| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a INSERM U.361, Maternité Port-Royal Cochin, Université René Descartes, 75014 Paris, France
b Laboratoire de Biochimie Endocrinienne, Faculté de Médecine Xavier Bichat, Université Denis Diderot, 75870 Paris, France
ABSTRACT
The status of the corticosteroid-binding globulin (CBG) at the fetomaternal interface, especially in the maternal intervillous blood space (I), was investigated and compared to that of CBG in the maternal (M) and fetal (umbilical arteries [A] and vein [V]) peripheral circulations at term. Immunoquantitation of plasma CBG showed that the CBG concentration in I was 30% less than that in M (P < 0.001) and threefold higher than that in umbilical cord blood (P < 0.001). The microheterogeneity of CBG studied by immunoaffinoelectrophoresis in the presence of concanavalin A and Western blotting indicated that the CBG in I was mainly of maternal origin and different from fetal CBG. A CBG mRNA, but no classic 50- to 59-kDa CBG, was found in isolated term trophoblastic cells. The steroid environment of the CBG in I differed greatly from that in the peripheral maternal and fetal circulations, because the progesterone:cortisol molar ratio in I was 75-fold higher than that in M and 7- to 10-fold higher than that in the fetal circulation. Binding studies revealed that the affinity constants of CBG for cortisol in I, A, and V were significantly lower than that in M plasma (P < 0.02) in their respective hormonal contexts. The binding parameters for I-CBG stripped of endogenous steroids and lipids were close to those for M-CBG but different from those of fetal CBG (P < 0.001). These data reflect the physiological relevance of the CBG-steroid interaction, especially with very CBG-loaded progesterone at the fetomaternal interface during late pregnancy.
cortisol, hormone action, parturition, placenta, placental transport, pregnancy, progesterone, steroid hormones
INTRODUCTION
The maternal intervillous blood space is a vital interface between the mother and the fetus from the first trimester of pregnancy until term [1]. At term, this space occupies 25%–40% of the hemomonochorial human placenta and bathes the chorial villi. It is delineated by a fetal epithelium, the syncytiotrophoblast, and is in close proximity to the maternal decidua via the placental basal plate. Almost all transfers between the maternal and fetal bloodstreams take place in this area [2]. The composition of the intervillous blood is still not completely known, but recent studies from our laboratory have shown that it contains large amounts of signaling factors of various origins [3]. Some, such as steroid hormones, endothelins, and eicosanoids, are mainly produced by the trophoblast [3], whereas others, such as polyunsaturated fatty acids (w6/w3), are of maternal nutritional origin [4].
All these signaling factors, especially the steroid hormones, have pleiotropic actions and play major roles in the chain of events by which a new human being enters the world. Thus, cortisol and progesterone are not only involved in development of the fetoplacental unit but also in adapting the maternal body to gestation and parturition. These hormones are mainly bound to a specific plasma binding protein, corticosteroid-binding globulin (CBG), whose concentration in the maternal peripheral plasma increases dramatically during gestation [5–11]. This protein has been thoroughly studied, and several excellent reviews on structural gene mapping of the CBG, regulation of CBG expression, and its physicochemical properties and physiological functions are available [7–9, 12, 13]. The primary roles of CBG are to protect circulating steroid hormones from catabolism and to ensure their transport to maternal and fetal target cells (i.e., uterus, lung, brain, kidneys, etc.), but the steroid-CBG complex also interacts with specific plasma membrane receptors, especially those of the human decidua [9, 11], and syncytiotrophoblast [14]. Such interactions lead to the activation of membrane signal transduction pathways and the production of a second messenger (i.e., cAMP), internalization of the steroid-CBG complex, or cleavage of the binding protein by proteases, with the local release of steroids [8–11, 15, 16]. Thus, CBG seems to play a subtle role in modulating steroid signals, especially at the fetomaternal interface.
Little information, however, is available regarding the CBG status (i.e., concentration, molecular heterogeneity, binding properties, environmental ligands) in the maternal intervillous blood during pregnancy. Therefore, we explored this pivotal compartment to check the relevance of the CBG-steroid interactions at the fetomaternal interface and to determine whether both maternal and fetal CBGs (from hepatic or trophoblastic source) are present in the maternal intervillous circulation. We examined women undergoing elective cesarean section at term, without labor, and measured the CBG concentration and spectrum of CBG isoforms at the fetomaternal interface. We then compared these profiles with those in the peripheral maternal and fetal circulations (umbilical arteries and vein) and with that in the plasma of nonpregnant women. We also measured the binding properties of CBG before and after removal of endogenous ligands to assess the influence of the steroid and lipid environments in each compartment on the CBG behavior. Finally, we used reverse transcriptase-polymerase chain reaction (RT-PCR) and Western blotting to test the capacity of trophoblastic cells from human term placenta to produce CBG mRNA and protein [17].
MATERIALS AND METHODS
Samples
This study was approved by the Consultative Committee for the Protection of Persons in Biomedical Research, Paris Cochin, France. Samples were collected from 22 women with uncomplicated pregnancies (38- to 40-wk amenorrhea) during elective cesarean section. The cesarean sections, which were carried out before the onset of labor, were performed because of a diagnosed cephalopelvic disproportion. The women were in good health, normotensive, and not on any medication (Table 1). Samples were also taken from seven healthy, nonpregnant women volunteers aged 25–30 yr who were nondrinkers, nonsmokers, and not on any medication. Informed written consent was obtained from all subjects.
|
Blood Cord blood samples were taken during delivery after clamping the cord, with the placenta still attached to the uterus. Maternal peripheral blood at the time of delivery was taken from the antecubital vein, and blood in the intervillous space was collected from the basal plate of the placenta. Blood samples from nonpregnant women were collected in the morning (0900 h). Blood was collected into tubes containing EDTA, then placed in an ice bath. Plasma was separated from the red blood cells by centrifugation (2000 rpm for 20 min at 4°C) within 30 min of the sample being taken and then stored at -80°C.
Placenta Placentas were aseptically obtained immediately after elective cesarean section from healthy mothers not on medication and in the 38th to 40th wk of amenorrhea.
Isolation of trophoblasts Cytotrophoblasts were isolated from placenta villous tissue by trypsin digestion and centrifugation through a continuous Percoll gradient (5%–70%) [18].
Purification of trophoblastic cells by negative selection Antibodies directed against contaminant membrane cell antigens (CD3, CD14, CD9, CD31, CD45; Dako, Versailles, France) were used together with magnetic Dynabeads (Dynal, Compiegne, France) coated with sheep anti-mouse immunoglobulin (Ig) G according to the manufacturer's instructions. Briefly, isolated trophoblasts were incubated with mouse anti-human CD3, CD14, and CD9 monoclonal antibodies at 4°C for 20 min. The cells (100 x 106 cells/ml) were washed two times in RPMI 1640/2% v/v fetal calf serum (FCS) and incubated with Dynabeads coated with anti-mouse antibody (one bead per two cells) for 15 min at 4°C. The supernatant containing unadsorbed cells was collected and centrifuged at 1000 x g, and the pelleted cells were suspended in medium. The cells (20 x 106 cells/ml) were then treated a second time with anti-CD3, anti-CD14, anti-CD9, anti-CD31, and anti-CD45 antibodies. Part of the unadhering cells, the purified trophoblastic cells, were directly used for RNA extraction and protein analysis. The remainder was seeded in tissue culture plates at approximately 2 x 106 cells/ml medium, in which they were maintained and renewed daily. The medium was Ham-F10 nutrient mixture with glutamine (Gibco Biotech Life Technology, Cergy Pontoise, France) containing 25 mmol/L Hepes, 100 µg/ml of streptomycin, 100 IU/ml of penicillin, and 15% FCS. Cells were incubated in a humidified atmosphere of 95% air/5% CO2 at 37°C for 2 days, by which time most of the cells had become organized into multinucleate syncytiotrophoblasts. The supernatant of the culture was collected, and the cells were scraped off and frozen in Trizol reagent (Gibco Biotech) for RNA extraction.
Immunoquantifications of Human Albumin and Corticosteroid-Binding Protein
Human plasma albumin and CBG were quantified by electroimmunodiffusion [19] using polyclonal anti-human albumin or anti-human CBG antibodies (Tebu; Nordic Immunological Laboratories, Tilburg, Denmark). The limit of CBG detection was 60 nM.
Western Blotting of CBG in Plasma and Trophoblastic Cells
Plasma samples and cytosol from isolated trophoblastic cells were subjected to SDS-PAGE (10%). Proteins adsorbed onto the gel were electrotransferred onto a nitrocellulose membrane and further immunodetected using a rabbit anti-human CBG antibody (Tebu, 1:500 v:v; Nordic Immunological Laboratories). Sites of primary antibody binding were also visualized with horseradish peroxidase-conjugated goat anti-rabbit IgG (1:5000 v:v; Amersham International, Buckinghamshire, UK). The immunoreactive bands was detected using the ECL blotting system (Amersham).
Characterization of CBG Glycan Isoforms
Crossed-immunoelectrophoresis was performed as previously described [20]. The CBG-lectin affinity was determined by affinoimmunoelectrophoresis. The plasma samples were separated in the first dimension by electrophoresis in a gel (1.2-mm deep, 10 g/L of agarose gel, in Tris-barbital-lactic acid electrode buffer [pH 8.6] at 14°C, 8 V/cm, 120 min) containing 0.53 g/L of concanavalin A (ConA; Sigma, Saint Quentin Fallavier, France) [4]. The reference samples were run in parallel lanes without additives. The second-dimension electrophoresis was at right angles to the first separation in the same buffer (14°C, 2 V/cm, 16 h) in agarose gel (10 x 10 cm) containing 70 g/L of 
-D-methyl-glucopyranoside and 0.3% polyclonal anti-human CBG antibody. The gels were dried and stained with Coomassie brilliant blue.
Immunoassay of Total Cortisol and Progesterone
Cortisol Plasma samples were assayed by fluorimetric polarization enzyme-linked immunoassay using a kit (Abbott, Rungis, France) and a TDX apparatus (Rungis, France) according to the manufacturer's recommendations. The anti-cortisol antibody used was 100% specific for cortisol, and the cross-reactivity with other steroids was less than 0.03%. The detection limit was 3.6 ng/ml (12 nM). Coefficients of intra- and interassay variations were 8.5% and 10% respectively.
Progesterone Samples (0.5 ml) were extracted three times with ethyl ether plus 1000 cpm of 1,2,6,7-3H-progesterone (95 Ci/mmol; Amersham) as an internal standard. The extracts were dried, taken up in human male plasma, and assayed using the 125I-progesterone COTRIA Kit (Biomerieux, Rungis, France) according to the manufacturer's recommendations. The detection limit was 0.05 ng/ml, and the intra- and interassay coefficients of variation were 6% and 8% respectively. All plasma samples were run at least in duplicate.
Gas Chromatographic-Mass Spectrometric Analysis of Progesterone
The progesterone immunoassay was validated with a reference method using isotope dilution-mass spectrometry and gas chromatography [21]. Plasma samples (50 µl) were diluted with 150 µl of water, spiked with 50 ng of 1,2-2H-progesterone, and left to equilibrate for 1 h at room temperature. The progesterone was then extracted with 1.5 ml of n-hexane. The organic phase was evaporated off, and the dry residue was derivatized with 50 µl of silylation reagent (N-methyl-N-trimethylsilyl-trifluoracetamide containing 0.3% trimethylsilyl iodide) at 70°C for 30 min. An aliquot of the resulting mixture (1 µl) containing progesterone trimethylsilyl enolether derivatives was injected onto a capillary column (25 m x 0.25 mm, coated with 5% phenyl-methylsilicone) directly coupled to the ion source of a quadropole mass spectrometer (Model 5970; Hewlett-Packard, Boulogne-Billancourt, France). Ion masses of 458 and 460 m/z, corresponding to those of derivatized progesterone, were recorded, and concentrations were calculated [21]. The accuracy error was 3%, and the interassay precision was 3.5%.
Binding Studies
Binding activity index of CBG for cortisol
The combining affinity index or "C value (L/g)" [22], which measures the ability of a mixture of proteins to bind a small molecule, is given by the expression
 |
Binding parameters The affinity constant (Ka) and the apparent concentration of binding sites (n) of three pools of plasma from maternal (peripheral and intervillous) and fetal (umbilical vein and arteries) circulations were estimated by Scatchard graphic analysis. Specific binding was estimated after subtraction of nonspecific binding, which was evaluated with excess unlabeled steroid [20]. Nonspecific bindings for each plasma sample studied did not exceed 2%–5% of total binding. Stripped endogenous ligands and unstripped plasma binding were measured by batchwise gel equilibration [23], in which fixed amounts of plasma from each maternofetal compartment (containing 8–10 nM CBG.) were added to tubes containing 0.2 g of Sephadex G-25, 2 ml of phosphate buffer (pH 7.4; 0.07 M KH2PO4/Na2HPO4), a fixed amount of 3H-cortisol (105 cpm = 0.7 nM), and increasing quantities of unlabeled cortisol (0.7–150 nM) [24–26]. The tubes were incubated for 1 h at 0°C. The parameters for cortisol binding to CBG were calculated for three pools of plasma (n = 5 for each) before and after treatment.
RNA Preparation and RT
Total RNA was extracted from placenta villi and cytotrophoblast cells using the Trizol reagent (Life Technologies, Cergy, France), and RT was performed as previously described [24].
RT-PCR and Southern Blot Analysis
The primers used to amplify human CBG cDNA were as follows: CBG upper oligonucleotide (sense), 5'-ATGACCTTGGAGATGTGCTG-3' (priming site in exon 4, nucleotides 929–948 [17]); and CBG lower (antisense), 5'-GGTCTCTTACACTGGGTTCA-3' (priming site in exon 5, nucleotides 1205–1224; Genosys Biotechnologies, Montigny-Le-Bretonneux, France). Briefly, 2.5 µl of cDNA reaction mixture were subjected to amplification in the presence of 0.2 µM sense and antisense oligonucleotide primers, 0.1 mM of each deoxynucleotide triphosphate (dNTP), 1 mM MgCl2, and 2.5 U of Taq polymerase (Life Technologies) in 25 µl of PCR buffer. Amplification consisted of 32 cycles with denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 1 min, with a final extension at 72°C for 10 min.
Lack of genomic DNA contamination was checked in all experiments by conducting a control reaction containing mRNA without reverse transcriptase. An endogenous gene, human ß2-microglobulin cDNA, was amplified in parallel using an additional pair of primers (Ozyme, Saint Quentin en Yvelines, France) to control for errors in the cDNA used in PCR reactions.
An aliquot (15 µl) of each PCR product was run on electrophoresis (3% NuSieve GTG agarose gel, FMC, in Tris-borate-EDTA buffer containing 0.01% ethidium bromide). A DNA nucleotide length standard (lambda 123-base pair [bp] DNA ladder; Life Technologies) was used to confirm the predicted PCR product sizes.
The PCR products were also checked after Southern blot analysis using an internal 5'-GTGGTCCATAAAGCTG-TGCT-3' oligonucleotide probe (nucleotides 1031–1050) specific for CBG that was labeled with fluorescein-11-deoxy-uridine triphosphate. Products were detected using an ECL 3'-oligolabeling detection ECL kit (RPN 2130; Amersham) according to the manufacturer's instructions.
Protein Determination
Protein concentrations were assayed according to the method described by Bradford [25] using human serum albumin as the standard.
Statistical Analysis
Results are expressed as the mean ± SEM. Data were analyzed by ANOVA for repeated measures after verification of their normal distribution. When the F value was significant (P < 0.05), Dunnett t-test was used for multiple comparisons of the means.
RESULTS
Albumin and CBG Immunoquantification
Figure 1 shows the uniform distribution of albumin in the maternal and fetal circulations of pregnant women at term (n = 22). Thus, the albumin concentrations in the maternal peripheral (M) and intervillous (I) blood and in the umbilical arteries (A) and vein (V) were not significantly different. The albumin concentration in M was also similar to that of nonpregnant women (NP). By contrast, the peripheral plasma CBG concentration was three- to fivefold higher in pregnant than in nonpregnant women. A transplacental concentration gradient of CBG was also found (Fig. 1 and Table 2). The CBG concentration in M was 1.5-fold higher (P < 0.003) than in I and fivefold higher (P < 0.001) than in V and A. The CBG concentration in A was not significantly different from that in V.
|
|
CBG Isoforms in the Intervillous Blood Space: Comparison with Peripheral Maternal and Fetal CBG
Western blotting Analysis of CBG in the various compartments of the maternofetoplacental unit (Fig. 2) indicated that the antiserum specifically recognized two main CBG electrophoretic isoforms from M, I, V, and A, with apparent molecular weights ranging from 50–60 kDa. The faster CBG isoform appeared to be present in a smaller amount.
|
Crossed immunoelectrophoresis Crossed immunoelectrophoresis of human CBG with a second-dimension gel containing anti-human CBG antibodies gave a single precipitation peak in NP, M, I, V, and A (Fig. 3). The patterns of immunoreactive CBG in M and I were similar. However, they differed from those of NP and fetal blood (i.e., A and V), because the CBGs from M and I were slightly more anionic than the corresponding CBGs from NP, A, and V.
|
The carbohydrate moiety of CBG enables it to bind to lectins such as ConA. Thus, CBG glycoforms were determined by their differential lectin-binding properties using crossed immunoaffinoelectrophoresis. Interaction with the lectin retarded glycoprotein migration in the first-dimension electrophoresis, so the multiple peaks probably reflected different degrees of interaction and/or different numbers of binding sites between ConA and human CBG.
The immunological behavior of the CBGs from the M and I plasma were identical, but they differed from those of the CBGs in the NP, A, and V fluids (Fig. 4). The ConA-CBG complexes (i.e., CBG ConA-reactive isoforms) formed immunoprecipitates with clearly lower mobilities than the immunoelectrophoretic patterns of CBG from M, I, V, A, and NP in the absence of lectin (upper CBG patterns). The maternal peripheral plasma CBG (i.e., M) and intervillous CBG (i.e., I) had five isoforms (down CBG patterns) that reacted differently with ConA: 1) ConA unreactive, 12% ± 2%; 2) ConA ± weak reactivity, 15% ± 2%; 3) ConA reactive +, 35% ± 3%; 4) Con A reactive ++, 23% ± 2%; and 5) ConA reactive +++, 16% ± 1%. By contrast, the CBG from NP, V, and A had similar patterns, with mainly very retarded ConA isoforms.
|
CBG Synthesis in Term Placenta Villi, Isolated Cytotrophoblastic Cells, and Syncytiotrophoblasts in Culture
Expression of the CBG gene was investigated using total RNA extracted from term placenta villi, isolated cytotrophoblastic cells, and syncytiotrophoblasts in culture. Human hepatoblastoma-derived (HepG2) cells were used as a positive control. The RT-PCR amplification using external oligonucleotides specific for a CBG sequence and Southern blotting and hybridization with an internal oligonucleotide specific for CBG sequence gave a labeled product of the predicted size (296 bp) in the term placenta villi (Fig. 5a, 1 and 2), cytotrophoblasts (Fig. 5b, 1 and 2), syncytiotrophoblasts (Fig. 5b, 3 and 4), and positive control HepG2 cells (Fig. 5a, 3 and Fig. 5b, 5). Omitting RT from the reacting mixture resulted in no amplification product. Positive amplification of the ß2-microglobulin gene led us to assert the integrity of mRNA in all of the samples analyzed and the effectiveness of RT (data not shown).
|
By contrast, we found no immunoreactive CBG in the medium of trophoblastic cells cultured for 24 h or in the cytosol of these cells using Laurell immunoelectrophoretic quantitation. Western blotting showed no classic (50–60 kDa) immunoreactive CBG protein in isolated trophoblastic cells (Fig. 2, T1 and T2). However, several immunoreactive bands of varying intensities were detected in the cytotrophoblastic cells. Only the fastest and most highly labeled, 42-kDa immunoreactive protein was found in all the placentas studied. This immunoreactive protein may be an unglycosylated CBG isoform as previously described in certain recombinant CBG mutant [26]
CBG-Steroid Interactions at the Fetomaternal Interface
Quantitation of total cortisol and progesterone The concentrations of cortisol and progesterone in M, V, A, and I indicated significant differences among the blood compartments (Fig. 6). The cortisol concentrations in the plasma of pregnant women were 1.5- to 2.0-fold higher than those in nonpregnant women; these results are in agreement with published data (see arrows NP in Fig. 6) [20]. The cortisol concentration in I was approximately 10-fold lower than in M (P < 0.001) and twofold lower than the cortisol concentrations in A (P < 0.01) and V (P < 0.001) (Fig. 6 and Table 2). The differences between cortisol in A and V were not significant.
|
The progesterone concentration in I was 10-fold higher (P < 0.001) than in M and five- to sevenfold higher than in the cord blood (P< 0.001) (Fig. 6 and Table 2). The progesterone concentration in M was also lower than that in cord blood (P < 0.001), and more progesterone was found in V than in A (P < 0.05). The very high concentration of progesterone in I was confirmed by gas chromatography/mass spectrometry. The progesterone concentrations in five pregnant women as determined by this method (I = 4720 ± 146 nM) were not significantly different from those as determined by radioimmunoassay (Table 2).
The progesterone:cortisol molar ratio in I was 75-fold higher than in M, whereas in the fetal circulation, this ratio was 5- to 10-fold greater than that in M (Table 2). Thus, the molar concentrations of the main ligands of CBG (i.e., progesterone and cortisol) in M were 30% lower than the CBG concentration, whereas their concentrations greatly exceeded (by two to fourfold) that of CBG in the other compartments. Such different hormonal environments (progesterone + cortisol/CBG) might influence the binding properties of the CBG in each compartment.
Binding properties of CBG in the maternal and fetal compartments Figure 7 (1) shows that, in a physiological context, the binding activity indices of plasma CBG for cortisol (L/g; expressed as C value = [B/U x [1/P]) were higher in M than in I, V, or A (P < 0.001). Thus, there appeared to be a discrepancy between the plasma CBG concentration in I (immunoquantitation), which was only 30% lower than that in M, and the observed reduction in the binding activity of the protein (10-fold lower) in this compartment.
|
These results could be caused by the CBG-binding sites being occupied by endogenous ligands or by the presence of nonbinding CBG variants. Because of the steroid hormone distribution of CBG in the maternal and fetal circulations reported above and the specific lipid microenvironment described in previous studies [4], the endogenous steroids were removed using AC (Fig. 7, 3), lipids were stripped with Lipidex (Fig. 7, 2), and the combined treatments were applied (Fig. 7, 4). The C values for CBG always increased (P < 0.001) after use of AC (Fig. 7, 3) or of AC plus Lipidex (Fig. 7, 4).
The changes in C values were small. Therefore, accurate Scatchard plots were obtained for 3H-cortisol binding to CBG from the pooled plasma for each compartment (Fig. 8).
|
The CBG-binding parameters before plasma treatment were deduced from Scatchard plots and are shown Figure 8 (1). They indicated that CBG from M has a four to fivefold greater Ka than I, V, and A, whereas the number of binding sites per mole CBG (n) did not significantly differ from one compartment to another: the value of n was less than one in all cases. The binding parameters of CBG changed significantly after removal of endogenous steroids and lipids. Steroid removal (Fig. 8, 3) always led to an increase in Ka without any significant change in the number of binding sites. Lipidex treatment alone (Fig. 8, 2) caused a small change in Ka, which increased 1.5- to 2-fold for I, V, and A. However, AC plus Lipidex (Fig. 8, 4) caused a marked increase in the number of binding sites for M and I, which tends to be one site per molecule of CBG. Removing steroids and lipids from fetal CBG did not change the number of binding sites, which remained less than one site per mole CBG, but it significantly increased the Ka of CBG for cortisol from A or V. Thus, removal of endogenous ligands revealed that fetal CBG had a Ka for cortisol two- to fourfold higher than that of CBG in M and I. The Ka for CBG in I was always twofold lower than that of CBG in M (P < 0.01) and fourfold lower than that of CBG in A or V (P < 0.001).
Previous studies have shown higher concentrations of polyunsaturated fatty acids (PUFAs) in I (PUFA:CBG molar ratio = 53) than in M (PUFA:CBG molar ratio = 10) [4] and have indicated that this special lipid environment affects conformation and binding properties of CBG [20, 27, 28]. Therefore, we explored the susceptibility of CBG to conformational changes induced by PUFAs. Exogenous arachidonic (C20:4) and docosahexaenoic (C22:6) fatty acids were added to plasma CBG from M and I, and the immunological behavior of the CBG molecules was investigated. Crossed immunoelectrophoresis (Fig. 9) demonstrated that PUFAs decreased the immunoreactivity of CBG from M (40%) to a lesser extent than that of CBG from I (80%), indicating that CBG from I was more sensitive to exogenous fatty acids. Adding fatty acids to CBG without any incubation had no effect on the immunorecognition of human CBG by anti-CBG antibodies (data not shown).
|
DISCUSSION
Our findings indicate a transplacental gradient of CBG concentration between the maternal and fetal circulations, with a notable quantity of CBG at the fetomaternal interface in human term pregnancy. Thus, the CBG level in the intervillous space is only slightly lower (20%–30%) than that in the maternal peripheral blood (P < 0.001) and approximately threefold more than that of the fetal circulation (P < 0.001).
It is generally agreed that CBG is mainly synthesized by the maternal and fetal livers, and that the placenta does not produce CBG to any significant degree, especially in mice and rabbit [16, 29]. However, it was recently shown that CBG mRNA is present in the human syncytiotrophoblast at term [17]. Thus, the CBG detected here at the maternofetal interface might be of maternal origin and/or result from the passage of fetal hepatic CBG through the placental villi, or it may even come from a trophoblastic source. Thus, it was important to determine the maternal and/or fetal origin of CBG at the fetomaternal interface, because the syncytiotrophoblast is in direct contact with the maternal intervillous blood and because we recently showed high concentrations of a fetal protein (i.e., alpha-fetoprotein) in this compartment [4]. Therefore, we compared the immunological behavior, glycan-microheterogeneity, and binding parameters of the maternal and fetal CBGs. The CBGs in the maternal intervillous and peripheral circulations seem to have similar immunoelectrophoretic profiles and are clearly more anionic than either fetal CBG in cord blood or adult CBG from nonpregnant women. Immunoaffinoelectrophoresis of CBG in the presence of the lectin ConA revealed the differential glycan-microheterogeneity of CBG in the maternal and fetal compartments. The CBG in the intervillous and in the peripheral maternal circulation both have five CBG isoforms and are clearly distinct from fetal and nonpregnant CBG. Our findings agree with those of previous studies, which showed the presence of a specific, ConA-unreactive variant of CBG in the maternal peripheral plasma of pregnant women and retroplacental blood [11, 14]. This suggests that the CBG at the fetomaternal interface is of maternal origin, but it does not rule out the presence of a small quantity of CBG from trophoblastic cells. This is why we looked for CBG synthesis in trophoblastic cells isolated from human term placenta. Both RT-PCR and Southern blotting detected CBG mRNA in cytotrophoblastic cells or syncytiotrophoblasts in culture. These data agree with in situ hybridization results indicating that human syncytiotrophoblast contain CBG mRNA [17]. However, in our experimental conditions, neither Laurell immunoquantitation nor Western blotting revealed any classic, 50- to 60-kDa CBG protein in trophoblastic cells, although SDS-PAGE constantly revealed an intense, immunoreactive, smaller, 42-kDa protein, which might well be an unglycosylated human CBG isoform [26]. This suggests that trophoblastic cells contribute to the production of CBG in the maternal intervillous space, but that the CBG is a variant that does not react with ConA and is a minor component in maternal blood.
Discrepancies were found between the CBG concentration and the CBG-binding activity index (C value [L/g]) in maternal and fetal blood compartments. Although the amounts of CBG and the CBG isoform distributions in maternal peripheral and intervillous blood are similar, these proteins have very different binding properties in their respective physiological ligand contexts. These differences, plus the existence of less than one binding site per mole CBG in all compartments, suggest that environmental hormonal or lipidic factors interfere with the binding of cortisol to CBG, especially in the maternal intervillous circulation.
We first analyzed the steroid hormone status of the maternal and fetal compartments to determine the physiological significance of the interactions between CBGs and their ligands. The intervillous circulation is distinct from the peripheral maternal circulation in having a high progesterone content and a low cortisol level. This low cortisol concentration probably results from a lack of placental 17-hydroxylase [30] or from activation of placental 11ß-hydroxysteroid dehydrogenase type 2, which transforms cortisol into cortisone [31].
The impact of endogenous plasma steroids on the binding properties of CBG was also evaluated. Removing steroids led to an increased Ka of CBG for cortisol, although less than one binding site per molecule of CBG was still found in all compartments. Such fractional numbers of binding sites on CBG molecules could result from the existence of nonbinding forms of CBG or of CBG-binding sites being occupied or masked by other endogenous ligands, such as PUFAs [20, 27, 28], which are present at a very high concentrations, especially in the intervillous circulation [4]. Good in vitro and in vivo evidence also indicates that such essential fatty acids act as endogenous conformational effectors and modulators of the binding of hormones to plasma proteins, such as sex steroid-binding protein and CBG [20, 27, 28]. Conformational change caused by a transient rise in the lipid microenvironment of the CBG may also decrease the immunoreactivity of the protein for the anti-CBG antibody. As a result, the CBG concentration in maternal intervillous blood may be underestimated. Immunological exploration provides added circumstantial evidence for fatty acid-mediated conformational changes in CBG, such as the immunoreactivity of CBG in maternal intervillous plasma, which was more readily altered by exogenous PUFAs than was that of maternal peripheral plasma. We have tentatively attributed these differences in CBG sensitivity to the pre-existing, threefold higher concentration of such fatty acids in the intervillous circulation [4].
The removal of endogenous steroids and lipids caused a more marked increase in the binding parameters of both maternal and fetal CBG molecules. Thus, the number of binding sites (n) for maternal peripheral and intervillous CBG increased to one binding site per mole CBG. However, the Ka of CBG in the maternal intervillous circulation was only approximately half that of peripheral maternal CBG (P < 0.01). This is in agreement with the observation of Werthamer et al. [32], indicating a placental CBG with a lower affinity for cortisol. The Ka of fetal CBG was significantly higher (P < 0.001) than those of maternal CBGs after endogenous ligand stripping, but the number of binding sites per molecule of fetal CBG remained less than one, suggesting the presence of inactive CBG isoforms in the fetal blood. Variant forms of CBG with abnormal steroid-binding characteristics have been described in pathological situations, such as septic shock [33, 34] or inflammation [15]. Several studies suggest that gestation and the mechanisms involved in the onset of labor and parturition are linked with inflammatory processes [35]. The CBG is a member of the serine protease-inhibitor family and is a substrate for elastase on the surface of neutrophils, and the activity of these cells increases during pregnancy [36]. Thus, the local action of a proteolytic enzyme may reduce the binding affinity or capacity of CBG in human intervillous and cord blood.
The presence of CBG and the unique hormone pattern of the maternal intervillous blood at term raise questions regarding both the spatiotemporal relationship of CBG with its hormone environment throughout pregnancy and the functional significance of the CBG-steroid interactions at the fetomaternal interface. Such interactions could be involved in the immune process of pregnancy, to allow survival of the semiallogeneic graft that constitutes the fetus [35], and also in the mechanisms of human parturition.
The dramatically high progesterone concentration in the maternal intervillous circulation at term strongly suggests that CBG, at this pivotal site, is highly loaded with progesterone, although CBG in the maternal peripheral circulation is probably loaded mainly with cortisol. The concentration of CBG ligands (i.e., cortisol and progesterone) in maternal intervillous blood and the fetal circulation exceeds greatly the CBG concentration, whereas an equimolar ratio of CBG to steroid is found in the maternal peripheral circulation. Approximately 85% of the progesterone and cortisol are bound in all compartments, suggesting that other plasma proteins, such as albumin and orosomucoid, bind these hormones despite their relatively low steroid-binding affinity. Their role must not be underestimated in the control and transport of steroids and their passage across cell membranes [5–7].
Several reports indicate that interaction of the CBG-steroid complex with specific membrane-binding sites is the first in a sequence of events leading to a specific cell response [8, 9, 37, 38]. Specific binding sites for the CBG-progesterone complex have been found on the plasma membranes of human decidua [9, 11], whereas the CBG-cortisol complex seems to be preferentially bound to membrane receptors on the syncytiotrophoblast [14, 37]. These membrane receptor-mediated mechanisms could be important for autocrine action and for the transfer of these hormones to the fetus. Cortisol is essential in fetal lung maturation [38]. The CBG-progesterone complex may also act as a shuttle for moving progesterone from the maternal intervillous circulation to target placental and uterine smooth muscle cells. This hypothesis is consistent with the recent finding of progesterone receptors in the muscle layer of human term fetoplacental vessels [24]. It suggests that progesterone is involved in the control of proliferation or differentiation and vasoactivity of the fetoplacental circulation, which are important events in placental vascular physiology and pathology, such as pre-eclampsia and intrauterine growth retardation. This may also occur on the maternal side, because progesterone receptors have been found in term decidual cells [39] and the myometria of pregnant women during late pregnancy [40].
The wide molecular and functional spectrum of human CBG at the fetomaternal interface is an indication of the great adaptive and/or interactive capacities of CBG and of its response to changes in its ligand environment during pregnancy. The diversity of the reciprocal intrinsic and extrinsic connections between the various CBG isoforms and steroid hormones in both space and time can modulate the activity of CBG, giving to it functional pluripotentiality throughout gestation. The present study has focused on the singularity of the human fetomaternal interface and the physiological significance of CBG-steroid interactions at this pivotal site, especially at the dawn of a new human life.
ACKNOWLEDGMENTS
We thank Dr. O. Parkes for editorial help.
FOOTNOTES
First decision: 10 August 2000.
1 Supported by the Institut National de la Santé et de la Recherche Médicale, and by University René Descartes Paris V (UER Cochin). 
2 Correspondence: C. Benassayag, INSERM U.361, Pavillon Baudelocque, 123, Boulevard de Port-Royal, 75014 Paris, France. FAX: 33 143 26 44 08; u361{at}cochin.inserm.fr
Accepted: October 13, 2000.
Received: July 10, 2000.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  M. Perusquia, E. Navarrete, J. Jasso-Kamel, and L. M. Montano Androgens Induce Relaxation of Contractile Activity in Pregnant Human Myometrium at Term: A Nongenomic Action on L-Type Calcium Channels Biol Reprod, August 1, 2005; 73(2): 214 - 221. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Mesiano Myometrial Progesterone responsiveness and the Control of Human Parturition Reproductive Sciences, May 1, 2004; 11(4): 193 - 202. [Abstract] [PDF]
|
|
|
|
 |
|
 J. M. Joyner, L. J. Hutley, A. W. Bachmann, D. J. Torpy, and J. B. Prins Greater replication and differentiation of preadipocytes in inherited corticosteroid-binding globulin deficiency Am J Physiol Endocrinol Metab, May 1, 2003; 284(5): E1049 - E1054. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
