| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Inserm,4 UMR 538, Faculté de Médecine Saint-Antoine,
Laboratoire de Biologie Cellulaire du Développement,5 CNRS, UMR7622, Université Pierre et Marie Curie, 75012 Paris, France
Unit of Developmental Genetics,6 Université Catholique de Louvain, B-1200 Brussels, Belgium
Department for Internal Medicine,7 Universitaetsklinikum Hamburg-Eppendorf, D-20246 Hamburg, Germany
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
cholesterol, cubilin, endocytosis, megalin, peri-implantation development
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although receptor-mediated endocytosis of the extracellular material appears to be essential, relatively little is known regarding the membrane receptors involved in nutrient uptake at the blastocyst stage. Cubilin, a 460-kDa multiligand endocytic receptor, is a key receptor for protein, vitamin, and lipid uptake [7, 8]. In the adult, it is localized at the apical pole of absorptive epithelia, including the ileum and the renal proximal convoluted tubule (PCT). In the ileum cubilin is the intrinsic factor-vitamin B12 receptor, and in the PCT it facilitates endocytosis of various proteins, including transferrin, albumin, and apolipoprotein A-I/high-density lipoprotein (apoA-I/HDL). In these sites, cubilin colocalizes and forms a complex with megalin, a 600-kDa protein member of the low-density lipoprotein (LDL) receptor gene family. The formation of the complex is essential for endocytosis of cubilin and its ligands because cubilin lacks a transmembrane domain [9]. Recent data suggest that megalin is not the only cubilin-interacting protein. Amnionless, a 50-kDa transmembrane protein, forms a high-affinity complex with cubilin in the adult PCT and ileum [10].
During development, cubilin is expressed in the visceral endodermal cells of the VYS of the rodent embryo together with amnionless and megalin and seems to be crucial for embryonic survival and normal growth [11, 12]. Administration of anti-cubilin antibodies to the pregnant rat induces in a dose-dependent manner embryonic resorptions or fetal malformations concerning essentially the rostral part of the embryo [11]. In vivo and in vitro, these antibodies disorganize the apical endocytic apparatus of the visceral endodermal cells [13] and probably interfere with uptake of essential maternal nutrients like cholesterol or lipophilic vitamins.
In this study, we hypothesized that cubilin was present and functional during the first stages of rodent embryonic development. Our results show that cubilin and megalin are first synthesized at the eight-cell stage and have very similar distributions in the pre- and early postimplantation embryo. Furthermore, we demonstrate that functional cubilin/ megalin endocytic complexes allow apoA-I internalization by the TE cells at the blastocyst stage, as well as cholesterol ester uptake in the VYS at early postimplantation stage, thus revealing the importance of cubilin for the nutrition of the peri-implantation embryo.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sheep and rabbit anti-megalin (1:500) and rabbit anti-cubilin (1:200 for immunocytological stainings of preimplantation embryos and 1:15 000 for immunohistochemistry) antisera were obtained as described elsewhere [14]. Rabbit anti-ezrin antibody was a gift from P. Mangeat (CNRS, UMR 5539, Montpellier, France) and was previously described [15]. Fluorescein isothiocyanate-conjugated anti-rabbit Ig, Cy-5-conjugated anti-sheep antibodies, and biotinylated secondary antibodies were from Jackson Laboratories (Jackson Immuno Research, Baltimore, MD). Blocking antibodies used were purified mouse monoclonal or rabbit polyclonal anti-cubilin (100 or 200 µg/ml) and rabbit polyclonal anti-megalin (200 µg/ml) and were characterized previously [11, 14]. Receptor-associated protein (RAP) was produced as a recombinant protein in Escherichia coli [16]. LDLs were kindly provided by J. Chapman (UMR 551, Paris, France). Apolipoprotein A-I (apoA-I) was provided by Dr. A. Kalopissis (INSERM U505, Paris, France) and conjugated to Alexa594 using the Alexa Fluor 594 Protein labeling Kit (Invitrogen, Cergy-Pontoise, France).
Doubly radiolabeled HDL (density = 1.063–1.21 g/ml) was prepared as described previously [17]. Briefly, apoE-deficient human HDL protein was labeled with [125I]-N-methyl-tyramine cellobiose-labeled apolipoprotein A-I ([125I]-NMTC-apoA-I). HDL-associated cholesteryl ester was labeled with [3H]-cholesteryl oleyl ether ([3H]-CE) (Amersham, UK). [125I]-NMTC and [3H]-CE are nondegradable intracellularly trapped tracers. Both of the tracers were incorporated in the HDL preparation by exchange procedures. The labeling procedure resulted in specific activities of 70–80 cpm/ng HDL protein for the protein tracer ([125I]-NMTC-apoA-I) and 12– 14 cpm/ng HDL protein for the lipid tracer ([3H]-CE). The protein content of the radiolabeled HDL was estimated by the method of Lowry.
Animals
Pregnant Wistar rats were from Harlan (France). Females OF1 mice were from Charles River (France). Mouse and/or rat embryos were analyzed from the two-cell stage to the eighth day postcoitum (dpc). Mouse embryos were used for the analysis of cubilin expression and function in the preimplantation period, whereas later stage analysis was performed on both mouse and rat embryos. This choice was made because of the better knowledge of mouse preimplantation development. All animal care and handling were performed according to institutional guidelines (Bureau for Experimental Animal research of Inserm and Department of Sciences de la Vie of CNRS) and to the French regulations on animal care during scientific experiments (décret no. 2001-464).
Recovery of Mouse Preimplantation Embryos
Nine- to 12-wk-old female OF1 mice were superovulated by intraperitoneal injection of 5 UI eCG (Intervet) and 5 UI hCG (Intervet) 48 h later. To obtain embryos, females were mated with OF1 males (ovulation/fertilization occurs about 12 h post-hCG). Mouse embryos were collected by flushing oviducts and were then cultured in T6 medium [18] supplemented with BSA under paraffin oil at 37°C in 5% CO2 in an air atmosphere. When necessary, zonae pellucidae were removed by a brief incubation in Tyrode acid solution [19].
Mouse Preimplantation Embryos Fixationand Immunocytological Staining
Mouse embryos were placed in specially designed chambers as previously described [20]. After centrifugation at 450 x g for 10 min at 37°C, samples were fixed in 3.7% formaldehyde in PBS for 30 min at 37°C, neutralized with 50 mM NH4Cl in PBS for 10 min, and postpermeabilized in 0.25% Triton X-100 in PBS for 10 min. Immunocytological staining was performed on fixed samples by incubation with the indicated antibodies in PBS/0.1% Tween 20/1.5% BSA for 1 h, followed by an incubation in the corresponding secondary antibody in PBS/Tween for 30 min. DNA was revealed by an incubation in propidium iodide (5 µg/ml) for 3 min. Samples were observed under a Leica TCS-SP confocal microscope. Omission of the primary antibodies gave no signal (data not shown).
Western Blot Analysis of Blastocysts and Yolk Sac
SDS-PAGE gel electrophoresis was performed with 4%–16% polyacrylamide gradient gels and 3% sodium dodecyl sulfate in the sample buffer. The proteins were electroblotted onto nitrocellulose membranes (Hyperbond C extra nitrocellulose; Amersham). After blotting, the membranes were blocked for 1 h in membrane buffer (MB; 20 mM CaCl2, 10 mM MgCl2, 100 mM Hepes, 1.4M NaCl, pH 7.8) containing 2% Tween and 4% low-fat dry milk, followed by washing three times for 25 min in MB containing 0.2% Tween. Subsequently, the blots were incubated with rabbit anti-rat cubilin antibodies diluted 1:1000 in MB containing 0.2% Tween, 2% low-fat dry milk, and 2 mM sodium azide overnight at 4°C. The blots were washed as described above and incubated for 1 h with alkaline phosphatase-conjugated secondary antibodies (donkey anti-rabbit IgG 1:7500; Promega, Madison, WI). After washing, NBT/BCIP color substrate (Nitro Blue Tetrazolium, 5-Bromo-4-Chloro-3-Indolyl-Phosphate) was added (Promega). As positive controls, Brown Norway rat yolk sac epithelial cells transformed with mouse sarcoma virus expressing mature cubilin and megalin were used [14].
Immunoelectromicroscopy
For immunoelectromicroscopy, preimplantation rat blastocysts were obtained 5 dpc by careful washing of the uterine cavity with 0.01 M periodate/0.075 M lysine containing 2% to 4% paraformaldehyde (PLP). Postimplantation embryos were obtained at 6 dpc by serial sectioning of uterine horns. Immuno-ultrastructural analysis was carried out on samples fixed with PLP, using pre-embedding techniques. Tissue slices prepared from uterine horns were immersed 1 h in PBS supplemented with 10% dimethyl sulfoxide. Small tissue blocks were frozen in isopentane. When required, 20-µm sections were cut in a cryostat from the tissue fragments and immediately immersed in PBS containing 1% BSA. Fibrin-embedded preimplantation whole blastocysts and uterine horns sections were incubated with 10 µg/ml of anti-cubilin mouse monoclonal antibodies. After extensive washing, sections were incubated for 5 h at 20°C with biotinylated sheep anti-mouse IgG followed by an incubation with peroxidase-labeled avidin at 20 µg/ml. Peroxidase activity was detected by incubation in diaminobenzidine chromogen (DAB, Sigma fast 3,3'-diaminobenzidine tablet sets) in Tris-HCl, 0.2 mM, pH 7.6, followed by H2O2 1:10 000 in the same solution for 5 min. After thorough washing, sections were fixed for 15 min with reduced osmium (150 mg potassium ferrocyanide, 3.75 ml 4% OsO4 for 10 ml final volume), dehydrated, and embedded in Epon 812. Semithin and thin sections were cut on a Reichert ultramicrotome. Ultrathin sections were examined on EM109 (Carl Zeiss) electron microscope. Controls for immunoperoxidase procedures included omission of the first antibody.
Immunohistochemistry
For light microscopy, rat and mouse embryos were collected, fixed in 75% ethanol, 2% formalin 40%, and 5% acetic acid and embedded in paraffin. Four-micrometer paraffin sections were placed on SuperFrost Plus glass slides (CML, France) and kept overnight (ON) at 50°C. The sections were dewaxed with toluene and rehydrated with distilled water through a series of alcohol solutions. Preliminary experiments showed that proteinase K treatment (10 µg/µl, 10 min at 37°C) was necessary to unmask epitopes for the proteins studied. The sections were then rinsed in TBS (0.15 M NaCl, 0.05 M Tris, pH 7.6) and incubated with blocking reagent 0.5% (from TSA Biotin System, Perkin Elmer) in TBS for 10 min at room temperature (RT), followed by incubation with the primary antibody diluted in TBS in a moist chamber at 4°C ON. A second goat anti-rabbit antibody was applied for 20 min at RT, followed by incubation with the biotinylated anti-goat antibody and avidin-horseradish peroxidase-labeled polymer. The sections were then exposed to a working solution containing the DAB chromogen (Sigma fast 3,3'-diaminobenzidine tablet sets) for 10 min at RT according to the manufacturer's instructions. Between each step, sections were rinsed three times for 5 min in TBS. The slides were counterstained with hematoxylin (Sigma), dehydrated through a series of alcohol solutions, and mounted in Eukitt (Labonord, Villeneuve d'Ascq, France).
Whole-Mount In Situ Hybridization
Mouse embryos collected at 6–8 dpc and rat embryos collected at 8– 10 dpc were fixed in 4% paraformaldehyde in PBS (pH 7.4) overnight at 4°C. The following day, they were washed in PBS with 0.1% Tween 20, dehydrated through a graded series of ethanol, and stored in 100% ethanol at –20°C. Whole-mount in situ hybridization (WM-ISH) was carried out using digoxigenin-labeled (Dig) riboprobes according to Wilkinson [21]. To generate sense or antisense cRNA probes, one microgram of plasmids containing the 3'-UTR or the CUB2 domain of cubilin cDNA were linearized using restriction enzymes, BamHI or XhoI. In vitro transcription was performed using the Roche Diagnostics kit and T7 or SP6 RNA polymerase in the presence of Dig-UTP. When necessary, embryos were subsequently sectioned on a vibratome, as previously described [22].
Uptake of ApoA-I-Alexa 594 by Mouse Blastocysts
Blastocysts devoid of zona pellucidae were incubated at 37°C in M2 [23] supplemented with BSA containing apoA-I-Alexa594 (30 µg/ml) for 10 min. After washing, embryos were either fixed and examined or incubated for a further 20 min at 37°C in M2 supplemented with BSA. For inhibition experiments, embryos were incubated 10 min with blocking antibodies (200 µg/ml) or RAP (1 µM) and then incubated 10 min in a solution containing the corresponding inhibitor and apoA-I-Alexa594 (30 µg/ml).
Uptake of Doubly Radiolabeled HDL by Rat Conceptuses
Cultures were carried out as previously reported [13, 24], with minor modifications. Wistar rat embryos were dissected free of the uterine wall on the 10th day of gestation (10 dpc) and placed after removal of the parietal layer of the yolk sac in round flasks containing 2 ml of heat-inactivated rat serum under constant rotation for 24 or 48 h, during which the concentration of O2 was progressively increased. After 24 or 48 h of culture (11 or 12 dpc, respectively), rat conceptuses (i.e., rat embryos surrounded by the visceral layer of the yolk sac) were removed from rat serum, washed in PBS, and put in flasks containing the incubation medium MEM/BSA 0.5% with various concentrations of doubly radiolabeled HDL. Three conceptuses were cultured for each experimental condition. For the time-dependence experiments, 11 dpc rat conceptuses were incubated for 1–3 h at 37°C under constant rotation in vials containing 30 µg/ml of doubly radiolabeled HDL. For dose-dependence experiments, 11 dpc rat conceptuses were incubated for 3 h with 1–60 µg/ml of doubly radiolabeled HDL. For inhibition experiments, we added indicated antibodies (100 or 200 µg/ml), RAP (1 µM), unlabeled HDL (400 µg/ml), and unlabeled LDL (200 µg/ml). Oxygen was added and the incubation was carried out under rotation for 3 additional h. At the end of the incubation period, the conceptuses were removed from the flasks, placed in PBS at 37°C, and examined to confirm normal development. All the conceptuses used had an active vitelline circulation. The conceptuses were washed three times in PBS/BSA 0.5% and twice in PBS alone. The visceral layer of the VYS was separated from the embryo and both the VYS and the embryo were dissolved in NaOH 0.1 N and sonicated. Aliquots were used for determination of [125I], [3H] (after lipid extraction) radioactivity, and protein concentration as above. The average protein content was 25–30 µg per VYS and around 35 µg per embryo after 24 h of culture (11 dpc). After 48 h of culture (12 dpc), the protein content was around 68 µg and 154 µg, respectively. To facilitate the quantitative comparison, the uptake of both radiolabels is shown as micrograms of HDL protein necessary to deliver the measured amount of internalized tracer as suggested by Pittman et al. [25]. The HDL cholesterol uptake is expressed as micrograms of HDL protein per milligram cell protein. Equal amounts of internalized tracers represent HDL endocytosis (i.e., holoparticle uptake).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Expression of cubilin in the peri-implantation embryo was studied using immunohistological, biochemical, and in situ hybridization techniques on both mouse and rat embryos. We first performed immunofluorescence and immunoperoxidase experiments on preimplantation embryos. Figure 1 shows cubilin expression from the two-cell stage to the blastocyst stage of the mouse embryo. No expression was found in the two-cell- (Fig. 1A) and four-cell-stage (Fig. 1B) embryos. Cubilin was first detected in the cytoplasm at the eight-cell stage (Fig. 1C) and in some, but not all, outer cells of the morula (16-cell stage) (Fig. 1D). The localization of the stained structures most likely corresponded to the endoplasmic reticulum (ER) (Fig. 1D'). At the 32-cell stage, all outer cells of the embryo, which are at the origin of the TE, were stained, whereas the internal cells were not stained. During the blastocyst stage (Fig. 1E), cubilin was predominantly localized at the apical membrane of the TE cells, which separate the internal cell mass cells from the maternal environment and constitute a characteristic epithelium with apical membranes facing outward. Megalin had a very similar distribution pattern at this (Fig. 1F) and all the previous stages (not shown). Immunoperoxidase electron microscopy confirmed the exclusive distribution of cubilin in TE cells. The discontinuous staining of the apical membrane most likely corresponds to clathrin coated pits and apical vesicles (Fig. 2A, arrowhead and arrow). In agreement, a single polypeptide of 460 kDa corresponding to mature cubilin was detected at the blastocyst stage by Western blot analysis (Fig. 3A).
|
|
|
At the next stage analyzed (4.5 dpc mouse, 6 dpc rat), primitive endoderm cells were positive for both cubilin and megalin. Immunoperoxidase electron microscopy (Fig. 2B) revealed cubilin staining in apical membrane invaginations, probably representing coated pits, in endocytic vesicles (Fig. 2B, arrow), ER (Fig. 2B, arrowhead), and Golgi apparatus. At the egg cylinder stage (6 dpc of the mouse, 7– 8 dpc of the rat embryo), the primitive endoderm has differentiated to visceral and parietal endoderm, which were immunoreactive with both anti-cubilin (Fig. 4B) and anti-megalin (data not shown) IgG. In both mouse and rat, there was no preferential localization to either the posterior or anterior side of the embryo. Staining was most striking in the columnar epithelium cells of the VE and persisted throughout development of the embryo. The typical apical staining is illustrated at 10–11 dpc of the rat embryo (Fig. 4, H and J). Confocal light microscopy demonstrated the overlapping expression of cubilin and megalin (Fig. 4J). From the early headfold stage (7.5 dpc of the mouse, 8–9 dpc of the rat embryo), ectodermal cells lining the proamniotic cavity (arrowheads in Fig. 4, B and E) were also positive for both cubilin and megalin. Cells of the emerging definitive endoderm were stained (not shown). No signal was detected at the node or the primitive streak. Trophoblastic cells and allantois were consistently negative. At 10 dpc of the rat and 8.5 dpc of the mouse, cubilin was also detected in the newly forming neuroepithelium (Fig. 4H, arrowhead). In these sites, the distribution of cubilin was identical to that of megalin (not shown) and concerned the apical pole of the neuroepithelial cells. Western blot analysis (Fig. 3B) confirmed the presence of cubilin in the VYS. In addition, the immunomorphological data were supported by WM-ISH in both mouse and rat embryos (at 8, 9, and 10 dpc of the rat and 6.5, 7.5, and 8.5 of the mouse) (Fig. 4, C, F, and I, respectively). Cubilin mRNA was detected in both the distal (extraembryonic) and proximal (embryonic) VE, cells although the signal was stronger in the distal VE (Fig. 4, C and F). Later, mRNA was also detected in the VYS and in the neuroepithelial cells (Fig. 4I).
|
Cubilin-Mediated Apolipoprotein A-I Uptakeat the Blastocyst Stage
The localization of the endocytic receptors cubilin and megalin at the apical membrane of the TE suggests interaction with components of the maternal environment, i.e., nutrients present in the endometrium and uterine glands. To test if cubilin expressed at the blastocyst stage was functional, we analyzed uptake of apoA-I, which is a constituent of the uterine fluid [3] and a high-affinity ligand for cubilin [26]. Purified fluorescent mouse apoA-I (apoA-I-Alexa594, 30 µg/ml) was incubated at 37°C with 64-cell-stage embryos devoid of zona pellucidae. After a short period (10 min) followed by extensive washing, apoA-I-Alexa594 was mainly found at the apical membrane of TE cells, where it colocalized with ezrin, an apical membrane marker of TE cells [15], and in some vesicles inside the cells (Fig. 5, A and B). In contrast, when the 10-min incubation with apoA-I-Alexa594 was followed by a 20-min chase at 37°C in medium devoid of ligand (Fig. 5C), apoA-I was detected in intracellular vesicles, distinct from the apical surface indicating internalization of apoA-I (Fig. 5C).
|
To test if the uptake of apoA-I was due to cubilin and see if megalin was involved in the process, we incubated embryos with blocking anti-cubilin (Fig. 5D) or anti-megalin antibodies (Fig. 5E) or RAP (Fig. 5F), a universal megalin inhibitor that also blocks apoA-I binding to cubilin [26] before the addition of apoA-I-Alexa594. Anti-cubilin antibodies, monoclonal or polyclonal, completely abolished apoA-I uptake, confirming that cubilin in the TE can bind and mediate endocytosis of apoA-I. Only a weak staining with apoA-I-Alexa594 was still observed after addition of the anti-megalin antibody (Fig. 5E), suggesting that, at the blastocyst stage, before implantation, cubilin endocytosis largely depends on megalin. Addition of RAP also completely blocked apoA-I endocytosis, confirming the roles of both cubilin and megalin in the process. Nonimmune IgG polyclonal or monoclonal did not interfere with apoA-I-Alexa594 uptake (not shown).
Cubilin-Mediated Uptake of HDL Cholesterol by the VYS Ex Vivo
It is established that maternal lipoproteins, mainly HDL, in rodents are a significant source of cholesterol for the developing embryo [27]. To quantify the amount of HDL cholesterol internalized by the VYS and assess the role of cubilin in this process, we followed the uptake of HDL radiolabeled on both the protein ([125I]-NMTC-apoA-I) and the cholesterol ester ([3H]-CE) components by rat conceptuses (embryos surrounded by the VYS) in ex vivo culture. Both tracers are not degradable and their accumulation in the VYS cells directly reflects the amounts internalized [17]. This culture model was used previously to demonstrate the teratogenic effects of the anti-cubilin antibodies [13].
HDL cholesterol uptake, according to [125I]-NMTC-apoA-I (solid bars) or [3H]-CE (open bars), was very efficient and increased almost linearly with time (Fig. 6A) or the HDL concentrations (Fig. 6B) used. Uptake was completely inhibited by an excess of unlabeled HDL (Fig. 6C, lane 2). It was mainly accounted for by endocytosis of HDL ([125I]-NMTC-apoA-I). The selective uptake did not exceed 10%–15% of the total cholesterol ester uptake ([3H]-CE) regardless of the concentration used. Similar results were obtained when HDL uptake was assessed after 48 h of ex vivo culture (equivalent to 12 dpc) (data not shown). As expected, under the present experimental conditions, the nonhydrolyzable tracers remained in the VYS and only trace amounts of radioactivity could be detected in the embryo (data not shown).
|
To analyze the role of cubilin and megalin in HDL uptake, we incubated the conceptuses with 30 µg/ml of doubly radiolabeled HDL for 3 h in the presence of relevant antibodies, RAP and LDL. Figure 6C, lane 3, shows that saturating concentrations of anti-cubilin antibody almost completely inhibited (80%) the uptake of both tracers. A set of experiments in which a lower amount of anti-cubilin antibody (100 µg/ml) was used resulted in lower inhibition of HDL uptake around 55% (data not shown). Anti-megalin antibodies had an inhibitory effect of around 50% (Fig. 6C, lane 4) whereas nonimmune IgG had no effect on the HDL uptake (Fig. 6C, lane 5). RAP, which inhibits ligand binding to both cubilin and megalin [28], also inhibited the uptake of HDL (Fig. 6C, lane 6). LDL, a ligand for megalin [29], did not alter the HDL uptake at a concentration of 200 µg/ ml (Fig. 6C, lane 7). We observed the same inhibition pattern for both of the developmental stages tested, 11 and 12 dpc (data not shown), suggesting that the cubilin/megalin complex is responsible for HDL cholesterol uptake in the developing VYS.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Our results show that cubilin, together with megalin, is expressed earlier than initially reported during embryonic development and most remarkably in structures that interface the mother and the fetus. Cubilin expression is first detectable in the biosynthetic apparatus at the eight-cell stage. During the next three rounds of division [30], the outer blastomeres differentiate to form a functional epithelium. At the 16-cell stage, tight junctions form between polarized outer cells, and cubilin first becomes restricted to the outer blastomeres and furthermore becomes detectable in the plasma membrane and endosomal-like structures. In the fully developed cavitated blastocyst, the TE cells form an outward-facing tight epithelium. These cells express cubilin on their apical membrane. As the embryo prepares for implantation, the primitive endoderm, which appears as a layer of cells on the blastocoelic surface of the ICM, is characterized by a high expression of both cubilin and megalin in membrane-coated pits and endosomes. The primitive endoderm rapidly differentiates into parietal and visceral endoderm, which express cubilin, megalin, and amnionless [11, 12]. The VE, which will become the VYS, is a monolayer of polarized cells, characterized by their high endocytic and degradative capacities [31]. It is indirectly exposed to the maternal circulation and constitutes the only functional interface between mother and fetus before placentation [5].
The VE is of major importance not only for embryonic metabolism but also for hematopoiesis, steroidogenesis, and a source of signals for neural induction, which begins in the anterior visceral endoderm [31]. As megalin can bind the N-terminal part of sonic hedgehog in vitro [32] and amnionless may interact with members of the bone morphogenetic proteins family (BMPs) [12], it is possible that these receptors are involved in neuroepithelial differentiation.
Cubilin-Mediated Endocytosis of Nutrientsin the Preimplantation Embryo
Prior to implantation, nutrition of the human and rodent embryos is essentially histiotrophic [33]. The secretions from the oviduct and the endometrium, which accumulate between the maternal and fetal tissues, contain a wide variety of proteins, including known ligands of cubilin such as albumin, transferrin, apoA-I, lipids, and sugars, which can be endocytosed and degraded by the TE of the blastocyst [4]. Small apoA-I-containing lipoproteins are also very likely to be present. Tracer studies have demonstrated that these proteins are a source for the amino acids necessary to cover the increased biosynthetic and developmental demands of the embryo [5, 24]. How these proteins become internalized is not known, but the results presented here suggest that their endocytosis is cubilin and megalin dependent. Both receptors are localized along the apical endocytic apparatus of the TE cells and are directly exposed to their ligands. We followed internalization of apoA-I, a high-affinity ligand for known cubilin [26]. Indeed, the only other known apoA-I interacting protein, scavenger receptor class B, type I, is not expressed at this stage [34]. In agreement with previous studies [26], monoclonal and polyclonal anti-cubilin antibodies inhibited apoA-I uptake in a concentration-dependent manner. Similarly, 100-fold excess of unlabeled apoA-I (not shown) or 1 µM of RAP, an inhibitor of apoA-I/cubilin/megalin interaction [26, 28], completely blocked endocytosis, confirming the specificity of apoA-I binding to cubilin. A strong inhibitory effect of anti-megalin antibodies suggests that megalin plays a role in this process, as previously demonstrated in the renal tubule [9]. Amnionless, a recently described partner of cubilin [10], is unlikely to be necessary for cubilin function in the blastocyst because it is first detected after implantation in the primitive endoderm [12]. It is thus possible that, until the blastocyst stage, megalin is the only molecular partner of cubilin. In addition to their role in the histiotrophic process, cubilin and megalin may have additional functions, such as blastocyst implantation, because uteroglobin, a powerful regulator of trophoblast proliferation and invasiveness, is a cubilin ligand [35].
Cubilin-Mediated Endocytosis of Nutrients in the Early Postimplantation Embryo
In rodents after implantation, the visceral layer of the yolk sac becomes crucial for protein and lipid endocytosis. Both are important for the rapidly dividing cells of the fetus because a deficient protein or lipid supply are not compatible with normal growth. Cubilin-mediated protein endocytosis by the VYS has been the object of many studies [13, 14], and we focused here on the implication of cubilin in the uptake of another essential nutrient, cholesterol. In the early postimplantation embryo, endogenous synthesis cannot provide the considerable amounts of cholesterol needed and thus maternal-fetal transport becomes essential [36]. Significant amounts of cholesterol are taken up as HDL, but controversies still remain concerning the mechanism of HDL internalization. Studies in the hamster using doubly labeled LDL and HDL suggest the existence of receptor-dependent and -independent processes [37], while a study using fluorescent DiI labeled HDL suggests that degradative endocytosis is at work [38]. However, although this approach is an elegant way to follow internalization of the protein/phospholipid constituents of HDL, it is not directly informative for HDL cholesterol internalization. To quantify HDL internalization by the VYS, we used HDL radiolabeled on both the cholesterol ester and apoA-I with nondegradable tracers. The uptake of similar amounts of apoA-I and cholesterol ester reported here clearly show that the main mechanism of cholesterol uptake by the VYS is endocytosis of HDL. A small part of cholesterol (15% of the total) is internalized independently of any protein, suggesting that, like in the hamster VYS, both endocytosis and selective lipid uptake operate in the rat VYS. Endocytosis of HDL is strongly inhibited by anti-cubilin antibodies, unlabeled HDL and RAP, and partially by anti-megalin IgG, showing that cubilin is the HDL receptor in the VYS.
Although not directly investigated, the presence of cubilin and megalin on the neuroectodermic cells facing the amniotic cavity may suggest internalization of lipoproteins of the amniotic fluid, which contains apoA-I (personal data). Such a possibility is in line with observations in mice invalidated for megalin, which present a holoprosencephaly possibly associated with a deficient cholesterol supply [39]. If the cubilin/megalin complex functions in the neuroectoderm like in the TE, it is very likely that invalidation of megalin affects cubilin function and impairs endocytosis of apoA-I-containing lipoproteins. It is interesting to note that impaired lipoprotein uptake also affects uptake of lipophilic vitamins such as vitamin E, A, or D [40], which are critical for normal embryonic growth.
In conclusion, the data presented here indicate that cubilin plays an important role during mammalian peri-implantation development. Its trophic function is essential for normal growth. The interactions of cubilin with megalin and amnionless, potential receptors for developmental important proteins, such as sonic hedgehog or members of the BMPs [12], may suggest that it is part of oligomeric complexes involved in signal transduction pathways.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence: Renata Kozyraki, INSERM UMR 538, 27 rue de Chaligny, 75012 Paris, France. FAX: 33 1 400 11 390; renata.kozyraki{at}chusa.jussieu.fr
3 E.A. and S.V. participated equally in the work
Received: 8 October 2004.
First decision: 29 October 2004.
Accepted: 7 December 2004.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  T. A. Pagler, S. Golsabahi, M. Doringer, S. Rhode, G. J. Schutz, M. Pavelka, C. Wadsack, M. Gauster, A. Lohninger, H. Laggner, et al. A Chinese Hamster Ovarian Cell Line Imports Cholesterol by High Density Lipoprotein Degradation J. Biol. Chem., December 15, 2006; 281(50): 38159 - 38171. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Kurimoto, Y. Yabuta, Y. Ohinata, Y. Ono, K. D. Uno, R. G. Yamada, H. R. Ueda, and M. Saitou An improved single-cell cDNA amplification method for efficient high-density oligonucleotide microarray analysis Nucleic Acids Res., March 17, 2006; 34(5): e42 - e42. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Yoshida and Y. Wada Transfer of maternal cholesterol to embryo and fetus in pregnant mice J. Lipid Res., October 1, 2005; 46(10): 2168 - 2174. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
