Keywords
endothelial cell activation, placental microvesicles, cytokines, antiangiogenic factors, eicosanoids, reactive oxygen species, lipid peroxidation
Editors’ comment: It was not until the early 1980s that physicians and scientists gained an appreciation of the physiological importance of the endothelium, that simple unicellular layer lining the luminal surface of blood vessels. Indeed, in Chesley’s first, single-authored edition of this text, the reference to this term was as it related to the “endotheliosis” lesion of the renal glomerulus. We now recognize that endothelial cells are critical sensors of the milieu interieur and potent regulators of vascular tone, organ perfusion, and ischemia. The “endothelial hypothesis” of preeclampsia etiology provides for a convergence of several factors thought to play fundamental roles in its pathogenesis: leukocytes, cytokines, fatty acids, oxygen free radicals, placental microvesicles, “antiangiogenic” factors, and autoantibodies are all considered. Moreover, a variety of therapeutic interventions has been conceived based on the principle of endothelial cell protection, and the clinical trials emanating from these concepts are reviewed in the chapter .
Introduction
The clinical manifestations of preeclampsia–eclampsia have been recognized since antiquity, but the pathophysiology of this syndrome remained completely obscure for nearly two millennia. Beginning in the mid-19th century, abnormal renal handling of nitrogen and water, referred to at that time as “dropsy,” was first reported among eclamptic women. Recognition that these signs were a manifestation of endothelial cell injury took another 90 years. Renal glomerular capillary endotheliosis, swelling of endothelial cytoplasm and obliteration of endothelial fenestrae, was originally observed in 1924 by Mayer, described further by Bell, and later refined by Spargo et al. The latter investigator was among the first to apply electron microscopy in this condition and is credited for introducing the terminology “glomerular capillary endotheliosis” as a characteristic feature of preeclampsia.
Investigators have documented morphological evidence of vascular injury in various organs of preeclamptic women. Sheehan and Lynch described hepatic periportal vascular lesions, characterized by arterial media infiltration and capillary thrombosis, in fatal cases of preeclampsia and eclampsia. Shanklin and Sibai observed ultrastructural defects in the mitochondria of myometrial venules in women with preeclampsia, further supporting a widespread endothelial cell disorder in this syndrome. As described in Chapter 6 , Chapter 10 , several animal models of preeclampsia, including those established in nonhuman primates, have manifestations of systemic endothelial cell activation. Moreover, Grundmann et al. demonstrated that concentrations of freely circulating endothelial cells, markers of vascular injury, were more than 4-fold higher in women with preeclampsia compared to normal pregnant controls.
Abnormal trophoblast invasion and impaired placental perfusion appear to be requisite precursors to the development of preeclampsia. If we imagine the placental bed as the root of the disease process, this might be a good place to begin to look for evidence of endothelial cell dysfunction. Indeed, the endothelium lining the uterine spiral arterioles, which normally undergoes denudation and replacement by invading endovascular cytotrophoblasts, may manifest the first pathological changes associated with preeclampsia. The failure of trophoblasts to undergo “pseudovasculogenesis” and assume an “endothelial” phenotype in preeclampsia is discussed in the accompanying Chapter 5 , Chapter 6 . Moreover, other vascular manifestations have been noted at this site. The acute atherosis lesion, initially described by Zeek and Assali and refined by Nadji and Sommers and DeWolf et al., demonstrates endothelial cell vacuolization, myointimal proliferation and foam cell infiltration of the tunica media. These histological changes also are observed in the vascular pathology of atherosclerosis ; the similarities suggest that endothelial cell dysfunction may be a mechanism responsible for diffuse vascular disease that may manifest in patients with preeclampsia.
Endothelial cells are strategically positioned at the interface between circulating blood and vascular smooth muscle or the extravascular space, where they occupy a surface area of more than 1000 m 2 . These cells are able to secrete a variety of signaling molecules directly into the circulation, potentially reaching every cell in the body. In turn, endothelial cells are themselves targets for cellular and soluble plasma constituents (e.g., cytokines, lipoproteins, platelets, leukocytes, placental microvesicles, antibodies, and other circulating factors). Endothelial cells modulate vascular tone, coagulation, permeability, and the targeting of immune cells. Under normal, physiological conditions the endothelium maintains homeostatic balance. Vascular tone is controlled by vasoconstrictors (e.g., endothelin and thromboxane A2) and vasodilators (e.g., nitric oxide (NO), prostacyclin (PGI2) and endothelial-derived hyperpolarizing factors (EDHF)). Hemostasis is maintained in equilibrium by procoagulant and anticoagulant influences and endothelial tight junctions control vascular permeability. Maternal endothelial cell dysfunction is believed to result in vasospasm, microthrombosis, and vascular permeability, which lead to the classical signs and symptoms observed in women with preeclampsia.
Findings relevant to endothelial cell biology in preeclampsia will be reviewed in this chapter in five parts. In Part I, we review the evidence favoring endothelial cell dysfunction and present a pivotal hypothesis for the pathophysiology of preeclampsia. Part II reviews the concept that circulating factors causally induce disturbances in maternal endothelial cell function. Part III proposes that prooxidant principles lead to a common convergence that creates an oxidative stress environment within maternal endothelia of preeclamptic women. Part IV discusses clinical trials that have been focused on improving vascular function. Finally, in Part V we conclude with a summary of the findings, speculations about future diagnostic and therapeutic approaches to preeclampsia, and some suggestions for the direction of new investigations into the mechanisms of this enigmatic condition.
Part I: Endothelial Cell Function and Preeclampsia
Over the past decade, the attention of scholars of preeclampsia has focused on the triad of impaired trophoblast invasion, uteroplacental ischemia, oxidative stress and generalized maternal endothelial cell activation or injury as major mechanistic factors in its pathogenesis. It is likely that several etiologies underlie the development of this syndrome, but these pathogenetic mechanisms are shared. As described in Chapter 5 (and detailed further in its new Appendix), failure of trophoblast differentiation, invasion, and vascular remodeling of the placental bed are now believed to be critical initiating factors in the development of preeclampsia and reduced angiogenic factor production and action ( Chapter 6 ).
We will review recent evidence that supports the concept that one of the most important targets of circulating “toxins” in preeclampsia is the maternal vascular endothelium, which ultimately leads to vascular oxidative stress. Indeed, the historical moniker “toxemia” should be revisited as an appropriate appellation for this syndrome. When clinically manifested in later pregnancy, the vessels of essentially all maternal organs are involved in preeclamptic women. Cerebral ischemia, edema and convulsions, pleural effusion and ascites, and hepatic dysfunction, in addition to the classical proteinuria and peripheral edema reflecting renal and subcutaneous vasculatures, respectively, are evidence of the many maternal vascular beds at risk. Endothelial cells within these compromised vessels not only lose their normal constitutive, homeostatic functions but also acquire new pathological properties (e.g., vasoconstrictor and procoagulant production) that increase vascular resistance and exacerbate ischemia. In this chapter we will present the hypotheses and evidence that support a major role of endothelial cell dysfunction due to a variety of factors in the circulation creating endothelial oxidative stress as a common downstream converging mechanism in the clinical manifestations of preeclampsia. Studies suggest that circulating factors are elaborated in response to uteroplacental ischemia. These factors are postulated to alter both maternal and placental endothelial cell phenotypes.
Endothelial Cell Dysfunction in Preeclampsia
Endothelial cell dysfunction and/or “activation” is a term used to define an altered state of endothelial cell differentiation, typically induced as a result of cytokine stimulation. It represents an inflammatory response to sublethal injury of these cells and is proposed to play a major role in the pathophysiology of atherosclerosis. Endothelial cell activation in preeclampsia may result from a variety of circulating factors, including many discussed in the previous chapters such as angiogenic factors, metabolic factors, and inflammatory mediators.
Endothelial activation is manifest biochemically by the synthesis and secretion of a variety of endothelial cell products including prostanoids, endothelin-1 (ET-1), platelet-derived growth factor (PDGF), fibronectin, selectins, and other molecules that influence vessel tone and remodeling. When these factors are elaborated in response to acute mechanical or biochemical endothelial cell damage they facilitate efficient wound healing. However, when activated by a chronic pathological process, such as preeclampsia, these responses can create a vicious circle of vasospasm, microthrombosis, and disruption of vascular integrity, creating serious physiologic disturbances, which persist until the inciting factor(s) is eliminated.
Some insights into the vascular pathophysiology of preeclampsia are derived from clinical experience with thrombotic microangiopathic disorders occurring during pregnancy. It has been noted that, while rare, pregnancy appears to predispose or exacerbate the development of these syndromes of microvascular thrombosis. Thrombotic thrombocytopenic purpura (TTP) is characterized by thrombocytopenia, microangiopathic hemolytic anemia, renal involvement, neurological symptoms, and fever. Like preeclampsia, the signs of TTP during pregnancy most commonly manifest in the late second trimester. Intravascular platelet agglutination, widespread hyaline thrombosis, and endothelial cell proliferation within capillaries and arterioles have been observed. PGI 2 normally suppresses platelet aggregation induced by cytokines and shear stress, but this vasodilator eicosanoid is reduced in TTP, in part, from endothelial apoptosis and also from accelerated metabolism.
Circulating Markers of Endothelial Cell Activation
Procoagulant Proteins and Plasminogen Activators
Preeclampsia is characterized by a maternal hypercoagulable state, resulting in intravascular coagulation, microthromboses in several organs including the placenta, with further impairment of the uteroplacental circulation. The clinical significance of this problem is reviewed in Chapter 17 , but hypercoagulability also underlies the principle of endothelial cell dysfunction in preeclampsia. Excessive fibrin deposition in the placenta was reported by Kitzmiller and Benirschke, suggesting that a disordered balance of placental coagulation and fibrinolysis may play a role in the activation of hemostasis. Roberts et al. posited that the hypercoagulable state was due, in part, to diffuse endothelial cell activation. The reduced expression of several relevant endothelial cell-associated anticoagulant proteins has been shown in preeclampsia, including antithrombin III, protein C, and protein S. The significance of these proteins as markers of maternal endothelial cell dysfunction is reviewed below.
Resistance to activated protein C is an inherited mutation of the coagulation factor V gene. The presence of the factor V Leiden mutation predisposes to thromboembolic events and its prevalence is increased in women with severe, early-onset preeclampsia compared to women with normal pregnancies. In addition, in women with severe, early-onset preeclampsia, 25% had evidence of functional protein S deficiency, 18% demonstrated hyperhomocysteinemia, and 29% had detectable anticardiolipin IgG and IgM antibodies.
Increased endothelial expression of other procoagulant proteins, including tissue factor, von Willebrand factor, platelet-activating factor, β-thromboglobulin, cellular fibronectin, and thrombomodulin, has also been reported. The latter two endothelial cell markers have been shown to differentiate preeclampsia from other forms of hypertension in pregnancy. Similarly, inhibitors of fibrinolytic or antithrombotic proteins appear to play a role in the imbalance of the coagulation cascade. Plasminogen activator inhibitor type 1 (PAI-1), whose synthesis during pregnancy is predominantly placental in origin, was observed to be increased in the plasma of women with preeclamptic pregnancies. Increased decidual and amniotic fluid concentrations of PAI-1 also have been reported in preeclampsia. Circulating levels of PAI-1, thrombomodulin and fibronectin were found to correlate directly with severity of the syndrome.
Endothelial Cell Adhesion Molecules
In addition to reduced anticoagulant synthesis, endothelial cells express extracellular matrix glycoproteins with procoagulant activities as a response to injury or activation. Two examples are fibronectin and von Willebrand factor. Both of these proteins are predominantly localized to the abluminal extracellular matrix of human endothelium, but as discussed below these can be actively secreted from endothelial cells under conditions of cellular activation. In addition, exposed fibronectin is stimulatory for neutrophil attachment and has important pathophysiological consequences.
Elevated concentrations of fibronectin in women with preeclampsia have been recognized for decades and have been confirmed in other clinical conditions associated with endothelial cell dysfunction. A specific cellular fibronectin isoform (cFN), which is non-hepatic in origin and almost exclusively localized to the vascular endothelium, has been of particular interest. It can be distinguished at the protein level by two extra domains (ED-A and ED-B) generated by differential mRNA splicing in endothelial cells. While it is a major component of the endothelial extracellular matrix, cFN is normally only a minor component of circulating fibronectin and thus is an accurate marker of endothelial cell injury.
Lockwood and Peters showed that ED-A fibronectin levels were increased in a cross-sectional study of women who later developed preeclampsia. Taylor et al. used a monoclonal antibody that recognizes a conformational epitope near the ED-B region of cFN and showed in a prospective, longitudinal study that plasma cFN concentrations were increased as early as the second trimester in women destined to develop preeclampsia but the analyte was not elevated in women developing transient hypertension without proteinuria.
The concentrations of other endothelial cell adhesion molecules, including VCAM-1, and P-selectin, are elevated in cases of preeclampsia. ICAM-1 and VCAM-1 also appear to have some value as predictors of preeclampsia as early as the midtrimester of pregnancy.
Mitogenic Activities and Growth Factors
Endothelial cells also respond to injury or activation with the release or secretion of mitogenic proteins or peptides. With acute vascular trauma, this response teleologically encourages the proliferation of vascular smooth muscle, allowing vessel remodeling and repair. However, in illnesses such as atherosclerosis or preeclampsia, mitogenic factors cause reduced blood flow by promoting vessel wall hypertrophy. Indeed, plasma from women with preeclampsia collected prior to delivery had more mitogenic activity than plasma from the same women obtained 48 hours post-partum. By contrast, plasma from normal pregnant women had essentially the same mitogenic activity before and after delivery. These “mitogenic indices” were elevated from as early as the first trimester of pregnancy, compared to women who proceeded to have normal pregnancy outcomes. Other studies indicate that the mitogenic activity is protease-, heat- and acid-labile and has an apparent molecular mass of ~150,000. To our knowledge the following hypothesis has not been tested directly, but activating Ang II type 1 receptor (AT1R) autoantibodies, which are present in women with preeclampsia as discussed in detail below, would be expected to have the identical biochemical characteristics since myocardial fibroblasts are induced to undergo mitogenesis through activation of the AT1R.
Growth Factor Binding Proteins in Preeclampsia
Based on the high apparent molecular mass of the mitogenic activator(s) described above, the plasma activity was postulated to be attributable to IGF binding protein complexes. However, most studies failed to show differences in circulating concentrations of these complexes in preeclampsia. In a longitudinal study comparing 20 primipara who developed preeclampsia with 20 matched, normal pregnant controls, de Groot et al. found that midtrimester maternal plasma IGFBP-1 concentrations were significantly reduced in the group of women who developed preeclampsia approximately 20 weeks later. They interpreted their data as reflecting reduced trophoblastic invasion of the maternal decidual stroma and decreased vascular deportation of IGFBP-1 in pregnancies that result in preeclampsia. This observation was confirmed by Hietala et al.; but it should be noted that some reports of third-trimester maternal serum IGFBP-1 show elevations in women with active, symptomatic severe preeclampsia.
The growth factor binding protein that to date has had the greatest impact on the preeclampsia field is soluble VEGF receptor 1 (sFlt-1). The history and role of this glycoprotein are presented comprehensively in Chapter 6 . Briefly, it is now broadly established that elevated sFlt-1 concentrations in women with preeclampsia are associated with decreased circulating levels of free VEGF and free PlGF. Soluble endoglin is a proteolytic cleavage product of the cell surface TGF-β co-receptor, released into the plasma by matrix metalloproteinases. This extracellular-domain protein binds BMP9 and BMP10 with high affinity, scavenging these TGF-β ligands in the circulation, blocking their angiogenic signaling in the microvasculature. Overall, many of the biomarkers of endothelial cell activation described above are also functional mediators of endothelial cell dysfunction. Their activities are described in the following section.
Part II: Circulating Factors Induce Endothelial Cell Dysfunction
Efforts to identify the putative circulating toxic factors responsible for endothelial cell dysfunction in preeclampsia have been ongoing for the past decade with disappointing results. The effort continues, but while specific candidates have been identified, most data suggest that multiple molecular species account for endothelial activation. Because of their direct contact with the vascular endothelial cell monolayer, various plasma constituents are likely candidates for endothelial cell activation in preeclampsia. Formed blood elements (e.g., platelets and neutrophils), placental membrane microvesicles, soluble proteins (e.g., autoantibodies and cytokines), lipids, cytokines, and matrix metalloproteinases (MMPs) have all been identified in the plasma of women with preeclampsia. In this section, we systematically discuss some of these postulated mediators of endothelial cell dysfunction in preeclampsia.
Formed Elements in Blood as Activators of Endothelium
Platelets
Clinical manifestations of abnormal platelet function include hypercoagulability, as reviewed in Chapter 17 . In vitro evidence of abnormal platelet aggregation in women with preeclampsia suggests that this phenonmenon could lead to endothelial cell activation. Prostacyclin receptor concentrations ( B max ) in platelet membranes were similar in women with normal pregnancy, preeclampsia, and transient hypertension of pregnancy. However, their binding affinity ( K a ) was considerably reduced in preeclampsia compared to the other two groups, resulting in enhanced aggregation. Retrospective cohort studies of women with renal disease in pregnancy indicate that the incidence of preeclampsia was significantly less common in women treated prophylactically with heparin or aspirin.
The platelet-specific NO donor, S -nitrosoglutathione, was administered to 10 women with severe preeclampsia at 21–33 weeks gestation. Significant, dose-dependent reductions in mean arterial pressure and uterine artery resistance indices were observed. Platelet activation, as measured by P-selectin expression, also was found to be decreased by S -nitrosoglutathione treatment. Thus, activated platelets may be one of the circulating factors in preeclampsia that mediate maternal endothelial cell activation and platelet-specific NO donors may prove beneficial in the management of severe preeclampsia.
Neutrophils
Innate immune cells bind to selectins on the surface of activated endothelial cells and modify the function of the vascular intima. Neutrophil activation was studied in 20 eclamptic and 10 preeclamptic patients and compared to 10 normotensive controls. Plasma levels of neutrophil elastase were increased significantly in eclamptic and preeclamptic women. Elastase values in cases of eclampsia were highly correlated with mean blood pressure, serum ET-1 levels, and endothelial cytotoxicity, measured by fura-2 release from human umbilical vein endothelial (HUVE) cell cultures. Lactoferrin also has been used as an indicator of neutrophil activation in normal and preeclamptic pregnancy. A comparative study between 40 normal and 42 preeclamptic women in the third trimester of pregnancy demonstrated that predelivery ratios of lactoferrin per neutrophil were higher in preeclamptic than in normal women.
These and other data suggest that neutrophil activation plays a role in this syndrome. Using whole blood cytometric analysis, Studena et al. observed an increase in leukocyte surface antigen expression in normal pregnancy relative to nonpregnant women and a further increase in activation antigen expression in cells from preeclamptic pregnancies. The findings support the hypothesis of Redman and Sargent that vascular inflammation plays a primary role. This theory is discussed in detail in Chapter 8 . By contrast, eosinophils do not appear to have a compelling effect in preeclampsia.
Activated monocytes and neutrophils release the hemoprotein myeloperoxidase (MPO). Interestingly, MPO induces low-density lipoprotein oxidation, activates metalloproteinases, and oxidatively consumes endothelium-derived NO, which are all reported to be involved in the vascular pathophysiology of preeclampsia (see sections below). Gandley et al. have shown that MPO levels are significantly increased in the circulation and placenta of women with preeclampsia and they speculate that MPO contributes to oxidative stress in the endothelium and placenta of women with preeclampsia.
Placental Membrane Microvesicles
Circulating syncytiotrophoblast microvesicles have been postulated as endothelial toxic factors derived from the placenta. These are discussed in detail in Chapter 8 . It was observed that dilutions of 20% sera from women with preeclampsia did not inhibit endothelial cell proliferation, whereas 20% plasma from the same preeclamptic women significantly suppressed endothelial cell growth compared with 20% normal pregnancy plasma. The suppression was even more pronounced in cases of severe preeclampsia. The authors discovered that syncytiotrophoblast microvesicles added to blood could not be recovered from serum, but only from plasma, and suggested that the plasma-borne endothelial cell suppressive factor in preeclampsia may be derived from particles released from the microvillous surface of preeclamptic placentas. Indeed, these microvesicles can be isolated by ultracentrifugation from the circulation of women with preeclampsia. These same particles have been shown to induce ultrastructural evidence of endothelial cell injury when incubated with isolated human arteries.
Endothelial Progenitor Cells (EPCs)
Recent research highlights the potential role of endothelial progenitor cells (EPCs) in the pathology of preeclampsia. EPCs encompass two distinct types of cells, circulating angiogenic cells (CACs) and endothelial colony forming cells (ECFCs), both of which are involved in de novo vessel formation and repair. ECFCs are highly proliferative and differentiate into mature endothelial cells at the site of vessel formation, while CACs are hematopoietic cells which promote migration and proliferation of ECFCs via the release of paracrine factors (reviewed in ). A decline in circulating EPCs is associated with endothelial dysfunction and cardiovascular disease. Compared to normal pregnancies, in which the level of circulating EPCs increases with gestational age, women with preeclampsia have significantly reduced numbers of EPCs. It has been suggested that limited bioavailability of NO, which is required for mobilization of EPCs, and an increase in antiangiogenic factors in preeclampsia may contribute to EPC-mediated endothelial dysfunction. Interestingly, diminished levels of EPCs persist in the circulation of preeclamptic mothers postpartum, and are suggested to be associated with long-term cardiovascular risk.
Immune Complexes: Antiphospholipid and Antiendothelial Cell Antibodies
As preeclampsia occurs more commonly in primigravidae and in women with underlying collagen-vascular diseases, an immunological component has long been suspected. Moreover, studies by Brosens et al. and Kitzmiller and Benirschke demonstrated histological changes in the placental beds of women with preeclampsia that resemble those of acute allograft rejection. Antipaternal epitope recognition of the placenta has been postulated as a stimulus to maternal immune complex deposition in preeclampsia.
This hypothesis is consistent with several epidemiological observations including a higher prevalence of preeclampsia in first pregnancies and when couples use condoms as their primary form of contraception. Some studies suggest that pregnancies conceived after ovum donation, where both gametes are immunologically foreign to the host, have a high incidence of preeclampsia. By contrast, previous exposure to foreign or paternal antigens appears to reduce the risk of preeclampsia. A history of prior blood transfusion, the practice of oral sex or an extended length of cohabitation prior to conception all are associated with a lower than expected prevalence of preeclampsia. These data have been comprehensively reviewed.
Elevated levels of IgG or IgM antibodies to cardiolipin and phosphatidylserine suggest that antiphospholipid antibodies may play a pathogenic role in some women with preeclampsia. In case-control studies, antineutrophil autoantibodies were significantly increased in preeclamptic and eclamptic subjects compared to normal controls ; particularly those associated with β 2 -glycoprotein I.
Rodgers et al. initially postulated that immune complexes might be involved in the process of endothelial cell activation, but were unsuccessful in detecting specific endothelial cell antigens by Western blotting with sera from preeclamptic women. However, using a more sensitive enzyme-linked immunosorbent assay (ELISA) method, Rappaport et al. observed elevated serum concentrations of antiendothelial cell antibodies in women with severe preeclampsia. It has been proposed that these antibodies interfere with endothelial cell function via inhibition of prostacyclin production or stimulation of procoagulant synthesis. However, prostacyclin production has been shown to be greater in cultured human endothelial cells incubated with plasma or sera from preeclamptic women than cells exposed to blood from normal pregnant controls. Preeclampsia plasma appeared to stimulate biosynthesis of a variety of prostanoids in several different cell types, including placental thromboxane synthesis, as suggested by Rote et al. to result in vasospasm and thrombosis. Peaceman and Rehnberg showed that the addition of 100 µM aspirin to IgG fractions from women with functional antiphospholipid antibodies inhibited thromboxane production by placental explants. Thus, autoantibodies may contribute functionally to immune cell activation and vascular and placental dysfunction associated with preeclampsia, and interference with these actions may be therapeutic.
Importantly, the identification of activating anti-AT1R autoantibodies (AT1-AA) adds a novel dimension to the potential role of immune complexes in preeclampsia. These immunoglobulins are found to be present in the serum of preeclamptic women at much higher concentrations than sera from nonpregnant or normal pregnant women. The antibodies bind and activate the AT1R, inducing AP-1 signaling and NF-κB activation. Reactive oxygen species, sFlt-1, and plasminogen activator inhibitor-1 are produced as a result. Their production and consequences are presented in detail in Chapter 15
Cytokines
Activation of the maternal immune system plays an important role in the development of preeclampsia. Excessive inflammation is central to this response and is believed to be a mediator of maternal endothelial dysfunction.
Activated immune cells including neutrophils, monocytes, NK cells, and T cells are also significantly elevated in women with preeclampsia ; leading to increased production of inflammatory molecules which further modulate the immune response. Elevated levels of many cytokines and chemokines have been identified in the maternal circulation of women with preeclampsia, including TNF-α, IL-6, IL-2, IL-8, IL-10, IP-10, MCP-1, and IL-12. Recent research shows that peripheral NK and T cells, although capable of producing VEGF, actually produce significantly less of this angiogenic factor under preeclamptic conditions. While the mechanism is unknown, this change further contributes to the angiogenic imbalance, promoting endothelial dysfunction. Many cytokines, particularly TNF-α, are known mediators of endothelial activation and dysfunction.
TNF-α has been postulated as an important mediator of endothelial cell activation in preeclampsia. Several studies have reported higher serum concentrations of TNF-α in preeclamptic women compared to normal pregnant women that normalize after delivery. Moreover, placentas from preeclamptic patients have increased mRNA and protein expression of TNF-α compared with placentas from normal pregnant women, implicating the placenta as a production source for this cytokine. However, based on uterine to peripheral vein TNF-α ratios, Benyo et al. suggested that the placenta is not the primary source of TNF-α in the maternal circulation and proposed that activated leukocytes and endothelium were responsible.
TNF-α alters the balance between endothelium-derived vasoconstrictors and vasodilators and impairs endothelium-dependent relaxation, in part, by activation of NAD(P)H oxidase, leading to production of superoxide anions that can scavenge NO. Moreover, TNF-α can induce vasoconstrictors such as ET-1, which are increased in preeclampsia patients. TNF-α can also stimulate mitochondrial production of free radicals. In addition, TNF-α increases the expression of endothelial adhesion molecules such as ICAM-1 and tissue factor, inhibits the thrombomodulin/protein C anticoagulation pathway, and blocks fibrin dissolution by stimulation of PAI-1.
Interestingly, TNF-α has been shown to have direct vascular effects during pregnancy. Chronic infusion of TNF-α into rats starting day 14 of gestation results in hypertension and decrease in renal iNOS expression. Furthermore, Vizi et al. reported that intraperitoneal injection of lipopolysaccharide, a potent stimulus for endogenous TNF-α production, induces greater TNF-α plasma levels and higher mortality in pregnant mice compared with non-pregnant animals. The data suggest that pregnancy is associated with a higher susceptibility to the effects of this cytokine, which may be a potential target for therapeutic intervention.
TNF-α also has been shown to activate type II phospholipase A2 (PLA2) and prostanoid biosynthesis in a variety of cultured cells. Thus, increased concentrations of PLA2 in placental homogenates from women with preeclampsia and elevated plasma levels of type II PLA2 observed in women with severe preeclampsia may be responses to increased TNF-α action in these women. Complicating these observations are reports that concentrations of soluble p55 TNF-α receptor, a TNF-α binding protein and potential inhibitor of TNF-α action, also are increased in preeclampsia serum.
Other cytokines, such as interleukin (IL)-2, a pleiotropic cytokine produced by activated T-lymphocytes, have potent autocrine and paracrine effects. Its ability to stimulate T cell, B cell and natural killer cell mitogenesis and activity have led investigators to believe that excessive IL-2 action may be involved in the pathogenesis of preeclampsia. It has been reported that normal maternal serum inhibits IL-2 production during pregnancy. Serum concentrations of the IL-2 receptor also have been evaluated as an indicator of preeclampsia risk.
Circulating plasma concentrations of IL-1 receptor antagonist, an inhibitor of IL-1 action, and IL-6 were noted to be increased in concentration by 77% and 24%, respectively, in women with preeclampsia compared to normal pregnant controls. It was suggested that these increased cytokine concentrations might contribute to the endothelial damage associated with preeclampsia. Other cytokines, including GM-CSF, G-CSF, IL-1, and IL-8, in maternal blood and amniotic fluid have been associated with preeclampsia and IUGR.
Circulating Lipids and Lipoproteins
Endothelial cell dysfunction may be caused by an imbalance between circulating VLDL particles or other lipids and a protective, basic isoform of plasma albumin (TxPA). One family with two cases of severe preeclampsia/eclampsia was found to have very high levels of Lp(a) lipoprotein. As the serum concentration of Lp(a) lipoprotein is genetically determined and the Lp(a) apolipoprotein has a close homology to plasminogen, very high levels of Lp(a) lipoprotein might interfere with the fibrinolytic/thrombolytic processes and could represent a genetically determined risk factor for preeclampsia.
Triglycerides are a major feature of metabolic syndrome, which is increasing due to the rising number of reproductive-aged women who are obese. In a systematic review by Ray et al., they determined that there was a consistent positive association between elevated maternal triglycerides and the risk of preeclampsia. Interestingly, normotensive pregnancies are characterized by progressive increases in serum LDL levels accompanied by parallel increases in serum triglyceride levels followed by sharp fall post-partum. Interestingly, increased triglyceride content of LDL results in smaller but denser LDL particles. These small dense LDL particles are more atherogenic and are more susceptible to oxidative modification. Sattar et al. first reported that smaller, denser LDL particles are increased in preeclamptic plasma when compared to normotensive controls which was subsequently confirmed by Hubel. In contrast to the above studies, one recent study has shown that plasma levels of oxLDL are in fact decreased in preeclampsia compared to age-matched normotensive pregnancy. The authors suggest that the decreased plasma oxLDL levels in preeclamptic women could be due to increased levels of autoantibody to oxLDL. A few studies have reported increased serum levels of autoantibodies to oxLDL while some studies have found no differences. As with many of the circulating factors measured in women with preeclampsia, differences in the literature may be attributed to differences in gestational age at sampling.
Oxidized LDL binds to its lectin-like oxidized LDL receptor (LOX-1) receptor and contributes from 50% to 70% of oxLDL uptake in endothelial cells, resulting in generation of superoxide anion via activation of NAD(P)H oxidase. While basal levels of LOX-1 expression are low, factors such as TNF-α, transforming growth factor-β, oxLDL, Ang II, ET-1, C-reactive protein, and 8-iso-prostane have all been shown to upregulate LOX-1 expression. Since all of these factors have been reported to be elevated in plasma of women with preeclampsia enhanced LOX-1 expression and uptake of oxLDL mediated via LOX-1 could be important in endothelial cell activation and/or dysfunction. LOX-1 mRNA expression showed highest expression in human placenta when compared to other human tissues. Halvorsen et al. identified the presence of LOX-1 receptor in trophoblastic cells and demonstrated that 8-iso-PGF2α, a product of lipid peroxidation and a marker of oxidative stress, could enhance oxLDL uptake via LOX-1 using a choriocarcinoma cell line. The Davidge laboratory recently demonstrated increased LOX-1 expression in the maternal vasculature of women with preeclampsia. They have further identified a decrease in endothelium-dependent relaxation in the vasculature of the reduced uteroplacental perfusion pressure (RUPP) model of preeclampsia that was accompanied by increased levels of LOX-1 and eNOS. Moreover, plasma from women with preeclampsia impaired endothelial-dependent vascular function that was prevented by LOX-1 inhibition. Taken together, these studies would further support the role of dyslipidemia in the pathophysiology of preeclampsia.
Non-Esterified Fatty Acids
Non-esterified fatty acids (NEFA) are organic acids containing an even number of carbon atoms. They are produced in the small intestine when fat is digested. Chains with only single bonds are “saturated” with hydrogen (e.g., palmitic and stearic acids) whereas chains with one or more double bonds are “unsaturated” (oleic (one double bond), linoleic (two double bonds), and linolenic (three double bonds)).
Several studies indicate that plasma NEFA are increased 2–3-fold in women with preeclampsia, even early in gestation and when controlled for race, BMI, and age. Interestingly, pregnant women with transient hypertension had plasma NEFA levels significantly lower than preeclamptic patients with similar degrees of hypertension. These observations reiterate that the pathophysiology of preeclampsia is more than pregnancy-induced hypertension.
NEFA levels in the plasma of pregnant women were found to be inversely proportional to a basic isoelectric isoform of albumin (p I =5.6) referred to as toxicity-preventing albumin (TxPA). TxPA was so named because it could prevent lipid-induced endothelial cell injury in vitro and in vivo . This relationship is consistent with the hypothesis that fatty acid binding to plasma albumin is responsible for its shift in p I . Vigne et al. directly quantified the amount of fatty acid bound to plasma albumin in normal pregnant and preeclamptic women, finding a 3-fold increase in the NEFA/albumin ratios (2.5 vs. 0.8) in severely preeclamptic vs. normal pregnant women. These values are very similar to the 1.6 and 0.9 molar ratios previously estimated by Endresen and colleagues. Albumin with<1.3 mol NEFA bound focuses at p I 5.6, whereas albumin with 2.5–10 mol NEFA focuses at p I 4.8, resulting in a conformational change that alters the redox activity and buffering capacity of plasma albumin. Recapitulating the elevated NEFA/albumin ratios observed in preeclamptic women demonstrated lipid accumulation and apoptosis in human endothelial cell cultures, apparently as a result of alterations in mitochondrial membrane potential. Plasma albumin plays an important role in safeguarding against the prooxidant effects of copper in the circulation through its Cys-34 residue, which directly participates in free radical scavenging. Modifications of Cys-34, resulting in the loss of its electron-donating SH-group, disrupt its antioxidant capacity. Kagan et al. have reported that the copper-binding capacity of albumin is impaired in women with preeclampsia. The high concentration (~600 μM) of this protein in the circulation and perpetual catalytic nature of the redox-cycling processes make these reactions a powerful source of oxidative stress. These data indicate that reducing antioxidants, such as ascorbate, cannot effectively prevent or ameliorate oxidative stress induced by Cu/albumin, as the continuous redox-cycling process depletes any exogenously added ascorbate and does not remove the source of redox-cycling. This observation may explain, in part, the failure of therapeutic trials of antioxidant vitamins to improve clinical outcomes in preeclampsia as discussed later in this chapter.
Among the essential fatty acids described above, linoleic acid (18:2 n −6) is necessary for the production of important long-chain polyunsaturated derivatives, particularly arachidonic acid (20:4 n −6). Not only is arachidonic acid a key structural element of cell membranes, it is the obligate precursor to a variety of vasoregulatory lipids. Cyclooxygenase and peroxidase catalyze the formation of prostaglandins, prostacyclin, and thromboxanes, whereas 5-lipoxygenase allows the production of leukotrienes. A third pathway is catalyzed by cytochrome P-450 epoxygenases, converting the substrate to epoxyeicosatrienoic acids (EETs). The latter are potent vasodilators and antiinflammatory compounds that can be inactivated by soluble epoxide hydrolase (sEH) to corresponding dihydroxyeicosatrienoic acids (DHETs). In a recent pilot study, urinary concentrations of total 14,15-DHET were found to be 2.6-fold higher in seven preeclamptic women at term, compared to nine uncomplicated, normotensive pregnant women ( P <0.02). The implication is that excessive catabolism of EETs in preeclampsia may contribute to reduced vasorelaxation and increased inflammation. These arachadonic acid metabolites also are known ligands for PPARs.
Peroxisome Proliferator-Activated Receptors (PPARs)
PPARs are major regulators of lipid and glucose metabolism, inflammation, and angiogenesis, which allow adaptation of the mother to the nutritional and perfusion requirements of the fetus. PPARs, members of the nuclear hormone receptor superfamily, are ligand-activated transcription factors. There are three PPAR isotypes, PPARα, PPARγ, and PPARβ/δ, which are structurally conserved across species.
The PPAR system is intimately involved in adipose and vascular tissues, and hence is highly relevant to preeclampsia. All three PPAR isotypes are expressed in human vascular and placental trophoblast cells. These receptors can be activated by natural ligands, e.g., prostaglandins (PGs), fatty acids and their derivatives, as well as by synthetic ligands. A number of naturally occurring PPAR ligands have been identified, including long-chain fatty acids (C16 and greater), eicosanoids such as 8(S)-HETE (PPARα) and 9- and 13-HODE (PPARγ), and PGs such as PGA1, which binds to PPARα and PPARβ/δ, and 15-deoxy-delta12, 14-prostaglandin J2 (15dPGJ2), which binds to PPARγ.
As preeclampsia is marked by hyperlipidemia, and characterized by a state of oxidative stress (see below), it has been postulated that pathological changes may be regulated by the PPAR system. The Taylor laboratory found that serum from women with severe preeclampsia had reduced levels of PPAR activating lipids compared with serum of parity and gestational age-matched, normal pregnant women. The reduction of transcriptional activity observed in trophoblast cells by sera of women with preeclampsia was shown for PPARγ and PPARα, but not for PPARβ/δ or RXR, and was manifested weeks and sometimes months before the clinical diagnosis was made. These results are consistent with other clinical evidence of a “hyperinflammatory” state of preeclampsia. It has been shown that the PPARγ agonist, rosiglitazone, ameliorated hypertension, improved vascular function and reduced the elevated microalbumin:creatinine ratio in pregnant rats that had undergone RUPP surgery. Some of these beneficial effects were abrogated in the presence of the heme oxygenase 1 inhibitor as a potential mechanism. Moreover, administration of a PPARγ antagonist induced features of preeclampsia (hypertension, proteinuria, and endothelial dysfunction) in healthy rats.
Genetic variations in the PPARγ gene may modify the risk to develop preeclampsia. The Pro467Leu mutation of PPARγ is a dominant negative mutant resulting from a C to T transition in exon 6. In a woman with this mutation, both of her two pregnancies were complicated by severe preeclampsia. However, a preliminary study of PPARγ gene variations in a Finnish population showed no association with preeclampsia or its severity.
Angiogenic Factors
That angiogenic factors play a critical role in the development of preeclampsia is now well accepted and this topic is reviewed comprehensively in the accompanying Chapter 6 . Vascular endothelial growth factor (VEGF) belongs to a family of glycoproteins that express their primary biologic effects by binding to VEGF receptor-1 (fetal liver tyrosine-like; Flt-1) and VEGF receptor-2 (kinase domain related receptor; KDR). VEGF promotes angiogenesis, induces endothelial cell production of vasodilatory NO and prostacyclin, and may be important in the regulation of vascular tone. Although VEGF has vasodilatory activity, some studies indicate detrimental effects following intravascular administration of this factor. VEGF injection increased ICAM-1 and VCAM-1 expression, leukocyte adhesion, vascular leakage, and inflammation. It is tempting to speculate that such proinflammatory actions of VEGF could be involved in the pathogenesis of endothelial dysfunction in preeclampsia.
Brockelsby et al. showed that incubation of myometrial resistance vessels with VEGF resulted in a blunted endothelium-dependent relaxation, mimicking the endothelial dysfunction observed in preeclampsia. By contrast, several studies suggest that low serum concentrations of VEGF-family angiogenic proteins are associated with preeclampsia. In view of the latter hypothesis it is interesting to note that the dose-limiting toxicities of bevacizumab (Avastin®), humanized anti-VEGF monoclonal antibodies used to treat colon, breast and other carcinomas, include hypertension and proteinuria.
Some of the conflicting reports regarding the role of VEGF in the pathophysiology of preeclampsia have been complicated by technical aspects of the available assays as well as differences in free vs. total VEGF measurements. Most investigators have failed to detect free VEGF in the maternal circulation. In a careful study of women undergoing IVF therapy, free VEGF was undetectable in the circulation within a month following embryonic implantation. Soluble Flt is a splice variant of the Flt-1 receptor that lacks transmembrane and cytoplasmic domains. sFlt-1 is produced by a number of organs including the placenta, lung, liver, kidney, and uterus. It acts as a potent VEGF and placental growth factor (PlGF) antagonist with an important role in preeclampsia pathophysiology. Maynard et al. performed gene expression profiling of placental tissue and demonstrated upregulated mRNA expression of sFlt-1 in preeclampsia and subsequently showed elevated serum levels of sFlt-1 levels in preeclamptic women, which has been confirmed by other investigators. There is a positive correlation between the severity of preeclampsia and serum sFlt-1 levels. Also, sFlt-1 levels vary inversely with serum VEGF and PlGF levels, an effect attributed to its adsorption of these factors. However, one study has shown decreased PlGF levels during the first trimester without an increase in sFlt-1 levels, suggesting that decreased PlGF levels observed in PE could also arise as a result of inadequate placental production. There is a strong association between decreased urinary PlGF and the subsequent development of preeclampsia. Several research groups are currently developing algorithms to use circulating sFlt-1 and PlGF concentrations as predictors of preeclampsia risk.
Abnormal expression of platelet-derived growth factor (PDGF), another vasculogenic factor primarily responsible for pericyte recruitment and new vessel maturation, also has been implicated in preeclampsia.
As noted above, endoglin is an extracellular component of the TGF-β receptor II complex. It is expressed in high concentrations by endothelial cells and syncytiotrophoblasts. Its primary function in endothelial cells appears to be in the regulation of TGF-β receptor II serine/threonine kinase activity. Endothelial cells that lack endoglin fail to migrate or proliferate. Mutations in endoglin have been associated with hereditary hemorrhagic telangiectasia (Weber-Osler-Rendu syndrome) and elevated serum concentrations of soluble endoglin are associated with preeclampsia.
Matrix Metalloproteinases
Matrix metalloproteinases (MMPs) are a large family of zinc-dependent endopeptidases expressed ubiquitously throughout the body. In vivo , MMP activity is regulated by the four tissue inhibitors of MMPs (TIMPs 1 through 4), which bind and inhibit the zinc-catlaytic domain of these enzymes (reviewed in ). An imbalance between MMP activity and TIMP inhibition has long been recognized to allow degradation of extracellular matrix, and in this manner contribute to vascular remodeling in both long-term physiological processes (e.g., embryogenesis ) and chronic pathological processes (e.g., heart failure, atherosclerosis ). More recently, however, MMPs have also been shown to cleave non-extracellular matrix proteins and contribute to acute processes such as vascular reactivity and platelet aggregation. Since the latter processes contribute to the pathology of preeclampsia, the potential role of MMPs in preeclampsia warrants attention.
MMP-2 may be of particular relevance in preeclampsia since its circulating levels have been found to be elevated in preeclamptic women compared to normal pregnant women. Myers et al. demonstrated that the MMP-2 to TIMP-1 ratio is imbalanced in women who develop preeclampsia. This imbalance preceded the onset of the condition and continued up to the time of delivery. Kolben et al. also found that circulating TIMP-1 is decreased in preeclampsia. The circulating levels of other TIMPs in preeclampsia still remain to be determined.
In vitro studies also suggest an imbalance between MMPs and TIMPs in preeclampsia. Human umbilical vein endothelial cells from preeclamptic pregnancies release more MMP-2 than endothelial cells from normal pregnancies. In this case, the increase in MMP-2 activity occurred with no change in TIMP-1 or -2 protein levels, predicting a net increase in proteolysis. These in vitro experiments also indirectly suggest that the source of increased circulating MMP-2 in preeclampsia may be the dysfunctional endothelium.
A number of studies have demonstrated that many of the circulating factors already discussed in this article increase MMP activity. For instance, VEGF was found to markedly increase MMP-2 release from endothelial cells in a time- and concentration-dependent manner. Other experiments indicate that peroxynitrite, TNF-α, Ang II, and oxLDL also can activate MMPs in vitro . Thus, it is possible that multiple molecular mediators of preeclampsia may converge to promote MMP production and release in preeclampsia.
The functional significance of elevated MMP-2 may lie in accumulating evidence suggesting that this enzyme can promote vasoconstriction. The Davidge laboratory demonstrated that MMP-2 can cleave big ET-1 to yield ET(1–32), a 32-amino-acid vasoconstrictor peptide. Recently, using the RUPP model, vasoconstriction to big ET was greater in the RUPP animals compared to normal pregnancy and MMP inhibition normalized this enhanced vasoconstriction response.
Endothelin
Endothelin is an endothelial cell marker of interest in preeclampsia. Of the three isoforms of endothelin, ET-1 is the predominant one produced in endothelial cells and exerts its biological actions in a paracrine fashion on subjacent vascular smooth muscle cells. Elevated ET-1 concentrations have been observed in clinical conditions associated with decreased intravascular volume such as that seen in preeclampsia. Moreover, ET-1 is a potent vasoconstrictor of the human uterine and renal vascular beds, both of which are affected in this syndrome.
There are two endothelin receptor isoforms: ETA and ETB. ETA has a selective affinity for ET-1, is found primarily on vascular smooth muscle cells, and is thought to mediate vasocontriction. ETB, which binds both ET-1 and ET-3 ligands with a similar affinity, is found on endothelial cells and may be involved in the autoregulation of NO and PGI2 release. In the human placenta, the ETB receptor isoform predominates, theoretically allowing vasodilatation of this important vascular bed.
Circulating concentrations of immunoreactive ET-1 have been shown to be elevated in women with overt signs of preeclampsia ; however, no differences were observed in ET-1 levels during the second trimester of pregnancy (several months prior to the onset of clinical symptomatology), compared to a matched, normal pregnant group. Elevated concentrations of ET-1 in the blood of women with preeclampsia have been confirmed by other groups, although Benigni et al. failed to find significant differences between women with preeclampsia and normal pregnant women. Although the results do not support an early, predisposing role for ET-1 in preeclampsia etiology, as a potent vasoconstrictor it likely has a role in increased vascular resistance and reduced perfusion to organs. This is supported by studies in the RUPP model whereby ET has a central downstream role for vasoconstriction. Thus, it has been suggested that manipulations of ET receptor may afford new strategies to control blood pressure and placental ischemia in preeclampsia.
Relaxin
In recent years, a definite link has been established between relaxin and NO in mediating the vasodilatation of pregnancy. Relaxin is typically considered a reproductive tract hormone active in mammalian parturition. However, new research indicates additional physiological roles for relaxin in the cardiovascular, renal and respiratory systems. Strong arguments to implicate the ETB receptor and NO pathway in relaxin action are presented in Chapter 16 .
Angiotensin II
Angiotensin II (Ang II) is well known for its vasoconstrictive properties. Since preeclampsia is a state of enhanced vasoconstriction and hypertension, one would expect higher Ang II levels in the plasma of women with preeclampsia. However, circulating Ang II levels in preeclamptic women have been reported to be similar or even decreased relative to normal pregnant controls. Thus, enhanced expression of the Ang II receptor (AT1R) has been invoked to explain the increased vascular tone in preeclampsia. Indeed, AT1R signaling in endothelial cells has been shown to result in cytokine production, increased extracellular matrix activity, and generation of reactive oxygen species through direct activation of endothelial NAD(P)H-oxidase. Interestingly, AT1R activation induces the production of TNF-α from endothelial cells leading to increased MMP-2 and decreased TIMP-2 release.
It has been reported that women with preeclampsia have increased expression of AT1R (specifically AT1R platelet binding sites), in comparison with normal pregnant women. Abdalla et al. found that increased formation of heterodimers between the bradykinin receptor and AT1R in preeclampsia resulted in enhanced Ang II signaling. These data support a role for AT1R levels in the alterations in Ang II sensitivity in preeclampsia that were observed in the seminal study by Gant et al. Their study demonstrated that normal pregnancy was associated with decreased sensitivity to the pressor effects of Ang II, whereas women who later developed preeclampsia showed increased sensitivity to Ang II starting at 22 weeks of gestation. Moreover, agonistic autoantibodies for AT1R (AT1-AA) have been measured in the circulation of women with preeclampsia that may account for enhanced AT1R activation in the absence of elevated Ang II levels in preeclampsia. These are discussed in detail in the accompanying Chapter 15 .
As mentioned briefly above, the discovery of circulating, agonistic autoantibodies directed against the AT1R represents a remarkable pathophysiological derangement in preeclampsia. The identification of these activating antibodies serves to align immunological theories with the rennin-angiotensin system. Similar effects were noted in a rat model of preeclampsia.
Toll-Like Receptors
Toll-like receptors (TLRs) modify the innate immune system via single membrane-spanning proteins that recognize ligands with structurally conserved molecular patterns. Pathogenic ligands of TLRs include bacterial cell lipopolysaccharides, viral double-stranded RNA, and fragmented host molecules such as fibrinogen and DNA with unmethylated CpG islands. These receptors are abundantly expressed in trophoblast and endothelial cells. TLR9, for example, has been identified in HUVE cells, where it transduces inflammatory changes in response to free, particularly hypomethylated, DNA. It has been demonstrated that concentrations of free fetal DNA in the maternal circulation may be more than 3-fold increased relative to the normal pregnant state, presumably as a result of placental villus shedding and widespread placental apoptosis of anchoring cytotrophoblasts. Furthermore, as placental DNA is globally hypomethylated, this is likely to be a potent stimulus to maternal endothelial TLR9, activating the downstream NF-κB pathway and leading to adhesion molecule expression and vascular permeability.
Part III: Oxidative Stress: A Point of Convergence for Endothelial Cell Dysfunction
As noted, circulating factors such as TNF-α, oxLDL, and AT1-AA acting on the vasculature result in oxidative stress, inflammation, and vasoconstriction, features that have been well documented in preeclampsia. Moreover, these factors can act individually and in concert to disturb maternal endothelial function. The following section will provide evidence to support the hypothesis that intracellular oxidative stress promotes endothelial dysfunction in preeclampsia.
Oxidative Stress as a Mediator of Endothelial Cell Dysfunction
Oxidative stress is an imbalance between prooxidant and antioxidant forces resulting in an overall oxidant insult. In preeclampsia, oxidative stress has been postulated to lead to altered endothelial cell function. Indeed, the prooxidant environment of endothelial cells is extensive, as these cells are constantly exposed to extracellular factors in the circulation that are capable of inducing an oxidative insult and can produce their own oxidants as well. Prooxidants include free radicals such as superoxide anions (O 2 · − ), hydroxyl radicals (OH·), nitric oxide (NO − ) and other reactive oxygen species (ROS) as well as reactive nitrogen species (RNS) such as peroxynitrite anion (ONOO − ). ROS and RNS are continuously produced in vivo by a number of cell types including endothelial cells. Oxidative stress is an example of a pathological process whereby multiple factors converge to cause endothelial cell dysfunction (see Fig. 9.1 ).
Reactive oxygen species contribute to normal metabolic processes in the body and have physiologic roles as second messengers in addition to their potentially pathogenic role. There are several metabolic pathways that can lead to the production of oxygen-derived free radicals. Mitochondria, endoplasmic reticulum, and nuclear membranes have been shown to produce superoxide anions as a consequence of autooxidation of electron transport chain components. Oxygen-derived free radicals also are produced as the result of arachidonate metabolism by prostaglandin H (PGH) synthase, lipoxygenase, and cytochrome P450. NO synthase can generate superoxide anions and hydrogen peroxide, particularly if the intracellular concentrations of l -arginine (the precursor for the synthesis of NO) or its co-factor tetrahydrobiopterin are low. Oxidation of hypoxanthine by xanthine oxidase also produces superoxide radicals within endothelial cells. It is important to note that oxygen-derived free radicals can be produced at a number of subcellular compartments, such as the mitochondria, endoplasmic reticulum, peroxisomes, phagosomes, plasma membrane, nuclear membrane, and cytoplasm. Therefore, the ability of antioxidants to quench free radicals depends on the production, localization as well as the type of oxidative insult.
Several factors in the circulation (e.g., neutrophils, xanthine oxidase, and cytokines) can act on endothelial cells to produce oxygen free radicals. This raises the possibility of free radical-dependent vascular dysfunction at sites distant from the primary source or insult. This is important since the placenta is central for the development of preeclampsia. As a consequence of endothelial cell dysfunction, the perfusion of many organs, including the placenta, is reduced. This, in turn, could lead to a feed-forward progression for further endothelial cell dysfunction thereby accelerating the symptoms of preeclampsia until the placenta has been removed.
In women with preeclampsia, superoxide generation from circulating neutrophils is enhanced. As previously noted, cytokines such as TNF-α can either directly or indirectly initiate oxidative stress. TNF-α, which is elevated in women with preeclampsia, has been shown to directly induce oxidative damage as well as increase endothelial cell-induced oxidation of LDL. TNF-α can also increase free radical production through the xanthine oxidase pathway and concentrations of the latter enzyme are increased in women with preeclampsia (see Fig. 9.1 ).
NAD(P)H oxidase appears to represent the most significant source of superoxide anion production in endothelial cells. NADPH oxidase is a complex membrane-integrated b-type cytochrome, cytochrome b558, which is composed of 91 and 22 kDa subunits (gp91phox and p22phox, respectively), and at least three cytosolic proteins (p47phox, p67phox, and p21rac). As previously noted, many of the circulating factors that are increased in women with preeclampsia can activate NADPH oxidase.
Free radicals thus generated have numerous deleterious downstream effects including lipid peroxidation of biological membranes and oxidative modification of lipoproteins. Some markers indicate that in women with preeclampsia, there is evidence for increased lipid peroxidation, whereas Regan et al. disputed this claim. However, their study only analyzed a single biochemical marker of oxidative stress, urinary 8,12-iso-iPF2alpha-VI, and hence their report should be interpreted with caution. Also, this marker has not been widely validated for oxidative stress. A subsequent study examined various biomarkers of oxidative stress in preeclampsia and concluded that preeclampsia is indeed a state of mild oxidative stress. Decreased plasma vitamin C levels and significantly elevated lipid hydroperoxides, indicative of oxidative stress, as well as elevated serum levels of soluble vascular cell adhesion molecule-1 (VCAM-1), all are consistent with ROS-induced endothelial dysfunction.
In preeclampsia, it has been suggested that certain antioxidant mechanisms are not adequate to compensate for an overwhelmed oxidant response. Re-establishing a balance between ROS and antioxidant protection in women with preeclampsia may protect the vascular endothelium. Unfortunately, this is a complex system and thus the results of antioxidant vitamin trials are equivocal, as discussed later in this chapter. Ultimately, antioxidant protection will depend on the compartment and type of oxidative insult imposed on a cell. Moreover, oxidative stress in preeclampsia is most likely the result of multiple prooxidant intracellular pathways. Understanding of the ultimate converging pathways leading to vascular oxidant stress and the downstream effects is needed so that rational interventions can be developed. Nonetheless, there is substantial evidence for elevated oxidants within the vascular endothelium of women with preeclampsia that can alter a number of vasoactive pathways, specifically involving endothelial-dependent relaxing pathways such as prostacyclin, NO, and endothelial-derived hyperpolarization.
Eicosanoid/Prostacylin Production
The production of eicosanoids, such as prostacyclin and thromboxane, is regulated by the availability of arachidonic acid and the activity of PGH synthase. Liberation of arachidonate from membrane phospholipids is mediated through phospholipases. The primary pathway is through activation of PLA2, which releases arachidonate directly from membrane phospholipids. Elevated lipid peroxides are known to increase PLA2 activity and play an important regulatory role in PGH synthase activity. The activity of PGH synthase requires low levels of lipid peroxides for activation and continued catalysis whereas higher levels of peroxides are inhibitory and can inhibit prostacyclin synthase.
Prostacyclin (PGI2) is known to cause vasorelaxation and inhibit platelet aggregation by receptor-mediated mechanisms. While cyclic (c)AMP acts as a second messenger for platelet aggregation, vasorelaxation by hyperpolarization may provide an explanation, in addition to stimulation of cAMP, for the PGI2 mechanism of action on blood vessels. PGI2 released from the endothelium is capable of relaxing vascular smooth muscle. PGI2 has a role in determining vasodilatation in different vascular beds, especially in relation to sex steroid status and in pregnancy.
In a rat model of oxidative stress (produced by depletion of vitamin E), increased PGH synthase-dependent vasoconstrictor modification of vascular responses was observed. Further, in rat models of hypertension (also believed to reflect a state of oxidative stress), a PGH synthase-dependent vasoconstrictor predominates to alter vascular function. However, one research group has studied the effect of indomethacin (an inhibitor of PGH synthase activity) on the acetylcholine- and bradykinin-induced relaxation response of subcutaneous arteries from preeclamptic and normal pregnant women. Indomethacin attenuated the relaxation response, but the degree of shift was not different between normotensive women and preeclamptic women.
Data suggest that there is an overall increase in the vasoconstrictor eicosanoids (as measured by their stable metabolites) in women with preeclampsia. For example, in the placentas of women with preeclampsia, more thromboxane is produced compared to prostacyclin. Urinary and plasma levels of thromboxane are elevated while prostacyclin levels are reduced in women with preeclampsia compared to women with uncomplicated pregnancies. However, the system is complex. Biphasic temporal patterns of PGI2 production in response to endothelial activators have been described. Wang et al . noted an initial increase followed by a persistent diminution of PGI2 production when bovine endothelial cells were exposed to hyperlipidemic sera. Similar kinetics of PGI2 release by HUVE cells were reported by Baker et al., after exposure to preeclampsia plasma. During the first 48 hours the secretion of PGI2 was enhanced over baseline, whereas more chronic exposure (>72 h) appeared to deplete PGI2 production.
Nitric Oxide
One mechanism by which oxygen free radicals may alter endothelial cell function is by scavenging NO, thereby reducing bioavailability of this potent vasodilator. Normal pregnancy is maintained in a state of vasodilatation, mediated in part by enhanced NO production. Nitric oxide synthase (NOS) catalyzes the conversion of l -arginine to l -citrulline and NO, a powerful vasorelaxing molecule. While a decrease in NO modulation of vascular tone could partly explain the pathogenesis of hypertension in preeclampsia, there are conflicting reports of increased, decreased or sometimes unchanged metabolites of NO in the maternal circulation. Decreased plasma levels of nitrite and nitrate have been attributed to increased plasma levels of S -nitrosoalbumin in preeclampsia. Elevated concentrations of non-esterified fatty acids alter the conformation of serum albumin and affect the redox potential of a critical cysteine residue in the protein. These changes impair the release of NO, ultimately resulting in reduced NO bioavailability. This decreased release of NO could also be partly due to decreased vitamin C levels, since S -nitrosoalbumin-mediated relaxation of resistance sized-vessels is enhanced by ascorbate.
Another pathway that may result in reduced NO levels is regulated by arginase. The two isoforms of this enzyme, arginase I and II, are found in cytosol and mitochondria, respectively. They catalyze the conversion of l -arginine to l -ornithine and urea. Arginase reciprocally regulates NOS, hence an increase in arginase activity could decrease l -arginine levels, the common substrate for arginase and NOS, and result in decreased NO production. Furthermore, a decrease in the substrate could result in uncoupling of endothelial NOS (eNOS) and generate superoxide anion. Arginase mRNA expression is increased in villous tissue of preeclamptic women, which inversely correlates with fetal l -arginine levels, suggesting excess consumption of the substrate. In preeclamptic women, arginase II expression, as assessed by immunohistochemistry, was found to be increased in syncytiotrophoblasts and endothelial cells of villous vessels of preeclamptic placenta. In regard to the maternal vasculature, arginase II expression is increased in arteries from omental fat biopsies of women with preeclampsia compared to normotensive pregnant women. Furthermore, HUVE cells treated with 2% plasma from preeclamptic women show increased arginase II expression and activity. Preeclamptic plasma also increased superoxide and peroxynitrite levels; however, inhibition of arginase or NOS reduced superoxide levels. Thus, arginase may result in eNOS uncoupling, contributing to oxidative stress by reducing the production of NO and promoting the production of superoxide. In addition, enhanced circulating asymmetric dimethylarginine (ADMA), an endogenous inhibitor of eNOS, has been observed in women with preeclampsia as yet another mechanism for reduced NO bioavailabilty in the maternal vasculature in preeclampsia.
Interestingly, endothelial cells acutely exposed to preeclamptic plasma showed an increase in NOS activity, enhanced eNOS protein expression, and stimulated nitrite/nitrate production. Although NO is an important vasorelaxant, an elevation of NO in the face of oxidative stress may be deleterious. NO can react with superoxide anions yielding the powerful oxidant peroxynitrite, which may alter vascular function. Myatt et al. have demonstrated increased peroxynitrite in the placentas of women with preeclampsia, while the Davidge laboratory demonstrated enhanced peroxynitrite formation in the vessels of women with preeclampsia. Moreover, circulating factors in women with preeclampsia can enhance peroxynitrite formation in cultured endothelial cells. These data imply that greater NO scavenging within the endothelium occurs in preeclampsia. It is interesting to note that decomposition of peroxynitrite at physiological pH also gives rise to nitrates and nitrites and thus may be a source of the elevated NO metabolites that have been observed in endothelial cells exposed to preeclamptic plasma. Peroxynitrite is not only an indicator of NO scavenging, but it also affects endothelial function. Peroxynitrite increases inducible NOS (iNOS) protein expression in isolated endothelial cells via activation of NF-κB. Thus a peroxynitrite-induced increase in NO could be deleterious in the face of increased oxidative stress, whereby there is a feed-forward mechanism to generate more peroxynitrite. Furthermore, peroxynitrite decreases prostacyclin synthase expression, which could alter the balance towards vasoconstriction.
Interestingly, it seems that in normal pregnancy there is a compensatory pathway that causes relaxation when NO and prostacyclin are inhibited, which may not occur in the vasculature of women with preeclampsia. This relaxation is mediated through endothelium-derived hyperpolarization.
Endothelium-Derived Hyperpolarization
Endothelium-dependent vasodilatation is primarily mediated by NO, prostaglandins, and endothelium-derived hyperpolarization factor (EDHF). However, the exact nature of EDHF is complex as there are multiple pathways or factors that result in hyperpolarization of the endothelial and/or smooth muscle cells thereby resulting in vasodilatation. Thus the current literature simply refers to this relaxation as EDHF with a number diverse factors/mechanisms involved, such as epoxyeicosatrienoic acid (EETs), cannabinoids, potassium ions, myoendothelial gap junctions, and hydrogen peroxide, which have all been identified as putative EDHF in various species and vascular beds.
Hyperpolarization generated in endothelial cells spreads to adjacent vascular smooth muscle cells through myoendothelial gap junctions. Moreover, calcium-activated potassium channels, most probably the SK4 (IKCa) and SK3 (SKCa) expressed either on the endothelium or on vascular smooth muscle cells, are the end-cellular gateway mediating hyperpolarization and subsequent EDHF-mediated relaxation. EDHF mechanisms have been reported to account for ~50% of endothelium-dependent vasodilatation in subcutaneous resistance arteries from healthy pregnant women and are dependent on Cx43 and gap junctions. In women with preeclampsia, decreased vasodilator responsiveness in myometrial arteries has been attributed to a reduction in EDHF contribution, potentially due to physical disruption of myoendothelial gap junctions that was also characterized by a diversification of EDHF mechanisms that includes EET and H 2 O 2 . Notably, an increase in superoxide anions that has been demonstrated in compromised pregnancies can produce H 2 O 2 by the enzyme superoxide dismutase, thus providing a potential explanation for the increased involvement of this molecule. Further, PlGF, an activator of EDHF, has been consistently shown to be reduced in women with preeclampsia. Thus EDHF via gap junctions and connexins are important mediators of pregnancy-induced vasodilatation that is impaired in women with preeclampsia.
In summary, numerous mechanisms act in concert to permit the vasodilatation featured in normal pregnancy, which is crucial for a successful outcome. Ultimately, there is reduced endothelium-dependent relaxation in the vasculature of women with preeclampsia. Enhanced generation of superoxide anion could scavenge NO, and this coupled with an absence of a compensatory pathway such as EDHF may lead to the symptoms observed in women with preeclampsia. Overall, it is likely that oxidative stress is a point of convergence for multiple factors affecting endothelial-mediated vasoactive pathways.
Other Vasodilators
Carbon monoxide and hydrogen sulfide are other vasodilators implicated in the pathophysiology of preeclampsia. Carbon monoxide is produced endogenously from hemeoxygenase-1 (HO-1), which is significantly decreased in placentas of women with preeclampsia. Women with preeclampsia have significantly decreased CO concentration in exhaled breath compared to normal healthy women. It is interesting to note that maternal cigarette smoking is associated with a lower risk of preeclampsia that may be due, in part, to CO (reviewed in ). In support of this hypothesis, Zhai et al. demonstrated an inverse association between CO concentration and preeclampsia risk.
In addition to NO and CO as gaseous mediators of vascular function, hydrogen sulfide (H 2 S) is another more recently identified gaseous mediator. It is produced endogenously from the activity of two enzymes, cystathionine-γ-lyase (CSE) and cystathionine-B-synthetase (CBS). CSE is the primary H 2 S-synthesizing enzyme in the vasculature. It was recently demonstrated that H 2 S is a powerful vasodilator of the placental vasculature and that expression of CSE is reduced in placentas from women with preeclampsia displaying abnormal umbilical artery Doppler waveforms. Moreover, plasma levels of H 2 S are significantly decreased in women with preeclampsia. Overall, it has been suggested that endogenous H 2 S is important for a healthy placental vasculature and may be impaired in women with preeclampsia.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
