Authors: Kelsey S. Schwartz, Anna E. Stanhewicz
Categories: Article
Source: Comprehensive Physiology
Doi: 10.1002/cphy.c240003
Authors: Kelsey S. Schwartz, Anna E. Stanhewicz
Preeclampsia, a pregnancy disorder characterized by de novo hypertension and maternal multisystem organ dysfunction, is the leading cause of maternal mortality worldwide and is associated with a fourfold greater risk of cardiovascular disease throughout the lifespan. Current understanding of the etiology of preeclampsia remains unclear, due in part to the varying phenotypical presentations of the disease, which has hindered the development of effective and mechanism-specific treatment or prevention strategies both during and after the affected pregnancy. These maternal sequelae of preeclampsia are symptoms of systemic vascular dysfunction in the maternal nonreproductive microvascular beds that drives the development and progression of adverse cardiovascular outcomes during preeclampsia. Despite normalization of vascular disturbances after delivery, subclinical dysfunction persists in the nonreproductive microvascular beds, contributing to an increased lifetime risk of cardiovascular and metabolic diseases and all-cause mortality. Given that women with a history of preeclampsia demonstrate vascular dysfunction despite an absence of traditional CVD risk factors, an understanding of the underlying mechanisms of microvascular dysfunction during and after preeclampsia is essential to identify potential therapeutic avenues to mitigate or reverse the development of overt disease. This article aims to provide a summary of the existing literature on the pathophysiology of maternal microvascular dysfunction during preeclampsia, the mechanisms underlying the residual dysfunction that remains after delivery, and current and potential treatments both during and after the affected pregnancy that may reduce microvascular dysfunction in these high-risk women.
Preeclampsia is a hypertensive disorder of pregnancy that complicates 2% to 8% of pregnancies globally and is the leading cause of maternal and perinatal death worldwide (80). In the United States, the rate of preeclampsia has been increasing steadily over the past 30 years (317), with a striking increase in the risk of severe preeclampsia since 1980 (13). Clinically, preeclampsia is characterized by de novo hypertension (systolic and/or diastolic pressures ≥140/90 mmHg, respectively) after the 20^th^ gestational week accompanied by proteinuria (0.3 g per 24 h) and/or symptoms of multisystem organ dysfunction such as thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or new-onset cerebral symptoms such as vision impairment and headache unresponsive to medication (Figure 1) (12). Preeclampsia can also present superimposed on chronic hypertension when the de novo development of the accompanying symptoms occurs in a pregnancy already complicated by chronic hypertension.
The exact etiology of preeclampsia is still currently undetermined and is the subject of ongoing investigation and debate. Historically, preeclampsia has been considered a consequence of inappropriate placentation, with the fetal and maternal symptoms occurring secondary to the poorly perfused placental circulation. Indeed, most cases occur during gestation and resolve following delivery of the placenta, and preclinical models of the disease demonstrate a clear role for placental ischemia and placental-derived factors that drive the maternal and fetal sequelae of preeclampsia. In this framework, poor uterine spiral artery remodeling results in a poorly perfused and hypoxic placental circulation, leading to the release of antiangiogenic factors, inflammatory cytokines, and reactive oxygen species, all of which enter the maternal circulation and cause widespread vascular endothelial dysfunction. However, more recently, it has become clear that the heterogeneity of clinical presentations that comprise preeclampsia indicates that it is more likely a maternal syndrome rather than a single disease. In response to emerging evidence from human studies of preeclampsia, two distinct phenotypes have been one that is placental in origin, manifesting early in pregnancy (<34 weeks gestation) that is referred to as “type 1” or “early-onset” preeclampsia, and one that is maternal in origin, characterized by an inability of the maternal circulation to appropriately meet the demands of pregnancy, manifesting later in pregnancy (≥34 weeks gestation) that is referred to as “type 2” or “late-onset” preeclampsia (241, 342). This pathophysiological distinction is likely an important one, driving differences in the underlying mechanisms of the maternal and fetal consequences of preeclampsia. However, given the relative recency with which this distinction has been made, few, if any, studies have examined mechanistic differences in the maternal microvascular dysfunction associated with each phenotype. This represents a major future direction for the field at large, but given the paucity of existing data speaking to these differences, we do not make the distinction between these phenotypes in the current article.
Despite remission of the clinical symptoms of preeclampsia (e.g., high blood pressure, proteinuria, and edema) at delivery or within 12 weeks postpartum, risk for maternal health complications remains elevated throughout the lifespan. Women with a history of preeclampsia have ≥4-fold risk of developing cardiovascular disease (CVD) (5, 26, 340). Notably, these women develop CVD at a younger age, and greater frequency, than women with a history of uncomplicated pregnancy (116, 161, 240). Consequently, women who had preeclampsia are far more likely to die of cardiovascular complications compared to women who had a normotensive pregnancy (26, 68, 203). Although the association between a prior preeclamptic pregnancy and increased risk for vascular diseases is evident, the putative mechanisms underlying this association remain unclear. This is further exacerbated by the various subtypes of preeclampsia that lead to a common clinical presentation during pregnancy (246). Evidence suggests there is a dose-response relationship between the severity of preeclampsia and relative risk (RR) of CVD sequelae (mild: RR 2.00, 1.83–2.19; RR 2.99, 2.51–3.58; and RR 5.36, 3.96–7.27, P < 0.0001) (191). However, this relationship is difficult to quantify due to the lack of standardization of preeclampsia severity. Additionally, racial and ethnic disparities in the incidence of preeclampsia persist in the prevalence of increased chronic disease risk postpartum. The prevalence of adverse cardiovascular events after preeclampsia is highest in Black or African American, Asian or Pacific Islander, American Indian, and Alaskan Native populations (49, 62, 182, 235, 248). Importantly, these differences exist in the absence of prepregnancy hypertension or other confounding conditions, suggesting that preeclampsia elicits more severe consequences that persist postpartum in these populations. The increased risk of CVD after preeclampsia holds significance as CVD is the leading cause of death among women in the United States (296), and women are significantly more likely to have asymptomatic CVD or present with atypical symptoms compared with men (95). Women who had preeclampsia demonstrate a 3.6-fold greater risk of having undiagnosed or uncontrolled hypertension compared to women who had a healthy pregnancy 20 years postpartum (52) and are far more likely to have asymptomatic heart failure (stage B), with early-onset preeclampsia conferring the greater risk, up to 18 years after the affected pregnancy (43, 192, 293). Although the lifetime maternal burden of preeclampsia is clear, the uncertainty of the underlying pathological mechanisms has hindered the assessment and application of therapeutic approaches to reduce or halt overt disease progression in these women.
In this article, we review the mechanisms driving maternal microvascular dysfunction during and after a preeclamptic pregnancy. We present preclinical and human data speaking to the pathophysiology of the progression of the maternal presentation of preeclampsia during pregnancy and the putative mechanisms driving accelerated CVD progression in women after a preeclamptic pregnancy. Finally, within each phase (pregnancy and postpartum), we review promising approaches for mechanism-specific therapy to improve maternal microvascular dysfunction for the treatment or prevention of acute and chronic disease.
The broad clinical definition of preeclampsia—encompassing renal, hepatic, and/or cerebral symptoms in addition to de novo hypertension—illustrates the heterogeneity of the presentation of the syndrome but also underscores the involvement of the nonreproductive maternal microvascular beds in the pathophysiology of the disease (Figure 2). Normal pregnancy leads to profound physiological changes in the maternal cardiovascular system that begin early in gestation and gradually increase and peak in the third trimester (179). Mediated in part by increases in circulating hormones estradiol, progesterone, and relaxin, pregnancy is associated with vasodilation of the systemic vasculature resulting in a ~30% to 40% drop in peripheral vascular resistance (58, 146, 179). These changes drive a characteristic decrease in arterial blood pressures, including systolic and diastolic blood pressures, mean arterial pressure, and central systolic blood pressure, which decline 5 to 10 mmHg below preconception values in uncomplicated pregnancy (110, 146). Conversely, blood pressure and peripheral vascular resistance are considerably higher in preeclampsia and increase above prepregnancy values (66). Interestingly, preeclamptic women have increased plasma testosterone and decreased estrogen and progesterone compared to normal pregnancy (149). However, the magnitude to which these differences in circulating hormones contribute to maternal endothelial dysfunction in preeclampsia is unclear. In normal pregnancy, the kidneys play a central role in the cardiovascular adjustments to pregnancy, and vasodilation of the renal circulation results in ~50% increase in renal plasma flow and glomerular filtration rate (GFR) compared to prepregnancy values (58, 309). Conversely, renal plasma flow and GFR are lower in preeclampsia (81), and in vivo measures of renal resistance and renal pressure pulsatility are higher in preeclampsia compared to normal pregnancy (17, 112). Considerably less is known about cerebral and hepatic hemodynamics in pregnancy and preeclampsia. However, it appears that cerebral blood flow velocity is reduced across gestation in normal pregnancy (24, 41, 266, 335, 348), while cerebral perfusion pressure is higher in preeclampsia compared to normal pregnancy (25, 167, 211, 245, 257, 307), and women with preeclampsia and accompanying cerebral symptoms have greater cerebral perfusion pressure compared to those with preeclampsia without cerebral symptoms (167). Furthermore, cerebral microvascular responses to vasodilatory stimuli are attenuated in vivo in women with preeclampsia (245, 257, 308). Hepatic blood flow remains unchanged in normal pregnancy, while ~20% to 30% of preeclamptic patients have elevated circulating liver enzymes secondary to vasospasm of the hepatic circulation, leading to increased hepatic microcirculatory resistance, sinusoidal obstruction, and ischemia (236, 282). Collectively, these functional changes in the nonreproductive maternal microvascular circulation demonstrate the involvement of these beds in the pathophysiology of the syndrome. Therefore, there is an urgent need to understand the mechanisms driving these changes to provide insight into the pathophysiology of the maternal syndrome and identify potential therapeutic approaches beyond early delivery of the placental-fetal unit.
Poor spiral artery remodeling and placental ischemia at the onset of preeclampsia trigger the release of antiangiogenic factors soluble fms-like tyrosine kinase-1 (sFlt-1) and soluble endoglin, increasing their concentrations in the maternal circulation and altering their ratio relative to proangiogenic vascular endothelial growth factors (VEGF) and, specifically, placental growth factor (PlGF) (169, 239, 292). Placental ischemia/hypoxia also triggers the release of inflammatory cytokines, hypoxia-inducible factors (HIF), reactive oxygen species (ROS), and the agonistic autoantibody for the angiotensin II type 1 receptor (AT1-AA), all of which enter the maternal circulation and lead to widespread endotheliosis and consequent hypertension with multisystem organ dysfunction characteristic of the syndrome. Here, we review the mechanisms contributing to maternal microvascular endothelial dysfunction during preeclampsia and the role of circulating factors in the pathogenesis of this endothelial dysfunction.
The vascular endothelium plays a dynamic role in the regulation of local blood flow and total peripheral resistance through modulation of vascular tone. In response to mechanical and chemical stimuli, vascular endothelial cells release vasoactive molecules termed endothelium-derived relaxing factors (EDRF) or endothelium-derived contracting factors (EDCF), which induce vascular smooth muscle (VSM) relaxation or contraction, respectively. The reduction in peripheral vascular resistance characteristic of normal pregnancy is driven, in large part, by increases in endothelium-dependent dilation favoring the endothelial production of, and VSM sensitivity to, EDRF (175). In preeclampsia, however, endothelial dysfunction, characterized by reduced synthesis of EDRF and increased synthesis and sensitivity to EDCF, may drive the increase in total peripheral resistance, hypertension, and end organ dysfunction characteristic of the syndrome (Figure 3).
Pregnant women with preeclampsia have reduced endothelium-dependent dilation, assessed via brachial artery flow-mediated dilation, compared to matched women during normal pregnancy (111, 140, 212). Isolated resistance arteries from pregnant women with preeclampsia have altered relaxation responses to endothelial stimulation with bradykinin (137, 142), and incubation with plasma from preeclamptic women reduces endothelium-dependent dilation in isolated myometrial arteries from normal pregnant women (115). Similarly, animal models of preeclampsia demonstrate reduced endothelium-dependent relaxation in the nonreproductive vasculature (69). This reduced endothelium-dependent relaxation in preeclampsia is mediated by reduced nitric oxide (NO) bioavailability, reduced prostacyclin (PGI2) synthesis, and reduced endothelium-derived hyperpolarizing factor (EDHF) production.
Nitric oxide is a potent vasodilator synthesized by nitric oxide synthase (NOS) in the endothelium. Synthesized NO diffuses to the VSM where it increases cyclic guanosine monophosphate (cGMP), which in turn decreases intracellular VSM calcium concentrations and causes vasorelaxation. Markers of NO metabolism, including serum nitrates and urinary excretion of cGMP, are increased in normal pregnancy (314) and attenuated in preeclampsia (263, 274). Vascular endothelial cells from preeclamptic women produce less NO than cells isolated from normal pregnant controls (145). Interestingly, placental endothelial NOS mRNA expression is reportedly higher in preeclampsia compared to normal pregnant patients (273), and bovine endothelial cells and pregnant rat uterine arteries cultured with plasma from preeclamptic women exhibit increased endothelial NOS expression compared to controls (75, 133). Reduced NO bioavailability in the setting of increased NOS expression is likely explained by increased scavenging of NO by reactive oxygen species (reviewed below), NOS uncoupling, or by functional opposition to vasodilation in the presence of greater vasoconstrictor tone. Systemic pharmacological inhibition of NOS during pregnancy induces a preeclampsia phenotype (i.e., hypertension, renal vasoconstriction, proteinuria, and intrauterine growth restriction) in murine models of the syndrome (76, 139, 205). Similarly, vascular studies of nonreproductive vessels from reduced uterine prefusion pressure (RUPP) rats demonstrate reduced endothelial NOS expression, NO synthesis, and endothelium-dependent (acetylcholine-mediated) dilation compared to normal pregnant rats (189). Collectively, these findings underscore the role of endothelium-derived NO in the maternal vascular adaptations to normal pregnancy and highlight its significance in the cardiovascular pathophysiology of preeclampsia.
Prostaglandin is a potent vasodilator produced by cyclooxygenases 1 and 2 in the arachidonic acid metabolic pathway. During normal pregnancy, urinary and fetoplacental concentrations of the PGI2 metabolite 6-keto-PGF1α are increased (108, 180), which may be reflective of a pregnancy-related upregulation of cyclooxygenase enzyme activity in the maternal vascular endothelial cells (127). Plasma and urinary 6-keto-PGF1α concentrations are reduced in preeclampsia and gestational hypertension (83, 108, 170, 344). Early studies suggested that umbilical and placental vascular prostacyclin is reduced in preeclampsia (243) and, more recently, that endothelial PGI2 synthesis is reduced in maternal resistance arteries isolated from preeclamptic women compared to normal pregnant controls (288). These data suggest that reductions in PGI2 synthesis contribute to the impaired relaxation of the maternal microvasculature during preeclampsia. However, the mechanisms driving this reduction are relatively unexplored in the context of the syndrome.
Endothelium-derived hyperpolarizing factor is a catch-all term used to describe endothelium-dependent relaxation of the vascular smooth muscle mediated by a mechanism separate from NO and PGI2. It is now known that EDHF actually includes more than one mechanism, which leads to the hyperpolarization of the underlying VSM cells either directly though myoendothelial gap junctions and/or accumulation of potassium ions or as a diffusible factor released by the endothelial cells (89). EDHF plays a large role in the endothelial control of vascular tone in resistance arterioles. In animal models, increased activity of endothelial large and small conductance Ca^2+^-activated K^+^ channels promotes EDHF-mediated dilation of uterine arteries (107). PlGF induces EDHF-dependent dilation via large and small conductance Ca^2+^-activated K^+^ channels in pregnant rats (183), and diffusible factors mediate increased EDHF-dependent dilation downstream of increasing progesterone (113) in mesenteric arteries from pregnant mice. In healthy pregnant women, EDHF contributes ~50% to endothelium-dependent (bradykinin-mediated) dilation assessed in isolated arterioles from subcutaneous fat biopsy and is mediated largely by gap junctions, specifically the gap junction protein connexin 43 (160, 176). In the same vessels isolated from women with preeclampsia, this EDHF-dependent dilation is reduced, and the mechanisms mediating EDHF-dependent dilation are altered, such that diffusible factors play a role in mediating the response (177). However, connexin 43 expression is not different between omental arteries isolated from women with normal pregnancy and preeclampsia (300), suggesting mechanisms upstream and/or downstream contribute to reduced EDHF-dependent dilation during preeclampsia. These data highlight a physiological role for increased EDHF-mediated dilation in normal pregnancy and suggest that EDHF-mediated responses may be reduced in preeclampsia.
Hydrogen sulfide has recently emerged as a gasotransmitter in the family of EDHF. Circulating hydrogen sulfide is lower in plasma from pregnant women with preeclampsia compared to normal pregnant controls (227, 324). Cystathionine-γ-lyase, the primary hydrogen sulfide-producing enzyme within the vasculature, is reduced in placental tissue in preeclampsia (324). Endothelial cells cultured with plasma from preeclamptic women have reduced expression of cystathionine-β-synthase, a rate-limiting enzyme in hydrogen sulfide synthesis (51). Interestingly, inhibition of hydrogen sulfide-producing enzymes in human umbilical endothelial cells increases the release of sFlt-1 and soluble endoglin, suggesting a role for hydrogen sulfide in preventing the pathogenesis of preeclampsia (324). Collectively, while there is a paucity of data speaking to the role of hydrogen sulfide as an EDHF in the maternal microvasculature, the existing data suggest that hydrogen sulfide plays a role in mediating reductions in vascular resistance in normal pregnancy, and that these pathways are disrupted, leading to increases in resistance in preeclampsia.
Pregnant women with preeclampsia have increased circulating concentrations of EDCF such as thromboxane A2 (TXA2) (59, 326) and endothelin-1 (ET-1) (28, 311). Similarly, animal models of preeclampsia demonstrate increased constrictor sensitivity to EDCF in maternal nonreproductive vascular beds. This increased endothelial synthesis of, and VSM cell sensitivity to, EDCF in preeclampsia likely contributes to the increased total peripheral resistance, hypertension, and end organ dysfunction characteristic of the syndrome.
Endothelin-1 is a potent vasoconstrictor produced in the vascular endothelial cells. ET-1-dependent vascular responses are mediated by the interaction of ET-1 with the ETB1 receptor on vascular endothelial cells, leading to NO synthesis and vasodilation, and the ETA and ETB2 receptors on VSM cells, leading to vasoconstriction. ET-1 concentrations are elevated in the circulation of women with preeclampsia compared with normal pregnant women (28, 311), and cultured endothelial cells stimulated with serum from preeclamptic women produce greater ET-1 concentrations than cells stimulated with serum from normal pregnant controls (262). In isolated resistance arteries from pregnant women, ET-1 induces a greater constrictor response in vessels from women with preeclampsia compared to normal pregnancy (337). Pretreatment with an ETA receptor inhibitor abolished this contraction, suggesting a role for this receptor subtype in enhanced ET-1-mediated vasoconstriction in preeclampsia (337). VSM ET-1 receptor expression and activity seem to be reduced in normal pregnant rats compared to never-pregnant controls, suggesting a role for changes in this system in the reduced peripheral resistance characteristic of normal pregnancy (8). Conversely, ET-1-dependent constriction is increased in isolated resistance arteries from the RUPP rat model of preeclampsia compared to sham controls (189). Downregulation of endothelial ETB1 receptor expression and activity contributes to this increased constrictor response (189), and pharmacological inhibition of the ETA receptor during pregnancy prevents maternal hypertension in this model (10, 289). These studies have also shown high mRNA concentrations of the ET-1 precursor preproendothelin in the renal cortex and medulla of the RUPP rat and a trend for improved renal function with ETA receptor blockade in this model (10). Collectively, the data demonstrate a substantial role for changes in ET-1 synthesis and signaling in mediating maternal microvascular dysfunction during preeclampsia.
Thromboxane A2 is a potent constrictor prostanoid synthesized in the arachidonic acid metabolic pathway by thromboxane synthase in platelets and in the vascular endothelium. Similar to the vasodilatory PGI2, TXA2 is produced downstream of cyclooxygenases 1 and 2. However, the effects of TXA2 oppose those of PGI2 in the VSM. Specifically, TXA2 induces platelet aggregation and VSM contraction and cell proliferation via its action on the VSM thromboxane receptor. Thromboxane synthase expression is increased in endothelial and VSM cells of omental arteries from women with preeclampsia compared to omental arteries from women with a normal pregnancy (206). Furthermore, TXB2, the stable metabolite of TXA2, is elevated in plasma and urine from women with preeclampsia compared to controls (59), suggesting that TXA2 contributes to maternal microvascular dysfunction during preeclampsia. Animal models of preeclampsia also demonstrate a role for TXA2 in maternal microvascular dysfunction during preeclampsia. Mesenteric arteries from pregnant rats with gestational hypertension secondary to an immunologic exposure produce more TXA2 (assessed as TXB2 concentrations) (216) and have greater constrictor responses to a pharmacological thromboxane receptor agonist (109) compared to normal pregnant controls. Conversely, thromboxane synthase knockout prevents hypertension and fetal growth restriction in high-salt-treated pregnant mice (219).
The vascular actions of TXA2 are opposed by NO and PGI2 in the microvasculature, and both NO and PGI2 desensitize VSM thromboxane receptors (103). Because both NO and PGI2 are reduced in preeclampsia, this lack of counterbalance in the local microvascular milieu may contribute to exaggerated TXA2 constrictor sensitivity. In support, NOS inhibition increases mesenteric artery constrictor sensitivity in isolated vessels from healthy pregnant rats but not in a rat model of preeclampsia (109). While circulating measures of NO, PGI2, and TXA2 metabolites suggest a similar imbalance in human preeclampsia, mechanistic studies of the interplay between these pathways in human maternal microvessels have not been conducted.
Placental ischemia secondary to shallow trophoblast invasion of the spiral arteries and poor placental perfusion in preeclampsia leads to the release of antiangiogenic and proapoptotic factors that enter the maternal circulation and contribute to endothelial dysfunction in the nonreproductive vascular beds (Figure 4) (18). In normal pregnancy, VEGF and PlGF support proper angiogenesis and modulate placental and maternal vascular endothelial function (231). However, in preeclampsia, alterations in the signaling of these molecules, particularly via increased circulating concentrations of soluble receptors sFlt-1 and soluble endoglin and their ratio to PlGF in circulation, lead to the development of placental and maternal vascular dysfunction and likely drive the pathogenesis of the maternal preeclamptic phenotype (231).
VEGF refers to a family of glycoproteins that bind their transmembrane receptors to induce embryonic vasculogenesis and angiogenesis and support endothelial cell survival and vascular homeostasis in mature vessels and tissues (213). PlGF is a VEGF homolog released by the placenta with similar bioactivity to VEGF (178). VEGF exerts its vascular effects through two endothelial transmembrane VEGF-receptor 1 (also known as Flt-1) and VEGF-receptor 2 (also known as KDR). While VEGF binds both Flt-1 and KDR, PlGF binds exclusively to Flt-1 (221), and both induce vasodilation via EDRF, NO, and PGI2 (98, 213). sFlt-1, the soluble form of Flt-1, sequesters VEGF and PlGF and blocks VEGF and PlGF-induced dilation of the maternal microvessels (187). sFlt-1 is elevated in the circulation of women with preeclampsia (143) and is associated with increased severity of the maternal syndrome (187). Circulating concentrations of free VEGF and free PlGF are reduced in women with preeclampsia (169), and this reduction is largely mediated by excess sFlt-1 binding and sequestering rather than decreased PlGF production (163). Indeed, circulating sFlt-1 and sFlt-1/PlGF ratio have recently been demonstrated in clinical trials to have strong predictive value in the diagnosis of preeclampsia (56, 84) and the prediction of severe features of the syndrome (292).
VEGF plays a major role in microvascular health and organ function and is a key mediator in the maintenance of glomerular capillaries and hepatic vasculature (132). As such, elevated sFlt-1, and subsequently reduced VEGF and PlGF, is a likely mediator of maternal microvascular dysfunction during preeclamptic pregnancy. In support, pregnant rodents infused with sFlt-1 have reduced circulating free VEGF and develop the hallmarks of preeclampsia, including maternal endothelial dysfunction, glomerular endotheliosis, hypertension, and fetal growth restriction (32, 46, 187). Conversely, infusion with PlGF or VEGF improves maternal blood pressure in nonhuman primates and rodent models of preeclampsia (174, 181, 277). In humans, high sFlt-1 and elevated sFlt-1/PlGF ratio are positively associated with circulating ET-1 (4, 311) and inversely related to NO formation in preeclampsia (253), suggesting that this angiogenic imbalance drives maternal microvascular dysfunction during preeclampsia.
Soluble endoglin is another antiangiogenic protein implicated in the antiangiogenic milieu and maternal endothelial dysfunction associated with preeclampsia (184). Endoglin is a transforming growth factor-β1 (TGF-β1) coreceptor expressed in mature vascular endothelial cells and required for endothelial homeostasis and survival (321). TGF-β1 signaling in mature vessels induces endothelium-dependent dilation via endothelial NO synthase. Soluble endoglin exerts antiangiogenic and pro-hypertensive effects by binding TGF-β1 and decreasing TGF-β1 bioavailability, subsequently reducing TGF-β1 binding with endothelial cell receptors. Through these mechanisms, soluble endoglin prevents TGF-β1-induced vasodilation, in part through downregulation of endothelial cell NO synthase (184, 310). Indeed, soluble endoglin is elevated in the circulation of women with preeclampsia compared to normal controls (4, 168) and is inversely related to NO formation in preeclampsia (253). Interestingly, it appears that soluble endoglin works with sFlt-1 to induce maternal microvascular dysfunction, as mice injected with both soluble endoglin and sFlt-1 during pregnancy develop a phenotype of preeclampsia with severe features that was absent with injection of each factor alone (310).
The renin-angiotensin-aldosterone system (RAAS) plays a key role in the maternal cardiovascular adaptations to pregnancy (254). Normal pregnancy is characterized by an increase in renin, angiotensin, and aldosterone, all of which support increases in plasma volume during gestation. However, these increases are balanced by a decreased sensitivity of the maternal vasculature to the vasoconstrictive effects of angiotensin II (ang II), supporting the characteristic reduction in total peripheral resistance despite an increase in RAAS activity. Conversely, women with preeclampsia have an increased pressor response to ang II (79, 94, 202), and the vasoconstrictor response to ang II is increased in omental arteries from women with preeclampsia compared to normal pregnant controls (210). Interestingly, concentrations of circulating RAAS components are not elevated in preeclampsia as seen in healthy pregnancy, and may even be decreased compared to normal pregnancy (269), implicating downstream mechanisms and/or inappropriate balance in vascular ang II signaling in the maternal vasculature during preeclampsia.
The vascular actions of ang II are mediated by the ang II type 1 receptor (AT1R) and the counterregulatory ang II type 2 receptor (AT2R) and Mas receptor. AT1R located on VSM induces vasoconstriction, inflammation, vascular growth, and reduced NO release when bound by ang II (90). These classical constrictor responses are counterbalanced by ang II binding to AT2R and angiotensin 1–7 binding to Mas on endothelial cells, which elicit NO-dependent vasodilation and anti-inflammatory, antioxidative, and antifibrotic responses (90, 201). AT1R expression is increased in placental VSM (162) and platelets (223) from women with preeclampsia, suggesting that increased receptor expression may contribute to increased sensitivity to ang II. In counterbalance to AT1R-mediated constriction, AT2R expression and sensitivity increase in the placenta and maternal vascular beds during normal pregnancy, reducing peripheral vascular resistance to facilitate enhanced blood flow (14, 198, 200, 284). Animal models suggest that AT2R expression and AT2R-mediated vasodilation are reduced in preeclampsia, contributing to vascular dysfunction and AT1R hypersensitivity (60, 61, 74). Furthermore, 5 days of the AT2R-specific agonist compound 21 upregulated AT2R, downregulated AT1R, and restored fetal and placenta weight in a testosterone-induced mouse model of preeclampsia (199). Additionally, 24 h of compound 21 treatment on ex vivo placental vessels and uterine artery endothelial cells isolated from preeclamptic women elicited similar beneficial changes in ang II receptor expression (197, 199). Immediately following delivery, women with preeclampsia exhibit increased AT1R and decreased AT2R expression in the placenta and vascular endothelial cells compared to women with a healthy pregnancy (130, 164). Collectively, these data suggest that an imbalance in the counterregulatory AT2R contributes to increased constrictor tone and maternal microvascular dysfunction during preeclampsia.
Circulating angiotensin 1–7 is increased during normal pregnancy and reduced during preeclampsia in humans (193). Renal angiotensin 1–7 is upregulated in normal pregnant rats compared to virgin nonpregnant rats and reduced in RUPP rats compared to normal pregnant controls (129). Uteroplacental Mas receptor expression is increased in early pregnancy in normal pregnant rats, but whether this is different in RUPP rats is equivocal (63, 343). Cotreatment with angiotensin 1–7 prevents damage to cultured human podocytes treated ex vivo with serum from preeclamptic women, suggesting that changes in circulating angiotensin 1–7 contribute to renal dysfunction associated with preeclampsia (294). Angiotensin 1–7 induce vasodilation in the microvasculature via endothelium-dependent NO synthesis (258, 264, 278), but whether this mechanism contributes to reduced vascular resistance in normal pregnancy or is dysregulated in preeclampsia remains to be determined.
Agonistic autoantibodies to the AT1R are elevated in plasma from women with preeclampsia (120, 270, 318, 319). Chronic infusion of AT1-AA into pregnant rats induces a preeclamptic phenotype (152, 331). Additionally, AT1-AA are shown to be increased endogenously in the RUPP model (119, 155), similar to that seen in women with preeclampsia. AT1-AA infusion during pregnancy in rats mediates increased maternal blood pressure via ET-1 (152), while blockade specific to AT1-AA or indirectly via AT1R inhibition with losartan prevents hypertension and improves pathophysiology in the RUPP rat model of preeclampsia (70, 301, 322). Endothelial responses to acetylcholine are reduced in isolated renal interlobar arteries from pregnant rats infused with AT1-AA (222), and AT1-AA infusion also impairs cerebral autoregulation in this model (327), highlighting the role of AT1-AA in microvascular dysfunction of the nonreproductive vascular beds. Indeed, endothelial cells treated with AT1-AA isolated from preeclamptic plasma release lactate dehydrogenase and have increased caspase activities (345), suggesting that AT1-AA directly contributes to endothelial cell dysfunction and injury in preeclampsia (222). AT1-AA increases ET-1 synthesis and oxidative stress and likely modulate vascular tone in pregnant women with preeclampsia via these mechanisms (152, 153). Additionally, AT1-AA infusion in pregnant rats also increases circulating sFlt-1 and soluble endoglin, suggesting that AT1-AA contributes to the detrimental vascular effects of these circulating antiangiogenic factors via their increased production (153).
Inappropriate immune activation has been proposed as a causative factor in the initiation of preeclampsia and the maternal microvascular dysfunction that drives the maternal presentation of the syndrome (151). Women with preeclampsia have elevated circulating inflammatory mediators (91, 242, 256), including TNF-α, IL-17 (204, 350), and LIGHT (325), a TNF superfamily member, and unchanged (204) or decreased (350) anti-inflammatory mediator IL-10. Furthermore, the magnitude of this immune imbalance has been demonstrated to associate with the severity of the maternal syndrome (350). Whether this increased inflammation directly impairs maternal microvascular function in humans is unexplored. However, RUPP rats have increased plasma TN-Fα (154) and IL-6 (316), and infusion of these factors into normal pregnant rats causes hypertension, proteinuria, and maternal vascular dysfunction (93, 156). Similarly, infusion of LIGHT into pregnant mice increases blood pressure and sFlt-1 expression and induces proteinuria (325). Both LIGHT and TNF-α infusion into pregnant rodents increase ET-1 protein expression in the maternal vasculature (154, 325), and ETA blockade prevents the development of a preeclamptic phenotype in pregnant rats exposed to TNF-α infusion (157). Similarly, TNF-α blockade with etanercept prevents the increase in blood pressure and proteinuria in pregnant RUPP rats, and endothelial cells treated with serum from RUPP rats plus etanercept secreted less ET-1 than cells treated with serum alone (154). Collectively, these data suggest that immune imbalance in preeclampsia contributes to maternal endothelial cell activation, through ET-1-dependent mechanisms, to cause maternal microvascular dysfunction and the clinical presentation of the syndrome.
Oxidative stress, characterized by an imbalance in reactive oxygen species production and antioxidant capacity to scavenge these species, leads to endothelial dysfunction through direct signaling by oxidative species or the disruption of normal signaling mechanisms favoring the production of EDCF and reducing production and bioavailability of EDRF in the vasculature (338). During preeclampsia, oxidative stress is elevated in placental and maternal tissues (297, 320) and may mediate maternal endothelial cell dysfunction and contribute to the pathophysiology of the syndrome. Specifically, the preeclamptic placenta demonstrates increased production of oxidant molecules including lipid peroxides, isoprostanes, and reactive oxygen species and reduced antioxidant mechanisms (78, 252, 320). Furthermore, the antioxidant capacity in maternal plasma is reduced in preeclampsia (207). Vessels isolated from women with preeclampsia have increased nitrotyrosine, a cellular marker of oxidative stress, and reduced superoxide dismutase compared to vessels from women with a normal pregnancy (247). VSM cells incubated with AT1-AA from preeclamptic women increased reactive oxygen species production mediated by NADPH oxidase (78). Furthermore, expression of arginase, an enzyme that competes with NO synthase for l-arginine and increases superoxide formation by NO synthase, is increased in maternal vessels from women with preeclampsia and contributes to increases in vascular oxidative stress (255). Preclinical models of preeclampsia similarly demonstrate that preeclampsia is characterized by maternal exposure to high oxidant stress. RUPP rats have increased plasma and renal markers of oxidative stress (21, 265) and lower plasma antioxidant capacity (21) compared to normal pregnant rats. Furthermore, the increase in oxidative stress can be attenuated by treatment with the superoxide scavenger tempol or pravastatin, both of which reduce the blood pressure response to reduced uterine perfusion during pregnancy (21, 265). Further work established that RUPP rodents have increases in mitochondrial oxidative stress in vascular- and renal-derived mitochondria, and treatment with the mitochondrial-specific antioxidants MitoQ and MitoTEMPOL prevents hypertension in this model (302). Interestingly, clinical trials of antioxidant vitamins C and E have been negative for the prevention of preeclampsia in pregnant humans, despite evidence that these antioxidant compounds are reduced in preeclampsia (230, 249). Similarly, treatment with pravastatin failed to reduce the incidence of term preeclampsia compared to placebo in pregnant women at high risk (82). Collectively, data from human and preclinical studies suggest that increases in oxidative stress contribute to maternal vascular dysfunction during preeclampsia. However, unlike preclinical models, effective approaches to target or reduce oxidative stress have failed to show efficacy in prevention or treatment of preeclampsia in human clinical trials.
Despite remission of the clinical symptoms after delivery, women with a history of preeclampsia have a four- to eight-fold higher lifetime risk of developing cardiovascular disease compared to women who had an uncomplicated pregnancy (Figure 5). This includes increased risks of fourfold for hypertension (26); twofold for ischemic heart disease, heart failure, stroke, and venous thromboembolism (126, 138, 340); and ninefold for early vascular aging (332). Importantly, these adverse vascular outcomes arise prematurely, ~10 years earlier than that normally experienced in women (116, 161, 240), and results in a higher overall mortality rate compared with women who had an uncomplicated pregnancy (26, 68, 203). These risks are evident within 1 to 3 years postpartum and remain decades later (116, 126, 203). In addition to the direct relation to increased CVD risk, a pregnancy complicated by preeclampsia is also associated with an increased risk of metabolic dysfunction (36, 217, 283, 285), including alterations in metabolic disease risk factors such as insulin resistance and dyslipidemia (65, 102, 124, 286, 336), all of which are traditional risk factors for CVD. Recently, the relation between preeclampsia and chronic renal disease has become increasingly evident. A major consequence of endothelial dysfunction during preeclampsia are impairments in renal function, exacerbated by the already augmented renal demands during healthy pregnancy. Indeed, there is a >9-fold risk of developing chronic renal disease in women who had preeclampsia, even when controlling for prepregnancy risk factors (20, 72, 87, 165, 323). In the absence of diagnosed renal disease, women with a history of preeclampsia have renal impairments including reduced renal blood flow, attenuated GFR, and proteinuria (2, 67, 131, 134, 185, 190, 286).
Although less than 6% of preeclampsia diagnoses progress into eclampsia (1), impairment in cerebrovascular structure and function persists in women with a history of preeclampsia, even in the absence of seizures during pregnancy (22, 71, 194, 229, 271). Changes in cerebral microstructure have been observed as white matter lesions after preeclampsia (271, 333), and women with a history of preeclampsia may have functional changes in the cerebral circulation such as increased blood brain barrier permeability (54) and reduced cerebrovascular reactivity (19, 267). Importantly, the severity of changes is proportional to the time since pregnancy (271), implying that persistent microvascular dysfunction continues to impair the cerebrovasculature in the absence of overt symptoms. Indeed, women who had preeclampsia consistently score worse on tests of cognitive function in the years and decades after delivery (9, 194, 195, 209, 267). Similarly, a history of preeclampsia is also associated with a two- to four-fold increased risk of retinal disorders and other ophthalmic complications (16, 23). Following preeclampsia, retinal endothelial dysfunctions (e.g., impairments in the retinal flicker response (47)) and alternations in retinal microvasculature (27, 275) are present in women with a history of preeclampsia. Interestingly, retinal microvascular endothelial dysfunction is an independent predictor of coronary artery disease (11, 339) and CVD (172, 347), further supporting the notion that reduced microvascular function in nonreproductive vascular beds persists long after the affected pregnancy.
It is clear that maternal vascular dysfunction persists post-partum in the absence of overt disease and likely contributes to the accelerated development of chronic disease in these women. Although delivery of the placenta resolves the primary syndrome in women, residual maternal vascular dysfunction remains in the absence of the placenta, suggesting that the long-term disease risk arises from dysfunction in the vasculature. As such, studies have focused on the mechanistic underpinnings of vascular dysfunction that persist after pregnancy but before the development of CVD, including components of the RAAS, ET-1, antiangiogenic factors such as sFlt-1, and chronic inflammation. However, further research is required to delineate precise mechanisms underpinning this dysfunction to identify viable therapeutic targets to reduce or prevent the development of overt disease. Few studies have examined residual microvascular dysfunction in postpartum women, and only recently have postpartum preclinical models emerged as a tool to delineate the mechanisms driving this dysfunction. Therefore, the following sections discuss the current understanding of pathways contributing to reduced vascular function and increased CVD risk postpartum, including preclinical models of postpartum vascular function and current human clinical trials aimed at restoring vascular function in these high-risk women.
Otherwise healthy women with a history of preeclampsia demonstrate reductions in endothelial function (214, 231, 328) to a magnitude consistent with reductions that are observed in populations with traditional CVD risk factors such as hypertension, dyslipidemia, and aging (105, 272, 341). It is well established that women who had preeclampsia have significant reductions in brachial artery flow-mediated dilation, a measure of NO-mediated endothelium-dependent vasodilation in conduit vessels, which is present within 1 year of delivery (208, 299), multiple years postpartum (44, 57, 101, 114), and even decades later (117). However, some studies have suggested that not all formerly preeclamptic women present with reductions in macrovascular endothelial function (128). Several studies have consistently demonstrated similar reductions in microvascular endothelial function, which has been shown to be a more potent predictor of 5-year CVD risk than endothelial function assessed in the conduit vasculature (159, 173, 214, 272). Indeed, endothelial dysfunction in the microvasculature precedes and predicts morphological changes in the conduit vessels (77, 159, 220, 228, 347). Importantly, vascular endothelial dysfunction is reversible with lifestyle or pharmacological interventions (15, 37, 73). Therefore, endothelial dysfunction present in otherwise healthy women with a history of preeclampsia highlights a window of opportunity for intervention to slow or prevent the development of overt CVD. In support, studies have suggested that this window of opportunity is open any time prior to overt disease development, from short postpartum through the 5^th^ decade of life (35, 215).
As previously discussed, the balance of EDRF and EDCF is disrupted during preeclampsia and plays a functional role mediating systemic microvascular dysfunction. Similarly, women with a history of preeclampsia have a proconstrictor vascular phenotype (215, 303). However, the mechanistic roles of EDRF and EDCF in postpartum vascular function after preeclampsia remain largely understudied (Figure 6).
Various studies have demonstrated that residual microvascular dysfunction in women after preeclampsia is endothelium mediated, including in the forearm (3, 88), capillaries (34), cutaneous (214, 234, 238, 278–281), and coronary beds (148). Specifically, our group has demonstrated that microvascular function as measured in the cutaneous circulation, a model of global microvascular function (77, 122), is attenuated in otherwise healthy women with a history of preeclampsia via reductions in endothelium- and NO-dependent dilation (234, 278–281). In support, serum nitrates, markers of NO metabolism, are lower in formerly preeclamptic women 3 years postpartum (101). The RUPP model to induce preeclampsia in rats demonstrates reduced endothelium-dependent dilation by reductions in NO bioavailability 3 months postpartum (45). Immunodeficient mouse models of preeclampsia have reduced acetylcholine-mediated vasorelaxation in mesenteric arteries (38) and internal carotid arteries (106) 4 weeks postpartum, mediated by reduced NO-dependent dilation secondary to reduced eNOS and increased iNOS expression in vascular endothelial cells (106). Similarly, mice with kidney complement knockout that develop a preeclampsia-like phenotype demonstrate reduced endothelium-dependent dilation in aortic rings (96) and mesenteric arteries (99, 287) and display reductions in glycocalyx function (99) after pregnancy. Collectively, these data demonstrate that reduced NO-dependent responses contribute to endothelial dysfunction in women with a history of preeclampsia. However, the molecular mechanisms underlying this dysfunction remain largely unclear and warrant further investigation.
Both animal and human studies have demonstrated dysregulation in the endothelin system as a mechanism contributing to vascular dysfunction after preeclampsia. Independent of pregnancy, healthy premenopausal women have reduced sensitivity to the vasoconstrictor actions of ET-1 (188, 295). Compared to age-matched men, premenopausal women have reduced circulating ET-1 (226) and greater expression and sensitivity of endothelial ETB1 (6, 86, 136). Activation of the vasodilatory ET-1 axis is upregulated in the presence of estradiol and progesterone (31, 268, 330) leading to a prodilatory vascular phenotype, while ET-1-mediated constriction enhances vascular tone in the presence of testosterone (224, 334), in OVX animal models (104, 141, 290), and in postmenopausal women (147, 329). However, like the mechanistic contribution of ET-1 to maternal vascular dysfunction during preeclampsia as described above, formerly preeclamptic women have increased vasoconstrictor sensitivity to ET-1 after pregnancy. Using the cutaneous circulation as a model, our group has shown that otherwise healthy women with a history of preeclampsia have an exaggerated vasoconstriction response to ET-1 compared to matched women with a history of uncomplicated pregnancy (280). Interestingly, we found that inhibiting the ETB receptor with BQ-788 decreased constrictor sensitivity in women with a history of preeclampsia, and that these women had reduced dilator responses to the ETB-specific agonist sarofotoxin. We also found no differences in ETA expression between groups but an increase in ETB expression in women who had preeclampsia, presumably representative of increased ETB2 on the VSM. Furthermore, pharmacological blockade of ETB improved endothelium- and NO-dependent dilation in women with a history of preeclampsia. Collectively, these data suggest that alterations in ET-1-mediated vascular responses after preeclampsia are mediated by changes in ETB receptor function and expression and contribute to endothelial dysfunction in these women. Notably, we found these alterations in the absence of differences in circulating ET-1 concentrations between groups. In agreement, other groups have reported no difference in circulating ET-1 6 years postpartum (158), suggesting that the role of ET-1 in maternal vascular dysfunction after pregnancy is mediated by changes in vascular sensitivity and receptor expression rather than increased circulating ET-1 concentrations.
In preclinical models, contrasting findings of the role of ET-1 have been reported in the postpartum period. Consistent with our findings, de Alwis et al. found that increased ET-1 concentrations present during pregnancy in the L-NAME mouse model of preeclampsia did not persist at 1 week postpartum and remained unchanged up to 10 weeks postpartum (76). However, they also found no significant differences in ETA or ETB gene expression at any point up to 10 weeks postpartum. The authors suggest that the L-NAME model may not be representative of human preeclampsia or the postpartum vasculature due to the cessation of L-NAME infusion at delivery likely allowing recovery in endothelial NO production (76). Conversely in the C1q knockout (C1qKO) mouse model of preeclampsia, serum ET-1 was increased during preeclampsia and 60 days postpartum compared to wild-type mice at both timepoints (96). Interestingly, hypertension in the preeclampsia-like mice resolved following delivery but progressively increased and was significantly elevated at 30 days postpartum and remained elevated through day 60. Another group utilizing a RAAS transgenic rat model to induce preeclampsia reported increased gene expression of ET-1 during preeclampsia that remained elevated 4 weeks postpartum (144). However, circulating ET-1 was not reported. Both studies utilized pravastatin as a treatment during pregnancy and found that treatment during pregnancy reduced serum ET-1 during preeclampsia and postpartum to similar concentrations measured in wild-type and control mice during pregnancy (96). Pravastatin did not influence ET-1 gene expression in transgenic rats at either timepoint (144). Collectively, the preclinical and human data suggest that alterations in ET-1 signaling are present after preeclamptic pregnancy, and contribute to endothelial dysfunction in women with a history of preeclampsia (280), and may contribute to increased lifetime CVD risk in these women.
Imbalance in the classical and counterregulatory RAAS axes is a putative mechanism underlying maternal microvascular dysfunction after preeclampsia (Table 1). Women with a history of preeclampsia are more likely to have salt-sensitive blood pressure responses, which are amplified by aberrant RAAS responsiveness to hypertensive stimuli. Women who had severe preeclampsia have higher ambulatory blood pressure and a blunted nighttime dip in blood pressures (2, 261), which is exacerbated on a high-salt diet (185), compared to controls with a history of normal pregnancy. This is accompanied by a rightward shift of the pressure-natriuresis relation in the preeclampsia group only (185), suggesting that formerly preeclamptic women require renal compensation on a high-salt diet, which may predispose these women to renal complications and early vascular aging.
A pivotal study by Saxena et al. demonstrated that women with a history of preeclampsia have an exaggerated pressor and adrenal response to systemic ang II infusion compared to those who had a normotensive pregnancy (260). These group differences were found only following a low-salt diet, as the pressor and renal (via aldosterone production) responses to systemic ang II were not different between postpartum groups following the high-salt diet. This suggests that women with a history of preeclampsia have an increased sensitivity to ang II even in the absence of the high-salt stimulus, and this increased basal ang II sensitivity likely contributes to the increased salt sensitivity of BP in these women after pregnancy. Similarly, Biwer et al. found that formerly preeclamptic women ~4 years postpartum have a greater salt sensitivity of systolic blood pressure than matched controls (32). The authors noted that while some women in the control group demonstrated salt-sensitive blood pressure responses, only a history of preeclampsia was collectively associated with a salt sensitivity of blood pressure postpartum. In the same study, mice exposed to sFlt-1-infusion during pregnancy had an increased pressor response to both high-salt diet and ang II infusion compared to control mice at 2 months postpartum (32). This is similarly demonstrated in the RUPP rat model of preeclampsia, which has increased blood pressure sensitivity to high-salt diet 3 weeks postpartum (186). These preclinical models suggest that increased sensitivity to pressor stimuli after preeclamptic pregnancy is mediated via dysregulated RAAS responsiveness and is attributable to the detrimental long-term effects of preeclampsia, independent of preexisting CVD risk.
Dysregulation of ang II signaling is one of the most studied mechanisms underlying vascular dysfunction in formerly preeclamptic women due to the key role the RAAS plays during pregnancy. To elucidate the mechanisms underlying increased ang II sensitivity postpartum, our group has demonstrated that increased ang II-mediated vasoconstriction is present in the cutaneous microvasculature, and that local inhibition of AT1R with losartan improved endothelium- and NO-dependent dilation in formerly preeclamptic women (281). Additionally, we found increased expression of AT1R in the skin, suggesting that increased AT1R expression contributes to increased ang II sensitivity postpartum. Interestingly, in preclinical models, ang II-mediated constriction was not different in aortic rings collected 6 weeks postpartum from rats treated with endotoxin while pregnant compared to postpartum control and never-pregnant rats (305). However, incubation of aortic rings with losartan reduced AT2R-mediated relaxation to ang II in formerly preeclamptic rats, but no difference was found between groups in AT1R-mediated contraction following incubation with PD-123319 (AT2R antagonist), suggesting a role for both AT1R and AT2R in ang II sensitivity after preeclampsia. Similarly, a greater pressor response to systemic ang II infusion and a significant relation between the change in MAP and the ratio of AT1R/AT2R expression were found in cutaneous biopsy samples from formerly preeclamptic women that were not observed in the matched normotensive postpartum or never-pregnant control groups (121). Although there were no differences in absolute ang II receptor expression between any of the groups (121), these findings suggest that differences in the sensitivity of AT1R and AT2R likely contribute to enhanced ang II vasoconstrictor responsiveness after preeclampsia.
We utilized angiotensin 1–7 to locally and acutely activate the vasodilatory Mas receptors in the cutaneous microvasculature and found that angiotensin 1–7 treatment reduced the exaggerated ang II-mediated constriction response and augmented endothelium- and NO-dependent dilation in women with a history of preeclampsia but had no effect in controls (278). Within our findings of exaggerated vasoconstrictor sensitivity postpartum, we demonstrated specificity to ang II-dependent mechanisms as we found no difference in the vasoconstriction response to the adrenergic agonist norepinephrine in formerly preeclamptic women (280, 281). Similar to ET-1, the increased vasoconstrictor sensitivity to ang II is not accompanied by increases in circulating ang II in postpartum women (121, 276, 304) or mice (96). Ang II binds with higher affinity to AT2R (40), but whether the increased vasoconstrictor sensitivity to ang II is due to both reduced AT2R expression and enhanced AT1R expression postpartum in women with a history of preeclampsia is unknown. Collectively, the data suggest that signaling downstream of ang II, such as differences in ang II receptor expression and sensitivity, mediates exaggerated ang II-dependent constriction in the maternal microvasculature postpartum.
The mineralocorticoid receptor (MR) is one of the terminal effectors of the RAAS and is involved in blood pressure control via renal sodium retention following activation by aldosterone binding (48). Aldosterone is synthesized and released by the adrenal cortex predominantly in response to ang II and K^+^ as well as adrenocorticotropin and vasopressin (29, 55). Although there are no differences in basal circulating aldosterone, women with a history of preeclampsia have a greater increase in aldosterone concentrations in response to systemic ang II infusion compared to matched postpartum controls (260). In the sFlt-1-treated pregnant mouse model of preeclampsia, postpartum dams exhibit augmented ang II-mediated increases in blood pressure, mediated by VSM MR-induced microvascular dysfunction and upregulation of AT1R expression (32). This response was not observed in MR knockout mice that received the same sFlt-1 infusion during pregnancy. The same study found that VSM cells exposed to sFlt-1 in vitro had enhanced MR transcriptional activity in the presence of ang II or aldosterone, which also resulted in increased AT1R expression and decreased AT2R expression (32). Together, these findings suggest that preeclampsia exposure, specifically high sFlt-1 during pregnancy, enhances future responsiveness to hypertensive stimuli via ang II-mediated increases in aldosterone and MR activity postpartum.
Another mechanistic underpinning of enhanced AT1R-mediated vascular dysfunction after pregnancy is an increase in circulating AT1-AA. Plasma AT1-AA contribute to maternal vascular dysfunction in pregnancies complicated by preeclampsia (120, 270) and remain elevated years postpartum (125, 244, 305). This may be due to the immunological memory of B2 cells (119). More recently, the long-term effects of elevated AT1-AA on vascular dysfunction have been demonstrated up to 16 weeks after preeclampsia in animals (39, 322). Rats induced preeclampsia via the RUPP procedure display increased mean arterial pressures, cardiac hypertrophy, and reduced cardiac mitochondrial function 10 weeks postpartum (39). However, in a group of RUPP rats that were also treated with an AT1-AA inhibitor during pregnancy, postpartum blood pressure and cardiac function were not different from postpartum controls (39), supporting the hypothesis that AT1-AA-induced vascular remodeling brought about during preeclampsia perpetuates persistent vascular damage postpartum. Another group utilizing AT1-AA injection during pregnancy found reductions in cardiac structure and function 16 weeks postpartum, including an increased susceptibility to ischemia reperfusion injury (322). Treatment with losartan during pregnancy ameliorated the long-term maladaptation of preeclampsia in rats (322). However, losartan is contraindicated in human pregnancy (30, 42) and is not feasible as a treatment during the affected pregnancy. Similarly, AT1-AA infusion during pregnancy was found to impair cerebral blood flow autoregulation in rats 10 week postpartum (53), suggesting that elevated AT1-AA during pregnancy has implications in the cerebrovasculature following preeclampsia. Additionally, AT1-AA increases inflammation and oxidative stress by generating ET-1, sFlt-1, and reactive oxygen species (78, 152, 270), further altering vascular function. Collectively, these data suggest that elevated AT1-AA during pregnancy contribute to microvascular dysfunction in nonreproductive vascular beds that persists postpartum, and that women who continue to produce AT1-AA in the years following the affected pregnancy may be at a greater risk of cardiovascular disease and have an even greater enhanced responsiveness to vascular stressors.
As described above, sFlt-1 is an antiangiogenic factor that is increased in ~90% of preterm cases of preeclampsia and is a major driver of maternal endothelial dysfunction during pregnancy. During preeclampsia, an increased sFlt-1/PlGF ratio is primarily driven by increased sFlt-1 (92), implying that sFlt-1 is both a biomarker for preeclampsia and an effector of the disease (187, 313). sFlt-1 is produced by the placenta, the major source during pregnancy, and endothelial cells (123, 187). As such, sFlt-1 can be produced following the delivery of the placenta and continue to affect the maternal endothelium postpartum (187). Concentrations of sFlt-1 decrease half as fast after delivery in pregnancy complicated by preeclampsia compared with normotensive pregnancy and remain elevated 48 h (232) and 1 week (225) postpartum. This suggests that sFlt-1 continues to be produced despite delivery of the fetal-placental unit and may promote vascular endothelial damage in the postpartum period. Several studies have found increased circulating sFlt-1 concentrations in the years following preeclampsia (7, 125, 150, 225, 298, 336), whereas others have reported no difference compared to women who had an uncomplicated pregnancy (97, 101, 218, 346). However, there are no differences in circulating PlGF after preeclampsia compared to normal pregnancy (7, 97, 150, 218, 346). Interestingly, women with a history of preeclampsia have an increase in sFlt-1 production in response to systemic ang II infusion, while women who had a healthy pregnancy have a reduced production of sFlt-1 to the same stimulus (260). Additionally, sFlt-1 concentrations have been shown to be positively correlated with markers of endothelial damage (225) and arterial aging (7, 97) after preeclampsia, suggesting that sFlt-1 contributes to maternal microvascular dysfunction following preeclampsia.
Preclinical models utilizing sFlt-1 injection during pregnancy demonstrate vascular dysfunction postpartum, including increases in vascular MR and AT1R signaling (32), reduced AT2R expression (32), increased vascular plaque inflammation (33), greater neuroinflammation and blood-brain barrier permeability (315), enhanced vascular responsiveness to future injury (32, 233), and an altered plasma proteome via enrichment of provascular disease and inflammatory proteins (50). Additionally, sFlt-1 concentrations are elevated postpartum in C1qKO mouse (96) and RAAS transgenic rat (144) models of preeclampsia. Collectively, the data from human and rodent models demonstrate that elevated sFlt-1 concentrations during pregnancy may remain elevated postpartum; however, the role of sFlt-1 on persistent microvascular function postpartum requires future study.
Women with a history of preeclampsia have elevated circulating inflammatory markers such as CRP (118, 124, 150, 291), IL-6 (35, 291), IL-6/IL-10 ratio (91), ICAM-1 and VCAM-1 (35, 259, 298), and lipoprotein-associated phospholipase A2 (Lp-PLA2) (351) in the absence of overt CVD or inflammatory disease. Similarly, an exaggerated inflammatory response to the seasonal flu vaccine has been observed in formerly preeclamptic women (306), suggesting that exaggerated reactivity to inflammatory stimuli may contribute to early vascular aging in these women. Our group has shown that 4 days of systemic treatment with the nonsteroidal anti-inflammatory drug salsalate improved endothelium-dependent vasodilation in otherwise healthy women with a history of preeclampsia (279), implying that tonic inflammation contributes to maternal microvascular dysfunction even after pregnancy. Animal models suggest that an elevated pro-inflammatory state persists after preeclampsia. In the sFlt-1 infusion model of preeclampsia, mice that appeared otherwise similar to the non-sFlt-1-infused controls have an exaggerated atherosclerotic inflammatory response to 8 weeks of a high-fat diet initiated immediately after delivery (33). sFlt-1-induced preeclamptic mice have increased vascular remodeling and an enhanced inflammatory response to vascular injury 2 months postpartum (233) and an increased pro-inflammatory plasma proteome 6 months postpartum (50). Additionally, immune suppression with abatacept during sFlt-1-induced preeclampsia prevented systemic inflammation and reductions in blood-brain barrier permeability 6 weeks postpartum (315). Compared to control pregnant mice, mice with immune deficiency during pregnancy have enhanced vascular remodeling 4 weeks postpartum (38). Similarly, a group utilizing STOX1 overexpressing mice to mimic preeclampsia found that reductions in vascular function and structure 8 months postpartum are mediated, in part, by a pro-inflammatory transcriptome profile (196). Inflammatory markers have been shown to be elevated in the C1qKO model 60 day postpartum (96), and there is increased neuroinflammation in RUPP rats 2 months postpartum, despite the remission of the preeclamptic phenotype (64). Additionally, COMT knockout mice treated with an anti-inflammatory apolipoprotein A1 mimetic at the end of pregnancy have improved arterial stiffness compared to untreated COMT knockout mice (171) although this study was limited to 10 days postpartum. Taken together, subclinical vascular dysfunction in otherwise healthy women with a history of preeclampsia that persists postpartum is likely mediated, in part, by an increased pro-inflammatory state causing or exacerbating vascular injury. Importantly, preclinical models suggest that this is due to insult during preeclampsia (i.e., independent of the preexisting risk factors before pregnancy) and may work in a feed-forward manner as shown by increases in circulating biomarkers postpartum.
Oxidative stress is elevated during preeclampsia but less is known about the pro-oxidative state postpartum. Increased oxidant stress is likely, but few studies have addressed oxidative stress directly. Measures of oxidative stress decline with and following delivery but have been shown to remain elevated 8 to 14 weeks postpartum in women who had preeclampsia compared to those with an uncomplicated pregnancy (237, 312). At 3 years postpartum, systemic antioxidant treatment with ascorbic acid improved reductions in brachial artery FMD in formerly preeclamptic women but had no effect in matched healthy women (57), suggesting that increased oxidative stress contributes to endothelial dysfunction after preeclampsia. Additionally, lymphoblasts collected from women with a history of preeclampsia have greater NADPH oxidase-mediated reactive oxygen species production than cells collected from women who had a normotensive pregnancy (166). Preclinical models demonstrate that treatment with l-citrulline, which reduces reactive oxygen species formation by increasing NO production, during preeclampsia improves vascular endothelial and glycocalyx function in mice 3 and 7 months postpartum (99). Mice fed a high-cholesterol diet during late pregnancy have increased oxidized low-density lipoproteins (oxLDL) (251), similar to that seen during human preeclampsia, and have elevated vascular oxidative stress, including increased nitrotyrosine and reduced superoxide dismutase, 3 months postpartum despite returning to a normal chow diet after pregnancy (250). Further, AT1-AA infusion and RUPP rodent preeclampsia models have increased renal and mitochondrial oxidative stress postpartum (39, 135, 349) and demonstrate that postpartum oxidative stress can be reduced with inhibition of SGLT-2 (349) or AT1-AA (39) during pregnancy. Similarly, treatment with pravastatin during preeclampsia in mice has been shown to improve postpartum endothelial function and markers of oxidative stress compared to untreated mice (96). As previously discussed, treatments to reduce oxidative stress during pregnancy have failed to translate to human studies; however, the impact of these therapies initiated during pregnancy or postpartum on postpartum microvascular function in humans remains unknown.
Preeclampsia is a complex and heterogeneous disorder of pregnancy with acute and chronic implications for maternal morbidity and mortality. During pregnancy, placental-derived antiangiogenic factors enter the maternal circulation and initiate a cascade of mechanisms driving widespread maternal microvascular endothelial dysfunction. The brain, kidneys, liver, and heart are particularly vulnerable to this dysfunction, and their injury characterizes the clinical presentation(s) of the syndrome. After pregnancy, despite delivery of the placenta and the remission of the clinical characteristics of preeclampsia, maternal microvascular dysfunction persists and likely contributes to the dramatic lifetime disease risk in women who experienced a preeclamptic pregnancy. Specifically, women with a history of preeclampsia are at greater risk for cardiovascular, renal, cerebrovascular, and neurological diseases. These short- and long-term consequences highlight the need for mechanism-specific approaches to prevent and treat preeclampsia during pregnancy and to prevent, slow, or treat lifetime disease progression in these women. Preclinical models of preeclampsia demonstrate that decreasing circulating placental-derived antiangiogenic factors, reducing oxidative stress and inflammation, restoring balance to the RAAS and endothelin systems, and stimulating the nitric oxide pathway all improve maternal and, in some cases, fetal outcomes. However, targeting these mechanisms in human clinical trials has proven ineffective or pharmacologically impossible (e.g., in the case of teratogenicity of RAAS and ET-1 inhibitors) to date. Recent advances in maternally sequestered pharmacologic strategies hold promise for the safe targeting of these mechanisms in pregnancy (85, 100, 174), but these drug delivery systems are not yet tested or approved for human use. Aspirin therapy initiated during high-risk pregnancy does increase gestational time and improve maternal and fetal preeclampsia outcomes. It is hypothesized that aspirin does target the putative endothelial mechanisms mediating maternal microvascular dysfunction during preeclampsia, but the mechanisms underlying the efficacy of aspirin treatment during pregnancy remain unexplored in in vivo human trials. Postpartum studies have suggested that reducing inflammation or oxidative stress, attenuating the endothelin system, and blocking or opposing AT1R-mediated constrictor responses all acutely improve maternal vascular endothelial function in humans. However, clinical trials for the long-term treatment and primary prevention of CVD in women with a history of preeclampsia are lacking. Continued basic science and clinical work identifying, examining, and targeting the mechanisms underlying maternal microvascular dysfunction during and after preeclampsia remain an active and promising area of research.