Female Sprague-Dawley rats weighing 220 ± 20 g, 8 weeks-old, were obtained from the Animal Center of Peking University Health Science Center with the certificate number SCXK 2006-0008. The animals were housed at 24 ± 1°C and relative humidity of 50 ± 1% under a 12 h light/dark cycle and given standard laboratory diet and water. The experimental procedures were in accordance with the recommendations of U.K. Animals (Scientific Procedures) Act, 1986 and associated guidelines, EU Directive 2010/63/EU for animal experiments. All animals were handled according to the guidelines of the Peking University Animal Research Committee. The experimental protocol was approved by the Committee on the Ethics of Animal Experiments of Peking University Health Science Center (LA2016314).

XYS granules was provided by Guangdong Yi Fang Pharmaceutical Co Ltd. (Guangzhou, China), which consists of radix bupleuri (chaihu) (16.2%), angelica (danggui) (16.2%), radix paeoniae alba (baishao) (16.2%), rhizoma atractylodis macrocephalae (baizhu) (16.2%), poria (fuling) (16.2%), ginger (shengjiang) (5.4%), mint (bohe) (5.4%), and glycyrrhizae (gancao) (8.1%). Antibodies against beta 2 adrenergic receptor (β2R), S6K I, phosphor-p70 S6K I (Thr229) and c-fos were purchased from Abcam (Cambridge, UK). Antibodies against Bcl-2 (B-cell lymphoma-2), Bax (B-cell lymphoma-2 associated X protein), cleaved caspase-3 (cleaved cysteinly aspartate specific proteinase-3), microtubule-associated protein light chain 3A (LC3A), LC3B, Akt and phosphor-Akt (Ser473) were purchased from Cell Signaling (Danvers, MA, USA). Antibody against dopamine beta hydroxylase (DβH) was purchased from Thermo Scientific (Rockford, IL, USA). (-)-Noradrenaline (NE) was from Sigma Chemical Co (St Louis, MO, USA). Pregnant mare serum gonadotropin was from ProSpec (Ness Ziona, Israel).

Animals underwent a week of adaptive breeding and estrous cycle determination, rats with normal estrous cycle were enrolled and divided into the following 6 groups randomly: (1) Control group, (2) Control+XYS (1.01 g/kg) group, (3) CUMS group, (4) CUMS+saline (NS) group, (5) CUMS+XYS (0.505 g/kg) group, and (6) CUMS+XYS (1.01 g/kg) group. The number of animals for determination of each parameter in each group is detailed in Table 1.

The dosage of XYS used in the experiments was calculated based on that for human, with 0.505 g/kg being equivalent to and 1.01 g/kg 2-fold the dose (27g crude drug/60 kg), respectively, used in clinic.

CUMS used in the present study consisted of 10 kinds of stressors: (1) 24 h food deprivation, (2) 24 h water deprivation with empty drinking bottles, (3) 24 h bedding deprivation, (4) 24 h wet bedding, (5) 24 h tilt cage (cages were tilted to 45° from the horizontal), (6) 2 min clamp tail (1 cm from the tail tip), (7) 2 h cold stress in low temperature incubator at 4°C, (8) 1 h shaker stress (140 rpm), (9) 5 min cold swim at 4–6°C, and (10) 3 h restraint. The animals in the CUMS group, CUMS+NS group, CUMS+XYS (0.505 g/kg) group and CUMS+XYS (1.01 g/kg) group were subjected to 3 weeks CUMS with two randomly selected stressors imposed on the animals every day. During the same period of time, the animals in Control group and Control+XYS (1.01 g/kg) group stood normally raised. In the week that followed, estrous cycle was determined for all animals, and only animals with disordered estrous cycle in CUMS groups were processed for treatment with or without XYS. XYS treatment started form the fifth week and lasted for 4 weeks, during this period, the animals in XYS treatment groups received XYS by gavage daily at a dose as indicated, while those in CUMS+NS group were given equivalent volume of NS in the same manner. XYS was suspended in NS before use. Estrous cycle was determined for each rat over the whole treatment period.

Female SD rats were administered by gavage with NS or XYS (1.01 g/kg) once a day for 3 days. 2 h after the final dose, the rats were sacrificed and blood was collected. The blood samples were centrifuged at 1,699 g for 15 min at 4°C, the serum was collected and incubated in a water bath at 56°C for 30 min for inactivation, and stored at −80°C before use (Cao et al., 2013).

Female SD rats postnatal 21 days were sacrificed by cervical dislocation under anesthesia and ovaries were excised 48 h after administration of pregnant mare serum gonadotropin (0.4 IU/g) by intraperitoneal injection. GCs were isolated by immersing the ovaries in 5% fetal bovine serum (FBS)-DMEM (dulbecco's modified eagle medium) and puncturing them with 26 gauge needles to release GCs (Abedini et al., 2016). GCs were seeded in 3.5 cm petri dish and cultured in 15% FBS-DMEM at 37°C with 5% CO₂ for 48 h, followed by incubation in 15% FBS-DMEM with NE (1 mM) (NE group) or without NE (Control group) for 24 h. A proportion of cells in NE group were transferred to 15% control rat serum-DMEM (NE+NS group) or 15% XYS (1.01 g/kg) rat serum-DMEM (NE+XYS (1.01 g/kg) group) incubating for additional 24 h.

Determination of the estrous cycle was performed daily for a week before CUMS and 5 weeks after CUMS using vaginal smears observed under a light microscope (Marcondes et al., 2002).

Rats were anesthetized by 20% urethane (10 ml/kg) at the end of the experiment. Ovaries were excised, rinsed in NS, fixed in 4% paraformaldehyde in 0.1 M phosphate buffer solution (pH 7.4). 5 μm paraffin sections were prepared and stained by hematoxylin and eosin (HE). The follicles were classified according to their hisotological characteristics (Cruz et al., 2012) and the number of primordial, primary, secondary, antral, cystic, and atretic follicles was counted for each ovary in a total of 10 sections with a 50 μm interval between the successive two.

At the end of the experiment, the rats were sacrificed and blood and ovaries were collected. The levels of NE in serum and ovarian tissue homogenate were determined by using ELISA kits (Eagle Biosciences, Nashua, NH, USA), and the levels of estradiol, progesterone and luteinizing hormone in serum were determined by using ELISA kits (R&D Systems, Minneapolis, MN, USA; Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer's instructions.

TUNEL staining was conducted for paraffin sections of rat ovary using an assay kit (Roche, Basel, Switzerland), according the manufactures instruction, and the nuclei were labeled with Hochest 33342 (Invitrogen, Camarillo, CA, USA). The number of the TUNEL positive cells in 5 fields was counted, and the average was calculated and expressed as cell number per field.

At the end of the experiment, the rats under anesthesia were infused via the left ventricle with 4% paraformaldehyde in 0.01 M phosphate buffer solution (pH 7.4), brain segments containing region of the locus coeruleus (bregma −9.80 to −10.04 mm) (Sabban et al., 2015) were removed and postfixed with the same fixative for 48 h, and immersed in 30% sucrose solution for at least 48 h at 4°C. The brain segments were cut into 10 μm sections with a cryostat microtome (CM1900; Leica, Nussloch, Germany). GCs were fixed with 4% paraformaldehyde in 0.1 M phosphate buffer solution for 20 min at the end of the experiment. The sections or cultured cells were permeabilized with 0.3% Triton X-100 for 30 min, then incubated with antibodies against β2R (1:50), LC3A (1:400), DβH (1:50) or c-fos (1:100) overnight at 4°C after blocked with goat serum, and then incubated with a fluorescent secondary antibody for 1.5 h at 37°C and Hochest 33342 (Invitrogen, Camarillo, CA, USA) for nuclei. Positive staining was examined by a laser scanning confocal microscope (TCS SP5, Leica, Mannheim, Germany).

Rats under anesthesia were sacrificed at the end of the experiment, the ovarian tissue and brain tissue containing region of locus coeruleus were removed and immediately frozen at −80°C till use. GCs were collected following incubation sequentially with NE and control or XYS rat serum. GCs, ovarian and brain tissue were homogenized in lysis buffer containing the protease inhibitor. Proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene fluoride (PVDF) membrane. After non-specific binding sites were blocked with 5% bovine serum albumin in Tris-buffered saline Tween (TBS-T), PVDF membranes were incubated overnight at 4°C with the primary antibodies against β2R (1:1000), Bax (1:1000), Bcl-2 (1:1000), cleaved caspase-3 (1:500), S6K I (1:1000), phosphor-S6K I (Thr229, 1:1000), Akt (1:1000), phosphor-Akt (Ser473, 1:2000), LC3A (1:1000), LC3B (1:1000), and GAPDH (1:2000) in diluent buffer (3% bovine serum albumin and 0.1% TBS-T). After rinsing with TBS-T for 3 times, PVDF membranes were incubated with secondary antibody (1:3000, Cell Signaling Technology, Danvers, MA, USA) for 1 h at room temperature and washed by TBS-T for 3 times. Antibody binding was detected by enhanced chemiluminescence detection kit (APPLYGEN, Beijing, China). Bands were scanned and evaluated by Bio-Rad Quantity One software for quantification (Bio-Rad, Richmond, CA, USA). All the western blot experiments were repeated at least 3 times.

All parameters were expressed as mean ± SE. Statistical analysis was performed using one-way ANOVA followed by Turkey test for multiple comparison. A probability less than 0.05 was considered to be statistically significant.

The estrous cycles of rats were assessed before and after CUMS. Figure 1A shows the representative profiles of estrous cycles in different groups. Rats of Control group (Figure 1Aa) and Control+XYS (1.01 g/kg) (Figure 1Ab) group revealed normal estrous cycles over the observation, while irregular estrous cycles were induced immediately after CUMS and persisted till the end of the experiment, if not interfered (Figure 1Ac). Post-treatment with XYS at 0.505 g/kg or 1.01 g/kg both recovered the irregular estrous cycles induced by CUMS starting from the third week of treatment (Figures 1Ad,e). Figure 1B is the quantification of the percentage of rats with normal estrous cycle during the post-treatment with XYS in different groups, which confirmed the results from the survey above.

Ovary morphology was examined by HE-staining and the representative images of different groups are presented in Figure 2. As compared to Control group (Figure 2a), numerous cystic follicles were observed in the ovaries of CUMS group (Figure 2b) and CUMS+NS group (Figure 2c). Noticeably, post-treatment with XYS (0.505 g/kg or 1.01 g/kg) significantly reduced the number of cystic follicles (Figures 2d,e). Illustrated in Figure 3A are the representative images of follicles at different developmental stages, and the percentage and number of the various follicles and corpus luteum are shown in Figure 3B and Table 2, respectively. At the end of CUMS, the number of both primary follicles and corpus luteum in the CUMS group was decreased, while the percentage of cystic follicles was increased. This tendency was further prominent 5 weeks after CUMS, accompanying with an increase in the percentage of primordial follicles and a decrease in the percentage of antral follicles. Post-treatment of XYS (0.505 g/kg and 1.01 g/kg) relieved the alternations mentioned above, suggesting the ameliorating effect of XYS on cystic follicles formation and follicle development abnormalities.

To examine the effect of XYS on the elevated activity of sympathetic nervous system induced by CUMS, NE level in rat serum and ovarian tissue was detected. Compared with Control group, the NE level in serum remained unchanged immediately after CUMS, but increased significantly at the end of the experiment (Figure 4A). On the other hand, NE level in ovarian tissue increased significantly after CUMS and decreased at the end of the experiment (Figure 4B). The alterations above were inhibited by post-treatment with XYS with significance.

Figures 4C–E present the results of the serum hormone level in different groups. Luteinizing hormone had no change immediately after CUMS challenge but increased at the end of the experiment if not interfered. This increase was protected by XYS treatment at 1.01 g/kg (Figure 4C). In contrast, the level of estradiol dramatically decreased after CUMS, but recovered completely at the end of the experiment, which was not affected by the treatment of XYS (Figure 4D). The level of progesterone in serum changed in response to CUMS challenge in a distinct manner in that it decreased markedly immediately after CUMS, which was persisted till the end of the experiment if not interfered, but prevented by treatment with XYS at 1.01 g/kg (Figure 4E).

To disclose the rationale behind the ameliorating effect of XYS on polycystic ovary formation and follicles development abnormalities, the expression of β2R in follicles was assessed by immunohistochemistry and western blot. Immunofluorescence staining (Figure 5A) shows that, compared with Control group, the expression of β2R increased in the primordial and primary follicles immediately after CUMS and further increased at the end of the experiment, which, however, was obviously relieved by post-treatment with XYS (Figures 5Aa1–e1,a2–e2). In accordance with the results in the ovarian tissue, NE exposure led to an increased expression of β2R in cultured granulosa cells (Figure 5B), which was attenuated by XYS. Western blot analysis (Figure 5C) shows a trend similar to immunofluorescence results, however there was no significant difference observed among the different groups.

In order to investigate whether the formation of cystic follicles is associated with apoptosis of granulosa cells, we conducted TUNEL staining for ovarian tissues. Compared with the Control group, the antral follicles, but not the secondary follicles, from CUMS group demonstrated a high number of apoptotic granulosa cells (Figure 6A). On the other hand, less TUNEL positive cells were observed in the antral follicles from rats after XYS post-treatment. A quantitative assessment confirmed this result showing that the number of TUNEL positive cells significantly increased immediately after CUMS challenge, but decreased spontaneously at the end of the experiment, a process that was significantly accelerated by treatment with XYS (Figure 6B). In line with this result, Western blot analysis revealed that the ratio of Bax and Bcl-2 and the expression of cleaved caspase-3 varied among groups in a similar manner as that for TUNEL positive cells (Figures 6C,D).

As an autophagy-related protein, LC3 is widely used as an autophagosome marker in mammalian cells (Klionsky et al., 2016). In view of the known involvement of autophagy in the development of follicles (Choi et al., 2010, 2013), we explored the expression of LC3 and the LC3-I to LC3-II conversion in the ovarian tissue by immunohistochemistry and Western blot. Compared with the Control group, the expression of LC3A in the granulosa cells of the antral and cystic follicles was elevated immediately after CUMS and at the end of the experiment, which was restored by treatment with XYS, however (Figure 7A). On the other hand, the conversion of LC3A-I to LC3A-II and LC3B-I to LC3B-II was increased as well at the end of the experiment, although this conversion did not change (LC3A-I to LC3A-II) or even decreased (LC3B-I to LC3B-II) immediately after CUMS. Nevertheless, the increased conversion of LC3 was inhibited significantly by post-treatment with XYS (Figures 7B,C).

In order to explore whether the increased NE level in serum and ovarian tissue induced by CUMS contributes to the autophagy of granulosa cells and the protective effect of XYS, we examined the expression and localization of LC3 in granulosa cells in vitro by immunofluorescence staining. As shown in Figure 8A, NE evoked an increase in the expression of LC3 in the cytoplasm of granulosa cells as compared with Control group, which was relieved by post-treatment with rat serum containing XYS (1.01 g/kg). The results of western blot analysis also confirmed that NE caused an elevated conversion of LC3A-I to LC3A-II and LC3B-I to LC3B-II in cultured granulosa cells (Figures 8B,C), which was inhibited by post-treatment with rat serum containing XYS (1.01 g/kg).

The mTOR pathway is known as a major suppressing signal of autophagy induction, which can be activated by PI3K/Akt signaling (Li et al., 2009; Qin et al., 2010). Moreover, both mTOR and PI3K/Akt pathway participate in the regulation of follicle development (Sobinoff et al., 2013). Thus, Western blot analysis was undertaken to assess the expression and phosphorylation of S6K I and Akt in ovarian tissues as well as in cultured granulosa cells from different groups. The results revealed that CUMS induced a decrease in phosphorylation of S6K I and Akt in ovarian tissues (Figures 9A,B), which was blunted significantly by post-treatment with XYS, suggesting the involvement of Akt/mTOR/S6K I pathway in attenuating effect of XYS on autophagy of ovarian tissues induced by CUMS. A similar result was observed in vitro in cultured granulosa cells exposed to NE (Figures 9C,D), hinting the key role of NE in CUMS challenge and the mechanism behind the effect of XYS.

Since locus coeruleus is the principal site for brain synthesis of noradrenaline, we thus detected DβH, the rate-limiting enzyme of NE synthesis, and c-fos protein, a marker of neuron activity, in locus coeruleus by immunohistochemistry and Western bolt. As shown in Figure 10, the expression of DβH and c-fos in locus coeruleus increased significantly at the end of the experiment, which was protected by post-treatment of XYS (1.01 g/kg), indicating the potential of XYS to regulate the activity of sympathetic nervous system.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer SH and handling Editor declared their shared affiliation.