

Males

Movement

Females 0 CM

0

Males

Females

Males

Females

SEX SEX 3. Effects of the aromatase inhibitor R76713 injected at different ages of the embryonic life on the sexual differentiation of masculine sexual behavior in quail. Each egg was injected with 10 /lg R76713 on Days 6 (D6), 9 (D9), 12 (D12), or 15 (D1.5) of incubation or with the control vehicle solution (Ctrl) on Day 6. Data presented are the total percentage of birds that performed at least one mount attempt or one cloaca1 contact movement during the behavioral tests performed in the arena (top graphs) and the mean frequency (2 SEM) per test of these behaviors (bottom graphs). The percentages of active birds were analyzed by the Fisher exact probability test, and the frequency data were analyzed by ANOVA followed when appropriate by Fisher PLSD test comparing groups two by two. Results of group comparisons are reported at the top of the corresponding bars as follows: *, P < .05 for the comparison with the control birds of the same sex; #, P < .05 for the comparison with males submitted to the same treatment. FIG.

during the incubation with the aromatase inhibitor, R76713, usually showed masculine sexual behavior (MA and CCM) when presented with a stimulus female. A weak activity (40% showing MA but no CCM) was still observed in females treated on Day 12. Treatment on Day 15 was apparently too late: these females showed no masculine behavior during the tests, just like the control females. The experimental treatments did not affect the percentage of male birds showing sexual behavior (see Fig. 3 for the detail of statistical comparisons). The analysis of the behavioral frequencies by two-way ANOVA (sex and treatment) confirmed these effects. Significant effects of the sex and of the interaction sex by treatment were observed for both MA (Sex: F1,73 = 30.128, P < .OOl; Interaction: F3,73= 7.704, P < .OOi) and CCM (Sex: F,,73 = 18.463, P < .OOl; Interaction: Fa,,, = 4.252, P = .004). No overall effect of the treatments was detected (F4,73 = 2.130, P = .086 and FJ,,~ = 1.231, P = .305 for MA and CCM, respectively). Fisher

SEXUAL

DIFFERENTIATION

OF QUAIL

191

PLSD tests were used to determine the origin of these effects. Females treated with R76713 on Day 12 or Day 15 of incubation were essentially similar to control females while those which received the treatment on Day 6 behaved exactly like males. The behavior of females treated on Day 9 was intermediate. No effect of the treatments could be detected in males except for a small decrease in CCM frequency observed in the D9 group (see Fig. 3 for the detail of statistical comparisons). The analysis of the behavioral data obtained during the tests in the home cage produced similar results. It is well established that weakly active birds are more likely to display masculine sexual behavior in this test situation (Schumacher and Balthazart, 1984). This was again confirmed here and the difference between the two test conditions was especially visible in the females for which the treatment with R76713 had been marginally active (injection on Day 12). Only 40% of them (2!5) showed MA in the arena test but 80% displayed this behavior in their home cage. None of the D12 females showed CCM in the arena but 4 out of 5 showed this behavior in the home cage. In this group of females, the frequency of both MA and CCM was also higher in the home cage situation than in the arena (data not shown). Experiment 3 Experiments 1 and 2 had shown that early treatment of female embryos with the aromatase inhibitor R76713 prevented their behavioral and morphological (cloaca1 gland area) demasculinization. The next experiment was performed to confirm that these effects were actually due to the removal of endogenous estrogens. These were therefore replaced in. R76713-treated birds and behavioral effects of these treatments were evaluated. Methods A total of 245 fertile eggs were randomly distributed into four experimental groups and set in the incubator. On Day 9 of incubation, they were injected with 10 pg of R76713 dissolved in 20 ~1 of saline/PEG400 (R76 group, N = 72), 25 pg EB dissolved in 50 ,xI oil (EB group, N = 47), or with both compounds at the same doses and dilutions (R76/EB group, N = 74) or with the solvent of both compounds [20 ~1 saline/PEG400 + 50 ~1 oil; control (C) group, N = 521. All birds were gonadectomized 25-27 days after hatching and implanted with T-filled Silastic capsules at 7 weeks of age. They were tested twice in the arena and once in their home cage during a 4-day period starting 9 days after the beginning of the T treatment. Final numbers of subjects in the different groups were 8 males and 18 females in the C group, 15 males and 15 females in the R76 group, 13 males and 5 females in the EB group, and 4 males and 4 females in the R76/EB group.

192

BALTHAZART,

DE CLERCK,

AND FOIDART

Results The hatching rate was again unequal in the four experimental groups (x’ = 26.939, P < .OOOl). This was due this time to a low success in the R76/EB group (C: 33/52, R76: 44/72, EB: 25/47, R76/EB: H/74). The low hatching rate in the R76/EB group was due to late mortality in egg (38 out of 74 eggs). When this group was excluded, hatching success was similar in the other three groups (x’ = 1.192, P = 551). Body weight of the newly hatched chicks was similar in all groups (F3.119 = 0.635. P = .594) at a value of 8.13 -+ 0.91 g (mean _+ SD). Body weight in the adults was analyzed by three-way ANOVA with the sex of the birds and the injections of R76713 (yes/no) or of EB (yes/no) as factors. This confirmed again the higher weight of the females (about 5%) compared to males (before T treatment: F1,74 = 5.244, P = .024; after T treatment: F3,ii9 = 5.207, P = .025). These analyses revealed no effect of the treatments with either R76713 or EB and no signifcant primary or secondary interaction between these factors. The three-way ANOVA of cloaca1 gland sizes measured at the end of the experiment revealed significant effects of the sex of the birds (F1,r4 = 37.026, P < .OOl) and of the EB treatment (F1,r4 = 32.440, P < .OOl) but not of R76713 (&, = 0.114, P = .736) nor of the primary or secondary interaction. Cloaca1 glands were usually larger in males than in females and they were decreased by EB injections in egg (see Fig. 1 for the detail of statistical comparisons by the Fisher PLSD test). The behavioral tests performed in the arena confirmed previously established facts concerning quail copulatory behavior: it is present in normal males but not in normal females, it is inhibited in males by EB injections in egg, and it is maintained in females that have been treated with R76713 in egg on Day 9 of incubation. In addition, the present experiment showed that the vigorous masculine behavior seen in R76713-treated females had been completely suppressed by the concurrent treatment with EB (see Fig. 4). These experimental effects were statistically established by the analysis of the percentage of birds showing MA or CCM (see top panels of Fig. 4 for the results of the Fisher exact probability tests) and by the analysis of the behavioral frequencies using three-way ANOVA as described above. These ANOVA confirmed the significant effects of the sex of the birds (MA: Fi,+, = 22.956, P < .OOl; CCM: &, = 6.242, P = .014; overall frequencies higher in males), of the EB treatment (MA: F,,7a = 52.963, P < .OOl; CCM: Fl,74 = 18.048, P < .OOl; suppression of behavior by EB injected in the egg), and of the interactions of sex with R76713 [MA: Fv-1 = 5.153, P = .026; CCM: Fl;1.,4= 1.693, P = .197; R76713 increases frequencies in females but not in males (not significant for CCM)] and of sex with EB (MA: F1,rd = 16.647, P < .OOl; CCM: F1,74 = 6.242,

SEXUAL

DIFFERENTIATION

193

OF QUAIL

Cloaca1Contact Movement

Mount Attempt 100

Females

Males

I

Males

Females

SEX

Ia R76

DC

D EB

Males

lil R?LIEB

Females

SEX

4. Effects of the aromatase inhibitor R76713, of &radio1 benzoate, and of the combined treatment on the sexual differentiation of masculine sexual behavior in quail. Eggs were injected on Day 9 of incubation with either 10 pg R76713 (R76) or 25 lug estradiol benzoate (EB). or with both compounds at the same doses (R76/EB), or with the control vehicle solution (C). Data presented are the total percentage of birds that performed at least one mount attempt or one cloaca1 contact movement during the behavioral tests performed in the arena (top graphs” and the mean frequency (&SEMI per test of these behaviors (bottom graphs). The percentages of active birds were analyzed by the Fisher exact probability test, the frequency data were analyzed by ANOVA followed when appropriate by Fisher PLSD test comparing groups two by two. Results of group comparisons are reported at the top of the corresponding bars as foilou;s: *, P < .05 for the comparison with the control birds of the same sex; #, P < .05 for the comparison with males submitted to the same treatment; A, P < .05 for the comparison with birds treated with R76713 alone. FIG.

P = .014; EB decreases behavior in males but not in females that are already inactive). The detailed comparisons of the experimental groups two by two using the Fisher PLSD test can be found in the bottom panels of Fig. 4 and confirm the general conclusions presented above. The results of the behavioral tests performed in the home cages led to similar conclusions and are therefore not presented here. Experiment 4 The first three experiments clearly demonstrated that the injection of an aromatase inhibitor in quail eggs maintains the capacity to display masculine copulatory behavior in the adult females hatched from these treated eggs. Experiment 3 also showed that the behavioral effects of R76713 were presumably the result of the decreased estrogen production in the embryo since concurrent injection on Day 9 of incubation of EB in R76713-treated eggs completely suppressed the effects of the aromatase

194

BALTHAZART,

DE CLERCK,

AND FOIDART

inhibitor. A last experiment was carried out to determine whether a delayed injection of EB would still produce the same behavioral effects. Methods A total of 223 fertile eggs were randomly distributed in four experimental groups and placed in the incubator. On Day 9 of incubation, all eggs were injected with R76713 (10 pg in 20 yl PEG400-saline). Two groups were then injected immediately after (Day 9) with either 25 pg EB dissolved in 50 ~1 oil (Group R76/EB.D9, N = 70) or 50 ,~l oil (Group R76/0il.D9, N = 46), while the other two groups received similar injections but these were performed on Day 14 only (Group R76/EB.D14, N = 67 and Group R76/0il.D14, N = 40). All birds were gonadectomized during the 4th week of post-hatching life (Days 26 or 27) and implanted with T-filled Silastic capsules (40 mm) at 6 weeks of age. They were tested for masculine sexual behavior twice in the arena and twice in their home cage during a ‘I-day period beginning 12 days after the implantation of the T-filled capsules. The final number of adult subjects in the different experimental groups was 19 males and 15 females for R76/EB.D9, 7 males and 13 females for R76/0il.D9, 20 males and 23 females for R76/EB.D14, and 10 males and 8 females for R76/0il.D14. Results Approximately 65% of the fertile eggs (146/223) hatched in this experiment and the hatching success was in the same order of magnitude in the four experimental groups (R76/EB.D9: 45/70; R76/0il.D9: 23/46; R76/EB.D14: 52/67; R76/0il.D14: 26/40, i.e., 50 to 77%). The eclosion rate was however statistically different between groups (x’ = 9.286, P = .026) due to higher success in the R76/EB.D14 group and lower success in the R76/0il.D9 group. Exclusion of one or the other of these groups from the analysis was sufficient to cancel the statistical differences between the remaining three groups of birds. The body weight of the chicks significantly differed according to their experimental group (F& = 4.862, P = .003). This was due exclusively to a higher weight in the R76/EB.D9 group (significantly different from the three other groups by the Fisher PLSD test; R76/EB.D9: 9.3 & 1.14, R76/0il.D9: 8.73 +- 1.17, 8.57 4 1.13 g, mean -C SD). R76/EB.D14: 8.5 2 1.01, R76/0il.D14: In adult birds, all quantitative data (body weight, cloaca1 gland area, behavioral frequencies) were analyzed by three-way ANOVA with the sex of the birds (male/female), the EB treatment (yes/no), and the age of the EB/Oil injections (Day 9/14) as factors. The analysis of body weights measured before the T treatment indicated no major effect and no interaction. After treatment with T, the ANOVA revealed a small effect of the sex (FI,107 = 4.036, P = .047; males: 204.14 -C 17.96, females

SEXUAL

DIFFERENTIATION

OF QUAIL

195

= 211.76 -+_23.97 g, mean +. SO) but not other major effect or interaction. Cloaca1 gland areas were significantly affected by the three main factors (Sex: Fl,107 = 31.143, P < .OOl, EB: Fl,107= 37.626, P < .001, Age: F1,107 -- 15.185, P < .OOl) as well as by several interactions between these factors (Sex x Age: Fl,107 = 4.429, P = .037; EB x Age: F1,107= 18.777, P < -001 and Sex x EB x Age: F,,107= 4.415, P = .038). The Fisher PLSD tests comparing two by two the experimental groups clearly identified the origin of these effects: glands were smaller in females than in males (except for group R76/0il.D14) and EB decreased gland size when injected on Day 9 but not on Day 14 (see Fig. 1 for the detail of the statistical comparisons). As expected based on the previous experiments, the percentage of birds performing MA or CCM in the behavioral tests (arena) was significantly affected by the embryonic treatments (Fig. 5). Since all birds had received an injection of R76713 on Day 9 of incubation, females were usually behaviorally active. The treatment with EB on Day 9 suppressed the behavior in males and also in females (suggesting again that the presence of copulatory behavior in R76713-treated females was caused by the removal of endogenous estrogens). Interestingly, the effects of the EB injection performed on Day 14 of incubation were different in males and females. While no behavioral consequence was detected in males in agreement with previous observations (Adkins, 1979; Schumacher et al., 1989), a decrease in the percentage of active females was observed. The magnitude of this effect was not sufficient to produce a significant difference in the comparison with the corresponding female controls (comparison of the R76/EB.D14 female with the R76/0il.D14 group indicated by * in the figure) but the effect was clearly suggested by the fact that the females treated with EB on Day 14 were significantly less active than the males receiving the same treatment (# in the figure) while the birds treated with oil at the same age were not. The analysis of the behavioral frequencies confirmed these effects. The frequencies of MA and CCM observed during the tests performed in the arena were similarly modified by the three main factors (for M-4: Sex: Fl,lo7 = 38.208, P = .005, EB: F 1,1o7= 19.893, F < .OOl, Age: FI,Io7 = 13.651, P < .OOl; for CCM: Sex: Fl.107= 13.847, P < .OOl, EB: F1,rO, = 32.948, P < .OOl, Age: Fl,107= 12.966, P < .OOl), by the interaction of EB and age (for MA: F,,,o, = 13.500, P-C .001; for CCM: Fl,Lo7= 4.200~ P = .043) and by the secondary interaction (Sex x EB ‘x Age; for MA: Fl,107= 7.310, P = .008; for CCM: Fl,107= 18.733, P < .OOl). In males, EB suppressed copulatory behavior if injected on Day 9 but not on Day 14. By contrast, in females, significant effects of the treatment were observed at both ages even if early injections had a more pronounced effect (see bottom of Fig. 5 for the detail of statistical comparisons by the Fisher

196

BALTHAZART,

DE CLERCK,

Mount Attempt

0

Males

Females

AND FOIDART

Cloaca1Contact Movement

Males

Females

SEX SEX FIG. 5. Effects of the aromatase inhibitor R76713, of estradiol benzoate, and of the combined treatment on the sexual differentiation of masculine sexual behavior in quail embryos that had been pretreated on Day 9 of incubation with the aromatase inhibitor R76713. All eggs were injected on Day 9 of incubation with 10 pg R76713 (R76). They received in addition one injection of 25 pg estradiol benzoate (EB) or of the control oil solution (Oil) either on Day 9 or on Day 14 of incubation. Data presented are the total percentage of birds that performed at least one mount attempt or one cloaca1 contact movement during the behavioral tests performed in the arena (top graphs) and the mean frequency (%SEM) per test of these behaviors (bottom graphs). The percentages of active birds were analyzed by the Fisher exact probability test, and the frequency data were analyzed by ANOVA followed when appropriate by Fisher PLSD test comparing groups two by two. Results of group comparisons are reported at the top of the corresponding bars as follows: *, P < .05 for the comparison with the control birds of the same sex (R76/0il at the same age); #! P < .05 for the comparison with males submitted to the same treatment; A, P < .05 for the comparison with birds of the same sex treated with EB on Day 9 (R76/EB.D9).

PLSD tests). It is important to notice that females injected with EB on Day 14 of incubation were here significantly less active than the control females injected with oil at the same age. The comparison of the percentage of active birds and of the behavioral frequencies indicates that the effects of the EB treatment on Day 14 were sexually differentiated: no behavioral consequence was detected in males while in females a significant behavioral inhibition could be observed. It was however clear that the earlier treatment (Day 9) with estrogen was more potent even in females. This latter conclusion was reinforced by the analysis of the behavioral data collected during the tests performed in the home cages, All major effects described above were also observed in this other condition but the suppression of behavior in females by EB injections on Day 14 was no longer so prominent. Females treated with EB

SEXUAL

DIFFERENTIATION

OF QUAIL

197

on Day 14 of incubation again displayed masculine copulatory behavior with frequencies that were reduced by comparison with the o&treated females but there was, in this test condition, no effect of EB on the percentage of active birds (data not shown). DISCUSSION Taken together, these four experiments clearly demonstrate that estrogens are the major and probably only hormonal stimulus that is responsible for the behavioral demasculinization in female quail. Previous work based on the injection of exogenous estradiol in male embryos had strongly suggested this possibility (Adkins, 1975, 1979) and one experiment using an antiestrogen in females had indicated that the mechanism suggested had indeed a physiological significance (Adkins, 1976). We now demonstrate that the removal of endogenous estrogens in females by means of a potent aromatase inhibitor completely suppresses the demasculinization of copulatory behavior. This clearly proves that estrogens are fully responsible for this process in physiological conditions. The fact that exogenous estrogens are able to restore the behavioral demasculinization of the R76713-treated females also supports this conclusion. The major fact observed in the present experiments is that the injection of a potent aromatase inhibitor into female quail eggs blocks the behavioral demasculinization, therefrom demonstrating that this process is the result of the embryonic exposure to estrogens. This effect is extremely robust and could be reproduced in four independent experiments. It is also important to notice that the females that had been treated with R76713 on Day 9 of incubation or before could not be distinguished behaviorally from normal males. The percentage of birds that showed copulatory behavior was similar in R76713-treated females and in intact males and the behavioral frequencies were identical in these two groups. The behavioral demasculinization of the females was compZetely blocked by the aromatase inhibitor and it therefore appears that this demasculinization is the exclusive consequence of an embryonic exposure to estrogens. Female brains can potentially express masculine copulatory behavior. The absence of these behavioral patterns in normal females is not the consequence of genetic differences but simply results from an ontogenetic differentiation process brought about by endogenous estrogens that are presumably of ovarian origin. These studies demonstrate that the aromatase inhibitor, R76713, represents an excellent experimental tool for the study of sexual differentiation It was used in the present studies to analyze the evolution in time of the behavioral demasculinization in female quail. Previous work had shown that exogenous estrogens are capable of demasculinizing male embryos if they are injected in egg on or before Day 12 of incubation (Adkins, 1979). This suggested by analogy that the sensitive period for sexual

198

BALTHAZART,

DE CLERCK,

AND

FOIDART

differentiation was closed on Day 12 in female embryos also (AdkinsRegan, 1983). However, more recent studies showed that if female quail are ovariectomized soon after hatching (Days 16-17 of incubation) they are still capable of exhibiting a weak masculine sexual behavior at least if tested in optimal conditions (Hutchison, 1978; Schumacher and Balthazart, 1984). They are therefore not fully demasculinized at hatching and actually a treatment with exogenous estrogens is able to complete this differentiation process if applied to ovariectomized females during the first 4 weeks post-hatch. Females are therefore still sensitive to the demasculinizing effects of estrogens during the post-hatching life while males are completely insensitive as early as Day 14 of incubation (Balthazart and Schumacher, 1984; Schumacher et al., 1989). The use of R76713 has allowed us to confirm here this sex difference. During Experiment 2, the evolution in time of the female demasculinization was analyzed by injecting female embryos with the aromatase inhibitor on different days during incubation. It was shown that early treatments are the most effective. Day 12 appears to be a critical age in this respect: if the differentiation is interrupted at that age by injection of the aromatase inhibitor, females are only partly demasculinized. Some of them are active at low frequency while others are not. The weak activity in the arena is clearly enhanced when a more appropriate test situation (Schumacher and Balthazart, 1984) such as the home cage is used. A treatment with R76713 performed on Day 15 is much less effective. Female demasculinization has progressed too far and/or the amount of estrogens that have been accumulated in the egg is already too high so that intense masculine behavior can no longer be elicited even if the synthesis of new estrogens is interrupted. However, some residual activity (mount attempts) is still present and will persist until the first few days after hatching. It will be revealed under optimal testing conditions. These data therefore show that the sexual differentiation of female behavior is a progressive phenomenon and that the most active phase for this process takes place between Days 9 and 15 of incubation. This in fact corresponds well with the period during which plasma levels of estradiol are increased in female embryos (Schumacher, Sulon, and Balthazart, 1988). This phenomenon is driven by estrogens as revealed by the blockage following injection of an aromatase inhibitor and the restoration following the treatment with exogenous estrogens. To obtain the sexual differentiation of a behavioral or morphological characteristics, the appropriate stimulus (estrogen in the present case) must be present during the period when the embryo is sensitive to its action (critical period). The differentiation and its evolution in time as described by Experiment 2 result from the summation of these two factors. Experiment 4 reveals in addition that, in quail, the sensitivity to the

SEXUAL

DIFFERENTIATION

OF

QUALL

199

differentiating action of estrogens is in itself a process that varies in time and that is sexually differentiated. During Experiments 3 and 4, the differentiation of females was blocked by an injection of R76713 on Day 9 of incubation and the effect of exogenous estrogens was subsequently tested. This confirmed that male embryos are no longer sensitive to estradiol effects after Day 12 of incubation. The injection of EB on Day 9 completely suppressed copulatory behavior in males but had absolutely no effect on Day 14. By contrast, in females, the sensitivity to estrogens is not lost so abruptly. A treatment with EB on Day 14 was not as effective as the same treatment applied on Day 9 but it still produced a significant demasculinization. The evolution in time of the sensitivity to estradiol is therefore sexually differentiated in quail embryos. Females retain a partial sensitivity longer than the males. This is in agreement with previous results from this laboratory that showed that a post-hatching treatment with estradiol was able to complete the demasculinization of neonatally ovariectomized females but had no effect in neonatally castrated males (Balthazart and Schumacher, 1984). Based on the available evidence, we have every reason to believe that the behavioral effects of R76’713 that were observed here are specific in nature. No overall effect on body weight of the treatments was observed in these experiments. It was confirmed that adult females are slightly heavier than males (significant in the four experiments) but no effect of the embryonic treatments on this variable could ever be detected. Similarly, the hatching success of the treated eggs was never influenced in a systematic way by the treatments. Significant differences in the proportion of eggs that hatched were detected between experimental groups. Nowever, these differences were not obviously related to a specific treatment. They concerned the eggs injected with 50 pg R76713 in Experiment I, with the control solution on Day 9 of incubation in Experiment 2, with the combination 10 pg R76713 + 25 pg EB in Experiment 3 and with oil in Experiment 4. The differential survival and hatching of the embryos is therefore not related to the hormonal treatment performed in ova. We did in fact notice during the study that the poor hatching systematically concerned a group of eggs placed in the same position in the incubator. The observed differences are likely to be due to small differences in temperature or humidity in a region of the incubator. It can therefore be safely said that the injections of EB and/or of the aromatase inhibitor have no negative effect on the general condition of the embryos. As far as the aromatase inhibitor, R76713, is concerned, this is also supported by the observation that no behavioral consequence of the R76713 injections could be detected in males whose copulatory behavior is supposed to develop in, the absence of estrogenic stimulation. Given that the behavioral effects observed following treatment with R76713 are positive in

200

BALTHAZART,

DE CLERCK,

AND FOIDART

nature (appearance of a behavior) and can be reversed by administration of exogenous estrogens, we conclude that they result from the suppression of endogenous estrogens following inhibition of aromatase activity. It had been previously reported that the cloaca1 gland, an androgendependent structure (Sachs, 1967), is significantly larger in male than in female quail even in gonadectomized adult birds exposed to the same levels of testosterone (Adkins, 1975; Balthazart et al., 1983). It was also shown that if male embryos are treated before Day 12 of incubation with estrogens, they will develop as adults a small gland similar in size to a female structure (their gland is demasculinized) (Schumacher et al., 1989). The present experiments univocally demonstrate that cloaca1 gland demasculinization in quail results from the embryonic exposure to estrogens. When the endogenous source of these steroids in females is suppressed by an aromatase inhibitor, they develop when treated with testosterone as adults a large gland identical to the gland of the males. This effect also can be reverted by a treatment with exogenous estrogens. The cloaca1 gland is therefore a good morphological marker of estrogen action in the egg and its differentiation essentially parallels the differentiation of masculine copulatory behavior. Most of the conclusions presented about behavior can be extended to the cloaca1 gland as far as the mechanisms of sexual differentiation are concerned. In conclusion, we have demonstrated here that the sexual differentiation in quail is primarily, perhaps exclusively, controlled by estrogen. The steroid is acting during two distinct stages: an early phase when the sensitivity to estrogens becomes differentiated and a latter phase during which the behavioral demasculinization is established. The notion that hormones might act in a biphasic manner during ontogeny to first sensitize brain cells and subsequently differentiate them had been proposed more than 10 years ago (Weisz and Ward, 1980; Harlan, Gordon, and Gorski, 1979). Circumstantial evidence suggests that this two-step model might be applicable to several experimental models concerning both mammals and birds (Baum, Erskine, Kornberg, and Weaver, 1990; Rhees, Shryne, and Gorski, 1990; Schumacher and Balthazart, 1985; Balthazart and Schumacher, 1987). The present data for the first time provide strong experimental evidence showing that the sensitivity to differentiating effects of estrogens is sexually differentiated during the late phase of the incubation in quail. The hormonal bases of the sexual differentiation in estrogen sensitivity are unknown and should now be investigated. We previously speculated that exposure to low doses of estradiol at a very early age was sensitizing the brain of females to the demasculinizing action of this steroid (Balthazart and Schumacher, 1987; Schumacher and Balthazart, 1985). This might also prolong their sensitivity. Experimental evidence supporting this proposition is lacking at present but could now be collected using the aromatase inhibitor, R76713, as an experimental tool. It should

SEXUAL

DIFFERENTIATION

201

OF QUAIL

indeed be possible to inhibit estradiol production very early on in female embryos (before Day 9) and then to test their sensitivity to estrogens. It is quite possible that if they are never exposed to estrogens, female embryos will then react like male embryos to experimental manipulations performed later in the embryonic or early post-hatching life. This type of experiment was previously impossible since it is technically impossible to surgically remove the source of steroids in a young avian embryo. R76713 represents a unique tool to investigate these questions. ACKNOWLEDGMENTS This work was supported by grants from the National Institutes of Health, Bethesda, MD (HD 22064) the Belgian FRFC (Nbr. 2.9003.91 and 9.4601.90), the EEC (SCl-U230-C/TT), and the University of Liege (Fonds Speciaux pour la Recherche) to J. B. We thank Professor E. Schoffeniels for his continued interest in our research and Dr. M. Schumacher (Universite de Paris Sud, Dept. Chimie Biologique) for his comments on an earlier version of this work. We are also grateful to Dr. R. De Coster (Janssen Research Foundation, Beerse, Belgium) for the generous supply of R76713.

REFERENCES Adkins, E. K. (1975). Hormonal basis of sexual differentiation Comp.

Physioi.

Psychol.

in the Japanese

quail.

J.

89, 61-71.

Adkins, E. K. (1976). Embryonic exposure to an antiestrogen masculinizes behavior of female quail. Physiol. Behav. 17, 357-359. Adkins. E. K. (1978). Sex steroids and the differentiation of avian reproductive behavior. Amer.

Zool.

18, 501-509.

Adkins, E. K. (1979). Effect of embryonic treatment with estradiol or testosterone on sexual differentiation of the quail brain. Nelaroendocrinology 29, 178-185. Adkins, E. K., and Adler, N. T. (1972). Hormonal control of behavior in the Japanese quail. .I. Comp. Physiol. Psychol. 81, 27-36. Adkins-Regan, E. (1983). Sex steroids and the differentiation and activation of avian reproductive behaviour. In J. Balthazart and R. Gilles (Eds.), Hormones and Behaviour in Higher Vertebrates, pp. 219-228. Springer-Verlag, Berlin. Adkins-Regan, E. (1990). Hormonal bases of sexual differentiation in birds. In J. Balthazart. (Eds.). Hormones, Brain and Behaviour in Vertebrates: 1. Sexual Wifferenriution, Neuroanatomical Aspects, Neurotransmitters and Neuropeptides. Comp. Physiol. Vol. 8, pp. 1-14. Karger. Basel. Adkins-Regan, E., Abdelnabi, M., Mobarak, M., and Ottinger, M. A. (1990). Sex steroid levels in developing and adult male and female zebra Iinches (Poephila guttutu). Gen. Comp. Endocrinol. 78, 93-109. Adkins-Regan, E., and Ascenzi; M. (1987). Social and sexual behaviour of male and female zebra finches treated with oestradiol during the nestling period. Anim. Behav. 35, 11001112. Adkins-Regan, E., and Ascenzi, M. (1990). Sexual differentiation of behaviour in the zebra finch: Effect of early gonadectomy or androgen treatment. Horm. Behav. 24, 114-127. Adkins-Regan, E. K., Pickett, P., and Koutnik. D. (1982). Sexual differentiation in quail: Conversion of androgen to estrogen mediates testosterone-induced demasculinization of copulation but not other male characteristics. Horm. Behav. 16, 259-278. Balthazart, J. (1989). Steroid metabolism and the activation of social behavior. In J. Bal-

202

BALTHAZART,

DE CLERCK,

AND FOIDART

thazart (Eds.), Advances in Comparative and Environmental Physiology, Vol. 3, pp. 105-159. Springer-Verlag. Berlin. Balthazart, J., Evrard, L., and Surlemont, C. (1990). Effects of the non-steroidal aromatase inhibitor, R76713 on testosterone-induced sexual behavior in the Japanese quail (Coturnix coturnix japonica). Horm. Behav. 24, 510-531. Balthazart, J., and Schumacher, M. (1984). Estradiol contributes to the postnatal demasculinization of female Japanese quail (Coturnix cotumix japonica). Homz. Behav. 18, 287-297. Balthazart, J., and Schumacher, M. (1987). A two-step model for sexual differentiation. In B. R. Komisaruk, H. I. Siegel, M. F. Cheng, and H. H. Feder (Eds.), Reproduction: A Behavioral and Neuroendocrine Perspective, pp. 308-324. Annals of the New York Academy of Sciences, New York. Balthazart, J., Schumacher, M., and Ottinger, M. A. (1983). Sexual differences in the Japanese quail: behavior, morphology and intracellular metabolism of testosterone. Gen. Comp. Endocrinol. 51, 191-207. Baum, M. J., Erskine, M. S., Kornberg, E., and Weaver, C. E. (1990). Prenatal and neonatal testosterone exposure interact to affect differentiation of sexual behavior and partner preference in female ferrets. Behav. Neurosci. 104, 183-198. De Coster, R., Wouters, W., Bowden, C. R., Vanden Bossche, H., Bruynseels, J., Tuman, R. W., Van Ginckel, R., Snoeck, E., Van Peer, A., and Janssen, P. A. J. (1990). New non-steroidal aromatase inhibitors: Focus on R76713. J. Steroid Biochem. 37,335341. Gibson, W. R., Follett, B. K., and Gledhill, B. (1975). Plasma levels of luteinizing hormone in gonadectomized Japanese quail exposed to short or to long daylengths. J. Endocrinol. 64, 87-101. Goy, R. W., and McEwen, B. S. (1980). Sexual Differentiation of the Brain. MIT Press, Cambridge, MA. Gurney, M. E. (1981). Hormonal control of cell form and number in the zebra iinch song system. J. Neurosci. 1, 658-673. Gurney, M. E., and Konishi, M. (1980). Hormone-induced sexual differentiation of brain and behavior in zebra finches. Science 208, 1380-1383. Harlan, R. E., Gordon, J. H., and Gorski, R. A. (1979). Sexual differentiation of the brain: Implications for neuroscience. In D. M. Schneider (Eds.), Review of Neuroscience. Vol. 4, pp. 31-71. Raven Press, New York. Hutchison, J. B., Wingfield, J. C., and Hutchison, R. E. (1984). Sex differences in plasma concentrations of steroids during the sensitive period for brain differentiation in the zebra finch. J. Endocrinol. 103, 363-369. Hutchison. R. E. (1978). Hormonal differentiation of sexual behavior in Japanese quail. Horn. Behav. 11, 363-387. Mathews, G. A., and Arnold, A. P. (1990). Antiestrogens fail to prevent the masculine ontogeny of the zebra finch song system. Gen. Comp. Endocrinol. 80, 48-58. Mathews, G. A., Brenowitz, E. A., and Arnold, A. P. (1988). Paradoxical hypremasculinization of the zebra finch song system by an antiestrogen. Horn. Behav. 22, 540551. Pohl-Apel, G., and Sossinka, R. (1984). Hormonal determination of song capacity in females of the zebra finch. 2. Tierpsychol. 64, 330-336. Rhees, R. W., Shryne, J. E., and Gorski, R. A. (1990). Termination of the hormonesensitive period for differentiation of the sexually dimorphic nucleus of the preoptic area in male and female rats. Dev. Brain Res. 52, 17-23. Sachs, B. D. (1967). Photoperiodic control of the cloaca1 gland of the Japanese quail. Science 157, 201-203.

SEXUAL

DIFFERENTIATION

OF QUAIL

203

Schumacher, M., and Balthazart, J. (1984). The postnatal demasculinization of sexual behavior in the Japanese quail. Harm. Behav. 18, 298-312. Schumacher, M., and Balthazart, J. (1985). Sexual differentiation is a biphasic process in mammals and birds. In R. Gilles and J. Balthazart (Eds.), Neurobiology: Current Comparative Approaches, pp. 203-219. Springer-Verlag, Berlin. Schumacher, M., Hendrick, J. C., and Bahhazart, J. (1989). Sexual differentiation in quail: Critical period and hormonal specificity. Norm. Behav. 23, 130-149. Schumacher, M., Sulon, J., and Balthazart, J. (1988). Changes in serum concentrations of steroids during embryonic and post-hatching development of male and female Japanese quail (Coturnix coturnix japonica). J. Endocrinol. 118, 127-134. Weisz, J., and Ward, I. L. (1980). Plasma testosterone and progesterone titers of pregnant rats, their male and female fetuses, and neonatal offspring. Endocrinology 106, X6316. Wilson, J. A., and Glick, B. (1970). Ontogeny of mating behavior in the chicken. Amer. J. Physiol. 218, 951-955. Wouters, W., De Coster, R., Krekels, M.. Van Dun, J., Beerens. D.. Haelterman, C., Raeymaekers, A., Freyne, E., Van Gelder, J., Venet, M., and Janssen, P. A. J. (1989). R76713, a new specific non-steroidal aromatase inhibitor. J. Steroid Biochem. 32, %I788.