Dose-Dependent Insulin Regulation of Insulin-Like Growth Factor Binding Protein-1 in Human Endometrial Stromal Cells Is Mediated by Distinct Signaling Pathways

IGF binding protein-1 (IGFBP-1) is a major product of decidualized human endometrial stromal cells and decidua, and as a modulator of IGF action and/or by independent mechanisms, it regulates cell growth and differentiation and embryonic implantation in these tissues. IGFBP-1 secretion is primarily stimulated by progesterone and cAMP and is inhibited by insulin and IGFs. The signaling pathways mediating the latter are not well defined, and the current study was conducted to determine which pathways mediate the effects of insulin on IGFBP-1 mRNA and protein expression by human endometrial stromal cells decidualized in vitro by progesterone. Cells were cultured and treated with different combinations of insulin; wortmannin, an inhibitor of the phosphatidylinositide-3-kinase (PI3-kinase) pathway; and PD98059, an inhibitor of the MAPK pathway. IGFBP-1 mRNA was determined by real-time PCR, and protein secretion in the conditioned medium was measured by ELISA. Activation of the PI3-kinase and the MAPK pathways was assessed by the detection of phosphorylated AKT and ERK in Western blots, respectively. Insulin inhibited IGFBP-1 mRNA and protein secretion in a dose-dependent fashion, with an ED₅₀ for the latter 0.127 ng/ml (21.6 pm). Inhibitor studies revealed that at low doses, insulin acts through the PI3-kinase pathway, whereas at higher levels it also activates the MAPK pathway in the inhibition of IGFBP-1. The data demonstrate that human endometrium is a target for insulin action in the regulation of IGFBP-1. At physiological levels insulin likely plays a homeostatic role for energy metabolism in the endometrium, and in hyperinsulinemic states, insulin action on the endometrium may activate cellular mitosis via the MAPK pathway and perhaps predispose this tissue to hyperplasia and/or cancer.

HUMAN ENDOMETRIUM UNDERGOES cyclic changes throughout the menstrual cycle in preparation for implantation and menstruation, accompanied by endometrial cellular mitosis, differentiation, and apoptosis. Insulin action in the endometrium and how it may affect these processes are not well understood, although members of the related IGF family have been extensively studied in this tissue (1–4). IGF binding protein (IGFBP)-1 is abundantly expressed in decidualized endometrial stromal cells and is believed to play an important role during embryonic implantation by inhibiting IGF actions and/or by other IGF independent effects. For example, IGF-I mediates the effects of estradiol on endometrial cellular mitosis during the proliferative phase of the cycle (5), and IGF-II is believed to mediate progesterone’s effects in the secretory phase (5). IGFBP-1, a major product of stromal cells in secretory endometrium and pregnancy decidua, mostly inhibits IGF actions in cycling endometrium or may regulate trophoblast invasion, either directly or indirectly, during early pregnancy (6). IGFBP-1 is known to be inhibited by insulin in endometrial stromal cells (7) and liver (8, 9). Endometrial disorders and placental dysfunction are common in hyperinsulinemic states, such as obesity and polycystic ovarian syndrome, with increased risks of endometrial hyperplasia, infertility, and fetal growth disorders. However, little is known about the mechanisms of insulin action and signaling pathways involved in regulating IGFBP-1 production in endometrial stromal cells.

Insulin is a pleiotropic hormone, capable of activating various intracellular pathways, having multiple effects, including glucose transport, regulation of gene expression, protein synthesis, and cell division. On binding insulin, the insulin receptor undergoes receptor autophosphorylation and enhanced tyrosine kinase activity. Subsequently intracellular substrates, e.g. insulin receptor substrate-1, are phosphorylated on tyrosine residues that serve as docking sites for downstream Src homology 2 domain containing proteins, including the p85 regulatory subunit of phosphatidylinositide 3-kinase (PI3-kinase) (6, 10, 11). Insulin activation of PI3-kinase results in phosphorylation and thereby activation of AKT (12). AKT, also known as protein kinase B (PKB) or Rac, plays a critical role in controlling the balance between cell proliferation, cell survival, and apoptosis in a variety of cells (12–16). In addition to activation of the PI3-kinase signaling pathway, insulin activates other pathways, including the MAPK, also known as ERK pathway (17). MAPK consists of three different isoforms (17) and plays a crucial role in various pathways, mediating signals of growth factors and G protein-coupled receptors to their intracellular targets (18). In addition to the PI3-kinase-dependent and PI3-kinase-independent pathways associated with the insulin receptor, insulin can act on cells by binding to the IGF type I receptor with a lower affinity, compared with the IGF peptides. Therefore, insulin may potentiate the action of IGFs by decreasing the major IGFBP in decidua and thereby increasing bioavailable IGF peptides and/or by directly activating the IGF receptors. Herein we have investigated insulin signaling pathways involved in the regulation of IGFBP-1 secretion by decidualized human endometrial stromal cells. The data support insulin acting primarily through the PI3-kinase pathway and, at higher concentrations also via the MAPK pathway, in regulating IGFBP-1.

Materials and Methods

Tissue specimens

Endometrial tissues were obtained in accordance with the guidelines of the Declaration of Helsinki. Written informed consent was obtained from subjects, and the study was approved by the Stanford University Committee on the Use of Human Subjects in Medical Research. Some tissues were also obtained through the Cooperative Human Tissue Network (Cleveland, OH). Histologically normal endometrial tissue samples were obtained from cycling, premenopausal patients (27–44 yr old) undergoing endometrial biopsy or hysterectomy for benign reasons, such as fibroids, pelvic organ prolapse, pelvic pain, or abnormal bleeding. All patients had regular menstrual cycles (25–35 d), were documented not to be pregnant, had no history of endometriosis, and had not been on hormonal treatment for at least 3 months before tissue sampling. Samples were collected at room temperature in DMEM, transported to the laboratory, and processed as described below.

Cell cultures

Endometrial tissue was subjected to collagenase digestion, and stromal cells were separated from epithelium, cultured, and passaged as previously described (19). Cells were used at passages 2–4 for these studies. Stromal cells were grown to confluence in 6-well tissue culture plates (Costar, Cambridge, MA) in DMEM (Life Technologies, Inc., Invitrogen Corp., Grand Island, NY) supplemented with 5 μg/ml insulin (Sigma, St. Louis, MO) and 10% heat-inactivated charcoal-stripped fetal bovine serum (Gemini BioProducts, Woodland, CA). Confluent cultures were then decidualized in vitro with 10 nm estradiol, 1 μm progesterone, and 20 ng/ml epidermal growth factor in serum-free and insulin-free decidualization medium (75% DMEM, 25% MCDB-105, 50 μg/ml ascorbic acid, 1 mg/ml BSA, 5 μg/ml transferrin) for 12–14 d until decidualization was established, as determined by increasing IGFBP-1 secretion. The epithelial growth factor supplementation was stopped 2–3 d before initiation of experimental condition started. Confluent cells were then cultured for 4 d in decidualizing medium with insulin (0, 0.1, or 1 ng/ml) and with or without the PI3-kinase inhibitor, wortmannin (Calbiochem, La Jolla, CA), the MAPK inhibitor PD98059 (Calbiochem), or both enzyme inhibitors. The serum-free decidualizing medium was renewed after 2 d, as was the insulin, in the groups indicated. For the insulin dose-response curve, insulin concentrations of 0.05, 0.1, 0.5, 1, 5, 10, 50, and 100 ng/ml were used in the absence of inhibitors. Every 2 d the culture medium was renewed and the conditioned media were collected and stored at −20 C for further analysis. Cells were harvested for RNA analysis after 4 d of treatment.

IGFBP-1 measurements

IGFBP-1 levels were assayed in duplicate in 2-d conditioned media from duplicate cultures for each treatment group, and serial dilutions were used as needed. Total IGFBP-1 ELISA kits from Diagnostics Systems Laboratories, Inc. (Webster, TX) were used. The kits use an enzymatically amplified two-step sandwich-type immunoassay. Intraassay coefficients of variation were 4.6, 2.5, and 1.7% for IGFBP-1 concentrations of 7.86, 31.16, and 106.57 ng/ml, respectively. Interassay coefficients of variation were 7.5, 30.48, and 100.92% for IGFBP-1 concentrations of 7.6, 6.8, and 6.2 ng/ml, respectively. Cells were considered decidualized if IGFBP-1 secretion was between 200 and 500 ng/ml after 14 d of estrogen and progesterone treatment.

RNA extraction and reverse transcription

Total RNA was isolated from cultured endometrial cells using Trizol (Invitrogen, Carlsbad, CA) following the manufacturer’s protocol. Isolated total RNA was then treated with deoxyribonuclease and purified by RNeasy Spin columns (Qiagen, Valencia, CA). RNA integrity was verified by agarose gel electrophoresis/ethidium bromide staining and spectrophotometry (OD_260/280 absorption ratio greater than 1.95). Total RNA (1 μg) was reverse transcribed by using an Omniscript kit (Qiagen), according to the manufacturer’s instructions with a 1:1 ratio of oligo (deoxythymidine)_16–18 and random hexamers (Invitrogen).

Real-time PCR

Real-time PCR were performed in quadruplicate using the QuantiTect SYBR Green PCR kit (Qiagen) following the manufacturer’s instructions. Reactions were conducted in the Mx4000 Q-PCR system (Stratagene, La Jolla, CA). Primers for reference and target genes were designed using the PCR design software (http://labtools.stratagene.com) and synthesized by Qiagen. The constructs are as follows: 18-S forward, 5′-gtaacccgttgaaccccatt-3-′ and reverse, 5′-ccatccaatcggtagtagcg-3′, resulting with 151-bp amplicon; IGFBP-1 forward, 5′-ctatgatggctcgaaggctc-3′ and reverse, 5′-ttcttgttgcagtttggcag-3′, with amplicon size of 156 bp. All assays were optimized for primer concentrations. For both the target and reference primer pair, a ratio of 1:1 showed the best amplification efficiency. Primers were used at the concentrations of 150 nm. In a 25-μl reaction volume, 1 μg of sample RNA and 12.5 μl of 1× QuantiTect SYBR Green PCR mix (Qiagen), 5 μl of a PCR primer mix (150 nm), and 6 μl RNase-free water were added. The thermal cycling conditions included an initial activation step at 95 C for 10 min, followed by 40 cycles of denaturation, annealing, and amplification (95 C for 30 sec, 60 C for 1 min, 72 C for 30 sec). PCR products were analyzed by thermal dissociation (55–95 C) with a fluorescence measurement at every 1 C increment. For each assay, no-template and no-reverse transcription controls were included. To determine the PCR efficiency of the reactions, a series of dilutions of the template was performed to generate standard curves. For each primer pair, 10 standard curves were generated, averaged, and incorporated in the formula. The corresponding efficiency of amplification (Eff) during the exponential phase was calculated according to the equation Eff = 10^[−1/slope] − 1.

Quantitative analysis was based on the relative quantification of IGFBP-1 mRNA in the samples treated with 0.1 and 1 ng/ml insulin alone and with addition of 45 μm PD 98059 alone, 500 nm wortmannin alone, or both inhibitors, relative to the decidualized nontreated controls. This ratio was corrected by the corresponding ratio of the 18-S reference gene. The relative expression ratio (R) of IGFBP-1 was calculated based on the corresponding efficiencies of amplification for each treatment, compared with the control, and the differences in threshold cycle values (ΔCt) using the following mathematical model: R = (1 + Eff_IGFBP-1)^{ΔCt IGFBP-1 (control-treatment)}/(1 + Eff_18-S)^{ΔCt 18-S (control-treatment)}. In this equation, R represents the ratio at which a given gene of interest is expressed in treatment relative to the control, when normalized by a reference gene. ΔCtGOI-IGFBP-1 and ΔCt18-S are the differences in threshold values between treatment and control samples for the gene of interest and the normalizer. Ct values were calculated by the Mx4000 software based on fluorescence intensity values after normalization with an internal reference dye and baseline correction. All samples for the normalizer 18-S group were diluted 1:50, and all the gene of interest samples were diluted 1:4 right before each quantitative PCR experiment.

Western immunoblots

Decidualized stromal cells were pretreated with decidualization media with or without either wortmannin (500 nm) or PD98059 (45 μm) for 30 min and then treated with insulin in varying doses for 10 min. Cells were washed and then lysed with lysis buffer [50 nm Tris-HCl (pH 8.0), 150 nm NaCl, 1 mm EDTA, 10 nm sodium pyrophosphate, 50 mm benzamidine, 10 μg/ml trypsin inhibitor, 4 μg/ml aprotinin, 2 mm phenylmethylsulfonyl fluoride, 0.1% Triton X-100, 0.1% Nonidet P-40]. Cell lysates were centrifuged at 14,000 × g for 10 min at 4 C (Centrifuge 5415D Eppendorf; Brinkmann, Westbury, NY), and the supernatant was used for the Western blots. The protein concentration in the supernatant was measured using the Bradford protein assay (Bio-Rad, Hercules, CA) according to the manufacturer’s instructions. Thirty micrograms of protein were denatured using Laemmli buffer with dithiothreitol (Bio-Rad), and then loaded and run on a 10% SDS-PAGE (Gradipore, San Diego, CA). Proteins were subsequently transferred onto a nitrocellulose transfer and immobilization membrane (Schleicher & Schuell Bioscience, Keene, NH). The membrane was incubated in TBS-T [20 mm Tris-HCl, 100 mm NaCl (pH 7.60), and 0.1% Tween 20] containing 5% nonfat dry milk for 1 h to block nonspecific binding sites. After three washes in TBS-T, the membranes were incubated overnight with the primary antibody against phospho-AKT (1:1000 dilution; Cell Signaling Technology, Beverly, MA) and anti-phospho-ERK (1:5000 dilution; Sigma), respectively, in TBS-T containing 5% nonfat dry milk. After overnight incubation, membranes were washed three times with TBS-T and incubated for 1 h with a matching alkaline phosphatase-conjugated secondary antibody (Amersham Bioscience, Little Chalfont Buckinghamshire, UK) diluted 1:1000 in TBS-T containing 5% nonfat dry milk. After several washings with TBS-T, bound antibodies were detected using ECL+Plus chemiluminescent detection system (enhanced chemiluminescence detergents (Amersham Bioscience) and exposed to x-ray films (Eastman Kodak, Rochester, NY).

Statistical analysis

IGFBP-1 secretion in each experimental group was compared with non-insulin-treated decidualized controls in the same experiment. A total of five experiments is included for statistical analysis. Experiments were performed in duplicate or triplicate. Results are expressed as the percentage of IGFBP-1 production, compared with that of controls in the same experiment. The mean percent of control IGFBP-1 secretion in each experimental condition is shown with sem. The Student’s t test was performed when comparing IGFBP-1 secretion by cells treated with insulin (0.1 or 1.0 ng/ml) vs. those treated with inhibitors (PD98059, wortmannin, or both) and with the same dose of insulin. P < 0.05 was used for significance.

For the real-time PCR data, each treatment, which represents the normalized repeated measures of expression data from all relevant experiments, was tested for significance, compared with the control by repeated-measures ANOVA (P < 0.05), with post hoc analysis using the conservative Bonferroni method. Statistical analysis was performed using SYSTAT (version 10.2.01; SYSTAT Software Inc., Point Richmond, CA).

Results

Dose response of IGFBP-1 suppression

An insulin dose-response curve was investigated using endometrial stromal cells from three different subjects. The IC₅₀ for each and the mean IC₅₀ values are shown in Fig. 1. The IC₅₀ for the individual experiments varied from 0.05 to 0.5 ng/ml (8.5–85 pm). The average IC₅₀ was 0.127 ng/ml (21.6 pm) of insulin. The physiological, pulsatile pattern of insulin secretion in nonobese humans, measured in the peripheral blood, varies from 0.348 ng/ml (59 pm) to 1.74 ng/ml (296 pm) (20). Thus, doses of 0.1 ng/ml (17 pm) and 1 ng/ml (170 pm) were chosen for the subsequent experiments so that the experimental conditions could approximate the IC₅₀ with the lower dose and be significantly above it with the higher dose. As expected, the degree of suppression of IGFBP-1 was related to the dose of insulin used; however, doses of insulin above 5 ng/ml caused no further suppression.

Fig. 1.

Insulin dose-response curve (n = 3). Decidualized endometrial stromal cells were treated (in duplicate) with insulin for 4 d, using varying doses (0.05–10 ng/ml). IGFBP-1 secretion was measured in conditioned media and compared with no-insulin controls. Each experiment is shown individually with dotted lines, and the average is shown in bold. The IC₅₀ is the dose of insulin leading to a 50% reduction of IGFBP-1 secretion. The range was 0.05–5 ng/ml of insulin, and the average of all three was 0.127 ng/ml (21.6 pm).

Activation of PI3-kinase and MAPK pathways by insulin: dose response

Figure 2A demonstrates activation of AKT within 10 min of treatment with insulin, an effect that is dose dependent. A band at 57 kDa for the active, phosphorylated form of AKT was observed at the lowest dose of insulin, 0.1 ng/ml. Figure 2B demonstrates a dose-dependent response to insulin, with phosphorylation of ERK1 and ERK2, members of the MAPK family exhibiting typical double bands at 44 and 42 kDa. In contrast to the pAKT activation, phosphorylation of ERKs occurs at a higher insulin level (1.0 ng/ml).

Fig. 2.

Phosphorylation of AKT or ERK1/ERK2 in response to insulin. A, Phosphorylation of AKT in decidualized endometrial stromal cells is induced by various insulin doses (0–1000 ng/ml) within 10 min of insulin treatment. Samples were resolved by SDS-PAGE and sequentially immunoblotted with anti-phospho-AKT (upper panel) and reprobed with anti-AKT (lower panel) antibodies. The positions of pAKT and AKT are indicated. The experiment shown is representative of three experiments. B, Phosphorylation of ERK in decidualized human endometrial stromal cells in response to various insulin doses (0.1–1000 ng/ml) within 10 min of insulin treatment, occurring from 1 ng/ml insulin to the highest dose (1000 ng/ml) of insulin but not when cells are treated with insulin as low as 0.1 ng/ml. Samples were resolved by SDS-PAGE and sequentially immunoblotted with anti-phospho-ERK (upper panel) and reprobed with anti-ERK (lower panel) antibodies. The positions of pERK and ERK are indicated. The experiment shown is representative of three such experiments.

Inhibition of PI3-kinase and MAPK pathways through their specific inhibitors

When either PI3-kinase and/or MAPK inhibitors were added 30 min before the insulin treatment (1 or 10 ng/ml), phosphorylation of AKT or ERK1/ERK2 was inhibited (Fig. 3). No phosphorylation of AKT was observed in cells pretreated with wortmannin (500 nm) and PD98059 (45 μm) together, before the insulin treatment, as well as when wortmannin (500 nm) was administered alone (Fig. 3A). These findings are similar to the non-insulin-treated controls, in which no bands at 57 kDa were detectable. Figure 3B shows similar results for the activation of ERK1/ERK2, in which phosphorylation is prevented by pretreatment with PD98059 (45 μm) and wortmannin (500 nm) in combination as well as with PD98059 (45 μm) alone. The two bands at 44 and 42 kDa only were observed in the control group with treatment of 1 or 10 ng/ml insulin for 10 min without inhibitors and in the group that was preincubated with the PI3-kinase inhibitor, wortmannin.

Fig. 3.

Effects of insulin signaling pathway inhibitors on phosphorylation of AKT and ERK1/2 in decidualized human endometrial stromal cells. A, Phosphorylation of AKT induced by two different insulin doses is inhibited by pretreatment of cells with either the insulin inhibitor combination of wortmannin (500 nm) and PD98059 (45 μm) or wortmannin alone. The activation of AKT is not sensitive to PD98059 treatment alone. Samples were resolved by SDS-PAGE and sequentially immunoblotted with anti-phospho-AKT (upper panel) and reprobed with anti-AKT (lower panel) antibodies. The positions of pAKT and AKT are indicated. The experiment shown is representative of five such experiments. B, Insulin induces ERK phosphorylation that is sensitive to the MAPK inhibitor PD98059 (45 μm) alone and in combination with the PI3-kinase inhibitor wortmannin (500 nm). The activation of ERK is not sensitive to the wortmannin treatment alone. Samples were resolved by SDS-PAGE and sequentially immunoblotted with anti-phospho-ERK (upper panel) and reprobed with anti-ERK (lower panel) antibodies. The positions of pERK and ERK are indicated. The experiment shown is representative of five such experiments.

Effects of inhibitors on IGFBP-1 secretion

IGFBP-1 protein secretion into conditioned media decreased in response to increasing concentrations of insulin (Fig. 4A). The decline was 72 ± 8 and 91 ± 6%, compared with estrogen and progesterone alone, in the 0.1 and 1 ng/ml treatment groups (P < 0.05). Figure 4A also demonstrates that when wortmannin pretreatment was used to block activation of the PI3-kinase pathway, decreased IGFBP-1 secretion was observed only at the higher level of insulin treatment. The decline in IGFBP-1 secretion with wortmannin pretreatment plus 1 ng/ml insulin was 72%, compared with 91% in the 1 ng/ml insulin treatment alone. Figure 4B demonstrates that when PD98059, a chemical inhibitor of MAPK activation, was used before adding the insulin, a similar decline in IGFBP-1 production was observed [68 and 83% in the low- and higher-dose insulin treatments (P > 0.05)]. Figure 4C shows that when the two inhibitors together were used, IGFBP-1 production was similar to control production with the 0.1 ng/ml insulin treatment and was reduced by only 21% with the higher concentration of 1 ng/ml. No reversal of inhibition was seen in any groups when insulin doses over 1 ng/ml were used, even when cells were treated with both inhibitors (data not shown).

Fig. 4.

Effects of insulin signaling pathway inhibitors on IGFBP-1 secretion from human decidualized endometrial stromal cells. A, IGFBP-1 secretion by decidualized stromal cells treated with insulin (0.1 and 1.0 ng/ml) with or without pretreatment by wortmannin (n = 5 in duplicate). IGFBP-1 in conditioned medium is shown as mean (±sem) percentage of decidualized controls in the same experiment. IGFBP-1 secretion was lower in both insulin treatment groups, and pretreatment with wortmannin (500 nm) significantly increased the IGFBP-1 secretion, compared with insulin treatment alone. In the higher insulin treatment group, IGFBP-1 secretion was significantly lower than decidualized controls, even with wortmannin treatment. B, IGFBP-1 secretion by decidualized stromal cells treated with insulin (0.1 and 1.0 ng/ml) with or without pretreatment by PD98059 (n = 5 in duplicate). IGFBP-1 secretion is shown as mean (±sem) IGFBP-1 secretion as a percentage of decidualized controls in the same experiment. IGFBP-1 secretion was lower in both insulin treatment groups, and no difference was observed with PD98059 treatment (45 μm). C, IGFBP-1 secretion by decidualized stromal cells treated with insulin (0.1 and 1.0 ng/ml) with or without pretreatment by PD98059 and wortmannin (n = 5 in duplicate). IGFBP-1 secretion is shown as mean (±sem) IGFBP-1 secretion as a percentage of decidualized controls in the same experiment. IGFBP-1 secretion was lower in both insulin treatment groups. Treatment with both inhibitors produced near control levels of IGFBP-1 secretion at both insulin treatment doses.

Quantitation of IGFBP-1 mRNA regulation in response to inhibitors

Real-time PCR experiments with human decidualized stromal cells and using specific primers for IGFBP-1 revealed expression of IGFBP-1 in all samples investigated after treatment with estrogen and progesterone. Table 1 shows real-time PCR data for each investigated subject’s RNA as mean Ct values ± sd. Treatment of cells with insulin over 48 h resulted in a dose-dependent decrease in IGFBP-1 mRNA (Fig. 5A). A dose of 0.1 ng/ml insulin resulted in a marked (26- to 54-fold) decrease (P < 0.0001) in the three different samples examined for IGFBP-1 expression, compared with the untreated control cells. At 1.0 ng/ml insulin highly significantly inhibited IGFBP-1 levels from a 70-fold decrease to a 235-fold decrease with P < 0.0001, compared with the untreated control (Fig. 5A). A no-template control was used to verify the quality and cDNA specificity of the primers. The integrity and relative amounts of these mRNAs were confirmed using 18-S as a constitutively, abundantly expressed marker.

Fig. 5.

Real-time PCR data demonstrating IGFBP-1 mRNA expression in decidualized endometrial stromal cells treated with insulin for 48 h with or without insulin inhibitors. All data are compared with the no-insulin-treated control (insulin 0) and are shown as percent of control. A, Fold change of IGFBP-1 expression in cells treated with 0, 0.1, and 1 ng/ml insulin for 48 h. Each line represents a different subject. B, Percent of control IGFBP-1 expression in cells (from the same subject) that were preincubated with the insulin inhibitor wortmannin (500 nm) 30 min before the insulin treatment. C, Results from a group that was pretreated with PD98059 (45 μm) 30 min before the insulin treatment at two different doses. D, IGFBP-1 levels shown as percent of control in cells treated with both insulin inhibitors together [wortmannin (500 nm) + PD98059 (45 μm)] 30 min before the insulin treatment.

TABLE 1.

Real-time PCR data showing mean Ct values for the gene of interest, IGFBP-1 in the different treatment groups, and insulin doses applied for all three different subjects investigated

Group/insulin dose (ng/ml)	Primer	Mean Ct ± sd
		Patient 1	Patient 2	Patient 3
Control
0	IGFBP-1	16.43 ± 0.99	18.69 ± 0.15	15.16 ± 0.15
	18-S	14.21 ± 0.13	15.48 ± 0.06	12.39 ± 0.23
0.1	IGFBP-1	24.14 ± 0.07	24.00 ± 0.16	20.43 ± 0.14
	18-S	16.07 ± 0.08	15.27 ± 0.26	11.86 ± 0.35
1.0	IGFBP-1	27.82 ± 0.85	28.90 ± 0.31	24.86 ± 0.27
	18-S	16.70 ± 0.28	19.21 ± 0.12	15.83 ± 0.53
Wortmannin
0	IGFBP-1	16.47 ± 0.13	16.30 ± 0.18	16.62 ± 0.12
	18-S	13.94 ± 1.54	14.98 ± 0.64	15.29 ± 0.08
0.1	IGFBP-1	16.51 ± 0.14	16.36 ± 0.07	16.59 ± 0.47
	18-S	15.43 ± 1.20	15.06 ± 0.31	14.95 ± 0.15
1.0	IGFBP-1	20.88 ± 0.14	21.11 ± 0.10	21.07 ± 0.13
	18-S	15.26 ± 0.79	15.36 ± 0.20	15.10 ± 0.17
PD98059
0	IGFBP-1	15.08 ± 0.28	15.16 ± 0.07	15.22 ± 0.04
	18-S	13.25 ± 0.15	13.22 ± 0.13	13.35 ± 0.08
0.1	IGFBP-1	23.23 ± 0.05	23.18 ± 0.14	23.16 ± 0.11
	18-S	16.05 ± 0.06	15.99 ± 0.11	15.83 ± 0.13
1.0	IGFBP-1	27.13 ± 0.39	27.46 ± 0.21	27.48 ± 0.32
	18-S	15.48 ± 0.10	15.88 ± 0.16	15.31 ± 0.10
W+ PD
0	IGFBP-1	14.63 ± 0.19	14.32 ± 0.32	14.97 ± 0.22
	18-S	13.19 ± 0.14	12.70 ± 0.54	13.04 ± 0.18
0.1	IGFBP-1	14.69 ± 0.18	14.84 ± 0.14	14.35 ± 0.25
	18-S	12.88 ± 0.09	13.23 ± 0.23	13.18 ± 0.24
1.0	IGFBP-1	15.20 ± 0.25	14.86 ± 0.25	15.36 ± 0.10
	18-S	12.99 ± 0.17	13.21 ± 0.20	13.25 ± 0.29

In the low-dose insulin treatment group, a 30-min pretreatment with the insulin inhibitor wortmannin (500 nm) led to a slight, nonsignificant increase in IGFBP-1 expression of 1-fold or less, compared with the insulin-untreated control in two of the three samples (Fig. 5B). The IGFBP-1 expression in the non-insulin-treated control group was comparable with that in cells treated with the low insulin dose in addition to wortmannin treatment. Although the insulin inhibitor wortmannin was able to reverse the negative effect of insulin on IGFBP-1 expression in the lower insulin (0.1 ng/ml) group, it was not able to reverse the effect on IGFBP-1 in the high insulin (1.0 ng/ml) group, with IGFBP-1 levels significantly inhibited (P < 0.0001) (Fig. 5B). Pretreatment of the cells for 30 min with the inhibitor PD98059 (45 μm) reversed the insulin effects in neither the high-dose nor low-dose insulin treatment group (Fig. 5C), as did wortmannin (Fig. 5B). A significant, 38-fold decrease of IGFBP-1 was observed in the PD98059 and low-dose insulin-treated group (P < 0.0001), and a marked (4505-fold) decrease in IGFBP-1 (P < 0.0001) was observed in the cells treated with the PD98059 and the high insulin dose (Fig. 5C). Pretreatment with both insulin inhibitors, wortmannin (500 nm) and PD98059 (45 μm) 30 min before the two different insulin doses were added, resulted in complete reversal of IGFBP-1 inhibition at both insulin doses. The expression of IGFBP-1 in the group with the combination of inhibitors showed no significant difference, compared with IGFBP-1 expression in the untreated control group (Fig. 5D), in both the low- or higher-dose insulin treatment groups.

Discussion

Insulin is known to inhibit IGFBP-1 secretion from human endometrial stromal cells, although the pathway to achieve this regulation is not well defined. Insulin is a pleiotropic hormone, exhibiting different effects and activating different intracellular signaling pathways, depending on its target tissue. Herein we found that insulin activates PI3-kinase and MAPK in decidualized human endometrial stromal cells, as evidenced by rapid phosphorylation of AKT and ERK1-ERK2, respectively. The dose of insulin required for activation of the pathways suggests that PI3-kinase is sensitive to both high and low levels of insulin, whereas MAPK is activated at higher insulin levels, compared with PI3-kinase. Inhibition of either pathway alone did not achieve full reversal of insulin action on IGFBP-1 secretion or transcription. However, when the inhibitors of both pathways were used together, a complete reversal of the insulin signal was achieved, indicating that insulin can act through both pathways to achieve the inhibition of IGFBP-1, at higher doses of insulin. This is the first report to show that insulin inhibits IGFBP-1 expression and secretion by decidualized stromal cells in a dose-dependent manner through a synergy of two different signaling pathways and highlights the importance of examining multiple insulin signaling pathways together. The fact that chemical activation occurs within minutes, whereas it takes hours to days to see a result at the level of IGFBP-1 protein secretion, underscores the complexity of insulin signaling and cellular effects.

Several studies support the finding that the intracellular signals resulting from insulin binding to its receptor are dependent on dose and tissue specificity. In addition, whereas activating a particular pathway suggests a biologic effect, it is important to study a biological function, such as cell growth or protein secretion or other ligand-induced, specific effects. For example, Boileau et al. (21) investigated the biological effect of insulin and its pathways in human placenta, using a choriocarcinoma cell line, JAr, and the effects of wortmannin and PD98059. They found that although PKB and MAPK were phosphorylated in response to insulin, the PI3-kinase-dependent phosphorylation of PKB did not lead to stimulation of glucose transport or glycogen synthesis, whereas activation of MAPK did (21). Nguyen et al. (22) used PI3-kinase and MAPK inhibitors to show that oncogenic forms of the insulin receptor activate both pathways, although they are not equally required. Growth of oncogene-transformed cells was inhibited with the PI3-kinase inhibition but only partially inhibited by treatment with the MAPK inhibitor PD98059 (22). The effect of blocking both pathways in this system was not examined, and, therefore, the potential additive effect of the two pathways could not be evaluated. In contrast to our study, Poretsky et al. (23), using human granulosa cells, did not find that wortmannin abolished the inhibitory effect of insulin on IGFBP-1 in this system. However, higher doses of insulin were used in that study. Another study in the human ovary demonstrated that insulin stimulates testosterone biosynthesis by human thecal cells by activating its own receptor and using inositoglycan mediators as the signal transduction system (24). These results in different tissues of the reproductive system in addition to our findings in human endometrium indicate that insulin acts via different signaling pathways, and the choice of the pathways used is not only dose dependent but tissue specific as well.

Cumulatively, these studies demonstrate that it is unlikely that insulin effects throughout the human body are mediated by stimulating just one pathway. Furthermore, that our experiments were conducted using several insulin doses higher than 1 ng/ml and no effect of the inhibitors was found highlights the importance of determining the lowest dose of insulin that may have a biological effect. Insulin at higher doses may activate additional pathways or induce adaptive changes in the cells that render different responses. In the current study, it is likely that insulin acts through its own receptor because effects of AKT and ERK phosphorylation and IGFBP-1 inhibition were observed at concentrations far below those at which cross-reaction with the type I IGF receptor may be observed (25, 26).

These studies indicate a redundancy in the insulin-stimulated cellular signaling pathways in regulating IGFBP-1. The endometrial stromal cells used in this study were from subjects who were not insulin resistant. Indeed, whether the endometrium is insulin resistant in hyperinsulinemic women is not known. However, redundancy in insulin signaling pathways in human endometrium may protect it from becoming insulin resistant, making it a potential target of high insulin in the circulation in hyperinsulinemic states. This could result in decreased IGFBP-1 secretion and increased IGF action. These local changes in growth factors may directly affect endometrial function or growth and may comprise a mechanism underlying clinical conditions, such as endometrial hyperplasia, implantation disorders, and spontaneous abortions in women with insulin resistance and hyperinsulinemia.

This work was supported by the National Institutes of Health Cooperative Program on Trophoblast-Maternal Tissue Interactions (National Institute of Child Health and Human Development Grant UO1 HD42298, to L.C.G.) and the German Research Foundation (Deutsche Forschungsgemeinschaft Grant HE 3544/1, to A.P.H.).

First Published Online December 21, 2004

1

R.B.L. and A.P.H. are equal contributors to this work.

Abbreviations:

 
• ΔCt,

  Differences in threshold cycle values;

 
• Eff,

  efficiency of amplification;

 
• IGFBP,

  IGF binding protein;

 
• PI3-kinase,

  phosphatidylinositide 3-kinase;

 
• PKB,

  protein kinase B;

 
• R,

  relative expression ratio.