- Email: [email protected]
Circulating concentrations of soluble endoglin (CD105) in fetal and maternal serum and in amniotic fluid in preeclampsia
Research
www. AJOG.org
BASIC SCIENCE: OBSTETRICS
Circulating concentrations of soluble endoglin (CD105) in fetal and maternal serum and in amniotic fluid in preeclampsia Anne Cathrine Staff, MD, PhD; Kristin Braekke, MD; Guro M. Johnsen, MSc; S. Ananth Karumanchi, MD; Nina Kittelsen Harsem, MD OBJECTIVE: We explored whether concentrations of soluble endoglin in fetal serum and amniotic fluid and in maternal serum were elevated in preeclampsia. STUDY DESIGN: Umbilical vein serum, amniotic fluid, and maternal serum from 42 preeclamptic and 43 uncomplicated pregnancies that were delivered by cesarean section were analyzed by enzyme-linked immunosorbent assay for soluble endoglin. RESULTS: Median maternal serum and amniotic fluid soluble endoglin
concentrations were elevated in preeclampsia, compared with control pregnancies (66.9 ng/mL vs 15.1 ng/mL; P ⬍ .001, and 1.9 ng/mL vs 0.6 ng/mL; P ⬍ .001). Low concentrations of soluble endoglin were found in fetal circulation, which did not differ between preeclampsia
and control pregnancies (5.0 ng/mL vs 4.7 ng/mL; P ⫽ .2). Maternal serum soluble endoglin levels correlated with circulating soluble fmslike tyrosine kinase 1 concentrations. CONCLUSION: We confirmed elevated soluble endoglin in maternal circulation in preeclampsia, which correlated with soluble fms-like tyrosine kinase 1 concentrations and soluble fms-like tyrosine kinase 1/placental growth factor ratio. The fetus appears not to contribute to elevated circulating maternal soluble endoglin concentrations in preeclampsia.
Key words: endoglin, soluble fms-like tyrosine kinase 1 (sFlt1), preeclampsia, oxidative stress, placental growth factor (PlGF)
Cite this article as: Staff AC, Braekke K, Johnsen GM, Karumanchi A, Harsem NK. Circulating concentrations of soluble endoglin (CD105) in fetal and maternal serum and in amniotic fluid in preeclampsia. Am J Obstet Gynecol 2007;197:176.e1-176.e6.
P
reeclampsia is a pregnancy complication that affects 3%-4% of all pregnancies and represents a major threat to maternal and fetal health.1 The exact pathophysiologic condition is unknown, but generalized endothelial dysfunction with systemic inflammatory response is thought to be the final common
From the Departments of Obstetrics and Gynecology (Drs Staff, and Harsem, and Ms Johnsen) and Pediatrics (Dr Braekke), Ullevaal University Hospital and Faculty of Medicine, University of Oslo, Oslo, Norway, and the Departments of Medicine and Obstetrics and Gynecology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA (Dr Karumanchi). Received Dec. 4, 2006; accepted Mar. 12, 2007. Reprints not available from the authors. Supported in part by research grants from the Eastern Norway Regional Health Authority (G.M.J.). 0002-9378/$32.00 © 2007 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2007.03.036
176.e1
pathway that leads to the maternal signs of preeclampsia with de novo hypertension and proteinuria in the second half of pregnancy.2 A shallow placentation, with abnormal invasion of cytotrophoblasts and incomplete remodeling of placenta-supplying maternal uterine spiral arteries, is central to the pathogenesis of preeclampsia.2 This results in a relative hypoxic uteroplacental circulation and augmented placental oxidative stress2 and thereby placental release of endothelial deranging factors to the maternal circulation. Pathogenic roles of soluble fms-like tyrosine kinase 1 receptor (sFlt1)3,4 and soluble endoglin (sEng)5,6 have been suggested in the development of preeclampsia. Both proteins are produced in excess from preeclamptic placentas and are found at elevated concentrations in maternal circulation in preeclampsia. Elevated serum concentrations of both proteins in the second trimester may predict the onset of preeclampsia.4,6,7 In vitro studies suggest that excess circulating sFlt1 and sEng in preeclampsia inhibit vascular endothelial growth factor
American Journal of Obstetrics & Gynecology AUGUST 2007
(VEGF)– and transforming growth factor beta (TGF-)–stimulation of endothelial-dependent nitrogen oxide activation and thereby oppose physiologic nitrogen oxide– dependent vasodilatation and vasomotor effects.5 Excess sFlt1 in maternal circulation binds the angiogenic factors VEGF and placental growth factor (PlGF) and thereby reduces their effects on maternal endothelium, with the ensuing hypertension and proteinuria.3 The effects of augmented sFlt1 concentrations and reduced VEGF action on the maternal vascular endothelium possibly could be due to impairment of VEGF-dependent activation of endothelial nitrogen oxide synthase.3,5 Endoglin, also named CD105, is a 180 kd protein that is localized on cell surfaces and is expressed highly on endothelial cells and on placental syncytiotrophoblasts.8 Endoglin functions as a coreceptor for the TGF- family. The TGF- receptors bind TGF-1 and -3, and endoglin modulates the action of these ligands, acting on intracellular signaling mediators and thereby altering
Basic Science: Obstetrics
www.AJOG.org transcriptional responses.9 Venkatesha et al5 recently demonstrated a novel, smaller (65 kd), placenta-derived soluble form of Eng (sEng) in the sera of pregnant women. Paralleling the sFlt1 findings, sEng is elevated in preeclamptic maternal serum and is reduced after delivery. The authors showed that sEng cooperated with sFlt1 in inducing endothelial dysfunction in vitro and severe preeclampsia-like illness in vivo in pregnant rats. SEng and sFlt1 blocked proangiogenic effects of TGF- and VEGF in vitro, respectively, and showed additive effects, which indicates that these soluble receptors may act in concert to disrupt endothelial integrity.5 The decline of sEng concentrations after delivery supports a placental origin of sEng in preeclampsia5 but cannot exclude the maternal vascular endothelium, circulating leukocytes, or the fetus as potential sources for elevated maternal serum concentration of sEng in preeclampsia. We previously have shown elevated sFlt1 concentration also in fetal circulation in preeclampsia, but at a much lower concentration than in the maternal circulation and therefore not likely to contribute to the maternal sFlt1 concentration in preeclampsia.10 The aim of the present study was to analyze concentrations of sEng in the fetal circulation and in matched maternal serum and amniotic fluid samples in preeclamptic, compared with uncomplicated, pregnancies. In addition, we wanted to explore a possible correlation between maternal sEng concentrations and previously measured sFlt1 and PlGF10 concentrations and 8-isoprostane,11 a well-established marker of oxidative stress.12
M ATERIALS AND M ETHODS Patient selection The study is part of an ongoing biobank collection of patient samples from complicated and uncomplicated pregnancies at Ullevaal University Hospital (Oslo, Norway). Women with singleton pregnancies, all delivered by cesarean section (n ⫽ 85), were included and comprised women with uncomplicated pregnancies (n ⫽ 42, control subjects) and pre-
eclamptic patients (n ⫽ 43). No women with chronic hypertension, renal disease, or diabetes mellitus were included. All patients were fasting; none of the patients were in active labor or had ruptured membranes or clinical signs of infection. Preeclampsia was defined as blood pressure augmentation after 20 weeks of gestation to ⬎140/90 mm Hg on ⱖ2 occasions 6 hours apart in a previously normotensive woman, combined with proteinuria. Proteinuria was defined as protein dip stick ⱖ1⫹ on ⱖ2 midstream urine samples 6 hours apart or a 24-hour urine excretion of ⱖ0.3 g protein in the absence of urinary infection. Severe preeclampsia was defined by the American College of Obstetricians and Gynecologists criteria,13 which includes women with blood pressure of ⱖ160 mm Hg systolic. The newborn infant birthweight percentiles were calculated according to national birth registry data.10 The study protocol was approved by the Regional Committee of Medical Ethics in Eastern Norway; informed written consent was obtained from each patient. Maternal blood samples were obtained before cesarean section in the 85 women (42 control subjects, 43 women with preeclampsia); amniotic fluid from 46 patients (16 control subjects, 30 women with preeclampsia) was sampled at cesarean section. After delivery of the placenta, 68 umbilical vein serum samples (delivering blood from the placenta to the fetus) were collected (36 control subjects, 32 women with preeclampsia). The processing of biological samples was described previously.10 Because sufficient quantities of serum from the umbilical arteries were not available, we analyzed fetal sEng concentrations in paired arterial and venous EDTAplasma samples from the umbilical cord in 13 random patients (7 controls, 6 women with preeclampsia).10
sEng determinations Enzyme-linked immunosorbent assay for human sEng was performed in duplicates according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN).
Research
Previous analyses In previous publications from our biobank of preeclamptic and uncomplicated cesarean deliveries, we have reported maternal concentrations of the oxidative stress marker 8-isoprostane (8iso-prostaglandin F2␣),11 the angiogenesis-associated proteins sFlt1 and PlGF,10 and calprotectin, a marker of inflammation.14 Because of the ongoing inclusion of patients to this biobank, not all patients in the present study were included in the previous studies.
Statistical analysis The results are presented as median values (and 95% CI of the median for laboratory results). Statistical analyses were performed with the Statistical Package for the Social Sciences software (version 14.0; SPSS Inc, Chicago, IL). Differences in continuous variables between groups were tested by nonparametric MannWhitney tests. Spearman’s correlation was used to calculate correlation coefficients. A probability level of ⬍.05 was considered statistically significant.
R ESULTS Clinical characteristics of the 85 pregnant women who were included in the sEng maternal serum analyses are shown in the Table. Fewer patients were included in the fetal serum (n ⫽ 68 patients) and amniotic fluid (n ⫽ 46 patients) analyses because of difficulties in obtaining these samples; however, this subset of patient group characteristics did not differ significantly from the results in the Table. Only 1 woman in the preeclampsia group and none of the women in the uncomplicated pregnancy group delivered a baby small for gestational age (below the 10th weight percentile). Of 43 preeclamptic patients, 34 patients had severe preeclampsia, whereas 9 patients had mild preeclampsia. All control subjects had negative urine dip stick protein readings, whereas 30 of the 43 preeclamptic patients had 3⫹ at delivery, 6 patients had 2⫹, and 7 patients had 1⫹ readings. Three of the patients with preeclampsia also had evidence of HELLP (hemolysis, elevated liver enzymes, and low platelets) syn-
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e2
Research
Basic Science: Obstetrics
www.AJOG.org
TABLE
Clinical characteristics of patient groups included in the sEng maternal serum analysis Control pregnancies (n ⴝ 42)
Characteristic
Preeclampsia (n ⴝ 43)
P value
Patient age at delivery (y)
31 (21-40)
31 (18-42)
.5
Body mass index before pregnancy (kg/m )
22.5 (17.4-29.4)
23.7 (18.9-41.1)
.03*
Body mass index at delivery (kg/m )
28.0 (20.3-37.5)
31.0 (22.0-45.9)
.02*
................................................................................................................................................................................................................................................................................................................................................................................ 2 ................................................................................................................................................................................................................................................................................................................................................................................ 2 ................................................................................................................................................................................................................................................................................................................................................................................
Parity
0 (0-1)
0 (0-3)
.05
................................................................................................................................................................................................................................................................................................................................................................................
Gestational age at delivery (weeks)
38.7 (34.3-41.7)
32.9 (24.9-38.7)
⬍.001*
................................................................................................................................................................................................................................................................................................................................................................................
Neonatal weight (g)
3456 (2575-4165)
1692 (540-3856)
⬍.001*
................................................................................................................................................................................................................................................................................................................................................................................
Neonatal weight percentile
53 (10-95)
37 (7-90)
.001*
................................................................................................................................................................................................................................................................................................................................................................................
Blood pressure (mm Hg)
.......................................................................................................................................................................................................................................................................................................................................................................
Systolic at ⬍20 wk
110 (80-135)
117 (90-135)
.02*
Diastolic at ⬍20 wk
65 (50-80)
70 (50-89)
.02*
Systolic at delivery
120 (90-140)
160 (145-220)
⬍.001*
Diastolic at delivery
70 (60-92)
105 (90-119)
⬍.001*
....................................................................................................................................................................................................................................................................................................................................................................... ....................................................................................................................................................................................................................................................................................................................................................................... ....................................................................................................................................................................................................................................................................................................................................................................... ................................................................................................................................................................................................................................................................................................................................................................................
Values shown are median and range (minimum and maximum values). * P ⬍.05, Mann-Whitney test.
drome.13 In the preeclampsia group, 38 of the 43 women were delivered prematurely before week 37 of gestation, and 28 women were delivered before week 34 of gestation. All women in the uncomplicated pregnancy group were delivered after week 37 of gestation, except for 1 woman at week 34 of gestation. The individual results from the sEng analyses are shown in Figure 1. The preeclampsia group had a 4.4-fold elevated median maternal serum sEng concentration (66.9 ng/mL; 95% CI, 51.0-79.7) compared with the control group (15.1 ng/mL; 95% CI, 12.9-17.8; P ⬍ .001). The umbilical vein serum samples had very low concentrations of sEng compared with the maternal serum concentrations, and the preeclampsia group did not differ significantly from the control group (preeclampsia, 5.0 ng/mL [95% CI, 4.6-5.2] vs control subjects, 4.7 ng/mL [95% CI, 4.3-5.0], P ⫽ .2). The amniotic fluids had even lower concentrations of sEng, but median sEng concentration of the preeclampsia group was statistically significantly 3.2-fold elevated as compared with the control group (preeclampsia, 1.9 ng/mL [95% CI, 1.1-4.3] vs control subjects, 0.6 ng/mL [95% CI, 0.4-0.8]; P ⬍ .001). There was a significant positive correlation between maternal serum and amniotic fluid sEng concentrations in pre176.e3
eclampsia (Spearman’s correlation, 0.5; P ⫽ .006). Fetal sEng concentration was 15% lower in plasma compared with corresponding serum sample (umbilical vein: median, 3.9 vs 4.6, respectively; P ⫽ .02). Median sEng concentration for the random 13 paired umbilical cord plasma samples did not differ significantly between artery and vein blood (3.8 ng/mL vs 3.9 ng/mL; P ⫽ .6). We previously analyzed maternal serum concentration of sFlt1 and free PlGF in 68 of the 85 women who were included in this study.10 We found a significant positive correlation between maternal sEng and maternal sFlt1 concentrations, both for the whole patient group and for the preeclampsia group (whole patient group: Spearman’s correlation coefficient, 0.8 [P⬍.001] and preeclampsia group separately: Spearman’s correlation coefficient, 0.5 [P⫽.038]; Figure 2). Maternal sEng concentrations were correlated inversely with PlGF concentrations for the preeclampsia group (Spearman’s correlation, 0.5; P ⫽ .01). In the present study, we calculated the ratio between previously measured maternal serum sFlt110 and maternal serum-free PlGF10 concentrations and found that the median sFlt1/PlGF ratio was significantly elevated in the preeclampsia group, com-
American Journal of Obstetrics & Gynecology AUGUST 2007
pared with the control group (126.4 [95% CI, 89.5-161.2] vs 18.5 [95% CI, 12.7-29.6]; P ⬍ .001). Also, there was a positive correlation between the maternal sEng concentrations and sFlt1/PlGF ratios, both for preeclampsia and the control group (both: Spearman’s correlation coefficient, 0.5; P ⬍ .005). There was no significant correlation between body mass index at delivery and sEng concentrations for either patient groups (both: P ⬎ .07). We found no difference in median maternal sEng concentrations for either the control or preeclampsia group when we compared women with a body mass index at delivery ⬍30 kg/m2 with a BMI at ⱖ30 kg/m2 (both: P ⬎ .2). For the preeclampsia group, there was a significant negative correlation between the maternal sEng concentrations and gestational length (Spearman’s correlation, – 0.4; P ⫽ .005). Also, for the 28 women in the preeclampsia group who delivered prematurely before week 34 of gestation, median serum sEng was elevated significantly, compared with those delivered from week 34 and onwards (n ⫽ 15; 76.4 vs 39.5 ng/mL; P ⫽ .006). There was no statistically significant difference in median maternal sEng concentration between the severe (n ⫽ 34 cases) and mild (n ⫽ 9 cases) preeclamp-
sEng concentration Maternal serum sEng, ng/mL
P <.0001 250 200 150 100 50 0 Controls
Fetal serum sEng, ng/mL
B
P =.2
10.0 7.5
FIGURE 2
Maternal serum sEng concentrations 30000
25000
20000
15000
10000
5000
0 0
50
100
150
200
250
Maternal serum sEng, ng/mL
5.0
C OMMENT 2.5 0.0 Controls
C Amniotic fluid sEng, ng/mL
Preeclampsia
is associated not only with augmented oxidative stress but also with a generalized systemic inflammatory response,2 and we showed in a previous publication that maternal circulating concentration of calprotectin, an inflammatory protein, is elevated in preeclampsia.14 We previously presented maternal calprotectin concentration in 61 of the 85 patients who were included in this study. We found a positive correlation for the preeclampsia group between calprotectin and sEng concentrations in the maternal circulation in the present study (Spearman’s correlation coefficient, 0.6; P ⫽ .001), but not for the control group (P ⫽ .2).
Maternal serum sFlt1, pg/mL
FIGURE 1
A
Research
Basic Science: Obstetrics
www.AJOG.org
Preeclampsia
P <.0001 10.0 7.5 5.0 2.5 0.0 Controls
Preeclampsia
sEng concentration (nanograms/milliliter) in A, maternal serum, in fetal serum from the umbilical vein (B, delivering blood from the placenta to the fetus), and in C, amniotic fluid. Horizontal bars represent median values.
sia group of preeclamptic women (71.9 vs 51.1 ng/mL; P ⫽ .3). We also previously presented maternal plasma concentration of the oxidative stress marker 8-isoprostane (8-isoprostaglandin F2␣) in 54 of the 85 women who were included in this study.10 For the preeclampsia group, there was a significant positive correlation between maternal sEng concentrations and 8-isoprostane plasma concentrations (Spearman’s correlation coefficient, 0.5; P ⫽ .038). For the control group, there was no such positive significant correlation (P ⫽ .1). Preeclampsia
Our study is, to our best knowledge, the first report on concomitant concentrations of maternal, fetal, and amniotic fluid concentrations of sEng in pregnancy. We found very low levels of sEng in umbilical vein serum without significant difference between cases and control subjects, with the inference that fetal blood is unlikely to contribute to the elevated sEng levels that are found in maternal blood in preeclampsia. Our results also support a strong correlation between elevated circulating maternal sEng concentrations and sFlt1 concentrations in the maternal circulation, as demonstrated previously by Venkatesha et al.5 In accordance with their study, we found in our Norwegian study population of preeclamptic women that elevated sEng in maternal serum was most often associated with elevated sFlt1 concentrations (Figure 2) and elevated sFlt1/PlGF ratios, which suggests a molecular link between these circulating proteins in preeclampsia. Interestingly, we found a higher concentration of sEng in the amniotic fluid in preeclampsia compared with control subjects, albeit at very low concentrations, which is in contrast to our and other authors’ previous findings of very high concentrations of sFlt1 in amniotic fluid,10 which suggests different transplacental transport for sFlt1 and sEng. Maternal sEng falls after delivery,5 and both fetus and placenta theoretically could contribute, directly or indirectly, to the elevated maternal sEng concentra-
Maternal serum sEng concentrations (nanograms/milliliter) are correlated positively with maternal serum sFlt1 concentrations (picograms/milliliter). The lines represent regression line and 95% CI of the regression line; the closed triangle represents preeclampsia; the open circle represents control pregnancies.
tions that are found before delivery in preeclampsia. We demonstrated very low concentrations of sEng in the fetal circulation and no difference between the preeclampsia and control groups. We therefore can exclude a substantial fetal contribution to the elevated maternal sEng concentration in established preeclampsia. Our findings, in light of previous publications that demonstrated endoglin production in syncytiotrophoblasts5,8 and endothelial cells5,8 and sEng production in term placentas,5 are supportive of a direct or indirect placental origin of the elevated maternal serum sEng concentration in preeclampsia. The direct placental contribution would be shedding of sEng from the placenta. Indirect placental contribution could include freeing of placenta compounds before delivery to the maternal circulation; compounds that could activate the maternal endothelium and thereby endothelial cell freeing of sEng. Several substances are produced in excess in preeclamptic placentas; some substances could be speculated to induce maternal endothelial cell sEng shedding, such as trophoblast debris2 and oxidized lipids,
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e4
Research
Basic Science: Obstetrics
including the bioactive 8-isoprostane,15,16 the latter is capable of inducing endothelial activation in vitro.12 Presently, it is unknown how endoglin is shed to the circulation in pregnancy and why it is found in excess concentration in preeclampsia. Betaglycan is another coreceptor of TGF-, which shares 71% sequence similarity on the transmembrane and intracellular domain with endoglin.17 In vitro, betaglycan is shed to a free form by matrix metalloproteinases18; it has been proposed that cell-bound endoglin could be shed after cleavage in a similar manner in preeclampsia.5 Hypoxia is known to upregulate endoglin expression.19,20 We speculate that augmented placental oxidative stress in preeclampsia could induce augmented sEng production and possibly even contribute to cleavage of syncytiotrophoblast cell-bound endoglin, releasing a smaller form to the maternal circulation. Alternatively, activated leukocytes in the maternal circulation could also contribute to the elevated sEng concentrations in preeclampsia. In this regard, peripheral blood mononuclear cells have been demonstrated recently to contribute to the circulating sFlt1 in preeclampsia.21 It is also interesting to note that we found a positive correlation between 8-isoprostane and sEng in the maternal circulation in preeclampsia that is similar to the positive correlation demonstrated in preeclampsia between maternal 8-isoprostane and sFlt concentrations in our previous study.10 Such an association between maternal sEng and 8-isoprostane concentrations in preeclampsia, which link elevated systemic maternal oxidative stress with an elevated systemic concentration of sEng, has not been demonstrated previously. We suggest that a hypoxic placenta in preeclampsia could directly (by producing more 8-isoprostane and sEng) and/or indirectly (by producing/releasing factors from the uteroplacental circulation that could induce the elevation of 8-isoprostane and sEng that is seen in the maternal circulation) contribute to both the observed elevated concentrations of the oxidized lipid 8-isoprostane and the sEng and sFlt1 proteins in the maternal circula176.e5
tion. All 3 substances are theoretically capable of inducing endothelial dysfunction,3,5,12 which is a typical feature of preeclampsia. The pathophysiologic role of 8-isoprostane is little explored in preeclampsia but has been shown in vitro to cause endothelial dysfunction and platelet activation and vasoconstriction,12 which are also main characteristics of preeclampsia. A weakness of our study is that the patient groups were not matched for gestational age. Our study involved analyses of circulating factors in samples from matched fetal, maternal, and amniotic fluid compartments; the patients who were included were all delivered by cesarean section, and uncomplicated “control” pregnancies that are delivered prematurely are not found easily. We have demonstrated that, in our preeclampsia group, there was a significant negative correlation between maternal sEng concentrations and gestational length. However, this probably reflects that women with more severe preeclampsia, who need premature delivery, have the highest sEng concentrations, because there is no evidence for sEng decreasing with augmenting gestational age. Levine et al6 recently demonstrated an increase in maternal sEng concentration both in uncomplicated and preeclamptic pregnancies during the third trimester but was less pronounced and at a later gestational age in uncomplicated pregnancies. Our study therefore probably underestimates the difference in median maternal sEng serum concentration between the patient groups. Also, similar to the findings of Levine et al,6 we did not find any statistically significant difference in maternal serum sEng in established preeclampsia of the third trimester between women with severe and mild preeclampsia. This is in contrast to a previous study of Venkatesha et al,5 who found more elevated sEng in established severe preeclampsia, but included fewer patients with severe preeclampsia than the present study and with significant differences in gestational ages between the mild and severe cases. We chose serum samples in our study instead of plasma samples because the results could be more easily comparable
American Journal of Obstetrics & Gynecology AUGUST 2007
www.AJOG.org with the 2 reports from other pregnancy cohorts.5,6 The 13 matched plasma samples from the fetal circulation were used to illustrate that there was not a great difference in sEng concentration between the blood deriving from the placental side (umbilical vein) and blood deriving from the fetal side (umbilical arteries), which supports our conclusion that the fetus does not contribute to the elevated maternal sEng concentration in established preeclampsia. The findings reported in this article shed important clues on the pathogenesis of preeclampsia. The syndrome of hypertension and proteinuria is observed in preeclamptic mothers, but not in their infants. The fetal growth restriction that accompanies severe preeclampsia is thought to be the result of the placental ischemia. These clinical observations suggest that the circulating factors that mediate the maternal syndrome, although of placental origin, should be restricted primarily to the maternal vascular compartment and not the fetal circulation. Both sFlt1 and sEng have been hypothesized to be central mediators of the maternal syndrome on the basis of clinical and rodent studies.3,5 The findings that the fetuses of preeclamptic mothers do not have high circulating concentrations of either sEng or sFlt1 is consistent with the idea that fetuses do not experience proteinuria or hypertension like the mothers because they are not exposed to high concentrations of antiangiogenic factors. There is a need for further understanding of preeclampsia, which affects mortality and morbidity rates of fertile women and their offspring both on a short-term and long-term basis, with augmented cardiovascular risk later in life for both the women with preeclampsia and their offspring.22,23 Our data do not indicate that sEng in fetal circulation is a relevant risk marker for cardiovascular disease in offspring of mothers with preeclampsia, nor do our data support a fetal contribution to the maternal elevated concentration of sEng in preeclampsia. A placental origin is likely, but additional contribution from activated maternal vasculature or leukocytes cannot be ruled out. f
www.AJOG.org REFERENCES 1. Duley L. Pre-eclampsia and the hypertensive disorders of pregnancy. Br Med Bull 2003;67:161-76. 2. Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science 2005; 308:1592-4. 3. Maynard SE, Min JY, Merchan J, et al. Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J Clin Invest 2003;111:649-58. 4. Levine RJ, Maynard SE, Qian C, et al. Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 2004;350:672-83. 5. Venkatesha S, Toporsian M, Lam C, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006;12:642-9. 6. Levine RJ, Lam C, Qian C, et al. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006;355:992-1005. 7. Vatten LJ, Eskild A, Nilsen TI, Jeansson S, Jenum PA, Staff AC. Changes in circulating level of angiogenic factors from the first to second trimester as predictors of preeclampsia. Am J Obstet Gynecol 2007;196:239.e1-6. 8. Gougos A, St Jacques S, Greaves A, et al. Identification of distinct epitopes of endoglin, an RGD-containing glycoprotein of endothelial cells, leukemic cells, and syncytiotrophoblasts. Int Immunol 1992;4:83-92. 9. Barbara NP, Wrana JL, Letarte M. Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members
Basic Science: Obstetrics of the transforming growth factor-beta superfamily. J Biol Chem 1999;274:584-94. 10. Staff AC, Braekke K, Harsem NK, Lyberg T, Holthe MR. Circulating concentrations of sFlt1 (soluble fms-like tyrosine kinase 1) in fetal and maternal serum during pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2005;122:33-9. 11. Harsem NK, Braekke K, Staff AC. Augmented oxidative stress as well as antioxidant capacity in maternal circulation in preeclampsia. Eur J Obstet Gynecol Reprod Biol 2006;128:209-15. 12. Patrono C, FitzGerald GA. Isoprostanes: potential markers of oxidant stress in atherothrombotic disease. Arterioscler Thromb Vasc Biol 1997;17:2309-15. 13. American College of Obstetricians and Gynecologists. Diagnosis and management of preeclampsia and eclampsia: ACOG practice bulletin no.: 33. Obstet Gynecol 2002; 99:159-67. 14. Braekke K, Holthe MR, Harsem NK, Fagerhol MK, Staff AC. Calprotectin, a marker of inflammation, is elevated in the maternal but not in the fetal circulation in preeclampsia. Am J Obstet Gynecol 2005;193:227-33. 15. Staff AC, Halvorsen B, Ranheim T, Henriksen T. Elevated level of free 8-iso-prostaglandin F2alpha in the decidua basalis of women with preeclampsia. Am J Obstet Gynecol 1999;181:1211-5. 16. Walsh SW, Vaughan JE, Wang Y, Roberts LJ. Placental isoprostane is significantly increased in preeclampsia. FASEB J 2000; 14:1289-96.
Research
17. Cheifetz S, Bellon T, Cales C, et al. Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J Biol Chem 1992;267:19027-30. 18. Velasco-Loyden G, Arribas J, Lopez-Casillas F. The shedding of betaglycan is regulated by pervanadate and mediated by membrane type matrix metalloprotease-1. J Biol Chem 2004;279:7721-33. 19. Docherty NG, Lopez-Novoa JM, Arevalo M, et al. Endoglin regulates renal ischaemia-reperfusion injury. Nephrol Dial Transplant 2006;21:2106-19. 20. Sanchez-Elsner T, Botella LM, Velasco B, Langa C, Bernabeu C. Endoglin expression is regulated by transcriptional cooperation between the hypoxia and transforming growth factor-beta pathways. J Biol Chem 2002; 277:43799-808. 21. Rajakumar A, Michael HM, Rajakumar PA, et al. Extra-placental expression of vascular endothelial growth factor receptor-1, (Flt-1) and soluble Flt-1 (sFlt-1), by peripheral blood mononuclear cells (PBMCs) in normotensive and preeclamptic pregnant women. Placenta 2005; 26:5633. 22. Roberts JM, Gammill H. Pre-eclampsia and cardiovascular disease in later life. Lancet 2005;366:961-2. 23. Vatten LJ, Romundstad PR, Holmen TL, Hsieh CC, Trichopoulos D, Stuver SO. Intrauterine exposure to preeclampsia and adolescent blood pressure, body size, and age at menarche in female offspring. Obstet Gynecol 2003;101:529-33.
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e6
www. AJOG.org
BASIC SCIENCE: OBSTETRICS
Circulating concentrations of soluble endoglin (CD105) in fetal and maternal serum and in amniotic fluid in preeclampsia Anne Cathrine Staff, MD, PhD; Kristin Braekke, MD; Guro M. Johnsen, MSc; S. Ananth Karumanchi, MD; Nina Kittelsen Harsem, MD OBJECTIVE: We explored whether concentrations of soluble endoglin in fetal serum and amniotic fluid and in maternal serum were elevated in preeclampsia. STUDY DESIGN: Umbilical vein serum, amniotic fluid, and maternal serum from 42 preeclamptic and 43 uncomplicated pregnancies that were delivered by cesarean section were analyzed by enzyme-linked immunosorbent assay for soluble endoglin. RESULTS: Median maternal serum and amniotic fluid soluble endoglin
concentrations were elevated in preeclampsia, compared with control pregnancies (66.9 ng/mL vs 15.1 ng/mL; P ⬍ .001, and 1.9 ng/mL vs 0.6 ng/mL; P ⬍ .001). Low concentrations of soluble endoglin were found in fetal circulation, which did not differ between preeclampsia
and control pregnancies (5.0 ng/mL vs 4.7 ng/mL; P ⫽ .2). Maternal serum soluble endoglin levels correlated with circulating soluble fmslike tyrosine kinase 1 concentrations. CONCLUSION: We confirmed elevated soluble endoglin in maternal circulation in preeclampsia, which correlated with soluble fms-like tyrosine kinase 1 concentrations and soluble fms-like tyrosine kinase 1/placental growth factor ratio. The fetus appears not to contribute to elevated circulating maternal soluble endoglin concentrations in preeclampsia.
Key words: endoglin, soluble fms-like tyrosine kinase 1 (sFlt1), preeclampsia, oxidative stress, placental growth factor (PlGF)
Cite this article as: Staff AC, Braekke K, Johnsen GM, Karumanchi A, Harsem NK. Circulating concentrations of soluble endoglin (CD105) in fetal and maternal serum and in amniotic fluid in preeclampsia. Am J Obstet Gynecol 2007;197:176.e1-176.e6.
P
reeclampsia is a pregnancy complication that affects 3%-4% of all pregnancies and represents a major threat to maternal and fetal health.1 The exact pathophysiologic condition is unknown, but generalized endothelial dysfunction with systemic inflammatory response is thought to be the final common
From the Departments of Obstetrics and Gynecology (Drs Staff, and Harsem, and Ms Johnsen) and Pediatrics (Dr Braekke), Ullevaal University Hospital and Faculty of Medicine, University of Oslo, Oslo, Norway, and the Departments of Medicine and Obstetrics and Gynecology, Beth Israel Deaconess Medical Center and Harvard Medical School, Boston, MA (Dr Karumanchi). Received Dec. 4, 2006; accepted Mar. 12, 2007. Reprints not available from the authors. Supported in part by research grants from the Eastern Norway Regional Health Authority (G.M.J.). 0002-9378/$32.00 © 2007 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2007.03.036
176.e1
pathway that leads to the maternal signs of preeclampsia with de novo hypertension and proteinuria in the second half of pregnancy.2 A shallow placentation, with abnormal invasion of cytotrophoblasts and incomplete remodeling of placenta-supplying maternal uterine spiral arteries, is central to the pathogenesis of preeclampsia.2 This results in a relative hypoxic uteroplacental circulation and augmented placental oxidative stress2 and thereby placental release of endothelial deranging factors to the maternal circulation. Pathogenic roles of soluble fms-like tyrosine kinase 1 receptor (sFlt1)3,4 and soluble endoglin (sEng)5,6 have been suggested in the development of preeclampsia. Both proteins are produced in excess from preeclamptic placentas and are found at elevated concentrations in maternal circulation in preeclampsia. Elevated serum concentrations of both proteins in the second trimester may predict the onset of preeclampsia.4,6,7 In vitro studies suggest that excess circulating sFlt1 and sEng in preeclampsia inhibit vascular endothelial growth factor
American Journal of Obstetrics & Gynecology AUGUST 2007
(VEGF)– and transforming growth factor beta (TGF-)–stimulation of endothelial-dependent nitrogen oxide activation and thereby oppose physiologic nitrogen oxide– dependent vasodilatation and vasomotor effects.5 Excess sFlt1 in maternal circulation binds the angiogenic factors VEGF and placental growth factor (PlGF) and thereby reduces their effects on maternal endothelium, with the ensuing hypertension and proteinuria.3 The effects of augmented sFlt1 concentrations and reduced VEGF action on the maternal vascular endothelium possibly could be due to impairment of VEGF-dependent activation of endothelial nitrogen oxide synthase.3,5 Endoglin, also named CD105, is a 180 kd protein that is localized on cell surfaces and is expressed highly on endothelial cells and on placental syncytiotrophoblasts.8 Endoglin functions as a coreceptor for the TGF- family. The TGF- receptors bind TGF-1 and -3, and endoglin modulates the action of these ligands, acting on intracellular signaling mediators and thereby altering
Basic Science: Obstetrics
www.AJOG.org transcriptional responses.9 Venkatesha et al5 recently demonstrated a novel, smaller (65 kd), placenta-derived soluble form of Eng (sEng) in the sera of pregnant women. Paralleling the sFlt1 findings, sEng is elevated in preeclamptic maternal serum and is reduced after delivery. The authors showed that sEng cooperated with sFlt1 in inducing endothelial dysfunction in vitro and severe preeclampsia-like illness in vivo in pregnant rats. SEng and sFlt1 blocked proangiogenic effects of TGF- and VEGF in vitro, respectively, and showed additive effects, which indicates that these soluble receptors may act in concert to disrupt endothelial integrity.5 The decline of sEng concentrations after delivery supports a placental origin of sEng in preeclampsia5 but cannot exclude the maternal vascular endothelium, circulating leukocytes, or the fetus as potential sources for elevated maternal serum concentration of sEng in preeclampsia. We previously have shown elevated sFlt1 concentration also in fetal circulation in preeclampsia, but at a much lower concentration than in the maternal circulation and therefore not likely to contribute to the maternal sFlt1 concentration in preeclampsia.10 The aim of the present study was to analyze concentrations of sEng in the fetal circulation and in matched maternal serum and amniotic fluid samples in preeclamptic, compared with uncomplicated, pregnancies. In addition, we wanted to explore a possible correlation between maternal sEng concentrations and previously measured sFlt1 and PlGF10 concentrations and 8-isoprostane,11 a well-established marker of oxidative stress.12
M ATERIALS AND M ETHODS Patient selection The study is part of an ongoing biobank collection of patient samples from complicated and uncomplicated pregnancies at Ullevaal University Hospital (Oslo, Norway). Women with singleton pregnancies, all delivered by cesarean section (n ⫽ 85), were included and comprised women with uncomplicated pregnancies (n ⫽ 42, control subjects) and pre-
eclamptic patients (n ⫽ 43). No women with chronic hypertension, renal disease, or diabetes mellitus were included. All patients were fasting; none of the patients were in active labor or had ruptured membranes or clinical signs of infection. Preeclampsia was defined as blood pressure augmentation after 20 weeks of gestation to ⬎140/90 mm Hg on ⱖ2 occasions 6 hours apart in a previously normotensive woman, combined with proteinuria. Proteinuria was defined as protein dip stick ⱖ1⫹ on ⱖ2 midstream urine samples 6 hours apart or a 24-hour urine excretion of ⱖ0.3 g protein in the absence of urinary infection. Severe preeclampsia was defined by the American College of Obstetricians and Gynecologists criteria,13 which includes women with blood pressure of ⱖ160 mm Hg systolic. The newborn infant birthweight percentiles were calculated according to national birth registry data.10 The study protocol was approved by the Regional Committee of Medical Ethics in Eastern Norway; informed written consent was obtained from each patient. Maternal blood samples were obtained before cesarean section in the 85 women (42 control subjects, 43 women with preeclampsia); amniotic fluid from 46 patients (16 control subjects, 30 women with preeclampsia) was sampled at cesarean section. After delivery of the placenta, 68 umbilical vein serum samples (delivering blood from the placenta to the fetus) were collected (36 control subjects, 32 women with preeclampsia). The processing of biological samples was described previously.10 Because sufficient quantities of serum from the umbilical arteries were not available, we analyzed fetal sEng concentrations in paired arterial and venous EDTAplasma samples from the umbilical cord in 13 random patients (7 controls, 6 women with preeclampsia).10
sEng determinations Enzyme-linked immunosorbent assay for human sEng was performed in duplicates according to the manufacturer’s instructions (R&D Systems, Minneapolis, MN).
Research
Previous analyses In previous publications from our biobank of preeclamptic and uncomplicated cesarean deliveries, we have reported maternal concentrations of the oxidative stress marker 8-isoprostane (8iso-prostaglandin F2␣),11 the angiogenesis-associated proteins sFlt1 and PlGF,10 and calprotectin, a marker of inflammation.14 Because of the ongoing inclusion of patients to this biobank, not all patients in the present study were included in the previous studies.
Statistical analysis The results are presented as median values (and 95% CI of the median for laboratory results). Statistical analyses were performed with the Statistical Package for the Social Sciences software (version 14.0; SPSS Inc, Chicago, IL). Differences in continuous variables between groups were tested by nonparametric MannWhitney tests. Spearman’s correlation was used to calculate correlation coefficients. A probability level of ⬍.05 was considered statistically significant.
R ESULTS Clinical characteristics of the 85 pregnant women who were included in the sEng maternal serum analyses are shown in the Table. Fewer patients were included in the fetal serum (n ⫽ 68 patients) and amniotic fluid (n ⫽ 46 patients) analyses because of difficulties in obtaining these samples; however, this subset of patient group characteristics did not differ significantly from the results in the Table. Only 1 woman in the preeclampsia group and none of the women in the uncomplicated pregnancy group delivered a baby small for gestational age (below the 10th weight percentile). Of 43 preeclamptic patients, 34 patients had severe preeclampsia, whereas 9 patients had mild preeclampsia. All control subjects had negative urine dip stick protein readings, whereas 30 of the 43 preeclamptic patients had 3⫹ at delivery, 6 patients had 2⫹, and 7 patients had 1⫹ readings. Three of the patients with preeclampsia also had evidence of HELLP (hemolysis, elevated liver enzymes, and low platelets) syn-
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e2
Research
Basic Science: Obstetrics
www.AJOG.org
TABLE
Clinical characteristics of patient groups included in the sEng maternal serum analysis Control pregnancies (n ⴝ 42)
Characteristic
Preeclampsia (n ⴝ 43)
P value
Patient age at delivery (y)
31 (21-40)
31 (18-42)
.5
Body mass index before pregnancy (kg/m )
22.5 (17.4-29.4)
23.7 (18.9-41.1)
.03*
Body mass index at delivery (kg/m )
28.0 (20.3-37.5)
31.0 (22.0-45.9)
.02*
................................................................................................................................................................................................................................................................................................................................................................................ 2 ................................................................................................................................................................................................................................................................................................................................................................................ 2 ................................................................................................................................................................................................................................................................................................................................................................................
Parity
0 (0-1)
0 (0-3)
.05
................................................................................................................................................................................................................................................................................................................................................................................
Gestational age at delivery (weeks)
38.7 (34.3-41.7)
32.9 (24.9-38.7)
⬍.001*
................................................................................................................................................................................................................................................................................................................................................................................
Neonatal weight (g)
3456 (2575-4165)
1692 (540-3856)
⬍.001*
................................................................................................................................................................................................................................................................................................................................................................................
Neonatal weight percentile
53 (10-95)
37 (7-90)
.001*
................................................................................................................................................................................................................................................................................................................................................................................
Blood pressure (mm Hg)
.......................................................................................................................................................................................................................................................................................................................................................................
Systolic at ⬍20 wk
110 (80-135)
117 (90-135)
.02*
Diastolic at ⬍20 wk
65 (50-80)
70 (50-89)
.02*
Systolic at delivery
120 (90-140)
160 (145-220)
⬍.001*
Diastolic at delivery
70 (60-92)
105 (90-119)
⬍.001*
....................................................................................................................................................................................................................................................................................................................................................................... ....................................................................................................................................................................................................................................................................................................................................................................... ....................................................................................................................................................................................................................................................................................................................................................................... ................................................................................................................................................................................................................................................................................................................................................................................
Values shown are median and range (minimum and maximum values). * P ⬍.05, Mann-Whitney test.
drome.13 In the preeclampsia group, 38 of the 43 women were delivered prematurely before week 37 of gestation, and 28 women were delivered before week 34 of gestation. All women in the uncomplicated pregnancy group were delivered after week 37 of gestation, except for 1 woman at week 34 of gestation. The individual results from the sEng analyses are shown in Figure 1. The preeclampsia group had a 4.4-fold elevated median maternal serum sEng concentration (66.9 ng/mL; 95% CI, 51.0-79.7) compared with the control group (15.1 ng/mL; 95% CI, 12.9-17.8; P ⬍ .001). The umbilical vein serum samples had very low concentrations of sEng compared with the maternal serum concentrations, and the preeclampsia group did not differ significantly from the control group (preeclampsia, 5.0 ng/mL [95% CI, 4.6-5.2] vs control subjects, 4.7 ng/mL [95% CI, 4.3-5.0], P ⫽ .2). The amniotic fluids had even lower concentrations of sEng, but median sEng concentration of the preeclampsia group was statistically significantly 3.2-fold elevated as compared with the control group (preeclampsia, 1.9 ng/mL [95% CI, 1.1-4.3] vs control subjects, 0.6 ng/mL [95% CI, 0.4-0.8]; P ⬍ .001). There was a significant positive correlation between maternal serum and amniotic fluid sEng concentrations in pre176.e3
eclampsia (Spearman’s correlation, 0.5; P ⫽ .006). Fetal sEng concentration was 15% lower in plasma compared with corresponding serum sample (umbilical vein: median, 3.9 vs 4.6, respectively; P ⫽ .02). Median sEng concentration for the random 13 paired umbilical cord plasma samples did not differ significantly between artery and vein blood (3.8 ng/mL vs 3.9 ng/mL; P ⫽ .6). We previously analyzed maternal serum concentration of sFlt1 and free PlGF in 68 of the 85 women who were included in this study.10 We found a significant positive correlation between maternal sEng and maternal sFlt1 concentrations, both for the whole patient group and for the preeclampsia group (whole patient group: Spearman’s correlation coefficient, 0.8 [P⬍.001] and preeclampsia group separately: Spearman’s correlation coefficient, 0.5 [P⫽.038]; Figure 2). Maternal sEng concentrations were correlated inversely with PlGF concentrations for the preeclampsia group (Spearman’s correlation, 0.5; P ⫽ .01). In the present study, we calculated the ratio between previously measured maternal serum sFlt110 and maternal serum-free PlGF10 concentrations and found that the median sFlt1/PlGF ratio was significantly elevated in the preeclampsia group, com-
American Journal of Obstetrics & Gynecology AUGUST 2007
pared with the control group (126.4 [95% CI, 89.5-161.2] vs 18.5 [95% CI, 12.7-29.6]; P ⬍ .001). Also, there was a positive correlation between the maternal sEng concentrations and sFlt1/PlGF ratios, both for preeclampsia and the control group (both: Spearman’s correlation coefficient, 0.5; P ⬍ .005). There was no significant correlation between body mass index at delivery and sEng concentrations for either patient groups (both: P ⬎ .07). We found no difference in median maternal sEng concentrations for either the control or preeclampsia group when we compared women with a body mass index at delivery ⬍30 kg/m2 with a BMI at ⱖ30 kg/m2 (both: P ⬎ .2). For the preeclampsia group, there was a significant negative correlation between the maternal sEng concentrations and gestational length (Spearman’s correlation, – 0.4; P ⫽ .005). Also, for the 28 women in the preeclampsia group who delivered prematurely before week 34 of gestation, median serum sEng was elevated significantly, compared with those delivered from week 34 and onwards (n ⫽ 15; 76.4 vs 39.5 ng/mL; P ⫽ .006). There was no statistically significant difference in median maternal sEng concentration between the severe (n ⫽ 34 cases) and mild (n ⫽ 9 cases) preeclamp-
sEng concentration Maternal serum sEng, ng/mL
P <.0001 250 200 150 100 50 0 Controls
Fetal serum sEng, ng/mL
B
P =.2
10.0 7.5
FIGURE 2
Maternal serum sEng concentrations 30000
25000
20000
15000
10000
5000
0 0
50
100
150
200
250
Maternal serum sEng, ng/mL
5.0
C OMMENT 2.5 0.0 Controls
C Amniotic fluid sEng, ng/mL
Preeclampsia
is associated not only with augmented oxidative stress but also with a generalized systemic inflammatory response,2 and we showed in a previous publication that maternal circulating concentration of calprotectin, an inflammatory protein, is elevated in preeclampsia.14 We previously presented maternal calprotectin concentration in 61 of the 85 patients who were included in this study. We found a positive correlation for the preeclampsia group between calprotectin and sEng concentrations in the maternal circulation in the present study (Spearman’s correlation coefficient, 0.6; P ⫽ .001), but not for the control group (P ⫽ .2).
Maternal serum sFlt1, pg/mL
FIGURE 1
A
Research
Basic Science: Obstetrics
www.AJOG.org
Preeclampsia
P <.0001 10.0 7.5 5.0 2.5 0.0 Controls
Preeclampsia
sEng concentration (nanograms/milliliter) in A, maternal serum, in fetal serum from the umbilical vein (B, delivering blood from the placenta to the fetus), and in C, amniotic fluid. Horizontal bars represent median values.
sia group of preeclamptic women (71.9 vs 51.1 ng/mL; P ⫽ .3). We also previously presented maternal plasma concentration of the oxidative stress marker 8-isoprostane (8-isoprostaglandin F2␣) in 54 of the 85 women who were included in this study.10 For the preeclampsia group, there was a significant positive correlation between maternal sEng concentrations and 8-isoprostane plasma concentrations (Spearman’s correlation coefficient, 0.5; P ⫽ .038). For the control group, there was no such positive significant correlation (P ⫽ .1). Preeclampsia
Our study is, to our best knowledge, the first report on concomitant concentrations of maternal, fetal, and amniotic fluid concentrations of sEng in pregnancy. We found very low levels of sEng in umbilical vein serum without significant difference between cases and control subjects, with the inference that fetal blood is unlikely to contribute to the elevated sEng levels that are found in maternal blood in preeclampsia. Our results also support a strong correlation between elevated circulating maternal sEng concentrations and sFlt1 concentrations in the maternal circulation, as demonstrated previously by Venkatesha et al.5 In accordance with their study, we found in our Norwegian study population of preeclamptic women that elevated sEng in maternal serum was most often associated with elevated sFlt1 concentrations (Figure 2) and elevated sFlt1/PlGF ratios, which suggests a molecular link between these circulating proteins in preeclampsia. Interestingly, we found a higher concentration of sEng in the amniotic fluid in preeclampsia compared with control subjects, albeit at very low concentrations, which is in contrast to our and other authors’ previous findings of very high concentrations of sFlt1 in amniotic fluid,10 which suggests different transplacental transport for sFlt1 and sEng. Maternal sEng falls after delivery,5 and both fetus and placenta theoretically could contribute, directly or indirectly, to the elevated maternal sEng concentra-
Maternal serum sEng concentrations (nanograms/milliliter) are correlated positively with maternal serum sFlt1 concentrations (picograms/milliliter). The lines represent regression line and 95% CI of the regression line; the closed triangle represents preeclampsia; the open circle represents control pregnancies.
tions that are found before delivery in preeclampsia. We demonstrated very low concentrations of sEng in the fetal circulation and no difference between the preeclampsia and control groups. We therefore can exclude a substantial fetal contribution to the elevated maternal sEng concentration in established preeclampsia. Our findings, in light of previous publications that demonstrated endoglin production in syncytiotrophoblasts5,8 and endothelial cells5,8 and sEng production in term placentas,5 are supportive of a direct or indirect placental origin of the elevated maternal serum sEng concentration in preeclampsia. The direct placental contribution would be shedding of sEng from the placenta. Indirect placental contribution could include freeing of placenta compounds before delivery to the maternal circulation; compounds that could activate the maternal endothelium and thereby endothelial cell freeing of sEng. Several substances are produced in excess in preeclamptic placentas; some substances could be speculated to induce maternal endothelial cell sEng shedding, such as trophoblast debris2 and oxidized lipids,
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e4
Research
Basic Science: Obstetrics
including the bioactive 8-isoprostane,15,16 the latter is capable of inducing endothelial activation in vitro.12 Presently, it is unknown how endoglin is shed to the circulation in pregnancy and why it is found in excess concentration in preeclampsia. Betaglycan is another coreceptor of TGF-, which shares 71% sequence similarity on the transmembrane and intracellular domain with endoglin.17 In vitro, betaglycan is shed to a free form by matrix metalloproteinases18; it has been proposed that cell-bound endoglin could be shed after cleavage in a similar manner in preeclampsia.5 Hypoxia is known to upregulate endoglin expression.19,20 We speculate that augmented placental oxidative stress in preeclampsia could induce augmented sEng production and possibly even contribute to cleavage of syncytiotrophoblast cell-bound endoglin, releasing a smaller form to the maternal circulation. Alternatively, activated leukocytes in the maternal circulation could also contribute to the elevated sEng concentrations in preeclampsia. In this regard, peripheral blood mononuclear cells have been demonstrated recently to contribute to the circulating sFlt1 in preeclampsia.21 It is also interesting to note that we found a positive correlation between 8-isoprostane and sEng in the maternal circulation in preeclampsia that is similar to the positive correlation demonstrated in preeclampsia between maternal 8-isoprostane and sFlt concentrations in our previous study.10 Such an association between maternal sEng and 8-isoprostane concentrations in preeclampsia, which link elevated systemic maternal oxidative stress with an elevated systemic concentration of sEng, has not been demonstrated previously. We suggest that a hypoxic placenta in preeclampsia could directly (by producing more 8-isoprostane and sEng) and/or indirectly (by producing/releasing factors from the uteroplacental circulation that could induce the elevation of 8-isoprostane and sEng that is seen in the maternal circulation) contribute to both the observed elevated concentrations of the oxidized lipid 8-isoprostane and the sEng and sFlt1 proteins in the maternal circula176.e5
tion. All 3 substances are theoretically capable of inducing endothelial dysfunction,3,5,12 which is a typical feature of preeclampsia. The pathophysiologic role of 8-isoprostane is little explored in preeclampsia but has been shown in vitro to cause endothelial dysfunction and platelet activation and vasoconstriction,12 which are also main characteristics of preeclampsia. A weakness of our study is that the patient groups were not matched for gestational age. Our study involved analyses of circulating factors in samples from matched fetal, maternal, and amniotic fluid compartments; the patients who were included were all delivered by cesarean section, and uncomplicated “control” pregnancies that are delivered prematurely are not found easily. We have demonstrated that, in our preeclampsia group, there was a significant negative correlation between maternal sEng concentrations and gestational length. However, this probably reflects that women with more severe preeclampsia, who need premature delivery, have the highest sEng concentrations, because there is no evidence for sEng decreasing with augmenting gestational age. Levine et al6 recently demonstrated an increase in maternal sEng concentration both in uncomplicated and preeclamptic pregnancies during the third trimester but was less pronounced and at a later gestational age in uncomplicated pregnancies. Our study therefore probably underestimates the difference in median maternal sEng serum concentration between the patient groups. Also, similar to the findings of Levine et al,6 we did not find any statistically significant difference in maternal serum sEng in established preeclampsia of the third trimester between women with severe and mild preeclampsia. This is in contrast to a previous study of Venkatesha et al,5 who found more elevated sEng in established severe preeclampsia, but included fewer patients with severe preeclampsia than the present study and with significant differences in gestational ages between the mild and severe cases. We chose serum samples in our study instead of plasma samples because the results could be more easily comparable
American Journal of Obstetrics & Gynecology AUGUST 2007
www.AJOG.org with the 2 reports from other pregnancy cohorts.5,6 The 13 matched plasma samples from the fetal circulation were used to illustrate that there was not a great difference in sEng concentration between the blood deriving from the placental side (umbilical vein) and blood deriving from the fetal side (umbilical arteries), which supports our conclusion that the fetus does not contribute to the elevated maternal sEng concentration in established preeclampsia. The findings reported in this article shed important clues on the pathogenesis of preeclampsia. The syndrome of hypertension and proteinuria is observed in preeclamptic mothers, but not in their infants. The fetal growth restriction that accompanies severe preeclampsia is thought to be the result of the placental ischemia. These clinical observations suggest that the circulating factors that mediate the maternal syndrome, although of placental origin, should be restricted primarily to the maternal vascular compartment and not the fetal circulation. Both sFlt1 and sEng have been hypothesized to be central mediators of the maternal syndrome on the basis of clinical and rodent studies.3,5 The findings that the fetuses of preeclamptic mothers do not have high circulating concentrations of either sEng or sFlt1 is consistent with the idea that fetuses do not experience proteinuria or hypertension like the mothers because they are not exposed to high concentrations of antiangiogenic factors. There is a need for further understanding of preeclampsia, which affects mortality and morbidity rates of fertile women and their offspring both on a short-term and long-term basis, with augmented cardiovascular risk later in life for both the women with preeclampsia and their offspring.22,23 Our data do not indicate that sEng in fetal circulation is a relevant risk marker for cardiovascular disease in offspring of mothers with preeclampsia, nor do our data support a fetal contribution to the maternal elevated concentration of sEng in preeclampsia. A placental origin is likely, but additional contribution from activated maternal vasculature or leukocytes cannot be ruled out. f
www.AJOG.org REFERENCES 1. Duley L. Pre-eclampsia and the hypertensive disorders of pregnancy. Br Med Bull 2003;67:161-76. 2. Redman CW, Sargent IL. Latest advances in understanding preeclampsia. Science 2005; 308:1592-4. 3. Maynard SE, Min JY, Merchan J, et al. Excess placental soluble fms-like tyrosine kinase 1 (sFlt1) may contribute to endothelial dysfunction, hypertension, and proteinuria in preeclampsia. J Clin Invest 2003;111:649-58. 4. Levine RJ, Maynard SE, Qian C, et al. Circulating angiogenic factors and the risk of preeclampsia. N Engl J Med 2004;350:672-83. 5. Venkatesha S, Toporsian M, Lam C, et al. Soluble endoglin contributes to the pathogenesis of preeclampsia. Nat Med 2006;12:642-9. 6. Levine RJ, Lam C, Qian C, et al. Soluble endoglin and other circulating antiangiogenic factors in preeclampsia. N Engl J Med 2006;355:992-1005. 7. Vatten LJ, Eskild A, Nilsen TI, Jeansson S, Jenum PA, Staff AC. Changes in circulating level of angiogenic factors from the first to second trimester as predictors of preeclampsia. Am J Obstet Gynecol 2007;196:239.e1-6. 8. Gougos A, St Jacques S, Greaves A, et al. Identification of distinct epitopes of endoglin, an RGD-containing glycoprotein of endothelial cells, leukemic cells, and syncytiotrophoblasts. Int Immunol 1992;4:83-92. 9. Barbara NP, Wrana JL, Letarte M. Endoglin is an accessory protein that interacts with the signaling receptor complex of multiple members
Basic Science: Obstetrics of the transforming growth factor-beta superfamily. J Biol Chem 1999;274:584-94. 10. Staff AC, Braekke K, Harsem NK, Lyberg T, Holthe MR. Circulating concentrations of sFlt1 (soluble fms-like tyrosine kinase 1) in fetal and maternal serum during pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2005;122:33-9. 11. Harsem NK, Braekke K, Staff AC. Augmented oxidative stress as well as antioxidant capacity in maternal circulation in preeclampsia. Eur J Obstet Gynecol Reprod Biol 2006;128:209-15. 12. Patrono C, FitzGerald GA. Isoprostanes: potential markers of oxidant stress in atherothrombotic disease. Arterioscler Thromb Vasc Biol 1997;17:2309-15. 13. American College of Obstetricians and Gynecologists. Diagnosis and management of preeclampsia and eclampsia: ACOG practice bulletin no.: 33. Obstet Gynecol 2002; 99:159-67. 14. Braekke K, Holthe MR, Harsem NK, Fagerhol MK, Staff AC. Calprotectin, a marker of inflammation, is elevated in the maternal but not in the fetal circulation in preeclampsia. Am J Obstet Gynecol 2005;193:227-33. 15. Staff AC, Halvorsen B, Ranheim T, Henriksen T. Elevated level of free 8-iso-prostaglandin F2alpha in the decidua basalis of women with preeclampsia. Am J Obstet Gynecol 1999;181:1211-5. 16. Walsh SW, Vaughan JE, Wang Y, Roberts LJ. Placental isoprostane is significantly increased in preeclampsia. FASEB J 2000; 14:1289-96.
Research
17. Cheifetz S, Bellon T, Cales C, et al. Endoglin is a component of the transforming growth factor-beta receptor system in human endothelial cells. J Biol Chem 1992;267:19027-30. 18. Velasco-Loyden G, Arribas J, Lopez-Casillas F. The shedding of betaglycan is regulated by pervanadate and mediated by membrane type matrix metalloprotease-1. J Biol Chem 2004;279:7721-33. 19. Docherty NG, Lopez-Novoa JM, Arevalo M, et al. Endoglin regulates renal ischaemia-reperfusion injury. Nephrol Dial Transplant 2006;21:2106-19. 20. Sanchez-Elsner T, Botella LM, Velasco B, Langa C, Bernabeu C. Endoglin expression is regulated by transcriptional cooperation between the hypoxia and transforming growth factor-beta pathways. J Biol Chem 2002; 277:43799-808. 21. Rajakumar A, Michael HM, Rajakumar PA, et al. Extra-placental expression of vascular endothelial growth factor receptor-1, (Flt-1) and soluble Flt-1 (sFlt-1), by peripheral blood mononuclear cells (PBMCs) in normotensive and preeclamptic pregnant women. Placenta 2005; 26:5633. 22. Roberts JM, Gammill H. Pre-eclampsia and cardiovascular disease in later life. Lancet 2005;366:961-2. 23. Vatten LJ, Romundstad PR, Holmen TL, Hsieh CC, Trichopoulos D, Stuver SO. Intrauterine exposure to preeclampsia and adolescent blood pressure, body size, and age at menarche in female offspring. Obstet Gynecol 2003;101:529-33.
AUGUST 2007 American Journal of Obstetrics & Gynecology
176.e6