Abstract
The use of non-steroidal anti-inflammatory drugs that target prostaglandin synthase (PTGS) enzymes has been implicated in miscarriage. Further, PTGS2-derived prostaglandins are reduced in the endometrium of patients with a history of implantation failure. However, in the mouse model of pregnancy, peri-implantation PTGS2 function is controversial. Some studies suggest that Ptgs2 −/− mice display deficits in ovulation, fertilization and implantation, while other studies suggest a role for PTGS2 only in ovulation but not implantation. Further, the uterine cell type responsible for PTGS2 function and the role of PTGS2 in regulating implantation chamber formation are not known. To address this, we generated tissue-specific deletion models of Ptgs2. We observed that PTGS2 ablation from the epithelium alone in Ltf cre/+ ; Ptgs2 f/f mice and in both the epithelium and endothelium of the Pax2 cre/+ ; Ptgs2 f/f mice does not affect embryo implantation. Further, deletion of PTGS2 in the ovary, oviduct and uterus using Pgr cre/+ ; Ptgs2 f/f does not disrupt pre-implantation events but instead interferes with post-implantation chamber formation, vascular remodeling and decidualization. While all embryos initiate chamber formation, more than half of the embryos fail to transition from blastocyst to epiblast stage, resulting in embryo death and resorbing decidual sites at mid-gestation. Thus, our results suggest no role for uterine epithelial PTGS2 in early pregnancy but instead highlight a role for uterine stromal PTGS2 in modulating post-implantation embryo and implantation chamber growth. Overall, our study provides clarity on the compartment-specific role of PTGS2 and provides a valuable model for further investigating the role of stromal PTGS2 in post-implantation embryo development.
Introduction
According to the American College of Obstetricians and Gynecologists, depending on maternal age, 9–40% pregnancies end in miscarriage, and only 10% of these losses are clinically recognized (American College of Obstetricians and Gynecologists' Committee on Practice Bulletins—Gynecology 2018, Dugas & Slane 2022). Additionally, 1–2% women experience recurrent pregnancy loss due to undetermined causes (Turesheva et al. 2023). Given the ethical considerations, human pregnancies cannot be studied directly. Thus, mice are often utilized as a model system to understand the early events of pregnancy. Recent advancements in 3D imaging methodology have been successfully applied to the pre-implantation stages of a mouse pregnancy, revealing phenomena that are challenging to uncover using traditional 2D histology. 3D imaging has revealed that embryo clusters enter the uterine environment at gestational day (GD) 3, ∼72 h after the mouse mating event. These embryos initially move together as clusters toward the middle of the uterine horn and then they undergo a bidirectional scattering movement followed by embryo spacing along the oviductal–cervical axis (Chen et al. 2013, Flores et al. 2020). At GD4, once the embryo arrives in the center of a flat peri-implantation region of the uterine lumen, a V-shaped embryo implantation chamber begins to form (Madhavan et al. 2022). This is concurrent with increased vascular permeability and sprouting angiogenesis at the embryo implantation sites at GD4 1800h (Madhavan et al. 2022, Massri et al. 2023). The proper formation of the embryo implantation chamber is critical as it facilitates embryo alignment along the mesometrial-anti-mesometrial axis, where the blastocyst’s inner cell mass faces the uterine mesometrial pole (Madhavan et al. 2022). Following embryo implantation, decidualization occurs, where stromal cells in the uterus become epithelialized, and embryos grow to the epiblast stage at GD5. Aberrations in events surrounding embryo implantation and decidualization can lead to a cascade of events that negatively impact subsequent pregnancy development, ultimately resulting in miscarriage and pregnancy loss (Cha et al. 2012).
Successful embryo implantation and maintenance of early pregnancy rely on a delicate interplay of numerous molecular mechanisms (Chen et al. 2013). Among these, prostaglandins (PGs), PGE2, PGI2 and PGF2 have emerged as critical mediators of reproductive success (Psychoyos et al. 1995, Wang & Dey 2006, Clark & Myatt 2008). PG synthesis begins with the phospholipase A2 enzyme cleaving arachidonic acid from the phospholipid bilayer. The prostaglandin synthase enzyme 1 (PTGS1) and PTGS2 convert arachidonic acid to PGH2. PGH2 is then converted to PGD2, PGE2, PGF2α, PGI2 and thromboxane (Funk 2001). Both PTGS1 and PTGS2 are glycosylated proteins with two catalytic sites: peroxidase and cyclooxygenase (thus the alternate names COX1 and COX2). These enzymes are similar at the amino acid level, but PTGS2 has an extra ‘side pocket’ that allows more space in the active site for substrate binding (Vecchio & Malkowski 2011). PTGS2 is often induced by cytokines, growth factors, hormones, inflammation and embryo attachment (Chakraborty et al. 1996, Ricciotti & FitzGerald 2011), while PTGS1 is constitutively expressed (Ricciotti & FitzGerald 2011).
Numerous studies have found evidence of PTGS1 and PTGS2 expression in human uterine compartments during implantation (Marions & Danielsson 1999). PTGS1 is expressed at a constant level in the human endometrium, while PTGS2 is expressed explicitly in the glandular epithelial cells and endothelial cells (Marions & Danielsson 1999) and the stromal cells (Stavreus-Evers et al. 2005). In addition, there is evidence for both PTGS1 and PTGS2 expression in the uteri of various species, including mice (Chakraborty et al. 1996), western spotted skunks (Das et al. 1999), baboons (Kim et al. 1999) and hamsters (Evans & Kennedy 1978, Wang et al. 2004b ). PTGS2 is expressed in the luminal epithelium and subepithelial stroma surrounding the blastocyst attachment site in the anti-mesometrial pole, and its expression is induced by the presence of the embryo (Chakraborty et al. 1996). Post-embryo implantation, PTGS1 is expressed in the secondary decidual zone; however, PTGS2 expression is localized at the mesometrial pole (Chakraborty et al. 1996).
Non-steroidal anti-inflammatory drugs (NSAIDs) that block PTGS1 and PTGS2 function are among the most common over-the-counter medications that women take during pregnancy (Thorpe et al. 2013). There is evidence for an 80% increased risk of miscarriage with the consumption of NSAIDs during pregnancy (Jackson-Northey & Evans 2002, Li et al. 2003, Li et al. 2018a ). PTGS1 has not been shown to have a role in pregnancy in women, and PTGS1-deficient mice do not display significant reproductive issues during pregnancy, except for prolonged parturition (Langenbach et al. 1995). On the other hand, studies in pregnant women who experience recurrent pregnancy loss or implantation failure after in vitro fertilization procedures demonstrate dysregulation in endometrial PTGS2 (Achache et al. 2010) and its derived prostaglandin PGI2 (Wang et al. 2010). Furthermore, genetic variations in the PTGS2 gene are associated with an increased risk of implantation failure among women going through assisted reproductive procedures (Salazar et al. 2010). In rodents, Lim et al. determined that PTGS2-deficient mice are infertile due to ovulation, fertilization and implantation deficits (Lim et al. 1997). While ovulation and fertilization defects are widely accepted, there is a controversy regarding the role of PTGS2 during embryo implantation (Lim et al. 1997, Cheng & Stewart 2003). Chang et al. reported that when wild-type blastocysts are transferred into PTGS2-deficient pseudopregnant uteri, a 24 h delay in decidualization is observed, but pregnancy proceeds to birth normally (Cheng & Stewart 2003). These data suggest that PTGS2 may not be essential for implantation, decidualization and overall pregnancy success. To explain the discrepancy between these studies, it has been proposed that a mixed mouse genetic background allows the upregulation of PTGS1 in PTGS2-deficient animals, and this PTGS1 may compensate for the loss of PTGS2 (Wang et al. 2004a ).
To resolve the controversy surrounding the function of PTGS2 in embryo implantation and to determine the compartment in which PTGS2 function is essential, we utilized the Cre-lox recombinase system (Kim et al. 2018). We deleted PTGS2 in the adult uterine epithelium using Ltf cre/+ (Daikoku et al. 2014), in the embryonic uterine epithelium and endothelium using Pax2 cre/+ (Ohyama & Groves 2004, Granger et al. 2024), and in the epithelial and stromal compartment of the uterus using Pgr cre/+ (Soyal et al. 2005, Madhavan & Arora 2022) (Table 1). We determined that PTGS2 function in the uterine epithelium and endothelium is not critical for implantation or pregnancy success. However, stromal PTGS2 is critical for post-implantation embryo and implantation chamber growth for continued pregnancy progression.
Mouse models used to study PTGS2 function in the uterus during implantation.
| Model | Tissue distribution | Deletion time |
|---|---|---|
| Ltf cre/+ ; Ptgs2 f/f | Uterine epithelium (LE, GE) | Adult (>8.5 weeks) |
| Pax2 cre/+ ; Ptgs2 f/f | Uterine epithelium (LE, GE) and endothelium | Embryonic (GD11.5) |
| Pgr cre/+ ; Ptgs2 f/f | Ovarian granulosa cells, oviductal epithelium, and muscle, uterine epithelium (LE, GE), circular smooth muscle and uterine stroma | Neonatal (P5–P21) |
LE, luminal epithelium; GE, glandular epithelium; GD, gestational day; P, postnatal day.
Materials and methods
Animals
We generated the Ptgs2 conditional deletion mice by breeding C57/bl6 Ptgs2 f/f (Ishikawa & Herschman 2006) with C57/bl6 Ltf cre/+ (Daikoku et al. 2014), mixed genetic background Pax2 cre/+ (Ohyama & Groves 2004) or C57/bl6 Pgr cre/+ (Soyal et al. 2005) mice (Table 1). For pregnancy studies, we set adult females at 6–10 weeks to mate with fertile males. For Ltf cre/+ ; Ptgs2 f/f , we mated them between 10 and 12 weeks, as PTGS2 deletion occurs in adult females (Daikoku et al. 2014). To create pseudopregnancy, we mated females with vasectomized males. The appearance of a vaginal plug was identified as GD GD0 1200h. We euthanized mice at several stages, including GD3 1200h and GD3 1800h, GD4 1800h, GD5.5, GD8.5 and GD12.5 or mice were allowed to go to term. We performed GD5.5, GD8.5 and GD12.5 dissections between 1300h and 1500h on the dissection day. To induce artificial decidualization, we used a non-surgical embryo transfer device, where we transferred 1 μL sesame oil and 3 μL PBS to a pseudopregnant mouse on either pseudopregnancy day 2 1800h or day 3 0800h. We euthanized the oil-stimulated pseudopregnant mice at day 3 of pseudopregnancy at 1200h or day 5.5 of pseudopregnancy. For GD4 and GD5, we euthanized the animals 10 min after a 0.15 mL intravenous injection of 1.5% Evans blue dye (MP Biomedicals, USA, ICN15110805). All mice were maintained on a 12 h light:12 h darkness cycle, and all mouse studies and protocols were approved by the Institutional Animal Care and Use Committee at Michigan State University.
Whole-mount immunofluorescence staining
As described previously (Arora et al. 2016, Flores et al. 2020, Madhavan et al. 2022), for whole-mount staining, we fixed dissected uteri in a mixture of cold DMSO:methanol (1:4). We hydrated the samples in a (1:1) methanol:PBST (PBS, 1% Triton) solution for 15 min, followed by a 15 min wash in PBST. We then placed the samples in a blocking solution (PBS, 1% Triton and 2% powdered milk) for 1 h at room temperature, followed by incubation with primary antibodies (Supplementary Table 1) in the blocking solution for seven nights at 4 °C. After washing twice with PBST solution for 15 min each and four additional times for 45 min each, we incubated the samples with Alexa Fluor-conjugated secondary antibodies for three nights or 72 h at 4 °C (Supplementary Table 1). Following the incubation, we washed the samples twice with PBST for 15 min each and four additional times for 45 min each and incubated the samples at 4 °C overnight with 3% H2O2 diluted in methanol. Finally, we washed the samples with 100% methanol three times, 30 min each, and cleared the tissues overnight with benzyl alcohol:benzyl benzoate (1:2) (Sigma-Aldrich, USA, 108006, B6630).
Cryo-embedding, cryo-sectioning and immunostaining
As described previously (Granger et al. 2024), we fixed uterine tissues in 4% PFA (paraformaldehyde) for 20 min and then incubated the samples with fresh 4% PFA overnight at 4 °C. The tissues were then washed three times with PBS for 5 min each and then incubated in 10% sucrose prepared in PBS at 4 °C overnight. We then transferred the samples to 20 and 30% sucrose solutions in PBS for 2–3 h each at 4 °C. Then we embedded the samples in Tissue-Tek OCT (Andwin Scientific, USA, 45831) and stored them at −80 °C. Cryo-sections of 7 µm thickness were mounted on glass slides (Fisher, USA, 1255015). For immunofluorescent staining, we allowed the slides to air dry for 15 min, then washed them three times with PBS for 5 min each and blocked with PBS +2% powdered milk +1% Triton solution for 20 min. After an additional three PBS washes, 5 min each, we stained the slides with primary antibodies (Supplementary Table 1) and incubated them at 4 °C overnight. The next day, we washed the slides with PBS for 3 × 5 min each, then incubated them with secondary antibodies and Hoechst (Supplementary Table 1) for 1 h at room temperature. Finally, after PBS washes, we added two drops of 20% glycerol in PBS to the slides followed by sealing the sections with glass coverslips.
In situ hybridization
We performed in situ hybridization on uterine sections using the RNAscope 2.5 HD Assay-RED kit (ACD Bio, USA, 322350), which also has immunofluorescence capabilities, as described previously (Granger et al. 2024). We aimed to detect Lif mRNA associated with the uterine glands at GD3 1800h. To detect Lif, we used the Mm-Lif probe (ACD Bio, 475841), and to label uterine glands, we included immunostaining for FOXA2 (Supplementary Table 1). The entire 3-day protocol was carried out according to the protocols provided by ACD Bio (322360-USM, MK 51-149 TN).
Serum progesterone measurement
After euthanizing the mouse, we collected 200–500 μL of blood samples and left them at room temperature for 30 min. Then, we centrifuged the samples for 15 min at 2,000 g, carefully separated the supernatant, and immediately stored the samples at −20 °C. Following sample collection and preservation, we sent the samples to the Ligand Assay and Analysis Core Laboratory in Charlottesville, VA, to determine progesterone levels. Samples were diluted at a ratio of 1:4, tested in triplicate to ensure accuracy, and the results were reported in ng/mL.
Oviduct flush and in vitro embryo culture
For the oviduct flush at GD1 1200h, we euthanized the female mice, excised both oviducts, and placed them in warm (37 °C) M2 medium (Sigma-Aldrich, M7167). We flushed each oviduct with approximately 300–500 μL of pre-warmed (37°) M2 medium using a blunted 30-gauge needle attached to a 1 mL syringe. We collected embryos and unfertilized eggs using a mouth pipette with a pulled glass capillary. After washing them 2–3 times in warm (37 °C) KSOM medium (Cytospring), we incubated them in 400–600 μL drop of KSOM media and placed them in a 37 °C jacketed incubator. We monitored embryonic development daily for 72 h and recorded the number of embryos reaching the 4-cell, 8-cell, morula and blastocyst stages (Frum & Ralston 2019).
RNA isolation, cDNA synthesis and quantitative PCR
We isolated uterine decidual tissues at GD5.5, snap-froze and stored the samples at −80 °C. We isolated total RNA from tissues using the TRIzol reagent (Invitrogen, USA, 15596019). Briefly, we homogenized the tissues in 1 mL TRIzol solution using the Bead Mill 4 homogenizer (Thermo Fisher Scientific, USA). Following phase separation with 500 μL chloroform, RNA was precipitated with isopropanol and washed with 75% ethanol. Then, we suspended the RNA in 50–100 μL RNase-free water (Invitrogen, AM9922). We measured the RNA concentration and purity using a NanoDrop 2000 spectrophotometer (Mettler Toledo, USA) with a concentration of at least 250 ng/μL. We performed first-strand cDNA synthesis from 1 μg RNA using a reverse transcriptase enzyme (Promega, USA, PRA5003). For qRT-PCR, we designed the primers using Primer3Plus and the NCBI website (Supplementary Table 2). We carried out the qRT-PCR reactions in triplicate for each sample using the QuantStudio 5 Real-Time PCR system (Applied Biosystems, USA) with a total reaction volume of 20 μL (10 μL SYBR Green (Thermo Fisher Scientific, A25742), 7.4 μL RNase- and DNase- free water, 1.6 μL primer and 1 μL cDNA). We used the comparative CT (ΔΔCt) method for gene expression analysis. We calculated the ΔCt for each sample by subtracting the Ct value of the Rpl19 gene from the Ct value of the target gene. We calculated the ΔΔCt by subtracting the mean ΔCt of the control group from the ΔCt of each sample. Fold change was calculated as 2(−ΔΔCt) (Livak & Schmittgen 2001).
Confocal microscopy
We used a Leica SP8 TCS white light laser confocal microscope utilizing 10× air to image whole uterine tissues or 20× water objective and a 7.0 μm Z-stack or system-optimized Z-stack to image the samples (Madhavan et al. 2022). Upon imaging, we imported the files (.LIF format) into Imaris v9.2.1 (https://imaris.oxinst.com/ (Bitplane; Oxford Instruments, UK) in 3D surpass mode. We created 3D renderings using surface modules.
Image analysis
Implantation chamber, luminal epithelium and embryo visualization
To visualize the implantation chamber, we used the CDH1 fluorescent signal for the luminal epithelium surface and the FOXA2 fluorescent signal for uterine glands. We isolated the luminal epithelium by subtracting the FOXA2-specific signal from the CDH1 signal. We used the Hoechst signal to locate embryos based on the inner cell mass signal, and we used the 3D rendering surface in IMARIS software to create the embryo surfaces. We used the measurement function in Imaris to measure the length of the implantation chamber.
Lif quantitation
As described (Granger et al. 2024), we used the FOXA2 signal to generate 3D surfaces of the glands' nuclei via the 3D surface function within the IMARIS software. Subsequently, we used the IMARIS masking function to produce a distinct channel for the Lif signal that lies beneath the previously established uterine gland 3D surface. Based on the new channel for the Lif signal, we created a new 3D surface of Lif. Following the creation of the 3D surfaces, we used the statistics function of Imaris to determine the 3D surface volume of both the glands and Lif. We used Microsoft Excel to calculate the Lif volume per uterine gland volume and plotted the results as Lif volume per uterine gland volume (FOXA2 signal) with normalized units.
Vessel density around the anti-mesometrial pole of the implantation chamber
We created a 3D rendering surface of blood vessels using a CD31 fluorescent signal and generated a channel in Imaris software to mask the surface of the blood vessels. For image segmentation, we imported 14 μm of the masked channel of vessels to ImageJ after adjusting the scale and applying the threshold function. Using vessel analysis and the Mexican Hat Filter Plugins in ImageJ (https://imagej.net/), we calculated the density and diameter of the blood vessels in the embryo implantation and inter-implantation site. For vessel density, the data are reported as the percentage of area occupied by blood vessels.
Statistical analysis
We used GraphPad Prism (https://www.graphpad.com/features (Dotmatics; GraphPad, USA) and Microsoft Excel to analyze the statistical differences between the treatment groups and plot our graphs. To analyze the difference between the two treatment groups, we employed the unpaired parametric two-tailed t-test. First, we tested the data for homogeneity of variance between the two treatments. If the variances were equal, we proceeded with a parametric two-tailed t-test. If the variances differed, we used the Mann–Whitney U test to compare the two treatment groups. We considered the data statistically different for P value <0.05 or less.
Results
Peri-implantation PTGS2 expression in embryo mediated and in oil-stimulated pregnancy
To determine which uterine cells might contribute to PTGS2 expression during peri-implantation stages, we performed expression analysis of PTGS2 in the uterine tract during peri-implantation stages utilizing natural and artificial pregnancy models. At GD3 1600h, when embryos are present in the uterus, PTGS2 is not expressed in the uterine luminal epithelium (Fig. 1A, A', B). mRNA expression of Ptgs2 has been reported in the luminal epithelium when a pseudopregnant uterus is stimulated with oil (Lim et al. 1997). We also observed PTGS2 protein expression in the uterine luminal epithelium 4 h after intraluminal oil stimulation of the pseudopregnant uterus at GD3 1200h (Fig. 1C, C', D). At GD4 1200h, when the embryo is at the center of the peri-implantation region, PTGS2 is expressed only in the luminal epithelium but not in the stroma (Madhavan et al. 2022). Following embryo implantation at GD4 1800h, PTGS2 is observed in the uterine subepithelial stroma surrounding the embryo implantation chamber (Fig. 1E, E', F), as reported previously (Chakraborty et al. 1996, Madhavan et al. 2022). At GD5.5, PTGS2 is expressed at the mesometrial pole surrounding the embryo implantation chamber, as reported previously (Chakraborty et al. 1996) and in uterine glands at the implantation chamber (Fig. 1G, G', H).
Timeline of uterine PTGS2 expression during peri-implantation stages and in an oil-stimulated pseudopregnancy. CDH1 and PTGS2 expression in pregnant wild-type uteri at GD3 1600h (A, A', B). PTGS2 expression in CDH1-positive cells in oil-stimulated uteri at day 3 of pseudopregnancy at 1200h, 4 h after oil stimulation (C, C', D). Three different regions from at least eight uterine horns were evaluated. PTGS2 expression in the subepithelial stroma surrounding the embryo implantation chamber at GD4 1800h (E., E', F). PTGS2 expression in the mesometrial pole and the uterine glands of the embryo implantation chambers at GD5.5 (G, G', H). At least two implantation sites from at least three uterine horns were analyzed. 7 µm XY slice (A, C, E, G). 105 μm XY slice (A', C', E', G'). 3D surface reconstruction (B, D, F, H). Scale bar: 200 μm. GD, gestational day; red arrowhead: embryo; white arrowheads: embryo implantation chamber.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
PTGS2 deletion in the uterine luminal epithelium and endothelium does not affect embryo implantation, embryo growth and pregnancy progression
To determine if the uterine epithelium is responsible for pre-implantation PTGS2 function, we generated tissue-specific deletion models of PTGS2 using Cre-lox recombinase methodology (Kim et al. 2018) (Supplementary Fig. 1A and B (see section on Supplementary materials given at the end of the article)). For adult uterine epithelial deletion, we used Ltf cre/+ ; Ptgs2 f/f mice (Daikoku et al. 2014) (Table 1), and for embryonic uterine epithelium and endothelial deletion, we used Pax2 cre/+ ; Ptgs2 f/f mice (Ohyama & Groves 2004) (Table 1). To confirm PTGS2 depletion in CDH1-positive uterine epithelial cells, we used oil-stimulated pseudopregnancies for both Ltf cre/+ ; Ptgs2 f/f and Pax2 cre/+ ; Ptgs2 f/f models (Fig. 2A, A', B, B', C, C'). At GD4 1800h, we observed the formation of the V-shaped embryo implantation chamber and stromal PTGS2 expression in control, Ltf cre/+ ; Ptgs2 f/f and Pax2 cre/+ ; Ptgs2 f/f mice (Fig. 2D, D', E, E', F, F'). At GD4 1800h, we observed no defects in the development of blastocyst in Pax2 cre/+ ; Ptgs2 f/f uteri (Fig. 2G and H and Table 3). Epithelial-specific and epithelial- and endothelial-specific PTGS2-deficient mutants displayed normal embryo spacing and increased vessel permeability at embryo implantation sites, as observed by the blue dye reaction at GD4 (Fig. 2I, J, K). At GD12.5, we observed that uteri from both mutants displayed embryos that had developed similarly to embryos from control uteri (Fig. 2J and L). Further, both Ltf cre/+ ; Ptgs2 f/f and Pax2 cre/+ ; Ptgs2 f/f mice were able to go to term with no significant effect on the duration of pregnancy or the number of pups born (Fig. 2K and L). Overall, our data suggest that the uterine epithelium and endothelium are not the sources of PTGS2-derived prostaglandin synthesis critical for implantation and pregnancy progression.
Conditional deletion of PTGS2 in the uterine epithelium and endothelium does not affect embryo implantation and pregnancy success. PTGS2 expression in CDH1-positive cells in oil-stimulated day 3 pseudopregnant Ptgs2 f/f uteri (A), Ltf cre/+ ; Ptgs2 f/f uteri (B) and Pax2 cre/+ ; Ptgs2 f/f uteri (C), 4 h after intraluminal oil stimulation. Three different regions from at least four uterine horns were evaluated. 7 µm XY slice (A, B, C). 105 μm XY slice (A', B', C'). PTGS2 expression in the subepithelial stroma in Ptgs2 f/f (D), Ltf cre/+ ; Ptgs2 f/f (E) and Pax2 cre/+ ; Ptgs2 f/f uteri (F) at GD4 1800h. 7 µm XY slice (D, E, F). 105 μm XY slice (D', E', F'). At least two implantation sites from three different uterine horns were analyzed. The top of the images represents the mesometrial pole, while the bottom represents the anti-mesometrial pole. Blastocyst-stage embryos in Ptgs2 f/f (G) and Pax2 cre/+ ; Ptgs2 f/f mice (H) at GD4 1800h. White dashed lines: blastocyst. Uteri with blue dye sites at GD4 1800h (I). Black asterisks: blue dye sites. Uteri with embryo sites at GD12.5 (J). Quantitation of blue dye sites at GD4 1800h, live embryos at GD12.5, and P0 pups in Ltf cre/+ ; Ptgs2 f/f mice (K) and in Pax2 cre/+ ; Ptgs2 f/f (L) with their respective controls. Each dot represents one mouse analyzed. Median values are shown. Data were analyzed using an unpaired parametric t-test. No significant differences were observed. Scale bars, A, B, C, A', B', C': 300 μm, D, E, F, D', E', F': 100 μm, G, H: 30 μm. LE, luminal epithelium.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
Stromal deletion of PTGS2 results in mid-gestation decidual resorption
To delete Ptgs2 in the granulosa cells of the preovulatory follicle and the corpus luteum, the epithelium and the myometrium of the oviduct (Soyal et al. 2005) and the circular smooth muscle, epithelium and stroma of the uterus (Soyal et al. 2005, Madhavan & Arora 2022) we utilized the progesterone-receptor-driven Cre (Pgr cre ) mouse line (Table 1, Fig. 3A, A', B, B' and Supplementary Fig. 2). We observed normal embryo spacing in Pgr cre/+ ; Ptgs2 f/f mice; however, embryo implantation was delayed, as observed using the blue dye reaction at GD4 1800h (Fig. 3C, D, G) (median blue dye sites in controls: 10, Pgr cre/+ ; Ptgs2 f/f : 7, P < 0.05). 24 h later at GD5.5, a similar number of decidual sites was observed in controls and Pgr cre/+ ; Ptgs2 f/f uteri (Fig. 3D and G). Complete loss of PTGS2 expression was observed in Pgr cre/+ ; Ptgs2 f/f implantation chambers at GD4 1800h (Fig. 3A and B) and at GD5.5 (Supplementary Fig. 2). To determine the cause for delayed implantation in mutant mice, we determined the mRNA expression of a critical glandular cytokine, leukemia inhibitory factor (Lif), at GD3 1800h. We observed reduced levels of Lif mRNA in FOXA2+ glandular epithelial uterine cells in Pgr cre/+ ; Ptgs2 f/f uteri (Supplementary Fig. 3A, B, C). However, we observed no differences in serum progesterone levels between control and mutant mice at GD3 and GD4 1800h (Supplementary Fig. 3D). Similar to GD5.5, at GD8.5, we observed no significant difference in the number of decidual sites between control and Pgr cre/+ ; Ptgs2 f/f uteri; however, we started to observe a few resorption sites in mutants (Fig. 3E and H). At GD12.5, the number of decidual sites was similar; however, we observed a significant number of resorbing decidua (50%) in mutant uteri (median live embryo number in control: 9, Pgr cre/+ ; Ptgs2 f/f : 4, P < 0.01) (Fig. 3F and H). Commensurate with the resorptions at mid-gestation, we observed a significant reduction in pups born to Pgr cre/+ ; Ptgs2 f/f females (48% pups loss) in comparison with controls (median live pup number in controls: 8, Pgr cre/+ ; Ptgs2 f/f : 4, P < 0.05) (Fig. 3I).
Pgr cre/+ ; Ptgs2 f/f mice display a delay in embryo implantation, mid-gestation decidual resorption, and pregnancy loss. PTGS2 expression in the subepithelial stroma surrounding the embryo implantation chamber in Ptgs2 f/f (A) and Pgr cre/+ ; Ptgs2 f/f (B) uteri at GD4 1800h. At least nine implantation sites were evaluated in at least two mice. 7 µm XY slice (A, B). 105 μm XY slice (A', B'). The top of the images represents the mesometrial pole, while the bottom represents the anti-mesometrial pole. Blue dye sites at GD4 1800h (C) and GD5.5 (D). Decidual sites at GD8.5 (E) and GD12.5 (F) in control and Pgr cre/+ ; Ptgs2 f/f uteri. Black asterisks: blue dye sites. Orange arrowheads: resorbed decidual sites. Quantification of blue dye sites at GD4 1800h and GD5.5 (G), decidual site numbers at GD8.5 and at GD12.5 (H) and live pups at P0 (I) in both groups. At least n = 3 mice were evaluated per genotype for each pregnancy stage. Each dot represents one mouse. Median values are shown. Data were analyzed using an unpaired parametric t-test. *P < 0.05, **P < 0.01. Scale bar for A, B, A', B': 100 μm.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
Abnormal embryo development in the post-implantation chamber of PTGS2-deficient uteri
To determine the first time point when embryo development is affected in Pgr cre/+ ; Ptgs2 f/f uteri, we examined embryo morphology at different time points during gestation. Ptgs2 −/− mice display a ∼30% reduction in the number of eggs ovulated per mouse and a complete failure of fertilization (Lim et al. 1997). Thus, we first examined the fraction of fertilized eggs in Pgr cre/+ ; Ptgs2 f/f mice. We performed an oviductal flush at GD1 1200h and cultured the embryos in vitro for 72 h. In control mice, we observed that 97.5% embryos were at the 2-cell stage at the time of oviductal flush. After 72 h of in vitro embryo culture, 18/39 (45%) embryos reached the morula stage and 21/39 (52.5%) reached the blastocyst stage (Table 2 and Fig. 4A, C, E, G, I, J). With Pgr cre/+ ; Ptgs2 f/f mice, we observed 12/56 (21.42%) unfertilized eggs, 4/56 (7.14%) 1-cell stage embryos and 40/56 (71.42%) 2-cell stage embryos at the time of oviductal flush. After 72 h of in vitro embryo culture, 8/56 (14.28%) embryos reached the morula stage, and 36/56 (64.28%) embryos reached the blastocyst stage. The 12/56 (21.42%) unfertilized eggs remained as such with no extrusion of polar body or cell division (Table 2 and Fig. 4B, D, F, H, I, J). Thus, Pgr cre/+ ; Ptgs2 f/f mice show a substantially improved fertilization rate compared to Ptgs2 −/− mice (Lim et al. 1997, Matsumoto et al. 2001). Overall, in our Pgr cre/+ ; Ptgs2 f/f model, we noted that once fertilization occurs, these embryos develop normally to the morula/blastocyst stage in vitro.
In vitro embryo culture of embryos flushed from Ptgs2 f/f and Pgr cre/+ ; Ptgs2 f/f uteri. Data are presented as n (%).
| Genotype | n | Embryo stage at oviductal flush (GD1 1200h) | Embryo stage after 72 h culture | ||||
|---|---|---|---|---|---|---|---|
| UF eggs | 1-cell embryo | 2-cell embryo | UF eggs | Morula | Blastocyst | ||
| Ptgs2 f/f | 5 | 1 (2.5) | 0 (0) | 39 (97.5) | 1 (2.5) | 18 (45) | 21 (52.5) |
| Pgr cre/+ ; Ptgs2 f/f | 8 | 12 (21.43) | 4 (7.14) | 40 (71.43) | 12 (21.43) | 8 (14.28) | 36 (64.28) |
Abbreviations – n, number of mice analyzed; GD, gestational day; h, hours; UF, unfertilized.
Stromal ablation of PTGS2 restricts embryo growth at post-implantation stages. Oviductal flush at GD1 1200h revealed 2-cell stage embryos in control (A) and 2-cell stage embryos and unfertilized eggs in Pgr cre/+ ; Ptgs2 f/f mice (B). 24, 48 and 72 h culture of flushed embryos/eggs in control (C, E, G) and Pgr cre/+ ; Ptgs2 f/f mice (D, F, H). Embryo development percentage at GD1 1200h (I) and at GD1 1200h + 72 h of culture (J). Blastocyst-stage embryos in control (K) and Pgr cre/+ ; Ptgs2 f/f mice (L) at GD3 1800h. Blastocyst-stage embryos in control (M) and blastocyst and abnormal embryos in Pgr cre/+ ; Ptgs2 f/f mice at GD4 1800h (N). Epiblast-stage embryos in control mice (O) and epiblast and abnormal embryos in Pgr cre/+ ; Ptgs2 f/f mice at GD5.5 (P). Red arrowheads: resorbing embryos. Comparison of embryo development percentage across GD1.5–GD5.5 (Q). Analysis was performed in uteri with embryos. At least n = 3 mice were analyzed per time point. Scale bars, A, B, C, D, E, F, G, H: 30 μm, K, L, M, N, O, P: 20 μm. Con, control; Mut, mutant Pgr cre/+ ; Ptgs2 f/f .
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
Next, we evaluated embryo development in our Pgr cre/+ ; Ptgs2 f/f model in vivo. We observed that for uteri with embryos at GD3 1800h, 95% embryos reached the blastocyst stage (Table 3 and Fig. 4K and L). However, at post-implantation stages at GD4 1800h, we observed that ∼62.5% embryos displayed embryo morphology that deviated from the typical elongated blastocyst (Table 3 and Fig. 4M and N). At GD5.5, we observed that 85% decidual sites had degrading embryos suggestive of pregnancy arrest (Table 3, Fig. 4O and P and Supplementary Fig. 4). Since some embryos displayed normal embryo development at GD5.5, we determined if there was compensatory upregulation of PTGS1 in the Pgr cre/+ ; Ptgs2 f/f uteri. PTGS1 was robustly expressed in the secondary decidual zone stroma around the embryo, in the glands at the anti-mesometrial pole, and in the glands and stroma in the inter-implantation region at GD5.5. No differences were observed between control and Pgr cre/+ ; Ptgs2 f/f deciduae (Supplementary Fig. 4). Our data suggest that embryonic growth restriction begins soon after implantation in Pgr cre/+ ; Ptgs2 f/f mice (Table 3 and Fig. 4Q).
Embryo development in Ptgs2 f/f and Pgr cre/+ ; Ptgs2 f/f mice.
| Stage/ Genotype | Mice, n | Uterine horns, n | Avg. embryo sites by blue dye | Embryo sites examined by morphology | Normal * n (%) | Abnormal † n (%) |
|---|---|---|---|---|---|---|
| GD3 1800h | ||||||
| Ptgs2 f/f | 5 | 5 | NA | 20 | 18 (90) | 2 (10) |
| Pgr cre/+ ; Ptgs2 f/f | 4 | 5 | NA | 19 | 18 (95) | 1 (5) |
| GD4 1800h | ||||||
| Ptgs2 f/f | 4 | 4 | 10.6 | 14 | 14 (100) | 0 (0) |
| Pgr cre/+ ; Ptgs2 f/f | 5 | 5 | 5.8 | 24 | 9 (37.5) | 15 (62.5) |
| Pax2 cre/+ ; Ptgs2 f/f | 4 | 5 | 5.85 | 32 | 29 (90.6) | 3 (9.4) |
| GD5.5 | ||||||
| Ptgs2 f/f | 3 | 3 | 9 | 9 | 8 (88.88) | 1 (11.11) |
| Pgr cre/+ ; Ptgs2 f/f | 5 | 5 | 9.6 | 20 | 3 (15) | 17 (85) |
Abbreviations – GD, gestational day; h, hours; NA, not applicable.
Normal: blastocyst or epiblast stage embryos with normal morphology.
Abnormal: embryos that deviate from the blastocyst morphology at GD3 1800h and GD4 1800h and resorbed epiblast observed at GD5.5.
Loss of stromal PTGS2 results in an abnormal implantation chamber, reduced implantation site vascular remodeling, and a poor decidualization response
Since we observed defects in the post-implantation embryo, we hypothesized that the implantation chamber and decidualization were the critical processes affected by the loss of PTGS2. We reconstructed the implantation chamber at GD4.5 and GD5.5 using 3D confocal imaging and image segmentation. At GD4 1800h, 13/14 embryos in control mice displayed a V-shaped chamber; however, in Pgr cre/+ ; Ptgs2 f/f mice, only 6/24 implantation chambers displayed a V-shape, while the remaining 18/24 embryos displayed either an asymmetric or an abnormal V-shaped chamber (Fig. 5A, B, C). At GD5.5, control mice displayed continued elongation of the V-shape chamber, while the chambers in the Pgr cre/+ ; Ptgs2 f/f uteri appeared shorter (Fig. 5D, E, F). The length of the implantation chamber in Pgr cre/+ ; Ptgs2 f/f mice was significantly lower than in control mice at both GD4 1800h (median chamber length in controls: 575.5 µm, Pgr cre/+ ; Ptgs2 f/f : 425.5 µm, P < 0.001) and GD5.5 (median chamber length in controls: 1007 μm, Pgr cre/+ ; Ptgs2 f/f : 589.5 µm, P < 0.0001) (Fig. 5G).
Abnormal embryo implantation chamber structure in Pgr cre/+ ; Ptgs2 f/f mice. At GD4 1800h, V-shaped implantation chambers (13/14) are observed in control mice (A) and 6/24 normal V-shaped implantation chambers (B) and 18/24 abnormally shaped implantation chambers (C) are observed in Pgr cre/+ ; Ptgs2 f/f mice. At GD5.5, elongated embryo implantation chambers (9/9) are observed in control mice (D) and 6/20 elongated (E) and 14/20 short implantation chambers (F) are observed in Pgr cre/+ ; Ptgs2 f/f mice. The top of the images represents the mesometrial pole, while the bottom represents the anti-mesometrial pole. Quantitation of implantation chambers in control and Pgr cre/+ ; Ptgs2 f/f mice at GD4 1800h and GD5.5 (G). At least n = 3 mice were evaluated per time point. Each dot represents one implantation chamber. Median values are shown. Data were analyzed using an unpaired parametric t-test. ***P < 0.001, ****P < 0.0001. Scale bar, A, B, C, D, E, F: 150 μm. Orange dashed lines: embryo implantation chamber; IC, implantation chamber.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
We also evaluated vascular development in the implantation and inter-implantation regions of the uterine horn at GD4 1800h. We observed a drastic decrease in vessel density surrounding the embryo implantation chamber in the mutant uteri compared to controls; however, the vessel density in the inter-implantation site remained comparable (Fig. 6A, B, C, D). Vessel diameter was similar in controls and mutants across both implantation and inter-implantation sites (Fig. 6A, B, D). CD31-positive cells accumulate around the implantation chamber (Govindasamy et al. 2021), and this expression overlaps with the PTGS2 expression domain. We observed that 8/8 implantation sites in control mice showed this CD31 signal around the implantation chamber (Fig. 6E, E', H), while only 4/10 implantation sites in the mutant showed a CD31 signal around the chamber (Fig. 6F, F', G, G', H).
Abnormal vascular development at the implantation site in Pgr cre/+ ; Ptgs2 f/f . CD31 expression in Ptgs2 f/f (A) and Pgr cre/+ ; Ptgs2 f/f (B) mice at GD4 1800h. Quantitation of vessel density (C) and vessel diameter (D) at embryo implantation sites and in inter-implantation sites (region between two implantation sites). CD31 expression around the embryo implantation chamber in Ptgs2 f/f (E, E') and Pgr cre/+ ; Ptgs2 f/f mice (F, F', G, G'). The top of the images represents the mesometrial pole, while the bottom represents the anti-mesometrial pole. Quantification of embryo implantation chambers with and without CD31 expression (H). n = 3 mice were evaluated per genotype. Each dot represents one implantation or inter-implantation site. Median values are shown. Data were analyzed using an unpaired parametric t-test. **P < 0.01. Scale bar, A, B: 200 μm, E, F, G, E', F', G': 100 μm. IS, implantation site; Inter-IS, inter-implantation site; IC, implantation chamber.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
Given the defects in the implantation chamber, we evaluated the expression of classic decidualization markers. Using qPCR, we observed a reduction in Bmp2 (P = 0.052) and Wnt4 (P < 0.05) transcripts at GD5.5 in Pgr cre/+ ; Ptgs2 f/f deciduae compared to controls (Fig. 7A). We also tested the decidual response of pseudopregnant control and mutant mice to an oil stimulus. We observed that, in comparison to the control uteri, intraluminal oil stimulation of the Pgr cre/+ ; Ptgs2 f/f uteri at pseudopregnancy day 2 failed to elicit a decidualization response at pseudopregnancy day 5.5 (Fig. 7B). Taken together, our data suggest that stromal PTGS2, or combined epithelial and stromal PTGS2, is crucial for post-implantation chamber growth, vessel remodeling surrounding the implantation chamber, and the initiation of decidualization, all of which are critical processes for successful pregnancy.
Decidualization failure in the stromal deletion model of PTGS2. Expression of decidualization markers was measured by qRT-PCR at GD5.5 (A). Artificial decidualization was induced by oil stimulation in pseudopregnant mice at pseudopregnancy day 2 1800h and analyzed at pseudopregnancy day 5.5 (B). At least three mice for each condition were analyzed. Each dot represents one mouse. Median values are shown. Data were analyzed using an unpaired parametric t-test. *P < 0.05. Scale bar, B: 1 cm. Black asterisks: decidual sites.
Citation: Reproduction 169, 4; 10.1530/REP-24-0408
Discussion
PTGS2-derived prostaglandins are functionally implicated in reproductive processes, including ovulation, fertilization, embryo implantation and decidualization (Kennedy 1977, Lim et al. 1997, Lim et al. 1999, Matsumoto et al. 2001). Despite these studies, there is still a debate in the literature regarding the role of PTGS2 in embryo implantation (Cheng & Stewart 2003). In this study, we used different tissue-specific ablation models of PTGS2 and showed that PTGS2 deletion in the uterine epithelium and endothelium does not impact pregnancy success. In contrast, deleting PTGS2 from both the epithelium and the stroma results in post-implantation embryonic growth restriction, defective implantation chamber growth and mid-gestation resorption. These results highlight a role for uterine stromal PTGS2 in post-implantation stages of embryo development and the initiation of decidualization but no critical role for PTGS2 in pre-implantation processes. During the drafting of this manuscript, (Aikawa et al. 2024) published their observations using Pgr cre/+ ; Ptgs2 f/f mice and their results are consistent with ours, suggesting a role for PTGS2 at the maternal–fetal interface. Given the debate on the role of PTGS2 function in murine pregnancy, consistent results with the tissue-specific deletion highlight a role for PTGS2 function independent of the mouse genetic background. The discussion below considers our study, as well as those by Aikawa et al. (2024).
Our data demonstrate differences between epithelial and combined epithelial and stromal deletion of PTGS2; however, it does not clearly distinguish between stromal cell-specific synthesis of prostaglandins and a requirement for a threshold level of prostaglandins at the embryo implantation site. Although there is a delay in implantation at GD4.5, embryo degradation is only observed at GD5.5 when PTGS2 is exclusively expressed in the stroma. This is highly suggestive of a critical role for stromal PTGS2 in post-implantation pregnancy progress. The phenotype in Pgr cre/+ ; Ptgs2 f/f mice also supports previous data highlighting unique contributions of stromal PTGS2. PTGS2-derived prostaglandins, particularly PGI2 and PGE2, are crucial regulators of vascular permeability and decidualization, with PGI2 working through the activation of PPARδ and IP (Lim et al. 1999). Chromatography/mass spectrometry analysis has revealed that PGI2 is the most abundant PG at embryo implantation sites, and PGI2 is primarily produced by stromal cells surrounding these implantation sites (Lim et al. 1999, Wang et al. 2007). In addition to PGI2, recent reports have highlighted other PGs, such as PGD2, which functions through its receptor located in the mesometrial region of stromal cells, and PGE2, which functions through its EP4 receptor located in the stromal anti-mesometrial region. The spatial and temporal regulation of these PTGS2-derived PGs remains high during the decidualization process, suggesting that these PGs play a critical role in the decidualization process (Sakamoto et al. 2024).
The delay in embryo implantation observed with combined epithelial and stromal PTGS2 deletion could be due to a greater reduction in local prostaglandin synthesis in the combined deletion compared to the epithelial-only deletion. This is supported by previous studies that demonstrate a need for threshold levels of prostaglandins for successful embryo implantation. For instance, pharmacological studies with the PTGS2-specific inhibitor Dup-697 demonstrate progressively severe effects on the implantation process with an increase in inhibitor dose (Lim et al. 1997). Kennedy et al. also report a direct relationship between prostaglandin concentration and vascular permeability (Kennedy 1979). Further, Wang et al. demonstrate that PTGS1 can rescue fertility in PTGS2-deficient mice in the CD1 mouse background, possibly suggesting that total prostaglandin levels must be maintained at a certain threshold to rescue fertility (Wang et al. 2004a ). Further studies are needed to distinguish between these possibilities of threshold levels of prostaglandins compared to stromal contributions of prostaglandin synthesis. The development of stromal cell-specific Cre lines, complemented with a quantitative assessment of compartment-specific PGs, would help address these questions.
Granulosa cell-specific deletion of PTGS2 does not produce ovulation and fertilization defects
PTGS2 is active in the ovaries during follicular development (Liu et al. 1997, Park et al. 2020), suggesting its importance during ovulatory processes. Clinical observations have reported luteinized unruptured follicle syndrome, characterized by the failure of follicle wall rupture despite a normal ovulatory cycle, in women who consume NSAIDs such as indomethacin or selective PTGS2 inhibitors (Micu et al. 2011). This condition results in infertility (Qublan et al. 2006). In rodents, indomethacin treatment during proestrus disrupts the follicle rupture process, resulting in ovulation failure (Gaytán et al. 2002). Furthermore, both in vitro and in vivo studies have demonstrated that PTGS2 inhibition through indomethacin and NS-398 treatment inhibited LH hormone induction of PGE2 production and thus decreased ovulation rates in rats (Mikuni et al. 1998). Ptgs2 −/− mice failed to produce PGs in response to gonadotropin stimulation and could not ovulate due to compromised cumulus expansion (Davis et al. 1999). This phenotype of failed ovulation occurs irrespective of the mouse genetic background. These diverse lines of studies underscore the indispensable role of PTGS2 in ovulation. PGR and PTGS2 are co-expressed in the mural granulosa cells of the preovulatory follicle following hCG stimulation (Zhang et al. 2023) and LH stimulation (Park et al. 2020). However, despite PTGS2 deletion in granulosa cells of the preovulatory follicle and the corpus luteum of the ovary (Soyal et al. 2005), Pgr cre/+ ; Ptgs2 f/f mice did not exhibit any ovulation failure. It is possible that Pgr cre may fail to delete Ptgs2 in all granulosa cells, resulting in residual PTGS2 expression and function during ovulation. Alternatively, serum PGs synthesized outside the ovary, oviduct and uterus may be responsible for the pro-inflammatory response resulting in ovulation. This will be a subject of future investigations.
Uterine epithelial PTGS2 does not contribute to embryo spacing and on-time embryo implantation
The endometrial epithelium has been recognized as a source of inducible PTGS2 and associated PGs, especially in the context of menstruation (Lundström et al. 1979). In addition, epithelial and endothelial PGs are thought to regulate smooth muscle contraction and relaxation (Félétou et al. 2011, Ruan et al. 2011). Inhibiting PG synthesis results in embryo crowding in pregnant rats (Kennedy 1977), and PGs are also critical for parturition (Aiken 1972, Reese et al. 2000), highlighting a possible link between epithelial PTGS2 and muscle contractility for embryo spacing and parturition. Our expression studies did not detect epithelial or endothelial PTGS2 during the preimplantation stage, although we did observe that PTGS2 is expressed in the luminal epithelium shortly after intraluminal stimulation with oil (Lim et al. 1997) and in the glands at the implantation chamber at GD5.5. Despite this, epithelial-only and epithelial and endothelial deletion of PTGS2 did not affect embryo spacing or on-time embryo implantation. Further, deletion of PTGS2 in the circular muscle in Pgr cre/+ ; Ptgs2 f/f did not affect embryo spacing, supporting that PTGS2 synthesized in the circular muscle, epithelium or endothelium is dispensable for uterine contractility critical for the initial phases of embryo movement.
Uterine stromal PTGS2 is critical for decidualization success
The previous literature suggests that implantation and decidualization failure in Ptgs2 −/− are not related to disruptions in ovarian steroid levels or genes related to implantation success, such as leukemia inhibitory factor (Lif) (Lim et al. 1997). Although progesterone levels were normal, we observed a significant reduction in Lif mRNA levels in our Pgr cre/+ ; Ptgs2 f/f model. Reduced levels of Lif can explain the delay in implantation and may also contribute to the absence of a decidualization response with an oil stimulus in this mutant. Delayed implantation may also explain the deviation of the embryo’s morphology compared to an elongated blastocyst at GD4. However, the absence of stromal PTGS2 at the anti-mesometrial pole of the implantation chamber in the Pgr cre/+ ; Ptgs2 f/f model is the most likely cause of poor elongation of the implantation chamber and degradation of the embryos at GD5.5. A defective chamber likely results in a ripple effect of decreased vascular remodeling in the decidua surrounding the implantation chamber and a reduction in the amount of decidualized stroma, leading to growth arrest in the embryo and failure of pregnancy progression. Our results suggest that elongation of the implantation chamber is critical for the transition of the embryo from an elongated blastocyst to an epiblast stage, highlighting a critical role for stromal PTGS2 in embryo-uterine communication at this stage of pregnancy.
Our results also highlight that once chamber formation begins and decidualization is initiated, the embryo is no longer needed for continuous expansion of the decidua. Even though 85% embryos displayed severe growth retardation at GD5.5, decidual expansion continued beyond GD8.5, and resorptions were observed at a significant level at GD12.5 when extraembryonic tissue contributions are required for the formation of the placenta. These data are in line with other models of decidualization where oil and beads (Herington et al. 2009, Chen et al. 2011) can stimulate the initiation of decidualization, and the decidua continues to expand in the absence of embryonic contributions until mid-gestation. It has been proposed that decidualization with a bead or oil is different from embryo-induced decidualization (Herington et al. 2009). The Pgr cre/+ ; Ptgs2 f/f mouse may be a good model to compare the growth of the decidua with and without a growing epiblast to explore the similarities and differences between the two decidualization processes.
Our data also highlight that even with complete ablation of stromal PTGS2, ∼50% embryos in the Pgr cre/+ ; Ptgs2 f/f uteri continue to develop beyond mid-gestation and are also born. PTGS2 may permit implantation chamber growth beyond a certain length. If the chamber is stochastically able to grow beyond this length (due to PTGS1 upregulation or other factors such as the expanding decidua), then PTGS2 in the stroma may no longer be required. It is also possible that the embryos that display a delay in implantation are susceptible to the absence of stromal PTGS2 during the elongation of the chamber. However, these different hypotheses need to be tested to determine why some embryos continue to grow despite the absence of stromal PTGS2.
Overlapping roles for PTGS1 and PTGS2 in murine implantation success and the role of mouse genetic background
Ptgs1 −/− mice on a 129/B6 mouse background have 32% lower vascular permeability and significantly lower PG levels (specifically 6-keto-PGF1α and PGE2). These mice also display an upregulation of PTGS2 expression during the pre-implantation stage (Reese et al. 1999). This indicates that PTGS2 can compensate for the function of PTGS1 (Reese et al. 2000). When Ptgs2 is inserted into the Ptgs1 locus, PTGS2 can compensate for PTGS1 loss and rescue the parturition defect observed in Ptgs1 −/− mice (Li et al. 2018b ). However, on a C57Bl6 mouse background, when Ptgs1 was placed in the Ptgs2 locus, PTGS1 failed to compensate for PTGS2 function, resulting in mice with implantation phenotypes similar to the Ptgs2 −/− mice (Lim et al. 1997, Li et al. 2018b ). It has been previously reported that on a mixed mouse genetic background, PTGS1 is upregulated in the Ptgs2 −/− mice, and these mice exhibit improved fertility compared to Ptgs2 −/− mice on a pure C57Bl6 mouse background (Wang et al. 2004a ). In our studies (on a C57Bl6 background), we did not observe a post-implantation compensatory increase in PTGS1 expression. Further, in our studies and those by Aikawa et al. (2024) (mouse background not specified), the Pgr cre/+ ; Ptgs2 f/f mice show ∼50% reduction in the number of pups at birth. Consistency among our studies suggests a post-implantation role for PTGS2 independent of mouse genetic background. Aikawa et al. also showed that depletion of both PTGS1 and PTGS2 (Ptgs1 −/− ; Pgr cre/+ ; Ptgs2 f/f mice, mouse background unknown) results in a complete failure of embryo implantation, with embryos floating in the uterus (Aikawa et al. 2024). Since embryos were presumably normal in these mice, the cause for a complete absence of implantation could be a lack of Lif; however, this needs to be tested. All of these studies highlight the interconnected roles of PTGS enzymes and for processes such as implantation and decidualization.
Conclusion
Our study highlights that PTGS2-derived PGs necessary for implantation do not come from uterine epithelial and endothelial sources. Our work provides definitive proof that stromal PTGS2 at the base of the embryo implantation chamber is critical for both the growth of the embryo and the elongation of the implantation chamber. Further work is needed to understand how stromal PTGS2 depletion affects the decidualization response and vascular remodeling and why a certain percentage of embryos can escape this requirement and go through gestation. Overall, this study distinguishes between the pre-implantation and post-implantation roles of PTGS2 and provides a valuable model for investigating the role of stromal PTGS2 without the need for embryo transfer to study the initiation of the decidualization process and how it relates to pregnancy success.
Supplementary materials
This is linked to the online version of the paper at https://doi.org/10.1530/REP-24-0408.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work.
Funding
This research was supported in part by the March of Dimes grant #5-FY20-209 and NIH R01HD109152 to RA, the Eunice Kennedy Shriver National Institute of Child Health & Human Development of the National Institutes of Health under award #T32HD087166 to NM, and award# R24 HD102061 to the University of Virginia Center for Research in Reproduction Ligand Assay and Analysis Core.
Author contribution statement
NM and RA conceptualized the study and designed the experiments. NM executed experiments. NM and RA validated the data and performed the analyses. NM and RA wrote and edited the manuscript. All authors reviewed and accepted the final version of the manuscript.
Acknowledgements
We thank Dr Harvey R. Herschman and Dr Srinivasa Reddy at UCLA for providing Ptgs2-floxed mice. We thank Curtis Chen for assistance with cryosectioning and Marcelio Shammami for assistance with the in vitro embryo culture protocol. We thank Dr. Gregory Burns, Dr. Nataki Douglas, Dr. Shuo Xiao and Dr. Asgerally Fazleabas for critical discussions related to the project.
References
Achache H , Tsafrir A , Prus D , et al. 2010 Defective endometrial prostaglandin synthesis identified in patients with repeated implantation failure undergoing in vitro fertilization. Fertil Steril 94 1271–1278. (https://doi.org/10.1016/j.fertnstert.2009.07.1668)
Aikawa S , Matsuo M , Akaeda S , et al. 2024 Spatiotemporally distinct roles of cyclooxygenase-1 and cyclooxygenase-2 at fetomaternal interface in mice. JCI Insight 9 e181865. (https://doi.org/10.1172/jci.insight.181865)
Aiken JW 1972 Aspirin and indomethacin prolong parturition in rats: evidence that prostaglandins contribute to expulsion of fetus. Nature 240 21–25. (https://doi.org/10.1038/240021a0)
American College of Obstetricians and Gynecologists' Committee on Practice Bulletins—Gynecology 2018 ACOG practice bulletin No. 200: early pregnancy loss. Obstet Gynecol 132 e197–e207. (https://doi.org/10.1097/AOG.0000000000002899)
Arora R , Fries A , Oelerich K , et al. 2016 Insights from imaging the implanting embryo and the uterine environment in three dimensions. Development 143 4749–4754. (https://doi.org/10.1242/dev.144386)
Cha J , Sun X & Dey SK 2012 Mechanisms of implantation: strategies for successful pregnancy. Nat Med 18 1754–1767. (https://doi.org/10.1038/nm.3012)
Chakraborty I , Das SK , Wang J , et al. 1996 Developmental expression of the cyclo-oxygenase-1 and cyclo-oxygenase-2 genes in the peri-implantation mouse uterus and their differential regulation by the blastocyst and ovarian steroids. J Mol Endocrinol 16 107–122. (https://doi.org/10.1677/jme.0.0160107)
Chen Q , Zhang Y , Peng H , et al. 2011 Transient {beta}2-adrenoceptor activation confers pregnancy loss by disrupting embryo spacing at implantation. J Biol Chem 286 4349–4356. (https://doi.org/10.1074/jbc.m110.197202)
Chen Q , Zhang Y , Elad D , et al. 2013 Navigating the site for embryo implantation: biomechanical and molecular regulation of intrauterine embryo distribution. Mol Aspects Med 34 1024–1042. (https://doi.org/10.1016/j.mam.2012.07.017)
Cheng JG & Stewart CL 2003 Loss of cyclooxygenase-2 retards decidual growth but does not inhibit embryo implantation or development to term. Biol Reprod 68 401–404. (https://doi.org/10.1095/biolreprod.102.009589)
Clark KE & Myatt L 2008 Prostaglandins and the reproductive cycle. In The Global Library of Women’s Medicine. Eds KG Danielsson & GC Di Renzo . The International Federation of Gynecology and Obstetrics (FIGO). (https://doi.org/10.3843/GLOWM.10314)
Daikoku T , Ogawa Y , Terakawa J , et al. 2014 Lactoferrin-iCre: a new mouse line to study uterine epithelial gene function. Endocrinology 155 2718–2724. (https://doi.org/10.1210/en.2014-1265)
Das SK , Wang J , Dey SK , et al. 1999 Spatiotemporal expression of cyclooxygenase 1 and cyclooxygenase 2 during delayed implantation and the periimplantation period in the Western spotted skunk. Biol Reprod 60 893–899. (https://doi.org/10.1095/biolreprod60.4.893)
Davis BJ , Lennard DE , Lee CA , et al. 1999 Anovulation in cyclooxygenase-2-deficient mice is restored by prostaglandin E2 and interleukin-1beta. Endocrinology 140 2685–2695. (https://doi.org/10.1210/endo.140.6.6715)
Dugas C & Slane VH 2022 Miscarriage (Archived). In StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing.
Evans CA & Kennedy TG 1978 The importance of prostaglandin synthesis for the initiation of blastocyst implantation in the hamster. J Reprod Fertil 54 255–261. (https://doi.org/10.1530/jrf.0.0540255)
Félétou M , Huang Y & Vanhoutte PM 2011 Endothelium-mediated control of vascular tone: COX-1 and COX-2 products. Br J Pharmacol 164 894–912. (https://doi.org/10.1111/j.1476-5381.2011.01276.x)
Flores D , Madhavan M , Wright S , et al. 2020 Mechanical and signaling mechanisms that guide pre-implantation embryo movement. Development 147 dev193490. (https://doi.org/10.1242/dev.193490)
Frum T & Ralston A 2019 Visualizing HIPPO signaling components in mouse early embryonic development. Methods Mol Biol 1893 335–352. (https://doi.org/10.1007/978-1-4939-8910-2_25)
Funk CD 2001 Prostaglandins and leukotrienes: advances in eicosanoid biology. Science 294 1871–1875. (https://doi.org/10.1126/science.294.5548.1871)
Gaytán F , Tarradas E , Morales C , et al. 2002 Morphological evidence for uncontrolled proteolytic activity during the ovulatory process in indomethacin-treated rats. Reproduction 123 639–649. (https://doi.org/10.1530/reprod/123.5.639)
Govindasamy N , Long H , Jeong HW , et al. 2021 3D biomimetic platform reveals the first interactions of the embryo and the maternal blood vessels. Dev Cell 56 3276–3287.e8. (https://doi.org/10.1016/j.devcel.2021.10.014)
Granger K , Fitch S , Shen M , et al. 2024 Murine uterine gland branching is necessary for gland function in implantation. Mol Hum Reprod. (https://doi.org/10.1093/molehr/gaae020)
Herington JL , Underwood T , McConaha M , et al. 2009 Paracrine signals from the mouse conceptus are not required for the normal progression of decidualization. Endocrinology 150 4404–4413. (https://doi.org/10.1210/en.2009-0036)
Ishikawa TO & Herschman HR 2006 Conditional knockout mouse for tissue-specific disruption of the cyclooxygenase-2 (Cox-2) gene. Genesis 44 143–149. (https://doi.org/10.1002/gene.20192)
Jackson-Northey K & Evans MF 2002 Taking NSAIDs during pregnancy. Is it safe? Can Fam Physician 48 483–485.
Kennedy TG 1977 Evidence for a role for prosaglandins in the initiation of blastocyst implantation in the rat. Biol Reprod 16 286–291. (https://doi.org/10.1095/biolreprod16.3.286)
Kennedy TG 1979 Prostaglandins and increased endometrial vascular permeability resulting from the application of artificial stimulus to the uterus of the rat sensitized for the decidual cell reaction. Biol Reprod 20 560–566. (https://doi.org/10.1095/biolreprod20.3.560)
Kim JJ , Wang J , Bambra C , et al. 1999 Expression of cyclooxygenase-1 and -2 in the baboon endometrium during the menstrual cycle and pregnancy. Endocrinology 140 2672–2678. (https://doi.org/10.1210/endo.140.6.6716)
Kim H , Kim M , Im SK , et al. 2018 Mouse Cre-LoxP system: general principles to determine tissue-specific roles of target genes. Lab Anim Res 34 147–159. (https://doi.org/10.5625/lar.2018.34.4.147)
Langenbach R , Morham SG , Tiano HF , et al. 1995 Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration. Cell 83 483–492. (https://doi.org/10.1016/0092-8674(95)90126-4)
Li DK , Liu L & Odouli R 2003 Exposure to non-steroidal anti-inflammatory drugs during pregnancy and risk of miscarriage: population based cohort study. Bmj 327 368–370. (https://doi.org/10.1136/bmj.327.7411.368)
Li DK , Ferber JR , Odouli R , et al. 2018a Use of nonsteroidal antiinflammatory drugs during pregnancy and the risk of miscarriage. Am J Obstet Gynecol 219 275.e1–275.e8. (https://doi.org/10.1016/j.ajog.2018.06.002)
Li X , Ballantyne LL , Crawford MC , et al. 2018b Isoform-specific compensation of cyclooxygenase (Ptgs) genes during implantation and late-stage pregnancy. Sci Rep 8 12097. (https://doi.org/10.1038/s41598-018-30636-x)
Lim H , Paria BC , Das SK , et al. 1997 Multiple female reproductive failures in cyclooxygenase 2-deficient mice. Cell 91 197–208. (https://doi.org/10.1016/s0092-8674(00)80402-x)
Lim H , Gupta RA , Ma WG , et al. 1999 Cyclo-oxygenase-2-derived prostacyclin mediates embryo implantation in the mouse via PPARdelta. Genes Dev 13 1561–1574. (https://doi.org/10.1101/gad.13.12.1561)
Liu J , Carrière PD , Doré M , et al. 1997 Prostaglandin G/H synthase-2 is expressed in bovine preovulatory follicles after the endogenous surge of luteinizing hormone. Biol Reprod 57 1524–1531. (https://doi.org/10.1095/biolreprod57.6.1524)
Livak KJ & Schmittgen TD 2001 Analysis of relative gene expression data using real-time quantitative PCR and the 2−ΔΔCT method. Methods 25 402–408. (https://doi.org/10.1006/meth.2001.1262)
Lundström V , Gréen K & Svanborg K 1979 Endogenous prostaglandins in dysmenorrhea and the effect of prostaglandin synthetase inhibitors (PGSI) on uterine contractility. Acta Obstet Gynecol Scand Suppl 58 51–56. (https://doi.org/10.3109/00016347909157790)
Madhavan M & Arora R 2022 Tracing early postnatal lineage of progesterone receptor in the mouse uterus. Mol Reprod Dev 89 69. (https://doi.org/10.1002/mrd.23557)
Madhavan MK , DeMayo FJ , Lydon JP , et al. 2022 Aberrant uterine folding in mice disrupts implantation chamber formation and alignment of embryo-uterine axes. Development 149 dev200300. (https://doi.org/10.1242/dev.200300)
Marions L & Danielsson KG 1999 Expression of cyclo-oxygenase in human endometrium during the implantation period. Mol Hum Reprod 5 961–965. (https://doi.org/10.1093/molehr/5.10.961)
Massri N , Loia R , Sones JL , et al. 2023 Vascular changes in the cycling and early pregnant uterus. JCI Insight 8 e163422. (https://doi.org/10.1172/jci.insight.163422)
Matsumoto H , Ma W , Smalley W , et al. 2001 Diversification of cyclooxygenase-2-derived prostaglandins in ovulation and implantation. Biol Reprod 64 1557–1565. (https://doi.org/10.1095/biolreprod64.5.1557)
Micu MC , Micu R & Ostensen M 2011 Luteinized unruptured follicle syndrome increased by inactive disease and selective cyclooxygenase 2 inhibitors in women with inflammatory arthropathies. Arthritis Care Res Hob 63 1334–1338. (https://doi.org/10.1002/acr.20510)
Mikuni M , Pall M , Peterson CM , et al. 1998 The selective prostaglandin endoperoxide synthase-2 inhibitor, NS-398, reduces prostaglandin production and ovulation in vivo and in vitro in the rat. Biol Reprod 59 1077–1083. (https://doi.org/10.1095/biolreprod59.5.1077)
Ohyama T & Groves AK 2004 Generation of Pax2-Cre mice by modification of a Pax2 bacterial artificial chromosome. Genesis 38 195–199. (https://doi.org/10.1002/gene.20017)
Park CJ , Lin PC , Zhou S , et al. 2020 Progesterone receptor serves the ovary as a trigger of ovulation and a terminator of inflammation. Cell Rep 31 107496. (https://doi.org/10.1016/j.celrep.2020.03.060)
Psychoyos A , Nikas G & Gravanis A 1995 The role of prostaglandins in blastocyst implantation. Hum Reprod 10 (Supplement 2) 30–42. (https://doi.org/10.1093/humrep/10.suppl_2.30)
Qublan H , Amarin Z , Nawasreh M , et al. 2006 Luteinized unruptured follicle syndrome: incidence and recurrence rate in infertile women with unexplained infertility undergoing intrauterine insemination. Hum Reprod 21 2110–2113. (https://doi.org/10.1093/humrep/del113)
Reese J , Brown N , Paria BC , et al. 1999 COX-2 compensation in the uterus of COX-1 deficient mice during the pre-implantation period. Mol Cell Endocrinol 150 23–31. (https://doi.org/10.1016/s0303-7207(99)00033-7)
Reese J , Paria BC , Brown N , et al. 2000 Coordinated regulation of fetal and maternal prostaglandins directs successful birth and postnatal adaptation in the mouse. Proc Natl Acad Sci U S A 97 9759–9764. (https://doi.org/10.1073/pnas.97.17.9759)
Ricciotti E & FitzGerald GA 2011 Prostaglandins and inflammation. Arterioscler Thromb Vasc Biol 31 986–1000. (https://doi.org/10.1161/atvbaha.110.207449)
Ruan YC , Zhou W & Chan HC 2011 Regulation of smooth muscle contraction by the epithelium: role of prostaglandins. Physiology 26 156–170. (https://doi.org/10.1152/physiol.00036.2010)
Sakamoto R , Fujiwara T , Kawano Y , et al. 2024 Uterine prostaglandin DP receptor-induced upon implantation contributes to decidualization together with EP4 receptor. J Lipid Res 65 100636. (https://doi.org/10.1016/j.jlr.2024.100636)
Salazar LA , Inostroza M , Jara C , et al. 2010 Association of -765G>C polymorphism of the COX-2 gene with recurrent embryo implantation failure in Southern Chilean women. Clin Chim Acta 411 1822–1824. (https://doi.org/10.1016/j.cca.2010.06.025)
Soyal SM , Mukherjee A , Lee KY , et al. 2005 Cre-mediated recombination in cell lineages that express the progesterone receptor. Genesis 41 58–66. (https://doi.org/10.1002/gene.20098)
Stavreus-Evers A , Koraen L , Scott JE , et al. 2005 Distribution of cyclooxygenase-1, cyclooxygenase-2, and cytosolic phospholipase A2 in the luteal phase human endometrium and ovary. Fertil Steril 83 156–162. (https://doi.org/10.1016/j.fertnstert.2004.06.057)
Thorpe PG , Gilboa SM , Hernandez-Diaz S , et al. 2013 Medications in the first trimester of pregnancy: most common exposures and critical gaps in understanding fetal risk. Pharmacoepidemiol Drug Saf 22 1013–1018. (https://doi.org/10.1002/pds.3495)
Turesheva A , Aimagambetova G , Ukybassova T , et al. 2023 Recurrent pregnancy loss etiology, risk factors, diagnosis, and management. Fresh look into a full box. J Clin Med 12 4074. (https://doi.org/10.3390/jcm12124074)
Vecchio AJ & Malkowski MG 2011 The structural basis of endocannabinoid oxygenation by cyclooxygenase-2. J Biol Chem 286 20736–20745. (https://doi.org/10.1074/jbc.m111.230367)
Wang H & Dey SK 2006 Roadmap to embryo implantation: clues from mouse models. Nat Rev Genet 7 185–199. (https://doi.org/10.1038/nrg1808)
Wang H , Ma WG , Tejada L , et al. 2004a Rescue of female infertility from the loss of cyclooxygenase-2 by compensatory up-regulation of cyclooxygenase-1 is a function of genetic makeup. J Biol Chem 279 10649–10658. (https://doi.org/10.1074/jbc.m312203200)
Wang X , Su Y , Deb K , et al. 2004b Prostaglandin E2 is a product of induced prostaglandin-endoperoxide synthase 2 and microsomal-type prostaglandin E synthase at the implantation site of the hamster. J Biol Chem 279 30579–30587. (https://doi.org/10.1074/jbc.m400573200)
Wang H , Xie H , Sun X , et al. 2007 Stage-specific integration of maternal and embryonic peroxisome proliferator-activated receptor delta signaling is critical to pregnancy success. J Biol Chem 282 37770–37782. (https://doi.org/10.1074/jbc.M706577200)
Wang Y , Zhao AM & Lin QD 2010 Role of cyclooxygenase-2 signaling pathway dysfunction in unexplained recurrent spontaneous abortion. Chin Med J Engl 123 1543–1547.
Zhang J , Goods BA , Pattarawat P , et al. 2023 An ex vivo ovulation system enables the discovery of novel ovulatory pathways and nonhormonal contraceptive candidates. Biol Reprod 108 629–644. (https://doi.org/10.1093/biolre/ioad009)
