Hypoxia-Inducible Factor-1 Expression in Human Endometrium and Its Regulation by Prostaglandin E-Series Prostanoid Receptor 2 (EP2
http://www.100md.com
《内分泌学杂志》
University of Edinburgh (H.O.D.C., J.O., T.A.H., L.B.), Centre for Reproductive Biology, The Queen’s Medical Research Institute
Medical Research Council Human Reproductive Sciences Unit (K.J.S., H.N.J.), The Queen’s Medical Research Institute, Edinburgh EH16 4TJ, Scotland, United Kingdom
Medical Research Council Centre for Inflammation Research (N.H.), University of Edinburgh, Edinburgh EH8 9AG, Scotland, United Kingdom
Abstract
The menstrual cycle is a complex interaction of sex steroids, prostanoids, and cytokines that lead to coordinated tissue degradation, regeneration and repair. The transcription factor hypoxia-inducible factor (HIF-1) plays critical roles in cellular responses to hypoxia, the generation of an inflammatory response and vasculogenesis through transcriptional activation of angiogenic genes. We hypothesize that HIF-1 is expressed in human endometrium and that locally synthesized prostaglandins (PGE2 and PGF2) regulate HIF-1 activity. Here we demonstrate that PGE2 up-regulates HIF-1 mRNA and protein via the E-series prostanoid receptor 2 (EP2), and this up-regulation is dependent on epidermal growth factor receptor kinase activity. We show the tight temporal-spatial confinement of HIF-1 protein expression in endometrium across the cycle. HIF-1 is expressed exclusively during the secretory and menstrual phases. Protein expression is maximal at progesterone withdrawal during the late secretory and menstrual phase. HIF-1 protein colocalizes with prostaglandin EP2 receptor in glandular cells. In contrast, HIF-1/aryl receptor nuclear translocator 1 expression occurs throughout the cycle but is maximal in glandular cells during the proliferative phase. This provides evidence for a role for HIF-1 in the menstrual cycle and demonstrates that HIF-1 activation in human endometrium may occur via a PGE2-regulated pathway and provides a coordinated pathway from progesterone withdrawal through to angiogenic gene expression via HIF-1.
Introduction
A UNIQUE FEATURE of the primate endometrium is the cycle-specific change in vascularization. Its response to progesterone withdrawal, which occurs at the end of a cycle in the absence of a pregnancy, results in menstruation in humans and a few Old World primates. The sequential effect of endogenous steroids, estrogen and progesterone, and withdrawal of progesterone on the human endometrium, have been extensively studied and reviewed (1, 2). The precise local mechanisms involved in the induction of menstruation have yet to be fully elucidated.
The classical experiments that resulted in the initial hypothesis regarding the mechanism of menstruation were conducted in rhesus monkeys. The direct observation of changes in intraocular implants of endometrium lead to the suggestion that the initiating events in menstruation were vasoconstriction of the spiral arterioles and a decrease in blood flow, resulting in tissue hypoxia followed by necrosis (3). However, there is evidence that menstrual bleeding in women is not merely due to a passive response to an ischemic injury but rather an active process of tissue fragmentation.
Hypoxia in endometrial tissues has since been considered a physiologic event in the premenstrual period and is a likely stimulus for angiogenesis. It has been demonstrated in vitro, in human endometrial stromal cells, that hypoxia and cAMP can induce production of the angiogenic factor, vascular endothelial growth factor (VEGF) (4).
Menstruation has been described as an inflammatory event with a complex interaction of sex steroids, prostanoids and cytokines leading to tissue degradation followed by a coordinated process of tissue regeneration and repair. Our studies, and those of others, of human endometrial tissue have shown that progesterone withdrawal leads to up-regulation of cyclooxygenase-2 (COX-2) and subsequent generation of the potent vasoactive prostaglandins PGE2 (a vasodilator) and PGF2 (a vasoconstrictor) (5, 6). Hypoxia-inducible factor (HIF)-1 is a heterodimeric nuclear transcription factor that mediates the effects of hypoxia (7). Under normoxic conditions the HIF-1 -subunit is hydroxylated at Pro402 or Pro564 by prolyl hydroxylases, polyubiquinated, and rapidly degraded in proteosomes. Under hypoxic conditions, the lack of hydroxylation prevents HIF-1 degradation allowing dimerization with HIF-1/ARNT1 (aryl receptor nuclear translocator 1) (8, 9) and nuclear translocation. In contrast to HIF-1, cytoplasmic HIF-1/ARNT1 is stable in normoxia and hypoxia. Downstream HIF-1 target genes include those encoding extracellular matrix remodeling/digesting proteinases (10) and angiogenic/tissue repair genes such as VEGF, connective tissue growth factor, endothelin and angiopoietin 2 (7, 11, 12). The role of angiogenesis, in particular, is of major significance in the cyclical repair and regeneration of the endometrium (11). VEGF expression in endometrial epithelium and stroma varies according to the stage of the menstrual cycle and may be regulated, at least in vitro, by steroids and hypoxia (4, 13).
Although hypoxia is the archetypal stimulus, HIF-1 gene and protein expression can also be induced through a number of inflammatory stimuli including TNF-, prostaglandins, and endotoxin (14, 15, 16). Recently, a series of studies performed in cancer cell lines indicated that the vasodilator PGE2, acting through specific prostanoid G protein-coupled receptors such as the EP2 receptor, generates VEGF expression via HIF-1 activation (15, 17, 18, 19). Hence, we hypothesize that there are potentially two pathways that link COX-2 up-regulation with downstream angiogenic gene expression; increased PGF2 expression leading to local ischemia/hypoxia and thus HIF-1 expression, and increased PGE2 expression directly inducing HIF-1 expression (Fig. 1). Because little is known with regard to its expression and regulation in healthy endometrium, the aims of this study were 1) to observe the expression of HIF-1 and HIF-1/ARNT1 in human endometrium across the endometrial cycle and 2) to define the regulation of HIF-1 expression in endometrial cells by PGE2. We have herein shown, for the first time, tight temporal-spatial confinement of HIF-1 protein expression across the cycle. HIF-1/ARNT1 protein expression was maximal in the glandular component during the proliferative phase at a time of postmenstrual repair. HIF-1 was expressed exclusively in the secretory and menstrual phases, with increasing intensity during progesterone withdrawal from the mid to late secretory phase with maximal expression at menstruation. HIF-1 expression was largely localized to the glandular epithelium, particularly in the uppermost subepithelial zones. We demonstrate colocalization of HIF-1 protein with the EP2 receptor in endometrial tissue. Finally, we show that PGE2 up-regulated HIF-1 mRNA and protein via the EP2 receptor and that this up-regulation was dependent upon epidermal growth factor receptor (EGFR) kinase activity. These observations suggest a role for HIF-1 in the menstrual cycle and demonstrate that endometrial HIF-1 activation may occur via a PGE2-regulated pathway.
Materials and Methods
Materials
All cell culture medium was purchased from Life Technologies (Paisley, Scotland, UK). Penicillin-streptomycin and fetal calf serum were purchased from PAA Laboratories Ltd. (Middlesex, UK). B-Actin mouse monoclonal antibody was purchased from Santa Cruz Biotechnology (Autogen-Bioclear, Wiltshire, UK). The polyclonal rabbit anti-EP2 receptor antibody was purchased from Cayman Chemical Co. (Alexis Corp., Nottingham, UK). The monoclonal mouse anti-HIF-1 and HIF-1 antibodies were purchased from Abcam Ltd. (Cambridge, UK). The antimouse alkaline phosphatase secondary antibodies, indomethacin, PBS, BSA, and PGE2 were purchased from Sigma Chemical Co. (Dorset, UK). The goat antimouse secondary antibody was purchased from Dako Corp. (Carpinteria, CA). The biotinylated horse antimouse IgG, biotinylated goat-antirabbit IgG, avidin and biotin were purchased from Vector Laboratories, Inc., (Peterborough, UK). ECF chemiluminescent system was purchased from Amersham Biosciences (Little Chalfont, Buckinghamshire, UK). AH 6809 (30 mM Stock in dimethylsulfoxide) and AG1478 (10 mM stock in dimethylsulfoxide) were purchased from Calbiochem (Nottingham, UK) and stored at –20 C.
Human endometrial tissue collection
Human endometrial biopsies were collected from patients undergoing hysterectomy or endometrial investigation for benign gynecological conditions (n = 54) that included pelvic pain, heavy bleeding, prolapse, and sterilization. Patients with endometriosis were excluded. Ethical approval was granted by the Lothian Research Ethics Committee, and written informed consent was obtained from each patient. All women reported regular menstrual cycles (25–35 d length) and had not taken any exogenous hormones or used an intrauterine device during the 3 months before obtaining the biopsy. At hysterectomy, a full thickness endometrial biopsy was taken from the uterine cavity of 33 of the above patients for immunohistochemical studies, including the functional and basal layer of the endometrium and adjacent myometrium. Endometrial biopsies were fixed in 4% neutral buffered formalin overnight at 4 C before routinely wax embedding using an 18-h cycle on a TP1050 processing machine (Leica Corp., Knowlhill, Milton Keynes, UK). In addition, endometrial tissue was collected from 23 of the above patients (with an endometrial tissue sampler (Pipelle, Laboratoire CCD, Paris, France); and either snap frozen in liquid nitrogen or placed in RNA Later, RNA stabilization solution [Ambion (Europe) Ltd., Cambridgeshire, UK] overnight at 4 C for subsequent RNA extraction. The biopsies were dated according to the criteria of Noyes et al. (20) based on the histological appearance, which was found to be consistent with the patient’s reported last menstrual period. Serum samples were collected from each patient at the time of endometrial biopsy collection for the determination of circulating serum progesterone and estradiol levels by RIA. A significant reduction in serum progesterone levels was detected in the late vs. the mid secretory phase (P < 0.01; see Table 1). Endometrial biopsies were classified as early proliferative (EP) (n = 5), mid proliferative (MP) (n = 5), late proliferative (LP) (n = 5), early secretory (ES) (n = 6), mid secretory (MS) (n = 5), late secretory (LS) (n = 5) or menstrual (M) (n = 2) for the immunohistochemical studies and proliferative (P) (n = 5), ES (n = 5), MS (n = 5), LS (n = 4), and M (n = 4) for the analysis of HIF-1 mRNA across the menstrual cycle.
HIF-1 and HIF-1/ARNT1 immunohistochemistry
Five-micrometer paraffin sections were dewaxed in Histoclear (National Diagnostics, Atlanta, GA) for 10 min before rehydration in descending grades of alcohol. Sections were washed with 0.01 M PBS (pH 7.4; Sigma) before pressure cooking in 0.01 M sodium citrate (pH 6) for 5 min at setting 2/high (Tefal, Clipso, Nottingham, UK) for antigen retrieval. After cooling for 20 min, the sections were again washed in PBS before blocking endogenous peroxidase activity by immersion in 3% hydrogen peroxide for 10 min at room temperature. After washing in PBS, the sections were incubated in avidin (Vector Laboratories) for 15 min, rinsed in PBS and then incubated in biotin (Vector Laboratories) for a further 15 min, all at room temperature. Slides were incubated in nonimmune rabbit serum (NRS) (where subsequent incubation was with the anti-HIF-1 antibody) or in nonimmune horse serum (NHS) (where subsequent incubation was with the anti-HIF-1 antibody) (NRS; Diagnostics Scotland, Carluke, Lanark, UK; NHS, Vector Laboratories) in PBS for 20 min at room temperature to block nonspecific binding of the primary antibody. The primary antibody, either monoclonal mouse anti-HIF-1 antibody (1:1000 dilution in NRS/PBS; IgG2b ab1) or monoclonal mouse anti-HIF-1 antibody (1:1000 dilution in NHS/PBS; IgG1 ab2771) was added to the slides and incubated overnight at 4 C. For the negative controls, the primary antibody was replaced with either nonimmune mouse IgG2b antibody (HIF-1) or IgG1 antibody (HIF-1) at a matched antibody concentration to the HIF-1 (1:40) or HIF-1 (1:1000) antibody. Subsequently, the sections were washed in PBS with added Tween 20 (PBST) before incubating in biotinylated horse antimouse antibody (Vector Laboratories) for 60 min at room temperature. After washing in PBST, an avidin-biotin-peroxidase complex (ABC-Elite; Vector Laboratories) was then applied for 60 min at room temperature. After a final wash with PBST, the chromagen 3, 3'-diaminobenzidine (Dako) was added and the reaction stopped with distilled H2O when nuclear staining was detected by inspection under the microscope. The sections were then counterstained with Harris’s hematoxylin, dehydrated, and finally mounted with Pertex (Cellpath plc, Hemel Hempstead, UK).
Scoring of immunoreactivity
Localization and intensity of immunostaining was evaluated blind by two independent observers using a semiquantitative scoring system with a three-point scale (0 = no staining, 1 = mild staining, 2 = strong staining for HIF-1) and a four-point scale (0 = no staining, 1 = mild staining, 2 = moderate staining and 3 =strong staining for HIF-1). This was applied to the glands and stromal cells in the functional and basal endometrial layers, respectively, as well as the surface epithelium. This scoring system has previously been validated (21) in a subset of tissue sections in which immunoreactivity was measured with a computerized image analysis system. A strong correlation between quantitative data derived from image analysis and subjective scores by trained observers was obtained. Statistical significance was determined using the Kruskal-Wallis nonparametric test (Instat, GraphPad Software, Inc., San Diego, CA). Dunn’s multiple comparison test was used for post hoc comparisons.
HIF-1-EP2 dual immunofluorescence
Paraffin tissue sections were dewaxed, pressure cooked, and endogenous peroxidase and biotin blocked as described above. Nonspecific binding of the primary antibody was blocked by incubating the sections for 20 min at room temperature in a 1:5 dilution of nonimmune goat serum (Diagnostics Scotland) with 5% BSA. The sections were then incubated with mouse monoclonal anti HIF-1 antibody (1:4000 dilution, IgG2b, ab1) and polyclonal rabbit anti-EP2 receptor antibody (1:50 dilution) at 4 C overnight. Negative controls were performed by incubating sections with a control mouse IgG2b antibody at a matched IgG concentration to the HIF-1 antibody (1:1300 dilution) and a control rabbit IgG at a matched concentration to the EP2 receptor antibody (1:100 dilution), or with the EP2 antibody preabsorbed overnight at 4 C with the peptide against which it had been raised (1:50 dilution). The sections were then washed in PBST before the addition of biotinylated goat-antirabbit IgG (1:200 dilution) and goat antimouse antibody (1:200 dilution) conjugated to horseradish peroxidase for 60 min at room temperature. After a further wash in PBST, the sections were incubated with tyramide cyanine 5 fluorescent complex (1:50 dilution, to detect HIF-1) for 10 min at room temperature (PerkinElmer Life Sciences, Boston, MA), washed in PBST, then incubated in a solution of avidin alexafluor 488 (1:200 dilution, Molecular Probes, Inc., Leiden, The Netherlands) to detect EP2 for 60 min at room temperature. The sections were then washed in PBST and incubated in a solution of biotin/ PBS for 20 min at room temperature (Vector Laboratories). Sections were then washed in PBST, and mounted in permafluor (Immunotech, Marseilles, France) and left to dry in the dark. The fluorescent images were collected on a Zeiss LSM 510 confocal laser scanning microscope (Carl Zeiss, Thornwood, NY). The avidin alexafluor 488 (EP2) was visualized using an argon laser with an excitation beam of 496nm, emitting at 519 nm and detected using a band pass filter from 505–550 nm. The cyanine 5 (HIF-1) was visualized using a Helium/neon 2 laser with an excitation beam of 649 nm, emitting at 670 nm and detected using a long pass filter of 650 nm.
Cell culture
Human Ishikawa endometrial adenocarcinoma cells (European Collection of Cell Culture, Centre for Applied Microbiology, Wiltshire, UK) were maintained in DMEM nutrient mixture F-12 with glutamax-1 and pyridoxine, supplemented with 10% fetal calf serum, and 1% antibiotics (stock 500 IU/ml penicillin and 500 μg/ml streptomycin) at 37 C and 5% CO2 (vol/vol). Stable EP2 transfectant cells were maintained under the same conditions with the addition of 200 μg/ml G418.
Transfection
The EP2 receptor sense (EP2S) stable cell line was constructed as described previously (19). Briefly, the EP2 receptor was isolated from proliferative phase human endometrial tissue and cloned into the eukaryotic expression vector pcDNA3.1 (Invitrogen, De Schelp, The Netherlands) containing a neomycin resistance gene expression cassette for G418 selection. The expression vector was transfected into Ishikawa cells and three independent clones, demonstrating elevated expression of EP2 receptor isolated and expanded for further investigation. We characterized all three clones and found that they all exhibited identical phenotypic and biochemical alterations. Therefore, the results of our studies reported here focused on one of these clones (clone S33).
Protein extraction
For HIF-1 studies, 2 x 106 cells were seeded in 10-cm dishes. The following day, cells were washed with PBS and incubated in serum-free culture medium containing penicillin/streptomycin and 8.4 μM indomethacin (a dual COX-enzyme inhibitor used to inhibit endogenous prostanoid biosynthesis) for at least 16 h. The next day, cells were pretreated with vehicle, 100 nM AG1478 (a specific inhibitor of EGFR kinase), 10 μM AH6809 (a specific EP2 receptor antagonist) for 1 h before stimulation with vehicle, 100 nM PGE2 or 200 μM desferrioxamine (DFO) for 48 h. After stimulation, cells were washed with ice-cold PBS. Cytosolic proteins were extracted with a cytosolic protein lysis buffer [10 mM HEPES (pH 7.8), 10 mM KCl, 2 mM MgCl2, 1 mM dithiothreitol, 0.1 mM EDTA, 1% Nonidet P-40] containing protease inhibitors (Complete mini protease inhibitor cocktail; Roche Diagnostics Ltd., Lewes, UK). Thereafter, the membrane fraction was pelleted by centrifugation at 14,000 x g for 2 min at 4 C. The clarified lysate was removed and stored and the nuclear fraction extracted with a nuclear protein lysis buffer [50 mM HEPES (pH 7.8), 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 1 mM dithiothreitol, 10% glycerol] containing protease inhibitors (Roche) followed by brief sonication. The protein content in the nuclear fraction was determined using protein assay kits (Bio-Rad, Hemel Hempstead, UK).
Immunoprecipitation and Western blot analysis
For HIF-1 and B-actin expression, a total of 50 μg of protein was resuspended in 20 μl of Laemmli buffer [125 mM Tris-HCl (pH 6.8), 4% SDS, 5% 2-mercaptoethanol, 20% glycerol and 0.05% bromophenol blue] and boiled for 5 min. Proteins were resolved on 4–12% Tris-glycine gels (NOVEX, Invitrogen), transferred onto polyvinylidene difluoride membrane (Millipore, Watford, UK) and subjected to immunoblot analysis. Membranes were blocked for 1 h at 25 C in 4% BSA diluted in TBST (50 mM Tris-HCl, 150 mM NaCl and 0.05% vol/vol Tween 20) and incubated with specific HIF-1 (1:200) and B-actin (1:500) primary antibody. After washing and incubating with alkaline-phosphatase-conjugated secondary antibodies (Sigma), immunoreactive proteins were visualized by the ECF chemiluminescent system according to the manufacturer’s instructions. HIF-1 proteins were revealed and quantified by, and normalized to, B-actin protein expression using a Typhoon PhosphorImager 9400 (Molecular Dynamics, Amersham Biosciences). Fold increase was calculated by dividing the relative expression of HIF-1 by the relative expression of B-actin. All data are presented as mean ± SEM.
Quantitative RT-PCR (Q-RT-PCR)
HIF-1 expression in Ishikawa cells and in endometrial tissue was measured by Q-RT-PCR (Taqman) analysis. Approximately 5 x 105 cells were seeded in 5-cm dishes and allowed to attach and grow overnight. The following day, cells were synchronized by serum withdrawal for at least 16 h in serum-free medium containing 8.4 μM indomethacin. The next day, cells were pretreated with vehicle, 100 nM AG1478, or 10 μM AH6809 for 1 h before stimulation with vehicle or 100 nM PGE2 for 48 h. After stimulation, cells were washed with ice-cold PBS. RNA was extracted from cells using Tri-reagent (Sigma) and from endometrium using either Tri-reagent or QIAGEN RNeasy columns (QIAGEN Ltd., West Sussex, UK) following the manufacturers’ guidelines. Once extracted and quantified, RNA samples were reverse transcribed using MgCl2 (5.5 mM), deoxy (d) nucleotide triphosphates (0.5 mM each), random hexamers (2.5 μM), ribonuclease inhibitor (0.4 U/μl) and multiscribe reverse transcriptase (1.25 U/μl; all from PE Biosystems, Warrington, UK). The mix was aliquoted into individual tubes, and 200–400 ng of RNA were added. After mixing, samples were incubated for 60–90 min at 25 C, 45 min at 48 C, and 95 C for 5 min. Thereafter, cDNA samples were stored at –20 C. A tube with no reverse transcriptase was included to control for DNA contamination. To measure cDNA expression, a reaction mix was prepared containing Taqman buffer (5.5 mM MgCl2, 200 μM dATP, 200 μM dCTP, 200 μM dGTP, 400 μM deoxyuridine triphosphate), ribosomal 18S forward, and reverse primers and probe (50 nM), forward and reverse primers for HIF-1 (300 nM), HIF-1 probe (100 nM), AmpErase Uracil N-glycosylase (Roche Diagnostics, Lewes, Suffolk, UK) (0.01 U/μl) and AmpliTaq Gold DNA Polymerase (0.025 U/μl; PE Biosystems). After mixing, 48 μl were aliquoted into separate tubes and 2 μl/replicate of cDNA added and mixed before placing duplicate 24-μl samples into a PCR plate. A no template control (containing water) was included in triplicate. PCR was carried out using an ABI Prism 7700 (Applied Biosystems, Warrington, UK). HIF-1 primers and probe for quantitative PCR were designed using the PRIMER express program (PE Biosystems). The sequence of the HIF-1 primers and probe were: forward, 5'-CGCATCTTGATAAGGCCTC-3'; reverse, 5'-AATCACCAGCATCCAGAAG-3'; probe (FAM labeled, 6-carboxyfluorescein) 5'-TCACACGCAAATAGCTGAT-3'. The ribosomal 18S primers and probe sequences were: forward, 5'-CGG CTA CCA CAT CCA AGG AA-3'; reverse, 5'-GCT GGA ATT ACC GCG GCT-3'; probe, (VIC-labeled, PE Biosystems) 5'-TGC TGG CAC CAG ACT TGC CCT C-3'. Data were analyzed and processed using Sequence Detector version 1.6.3 (PE Biosystems). Expression of HIF-1 was normalized to RNA loading for each sample using the 18S ribosomal RNA as an internal standard.
For HIF-1 expression in Ishikawa cells results are expressed as fold increase where relative expression of HIF-1 in cells treated with PGE2 was divided by the relative expression in vehicle-treated cells. Data are presented as mean ± SEM. For endometrial tissue, results were expressed as quantity relative to a comparator, which was a sample of RNA taken from the proliferative stage of the menstrual cycle. Significant difference was determined using one-way ANOVA, and individual differences were described using the least significant difference post hoc multiple comparison (SPSS, Inc., Chicago, IL).
Results
Spatial and temporal distribution of HIF-1 protein expression (Figs. 2, a–c, and 3)
Immunostaining for HIF-1 was exclusively detected in cell nuclei consistent with the fact that stable HIF-1 is rapidly translocated to the nucleus (Fig. 2, b and c). There was no immunostaining in the negative controls (Fig. 2c, inset). During the proliferative phase, immunostaining was absent in all glands and stromal cells and negligible immunoreactivity was observed in the surface epithelium (Fig. 2a).Glandular cells and stromal cells in the functional layer (F) stained most intensely in samples collected from the menstrual stage of the cycle (Fig. 2c). HIF-1 was detected in the glands of the basal layer (B) only in the late secretory and menstrual phase (Fig. 2c). In none of the samples was there any immunostaining of the stromal cells in the basal layer (Fig. 2, a–c).
After semiquantitative evaluation and statistical analysis, the Kruskal-Wallis test, showed a significant difference (P < 0.01) in the intensity of nuclear staining in samples from different stages of the menstrual cycle (Fig. 3, a and b). The difference in intensity across the cycle stages was significant for both the glands (P < 0.01; Fig. 3a) and the stromal cells (P < 0.01; Fig. 3b). Using Dunn’s multiple comparison test, there was a significant difference in staining of the stromal cells in the functional layer in the menstrual phase when compared with the proliferative, early and mid secretory phases (Fig. 3b; P < 0.05). Glands in the basal layer in the menstrual phase also exhibited positive immunoreactivity (Fig. 3d; P < 0.05).
Spatial and temporal distribution of HIF-1/ARNT1 protein expression (Figs. 2, d–f, and 4)
In the functional layer (F) of endometrium, HIF-1/ARNT1 protein was observed in the nuclei of surface and glandular epithelial cells and stroma. Protein expression in the functional layer was greater in the proliferative compared with mid and late secretory phases of the cycle (Fig. 4, a–c; P < 0.05). In the basal region (B) of endometrium HIF-1/ARNT1 protein was greater in glandular cells in the proliferative phase (Fig. 4d; P < 0.05). Stromal cell expression of HIF-1 protein in the functional layer was also greater in the proliferative phase (Fig. 4c). Thus, overall HIF-1/ARNT1 protein immunoreactivity was maximally intense in the proliferative phase. Furthermore, HIF-1/ARNT1 protein was expressed in stromal cells at this cycle stage when HIF-1 immunoreactivity was negligible.
Expression of HIF-1 mRNA in human endometrium by Q-RT-PCR (Fig. 5)
HIF-1 mRNA was present at low levels in the proliferative and early secretory stages of the menstrual cycle. By the mid secretory phase, levels of HIF-1 had risen significantly when compared with both the proliferative (P < 0.05) and early secretory phase samples (P < 0.01). Moderate levels of HIF-1 were also present in the late secretory and menstrual phases of the cycle.
HIF-1 expression is up-regulated by PGE2-EP2 receptor interaction in endometrial epithelial cells
The role of PGE2-EP2 receptor signaling on the expression of HIF-1 mRNA was investigated by Q-RT-PCR analysis after stimulation of wild-type and EP2S Ishikawa cells with 100 nM PGE2 or vehicle (48 h). As shown in Fig. 6, PGE2 stimulation resulted in a 1.6 ± 0.03-fold increase in the expression of HIF-1 mRNA in Ishikawa EP2S cells (P < 0.05). No increase in HIF-1 expression was observed in wild-type cells treated with PGE2. Cotreatment of cells with the selective EP2 receptor antagonist (AH6809) and the EGFR kinase inhibitor AG1478 abolished the EP2 receptor-mediated elevation in HIF-1 mRNA expression (P < 0.05).
To determine whether PGE2 activation of the EP2 receptor promoted the expression of HIF-1 protein, we treated Ishikawa wild-type and EP2S cells with 100 nM PGE2 or vehicle in the presence or absence of the selective EP2 receptor antagonist (AH6809) and EGFR kinase inhibitor (AG1478) for 48 h and measured HIF-1 expression by Western blot analysis. In parallel, cells were treated with DFO, an iron chelator that inactivates prolyl hydroxylases, hence conferring HIF-1 stabilization. An increase in content of HIF-1 was observed in Ishikawa EP2S cells (Fig. 7, lane 7, P < 0.001) after treatment with PGE2 compared with cells stimulated with vehicle alone (Fig. 7, lane 6). No such increase in the expression of HIF-1 protein was observed in wild-type cells in response to PGE2 (Fig. 7, lane 2). Preincubation of cells with AH6809 (Fig. 7, lane 8) and AG1478 (Fig. 7, lane 9) abolished the PGE2-mediated elevation in HIF-1 protein expression in EP2S cells. Stimulation of Ishikawa cells with DFO resulted in a 4.8 ± 1.1 (Fig. 7, lane 5, P < 0.001) and 3.5 ± 0.2 (Fig. 7, lane 10, P < 0.001)-fold increase in the expression of HIF-1 in wild-type and EP2S cells, respectively.
Colocalization of HIF-1 protein and EP2 receptor protein in human endometrium
Dual immunofluorescence has demonstrated colocalization of HIF-1 and EP2 receptor proteins to the glandular epithelial cells (Fig. 8). The EP2 receptor protein is expressed at all stages of the menstrual cycle in the cytoplasm of the glandular epithelium and vascular endothelium (22). As described earlier HIF-1 protein is expressed particularly in the late secretory and menstrual phases in the nuclei of glandular cells.
Discussion
In this study, we have demonstrated tight temporal and spatial confinement of HIF-1 expression across the menstrual cycle. HIF-1 protein was expressed with increasing intensity from premenstrual (secretory) through to menstrual phase human endometrium. No HIF-1 immunoreactivity was detected in endometrial biopsies from the proliferative phase. Within the endometrial layers, the strongest immunoreactivity observed was in the uppermost endometrial zones (functional layer), with staining localized to glandular epithelium, surface epithelium and some stromal cells. HIF-1 mRNA expression peaked in the mid secretory phase, before maximal HIF-1 protein expression. The temporal and spatial distribution of HIF-1 protein and mRNA detected in this study would thus seem to be related to the process of menstruation. HIF-1 appears to be most strongly expressed during the perimenstrual phase in those endometrial layers that will be sloughed off at menstruation. In contrast, HIF-1/ARNT1 protein expression was maximal in the glandular component during the proliferative phase, that is, at a time of postmenstrual repair. Endometrial tissue collection at the time of hysterectomy permits for maintenance of the spatial orientation of the tissue. This is not possible with endometrial curettage sampling.
One previous study has also reported on the expression of HIF-1 proteins in human endometrium (23). The authors failed to detect HIF-1 nuclear staining in any endometrial samples, finding only fairly minor cytoplasmic staining without a specific temporal or spatial pattern. HIF-1 expression was also reported and found to be predominantly epithelial and strongest after ovulation, similar to our observations herein. The discrepancies in findings of HIF-1 expression may reflect the use of different primary antibodies and different methods of tissue collection.
The presence of HIF-1 protein in the glandular epithelium may be responsible for the up-regulation of VEGF that has been reported at the time of progesterone withdrawal in the uppermost zones of the endometrium (13, 24).
HIF-1 is recognized as a critical mediator of hypoxic responses, and hypoxia is certainly the best studied and probably the most potent stimulus for HIF-1 induction. Virtually all cell types appear to express stable HIF-1 protein under conditions of reduced oxygen tension, implying that the oxygen-sensing apparatus is ubiquitous. For this reason, the detection of HIF-1 in vitro or in animal models is often considered to be an indicator of cell or tissue hypoxia. However, the observations in the present study suggest that within the endometrium, hypoxia may not be the only or indeed the most important stimulus for HIF-1 expression. During the premenstrual and menstrual phase, HIF-1 was found to be strongly expressed in glandular epithelium, whereas adjacent stromal cells stained weakly or not at all. Because adjacent cells are unlikely to be exposed to significantly different oxygen tensions, this observation implies an alternative receptor-specific pathway to HIF-1 induction. We have shown that the temporal expression of HIF-1 in the endometrium coincides with progesterone withdrawal (see Table 1). We have previously shown that this latter event results in expression of COX-2 and subsequent generation of the potent vasoactive prostaglandins PGE2 and PGF2 (1, 25). PGE2 is a vasodilator and inducer of vascular leak, whereas PGF2 is a potent vasoconstrictor. Although the latter may induce HIF-1 expression through local tissue ischemia, we here have shown that endometrial HIF-1 expression may be regulated directly by PGE2 as HIF-1 protein colocalized with the EP2 receptor and PGE2 promoted HIF-1 mRNA and protein expression via the EP2 receptor in normoxic conditions.
PGE2 couples to four pharmacologically classified subtypes of G protein-coupled receptor (EP1-EP4) (26). These subtypes may be coexpressed on the same cell or on adjacent cells, indicating an autocrine/paracrine control of an autocoid biosynthesized and released in close proximity to the site of its action (27). Thus, the signal transduction from PGE2-EP ligand-receptor binding to activation of HIF-1 and VEGF expression is likely to be complex. In the present study, we have demonstrated that HIF-1 mRNA and protein expression are induced in endometrial epithelial cells by PGE2 via the EP2 receptor. These findings are in agreement with a similar study by Liu et al. (28), where HIF-1 expression in human prostate cancer cells was induced, with EP2 and EP4 (not EP3) receptor-selective agonists, demonstrating a role for specific EP receptors in the PGE2-regulation of HIF-1 expression. Inhibition of EGFR kinase activity abolished the PGE2-induced expression of HIF, indicating that trans-phosphorylation of the EGFR is also a requirement for the PGE2-induced expression of HIF via the EP2 receptor in endometrial epithelial cells. In a recent study, we have shown that EP2 receptor localizes to glandular epithelial cells and no variation in expression is detected in endometrial biopsies collected across the menstrual cycle (22). Moreover, PGE2-EP2 receptor signaling promotes the expression and release of VEGF from endometrial epithelial cells via the intracellular trans-phosphorylation of the EGFR and activation of the ERK1/2 signaling pathways (19). Our data herein, by inference from the above, strongly suggests this pathway is dependent on HIF-1. PGE2 is also synthesized across the menstrual cycle, albeit its pattern of synthesis may vary depending on the stage of the menstrual cycle. Hence, it is our belief that the EP receptors can be activated during the late secretory phase to affect HIF expression. Recent studies of human colonic cancers have demonstrated a correlation between the expression of the COX-2 enzyme, and its biosynthesized product prostaglandin E2 with VEGF expression and angiogenesis (29, 30) and inhibition of HIF-1 expression by RNA interference blocked the induction of VEGF mRNA (15). Further studies elucidating the relationship between EP2 receptor signaling, HIF and VEGF in endometrial epithelial cells may provide insight into some of the complex processes that occurs in the endometrium during menstruation, such as neovascularization and endometrial regeneration.
HIF-1/ARNT1 is considered to be a constitutively expressed cytoplasmic protein and nuclear HIF-1 translocation is dependent upon its dimerization with HIF-1/ARNT1. Therefore, the observation that HIF-1/ARNT expression varied in a specific temporal and spatial distribution during the menstrual cycle, and that the temporal pattern of expression was reciprocal to that seen with HIF-1, was unexpected. For example, HIF-1 expression in glandular epithelium was strongest during the proliferative phase when HIF-1 expression was absent. HIF-1/ARNT1 is a basic helix-loop-helix/PER-ARNT-SIM-harboring protein and is also expressed as two further isoforms, ARNT2 and ARNT3. All three ARNT proteins may dimerize with HIF-1 (7). The monoclonal antibody used in our current studies is specific for HIF-1/ARNT1 and does not detect ARNT2 or ARNT3. Further studies are required to determine the expression and role of ARNT2 and 3 in human endometrium. However, recent data on HIF expression in the preimplantation mouse uterus supports the notion that specific ARNT family members are differentially expressed in uterine tissue and may only dimerize with HIF-1 at specific stages of implantation (31). In addition to a role in HIF-1mediated pathways, the ARNT family is a critical component of the xenobiotic detoxification pathway. The aryl hydrocarbon (Ahr) receptor binds to nuclear ARNTs and induces the transcription of genes encoding xenobiotic-metabolizing enzymes. Little is known about the functional relevance of this pathway in endometrium However, AhR-deficient mice have a complex reproductive deficit and abnormal uterine physiology (32, 33, 34, 35), suggesting that the Ahr pathway may play an integral part in the normal reproductive physiology and not just in xenobiotic biology. Interestingly, the pattern of HIF-1/ARNT1 expression observed in our studies, that is, epithelial zone and predominantly proliferative phase, is broadly similar to the pattern of AhR expression described in human endometrium (36). The implication is that in human endometrium, ARNT1 may play a functional role in the AhR-mediated pathway rather than HIF-1mediated pathways.
Our finding that HIF-1 expression occurred maximally at the menstrual stage is open to interpretation. It is possible that its expression is of particular significance in initiating the regeneration and repair of the endometrium at the beginning of the subsequent cycle. VEGF and a number of other angiogenic genes have been observed in several model systems to be transcriptionally regulated by HIF-1 (7). A positive correlation between HIF-1 expression, VEGF expression and increased tissue microvessel density in human endometrial carcinoma has been established (37). In addition, vasculoprotective genes such as hemoxygenase-1 and inducible nitric oxide synthase are also HIF-1 regulated and may serve to limit tissue injury during ischemia/reperfusion in the endometrium. Although our studies demonstrate that HIF-1 expression may be induced in an endometrial cell line after PGE2 stimulation under normoxic conditions, this does not imply that hypoxia is not relevant to HIF-1 signaling in intact endometrial tissue. We hypothesize that dual regulation of HIF-1 in endometrium may be present through both a hypoxia-independent, PGE2-dependent mechanism and a classical hypoxia-dependent mechanism (see Fig. 1).
In conclusion, this study demonstrates the cell- and stage-specific expression of HIF-1 across the human menstrual cycle and implicates a mechanism for HIF-1 generation through PGE2 stimulation via the EP2 receptor. Further studies are required to determine the additional role of tissue hypoxia in HIF-1 regulation and to clarify the precise functional role of this transcription factor in initiating menstruation or in the subsequent regeneration and repair of the endometrium.
Acknowledgments
We acknowledge the advice provided by Professor R. W. Kelly (Medical Research Council Human Reproductive Sciences Unit, Edinburgh) on the Q-RT-PCR studies. We also acknowledge the assistance of our clinical research nurses Catherine Murray and Sharon McPherson for assistance with patient recruitment and endometrial tissue collection. The secretarial assistance of Kate Williams and Meg Anderson is acknowledged as well as the help provided by Ted Pinner and Corrine McLeod with the illustrations.
Footnotes
This work was supported by UK Medical Research Council Programme Grant No. 0000066.
Disclosure statement: J.O., T.A.H., L.B., K.J.S., H.N.J., and N.H. have nothing to declare. H.O.D.C. has received support, for invited presentations at scientific meetings on management of bleeding problems, from Schering (but not on the topic of the scientific content of this manuscript).
First Published Online November 10, 2005
Abbreviations: Ahr, Aryl hydrocarbon receptor; ARNT1, aryl receptor nuclear translocator 1; COX-2, cyclooxygenase-2; DFO, desferrioxamine; EGFR, epidermal growth factor receptor; EP2, prostaglandin E-series prostanoid receptor 2; HIF-1, hypoxia-inducible factor; NHS, nonimmune horse serum; NRS, nonimmune rabbit serum; PBST, PBS with Tween 20; PGE2 and PGF2, locally synthesized prostaglandins; Q-RT-PCR, quantitative RT-PCR; VEGF, vascular endothelial growth factor.
Accepted for publication October 18, 2005.
References
Critchley HO, Kelly RW, Brenner RM, Baird DT 2001 The endocrinology of menstruation—a role for the immune system. Clin Endocrinol (Oxf) 55:701–710
Critchley HK, Kelly RW, Brenner RM, Baird DT 2003 Antiprogestins as a model for progesterone withdrawal. Steroids 68:1061–1068
Markee JE 1940 Menstruation in intraocular endometrial transplants in the rhesus monkey. Contrib Embryol 177:211–308
Popovici RM, Irwin JC, Giaccia AJ, Guidice LC 1999 Hypoxia and cAMP stimulate vascular endothelial growth factor (VEGF) in human endometrial stromal cells: potential relevance to menstruation and endometrial regeneration. J Clin Endocrinol Metab 84:2245–2248
Critchley HO, Kelly RW, Lea RG, Drudy TA, Jones RL, Baird DT 1996 Sex steroid regulation of leukocyte traffic in human decidua. Hum Reprod 11:2257–2262
Sugino N, Karube-Harada A, Taketine T, Sakata N, Nakamura Y 2004 Withdrawal of ovarian steroids stimulates prostaglandin F2alpha production through nuclear factor-B activation via oxygen radicals in human endometrial stromal cells: potential relevance to menstruation. J Reprod Dev 50:215–225
Semenza GL 2000 HIF-1: mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 88:1474–1480
Pugh CW, Ratcliffe PJ 2003 Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med 9:677–684
Poellinger L, Johnson RS 2004 HIF-1 and hypoxic response: the plot thickens. Curr Opin Genet Dev 14:81–85
Krishnamachary B, Berg-Dixon S, Kelly B, Agani F, Feldser D, Ferreira G, Iyer N, LaRusch J, Pak B, Taghavi P, Semenza GL 2003 Regulation of colon carcinoma cell invasion by hypoxia-inducible factor. Cancer Res 63:1138–1143
Smith SK 2001 Regulation of angiogenesis in the endometrium. Trends Endocrinol Metab 12:147–151
Higgins DF, Biju MP, Akai Y, Wutz A, Johnson RS, Haase VH 2004 Hypoxic induction of CTGF is directly mediated by HIF-1. Am J Physiol Renal Physiol 287:F1223–F1232
Sharkey AM, Day K, McPherson A, Malik S, Licence D, Smith SK, Charnock-Jones DS 2000 Vascular endothelial growth factor expression in human endometrium is regulated by hypoxia. J Clin Endocrinol Metab 85:402–409
Haddad JJ, Land SC 2001 A non-hypoxic, ROS-sensitive pathway mediates TNF--dependent regulation of HIF-1. FEBS Lett 505:269–274
Fukuda R, Kelly B, Semenza GL 2003 Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1. Cancer Res 63:2330–2334
Blouin CC, Page EL, Soucy GM, Richard DE 2004 Hypoxic gene activation by lipopolysaccharide in macrophages: implication of hypoxia-inducible factor 1. Blood 103:1124–1130
Seno H, Oshima M, Ishikawa T, Oshima H, Takaku K, Chiba T, Narumiya S, Taketo MM 2002 Cyclooxygenase-2 and prostaglandin E2 receptor EP2-dependent angiogenesis in APC 716 mouse intestinal polyps. Cancer Res 62:506–511
Eibl G, Bruemmer D, Okada Y, Duffy JP, Law RE, Reber HA, Hines OJ 2003 PGE2 is generated by specific COX-2 activity and increases VEGF production in COX-2-expressing human pancreatic cancer cells. Biochem Biophys Res Commun 306:887–897
Sales KJ, Maudsley S, Jabbour HN 2004 Elevated prostaglandin EP2 receptor in endometrial adenocarcinoma cells promotes VEGF expression via cAMP-mediated transactivation of the EGF Receptor and ERK1/2 signalling pathways. Mol Endocrinol 18:1533–1545
Noyes RW, Hertig AT, Rock J 1950 Dating the endometrial biopsy. Fertil Steril 1:3–25
Wang H, Critchley HOD, Kelly RW, Shen D, Baird DT 1998 Progesterone receptor subtype B is differentially regulated in human endometrial stroma. Mol Hum Reprod 4:407–412
Milne SA, Perchick GB, Boddy SC, Jabbour HN 2001 Expression, localization, and signalling of PGE2 and EP2/EP4 receptors in human nonpregnant endometrium across the menstrual cycle. J Clin Endocrinol Metab 86:4453–4459
Zhang J, Salamonsen LA 2002 Expression of hypoxia-inducible factors in human endometrium and suppression of matrix metalloproteinases under hypoxic conditions do not support a major role for hypoxia in regulating tissue breakdown at menstruation. Hum Reprod 17:265–274
Nayak NR, Brenner RM 2002 Vascular proliferation and vascular endothelial growth factor expression in the rhesus macaque endometrium. J Clin Endocrinol Metab 87:1845–1855
Baird DT, Cameron ST, Critchley HO, Drudy TA, Howe A, Jones RL, Lea RG, Kelly RW 1996 Prostaglandins and menstruation. Eur J Obstet Gynecol Reprod Biol 70:15–17
Coleman RA, Smith WL, Narumiya S 1994 International Union of Pharmacology classification of prostanoid receptors: properties, distribution, and structure of the receptors and their subtypes. Pharmacol Rev 46:205–229
Ashby B 1998 Co-expression of prostaglandin receptors with opposite effects: a model for homeostatic control of autocrine and paracrine signaling. Biochem Pharmacol 55:239–246
Liu XH, Kirschenbaum A, Lu M, Yao S, Dosoretz A, Holland JF, Levine AC 2002 Prostaglandin E2 induces hypoxia-inducible factor-1 stabilization and nuclear localization in a human prostate cancer cell line. J Biol Chem 277:50081–50086
Cianchi F, Cortesini C, Bechi P, Fantappie O, Messerini L, Vannacci A, Sardi I, Baroni G, Boddi V, Mazzanti R, Masini E 2001 Up-regulation of cyclooxygenase 2 gene expression correlates with tumor angiogenesis in human colorectal cancer. Gastroenterology 121:1339–1347
Tsujii M, Kawano S, Tsuji S, Sawaoka H, Hori M, DuBois RN 1998 Cyclooxygenase regulates angiogenesis induced by colon cancer cells. Cell 93:705–716
Daikoku T, Matsumoto H, Gupta RA, Das SK, Gassmann M, DuBois RN, Dey SK 2003 Expression of hypoxia-inducible factors in the peri-implantation mouse uterus is regulated in a cell-specific and ovarian steroid hormone-dependent manner. Evidence for differential function of HIFs during early pregnancy. J Biol Chem 278:7683–7691
Fernandez-Salgureo P, Pineau T, Hilbert DM, McPhail T, Lee SS, Kimura S, Nebert DW 1995 Immune system impairment and hepatic fibrosis in mice lacking the dioxin-binding Ah receptor. Science 268:722–726
Pocar P, Fischer B, Klonisch T, Hombach-Klonisch S 2005 Molecular interaction of the aryl hydrocarbon receptor and its biological and toxicological relevance for reproduction. Reproduction 129:379–389
Mimura J, Yamashita K, Nakamura K, Morita M, Takagi TN, Nakao K, Ema M, Sogawa K, Yasuda M, Katsuki M, Fujii-Kuriyama Y 1997 Loss of teratogenic response to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in mice lacking the Ah (dioxin) receptor. Genes Cells 2:645–654
Fernandez-Salguero PM, Ward JM, Sundberg JP, Gonzalez FJ 1997 Lesions of aryl-hydrocarbon receptor-deficient mice. Vet Pathol 34:605–614
Kuchenhoff A, Seliger G, Klonisch T, Tscheudschilsuren G, Kaltwasser P, Seliger E, Buchmann J, Fischer B 1999 Arylhydrocarbon receptor expression in the human endometrium. Fertil Steril 71:354–360
Sivridis E, Giatromanolaki A, Koukourakis MI, Georgiou L, Anastasiadis P 2002 Patterns of episialin/MUC1 expression in endometrial carcinomas and prognostic relevance. Histopathology 40:92–100(Hilary O. D. Critchley, Julia Osei, Tere)
Medical Research Council Human Reproductive Sciences Unit (K.J.S., H.N.J.), The Queen’s Medical Research Institute, Edinburgh EH16 4TJ, Scotland, United Kingdom
Medical Research Council Centre for Inflammation Research (N.H.), University of Edinburgh, Edinburgh EH8 9AG, Scotland, United Kingdom
Abstract
The menstrual cycle is a complex interaction of sex steroids, prostanoids, and cytokines that lead to coordinated tissue degradation, regeneration and repair. The transcription factor hypoxia-inducible factor (HIF-1) plays critical roles in cellular responses to hypoxia, the generation of an inflammatory response and vasculogenesis through transcriptional activation of angiogenic genes. We hypothesize that HIF-1 is expressed in human endometrium and that locally synthesized prostaglandins (PGE2 and PGF2) regulate HIF-1 activity. Here we demonstrate that PGE2 up-regulates HIF-1 mRNA and protein via the E-series prostanoid receptor 2 (EP2), and this up-regulation is dependent on epidermal growth factor receptor kinase activity. We show the tight temporal-spatial confinement of HIF-1 protein expression in endometrium across the cycle. HIF-1 is expressed exclusively during the secretory and menstrual phases. Protein expression is maximal at progesterone withdrawal during the late secretory and menstrual phase. HIF-1 protein colocalizes with prostaglandin EP2 receptor in glandular cells. In contrast, HIF-1/aryl receptor nuclear translocator 1 expression occurs throughout the cycle but is maximal in glandular cells during the proliferative phase. This provides evidence for a role for HIF-1 in the menstrual cycle and demonstrates that HIF-1 activation in human endometrium may occur via a PGE2-regulated pathway and provides a coordinated pathway from progesterone withdrawal through to angiogenic gene expression via HIF-1.
Introduction
A UNIQUE FEATURE of the primate endometrium is the cycle-specific change in vascularization. Its response to progesterone withdrawal, which occurs at the end of a cycle in the absence of a pregnancy, results in menstruation in humans and a few Old World primates. The sequential effect of endogenous steroids, estrogen and progesterone, and withdrawal of progesterone on the human endometrium, have been extensively studied and reviewed (1, 2). The precise local mechanisms involved in the induction of menstruation have yet to be fully elucidated.
The classical experiments that resulted in the initial hypothesis regarding the mechanism of menstruation were conducted in rhesus monkeys. The direct observation of changes in intraocular implants of endometrium lead to the suggestion that the initiating events in menstruation were vasoconstriction of the spiral arterioles and a decrease in blood flow, resulting in tissue hypoxia followed by necrosis (3). However, there is evidence that menstrual bleeding in women is not merely due to a passive response to an ischemic injury but rather an active process of tissue fragmentation.
Hypoxia in endometrial tissues has since been considered a physiologic event in the premenstrual period and is a likely stimulus for angiogenesis. It has been demonstrated in vitro, in human endometrial stromal cells, that hypoxia and cAMP can induce production of the angiogenic factor, vascular endothelial growth factor (VEGF) (4).
Menstruation has been described as an inflammatory event with a complex interaction of sex steroids, prostanoids and cytokines leading to tissue degradation followed by a coordinated process of tissue regeneration and repair. Our studies, and those of others, of human endometrial tissue have shown that progesterone withdrawal leads to up-regulation of cyclooxygenase-2 (COX-2) and subsequent generation of the potent vasoactive prostaglandins PGE2 (a vasodilator) and PGF2 (a vasoconstrictor) (5, 6). Hypoxia-inducible factor (HIF)-1 is a heterodimeric nuclear transcription factor that mediates the effects of hypoxia (7). Under normoxic conditions the HIF-1 -subunit is hydroxylated at Pro402 or Pro564 by prolyl hydroxylases, polyubiquinated, and rapidly degraded in proteosomes. Under hypoxic conditions, the lack of hydroxylation prevents HIF-1 degradation allowing dimerization with HIF-1/ARNT1 (aryl receptor nuclear translocator 1) (8, 9) and nuclear translocation. In contrast to HIF-1, cytoplasmic HIF-1/ARNT1 is stable in normoxia and hypoxia. Downstream HIF-1 target genes include those encoding extracellular matrix remodeling/digesting proteinases (10) and angiogenic/tissue repair genes such as VEGF, connective tissue growth factor, endothelin and angiopoietin 2 (7, 11, 12). The role of angiogenesis, in particular, is of major significance in the cyclical repair and regeneration of the endometrium (11). VEGF expression in endometrial epithelium and stroma varies according to the stage of the menstrual cycle and may be regulated, at least in vitro, by steroids and hypoxia (4, 13).
Although hypoxia is the archetypal stimulus, HIF-1 gene and protein expression can also be induced through a number of inflammatory stimuli including TNF-, prostaglandins, and endotoxin (14, 15, 16). Recently, a series of studies performed in cancer cell lines indicated that the vasodilator PGE2, acting through specific prostanoid G protein-coupled receptors such as the EP2 receptor, generates VEGF expression via HIF-1 activation (15, 17, 18, 19). Hence, we hypothesize that there are potentially two pathways that link COX-2 up-regulation with downstream angiogenic gene expression; increased PGF2 expression leading to local ischemia/hypoxia and thus HIF-1 expression, and increased PGE2 expression directly inducing HIF-1 expression (Fig. 1). Because little is known with regard to its expression and regulation in healthy endometrium, the aims of this study were 1) to observe the expression of HIF-1 and HIF-1/ARNT1 in human endometrium across the endometrial cycle and 2) to define the regulation of HIF-1 expression in endometrial cells by PGE2. We have herein shown, for the first time, tight temporal-spatial confinement of HIF-1 protein expression across the cycle. HIF-1/ARNT1 protein expression was maximal in the glandular component during the proliferative phase at a time of postmenstrual repair. HIF-1 was expressed exclusively in the secretory and menstrual phases, with increasing intensity during progesterone withdrawal from the mid to late secretory phase with maximal expression at menstruation. HIF-1 expression was largely localized to the glandular epithelium, particularly in the uppermost subepithelial zones. We demonstrate colocalization of HIF-1 protein with the EP2 receptor in endometrial tissue. Finally, we show that PGE2 up-regulated HIF-1 mRNA and protein via the EP2 receptor and that this up-regulation was dependent upon epidermal growth factor receptor (EGFR) kinase activity. These observations suggest a role for HIF-1 in the menstrual cycle and demonstrate that endometrial HIF-1 activation may occur via a PGE2-regulated pathway.
Materials and Methods
Materials
All cell culture medium was purchased from Life Technologies (Paisley, Scotland, UK). Penicillin-streptomycin and fetal calf serum were purchased from PAA Laboratories Ltd. (Middlesex, UK). B-Actin mouse monoclonal antibody was purchased from Santa Cruz Biotechnology (Autogen-Bioclear, Wiltshire, UK). The polyclonal rabbit anti-EP2 receptor antibody was purchased from Cayman Chemical Co. (Alexis Corp., Nottingham, UK). The monoclonal mouse anti-HIF-1 and HIF-1 antibodies were purchased from Abcam Ltd. (Cambridge, UK). The antimouse alkaline phosphatase secondary antibodies, indomethacin, PBS, BSA, and PGE2 were purchased from Sigma Chemical Co. (Dorset, UK). The goat antimouse secondary antibody was purchased from Dako Corp. (Carpinteria, CA). The biotinylated horse antimouse IgG, biotinylated goat-antirabbit IgG, avidin and biotin were purchased from Vector Laboratories, Inc., (Peterborough, UK). ECF chemiluminescent system was purchased from Amersham Biosciences (Little Chalfont, Buckinghamshire, UK). AH 6809 (30 mM Stock in dimethylsulfoxide) and AG1478 (10 mM stock in dimethylsulfoxide) were purchased from Calbiochem (Nottingham, UK) and stored at –20 C.
Human endometrial tissue collection
Human endometrial biopsies were collected from patients undergoing hysterectomy or endometrial investigation for benign gynecological conditions (n = 54) that included pelvic pain, heavy bleeding, prolapse, and sterilization. Patients with endometriosis were excluded. Ethical approval was granted by the Lothian Research Ethics Committee, and written informed consent was obtained from each patient. All women reported regular menstrual cycles (25–35 d length) and had not taken any exogenous hormones or used an intrauterine device during the 3 months before obtaining the biopsy. At hysterectomy, a full thickness endometrial biopsy was taken from the uterine cavity of 33 of the above patients for immunohistochemical studies, including the functional and basal layer of the endometrium and adjacent myometrium. Endometrial biopsies were fixed in 4% neutral buffered formalin overnight at 4 C before routinely wax embedding using an 18-h cycle on a TP1050 processing machine (Leica Corp., Knowlhill, Milton Keynes, UK). In addition, endometrial tissue was collected from 23 of the above patients (with an endometrial tissue sampler (Pipelle, Laboratoire CCD, Paris, France); and either snap frozen in liquid nitrogen or placed in RNA Later, RNA stabilization solution [Ambion (Europe) Ltd., Cambridgeshire, UK] overnight at 4 C for subsequent RNA extraction. The biopsies were dated according to the criteria of Noyes et al. (20) based on the histological appearance, which was found to be consistent with the patient’s reported last menstrual period. Serum samples were collected from each patient at the time of endometrial biopsy collection for the determination of circulating serum progesterone and estradiol levels by RIA. A significant reduction in serum progesterone levels was detected in the late vs. the mid secretory phase (P < 0.01; see Table 1). Endometrial biopsies were classified as early proliferative (EP) (n = 5), mid proliferative (MP) (n = 5), late proliferative (LP) (n = 5), early secretory (ES) (n = 6), mid secretory (MS) (n = 5), late secretory (LS) (n = 5) or menstrual (M) (n = 2) for the immunohistochemical studies and proliferative (P) (n = 5), ES (n = 5), MS (n = 5), LS (n = 4), and M (n = 4) for the analysis of HIF-1 mRNA across the menstrual cycle.
HIF-1 and HIF-1/ARNT1 immunohistochemistry
Five-micrometer paraffin sections were dewaxed in Histoclear (National Diagnostics, Atlanta, GA) for 10 min before rehydration in descending grades of alcohol. Sections were washed with 0.01 M PBS (pH 7.4; Sigma) before pressure cooking in 0.01 M sodium citrate (pH 6) for 5 min at setting 2/high (Tefal, Clipso, Nottingham, UK) for antigen retrieval. After cooling for 20 min, the sections were again washed in PBS before blocking endogenous peroxidase activity by immersion in 3% hydrogen peroxide for 10 min at room temperature. After washing in PBS, the sections were incubated in avidin (Vector Laboratories) for 15 min, rinsed in PBS and then incubated in biotin (Vector Laboratories) for a further 15 min, all at room temperature. Slides were incubated in nonimmune rabbit serum (NRS) (where subsequent incubation was with the anti-HIF-1 antibody) or in nonimmune horse serum (NHS) (where subsequent incubation was with the anti-HIF-1 antibody) (NRS; Diagnostics Scotland, Carluke, Lanark, UK; NHS, Vector Laboratories) in PBS for 20 min at room temperature to block nonspecific binding of the primary antibody. The primary antibody, either monoclonal mouse anti-HIF-1 antibody (1:1000 dilution in NRS/PBS; IgG2b ab1) or monoclonal mouse anti-HIF-1 antibody (1:1000 dilution in NHS/PBS; IgG1 ab2771) was added to the slides and incubated overnight at 4 C. For the negative controls, the primary antibody was replaced with either nonimmune mouse IgG2b antibody (HIF-1) or IgG1 antibody (HIF-1) at a matched antibody concentration to the HIF-1 (1:40) or HIF-1 (1:1000) antibody. Subsequently, the sections were washed in PBS with added Tween 20 (PBST) before incubating in biotinylated horse antimouse antibody (Vector Laboratories) for 60 min at room temperature. After washing in PBST, an avidin-biotin-peroxidase complex (ABC-Elite; Vector Laboratories) was then applied for 60 min at room temperature. After a final wash with PBST, the chromagen 3, 3'-diaminobenzidine (Dako) was added and the reaction stopped with distilled H2O when nuclear staining was detected by inspection under the microscope. The sections were then counterstained with Harris’s hematoxylin, dehydrated, and finally mounted with Pertex (Cellpath plc, Hemel Hempstead, UK).
Scoring of immunoreactivity
Localization and intensity of immunostaining was evaluated blind by two independent observers using a semiquantitative scoring system with a three-point scale (0 = no staining, 1 = mild staining, 2 = strong staining for HIF-1) and a four-point scale (0 = no staining, 1 = mild staining, 2 = moderate staining and 3 =strong staining for HIF-1). This was applied to the glands and stromal cells in the functional and basal endometrial layers, respectively, as well as the surface epithelium. This scoring system has previously been validated (21) in a subset of tissue sections in which immunoreactivity was measured with a computerized image analysis system. A strong correlation between quantitative data derived from image analysis and subjective scores by trained observers was obtained. Statistical significance was determined using the Kruskal-Wallis nonparametric test (Instat, GraphPad Software, Inc., San Diego, CA). Dunn’s multiple comparison test was used for post hoc comparisons.
HIF-1-EP2 dual immunofluorescence
Paraffin tissue sections were dewaxed, pressure cooked, and endogenous peroxidase and biotin blocked as described above. Nonspecific binding of the primary antibody was blocked by incubating the sections for 20 min at room temperature in a 1:5 dilution of nonimmune goat serum (Diagnostics Scotland) with 5% BSA. The sections were then incubated with mouse monoclonal anti HIF-1 antibody (1:4000 dilution, IgG2b, ab1) and polyclonal rabbit anti-EP2 receptor antibody (1:50 dilution) at 4 C overnight. Negative controls were performed by incubating sections with a control mouse IgG2b antibody at a matched IgG concentration to the HIF-1 antibody (1:1300 dilution) and a control rabbit IgG at a matched concentration to the EP2 receptor antibody (1:100 dilution), or with the EP2 antibody preabsorbed overnight at 4 C with the peptide against which it had been raised (1:50 dilution). The sections were then washed in PBST before the addition of biotinylated goat-antirabbit IgG (1:200 dilution) and goat antimouse antibody (1:200 dilution) conjugated to horseradish peroxidase for 60 min at room temperature. After a further wash in PBST, the sections were incubated with tyramide cyanine 5 fluorescent complex (1:50 dilution, to detect HIF-1) for 10 min at room temperature (PerkinElmer Life Sciences, Boston, MA), washed in PBST, then incubated in a solution of avidin alexafluor 488 (1:200 dilution, Molecular Probes, Inc., Leiden, The Netherlands) to detect EP2 for 60 min at room temperature. The sections were then washed in PBST and incubated in a solution of biotin/ PBS for 20 min at room temperature (Vector Laboratories). Sections were then washed in PBST, and mounted in permafluor (Immunotech, Marseilles, France) and left to dry in the dark. The fluorescent images were collected on a Zeiss LSM 510 confocal laser scanning microscope (Carl Zeiss, Thornwood, NY). The avidin alexafluor 488 (EP2) was visualized using an argon laser with an excitation beam of 496nm, emitting at 519 nm and detected using a band pass filter from 505–550 nm. The cyanine 5 (HIF-1) was visualized using a Helium/neon 2 laser with an excitation beam of 649 nm, emitting at 670 nm and detected using a long pass filter of 650 nm.
Cell culture
Human Ishikawa endometrial adenocarcinoma cells (European Collection of Cell Culture, Centre for Applied Microbiology, Wiltshire, UK) were maintained in DMEM nutrient mixture F-12 with glutamax-1 and pyridoxine, supplemented with 10% fetal calf serum, and 1% antibiotics (stock 500 IU/ml penicillin and 500 μg/ml streptomycin) at 37 C and 5% CO2 (vol/vol). Stable EP2 transfectant cells were maintained under the same conditions with the addition of 200 μg/ml G418.
Transfection
The EP2 receptor sense (EP2S) stable cell line was constructed as described previously (19). Briefly, the EP2 receptor was isolated from proliferative phase human endometrial tissue and cloned into the eukaryotic expression vector pcDNA3.1 (Invitrogen, De Schelp, The Netherlands) containing a neomycin resistance gene expression cassette for G418 selection. The expression vector was transfected into Ishikawa cells and three independent clones, demonstrating elevated expression of EP2 receptor isolated and expanded for further investigation. We characterized all three clones and found that they all exhibited identical phenotypic and biochemical alterations. Therefore, the results of our studies reported here focused on one of these clones (clone S33).
Protein extraction
For HIF-1 studies, 2 x 106 cells were seeded in 10-cm dishes. The following day, cells were washed with PBS and incubated in serum-free culture medium containing penicillin/streptomycin and 8.4 μM indomethacin (a dual COX-enzyme inhibitor used to inhibit endogenous prostanoid biosynthesis) for at least 16 h. The next day, cells were pretreated with vehicle, 100 nM AG1478 (a specific inhibitor of EGFR kinase), 10 μM AH6809 (a specific EP2 receptor antagonist) for 1 h before stimulation with vehicle, 100 nM PGE2 or 200 μM desferrioxamine (DFO) for 48 h. After stimulation, cells were washed with ice-cold PBS. Cytosolic proteins were extracted with a cytosolic protein lysis buffer [10 mM HEPES (pH 7.8), 10 mM KCl, 2 mM MgCl2, 1 mM dithiothreitol, 0.1 mM EDTA, 1% Nonidet P-40] containing protease inhibitors (Complete mini protease inhibitor cocktail; Roche Diagnostics Ltd., Lewes, UK). Thereafter, the membrane fraction was pelleted by centrifugation at 14,000 x g for 2 min at 4 C. The clarified lysate was removed and stored and the nuclear fraction extracted with a nuclear protein lysis buffer [50 mM HEPES (pH 7.8), 50 mM KCl, 300 mM NaCl, 0.1 mM EDTA, 1 mM dithiothreitol, 10% glycerol] containing protease inhibitors (Roche) followed by brief sonication. The protein content in the nuclear fraction was determined using protein assay kits (Bio-Rad, Hemel Hempstead, UK).
Immunoprecipitation and Western blot analysis
For HIF-1 and B-actin expression, a total of 50 μg of protein was resuspended in 20 μl of Laemmli buffer [125 mM Tris-HCl (pH 6.8), 4% SDS, 5% 2-mercaptoethanol, 20% glycerol and 0.05% bromophenol blue] and boiled for 5 min. Proteins were resolved on 4–12% Tris-glycine gels (NOVEX, Invitrogen), transferred onto polyvinylidene difluoride membrane (Millipore, Watford, UK) and subjected to immunoblot analysis. Membranes were blocked for 1 h at 25 C in 4% BSA diluted in TBST (50 mM Tris-HCl, 150 mM NaCl and 0.05% vol/vol Tween 20) and incubated with specific HIF-1 (1:200) and B-actin (1:500) primary antibody. After washing and incubating with alkaline-phosphatase-conjugated secondary antibodies (Sigma), immunoreactive proteins were visualized by the ECF chemiluminescent system according to the manufacturer’s instructions. HIF-1 proteins were revealed and quantified by, and normalized to, B-actin protein expression using a Typhoon PhosphorImager 9400 (Molecular Dynamics, Amersham Biosciences). Fold increase was calculated by dividing the relative expression of HIF-1 by the relative expression of B-actin. All data are presented as mean ± SEM.
Quantitative RT-PCR (Q-RT-PCR)
HIF-1 expression in Ishikawa cells and in endometrial tissue was measured by Q-RT-PCR (Taqman) analysis. Approximately 5 x 105 cells were seeded in 5-cm dishes and allowed to attach and grow overnight. The following day, cells were synchronized by serum withdrawal for at least 16 h in serum-free medium containing 8.4 μM indomethacin. The next day, cells were pretreated with vehicle, 100 nM AG1478, or 10 μM AH6809 for 1 h before stimulation with vehicle or 100 nM PGE2 for 48 h. After stimulation, cells were washed with ice-cold PBS. RNA was extracted from cells using Tri-reagent (Sigma) and from endometrium using either Tri-reagent or QIAGEN RNeasy columns (QIAGEN Ltd., West Sussex, UK) following the manufacturers’ guidelines. Once extracted and quantified, RNA samples were reverse transcribed using MgCl2 (5.5 mM), deoxy (d) nucleotide triphosphates (0.5 mM each), random hexamers (2.5 μM), ribonuclease inhibitor (0.4 U/μl) and multiscribe reverse transcriptase (1.25 U/μl; all from PE Biosystems, Warrington, UK). The mix was aliquoted into individual tubes, and 200–400 ng of RNA were added. After mixing, samples were incubated for 60–90 min at 25 C, 45 min at 48 C, and 95 C for 5 min. Thereafter, cDNA samples were stored at –20 C. A tube with no reverse transcriptase was included to control for DNA contamination. To measure cDNA expression, a reaction mix was prepared containing Taqman buffer (5.5 mM MgCl2, 200 μM dATP, 200 μM dCTP, 200 μM dGTP, 400 μM deoxyuridine triphosphate), ribosomal 18S forward, and reverse primers and probe (50 nM), forward and reverse primers for HIF-1 (300 nM), HIF-1 probe (100 nM), AmpErase Uracil N-glycosylase (Roche Diagnostics, Lewes, Suffolk, UK) (0.01 U/μl) and AmpliTaq Gold DNA Polymerase (0.025 U/μl; PE Biosystems). After mixing, 48 μl were aliquoted into separate tubes and 2 μl/replicate of cDNA added and mixed before placing duplicate 24-μl samples into a PCR plate. A no template control (containing water) was included in triplicate. PCR was carried out using an ABI Prism 7700 (Applied Biosystems, Warrington, UK). HIF-1 primers and probe for quantitative PCR were designed using the PRIMER express program (PE Biosystems). The sequence of the HIF-1 primers and probe were: forward, 5'-CGCATCTTGATAAGGCCTC-3'; reverse, 5'-AATCACCAGCATCCAGAAG-3'; probe (FAM labeled, 6-carboxyfluorescein) 5'-TCACACGCAAATAGCTGAT-3'. The ribosomal 18S primers and probe sequences were: forward, 5'-CGG CTA CCA CAT CCA AGG AA-3'; reverse, 5'-GCT GGA ATT ACC GCG GCT-3'; probe, (VIC-labeled, PE Biosystems) 5'-TGC TGG CAC CAG ACT TGC CCT C-3'. Data were analyzed and processed using Sequence Detector version 1.6.3 (PE Biosystems). Expression of HIF-1 was normalized to RNA loading for each sample using the 18S ribosomal RNA as an internal standard.
For HIF-1 expression in Ishikawa cells results are expressed as fold increase where relative expression of HIF-1 in cells treated with PGE2 was divided by the relative expression in vehicle-treated cells. Data are presented as mean ± SEM. For endometrial tissue, results were expressed as quantity relative to a comparator, which was a sample of RNA taken from the proliferative stage of the menstrual cycle. Significant difference was determined using one-way ANOVA, and individual differences were described using the least significant difference post hoc multiple comparison (SPSS, Inc., Chicago, IL).
Results
Spatial and temporal distribution of HIF-1 protein expression (Figs. 2, a–c, and 3)
Immunostaining for HIF-1 was exclusively detected in cell nuclei consistent with the fact that stable HIF-1 is rapidly translocated to the nucleus (Fig. 2, b and c). There was no immunostaining in the negative controls (Fig. 2c, inset). During the proliferative phase, immunostaining was absent in all glands and stromal cells and negligible immunoreactivity was observed in the surface epithelium (Fig. 2a).Glandular cells and stromal cells in the functional layer (F) stained most intensely in samples collected from the menstrual stage of the cycle (Fig. 2c). HIF-1 was detected in the glands of the basal layer (B) only in the late secretory and menstrual phase (Fig. 2c). In none of the samples was there any immunostaining of the stromal cells in the basal layer (Fig. 2, a–c).
After semiquantitative evaluation and statistical analysis, the Kruskal-Wallis test, showed a significant difference (P < 0.01) in the intensity of nuclear staining in samples from different stages of the menstrual cycle (Fig. 3, a and b). The difference in intensity across the cycle stages was significant for both the glands (P < 0.01; Fig. 3a) and the stromal cells (P < 0.01; Fig. 3b). Using Dunn’s multiple comparison test, there was a significant difference in staining of the stromal cells in the functional layer in the menstrual phase when compared with the proliferative, early and mid secretory phases (Fig. 3b; P < 0.05). Glands in the basal layer in the menstrual phase also exhibited positive immunoreactivity (Fig. 3d; P < 0.05).
Spatial and temporal distribution of HIF-1/ARNT1 protein expression (Figs. 2, d–f, and 4)
In the functional layer (F) of endometrium, HIF-1/ARNT1 protein was observed in the nuclei of surface and glandular epithelial cells and stroma. Protein expression in the functional layer was greater in the proliferative compared with mid and late secretory phases of the cycle (Fig. 4, a–c; P < 0.05). In the basal region (B) of endometrium HIF-1/ARNT1 protein was greater in glandular cells in the proliferative phase (Fig. 4d; P < 0.05). Stromal cell expression of HIF-1 protein in the functional layer was also greater in the proliferative phase (Fig. 4c). Thus, overall HIF-1/ARNT1 protein immunoreactivity was maximally intense in the proliferative phase. Furthermore, HIF-1/ARNT1 protein was expressed in stromal cells at this cycle stage when HIF-1 immunoreactivity was negligible.
Expression of HIF-1 mRNA in human endometrium by Q-RT-PCR (Fig. 5)
HIF-1 mRNA was present at low levels in the proliferative and early secretory stages of the menstrual cycle. By the mid secretory phase, levels of HIF-1 had risen significantly when compared with both the proliferative (P < 0.05) and early secretory phase samples (P < 0.01). Moderate levels of HIF-1 were also present in the late secretory and menstrual phases of the cycle.
HIF-1 expression is up-regulated by PGE2-EP2 receptor interaction in endometrial epithelial cells
The role of PGE2-EP2 receptor signaling on the expression of HIF-1 mRNA was investigated by Q-RT-PCR analysis after stimulation of wild-type and EP2S Ishikawa cells with 100 nM PGE2 or vehicle (48 h). As shown in Fig. 6, PGE2 stimulation resulted in a 1.6 ± 0.03-fold increase in the expression of HIF-1 mRNA in Ishikawa EP2S cells (P < 0.05). No increase in HIF-1 expression was observed in wild-type cells treated with PGE2. Cotreatment of cells with the selective EP2 receptor antagonist (AH6809) and the EGFR kinase inhibitor AG1478 abolished the EP2 receptor-mediated elevation in HIF-1 mRNA expression (P < 0.05).
To determine whether PGE2 activation of the EP2 receptor promoted the expression of HIF-1 protein, we treated Ishikawa wild-type and EP2S cells with 100 nM PGE2 or vehicle in the presence or absence of the selective EP2 receptor antagonist (AH6809) and EGFR kinase inhibitor (AG1478) for 48 h and measured HIF-1 expression by Western blot analysis. In parallel, cells were treated with DFO, an iron chelator that inactivates prolyl hydroxylases, hence conferring HIF-1 stabilization. An increase in content of HIF-1 was observed in Ishikawa EP2S cells (Fig. 7, lane 7, P < 0.001) after treatment with PGE2 compared with cells stimulated with vehicle alone (Fig. 7, lane 6). No such increase in the expression of HIF-1 protein was observed in wild-type cells in response to PGE2 (Fig. 7, lane 2). Preincubation of cells with AH6809 (Fig. 7, lane 8) and AG1478 (Fig. 7, lane 9) abolished the PGE2-mediated elevation in HIF-1 protein expression in EP2S cells. Stimulation of Ishikawa cells with DFO resulted in a 4.8 ± 1.1 (Fig. 7, lane 5, P < 0.001) and 3.5 ± 0.2 (Fig. 7, lane 10, P < 0.001)-fold increase in the expression of HIF-1 in wild-type and EP2S cells, respectively.
Colocalization of HIF-1 protein and EP2 receptor protein in human endometrium
Dual immunofluorescence has demonstrated colocalization of HIF-1 and EP2 receptor proteins to the glandular epithelial cells (Fig. 8). The EP2 receptor protein is expressed at all stages of the menstrual cycle in the cytoplasm of the glandular epithelium and vascular endothelium (22). As described earlier HIF-1 protein is expressed particularly in the late secretory and menstrual phases in the nuclei of glandular cells.
Discussion
In this study, we have demonstrated tight temporal and spatial confinement of HIF-1 expression across the menstrual cycle. HIF-1 protein was expressed with increasing intensity from premenstrual (secretory) through to menstrual phase human endometrium. No HIF-1 immunoreactivity was detected in endometrial biopsies from the proliferative phase. Within the endometrial layers, the strongest immunoreactivity observed was in the uppermost endometrial zones (functional layer), with staining localized to glandular epithelium, surface epithelium and some stromal cells. HIF-1 mRNA expression peaked in the mid secretory phase, before maximal HIF-1 protein expression. The temporal and spatial distribution of HIF-1 protein and mRNA detected in this study would thus seem to be related to the process of menstruation. HIF-1 appears to be most strongly expressed during the perimenstrual phase in those endometrial layers that will be sloughed off at menstruation. In contrast, HIF-1/ARNT1 protein expression was maximal in the glandular component during the proliferative phase, that is, at a time of postmenstrual repair. Endometrial tissue collection at the time of hysterectomy permits for maintenance of the spatial orientation of the tissue. This is not possible with endometrial curettage sampling.
One previous study has also reported on the expression of HIF-1 proteins in human endometrium (23). The authors failed to detect HIF-1 nuclear staining in any endometrial samples, finding only fairly minor cytoplasmic staining without a specific temporal or spatial pattern. HIF-1 expression was also reported and found to be predominantly epithelial and strongest after ovulation, similar to our observations herein. The discrepancies in findings of HIF-1 expression may reflect the use of different primary antibodies and different methods of tissue collection.
The presence of HIF-1 protein in the glandular epithelium may be responsible for the up-regulation of VEGF that has been reported at the time of progesterone withdrawal in the uppermost zones of the endometrium (13, 24).
HIF-1 is recognized as a critical mediator of hypoxic responses, and hypoxia is certainly the best studied and probably the most potent stimulus for HIF-1 induction. Virtually all cell types appear to express stable HIF-1 protein under conditions of reduced oxygen tension, implying that the oxygen-sensing apparatus is ubiquitous. For this reason, the detection of HIF-1 in vitro or in animal models is often considered to be an indicator of cell or tissue hypoxia. However, the observations in the present study suggest that within the endometrium, hypoxia may not be the only or indeed the most important stimulus for HIF-1 expression. During the premenstrual and menstrual phase, HIF-1 was found to be strongly expressed in glandular epithelium, whereas adjacent stromal cells stained weakly or not at all. Because adjacent cells are unlikely to be exposed to significantly different oxygen tensions, this observation implies an alternative receptor-specific pathway to HIF-1 induction. We have shown that the temporal expression of HIF-1 in the endometrium coincides with progesterone withdrawal (see Table 1). We have previously shown that this latter event results in expression of COX-2 and subsequent generation of the potent vasoactive prostaglandins PGE2 and PGF2 (1, 25). PGE2 is a vasodilator and inducer of vascular leak, whereas PGF2 is a potent vasoconstrictor. Although the latter may induce HIF-1 expression through local tissue ischemia, we here have shown that endometrial HIF-1 expression may be regulated directly by PGE2 as HIF-1 protein colocalized with the EP2 receptor and PGE2 promoted HIF-1 mRNA and protein expression via the EP2 receptor in normoxic conditions.
PGE2 couples to four pharmacologically classified subtypes of G protein-coupled receptor (EP1-EP4) (26). These subtypes may be coexpressed on the same cell or on adjacent cells, indicating an autocrine/paracrine control of an autocoid biosynthesized and released in close proximity to the site of its action (27). Thus, the signal transduction from PGE2-EP ligand-receptor binding to activation of HIF-1 and VEGF expression is likely to be complex. In the present study, we have demonstrated that HIF-1 mRNA and protein expression are induced in endometrial epithelial cells by PGE2 via the EP2 receptor. These findings are in agreement with a similar study by Liu et al. (28), where HIF-1 expression in human prostate cancer cells was induced, with EP2 and EP4 (not EP3) receptor-selective agonists, demonstrating a role for specific EP receptors in the PGE2-regulation of HIF-1 expression. Inhibition of EGFR kinase activity abolished the PGE2-induced expression of HIF, indicating that trans-phosphorylation of the EGFR is also a requirement for the PGE2-induced expression of HIF via the EP2 receptor in endometrial epithelial cells. In a recent study, we have shown that EP2 receptor localizes to glandular epithelial cells and no variation in expression is detected in endometrial biopsies collected across the menstrual cycle (22). Moreover, PGE2-EP2 receptor signaling promotes the expression and release of VEGF from endometrial epithelial cells via the intracellular trans-phosphorylation of the EGFR and activation of the ERK1/2 signaling pathways (19). Our data herein, by inference from the above, strongly suggests this pathway is dependent on HIF-1. PGE2 is also synthesized across the menstrual cycle, albeit its pattern of synthesis may vary depending on the stage of the menstrual cycle. Hence, it is our belief that the EP receptors can be activated during the late secretory phase to affect HIF expression. Recent studies of human colonic cancers have demonstrated a correlation between the expression of the COX-2 enzyme, and its biosynthesized product prostaglandin E2 with VEGF expression and angiogenesis (29, 30) and inhibition of HIF-1 expression by RNA interference blocked the induction of VEGF mRNA (15). Further studies elucidating the relationship between EP2 receptor signaling, HIF and VEGF in endometrial epithelial cells may provide insight into some of the complex processes that occurs in the endometrium during menstruation, such as neovascularization and endometrial regeneration.
HIF-1/ARNT1 is considered to be a constitutively expressed cytoplasmic protein and nuclear HIF-1 translocation is dependent upon its dimerization with HIF-1/ARNT1. Therefore, the observation that HIF-1/ARNT expression varied in a specific temporal and spatial distribution during the menstrual cycle, and that the temporal pattern of expression was reciprocal to that seen with HIF-1, was unexpected. For example, HIF-1 expression in glandular epithelium was strongest during the proliferative phase when HIF-1 expression was absent. HIF-1/ARNT1 is a basic helix-loop-helix/PER-ARNT-SIM-harboring protein and is also expressed as two further isoforms, ARNT2 and ARNT3. All three ARNT proteins may dimerize with HIF-1 (7). The monoclonal antibody used in our current studies is specific for HIF-1/ARNT1 and does not detect ARNT2 or ARNT3. Further studies are required to determine the expression and role of ARNT2 and 3 in human endometrium. However, recent data on HIF expression in the preimplantation mouse uterus supports the notion that specific ARNT family members are differentially expressed in uterine tissue and may only dimerize with HIF-1 at specific stages of implantation (31). In addition to a role in HIF-1mediated pathways, the ARNT family is a critical component of the xenobiotic detoxification pathway. The aryl hydrocarbon (Ahr) receptor binds to nuclear ARNTs and induces the transcription of genes encoding xenobiotic-metabolizing enzymes. Little is known about the functional relevance of this pathway in endometrium However, AhR-deficient mice have a complex reproductive deficit and abnormal uterine physiology (32, 33, 34, 35), suggesting that the Ahr pathway may play an integral part in the normal reproductive physiology and not just in xenobiotic biology. Interestingly, the pattern of HIF-1/ARNT1 expression observed in our studies, that is, epithelial zone and predominantly proliferative phase, is broadly similar to the pattern of AhR expression described in human endometrium (36). The implication is that in human endometrium, ARNT1 may play a functional role in the AhR-mediated pathway rather than HIF-1mediated pathways.
Our finding that HIF-1 expression occurred maximally at the menstrual stage is open to interpretation. It is possible that its expression is of particular significance in initiating the regeneration and repair of the endometrium at the beginning of the subsequent cycle. VEGF and a number of other angiogenic genes have been observed in several model systems to be transcriptionally regulated by HIF-1 (7). A positive correlation between HIF-1 expression, VEGF expression and increased tissue microvessel density in human endometrial carcinoma has been established (37). In addition, vasculoprotective genes such as hemoxygenase-1 and inducible nitric oxide synthase are also HIF-1 regulated and may serve to limit tissue injury during ischemia/reperfusion in the endometrium. Although our studies demonstrate that HIF-1 expression may be induced in an endometrial cell line after PGE2 stimulation under normoxic conditions, this does not imply that hypoxia is not relevant to HIF-1 signaling in intact endometrial tissue. We hypothesize that dual regulation of HIF-1 in endometrium may be present through both a hypoxia-independent, PGE2-dependent mechanism and a classical hypoxia-dependent mechanism (see Fig. 1).
In conclusion, this study demonstrates the cell- and stage-specific expression of HIF-1 across the human menstrual cycle and implicates a mechanism for HIF-1 generation through PGE2 stimulation via the EP2 receptor. Further studies are required to determine the additional role of tissue hypoxia in HIF-1 regulation and to clarify the precise functional role of this transcription factor in initiating menstruation or in the subsequent regeneration and repair of the endometrium.
Acknowledgments
We acknowledge the advice provided by Professor R. W. Kelly (Medical Research Council Human Reproductive Sciences Unit, Edinburgh) on the Q-RT-PCR studies. We also acknowledge the assistance of our clinical research nurses Catherine Murray and Sharon McPherson for assistance with patient recruitment and endometrial tissue collection. The secretarial assistance of Kate Williams and Meg Anderson is acknowledged as well as the help provided by Ted Pinner and Corrine McLeod with the illustrations.
Footnotes
This work was supported by UK Medical Research Council Programme Grant No. 0000066.
Disclosure statement: J.O., T.A.H., L.B., K.J.S., H.N.J., and N.H. have nothing to declare. H.O.D.C. has received support, for invited presentations at scientific meetings on management of bleeding problems, from Schering (but not on the topic of the scientific content of this manuscript).
First Published Online November 10, 2005
Abbreviations: Ahr, Aryl hydrocarbon receptor; ARNT1, aryl receptor nuclear translocator 1; COX-2, cyclooxygenase-2; DFO, desferrioxamine; EGFR, epidermal growth factor receptor; EP2, prostaglandin E-series prostanoid receptor 2; HIF-1, hypoxia-inducible factor; NHS, nonimmune horse serum; NRS, nonimmune rabbit serum; PBST, PBS with Tween 20; PGE2 and PGF2, locally synthesized prostaglandins; Q-RT-PCR, quantitative RT-PCR; VEGF, vascular endothelial growth factor.
Accepted for publication October 18, 2005.
References
Critchley HO, Kelly RW, Brenner RM, Baird DT 2001 The endocrinology of menstruation—a role for the immune system. Clin Endocrinol (Oxf) 55:701–710
Critchley HK, Kelly RW, Brenner RM, Baird DT 2003 Antiprogestins as a model for progesterone withdrawal. Steroids 68:1061–1068
Markee JE 1940 Menstruation in intraocular endometrial transplants in the rhesus monkey. Contrib Embryol 177:211–308
Popovici RM, Irwin JC, Giaccia AJ, Guidice LC 1999 Hypoxia and cAMP stimulate vascular endothelial growth factor (VEGF) in human endometrial stromal cells: potential relevance to menstruation and endometrial regeneration. J Clin Endocrinol Metab 84:2245–2248
Critchley HO, Kelly RW, Lea RG, Drudy TA, Jones RL, Baird DT 1996 Sex steroid regulation of leukocyte traffic in human decidua. Hum Reprod 11:2257–2262
Sugino N, Karube-Harada A, Taketine T, Sakata N, Nakamura Y 2004 Withdrawal of ovarian steroids stimulates prostaglandin F2alpha production through nuclear factor-B activation via oxygen radicals in human endometrial stromal cells: potential relevance to menstruation. J Reprod Dev 50:215–225
Semenza GL 2000 HIF-1: mediator of physiological and pathophysiological responses to hypoxia. J Appl Physiol 88:1474–1480
Pugh CW, Ratcliffe PJ 2003 Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med 9:677–684
Poellinger L, Johnson RS 2004 HIF-1 and hypoxic response: the plot thickens. Curr Opin Genet Dev 14:81–85
Krishnamachary B, Berg-Dixon S, Kelly B, Agani F, Feldser D, Ferreira G, Iyer N, LaRusch J, Pak B, Taghavi P, Semenza GL 2003 Regulation of colon carcinoma cell invasion by hypoxia-inducible factor. Cancer Res 63:1138–1143
Smith SK 2001 Regulation of angiogenesis in the endometrium. Trends Endocrinol Metab 12:147–151
Higgins DF, Biju MP, Akai Y, Wutz A, Johnson RS, Haase VH 2004 Hypoxic induction of CTGF is directly mediated by HIF-1. Am J Physiol Renal Physiol 287:F1223–F1232
Sharkey AM, Day K, McPherson A, Malik S, Licence D, Smith SK, Charnock-Jones DS 2000 Vascular endothelial growth factor expression in human endometrium is regulated by hypoxia. J Clin Endocrinol Metab 85:402–409
Haddad JJ, Land SC 2001 A non-hypoxic, ROS-sensitive pathway mediates TNF--dependent regulation of HIF-1. FEBS Lett 505:269–274
Fukuda R, Kelly B, Semenza GL 2003 Vascular endothelial growth factor gene expression in colon cancer cells exposed to prostaglandin E2 is mediated by hypoxia-inducible factor 1. Cancer Res 63:2330–2334
Blouin CC, Page EL, Soucy GM, Richard DE 2004 Hypoxic gene activation by lipopolysaccharide in macrophages: implication of hypoxia-inducible factor 1. Blood 103:1124–1130
Seno H, Oshima M, Ishikawa T, Oshima H, Takaku K, Chiba T, Narumiya S, Taketo MM 2002 Cyclooxygenase-2 and prostaglandin E2 receptor EP2-dependent angiogenesis in APC 716 mouse intestinal polyps. Cancer Res 62:506–511
Eibl G, Bruemmer D, Okada Y, Duffy JP, Law RE, Reber HA, Hines OJ 2003 PGE2 is generated by specific COX-2 activity and increases VEGF production in COX-2-expressing human pancreatic cancer cells. Biochem Biophys Res Commun 306:887–897
Sales KJ, Maudsley S, Jabbour HN 2004 Elevated prostaglandin EP2 receptor in endometrial adenocarcinoma cells promotes VEGF expression via cAMP-mediated transactivation of the EGF Receptor and ERK1/2 signalling pathways. Mol Endocrinol 18:1533–1545
Noyes RW, Hertig AT, Rock J 1950 Dating the endometrial biopsy. Fertil Steril 1:3–25
Wang H, Critchley HOD, Kelly RW, Shen D, Baird DT 1998 Progesterone receptor subtype B is differentially regulated in human endometrial stroma. Mol Hum Reprod 4:407–412
Milne SA, Perchick GB, Boddy SC, Jabbour HN 2001 Expression, localization, and signalling of PGE2 and EP2/EP4 receptors in human nonpregnant endometrium across the menstrual cycle. J Clin Endocrinol Metab 86:4453–4459
Zhang J, Salamonsen LA 2002 Expression of hypoxia-inducible factors in human endometrium and suppression of matrix metalloproteinases under hypoxic conditions do not support a major role for hypoxia in regulating tissue breakdown at menstruation. Hum Reprod 17:265–274
Nayak NR, Brenner RM 2002 Vascular proliferation and vascular endothelial growth factor expression in the rhesus macaque endometrium. J Clin Endocrinol Metab 87:1845–1855
Baird DT, Cameron ST, Critchley HO, Drudy TA, Howe A, Jones RL, Lea RG, Kelly RW 1996 Prostaglandins and menstruation. Eur J Obstet Gynecol Reprod Biol 70:15–17
Coleman RA, Smith WL, Narumiya S 1994 International Union of Pharmacology classification of prostanoid receptors: properties, distribution, and structure of the receptors and their subtypes. Pharmacol Rev 46:205–229
Ashby B 1998 Co-expression of prostaglandin receptors with opposite effects: a model for homeostatic control of autocrine and paracrine signaling. Biochem Pharmacol 55:239–246
Liu XH, Kirschenbaum A, Lu M, Yao S, Dosoretz A, Holland JF, Levine AC 2002 Prostaglandin E2 induces hypoxia-inducible factor-1 stabilization and nuclear localization in a human prostate cancer cell line. J Biol Chem 277:50081–50086
Cianchi F, Cortesini C, Bechi P, Fantappie O, Messerini L, Vannacci A, Sardi I, Baroni G, Boddi V, Mazzanti R, Masini E 2001 Up-regulation of cyclooxygenase 2 gene expression correlates with tumor angiogenesis in human colorectal cancer. Gastroenterology 121:1339–1347
Tsujii M, Kawano S, Tsuji S, Sawaoka H, Hori M, DuBois RN 1998 Cyclooxygenase regulates angiogenesis induced by colon cancer cells. Cell 93:705–716
Daikoku T, Matsumoto H, Gupta RA, Das SK, Gassmann M, DuBois RN, Dey SK 2003 Expression of hypoxia-inducible factors in the peri-implantation mouse uterus is regulated in a cell-specific and ovarian steroid hormone-dependent manner. Evidence for differential function of HIFs during early pregnancy. J Biol Chem 278:7683–7691
Fernandez-Salgureo P, Pineau T, Hilbert DM, McPhail T, Lee SS, Kimura S, Nebert DW 1995 Immune system impairment and hepatic fibrosis in mice lacking the dioxin-binding Ah receptor. Science 268:722–726
Pocar P, Fischer B, Klonisch T, Hombach-Klonisch S 2005 Molecular interaction of the aryl hydrocarbon receptor and its biological and toxicological relevance for reproduction. Reproduction 129:379–389
Mimura J, Yamashita K, Nakamura K, Morita M, Takagi TN, Nakao K, Ema M, Sogawa K, Yasuda M, Katsuki M, Fujii-Kuriyama Y 1997 Loss of teratogenic response to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in mice lacking the Ah (dioxin) receptor. Genes Cells 2:645–654
Fernandez-Salguero PM, Ward JM, Sundberg JP, Gonzalez FJ 1997 Lesions of aryl-hydrocarbon receptor-deficient mice. Vet Pathol 34:605–614
Kuchenhoff A, Seliger G, Klonisch T, Tscheudschilsuren G, Kaltwasser P, Seliger E, Buchmann J, Fischer B 1999 Arylhydrocarbon receptor expression in the human endometrium. Fertil Steril 71:354–360
Sivridis E, Giatromanolaki A, Koukourakis MI, Georgiou L, Anastasiadis P 2002 Patterns of episialin/MUC1 expression in endometrial carcinomas and prognostic relevance. Histopathology 40:92–100(Hilary O. D. Critchley, Julia Osei, Tere)