Pathophysiology of preeclampsia: links with implantation disorders

European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

Review

Pathophysiology of preeclampsia: links with implantation disorders Philippe Merviela,b,*, Lionel Carbillonb,c, Jean-Claude Challierb, Miche`le Rabreaud, Michel Beaufilse, Serge Uzana,b a

Department of Gynecology, Obstetrics and Reproductive Medicine, Hospital Tenon, 4 rue de la Chine, 75020 Paris, France b UPRES Physiology of Implantation and Development (EA 2396), Hospital Tenon, 4 rue de la Chine, 75020 Paris, France c Department of Gynecology and Obstetrics, Hospital Jean-Verdier, Avenue du 14 Juillet, 93143 Bondy, France d Institut d’Histo-Cyto-Pathologie, ZA du Limancet, 114-116 Avenue Leon Blum, 33495 Le Bouscat Cedex, France e Department of Internal Medicine, Hospital Tenon, 4 rue de la Chine, 75020 Paris, France Received 5 May 2003; accepted 22 December 2003

Abstract The phenomenon of implantation anchors the embryo into the uterine wall and produces a hemochorial placenta that maintains the pregnancy and fetal growth. Implantation and placentation are intimately linked and cannot be dissociated either in time or in space. Preeclampsia is characterized by hypertension and proteinuria. It is secondary to an anomaly of the invasion of the uterine spiral arteries by extra-villous cytotrophoblast cells, associated with local disruptions of vascular tone, of immunological balance and inflammatory status, and sometimes with genetic predispositions. Preeclampsia is a disease of early pregnancy, a form of incomplete spontaneous abortion, but is expressed late in pregnancy. Aspirin may play a favorable role in implantation which is related to the genesis of preeclampsia and some cases of intra-uterine growth restriction. The most important points in obtaining a preventive effect from low-dose aspirin during the pregnancy are early treatment (before 13 weeks of gestation) and the prescription of a sufficient dose (more than 100 mg per day). # 2004 Published by Elsevier Ireland Ltd. Keywords: Aspirin; Extra-villous cytotrophoblast; Implantation; Preeclampsia

1. Introduction Implantation of the human embryo leads to the invasion by the extraembryonic trophoblast of the endometrium and to the colonization of the uterine arteries. This phenomenon enables the embryo to be anchored in the uterine wall and thus makes possible the maintenance of the pregnancy and fetal growth via the placenta. For this semi-allogenic graft to occur, the endometrium must first undergo structural and biochemical modifications, called decidualization. An implantation window therefore occurs, the result of the synchronization between embryonic development and endometrial maturity. The interactions between the embryo and various components of the decidua then lead to implantation in the strict sense of the term. Finally, the trophoblastic cells transform the uterine vascular system;

*

Corresponding author. Tel.: þ33-1-56-01-68-76; fax: þ33-1-56-01-60-62. E-mail address: [email protected] (P. Merviel). 0301-2115/$ – see front matter # 2004 Published by Elsevier Ireland Ltd. doi:10.1016/j.ejogrb.2003.12.030

this change, together with the action of vasomotor factors, will adequately nourish the fetal-placental unit [1]. Accordingly, implantation and placentation are actually two aspects of the same phenomenon, in different places and moments of the pregnancy [2]. Preeclampsia is characterized by hypertension (>140/ 90 mmHg) and proteinuria (>0.3 g/l). It is secondary to an anomaly of the invasion of the uterine spiral arteries by cytotrophoblast cells. It may be considered finally as one form, incomplete, of spontaneous abortion, in view of the many similarities between them (and their frequent association in the same woman). Like spontaneous abortion, preeclampsia is a disease of early pregnancy and even precedes implantation, beginning at the follicular phase. It is, however, only expressed late in pregnancy (from the second-trimester), after the activation of a cascade of events has finally led to the appearance of clinical signs. The aim of this review is to reconsider the current theories about preeclampsia, by looking at it together with the mechanisms of embryo implantation, and to try to identify the therapeutic implications.

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

2. An anomaly of the extra-villous cytotrophoblastic invasion of the uterine spiral arteries The extra-villous cytotrophoblast cells (EVCT) must invade the decidua before they modify the walls of the spiral arteries [3,4]. The trophoblast behaves like a ‘‘pseudotumor’’ invading the endometrium, which tolerates it in a controlled way. Unlike an invasive tumor or an inflammatory reaction, however, implantation is an invasion controlled in time and space. Any anomaly between the factors promoting and those limiting this invasion may cause a pregnancyrelated disease. Preeclampsia is one of these [5]. To invade the decidua, the trophoblast cells need both to recognize the different components of the membrane and of the extracellular matrix ((ECM) integrins, cadherin) and break them down (metalloproteases). To control this invasion, the endometrium modifies the composition of its extracellular matrix (ECM), secretes transforming growth factor (TGF) b and tissue metalloprotease inhibitors (TIMP) [6]. Moreover, the decidua is colonized by immune system cells (NK cells, lymphocytes and macrophages) that are responsible for the local production of cytokines that promote or inhibit the trophoblastic invasion. Except during pregnancy, the ECM is composed of collagens I, III, Vand VI, fibronectin and periglandular tenascin deposits. During decidualization, the endometrial stromal (decidual) cells produce a pericellular matrix composed of collagen IV, laminin and heparan sulphate; substantial hydration of the stroma occurs at the same time. This change in the composition of the ECM and its hydration make it easier for the EVCT to invade the decidua. The modified ECM also establishes close contacts with the lymphoid cells present in the decidua, thereby increasing the cellular interaction between trophoblastic and lymphoid cells. The adhesion of EVCT to ECM components (collagen IV, laminin, proteoglycans, heparan sulphate, entactin and fibronectin) requires the intervention of receptors on the plasmic membrane that enable the cell to identify and then bind to them: these receptors are the integrins and cadherins. Integrins are heterodimeric glycoproteins with two subunits: a and b. Their combination forms many integrins that bind to various components of the ECM. A ‘‘switch’’ changes the profile of the integrins as the cytotrophoblast cells change from proliferative (at the base of anchoring villi) to interstitial (in the deepest portion of the decidua and spiral arteries). Thus, proliferative EVCT expresses only integrin a6b4 (receptor for laminin, a component of basement membrane), then as they migrate (becoming invasive) they acquire the ability to express integrin a5b1 (fibronectin receptor). When the cytotrophoblastic cells become interstitial, integrins a1b1 (laminin and type I and IV collagen receptors), avb1 and avb3 (vitronectin receptor) also appear [7]. Because the ECM is organized in a tridimensional network that prevents passive cell migration, the adhesion of trophoblast cells to the ECM components is necessary but

135

not sufficient to guarantee their invasiveness. Accordingly, the trophoblast cell must be able to proteolyse the ECM conponents. Matrix metalloproteases (MMP) are endoproteinases that require the presence of Ca2þ and Zn2þ ions; they include 13 members in three families: collagenases (MMP 1, 8, 13), which break down types I and III collagens; gelatinases A (MMP 2) and B (MMP 9), which break down gelatin, collagen IV and elastin; and the stromelysins (MMP 3, 7, 10 and 11), which have a broader spectrum. These enzymes are regulated by their activation level (most are secreted as the inactive forms, zymogens) and by the presence of specific tissue inhibitors, the TIMPs (tissue inhibitors of metalloproteinases, secreted by the decidua), which block the active site of the enzyme. The trophoblast cells secrete some MMPs during the first-trimester. These include gelatinase B, which plays a primordial role in the invasion. TIMP 1 expression, which is highest at term, balances gelatinase B activity and thereby controls the invasiveness of the trophoblast. In addition, integrins can modulate MMP expression. TGF b is a growth factor expressed at the feto-maternal interface by the decidua, from the first-trimester through term. It inhibits the proliferation and invasion of the trophoblast. TGF b1 is essentially expressed by the villi, and TGF b2 by the decidua. TGF b1 promotes ECM formation, in particular collagen and fibronectin, inhibits plasminogen activator production, induces TIMP 1 expression, and reduces EVCT migration by overexpressing a5b1. This overexpression makes the EVCT more adherent to the ECM and activates the differentiation of the cytotrophoblast into non-invasive syncytiotrophoblast. This endometrial maturation is also linked to the growth of the spiral arteries (branches of the uterine arteries), which will then carry maternal blood toward the intervillous spaces of the placenta. The growth and structure of the spiral arteries depend on ovarian hormonal secretions [8]. Under the influence of estrogens, their diameter increases as they grow longer and become progressively twisted [9]. This endothelial proliferation continues during the luteal phase and the first weeks of gestation. Growth factors also play a role in neoangiogenesis: fibroblast growth factor (FGF) b, a powerful angiogenic factor, is increased by estradiol and inhibited by progesterone; vascular endothelial growth factor (VEGF), stimulated by estrogens and hypoxia, is mitogenic for endothelial cells and increases vascular permeability; and platelet-derived growth factor (PDGF), which contributes to angiogenesis and to the growth of smooth muscle cells. Finally, other factors, including TGF b, tumor necrosis factor (TNF) a, interleukin (IL)-1 and IL-6, also participate in this angiogenesis. This neoangiogenesis can be disrupted by conditions that are accompanied by microangiopathy, such as insulin-dependent or gestational diabetes or chronic or gestational hypertension after a kidney transplant. The establishment of the uteroplacental vascular system begins with the invasion of the maternal decidua by the

136

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

EVCT. Two successive and interdependent phenomena are necessary to accomplish the complete transformation of the uterine spiral arteries by the trophoblastic cells [10,11]:

3. The first vascular invasion by the cytotrophoblast During the first-trimester (from 8 until 12 weeks, approximately), the EVCT sheathes the outer wall of the decidual capillaries and the intra-endometrial branches of the uterine spiral arteries, thereby creating a trophoblastic shell around these vessels. The trophoblast cells invade from the exterior towards the inner capillary walls, where they are organized in loose but interrelated clusters of trophoblastic shell. These intravascular plugs obstruct almost all of the decidual capillaries. The plugs are more of a filter than a barrier [12]. Nonetheless the permeability of these plugs enables plasma, with some maternal red blood cells, to diffuse past them towards pools of blood that result from the vascular invasion and which are the future intervillous spaces [13]. 3.1. Oxygen partial pressure This anatomical phenomenon results in the increase of oxygen partial pressure (PO2 ) upstream from these plugs and its decrease downstream from them. The increased PO2 observed upstream diminishes lipid peroxidation in the endothelial cells of the intramyometrial spiral arteries, which in turn is translated into an increase in prostacyclin (so called prostaglandin I2 or PGI2) and a diminution of thromboxan A2 (TXA2), and the consequent increase in the vasodilatation of these vessels. The increase in PO2 also diminishes production of endothelin-1 (ET-1), which is vasoconstrictive. Downstream from these plugs, the reduced PO2 works toward guaranteeing the best possible environment for the embryo’s organogenesis. Inversely, the high pressure upstream from the plugs increases the release of nitric oxide (NO) by the endothelins of the myometrial spiral arteries and thus helps to further increase local vasodilatation [14,15] showed that placental PO2 , while lower than endometrial PO2 during the first-trimester, but between 8 and 12 weeks, PO2 increases progressively. Moreover, because first-trimester embryos lack defense systems against oxygen free radicals, this low PO2 level protects their tissues against the harmful effects of oxygen [16,17]. Finally, embryos at this term have embryonic hemoglobin, which has a greater affinity for oxygen in low partial pressure conditions, such as those encountered in plasma. 3.2. Hemostasis in the vascular spaces The existence of plugs in the endometrial capillaries should theoretically be accompanied by a stacking of maternal red blood cells upstream and the appearance

of extensive thrombi. This is not observed in vivo, however, because of systems that regulate local hemostasis: thrombomodulin, tissue factor and plasminogen activator. These local factors work together to ensure that blood flows through the uterine spiral arteries and to prevent extravasation following the EVCT invasion. Thrombomodulin (TM) is secreted by the endothelial cells and activates protein C, which has a proteolytic activity and inhibits the formation of blood clots. By its anticoagulant action, TM prevents the formation of intravascular thromboses. Tissue factor (TF) is a pro-coagulant factor located on the membranes of endometrial stromal cells (during the secretory phase) and of perivascular decidual cells. Stimulable by progesterone, it contributes to the perivascular endometrial hemostasis necessary after the EVCT vascular invasion (by synthesizing thrombin, which transforms fibrinogen into fibrin). At the same time, fibrinolysis in the decidua is inhibited by the activation of plasminogen activator inhibitor type 1 (PAI-1) and the diminution of tissue type and urokinase-type plasminogen activators (tPA, uPA). Some authors have hypothesized that arterial–venous shunts may exist upstream from the plugs, which might explain the diminution of maternal blood intake at the plugs and thus the absence of thrombi. This has never been shown clearly during pregnancy. Remember that the uterine spiral arteries are connected in parallel to the uterine radiate arteries and are not the final branches of the latter. Accordingly, supplemental blood reaching the placenta during pregnancy (in connection with the opening of vascular space and the increase in the maternal heart rate) is distributed evenly throughout all the spiral arteries; the blood influx is thus moderated. In these conditions arterial–venous shunts do not seem strictly necessary. Recently, Schaaps [18] pointed out the importance of the myometrial arterial network set up from the beginning of pregnancy. Ultrasound studies show the significant diminution of this network, parallel to the uterine axis, in cases of preeclampsia. It appears to cushion the mother’s bouts of hypertension by functioning as a vascular valve, but also ensures better distribution of the blood flow and adequate nutrition for fetal growth. 3.3. Hemodynamic protection of the embryo The plugs in the decidual capillaries provide hemodynamic protection for the embryo, by preventing strong vascular pressure in the blood lakes. The increased sinuosity of the spiral arteries at the beginning of pregnancy (which have a damper effect on maternal blood flow) and the extraembryonic coelom also play a role in this protection. In cases of spontaneous abortions, it is frequently observed that due to the absence of intravascular plugs, the mother’s blood has flooded the intervillous lakes [19]. The leads to the cessation of the embryonic-placental circulation and the death of the embryo.

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

3.4. Early maternal–embryonic exchanges, followed by embryonic–maternal exchanges At the beginning of pregnancy, the embryo evacuates its wastes towards the yolk sac. There is no embryonic circulation and the pressure in the villous capillaries is less than that in the blood lakes. The exchanges therefore occur solely towards the embryo from the mother (transfer of nutrients and oxygen). From 4 to 5 weeks, fetal heart activity begins and pressure in the villous capillaries increases, thereby enabling exchanges from the embryo to the mother. The plugs in the maternal vessels protect these embryonic– maternal exchanges, for a substantial increase in the blood lake pressure would interrupt them and cause a vascular collapse of the villous vessels (and lead the embryo to stop thriving). The first trophoblastic invasion can be observed from 5 weeks in the intra-endometrial arteries. Between 5 and 8 weeks, the plugs obstruct the vascular lumina almost completely and prevent the passage of maternal blood to the intervillous lakes; they then progressively disaggregate from week, 8–13.

4. The second vascular invasion by the cytotrophoblast The second trophoblastic invasion of the intramyometrial spiral arteries thus occurs between 13 and 18 weeks, at which time it is totally completed. Because these are continuous phenomena, an intramyometrial vascular invasion can sometimes be observed before 13 weeks [20]. Starting at 8 weeks and through 13 weeks, the trophoblastic shell surrounding the decidual spiral arteries becomes discontinuous, persisting only at the anchoring villi of the placenta; this induces the progressive release of the intravascular plugs. A portion of the trophoblastic cells from the plugs will move backwards to colonize the inner wall of the intramyometrial spiral arteries and then penetrate into the thickness of the vascular wall [21]. This intraparietal encroachment causes the endothelial cells and the smooth muscle cells of the tunica media and the internal elastic layer to disappear progressively [22]. The latter is replaced by a fibrin deposit that deprives the vessels of their contractility. The trophoblast cells progressively develop an endothelial phenotype because of the switch from E- to VE-cadherin and the acquisition of endothelial cell molecules such as VCAM-1 and PECAM-1 [23,24]. Decreased resistance in the uterine arteries occurs and starts the continuous blood flow through the intervillous spaces that is necessary to fetal growth during the second and thirdtrimesters. Intervillous blood flow increases at around 10– 12 weeks of gestation and results in exposure of the trophoblast to increased oxygen tension (PO2 ) [25]. Expression of hypoxia inductible factor-1 (HIF-1) and TGFb3 (an inhibitor of early trophoblast differentiation) is high in early pregnancy and falls at around 10 weeks of gestation when

137

placental PO2 levels are believed to increase. When the oxygen tension fails to increase, HIF-1 and TGFb3 expression remain high, resulting in shallow trophoblast invasion and predisposing the pregnancy to preecclampsia [26]. Effective fetal–maternal interactions during early placentation are critical for a successful pregnancy [25]. These anatomical and hemodynamic processes can be seen in a doppler study by the disappearance of the uterine artery notch, by the increased diastolic flow through these arteries, and by the blood flow that appears in the intervillous spaces. This progressive replacement of the collagen and elastin frame and the transformation of the intramyometrial spiral arteries is most often completed at 18 weeks, but sometimes requires several additional weeks (notches not infrequently disappear between the 22- and 26-week doppler ultrasounds). The defective development of the myometrial arterial network, as described by Schaaps [18], also plays a role in the persistence of the notches in preeclampsia. At the same time, the endothelial cells detach from the uterine veins and arteries, proliferate, and migrate towards the internal face of the intervillous spaces, thus separating the fetal circulation from the maternal blood by a double cellular layer, trophoblastic and endothelial. During this period, the fetus acquires fetal hemoglobin (HbF) with oxygen-uptake capacities in line with its greater needs for growth.

5. Vasomotor factors Pregnancy is associated with a diminution in blood pressure, a drop in systemic and uterine vascular resistance and a reduced response to various vasopressor (vasoconstrictive) agents. The uteroplacental vessels are subjected to various factors that regulate their vascular tone. Even before implantation, the sexual hormones regulate the balance in the uterus between vasoconstricting and vasodilating agents. During the follicular phase, NO secretion (together with 17ß-estradiol) increases and then diminishes during the luteal phase. Similarly, endothelin-1 (ET-1) also diminishes during the follicular phase. Estradiol has an indirect vasodilating action (NO, ET-1 and prostaglandins) and inhibits vasoconstrictors, while progesterone functions as an estradiol antagonist. Nonetheless progesterone alone or combined with estrogens inhibits the vasoconstrictive response to angiotensin II (AII) [27]. Schematically, uterine vascular tone is regulated principally by two opposing vasomotor systems: a vasoconstrictor system (endothelin/enkephalinase) and a vasodilator system (nitric oxide/guanylate cyclase). These agents are involved in the placental regulation of the vascular flow: they set up a local balance that enables adequate blood intake. Some diseases, such as the vascular complications of hypertension, may be due to the deregulation of these systems [28]. It is now generally agreed that endothelial cells, together with endocrine and nervous system factors, participate in

138

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

regulating vasomotor tone [29,30]. They also ensure the continuous inhibition of platelet aggregation. 5.1. Endothelins Endothelin (ET) is one of the most powerful vasoconstrictive factors in the fetal-placental circulation. It is a 21amino-acid peptide, existing in three isoforms (ET-1, ET-2 and ET-3); its target is the spiral artery endothelial cell. It is synthesized through post-translational maturation: preproET-1 is transformed into proET-1, then into big ET-1 and finally into ET-1, through endothelin conversion enzyme (ECE). Its receptors (ETA, ETB) are coupled with a protein G: ET-1 coupled with ETA causes vasoconstriction via Ca2þ and phospholipase C (PLC) (principal action); and ET-1 coupled with ETB is vasodilating through the action of NO and PGI2 [31]. Endothelin is broken down by the enzyme enkephalinase, whose production is induced by progesterone. Thus, in the middle of the luteal phase, at the progesterone peak, the ratio of endothelins to enkephalinase favors the breakdown of the endothelins and therefore contributes to the absence of vasoconstriction of the implantation site vessels. Moreover, the hypothesis of the involvement of ET1 in disorders such as intra-uterine growth retardation and preeclampsia seems to be confirmed by the increase in the concentration of immunoreactive endothelin in both the umbilical vessels and the maternal blood in these cases and by the increase in the expression of the ET-1 gene in the placental villi in preeclampsia [32,33]. 5.2. Nitric oxide When its integrity is intact, endothelin produces a highly labile vasodilator product that allows local up-regulation of the blood flow [34]. This factor, first called endothelial derived relaxing factor (EDRF), has since been identified as nitric oxide [35]. This gas, diffusible through cell membranes, penetrates the smooth muscle cells surrounding the vessels and there reaches its target, guanylate cyclase. This enzyme is required for the formation of cGMP, which activates the intracellular protein kinases and keeps the smooth muscle cells relaxed. NO is synthesized from Larginine in the presence of NO synthase (NOS), which exists in three isoforms. Two, endothelial (eNOS) and neuronal, are constitutive and require calcium for their action. Several stimuli (acetylcholine, bradykinin, serotonin, ATP, vascular shearing forces) increase eNOS activity, ensuring self-regulation of the blood flow [36]. In women, eNOS has been found in the uterine and umbilical arteries, chorionic vessels and placental villi [37]; during normal pregnancy, urinary excretion of cGMP, nitrates (NO2) and nitrites (NO3) (both NO metabolites) increases. NO relaxes the vascular smooth muscle cells by activating cGMP [38]. NO vasodilatation depending on the spiral arteries appears to precede the trophoblastic invasion. NO is produced by the endothelial cells after stimulation of the receptors (associated with a

protein G) by acetylcholine and bradykinin [39]. Modifications of the sex hormones disrupt NO secretion. Pregnancy is associated with increased NO production. Gude et al. [40] showed that administration of inhibitors of either NOS or guanylate cyclase amplifies the vasoconstriction observed with ET-1 and TXA2. Other mechanisms are probably also involved: sensitivity and affinity of the endothelial receptors, effects on intermediate levels of the NO chain, modulation of proteins G. NO is a powerful vasodilator released by the endothelial cells. Experiments performed with glyceryl dinitrate, an NO donor, and prostacyclin show that administration of the former during the first-trimester diminishes uterine artery vascular resistance while the latter has no effect. Ramsay et al. [41] administered glyceryl trinitrate (GTN) between 8 and 10 weeks: a decrease in peripheral resistance and an increase in blood flow were then observed. Similarly, GTN administered between 24 and 26 weeks to women with an elevated uterine artery resistance index and a bilateral notch decreased the index. Moreover, there were no modifications of the doppler umbilical index or of maternal blood pressure or pulse. Thaler et al. [42] administered an NO donor (isosorbide dinitrate) to women between 17 and 24 weeks and observed a diminution of the resistance index for both the uterine and umbilical arteries. 5.3. Vascular effects of steroid hormones [43,44] Estrogens (in particular estradiol) have a direct (nongenomic) and indirect (by protein synthesis) vasodilating effect on the uterine arteries, in the latter case by the intermediary of ET-1, NO and prostaglandins [45]. This vasodilation seems local, because it does not occur in the renal or mesenteric arteries when estrogens are administered. Moreover this effect is independent of the adrenergic, histaminic and cholinergic receptor system. Estrogens facilitate Ca2þ entry into the endothelial cell (either directly or by the intermediary of acetylcholine receptors); this leads to NO synthesis by stimulation of NO synthase and then to a relaxant effect on the muscle via cGMP [46]. Estradiol affects endothelin synthesis by inhibiting ECE and its transformation of big ET-1 into ET-1. ET-1 synthesis is thus reduced [47]. Progesterone modulates the effect of vasopressin on the uterine arteries; the effect depends on the progesterone concentration: it is vasodilating alone or combined with micromolar estradiol doses, but vasoconstrictive at lower doses (1–10 nM). 5.4. Prostaglandins Other factors also participate in the regulation of vascular tone. Prostaglandins are molecules derived from the metabolism of arachidonic acid; they have auto/paracrine activity and are thus either vasoconstrictive (PGF2a, TXA2) via an increase in intracellular Ca2þ and protein kinase C (PKC)

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

stimulation, or vasodilating (PGE2, PGD2, PGI2), with action mediated by adenylate cyclase and cAMP formation [48]. Prostaglandins also modulate the action of some hormones, amplifying or inhibiting their effects. In particular, PGI2 diminishes the sensitivity of maternal vessels to angiotensin II; it accounts for the resistance observed to the vasoconstrictive action of AII during normal pregnancy. Moreover, some agents, such as interleukin 1, gonadotrophin releasing hormone (GnRH) and corticotropin releasing hormone (CRH), stimulate prostaglandin synthesis [49]. During a normal pregnancy, the ratio of prostacyclin to thromboxan increases progressively throughout the pregnancy. Prostacyclin is produced by vascular endothelin and trophoblastic cells. Its action is vasodilating (relaxes vessels contracted by angiotensin II), muscle-relaxing, and antiaggregating. Thromboxan A2 is synthesized by the platelets and is a powerful vasoconstrictor that also promotes platelet aggregation and uterine contractility. During a normal pregnancy, the intravascular plugs formed by the extra-villous cytotrophoblast cells at the junction between the spiral arteries and the intervillous spaces inhibit membrane lipid peroxidation and diminish TXA2 levels [50]. The hCG secreted by the trophoblastic cells also participates in regulating the vascular bed, by promoting PGI2 synthesis and diminishing that of PGE2 and TXA2. It also stimulates placental production of prorenin. Accordingly, the predominance of PGI2 action compared with that of TXA2 (increase in the PGI2/TXA2 ratio) causes vasodilation and decreases vascular resistance. The absence of these plugs in preeclampsia, caused by a defect in the EVCT vascular invasion, reduces the PGI2/TXA2 ratio (decreases PGI2 synthesis and increases that of TXA2), leading to an augmentation in both the sensitivity of maternal vessels to angiotensin II and vascular resistance, to the absence of vasodilatation, and to the formation of localized microthromboses [51]. Moreover, the vasomotor action of the prostaglandins can be modified by vascular alterations (fibrin deposits, endothelial lesions, etc.). Studies show these vascular lesions in the placentas of 57% of intra-uterine growth retardation (IUGR) cases and 74% of women with hypertension. They thus further reduce the perfusion of the intervillous spaces. 5.5. ACTH and CRF During pregnancy, adrenocorticotropin hormone (ACTH) is present simultaneously in the maternal and fetal circulations [52]; it enters the former from the syncytiotrophoblast and the latter from the fetal pituitary gland. Its vasodilating action appears to be mediated by either mastocyte degranulation or the release of histamine or progesterone (all vasodilating). ACTH is a more powerful vasodilator (187) than PGI2 [53]. Its action is not mediated by NO, cGMP or a prostaglandin, but involves a specific receptor. In pregnancies complicated by preeclampsia or IUGR, ACTH levels in the umbilical artery

139

are elevated, in response to stress and to the hypoxemia that follows the diminished blood flow. Corticotropin releasing factor (CRF) is a powerful vasodilator (50/PGI2); its effect on the vascular musculature is mediated by the NO–cGMP pair [54]. It increases regularly in the fetal-placental circulation from 7 to 8 weeks until term. It is a 41-amino-acid peptide that stimulates the secretion of ACTH and of ß-endorphins in the placenta and the maternal and fetal pituitary glands, as well as the production of prostaglandins in the decidua and membranes. In vitro, PGE2, PGF2, norepinephrine, acetylcholine, oxytocin, IL-1, angiotensin II, arginine vasopressin, neuropeptide Y (NPY) and the glucocorticoids stimulate CRF secretion, while progesterone and NO inhibit it. There is, nonetheless, no synergy during pregnancy between arginine vasopressin and CRF as to ACTH secretion, and dexamethasone does not inhibit CRF. Its receptor has been found on syncytiotrophoblastic and endothelial cells (eNOS located in the latter). Its level is higher during pregnancies complicated by hypertension or IUGR [55], in response to a diminished placental vascular flow, but its vasodilating action is reduced because of the endothelial alterations [56–58]. 5.6. Neuropeptides The vasomotor peptides such as vaso-intestinal peptide (VIP), substance P and calcitonin gene-related peptide (CGRP), which are synthesized from nerve fibers near the uterine vessels, also help regulate the uteroplacental blood flow. Their action may be mediated by the NO– cGMP pair. VIP and substance P (SP) exert a vasodilating action on uterine vessels contracted with vasopressin (no effect on basal parietal tension) and on villous vessels contracted with PGF2a. This action is not mediated by acetylcholine, adrenaline or prostaglandin receptors. In humans, CGRP appears to be a growth factor for the endothelial cells of the umbilical vein and an angiogenic factor in ischemic conditions [59,60]. It is a 37-amino-acid peptide derived from the calcitonin/CGRP gene (chromosome 11) by alternative mRNA splicing. The concentration of immunoreactive CGRP increases during pregnancy (multiplied by 2–3 at term) and diminishes during the first 7 days postpartum [61]. The uterine arteries also appear to be more sensitive to the effect of CGRP during pregnancy [62]. It is secreted by nonadrenergic–noncholinergic A (@) and C nerve fibers located at the artery adventitia-media junction, in association with substance P. Its decay is associated with enkephalinase, competing with substance P; this explains the CGRP/substance P potentiation. Two types of receptors are present throughout all of CGRP 1R and CGRP 2R; they contain 7 transmembrane passages coupled with protein G. Its endothelin-independent action binds onto the receptors of the smooth muscle cells, activates adenylate cyclase and produces cAMP as well as blocking calcium L channels, thereby reducing the intracellular Ca2þ concentration [63].

140

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

The result is powerful vasodilation of the uterine arteries. CGRP also acts via NOS, increasing vasodilatation by releasing cGMP. ß Adrenergic innervation diminishes during pregnancy (at term it is only 2% of that in non-pregnant women), while CGRP innervation increases: both factors help maintain uterine vasodilatation [64]. Experiments in rats show that CGRP may inhibit the hypertension induced by L-NAME (analogue of L-arginine, inhibiting NOS) administration and reduce fetal mortality associated with hypertensive syndromes, but it does not improve birth weight [65]. The explanation is that CGRP does not cross the placenta; it corrects the NO deficit on the mother’s side but not in the fetus. 5.7. The effect of calcium The beneficial effects of calcium supplementation in reducing the risk of preeclampsia suggests that Ca2þ plays a role [66–68]. The platelets of a woman with preeclampsia, compared to those of a woman in a normal pregnancy, contain more intracellular calcium [69], probably related to an increase in the voltage-dependent calcium channel activity [70]. Pregnancy and increased blood estradiol levels affect these channels and reduce the concentration of intracellular Ca2þ.

6. Case of preeclampsia In preeclampsia, decidual resistance, more powerful than the trophoblastic invasion, prevents the EVCT from reaching the spiral arteries. Placentas of women with preeclampsia express lower levels of matrix metalloproteinase (MMP) 9, human lymphocyte antigen (HLA) G, placental lactogen hormone (HPL) and a1ß1 than those of women with normal pregnancies; the integrin a4ß5 level is stable and that of a1ß1 increases [71]. Moreover, the switch from E-cadherin to VE-cadherin does not occur, nor are VCAM-1 and PECAM-1 produced [72]. These phenomena testify that these cytotrophoblasts have lost their capacity for deep invasion. Invasive EVCT dedifferentiates into syncytium (giant cells) that thereby lose their penetrating power; the increase in giant cells thus expresses this initial impulse [73,74]. A related finding is the higher frequency of preeclampsia and intra-uterine growth restriction among nulliparas (75% of the cases): this may be associated with the fact that the arteries colonized in a first pregnancy can be invaded more easily during subsequent pregnancies. The role of the decidual natural killer (NK) cells may explain why a subsequent invasion is facilitated; they can be thought of as the endometrial memory of paternal antibodies. Preeclampsia was similar in nulliparas and in the multiparas who had changed partners (3.2 and 3%), but lower among multiparas with the same partner (1.9%). A similar finding is noted in pregnancies after ovocyte donation, sperm donation, or a long period of contraceptive use. Accordingly we must

consider preeclampsia more as a disease of primipaternity than primigravidity [75]. In preeclampsia, the second trophoblastic invasion either does not occur or is incomplete because of the lack of intravascular plugs. This is expressed by the persistence of uterine vasoconstriction. Blood intake into the intervillous spaces is diminished and fetal growth retardation ensues. The downstream consequence of this vasoconstriction is hypoxia, with an increase in lipid peroxidation [76,77] and in the TXA2/PGI2 ratios [78], both of which accentuate vasoconstriction and platelet aggregation. Thromboses and disseminated fibrin deposits are usually found in the placenta in this disease [79]. Moreover, downstream hypoxia increases ET-1 production and diminishes that of NO (also related to the reduction of the mechanical force of the artery wall). Evidence of intravascular plugs is not found in preeclampsia. This absence explains why PO2 in the decidual spiral arteries is on the whole lower than that observed upstream from these plugs: this results in increased lipid peroxidation [80] and a decreased PGI2/TXA2 ratio, with vasoconstriction and platelet aggregation. This diminution of PO2 also causes an increase in ET-1 and (in combination with the diminution of the mechanical forces on the vascular wall) a decrease in NO in the myometrial and decidual spiral arteries. The stimulant effect of ET-1 on the release of NO partly compensates for this NO decrease. Moreover the absence of plugs (resulting in relative high pressure in the blood lakes) is responsible for the increase in the rate of spontaneous abortions and the fetal ‘‘failure to thrive’’ observed in patients at risk of preeclampsia [81]. Preeclampsia is thus characterised by an increase in systemic vascular resistance and in vascular reactivity and by a change in the distribution of the pelvic blood flow that precedes the onset of hypertension [82]. All of these suggest a dysregulation in the normal vasomotor factors of pregnancy. During preeclampsia, the sympathetic system/normal pregnancy is activated [83]. Some older studies suggest that preeclampsia is associated with low serum estradiol levels, which are also found among women who live at altitudes higher than 3000 m; they have an increased risk of preeclampsia and lower estradiol levels before the onset of preeclampsia [84]. When preeclampsia occurs, modifications of the vasomotor imbalance and of the Ca2þ ion homeostasis can be observed. The increase in the vasoconstrictor/vasodilator ratio is proportional to the severity of the syndrome, and NO metabolite levels also diminish [85]. Rats treated with an NO inhibitor, nitro-L-arginine methyl ester (L-NAME) develop hypertension and proteinuria; the growth of their fetuses is retarded [86,87]. Grunewald et al. [88] administered GTN by IV to women with preeclampsia and observed that both systolic and diastolic blood pressure fell significantly, and that the umbilical resistance index also fell. Lees et al. [89] treated these women with an NO donor (Snitrosoglutathione, GSNO) and observed decreases in mean

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

blood pressure, platelet activation and uterine resistance. The NO donors thus seem therapeutically interesting, because they reduce BP, inhibit platelet activation and improve uterine and fetal hemodynamics [90]. Moreover GTN increases the blood flow in the uterine arteries without modifying either heart rate or systemic blood pressure, as occurs during the second-trimester; this suggests that this effect on uterine artery vasodilatation may be a priority for NO. Finally chronic hypoxia, which results from a placentation defect, can induce the transcription of some genes [16], including that of ET-1 (vasoconstrictor), angiotensin conversion enzyme (hypertensive), plasminogen activator [91] (stimulating formation of active TGF ß, which inhibits the EVCT invasion) and cyclooxygenase-1 (COX-1, an enzyme involved in prostaglandin production).

7. Disruption of the uterine immune balance In theory, an implanting embryo could be the target of various types of immunological aggression: standard cellmediated lysis, lysis by cytotoxic antibodies associated with complement and NK cell lysis. During pregnancy, the embryo is protected from these dangers by its own early antigenicity, by cytokine secretion and local immunosuppressors, intrinsic resistance to cellular lysis, and the MCPDAF system. Two phases occur: an initial maternal immunological reaction to the allograft, followed by the development of allogenic tolerance [92].

141

metrial or trophoblast cells. Beginning in the preimplantation period, lymphocytes pour into the uterus, where the levels of macrophage colony stimulating factor (M CSF or CSF 1), granular cell and macrophage colony timulating factor (GM CSF), granular cell colony stimulating factor (G CSF), tumor necrosis factor (TNFa), interleukin-6 (IL-6) and IL-1 increase substantially; they indicate the presence of cellular and humoral inflammatory reactions. 9.1. Cytokines involved in the inflammatory reaction IL-1 is detected in the decidua, and its receptors in the epithelial cells of the endometrium. The embryo also secretes IL-1, which suggests that it controls its own development via interferon (IFN), which is a trophic factor for the trophoblast and for which the embryo has receptors. It appears to stimulate the production of IFN, IL-6 and PGE2 and to enable the expression of class II HLA antigens (HLA-DR) in the uterus. 9.2. Cytokines involved in trophoblast development CSF 1, GM-CSF and G-CSF are secreted in large quantities in the decidua, and their receptors are expressed by the early trophoblast. CSF 1 favors the growth of the trophoblast. GM CSF plays a direct role in the attachment and growth of the trophoblast and in the survival of the embryo. This factor is secreted by large granular lymphocytes (LGL) CD 56þ, present in the decidua. G CSF receptors, necessary for implantation, are present in the trophoblast and the decidua. 9.3. Immunosuppressive cytokines

8. Antigenicity of the early embryo The trophoblastic cells in contact with maternal blood do not possess class I or II HLA antigens; in this, they are unlike the extra-villous cytotrophoblast cells at the top of the trophoblastic columns or in the spiral arteries, both of which are rich in class I HLA (HLA A–C). This HLA class presents peptide autologous antigens to lymphocytes T (LT). Three other class I antigens have also been identified: HLA E, F and G. While HLA E and F have been found in numerous tissues, both fetal and adult, HLA G is expressed solely in the extra-villous cytotrophoblast cells, at the fetal–maternal interface (where no other class I or II antigens are found). This particular expression of HLA G plays a role in the tolerance of the semi-allogenic graft that is the pregnancy (NK response is inhibited by the disguise of trophoblastic cells as self cells).

9. Cytokine and local immunosuppressor secretion Cytokines may be produced by the secretion of leukocytes infiltrated into the decidua or may be synthesized by endo-

IL-6 inhibits expression of IL-2 receptors, which would otherwise support the proliferation of cytotoxic cells (lymphocytes T and NK cells), of B lymphocytes and of antibody-dependent cell-mediated cytotoxicity. IL-10 plays a key role in preventing embryo resorption by antagonising IFNg and TNFa.

10. Immunosuppressive hormones Progesterone also participates in maintaining the semiallograft, because of its local anti-inflammatory activity: it can inhibit phagocytosis and lymphocyte proliferation in the uterus, either directly by blocking the LT CD4þ activity and proliferation induced by IL-1, or indirectly, by inducing the release of two immunosuppressor factors by the lymphocytes: T-suppressor-induced factor (TSIF), whose antibodies can abort a pregnancy in mice, and progesterone-induced blocking factor (PIBF), which blocks the lysis of embryo fibroblasts by NK cells and inhibits mixed lymphocyte reactions by preventing TNFa secretion by cytotoxic cells (LT, NK, . . .). Progesterone also has an immunosuppressive

142

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

action that works in synergy with prostaglandins E (PGE) (and inhibits lymphocyte T proliferation). E2 prostaglandin (PGE2) inhibits the proliferation of LT CD4þ cells and their secretion of lymphokines (IL-2, TNFa and IFNg); at the same time it (like IL-6) reduces expression of IL-2 receptors on activated LT. PGE2 inhibits the production of IL-12 by antigen-activated monocytes.

11. Intrinsic resistance to cellular lysis This anti-complement activity is caused by the embryo’s expression of anti-complement molecules: membrane complement proteins (MCP or CD 46), protectin (CD 59), which blocks complement binding on the constant region of the antibody chain; and DAF (CD 55), which accelerates complement destruction. These molecules are expressed on the membrane of gametes, fertilized eggs, and blastocysts. They thus transform any cytotoxic antibody into a blocking antibody, a sub-class of facilitating antibodies. The concealment of the antigens minimizes the risks that they will be destroyed by the LT. The concept of humorally-mediated abortions is theoretically possible if these molecules are expressed slightly, badly, or not at all, or are expressed at the wrong stage or in the wrong location.

12. Cases of preeclampsia and maternal autoimmune syndrome Cellular rejection involves a cascade of events that include the participation of lymphocytes B and the production of humoral antibodies. Their action on local vessels may not only explain the elementary lesions observed but also induce a maternal autoimmune syndrome that may be reactivated in subsequent pregnancies or capable of developing on its own. This initial impulse of immune rejection, which corresponds to an antigen–antibody confrontation that is particular in its genetic determinism, appears nonetheless compatible with the normal development of pregnancy until the end of the second-trimester. The purely maternal or obstetrical environmental context sometimes involves confirmed immune disease, lupus or the anti-phospholipid syndrome (APLS). In this respect, preeclampsia is similar to some repeated spontaneous abortions. In addition, morphologically, anomalies of the same type, which associate cellular reactions and elementary vascular lesions can be observed in some types of abortions. These observations and the notion of a sequence of preeclampsia–early repeated abortions in the same patients suggests that they may be two diseases on one continuum. During preeclampsia, expression of HLA G antigen in the trophoblastic cells diminishes. In such cases, HLA G gene mutations have been found in African–American populations, although they are extremely rare in the white population. It is observed a diminution in the HLA G mRNA in the

trophoblast of patients with preeclampsia, proportional to the reduction in the number of trophoblastic cells observed in this disease. On the other hand, the existence of an HLA C antigen is proved in the trophoblastic cells; it may act synergistically with HLA G to enable NK cell recognition and thus prevent cell lysis. The decidua contains a large population of CD 56þ cells, which are markers for LGL. These cells resemble NK cells without expressing the same activity. Also found there are macrophages (19%) and T lymphocytes (8%). As pregnancy progresses, macrophages and lymphocytes T remain constant while the number of LGL falls. HLA G can inhibit LGL NK activity. The LGL cells have a particular NK phenotype and play a role in the phenomena of graft immunity. They represent 2% of the circulating blood lymphocytes. One role of these uterine myeloid cells is to produce cytokines. Some, such as CSF-1, GM-CSF, IL-1, TNFa and IFNg are known to regulate blastocyst attachment as well as trophoblast cell proliferation and invasion. TNFa, IFNg and TGFb1 inhibit trophoblastic proliferation while IL-1, GM-CSF and IL-6 increase it. These notions are related to the more general concept of the Th1–Th2 systems. Th1 groups T helper cells that secrete IL-2 and IFN g (cell-mediated immunity), while Th2 produces IL-4, IL-6 and IL-10 (humoral-mediated immunity). A successful pregnancy requires that the Th2 system be dominant. During preeclampsia, this relation is inversed (Th1 > Th2), thereby increasing the rate of destruction of trophoblastic cells. They therefore cannot fully play their necessary roles: anchorage of the embryo in the decidua (as shown by the high rate of spontaneous abortions in women with preeclampsia), modification of the uterine spiral arteries, etc. Different molecules are involved in the immunoendocrine regulation of implantation: prostaglandin E2, GM-CSF, IL10 (which reduces the production of the Th1 system and stimulates ACTH production) and TGFb. Anomalies in the production of PGE2 and TGFb have been observed during preeclampsia.

13. The genetic theory of preeclampsia [93,94] There is no single gene for preeclampsia, but probably a group of maternal genetic polymorphisms that, when associated with environmental factors, predispose her to this disease. The hypothesis of recessive transmission of maternal genes seems the most probable [95]. Moreover the fetus’ genes too appear to contribute to the development of preeclampsia. This hypothesis is confirmed by the finding that two identical twins do not have the same risk of preeclampsia [96]. Environmental factors may affect the expression of these ‘‘predisposition’’ genes. Preeclampsia at a microscopic level is morphologically expressed as a disease of the vascular endothelium [79]. Researchers are beginning to understand the initial pathophysiology. Two types of mechanisms seem to be associated.

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

First, the spiral arterioles appear to develop insufficiently, with an incomplete transformation of the decidual vessels in the nidation area and conservation of the myometrial vessel tunica muscularis, which is sensitive to vasoconstrictive amines. Second, lesions resembling those of acute atherosis appear on the vessel walls. Accordingly, from the beginning, a physical, cellular cause and a chemical, humoral cause seem to be associated. The migration of cytotrophoblasts towards the vessels is normally accompanied by a transformation in the cellular phenotype, which changes from epithelial to endothelial. While the increasing rarity of these cells in preeclampsia is generally accepted, the integrity of the cellular transformation during the migration has not yet been demonstrated. The inflammatory alterations of the spiral arterioles outside the area of nidation are related (local reactions to substances resulting from the initial implantation anomaly). The causal pathogenic mechanism currently in favour involves the destruction of infiltrating cytotrophoblasts by a series of reactions that bring into the picture the NK CD 56þ lymphocytes specific to the decidua and the LT. It is therefore the presence of these invading cytotrophoblastic cells with a particular immune profile that is responsible for the rejection reaction. The involvement of endometrial NK brings to mind first an anomaly of its specific target, HLA-G. This immune-type inflammatory stage is more easily observed in some clinical contexts that suggest a maternal recessive gene whose transmission to the fetus involves a reaction rejection only if the father passed on the same anomalous gene. The antigenic presentation of the cytotrophoblastic cells of the fetal envelope thus become unacceptable for the mother. The fetal genetic contribution to the development of preeclampsia is confirmed by findings that two identical twins do not present the same risks. Accordingly, preeclampsia may result from a hereditary disease associated with a recessive maternal gene; the disease expression thus depends on the father. Women born from pregnancies complicated by preeclampsia are at higher risk of this complication themselves; the same is true for the daughters-in-law of women who have had preeclampsia. Similarly, the risk of this complication for a woman whose partner has already had a child with another woman in a pregnancy with preeclampsia is twice as high as the risk among women with no family history on any side. There is thus a clear paternal role in the genesis of this complication [97], as there is in the phenomenon of implantation (molar pregnancies, where paternal uniparental disomy is observed). It is thus highly likely that preeclampsia involves a paternal genomic imprint of certain genes: IGF2, allele T235 of the angiotensine gene, Factor V Leiden, and methyl tetra-hydrofolate reductase (MTHFR). There are others candidate genes, located on chromosomes 1, 3, 4, 9 and 18. Some disturbances are not necessarily secondary to genetic anomalies. This is the case for example for superoxide dismutase (SOD): its expression is reduced during preeclampsia, with-

143

out any modification of the Cu–Zn SOD gene (on chromosome 21) [98]. In preeclampsia, levels of soluble TNFa (sTNFp55) receptor increase before clinical signs appear; nonetheless for neither preeclampsia nor hemolysis-enzymes liver-low platelets (HELLP) syndrome is there any activation of the promoter gene also located on chromosome 6 [93]. The candidate genes situated on chromosome 1 include the genes for factor V and for methylene tetrahydrofolate reductase (MTHFR). Factor V is pro-coagulant, but this effect is normally counterbalanced by activated protein C. Approximately 2–7% of the population carries a mutation of factor V called a ‘‘Leiden mutation’’ (FVL) that makes them resistant to activated protein C. This resistance is present among 20% of the women with preeclampsia [99]; heterozygosity for this mutation is found in 9% of women with preeclampsia and only 4.2% of the general population. Additionally, Rigo et al. [100] found relatively high rates of HELLP-syndrome patients among those with the FVL mutation. Similarly, a mutation of the MTHFR gene (C677T) has been found that can reduce enzyme activity and increase plasma concentrations of homocysteine, as has been described in preeclampsia. Both monozygous and heterozygous status predispose the carrier to preeclampsia but are not prerequisites for it. However, O’Shaughnessy et al. [101] did not findany association of preeclampsia with either FVL or MTHFR. Additionaly, an association between preeclampsia and polymorphisms in the prothrombin gene described by Kupferminc [102] was not found by others [103].The angiotensin gene is also on chromosome 1.

14. The inflammatory theory of preeclampsia [104] Preeclampsia is a disease characterised by a generalized dysfunction of the endothelial cell, linked to several factors: fatty acids, lipoproteins, lipid peroxide, TNFa, decay products of fibronectin and microvillous fragments of syncytiotrophoblastic cells. All these factors together result from a generalized intravascular inflammatory response present during pregnancy but exacerbated in preeclampsia. During inflammation, leukocyte adhesion proteins in the vascular system increase, stimulated too early by thrombin and histamine and then in the hours that follow by IL-1 or TNFa. Vascular permeability then increases, together with extravasation, a cellular chemotaxis with phagocytosis. During preeclampsia, granulocyte and monocyte activation occur together with the endothelial dysfunction; this increases the level of adhesion molecules (CD11b and CD64) or other factors (L-selectin and HLA-DR). These cells also cause an increase in TNFa, IL-6 and phospholipase A2 (important inflammatory reaction mediators) and they produce and secrete oxygen free radicals. During preeclampsia, these radicals increase as does the expression

144

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

of CD11b and CD64 (phagocytes), while L-selectin (granulocytes) and HLA-DR (monocytes) decrease. These disturbances also occur during normal pregnancies but are significantly less important. Moreover, activated neutrophils produce some proteases, including elastase, which increase during preeclampsia. Elastase is associated with increased production of endothelins and factor VIII and plays a role in the endothelial alterations observed in this disease.

15. Targets of the action of preventive treatment with aspirin and other therapies [105,106] The lack of vasodilatation associated with the absence or incompleteness of the EVCT invasion of the spiral arteries and the resulting microthromboses associated with endothelial lesions lead to anomalies in the regulation and synthesis of prostaglandins, in particular, prostacyclin (PGI2, a vasodilator) and thromboxan A2 (TXA2). Prostaglandins are metabolites of arachidonic acid by the intermediary of cyclooxygenase, and aspirin is an inhibitor of cyclooxygenase (COX). By diminishing TXA2, it modifies the PGI2/ TXA2 ratio and therefore tends to re-establish the physiological balance disturbed by the failed invasion [107]. The continuous administration of aspirin at doses that range from 0.3 to 1.5 g/kg per day seems to inhibit platelet COX activity more effectively than endothelial cell COX activity. Endothelial cells are nucleated and can resynthesize COX, unlike the anuclear platelets, which are permanently inactivated. Only new platelets, formed from megacaryocytes, can renew TXA2, but they then become incapable of producing more because the aspirin administration is repeated. Inversely, endothelial cells renew PGI2 fairly rapidly. Accordingly, it is appropriate to prescribe a preventive treatment with aspirin in pregnancies where the women have poor obstetrics histories of previous early severe preeclampsia or severe intra-uterine growth retardation defined as <3rd percentile. However, an abnormal result for the angiotensin II test [108] (IV angiotensin dose necessary to increase diastolic blood pressure by 20 mmHg, test abnormal if the result <10 ng/(kg min)) or an anti-phospholipid syndrome (where the anticardiolipin antibodies can interfere with endothelial synthesis of PGI2) were indications for lowdose aspirin prevention. Low-dose aspirin therapy prevention is ineffective among women with underlying medical illness i.e. chronic hypertension, chronic nephropathy or kidney transplant, diabetes [109]. There is no convincing data to prescribe low-dose aspirin in patients with abnormal uterine doppler findings between 22 and 24 week gestation [110]. It appears necessary to begin the treatment as early as possible: from 13 weeks or even earlier (depending on the pathophysiologic bases explained above) when the indication is associated with the obstetrical history. This early prescription will have as its goal to limit, but not to prevent,

the cascade of biological events in the mother (increase of renin, angiotensin, and aldosterone) that follow the increase in vascular resistance and the obstetrical complications that can result from it. This treatment should be continued until a term of 35 weeks. In some indications, this treatment can begin even before conception (for autoimmune diseases such as anti-phospholipid syndrome), combined with heparin or corticosteroid therapy. Although trials have been performed with varying doses of aspirin (from 50 to 150 mg), the optimal dose does not seem to have been established. Nonetheless, the initial dose usually prescribed is from 100 to 150 mg per day. An aspirin prescription should be based upon the mother’s weight and the performance of an Ivy bleeding time test. The bleeding time must be tested before the aspirin treatment begins and must be less than 8 min. After 10–15 days of treatment, bleeding time must be tested again (and must still be less than 8 min) to see whether a dose adjustment is needed (50 mg per day or even every other day). Inversely, a variation of more than 2 min in the bleeding time before and after the aspirin prescription demonstrates some platelet anti-aggregation efficacy, which seems to reduce the risk that pregnancy-related disorders (preeclampsia, IUGR, or abruptio placentae) will recur [111]. In conclusion, an analysis of the literature shows that the two most important points in obtaining a preventive effect from low-dose aspirin during the pregnancy of a woman at risk of vascular accidents (preeclampsia, IUGR) are early treatment (between 8 and 15 weeks, or even around the period of conception) and the prescription of a sufficient dose (100–150 mg per day).

16. Conclusions Preeclampsia is therefore only a form of spontaneous abortion, incomplete because it involves only the vascular face of the implanted embryo. Along with the anomaly of the Abnormal vascular invasion by extra-villous cytotrophoblast: Only 30-50% of uterine spiral arteries invaded (< plugs)

Abnormal vascular tone: > endothelins, < nitric oxide; > thromboxan A2, < prostacyclin

PREECLAMPSIA

Immunologic disorders: Th1 predominant / Th2; < HLA G

Genetic anomalies: > TNF α gene factor V Leiden ?, MTHFR ?

Inflammatory theory: dysfunction of the endothelial cells: > TNFα , lipid peroxides and fatty acids <: decreased expression; >: increased expression

Fig. 1. Pathophysiologic mechanisms of preeclampsia.

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

trophoblastic invasion of the uterine spiral arteries, we find local disruptions of vascular tone, of immunological balance, and inflammatory status, sometimes associated with genetic predispositions [112] (Fig. 1). Preeclampsia, a disease defined in the second and third-trimesters of pregnancy, is multifactorial from a pathophysiologic point of view, thus complicating its prediction and treatment as well as the efficacy of preventive measures against it. Acknowledgements The authors thank Pr Philippe Bouchard, Department of Endocrinology, Hospital St. Antoine, Paris (France) and Pr Jean-Michel Foidart, Department of Gynecology and Obstetrics, Hospital La Citadelle, Liege (Belgium) for their comments, suggestions and critical reading of the manuscript. References [1] Brosens I, Robertson WB, Dixon HG. The physiological response of the vessels of the placenta bed to normal pregnancy. J Pathol Bact 1967;93:569–79. [2] Cross JC, Werb Z, Fisher SJ. Implantation and placenta: key pieces of the development puzzle. Science 1994;166:1508–18. [3] Pijnenborg R. Establishment of uteroplacental circulation. Reprod Nutr Develop 1988;28:1581–6. [4] Strickland S, Richards WG. Invasion of the trophoblast. Cell 1992;71:355–7. [5] Zhou Y, Fisher SJ, Janatpour M, et al. Human cytotrophoblasts adopt a vascular phenotype as they differentiate: a strategy for successful endovascular invasion? J Clin Invest 1997;99:2139–51. [6] Giudice LC, Saleh W. Growth factors in reproduction. Trends Endocrinol Metab 1995;6:60–9. [7] Damsky CH, Librach C, Lim KH, et al. Integrin switching regulates normal trophoblast invasion. Development 1994;120:3657–66. [8] Flamme I, Risau W. Mechanism of blood vessel formation. Ann Anat 1995;177:493–502. [9] Kaiserman-Abramof IR, Padykula HA. Angiogenesis in the postovulatory primate endometrium: the coiled arteriolar system. Anat Rec 1989;224:479–89. [10] Hustin J. Vascular physiology and pathophysiology of early pregnancy. In: Bourne T, Jauniaux E, Jurkovic D, editors. Transvaginal colour doppler. Heidelberg: Springer; 1995. p. 47–56. [11] Hustin J, Schaaps JP, Lambotte R. Anatomical studies of the uteroplacental vascularization in the first trimester of pregnancy. Trophoblast Res 1988;3:49–60. [12] Jaffe R, Jauniaux E, Hustin J. Maternal circulation in the first trimester human placenta—myth or reality? Am J Obstet Gynecol 1997;176:695–705. [13] Meekins JW, Luckas MJM, Pijnenborg R, Mc Fadyen IR. Histological study of decidual spiral arteries and the presence of maternal erythrocytes in the intervillous space during the first trimester of normal human pregnancy. Placenta 1997;18:459–64. [14] Nanaev A, Chwalisz K, Frank HG, Kohnen G, Hegele-Hartung C, Kaufmann P. Physiological dilation of uteroplacental arteries in the guinea pig depends on nitric oxide synthase activity of extravillous trophoblast. Cell Tissue Res 1985;282:407–21. [15] Rodesch F, Simon P, Donner C, Jauniaux E. Oxygen measurements in the materno-trophoblastic border during early pregnancy. Obstet Gynaecol 1992;80:283–5.

145

[16] Halliwell B. Free radicals, antioxidants, and human diseases: curiosity, cause or consequence? Lancet 1994;344:721–4. [17] Poranen AK, Ekblad U, Uotila P, Ahotupa M. Lipid peroxidation and antioxidants in normal and pre-eclamptic pregnancies. Placenta 1996;17:401–5. [18] Schaaps JP. Three-dimensional doppler ultrasonography of the uterus and ovary. J Gynecol Obstet Biol Reprod 2000;29:264–8. [19] Jaffe R, Dorgan A, Abramowicz JS. Color doppler imaging of uteroplacental circulation in the first trimester: value in prediction pregnancy failure or complication. AJR 1995;164:1255–8. [20] Jauniaux E, Jurkovic D, Campbell S. Current topic: in vivo investigation of the placental circulation by doppler echography. Placenta 1995;16:323–31. [21] Kam EP, Gardner L, Loke YW, King A. The role of trophoblast in the physiological change in decidual spiral arteries. Hum Reprod 1999;14:2131–8. [22] Lin S, Shimizu I, Suehara N, Nakayama M, Aono T. Uterine artery doppler velocimetry in relation to trophoblast migration into the myometrium of the placental bed. Obstet Gynecol 1995;85:760–5. [23] Burrows TD, King A, Loke YW. Expression of adhesion molecules by endovascular trophoblast and decidual endothelial cells: implications for vascular invasion during implantation. Placenta 1994;15:21– 33. [24] Vicovac L, Aplin JD. Epithelial-mesenchymal transition during trophoblast differentiation. Acta Anat 1996;156:202–16. [25] Aplin JD. Hypoxia and human placental development. J Clin Invest 2000;105:559–60. [26] Caniggia I, Mostachfi H, Winter J, et al. Hypoxia-inductible factor1 mediates the biological effects of oxygen on human trophoblast differentiation through TGFb3. J Clin Invest 2000;105:577–87. [27] Myatt L. Current topic: control of vascular resistance in the human placenta. Placenta 1992;13:329–41. [28] Seligman SP, Buyon JP, Clancy RM, Young BK, Abramson SB. The role of nitric oxide in the pathogenesis of preeclampsia. Am J Obstet Gynecol 1994;171:944–8. [29] Barnea ER. Reduced catechol metabolism in the first trimester placenta of cigarette smokers. In: Barnea ER, Check JH, Grudzinskas JG, Maruo T, editors. Implantation and early pregnancy in humans. Carnforth: Parthenon Publishing; 1993. p. 387–98. [30] Shepherd RW, Stanczyk FZ, Betha CL, Novy MJ. Fetal and maternal endocrine responses to reduced uteroplacental blood flow. J Clin Endoc Metab 1992;75:301–7. [31] Collett GP, Kohnen G, Campbell S, Davenport AP, Cameron AD, Cameron IT. Localisation of endothelin receptors in human uterus throughout the menstruel cycle. Mol Hum Reprod 1996; 2:439–44. [32] Mac Mahon LP, Redman CWG, Firth JD. Expression of the three endothelin genes and plasma levels of endothelin in pre-eclamptic and normal gestations. Clin Sci 1993;85:417–24. [33] Mac Queen J, Kingdom JC, Connell JM, Whittle MJ. Fetal endothelin levels and placental endothelin receptors in intrauterine growth retardation. Obstet Gynecol 1993;82:992–8. [34] Palmer RMJ, Ferrige AG, Moncada S. Nitric oxide release accounts for the biological activity of endothelium-derived relaxing factor. Nature 1987;327:524–6. [35] Ignarro LJ. Biosynthesis and metabolism of endothelium-derived nitric oxide. Ann Rev Pharmacol Toxicol 1990;30:535–69. [36] Miyahara K, Kawamoto T, Sase K, et al. Cloning and structural characterization of the human endothelial nitric-oxide synthase gene. Eur J Biochem 1994;223:719–26. [37] Myatt L, Brockman DE, Eis AL, Pollock JS. Immunohistochemical localization of nitric oxide synthase in the human placenta. Placenta 1993;14:487–95. [38] Myatt L, Brewer A, Brockman DE. The action of nitric oxide in the perfused human fetal-placental circulation. Am J Obstet Gynecol 1991;164:687–92.

146

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147

[39] Jovanovic A, Grbovic L, Tilic I. Predominant role for nitric oxide in the relaxation induced by acetyl-choline in human uterine artery. Hum Reprod 1994;9:387–93. [40] Gude NM, Xie CY, King RG. Effects of eicosanoid and endothelial cell derived ralaxing factor inhibition on fetal vascular tone and responsiveness in the human perfused placenta. Trophoblast Res 1993;7:133–45. [41] Ramsay B, De Belder A, Campbell S, Moncada S, Martin JF. A nitric oxide donor improves uterine artery diastolic blood flow in normal early pregnancy and in women at high risk of preeclampsia. Eur J Clin Invest 1994;24:76–8. [42] Thaler I, Amit A, Jakobi P, Itskowitz-Eldor J. The effect of isosorbide dinitrate on uterine artery and umbilical artery flow velocity waveforms at mid-pregnancy. Obstet Gynecol 1996; 88:838–43. [43] Mabry White M, Zamudio S, Stevens T. Estrogen, Progesterone, and vascular reactivity: potential cellular mecanisms. Endoc Rev 1997;16:739–51. [44] Perrot-Applanat M, Deng M, Fernandez H, Lelaidier C, Meduri G, Bouchard P. Immunohistochemical localization of estradiol and progesterone receptors in human uterus throughout pregnancy: expression in endometrial blood vessels. J Clin Endoc Metab 1994;78:216–24. [45] Liao QP, Buhimschi I, Saade G. Regulation of vascular adaptation during pregnancy and post-partum. Effects of nitric oxide inhibition and steroid hormones. Hum Reprod 1996;11:2777–84. [46] Weiner CP, Lizasoain I, Baylis SA, Knowles RG, Charles IG, Moncada S. Induction of calcium-dependant nitric oxide synthases by sex hormones. Proc Natl Acad Sci USA 1994;91:5212–6. [47] Hinson JP, Vinson GP, Kapas S, Teja R. The relationship between adrenal vascular events and steroid secretion: the role of the mast cell and endothelin. J Steroid Biochem Mol Biol 1991;40:381–9. [48] Smith WL, Marnett LJ, Dewitt DL. Prostaglandin and thromboxan biosynthesis. Pharmacol Ther 1991;49:153–79. [49] Ding ZQ, Rowe J, Sinosich MJ, Saunders DM, Gallery ED. In-vitro secretion of prostanoids by placental villous cytotrophoblasts in pre-eclampsia. Placenta 1996;17:407–11. [50] Davidge ST, Hubel CA, Mc Laughlin MK. Lipid peroxidation increases arterial cyclooxygenase activity during pregnancy. Am J Obstet Gynecol 1994;170:215–22. [51] Wetzka B, Charnock-Jones DS, Viville B, et al. Expression of prostacyclin and thromboxane synthases in placenta and placental bed after pre-eclamptic pregnancies. Placenta 1996;17:573–81. [52] Barnea ER, Lavy G, Fakih H, De Cherney AH. The role of ACTH in placenta steroidogenesis. Placenta 1986;7:307–13. [53] Clifton VL, Read MA, Boura ALA, Robinson PJ, Smith R. Adrenocorticotropin causes vasodilatation in the human fetalplacenta circulation. J Clin Endoc Metab 1996;81:1406–10. [54] Clifton VL, Read MA, Leitch IM, et al. Corticotropin-releasing hormone-induced vasodilatation in the human fetal-placenta circulation: involvement of the nitric oxide-cyclic guanosine 30 ,50 monophosphate-mediated pathway. J Clin Endoc Metab 1995;80:2888–93. [55] Goland RS, Jozak S, Warren WB, Conwell IM, Stark RI, Tropper PJ. Elevated levels of umbilical cord plasma corticotropin-releasing hormone in growth-retarded fetuses. J Clin Endoc Metab 1993;77:1174–9. [56] Bodelsson G, Stjernquist M. Smooth muscle dilatation in the human uterine artery induced by substance P, vasoactive intestinal polypeptide, calcitonin gene-related peptide and atrial natriuretic peptide: relation to endothelium-derived relaxing substances. Hum Reprod 1992;7:1185–8. [57] Graf AH, Hutter W, Hacker GW, et al. Localization and distribution of vasoactive neuropeptides in the human placenta. Placenta 1996;17:413–21. [58] Hansen V, Maigaard S, Allen J, Forman A. Effects of vasoactive intestinal polypeptide and substance P on human intramyometrial

[59] [60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

[73]

[74]

[75]

[76]

[77] [78]

arteries and stem villous arteries in term pregnancy. Placenta 1988;9:501–6. Carter TL, Park P, Pratt K. Effects of CGRPa on the migration of endothelial cells of human origin. FASEB J 1993;7:125. Datta HK, Rafter PW, Wimalawansa SJ. CGRP and amylinamide display a proliferative effect on human umbilical endothelial cells and osteoblast-like osteosarcoma cell-line. Biochem Soc Trans 1990;18:1276–8. Stevenson JC, MacDonald DWR, Warren RC, Brooker MW, Whitehead MI. Increased concentration of calcitonin gene-related peptide during normal human pregnancy. Br Med J 1986;293: 1329–30. Nelson SH, Steinsland OS, Suresh MS. Possible physiologic role of calcitonin gene-related peptide in the human uterine artery. Am J Obstet Gynecol 1993;168:605–11. Wimalawansa SJ. Anti-hypertensive effects of oral calcium supplementation is mediated through the potent vasodilatator CGRP. Am J Hypertens 1993;6:996–1002. Wimalawansa SJ. Calcitonin gene-related peptide and its receptors: molecular genetics, physiology, pathophysiology and therapeutic potentials. Endoc Rev 1996;17:533–85. Yallampalli C, Dong YL, Wimalawansa SJ. Calcitonin gene-related peptide reverses the hypertension and significantly decreases the fetal mortality in pre-eclampsia rats induced by NG-nitro-L-arginine methyl ester. Hum Reprod 1996;11:895–9. Belizan JM, Villar J, Gonzales L, Campodonico L, Bergel E. Calcium supplementation to prevent hypertensive disorders of pregnancy. N Engl J Med 1991;325:1399–405. Knight KB, Keith RE. Calcium supplementation on normotensive and hypertensive pregnant women. Am J Clin Nutr 1992;55: 891–5. Sanchez-Ramos L, Briones DK, Kaunitz AM, Delvalle GO, Gaudier FL, Walker CD. Prevention of pregnancy-induced hypertension by calcium supplementation in angiotensin IIsensitive patients. Obstet Gynecol 1994;84:349–53. Kilby MD, Pipkin FB, Symonds EM. Calcium and platelets in normotensive and hypertensive human pregnancy. J Hypertens 1992;10:997–1003. Ebeigbe AB, Ezimokhai M, Aloamaka CP. Responses of arterial smooth muscle from normotensive and pre-eclamptic subjects to the calcium channel agonist, Bay K 8644. Res Exp Med (Berl) 1987;187:461–8. Merviel P, Challier JC, Carbillon L, Foidart JM, Uzan S. The role of integrins in human embryo implantation. Fetal Diagn Ther 2001;16:364–71. Zhou Y, Damsky CH, Chiu K, Roberts JM, Fisher SJ. Preeclampsia is associated with abnormal expression of adhesion molecules by invasive cytotrophoblasts. J Clin Invest 1993;91:950–60. Fox H. Effect of hypoxia on trophoblast in organ culture: a morphologic and autographic study. Am J Obstet Gynecol 1970; 107:1058–64. Genbacev O, Joslin R, Damsky CH, Polliotti BM, Fisher SJ. Hypoxia alters early gestation human cytotrophoblast differentiation/invasion in vitro and models the placental defects that occur in preeclampsia. J Clin Invest 1996;97:540–50. Walsh SW. Preeclampsia: an imbalance in placental prostacyclin and thromboxane production. Am J Obstet Gynecol 1985;152: 335–40. Maseki M, Nishigaki I, Hagihara M, Tomoda Y, Yagi K. Lipid peroxide levels and lipid content of serum lipoprotein fractions of pregnant subjects with and without preeclampsia. Clin Chim Acta 1981;115:155–61. Nelson GH, Zuspan FP, Mulligan LT. Defects of lipid metabolism in toxemia of pregnancy. Am J Obstet Gynecol 1966;94:310–5. Fitzgerald DJ, Rocki W, Murray R, Mayo G, Fitzgerald GA. Thromboxane A2 synthesis in pregnancy-induced hypertension. Lancet 1990;335:751–4.

P. Merviel et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 115 (2004) 134–147 [79] Tsukimori K, Maeda H, Shingu M, Koyanagi T, Nobunaga M, Nakano H. The possible role of endothelial cells in hypertensive disorders during pregnancy. Obstet Gynecol 1992;80:229–33. [80] Davidge ST, Hubel CA, Brayden RD, Capeless EC, Mc Laughlin MK. Sera antioxidant activity in uncomplicated and preeclamptic pregnancies. Obstet Gynecol 1992;79:897–901. [81] Jauniaux E, Watson AL, Hempstock J, Bao YP, Skepper JN, Burton GJ. Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure. Am J Pathol 2000;157:2111–22. [82] Oian P, Kjeldsen SE, Eide I, Maltau JM. Increased arterial catecholamines in pre-eclampsia. Acta Obstet Gynecol Scand 1986;65:613–7. [83] Mochizuki M, Morikawa H, Yamasaki M, Maruo T. Vascular reactivity in normal and abnormal gestation. Am J Kidney Dis 1991;17:139–43. [84] Zamudio S, Leslie K, White M, Hagerman DD, Moore LG. Low serum estradiol and high serum progesterone concentrations characterize hypertensive pregnancies at high altitude. J Soc Gynecol Invest 1994;1:197–205. [85] Brennecke SP, Gude NM, Dilulio JL, King RG. Nitric oxide synthase activity in human placental villous tissue is reduced in preeclampsia [Abstract 1]. In: Proceedings of the Ninth International Congress for the Study of Hypertension in Pregnancy; 1994. p. 87. [86] Buhimschi I, Yallampalli C, Chwalisz K, Garfield RE. Preeclampsia-like conditions produced by nitric oxide inhibition: effects of L-arginine, D-arginine and steroid hormones. Hum Reprod 1995;10:2723–30. [87] Molnar M, Suto T, Toth T, Hertelendry F. Prolonged blockade of nitric oxide synthesis in gravid rats produces sustained hypertension, proteinuria thrombocytopenia and intrauterine growth retardation. Am J Obstet Gynecol 1994;170:1458–70. [88] Grunewald C, Kublickas M, Carlstrom K, Lunell NO, Nisell H. Effects of nitroglycerin on the uterine and umbilical circulation in severe preeclampsia. Obstet Gynecol 1995;86:600–4. [89] Lees C, Langford E, Brown AS, et al. The effects of Snitrosoglutathione on platelet activation, hypertension and uterine and fetal doppler in severe preeclampsia. Obstet Gynecol 1996;88:14–9. [90] Moncada S, Higgs EA. Molecular mechanisms and therapeutic strategies related to nitric oxide. FASEB J 1995;9:1319–30. [91] Estelles A, Gilabert J, Grancha S, et al. Abnormal expression of type 1 plasminogen activator inhibitor and tissue factor in severe preeclampsia. Thromb Haemost 1998;79:500–8. [92] Wegman TG. Foetal protection against abortion: is it immunosuppression or immunostimulation? Am Immunol Inst Pasteur 1984;135:309–11. [93] Pipkin FB. Risk factors for preeclampsia. N Engl J Med 2001;344:925–6. [94] Roberts JM, Cooper DW. Pathogenesis and genetics of preeclampsia. Lancet 2001;357:53–6. [95] Liston WA, Kilpatrick DC. Is genetic susceptibility to preeclampsia conferred by homozygosity for the same single

[96] [97]

[98]

[99]

[100]

[101]

[102]

[103]

[104]

[105] [106]

[107] [108]

[109] [110]

[111]

[112]

147

recessive gene in mother and fetus? Br J Obstet Gynaecol 1991;98:1079–86. Thornton JG, Onwude JL. Pre-eclampsia: discordance among identical twins. Br Med J 1991;303:1241–2. Esplin MS, Fausett MB, Fraser A, et al. Paternal and maternal components of the predisposition to preeclampsia. N Engl J Med 2001;344:867–72. Chen G, Wilson R, Boyd P, et al. Normal superoxide dismutase (SOD) gene in pregnancy-induced hypertension: is the decreased SOD activity a secondary phenomenon? Free Radic Res 1994;21:59–66. Lachmeijer AM, Dekker GA, Pals G, Aarnoudse JG, Ten Kate LP, Arngrimsson R. Searching for preeclampsia genes: the current position. Eur J Obstet Gynecol Reprod Biol 2002;105:94–113. Rigo J, Nagy B, Fintor L, Tanyi J, Beke A, Karadi I. Maternal and neonatal outcome of preeclamptic pregnancies: the potential roles of factor V Leiden mutation and 5,10 methylenetetrahydrofolate reductase. Hypertens Pregnancy 2000;19:163–72. O’Shaughnessy KM, Fu B, Ferraro F, Lewis I, Downing S, Morris NH. Factor V Leiden and thermolabile methylenetetrahydrofolate reductase gene variants in an East Anglian preeclampsia cohort. Hypertension 1999;33:1338–41. Kupferminc MJ, Eldor A, Steinman N, et al. Increased frequency of genetic thrombophilia in women with complications of pregnancy. N Engl J Med 1999;340:9–13. Livingston JC, Barton JR, Park V, Haddad B, Phillips O, Sibai BM. Maternal and fetal inherited thrombophilias are not related to the development of severe preeclampsia. Am J Obstet Gynecol 2001;185:153–7. Redman CWG, Sacks GP, Sargent IL. Preeclampsia: an excessive maternal inflammatory response to pregnancy. Am J Obstet Gynecol 1999;180:499–506. Merviel P, Uzan S. Aspirine et grossesse. Sang Thromb Vaiss 1996;8:47–54. Uzan S, Merviel P, Kadoch O, et al. Aspirin and prevention of vascular complications: there are still indications. Ultrasound Obstet Gynecol 2000;15:4–6. Vayne JR. Inhibition of prostaglandin synthesis as a mechanism of action of aspirin-like drugs. Nature 1971;231:232–5. Gant NF, Daley GL, Chand S, Whalley PJ, Mac Donald PC. A study of angiotensin II pressor response throughout primigravid pregnancy. J Clin Invest 1973;52:2682–9. Heyborne KD. Preeclampsia prevention: lessons from the low-dose aspirin therapy trials. Am J Obstet Gynecol 2000;183:523–8. Subtil D, Goeusse P, Houfflin-Debarge V, et al. Randomised comparison of uterine artery Doppler and aspirin (100 mg) with placebo in nulliparous women: the Essai Re´ gional Aspirine Me`re-Enfant study (part 2). Br J Obstet Gynaecol 2003;110: 485–91. Dumont A, Flahault A, Beaufils M, Verdy E, Uzan S. Effect of aspirin in pregnant women is dependent on increase in bleeding time. Am J Obstet Gynecol 1999;180:135–40. Walker JJ. Pre-eclampsia. Lancet 2000;356:1260–5.