Our study group consisted of 20 infertile patients with PCOS without metabolic syndrome and 19 normally ovulating women diagnosed with male factor or tubal infertility who were undergoing their first cycle of in vitro fertilization (IVF) or intracytoplasmic sperm injection (ICSI) from September 2019 to August 2020 in the Reproductive Medical Center of Anhui Provincial Hospital. All patients provided written informed consent before participation, and all procedures of the study were approved by the human ethics committee of Anhui Provincial Hospital (approve ID: 20170108). The ages and body mass index (BMI) values of the two groups were comparable.

For each patient, FF samples were aspirated from follicles with a diameter $\ge$18 mm under the guidance of transvaginal ultrasound, and were all free of blood contamination. After oocytes were isolated, the FF samples were transported on ice to the laboratory immediately, and then centrifuged at 240 $\times$ g at 4 ${}^{\circ }$C for 10 min to isolate supernatant for TNF-$\alpha$ detection and granulosa cells for culture.

Twelve histological specimens of normal ovaries were retrieved from storage at the Department of Pathology of China-Japan Friendship Hospital. These ovarian tissues were obtained from patients of childbearing age who underwent hysterectomy and double appendectomy for invasive cervical cancer. The project was approved by the Ethics Committee of China-Japan Friendship Hospital (approval number: 2020-28-k20).

Granulosa cells from FF samples were washed twice with 3 mL culture medium (serum-free DMEM/F12). After centrifuged at 1000 $\times$ g at 4 ${}^{\circ }$C for 10 min, the supernatant fluid was discarded and the cells were resuspended. Cell viability and cell count were checked by 2% Trypan Blue staining. Then, the granulosa cells were diluted and cultured in DMEM/F12 (11320-033, Gibco, USA) containing 25 mM HEPES, 2 mM L-glutamine, and 10% fetal bovine serum in 6-well plates (Nunc, Thermo Fisher, USA) at 37 ${}^{\circ }$C with 5% CO${}_2$. When the cells adhered to the bottom of the plates, the medium was refreshed with serum-free DMEM/F12. As for E2 detection, 10${}^{-5}$ M testosterone was added as the substrate. Then, the cells were treated with varying concentrations (0, 0.2, 0.4, 0.8, 2.0, 4.0 ng/mL) of recombinant human TNF-$\alpha$ (Cat. 300-01A, PeproTech, Rocky Hill, NJ, USA). After 48 h, the medium was collected as conditioned medium, which was frozen at –70 ${}^{\circ }$C until E2 and inhibin assays were performed.

Granulosa cells with a concentration of 5 $\times$ 10${}^3$ cells/well were seeded in 96-well plates and treated with different concentration of TNF-$\alpha$ as described above. Cell viability was evaluated using a WST-8 Cell Counting Kit (CCK-8, Dojindo, Kyushu, Japan) per the manufacturer’s instructions at 24, 48, 72, and 96 h.

Four-micrometer-thick tissue sections were cut from paraffin-embedded tissue blocks, mounted onto glass slide, deparaffinized in xylene and rehydrated in graded ethanol series sequentially. Then the slides were immersed in 0.01 M citrate buffer (pH 6.0) buffer and placed in microwave oven for 10 min for antigen retrieval. After cooling to 20–28 ${}^{\circ }$C, endogenous peroxidase was blocked with 3% H${}_2$O${}_2$ for 10 min. For immunocytochemistry, granulosa cells were grown on coverslips to semiconfluency, fixed with cold acetone for 10 min and rinsed with PBS. After blocking non-specific sites with 5% normal goat serum for 1 h at 20–28 ${}^{\circ }$C, the sections were incubated overnight at 4 ${}^{\circ }$C in a humidified chamber with primary antibodies as follows: anti-TNF-$\alpha$ at a 1:200 dilution (Rabbit monoclonal [TNF/1500R], ab270264), anti-TNF receptor 1 at a 1:500 dilution (Rabbit polyclonal, ab19139), anti-TNF receptor 2 at a 1:200 dilution (Rabbit monoclonal [EPR1653], ab109322), and Ki-67 at a 1:200 dilution (Rabbit monoclonal [SP6], ab16667). All of the primary antibodies were purchased from Abcam (USA). For the negative controls, preimmune rabbit serum were used in place of the primary antibodies. After rinsing with PBS, the sections were sequentially treated with biotinylated secondary antibodies at 20–28 ${}^{\circ }$C for 1 h, and diaminobenzidine substrate chromogen system to visualize specific staining. For counterstaining, the sections were dipped in Mayer’s hematoxylin for 30 s.

The concentrations of TNF-$\alpha$ in the supernatants of FF and the concentrations of estradiol and inhibin in the conditioned medium were measured with ELISA kits according to the manufacturers’ instructions. The kit for TNF-$\alpha$ (Cat. #BMS223HS, Thermo Fisher Scientific, Waltham, MA, USA) and the kit for inhibin-$\alpha$ (Cat. SEA395Hu, Cloud-Clone Corporation, Wuhan, China) were sandwich enzyme immunoassay for quantitative determination while the kits for estradiol (Cat. #KAQ0621, Thermo Fisher Scientific) was a competitive binding immunoassay. Bring the strips and reagents of the kit and samples to 20–28 ${}^{\circ }$C before use. The serum-free medium and sample diluent were used as sample control and blank, respectively. Each sample, standard, sample control and blank were assayed in triplicate. At the end of reaction, all the wells were analyzed by spectrophotometry and read the absorbance of each well at 450 nm. The kit for TNF-$\alpha$ had a sensitivity of 0.13 pg/mL and precision (CV%) of 8.5% (intra-assay) and 9.8% (inter-assay). The sensitivity, intra-assay and inter-assay CV for the estradiol kit and inhibin-$\alpha$ kit were 5 pg/mL, 4.3%, 6.1% and 5.5 pg/mL, 10%, 12%, respectively. Standard curves and sample values were plotted using GraphPad Prism Software (Version 6.0, San Diego, CA, USA).

Total RNA was isolated with Trizol reagent (15596018, Life Technologies, Carlsbad, CA, USA) and reverse-transcribed into the first strand of cDNA by using SuperScript III reverse transcriptase (2680, TaKaRa, Kyoto, Japan). Then the cDNA were used for real-time PCR with specific primers (Table 1). Real-time PCR was performed on a CFX96 Touch instrument (Bio-Rad) using the iQ${}^{\text{TM}}$ SYBR Green Supermix kit (170-8880, Bio-Rad, USA). The cycling conditions of PCR included 95 ${}^{\circ }$C for 15 s, 60 ${}^{\circ }$C for 30 s, repeating 40 cycles. Data were normalized to GAPDH levels [19].

Antibodies specific to TNF receptor 1 (Rabbit polyclonal, ab19139, 1:1000), TNF receptor 2 (Rabbit monoclonal [EPR1653], ab109322, 1:1000), Caspase 3 (Rabbit monoclonal [EPR18297], ab184787, 1:1000), and aromatase (Rabbit polyclonal, ab18995, 1:500) were obtained from Abcam. Anti-GAPDH antibody (Rabbit monoclonal [14C10], #2118, 1:2000) was purchased from Cell Signaling Technology.

The expression of these proteins was detected by immunoblotting as previously described [20]. The granulosa cells were lysed using RIPA lysis buffer (Cat. P0013C, Beyotime Biotechnology, Beijing, China). The protein concentration of the whole cell lysates was measured with a bicinchoninic acid protein assay kit (Pierce Biotechnology). Proteins (30 $\mu$g) were mixed with 5$\times$ SDS-PAGE loading buffer, then loaded and separated on 10% SDS-PAGE gels and subsequently electrotransferred to polyvinylidene fluoride (PVDF) membranes (IPFL00010, Merck Millipore, Kenilworth, NJ, USA). After blocking with 5% nonfat milk for 1 h at 20–28 ${}^{\circ }$C, the PVDF membranes were incubated with primary antibodies on a rocker platform at 4 ${}^{\circ }$C overnight. Then, the PVDF membranes were washed with 0.1% Tween 20-containing Tris-buffered saline (TBS-T) three times, and incubated with HRP-linked Goat anti-Rabbit IgG antibody (#7074, Cell Signaling) at 20–28 ${}^{\circ }$C for 1 hour, and developed with chemiluminescent substrate (34080, Thermo) after washing with TBS-T four times. Autoradiography was performed with chemiluminescence image analysis system (Tanon 5200, Beijing, China).

Cell apoptosis were detected with an Annexin V-FITC/PI Apoptosis Detection Kit (Cat. BMS500FI, eBioscience). After treatment with varying concentrations of recombinant human TNF-$\alpha$ for 96 h, 2 $\times$ 10${}^5$ granulosa cells were harvested and stained with 5 $\mu$L of Annexin V-FITC and 5 $\mu$L of PI for 15 min in the dark at 20–28 ${}^{\circ }$C. The cells were then washed, resuspended with 1$\times$ PBS containing 2% FBS, and analyzed by flow cytometry (FACS Calibur, BD).

All the experiments were repeated three times. SPSS 19.0 software (IBM Corp., Chicago, IL, USA) was used for data analysis. The data are presented as the mean $\pm$ standard deviation. As the concentration of TNF-$\alpha$ in the follicular fluid from PCOS patients was quite different, the data were checked the normal distribution and variance homogeneity before comparisons. The statistical significance of differences between groups was performed by Student’s unpaired t test. One-way ANOVA was applied for multiple comparisons, further LSD test was used to inter-group comparisons. A value of p 0.05 was considered as statistical significance.

TNF-$\alpha$ detection with immunohistochemistry staining in specimens of normal ovaries showed that the level of TNF-$\alpha$ increased gradually in human follicles, in accordance with the development of follicles (Fig. 1A). However, a significantly higher TNF-$\alpha$ level in the FF was observed for the PCOS group than for the normal group (Fig. 1B), which to some extent contributed to granulosa cell apoptosis and follicular atresia.

Fig. 1.

TNF-$\alpha$ expression in human follicles. (A) Immunohistochemistry showed TNF-$\alpha$ expression in all stages of developmental follicles (1–4). 5, negative control. n = 12. (B) Quantitative detection of TNF-$\alpha$ in FF from healthy people (n = 19) and PCOS patients (n = 20). The data expressed as the mean $\pm$ SD. Bar = 20 $\mu$m. **, p 0.01, vs. health group.

TNF-$\alpha$ exerts its functions in the ovary through two types of receptors, TNFR1 and TNFR2. As shown in Fig. 2A,B, the expression of TNFR1 and TNFR2 showed no significant differences in human granulosa cells. Treatment with varying concentrations of TNF-$\alpha$ also caused no evident effect on the expression of TNFR1 and TNFR2. However, the expression of TNFR1 appeared to increase in accordance with follicular development, while the expression of TNFR2 remained constant at all stages in follicles (Fig. 2C).

Fig. 2.

TNFR expression in cultured granulosa cells and developmental follicles in ovarian tissues. (A) Transcript levels and (B) protein expression of TNFR1 and TNFR2 in granulosa cells after TNF-$\alpha$ treatment. n = 19. (C) Immunohistochemistry showing TNFR1 and TNFR2 expression in ovarian tissues. The data expressed as the mean $\pm$ SD. Bar = 20 $\mu$m. n = 12.

As shown in Fig. 3A,B, both mRNA and protein expression of P450 aromatase (P450arom), which catalyzes the formation of aromatic C18 estrogen from C19 androgens, was gradually inhibited by TNF-$\alpha$ in a dose-dependent manner (from 0.4 to 4 ng/mL) after 24 h of treatment. In addition, the concentration of E2 was decreased in the same way in the conditioned medium 48 h after TNF-$\alpha$ treatment (Fig. 3C). However, surprisingly, P450arom expression was significantly upregulated meanwhile E2 concentration was also increased by treatment with 0.2 ng/mL TNF-$\alpha$ for 48 h (Fig. 3A–C).

Fig. 3.

P450arom expression and E2 secretion in granulosa cells treated with varying concentrations of TNF-$\alpha$. (A) Relative mRNA levels and (B) relative protein levels of P450arom in granulosa cells treated with 0.2, 0.4, 0.8, 2, and 4 ng/mL TNF-$\alpha$ for 24 h. (C) E2 concentrations in the supernatants of cultured granulosa cells treated with 0.2, 0.4, 0.8, 2, and 4 ng/mL TNF-$\alpha$ for 48 h. The data expressed as the mean $\pm$ SD. n = 5. *, p 0.05; **, p 0.01, vs. untreated group.

In this study, our data showed that TNF-$\alpha$ can regulate inhibin A expression and secretion in granulosa cells in a dose-dependent manner, similar to the regulation of P450arom expression (Fig. 4A,D). The transcription of inhibin $\alpha$- and $\beta$A-subunit were elevated significantly in granulosa cells treated with 0.2 ng/mL TNF-$\alpha$ for 24 h and decreased dramatically as the dose of TNF-$\alpha$ increased (Fig. 4B,C). At a concentration of 4 ng/mL, inhibin A was significantly inhibited by TNF-$\alpha$. The expression of $\beta$B subunit mRNA, to some extent, was also altered by TNF-$\alpha$, although no significant difference was observed. Accordingly, the secretion of inhibin A was also markedly increased by 0.2 ng/mL TNF-$\alpha$ treatment for 48 h, but it decreased with increasing TNF-$\alpha$ concentrations (Fig. 4D).

Fig. 4.

Inhibin A expression and secretion in granulosa cells treated with varying concentrations of TNF-$\alpha$. Relative mRNA levels of subunits $\alpha$ (A), $\beta$A (B), and $\beta$B (C) of inhibin in granulosa cells treated with 0.2, 0.4, 0.8, 2, and 4 ng/mL TNF-$\alpha$ for 24 h. (D) Inhibin A concentrations in the supernatants of cultured granulosa cells treated with 0.2, 0.4, 0.8, 2, and 4 ng/mL TNF-$\alpha$ for 48 h. The data expressed as the mean $\pm$ SD. n = 5. *, p 0.05; **, p 0.01, vs. untreated group.

To assess the effect of TNF-$\alpha$ on granulosa cell growth, a CCK-8 assay was used to evaluate cell viability at different time points. The data showed that 0.2 ng/mL TNF-$\alpha$ significantly promoted cell proliferation in a time-dependent manner, while a high dose of 4 ng/mL led to evident cell inhibition. Compared with the untreated group, the group treated with the medium dose (0.8 ng/mL) showed no apparent difference in cell growth (Fig. 5A). Ki-67 in granulosa cells was expressed in a similar way after exposure to varying concentrations of TNF-$\alpha$ for 96 h (Fig. 5B). Furthermore, the apoptosis assay results showed that 4 ng/mL TNF-$\alpha$ induced significant apoptosis and upregulation of cleaved-caspase3 (Fig. 5C,D,E), which may have contributed to some extent to the notable decreases in E2 and inhibin A secretion.

Fig. 5.

TNF-$\alpha$ regulates granulosa cell proliferation and apoptosis. (A) Cell viability of granulosa cells treated with 0, 0.2, 0.8, and 4 ng/mL TNF-$\alpha$ for 24, 48,72, and 96 h. Ki67 expression (B), cell apoptosis (C), and cleaved-caspase 3 expression (D,E) in granulosa cells after treatment with 0, 0.2, 0.8, and 4 ng/mL TNF-$\alpha$ for 96 h. The data expressed as the mean $\pm$ SD. Bar = 20 $\mu$m. n = 9. **, p 0.01, vs. untreated group.