| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Administration on Endometrium of Nonpregnant Ewes: A Comparison with Pregnant Ewes
Departments of Animal Sciences (Y.C., J.A.G., E.A., A.M.W., R.M.R.), Biochemistry (R.M.R.), Statistics (L.B.H.), and Biomedical Sciences (C.S.R.), University of Missouri, Columbia, Missouri 65211; Department of Animal Sciences (A.D.E.), University of Florida, Gainesville, Florida 32611; Monsanto Company (N.P.), St. Louis, Missouri 63167; and Atwater-Merced Veterinary Clinics (M.P.A.), Merced, California 95348
Address all correspondence and requests for reprints to: R. M. Roberts, 240b Christopher S. Bond Life Sciences Center, 1201 East Rollins Road, University of Missouri-Columbia, Columbia, Missouri 65211. E-mail: robertsrm{at}missouri.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(IFNT) alters maternal physiology to accommodate a pregnancy. We hypothesized that the effectiveness of IFNT on extending corpus luteum (CL) life span in nonpregnant ewes would depend upon the dose and manner of administration and would be correlated with the response in gene expression in endometrium. We anticipated that IFNT, whether administered im or by uterine infusion, would mimic changes observed in pregnancy. Ewes were assigned to five treatments: 1) uterine infusion of saline; 2) uterine infusion of ovine IFNT4 (200 µg/d); 3) saline im injection; 4) im injection of IFNT4 at low dose (200 µg/d); and 5) high dose (2 mg/d). CL life span was increased in groups 2 and 5, but not in 1, 3, and 4. Endometrial RNA extracted from groups 1–5 on d 14 and from d 14 pregnant and nonbred (cyclic) ewes was used to assess expression of 70 genes on microarrays. When pregnant and cyclic ewes were compared, 30 genes were up-regulated and nine down-regulated during pregnancy. Responses were slightly less in groups 2 and 5 but were much lower in group 4. The majority of the highly up-regulated genes were associated with antiviral responses. Those down-regulated included ones for IGF-II, hypoxia-inducible factor 1
, oxytocin receptor, prostaglandin F synthase, and cyclooxygenase-2. Quantitative PCR for selected genes confirmed these data and revealed that similar gene expression changes occurred in the CL of pregnant and group 2 ewes. IFNT treatment mimics pregnancy, but relatively high doses of im-injected IFNT are required to elicit a full endometrial response. |
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
(PGF2) from the uterine endometrium during the late luteal phase of the estrous cycle (1, 2, 3). Similarly, a pregnancy will fail if the embryo fails to send a signal adequately robust to prevent the pulsatile PGF2 release. The signal in ruminants is a type I interferon (IFN), termed IFN- (IFNT), released by the conceptus, beginning at the blastocyst stage of development and continuing until the trophoblast is firmly attached to the wall of the uterus. IFNT is produced by trophectoderm (4, 5) and suppresses the pulsatile release of PGF2 from the endometrium (1, 2, 3). As a consequence of its effects, IFNT is generally regarded as the primary, although not necessarily the only, conceptus factor that rescues the CL and extends the functional life span of the CL.
Two different, but not mutually exclusive, mechanisms have been proposed to explain how IFNT affects prostaglandin metabolism. The first is a reduction of oxytocin receptor (OTR) number on the uterine epithelium as an outcome of the down-regulation of estrogen receptor (ER) by IFNT (6). The OTR down-regulation prevents the hormone, oxytocin, from mediating PGF2 release (7, 8, 9, 10). The second proposed mechanism is the down-regulation of enzymes involved in the synthesis of PGF2. Some investigators have reported an alteration in the prostaglandin E2 (PGE2): PGF2 ratio in favor of PGE2 during pregnancy (11, 12) and in response to IFNT in endometrial cells in bovine uterine endometrium primary cultures (13, 14, 15). Because PGE2 acts to maintain CL function (16, 17, 18), IFNT action under this mechanism could be regarded as both luteoprotective and luteotrophic. In primary endometrial cultures, IFNT has been reported to reduce the expression of PGE2-9-ketoreductase, an enzyme that catalyzes the production of PGE2 to PGF2 (19). At low concentrations, IFNT also down-regulates the cyclooxygenase-2 (COX-2) gene, which is responsible for controlling the synthesis of prostaglandins, in primary bovine endometrial cell cultures (20, 21) and a bovine endometrial cell line (22, 23). However, at high concentrations, IFNT has been reported to up-regulate COX-2 in endometrial explants (15), in primary bovine endometrial cells (13), and during early pregnancy (24).
Although there have been many papers showing that infusion of preparations containing IFNT (25, 26, 27, 28, 29, 30, 31, 32, 33) and other type I IFN (34, 35, 36) into the uterus can extend the functional life span of the CL in both sheep and cattle, there have been several other reports demonstrating that im injection of type I IFNs can also extend estrous cycle length, although usually not as effectively as intrauterine infusion (34, 35, 37, 38, 39). There has been particular interest in im administration because it could provide a simple means of improving pregnancy success by augmenting the IFNT produced from the conceptus at a time when embryonic loss is high (37, 38, 40). By contrast, one report suggests that sc injection of IFNT does not suppress expression of endometrial ER and OTR and that any extension of the estrous cycle that occurs from this route of administration might not be a reflection of events that normally occur during pregnancy (41). Accordingly, it remains unclear whether the ability of injected type I IFN to prolong CL life span in cattle and sheep is the result of the IFN acting on the endometrium directly or through unintentional side effects, e.g. by inducing fever (42) or by acting on the pituitary or ovary. There has been one early report, however, of the up-regulation of an unidentified secreted protein in the ovine uterus in response to im injected IFNT (43), suggesting that injected IFNT reaches its target.
In this paper, we have tested three hypotheses. The first is that the effectiveness of im injection on extending CL life span in sheep most likely depends upon the dose employed, acknowledging the likelihood that there will be far greater dilution of the IFN when it is injected into the body of the animal compared with it being infused directly into the uterine lumen. The second hypothesis is that the effectiveness of IFN administration will be correlated with the response it causes in the endometrium. Here, we followed the changes of approximately 70 genes considered as possible genes regulated by type I IFN (so called IFN-stimulated genes, ISGs) by using customized microarrays. Third, we hypothesized that the effects of IFNT on endometrium would mimic those observed in pregnancy.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Antiviral assay
Antiviral assays were carried out as described previously (32) on Madin-Darby bovine kidney cells (ATCC no. CCL22; American Type Culture Collection, Manassas, VA). Cells were exposed to 3-fold serial dilutions of ovIFNT4 in culture medium containing 10% fetal bovine serum. After 24 h, cells were challenged with vesicular stomatitis virus for 1 h in serum-free medium. The cells were cultured for an additional 20 h and were then stained with gentian violet. The laboratory standard used was a recombinant bovine IFNT1 preparation (5.4 x 107 IU/mg) that had been standardized against a human IFN reference reagent provided by the Antiviral Substance Program, National Institute of Allergy and Infectious Diseases, National Institutes of Health (Bethesda, MD).
Experimental design and treatments
Endometrial tissues from three pregnant and three cyclic (nonbred) ewes were collected at d 14 of the estrous cycle. Pregnancy was confirmed by the presence of a conceptus. After hysterectomy, representative regions of caruncular and intercaruncular zones of the endometrium were fixed in Bouin solution (Sigma, St. Louis, MO). The remaining endometrial tissues were frozen in liquid nitrogen and stored at –80 C for RNA extraction.
Cross-bred ewes from the University of Missouri Sheep Farm were synchronized by two im injections of a prostaglandin analog (Lutalyse; Pharmacia & Upjohn Co., Kalamazoo, MI) 9 d apart (see supplemental Table 1 on The Endocrine Society’s Journals Online web site at http://endo.endojournals.org). An intrauterine catheter was introduced into the uterine horn ipsilateral to the functional CL of randomly selected ewes as described previously (32, 33). Recombinant ovIFNT4 (specific antiviral activity 2 x 108 IU/mg) was infused twice daily in a sterile solution in 1x PBS (pH 7.2) containing 0.1% (wt:vol) ovine serum albumin (OSA). Controls received PBS containing 0.1% (wt:vol) OSA. Ewes were assigned randomly to receive the following treatments: 1) control for intrauterine treatment (PBS containing 0.1% [wt/vol] OSA) (n = 8); 2) intrauterine infusion of 200 µg/d ovIFNT4 (n = 8); 3) control im injection (PBS containing 0.1% [wt/vol] OSA) (n = 8); 4) im injection of 200 µg/d ovIFNT4 (n = 8); and 5) im injection of 2 mg/d ovIFNT4 (n = 8).
Uterine infusions and im injections of IFNT4 and control fluids were administered at 0700 and 1900 h from d 11–17 after estrus. Four ewes from each uterine infusion group and three ewes from each im injection experimental and control group were killed on the morning of d 14 after estrus 3 h after they had received the morning treatment. The uterine horn with the catheter was collected from each ewe. The CL were collected ipsilateral to the catheterized horn. Liver, hypothalamus, and pituitary gland were collected from one ewe in each treatment group. Endometrial tissues were collected from the mesometrial side. All tissues were quick frozen in liquid nitrogen and stored at –80 C for subsequent RNA extraction.
Blood samples were collected from the jugular vein at d 3, 7, and 10, and daily thereafter until ewes were killed or observed in estrus. Serum was stored at –20 C until analyzed for progesterone concentration. The experiment was terminated at d 42, despite two ewes not returning to estrus by that time point. The procedures used in this study were carried out under approved protocol 2745 of the University of Missouri Animal Care and Use Committee.
Statistical analysis for estrous cycle extensions study
CL life span was defined as the number of d from synchronized estrus to a decrease in serum progesterone to less than 0.5 ng/ml. The results are presented as mean ± SEM. Differences between IFNT treatment regimes were contrasted by using least-squares ANOVA after data had been ranked nonparametrically. Pair-wise comparisons were used to detect differences between individual treatment means (SAS software analysis; SAS Institute, Cary, NC) (44).
The bovine custom microarray
No ovine or bovine microarrays were available at the time this study was initiated in 2002. In addition, the database for ovine express sequence tags (ESTs) was limited. Accordingly, a custom bovine microarray was developed in our laboratory for this experiment. A total of 70 bovine ESTs that had sequence similarity with known human ISGs (45) and that were likely orthologs for the ovine genes were selected (see supplemental data). The majority of the ESTs used were obtained from MARC (USDA, US Meat Animal Research Center, Clay Center, NE). These ESTs were generated from cDNA libraries produced from endometrium, ovary, embryos, and other tissues (46). Another group of ESTs were obtained from a bovine mammary gland library generated at Monsanto Co. (St. Louis, MO). Others were selected from public databases.
The inserted cDNA fragments were PCR-amplified in 100-µl reactions containing 0.25 mM deoxynucleotide triphosphate and 0.5 µM M13 forward and reverse primers by using Klentaq (0.5 U) (AB Peptides, St. Louis, MO). PCR conditions included a 95 C initial denaturation for 5 min followed by 35 cycles of 95 C denaturation (30 sec), 60 C annealing (30 sec) and 72 C extension (2 min). A PCR product for each clone was verified by visualization on an agarose gel. PCR products were purified by means of a PCR purification kit (Promega, Madison, WI). The purified PCR products were dried in a CentriVap vacuum centrifugation system (Labconco Corp., Kansas City, MO) and resuspended in 3x standard saline citrate (SSC).
Bovine custom arrays were prepared by printing 70 PCR products of the ESTs. Exogenous spiking controls were used for assist data analysis (47). Briefly, Arabidopsis control cDNA plasmids [The Institute for Genomic Research (TIGR), Rockville, MD] were used to synthesize cRNA to spike into RNA samples. These control cRNAs were added to the Cy3 and Cy5 labeling reactions at equal concentrations. PCR amplicons from the Arabidopsis control cDNA plasmids (n = 10) were also printed on each array to provide a normalization control. Blank spots and 3x SSC were used as negative controls. Cot-1 DNA (Applied Genetics Laboratories, Melbourne, FL) and polyadenylic acid were used to detect nonspecific hybridization.
Gold Seal glass microscope slides (Fisher Scientific, Hampton, NH) were coated with 0.02% poly-L-lysine (Sigma, St. Louis, MO) in 0.05x PBS. A robotic microarray printer was used to spot the PCR products onto poly-lysine slides. The bovine custom arrays had a total of 768 spots and were organized in 16 blocks arrayed in four rows and four columns. Each block had 48 spots arranged in four rows and 12 columns. Each clone had four replicates on each slide.
A reference design was used in the presented study, in which all samples were compared with a common reference sample. Gene expression differences between samples were compared indirectly. The common reference was derived from untreated ewe liver RNA prepared in the lab. The microarray hybridization procedures have been described previously (48). Briefly, total RNA was isolated by using RNA STAT-60 (Tel-Test, Friendswood, TX). RNA (10 µg) was reverse-transcribed by using SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA). The cDNA from samples and reference were labeled separately with either Cyanine 3 (Cy3) or Cy5 fluorescence dye (Amersham Biosciences, Piscataway, NJ). The resulting labeled probes were purified, dried by CentriVap and quantitated by measuring the absorbance at 532 nm (Cy 3) and 635 nm (Cy 5), respectively. Equal amounts of each sample were resuspended in 15 µl hybridization solution [50% formamide, 5x SSC, 0.1% sodium dodecyl sulfate (SDS), 10 µg ovine Cot-1 DNA, and 10 µg polyadenylic acid in water]. Samples were then denatured at 95 C for 3 min and then applied to the microarrays under 22 x 20 mm Lifter Slips (Erie Scientific, Portsmouth, NH). After incubation at 42 C for 16 h, the arrays were sequentially washed with 1x SSC/0.2% SDS, 0.1x SSC/0.2% SDS and 0.1x SSC solutions. The arrays were dried by centrifugation at 1000 x g.
Hybridization images were immediately scanned by a GenePix 4000B laser scanner (Axon Instrument, Union City, CA). Hybridization intensities were analyzed with the GenePix Pro 4.0 computer software program. Hybridization of each sample to reference was repeated twice with a total of three hybridizations.
Hybridization intensities were normalized by using GenePix Pro 4.0 software (Axon Instrument). The external controls (plant genes) were used as normalization features (47). The result file for each hybridization was imported into Excel. Spots with poor quality and control spots were filtered out. The median values obtained from all three replicates after data filtering were analyzed by using the software known as Significance Analysis of Microarrays (SAM) (49). Analysis was done by using the two class-unpaired option and k-nearest neighbors method for missing values, with 300 permutations of the data. A clustering analysis of the differentially expressed genes was conducted by using the program Cluster and Tree View (50).
Quantitative real-time PCR analysis
The relative expression levels of five selected ovine genes were determined by real-time PCR in part to confirm the reproducibility of the microarray and also to provide more quantitative data for genes of particular interest. Primers specific for the genes for ubiquitin cross-reactive protein (UCRP), COX-2, OTR, IGF-II, and hypoxia-inducible factor 1, -subunit (HIF-1) were designed with Primer Express software (Applied Biosystems, Foster City, CA). The primer pairs are listed in supplemental Table 2. No particular attempt was made to design primers from different introns. The RNA, which had been treated with deoxyribonuclease I (Ambion, Austin, TX), was the same as that used in the microarray analysis. To test whether any intact DNA was present, we carried out RT-PCR analysis with primers for ß-actin that were in separate exons and whose amplification products would have crossed intron boundaries. No genomic DNA contamination was apparently present in any of the RNA samples.
Real-time PCR was carried out on an ABI prism 7500 sequence detection system (Applied Biosystems, Foster City, CA). SuperScript III reverse transcriptase (Invitrogen) and an oligo(deoxythymidylic acid) primer were used to transcribe the previously extracted total RNA into cDNA. Each real-time PCR contained 20 ng of synthesized cDNA, 2x SYBR Green Master Mix (QIAGEN, Valencia, CA), and a final concentration of 300 nM of forward and reverse primers for the candidate genes. Reactions were run for 40 cycles (95 C for 15 sec, 60 C for 1 min) after an initial 2 min at 50 C for enzyme activation and 10 min incubation at 95 C. A melting curve was run at the end to verify that only a single amplicon existed. The threshold cycle (CT), which indicates the relative abundance of a particular transcript, was calculated for each reaction by the ABI prism 7500 sequence detection system. Expression level of ovine ribosomal protein L19 (RPL19) was used as an endogenous control. In a preliminary study, the levels of several housekeeping genes [ribosomal protein L19 (PRL19), glyceraldehyde-3-phosphate dehydrogenase, and ß-actin] were measured in each sample by using specific primers (supplemental Table 2). RPL19 transcripts exhibited the least variation between tissues and between treatments, and were subsequently used as the endogenous controls. All transcript concentrations were normalized to that of ribosomal protein L19 mRNA expression. The relative quantitation of candidate gene expression in each treatment groups was determined by the comparative CT method (2–
CT) as described in the user bulletin no. 2 ABI Prism 7500 Sequence Detection System:
 |
RNA from individual ewes was isolated and analyzed separately to measure relative gene expression level in each IFNT treatment group. Each of the five transcripts was amplified in triplicate in a single PCR run. The entire experiment was then repeated with RNA from two additional ewes from the same treatment group to provide a total of three replicates, each carried out in triplicate for each treatment group. Real-time PCR quantification of gene expression level in each treatment was the mean of three real-time PCR experiments from three ewes, respectively. Differences in relative mRNA expression between experimental groups were assessed by one-way ANOVA, followed by pair-wise comparison by using least significant difference test (51). All experimental data are shown as the mean ± SEM. Values were considered significantly different at P < 0.05.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
on estrous cycle length in cyclic ewes infusion group because she had an anomalous progesterone profile. As anticipated from several previous studies (30, 32, 52), twice daily uterine infusions of 100 µg of IFNT (group 2; n = 3) extended estrous cycle length (32.7 ± 9.3 d), although there was considerable variation among the ewes, with two animals showing extensions of at least 42 d and the remaining ewe showing no apparent extension at all. The low-dose (200 µg/d) IFNT im injection (group 4; n = 5) did not influence average CL life span (17.2 ± 1.6 d) when compared with controls, although one ewe had an estrous cycle length of 23 d, suggesting that this dose might be marginally effective in some animals. By contrast, the high-dose (2 mg/d) im injection (group 5; n = 5) significantly extended cycle length to about 22 d (21.8 ± 1.9 d), with only one ewe failing to respond. Parametric ranking indicated no differences in cycle length between the controls (groups 1 and 3) and group 4 (low-dose injection group). Rank analysis indicated that groups 2 and 5 had a significant extension in estrous cycle length compared with controls (P < 0.06). The results suggest that im injections of IFNT can cause CL life span extension, although higher doses must be employed to be as effective as intrauterine infusion where the uterine endometrium becomes the immediate target for the IFNT.
Microarray analysis of ISGs expression in ovine uterine endometrium
Approximately half of the ewes in the experiment described above were killed at d 14 after previous estrus, 3 d after beginning the experimental and control treatments. Although there is no way of predicting which of these animals administered IFNT would have responded to IFNT, by showing increased estrous cycle length, the up-regulation of ISGs would be a clear indication of whether or not the tissue had been exposed to the IFN. RNA was also collected from a separate group of ewes at d14 of pregnancy, from which conceptuses were successfully flushed, and from nonbred (cyclic) ewes at d 14 of their estrous cycles.
Ovine endometrium RNA samples were analyzed for differences in expression of ISGs by using custom microarrays representing genes known to respond to type I IFN in human cells and other situations (supplemental Table 4). These custom microarrays, because of their low number and cost, allowed us to conduct multiple replicates and to compare at least three ewes in each experimental and control group, thereby providing a robust statistical treatment of the data. Importantly, cluster analysis separated the three experimental treatments (pregnant, P; IFNT infusion, group 2; high-dose IFNT injection, group 5) into three separate but closely related clusters (Fig. 1
). The other experimental group (low-dose IFNT injection) clustered as an outgrouping from the first three but was clearly distinct from the two control groups (infusion control and im injection control), which clustered tightly with the cyclic ewes. Importantly, all ewes appeared to respond to treatment with IFNT, whether by injection or intrauterine infusion, although there was considerable variation in the responses of individual genes within treatment groups (Fig. 1).
|
and 2, respectively. The complete data for all the genes on the microarray, plus annotation information, are provided in supplemental data. When pregnant and cyclic ewes were compared, 30 of the genes represented on the microarray were up-regulated during pregnancy (Table 1 and Fig. 1
), whereas another 11 showed a reduction in expression in terms of mRNA concentration (Table 2 and Fig. 1). Of the remaining 25 genes, none differed significantly in expression (FDR > 2%) between the pregnant and nonpregnant ewes (supplemental Table 4). The genes that demonstrated the largest differences in expression between the pregnant and nonpregnant state were those that encoded-major histocompatability complex (MHC) class I heavy chains (
26-fold), (2'-5') oligo A synthetase (23- to 25-fold, depending on the probe), UCRP (14-fold), 9–27 protein (11-fold), and complement component C1r (9-fold). All but the latter are classical ISGs, typically stimulated in response to either type I IFN or virus (53). Of the 30 genes up-regulated during pregnancy, 28 were also increased in response to infusion of IFNT into the uterine lumen (group 2 vs. group 1 infusion controls), including the top five stimulated genes listed above. In all but three instances, expression changes among these 28 genes were lower in the IFNT-infused group than in the pregnant animals. Injected IFNT at the high dose provided a similar response to the IFN uterine infusion, although in this case 21 rather than 28 genes were up-regulated significantly in response to treatment. The low-dose injection had, as anticipated, the smallest effect, with only 10 genes registering an increase in expression in uterine endometrium.
|
|
and Fig. 1) were also down-regulated in response to IFN infusion into the uterus. The response was again not as great in the ewes injected with high doses of IFNT (Table 3 and Fig. 1), with only four (PGF synthase, COX-2, IGF-II, and HIF-1) of the 11 down-regulated genes noted during pregnancy showing a significant reduction in expression. The low-dose injection again had a minimal effect. One interesting, possibly important, feature of these data was that four down-regulated genes (clathrin-like protein, ZnT-3, N-sulfoglucosamine sulfohydrolase, and tubulin-folding cofactor 3) showed responses only in the pregnant ewes (Table 2 and Fig. 1).
|
and 4 and Fig. 1). By contrast, the genes for ER- and ER-ß showed no significant change either in response to IFN infusion and im injection or in pregnant vs. cyclic animals.
|
were selected for real-time PCR analysis of their endometrial mRNA concentrations to validate the gene expression patterns evaluated on the microarray (Table 3). Concentration of each tested mRNA was very similar among the control samples (d 14 cyclic, group 1, and group 3) leading confidence to the analysis. UCRP and OTR were chosen because the former is well established as being highly up-regulated and the latter, although not included on the microarrays, down-regulated in response to IFNT in ovine endometrium. Both have been implicated in the overall process of maternal recognition of pregnancy in sheep (54, 55). As noted in the microarray analysis, UCRP was up-regulated during pregnancy and after all three IFNT treatments in the order pregnancy > IFNT infusion > high-dose injection > low-dose injection. As anticipated, OTR mRNA concentrations were strongly down-regulated in response to pregnancy and IFNT exposure, with the low-dose injection having the smallest effect.
COX-2 was included in the real-time PCR analysis because its response to IFNT in endometrial cells and tissues has been controversial (see introductory text). The data validated those observed in the microarrays. COX-2 expression was strongly down-regulated during pregnancy and in response to IFNT infusion but less so after IFNT injection (Tables 2 and 3). Consistent with the microarray data (Table 2), the low-dose injection provided no significant down-regulation of COX-2.
HIF-1 has been found to be responsive to IFN-ß in human cell line (45) and for this reason had been included on the microarray. In situ hybridization revealed that its expression, which, as far as we are aware, had not been studied previously in the sheep uterus, was confined largely to the surface and glandular epithelium and to the stromal tissue of caruncles (data not shown). Real-time PCR validated the microarray data and also indicated that the magnitude of its response to treatment was comparable with that of COX-2 (Table 3). In particular, HIF-1 was reduced in expression by approximately 90% in response to pregnancy and IFNT infusion.
Finally, the unexpected down-regulation of IGF-II expression in endometrium was confirmed in this study. As with the other genes, there was a remarkable correspondence between the magnitude of the change noted during pregnancy and that which occurred in response to IFNT infusion (Table 3). Again, IFNT injection at the high dose mimicked the outcome of IFNT infusion, although the magnitude of the response was less. Low-dose injection once more provided only a marginal response.
In situ hybridization analysis indicated that the mRNAs for HIF1- and IGF-II, like that for UCRP, were concentrated in the surface and upper glandular epithelium and to a lesser extent in the stromal cells of caruncles (data not shown), a distribution almost identical with that of that of the IFN receptor subunits (56).
Together these data provide confidence in the more extensive data set obtained with microarray experiments. They also confirm the reliability of the results on COX-2 mRNA down-regulation.
ISG expression in extrauterine tissues
Real-time PCR was used to assess the mRNA concentration changes that occurred in a number of nonuterine tissues in response to IFNT administration. The tissues selected for analysis were CL, liver, pituitary, and hypothalamus. Although the hypothalamus-pituitary axis serves as neural control center for gonadal hormones, these tissues were not anticipated to respond to any of the treatments because they should have been protected by the blood-brain barrier. The CL, on the other hand, has been reported to show elevation of some ISGs during pregnancy and administration of IFNT (41), and, along with the liver, would be expected to respond to injected IFN in a dose-dependent manner. These tissues were only recovered from groups 1–5 and not from the d-14 pregnant and cycling animals.
As predicted, neither pituitary nor hypothalamus responded to any of the treatments. None of the five selected genes showed any significant changes in expression (data not shown). The CL, by contrast, showed almost identical expression changes relative to controls as the uterine endometrium (Table 4). Importantly, the ranking in terms of gene expression change was IFNT infusion > high-dose injection > low-dose injection, suggesting that IFNT introduced into the uterine lumen was able to mediate molecular events in the CL with relative ease.
A liver sample was only collected from one animal of each group. Real-time PCR analysis of mRNA concentrations in the three IFN-treated animals, and the infusion control was performed to show whether gene expression in liver was affected by IFNT treatments (Table 5). UCRP was up-regulated by im injection of the higher amounts of IFNT, but not by intrauterine infusion, suggesting that IFNT does not readily escape the uterus and enter the systemic circulation. OTR was down-regulated in the ewe administered IFNT by im injection, but the three other genes that were responsive in the endometrium showed no response in the liver to any of the treatments.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Our second hypothesis was that the response of the ewes in terms of estrous cycle extension would be correlated with those noted in gene expression changes in the endometrial tissues of treated ewes. This hypothesis was also supported. Whereas injection of the high amounts of IFNT gave a response only slightly less than group 2 (uterine infusion group), injection of 200 µg IFNT had only a modest effect on gene expression, with many of the candidates chosen not showing any significant changes when assayed by microarray (Tables 1 and 2). Of course it is not possible in these experiments to predict which of the ewes in groups 2 and 5 might have gone into an extended period of pseudopregnancy, as they were killed at d 14. There was, however, noticeable variation among some of the normally up- or down-regulated genes to the treatments (Fig. 1). Whether any of these genes are indicators of ewes likely to respond to IFNT by showing estrous cycle extension remains to be determined.
Our third hypothesis was that responses to exogenous IFNT by either im injection or uterine infusion would mimic the changes in endometrial gene expression caused as a result of pregnancy. These similarities have already been remarked upon. Indeed, except for the four down-regulated genes discussed above, the patterns of regulation were quite similar among treatments, although they were always greatest in the pregnant ewes and lowest in the ewes receiving im injections. Together, these data suggest that high doses of injected IFNT exert effects on the nonpregnant uterus quite similar to those seen in pregnancy. On these grounds, therefore, the strategy of using injected IFNT to boost the endogenous IFNT signal from the conceptus to increase pregnancy success seems not unreasonable. On the other hand, the amount of IFNT injected would have to be high to be effective, and such levels can cause serious side effects in the animals (39, 42, 58, 61). In addition, some genes regulated during pregnancy appeared not to change expression after IFNT administration to nonpregnant ewes (Fig. 1), suggesting that there is an interaction between IFNT and some other component of the conceptus in eliciting an endometrial response. Another possibility is that the changes are independent of IFNT, depending instead on other factors from the trophoblast.
Our interesting observation was that IFNT infused into the uterine lumen altered the expression of IFN-responsive genes in the CL. There was already evidence that ISGs can become up-regulated at sites outside the endometrium. Mx protein, for example, is induced in the uterine epithelium, stroma, and myometrium (62, 63), and in peripheral blood mononuclear cells (64) during early pregnancy in sheep. Expression of Mx mRNA and UCRP mRNA is increased in CL in response to IFNT (41). Our experiments confirmed these earlier observations and showed that the responses in the CL were of comparable magnitude to those occurring in endometrium, a surprising outcome. How the IFNT placed in the uterine lumen reaches the CL is mysterious. The concentration of the protein in the uterine vein of pregnant ewes, as determined by the antiviral activity, has been reported to be slightly elevated (38), but this material would be directed away from the ovary and not toward it. In any case, IFNT is undetectable in blood collected from the jugular vein of pregnant ewes, whereas it remains elevated for several hours in ewes given im injections of IFNT (38). One explanation for the effects on the CL is that IFNT originating in the uterus reaches the ovary by a route that does not involve the maternal blood stream, e.g. through the lymphatic system. Alternatively, the effects may be indirect, e.g. through the movement of immune cells. The fact that there were changes in expression of IFNT-target genes in the CL as well as in the endometrium, raises the possibility that these responses in the CL could contribute to the antiluteolytic action of the IFNT.
The genes selected for these experiments were chosen largely because of the availability of the bovine ESTs for genes whose activity had previously been shown to be either up-regulated by IFNT (65) or regulated by either IFN- or IFN-ß in human cells (IFN stimulated genes or ISGs) (18, 45, 65) (Tables 1 and 3 and supplemental Table 4). A few others were included, such as the OTR, ER, and COX-2 because they had been implicated in the antiluteolytic activity of IFNT (6, 20). We discuss these further later in this Discussion. Several of the genes, such as those encoding Mx (62), 2',5'-oligoadenylate synthetase (66, 67), UCRP/ISG17 (UCRP/ IFN-stimulated gene 17) (54, 68), ß2-microglobulin (69), and IFN regulatory factor (IRF)-1 and -2 (70), have been demonstrated previously to be up-regulated by IFNT in endometrium, although no large-scale comparative study has been performed until now. Most of these genes are classic ISGs and are generally regarded as part of the early response to viral infection. Among some of the other genes selected for the microarray, a few are transcription factors known to control the downstream actions of the IFN (supplemental Table 4), whereas another, a MAPK, has been linked to IFN signaling (22), but is not part of the classical Janus kinase-signal transducer and activator of transcription signal (STAT) transduction pathway.
Another group of genes we included in our analysis have been reported to respond to type I IFN action in a large scale survey (45, 71), but there has been little or no follow up to the initial report and no confirmation of the response in nonhuman cells. The discovery that the genes for IGF-II and HIF-1 were significantly down-regulated by IFNT in endometrium and CL is of particular interest. The timing and cellular localization of IGF-II mRNA in human endometrium has suggested that this protein drives proliferation of endometrial stromal cells (72) and promotes decidualization and the migration of endothelial cells (73). HIF-1 also regulates angiogenesis and has been implicated in implantation in the mouse (74). The down-regulation of these two genes in sheep endometrium during the periimplantation period and in response to exogenous IFNT may reflect the need to avoid rather than to induce a decidual response because the sheep trophoblast is essentially noninvasive and does not penetrate the endometrial stromal tissues.
Two other genes significantly down-regulated by IFNT in endometrium were the ones encoding PGF synthase and COX-2, the inducible form of prostaglandin endoperoxidase H synthase (Tables 2 and 3). By contrast, the products of two other genes involved in prostaglandin metabolism, PGDH, and PGE2 synthase showed increased activity (Table 1). Logically, the anticipated outcome of these changes would be the increase production of PGE2 relative to PGF2. These data, on the surface, appear to be somewhat inconsistent with the observation that the amount of PGF2 released tends to increase as the pregnancy proceeds (75), even though the pulsatile pattern of PGF2 release into the utero-ovarian vein normally seen toward the end of an estrous cycle is down-modulated. However, detailed comparisons of events occurring on the morning of d 14 have not to our knowledge been made, and a reduction in the synthesis of PGF2 in both endometrium and CL at that time could have a significant role in protecting the CL from regression. That a decline in progesterone can precede the initiation of major PGF2 pulses is also well established (11) and indicates that at this stage of pregnancy the CL is poised to regress and is vulnerable to the onset of luteolytic events. Wiltbank and colleagues (76, 77), for example, have suggested that intraovarian as well as endometrial processes must exist to protect the CL of pregnancy.
Our observation that the COX-2 transcripts, which are localized predominantly in the surface and glandular epithelium of ovine endometrium (data not shown), are markedly reduced in amount in the d-14 pregnant ewe and by IFNT administration is controversial (78). COX-2 is regarded as the rate-limiting enzyme for PGF2 synthesis in endometrium, catalyzing the conversion of arachidonic acid to prostaglandin H2 (PGH2), the precursor of both PGE2 and PGF2. However, there are conflicting reports on the effects of IFNT on the expression of COX-2. For example, IFNT has been reported to increase both COX-2 mRNA (78) and protein expression (24, 79) in early pregnancy and in bovine endometrial cells (13), whereas others have noted a decline in COX-2 mRNA expression in response to IFNT in bovine endometrial cells (21, 23). In another study, Kim et al. (78) compared pregnant and cyclic ewes at d 10, 12, and 14 after previous estrus and assessed COX-2 expression levels by semiquantitative RT-PCR. Rather than showing a decreased amount of COX-2 mRNA in pregnant vs. nonpregnant endometrium, Kim et al. noted that expression was higher in the pregnant ewes. They also noted that intrauterine infusion of IFNT had no effect on COX-2 endometrial mRNA concentrations and appeared not to reduce the amount of COX-2 protein, as assessed by immunocytochemistry. We are unable to explain these differences. Our microarray analyses, which employed multiple animals, at least three microarrays for each animal comparison, four replicates of each gene on every microarray, and several housekeeping genes selected on the basis of showing no significant change among treatments, revealed a significant reduction in COX-2 transcripts in pregnant ewes and after either IFNT infusion or IFNT injection. These data for the ewes receiving IFNT treatments were confirmed by real-time PCR. By contrast, our results confirm that OTR mRNA is markedly down-regulated in response to IFNT (Table 4), although there was no change in the concentration of ER genes relative to controls either during pregnancy or after IFNT treatment (supplemental data). Because the ER and ERß transcripts were of relatively low abundance as gauged by the hybridization signals of the ovine transcripts to the bovine cDNAs on the array, the microarrays might not provide a sufficiently sensitive measure of transcript concentrations. However, others (10) failed to note regulation of ER by IFNT in bovine endometrium. Together, these results suggest that the mechanism controlling oxytocin receptor expression remains to be addressed further.
In summary, these results show that both uterine infusion and im injection of IFNT into nonpregnant ewes evoke many of the same endometrial responses as pregnancy, although on a lesser scale. The experiments clearly confirm that IFNT administration and the state of pregnancy leads to the marked down-regulation of both COX-2 and OTR gene expression at d 14 after previous estrus and demonstrate for the first time a reduced expression of two other genes (those for IGF-II and HIF-1) that have been implicated in inducing a decidual response in other species. The microarray data, however, provide only limited information, and further biochemical, genetic, and physiological approaches are needed to link any of the changes to reproduction.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
Gene symbols are in italics with all letters capitalized; the protein designations are the same as the gene symbol but not in italics.
We declare there is no conflict of interest that would prejudice the impartiality of this scientific work.
First Published Online February 9, 2006
Abbreviations: CL, Corpus luteum; COX-2, cyclooxygenase-2; CT, threshold cycle; ER, estrogen receptor; ESTs, express sequence tags; FDR, false discovery rate; GE, glandular epithelium; GST, glutathione-S-transferase; HIF-1, hypoxia-inducible factor 1, -subunit; IFN, interferon; IFNT, IFN-; IRF, IFN regulatory factor; ISGs, IFN-stimulated genes; MHC, major histocompatability complex; OSA, ovine serum albumin; OTR, oxytocin receptor; ovIFNT, ovine IFNT; PGDH, prostaglandin dehydrogenase; PGE2, prostaglandin E2; PGF2, prostaglandin F2; RPL19, ribosomal protein L19; SAM, Statistical Analysis of Microarray; SDS, sodium dodecyl sulfate; SSC, standard saline citrate; STAT, signal transducer and activator of transcription signal; UCRP, ubiquitin cross-reactive protein.
Received October 14, 2005.
Accepted for publication January 31, 2006.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
regulates expression of endometrial receptors for estrogen and oxytocin but not progesterone. Biol Reprod 53:732–745[Abstract]-Interferon: pregnancy recognition signal in ruminants. Proc Soc Exp Biol Med 213:215–229[CrossRef][Medline] production and uterine hormone receptors during early pregnancy. J Reprod Fertil Suppl 54:317–328[Medline] and 6-keto-prostaglandin F1 in the utero-ovarian venous plasma of nonpregnant and early pregnant ewes. Biol Reprod 30:936–944[Abstract] and PGE2 production in sheep. J Reprod Fertil 101:737–741 production in endometrial epithelial cells in vitro: role of cyclooxygenase-2. Endocrinology 138:4798–4805 in bovine endometrial cells from caruncular and inter-caruncular areas. Biol Reprod 59:241–247 increases PGE2 production and COX-2 gene expression in the bovine endometrium in vitro. Mol Cell Endocrinol 132:117–126[CrossRef][Medline] on prostaglandin biosynthesis, transport, and signaling at the time of maternal recognition of pregnancy in cattle: evidence of polycrine actions of prostaglandin E2. Endocrinology 145:5280–5293. Biol Reprod 62:125–131 in bovine endometrial cells. Endocrinology 139:2293–2299 in bovine endometrial cells. Biol Reprod 60:656–663 stimulates the Janus kinase-signal transducer and activator of transcription pathway in bovine endometrial epithelial cells. Biol Reprod 64:654–665 suppresses prostaglandin F2 secretion independently of the mitogen-activated protein kinase and nuclear factor 
B pathways. Biol Reprod 64:965–973. Biol Reprod 70:54–64 of cows in response to recombinant interferon-. J Dairy Sci 78:1921–1931[Abstract] release by oestradiol and oxytocin. J Reprod Fertil 76:425–433-interferon (trophoblastin) in sheep. J Endocrinol 127:R5–R8 does not affect biological activity. Biol Reprod 58:1463–1468 genes are not expressed identically and their protein products display different activities. Biol Reprod 58:566–573 in the periimplantation conceptus: sequence relationships and comparative biological activities. Biol Reprod 61:1592–1600 1 on luteal lifespan in cattle. J Dairy Sci 72:1859–1865 I at the time of maternal recognition of pregnancy inhibits prostaglandin F2 secretion and causes luteal maintenance in cyclic ewes. J Reprod Fertil 94:489–500 and - in sheep. Biol Reprod 73:1087–1093 I1 during the period of maternal recognition of pregnancy. J Anim Sci 68:2766–2770[Abstract] I1 on estrous cycle length and pregnancy success in sheep. J Reprod Fertil 91:347–356 with mutations near the carboxyl terminus on luteal life span in sheep. Biol Reprod 56:214–220[Abstract] and other cytokines in early pregnancy. Reprod Fertil Dev 9:355–380[CrossRef][Medline] on the endometrium of cyclic ewes. Biol Reprod 61:464–470 on body temperature and plasma progesterone concentrations in cyclic dairy cows. J Dairy Sci 78:1470–1476[Abstract] I1 in nonpregnant ewes. J Reprod Fertil 93:367–374, ß, or 
using oligonucleotide arrays. Proc Natl Acad Sci USA 95:15623–15628 induce conjugation of bovine ubiquitin cross-reactive protein to cytosolic uterine proteins. Biol Reprod 58:898–904 alters oestrogen receptor and oxytocin receptor expression in the endometrium of cyclic ewes. J Mol Endocrinol 15:203–220 by intrauterine and intramuscular administration of recombinant bovine interferon- to cows. J Reprod Fertil 93:375–384 on circulating concentrations of progesterone, cortisol, luteinizing hormone, and antiviral activity; interestrous interval; rectal temperature; and uterine response to oxytocin in cyclic ewes. Biol Reprod 57:621–629[Abstract] differentially induces lymphopenia and neutropenia in lambs. J Interferon Cytokine Res 18:731–737[Medline] in the uterus during early pregnancy. J Reprod Fertil Suppl 54:329–339[Medline] I1. J Reprod Fertil 93:599–607. Biol Reprod 58:1154–1162 independent of signal transducer and activator of transcription 1. Endocrinology 144:5203–5214 synthesis and release in the ewe during the initial establishment of pregnancy. J Reprod Fertil 83:527–536 regulates distinct physiological changes in early and mid-cycle bovine corpora lutea. Biol Reprod 58:346–352 on expression of cyclooxygenase two (COX-2) in ovine endometrium. Reprod Biol Endocrinol 1:58[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  T. Ezashi, P. Das, R. Gupta, A. Walker, and R. M. Roberts The Role of Homeobox Protein Distal-Less 3 and Its Interaction with ETS2 in Regulating Bovine Interferon-Tau Gene Expression-Synergistic Transcriptional Activation with ETS2 Biol Reprod, July 1, 2008; 79(1): 115 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F. Oliveira, L. E. Henkes, R. L. Ashley, S. H. Purcell, N. P. Smirnova, D. N. R. Veeramachaneni, R. V. Anthony, and T. R. Hansen Expression of Interferon (IFN)-Stimulated Genes in Extrauterine Tissues during Early Pregnancy in Sheep Is the Consequence of Endocrine IFN-{tau} Release from the Uterine Vein Endocrinology, March 1, 2008; 149(3): 1252 - 1259. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. E Spencer, O. Sandra, and E. Wolf Genes involved in conceptus-endometrial interactions in ruminants: insights from reductionism and thoughts on holistic approaches Reproduction, February 1, 2008; 135(2): 165 - 179. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Das, T. Ezashi, R. Gupta, and R. M. Roberts Combinatorial Roles of Protein Kinase A, Ets2, and 3',5'-Cyclic-Adenosine Monophosphate Response Element-Binding Protein-Binding Protein/p300 in the Transcriptional Control of Interferon-{tau} Expression in a Trophoblast Cell Line Mol. Endocrinol., February 1, 2008; 22(2): 331 - 343. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Kashiwagi, C. M. DiGirolamo, Y. Kanda, Y. Niikura, C. T. Esmon, T. R. Hansen, T. Shioda, and J. K. Pru The Postimplantation Embryo Differentially Regulates Endometrial Gene Expression and Decidualization Endocrinology, September 1, 2007; 148(9): 4173 - 4184. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Y. Chen, E. Antoniou, Z. Liu, L. B Hearne, and R M. Roberts A microarray analysis for genes regulated by interferon-{tau} in ovine luminal epithelial cells Reproduction, July 1, 2007; 134(1): 123 - 135. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
