help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kao, L. C.
Right arrow Articles by Giudice, L. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Kao, L. C.
Right arrow Articles by Giudice, L. C.
Endocrinology Vol. 143, No. 6 2119-2138
Copyright © 2002 by The Endocrine Society

REPRODUCTION-DEVELOPMENT

Global Gene Profiling in Human Endometrium during the Window of Implantation

L. C. Kao, S. Tulac, S. Lobo, B. Imani, J. P. Yang, A. Germeyer, K. Osteen, R. N. Taylor, B. A. Lessey and L. C. Giudice

Department of Gynecology and Obstetrics (L.C.K., S.T., S.L., B.I., J.P.Y., A.G., L.C.G.), Stanford University, Stanford, California 94305; Department of Obstetrics & Gynecology (K.O.), Vanderbilt University, Nashville, Tennessee 37232; Department of Obstetrics, Gynecology, and Reproductive Sciences (R.N.T.), the University of California, San Francisco, California 94143; and Department of Obstetrics & Gynecology (B.A.L.), University of North Carolina, Chapel Hill, North Carolina 27599

Address all correspondence and requests for reprints to: Linda C. Giudice, M.D., Ph.D., Center for Research on Women’s Health and Reproductive Medicine, Division of Reproductive Endocrinology and Infertility, Department of Gynecology and Obstetrics, Stanford University, Stanford, California 94305-5317. E-mail: . giudice{at}stanford.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
Implantation in humans is a complex process that is temporally and spatially restricted. Over the past decade, using a one-by-one approach, several genes and gene products that may participate in this process have been identified in secretory phase endometrium. Herein, we have investigated global gene expression during the window of implantation (peak E2 and progesterone levels) in well characterized human endometrial biopsies timed to the LH surge, compared with the late proliferative phase (peak E2 level) of the menstrual cycle. Tissues were processed for poly(A+) RNA and hybridization of chemically fragmented, biotinylated cRNAs on high density oligonucleotide microarrays, screening for 12,686 genes and expressed sequence tags. After data normalization, mean values were obtained for gene readouts and fold ratios were derived comparing genes up- and down-regulated in the window of implantation vs. the late proliferative phase. Nonparametric testing revealed 156 significantly (P < 0.05) up-regulated genes and 377 significantly down-regulated genes in the implantation window. Up-regulated genes included those for cholesterol trafficking and transport [apolipoprotein (Apo)E being the most induced gene, 100-fold], prostaglandin (PG) biosynthesis (PLA2) and action (PGE2 receptor), proteoglycan synthesis (glucuronyltransferase), secretory proteins [glycodelin, mammaglobin, Dickkopf-1 (Dkk-1, a Wnt inhibitor)], IGF binding protein (IGFBP), and TGF-ß superfamilies, signal transduction, extracellular matrix components (osteopontin, laminin), neurotransmitter synthesis (monoamine oxidase) and receptors ({gamma} aminobutyric acid A receptor {pi} subunit), numerous immune modulators, detoxification genes (metallothioneins), and genes involved in water and ion transport [e.g. Clostridia Perfringens Enterotoxin (CPE) 1 receptor (CPE1-R) and K+ ion channel], among others. Down-regulated genes included intestinal trefoil factor (ITF) [the most repressed gene (50-fold)], matrilysin, members of the G protein-coupled receptor signaling pathway, frizzled-related protein (FrpHE, a Wnt antagonist), transcription factors, TGF-ß signaling pathway members, immune modulators (major histocompatibility complex class II subunits), and other cellular functions. Validation of select genes was conducted by Northern analysis and RT-PCR using RNA from endometrial biopsies obtained in the proliferative phase and the implantation window and by RT-PCR using RNA from cultured endometrial epithelial and stromal cells. These approaches confirmed up-regulation of genes corresponding to IGFBP-1, glycodelin, CPE1-R, Dkk-1, mammaglobin, and ApoD and down-regulation for PR membrane component 1, FrpHE, matrilysin, and ITF, as with the microarray data. Cultured endometrial epithelial cells were found to express mRNAs for glycodelin, CPE-1R, Dkk-1, the {gamma} aminobutyric acid A receptor {pi} subunit, mammaglobin, matrilysin, ITF and PR membrane component 1. The expression of IGFBP-1, CPE1-R, Dkk-1, and ApoD mRNAs increased upon decidualization of stromal cells in vitro with progesterone after E2 priming, whereas FrpHE decreased, consistent with the microarray results. Overall, the data demonstrate numerous genes and gene families not heretofore recognized in human endometrium or associated with the implantation process. Reassuringly, several gene products, known to be differentially expressed in the implantation window or in secretory endometrium, were verified, and the striking regulation of select secretory proteins, water and ion channels, signaling molecules, and immune modulators underscores the important roles of these systems in endometrial development and endometrial-embryonic interactions. In addition, the current study validates using high density oligonucleotide microarray technology to investigate global changes in gene expression in human endometrium.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
IMPLANTATION IN HUMANS involves complex interactions between the embryo and the maternal endometrium. Histologic examination of early human pregnancies reveals distinct patterns of blastocyst attachment to the endometrial surface and the underlying stroma (1, 2), supporting a model of implantation in humans in which the embryo apposes and attaches to the endometrial epithelium, traverses adjacent cells of the epithelial lining, and invades into the endometrial stroma. The endometrium is receptive to embryonic implantation during a defined window that is temporally and spatially restricted. Temporal definition of the window of implantation in human endometrium derives from several sources. Early studies suggest that the window resides in the midsecretory phase because embryos identified in secretory phase hysterectomy specimens were all free-floating before d 20 of the cycle and were all attached when specimens were obtained after d 20 (1). In addition, the temporal and spatial appearance of epithelial dome-like structures (pinopodes) support a receptive phase of embryonic implantation because they appear on cycle d 20–24, correlate with implantation sites, and are believed to participate in attachment of the embryo to the epithelium (3). A recent report (4) demonstrates a high success (84%) of continuing pregnancy for embryos that implant between cycle d 22–24 (postovulatory d 8–10), compared with 18% when implantation occurred 11 d or more after ovulation. Together these data suggest that the window of implantation in humans spans cycle d 20–24 and involves the epithelium and subsequently underlying stroma.

Molecular definition of the window of implantation in human endometrium has been more difficult to define and derives primarily from animal models and clinical specimens (see Refs. 5 and 6 for reviews). These studies have revealed a limited number of potential molecular markers of the implantation window and of uterine receptivity to embryonic implantation. Animal models of homologous recombination and gene knockouts that demonstrate an implantation-based infertility phenotype provide important insight into potential markers for uterine receptivity and participants in the molecular mechanisms occurring during embryonic implantation into the maternal endometrium (5, 6, 7). By translation from such models and building upon a literature of known expressed genes and proteins and uniquely expressed secretory proteins in human endometrium, the expression of several molecules has been found to be specifically and temporally expressed within and framing the window of implantation in humans (7), suggesting their functionality in the implantation process. The molecular dialogue that occurs between the endometrium and the implanting conceptus involves cell-cell and cell-extracellular matrix interactions, mediated by lectins, integrins, matrix degrading enzymes and their inhibitors, and a variety of growth factors and cytokines, their receptors and modulatory proteins. Of note are molecules that participate in attachment of an embryo to the maternal endometrial epithelium, including carbohydrate epitopes (e.g. H-type 1 antigen), heparan sulfate proteoglycan, mucins, integrins (especially {alpha}vß3, {alpha}4ß1), and the trophin-bystin/tastin complex (reviewed in Refs. 5 and 6). Molecules that participate in embryonic attachment to the epithelium and subsequent signaling between epithelium and stroma have been deduced from knockout studies of a given gene in mice that result in absence of embryonic attachment to the epithelium and loss of decidualization of the stroma. These molecules include leukemia inhibitor factor, the homeobox genes, HoxA-10 and HoxA-11, and cyclooxygenase 2 (reviewed in Ref. 7).

In the pregenomic era, a one-by-one approach has been useful to reveal select candidates for uterine receptivity. In the current study, we report global gene profiling, using high density oligonucleotide microarray technology, of human endometrium during the window of implantation. In addition, we have pursued, in well characterized human endometrial tissue and isolated endometrial stromal and epithelial cells, validation of select gene expression by RT-PCR and Northern analysis. This combined effort has resulted in the identification of genes, gene families, and signaling pathways that are candidates for uterine receptivity and molecular mechanisms underlying the process of human implantation. The endometrial signature of genes during the window of implantation provides an opportunity to design diagnostic (screening) tests for patients with infertility and endometrial disorders and for targeted drug discovery for treating implantation-based infertility, other endometrial disorders, and endometrial-based contraception.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Tissue specimens and cell culture
Tissues.
Endometrial biopsies were obtained from normally cycling women after informed consent, under an approved protocol by the Stanford University Committee on the Use of Human Subjects in Medical Research, and the Human Subjects Committees at the University of North Carolina, Vanderbilt University, and the University of California at San Francisco. All specimens were obtained in accordance with the Declaration of Helsinki. A total of 28 biopsy samples were obtained from two time points of the menstrual cycle and used in this study: 10 in the late proliferative phase (peak circulating E2 levels; cycle d 8–10), and 18 during the window of implantation [midsecretory phase (peak E2 and progesterone)], which were timed to the LH surge (LH + 8 to LH +10, where LH = 0 is the day of the LH surge). Timing to the LH surge assured sampling during the window of implantation. Of the 28 biopsies, 11 (4 in late proliferative phase and 7 window of implantation) were used for microarray studies, 5 secretory specimens were used exclusively for cell isolation and culture, and 12 were used for Northern analysis and RT-PCR validation (vide infra). Different samples were used for the microarrays and the validation studies. Subjects ranged in age between 28 and 39 yr of age, had regular menstrual cycles (26–35 d), were documented not to be pregnant, and had no history of endometriosis. Endometrial biopsies were performed with Pipelle catheters under sterile conditions, from the uterine fundus. A portion of each sample was processed for histologic confirmation, and the remainder was processed for cell culture or immediately frozen in liquid nitrogen for subsequent RNA isolation. Cycle stages for all specimens (proliferative and midsecretory) were histologically confirmed (8) independently by three observers: L.C.G., B.A.L., and an independent pathologist.

Cell culture.
Five midsecretory specimens were used for cell isolation and culture for this study. Tissue was subjected to collagenase (Sigma, St. Louis, MO) digestion, and stromal cells were separated from epithelium, as previously described (9). Initially, stromal cells were centrifuged and the resulting pellet was resuspended in DMEM/10% FBS. The cells were then preplated in 10-cm standard culture plates in DMEM/F12 media for 1 h at 37 C, and the media were then replaced with DMEM/10% FBS. Glands retained on the filter were backwashed into sterile tubes, washed with PBS three times, centrifuged and resuspended in MCDB-105. Endometrial stromal cells were plated and passaged in standard tissue culture plates at a density of 2–3 x 105/10 cm plate and cultured in phenol-red-free, high-glucose DMEM/MCDB-105 medium with 10% charcoal-stripped FBS, insulin (5 µg/ml), gentamicin, penicillin and streptomycin, as described (9). Stromal cells were used at passages 2–6 for these studies. Endometrial epithelial cells were plated in two chamber collagen type I-coated chamber slides (Costar, Cambridge, MA) and cultured in MEM{alpha} with 10% charcoal-stripped FBS at 37 C in 9% CO2 for up to 1 wk (10). Purity was established by vimentin and cytokeratin immunostaining. The culture medium was renewed every 2 d, and the cells were harvested for RNA analysis at the end of the culture period.

Gene expression profiling
RNA preparation/target preparation/array hybridization and scanning.
For microarray analysis, 4 late proliferative phase samples and 7 window of implantation samples were used. Each endometrial biopsy sample was processed individually for microarray hybridization (samples were not pooled) following the Affymetrix (Affymetrix, Santa Clara, CA) protocol. Poly(A)+ RNA was initially isolated from the tissue samples using Oligotex Direct mRNA isolation kits (QIAGEN, Valencia, CA), following the manufacturer’s instructions. Specimens (120–260 mg) yielded between 1–8 µg poly(A)+ RNA and the purity of isolated mRNAs was evaluated spectrophotometrically by the A260/A280 ratio. A T7-(deoxythymidine)24 oligo-primer was used for double-stranded cDNA synthesis by the Superscript Choice System (Life Technologies, Inc., Gaithersburg, MD). In vitro transcription was subsequently carried out with Enzo BioArray High Yield RNA T7 Transcript Labeling Kits (Enzo, Farmingdale, NY). Additional cRNA clean-up was performed using RNeasy spin columns (QIAGEN), before chemical fragmentation with 5x fragmentation buffer (200 mM Tris, pH 8.1; 500 mM KOAc; 150 mM MgOAc). After chemical fragmentation, biotinylated cRNAs were mixed with controls and were hybridized to Affymetrix Genechip Hu95A oligonucleotide microarrays [corresponding to 12,686 human genes and expressed sequence tags (ESTs)] on an Affymetrix fluidics station at the Stanford University School of Medicine Protein and Nucleic Acid (PAN) Facility. Fluorescent labeling and laser confocal scanning were conducted in the PAN Facility and generated the data for analysis.

Data analysis.
One of the most critical steps in microarray profiling experiments is accurate assessment of the expression ratios between the sample and the reference, because most subsequent analyses depend on the accuracy of these ratios. The observed signal is comprised of the true expression level with noise due to background and noise due to experimental variations from the probe preparation and hybridization efficiency. Due to variations in the hybridization and scanning processes, several approaches for data analysis have been devised to compensate for these differences. Two major steps are: 1) to eliminate weak expressions that are statistically too close to the background estimate to avoid the detrimental effects on the ratios; and 2) to adjust the expression of each gene by the overall expression of signals on a specific chip. In the current study the data were analyzed with GeneChip Analysis Suite version 4.01 (Affymetrix), GeneSpring version 4.0.4 (Silicon Genetics), and Microsoft Corp. Excel/Mac2001 software. Expression profile data were first prepared using GeneChip Microarray Analysis Suite and subsequently exported to GeneSpring for further analysis. The GeneSpring version 4.0.4 software allows rank-sum normalization and statistical analysis. Initially, within each hybridization, the 50th percentile of all measurements was used as a positive control, and each measurement for each gene was divided by this control. The bottom tenth percentile was used for background subtraction. Between different hybridization outputs/arrays, each gene was normalized to itself by making a synthetic positive control for that gene comprised of the median of the gene’s expression values over all samples of an experimental group, and dividing the measurements for that gene by this positive control, as per the manufacturer’s instructions. Mean values were then calculated among individual experimental groups for each gene probe-set, and between-group fold-change ratios [i.e. window of implantation (n = 4) : late proliferative phase (n = 7) ratios] were derived. A difference of 2-fold was applied to select up-regulated and down-regulated genes, as described (11, 12). Because the data were not normally distributed, nonparametric testing was also conducted using the Mann-Whitney U test to calculate P values, and applying P < 0.05 to assign statistical significance between the two groups, as described (11). To assess chip-to-chip variability, preliminary experiments were conducted in which RNA from one tissue sample was subjected to two independent hybridizations. Less than 2.7% of the total genes on the array showed more than 3-fold variation, providing a greater than 95% confidence level, consistent with the manufacturer’s claims for chip-to-chip variability (12, 13).

Validation of gene expression data
RT-PCR.
Genes of different expression fold changes were randomly selected for validation by RT-PCR and/or Northern analyses. Total RNA from cultured endometrial epithelial cells, stromal cells or whole endometrial tissue was isolated using Trizol (Life Technologies, Inc.) protocol, then treated with deoxyribonuclease (QIAGEN) and purified by RNeasy Spin Columns (QIAGEN). RT was first performed with Omniscript kit (QIAGEN) for 1 h at 37 C, followed by PCR in a 50-µl reaction volume with Taq polymerase (QIAGEN) and specific primer pairs using the Eppendorf Mastercycler Gradient (Brinkmann Instruments, Westbury, NY). The amplification cycle consisted of a hot start at 94 C for 2 min followed by 35 cycles of denaturation at 94 C for 1 min, annealing at 58 C for 1 min and extension at 72 C for 1 min. Specific primer pairs (Table 1Go) were synthesized by the PAN Facility, Stanford University School of Medicine, and were used at 25 pmol per reaction. Sequences were derived from public databases, and all PCR products were confirmed by the Stanford PAN Sequencing Facility. Subcloning by TA cloning into pGEM Teasy (Promega Corp., Madison, WI) or pDrive Cloning Vector (QIAGEN) were performed to generate specific probes for Northern analyses.


View this table:
[in this window]
[in a new window]
  		Table 1. Oligonucleotide primers with predicted respective PCR product sizes

 
Northern analysis.
Twelve endometrial biopsy samples were used for these studies, six from the late proliferative phase and six during the window of implantation. Total RNAs (10–20 µg) were electrophoresed on 1% formaldehyde agarose gels and transferred to nylon membranes for Northern analyses. Specific P32-labeled cDNA probes, ranging 400–900 bp, were generated using Ready-to-Go random primer kit (Amersham Pharmacia Biotech, Peapack, NJ) and 32{alpha}P-deoxy-CTP (NEN Life Science Products, Boston, MA). Membranes were prehybridized at 68 C for 30 min in ExpressHyb buffer (CLONTECH Laboratories, Inc., Palo Alto, CA) and hybridization carried out for another hour at 68 C using ExpressHyb buffer containing 1–2 x 106 cpm/ml of labeled probe. Washing was subsequently carried out according to the manufacturers’ instructions. Membranes were exposed to Kodak (Rochester, NY) MS x-ray films, and densitometry performed with Bio-Rad Laboratories, Inc. (Hercules, CA) GS-710 Imaging Densitometer and analyzed by its accompanied software Quantity One, version 4.0.2. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) mRNA densitometric readings were used to normalize signals for genes expressed on a given Northern blot. These normalized data were used to calculate relative mRNA expression intensities between the window of implantation and proliferative phase endometrium samples on the same blot. Mean values of relative expression intensities from different Northern blots were used for final data presentation. Stripping and reprobing were performed using the same membranes.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Data analysis
The data were analyzed with GeneChip Analysis Suite version 4.01, GeneSpring version 4.0.4, and Microsoft Corp. Excel/Mac2001 software, as described in Materials and Methods. Figure 1Go shows the scatter plot of the normalized data for all genes and all experiments for samples in the proliferative phase and the secretory phase (window of implantation), underscoring that the data are not normally distributed. Fold-change ratios between groups (i.e. window of implantation:late proliferative phase ratios) were subsequently derived, and a difference of 2-fold, a generally adopted fold-change difference for oligonucleotide microarray profile analysis (11, 12), was applied to select up-regulated and down-regulated genes. Nonparametric testing was further applied, using a P value of 0.05 to identify statistical significance between the two groups (13). Figure 2GoGo, panels A and B, shows the scatter plots of genes up- and down-regulated by at least 2-fold before nonparametric testing, and panels C and D demonstrate individual genes up- and down-regulated by at least 2-fold after nonparametric testing. With this strategy, we identified, during the window of implantation, 156 genes that were significantly up-regulated, of which 40 were ESTs, and 377 genes that were significantly down-regulated, of which 153 were ESTs. Tables 2GoGo and 3GoGoGoGo show, in descending order, respectively, the fold increase and fold decrease, the P values (P < 0.05), and the GenBank accession numbers for the 116 specifically up-regulated genes and the 224 down-regulated genes in the window of implantation in human endometrium, compared with the late proliferative phase, according to clustering assignments (vide infra).



View larger version (16K):
[in this window]
[in a new window]
  		Figure 1. Scatter plots of the composite normalized data, from all microarray experiments, were segregated into two groups: the late proliferative phase (P) and the window of implantation during the secretory phase (S Nc). Relative intensity of each gene, after normalization, is represented by the vertical axis. The colors are based on relative intensity of each gene, red being greater than the value of one and blue being less than one, during the proliferative phase.

 


View larger version (12K):
[in this window]
[in a new window]
  		Figure 2A. Genes up- and down-regulated by at least 2-fold in the window of implantation vs. the late proliferative phase. A and B, Scatter plots of genes up-regulated (A) and down-regulated (B) before nonparametric testing. C and D, Line graphs of genes up-regulated (C) and down-regulated (D) after nonparametric testing. Relative intensity of each gene, after normalization, is represented by the vertical axis. The colors are based on relative intensity of each gene, red being greater than the value of one and blue being less than one, during the proliferative phase.

 


View larger version (39K):
[in this window]
[in a new window]
  		Figure 2B. Continued

 

View this table:
[in this window]
[in a new window]
  		Table 2A. Genes up-regulated during the window of implantation

 

View this table:
[in this window]
[in a new window]
  		Table 2B. Continued

 

View this table:
[in this window]
[in a new window]
  		Table 3A. Genes down-regulated during the window of implantation

 

View this table:
[in this window]
[in a new window]
  		Table 3B. Continued

 

View this table:
[in this window]
[in a new window]
  		Table 3C. Continued

 

View this table:
[in this window]
[in a new window]
  		Table 3D. Continued

 
Clustering
The stringent data filtering for significant and consistent changes permitted identification of biologically relevant gene clustering in human endometrium during the window of implantation vs. the late proliferative phase. We performed unsupervised cluster analysis (14), based on National Center for Biotechnology Information/Entrez/Online Mendelian Inheritance in Man database search, which allowed grouping of genes into several categories (Tables 2GoGo and 3GoGoGoGo). The most markedly up-regulated genes (categories in descending order of maximal fold change) include those involved in cholesterol trafficking and transport [apolipoprotein (Apo)E and D], PG biosynthesis and action (PLA2 and the PGE2 receptor), proteoglycan synthesis (glucuronyltransferase I), and a variety of secretory proteins, including glycodelin (pregnancy-associated endometrial {alpha}2 globulin), mammaglobin (a member of the uteroglobin family), members of the Wnt regulation pathway [Dickkopf-1 (Dkk-1)], IGFBP family, and TGF-ß superfamily. Additional genes were up-regulated, including G0S2 (a cell cycle switch protein), several genes involved in signal transduction, nitric oxide metabolism (arginase II), and extracellular matrix components/cell adhesion molecules, including osteopontin and laminin subunits. Also, of note are the marked up-regulation of genes for neuromodulator synthesis/receptors (GABAA receptor {pi} subunit), immune modulators [e.g. natural killer-associated transcript 2, members of the complement family, and interferon-induced genes (interferon {gamma}-inducible indoleamine 2,3-dioxygenase; IDO], genes involved in detoxification (several types of metallothioneins and glutathione peroxidase), phospholipid binding proteins (annexins), as well as some proteases, transcription factors and structural/cytoskeletal proteins. Among several gene families not heretofore known to exist in endometrium are members of water and ion transport that are common to the gastrointestinal epithelial mucosa and other mucosal surfaces [e.g. Clostridia Perfringens Enterotoxin (CPE) 1 receptor (CPE-1R) and the sulfonylurea receptor (K+ ion channel)]. Several genes for other cellular functions were also up-regulated.

The most abundantly down-regulated genes involved secretory proteins, including ITF (a member of a family of proteins that maintains intestinal luminal epithelial cell integrity) and proteases, such as matrilysin [matrix metalloproteinase 7 (MMP)], dipeptidyl aminopeptidase, and carboxypeptidase E. Also markedly down-regulated genes included those for G protein-coupled receptor signaling: G protein-coupled receptor kinase and G protein {gamma}-11 subunit. Also, marked down-regulation of calcineurin (a protein involved in Ca2+ signaling) was observed, as well as some members of the Wnt pathway [frizzed related protein (FrpHE) and secreted frizzled-related protein], genes for TGF-ß signaling (homolog of the Drosophila mothers against decapentaplegic 1), the PPAR and members of the fibroblast growth factor receptor family. Select extracellular matrix/cell adhesion molecules were down-regulated, including {alpha}-1 type XVI collagen and the extracellular matrix protein, tenascin-C. Numerous genes corresponding to transcription factors were found to be down-regulated during the implantation window, including homeobox genes (muscle segment homeobox-2 and homeobox-7), Kruppel family of zinc finger proteins, the early response gene protein [erythroblastosis virus E26 oncogene (ets)-related gene], several proto-oncogenes (c-fos, B lymphoma Mo-MLV insertion region, and others), apoptosis/inhibitors (TNF-related apoptosis-inducing ligand receptor 2), and immune modulators (MHC class II subunits). Of interest is the observed down-regulation of vasoactive substances [endothelin 3 and vascular endothelial growth factor (VEGF)], several cell cycle regulators, and genes whose products have relevance to steroid hormone actions [putative progesterone binding protein/PR membrane component 1 (PGRMC1) and steroid receptor coactivator 1e]. Down-regulation of several transporters and calcium channel subunits, structural and cytoskeletal proteins, and ion binding proteins were observed, as well as genes for other cellular functions.

Because clinical endometrial biopsy samples contain a mixture of different cell populations, including glandular and surface epithelial cells, stromal cells, and vascular, smooth muscle, and blood cell components, it is anticipated that many genes and gene families participating in different processes would be represented in the microarray data. In addition, previously documented genes in these cellular components would be anticipated to be detected in the GeneChip analysis. Reassuringly, many genes known to be significantly up-regulated in human endometrium during the secretory phase were up-regulated more than 2-fold and included (Table 2GoGo): pregnancy-associated endometrial {alpha}2-globulin [glycodelin, an exclusively endometrial epithelial cell product predominantly expressed in the secretory phase of the cycle (15)]; IGFBP-2, which is exclusively an endometrial stromal cell product up-regulated in the secretory phase (16); osteopontin, an endometrial epithelial-specific protein (17); PGE2 receptor; TGFß-type II receptor; and IL-15 (18, 19) and its receptor (20, 21). Others (reviewed in Refs. 5 and 6), such as IGF-II, plasminogen activator inhibitor, urokinase receptor, tissue inhibitor of metalloproteinase-3, fibroblast growth factor (FGF)-6, and FGF-8, and IGFBP-1 were also up-regulated, although they did not reach statistical significance by nonparametric testing. For the genes down-regulated in the window of implantation (Table 3GoGoGoGo), significant down-regulation of matrilysin (MMP-7) and tenascin-C was detected, consistent with previous studies demonstrating their decreased expression in secretory, compared with proliferative phase, endometrium (22, 23).

Validation of gene expression
We validated expression of select up-regulated and down-regulated genes using two approaches: RT-PCR and Northern analyses with RNAs from late proliferative and window of implantation endometrial biopsy tissue samples and RT-PCR using RNA isolated from cultured human endometrial glandular and stromal cells. The results are shown in Figs. 3–5GoGoGo. For the RT-PCR studies, the primer sets are shown in Table 1Go. Although quantitative PCR was not performed, the RT-PCR data in Fig. 3Go demonstrate clearly in the implantation window compared with late proliferative phase endometrium, up-regulation of IGFBP-1, glycodelin, CPE-1R, Dkk-1, GABAAreceptor {pi} subunit, mammaglobin, and ApoD, and down-regulation of PGRMC1, frpHE, matrilysin, and intestinal trefoil factor (ITF). These data are consistent with the observations from the microarray data (Tables 2GoGo and 3GoGoGoGo).



View larger version (30K):
[in this window]
[in a new window]
  		Figure 3. Validation of selected genes more than 2-fold up- or down-regulated during the window of implantation in human endometrium by RT-PCR. Endometrial biopsy samples from late proliferative phase (n = 3) and the window of implantation (n = 3) were processed for total RNA, and representative results are shown. RT-PCR was conducted with specific primer sets shown in Table 1Go, using the samples from the proliferative phase (lanes a) or window of implantation (lanes b). Appropriate-sized products corresponding to GAPDH (lanes 1a, 1b), IGFBP-1 (lanes 2a, 2b), glycodelin (lanes 3a, 3b), CPE-1R (lanes 4a, 4b), Dkk-1 (lanes 5a, 5b), GABAA R {pi} (lanes 6a, 6b), mammaglobin (lanes 7a, 7b), ApoD (lanes 8a, 8b), PGRMC-1 (lanes 9a, 9b), FrpHE (lanes 10a, 10b), matrilysin (lanes 11a, 11b), and ITF (lanes 12a, 12b) are shown.

 


View larger version (21K):
[in this window]
[in a new window]
  		Figure 4. Northern analysis demonstrating up-regulation of Dkk-1, IGFBP-1, GABAA R {pi} subunit, glycodelin, and down-regulation of PGRMC-1, matrilysin and FrpHE in the secretory phase (implantation window, lane c), compared with the proliferative phase (lane b). Placental basal plate with decidua (lane a) is shown as a positive control for Dkk-1 and IGFBP-1 on the left panel. GAPDH hybridization of respective blots are shown for comparison.

 


View larger version (17K):
[in this window]
[in a new window]
  		Figure 5. Expression of selected genes in cultured human endometrial epithelial (A) and stromal (B) cells by RT-PCR. A, Lane 1, GAPDH (control); lane 2, Glycodelin; lane 3, CPE-1R; lane 4, Dkk-1; lane 5, GABAA R {pi} subunit; lane 6, Mammaglobin; lane 7, Matrilysin; lane 8, ITF; lane 9, PGRMC-1. B, Demonstrates RT-PCR products using endometrial stromal cells nondecidualized (lanes "a") or decidualized (lanes "b") with progesterone after E2 priming, as described in Material and Methods. Lanes 1a and 1b, GAPDH (control); lanes 2a and 2b, IGFBP-1; lanes 3a and 3b, CPE-1R; lanes 4a and 4b, Dkk-1; lanes 5a and 5b, ApoD; lanes 6a and 6b, FrpHE. Experiments were conducted with isolated cells from five different samples. Representative results are shown.

 
Northern analyses were conducted to validate further select changes in gene expression in the implantation window vs. the late proliferative phase, and a representative set of Northern blots is demonstrated in Fig. 4Go. Densitometric analyses were conducted and signals were normalized to GAPDH for samples on the same blots. Ratios between the window of implantation vs. proliferative endometrium gene expression were calculated for each mRNA on each blot. Mean values of relative expression intensities from n = 3 different blots were then calculated. These data demonstrate up-regulation of Dkk-1 (3.2-fold), IGFBP-1 (2.8-fold), GABAA receptor {pi} subunit (24.6-fold), and glycodelin (3.1-fold), and the down-regulation of PGRMC-1 (1.3-fold), matrilysin (20.0-fold), and FrpHE, 50-fold. The data are consistent with and validate those obtained through the microarray expression profiling analysis, although the levels of fold-change are not the same as in the microarray analysis (Tables 2GoGo and 3GoGoGoGo) and would not be expected to be the same (see Discussion).

RT-PCR experiments using RNA from cultured endometrial epithelial and stromal cells and the primers listed in Table 1Go, revealed the following (Fig. 5Go): PCR products corresponding to glycodelin (positive control), the CPE-1R, Dkk-1, the GABAA receptor {pi} subunit, mammaglobin, matrilysin, ITF, and PGRMC-1 were all expressed in human endometrial epithelial cells (panel A), demonstrating their expression in this cell type in human endometrium. With cultured human endometrial stromal cells (Fig. 5BGo), up-regulation of the CPE-1R, Dkk-1, ApoD, and IGFBP-1 (positive control) and down-regulation of frpHE upon in vitro decidualization with progesterone after E2 priming were observed. The GABAA receptor {pi} subunit, mammaglobin, matrilysin, ITF, and PGRMC-1 were not detected in isolated and cultured endometrial stromal cells before or after decidualization (data not shown).


   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
Molecular mechanisms that involve apposition, attachment, and intrusion of an implanting embryo into human endometrium are beginning to be appreciated. Most of what is believed to occur during human implantation is derived from animal models that have been invaluable, especially when the reproductive phenotype involves implantation failure (7). A limiting factor in research with human endometrium has been the availability of appropriately characterized clinical specimens. Herein, we have presented global gene profiling of well characterized human endometrial biopsy samples that were obtained during the window of implantation, defined by timing to the LH surge and histologically confirmed (8). About one-third of the samples we collected had to be discarded because their histology was not consistent with normal temporal development in the cycles in which they were obtained. This observation underscores the need for precise histologic confirmation of endometrial samples before analysis. The approach taken in this study also demonstrates that reproducible and sensitive experimental methodology for global gene analysis can be applied to small quantities of human endometrial tissue. Similar approaches have been successfully pursued with whole tissues [e.g. developing rat kidney (24)] and breast cancer biopsies (25). The current study used high-density oligonucleotide microarray expression profiling that allowed profiling and interrogation of expression of 12,686 full-length genes and ESTs in human endometrial biopsies. The microarray technologies and the data presented herein highly support the use of this powerful approach to investigate molecular candidates involved in human uterine receptivity. Results from the human genome project suggest that we have interrogated about one third to one half of existing human genes, and thus, investigation of additional genes, as well as their validation, present a formidable task and await further investigation.

Endometrial biopsy specimens contain several cell populations and may differ in their complement of such populations. This heterogeneity may contribute to differences observed in relative expression of select genes between the implantation window and the late proliferative phase as assessed by the microarray approach vs. Northern analysis or RT-PCR approaches. In addition, because different samples were used for the microarrays and the validation studies, subject-to-subject biologic variation in samples obtained in the same phase of would be anticipated. Also, in the microarray analysis, the mean of an individual gene readout from the samples in the window of implantation was compared with the mean of the same gene readout of the proliferative phase samples; whereas, in calculating the fold-change for a given gene analyzed by Northern analysis, the mean of the densitometric OD readings were calculated after normalization to GAPDH. We speculate that differences in the fold-change values between the two methodologies may also be due to the lower abundance of specific mRNAs in relation to the highly abundant GAPDH mRNA, especially because the microarray profile represents true abundance of each mRNA species globally within the tissue, whereas Northern analysis reflects mRNAs of higher abundance and is poor in detecting very low abundance transcripts.

While this heterogeneity among samples may influence the relative expression of some genes, we confirmed previously documented genes within the window of implantation, such as the endometrial glandular-specific glycodelin and the stromal cell-specific, IGFBP-1 and IGFBP-2. Other molecules, such as the PGE2 receptor, IL-15, and the TGF-ß type II receptor have all been reported to be up-regulated in human endometrium during the secretory phase in various cell types (5, 6, 7), further validating the approach taken herein. Down-regulation of matrix metalloproteinase-7 (matrilysin) has been demonstrated previously (22), as has the down-regulation of tenascin-C, a multifunctional extracellular matrix glycoprotein that is regulated by multiple soluble factors, integrins, and mechanical forces, and known to be highly expressed in the proliferative phase of the menstrual cycle compared with the secretory phase (23). While the data presented contribute to the molecular signature of the endometrium that defines the state of receptivity to embryonic implantation, localization of cell type expression for select genes is clearly needed and is underway in our laboratories. Also, while the validation studies presented herein support cell-specific expression for a few, selected genes and validate in a limited fashion the microarray data, they do not represent the full spectrum of cell types in the endometrium, underscoring the need for in situ hybridization studies and subsequent protein demonstration. In addition, endometrial proteins that require posttranslational modifications for their activity are not revealed by gene profiling techniques and require alternative methods of investigation.

The choice to compare gene expression profiles in the window of implantation to the late proliferative phase was made to focus on the comparison of genes expressed during peak exposure of the endometrium to E2 and progesterone (window of implantation) vs. peak E2 (late proliferative phase). While many of the genes are known to be regulated by progesterone directly, e.g. glycodelin, IGFBP-1 and tissue inhibitor of metalloproteinase-3, regulation of others during the window of implantation likely derive from progesterone-induced (or suppressed) paracrine products that are mediators of the progesterone response. The finding of unique gene families, not previously known to be expressed in human endometrium or to be regulated by progesterone provides new avenues of investigation, which is an advantage of this unbiased technique and which transcends the goals of the current investigation to other fields. In addition, why more genes are down-regulated than up-regulated is an interesting observation, and we speculate that because during the implantation window, compared with the late proliferative phase, ERs are down-regulated in endometrial epithelial cells, genes that were up-regulated by E2 in the proliferative phase are now down-regulated due to the loss of E2 action. In addition, direct down-regulation by progesterone and multiple progestomedins during the implantation window may result in more down-regulated genes compared with up-regulated genes.

Up-regulated genes
Several genes and gene families that are up-regulated in the window of implantation warrant further discussion. ApoE is the most abundantly (100-fold) up-regulated gene in the window of implantation. It binds hydrophobic molecules and is important in cholesterol transport and trafficking (26). Local production of ApoE in steroidogenic tissues, particularly the ovary, has been reported, through mechanisms involving the low density lipoprotein receptor family (27). The high expression of ApoE (and ApoD) in the endometrium suggests an important role for it in cholesterol transport in this tissue, perhaps for steroid hormone biosynthesis or steroid hormone binding.

PLA2, the second most abundantly (18-fold) up-regulated gene in the window of implantation, belongs to a family of enzymes (secreted, membrane bound, Ca2+ dependent) that catalyze the hydrolysis of membrane glycerophopholipids, resulting in the release and metabolism of arachadonic acid and generation of lipid signals: platelet-activating factor, lysophosphatidic acid, PG and leukotrienes (see Ref. 28 for review). The importance of PG action during the window of implantation is underscored by the concomitant (4-fold) up-regulation of the PGE2 receptor (Table 2GoGo). PLA2 is also involved in calcium influx into nonexcitable cells (29) and in the modulation of TNF-{alpha} and IL-1ß-induced NF-{kappa}B activation (30), which is important in endometrial function (31). PGs are important for vascular permeability and endometrial decidualization (see Refs. 5 and 6 for reviews). Further definition of mechanisms underlying PLA2 and PG actions during the implantation window are major challenges for further investigation.

Of interest in the implantation window is the finding of expression and marked (12-fold) up-regulation of mammaglobin, classically known as a breast-specific uteroglobin family member (32). Mammaglobin B has been identified in rat uterus (33), and uteroglobin has been well characterized in rabbit and human endometrium and is known to be regulated by progesterone (reviewed in Ref. 34). Several properties of members of the uteroglobin family have been identified, including serving as a substrate for transglutaminases and acting as an antiinflammatory agent by inhibiting PLA2 (35). It is striking that both PLA2 and mammaglobin, a putative inhibitor of this enzyme, are so markedly up-regulated during the window of implantation in human endometrium. Of course, mammaglobin, a member of the secretory lipophilin family of proteins that are prominent in glandular secretions and hormone-responsive tissues (36), may have other functions, yet to be identified in the implantation window in human endometrium.

Another inhibitor of PLA2 is annexin IV, a member of the annexins or lipocortin family of calcium-dependent phospholipid-binding proteins (37). Annexin IV, also known as placental anticoagulant protein II, has anticoagulant activity, as well. The up-regulation of annexin IV, annexin II, and lipocortin-2 in the implantation window underscores the importance of regulating PLA2 activity and maintaining an environment for anticoagulation during implantation.

Pregnancy-associated endometrial {alpha}2 globulin, also known as glycodelin, is an endometrial epithelial-specific protein and is up-regulated in human endometrium during the periimplantation period and in the late secretory phase (15). Data in this study support these well established observations. Glycodelin belongs to a family of lipocalins that participate in regulation of the immune response that also includes {alpha}1 microglobulin and the {gamma} chain of complement factor 8 (38). The lipocalins typically bind small hydrophobic molecules, like retinol and RA, although glycodelin does not bind these molecules (15).

The finding of members of the Wnt family is surprising. Of particular interest is the marked up-regulation of Dickkopf-1 (an inhibitor of Wnt signaling) and of LRP (low density lipoprotein receptor-like protein), and the down-regulation of frizzled related protein (FrpHE), also an inhibitor of Wnt signaling (39). Dickkopf-1 inhibits Wnt signaling by binding LRP5/6 (40), and FrHPE inhibits Wnt action by competitive binding to Wnt ligand(s). Wnt7A (-/-) null mice are infertile and have complete absence of uterine glands and a reduction in mesenchymally derived uterine stroma (41). We have localized Wnt7A exclusively to epithelium and frizzled receptor to epithelium and stroma in human endometrium (Tulac, S., L. C. Kao, and L. C. Giudice, manuscript in preparation). It is possible that the Wnt family may play a role in epithelial-embryo and/or epithelial-stromal interactions and thus in uterine receptivity. The role of the Wnt family in human endometrium and implantation is currently under investigation in our laboratories.

Proteoglycans, extracellular matrix (ECM) proteins, and cell surface glycoproteins function in epithelial-embryonic interactions. The ECM is also a reserve of many peptide growth factors and angiogenesis modulators, underscoring the importance of its regulation in events occurring in the endometrium. Of particular interest is the marked (16-fold) up-regulation of glucyronyltransferase I, a central enzyme in heparan/chondroitin sulfate and other proteoglycan biosynthesis (42, 43). Also, significantly up-regulated (8-fold) during the window of implantation is the ECM protein, osteopontin, known to be progesterone-regulated and up-regulated in the midsecretory phase in human endometrium (17). Osteopontin has been postulated to bridge embryo-epithelial attachment. Also, we found up-regulation of laminin B, and proline-rich protein, an ECM protein commonly found in intestinal epithelium (44).

A number of genes involved in immune modulation deserve special mention, although their cellular expression has not yet been determined. These include (also see Table 2GoGo): natural killer-associated transcript 2, members of the complement family (including adipsin, which is the same as complement D (45), decay-accelerating factor, and complement 1r), IL-15 (18, 19) and its receptor (20, 21), NKG5 [an NK and T-cell specific gene strongly up-regulated upon cell activation (46)], interferon {gamma}-inducible IDO, interferon regulatory factor 5, and lymphotaxin/SCM1ß (expressed in NK cells; Ref. 47). Some of these immune modulators are well characterized in human endometrium and have functions related to NK cell differentiation [e.g. IL-15 (18, 19, 20, 21, 48)] and complement action (49, 50, 51, 52), and may play key roles in immune tolerance of an implanting embryo [e.g. IDO may have a role in the prevention of allogeneic fetal rejection by tryptophan catabolism (53)]. The impressive regulation of immune modulators underscores the need for further investigation into this important group of gene families in the implantation process, especially in view of the controversies currently surrounding immune-based therapies for some infertility patients.

Of interest are transport proteins for water and ions that are common to kidney and gastrointestinal epithelium (Table 2GoGo). It is reasonable that mechanisms are conserved for water and ion transport—whether they be in the gut, the kidney, or the endometrium. Finding expression of these transporters and their marked up-regulation during the implantation window likely reflects the importance of water (and ion) shifts that take place across the epithelium and the importance of endometrial stromal edema that occur during the window of implantation (8, 54). The gene for the CPE-1R, for example, was up-regulated 4-fold in the implantation window. This receptor is a tight junction protein component that forms pores for water transport in the gut (55), and in response to CPE results in massive water shifts into the intestinal lumen. It has been found to be abundantly expressed in the gastrointestinal tract and in the uterus (56). Whether this receptor is involved in water transport that occurs during the midsecretory phase, is unknown. Further, its endogenous ligand in the endometrium (and its true function) await definition. The finding of the CPE-1R and of a membrane protein potassium channel—the sulfonyl urea receptor (57), open new avenues of investigation in endometrial biology, focusing on, e.g. signaling from an embryo involving ion fluxes, with appropriate channels in place for such interactions.

Genes for members of the metallothionein family of proteins that are involved in detoxification and zinc binding are up-regulated 2.3- to 5.8-fold during the implantation window. In zinc deficiency and in metallothionein knockout mice, there is an alteration of Th1 and Th2 cytokines (58). Because the ratio of Th1 to Th2 is believed to be important for successful implantation in humans (see Ref. 59 for review), this gene family may provide a mechanism to regulate the immune balance for embryonic tolerance during implantation. Also of note is the up-regulation of genes governing intracellular Ca2+ signaling and Ca2+ homeostasis [annexin II (60)], underscoring the importance of Ca2+ in the implantation window (5, 6). Genes whose products are involved in G protein-coupled receptor desensitization, e.g. ß-arrestin, ß-adaptin, and clathrin (61), are up-regulated, supporting attenuated G protein-coupled receptor signaling in the implantation window. Cyclophilins are up-regulated during the implantation window, and since they bind with Hsp 90 to inactivate steroid hormone receptors (62), they may contribute to the observed down-regulation of the estrogen receptor in endometrial epithelium between cycle d 20–24 (63).

Up-regulation of the GABAA receptor {pi} subunit and documentation of its epithelial origin in human endometrium during the implantation window (Table 2GoGo and Figs. 3–5GoGoGo) raise the issue of the role of neurotransmitters and of progesterone metabolism in this tissue. The GABAA receptor has been reported in rat uterus (64) and is important in the binding of reduced metabolites of progesterone in this tissue. Whether this is important in human endometrium remains to be determined. The observations of up-regulation of monoamine oxidase (important in norepinephrine synthesis) and diamine oxidase, well recognized in human endometrium (28, 65, 66, 67), underscore the need to reach beyond conventional thinking about mechanisms operating in endometrial development and perhaps embryo-endometrial interactions. Cellular localization of these genes and their ligands (e.g. for the GABAA receptor {pi} subunit) clearly need further definition. However, these findings and our recent findings of neuromodulators and their receptors in decidualized human endometrial stromal cells (68) underscore further consideration of neurotransmitter receptors participating in signals from an implanting embryo during nidation into the endometrium. Some of these receptors may have other functions, as has been shown for dopamine and morphine, stimulating nitric oxide production by human endometrial glandular epithelial cells in culture (69).

Down-regulated genes
ITF, a member of a family of secreted proteins that are expressed in the epithelial mucosal layer of the small intestine and colon, is the most markedly (50-fold) down-regulated gene in human endometrium during the window of implantation (Table 3GoGoGoGo). Studies with ITF null (-/-) mice support a central role for ITF in maintenance and repair of the intestinal mucosa (70). Whether an analogous role is present in endometrium warrants further investigation.

Other markedly down-regulated genes include some that are involved in G protein-coupled receptor signaling: G protein-coupled receptor kinase (23-fold reduction); HM145 (a G protein-coupled receptor for leukocyte chemoattractants (71), 11-fold reduction), and the G protein {gamma} 11 subunit (4.7-fold reduction). Down-regulation of this signaling pathway raises questions of identifying ligand/receptor complexes using this pathway and why their down-regulation is important during the implantation window. This is notable, especially because this apparently is coordinated with up-regulation of G protein receptor inhibitory factors (vide supra).

Several peptidases were also found to be down-regulated during the implantation window (Table 3GoGoGoGo), including, matrilysin (24-fold), dipeptidyl amino peptidase (10-fold), carboxypeptidase E (9.7-fold), and cathepsin F (3-fold), suggesting that proteolysis is minimized during this part of the menstrual cycle. As has been shown for MMPs (reviewed in Ref. 72), inhibition of MMPs may be critical to the maintenance of endometrial tissue architecture, very important during the implantation window. Dipeptidyl amino peptidase is a brush-border membrane-bound enzyme in the kidney proximal tubule and has been implicated in regulation of the biologic activity of multiple hormones and chemokines. Carboxypeptidase E is a regulated secretory pathway sorting receptor which regulates hormone, neuropeptide, and granin secretion in a calcium-dependent manner, important in prohormone processing, including proinsulin and neurotransmitters (73). Down-regulation of these enzymes may be part of a local control mechanism for regulating peptide activity within the endometrium.

Several other genes were also markedly down-regulated, including MSX-2 (a homeobox gene, 9-fold), genes involved in calcium and ion transport, and calcineurin, a protein involved in Ca2+ signaling (7.5-fold). Calcineurin is important in the activation of T cells (74). Antigen recognition by T cell receptors initiates signal transduction resulting in activation of tyrosine kinases, followed by PLC phosphorylation. This causes phosphatidyl inositol phosphate phosphorylation to phosphatidyl inositol phosphate 3, elevating intracellular Ca2+ and 1,2-diacylglycerol. Through the increased level of free Ca2+, a complex of calmodulin and calcineurin is formed. Calcineurin is a Ca-/calmodulin-dependent ser-thr phosphatase and dephosphorylates the nuclear factor of activating T cells (NF-AT). In the dephosphorylated form, NF-AT crosses into to the nucleus to function as a transcription activator for IL-2 expression. Down-regulation of calcineurin in endometrium would suggest limitation of NF-AT activation in this tissue. In addition, several transcription factors are down-regulated. Of note is the erg protein, a member of the ets family, important in regulation of extracellular matrix (75). With the dynamic changes in the extracellular matrix that occur in endometrium during the window of implantation and during early pregnancy, ets family members may play an important role.

Semaphorin E and semaphorin III family homologs were found to be down-regulated (6- and 3-fold, respectively) during the implantation window. Semaphorin III, interacting with its receptor, can result in either chemorepulsion or chemoattraction of developing axons, depending on levels of cellular cGMP (76). Finding the semaphorins and neurotransmitter receptors, as described above, suggests that perhaps we should be looking at other systems, such as ion signaling and chemoattractants/repellants for mechanisms to guide an embryo within the endometrium, analogous to neurotransmitter and semaphorin action in the neuronal system.

Of note also is down-regulation during the implantation window of the vasoactive factor, endothelin 3, and the angiogenic factor, VEGF (Table 3GoGoGoGo). Minimizing vasoconstriction is teleologically sound during a period that requires enhanced blood flow to the conceptus. Why VEGF is down-regulated is not clear, and conflicting reports have been reported on cyclic variations of this angiogenic factor in human endometrium (see Ref. 77 for review). However, Semaphorin III and VEGF compete for the same receptor, neuropilin-1 and this interaction results in inhibition of aortic endothelial cell migration (78). Interactions between the angiogenic system and the neuronal guidance system suggest potential new mechanisms for regulation of cellular motility in the endometrium during the implantation window, if indeed this extrapolation can be made.

The current study opens new conceptual approaches to mechanisms involved in the steroid hormone-dependent differentiation of the endometrium in the secretory phase of the menstrual cycle and mechanisms underlying endometrial development optimal for embryonic implantation and for embryo-endometrial interactions. While much validation of cell-specific expression of various genes described herein remains to be determined, the classes of molecules described herein support the following model. As an embryo attaches to the endometrial epithelium, bridging to cell surface carbohydrates and proteins is important, and mechanisms must be in place in the maternal endometrium for synthesis of these molecules. Once attachment occurs, a set of mechanisms is likely to be put into motion for endometrial-stromal interactions, intrusion of the trophoblast into the stromal compartment, and guidance of the trophoblast to the maternal spiral arteries, while maintaining integrity of the ECM and anticoagulation. It is envisioned that embryo-endometrial interactions involve ion transport and signaling through paracrine mechanisms via growth factors and cytokine families, as well as adaptation of guidance mechanisms similar to those used in angiogenesis and neuronal migration to target the trophoblast through the stroma to reach to the maternal vasculature. The immune system must facilitate tolerance of the implanting allograph and other protective mechanisms (e.g antibacterial, detoxification) are likely to be important to maximize viability of the implanting conceptus. While this model is rudimentary at best, it provides a framework for the role of the genes identified in this study in these processes. It is important to note that despite the anticipated interactions between the endometrium and the conceptus, based on gene expression in the endometrium during the implantation window described herein, the microarray approach provides a static snapshot of gene expression in the endometrium in the absence of an embryo and does not reveal the dynamic dialog that occurs minute-to-minute during embryonic implantation between the endometrium and the embryo. Nonetheless, it does provide insight into the molecular pathways, molecular signals, and physiologic processing that await an embryo should nidation occur.

Validation of functions for genes in the window of implantation will derive, in the future, from animal models of homologous recombination and gene knockouts, transgenic mice, studies in nonhuman primates, and other species whose endometrium and implantation processes are similar to those in humans, and further studies in human endometrial disorders related to implantation-based infertility. We believe that the current study provides the basis for defining markers of uterine receptivity during the window of implantation in human endometrium. Recent applications of global gene profiling relevant to implantation include a study by Reece et al. (79) in which uterine genes and gene families were characterized in mice during implantation in a variety of pregnancy models, and by Aronow et al. (80) on genes involved in human trophoblast differentiation. Information from these studies and the current study in human endometrium should further advance our knowledge about implantation in humans. However, just because a gene is regulated in the implantation window does not necessarily mean that it is important in the implantation process, and differences between coincidence and function await further validation.

The data presented herein offer the opportunity to develop an endometrial database of genes expressed during the window of implantation. The current study validates using microarray technology to investigate global changes in gene expression in human endometrium and can be extrapolated to defining the genetic profiles during the proliferative phase, periovulatory phase, and during the late secretory phase in the absence of implantation and in preparation for menstrual desquamation. Such an approach sets the stage for further investigation of global changes in gene expression in disorders of the endometrium, including implantation-related infertility (as in women with endometriosis), evaluation of the endometrium for normalcy in women with hyperandrogenic disorders, in normovulatory women in response to therapeutics in which the endometrium is targeted (or as a side effect of other therapies), as well as endometrial hyperplasia and endometrial cancers. Finally, this study sets the stage to develop a screen for candidate genes in patients with infertility and for targeted drug discovery for enhancing (or inhibiting) implantation for infertility treatment (or contraception).


   Acknowledgments
 


   Footnotes
 
This research was supported as a Collaborative Research Initiative of the Specialized Cooperative Centers Program in Reproduction Research through cooperative agreement U54 HD31398 (to L.C.G.) and the NIH Office of Women’s Health Research (to L.C.G., B.A.L., K.O., R.N.T.), the NIH Women’s Reproductive Health Research Career Development Program (to L.C.K.), and the German Research Foundation [Deutsche Forschungsgemeinschaft GE 1173/1-1 (to A.G.)].

Abbreviations: Apo, Apoliprotien; CPE, Clostridia Perfringens Enterotoxin; CPE-1R, Clostridia Perfringens Enterotoxin 1 receptor; Dkk-1, Dickkopf-1; ECM, extracellular matrix; ESTs, expressed sequence tags; ets, erythroblastosis virus E26 oncogene; FGF, fibroblast growth factor; FrpHE, frizzled-related protein; GABAAR, {gamma} aminobutyric acid A receptor {pi} subunit; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IDO, interferon {gamma}-inducible indoleamine 2,3-dioxygenase; IGFBP, IGF binding protein; ITF, intestinal trefoil factor; MHC, major histocompatibility complex; MMP-7, matrix metalloproteinase 7; NF-AT, nuclear factor of activating T cells; PAN, Protein and Nucleic Acid Facility; PGRMC1, PR membrane component 1; VEGF, vascular endothelial growth factor.

Received October 31, 2001.

Accepted for publication March 15, 2002.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Hertig A, Rock JA 1956 A description of 34 human ova within the first 17 days of development. Am J Anat 98:435–494
  2. Lindenberg S 1991 Experimental studies on the initial trophoblast endometrial interaction. Danish Med Bull 38:371–380
  3. Nikas G 1999 Cell surface morphological events relevant to human implantation. Hum Reprod 14(Suppl 2):37–44
  4. Wilcox AJ, Baird D, Weinberg CR 1999 Time of implantation of the conceptus and loss of pregnancy. N Engl J Med 340:1776–1779
  5. Lessey BA 2000 The role of the endometrium during embryo implantation. Human Reprod 15:39–50
  6. Giudice LC 2002 Implantation and endometrial function. In: Fauser BCJM, ed. Molecular biology in reproductive medicine. New York: Parthenon Publishing; 439–465
  7. Paria BC, Lim H, Das SK, Reese J, Dey SK 2000 Molecular signaling in uterine receptivity for implantation. Semin Cell Dev Biol 11:67–76[CrossRef][Medline]
  8. Noyes RW, AI Hertig, J Rock 1950 Dating the endometrial biopsy. Fertil Steril 1:3–25
  9. Irwin JC, Utian WH, Eckert RL 1991 Sex steroids and growth factors differentially regulate the growth and differentiation of cultured human endometrial stromal cells. Endocrinology 129:2385–2392[Abstract]
  10. Taylor RN, Vigne J-L, Zhang P, Hoang P, Lebovic DI, Mueller MD 2000 Effects of progestins and relaxin on glycodelin gene expression in human endometrial cells. Am J Obstet Gynecol 182:841–849[CrossRef][Medline]
  11. Claverie J-M 1999 Computational methods for the identification of differential and coordinated gene expression. Hum Mol Genet 8:1821–1833[Abstract/Free Full Text]
  12. Lipshutz RJ, Fodor SP, Gingeras TR, Lockhart DJ 1999 High density synthetic oligonucleotide arrays. Nat Gene 21(Suppl):20–24
  13. Wodicka L, Dong H, Mittmann M, Ho MH, Lockhart DJ 1997 Genome-wide expression monitoring in Saccharomyces cerevisiae. Nat Biotechnol 15:1359–1367[CrossRef][Medline]
  14. Raychaudhuri S, Sutphin PD, Chang JT, Altman RB 2001 Basic microarray analysis: grouping and feature reduction. Trends Biotechnol 19:189–193[CrossRef][Medline]
  15. Seppala M, Koistinen H, Koistinen R 2001 Glycodelins. Trends Endocrinol Metab 12:111–117[CrossRef][Medline]
  16. Giudice LC, Milkowski DA, Lamson G, Rosenfeld RG, Irwin JC 1991 Insulin-like growth factor binding protein-2 (IGFBP-2) and IGFBP-3 in human endometrium: steroid-dependent messenger RNA expression and protein synthesis. J Clin Endocrinol Metab 72:779–787[Abstract]
  17. Apparao KBC, Murray MJ, Fritz MA, Meyer WR, Chambers AF, Truong PR, Lessey BA 2001 Osteopontin and its receptor {alpha}vß3 integrin are co-expressed in the human endometrium during the menstrual cycle but regulated differentially. J Clin Endocrinol Metab 86:4991–5000[Abstract/Free Full Text]
  18. Kitaya K, Yasuda J, Yagi I, Tada Y, Fushiki S, Honjo H 2000 IL-15 expression at human endometrium and decidua. Biol Reprod 63:683–687[Abstract/Free Full Text]
  19. Okada S, Hidetaka O, Sanezumi M, Nakajima T, Yasuda K, Kanzaki H 2000 Expression of interleukin-15 in human endometrium and decidua. Mol Hum Reprod 6:75–80[Abstract/Free Full Text]
  20. Verma S, Hiby SE, Loke YW, King A 2000 Human decidual natural killer cells express the receptor for and respond to the cytokine interleukin 15. Biol Reprod 62:959–968[Abstract/Free Full Text]
  21. Li XF, Charnock-Jones S, Zhang E, Hiby S, Malik S, Day K, Licence D, Bowen JM, Gardner L, King A, Loke YW, Smith SK 2001 Angiogenic growth factor messenger ribonucleic acids in uterine natural killer cells. J Clin Endocrinol Metab 86:1823–1834