Biology and Medicine
Editor-in-Chief: Eytan R. Barnea MD, FACOG
SIEP publications are protected by copyright and all
rights are reserved. SIEP publications may not be reproduced in any form or by any
means without written permission from the copyright owner. This includes the posting of
electronic files on the Internet, transferring electronic files to other persons,
distributing printed output, and photocopying. Requests for authorization to make
photocopies should be directed to: SIEP, 1697 Lark Lane, Cherry Hill, NJ 08003-3157
Tel: (856) 429 2699, Fax: (856) 429 7414, Email: barnea@earlypregnancy.org.
|
EARLY
PREGNANCY: Biology and Medicine Editor-in-Chief: Eytan R. Barnea MD, FACOG |
July 2000
Volume IV, Number 3
ISSN: 1537-6583
Pages: 176-190
Alison J. Trew1, Gendie E. Lash2
and Philip N. Baker2
1 School of Biochemistry and Molecular Biology, University of Leeds,
Leeds, LS2 9JT, United Kingdom, 2School of Human Development, Department of
Obstetrics and Gynaecology, University of Nottingham, City Hospital, Hucknall Road,
Nottingham, NG5 1PB, United Kingdom
Short Title: In vitro trophoblast invasion model
Key Words: cytotrophoblast, cell invasion, Matrigel Invasion Chamber, vascular endothelial growth factor, epidermal growth factor
Correspondence: Professor
PN Baker, School of Human Development, Department of Obstetrics and Gynaecology,
University of Nottingham, City Hospital, Hucknall Road, Nottingham, NG5 1PB, United
Kingdom, Tel: 0115 9627914, Fax: 01159627670,
Email: philip.baker@nottingham.ac.uk
Acknowledgements: We are grateful for
the support of Action Research and the Joan Dawkins Fellowship (BMA).
Pre-eclampsia, a multisystemic disorder characterised by hypertension and proteinuria, and intrauterine growth restriction (IUGR) are responsible for considerable perinatal mortality and morbidity. Although both complications of pregnancy are often only apparent in late pregnancy, there is accumulating evidence that they share a common pathogenesis, whereby deficient cytotrophoblast invasion of the placental bed spiral arteries in early pregnancy leads to a poorly perfused feto-placental unit (Lunell et al., 1982). Compared with normal pregnancy, cytotrophoblast cells in pre-eclampsia show an increased ability to proliferate (Redline and Patterson, 1995) and the trophoblast migration and invasion that does occur is largely restricted to the decidual portion of the spiral arteries (Robertson et al., 1975; Brosens et al., 1977).
The mechanisms underlying this deficient trophoblast invasion are presently unclear. Many growth factors (especially angiogenic growth factors) have been implicated in playing a role in both the autocrine and paracrine regulation of cytotrophoblast differentiation and function. Much of the evidence for the involvement of growth factors in cytotrophoblast function is derived from expression and localisation studies of these factors and their receptors in either the developing placenta or the surrounding maternal tissues (Simon et al., 1994; King et al., 1995; Bennet et al., 1996). However, very little is known about the biological role of these cytokines in cytotrophoblast invasion.
Many studies into the biological activity of these cytokines on cytotrophoblast invasion have used an in vitro invasion model. This model uses purified first trimester cytotrophoblast cells which are then cultured on Matrigel® -coated transwell inserts containing PET filters with 8 µm pores. Insulin-like growth factor II and its associated binding protein, insulin-like growth factor binding protein 1, interleukin-1ß and EGF have all been shown to increase cytotrophoblast invasion through Matrigel® , whilst the platelet-derived growth factors (PDGF-AA and PDGF-BB) and tumour necrosis factor-a were found to have no effect (Bass et al., 1994; Librach et al., 1994; Irving and Lala, 1995). In contrast transforming growth factor-ß has been found to reduce in vitro trophoblast invasion (Irving and Lala, 1995), although Bass et al. (1994) found that TGF-ß had no effect on cell invasion. VEGF and placental growth factor (PlGF) have both been shown to have no effect on trophoblast invasion using this model (Athanassiades et al., 1998; Athanassiades and Lala, 1998). The pathways through which these factors act are very poorly understood. However, IL-1ß has shown to stimulate trophoblast secretion of the 92 kDa type IV collagenase, matrix metalloproteinase (MMP)-9 as well as metalloproteinase activity (Librach et al., 1994). This suggests that growth factors may control trophoblast invasiveness through regulation of adhesion molecules or that matrix components may be responsible for insufficient trophoblast invasion in diseases characterised by poor placentation (Damsky et al., 1992; Zhou et al., 1993).
It was the intention of this study to re-investigate the methods used to visualise invaded cells in the Matrigel® invasion model using standard and confocal fluorescent microscopy. For this study two growth factors were used, one which had previously been shown to affect cytotrophoblast cell invasion (EGF) and one which had not (VEGF).
Materials and Methods
Cytotrophoblast isolation and culture
Placentae were obtained immediately after first trimester (8-12 week) terminations
and placed in Ham’s F10 media with 1000U/ml penicillin, 1 mg/ml streptomycin, 2.5
mg/ml amphotericin-B, 2 mM glutamine, 1U/ml heparin on ice. Cytotrophoblast cells were
isolated from pools of multiple first trimester placentae (2-4 patient samples) using
methods established by Kliman et al. (1986) and widely used for the isolation of
cytotrophoblast cells from term placenta. In brief, the membranes and connective tissue
were removed and the remaining tissue was digested with trypsin (0.25%) and DNase (0.2
mg/ml) in pre-warmed Hanks’ solution containing 25 mM HEPES in a shaking water bath
at 37° C for 30 minutes. The tissue fragments were allowed to settle and the supernatant
was layered over 5 ml aliquots of newborn calf serum and centrifuged at 1000 x g for 10
minutes at room temperature. The pelleted cells were resuspended in Dulbecco’s
Modified Eagle’s Medium (DMEM) at room temperature. The remaining placental tissue
was subjected to the digestion procedure two more times with the addition of fresh
trypsin-DNase solution. All of the resultant cell suspensions were pooled, centrifuged at
1000 x g, and resuspended in 6 ml DMEM. This suspension was divided in half and layered
over two discontinuous Percoll density gradients (5ml 70%, 4 ml 55%, 4 ml 40% and 5 ml 30%
Percoll in Hanks’ solution). The Percoll gradients were centrifuged at 12000 x g at
room temperature for 30 minutes. The band of cells between the 40% and 55% Percoll layers
contained an enriched population of cytotrophoblast cells. These cells were collected and
washed once in culture medium. In all cases remaining leucocytes and mesenchymal cells
were removed by incubation with immunomagnetic beads (Dynal UK Ltd.) coated with antibody
to leucocyte common antigen (LCA, CD45) and antibody to vimentin for 30 minutes at 4° C
(both antibodies were obtained from Dako UK Ltd.). The resulting cells were cultured in
DMEM:Ham’s F12 (1:1) with 25 mM HEPES (pH 7.4), 2 mM glutamine and antibiotics, and
either 10% fetal calf serum or 2% serum supplement (TCH Supplement, ICN) in a humidified
5% CO2/95% air incubator at 37° C.
Fibroblast isolation and culture
Fibroblast cells were isolated from first trimester placentae at the same time as
the cytotrophoblast cells were isolated. On a Percoll gradient fibroblast cells collect at
concentrations less than 40% Percoll (the top two layers). These layers were removed and
treated in a similar manner as the cytotrophoblast cells (without immunoadsorption) as
described above. The fibroblast cells were cultured under the same conditions as the
cytotrophoblast cells.
Immunostaining
To determine the purity of the cytotrophoblast or fibroblast cell population, cells
were plated on a 24-well plate at 106 cells per well in media containing 10%
fetal calf serum. On the day after cell isolation, the cells were washed twice with warm
phosphate-buffered saline (PBS), fixed with methanol for 10 minutes, washed twice with PBS
and stained with antibodies to cytokeratin which reacts with cytotrophoblast cells but not
other villous components, (MNF116, Dako UK Ltd.), vimentin which is expressed by
mesenchymal cells (V9, Dako UK Ltd.), leucocyte common antigen (LCA/CD45) (T29/33, Dako UK
Ltd.) or von Willebrand factor which reacts with endothelial cells (vWF) (F8/86, Dako UK
Ltd.). Briefly, the cells were pre-incubated for 30 minutes in 20% goat serum in PBS, then
incubated for 1 hour with either anti-cytokeratin diluted 1:50 (v/v) in 20% goat serum,
vimentin (1:20), LCA (1:10) or vWF (1:25). The cells were washed in PBS (three times for
10 minutes) and incubated for 30 minutes with fluorescein-conjugated goat anti-mouse IgG
(Sigma Chemical Co., Poole, UK) diluted 1:50 (v/v) in 10% goat serum. Cells were viewed
with an inverted fluorescent microscope (Nikon DIAPHOT 300). The number of cells in five
fields of view in each well were counted. The purity of the cytotrophoblast cells within
the isolated cells was estimated from the proportion of cytokeratin positive cells
recorded. The purity of the fibroblast cells isolated was estimated from the proportion of
vimentin positive cells recorded. The Matrigel® -coated transwell inserts were
stained in a similar manner using anti-cytokeratin for the cytotrophoblast experiments and
anti-vimentin for the fibroblast experiments.
Invasion assay
Cytotrophoblast (2.5 x 105 cells) in DMEM/F12 media containing 2% serum
supplement were cultured on Matrigel® -coated transwell inserts (6.5 mm
diameter) containing 8 µm pore size polycarbonate membrane (BioCoat Matrigel® Invasion
Chamber, Becton Dickinson) for 12 hours to 5 days as indicated in the text. Fibroblasts
(2.5 x 105 cells) in DMEM/F12 media containing either 0.5% or 10% FCS were
cultured in the same manner as the cytotrophoblast cells for 4 days. Cytotrophoblast cells
were either cultured in media alone or with the addition of 10ng/ml EGF (Sigma Chemical
Co., Poole, UK) or 5ng/ml VEGF (PeproTech, London, UK) in accordance with previous studies
(Bass et al., 1994; Athanassiades et al., 1998). After incubation the
cultures were rinsed in PBS and the non-invading cells were removed from the upper surface
of the membrane by wiping the surface with a muslin wrapped cotton bud. The cells were
fixed in methanol for 10 minutes and washed twice in PBS. The filters were dissected from
the inserts with a scalpel blade and stained with anti-cytokeratin antibody
(cytotrophoblasts) or anti-vimentin (fibroblasts) as described above. Invasion was
quantified by counting the number of cells and cell processes in five fields of view at
magnification x 100 on each membrane using a fluorescent microscope. Each field of view at
this magnification showed approximately on ninth of the total area of the membrane.
Cytotrophoblast cells were incubated on Matrigel® -coated transwells over a five day period in culture medium containing 2% serum supplement. Media was collected from the lower well beneath the transwell membrane and was centrifuged at 1300 rpm for 4 minutes in a cytospin. The slides were fixed and stained with anti-cytokeratin as described above.
To determine the interassay variation, cytotrophoblast cells were incubated on Matrigel® -coated transwell inserts for five days in media containing 2% serum supplement. Three replicate membranes were fixed and stained every 24 hours. Cytokeratin-positive cells on the underside of the membrane were counted. Five fields of view were counted on each membrane. The same analysis was performed on cytotrophoblast cells incubated for five days in media supplemented with 10 ng/ml EGF (Sigma Chemical Co., Poole, UK). The coefficient of interassay variation was 9%. Since the number of wells that can be used in any experiment is limited by the number of cells isolated from placental tissue, replicate wells were not used in subsequent experiments. The coefficient of intra-assay variation was 21%.
Growth factor reduced Matrigel®
BioCoat transwell inserts (6.5 mm diameter) containing an 8 µm pore size membrane
were coated with either 15 µl of pre-cooled Matrigel® (Constituent EGF
0.5-1.3 ng/ml) or 15 µl Growth Factor-Reduced Matrigel® (GFR-Matrigel®)
(constituent EGF less than 0.5 ng/ml). Cytotrophoblast cells (2.5 x 105 cells)
were cultured on these inserts for 72 hours in media containing 2% serum supplement. In
addition, transwell inserts were coated with Matrigel® and GFR-Matrigel®
containing 5 ng/ml anti-EGF (PeproTech) to neutralise the potential effect of constituent
EGF. Cytotrophoblast cells (2.5 x 105 cells) were cultured on these inserts for
72 hours. The medium added to the top and bottom of these cultures contained 2% serum
supplement and 5 ng/ml anti-EGF.
Confocal microscopy
Cytotrophoblast cells were incubated in two Matrigel® -coated
transwells for 72 hours in medium containing 2% serum supplement. The membrane was stained
with anti-cytokeratin and examined under a confocal microscope (Leica DM RBE). The upper
surface of one of the transwell membranes was wiped with a cotton swab but in the second
transwell the cells on the upper surface of the membrane were not wiped off, hence the
Matrigel® was intact. The area examined included the underside as well as the
upper surface of the membrane and its Matrigel® coating. The image was black
unless fluorescent cytokeratin-positive cells were present. In addition a Matrigel®
-coated transwell was incubated for 72 hours in identical culture media but without any
cells seeded on the upper surface of the transwell membrane.
Statistical analysis
Statistical analysis of data was performed using Mann-Whitney U test. All data are
expressed as medians +/- interquartile ranges. Statistical significance was set at
p<0.05.
The fibroblast cells isolated from first trimester placentae were 100% vimentin positive, with no staining with anti-cytokeratin or anti-CD45.
Cell invasion
Fluorescent Microscopy
An initial experiment using fluorescent microscopy to count cytokeratin-positive
cells indicated that EGF significantly increased the rate of cytotrophoblast invasion by
approximately two-fold at 24 hours (p<0.01, Mann-Whitney U test), but at 48 hours EGF
no longer had any significant effect on invasion (Figure 1).
In addition VEGF significantly increased cell invasion at 24 hours (p<0.01,
Mann-Whitney U test), also by two-fold. However, at 48 hours VEGF appeared to have a
significant inhibitory effect on cell invasion (p<0.001, Mann-Whitney U test).
In four subsequent experiments cell invasion was measured at 12, 24, 36 and 48 hours (Figure 2). EGF significantly increased the number of invaded cells on the underside of the membrane only at 12 hours (p<0.05, Mann-Whitney U test). VEGF did not significantly increase or decrease the number of invaded cells counted on the underside of the membrane at any incubation period. However, there was no consistency in the observed effects of either EGF or VEGF at 12, 24, 36 and 48 hours between the four experiments: at 12 hours EGF significantly increased cell invasion in only 1 of 4 experiments and at 48 hours cell invasion was significantly increased with EGF in 2 of 4 experiments but was reduced in 1 of 4 experiments (p<0.05, Mann-Whitney U test). VEGF significantly reduced cell invasion in 1 of 4 experiments at 48 hours. Notably, cell invasion in the control wells increased over 48 hours in only one of the experiments. In two experiments the median number of cells counted on the underside of the membrane at 48 hours was less than at 36 hours. In contrast, cells incubated for longer periods demonstrated an increasing cell invasion over 4 days, but a decrease in the number of cells on the underside of the membrane on day 5 compared with day 4.
In three experiments media from the lower well was fixed on slides and stained with anti-cytokeratin to determine whether invaded cells were present in the media beneath the membrane. Positive cytokeratin-staining was observed in media collected after 12 hours incubation up to 5 days incubation in all three experiments (Figure 3). However, cell bodies were smaller than the cytotrophoblast cells seeded on the upper surface of the membrane suggesting either that cells were present in the lower well but had lysed, or that some component in the culture media reacts positively with anti-cytokeratin, or that there is some cross-reactivity between the IgG-FITC conjugate and a component in the culture media. Aliquots of the same media tested for trypan blue exclusion indicated that the media obtained from the lower well did not contain any whole cells, either viable or dead. Cells were not attached to the surface of the lower well in any of the experiments.
Fibroblast cells were able to successfully invade through the of Matrigel® -coated transwells in response to 10% FCS over a 4 day period (Figure 4). There was no evidence of cells in the media of the bottom chamber of the culture dish suggesting that all cells were attached to the lower part of the membrane.
Growth Factor Reduced Matrigel®
The number of cells detected on the membranes of Matrigel® -coated
transwells and GFR-Matrigel® -coated transwells was not significantly
different from commercially available invasion chambers at 72 hours (Figure 5). There was no significant difference between
Matrigel® -coated wells with and without anti-EGF or between GFR-Matrigel®
-coated wells with and without anti-EGF.
Fluorescent Confocal Microscopy
Cytotrophoblast cells that had been incubated in Matrigel® invasion
chambers for 72 hours were stained with anti-cytokeratin and examined under a confocal
microscope. In two separate experiments cytokeratin-positive cells were observed on the
upper surface of the membrane but nothing was visible within the Matrigel® ,
in the pores of the membrane, or on the underside of the membrane (Figure 6). A high background fluorescence was observed
within these layers from the fluorescent cells on the surface of the Matrigel. A duplicate
membrane was examined that was stained with antibody to cytokeratin after the Matrigel on
the upper surface had been wiped off the membrane as described earlier. There was no
evidence of any cytokeratin-positive cells on the lower surface of the membrane or on the
upper surface but some pores within the membrane did fluoresce for approximately three
serial 1 µm sections. However, the same pores were negative for cytokeratin above and
below this 3 µm portion of the membrane. In addition, Matrigel® -coated
transwells were incubated for 72 hours in culture media without cytotrophoblast cells
seeded on the upper surface of the Matrigel. Staining with anti-cytokeratin was positive
within some of the pores but there was no evidence of a cell emerging from either side of
the same pores.
Discussion
Our initial data suggested that EGF significantly increased the rate of cytotrophoblast invasion by approximately two-fold at 24 hours but by 48 hours no effect was observed. Thus we believed that we were able to reproduce the findings of Bass et al. (1994). In the same experiment VEGF produced a significant decrease in cell invasion at 48 hours which we thought might be due to the invaded cells that had emerged on the underside of the membrane becoming detached. However, we were unable to repeat this finding, and in subsequent experiments the number of invaded cells in the control wells was sometimes less at 48 hours than at earlier time points. In cells cultured for longer incubation periods the number of cells on the underside of the membrane increased with time but was less at 5 days compared with 4 days. Since positive cytokeratin staining was observed on the lower side of the membrane in all of these experiments, this suggested that invaded cells could detach themselves from the membrane and be released into the media of the lower well. We initially thought that the variation in our results reflected biological variations in cell viability, cell invasion and growth factor expression on the cell surface between different cell preparations. Because we were unable to establish a trend in any one of the experiments we performed using BioCoat Invasion Chambers® or between the laboratory prepared Matrigel® -coated transwells and the GFR-Matrigel® -coated transwells, we postulated that the growth factor receptors expressed on the cytotrophoblast cell surface were maximally stimulated by EGF and other growth factors that are present within the Matrigel® . Hence the levels of exogenous growth factor in the culture media could have no additional effect.
If cytotrophoblast cells invaded Matrigel® and were released into the lower well, we expected to find positive cytokeratin staining of cells within the media. Although positive staining was observed it was not typical of cytokeratin staining of trophoblast cells isolated from first trimester placentae. This and trypan blue analysis suggested that the staining was either dead cellular fragments or artefactual staining. Confocal microscopy demonstrated unequivocally that the cytokeratin staining was indeed an artefact produced in the method since positive staining was observed on membranes that were cultured without ever being in contact with a cytotrophoblast cell.
These findings suggest that cytotrophoblast cells isolated from first trimester placentae do not invade Matrigel® or GFR-Matrigel® whereas fibroblast cells from the same preparation do. Thus, we are at odds with the previous studies that describe an in vitro model of trophoblast invasion. The viability of cells isolated from first trimester placenta in our laboratory in most cases was 100%, and we achieved on average 94% pure trophoblast cells within each isolation. Thus, our method of cell isolation appears to concord with the current literature. The work carried out in our laboratory used the same culture media containing a serum supplement, transwell inserts containing 8 µm pores, and laminin-rich extracellular matrix substrate Matrigel® . However, we feel that we have demonstrated that these cells do not invade Matrigel® -coated transwells. In a different assay system using first trimester trophoblast cells isolated in the same manner and EGF and VEGF at the same concentrations we have shown that both growth factors increase the rate of syncytialisation in these cells (Strachan et al., 1999). This supports that these growth factors do have biological activity in this cell type although not in the assay system described in this study.
Different methods of quantification have been employed in previous in vitro studies of trophoblast invasion using the Matrigel® system. Several studies have examined filters sputter-coated with gold palladium in a scanning electron microscope to produce photographs that can be analysed to estimate the surface area of the filter covered by invasive cells (Bass et al., 1994; Librach et al., 1994; Damsky et al., 1994). In a recent publication confocal microscopy was used to count the number of fluorescently labelled cells viewed in serial 1 µm sections parallel to the membrane (Genbacev et al., 1996). In contrast some groups have used Wright-Giemsa staining or haematoxylin-eosin staining and counted stained cells under a light microscope (Shimonovitz et al., 1994), however this method is only valid for a known pure cell population. Our results caution that cytokeratin-positive staining observed under a fluorescent microscope is not necessarily indicative of epithelial cells and is probably a staining artefact of the Matrigel® insert itself. In the light of our findings we suggest that more than one method of cell visualisation is used when assessing the invasiveness of cells in the Matrigel® assay.
Due to the described problems with the Matrigel® assay system it is possible that we need to start finding alternative models. A different model has just been described where trophoblast cells are attached to a cytocarrier bead which is then embedded into a fibrin clot, the growth factor of interest is then added to the media overlaying the clot and the number and length of processes assessed. This method has been used to describe both a positive and a negative effect of growth factors on trophoblast invasion using hepatocyte growth factor (Cartwright et al., 1999) and VEGF (Lash et al., 1999) respectively.
For personal use. Only reproduce with permission from SIEP.
Lash, G.E., Cartwright, J.E., Whitley, G.StJ., Trew, A.J. and Baker, P.N. (1999). The effects of angiogenic growth factors on extravillous trophoblast invasion and motility. Placenta, In Press
Librach, C.L., Feigenbaum, S.L., Bass, K.E., Cui, T.Y., Verastas, N., Sadovsky, Y., Quigley, J.P., French, D.L. and Fisher, S.J. (1994). Interleukin-1 beta regulates human cytotrophoblast metalloproteinase activity and invasion in vitro. J. Biol. Chem. 269, 17125-17131
Lunell, N.O., Nylund, L.E., Lewander, R. and Sarby B. (1982). Uteroplacental blood flow in pre-eclampsia measurements with indium-113m and a computer-linked gamma camera. Clin. Exp. Hypertens. 1, 105-117
Redline, R.W. and Patterson, P. (1995). Pre-eclampsia is associated with an excess of proliferative, immature intermediate trophoblasts. Human Pathol. 26, 594-600
Robertson, W.B., Brosens, I. and Dixon, G. (1975). Uteroplacental vascular pathology. Eur. J. Obstet. Gynaecol. Reprod. Biol. 5, 47-65
Shimonovitz, S., Hurwitz, A., Dushnik, M., Anteby, E., Geva-Eldar, T. and Yagel, S. (1994). Developmental regulation of the expression of 72 and 92 kd type IV collagenases in human trophoblasts: A possible mechanism for control of trophoblast invasion. Am. J. Obstet. Gynaecol. 171, 832-838
Simon, C., Frances, A., Piquette, G., Hendrickson, M., Milki, A. and Polan, M.L. (1994). Interleukin-1 system in the materno-trophoblast unit in human implantation: Immunohistochemical evidence for autocrine/ paracrine function. J. Clin. Endocrinol. Metab. 78, 847-854
Strachan, B.K., Lash, G.E. and Baker, P.N. (1999). The syncytialisation of isolated trophoblast cells is enhanced by vascular endothelial growth factor but not by placental growth factor. J. S. G. I. 6(1s), 218A
Zhou, Y., Damsky, C.H., Chiu, K., Roberts, J.H. and Fisher, S.J. (1993). Preeclampsia is associated with abnormal expression of adhesion molecules by invasive cytotrophoblasts. J. Clin. Invest. 91, 950-960
Effect of EGF (10ng/ml) and VEGF (5ng/ml) on cytotrophoblast cells cultured for 48 hours on Matrigel® -coated transwells. Data from one experiment: values are medians and interquartile ranges of the number of cells counted on the lower side of the membrane in five fields of view per membrane. * p<0.05, * * p<0.01 (Mann-Whitney U test).
Effect of EGF (10ng/ml) and VEGF (5ng/ml) on cytotrophoblast cells cultured for 48 hours on Matrigel® -coated transwells. Data from four experiments using one membrane per experiment: values are medians and interquartile ranges of the number of cells counted on the lower side of the membrane in five fields of view per membrane. * p<0.05, * * p<0.01 (Mann-Whitney U test).
Cytokeratin staining of a cytospin of culture media from below the transwell membrane: cytotrophoblast cells incubated for 5 days on Matrigel® -coated transwells. Media from the lower chamber of a transwell fixed onto slides and stained with anti-cytokeratin antibody and examined by a fluorescent microscope (100x).
Effect of fetal calf serum (FCS) (0.5 and 10%) of fibroblast cells cultured for 80 hours on Matrigel® -coated transwells. Data from one experiment using three membranes per experiment: values are medians and interquartile ranges of the number of cells counted on the lower side of the membrane in five fields of view per membrane. p<0.0001 (Mann-Whitney U test).
Effect of anti-EGF (5ng/ml) on cytotrophoblast invasion in vitro on Matrigel® -coated transwells (M) and Growth Factor-Reduced Matrigel® -coated transwells (GFR). Results are compared with invasion on a commercially available invasion chamber (IC). Values are medians and interquartile ranges of the number of cells counted on the lower side of the membrane in five fields of view per membrane.
Cytotrophoblast cells cultured for 3 days on Matrigel® -coated transwells. Cultures were fixed and stained with anti-cytokeratin antibody and membranes were inverted on a microscope slide. Serial 1µm sections that were parallel to the membrane were examined using a confocal microscope (40x). A, the monolayer of cells on the upper surface of the membrane remain intact (top left frame: cell monolayer on upper surface of Matrigel® ; bottom right frame: lower surface of membrane). B, cells and Matrigel® on upper surface of the membrane have been removed (top left frame: upper surface of the membrane; bottom right frame: lower surface of the membrane).
Figure 6A
Figure 6B
