Placental Vasculogenesis

During embryonic development, the blastocyst differentiates into two cell populations: an outer polarized trophoectoderm and the non-polarized inner cell mass. The trophoectoderm gives rise to extra-embryonic membranes, whereas the inner cell mass is destined to form the embryo proper.

Extra-embryonic mesenchymal cells give rise to cores that penetrate into the center of the cytotrophoblast columns. These mesenchymal cells differentiate into endothelial cells, forming the first capillaries of the placental vasculature. In humans, placental vasculogenesis is evident by approximately 21–22 days post-conception. At this stage, cords of hemangiogenic cells are present and some demonstrate primitive lumen formation. These cords further develop so that by approximately day 32 post-conception, most villi show the presence of capillary structures. The highly proliferative trophoblasts ultimately invade the entire endometrium, the outer one-third of the myometrium, and the maternal circulation. Hypoxia is an important driving force for trophoblast proliferation.

As trophoblasts invade the uterine vasculature, they are exposed to increasing concentrations of oxygen, at which time they exit the cell cycle and differentiate. Further development leads to the penetration of cytotrophoblastic cones into the syncytiotrophoblastic mass and the development of lacunae which eventually become the intervillous space. Continuing growth and differentiation of the trophoblasts leads to branching of the trophoblast villi and the shaping of a placental labyrinth or intervillous space, where the fetal/maternal exchange of oxygen and nutrients occurs. Trophophoblasts invade maternal tissue with a variable depth of invasion among species, the deepest known being in humans.

Maternal Vascular Remodeling

The formation of adequate maternal–placental circulation requires remodeling of maternal blood vessels (namely, the spiral arteries). In humans, during the mid-late first trimester, the trophoblasts invade deeply through the endometrium and into the superficial part of the myometrium, completely remodeling the proximal ends of the maternal spiral arteries. Through the open endings of the maternal vessels that are created by trophoblast invasion, maternal blood is released into the intervillous space, flows around the trophoblast villi, and is drained by spiral veins (see Fig. 6.1 ).

A schematic of placental vascular remodeling in health (upper panel) and in disease – preeclampsia (lower panel). Exchange of oxygen, nutrients, and waste products between the fetus and the mother depends on adequate placental perfusion by maternal spiral arteries. Blood from the intervillous space is returned to the mother’s circulation via spiral maternal veins noted above. In normal placental development, cytotrophoblasts of fetal origin invade the maternal spiral arteries, transforming them from small-caliber resistance vessels to high-caliber capacitance vessels capable of providing adequate placental perfusion to sustain the growing fetus. During the process of vascular invasion, the cytotrophoblasts undergo a transformation from an epithelial to an endothelial phenotype, a process referred to as “pseudovasculogenesis” (upper panel). In preeclampsia, cytotrophoblasts fail to adopt an invasive endothelial phenotype. Instead, invasion of the spiral arteries is shallow and they remain small-caliber, resistance vessels (lower panel). This is thought to lead to placental ischemia and secretion of antiangiogenic factors.

Figure reproduced with permission from Lam et al. (This figure is reproduced in color in the color plate section.)

When cytotrophoblasts invade maternal spiral arteries, they replace the luminal endothelial cells, a process common in all species with hemochorial placentation. During this process, the endovascular cytotrophoblasts convert from an epithelial to endothelial phenotype, a process referred to as pseudovasculogenesis or vascular mimicry ( Fig. 6.1 ; see also Chapter 5 ). Thus cytotrophoblast stem cells lose epithelial markers, such as E-cadherin and α ₆ β ₃ integrin, and gain endothelial markers, such as vascular endothelial-cadherin (VE-cadherin) and α _v β ₃ integrin.

In addition to replacing maternal endothelial cells, cytotrophoblasts also remodel the highly muscular tunica media of spiral arteries, a process that is dependent on the enzyme membrane metalloproteinase 9 (MMP-9). This transforms the maternal high resistant vessels into larger low-resistance capacitance vessels. Uterine blood flow during pregnancy increases more than 20-fold, and the functional consequence of spiral artery remodeling maximizes the capacity of the maternal–placental circulation by providing sufficient blood supply for placenta and fetus at low blood pressure. Of note, spiral arteries from both the implantation and non-implantation regions display these physiological changes.

The mechanisms underlying these changes remain unknown. Placental oxygen tension has been suggested to be one of the major regulators of cytotrophoblast migration and differentiation. It has further been hypothesized that decidual natural killer (NK) cells and/or activated macrophages play a role in this vascular remodeling.

Fetal Circulation and Placental Villous Angiogenesis

The fetal circulation enters the placenta via the umbilical vessels. Inside the placenta, fetal vessels branch successively into units within the cotyledons and then into capillary loops within the chorionic villi. From post-conception day 32 until the end of the first trimester, the endothelial tube segments formed by vasculogenesis in the placental villi are transformed into primitive capillary networks by the balanced interaction of two parallel mechanisms; (a) elongation of pre-existing tubes by non-branching angiogenesis, and (b) ramification of these tubes by lateral sprouting (sprouting angiogenesis). A third process, termed intussusceptive microvascular growth, rarely contributes. In the third month of pregnancy, some of the centrally located endothelial tubes of immature intermediate villi achieve large diameters of 100 µm and more. Within a few weeks, they establish thin media- and adventitia-like structures by concentric fibrosis in the surrounding stroma and by differentiation of precursor pericytes and smooth muscle cells expressing α- and γ-smooth muscle actins in addition to vimentin and desmin. This is followed quickly by the expression of smooth muscle myosin. These vessels are forerunners of the villous arteries and veins and are developmentally regulated through the platelet-derived growth factor (PDGF) pathway.

After post-conception week 24 and continuing through term, patterns of villous vascular growth switch from the prevailing branching angiogenesis to a prevalence of non-branching angiogenesis. Analysis of proliferation markers at this stage reveals a relative reduction of trophoblast proliferation and an increase in endothelial proliferation along the entire length of these villous structures, resulting in non-sprouting angiogenesis by proliferative elongation. The final length of these peripheral capillary loops exceeds 4000 µm and they grow at a rate which exceeds that of the villi themselves, resulting in coiling of the capillaries. The looping capillaries bulge towards, and obtrude into, the trophoblastic surface and thereby contribute to formation of the terminal villi. Each of the latter is supplied by one or two capillary coils and is covered by an extremely thin (<2 µm) layer of trophoblasts that contributes to the so-called vasculosyncytial membranes. These are the principal sites of diffusional exchange of gases between mother and fetus. Normally, the capillary loops of 5–10 such terminal villi are connected to each other in series by the slender, elongated capillaries of the central mature intermediate villus.

The fetal vessels (chorionic vessels) from the individual cotyledons of the placenta unite at the placental surface to form the umbilical vessels that then traverse the umbilical cord. The umbilical cord consists of one vein and two arteries. The connective tissue surrounding these vessels in the umbilical cord is referred to as Wharton’s jelly. Most umbilical cords are twisted at birth, probably related to fetal activity in utero . The umbilical vein carries oxygenated blood from the placenta to the fetus and the umbilical arteries carry deoxygenated blood back to the placenta.

Angiogenic Factors and Placentation

Placental vascularization involves a complex interaction of several regulatory factors. The list of pro- and antiangiogenic molecules involved in placental vascular development has expanded exponentially over the past decade and is reviewed in detail elsewhere. The families of vascular endothelial growth factor (VEGF) and angiopoietin (Ang) gene products are those most extensively studied.

VEGF-A was initially defined, characterized, and purified for its ability to induce vascular leak and permeability, as well as for its ability to promote vascular endothelial cell proliferation. Members of this family include VEGF-A, VEGF-B, VEGF-C, VEGF-D and placental growth factor (PlGF). VEGF-A is an endothelial-specific mitogen and survival factor that exists in four major isoforms: VEGF121, VEGF165, VEGF189, and VEGF206. Inactivation of a single VEGF allele results in embryonic lethality in heterozygous embryos at days 11–12. Additionally, significant defects in placental vasculature are observed, implicating VEGF in placental vascular development. The high-affinity receptor tyrosine kinases for VEGF-A include Flt-1 (also referred to as VEGFR-1) and Kinase-insert Domain containing Receptor (KDR, human)/Flk-1 (murine), also known as VEGFR-2. KDR mediates the major growth and permeability actions of VEGF, whereas Flt-1 may have a negative role, either by acting as a decoy receptor or by suppressing signaling through KDR. Placental growth factor (PlGF), the first VEGF relative identified, was found to be abundantly expressed in the placenta. It acts by binding to Flt-1 but not KDR.

PIGF can potentiate the angiogenic activity of VEGF. Although reproduction and embryonic angiogenesis defects were not observed in mice with an isolated PIGF−/−genotype, the cross-sectional area of the midpregnancy placentas was decreased by 40% in PlGF null mice. VEGF-C and VEGF-D, based on their ability to bind the lymphatic-specific Flt-3 receptor (also known as VEGFR-3), appear to be important for lymphatic development.

Alternative splicing of Flt-1 results in the production of a truncated, endogenously secreted antiangiogenic protein referred to as sFlt-1, which lacks the cytoplasmic and transmembrane domain but retains the ligand-binding domain. Thus, sFlt-1 can antagonize VEGF and PlGF by binding to them and preventing interaction with their endogenous full-length receptors. There is substantial evidence that increased production of sFlt-1 plays a major pathogenic role in the severe endothelial dysfunction of preeclampsia (discussed below).

The Tie-2 receptor binds a family of four ligands termed angiopoietins (Ang-1 to Ang-4). Ang-1-mediated activation of Tie-2 promotes endothelial survival and capillary sprouting. The effects of Ang-1 and VEGF on sprouting are synergistic, reflecting different receptor and signal transduction pathways. In addition to its direct effects on endothelial cells, Tie-2 activation has been reported to induce maturation of adjacent smooth muscle and pericyte precursors via paracrine action of endothelial-derived factors such as PDGF-B. Ang-1 has also been shown to inhibit capillary permeability, preventing plasma leakage in response to both VEGF and histamine. In the presence of abundant VEGF, Ang-2 is thought to destabilize vascular networks and facilitate sprouting, e.g., during tumor growth. Conversely, with low ambient VEGF levels, Ang-2 may cause vessel regression, e.g., in corpus luteum involution.

There are few in vivo functional studies in humans demonstrating how placentation is regulated on a molecular basis by vascular growth factors. Nevertheless what is known about the expression patterns of angiogenic factors provides insights into potential molecular regulatory processes. VEGF and its receptors have been localized in the placenta of both primates and humans. VEGF is expressed in maternal decidual cells and in invading cytotrophoblasts. As for the two VEGF receptors, only Flt-1 expression is localized in the invading extravillous trophoblast during early pregnancy. Such observations raise the possibility that VEGF-Flt-1 interactions contribute to early trophoblast invasion. Later in gestation, VEGF is localized to cytotrophoblast cells, and Flt-1 to extravillous trophoblasts, all within the trophoblast tree or columns. This pattern suggests that in addition to regulating trophoblast invasion, the VEGF-Flt-1 axis plays an important role in the coordination of trophoblast differentiation and migration. Interestingly, more recently sFlt-1 has been primarily localized to the syncytiotrophoblast, suggesting that its primary role may be secretion into the maternal bloodstream to regulate systemic vascular homeostasis.

VEGF-A induces cytotrophoblast invasion in vitro , an effect blocked by exogenous sFlt-1. Although homozygous knock-out studies for both Flt-1 and Flk-1 in mice show defective fetal and placental vasculogenesis and angiogenesis, we are unaware of definitive in vivo evidence of impaired trophoblast invasion. This may be because the normal invasion of trophoblasts is relatively shallow even at baseline in rodents, contrasting with human placentas, where invasion is robust. In vitro studies using trophoblast cultures suggest that PlGF and VEGF-C (two related growth factors) are also expressed by the invading cytotrophoblasts and may contribute to cytotrophoblast invasion and differentiation via Flt-1 and Flt-3 signaling.

The development of the villous tree occurs contemporaneously with the formation of the fetoplacental vascular system. The latter process involves differentiation and proliferation of fetal endothelial cells, tubule formation and vessel stabilization. The KDR receptor is exclusively expressed on endothelial or mesenchymal cells from which endothelial cells differentiate, whereas VEGF expression is localized to the trophoblast. This suggests that endothelial cell differentiation, migration and proliferation, essential steps for building the primary vascular network, are mediated by VEGF/KDR ligand–receptor signaling in a paracrine system.

The angiopoietins are also expressed during the early placentation period in marmosets, indicating their involvement in the regulation of trophoblast growth. In this species Ang-1 is highly expressed in the syncytiotrophoblast, whereas its receptor, Tie-2, is located in the cytotrophoblast. Similar observations have been made in the human placenta, where Ang-1 has been shown to stimulate trophoblast growth and migration in vitro , and Ang-1 gene expression increases as gestation progresses. Thus, Ang-1 expression appears to trigger the in-growth of cytotrophoblast cones into the syncytiotrophoblast, whereas its relatively higher expression later in pregnancy may be required for branching of the villi and shaping the intervillous space.

Fetal capillaries within the placental villi are thin-walled in order to allow oxygen diffusion, while the chorion vessels are stabilized by a thick wall of pericytes, smooth muscle cells or both that ensure their task of collecting and draining fetal placental blood. Tie-2 has been shown to be expressed at high levels in the endothelium of chorion vessels, and at low levels in the fetal capillaries of the villi. These results suggest that the Ang-1/Tie-2 ligand–receptor pair acts on the fetal vasculature, especially on chorion vessels to induce stability. In humans, Ang-1 is secreted into the media of stem villus vessels at term, which is also consistent with its reported paracrine role in vessel maturation and stabilization.

Maternal vessels must be remodeled to attain an effective uteroplacental circulation. In humans, the trophoblast invasion process is so deep that the proximal parts of maternal spiral arteries become completely digested and the intervillous space is filled by their open endings. In addition, with Ang-2 produced by the cytotrophoblast and Tie-2 expressed in the maternal endothelium, a paracrine mechanism for maternal vascular remodeling is established.

In non-human primates, highest expression of Ang-2 mRNA is detected during early gestation, when maternal vascular remodeling takes place. Ang-2 is a plausible candidate for induction of maternal vascular transformation because it destabilizes the vasculature. This destabilization process is a local event, which may be driven in part by Ang-2/Tie-2. In humans, although the remodeling process of maternal vessels may be supported by Ang-2/Tie-2, it is primarily driven by aggressive trophoblast invasion. Finally, TGF-β1 and 3, two angiogenic morphogens made in the maternal decidua and the syncitiotrophoblast, also affect cytotrophoblast invasion and differentiation through endoglin signaling.

In summary, substantial evidence supports a critical role for angiogenic growth factors in hemochorial placentation of both humans and other primates. In our estimation, the most comprehensive studies investigating VEGF and angiopoietins, and their receptors across gestation, has been undertaken in marmoset monkeys, with results indicating a tight spatial and temporal regulation of placentation and angiogenesis. These findings led to the hypothesis that VEGF/Flt-1 and Ang1/Tie-2 pairs are critically involved in trophoblast differentiation and invasion; VEGF/KDR and Ang-1/Tie-2 may trigger fetoplacental vascular development, whereas Ang-2/Tie-2 may support the remodeling processes of the maternal vasculature.

The characterization of antiangiogenic proteins in the placenta, in contrast to pro-angiogenic proteins, has received less attention. One important antiangiogenic protein that has been well studied is sFlt-1, a potent antiangiogenic molecule and produced in abundant quantities within the placenta. Its production, predominantly from cytotrophoblasts, increases as pregnancy progresses. Other antiangiogenic proteins expressed in the placenta include thrombospondin-1, endostatin, and truncated fragments of prolactin, but their roles during normal placentation are unclear. More recently, soluble endoglin, a novel antiangiogenic protein made by the syncytiotrophoblast that acts by inhibiting TGF-β signaling, was reported to play an important role in the pathogenesis of the maternal syndrome of preeclampsia. However, the precise role of soluble endoglin during normal placentation remains unknown.

In general, the pro-angiogenic proteins are most highly expressed early in pregnancy and probably account for both placental angiogenesis and the increase in placental mass that accompanies fetal development. Toward term, antiangiogenic factors are increasingly expressed, possibly in preparation for delivery. In addition to the gestational age-dependent distribution of these proteins, there are also regional differences in their expression. For example, term human placental extracts derived from the decidual plate express stronger antiangiogenic activity compared with chorionic villus extracts.

Natural Killer Cells and Placental Vascular Development

Natural killer (NK) cells are considered important mediators of innate immunity, and those present at the maternal–fetal interface have been recently noted to play key roles during normal placental vascular remodeling. During the first trimester of human pregnancy, uterine NK (uNK) cells are a major cell population at the maternal–fetal interface, accounting for 70% of the local lymphocytes. In contrast, peripheral blood NK cells (pNK) account for only up to 15% of circulating lymphocytes. The abundance of NK cells in the decidua has prompted the idea that these cells might play a role in pregnancy support and maintenance. Initial evidence came from studies involving NK-cell-deficient mice, which manifested defective decidual vessel remodeling. Several candidate molecules of NK cell origin, such as gamma interferon, have been proposed to account for the NK-cell-mediated vascular remodeling; however, definitive evidence of this mechanism in humans is still lacking. Recent genetic studies suggest that susceptibility to preeclampsia may be influenced by polymorphic HLA-C ligands and killer cell receptors (KIR) present on NK cells. Preeclampsia is much more prevalent in women homozygous for the inhibitory KIR haplotypes (AA) than in women homozygous for the stimulator KIR BB. The effect is strongest if the fetus is homozygous for the HLA-C2 haplotype. Based on these observations, it is believed that in normal pregnancies uNK cell activation, through interaction with HLA-C on extravilous trophoblasts, promotes placental development and maternal decidual spiral artery modification by extra-villous cytotrophoblasts. Insufficient uNK cell activation might therefore blunt this process, resulting in incomplete decidual artery remodeling, thereby increasing the risk of pregnancy complications. Recent evidence showing secretion of trophoblast migration promoting factors and angiogenic factors by human uNK cells upon activation supports this hypothesis. Interestingly, consistent with this hypothesis, uNK cell interactions with trophoblasts were found to be impaired in humans at risk of subsequent preeclampsia.

Soluble Antiangiogenic Factors in Preeclampsia

The role of maternal endothelial dysfunction in producing the clinical manifestations of preeclampsia has been studied extensively since the 1980s. Roberts et al. proposed that endothelial cell injury caused increased sensitivity to pressor agents, vasoconstriction, and activation of the coagulation cascade that served as the basis of preeclampsia. Evidence of endothelial involvement in the kidneys was noted as early as the 1920s and by the 1960s, through the lens of electron microscopy, “glomerular capillary endotheliosis” became considered the characteristic renal lesion of preeclampsia. The lesion, described in detail in Chapter 15 , Chapter 16 , is characterized by occlusion of the glomerular capillaries by swollen endothelial cells.

Blood from preeclamptic patients demonstrates markers of endothelial injury, including increased levels of von Willebrand Factor, cellular fibronectin, and thrombomodulin. Circulating prostacyclins (normally produced in healthy endothelial cells) are decreased in preeclamptic patients. Blood vessels from preeclamptic patients reveal decreased endothelial-mediated vasodilator ability, whereas plasma endothelin-1 levels are increased. In summary, extensive data from multiple studies support the notion that the maternal serum in preeclampsia has soluble factors that mediate endothelial dysfunction. Moreover, serum or plasma from preeclampsia patients alters endothelial cell phenotype in vitro , including altered expression of vascular cell adhesion molecules, nitric oxide, and changes in prostaglandin balance. Interestingly, preeclamptic patients have also been noted to have decreased skin capillary density compared with healthy pregnant patients, a finding suggesting generalized angiogenesis may be defective in preeclampsia. Finally, women with a prior history of preeclampsia have an increased risk of remote ischemic heart disease, stroke, and hypertension. The association between preeclampsia and subsequent development of cardiovascular disease also points to the existence of systemic endothelial dysfunction in these individuals.

Soluble Fms-Like Tyrosine Kinase I (sFLT-1 or sVEGFRI)

In early 2000, studies were designed whose aim was to identify the soluble factors that mediate maternal endothelial dysfunction in preeclampsia. Gene expression profiling of placental tissue from women with and without preeclampsia, using microarray chips, revealed that messenger RNA for sFlt-1 was dramatically upregulated in preeclamptic placentas. sFlt-1 is a secreted protein, derived from a splice variant of the VEGF (vascular endothelial growth factor) receptor Flt-1 mRNA that lacks the transmembrane and cytoplasmic domain of the membrane-bound receptor. Circulating in the blood, it acts as a potent antagonist to VEGF and PlGF, factors made by normal placenta. VEGF is also constitutively expressed by glomerular podocytes as well as by other supporting cells and is thought to maintain fenestrations within the glomerular and hepatic sinusoidal endothelia. Systemic levels of sFlt-1 in patients with preeclampsia are greatly increased prior to delivery and decrease to baseline 48–72 h after delivery. Using the endothelial tube formation assay, an established in vitro model of angiogenesis, the authors observed that serum from patients with preeclampsia inhibited endothelial tube formation, but by 48 h postpartum this antiangiogenic effect had disappeared from the serum, suggesting that the inhibition of angiogenesis was caused by a circulating factor released from the placenta. When sFlt-1 was added to serum from normotensive pregnant subjects (at concentrations mimicking levels measured in subjects with preeclampsia), endothelial tube formation was blocked, recreating the effects noted with serum from preeclamptic patients. These effects could be reversed by adding exogenous VEGF and PlGF. These findings were also confirmed using supernatants from cultured preeclamptic villous explants. Hence, it is believed that an excess of circulating sFlt-1 levels may lead to an antiangiogenic state, causing endothelial dysfunction, and the clinical syndrome of preeclampsia.

Contemporaneous studies evaluating the expression of placental growth factor (PlGF) concentrations in the plasma of normal and preeclamptic pregnancies indicated that this also is an attractive biomarker. High sFlt-1 levels correlate inversely with decreased free PlGF levels during preeclampsia. In one report, uterine vein sFlt-1 concentrations were 4- to 5- fold higher than peripheral venous levels, suggesting that the predominant source of maternal sFlt-1 was of placental origin. More recently, several new isoforms of sFlt-1 with varying mRNA lengths have been described to be upregulated in preeclamptic placentas. In particular, a novel isoform of sFlt-1, referred to as sFlt-1-14, was recently described, expressed only in humans and not in rodents. Interestingly, this sFlt-1-14 has been noted to be the predominant VEGF inhibitor circulating in preeclamptic circulation. Furthermore, the source of sFlt-1-14 was located to the syncitial knots, suggesting that degeneration and/or aging of the placenta may be one mechanism for the upregulated sFlt-1 in preeclampsia. Work from numerous groups (summarized in Chapter 8 ) suggests that syncytial knots in the preeclamptic tissue are shed systemically in circulation and that circulating syncytial microparticles carry sFlt-1 to distal target organs.

Exogenous gene transfer of sFlt-1 into pregnant rats using an adenoviral vector produced hypertension, proteinuria, and glomerular endotheliosis, the renal lesion characteristic of preeclampsia. The glomerular lesion in these experimental animals, consisting of severe glomerular endothelial swelling (with loss of endothelial fenestrae) and generally preserved foot processes in the setting of heavy proteinuria, is striking in its resemblance to the renal histological findings in human preeclampsia (see Fig. 6.2 ). Some of these findings have now been reproduced in mice administered adenovirus expressing sFlt-1 or recombinant sFlt-1 protein.

Glomerular endotheliosis. (A) Normal human glomerulus. (B) Human preeclamptic glomerulus of 33-yr-old woman with a twin gestation and severe preeclampsia at 26 weeks gestation. The urine protein/creatinine ratio was 26 at the time of biopsy. (C) Electron microscopy of a glomerulus from the same patient. Note the occlusion of the capillary lumens by the swollen cytoplasm of endocapillary cells. Podocyte cytoplasm shows protein resorption droplets but relatively intact foot processes. Original magnification 1500×. (D) Control rat glomerulus: note normal cellularity and open capillary loops. (E) sFlt-1 treated rat: note similar occlusion of the capillary lumens by swollen endothelial cells with minimal increase in cellularity. (F) Electron microscopy of a sFlt-1 treated rat: note similar occlusion of capillary loops by swollen endocapillary cell cytoplasm accompanied by the relative preservation of podocyte foot processes. Original magnification 2500×. All light photomicrographs are of H&E sections taken at the identical original magnification of 40×.

Figure reproduced with permission from Karumanchi et al. (This figure is reproduced in color in the color plate section.)

The effects described above were also observed with sFlt-1 gene transfer in non-pregnant animals, suggesting that the effects of sFlt-1 on the maternal vasculature were direct and not dependent on the presence of a placenta. Furthermore, when a soluble form of VEGF receptor-2 (sFlk-1) (which does not antagonize PlGF) was given exogenously it did not induce a preeclamptic phenotype in pregnant rats, suggesting that antagonism of both VEGF and PlGF is necessary to induce the maternal syndrome. Finally, exogenous administration of VEGF-121, a soluble, circulating isoform of VEGF that lacks the heparin binding domain, rescued the preeclamptic phenotype in animal models of preeclampsia without adverse effects to the fetus. Similarly recombinant PlGF has also been shown to ameliorate preeclamptic signs and symptoms in the sFlt-1 overexpression mouse model of preeclampsia. Hence, it has been concluded that excess production of sFlt-1 by the placentas of preeclamptic women might be responsible for the hypertension and proteinuria by VEGF and PlGF signaling and that reduction of sFlt-1 may be one strategy to combat preeclampsia. It is still unknown whether the excess sFlt-1 made in preeclampsia is a primary phenomenon or secondary to a pathophysiological trigger such as placental ischemia or agonistic autoantibodies to the angiotensin 1 receptor (reviewed in Chapter 15 ) or impaired hemoxygenase expresion.

Several risk factors for preeclampsia can also be correlated with increased sFlt-1 levels. These include multigestational pregnancies, high-altitude pregnancies, trisomy 13, and nulliparity. Furthermore, the low level of circulating sFlt-1 typically found in smokers may explain the surprising decreased incidence of preeclampsia in this group.

Antagonism of VEGF and PlGF may have a pathogenic role in the hypertension and proteinuria noted in preeclampsia. VEGF induces nitric oxide and vasodilatory prostacyclins in endothelial cells, suggesting a role in decreasing vascular tone and blood pressure. Exogenous VEGF has been noted to accelerate renal recovery in rat models of glomerulonephritis and experimental thrombotic microangiopathy. More recently, exogenous VEGF was shown to ameliorate cyclosporine-related hypertension, endothelial dysfunction, and nephropathy. The tissues targeted in preeclampsia (such as the glomerulus or the hepatic sinusoids) have fenestrated endothelia, and it has been shown that VEGF induces endothelial fenestrae in vitro . Additionally reduction by 50% of VEGF production in+/−transgenic mouse glomerulus leads not only to glomerular endotheliosis but also to loss of glomerular endothelial fenestrae. Interestingly, several antiangiogenic compounds, including anti-VEGF antibodies used to treat malignancy-related angiogenesis, have also been associated with hypertension, proteinuria and reversible posterior leukoencephalopathy. Collectively, these data suggest that VEGF is important not only in blood pressure regulation, but also in maintaining the integrity of the glomerular filtration barrier. Thus antagonism of VEGF signaling, such as with excess sFlt-1, might lead to endothelial dysfunction, proteinuria, and hypertension. Finally, the excess sFlt-1 production in preeclampsia is consistent with the evolutionary explanation for preeclampsia pathogenesis that has been hypothesized by Haig, namely to improve fetal nutrient delivery by increasing maternal peripheral vascular resistance.

Soluble Endoglin

A large body of human and animal studies suggest that another placental antiangiogenic protein, soluble endoglin (sEng), may also contribute to the pathogenesis of preeclampsia. Soluble Eng was also elevated in patients with preeclampsia during clinical disease and prior to onset of symptoms (see Fig. 6.3 ).

Mean levels of sFlt-1/PlGF and soluble endoglin (sEng) and by weeks before the onset of preeclampsia. (A) This panel shows the mean concentrations of sFlt-1/PlGF according to the number of weeks before the onset of preterm preeclampsia (PE<37 weeks) and the mean concentrations in normotensive controls with appropriate- or large-for-gestational-age infants. Control specimens were matched within 1 week of gestational age to specimens from women who later developed preterm preeclampsia. (B) This panel shows the mean levels of sEng in case and control specimens shown in panel A according to the number of weeks before the onset of preterm preeclampsia (PE<37 weeks).

Figures reproduced with permission from Levine et al. and Hagmann et al. (This figure is reproduced in color in the color plate section.)

Endoglin (Eng) is an angiogenic receptor expressed mainly on the surface of endothelial cells and placental syncytiotrophoblast. Eng acts as a co-receptor for transforming growth factor-β (TGF-β, a potent pro-angiogenic molecule) signaling in endothelial cells. We recently observed that the Eng mRNA, like that of sFlt-1, is upregulated in the placentas of preeclamptic women. It was further noted that the soluble endoglin (sEng) is released in excess quantities into the circulation of preeclamptic patients. Soluble endoglin acts as an antiangiogenic factor by inhibiting TGF-β signaling in the vasculature. sEng amplifies the vascular damage mediated by sFlt-1 in pregnant rats, inducing a more severe preeclampsia-like syndrome with features of the HELLP syndrome. Furthermore, overexpression of sFlt-1 and sEng in rodents was also found to induce focal vasospasm, hypertension and increased vascular permeability associated with brain edema, producing images reminiscent of reversible posterior leukoencephalopathy associated with human eclampsia (reviewed in Chapter 13 ) (see Fig. 6.4 ). The contributions of sEng and sFlt-1 to the pathogenesis of the maternal syndrome of preeclampsia are at least in part, related to their inhibition of VEGF, PlGF and TGF-β stimulation of eNOS activation and vasomotor effects (see Fig. 6.5 ). Another candidate molecule that may be central to the pathogenesis of the pro-coagulant state and the thrombocytopenia noted in preeclampsia is prostacyclin (PGI ₂ ). Both VEGF and TGFβ1 have been shown to stimulate the production of the anti-thrombotic prostacyclin-PGI ₂ . Clinical studies demonstrating decreased endothelial PGI ₂ production, even before onset of clinical preeclampsia, support this hypothesis. In summary, maternal endothelial dysfunction caused by placenta-derived soluble factors such as sFlt-1 and sEng is emerging as the final common pathway that mediates the clinical syndrome of preeclampsia. Future studies are needed to clarify the molecular nature of the circulating sEng protein and its downstream signaling pathways that mediate endothelial dysfunction.

Cerebral edema in eclamptic subjects and in animal models of preeclampsia/eclampsia.

sFlt-1 and sEng causes endothelial dysfunction by antagonizing VEGF and TGF-β signaling. There is mounting evidence that VEGF and TGF-β are required to maintain endothelial health in several tissues including the kidney and perhaps the placenta. During normal pregnancy, vascular homeostasis is maintained by physiological levels of VEGF and TGF-β signaling in the vasculature. In preeclampsia, excess placental secretion of sFlt-1 and sEng (two endogenous circulating antiangiogenic proteins) inhibits VEGF and TGF-β1 signaling respectively in the vasculature. This results in endothelial cell dysfunction, including decreased prostacyclin, nitric oxide production and release of procoagulant proteins.

Figure reproduced with permission from Karumanchi and Epstein. (This figure is reproduced in color in the color plate section.)

Upstream Pathways and Mechanisms of Preeclampsia

The studies described above are compelling and support the hypothesis that the maternal syndrome of preeclampsia, particularly in its more severe forms, is “an antiangiogenic state”. Placental hypoxia has been suggested as the cause of excess antiangiogenic factors, but the precise mechanisms of excess sFlt-1 and sEng production by this organ are not known. Various pathways including deficient heme-oxygenase (HO) expression, angiotensin receptor 1 autoantibodies, oxidative stress, inflammation, altered natural killer cell signaling and, more recently, deficient catechol- O -methyl transferase have been proposed to play key roles in inducing placental disease. Of these factors, alterations in the hemoxygenase system have emerged as an important upstream regulator of placental health. In cell and organ culture studies, HO1 and its downstream metabolite carbon monoxide act as vascular protective factors by inhibiting the production of sFlt-1 and sEng. Women with preeclampsia exhale less carbon monoxide than women with normal pregnancies and decreased HO1 has been noted in the placentas as early as the first trimester of pregnancy. These findings may also explain why cigarette smoking is paradoxically associated with a protective effect in preeclampsia More recently, animal studies suggest that dysregulated transthyretin or low levels of the enzyme cystathionine gamma lyase (which regulates hydrogen sulfide production) may contribute to preeclampsia by disrupting placental angiogenesis.

There are other questions to be resolved. Although sFlt-1 levels are increased in most patients with preeclampsia, they are not universally elevated in every diagnosed case. This is especially true when the disease is mild. Moreover, the relationship of sFlt-1 to some of the known risk factors for preeclampsia (such as obesity, and preexisting hypertension) is unclear. One hypothesis proposes that there is a threshold for sFlt-1 to cause disease. This theory evolves from the fact that sFlt-1 levels increase across gestation in all pregnant women. Thus it is theorized that if circulating sFlt-1 concentrations remain below a certain level, pregnancy proceeds normally, but when values exceed the cutoff level preeclampsia develops. In this respect, risk factors effectively lower the threshold, rendering women more sensitive to sFlt-1 and therefore resulting in preeclampsia despite levels that match those of normal pregnancy.

Still another reason for the appearance of preeclampsia phenotypes without increased sFlt-1 levels may be related to the fact that the disease may be misclassified in some patients. This is not entirely surprising as renal biopsy studies by Fisher et al. in 1981 had already suggested that ~15% of such diagnoses were incorrect in nulliparous women, and>40% of multiparas were misdiagnosed. Misclassification of preeclampsia diagnosis can be due to the presence of preconception chronic hypertension and/or obesity with minimal glomerulosclerosis, insufficient to produce proteinuria. A paucity of of major adverse maternal or fetal outcomes in preeclamptics characterized by normal circulating angiogenic profiles supports this latter hypothesis.

If antiangiogenic factor production is an important cause of preeclampsia, there might be at least two kinds of predisposing factors. One could involve overproduction of sFlt-1 and sEng and occur in instances such as multiple gestation, hydatiform mole, trisomy 13, and, possibly, nulliparity. Another set of predisposing factors would include disorders such as chronic hypertension or thrombophilia that sensitize the maternal vascular endothelium to the antiangiogenic effects of sFlt-1 and sEng. We do not yet know whether diabetes, hypertension and preexisting renal disease predispose to preeclampsia by increasing the production of sFlt-1 and sEng or by sensitizing the vascular endothelium to their presence, or, as discussed above, are likely to be misdiagnosed because of lack of specific criteria for the diagnosis of superimposed preeclampsia.

Hypoxia is known to increase the production of sFlt-1 by placental trophoblasts and to inhibit their production of PlGF, so that placental ischemia might trigger the preeclamptic syndrome. There is strong evidence for placental ischemia in many patients with preeclampsia but not in all. Whereas growth restriction, a corollary of placental ischemia, frequently accompanies preeclampsia, as many as one-third of the neonates of preeclamptic women are large for gestational age. Placental infarction unaccompanied by preeclampsia is a common finding in mothers with sickle cell anemia and those with fetuses who have intrauterine growth retardation. Placental overproduction of sFlt-1, whatever its cause, might impair angiogenesis locally and result in placental ischemia, thereby initiating a vicious circle leading to even more sFlt-1 production.

We propose that three factors conspire, in variable degree, to produce the clinical syndrome of preeclampsia: (1) a disturbance in the balance of circulating factors controlling angiogenesis/antiangiogenesis, attributable to placental overproduction of sFlt-1 and underproduction of PlGF, (2) increased vascular endothelial sensitivity to such factors, and (3) placental ischemia exaggerating the imbalance in angiogenic factors by further inducing the production of antiangiogenic factors. Human pregnancy is initially characterized by rapid angiogenesis, localized to the placenta, followed, as pregnancy concludes, by a slow regression of blood vessel growth. It would therefore not be surprising if derangements of this remarkably complex process might occur and lead to the systemic manifestations of preeclampsia.

Angiogenesis and the Remote Consequences of Preeclampsia

Considerable evidence suggests that preeclampsia is a marker of remote cardiac and vascular diseases, but whether preeclampsia per se is a cause of future hypertension and heart disease or whether this represents the sequestration of women in the preeclamptic population to be predestined to these disorders has yet to be determined. In this respect, one small study published in 1964 suggested that the tendency to late hypertension is the result of preeclampsia, rather than a heritable tendency, demonstrated by the absence of hypertension in the siblings of preeclamptic patients. It has also been suggested that subtle renal injury, such as that induced by preeclampsia, can eventually lead to the development of chronic hypertension in animal models. Recently sFlt-1 has been suggested as pathogenic in peripartum cardiomyopathy (PPCM), a rare complication of pregnancy that is often associated with preeclampsia. It is also tempting to speculate that the long-term cardiovascular complications noted in some patients who had preeclampsia may be due to a chronic antiangiogenic state resulting from polymorphisms in genes such as sFlt-1. Alternatively, persistence of autoantibodies against the AT1 receptor could be responsible for the long-term cardiovascular sequlae (see Chapter 15 ).

Additionally, patients with preeclampsia are said to have a decreased long-term incidence of malignancy, a provocative observation disputed by some, that may suggest that a long-term antiangiogenic state associated with preeclampsia may provide a protective milieu.

Role of Angiogenic Biomarkers in Preeclampsia

Over the past decade, numerous cross-sectional and longitudinal studies have shown that high sFlt-1 and low PlGF are present during preeclampsia and prior to clinical disease (see Fig. 6.3 ).

The recent availability of automated platforms has allowed researchers to validate these biomarkers in several cohort studies. A number of studies have demonstrated the ability of sFlt-1 and PlGF to identify women with preeclampsia using newly developed automated assays with excellent sensitivities and specificities for preterm preeclampsia, More recently, several reports have suggested that angiogenic factors measured in the last trimester can predict development of severe preeclampsia-related adverse outcomes with very high sensitivity and specificity. These algorithms appear to distinguish patients whose course will be relatively benign from those likely to experience severe complications. Measurement of angiogenic factors appears to be useful, especially in distinguishing preterm gravidas who can be managed conservatively from those with substantial risk of developing adverse outcomes and requiring close surveillance and a low threshold for delivery. In addition, circulating angiogenic factors may have the specificity to differentiate preeclampsia from other diseases that mimic preeclampsia such as chronic hypertension, gestational hypertension, and other chronic kidney diseases.

Therapeutic Strategies for Preeclampsia

Studies of angiogenic pathways are helping devise specific therapies for preeclampsia. In a pilot study limited to three severe early preeclamptics (24–32 week gestation) Thadhani et al. depleted sFlt-1 30% by dextran sulfate apheresis and prolonged pregnancy by 2–4 weeks. If confirmed this approach could lead to targeted therapy for a specific group of patients with premature disease and an abnormal angiogenic profile. More recently statin therapy that promotes angiogenesis was shown to prevent or ameliorate disease in an animal model of preeclampsia. Pilot human trials to test the safety and efficacy of statins in severe preeclampsia are ongoing. Relaxin, a naturally occurring vasodilator of pregnancy and which induces local VEGF and endothelial progenitor cells, is also being investigated for its therapeutic potential in preeclampsia. Finally, dietary choline supplementation, shown to reduce placental sFlt-1 expression, has been suggested as a strategy to improve placental angiogenesis. The future for specific therapies that reduce sFlt-1 production, antagonize its actions or enhance PlGF or VEGF levels is therefore promising.

Anti-Angiogenic Versus Normal Angiogenic Forms of Preeclampsia

As of 2014, there have been discussions described in several reviews concerning whether preeclampsia is “a syndrome of multiple disorders or multiple subtypes” or a “single homogenous disease”. Characterization of circulating angiogenic factors in clinical studies of preeclampsia has provided some answers to this debate.

Based on epidemiological and experimental studies, we believe that alterations in angiogenic factors are both necessary and sufficient for the development of preeclampsia phenotypes, and their related complications. When preeclamptic patients with anti-angiogenic (defined by high sFlt-1/PlGF ratio≥85) and normal angiogenic plasma profiles (defined by sFlt-1/PlGF ratio<85) were studied, the serious adverse outcomes traditionally ascribed to preeclampsia were only found in subjects with abnormalities in angiogenic factors. These data suggested that patients presenting signs and symptoms of preeclampsia but with relatively normal angiogenic profiles may have been misclassified. This was not surprising as patients who presented with non-angiogenic forms of preeclampsia were obese or had other chronic conditions such as diabetes or chronic hypertension and therefore were less likely to be reliably diagnosed based on the traditional but nonspecific criteria of blood pressure and proteinuria. While there may be different causes that stimulate the placenta to produce an excess of antiangiogenic factors, the latter alone may be sufficient to produce preeclamptic phenotypes. Thus, severe complications of preeclampsia are directly correlated with more dramatic alterations in circulating angiogenic factors.

In contrast, it has been argued by others that preeclampsia is a heterogenous disease with multiple subtypes. Despite decades of research, no single test has been able to predict preeclampsia with the appropriate likelihood ratios required for an adequate clinical test. Because various forms exist (e.g., placental or maternal [discussed in Chapter 8 ]), the search for a single marker of preeclampsia may be futile. Unfortunately, it is very difficult to obtain the evidence necessary to clarify this debate; in developed nations women classified as preeclamptic are not permitted to develop adverse outcomes unless they present very prematurely, and aggressive delivery of patients with suspected preeclampsia is the common practice.

Of considerable interest relative to the two views discussed above is that increasing sFlt-1 and sEng levels in experimental animals produces the histological lesions described in kidney, and liver tissue of a woman dying of either preeclampsia or eclampsia reported by Sheehan and Lynch in 1973. Taken together with human epidemiological studies, these data support the hypothesis that increased circulating antiangiogenic factors are necessary to induce the most severe forms of preeclampsia.

Prospective clinical trials are still needed to evaluate whether subjects with a normal angiogenic profile can be expectantly managed and delivered at 37 weeks or beyond. These tests should be easier to perform now that precise automated angiogenic factor assays that produce rapid results are available.

In summary, measuring plasma circulating angiogenic factors may provide a new approach to identify what we consider the “true or multi-systemic preeclampsia”. It may also improve classification and lead to better research on causality, prediction, and management of both immediate and remote outcomes.