This study was carried out in accordance with the recommendations of “experimental animals guidelines established by the Laboratory Animal Research Institute”. The protocol was approved by “Ethical Committee of the Nanjing Medical University”. Female ICR mice (Oriental Bio Service Inc., Nanjing) at 12 weeks of age (30 ± 2 g) were housed in stainless steel cages with wood bedding to minimize additional exposure to endocrine disrupting chemicals (temperature 23 ± 2°C, humidity 55 ± 5%, 12:12 h light/dark cycle, and lights from 06:00) in Animal Research Center of Nanjing Medical University. They received food and water ad libitum. Their body weight was measured every day. All efforts were made to minimize animal suffering. Every early morning (09:00 h), estrous cyclicity was examined using vaginal cytology (Caligioni, 2009).

TCS (>99% purity) was purchased from Sigma-Aldrich (Sigma-Aldrich Inc., St. Louis, MO, USA). TCS was dissolved in dimethyl sulfoxide (DMSO), and then diluted with corn oil (final concentration of 0.5% DMSO). After 3–4 regular estrous cycles were determined, the mice were given the oral intake of TCS at doses of 1, 10 and 100 mg/kg per day at 08:00 h. A recent study (Wang et al., 2015) reported that the urinary TCS level in gestational mice receiving 10 mg/kg TCS is equivalent to high urinary TCS levels of spontaneous abortion patients. Thus, these doses resemble the exposure level to TCS in spontaneous abortion patients. Control mice were treated with oral intake of 0.5% DMSO.

To measure urinary TCS levels, each mouse was housed in a metabolic cage for 5 days. The urine samples (0.2–0.3 ml/mouse) within 12 h after the administration of TCS were collected and stored at −80°C until measurement. The TCS concentrations of urinary (1 ml/mouse) were measured using an established method (Wang et al., 2017). Briefly, the urine samples hydrolyzed with β-glucuronides (Type H-1 from Helix Pomatia, Sigma-Aldrich, St. Louis, MO, USA) were concentrated by a solid phase extraction (SPE; 500 mg/3 mL; Supelco, ENVI-18) and analyzed using liquid chromatography electrospray ionization tandem mass spectrometry (HPLC-MS/MS, Agilent 1290-6490, Agilent Technologies, Santa Clara, CA, USA). Analysts were blinded to all information concerning subjects during the tests.

The mice at diestrus were anesthetized with intraperitoneal (i.p.) injection chloral hydrate (400 mg/kg). Both ovaries were dissected and fixed in Bouin’s fluid. The samples were dehydrated through a graded series of alcohol, cleared in xylene, and then embedded in paraffin wax. After the sections (5 μm) were deparaffined and rehydrated, the sections were stained with hematoxylin and eosin (HE). The classification of follicular stages was made following the morphological criteria as described previously (Myers et al., 2004). Follicles were counted using a conventional light microscope (Olympus DP70, Japan) with a 40× objective. Antral follicles (early antral, antral and preovulatory follicles) and corpora lutea were counted in every 6th section (30 μm apart). Then, the numbers of antral follicles and corpora lutea were multiplied by 6 to give a total number in each ovary.

Mice were anesthetized with chloral hydrate (400 mg/kg, i.p.) and perfused transcardially with 4% paraformaldehyde. Brains were transferred gradually into 15% and 30% sucrose until they settled. Sections (40 μm thick) through the AVPV (Bregma +0.50 to +0.02 mm) the ARC area (Bregma −1.46 to −1.70 mm; Marraudino et al., 2017) were cut using a cryostat. Free-floating sections were incubated in 0.5% sodium metaperiodate for 20 min and then in 1% sodium borohydride for 20 min. The sections were pre-incubated with 1% normal fetal goat serum for 60 min, and then incubated in rabbit anti-kisspeptin polyclonal antibody (1:1000, Catalog# AB9754, Millipore, Billerica, MA, USA) at 4°C for 24 h. Then, the sections were treated with biotin-conjugated goat anti-rabbit IgG (1:400; vector laboratories, Burlingame, CA, USA) at 37°C for 2 h. The immune-reactivity was visualized with the standard avidin-biotin complex reaction with Ni-3, 3-diaminobenzidine (DAB, Vector Laboratories). In every experiment, incubation of sections without the primary antibody served as negative controls for immunohisto-chemistry. Kisspeptin-positive (kisspeptin⁺) cells in AVPV (AVPV-kisspeptin⁺ cells) and ARC (ARC-kisspeptin⁺ cells) were observed by conventional light microscope (Olympus DP70; Olympus, Tokyo, Japan) with a 40× objective.

Orbital blood (~300 μl) was obtained under anesthetized conditions with chloral hydrate (400 mg/kg, i.p.) at 1600–1700 h. Serum (~100 μl) was separated by centrifugation at 4°C and stored at −80°C until assay. The levels of serum estradiol (E2), progesterone (P4), LH and follicle-stimulating hormone (FSH) were measured on the day of diestrus (group I). The measurement of T3, T4, thyroid stimulating hormone (TSH), thyrotropin-releasing hormone (TRH) and PRL had no limit for any estrous cyclicity (group II). The serum sample (5 μl) was needed for each assay (E2, P4, LH, FSH, PRL, T3, T4, TSH or TRH) using commercial enzyme-linked immunosorbent assay (ELISA) kits (Uscn Life Science Inc., Houston, TX, USA). The measurement of each sample was repeated 2 times to obtain an average value. The sensitivities were 47.1 pg/ml for T3, 1.4 ng/ml for T4, 19.3 pg/ml for TSH, 0.18 μIU/ml for TRH, 2.0 pg/ml for E2, 0.2 ng/ml for P4, 0.2 ng/ml for LH, 0.4 ng/ml for FSH and 0.4 ng/ml for PRL, respectively. The intra- and inter-assay coefficients of variation were 4.5% and 7.2% for T3, 4.3% and 7.5% for T4, 3.2% and 9.5% for TSH, 5.6% and 7.2% for TRH, 6.0% and 5.8% for E2, 5.8% and 8.4% for P4, 5.5% and 8.9% for LH, 4.3% and 10.3% for FSH, 4.7% and 4.9% for PRL. For determination of the LH surge, the repetitive blood sampling was undertaken at 1600, 1700 and 1800 h, respectively, on the day of proestrus (group III). The mice were anesthetized with ketamine (80 mg/kg) and xylazine (4 mg/kg) and a needle was inserted into a caudal vein at 1400 h. The mice were gently restrained in a cardboard tube, and the blood sample (~50 μl per time) was collected without anesthesia using heparinized syringes. After each blood collection, an equivalent volume of heparinized saline (5 U/ml normal saline; CP Pharmaceuticals Ltd, Wrexham, UK) was injected. Serum (~15 μl) was stored at −80°C for subsequent ELISA of LH.

The POA area (0.76 mm anterior to Bregma and 0.50 mm posterior to Bregma; Mayer and Boehm, 2011), the AVPV area (0.50 mm anterior to Bregma and 0.02 mm posterior to Bregma) at proestrus and the ARC area (−1.46 mm anterior to Bregma and −1.70 mm posterior to Bregma; Marraudino et al., 2017) at diestrus were collected from the frozen slices (200 μm thick) of brain using 16-gauge stainless steel tubing, and then stored at −80°C until assay. Total RNA of POA, AVPV or ARC regions was isolated using Trizol reagent (Invitrogen, Camarillo, CA). RNA (1 μg) was reverse-transcribed into cDNA using a Prime-Script RT reagent kit (Takara) for quantitative PCR (ABI Step One Plus) in the presence of a fluorescent dye (SYBR Green I; Takara). The synthesized cDNA was stored at −20°C until qRT-PCR was performed. The following primers were used for real-time PCR as described previously (Xi et al., 2011): GnRH F-GGGAAAGAGAAACACTGAACAC, R-GGACAGTACATTCGAAGTGCT; kiss1 F-GAATGATCTCAATGGCTTCTTGG, R-TTTCCCAGGCATTAACGAGTT; GAPDH F-ACCACAGTCCATGCCATCAC, R-TCCACCACCCTGTTGCTGTA. All samples were run in triplicate for each gene and for GAPDH (as housekeeping gene). There was no difference in GAPDH expression among the groups. The relative expression of genes was determined using the 2^−∆∆ct method with normalization to GAPDH expression. On the basis of melting curve analyses, there were no primer dimers or secondary products formed.

Levothyroxine (L-T4; Sigma-Aldrich Inc., St. Louis, MO, USA) dissolved in 0.9% saline was subcutaneously (s.c.) injected at dose of 20 μg/kg/day (Cao et al., 2017). Quinpirole (Quin; Sigma-Aldrich Inc., St. Louis, MO, USA) was dissolved in 0.9% saline and injected (i.p.) at dose of 2 mg/kg (Zhang et al., 2016).

Kisspeptin-10 [Kp-10, KiSS-1 (112–121)/metastin (45–54; human)] (Sigma-Aldrich Corp) was dissolved in DMSO, and then was diluted by 0.9% saline to a final concentration of 0.5% DMSO. For repeated intracerebroventricular (i.c.v.) injection of kisspeptin-10, the mice were anesthetized with chloral hydrate (400 mg/kg, i.p.) and then placed into a stereotaxic instrument (Stoelting, Wood Dale, IL, USA). A small hole (2 mm diameter) was drilled in the skull using a dental drill. A guide cannula (26-gauge, Plastics One, Roanoke, VA, USA) was implanted into the right lateral ventricle (0.3 mm posterior, 1.0 mm lateral and 2.5 mm ventral to Bregma) and anchored to the skull with three stainless steel screws and dental cement. On day 3 after surgery, the dummy cannula was removed from the guide cannula, and replaced by infusion cannulas (30 gauge) connected by polyethylene tubing (PE10; Becton Dickinson, Sparks, MD, USA) with a stepper-motorized micro-syringe (Stoelting, Wood Dale, IL, USA). The kisspeptin-10 (1 nmol/3 μl; Gottsch et al., 2004) was injected daily for successive 7 days. The injection of kisspeptin-10 was given at 30 min after TCS administration. This dose was selected on the basis of the previous report that kisspeptin-10 potently elicits the LH secretion (Navarro et al., 2004). The mice treated with injection (i.c.v.) of vehicle (0.9% saline) were served as the control group.

Mice were ovariectomized (OVX) under the anesthetized conditions with intramuscular injections of ketamine (80 mg/kg) and xylazine (4 mg/kg) at day 43 of TCS exposure. After surgery, the OVX mice received a subcutaneous implant of a silastic tubing (1.57 mm inside diameter; 3.18 mm outside diameter; 10 mm in length; Dow Corning, Midland, MI, USA) that was filled with 20 μg/ml of E2 in olive oil. The E2-treatment for 5 days produced a physiological level of serum E2 (20.71 ± 6.25 pg/ml) in adult female mice. On day 6 after implanting silastic tubing, the mice were given the injection (s.c.) of E2 at doses of 100 μg/kg for two consecutive days to produce a preovulatory high level of E2 (314.38 ± 66.91 pg/ml) that could exert a positive feedback action in AVPV-kisspeptin neurons (Wang et al., 2014). The silastic tubing filled with vehicle was implanted in OVX mice as the control group.

All group data in the Figures 1–5 are expressed as the mean ± standard error of the mean (SEM); all group data in Tables 1, 2 are expressed as the mean ± standard deviation (SD). All statistical analyses were performed using SPSS software, version 16.0 (SPSS Inc., Chicago, IL, USA). When analyzing one-variable experiments with more than two groups, differences among means were analyzed using one-factor analysis of variance (ANOVA) followed by Bonferroni post hoc tests. Differences at P < 0.05 were considered statistically significant.

Female mice (12 weeks old) were treated with the oral gavage of TCS at 1, 10 and 100 mg/kg/day for 50 days, hereafter referred to as 1-TCS, 10-TCS, or 100-TCS mice (Figure 1). In comparison with controls (0.95 ± 0.28 ng/ml), the urinary TCS levels were increased by approximately 4-fold in 1-TCS mice (3.82 ± 1.59 ng/ml), 20-fold in 10-TCS mice (21.09 ± 5.78 ng/ml) and 40-fold in 100-TCS mice (40.82 ± 4.48 ng/ml), respectively. One-way repeated measures ANOVA revealed that TCS exposure did not affect body weight (F_(3,76) = 1.129, P > 0.05; Figure 2A): mean body weights in 1-TCS mice (37.40 ± 2.67 g), 10-TCS mice (37.69 ± 2.78 g) and 100-TCS mice (36.60 ± 2.49 g) were not significantly different than control mice (37.81 ± 2.24).

The estrous cycle was monitored using the vaginal smear test. As shown in Figure 2B, mice with regular 4–5 day cycles consisting of 1 day in proestrus followed by 1 day in estrus and 2–3 days in diestrus (including 1 day of metestrus) were called “regular cyclers”. Estrous cycle length (F_(3,76) = 18.31, P < 0.01; Figure 2C), and specifically diestrus length (F_(3,76) = 24.17, P < 0.01; Figure 2D) was altered by the TCS exposure. Approximately 65% of 10-TCS mice (P < 0.01) and 90% of 100-TCS mice (P < 0.01) had a persistent diestrus (lasting more than 5 days) starting from the 2nd week of TCS exposure.

Ovary weights of 1-TCS mice (7.48 ± 1.55 mg), 10-TCS mice (7.69 ± 2.18 mg) and 100-TCS-mice (7.92 ± 1.44 mg) were not significantly different than control mice (8.12 ± 1.40 mg, P > 0.05, n = 20). To examine the possible influence of TCS on follicle development and ovulation, we counted the number of antral follicles and corpora lutea at diestrus using morphological criteria. As shown in Figure 3A, ovaries from control mice appeared typical for the diestrus stage. Numerous corpora lutea were observed in these ovaries, some showing clear evidence of recent ovulation. The numbers of antral follicles (F_(3,36) = 5.675, P < 0.05; Figure 3A-i) and corpora lutea (F_(3,36) = 3.371, P < 0.05; Figure 3A-ii) were affected by the TCS exposure, where the 10-TCS mice and 100-TCS mice showed a significant reduction in the number of antral follicles (P < 0.05) and corpora lutea (P < 0.05).

To explore the underlying mechanisms of TCS-caused persistent diestrus and ovary dysfunction, we measured the levels of reproductive hormones (Figure 1). At diestrus, the levels of serum FSH (F_(3,36) = 3.349, P < 0.05; Table 1), LH (F_(3,36) = 2.889, P < 0.05), P4 (F_(3,36) = 4.035, P < 0.05) and PRL (F_(3,76) = 3.376, P < 0.05) and GnRH mRNA (F_(3,36) = 9.422, P < 0.01), were all affected by TCS exposure. Specifically, 10-TCS mice and 100-TCS mice showed a modest but significant decrease in the levels of LH (P < 0.05 and P < 0.01), FSH (P < 0.05), P4 (P < 0.05) and GnRH mRNA (P < 0.01) compared to controls, which were associated with an obvious increase in the level of PRL (P < 0.05). Although the TCS exposure had a tendency to decrease the serum E2 level, this difference did not reach statistical significance (F_(3,36) = 2.083, P > 0.05). A surge-like LH release (LH surge) was observed between 1600 and 1700 in proestrus control mice and 1-TCS mice, but not in 10-TCS mice and 100-TCS mice (Figure 4A).

To further investigate the targets of TCS-reduced HPG axis and LH surge production, we examined kisspeptin expression in AVPV of proestrus mice and ARC of diestrus mice. In proestrus control mice, the kisspeptin+ cells in AVPV (AVPV-kisspeptin+) were located along the third ventricle (Figure 4B). In comparison with control mice, the immunoreactivity of AVPV-kisspeptin+ cells in 10-TCS mice and 100-TCS mice was significantly reduced. In diestrus control mice, a large number of kisspeptin+ cells were observed in ARC (ARC-kisspeptin+ cells, Figure 4E). The immunoreactivity of ARC-kisspeptin+ cells in diestrus 10-TCS mice and 100-TCS mice was lower than that in control mice. In addition, either the level of AVPV-kiss1 mRNA at proestrus (F_(3,36) = 11.512, P < 0.01; Figure 4C) or the level of ARC-kiss1 mRNA at diestrus (F_(3,36) = 3.614, P < 0.05; Figure 4F) in 10-TCS mice and 100-TCS mice were significantly reduced. To avoid the influence of gonadal hormones and estrous cycles in kisspeptin expression, mice were OVX mice and then treated with E2 (100 μg/kg) to produce a preovulatory high level of E2 (Murphy, 2005). As shown in Figure 4D, the administration of E2 in control mice exerted a positive feedback regulation in the AVPV-kisspeptin expression (P < 0.01, n = 10). In comparison with OVX control mice, the level of AVPV-kiss1 mRNA in OVX 10-TCS mice was reduced by approximately 40% (P < 0.01, n = 10). Although the high dose of E2 could elevate the level of AVPV-kiss1 mRNA in OVX 10-TCS mice (P < 0.01, n = 10), their level of AVPV-kiss1 mRNA still was less than half of E2-treated control mice (P < 0.01, n = 10).

To further explore the underlying mechanisms of TCS-enhanced secretion of PRL, we measured serum total thyroid hormones (T3 and T4), TSH and TRH (Figure 1). As shown in Table 1; the levels of T3 (F_(3,76) = 5.170, P < 0.01) and T4 (F_(3,76) = 15.267, P < 0.01), TSH (F_(3,76) = 3.919, P < 0.05) and TRH (F_(3,76) = 3.882, P < 0.05) were each affected by TCS exposure. While T3 and T4 were substantially reduced in 10-TCS (P < 0.05) and 100-TCS (P < 0.01) mice, TSH (P < 0.05) and TRH (P < 0.05) were increased. Interestingly, the administration of L-T4 at the dose of 20 μg/kg for 20 days starting from day 30 of 10-TCS exposure (Figure 1) reduced the previously observed increases in the levels of TSH (P < 0.05, n = 10; Table 2), TRH (P < 0.05, n = 10) and PRL (P < 0.05, n = 10).

Thyroid hormone replacement in TCS mice corrected their hyperprolactinemia, next experiments were designed to examine whether the hyperprolactinemia causes the suppression of kisspeptin expression and hypothalamic-pituitary-reproductive endocrine. The results showed that the treatment of 10-TCS mice with L-T4 for 20 days was able to correct the decrease in the level of AVPV-kiss1 mRNA at proestrus (P < 0.05, n = 10; Figure 5A) or ARC-kiss1 mRNA at diestrus (P < 0.05, n = 10; Figure 5B), which was accompanied by a recovery of GnRH mRNA (P < 0.05, n = 10; Table 2), serum LH (P < 0.05, n = 10) and FSH (P < 0.05, n = 10).

Until now, there have been no commercially available the PRL receptor antagonists (Lan et al., 2017). To further determine the involvement of TCS-induced hyperprolactinemia in the down-regulation of kisspeptin-reproductive endocrine, we used a type-2 dopamine receptors agonist quinpirole (Quin), because a recent study (Nakano et al., 2010) has reported that TRH-induced PRL release is inhibited by the activation of type-2 dopamine receptors. As expected, the administration of quinpirole (2 mg/kg) for 20 days in 10-TCS mice could prevent the increase in the level of serum PRL (P < 0.05, n = 10; Table 2), but it had no effects on the elevation of TSH (P > 0.05, n = 10) and TRH (P > 0.05, n = 10). Similarly, the administration of quinpirole for 20 days in 10-TCS mice restored the levels of AVPV-kiss1 mRNA (P < 0.05, n = 10), ARC-kiss1 mRNA (P < 0.05, n = 10) and GnRH mRNA (P < 0.01, n = 10), as well as serum LH (P < 0.01, n = 10; Table 2) and FSH (P < 0.05, n = 10).

To demonstrate that TCS-induced hyperprolactinemia through reducing kisspeptin expression suppresses the reproductive endocrine, the 10-TCS mice were given the injection (i.c.v.) of the GPR45 agonist kisspeptin-10 (Kp-10) for 7 days (Figure 1). The results showed that the application of kisspeptin-10 was able to correct the decline of GnRH mRNA (P < 0.05, n = 10; Table 2), LH (P < 0.05, n = 10) and FSH (P < 0.05, n = 10) without altering the increased levels of TSH (P > 0.05, n = 10) and TRH (P > 0.05, n = 10), or PRL (P > 0.05, n = 10).

To confirm whether the TCS-suppressed kisspeptin neurons caused persistent diestrus and ovary dysfunction, 10-TCS mice were treated with L-T4 or quinpirole for 20 days after examined persistent diestrus (Figure 1). The results showed that treatment with L-T4 or quinpirole could return the diestrus lengths to normal (P < 0.05 and P < 0.01, n = 10; Figure 5C) and recover the numbers of antral follicles (P < 0.05, n = 10; Figure 5D) and corpora lutea (P < 0.05, n = 10; Figure 5E).