| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Schepens Eye Research Institute,4 Department of Ophthalmology, Harvard Medical School, Boston, Massachusetts 02114
Departments of Clinical Science, Intervention, and Technology5 and Woman and Child Health,6 Division of Obstetrics and Gynecology, Karolinska Institutet, Stockholm 171 77 Sweden
ABSTRACT
In order for the preimplantation embryo to implant into the uterus, the trophoblast cells must initially adhere to the uterine epithelial surface. In preparation, the luminal secretory cells of the epithelium lose their nonadhesive character and their surface microvilli and bulge into the lumen, forming uterodomes (pinopodes; uterodome is used instead of pinopode, since in humans the surface membrane exocytoses rather than endocytoses (Murphy, Hum Reprod 2000; 15:2451–2454). Previous research has led to the hypothesis that loss of the nonadhesive membrane-spanning mucin MUC1 from the uterodome surface allows trophoblast adherence. Immunofluorescence microscopic assay of luminal epithelia on human uterine biopsies taken from LH+0 to LH+13 show that another membrane-spanning mucin, MUC16, was lost from uterodome surfaces in all samples taken during the receptive phase, LH+6 to LH+8 (n = 12), and that MUC1 was present on uterodomes in 4 of 12 samples and on all ciliated cells of the epithelium in the receptive phase. Short interfering RNA (siRNA) knockdown of MUC16 in a uterine epithelial cell line ECC-1 that, like uterine epithelium, expresses MUC16 and MUC1 allowed increased adherence of cells of a trophoblast cell line. In parallel experiments, siRNA knockdown of MUC1 did not affect trophoblast cell adherence. These data indicate that MUC16 is a membrane component of the nonreceptive luminal uterine surface, which prevents cell adhesion, and that its removal during uterodome formation facilitates adhesion of the trophoblast.
female reproductive tract, implantation, trophoblast, uterus
In order for a mammalian preimplantation embryo to continue its development past the blastocyst stage, it must attach to the luminal epithelium of the uterus and implant into the uterine wall. Events orchestrating implantation are a complex, successive set of genetic and cellular interactions involving both hormonal preparation of the receptive uterine surface and subsequent interactions between the developing blastocyst and the uterus (for review, see Wang and Dey [1]). Recent studies profiling gene and protein expression in the uterus and preimplantation embryo, in combination with gene knockout models in mice, have provided a wealth of candidate molecules that may be important in human implantation (for review, see Wang and Dey [1], Giudice [2], and Horcajadas et al. [3]). Despite these advances, there is little direct information regarding the molecular changes that occur on the apical surface of human uterine epithelial cells in preparation for the attachment and subsequent implantation of the blastocyst.
Prior to the receptive period, the glycocalyx present on the uterine luminal epithelial surface provides a barrier to cell adherence, and is thus a nonadhesive surface [4]. It has been hypothesized that loss of the antiadhesive surface is required for attachment of the blastocyst (for review, see Aplin [5]). There are two types of cells in the uterine epithelium that form the luminal uterine surface—the secretory cells and the ciliated cells [6]. It is well documented across mammalian species, including humans [6–10], that the apical surfaces of secretory cells of the luminal uterine epithelium undergo dramatic morphologic change in preparation for the receptive period, which in humans is associated with concentration of progesterone receptors [6]. The surface of the secretory cells bulges out into the lumen to form uterodomes, the microvilli present on the apical cell membrane are lost, and the surface of the uterodomes appears smooth and flat [11, 12]. Ciliated cells intercalated within the secretory cell population do not change their morphology during the receptive phase [9]. In the human, the apical surface morphologic changes begin in the secretory cells around LH+6, reach their maximum between days LH+7 and LH+9, and then regress [6]. Although the alterations in the molecular character of the membrane that correlate to the morphologic change and render the cell surfaces adhesive to trophectoderm cells have been the subject of recent research (for review, see Wang and Dey [1], Aplin [5], Murphy [12], and Aplin and Kimber [13]), less emphasis has been placed on the changes that occur in the membrane that make it lose its nonadhesive character.
Interest in the change from nonadhesive to adhesive surface character of the endometrium has, to date, centered primarily on MUC1, one of the membrane-associated mucins present in the glycocalyx of wet-surfaced epithelia [14–16], including the uterine surfaces of many species (for review, see Murphy [12]). Studies in mice report that the mucin disappears from the uterine surface at the receptive phase [15, 17] and that the removal of the ectodomain of the mucin may be by the proteolytic action of TNF-converting enzyme (TACE), an extracellular metalloprotease-disintegrin tumor necrosis factor (TNF)-converting enzyme, also known as ADAM 17 [18]. In humans, there are contradictory reports that MUC1 is increased during the receptive phase [19], is lost from the uterodome surface [20, 21] or, as shown using an in vitro technique, is lost locally in the region of the implanting blastocyst [22].
In addition to MUC1, another membrane-associated mucin, MUC16, previously known as the ovarian tumor cell marker CA125 [23, 24], has been reported to be expressed by the normal human uterine epithelium and uterine adenocarcinomas [25–27]. Compared with MUC1, MUC16 has a larger glycosylated ectodomain [14, 23, 24]. Like MUC1, its ectodomain is shed from cell surfaces and, since it too is localized in the glycocalyx, MUC16 could potentially contribute to the formation of a nonadhesive barrier that prevents cell adherence [28–30]. Since the distribution of MUC16 during the prereceptive and receptive stages of the uterine epithelium has not been examined, we have screened human biopsies taken through the cycle by immunohistochemistry and compared the MUC16 distribution to that of MUC1. Finding loss of MUC16 from the uterine luminal surface during the receptive phase, we assayed the effect of siRNA knockdown of MUC16 and MUC1 in a uterine cell line (ECC-1) on trophoblast cell adhesion.
Tissues used for these studies were used in previous work [6, 31, 32]. Tissue collection was done in compliance with the principles expressed in the Declaration of Helsinki and with the approval of the ethics committees at the Karolinska University Hospital (Stockholm, Sweden) and the Schepens Eye Research Institute (Boston, MA). All participating women gave informed consent to participate. Endometrial biopsies were obtained from healthy, normally cycling women ages 21–44 yr, all of whom were proven fertile. None of the women had been using steroid hormone contraceptives or an intrauterine device for 3 mo or more before the study, nor had they been pregnant or had a history of inflammatory pelvic disease within the previous 12 mo. The biopsies originated from the anterior wall of the uterine cavity and were obtained without dilatation of the cervix using a Randall curette (Stille Werner AB, Stockholm, Sweden). Tissues were collected for immunohistochemistry and RNA analysis.
For biopsies taken for RNA analysis, cycle day was determined by the LH surge date, as measured in serum. Fourteen samples were collected and assayed; two were at the proliferative or pre-LH surge stage, four at early secretory (LH+1 to LH+5), four at mid-secretory (LH+6 to LH+8), and four at late secretory (LH+9 to LH+13).
Twenty-four biopsies were taken for immunohistochemistry and scanning electron microscopy. The day of the LH surge was identified by each participant by morning urine assay (Clearplan; Unipath Ltd., Bedford, UK). Of the biopsies, three were proliferative, one LH+0; one LH+2; one LH+3; one LH+4; one LH+6; nine LH+7; two LH+8; one LH+9; two LH+10; one LH+11; and one LH+13. In eight of the biopsies from the mid-luteal phase (LH+6 to LH+8), uterodomes were present, as determined by scanning electron microscopy [6]; in one sample, uterodome presence was determined by confocal microscopy.
Immunohistochemistry and Confocal Analysis
Staged endometrial biopsies were fixed in 4% (w/v) formalin, dehydrated in a series of ETOHs, and embedded in paraffin. Serial sections (4 µm) were cut and processed for routine immunofluorescence microscopy using antibodies to MUC1 (anti-MUC1 episialin, clone 214D4; Upstate, Lake Placid, NY [33]) and MUC16 (MUC16-anti-human CA125, clone OC125; DAKO Corp., Carpinteria, CA). Antigen retrieval methods were used for MUC16 localization as recommended by the manufacturer. As control, antigen retrieval methods were applied to sections used for MUC1 localization to make certain that the method did not alter its localization. The secondary antibody was affinity-purified fluorescein isothiocyanate (FITC)-conjugated donkey anti-mouse IgG. Sections incubated with secondary antibody only served as control. Nonimmune primary antibodies of the same isotype as that of the MUC1 and MU16 antibodies (IgG1; Chemicon, Billerica, MA) were also used on sections for both localizations as control. Images of MUC1 and MUC16 antibody binding were taken on adjacent sections of all tissues. Images were acquired using either a Zeiss Photoscope III equipped with a mercury bulb and FITC and tetramethylrhodamine isothiocyanate (TRITC) filter sets or a Leica TCS-SP2 confocal scanning laser microscope equipped with a PLAP0100X 1.4 oil objective lens.
Confocal images of FITC- and TRITC-labeled sections were acquired simultaneously (in two channels) using a double dichroic filter and both an argon/krypton laser (488-nm wavelength) and a helium neon laser (543-nm wavelength). Sections were mounted with Vectashield containing propidium iodide (Vector Laboratories, Burlingame, CA). The propidium iodide stains nuclei, allowing for demonstration of position of antibody binding relative to tissue structure. Images were acquired using a Standard Bit mode XYZ and a high-resolution (1024 x 1024) format.
RNA was extracted from endometrial biopsies using the SV total RNA isolation system (SDS; Promega, Madison, WI) according to the manufacturer's instructions. First-strand cDNA was synthesized according to the manufacturer's protocol from 1 µg DNase-treated RNA with random primers using SuperScript II-Reverse Transcriptase (Invitrogen, Carlsbad, CA) followed by RNase H treatment as previously described [34].
Relative expression of MUC1 and MUC16 was determined using real-time quantitative RT-PCR performed, as previously described [35, 36], on the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA) using TaqMan chemistry and the delta Ct method (User Bulletin 2; Applied Biosystems) [37]. The level of mucin mRNA expressed at the four stages of the cycle were compared using proliferative phase samples as the calibrator [35]. Statistical comparisons of the real-time RT-PCR results were performed using the Fisher protected least-significant difference test (Statview 5.0 for MacIntosh; SAS Institute Inc., Cary, NC). P < 0.05 was considered significant. Primers and probes for MUC1 [38] and MUC16 [28] were reported.
Knockdown of MUC16 and MUC1 in ECC-1 Cells
Cell line. A uterine epithelial cell line designated ECC-1 [39] and obtained from Dr. Bruce Lessey, Oncology Research Institute (Greenville, SC) was used for testing the effect of siRNA knockdown of MUC16 and MUC1 on trophoblast cell adhesion. Choice of the cell line is appropriate, since the cells exhibit characteristics of luminal epithelial cells [40]. ECC-1 cells were grown in Dulbecco modified Eagle medium/Ham nutrient mixture F12 (DMEM/F12; Hyclone, Logan, UT) plus 10% heat-inactivated fetal calf serum and 100 units penicillin/streptomycin (DMEM/F12-Complete; Hyclone). In preliminary real-time RT-PCR experiments, MUC16 and MUC1 mRNA were detected in the cell line (data not shown). Since preliminary transient transfections yielded a low percentage of MUC16 knockdown, and since optimal mucin expression requires culture to cell confluence, a longer time period than transient transfection knockdown of gene expression allows, stable transfections were required.
MUC16 siRNA transfection. Three siRNA sequences targeting human MUC16 were chosen using software from OligoEngine Inc. (Seattle, WA). A BLAST homology search indicated that the selected siRNAs would not affect any other known human gene. The sequences are: no. 1, sense, 5'-TCCGCTAGCGCAGCAACATCAA-3', antisense, 5'-GGATTGATGTTGCTGCGCTAGC-3'; no. 2, sense, 5'-TCGACAGCTAGCGCAGCAGCAACAC-3', antisense, 5'-TCGAGTCTTGCTGCGCTAGCTG-3'; no. 3, sense, 5'-TAACCATCACCACCCAAAC-3', antisense 5'-GTTTGGGTGGTGATGGTTA-3'. Custom complimentary oligonucleotides were synthesized, annealed, and ligated into pSUPER.retro.puro (OligoEngine), a mammalian expression vector that directs intracellular synthesis of siRNA transcripts [41]. The ligated vectors were transformed into chemically competent Escherichia coli (DH5
cells), correct orientation of the hairpin siRNA insert was verified with DNA sequencing, and large-scale cultures were grown to obtain a sufficient amount of plasmid vector. Using PolyFect transfection reagents (Qiagen, Valencia, CA), the vectors were transfected into 293–10A1 cells, a packaging cell line that produces high-titer retrovirus in culture (ATCC, Manassas, VA).
ECC-1 cells were seeded in 24-well plates at 2 x 104 cells per well and grown to 40%–50% confluence in serum-free growth medium. Tissue culture medium from transfected 293–10A1 cells [42] was filtered 48 h after transfection, and the viral supernatant was used to infect cultures of the subconfluent ECC-1 cells after the addition of 4 µg/ml polybrene. Isolated clones were obtained using antibiotic selection (puromycin) and further expanded to confluence to obtain stably transfected cells. The transfected ECC-1 cells were grown to confluence in DMEM/F12-Complete (Hyclone). RNA was extracted from the cells, and knockdown of MUC16 expression was verified by real-time RT-PCR. MUC16, MUC1, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) RNA levels as well as MUC16 and MUC1 protein levels were compared in stably transfected, nontransfected, and empty vector transfected cells. All cells were plated at equal density and grown to confluence. Protein was extracted and separated on 1% agarose gels followed by transfer to nitrocellulose with vacuum blotting [43]. Immunoblots were performed using antibodies to MUC16 (OC125), MUC1 (214D4), and GAPDH, as previously described [36, 43]. Comparison of MUC16 and MUC1 protein levels was done by densitometry [36, 43].
MUC1 siRNA transfection. Techniques employed for siRNA knockdown of MUC1 in ECC-1 cells were previously reported for knockdown of MUC1 in a pancreatic cancer cell line (S2–103) [44]. Plasmids (pSUPER.neo+GFP; OligoEngine) containing two sequences of MUC1-specific siRNAs (designated seq 1 and seq 2) plus empty plasmid were generously provided by Dr. Michael Hollingsworth (University of Nebraska Medical Center/Eppley Cancer Center, Omaha, NE). Plasmids were amplified in chemically competent E. coli (DH5 cells). ECC-1 cells were seeded in 12-well plates at 2 x 104 cells per well and were grown to 40% confluence. Medium was removed, and cells were washed three times in PBS, then incubated for 5 h with 1 µg plasmids plus 20 µg PolyFect transfection reagent in DMEM/F12. Cells were then washed in DMEM/F12 and cultured for 24 h in DMEM/F12-Complete. Transfected cells were then selected by culture in the presence of 500 µg neomycin per milliliter media (Sigma-Aldrich, St. Louis, MO) for 48 h, followed by serial dilutions of the cells to obtain single-cell clones. Cells then were grown in DMEM/F12-Complete for trophoblast adhesion experiments. Controls for the experiments were nontransfected ECC-1 cells and those transfected with empty plasmids.
Trophoblast adhesion assay. Transfected ECC-1 cells were seeded in one-well chamber slides and grown to confluence in DMEM/F12-Complete. A suspension of cells of a telomerase-transformed trophoblast cell line (SW-71) cultured initially in DMEM/F12-Complete [45] was prepared in serum-free DMEM/F12 after detachment with enzyme-free cell dissociation buffer (Invitrogen). The cells were washed in DMEM/F12 and labeled with fluorescent dye, 6-CFDA (0.8 mg/ml media; Invitrogen) for 30 min at 37°C. Before incubation with labeled SW-71 cells, the transfected ECC-1 cells and controls (nontransfected and empty vector-transfected cells) were washed with 1x HEPES buffer. The labeled SW-71 cells were washed in DMEM/F12, and 5 x 105 cells in 1 ml cell suspension was incubated in each well with confluent transfected ECC-1 cells and controls at 37°C for 1 h [46]. The cell suspension was removed, and the slides were gently washed five times with PBS. Cells were fixed with 4% (w/v) paraformaldehyde and photographed at 10x using fluorescence and phase-contrast microscopy. Eight pictures were taken of each slide (n = 6), and the number of bound SW-71 cells was manually counted and expressed as a number of adherent SW-71 trophoblast cells per 10x field (0.06-mm2 culture surface). Statistical comparisons of adherent cells were performed using the Fisher protected least-significant difference test (Statview 5.0). P < 0.05 was considered significant.
Localization of MUC16 in Human Uterine Epithelia
Antibodies to MUC16 localized to apical membranes of both luminal and glandular epithelia in three of three proliferative phase biopsies and four of four early secretory (LH+0 to LH+4) biopsies (Fig. 1 and Table 1). The binding appeared continuous across the surface of the luminal epithelium. MUC16 was present along apical surfaces in six of seven samples of the glandular epithelia of the proliferative and early secretory samples (one biopsy had no glandular tissue; data not shown). MUC1 showed binding patterns comparable to MUC16 in these specimens, except that binding was in some cases intermittent along the epithelial surface (compare Fig. 1, B to D).
|
|
Apical cells of all 12 samples taken during the receptive, mid-secretory stages, LH+6 to LH+8, lacked MUC16 on their luminal epithelial surfaces (Fig. 2, A, C, and E, and Table 1). The lack of MUC16 antibody binding correlated with the presence of uterodomes on the secretory cells (Table 1). In 6 of 12 mid-secretory samples, MUC16 was present on glandular epithelium (Fig. 2A and Table 1). In contrast to MUC16, MUC1 was present on the luminal surface of all 12 mid-secretory stage samples. In four of the samples, MUC1 appeared on uterodomes (Fig. 2D), and in three samples, binding appeared weak and intermittent. In all of the samples, binding of the MUC1 antibody was present on ciliated cells, and in two of the samples, only ciliated cells were labeled (Fig. 2F and Table 1). Of the 12 samples, 11 had MUC1 on their glandular epithelial surfaces (Fig. 2B and Table 1).
|
In late secretory samples, LH+9 to LH+13, MUC16 was again present, with four of the five samples showing luminal antibody binding (Fig. 2G and Table 1). The one sample that did not show binding was the LH+13 sample, the latest cycle stage examined. Similarly, MUC1 binding also was present along apical cell surfaces of the luminal epithelium in late secretory samples, and the binding was present on secretory cells (Fig. 2H). Glandular epithelia in all samples had both MUC16 and MUC1 on their apical surfaces (Table 1 and data not shown).
In summary, these immunohistochemical data show that MUC16 is lost from uterodome surfaces present in samples obtained from the receptive phase, LH+6 to LH+8. MUC1 was present on all of the samples with uterodomes but was particularly evident on the ciliated cells intercalated between uterodomes. Glandular epithelia retained both mucins during the receptive phase.
MUC16 and MUC1 RNA in Endometrium
Messenger RNA levels of MUC16 were significantly increased in late secretory samples compared with early and mid-secretory samples (Fig. 3). The loss of MUC16 from luminal surfaces at mid-secretory stage was not reflected in a significant decrease in MUC16 mRNA compared with earlier stages (Fig. 3). MUC1 showed a trend toward increased mRNA with cycle, but the difference in levels did not reach significance, perhaps due in part to a low number of proliferative samples available for assay (n = 2).
|
Knockdown of MUC16 and MUC1 with RNAi and its Effect on Trophoblast Adherence
To determine the effect of loss of MUC16 on trophoblast adherence, we developed an in vitro adherence assay using the trophoblast cell line SW-71, a recently described telomerase-transformed cell line [45], and ECC-1 cells that express both MUC16 and MUC1 (Fig. 4). The ECC-1 cell line appears ideal since it polarizes in culture and expresses MUC16 and MUC1 apically (Fig. 4, A–C) similar to native uterine epithelia.
|
Two of the three MUC16 siRNA sequences were found to be effective in knocking down expression in ECC-1 cells (the third was inactive; data not shown). Short interfering RNA sequences 1 and 2 significantly reduced MUC16 mRNA by 45% and 89%, respectively, with no effect on MUC1 mRNA compared with nontransfected and vector-transfected controls (n = 3, P < 0.01; Fig. 5A). In addition, the expression of MUC16 protein was decreased compared with controls in the siRNA-expressing ECC-1 cells. Short interfering RNA sequences 1 and 2 reduced MUC16 protein by 19% and 86%, respectively, compared with controls (n = 3, P < 0.01; Fig. 5, B and C). MUC1 protein levels were not significantly affected by the MUC16 siRNA (Fig. 5, B and C). Immunofluorescence microscopy performed on confluent cultures of nonpermeabilized ECC-1 cells also confirmed that there is reduced cell surface localization of MUC16 in the transfected cultures compared with control (compare Fig. 4, D and E to F).
|
Of the two MUC1 siRNA sequences stably transfected into the ECC-1 cells, sequence 2 was found to be the most effective, with 71% reduction of MUC1 mRNA and 87% reduction of MUC1 protein (Fig. 5) compared with 34% reduction of RNA and 17% reduction of protein by siRNA sequence 1. MUC16 mRNA and protein levels were not affected by MUC1 knockdown (Fig. 5, A and C). Knockdown of MUC1 in the ECC-1 cells was confirmed by immunofluorescence microscopy on nonpermeabilized cells (compare Fig. 4, G and H to I).
To test the effect of MUC16 and MUC1 on adherence of trophoblast cells, parallel adhesion assays were performed using the ECC-1 cells stably transfected with MUC16 siRNA sequence 2 that exhibited 86% knockdown of MUC16 protein, and ECC-1 cells stably transfected with MUC1 siRNA sequence 2 that knocked down 87% of MUC1 protein (Fig. 6). The number of adherent CFDA-labeled trophoblast cells (Fig. 6E) was significantly higher (approximately 3.8-fold higher in the transfected cultures of ECC-1 cells expressing MUC16 siRNA sequence 2) compared with nontransfected and vector-transfected controls (n = 6, P < 0.01). There was no statistical difference in the number of adherent trophoblast cells on the ECC-1 cells expressing MUC1 siRNA sequence 2 compared with controls (Fig. 6E). These data suggest that loss of MUC16 may be necessary to facilitate adherence and subsequent implantation of the blastocyst.
|
The data from this study demonstrate for the first time a loss of the membrane-associated mucin MUC16 from the surface of uterodomes—the apical cell surfaces of human luminal epithelium that form during the receptive phase of the female cycle. These data, in addition to the data from in vitro assays that showed that trophoblast cell adhesion to ECC-1 cells was significantly increased when MUC16 protein was reduced by siRNA methods, indicate that MUC16 contributes to the barrier on the endometrial surface and prevents cell adhesion in the nonreceptive uterus. Thus, MUC16 removal on uterodomes may facilitate implantation. Our data indicate the loss of the membrane-associated mucin MUC1 from some of the biopsies with uterodome surfaces, as reported by Horne et al. [21], but also demonstrate retention of MUC1 on some uterodome surfaces and on the ciliated cells of the uterine epithelium during the receptive phase. The loss of MUC16 appears, however, to be more complete than that of MUC1, and the in vitro assays showing no increase in trophoblast adherence to cells in which MUC1 protein was reduced by siRNA methods suggest that MUC16 is a more effective barrier to trophoblast adherence.
MUC16 and MUC1 may perform tissue-specific barrier functions due to the similarity in the presence of tandem repeat sequences of amino acids in their protein backbones, which are rich in serine, and threonine, and are thus heavily O-glycosylated, but they each have unique structural features and characteristics that suggest additional unique functions [14]. For example, the glycosylation of the two mucins has been shown to differ in ocular surface epithelia, which may impart a unique function [28]. In addition, although both mucins have short cytoplasmic tails, their sequences differ (compared with databases). Because the cytoplasmic tail of MUC1 has been studied in detail, MUC1 has been described as a multifunctional molecule, with its extracellular domain providing a nonadhesive surface glycocalyx function and its short cytoplasmic tail functioning as a signaling component [14]. Data from studies in breast tumor cells suggest that the cytoplasmic tail can be phosphorylated, can associate with beta catenin, and can, when freed from the surface membrane, be transported with the catenins into the nucleus of the cell [47]. To our knowledge, the cytoplasmic tail of MUC16 has not been investigated as to signaling function. Perhaps the two mucins have, in addition to nonadhesive/barrier function, other unique functions in the human endometrium. There remains the possibility that additional membrane-spanning mucins are expressed by the human uterine surface (e.g., MUC4) and that they also may contribute to the adherence barrier on the uterine luminal surface.
The mechanism of loss of MUC16 and MUC1 from the luminal epithelial surface in mid- secretory/receptive phase in the human is not well understood. The possibilities include downregulation of the MUC genes, metabolic turnover, and/or ectodomain shedding of the mucin from the cell surface, the latter being a common feature of membrane-associated mucins (for review, see Hollingsworth and Swanson [14]). The data on MUC16 mRNA levels in this study suggest a comparable amount of mRNA in early secretory and mid-secretory stages, without an abrupt downregulation in mid-secretory stage that would account for loss of luminal MUC16. The significant increase in MUC16 mRNA at late secretory stage does not correspond to loss of MUC16 at midcycle but could account for the return of luminal MUC16 after the receptive phase. It should be noted, however, that the biopsy samples included both glandular and luminal epithelia, and MUC16 protein was present in glandular epithelium of 6 of 12 mid-secretory samples (LH+6 to LH+8) assayed by immunohistochemistry (Table 1). Presence of glandular mRNA for the mucin may obscure loss of mRNA in luminal epithelium. It should also be noted that due to the difficulty in obtaining human samples, the number of samples assayed is small and, as we have previously noted, there are large variations in mucin gene expression profiles in human individuals' cervical tissue over the cycle [48].
Messenger RNA for MUC1, as measured by real-time RT-PCR, showed a pattern similar to MUC16 mRNA in early secretory and mid-secretory stages, with comparable amounts of mRNA. These data and data showing an increase in mRNA (although not significant) in late secretory stage are not comparable to those in a previous report using densitometric analysis of Northern blots, which showed that MUC1 expression was greatest in early secretory stage, then gradually declined in mid-secretory and late secretory stages [49]. As with MUC16, MUC1 loss from some samples in mid-secretory stage was not reflected in dramatic mRNA decrease. However (also like MUC16), MUC1 protein was present in glandular tissue of 11 of 12 samples assayed by immunohistochemistry. To determine whether the mRNA is downregulated in luminal epithelium, a more precise measurement of mRNA in these cells alone is required.
The possibility of enhanced shedding of the mucins from the cell surface with uterodome formation may be likely, since both MUC16 and MUC1 are shed from cell surfaces and are normally present as soluble forms in secretions of epithelia [14]. The mechanism of shedding of those membrane-associated mucins that have been studied is by sheddases. Specific sheddases for MUC16 have not been reported, but in studies of the HES uterine epithelial cell line, Thathiath et al. [18] demonstrated that MUC1 shedding induced by PMA (phorbol 12-myristate 13-acetate) was mediated by TNF-converting enzyme (TACE), also known as ADAM 17 as well as MT1-MMP. As demonstrated by immunohistochemistry, TACE is expressed in human luminal endometrial epithelia at mid-luteal phase, and it forms a stable association with MUC1 in isolates from the HES cells [18].
If MUC16 were shed from the surface of the luminal epithelium with uterodome formation, an increase in MUC16 in cervical mucus might occur. Previous studies have demonstrated the presence of soluble MUC16, or CA125 antigen, in cervical mucus and sera of normally cycling women. Analysis of CA125 levels in cervical mucus of 20 normally cycling women showed high concentrations of the mucin in cervical mucus, with an increase in amount during the periovulatory period [50]. Another study measured CA125 serum levels of 33 women before and during hormonally induced cycles and found a rise in serum levels from late proliferative to early secretory phase [51]. A third study of 132 women found that CA125 concentrations in uterine flushings varied throughout the menstrual cycle, with highest concentrations occurring during the mid-follicular phase (Days 6–10) [52]. These studies did not show a dramatic rise in cervical CA125 levels just prior to the receptive phase but, as with the RNA data described above, glandular contribution of the mucin may obscure MUC16 shedding from the luminal surface. In addition, cervical epithelia express MUC16 [27], and thus will contribute to the MUC16 in the cervical mucus samples.
The in vitro MUC16 and MUC1 knockdown data provided herein indicate that loss of MUC16 enhances trophoblast adherence and that the molecule provides a barrier to blastocyst adherence. The fact that knockdown of MUC1 in the ECC-1 cells did not effect trophoblast cell adherence suggests that blastocysts may adhere even if MUC1 is present. The presence of both mucins may, however, be additive in preventing adherence in vivo. As with all in vitro assays, the level to which the cells in vitro mimic the concentration of mucins on the native epithelial surface is uncertain.
In summary, data from these studies indicate that MUC16 contributes to formation of a barrier to blastocyst adherence, and that the loss of MUC16 from the uterodome facilitates adherence in human luminal uterine epithelium.
ACKNOWLEDGMENTS
We wish to thank Ms. Beatriz Perez for her technical assistance.
FOOTNOTES
3 These authors contributed equally to this work. 
1Supported by National Institutes of Health/National Eye Institute (NIH/NEI) R01 EY03306 to I.K.G., and NIH/NEI F32 EY016937 to T.B.
3These authors contributed equally to this work.
Correspondence: 2Ilene K. Gipson, Schepens Eye Research Institute, 20 Staniford St., Boston, MA 02114. FAX: 617 912 0126; e-mail: Ilene.Gipson{at}schepens.harvard.edu
Received: 26 October 2006.
First decision: 10 November 2006.
Accepted: 26 September 2007.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
