TY - JOUR
AU - Westwood, M.
AB - Abstract During pregnancy, insulin-like growth factors (IGFs) are important for growth of fetal and maternal tissues. One of the IGF binding proteins, IGFBP-1, is thought to regulate their activity within the local environment of the placenta. IGFBP-1 usually exists as a phosphorylated, high affinity species, which sequesters IGFs, thereby inhibiting their actions. This study has investigated the mechanisms that release IGF from IGFBP-1 at the maternal–fetal interface. Under basal conditions, human decidualized endometrium produces both non-phosphorylated (np) and phosphorylated (p) isoforms of IGFBP-1; however, in the presence of IGF-II, which is a trophoblast secretory product, npIGFBP-1 was preferentially produced. Furthermore, we found that trophoblast, presumably via placental alkaline phosphatase, can de-phosphorylate pIGFBP-1. Since npIGFBP-1 has decreased affinity for IGF-I, these effects should enhance IGF-I bioavailability. In addition, we found that decidual cells produce a protease, which cleaves IGFBP-1, but only when it is non-phosphorylated; [125I]-npIGFBP-1 is proteolysed into 14 and 17 kDa fragments which have markedly reduced affinity for IGF. We therefore propose paracrine modulation of IGFBP-1 at the maternal–fetal interface involving a multi-step process of de-phosphorylation and proteolysis; this will result in enhanced IGF bioavailability and is likely to represent an important mechanism for controlling fetal and maternal tissue growth. decidua, IGFBP-1, phosphorylation, proteolysis, trophoblast Introduction Insulin-like growth factors (IGFs) and the binding proteins (IGFBPs) that control their activity are crucial for fetal growth and development. This is shown by studies in mice, where ablation of either the IGF-I or IGF-II gene reduces fetal size to 60% that of normal littermates (Powell et al., 1993). If both genes are knocked out, then pup size is reduced to 30% (Liu et al., 1993). In humans, the IGF axis is disrupted in disorders of fetal growth; IGF-I concentrations are decreased in cord sera from intra-uterine growth restricted (IUGR) or small for gestational age (SGA) fetuses (Giudice et al., 1995) and are increased in large for gestational age (LGA) newborns (Giudice et al., 1995). Successful fetal growth requires a developmentally and functionally adequate maternal–fetal interface (Han, 1993). Han et al. (1996) have shown that IGF-I and IGF-II, though especially the latter, are present in placental trophoblasts and fetal membranes from as early as 6 weeks. IGF-II expression is most abundant in the trophoblastic columns of the anchoring villi, particularly in cells at the leading edge of the column, suggesting that IGF-II may have a role in trophoblast invasion of the endometrium (Han et al., 1996). Both peptides are absent from decidualized endometrium; however IGFBP-1, an important IGF binding protein in pregnancy, is a major secretory product of decidual cells (Waites et al., 1988). It is also the predominant IGFBP in amniotic fluid (Drop et al., 1984), a major IGF binding species in fetal plasma (Drop et al., 1984) and its concentrations are increased in the maternal circulation during pregnancy. At term, maternal (Howell et al., 1985) and fetal IGFBP-1 (Verhaeghe et al., 1993) concentrations are negatively correlated with birth weight and in pregnancies complicated by IUGR or pre-eclampsia, raised maternal and fetal IGFBP-1 concentrations have been observed (Giudice et al., 1997). These studies all point to a role for IGFBP-1 as a local modulator of IGF action in fetal growth. In addition, IGFBP-1 may be able to have IGF-independent effects, since it has an RGD sequence which can interact with the α5β1 integrin present on cytotrophoblast cells (Irwin and Giudice, 1998). We have shown that, in the circulation, IGFBP-1 normally exists in a highly phosphorylated state (pIGFBP-1) (Westwood et al., 1994). However, during pregnancy, we found non-phosphorylated isoforms of IGFBP-1 (npIGFBP-1) in the maternal circulation (Westwood et al., 1994). Since pIGFBP-1 has a higher affinity for IGF-I than npIGFBP-1 (Jones et al., 1991; Westwood et al., 1997), changes in IGFBP-1 phosphorylation status may alter IGF-I bioavailability. However, phosphorylation does not affect the affinity of IGFBP-1 for IGF-II (Westwood et al., 1997) and, hence, promoting the presence of npIGFBP-1 at the fetal–maternal interface will not affect the actions of the predominant IGF at this site. Another mechanism for releasing IGFs from binding proteins is proteolysis. Pregnancy-associated proteolysis of several other binding proteins, particularly IGFBP-3, has been comprehensively described (Davies et al., 1991). These binding proteins are cleaved within the maternal circulation by enzymes that appear to have no affect on IGFBP-1. However, plasmin (Frost et al., 1993), stromelysin-3 (Manes et al., 1997) and amniotic fluid (Binoux et al., 1994) do have IGFBP-1 proteolytic activity, though little is known of the physiological relevance of an IGFBP-1 protease. This study aimed to investigate mechanisms for producing npIGFBP-1 at the fetal–maternal interface and to explore the possibility that IGFBP-1 proteolysis at this site might be a mechanism for increasing IGF-II bioavailability. Materials and methods Tissues and cell culture Human first trimester decidualized endometrial cells Samples of normal human first trimester decidua parietalis (8–12 weeks gestation) were obtained (with local ethical committee approval) after termination of pregnancy and carefully examined under a dissecting microscope to ensure absence of villous material. In accordance with our previously described method (Vicovacet al., 1994), decidual tissue was incubated at 37°C with 0.02% protease for 15 min followed by 0.2% hyaluronidase and 0.25% collagenase for 2 h. The resulting suspension was filtered initially through a 100 μm nylon sieve to remove undigested tissue fragments and then through a 40 μm sieve to retain whole glands and cell aggregates as previously shown for endometrium. Cells were then resuspended in 25% Percoll (Amersham Pharmacia Biotech UK Ltd, Bucks, UK), and layered over 60% Percoll. After centrifugation at 670 g for 30 min, cells at the 25/60% interface were collected, washed three times in phosphate-buffered saline (PBS), counted and plated at 0.5×106/well in 6-well plates. Cultures were maintained in Dulbecco's modified Eagle's medium (Sigma, Poole, Dorset, UK) (DMEM)/10% fetal calf serum (FCS)/100 μg/ml streptomycin (Sigma) and 100 IU/ml penicillin Sigma at 37°C in 5% CO2. Cultures were established on three separate occasions and each variable was tested in triplicate. BeWo choriocarcinoma cells BeWo cells were obtained from the European Collection of Animal Cell Culture (Porton Down, UK). They were cultured in an equal mixture of DMEM and Ham's F12 containing 10% FCS, 2 mmol/l glutamine, 5 μg/ml gentamicin and 100 μg/ml streptomycin at 37°C in 5% CO2. Methods Biochemical characterization of IGFBP-1 phosphorylation status IGFBP-1 phosphorylation status was determined using our previously described method of immunoprecipitation followed by n-octyl glucoside electrophoresis and Western ligand blotting (Westwood et al., 1994). Samples (adjusted to contain equal protein concentrations) were incubated with anti-IGFBP-1 (6303; a kind gift of Medix Biochemica, Kauniainen, Finland) antibody at 4°C overnight and then anti-mouse immunoglobulin (Ig)G antibody (Sac Cel; IDS, Tyne & Wear, UK) was added for 1 h at room temperature. Bound antibody was separated by centrifugation for 10 min and the precipitated proteins were washed in PBS/0.25% bovine serum albumin (BSA)/0.1% Tween 20 prior to resuspension in gel loading buffer. All samples were then boiled for 5 min. Electrophoresis was performed using stacking (4%) and resolving (12%) gels containing 20 mmol/l of the non-ionic detergent n-octyl glucoside (n-OG). Following overnight transfer onto nitrocellulose membranes, proteins were revealed by incubation with 150 000 cpm/ml [125I]-IGF-I (4 h at 25°C) and autoradiography. Radioimmunoassay of IGFBP-1 The concentrations of IGFBP-1 produced by decidualized endometrial cells in response to 7 days incubation with 0–500 ng/ml IGF-I, IGF-II or insulin were determined using our previously reported radioimmunoassay which measures all isoforms of IGFBP-1 (Westwood et al., 1997). Unless otherwise stated all chemicals used were obtained from Sigma. Purification of highly phosphorylated IGFBP-1 Highly phosphorylated IGFBP-1 was purified as previously described (Westwood et al., 1997) from plasma donated by healthy female subjects taking a combined oral contraceptive pill, since circulating concentrations of this isoform are elevated in these subjects (Westwood et al., 1999). Unless otherwise stated all chemicals used were obtained from Sigma. IGFBP-1 protease assay To determine whether IGFBP-1 is susceptible to proteolysis, samples reflecting IGFBP-1 sites of synthesis (conditioned medium from HepG2 and first trimester decidual cells) or potential target tissues [conditioned medium from first trimester trophoblast and 3T3-L1 adipocytes (pre- and post-differentiation)] were analysed on three occasions using a method based on an IGFBP-3 protease assay (Lamson et al., 1991). Normal plasma, pregnancy plasma, amniotic fluid and plasmin (0.025 IU) were used as controls (Frost et al., 1993; Binoux et al., 1994). Briefly, 30 000 cpm [125I]-recombinant human IGFBP-1 (iodinated by the chloramine T method) was incubated with a 5–40 μl sample in 0.5 mmol/l calcium chloride/PBS for 16 h at 37°C. The reaction was stopped by the addition of gel loading buffer and boiling for 5 min and the samples were subjected to sodium dodecyl sulphate (SDS) electrophoresis (15% resolving gel) and autoradiography. Characterization of IGFBP-1 protease To characterize the IGFBP-1 protease present in decidual cell conditioned medium, samples were incubated as described above in the presence of a range of protease inhibitors (Boehringer Manheim, Lewes, East Sussex). In addition, zymography was used to estimate the molecular weight of the protease. Samples were electrophoresed through a 10% poly-acrylamide/SDS gel containing 2 μg/ml recombinant human IGFBP-1. 50 mmol/l Tris, pH 8.0/5 mmol/l CaCl2 was used to activate the protease and simultaneously transfer the IGFBP-1 fragments to polyvinyldifluoridine (PVDF) membranes by capillary action (24 h at 37°C). The fragments were then visualized by Western immunoblotting. Western immunoblotting Non-specific binding was blocked by incubating membranes with 0.15 mol/l NaCl/1% BSA for 1 h at room temperature. They were then exposed to a 1/1000 dilution of anti-IGFBP-1 antibody (6303 or 6305) for 4 h at 25°C, followed by anti-mouse IgG linked to horseradish peroxidase for 2 h. Chemiluminescence (ECL, Amersham, UK) was used to reveal bound antibody. Effect of IGFBP-1 phosphorylation status on proteolysis Normal human plasma was used as the source of phosphorylated IGFBP-1. This was found by radioimmunoassay (Westwood et al., 1994) to contain 200 μg/l of IGFBP-1 and n-OG electrophoresis and Western ligand ([125I]-IGF-I + [125I]-IGF-II) blotting was used to confirm that only the highly phosphorylated isoform was present. Recombinant IGFBP-1, used as the non-phosphorylated isoform, was spiked at 200 μg/l into plasma stripped of endogenous IGFBP-1 by immunoaffinity chromatography (Westwood et al., 1997). Both IGFBP-1 preparations were incubated with decidual conditioned medium (50 μl) or plasmin (0.025 IU) for 16 h at 37°C and then the reaction was stopped by the addition of SDS loading buffer. Following electrophoresis through SDS–polyacrylamide (15%) gels, the proteins were transferred to nitrocellulose and IGFBP-1 detected by Western immunoblotting. This experiment was repeated on three separate occasions. Results Characterization of IGFBP-1 isoforms at the maternal–fetal interface IGFBP-1 production by human first trimester decidualized endometrial cells Under basal conditions, decidual cells produced the highly phosphorylated isoform found in the normal human circulation as well as the non- and lesser- phosphorylated IGFBP-1 isoforms characteristic of human pregnancy plasma (Figure 1A). In order to investigate the mechanisms controlling production of these isoforms, decidual cells were incubated for 7 days with factors (IGF-I, IGF-II and insulin; 0–500 ng/ml) known to regulate IGFBP-1 concentrations (Thrailkill et al., 1990; Irwin et al., 1993). In each of three experiments all of the factors stimulated IGFBP-1 production at lower concentrations, but inhibited IGFBP-1 productions at higher concentrations (Figure 1B). More interesting however, was our observation that these factors had a differential effect on IGFBP-1 phosphorylation status (Figure 1A). In response to IGF-II, the predominant IGF produced by trophoblast, decidualized endometrial cells produced relatively more non-phosphorylated IGFBP-1 than the highly phosphorylated isoform (Figure 1), whereas neither IGF-I (Figure 1A) nor insulin (data not shown) had this effect. IGFBP-1 production and modulation by trophoblast cells IGFBP-1 was not detectable by RIA in medium conditioned by first trimester villous trophoblast explants, differentiating term cytotrophoblasts (24 and 96 h) nor BeWo choriocarcinoma cells, confirming that human trophoblasts do not produce IGFBP-1 (Han et al., 1996). However, in the light of the results described above, we considered other mechanisms by which trophoblast could modulate decidual cell IGFBP-1 in order to increase the presence of npIGFBP-1 in the paracrine environment. Trophoblast and BeWo cells express placental alkaline phosphatase (PLAP) and thus we investigated whether IGFBP-1 phosphorylation status was altered on exposure to trophoblast cells; BeWo cells were cultured (n = 3) with 100 ng/ml highly phosphorylated IGFBP-1; n-OG electrophoresis and Western ligand blotting was used to analyse the conditioned medium. Figure 2 demonstrates that after 24 h, non- and lesser-phosphorylated isoforms appeared. Direct exposure to cells is necessary for de-phosphorylation to occur; incubation of highly phosphorylated IGFBP-1 with BeWo conditioned medium did not result in production of non- and lesser-phosphorylated isoforms (Figure 2). Proteolysis of IGFBP-1 at the maternal–fetal interface Source of IGFBP-1 protease IGFBP proteolysis is a well-recognized mechanism for controlling IGF bioavailability yet little is known about the susceptibility of IGFBP-1 to enzymatic cleavage. Therefore, we investigated if [125I]-IGFBP-1 could be cleaved by proteases in tissues producing IGFBP-1 (decidual cells and a hepatoma cell line, Hep G2) or at its potential sites of action (trophoblast, fibroblasts and adipocytes), or in biological fluids known to contain proteases for other IGFBPs (amniotic fluid, pregnancy and normal serum; Davies et al., 1991; Claussen et al., 1994). Plasmin was used as a positive control (Frost et al., 1993). Figure 3, which is representative of three experiments, shows [125I]-IGFBP-1 running at 30 kDa as well as some smaller molecular weight fragments that are secondary to a degree of peptide destruction upon iodination; storage at –20°C or incubation at 37°C had no effect upon [125I]-IGFBP-1 stability. [125I]-IGFBP-1 was significantly proteolysed by plasmin, as monitored by the reduction in intensity of the 30 kDa band. Proteolysis occurred to a similar degree in the presence of amniotic fluid, medium conditioned by first trimester decidualized endometrial cells or Hep G2 cells. Trophoblast, fibroblasts and adipocytes appear not to produce an IGFBP-1 protease and unlike other IGFBPs, IGFBP-1 was not cleaved when incubated with pregnancy or normal sera. Specificity of IGFBP-1 protease In order to determine whether the phosphorylation status of IGFBP-1 affects its susceptibility to proteolysis, plasma containing either npIGFBP-1 or pIGFBP-1 at 200 μg/l was incubated on three occasions with conditioned medium from decidualized endometrial cells or plasmin at 37°C overnight. The fate of the IGFBP-1 isoforms was then investigated by Western immunoblotting. Figure 4 shows that non-phos- phorylated IGFBP-1 is completely proteolysed by the protease activity in decidual conditioned medium or by plasmin. In contrast, pIGFBP-1 is resistant to both proteases. Characterization of IGFBP-1 protease Substrate zymography and inhibitor studies were performed in order to learn more about the IGFBP-1 protease activity present in amniotic fluid and conditioned medium from decidual cells. Figure 5 shows that decidual conditioned medium contains an IGFBP-1 protease that co-migrates with plasmin. In addition, both decidual medium and amniotic fluid contain high molecular weight proteases of 100–200 kDa. When [125I]-IGFBP-1 was incubated with amniotic fluid or conditioned medium from decidual cells in the presence of inhibitors of aspartate, cysteine, serine or metalloproteases, both activities were sensitive to inhibitors of serine proteases [chymostatin, 4-(2-aminoethyl)benezenesulfonyl fluoride (AEBSF) and phenyl methyl sulphonyl fluoride (PMSF)], while the activity in decidual conditioned medium was also partially inhibited by the metalloprotease inhibitor, EDTA (Figure 6). Neither antipain, bestain, E-64, leupeptin, pepstatin, phosphoramidon, nor aprotinin inhibited the digestion of [125I]-IGFBP-1 in any of the three experiments performed. Consequences of IGFBP proteolysis In order to determine whether proteolysis of IGFBP-1 could have physiological consequences for IGF bioavailability, the protease from decidual cell conditioned medium was allowed to cleave the endogenous IGFBP-1 by incubating the medium overnight at 37°C. IGFBP-1 was then analysed by Western ligand and immunoblotting. The representative (n = 3) immunoblot depicted in Figure 7 shows that the majority of IGFBP-1 in decidua-conditioned medium is cleaved resulting in the appearance of fragments with a molecular weight of ~17 kDa and that this proteolysis can be inhibited with PMSF. The ligand blot shown in Figure 7 demonstrates that when the same blot is probed with a mixture of [125I]-IGF-I and [125I]-IGF-II, these ligands bind intact IGFBP-1 but not the 17 kDa fragments. Discussion This study supports the hypothesis of IGF/IGFBP interactions at the feto–maternal interface. We have shown that decidual IGFBP-1 production and phosphorylation status is regulated by factors known to be in the paracrine environment, resulting in the predominance of an IGFBP-1 isoform which has an increased susceptibility to proteolysis and hence the potential for increased IGF bioavailability. Both IGF-I and IGF-II mRNA, the latter in greater abundance, are present at the feto–maternal interface and are reported to be entirely of fetal origin (Wang et al., 1988; Zhou and Bondy, 1992; Han et al., 1996). We, like others, have found that both these peptides regulate the concentrations of IGFBP-1 produced by decidualized endometrial cells. However, our study has also shown a differential effect on the phosphorylation status of IGFBP-1, with increased concentrations of the non-phosphorylated isoform in the presence of IGF-II but no difference with IGF-I. The signalling pathway by which IGF-II mediates this increase in npIGFBP-1 production by decidual cells has not been investigated, but our data suggests that it does not involve the type1 IGF receptor since IGF-I, which also signals through this receptor did not have a similar effect on decidual IGFBP-1 phosphorylation status. Recent data has demonstrated that IGF-II can signal through the type 2 IGF receptor in endometrial epithelial cells (Badinga et al., 1999) and there is evidence from the IGF knock-out studies to suggest that in placenta, IGF-II signalling is mediated, at least in part, by mechanisms independent of the type 1 IGF receptor (Baker et al., 1993). Whatever the mechanism of IGF-II signalling, it is not clear whether the increase in npIGFBP-1 is caused by modification of intracellular kinase pathways or by increasing extracellular de-phosphorylation. The fact that decidual cells express alkaline phosphatase (Galski et al., 1982) and that highly phosphorylated IGFBP-1 isoform is still detectable in medium conditioned by cells cultured with IGF-II, support the latter hypothesis. We also demonstrated that trophoblast cells can de- phosphorylate the highly phosphorylated variant of IGFBP-1. Trophoblast (Webb et al., 1985) and BeWo (unpublished observations) express placental alkaline phosphatase and it may be via this enzyme, that trophoblast can increase the concentrations of non-phosphorylated IGFBP-1 and thus IGF-I bioavailability at the feto–maternal interface. De-phosphorylation occurred only if the IGFBP-1 was directly exposed to cells. This may simply be because placental alkaline phosphatase (a GPI-anchored protein) is not released into medium by these cells or, more interestingly, it may imply that IGFBP-1 must be localized to the cell surface in order for de-phosphorylation to occur. IGFBP-1 has an RGD site and can interact with the α5β1 integrin present on trophoblast and this potential mechanism for controlling phosphorylation status merits further investigation. Increased IGF-I bioavailability may be necessary for decidua or trophoblast growth and metabolism. IGF-I stimulates nutrient transport across the placenta (Kniss et al., 1994) and recently, it has been demonstrated that this is affected by IGFBP-1 phosphorylation status (Yu et al., 1998); npIGFBP-1 enhanced IGF-I stimulated [3H]-γ-amino isobutyric acid uptake by human trophoblast cells whilst phosphorylated IGFBP-1 inhibited this effect. In addition, we have recently found that in type 1 diabetic pregnancy, elevated phosphorylated IGFBP-1 and an increased ratio of pIGFBP-1 to the non- and lesser-phosphorylated isoforms, are associated with low birthweight (Gibson et al., 1999). We speculate that the capacity of the fetal–maternal unit to produce non-phosphorylated IGFBP-1 may significantly affect fetal nutrition and growth by increasing IGF-I bioavailability. We have also found that changing IGFBP-1 phosphorylation status does not affect its affinity for IGF-II (Westwood et al., 1997), the predominant IGF peptide at the fetal–maternal interface, which led us to look for another mechanism for increasing IGF-II bioavailability. Proteolysis of the other IGFBPs, is an important mechanism for regulating IGF actions, yet apart from a report of an IGFBP-1 protease in amniotic fluid (Binoux et al., 1994), little is known about cellular IGFBP-1 proteolysis. Our study has shown that as well as amniotic fluid, medium conditioned by decidualized endometrial but not trophoblast cells contains a protease capable of cleaving IGFBP-1 into fragments which do not bind to IGF in Western ligand blot analysis. However, we also found that IGFBP-1 is only susceptible to proteolysis when it is in the non-phosphorylated state. Therefore, we propose a model in which trophoblast and decidua act in mutual harmony to direct IGF bioavailability in the local environment. We propose that maternal decidua is stimulated by trophoblast derived IGF-II to preferentially produce non-phosphorylated IGFBP-1 and that trophoblast expression of placental alkaline phosphatase will increase further the preponderance of non-phosphorylated IGFBP-1 by de-phosphorylating circulating pIGFBP-1. We speculate that this will increase IGF-I bioavailability, which, as discussed, may be important for regulating metabolism at the fetal–maternal interface. In addition it will also act as a mechanism for making IGFBP-1 susceptible to proteolysis by an enzyme produced in the maternal decidualized endometrium. We have found that like other IGFBPs, once cleaved, IGFBP-1 can no longer bind IGF and this may therefore be a way of increasing the bioavailability of IGF-II as well as IGF-I. Recently, Manes et al. have shown that in contrast to intact IGFBP-1, IGFBP-1 fragments generated by stromelysin-3 cannot inhibit IGF-induced proliferation in breast adenocarcinoma cells (Manes et al., 1997). We have not assessed the extent of IGFBP-1 proteolysis in vivo, though obviously not all IGBP-1 is cleaved since intact IGFBP-1 can be detected in amniotic fluid and our decidual cell cultures. This poses interesting questions regarding the control of proteolysis and may suggest the presence of inhibitors. Alternatively, the ratio of intact to fragmented IGFBP-1 may depend on the balance between IGFBP-1 production/transport and proteolysis. Nonetheless, since IGFBP-1 is the predominant IGF binding species at the maternal–fetal interface, even small changes to the proportion of proteolysed IGFBP-1 could have significant consequences for IGF bioavailability. IGFBP proteolysis may be a general feature of pregnancy, since IGFBP-3 is proteolysed in the maternal circulation, presumably to increase IGF activity in order to meet the increased metabolic demands placed on the mother, and the proportion of IGFBP-1 in the non-phosphorylated form in the maternal circulation could contribute to this phenomenon. However, there are clearly different enzymes involved in the protolysis of the two binding proteins, since in our study, pregnancy plasma did not cleave [125I]-IGFBP-1 and we could not detect the presence of IGFBP-1 fragments in the maternal circulation. Furthermore, other authors (Irwin et al., 2000) have recently reported a trophoblast-derived IGFBP-3 protease which is inactive towards IGFBP-1 and it has been suggested that individual enzymes are specific for each of the binding proteins (Maile and Holly, 1999). Additionally, proteolysis of IGFBP-1 may provide a mechanism for regulating some IGF-independent action of IGFBP-1. Non-phosphorylated IGFBP-1 has been shown to increase CHO cell migration via its RGD interaction with α5β1 integrin (Jones et al., 1993), though whether or not this affect extends to the migration of trophoblast is controversial. IGFBP-1 has been reported to stimulate human trophoblast cell migration and invasion, both independently and by enhancing IGF-II actions (Irving and Lala, 1995; Hamilton et al., 1998). If true, then proteolysis of IGFBP-1 may be necessary for curbing trophoblast invasion of the decidualized endometrium, especially if association with integrin is necessary for de-phosphorylation and hence proteolysis to occur. However, data from a co-culture model of human cytotrophoblast and decidualized human endometrial stromal cells (Irwin and Giudice, 1998) suggest that IGFBP-1 acts as a barrier to trophoblast migration since cytotrophoblasts did not penetrate the decidual multilayer in the presence of IGFBP-1 (Irwin et al., 1994). Under such circumstances, proteolysis of IGFBP-1 via a sequence of events initiated by production of IGF-II may be necessary to overcome maternal restraint. It is possible that resistance to invasion occurs when trophoblast reaches the myometrium, but we have not investigated proteolysis of IGFBP-1 at this site. Whilst we speculate that de-phosphorylation followed by proteolysis of IGFBP-1 might offer two `tiers' of control over IGF bioavailability, it is also possible that de-phosphorylation of IGFBP-1 serves merely as a switch for susceptibility to proteolysis. IGFBP post-translational modifications are known to affect sensitivity to proteases, for example the non-glycosylated variant of IGFBP-3 is degraded more readily than the glycosylated isoform (Kubler et al., 1998) and IGFBP-5 is protected from proteolysis when bound to the extracellular matrix (Arai et al., 1994). This study has partially characterized the protease(s) present in amniotic fluid and conditioned medium from decidualized endometrial cells. It is not clear whether the two sources contain the same protease, though we have found them to have similar molecular weight and to be blocked by inhibitors of serine proteases and, to a degree, metalloproteases. Furthermore, Manes et al. (1997) have shown that MMP-11 (stromelysin-3) can cleave IGFBP-1 in vitro. This enzyme is present at the maternal–fetal interface (Maquoi et al., 1997), though studies to identify the relevant proteases for IGFBP-1 action at this site are in progress. In summary we have shown that decidualized endometrial cells produce both highly phosphorylated and non-phosphorylated IGFBP-1. Both IGF-I and IGF-II regulate IGFBP-1 production but only IGF-II can influence phosphorylation status by preferentially increasing npIGFBP-1 concentrations. NpIGFBP-1 is susceptible to a protease produced by decidualized endometrium and this results in IGFBP-1 fragments which do not bind IGF. We therefore propose a paracrine interaction in which trophoblast-directed modulation of IGFBP-1 regulates actions of the IGF axis at the feto–maternal interface. Figure 1. View large Download slide View large Download slide (A) Representative characterization (n = 3) of insulin-like growth factor binding protein (IGFBP)-1 isoforms present in medium conditioned by human first trimester decidualized endometrial cells cultured with increasing concentrations of IGF-I or IGF-II for 7 days. Samples were immunoprecipitated with monoclonal antibody 6303 and subjected to n-octyl glucoside (n-OG) electrophoresis and Western ligand blotting with [125I]-IGF-I. The phosphorylation pattern of human recombinant IGFBP-1 and IGFBP-1 from normal plasma is shown for comparison. (B) Mean (± SEM) concentrations of IGFBP-1 secreted by first trimester human decidualized endometrial cells in response to 7 days treatment with 0–500 ng/ml IGF-I (▪), IGF-II (□) or insulin ( ). Figure 1. View large Download slide View large Download slide (A) Representative characterization (n = 3) of insulin-like growth factor binding protein (IGFBP)-1 isoforms present in medium conditioned by human first trimester decidualized endometrial cells cultured with increasing concentrations of IGF-I or IGF-II for 7 days. Samples were immunoprecipitated with monoclonal antibody 6303 and subjected to n-octyl glucoside (n-OG) electrophoresis and Western ligand blotting with [125I]-IGF-I. The phosphorylation pattern of human recombinant IGFBP-1 and IGFBP-1 from normal plasma is shown for comparison. (B) Mean (± SEM) concentrations of IGFBP-1 secreted by first trimester human decidualized endometrial cells in response to 7 days treatment with 0–500 ng/ml IGF-I (▪), IGF-II (□) or insulin ( ). Figure 2. View largeDownload slide Characterization of insulin-like growth factor binding protein (IGFBP)-1 isoforms produced in response to incubating highly-phosphorylated IGFBP-1 with BeWo cells or BeWo-conditioned medium. Samples were immunoprecipitated with monoclonal antibody 6303 and subjected to n-octyl glucoside (n-OG) electrophoresis and Western ligand blotting with [125I]-IGF-I. Non-phosphorylated IGFBP-1 is only observed when pIGFBP-1 is incubated with cells; incubation with BeWo-conditioned medium does not result in de-phosphorylation, suggesting the involvement of a cell membrane associated phosphatase. Lane 1 = migratory pattern of non-phosphorylated (human recombinant) IGFBP-1. Lanes 2–4 = IGFBP-1 isoforms present in amniotic fluid, normal human plasma and conditioned medium from human decidual cells. Lane 5 = BeWo cells do not produce any imunoreactive IGFBP-1. This blot is representative of the results obtained from three experiments. Figure 2. View largeDownload slide Characterization of insulin-like growth factor binding protein (IGFBP)-1 isoforms produced in response to incubating highly-phosphorylated IGFBP-1 with BeWo cells or BeWo-conditioned medium. Samples were immunoprecipitated with monoclonal antibody 6303 and subjected to n-octyl glucoside (n-OG) electrophoresis and Western ligand blotting with [125I]-IGF-I. Non-phosphorylated IGFBP-1 is only observed when pIGFBP-1 is incubated with cells; incubation with BeWo-conditioned medium does not result in de-phosphorylation, suggesting the involvement of a cell membrane associated phosphatase. Lane 1 = migratory pattern of non-phosphorylated (human recombinant) IGFBP-1. Lanes 2–4 = IGFBP-1 isoforms present in amniotic fluid, normal human plasma and conditioned medium from human decidual cells. Lane 5 = BeWo cells do not produce any imunoreactive IGFBP-1. This blot is representative of the results obtained from three experiments. Figure 3. View largeDownload slide Proteolysis of [125I]-np insulin-like growth factor binding protein (IGFBP)-1. Radiolabelled npIGFBP-1 was incubated with plasmin (lanes 3 and 4), serum from non-pregnant (lane 5) or pregnant (lane 6) women, medium conditioned by 3T3-L1 adipocytes (lane 7), fibroblasts (lane 8), primary trophoblast (lane 9), BeWo choriocarcinoma cells (lane 10), decidual cells (lane 11) and HepG2 cells (lane 12) or amniotic fluid (lane 13) overnight at 37°C. Samples were then analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography. Intact IGFBP-1, running at ~30 kDa is indicated by the arrow. [125I]-IGFBP-1 stored at −20°C (lane 1) and incubated alone at 37°C (lane 2) were included as controls. Similar results were achieved in two other experiments. Figure 3. View largeDownload slide Proteolysis of [125I]-np insulin-like growth factor binding protein (IGFBP)-1. Radiolabelled npIGFBP-1 was incubated with plasmin (lanes 3 and 4), serum from non-pregnant (lane 5) or pregnant (lane 6) women, medium conditioned by 3T3-L1 adipocytes (lane 7), fibroblasts (lane 8), primary trophoblast (lane 9), BeWo choriocarcinoma cells (lane 10), decidual cells (lane 11) and HepG2 cells (lane 12) or amniotic fluid (lane 13) overnight at 37°C. Samples were then analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography. Intact IGFBP-1, running at ~30 kDa is indicated by the arrow. [125I]-IGFBP-1 stored at −20°C (lane 1) and incubated alone at 37°C (lane 2) were included as controls. Similar results were achieved in two other experiments. Figure 4. View largeDownload slide The effect of phosphorylation status on insulin-like growth factor binding protein (IGFBP)-1 susceptibility to proteolysis. Phosphorylated or non-phosphorylated IGFBP-1 was `spiked' into decidual cell conditioned medium (stripped of endogenous IGFBP-1) or plasmin (0.025 IU) at 200 ng/ml, incubated at 37°C overnight and analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and Western immunoblotting. Non-phosphorylated IGFBP-1 was cleaved by both sources of proteolytic activity whereas the phosphorylated isoform was resistant to proteolysis. This blot is shown as a representative of three experiments. Figure 4. View largeDownload slide The effect of phosphorylation status on insulin-like growth factor binding protein (IGFBP)-1 susceptibility to proteolysis. Phosphorylated or non-phosphorylated IGFBP-1 was `spiked' into decidual cell conditioned medium (stripped of endogenous IGFBP-1) or plasmin (0.025 IU) at 200 ng/ml, incubated at 37°C overnight and analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and Western immunoblotting. Non-phosphorylated IGFBP-1 was cleaved by both sources of proteolytic activity whereas the phosphorylated isoform was resistant to proteolysis. This blot is shown as a representative of three experiments. Figure 5. View largeDownload slide Insulin-like growth factor binding protein (IGFBP)-1 zymogram demonstrating the molecular weight of plasmin and the proteolytic activity in amniotic fluid and conditioned medium from decidualized endometrial cells. This is a representative zymogram from three separate experiments. Figure 5. View largeDownload slide Insulin-like growth factor binding protein (IGFBP)-1 zymogram demonstrating the molecular weight of plasmin and the proteolytic activity in amniotic fluid and conditioned medium from decidualized endometrial cells. This is a representative zymogram from three separate experiments. Figure 6. View large Download slide View large Download slide Inhibition of [125I]- insulin-like growth factor binding protein (IGFBP)-1 proteolytic activity. Radiolabelled non-phosphorylated IGFBP-1 was incubated on three occasions with (A) amniotic fluid or (B) conditioned medium from decidualized endometrial cells at 37°C overnight in the absence or presence of a range of protease inhibitors: lane 1 = 50 μg/ml antipain; lane 2 = 40 μg/ml bestatin; lane 3 = 50 μg/ml chymostatin; lane 4 = 10 μg/ml E-64; lane 5 = 0.5 μg/ml leupeptin; lane 6 = 0.7 μg/ml pepstatin; lane 7 = 200 μg/ml phosphoramidon; lane 8 = 1 mg/ml AEBSF; lane 9 = 0.5mg/ml EDTA; lane 10 = 2 μg/ml aprotinin; lane 11 =10 mmol/l phenyl methyl sulphonyl fluoride (PMSF). Samples were then analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography. Intact IGFBP-1, running at ~30 kDa is indicated by the arrow. [125I]-IGFBP-1 stored at –20°C (lane 12) and incubated alone at 37°C (lane 13) were included as controls. Representative autoradiographs are presented. Figure 6. View large Download slide View large Download slide Inhibition of [125I]- insulin-like growth factor binding protein (IGFBP)-1 proteolytic activity. Radiolabelled non-phosphorylated IGFBP-1 was incubated on three occasions with (A) amniotic fluid or (B) conditioned medium from decidualized endometrial cells at 37°C overnight in the absence or presence of a range of protease inhibitors: lane 1 = 50 μg/ml antipain; lane 2 = 40 μg/ml bestatin; lane 3 = 50 μg/ml chymostatin; lane 4 = 10 μg/ml E-64; lane 5 = 0.5 μg/ml leupeptin; lane 6 = 0.7 μg/ml pepstatin; lane 7 = 200 μg/ml phosphoramidon; lane 8 = 1 mg/ml AEBSF; lane 9 = 0.5mg/ml EDTA; lane 10 = 2 μg/ml aprotinin; lane 11 =10 mmol/l phenyl methyl sulphonyl fluoride (PMSF). Samples were then analysed by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS–PAGE) and autoradiography. Intact IGFBP-1, running at ~30 kDa is indicated by the arrow. [125I]-IGFBP-1 stored at –20°C (lane 12) and incubated alone at 37°C (lane 13) were included as controls. Representative autoradiographs are presented. Figure 7. View largeDownload slide Effect of proteolysis on insulin-like growth factor binding protein (IGFBP)-1 ability to bind ligand (representative of three experiments). IGFBP-1 fragments (≈17 kDa) generated by incubating conditioned medium from decidualized endometrial cells at 37°C overnight are detected by Western immunoblotting but do not bind a mixture of radiolabelled IGF-I and -II. Intact IGFBP-1 (30 kDa) can be visualized by both blotting techniques. IGFBP-1 proteolysis can be inhibited by the addition of 10 mmol/l phenyl methyl sulphonyl fluoride (PMSF). Figure 7. View largeDownload slide Effect of proteolysis on insulin-like growth factor binding protein (IGFBP)-1 ability to bind ligand (representative of three experiments). IGFBP-1 fragments (≈17 kDa) generated by incubating conditioned medium from decidualized endometrial cells at 37°C overnight are detected by Western immunoblotting but do not bind a mixture of radiolabelled IGF-I and -II. Intact IGFBP-1 (30 kDa) can be visualized by both blotting techniques. IGFBP-1 proteolysis can be inhibited by the addition of 10 mmol/l phenyl methyl sulphonyl fluoride (PMSF). 5 To whom correspondence should be addressed at: Endocrine Sciences Research Group, University of Manchester, Oxford Road, Manchester, M13 9PT, UK. E-mail: melissa.westwood@man.ac.uk The authors gratefully acknowledge the financial support of Salford Royal Hospitals NHS Trust, Wellbeing and The Royal Society and thank Ms Justine Allen for her expert technical assistance. References Arai, T., Arai, A., Busby-WH, J. et al. ( 1994) Glycosaminoglycans inhibit degradation of insulin-like growth factor binding protein-5. Endocrinology , 135, 2358–2363. Google Scholar Badinga, L., Song, S., Simmen, R.C.M. et al. ( 1999) Complex mediation of uterine endometrial epithelial cell growth by insulin-like growth factor-II (IGF-II) and IGF binding protein-2. J. Mol. Endocrinol. , 23, 277–285. Google Scholar Baker, J., Liu, J.-P., Robertson, E. et al. ( 1993) Role of insulin-like growth factors in embryonic and postnatal growth. Cell , 75, 73–82. Google Scholar Binoux, M., Lalou, C., Lassarre, C. et al. (1994) Regulation of IGF bioavailability by IGFBP proteases. In Baxter, R.C., Gluckman, P.D., and Rosenfeld, R.G. (eds), The Insulin-Like Growth Factors and Their Regulatory Proteins. Elsevier Science B.V., Amsterdam, The Netherlands, 217 pp. Google Scholar Claussen, M., Zapf, J. and Braulke, T. ( 1994) Proteolysis of insulin-like growth factor binding protein-5 by pregnancy serum and amniotic fluid. Endocrinology , 134, 1964–1966. Google Scholar Davies, S.C., Holly, J.M., Coulson, V.J. et al. ( 1991) The presence of cation-dependent proteases for insulin-like growth factor binding proteins does not alter the size distribution of insulin-like growth factors in pregnancy. Clin. Endocrinol. , 34, 501–506. Google Scholar Drop, S.L., Kortleve, D.J., Guyda, H.J. et al. ( 1984) Immunoassay of a somatomedin-binding protein from human amniotic fluid: levels in fetal, neonatal, and adult sera. J. Clin. Endocrinol. Metab. , 59, 908–915. Google Scholar Frost, V.J., Macaulay, V.M., Wass, J.A.H. et al. ( 1993) Proteolytic modification of insulin-like growth factor binding proteins: comparison of conditioned media from human cell lines, circulating proteases and characterized enzymes. J. Endocrinol. , 138, 454–554. Google Scholar Galski, H., Weinstein, D., Abraham, K. et al. ( 1982) The in vitro synthesis and secretion of alkaline phosphatase from first trimester human decidua. J. Obstet. Gynecol. Reprod. Biol. , 14, 1–11. Google Scholar Gibson, J.M., Westwood, M., Lauszus, F.F. et al. ( 1999) Phosphorylated insulin-like growth factor binding protein-1 is increased in pregnant diabetic subjects. Diabetes , 48, 321–326. Google Scholar Giudice, L.C., de-Zegher, F., Gargosky, S.E. et al. ( 1995) Insulin-like growth factors and their binding proteins in the term and preterm human fetus and neonate with normal and extremes of intrauterine growth. J. Clin. Endocrinol. Metab. , 80, 1548–1555. Google Scholar Giudice, L.C., Martina, N.A., Crystal, R.A. et al. ( 1997) Insulin-like growth factor binding protein-1 at the maternal–fetal interface and insulin-like growth factor-I, insulin-like growth factor-II, and insulin-like growth factor binding protein-1 in the circulation of women with severe preeclampsia. Am. J. Obstet. Gynecol. , 176, 751–757. Google Scholar Hamilton, G.S., Lysiak, J.J., Han, V.K. et al. ( 1998) Autocrine-paracrine regulation of human trophoblast invasiveness by insulin-like growth factor (IGF)-II and IGF-binding protein (IGFBP)-1. Exp. Cell Res. , 244, 147–156. Google Scholar Han, V.K. (1993) Growth factors in placental growth and development. In Rice, G.E. and Brennecke, S.P. (eds), Molecular Aspects of Placental and Fetal Membrane Autocoids. CRC Press, Boca Raton, FL, USA, 395 pp. Google Scholar Han, V.K., Bassett, N., Walton, J. et al. ( 1996) The expression of insulin-like growth factor (IGF) and IGF-binding protein (IGFBP) genes in the human placenta and membranes: evidence for IGF-IGFBP interactions at the feto–maternal interface. J. Clin. Endocrinol. Metab. , 81, 2680–2693. Google Scholar Howell, R.J.S., Perry, L.A., Choglay, N.S. et al. ( 1985) Placental protein 12 (pp12): a new test for the prediction of the small-for-gestational age infant. Br. J. Obstet. Gynaecol. , 92, 1141–1144. Google Scholar Irving, J.A. and Lala, P.K. ( 1995) Functional role of cell surface integrins on human trophoblast cell migration: regulation by TGF-beta, IGF-II, and IGFBP-1. Exp. Cell Res. , 217, 419–427. Google Scholar Irwin, J.C., de-las Fuentes L. and Giudice, L.C. ( 1994) Growth factors and decidualization in vitro. Ann. N.Y. Acad. Sci. , 734, 7–18. Google Scholar Irwin, J.C., de-las Fuentes, L., Dsupin, B.A. et al. ( 1993) Insulin-like growth factor regulation of human endometrial stromal cell function: coordinate effects on insulin-like growth factor binding protein-1, cell proliferation and prolactin secretion. Regul. Pept. , 48, 165–177. Google Scholar Irwin, J.C. and Giudice, L.C. ( 1998) Insulin-like growth factor binding protein-1 binds to placental cytotrophoblast α5β1 integrin and inhibits cytotrophoblast invasion into decidualised endometrial stromal cultures. GH & IGF Res. , 8, 21–31. Google Scholar Irwin, J.C., Suen, L.-F., Cheng, B.-H. et al. ( 2000) Human placental trophoblasts secrete a disintegrin metalloproteinase very similar to the insulin-like growth factor binding protein-3 protease in human pregnancy serum. Endocrinology , 141, 666–674. Google Scholar Jones, J.I., D'Ercole, A.J., Camacho, H.C. et al. ( 1991) Phosphorylation of insulin-like growth factor (IGF)-binding protein 1 in cell culture and in vivo: effects on affinity for IGF-I. Proc. Natl Acad. Sci. USA , 88, 7481–7485. Google Scholar Jones, J.I., Gockerman, A., Busby-WH, J. et al. ( 1993) Insulin-like growth factor binding protein 1 stimulates cell migration and binds to the alpha 5 beta 1 integrin by means of its Arg-Gly-Asp sequence. Proc. Natl Acad. Sci. USA , 90, 10553–10557. Google Scholar Kniss, D.A., Shubert, P.J., Zimmerman, P.D. et al. ( 1994) Insulin-like growth factors. Their regulation of glucose and amino acid transport in placental trophoblasts isolated from first-trimester chorionic villi. J. Reprod. Med. , 39, 249–256. Google Scholar Kubler, N., Cowell, S., Zapf, J. et al. ( 1998) proteolysis of insulin-like growth factor binding proteins by a novel 50 kilodalton metalloproteinase in human pregnancy serum. Endocrinology , 139, 1556–1563. Google Scholar Lamson, G., Giudice, L.C. and Rosenfeld, R.G. ( 1991) A simple assay for proteolysis of IGFBP-3. J. Clin. Endocrinol. Metab. , 72, 1391–1393. Google Scholar Liu, J.P., Baker, J., Perkins, A.S. et al. ( 1993) Mice carrying null mutations of the genes encoding insulin-like growth factor I (Igf-1) and type 1 IGF receptor (Igf1r). Cell , 75, 59–72. Google Scholar Manes, S., Mira, E., Barbacid, M.M. et al. ( 1997) Identification of insulin-like growth factor-binding protein-1 as a potential physiological substrate for human stromelysin-3. J. Biol. Chem. , 272, 25706–25712. Google Scholar Maile, L.A. and Holly, J.M.P. ( 1999) Insulin-like growth factor binding protein (IGFBP) proteolysis: occurance, identification, role and regulation. GH & IGF Res. , 9, 85–95. Google Scholar Maquoi, E., Polette, M., Nawrocki, B. et al. ( 1997) Expression of stromelysin-3 in the human placenta and placental bed. Placenta , 18, 277–285. Google Scholar Powell, B.L., Hollingshead, P., Warburton, C. et al. ( 1993) IGF-I is required for normal embryonic growth in mice. Genes Dev ., 7, 2609–2617. Google Scholar Thrailkill, K.M., Clemmons, D.R., Busby, W.-H.J. et al. ( 1990) Differential regulation of insulin-like growth factor binding protein secretion from human decidual cells by IGF-I, insulin and relaxin. J. Clin. Invest. , 86, 878–883. Google Scholar Verhaeghe, J., Van Bree, R., Van Herck, E. et al. ( 1993) C-peptide, insulin-like growth factors I and II, and insulin-like growth factor binding protein-1 in umbilical cord serum: correlations with birth weight. Am. J. Obstet. Gynecol. , 169, 89–97. Google Scholar Vicovac, L.M., Starkey, P.M. and Aplin, J.D. ( 1994) Comment: effect of cytokines on prolactin production by human decidual stromal cells in culture: studies using cells freed of bone marrow-derived contaminants. J. Clin. Endocrinol. Metab. , 79, 1877–1882. Google Scholar Waites, G.T., James, R.F. and Bell, S.C. ( 1988) Immunohistological localization of the human endometrial secretory protein pregnancy-associated endometrial alpha 1-globulin, an insulin-like growth factor-binding protein, during the menstrual cycle. J. Clin. Endocrinol. Metab. , 67, 1100–1104. Google Scholar Wang, C.-Y., Daimon, M., Shen, S.-J. et al. ( 1988) Insulin-like growth factor-I messenger ribnucleic acid in the developing human placenta and in term placenta of diabetics. Mol. Endocrinol. , 2, 217–229. Google Scholar Webb, P.D., Evans, P.W., Molloy, C.M. et al. ( 1985) Biochemical studies of human placental microvillous plasma membrane proteins. Am. J. Reprod. Immunol. Microbiol. , 8, 113–119. Google Scholar Westwood, M., Gibson, J.M., Davies, A.J. et al. ( 1994) The phosphorylation pattern of insulin-like growth factor-binding protein-1 in normal plasma is different from that in amniotic fluid and changes during pregnancy. J. Clin. Endocrinol. Metab. , 79, 1735–1741. Google Scholar Westwood, M., Gibson, J.M., Pennells, L.A. et al. ( 1999) Plasma insulin-like growth factors and binding proteins are modified during oral contraceptive usage and the normal menstrual cycle. Am.J. Obstet. Gynecol. , 180, 530–536. Google Scholar Westwood, M., Gibson, J.M. and White, A. ( 1997) Purification and characterization of the insulin-like growth factor-binding protein-1 phosphoform found in normal plasma. Endocrinology,  138, 1130–1136. Google Scholar Yu, J., Iwashita, M., Kudo, Y. et al. ( 1998) Phosphorylated insulin-like growth factor (IGF)-binding protein-1 (IGFBP-1) inhibits while non-phosphorylated IGFBP-1 stimulated IGF-I induced amino acid uptake by cultured trophoblast cells. GH & IGF Res. , 8, 65–70. Google Scholar Zhou, J. and Bondy, C.A. ( 1992) Insulin-like growth factor-II and its binding proteins in placental development. Endocrinology , 131, 1230–1240. Google Scholar © European Society of Human Reproduction and Embryology
TI - Regulation of IGF bioavailability in pregnancy
JF - Molecular Human Reproduction
DO - 10.1093/molehr/7.1.79
DA - 2001-01-01
UR - https://www.deepdyve.com/lp/oxford-university-press/regulation-of-igf-bioavailability-in-pregnancy-8SwxJh8bf4
SP - 79
EP - 87
VL - 7
IS - 1
DP - DeepDyve
ER -