Hum. Reprod. Advance Access originally published online on January 12, 2006
Human Reproduction 2006 21(5):1299-1304; doi:10.1093/humrep/dei489
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Stellate transformation of invasive trophoblast: a distinct phenotype of trophoblast that is involved in decidual vascular remodelling and controlled invasion during pregnancy
1 Department of Obstetrics & Gynecology, 2 Graduate Institute of Anatomy and Cell Biology and 3 Graduate Institute of Immunology, National Taiwan University Hospital and National Taiwan University College of Medicine, Taipei, Taiwan
4 To whom correspondence should be addressed at: Department of Anatomy and Cell Biology, College of Medicine, National Taiwan University, No. 1 Jen-Ai Road, Sect. 1, Taipei 100, Taiwan. E-mail: clc{at}ha.mc.ntu.edu.tw
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
BACKGROUND: Successful implantation relies on the tightly regulated invasion of extravillous trophoblasts (EVTs). However, little is known about their phenotypic differentiation and relevant motile behaviour. Furthermore, the cell–cell interactions between EVTs and decidual arterioles during physiological transformation are also poorly understood. METHODS: A total of 128 decidual specimens from early and late gestations containing components of EVTs and spiral arterioles were investigated using immunohistochemistry and periodic acid–Schiff reaction. RESULTS: Unipolar, tadpole-like EVTs are observed throughout the interstitial area, with a tendency to decrease along the invasive pathway. The stellate differentiation of the EVTs is identified around and inside decidual arterioles or in the third-trimester myometrium. Furthermore, stellate transformation of EVTs precedes its interactions with the decidual arteriole. These specialized stellate trophoblasts invade and infiltrate the tunica media, accompanying lacuna formation inside the vessel wall and perturbation of actin fibre alignment of the tunica media. CONCLUSION: Stellate transformation of trophoblasts may explain controlled invasion of EVTs and probably plays a key role in initiating cell–cell interaction in decidual vascular remodelling.
Key words: interstitial trophoblast/pregnancy/stellate transformation/vascular remodelling
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
How human extravillous trophoblasts (EVTs) perform controlled invasion and decidual vascular remodelling remains poorly understood. During interstitial invasion, a subset of these EVTs encroaches on the maternal endometrium, commingles with resident decidual and immune cells and eventually stops at the inner third of the myometrium (Pijnenborg et al., 1981
). During endovascular invasion, masses of cytotrophoblasts migrate into the tip of the uterine vessels, eventually remodelling the vasculature near the end of the first trimester (Benirschke and Kaufmann, 1994). Together, these two functional components of cytotrophoblast invasion anchor the placenta to the uterus and permit a steady increase in the supply of maternal blood to the growing fetus.
Most studies concerning EVTs have been performed on trophoblastic cell columns (Damsky et al., 1994; Genbacev et al., 1997) or villous explants growing on Matrigel (Caniggia et al., 1999; Leach et al., 2004). The issue of whether these observed interstitial trophoblasts represent the final stage of differentiation of this EVT population is still unresolved. Furthermore, the elaboration of the mechanisms underlying the cessation of EVT invasion along the invasive pathway also remains out of reach.
In vivo, EVTs penetrate the decidua and the first third of the myometrium while simultaneously altering the maternal vasculature to achieve the low-resistance and high-flow circulation of the intervillous space in the fetal placenta. The accumulation of current knowledge has provided consensus that the interstitial EVT plays an essential role in remodelling of the uterine decidual arterioles (Zhou et al., 1997, 1998; Kam et al., 1999; Dunk et al., 2003). However, these studies are mostly confined to determination of the causal relationship between the presence of interstitial EVTs surrounding the spiral artery and subsequent media transformation of the vessel wall. To better characterize these cellular interactions and the entire dynamic process, we used decidual curettage containing both elements of the fetal trophoblast and maternal decidual arteries to further investigate changes in their cellular aspects at the implantation site. Particular emphasis is also placed on the unique stellate transformation of EVTs and its inferred phenotypic characterization, and contribution to the controlled invasion and media alteration of decidual spiral arterioles during pregnancy.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Placental bed sampling
From January 2003 to April 2004, placental bed tissues were obtained during elective abortion from normal pregnancies because of psychosocial indications or from third-trimester Caesarean sections performed at the National Taiwan University Hospital (NTUH). Informed consents were obtained from all patients before the operation. The experiment protocol was also approved by the ethical committee of the NTUH. A total of 128 placental bed tissues ranging from 6 to 38 weeks of gestation ascertained to have decidual components were studied.
The placental bed tissues were washed three times in cold phosphate-buffered saline (PBS) to remove blood clots and then immediately frozen in liquid nitrogen within 20 min of curettage. Cryostat blocks were sliced to a thickness of 10 µm and then attached to the aminoalkylsilane-coated slides, air-dried and stored at –20°C until examination.
Antibodies, tissue preparation and immunohistocytochemistry
Identifications of trophoblasts and the muscular layer of the spiral artery at the fetal–maternal interface were based on the immunolocalization of cytokeratin 7 (CK7) for the trophoblasts and Phalloidin staining for the actin component of the muscular layer. The decidual cells were microscopically identified based on their typical features, which include enlarged and vesiculated nuclei and swollen cytoplasm. Sections prepared for immunohistochemical staining were fixed in 4°C, 4% paraformaldehyde for 10 min and then washed three times in PBS (pH 7.2) for 10 min. Prior to incubation with primary antibodies, sections were incubated for 30 min with 0.3% Triton and 10% goat serum in order to reduce the non-specific reactions of the secondary antibodies. After washing three times with PBS for 10 min, sections were incubated in a 1:100 dilution of the monoclonal mouse anti-human CK7 antibody (Sigma, MD, USA) at 4°C overnight. Omission of the primary antibody was used for the negative controls. The next day, a 1:200 dilution of Rhodamine-Phalloidin (Molecular Probes Inc., OR, USA) and a 1:1000 dilution of Hoechst 33342 reagents were applied after careful washing with PBS. The reaction was terminated 60 min later and thoroughly rewashed by PBS, finally mounted in Crystal-mount (Biomeda, CA, USA).
These slides were also stained with diaminobenzidine tetrahydrochloride (DAB; Sigma) reagent for bright-field examination. After incubation with the same primary antibody (anti-CK7), slides were washed thoroughly in PBS. Secondary antibody, biotinylated horse anti-mouse immunoglobulin G (Vector, CA, USA) at 1:200 dilution was added and incubated for 30 min. After careful washing in PBS, the avidin-biotin-peroxidase complex (ABC reagent; Vector) was prepared by adding single drops of Reagent A and of Reagent B to 2.5 ml of PBS, and it was then left at room temperature for 30 min. Slides were incubated with ABC reagent for 30 min and then washed with PBS. The peroxidase activity was demonstrated using DAB (Sigma), according to the manufacturer’s instructions. Slides were counterstained in haematoxylin or the periodic acid–Schiff (PAS) reagents to allow fibrinoid identification in the transformed spiral arteries.
Observations
Sections of decidual sampling after immunostaining were examined using a confocal microscope (Leica TCS-SP2, Heidelberg, Germany) equipped with an argon–krypton ion laser or a fluorescence microscope (Leica DMR, Microsystems, Weiziar GmbH, Germany) with a photographic system and a Nikon D1x digital camera (Nikon Co., Tokyo, Japan).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Specimens of placenta bed from 128 pregnancies met the following inclusion criteria: (i) first-trimester pregnancy (n = 113) with positive fetal pole and cardiac action and (ii) third-trimester pregnancy (n = 15).
Morphological characterization of interstitial EVTs along the invasive pathway in the first trimester
Under microscopy examination, clusters of EVTs in cell columns appeared as round or polygonal shape, attaching to each other or grouping in strings (Figure 1A). Those EVTs invading the interstitial area manifested an architecture remodelling from the round or polygonal shape to a unipolar configuration (eccentric nucleus) with a dominant pseudopod (Figure 1B). This architecture mimicked the extending neurite with a path-finding axon or the translocating fibroblast with rapid actin polymerization in the leading edge (Machesky, 2000). These tadpole-like EVTs could be seen throughout the entire interstitial area but were denser in the superficial decidua. Particularly, pseudopodia of these interstitial EVTs (InEVTs) were drawn up in orderly ranks compared with the randomly directed pseudopodia of InEVTs in the late gestation. In the deep decidua, most of the InEVTs demonstrated an intermediate phenotype between tadpole shape and round architecture. Their dominant and extended pseudopodia were either absent or inconspicuous (Figure 1C). EVT clusters were also identified between the myometrial fibres that border the deep decidua (Figure 1D). The EVT architecture in this region was transformed into round or polygonal shape again, as seen in the cell columns (Figure 1A). In brief, the presence of tadpole-like EVTs showed a tendency to decrease along the invasive pathway. The percentages of distribution for tadpole-like EVTs in different regions of placental bed are summarized in Table I.
|
|
Stellate transformation of EVT in the third trimester
The morphology of the third-trimester EVTs was also investigated using placental bed biopsies or en bloc specimens from Caesarean hysterectomy because of gynaecological indications. In the third trimester, the number and thickness of the trophoblastic cell columns were apparently decreased in comparison with the first trimester. These EVTs were remodelled into a stellate shape with five to 10 short cellular processes (Figure 1E). They were identified discretely in the loose fibrinoid layer of the basal plate, mostly residing in clear lacuna (Figure 1F). In the superficial decidua, the morphology of tadpole-shaped InEVTs was the same as those of the first trimester (Figure 1G). Nevertheless, the arrangement of pseudopodia in these InEVTs was rather chaotic. Compared with the first trimester, only scattered stellate-shaped EVTs were observed in the myometrium at term. These EVTs, which were several times larger than the InEVTs found in other areas, exclusively exhibited a centrally located nucleus arborized with 10 or more dendritic processes (Figure 1H).
The monoclonal antibody anti-CK7 also reacted with the residues of the maternal endometrial glands. However, these glands usually resided deep in the decidua. Furthermore, the glandular cells were arranged in a rosette or fence shape. Thus, they could be easily distinguished from the InEVTs.
Interactions between EVTs and the decidual arterioles in the first trimester
Before 7 weeks of gestation, there was no apparent interaction between these invasive forms of InEVTs and the decidual spiral arterioles. In some areas, however, many of the invasive InEVTs targeted and approached the surroundings of the decidual spiral arterioles (Figure 2A). Nonetheless, there was no change to the architecture of either the cytotrophoblasts or the vessel itself at this stage. Notably, the density of these InEVTs was not particularly high in the perivascular area.
|
Interactions between invasive InEVTs and maternal arterioles were particularly evident between 7 and 9 weeks of gestation. During this stage, aggregates of InEVTs were observed to colonize around the decidual spiral arterioles (Figure 2B). Some of these InEVTs in the perivascular area withdrew their dominant pseudopod and remodelled themselves into stellate forms (Figure 2C). Besides, some of the cellular processes of InEVTs were found to penetrate into the vessel wall. Occasionally, the cell body of these stellate InEVTs was completely buried in the tunica media of the decidual spiral arterioles (Figure 2D). Notably, only those InEVTs that had undergone stellate transformation could be found inside the tunica media. Up to the end of 8 weeks of gestation, more and more InEVTs achieved stellate transformation, invaded the decidual spiral arterioles and harboured inside the tunica media (Figure 2E). However, major changes in the arteriole wall, such as thinning or vacuolization, did not occur at this gestational age.
Investigating specimens obtained between the end of 8 weeks and the beginning of 9 weeks of gestation age, we found that stacks of InEVTs harboured around the decidual spiral arterioles and several layers of stellate InEVTs jammed inside the tunica media (Figure 2F). Using Rhodamine-Phalloidin examination, we found that the vessel wall itself appeared thinning in comparison with vessels without InEVT invasion (data not shown). All the InEVT–vessel interactions could be sometimes observed on the same slide at this gestational age, reflecting a gradual transition rather than an abrupt process of the InEVT–arteriole interaction. In brief, the percentages of the stellate trophoblasts appearing in perivascular region apparently step up with the increasing gestational age, particularly after 8 weeks’ gestation (Table II).
|
In comparison with the earlier developmental changes described above, the most pronounced event after 9 weeks of gestation was the appearance of abundant endovascular EVTs (EnEVTs) inside the lumen of the decidual spiral arteriole (Figure 2G). After extensive examination, it was noted that this retrograde EnEVT migration only occurred in those arterioles with stellate InEVT infiltration. This retrograde migration of EnEVTs was never identified in the vessels without stellate InEVT invasion. Meanwhile, the tendency to decreasing stellate InEVTs was also noted in the perivascular region after retrograde EnEVT invasion. In these fully transformed vessels, the vascular lumen was filled with plugs of endovascular trophoblasts. Only scattered stellate InEVTs with weak reactivity of anti-CK7 could be observed, leaving a largely acellular area around the vessel (Figure 2H). This acellular area could be confirmed nucleus negative using haematoxylin staining (data not shown).
We also examined the PAS reaction in these interactive steps. Negative PAS reaction was evidenced by purplish red in the vessel without stellate InEVT invasion. However, a bright magenta-coloured vessel wall where the InEVTs infiltrated the tunica media indicated a positive PAS reaction (Figure 3A and B). Notably, PAS staining revealed that some of these stellate InEVTs resided in large, clear lacunae of tunica media (Figure 3B), indicating the deposition of extracellular matrix from these stellate InEVTs.
|
The stellate InEVTs in the perivascular area were also examined using confocal laser microscopy. The confocal examination demonstrated that these stellate InEVTs were buried in the tunica media of the decidual spiral arterioles, and not an artefactual overlap because of slice thickness (Figure 4A). The F-actin staining was localized in the cortex of the entire cell body, and the ruffles, numerous fine bursts and dendritic processes emanating from its surface, as the actin distribution of a ‘pause and explore’ fibroblast (Machesky, 2000). The arteriole endothelium was not disrupted before the retrograde EnEVT migration. Particularly, the muscle fibres of tunica media were perturbed in the presence of the stellate InEVTs. Furthermore, the tadpole-shaped InEVTs in the interstitial area and the large stellate InEVTs surrounding the decidual arterioles could be observed on the same slide (Figure 4B), confirming the fact that the stellate transformation of the InEVTs was not an artefact of specimen preparation or staining.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Although this InEVT invasion pathway has already been well described in the literature, little is known about the motile behaviour of InEVTs at either location. A recent in vitro study of dendritic cell (DC) differentiation (Shutt et al., 2000
) may serve as a model to enhance the understanding of the in vivo counterpart of the InEVT invasion pathway. The InEVTs exiting the trophoblastic cell column and invading the interstitial area (Figure 1B) simulate the process of extravasation of undifferentiated, highly motile DCs from the bloodstream and migration into the peripheral tissue (phase I DCs), whereas EVTs harbouring inside the fibrinoid of cell columns (Figure 1E and F) and myometrium (Figure 1H) of the third trimester, or perivascular area beyond the end of 8 weeks’ gestation (Figure 2C–H), show the arborized, stellate characteristics of DCs that stay in tissue for antigen capture (phase II DCs). Morphologically, in phase I DCs, a dominant pseudopod is extended in the direction of cellular translocation, as in the phenotype of Dictyostelium amoebae (Wessels and Soll, 1998) or the unipolar, tadpole-shaped InEVT examined in this study. Using computer-aided motion analysis, persistence and directionality of phase I DC migration was observed. Notably, although absent or minimal net translocation was revealed for phase II stellate DCs, major dendritic processes of them were extended in a spatially random fashion, reflecting the non-directional movement of the cellular centroid (Shutt et al., 2000). These dynamic analyses of the distinct DC phenotypes may help infer the unique behavioural states and in vivo function underlying the architecture changes of InEVTs in response to various cues along their invasive pathway.
Until recently, debate still existed with respect to the process underlying the physiological changes to the spiral arteries during pregnancy (Kaufmann et al., 2003). On the basis of human studies, most researchers suggest that InEVTs invade the decidual vasculature and enter the lumen, being the leading cause of the loss of normal musculoelastic structure of spiral arteries. In fact, although extensive infiltration of tunica media by stellate EVTs was observed in this study, breaching endothelium by them or even protruding into vascular lumen was not identified. Moreover, the morphology of endovascular trophoblasts more or less resembles EVTs residing in the proximal cell column (mostly round shape) but differs greatly from intramural trophoblasts (stellate shape) (Figure 3A). It seems more reasonable to assume that the endovascular trophoblast is derived from a source other than intramural trophoblasts. A combination of both hypotheses had been suggested (Kam et al., 1999) and was concordant with our observations in the present study. In this conceptualization, InEVT prompts fibrinoid change in the smooth-muscle layer of the spiral artery, and EnEVT replaces the resident endothelial cells. These conflicts concerning physiological changes to the spiral arteries during pregnancy highlight the need for further elaboration of the mechanisms underlying the critical cell–cell interactions occurring in vascular remodelling.
This stellate transformation of InEVT serves as a starting point for elucidation of the tightly regulated invasion process and decidual vascular remodelling. We have assumed two important in vivo functions for this stellate InEVT transformation. First, InEVTs performing stellate transformation lose their motile behaviour and invasion capability, which is characteristic of phase II DCs (Shutt et al., 2000). In the first trimester, tadpole-shaped InEVTs possibly may acquire actin assembly for both protrusion and retraction of cell body in response to the cue present in the interstitial area. Spatial gradients of attractants from respective target tissues should suppress lateral pseudopod formation, as is the case for Dictyostelium amoeba (Wessels and Soll, 1998), aligning these InEVTs in a queue (Figure 1B). Once these invasive InEVTs approach the spiral artery, they transform into the stellate phenotype and lose their motility, possibly due to the homogenized gradient of cues in the surroundings of the decidual vessel. This finding may explain the results of a previous study, where it was demonstrated that InEVTs usually colonize around the decidual spiral artery (Kam et al., 1999). Second, matrix-type fibrinoid of the placental bed is derived from the secretory product of invasive EVTs (De Wolf et al., 1973; Kaufmann et al., 1996). Interestingly, we found that most of the locations of matrix-type fibrinoid (fibrinoid of cell column and intramural fibrinoid) are co-localized with the distribution of stellate EVTs, suggesting interrelations between the specific in vivo functions of stellate EVTs and matrix-type fibrinoid. For example, the clear lacuna (Figure 3B) produced via the deposition of fibrinoid by stellate EVTs may further prompt perturbation of actin fibres in the vessel media (Figure 4A), contributing to the subsequent formation of a flaccid sac-like decidual vessel. Likewise, the intramural fibrinoid may also be involved in inducing the invasive and adhesive properties of endovascular trophoblasts via the changes of oncofetal fibronectin and intergrin repertoires.
In summary, this study has characterized the unique spatio-temporal regulation of InEVT phenotypic differentiation and the delicate cell–cell interaction during spiral artery transformation. Our observations also indicate that the stellate transformation probably represents a ‘pause and explore’ phenotype of InEVT that explains the confinement of the InEVT invasion within the superficial myometrium and probably sheds light on how InEVT elicits cell–cell interactions in decidual vascular remodelling. Clearly, many questions arise to be clarified in the future, such as molecular mechanisms of the stellate transformation of InEVT and also cues for EnEVT invasion.
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
This article was sponsored by a grant from National Science Council of Taiwan (92-2314-B-002-155).
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAcknowledgementsReferences |
|---|
Benirschke K and Kaufmann P (1994) Non-villous part of the placenta. In Benirschke K and Kaufmann P (eds) Pathology of the Human Placenta. Spring-Verlag, New York, pp. 182–267.
Caniggia I, Mostachfi H, Winter J, Gassmann M, Lye SJ, Kuliszewski M and Post M (1999) Hypoxia-inducible factor-1 mediates the biological effects of oxygen on human trophoblast differentiation through TGF
3. J Clin Invest 105,577–587.
Damsky CH, Librach C, Lim KH, Fitzgerald ML, McMaster MT, Janatpour M, Zhou Y, Logan SK and Fisher SJ (1994) Integrin switching regulates normal trophoblast invasion. Development 120,3657–3666.[Abstract]
De Wolf F, De Wolf-Peeters C and Brosens I (1973) Ultrastructure of the spiral arteries in the human placental bed at the end of the normal pregnancy. Am J Obstet Gynecol 117,833–848.[Web of Science][Medline]
Dunk C, Petkovic L, Baczyk D, Rossant J, Winterhager E and Lye S (2003) A novel in vitro model of trophoblast-mediated decidual blood vessel remodeling. Lab Invest 83,1821–1828.[CrossRef][Web of Science]
Genbacev O, Zhou Y, Ludlow JW and Fisher SJ (1997) Regulation of human placental development by oxygen tension. Science 277,1669–1672.
Kam EP, Gardner L, Loke YW and King A (1999) The role of trophoblast in the physiological change in decidual spiral arteries. Hum Reprod 14,2131–2138.
Kaufmann P, Huppertz B and Frank HG (1996) The fibrinoid of the human placenta: origin, composition and functional relevance. Ann Anat 178,485–501.[Web of Science][Medline]
Kaufmann P, Black S and Huppertz B (2003) Endovascular trophoblast invasion: implications for the pathogenesis of intrauterine growth retardation and preeclampsia. Biol Reprod 69,1–7.
Leach RE, Kilburn B, Wang J, Liu Z, Romero R and Armant DR (2004) Heparin-binding EGF-like growth factor regulates human extravillous cytotrophoblast development during conversion to the invasive phenotype. Dev Biol 266,223–237.[CrossRef][Web of Science][Medline]
Machesky LM (2000) Putting on the brakes: a negative regulatory function for Ena/VASP proteins in cell migration. Cell 101,685–688.[CrossRef][Web of Science][Medline]
Pijnenborg R, Bland JM, Robertson WB, Dixon G and Brosens I (1981) The pattern of interstitial trophoblastic invasion of the myometrium in early human pregnancy. Placenta 2,303–316.[Web of Science][Medline]
Shutt DC, Daniels KJ, Carolan EJ, Hill AC and Soll DR (2000) Changes in the motility, morphology, and F-actin architecture of human dendritic cells in an in vitro model of dendritic cell development. Cell Motil Cytoskeleton 46,200–221.[CrossRef][Web of Science][Medline]
Wessels D and Soll DR (1998) Computer-assisted analysis of cytoskeleton mutants of Dictyostelium. In Soll DR and Wessels D (eds) Motion Analysis of Living Cells. Wiley-Liss Inc., New York, pp. 101–140.
Zhou Y, Damsky CH and Fisher SJ (1997) Preeclampsia is associated with failure of human cytotrophoblasts to mimic a vascular adhesion phenotype. One cause of defective endovascular invasion in this syndrome? J Clin Invest 99,2152–2164.[Web of Science][Medline]
Zhou Y, Genbacev O, Damsky CH and Fisher SJ (1998) Oxygen regulates human cytotrophoblast differentiation and invasion: implications for endovascular invasion in normal pregnancy and in pre-eclampsia. J Reprod Immunol 39,197–213.[CrossRef][Web of Science][Medline]
Submitted on June 16, 2005; resubmitted on October 18, 2005; accepted on December 13, 2005.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||