| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Vincent Center for Reproductive Biology, Vincent Obstetrics and Gynecology Service, Massachusetts General Hospital, and Department of Obstetrics, Gynecology and Reproductive Biology, Harvard Medical School, Boston, Massachusetts 02114
ABSTRACT
Signaling mechanisms coordinating uterine angiogenesis and tissue remodeling during decidualization are not completely understood. Prostanoid signaling is thought to play a functionally important role in each of these events. In the present study, we demonstrate that the subfamily of G-protein-coupled receptors that binds and becomes activated by the terminal signaling lipid in the sphingolipid pathway, sphingosine-1-phosphate (S1P), were expressed during uterine decidualization. Three of the five known S1P receptors, termed endothelial differentiation genes (Edg; Edg1, Edg3, and Edg5) were upregulated in the uterine deciduum from Day of Pregnancy (DOP) 4.5 to 7.5, while Edg6 and Edg8 expression remained unchanged. Consistent with angiogenesis in general during decidualization, we believe EDG1 and EDG5 to be regulated by the embryo because no microvascular expression for these receptors was observed in oil-induced deciduomas. Observed expression of EDG1 and EDG5 showed a similar expression pattern to that previously reported for prostaglandin-endoperoxide synthase 2 (PTGS2), transitioning from the sublumenal stromal compartment in the antimesometrial pole (DOP 5) to the microvasculature of the mesometrial pole (DOP 7). Furthermore, these two receptors colocalized with PTGS2 at three additional sites at the maternal:fetal interface throughout pregnancy. Treatment of cultured predecidualized stromal cells with S1P resulted in upregulation of Ptgs2 mRNA and PTGS2 protein, but not the downstream enzyme prostacyclin synthase. These combined results suggest the existence of a link between the sphingolipid and prostanoid signaling pathways in uterine physiology, and that, based on their expression pattern, S1P receptors function to coordinate uterine mesometrial angiogenesis during the implantation phase of early gestation.
angiogenesis, cyclooxygenase-2, decidua, developmental biology, G-protein-coupled receptor, implantation, pregnancy, signal transduction, sphingosine-1-phosphate, uterus
INTRODUCTION
Implantation of the embryo in the uterus is a complex process, and its failure is broadly considered an impediment to successful pregnancy and improving assisted reproduction [1]. In many species, including humans and laboratory rodents, the endometrial stromal compartment undergoes a postnatal developmental paradigm during implantation called decidualization [2, 3]. In mice, decidualization begins shortly after the blastocyst adheres to the uterine epithelial lining, resulting in rapid proliferation and differentiation of the underlying stroma with subsequent displacement and erosion of the epithelium. The expanded stromal tissue, referred to as the deciduum, serves a variety of functions that are essential for the establishment and maintenance of pregnancy. The deciduum is by definition a secretory tissue in that it produces a variety of important endocrine and paracrine signaling molecules, such as prolactin-related proteins [4], interleukins [5], cytokines [5], and prostanoids [6], among others. Additional known functions of the deciduum include its immunosuppressive actions [7, 8], control of trophoblast growth and cell migration [9], nutrient value following programmed cell death for the expanding trophectoderm [2, 3], and it also provides a vascular network for nutrient/gas exchange for the embryo before development of the placenta [10]. Early decidualization is controlled primarily by maternal factors because this developmental process can be induced in the absence of an embryo. Indeed, the hormonally primed uterus can be stimulated to undergo decidualization upon provision of a mechanical stimulus, such as intrauterine infusion of sesame oil or by scratching the endometrium. The resultant endometrial structure (i.e., deciduoma) is similar to the embryo-induced deciduas in many respects, but lacks substantial mesometrial vasculature. In addition to nutrient/gas exchange, the vascular component of the deciduum also serves as a conduit for invading trophoblast cells that use the maternal blood vessels as a blueprint for the developing placental vasculature. As such, faulty uterine angiogenesis during decidualization will likely result in development of an inadequate placenta. To date, three signaling pathways that coordinate uterine angiogenesis in a number of species have been studied in some detail. These include the vascular endothelial growth factor (VEGF) [11], angiotensin [12], and prostanoid [13] signaling pathways. Using mutant mice, the prostanoid pathway has been shown to be functionally important for stromal cell decidualization [14], as well as mesometrial vascular development via activation of the peroxisome proliferator-activated receptor delta [15], and VEGF signaling [13].
It has been known for over a decade that the proapoptotic sphingolipid ceramide induces expression of prostaglandin-endoperoxide synthase 2 (Ptgs2) and synthesis of prostaglandin E2 (PGE2) [16]. However, only within the past few years has it been demonstrated that other sphingolipids also regulate prostanoid synthesis. The downstream metabolic by-product of ceramide, sphingosine-1-phosphate (S1P), upregulates PTGS2 protein expression and PGE2 synthesis in human amniotic Wistar Institute Susan Hayflich (WISH) cells [17]. In human smooth muscle cells, S1P increased expression of prostaglandin I2 (PGI2 or prostacyclin) synthase and production of PGI2, thus establishing a clear link between sphingolipids and prostanoids [18].
Because prostanoid signaling plays such a prominent role in uterine decidualization and angiogenesis and serves an apparently important immunomodulatory function, we sought to identify other, as yet, uncharacterized signaling mechanisms in the uterus that regulate prostanoid metabolizing enzymes during early gestation. Because of the previously established connection between prostaglandin and sphingolipid pathways, one objective here was to begin studying expression of the five known high-affinity G-protein-coupled S1P receptors in the uterus during the implantation phase of pregnancy. It was also of interest to firmly establish a link, either in parallel or series, between the sphingolipid and prostanoid pathways in uterine stromal cells of pregnancy.
MATERIALS AND METHODS
Animals
Sexually mature female and male ICR mice were paired to establish pregnancy. Female mice were considered Day of Pregnancy (DOP) 0.5 upon observation of a vaginal semenal plug. Uterine tissue or whole implantation sites were collected on DOPs 3.5, 4.5, 5, 5.5, 7.5, and 9.5 and prepared for cell culture, RNA isolation, or immunohistochemistry, as described below. To generate oil-induced deciduomas, female mice were placed with vasectomized male mice until a vaginal semenal plug was observed on Day 0.5 of pseudopregnancy. Sesame oil (10 µl) was then infused into the uterine lumen immediately below the utero-tubal junction to induce decidualization on Day 4.5 of pseudopregnancy. Uterine tissue was collected 72 h later, a time corresponding to DOP 7.5. All animal protocols were reviewed and approved by the Massachusetts General Hospital institutional care and use committee and were performed in accordance with the Public Health Service Policy on Humane Care and Use of Laboratory Animals.
RNA Isolation, RT-PCR, and Northern Blot Analysis
Total cellular RNA was isolated from stromal cells grown in culture or from implantation sites on DOP 4.5, 7.5, or 9.5 using Tri-reagent as described by the manufacturer (Sigma Chemical Co., St. Louis, MO). On DOP 4.5, endometrial tissue was enriched by scraping the uterus with a scalpel blade following identification of implantation sites with the blue-dye method (intravenous injection of 100 µl 0.1% Evans blue dye in saline, wt/vol). On DOP 7.5 and 9.5, decidualized endometrial tissue was enriched by scraping the deciduum from the myometrium and microdissecting away the embryos and extraembryonic membranes. After DNase I (Promega, Madison, WI) treatment to remove possible traces of contaminating genomic DNA, 2 µg of total RNA was reverse transcribed using oligo (dT) primer (Promega) and SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA). Mock reverse transcriptase reactions were performed for each RNA sample in which reverse transcriptase was omitted from the reaction to further confirm absence of contaminating genomic DNA. Semiquantitative PCR was performed by first amplifying message for the ribosomal protein L7 (Rpl7) in all RNA samples. The expression level of Edg1, Edg3, Edg5, Edg6, Edg8 from decidual tissues on DOPs 4.5–9.5 and untreated predecidualized stromal cell cultures, as well as Ptgs2 and PGI2 synthase from primary cultures in response to S1P treatment, was then evaluated. All PCR products were sequenced to confirm primer specificity following amplification with the primer sets listed in Table 1. Northern blot analysis, as described in detail elsewhere [19], was then used to validate RT-PCR results for Edg1 and Edg5. Briefly, 10 µg of total RNA was separated by electrophoresis on a formaldehyde denaturing gel. The RNA was transferred to nylon membrane (Amersham Biosciences, Piscataway, NJ) and then probed with radiolabeled (dCTP; Perkin Elmer; Boston, MA) Edg1, Edg5, or Ptgs2 cDNA (Random Primers Labeling Kit; Invitrogen). Blots were then reprobed with a radiolabeled 18S rRNA cDNA (Ambion; Austin, TX), which served as an internal control.
|
Immunohistochemistry
Dissected tissues were fixed in buffered 4% (wt/vol) paraformaldehyde followed by paraffin embedding. Medial sections (6 µm) of implantation sites were deparaffinized with xylenes followed by graded rehydration in ethanol (100%, 95%, 80%, and 70%; vol/vol) and distilled water. Following peroxidase quenching (5 min in methanol containing 3% hydrogen peroxide, vol/vol), antigens were unmasked by high temperature (10-min boiling in microwave) antigen retrieval in which sections were immersed in 10 µM sodium citrate buffer (pH 6.5) [20]. After equilibration in phosphate buffered saline (PBS) and blocking (2% bovine serum albumin [wt/vol] and 1% normal donkey serum [vol/vol]), sections were incubated overnight at 4°C with primary antibody diluted (EDG1, EDG5, and PTGS2; 1:200; Santa Cruz Biotechnologies, Santa Cruz, CA) in TNK buffer (0.1 M Tris, pH 7.6; 0.55 M NaCl; 0.01 M KCl). Sections were then washed (PBS, 3 x 10 min each) and incubated for 1 h at room temperature with a 1:200 dilution of biotinylated secondary antibody (1:200; Santa Cruz Biotechnologies), followed by washing and a final incubation with horseradish peroxidase-conjugated streptavidin for 45 min at room temperature (Vector Laboratories; Burlingame, CA). After washing as before, sections were exposed to 3,3'-diaminobenzidine substrate for 2–10 min, counterstained with hematoxylin, dehydrated in ethanol and xylenes, and mounted for light microscopy. For confocal microscopy, Alexa 488- and 568-conjugated secondary antibodies (Invitrogen) were used in place of biotin-conjugated secondary antibodies. Antigen specificity was confirmed with negative controls in which tissue sections were prepared as described above, but without primary antibody.
In Vitro Predecidualized Stromal Cell Cultures
Stromal cell cultures were established using uterine stromal tissue isolate from DOP 3.5 of pregnancy. Uteri were removed from female mice, cut longitudinally to expose the endometrium, and then placed in serum-free Dulbecco modified Eagle medium (DMEM) containing 6 mg/ml dispase and 25 mg/ml pancreatin (1 h at 4°C, 1 h at 22°C, 10 min 37°C; Invitrogen) to remove the epithelium [21]. Partially digested uteri were then washed twice in DMEM and placed in DMEM containing 0.5 mg/ml collagenase (Sigma Chemical Co.). After incubation for 45 min at 37°C, tubes were mixed gently for 10–12 sec until the supernatant became turbid with dispersed endometrial stromal cells. The supernatant was collected and filtered through a nylon filter (40-µm pore size) to remove cell clusters. The resultant supernatant was then placed in a sterile tube, centrifuged at 450 x g for 10 min at room temperature, and the pellet of cells was rinsed once with DMEM. After the second centrifugation, cells were resuspended in complete culture medium [DMEM, 10% charcoal stripped heat-inactivated serum (vol/vol), 10 ng/ml progesterone, 1% amphoteracin B (vol/vol), 100 µg/ml streptomycin, and 100 U/ml penicillin) and plated at a density of 25000 cells/cm2. Culture dishes were placed at 37°C for 30 min to allow stromal cells to adhere. Medium containing endothelial cells, dead cells, and cellular debris was removed and replaced with fresh medium. Adherent predecidualized stromal cells were cultured for approximately 48 h (70–80% confluence), at which time new medium was provided followed by treatment with S1P (1µM).
Western Blot Analysis
Protein lysates were collected for Western blot analysis as described in detail [22, 23]. Following separation by SDS-PAGE using the NuPage System (Invitrogen), proteins (15 µg/lane) were transferred (100 V, 1 h) to polyvinylidene difluoride membranes. Nonspecific binding was blocked with 5% fat-free milk (wt/vol) in TBST buffer (50 mM Tris-HCl [pH 7.5] 0.15 NaCl, 0.05% Tween-20 [vol/vol]) for 1 h at room temperature. The phosphorylated (i.e., active) forms of mitogen activate protein kinases (MAPK) 1, 3, 8, 9, 14, and AKT and were detected by Western blot analysis with commercial polyclonal antibodies (phospho-MAPK1/3, 1:5000, Promega; pan-MAPK14, 1:1000, Cell Signaling, Beverly, MA; phospho-MAPK8/9, 1:100, Santa Cruz Biotechnology; phospho-Akt, 1:1000, Santa Cruz Biotechnology; overnight incubation at 4°C). PTGS2 protein was detected with polyclonal antiserum at a dilution of 1:200 (Santa Cruz Biotechnology). Membranes were then washed (3 x 10 min each) in TBST buffer and incubated with anti-rabbit IgG horseradish peroxidase conjugate or anti-mouse IgG horseradish peroxidase conjugate (1:200; Santa Cruz Biotechnology) for 1 h at room temperature. The membranes were washed with TBST as before and bound antibody was detected using enhanced chemiluminescent reagents based on the manufacturer's recommendations (Amersham). To verify equal protein loading, membranes were then stripped and reprobed with actin antibody (1:1000; Santa Cruz Biotechnologies).
Endothelial Cell Isolation
Decidual tissue, dissected free of embryo, extraembryonic membranes, and myometrium, was obtained from the mesometrial pole of Day 7.5 implantation sites (experiment completed in duplicate). The decidual tissue was minced with scalpel blades and placed in DMEM containing 0.2 mg/ml collagenase (wt/vol) and allowed to incubate at 37°C for 30 min. After brief agitation to dissociate the tissue completely, the cell suspension was filtered through a nylon mesh (40-µm pore size). Following centrifugation, cells were suspended in PBS containing 0.1% bovine serum albumin and precleared antiplatelet/endothelial cell adhesion molecule 1 (PECAM1) antibody (1.25 µg/106 cells). The cell suspension was allowed to incubate on a rocker at 4°C for 30 min. Following centrifugation and a single wash with PBS, precleared immunobeads coupled to protein G (Dynal Invitrogen Corporation; Carlsbad, CA) were added to the cells at an estimated concentration of four beads/cell. Cells were then rinsed as before following a 30-min incubation at 4°C. RNA was isolated from both the PECAM1-positive and PECAM1-negative decidual cell fractions for RT-PCR analysis using standard procedures described above.
Experimental Replication and Statistical Analysis
Each experiment was independently replicated at least three times, with different mice being used in each experiment. All graphs represent the mean + SEM from replicated experiments. Micrographs, autoradiographs, and scanned images shown in figures are representative of at least three independent experiments. Assignment to DOP or treatments was made at random. Raw data were analyzed by one-way analysis of variance, and Duncan multiple range test or Tukey multiple comparison test were used to identify significant differences in mean expression values during DOP or in response to S1P treatment, where P < 0.05 was chosen to indicate a statistically significant difference.
RESULTS
Uterine Decidual S1P Receptor mRNA Expression
Most cellular responses to S1P are mediated through activation of one or more of five known high-affinity G-protein-coupled receptors called S1P receptors. Our first experiment was to demonstrate expression of S1P receptors at the maternal:embryo interface throughout decidualization (DOP 4.5–9.5). It was revealed through RT-PCR that all five receptors were expressed in the deciduum on Days 4.5, 7.5, and 9.5 (Fig. 1a). However, only Edg1, Edg3, and Edg5 were developmentally regulated, showing an increase in expression between DOPs 4.5 and 7.5. Receptors Edg6 and Edg8, while expressed at very low levels and generally confined to immune cells in other tissues [32], were not upregulated in response to decidualization. Northern blot analysis (Fig. 1b) was then used to validate semiquantitative RT-PCR results. Edg1 and Edg5 were found by Northern blotting to be abundantly expressed and significantly (P < 0.01, Fig. 1c) upregulated from DOP 4.5 to DOP 7.5.
|
Immunolocalization of EDG1 and EDG5
To identify the decidual cellular constituent(s) expressing EDG1 and EDG5 protein, both of which were developmentally upregulated at the mRNA level during decidualization, immunohistochemical analysis was performed on medial, longitudinal sections of implantation sites from DOP 4.5 to DOP 9.5. Of note, because expression for both S1P receptors was identical during all stages of implantation studied, we show and describe data here for only EDG1. EDG1 was weakly expressed in the lumenal and glandular epithelium on DOP 4.5, with more intense staining in the sublumenal antimesometrial stromal cell compartment (Fig. 2a). By DOP 7.5, EDG1 expression transitioned from the antimesometrial compartment to the microvasculature of the mesometrium (Fig. 2b), with retention of some weak staining in decidualizing stromal cells throughout the deciduum. The most intense staining was detected in uterine cells directly adjacent to giant trophoblast cells. On DOP 9.5, EDG1 expression was observed in what appeared to be the microvascular network of the mesometrial deciduum (Fig. 2c). No staining was observed in sections in which primary antibodies were omitted (negative controls, see Fig. 2d). To demonstrate that Edg1 and Edg5 were expressed in endothelial cells, we isolated decidual PECAM1-postitive cells from the mesometrial pole of Day 7.5 implantation sites using a magnetic immunobead approach. As shown in Figure 2e, PECAM1-positive cells expressed Edg1 and Edg5, while PECAM1-negative cells expressed somewhat lower levels of each transcript. These data are consistent with our immunohistochemical results, thus supporting a role for EDG1 and EDG5 in both endothelial and stromal cell functions.
|
Embryonic Regulation of S1P Receptor Expression
In this experiment, we compared expression of EDG1 and EDG5 in embryo-induced deciduas of pregnancy with oil-induced deciduomas of pseudopregnancy. This comparative model can be used effectively to identify target genes induced in the uterus by embryonic signaling factors. While EDG1 was only weakly expressed in antimesometrial decidualized stromal cells on DOP 7.5, expression of this receptor became elevated in the mesometrial microvasculature (Figs. 2b and 3a). Interestingly, in the absence of an embryo, EDG1 was not expressed at the mesometrial pole of the oil-induced deciduoma on Day 7.5 of pseudopregnancy (Fig. 3b). Parallel results were obtained for EDG5 (data not shown).
|
Colocalization of EDG1 and EDG5 with PTGS2 at the Maternal:Embryo/Fetal Interface
The spatiotemporal expression of EDG1 and EDG5 noted in the previous experiment was similar to that described for the prostanoid synthetic enzyme prostaglandin-endoperoxide synthase 2 (PTGS2, [24]). Based on findings that S1P induces expression of PTGS2 in other cell types [17, 18], we hypothesized that EDG1 and EDG5 colocalize with PTGS2 at the maternal:embryo/fetal interface. Using immunofluorescent confocal microscopy, EDG1 expression overlapped with PTGS2 in decidualizing stromal cells immediately surrounding the embryo on DOP 5, particularly on the mesometrial side of the embryo (Fig. 4a). In agreement with PTGS2 expression, EDG1 and EDG5 were richly expressed in the mesometrial microvasculature on DOP 7.5 with only weak staining being retained in surrounding stromal cells (Fig. 4b). Placental tissue (DOP 14.5) contained a number of small, highly vascularized islets that were positive for EDG1, EDG5, and PTGS2. By late gestation (DOP 17.5), all three proteins were readily detected in overlapping fashion in fetal bone (Fig. 4, bottom panels) and skeletal muscle (not shown).
|
Regulation of PTGS2 in Primary Predecidualized Stromal Cells by S1P
Prostacyclin is the most abundant, and perhaps most significant, prostanoid produced during decidualization [15]. In light of the positive correlation observed in expression of EDG1/EDG5 and PTGS2, we next cultured primary predecidualized stromal cells from DOP 3.5 to determine if S1P regulated PTGS2 expression, as was shown in other tissues [17, 18]. In response to 1 µM S1P, a dose typically used in cell culture (e.g., [18]), Ptgs2 mRNAs increased in cultured decidual cells at 0.5 h and reached peak levels by 1.5 h, while prostacyclin synthase (Pgis) gene expression remained unchanged by S1P treatment (Fig. 5a). PTGS2 protein levels increased in S1P-treated stromal cells by at least twofold compared with vehicle-treated cells (Fig. 5b). Increased expression of PTGS2 by S1P is likely mediated by EDG1, EDG3, and/or EDG5 because only these receptors, but not EDG6 and EDG8, were expressed in cultured predecidualized stromal cells (see Fig. 6c).
|
|
S1P-Activated Signal Transduction in Uterine Stromal Cells
The objective of this final experiment was to identify intracellular signal transduction pathways activated by S1P in uterine predecidualized stromal cells. Ligand-bound S1P receptors are known to activate a number of intracellular signaling pathways, including the MAPK and Akt pathways [reviewed in 25]. Furthermore, PTGS2 is regulated by members of the MAPK signaling family during decidualization in mice [26], as well as by AKT in other cell types [27, 28]. Provision of S1P to predecidual stromal cells resulted in the rapid and transient activation of MAPKs 1 and 3 (Fig. 6a). The MAPK1/3 phosphorylation remained significantly higher than control (time zero) levels from 2 to 5 min (P < 0.05). The remaining signaling pathways were not activated by S1P significantly (P > 0.05) above control values within the 60-min time course. Pharmacological inhibition of MAPK signaling using the MAPK kinase 1 (MAP2K1) inhibitor PD98059 attenuated (P < 0.05) S1P-induced Ptgs2 gene expression (Fig. 6b), suggesting that MAPK1/3 signaling is required for induction of Ptgs2 gene expression by S1P.
DISCUSSION
In the present study, we established that the sphingolipid pathway, as with many other previously characterized signal transduction pathways, regulates PTGS2 expression in uterine predecidualized stromal cells of early gestation. We also lay the foundation for understanding how other, previously uncharacterized, signaling factors potentially contribute to uterine angiogenesis during pregnancy, as we have characterized S1P receptor expression throughout the implantation phase of pregnancy. In other tissues, S1P receptors, when activated by S1P, function diversely to coordinate angiogenesis [29], lymphocyte egress [30], chemotaxis [31], proliferation [32], cell survival [32], immunomodulation [33], and, more recently, prostanoid synthesis [16–18]. The first S1P receptor cloned, called endothelial differentiation gene 1, or Edg1, was established as a highly inducible and abundant G-protein-coupled orphan receptor using subtraction hybridization during endothelial cell differentiation [34, 35]. Later, Lee et al. identified S1P as the ligand for EDG1 in which S1P caused cell-cell aggregation and enhanced expression of cadherins by a Rho-dependent mechanism [36]. Since then, four other S1P receptors have been cloned in humans and mice that bind S1P with high affinity [reviewed in 25]. Receptors EDG1, EDG3, and EDG5 are best known for directing angiogenesis and vascular maturation in a number of tissues [29, 37], and function redundantly or coordinately during embryonic angiogenesis. Indeed, double- (Edg1/Edg3) and triple- (Edg1/Edg3/Edg5) mutant mice exhibit similar, but increasingly exaggerated, vascular defects during embryogenesis compared with Edg1 single-mutant embryos [38], which are embryonic lethal around Day 13.5 of pregnancy due to incomplete vascular maturation [37]. EDG1, EDG3, and EDG5 are widely expressed in most tissues, whereas EDG6 or EDG8 expression is restricted essentially to lymphoid and neural tissues [39]. Further evidence that EDG1, 3, and 5 exhibit redundant functions come from studies of downstream signaling pathways, in which all three activated receptors initiate Gi/o-dependent events, and EDG5 and EDG3 are also linked to Gq/12 [25]. In addition to vascular defects, edg1-deficient embryos display abnormal limb development coupled with altered expression of the hypoxia inducible factor-1
and its response gene VEGF in developing limbs [40]. Consistent with the known functions of EDG1, EDG3, and EDG5, we found expression of these receptors, but not EDG6 and EDG8, to be developmentally upregulated in the uterus from DOP 4.5 to 7.5, a period of time in which extensive angiogenesis is observed in the uterus. Expression of EDG1 and EDG5 was most evident in the microvasculature and stromal cells immediately overlying trophoblast giant cells of the mesometrial pole, with only very weak staining in the surrounding decidual cells. It is suggested that EDG1, EDG3, and EDG5 play a role in decidual angiogenesis, placing the S1P:S1P receptor signaling pathway along side the VEGF and angiopoeitin pathways [10, 11] as a potential mediator of uterine angiogenesis. Although not the focus of the present study, it is possible that S1P- and VEGF-signaling pathways signal sequentially in the uterus, as has been shown in other model systems [40, 41]. Presently, it is not clear if S1P receptors are functionally required for uterine angiogenesis; although a related receptor, lysophosphatitic acid (LPA) 3, which binds to and becomes activated by its lipid ligand LPA, is essential for PTGS2 regulation in the sublumenal compartment of the early implantation site, as well as for blastocyst spacing along the uterus [42]. The embryonic lethality of Edg1-null female mice precludes the use of this knockout line for postnatal investigations. Edg3 and Edg5 single-mutant female mice evidently deliver normal litters [38, 43; unpublished results]; however, Edg3/Edg5 double-mutant female mice, when bred to wild-type males, are subfertile in that litter sizes are reduced by about two pups per litter [43]. Prostanoids play a pivotal role in establishing and maintaining pregnancy, particularly during decidualization, and serve immunomodulatory and angiogenic roles [44]. A number of cytokines have been shown to regulate production of prostanoids, including interleukin (IL)-1ß [45], IL-1 [46], leukemia inhibitory factor [47, 48], or members of the epidermal growth factor family [44], all of which increase protein levels of the prostaglandin rate-limiting enzyme PTGS2. Classical studies with pharmacological agents demonstrated that decidualization could be blocked with PTGS2 inhibitors or initiated in steroid-hormone-primed females mice upon provision of prostaglandins in vivo. More recently, mutant mice deficient in PTGS2 were shown to exhibit faulty decidualization [49] due in part to disrupted VEGF signaling and attenuated mesometrial angiogenesis [13], a phenotype evidently dependent on genetic background [49]. Other studies highlight the functional importance of prostacyclin as the major prostanoid required for implantation both in terms of abundance and functional significance. Prostacyclin signals intracellularly through the peroxisome proliferator-activated receptor-delta transcription factor [15].
In our EDG1 and EDG5 IHC experiments, we noted a similar decidual expression pattern of these receptors to that previously described for PTGS2. A literature search revealed that members of the sphingolipid pathway upregulates PTGS2 expression and prostanoid synthesis in other systems. For example, tumor necrosis factor-alpha-induced PTGS2 expression and prostaglandin E2 synthesis are dependent on activation of sphingosine kinase 1 [50], an enzyme that catalyzes conversion of sphingosine to S1P. Amniotic fluid harbors S1P that modulated PTGS2 expression in human amnion-derived WISH cells [17]. More recently, Pettus et al. demonstrated that S1P and ceramide-1-phosphate (C1P) elevate prostaglandin E2 production in coordinated, but distinct, pathways, in that S1P mediates the effects of cytokines on PTGS2 induction, while C1P is required for activation and translocation of cytoplasmic phospholipase A2 [51]. Moreover, ceramide stimulated prostaglandin production in human amnion and uterine decidual cells [52]. Collectively, these studies suggest that the sphingolipid pathway serves an intermediary role in cytokine-initiated prostanoid synthesis. The potential for S1P to regulate PTGS2 also exists at sites within the placenta and fetus, as EDG1 and EDG5 were both found to colocalize with PTGS2 in placental vascular islets and fetal connective tissues. To confirm that S1P regulates PTGS2 expression, cultured predecidualized stromal cells obtained from DOP 3.5 were treated with S1P. While we noted a substantial increase in Ptgs2 mRNA in response to S1P, a finding that was dependent on MAPK1/3 signaling, only a twofold increase in PTGS2 protein was observed. Disparity in the induction of Ptgs2 mRNA and PTGS2 protein by sphingolipids was similarly reported by Kirtikara et al. In their study, the S1P precursor ceramide induced Ptgs2 mRNA fivefold in fibroblasts, but did not significantly change PTGS2 protein levels, unlike IL-1ß, which resulted in a dramatic accumulation of PTGS2 protein [53]. It was concluded that ceramide played an accessory or enhancer role in cytokine-induced prostaglandin synthesis. While not tested in our study, we suspect a similar mechanism is at play in predecidualized stromal cells. S1P also increases production of prostacyclin, a downstream by-product of PTGS2 metabolism, in smooth muscle cells [18]. Because prostacyclin is the major prostanoid produced in the decidualizing uterus, we investigated mRNA expression of prostacyclin synthase, but found that this enzyme was not regulated by S1P at the transcriptional level in uterine stromal cells of pregnancy. As such, consistent with their low level of expression in the stromal cell compartment, S1P receptors may play only an ancillary role in regulating prostanoids in the stromal compartment; however, S1P may be important for coordinating prostanoid synthesis in PECAM1-positive cells or any of the other compartments in which abundant S1P receptor:PTGS2 colocalization was observed. Functional studies with S1P receptor-deficient mice should yield more mechanistically relevant data.
Should S1P receptors be functionally required for the establishment or maintenance of pregnancy, it is reasonable to speculate that S1P is generated at the implantation site, allowing for paracrine-mediated (i.e., local), rather than endocrine-mediated (i.e., systemic), signaling. Indeed, we have established that the entire sphingolipid-metabolizing pathway is upregulated in the deciduum in response to embryonic factors and that levels of sphingolipid metabolites, such as ceramide, increase nearly threefold from DOP 4.5 to DOP 7.5 (unpublished result). Others have shown that acid ceramidase, an enzyme that converts ceramide to sphingosine, is upregulated in the uterus at the implantation site [54]. As such, we propose a model whereby the embryo signals uterine angiogenesis through elevated local S1P synthesis and increased S1P-receptor expression. Production of S1P in the deciduum could also act locally on S1P receptors to coordinate vascular maturation or permeability [32, 37], cell adhesion [37], or to serve immunomodulatory roles such as lymphocyte egress [30] or regulation of cytokine secretion [33]. Finally, maternally derived S1P may signal in the trophoblast or embryo. It will be interesting to determine if S1P controls trophoblast outgrowth, either proliferation of trophoblast cells or their expression of adhesion molecules. Others have shown that the lysophospholipid lysophosphatidic acid, which signals through G-protein-coupled receptors homologous to S1P receptors, accelerates blastocyst differentiation through induction of calcium transients and heparin-binding epidermal growth factor autocrine signaling [55]. It will also be of interest to identify those embryonic factors that regulate S1P receptor expression in the uterus.
In summary, we have established a link between the sphingolipid and prostanoid signaling pathways in that S1P upregulated PTGS2 in cultured predecidualized uterine stromal cells of early pregnancy. S1P clearly activated MAPKs 1 and 3, which were shown to be required for S1P-induced PTGS2 expression. Further, S1P receptors EDG1 and EDG5 colocalized with PTGS2 in multiple compartments at the maternal:embryo/fetal interface throughout pregnancy. Using RT-PCR and immunohistochemistry, an initial expression profile was established for receptors that bind S1P with high affinity during implantation, suggesting roles in both prostanoid synthesis and uterine angiogenesis. Based on observations made from our gene expression and cell culture experiments in vitro, functional studies are being completed in our laboratory to test the hypothesis that S1P and cognate G-protein-coupled receptors are yet another signaling pathway that functions to regulate uterine angiogenesis during pregnancy.
FOOTNOTES
2 Correspondence: James K. Pru, Vincent Center for Reproductive Biology, Department of Obstetrics and Gynecology, Massachusetts General Hospital, Room 6613B, Building 149, 149 13th Street, Charlestown, MA 02129. FAX: 617 724 9935; jpru{at}partners.org 
1 Supported in part by Vincent Memorial Research Funds and NIH Grant R01-ES012070 to J.K.P.
Received: 16 August 2005.
First decision: 6 September 2005.
Accepted: 28 November 2005.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  X. Ye Lysophospholipid signaling in the function and pathology of the reproductive system Hum. Reprod. Update, June 25, 2008; (2008) dmn023v2. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. Kaneko-Tarui, L. Zhang, K. J Austin, L. E Henkes, J. Johnson, T. R Hansen, and J. K Pru Maternal and Embryonic Control of Uterine Sphingolipid-Metabolizing Enzymes During Murine Embryo Implantation Biol Reprod, October 1, 2007; 77(4): 658 - 665. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y.-J. Jeng, V. R. Suarez, M. G. Izban, H.-Q. Wang, and M. S. Soloff Progesterone-induced sphingosine kinase-1 expression in the rat uterus during pregnancy and signaling consequences Am J Physiol Endocrinol Metab, April 1, 2007; 292(4): E1110 - E1121. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
