Correspondence: Dr M. Knight, Nuffield Department of Obstetrics and Gynaecology, John Radcliffe Hospital, Oxford OX3 9DU.
Objective To investigate whether syncytiotrophoblast microvilli (STBM) are shed into the maternal circulation in increased amounts in pre-eclamptic pregnancies as a possible cause of maternal vascular endothelial dysfunction.
Design A time-resolved fluoroimmunoassay was developed to measure STBM levels in peripheral and uterine venous plasma from normal pregnant and pre-eclamptic women. Three colour flow cytometry was used to assess the microparticulate nature of the STBM in pregnancy plasma. The effects of these plasmas on endothelial cell proliferation was compared and a correlation with the levels of STBM detected was sought.
Setting A laboratory investigation using clinical samples obtained from an obstetric practice in a teaching hospital.
Samples Peripheral venous plasma from 20 women with established pre-eclampsia, 20 normal pregnant women matched for age, gestation and parity, and 10 nonpregnant women of reproductive age. Paired uterine and peripheral venous plasma taken at caesarean section from 10 women with pre-eclampsia and 10 unmatched normal pregnant women.
Results STBM were detected in the plasma of pregnant women by both flow cytometry and time-resolved fluoroimmunoassay. Significantly higher levels of STBM were found in women with established pre-eclampsia (P= 0.01). STBM concentrations were higher in uterine venous plasma than in concurrently sampled peripheral venous plasma, confirming their placental origin. A significant correlation was found between the amount of STBM in the plasma and endothelial cell inhibitory activity.
Conclusions STBM are shed into the maternal circulation (microvillous deportation) and are present in significantly increased amounts in pre-eclamptic women. They may contribute to the endothelial dysfunction underlying the maternal syndrome of pre-eclampsia.
Pre-eclampsia is a multisystemic disorder of human pregnancy which can have severe consequences for both mother and infant. The cause remains unknown. The maternal syndrome is characterized by raised blood pressure, proteinuria, oedema, and other features which may include coagulation, renal or liver disturbance or focal cerebral ischaemia leading to eclampsia. It is now widely accepted that the maternal syndrome may be explained by a generalized vascular endothelial cell disturbance, caused by a circulating factor1. The nature of this factor is unknown, but pre-eclampsia is specific to pregnancy. Since pre-eclampsia can occur with a hydatidiform mole2 when placental tissue alone is present, the placenta is clearly implicated.
There is evidence that an endothelial disrupting activity is present on the syncytiotrophoblast microvillous membrane. We have previously shown that syncytiotrophoblast microvilli (STBM) prepared from normal placentas both inhibit the proliferation of endothelial cells and disrupt their growth as a monolayer in culture3. More recently we have found that STBM also inhibit endothelial cell-dependent relaxation of small arteries in vitro4 and stimulate the release of endothelin 1, a potent vasoconstrictor, from bovine endothelial cells (R. Corder, unpublished observation). We have also shown that a similar endothelial inhibitory activity is present in the plasma of pre-eclamptic women at significantly higher levels than gestationally matched normal controls5. This has lead us to speculate that the endothelial damage seen in pre-eclampsia results from the shedding of higher levels of STBM into the circulation compared with normal pregnancy.
STBM could enter the circulation either by way of whole trophoblast cells or as microvillous fragments. We have confirmed previous reports that syncytiotrophoblast is shed from the placenta in increased amounts in pre-eclampsia. Few of these cells appear to reach the peripheral circulation and are therefore unlikely to be responsible for the generalised vascular endothelial disturbance (M. Johansen, unpublished observation). Evidence that individual STBM may be shed is provided by studies of the syncytiotro-phoblast in pre-eclampsia which have shown the microvilli to be structurally abnormal and reduced in number7. To investigate this hypothesis, we developed an assay to detect and quantify STBM in the maternal circulation and sought to correlate the levels of STBM measured in vivo with plasma endothelial cell inhibitory activity in vitro.
STBM were prepared from term placentas delivered by elective caesarean section as described previously3. Red blood cell membranes were prepared from the blood of nonpregnant females by the method of Khalfoun et al.8 All membranes were resuspended in phosphate buffered saline (PBS) and quantified by the bicinchoninic protein assay (Pierce, Illinois, USA).
Samples were taken from women who gave their consent after an explanation of the purpose of the study. All studies were approved by the Central Oxford Regional Ethics Committee. Pre-eclampsia was defined as new hypertension of > 90 mmHg diastolic and at least 20 mmHg higher than the first recorded blood pressure, together with proteinuria of ≥ 500 mg/24 h or ++ on dipstick testing on two occasions at least six hours apart. Both hypertension and proteinuria regressed after delivery. Twenty millilitres of blood were taken from an antecubital vein for all peripheral samples, and if delivery was by caesarean section, up to 15 ml of blood was drawn from a uterine vein before incision of the uterus. All samples were taken into lithium heparin vacutainers (Becton-Dickinson, Oxford, UK).
Antecubital vein blood was taken antenatally from 20 normal pregnant women who had never had hypertension or proteinuria, matched for age (± 4 years) gestation (± 13 days) and parity (0, 1–3, 4 +) with 20 pre-eclamptic women, and from 10 nonpregnant women of reproductive age. In addition to sampling of peripheral blood at the time of diagnosis, uterine and peripheral venous blood was also taken at delivery from a further 10 women with pre-eclampsia undergoing caesarean section, together with a peripheral venous sample at 48 h postpartum. Uterine and peripheral blood samples was also taken from 10 women undergoing caesarean section for reasons other than pre-eclampsia (for example, breech position, previous caesarean section). Case matching for gestational age or parity was not possible in these cases. The characteristics of pre-eclamptic and control women are shown in Tables 1 and 2.
Table 1. Subjects in peripheral vein blood study. Data are shown as mean [standard error] unless otherwise indicated and were comparing using the Wilcoxon ranked pairs test unless indicated by * where Fishers exact test was used. NS = no significant difference.
Pre-eclampsia (n = 20)
Controls (n -= 20)
First diastolic pressure (mmHg)
Maximum diastolic pressure (mmHg)
Proteinuria (++ or >): n (%)
Gestation at sampling (days)
Gestation at delivery (days)
Birthweight of infant (g)
Table 2. Subjects in uterine vein blood study. Data are shown as mean [standard error] unless otherwise indicated and were compared using Wilcoxon ranked pairs test unless indicated by * where Fishers exact test was used. NS = no significant difference.
Pre-eclampsia (n = 10)
Controls (n = 10)
First diastolic pressure (mmHg)
Maximum diastolic pressure (mmHg)
Proteinuria (++ or >): n (%)
Gestation at sampling/delivery (days)
Birthweight of infant (g)
Ten millilitres of plasma was diluted 1 in 2 with phosphate buffered saline (Sigma Chemical Co, Poole, Dorset, UK) and centrifuged at 150,000 g for 45 min at 4 °C to pellet any STBM. The pellet was resuspended in 350 μl assay buffer (Wallac, Milton Keynes, UK) for assay of STBM, or in 50 μl PBS/0.1% bovine serum albumin (Sigma Chemical Co.) for flow cytometry. Samples were used fresh for the fluoroimmunoassay and flow cytometry, or frozen at -20 °C before measurement of endothelial cell inhibitory activity.
Hybridoma culture and antibody labelling
Hybridoma cells were cultured to exhaustion in RPMI medium (Gibco BRL, Paisley, Scotland) containing 100 μl/ml benzylpenicillin, 100 μg/ml streptomycin and 20% fetal calf serum (Gibco BRL). IgG was then isolated from the culture supernatant by affinity chromatography using protein-G conjugated beads (Pharmacia-Biotech, St Albans, Herts, UK). Purified antibody was labelled for the time-resolved fluoroimmunoassay with 14 europium atoms per molecule using a DELFIA europium-labelling kit (Wallac)9. For use in three colour flow cytometry, antibodies were biotinylated by incubating 1 mg antibody in 0.1 M borate buffer with 250 pg succinimidobiotin for 4 h at 18–20 °C. The reaction was stopped by addition of 1 M NH4 Cl and uncoupled biotin removed by dialysis against PBS.
Single colour flow cytometry
A panel of anti-trophoblast antibodies (Table 3) was tested for reactivity against STBM by flow cytometry; 30 pg STBM was incubated with 50 p1 of each anti-body diluted in PBS/0.1% bovine serum albumen at 4 °C. After a 30 min incubation, 2 μl FITC-conjugated goat anti-mouse immunoglobulin (Dako, High Wycombe, UK) was added to each tube and incubated for a further 30 min. Samples were then diluted with 100 μl PBS/0.1% bovine serum albumen before analysis on a Coulter Epics Elite flow cytometer. The discrimination cut off for the forward angle light scatter as set low to allow detection of particles < 1 μ in diameter. The fluorescence intensity (mean channel rightness) and percentage of positive particles was assessed for each antibody.
Table 3. Antibody screening by STBM flow cytometry.
Ninety-six well, low fluorescence plates (Nunc, Denmark) were coated overnight at 4 °C with anti-body diluted in tris-buffered saline. Plates were blocked for 3 hours with tris-buffered saline/10% horse serum (Gibco) and then washed five times with TBS/0.05% Tween 20 (Sigma). Samples were incubated overnight at 18–20 °C in triplicate, together with standards of 200 ng-50 μg/ml STBM. Plates were washed five times with TBS/Tween and then incubated for 2 hours at 18–20 °C with europium-labelled antibody diluted in assay buffer. After a further 10 washes with TBS/Tween, plates were developed with enhancement solution (Wallac) 100 μl/well. All incubations were carried out with agitation. Plates were read on a 1234 DELFIA research fluorometer (Wallac), and results calculated using MultiCalc 2.0 software.
The highest mean channel brightness in STBM flow cytometry was obtained from the antibodies NDOG1, NDOG210, H31711 (gift of Dr P. J. McLaughlin, University of Liverpool) and ED82212 (gift of Dr D. Sooranna, Chelsea and Westminster Hospital, London) (Table 3). These antibodies were tested in various combinations to establish a fluoroimmunoassay. The most sensitive and specific combination was H3 17 as the capture antibody with NDOG2 as the reporter (data not shown). Both H317 and NDOG2 bind placental alkaline phosphatase, therefore the specificity and sensitivity of the assay were assessed using STBM, red blood cell membranes and free placental alkaline phosphatase (Calbiochem Nottingham, UK) seeded into male blood and recovered by centrifugation. Six patient plasma pellets were measured both in the H317 FIA and in a fluoroimmunoassay using ED822 as the capture antibody in order to establish the presence of a trophoblast antigen other than placental alkaline phosphatase on the particles detected. The antigen to which ED822 binds is not known, but it does not bind placental alkaline phosphatase, as evidenced by the different tissue staining pattern, and failure to inhibit the reaction of ED822 with STBM by prior reaction with free placental alkaline phosphatase.
Three colour flow cytometry
Three colour flow cytometry was also used to analyse particles pelleted from plasma. FITC-conjugated anti-glycophorin A and anti-glycoprotein IIb were used simultaneously to detect microparticles of red cell and platelet origin respectively in order to exclude these from the analysis. The remaining particles were labelled with two different anti-syncytiotrophoblast antibodies, ED822 (allophycocyanin channel) and NDOG2 (phycoerythrin channel) in order to establish whether microparticles expressing both antigens on their surface could be detected.
Each resuspended plasma pellet sample was divided into two 25 μl aliquots. The first aliquot was labelled with nonspecific mouse immunoglobulins in order to detect background antibody binding. The second aliquot was labelled with the two different anti-trophoblast antibodies, together with a cocktail of anti-glycophorin A and anti-glycoprotein IIb. Two aliquots of 30 pg of STBM were similarly labelled. Each sample was incubated for 30 min with the primary antibody at 10 μg/ml in PBS/0.1% bovine serum albumen (purified pooled mouse IgG (Coulter) or ED822, followed by a 30 min incubation with allophycocyanin conjugated sheep anti-mouse Ig (anti mouse-APC, Molecular Probes, Leiden, The Netherlands), 2 pg per sample. Unbound anti mouse-APC was blocked by addition of 10 pg pooled purified mouse immunoglobulin (Sigma Chemical Co) for 15 min. Samples were further incubated with biotinylated mouse IgG or biotinylated NDOG2, 10 μg/ml for 30 min, followed by phycoerythrin-conjugated streptavidin (Sigma Chemical Co), 2 μl. Finally, particles were incubated with FITC conjugated mouse immunoglobulins (Dako) 2 μg, or anti-glycophorin A-FITC 1 μg and anti-glycoprotein IIb-FITC 1 μg. All incubations were carried out in the dark on ice. Single colour control samples of STBM and red blood cell membranes were included in each experiment.
Data were analysed using Coulter Elite and Reallist software. An amorphous gate was used to select the events of appropriate size and granularity characteristics to detect subcellular particles; a linear gate was used to exclude FITC positive events (ie, platelet and erythrocyte microvesicles); a quadratic gate was then used to determine particles positive for both anti-trophoblast antibodies (phycoerythrin and allophycocyanin channels). The size of the particles detected was calculated by comparison with the forward angle light scatter signal obtained from beads of standard 10 pm size (standard-brite beads, Coulter) and 3.6 pm size (flow-set beads, Coulter).
Plasma pellets from samples of STBM seeded into nonpregnant female blood, unseeded female blood, blood from pre-eclamptic women and blood from normal pregnant women were measured.
Endothelial cell proliferation assay
The endothelial cell inhibitory activity of control pregnant and pre-eclampsia plasmas was assessed by 3 H-thymidine incorporation as described previously5. Each plasma was tested in five replicates, and compared with 10 replicates of a standard plasma pool included on each plate. The median test counts were expressed as the percentage of the mean count of the plasma pool to give a proliferation index. The correlation between the proliferation index and the STBM levels detected in each sample was then assessed.
Paired data were compared using the Wilcoxon ranked pairs test. Unpaired data were compared using the Mann-Whitney U test. Unpaired categorical data were compared using Fishers exact test, and paired data with McNemar's test with continuity correction. Correlation between data sets was investigated using Spearman's rank correlation. Significance was assumed if P < 0.05.
The fluoroimmunoassay was linear over the range of STBM concentrations 50 ng/ml to 5 μg/ml [r= 0.99, P < 0.001 (linear regression by least squares method of log transformed data)], with interassay variabilities of 7% (low range), 6% (mid-range) and 5% (high range) (n = 10). Corresponding intra-assay variabilities were 7%, 11% and 8% (n = 8). No signal was detected in the assay from red blood cell membranes, free placental alkaline phosphatase, male plasma pellets (n = 10) or nonpregnant female plasma pellets (n = 10), confirming the specificity of the assay. The sensitivity of the assay including the centrifugation/concentration step was < 2 ng STBM protein/ml plasma.
STBM activity was detected in the plasma pellets of 17/20 pre-eclamptic women compared with 10/20 normal pregnant women (P < 0.05, McNemar's test with continuity correction), but none was found in the plasma of nonpregnant women (Fig. 1). There were significantly higher levels (P= 0.01) in the pre-eclampsia plasmas (median equivalent to 7.5 ng STBM protein/ml plasma (range 0–39.4)) compared with the normal pregnant controls (median 1.2 ng STBM/ml (range 0–18.3)).
No correlation was found between the STBM levels detected and routine measures of disease severity, such as plasma urate and aspartate transferase levels, platelet counts, 24 hour urinary protein and maximum diastolic blood pressure. In both groups of women a trend was seen towards increasing STBM levels with increasing gestational age at sampling. This was significant in the control group (r= 0.48, P < 0.05) but did not reach significance in the pre-eclamptic group (r= 0.26, P > 0.1).
The levels of STBM in patient samples measured with both H317 and ED822 assays correlated significantly (r= 0.86, P < 0.05), confirming that the particles detected have at least two different trophoblast antigens on their surface.
STBM levels were found to be significantly higher in uterine vein plasma compared with peripheral vein plasma in both pre-eclamptic and control pregnant women (28.7 ng/ml vs 10.0 ng/ml in pre-eclampsia plasmas, P < 0.01; 9.1 ng/ml vs 2.3 ng/ml in normal pregnant patients, P < 0.05) (Fig. 2). Additionally, levels in uterine vein plasma were significantly different between the two groups (P < 0.01, n= 10).
STBM levels were measured in peripheral plasma of 10 pre-eclamptic women at diagnosis (between 1 and 14 days before delivery). STBM levels had increased at the time of delivery by caesarean section, were highest in uterine vein plasma, and in all but one woman had disappeared by 48 h postpartum (Fig. 3).
Flow cytometry was used to confirm the particulate nature of the STBM detected in the peripheral plasma of pre-eclamptic women. Using this method, it was possible to resolve artificially prepared STBM as a distinct population of particles of calculated size range 0.2–1.2 pm in diameter, mean 0.4 μm, positively labelled with both anti-trophoblast antibodies (Fig. 4a). STBM seeded into male blood and recovered by ultracentrifugation gave a similar picture (Fig. 4b), as did the pelleted material from the plasma of a women with HELLP syndrome (Fig. 4c).
The trophoblast microparticles in this women were of mean size 0.6 pm, strongly suggesting their microvillous origin. These particles did not appear when the same plasma pellet was labelled with nonspecific mouse immunoglobulins (Fig. 4d). Trophoblast microparticles were not detected in the pelleted material from a normal pregnant woman (Fig. 4e) or a nonpregnant subject (Figure 4f). Trophoblast microparticles were detected in 4/6 pre-eclamptic plasmas, 2/6 normal pregnant plasmas and 0/6 nonpregnant plasmas.
Pre-eclampsia plasmas were confirmed to be more inhibitory than plasmas from normal pregnant women (Fig. 5a). A significant negative correlation was found between the endothelial proliferation index of each plasma and the STBM levels measured by FIA (Fig. 5 b, r= -0.38, P= 0.02), indicating that STBM present in the plasma may be responsible for some of the inhibitory activity.
We have developed a sensitive time-resolved fluoroimmunoassay to enable us to detect microparticulate trophoblast material in the plasma of pregnant women. The narrow emmission spectrum of the europium label used for detection, combined with time-resolved fluorescence measurement enhanced the sensitivity of the assay. The use of a two-site assay combined with analysis of plasma ultracentrifugates enabled us to detect microparticulate trophoblast without interference free trophoblast antigens. Our results demonstrated the presence of circulating syncytiotrophoblast microparticles in the peripheral blood of both preeclamptic and normal pregnant women. These particles are present in significantly higher amounts in the blood of pre-eclamptic women, suggesting that they may play a role in the generalised endothelial disturbance characteristic of the syndrome. Higher levels of STBM in uterine vein plasmas compared with matched peripheral plasmas, is consistent with their placental origin.
The strongest evidence that the circulating trophoblast material is particulate is provided by the flow cytometric data. Analysis of pelleted material from pregnancy plasmas has revealed the presence of a group of particles which have four different characteristics in common with STBM prepared artificially. These particles are of the same size and granularity as STBM, and also possess on their surface two syncytiotrophoblast antigens detected by ED822 and NDOG2 antibodies. Additionally, these particles do not express either platelet (glycoprotein IIb) or erythrocyte (glycophorin A)antigens, and are therefore unlikely to originate from red blood cells or platelets. This property enabled us to use three-colour flow cytometry to exclude from the analysis platelet microvesicles, known to be present in pre-eclampsia plasma17, and fragments of red cell membranes which may be present in women with haemolysis as part of the HELLP syndrome. The use in combination of both flow cytometric and FIA techniques establishes the presence of STBM in the circulation of both pre-eclamptic and normal pregnant women, and that STBM are shed in higher amounts in pre-eclamptic pregnancy compared with normal pregnancy.
There was a significant but weak correlation of STBM levels with in vivo plasma endothelial inhibitory activity, but no apparent correlation with clinical indices of disease severity. The former observation suggests that STBM are unlikely to be the sole cause of the endothelial cell disturbance of pre-eclampsia, although investigation of a relation between STBM levels and in vivo markers of endothelial damage in plasma; for example, von Willebrand's factor or endothelin-1, may help clarify the role of STBM in inducing this endothelial dysfunction.
There is some indirect evidence from earlier studies to suggest that STBM may be deported in pregnancy. Actin is a component of the STBM cytoskeleton18. Circulating actin complexes have been found in pregnancy and postulated as being derived from deported trophoblast19. These complexes were not measured, but were more easily visualised in western blots of pre-eclampsia sera than sera from normal pregnant women. Tissue polypeptide antigen, a complex of degraded cytokeratins, has also been found to be increased in the plasma of women with pre-eclampsia, and further raised in women with HELLP syndrome20. The syncytiotrophoblast has been shown to be a rich source of tissue polypeptide antigen21, and its presence within the circulation may indicate STBM breakdown. That STBM could not be detected in all women with pre-eclampsia may indicate that they are released episodically, perhaps associated with acute placental events such as ischaemia or infarction. The endothelial disrupting effect of STBM takes several hours to develop (A. Kr. Smarason, unpublished observation), hence it is not surprising that a simple relation between STBM levels and clinical parameters was not observed. STBM levels measured in plasma represent a single snapshot of an acute event, whereas changes in clinical parameters will reflect more chronic events, and thus there is unlikely to be a direct relation between the variables.
Once released into the circulation, STBM come into contact not only with endothelium but also maternal leucocytes and erythrocytes. These cells may bind and transport STBM, and their associated endothelial inhibitory factor around the circulation, or phagocytes may process the STBM to release components including the endothelial inhibitory factor in a free form. Leucocyte interaction with STBM may play an additional role in amplifying any effect of the STBM endothelial disrupting factor by release of harmful cytokines into the circulation, which activate endothelium. Several studies have shown that maternal leucocytes are activated in pre-eclampsia22–24, (G. P. Sachs, unpublished observation) and that there are raised plasma levels of TNF-α25,26, a potent stimulator of endothelial cells.
Our study shows for the first time that microvillous deportation occurs in vivo. It is less certain that STBM are directly responsible for the endothelial cell inhibitory activity seen in pre-eclampsia plasmas. Further work is required to elucidate the role of possible leucocyte-STBM interactions, which could lead to cytokine release and amplification of endothelial damaging activity, or cause processing of STBM and release of the inhibitory factor in a free form. Investigation of the element of STBM which disrupts endothelial function will further clarify the relevance of microvillous deportation to the pathogenesis of pre-eclampsia.
This work was supported by funding from the Medical Research Council (Grant No. G9409257), the Wellcome Trust (Reference No. 037358/Z/92) and the Oxford University Medical Research Fund. We would like to thank Dr C. Potter for advice on the time-resolved fluoroimmunoassay and Wallac Ltd for use of the DELFIA 1234 research fluorometer; Dr J. Morris for help with endothelial cell culture; Mrs E. Coghill for assistance with flow cytometry; and the staff and patients of the High Risk Pregnancy Unit of the John Radcliffe Hospital for patient recruitment and blood samples.