Insulin Reduces Reaction of Follicular Granulosa Cells to FSH Stimulation in Women With Obesity-Related Infertility During IVF

Abstract

Context

Women with obesity usually need larger doses of FSH for ovarian stimulation, resulting in poor outcomes; however, the mechanism is still unclear.

Objective

To investigate the molecular regulation of FSH receptor (FSHR) expression associated with obesity.

Design

Case-control study to improve in vitro fertilization (IVF) outcomes.

Patients

Women with obesity (82) and women who were overweight (457) undergoing IVF and 1790 age-matched controls with normal weight from our reproductive medicine center.

Main Outcome Measures

FSHR expression was decreased in parallel with body mass index (BMI), whereas the estradiol (E₂) level on the human chorionic gonadotropin (hCG) trigger day was significantly lower.

Results

FSHR expression in human granulosa cells (hGCs), both mRNA (P = 0.02) and protein (P = 0.001) levels, was decreased in women who were overweight or obese. Both insulin (P < 0.001) and glucose (P = 0.0017) levels were positively correlated with BMI in fasting blood and follicle fluids (FFs) but not with FFs leptin level. We treated human granulosa-like tumor cells (KGN) cells with insulin; E₂ production was compromised; the level of phosphorylated (p)-protein kinase B (p-Akt2) decreased, whereas p-glycogen synthase kinase 3 (GSK3) increased; and there were similar changes in hGCs from women with obesity. Stimulated hGCs from women with obesity with compound 21 (CP21), an inhibitor of GSK3β, resulted in upregulated β-catenin activation and increased FSHR expression. CP21 also increased the expression of insulin receptor substrate 1 and phosphatidylinositol 3-kinase (PI3K), as well as p-Akt2.

Conclusions

Women with obesity in IVF were associated with reduced FSHR expression and E₂ production caused by a dysfunctional insulin pathway. Decreased FSHR expression in hGCs from women with obesity and insulin-treated KGN cells could be rescued by an inhibitor of GSK3β, which might be a potential target for the improvement of the impaired FSH-stimulation response in women with obesity.

The increasing epidemic of obesity with its related metabolic disorders has become a worldwide public health problem. Human obesity is a result of increased energy intake and decreased energy expenditure, resulting in a considerable increase in adipose tissue that is generally harmful to health. Obesity negatively affects female reproductive health, including increased risks of menstrual dysfunction, anovulation, and other fertility problems. Furthermore, there is increasing evidence that the success rate with ovulation induction and assisted reproductive technology (ART), such as in vitro fertilization (IVF), is generally lower in infertile women who are overweight or obese. Obesity increases the costs of infertility treatments with more doses of gonadotropins (Gns) needed, and reduces their effectiveness with lower embryo implantation and pregnancy rates (1–3) but higher miscarriage rates (4), which in turn, reduces conception rates following ART (5). Such women usually need larger doses of Gns for ovarian stimulation, but we still do not know if this is because of an elevated body mass index (BMI) or a result of their poor response to Gn stimulation (6–8).

In mammals, a mature antral follicle with a single oocyte enclosed by granulosa cells (GCs) develops from a primordial follicle, which is systematically controlled by subtle regulatory mechanisms. Normal proliferation of GCs is essential for follicular development, which in turn, influences the quality of oocytes and subsequent embryonic development. Several studies have shown that an altered follicular microenvironment in women who are overweight or obese might exert a detrimental effect on GCs function by inhibiting estrogen, particularly estradiol (E₂), production (9–12). According to the classic “two cells–two Gns” theory, FSH mainly affects the proliferation and differentiation of GCs. FSH binds to its receptor (FSHR) localized in GCs and activates intracellular cAMP and protein kinase A (PKA) to affect CYP19A1 (Cytochrome P450 19A1) gene-encoded aromatase expression via the activation of cAMP-responsive transcriptional regulatory proteins (13, 14). Although regulation of FSHR transcription is not well understood, studies with transgenic mice have indicated that FSHR transcription is also highly dependent on regulatory elements that lie distal to the promoter region (15, 16).

Despite the importance and insidiousness of obesity on female fecundity, the molecular underlying mechanism(s) of obesity-related reproductive dysfunction, which particularly impaired the effect of controlled ovarian hyperstimulation (COH), are yet to be fully elucidated. The etiologies of obesity-related infertility are myriad; however, one common underlying feature associated with obesity is hyperinsulinemia or even insulin resistance (IR). Insulin binding to its receptor initiates a signal transduction, resulting in subsequent activation of the pathway (17). This pathway includes activation of phosphatidylinositol 3-kinase (PI3K) signaling, which in turn, phosphorylates its substrate protein kinase B (Akt2) (18). A recent series of reports has revealed that FSH also can activate the PI3K/Akt pathway through a mechanism that may or may not involve PKA (15, 19, 20). Some of the molecular events downstream of PI3K/Akt2 activation by FSH in GCs have been elucidated, and one of the major substrates of Akt2 is glycogen synthase kinase 3β (GSK3β). Activated Akt2 inhibits GSK3β by phosphorylation of an N-terminal serine residue, thereby modulating various cellular processes, including glycogen metabolism (21). One substrate of GSK3β is β-catenin, a multifunctional protein that is both a structural component of cell–cell adhesion structures and an important signal transduction effector (22). In addition, FSH could stimulate the transcriptional activity of β-catenin by the promotion of the PKA-dependent or PI3K-independent phosphorylation pathway. Then, Akt2/GSK3β/β-catenin therefore represents an important point of convergence and crosstalk between the insulin and FSH pathways.

In this study, we hypothesize that in women who are overweight or obese, insulin pathway dysfunction may interfere with FSH-stimulated signal transduction, resulting in decreased production of E₂ and even interfere with the expression of FSHR. To test this hypothesis, we described the presence of metabolically important parameters in the serum and follicular fluid (FF) of women undergoing COH. Furthermore, we mainly focused on the exploration of the underlying mechanisms for BMI-related changes negatively linked with poor reproductive outcomes in women with obesity and attempted to find some rescue measure to increase their sensitivity to FSH stimulation.

Materials and Methods

Ethics approval

This study was approved by the Ethics Committee of the Third Affiliated Hospital of Guangzhou Medical University, and all of the participants gave their written, informed consent about the use of clinical data, blood, and FF samples containing human chorionic Gn (hCG) undergoing IVF. All experimental procedures were approved by the Institutional Review Board, and signed, informed consent was obtained from each patient.

Study design and collection of IVF clinical data

This was a retrospective cohort study of fresh IVF cycles carried out from 2013 to 2016 at our Reproductive Medicine Center. All patients had normal karyotypes and were undergoing COH at our Reproductive Medical Center for the first IVF cycle. Women with non-ovarian infertile factors were recruited, including male or tubal factor, such as hydrosalpinx and fallopian tubal obstruction infertility, requiring ART treatment. Patients were excluded if they had polycystic ovarian syndrome, or damaged ovarian function before the study caused by surgery or the use of harmful medications. Patients were then grouped as normal weight (BMI 18.5–23.9 kg/m²), overweight (BMI 24–27.9 kg/m²), or obese (BMI ≥ 28 kg/m²), based on guidelines for Asian populations (Chinese, Singaporeans, and Indonesians). Only women aged 29–32 years were included in the study, which included 457 subjects who were overweight; 82 subjects with obesity; and 1790 age-matched subjects of normal weight. Patients in all three groups underwent a long treatment protocol with Gn-releasing hormone agonist administration in the midluteal phase at day 21. The baseline FSH levels in blood were <10 IU/L before the start of COH, followed by ovarian stimulation with recombinant human FSH (rFSH; follitropin alfa; Merck, Geneva, Switzerland) at day 3. Height and weight were measured, and fasting blood was collected on the initiation day of COH.

FFs and hGCs collection

Preovulatory ovarian FFs (60 who were overweight, 35 with obesity, with 60 age-matched women of normal weight) were collected during transvaginal ultrasound-guided oocyte retrieval. At the time of oocyte retrieval, a single FF aspirate per patient from the first large follicle (diameter >14 mm) was collected to prevent blood contamination. Only FF samples that were found to be free of blood upon macroscopic analysis were retained for further analyses. Each FF sample was centrifuged at 1000 g for 10 minutes, and the supernatant was collected and then stored at −80°C for further analysis. Isolation of hGCs was performed as reported previously (23). The pellets were resuspended in PBS with 0.2% hyaluronidase (MilliporeSigma, Burlington, MA) and incubated at 37°C for 30 minutes. The suspension was layered over Ficoll-Paque (GE Healthcare, Buckinghamshire, United Kingdom) and centrifuged at 700 g for 30 minutes. The hGCs were collected from interphases and were further washed with PBS before culture in DMEM/F12 medium (Thermo Fisher Scientific, Gibco, Grand Island, NY) containing 10% fetal bovine serum (FBS; Gibco) and 1% antibiotics (100 U/mL penicillin and 0.1 mg/mL streptomycin; Gibco) in six-well plates at a density of 2 × 10⁵ per well at 37°C in a humidified atmosphere containing 5% CO₂. All hGCs used in this study were precultured for at least 1 day before treatment.

E₂ assays

Concentrations of E₂ in FFs and hGCs culture medium were measured using a rodent E₂ ELISA test kit (R&D Systems, Abingdon, United Kingdom), according to the manufacturer’s instructions. Absorbance was read at 450 nm on an Epoch multivolume spectrophotometer system (BioTek, Beijing, China). The sensitivity limit of the assay is 1 pg/mL. The performance characteristics of the ELISA assays ranged from 4% to 10% for the intra-assay coefficient of variation (CV) and 7%–10% for the interassay CV.

Aromatase activity assay

Aromatase enzyme activity was determined in cell cultures by measuring the conversion of testosterone (T) into E₂, as described (24). The hGCs, collected from FF samples (15 overweight, 15 obese, with 15 age-matched normal weight), were counted and plated at a density of 3 × 10⁵ cells/well in DMEM/F12 (Thermo Fisher Scientific, Gibco) medium containing 5% FBS. After incubation for 24 hours, culture medium was replaced into phenol red-free DMEM (Thermo Fisher Scientific, Gibco) medium with 2% charcoal-stripped FBS (Biological Industries, BI, Cromwell, CT) containing the aromatase substrate T (100 nM; MilliporeSigma, Burlington, MA). After 48 hours of incubation, also along with rFSH, which could maintain the physiological characteristics of ovarian follicle GCs in vivo, the media were collected and stored at −20°C until further analysis. E₂ concentrations were measured using the method described above. Protein concentrations were determined by the Bradford method, with bovine serum albumin as the standard. Aromatase activity was calculated as picograms of E₂ synthesized per hour per milligram of protein and expressed as a percentage of control.

FSH, insulin, glucose, and leptin assays

Concentrations of FSH, insulin, glucose, and leptin in fasting peripheral serum and FF samples and glucose in human granulosa-like tumor cells (KGN) cell culture media were measured using a commercially available ELISA kit (R&D Systems, Abingdon, United Kingdom), according to the manufacturer’s instructions. Absorbance was read at 450 nm on an Epoch multivolume spectrophotometer system (BioTek, Beijing, China). The sensitivities for various assays were 0.1 mIU/mL for FSH, 0.1 mIU/mL for insulin, 1 mmol/mL for glucose, and 0.1 ng/mL for leptin, whereas the intra-assay and interassay CV both were <10.0%. Homeostatic model assessment index 2 for IR (HOMA2-IR) and HOMA percent insulin secretion (HOMA%S) were both calculated using the software download from website of the Diabetes Trials Unit (The Oxford Centre for Diabetes, Endocrinology and Metabolism, Oxford, United Kingdom; https://www.dtu.ox.ac.uk/homacalculator/download.php).

Insulin treatment model of KGN cells

For in vitro studies to evaluate the mechanism of insulin on FSH-stimulated signal pathways and E₂ production, we used a steroidogenic human granulosa-like tumor cell line KGN. KGN cells are undifferentiated and maintain the physiological characteristics of ovarian cells and normal expression of the functional FSHR and CYP19A1 gene-encoded aromatase (25). KGN cells were grown in DMEM/F12 (Gibco), supplemented with 10% FBS (Gibco) and 1% antibiotics (100 U/mL penicillin and 0.1 mg/mL streptomycin; Gibco), and cultured at 37°C under 5% CO₂. Cells were equally distributed onto a six-well plate and rinsed with PBS (Gibco) after 12 hours of incubation with full medium. Then cells were serum starved for 12 hours in DMEM/F12 medium with 0.5% charcoal-stripped FBS (Biological Industries, Cromwell, CT) and exchanged for fresh medium. According to the range of insulin and FSH concentrations detected in FFs, KGN cells were stimulated with human recombinant insulin (MilliporeSigma) at concentrations of 0.01, 0.05, 0.1, and 0.5 μg/mL, and full medium was used as a positive control. To evaluate the effect of hyperinsulinemia or even IR on FSH-stimulated mRNA expression of FSHR signal pathway genes, KGN cells were treated with insulin (0.01, 0.05, 0.1, and 0.5 μg/mL) or without insulin (0) for 24 hours, together with a fixed concentration (10 IU/mL) of rFSH. To evaluate the effect of insulin on FSH-stimulated protein expression, KGN cells were preincubated for 48 hours with the same condition.

Culture and treatment of hGCs from women with obesity Using CP21

To evaluate the mechanism of GSK3β, especially acting via β-catenin on regulating FSHR expression, hGCs were treated with a GSK3β inhibitor, compound 21 (CP21; or CP21R7; Selleck Chemicals, Houston, TX) (26), dissolved in dimethyl sulfoxide. Primary hGCs from obese patients (n = 35) were plated in six-well plates at 2 × 10⁵ per well at 37°C under 5% CO₂. After starving for 12 hours, the cells were treated without (0 or dimethyl sulfoxide) or with CP21 at concentrations of 3, 5, 10, and 15 μg/mL, together with incubation of rFSH (10 IU/mL) for 24 hours to test the FSH-stimulated mRNA expression of FSHR signal pathway genes or for 48 hours to test the protein expression level.

RNA extraction and qRT-PCR

Total RNA was isolated from KGN cells and hGCs using Trizol-based standard extraction protocol (Thermo Fisher Scientific, Life Technologies, Waltham, MA), and the RNA concentration was quantified using a NanoDrop 2000 (Thermo Fisher Scientific). One microgram of total RNA was reversely transcribed in a 40-μL volume using the ReverTra Ace quantitative PCR (qPCR) RT MasterMix with genomic DNA remover (Toyobo, Osaka, Japan). Performed with AceQ qPCR SYBR Green Master Mix (Vazyme, Nanjing, China), qPCR reaction was carried out at a final volume of 20 μL containing 1 μL each cDNA, 10 μL 2× SYBR Green PCR Master Mix, 7.4 μL ultrapure water, and 0.4 μM sense and antisense primers to quantitate mRNA levels. The primers were designed with computer assistance, according to GenBank (27). Negative controls, including a no-template control and a negative RT control, were included on every plate for every primer set. Cycling conditions were as follows: 15 minutes at 94°C, followed by 40 cycles of 15 seconds at 94°C, 30 seconds at 60°C, and 45 seconds at 72°C. A final extension was performed for 10 minutes at 72°C. Each sample in every group was measured in triplicate or quadruplicate. All amplifications were carried out on a StepOne Plus system (Thermo Fisher Scientific, Applied Biosystems). The specificity of qPCR amplification was verified by the performance of a melting curve analysis and agarose gel electrophoresis. Relative gene expression was calculated by the efficiency-corrected ∆∆Ct method, normalized to the level of glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA used as an internal standard. The expressions of the FSHR, luteinizing hormone receptor (LHR), Steroidogenic Acute Regulatory Protein (StAR), CYP17A1, and CYP19A1 genes were examined as steroidogenesis in hGCs and insulin receptor substrate (IRS-1), Akt2, GSK3β, and β-catenin were examined as markers of FSHR function.

Western blot analysis

The KGN cells and hGCs were rinsed with ice-cold PBS, lysed in radioimmunoprecipitation buffer (Thermo Fisher Scientific) containing phosphatase and protease inhibitors (Thermo Fisher Scientific) for 30 minutes on ice, then centrifugation was performed at 12,000 g for 15 minutes at 4°C. The supernatants were recovered, and the protein concentrations were measured using protein assay reagents (Thermo Fisher Scientific). Equivalent amounts of protein were separated by 4% to 10% SDS-PAGE and then electrophoretically transferred onto polyvinylidene fluoride membranes (Merck Millipore) after blocking in 5% nonfat dry milk in 50 mM Tris, 150 mM NaCl, and 0.1% Tris-buffered saline–Tween 20 (TBST) at room temperature for 1 hour and incubated with the appropriate primary antibody (27), diluted in TBST with 5% bovine serum albumin overnight at 4°C with agitation. After complete washing in TBST, films were incubated with peroxidase-conjugated IgG secondary antibody (antirabbit, 1:5000; Thermo Fisher Scientific) at room temperature for 2 hours, washed in TBST, and developed with SuperSignal enhanced chemiluminescence reagents (Thermo Fisher Scientific), according to the manufacturer’s instructions. Finally, the images were acquired using a ChemiDoc MP system (Bio-Rad Laboratories, Hercules, CA) and quantified using ImageLab software. Membranes were reprobed using different primary antibodies after stripping at room temperature for 15 minutes in stripping buffer (Thermo Fisher Scientific). The data from the Western blots represent the mean densitometry measurements taken from all individual experiments. The densitometry analysis of the phosphorylated forms of the proteins (e.g., p-Akt2, p-GSK3β, and p-β-catenin) was done relative to their total form, and the others were normalized against GAPDH as the control.

Statistical analysis

All results are shown as the means ± SD. With the use of GraphPad Prism 6 software, data were analyzed using the Student t test for paired comparisons and ANOVA for multiple comparisons. Significance was set at P < 0.05.

Results

Basic characteristics of patients with infertility and IVF cycle outcomes

With the limitation of the age bracket to patients aged 29 to 32 years, 2329 patients, including 457 with an overweight BMI; 82 with obesity; and 1790 age-matched women of normal weight undergoing IVF for the first time, were analyzed. The basic patient characteristics are shown in Table 1. The data are presented as the means ± SD. With the adjustment for female age, the data for BMI, body surface area (BSA), basal FSH, luteinizing hormone (LH), E₂, T, and anti-Müllerian hormone levels were significantly different among the groups.

Table 1.

Demographic and Basic Reproductive Characteristics, According to Female BMI

Parameter	Normal (n = 1790)	Overweight (n = 457)	Obesity (n = 82)	P Value
	BMI 18.5–23.9 kg/m2	BMI 24–27.9 kg/m2	BMI ≥ 28 kg/m2
Age, y	29.91 ± 0.79	30.11 ± 0.80	30.24 ± 0.76	0.321
BMI, kg/m2	20.93 ± 1.45	25.44 ± 1.10	30.53 ± 3.51	0.002
BSA, m2	1.62 ± 0.10	1.75 ± 0.09	1.90 ± 0.15	0.005
FSH, U/L	5.58 ± 1.43	5.09 ± 1.34	5.24 ± 1.55	0.007
LH, U/L	3.49 ± 2.34	3.04 ± 1.68	2.84 ± 1.35	0.014
E2, pM	143.83 ± 82.02	121.89 ± 88.36	119.79 ± 74.35	0.003
T, nM	1.48 ± 0.81	1.54 ± 1.36	1.65 ± 0.89	0.019
AMH, ng/mL	4.75 ± 2.84	4.30 ± 2.88	4.06 ± 3.08	0.007
AFC	16.69 ± 7.07	16.41 ± 5.40	16.35 ± 4.43	0.239

Values are shown as means ± SD.

Abbreviations: AFC, antral follicle count; AMH, anti-Müllerian hormone; BSA, body surface area.

When oocyte numbers, FSH initial dosage and duration, and total dosage were compared among the different BMI groups, we found that more FSH was required for women who were overweight or obese, but fewer oocytes were retrieved when compared with the group of women with normal weight, shown in Table 2. We also found that the E₂ level, on the day of administration of hCG to trigger ovulation, was significantly lower, and the FSH dose administered per BSA was elevated along with higher BMI values.

Table 2.

IVF Cycle Characteristics, According to Female BMI

Parameter	Normal (n = 1790)	Overweight (n = 457)	Obesity (n = 82)	P Value
	BMI 18.5–23.9 kg/m2	BMI 24–27.9 kg/m2	BMI ≥ 28 kg/m2
FSH initiate dosage, IU	162.53 ± 47.68	181.82 ± 50.69	193.6 ± 52.82	0.002
FSH duration, d	12.38 ± 1.92	13.25 ± 2.41	13.27 ± 2.02	0.001
FSH dosage, IU	2295.70 ± 838.47	2707.90 ± 909.11	3030.30 ± 1015.54	0.001
FSH dosage per BSA, IU/m2	1440.28 ± 515.09	1550.20 ± 522.25	1598.90 ± 529.45	0.006
E2 rise, pM	12,948 ± 5096.31	11,501 ± 4801.42	10,999 ± 8054.99	0.004
E2 growth per BSA, pM	6.37 ± 3.09	4.80 ± 2.81	4.31 ± 3.47	0.001
hCG E2, pM	12,576 ± 5052.91	11,530 ± 4777.61	11,058 ± 8056.65	0.011
hCG >16 mm AFC	5.72 ± 5.18	5.36 ± 4.68	4.9 ± 4.43	0.049
Oocytes retrieved	13.83 ± 6.49	13.13 ± 6.73	12.18 ± 5.91	0.034

Values are shown as means ± SD.

Lower FSHR expression of hGCs in infertile women with obesity

First, we examined FSH concentrations in blood and FF samples in each IVF group. As shown in Fig. 1A, we did not find differences among the three groups. To determine whether the presence of elevated BMI was associated with aberrant function of the FSHR, the FSHR gene expression level in hGCs from women who were overweight or obese was compared with that of the normal-weight group. Expression of the FSHR in hGCs at both mRNA (P = 0.02) and protein (P = 0.001) levels, was significantly lower in patients with elevated BMI compared with normal-weight controls (Fig. 1B).

Figure 1.

Elevated BMI associated with reduced FSHR expression and lower E₂ concentrations with declined FSH-related gene expression in hGCs from women with infertility during IVF. (A) FSH concentration in blood (b) and FFs. (B) The comparison of FSHR mRNA and protein levels in hGCs among the three groups and the most representative image of Western blotting. (C) E₂ concentrations in FFs from women with infertility were correlated with elevated BMI. (D) Steroidogenesis-related genes, such as LHR, CYP17A1, and CYP19A1 mRNA expression levels, were decreased in hGCs from women with obesity rather than the group of women with normal weight without affecting StAR mRNA expression. Data were normalized against the corresponding levels of GAPDH mRNA. Values represent means ± SD of quadruplicate trials relative to the mean value of the control. N, normal-weight group; OB, obesity group; OW, overweight group. *P < 0.05 different from control, **P < 0.01 significantly different from control.

Steroidogenesis-related genes, such as LHR, CYP17A1, and CYP19A1 mRNA expression levels, were decreased in hGCs from women with obesity-related infertility without affecting StAR mRNA expression

To determine whether the presence of elevated BMI was associated with the compromised production of E₂ in women undergoing COH, E₂ concentrations in FF samples were compared among the overweight, obese, and control groups. The E₂ level in FFs was significantly decreased with elevated BMI (Fig. 1C). To elucidate the underlying mechanism by which obesity might interfere with E₂ production, we examined mRNA expression of steroidogenesis-related genes in hGCs. We found that the mRNA expression level of steroidogenesis-related genes, such as LHR, CYP17A1, and CYP19A1, shows a statistically significant decrease in hGCs from women with obesity (P < 0.05). It is worth mentioning that the CYP19A1 gene mRNA expression was significantly lower in hGCs from women with obesity, rather than the other groups (P < 0.01; Fig. 1D).

Hyperinsulinemia or even IR exist, both in fasting blood and FFs from women with obesity

Analysis of fasting blood and FF constituents on hCG trigger day showed that increased BMI was correlated with elevated levels of insulin (P < 0.001) and glucose (P = 0.0017; Fig. 2A and 2B) but not with FF-leptin (P = 0.172; Fig. 2C). As shown in Fig. 2D, HOMA2-IR levels in women with obesity were elevated (r = 0.8904; P = 0.0001), with decreasing insulin sensitivity (r = −0.8660; P = 0.0002), suggesting that hyperinsulinemia or even IR may be present in peripheral blood and FFs from women with obesity.

Figure 2.

Fasting blood (b) and FF insulin, glucose, and leptin level in women with varying BMI. Increasing BMI was associated with increased (A) b-insulin (r = 0.7229; P = 0.0001) and FF-insulin (r = 0.6465; P = 0.0002), (B) b-glucose (r = 0.7216; P = 0.0005) and FF-glucose (r = 0.4645; P = 0.0017), and (C) b-leptin (r = 0.5768; P = 0.001) and FF-leptin (r = 0.2326; P = 0.1721). (D) HOMA2-IR and HOMA%S were calculated using the HOMA2 model download software [available from Oxford Centre for Diabetes, Endocrinology and Metabolism (www.OCDEM.ox.ac.uk)], and the HOMA2-IR levels in women with obesity were elevated (r = 0.8904; P = 0.0001) with decreasing insulin sensitivity (r = −0.8660; P = 0.0002).

FSH-stimulated expression of aromatase protein level and enzyme activity were both decreased in hGCs from women with obesity-related infertility

We firstly examined the effect of obesity on the protein expression level of CYP19A1-encoded aromatase using Western blot analysis. The protein level of aromatase in hGCs from women who were overweight or obese was obviously lower than in the normal-weight group (P < 0.01; Fig. 3A). To assess FSH-related aromatase enzyme activity, hGCs were incubated for 48 hours with rFSH stimulation, which also could maintain the physiological characteristics of ovarian follicle GCs in vivo. Then, we found that even treated with rFSH, the aromatase protein level and enzyme activity were both significantly reduced in hGCs from women who were overweight, especially women with obesity compared with hGCs from normal weight (P < 0.01; Fig. 3A).

Figure 3.

Expression of FSH- and insulin-responsive protein in hGCs from women who were normal weight, overweight, and who had obesity. (A) FSH-stimulated expression of aromatase protein level and enzyme activity was decreased in hGCs from the women with obesity and infertility. Aromatase activity as the percentage of control (±SD) in hGCs between the group of women who were overweight and the group of women with obesity. (B) The reduced p-Akt2 and increased p-GSK3β level, showing dysfunctional insulin signal pathway in hGCs from the women with obesity-related infertility. *P < 0.05 different from control, **P < 0.01 significantly different from control.

Insulin-responsive expression of p-Akt2 was reduced with activated GSK3β increased in hGCs from women with obesity-related infertility

We also examined the p-Akt2 and p-GSK3β as activated markers of the FSH-stimulated insulin signal pathway. As shown in Fig. 3B, hGCs from women who were overweight or obese showed reduced p-Akt2 and increased p-GSK3β levels, suggesting a dysfunctional insulin signal pathway existed in hGCs from infertile women who were overweight or obese undergoing COH treatment.

Hyperinsulinemia or even IR plays an important role in the poor response to FSH stimulation using KGN cell models

To elucidate that the underlying mechanism by hyperinsulinemia or even IR interferes with FSH stimulation and FSHR expression, we examined the regulation on the FSH pathway in KGN cells by insulin treatment to simulate the follicular microenvironment of hGCs. According to the insulin concentration detected in FFs, we chose different doses of insulin, without or with a normal rFSH supplement, to stimulate the cells. We found that KGN cells would take up less glucose as a result of exposure under higher prolonged levels of insulin concentration in vitro, which will be reflected as a higher level of glucose and could be detected in the culture media (Fig. 4A). Next, we analyzed the ability of insulin and rFSH-stimulated KGN cells to produce E₂ and found that rFSH could stimulate E₂ production, whereas the E₂ level was obviously deceased with an elevated insulin concentration (Fig. 4B). Similar to the results of hGCs, hyperinsulin could decrease protein levels of FSH-stimulated CYP19A1-encoded aromatase and FSHR (Fig. 4C– E). Of note, we found that the p-Akt2 was decreased along with increased p-GSK3β in these KGN cells used for hyperinsulinemia or even an IR model (Fig. 4C, 4F, and 4G), consistent with the results obtained with cultured hGCs.

Figure 4.

Hyperinsulinemia or even IR inhibits FSHR expression and the FSH-stimulated signal pathway of insulin-treated KGN cells. (A) KGN cells were treated with different concentrations of insulin together with rFSH, as 0.01 g/mL compared with the insulin level in FFs of women with normal BMI and 0.05 vs OW and 0.5 vs OB. The glucose level in the supernatant of KGN cell culture medium. (B) The E₂ level in supernatant of KGN cell culture medium. (C) Western blot analysis was performed among different concentrations of insulin-treated groups. GAPDH was used as a loading control. (D and E) Expression levels of FSHR and aromatase protein in KGN cells with different concentrations of insulin-treated groups. (F and G) The p-Akt2 and p-GSK3β protein in KGN cells with different concentrations of insulin-treated groups. t, total. *P < 0.05 different from control, **P < 0.01 significantly different from control.

GSK3β is essential for reduced FSHR expression of hGCs from women with obesity and infertility

Although dysregulated Akt/GSK3β signaling is clearly in both obese hGCs and KGN cells treated as the hyperinsulinemia or even IR model, whether the insulin and FSH pathways can interact or synergize in hGCs to follicular development or in abnormal pathological conditions, such as obesity, remains unknown. To clarify the underlying mechanism in which GSK3β interferes with FSH-stimulated E₂ production and FSHR expression, we used CP21 as a GSK3β inhibitor. When we treated hGCs from women with obesity with CP21, the GSK3β mRNA level was decreased, along with the time of culture and the increased concentration of CP21 (Fig. 5A). In addition, the mRNA expression level of β-catenin was increased along with the dose of CP21 (Fig. 5B), confirming that the activity of GSK3β was inhibited. We postulate that CP21 might inhibit activation of GSK3β to reduce the degradation of β-catenin, causing newly synthesized β-catenin to accumulate and translocate to the nucleus to activate target genes, including FSHR. To test this hypothesis, we examined the effects of CP21 on activation of β-catenin, using an antibody directed against different p-β-catenin sites. As shown in Fig. 4C, the stimulation of the hGCs with CP21 could inhibit p-GSK3β and increase the p-β-catenin at sites Ser552 and Ser675 (Fig. 5C–5E). As shown in Fig. 5F, treatment with CP21 significantly increased the FSHR expression level in hGCs from women with obesity. We also found that treatment of these hGCs with CP21 could increase the expression of IRS-1, PI3K, and p-Akt2, without affecting the expression of total Akt (Fig. 5C and 5G).

Figure 5.

GSK3β is essential for reduced FSHR expression of hGCs from women with obesity and infertility. (A) The hGCs were preincubated with different dosages of CP21, together with rFSH, and the GSK3β mRNA expression was decreased. (B) The CP21 increased mRNA expression of β-catenin. (C) Western blot analysis was performed among different dosages of CP21 groups; GAPDH was used as a loading control. (D) The p-GSK3β protein was decreased along with the increased dosage of CP21. (E) The expression level of phosphorylation site 552 and 675 of β-catenin was upregulated, along with the increased dosage of CP21, as transcriptional activation function to target genes includes FSHR. (F and G) The protein phosphorylation level of FSHR and IRS-1 among different dosages of CP21 groups. NC, non-specific control. *P < 0.05 different from control, **P < 0.01 significantly different from control.

Discussion

Here, we aimed to understand the GC-derived effects of women with obesity-related infertility on FSH-stimulated follicular development. We first conducted a detailed phenotypic assessment of poor FSH response through FSHR expression and FSHR signal pathway in FFs and hGCs from women who were overweight or obese and identified a notable reduction in FSHR protein production. Furthermore, we found that insulin related with obesity might play an important role in the underlying mechanism and the important effects of GSK3β on FSHR expression. Finally, we demonstrated that suppression of GSK3β in hGCs from such women could lead to partial overexpression of the FSHR protein, with β-catenin accumulating and being translocated to the nucleus for transcriptional activity.

Elevated BMI is one of the reasons for reduced fertility in women, and ART has become a standard treatment option for them. Although previous studies have shown that female obesity has a pernicious influence on COH response and outcomes in IVF, the mechanisms leading to a poor outcome in such cases are poorly understood.

Here, we first observed the effect of being overweight or obese on IVF treatment responses and outcomes and then attempted to reveal the possible mechanisms for women’s reduced response to FSH stimulation. With the consideration that ovarian function and female fertility are age related, we only included women aged 29 to 32 years. We found that the basic E₂ level and numbers of oocytes retrieved were lower in the overweight and obese groups than in the normal-weight group, indicating that follicular function might be impaired in these women with obesity. Then, we analyzed FSH doses, the number of days of stimulation, and E₂ levels and found that the women’s response to FSH decreased with elevated BMI, indicating that the need for FSH stimulation in such women was greater than in women with normal weight. Some previous investigations have indicated that there might be an optimal body weight for reproductive function in women and an increased risk of anovulatory infertility in either who are underweight or overweight/obese (7, 28). Reduced fecundity in women who are overweight might be related to multiple endocrine, adipokine, and metabolic alterations that affect follicle growth, embryo development, and implantation (2, 29). There is also evidence that E₂ levels can decrease with increased BMI (30–34). Although previous investigations have reported that elevated BMI is associated with the requirement for higher total doses (and/or duration) of FSH stimulation (35–40), the etiology of such reduced follicular function parameters is still not completely understood.

Here, we compared FSHR expression in hGCs among patient groups with different BMI values. We found that expression of the FSHR was significantly lower in hGCs from women with infertility who were overweight/obese at both mRNA and protein levels. We also demonstrated that E₂ levels in FF samples were significantly decreased with elevated BMI. To examine the underlying mechanisms by which obesity impairs E₂ production in hGCs, we examined expression level of steroidogenesis-related genes, such as LHR, StAR, CYP17A1, and CYP19A1, as well as aromatase enzyme activity. We found that elevated BMI was associated with reduced mRNA expression of CYP17A1 and CYP19A1, along with CYP19A1 protein expression level and aromatase enzyme activity in hGCs. Consistently, these findings suggest that obesity impairs the FSH-stimulated steroidogenic function of hGCs. Taken together, these results provide powerful support for the idea that a reduced FSHR expression level in hGCs from overweight/obese women would be responsible for the poor response to FSH stimulation and decreased E₂ production.

Insulin was initially identified for its roles in regulating carbohydrate, fat, and protein metabolism in muscle, liver, and adipose tissues. Basal plasma levels of insulin also correlate with body fat mass. However, studies from the last two decades have rapidly expanded the range of actions of insulin, including regulation of steroidogenesis in ovarian cells in vitro and in the stromal and follicular compartments of human and murine ovaries (41–45). Insulin levels were increased, along with those of glucose, in peripheral fasting serum and FF samples from overweight/obese women undergoing IVF, showing that hyperinsulinemia or even IR might arise from defective insulin action or signal pathway. To clarify the underlying mechanisms by which insulin interferes with FSH response, we examined the regulation of insulin on FSH reaction and FSHR transduction pathways using the KGN cell line in vitro study. Our results, observed in hGCs from women with obesity/overweight, show that insulin could decrease FSHR and aromatase protein levels. According to the insulin and FSH concentration detected in FF samples from women with infertility during IVF, we tried to simulate the microenvironment of follicle development to KGN cells in vitro. Consequently, the activation of insulin could be weakened with the decreased p-Akt2 and increased p-GSK3β levels, both found in hGCs from women with obesity and KGN cells treated with insulin. We found substantial increases in p-GSK3β levels in hGCs from obese/overweight women and in KGN cells.

GSK-3β is a proline-directed serine-threonine kinase that was initially identified as an enzyme that was able to deactivate glycogen synthase phosphorylation (46, 47). In addition to the inhibition of cellular responses to insulin (for example, inhibition of GSK3β is required for insulin stimulation of glycogen synthesis), this enzyme influences cell division, growth, and development as an endogenous inhibitor of canonical Wnt (Wingless-related integration site) signaling. The finding that systemic inhibition of GSK3β improves whole-body glucose homoeostasis (48–50) implies that GSK3β exerts a tonic inhibitory effect on glucose metabolism. Importantly, GSK3 has pleiotropic roles in WNT signaling by activating its pathway at the receptor level, whereas also acting as a negative regulator of β-catenin, and this makes it a problematic target clinically (51–53). Therefore, p-GSK3β by Akt results in the hypophosphorylation, stabilization, and accumulation of β-catenin, which subsequently translocates to the cell nucleus and associates with various transcription factors to modulate the transcriptional activity of specific target genes, such as CYP19A1 and FSHR. To explore the underlying effects of GSK3β on reduction of FSHR expression, we chose an inhibitor to suppress the activity of GSK3β. CP21 is a potent and selective GSK3β inhibitor that can potently activate the canonical WNT signaling pathway (54). We found that CP21 could efficiently cause a decrease of p-GSK-3β (Tyr216) and an increase of p-Ser552-β-catenin and p-Ser675-β-catenin. These act in transcriptional activation with target genes, including FSHR and CYP19A1 (55). We also observed increased levels of the FSHR protein in hGCs from obese/overweight women after treatment with CP21, along with increased levels of IRS-1 and PI3K, suggesting that the insulin signal pathway might also be activated in hGCs from women with obesity by treatment with CP21. The effects on the improvement of the insulin pathway remain to be determined but are likely to be substantial. Furthermore, informative research into the underlying mechanisms responsible for BMI-related changes will help women with obesity with better COH outcomes.

Conclusions

We have demonstrated that elevated BMI was associated with reduced FSHR expression in hGCs; E₂ synthesis-related genes, especially CYP19; and decreased E₂ production. These might associate with insulin pathway dysfunction as decreased p-Akt2 in hGCs from women with obesity, in turn, activating GSK3β. Increased GSK3β activity leads to negative- regulated transcriptional activation of β-catenin (Fig. 6A). In hGCs from obese infertility women, the increased p-GSK3β could be inhibited by CP21 treatment, and transcriptional activation of β-catenin would be increased subsequently. As a result, the reduced FSHR expression associated with obesity could be rescued (Fig. 6B).

Figure 6.

Diagram illustrates how insulin reduces reaction of follicular hGCs to FSH stimulation in women with obesity-related infertility. (A) In hGCs from obese infertility women and insulin-treated KGN cells, the inhibition of Akt2 on GSK3β activity was reduced, which may affect the transcriptional activity of β-catenin, resulting in subsequent decreased FSHR expression. (B) The increased p-GSK3β could be inhibited by CP21, and the obesity-related FSHR expression could be rescued. Somehow, the insulin signal pathway could be improved by increased Akt2 activity and IRS-1 protein. Targeted drug, such as CP21, might be a potential therapy for improvement of the obesity-associated FSHR expression and FSH reaction in women with infertility undergoing IVF. CRE, cAMP-Response Element; P, phosphorylation; SF1, steroidogenic factor 1; TCF, transcription factor.

In summary, we support a potential mechanism by which periconceptional hyperinsulinemia, associated with obesity, could act on the E2 production and FSHR expression. This study offers insights into the etiology of obesity interference with the FSH pathway and provides a potential target for improvement of the effectiveness of COH treatment. Further research into the underlying mechanisms responsible for poor reproductive outcomes in women with obesity will help to improve the metabolic disorders, which put women with obesity at risk for poor reproductive performance.

Abbreviations:

Abbreviations:
 
• Akt2

  protein kinase B

 
• ART

  assisted reproductive technology

 
• BMI

  body mass index

 
• BSA

  body surface area

 
• COH

  controlled ovarian hyperstimulation

 
• CP21

  compound 21

 
• CV

  coefficient of variation

 
• CYP19A1

  Cytochrome P450 19A1

 
• CYP17A1

  Cytochrome P450 17A1

 
• E₂

  estradiol

 
• FBS

  fetal bovine serum

 
• FF

  follicle fluid

 
• FSHR

  FSH receptor

 
• GAPDH

  glyceraldehyde 3-phosphate dehydrogenase

 
• GC

  granulosa cell

 
• Gn

  gonadotropin

 
• GSK3

  glycogen synthase kinase 3

 
• hCG

  human chorionic gonadotropin

 
• hGCs

  human granulosa cells

 
• HOMA%S

  homeostatic percent insulin secretion

 
• HOMA2-IR

  homeostatic model assessment index 2 for insulin resistance

 
• IR

  insulin resistance

 
• IRS-1

  insulin receptor substrate 1

 
• IVF

  in vitro fertilization

 
• KGN

  human granulosa-like tumor cells

 
• LH

  luteinizing hormone

 
• LHR

  luteinizing hormone receptor

 
• p

  phosphorylated

 
• PI3K

  phosphatidylinositol 3-kinase

 
• PKA

  protein kinase A

 
• qPCR

  quantitative PCR

 
• rFSH

  recombinant human FSH

 
• T

  testosterone

 
• TBST

  Tris-buffered saline–Tween 20

Acknowledgments

The authors thank the investigators, staff, and participants of the studies for their valuable contributions.

Financial Support: Support for this work was provided by the Innovation of Science and Technology Commission of Guangzhou (Grant 201604020075 to J.-Q.L.); China Postdoctoral Science Foundation (Grant 2016M590767 to P.X.); and National Key Research and Development Program of China (Grant 2017YFC1001400 to D.-J.C.).

Disclosure Summary: The authors have nothing to disclose.