A biological model for the regulation of peri-implantational hemostasis and menstruation

OBJECTIVE

To delineate the physiologic mechanisms whereby the human endometrium maintains hemostasis during endovascular trophoblast invasion but permits menstrual hemorrhage.

METHODS

Experimental results are presented that are relevant to developing a comprehensive biological model for studying peri-implantational hemostasis and menstruation.

RESULTS

A marked increase in the expression of tissue factor (TF) and type-1 plasminogen activator inhibitor (PAI-1) and an inhibition of tissue-type and urokinase-type plasminogen activators (tPA and uPA, respectively), matrix metalloproteinases (MMP), and endothelin-1 (ET-1) expression accompany progestin-induced decidualization of estrogen-primed endometrial stromal cells both in vivo and in vitro. The presence of these important regulators of hemostasis, fibrinolysis, extracellular matrix (ECM) turnover, and vascular tone in decidualized human stromal cells and decidual cells isolated from gestational endometrium suggests a mechanism to explain the absence of hemorrhage during invasion of the endometrial vasculature by trophoblasts. Conversely, progesterone withdrawal reduces TF and PAI-1 expression and increases tPA, uPA, MMP, and ET-1 expression accounting for the hemorrhage, enhanced fibrinolysis, ECM degradation, and ischemic spiral arterial vascular injury characterizing menstruation.

CONCLUSION

Perivascular decidualized endometrial stromal cells are spatially and temporally positioned to promote endometrial hemostasis during implantation but, paradoxically, promote the hemorrhage of menstruation via their hormone-regulated expression of hemostatic, proteolytic, and vasoactive proteins.

Corticotropin-releasing factor and its binding protein: maternal serum levels in term and preterm deliveries

OBJECTIVE

The primary objective of this investigation was to evaluate whether maternal serum corticotropin-releasing factor levels during pregnancy were predictive of spontaneous preterm delivery.

STUDY DESIGN

Maternal serum levels of corticotropin-releasing factor and its binding protein were measured from 20 weeks of gestation in a cross-sectional study of 396 asymptomatic women at high risk for preterm delivery.

RESULTS

Gestational age-specific corticotropin-releasing factor levels were not consistently or substantially increased for preterm compared with term deliveries, whether preterm delivery was due to preterm labor or preterm premature rupture of membranes. The binding protein for corticotropin-releasing factor did not vary according to gestational age until term, when it dropped substantially.

CONCLUSION

Serum corticotropin-releasing factor levels do not appear to be an important predictor of preterm birth in asymptomatic patients who subsequently have either preterm labor or preterm premature rupture of membranes. Nevertheless, the drop in the corticotropin-releasing factor binding protein level at term suggests that the bioavailability of corticotropin-releasing factor increases as parturition approaches.

Production of premature delivery in pregnant rhesus monkeys by androstenedione infusion

The endocrine mechanism involved in term and preterm delivery in primates, including pregnant women, are poorly understood. In the term monkey, fetal plasma androgen concentration rises to two hundred times the maternal concentration which remains unchanged. Placental conversion of androgen to estrogen results in increased maternal plasma estrogen concentration at term in both pregnant nonhuman primates and women. In the present study, continuous infusion of androstenedione to 0.8 gestation monkeys resulted in the premature occurrence of labor-type myometrial activity and increases in maternal plasma estrogen, oxytocin and amnion fibronectin concentrations similar to those measured at normal-term labor. Androstenedione induction of these normal-term biochemical and endocrine changes accompanied by fetal membrane rupture, cervical dilatation and live delivery provides a rich opportunity to study the molecular and physiological mechanisms of both term and preterm labor in primates.

Expression of 11 beta-hydroxysteroid dehydrogenase during decidualization of human endometrial stromal cells

This study evaluated the expression of the corticosteroid-metabolizing enzyme 11 beta-hydroxysteroid dehydrogenase (11 beta HSD) during in vitro decidualization of human endometrial stromal cells. The cultured stromal cells displayed both NADP(+)-dependent (type 1) and NAD(+)-dependent (type 2) 11 beta HSD activities under basal conditions. Although the cells did not respond to estradiol (E2) added alone, catalytic levels of both isoforms were enhanced by medroxyprogesterone acetate (MPA) and further enhanced by E2 plus MPA. Type I messenger RNA (mRNA) was undetected by Northern analysis of total RNA, but was evident as a 1.5-kilobase band in polyadenylated selected RNA from E2- plus MPA-treated cultures. Use of RT-PCR to augment the sensitivity of mRNA detection revealed the presence of type I mRNA as a faint band in the MPA-treated cultures and as an intense band in the E2- plus MPA-treated cultures. Thus, type I mRNA is present as a low abundance message in the cultured stromal cells whose steady state levels parallel progestin-enhanced enzyme activity. As the expression of several progestin-regulated decidualization markers is also augmented by E2, the results of the present study reveal a correlation between enhanced 11 beta HSD expression and the decidualization reaction. Time-course measurements indicated that elevated 11 beta HSD expression is an early event in the decidualization response, which precedes E2- plus MPA-enhanced PRL production by several days. Clear dose-response effects on both type 1 and type 2 11 beta HSD activities were obtained in cells incubated with 10(-8) mol/liter E2 added together with MPA at concentrations that approximated circulating progesterone levels from the luteal phase (10(-9) mol/liter) through pregnancy (10(-7) mol/liter). Corticosteroids are thought to exert toxic and teratogenic effects on the implanting embryo and could influence trophoblast invasion by regulating extracellular matrix turnover. Therefore, the novel finding that decidualization involves marked enhancement of the corticosteroid-metabolizing capacity of stromal cells suggests a mechanism by which decidual cells could affect the health and invasiveness of implanting trophoblastic cells.

Effects of thrombin on steroid-modulated cultured endometrial stromal cell fibrinolytic potential

By virtue of their unique chronic expression of tissue factor, the primary initiator of hemostasis, decidualized endometrial stromal cells are capable of significant thrombin generation after vascular disruption. In addition to its potent procoagulant effects, thrombin modifies endothelial and glomerular cell fibrinolytic activity. Therefore, we evaluated whether thrombin affected the expression of endometrial stromal cell urokinase-type (uPA) and tissue-type (tPA) plasminogen activators and their primary inhibitor, type 1 plasminogen activator inhibitor (PAI-1), and whether ovarian steroids modulated putative thrombin effects. Confluent stromal cell cultures were incubated in a defined medium containing vehicle control, 10(-8) mol/L estradiol (E2), 10(-7) mol/L medroxyprogesterone acetate (MPA), or E2 plus MPA for 4 days. The medium was then collected and exchanged for medium containing the corresponding steroids with or without thrombin and the specific thrombin inhibitor, D-phenyl-alanyl-propyl-arginine-chloromethyl ketone, for an additional 24 h. The conditioned medium was then collected and analyzed for immunoreactive (ir) uPA, tPA, and PAI-1 by enzyme-linked immunosorbent assay and for PA activity by chromogenic assay, whereas Northern analysis of the cells was employed to evaluate the expression of thrombin receptor, uPA, tPA, and PAI-1 messenger ribonucleic acid (mRNA) species. The latter studies revealed that confluent cultures incubated in defined medium expressed the 3.45-kilobase thrombin receptor message. Steady state levels of thrombin receptor mRNA were unaffected by exogenous steroids. Thrombin added in the absence of exogenous steroids elevated concentrations of ir tPA, uPA, and PAI-1 compared with control cultures. Conversely, in the absence of added thrombin, MPA added alone or together with E2 inhibited levels of ir tPA and uPA while stimulating PAI-1 levels despite the lack of a response to E2 alone. Interestingly, thrombin counteracted this progestin inhibition of tPA and uPA expression and augmented the progestin-enhanced expression of PAI-1. Northern analysis revealed that steady state levels of tPA and uPA mRNA were also enhanced by thrombin in both control and steroid-containing cultures. Net PA activity reflects the balance between PA and PAI-1. In the absence of thrombin, there is virtually no detectable tPA activity and minimal uPA activity in progestin-exposed cultures. However, thrombin elicited significant increases in tPA and uPA activity in control and E2-treated cultures. Despite the molar excess of PAI-1 in MPA-treated and E2- plus MPA-treated cultures, thrombin reversed progestin inhibition of PA activity. Predictably, the addition of D-phenyl-alanyl-propyl-arginine-chloromethyl ketone, blocked the effects of thrombin on PAI-1, tPA, and uPA protein and mRNA expression and PA activity. In summary, thrombin enhances endometrial stromal cell fibrinolytic and extracellular matrix-degrading protease activity in vitro. Such processes occurring in vivo would probably play a role in menstruation and abnormal uterine bleeding.

Corticotropin-releasing hormone and related pituitary-adrenal axis hormones in fetal and maternal blood during the second half of pregnancy

There is little information available concerning the ontologic development of the human hypothalamic-pituitary-adrenal (HPA) axis nor of the potential interactions among fetal, maternal and placental-derived HPA axis hormones. This study evaluated levels of these hormones in matched maternal and fetal pairs during the second half of uncomplicated pregnancies. Immunoassays were used to measure serum concentrations of corticotropin-releasing hormone (CRH), adrenocorticotropin (ACTH) and cortisol in 104 matched fetal and maternal blood samples. Fetal specimens were obtained by percutaneous umbilical blood sampling (PUBS) between 18 and 40 weeks in patients whose pregnancies resulted in healthy, term infants. Correlations among these hormones, and the effect of gestational age were assessed. Maternal CRH concentrations [median (range)] [1.10 ng/ml (0.15 to 23.69)] were significantly greater than fetal values [0.35 ng/ml (0.07 to 1.0)]. Levels of maternal CRH (r = 0.73; p < 0.001) but not fetal CRH (r = 0.01; p = 0.98) correlated with gestational age. Maternal ACTH decreased (r = -0.21; p = 0.04) while fetal ACTH increased (r = 0.35; p < 0.003) with gestational age. Both maternal (r = 0.45; p < 0.001) and fetal (r = 0.57; p < 0.001) cortisol levels increased with gestational age. Maternal serum CRH values correlated best with fetal cortisol (r = 0.40; p = 0.0002) and correlated modestly with maternal cortisol (r = 0.28; p = 0.01), fetal ACTH (r = 0.24; p = 0.03) and fetal CRH (r = 0.23; p = 0.04); but not with maternal ACTH (r = -0.12; p = 0.3). Maternal CRH concentrations increase in the third trimester and correlate with rising fetal cortisol levels.

The diagnosis of preterm labor and the prediction of preterm delivery

Clinical approaches to the diagnosis of PTL and the prediction of PTD are complicated by the absence of a gold standard for the pathogenic process leading to PTD. There is also substantial overlap between the signs and symptoms of PTL and impending PTD, and the normal processes of pregnancies destined to remain uncomplicated (e.g., our inability to convincingly differentiate PTL from Braxton-Hicks contractions). Our emphasis on the diagnosis of PTL rather than the pathogenic process preceding PTD also results in failure to detect the 50% of spontaneous PTDs in which uterine contractions follow PPROM. Thus, clinical predictors of incipient PTD including cervical change, uterine contractions, vaginal bleeding, risk scoring schemes, and fetal breathing activity, either have poor sensitivity or specificity, or are accurate only at late stages in the pathogenic process. The most promising approaches to the detection of impending PTD are laboratory indices of the putative pathogenic processes including: maternal serum or plasma CRH, salivary E3, serum collagenase and cervicovaginal cytokines, granulocyte elastase, and FFN levels. However, even if these indices prove sensitive, specific, and early predictors of PTD, they will be useful only if more appropriate therapies are found to treat patients. The latter will depend on addressing the primary causes of chorionic-decidual cell activation (e.g., infection, stress, utero-placental ischemia, hemorrhage, endocrinopathies).

Accuracy of ultrasound estimation of fetal weight in multiple pregnancies

Our objective was to assess the accuracy of ultrasonographic estimation of fetal weight in twins and triplets as compared to singleton pregnancies. Retrospective analysis was undertaken of ultrasound data of all fetuses who underwent an examination within 1 week of delivery (singletons 1832, twins 518, triplets 51). At birth weights below 2500 g, there was a significant overestimation of fetal weight in twins as compared to singletons, but the accuracy of the estimate was the same, except in twins between 1500 and 2499 g, when the weight was based on abdominal circumference and femur length alone. At birth weights of more than 2500 g, no difference was detected between twins and singletons. At all birth weights below 2500 g, the accuracy of weight estimation in triplets was equal to that in singletons and there were no triplets above this weight. We conclude that ultrasonographic estimation of fetal weight is as accurate in twins and triplets as it is in singletons.

Opposing actions of transforming growth factor-beta and glucocorticoids in the regulation of fibronectin expression in the human placenta

Alterations in the expression of extracellular matrix (ECM) proteins in the placenta and fetal membranes have been linked to parturition whether occurring before or at term. In the present study, we examined the individual and combined effects of transforming growth factor (TGF)-beta and dexamethasone (DEX) on the expression of oncofetal fibronectin (onfFN), i.e. a major ECM protein synthesized by placenta, in cytotrophoblasts isolated from human term placentas to establish a model system from which to evaluate the actions of positive and negative regulators of ECM protein expression in the human placenta. Cytotrophoblasts were maintained for 21=62 h in medium supplemented with 4% charcoal-stripped calf serum in the presence or absence of TGF-beta (2 ng/mL) and DEX (10(-7) mol/L). Levels of onfFN in culture media were determined by immunoassay. TGF-beta treatment alone induced approximately a 150% increase in media levels of onfFN after 21 and 45 h of culture when compared with control, whereas DEX treatment alone reduced levels of onfFN to 15% of control levels. Media levels of onfFN in cells treated with both TGF-beta and DEX were 40-90% of control levels. Similarly, treatment of cells with TGF-beta alone promoted a 100-250% increase in rates of FN synthesis and levels of FN messenger ribonucleic acid, whereas DEX treatment alone reduced these indices of FN expression to approximately 10% of control levels. In cells treated with TGF-beta and DEX, levels of ECM protein synthesis and FN messenger ribonucleic acid were between 30 and 100% of control values. Similar patterns of regulation of FN expression by TGF-beta and DEX were observed when experiments were carried out in serum-free medium. Our results suggest that during pregnancy, TGF-beta and glucocorticoids may be important opposing physiological regulators of placental ECM protein expression.

Plasminogen activator activity during decidualization of human endometrial stromal cells is regulated by plasminogen activator inhibitor 1

Progesterone acts on the estradiol (E2)-conditioned human endometrium to induce decidualization of stromal cells. Consistent with these differential hormone actions in vivo, progestins regulate several end points of decidualization in human endometrial stromal cell monolayers, and E2 augments the effects of progestin. This study shows that in vitro decidualization of the stromal cells is accompanied by diminished plasminogen activator (PA) expression. Polyacrylamide gel electrophoretic separation after immunoprecipitation of biosynthetically labeled PAs revealed that medroxyprogesterone acetate (MPA) lowered levels of secreted tissue type PA (tPA) at 67 kilodaltons and urokinase type PA (uPA) at 55 kilodaltons. These levels were reduced further by E2 plus MPA despite a lack of response to E2 alone. Although tPA activity was readily measured by a chromogenic assay, detection of uPA activity required prior activation, indicating that uPA is released as the pro-uPA zymogen. Comparisons of levels of immunogenic PAs, as measured by specific enzyme-linked immunosorbent assays, with the corresponding catalytic activities revealed selective progestational inhibition of PA activity vs. antigen after 3 days of experimental incubation. Thus, 10(-7) mol/L MPA produced about a 2-fold greater reduction of levels of PA activity than that of its corresponding antigen. More strikingly, 10(-8) mol/L E2 plus 10(-7) mol/L MPA virtually eliminated both tPA activity (99% inhibition; P < 0.005) and uPA activity (93% inhibition; P < 0.005); the reductions in levels of the corresponding antigens were only about 50% of the control levels and did not attain statistical significance. Only after 3-6 days of incubation with E2 plus MPA was statistically significant inhibition achieved for immunogenic levels of both tPA (P < 0.05) and uPA (P < 0.005). Preferential inhibition of levels of PA activities compared with those of the corresponding PA antigens reflects the action of the potent PA inhibitor PAI-1. Thus, the concentration of PAI-1 in the stromal cell-conditioned medium at the end of 0-3 days exceeded those of tPA and uPA, respectively, by 28- and 12-fold in response to MPA and by 52- and 25-fold in response to E2 plus MPA. Extrapolation of these in vitro results to the events of the luteal phase, whose steroidal milieu is mimicked by E2 plus MPA, indicates that decidual cell-derived PAI-1 is a key regulator of proteolytic degradation of extracellular matrix and fibrinolysis during implantation and menstruation.

Reduction of extracellular matrix protein expression in human amnion epithelial cells by glucocorticoids: a potential role in preterm rupture of the fetal membranes

Low levels of expression of extracellular matrix (ECM) proteins in chorioamniotic membranes is a characteristic of prematurely ruptured membranes, a condition associated with 40% of preterm deliveries. In light of the rise in levels of glucocorticoids (GC) in amniotic fluid associated with preterm labor, in the present study we examined the effects of GC on the expression of major ECM proteins in cultures of amnion epithelial cells recovered after digestion of human term amnions. Amnion cells were maintained with and without 10(-7) mol/L dexamethasone (DEX), and levels of the ECM protein fibronectin (FN) were determined by enzyme-linked immunosorbent assay. DEX treatment reduced FN expression in amnion epithelial cells to 15-30% of control levels and reduced FN expression in placental cells to 30-50% of control levels. Conversely, DEX treatment weakly stimulated FN expression in chorion cell cultures. DEX treatment did not affect the total level of amnion cell protein, indicating that the effects of DEX in amnion cells did not result from a general reduction in protein synthesis. Cortisol and DEX reduced FN expression in amnion cells, with half-maximal effective concentrations of approximately 60 and 8 nmol/L, respectively. In immunoprecipitation studies, DEX treatment reduced FN and collagen III synthesis to 20% of control levels, suggesting that GC may coordinately reduce the synthesis of major ECM proteins in amnion cells. Similarly, DEX treatment reduced the levels of FN messenger ribonucleic acids in amnion cells to approximately 15% of control levels. DEX treatment also promoted a marked reduction in FN expression in amnion cells cultured in serum-free medium to 10-50% of control levels. Our results indicate that GC negatively regulate ECM protein expression in amnion epithelial cells, suggesting a potential role in the genesis of altered fetal membrane ECM protein expression associated with prematurely ruptured membranes.

Incidence of pregnancy-induced hypertension among gestational diabetics

Although hypertensive disorders of pregnancy are more likely to occur in pregestational diabetics, the question of whether they occur more frequently in gestational diabetics or certain subsets of gestational diabetics remains unclear. This study compared 197 gestational diabetics with 197 control patients matched on the basis of age, race, parity, and prepregnancy weight. No significant difference was found between the two groups in the incidence of either pregnancy-induced hypertension or preeclampsia. There was, however, a small but significant elevation in mean arterial blood pressure in the third trimester in gestational diabetics compared with control patients (90.1 versus 87.5 mm Hg; p = 0.006). Mean arterial pressures were also higher in diabetic patients on insulin compared with those on diet, and higher in diabetic patients diagnosed early (less than 24 weeks) compared with those diagnosed late (more than 24 weeks) in pregnancy; however, there were larger numbers of chronic hypertensives in these two groups. We conclude that gestational diabetics do not develop pregnancy-induced hypertension more frequently. Small increases in blood pressure late in pregnancy in these patients achieve statistical significance, but their clinical relevance is unclear.

Maternal serum levels of alpha-fetoprotein and fetal fibronectin as markers of labor in the third trimester

Unexplained elevations of maternal serum alpha-fetoprotein (MSAFP) in the second trimester have been found to be associated with a two- to fourfold increase in the rate of preterm delivery, but the sensitivity is low. Therefore, we reasoned that MSAFP levels in the third trimester could prove to be a more useful biochemical marker to predict labor. The presence of placental and membrane-derived fetal fibronectin (FFN) in cervicovaginal secretions has recently been shown to predict preterm delivery with a sensitivity of 81.7% and specificity of 82.5%. We postulated that damage to membranes and microscopic breakdown of fetomaternal blood barrier during labor might result in release of AFP or FFN into maternal serum. Maternal serum alpha fetoprotein and fetal fibronectin levels were measured prospectively in 29 patients in active labor at term and in 25 controls undergoing elective cesarean section. Neither MSAFP nor serum FFN levels were associated with labor at term. We did, however, note significantly higher MSAFP levels in mothers bearing male fetuses versus female fetuses (p < 0.01). Since the current literature supports a sex difference in the MSAFP levels in the second trimester, this does not appear to change as gestation advances. Further studies are needed to determine if, in addition to maternal weight and race, MSAFP levels should be also adjusted for fetal sex diagnosed on sonography.

Biological mechanisms underlying the clinical effects of RU 486: modulation of cultured endometrial stromal cell plasminogen activator and plasminogen activator inhibitor expression

The abortifacient and menstrual effects of the potent antiprogestin, RU 486, are associated with both endometrial hemorrhage and extracellular matrix (ECM) degradation. Such processes reflect reduced perivascular decidual cell hemostatic and increased ECM-degrading protease activity. Therefore, we assessed the effects of RU 486 administration on the expression of immunoreactive (ir) endometrial stromal cell urokinase-type (uPA) and tissue-type (tPA) plasminogen activator and their activities as well as levels of ir type 1 plasminogen activator inhibitor (PAI-1) using a well characterized in vitro model of decidualization. Thus, confluent stromal cell cultures were exposed to vehicle control, 10(-8) mol/L estradiol (E2), 10(-7)-10(-8) mol/L medroxyprogesterone acetate (MPA), E2 plus MPA, or 10(-6)-10(-7) mol/L RU 486 alone or in combination with MPA or E2 plus MPA for 3-4 days. Compared to the vehicle control, E2 and RU 486 used alone had no effect on levels of ir PAI-1, uPA, or tPA or on PA activity in the conditioned medium. In contrast, MPA and E2 plus MPA decreased ir uPA and tPA levels and their corresponding activities, whereas MPA increased, and E2 plus MPA further increased ir PAI-1 release. These effects of progestin were blocked by a log higher concentration of RU 486. Similar results were obtained for steady state PAI-1 messenger ribonucleic acid levels. To determine if RU 486 reversed progestin-inhibited stromal cell uPA and tPA release and progestin-enhanced PAI-1 expression, confluent cultures were exposed to 10(-8) mol/L E2 plus 10(-7) mol/L MPA for 10 days, washed, and reexposed to E2 plus MPA, steroid-free medium, or RU 486 for 3-5 or 9-11 days. Compared with cultures maintained in E2 plus MPA for 3-5 days, withdrawal to a steroid-free medium failed to affect stromal cell ir PAI-1, uPA, or tPA levels. In contrast, exposure to RU 486 for 3-5 days increased ir uPA and tPA levels 5- to 8-fold (P < 0.02) while reducing PAI-1 levels by 85% (P < 0.04). By 9-11 days of treatment, steroid withdrawal and RU 486 exerted similar effects on ir PAI-1, tPA, and uPA levels. Comparable results were obtained for PAI-1, uPA, and tPA steady state messenger ribonucleic acid levels. These findings indicate that RU 486 blocks and reverses progestin-inhibited PA expression, suggesting a mechanism for RU 486-induced endometrial hemorrhage and ECM dissolution.

Alloimmune thrombocytopenia: fetal and neonatal losses related to cordocentesis

OBJECTIVE

This report describes the increased risks of cordocentesis in fetuses affected with alloimmune thrombocytopenia.

STUDY DESIGN

As part of a multicenter treatment study clinical and laboratory data from five pregnancies with alloimmune thrombocytopenia in which there was a fetal or neonatal loss associated with cordocentesis were reviewed. The fetal or neonatal deaths were all thought to be a result of exsanguination. These fetuses were compared with a group of 44 affected fetuses who underwent the same procedure but who survived. The data were analyzed by the Wilcoxon rank-sum test and the two-tailed Fisher's exact test. A p value < 0.05 was considered significant.

RESULTS

The mean platelet count at cordocentesis was significantly lower in the cases than in the controls (5.8 vs 32.8 x 10(9)/L, p = 0.005). The incidence of antenatal intracranial hemorrhage in the untreated sibling of the prior affected pregnancy was significantly greater in the cases than in the controls (two of five vs one of 42, p = 0.02).

CONCLUSION

Fetuses affected with alloimmune thrombocytopenia are at increased risk for fatal exsanguination associated with cordocentesis.

Fetal and maternal plasma endothelin levels during the second half of pregnancy

OBJECTIVE

Our objective was to evaluate maternal and fetal endothelin concentrations in uncomplicated pregnancies across the second half of pregnancy.

STUDY DESIGN

Paired (n = 64) maternal venous and fetal umbilical venous or arterial samples were obtained during cordocentesis. In addition, eight neonatal umbilical vein samples were obtained immediately after delivery. Samples were assessed for hematocrit and pH, and concentrations of endothelin were measured by sensitive enzyme immunoassay.

RESULTS

No significant correlation was found between either fetal or maternal endothelin levels and gestational age (r = 0.01, p = 0.91 and r = 0.07, p = 0.5, respectively). Fetal plasma endothelin concentrations were significantly lower than neonatal umbilical vein endothelin levels [median 2.5 pg/ml (range 0.9 to 5.73) vs 15.77 pg/ml (8.12 to 19.58), respectively; p < 0.0001] but significantly higher than maternal levels [1.3 pg/ml (0.8 to 3.25); p < 0.0001]. In addition, endothelin values were higher in the umbilical artery than in the umbilical vein, but this difference failed to achieve statistical significance [2.89 pg/ml (1.61 to 5.73) vs 2.29 pg/ml (0.9 to 5.70), respectively; p = 0.06]. No correlation was noted between fetal and maternal endothelin levels (r = 0.12, p = 0.36).

CONCLUSION

Fetal endothelin levels were significantly higher than maternal levels, but neither correlated with gestational age across the second half of pregnancy.

Reduction of annexin-V (placental anticoagulant protein-I) on placental villi of women with antiphospholipid antibodies and recurrent spontaneous abortion

OBJECTIVE

The mechanism by which antiphospholipid antibodies are associated with pregnancy loss and thromboembolic conditions has yet to be elucidated. Annexin-V, an anticoagulant phospholipid-binding protein, is normally present in syncytiotrophoblasts lining the placental villi, where it may play a role in the maintenance of intervillous blood fluidity. We therefore investigated the distribution of annexin-V in placentas of patients with antiphospholipid antibodies in situ and then used short-term villous cultures to study the direct effect of antiphospholipid antibodies on the immunolocation of annexin-V.

STUDY DESIGN

We performed a blinded study by means of computerized morphometric analysis of placental tissues that were stained for annexin-V with affinity-purified polyclonal antibody in an avidin-biotin peroxidase system. The distribution of villous surface annexin-V on cross sections of placentas of patients with antiphospholipid antibodies was compared with that of placentas from patients with uncomplicated pregnancies, elective abortions, and pregnancy losses not associated with antiphospholipid antibodies (n = 8 for each group). We quantitated villous surface annexin-V in cultured placental villi that were incubated with antiphospholipid antibodies immunoglobulin G compared with normal immunoglobulin G and measured annexin-V levels by enzyme-linked immunosorbent assay in conditioned media and in the villi.

RESULTS

The mean villous surface annexin-V of the group with antiphospholipid antibodies was 26.2% +/- 17% (SD) versus 93.9% +/- 5.7% in the normal control group (p < 0.0001). Villi from patients undergoing elective abortions and with pregnancy losses that were not attributed to antiphospholipid antibodies also showed higher mean villous surface annexin-V levels (86.9% +/- 10.6% and 83.5% +/- 11.3%, respectively, p < 0.0001). Organ culture of normal placental villi with affinity-purified immunoglobulin G from patients with antiphospholipid antibodies showed a dose-dependent decrease of villous surface annexin-V over a concentration range of 1.5 micrograms/ml to 1.5 mg/ml. Annexin-V concentrations in conditioned media were significantly lower in the presence of antiphospholipid antibodies immunoglobulin G compared with normal immunoglobulin G (49.4 +/- 8.9 ng/gm wet weight vs 57.2 +/- 11.5 ng/gm, respectively, p < 0.05). In contrast, the mean level of annexin-V in placental villi incubated with antiphospholipid antibodies immunoglobulin G was greater than in villi incubated with normal immunoglobulin G, 1328 +/- 130 ng/gm wet weight versus 1183 +/- 165 ng/gm (p < 0.02).

CONCLUSIONS

Patients with antiphospholipid antibodies and a history of previous pregnancy losses have a significant reduction in annexin-V immunostaining on placental villous surfaces, and antiphospholipid antibodies immunoglobulin G can directly decrease levels of villous surface annexin-V on cultured placental villi. Assays of annexin-V in the conditioned media and cell pellets of cultured placental villi suggest that the mechanism for antiphospholipid antibodies-mediated reduced annexin-V surface staining is an inhibition of annexin-V transport to the villous surface rather than displacement by antiphospholipid antibodies from the surface. This antiphospholipid antibodies-induced deficiency of placental surface annexin-V may contribute to the placental thrombosis observed in these patients.

Platelet disorders in pregnancy: implications for mother and fetus

Platelet disorders in pregnancy are not uncommon. Most often, obstetricians are faced with a patient with undiagnosed thrombocytopenia and have the responsibility of deciding if the condition is ITP, gestational thrombocytopenia, or a process related to pre-eclampsia. Correct diagnosis is important because ITP can be associated with fetal thrombocytopenia, making route of delivery important. In alloimmune thrombocytopenia, the mother develops antibodies to a specific platelet antigen present on the fetal platelet but absent on her own. Severe AIT can cause intracranial hemorrhage and have disastrous consequences for the fetus and neonate. In preliminary studies antenatal intravenous gamma globulin therapy has shown promise in preventing the development of intracranial hemorrhage and ameliorating fetal thrombocytopenia. Essential thrombocythemia with a platelet count of greater than 600 x 10(9) platelet/L can occur in pregnancy; therapy consists of antiplatelet aggregating agents such as aspirin, and plateletpheresis. Platelet function disorders can be acquired or inherited. Acquired platelet dysfunction disorders are usually caused by drugs such as aspirin or indomethacin, or by a systemic disease. Hereditary disorders of platelet function can be diagnosed in utero, but cordocentesis may represent an unacceptably high risk. For prenatal diagnosis, other methods, such as chorionic villus sampling or amniocentesis, should be investigated as an alternative to the potentially high risk of cordocentesis.

Increased interleukin-6 concentrations in cervical secretions are associated with preterm delivery

OBJECTIVE

This study sought to determine whether elevated concentrations of interleukin-6 in the cervical and vaginal secretions of patients between 24 and 36 weeks' gestation predicted subsequent preterm delivery and/or identified those preterm deliveries associated with maternal infectious morbidity.

STUDY DESIGN

A cohort study was undertaken with cervical and vaginal samples collected from 161 consenting patients seen at 3- to 4-week intervals between 24 and 36 weeks. Levels of interleukin-6 were measured by immunoassay. Demographic, obstetric, neonatal, and laboratory data were analyzed by Fisher's exact test, Student t test, or Wilcoxon rank sum test, linear and multiple logistic regression, and receiver-operator characteristic curve analysis.

RESULTS

There were 4.2-fold and 3.4-fold increases in maximal cervical and vaginal interleukin-6 concentrations, respectively, among patients with preterm deliveries versus term deliveries. The receiver-operator characteristics curve analysis indicated that a single cervical interleukin-6 value > 250 pg/ml of sample buffer, present between 24 and 36 weeks' gestation, optimally identified patients with subsequent preterm deliveries versus term deliveries (sensitivity 50.0%, 95% confidence interval 33.2% to 66.8%; specificity 85.0%, 95% confidence interval 78.8% to 91.2%; positive predictive value 47.2%, 95% confidence interval 30.9% to 63.5%; negative predictive value 86.4%, 95% confidence interval 80.4% to 92.4%). The optimal vaginal interleukin-6 cutoff value (> 125 pg/ml) proved less sensitive (45.5%, 95% confidence interval 28.5% to 62.4%) but equally specific (86.6%, 95% confidence interval 80.7% to 92.5%). Multiple logistic regression indicated that a cervical interleukin-6 level > 250 pg/ml was an independent predictor of preterm delivery (adjusted odds ratio 4.8, 95% confidence interval 1.7 to 14.3). Cervical interleukin-6 levels did not correlate with cervical change or gestational age at sampling. Among patients delivered preterm there were no differences in the mean white blood cell count on admission or the prevalence of vaginal pathogens, alkaline vaginal pH, chorioamnionitis, or endometritis among patients with cervical interleukin-6 values > 150 or < or = 250 pg/ml.

CONCLUSIONS

Cervical interleukin-6 is a relatively insensitive, although fairly specific discriminator of patients with subsequent preterm deliveries. Among patients delivered preterm, elevated cervical interleukin-6 values are not apparently associated with maternal infectious morbidity.

Retinal and pregnancy outcomes in the presence of diabetic proliferative retinopathy

The objective of this study was to examine the retinal and pregnancy outcomes of pregnancies complicated by advanced diabetic retinopathy. Twenty pregnancies complicated by advanced diabetic retinopathy were included in this retrospective study. The data were analyzed to determine trends in perinatal outcome and to document the ophthalmologic performance. Ophthalmologic management included frequent funduscopic examinations by ophthalmologists of the Yale Retina Center. Among the 20 pregnancies, spontaneous abortion occurred in 2 (10%) and stillbirth in 1 (5%); the remaining 17 (85%) pregnancies culminated in live births at a mean gestational age of 36 weeks (+/- 2.3 SD), with a mean birth weight of 2,620 g (+/- 834 SD). The perinatal survival rate was 94%. Photocoagulation therapy was necessary prior to pregnancy in 45%, during pregnancy in 60% and postpartum in 65%. No pregnancies were terminated because of progressive visual changes that did not respond to photocoagulation therapy. Retinal status should not preclude pregnancy since contemporary methods of management can result in satisfactory retinal and pregnancy outcomes even in the presence of advanced diabetic microvascular disease.

Negative regulation of placental fibronectin expression by glucocorticoids and cyclic adenosine 3',5'-monophosphate

Fibronectin is a ubiquitous extracellular matrix (ECM) protein known to play a critical role in cell adhesion. In the present study we dissected the effects of glucocorticoids and cyclic adenosine 3',5'-monophosphate (cAMP) on FN expression in cultures of cytotrophoblasts isolated from human term placentas to identify compounds which may influence uterine-placental adherence. Based on immunoassay data, relative to controls, glucocorticoid treatment (1-1000 nM) of cytotrophoblasts specifically inhibited media levels of oncofetal FN (i.e., FNs bearing an oncofetal epitope) 65-92%. Treatment of cytotrophoblasts with androgens, estrogens, and progestins (1-1000 nM) did not markedly affect onfFN expression. Corticotropin-releasing hormone (CRH) treatment (200 nM) alone had no effect on levels on onfFN. In combination experiments using 100 nM dexamethasone (DEX), 1000 nM medroxyprogesterone acetate (MPA), 10 nM estradiol (E2) and 200 nM corticotropin-releasing hormone CRH, we observed that DEX treatment also promoted approximately an 85% reduction in media levels of onfFN. This indicated that glucocorticoids profoundly suppress FN expression in the presence of high concentrations of other steroids and pregnancy-associated paracrine effectors. To examine the influence of ECM protein composition on glucocorticoid-mediated suppression of onfFN expression, cells were inoculated on untreated culture wells or on wells coated with FN, laminin, or collagen I. We observed that DEX treatment downregulated onfFN levels 70-85% under each of these conditions, suggesting that glucocorticoid effects on FN expression were not dependent on the presence of an exogenous ECM. Treatment of cytotrophoblasts with 8-bromo-cAMP resulted in a dose-dependent reduction in onfFN expression to 3% of control levels with an EC50 of 150 nM. Based on Northern blotting, treatment of cytotrophoblasts with 100 nM DEX, 1 mM 8-bromo-cAMP, or 2 nM relaxin inhibited steady state levels of FN mRNA approximately 90% relative to controls. Our results suggest that during pregnancy glucocorticoids and compounds that alter intracellular concentrations of cAMP may profoundly suppress FN expression and therefore may have dramatic effects on uterine-placental adherence.

The role of progestationally regulated stromal cell tissue factor and type-1 plasminogen activator inhibitor (PAI-1) in endometrial hemostasis and menstruation

The physiologic mechanisms whereby the human endometrium maintains hemostasis during endovascular trophoblast invasion, yet permits menstrual hemorrhage, are unknown. This paradoxical relationship was investigated by evaluating endometrial expression of tissue factor (TF), the primary initiator of hemostasis, and plasminogen activator inhibitor-1 (PAI-1), the primary inhibitor of fibrinolysis. We observed increased immunostaining for TF and PAI-1 in sections of decidualized stromal cells from luteal phase and gestational endometrium. To determine whether TF and PAI-1 expression are directly linked to decidualization, both endpoints were monitored in a well described in vitro model of decidualization. Thus, confluent stromal cell cultures were exposed to vehicle control, 10(-8) M estradiol (E2), 10(-8) to 10(-6) M medroxyprogesterone acetate (MPA) or both E2 + MPA for 2-24 days in serum-containing or defined media. The progestin enhanced the content of stromal cell-associated immunoreactive and functionally active TF and PAI-1 released into the medium and elevated levels of stromal cell TF and PAI-1 mRNA. While E2 alone was ineffective, it greatly augmented MPA-enhanced TF and PAI-1 protein and mRNA content. Dose-dependent effects on TF and PAI-1 content were observed between 10(-8) to 10(-6) M MPA +/- E2. Similar results were observed for decidual cells derived from first trimester endometrium and cultured in type 1 collagen gels. Following optimal induction of TF and PAI-1 expression by E2 + MPA in stromal cell cultures, removal of these steroids greatly reduced levels of both TF and PAI-1 protein and mRNA within 4 days. These studies suggest a mechanism whereby endometrial hemostasis is maintained during trophoblast invasion yet reduced at the end of nonfertile cycles to permit menses.

Biological mechanisms underlying RU 486 clinical effects: inhibition of endometrial stromal cell tissue factor content

Despite the pronounced hemorrhagic effects of RU 486 administration on luteal phase and early gestational endometrium, no information is available on the effect of RU 486 on endometrial hemostatic potential. The expression of endometrial stromal cell tissue factor (TF), the primary initiator of hemostasis, has been shown to be progestationally regulated in vivo and in vitro. To evaluate the effects of RU 486 on progestin-enhanced TF expression, confluent stromal cell cultures derived from proliferative phase endometria were exposed to vehicle control, 10(-8) mol/L estradiol (E2), 10(-6) mol/L dexamethasone, 10(-7) mol/L medroxyprogesterone acetate (MPA), E2 plus MPA, E2 plus 10(-6) mol/L progesterone (P), or 10(-6) mol/L RU 486 alone or with E2 plus MPA or E2 plus P for 3-4 days. Compared to the vehicle control, E2, dexamethasone, and RU 486 alone had no effect on the content of immunoreactive and functionally active TF protein, whereas MPA increased and the combination of E2 and MPA further increased TF protein content. Similarly, E2 and P enhanced the stromal cell TF content. These progestin effects were blocked by RU 486. Similar results were obtained for steady state TF messenger ribonucleic acid (mRNA) levels. Possible RU 486-mediated reversal of progestin-enhanced stromal cell TF expression was assessed by incubating confluent cultures in E2 plus MPA for 3-10 days to enhance TF content, then washing the cultures and reexposing them to either E2 plus MPA or to RU 486 alone or with E2 plus MPA for 3, 4, or 7 days. Exposure to RU 486 alone or with E2 plus MPA greatly reduced levels of stromal cell TF protein and mRNA expression compared to those in cultures maintained in E2 plus MPA. These findings demonstrate that RU 486 not only blocks but also reverses in vitro progestin-enhanced stromal cell TF protein and mRNA expression, suggesting an additional mechanism for RU 486-induced menses and early abortion.