Tissue Collection and Preservation

All experimental procedures were carried out in accordance with animal welfare legislation and were approved by respective authorities of the Justus-Liebig University in Giessen, Germany, and University of Ankara, Turkey, where the tissue materials had been collected and used in our previous studies [7, 14, 17].

Tissue material was collected surgically by ovariohysterectomy (OHE), from 18 crossbred, healthy bitches, 2–8 yr of age, which were included in the study and allotted to 4 experimental groups according to stage of pregnancy: preimplantation (Days 8–12, n = 5), postimplantation (Days 18–25, n = 5), mid-gestation (Days 35–40, n = 5), and prepartum luteolysis (n = 3). The mating day was designated as Day 0 of pregnancy for all bitches and was 2–3 days after ovulation, which was assessed by measurements of P4 (>5 ng/ml in peripheral blood) and by evaluation of vaginal smears. Preimplantation status was verified by detecting the presence of embryos in uterine flushings, with a stereomicroscope. According to descriptions by Amoroso [18] and Kehrer [19], the postimplantation and mid-gestation stages of canine pregnancy are characterized by the formation or presence, respectively, of fully developed uteroplacental compartments. The day of prepartum luteolysis was detected by regular measurements of peripheral P4 concentration at 6-h intervals beginning with Day 58 of gestation; OHE was performed after maternal P4 concentrations dropped below 3 ng/ml in 2 consecutive measurements.

For immunohistochemistry (IHC) and nonradioactive in situ hybridization (ISH), uterine samples (at preimplantation stage of pregnancy) were collected immediately after OHE. Following implantation, uteroplacental compartments, understood as full thickness, middle part of the girdle, avoiding the marginal hematoma, were collected. All tissues were cleaned of surrounding connective tissues and fixed with 10% neutral phosphate-buffered formalin for 24 h at +4°C. Then, samples were washed daily with fresh phosphate-buffered saline (PBS) for 1 wk, dehydrated in a graded ethanol series, and embedded in a paraffin equivalent Histo-Comp (Vogel, Giessen, Germany). For total RNA isolation and Western blotting applications, samples from the uterus and uteroplacental compartments were immersed in RNAlater (Ambion Biotechnology GmbH, Wiesbaden, Germany) for 24 h at +4°C and subsequently stored at −80°C until analysis.

Construction of PTGES Expressing EYFP- and His-Tagged Plasmids

The Gateway cloning technology (Invitrogen, Carlsbad, CA) was used for cloning of PTGES expression plasmids according to the manufacturer's protocol and as described previously [7]. Briefly, using the canine-specific sequence of PGES (PTGES) as a template (GenBank accession number NP_001116326; Canis lupus familiaris prostaglandin E synthase [PTGES] mRNA, complete cds) [9], primers were designed, including attB sequences that were attached to forward and reverse oligonucleotides to amplify open reading frame, including PCR products suitable for BP recombination. The stop codon was removed from the reverse primer in order to enable the subsequent synthesis of fusion protein-expressing constructs. The following sequences were used (regions showing gene-specific sequences are in boldface type): attB1-PTGES forward, 5′-GGG GAC AAG TTT GTA CAA AAA AGC AGG CTT AAT GCC TCC CCC TGT CCT GGC; and attB2-PTGES reverse, GGG GAC CAC TTT GTA CAA GAA AGC TGG GTT CAG GTG GCA GGC TGC TTC CC.

The PCR reaction was run with Phusion High-Fidelity DNA hot start polymerase (New England Biolabs, Frankfurt, Germany) in an Eppendorf Mastercycler (Vaudaux-Eppendorf AG, Basel, Switzerland) under conditions previously described [7]. After purification with a gel extraction system (Qiaex II; Qiagen GmbH, Hilden, Germany), the attB-flanked PCR products were used for BP recombination with attP-containing pDONOR221 vector with BP clonase II. The entry clone obtained was amplified in One Shot ccdB Survival 2 T1R-competent cells (Invitrogen) on kanamycin-containing agar plates and purified using the PureYield Plasmid Miniprep System (Promega, Dübendorf, Switzerland). Afterward, entry clones containing attL sites were recombined with pHSV-EYFP-Rfc-C1 and pHSV-V5His-Rfc-C1 expression vectors [20] in the presence of LR clonase II (Invitrogen) to produce plasmids expressing the EYFP (Enhanced Yellow Fluorescent Protein) and His-tagged canine-specific PTGES fusion proteins. The generated expression vectors were amplified in One Shot ccdB Survival 2 T1R Competent Cells (Invitrogen), purified with the PureYield Plasmid Miniprep System (Promega), control digested to confirm the presence of insert, and finally commercially sequenced on both strands (Microsynth, Balgach, Switzerland).

Transient Expression and Subcellular Localization of Recombinant PTGES Protein

The Vero 2.2 cell line served as a heterologous system according to the previously described protocol [7]. Cells were cultured in Dulbecco Modified Eagle Medium-high glucose (Chemie Brunschwig AG, Basel, Switzerland) containing 2 mM l-glutamine, heat-inactivated 2% fetal bovine serum (Chemie Brunschwig AG), and 1% penicillin-streptomycin (Chemie Brunschwig) in humidified air containing 5% CO₂ at 37°C. Transfection of pHSV-EYFP-Rfc-C1-PTGES and pHSV-V5His-Rfc-C1-PTGES expression vectors was achieved with FuGENE HD-transfection reagent (Roche Diagnostics, Mannheim, Germany) at a ratio of 1 μg of DNA to 3.5 μl of reagent. Transfection efficiency was checked by inspecting the pHSV-EYFP-Rfc-C1-PTGES-transfected cells under a fluorescent microscope. For subsequent Western blot analysis, cells were washed with ice-cold sterile PBS and harvested using a NET-2 lysis buffer (50 mM Tris-HCl, pH 7.4, 300 mM NaCl, 0.05% NP-40) containing protease inhibitor cocktail (10 μl/ml; Sigma-Aldrich, Chemie GmbH, Buchs, Switzerland).

The subcellular localization of PTGES-EYFP recombinant proteins was assessed in cells grown on sterile cover glasses, using a Leica SP2 model confocal microscope (Leica Microsystems AG, Heerbrugg, Switzerland). Cells transiently expressing the fluorescent recombinant proteins were fixed with formaldehyde (2% final concentration) for 10 min at 37°C. For permeabilization, PBS-0.1% Triton X-100 was used for 2 min at room temperature, followed by 3 washing steps of 5 min each with PBS. Cells were then incubated for 20 min at room temperature with the respective conjugates (i.e., Alexa Fluor 594-conjugated concanavalin A [20 μg/ml; Invitrogen]) for detecting endoplasmic reticulum (ER) or with Alexa Fluor 633-conjugated wheat germ agglutinin (5 μg/ml; Molecular Probes, Inc., Leiden, Netherlands) for detecting Golgi apparatus. Then, 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) was added to the conjugate solution to achieve nuclear staining. Cells were then washed 3 times for 10 min each with H₂O at room temperature. Finally, cover glasses with cells were mounted with Glycergel (Dako North America, Inc., Carpinteria, CA) on microscope slides, and signals were visualized using confocal microscopy.

Purification of PTGES-His-Tagged Recombinant Protein, Its Enzymatic Activity, and Determination of the Biochemical Potential of Uteroplacental Homogenates to Assist in the Conversion of PGE2 to PGF2α by 9K-PGE2 Reductase Activity

Heterogeneously expressed His-tagged PTGES protein was purified with nickel-nitrilotriacetic acid magnetic agarose beads (Qiagen), using the manufacturer's protocol. An ER isolation kit (Sigma-Aldrich) was used, following the provider's instructions, to isolate microsomal fractions from tissue homogenates obtained from dogs during prepartum luteolysis. Placental and uterine tissue compartments were separated macroscopically and then subjected to the experimental procedure.

The biochemical activities of the recombinant PTGES, as well as that of the uteroplacental microsomes, were determined by applying a recently elaborated protocol [7], with some minor modifications. NADPH (1.0 μM [Sigma-Aldrich]) was used as a cofactor in reactions containing 0.5 μg recombinant protein or uterine and placental microsomal fractions in 100 μl of 50 mM Tris HCl, pH 7.5. As a substrate, 0.001–0.1 μM PGH2 or PGE2 was used. For negative controls, reactions were performed omitting proteins, and detected background signals were subtracted from the readouts obtained in experiments run with proteins. Assays were carried out at 37°C for 1–10 min. Each experiment was repeated at least three times. The following enzymatic conversions were assessed: PGH2 into PGE2 and PGF2α for the PTGES recombinant protein, and PGE2 to PGF2α for the uterine and placental homogenates. The PGF2α and PGE2 enzyme immunoassay kits (Cayman Chemical Co., Ann Arbor, MI) were used according to the manufacturer's instructions. The intra- and interassay coefficient of variation values were approximately 3%. In a final step, plates were read photometrically using a SpectraMax Plus384 absorbance microplate reader (Bucher Biotec AG, Basel, Switzerland) set at 405 nm.

Protein Preparation and Western Blot Analysis

Protein extraction and SDS-PAGE were performed according to our previously published protocol [2122–23]. The NET-2 lysis buffer containing 10 μl/ml protease inhibitor cocktail (Sigma-Aldrich) was used for protein preparation. Proteins were electroblotted onto polyvinylidene difluoride membranes (Bio-Rad Laboratories, Inc., Hercules, CA). The horseradish peroxidase (HRP)-conjugated secondary antibodies were detected with Immun-Star Western C chemiluminescent kit substrate (Bio-Rad) and SuperSignal West Femto Chemiluminescent Substrate (Pierce Biotechnology, Rockford, IL) according to the manufacturers' protocols. Finally, immunoreactive bands were detected using ChemiDoc XRS+ System and Image Lab (Bio-Rad).

Canine-specific guinea pig polyclonal affinity-purified anti-PTGES antibody was generated (Eurogentec SA, Seraing, Belgium) against RSDQDVDRCLRAHRND-C terminal amino acids 61–76 of the canine PTGES protein (GenBank accession no. NP_001116326). Antibody was applied at a 1:25 dilution in the Western blot analysis. Additional antibodies used for this study were rabbit polyclonal, affinity-purified immunoglobulin G (IgG) against EP2 (1:500 dilution) and EP4 (1:500 dilution), together with respective blocking peptides (Cayman Chemical Co.); mouse monoclonal anti-EYFP (1:1000 dilution; product no. 632380; code JL-8; Clontech Laboratories, Saint-Germain-en-Laye, France); monoclonal mouse anti-His (product no. R940-25; 1:100 dilution; Invitrogen); mouse monoclonal anti-β-actin antibody (code sc-69879; 1:1000 dilution; Santa Cruz Biotechnology, Dallas, TX); secondary HRP goat anti-mouse IgG (1:15 000; code W402B; Promega); secondary donkey anti-rabbit HRP-labeled secondary IgG (1:15 000; Pierce Biotechnology, Rockford, IL); and anti-guinea pig IgG conjugated to HRP (code A5545; 1:15 000; Sigma-Aldrich).

Semiquantitation of PTGER2 (EP2) and PTGER4 (EP4) protein expression levels was achieved by reblotting the Western blot membranes with anti-β-actin antibody, used as a loading control. ImageJ software (US National Institutes of Health, Bethesda, MD) was applied to determine the optical density of bands. The values are presented as the ratio of target protein optical density to β-actin optical density.

Immunohistochemistry

The IHC procedure was performed using our standard immunoperoxidase method as reported previously [10, 23]. Briefly, formalin-fixed paraffin-embedded tissues were cut (2–3 μm), mounted on microscope slides (Super Frost; Menzel-Glaeser, Braunschweig, Germany), deparaffinized in xylene, rehydrated in a graded ethanol series, and washed under running tap water. For antigen retrieval, slides were preincubated in 10 mM citrate buffer, pH 6.0, at room temperature prior to microwave irradiation at 560 W for 15 min. Endogenous peroxidase activity was quenched by immersing the sections in 0.3% hydrogen peroxide in methanol. Nonspecific binding sites were blocked by incubation with either 10% horse serum or 10% normal goat serum, depending on the secondary antibody used. Then, respective primary antibodies were applied overnight at +4°C. The primary antibodies used for IHC were canine-specific guinea pig polyclonal affinity purified anti-PTGES (the same as those used for Western blotting; 1:300 dilution; Eurogentec); rabbit polyclonal anti-EP2 (1:200 dilution), and anti-EP4 (1:100 dilution; Cayman Chemical Co.); goat polyclonal affinity purified anti-PGT IgG (1:100 dilution; Santa Cruz Biotechnology); monoclonal mouse anti-vimentin (1:100 dilution) and polyclonal rabbit anti-cytokeratin (1:300 dilution; Dako). Samples omitting the primary antibody and antiserum-specific isotype controls applied at the same dilution and protein concentration as the primary antibody served as negative controls. The following secondary antibodies were used for 30 min at room temperature at 1:100 dilution: biotinylated goat anti-guinea pig IgG BA-7000; goat anti-rabbit IgG BA-1000; horse anti-goat IgG BA-9500; and horse anti-mouse IgG BA2000 (all from Vector Laboratories Inc., Burlingame, CA). Subsequently, slides were incubated for 30 min with the streptavidin-peroxidase Vectastain ABC kit (Vector Laboratories) to enhance the signals. Peroxidase activity was detected using liquid DAB+ substrate kit (Dako). Slides were counterstained with hematoxylin and mounted in Histokit (Assistant, Osterode, Germany).

Total RNA Isolation and Real-Time (TaqMan) PCR

Total RNA extraction was performed using TRIzol reagent (Invitrogen) following the manufacturer's protocol. Concentrations of RNA were determined with a NanoDrop 2000C spectrophotometer (Thermo Fisher Scientific AG, Reinach, Switzerland). Genomic DNA contamination was removed with the RQ1 RNase-free DNase (Promega) following the manufacturer's instructions. Reverse transcription was carried out using random hexamers as primers and other reagents (Applied Biosystems, Foster City, CA) as previously reported [10, 17].

The expression of PGT mRNA in the preimplantation uterus and uteroplacental compartments during pregnancy and normal and induced parturition were analyzed by real-time (TaqMan) PCR using an automated fluorimeter ABI PRISM 7500 Sequence Detection System (Applied Biosystems). The previously described protocol was applied [10, 17]. Each sample was analyzed in duplicate. Autoclaved water instead of cDNA and the so-called minus-reverse transcription controls were run for negative controls. Reactions were set up in mixtures consisting of 200 nM TaqMan probe, 300 nM of each primer, 12.5 μl of Fast Start universal probe Master (ROX; Roche Diagnostics); and 5 μl of cDNA, corresponding to 100 ng total RNA. The cycling conditions were denaturation at 95°C for 10 min, 40 cycles at 95°C for 15 sec, and 1 cycle at 60°C for 60 sec. GAPDH, 18S rRNA, and cyclophilin A (PPIA) were used for normalization of target gene expression in the comparative cycle threshold method (ΔΔCT), ensuring approximately 100% reaction efficiency, according to the manufacturer's protocols for the ABI PRISM 7500 Sequence Detection System and as described previously [10, 17]. The relative gene expression was calculated using the sample with the lowest concentration of target gene as a calibrator.

The primers and probes for PGT, GAPDH, and 18S rRNA were designed based on the canine-specific sequences of PGT (GenBank accession number NM_001011558), GAPDH (GenBank accession number AB028142), and 18S rRNA (GenBank accession number FJ797658). The sequences of primers and probes were PGT forward: 5′-TGC AGC ACT AGG AAT GCT GTT C-3′; PGT reverse: 5′-GGG CGC AGA GAA TCA TGG A-3′; PGT TaqMan probe: 5′-TCT GCA AAC CAT TCC CCG CGT G-3′ (length of the amplicons was 116 base pairs [bp]); GAPDH forward: 5′-GCT GCC AAA TAT GAC GAC ATC A-3′; GAPDH reverse: 5′-GTA GCC CAG GAT GCC TTT GAG-3′; GAPDH TaqMan probe: 5′-TCC CTC CGA TGC CTG CTT CAC TAC CTT-3′ (length of the amplicons was 75 bp); 18S rRNA forward: 5′-GTC GCT CGC TCC TCT CCT ACT-3′; 18S rRNA reverse: 5′-GGC TGA CCG GGT TGG TTT-3′; 18S rRNA TaqMan probe: 5′-ACA TGC CGA CGG GCG CTG AC-3′ (length of amplicons was 125 bp). A commercially available gene-specific TaqMan probe and primer sets of canine-specific cyclophilin A (TaqMan system) were purchased from Applied Biosystems (product no. Cf03986523-gH). Selected PCR products were commercially sequenced (Microsynth) in order to confirm the specificity of amplicons.

In Situ Hybridization

The expression of PGT was additionally localized in the canine uteroplacental compartments at the mRNA level by using a nonradioactive ISH method with formalin-fixed paraffin-embedded sections as described previously [23]. The 348-bp PCR product, amplified with the following primers: PGT forward: 5′-GGA TGA AGC AAG GAA GAT G-3′; and PGT reverse: 5′-ATG GAG ATG GTG ATG ATG G-3′, served as the template for synthesis of the digoxigenin (DIG)-labeled cRNA probes. Detection of DIG-labeled riboprobes was done after incubating slides with alkaline phosphatase-conjugated, sheep anti-DIG Fab fragments (Roche Diagnostics) diluted 1:5000. Signals were detected using the substrate 5-bromo-4-chloro-3-indolyl phosphate and Nitroblue Tetrazolium (NBT/BCIP; Roche Diagnostics).

Data Evaluation

The effects of time on PGT mRNA expression and PTGER2 (EP2) protein expression were assessed by applying a parametric one-way analysis of variance (ANOVA), followed by a Tukey-Kramer multiple comparisons post-test. Due to the uneven distribution of the data obtained for PGTER4 (EP4) protein expression, the Kruskal-Wallis test (a nonparametric ANOVA) followed by Dunn multiple comparison test was performed. Numerical data are presented as means ± standard deviations (S.D.). GraphPad version 3.06 software (GraphPad Software, San Diego, CA) was used. A P value of <0.05 was considered statistically significant.

Expression and Localization of PTGES in Canine Uteroplacental Compartments at Selected Time Points During Pregnancy and at Prepartum Luteolysis

The pHSV-EYFP-Rfc-C1-PTGES and pHSV-V5His-Rfc-C1-PTGES vectors were transfected into Vero 2.2 cells, which served as a heterologous expression system for synthesis of recombinant canine EYFP-tagged and His-tagged PTGES proteins. Transfection efficiency was assessed by monitoring the cellular expression of fluorescent EYFP-PTGES protein (Fig. 1, A and B). Cells were harvested when high transfection efficiency was detected, approximately 24–48 h after transfection, and cell lysates were used in Western blot analysis. Anti-His and anti-EYFP antibodies were used, allowing estimation of the molecular weight of both proteins. Whereas the PTGES-EYFP protein was detected as an immunoreactive band of approximately 44 kDa, the actual size of canine PTGES, determined based on the His-tagged protein molecular weight, was approximately 18 kDa (Fig. 1C). In the next step, the recombinant cell homogenates were probed with the new custom-made anti-PTGES antibody, revealing results similar to those obtained with the anti-EYFP and anti-His antibodies and confirming its specificity (Fig. 1D). In all experiments, nontransfected control Vero cells were used as a negative control, displaying background staining that could be differentiated from the positive signals achieved with the specific antibodies. The efforts to detect Western blot PTGES signals in the native tissue homogenates obtained from uteroplacental compartments from different gestational stages failed, which presumably can be attributed to the lower detection limit of the antibody in these complex tissue homogenates.

FIG. 1

Transfection of Vero 2.2 cells with pHSV-EYFP-Rfc-C1-PTGES (for EYFP-PTGES recombinant fusion-protein expression) and pHSV-V5His-Rfc-C1-PTGES (for His-PTGES recombinant fusion-protein expression) vectors. Validation of canine-specific anti-PGES antibody. A) Brightfield microscopy view of Vero 2.2 cells 24 h after transfection. B) pHSV-EYFP-Rfc-C1-PTGES expression in transfected Vero 2.2 cells. White arrows indicate fluorescing cells expressing EYFP-PTGES (green). C) Expression of EYFP-PTGES (∼44 kDa) and His-PTGES (∼18 kDa) fusion proteins in Vero 2.2 cell lysates was analyzed by Western blot analysis using anti-EYFP and anti-His antibodies. A lysate from nontransfected Vero 2.2 cells was used as a negative control; β-actin expression (45 kDa) served as a loading control. D) Western blot analysis of His-PTGES and EYFP-PTGES recombinant protein expression in Vero 2.2 cell lysates stained with custom-made canine-specific anti-PTGES antibody; nontransfected Vero 2.2 cells were used as negative control, and β-actin (45 kDa) was used as an internal loading control.

The subcellular localization of PTGES-EYFP recombinant protein was visualized using confocal microscopy (Fig. 2). In order to differentiate between the different cellular components, ER was detected by concanavalin A staining, and Golgi apparatus was revealed by wheat germ agglutinin staining. The localization of heterogeneously expressed PTGES-EYFP matched the concanavalin A staining, indicating association of the recombinant protein with the ER (Fig. 2). No overlapping was observed for the expression of PTGES-EYFP and staining of the Golgi apparatus.

FIG. 2

Subcellular localization of EYFP-PTGES recombinant protein in pHSV-EYFP-Rfc-C1-PTGES-transfected Vero 2.2 cells. Endoplasmic reticulum (ER) was detected by concanavalin A staining and Golgi apparatus was visualized by wheat germ agglutinin staining. A) Expression of canine recombinant EYFP-PTGES fusion protein (green). B) ER (red) is shown. C) Golgi apparatus (yellow). D) EYFP-PTGES/ER merged image indicate the localization of PTGES to microsomes. E) EYFP-PTGES/Golgi apparatus merged image; there is no overlapping. F) EYFP-PTGES/ER/Golgi apparatus merged image. G and H) Higher magnification shows the subcellular localization of recombinant EYFP-PTGES fusion protein (G) and ER (H). Nuclear staining was achieved with DAPI (see Materials and Methods).

As determined by IHC, prior to implantation, PTGES was predominantly localized in the uterine surface (luminal) epithelial cells, superficial and deep uterine glands, and myometrium (Fig. 3, A and B). Staining was also visible in vascular endothelial cells. Weaker signals were detected in uterine stromal cells. Following placentation and at mid-gestation, similar uterine localization at placentation sites was seen, with distinctly stronger signals observed in the superficial uterine glands enlarging to the so-called glandular chambers than in the deep uterine glands (Fig. 3, C and D). In the clearly visible placental labyrinth, intense signals were localized in the fetal trophoblast cells from postimplantation until prepartum luteolysis (Fig. 3, E–G). No signals or only weak signals were detectable in the maternal part of the placenta, the so-called decidual cells (Fig. 3, E–G).

FIG. 3

Immunohistochemical localization of PTGES in the canine early pregnant uterus and uteroplacental compartments at selected time points during pregnancy: at the preimplantation stage (A and B), in the uteroplacental compartments at midgestation (C, D, and E), and in the uteroplacental compartments during prepartum luteolysis (F and G). A and B) Preimplantation PTGES is localized to the surface (luminal) epithelial cells (solid arrows [A]), glandular epithelial cells of the superficial and deep uterine glands (open arrows [A and B]), myocytes (solid arrowhead [B]), and vascular endothelial cells (solid arrow [B]). After implantation, within the uteroplacental compartments, endometrial signals are localized to the superficial uterine glands (the so-called glandular chambers) (solid arrows [C]), deep uterine glands (solid arrowheads [D]), endometrial stromal cells (open arrow [C and D]), and vascular endothelial cells (open arrowheads [D]). In the placental labyrinth, fetal trophoblast cells stain strongly (open arrows [E, F, and G]), and maternal decidual cells stain weakly (solid arrows [E and G]) for PTGES. There is no background staining in the isotype control (inset [G]).

Biochemical Activities of Canine PTGES and 9K-PGR Activity of Uteroplacental Homogenates

Canine PTGES-His-tagged fusion protein and the uterine and placental microsomal fractions were purified as described in Materials and Methods. Canine PTGES revealed synthetic activity toward conversion of PGH2 to PGE2 but not toward PGF2α, thereby confirming its biochemical identity. The maximal enzymatic activity, yielding 10.22 ± 2.72 ng/μg protein, was observed with 0.1 μM PGH2 within 10 min. Under the same experimental conditions, the microsomal fractions isolated from uteroplacental compartments during prepartum luteolysis showed the maximum conversion rate of PGE2 to PGF2α, displaying values of 22.8 ± 4.48 ng/μg protein and 25.15 ± 2.98 ng/ μg protein, within 10 min for the uterine compartments at placentation sites and the placenta, respectively. No activity was detected in samples from which the cofactor NADPH was omitted. Further detailed investigations on enzyme kinetics were not within the scope of the present study.

Spatiotemporal, Uteroplacental Expression of PTGER2 (EP2), PTGER4 (EP4), and PGT

Both the PTGER2 (EP2) and PTGER4 (EP4) proteins were clearly detectable in the uteroplacental compartments at all time points investigated. The expression of PTGER4 was time-dependent (P = 0.03) and was significantly decreased (P < 0.05) during prepartum luteolysis (Fig. 4B). A similar expression pattern was observed for PTGER2 expression, but due to very high individual variations, the effects observed were statistically not significant (P = 0.11) (Fig. 4A). The specific signals of either of the proteins were significantly diminished when the antibodies were blocked with epitope-specific blocking peptides (Fig. 4, A and B).

FIG. 4

Expression of PTGER2/EP2 (A) and PTGER4/EP4 (B) proteins in uteroplacental units throughout pregnancy was detected by Western blot analysis. Epitope-blocking peptide was used for both antibodies in order to block specific EP2 and EP4 signals. Representative immunoblots are shown. β-actin (45 kDa) served as the internal loading control.

At the cellular level, as determined by IHC, during preimplantation, PTGER2 and PTGER4 were localized to the endometrial surface epithelium, superficial and deep uterine glands, and myometrium (Figs. 5, A and B, and 6, A and B). After implantation, during mid-gestation and during prepartum luteolysis, a similar localization pattern was seen for both receptors in the uterine parts of the uteroplacental compartments (i.e., superficial uterine glands, the so-called glandular chambers) and deep uterine glands (Figs. 5, C, D, H, and I, and 6, C, D, H, and I). Furthermore, immunoreactive signals for both receptors were detected in uterine stromal cells (Figs. 5, C and D, and 6, C and D). Following placentation, anti-pan-cytokeratin and anti-vimentin staining were performed in consecutive sections in order to more easily distinguish between cell types within the canine placenta (i.e., endothelial, trophoblast and decidual cells (Figs. 5, F and G, and 6, F and G). In the placental labyrinth, stronger signals were localized in the fetal trophoblast cells, especially in the syncytiotrophoblast and invading columnar trophoblast at the base of the placental labyrinth, whereas weaker signals were observed in maternal stroma-derived decidual cells and endothelial cells (Figs. 5, E and J, and 6, E and J).

FIG. 5

Immunohistochemical localization of PTGER2/EP2 in the canine pregnant uterus throughout pregnancy. At the preimplantation stage (A and B), in the uteroplacental compartments at midgestation (C, D, and E), and in the uteroplacental compartments during prepartum luteolysis (H, I, and J). A and B) Preimplantation endometrial PTGER2/EP2 expression is localized to the surface (luminal) epithelial cells (solid arrows [A]), glandular epithelial cells of the superficial and deep uterine glands (open arrows [A and B]) and myocytes (solid arrows [B]). Following implantation, within the uteroplacental compartments, endometrial signals are localized to the superficial uterine glands (the so-called glandular chambers) and epithelial cells of the deep uterine glands (solid arrows [C and D]), endometrial stromal cells (open arrow [C]), and vascular endothelial cells (open arrowheads [D]). A similar localization pattern is observed during prepartum luteolysis in the endometrium: superficial endometrial glands (glandular chambers) and deep uterine glands (solid arrows [H and I]), endometrial stromal cells (open arrows [I]), and vascular endothelial cells (open arrowheads [I]). In the placental labyrinth, fetal trophoblast cells stain strongly (open arrows [E and J]) for PTGER2/EP2. Solid arrowhead in J indicates maternal endothelial cells. The detection of cytokeratin (F) and vimentin (G) was performed in consecutive sections in order to more easily differentiate cell types within the canine placenta, i.e., endothelial, trophoblast, and decidual cells. Trophoblast cells stain positively for cytokeratin (open arrows [F]), while endothelial cells (open arrowheads [G]) and decidual cells (solid arrowheads [G]) stain positively for vimentin. There is no background staining in the isotype control (inset [J]).

FIG. 6

Immunohistochemical localization of PTGER4/EP4 in the canine pregnant uterus at selected periods of pregnancy: at the preimplantation stage (A and B), in the uteroplacental compartments postimplantation (C, D, and E), and in the uteroplacental compartments during prepartum luteolysis (H, I, and J). A and B) Preimplantation endometrial PTGER4/EP4 expression is localized to the surface (luminal) epithelial cells (solid arrows [A]), glandular epithelial cells of the superficial and deep uterine glands (open arrows [A and B]) and myocytes (solid arrows [B]). Following implantation, within the uteroplacental compartments, endometrial signals are localized to the superficial uterine glands (the so-called glandular chambers) (solid arrows [C]), epithelial cells of the deep uterine glands (solid arrowheads [D]), endometrial stromal cells (open arrows [C and D]), and vascular endothelial cells (open arrowheads [D]). Strong endometrial signals are also observed during prepartum luteolysis: in the superficial endometrial glands (glandular chambers) (solid arrows [H), deep uterine glands (solid arrows [I]), endometrial stromal cells (open arrows [H and I]), and vascular endothelial cells (solid arrowheads [H and I]). In the placental labyrinth, invading trophoblast at the base of the placental labyrinth surrounding maternal blood vessels (open arrows [E and J]) and maternal endothelial cells (open arrowheads [E and J]) stain positively for EP4. Solid arrowhead in E indicates maternal decidual cell. The detection of cytokeratin (F) and vimentin (G) was performed in consecutive sections in order to make it easier to differentiate cell types within the canine placenta (i.e., endothelial, trophoblast, and decidual cells). Trophoblast cells stain positively for cytokeratin (open arrows [F]), and endothelial cells (open arrowheads [G]), and decidual cells (solid arrowheads [G]) stain positively for vimentin. There is no background staining in the isotype control (inset [J]).

Because the anti-PGT antibody proved unsuitable for Western blot analysis, the quantitation of gene expression was performed by using semi-quantitative PCR. Consequently, PGT mRNA expression was detected at all selected time points during canine gestation, revealing a significant effect of time (P < 0.02). Its expression was more or less constant until mid-gestation, then showed significantly (P < 0.05) elevated levels prior to parturition, during prepartum luteolysis (Fig. 7). The protein was colocalized with PTGES, PTGER2, and PTGER4 in the endometrial superficial and deep uterine glands, vascular endothelial cells (Fig. 8, A and B), and prior to implantation, in the endometrial surface (luminal) epithelial cells; strong signals were localized in the myometrium (Fig. 8B). Following implantation and placentation endometrial signals in the uteroplacental compartment were maintained in the superficial uterine glands (the so-called glandular chambers), deep uterine glands, and vascular endothelial cells (Fig. 8, C and D). Similar endometrial localization patterns were observed during prepartum luteolysis (Fig. 8, F and G). The placental signals were predominantly localized in the fetal trophoblast cells (Fig. 8E) and weaker signals or no signals were localized in maternal decidual cells. Additionally, at the time of prepartum luteolysis, clearly visible signals were localized in maternal decidual cells (Fig. 8H). A similar signal localization pattern was observed at the mRNA level by in-situ hybridization (Fig. 9, A–E).

FIG. 7

Mean ± SD expression of prostaglandin transporter (PGT) as determined by real time (TaqMan) PCR in the canine uteroplacental units at selected time points during pregnancy. Bars with different letters differ at P < 0.05.

FIG. 8

Immunohistochemical localization of prostaglandin transporter (PGT) in the canine pregnant uterus at selected periods of pregnancy. Preimplantation stage (A and B), uteroplacental compartments postimplantation (C–E), and the uteroplacental compartments during prepartum luteolysis (F–H). A) Preimplantation endometrial PGT expression is localized to the surface (luminal) epithelial cells (weaker signals, solid arrows [A]), glandular epithelial cells of the superficial glands and deep uterine glands (weaker signals, open arrows [A and B]) and vascular endothelial cells (solid arrowhead [B]). Following implantation, within the uteroplacental compartments, endometrial signals are localized to the superficial uterine glands (the so-called glandular chambers) (solid arrows [C]), deep uterine glands (open arrows [D]), media of vessels and vascular endothelial cells (open arrowhead [D]). A similar endometrial localization pattern is observed during prepartum luteolysis; pointed out are superficial endometrial glands (glandular chambers) (solid arrows [F]) and deep uterine glands (solid arrows [G]). In the placental labyrinth, fetal trophoblast cells (solid arrowheads [E and H]) and maternal decidual cells (open arrowheads [E and H]) stain positively for PGT. There is no background staining in the isotype control (inset [H]).

FIG. 9

Localization of prostaglandin transporter (PGT) mRNA as determined by in situ hybridization in the uteroplacental compartments during mid-gestation (A and B) and prepartum luteolysis (C–E). A) Signals are localized to the superficial uterine glands (open arrows) and endometrial stromal cells (solid arrowheads). B) Signals are localized to the deep uterine glands (solid arrows) and myocytes (open arrows). A similar endometrial localization pattern was observed during prepartum luteolysis; pointed out are superficial endometrial glands (glandular chambers; open arrows [C]) and deep uterine glands (open arrowheads [D]). E) The placental expression of PGT mRNA during prepartum luteolysis, localized in the fetal trophoblast cells (open arrowheads), maternal decidual cells (solid arrowheads) and capillary pericytes (solid arrows). There is no background staining in the negative control (inset to E).