Human chorionic gonadotropin (hCG) displays a major role in pregnancy initiation and progression and is involved in trophoblast differentiation and fusion. However, the site and the type of dimeric hCG production during the first trimester of pregnancy is poorly known. At that time, trophoblastic plugs present in the uterine arteries disappear, allowing unrestricted flow of maternal blood to the intervillous space. The consequence is an important modification of the trophoblast environment, including a rise of oxygen levels from about 2.5% before 10 wk of amenorrhea (WA) to ∼8% after 12 WA. Two specific β-hCG proteins that differ from three amino acids have been described: type 1 (CGB7) and type 2 (CGB3, -5, and -8). Here, we demonstrated in situ and ex vivo on placental villi and in vitro in primary cultures of human cytotrophoblasts that type 1 and 2 β-hCG RNAs and proteins were expressed by trophoblasts and that these expressions were higher before blood enters in the intervillous space (8–9 vs. 12–14 WA). hCG was immunodetected in villous mononucleated cytotrophoblasts (VCT) and syncytiotrophoblast (ST) at 8–9 WA but only in ST at 12–14 WA. Furthermore, hCG secretion was fourfold higher in VCT cultures from 8–9 WA compared with 12–14 WA. Interestingly, VCT from 8–9 WA placentas were found to exhibit more fusion features. Taken together, we showed that type 1 and type 2 β-hCG are highly expressed by VCT in the early first trimester, contributing to the high levels of hCG found in maternal serum at this term.
- gene expression
- cell fusion
the chorionic villus represents the structural and functional unit of the human placenta and is bathed in maternal blood via the spiral arteries within the intervillous space at the end of the first trimester of pregnancy. Indeed, during the early first trimester, blood flow to the placenta is severely restricted due to obstruction of the uterine vessels by migrating trophoblastic cells (8). On physiological remodeling of these arteries, trophoblastic plugs disappear, allowing the unrestricted flow of maternal blood to the intervillous space. Thus, before around 10 wk of amenorrhea (WA) oxygen concentrations in the intervillous space are ∼2.5% (Po2 of 15 mmHg), whereas, after the onset of maternal blood flow to the placenta (at about 12 WA), oxygen concentrations in the intervillous space rise to ∼6–8% corresponding to a Po2 of about 60 mmHg (7, 22, 23). However, the rise of oxygen is not the only change. Indeed, the entry of maternal blood is also accompanied by modification of flow and by the presence of all the components that maternal blood can contain. It is a radical modification of the environment for the chorionic villi.
The mononucleated villous cytotrophoblasts (VCT) that cover the floating chorionic villi fuse to form a multinucleated syncytiotrophoblast (ST) that is in direct contact with the maternal blood. ST is involved in the exchange of gases and nutrients between the mother and the fetus. The ST represents also the endocrine tissue of the placenta, secreting a large amount of hormones, including human chorionic gonadotropin (hCG). The variation of hCG gene expression and secretion during these early gestation events is poorly described.
After implantation, hCG is the first trophoblast signal detected in the maternal blood and is absolutely necessary for placental development. Maintenance of pregnancy during early gestation depends on the synthesis of hCG, which prevents regression of the corpus luteum (16), allowing the maintenance of ovarian progesterone secretion (20). In addition to its well-established endocrine role, hCG has a role in human trophoblast differentiation in a paracrine manner, as well as fusion (40) and invasion (14), and in angiogenesis (3, 41).
hCG is a glycoprotein hormone composed of an α-subunit that is common to pituitary gonadotropins such as luteinizing hormone (LH), follicle-stimulating hormone, and thyroid-stimulating hormone, and a β-subunit that confers the biological specificity of the hormone and represents a step limiting for hCG synthesis. hCG and its free β-subunit are detected in the maternal blood from the 2nd wk of pregnancy, and their levels increase until reaching a peak at 10–12 wk and then decrease gradually, whereas α-hCG levels increase progressively up to term (20). A single gene located on chromosome 6q21.1-q23 encodes the CGA subunit. The CGB subunit molecule is encoded by any one of the six nonallelic genes CGB8, CGB7, CGB5, CGB3, CGB2, and CGB1 present on chromosome 19q13.32 (5). CGB1 and CGB2 are two pseudogenes, whereas the other CGB genes are regrouped in two subtypes: type 1 (CGB7) and type 2 (CGB3, -5, and -8), which correspond to two different proteins that differ from three amino acids. Little is known about the expression of β-hCG type 1 and 2 during villous trophoblast differentiation before and after oxygenation of the intervillous space.
The objective of this study was to investigate in early and late first-trimester placentas: 1) the type of the β-subunit hCG gene expression in placental tissues and 2) the expression and secretion in vitro of these β-subunits during villous trophoblast differentiation from a mononucleated villous trophoblastic cell to a ST.
MATERIALS AND METHODS
Isolation, purification, and primary culture of VCT.
Placental tissues were obtained from the Department of Obstetrics and Gynecology at Saint Vincent Hospital (Paris, France) during the first trimester (8–14 WA) following legal voluntary interruption of pregnancy in patients. These biological samples were obtained following informed patient written consent and approval from our local ethics committee (CCPPRB, Paris Cochin, N°18–05, Paris, France).
The tissue was washed in Ca2+- and Mg2+-free Hanks' balanced salt solution (HBSS) supplemented with 100 IU/ml penicillin and 100 μg/ml streptomycin. Chorionic villi were dissected and rinsed before storing at −80°C or cell isolation. VCT isolation was based on the methods of Kliman et al. (25) and is a modified version of Tarrade et al. (35) and Handschuh et al. (15). After dissection, the chorionic villi were incubated in HBSS (5 ml/g) containing 0.1% trypsin (Difco Laboratories, Detroit, MI), 4.2 mM MgSO4, 25 mM HEPES, and 50 U/ml DNase type IV (Sigma, Saint-Quentin Fallavier, France) for 15 min at 37°C without agitation, and this was repeated four times with the same trypsin solution. The three first 15-min trypsin digestions, containing a mix of extravillous cytotrophoblasts and VCT, were discarded, and the two last ones, containing a majority of VCT, were kept and pooled. The chorionic villi were finally washed with warm HBSS (37°C). Each time, the supernatant containing VCT was collected after tissue sedimentation, filtered (100-μm pores), and incubated on ice with 10% FCS (vol/vol) and 10% DMEM to stop trypsin activity. After purification by Percoll gradient fractionation, VCT were diluted to a concentration of 1.25 × 106 cells/ml in DMEM supplemented with 10% FCS, 2 mM glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin and plated at 150,000 cells/cm2 on 35- or 60-mm Techno Plastics Products culture dishes. After 2 h of culture in 5% CO2 at 37°C, VCT were carefully washed two times to eliminate nonadherent cells. Purified VCT cultures were characterized by cytokeratin 7 (CK7) immunolabeling and by the ability to the aggregate at 48 h and to form ST at 72 h. Cells were trypsinized and harvested, and cell pellets were frozen at −80°C until protein or RNA extraction.
Quantitative real-time PCR.
Relative real-time PCR (Q-PCR) was used to compare expression of hCG genes in primary-cultured VCT and ST from early (n = 7) or late (n = 5) first-trimester placenta or directly on placental tissues from early (n = 7) or late (n = 7) first-trimester placenta. Total RNA was extracted using the acid-phenol guanidium method (TRIzol reagent; Invitrogen) and purified with the RNeasy MinElute Cleanup Kit (Qiagen) following the manufacturer's instructions. cDNA synthesis was performed using the Superscript III Reverse Transcriptase kit (Invitrogen). Total RNA (0.5 or 1 μg) was heated to 65°C for 5 min with 0.1 μg of random primers and 500 μM of each dNTP, chilled on ice, briefly centrifuged, and reverse transcribed in a final volume of 20 μl containing 1× RT buffer (3 mM MgCl2, 75 mM KCl, and 50 mM Tris·HCl, pH 8.3), 40 units of RNaseOUT (Invitrogen), 20 mM dithiothreitol, and 200 units of Superscript III RNase H-RT. Samples were incubated at 25°C for 2 min and 42°C for 50 min, inactivated at 70°C for 15 min, and then cooled at 5°C. The primers used for this study are described in Table 1 and were obtained from Eurogentec (Angers, France). We tested by RT-PCR (Go Taq flexi DNA Polymerase; Promega) three primer couples and two primer couples for type 1 β-hCG and type 2 β-hCG genes, respectively, and actin primers for endogenous control. We chose the primer couple giving the best amplification for each type gene, and we validated its specificity by sequencing (GATC Biotech, Konstanz, Germany). All Q-PCR reactions were performed in an ABI Prism 7900 Real Time Detection system (Applied Biosystems, Courtaboeuf, France) using a cDNA dilution corresponding to 25 ng of RNA with the SYBR Green PCR kit (Eurogentec). Five or seven experiments were performed in duplicate for each gene; the initial denaturation step at 95°C for 10 min was followed by 40 cycles at 95°C for 15 s and 65°C for 1 min followed by a three-step dissociation of 15 s at 95°C, 30 s at 60°C, and 15 s at 95°C to verify the specificity of the primer binding. CK7, a trophoblast intermediate filament protein, and ribosomal 18S (Kit Eurogentec) were used as endogenous RNA controls for normalization and gave similar results.
For total protein content, cell pellets were incubated with 200 μl of RIPA buffer [20 mM Tris·HCl (pH 8), 150 mM NaCl, 4 mM EGTA, and 1% Triton X-100] with antiproteolytic cocktail in ice for 30 min and vortexed every 10 min. Cell lysates were centrifuged at 13,000 g for 20 min, and the supernatants containing the protein lysates were kept. Protein lysates (30 μg) were incubated in 4× lithium dodecyl sulfate, 10× reducing agent, and milliQ water in a 40-μl final volume. After 10 min at 60°C, the dissociated proteins were loaded on a gradient of 4–12% acrylamide gel. Proteins were transferred to nitrocellulose membranes and incubated with 5% nonfat milk in PBS-0.1% Tween 20 for 2 h at room temperature. Next, the membranes were incubated with the specific antibody (Table 2) in 0.5% nonfat milk in PBS-Tween overnight at 4°C followed by horseradish peroxidase-conjugated goat anti-rabbit or anti-mouse IgG antibodies and enhanced chemiluminescence reagents. The different antibodies used were directed against β- and dimeric hCG (A0231, 1:2,000, rabbit polyclonal antibody; Dako, Trappes, France) and against free type 2 β-hCG (FBT11-II, 1:500; Table 2). Murine antibody FBT11-II, elicited against purified β-hCG subunit (CR 129), is directed to a discontinuous epitope encompassing residues 1 through 7 and 82 through 92 of β-hCG (4). This antibody was shown to be specific to type 2 β-hCG (Aldaz-Carroll L, personal communication).
Total hCG (hCG) and free β-hCG were determined on 24-h culture medium collected after 24, 48, and 72 h of cell culture using the chemiluminescent immunoassay analyzer ACS-180SE system (Bayer Diagnostics, Westwood, CA). The assay used for total hCG quantification was a sandwich immunoassay that recognized free β-subunit, nicked and intact hCG. Two antibodies were used: a polyclonal goat anti-hCG antibody and a mouse monoclonal anti-hCG antibody, which were raised against two different epitopes of the β-subunit of hCG. Performance characteristics were as follows: range 2–1,000 IU/l; within-run precision 2.8% and between-run precision 5.6%; functional detection limit 1 IU/l. Samples were diluted (1:100) in the reaction buffer when necessary. Free type 2 β-hCG were quantified in the same cell supernatants by immunoassay. To compare the secretion of hCG in early and late first-trimester VCT supernatants, results were converted to nanograms per milliliter (9.3 mIU/ml hCG correspond to 1 ng/ml; WHO).
Villi from placental tissues from 8–9 WA (n = 5) and from 12–14 WA (n = 3) were fixed and embedded in parablast, and serial sections were immunostained as described elsewhere (15). The two different antibodies used were directed against β- and dimeric hCG (A0231, 1:2,000, rabbit polyclonal antibody; Dako) or against free type 2 β-hCG (FBT11-II, 1:500) (Table 2).
Determination of cell fusion index.
Syncytium number and their size were analyzed on fixed cells (8 min in methanol at −20°C) after immunolabeling of the intercellular boundaries with a mouse anti-desmoplakin antibody (IgG, 1:200; Abcam, Cambridge, UK) and nuclei with DAPI (Fig. 1A). After a 72-h culture, the number of mononuclear cells and syncytia and the number of nuclei in syncytia were counted, and the fusion index (N − S)/T was determined (N is nuclei number in syncytia, S is number of syncytia, and T is total number of nuclei) (32). For each group of term, the experiment was performed with three different primary cultures obtained from three different placentas and run in duplicate each time.
Values represent means ± SE of 5–10 separate primary cultures obtained from 7 different placental tissues. Statistical analysis was performed using GraphPad Prism software (GraphPad, La Jolla, CA). The Mann and Whitney test was used to test differences between groups. ANOVA followed by the Kruskal-Wallis test was used to compare protein secretion from 24-, 48-, and 72-h trophoblast cultures. Results were considered significant when P < 0.05.
Analysis of type 1 and type 2 β-hCG gene expression in chorionic villi from early and late first trimester.
Specific primers for each β-hCG type were designed and validated after sequencing of RT-PCR products obtained with total RNA extracted from villous trophoblasts and the T24 cell line. The human bladder carcinoma cell line T24 was used as a negative control for type 2 β-hCG gene expression (2). Figure 1A depicts a representative 2% agarose gel showing the expression of type 1 and type 2 β-hCG transcripts in the trophoblast (ST) and of only type 1 in the T24 cell line. To assess levels of type 1 and type 2 hCG gene expressions ex vivo, placental villi from 8–9 and 12–14 WA were used for real-time Q-PCR and Western blot analysis. Total hCG and type 1 and type 2 β-hCG transcript levels in placental villi at both terms were analyzed by relative quantitative RT-PCR and normalized to CK7, a specific trophoblast intermediate filament protein in placenta, or to 18S ribosomal mRNA, which gave similar results (data not shown). The results shown are normalized to CK7 to reflect the trophoblast mass within the placental villi. Figure 1B shows that type 1 and type 2 β-hCG transcript levels were both expressed by chorionic villi and higher in 8–9 WA compared with 12–14 WA. The most important difference between term groups was for type 1 β-hCG RNA levels, which were 12-fold higher in early placentas. hCG protein expression was analyzed by Western blot using an anti-total (anti-type 1 and anti-type 2) β-hCG antibody (A0231) and a specific anti-β-hCG type 2 antibody (FBT11). We observed that expression of total and type 2 β-hCG proteins was also higher in early than in late chorionic villi (Fig. 1C).
Analysis of type 1 and type 2 β-hCG transcript levels in trophoblast primary cultures.
Mononucleated VCT were isolated from early (8–9 WA) and late (12–14 WA) first-trimester human chorionic villi and differentiated in vitro into multinucleated ST over a 72-h culture period. As shown in Fig. 2A, mononucleated VCT fused to form a multinucleated ST at 72 h as shown after desmoplakin immunolabeling and nuclei staining with DAPI. Total hCG and type 1 and type 2 β-hCG transcript levels in trophoblasts at both terms were analyzed by relative quantitative RT-PCR and normalized to 18S ribosomal mRNA. CK7 was also used as endogenous RNA controls and gave similar results (data not shown). We first compared hCG gene expressions during trophoblast differentiation in vitro (Fig. 2B). In both 8–9 WA and 12–14 WA trophoblasts, type 1 and type 2 β-hCG RNA were more expressed in ST compared with mononucleated VCT (∼10-fold). However, total β-hCG gene expression during the formation of the ST was 8- and 22-fold higher in early and late first trimester, respectively. We next compared β-hCG gene expressions between terms in either mononucleated VCT or ST (Fig. 2C). In 8–9 WA mononucleated VCT, types 1 and 2 β-hCG and total hCG were two- to threefold higher than in 12–14 WA VCT, but no significant difference was observed in ST irrespective of the gestational age. The higher expressions of both type 1 and 2 β-hCG in VCT from 8–9 WA are in agreement with results obtained with placental villi (cf. Fig. 1B).
The comparison of cycle threshold (Ct) values gives an evaluation of transcript expression levels of each β-CG type in the same cells. In all conditions, whether it was comparing VCT with ST or comparing between term groups, we observed a difference of 7–10 Ct values between type 2 (Ct value range: 17–20) and type 1 (Ct value range: 26–30), indicating that type 2 should be more expressed than type 1 β-hCG in our experimental conditions.
hCG protein expression and secretion in trophoblast primary cultures from early and late first-trimester human placentas.
Figure 3A depicts the whole cell content in Cu,Zn-superoxide dismutase (SOD1), an oxygen-inducible gene, in freshly isolated mononucleated VCT (2 h) and in vitro differentiated ST (72 h) from the two terms. Western blot analysis revealed that at 12–14 WA SOD1 is higher (∼2-fold after normalization to actin content) in both VCT and ST compared with trophoblasts from 8–9 WA. Thus, SOD1 was and remained increased in trophoblasts after isolation (2 h) and during the 72-h in vitro culture period in late first-trimester placentas corresponding to the entrance of the maternal blood flow and the increase of oxygen in the intervillous space. No significant variation was found between 2 and 72 h whatever the term. We also studied by immunoblotting the expression of hypoxia-inducible factor (HIF)-1α another oxygen-inducible and very labile protein, but it remained undetectable in our conditions.
We next investigated hCG contents in whole cell lysates from the same trophoblast primary cultures by Western blot analysis using A0231 and FBT11 antibodies. As shown in Fig. 3B, type 2 β-hCG was immunodetected in ST at both terms but not in VCT. In contrast, type 1 and type 2 β-hCG were immunodetected in both VCT (2 h) and ST (72 h) at 8–9 WA but only in ST at 12–14 WA. These results show that in vitro only the VCT from early but not from late first-trimester placentas have the ability to produce hCG. We next analyzed the secretion of total hCG and type 2 free β-hCG by trophoblasts during differentiation in vitro from VCT (24 h) to ST (72 h) by collecting cell culture media samples every 24 h. As shown in Fig. 3C, left, villous trophoblasts from both terms secreted significant increasing amounts of total hCG during the formation of the ST (P < 0.001), but hCG secretion was increased fourfold in 24-h VCT cultures from 8–9 WA (92 ± 19 ng/ml) compared with 12–14 WA (27 ± 5 ng/ml) (P < 0.01). Thus, during differentiation into ST, hCG secreted in culture media reached the same levels at 72 h independently from gestational age. However, hCG increased 25-fold in primary cultures from late placentas, whereas, in cultures from the early ones, only a sixfold increase was observed. These secretion results are in agreement with those obtained for intracellular hCG by Western blot (Fig. 3B, left). In contrast to total hCG, when type 2 free β-hCG was quantified in the same cell supernatants, no variation on its secretion was observed during the 72-h culture period at both term groups (Fig. 3C, right). However, secretion of free type 2 β-hCG was significantly higher at any stage of the culture in trophoblasts from 8–9 WA compared with 12–14 WA.
Immunodetection of total hCG and type 2 β-hCG in placental tissue sections from early and late first trimester.
Type 1 and type 2 β-hCG and dimeric hCG were immunodetected in situ with a polyclonal anti-hCG antibody (A0231). The monoclonal antibody (FBT11) was used for immunodetection of type 2 β-hCG. Figure 4 shows that type 1 and type 2 β-hCG were mostly located in the ST at both gestational ages but also present in mononucleated VCT from early first-trimester placenta sections, whereas they were almost undetectable in mononucleated VCT from late ones. The same distribution was observed with type 2 β-hCG. Isotypic controls for each antibody (mouse and rabbit IgG) used at the same concentrations showed no labeling.
Cell fusion index in villous trophoblast cultures from early and late first-trimester placentas.
Mononucleated VCT were isolated from 8–9 or 12–13 WA placentas and allowed to differentiate in vitro to aggregate and fuse. After a 72-h culture period, the number of syncyciotrophoblasts containing at least three nuclei and of mononucleated cells were estimated by the cellular distribution of nuclei and desmoplakin, and counting. The fusion index was determined as described in materials and methods and was 17% higher in early (0.75 ± 0.025) compared with late (0.59 ± 0.03; P < 0.05) trophoblast primary cultures (Fig. 5).
During the first trimester of pregnancy, a major consequence of arterial remodeling is the entry of maternal blood in the intervillous space. The mostly studied consequence is the rise of oxygen from 2.5% to reach ∼6–8% at the end of the first trimester (7, 8, 22, 23). Nevertheless, it is evident that the entry of maternal blood changes the entire environment of the trophoblast and not only the oxygen tension. The incidence of this important physiological event on the expression of β-hCG genes and the formation of the ST, which is the exchange and endocrine tissue of the human placenta, remains poorly known. In these studies, we showed for the first time that in vitro, ex vivo, and in situ, both type 1 and type 2 β-hCG genes are differentially expressed and located in VCT depending on gestational age, i.e., before or after the entrance of maternal blood flow in the intervillous space. We provided evidence that the mononucleated VCT, from early placentas, have the capacity to express and secrete hCG, whereas only the ST produced hCG in late first-trimester placentas, as well established in term placentas (11). Furthermore, hCG proteins and RNAs are more expressed in early than in late first-trimester placental tissues. These observations might contribute to the higher level of hCG in maternal serum during the first trimester (20).
β-hCG is encoded by a cluster of six genes, with the CGB5 and CGB8 genes being the most actively transcribed, contributing ∼60–80% to the total pool of β-subunit mRNA transcripts in the placenta (5, 28, 34). CGB3, CGB5, and CGB8 encode for the same protein (type 2) and CGB7 for a different protein (type 1). So far, the expression of these two genes during villous trophoblast differentiation in vitro and in whole placental tissue has never been investigated. We found that both type 1 and type 2 β-hCG gene expression increased during trophoblast differentiation in vitro and are higher in early first-trimester placentas. We also confirmed that CGB3, -5, and -8 is more expressed than CGB7 irrespective of the gestational age (5). Of particular interest is the higher expression of type 1 and type 2 CGB genes in the mononucleated cytotrophoblasts in early pregnancies. Factors involved in the upregulation of CGB gene expression in mononucleated cytotrophoblastes from the early first trimester remain to be determined as the subcellular location of type 1 β-hCG protein in human placenta, but specific tools are not commercially available so far. The biological function of type 1 β-hCG product in villous trophoblasts is unknown, but it may vary because of the different amino acid sequences, especially a prolin is replaced by a methionin, which may modify tertiary structure and/or glycosylation.
Glycosylation of hCG is known to modulate its biological activity. Hyperglycosylated hCG produced by human invasive trophoblasts stimulates the cell invasion process through a LH-CG receptor-independent pathway (13, 14). Abnormal glycosylation of hCG in trisomy 21-affected pregnancies is associated with a decreased stimulation of LH-CG receptor (9). In addition, glycosylation of hCG may play a role in its intracellular trafficking. The localization of the β-subunits of hCG in the VCT in early placenta suggests a possible role of hCG in the embryonic compartment.
Knowing the essential paracrine role of hCG in the formation of the ST through binding to LH/hCG receptor and activation of the cAMP cascade (17), the other interesting observation of our in vitro study is that, in early placentas, the cell fusion index is higher, reflecting the number of larger STs. We suggest that the production of hCG by mononucleated VCT from early first trimester might participate in the increased fusion process observed in vitro. We also observed that, in early first trimester, SOD1 expression was lower. Interestingly, SOD1, an oxygen-regulated gene (37), overexpression was shown to impair hCG secretion and ST formation (10).
In our study, human primary trophoblastic cells were isolated from different gestational age placentas and cultured under standard tissue culture conditions (20% O2), which is considered hyperoxic for in vivo conditions of gestation. Interestingly, we observed that, in cytotrophoblast primary cultures, expression of SOD1 and hCG genes remained differentially expressed between early and late first-trimester placentas even after a 72-h culture period under 20% O2. SOD1 protein levels were low all along the culture in 8–9 WA trophoblasts, suggesting that the isolated cells kept in vitro the memory of the in vivo cell state. We also observed that hCG gene expression remained elevated in trophoblast cultures from early placentas, at least during the two first days, similarly to what was found in chorionic villi.
Few studies have investigated gene expressions before and after maternal blood entry in the intervillous space. Expression of the immunoinhibitor CD274 was shown to be low in early first trimester and to increase around the onset of the second trimester (18). A different regulation in trophoblastic endocrine gland-derived (EG) vascular endothelial growth factor (VEGF) secretion by hCG via EG-VEGF receptors was also reported between early and late first-trimester placentas (6).
The rise of oxygen in the intervillous space is the most known consequence of maternal blood entry, so our observations raise the question of its role in the regulation of placental genes, including hCG. Most studies have investigated the role of hypoxia in vitro on hCG expression in term cytotrophoblasts or trophoblast cell lines and found a decrease in hCG production. Term villous trophoblasts cultivated under 8% O2 secrete less hCG and have a lower capacity to fuse in vitro than those cultivated under 20% O2 (1). Similar observations were reported using the choriocarcinoma cell line Bewo (24), with an even lower hCG secretion rate under 2% O2 (27). These in vitro results are in discrepancy with our ontogeny studies and suggest that not only the oxygen tension is involved. Similar results were observed for transthyretin, a carrier of thyroid hormone and vitamin A. It was shown to be expressed in the human placenta from 6 wk gestation and to rise during the first trimester at a time when placental oxygen tensions are also rising (30). However, the same authors showed that transthyretin expression was increased in vitro by hypoxia, suggesting that in situ observations cannot be directly related to the effect of oxygen (31).
Oxygen levels control gene expression through HIF-1-dependent or -independent pathways. HIF-1α and HIF-2α are constitutively expressed during the first trimester of pregnancy and might correspond to a relatively hypoxic environment at this term (12). Interestingly, it was shown in nontrophoblastic cells that these two factors are not similarly regulated by hypoxia, depending of severity and length of the hypoxic period (19, 38). In silico analysis of type 1 (CGB7) and type 2 (CGB3, CGB5, and CGB8) promoters revealed consensus response elements for HIF (hypoxia-responsive elements). Together, these observations are in favor for a role of oxygen levels in the regulation of hCG gene expression. However, our results are in discrepancy with those obtained with cell line or term trophoblasts cultured in hypoxic conditions. Indeed, our ex vivo and in vitro study demonstrates that villous trophoblasts from early first-trimester placentas, with low oxygen levels, produce more hCG and larger ST than those from late first-trimester tissues with higher oxygen levels. These observations suggest that the environment, not only oxygen levels, of the trophoblast in vivo may be essential for hCG expression. HIF was shown to be regulated by oxygen but also by nonhypoxic stimuli, including growth factors and hormones (33). Similarly, the gene expression profiling of the maternal-fetal interface reveals dramatic changes between the second and third trimester of pregnancy, whereas no dramatic changes in oxygen levels are observed (39). A recent study shows that gene expression patterns in trophoblasts cultured under 8 or 20% oxygen are not markedly different and did not fully capture expression patterns observed in vivo (29). The observations that the difference of physiological cellular microenvironment, before and after arrival of maternal blood flow in the intervillous space, is maintained in vitro suggest epigenetic regulation involving different methylation patterns (36) or chromatin condensation. Chromatin status will allow or not the accessibility to their binding sites of transcription factors involved in the regulation of β-hCG in villous trophoblast at term such as activator protein-2 and specificity protein-1 (26).
In conclusion, type 1 and type 2 β-hCG are produced and secreted by human villous trophoblasts and highly expressed, predominantly in mononucleated cells, before the entrance of maternal blood flow in the intervillous space, and this could be in part dependent on the oxygen levels.
This work was funded by Inserm and Université Paris Descartes and “la Caisse d'Assurance Maladie des Professions Libérales de Provinces” CAMPLP, 75578 Paris cedex 12, France.
No conflicts of interest, financial or otherwise, are declared by the authors.
Author contributions: M.C. and T.F. conception and design of research; M.C., S.B., N.S., J.G., P.M., and L.A.-C. performed experiments; M.C., S.B., N.S., and T.F. analyzed data; M.C., S.B., L.A.-C., and T.F. interpreted results of experiments; M.C. and S.B. prepared figures; M.C. and T.F. drafted manuscript; M.C., S.B., N.S., L.A.-C., D.E.-B., and T.F. edited and revised manuscript; M.C., S.B., N.S., J.G., P.M., L.A.-C., D.E.-B., and T.F. approved final version of manuscript.
We thank the Department of Obstetrics and Gynecology of Saint Vincent Hospital, Paris, France, for providing placental tissues. We are grateful to Audrey Chissey and Alicia Grosso for expertise in trophoblastic cell cultures and Sophie Richon for free type 2 β-hCG immunoassay.
- Copyright © 2012 the American Physiological Society