| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
inHuman Endometrium1
a Department of Obstetrics and Gynecology, Division of Reproductive Endocrinology and Infertility, Duke University Medical Center, Durham, North Carolina 27710
b Laboratory of Molecular Carcinogenesis, and
c Laboratory of Reproductive and Developmental Toxicology, NIEHS, NIH, Research Triangle Park, North Carolina 27709
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) in human endometrium, uterine luminal fluid proteins were retrieved by lavage followed by collection of the adjacent endometrium at hysterectomy. In the endometrium we observed the presence of the full-length transmembrane TGF protein and the phosphorylation of its only known receptor, the epidermal growth factor receptor (EGFR), by immunoprecipitation-Western blot; TGF mRNA via reverse transcription-polymerase chain reaction; and immunolocalization of TGF to the surface endometrium adjacent to the uterine lumen. Despite this demonstration of TGF in functional endometrium, we could not detect measurable amounts of TGF in any of the 16 endometrial washings by either immunoprecipitation-Western blot or by ELISA. Recovery rate for intraluminal fluid spiked with TGF control peptide was 93.4–97%. The inability to detect TGF in intraluminal fluid despite its high concentration in cells directly adjacent to the uterine lumen, along with the absence of any cleaved TGF species identified in the endometrium, suggests that TGF signals its receptor as a transmembrane ligand. Since the EGFR is present in the endometrium and on the surface of embryos, these data are consistent with a juxtacrine mode of signaling for TGF between endometrial cells, and between the luminal surface epithelium and preimplantation embryos.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) action has involved the synthesis of a transmembrane precursor molecule, the cleavage of an active peptide from the extracellular portion of this precursor, and the migration of a soluble, active peptide to a distant receptor. Much of this biology was elucidated in malignant cells. Recently, evidence supporting a juxtacrine mode of signaling for TGF has emerged, which raises the question as to whether the predominant mode of signaling for TGF in nontransformed tissue is via the full-length transmembrane ligand under conditions that require cell-to-cell contact.
There is evidence that the full-length transmembrane TGF, anchored in the membrane of the uterine luminal epithelium, signals the epidermal growth factor receptor (EGFR) on the mouse blastocyst through direct cell-to-cell contact [1–5]. Induction and peak levels of TGF mRNA [1], along with full-length transmembrane protein [2], is found in the luminal epithelium during implantation. This endometrial TGF expression is temporally coupled with a corresponding rise in blastocyst expression of EGFR [3, 4] and the preferential localization of functional EGFR on the basolateral surface of the blastocyst trophectoderm [5]. Evidence that TGF/EGFR signaling may be important in the human includes the observation of persistent expression of EGFR mRNA in the polar trophectoderm of human blastocysts into the periimplantation time frame [6], the immunolocalization of EGFR in human implantation trophoblast [7], and the cyclic changes of TGF expression in the endometrial surface epithelium during the secretory phase of the menstrual cycle when embryos are present in the uterus [8].
Full-length transmembrane TGF is the predominant form of TGF in a variety of normal cell types [9]. The cytoplasmic domain of TGF is highly conserved among species [10, 11], which suggests a defined biologic function. Recent work has demonstrated the potential for bidirectional signaling when transmembrane TGF binds the EGFR, with both the extracellular domain inducing a mitogenic response through the EGFR and the cytoplasmic domain becoming associated with a protein kinase complex [12, 13] analogous to the p56lck association of the T-cell membrane glycoproteins CD4 and CD8 [14–16]. This type of bidirectional signaling for TGF and the EGFR has potentially significant implications for reproductive biology in humans. Once a competent embryo expressing EGFR makes contact with the full-length transmembrane TGF in the recipient endometrium, not only are embryonic growth and differentiation fostered, but so is the potential for the embryo to induce a mitogenic response in the tissue that it is invading. This type of bidirectional signaling may also be operative between endometrial cells during cyclic endometrial regeneration. Significant is the fact that this bidirectional signaling would not be possible if the extra-cellular peptide were liberated.
Juxtacrine signaling of the EGFR under in vitro conditions whereby the transmembrane TGF could not be cleaved has been demonstrated with human fibroblasts [17], a human epidermoid carcinoma cell line [17, 18], and a murine hematopoietic cell line [19]. The mechanism in vivo whereby TGF signals its receptor in human endometrium is unknown. Understanding which type of intercellular communication is predominately operative, either through a soluble factor or through the full-length transmembrane protein, each with very different biologic implications, is important for understanding implantation and the biology of the endometrium.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The inclusion criteria for entry into this study included women of reproductive age (28–48 yr old, mean age 39, median age 40) with spontaneous menstrual cycles occurring every 26–35 days, who were scheduled for a hysterectomy. Exclusion criteria included 1) any use of hypothalamic, pituitary, or ovarian reproductive hormones or analogues within 6 mo; 2) past or present evidence of endometrial pathology; 3) a history of dysfunctional uterine bleeding; or 4) submucous leiomyomas. This study was done in accordance with the guidelines of the Internal Review Board committee at Duke University Medical Center.
Acquisition of Intraluminal Fluid Sample
Immediately after the uterus was handed from the operative field, the cervix was removed from the uterine corpus at the level of the internal os, the uterus was tilted fundus down, and the uterine cavity was gently lavaged with PBS, pH 7.4, using a Pasteur pipette. The uterus was then bivalved. The intact endometrium was once again gently irrigated, and the run-off was collected. The total lavage volume from each patient was 1.8–2 ml. Lavaged intraluminal fluid specimens were immediately immersed in the operating room in liquid nitrogen at -70°C and stored in the laboratory at -20°C, and all samples were assayed within 6 mo of collection. A full-thickness endometrial specimen was then obtained and fixed in formaldehyde for determination of the stage of the menstrual cycle. All endometrial specimens were obtained from the fundus in a uniform fashion. The same board-certified pathologist performed a blinded histological analysis of all endometrial samples.
Immunoprecipitation and Western Blot Analysis of Intraluminal Fluid
Total protein concentration for each lavage specimen was determined with the Pierce (Rockford, IL) BCA Protein Assay. One-milliliter intraluminal samples were then prepared by bringing them to a final concentration of 1% BSA, 1% SDS and 2.5 mM EDTA, and then boiling for 10 min. An equal volume of solution B (0.2% SDS, 0.5 mM EDTA, 0.8 mM dithiothreitol, 0.8% Triton X-100, and 0.3 M sodium chloride in 40 mM Tris-HCl buffer pH 7.5) was added, and the solution was incubated with 4 mg protein-A sepharose (Pharmacia, Piscataway, NJ) for 5 h at 4°C. The supernatant was retrieved after centrifugation for 2 min at 700 x g, and antibody against the amino terminus of the mature TGF protein (Ab-1; Oncogene Research Products, Cambridge, MA) was added at a concentration of 0.1 µg/ml and incubated for 4 h at 4°C. After formation of the antigen-antibody complex, 4 mg of protein-A sepharose was added and incubated overnight at 4°C. After centrifugation for 2 min at 700 x g, the pellet was resuspended in 40 µl double-strength Laemli sample buffer, and the proteins were resolved on a 16.5% Tris-Tricine SDS-PAGE gel. The proteins in these gels were then transferred to nitrocellulose membrane, and the membranes were blocked with Tris-buffered saline (TBS), 0.1% Tween 20, 10% nonfat dried milk and probed with the appropriate antibodies. Immunoreactive proteins were detected with enhanced chemiluminescence (Amersham Life Science, Arlington Heights, IL). The recombinant TGF used as a control in all immunoprecipitation-Western blots of intraluminal fluid and endometrium was an undenatured protein (Oncogene Research Products) with a shelf life for bioactivity in excess of one year when stored at -20°C.
ELISA of Intraluminal Fluid
Intraluminal samples were pooled according to phase of the menstrual cycle and assayed with a TGF quantitative ELISA (Oncogene Research Products; published sensitivity of 10 pg/ml) according to the manufacturer's protocol. Briefly, polyclonal antibodies corresponding to epitopes from the 50-amino acid extracellular region as well as epitopes corresponding to the full-length molecule were immobilized onto the surface of the microtiter plate wells (to capture antigen), the wells were rinsed, and then uterine lavage samples that had been mixed with biotinylated goat anti-human TGF antibodies were added to the wells. After the 3-h incubation, the wells were washed (x3) and rinsed (x2) to remove any unbound material. Streptavidin-horseradish peroxidase was then incubated with the samples for 30 min. After washing (x3) and rinsing (x2), the chromogenic substrate O-phenylenediamine was incubated with the samples in the dark, and then 4 N sulfuric acid was added to stop the reaction. Spectro-photometric absorbency of light at 490 nm for the intraluminal samples was quantified against a standard curve using known amounts of lyophilized control TGF peptide reconstituted in dH2O and then serially diluted in assay buffer to concentrations of 250, 125, and 63 pg/ml. In addition, control samples were set up at concentrations of 63, 125, and 250 pg/ml reconstituted with a 1:1 mix of intraluminal fluid and assay buffer to rule out the possibility of a factor being present in the intraluminal fluid that was interfering with the ability of the assay to detect TGF. All assays were carried out in duplicate.
Immunoprecipitation and Western BlotAnalysis of Endometrium
The endometrium was gently scraped from the fundus using a scalpel in a uniform fashion and was immediately placed into solubilization buffer (1% Triton X-100, 2 mM EDTA, 2 mM EGTA, 1 mM Na3VO4, 20 mM NaF, 50 µM Na2Mo4O4, 20 µg/ml aprotinin, 20 µg/ml leupeptin, and 15 µM [4-amidinophenyl]-methanesulfonyl fluoride in 20 mM Hepes, pH 7.4) on ice. The tissue was then homogenized with a Virtishear (Virtis Corp., Gardiner, NY). After centrifugation at 21 000 x g for 1 min, the supernatant was drawn off into aliquots. Total protein concentration was determined with the Pierce BCA Protein Assay.
Protein-A sepharose beads, 10 mg, were washed twice with 1 ml inositol 1,4,5-triphosphate buffer (50 mM Tris pH 8.5, 150 mM NaCl, 1 mM EDTA, and 0.5% Triton X-100) to remove unconjugated protein-A. These beads were then added to the endometrial homogenate and placed at 4°C for 20 min on a rocker platform to clear the homogenate of any nonspecific binding to the beads. The cleared supernatant was then incubated for 1 h with the appropriate immunoprecipitating antibody. For the EGFR, a rabbit polyclonal antiserum was raised against a peptide corresponding to the extreme C-terminus (residues 1157–1186) of the rat EGFR, and affinity-purified (courtesy of R.P. DiAugustine, LMC, NIEHS, Research Triangle Park, NC); the commercial source and epitope specificity of all other antibodies is noted elsewhere.
After the 1-h incubation, 10 mg protein-A sepharose (prewashed as above x2) was added to the antibody-antigen complex and incubated on a platform rocker for 3 h at 4°C. Samples were then centrifuged at 10 000 x g for 30 sec, and the pellet was washed three times in buffer with a progressively lower salt concentration (Tris 50 mM to 10 mM; NaCl 500 mM to none; EDTA 1 mM to none) and with a change in detergent from 0.5% Triton X-100 to 0.1%. The pellet was then brought up in 35 µl of double-strength Laemli sample buffer and boiled for 5 min, and the supernatant was stored at -20°C. The samples were resolved on SDS-PAGE gels—15% for the TGF blots and 7.5% for the EGFR blots. Membranes were blocked with TBS, 0.1% Tween 20, 5% bovine serum albumin and probed with the appropriate antibodies. Immunoreactive proteins were detected with enhanced chemiluminescence.
Reverse-Transcriptase Polymerase Chain Reaction(RT-PCR) to Amplify TGF mRNA from the Endometrium
Endometrial tissue was obtained as above and homogenized with the Virtishear in a guanidine HCl buffer. Total RNA was isolated via sodium acetate/ethanol precipitation reactions followed by organic extractions with chloroform/butanol.
All chemicals were supplied by J.T. Baker Co. (Phillipsburg, NJ), Sigma Chemical Co. (St. Louis, MO), and United States Biochemical Corp. (Cleveland, OH). RNA was resuspended in nuclease-free water and stored at -20°C. Optical density readings were measured with a Beckman spectrophotometer (Schaumburg, IL). Twenty micrograms of total RNA isolated was run on an electrophoresis gel in order to ensure its purity and demonstrate the absence of degradation.
RT-PCR reagents were supplied by Lifecodes Corp. (Stamford, CT) and Perkin-Elmer Corp. (Norwalk, CT). For 8 µg of total RNA isolated from the human endometrial tissue, 1 µg of oligo(dT)16 was added, incubated for 5 min at 68°C, and then placed on ice for 2 min. For the negative control, nuclease-free water was substituted for the RNA. The following reagents were utilized per reverse transcription reaction: 10 µl dNTPs, 2 µl 10-strength PCR buffer, 4 µl 25 mM MgCl2, 1 µl RNase inhibitor, and 1 µl MuLV. The reactions were heated for 1 h at 42°C and then for 10 min at 95°C. To 1 µl of the resultant cDNA, the following PCR reagents were added per reaction: 5 µl dNTPs, 5 µl 10-strength PCR buffer, 1 µl each of 3' and 5' primer, and 0.4 µl Taq polymerase; the solution topped off with 20 µl of mineral oil. The human TGF primer set along with positive control cDNA was supplied by Clontech Laboratories (Palo Alto, CA). A Hybaid thermocycler (distributed by Vanguard International, Neptune, NJ) was programmed according to Clontech Lab's primer set specifications: 45 sec at 94°C, 45 sec at 60°C, and 2 min at 72°C, for 35 cycles; 7 min at 72°C; and 7 min at 25°C and the resultant RT-PCR product was then stored on ice. Five microliters of the RT-PCR DNA sample was loaded on a 1.8% agarose (FMC Corp., Rockland, ME) gel in Tris-borate-EDTA with 0.1 µg/ml ethidium bromide.
Immunohistochemistry
Paraffin-embedded sections of the endometrium were incubated at 65°C for 10 min before undergoing deparaffinization and rehydration. Twenty-percent glacial acetic acid at 4°C was used for endogenous alkaline phosphatase blocking before rinsing with room temperature water and re-equilibrating with TBS-0.1% BSA. All subsequent incubations and washes were performed at room temperature. Slides were then incubated for 30 min with 5% normal goat serum. TGF antibody (Ab-2; Oncogene Research Products), at a concentration of 7 µg/ml in TBS-0.1% BSA, was incubated for 1.5 h. Negative controls were included that underwent identical treatment except for the omission of the primary antibody, Ab-2. After rinsing in TBS-0.1% BSA, the slides were then incubated with anti-mouse IgG Link (BioGenex Laboratories, San Ramon, CA) at a 1:40 dilution for 20 min. Repeated rinses and incubation with 1:20 dilution of alkaline phosphatase-conjugated streptavidin Label (BioGenex Labs) followed. After a final rinse with TBS, Chromagen substrate (BioGenex Labs) was applied for 6 min.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
protein in intraluminal fluid by immunoprecipitation-Western blotting, titration experiments were initially performed with the recombinant 50-amino acid TGF control peptide. TGF was consistently immunoprecipitated to a sensitivity of 0.75 µg/ml, allowing us to detect TGF if it was 
0.03%, 0.038%, 0.18%, 0.016%, and 0.06% of total protein in the intraluminal fluid samples dated, respectively, early proliferative, mid-proliferative, Day 18, Day 23, and Day 26. As lanes 2–6 demonstrate in Figure 1, no mature TGF was detected in the intraluminal fluid. In lane 1, a band was observed at 5.5 kDa that consisted of 2 µg of recombinant TGF immunoprecipitated from a total reaction volume of 2 ml. Further, a band was also observed at 5.5 kDa in lane 7 that consisted of 2 µg of recombinant TGF immunoprecipitated from a total reaction volume of 2 ml after incubation for 4 h at 37°C in 0.9 ml of intraluminal fluid from a uterus in which the endometrium was dated as mixed secretory/proliferative. This lane served to control for the possibility that the absence of a band in our experimental lanes was a result of digestion of the cleaved fragment by any proteases or elastases resident in human intraluminal fluid.
|
To address the concern that the immunoprecipitation-Western approach (Fig. 1) was not sufficiently sensitive to detect secreted TGF in uterine lavage fluid, we turned to ELISA because of a > 10 000-fold improvement in sensitivity. Uterine luminal fluid was assayed with ELISA from the uteri of 16 women. The first ELISA run assayed the intraluminal fluid of 9 women whose corresponding endometria were dated as early menstrual, late menstrual, early proliferative, mid-proliferative, late proliferative, Day 18, Days 20–22, Day 23, or Day 27. All of the samples were run in duplicate. The standard curve for this assay demonstrated an excellent quadratic curve fit with a correlation coefficient of 0.999. In no sample was TGF detected. Further evidence that the intraluminal samples are negative for any detectable TGF is the fact that all values fell below one of the two optical densities for the blank control containing assay buffer without any TGF control added.
The intraluminal fluid from the uteri of an additional 7 patients was run in a second ELISA assay. The corresponding endometria from these 7 patients were dated as menstrual not otherwise specified (n = 2), early proliferative (n = 3), mid-proliferative (n = 1), or Days 20–22 (n = 1). All of the samples were run in duplicate. The standard curve for this assay again demonstrated an excellent quadratic curve fit with a correlation coefficient of 0.999. In no sample was TGF detected. Further, all sample values were extremely close to the identical negative control values.
To rule out the possibility that a factor present in the intraluminal fluid was interfering with the ability of the assay to detect TGF, known quantities of TGF were added to a 1:1 mixture of intraluminal fluid:assay buffer and analyzed. These controls were a component of the second ELISA run and, as such, utilized the same precise standard curve. All controls were run in duplicate. Recovery was 96%, 93.4%, and 97% of expected for, respectively, 63, 125, and 250 pg/ml, for an average recovery of rate of 95%. The ELISA data supported and extended our results with immunoprecipitation-Western analysis and suggest that TGF is not present in uterine luminal fluid in amounts detectable by either of these two methods.
Next, we evaluated the endometrial tissue of the lavaged uteri to confirm the presence of TGF protein in these specimens. In Figure 2, lanes 2 and 4–6, is a mid-luteal (Days 20–22) endometrial sample immunoprecipitated with polyclonal anti-sera (M301100S; Biodesign International, Kennebunk, ME) capable of immunoprecipitating both the soluble extracellular 5.5-kDa fragment as well as the full-length transmembrane form of TGF. This experiment demonstrated the epitope specificity of the antibodies used to probe the blot in Figure 3. A 28-kDa species of TGF was identified in the endometrial homogenate when probed with two different commercial antisera that recognize epitopes at opposite ends of the molecule, namely, the extracellular and cytosolic regions. The monoclonal antibody Ab-1 (Oncogene Research Products) recognized an epitope on the 5.5-kDa TGF control peptide in Figure 2, lane 1, and the same epitope on the 28 kDa full-length transmembrane TGF immunoprecipitated from the Day 20–22 endometrium in lane 2. In lanes 3–6, antiserum to the cytosolic epitope(s) of TGF (M30102S) was used on the Western blot. This antiserum to cytosolic epitope(s), shown in Lane 3, did not recognize the 5.5-kDa control peptide as expected yet recognized the cytosolic region of the full-length 28-kDa transmembrane TGF immunoprecipitated from the Day 20–22 endometrium in lane 4. In lane 5, the supernatant of the immunoprecipitation reaction for lane 4 demonstrated efficient immunoprecipitation by the antisera M30100S. In lane 6, the crude endometrial homogenate before immunoprecipitation was probed with the antiserum to the cytosolic epitope(s). If the 28-kDa band observed in lane 4 was due to nonspecific secondary binding to immunoglobulin, one would not observe the 28-kDa band seen in lane 6.
|
|
Demonstrated in Figure 3 is an immunoprecipitation-Western blot of four endometrial samples spanning the menstrual cycle: early proliferative, mid-proliferative, mid-luteal (Days 20–22), and late luteal (Day 27). These samples were immunoprecipitated with antisera to TGF and then probed separately with antibodies recognizing the amino (extracellular) terminus (lanes 1–4) and then the carboxyl (cytosolic) terminus (lanes 5–8) of the transmembrane molecule. Total starting protein for immunoprecipitation was 0.5 mg for each of the eight samples, and all samples were run in the same electroblotting tank to control for protein transfer efficiency. Consistently observed was the 28-kDa band throughout the cycle. The absence of a band at 5.5 kDa in lanes 1–4 demonstrates the absence of a cleaved extracellular TGF species present in the endometrium. This is supported by the absence of a postcleavage cytosolic remnant of any molecular size less than 28 kDa in lanes 5–8. For example, if the 5.5-kDa extracellular protein was cleaved, with amino terminus glycosylation accounting for an additional 4.5 kDa, then one would expect to see a postcleavage remnant of approximately 18 kDa identified by the antibody to the cytosolic tail of the molecule used in lanes 5–8. Further, despite the presence of the full-length TGF protein in the endometrium of these four uteri, no TGF was detected in the intraluminal fluid from the same uteri when analyzed via ELISA.
With regard to the only known receptor for TGF, the EGFR was appropriately identified to migrate at 170 kDa and as such was shown to be present in all four endometrial samples spanning the menstrual cycle (Fig. 4A, lanes 1–4; early proliferative, mid-proliferative, mid-luteal Days 20–22, and late luteal Day 27, respectively). In panel B, an antibody that recognizes phosphorylated tyrosine residues identified the EGFR at 170 kDa from the same four endometrial samples spanning the menstrual cycle (lanes 1–4; early proliferative, mid-proliferative, mid-luteal Days 20–22, and late luteal Day 27, respectively). This demonstrates the autophosphorylation (activation) of the EGFR.
|
Providing evidence of active gene transcription for TGF in the collected tissue samples, Figure 5 demonstrates the TGF RT-PCR product of RNA extracted from endometrium dated as mid-luteal (d20–22). A single band was observed to migrate with the positive control for TGF, demonstrating active mRNA production in the endometrium from a uterus in which no TGF was detected in the intraluminal fluid via ELISA.
|
Immunolocalization of TGF to the surface endometrium directly adjacent to the lumen is demonstrated in Figure 6A. Figure 6B shows a sequential cut of the same endometrial paraffin block. This negative control was run under the same experimental conditions as the section shown in Figure 6A except for the omission of the primary antibody to TGF. Notable is the lack of surface staining for TGF, which is seen in Figure 6A. Again, the uterus was lavaged before the endometrium was obtained for this paraffin section and was included in the intraluminal fluid analysis. Despite the immunolocalization of TGF to the cells directly adjacent to the uterine lumen, no TGF was detected by ELISA.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The importance of elucidating the interplay between receptors on the embryo and growth factor ligands in the endometrium is underscored by the high rate of failed human implantation despite successful fertilization. Developmental and/or implantation failure accounts for the calculated loss rate through 6 wk of at least 73% in natural single conceptions [22]. Insights into endometrial regulatory events that include the potential for bidirectional signaling of TGF with the possibility of modulating the extracellular and/or cytosolic signaling to foster optimal nidation, may facilitate improved implantation rates in the face of unexplained recurrent miscarriage and allow for greater precision with embryo transfer and lower multiple gestation rates.
If endometrial TGF plays a role in the preparation of embryos and/or the endometrium for implantation, this could occur either by the release of a soluble form of TGF into the uterine lumen or by direct contact between the transmembrane form of TGF on the endometrial cell and the EGFR on an adjacent embryo or endometrial cell. In the same uteri from which we collected uterine luminal fluid, we confirmed active DNA transcription of TGF via the detection of mRNA by RT-PCR and the presence of TGF protein by Western blot and immunohistochemistry. This protein was seen to migrate at 28 kDa when identified on immunoblot with antibodies that recognize epitopes on either end of the molecule, namely, an antibody to an epitope within the extracellular 50-amino acid fragment as well as an antiserum that recognizes an epitope(s) in the cytosolic region. We found TGF protein in the surface epithelium directly adjacent to the uterine lumen but were unable to detect TGF in the uterine luminal fluid.
Not finding a soluble form of TGF in the uterine lumen implies that it is either not present or that it is present in a lower concentration. Casslen and Ohlsson found that fluid secreted into the uterine lumen varied in volume from 20 to 100 µl throughout the menstrual cycle [23]. Considering that we used a maximum lavage volume of 2 ml (1:100–1:20 dilution), we were able to detect TGF to an undiluted concentration of 1000–200 pg/ml. In fibroblasts transfected with an EGFR expression plasmid, it was demonstrated that a minimal concentration of 5000–10 000 pg/ml of the 50-amino acid soluble form of TGF was required to obtain detectable tyrosine phosphorylation of the EGFR [17]. Our ability to detect TGF in the intraluminal fluid is, therefore, well below the demonstrated minimal threshold for bioactivity in these fibroblasts, with the added advantage of over-expressed receptors. To date, we are aware of no published report of TGF in human intraluminal fluid. By way of comparison, a protein known to be secreted by the endometrium into the luminal fluid, namely pregnancy-associated endometrial -2-globulin, has been demonstrated in intraluminal fluid at a concentration of 2–4 µg/ml [24]. This protein concentration is 200 000- to 400 000-fold in excess of the detection limits of our assay.
The absence of detectable TGF in the intraluminal fluid of the human uterus suggests that the signaling between endometrial TGF and the EGFR may occur through a juxtacrine or direct cell-to-cell contact mechanism. The demonstration of mRNA production for TGF and immunohistochemical localization of TGF to the surface endometrium directly adjacent to the uterine lumen throughout the menstrual cycle serves to complement the protein and ELISA data. The absence of a cleaved TGF species in the intraluminal fluid directly adjacent to tissue with active DNA transcription of TGF and active protein expression of TGF argues against the release of a soluble form of TGF as the mechanism for signal transduction in normal cycling endometrium. This mode of signal transduction is further supported by the notable absence of any soluble amino terminus species and the absence of any postcleavage cytosolic remnant in the immunoblot preparations. Our study does not rule out or directly address the possibility that a preimplantation embryo could trigger the local release of soluble TGF. The endometrium functions all the way up to embryo contact in cyclic anticipation of an embryo. Cyclic follicular growth and luteal secretion can continue throughout a woman's reproductive years regardless of fertilization and the subsequent intraluminal presence of an embryo. The ELISA, immunohistochemical, and immunoblot data from this study provide evidence of the intact full-length ligand in high concentration at the surface of the recipient endometrium in anticipation of the embryo.
Studies of uterine luminal fluid are limited because it is has been difficult to obtain uterine luminal fluid atraumatically. The literature contains several reports of studies in which intraluminal fluid was obtained by uterine lavage. Dawood and Fazleabas [25] reported a 77–100% recovery rate after uterine lavage with a double-barrel cannula at the time of laparoscopy in 45 women at various times in the menstrual cycle. Young et al. [26] demonstrated the reliable recovery of secreted endometrial proteins obtained by using an insemination catheter and a 2-ml saline intraluminal lavage. Using this collection method, these authors demonstrated a midluteal suppression of progesterone-associated endometrial protein elaborated into the intraluminal fluid after treatment with high-dose ethinyl estradiol-norgestrel emergency contraception. While both of these methods have the advantage of keeping the uterus in situ, they may cause losses of luminal fluid through the proximal tubal ostia and cervix and have the potential for trauma to the luminal epithelium. Trauma is particularly important when studying a protein expressed by the surface epithelial cells. In the study reported here, we used a method of collecting human uterine luminal fluid at hysterectomy without disrupting the surface epithelium. Lack of surface trauma is supported by the absence of endometrial membrane TGF in the lavage specimens and the retention of normal endometrial surface architecture as evidenced by apical surface staining for TGF in paraffin sections of the same endometrial tissue that was lavaged.
TGF was first identified as a polypeptide secreted from retrovirally transformed fibroblasts [27]. Secreted TGF is classically described as a water-soluble, acid-stable polypeptide that is cleaved from the full-length transmembrane form of the molecule. The smallest form is 50 amino acids long and shares about 30% structural similarity with EGF [28]. Derynck and coworkers [29] sequenced the cDNA for the transmembrane form of TGF from a renal carcinoma cell line and described a 160-amino acid protein, with an extracellular amino terminus containing the soluble 50-amino acid sequence, a 23-amino acid hydrophobic region that spans the membrane, and a 39-amino acid cytosolic region. Species of higher molecular weight have been observed, and this has been attributed to variable N- and O-glycosylation of the amino terminus [8, 10, 30–32]. There is evidence that the degree of glycosylation is dependent on specific tissue type [33]. Full-length transmembrane TGF of 18–20 kDa has been isolated from a variety of cell sources [10, 30, 31], a 25-kDa TGF species has been isolated from a transformed prostatic carcinoma cell line [32], and a 28-kDa species has been isolated from human endometrial membranes [8].
Several groups have demonstrated that the full-length transmembrane form of TGF is capable of activating the EGFR. Brachmann et al. [17] performed site-directed mutagenesis of the DNA coding for the full-length transmembrane TGF, resulting in a ligand that could not be cleaved to liberate a soluble form of the growth factor. The SV40 plasmid construct was then transfected into Chinese hamster ovary cells (CHO) that were devoid of EGF receptors. Solubilized full-length ligand was demonstrated to activate the EGFR as evidenced by receptor autophosphorylation in intact live cells, both A431 (human epidermoid carcinoma cell line rich in EGF receptor) and HERc (NIH 3T3 fibroblasts transfected with an EGFR expression plasmid), without the possibility of immunoprecipitated EGFR coming from the CHO cells. Further, this group demonstrated activation of the EGFR by the transmembrane form of TGF under conditions that required cell-to-cell contact. A monolayer of HERc cells, which do not secrete TGF, were layered over with cells expressing only the full-length transmembrane TGF ligand. Activation of the EGFR was demonstrated by autophosphorylation in the fixed monolayer as well the demonstration of c-fos induction, a recognized indication of downstream mitogenic activity. Similarly, Wong et al. [18] developed in baby hamster kidney cells a mutated transmembrane TGF that could not be cleaved, incubated these intact cells with subconfluent A431 cells, and demonstrated the activation of the EGFR via autophosphorylation with an associated increase in intracellular calcium uptake. In addition, Anklesaria and coworkers [19] took the murine hematopoietic progenitor cell line 32D, which expressed EGFR but had lost the ability to adhere, proliferate, or survive on the murine bone marrow stromal cell line GB1/6, and recovered this ability by expressing the full-length transmembrane TGF in the GB1/6 cell line. 32D cell islands were observed attached to the monolayer of GB1/6 cells containing the full-length transmembrane TGF with the induction of sustained proliferation as evidence of mitogenic response. This response was inhibited with the addition of soluble EGFR ligands. To consider the 32D islands as analogous to the embryo at nidation and the GB1/6 cells containing the full-length transmembrane TGF as the surface epithelium is enticing. At a minimum, however, these experiments [17–19] independently verify in a variety of tissues that TGF is biologically active while anchored as a transmembrane ligand.
There are additional observations that converge on the possibility that in certain nontransformed tissues, the full-length transmembrane TGF may be signaling its receptor as an intact molecule. Important considerations include the variable rate and incidence of cleavage of the 50-amino acid soluble peptide dependent on the tissue being studied, the fact that cleavage was first described and seen most consistently in transformed tissues, the inefficiency of cleavage as evidenced by the elaboration of a heterogeneous group of postcleavage species, and the evolutionary conservation of the full-length transmembrane TGF. Indeed, it would be most favorable for embryo cell-to-endometrial cell recognition to foster optimal nidation if the growth factor were anchored in the recipient endometrium. Further, the orderly proliferation of surface epithelium required to be renewed each cycle invokes this mechanism nicely as mitogenesis is signaled as a result of neighboring cell-to-cell contact to promote a monolayer of renewed epithelium.
Previous investigators [34–36] have reported the presence of EGFR in normal endometrium with suggestions of cyclic fluctuation in receptor concentration. Our findings are not inconsistent with these results. The intent of our experiments was not to identify variations in receptor or ligand concentration but rather to detect evidence of receptor activation and to identify which species of ligand was present at the time of this activation. Our results do demonstrate the presence of full-length TGF in the endometrium, with the notable absence of a soluble amino terminus species and any postcleavage cytosolic remnant for TGF, while concurrently demonstrating the autophosphorylation of EGFR in the same tissue. The demonstration of activated EGFR throughout the menstrual cycle in conjunction with the demonstration of the full-length transmembrane TGF species resident in the endometrium at the same time throughout the menstrual cycle suggests that a juxtacrine mode of signaling is operative for the TGF/EGFR pathway in nontransformed, normal cycling endometrium.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence and current address: Mark R. Bush, Division of Reproductive Endocrinology and Infertility, MCHG-OG, Madigan Army Medical Center, Tacoma, WA 98431. FAX: (253) 968-2558; e-mail:MAJ_Mark_Bush{at}smtplink.mamc.amedd.army.mil
Accepted: August 11, 1998.
Received: March 3, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
in the mouse uterus during the peri-implantation period. Biol Reprod 1991; 45:365–372.[Abstract] precursors in the mouse uterus during the periimplantation period and after steroid hormone treatments. Biol Reprod 1994; 50:481–491.[Abstract] and a functional epidermal growth factor receptor (EGFR) on the basolateral surface of the trophectoderm in the mouse blastocyst. Dev Biol 1992; 154:396–409.[CrossRef][Medline], and EGFR expression in human preimplantation embryos. Development 1995; 121:299–307.[Abstract]: a transmembrane, surface epithelial protein that transiently disappears during the midsecretory phase of the menstrual cycle. J Soc Gynecol Invest 1997; 4:160–166.[CrossRef][Medline]. Adv Cancer Res 1992; 58:27–52.[Medline] species are derived from a glycosylated and palmitoylated transmembrane precursor. Cell 1987; 48:429–440.[CrossRef][Medline]. Nature 1985; 313:489–491.[CrossRef][Medline] precursor with a protein kinase complex. J Cell Biol 1994; 125:903–916. are palmitoylated and mediate cytoplasmic protein association. J Biol Chem 1996; 271:28502–28508. precursors activate EGF/TGF- receptors. Cell 1989; 56:691–700.[CrossRef][Medline] precursor expressed on the cell surface binds to the EGF receptor on adjacent cells, leading to signal transduction. Cell 1989; 56:495–506.[CrossRef][Medline] to epidermal growth factor receptors promotes cell proliferation. Proc Natl Acad Sci USA 1990; 87:3289–3293.-2-globulin, a glycosylated ß-lactoglobulin homologue, in the human fetus and adult employing monoclonal antibodies. Hum Reprod 1990; 5:487–493.: precursor structure and expression in E. coli. Cell 1984; 38:287–297.[CrossRef][Medline]. J Biol Chem 1988; 263:3924–3929.. Nature 1987; 326:883–885.[CrossRef][Medline]. Cell 1988; 54:593–595.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  L. G. Sheffield and J. J. Gavinski Proteomics Methods for Probing Molecular Mechanisms in Signal Transduction J Dairy Sci, July 1, 2003; 86(13_suppl): E115 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. G. Sheffield and J. J. Gavinski Proteomics methods for probing molecular mechanisms in signal transduction J Anim Sci, March 1, 2003; 81(suppl_3): 48 - 57. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. T. Teng, W. Gladwell, C. Beard, D. Walmer, C. S. Teng, and R. Brenner Lactoferrin gene expression is estrogen responsive in human and rhesus monkey endometrium Mol. Hum. Reprod., January 1, 2002; 8(1): 58 - 67. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. G. Richards, D. M. Klotz, M. R. Bush, D. K. Walmer, and R. P. DiAugustine E2-Induced Degradation of Uterine Insulin Receptor Substrate-2: Requirement for an IGF-I-Stimulated, Proteasome-Dependent Pathway Endocrinology, September 1, 2001; 142(9): 3842 - 3849. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
