| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
a Prince Henry's Institute of Medical Research, Clayton, Victoria, 3168, Australia
b Institute of Reproduction and Development, Monash Medical Centre, Clayton, Victoria, 3168, Australia
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Ets (E26 transformation-specific) family of transcription factors are involved in proliferation, differentiation, and tissue remodeling during embryonic development. The first member of this family to be identified was the viral oncogene v-ets from the avian erythroblastosis virus, E26 [1, 2]. Subsequently, the human cellular homologue, Ets1, and other Ets family members have been identified including Ets2 [3, 4], Erg [5], and Elf1 [6]. The Ets proteins encoded by these genes each contain a highly conserved region of 85 amino acids termed the Ets domain [7], which binds purine-rich core sequences (GGAA/T) in the promoter and enhancer regions of target genes. Genes with Ets-binding sites include those for cytokines such as granulocyte-macrophage-colony-stimulating factor (GM-CSF), interleukin (IL)-12, and IL-1; the apoptosis-related genes p53 and bcl-2; and many of the matrix metalloproteinases.
The expression of Ets family members during murine embryonic development has been widely investigated. Ets1 expression was found to be restricted to tissues undergoing morphogenesis and tissue remodeling, whereas Ets2 showed a widespread expression in all neonatal organs examined [8–10]. PEA3, ER81, and ERM, which share high sequence identity in their Ets domains, are localized in many of the same developing organs, yet they display unique cell-type-specific expression patterns [11]. In the adult mouse, Ets1 is expressed at high levels predominantly in the thymus, whereas Ets2 expression can be found to varying degrees in a wide range of tissues and cell types [8, 10, 12]. Elf1, which was previously thought to be lymphoid- and myeloid-specific [6], was recently detected at high levels in adult murine testis, liver, kidney, small intestine, and bladder, and in epithelial cells of the gut and skin [13]. Together, these data highlight the predominant colocalization of Ets proteins yet suggest unique roles for specific Ets family members at different stages of development and in adult tissues.
A number of genes containing Ets-binding sites are differentially regulated in the endometrium during the normal menstrual cycle, but little is known of the molecular mechanisms controlling their transcription. In this study, we investigated the spatial and temporal protein expression patterns of Ets1, Ets2, and Elf1 in the normal human endometrium. We demonstrate that not only are all three proteins expressed in this tissue but also that they have highly specific and different cellular localizations, which change with the phase of the menstrual cycle.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Endometrial tissue was obtained by curettage from consenting women undergoing infertility investigation or tubal ligation, with no evidence of endometrial dysfunction, as assessed by pathological examination (approved by the Human Ethics Committee at the Monash Medical Centre, Clayton, Australia). Endometrial samples used for immunohistochemistry were fixed in 10% buffered formalin at 4°C overnight, washed in Tris-buffered saline (TBS), and embedded in paraffin wax. Samples used for cell culture were collected into Dulbecco's minimum essential medium (MEM; Trace Biosciences, Sydney, Australia) for transport to the laboratory.
Immunohistochemistry
Immunohistochemical detection was performed on 5-µm sections using the TSA-Indirect Kit (NEN Life Science Products, Boston, MA) according to the manufacturer's instructions. Affinity-purified, rabbit polyclonal antibodies against Ets1, Ets2, and Elf1 (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were diluted in blocking buffer provided with the kit to concentrations of 1:1000, 1:500, and 1:750, respectively. Incubation was at 4°C overnight for all three antibodies. Secondary antibody was a 1:200 dilution of biotinylated donkey anti-rabbit IgG (DAKO, Glostrup, Denmark), and incubation was for 45 min at room temperature. Visualization used the liquid DAB-Plus Substrate Kit (Zymed Laboratories, South San Francisco, CA) at 1:5 dilution of diaminobenzidine (DAB)/chromogen mixture for 3 min followed by a 1:2 dilution of DAB enhancer for 1 min. Sections were counterstained with Mayer's hematoxylin, and coverslips were mounted with DPX (BDH Laboratory Supplies, Poole, Dorset, England).
For every tissue, a second section on the same slide was used as a negative control for which normal rabbit immunoglobulin IgG (DAKO) was diluted to the same final protein concentration as the primary antibody and substituted for primary antibody. In addition, one positive control slide each of rat thymus and Day 28 endometrium was included in each staining run as positive controls for Ets proteins and to provide quality control between experimental runs. As an additional control, preadsorption was performed by incubating the Ets1, Ets2, and Elf1 antibodies with the corresponding peptides (Santa Cruz Biotechnology, Inc.) at a ratio of 1:100 (w:w) for 7 days at 4°C before immunohistochemistry.
Quantitative analysis of the number of positively stained cells was undertaken using an Olympus (Woodbury, NY) BX-50 microscope and a x40 objective. The image was captured using a Pulnix TMC-6 video camera (Sunnyvale, CA) coupled to a Pentium PC computer using a Screen Machine II fast multimedia video adaptor (FAST Multimedia AG, Munich, Germany). A software package (Olympus DK CASTGRID V1.10, Olympus, Denmark) was used to generate a counting frame (14565 µm2) directly onto the video screen. Fields to be counted were selected using a systematic uniform sampling scheme generated by the CASTGRID V1.10 computer program with the aid of a motorized stage (Multicontrol 2000; ITK, Ahornweg, Germany). The total number of cells and the number of positive cells (excluding intravascular cells) were counted for each section. Positive cells were assigned an intensity of 1–2+ or 3–4+ in relation to those assigned to the positive control endometrial sample. Cell counting was performed by the same observer, with no knowledge of the patients' clinical characteristics.
Differences in the proportion of stromal or glandular epithelial (GE) cells and the intensity of staining at different times in the menstrual cycle and between cell types was assessed using Student's two-tailed t-test. Differences were taken as significant when p < 0.05. Results are presented as mean ± SEM.
Isolation and Culture of Human Endometrial Stromal and Epithelial Cells
Cells were prepared from endometrial tissue as described previously in order to obtain > 95% pure populations of endometrial GE and stromal cells in sufficient number [14, 15]. Briefly, chopped tissue was digested with bacterial collagenase type III (Worthington Biochemical Corporation, Freehold, NJ) at a concentration of 45 IU/ml, in the presence of 3.5 µg/ml deoxyribonuclease (Boehringer Mannheim Biochemica, Mannheim, Germany) for 40 min at 37°C in a shaking incubator, filtered sequentially through 45- and 10-µm nylon filters to remove glands, and centrifuged on Ficoll-Paque (Pharmacia, Uppsala, Sweden) to pellet erythrocytes. Stromal cells were collected from the Ficoll interface, and epithelial glands were recovered from the filters by backwashing [15]. The resulting stromal cells or glands were resuspended in a 1:1 mixture of Dulbecco's MEM and Ham's F-12 medium (Trace Biosciences) with 10% fetal calf serum (Trace Biosciences) and 1% antibiotics (penicillin, streptomycin, and fungizone; CSL, Melbourne, Australia), and grown until confluent with changes of medium every 48 h. Medium was then removed, and cells were scraped into lysis buffer (1% SDS, 0.06 M Tris, 10% glycerol), and stored at -20°C.
Western Blot Analysis
Frozen, cultured endometrial cells in lysis buffer were thawed, and total protein concentration was determined using the micro-bicinchoninic acid (BCA) protein assay (Pierce, Rockford, IL) according to the manufacturer's instructions. Samples each containing 6 µg of protein, together with molecular weight standards, were subjected to SDS-PAGE on 10% gels under reducing conditions, and the proteins were transferred to Hybond-P membranes (Amersham Life Science Ltd, Little Chalfont, Buckinghamshire, UK). Jurkat T cells were used as positive controls for evaluation of Ets1 and Elf1, and transfected NIH3T3 cells that overexpress Ets2 (Ets2/3T3) [16] were used as the positive control for Ets2 evaluation. After nonspecific sites were blocked with 10% skim milk powder in TBS with 0.1% Tween-20 (TBS-T) for 1 h, blots were incubated overnight with affinity-purified rabbit antisera to Ets1, Ets2 (both diluted 1:2000), or Elf1 (1:1000; Santa Cruz Biotechnology, Inc.), diluted in 5% skim milk powder in TBS-T. The blot was subsequently washed 3 times in TBS-T for 10 min each and incubated with donkey anti-rabbit IgG conjugated to horseradish peroxidase (Amersham Life Science) diluted 1:2000 in 5% skim milk powder in TBS-T. Chemiluminescence was detected using the ECL Plus Kit (Amersham Life Science).
Blots were exposed to Kodak X-OMAT AR film (Eastman Kodak, Rochester, NY) for 30 sec to 2 min and developed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Both the proportion of cells stained and the relative intensity of staining were examined for the GE and stromal compartments for each protein and are represented graphically in Figure 1. Photomicrographs of immunohistochemistry sections showing representative tissues from the menstrual cycle stained with Elf1, Ets1, and Ets2 are shown in Figure 2. All three proteins were found within endometrial tissue across the entire menstrual cycle in both GE and stromal cells; however, the proportion of cells stained in each cell type and the intensity of staining varied for each protein (Fig. 1).
|
|
Elf1 Immunoreactive Elf1 was found across the entire menstrual cycle in both GE and stromal cells (Fig. 1A).
Glandular epithelium From the early proliferative (P) to late secretory (S) phase, there was significantly more Elf1 in glandular epithelium than in stroma (p < 0.05). Elf1 was present in > 75% of GE cells during these phases. The predominance of Elf1 in GE cells can be seen in Figure 2D, which shows strong GE staining of Elf1 in Day 14 endometrium, with relatively little staining in the stromal cells.
The proportion of GE cells stained for Elf1 was significantly reduced during the menstrual phase compared to the other phases of the cycle (p < 0.05), and there was no significant difference at this time between the percentage of Elf1 in GE and stromal cells (Fig. 1A). This is evident in Figure 2A, which shows a section of endometrial tissue from Day 1 of the cycle with very little staining of Elf1 in the GE or stromal compartments. The percentage of GE cells stained with high (3–4+) compared to low (1–2+) intensity was not significantly different during the menstrual, early P, early S, and mid S stages of the menstrual cycle, but a significantly greater proportion of GE cells stained strongly (3–4+) for Elf1 during the mid-late P and late S phases (p < 0.05; Fig. 1A).
Stroma From the early P to the late S phases of the menstrual cycle, Elf1 was expressed in only 20–35% of stromal cells. There was a relatively equal proportion of cells stained with low (1–2+) and high (3–4+) intensity during these phases (Fig. 1A). Consistent with GE cells, the percentage of stromal cells containing Elf1 was very low in the menstrual phase of the cycle (16%), and the majority of these cells (90%) stained with low intensity (Figs. 1A and 2A).
Ets1 Ets1 was present in both GE cells and stromal cells across all phases of the menstrual cycle (Fig. 1B).
Glandular epithelium Ets1 was significantly more abundant in GE cells than in stromal cells from the mid-late P phase to the mid S phase (p < 0.05; Figs. 1B and 2E). There was no significant difference between GE and stromal cells for the proportion of cells stained in the menstrual, early P, or late S phases (Fig. 1B).
The overall pattern of Ets1 expression in GE cells was similar to that of Elf1. However, the proportion of stained cells was much lower (< 50%), and the intensity of Ets1 staining was predominantly low (1–2+) with < 8% of GE cells staining with high intensity (3–4+) at any stage of the cycle (Fig. 1B). Staining of Ets1 in GE cells was almost completely abolished during the menstrual phase, in which the proportion of GE cells containing Ets1 was significantly lower than in all other phases of the cycle (p < 0.05; Figs. 1B and 2B).
Stroma The percentage of stromal cells containing Ets1 was not significantly altered across the menstrual cycle; however, there was a trend towards increasing Ets1 across the cycle (Fig. 1B). Consistent with results in the GE compartment, the presence of Ets1 was lowest in the menstrual phase (Figs. 1B and 2B).
Ets2 Ets2, unlike Elf1 and Ets1, was expressed equally in stromal and GE cells with no significant dominance in one cell type at any stage of the menstrual cycle (Figs. 1C and 2F). Consistent with Elf1 and Ets1, Ets2 was almost completely absent during the menstrual phase in both GE and stromal cells (Fig. 2C). The intensity of staining was significantly low (1–2+) in both GE cells and stromal cells at all stages of the cycle (p < 0.05; Fig. 1C).
Immunolocalization of Ets Proteins in Endometrial Decidua and Luminal Epithelium
In late secretory endometrial tissues, which contain many decidualized stromal cells, Elf1, Ets1, and Ets2 were all found at high intensity (3–4+) in most cells of this type, accounting for most of the stromal staining seen at this stage (Fig. 2, G–I).
Ets1, Ets2, and Elf1 immunostaining was seen in the luminal epithelium predominantly in cells of lymphomyeloid origin migrating between luminal epithelial cells (data not shown).
Preadsorption and Controls
Staining of Elf1, Ets1, and Ets2 was abolished in sections in which the antibodies for these proteins were preadsorbed with the corresponding peptides (Fig. 2, K–M). Positive staining was observed in rat thymus tissue incubated with Elf1 (Fig. 2J), Ets1, or Ets2 (data not shown). No staining was observed when sections were incubated with rabbit IgG instead of Ets primary antibodies (Fig. 2N).
Western Analysis of Ets Proteins in Cultured Endometrial Cells
Western blotting was performed on cellular lysates of purified epithelial and stromal endometrial cells to confirm the cellular distribution of Ets proteins in the human endometrium and to ascertain the molecular masses of isoforms of Ets proteins in these cells. Endometrial samples from Days 9–22 of the menstrual cycle were used since cells derived from tissue in the menstrual and late secretory phases are difficult to culture.
Elf1 Western analysis revealed that the Elf1 protein (which migrates as a 97-kDa doublet by SDS-PAGE) was present in both GE cells and Jurkat T cells used as a positive control (Fig. 3A). Elf1 was faintly detectable in the stromal cell lysates (Fig. 3A). This result confirmed the cellular localization obtained by immunohistochemistry.
|
Ets1 Western blot analysis using the Ets1 polyclonal antibody demonstrated bands at 39, 42, and 51 kDa in Jurkat T-cells used as a positive control (Fig. 3B, lane 1). p51 is the full-length isoform of Ets1, and p42 is a product of alternatively spliced mRNA that lacks exon VII [17]. p39 has previously been shown to be derived from p42 by covalent modification [18]. In endometrial stromal and GE cells (Fig. 3B, lanes 2 and 3), p42 and p51 were detected. The abundance of p51 was much higher than p42 in both stromal and GE cells. In addition, a 52-kDa product of intensity similar to that of p51 was seen in endometrial stromal cells (Fig. 3B, lane 2) but was not detected in endometrial GE cells (Fig. 3B, lane 3). This protein is a phosphorylated form of p51 [17]. Unlike the control T cells, p39 was not detected in either endometrial GE or stromal cells.
Ets2 Analysis of Ets2/3T3 cells, endometrial stromal cells, and GE cells with the Ets2 polyclonal antibody revealed two proteins at 54 kDa and approximately 52 kDa (Fig. 3C). The 54-kDa protein has previously been identified as the full-length isoform of Ets2 [18]. p52 has not previously been characterized; however, this protein may be a phosphorylated form of Ets2 [19]. The abundance of p52 was substantially higher in GE cells than in stromal cells and was the predominant form of Ets2 in all of the cell types examined. Overall, there was less Ets2 protein in stromal cells than in the other two cell types.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Immunohistochemical analysis of Elf1 in the human endometrium showed that this transcription factor was localized predominantly in the GE cells, with very little immunoreactivity in stromal cells. Western blot analysis confirmed this localization pattern and also showed the presence of an Elf1 doublet at approximately 97 kDa. Elf1 is known to be expressed at high levels in normal lymphoid cells and cell lines. Examination of Elf1 expression in normal mouse tissues, murine cell lines, human adult and fetal tissues, and human cell lines indicates that Elf1 expression is highest in lymphocytes, with moderate expression in hematopoietic cell lineages [20]. Recently, Elf1 was also shown to be expressed in epithelial cells of the developing and adult mouse respiratory tract, gastrointestinal and urogenital tracts, and skin [13]. These results suggest dominant expression of Elf1 in the epithelial lineage of an epithelial-mesenchymal system. This implicates Elf1 as a potential regulator of epithelial cell restricted genes. One such gene, characterized by its exclusive localization to GE cells in the human endometrium, is the gene encoding matrix metalloproteinase (MMP)-7 (matrilysin) [21]. Recent studies on the interactions between Ets proteins and MMPs have revealed that most of the MMPs contain one or more functional binding sites for Ets proteins, and some of these proteins have been shown to transactivate MMP promoters in vitro [22–26]. It is therefore possible to speculate that Elf1 may be involved in the regulation of MMP-7 in endometrial GE cells. In contrast to Elf1, Ets1 and Ets2 show a much wider distribution in the human endometrium. Our results show that Ets2 is equally abundant in both GE and stromal cells. This is characteristic of previous localization patterns of Ets2, in which it has been found at various levels of expression in all proliferating tissues and cell lines examined to date [8]. Detailed analysis of Ets2 isoforms in cultured endometrial stromal and GE cells by Western blotting revealed that there are two immunoreactive forms of Ets2 (p52 and p54). p52 has only recently been reported [19] and may be produced by posttranslational modification of the well-characterized 54-kDa form. Since p52 is the predominant Ets2 protein in both endometrial stromal and GE cells, it is likely that this modification is important to the biological activity of Ets2 in the endometrium. It is also likely that p52 Ets2 constitutes the active isoform in GE cells, where it is significantly more abundant than in stromal cells. Ets2 has previously been implicated in regulation of cell cycle processes, and this may also be the case in human endometrium. Recently, Yamamoto et al. [27] showed that a targeted deletion of ets2 in mice resulted in growth retardation and then embryonic death. This defect in embryonic development may be due, in part, to a consequent defect in transcriptional regulation of epidermal growth factor (EGF), transforming growth factor 
(TGF), and the MMP-13 (homologue to human MMP-1), MMP-3 and MMP-9 by Ets2 [27]. TGF, EGF, and EGF receptors are normally ubiquitously expressed throughout the endometrium, in both the proliferative and secretory phases of the menstrual cycle (see [28] for review). These expression patterns correlate with that found for Ets2 in this study. MMPs 1, 3 and 9 are found only in endometrial stroma (see [29] for review), suggesting that stromal Ets2 may contribute to their expression but is not sufficient to induce epithelial MMP expression.
Western analysis of Ets1 showed that alternative splicing and differential phosphorylation may modify its function in the human endometrium. In endometrial stromal cells, the Ets1 isoforms p42 (alternatively spliced), p51 (full-length), and p52 (phosphorylated) were present. However, in endometrial GE cells, p52 was not detected. It is possible that phosphorylation of Ets1 in stromal cells is an intrinsic mechanism necessary for its transcriptional activity, which does not occur in the GE compartment.
Given the differences in cell-type expression of the three Ets proteins, it is interesting that their expression levels are modified similarly during the menstrual cycle. First, Ets1, Ets2, and Elf1 were all found in significantly low abundance during the menstrual phase of the cycle. In addition, all were found in high abundance and intensity in decidualized cells of late secretory phase endometrium. These cells secrete a range of proteins characteristic of their differentiated state including insulin-like growth factor binding protein 1, prolactin, relaxin, desmin, hsp27, laminin, and the tissue inhibitors of MMPs (TIMPs) [30, 31]. Logan et al. [32] have shown previously, through cotransfection of Ets1 and AP-1 expression vectors, that these proteins interact to produce up-regulation of the TIMP1 promoter. Since the regulatory mechanisms of TIMP regulation are not completely understood, it is likely that Ets proteins are involved in the regulation of TIMPs and probably other proteins characteristically secreted by endometrial decidual cells. It has already been demonstrated in other systems that a variety of members from the Ets family of transcription factors can be localized to the same tissue and regulate a variety of gene promoters, in vitro, through the same DNA binding site. It is therefore not surprising that different members of the Ets family are present in the endometrium. Specific transcriptional activation or repression by each particular Ets protein may involve a variety of regulatory mechanisms to allow competition for Ets-binding sites. Further detailed in vitro analysis of the interactions between these Ets members and factors known to be important in endometrial remodeling will provide important information on the specific roles for these transcription factors in this tissue.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Lynette M. Kilpatrick, Prince Henry's Institute of Medical Research, PO Box 5152, Clayton, Victoria, 3168, Australia. FAX: 61 3 95946125; lynette.kilpatrick{at}med.monash.edu.au
Accepted: February 17, 1999.
Received: November 2, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and tumor necrosis factor ß. J Clin Endocrinol Metab 1994; 79:530–536.[Abstract]This article has been cited by other articles:
|
|
 |
|
  N. Tawadros, L.A. Salamonsen, E. Dimitriadis, and C. Chen Facilitation of decidualization by locally produced ghrelin in the human endometrium Mol. Hum. Reprod., July 1, 2007; 13(7): 483 - 489. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. A. Kessler, J. K. Schroeder, A. K. Brar, and S. Handwerger Transcription factor ETS1 is critical for human uterine decidualization Mol. Hum. Reprod., February 1, 2006; 12(2): 71 - 76. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
