Effects of estrogens and antiestrogens on gonadal sex differentiation and embryonic development in the domestic fowl (Gallus gallus domesticus)

Dosing

All experiments were carried out with respect for the principles of laboratory animal care, in accordance with the European Communities Council Directive of 24 November 1986 (86/609/EEC) and the German Animal Welfare Act. Fertilized eggs of white Leghorn (G. domesticus) were obtained from a local breeder (LSL Rhein-Main Geflügelvermehrungsbetrieb, Dieburg, Germany). The total number of eggs per experiment and treatment-group are shown in Table 1 for fulvestrant and Table 2 for tamoxifen, including the treatment-groups with parallel sole and co-exposure to 17α-ethinylestradiol (EE₂). The testing of tamoxifen was conducted in a series of four experiments while fulvestrant was tested in a single experiment. Tamoxifen was tested in four experiments to ensure a higher degree of replication for an up to then not tested class of EDCs with unknown effects. The eggs were incubated at 37.5 ± 0.5 °C and 60 ± 10% relative humidity and turned over eight times a day in a fully automated incubator (J. Hemel Brutgeräte, Verl, Germany).

Fulvestrant (CAS: 129453-61-8; purity: ≥98%), tamoxifen (CAS: 10540-29-1; purity: ≥99%) and EE₂ (CAS: 57-63-6; purity: ≥98%) were purchased from Sigma Aldrich Chemie GmbH (München, Germany). Fulvestrant (applied doses: 0.1, 1, 10 µg/g egg), tamoxifen (applied doses: 0.001, 0.01, 0.1, 1, 10 µg/g egg) and EE₂ (applied dose: 20 ng/g egg) were dissolved alone or in combination in 60 µL of the solvent dimethyl sulfoxide (DMSO; CAS: 67-68-5; purity: 99.5%; Applichem, Darmstadt, Germany). Eggs were yolk-injected on day 1 of incubation and further processed until dissection as described elsewhere (Jessl, Scheider & Oehlmann, 2018).

Dissection, tissue preparation and evaluation

On day 19 of incubation embryos were dissected. Deformations of body and internal organs were recorded with special focus on ovaries and testes. Gonads were photographed (Diskus, Carl H. Hilgers, Königswinter, Germany) for further analysis of the gonad surface area, in which the entire visible surface of each single gonad was determined with an image editing program (Fiji is just ImageJ, Open Source). After dissection gonads were fixed in Bouin’s solution, which was removed by repeated rinsing with 80% ethanol after 24 h. Ethanol was removed by saccharose solution (10, 20 and 30% in phosphate buffered saline). Gonads were embedded in Tissue-Tek^® (Sakura Finetek Europe B.V., Alphen aan den Rijn, Netherlands) and sectioned (6 µm) by a freeze microtome (Microm HM 500 O, Thermo Fisher Scientific Germany, Bonn, Germany) at −23 °C. Tissue sections were stained with hematoxylin and eosin.

Measurements and statistics

Histological examination was performed using a light microscope (Olympus BX50, Olympus, Tokyo, Japan) and a camera (JVC Digital Camera, KY-F75U, Yokohama, Japan). Gonadal cortex thickness of both sexes and male percentage of seminiferous tubules in left testes were measured (Fiji is just ImageJ, Open Source). Ten sections for each embryo were evaluated which were exclusively taken from the gonad’s middle sectional plane as described previously (Jessl, Scheider & Oehlmann, 2018).

For the test series with the combination of tamoxifen and EE₂, the results of four test runs and for the combination of fulvestrant and EE₂, the results of a single test run were merged and analyzed. For the endpoints, gonadal cortex thickness and percentage of seminiferous tubules as well as gonad surface area data were normalized to the solvent control. Data were analyzed using Fisher’s exact test, one-way ANOVA with Dunnett’s multiple comparison test or Kruskal-Wallis test with Dunn’s multiple comparison test with GraphPad Prism 5.01 (GraphPad Software Inc., San Diego, USA).

Determination of sexual genotype

DNA isolation was performed using a tissue sample from the heart taken during dissection. All embryos were typed for their sexual ZZ or ZW genotype, using the PCR-based method of Fridolfsson & Ellegren (1999) with the primers 2550F (5′-GTT ACT GAT TCG TCT ACG AGA-3′) and 2718R (5′-ATT GAA ATG ATC CAG TGC TTG-3′) and a modified protocol. Both primers mark two CHD1 introns, located on the Z (CHD1Z, 600 bp) and W chromosome (CHD1W, 450 bp). Thermal cycling was composed of DNA polymerase activation at 95 °C for 15 min followed by 40 cycles of denaturation at 95 °C for 30 s, annealing at 52 °C for 30 s, elongation at 72 °C for 40 s and a final extension step at 72 °C for 5 min followed by a melt curve (60–95 °C with a heating rate of 0.2 °C/s). Following qPCR-mediated amplification with EvaGreen^® dye, all amplicons showed two different melting peaks based upon the different melting properties of double-stranded DNA, which resulted in characteristic bands for each sex (Chen et al., 2012). Both sexes had a single 600-bp CHD1-Z specific fragment with a melting temperature of ∼84 °C. Females had an additional 450-bp CHD1-W female-specific fragment with a melting temperature of ∼82 °C.

Effects of in ovo exposure to fulvestrant and EE₂

Embryonic mortality and malformations

The number of fertilized eggs, embryonic mortality and malformations per treatment-group are presented in Table 1. The fertility rate of the individual groups was at least 82% with a mean total fertility rate of 92% for the whole experiment. The solvent control group showed the lowest mortality of 15%, followed by the untreated control group with 18% mortality. Sole treatment to fulvestrant as well as co-exposure to all concentrations of fulvestrant plus EE₂ showed mortality rates up to 35%. Sole treatment to EE₂ caused 46% mortality being significantly different from the control (p < 0.05) (Fig. 1A).

While no malformations were detected in the control groups, all substance treated groups showed one to two malformed embryos. The EE₂-treated group 1 of 22 embryos (4.55%) showed a single malformation which was found to be celosomia. Examining all groups receiving different concentrations of fulvestrant, four of 60 embryos (6.67%) showed malformations which were mainly found to be celosomia or malformations of the eyes or the beak. Examining all groups co-exposed to fulvestrant and EE₂, five of 60 embryos (8.33%) showed malformations which were mainly found to affect the beak, the eyes, the limbs, the brain or celosomia. Compared to the control, none of the tested groups showed a substance-induced increase in the rate of malformation.

Morphological observation of the gonads—gonad surface area

In females sole exposure to EE₂ or all concentrations of fulvestrant as well as co-exposure to EE₂ and lower concentrations of fulvestrant (0.1 and 1 µg/g egg) had no statistically significant effect on left and right gonad surface area. Co-exposure to EE₂ and 10 µg fulvestrant/g egg, however, resulted in a statistically significant increase of the left gonad surface area (p < 0.001) (Fig. 1B and Table 3).

Table 3:

Gonad surface area, gonadal cortex thickness and percentage of seminiferous tubules of chicken embryos after in ovo exposure to fulvestrant (Ful, 0.1, 1, 10 µg/g egg) and EE₂ (20 ng/g egg) or co-exposure to all concentrations of fulvestrant and EE₂.

Sex	Group	Gonad surface area		Cortex thickness (µm)	Seminiferous tubules (%)
		left (mm2)	right (mm2)
Male	NC	4.22 ± 0.95	3.92 ± 0.57	10.1 ± 0.40	28.3 ± 4.49
	SC	3.99 ± 0.46	3.91 ± 0.30	10.7 ± 0.73	30.2 ± 1.83
	Ful 0.1	4.24 ± 0.42	3.55 ± 0.24	11.5 ± 0.61	33.8 ± 3.56
	Ful 1	3.90 ± 0.36	3.49 ± 0.13	11.6 ± 1.30	32.0 ± 2.60
	Ful 10	3.99 ± 0.82	3.24 ± 0.66	12.0 ± 1.54	30.6 ± 3.98
	EE2	6.66 ± 0.82c	3.79 ± 0.23	44.6 ± 39.3a	27.3 ± 4.65
	Ful 0.1 +EE2	5.23 ± 1.23	3.71 ± 0.89	25.0 ± 18.3	26.4 ± 1.23
	Ful 1 +EE2	6.21 ± 1.09b	3.75 ± 0.67	47.6 ± 49.4a	23.6 ± 2.82a
	Ful 10 +EE2	4.12 ± 0.53	3.29 ± 1.05	10.4 ± 0.91	31.4 ± 2.70
Female	NC	10.3 ± 0.69c	2.38 ± 0.34b	143 ± 33.9	–
	SC	8.23 ± 0.95	1.80 ± 0.34	154 ± 30.0	–
	Ful 0.1	9.35 ± 0.72	1.84 ± 0.68	149 ± 15.0	–
	Ful 1	8.80 ± 1.03	2.10 ± 0.40	145 ± 32.7	–
	Ful 10	8.91 ± 1.20	2.01 ± 0.37	149 ± 19.5	–
	EE2	8.40 ± 0.91	1.81 ± 0.27	133 ± 15.0	–
	Ful 0.1 +EE2	8.50 ± 1.98	1.69 ± 0.45	153 ± 8.89	–
	Ful 1 +EE2	8.74 ± 1.12	1.67 ± 0.37	146 ± 37.2	–
	Ful 10 +EE2	11.0 ± 1.37c	1.68 ± 0.41	153 ± 21.5	–

DOI: 10.7717/peerj.5094/table-3

Notes:

Statistical analysis by one-way ANOVA with Dunnett’s multiple comparison test. Lowercase indicate significant differences compared to the solvent control (SC). NC, untreated control. Level of significance: ^ap < 0.05; ^bp < 0.01; ^c0.001.

In males fulvestrant did not affect left and right gonad surface area at all applied sole concentrations. EE₂ administered alone or in combination with lower concentrations of fulvestrant (1 µg/g egg) resulted in a significant increase in the left gonad surface area (p < 0.001 and p < 0.01, respectively). The affected left testes showed a female-like shape and a well visible female-typical thickened cortex region when viewed under a stereomicroscope (Fig. 2). However, a concentration of 10 µg fulvestrant/g egg completely antagonized the EE₂-induced feminization of genetic males.

Effects of in ovo exposure to tamoxifen and EE₂

Embryonic mortality and malformations

The number of fertilized eggs, embryonic mortality and malformations per treatment-group are presented in Table 2. The fertility rate of the individual groups was at least 88% with a total fertility rate of 92% for the whole experiment. The unmanipulated negative control showed the lowest mortality of nearly 8% and differed significantly from the solvent-treated control with 28% mortality (p < 0.001). Sole exposure to tamoxifen showed mortality rates up to 56%. Only the group treated with 10 µg tamoxifen/g egg showed a significant difference to the solvent-treated control (p < 0.01). Sole exposure to 20 ng EE₂/g egg caused 24% mortality. Groups co-exposed to tamoxifen and EE₂ showed mortality rates up to 40% with a significant deviation to the solvent-treated control for the group co-exposed to EE₂ and 0.001 µg tamoxifen/g egg (p < 0.05) (Fig. 3A).

While no malformations were detected in the untreated negative control group, four of 75 embryos (5.33%) of the solvent-treated control showed malformations which affected the eyes, the beak, the extremities or edema or celosomia. In the EE₂-treated group three of 45 embryos (6.67%) showed malformations of the eyes, the beak or celosomia. Examining all groups receiving different concentrations of tamoxifen, five of 156 embryos (3.21%) showed malformations which were found to be celosomia, malformations of the vertebral column, the beak or the brain. There was also one egg containing two embryos conjoined at the head, showing various malformations. Examining all groups co-exposed to EE₂ and different concentrations of tamoxifen, 10 of 190 embryos (5.26%) showed malformations which were found to be celosomia, malformations of the beak or the extremities or a missing mesonephros. Compared to the untreated negative control group, the solvent-treated group and the substance-treated groups receiving EE₂ or a combination of EE₂ and 10 µg tamoxifen/g egg showed a statistically significant effect (p < 0.05). On the contrary, there were no significant differences between the substance-treated groups and the solvent control.

Cortex thickness and percentage of seminiferous tubules

In ovo exposure to all concentrations of tamoxifen as well as co-exposure to tamoxifen and EE₂ resulted in a statistically significant concentration-dependent decrease of the left ovarian cortex thickness in females (Fig. 3C). Remarkably, several females exposed to the highest concentration of tamoxifen (10 µg/g egg) also showed an altered distribution of the ovarian cortex region (Fig. 4), which could not be detected at lower concentrations. The affected left ovaries were no longer covered by a continuous ovarian cortex but exhibited larger regions of uncovered medulla, partly resembling a testis. In none of any other experiments, even with different endocrine active compounds, was such a phenomenon detected (Scheider et al., 2014; Scheider et al., 2018; Jessl, Scheider & Oehlmann, 2018).

In males the testicular cortex thickness or the percentage of seminiferous tubules were not affected by sole tamoxifen-exposure. Administration of EE₂ alone or in combination with lower concentrations of tamoxifen (0.001 and 0.01 µg/g egg) resulted in a statistically significant increase in testicular cortex thickness by up to 566%, which was completely suppressed at higher concentrations of tamoxifen (0.1, 1 and 10 µg/g egg). By administration of EE₂ alone or in combination with lower concentrations of tamoxifen (0.001, 0.01 µg/g egg) the percentages of seminiferous tubules were significantly decreased (p < 0.001). However, higher concentrations of tamoxifen (0.1, 1 and 10 µg/g egg) could partly (percentage of seminiferous tubules) or completely (testicular cortex thickness) antagonize the EE₂-induced feminization of genetic males (Table 4).

Embryonic mortality and malformations

In the present study fertilization rates were comparable for all experiments with values around 90%. These values are consistent with the results of Romanoff & Romanoff (1972), which described a fertilization rate of almost 89% in an experiment examining 202 breeding hens. The detection of unfertilized eggs is of importance here since the calculation of mortality and the incidence of malformations is based on the number of fertile eggs. While the untreated control showed mortality rates below 20% in all experiments, in 80% (four of five) of the experiments the mortality in the solvent control was below 30%. Based on the data of various publications (Romanoff & Romanoff, 1972; Wyatt & Howarth, 1976; DeWitt, Meyer & Henshel, 2005a; DeWitt, Meyer & Henshel, 2005b) and our own investigations (Jessl, Scheider & Oehlmann, 2018), the natural mortality of unmanipulated chicken embryos is expected to be about 20%. Since it is known that carrier substances in general, or DMSO in the present case, can affect the survival of chicken embryos, for example through their intrinsic toxicity, volume, or density, an increase in mortality compared to the untreated control was expected (Caujolle et al., 1967; Carew & Foss, 1972; Landauer & Salam, 1972; Morgan, 1974; Wyatt & Howarth, 1976). Not only in the present study (110 untreated and 95 solvent-treated individuals in five experiments), but also in further investigations (Jessl, Scheider & Oehlmann, 2018; 256 untreated and 258 solvent-treated individuals in 15 experiments), all of the experiments showed mortality rates below 20% for the untreated control and a predominant proportion of the experiments (87%; 13 of 15 experiments) showed mortality rates below 30% for the solvent control. Due to the large sample size and the consistent results, we assume that the shown values can be reliably used for the comparison to substance-treated groups.

In general, the treatment of chicken embryos with any of the compounds fulvestrant or tamoxifen alone or in combination with EE₂ led to an increase in mortality as expected. However, significant deviations in mortality from the solvent control were rare and found to be concentration-independent. Groups treated with a defined concentration of fulvestrant or tamoxifen responded similarly to groups that were co-exposed to the same substance plus EE₂. The absence of a statistically significant difference between solely treated and co-exposed groups of the same substance and concentration suggests that the additional administration of EE₂ does not necessarily result in an increase in mortality, as one might have expected. Despite the increase in mortality in some groups of the experiments, a sufficient number of vital embryos remained for subsequent histological studies.

Analyzing the frequency of malformations in the five experiments, no malformations were found in the untreated control group with 110 embryos while in the solvent-treated control group four of 95 embryos (4.21%) exhibited malformations. To some extent, this statistically significant increase in the rate of malformation from untreated to solvent-treated control group reflects a weak teratogenic potential of DMSO as noted by Wyatt & Howarth (1976). However, this value is still in the range of the reported spontaneous malformation rate in chicken embryos of 2–7% (Alsop, 1919; Byerly, 1930; Caujolle et al., 1967). In this context, the incidence of malformations in the solvent-treated control can be considered as inconspicuous.

Since there are no statistically significant differences in the frequency of malformations between the individual concentrations of the respective substances fulvestrant, tamoxifen and/or EE₂ in all experiments, they are combined in groups representing the respective substance (EE_2total, fulvestrant_total, fulvestrant + EE_2total, tamoxifen_total and tamoxifen + EE_2total). While almost all of the substance groupings show significantly higher malformation rates than the untreated control (EE_2total, fulvestrant_total and tamoxifen + EE _2total: p < 0.05; fulvestrant + EE _2total: p < 0.01), none of them are significantly different from the solvent-treated control. In addition, there are no statistically significant differences between the respective substance groupings. However, some of the substance-treated groupings show a marginally (EE_2total, tamoxifen + EE_2total; p > 0.05) or significantly (fulvestrant_total: p < 0.05) increased incidence of celosomia, compared to the solvent-treated control. In addition, exencephalia are exclusively found in substance-treated groups, while the formation of edema is completely absent. Various malformations have already been described, among them, for example, terata of the eyes, the beak, the brain or the formation of celosomia (Byerly, 1930; Caujolle et al., 1967). These terata largely coincide with the malformations found in the solvent- and substance-treated groups. There is no statistical evidence that treatment with fulvestrant, tamoxifen and/or EE₂ specifically favors particular terata or generally results in increased malformation rates.

Morphological observations of the gonads—gonad surface area

Since it is known that substances with endocrine potential are able to induce morphological changes in the sex organs of birds (Scheib, 1983; Berg et al., 1998; Berg et al., 1999; Berg et al., 2001; Berg, Halldin & Brunstrom, 2001; Matsushita et al., 2006; Razia et al., 2006), the chicken embryo appears as a suitable model for the study of early sexual development and the potential impact of EDCs impacting the reproductive axis. In birds, the genetic male is homozygous (ZZ), while the genetic female is heterozygous (ZW). During embryonic development the differentiation in one of the sexes depends on the level of circulating steroid hormones, mainly estrogen (Bruggeman, Van As & Decuypere, 2002). Without estrogen or external influences the undifferentiated gonads of genetic males develop into testes. In genetic females, the key enzyme P450 aromatase (P450arom) is synthesized and testosterone is metabolized to estrogen (Kagami & Hanada, 1997; Ayers, Sinclair & Smith, 2013; Scheider et al., 2014). Since P450arom is not synthesized in male gonads (Ayers, Sinclair & Smith, 2013; Scheider et al., 2014), constitutional estrogen concentration in testes is very low (Woods & Erton, 1978; Tanabe et al., 1979; Tanabe, Yano & Nakamura, 1983) and not sufficient to cause a feminization. Though, for a short time during embryonic development the estrogen receptor (ER) is detectable in male gonads, making males basically vulnerable to estrogen (Gasc, 1980; Smith, Andrews & Sinclair, 1997; Nakabayashi et al., 1998). The artificial presence of estrogen or estrogen-active EDCs at this critical time point therefore causes the differentiation of genetic males towards the phenotypically female sex as demonstrated in the present study and various other studies on domestic fowl (Gallus g. domesticus) and japanese quail (Coturnix japonica) (Sotonyi & Csaba, 1986; Etches & Kagami, 1997; Berg et al., 1998; Berg et al., 1999; Berg et al., 2001; Berg, Halldin & Brunstrom, 2001; Shibuya et al., 2004).

Furthermore, it is shown that the injection of interfering substances into fertilized eggs, such as aromatase inhibitors, can result in a stop of estrogen synthesis leading to a masculinization of ovaries (Elbrecht & Smith, 1992; Burke & Henry, 1999). Not only aromatase inhibitors, but also antiestrogens are capable of affecting female gonad differentiation in birds. Tamoxifen, a selective ER modulator, interacts with various proteins involved in the transcription of estrogen-regulated genes (MacGregor & Jordan, 1998). In the present study sole exposure to high concentrations of tamoxifen resulted in a significant decrease of the left ovarian surface area. This data coincide with the studies of Scheib (1983) on quail embryos, which show that treatment with tamoxifen reduces the size of the left ovary and disturbs the correct formation of the left ovarian cortex. Surprisingly, sole exposure to all concentrations of fulvestrant had no such effect on female ovarian surface area. Due to its considerably higher affinity to ER compared to tamoxifen (Wakeling & Bowler, 1987; Wakeling, Dukes & Bowler, 1991) and the fact that fulvestrant-treatment results in a complete inhibition of the estrogen signaling pathway (Osborne et al., 1995; Wakeling, 1995; Wardley, 2002) we expected fulvestrant as the more potent antiestrogen. The difference between fulvestrant and tamoxifen regarding their effect on embryonic ovarian development of Gallus domesticus could be due to their mode of action, since tamoxifen is a partial ER antagonist, depending on the target tissue, and thus has both antiestrogenic and estrogen-like activity while fulvestrant is a pure or competitive ER antagonist without intrinsic activity (Webb et al., 1995; MacGregor & Jordan, 1998; Bentrem et al., 2001; Shou et al., 2004). However, it is unknown whether these differences can impact the activity in ovo.

The present study also demonstrates a concentration-dependent neutralization of the feminizing effect of EE₂ with increasing concentration of fulvestrant or tamoxifen. In genetic females co-exposure to EE₂ and higher concentrations of fulvestrant or tamoxifen led to a statistically significant increase in female left gonad surface area. This data coincide with the study of Scheib (1983) which shows that the feminizing effect of diethylstilbestrol on male quail embryos could be largely compensated by simultaneous tamoxifen-treatment. While in the experiments of Scheib (1983) sole treatment with the estrogenic compound diethylstilbestrol reduces the volume of the left ovary but does not affect its differentiation, co-exposure to tamoxifen and the estrogen results in a reduction of the tamoxifen-related effects in females. It can be assumed that the competition between estrogen and antiestrogen at the binding site of the ER is influenced by their differing concentrations but also by their specific binding affinity to the receptor. Blair et al. (2000) reported relative binding affinities to the ER from rat uterine cytosol preparations of 190%, 1.6% and 37.5% for EE₂, tamoxifen and fulvestrant, respectively. Because both antiestrogens were applied at doses up to factor 500 higher than the estrogen, it is likely that the estrogen is successively displaced from the ER, resulting in a neutralization of the estrogenic response.

Histological observations of the gonads—left testis and ovary

The results of the present study show that in ovo exposure to EE₂ leads to a reduction of the ovarian cortex in females. In males EE₂ causes a significantly thickened testicular cortex with oocyte-like cells and female-like structure.

Furthermore, sole treatment with fulvestrant or tamoxifen does not significantly affect differentiation of male testes, which coincides with the studies of Salzgeber, Reyssbrion & Baulieu (1981) and Scheib & Baulieu (1981). Their experiments on chicken and quail show that tamoxifen especially affects female but not male gonadogenesis. Thus, tamoxifen-treatment especially disturbed the formation of the left ovarian cortex (Scheib, 1983). In the present study, exposure to 10 µg tamoxifen/g egg results in an altered distribution of the left ovarian cortex which might be based on the specific mechanism of action of tamoxifen. Tamoxifen is a partial agonist of the ER whose effect always depends on the specific tissue along with the intrinsic activity of tamoxifen. As ER alpha is primarily expressed in the gonadal cortex (Nakabayashi et al., 1998), its blockage at high doses might have contributed to the suppression of the cortical hypertrophy in the ovarian differentiation and its unbalanced dispersal. This might lead to the speculation that in higher doses tamoxifen might improve sterility of the respecting females, as the oogonia develop in the border between cortex and medulla (Gonzalez-Moran, 2011).

Since the ER is largely irrelevant for the normal differentiation of the male sex, fulvestrant and tamoxifen remain without statistically significant effect for males. However, the feminization of genetic males of domestic fowl and quail caused by estrogenic substances can be effectively compensated by antiestrogens (Samsel, Zeis & Weniger, 1982; Scheib, 1983) which agrees with the data of the present study. It can be assumed that both estrogenic and antiestrogenic substances compete for the binding site of the ER. Accordingly, a high estrogen-concentration leads to its preferential binding to the ER and results in the feminization of male gonads. With increasing concentration of the antiestrogen the ER is successively blocked. This results in a compensation of the feminizing effects as the ER and subsequent signal cascades are not affected by the estrogen.