| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Pregnancy; |
CNRS UMR6247-GreD,3 Centre de Recherche en Nutrition Humaine d'Auvergne,4 Clermont Université, 63177 Aubière, France
Centro de Investigaciones Biologicas,5 Consejo Superior de Investigaciones Cientificas (CSIC), and Center for Biomedical Research on Rare Diseases (CIBERER), 28040 Madrid, Spain
ABSTRACT
Human implantation involves invasion of the uterine wall and remodeling of uterine arteries by extravillous cytotrophoblasts. Defects in these early steps of placental development lead to poor placentation and are often associated with preeclampsia, a frequent complication of human pregnancy. One of the complex mechanisms controlling trophoblast invasion involves the activation of the liver X receptor beta (or NR1H2, more commonly known as LXRbeta) by oxysterols known as potent LXR activators. This activation of LXRbeta leads to a decrease of trophoblast invasion. The identification of new target genes of LXR in the placenta could aid in the understanding of their physiological roles in trophoblast invasion. In the present study, we show that the endoglin (ENG) gene is a direct target of the liver X receptor alpha (NR1H3, also known as LXRalpha). ENG, whose gene is highly expressed in syncytiotrophoblasts, is part of the transforming growth factor (TGF) receptor complex that binds several members of the TGFbeta superfamily. In the human placenta, ENG has been shown to be involved in the inhibition of trophoblast invasion. Treatment of human choriocarcinoma JAR cells with T0901317, a synthetic LXR-selective agonist, leads to a significant increase in ENG mRNA and protein levels. Using transfection and electrophoretic mobility shift assays, we demonstrate that LXR (as a heterodimer with the retinoid X receptor) is able to bind the ENG promoter on an LXR response element and mediates the activation of ENG gene expression by LXRalpha in JAR cells. This study suggests a novel mechanism by which LXR may regulate trophoblast invasion in pathological pregnancy such as preeclampsia.
endoglin, gene regulation, human syncytiotrophoblast, implantation, liver X receptor, LXR, oxysterols, syncytiotrophoblast
The liver X receptors (LXRs) NR1H3 and NR1H2, more commonly known as LXRalpha and LXRbeta, respectively, belong to a subclass of nuclear receptors that bind to naturally occurring oxidized forms of cholesterol, known as oxysterols, and activate target gene expression [1]. LXRalpha expression is predominant in tissues showing high lipid metabolism, including liver, intestine, adipose tissue, adrenals, kidney, and testis, whereas LXRbeta is ubiquitously expressed [1]. They both bind DNA as obligate heterodimers with the retinoid X receptor (RXR) for the 9-cis retinoic acid. LXR/RXR classically binds with high affinity to DNA sequences called LXRE (liver-X-responsive element), composed of two direct repeats (DR) of a hexanucleotide motif separated by four nucleotides (DR4) [1]. Various studies, essentially performed with LXR-deficient mice, pointed out that LXRs were physiologically involved in the regulation of several metabolisms such as cholesterol and fatty acid metabolisms, inflammation, glucose homeostasis, epidermal differentiation, and fertility (for a review see [2]). Brain-penetrable LXR agonists or modulators have also been suggested as useful synthetic agents for the treatment and/or prevention of Alzheimer disease [3]. It is of interest to screen for new target genes in order to understand the physiological roles of LXRs as well as to prevent any side effects when the synthetic LXR modulators are available for the treatment of human diseases. Placenta is an organ in which oxysterols are detected at a high concentration range [4]. We previously determined that both LXRs were present in human placenta at various stages of gestation as well as in the human placental trophoblast cell lines JAR, JEG, and BeWo [5]. We hypothesized that LXRs could have some physiological roles during early human placentation and in fetal nutrition, especially concerning cholesterol supply to the fetus [5]. So far, only a potential role for LXR in the regulation of lipogenesis in the placenta and a possible involvement in differentiation of trophoblast cells have been suggested [6]. Other studies showed that the placental implantation process was modulated by LXR [7]. Moreover, several LXR target genes (apoliporotein E [APOE], sterol regulatory element binding transcription factor 1 [SREBF1], ATP-binding cassette A1/G1 [ABCA1/G1]) have been demonstrated to be expressed in the placenta and could provide a means to control delivery of maternal lipids to the fetus [8]. Altogether these data led us to suspect that LXR might play a crucial role in several mechanisms occurring in placental development and/or function. Identification of new target genes of LXRs in the placenta could help to understand their physiological functions in that organ and open a new field of investigation to understand some obstetrical pathologies, such as preeclampsia. We developed a rapid in silico screening method to identify new putative LXR target genes based on the presence of LXRE motifs in their promoters. Among the various genes identified in this screen, we focused our study on endoglin (ENG) because of its involvement in placental implantation [9], a process that was previously shown to be modulated by oxysterols and LXRs [7]. In this article, we show that the gene encoding endoglin is a direct target of LXRalpha in human syncytiotrophoblast cells.
In Silico Screening for LXR Target Genes
To identify putative LXR target genes in the placenta, we used the FastM computer program to associate the LXRE and SREBF1 position-specific scoring matrices in a virtual transcriptional model in order to screen the genomic databases with the ModelInspector program. The distance between the two elements ranged from 5 to 500 bp in any orientation. Using Bibliosphere software, bibliographical and gene ontology filters (biological process, tissues expression) were applied to identify genes that were expressed in placenta and whose products were involved in cholesterol, steroid, or lipid metabolisms. All the programs used for this in silico approach were included in the Genomatix package (Genomatix softwares, München, Germany).
Ligands 22(R) hydroxycholesterol (22(R)OH), 25 hydroxycholesterol (25OH), and 9-cis retinoic acid (9-cis RA) were purchased from Sigma-Aldrich (L'Isle D'Abeau, France). LG268 was obtained from Ligand Pharmaceutical (La Jolla, CA) and T0901317 was purchased from Cayman Chemical (Montigny le Bretonneux, France). All ligands were diluted in dimethyl sulfoxide (DMSO). Plasmids used in transfection assays (pCMX-LXRalpha, pCMX-LXRbeta, pCMX-RXRalpha, and pCMX) have been described elsewhere [10]. The plasmids pCMX-FXR and pSG5-PXR were generous gifts from Dr. Mangelsdorf [11] and Dr. JM Pascussi [12], respectively. The reporter construct pLuc–1950/+350 (or pLuc-proximal promoter) was derived from the human ENG promoter as described elsewhere [13]. Briefly, PCR was carried out in the presence of sequence-specific primers flanked by HindIII/XhoI sites for directional cloning and inserted in the reporter luciferase vector pXP2. The luciferase reporter containing ENG distal sequence (–5333 to –3535) was obtained by cloning this sequence into the reporter luciferase vector pGL3-basic vector (Promega, Charbonnieres, France). To amplify this sequence, PCR was carried out from genomic purified DNA of human blood cells in the presence of specific primers: (–5333Fw): 5'-GCGACCCCACACCTAATTT-3' and (–3535Rv): 5'-GCCACTGAAGGAATTTCAG-3'. The resulting 1799-bp fragment was cloned into pGEM-T easy (NcoI/SacI) and also in pGL3-basic (Promega) vectors. The various truncated fragments of this distal sequence were obtained by PCR amplification using the specific primers: (–4735Rv): 5'-TTCAAGGCTGCAGTAAGCTG-3'; (–4107Fw): 5'-GAATGTTCTGGAAGCCAAGG-3'; (–4971Fw): 5'-TAGTTGAGACGGGGTTTTGC-3'; (–4143Rv): 5'-ACTTCAGTGGCTGAGGCTGT-3' and (–4581Fw): 5'-CAACAGATGAGCTTGTGACGA-3'.
Cell Culture and Transfection and Reporter Gene Assays
Human placental JAR cells (HTB-1AA obtained from American Type Culture Collection, Rockville, MD) were cultured at 37°C in an atmosphere of 5% CO2 in RPMI medium containing 100 µg/ml gentamicin supplemented with 10% fetal calf serum (FCS). On Day 1 of transfection assay, JAR cells were seeded at 3 x 105 cells per well in six-well plates and allowed to adhere. Eight to ten hours later, cells were washed and transfected in serum-free medium with 1 µg per well of DNA (800 ng of luciferase reporter plasmids together with 100 ng of pCMX-RXRalpha and 100 ng of pCMX-LXRalpha/beta or pCMX-FXR or pSG5-PXR) using Metafecten (Biontex Martinsried, Planegg, Germany). On Day 2, cells were washed twice with 1x PBS, and 1 ml of serum-free medium without antibiotics containing DMSO or T0901317 (1 µM) was applied. Twenty-four hours later, cells were harvested for luciferase activity (Yelen, Ensues-la-Redonne, France). Luciferase activity was expressed relative to nontreated cells as mean ± SD. Statistical significance was assessed with the Student t-test.
RNA Isolation and Northern Blot Analyses
JAR cells were cultured in serum-free medium for 24 h prior to the addition of ligands. Cells were then incubated in the presence of T0901317 or DMSO (vehicle) for 1 to 24 h, and total RNA was isolated using TRIzol reagent (Invitrogen, Cergy Pontoise, France). Thirty micrograms of total RNA (per lane) was separated on 1% agarose/formaldehyde gels, transferred to nylon membranes (HybondN+; Amersham Pharmacia Biotech, Orsay, France) and fixed by heating at 80°C for 2 h. Random primed [
-32P] dCTP-radiolabeled probes were generated using the DNA labeling kit (Amersham Pharmacia Biotech). Human ENG cDNA probe (+629/+1069) was amplified by PCR using two specific oligonucleotides (Fw: 5'-AGAGGTGCTTCTGGTCCTCA-3' and Rv: 3'-AGTTCCACCTTCACCGTCAC-5'). Results were quantified in a phosphorimager (Biorad, Marnes la Coquette, France) and graphed relative to the ribosomal 18S.
JAR cells were cultured in serum-free medium for 24 h prior to the addition of ligands. Vehicle (DMSO) or various natural (22(R)OH, 25OH, 9-cis RA) or synthetic ligands (T0901317 or LG268) were added at 1 µM for 18 h. Total protein extracts were prepared from treated JAR cell cultures (4–5 x 106 cells). Briefly, cells were rinsed once with PBS and scraped in 100 µl lysis buffer (20 mM Hepes, 0.42 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, glycerol 25%, NP40 1% [v/v]) supplemented with protease inhibitors (PMSF, 1x Complete [Roche Molecular Biochemicals, Meylan, France]). After sonication, homogenates were centrifuged at 4°C for 10 min at 15 000 x g. Thirty micrograms of total protein was subjected to denaturing SDS-PAGE and transferred to nitrocellulose membrane (Hybond-ECL; GE Healthcare, Uppsala, Sweden). After 1-h saturation with Tris-buffered saline-Tween 0.2 % (v/v) containing 5% (w/v) of skimmed and dried milk, membranes were incubated overnight at 4°C with primary polyclonal antibodies raised against either ENG (1:500; BD Biosciences, Le Pont de Claix, France) or human anti ABCA1 (1:500; Novus Biologicals) or actin (1:2000; P.A.R.I.S, Compiègne, France), followed by 1-h incubation with a peroxidase-conjugated anti-rabbit IgG (1:10 000; P.A.R.I.S). Peroxidase activity was detected with the Western Lightning System (Perkin-Elmer Life Sciences, Courtaboeuf, France). Protein levels were expressed relative to nontreated cells as mean ± SD. Statistical significance was assessed with the Student t-test.
JAR cells were cultured in six-well plates (3 x 105 cell per well) on precoated micro coverslips with RPMI containing 10% FCS, glutamine, and gentamycine. Eight hours later, medium was replaced by serum-free RPMI medium without antibiotics for 12 h. Induction of cells was performed with 25OH, 22(R)OH, or T1317 at 1 µM for 18 h. Cells were then rinsed once with PBS, fixed for 10 min in 4% paraformaldehyde, and incubated in PBS with 2% BSA (w/v) for 1 h. Primary antibody toward ENG (1:200; DAKO, Trappes, France) was added for 1 h at room temperature in wet atmosphere. After several washes in PBS with 2% BSA, secondary fluorescein isothiocyanate-conjugated antibody (1:100; P.A.R.I.S) was added for 1 h at room temperature. Nuclei were counterstained with DAPI (4',6'-diamidino-2-phenylindole) coloration.
Electrophoretic Mobility Shift Assay
Total protein extracts were prepared from JAR cell cultures (5 x 106 cells). Briefly, cell monolayers cultured in 10-cm-diameter plates were rinsed once with PBS and scraped in 100 µl lysis buffer (20 mM Hepes, 0.42 M NaCl, 1.5 M MgCl2, 0.2 mM EDTA, glycerol 25%, NP40 1% [v/v]) supplemented with protease inhibitors (PMSF; 1x Complete). The cells were sonicated, and homogenates were centrifuged 10 min at 15 000 x g at 4°C. After centrifugation the supernatant was aliquoted and kept at –80°C until use. Total protein extracts were prepared from wild type, Lxr alpha–/–, Lxr beta–/–, or Lxr alpha beta–/– mice livers by crushing 1-g liver pieces with a Dounce (MC2, Clermont-Ferrand, France) in 1.5 ml high-salt buffer. The extracts were kept for 20 min on ice and centrifuged twice at 15 000 x g for 20 min. The supernatants were aliquoted and stored in a saline buffer (20 mM Hepes, glycerol 20%, 0.1 M KCl, 0.2 mM EDTA, 0.5 mM PMSF, 0.5 mM dithiothreitol [DTT], 1x Complete) at –80°C until use. Probes were labeled with T4 polynucleotide kinase and [
-32P dATP] (GE Healthcare) and purified (Qiaquick Nucleotide Removal Kit; QIAGEN, Courtaboeuf, France). Binding reactions were performed with 10 µg of protein extract and 0.05 pmol of each ENG-LXRE oligonucleotide (Table 1) or APOC1-LXRE oligonucleotide (5'-GCTGCCAGGGTCACTGGCGGTCAAAGGCAG-3') in the presence or absence of 50-fold molar excess of unlabeled competitor in a 1x binding buffer (20 mM Hepes, 0.2 mM EDTA, 100 mM NaCl, 100 mM KCl, 10 mM MgCl2, 8 mM spermidine, 4 mM DTT, 200 µg/ml BSA, and 8% Ficoll [Sigma-Aldrich]) and 2 µg poly(deoxyinosinic-deoxycytidylic) acid. The nonspecific competitor used in this assay was the mouse Sp1-specific sequence 5'-GGTAAGAGCCCGCCTCCTTTATCC-3'. Complexes were resolved by electrophoresis through 6% nondenaturing polyacrylamide gels in 0.5x Tris-borate-EDTA buffer. The gels were dried and autoradiographed.
|
Site-directed mutagenesis of ENG-LXRE4 motif was performed using Quick Change II Site-Directed Mutagenesis Kit (Stratagene, Amsterdam, Netherlands) according to the manufacturer's recommendations with the primer LXRE4mut (Table 1). The mutation of the selected and amplified plasmid was verified by sequencing before its use in transfection assays.
A Large 6-kb Region Upstream of the Transcription Initiation Site of the ENG Gene Contains Six Putative LXREs
Because several LXR target genes have been shown to be regulated by SREBF1 through binding sites (SRE) located in the vicinity of the LXRE sites, a pattern associating both LXRE and SRE matrices was designed (FastM; Genomatix). Genomic libraries were screened in silico using the 5'UTR and promoters subdivision database, and a list of several hundreds of genes was obtained. The list was refined by focusing on genes involved in cholesterol/oxysterol or lipid metabolism, atherosclerosis, or steroidogenesis. ENG was selected because of its expression in the placenta, including syncytiotrophoblasts [14], and its involvement in trophoblast invasion [9], a process partially inhibited by oxysterols [7]. Computer analyses identified six putative LXRE sites in the 6-kb region located upstream of the transcription initiation site (Table 1). LXRE1 and LXRE2 were located in the proximal promoter (–648/–632 and –1072/–1056, respectively), while LXRE3 to 6 were detected in a region ranging from –5333 to –3535 (or distal promoter). LXRE1, LXRE3, LXRE5, and LXRE6 were located within Alu sequences specific of primate genomes. Interestingly, these Alu sequences belong to the AluSx subfamily also found in the promoters of ABCA1 [15] and APOE [16] human genes, two LXR target genes. The presence of these very well-conserved LXRE sites in the upstream region suggested that ENG could be an LXR target gene in the placenta.
ENG Protein and mRNA Levels Are Increased by LXR or RXR Natural and Synthetic Ligands
Since ENG is highly expressed in human syncytiotrophoblasts [14] and JAR choriocarcinoma cells [17], the effects of natural (22(R)OH and 25OH) and synthetic (T0901317) LXR ligands as well as natural (9-cis RA) and synthetic (LG268) RXR ligands were tested in JAR cells for 18 h. Western blot analyses demonstrated that incubation with either LXR or RXR ligands induced a strong accumulation of ENG in treated cells compared to untreated cells (Fig. 1, A–C). LXR ligands induced a 1.8- to 6-fold increase of ENG levels, whereas RXR-ligands showed an increase of 4- to 5-fold. The simultaneous treatment with both LXR and RXR ligands resulted in a more effective stimulation of ENG expression (from 4- to 7-fold), as expected for these nuclear receptors [18]. As a control, ENG protein levels appeared to be increased by LXR upon ligand stimulation in a fashion similar to protein levels of ABCA1 (Fig. 1D), a well known LXR target [19]. The fact that ENG was highly accumulated upon treatment with T0901317 or natural LXR ligands was in agreement with immunohistochemistry experiments showing an accumulation of ENG in the cell membranes of JAR cells treated with 25OH, 22(R)OH, or T0901317 (Fig. 2). Given the optimal response obtained with the LXR agonist T0901317, this stimulus was selected for further experiments. In order to determine the level of regulation underlying the ENG protein accumulation, Northern blot analyses of ENG mRNA were performed in JAR cells treated with T0901317 for 1–24 h. As shown in Figure 3, treatment with the LXR agonist consistently increased ENG mRNA levels from 1.4- to 2-fold in a time-dependent manner. Taken together, these results show that LXR agonists lead to increased ENG mRNA and protein levels.
|
|
|
LXRE4 Specifically Mediates Regulation of ENG Gene Expression by LXRalpha
The transcriptional regulation of ENG gene expression via LXR was explored in transient transfection assays using various truncated portions of the ENG promoter together with expression vectors encoding LXRalpha and RXRalpha (Fig. 4). Since LXR-dependent effects on gene expression are mediated by LXRE motifs, these constructs were designed to contain one to four of the LXREs shown in Table 1. Thus, the ENG promoter constructs –1950/+350 (LXRE1 and LXRE2), –5333/–3535 (LXRE3, LXRE4, LXRE5, and LXRE6), –5333/–4735 (LXRE5 and LXRE6), –4107/–3535 (LXRE3), –4971/–4143 (LXRE4 and LXRE5), and –4581/–4143 (LXRE4) containing different combinations of LXRE motifs were used (Fig. 4A). Transfection experiments with the –1950/+350 promoter fragment revealed that LXRE1 and LXRE2 motifs were not responsive to T0901317. By contrast, the –5333/–3535 promoter fragment drove a 2.5-fold increase of the reporter gene activity in response to the LXR agonist, suggesting that at least one of the LXRE motifs (nos. 3–6) included in this fragment was functional. To pinpoint the functional element(s), deletions constructs –5333/–4735, –4107/–3535, –4971/–4143, and –4581/–4143 were used in similar experiments. Only LXRE4-containing fragments –5333/–3535, –4971/–4143, and –4581/–4143 showed a 2.5- to 2.8-fold induction of the luciferase activity in the presence of T0901317 (Fig. 4A). This strongly suggested that LXRE4 was the only functional motif, although RXR/LXR heterodimer was able to bind all the identified ENG LXRE sequences in electrophoretic mobility shift assay (EMSA; data not shown). To confirm that LXRE4 was involved in the transduction of the T0901317 signal, site-directed mutagenesis of this response element was performed. Transfection with the corresponding mutant of the –4581/–4143 reporter vector showed that the T0901317-dependent induction was abolished (1.2-fold versus 2.6-fold). Altogether, these data demonstrated that LXRE4 mediates the activation of the reporter gene expression by T0901317.
|
Because T0901317 is also a high-affinity ligand for the xenobiotic receptor pregnane X receptor (NR1I2) [20] and the farnesoid X receptor (NR1H4) [21], the regulation of ENG promoter by LXRalpha could not be definitely ascertained. To address this issue, transient transfections were performed with the –5333/–3535 fragment (Fig. 4A) and expression vectors for PXR or FXR together with RXR expression vector. Figure 4B shows that incubation with T0901317 did not activate the gene reporter via PXR or FXR, confirming the specific involvement of LXRalpha in the activation of ENG expression by T0901317. Among the various LXR target genes, some genes have been shown to either specifically respond to LXRalpha or LXRbeta [18, 22]. For this reason, we aimed to study whether ENG promoter activity could be induced by LXRalpha and/or LXRbeta in the human cell line JAR. As shown in Figure 4C, T0901317 failed to induce the expression of the reporter gene when LXRbeta was cotransfected together with RXRalpha, indicating that none of the putative LXREs, including LXRE4, was able to mediate LXRbeta activation of the reporter constructs. By contrast, upon cotransfection of LXRalpha with RXRalpha, T0901317 stimulated the promoter activity 2.5-fold. Altogether, we concluded that ENG was a target gene for LXRalpha in the human trophoblast cell line JAR.
The RXR/LXR Heterodimer Binds to ENG-LXRE4
EMSA experiments were performed in order to test the ability of the RXR/LXR heterodimer to bind ENG-LXRE4 sequence (Fig. 5A). A retarded complex was observed when radio-labeled ENG-LXRE4 oligonucleotide was incubated with liver nuclear extracts from wild-type mice (lane 1). Formation of this complex was abolished when liver nuclear extracts from Lxralpha beta –/– mice were used (lane 4) and was efficiently competed by addition of unlabeled ENG-LXRE4 oligonucleotide (S, lane 6) but not by addition of the nonspecific SP1 oligonucleotide (NS, lane 5). This indicated that the RXR/LXR heterodimer was able to bind to the ENG-LXRE4 sequence. In order to determine if both LXRalpha and LXRbeta isoforms were able to bind to the ENG-LXRE4 sequence, EMSA was also performed with liver nuclear extracts from Lxralpha–/– or Lxrbeta–/– mice. Complexes formed by nuclear extracts from both genetic backgrounds (lanes 3 and 2) were observed, though they were weaker than the complex formed by wild-type extract (lane 1). This indicated that both isoforms were able to bind to ENG-LXRE4 with similar affinity in vitro, although only LXRalpha was able to activate ENG promoter through LXRE4 in reporter assays. As a positive control, APOCI-LXRE oligonucleotide, known to bind the LXR/RXR heterodimer, was used as a labeled probe (Fig. 5B). As expected, a retarded complex was observed with JAR cells extracts and with liver protein extracts from wild-type mice (wt) but not with liver protein extracts from Lxralpha beta –/– (Lxrab–/–). This complex was efficiently competed by addition of unlabeled APOCI-LXRE (specific, S) but not by addition of the nonspecific SP1 oligonucleotide (NS). A similar pattern was observed with each of the ENG-LXREs 1, 2, 3, 5, and 6, though with different affinities (Fig. 5B). These results indicated that LXR/RXR was able to bind all the identified LXREs, although only one of them was able to activate ENG promoter.
|
LXRs are considered interesting putative pharmacologic targets to prevent and/or treat pathological lipid homeostasis disorders. Consequently, in order to identify and characterize all metabolic pathways in which LXRs are involved and to avoid any possible side effects of the LXR agonists used as therapeutic drugs, it is of great interest to develop a method for the rapid screening of new target genes. The in silico method used in this paper has proved to be reliable since ENG was identified as a LXR target gene. However, as shown by the fact that only one out of the six identified LXRE appeared to be functional, further experiments are absolutely needed to test the functionality of the putative LXREs and to confirm the identification of the gene as a bona fide LXR target.
Our present study identified ENG as a direct target gene of LXR. ENG (CD105) is a membrane cell protein that is part of the complex receptor for tumor growth factor-beta. It is highly expressed, among other organs, in the placenta, notably in syncytiotrophoblasts, which compose the first generation of a multinucleated trophoblastic layer during implantation [14]. Endoglin is also transiently up-regulated in extravillous trophoblasts differentiating along the invasive pathway [9]. A major function of the ENG/TGFbeta signaling pathway in the placenta is to control proliferation, migration, and invasiveness of normal human trophoblast cells [9]. In addition, ENG may be released from the cellular membrane into circulation as soluble endoglin (sENG). Plasma or serum levels of sENG are elevated in patients with atherosclerosis and correlate with total cholesterol levels [23].
In this paper, we showed that the expression of the ENG gene was up-regulated at both mRNA and protein levels by administration of T0901317 in human choriocarcinoma JAR cells. Among the six DR4 elements identified in silico as putative LXREs, we showed that, although they were all able to bind the LXR/RXR heterodimer, only one of them appeared to be a functional LXRE in transfection assays. We demonstrated that both the LXRalpha and LXRbeta isoforms were able to bind this functional LXRE in vitro, but that surprisingly, LXRbeta was unable to mediate T0901317 activation of a luciferase reporter gene. Although both receptors can bind the same DNA sequence in vitro, LXRalpha and LXRbeta can differentially regulate gene expression [24]. Our results show that in JAR placental cells where both are expressed [5], LXRalpha and LXRbeta clearly do not have overlapping roles, as observed in other tissues such as the liver [25] and uterus [26]. These data strongly support the existence of specific molecular mechanisms leading to LXRalpha transactivation, such as the involvement of specific cofactors for each isoform. Such factors remain to be discovered.
Several lines of evidence support the involvement of the LXR pathway in trophoblast biology: 1) trophoblast invasion is accompanied by an increased degradation of extracellular matrix proteins by members of the matrix metalloproteinases (MMPs) family [9], and expression of MMP9 is regulated in macrophages by a mechanism dependent on LXR activation [27]; 2) T0901317 and oxidized LDLs (oxLDL), rich in oxysterols, significantly reduce trophoblast invasion via a mechanism involving LXRbeta [7]; and 3) LXRbeta is able to modulate TGFbeta signaling [28]. It is unlikely that the LXR pathway is a primary regulator of trophoblast invasion since Lxr-null mice show no obvious defect in placental implantation (authors' unpublished observation). In this context, the physiological relevance of ENG regulation by LXR remains to be established. Nonetheless, other studies have shown that overexpression of ENG in mouse fibroblasts [29] or endothelial cells [30] led to decreased cell migration. Furthermore, it has been hypothesized that a lower ENG expression decreased cell adhesion and facilitated prostatic cell migration and invasion [31]. These studies thus establish that ENG inhibits migration and invasion of various cell types such as fibroblasts, endothelial cells, and prostate epithelium or prostate cancer cells. The present study demonstrates that ENG is up-regulated by LXR in choriocarcinoma cells. This up-regulation of ENG might inhibit trophoblast invasion. A previous study already linked trophoblast invasion and LXR [7]. These authors showed that treatment of extravillous cytotrophoblasts with T0901317 or oxLDL inhibited trophoblast invasion. They assumed that this effect was mediated by oxLDL internalization via LOX1 receptor (OLR1), whose expression was increased by oxLDL. Another study hypothesized that LXR inhibited the expression of MMP9, an enzyme implicated in extracellular matrix degradation, which is an essential process of normal trophoblast invasion in placental implantation [27]. Normal physiological placental invasion needs to be spatially and temporally regulated, reaching a peak at 12 wk of gestation and rapidly declining thereafter. Indeed, when extravillous trophoblasts reach the first third of endometrium, invasion needs to be stopped. This normal process is under the control of several inhibitory systems involving the TGFbeta receptor complex and its antiproliferative, antimigratory, and anti-invasive signals [32]. One of them specifically involves membrane ENG and its ligands TGFbeta 1 and TGFbeta 3 [9]. It leads to the inhibition of trophoblast outgrowth and migration by stimulating the expression of MMP9 inhibitors. In that way, invasion is stopped in the syncytium and also in the extravillous trophoblasts that have invaded maternal tissue and anchored the placenta. In our experiments, we demonstrated that membrane ENG was up-regulated by LXR. Knowing that normal pregnancy is always accompanied by an enhanced oxidation of LDL particles [33] and that, in preeclamptic pregnancies, there is evidence of abnormally enhanced production of lipid peroxides in maternal blood [34], one can hypothesize that an increased internalization of oxLDL in trophoblasts might activate LXR and increase membrane ENG level. This consequently would enhance the TGFbeta 1/3 inhibitory signal. This could slow down trophoblast invasion, resulting in precocious placental adherence disruption and enhancement of the preeclamptic symptoms. Indeed a shallow placentation with abnormal invasion of cytotrophoblasts and incomplete remodeling of uterine spiral arteries is observed in preeclampsia [35]. Interestingly, increased levels of sENG, secreted by the placenta, have been detected in the maternal circulation before the onset of preeclampsia. It has been postulated that this sENG plays a pathogenic role contributing to systemic endothelial dysfunction, resulting in hypertension and other systemic manifestations of preeclampsia (for a review see [36]). In summary, our data suggest that inappropriate expression and/or function of ENG that could result from up-regulation by LXR might contribute to major complications of pregnancy, such as preeclampsia.
ACKNOWLEDGMENTS
We thank Angelique de Haze for her excellent technical assistance, Dr. Pierre Val (UMR CNRS6245, Aubière, France) for helpful suggestions about mutagenesis and EMSA experiments and reading of the manuscript, Dr. D.J. Mangelsdorf (Howard Hughes Medical Institute, Dallas, TX) for providing the expression vectors for LXRs, RXR, and FXR, Dr. J.M. Pascussi (INSERM, UMR632, Montpellier, France) for providing the expression vector pSG5-PXR, and Dr. G. Veyssière and members of the Chester's lab for critically reading the manuscript. LG268 was kindly provided by Dr. M. Leibowitz (Ligand Pharmaceuticals Inc., San Diego, CA).
FOOTNOTES
1Supported by grants from the Centre National de la Recherche Scientifique, the Université Blaise Pascal, the Fondation Danone, the Fondation pour la Recherche Médicale INE2000-407031/1, and the Fondation BNP-Paribas. K.M. is a recipient of a doctoral fellowship from the Ministère de l'Education Nationale de Recherche et de la Technologie. C.B. is supported by grants from Ministerio de Educacion y Cienza of Spain (SAF2004-01390 and SAF2007-61827). 
Correspondence: 2Jean-Marc Lobaccaro, CNRS UMR6247 and Centre de Recherche en Nutrition Humaine d'Auvergne, Clermont Université, 24 avenue des Landais, 63177 Aubière, France. FAX: 33 473 407 042; e-mail: j-marc.lobaccaro{at}univ-bpclermont.fr
Received: 5 November 2007.
First decision: 26 November 2007.
Accepted: 7 February 2008.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
