PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
 
Reprod Sci. Author manuscript; available in PMC 2010 June 8.
Published in final edited form as:
PMCID: PMC2882188
NIHMSID: NIHMS92060

Endometrial Gene Expression in Early Pregnancy: Lessons From Human Ectopic Pregnancy

Ricardo F. Savaris, MD, PhD, MSc, Amy E. Hamilton, BS, Bruce A. Lessey, MD, PhD, and Linda C. Giudice, MD, PhD, MSc

Abstract

Human endometrium undergoes modifications in preparation for embryonic implantation. This study investigated in vivo the endocrine effects of pregnancy on the endometrium, using the model of ectopic pregnancy. Endometrial biopsies from 9 subjects with ectopic pregnancy (Preg) were compared with 8 and 6 samples of mid and late secretory endometrium, respectively. After hybridizing with Affymetrix HGU133 Plus 2 chips, data were analyzed using GeneSpring GX and Ingenuity Pathways Analysis. From 54 675 genes, 3021 genes were significantly differentiated when mid-secretory endometrium was compared with the Preg (Volcano plot; P < .05, ≥2-fold change). The complement and coagulation cascade, phospholid degradation, glycosphingolipid biosynthesis (globoseries), retinol metabolism, antigen presentation pathway, glycosphingolipid biosynthesis, and O-glycan biosynthesis were main significant canonical pathways found in Preg samples. Validation was done with reverse transcriptase polymerase chain reaction. In conclusion, the ectopic embryo has a significant impact, by an endocrine mechanism, on endometrium, when compared with the window of implantation.

Keywords: Microarray, ectopic pregnancy, transcriptome

Human endometrium is a steroid-responsive tissue that undergoes histologic, structural, biochemical, and programmed gene expression changes throughout the menstrual cycle in preparation for embryonic implantation.13

In the very early stages of implantation, embryo attachment to the endometrium and penetration into the endometrial stromal compartment for further growth and development likely occur primarily by paracrine signaling between cells of the conceptus and the decidua.4 As gestation proceeds, there are multiple interactions occurring simultaneously. For example, human chorionic gonadotropin (hCG) is secreted from the preimplantation embryo and likely serves paracrine functions at the time of nidation, as well as its known endocrine function of stimulating progesterone production by the corpus luteum for pregnancy maintenance. In addition, there are multiple placental products found in the maternal circulation early in gestation (eg, steroid hormones, placental growth hormone, adrenocorticotropic hormone [ACTH], human placental lactogen, dimeric inhibin, pregnancy-associated plasma protein-A [PAPP-A], and pregnancy-specific glycoproteins, among others5) that likely are endocrine signals from the conceptus to other tissues in the maternal host, including the endometrium.

We and others have defined the transcriptome of human endometrium, with a focus on the window of receptivity (mid-secretory endometrium, MSE) in nonconception cycles.69 The endometrial transcriptome, biological processes, and biochemical pathways beyond this time period, in the absence of pregnancy, reflect molecular preparation for tissue desquamation and are uninformative about changes in a conception cycle. In a conception cycle, extending our knowledge into the late secretory phase (week 4 and beyond), has been difficult to investigate because of ethical concerns and confounding contributions of the conceptus/placenta in whole tissue analyses. Thus, alternative approaches have been pursued to unravel trophoblast/endometrium cross-talk, including coculture of human endometrial cells and trophoblast,10,11 effects of media from human trophoblasts on decidualized stromal fibroblasts,10 animal models,12 and molecular profiling of first trimester abortion specimens.13

Ectopic pregnancy, where the conceptus is extrauterine, is a useful model to study the changes in the intrauterine endometrium during pregnancy. In this clinical setting, the endocrine effects of pregnancy on the endometrial transcriptome can be investigated without the confounding effects of placental factors or direct paracrine trophoblast–decidua interactions. Herein, we have compared the endometrial transcriptome from women with tubal gestation to that of endometrium of normal, nonpregnant women during the middle (and late) secretory phase of the menstrual cycle. While acknowledging the limitations of this model, these data have relevance to pregnancy maintenance in normal and abnormal settings and give insight into the endometrial changes that occur inside the uterus during the first trimester of pregnancy without the presence of the embryo inside the uterine cavity. In addition, this study has generated provocative hypotheses for subsequent research.

MATERIALS AND METHODS

Sample Collection and Processing

Endometrial samples from pregnant women (n = 9) were obtained from normally cycling subjects undergoing laparoscopic surgery for ectopic pregnancy at Hospital de Clínicas de Porto Alegre, Brazil. After written informed consent, endometrial samples were obtained by suction using a Karman plastic curette (MedGyn Products Inc, Lombard, Illinois), after visual confirmation of the ectopic pregnancy at laparoscopy, but before starting the oviductal surgery. Specimens were snap frozen in liquid nitrogen and stored at −80°C until microarray analysis. Additional samples (n = 4) were fixed in 10% formalin and paraffin embedded for microarray validation through immunohistochemistry. These additional samples will permit the generalization of the microarray data. All patients were normo-ovulatory, had not received methotrexate treatment, or other hormonal medications within 3 months of the surgery, and had no or light uterine bleeding. All histology reports from the endometrial biopsy derived from cases of ectopic pregnancy showed the typical decidual changes in endometrial epithelium.

Gene array data from normally cycling women were obtained from the databank of our previous study.14 The final experiment compared 3 conditions: MSE (n = 8); endometrium from subjects with tubal ectopic pregnancy (Preg; n = 9), and late secretory endometrium (LSE; n = 6). Subject characteristics are shown in Table 1. Samples from LSE were included mainly for ingenuity pathway analysis. This premenstrual phase represents a nonconception cycle, and variations of expression of a certain gene could be seen from the MSE to Preg (conception cycle), or to LSE (nonconception cycle).

Table 1
Characteristics of the Samples

Sample Size and Ethical Issues

This study was submitted and approved by the Brazilian National Committee on Ethics in Research (CONEP 1731/2005) and by the Committee on Human Research at University of California, San Francisco (H52883-31577-01). Sample size was calculated according to the literature,15 considering the median of a standard deviation of 0.2 for HG U133 Plus 2 microarray data set, a minimal 2-fold change for considering a true effect size, a proportion of nondifferentially expressed genes of 0.9, a power of 0.8 and a false discovery rate (FDR) of 0.05. The sample size for these parameters yielded at least 6 samples per group.15

RNA Preparation/Target Preparation/Array Hybridization and Scanning

Total RNA extraction was performed from each sample using Trizol reagent (Invitrogen, Carlsbad, California) according to the manufacturer’s instruction. The RNA preparations were DNase treated and further purified using the RNeasy Mini Kit (QIAGEN, Valencia, CA). The purity was analyzed by both the 260/280 absorbance ratio as well as gel electrophoresis on 1.5% agarose gels. Samples were stored in RNase-free water until further analysis. First-strand and second-strand cDNAs were prepared as described previously.14 Individual samples were hybridized overnight to high-density human genome (HG) U133 Plus 2.0 Arrays (Affymetrix, Inc, Santa Clara, CA), containing 54 675 gene probe sets. Subsequently, the chips were scanned using an HR3000 scanner, and the data were extracted using the Affymetrix GeneChip Operating Software (GCOS) version 1.1.

Microarray Gene Expression Data Analysis

The intensity values of different probe sets (genes) generated by Affymetrix GCOS were imported into GeneSpring GX version 7.3.1 software (Agilent Technologies, Inc, Santa Clara, CA) for data analysis. The data files (.CEL files) containing the probe level intensities from our previous study,14 that is, the early proliferative phase, mid and late secretory phase, plus the new data from the endometrium in ectopic pregnancies were preprocessed by Robust Multichip Average (RMA), according to the gene information from the array. The RMA analyses converted the probe-level expression data into gene-level expression data that were normalized, which is the most consistent “good performance.”16 The RMA normalization steps ensured that the distribution of the expression values was comparable across the HG U133 Plus 2 chips. An additional normalization step (per gene: normalize to the median) was used, which ensured that the expression value for one gene across the different conditions was centered on 1, by dividing the expression value by the median value of the expression values for that gene across conditions. This ensured that genes that do not change across conditions got a normalized expression value of 1, allowing for easy visual detection of differentially expressed genes. After identifying sample replicates, gene signals were transformed into a log ratio for normal distribution. Only the data from MSE, LSE, and Preg were used. Quality control was performed filtering the genes below 3 standard errors of the chips. Further filtering was performed to eliminate outlier genes within samples, by using the 95% confidence interval, resulting in the final working gene list consisting of 38 704 genes.

Hierarchical Clustering

The working gene list was submitted to hierarchical clustering to identify samples with similar patterns of gene expression. The smooth correlation for distance measure algorithm (GeneSpring) was used.

Statistical Analysis

The Volcano plot was applied to gene lists from 2 conditions. The values of fold change and P values between these 2 conditions were plotted for each gene.17 The conditions Preg versus MSE were analyzed. The Volcano plot considered genes with equal variance, and genes that had ≥2-fold change and a P value <.05 in the Student’s t test were selected. The Benjamini–Hochberg multiple testing correction for FDR at 0.05 was used. A separate analysis, using nonparametric statistics between Preg versus MSE, was performed to confirm the gene list generated by the parametric analysis with Volcano plot. The same settings on fold change and FDR were used. Concordance was identified by the kappa coefficient (κ), which measures pairwise agreement among a set of category judgments, correcting for expected chance agreement. A κ value of <.20 indicates weak correlation, .21 to .40 fair, .41 to .60 moderate, .61 to .80 good, and .81 to 1 excellent correlation.18

Further comparisons were made between successive cycle phases using 1-way analysis of variance (ANOVA): MSE, LSE, and Preg with a P < .05 for Ingenuity Pathway Analysis (IPA). The Benjamini–Hochberg multiple testing correction for FDR at 0.05 and Tukey as the post hoc test, resulted in a gene list of 32 912 genes that were further filtered on ≥2-fold change, comparing MSE versus Preg and LSE for IPA. Preliminary data obtained from ANOVA were further analyzed through IPA 5.0 (Ingenuity Systems, http://www.ingenuity.com) for each individual condition and with comparisons between them.

Canonical pathway analysis: entire data set

Canonical pathway analysis identified pathways from the IPA library of canonical pathways that were most significant to the data set. Genes from the data set that met the 2.0 cutoff of log ratio with a P value <.05 and were associated with a canonical pathway in the Ingenuity Pathways Knowledge Base were considered. The significance of the association between the data set and the canonical pathway was measured in 2 ways: (1) A ratio of the number of genes from the data set that map to the pathway divided by the total number of genes that map to the canonical pathway is displayed; (2) Fisher’s exact test was used to calculate a P value determining the probability that the association between the genes in the dataset and the canonical pathway is explained by chance alone.

Pathways graphical representation

IPA Pathways are a graphical representation of the molecular relationships between genes/gene products (http://www.ingenuity.com/company/pdf/Citation_Guidelines_2005-09-13.pdf).

Validation of Microarray Data by Real-Time Polymerase Chain Reaction

First-strand cDNA was generated from 1 μg of total RNA using the iScript cDNA synthesis Kit (Bio-Rad, Hercules, CA). Genes of different expression fold changes in Preg and MSE were selected for validation by real-time polymerase chain reaction (PCR), as described previously,14 with some modifications. Briefly, real-time PCR was performed using the same samples from Preg (n = 9) and MSE (n = 8) and were performed in duplicate using the Brilliant SYBR Green PCR kit (Stratagene, La Jolla, CA) according to the manufacturer’s specifications. Ribosomal protein L19 (RPL19) was the chosen normalizer.14 Primers are listed in Table 2. Primer performance and quality were accessed by standard curves with serial dilutions of pooled samples, as described.14 The efficiencies of amplification (EFF) for each gene were obtained from Mx3005P software (Stratagene, La Jolla, CA) in the exponential phase of the amplification curve. PCR thermal profiles were run using the Mx3005P (Stratagene).14 Primers with standard curves that showed 100% amplification efficiency (±10%) were used in subsequent sample analyses. Ct values were calculated by the Mx3005P software based on fluorescence intensity values after normalization with an internal reference dye and baseline correction.14 Statistical analysis of the RT-PCR data were performed using the Mann–Whitney test of the relative expression of each group, with P < .05.

Table 2
Primers Used for Microarray Gene Expression Validation

Immunohistochemistry

Additional formalin-fixed paraffin-embedded endometrial samples (n = 4 Preg and n = 4 MSE) were used for immunostaining the highest gene expressed in our microarray analysis (matrix metalloproteinase-7, MMP7). Archived breast cancer samples were used as positive controls. Negative control was obtained by using low concentration (2 μg/mL) of primary antibody. Sections were cut (8 μm), deparaffinized, hydrated and treated with 0.3% H2O2–methanol for 30 minutes at 25°C to reduce endogenous peroxidase activity. Nonspecific sites were blocked with 2% normal goat serum. After 3 × 5 minutes phosphate buffered saline (PBS) washings, mouse monoclonal anti-MMP7 active form (IM47, 15 μg/mL; Calbiochem, San Diego, CA) was incubated overnight at 4°C in a humidity chamber. After additional 3 × 5 PBS washings, another blocking with 2% normal goat serum followed by 1 hour of incubation with biotinylated anti-mouse IgG (BA-9200, Vector Laboratories, Burlingame, CA) at 25°C was performed. Slides were incubated with avidin DH-biotinylated horseradish peroxidase H complex (Vectastain Standard ABC kit, Vector Laboratories) for 1 hour at 25°C followed by immersion in 3,3′-diaminobenzidine tetrahydrochloride (Aldrich Chemical Co, St Louis, Missouri) at 150 mg/200 mL 0.05 M Tris buffer containing 0.002% H2O2 for 10 minutes with constant stirring. Sections were exposed to osmium vapors and counterstained with 0.05% toluidine blue in 30% ethanol, dehydrated, cleared in xylene and mounted with Permount (Fisher Scientific, Hanover Park, IL). Photographs were taken using a SPOT-4 Megapixel Digital Color Camera System (Diagnostic Instruments, Inc, Sterling Heights, Michigan) attached to a Nikon ECLIPSE E600 microscope and prepared using SPOT image processing software.

RESULTS

Gene Expression Data Distribution and Hierarchical Clustering

The gene expression data were normally distributed with log ratio transformation in all 3 physiologic groups. The concordance in the gene list generated between parametric and nonparametric analyses was excellent, with the κ value at 0.95. Figure 1 shows the dendrogram of sample clustering and the heatmap of gene expression for all 3 physiologic groups (MSE, LSE, and Preg). The hierarchical tree with branches indicates the relationships between the different samples, using Pearson correlation for similarity measures, and demonstrates that sample groups segregate into 3 main branches, showing similarity within the samples within each physiologic state but clear differences between sample groups.

Figure 1
Hierarchical clustering. The heat map represents the relative expression levels of genes in each individual sample. Each horizontal line represents a single gene and each column represents a single sample. Samples cluster according to the condition (MSE, ...

Volcano Plot/ANOVA

When Preg was compared with MSE, the parametric Volcano plot/ANOVA analysis generated a gene list of 3021 genes, of which 1990 were upregulated and 1031 were downregulated. Analysis of LSE versus MSE yielded a list of 973 genes, of which 515 were upregulated and 458 were downregulated. The genes that are upregulated and downregulated ≥4-fold are in listed in Tables 3 and and4,4, respectively.

Table 3
Genes Upregulated Between Preg and MSE
Table 4
Genes Downregulated in Preg Versus MSE

Ingenuity Pathways Analysis

ANOVA analysis comparing the MSE, LSE, and Preg groups and considering only a P < .05, without considering a fold change, resulted in preliminary composite list of 32 912 genes for uploading to IPA, which can show the changes of the genes over these 3 conditions. These genes with their corresponding gene identifiers, log ratio expression values, and P values were uploaded into the application, and after using a 2-fold cutoff, the analysis identified 774 genes eligible for generating networks.

Canonical pathways analysis

From all canonical pathways from the IPA, 7 pathways reached the 2-fold cutoff in the Preg group. Natural killer cell signaling pathway had a significant increase in LSE, but no expression in Preg (Figure 2). From these pathways, the complement system has the highest significant expression (Figure 2). In the early stages of pregnancy, although factor B is upregulated, the alternate pathway is inhibited by complement factor H (CFH), which has a 4-fold increase (P < .0001). There is no expression of activated C3. The classical pathway is not active either because C4b is not expressed or regulated. The membrane attack complex, composed of C5–C9 (in the late steps of complement activation), did not reach the 2-fold preestablished cutoff (Figure 3). The major histocompatibility complex, class I, E in the antigen presentation pathway (Figure 4) was upregulated in the Preg samples.

Figure 2
Canonical pathways. Ingenuity Pathways Analysis identified the most significant genes from the data set. The bars represent the canonical pathway, which is associated with the genes uploaded from the gene list generated from GeneSpring that met the 2.0 ...
Figure 3
The complement cascade. This figure represents the complement cascade with 3 pathways of the complement cascade represented. The lectin and classical pathways are antigen–antibody complex related, whereas the alternative is continuously activated ...
Figure 4
Representation of the antigen presentation pathway. The expression of the HLA-E is higher in Preg, whereas it is downregulated in MSE and LSE. HLA-E inhibits the activation of the natural killer (NK) cells, thus modulating the immune system. The intensity ...

Validation of Microarray Data by Real-Time PCR

Real-time PCR analysis confirmed the microarray data of regulation of a subset of 21 genes (except CFI, which was only identified in IPA analysis by statistical difference, not by fold change). Statistical analyses for each validated gene using all samples (n = 19) are listed for comparison in Table 5.

Table 5
Real-Time Relative Quantification Transcript for Preg and MSE Samples

Validation of Microarray Data by Immunohistochemistry

The immunostaining for active form of MMP7 was present in the stromal and glandular cells of Preg samples, while in MSE, weak or absent expression was found in stromal cells (Figure 5). Positive and negative controls stained positively and negatively, respectively (data not shown).

Figure 5
Immunostaining for active form of MMP7 in mid-secretory endometrium (MSE) (A) and Preg (B). Note the immunostaining of decidualized endometrium cells in the stroma (arrow) in the presence of ectopic pregnancy (B), whereas no staining is observed in the ...

COMMENT

The endometrial changes that occur between the mid-secretory phase of the menstrual cycle until the pregnancy have been difficult to extrapolate given the ethical and scientific limitations involved in studying early pregnancy. Several array data sets compiled from human pregnancy are available, but they used in vitro models,10,11 or in vivo models in the presence of the trophoblast.13 By analyzing the endometrium from the uterine cavity in the setting of a tubal pregnancy, we were able to identify dramatic changes that reflect the endocrine milieu of pregnancy without the contribution of the conceptus inside the uterine cavity and this may be the reason why many differences were found comparing our model to the published data. For instance, Chen et al13 used a human in vivo model comparing the decidua gene profile from an intrauterine pregnancy to the transcriptome of the trophoblast, and they presented a list of 48 genes that were upregulated in the decidua. From these genes, only 7 were in common with our gene list (IGFBP1; thrombin receptor-like 2; protein tyrosine phosphatase, receptor type; AXL receptor tyrosine kinase; CD54; NADH dehydrogenase [ubiquinone] 1; and microsomal glutathione S-transferase 1). Furthermore, the highly expressed presence of insulin receptor found by Chen et al13 and Popovici et al,11 who used a coculture model, was not found in our study, underlying the difference made by the local presence of the trophoblast. The comparison between the transcriptome of MSE to the decidua in a setting of an ectopic pregnancy gives valuable insight into the modifications of genes, gene families, and pathways that occur from the window of implantation (MSE) to the early gestation period as a result of secretion of products from the ectopic trophoblast. We are not able to distinguish if the differences found are derived from the products of the corpus luteum or from the ectopic trophoblast and embryo. A study having women taking estradiol and progesterone for 6 weeks could help explain it, but it could face ethical concerns. The numerous genes, gene families, and canonical pathways that are upregulated in the decidua in the setting of tubal gestation, compared with MSE, are subject to several interpretations. Some of the genes and pathways may be regulated in an endocrine fashion by products from the corpus luteum and the ectopic placenta, or from a prolonged exposure of their products. Caution regarding interpretation of these results is required given that endocrine parameters in ectopic pregnancy may vary from that of normal pregnancies. Some may be absent or have limited expression in endometrial decidua with ectopic gestation because they ordinarily result from paracrine interactions between the trophoblast and the decidual cells in a normal pregnancy. There may be a threshold value at which gene expression in the decidua is maintained or even stimulated, and below which, gene expression may be suppressed. Also, the presence of a placenta in the fallopian tube may result in paracrine signaling between the trophoblast and oviductal cells that could result in secreted products (eg, cytokines, chemokines) affecting the uterine decidual transcriptome. Of note, the majority of our LSE samples had leiomyomas, which could be a limitation of our study. Because our main analysis was based on MSE versus Preg, the potential impact of this is minimal based on our final analysis.

In this study, we compared gene expression profiles and pathways of intrauterine decidua in the presence of ectopic pregnancy with 6.5 ± 1.04 (mean ± SEM) weeks with MSE, which is the third week of the menstrual cycle (peak of progesterone) in the nonpregnant state. It has being reported that between 5 and 9 weeks of gestation, circulating progesterone and hCG levels are lower in ectopic pregnancy, compared with normal intrauterine pregnancy.19 Mean values of progesterone and β-hCG in the latter were 21.3 ± 6.8 ng/mL and 37 533 ± 30 133 mUI/mL, compared with ectopic pregnancy values of 6.3 ± 3.1 ng/mL and 4106 ± 1762 mUI/mL, respectively. However, in our samples, we found mean hCG levels of 16 665 ± 26 267 mUI/mL from the available data (n = 8). Mean progesterone in the mid-luteal phase of a nonconception cycle is 10.83 ng/mL.20 Although we did not measure progesterone levels in our subjects, which precludes an objective conclusion about the role of progesterone over the endometrium in this setting, the overall gene profiles and regulated pathways observed herein are not consistent with global progesterone withdrawal at the level of the endometrium (as in LSE in nonconception cycles14,21) or with an inadequate progesterone response. In support of this is the observation (Table 3) that IGFBP-1, IGFBP-6, somatostatin, HTR2B, HLA-DQB1, RGS2, HBB, IL1R type II, and ADAMTS2 are all upregulated in Preg decidua versus MSE, and are known to be upregulated in either endometrial stromal fibroblasts in response to progesterone22 or in MSE in nonconception cycles.14,23 Another possibility for these findings is a prolonged progesterone exposure.

A study that most resembles our model was used by Sherwin et al.12 They used intrauterine infusion of hCG in the baboon, and found regulation of numerous genes in endometrium that are important in embryo attachment, endometrial remodeling, antioxidant defense, and modulation of the immune response. Our 6-week decidua data demonstrate that 30 genes out of the 61 genes regulated by hCG observed by Sherwin et al12 are in common. Of note, MMP7, members of the complement and coagulation cascades, and interleukin-8 (IL-8) are among these (Table 6). The similarities are even greater if the pathways involved are considered, such as the complement cascade and the Wnt pathway (see below). Yet, IL-6 levels reported by Sherwin et al12 were not found in our study. IL-6 was significantly upregulated in MSE, compared with Preg samples, by RT-PCR analysis (Table 5), for unclear reasons. Furthermore, our results are in accordance to those found by Tabibzadeh et al,24 in that IL-6 mRNA expression increases progressively across the menstrual cycle (in our microarray data: MSE:1.122, LSE:1.625, Preg:1.879), although they were not statistically significant (P = .3).

Table 6
Comparison Between the Gene List From Sherwin et al12 and Savaris et al (This Study)

The different genes across the 3 different conditions (MSE, LSE, and Preg) were used for IPA analysis, where the variations of the genes across the conditions can be seen. We decided to validate the main genes identified in the microarray analysis that were relevant for endometrial receptivity or were derived from IPA analysis, such as ETS-1, which could explain the upregulation of MMP7. Of note, MMP7 is one of the products of the Wnt/β-catenin pathway, and this pathway is probably downregulated, and Sherwin et al12 reported the expression of this MMP. Because of these, we decided to validate MMP7 by IHC. Genes with a wide function, such as cytochrome P450, were precluded from the validation. However, we focus our discussion on the main processes between MSE and Preg samples because they represent the physiological timeline from uterine receptivity through early pregnancy.

DECIDUALIZATION/APOPTOSIS AND PROLIFERATION PREVENTION

Inhibin βA/Somatostatin (SST)/Forkhead Box O1 (FOXO1A)/IGFBP1/TIMP3

Inhibin βA is upregulated in early pregnancy, and this finding is corroborated by Otani et al,25 who reported intense immunoreactivity in decidual stromal cells during early pregnancy. Inhibin βA promotes decidualization per se.26 Progesterone (and activation of the protein kinase A pathway) is a regulator of FOXO1A, prolactin, IGFBP1, TIMP3, SST, endometrial bleeding-associated factor (EBAF), and BIM (bcl-2 interacting mediator of cell death) mRNAs in endometrial stromal fibroblasts27,28 in vitro and in vivo studies, and all are markedly induced during decidualization process.14 Herein, mRNA levels of FOXO1A, IGFBP1, TIMP3, and SST were all upregulated in our microarray data, and they were validated by RT-PCR analyses, compared with MSE, except for TIMP3 that did not reach statistical significance. BIM mRNA levels were upregulated in microarray analysis as well. Interestingly, placental corticotrophin releasing hormone (pCRH), whose plasma levels progressively increase during pregnancy,29 increases FOXO1A and IGFBP1 mRNA expression through signaling via the glucocorticoid receptor,30 which is upregulated in our microarray analysis (1.4-fold change; P = .03), suggesting that the decidua may be a target for placental CRH in an endocrine fashion. Labied et al28 have shown that withdrawal of progesterone from human decidualized endometrial stromal cells elicited reaccumulation of FOXO1A in the nucleus and apoptosis, whereas in the presence of progesterone, FOXO1A is accumulated in the cytoplasm of the cells and apoptosis is precluded. Thus, these data are consistent with our findings, and we may infer that progesterone levels found in early pregnancy may preclude endometrial cellular apoptosis through FOXO1A. Furthermore, endometrial proliferation is precluded by inhibition of the Wnt/β-catenin pathway, through the upregulation of DKK-3 and SFRP1, as has been shown by Tulac et al31 in an in vitro study.

IMMUNOLOGICAL PROCESSES

Immune interactions at the maternal–fetal interface play an important role in implantation biology. Some participants have been found to be regulated in Preg versus MSE, including members of the complement cascade, such as complement factor H (CFH), and from the antigen presentation pathway (HLA-E). The complement inhibition is an absolute requirement for normal pregnancy, where Crry deficiency in utero leads to progressive embryonic lethality by activation of C3 fragments and polymorphonuclear leukocytes.32 In humans, during the window of implantation, maximal DAF expression, at the mRNA and protein levels, occurs by day 22 of the menstrual cycle and is sustained thereafter in whole tissue and in endometrial luminal and glandular epithelium.33 Nevertheless, our results showed that only the alternative complement cascade is inhibited by CFH, in decidua in the presence of ectopic pregnancy, as demonstrated by microarray and RT-PCR analysis. The alternative pathway is related to autoantibodies and the initiation of the inflammatory cascade.34 Surprisingly, the lectin and the classical pathways are not inhibited. These pathways are activated by bacterial surface and antigen–antibody complex, respectively,35 showing that the decidua is prepared to defend against bacterial infections. Similarly, another antibacterial mechanism, through lipolcain-2 (LCN2) was found in our microarray and RT-PCR analysis. LCN2 modulates inflammation/infection, and prevents growth of microorganisms by depriving them of iron.36 Furthermore, LCN2 is upregulated by trophoblast products, such as glucocorticoid, retinoic acid,37 IGF-I,38 E2,39 and progesterone,40 and downregulated by activation of the Wnt pathway.41 All these data are in accordance with the current findings. The physiologic importance of LCN2 in pregnancy can be derived from the genotype of LCN2 −/− knockout mice.42 LCN2 −/− females have a normal phenotype but are less fertile than wild-type mice. In wild-type mice, LCN2 mRNA and protein in the decidua remain high immediately after fertilization and early pregnancy (on days 1 and 2 of a 20-day gestation), and decline with progression of pregnancy.43 These findings are comparable to ours in early pregnancy.

It is well established that uterine natural killer (NK) cells are particularly abundant around the time of implantation and during early pregnancy.44 However, the IPA analysis showed a significant reduction of NK cells signaling in Preg samples, compared with LSE as shown in Figure 2. A possible explanation for this finding can be related to increased expression of HLA-E, a nonclassical major histocompatibility complex (MHC) class I molecule.45 HLA-E has been related to the inhibition of the killing activity of NK cells.46 HLA-E complex is elevated in Preg samples, and the presence of HLA-E in decidua may protect an incoming embryo from NK cell attack. Indeed, there is growing evidence showing that decidual NK cells control trophoblast invasion, and this control occurs through the release of an interferon-inducible protein-10 (IP-10), and IL-8.47 In our analysis, IL-8 mRNA is significantly upregulated in Preg samples, compared with the MSE, which is consistent with the results found by Critchley et al.21 These authors demonstrated IL-8 immunostaining primarily in perivascular sites in decidua derived from ectopic pregnancy.21 Upregulation of IL-8 has been reported at the site of endometrial and embryo interactions in most,1012 but not all13 studies, and hCG upregulates its expression.12 IL-8 has a potent chemotactic activity on neutrophils migration,48 and this migration involves the expression of calgranulin A/B complex (S100A8/A9).49 This complex is upregulated in our microarray analysis, and S100A8 was confirmed by RT-PCR.

EXTRACELLULAR REMODELING AND ANGIOGENESIS

MMP7 and MMP11 are the most highly regulated genes in the Preg samples (Table 3). They are known to be downregulated in endometrial epithelium and stromal fibroblasts, respectively, in response to progesterone, although upregulation of MMP7 in response to hCG was demonstrated in the baboon model described above.12 MMPs are regulated by transforming growth factor beta (TGFβ) family members, and are activated by other proteases and inhibited by tissue inhibitor of metalloproteases (TIMPs).50 Immunohistochemical staining of endometrial decidua derived in the setting of tubal pregnancy confirmed the expression of the active form of this collagenase (Figure 5). Interestingly, it is one of the products of the Wnt/β-catenin signaling pathway activation.51 However, because DKK-3 and SFRP1, inhibitors of the Wnt/β-catenin pathway, are upregulated in the decidua of ectopic pregnancy (Table 5), the marked upregulation of MMP7 is likely controlled by other mechanisms. Therefore we decided to validate the active form of the protein by IHC. Candidates include v-ets erythroblastosis virus E26 oncogene homolog 1 (avian) (ETS-1), which was confirmed by RT-PCR analysis, and retinoic acid, which is highly expressed in Preg samples (Figure 2). Both ETS-1 and retinoic acid are known stimulators of MMP7,52 as well hCG.12 With regard to MMP11, Popovici et al11 did not find regulation of MMP11 in a coculture model of trophoblasts and endometrial stromal cells, and further studies are needed to understand the regulation of this MMP in decidua.

Periostin was highly upregulated in Preg samples by microarray analysis (7.4-fold) and confirmed by RT-PCR (14.11-fold). There are no published data on POSTN in the human endometrium, although this secreted and soluble extracellular matrix protein53 has been identified in a wide range of normal adult tissues,5457 and cancers.58,59 Its function is related to extracellular matrix remodeling,60 enhancement of motility and invasiveness of cancer cells, and is a ligand for the integrins αvβ3 and αvβ5.54 It may serve similar functions for trophoblasts in decidua. TGFβ1 increases the expression of periostin,55 and TIMP-2 and periostin are codistributed on extracellular matrix.61 Herein, TGFβ1 was found to be upregulated 2.27-fold (P = .0007) and TIMP2 2.67-fold (P = .007) in Preg samples, consistent with a role and distribution of POSTN in other systems.

Along with its coagulation activity, PAI-1 exerts control over trophoblast invasion.62 In our study, PAI-1 was upregulated in Preg samples. Lockwood et al63 demonstrated that PAI-1 protein and mRNA expression is enhanced by estradiol and medroxyprogesterone acetate in decidualized stromal cells from luteal phase and from first trimester endometrium cultured in type 1 collagen gels. PAI-1 is important for pregnancy maintenance, as women with homozygosity for PAI-1 polymorphism have a higher relative risk of early pregnancy loss.64 The upregulation of PAI-1 without the local presence of the trophoblast is a new finding and needs further investigation.

Thus, MMPs play a key role in angiogenesis by degrading the extracellular matrix and permitting the migration and tube formation of endothelial cells,65 and IL-8 induces angiogenesis.66 Lumican is involved in extracellular matrix remodeling,67 as are MMP7, MMP11, periostin, and IL-8. These observations underscore the importance of extracellular matrix remodeling in the decidua in preparation of the highly invasive trophoblast.

In conclusion, we have demonstrated that the transcriptome of the decidua in the setting of an ectopic pregnancy is likely influenced by trophoblast products in an endocrine fashion, compared with nonpregnant, secretory endometrium. The validated genes in this study are related to the coagulation cascade and trophoblast invasion (PAI-1), immunological processes (HLA-E, IL-8, S100A8, CFH), endometrial tissue remodeling (MMP7, 11, periostin, lumican, IL-8), decidualization, and apoptosis/proliferation prevention (FOXO1A, IGFBP-1, SST, DKK3, SFRP1). Our findings are in accordance with studies by others that used in vitro human and animal models, suggesting that although paracrine interactions are important in the invasive phase of implantation, endocrine effects are also likely important, as well. We also have found genes that have not been described before in the endometrium of early human pregnancy, such as complement factor H, periostin, lipocalin 2, and lumican, as well as CFI in MSE, opening new fields for further investigation. Importantly, in our list of upregulated genes in Preg versus MSE, we did not observe upregulation of GRO-1, DKK-1, and TIMP-3, previously observed in cultured cells of human endometrial stromal fibroblast in the presence of trophoblast,13 and in stromal fibroblasts treated with trophoblast culture medium.10 These observations strongly support either paracrine regulation of these factors or a threshold of response to progesterone with regard to specific gene expression. This is in contrast to IL-8, FOXO1A, SST, and IGFBP1, which appear to be regulated by endocrine factors, such as progesterone, hCG, and other placental products that prepare the endometrium for implantation.

Acknowledgments

Supported by the NICHD/NIH through cooperative agreement 1U54HD055764-01 (LCG) and U54-HD35041 (BAL) as part of the Specialized Cooperative Centers Program in Reproduction and Infertility Research (LCG) and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) BEX3656-06-3 (RFS). Samples were collected at Hospital de Clínicas de Porto Alegre. Data analysis and RT-PCR were performed at University of California, San Francisco. Immunohistochemistry was performed at Greenville Hospital System.

References

1. Noyes RW, Hertig AT, Rock J. Dating the endometrial biopsy. Fertil Steril. 1950;1:3–25.
2. Lessey BA, Damjanovich L, Coutifaris C, Castelbaum A, Albelda SM, Buck CA. Integrins adhesion molecules in the human endometrium. Correlation with the normal and abnormal menstrual cycle. J Clin Invest. 1992;90:188–195. [PMC free article] [PubMed]
3. Achache H, Revel A. Endometrial receptivity markers, the journey to successful embryo implantation. Hum Reprod Update. 2006;12:731–746. [PubMed]
4. Giudice LC, Irwin JC. Roles of the insulinlike growth factor family in nonpregnant human endometrium and at the decidual:trophoblast interface. Semin Reprod Endocrinol. 1999;17:13–21. [PubMed]
5. Mooney SB, Giudice LC. Fetal Endocrinology. 2002. Endotext.com.
6. Carson DD, Lagow E, Thathiah A, et al. Changes in gene expression during the early to mid-luteal (receptive phase) transition in human endometrium detected by high-density microarray screening. Mol Hum Reprod. 2002;8:871–879. [PubMed]
7. Horcajadas JA, Riesewijk A, Martín J, et al. Global gene expression profiling of human endometrial receptivity. J Reprod Immunol. 2004;63:41–49. [PubMed]
8. Kao LC, Tulac S, Lobo S, et al. Global gene profiling in human endometrium during the window of implantation. Endocrinology. 2002;143:2119–2138. [PubMed]
9. Ponnampalam AP, Weston GC, Trajstman AC, Susil B, Rogers PA. Molecular classification of human endometrial cycle stages by transcriptional profiling. Mol Hum Reprod. 2004;10:879–893. [PubMed]
10. Hess AP, Hamilton AE, Talbi S, et al. Decidual stromal cell response to paracrine signals from the trophoblast: amplification of immune and angiogenic modulators. Biol Reprod. 2007;76:102–117. [PubMed]
11. Popovici RM, Betzler NK, Krause MS, et al. Gene expression profiling of human endometrial-trophoblast interaction in a coculture model. Endocrinology. 2006;147:5662–5675. [PubMed]
12. Sherwin JR, Sharkey AM, Cameo P, et al. Identification of novel genes regulated by chorionic gonadotropin in baboon endometrium during the window of implantation. Endocrinology. 2007;148:618–626. [PubMed]
13. Chen HW, Chen JJ, Tzeng CR, et al. Global analysis of differentially expressed genes in early gestational decidua and chorionic villi using a 9600 human cDNA microarray. Mol Hum Reprod. 2002;8:475–484. [PubMed]
14. Talbi S, Hamilton AE, Vo KC, et al. Molecular phenotyping of human endometrium distinguishes menstrual cycle phases and underlying biological processes in normo-ovulatory women. Endocrinology. 2006;147:1097–1121. [PubMed]
15. Liu P, Hwang JT. Quick calculation for sample size while controlling false discovery rate with application to microarray analysis. Bioinformatics. 2007;23:739–746. [PubMed]
16. Breitling R. Biological microarray interpretation: the rules of engagement. Biochim Biophys Acta. 2006;1759:319–327. [PubMed]
17. Cui X, Churchill GA. Statistical tests for differential expression in cDNA microarray experiments. Genome Biol. 2003;4:210. [PMC free article] [PubMed]
18. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. Fam Med. 2005;37:360–363. [PubMed]
19. Soriano D, Hugol D, Quang NT, Darai E. Serum concentrations of interleukin-2R (IL-2R), IL-6, IL-8, and tumor necrosis factor alpha in patients with ectopic pregnancy. Fertil Steril. 2003;79:975–980. [PubMed]
20. Sallam HN, Sallam A, Ezzeldin F, Agamia AF, Abou-Ali A. Reference values for the midluteal plasma progesterone concentration:evidence from human menopausal gonadotropin-stimulated pregnancy cycles. Fertil Steril. 1999;71:711–714. [PubMed]
21. Critchley HO, Jones RL, Lea RG, et al. Role of inflammatory mediators in human endometrium during progesterone withdrawal and early pregnancy. J Clin Endocrinol Metab. 1999;84:240–248. [PubMed]
22. Tierney EP, Tulac S, Huang ST, Giudice LC. Activation of the protein kinase A pathway in human endometrial stromal cells reveals sequential categorical gene regulation. Physiol Genomics. 2003;16:47–66. [PubMed]
23. Burney RO, Talbi S, Hamilton AE, et al. Gene expression analysis of endometrium reveals progesterone resistance and candidate susceptibility genes in women with endometriosis. Endocrinology. 2007;148:3814–3826. [PubMed]
24. Tabibzadeh S, Kong QF, Babaknia A, May LT. Progressive rise in the expression of interleukin-6 in human endometrium during menstrual cycle is initiated during the implantation window. Hum Reprod. 1995;10:2793–2799. [PubMed]
25. Otani T, Minami S, Kokawa K, Shikone T, Yamoto M, Nakano R. Immunohistochemical localization of activin A in human endometrial tissues during the menstrual cycle and in early pregnancy. Obstet Gynecol. 1998;91(5 Pt 1):685–692. [PubMed]
26. Jones RL, Salamonsen LA, Zhao YC, Ethier JF, Drummond AE, Findlay JK. Expression of activin receptors, follistatin and betaglycan by human endometrial stromal cells; consistent with a role for activins during decidualization. Mol Hum Reprod. 2002;8:363–374. [PubMed]
27. Grinius L, Kessler C, Schroeder J, Handwerger S. Forkhead transcription factor FOXO1A is critical for induction of human decidualization. J Endocrinol. 2006;189:179–187. [PubMed]
28. Labied S, Kajihara T, Madureira PA, et al. Progestins regulate the expression and activity of the forkhead transcription factor FOXO1 in differentiating human endometrium. Mol Endocrinol. 2006;20:35–44. [PubMed]
29. Sasaki A, Shinkawa O, Yoshinaga K. Placental corticotropin-releasing hormone may be a stimulator of maternal pituitary adrenocorticotropic hormone secretion in humans. J Clin Invest. 1989;84:1997–2001. [PMC free article] [PubMed]
30. Rogatsky I, Wang JC, Derynck MK, et al. Target-specific utilization of transcriptional regulatory surfaces by the glucocorticoid receptor. Proc Natl Acad Sci U S A. 2003;100:13845–13850. [PubMed]
31. Tulac S, Nayak NR, Kao LC, et al. Identification, characterization, and regulation of the canonical Wnt signaling pathway in human endometrium. J Clin Endocrinol Metab. 2003;88:3860–3866. [PubMed]
32. Xu C, Mao D, Holers VM, Palanca B, Cheng AM, Molina H. A critical role for murine complement regulator crry in fetomaternal tolerance. Science. 2000;287:498–501. [PubMed]
33. Young SL, Lessey BA, Fritz MA, et al. In vivo and in vitro evidence suggest that HB-EGF regulates endometrial expression of human decay-accelerating factor. J Clin Endocrinol Metab. 2002;87:1368–1375. [PubMed]
34. Low JM, Moore TL. A role for the complement system in rheumatoid arthritis. Curr Pharm Des. 2005;11:655–670. [PubMed]
35. Markiewski MM, Lambris JD. The role of complement in inflammatory diseases from behind the scenes into the spotlight. Am J Pathol. 2007;171:715–727. [PubMed]
36. Borregaard N, Theilgaard-Mönch K, Cowland JB, Ståhle M, Sørensen OE. Neutrophils and keratinocytes in innate immunity—cooperative actions to provide antimicrobial defense at the right time and place. J Leukoc Biol. 2005;77:439–443. [PubMed]
37. Bratt T. Lipocalins and cancer. Biochim Biophys Acta. 2000;1482:318–326. [PubMed]
38. Sørensen OE, Cowland JB, Theilgaard-Mönch K, Liu L, Ganz T, Borregaard N. Wound healing and expression of antimicrobial peptides/polypeptides in human keratinocytes, a consequence of common growth factors. J Immunol. 2003;170:5583–5589. [PubMed]
39. Seth P, Porter D, Lahti-Domenici J, Geng Y, Richardson A, Polyak K. Cellular and molecular targets of estrogen in normal human breast tissue. Cancer Res. 2002;62:4540–4544. [PubMed]
40. Yao MW, Lim H, Schust DJ, et al. Gene expression profiling reveals progesterone-mediated cell cycle and immunoregulatory roles of Hoxa-10 in the preimplantation uterus. Mol Endocrinol. 2003;17:610–627. [PubMed]
41. Ziegler S, Röhrs S, Tickenbrock L, et al. Lipocalin 24p3 is regulated by the Wnt pathway independent of regulation by iron. Cancer Genet Cytogenet. 2007;174:16–23. [PubMed]
42. Berger T, Togawa A, Duncan GS, et al. Lipocalin 2-deficient mice exhibit increased sensitivity to Escherichia coli infection but not to ischemia-reperfusion injury. Proc Natl Acad Sci U S A. 2006;103:1834–1839. [PubMed]
43. Huang HL, Chu ST, Chen YH. Ovarian steroids regulate 24p3 expression in mouse uterus during the natural estrous cycle and the preimplantation period. J Endocrinol. 1999;162:11–19. [PubMed]
44. King A, Burrows T, Verma S, Hiby S, Loke YW. Human uterine lymphocytes. Hum Reprod Update. 1998;4:480–485. [PubMed]
45. Huddleston H, Schust DJ. Immune interactions at the maternal-fetal interface: a focus on antigen presentation. Am J Reprod Immunol. 2004;51:283–289. [PubMed]
46. Lee N, Llano M, Carretero M, et al. HLA-E is a major ligand for the natural killer inhibitory receptor CD94/NKG2A. Proc Natl Acad Sci U S A. 1998;95:5199–5204. [PubMed]
47. Le BP, Tabiasco J. Killers become builders during pregnancy. Nat Med. 2006;12:991–992. [PubMed]
48. Kobayashi Y. The role of chemokines in neutrophil biology. Front Biosci. 2008;13:2400–2407. [PubMed]
49. Foell D, Wittkowski H, Vogl T, Roth J. S100 proteins expressed in phagocytes: a novel group of damage-associated molecular pattern molecules. J Leukoc Biol. 2007;81:28–37. [PubMed]
50. Dong JC, Dong H, Campana A, Bischof P. Matrix metallo-proteinases and their specific tissue inhibitors in menstruation. Reproduction. 2002;123:621–631. [PubMed]
51. Zhai Y, Wu R, Schwartz DR, et al. Role of beta-catenin/T-cell factor-regulated genes in ovarian endometrioid adenocarcinomas. Am J Pathol. 2002;160:1229–1238. [PubMed]
52. Raouf A, Li V, Kola I, Watson DK, Seth A. The Ets1 protooncogene is upregulated by retinoic acid: characterization of a functional retinoic acid response element in the Ets1 promoter. Oncogene. 2000;19:1969–1974. [PubMed]
53. Takeshita S, Kikuno R, Tezuka K, Amann E. Osteoblast-specific factor 2: cloning of a putative bone adhesion protein with homology with the insect protein fasciclin I. Biochem J. 1993;294(Pt 1):271–278. [PubMed]
54. Gillan L, Matei D, Fishman DA, Gerbin CS, Karlan BY, Chang DD. Periostin secreted by epithelial ovarian carcinoma is a ligand for alpha(V)beta(3) and alpha(V)beta(5) integrins and promotes cell motility. Cancer Res. 2002;62:5358–5364. [PubMed]
55. Horiuchi K, Amizuka N, Takeshita S, et al. Identification and characterization of a novel protein, periostin, with restricted expression to periosteum and periodontal ligament and increased expression by transforming growth factor beta. J Bone Miner Res. 1999;14:1239–1249. [PubMed]
56. Sasaki H, Lo KM, Chen LB, et al. Expression of periostin, homologous with an insect cell adhesion molecule, as a prognostic marker in non-small cell lung cancers. Jpn J Cancer Res. 2001;92:869–873. [PubMed]
57. Litvin J, Chen X, Keleman S, Zhu S, Autieri M. Expression and function of periostin-like factor in vascular smooth muscle cells. Am J Physiol Cell Physiol. 2007;292:C1672–C1680. [PubMed]
58. Grigoriadis A, Mackay A, Reis-Filho JS, et al. Establishment of the epithelial-specific transcriptome of normal and malignant human breast cells based on MPSS and array expression data. Breast Cancer Res. 2006;8:R56. [PMC free article] [PubMed]
59. Kudo Y, Siriwardena BS, Hatano H, Ogawa I, Takata T. Periostin: novel diagnostic and therapeutic target for cancer. Histol Histopathol. 2007;22:1167–1174. [PubMed]
60. Oka T, Xu J, Kaiser RA, et al. Genetic manipulation of periostin expression reveals a role in cardiac hypertrophy and ventricular remodeling. Circ Res. 2007;101:313–321. [PMC free article] [PubMed]
61. Yoshiba N, Yoshiba K, Hosoya A, et al. Association of TIMP-2 with extracellular matrix exposed to mechanical stress and its co-distribution with periostin during mouse mandible development. Cell Tissue Res. 2007;330:133–145. [PubMed]
62. Floridon C, Nielsen O, Hølund B, et al. Does plasminogen activator inhibitor-1 (PAI-1) control trophoblast invasion? A study of fetal and maternal tissue in intrauterine, tubal and molar pregnancies. Placenta. 2000;21:754–762. [PubMed]
63. Lockwood CJ, Krikun G, Papp C, et al. The role of progestationally regulated stromal cell tissue factor and type-1 plasminogen activator inhibitor (PAI-1) in endometrial hemostasis and menstruation. Ann N Y Acad Sci. 1994;734:57–79. [PubMed]
64. Dossenbach-Glaninger A, van Trotsenburg M, Dossenbach M, et al. Plasminogen activator inhibitor 1 4G/5G polymorphism and coagulation factor XIII Val34Leu polymorphism: impaired fibrinolysis and early pregnancy loss. Clin Chem. 2003;49:1081–1086. [PubMed]
65. Smith SK. Regulation of angiogenesis in the endometrium. Trends Endocrinol Metab. 2001;12:147–151. [PubMed]
66. Strieter RM, Kunkel SL, Elner VM, et al. Interleukin-8. A corneal factor that induces neovascularization. Am J Pathol. 1992;141:1279–1284. [PubMed]
67. Iozzo RV. The family of the small leucine-rich proteoglycans: key regulators of matrix assembly and cellular growth. Crit Rev Biochem Mol Biol. 1997;32:141–174. [PubMed]