| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Department of Obstetrics and Gynecology, The University of Iowa, Iowa City, Iowa 52240
b Holden Comprehensive Cancer Center at the University of Iowa, Iowa City, Iowa 52240
c Department of Anatomy and Cell Biology, The University of Iowa, Iowa City, Iowa 52242
ABSTRACT
In vitro morphogenesis of epithelial cells to form tube-like structures is regulated by hepatocyte growth factor-scatter factor (HGF/SF). The placenta is a rich source of HGF/SF, and its absence in mice has been shown to lead to impaired placental growth and embryonic death. There is no information in the literature regarding in vitro morphogenesis of human cytotrophoblasts or the effect of HGF/SF on this process. In this study, cytotrophoblasts were isolated from human placentae obtained from all three trimesters of gestation and cultured on the recombinant basement membrane matrix (Matrigel). Under these conditions, cytotrophoblasts participated in morphogenetic events including formation of spheroid-like structures, radial linear processes with branching, and invaded Matrigel and formed large, tube-like structures. The presence of a developing lumen was documented in the linear projections arising from spheroids and in the tube-like structures by both confocal and transmission electron microscopy. Immunohistochemistry was used to characterize the phenotype of the cells, and staining with anti-cytokeratin and anti-E-cadherin antibodies confirmed the presence of cytotrophoblasts in both the spheroids and tube-like structures. Recombinant HGF (rHGF) significantly increased the invasive activity of cytotrophoblasts isolated from the first and second (P < 0.001) and third trimesters (P < 0.01). In addition, rHGF significantly increased the percentage of spheroids with branching processes in the first and second trimesters (P < 0.05). Anti-HGF antibody inhibited both these effects in a dose-dependent manner, indicating the specificity of the above findings. This study provides new evidence indicating that HGF/SF regulates invasion and branching morphogenesis of cytotrophoblasts throughout gestation, with maximum effects in the first and second trimester. These findings may help to elucidate the importance of the reduced expression of HGF/SF identified in placentae from women with preeclampsia or intrauterine growth restriction and suggest that HGF/SF may serve as an important candidate in therapeutic intervention strategies.
cytokines, developmental biology, growth factors, placenta, pregnancy
INTRODUCTION
The formation of epithelial organs during embryonic development involves an ordered sequence of morphogenetic processes. The early appearance of these organs is that of highly arborized duct-like structures containing a central lumen and subsequent formation of precisely organized multicellular structures. Tissue- and function-specific morphogenesis has been described in organs such as the kidney, lungs, thyroid, and mammary gland [1–3]. Furthermore, there is in vitro evidence to indicate that epithelial-mesenchymal interactions are critical for proper morphogenesis [4]. Epithelial tubular morphogenesis and differentiation are mediated by a number of factors, including cytokines such as hepatocyte growth factor (HGF) [5], extracellular matrix components such as laminin-1 [6], and integrins [7]. Hepatocyte growth factor, also called scatter factor (SF), is a multifunctional cytokine that promotes growth, differentiation, invasion, angiogenesis, and morphogenesis [2, 4, 5]. The stimulatory effects of HGF/SF on linear and branching morphogenesis in kidney and breast cells have been described in detail [5]. However, there is no information in the literature describing in vitro morphogenesis by cytotrophoblasts isolated from the human placenta or the putative effects of HGF/SF on this process.
The gross structure of the human placenta has been studied in detail using sophisticated modalities for ultrastructural analysis and stereologic studies [8–10]. In order to correlate the structural development of the placenta with pathological states and clinical outcomes, ultrasound and doppler waveform have been extensively used [11, 12]. In vitro differentiation of villus cytotrophoblasts along the invasive pathway has also been studied using isolated cells [13] and placental explant models [14]. The ability to recapitulate placental morphogenesis in vitro would provide another model to better understand the effects of a variety of agents on placental development and elucidate their mechanism of action. In this paper we describe the events involved in cytotrophoblast morphogenesis in a three-dimensional matrix model and have examined the effects of a known morphogen, HGF/SF, to validate this system.
The biological activities of HGF/SF are mediated via c-met, a transmembrane tyrosine kinase, encoded by the c-met protooncogene. In contrast to HGF/SF, which is mainly produced by cells of mesenchymal origin, c-met protein and message are expressed in most circumstances by cells of epithelial origin [15]. One of the richest sources of HGF/SF is the human placenta [16–18], where HGF/SF mRNA is strongly localized to the villous mesenchymal core [19, 20], and the c-met transcript is detected in the trophoblast layer with low levels throughout the villous core. Production of HGF/SF in vitro and mRNA expression have been shown to be maximum in the second trimester and then decrease at term [21]. Other studies have shown HGF/SF to be a potent mitogen [19] and to stimulate in vitro invasion of first and second trimester cytotrophoblasts in a dose-dependent manner [22]. These varied effects of HGF/SF are potentially important in placentation, a process requiring proliferation and cellular orientation of trophoblast, mesenchymal, and endothelial cells.
In mice, both HGF/SF or c-met gene knockout have similar effects on embryogenesis [23–25]. Homozygous mutant embryos lacking HGF had markedly smaller placentae secondary to reduced numbers of labyrinthine trophoblast cells and did not survive beyond Embryonic Day 17.5 [25]. The network of embryonic vessels and maternal sinuses was also poorly developed. The placental abnormalities were detected as early as Day 10.5, and embryonic death was ascribed to placental insufficiency and liver abnormalities. Maternally derived HGF/SF also stimulates growth of trophoblasts in early postimplantation mouse embryos [26]. The combined data from the above studies suggest that HGF/SF is an essential mediator of allantoic-trophoblastic interaction required for mouse placental organogenesis.
In human pregnancies, there is indirect evidence to suggest that HGF/SF may contribute to diseases associated with placental insufficiency such as intrauterine growth restriction (IUGR) and preeclampsia. Intrauterine growth restriction is characterized by pregnancies complicated by markedly decreased fetal weight, small placentae with few terminal villi, and reduction in villous trophoblast proliferation. Reduced expression of HGF/SF in placentae from IUGR pregnancies when compared to age-matched controls has been described [27, 28], although c-met expression remained unchanged in these placentae. Further, there is recent evidence that HGF/SF expression and secretion are also markedly reduced in placentae from pregnancies complicated by preeclampsia [22, 29]. However, there is currently no direct evidence demonstrating that HGF/SF regulates trophoblast morphogenesis in the human placenta. Thus, the aims of our study were first, to determine if human placental morphogenesis can be demonstrated in vitro; and second, to determine whether HGF/SF regulates cytotrophoblast morphogenesis in this in vitro model.
MATERIALS AND METHODS
Cytotrophoblast Isolation
Placentae were obtained immediately after first and second trimester terminations (7–20 wk) and after delivery by cesarean section at term (37–40 wk). The University of Iowa Institutional Review Board approved this study. Cytotrophoblasts were isolated on the same day by sequential enzymatic digestion using previously published methods (first and second trimester [30], third trimester [31]) with some modifications. Briefly, the placental tissue was rinsed and cut into fine pieces that were digested once with collagenase (2 mg/ml) and then with three cycles of 8% trypsin (Sigma, St. Louis, MO). After each cycle of digestion, the cells in the supernatant were pelleted by centrifugation and collected. The cytotrophoblasts were then filtered to separate them from the undigested tissue, and contaminating red and white blood cells were further removed by using a 20–70% layered Percoll gradient (Sigma). Any remaining leukocytes were removed by incubating the cytotrophoblast-enriched Percoll gradient fraction with magnetic beads coated with an antibody to CD45 at a density of 5 beads/cell for 20 min at 4°C (Dynal Inc., Lake Success, NY). The viable cytotrophoblasts were then counted using trypan blue exclusion. After each isolation, cytospins were performed and the cells were stained with mouse anti-human cytokeratin antibody (Dako, Carpinteria, CA) to estimate the percentage of cytokeratin-positive cells in the preparation.
Tubular Morphogenesis Assay
Isolated cytotrophoblasts (5 x 105) were plated on the surface of undiluted Matrigel (Collaborative Biomedical, Bedford, MA)-coated transwells (12 mm; Corning Costar Corp., Cambridge, MA) inserted in 24-well tissue culture plates. The cells were cultured in 500 µl Ham F12-Dulbecco minimum essential medium (1:1) supplemented with 10% fetal calf serum under standard tissue culture conditions. The test wells had either recombinant HGF (rHGF; Collaborative Biomedical) added to the media at a concentration of 20 ng/ml or anti-human HGF monoclonal antibody (R&D Systems, Minneapolis, MN) added at a concentration of 3 µg/ml or 6 µg/ml. Daily observations were made by an inverted photomicroscope to document the developmental stages of tubular morphogenesis. Quantification of branching was performed on Day 12 of culture by estimating the percentage of spheroids with linear processes showing terminal branching in 10 different fields at 100x magnification for each transwell. Three replicates were included within each experiment, and three different preparations of cytotrophoblasts isolated from all three trimesters of gestation were analyzed.
Invasion Assay
Cell invasion was measured using the MICS (membrane invasion culture system) chambers comprising of 12 wells in a Boyden-like chamber [32]. Cytotrophoblasts (1 x 105) in serum-free medium (containing Mito+; Collaborative Biomedical) were seeded into the upper wells of the MICS chamber containing a polycarbonate filter with 10-µm pores coated with a mixture of 50 µg/ml laminin and 50 µg/ml collagen IV in a 2-mg/ml gelatin base. The media in the test wells contained 20 ng/ml rHGF or HGF with anti-HGF antibody (3 or 6 µg/ml). After incubating the cells at 37°C for 60 h, the wipe and flip method was used to gently remove cytotrophoblasts on the upper side of the filter with a cotton swab. The filter was then fixed in 100% ice-cold methanol and mounted with the lower side up onto microscope slides. The cells were stained with anti-cytokeratin antibody (Dako) conjugated with rhodamine. The cytokeratin-positive cells that had entirely invaded the filter were counted by confocal microscopy. Three replicate wells were analyzed in each experiment and three different preparations of cytotrophoblasts isolated from all three trimesters of gestation were analyzed.
Immunohistochemistry
To determine the histological and phenotypic characteristics of the cell aggregates and tube-like structures, the Matrigel-coated filters were cut out of the transwells and fixed in formalin, embedded in paraffin, and sectioned at 4-µm thickness. The slides were deparaffinized with xylene, and antigen recovery was performed with citrate buffer (pH 6.0).
The following antibodies were used for immunohistochemistry: mouse anti-human endothelial cell (CD31, dilution 1:100; Dako), mouse anti-human cytokeratin (dilution 1:150; Dako), anti-vimentin clone V-9 (dilution 1:150; Sigma), mouse anti-human E-cadherin (dilution 1:70; Transduction Laboratories, Lexington, KY), and rabbit anti-human h-met (dilution 1:100; Santa Cruz Biotechnology, Inc., Santa Cruz, CA).
Single immunohistochemical staining was performed on the Dako Autostainer using the LSAB+ System, alkaline phosphatase (Dako) to detect CD31, cytokeratin, and vimentin and the RTU Vectastain Universal Elite ABC kit (Vector Laboratories, Burlingame, CA) to detect E-cadherin and c-met. In brief, slides for the RTU kit only were treated for 10 min using 0.03% hydrogen peroxide and slides for both LSAB+ and RTU kits were blocked for 7 min in Protein Block Serum-Free (Dako). The sections were incubated in the primary antibody for 45–60 min, rinsed in Tris-buffered saline with Tween 20 (TBST), and then incubated in the LSAB+ Link or RTU Vectastain Secondary for 30 min. After another rinse in TBST the sections were incubated in the LSAB+ streptavidin-alkaline phosphatase or RTU Vectastain Elite ABC for 30 min. Color was produced by application of Vector Red Substrate (Vector Laboratories) for the LSAB+ kit or DAB+ (Brown) substrate (Dako) for the Elite ABC kit.
Dual immunohistochemical staining using both anti-cytokeratin and anti-vimentin antibodies was performed using Dako Envision Doublestain System (Dako). Briefly, the sections were blocked in 0.03% hydrogen peroxide for 10 min, rinsed with TBST, and incubated with the first primary antibody (Vimentin) for 45 min. Peroxidase-labeled polymer was applied for 10 min followed by Vector VIP substrate (purple) for 5 min. The slides were rinsed and a double stain block was applied for 3 min. The sections were then incubated with the anti-cytokeratin antibody for 45 min, followed by alkaline phosphatase-labeled polymer for 10 min and Vector Red substrate for 7 min. All sections were counterstained with methyl green. Negative controls were included in each experiment, and sections of human placental tissue from all three trimesters were used as the positive controls.
Confocal Microscopy
Matrigel-coated filters were cut out of the transwells, fixed in Pen-Fix (Richard Allan, Kalamazoo, MI), and placed on a slide maintaining its original orientation. The filters were stained with cytokeratin antibody conjugated with rhodamine for 1 h. Tube-like structures were viewed with a Bio-Rad 1024 laser scanning confocal microscope (Bio-Rad, Hercules, CA) equipped with a krypton-argon laser.
Scanning and Transmission Electron Microscopy
Matrigel-coated transwell filters were fixed in half-strength Karnovsky solution (2.5% glutaraldehyde; 1.5% paraformaldehyde in 0.1 M sodium cacodylate buffer) and postfixed for 20 min at room temperature in 1% osmium tetroxide with 1.5% potassium ferrocyanide in 0.1 M cacodylate buffer. Cells were rinsed and dehydrated using gradually increasing concentrations of ethanol from 50% to 100%. A transition reagent, hexamethyldisilazane, was used to reduce surface tension formed during the liquid-to-gas phase. The cells were then mounted onto aluminum stubs and sputter-coated with gold-palladium. Samples were imaged using a Hitachi 4000 scanning electron microscope with a field emission electron source (Hitachi Scientific Instruments, Tokyo, Japan).
For transmission electron microscopy, the filters were stained en-bloc for 48 h in 2.5% uranyl acetate at 37°C, dehydrated in a series of ethanol concentrations, and embedded in LX-112 epoxy embedment media (Ladd Research Industries, Inc., Williston, VT). The resulting blocks were thin-sectioned at 100-nm thickness, counterstained with Reynold lead citrate for 5 min, and viewed with a Hitachi H-7000 transmission electron microscope.
Statistical Analysis
The data are presented as mean ± SEM. The effects of treatment (HGF and antibodies) on the response variable was tested using ANOVA for a mixed effects model for the invasion data. The one-way ANOVA was used to determine the differences in the morphogenesis assay. The P values for these test statistics were adjusted using the Bonferroni method. Values were significantly different at P < 0.05.
RESULTS
Tubular Morphogenesis
Characterization of the cytotrophoblast phenotype is summarized in Table 1. The sequence of morphogenetic events observed when cytotrophoblasts were cultured on Matrigel-coated filters is shown in Figure 1. Cytotrophoblast preparations with >90% cytokeratin-positive cells as determined by immunostaining of cytospins were used in these experiments. Within 24–48 h, most cytotrophoblasts aggregated to form spheroid-like structures on the upper surface of the filter (Fig. 1A). Linear processes were seen to arise from and interconnect the spheroid-like structures (Fig. 1B). The linear projections from some of the spheroids increased in length and developed extensive branching patterns over 7–14 days in culture (Fig. 1C). In addition, single cells and the entire spheroid-like structures invaded the underlying Matrigel over time. By the third week of culture, some spheroids were connected by larger tube-like structures that also displayed branching patterns (Fig. 1, D–F). Cytotrophoblasts isolated from all three trimesters participated in all these biological events; however, the time course of sequential development was slower with cytotrophoblasts isolated from term placentae. By the third week in culture, there was extensive overlap of branching processes in transwells containing first and second trimester cytotrophoblasts. The cytotrophoblasts isolated from the third trimester, however, had fewer spheroids with branching processes at this later time point. Formation of linear and branching processes was not observed when the cytotrophoblasts were cultured on collagen I-coated filters or cultured directly within three-dimensional Matrigel gels (data not shown).
|
|
Evidence of Lumen Formation
The formation of three-dimensional, tube-like, linear and branching structures by cytotrophoblasts isolated from all three trimesters was confirmed by several techniques (Fig. 2). Histological sections of cytotrophoblasts cultured on Matrigel-coated filters, when stained with hematoxylin and eosin showed small and large tube-like structures lined by a single layer of cells (Fig. 2A). Scanning electron microscopy was used to visualize spheroids and the linear cellular processes arising from them (Fig. 2, B and C). The presence of a lumen noted within one of these processes is shown in Figure 2D. Confocal microscopy was used to identify the initial development of a lumen within cytokeratin-labeled, solid-appearing radial processes (Fig. 2E) and to confirm the presence of a lumen in the large tube-like structures (Fig. 2F). Transmission electron microscopy showed that a single layer of uninucleate cells lined the lumen of the large tube-like structures (Fig. 2G). Microvilli were seen at the luminal surface of individual cells, and desmosomes were present between adjacent cells. There was evidence of cellular debris associated with probable apoptosis within the lumen of some tubes.
|
Immunohistology
Immunohistochemistry was used to characterize the phenotype of the cells isolated from all three trimesters participating in the morphogenetic events described throughout the study, and the findings have been summarized in Table 1. Early cytotrophoblast aggregates that formed on the surface of the Matrigel-coated filters (Day 5 of culture) stained positively for the epithelial cell marker cytokeratin (Fig. 3A), and most of these cells were also positive for E-cadherin (Fig. 3B). Dual immunostaining for cytokeratin and vimentin showed that the tube-like structures that formed within the matrix after 2–3 wk of culture were lined by cells that predominantly stained with anti-cytokeratin antibody (Fig. 3C). These tube-like structures also stained positively with anti-E-cadherin (Fig. 3D) and anti-c-met (Fig. 3E) antibodies but were negative for CD31 (Fig. 3F). Dual immunostaining of cross sections of the invading spheroid-like structures showed some cells labeled with both cytokeratin and vimentin antibodies (Fig. 3G). Cells remaining on the surface of the Matrigel, and not participating in aggregation or invasion after 2–3 wk stained predominantly with anti-cytokeratin antibody (red). The single cells invading the extracellular matrix at this time were positive for vimentin (purple). No staining was detected in the negative controls on deletion of the primary antibody (Fig. 3H). Similar phenotypic characteristics were noted by immunohistochemistry when morphogenetic events were studied in all three trimesters.
|
Effect of rHGF on Cytotrophoblast Invasion
The effect of adding rHGF to serum-free medium containing cytotrophoblasts isolated from all three trimesters was tested with respect to the cell's invasive ability (Fig. 4). Cytotrophoblast preparations with >95% cytokeratin-positive cells as determined by immunostaining of cytospins were used in these experiments. The number of cytokeratin-positive cells that invaded each filter on addition of 20 ng/ml rHGF or rHGF with anti-HGF antibody (3 and 6 µg/ml) were counted by confocal microscopy. In the presence of rHGF, the number of cytotrophoblasts invading the filters significantly increased in all trimesters with the least response in the third trimester (first and second trimester P < 0.001, third trimester P < 0.01). The specificity of this effect was demonstrated by inhibiting invasion in a dose-dependent manner by the addition of the neutralizing HGF antibody. The inhibition of the stimulatory effect of rHGF by the antibody was statistically significant at a dose of 6 µg/ml in all three groups (first trimester and second trimester P < 0.001, third trimester P < 0.005). Overall, cytotrophoblasts isolated from the third trimester demonstrated a decrease in invasive ability by 52% compared to those cells isolated from the first and second trimesters.
|
Effect of rHGF on Cytotrophoblast Morphogenesis
In the first week of culture, aggregation of cytotrophoblasts to form spheroid-like structures with interconnecting linear processes was observed (Fig. 5A). In the presence of anti-HGF antibody aggregation of cytotrophoblasts occurred; however, there were fewer linear processes interconnecting these spheroids (Fig. 5B). The effect of rHGF (20 ng/ml) and its neutralizing antibody (3 and 6 µg/ml) on cytotrophoblast morphogenesis was further assessed by counting the number of spheroids with terminal branching of the linear processes after 12 days in culture (Fig. 5, C–F, and Fig. 6). This time period was chosen because it demonstrated linear growth associated with branching without extensive overlapping of radial processes from adjacent spheroids, thus allowing accurate assessment of terminal branching. Addition of rHGF to the culture media significantly increased the formation of branching processes in the first and second trimesters as compared to controls (P < 0.05, Fig. 5, C versus D, and Fig. 6). Overall, there was a dose-related response to the addition of anti-HGF antibody, with maximal inhibition of branching morphogenesis achieved at 6 µg/ml (Fig. 5, E and F, and Fig. 6). Also, fewer branching processes were seen to arise from spheroids in control transwells containing third trimester cytotrophoblasts compared to first and second trimester cytotrophoblasts (Fig. 6). When these experiments were continued beyond 12 days, rHGF did not appear to increase the formation of the large tube-like structures, although the cells lining the lumen of these tubes stained positively for c-met (shown in Fig. 3F).
|
|
DISCUSSION
This study demonstrates for the first time that placental morphogenesis can be recapitulated in vitro using a three-dimensional matrix and is regulated by HGF/SF. The in vitro developmental program of cytotrophoblasts includes a progression of events starting with cellular aggregation to the formation of spheroid-like structures and then invasion into the matrix. From the spheroid-like structures arise linear processes that are capable of branching, and the invading cells can reorganize to form large tube-like structures. Using a number of different methods we confirmed that these cord-like linear structures were initially solid and then subsequently formed a central lumen, hence engaging in tubulogenesis. Phenotypic characterization by immunohistochemistry identified the cells participating in the formation of spheroid and tube-like structures to be epithelial in origin as they stained positively with antibodies to cytokeratin, E-cadherin, and c-met. These cells were not positive for CD31, indicating that endothelial cells were not involved in the formation of any of these structures. Cytotrophoblasts isolated from all three trimesters participated in these biological processes including the formation of large tube-like branching structures. However cytotrophoblasts isolated from term placentae, demonstrated decreased invasive ability and slower progression through the various morphogenetic events described above. The morphogenesis assay was used to examine the effects of rHGF and its antibody on cytotrophoblast development. Recombinant HGF significantly increased the invasive ability of cytotrophoblasts and the number of branching processes arising from individual spheroids in the first and second trimesters. The specificity of both these effects was confirmed by the addition of a neutralizing antibody to HGF that demonstrated a dose-dependent inhibition of these biological events. This is the first description of HGF/SF regulating cytotrophoblast morphogenesis. It has been previously demonstrated that cytotrophoblasts when cultured in vitro will aggregate to form spheroid-like structures within the first 24 h [13]. Our study has examined this developmental process over a prolonged culture period and in cytotrophoblasts from all three trimesters of gestation. The ability of cytotrophoblasts to aggregate and invade the extracellular matrix appears to provide the three-dimensional environment required for tubular morphogenesis. This is supported by the observation that cytotrophoblasts directly cultured within three-dimensional Matrigel gels remained as single cells and did not exhibit any of the above events, possibly suggesting that the process of aggregation and invasion selects a subset of cells capable of participating in morphogenesis. Also, cells cultured on collagen I gels did not recapitulate the morphogenetic events but resulted in marked contraction of the reconstituted matrix (data not shown). These findings suggest that cytotrophoblasts have the capability of remodeling their extracellular matrix environment. However, the collagen I gels, being devoid of growth factors and multiple matrix components, could not induce the dynamic morphogenetic events in cytotrophoblasts, as demonstrated by the Matrigel matrix.
In this study we examined two biological effects of HGF/SF on cytotrophoblasts isolated from first, second, and third trimesters namely, invasion and morphogenesis. HGF/SF has previously been shown to stimulate in vitro invasion of first and second trimester cytotrophoblasts [22]. The increased invasion has been shown to be secondary to activation of the c-met receptor that mediates all the biological activities of HGF/SF and is associated with the expression of matrix metalloprotease (MMP)-9 [29]. We systematically investigated the effects of rHGF and its neutralizing antibody on invasion of cytotrophoblasts throughout gestation. We found that in comparison to first and second trimester cytotrophoblasts the overall ability of third trimester cytotrophoblasts to invade is decreased and their increase in invasive ability in response to rHGF was also reduced. These findings may be secondary to the reduced expression of c-met that has been reported at term [21]. Other cytokines such as transforming growth factor (TGF) ß, epidermal growth factor (EGF), and interleukin-1ß have also been shown to increase the invasive capacity of cytotrophoblasts in vitro [33, 34]. There is evidence to suggest that in vitro production of HGF/SF from villous explants derived from preeclamptic patients is lowered [29], and HGF/SF mRNA expression is decreased in placentae from patients with preeclampsia as compared to controls [22, 35]. In the previous studies, the majority of the placentae analyzed were from the third trimester, and in the current study we have shown that anti-HGF can specifically decrease invasion of cytotrophoblasts derived from the third trimester. These findings, collectively, may therefore provide an explanation for the shallow cytotrophoblast invasion observed in placentae from women with preeclampsia.
In the present study, we tested the hypothesis that HGF/SF may also regulate placental morphogenesis by increasing branching of linear processes arising from the spheroid-like structures. To determine the precise effect of HGF/SF on formation of branching structures, we used the morphogenesis assay to determine the percentage of spheroids with evidence of terminal branching. We observed a significant increase in branching processes in response to rHGF in cytotrophoblasts isolated from the first and second trimester. The lower numbers in the third trimester may reflect the slower progression through the various morphogenetic stages noted in vitro or a decreased response secondary to lower c-met receptor levels [21]. Inhibition of branching by the addition of a neutralizing antibody to HGF was dose-dependent and illustrates the specific role of HGF/SF in the regulation of cytotrophoblast branching morphogenesis. Interestingly, decreased HGF immunostaining has been observed in the villous stroma of placentae obtained from patients with IUGR [28] with no change in c-met expression. Our data demonstrate a direct effect of rHGF on cytotrophoblast morphogenesis and hence provide a possible explanation for the smaller placental size and decreased villous tree elaboration noted in placentae representing IUGR. In mice, HGF/SF knockout experiments directly demonstrate the relationship between HGF/SF and placental size [25]. Recently, the same authors reported that injection of HGF on Embryonic Day 9.5 into the amniotic cavity of HGF (-/-) embryos rescued the placental defect and resulted in survival up to term [36]. The possibility of using HGF/SF as a therapeutic intervention in clinical situations such as preeclampsia and IUGR remains to be explored.
The role of HGF/SF as a mediator of epithelial-stromal interactions and a promoter of branching, duct-like patterns has been reported in cells derived from other tissues. In normal mammary gland epithelial cells, rHGF (20 ng/ml) stimulated cord formation in a dose-dependent manner with a 77-fold increase in mean additive cord length [5]. TGF
and EGF also induced cord formation, although to a lesser extent than HGF/SF, whereas insulin-like growth factor-II, basic fibroblast growth factor, and platelet-derived growth factor did not enhance cord formation. Further, 17ß-estradiol, insulin, progesterone, and prolactin did not significantly modify the morphogenetic properties of these cells, whereas hydrocortisone greatly potentiated the tubulogenic effects of HGF/SF by increasing tube length and lumen formation. The effect of protease inhibitors on tubule formation was studied in a kidney cell line [1] and showed that in the presence of a synthetic collagenase inhibitor, extension of cytoplasmic processes was visualized; however, no branching occurred. Similarly, serine protease inhibitors suppressed tube formation in fibrin gel cocultures, indicating that different classes of proteases are essential for morphogenesis depending on the composition of the substrate. Although our current study did not address the involvement of MMPs, it is clear from work in our laboratory that they are important for the formation and remodeling of tubular networks [37].
The tube-like structures formed within the extracellular matrix recapitulate primitive villous growth in vitro. Villous growth in vivo occurs by formation of new terminal villi that sprout from intermediate and stem villi. The sevenfold increase in placental weight from early second trimester to term is mainly attributed to a tremendous growth in terminal villi, especially an increase in their length from 20 wk onward [38]. This results in the expansion of the villous component of the placental parenchyma providing effective exchange surfaces. In our experiments, tubular morphogenesis was demonstrated by cytotrophoblasts derived from all three trimesters, although the time course of morphogenetic events visualized with term cytotrophoblasts was slower compared to cytotrophoblasts isolated from first and second trimester. Further corroboration was shown by our in vitro data that demonstrate a decrease in invasive ability of cytotrophoblasts with an increase in the gestational age of the placenta.
In the in vitro morphogenesis model presented herein, some of the cytotrophoblasts invading the Matrigel stained for both cytokeratin and vimentin intermediate filaments. The coexpression of epithelial and mesenchymal intermediate filaments, although a novel finding for cytotrophoblasts, has been described in highly invasive and metastatic epithelial and nonepithelial neoplasms such as breast and melanoma [39]. However, this phenomenon of epithelial-to-mesenchymal conversion has not been previously described in the human placenta. The ability to coexpress vimentin and keratin intermediate filaments offers a selective advantage to tumor cells in their migratory and invasive function due to unique interactions between cell surface receptors, the cytoskeleton, and extracellular matrix. There are data to suggest that integrins are the most likely receptors to interact with intermediate filaments and their linking proteins [40]. Further, the matrix environment can modulate the integrin expression patterns and play a key role during tubular morphogenesis as shown in fetal tracheal epithelium [7]. There is evidence to suggest that invading cytotrophoblasts, both in vivo and in vitro, alter their integrin expression to mimic a more invasive phenotype, i.e., from 5 to 1 integrin [41], that may augment their migratory and invasive potential, as has been shown in aggressive, interconverted tumor cells [39].
Several in vitro studies have shown that reconstituted matrices, such as collagen or basement membrane gels, promote the organization of epithelial cells into three-dimensional tube-like structures. This is the first report of experimental conditions stimulating a component of the architecture of the human placenta. Using the data from our current study and the literature, we have proposed a hypothetical model to elucidate the role of HGF/SF in the two clinical conditions discussed above namely, preeclampsia and IUGR (Fig. 7). HGF/SF produced by villous stromal cells acts via c-met to stimulate mitosis, invasion, and morphogenesis. Decreased levels of HGF/SF may reduce proliferation and morphogenesis, resulting in the diminished placental surface area associated with pregnancies complicated with IUGR. In addition, reduced levels of HGF may decrease invasion of cytotrophoblasts, a finding associated with preeclampsia. Further studies are ongoing to determine the signaling cues that are critical to HGF/SF regulation of cytotrophoblast morphogenesis and the clinical implications of abnormal levels of HGF/SF during development.
|
ACKNOWLEDGMENTS
The authors thank Katherine Walters and Dr. Paul Heidger for their expertise with confocal microscopy and transmission electron microscopy, Drs. Susan Fisher and Harvey Kliman for their guidance with cytotrophoblast isolation and culture, and Dr. Dawn Kirschmann for graphic assistance.
FOOTNOTES
First decision: 28 March 2001.
1 This work was supported by a grant to A.D. from the Reproductive Scientist Development Program through National Institutes of Health grant 5K12HD00849 and the American Society for Reproductive Medicine. 
2 Correspondence: Anuja Dokras, Dept. of Obstetrics and Gynecology, University of Iowa, 200 Hawkins Drive, Iowa City, IA 52242-1080. FAX: 319 353 6659; anuja-dokras{at}uiowa.edu
Accepted: June 5, 2001.
Received: March 12, 2001.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  A. P. Petty, K. L. Garman, V. D. Winn, C. M. Spidel, and J. S. Lindsey Overexpression of Carcinoma and Embryonic Cytotrophoblast Cell-Specific Mig-7 Induces Invasion and Vessel-Like Structure Formation Am. J. Pathol., May 1, 2007; 170(5): 1763 - 1780. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Mori, G. Ishikawa, S.-S. Luo, T. Mishima, T. Goto, J. M. Robinson, S. Matsubara, T. Takeshita, H. Kataoka, and T. Takizawa The Cytotrophoblast Layer of Human Chorionic Villi Becomes Thinner but Maintains Its Structural Integrity During Gestation Biol Reprod, January 1, 2007; 76(1): 164 - 172. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-T. Chen, T.-S. Lin, K.-C. Chow, H.-H. Huang, S.-H. Chiou, S.-F. Chiang, H.-C. Chen, T.-L. Chuang, T.-Y. Lin, and C.-Y. Chen Cigarette Smoking Induces Overexpression of Hepatocyte Growth Factor in Type II Pneumocytes and Lung Cancer Cells Am. J. Respir. Cell Mol. Biol., March 1, 2006; 34(3): 264 - 273. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
