Human Reproduction vol.7 suppl.l pp.31-38, 1992
Endometrial proteins: A reappraisal
Markku Seppala, Mervi Julkunen, Leena Riittinen and Riitta Koisrinen Department I of Obstetrics and Gynaecology, Helsinki University Central Hospital SF-OO29O Helsinki, Finland
Key Words: endometrion, protein, PPM, IGF
Introduction
The human uterus has been considered as an end organ which does not send feed-back signals to other parts of the body. Recent evidence indicates that this may not be entirely the case. The uterus responds to various endocrine stimuli by secreting local mediator substances and proteins which can also be found in peripheral blood. A physiological cause of infertility may lie in uterine responsiveness. The hormonal stimulus may be inadequate, or the uterine responsiveness not synchronized with embryonic development, even in the presence of apparently adequate levels of ovarian steroid hormones. Therefore, understanding uterine physiology is the key in any attempt at improving fertility since adequate embryo-uterine interactions are essential. Primary dysfunction of the endometrium may be associated with unexplained infertility, even when progesterone secretion is normal. This was indicated in a recent study by Graham et al. (1990), in which immunohistochemistry with monoclonal antibody D9B1 was used to assess the production and secretion of an oligosaccharide epitope produced by endometrial glands between two and seven days after the luteinizing hormone surge. In women with infertility, the © Oxford University Press
Endometrial cell types
The endometrium contains epithelial cells, stromal cells, endothelial cells of blood vessels, transiently resident cells, the extracellular matrix and connective tissue. In the proliferative phase, both the glands and the stroma contain oestrogen and progesterone receptors. After ovulation, the glandular receptors disappear under the influence of progesterone, while stromal receptors are maintained (Bouchard et al., 1991). Endometrial biopsies obtained during oestrogen replacement for the establishment of a receptive endometrium in women with premature menopause have revealed the expression of an oestrogen receptor-related antigen in the cytoplasm of the epithelium. After three days of progesterone replacement, the endometrium showed normal secretory features and expression of the oestrogen receptor-related antigen now appeared in the stroma as well as in the glands (Critchley et al., 1990a). Stromal oedema is maximal from days LH + 6 to LH + 8, which is believed to correspond to the beginning of the so called "implantation window" (Johannisson, 1991). In the secretory phase, the actions of oestrogen and progesterone on epithelial cells are likely to be mediated in a paracrine fashion by factors secreted from the neighbouring stromal cells.
Growth factors
The endometrium produces several growth factors such as epidermal growth factor (EGF), transforming growth factors (TGFa, TGFpi), platelet-derived growth factor (PDGF) and insulin-like growth factor-I (IGF-I). EGF is a product of the uterine epithelial cells. There are EGF-binding sites in the human uterus (Hoffmann et al., 1984), both in the glands and in the stroma (Bergchuck et al., 1989; Smith et al., 1991). Uterine EGF secretion is induced by oestrogen (Paria and Dey, 1990) and EGF is involved in mediating 31
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
Uterine factors influence reproduction at the macro-anatomy level, and the effects of hormonal steroids on endometrial morphology are well recognized in the histopathological diagnosis of dysfunctional bleeding and infertility. During the past decade, attention has been paid to endometrial protein synthesis and secretion with respect to endocrine stimuli and implantation, and to the paracrine/autocrine effects of endometrial peptide growth factors, their binding proteins and other factors. The emphasis of this presentation is on protein secretion of the secretory endometrium, in which progesterone plays a pivotal role. Insulin-like growth factors have receptors on the endometrium, and IGF-binding proteins, stimulated by progesterone, modulate the effects of IGFs locally. Also other protein products of the secretory endometrium have been reviewed in this communication, with special emphasis on studies of a progesterone-associated endometrial protein which has many names in the literature, such as PEP, PP14, o2-PEG and AUP. Extensive studies are ongoing in many laboratories to elucidate the regulation, function, interplay at tissue and cellular levels, and clinical significance of these proteins.
synthesis and secretion of this epitope was significantly reduced and delayed in the presence of normal serum progesterone levels. Drugs used to improve pregnancy rates in patients undergoing IVF may have adverse effects on the endometrium. In their recent work, Paulson and coworkers (1990) observed that a woman whose uterus was not subject to the effects of fertility drugs had an implantation rate three times higher than women whose uterine environment had been influenced by drugs given to improve fertility. This presentation will address endometrial protein secretion during the peri-implantation period. Of particular interest are paracrine and autocrine mechanisms, including mediators of hormone action, such as growth factors, their receptors and binding proteins. The emphasis of this presentation is on two major proteins of secretory endometrium, namely insulin-like growth factor binding protein-1 and placental protein 14.
M.Seppilfl at al
IGF-binding proteins
Besides their endometrial receptors, IGFs are bound to specific binding proteins (IGFBPs). At least six different types of IGFbinding proteins have been reported. IGFBP-1 has appeared in the literature under various names, such as amniotic fluid somatomedin-binding protein (Povoa et al., 1984), a,pregnancy-associated endometrial protein (Table 1) (Bell and Bohn, 1986) and placental protein 12 (Rutanen et al., 1982; 1984b; Koistinen et al., 1986). Julkunen and coworkers (1988a) have isolated a human decidual cDNA encoding IGFBP-1 and obtained its complete primary structure. This sequence does not have homology to the sequence of type 1 or type 2 IGF receptors (Lee et al., 1988). Secretory and decidual cells of the human endometrium produce IGFBP-1. This has been verified by incorporation of labelled methionine (Rutanen et al., 1985; 1986; Bell, 1986), as well as by the detection of specific IGFBP-1 mRNA (Julkunen et al., 1988a). IGFBP-1 mRNA has been localized by in situ hybridization to stromal cells in the secretory phase, whereas no message has been found in secretory glands or in any cell type of proliferative stage endometrium (Julkunen et al., 1990). No corresponding protein has been detected in proliferative 32
endometrium either (Rutanen et al., 1984a; Wahlstrom and SeppalM, 1984). After ovulation, the IGFBP-1 concentration in the tissue increases towards the end of the luteal phase, and progesterone and progestins enhance IGFBP-1 release from cultured endometrial tissue explants and cultured stromal cells (Rutanen et al., 1986; Bell et al., 1991). Besides progestins, relaxin is a potent stimulator of human endometrial stromal cell IGFBP secretion (Bell et al., 1991). IGFBP-1 inhibits the binding of IGF-I to its endometrial receptor (Rutanen et al., 1988). In view of these studies, it appears that IGFBP-1, whose synthesis in the endometrium is triggered or enhanced by progesterone, could be one of the factors which locally block the mitogenic effect of IGF-I, thereby participating in the proliferative/secretory transition of the endometrium. We and others have found that IGFBP-1 is detectable by immunoperoxidase staining in the secretory endometrium from day 4 after ovulation onwards, i.e., before implantation is expected to take place in a fertile cycle (Wahlstrom et al., 1985). Staining of IGFBP-1 is found mainly in the stroma (Waites et al., 1989). While there is an early report describing IGFBP-1/ PP12 in the glands (Wahlstrom and SeppSIS, 1984), it is quite clear now that glandular staining cannot be a sign of synthesis, because the glands contain no specific mRNA (Julkunen et al., 1990). However, it is noteworthy that in the baboon, IGFBP-1 is mainly localized in the endometrial glands (Fazleabas et al., 1989). It is possible that IGFBP-1 plays a role in a paracrine interaction between the stroma and the glands, by which the glands may acquire the IGFBP-1 detected by some authors. A rat homologue to human IGFBP-1 has been identified and its cDNA has been cloned from rat decidual cDNA (Murphy et al.. 1990). The rat IGFBP-1 has 66% sequence homology with human IGFBP-1, and the rat IGFBP-1 cDNA hybridizes with a 1.6 kilobase transcript in rat uterus, liver, kidney and brain (Murphy et al., 1990). Human IGFBP-1 transcripts have been found in the liver (Julkunen et al., 1988a), endometrium and decidua (Julkunen et al., 1988a; 1990), ovary (Koistinen et al., 1990) and kidney (Suikkari et al., 1992). The human IGFBP-1 was first considered independent of growth hormone. In hypophysectomized rats, hepatic IGFBP-1 expression was up-regulated in growth hormone deficiency, and downregulated after a single i.p. injection of growth hormone (Seneviratne et al., 1990). In a recent placebo-controlled study on infertile women, Tapanainen and coworkers (1991) found that growth hormone suppresses the circulating level of human IGFBP-1 indicating that, as in the rat, the human serum IGFBP-1 level is not entirely independent of growth hormone. The effects of decidualization on the uterine expression of IGF-I and IGFBP-1 have been studied in the hypophysectomized-ovariectomized rat (Croze et al., 1990). The changes in IGF-I mRNA abundance appear to be related to oestradiol injections. IGFBP-1 mRNA has been reported to be undetectable in the early decidualization process, whereas it reaches maximal levels on day six. These experiments also show that, in the rat, uterine IGFBP-1 expression is independent of the pituitary during decidualization (Croze et al., 1990), contrary to observations in the liver (Murphy et al., 1990). The liver is assumed to be the major source of circulating
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
the oestrogen-induced uterine growth (Brigstock el al., 1989). Oestradiol probably acts by causing cleavage of the EGF precursor to mature EGF, and by stimulating synthesis of the EGF receptor (DiAugustine, et al, 1988). Experiments in ovariectomized mice have shown that EGF can replace oestrogen in the stimulation of genital tract growth and differentiation (Nelson et al., 1991). EGF-induced mitogenesis does not require pituitary or adrenal hormones, and EGF mimics oestrogen in the induction of uterine lactoferrin (Nelson et al., 1991). Insulin-like growth factors, IGF-I and IGF-II, are mitogenic peptides which potentiate the activity of other mitogens, such as EGF and PDGF. Growth hormone acts by increasing IGF-I synthesis in the liver, and probably also at other sites in the body. IGFs are believed to play a part in endometrial differentiation and trophoblast growth. In the rat, uterine cells produce IGF-I, and oestradiol stimulates expression of IGF-I mRNA (Murphy and Friesen, 1988). Besides IGF-I, endometrial glands and surface epithelium contain growth hormone receptors and type 1 IGF receptors (Lobie et al., 1990). Both oestradiol and growth hormone increase IGF-I production by the uterus. Thus, the uterus appears to be a site of direct action of growth hormone. An example of the effects of IGF-I on uterine protein synthesis is the stimulation of the synthesis and release of prolactin from human decidual cells (Thraikill et al., 1988). Our recent studies have shown that one of the IGF-binding proteins, IGFBP-1, inhibits the IGF-I induced enhancement of prolactin release by decidua (Seppala et al., 1991). Both IGF-I and IGF-II bind to specific receptors on the human endometrium (Rutanen et al., 1988). No sequence homology has been found between type 1 and type 2 IGF receptors and the insulin receptor (Rinderknech and Humbel, 1978a,b).
Endometrial proteins
in stromal cells (Giudice et al, 1991). Oestradiol and progestins enhance IGFBP-3 synthesis, although the enhancement of IGFBP-3 is not so marked as is true of IGFBP-2 (Giudice et al, 1991). Thus, the human endometrium contains transcripts for IGFBP-1, IGFBP-2 and IGFBP-3, and this tissue has the capacity to translate these mRNAs into their respective proteins (Julkunen et al, 1988a; Rutanen et al, 1986; Julkunen et al., 1990; Giudice et al., 1991). Different types of binding proteins are synthesized in proliferative and secretory phases, and, besides the stimulatory effect on prolactin (Chen et al., 1989) and aromatase activity (Tseng et al., 1987), progesterone or progestins appear to be the key steroids enhancing the synthesis of the three IGF-binding proteins. The biological significance of the changing pattern of endometrial IGF-binding proteins during different phases of the cycle is of great interest as regards their biological role as local regulators of IGF action. Glandular secretory protein PP14
Like IGFBP-1, this protein, too, has several names in the literature. Thus, PP14 (placental protein 14), PEP (progesterone-associated endometrial protein), AUP (alpha-uterine protein) and a2-PEG (pregnancy-associated endometrial a2globulin) appear to be the same protein (see Bell and Bohn, 1986; Julkunen etal, 1986b; Sutcliffeef al., 1982). Any of these names can be found in the literature (see Table I).
Table I. Human endometrial proteins Prolactin (Maslar and Riddick, 1979) Alkaline phosphatase (Galski et al., 1982) 17p-hydroxysteroid dehydrogenase (Strinden and Shapiro, 1983) Diamine oxidase (Holinka and Gurpide, 1984) • • * *
Placental protein 12 (Bohn el al., 1982; Rutanen el al., 1985) Endometrial protein 14 (Bell, 1985) Pregnancy-associated endometrial al-globulin (Bell, 1986) Insulin-like growth factor-binding protein-1 (Koistinen etal., 1986)
•* " •• *• •* •* **
Progesterone-dependent endometrial protein (Joshi et al., 1980) Chorionic a2-microglobulin (Petrunin et al., 1980) Alpha-uterine protein (SutclifTe et al., 1982) Endometrial protein 15 (Bell et al., 1985) Placental protein 14 (Bohn etal., 1982; Julkunen et al., 1986; 1988) a2-pregnancy-associated endometrial globulin (Bell, 1986) Endometrial B-lactoglobulin homologue (Huhtala et al., 1987)
Pregnancy associated plasma protein A (SjOberg et al., 1984) Placental protein 5 (Butzow et al., 1986) Endometrial proteins 1-17 (Bell, 1986) Albumin (Fay et al.. 1990) Alpha 1-antitrypsin (Fay et al., 1990) Ceruloplasmin (Fay el al.. 1990) Bcta-lipoprotein (Fay et al., 1990) Alpha2-macroglobulin (Fay el al., 1990) Fibronectin (Fay el al., 1990) Complement factors C3 and C4 (Fay el al., 1990) *, likely to be the same proteins **, likely to be the same proteins
33
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
IGFBP-1 in nonpregnant women. The serum IGFBP-1 concentration is inversely correlated with that of insulin (Suikkari et al, 1988), which suppresses release of IGFBP-1 from HepG2 liver cancer cells (Singh et al, 1990) and from the endometrium (Thrailkill et al., 1990). IGF-I has been found to suppress IGFBP-1 release from human endometrium and relaxin stimulates the same (Thrailkill et al., 1990). The human granulosa cells are yet another site of IGFBP-1 production (Suikkari et al., 1989; Koistinen et al., 1990). Thus, in view of the multiple sites of synthesis and regulation of IGFBP-1 secretion by many factors which are not menstrual cycle-dependent, it is not surprising that no consistent variation has been found in the circulating IGFBP-1 levels during the menstrual cycle (Suikkari et al., 1987). However, there is a significant diurnal variation, the highest levels occurring at night (Rutanen et al., 1984b; Baxter and Cowell, 1987). During pregnancy, the levels are highest at 18-28 weeks (Rutanen et al., 1982). In the third trimester, maternal serum levels of IGFBP-1 become elevated in fetal growth retardation (Howell et al., 1985) and pre-eclampsia (lino et al, 1986), and it has been suggested that the rise in IGFBP-1 could be a primary factor in retardation of fetal growth (Wang et al, 1991). Fasting has been shown to increase rat hepatic IGFBP-1 mRNA (Murphy et al, 1991), and fasting also increases the serum IGFBP-1 level in obese women (Kiddy et al., 1989). The latter authors have ascribed this effect to decreased serum insulin levels during fasting, as insulin has been found to suppress IGFBP-1 secretion in vivo (Suikkari et al., 1988) and in cultured human primary liver cancer, HepG2, cells (Singh et al, 1990). IGFBP-2 is the major IGF-binding protein in cerebrospinal fluid (Binoux et al. 1982). IGFBP-2 mRNA has also been detected in the human endometrium by Guidice and coworkers (1991), who, using a cDNA probe specific for IGFBP-2, found that a 1.4 kilobase mRNA was differentially expressed in proliferative and secretory endometrium. Unlike IGFBP-1 mRNA, IGFBP-2 mRNA transcripts have been detected in both proliferative and secretory endometrium, their expression being higher in the secretory phase. Progesterone markedly stimulates endometrial synthesis of IGFBP-2. This effect can be blocked by a progesterone antagonist RU 486 (Giudice et al., 1990), suggesting progesterone-dependence of IGFBP-2 synthesis. Endometrial stromal cells synthesize IGFBP-2, and this synthesis is markedly enhanced by oestradiol and progesterone (Giudice et al, 1991). IGFBP-3 is the major IGF carrier in serum (Baxter, 1988). Full length cDNA clones of this protein, also called growth hormone dependent IGFBP, or BP-53, have been isolated and the complete sequence has been deduced (Wood et al., 1988). There is 33% sequence homology between IGFBP-3 and IGFBP-1, including conservation of all 18 cysteine residues (Wood et al., 1988). Besides its synthesis in the liver, IGFBP-3 is the major IGFBP synthesized by term human placenta (Deal et al, 1990). Messenger RNA encoding this protein has also been detected in the endometrium, where it was first reported to be localized to proliferative glands (Giudice et al., 1990). Endometrial stromal cell cultures have subsequently been established and these studies have shown constitutive synthesis and secretion of IGFBP-3
M-Seppilfi at aL
Table II. PP14 concentration (median values and range) in endometrial tissue relative to the phase of differentiation (Data from Julkunen, 1986)
(Julkunen et al, 1986d; 1990). This is not unexpected in view of the Mullerian origin of both the uterus and the Fallopian tubes. Yet it is not clear whether PPM is entirely endometriumspecific, because sufficient numbers of various tissues from various phases of the menstrual cycle have not been studied. Serum PP14 levels related to hormone stimulation and cyclical changes The tissue PPM concentration is reflected in circulating levels. The lowest levels are seen at the time of ovulation. The levels rise during the last week of the luteal phase and peak at the onset of menstruation (Julkunen et al, 1986c). The mean doubling time is three days. No similar pattern is seen in anovulatory cycles. At the onset of menstruation, there is virtually no overlap in the levels between ovulatory and anovulatory cycles. Therefore, it is possible to tell from an elevated serum PPM level taken at the onset of menstrual bleeding whether the cycle was ovulatory and whether the endometrium had responded to elevated levels of progesterone or not. In women with an inadequate luteal phase, the serum progesterone and PEP levels are low (Joshi et al.. 1988). We have addressed the question of the magnitude of the uterine contribution to circulating PPM levels (Seppala' et al., 1988a). This was studied in two groups of postmenopausal women receiving similar oestrogen-progestogen replacement therapy. One group had an intact uterus, while the other group had undergone hysterectomy. As expected, the PPM levels were initially low in either group. After three weeks' cyclical oestrogenyprogestogen treatment, the levels in women with an intact uterus were significantly higher than those in hysterectomized women. We have also measured serum PPM levels in women after hysterectomy plus ovariectomy, and there was no significant difference from the levels measured after hysterectomy alone. The PPM responses were quite variable in the intact uterus group when the treatment ceased. Some women exhibited a greater than 100% rise, whereas others showed no rise at all. This demonstrates that the postmenopausal uterus may respond to hormone stimulation by PPM production, but the responses may vary considerably. In women with unexplained infertility, micronized oral progesterone increases the serum PPM concentration (Seppala' et al., 1987). We have also found that suppression of the serum prolactin level by bromocriptine is followed by elevated levels of oestradiol on day 9, and elevated levels of PPM on day 22 of the cycle. It seems that the higher the serum oestradiol level is on day 9, the higher is the PPM level in the late luteal phase. Clearly, either oestrogen or suppression of prolactin were responsible for the elevated serum PPM level in the luteal phase (Seppala et al.. 1989).
Endometrial histology
PP14 Hg/g protein
Proliferative
0.15 (
Endometrial proteins: A reappraisal
Markku Seppala, Mervi Julkunen, Leena Riittinen and Riitta Koisrinen Department I of Obstetrics and Gynaecology, Helsinki University Central Hospital SF-OO29O Helsinki, Finland
Key Words: endometrion, protein, PPM, IGF
Introduction
The human uterus has been considered as an end organ which does not send feed-back signals to other parts of the body. Recent evidence indicates that this may not be entirely the case. The uterus responds to various endocrine stimuli by secreting local mediator substances and proteins which can also be found in peripheral blood. A physiological cause of infertility may lie in uterine responsiveness. The hormonal stimulus may be inadequate, or the uterine responsiveness not synchronized with embryonic development, even in the presence of apparently adequate levels of ovarian steroid hormones. Therefore, understanding uterine physiology is the key in any attempt at improving fertility since adequate embryo-uterine interactions are essential. Primary dysfunction of the endometrium may be associated with unexplained infertility, even when progesterone secretion is normal. This was indicated in a recent study by Graham et al. (1990), in which immunohistochemistry with monoclonal antibody D9B1 was used to assess the production and secretion of an oligosaccharide epitope produced by endometrial glands between two and seven days after the luteinizing hormone surge. In women with infertility, the © Oxford University Press
Endometrial cell types
The endometrium contains epithelial cells, stromal cells, endothelial cells of blood vessels, transiently resident cells, the extracellular matrix and connective tissue. In the proliferative phase, both the glands and the stroma contain oestrogen and progesterone receptors. After ovulation, the glandular receptors disappear under the influence of progesterone, while stromal receptors are maintained (Bouchard et al., 1991). Endometrial biopsies obtained during oestrogen replacement for the establishment of a receptive endometrium in women with premature menopause have revealed the expression of an oestrogen receptor-related antigen in the cytoplasm of the epithelium. After three days of progesterone replacement, the endometrium showed normal secretory features and expression of the oestrogen receptor-related antigen now appeared in the stroma as well as in the glands (Critchley et al., 1990a). Stromal oedema is maximal from days LH + 6 to LH + 8, which is believed to correspond to the beginning of the so called "implantation window" (Johannisson, 1991). In the secretory phase, the actions of oestrogen and progesterone on epithelial cells are likely to be mediated in a paracrine fashion by factors secreted from the neighbouring stromal cells.
Growth factors
The endometrium produces several growth factors such as epidermal growth factor (EGF), transforming growth factors (TGFa, TGFpi), platelet-derived growth factor (PDGF) and insulin-like growth factor-I (IGF-I). EGF is a product of the uterine epithelial cells. There are EGF-binding sites in the human uterus (Hoffmann et al., 1984), both in the glands and in the stroma (Bergchuck et al., 1989; Smith et al., 1991). Uterine EGF secretion is induced by oestrogen (Paria and Dey, 1990) and EGF is involved in mediating 31
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
Uterine factors influence reproduction at the macro-anatomy level, and the effects of hormonal steroids on endometrial morphology are well recognized in the histopathological diagnosis of dysfunctional bleeding and infertility. During the past decade, attention has been paid to endometrial protein synthesis and secretion with respect to endocrine stimuli and implantation, and to the paracrine/autocrine effects of endometrial peptide growth factors, their binding proteins and other factors. The emphasis of this presentation is on protein secretion of the secretory endometrium, in which progesterone plays a pivotal role. Insulin-like growth factors have receptors on the endometrium, and IGF-binding proteins, stimulated by progesterone, modulate the effects of IGFs locally. Also other protein products of the secretory endometrium have been reviewed in this communication, with special emphasis on studies of a progesterone-associated endometrial protein which has many names in the literature, such as PEP, PP14, o2-PEG and AUP. Extensive studies are ongoing in many laboratories to elucidate the regulation, function, interplay at tissue and cellular levels, and clinical significance of these proteins.
synthesis and secretion of this epitope was significantly reduced and delayed in the presence of normal serum progesterone levels. Drugs used to improve pregnancy rates in patients undergoing IVF may have adverse effects on the endometrium. In their recent work, Paulson and coworkers (1990) observed that a woman whose uterus was not subject to the effects of fertility drugs had an implantation rate three times higher than women whose uterine environment had been influenced by drugs given to improve fertility. This presentation will address endometrial protein secretion during the peri-implantation period. Of particular interest are paracrine and autocrine mechanisms, including mediators of hormone action, such as growth factors, their receptors and binding proteins. The emphasis of this presentation is on two major proteins of secretory endometrium, namely insulin-like growth factor binding protein-1 and placental protein 14.
M.Seppilfl at al
IGF-binding proteins
Besides their endometrial receptors, IGFs are bound to specific binding proteins (IGFBPs). At least six different types of IGFbinding proteins have been reported. IGFBP-1 has appeared in the literature under various names, such as amniotic fluid somatomedin-binding protein (Povoa et al., 1984), a,pregnancy-associated endometrial protein (Table 1) (Bell and Bohn, 1986) and placental protein 12 (Rutanen et al., 1982; 1984b; Koistinen et al., 1986). Julkunen and coworkers (1988a) have isolated a human decidual cDNA encoding IGFBP-1 and obtained its complete primary structure. This sequence does not have homology to the sequence of type 1 or type 2 IGF receptors (Lee et al., 1988). Secretory and decidual cells of the human endometrium produce IGFBP-1. This has been verified by incorporation of labelled methionine (Rutanen et al., 1985; 1986; Bell, 1986), as well as by the detection of specific IGFBP-1 mRNA (Julkunen et al., 1988a). IGFBP-1 mRNA has been localized by in situ hybridization to stromal cells in the secretory phase, whereas no message has been found in secretory glands or in any cell type of proliferative stage endometrium (Julkunen et al., 1990). No corresponding protein has been detected in proliferative 32
endometrium either (Rutanen et al., 1984a; Wahlstrom and SeppalM, 1984). After ovulation, the IGFBP-1 concentration in the tissue increases towards the end of the luteal phase, and progesterone and progestins enhance IGFBP-1 release from cultured endometrial tissue explants and cultured stromal cells (Rutanen et al., 1986; Bell et al., 1991). Besides progestins, relaxin is a potent stimulator of human endometrial stromal cell IGFBP secretion (Bell et al., 1991). IGFBP-1 inhibits the binding of IGF-I to its endometrial receptor (Rutanen et al., 1988). In view of these studies, it appears that IGFBP-1, whose synthesis in the endometrium is triggered or enhanced by progesterone, could be one of the factors which locally block the mitogenic effect of IGF-I, thereby participating in the proliferative/secretory transition of the endometrium. We and others have found that IGFBP-1 is detectable by immunoperoxidase staining in the secretory endometrium from day 4 after ovulation onwards, i.e., before implantation is expected to take place in a fertile cycle (Wahlstrom et al., 1985). Staining of IGFBP-1 is found mainly in the stroma (Waites et al., 1989). While there is an early report describing IGFBP-1/ PP12 in the glands (Wahlstrom and SeppSIS, 1984), it is quite clear now that glandular staining cannot be a sign of synthesis, because the glands contain no specific mRNA (Julkunen et al., 1990). However, it is noteworthy that in the baboon, IGFBP-1 is mainly localized in the endometrial glands (Fazleabas et al., 1989). It is possible that IGFBP-1 plays a role in a paracrine interaction between the stroma and the glands, by which the glands may acquire the IGFBP-1 detected by some authors. A rat homologue to human IGFBP-1 has been identified and its cDNA has been cloned from rat decidual cDNA (Murphy et al.. 1990). The rat IGFBP-1 has 66% sequence homology with human IGFBP-1, and the rat IGFBP-1 cDNA hybridizes with a 1.6 kilobase transcript in rat uterus, liver, kidney and brain (Murphy et al., 1990). Human IGFBP-1 transcripts have been found in the liver (Julkunen et al., 1988a), endometrium and decidua (Julkunen et al., 1988a; 1990), ovary (Koistinen et al., 1990) and kidney (Suikkari et al., 1992). The human IGFBP-1 was first considered independent of growth hormone. In hypophysectomized rats, hepatic IGFBP-1 expression was up-regulated in growth hormone deficiency, and downregulated after a single i.p. injection of growth hormone (Seneviratne et al., 1990). In a recent placebo-controlled study on infertile women, Tapanainen and coworkers (1991) found that growth hormone suppresses the circulating level of human IGFBP-1 indicating that, as in the rat, the human serum IGFBP-1 level is not entirely independent of growth hormone. The effects of decidualization on the uterine expression of IGF-I and IGFBP-1 have been studied in the hypophysectomized-ovariectomized rat (Croze et al., 1990). The changes in IGF-I mRNA abundance appear to be related to oestradiol injections. IGFBP-1 mRNA has been reported to be undetectable in the early decidualization process, whereas it reaches maximal levels on day six. These experiments also show that, in the rat, uterine IGFBP-1 expression is independent of the pituitary during decidualization (Croze et al., 1990), contrary to observations in the liver (Murphy et al., 1990). The liver is assumed to be the major source of circulating
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
the oestrogen-induced uterine growth (Brigstock el al., 1989). Oestradiol probably acts by causing cleavage of the EGF precursor to mature EGF, and by stimulating synthesis of the EGF receptor (DiAugustine, et al, 1988). Experiments in ovariectomized mice have shown that EGF can replace oestrogen in the stimulation of genital tract growth and differentiation (Nelson et al., 1991). EGF-induced mitogenesis does not require pituitary or adrenal hormones, and EGF mimics oestrogen in the induction of uterine lactoferrin (Nelson et al., 1991). Insulin-like growth factors, IGF-I and IGF-II, are mitogenic peptides which potentiate the activity of other mitogens, such as EGF and PDGF. Growth hormone acts by increasing IGF-I synthesis in the liver, and probably also at other sites in the body. IGFs are believed to play a part in endometrial differentiation and trophoblast growth. In the rat, uterine cells produce IGF-I, and oestradiol stimulates expression of IGF-I mRNA (Murphy and Friesen, 1988). Besides IGF-I, endometrial glands and surface epithelium contain growth hormone receptors and type 1 IGF receptors (Lobie et al., 1990). Both oestradiol and growth hormone increase IGF-I production by the uterus. Thus, the uterus appears to be a site of direct action of growth hormone. An example of the effects of IGF-I on uterine protein synthesis is the stimulation of the synthesis and release of prolactin from human decidual cells (Thraikill et al., 1988). Our recent studies have shown that one of the IGF-binding proteins, IGFBP-1, inhibits the IGF-I induced enhancement of prolactin release by decidua (Seppala et al., 1991). Both IGF-I and IGF-II bind to specific receptors on the human endometrium (Rutanen et al., 1988). No sequence homology has been found between type 1 and type 2 IGF receptors and the insulin receptor (Rinderknech and Humbel, 1978a,b).
Endometrial proteins
in stromal cells (Giudice et al, 1991). Oestradiol and progestins enhance IGFBP-3 synthesis, although the enhancement of IGFBP-3 is not so marked as is true of IGFBP-2 (Giudice et al, 1991). Thus, the human endometrium contains transcripts for IGFBP-1, IGFBP-2 and IGFBP-3, and this tissue has the capacity to translate these mRNAs into their respective proteins (Julkunen et al, 1988a; Rutanen et al, 1986; Julkunen et al., 1990; Giudice et al., 1991). Different types of binding proteins are synthesized in proliferative and secretory phases, and, besides the stimulatory effect on prolactin (Chen et al., 1989) and aromatase activity (Tseng et al., 1987), progesterone or progestins appear to be the key steroids enhancing the synthesis of the three IGF-binding proteins. The biological significance of the changing pattern of endometrial IGF-binding proteins during different phases of the cycle is of great interest as regards their biological role as local regulators of IGF action. Glandular secretory protein PP14
Like IGFBP-1, this protein, too, has several names in the literature. Thus, PP14 (placental protein 14), PEP (progesterone-associated endometrial protein), AUP (alpha-uterine protein) and a2-PEG (pregnancy-associated endometrial a2globulin) appear to be the same protein (see Bell and Bohn, 1986; Julkunen etal, 1986b; Sutcliffeef al., 1982). Any of these names can be found in the literature (see Table I).
Table I. Human endometrial proteins Prolactin (Maslar and Riddick, 1979) Alkaline phosphatase (Galski et al., 1982) 17p-hydroxysteroid dehydrogenase (Strinden and Shapiro, 1983) Diamine oxidase (Holinka and Gurpide, 1984) • • * *
Placental protein 12 (Bohn el al., 1982; Rutanen el al., 1985) Endometrial protein 14 (Bell, 1985) Pregnancy-associated endometrial al-globulin (Bell, 1986) Insulin-like growth factor-binding protein-1 (Koistinen etal., 1986)
•* " •• *• •* •* **
Progesterone-dependent endometrial protein (Joshi et al., 1980) Chorionic a2-microglobulin (Petrunin et al., 1980) Alpha-uterine protein (SutclifTe et al., 1982) Endometrial protein 15 (Bell et al., 1985) Placental protein 14 (Bohn etal., 1982; Julkunen et al., 1986; 1988) a2-pregnancy-associated endometrial globulin (Bell, 1986) Endometrial B-lactoglobulin homologue (Huhtala et al., 1987)
Pregnancy associated plasma protein A (SjOberg et al., 1984) Placental protein 5 (Butzow et al., 1986) Endometrial proteins 1-17 (Bell, 1986) Albumin (Fay et al.. 1990) Alpha 1-antitrypsin (Fay et al., 1990) Ceruloplasmin (Fay el al.. 1990) Bcta-lipoprotein (Fay et al., 1990) Alpha2-macroglobulin (Fay el al., 1990) Fibronectin (Fay el al., 1990) Complement factors C3 and C4 (Fay el al., 1990) *, likely to be the same proteins **, likely to be the same proteins
33
Downloaded from http://humrep.oxfordjournals.org/ at Deakin University Library on March 16, 2016
IGFBP-1 in nonpregnant women. The serum IGFBP-1 concentration is inversely correlated with that of insulin (Suikkari et al, 1988), which suppresses release of IGFBP-1 from HepG2 liver cancer cells (Singh et al, 1990) and from the endometrium (Thrailkill et al., 1990). IGF-I has been found to suppress IGFBP-1 release from human endometrium and relaxin stimulates the same (Thrailkill et al., 1990). The human granulosa cells are yet another site of IGFBP-1 production (Suikkari et al., 1989; Koistinen et al., 1990). Thus, in view of the multiple sites of synthesis and regulation of IGFBP-1 secretion by many factors which are not menstrual cycle-dependent, it is not surprising that no consistent variation has been found in the circulating IGFBP-1 levels during the menstrual cycle (Suikkari et al., 1987). However, there is a significant diurnal variation, the highest levels occurring at night (Rutanen et al., 1984b; Baxter and Cowell, 1987). During pregnancy, the levels are highest at 18-28 weeks (Rutanen et al., 1982). In the third trimester, maternal serum levels of IGFBP-1 become elevated in fetal growth retardation (Howell et al., 1985) and pre-eclampsia (lino et al, 1986), and it has been suggested that the rise in IGFBP-1 could be a primary factor in retardation of fetal growth (Wang et al, 1991). Fasting has been shown to increase rat hepatic IGFBP-1 mRNA (Murphy et al, 1991), and fasting also increases the serum IGFBP-1 level in obese women (Kiddy et al., 1989). The latter authors have ascribed this effect to decreased serum insulin levels during fasting, as insulin has been found to suppress IGFBP-1 secretion in vivo (Suikkari et al., 1988) and in cultured human primary liver cancer, HepG2, cells (Singh et al, 1990). IGFBP-2 is the major IGF-binding protein in cerebrospinal fluid (Binoux et al. 1982). IGFBP-2 mRNA has also been detected in the human endometrium by Guidice and coworkers (1991), who, using a cDNA probe specific for IGFBP-2, found that a 1.4 kilobase mRNA was differentially expressed in proliferative and secretory endometrium. Unlike IGFBP-1 mRNA, IGFBP-2 mRNA transcripts have been detected in both proliferative and secretory endometrium, their expression being higher in the secretory phase. Progesterone markedly stimulates endometrial synthesis of IGFBP-2. This effect can be blocked by a progesterone antagonist RU 486 (Giudice et al., 1990), suggesting progesterone-dependence of IGFBP-2 synthesis. Endometrial stromal cells synthesize IGFBP-2, and this synthesis is markedly enhanced by oestradiol and progesterone (Giudice et al, 1991). IGFBP-3 is the major IGF carrier in serum (Baxter, 1988). Full length cDNA clones of this protein, also called growth hormone dependent IGFBP, or BP-53, have been isolated and the complete sequence has been deduced (Wood et al., 1988). There is 33% sequence homology between IGFBP-3 and IGFBP-1, including conservation of all 18 cysteine residues (Wood et al., 1988). Besides its synthesis in the liver, IGFBP-3 is the major IGFBP synthesized by term human placenta (Deal et al, 1990). Messenger RNA encoding this protein has also been detected in the endometrium, where it was first reported to be localized to proliferative glands (Giudice et al., 1990). Endometrial stromal cell cultures have subsequently been established and these studies have shown constitutive synthesis and secretion of IGFBP-3
M-Seppilfi at aL
Table II. PP14 concentration (median values and range) in endometrial tissue relative to the phase of differentiation (Data from Julkunen, 1986)
(Julkunen et al, 1986d; 1990). This is not unexpected in view of the Mullerian origin of both the uterus and the Fallopian tubes. Yet it is not clear whether PPM is entirely endometriumspecific, because sufficient numbers of various tissues from various phases of the menstrual cycle have not been studied. Serum PP14 levels related to hormone stimulation and cyclical changes The tissue PPM concentration is reflected in circulating levels. The lowest levels are seen at the time of ovulation. The levels rise during the last week of the luteal phase and peak at the onset of menstruation (Julkunen et al, 1986c). The mean doubling time is three days. No similar pattern is seen in anovulatory cycles. At the onset of menstruation, there is virtually no overlap in the levels between ovulatory and anovulatory cycles. Therefore, it is possible to tell from an elevated serum PPM level taken at the onset of menstrual bleeding whether the cycle was ovulatory and whether the endometrium had responded to elevated levels of progesterone or not. In women with an inadequate luteal phase, the serum progesterone and PEP levels are low (Joshi et al.. 1988). We have addressed the question of the magnitude of the uterine contribution to circulating PPM levels (Seppala' et al., 1988a). This was studied in two groups of postmenopausal women receiving similar oestrogen-progestogen replacement therapy. One group had an intact uterus, while the other group had undergone hysterectomy. As expected, the PPM levels were initially low in either group. After three weeks' cyclical oestrogenyprogestogen treatment, the levels in women with an intact uterus were significantly higher than those in hysterectomized women. We have also measured serum PPM levels in women after hysterectomy plus ovariectomy, and there was no significant difference from the levels measured after hysterectomy alone. The PPM responses were quite variable in the intact uterus group when the treatment ceased. Some women exhibited a greater than 100% rise, whereas others showed no rise at all. This demonstrates that the postmenopausal uterus may respond to hormone stimulation by PPM production, but the responses may vary considerably. In women with unexplained infertility, micronized oral progesterone increases the serum PPM concentration (Seppala' et al., 1987). We have also found that suppression of the serum prolactin level by bromocriptine is followed by elevated levels of oestradiol on day 9, and elevated levels of PPM on day 22 of the cycle. It seems that the higher the serum oestradiol level is on day 9, the higher is the PPM level in the late luteal phase. Clearly, either oestrogen or suppression of prolactin were responsible for the elevated serum PPM level in the luteal phase (Seppala et al.. 1989).
Endometrial histology
PP14 Hg/g protein
Proliferative
0.15 (
