Abstract

In cattle, elevated concentrations of circulating progesterone (P4) in the immediate postconception period are associated with advanced conceptus development, while low P4 is implicated as a causative factor in low pregnancy rates observed in dairy cows. This study aimed to: 1) describe the transcriptional changes that occur in the bovine endometrium during the estrous cycle, 2) determine how elevated P4 affects these changes, 3) identify if low P4 alters the expression of these genes, and 4) assess the impact that low P4 has on conceptus development. Relatively few differences occurred in endometrial gene expression during the early luteal phase of the estrous cycle (Day 5 vs. 7), but comparison of endometria from more distant stages of the luteal phase (Day 7 vs. 13) revealed large transcriptional changes, which were significantly altered by exogenous supplementation of P4. Induction of low circulating P4 altered the normal temporal changes in gene expression, and these changes were coordinate with a delay in the down-regulation of the PGR from the LE and GE. Altered endometrial gene expression induced by low P4 was associated with a reduced capacity of the uterus to support conceptus development after embryo transfer on Day 7. In conclusion, the present study provides clear evidence that the temporal changes in the transcriptome of the endometrium of cyclic heifers are sensitive to circulating P4 concentrations in the first few days after estrus. Under low P4 conditions, a suboptimal uterine environment with reduced ability to support conceptus elongation is observed.

Introduction

Progesterone (P4) plays a key role in the reproductive events associated with the establishment and maintenance of pregnancy. Elevated concentrations of circulating P4 in the immediate postconception period are associated with an advancement of conceptus elongation [13], an increase in interferon-tau production [4, 5], and higher pregnancy rates in cattle and sheep [68]. In cattle, approximately 40% of conception loss is estimated to occur in the period from Day 8 to Day 16 of pregnancy (Day 0 = ovulation) [9]; a considerable proportion of this loss may be attributable to inadequate circulating P4 concentrations and the subsequent downstream consequences on endometrial gene expression [10] and histotroph secretion into the uterine lumen [11].

Preparation of the uterine endometrium for embryo attachment and implantation in all studied mammals, including ruminants, involves carefully orchestrated spatiotemporal alterations in transcriptomic profiles. In both cyclic and pregnant animals, similar changes occur in endometrial gene expression up to initiation of conceptus elongation, suggesting that the default mechanism in the uterus is to prepare for an expected pregnancy [12]. This is supported by the fact that it is possible to transfer an embryo to a synchronous uterus 7 days after estrus and establish a higher rate of pregnancy, as is routine in commercial bovine embryo transfer. It is only in association with maternal recognition of pregnancy, which occurs on approximately Day 16 in cattle, that significant changes in the transcriptomic profile are detectable between cyclic and pregnant endometria [12], when the endometrium responds to increasing concentrations of interferon-tau secreted by the filamentous conceptus. Elevated P4 advances the transcriptomic changes in the endometrium, which normally occur during pregnancy, resulting in enhanced conceptus elongation [10]. Interestingly, we have shown that the embryo does not have to be present in the uterus during the period of P4 elevation in order to benefit from it, supporting the concept that the positive effect on conceptus growth is mediated via P4-induced changes in the endometrial transcriptome [13].

Low P4 concentrations have been implicated as a causative factor in the low pregnancy rates observed in dairy cows [14]. Recent evidence from our group has shown that lower circulating concentrations of P4 in postpartum dairy cows are associated with an impaired ability of the oviduct/uterus to support embryo development compared with that of dairy heifers [15]. However, the mechanism of action of low P4 concentrations on reduced embryo/conceptus development and survival are not known. We have recently developed a low-P4 model whereby injection of PGF2α on Days 3, 3.5, and 4 of the estrous cycle induces a significant reduction in circulating P4 concentrations [16]. The use of this model, which is independent of the many confounding factors that affect pregnancy rates in dairy cows, can facilitate the understanding of the mechanisms of action of low P4 on conceptus development. Thus, it was hypothesized that, in heifers with induced low P4, sequential modulations of the endometrial transcriptome will occur, leading to a suboptimal uterine environment incapable of supporting normal conceptus development. The specific objectives of this study were to: 1) describe the changes that occur in the endometrial transcriptome during the bovine estrous cycle, (2) determine how elevated P4 affects the temporal pattern of gene expression in the endometrium of cyclic heifers, (3) determine if the expression of these genes is altered in heifers with low P4, and (4) determine the consequences of low P4 for conceptus development following embryo transfer.

Materials and Methods

All experimental procedures involving animals were licensed by the Department of Health and Children, Ireland, in accordance with the Cruelty to Animals Act (Ireland 1897) and the European Community Directive 86/609/EC, and were sanctioned by the Animals Research Ethics Committee, University College Dublin, Ireland. Unless otherwise stated, all reagents were sourced from Sigma-Aldrich (Dublin, Ireland). All GeneChip products were from Affymetrix (Santa Clara, CA).

Experiment 1

The aim of this experiment was to describe global changes in the endometrial transcriptome during the luteal phase of the bovine estrous cycle. It was carried out as part of a larger study identifying changes in the endometrial transcriptome under varying conditions of pregnancy and/or P4 status with the data relating to pregnant animals already published [10]. The days for tissue collection were chosen to coincide with key stages of embryo development in pregnant animals: 1) Day 5 = 16-cell/early morula-stage embryos, 2) Day 7 = blastocyst stage, 3) Day 13 = initiation of conceptus elongation, and 4) Day 16 = filamentous conceptus at the time of maternal recognition of pregnancy. As such, tissues derived from the noninseminated, cyclic control heifers were taken at the same time points; however, only data relating to cyclic heifers are presented in this paper. The experimental model used has been previously described [3, 10]. Briefly, estrus was synchronized in crossbred beef heifers (n = 263) using an 8-day CIDR treatment (controlled internal drug release device, 1.94 g P4; InterAg, Hamilton, New Zealand) with an intramuscular injection of a PGF2α analogue (2 ml Estrumate containing 0.5 mg cloprostenol; Shering-Plough Animal Health, Hertfordshire, UK) 1 day before CIDR removal. Only those heifers that came into standing estrus (Day 0; n = 210) within a narrow window (36–48 h after CIDR removal) were used. One-third of them were randomly assigned to be noninseminated cyclic heifers. In order to elevate circulating concentrations of P4, approximately one-half of the heifers received a P4-releasing intravaginal device (PRID, containing 1.55 g P4; CEVA Animal Health Ltd., Chesham, U.K.) on Day 3 of the estrous cycle. Thus, this generated two treatment groups: 1) cyclic heifers with normal P4, and 2) cyclic heifers with high P4. Blood samples were taken from all heifers twice daily from Day 0 to Day 7 after estrus, and once daily thereafter until slaughter, to characterize the postovulatory changes in concentrations of P4 in serum; these profiles have been previously published [3], and confirm that PRID treatment, as expected, significantly increased P4 concentrations between Day 3.5 and Day 8. Heifers were slaughtered on Day 5 (n = 16), 7 (n = 15), 13 (n = 12), or 16 (n = 11) of the luteal phase of the estrous cycle.

Tissue collection

Within 30 min of slaughter, the reproductive tract of each heifer was recovered and placed on ice. The uterine horn ipsilateral to the corpus luteum (CL) was flushed with 20 ml of PBS containing 5% fetal calf serum (FCS). Strips of endometrial tissue (∼300 mg) were removed from the ipsilateral horn and immediately immersed in 1:5 w/v of RNAlater. Endometrial samples were stored at 4°C for 24 h. Excess RNAlater was removed and samples were placed into new RNase DNase-free tubes and stored at −80°C until subsequent RNA extraction.

RNA extraction and microarray hybridization

RNA extraction and subsequent microarray analyses was carried out as described previously [10] on endometria from cyclic heifers with high or normal P4 on Days 5, 7, 13, and 16 of the estrous cycle (n = 5 animals per treatment per time point; n = 40 microarrays in total). Total RNA was extracted from 100 mg of endometrial homogenate using the Trizol reagent (Invitrogen, Carlsbad, CA) per the manufacturer's instructions. On-column DNase digestion and RNA cleanup was performed using the Qiagen mini kit (Qiagen, Crawley, West Sussex, U.K.) and RNA quality and quantity was assessed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA).

All microarray protocols were carried out by Almac Diagnostics Ltd. (Craigavon, Armagh, Northern Ireland), as previously described [10]. Briefly, 2 μg of total RNA was converted to cDNA via first- and second-strand synthesis using the GeneChip Expression 3′-Amplification One-Cycle cDNA Synthesis kit and the GeneChip Eukaryotic PolyA RNA Control Kit. This double-stranded cDNA was cleaned up using the GeneChip Sample Cleanup Module. Biotin-labeled cRNA was synthesized using the GeneChip Expression 3′-Amplification in vitro transcription (IVT) labeling kit and unincorporated NTPs were removed using the GeneChip Sample Cleanup Module. Complimentary RNA quality was assessed using an Eppendorf biophotometer (Eppendorf International, Westbury, NY) and the Agilent 2100 Bioanalyzer. Complimentary RNA (25 μg) was fragmented by incubation with 5× fragmentation buffer and RNase-free water at 94°C for 35 min. The quality of the 35–200 base fragments generated for hybridization was assessed using the Agilent 2100 Bioanalyzer. Once the yield was adjusted, 15 mg of fragmented cRNA was made up in hybridization cocktail, per the Affymetrix technical manual, corresponding to a 49-format (standard)/64-format array. The hybridization cocktail was added to the GeneChip Bovine Genome Array and hybridized for 16 h at 45°C, washed, and stained and scanned using the GeneChip Scanner 3000.

Experiment 2

The aim of Experiment 2 was to determine how the temporal changes, identified in Experiment 1, were affected by low P4 concentrations. The estrous cycles of an additional batch of crossbred beef heifers were synchronized as described in Experiment 1 (with the exception that CIDRs, containing 1.38 g P4, were used; Pfizer Animal Health, Ireland), and only heifers detected in standing estrus (n = 40) were used. All heifers were randomly assigned to one of two treatments: 1) cyclic heifers with normal P4 (n = 12), and 2) cyclic heifers with low P4 (n = 28). To achieve low P4 concentrations, heifers were given three intramuscular injections of 0.5 mg of a PGF2α analogue (Estrumate) on Days 3, 3.5, and 4 of the estrous cycle, as previously described [16]. Increased numbers of heifers were assigned to the low P4 group, as previous studies have shown that a proportion of treated animals either do not respond to the PGF2α administration or undergo complete luteolysis [16]. All heifers were blood sampled daily from estrus until slaughter on either Day 7 or Day 13 of the estrous cycle to monitor P4 concentrations. These days were chosen to coincide with the timing of the down-regulation of the nuclear P4 receptor (PGR) in the luminal epithelium (LE), a key event required for the normal temporal regulation of gene expression in the endometrium [17].

Tissue collection and serum P4 analysis

At slaughter, all reproductive tracts were processed as described in Experiment 1 above. Additionally, for immunohistochemistry, whole-uterine cross-sections were removed from the tip of each ipsilateral uterine horn and immediately preserved in 10% buffered formalin for 24 h. Samples were then processed by dehydration through a series of ascending concentrations of ethanol, cleared in xylene, and mounted in paraffin wax blocks. Serum P4 concentrations were determined using the solid-phase RIA Coat-A-Count P4 kit (Siemens Medical Solutions Diagnostics, Los Angeles, CA). Serum (100 μl) was placed in antibody-coated tubes in duplicate. A further 750 μl of iodinated P4 tracer was added to each tube. The tubes were then incubated at 37°C for 1 h for the rapid P4 assay or at room temperature for 3 h for the standard assay. Tubes were decanted and counted for 90 sec using a Wallac 1470 gamma counter (Wallac/Perkin Elmer). Recovery of added mass of 40 ng P4 was 99.8%. Cross-reactivity with 17-α-hydroxy P4, 5-α-dihydroxy P4, corticosterone, cortisone, and hydrocortisone was 19.5%, 13.4%, 0.57%, 0.04%, and 0.03%, respectively. Parallelism studies between the standard curve and diluted serum indicated no interference with serum. The intra-assay coefficients of variation (CVs) were 16.3%, 9.9%, and 8.3%, and interassay CVs were 16.5%, 10.1%, and 8.3% for low-, medium-, and high-quality control serum pools, respectively. The assay sensitivity was 0.03 ng P4/ml serum. Animals in the low-P4 group that did not respond to PGF2α administration or those that underwent luteolysis were discarded. This generated five animals per time point for control heifers, and seven animals per time point for heifers with low-P4 concentrations.

RNA isolation and quantitative real-time PCR analysis

Quantitative real-time PCR (Q-RT-PCR) was performed on endometrial tissues for genes with the largest P4-induced fold change differences on Days 5 and 7 of the estrous cycle, as determined in Experiment 1 (Table 1), and on other candidate genes identified from our previous studies, to contribute to an optimal uterine environment [18]. In order to ensure that alterations in gene expression in the endometrium were due to P4 treatment, all Q-RT-PCR was performed on a separate set of endometrial tissue from that which was used for microarray gene expression analysis. RNA was extracted as described for Experiment 1. Total RNA (5 μg) was converted to cDNA using Superscript III (Invitrogen, Carlsbad, CA) and random hexamers, per the manufacturer's instructions. Complimentary DNA was then quantified and diluted to give a final concentration of 10 ng/μl. All primers were designed using Primer Express Software (Applied Biosystems, Foster City, CA; Table 2), and Q-RT-PCR was carried out on the 7500 Fast Real-Time PCR System (Applied Biosystems). Reactions, consisting of 50 ng of cDNA, 300 nmol concentration of primers, and 10 μl Sybrgreen mastermix (Applied Biosystems), were carried out in duplicate with a final reaction volume of 20 μl made up with RNase DNase-free H2O. All reactions were carried out under the following cycling conditions: 50°C for 2 min, 95°C for 10 min, and 40 cycles at 95°C for 15 sec and 60°C for 1 min, with the inclusion of a dissociation curve to ensure specificity of amplification. A standard curve was included for each gene of interest, as well as for the normalizer gene, to obtain primer efficiencies.

Table 1

Entrez gene symbol and gene description for genes altered to the greatest extent by high P4 on Days 5 and 7 of the estrous cycle and used for Q-RT-PCR analysis in heifers with low P4 (experiment 2).

Entrez gene ID Gene description Day of cycle Experiment 1: Affymetrix microarraya 
Fold change (high P4) P value 
DKK1 Dickkopf homolog 1 −3.06 <0.05 
DGAT2 Diacylglycerol O-acyltransferase homolog 2 2.94 <0.05 
DGAT2 Diacylglycerol O-acyltransferase homolog 2 3.65 <0.01 
CYP26A1 Cytochrome P450, family 26, subfamily A, polypeptide 1 3.78 <0.05 
ASGR2 Asialoglycoprotein receptor 2 3.32 <0.01 
HAVCR1 Hepatitis A virus cellular receptor 1 N-terminal domain containing protein 2.58 <0.01 
FABP3 Fatty acid binding protein 3, muscle and heart (mammary-derived growth inhibitor) −3.03 <0.01 
PRC1 Protein regulator of cytokinesis 1 −3.07 <0.01 
SLC2A5 Solute carrier family 2 (facilitated glucose/fructose transporter), member 5 −3.58 <0.05 
Entrez gene ID Gene description Day of cycle Experiment 1: Affymetrix microarraya 
Fold change (high P4) P value 
DKK1 Dickkopf homolog 1 −3.06 <0.05 
DGAT2 Diacylglycerol O-acyltransferase homolog 2 2.94 <0.05 
DGAT2 Diacylglycerol O-acyltransferase homolog 2 3.65 <0.01 
CYP26A1 Cytochrome P450, family 26, subfamily A, polypeptide 1 3.78 <0.05 
ASGR2 Asialoglycoprotein receptor 2 3.32 <0.01 
HAVCR1 Hepatitis A virus cellular receptor 1 N-terminal domain containing protein 2.58 <0.01 
FABP3 Fatty acid binding protein 3, muscle and heart (mammary-derived growth inhibitor) −3.03 <0.01 
PRC1 Protein regulator of cytokinesis 1 −3.07 <0.01 
SLC2A5 Solute carrier family 2 (facilitated glucose/fructose transporter), member 5 −3.58 <0.05 
a

Fold change (high P4 compared with normal P4 heifers) and significance (P < 0.05) are given for each gene as determined by microarray analysis.

Table 2

Gene accession number, Entrez gene symbol, gene name, and forward and reverse primer sequences for all genes used for Q-RT-PCR.

Accession no. Entrez gene symbol Gene name Forward primer Reverse primer 
AV607592 ACTB Actin, beta CGCCATGGATGATGATATTGC AAGCCGGCCTTGCACAT 
NM_001075144 ANPEP Alanyl (membrane) aminopeptidase ATCCGGATGCTCTCGAATTTC TCTGATAGGCAAAGGTCTGCAA 
NM_174030 CTGF Connective tissue growth factor CGTGTGCACCGCTAAAGATG TCCGCTCTGGTACACAGTTCCT 
NM_001075120 LPL Lipoprotein lipase CAGGTCGAAGTATCGGAATCCA GAAAGTGCCTCCGTTAGGGTAAA 
NM_180998 LTF Lactotransferrin GAAGGTAGATTCGGCGCTGTA CAGTTTCCCTGAGGTTCTTCAAG 
XM_583951 PGR Predicted: Bos taurus progesterone receptor, mRNA GAGAGCTCATCAAGGCAATTGG CACCATCCCTGCCAATATCTTG 
NM_205793 DGAT2 Bos taurus diacylglycerol O-acyltransferase homolog 2 (mouse) (DGAT2), mRNA GGTGCTACAATGGGTCCTGT TGAAGTAGAGCACGGCAATG 
XM_508572 DKK1 Dickkopf homolog 1 (Xenopus laevis) GAAGGTTCTGCCTGTCTTCG TGCACACTTGACCTTCCTTG 
XM_874268 CYP26A1 Cytochrome P450, family 26, subfamily A, polypeptide 1 TGCCGATATCTTCACCAACA TCCTCCAAATGGAATGAAGC 
NM_001075952 ASGR2 Similar to Asialoglycoprotein receptor 2 (ASGP-R 2) (ASGPR 2) (Hepatic lectin H2) GACTTCTGGAAGCCAAGCTG ATGGGGAAGTGCTTCAGATG 
NM_001046523 HAVCR1 Bos taurus hepatitis A virus cellular receptor 1 N-terminal domain containing protein (MGC137099), mRNA GCTTTCAGAAACCCTGCTTG GCCAGATTTCCCTACGTTCA 
NM_174313 FABP3 Fatty acid binding protein 3, muscle and heart (mammary-derived growth inhibitor) AAATTCTCCTGGGGTCAGGT GCCTTGGCTCTGCTTTATTG 
NM_001076543 PRC1 Protein regulator of cytokinesis 1 AAAGCATCTAGGCGTGAGGA GTGTCTGGCGTGTGCTCTAA 
NM_001101042 SLC2A5 Solute carrier family 2 (facilitated glucose/fructose transporter), member 5 AGTCTCCTGGCAAACGAAGA AAGAAGGGCAGGAAGAGGAG 
NM_177518.1 AGPAT1 1-Acylglycerol-3-phosphate O-acyltransferase 1 (lysophosphatidic acid acyltransferase, alpha) TCCCTCATCCTCAGTGGTTC CGTGGGAGTGAGGAATTGTT 
NM_001075932.1 AGPAT5 1-Acylglycerol-3-phosphate O-acyltransferase 5 (lysophosphatidic acid acyltransferase, epsilon) CTGAGGCTGGAAGGAAACTG AATTCAGGAGCAATCGGATG 
NM_001034036.1 PPARA Bos taurus peroxisome proliferator-activated receptor alpha (PPARA), mRNA CCCTCTTTGTGGCTGCTATC GCACAATACCCTCCTGCATT 
NM_001083636.1 PPARD Bos taurus peroxisome proliferator-activated receptor delta (PPARD), mRNA ATGCCCTGGAACTTGATGAC TCCACCTGAGACACGTTCAT 
NM_181024.2 PPARG Bos taurus peroxisome proliferator-activated receptor gamma (PPARG), mRNA GATCTTGACGGGAAAGACGA GGGGACTGATGTGCTTGAAC 
XM_881943.3 RXRA PREDICTED: Bos taurus retinoid X receptor, alpha, transcript variant 4 (RXRA), mRNA GAGGCGTACTGCAAACACAA AAGAGGTGCTCCAGGCACT 
NM_001083640.1 RXRB Bos taurus retinoid X receptor, beta (RXRB), mRNA CCCTGGAGACCTACTGCAAA GCCTATGGACCTGAGAGCAG 
Accession no. Entrez gene symbol Gene name Forward primer Reverse primer 
AV607592 ACTB Actin, beta CGCCATGGATGATGATATTGC AAGCCGGCCTTGCACAT 
NM_001075144 ANPEP Alanyl (membrane) aminopeptidase ATCCGGATGCTCTCGAATTTC TCTGATAGGCAAAGGTCTGCAA 
NM_174030 CTGF Connective tissue growth factor CGTGTGCACCGCTAAAGATG TCCGCTCTGGTACACAGTTCCT 
NM_001075120 LPL Lipoprotein lipase CAGGTCGAAGTATCGGAATCCA GAAAGTGCCTCCGTTAGGGTAAA 
NM_180998 LTF Lactotransferrin GAAGGTAGATTCGGCGCTGTA CAGTTTCCCTGAGGTTCTTCAAG 
XM_583951 PGR Predicted: Bos taurus progesterone receptor, mRNA GAGAGCTCATCAAGGCAATTGG CACCATCCCTGCCAATATCTTG 
NM_205793 DGAT2 Bos taurus diacylglycerol O-acyltransferase homolog 2 (mouse) (DGAT2), mRNA GGTGCTACAATGGGTCCTGT TGAAGTAGAGCACGGCAATG 
XM_508572 DKK1 Dickkopf homolog 1 (Xenopus laevis) GAAGGTTCTGCCTGTCTTCG TGCACACTTGACCTTCCTTG 
XM_874268 CYP26A1 Cytochrome P450, family 26, subfamily A, polypeptide 1 TGCCGATATCTTCACCAACA TCCTCCAAATGGAATGAAGC 
NM_001075952 ASGR2 Similar to Asialoglycoprotein receptor 2 (ASGP-R 2) (ASGPR 2) (Hepatic lectin H2) GACTTCTGGAAGCCAAGCTG ATGGGGAAGTGCTTCAGATG 
NM_001046523 HAVCR1 Bos taurus hepatitis A virus cellular receptor 1 N-terminal domain containing protein (MGC137099), mRNA GCTTTCAGAAACCCTGCTTG GCCAGATTTCCCTACGTTCA 
NM_174313 FABP3 Fatty acid binding protein 3, muscle and heart (mammary-derived growth inhibitor) AAATTCTCCTGGGGTCAGGT GCCTTGGCTCTGCTTTATTG 
NM_001076543 PRC1 Protein regulator of cytokinesis 1 AAAGCATCTAGGCGTGAGGA GTGTCTGGCGTGTGCTCTAA 
NM_001101042 SLC2A5 Solute carrier family 2 (facilitated glucose/fructose transporter), member 5 AGTCTCCTGGCAAACGAAGA AAGAAGGGCAGGAAGAGGAG 
NM_177518.1 AGPAT1 1-Acylglycerol-3-phosphate O-acyltransferase 1 (lysophosphatidic acid acyltransferase, alpha) TCCCTCATCCTCAGTGGTTC CGTGGGAGTGAGGAATTGTT 
NM_001075932.1 AGPAT5 1-Acylglycerol-3-phosphate O-acyltransferase 5 (lysophosphatidic acid acyltransferase, epsilon) CTGAGGCTGGAAGGAAACTG AATTCAGGAGCAATCGGATG 
NM_001034036.1 PPARA Bos taurus peroxisome proliferator-activated receptor alpha (PPARA), mRNA CCCTCTTTGTGGCTGCTATC GCACAATACCCTCCTGCATT 
NM_001083636.1 PPARD Bos taurus peroxisome proliferator-activated receptor delta (PPARD), mRNA ATGCCCTGGAACTTGATGAC TCCACCTGAGACACGTTCAT 
NM_181024.2 PPARG Bos taurus peroxisome proliferator-activated receptor gamma (PPARG), mRNA GATCTTGACGGGAAAGACGA GGGGACTGATGTGCTTGAAC 
XM_881943.3 RXRA PREDICTED: Bos taurus retinoid X receptor, alpha, transcript variant 4 (RXRA), mRNA GAGGCGTACTGCAAACACAA AAGAGGTGCTCCAGGCACT 
NM_001083640.1 RXRB Bos taurus retinoid X receptor, beta (RXRB), mRNA CCCTGGAGACCTACTGCAAA GCCTATGGACCTGAGAGCAG 

Protein localization by immunohistochemistry

Localization of protein for the nuclear PGR was carried out by immunohistochemistry, as previously described [19], using monoclonal mouse antibodies against both the A and B isoforms (PGR-AB) or just the B isoform (PGR-B) (Laboratory Instruments and Service Centre, Ashbourne, Ireland). Briefly, 5-μm sections of uterine horn were cut from paraffin-embedded blocks, mounted on glass slides coated with 98% 3-aminopropyl triethoxy-silane, and dried overnight at 56°C. Unless otherwise stated, all washes were carried out with 0.05 M Tris-buffered saline (TBS; pH 7.7) twice for 5 min at room temperature in a humid chamber. All slides were deparaffinized with xylene, rehydrated through a series of graded ethanol washes (70%–100% ethanol) and antigen retrieval performed (slides were heated for 20 min in 0.01 M sodium citrate buffer, pH 6.0 and cooled for 20 min). Endogenous peroxidase was then blocked, (30 min in a 1% hydrogen peroxide in methanol solution) and nonspecific binding inhibited by incubating the slides with 2% normal rabbit serum in 0.05 M TBS v/v for 30 min. Slides were then incubated with their respective primary antibodies for the optimized time and concentrations (PGR-AB, 4 μg/ml; PGR-B, 5 μg/ml; 4°C overnight [19]). After overnight incubation, all slides were washed and incubated for a further 45 min with a rabbit anti-mouse polyclonal biotinylated secondary antibody (0.013 μg/ml). All slides were then washed, incubated for 30 min with avidin-biotin complex (Vectastain Elite ABC Kit; Vector Labs, Peterborough, U.K.) and washed again. The 3,3′-diaminobenzidine tetrahydrochloride chromogen substrate was then added to the slides for 10 min, flushed with distilled water for 7 min, and dehydrated through a series of increasing ethanol concentrations (70%–100%; 5 min), and subsequently cleared in two successive xylene containers for 10 min each. Finally, all slides were mounted using DPX (AGB Scientific Ltd., Dublin, Ireland) and observed under 10× magnification. Intensity of staining of the LE, superficial glands (SGs), deep glands (DGs), stroma (STR), and myometrium (MYO) was determined using a digital camera, with four images captured per tissue section (two images showing the LE, SGs, and STR, and two images showing the DGs and MYO). Intensity of staining for all the regions was determined using Image-Pro Plus software (version 6.2; MediaCybernetics, Bethesda, MD), as previously described [19].

Experiment 3

To test the hypothesis that a modified uterine environment induced by low circulating P4 decelerates posthatching conceptus elongation, Day 7 blastocysts produced in vitro, in the absence of P4, were transferred to synchronized recipient heifers (n = 20 blastocysts per recipient) with normal (n = 9) or low (n = 11) P4 concentrations. Cross-bred beef heifers of similar average age (2.2 ± 0.23 yr) and weight (484 ± 8.58 kg) were synchronized using a CIDR device in conjunction with PGF2α administration, as described in Experiment 1. Low P4 was achieved by administration of PGF2α on Days 3, 3.5, and 4 of the estrous cycle, as described in Experiment 2, and was confirmed in blood samples that were collected once per day from the onset of estrus until slaughter on Day 14. Samples collected from Day 0–6 were analyzed for P4 concentrations on Day 6 using a room temperature rapid version of the assay described for Experiment 2 to ensure that the selected recipients (n = 11) had low P4 concentrations within the required range (>1 SD from the controls).

All heifers were scanned on Days 3, 5, and 6 postestrus using a portable Voluson I scanner (GE Healthcare, Chalfont St. Giles, UK) with a linear (RSP 6–16 MHz) transducer. Blood flow around the CL was measured using Doppler color flow. All Doppler images were “frozen” and stored for subsequent analysis using the software, Pixel Flux (Chameleon-Software, Leipzig, Germany). This program was used to evaluate the area of color pixels (cm2) in the CL as a semiquantitative measurement of luteal blood flow. Serum P4 concentrations were measured as described in Experiment 2. The intra-assay CVs for low-, medium-, and high-quality control serum pools were 8.5%, 4.8%, and 4.9%, and interassay CVs (n = 3) for the same samples were 10%, 4.9%, and 5%, respectively.

Following slaughter, the reproductive tracts of heifers were recovered and transported back to the laboratory, where both uterine horns were flushed with 40 ml PBS supplemented with 1% FCS to recover the conceptuses. The presence and number of conceptuses were determined using a stereomicroscope, and their dimensions were measured in a Petri dish over a transparent, graduated grid (1-mm graduations). All CL were dissected out of the ovaries and weighed.

Data Analyses

For analysis of the array data, raw signal intensities were read into R and preprocessed using functions of both affy and GCRMA packages of the BioConductor project [20]. Correspondence analysis [21] was performed to determine the greatest source of variation in the tissue samples. Lists of differentially expressed genes (DEGs) were identified using the Limma package [22], employing linear modeling and an empirical Bayes framework to shrink the variance of measurements on each probe set. A modified t-test was then carried out, and all P values were adjusted for multiple testing using the Benjamini and Hochberg false discovery rate method. Lists of DEGs for the various comparisons were chosen on the basis of an adjusted P < 0.05. For gene ontology (GO) overrepresentation analysis, P < 0.01 was used as the cutoff point. Analysis of the GO terms was performed using the GOstats package of Bioconductor [23]. The chip probes were first filtered as outlined in GOstats vignette providing a “gene universe” that represents the expressed genes exhibiting some variability across all experimental conditions. Filtering reduced the amount of false positives resulting from the analysis (i.e., GO terms marked as statistically significant, when, in truth, they are not). For each list of significant probes generated from the microarray analysis (DEGs), a conditional hypergeometric statistical test was performed using a cutoff P value of 0.01. This selected the overrepresented GO nodes (i.e., those associated with the probe list more than would be expected by chance based on the “gene universe”) while taking into account the structural relationship between terms in the GO graph.

Ingenuity pathway analysis (IPA; http://www.ingenuity.com) was carried out for each list of DEGs generated from different comparisons. To identify the pathways that were associated with each list of DEGs, a P value for a given function was calculated by considering the number of functional analysis molecules that participate in that function (e.g., ligand, receptor, etc.) and the total number of molecules that are known to be associated with that function/pathway in the Ingenuity Knowledge Base. Functions and pathways with P values less than 0.05 were considered significant for a given list of DEGs. The ratio indicates the number of genes for a particular pathway/ontology in a gene list compared to the total number of genes in a given pathway.

Interaction networks for each list of DEGs were generated by identifying genes that serve as molecules of interest. These molecules of interest interact with other molecules in the Ingenuity Knowledge Base, and were identified as network-eligible molecules and used as “seed” molecules to generate a network. These network-eligible molecules were combined into networks that maximize their specific connectivity, which was identified as their interconnectedness with each other relative to all molecules to which they were connected in the Ingenuity Knowledge Base. Additional molecules from the Ingenuity Knowledge Base were used to specifically connect two or more smaller networks and merge them into a larger one, with a limit of 35 molecules for each interaction network generated. The networks generated were scored based on the number of network-eligible molecules they contain. The network score is based on the hypergeometric distribution, and is calculated with the right-tailed Fisher exact test. The higher the score, for a given interaction network, the lower the probability of finding the observed number of network-eligible molecules in a given network by chance.

Differences in serum P4 concentrations between groups were analyzed by ANOVA. CL blood flow was analyzed using a repeated measures model that was developed using the MIXED procedure in SAS 9.1. Differences in mean conceptus length, width, area, and number were analyzed by Student t-test [24]. Similarly, differences in gene expression, using the relative standard curve method, and differences in protein intensities were analyzed using the Student t-test.

Results

Experiment 1

Principal component analysis revealed that the source of greatest variation in the transcriptional profile of the endometria is day of the estrous cycle. Endometria recovered from cyclic heifers on Days 5 and 7 clustered together, while endometria from Days 13 and 16 clustered together (Fig. 1), with Day-13 and Day-16 endometria clustering together to a greater extent than the Day-5 and Day-7 endometria. To identify the specific genes altered by elevated P4 on a particular day of the estrous cycle, the endometrial transcriptome of cyclic heifers with high and normal P4 was compared on Days 5, 7, 13, and 16. Relatively low numbers of genes were altered by high P4 on a given day of the estrous cycle (36 on Day 5, 910 on Day 7, 81 on Day 13, and 716 on Day 16). On Day 5 of the estrous cycle, the expression of diacylglycerol O-acyltransferase 2 (DGAT2), TCDD-inducible poly(ADP-ribose) polymerase (TIPARP), dickkopf homologue 1 (DKK1), and aldo-keto reductase family 1, member C1-like (LOC782061) were altered to the greatest extent (Fig. 2 and Supplemental Table S1A available at www.biolreprod.org). On Day 7, elevated P4 altered the expression of 910 genes in total (Fig. 2 and Supplemental Table S1B) with the expression of cytochrome P450, family 26, subfamily A, polypeptide 1 (also called cytochrome P450 retinoic acid [RA] hydroxylase) (CYP26A1), DGAT2, and asialoglycoprotein receptor 2 (ASGR2) increased, and fatty acid binding protein 3, muscle and heart (mammary-derived growth inhibitor) (FABP3), protein regulator of cytokenesis 1 (PRC1), and solute carrier family 2 (facilitated glucose/fructose transporter), member 5 (SLC2A5) decreased to the greatest degree.

Fig. 1

Principal component analysis (PCA) plot indicating the source of greatest variation in the overall transcriptional profile of all 40 cyclic endometrial samples. Samples recovered from cyclic endometria on Days 5 and 7 cluster together, and samples recovered from endometria on Days 13 and 16 of the estrous cycle cluster together, indicating the source of greatest variation in endometrial gene expression is time.

Fig. 1

Principal component analysis (PCA) plot indicating the source of greatest variation in the overall transcriptional profile of all 40 cyclic endometrial samples. Samples recovered from cyclic endometria on Days 5 and 7 cluster together, and samples recovered from endometria on Days 13 and 16 of the estrous cycle cluster together, indicating the source of greatest variation in endometrial gene expression is time.

Fig. 2

Numbers of DEGs that were temporally regulated in the endometrium of cyclic beef heifers with normal (CN) or high (CH) concentrations of circulating P4. Numbers indicate the number of different genes for each comparison. Venn diagrams indicate the number of common and unique genes that change across time between treatments.

Fig. 2

Numbers of DEGs that were temporally regulated in the endometrium of cyclic beef heifers with normal (CN) or high (CH) concentrations of circulating P4. Numbers indicate the number of different genes for each comparison. Venn diagrams indicate the number of common and unique genes that change across time between treatments.

In contrast to the low number of differences in gene expression on a given day, there were a large number of genes temporally regulated (i.e., changed between successive days of the estrous cycle). In particular, the longer the time interval between days, the greater the number of DEGs (e.g., for the high P4 group, Day 5 vs. Day 7: 3652 probe sets; Day 7 vs. Day 13: 6411 probe sets [Fig. 2]). Between Day 5 and Day 7 of the estrous cycle, heifers with normal P4 exhibited negligible differences in endometrial gene expression (seven DEGs) compared with those with elevated P4 in which the expression of 3654 genes was altered (Supplemental Table S2). None of the DEGs between Days 5 and 7 in normal P4 heifers were represented in any of the GO SLIM categories, and no IPA canonical pathways or interaction analysis were determined. In contrast, the genes that were altered by high P4 between Days 5 and 7 of the estrous cycle were represented in all GO SLIM categories, including those involved in transport, protein transport, and transcription (Fig. 3A). IPA analysis revealed that the genes that were altered due to high P4 between Days 5 and 7 of the estrous cycle were associated with the top five canonical pathways of mitotic roles of polo-like kinase (7 DEGs in this pathway from a total of 62; cell cycle: G2/M DNA damage checkpoint regulation [4/44], arginine and proline metabolisim [5/183], ATM signaling [4/53], and urea cycle and metabolism of amino groups [3/80]). IPA network interaction analysis revealed five interaction networks, the first of which involves cell cycle, cancer, and genetic disorder genes. This interaction network had only four genes the expression of which decreased due to high P4 concentrations on Day 7 compared with Day 5 of the estrous cycle (CYP26A1, CPN1, SQSTM1, and GRB7), with all other DEGs increased in expression (Fig. 4). Interaction networks two, three, four, and five involved molecules associated with: 1) cell cycle, cellular assembly and organization, DNA replication, recombination, and repair, 2) cell cycle, cellular movement, and cancer, 3) cell cycle, cellular assembly and organization, DNA replication, recombination, and repair, as well as 4) DNA replication, recombination, and repair, cell cycle, and cellular assembly and organization (Supplemental Fig. S1, A–D).

Fig. 3

Bar graph representing GO slim terms and numbers of DEGs associated with each term between Days 5 and 7 (A), Days 7 and 13 (B), and Days 13 and 16 (C) in the endometrium of cyclic beef heifers with normal (CN, black bars) or high (CH, open bars) concentrations of circulating P4. GO slim terms are only included when five or more genes are associated with the term in at least one category.

Fig. 3

Bar graph representing GO slim terms and numbers of DEGs associated with each term between Days 5 and 7 (A), Days 7 and 13 (B), and Days 13 and 16 (C) in the endometrium of cyclic beef heifers with normal (CN, black bars) or high (CH, open bars) concentrations of circulating P4. GO slim terms are only included when five or more genes are associated with the term in at least one category.

Fig. 4

Graphical representation of an ingenuity pathway interaction network analysis generated from the DEGs between Days 5 and 7 of the estrous cycle in heifers with high P4. These DEGs were involved in cell cycle, cancer, and genetic disorder are displayed as colored shapes. The gray nodes are indicative of protein complexes and/or nonspecified members of a gene family that are important components of the interaction network, but are not DEGs. The network displays nodes (genes/gene products) and edges (the biological relationship between nodes). The color intensity of the nodes indicates the fold change (increase [red] or decrease [green]) associated with a particular gene on Day 7 compared with Day 5. A solid line indicates a direct interaction between nodes (genes/gene products), and a dashed line indicates an indirect relationship between nodes. The shape of the node is indicative of its function (see Supplemental Key, available online).

Fig. 4

Graphical representation of an ingenuity pathway interaction network analysis generated from the DEGs between Days 5 and 7 of the estrous cycle in heifers with high P4. These DEGs were involved in cell cycle, cancer, and genetic disorder are displayed as colored shapes. The gray nodes are indicative of protein complexes and/or nonspecified members of a gene family that are important components of the interaction network, but are not DEGs. The network displays nodes (genes/gene products) and edges (the biological relationship between nodes). The color intensity of the nodes indicates the fold change (increase [red] or decrease [green]) associated with a particular gene on Day 7 compared with Day 5. A solid line indicates a direct interaction between nodes (genes/gene products), and a dashed line indicates an indirect relationship between nodes. The shape of the node is indicative of its function (see Supplemental Key, available online).

Although there were considerable differences in gene expression profiles of heifers between Days 7 and 13 of the estrous cycle, P4 supplementation considerably increased this number compared with that in normal P4 heifers (6411 vs. 3969, respectively). There were more genes represented in each of the GO SLIM categories, with the exception of cell cycle, protein modification, and biosynthetic process (Supplemental Table S3; Fig. 3B). IPA analysis revealed the top five canonical pathways associated with these DEGs in heifers with normal P4 concentrations were neuregulin signaling (5/103), atherosclerosis signaling (5/113), role of macrophages, fibroblasts, and endothelial cells in rheumatoid arthritis (8/357), thrombopoietin signaling (3/59), and HGF signaling (4/103). The network interaction analysis revealed that the five networks contained genes, the functions of which were involved in: 1) lipid metabolism, molecular transport, and small molecule biochemistry, 2) cell cycle, embryonic development, and renal and urological system development and function, 3) carbohydrate metabolism, embryonic development, and organ development, 4) cancer, gastrointestinal disease, and cellular development, and 5) cellular development, cellular growth and proliferation, and connective tissue development and function (Supplemental Fig. S2, A–E).

In contrast, the top five canonical pathways associated with the DEGs due to high P4 between Days 7 and 13 of the estrous cycle were involved in arachidonic acid metabolism (9/224), glycosphingolipid biosynthesis-lactoseries (3/27), Fcg receptor-mediated phagocytosis in macrophages and monocytes (6/101), glycerophospholipid metabolism (7/192), and endothelin-1 signaling (8/185). Interaction analysis of these DEGs revealed networks the genes of which function in 1) cellular movement, carbohydrate metabolism, and cell-to-cell signaling and interaction, 2) lipid metabolism, small molecule biochemistry, and vitamin and mineral metabolism, 3) cell cycle, cellular development, and tumor morphology, 4) renal tubule injury, genetic disorder, and hematological system development and function, as well as 5) carbohydrate metabolism, small molecule biochemistry, and cell death (Supplemental Fig. S3, A–E).

Finally, analysis of the differences between Days 13 and 16 of the luteal phase of the estrous cycle revealed few DEGs in heifers with normal (160 DEGs) or high P4 (85 DEGs) (Fig. 3C; Supplemental Table S3). In heifers with normal P4, only three DEGs were associated with the canonical pathways glycine, serine, and threonine metabolism and TR/RXR activation. Only two DEGs were associated with the pathways of nitrogen metabolism, arginine and proline metabolisim, as well as glycerolipid metabolism. Only the top three interaction networks had more than 10 DEGs associated with the network (lipid metabolism, small molecule biochemistry, molecular transport, tissue development, cellular growth and proliferation, lymphoid tissue structure and development and lipid metabolism, molecular transport, and small molecule biochemistry; Supplemental Fig. S4, A–C). Interestingly, there were no canonical pathways or interaction networks associated with the DEGs from heifers with high P4 concentrations between Days 13 and 16 of the estrous cycle.

Experiment 2

Repeated administration of PGF2α to heifers resulted in significantly lower concentrations of serum P4 compared with untreated control heifers on all days studied (P > 0.05; Fig. 5).

Fig. 5

P4 profiles for control heifers with normal circulating P4 (CN) and those induced to have low P4 by repeated administration of PGF2α (CL) (Experiment 2).

Fig. 5

P4 profiles for control heifers with normal circulating P4 (CN) and those induced to have low P4 by repeated administration of PGF2α (CL) (Experiment 2).

The expression of ASGR2, DKK1, and hepatitis A virus cellular receptor 1 (HAVCR1 or MGC137099) was significantly lower on Day 7, and SLC2A5 expression was significantly higher in heifers with low P4 compared with controls (P > 0.05; Fig. 6A). In contrast, on Day 13, ASGR2 and DGAT2 expression increased in the low-P4 group, and FABP3 expression was significantly decreased compared with controls (P > 0.05; Fig. 6B). Low P4 did not affect the expression of CYP26A1 or PRC1 on either day examined (P > 0.05). Given that DGAT2 and FABP3 expression was delayed in low-P4 heifers, expression of genes involved in their respective pathways was characterized.

Fig. 6

Q-RT-PCR analysis of eight candidate genes identified from Experiment 1 in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

Fig. 6

Q-RT-PCR analysis of eight candidate genes identified from Experiment 1 in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

Analysis of genes involved in the FABP3 pathway indicated that peroxisome proliferator-activated receptor alpha (PPARA), PPAR delta (PPARD), retinoid X receptor alpha (RXRA), and RXR beta (RXRB) displayed similar expression patterns, with lower expression on Day 7 in low-P4 heifers compared with controls (P > 0.05; Fig. 7A). Only PPARA and PPARD were significantly altered on Day 13 by low P4, with an increase in expression compared with controls (P > 0.05; Fig. 7B). RXR gamma (RXRG) was not detectable in the endometrium, and PPAR gamma (PPARG) was not affected by the P4 status of the heifers (P > 0.05).

Fig. 7

Q-RT-PCR analysis of seven candidate genes involved in the DGAT2 and FABP3 pathways in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

Fig. 7

Q-RT-PCR analysis of seven candidate genes involved in the DGAT2 and FABP3 pathways in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

For the genes involved in the DGAT2 pathway, 1-acylglycerol-3-phosphate O-acyltransferase 1 (lysophosphatidic acid acyltransferase alpha) (AGPAT1) and 1-acylglycerol-3-phosphate O-acyltransferase 5 (lysophosphatidic acid acyltransferase epsilon) (AGPAT5) expression was higher on Day 7 in low-P4 heifers compared with controls, but, on Day 13, only AGPAT1 expression was significantly higher compared with controls (P > 0.05; Fig. 7, A and B, respectively). There was no significant effect of P4 concentration on the expression of glycerol-3-phosphate acyltransferase, mitochondrial (GPAT), diacylglycerol kinase alpha 80 kDa (DGKA), monoacylglycerol O-acyltransferase 1 (MOGAT1), monoacylglycerol O-acyltransferase 2 (MOGAT2), 1-acylglycerol-3-phosphate O-acyltransferase 3 (AGPAT3), 1-acylglycerol-3-phosphate O-acyltransferase 4 (lysophosphatidic acid acyltransferase delta) (AGPAT4) or 1-acylglycerol-3-phosphate O-acyltransferase 6 (lysophosphatidic acid acyltransferase zeta) (AGPAT6) members of the DGAT2 pathway (P > 0.05; data not shown).

To identify if low P4 concentrations affected the expression of PGR and genes, whose expression is altered in association with the loss of the PGR from the LE and GE [18], Q-RT-PCR was performed and the localization of the PGR protein was analyzed by immunohistochemistry. PGR mRNA was more abundant (P < 0.05) in heifers with low P4 on Day 7 compared with controls, but was not different on Day 13 (Fig. 8, A and B). Expression of alanyl (membrane) aminopeptidase (ANPEP) and lipoprotein lipase (LPL) was lower (P > 0.05) on Day 7 in heifers with low P4 compared with controls (Fig. 8, A and B). In contrast, by Day 13, ANPEP and LPL expression increased, while connective tissue growth factor (CTGF) expression decreased (P > 0.05) in the low-P4 group compared with controls.

Fig. 8

Q-RT-PCR analysis of five candidate genes, the expression of which has been shown previously to be important in an optimal uterine environment [18, 19], in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

Fig. 8

Q-RT-PCR analysis of five candidate genes, the expression of which has been shown previously to be important in an optimal uterine environment [18, 19], in the endometrium of cyclic beef heifers with normal (CN, black bars) or low (CL, open bars) concentrations of circulating P4 on Day 7 (A) and Day 13 (B) of the estrous cycle. *Significant difference (P > 0.05).

In heifers with normal and low P4 concentrations, both isoforms of the PGR localized to the LE, SG, DG, STR, and MYO on Day 7 of the estrous cycle, with a complete loss of PGR in the LE and SG on Day 13 in both groups (Fig. 9A). Analysis of the intensity of staining revealed significantly more PGR in the LE and SG of heifers with low P4 on Day 7 compared with their normal P4 counterparts (Fig. 9B), with no effect on PGR staining intensity in the MYO (data not shown), indicating a delay in the normal loss of this receptor from these tissues.

Fig. 9

A) Representative images depicting the localization of PGR-AB protein in the LE, SGs, and STR of the bovine endometrium on Days 7 and 13 of the estrous cycle of heifers with normal (CN) or low (CL) P4. Original magnification ×10. B) Graphical representation of the intensity of PGR-AB protein isoform localized within the LE or SG in uterine cross sections of cyclic heifers with normal (CN, black bars) or low (CL, open bars) P4 on Days 7 and 13 of the estrous cycle. *Significant difference in protein intensities between P4 treatments (P < 0.05).

Fig. 9

A) Representative images depicting the localization of PGR-AB protein in the LE, SGs, and STR of the bovine endometrium on Days 7 and 13 of the estrous cycle of heifers with normal (CN) or low (CL) P4. Original magnification ×10. B) Graphical representation of the intensity of PGR-AB protein isoform localized within the LE or SG in uterine cross sections of cyclic heifers with normal (CN, black bars) or low (CL, open bars) P4 on Days 7 and 13 of the estrous cycle. *Significant difference in protein intensities between P4 treatments (P < 0.05).

Experiment 3

Recipient heifers in the low-P4 group had lower (P > 0.05) daily P4 concentrations from Day 5 until slaughter compared with concentrations of P4 in control recipients (Fig. 10A). Moreover, low-P4 recipients displayed reduced blood flow around the CL on Day 6 compared with control heifers (P > 0.05; Fig. 10B), but there was no significant effect on CL weight.

Fig. 10

A) Mean serum P4 concentrations ± SEM (ng/ml) in control heifers (n = 9) and those administered PGF2α on Days 3, 3.5, and 4 postestrus (n = 11) to induce low P4. Treatment with PGF2α resulted in a significant decrease in P4 concentrations in treated animals from Day 5 (24 h after last PGF2α injection) until slaughter on Day 14, indicated by an asterisk (*P > 0.05). B) Blood flow intensity and representative images through a CL on Day 6 of the estrous cycle in a heifer induced to have low P4 (left) compared with a control heifer (right), as measured by color Doppler ultrasound.

Fig. 10

A) Mean serum P4 concentrations ± SEM (ng/ml) in control heifers (n = 9) and those administered PGF2α on Days 3, 3.5, and 4 postestrus (n = 11) to induce low P4. Treatment with PGF2α resulted in a significant decrease in P4 concentrations in treated animals from Day 5 (24 h after last PGF2α injection) until slaughter on Day 14, indicated by an asterisk (*P > 0.05). B) Blood flow intensity and representative images through a CL on Day 6 of the estrous cycle in a heifer induced to have low P4 (left) compared with a control heifer (right), as measured by color Doppler ultrasound.

Recovery rate of conceptuses on Day 14 was higher from normal P4 heifers (9.6 ± 1.7) compared with the recovery rate from recipients with low P4 (5.4 ± 1.7), although this was not statistically significant (P > 0.09). Mean conceptus length and area were significantly decreased (P > 0.05) in conceptuses recovered from low-P4 recipients compared with control recipients (Fig. 11, A and B).

Fig. 11

Mean ± SEM Day-14 embryo dimensions (A) and representative images of Day-14 elongating embryos (B) following transfer of Day-7 blastocysts to synchronized recipient heifers with normal or low circulating P4 and recovery at slaughter on Day 14. Original magnification ×4. *Significant difference in embryos recovered from normal compared to low P4 recipients (P < 0.05).

Fig. 11

Mean ± SEM Day-14 embryo dimensions (A) and representative images of Day-14 elongating embryos (B) following transfer of Day-7 blastocysts to synchronized recipient heifers with normal or low circulating P4 and recovery at slaughter on Day 14. Original magnification ×4. *Significant difference in embryos recovered from normal compared to low P4 recipients (P < 0.05).

Discussion

In the current study, the hypothesis tested was that alterations in circulating P4 concentrations (to induce either high or low circulating P4) in heifers alters the normal sequential changes that occur in the endometrial transcriptome, leading to an altered uterine environment with consequences for normal conceptus development. The main findings were that: 1) only minor differences occurred in endometrial gene expression profiles of heifers when short time intervals were compared during the early luteal phase under normal concentrations of P4 (Day 5 vs. Day 7), but much larger differences were identified when endometria from more distant stages of the luteal phase (Day 7 vs. Day 13) were compared; 2) exogenous supplementation of P4 that elevated circulating concentrations from Day 3 to Day 8 markedly altered the expression of a large number of genes at all four stages of the luteal phase examined; 3) induction of low P4 concentrations in serum altered the normal temporal changes that occurred in the expression of genes in the endometrium, mainly by delaying their expression; 4) this delay in temporal changes in endometrial gene expression was correlated with a delay in the down-regulation of the PGR from the LE and GE; and 5) the altered endometrial gene expression induced by low P4 was associated with a reduced capacity of the uterus to support conceptus development after multiple embryo transfer on Day 7.

Insertion of a PRID on Day 3 after estrus elevated serum P4 concentrations by approximately 1–2 ng/ml (well within physiological ranges). This led to the altered expression of over 3500 genes between two time points relatively close together (Day 5 vs. Day 7). Of these P4-induced genes, over 1000 were also differentially expressed between Days 7 and 13 in animals with normal P4, indicating that elevated P4 advanced endometrial gene expression in cyclic heifers, consistent with our previous observations in pregnant heifers [10]. When gene expression patterns of endometria from given days were compared, relatively few genes were altered by elevated P4 on Day 5, but, by Day 7 (by which time serum P4 concentrations had been significantly elevated for 4 days [3]), a significant divergence in the pattern of gene expression in the endometrium was apparent. Moreover, some of the genes with the largest fold change differences (e.g., DGAT2 and FABP3) may contribute to the composition of the histotroph. Both DGAT2 and FABP3 are known to be involved in triglyceride synthesis and transport (discussed subsequently here in further detail). CYP26A1 is known to be involved in RA metabolism predominantly through eliminating RA [25]. The balance of RA concentration in the uterine environment has been well characterized, and the modulation of RA concentrations during pregnancy has been well established [2629]. ASGR2 encodes for a protein that, in conjunction with ASGR1, forms ASGR, which is involved in mediating the endocytosis of serum glycoproteins [30], and may well play a role in trafficking glycoproteins into the uterine lumen. The expression of members of the glucose and fructose transporters in the uterine endometrium of ruminants, and their increased expression associated with increased P4 and interferon-tau, has been recently documented [18, 31]. SLC2A5 is a facultative fructose transporter [32] known to be expressed in fast-cleaving embryos to enhance fructose uptake for nucleotide synthesis [33]. The down-regulation of SLC2A5 in the uterine endometrium may ensure that available fructose is being utilized by the embryo, and not the uterine endometrium.

The main control mechanism that ensures the sequential modulation of endometrial gene expression is the circulating concentration of P4 via the down-regulation of its own receptor from the LE and GE. Our previous studies identified that the expression pattern of a number of genes in pregnant heifers is temporally correlated with the early down-regulation of the PGR, and that elevated P4 alters the timing and duration of expression of these genes [10, 19]. In the present study, a similar panel of genes that were altered by high P4 in endometria of cyclic heifers was identified, and the question of how the expression pattern of these genes differed in heifers with low P4 was addressed (Table 2). In the low-P4 model, a number of the transcripts studied displayed a delay in their expression pattern (e.g., in heifers with normal or high P4 concentrations, the expression of a given gene was high on Day 7, whereas, in heifers with low P4, the expression of the same gene was very low or not detectable). From Experiment 1, the expression of DGAT2 and ASGR2 was high on Day 7 in heifers with high P4, and FABP3 expression was low on Day 7 in heifers with high P4 compared with controls. However, in Experiment 2, DGAT2 and ASGR2 expression was high and FABP3 expression was low on Day 13 in heifers with low P4 only. In a normal P4 environment; in other words, the expression pattern of these genes in a normal P4 environment on Day 7 was only observed on Day 13 of the estrous cycle in the endometria from heifers exposed to a low-P4 environment. The fact that not all genes identified in the microarray study displayed this same delay in expression in the low-P4 model indicates that either these genes were not directly regulated by P4 or their expression patterns were not critically dependent on sustained circulating concentrations of P4.

Despite the importance of P4 for the establishment and maintenance of pregnancy in mammals, paradoxically, endometrial epithelia cease expressing PGR before implantation in all mammals studied [34]. This loss of PGR is a key checkpoint allowing genes necessary for conceptus development to be switched on and for those genes that are not required, or impede conceptus development/implantation, to be switched off [17]. Analysis of genes previously identified as altered in association with the loss of PGR [18], and therefore important for providing an appropriate uterine environment, revealed that, in the low-P4 model, their expression levels were either low on the appropriate day of the estrous cycle (ANPEP and LPL on Day 7; CTGF on Day 13), or their down-regulation was not complete at the appropriate time (ANPEP and LPL on Day 13). ANPEP is a peptidase that cleaves neutral amino acids from peptides [35], and its expression is affected by P4 in both cattle and humans [35, 36]. ANPEP may be involved in preventing the cleavage of proteins secreted into the uterine lumen, or may be associated with allowing increased neutral amino acids into the uterine lumen, both of which would alter histotroph composition. LPL is involved in delivering triacyglycerol (TAG) to tissues [37], and may supply the uterine histotroph with TAG, an energy source that the embryo can use at this stage of development [38]. CTGF, a protein involved in cell proliferation, migration, and adhesion, was previously identified as up-regulated in the uterine endometrium during the luteal phase of the estrous cycle/early pregnancy in many species, and was detected in the lumen of the uterus [2, 3942]. Both elevated P4 and IFNT concentrations increase the expression of endometrial CTGF in cattle and sheep [2, 36], indicating that it may be involved in proliferation of the conceptus trophectoderm; thus, in animals in which high P4 was induced, CTGF-induced trophectoderm proliferation may contribute to advanced conceptus development. It was clear from the current data that low P4 concentrations caused a disruption in the temporal regulation of these genes previously identified as required for an optimum uterine environment, due, in part, to delayed down-regulation of the PGR from the LE on Day 7 of the estrous cycle in heifers with low P4. Exposure of the uterine endometrium to sustained concentrations of P4 induced the loss of PGR from the LE and GE, allowing the tightly regulated temporal changes in endometrial gene expression to occur [34, 43]. Our data suggest that, in heifers with low P4, the sustained concentrations of P4 required to begin inducing the down-regulation of the PGR from the LE and GE were not sufficient at Day 7, leading to elevated abundance of the PGR protein in low-P4 heifers. This would lead to a delay in the temporal changes of genes required to provide an optimal uterine environment (e.g., ANPEP, CTGF, LPL), independent of the presence of a developing embryo/conceptus. The loss of the PGR from the LE occurred by Day 13 in heifers with both normal and low P4. This was not surprising, given that, in this model of low P4, the postovulatory rise in P4, although delayed, does still occur.

Induction of low circulating concentrations of P4 by administration of PGF2α on Days 3, 3.5, and 4 led to a significant reduction in CL blood flow on Day 6, supporting the use of this model as an effective method to induce low P4 concentrations in heifers, while still maintaining an output of P4 from the CL that parallels normal physiological events. Furthermore, CL size at slaughter on Day 14 was not different between heifers with low and normal P4 (data not shown), supporting the previous hypothesis that the low P4 induced in this model was due to a change in vascularization of the CL, rather than CL size per se [16]. To prove that low P4 has a functional effect on the endometrium, thereby affecting conceptus development, in vitro-produced blastocysts were transferred to the uteri of recipient heifers with normal or low P4 using a multiple embryo transfer model previously validated with animals induced to have elevated P4 [44]. Twenty Day-7 blastocysts were transferred per recipient, and all animals were slaughtered on Day 14 to assess conceptus development. As indicated previously, Day 14 was chosen for slaughter, as it represents the initiation of elongation in cattle—a critical checkpoint in development. In addition, previous personal experience has shown that, in the case of multiple transfers of embryos, separating rapidly elongating conceptuses after Day 14 can be extremely difficult. Lastly, it is known that very little embryonic loss occurs after Day 14 in cattle [45], and, thus, survival to Day 14 represents a useful measure of developmental competence. Recovery rate on Day 14 was lower in the low-P4 heifers compared with controls, and the mean conceptus length and area were also significantly reduced in the presence of low P4, indicating a functional consequence for conceptus development.

The potential implications of circulating P4 on the composition of oviduct and uterine fluid are interesting. Hugentobler et al. [46] characterized the effects of changes in systemic P4 on amino acid, ion, and energy substrate composition of oviduct and uterine fluids on Days 3 and 6, respectively, of the estrous cycle. Infusion of P4 had no effect on oviduct fluid secretion rate; however, uterine fluid secretion rate was lowered. P4 increased uterine glucose, decreased oviduct sulphate and, to a lesser extent, oviduct sodium, but had no effect on any of the ions in the uterus. The most marked effect of P4 was on oviductal amino acid concentrations, with a 2-fold increase in glycine, while in the uterus, only valine was increased. These data provide evidence of P4 regulation of oviduct amino acid concentrations in cattle, and may partly explain the differences in embryo development observed in Experiment 3.

In conclusion, the present study provides clear evidence for a temporal change in the transcriptomic signature of the bovine endometrium, which is sensitive to the concentrations of circulating P4 in the first few days after estrus, and which can, under conditions of low P4, lead to a suboptimal uterine environment and a reduced ability to support conceptus elongation.

Acknowledgments

The authors acknowledge the technical contributions of Fiona Carter, Sandra Aungier, Charlotte Lewis, Alison Lee, and Lilian Okumu in the conduct of animal experiments, and Penny Furney and Shay McDonnell in the conduct of progesterone radioimmunoassays.

References

1
Garrett
JE
Geisert
RD
Zavy
MT
Morgan
GL
Evidence for maternal regulation of early conceptus growth and development in beef cattle.
J Reprod Fertil
 
1988
;
84
:
437
446
.
2
Satterfield
MC
Song
G
Kochan
KJ
Riggs
PK
Simmons
RM
Elsik
CG
Adelson
DL
Bazer
FW
Zhou
H
Spencer
TE
Discovery of candidate genes and pathways in the endometrium regulating ovine blastocyst growth and conceptus elongation.
Physiol Genomics
 
2009
;
39
:
85
99
.
3
Carter
F
Forde
N
Duffy
P
Wade
M
Fair
T
Crowe
MA
Evans
AC
Kenny
DA
Roche
JF
Lonergan
P
Effect of increasing progesterone concentration from Day 3 of pregnancy on subsequent embryo survival and development in beef heifers.
Reprod Fertil Dev
 
2008
;
20
:
368
375
.
4
Mann
GE
Lamming
GE
Relationship between maternal endocrine environment, early embryo development and inhibition of the luteolytic mechanism in cows.
Reproduction
 
2001
;
121
:
175
180
.
5
Mann
GE
Fray
MD
Lamming
GE
Effects of time of progesterone supplementation on embryo development and interferon-tau production in the cow.
Vet J
 
2006
;
171
:
500
503
.
6
Stronge
AJ
Sreenan
JM
Diskin
MG
Mee
JF
Kenny
DA
Morris
DG
Post-insemination milk progesterone concentration and embryo survival in dairy cows.
Theriogenology
 
2005
;
64
:
1212
1224
.
7
McNeill
RE
Diskin
MG
Sreenan
JM
Morris
DG
Associations between milk progesterone concentration on different days and with embryo survival during the early luteal phase in dairy cows.
Theriogenology
 
2006
;
65
:
1435
1441
.
8
Ashworth
CJ
Sales
DI
Wilmut
I
Evidence of an association between the survival of embryos and the periovulatory plasma progesterone concentration in the ewe.
J Reprod Fertil
 
1989
;
87
:
23
32
.
9
Thatcher
WW
Meyer
MD
Danet-Desnoyers
G
Maternal recognition of pregnancy.
J Reprod Fertil Suppl
 
1995
;
49
:
15
28
.
10
Forde
N
Carter
F
Fair
T
Crowe
MA
Evans
AC
Spencer
TE
Bazer
FW
McBride
R
Boland
MP
O'Gaora
P
Lonergan
P
Roche
JF
Progesterone-regulated changes in endometrial gene expression contribute to advanced conceptus development in cattle.
Biol Reprod
 
2009
;
81
:
784
794
.
11
Bazer
FW
Vallet
JL
Roberts
RM
Sharp
DC
Thatcher
WW
Role of conceptus secretory products in establishment of pregnancy.
J Reprod Fertil
 
1986
;
76
:
841
850
.
12
Forde
N
Carter
F
Fair
T
Crowe
MA
Evans
ACO
Spencer
TE
Bazer
FW
O'Gaora
P
McBride
R
Boland
MP
Lonergan
P
Roche
JF
Effect of pregnancy and progesterone on gene expression in the uterine endometrium of cattle. In: Abstracts of the 42nd Annual Meeting of the Society for the Study of Reproduction, July 18–22, 2009, Pittsburgh, Pennsylvania.
Biol Reprod
 
2009
;
Special Issue: Abstract 37.
13
Clemente
M
de La Fuente
J
Fair
T
Al Naib
A
Gutierrez-Adan
A
Roche
JF
Rizos
D
Lonergan
P
Progesterone and conceptus elongation in cattle: a direct effect on the embryo or an indirect effect via the endometrium?
Reproduction
 
2009
;
138
:
507
517
.
14
Diskin
MG
Morris
DG
Embryonic and early foetal losses in cattle and other ruminants.
Reprod Domest Anim
 
2008
;
43
(suppl 2):
260
267
.
15
Rizos
D
Carter
F
Besenfelder
U
Havlicek
V
Lonergan
P
Contribution of the female reproductive tract to low fertility in postpartum lactating dairy cows.
J Dairy Sci
 
2010
;
93
:
1022
1029
.
16
Beltman
ME
Roche
JF
Lonergan
P
Forde
N
Crowe
MA
Evaluation of models to induce low progesterone during the early luteal phase in cattle.
Theriogenology
 
2009
;
72
:
986
992
.
17
Bazer
FW
Wu
G
Spencer
TE
Johnson
GA
Burghardt
RC
Bayless
K
Novel pathways for implantation and establishment and maintenance of pregnancy in mammals.
Mol Hum Reprod
 
2010
;
16
:
135
152
.
18
Forde
N
Spencer
TE
Bazer
FW
Song
G
Roche
JF
Lonergan
P
Effect of pregnancy and progesterone concentration on expression of genes encoding for transporters or secreted proteins in the bovine endometrium.
Physiol Genomics
 
2010
;
41
:
53
62
.
19
Okumu
LA
Forde
N
Fahey
AG
Fitzpatrick
E
Roche
JF
Crowe
MA
Lonergan
P
The effect of elevated progesterone and pregnancy status on mRNA expression and localisation of progesterone and oestrogen receptors in the bovine uterus.
Reproduction
 
2010
;
140
:
143
153
.
20
Gentleman
RC
Carey
VJ
Bates
DM
Bolstad
B
Dettling
M
Dudoit
S
Ellis
B
Gautier
L
Ge
Y
Gentry
J
Hornik
K
Hothorn
T
et al.  
Bioconductor: open software development for computational biology and bioinformatics.
Genome Biol
 
2004
;
5
:
R80
.
21
Fellenberg
K
Hauser
NC
Brors
B
Neutzner
A
Hoheisel
JD
Vingron
M
Correspondence analysis applied to microarray data.
Proc Natl Acad Sci U S A
 
2001
;
98
:
10781
10786
.
22
Smyth
GK
Limma: linear models for microarray data.
In:
Gentleman
R. VC
Dudoit
S
Irizarry
R
Huber
W
(Eds.),
Bioinformatics and Computational Biology Solutions Using R and Bioconductor
 .
New York:
Springer;
2005
:
397
420
.
23
Falcon
S
Gentleman
R
Using GOstats to test gene lists for GO term association.
Bioinformatics
 
2007
;
23
:
257
258
.
24
Snedecor
GW
Cochran
WG
Statistical Methods
 .
Ames, IA:
Iowa State University Press;
1989
.
25
Lutz
JD
Dixit
V
Yeung
CK
Dickmann
LJ
Zelter
A
Thatcher
JE
Nelson
WL
Isoherranen
N
Expression and functional characterization of cytochrome P450 26A1, a retinoic acid hydroxylase.
Biochem Pharmacol
 
2009
;
77
:
258
268
.
26
McNeill
RE
Sreenan
JM
Diskin
MG
Cairns
MT
Fitzpatrick
R
Smith
TJ
Morris
DG
Effect of systemic progesterone concentration on the expression of progesterone-responsive genes in the bovine endometrium during the early luteal phase.
Reprod Fertil Dev
 
2006
;
18
:
573
583
.
27
Liu
KH
Godkin
JD
Characterization and immunolocalization of bovine uterine retinol-binding protein.
Biol Reprod
 
1992
;
47
:
1099
1104
.
28
Harney
JP
Ott
TL
Geisert
RD
Bazer
FW
Retinol-binding protein gene expression in cyclic and pregnant endometrium of pigs, sheep, and cattle.
Biol Reprod
 
1993
;
49
:
1066
1073
.
29
MacKenzie
SH
Roberts
MP
Liu
KH
Dore
JJ
Godkin
JD
Bovine endometrial retinol-binding protein secretion, messenger ribonucleic acid expression, and cellular localization during the estrous cycle and early pregnancy.
Biol Reprod
 
1997
;
57
:
1445
1450
.
30
Stockert
RJ
The asialoglycoprotein receptor: relationships between structure, function, and expression.
Physiol Rev
 
1995
;
75
:
591
609
.
31
Gao
H
Wu
G
Spencer
TE
Johnson
GA
Bazer
FW
Select nutrients in the ovine uterine lumen. II. Glucose transporters in the uterus and peri-implantation conceptuses.
Biol Reprod
 
2009
;
80
:
94
104
.
32
Corpe
CP
Bovelander
FJ
Munoz
CM
Hoekstra
JH
Simpson
IA
Kwon
O
Levine
M
Burant
CF
Cloning and functional characterization of the mouse fructose transporter, GLUT5.
Biochim Biophys Acta
 
2002
;
1576
:
191
197
.
33
Gutierrez-Adan
A
Rizos
D
Fair
T
Moreira
PN
Pintado
B
de la Fuente
J
Boland
MP
Lonergan
P
Effect of speed of development on mRNA expression pattern in early bovine embryos cultured in vivo or in vitro.
Mol Reprod Dev
 
2004
;
68
:
441
448
.
34
Bazer
FW
Wu
G
Spencer
TE
Johnson
GA
Burghardt
RC
Bayless
K
Novel pathways for implantation and establishment and maintenance of pregnancy in mammals.
Mol Hum Reprod
 
2010
;
16
:
135
152
.
35
Sjostrom
H
Noren
O
Olsen
J
Structure and function of aminopeptidase N.
Adv Exp Med Biol
 
2000
;
477
:
25
34
.
36
Forde
N
Spencer
TE
Bazer
FW
Song
G
Roche
JF
Lonergan
P
Effect of pregnancy and progesterone concentration on expression of genes encoding for transporters or secreted proteins in the bovine endometrium.
Physiol Genomics
 
2009
;
41
:
53
62
.
37
Mead
JR
Irvine
SA
Ramji
DP
Lipoprotein lipase: structure, function, regulation, and role in disease.
J Mol Med
 
2002
;
80
:
753
769
.
38
Ferguson
EM
Leese
HJ
A potential role for triglyceride as an energy source during bovine oocyte maturation and early embryo development.
Mol Reprod Dev
 
2006
;
73
:
1195
1201
.
39
Brigstock
DR
Steffen
CL
Kim
GY
Vegunta
RK
Diehl
JR
Harding
PA
Purification and characterization of novel heparin-binding growth factors in uterine secretory fluids: identification as heparin-regulated Mr 10,000 forms of connective tissue growth factor.
J Biol Chem
 
1997
;
272
:
20275
20282
.
40
Klein
C
Bauersachs
S
Ulbrich
SE
Einspanier
R
Meyer
HH
Schmidt
SE
Reichenbach
HD
Vermehren
M
Sinowatz
F
Blum
H
Wolf
E
Monozygotic twin model reveals novel embryo-induced transcriptome changes of bovine endometrium in the preattachment period.
Biol Reprod
 
2006
;
74
:
253
264
.
41
Surveyor
GA
Wilson
AK
Brigstock
DR
Localization of connective tissue growth factor during the period of embryo implantation in the mouse.
Biol Reprod
 
1998
;
59
:
1207
1213
.
42
Uzumcu
M
Homsi
MF
Ball
DK
Coskun
S
Jaroudi
K
Hollanders
JM
Brigstock
DR
Localization of connective tissue growth factor in human uterine tissues.
Mol Hum Reprod
 
2000
;
6
:
1093
1098
.
43
Spencer
TE
Sandra
O
Wolf
E
Genes involved in conceptus-endometrial interactions in ruminants: insights from reductionism and thoughts on holistic approaches.
Reproduction
 
2008
;
135
:
165
179
.
44
Clemente
M
de La Fuente
J
Fair
T
Al Naib
A
Gutierrez-Adan
A
Roche
J
Rizos
D
Lonergan
P
Progesterone and conceptus elongation in cattle: a direct effect on the embryo or an indirect effect via the endometrium?
Reproduction
 
2009
;
138
:
507
517
.
45
Dunne
LD
Diskin
MG
Sreenan
JM
Embryo and foetal loss in beef heifers between Day 14 of gestation and full term.
Anim Reprod Sci
 
2000
;
58
:
39
44
.
46
Hugentobler
SA
Sreenan
JM
Humpherson
PG
Leese
HJ
Diskin
MG
Morris
DG
Effects of changes in the concentration of systemic progesterone on ions, amino acids and energy substrates in cattle oviduct and uterine fluid and blood.
Reprod Fertil Dev
 ;
22
:
684
694
.

Author notes

1
Supported by Science Foundation Ireland grants 06/INI/B62 and 07/SRC/B1156 (the opinions, findings, and conclusions or recommendations expressed in this article are those of the authors, and do not necessarily reflect the views of the Science Foundation Ireland).
3
These authors contributed equally to this work.