Pregnancy demands major changes of the cardiovascular system, and this involves, among others, activation of the RAAS (renin–angiotensin–aldosterone system), allowing an aldosterone-dependent increase in volume. Remarkably, a relative resistance to the pressor response of AngII (angiotensin II) develops simultaneously to prevent the increase in blood pressure that would normally accompany RAAS activation. The increase in volume, the degree of RAAS activation and the diminished pressor response to AngII are less pronounced in pre-eclampsia. However, animal models displaying excessive RAAS activation also result in a pre-eclampsia-like syndrome, and the aldosterone/renin ratio is elevated in pre-eclampsia compared with a normal pregnancy. New insights into the pathogenesis of pre-eclampsia have revealed a major role for VEGF (vascular endothelial growth factor), VEGF-inactivating sFlt-1 (soluble fms-like tyrosine kinase-1) and AT1 (angiotensin II type 1) receptor autoantibodies. The last mentioned activate AT1 receptors, thereby potentially suppressing circulating renin and aldosterone. VEGF, both directly and indirectly (by increasing capillary density), affects adrenal aldosterone synthesis. The present review summarizes all of the recent findings regarding RAAS regulation in pre-eclampsia compared with normal pregnancy, concluding that factors such as sFlt-1 and AT1 receptor autoantibodies disturb the delicate balance that normally results in a volume increase and a diminished vasoconstrictor response to AngII in pregnant women. It is possible that there are non-parallel changes in the circulating and renal RAAS in pre-eclampsia, which are potentially reflected by the urinary levels of renin.
Pre-eclampsia is a condition unique to pregnancy characterized by new-onset hypertension and proteinuria after 20 weeks of gestation. Despite many years of research, the aetiology of the disease is still unknown. Hypertensive disorders, including pre-eclampsia, are among the most common complications of pregnancy, and they constitute a major cause of maternal, fetal and neonatal morbidity and mortality worldwide. Management of pre-eclampsia consists of treating the elevated blood pressure and the prevention of seizures when pre-eclampsia progresses to eclampsia, but there is no definite cure other than delivery .
Currently, two stages in the development of pre-eclampsia are distinguished. The first stage consists of poor placentation as a result of abnormally shallow cytotrophoblast invasion in the maternal spiral arteries supplying the placenta. As a consequence the extent of the widening of the spiral arteries is insufficient, causing impaired blood flow to the placenta. This leads to the second stage of pre-eclampsia, which consists of repeated periods of placental hypoxia, resulting in an increased production of ROS (reactive oxygen species), HIF-1α (hypoxia-inducible factor-1α) and TGF-β1 (transforming growth factor-β1). These factors, in turn, increase the production of a splice variant of the VEGF (vascular endothelial growth factor) receptor sFlt-1 (soluble fms-like tyrosine kinase-1). sFlt-1 can be found at high concentrations in the maternal circulation during pre-eclampsia and acts as an anti-angiogenic factor by binding free VEGF and PlGF (placental growth factor). A decreased availability of these factors affects the growth and development of the placenta and fetus, but it also affects the mother and is considered to be responsible for the clinical features of pre-eclampsia when present at increased amounts in the maternal circulation (Figure 1). In particular, the health of endothelial cells and kidney function including the maintenance of the glomerular filtration barrier strongly depend upon these growth factors .
Vascular system and perfusion of the placenta in a pre-eclamptic and a healthy pregnancy
Increased levels of sFlt-1 can be detected 5 weeks prior to the onset of the clinical symptoms, and decreased levels of PlGF can be detected even earlier. Many studies have already been performed investigating sFlt-1 and PlGF for the prediction and diagnosis of pre-eclampsia with promising results. The highest sensitivity and specificity for predicting and diagnosing pre-eclampsia are obtained using the sFlt-1/PlGF ratio, where a ratio above 85 was used as the cut-off for a positive test .
Since there is no cure or therapy to prevent pre-eclampsia, the usefulness of the test is currently limited to predict a poor pregnancy outcome in patients with pre-eclampsia and to distinguish between other diseases in pregnancy with pre-eclampsia-like symptoms [3,4].
Animal models confirm the importance of sFlt-1 in the pathogenesis of pre-eclampsia. In rats where the perfusion of the placenta is surgically reduced by partially clamping the blood flow to the placenta, increased sFlt-1 levels have been reported . Furthermore introducing sFlt-1 via a viral factor in pregnant rats gives a pre-eclampsia-like syndrome, with hypertension and proteinuria .
The RAAS (renin–angiotensin–aldosterone system) (Figure 2) is involved in both stages of the disease. Importantly, it is generally believed that some RAAS components, prorenin in particular, are synthesized in the uteroplacental unit, thus allowing local AngII (angiotensin II) production independently of AngII production in the systemic circulation (for an extensive overview see Herse et al. ).
The RAAS in pre-eclampsia compared with a normal pregnancy
The present review discusses pre-eclampsia-related alterations in various components of the RAAS, some of which were first reported >20 years ago, focusing in particular on their recently elucidated link with placental hypoxia and anti-angiogenic factors. Given its role as a regulator of blood pressure and fluid homoeostasis, changes in the RAAS should be evaluated in view of the changes in the cardiovascular system that occur in normal pregnancy and in a pregnancy complicated by pre-eclampsia. The haemodynamic changes and changes in the extracellular volume occurring during pregnancy and pre-eclampsia will therefore be reviewed first.
HAEMODYNAMIC CHANGES IN PREGNANCY AND PRE-ECLAMPSIA
Pregnancy demands major changes in the cardiovascular system. During pregnancy there is a 30–50% increase in the extracellular fluid and a 30–40% increase in plasma volume. The increase in plasma volume is extremely important as it is a major determinant of organ perfusion. The driving force for the increase in extracellular volume seems to be a decrease in the systemic vascular resistance, as reflected by a fall in the systolic and diastolic blood pressure in early gestation. It is generally believed that vasodilating factors such as nitric oxide play an important role in the decrease in vascular resistance . Underfilling of the circulation owing to generalized vasodilation may result in a compensatory activation of the RAAS, leading to water and sodium retention (Figure 1). Related to the increase in plasma volume and vasodilation, cardiac output increases by 30–50% during pregnancy via increases in both stroke volume and heart rate . Furthermore, the renal blood flow and glomerular filtration rate markedly increase during pregnancy, and peak at approximately 50% above non-pregnant levels in the second trimester . As compared with normal pregnancy, the expansion of plasma volume and the decrease in vascular resistance are less pronounced in pre-eclampsia. This decreased expansion in plasma volume can be detected as early as weeks 14–17 of gestation . Most probably as a direct consequence of the decreased plasma volume, cardiac output is also lower in pre-eclampsia than in normal pregnancy . Uterine and umbilical artery Doppler flow measurements have revealed that pregnancies complicated either by pre-eclampsia or growth restriction are characterized by a compromized uteroplacental flow , suggesting that the normally occurring expansion of plasma volume and increase in cardiac output are essential for the maintenance of a sufficient uteroplacental flow.
In summary, besides hypertension and proteinuria, pre-eclampsia, as compared with normal pregnancy, is characterized by a relatively high vascular resistance, a reduced intravascular volume, a reduced cardiac output and a decrease in uteroplacental flow. Remarkably, despite this reduction in circulating volume, most components of the circulating RAAS are down-regulated in pre-eclampsia compared with normal pregnancy. Whether this is a physiological response to the increased blood pressure or an active contributor to the pathophysiology of pre-eclampsia is an unresolved question.
Oestrogens stimulate the synthesis of angiotensinogen, resulting in an increase in the levels of circulating angiotensinogen during the first 20 weeks of gestation. Angiotensinogen levels in pre-eclampsia are comparable with those in normal pregnancy; however, there appears to be a relative increase in the so-called high-molecular-mass form of angiotensinogen in pre-eclampsia. When expressed as a percentage of the total angiotensinogen, this form is present at <5% in non-pregnant women. In normal pregnancy, it increases, possibly owing to release from the placenta, to 16% , whereas in women with pregnancy-induced hypertension it is 28% . Recently, it was reported that high-molecular-mass angiotensinogen is a complex of proMBP (proform of eosinophil major basic protein) with normal angiotensinogen . Importantly, the kinetics of the cleavage by renin of proMBP–angiotensinogen are dramatically impaired and may contribute to the reduced PRA (plasma renin activity) observed in pre-eclampsia .
Besides formation of this complex, two isoforms of the angiotensinogen molecule were found when solving its crystal structure at a 2.1 Å (1 Å=0.1 nm) resolution . These two forms represent the reduced and oxidized forms of a labile disulfide bridge in the molecule. Oxidation results in a structural change which preferentially interacts with (pro)renin receptor-bound renin. Given the increased oxidative stress in pre-eclampsia, the amount of oxidized angiotensinogen in this disease is increased . However, the resulting enhanced angiotensin generation by renin, when interacting with oxidized angiotensinogen, contrasts with the reduction in PRA that is normally observed in pre-eclampsia . Combined with recent studies raising doubt on the role of the (pro)renin receptor as a renin-binding receptor in vivo , the clinical significance of these findings therefore remains uncertain.
Polymorphisms of the angiotensinogen gene that increase plasma angiotensinogen levels have also been associated with pre-eclampsia . The underlying mechanism remains unclear, but it might involve the reduction in renin that is the consequence of such elevated angiotensinogen levels . Interestingly, in this regard, pregnant mice overexpressing murine angiotensinogen fail to maintain their volume overload, possibly because they were unable to up-regulate renin expression in the distal nephron . Whether this has relevance to the clinical syndrome of pre-eclampsia remains to be determined. Mating of female mice overexpressing human angiotensinogen with males expressing human renin (both of which cannot react with their mouse counterparts) leads to a pre-eclampsia-like syndrome with hypertension, proteinuria and the well-known histological abnormalities of the kidney . Importantly, this model relies on the release of placenta-derived human renin into the circulation, where it reacts with maternal angiotensinogen. When mating female mice overexpressing human renin with male mice overexpressing human angiotensinogen no such phenotype occurred. A similar transgenic animal model in rats also resulted in impaired spiral artery remodelling and a reduced placental perfusion . However, it should be realized that in both of these models the circulating RAAS activity is excessively high , as opposed to the human situation, were pre-eclampsia associates with a relatively low degree of RAAS activity.
PRORENIN AND RENIN
In early pregnancy there is a 4–5-fold increase in the levels of renin's inactive precursor prorenin. This prorenin is mainly derived from the ovaries and, to a lesser extent, from the uteroplacental unit and the kidneys [25–27]. Why the ovaries synthesize and release prorenin in such massive amounts is still unknown. Following the discovery of the (pro)renin receptor , it was thought that this receptor might bind and activate prorenin in vivo, thereby for the first time providing a function for prorenin. However, the nanomolar affinity of this receptor for prorenin implies that its (picomolar) concentrations in the plasma are several orders below the levels required for binding, even in pregnancy . Nevertheless, this receptor might play a role in prorenin-synthesizing organs, for instance the ovary, where such concentrations are likely to occur. Importantly, prorenin levels are elevated in pre-eclampsia compared with normal pregnancy, and, in women with Type 1 diabetes, high plasma prorenin associates with an increased risk for the development of pre-eclampsia .
Renin, unlike prorenin, is exclusively derived from the kidneys. Its levels are elevated in pregnancy, most probably as a compensatory mechanism in response to the fall in vascular resistance and blood pressure at the beginning of pregnancy. Renal biopsies taken from patients with pre-eclampsia a few days after pregnancy contained almost no renin-positive juxtaglomerular cells . This is unlikely to be the result of depletion, since, if anything, renin is lower in pre-eclampsia than in a normal pregnancy. Thus it seems that renin synthesis in pre-eclampsia is suppressed. These low(er) renin levels in pre-eclampsia are counterintuitive in the face of the reduced plasma volume in this condition. Potential renin suppressors are the elevated atrial natriuretic peptide levels in pre-eclampsia , the increased sensitivity to AngII and AT1-AAs [agonistic autoantibodies against the AT1 receptor (AngII type 1 receptor)]. The last two interfere with the negative-feedback loop between renin and AngII, since AT1 receptor stimulation suppresses renin release. Finally, VEGF blockade with the angiogenesis inhibitor sunitinib not only increases blood pressure and induces a pre-eclampsia-like syndrome [involving renal dysfunction, proteinuria, glomerular endotheliosis and activation of the ET-1 (endothelin-1) system], but also suppressed renin . Thus VEGF removal by the elevated sFlt-1 levels might be a further explanation for the decrease in renin in pre-eclampsia.
As a result of the increased levels of angiotensinogen and renin, AngII is increased in a normal pregnancy. Not surprisingly, given the lower levels of angiotensinogen and renin, AngII is lower in pre-eclampsia compared with normal pregnancy. Interestingly, historical experiments by Gant et al.  showed that twice as much AngII is needed to get a 20 mmHg increase in blood pressure in pregnant compared with non-pregnant women. Women with pre-eclampsia, on the other hand, do not show this resistance to AngII, which can already be observed as early as week 10 of gestation and thus well before the onset of clinically apparent symptoms. This increased sensitivity was still present 8 months after pregnancy . There are several explanations for this increase in sensitivity. First, the adipose tissue of patients with pre-eclampsia displays elevated AT1 receptor expression. Increased AT1 receptor expression may also be present in other tissues. . Secondly, pre-eclampsia is characterized by an increased heterodimerization of the AT1 receptor with the bradykinin type 2 receptor. Since this prevents AT1 receptor internalization, the AngII response cannot be diminished by removing the AT1 receptors from the membrane . Thirdly, the balance between constrictor AT1 receptors and dilator AT2 receptors (AngII type 2 receptors) receptors may be disturbed . Finally, it has been reported that increased neutrophil infiltration [resulting in excessive ROS production and RhoA (Ras homologue family member A) kinase activation] leads to increased vascular reactivity in pre-eclampsia . Of note, since this conclusion was reached by constructing two AngII concentration–response curves in a row in human omental arteries, without correcting for tachyphylaxis, it needs to be interpreted with caution .
AT1 RECEPTOR AUTOANTIBODIES
In 1999, Wallukat et al.  reported that autoantibodies that stimulate the AT1 receptor are present in women with pre-eclampsia, raising the possibility that pre-eclampsia is an autoimmune disease. AT1-AA levels have been shown to correspond with disease severity . There are multiple ways by which AT1-AA might contribute to the pathogenesis of pre-eclampsia. First, AT1-AA will induce vasoconstriction via stimulation of AT1 receptors on vascular smooth muscle cells. Secondly, AT1-AA–AT1 receptor interaction up-regulates PAI-1 (plasminogen-activator inhibitor-1), sFlt-1 and NADPH oxidase, which are all linked to the pathogeneses of the disease . Indeed, NADPH up-regulation enhances ROS formation , whereas PAI-1 decreases trophoblast invasion, resulting in decreased placental function. The question arises as to why these antibodies are produced in the first place. Initially it was thought that the AT1-AA were cross-reacting against the antigenic region of the parvovirus B19; however, no evidence for this theory could be found subsequently [42,43]. An activation of the immune system seems to play an important role, since treating rats with rituximab prevented the mobilization of B-lymphocytes, thereby suppressing the production of AT1-AA . Furthermore, infusion of inflammatory cytokines into pregnant rats results in the production of AT1-AA . Moreover, the question of why the autoimmune response is directed against the AT1 receptor, and whether there are other vasoactive autoantibodies in pre-eclampsia, remains to be elucidated.
ACE2 AND ANGIOTENSIN-(1–7)
Ang-(1–7) [angiotensin-(1–7)] is an angiotensin metabolite, generated by ACE2 (angiotensin-converting enzyme 2) from AngI (angiotensin I) or AngII (Figure 2), that has received a lot of attention, in part, because binding to its receptor Mas counteracts the AT1-receptor-mediated effects . In parallel with AngII, its levels are elevated during normal pregnancy, but relatively reduced in pre-eclampsia . The simplest explanation is that this is the consequence of a reduced degree of RAAS activity. Interestingly, pregnant ACE2-knockout mice display a higher blood pressure during pregnancy, a reduced gain of weight (indicative of a reduced volume increase) and deliver smaller pups . This may be due to the expected increase in AngII (in the absence of its metabolism by ACE2), a reduction in the levels of Ang-(1–7) or both.
In apparent contrast with these observations, Sykes et al.  showed in early pregnancy that women who developed pre-eclampsia had higher Ang-(1–7) levels than women who remained healthy during pregnancy. Further analysis revealed that this difference was only present in women with female fetuses . In this regard, it should be realized that Ang-(1–7) is also associated with the inhibition of angiogenesis in animal models for tumour growth . Obviously, angiogenesis is an essential aspect of pregnancy, and whether Ang-(1–7) is good or bad in this situation remains to be determined.
Consistent with the changes in renin and AngII, aldosterone is increased during pregnancy and is relatively low in pregnancies complicated by pre-eclampsia. A (too) low aldosterone level might be responsible for the decreased volume expansion in pre-eclampsia, and the resultant poor placental perfusion. Indeed, the proliferation of cultured human trophoblasts increased upon aldosterone stimulation, and aldosterone levels correlated with placental weight in both rats and humans . Furthermore, blocking mineralocorticoid receptors with spironolactone resulted in a decreased umbilical flow in mice , either owing to interference with the local vascular effects of aldosterone  or to a general reduction in volume.
In line with these observations, gain-of-function variants of the CYP11B2 (cytochrome P450, family 11, subfamily B, polypeptide 2; also known as aldosterone synthase) gene reduced the risk of developing pre-eclampsia . Furthermore, an increase in volume transiently decreases blood pressure in pre-eclampsia , and supplemental NaCl exerts similar hypotensive effects in women that display neither the increase in aldosterone production nor the blood pressure fall expected during pregnancy .
The aldosterone/renin ratio is elevated in normal pregnancy. This suggests that additional factors on top of AngII determine aldosterone synthesis. Indeed, a recent study has demonstrated that VEGF stimulates aldosterone production by adrenocortical cells . Furthermore, in rats overexpressing the VEGF-inactivating sFlt-1, capillary rarefaction of the adrenal glands was observed and plasma aldosterone and sFlt-1 levels were inversely correlated . This suggests that VEGF can also stimulate aldosterone production indirectly, i.e., by enhancing adrenal capillary density . These observations suggest that VEGF might be the additional aldosterone-stimulating factor in pregnancy. Simultaneously, they offer an explanation for the lower aldosterone levels in pre-eclampsia.
In line with the above studies, infusion of IgG from women with pre-eclampsia (most probably containing AT1-AA) resulted in hypertension and proteinuria in pregnant mice . Interestingly, in non-pregnant mice, such an IgG injection resulted in an increase in aldosterone (most probably via stimulation of the AT1 receptor), whereas in pregnant mice it led to a decrease in aldosterone production, most probably because infusion of AT1-AA stimulated the production of sFlt-1. As described above, sFlt-1 facilitates the capillary rarefaction of the adrenal glands and VEGF infusion could counteract this phenomenon .
Buhl et al.  reported another mechanism that might explain the decrease in aldosterone levels in pre-eclampsia. They observed that urinary plasminogen is greatly elevated in pre-eclampsia, and correlated negatively with urinary aldosterone. This was due to plasminogen proteolysis and the subsequent stimulation of ENaC (epithelial sodium channel) by plasmin, resulting in sodium retention and aldosterone suppression.
Although a major role for the (relative) lack of aldosterone in the pathogenesis of pre-eclampsia may be present, it has also been suggested that the plasma volume changes precede the changes in aldosterone, both in pre-eclampsia and in patients with fetal growth restriction . Moreover, although aldosterone levels in pre-eclampsia are lower than in normal pregnancy, relative to renin they are actually higher . This implies that aldosterone stimulants other than VEGF and renin come into play, e.g. AT1-AA, the increased sensitivity to AngII and/or ET-1.
CONCLUSIONS AND REMAINING QUESTIONS
A key event in the pathogenesis of pre-eclampsia is the decreased perfusion of the placenta. Most probably this is the result of impaired widening of the maternal spiral arteries . In the second phase underfilling of the maternal vascular system may contribute to the pathogenesis of the disease. At this second stage there is systemic endothelial dysfunction, causing high systemic vascular resistance and an increase in blood pressure. In combination with an increase in vascular permeability and a decrease in oncotic pressure, circulating volume is decreased further, thereby worsening the placental perfusion. In turn, the placenta will produce more vasoactive factors such as sFlt-1, AT1-AA and ROS, causing further vasoconstriction and thereby resulting in a vicious cycle.
In line with this theory is that interrupting this vicious circle by increasing the circulating volume with intravenous fluid replenishment can temporarily lower blood pressure. However, this is unlikely to be effective at the later stages of the disease, because then vascular leakage and hypo-albuminaemia, resulting in a decreased plasma oncotic pressure, reduce the capacity to retain fluid in the intravascular compartment.
The RAAS can interfere in two ways with this vicious circle, via sodium and water retention (via aldosterone and directly via AngII) and via an increase in vascular resistance. The first part might be beneficial, since an increase in volume in view of maintaining tissue perfusion is crucial. The second part might be deleterious since the increase of vascular resistance is part of this vicious circle.
In pregnancy, the RAAS supports the increase in volume by increasing the production of aldosterone, while simultaneously developing a relative resistance to the pressor response of AngII . How the latter occurs is not completely clear. Possibly, high levels of aldosterone counteract AngII-induced vasoconstriction  and/or there is an up-regulation of vasodilator AT2 receptors . Importantly, the increase in aldosterone exceeds that in renin and AngII, and the resulting increase in the aldosterone/renin ratio suggests that other aldosterone stimulants have come into play. Indeed, one such new player is VEGF .
In pre-eclampsia, the RAAS is suppressed and, as we described previously, this is somewhat unexpected in view of the reduced circulating volume. It is hard to say whether the suppression of the RAAS is inappropriate and deleterious or a healthy protection against a further increase in blood pressure. Indeed, transgenic animal models with an excessive RAAS activity show almost all of the features of a pregnancy complicated with pre-eclampsia [22,23]. We propose that in pre-eclampsia the RAAS has lost the delicate balance that in normal pregnancy results in a volume increase and a diminished vasoconstrictor response to AngII. RAAS suppression could be a logical consequence of both the increased pressor response to AngII and the elevated levels of AT1-AA in pre-eclampsia. Clearly, a crucial question is why the pressor response to AngII remains enhanced. A possibility is that either there are no counterbalancing AT2 receptors and/or that the AT2 receptors display a constrictor phenotype in this disorder . When suppressing the RAAS, aldosterone levels should decrease in parallel. Yet in pre-eclampsia the aldosterone/renin ratio is increased even further as compared with normal pregnancy. This is not due to VEGF-dependent aldosterone production, since VEGF activity, if anything, is diminished in pre-eclampsia . It may relate to the increased pressor response to AngII and the elevated levels of AT1-AA, which might keep aldosterone levels, at least relative to renin, elevated, even when circulating AngII levels are low. Furthermore, there may be alternative aldosterone stimulants such as ET-1, as the ET-1 system is activated in pre-eclampsia . Future studies should therefore carefully determine the factors that affect plasma aldosterone (renin, AngII, VEGF, AT1-AA, ET-1 and the sFlt-1/PlGF ratio) to obtain a full understanding of its ‘inappropriate’ level in pre-eclampsia. In addition, given the observation that pregnant mice overexpressing murine angiotensinogen were unable to up-regulate renin expression in the distal nephron (as normally occurs in pregnancy)  and consequently develop hypovolaemia, it might be worthwhile to study the renal RAAS in pre-eclampsia in more detail, e.g. by measuring renin in urine. Evidence indicates that urinary renin levels reflect renal RAAS activity more accurately than plasma renin levels [61,62]. Such information will tell us to what degree changes in the circulating and renal RAAS run in parallel in normal pregnancy and pre-eclampsia. This is particularly relevant in view of our observation that VEGF blockade with sunitinib suppresses renin and simultaneously induces a pre-eclampsia-like syndrome .
angiotensin-converting enzyme 2
- AT1 receptor
AngII type 1 receptor
agonistic autoantibody against AT1 receptor
- AT2 receptor
AngII type 2 receptor
placental growth factor
plasma renin activity
proform of eosinophil major basic protein
reactive oxygen species
soluble fms-like tyrosine kinase-1
vascular endothelial growth factor