Conceptus estrogen and prostaglandins provide the maternal recognition of pregnancy signal to prevent luteolysis during early pregnancy in the pig^†

Abstract

Conceptus estrogens and prostaglandins have long been considered the primary signals for maternal recognition of pregnancy (MRP) in the pig. However, loss-of-function studies targeting conceptus aromatase genes (CYP19A1 and CYP19A2) and prostaglandin–endoperoxide synthase 2 (PTGS2) indicated that conceptuses can not only signal MRP without estrogens or prostaglandins but can maintain early pregnancy. However, complete loss of estrogen production leads to abortion after day 25 of gestation. Although neither conceptus estrogens nor prostaglandins had a significant effect on early maintenance of corpora lutea (CL) function alone, the two conceptus factors have a biological relationship. To investigate the role that both conceptus estrogens and prostaglandins have on MRP and maintenance of pregnancy, a triple loss-of function model (TKO) was generated for conceptus CYP19A1, CYP19A2, and PTGS2. In addition, a conceptus CYP19A2^−/− model (A2KO) was established to determine the role of placental estrogen during later pregnancy. Estrogen and prostaglandin synthesis were greatly reduced in TKO concept uses which resulted in a failure to inhibit luteolysis after day 15 of pregnancy despite the presence of conceptuses in the uterine lumen. However, A2KO placentae not only maintained functional CL but were able to maintain pregnancy to day 32 of gestation. Despite the loss of placental CYP19A2 expression, the allantois fluid content of estrogen was not affected as the placenta compensated by expressing CYP19A1 and CYP19A3, which are normally absent in controls. Results suggest conceptuses can signal MRP through production of conceptus PGE or stimulating PGE synthesis from the endometrium through conceptus estrogen. Failure of conceptuses to produce both factors results in failure of MRP and loss of pregnancy.

Graphical Abstract

Open in new tabDownload slide

Introduction

Maternal recognition of pregnancy (MRP) in mammals is a highly regulated and localized pathway in which the conceptus alters the endometrial environment to favor fetal development and pregnancy to term. The timing and mechanisms controlling MRP varies among species is initiated on day 12 of gestation in the pig [1–3]. The endocrine/exocrine theory [2] has been a long accepted and established model by which the pig conceptuses signal MRP and prevent regression of the corpora lutea (CL). The endocrine/exocrine theory proposed that in the presence of conceptus estrogen secretion on day 12 of pregnancy, endometrial prostaglandin F2α (PGF2α) secretion is rerouted away from the uterine vasculature, where it would typically be transported both locally and systemically to the ovary [4], to being sequestered within the uterine lumen to prevent luteolysis [5]. For over 45 years after publication of the endocrine/exocrine theory [2], conceptus estrogen was considered the primary MRP signal in the pig.

The pig conceptus expresses three aromatase paralogs, CYP19A1, CYP19A2, and CYP19A3 [6, 7] that stimulate early estrogen production. These isoforms are considered the conceptus, placental, and gonadal isoforms, respectively [8]. To better understand the role of conceptus estrogens on MRP, a loss-of-function study targeted the conceptus form of aromatase, CYP19A1 [9]. Although conceptus estrogen synthesis was inhibited, CYP19A1 null pregnancies not only maintained functional CL but were able to maintain pregnancy to day 24–32 of gestation. However, all CYP19A1 null pregnancies were then lost or aborted. Thus, conceptus estrogen is not the only factor involved with CL maintenance in the pig, but is essential to maintain pregnancy to term.

Vanselow et al. [8] evaluated Meyer et al. [9] sequence data and suggested that gene editing had targeted the placental aromatase paralog, CYP19A2, rather than the anticipated CYP19A1 gene. Subsequent resequencing of the CYP19A1^−/− embryos established that the CYP19A1 guide RNA had actually targeted both CYP19A1 and CYP19A2 in the loss-of-function study [10]. Lucas et al. [10] demonstrated that CYP19A1 was the main isoform expressed from days 12–14 and on day 21 with lower levels of CYP19A2 expression during these time periods. CYP19A2 was the primary aromatase paralog expressed from days 21 to 54 while CYP19A3 is primarily expressed in the gonads. Although the collective role of the aromatase isoforms during maternal recognition and the expression patterns of each aromatase paralog that form across early pregnancy has now been established, the individual role of each paralog during early gestation is still unknown. In retrospect, it was fortunate that both conceptus CYP19A1 and CYP19A2 were edited as conceptus and placental estrogen secretion was lost in the Meyer et al. study [8], indicating the estrogen was not essential for MRP in the pig.

Beside estrogen production, conceptus synthesis of prostaglandins (PG), specifically prostaglandin E2 (PGE2), has also been an established model to prevent CL regression in the pig [11]. The uterine luminal content of PGF2α and PGE increase following rapid conceptus elongation on day 12 and attachment to the uterine surface epithelium from day 13 to 18 [12, 13]. During this period of conceptus development, PGE content in the uterine lumen is nearly twice the content of PGF2α in pregnant gilts [11, 12]. The larger ratio of PGE to PGF2α is proposed to serve a luteoprotective role to prevent CL regression [11, 14]. Pfeiffer et al. [15] ablated early conceptus PG production by targeting the conceptus prostaglandin synthase 2 (PTGS2) gene. In this study, PTGS2^−/− conceptuses were capable of signaling MRP and survived beyond day 30 of gestation. Conceptus knockout of both CYP19A1, CYP19A2, [9] and PTGS2 [15] expression leaves a question concerning how luteolysis is inhibited during early pregnancy in the pig. Interleukin 1β2 (IL1B2) and interferon γ (IFNG) are produced by pig conceptuses during MRP but function more in a role for stimulating rapid trophoblast elongation and immune regulation, which are essential for maintenance of pregnancy unrelated to CL maintenance [16].

When comparing the estrogen [9] and PG [15] studies, a link between the two knock-out models were discovered. The content of PGE2 within the uterine flushing’s was comparable between CYP19A1/CYP19A2 null and wild type pregnancies, and despite the ablation of conceptus prostaglandin synthesis, PGE2 was not only present but elevated in uterine flushings of recipients containing PTGS2 null embryos. PGE2 can be produced and released through two different sources. Pig conceptuses express PTGS2 [17, 18] and release PGE2 during the period of MRP [15, 19] while conceptus estrogen or exogenous estrogen is able to stimulate production of prostaglandins, specifically PGE2 from the endometrium [20–23]. By understanding PGE synthesis sources and comparing the effects of each loss-of-function model, it was hypothesized that conceptus aromatases (CYP19A1 and CYP19A2) and conceptus PTGS2 can compensate for one another to ensure either conceptus or endometrial PGE is produced and released to prevent luteolysis.

The objectives of the present studies were to (1) evaluate the role of placental estrogen synthesis from days 21 to 35 in maintenance of embryo development and survival through gene editing of the placental aromatase isoform CYP19A2 and (2) utilize a triple loss-of-function study of CYP19A1, CYP19A2, and PTGS2 to evaluate the combined effects of conceptus estrogen and PGs on luteal function in the pig. By utilizing a triple knock-out model, both conceptus and estrogen-stimulated endometrial PGE synthesis pathways were inhibited which addressed the question of PGE being the key to maternal recognition in the pig.

Materials and methods

All procedures used in this study were conducted in accordance with the Guide for Care and Use of Agriculture Animals in Research and Teaching and approved by the University of Missouri-Columbia Institutional Animal Care and Use Committee under Protocol 8813.

Animals

Recipient gilts utilized for embryo transfer and bred wild-type (WT) gilts serving as normal controls were Large White × Landrace crossbred gilts. All recipients and bred females were of similar age (8–10 months) and weight (100–130 kg). Gilts were observed for estrous behavior twice daily with the onset of estrus designated day 0 of the estrous cycle. Bred gilts were artificially inseminated with semen collected from mature boars 24 and 36 h from detection of estrus.

CRISPR /Cas9 design and production of the CYP19A2^−/− and the CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− cell lines

Aromatase paralog sequences for CYP19A1, CYP19A2, and CYP19A3 are all located on chromosome 1 and closely aligned within the pig genome (CYP19A1: 120,596,235 -120,627,806/ CYP19A2: 120652172–120695899/ CYP19A3:120476890–120556103). CYP19A1 contains 10 exons whereas CYP19A2 and CYP19A3 contain 9 exons. The single porcine PTGS2 gene is located on chromosome 9 (127,850,164 – 127,858,866) and contains 10 exons. Broad Institute design tool (https://broadinstitute.org/gpp/public/analysis-tools/sgrna-design) was used to design one pair of guide RNA (gRNAs) that target exon 4 of porcine CYP19A1 and CYP19A2 genes simultaneously (Supplementary Table S1). The same protocol was used to design a second pair of gRNA that targeted exon 1 of the PTGS2 gene. To minimize off targeting events, each gRNA was tested by using NCBI’s Nucleotide BLAST Tool (https://blast.ncbi.nlm.nih.gov/Blast.cgi) to search for similar sequences in the Sus scrofa genome.

In an attempt to obtain a model for the triple CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− loss-of-function study, two sets of CRISPR gRNAs specific for all three genes, as described above, were inserted into Integrated DNA technology gBlock Gene Fragments and injected into single cell fertilized oocytes as previously described [24]. Injected oocytes were cultured to day 6 as previously described [25] and morula and blastocyst stage embryos (30–50) were co-transferred to a recipient gilt on day 4 or 5 of estrous cycle [26] with wild-type in vitro produced embryos to ensure maintenance of gestation to day 30 [27]. On day 30 of pregnancy, the recipient gilt was euthanized, and fetal fibroblasts were harvested from each of the nine fetuses. Fetal fibroblast cells were lysed for PCR amplification (Supplementary Figure S1A) and PCR products were submitted to Azenta Life Sciences for high throughput amplicon seq (NGS). Although none of the fetuses possessed a triple CYP19A1/CYP19A2/PTGS2 knockout, one fetus (Fetus 4) possessed wild-type alleles for CYP19A1, a biallelic 10 base pair and 8 base pair insertion for CYP19A2 (Supplementary Figure S1B); and a heterozygous wild-type/160 base pair deletion for PTGS2 (Supplementary Figure S1C). NGS sequencing was confirmed by using PCR amplification of fetal fibroblast lysate, cloning of the PCR product into a pCR4-TOPO TA vector, which was submitted to the University of Missouri DNA Core Facility for Sanger Sequencing. With at least one functional allele for both CYP19A1 and PTGS2, but a biallelic edit for CYP19A2, the CYP19A2^−/− fetal fibroblasts were utilized to provide a model for the CYP19A2^−/− loss-of-function study.

Because the initial attempt to obtain a triple CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− knockout through direct zygote injection only provided CYP19A2^−/− fibroblasts to investigate the role of estrogen production during later placental formation, we chose to transfect wildtype porcine fibroblasts via electroporation [28] to create a CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− cell line. Wildtype fetuses were harvested on day 35 of gestation as previously described [28]. Fetal fibroblasts were then dissociated with Accutase (STEMCELL Technologies, 07922) and frozen in 90% Fetal Bovine Serum (FBS) and 10% dimethyl sulfoxide. After 24 h, at −80°C, the cryotubes were transferred to the liquid nitrogen. Annealed gRNA oligonucleotides were inserted into a pX330 plasmid (Addgene, 42230) containing the Cas9 protein coding sequence and cloned into ampicillin-resistant Escherichia coli. Cutting efficiency of each guide was tested previously at different concentrations of 2, 4, or 8 μg total [9, 15]. Guides were transfected into porcine fetal-derived fibroblast cells at a concentration of 4 μg (2 μg per gRNA) to ensure optimal cutting efficiency. Briefly, porcine fetal fibroblast cells (final concentration of 1 × 10⁶/mL) were transfected in 75% cytosalts (120 mM KCl, 0.15 mM CaCl₂, 10 mM K₂HPO₄ [pH 7.6], and 5 mM MgCl₂) and 25% growth factor formulation (Opti-MEM; Gibco, Cat #31985070). Then, two hundred microliters of cells (2 × 10⁵ cells) were electroporated with three consecutive 250-V, 1-ms square wave pulses in a 2-mm gap cuvette in the presence of the CRISPR constructs. Immediately after electroporation, the cells were resuspended in complete media containing Dulbecco Modified Eagle medium (DMEM; Gibco, 11885084) (DMEM supplemented with 0.5% Glutamax (Gibco, 35050061), 12% FBS (Gibco, Grand Island, NY), and 2.0 ng/mL Fibroblast Growth Factor (Sigma-Aldrich, F0291). Four hundred cells were seeded per 100 mm dish for colony formation [28].

Clonal expansion and colony screening

Following culture for 10 days, individual colonies were collected, lysed, and DNA utilized for PCR amplification by using primers specifically designed to amplify the location of gRNA. The PCR product was run on a 2% ethidium bromide gel, and bands were assessed for insertions or deletions. The PCR amplicon from each colony that showed the expected size deletion was cloned into a pCR2.14-TOPO TA vector (Invitrogen, 450641) and topo cloned colonies were submitted to the University of Missouri DNA Core Facility for Sanger Sequencing to identify edits on each allele. No WT alleles were detected for all three genes (CYP19A1, CYP19A2, and PTGS2) in “Colony 151”, after 977 colonies were screened (Supplementary Figure S2A). Colony 151 contains a biallelic 40 base pair deletion for CYP19A1 (Supplementary Figure S2B), a 40 base pair deletion for CYP19A2 (Supplementary Figure S2C), and a 160 base pair deletion on one allele with a 163 base pair deletion on the second allele for gene PTGS2 (Supplementary Figure S2D).

Somatic cell nuclear transfer

Fetal-derived fibroblast cells [29] containing CYP19A2^−/− or CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− gene edits were grown to 80% confluency and used as donor cells for somatic cell nuclear transfer (SCNT) as previously described [30] with modifications to the maturation system [31]. To produce donor embryos for SCNT, ovaries were obtained from a local abattoir or purchased from Applied Reproductive Technology (Madison, WI). Oocytes were matured in vitro, and the cumulus cells were removed with 0.1% hyaluronidase and gentle vortexing. Oocytes were screened, and mature donor oocytes were selected for SCNT if the first polar body had been extruded. The polar body and metaphase II plate were mechanically removed from the oocyte, and donor fetal fibroblast cells (CYP19A2^−/− or CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/−) were inserted into the perivitelline space by micromanipulation [32]. Donor oocytes and fetal fibroblasts were electrically fused by using two DC pulses at 1.2 kV/cm for 30 μs by using a BTX Electro Cell Manipulator (Harvard Apparatus; Holliston, MA). After fusion, the cloned zygotes were chemically activated [33] and cultured [34] for 6 days. Culture media (500 μL/well) was collected to determine prostaglandin synthesis following development to blastocysts on days 5–6 (60 blastocysts/well; in vitro fertilized WT n = 7 wells and CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/−; n = 13 wells).

Embryo transfer

Recipient gilts were monitored for signs of estrus with the onset of estrus designated as day 0 of the estrous cycle. Day 6 morula to blastocyst stage embryos (n = 30–50) were surgically transferred into the oviduct of recipient gilts near the ampullary–isthmic junction on either day 3, 4, or 5 of the estrous cycle as previously described [26]. Blood was collected from the jugular vein via vacutainer tube aspiration for serum progesterone analysis from embryo transfer recipient gilts during early conceptus development and establishment of pregnancy (days 14 and 16) in recipients carrying CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− embryos to determine if luteolysis was occurring.

Embryo collection

Recipient and WT gilts were euthanized via jugular vein injection of Euthasol (Virbac AH, Inc, 200071). The uteri were collected from WT (n = 3) and CYP19A2^−/− (n = 3) recipient gilts at day 15 of gestation. Three WT gilts and seven CYP19A2^−/− recipient gilts were allowed to continue pregnancy to day 30 of gestation before collection. Recipient gilts carrying CYP19A1^−/−/ CYP19A2^−/−/PTGS2^−/− conceptuses were collected on day 14 (n = 4) and day 16 (n = 6). Recipients which had CYP19A1^−/−/CYP19A2^−/−/PTGS2^−/− embryos were collected on day 14 of gestation when CYP19A1 expression would be the highest to determine if conceptus estrogen synthesis was knocked out and day 16 when normal luteolysis is initiated in cyclic gilts. Collection CYP19A2^−/− recipients was delayed to day 15, when CYP19A2 expression is initiated (CYP19A1 expression is low to absent) to determine if conceptus CYP19A2 mRNA expression was effectively knocked out during the second increase in estrogen secretion.

Upon respective collection days, the ovaries and uteri of recipients were harvested and kept on ice to be processed. The uteri were rinsed, and the broad ligament was trimmed from the uterine horns and body. Each uterine horn was clamped approximately 4–6 inches from the cervix, and conceptuses were flushed twice from each individual horn by using 30 mL of sterile Dulbecco’s phosphate-buffered saline (DPBS). Conceptuses retrieved from the uterine lumen flushing (ULF) were washed again with DPBS and examined for their morphology and viability. A subset of day 14 or 15 filamentous conceptus tissue in WT (~ 60 mg wet weight) was immediately placed into 2 mL of DMEM culture media supplemented with 0.03 g/mL bovine serum albumin (Sigma-Aldrich, A7906) and 10 μg/mL gentamicin (Gibco, 2517930) following flushing from the uterine horn. Conceptuses in the culture media (5 to 6 cultures/recipient) were placed in a humidified atmosphere of 5% O₂, 5% CO₂ at 37°C. Conceptus culture media (200 μL) was collected and snap frozen after 24 h of culture for analysis of conceptus PG, and estradiol-17β (E2) production. A subset of conceptus tissue not utilized for in vitro culture and endometrial samples were taken from six random spots along the mesometrial side of the uterine horn and were placed into individual 1.5 mL tubes, flash frozen in liquid nitrogen and stored at −80°C until used for RNA and DNA extraction. Chilled ULF were centrifuged at 3000 g for 10 min to remove uterine debris flash frozen and stored at −80°C for PG, E2, testosterone, IL1B2, and IFNG analyses. Cross sections of the uterine horn were taken from the 4–6 inches closest to the cervix that still contained conceptus tissue and preserved in 10% buffered formalin for 24 h. Samples were then stored in 70% ethanol for immunohistochemistry, which was performed by the VMDL core at the University of Missouri.

For the day 32 WT (n = 3) and CYP19A2 ^−/− (n = 2) uterine collections, the broad ligament was removed from the uterus and the uterine horns were then opened along the antimesometrial border to dissect out each individual placenta and fetus. Fetal, placental, and endometrial tissues were collected, flash frozen, and stored at −80°C until RNA/DNA extraction. Allantoic fluid samples were collected from each fetus, flash frozen and stored at −80°C until used for analysis of E2. Fetal fibroblast cells were collected from each fetus and either lysed for PCR amplification and genotyping or culture propagated and stored in liquid nitrogen. Uterine sections (3 per horn) containing attached placental tissue were removed and fixed in 10% buffered formalin for 24 h. Samples were then stored in 70% ethanol until used for immunohistochemistry.

RNA isolation

Conceptus, placental, and endometrial RNA was extracted using the RNEasy Plus Universal Mini Kit (Qiagen, 73404) according to manufacture instructions. Samples were subjected to an on-column DNase treatment. Quantity and quality of RNA in the samples were established by using a NanoDrop ND-1000 spectrophotometer and visualization of RNA bands with a 1% ethidium bromide integrity gel.

Reverse transcription and real-time qPCR gene expression analysis

Total pooled samples of conceptus tissue (four per recipient), placenta (two per recipient), and endometrial (two per recipient) RNA (1 μg) were reverse transcribed in a 20 μL reaction mix by using iScript Reverse Transcription Supermix for Rt-qPCR (Bio-Rad, 1708841). Amplification of cDNA was performed in an Eppendorf Mastercycler Pro and primed for 5 min at 25°C, reverse transcribed for 20 min at 46°C, and inactivated for 5 min 95°C. Complementary DNA was stored at −20°C until used for real-time qPCR. Reactions were performed and quantified by using the CFX384 Real-Time System (Bio-Rad). Primers were designed based on the S. scrofa  GenBank Ref-Seq mRNA spanning exon-exon junctions whenever possible. Conceptus gene expression was measured by using primers specific for prostaglandin–endoperoxide synthase (PTGS1, PTGS2), interferon (IFNG, IFND), pregnancy associated glycoprotein 6 (PAG6), interleukin 1B2 (ILIB2), aldo-keto reductase family 1 member b (AKR1B1), and 15-hydroxyprostaglandin dehydrogenase (HPGD) (Supplementary Table S2). Endometrial gene expression was measured by using primers (Supplementary Table S3) specific for prostaglandin E synthase 1 (PTGES1), PTGES2, PTGES3, PTGS1, PTGS2, prostaglandin I synthetase (PTGIS), prostaglandin F synthetase (PGFS), ATP-binding cassette, subfamily C, member 4 (ABCC4) and solute carrier organic anion transporter family, member 2A1 (SLCO2A1) and 3A1 (SLCO3A1). Placental gene expression of CYP19A1, CYP19A2, and CYP19A3 in d32 WT and CYP19A2^−/− samples was measured as described previously [10]. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a housekeeping gene in all samples and was not statistically different (P > 0.05) among three treatments. Relative expression of sequence-specific products was obtained after normalization of the target gene cycle threshold (Ct) value by the GAPDH Ct by using the ^-2ΔCT method. Fold change was calculated compared to WT ΔCT average.

Prostaglandin assay

Prostaglandins in conceptus culture media and ULF were analyzed by using a Parent Prostaglandin assay utilized by the Eicosanoid Core Laboratory at Vanderbilt University Medical Center (Nashville, TN). To quantify eicosanoids in ULF samples, 100 μL of fluid was placed in a microcentrifuge tube containing 25% methanol in water (500 μL) and internal standard (d₄-PGE₂ and d₄-LTB₄, 1 ng each). The sample was vortexed and then extracted on an Oasis MAX uElution plate (Waters Corp., Milford, MA) as follows. Sample wells were first washed with methanol (200 μL) followed by 25% methanol in water (200 μL). The sample was then loaded into the well and washed with 600 μL 25% methanol. Eicosanoids were eluted from the plate with 30 μL of 2-propanol/acetonitrile (50/50, v/v) containing 5% formic acid into a 96-well elution plate containing 30 μL water in each well. Samples were analyzed on a Waters Xevo TQ-S micro triple quadrupole mass spectrometer connected to a Waters Acquity I-Class UPLC (Waters Corp., Milford, MA). Separation of analytes was obtained by using an Acquity PFP column (2.1 × 100 mm) with mobile phase A being 0.01% formic acid in water and mobile phase B-acetonitrile. Eicosanoids were separated by using a gradient elution beginning with 30% B going to 95% B over 8 min at a flow rate of 0.250 mL/min.

Measurement of interferon gamma, interleukin-1B, progesterone, testosterone, estradiol-17β, and total prostaglandins

Interferon gamma (IFNG), IL1B, progesterone, testosterone, and total prostaglandins were measured by using a porcine IFNG enzyme-linked immunosorbent assay (ELISA) kit (Sigma-Aldrich, Cat# RAB0226), porcine IL1B ELISA kit (Abcam, Cat# ab100754), human progesterone ELISA kit (Abcam, Cat# ab108670), testosterone ELISA kit (Abcam, ab108666), and total prostaglandin screening ELISA kit (Cayman Chemical, 514012). IFNG, IL1B and testosterone were measured in ULF. Progesterone was measured in blood plasma. Total prostaglandins were measured in blastocyst culture media, conceptus culture media, and ULF. ELISA assays were performed according to manufacturer’s instructions. Concentrations were determine by a standard curve plotted in regression curves (R² > 0.99). Standards were provided in the kits.

Estradiol-17β concentrations were measured in conceptus culture media, ULF, and allantoic fluid samples by using radioimmunoassay as described previously [15, 35]. The sensitivity of the E2 assay was 0.5 pg/mL.

Immunohistochemistry

Immunohistochemistry was performed on endometrium containing SCNT-derived conceptuses on either day 14 for CYP19A1^−/−/ CYP19A2^−/−/PTGS2^−/− or 15 for CYP19A2 ^−/− and WT conceptuses, and day 30 to 32 CYP19A2 ^−/− and WT placental-uterine sections to visualize CYP19A and PTGS2 cellular localization. Formalin-fixed uterine cross-sections were processed, and paraffin embedded in the Sakura Tissue-Tek VIP 6 AI vacuum infiltration processor and the Sakura Tissue-Tek Embedding center. Sections from the paraffin blocks were cut at 5 μm thickness using Leica RM2255 Microtomes. All immunohistochemistry procedures were performed by the VMDL core at the University of Missouri. In brief, sections (5 μm) were deparaffinized, pretreated with Diva Declocker (Biocare Medical, DV2005L2J) for epitope retrieval by incubating sections for 20 min in boiling Reveal Decloaker 10x (pH 6.0; RV1000M, Biocare Medical). After cooling to room temperature, sections were rinsed twice in PBS (pH 7.5) and blocked with Background Sniper (Biocare Medical) at room temperature (RT) for 10 min and then incubated 30 min at RT with either rabbit anti-COX-2 polyclonal antibody (1:250 dilution; 160106, Cayman Chemical) or CYP19A rabbit anti-aromatase polyclonal antibody (1:100 dilution; PA1–21398, Invitrogen). Primary antibodies were diluted in Da Vinci Green Diluent (Biocare, PD900H). Slides were then incubated with Dako EnVision–labeled polymer antirabbit (Dako, K4003) for 30 min, and Romulin Red (Biocare, RAEC810). Sections were washed with PBS and incubated for 30 min with rabbit HRP-conjugated secondary antibody (1:400 dilution; Agilent, K400311–2). Staining was developed with Romulin Red (Biocare, RAEC810) for 10 min. Sections that did not receive the primary antibody were used as a negative control. All samples were counterstained with hematoxylin. Images were captured with a Leica DM5500 B upright microscope and Leica DFC450 C camera using Leica Application Suite X (LAS X).

Statistical analysis

Statistical analysis of continuous variables performed by using GraphPad Prism 10. Normality of residues was checked by Shapiro–Wilk test, and variables were transformed when necessary to fulfill assumptions. Outliers identified using the ROUT method (GraphPad Prism) were excluded from analyses and final figures. Real time-qPCR expression in day 32 endometrium samples and total prostaglandins in blastocyst culture media was analyzed by using an unpaired T-test by conceptus genotype (WT or CYP19A2 ^−/−). All other data was evaluated using an ordinary one-way ANOVA by conceptus genotype (WT vs CYP19A2 ^−/− vs CYP19A1^−/−/ CYP19A2^−/−/PTGS2^−/−). When significant, means across groups were compared by using Tukey’s test. P-values <0.05 indicate a significant difference. Data are shown as mean ± SEM.

Results

In vitro development of CYP19A2^−/− (A2KO) and CYP19A1^−/−/ CYP19A2^−/−/PTGS2^−/− (TKO) embryos

We successfully created an A2KO line where one copy of CYP19A2 had a 10 base pair insertion, and the second allele contained an 8 base pair insertion. Wild-type alleles were detected for CYP19A1 and PTGS2 exhibited a heterozygous deletion of 160 base pair. Additionally, a TKO line was created, where no WT alleles were detected for all three genes (CYP19A1, CYP19A2, and PTGS2). All edits described for A2KO, and TKO lines resulted in frameshifts in the translational reading frame that would cause an enzyme deficiency. The blastocyst development rate for A2KO and TKO SCNT embryos was 28.5% and 28.8%, respectively. SCNT-derived pig embryos, on average, develop to the blastocyst stage 30% of the time [36], making early embryo development comparable between knockouts and controls.

Uterine flushing of recipient gilts containing WT, A2KO, and TKO embryos on days 14 and 15 of gestation

At day 15 of gestation, recipients carrying WT and A2KO embryos contained elongated filamentous conceptuses in the uterine lumen (Figure 1). TKO embryos collected at day 14 had also undergone elongation and were morphologically normal (Figure 1). When flushed from the uterine horn, the amount of conceptus tissue present was similar between A2KO and TKO groups but was visually greatly less than the large mass of conceptus tissue from obtained from WT recipient (Figure 1) females. The differences in conceptus quantity were expected as number of cloned embryos that survive (8 to 10) will be reduced compared with WT which can have 16 to 20 embryos developing. In addition, cloned embryos are 24 to 36 h behind in development compared with WT. Morphologically normal CL were present in the ovaries of all bred and recipient gilts collected on days 14 and 15 (Figure 1).

Conceptus IL1B2 and IFNG gene expression and protein secretion on either day 14 and 15 of gestation

Relative abundance of IL1B2 mRNA (Figure 2A) was not different between WT and A2KO conceptuses. However, TKO conceptus IL1B2 abundance was 200 and 1000-fold greater compared with WT (P < 0.0001) and A2KO conceptuses (P < 0.0002), respectively. The greater gene expression in TKO conceptuses is a day effect resulting from TKO conceptuses having just undergone rapid trophoblast elongation on day 14 compared to WT and A2KO filamentous conceptuses which were collected on day 15. IL1B2 content in the ULF of day 15 WT, day 15 A2KO, and day 14 TKO was not different (Figure 2B). Relative abundance of conceptus IFNG mRNA (Figure 2C) was approximately 2-fold greater in TKO conceptuses compared to WT (P < 0.05) and A2KO conceptuses (P < 0.01). However, IFNG content in the ULF was similar between treatment groups (Figure 2D).

Testosterone and estradiol-17B in conceptus culture media and uterine luminal flushings

As expected, overall steroidogenesis was not affected by editing conceptus CYP19A1 and CYP19A2 expression. The total content of testosterone in ULF of WT (439.7 + 31.2 pg), A2KO (320.7 + 48.0 pg) and TKO (320.5 + 53.3 pg) recipients was not significantly different (P > 0.10). However, the concentration of E2 in conceptus culture media was affected by genotype. In vitro production of E2 was greater in WT compared with A2KO (P < 0.001) and TKO (P < 0.0001) conceptus cultures (Figure 3A). The concentration of E2 was also lower in TKO compared to A2KO conceptus (P < 0.02) culture media. Although the concentration of E2 was significantly lower in conceptus culture media of TKO compared to WT and A2KO conceptuses, there were a few replicate dishes from two of four TKO recipients that contained unexpectedly high levels of E2 (see individual culture well concentrations Figure 3A). These replicates were not excluded from the analysis and account for most of the E2 measured in this group. Transcripts for CYP19A1 were detected by end-point PCR analysis in conceptus tissue from two of the four TKO pregnancies (Supplementary Figure SS3). The results support the possibility the unexpected conceptus E2 production in these pregnancies were caused by presence of parthenotes in addition to TKO conceptuses. Parthenogenesis is a known risk factor when utilizing SCNT for studies on early development in the pig, as parthenotes can develop up to day 30–35 of gestation before degeneration [37]. Although having some parthenogenic conceptuses can cause skewed data and results, genotyping of additional conceptuses from these pregnancies has confirmed the anticipated edits for CYP19A1. If these cultures were excluded from the analysis the concentration of conceptus E2 would have been 50% lower. Total content of E2 in ULF of recipient gilts was significantly greater in WT compared to A2KO (P < 0.0002) and TKO (P < 0.0001) recipients (Figure 3B). There was no significant difference in E2 between the ULF of recipients carrying A2KO and TKO conceptuses.

Conceptus in vitro PG synthesis and content in uterine luminal flushings

Concentration of total PGs in culture media was significantly lower (P < 0.0001) for TKO (0.02 + 0.006 ng/mL) compared to control (0.68 ng + 0.17 ng/mL) blastocysts. The concentration of PG in day 14/15 conceptus culture media was greatest in WT compared with A2KO (P < 0.0003) and TKO (P < 0.0001) conceptuses (Figure 3C). The concentration of PG in culture media was also lower in TKO compared with A2KO (P < 0.0001) conceptuses. Total ULF PG content was greatest in WT compared with A2KO (P < 0.0005) and TKO (P < 0.0001) recipients while ULF PG was lower (P < 0.10) in TKO compared to A2KO recipients (Figure 3D).

The concentration of PGE in the culture media of WT conceptuses (Figure 3E) was greater compared to A2KO (P < 0.03) and TKO conceptuses (P < 0.0002). The PGE concentration was also greater in culture media of A2KO compared with TKO conceptuses (P < 0.06). Total ULF PGE content was greater in WT compared with A2KO (P < 0.02) and TKO (P < 0.0002) recipients while content was lower in TKO (Figure 3F) compared with A2KO (P < 0.02) recipients. While the concentration of PGF2α in culture media was not statistically different between TKO and A2KO conceptuses, the concentration of PGF2α was greater in WT compared with A2KO (P < 0.0001) and TKO (P < 0.0001) cultures (Figure 3G). The total ULF content of PGF2α was not statistically different between WT and A2KO recipients (Figure 3H) but was greatly lower in TKO compared to WT (P < 0.02) and A2KO recipients (P < 0.002).

The total culture media content of six keto PGF1α was similar between WT and A2KO (Figure 3I) but was lower in TKO compared to WT (P < 0.0006) and A2KO (P < 0.004) conceptuses. The total UFL content of six keto PGF1α was not different for the treatment groups (Figure 3J). The in vitro culture media concentration of PGJ2 was not statistically different between WT and A2KO conceptuses (Figure 3K) but was lower in TKO compared to both WT (P < 0.0001) and A2KO (P < 0.0003) conceptuses. The total content of PGJ2 in ULF was not significantly different between WT and A2KO or A2KO and TKO recipients (Figure 6L). However, ULF PGJ2 content of WT was greater compared to TKO (P < 0.02) recipients. No statistically significant difference was detected among treatment groups for TxB2 concentration in conceptus culture media (Figure 3M), or total ULF content of TxB2 (Figure 3N). Conceptus synthesis of 15 keto PGE2 was low to undetectable in culture media. The content of 15 keto PGE2 in the ULF was not statistically different between WT (34.8 + 6.4 ng), and A2KO (34.8 + 11.8 ng) compared with TKO (11.3 + 3.1 ng) recipient gilts P > 0.10).

Conceptus and endometrial gene expression on days 14 and 15

The relative abundance of conceptus mRNA for PTGS1 (Figure 4A) was not affected by genotype. Relative abundance of PTGS2 (Figure 4B) was significantly different between WT and TKO conceptuses (P < 0.03). Conceptus PAG6 expression was lower for TKO compared with WT (P < 0.07) and A2KO (P < 0.05) conceptuses (Figure 4C). Abundance of IFND mRNA was not affected by genotype (Figure 4D). Expression of AKR1B1 (Figure 4E) was significantly greater in A2KO compared to TKO conceptuses (P < 0.07). A genotype affect was not detected for endometrial mRNA abundance for PTGES (Figure 4F), PTGES2 (Figure 4G), PTGES3 (Figure 4H), and PTGS1 (Figure 4I). Endometrial expression of PTGS2 (Figure 4J) was only greater in TKO compared to WT recipients (P < 0.04). Expression of PTGIS mRNA was not affected by genotype (Figure 4K) while PGFS was not statistically different between A2KO and TKO endometrium, and endometrial WT expression was greater compared with A2KO (P < 0.0003) and TKO (<0.0001) endometrium (Figure 4L). The relative abundance of endometrial SLCO2A1 mRNA was not different between treatment groups (Figure 4M) but endometrial mRNA for SLCO3A1 (Fig. 4N) was greater in WT compared to A2KO (P < 0.0009) and TKO (P < 0.0008) recipient endometrium. Also, endometrial mRNA abundance was greater in WT recipients compared with A2KO (P < 0.002) and TKO (P < 0.002) recipients (Figure 4O).

Immunohistochemistry

Expression of CYP19A protein was detected in the trophoblast of WT conceptuses on day 15 of pregnancy while there was little to no staining for CYP19A in the trophoblast of A2KO and TKO conceptuses (Figure 5A). On day 30–32, CYP19A staining was present at the placental/maternal luminal epithelial interface of WT and A2KO recipients.

The endometrial epithelium of all recipients expressed PTGS2 on day 14–15 and 30–32 of pregnancy (Figure 5B). Although the trophoblast of WT and A2KO expressed PTGS2 on day 15 of pregnancy, PTGS2 was absent in TKO conceptuses. PTGS2 was present in the chorion of both WT and A2KO placenta.

Uterine flushing of recipient gilts containing day 16 TKO embryos

On day 16, the CL present on the ovaries of all TKO recipients (n = 6) were regressing (Figurre 6A) compared to the day 15 A2KO and WT recipients. Elongated intact TKO conceptuses were flushed from the uterine lumen of five recipients while spherical trophoblastic vesicles were flushed from one recipient on day 16 (Figure 6B). Plasma progesterone concentrations indicated that luteolysis was initiated as progesterone significantly (P < 0.0001) decreased from 29.7 + 3.5 ng/mL on day 14 to 7.0+ 2.8 ng/mL on Day 16. Plasma progesterone concentrations had decreased from day 14 to 16 in all TKO recipients collected (Figure 6C). Plasma progesterone concentration of two recipients indicated the CL had fully regressed. Thus, despite the presence of viable TKO conceptuses in the uterine lumen, the CL were regressing or had regressed.

Analysis of allantoic fluid E2 concentrations in WT and A2KO recipient gilts

All WT pregnancies (n = 3) were maintained to day 32 while two A2KO gilts were pregnant on day 32 and 4 gilts returned to estrus beyond the length of a normal day 21 estrous cycle on day 26, 27, 34, and 36. Pregnancy was confirmed by ultrasonography in all three WT and two A2KO gilts on at days 25 and 29 of gestation (Figure 7A and B). Both recipients carrying A2KO embryos possessed healthy fetuses with morphologically normal developing placentae (Figure 7C). The number of embryos collected from the uterus of the three WT pregnancies was 10, 17, and 18 while uteri of the two A2KO recipient gilts contained 8 and 9 embryos. The average allantoic fluid volume was 140 and 244 mL for the embryos of WT and A2KO embryos, respectively.

The concentration of E2 in the allantoic fluid of A2KO embryos was significantly greater (P < 0.0001) compared with WT embryos (Figure 8A). This result was unexpected since CYP19A2 is considered the major aromatase gene expressed by the placenta during this period of pregnancy [10]. Genome sequencing reconfirmed the expected 10 base pair and 8 base pair insertions for gene CYP19A2 in each of the A2KO fetuses (Figure 8B). While the expression of dysfunctional CYP19A2 mRNA was expected, end-point PCR amplification by using primers specific for CYP19A1, and CYP19A3 detected expression of both CYP19A1 and CYP19A3 mRNA in A2KO placentae with no expression in WT controls (Figure 8C). Thus, CYP19A1 and CYP19A3 expression appears to be upregulated to compensate for the loss of CYP19A2 activity in A2KO fetuses/placentae.

Endometrial gene expression on day 32

Endometrial mRNA expression of PGFS (Figure 9A), PTGS1 (Figure 9B), PTGES2 (Figure 9E), PTGES3 (Fig. 9F), and PTGIS (Figure 9G) were not significantly different by genotype. Relative abundance of PTGS2 mRNA was significantly higher (P < 0.01) in the endometrium of A2KO recipients compared with WT gilts (Figure 9C). There was also a significant increase (P < 0.04) of PTGES expression in the endometrium of WT and A2KO recipients (Figure 9D).

Discussion

Maternal recognition of pregnancy, first coined by Rodger V. Short in 1969 [38], refers to a chemical signal from the conceptus which maintains the function of CL past the point of cyclic regression. In the pig, the importance of conceptus estrogens during early pregnancy, and their role during the period of maternal recognition, has been highlighted for nearly 70 years.

Kidder et al. [39] was the first to demonstrate the role of estradiol in luteal maintenance of the pig by administering a synthetic non-steroidal estrogen to cycling gilts on day 11 of estrus and extending the estrous cycle up to 25 days which was later confirmed by Gardner et al. [40]. Perry et al. [41] established that early pig conceptuses had the capacity to synthesize estrogens through aromatase conversion of androstenedione and dehydroepiandrosterone. These results were confirmed [42] and extended to later stages of conceptus and placental development [43]. The timing behind estrogen production from the conceptus was further categorized by Zavy et al. [12] and Geisert et al. [13]. Conceptus in utero estrogen synthesis and secretion during early conceptus development and uterine attachment is biphasic, coinciding with elongation of the conceptus on days 11–12 of gestation and again during trophoblast-endometrial attachment and placental membrane formation from days 15 to 30. Administration of exogenous estrogen from days 11 to 15 [44] or days 11 and 14 to 18 of the estrous cycle to mimic the conceptus biphasic release of estrogen [3] induces pseudopregnancy which could be extended beyond 60 days. Thus, estrogen can prevent luteolysis and ensure long term maintenance of pregnancy in the pig.

To better understand initiation of luteolysis in the pig and how conceptus estrogens played a role in controlling these mechanisms, estrogens and prostaglandins were measured in the utero-ovarian vein plasma of pregnant and cyclic gilts [45]. It was discovered that the uterus, specifically the luminal and glandular epithelium, secreted pulses of PGF2α into the uterine vasculature at day 15 of the estrous cycle, initiating luteal regression. However, this release of PGF2α was not observed in the utero-ovarian circulation of pregnant females [45]. Frank et al. [44, 46] expanded on these findings, and found that exogenous estrogen administration significantly lowered concentrations of PGF2α in the utero-ovarian vasculature, which closely mirrored the prostaglandin profiles of pregnant gilts in the Moeljono study [45]. However, although the vascular concentrations of PGF2α were lower in estrogen treated gilts, the uterine luminal content of PGF2α was greatly increased from days 15 to 18 compared with controls [44, 46]. The “redirection” of PGF2α into the uterine lumen provided the basis for the endocrine/exocrine theory of MRP [2].

To directly investigate the role of conceptus estrogen during the period of maternal recognition in the pig, CRISPR/CAS9 gene editing was utilized to generate a loss of function study for conceptus aromatase gene CYP19A1 [9]. By targeting CYP19A1, Meyer was able to generate conceptuses that were unable to synthesize estrogens, and thus, in theory, would not provide the signal for MRP. Surprisingly, CYP19A1^−/− embryos could not only signal MRP, but recipient gilts were able to maintain pregnancy to around day 25 of gestation. However, all pregnancies were lost between days 25 and 32 of gestation. Thus, early conceptus estrogen is not essential to early embryo development or MRP in the pig but is essential for maintenance of pregnancy to term.

It is important to note that the pig does not possess a single aromatase gene but rather three aromatase paralogs: CYP19A1, CYP19A2, and CYP19A3 [7, 47]. These paralogs are currently classified as the conceptus, placental, and gonadal forms, respectively, of aromatase in the pig [6]. The expression patterns of the three aromatase isoforms have been characterized across early pregnancy in the pig [10]. CYP19A1, the conceptus paralog, is primarily expressed in early spherical to filamentous day 12 stage embryos and day 15 conceptuses. The placental isoform, CYP19A2, is lower but detectable in spherical and filamentous day 12 conceptuses and is the primary isoform expressed from days 21 to 54. Lastly, CYP19A3 expression was found primarily in the fetal ovary and testes [10], confirming that it is indeed the major isoform expressed by the gonads [6]. Vanselow et al. [8] indicated that the gRNA used in the Meyer study edited CYP19A2 rather than the CYP19A1 paralog. Fortunately, resequencing of Meyer et al. [9] conceptuses established that both CYP19A1 and CYP19A2 were edited [10] allowing for a complete loss of conceptus and placental estrogen production and solidifying the conclusion that neither CYP19A1 nor CYP19A2 are essential to maternal recognition in the pig but are necessary for maintenance of pregnancy past day 32.

While the Meyer et al. [9] study highlighted the effects of losing both CYP19A1 and CYP19A2 function during early pregnancy, the individual roles and relationships of the three aromatase isoforms is not well understood. The present study provided a model to directly address the question concerning the essential need of placental CYP19A2 expression for estrogen production in maintenance of pregnancy beyond 25 days of gestation. It was hypothesized that CYP19A2^−/− (A2KO) conceptuses would lose the ability to produce estrogen past day 21 of gestation and pregnancy would be terminated by day 30 of gestation as occurred in the Meyer et al. [9] study.

Wild-type and A2KO conceptuses were collected on day 15 of pregnancy to ensure that A2KO conceptuses could maintain corpora luteal function past the point of luteal regression. Due to a functional copy of CYP19A1, A2KO conceptuses were anticipated to produce estrogen during early development on day 15, although less since the gene edit makes CYP19A2 non-functional. Estradiol in the conceptus culture media and ULF was significantly less in A2KO pregnancies when compared to artificially inseminated WT pregnancies. The decrease in estradiol not only reflects loss of CYP19A2 activity but also an effect of cloning and in vitro culture. Cloned embryos develop approximately 36 to 48 h slower than WT embryos [48], which means that the day 15 A2KO embryos are developmentally staged to be comparable to day 13.5 of development. Due to the biphasic release pattern of conceptus estrogens (days 11–12 and 15 to 30), A2KO embryos would have begun to decrease estrogen production following elongation while WT conceptus estrogen synthesis is increasing during the second phase of estrogen production.

All WT pregnancies and two A2KO recipients-maintained pregnancy to day 32. Four other A2KO recipients extended beyond day 21 of a normal estrous cycle but did not maintain the pregnancy, similar to results of Meyer et al. [9]. The embryos and placenta of the day 30 A2KO recipient pregnancies were morphologically normal and healthy. Surprisingly, E2 was not only present in the allantoic fluid of the A2KO embryos but was significantly greater than E2 levels produced by WT embryos. While CYP19A2 (placental aromatase) is normally the only aromatase paralog expressed in the WT placenta [10]; in the present study, CYP19A1 and CYP19A3 were expressed in the placenta of A2KO fetuses. Immunohistochemistry detected expression of CYP19A (aromatase) and PTGS2 in the A2KO and WT placentae. Thus, CYP19A2 may play a role in downregulating the expression of the other two aromatase paralogs, and/or when CYP19A2 function is lost the conceptus and gonadal paralogs will be expressed to ensure estrogen synthesis occurs and pregnancy is maintained to term. These results indicate the adaption of placenta to meet the essential need for continued production of estrogens during the period of placental development and throughout pregnancy.

After the Meyer et al. [9] study indicated that conceptus estrogen production is not the sole regulator of preventing luteolysis during maternal recognition in the pig; it was unclear what other conceptus factors could be contributing to luteal maintenance. Ford et al. [11] indicated that conceptus prostaglandins, specifically PGE2 played a luteotropic role during early pregnancy. To better understand the role of conceptus prostaglandins during maternal recognition of pregnancy, Pfeiffer et al. [15] targeted conceptus PTGS2 and ablated early conceptus prostaglandin synthesis. PTGS2^−/− recipient pregnancies were maintained to day 30 of gestation and indicated that the loss of early conceptus prostaglandin synthesis was not essential to maternal recognition or maintenance of pregnancy to day 30 of gestation.

Although pregnancies were maintained to day 30 in both the Meyer et al. [9] and Pfeiffer et al. [15] studies despite the genetic modifications, both conceptus estrogen and prostaglandins may serve a role in preventing luteolysis in the pig. Luminal content of PGE2 is comparable between Meyer’s CYP19A1^−/−/ CYP19A2^−/−, Pfeiffer’s PTGS2^−/−, and their WT pregnancies on Day 14. Increased luminal content of PGE2 is achieved through two sources (1) PTGS2 in the conceptus converts arachidonic acid into the precursors necessary to produce conceptus prostaglandins (PGE), prostacyclin and thromboxanes and (2) estradiol from the conceptus stimulates prostaglandin synthesis from the endometrium. The conglomerate effect of conceptus E2 and PG role in endometrial function is also supported by the study of Kaczynski et al. [49] who suggested that estrogen or PGE2 alone did not mimic the endometrial transcriptome expression on day 12 of pregnancy. However, endometrial transcriptome of cyclic gilts administered both estradiol and PGE2 were similar to day 12 of pregnancy [49]. Given these findings, it was hypothesized that either conceptus estradiol or prostaglandins can act as maternal recognition signals to ensure proper PGE2 production and maintain CL.

The triple loss-of-function model targeting conceptus CYP19A1, CYP19A2, and PTGS2 was established to determine if loss of both conceptus estrogen and prostaglandin production would inhibit proper PGE production and prevent luteal maintenance past day 15 of pregnancy. TKO pregnancies were collected at day 14 of gestation to ensure that TKO conceptuses were developmentally competent up to the point that luteolysis would be initiated in cyclic pigs. Conceptuses were collected on day 14 which would be close to completion of rapid conceptus elongation and the time of highest production of E2 during the first phase of E2 synthesis. Despite the loss of conceptus estrogens and prostaglandins, TKO conceptuses were still present and viable in the uterine lumen at day 14 gestation. Immunohistochemistry confirmed the lack of CYP19A and PTGS2 production in TKO conceptuses. Conceptus expression and production of IL1B2, IFNG and PAG which contribute to conceptus elongation, immune regulation, and proinflammatory responses during early pregnancy were not compromised. As anticipated, testosterone present in the uterine lumen was comparable between all groups, indicating that androgen precursors for the steroidogenesis pathway were present in all pregnancies. Despite the availability of testosterone, estradiol was significantly lower in conceptus culture media and uterine luminal flushings of TKO groups meaning that knockout conceptuses were incapable of converting androgens into estrogens due to loss of aromatase.

Because functionality of PTGS2 is a rate limiting step of prostaglandin synthesis, conceptus prostaglandin production was significantly reduced for total PG and downstream production of specifically PGE, PGF2α, PGJ2, and PGI2 which was also reflected in the UFL of TKO recipient gilts. We also observed a decrease in prostaglandins in conceptus culture media and uterine flushings from A2KO embryos. Based on our experience we do believe that decreased prostaglandin levels were due to the delayed development of clones, however we did not evaluate if PTGS2 is haploinsufficient for prostaglandin production.

The endocrine/exocrine model for MRP (inhibiting luteolysis) is based on the sequestering of PG, specifically PGF2α in the uterine lumen of pregnancy to prevent movement into the uterine vasculature to stimulate CL regression. With the loss of conceptus estrogen and PG synthesis in TKO conceptuses, PG was not sequestered in the uterine lumen compared with WT and A2KO recipients and luteolysis occurred on day 16 in the presence of viable conceptuses. It is apparent that although conceptus estrogen and prostaglandin production alone are not essential to early embryo development, luteolysis occurs in the absence of both conceptus factors. Although production and contribution of either conceptus or endometrial PGJ2 and PGI2 [49] should not be disregarded as serving a role in early establishment of pregnancy, either direct PGE2 production from the conceptus or conceptus estrogen stimulated PGE synthesis by the endometrium appear to be the major regulators of CL maintenance in the pig. The lack of an estrogen induced endometrial upregulation of PTGES1, PTGES2, and PTGES3 expression indicates conceptus E2 does not alter expression of PTGES in pregnant animals to increase PGE synthesis. It is possible that estrogen and PGE may serve to activate PG transporters in the epithelium to sequester PG within the uterine lumen [50]. Seo et al. [50] indicated that pig endometrium expressed mRNA for both ABCC4 (also known as multidrug resistant protein 4) and SLCO2A1 (also known as prostaglandin transporter, PGT) with expression being most abundant on day 12 of pregnancy. ABCC4 protein localized mainly to endometrial LE and GE, whereas SLCO2A1 localizes primarily to endothelial cells of blood vessels. Single cell analyses of day 15 pregnant endometrium localized SLCO2A1 in endometrial endothelial cells while ABCC4 and SLCO3A1 (also known as organic anion transporter) expression is primarily in GE and ciliated LE (Sponchiado MS and Geisert RD unpublished data). Endometrial expression of ABCC4 and SLCO3A1 were greater in WT compared to A2KO and TKO recipients. Both ABCC4 and SLCO3A1 transports PGE1 and PGE2 with the highest affinity [51–53] which would assist in providing greater movement of PGE into the uterine vasculature to protect the CL from luteolysis. Interestingly, SLCO3A1 can also regulate inflammation through activation of nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) activity [54]. The role of PG transport in the endometrium needs to be more thoroughly investigated.

It is apparent that although conceptus estrogen and prostaglandin production alone are not essential to early embryo development, luteolysis occurs in the absence of these two conceptus factors. We interpret our results to suggest that, after nearly 40 years, an adaption can be made to the endocrine/exocrine theory in which, rather than estrogen alone, both conceptus estrogens and prostaglandins ensure proper PGE production for luteal maintenance past day 15 of gestation which serves as the true maternal recognition signals to prevent luteolysis during pregnancy in the pig.

Acknowledgment

We acknowledge technical assistance provided by Dr. Kristen Whitworth and acquisition of the porcine fetal-derived fibroblast cells from the National Swine Resource and Research Center (U42OD011140). The authors would like to thank Jason Dowell for assistance with embryo transfers and Melissa S. Samuel for help with the surrogates before conceptus collections.

Author contributions

R.M.S., C.G.L., M.S., R.D.G., K.D.W., and R.S.P. contributed to the conception of the study and interpretation of the results. R.M.S., C.G.L., and R.D.G. wrote the paper and M.F.S., R.S.P., and K.D.W. participated in paper revision. R.M.S., C.G.L., and K.D.W. designed guide RNAs and assisted in colony growth and selection. E.M.K., L.D.S., and R.D.G. generated cloned embryos and preformed the embryo transfers. R.M.S., C.G.L., E.K.M., and M.S. assisted with animal uterine collection and sampling of conceptus and endometrial tissue. R.M.S., C.G.L., and M.S. performed RT-PCR, and immunohistochemical analyses. R.M.S., M.F.S., and M.C.L. performed hormone assays. M.S. and M.C.L. performed the statistical analyses of the experimental data. All authors read and approved the final manuscript.

Conflict of Interest: The authors have declared that no conflict of interest exists.

Data availability

The data underlying this article are available in the article and in its online supplementary material.

Footnotes

References