Expression of endometrial glycogen synthase kinase-3β protein throughout the menstrual cycle and its regulation by progesterone

Abstract

Glycogen synthase kinase-3β (GSK-3β) is a serine/threonine kinase that plays a role in glycogen synthesis by inhibiting glycogen synthase (GS) through phosphorylation. We hypothesized that GSK-3β by virtue of its role in glycogen synthesis through the inhibition of GS will play a role in the preparation of the endometrium for blastocyst implantation. Immunohistochemical (IHC) analysis and Western blot analysis (WBA) detected GSK-3β in the endometrium, myometrium, Fallopian tube and ovary. WBA showed more than 5-fold higher endometrial expression of the phosphorylated GSK-3β (pGSK-3β) isoform (inactive) in the secretory phase as compared with the proliferative phase (P < 0.001), whereas no differences in total GSK-3β expression were detected. IHC analysis confirmed the WBA and showed marked expression of pGSK-3β predominantly in glandular epithelial cells in early and mid secretory endometrium with scant expression during the proliferative phase. In in vitro experiments using human endometrial-derived epithelial cell line (HES), progesterone did not alter total GSK mRNA or protein expression. However, progesterone induced a dose-dependent increase in the expression of pGSK-3β, which could be blocked by RU486. Cyclic expression of GSK-3β’s active and inactive forms in the endometrium suggests that sex hormones regulate the expression of this enzyme. In vitro experiments demonstrate that progesterone through receptor-mediated mechanisms induces phosphorylation of endometrial GSK-3β.

Introduction

The process of implantation requires a hospitable uterine environment. One of the key metabolic changes that occurs during the peri-implantation period and throughout pregnancy in discrete cell populations in the placenta is the rise in endometrial glycogen content (Milwidsky et al., 1980). Although during implantation this is catalysed by glycogen synthase (GS) and is regulated by progesterone (Shapiro et al., 1980; Secchi et al., 1987; Su et al., 1996), it is conceivable that this same regulation is operative later in the decidualized epithelium and the placenta and that glycogen plays a critical role in ensuring the development of a full-term normal weight pregnancy (Robertson et al., 1999; Yang et al., 2003). One mechanism by which the activity of GS may be regulated is through its phosphorylation by the enzyme glycogen synthase kinase-3 (GSK-3), a multifunctional serine/threonine kinase ubiquitously expressed in eukaryotic cells and existing in mammals as two isoforms (α and β) (Embi et al., 1980; Hemmings et al., 1981; Woodgett, 1990). Both isoforms through the phosphorylation of GS will render it inactive. Therefore, inhibition of GSK-3 isoforms is crucial for the activation of GS and the synthesis of glycogen.

The activity of GSK-3β is controlled through two alternative canonical pathways, the first of which reviewed elsewhere (Moon et al., 1997; Cohen and Frame, 2001; Dominguez and Green, 2001; Frame and Cohen, 2001; Woodgett, 2001) is through Wnt signalling and leads to a non-phosphorylative inactivation of GSK-3β, stabilization of β-catenin and transcriptional activation of TCF/LEF-responsive genes. An alternative mechanism regulating predominantly the metabolic and cell survival functions of GSK-3β is through an inactivating phosphorylation at serine 9 position for GSK-3β and serine 21 for GSK-3α (Cohen and Frame, 2001; Dominguez and Green, 2001; Woodgett, 2001). Several kinases may mediate this phosphorylation including PKA, PKC and AKT (PKB). The role of insulin in the inactivation of GSK-3α and GSK-3β and the subsequent increase in GS activity occurs through AKT (Cohen and Frame, 2001; Woodgett, 2001). Although most of the literature has concerned itself with the study of GSK-3β, many but not all of the roles of GSK-3β are shared with GSK-3α.

GSK-3β is a highly conserved enzyme with key roles in numerous biological processes reviewed elsewhere (Cohen and Frame, 2001; Woodgett, 2001). Some of the most studied are its roles in glucose and energy homeostasis, cell fate decisions such as body patterning of the Drosophila embryo (Siegfried et al., 1992; Siegfried and Perrimon, 1994), formation of the ventral–dorsal axis in the Xenopus embryo (Woodgett, 1994; Klein and Melton, 1996; Stambolic et al., 1996; Moon et al., 1997) and regulation of meiotic entry in yeast and possibly rodents (Bowdish and Mitchell, 1993; Bowdish et al., 1994; Malathi et al., 1997, 1999; Guo et al., 2003). Interestingly, GSK-3β inactivation is required and is rate limiting for progesterone-mediated Xenopus oocyte maturation (Fisher et al., 1999; Nebreda and Ferby, 2000). During oocyte maturation, both insulin and progesterone through non-genomic mechanisms induce serine phosphorylation and inactivation of GSK-3β (Sarkissian et al., 2004). At the level of the endometrium, the actions of GSK-3β are complex. Components of the Wnt-signalling pathway, with the exception of GSK-3β, are regulated transcriptionally in the various phases of the menstrual cycle suggesting a role in peri-implantation events (Tulac et al., 2003). Aside from a role in uterine development, canonical Wnt signalling with inhibition of GSK-3β and nuclear translocation of β-catenin is essential for the estrogen receptor (ER)-independent estrogen-regulated growth of epithelial cells of the endometrium (Hou et al., 2004). Other evidence suggests opposing effects of estradiol (E₂) and progesterone through the PI3-kinase-AKT-Ser-9 inactivation of GSK-3β on the proliferation of endometrial epithelial cells (Chen et al., 2005). Additionally, progesterone up-regulates Dickkopf-1 (DKK-1), a potent inhibitor of Wnt signalling in the stromal cells of the secretory phase human endometrium (Tulac et al., 2006).

The purpose of this study was to extend our knowledge of the role of these pathways in decidualization of the human endometrium and to test the hypothesis schematically depicted in Figure 1). On the basis of this hypothesis, progesterone regulates the activity of GSK-3β through phosphorylation, such that during the secretory phase when expected implantation occurs and endometrial glycogen stores are high, a greater proportion of GSK-3β is in the inactive (phosphorylated) form. This would result in a greater proportion of GS to be in the non-phosphorylated (active) state and hence would lead to increased glycogenesis. To test this hypothesis, we determined the endometrial expression of the inactive and total (predominantly active) forms of GSK-3β throughout the menstrual cycle and examined the influence of progesterone and its antagonist RU486 on in vitro expression of both forms of the enzyme in cultured endometrial [human endometrial-derived epithelial cell line (HES)] cell line. Our data confirm our hypothesis that progesterone through its receptor facilitates glycogen synthesis in the secretory phase human endometrium through the phosphorylative inhibition of GSK-3β.

Materials and methods

Specimens

Tissues used for this study were collected at two institutions including the University of Wisconsin, Madison, and Harbor-UCLA Medical Center Torrance. Institutional Review Board (IRB) approval at both institutions was obtained. In the case of tissues for WBA, samples were obtained from premenopausal women undergoing hysterectomy for benign gynaecologic indications, including pelvic organ prolapse, symptomatic myoma and dysmenorrhoea. The mean ages of these women were 36 and 39 years for proliferative and secretory phase, respectively. Tissues used in the immunohistochemical studies were provided by the department of Pathology at Harbor-UCLA Medical Center and were derived from women undergoing surgery for benign gynaecologic indications. Histological dating of tissues was done according to the criteria of Noyes et al. (1950).

Real-time RT–PCR

For real-time PCR, cDNA was generated with 0.5 µg of total RNA using Omniscript Reverse Transcription (RT) reagent (QIAGEN). The RNA was incubated in 20 µl of a RT reaction mixture [×1 RT buffer, 0.5 mM dNTP, 1 µM random hexamers, 10 U Rnasin (RNase inhibitor) and 4 U Omniscript reverse transcriptase] at 37°C for 1 h. The reverse transcriptase was inactivated by heating at 95°C for 5 min. PCR was performed in 96-well optical reaction plates (Applied Biosystems) on cDNA equivalent to 25 ng RNA in a volume of 25 µl, containing ×1 reaction buffer (EUROGENTEC); forward and reverse appropriate primers were selected from Assay on Demand (Applied Biosystems). Real-time PCR was performed for GSK and glyceraldehyde-3-phosphatedehydrogenase (GAPDH) gene, using the ABI-Prism 7700 Sequence System (Applied Biosystems) using the following conditions: 10 min at 95°C for one cycle and 15 s at 95°C, 1 min at 60°C for 40 cycles. Primer sequences for GSK were as follows: GSK sense: 5′ATTACGGGACCCAAATGTCA3′; GSK antisense: 5′ATTGGTCTGTCCACGGTCTC3′; GAPDH sense: 5′ATC ACTGCCACCCAGAAGACT3′; GAPDH antisense: 5′CATGCCACTGAG CTTCCCGTT3′.

Control PCR was done with water replacing cDNA. All the controls gave a threshold cycle (C_T) value of 40, indicating no detectable PCR product under these cycle conditions. ABI Sequence Detection System 1.6 software (Applied Biosystems) was used to determine the cycle number at which fluorescence emission crossed the automatically determined threshold level (C_T). All samples were run in triplicate.

The cycle at which each reaction reached threshold fluorescence was defined as the C_T for that reaction. The results were analysed using comparative method and the values were normalized to GAPDH RNA expression by subtracting mean C_T of GAPDH from mean target C_T for each sample, to obtain the mean ΔC_T. The mean ΔC_T values were then converted into fold change based on a doubling of PCR product in each PCR cycle, according to the manufacturer’s guidelines.

WBA

Cells and tissue were sonicated in protein lysis buffer and protein concentrations determined by the BCA Protein Assay kit (Pierce, Rockford, IL, USA). For each sample, 35 µg of protein was separated on a 7.5% polyacrylamide gel. The separated proteins were transferred electrophoretically to Immobilon-P membranes (Millipore Corp., New Bedford, MA, USA). Membranes were blocked for 2 h in a 5% milk buffer before overnight incubation with GSK-3β and phospho-GSK-3β antibodies at primary dilution of 1 : 1000 (Cell Signaling Technology, Inc., Beverly, MA). The blots were subjected to enhanced chemiluminescence (ECL Western Blotting Detection System, Amersham Corp., Arlington Heights, IL, USA), with enzyme conjugate anti-mouse IgG horseradish peroxidase as a secondary antibody. Blots were then exposed to autoradiography film. The resulting bands corresponding to ∼48 kDa were compared by scanning densitometry. To ensure equal loading, we stripped protein blots and reprobed them for GAPDH.

Immunohistochemistry

Tissues were fixed in formalin and embedded in paraffin, and 5 µm sections were cut and mounted onto slides. Sections were then de-paraffinized, rehydrated and pre-incubated with 3% horse serum in phosphate-buffered saline (PBS) for 20 min, followed by 4 h of incubation at room temperature with affinity-purified polyclonal antibodies (Cell Signaling Technology, Inc., Beverly, MA, USA) to total GSK-3β (1 : 50 dilution) and phosphorylated GSK-3β (pGSK)-3β (1 : 25 dilution). The IgG concentration for GSK-3β was 140 µg/ml and for pGSK-3β was 53 µg/ml. The antibodies used are specific to the two β isoforms of the protein and do not cross-react with GSK-3α but do cross-react with the human, rodent and rabbit GSK-3β protein. For signal detection, we used the avidin–biotin–peroxidase (ABC) technique with diaminobenzidine as the substrate. Sections were counterstained with haematoxylin. Negative control experiments were performed in all IHC studies by omitting the primary antibody and substituting it with normal rabbit serum. Positive control studies were performed with samples of adult rodent testes, shown previously to express GSK-3β in both the early meiotic prophase (preleptotenes, leptotenes) and Sertoli cells of the seminiferous tubules (Guo et al., 2003).

Cell culture

HES cells provided by Dr Douglas A. Kniss (The Ohio State University, Columbus, OH, USA) were cultured in Medium 199 (M199, GIBCO) supplemented with 10% fetal bovine serum, 10 mM l-glutamine and 0.5 mM l-Arg in a 5% CO₂-humidified atmosphere. Cells were used after three passages. At the time of experiments, cells were detached by trypsinization. About 1 × 10⁵ cells/well were plated into a six-well tissue culture plate and cultured in growth medium for 24 h. Experiments were carried out using at least 70% confluent cells in M199. After 24 h of serum starving, medium was replaced with experimental medium consisting of M199 with 2% charcoal-treated fetal bovine serum supplemented with 10 mM l-glutamine and 0.5 mM l-Arg. To determine the effects of progesterone and RU486 on GSK-3β expression, cells were treated with vehicle or progesterone (10^-8–10^–6 M) and harvested at 24 h. The progesterone antagonist RU486 (10^–5 M) was added 2 h before the addition of progesterone. Control wells contained the vehicle diluent.

Statistical analysis

The WBA data comparing proliferative and secretory GSK-3β levels were analysed by Student’s t-test. The culture data for either mRNA or protein were analysed by one-factor ANOVA followed by the Student–Newman–Keuls test for in between group comparisons. P values <0.05 were considered significant.

Results

Expression and distribution of total GSK-3β in human female reproductive tissues

WBA (Figure 2A) showed the highest total GSK-3β expression in the endometrium, followed by myometrium, ovary and Fallopian tubes. IHC analysis of the reproductive tract for GSK-3β demonstrated expression in the endometrium (Figure 2D), ovary (Figure 2E) and Fallopian tubes (Figure 2F). Endometrial tissue showed staining predominantly in the glandular epithelial cells of the endometrium (Figure 2D). Staining was also observed in luteinized stromal cells of the ovary (Figure 2E ), epithelial cells and smooth muscle cells of the Fallopian tubes (Figure 2F). In the myometrium (data not shown), diffuse staining was noted in smooth muscle cells. The negative control showed no staining (Figure 2B), whereas the positive control (Figure 2C) showed the expected previously described GSK-3β staining in seminiferous tubules and Sertoli cells (Guo et al., 2003).

Comparative quantitative total and Ser9 GSK-3β expression in human endometrium in various stages of the menstrual cycle

WBA for total and pGSK-3β in endometria obtained from proliferative phase (n = 4) and secretory phase (n = 5) is shown in Figure 3A and B, respectively. The expression of total GSK-3β protein did not vary throughout the menstrual cycle. However, expression of pGSK-3β varied with more than a 5-fold increase in its expression in secretory phase samples as compared with the proliferative phase (Figure 3B) (P < 0.001). IHC analysis for GSK-3β and pGSK-3β was performed in endometria from various stages of the menstrual cycle. IHC analysis for total GSK-3β demonstrated uniform staining of epithelial cells throughout the menstrual cycle (data not shown). Expression of pGSK-3β was not uniformly detectable in controls irrespective of stage of the menstrual cycle (Figure 4 , panels B, C, F and J). Consistent with total GSK-3β results, pGSK-3β is restricted in its expression by IHC analysis to the epithelial compartment (Figure 4, panels G–L), and consistent with western blotting, its expression in the proliferative phase epithelial cells is marginal (panels B and C) but peaks shortly after ovulation on day 17 secretory endometrium (panels H and M). Elevation is sustained but less pronounced in the late secretory phase (Figure 4, panels K, L and N).

In vitro studies: progesterone regulates GSK-3β phosphorylation in HES cells

In cell culture experiments using HES cells, progesterone at doses of 10^–6 and 10^–8 M did not alter total GSK-3β mRNA as determined by real-time RT-PCR (Figure 5A). Progesterone also did not alter total GSK-3α protein levels (data not shown) but stimulated pGSK-3β protein expression (Figure 5B). This effect of progesterone could be blocked by the anti-progestin RU486 suggesting that inactivation of GSK-3β through phosphorylation is mediated through progesterone and its receptor (Figure 5B). This experiment also indicates under basal conditions phosphorylation of a portion of GSK-3β occurs and this can be further stimulated by progesterone.

Discussion

The hormonal control of endometrial receptivity is a complex process leading to changes in both the epithelial and the stromal compartments of the post-ovulatory endometrium. Some of these changes include the development of the post-ovulatory triad which includes the presence of giant mitochondria, subnuclear glycogen and a nuclear channel system (Noyes et al., 1950; Dockery et al., 1988; Demir et al., 2002). Although it is generally accepted that the glycogen secretion in the secretory endometrium is under hormonal control, little is known about the mechanism of this control and the implantation process. More is known about the regulation of glycogen synthesis post-implantation and its critical role in rodent early pregnancy. Recently, AKT2 (PKB), one of the most prominent regulators of inactivation of GSK-3β through serine phosphorylation, was implicated in the regulation of glycogen synthesis. In AKT2-deficient mice, there is loss of ‘glycogen’ cells in the syncytiotrophoblast, leading to placental malfunction, fetal growth impairment and premature pregnancy loss (Yang et al., 2003). This study demonstrates that GSK-3β protein is expressed throughout the reproductive tract with the greatest expression in the endometrium. Furthermore, both IHC analysis and WBA showed that although total GSK-3β protein and mRNA expression does not show cyclic variation, the phosphorylated form of the enzyme varies during the cycle, with significantly higher expression during the secretory phase when progesterone levels are known to be elevated. The lack of menstrual cycle variance in total GSK-3β mRNA/protein expression is in agreement with the findings of Tulac et al. (2003). Our in vitro studies indicate that progesterone although unable to activate transcriptional events leading to the increase in GSK-3β mRNA steady state or the translated total protein levels can directly act on the endometrial cells to up-regulate through post-translational modification the expression of phosphorylated GSK-3β. These data support the hypothesis that progesterone induces the phosphorylation of GSK-3β thereby rendering it inactive. Active GSK-3β inactivates through phosphorylation the rate-limiting enzyme in glycogen synthesis, GS at three distinct residues (Ferrer et al., 2003). Inactivation of GSK-3β would relieve inhibition of GS leading to greater glycogen synthesis during the secretory phase. On the basis of our proposed model (Figure 1), aberrations in the inactivation of GSK-3β could result in inadequate glycogen production and potentially failure of implantation. The endometrial glycogen content is an important determinant of fertility, and earlier studies showed decreased endometrial glycogen stores in a subset of infertility patients (Maeyama et al., 1977). It may be speculated that the lower endometrial glycogen levels in these patients could be secondary to defects in GSK-3β phosphorylation. Another condition in which altered GSK-3β phosphorylation could be causing infertility is in patients with polycystic ovary syndrome (PCOS) with hyperinsulinaemia. Insulin is known to down-regulate GSK-3β activity through Ser-9 phosphorylation in fat, liver and muscle tissues in diabetes- and obesity-prone C57BL/6J mice (Eldar-Finkelman et al., 1999). We speculate that insulin resistance along with anovulation and absence of luteal phase progesterone in patients with PCOS could blunt the endometrial expression of pGSK-3β in luteal phase along with glycogen synthesis and contribute to the higher incidence of implantation failures previously reported in these patients (Glueck et al., 2002). Modulators of GSK-3β, which have recently become available (Cohen and Goedert, 2004) if made clinically available, could provide a means for manipulating endometrial GSK-3β and correcting defects in its expression. Although our present work has focused on the role of GSK-3β in glycogen metabolism in the secretory endometrium, this kinase has many other downstream substrates (Cohen and Frame, 2001) including transcription factors and proteins involved in both cell cycle regulation and apoptosis. We cannot exclude other consequences of progesterone-mediated inactivation of GSK-3β in the secretory human endometrium. It has been previously shown that the activity of the GS is 3-fold higher in the glandular versus the stromal cells of the human luteal phase endometrium and a strong expression of glycogen phosphorylase is present in the stroma (Souda et al., 1985). Although GSK-3β is the dominant regulator of GS, other factors contribute to its regulation. In mice lacking Glut4 glucose transporter, GS is regulated positively through a variety of factors such as hexokinase II, glucose-6-phosphate and a simultaneous decrease in the activity of glycogen phosphorylase (Kim et al., 2005). These data together with the apparent absence of pGSK-3β in the stromal cells of the endometrium suggest differential mechanisms of regulation of glycogen synthesis in each of these cell types.

GSK-3β is expressed in reproductive tissues such as the Fallopian tube and ovary raising questions regarding its physiological role in these sites. Its homology to yeast meiotic regulators, its expression profile in the male testes at initiation of meiosis and its inhibition in vitro of S phase meiosis I implicate it in the initiation of meiosis in rodents (Guo et al., 2003). Its expression profile in the female genital ridge (unpublished observation) could potentially implicate it in a similar role in the initiation of meiosis in female oocytes in utero. Others have reported that GSK-3β inhibits progesterone-mediated Xenopus oocyte maturation (Fisher et al., 1999; Nebreda and Ferby, 2000) and that progesterone inactivates GSK-3β in immature, G2-arrested Xenopus oocytes leading to their maturation and resumption of G2 to M transition of meiosis I. In this model, GSK-3β is critical for the progesterone-mediated resumption of meiosis in arrested oocytes. These studies, along with our own, demonstrate the multifunctional role of GSK-3β in a variety of reproductive tissues.

In summary, Wnt signalling and GSK-3β are widely expressed in the female reproductive tract. Both Wnt and PI3-kinase/AKT signalling are regulated by E₂ and progesterone and play important roles in the cyclical changes in the human endometrial epithelium. Here, we add to this body of knowledge showing that the GSK-3β phosphorylated form shows cyclic variation in the endometrium and phosphorylation of GSK-3β in the endometrium is regulated by progesterone and that this inhibtion of GSK-3β is temporaly and cellularly coinicident with the increased glycogen synthesis in the luteal phase epithelium. Additional studies in patients with various causes of implantation failure should help define the physiological role of this enzyme in endometrial physiology and pathology.

Acknowledgements

This work was supported in part by NIH RO3HD41409-01 to O.K.

References

Author notes

1Department of Internal Medicine, 2Department of Obstetrics and Gynecology and 3Department of Pathology, Harbor-UCLA Medical Center, Torrance, CA, USA