MUC1 and endometrial receptivity

Molecular Human Reproduction vol.4 no.12 pp. 1089–1098, 1998

MUC1 and endometrial receptivity

M.Meseguer1,2, A.Pellicer1,2 and C.Simoń1,2,3 1Department

of Pediatrics, Obstetrics and Gynecology, Valencia University School of Medicine and 2Instituto Valenciano de Infertilidad Valencia, Spain

3To

whom correspondence should be addressed at: Guardia Civil 23, 46020 Valencia, Spain

Mucins, best known as the major constituent of mucus, are a family of highly glycosylated, high molecular weight (≥ 200 kDa) glycoproteins present on the surface of human epithelial cells. MUC1 has the features of an integral membrane protein. It has an extracellular tandem repeat domain that forms the major part of the core protein, and results in a highly repetitive structure, which is extremely immunogenic. In the protein there is also a proteolytic cleavage site reported in the proximal extracellular domain. The secreted form of MUC1 lacks the cytoplasmic tail, but it is not clear whether this results from alternative splicing or proteolysis and release of the free extracellular domain. The locus of the MUC1 gene is on band 21 of the long arm of chromosome 1 (1q21). Anti-adhesion properties of this mucin are probably the result of the unique structure of the molecule. In mouse uterine epithelium, the homologue MUC1 is regulated with reduced expression in the implantation period, but in humans, expression is high during the peri-implantation period. MUC1 may inhibit the interaction between trophoblast and apical epithelium adhesion molecules at the time of implantation, giving the possibility of forming a uterine barrier for implantation. Key words: anti-adhesion/endometrium/implantation/menstrual cycle/MUC1

Introduction Human endometrium is the most fascinating tissue model for the study of sex steroid-morphophysiological interactions. In recent years, there has been a veritable explosion of research in ultrastructure, steroidobiochemistry, immunohistochemistry, protein chemistry, and molecular biology of human and animal endometria (Ferenczy and Gurlnick, 1983; Bulletti et al., 1991; Seppa¨la¨ et al., 1991; Yen et al., 1991; Giudice, 1994). The results of these studies show that the biological phenomena operating in the endometrium are inter-related, having as their primary goal the development of an appropriate environment for the implantation of the conceptus. The cyclic changes in the endometrium are under the control of ovarian oestrogens and progesterone via their respective specific receptors as well as other peptides and enzymes. The endometrial cycle follows a precisely regulated series of morphophysiological events characterized by proliferation, secretory differentiation of oestrogen-primed endometrium and in the absence of conception, degeneration and regeneration (Giudice and Ferenczy, 1996). In oestrous mammals the luteal phase is vastly reduced, there is little or no proliferation with no abrupt changes and, therefore, no menstruation. In some oestrous species, e.g. the rabbit, ovulation is triggered by sexual intercourse during the oestrous phase (period when the follicle is ready for ovulation). In the rat, the cycle repeats every 4 days, and includes dioestrous, pro-oestrous, oestrous and meta-oestrous phases; ovulation is spontaneous (it takes place in the oestrous phase), but not the activation of the corpus luteum, which depends on nervous signals from the mechanical stimulation of the cervix at about the time of ovulation (Eckert et al., 1990). © European Society of Human Reproduction and Embryology

Embryo implantation is a highly controlled physiological process (Simoń et al., 1996). In rodents, implantation is known to occur only during a certain time in pregnancy referred to as the ‘window’ of implantation (Iizuka et al., 1988). It occurs on day 4 of pregnancy, and the receptive phase lasts ,24 h (Aplin, 1996). Prior to this phase, the uterus is refractory to embryo attachment and subsequent implantation. The opening of this window and the process of implantation are known to be controlled by ovarian steroid hormones. In various primate species, the uterus enters the receptive phase ~8–10 days postovulation (in a hormonal milieu rich both in oestrogen and progesterone) (Wolf et al., 1993; Knobil and Neill, 1994). In women, the window appears to last ~5 days, from days 20– 24 of the cycle; measured from the luteinizing hormone (LH) peak, which precedes ovulation by ~36 h, and gives a receptive phase lasting from approximately day LH17 to day LH111 (physiological implantation window). In humans, the operational definition (in terms of pregnancy success after embryo replacement) of the beginning of the receptive phase (clinical implantation window) is not as precise as the end; i.e. embryos replaced before day 20 may implant, while those replaced after day 24 will not (Bergh and Navot, 1992) Mucins are best known as the major constituent of mucus. They are a family of highly glycosylated, high molecular weight (250–500 kDa) glycoproteins present on the surface of human epithelial cells from the mammary glands, salivary glands, digestive tract, respiratory tract, kidney, bladder, prostate, uterus and rete testis. Mucins consist of a non-globular, thread-like protein backbone, which contains both highly glycosylated and non-glycosylated regions. Most of the carbo1089

M.Meseguer, A.Pellicer and C.Simoń

hydrate chains are attached by an O-glycosidic linkage between an N-acetyl galactosamine and the side-chain of serine or threonine, although some oligosaccharides are attached by a linkage between the nitrogen of asparagine and N-acetylglucosamine (N-glycosidic bond) (Devine and MacKenzie, 1992). Mucin core proteins contain high levels of serine, threonine, alanine, glycine and proline residues. This produces molecules with highly extended structures (Jentoft, 1990); as many as one in three amino acids may be glycosylated in mucins and the majority of the oligosaccharides are composed of N-acetyl galactosamine, N-acetyl glucosamine, galactose, fructose and neuraminic acid (sialic acid) (Devine and MacKenzie, 1992). Oligosaccharide assembly is primarily localized to the Golgi apparatus (Pimental et al., 1996). Complementary DNA clones have been isolated from nine human epithelial mucin genes and designated as MUC1-4, MUC5B, MUC5AC, and MUC6-8 (Gendler and Spicer, 1995). These glycoproteins fall into two categories: cell surface associated and true secretory mucins. The latter (MUC-2, -3, 4, -5, -6) self-associate, usually by disulphide bonding, to form large oligomers (Sheehan et al., 1990, 1991). Highly O-glycosylated and hydrated domains lie between the disulphide knots and display extended conformations (Sheehan et al., 1990; Devine and MacKenzie, 1992), so that the mucins in mucus serve as lubricants and protect the underlying epithelial cells (Hilkens et al., 1992). A feature of the O-glycosylated, serine and threonine-rich domains of mucins is the presence of repeat sequences whose length varies from 11 to 81 residues (Devine and MacKenzie, 1992; Straus and Dekker, 1992). High molecular weight secretory mucin is known to be produced by cervical tissue and to exhibit menstrual cycle stage-specific physical properties (Sheehan et al., 1990). It is likely that the endometrium also produces one or more large secretory mucins (Van Kooig et al., 1982). The cell surface associated mucins include leukosialin (CD43), glycophorin, ASGP-1 (ascites sialoglycoprotein 1), epiglycanin and MUC1. The latter molecule is also known as episialin, polymorphic epithelial mucin (PEM), epithelial membrane antigen (EMA), as well as by a variety of other names (Hilkens et al., 1992). Mucins present at the cell surface could act as a barrier between the cell membrane and the external environment, shielding the cell from micro-organisms, toxins or proteolytic attack (Devine and MacKenzie, 1992). The association between mucins and malignancy has been well documented. Apart from increased synthesis/secretion of mucins, altered glycosylation is probably responsible for many antigenic differences between carcinomas and adjacent normal epithelial tissue (Devine et al., 1990).

Structure of MUC1 Polyclonal and monoclonal antibodies to the MUC1 core protein were developed and used to screen a lgt11 expression library of cDNA made from mRNA of the mammary tumour cell line MCF-7. Seven cross-reacting clones were isolated, with inserts of 0.1–1.8 kb long. RNA blot analysis, using as a probe the 1.8 kb insert subcloned in plasmid pUC8 (pMUC10), revealed transcripts of 4.7 and 6.4 in this tumour cell line and 1090

restriction enzyme analysis detected a restriction fragment length polymorphism when human genomic DNA was digested with EcoRI or HinfI (Gendler et al., 1987). The locus of the MUC1 gene is on band 21 of the long arm of chromosome 1 (1q21) (Gendler et al., 1990a). In the 59 non-coding region of the MUC1 gene there are numerous base sequences that may potentially act as regulatory sites. Among these, there are possible consensus sequences for binding of oestrogen- and progesterone-receptor complexes (Lancaster et al., 1990), but their activity has not been demonstrated directly. A large domain in the gene containing a variable number of 60 bp tandem repeats causes the MUC1 locus to be a hypervariable ‘minisatellite’ region of human DNA similar to others described by several groups, but which is novel in that it is transcribed and translated. The same polymorphism is demonstrable in the expressed gene product (Swallow et al., 1987). This gene is composed of seven exons and varies in size from 4 to 7 kb, depending on the number of tandem repeats in exon 2 (Ligtenberg et al., 1990). Two different splice variants that produce two very similar protein forms have been found. The sole difference is the use of two alternative splice acceptor 59 sites for exon 2, located 27 bases apart. Therefore, the length of the putative translation products differs by only nine amino acids. However, the alternative splicing event affects the signal sequences of the gene products. According to the predictive method of Von Heijne (1986), the signal peptide of variant A is 22 amino acids long and is cleaved between the threonine and alanine residues present in the additional nine amino acids. The cleavage site of variant B is located between the glycine and serine residues 23 amino acids downstream from the translational start codon (AUG: Met), resulting in a different amino terminus of the mature glycoprotein (Ligtenberg et al., 1990) (Figure 1). There is unconfirmed evidence for another alternative spliced form that lacks the transmembrane domain (Williams et al., 1990). MUC1 has the features of an integral membrane protein, consisting of three distinct regions: (i) the amino terminal region containing a hydrophobic signal sequence of 13 amino acids long after the first seven aminoacids, and degenerate tandem repeats; (ii) the tandem repeat region consisting of well-conserved 20 amino acid repeats; and (iii) the carboxylterminal region containing degenerate tandem repeats, a transmembrane domain (hydrophobic) of 31 amino acids and a 69 amino acid cytoplasmic tail (Gendler et al., 1990b). The number of tandem repeats per molecule varies between approximately 21 and 125, the most common allelic variants being 41 and 85 (Ligtenberg et al., 1990). This domain forms the major part of the core protein, and results in a highly repetitive structure, which is extremely immunogenic (Figure 2). The sequence of the 20 amino acid tandem repeats unit corresponds to what might be expected for a protein, which is extensively O-glycosylated. Five serines and threonines, four of which are in doublets, are found in this repeat unit and these potential glycosylation sites are separated by regions rich in prolines, which would allow for up to one-fourth of the amino acids to be glycosylated. The molecular weight of


Figure 1. Organization of genes and mRNA encoding MUC1. Schematic representation of the fragment on which the MUC1 gene is located (top); filled boxes indicate exon sequences; in exon 2, repeats are indicated by the waved box. The general structure of MUC1 mRNA and protein product (middle), exons are indicated with numbered boxes; the hatched box represents the signal sequence, and the black filled box represents the transmembrane region. Magnification of the region involved in alternative splicing (bottom), part of the genomic fragment is indicated in the middle, part of the splice variant 1 containing the additional 27 bp (striped box) and part of variant 2 lacking this region, (based on data from Gendler et al., 1987; 1990; Swallow et al., 1987; Lancaster et al., 1990; Ligtenberg et al., 1990).

Figure 2. Diagram of MUC1 protein backbone (based on data from Gendler et al., 1990; Lan et al., 1990; Ligtenberg et al., 1990, 1992; Aplin et al., 1994)

the peptide backbone of an apparent 120 000–225 000 kDa is in accord with a protein of apparently 240 000–450 000 kDa containing ~50% of carbohydrate by weight (Gendler et al., 1990b). A proteolytic cleavage site has also been reported in the membrane proximal extracellular domain (Ligtenberg et al., 1992). The secreted form of MUC1 lacks the cytoplasmic tail (Boshell et al., 1992), but it is not clear whether this is the result of alternative splicing or proteolysis and release from the free extracellular domain (Aplin et al., 1994). There are several associations systems between proteins and the lipid membrane. Transmembrane proteins cross the plasma membrane once or several times like an α-helix structure which is 20 or 25 amino acids long. These hydrophobic structures interact with the lipid hydrophobic tails inside the lipid bilayer. Some of these proteins are covalently attached

to the fatty acids of the cytoplasmic monolayer (Magee et al., 1989), and this increases the interaction and stabilizes the binding of the protein to the membrane. The only three cysteine residues occurring in the translation product are present in the transmembrane region. It is postulated that a covalent bond is formed between a cystein sulphidric side-chain from MUC1 transmembrane region and the cytoplasmic monolayer fatty acids. If this is the case, the lipid bilayer anchorage would be reinforced, since the transmembrane domain of MUC1 is 31 amino acids long and this protein probably crosses the plasma membrane only once. Since MUC1 does not contain cysteine residues in the extracellular domain, it cannot form oligomers by disulphide bonding. This contrasts with the oligomerization of the gelforming mucins that are present in the secretions from specialized epithelial cells in many exocrine tissues, such as the salivary gland (Ligtenberg et al., 1990). The sequence similarity between mouse MUC1 and the human protein is only 34% in the tandem repeat domain, showing mainly conservation of serines and threonines (presumed sites of O-linked carbohydrate attachment). The similarity rises to 87% in the transmembrane and cytoplasmic domains, suggesting that these regions may be functionally important. Interestingly, the mouse homologue, unlike its human counterpart, does not exhibit a variable number of tandem repeats (Spicer et al., 1991). A series of studies was performed to monitor aspects of assembly, intracellular transport, and secretion of MUC1 and other mucins in polarized uterine epithelial cells, and a model based on these results of the metabolism of MUC1 was put forward: MUC1 is synthesized in the rough endoplasmic reticulum, passes through the smooth endoplasmic reticulum and the Golgi apparatus where it is glycosylated. Thereafter it is transferred to the cell surface. A fraction (30–50%) of uterine epithelial cell mucins appears to be released from the apical surface into the apical media or lumen. The remainder is presumably degraded intracellularly following endocytosis (Figure 3) (Pimental et al., 1996).

MUC1 detection in the endometrium Several antibodies that detect various MUC1 epitopes have been developed (Table I): HMFG-1, HMFG-2 and SM-3 are monoclonal antibodies [immunoglobulin (Ig)G] developed in mouse (Arklie et al., 1981; Taylor-Papadimitriou et al., 1981), the minimum amino acid sequence required to form a reactive epitope is for the first PDTR, for the second DTR and for the third PDTRP contained within the 20 amino acid tandem repeat region (Arklie et al., 1981; Taylor-Papadimitriou et al., 1981, 1991; Girling et al., 1989; Devine et al., 1992). BC1, BC2 and BC3 are mouse monoclonal antibodies (first and second IgG, third IgM) and have the same minimum peptide epitope within the tandem repeat region. Individual differences in the reactivity of the antibodies were noted, they recognize related but not identical epitopes, co-expressed on the same molecule (Xing et al., 1989, 1990). Binding of each antibody can be influenced by the level of glycosylation of the core protein; in particular, HMFG-1 binding to endometrial MUC1 is increased by pre-treatment with sialidase, presumably 1091


Figure 3. Model of MUC1 metabolism in uterine epithelial cells (based on data from Pimental et al., 1996). SER 5 smooth endoplasmic reticulum; RER 5 rough endoplasmic reticulum.

Table I. Antibodies against MUC1 epitopes used in endometrium Monoclonal antibody

Type

HMFG-1

Mouse IgG HMFG-2 Mouse IgG SM3 Mouse IgG BC-1 Mouse IgG BC-2 Mouse IgG BC-3 Mouse IgM NCRC-11 Mouse IgM CT-1 Polyclonal Rabbit IgG 5D4 D9B1

Mouse IgG Mouse IgM

Epitope

Localization

PDTR

VNTR of core

DTR

VNTR of core

PDTRP

VNTR of core

APDTR

VNTR of core

APDTR

VNTR of core

APDTR

VNTR of core

RPA(P)

VNTR of core

–

Cytoplasmic tail

Highly sulphated keratin chains Sialic acida

The last 17 C-terminal amino acids Associated with extracellular domain Associated with extracellular domain

Ig 5 immunoglobulin; VNTR 5 Variable number of tandem repeats aD9B1 epitope is a terminal non-reducing sialic acid residue on a keratan sulphate type glycan; it includes this sialic acid as well as adjacent sugar residues.

because of steric hindrance by short oligosaccharides of the NeuNAc-GalNAc (sialyl-Tn) type attached either to the threonine residue in the peptide epitope or at an adjacent serine or threonine (Devine et al., 1990; Hey et al., 1994). NCRC11 is a mouse monoclonal antibody (IgM); its defined antigen is expressed on the epitopes of HMFG-1 and HMFG-2 antibodies. In competitive binding inhibition assays, these antibodies partially inhibited the binding of NCRC-II antibody to antigen, suggesting that the epitopes involved are topographically closely associated (Price et al., 1985). 1092

CT-1 is a rabbit polyclonal antibody to the carboxyl-terminal 17 amino acids of the human MUC1 mucin. This antibody reacts with tissues from many mammalian species suggesting that the cytoplasmic portion of the molecule is well-conserved (Pemberton et al., 1992). 5D4 is a monoclonal antibody that binds to highly sulphated keratan chains (Mehmet et al., 1986). Its reactivity with high molecular weight glycoconjugates in the endometrium has been demonstrated by Western blotting (Hoadley et al., 1990). D9B1 is a monoclonal antibody that recognizes high molecular weight glycoprotein-associated sialylated oligosaccharide that is a secretory product of endometrial epithelium (Smith et al., 1989). The presence of MUC1 in the endometrium has been examined at both protein and mRNA levels. At protein level, there are semi-quantitative immunohistochemistry studies on endometrial tissue sections, endometrial explants and also invitro cultured uterine epithelial cells (Aplin et al., 1988, 1994; Campbell et al., 1988; Smith et al., 1989; Hoadley et al., 1990; Graham et al., 1991; Rye et al., 1993; Hey et al., 1994, 1995; Pimental et al., 1996). Also, quantitative studies with sodium dodecyl sulphate–polyacrylamide gel electrophoresis and Western blotting of secretory endometrial extract have been performed (Hey et al., 1994), including enzyme-linked immunosorbent assays (ELISA) in uterine flushings (Hey et al., 1995).

MUC1 regulation during the menstrual cycle: hormonal control The endometrium is a target tissue for the action of steroid hormones that stimulate differentiation of the tissue required for normal reproductive function (Psychoyos, 1973). MUC1 has been detected in the endometrium by several methods and its expression varies within the menstrual cycle; moreover these variations are not the same in all mammals (Table II). In mouse, MUC1 protein observed by immunofluorescence shows intense staining in the luminal and glandular epithelial cells of the uterus in pro-oestrous, oestrous and meta-oestrous metasections. Staining persists in glandular epithelia at the dioestrous stage, but is noticeably reduced in the luminal epithelia. The staining appears localized to the apical part of both the luminal and glandular epithelial cells (Surveyor et al., 1995). During early pregnancy, there is an intense staining in both epithelia, at day 3 of pregnancy staining is reduced in the luminal epithelium, and by day 4 staining is almost completely lost in both epithelia. This pattern persists until at least day 6 of pregnancy (Surveyor et al., 1995). Uterine MUC1 mRNA is highly abundant at oestrous and remains high until day 4, when relative MUC1 mRNA levels decline drastically (Surveyor et al., 1995). Oestrogen stimulates MUC1 mRNA expression and an antioestrogen antagonizes this effect, suggesting an involvement of the oestrogen receptor (Surveyor et al., 1995). In contrast, progesterone alone has no effect, but it antagonizes the stimulatory effect of oestrogen. In contrast with the uterus, MUC1 expression in cervical and vaginal epithelia does not appear to be significantly down-regulated during the periimplantation period (Surveyor et al., 1995).


Table II. Summary of MUC1 presence and expression in mammalian endometrium Study

Species

Presence in luminal epithelium

Presence in glandular epithelium

mRNA expression

Surveyor et al., 1995

Mouse

High during pro-oestrous, oestrous, meta-oestrous and early PG

High during oestrous cycle and early pregnancy

High at oestrous and at early pregnancy until day 4

Low at dioestrous and from day 3 of PG until at least day 6

Low from day 4 of pregnancy until at least day 6

Low at day 4 of pregnancy

High at oestrous and from day 8 of PG in inter-implantation sites.

High at oestrous and do not decrease at day 5 of pregnancy

Decreases 57% from oestrous to day 5 of pregnancy

DeSouza et al., 1997

Rat

Low from day 5 until day 8 of pregnancy Bowen et al., 1996

Pig

High at day 0 and 4 of oestrous cycle. Low on day 8 and undetectable levels from days 0 to 15 of the oestrous cycle or PG.

High at days 0 and 4 of oestrous cycle. Low on day 8 and undetectable levels from days 10 to 15 of the oestrous cycle or PG.

Hild-Petito et al., 1996

Baboon

Low during follicular phase and at days 10–12 post-ovulation

High during follicular phase and do not vary substantially throughout the luteal phase

Hewetson et al., 1997 Hoffman et al., 1998

Rabbit

Low at oestrous and at 7.25 day of PG in implantation sites. High at PSP, PG and at 7.25 days in nonimplantation regions.

Low at oestrous, PG and PSP.

Increases from oestrous levels six times at PSP and 30 times in preimplantation uteri. Reduced at implantation period in implantation sites

Several

Human

–

Low throughout proliferative phase. High during secretory phase. 2–3 days after the LH peak

Increased abundance from the proliferative to secretory stage. Decreases in late secretory phase.

High at 5–8 days post-ovulation

PG 5 pregnancy; PSP 5 pseudopregnancy.

The highest levels of MUC1 mRNA and protein coincide with the highest levels of oestrogen in plasma, and this occurs at the pro-oestrous and oestrous phases; however, in ovariectomized mice without hormonal replacement the endometrium still expresses MUC1 at very low levels (Surveyor et al., 1995). Expression of rat MUC1 protein appears to decrease in the luminal, but not glandular, uterine epithelium at the receptive phase (day 5 of pregnancy), nonetheless, rat luminal epithelium still expresses MUC1 in a few punctate foci during this phase (DeSouza et al., 1998). The pattern of MUC1 expression in porcine uterine epithelium is similar to that reported for rodents (Bowen et al., 1996). In mice and pigs, a higher loss of MUC1 expression occurs in luminal uterine epithelium, while in glandular epithelium it decreases only in pigs. However, in rat no dramatic decrease has been demonstrated in uterine MUC1 protein expression during the oestrous cycle. These observations indicate that uterine control of MUC1 expression is speciesspecific and is not really predictable from a phylogenetic standpoint (DeSouza et al., 1998). In rabbit, MUC1 mRNA expression has been investigated in endometrial cells from oestrous animals, those receiving progesterone for 5 days, and oestrous animals receiving progesterone for 5 days followed by oestradiol for 5 days. Interestingly, MUC1 is increased by progesterone and returned to oestrous levels by sequential treatment with oestrogen (Hewetson et al., 1997). mRNA is elevated 2–6-fold over oestrous levels in endometrium of pseudopregnant females and 30-fold in preimplantation stage uteri. This high level of MUC1 expression continues during the implantation period, but is drastically reduced in the

endometrium from implantation sites. At implantation, MUC1 protein in the uterus is markedly reduced or absent in the luminal epithelium in comparison with the blastocyst, as determined by CT-1 antibody. Short-term co-culture of uterine epithelial cells with trophoblastic vesicles derived from blastocyst also results in a local reduction in apical epithelial MUC1 protein (Hoffman et al., 1998). With the CT-1 antibody, it is possible that the additional reduction of MUC1 is due to localized proteolytic cleavage of the extracellular domain of this mucin, and removal of this segment from the apical cytoplasm after cleveage (Pimental et al., 1996). Proteolytic cleavage may be due to the trophoblast of implanting blastocysts, which is known to secrete several proteases (Denker, 1977). In the baboon (Papio anubis), elegant studies have been performed by Hild-Petito et al. (1996). Adult female baboons were studied during the menstrual cycle, and after ovariectomy with or without steroid treatment and in correlation to progesterone receptor expression in epithelial cells. MUC1 expression is very low in luminal and glandular epithelia in ovariectomized animals, but animals under conditions dominated by oestrogen influences, (i.e. those in the late follicular phase or those that are ovariectomized and given oestrogen treatment), display moderate reactivity in basal glandular epithelium but little or no reactivity in either the functionalis region or the surface epithelium. MUC1 expression in the glandular epithelia persists under the influence of progesterone and does not vary substantially throughout the luteal phase. Baboons receiving a combination of oestrogen and progesterone initially displayed strong surface expression of MUC1 (animals on days 5–8 1093


post-ovulation or ovariectomized, oestrogen-primed animals given 7 days of combined hormone treatment). However, there is a decrease in its expression in surface epithelium at days 10–12 post-ovulation or after 14 days of combined oestrogen and progesterone treatment in ovariectomized oestrogenprimed baboons. Using anti-oestrogens, the authors conclude that oestrogen is not required for maintenance of MUC1 expression in either luminal or glandular epithelium. Progesterone receptor is expressed in both surfaces on day 5 in a manner consistent with strong MUC1 expression in both compartments, and by day 12 progestin receptors are depleted from surface epithelium. Treatment with an antiprogestin inhibits the downregulation of both MUC1 and progesterone receptors in the surface epithelium (Hild-Petito et al., 1996). In humans, MUC1 regulation contrasts with that reported in mouse. Immunohistochemistry using HMFG-1, HMFG-2 and BC3 in the proliferative phase of the cycle, stain endometrial glands, but they do not stain with antibody SM3. Diffuse staining is seen throughout the cytoplasm of certain cells, while strong and often continuous staining of the apical surface is observed. The secretory phase of the cycle is characterized by detection of SM3 epitope and increased expression of the epitopes recognized by HMFG1, HMFG2 and BC3 (Hey et al., 1994). On day LH12, apical cell surface staining is predominant, with low levels in the gland secretions. A proportion of glands remains unstained with HMFG-1, whereas BC3 detects immunoreactivity. By day LH 1 3, prominent basal intracellular immunoreactivity in glands is observed with both monoclonal antibodies, and continues on days LH14, LH15 and LH16, after which staining becomes predominantly confined to intraluminal secretions. Monoclonal antibody HMFG1 produces consistently lower reactivity than BC3 (Hey et al., 1995). With biopsies taken on day LH17 (the approximate time of blastocyst implantation), the distribution of MUC1 core protein was studied using two antibodies, BC3 and HMFG1. In each case, a significant fraction of the glands (77 and 47% respectively) principally exhibits immunoreactivity in intraluminal secretions. Other glands are stained intracellularly or unstained (Serle et al., 1994). Endometrial secretions undergo steroid hormone-induced alterations in sialylation (Seif et al., 1989; Smith et al., 1989; Hoadley et al., 1990; Aplin, 1991). After using sialidase (to reveal masked epitopes), the proportion of glands with strongly stained intraluminal secretions increases from 47 to 57% for HMFG1, indicating that MUC1 is sialylated in the mid-luteal phase (Hey et al., 1994; Serle et al., 1994). This increase in secreted immunoreactivity is mainly due to a decrease in the proportion of HMFG1-negative glands exhibiting only apical surface staining. Nonetheless, some glands are BC3-negative and remain HMFG1-immunonegative after sialidase exposure, indicating that little MUC1 is being expressed. This suggests that there is intrinsic heterogeneity in mucin expression in glands in normal endometrium (Smith et al., 1989; Hoadley et al., 1990; Hey et al., 1994). Immunohistochemical techniques with HMFG1 and NCRC11 antibodies show a contrary pattern of expression during the proliferative phase. HMFG-1 shows up strongly 1094

reactive, localized at the apical surface of glands, while with NCRC11 no staining is observed. In the secretory phase, HMFG1 reduces its immunoreactivity until it disappears in the mid-secretory phase (days 19–23), whereas NCRC11 increases its staining, being intense in the mid-phase, lowering in the late secretory phase, the minimum staining being on day 28 in most of the glands. In this case the pattern of expression of the detected epitope by HMFG1 is altered with respect to the results presented by other authors, even though the altered binding patterns of antibodies and their epitopes in human endometrium during the menstrual cycle are not novel phenomena, and indeed are not surprising considering the nature and function of this tissue ( Rye et al., 1993). Evidence exists that tissue-specific glycosylation occurs in MUC1, for example, the pancreatic form of the molecule is considerably larger than the breast or endometrial product, despite having an essentially identical transcript (Lan et al., 1990). It has been demonstrated that endometrial MUC1 carries highly sulphated lactosaminoglycan chains and sialokeratan sulphate epitopes recognized by 5D4 and D9B1 antibodies respectively (Aplin et al., 1998). Studies on the behaviour of the D9B1 epitope during the cycle indicate that it is absent in the normal proliferative phase, and is produced and secreted in epithelial cells in the secretory phase, achieving maximal levels in secretions 6–7 days after the LH peak (Smith et al., 1989). Production of the epitope is shown to depend on progesterone (Graham et al., 1991). Electron microscopic immunolocalization studies using antibody D9B1 have demonstrated the presence of this epitope in the glycocalyx at the apical cell surface of gland cells, including microvilli (Aplin et al., 1994). Immunohistochemistry, using antibody 5D4 to keratan sulphate, indicates that this not only is a cell surface-associated and secretory product of endometrium, but it also shows a hormonal pattern of regulation that closely resembles that of the D9B1 epitope (Aplin et al., 1988). Thus keratan sulphate is detectable in association with a fraction of the gland cells in the proliferative phase, but with a very significant increase beginning 2–3 days after the LH peak. From LH13 to LH15 days, intracellular immunoreactivity is preponderant in the gland cells. This appears at the cell base but is subsequently mobilized through the apical cytoplasm and into gland secretions. Maximal immunoreactivity in the secretions occurs in the period from LH16 to LH18 days, coinciding with the expected time of implantation (Aplin et al., 1994). Interestingly 5D4 epitope is abundant at the luminal epithelial surface until the implantation phase, when it disappears first from groups of cells and thereafter altogether (Aplin et al., 1998). Endometrial epithelium also expresses Sialyl-Lewis X (SLeX) and SialylLewis A (SLeA), with a distribution and pattern during menstrual cycle similar to that reported for MUC1. It has been demonstrated using Western blotting and double determinant ELISA of uterine flushings that SLeX is associated with MUC1 core protein, as well as SLeA in two endometrial carcinoma cell lines (HEC1A and HEC1B) (Hey and Aplin, 1996). The concentrations of secretory MUC1, measured by ELISA in timed uterine flushings obtained from normal fertile subjects at various days during the luteal phase indicate that MUC1 is


present at lower concentrations in luteal phase flushings performed before LH17 than after LH17. Both luminal and glandular epithelia are immunopositive for MUC1 in the secretory phase, but it is probable that the major source is the glandular compartment. Although is not possible to draw firm conclusions from static measurements on the kinetics of escape of the secretory components from glands, it is to be noted that an increase in immunostaining of glandular lumen on days LH13, LH14 is followed by an increase in immunoreactivity in flushings on day LH17. Thus, on day LH 17 (when implantation of blastocysts would be expected in a conception cycle), MUC1 seems to be consistently increased and becomes detectable in the uterine lumen. It is also interesting that repeating the flushing procedure every 2 days in the period after day LH17 renders high MUC1 concentrations in all the samples, suggesting that the diffusion from the glands at this stage of the cycle is fast enough to restore luminal composition within 2 days. In women suffering from recurrent spontaneous miscarriage, the concentration of MUC1 was significantly lower than in the controls on day LH110 (Hey et al., 1995). Labelling studies with [35S]-sulphate in endometrial explants of secretory endometrium demonstrates that endogenous synthesis of MUC1 occurs in the tissue and that the product is sulphated (Aplin et al., 1994). In primary endometrial epithelial cell culture (polarized cells, cultured on a matrigel-coated glass coverslip), intracellular heterogeneity is evident from both polypeptide and glycan expression (Campbell et al., 1988). MUC1 has been demonstrated at the cell surface as well as intracellularly. Secreted MUC1 has been detected in conditioned medium from epithelial cell monolayers (Aplin et al., 1994).

Discussion and future prospects The anti-adhesion properties of MUC1 are probably the result of the unique structure of the molecule. As we have discussed before, epsialin towers 200–500 nm above the plasma membrane, whereas most proteins at the cell surface remain inside the boundaries of the glycocalyx, which is ~10 nm thick (Figure 4). It may be postulated that a high density of MUC1 at the cell surface prevents cellular adhesion by masking adhesion molecules, and preventing interactions between glycoproteins on opposite membranes (Hilkens et al., 1992), thus reducing cell–cell interactions as well as integrin mediated extracellular matrix adhesion. The abundance of sialic acid residues contributes to the anti-adhesion properties of MUC1, since these residues will make the glycan proteins more bulky and thus may contribute to the rigidity of the molecule. Alternatively, many sialic acid and sulphate residues would give the glycoprotein a strong negative charge, also contributing towards the anti-adhesion effect by charge repulsion (Coulombic repulsion) of the negatively charged trophectodermal surface (Hilkens et al., 1992; Aplin et al., 1994). Subsequently, MUC1 may inhibit the interaction between embryo and maternal apical epithelium adhesion molecules during implantation, because of the possibility to create a uterine barrier for implantation

Figure 4. Proposed structure of a mature MUC1 molecule. It extends above the glycocalyx and varies depending on number of tandem repeats. and indicate O and N-linked carbohydrates respectively. The extracellular domain of MUC1 is predicted to extend 200–500 nm above the cell membrane. For comparison the thickness of the glycocalyx is indicated (based on from Jenfot et al., 1990; Hilkens et al., 1992; Aplin et al., 1994).

(Figure 5) (Aplin et al., 1996). In mouse and rat uterine epithelium, the homologue MUC1 is hormonally regulated, with reduced expression in the implantation period (Braga et al., 1996; De Souza et al., 1998). Thus, MUC1 may play a role in defining the ‘receptive window’, although it does not seems to be responsible for subsequent loss of receptivity, since it does no reappear in the epithelial compartment after day 5 (Aplin, 1997). Nonetheless, reduced mucin expression alone is not sufficient to generate a receptive uterine state; other events such as blastocyst activation (Paria et al., 1993) and/or induction of attachmentpromoting molecules at the uterine epithelial cell surface (Surveyor et al., 1995), play a role in receptivity. In humans, MUC1 expression is high during the implantation window. This represents a contradiction if MUC1 is indeed an anti-adhesion molecule. The intrinsic heterogeneity present in this molecule may allow a local mechanism to define a receptive site within the endometrium. It is also possible that MUC1-associated glycans could be recognized by the embryo. There are several candidate proteins that bind carbohydrates, e.g. selectins, cell surface glycosyltransferases or lectins. This close relationship has never been demonstrated. Nevertheless, lectins present on the trophectoderm could also be candidates for the mediation of the cell–cell interactions involving a carbohydrate ligand present on the surface of endometrial epithelial cells as has been reported in mice (Poirier and Robertson, 1993). It is becoming increasingly apparent that adhesion molecules involved in cellular adhesion to other cells and to the extracellular matrix are crucial to the adhesion of the blastocyst to the maternal endometrium. These first interactions could be followed by integrin- or cadherin-mediated adhesion in a cascade (Aplin, 1997). Evidence for an integrin or cadherin-mediated cascade is the presence of integrins at uterine epithelial cells (Lessey et al., 1992; Tabibzadeh, 1992; Klentzeris et al., 1993), and also in the 1095


Figure 5. Model of MUC1 function as an anti-adhesion molecule during embryo attachment. The highly extended structure of MUC1 prevents the embryo from attaching to the apical surface of uterine epithelial cells (UEC). Loss of MUC1 permits access to apical cell surface receptors on the uterine epithelium that supports embryo attachment during the receptive phase (based on data reported from Hilkens et al., 1992; Aplin et al., 1994, 1996; Surveyor et al., 1995).

embryo development (Fusi et al., 1992; Turpeenniemi-Hujanen et al., 1992; Campbell et al., 1995), and further the embryonic regulation of endometrial integrins has been reported (Simoń et al., 1997). Cadherin expression has also been reported in the endometrial epithelium (Tabibzadeh et al., 1995; MacCalman et al., 1996) and in the trophoblast (Coutifaris et al., 1991; MacCalman et al., 1996). This adhesion function is not necessarily inconsistent with an inhibitory role, since there is high microvariability of glycosylation, suggesting that perhaps specific areas of the endometrium could be recognition structures, while other regions would be inhibitory (Aplin, 1997). In the baboon, MUC1 is also expressed at relatively high levels in the implantation phase, a species in which the anatomy of implantation is relatively similar to human (Hild-Petito et al., 1997). It is interesting that implantation is much more efficient in the mouse compared to humans; the human uterus might impose a barrier to the implanting embryo. One reason for this conjecture might be the relatively high proportion of abnormal embryos observed in human (Hustin et al., 1990). The pregnancy process implies expenditure of time and energy and should not be wasted in conceptions that are destined to fail. Implantation could thus be a natural selection process wherein only good quality embryos succeed (Aplin et al., 1996). In baboons, there are interesting differences depending upon the level of the social hierarchy. High status females appear to conceive more easily, but this is related to increased abortion rates (Packer et al., 1995). In humans, women with recurrent spontaneous miscarriges show reduced levels of secreted MUC1 (Hey et al., 1995). Such data might imply that this selection process is deficient, permitting embryos to implant that are not competent to reach term. An important consequence of these events is that an abnormal maternal environment leads to the survival of embryos with intrinsic abnormalities (Aplin, 1996). Keratan sulphate has been reported to disappear during the implantation period (Aplin et al., 1998), and this might 1096

imply a global loss of MUC1 volume, and subsequently a decrease in its rigidity that could permit embryo adhesion. All these possibilities augur well for future prospects, because depending on the function of MUC1 on the endometrial surface, its presence or absence will be important in embryo implantation. If we could control the levels of MUC1 expression, not only in surface epithelium but also in secretions, we might favour the rates of pregnancy in patients undergoing assisted reproduction techniques by decreasing its levels or, perhaps, in women suffering from recurrent miscarriage, increasing this expression could help prevent abortion. We need to consider the benefits of increasing the rates of implantation with regard to the disadvantages of increased miscarriage. On the other hand, MUC1 may be regulated by steroid-independent mechanisms, making its regulation difficult. The first report of endometrial MUC1 down-regulation by the embryo was reported in rabbits (Hoffman et al., 1998), a species in which hormonal regulation is similar to that in humans. This could be the regulation mechanism in human embryos, it is possible that healthy embryos have the ability to decrease MUC1 at the implantation site. If MUC1 does not inhibit the adhesion process and functions in the initial attachment of the embryo to the apical surface, the embryo could then locally up-regulate MUC1.

References Aplin, J.D. (1996) The cell biology of human implantation. Placenta, 17, 269–275. Aplin, J.D. (1991) Glycans as biochemical markers of human endometrial secretory differentiation. J. Reprod. Fertil., 91, 525–541. Aplin, J.D. (1997) Adhesion molecules in implantation. J. Reprod. Fertil., 2, 84–93. Aplin, J.D., Hoadley, M.E. and Seif, M.W. (1988) Hormonally regulated secretion of keratan sulphate by human endometrial epithelium. Biochem. Soc. Trans., 17, 136–137.

MUC1 and endometrial receptivity Aplin, J.D., Seif, M.W., Graham, R.A. et al. (1994) The endometrial cell surface and implantation. Expression of the polymorphic mucin MUC-1 and adhesion molecules during the endometrial cycle. Ann. N.Y. Acad. Sci., 734, 103–121. Aplin, J.D., Spanswick, C., Behzad, F. et al. (1996) Integrins b5, b3 and α5 in human and mouse endometrium: expression in stromal and glandular cells. Mol. Hum. Reprod., 2, 527–534. Aplin, J.D., Hey, N.A. and Graham, R.A. (1998) Human endometrial MUC1 carries keratan sulphate: characteristic glycoforms in the luminal epithelium at receptivity. Glycobiology, 8, 269–276. Arklie, J., Taylor-Papadimitriou, J., Bodmer, W. et al. (1981) Differentiation antigens expressed by epithelial cells in the lactating breast are also detectable in breast cancers. Int. J. Cancer, 28, 23–29. Bergh, P.A. and Navot, D. (1992) The impact of embryonic development and endometrial on the timing of implantation. Fertil. Steril., 58, 537–542. Boshell, M., Lalani, E.N., Pemberton, L. et al. (1992) The product of the human MUC1 gene when secreted by mouse cells transfected with the full length cDNA lacks the cytoplasmic tail. Biochem. Biophys. Research. Comm., 185, 1–8. Bowen, J.A., Bazer, F.W., Burghardt, R.C. et al. (1996) Spatial and temporal analyses of integrin and MUC1 expression in porcine uterine epithelium and trophoectoderm in vivo. Biol. Reprod., 55, 1098–1106. Braga, V.M.M. and Gendler, S.J. (1993) Modulation of MUC1 mucin expression in the mouse uterus during the estrous cycle early pregnancy and placentation. J. Cell Sci., 105, 397–405. Bulletti, C. and Gurpide, E. (eds) (1991) The primate endometrium. Ann. N.Y. Acad. Sci., 6–27. Campbell, S., Seif, M.W., Aplin, J.D. et al. (1988) Expression of a secretory product by microvillious and ciliated cells of the human endometrial epithelium in vivo and in vitro. Hum. Reprod., 3, 927–934. Campbell, S., Swann, H.R., Seif, M.W. et al. (1995) Cell adhesion molecules on the oocyte and preimplantation human embryo. Hum. Reprod., 10, 1571–1578. Coutifaris, C., Kao, L.C., Sehdev, H.M. et al. (1991) E-Cadherin expression during the differentiation of human trophoblast. Development, 113, 767–777. Denker, H.W. (1977) Implantation: the role of proteinases and blaockage of implantation by proteinase inhibitors. Adv. Anat. Embryol. Cell. Biol., 53, 1–123. DeSouza, M.M., Mani, S.K., Julian, J. et al. (1998) Reduction of MUC1 expression during the receptive phase in the rat uterus. Biol. Reprod., 58, 1503–1507. Devine, P. and Mackenzie, I.F.C. (1992) Mucins: structure, function and association with malignancy. Bioassays, 14, 619–625. Devine, P.L., Warren, J.A., Ward, B.J. et al. (1990) glycosylation and exposure of tumour-associated epitopes on mucins. J. Tumour Marker Oncol., 5, 11–26. Eckert, R., Randall, D. and Augustine, G. (eds) (1990) Fisiologıá Animal: Mecanismos y Adaptaciones. McGraw-Hill-Interamericana de Espanã, Madrid, Spain, pp. 319–321. Ferenczy, A. and Gurlnick, M. (1983) Endometrial microstructure: structure function relationships throughout the menstrual cycle. Semin. Reprod. Endocrinol., 1, 205. Gendler, S.J. and Spicer, A.P. (1995) Epithelial mucin genes. Ann. Rev. Physiol., 57, 607–634. Gendler, S.J., Burchell, J.M., Duhig, T. et al. (1987) Proc. Natl Acad. Sci. USA, 84, 6060–6064. Gendler, S.J., Cohen, E.P., Craston, A. et al. (1990a) The locus of the polymorphic epithelial mucin (PEM) tumour antigen on chromosome 1q21 shows a high frequency of alteration in primary human breast tumours. Int. J. Cancer, 45, 431–435. Gendler, S.J., Lancaster, C.A., Taylor-Papadimitriou, J. et al. (1990b) Molecular cloning and expression of human tumour associated polymorphic epithelial mucin. J. Biol. Chem., 265, 15286–15293. Girling, A., Bartkova, J., Burchell, J. et al. (1989) A core protein epitope of the polymorphic epithelial mucin detected by the monoclonal antibody SM3 is selectively exposed in a range of primary carcinomas. Int. J. Cancer, 43, 1072–1076. Giudice, L.C. (1994) Growth factors and growth modulators in human uterine endometrium: their potential relevance to reproductive medicine. Fertil. Steril., 61, 1–17. Giudice, L.C. and Ferenczy, A. (1996) The endometrial cycle. In Adashi, E.Y., Rock, J.A. and Rosenwaks, Z. (eds) Reproductive Endocrinology, Surgery, and Technology. Vol. 1. Lippincott-Raven Publishers, Philadelphia/ New York, USA, pp. 272–300.

Graham, R.A., Li, T.C., Seif, M.W. et al. (1991) The effects of the antiprogestin RU486 (Mifepristone) on an endometrial secretory glycan; an immunocytochemical study. Fertil. Steril., 55, 1132–1136. Fusi, F.M., Vignalli, M., Busacca, M. et al. (1992) Evidence of the presence of an integrin on the oolema of unfertilised human oocytes. Mol. Reprod. Dev., 31, 215–222. Hewetson, A. and Chilton, B.S. (1997) Molecular cloning and hormonedependent expression of rabbit MUC1 in the cervix and uterus. Biol. Reprod., 57, 468–477. Hey, N.A., Graham, R.A., Seif, M.W. et al. (1994) The polymorphic epithelial mucin MUC1 is regulated with maximal expression in the implantation phase. J. Clin. Endocrinol. Metab., 78, 337–342. Hey, N.A., Li, T.C., Devine, P.L. et al. (1995) MUC1 in secretory phase endometrium: expression in precisely dated biopsies and flushings from normal and recurrent miscarriage patients. Hum. Reprod., 10, 2655–2662. Hey, N.A. and Aplin, J.D. (1996) Siayl-Lewisx and Siayl-Lewisa are associated with MUC1 in human endometrium. Glycoconj. J., 13, 769–779. Hild-Petito, S., Fazleabas, A.T., Julian, J. et al. (1996) Mucin (MUC1) expression is differentially regulated in uterine luminal and glandular epithelia of the baboon (Papio anubis). Biol. Reprod., 54, 939–947. Hilkens, J., Ligtenberg, M.J.L., Vos, H.L. and Litvinov, S.V. (1992) Cell membrane-associated mucins and their adhesion modulating property. Trends Biochem. Sci., 17, 359–363. Hoadley, M.E., Seif, M.W. and Aplin, J.D. (1990) Menstrual cycle-dependent expression of keratan sulphate in human endometrium. Biochem. J., 266, 757–763. Hoffman, L.H., Olson, G.E., Carson, D.D. et al. (1998) Progesterone and implanting blastocyst regulate MUC1 expression in rabbit uterine epithelium. Endocrinology, 139, 266–271. Hustin, J., Jauniaux, E. and Schaaps, J.P. (1990) Histological study of the materno–embryonic interface in the spontaneous abortion. Placenta, 11, 477–486. Iizuka,R. and Semm,K. (eds) (1988) Human Reproduction. Current Status/ Future Prospect. Excepta Media (International Congress), Amsterdam-New York-Oxford, pp. 231–232. Jentoft, N. (1990) Why are proteins O-glycosylated? Trends. Biochem. Sci., 15, 291–294. Klentzeris, L.D., Bulmer, J.N., Tredosiewicz, L.K. et al. (1993) b1 integrin cell adhesion molecules in the endometrium of fertile and unfertile women. Hum. Reprod., 8, 1223–1230. Knobil, E. and Neill, J.D. (eds) (1994) The Physiology of Reproduction. 2nd edn. Raven Press Ltd, New York, USA, pp. 541–569. Lan, M.S., Batra, S.K., Qi, W.N. et al. (1990) Cloning and sequencing of the human pancreatic tumour mucin cDNA. J. Biol. Chem., 265, 15286–15293. Lancaster, C.A., Peat, N., Duhig, T. et al. (1990) Structure and expression of the polymorphic epithelial gene: an expressed VNTR unit. Biochem. Biophys. Res. Com., 173, 1019–1029. Lessey, B.A., Damjanovich, L., Coutifaris, C. et al. (1992) Integrin adhesion molecules in the human endometrium. Correlation with the normal and abnormal menstrual cycle. J. Clin. Invest., 90, 188–195. Ligtenberg, M.J.L., Kruishaar, L., Buijs, F. et al. (1992) Cell associated epsialin is a complex containing two proteins derived for a common precursor. J. Biol. Chem., 267, 6171–6177. Ligtenberg, M.J.L., Vos, H.L., Gennissen, A.M.C. et al. (1990) Epsialin, a carcinoma-associated mucin, is generated by a polymorphic gene encoding splice variants with alternative amino termini. J. Biol. Chem., 265, 5573– 5578. MacCalman, C.D., Furth, E.E., Omigbodun, A. et al. (1996) Regulated expression of cadherin-11 in human epithelial cells: a role for cadherin-11 in trophoblast endometrium interactions? Dev. Dyn., 206, 201–211. Magge, A.I., Gutierrez, L., Marshall, C.J. et al. (1989) Targeting of oncoproteins to membranes by fatty acylation. J. Cell. Sci., 11, (Suppl.), 149–160. Mehmet, H., Scudder, P., Tang, P.W. et al. (1986) The antigenic determinants recognised by three monoclonal antibodies to keratan sulphate involve sulphated heptaor larger oligosaccharides of the poly (N-acetyllactosamine) series. Eur. J. Biochem., 157, 385–391. Packer, C., Collins, D.A., Sindimwo, A. et al. (1995) Reproductive constraints on aggressive competition by female baboons. Nature, 373, 60–63. Paria, B.C., Das, S.K., Andrews, G.K. et al. (1993) Expression of the epidermal growth factor receptor gene is regulated in mouse blastocysts during delayed implantation. Proc. Natl. Acad. Sci. USA, 90, 55–59. Pemberton, L., Taylor-Papadimitriou, J. and Gendler, S.J. (1992) Antibodies to the cytoplasmic domain of the MUC1 mucin show conservation throughout mammals. Biochem. Biophys. Res. Commun., 185, 167–175.

1097

M.Meseguer, A.Pellicer and C.Simoń Pimental, R.A., Julian, J., Gendler, S.J. et al. (1996) Synthesis and intracellular trafficking of MUC1 and mucins by polarised mouse uterine cells. J. Biol. Chem., 271, 28128–28137. Poirier, F. and Roberston, E.J. (1993) Normal developments of mice carrying a null mutation in the gene encoding the L14 S-type lectin. Development, 119, 1229–1236. Price, M.R., Edwards, S., Owainati, A. et al. (1985) Multiples epitopes on a human breast-carcinoma-associated antigen. Int. J. Cancer, 36, 567–574. Psychoyos, A. (1973) Hormonal control of ovoimplantation. Vitamin Horm., 31, 201–256. Rye, P.D., Bell, S.C. and Walker, R.A. (1993) Immunohistochemical expression of tumour-associated glycoprotein and polymorphic epithelial mucin in the human endometrium during the menstrual cycle. J. Reprod. Fertil., 97, 551–556. Seif, M.W., Aplin, J.D., Foden, L.J. et al. (1989) A novel approach for monitoring the endometrial cycle and detecting ovulation. Am. J. Obstet. Gynecol., 160, 357–362. Seppa¨la¨, M., Julkunen, M., Riitinen, L. and Koistinen, R. (1991) Endometrial proteins: a reappraisal. Hum. Reprod., 7, 31. Serle, E., Aplin, J.D., Li, T.C. et al. (1994) Endometrial differentiation in the peri-implantation phase of women with recurrent miscarriage: a morphological and immunohistochemical study. Fertil. Steril., 62, 989–996. Sheehan, J.K. and Carlstedt, I. (1990) Electronic microscopy of cervical mucus glycoproteins and fragments therefrom. Biochem. J., 265, 169–178. Sheehan, J.K., Boot-Handford, R.P., Chantler, E. et al. (1991) Evidence for shared epitopes within the naked protein domains of human mocus glycoproteins. Biochem. J., 274, 293–296. Simoń, C., Gimeno, M.J., Mercader, A. et al. (1996) Cytokines adhesion molecules-invasive proteinases. The missing paracrine/autocrine link in embryonic implantation? Mol. Hum. Reprod., 6, 405–424. Simoń, C., Gimeno, M.J., Mercader, A. et al. (1997) Embryonic regulation of b3, α4 and α1 in human endometrial epithelial cells in vitro. J. Clin. Endocrinol. Metab., 82, 2607–2616. Smith, R.A., Seif, M.W., Rogers, A.W. et al. (1989) The endometrial cycle: the expression of a secretory component correlated with the luteinizing hormone peak. Hum. Reprod., 4, 236–242. Spicer, A.J., Parry, G., Patton, S. and Gendler, S.J. (1991) Molecular cloning and analysis of the mouse homologue of the tumour-associated mucin, MUC1, reveals conservation of potential o-glycosylation sites, transmembrane, and cytoplasmic domains and a loss of minisatellite-like polymorphism. J. Biol. Chem., 266, 15099–15109. Straus, G.J. and Deckker, J. (1992) Mucin-type glycoproteins. Crit. Rev. Biochem. Mol. Biol., 27, 57–92. Surveyor, G.A., Gendler, S.J., Pemberton, L. et al. (1995) Expression and steroid hormonal control of MUC1 in the mouse uterus. Endocrinology, 136, 3639–3647. Swallow, D.M., Gendler, S., Griffiths, B. et al. (1987) The human tumour associated epithelial mucin are coded by an expressed hypervariable gene locus PUM. Nature, 328, 82–84. Tabibzadeh, S. (1992) Patterns of expression of integrin molecules in human endometrium during the menstrual cycle. Hum. Reprod., 7, 876–882. Tabibzadeh, S., Babaknia, A., Kong, Q,F. et al. (1995) Menstruation is associated with disordered expression of desmoplakin I/II cadherins and catenins and conversion of F-actin to G-actin patterns in endometrial epithelium. Hum. Reprod., 10, 776–784. Taylor-Papadimitriou, J. (1991) Report on the first international workshop on carcinoma-associated mucins. Int. J. Cancer, 49, 1–5. Taylor-Papadimitriou, J., Peterson, J.A., Arklie, J. et al. (1981) Monoclonal antibodies to epithelium specific components of the human milk fat globule membrane: production and reaction with cells in culture. Int. J. Cancer, 28, 17–21. Turpeenniemi-Hujanen, T., Roonberg, L., Kauppila, A. et al. (1992) Laminin in the human embryo implantation: analogy to the invasion by malignant cells. Fertil. Steril., 58, 105–113. Van Kooig, R.J., Roelofs, H.J.M., Kathan, G.A.M. and Kramer, M.J. (1982) Synthesis of a mucous glycoprotein in the human uterus. Eur. J. Obstet. Gynecol. Reprod. Biol., 14, 191–197. Von Heijne, G. (1986) A new method for predicting signal sequence cleacage site. Nucleic Acids Res., 14, 4683–4690. Wesseling, J., van der Valk, S.W., Vos, H.L. et al. (1995) Epsialin (MUC1) overexpression inhibits integrin-mediated cell adhesion to extracellular matrix components. J. Cell Biol., 129, 255–265. Williams, C.J., Wreschner, D.H., Tanaka, A. et al. (1990) Multiple protein forms of the human associated epithelial membrane antigen are generated

1098

by alternative splicing and induced by hormonal stimulation. Biochem. Biophys. Research. Comm., 179, 1331–1338. Wolf, R.L., Stouffer, R.L. and Brenner, R.M. (eds) (1993) In Vitro Fertilization and Embryo Transfer in Primates. Springer-Verlag, New York, USA, pp. 145–157. Xing, P.-X., Reynolds, K., Tjandra, J.J. et al. (1990) Synthetic peptides reactive with anti-human milk fat globule membrane monoclonal antibodies. Cancer Res., 50, 89–96. Xing, P.-X., Tjandra, J.J., Stacker, S.A. et al. (1989) Monoclonal antibodies reactive with mucin expressed in breast cancer. Immunol. Cell. Biol., 67, 183–185. Yen, S.S.C. and Jaffe, R.B. (eds) (1991) Reproductive Endocrinology. W.B.Saunders, Philadelphia, USA, pp. 309–356. Received on April 15, 1998; accepted on August 28, 1998