| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Research Article |
Prince Henry's Institute of Medical Research,3 Clayton, Victoria 3168, Australia
Department of Obstetrics and Gynecology,4 Monash University, Clayton, Victoria 3168, Australia
ABSTRACT
Human embryo implantation is a complex process involving blastocyst attachment to the endometrial epithelium and subsequent trophoblast invasion of the decidua. Chemokines, critical regulators of leukocyte migration, are abundant in endometrial epithelial and decidual cells at this time. We hypothesized that endometrial chemokines stimulate trophoblast invasion. Chemokine receptors CX3CR1 and CCR1 were immunolocalized in human first-trimester implantation sites, specifically to endovascular extravillous trophoblasts, but not to the invading interstitial EVTs (iEVTs), with weak staining also on syncytium. CCR3 was localized to invading iEVTs and to microvilli on the syncytial surface. Expression of CX3CL1 (fractalkine), CCL7 (MCP-3), and their receptors (CX3CR1, CCR1, CCR2, CCR3, and CCR5) mRNA was examined in cellular components of the maternal-embryonic interface by RT-PCR. Both chemokines were abundant in entire endometrium and placenta, endometrial cells (primary cultures and HES, a human endometrial epithelial cell line) and trophoblast cell lines (JEG-3, ACIM-88, and ACIM-32). Chemokine receptor mRNA was expressed by placenta and trophoblast cell lines: CCR1 by all trophoblast cell types, whereas CCR2, CCR3, and CX3CR1 were more variable. CX3CR1, CCR1, CCR2, and CCR5 were also expressed by endometrial cells. Migration assays used the trophoblast cell line most closely resembling extravillous cytotrophoblast (AC1M-88). Trophoblast migration occurred in response to CX3CL1, CCL14, and CCL4, but not CCL7. Endometrial cell-conditioned media also stimulated trophoblast migration; this was attenuated by neutralizing antibodies to CX3CL1 and CCL4. Thus, chemokines are expressed by maternal and embryonic cells during implantation, whereas corresponding receptors are on trophoblast cells. Promotion of trophoblast migration by chemokines and endometrial cell conditioned medium indicates an important involvement of chemokines in maternal-fetal communication.
cytokines, implantation, placenta, pregnancy, trophoblast
INTRODUCTION
Human embryo implantation is a complex process, requiring synchronous development of a receptive endometrium and an activated blastocyst. The human endometrium becomes receptive to blastocyst implantation 6–8 days following ovulation, and remains so for around 4 days known as the "implantation window." Embryo implantation involves apposition of the hatched blastocyst to the endometrium, followed by attachment to and penetration of the epithelium and subsequent invasion into the underlying decidual tissues. Cross-communication between maternal and fetal cells is essential for successful implantation. Because no direct contact occurs during the apposition phase, communication must be in the form of soluble mediators such as cytokines and growth factors. Subsequently there is direct contact between endometrial epithelial cells (EEC) and the trophectoderm during the adhesion phase [1] and between trophoblast and decidual cells once the epithelium is penetrated.
Following syncytium formation, cytotrophoblast cells penetrate through the syncytium; some remain at the tips of the villi as a solid core, known as the cytotrophoblast cell column (CC), which anchors the placental villi to the decidua. These columns spread laterally and form the cytotrophoblast shell, from which invading extravillous cytotrophoblast cells (EVTs) originate. A subpopulation of EVTs, interstitial EVTs (iEVTs), specifically migrates through the decidua and home to the vicinity of maternal spiral arteries, where it has remarkable remodeling effects. Endovascular EVTs (vEVTs) invade the arteries at the point where the shell contacts the opening of the maternal vessels and aggregate, forming plugs [2, 3], as summarized in Figure 1. Extensive communication between trophoblast subtypes and endometrial cells is required for the entire invasive process and hence for successful implantation.
|
Chemokines, a family of small chemotactic cytokines, are best known for their functions in leukocyte trafficking. However, roles in reproductive events such as ovulation, menstruation, embryo implantation, and parturition, and disease conditions like endometriosis and cancer, have more recently been recognized [4–7]. Similarities between trophoblast invasion and leukocyte chemotaxis suggest that common systems may be in play during these events [8, 9]. Trophoblast invasion is an integrin-dependent process [3] that proceeds in the presence of a specific chemokine-cytokine-rich microenvironment, derived from the endometrial cells, particularly the epithelium and decidua [4, 10]. Certain chemokine receptors have been identified both on human blastocyst and on trophoblast cells [11, 12]. Collectively, these findings suggest that similar mechanisms to those used for leukocyte migration may be important during implantation and placentation. Our recent identification of chemokines maximally produced at the time of implantation and during early pregnancy, predominantly by decidualized stroma and epithelial cells [4, 13], further indicates a role for chemokines during early pregnancy. CX3CL1, CCL7, CCL14, and CCL4 were among the most abundant chemokines produced during the midsecretory phase; a possible function during implantation could therefore be proposed for the above chemokines. In this study, we investigated the expression and cellular localization of the receptors for these chemokines (CCR1, 2, 3, 5, and CX3CR1) in human first-trimester implantation sites and their expression in related cell lines. After careful selection of a trophoblast cell line, whose chemokine receptor phenotype closely approximated the in vivo situation, the ability of the above chemokines to stimulate trophoblast migration was assessed in vitro. A potential stimulatory role for endometrial secreted products (epithelial and decidual) in trophoblast invasion was also examined in the absence and presence of neutralizing antibodies against selected chemokines. These studies provide further evidence for an important role for endometrially-derived chemokines in promoting and directing trophoblast migration during implantation.
MATERIALS AND METHODS
Tissue Collection and Patient Details
Ethical approval was obtained from appropriate Institutional Ethics Committees for all tissue collections and written informed consent was obtained from all subjects.
Nonpregnant and Pregnant Human Endometrium. Human endometrial biopsies were obtained during the proliferative and secretory stages of the menstrual cycle from normal fertile women undergoing curettage following laparoscopic sterilization or assessment of tubal patency. Patients with uterine abnormalities such as leiomyomas, endometrial polyps, or endometriosis, or those who had received steroid hormone therapy in the last 6 mo, were excluded. Cycle stage was confirmed by histological dating, according to standard criteria [14]. First-trimester implantation sites (containing decidua and placenta [PL]; n = 6) were collected by curettage before termination of pregnancy (gestation 8–12 wk) by vacuum aspiration. Tissue was collected in 10% buffered formalin at 4°C overnight (for wax embedding), RNA later (Ambion, Inc., for subsequent RNA extraction), or in a 1:1 mixture of Dulbecco's minimum essential medium/Hams F12 (DMEM/F12; Trace Biosciences; for tissue culture).
Immunohistochemistry
CX3CL1 receptor (CX3CR1) was localized using rabbit polyclonal anti-human CX3CR1 (ab7201; Abcam Limited), as described previously [13]. In brief, (n = 6) serial paraffin sections (2µm) were dewaxed in Histosol (Sigma Chemical Co.) and rehydrated through descending grades of alcohol to dH2O. Sections were microwaved at high power (1000 W) in 0.01 mol/L sodium citrate buffer (pH 6.0) for 2 x 5 min, incubated in the hot buffer for a further 20 min. Endogenous peroxidase activity was quenched with 3% H2O2 in dH2O for 10 min at room temperature. Nonspecific binding was blocked by incubation with a nonimmune block (10% normal goat serum; 2% normal human serum). Primary antibody was applied at 2 µg/ml overnight (16–18 h) at 4°C, biotinylated swine anti-rabbit IgG (E353; DAKO) was applied for 30 min, followed by avidin/biotin conjugated with horseradish peroxidase (DAKO). Positive localization of CX3CR1 protein was identified by the application of the peroxidase substrate 3, 3'-diaminobenzidine (DAKO). Tissue sections were counterstained with Harris' hematoxylin, dehydrated through ascending grades of ethanol, and mounted. Negative controls were included for each section in which rabbit immunoglobulin (DAKO) was substituted at matching concentration to the primary antibody.
CCR1 protein localization was examined using rabbit polyclonal anti-human CCR1 (ab1681; Abcam Limited). The immunostaining protocol was identical to that used for CX3CR1 except for the primary antibody.
CCR3 was localized using a rabbit polyclonal antibody kindly provided by Dr. B. Daugherty (Merck Research Laboratories). The immunostaining protocol (described in detail in [15]) was similar to that for CX3CR1 with the following exceptions; no antigen retrieval was necessary for detection, normal goat serum was used to block nonspecific binding, and biotinylated goat anti-rabbit IgG (Vector Laboratories Inc.) was used as the secondary antibody. Negative controls were included for each tissue section.
Immunohistochemistry for cytokeratin and human leukocyte antigen-G (HLA-G) was conducted on (n = 6) serial (2 µm) sections of first-trimester human implantation sites to confirm trophoblast cell identity. Cytokeratin and HLA-G were used on alternating sections with the various chemokine receptors. Both cytokeratin and HLA-G were used in a protocol similar to that described above except, anti-human cytokeratin (CAM 5.2; Becton Dickinson Immunocytochemistry Systems) was used and prediluted according to manufacturer's instructions. Tissue was digested with trypsin (Sigma; 0.2% in 0.2% calcium chloride/TBS) for 10 min at 37°C before primary antibody application.
HLA-G (16G1) was localized using a monoclonal antibody kindly provided by Dr. D. E. Geraghty (Fred Hutchinson Cancer Research Center, Seattle, WA). Tissue was microwaved at medium-high power (1000 W) in 0.01 mol/L sodium citrate buffer (pH 6.0) for 5 min, followed by incubation in the hot buffer for a further 20 min before application of primary antibody at 5 µg/ml as previously described [16]. The secondary antibody used with cytokeratin and HLA-G was biotinylated horse anti-mouse IgG secondary antibody and was also applied for 30 min.
Isolation and Culture of Human Endometrial Stromal Cells and Human Endometrial Epithelial Cells
Endometrial stromal cells (ESCs) and EECs were isolated from tissue by enzymatic digestion and filtration as described previously [17, 18]. In brief, minced tissue was digested with bacterial collagenase type III (Worthington Biochemical Corporation) at a concentration of 45 IU/ml in the presence of 3.5 µg/ml deoxyribonuclease (Boehringer Mannheim Biochemica) for 30–45 min at 37°C. The digested tissue was filtered sequentially through 45- and 10-µm nylon filters to remove glands, and centrifuged. Epithelial glands were recovered from the filters by backwashing [19] and stromal cells were collected from the cell pellet. EECs and ESCs were resuspended in DMEM/F12 with 10% charcoal-stripped fetal calf serum (CS-FCS) and 1% antibiotics. Stromal cells were then plated at a density of 2.5 x 105 cells/well in a 24-well plate for 30–60 min to adhere to the plate; cells that did not adhere were washed off and stromal cells were grown to confluence. Once confluent, stromal cells were washed in PBS and changed to serum-free medium containing transferrin (10 µg/ml), sodium selenite (25 ng/ml), linoleic acid (4.7 µg/ml), BSA (1 mg/ml) (all from Sigma Diagnostics) and insulin (5 µg/ml) (Actrapid; Novo-Nordisk) (TSL-I) and cultured for 48 h; medium was collected and stored as aliquots at –20°C for subsequent experiments and cells were retained for RNA extraction. Stromal cell isolation yields approximately 95% pure stromal cell preparations. Epithelial cells were further purified by selective adherence. Cells were plated in a 35-mm plastic culture dish, contaminating stromal cells were allowed to adhere for 30–60 min, and nonadhering epithelial cells were washed off. This step was repeated, and pure epithelial cells were then plated in DMEM/F12 with 10% CS-FCS for 2–3 days, after which medium was changed to DMEM/F12/TSL-I for 24 h. Medium and cells were harvested as above for subsequent experiments. Cells were collected for RNA extraction. Epithelial cell isolation yields approximately 85% pure epithelial cell preparations. Cell purity was assessed by immunostaining for cytokeratin (antibody SC-8018; Santa Cruz Biotechnologies) and vimentin (antibody SC-6260; Santa Cruz Biotechnologies) as described previously [20].
In vitro decidualization. Once confluent, ESCs were washed in PBS and maintained in serum-free DMEM/F12/TSL for 2 days. Cells were then decidualized as previously described [18] by treatment with estradiol (E2, 10–8 M; Sigma) in the presence (or absence) of medroxy-progesterone acetate (10–7 M; Sigma) for 10 days, with medium change every 2 days. Decidualization was confirmed by ELISA for prolactin (Bioclone Aust.; data not shown). Conditioned media and cells were harvested for subsequent analyses.
Additional nondecidualized stromal cells and an irrelevant pancreatic (INS-1) cell line were cultured in DMEM/F12/TSL-I for 48 h to provide control conditioned media.
Cell Lines
Three trophoblast cell lines—the human choriocarcinoma JEG-3 [21] and two human choriocarcinoma-primary trophoblast hybrids, AC1M-32 and AC1M-88 [22]—and a human endometrial epithelial (HES) cell line, a gift from Dr. D. A. Kniss (Dept. of Obstetrics and Gynecology, The Ohio State University), [23, 24] were cultured in RPMI-1640 (Sigma-Aldrich) with 10% CS-FCS for 2–3 passages following thawing. Once 80% confluent, cells were harvested and pelleted for RNA extraction. AC1M-88 cells were also cultured to 80% confluence and media was then changed to RPMI/TSL-I. Cells were cultured for 48 h, harvested and pelleted for subsequent RNA extraction. An irrelevant pancreatic cell line (INS-1) was also grown to 80% confluence and media was then changed to RPMI/TSL-I; media was collected for subsequent experiments.
RNA Extraction, Purification, and Quantitation
Total RNA was isolated from cultured trophoblast cell lines (JEG-3, AC1M-32, and AC1M-88), HES, isolated EECs, and decidualized ESCs (dESCs) using an RNeasy MiniKit (Qiagen GmbH) according to the manufacturer's instructions. Total RNA was extracted from midsecretory endometrial (n = 5) and placental (n = 5) samples by homogenization in Trizol reagent (Qiagen Sciences), according to the manufacturer's instructions, with the exception of an additional chloroform extraction step to minimize carryover of phenol into the precipitated RNA. All samples were treated with RNase-free DNase (Ambion) to remove any genomic DNA contamination, and analyzed by spectrophotometry to determine RNA concentration, yield, and purity. Any samples with ratios of A260/280 < 1.7 or A230/280 > 1 were purified through RNeasy spin columns (Qiagen) according to manufacturer's instructions and reanalyzed. One microgram of purified RNA was then run on a 1% agarose (Roche) gel to ensure integrity of rRNA subunits.
RT-PCR
Expression of mRNA for CX3CL1, CCL7, their receptors and 18s rRNA was assessed in three trophoblast cell lines (JEG-3, AC1M-32, and AC1M-88), in HES, and in EECs, midsecretory (MS) phase endometrium and first-trimester PL using RT-PCR. CX3CL1 and CCL7 were also assessed in dESCs. One microgram of RNA sample was reverse transcribed using AMV-RTase (Promega) and 100 ng random hexanucleotide primers (Amersham Biosciences); cDNA generated was subsequently amplified by PCR for each chemokine/receptor and 18s using a Hybaid Express block cycler (40 cycles; Hybaid Ltd.), following optimization of PCR conditions for chemokine receptors (Table 1), and under conditions previously reported for CX3CL1 and CCL7 [4] and 18s rRNA [25]. Products were electrophoresed on a 2% agarose gel, the bands excised, and cDNA purified using UltraClean GelSpin columns (MoBio Laboratories Inc.). The resultant PCR-generated cDNAs were quantitated by spectrophotometry and sequenced to confirm identity. Semiquantified data are presented as a percentage of total expression following densitometric analysis.
|
Trophoblast Migration Assay
Chemokines. Based on the PCR data, the trophoblast cell line AC1M-88 was selected for migration studies. Freshly thawed cells were passaged 2–3 times and grown to 80% confluence before serum starvation for 18–24 h in RPMI/TSL-I. Cells were harvested with 0.05% trypsin (Sigma) and trypsin activity then quenched by the addition of RPMI with 2% CS-FCS. After two washes with PBS cells were resuspended at a concentration of 0.25–1.0 x 106 cells per ml in RPMI/TSL-I.
A migration assay kit, ECM 510 QCM Chemotaxis 96-well (8 µm) Cell Migration Assay (Chemicon), was used according to manufacturer's instructions. RPMI/TSL-I with or without (carrier-free) recombinant human chemokines (CX3CL1, CCL7, CCL14, and CCL4 [1–100 nM]; R and D Systems Inc.) were added to the lower chambers of a 96-well plate (5 wells per treatment). The cell culture insert (8-µm-diameter pores) was placed in each well, forming the upper chamber; cells were seeded into each well of the upper chamber (2.5–7.5 x 104 cells in 100 µl of RPMI). The chamber was incubated for 22 h at 37°C in a CO2 incubator to allow migration through the membrane. Cells remaining in the upper chamber were removed and migration through the membrane was assessed by detaching cells on the lower surface of the membrane with cell detachment buffer. Migrated cells were lysed and quantitated using the fluorescent dye CyQuant GR Dye, which exhibits a strong fluorescence enhancement when bound to cellular nucleic acids. Cell migration was assessed using a fluorescent plate reader, with a 480/520-nm filter set. Mean migration from two or three separate experiments (n 
5 wells/dose per experiment) was combined and expressed as mean fold change from control.
Conditioned medium. In subsequent experiments, conditioned media from EEC or decidualized and nondecidualized ESC cells were placed in the lower chamber of the assay, with replicates of 5–12 wells per treatment. Conditioned medium from the pancreatic cell line INS-1 was used as an additional control. For neutralization experiments, neutralizing antibodies against human CX3CL1, CCL4, and CCL7 (R and D Systems) were preincubated at 37°C with EEC-conditioned media (2 patients pooled) for 1 h before performing migration assays. The conditioned media containing neutralizing antibodies was placed in the lower chamber of the migration assay and trophoblast cells in the upper chamber. Migration conditions and analysis were identical to those described above. Each migration assay experiment was repeated 2–3 times.
Statistical Analysis
Results for the migration assays are expressed as mean ± SEM fold change for each treatment compared to control. Statistical analysis was performed using one-way ANOVA, followed by Tukey's post hoc test (P< 0.05 was taken as significant) after testing for normal distribution using PRISM version 3.00 for Windows (GraphPad).
RESULTS
Immunolocalization of Chemokine Receptors in Human Implantation Sites
To localize chemokine receptors at the feto-maternal interface, tissue sections from first-trimester human implantation sites were stained for CCR1, CCR3, and CX3CR1. In the maternal components of the tissue, immunoreactive CCR1 protein was localized to glandular (Fig. 2A) and luminal epithelium (data not shown), leukocyte subpopulations (Fig. 2B), and decidualized stromal cells (Fig. 2C). Weak expression was seen in the syncytio- and cytotrophoblast layer and the cell column (Fig. 2D). Similarly to previous findings [12] CCR1 protein was predominantly localized on the EVTs in the maternal arteries, vEVTs (Fig. 2E), but not on invading iEVTs (data not shown). Trophoblast identity was confirmed by cell morphology and serial staining for cytokeratin (Fig. 2F) and HLA-G (Fig. 2G).
|
Immunoreactive CCR3 protein was localized to glandular epithelium (Fig. 2H), leukocyte subpopulations, endothelial and vascular smooth muscle cells of endometrial blood vessels (Fig. 2I), and decidualized stromal cells (Fig. 2J), consistent with previous findings [15]. CCR3 protein was also localized to the invading iEVTs (Fig. 2, I, L, and N). Intense staining was observed on the microvilli on the apical surface of the syncytium, with weak staining observed in the villous cytotrophoblast layer (Fig. 2K) and the cell column (data not shown). Trophoblast identity was confirmed by serial cytokeratin staining (Fig. 2, M and O).
CX3CR1 protein was also localized to vEVTs (Fig. 2P), and trophoblast identity was confirmed by serial cytokeratin (Fig. 2Q) and HLA-G (Fig. 2R) staining. Expression was absent on the iEVTs (Fig. 2S) confirmed by serial staining for cytokeratin (Fig. 2T); weak staining was also observed on the syncytium and the cell column (Fig. 2U). Trophoblast identity was confirmed by serial cytokeratin staining (Fig. 2V).
Expression of CX3CL1, CCL7, and Receptor mRNA in Cell Lines and Primary Tissues Representing Components of the Feto-Maternal Interface
To assess the mRNA for CX3CL1, CCL7, and their receptors, RT-PCR was performed on three trophoblast cell lines (JEG-3, AC1M-32, and AC1M-88), and on HES cells, EECs, decidualized ESCs, and samples of midsecretory phase endometrium and first-trimester PL. Specific bands were detected for CX3CL1, CX3CR1, CCL7, CCR1, CCR2, CCR3, and CCR5 (Fig. 3), and semiquantitiation of data is shown in Table 2, where densitometric analysis was performed on resultant PCR products (Fig. 3). Each mRNA was assessed independently; levels were normalized to 18s for each sample. Results are shown in Table 2 as a percentage of maximal chemokine or receptor expression within each mRNA examined, with highest expression designated 100.
|
|
CX3CL1 mRNA was present in all three trophoblast cell lines but was most abundant in AC1M-32. It was also abundant in EEC, MS-phase endometrium, and dESCs, and was detectable in first-trimester PL (Fig. 3 and Table 2). Expression of CX3CL1 receptor CX3CR1 mRNA was easily detected in two of the trophoblast cell lines, but was most abundant in the AC1M-88. CX3CR1 mRNA was also detected in both HES and EECs. Abundant expression was observed in MS-phase endometrium and in first-trimester PL (Fig. 3 and Table 2).
CCL7 was not expressed by JEG-3 or AC1M-88 cells, but was easily detectable in the trophoblast cell line AC1M-32, the HES cell line, EECs, MS endometrium, first-trimester PL, and dESCs. CCR1 mRNA was detected in all three trophoblast cell lines, the HES cell line, and EECs. CCR1 expression was present in MS-phase endometrial samples and first-trimester PL. CCR2 and CCR3 mRNA was also detected in the trophoblast cell lines, with abundant CCR3 expression in the AC-1M88 cells. CCR2 was detected in the HES cell line, in MS-phase endometrium, and in first-trimester PL, but not in EECs. CCR3 was also detected in first-trimester PL, but not in HES, EECs, or MS-phase endometrium. CCR5 was detectable in the trophoblast cell lines, EECs, MS endometrium, in first-trimester PL, but not in HES (Fig. 3 and Table 2). Semiquantitative results are summarized in Table 2, showing percentage of total mRNA expression after normalization for 18s.
Effect of CX3CL1, CCL7, CCL14, and CCL4 on Trophoblast Migration
Based on the chemokine receptor expression profile, AC1M-88 cells were selected for use in migration assays as most representative in this respect of trophoblasts "in situ." Because of abundant expression of CCR1 mRNA in all cell types examined at the feto-maternal interface, the CCR1 ligands CCL14 and CCL4 were selected in addition to CX3CL1 and CCL7 for assessment of their effects on trophoblast cell migration. Preliminary dose-response experiments (1–100 nM) were performed for each chemokine. Representative data for CCL4 is shown in Fig. 4A. Significant migration was seen at 100 nM, although a lesser nonsignificant response is observed at lower doses. Similar results were obtained for the other chemokines. 100 nM was thus chosen for further experiments.
|
In the presence of CX3CL1 (100 nM) the number of trophoblast cells that migrated to the lower chamber was significantly increased in comparison to control (P < 0.05; RPMI/TSL-I alone; Fig. 4B). CCL14 and CCL4 likewise stimulated significant trophoblast migration (P < 0.05; Fig. 4B). CCL7 was without significant effect.
Influence of Endometrial Epithelial Conditioned Medium on Trophoblast Migration
The influence of endometrial epithelial factors on trophoblast migration was examined using conditioned medium from EECs. The number of trophoblast cells that migrated in response to the epithelial conditioned medium (epithelial cells derived from three individual patients) was significantly elevated compared to control (P < 0.05)(RPMI/TSL-I, alone; Fig. 5A). Conditioned media from irrelevant cells (pancreatic cells) had no effect on trophoblast migration compared to control (Fig. 5A).
|
Influence of Endometrial Decidualized and Nondecidualized Stromal Cell Conditioned Medium on Trophoblast Migration
The influence of stromal factors on trophoblast migration was also examined using conditioned medium from decidualized and nondecidualized ESCs. Trophoblast migration was slightly increased in response to medium from nondecidualized ESCs (stromal cells derived from two patients; pooled medium), when compared to control (RPMI/TSL-I alone; Fig. 5B). However, in response to conditioned medium from decidualized cells (stromal cells derived from four individual patients), the number of trophoblast cells that migrated was significantly increased when compared to control (P < 0.05; Fig. 5B).
Individual Chemokine Contribution to Epithelial Factors Promoting Trophoblast Migration
To determine the potential contribution of individual chemokines to the significant trophoblast migration in response to endometrial epithelial conditioned medium, neutralizing antibodies specific for human CX3CL1, CCL4, and CCL7 were added to pooled medium from the same 3 patients used in previous EEC conditioned medium experiments. In the presence of anti-CX3CL1, a significant reduction in trophoblast migration was observed when compared to control (P < 0.05; EEC conditioned medium alone; Fig. 6). Similarly, anti-CCL4 also significantly decreased trophoblast migration in response to EEC conditioned medium (P < 0.05). Anti-CCL7 neutralizing antibody had no effect on trophoblast migration (Fig. 6).
|
DISCUSSION
Extensive crosstalk between fetal and maternal cells during embryo implantation is well recognized as an essential requirement for the establishment of pregnancy. Recent identification of a number of chemokines and their receptors at the feto-maternal interface suggests roles for chemokines in regulating the processes that occur during implantation [4, 10, 26]. The present study identified mRNA for the chemokines CX3CL1 and CCL7, and their receptors CX3CR1, CCR1, CCR2, CCR3, and CCR5 in human endometrium during the window of implantation and in first-trimester PL. Furthermore immunoreactive CCR1, CCR3, and CX3CR1 were shown in maternal and fetal cells in human first-trimester implantation sites. Chemokine and receptor expression was also demonstrated in a number of cell lines representing the components of the feto-maternal interface. Following careful selection, human endometrial epithelial cells and the human choriocarcinoma-primary trophoblast cell line AC1M-88 were chosen because of close resemblance to in vivo cell phenotypes in terms of chemokine and receptor expression. Using AC1M-88 cells in an in vitro trophoblast migration assay, significant trophoblast migration occurred in response to recombinant CX3CL1, CCL14, and CCL4. Importantly, epithelial and decidualized stromal cell conditioned media potently stimulated trophoblast migration and this was partially blocked by neutralization of CX3CL1 and CCL4.
In the early stages of placentation, cytotrophoblast cells enter two main pathways of differentiation, villous and extravillous. The villous cytotrophoblasts line the chorionic villi; these fuse to generate the syncytiotrophoblast,which mediates nutrient, gas, and waste exchange between fetal and maternal blood. In points of contact between villous tips and the decidua, cytotrophoblast cells form anchoring cell columns, establishing a physical connection between the fetus and the mother. At the distal portion of the cell column, cytotrophoblasts differentiate into an invasive phenotype and then either invade maternal blood vessels and become vEVTs, or migrate through the decidua as iEVTs. Initially, vEVTs form plugs in maternal arteries; they then actively remodel maternal arteries by destroying the muscular wall and replacing the endothelial lining, increasing the blood flow to the PL. Interstitial EVTs cluster outside vessels and probably assist in vascular remodeling [27, 28]. How cytrophoblasts inhabit the maternal arteries and serve as endothelium is unknown. We have shown here that chemokines may be involved in the targeted migration of EVT populations through the decidua and specifically to the maternal arteries, because these cells express chemokine receptors in early human implantation sites. As expected, CCR1 was localized to leukocytes, but it was also present on vEVTs, and to a lesser extent on the syncytiotrophoblast, in accordance with the previous finding of CCR1 expression on invasive EVTs [12]. We have previously demonstrated the production of a number of CCR1 ligands (CCL7, CCL14, CCL16, and CCL4) by endometrial epithelium, decidualized stromal cells, and vascular endothelium [4], although it is important to note that these ligands bind other chemokine receptors (for example, CCL4 binds both CCR1 and CCR5 [29]).
The CX3CL1 receptor CX3CR1 was also present on vEVTs and weakly stained in the syncytiotrophoblast layer. Although CX3CR1 mRNA expression has been detected in trophoblast cells [10], this is the first demonstration that it specifically localizes to vEVTs. CX3CL1, the only known ligand for CX3CR1, is predominantly localized to endometrial epithelium, decidualized cells, and leukocytes, with maximal expression in the midsecretory phase and in early pregnancy decidua [13].
In contrast to CCR1 and CX3CR1, CCR3 was localized to iEVTs invading the decidua and to microvilli on the apical surface of the syncytium, but was absent in the syncytio-cytotrophoblast layer. CCR3 protein was also localized to the cell column. CCR3 ligands (CCL7 and CCL11) are produced by the endometrium, particularly during the midsecretory phase, by decidualized stromal cells, endometrial epithelium, and vasculature [4, 15]. Although CCR3 has previously been described in endometrial cells, this is the first demonstration of this chemokine receptor on trophoblast cells.
Thus, the chemokines CX3CL1 and CCL7 are produced by endometrial cells, particularly epithelial and decidualized stromal cells in the endometrium. The receptors for these chemokines are present in adjacent trophoblast cells, supporting a role for endometrial chemokines in trophoblast differentiation and directed trophoblast invasion. Further, CX3CL1 derived from subpopulations of uterine natural killer cells and macrophages in decidualized zones [13] could also contribute to regulation of CX3CR1-positive EVTs. CCR1 ligands CCL14 and CCL4 are likewise produced by human endometrium [4, 12, 30]. Previously, expression of chemokine receptors CCR2B and CCR5, which bind a number of chemokines including CCL7, has been detected in human blastocyst.
To confirm the functional significance of chemokine receptor expression patterns, we demonstrated significant trophoblast migration in response to recombinant CX3CL1, CCL14, and CCL4, demonstrating for the first time that trophoblast migrates in response to CX3CL1 and CCL4. Previously, CCL14 has been shown to promote significant trophoblast migration, using a similar in vitro system [12]; this is confirmed here.
Thus, epithelial and decidual derived CX3CL1, CCL14, and CCL4 can contribute to trophoblast invasion by promoting their migration. Chemokines are likely to play a role in trophoblast migration in each phase of implantation, from early apposition to invasion into blood vessels. It is therefore not surprising that conditioned culture medium from EECs and decidualized ESCs induced high chemotaxis of trophoblast cells compared to nondecidualized ESCs stromal cells, which do not produce these chemokines. This was due at least in part to CX3CL1 and CCL4, as shown by the reduction in migration in the presence of specific neutralizing antibodies. A similar role for CCL14 can be envisaged, but could not be examined in the present study because of unavailability of a specific neutralizing antibody. However, reduced trophoblast migration was observed when CCL14 was inactivated, when compared to intact CCL14 [12].
Previous studies showed a changing basal to apical localization of chemokines in endometrial epithelial cells, as the cycle progresses from the proliferative to the secretory phase, indicating secretion of chemokines into the uterine lumen during the secretory phase [4, 13]. Further, chemokines are apically secreted by cultured endometrial epithelial cells [31]. Thus, in vivo, chemokines (such as CX3CL1 and CCL4) may act on the blastocyst before and during early stages of implantation to promote early trophoblast function. Later in implantation, chemokines secreted from decidual cells could promote trophoblast invasion through the decidua. Furthermore, there is evidence for the continued involvement of epithelial gland products for the future development of the PL throughout the first trimester of pregnancy [32].
The mechanisms underlying EVTs' homing to the maternal vasculature are not well understood, and thus the likelihood that chemokines contribute to the directional migration and homing of trophoblasts to maternal arteries is important. Previously, we have shown specific localization of CX3CL1, CCL14, CCL16, CXCL8, CCL11, and CCL22 to endometrial vasculature [4, 13]. Although the present study did not examine the effect of endothelial conditioned media on trophoblast migration in the present study, the production of numerous chemokines by spiral arterioles suggests a direct role at least for CX3CL1, CCL14, and CCL16 in regulating trophoblast migration to these sites. Thus, endometrially derived chemokines may act at different sites as trophoblast invasion progresses.
Using first-trimester implantation sites, we have shown CCR1, CCR3, and CX3CR1 protein localized to leukocyte subpopulations, endometrial epithelium, and decidualized stromal cells in the maternal compartments, consistent with previous findings in nonpregnant endometrium [13, 15]. A role for chemokines and receptor expression by nonmigrating cells is unknown. As maximal expression of both ligand and receptor is present on the endometrial epithelium during the secretory phase, an autocrine role for the chemokines in regulating adhesion molecule expression to create a receptive endometrium can be proposed.
In conclusion, we have demonstrated that chemokine-receptor networks are present at the feto-maternal interface, and in particular that invasive trophoblasts possess receptors for a number of the chemokines produced by the human endometrium. A direct role for CX3CL1, CCL14, and CCL4 can be predicted in stimulating trophoblast migration because of their appropriate ligand/receptor expression in vivo and the functional effects exhibited in vitro.
ACKNOWLEDGMENTS
The authors would like to thank Prof. Euan Wallace and Dr. Ursula Manuelpillai (Monash University, Dept. of Obstetrics and Gynaecology) for provision of first-trimester implantation site blocks.
FOOTNOTES
1 Supported by the National Health and Medical Research Council of Australia (#241000 and #143798), a Prince Henry's Institute of Medical Research Ph.D. scholarship, and the Flew Foundation. 
2 Correspondence: Natalie J. Hannan, PHIMR, P.O. Box 5152, Clayton, Victoria 3168, Australia. FAX: 61 3 9594 6125; natalie.hannan{at}phimr.monash.edu.au
Received: 13 July 2005.
First decision: 3 August 2005.
Accepted: 30 January 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  A. Tapia, L. A. Salamonsen, U. Manuelpillai, and E. Dimitriadis Leukemia inhibitory factor promotes human first trimester extravillous trophoblast adhesion to extracellular matrix and secretion of tissue inhibitor of metalloproteinases-1 and -2 Hum. Reprod., August 1, 2008; 23(8): 1724 - 1732. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Dominguez, S. Martinez, A. Quinonero, F. Loro, J.A. Horcajadas, A. Pellicer, and C. Simon CXCL10 and IL-6 induce chemotaxis in human trophoblast cell lines Mol. Hum. Reprod., July 1, 2008; 14(7): 423 - 430. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. J Hannan and L. A Salamonsen CX3CL1 and CCL14 Regulate Extracellular Matrix and Adhesion Molecules in the Trophoblast: Potential Roles in Human Embryo Implantation Biol Reprod, July 1, 2008; 79(1): 58 - 65. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 T. E Spencer, O. Sandra, and E. Wolf Genes involved in conceptus-endometrial interactions in ruminants: insights from reductionism and thoughts on holistic approaches Reproduction, February 1, 2008; 135(2): 165 - 179. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. Paiva, L. A. Salamonsen, U. Manuelpillai, C. Walker, A. Tapia, E. M. Wallace, and E. Dimitriadis Interleukin-11 Promotes Migration, But Not Proliferation, of Human Trophoblast Cells, Implying a Role in Placentation Endocrinology, November 1, 2007; 148(11): 5566 - 5572. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. G Lea and O. Sandra Immunoendocrine aspects of endometrial function and implantation Reproduction, September 1, 2007; 134(3): 389 - 404. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
-CX3CL1,